Exploring the biotransformation of N-(2-hydroxyphenyl)-2-propylpentanamide (an aryl valproic acid derivative) by CYP2C11, using in silico predictions and in vitro studies

Research Paper

Exploring the biotransformation of

N-(2-hydroxyphenyl)-2-propylpentanamide (an aryl valproic

acid derivative) by CYP2C11, using in silico predictions and

in vitro studies

Jessica Elena Mendieta-Wejebe^a, Arianna Silva-Trujillo^a, Martiniano Bello^b,

Humberto L. Mendoza-Figueroa^b, Norma Lizeth Galindo-Alvarez^b, Arnulfo Albores^c,

d

a

e

ꢀ

Feliciano Tamay-Cach , Martha Cecilia Rosales-Hernandez , Aurelio Romero-Castro and

b

ꢀ

Jose Correa-Basurto

a

b

ꢀ

~

ꢀ

Laboratorio de Biofısica y Biocatalisis, Laboratorio de Diseno y Desarrollo de Nuevos Farmacos e Innovacion Biotecnologica (Laboratory for the

c

ꢀ

Design and Development of New Drugs and Biotechnological Innovation), Escuela Superior de Medicina, Seccion de Toxicologıa, Centro de

d

ꢀ

Investigacion y de Estudios Avanzados, Laboratorio de Investigacion Bioquımica, Escuela Superior de Medicina, Instituto Politecnico Nacional,

e

ꢀ

Ciudad de Mexico and Division de Ciencias de la Salud, Universidad de Quintana Roo, Chetumal, Mexico

Keywords

Abstract

CYP2C11; HO-AAVPA; rat liver microsomes;

valproic acid

Objectives N-(2-hydroxyphenyl)-2-propylpentanamide (HO-AAVPA), a deriva-

tive of valproic acid (VPA), has been proposed as a potential anticancer agent

due to its improved antiproliferative effects in some cancer cell lines. Although

there is evidence that VPA is metabolized by cytochrome P450 2C11 rat isoform,

HO-AAVPA CYP-mediated metabolism has not yet been fully explored. There-

fore, in this work, the biotransformation of HO-AAVPA by CYP2C11 was inves-

tigated.

Correspondence

Jessica Elena Mendieta-Wejebe, Laboratorio

ꢀ

de Biofısica y Biocatalisis, Escuela Superior

ꢀ

de Medicina, Instituto Politecnico Nacional,

ꢀ

Plan de San Luis y Salvador Dıaz Miron s/n,

ꢀ

Casco de Santo Tomas, Ciudad de Mexico

ꢀ

11340, Mexico.

Methods Kinetic parameters and spectral interaction between HO-AAVPA and

CYP were evaluated using rat liver microsomes. The participation of CYP2C11 in

metabolism of HO-AAVPA was conﬁrmed by cimetidine (CIM) inhibition assay.

Docking and molecular dynamics simulations coupled to MMGBSA methods

were used in theoretical study.

E-mail: jesmenwej@yahoo.com

ꢀ

~

Jose Correa Basurto, Laboratorio de Diseno

ꢀ

y Desarrollo de Nuevos Farmacos e

ꢀ

Innovacion Biotecnologica (Laboratory for

the Design and Development of New Drugs

and Biotechnological Innovation), Escuela

Key ﬁndings HO-AAVPA is metabolized by CYP enzymes (K_M= 38.94 µM),

yielding a hydroxylated metabolite according to its HPLC retention time

(5.4 min) and MS analysis (252.2 m/z). In addition, CIM inhibition in rat liver

microsomes (K_i= 59.23 µM) conﬁrmed that CYP2C11 is mainly involved in

HO-AAVPA metabolism. Furthermore, HO-AAVPA interacts with CYP2C11 as

a type I ligand. HO-AAVPA is stabilized at the CYP2C11 ligand recognition site

through a map of interactions similar to other typical CYP2C11 substrates.

Conclusion Therefore, rat liver CYP2C11 isoform is able to metabolize HO-

AAVPA.

ꢀ

Superior de Medicina, Instituto Politecnico

ꢀ

Nacional, Plan de San Luis y Salvador Dıaz

ꢀ

Miron, Ciudad de Mexico 11340, Mexico.

E-mail: corrjose@gmail.com

Received January 6, 2020

Accepted March 21, 2020

doi: 10.1111/jphp.13270

demonstrated that VPA is hepatotoxic due to its reactive

metabolites 4-ene-valproic acid (4-ene-VPA) and 2,4-di-

Introduction

Cancer is a major disease that affects millions of people

worldwide. Current treatments are usually complex, and

some are highly toxic, causing several side effects.^[1]Hence,

the development of therapeutic alternatives with lower tox-

icity has become a priority. Valproic acid (VPA) has been

used for the treatment of some types of cancer, including

haematologic malignancies.^[2]However, some studies have

ene-valproic acid (2,4-diene-VPA), which are enzymatically

generated by cytochrome P-450 (CYP). Moreover, VPA

metabolism induces hepatocyte membrane lysis and ROS

formation.^[3,4]Therefore, N-(2-hydroxyphenyl)-2-propy-

lpentanamide (HO-AAVPA; Scheme 1), an aryl derivative

of VPA, has been proposed from theoretical studies by our

research group.^[5]HO-AAVPA could be a potential

1

Biotransformation of HO-AAVPA by CYP450

Jessica Elena Mendieta-Wejebe et al.

Thus, due to the promising application of HO-AAVPA

in cancer therapy, this study aimed to investigate its meta-

bolism by the CYP2C11 isoform. Experimentally, the HO-

AAVPA metabolic proﬁle was assessed using rat liver

microsomes, and the obtained metabolites were analysed

by liquid chromatography coupled to mass spectrometry

(LC-MS). Interaction modes and binding free energy values

between the HO-AAVPA and CYP2C11 isoforms were eval-

uated by docking and molecular dynamics (MD) simula-

tions with the molecular mechanics generalized-born

surface area (MMGBSA) approach. Finally, two HO-

AAVPA metabolites were synthesized to be compared with

those obtained from rat liver microsomes and those pre-

dicted by the in silico methods.

Scheme 1 Chemical structure of N-(2-hydroxyphenyl)-2-propylpen-

tanamide (HO-AAVPA).

anticancer agent due to its antiproliferative effects in some

cancer cell lines, such as cervical cancer (HeLa), undifferen-

tiated sarcoma (A204) and some breast cancer cells.^[5]In

these assays, it was shown that HO-AAVPA has a higher

antiproliferative effect than VPA (patent: MX 363005 B).^[5]

The enzymatic (metabolic) stability assessment of a new

drug is an essential step in the early stages of its develop-

ment since the optimization of its pharmacokinetic proper-

ties is very important for continuation of the

pharmacological evaluation. Hepatic metabolism is the

main pathway for drug elimination, where cytochrome

(CYP) enzymes, which are located in the hepatocyte endo-

plasmic reticulum, play a major role. The stages of CYP

catalysis inherently involve the binding of the substrate to

the active site, causing a change in the redox potential of

the ferric ion (Fe³⁺) in the haem group, triggering a series

of complex oxidation reactions: hydroxylation, formation

of C–C bonds and cleavage of the corresponding substrate,

among others.^[6]It has been demonstrated that CYP2C9 is

one of the main isoforms involved in drug biotransforma-

tion in humans.^[7]Additionally, it is well known that rat

CYP homolog models can be used for human metabolism

studies.^[8]It has been found that the best homolog of the

human isoform CYP2C9 in rats is CYP2C11.^[9]CYP2C11 is

approximately 50% of the total CYP isoform content in the

rat liver microsomal system and is responsible for the

hydroxylation of testosterone at the 2-a and 16-a posi-

tions,^[10]as well as VPA metabolism.^[9]In addition, CYP

enzymes are susceptible to induction or inhibition by sev-

eral xenobiotics, leading to drug–drug interactions when

they are coadministered.^[11]For example, it has been

demonstrated that low concentrations of dexamethasone

(DEX) induce the expression of CYP2C11,^[12]whereas

cimetidine (CIM) is a speciﬁc CYP2C11 inhibitor.^[13,14]

CYP enzymes bind multiple substrates by two interaction

modes known as type I and type II, which have been exten-

sively studied using ultraviolet-visible (UV-Vis) spec-

trophotometry. Type I ligands are characterized by the

appearance of an absorbance minimum at 420 nm and a

peak at 385–390 nm, whereas the type II ligands show an

absorption peak at approximately 430 nm and a minimum

at 390 nm.^[15]

Materials and Methods

Animals

Male rats of the Sprague-Dawley strain (180 ꢀ 20 g) were

obtained from Harlan Laboratories (Indianapolis, IN,

USA). One week before the experimental assays, animals

were allowed to acclimatize and fed with rat chow and

drinking water ad libitum during this period. Animals were

divided into two groups: Group I (n = 5), without treat-

ment; and Group II (n = 5), treated with an intraperitoneal

(i.p.) injection of DEX (40 mg/kg per day).^[16]At the end

of the experimental, animals were euthanized using pento-

barbital, immediately after which the livers were rapidly

removed and placed on ice for microsomal preparation. All

procedures described in this study were conducted in

accordance with the Mexican Ofﬁcial Standard NOM-062-

ZOO-1999, Technical Speciﬁcations for Production, Care

and Use of Laboratory Animals. Furthermore, the protocol

was approved by the Ethics Committee for the Care and

Use of Laboratory Animals (Approval number: ESM.CIC-

UAL-02/27-07-2015) of the Escuela Superior de Medicina-

IPN (http://www.sepi.esm.ipn.mx/Paginas/CICUAL.aspx).

Reagents and standards

Glucose 6-phosphate (G6P), glucose 6-phosphate dehydro-

genase (G6PDH), magnesium chloride (MgCl₂.6H₂O),

b-nicotinamide adenine dinucleotide phosphate (NADP⁺),

(ꢁ)-nicotine hydrogen tartrate salt, diclofenac sodium salt,

DEX, CIM, estradiol, dimethyl sulfoxide (DMSO) and

acetic acid (C₂H₄O₂) were reagent grade, while the acetoni-

trile (ACN) and the deionized water used were HPLC

grade. All reagents and solvents were purchased from

Sigma-Aldrich (St Louis, Mo, USA). HO-AAVPA was syn-

thesized in our laboratory,^[5]and its purity was >97%, as

conﬁrmed by high-performance liquid chromatography

(HPLC; Agilent inﬁnity 1260) analysis. Its authenticity was

2

Jessica Elena Mendieta-Wejebe et al.

Biotransformation of HO-AAVPA by CYP450

determined using mass spectrometry (MS), nuclear mag-

netic resonance (NMR, ¹H and ¹³C) and Fourier transform

infrared (FTIR) spectroscopies. These analyses were per-

formed at the Centro de Nanociencias y Micro-Nanotec-

USA). The depletion constant (k_dep) was estimated at

different compound concentrations with linear regres-

sion analysis, according to the ﬁrst-order equa-

tion C = C₀(e^ꢁk_dept). The Michaelis–Menten parameters

(V_maxand K_M) were determined based on the non-linear

regression of the plot obtained from the product formation

velocity vs the substrate concentration using the Obach–

Reed–Hagen equation: k_dep = k_dep_([S]?0)[1ꢁ([S]/

[S] + K_M)], where [S] is the substrate concentration. The

V_maxwas calculated based on the equation k_dep_([S]?

ꢀ

nologıa-IPN (https://www.ipn.mx/nanocentro/).

Microsomal preparations

Rat liver microsomes were prepared according to the

method described by Manno et al.^[17]Samples were

stored in 0.1 ^Mphosphate buffer, pH 7.4 with 20% glyc-

erol at ꢁ80°C until use. The protein concentration in the

microsomal preparation was determined with the com-

mercial Bradford kit (Protein Determination Kit, Item

No. 704002, Cayman Chemical). The CYP content was

estimated following the method used by Omura and

Sato, 1964.^[15]

= V_max/K_M.^[21]

0)

Inhibition of microsomal CYP2C11 activity

by cimetidine

The effect of CIM on HO-AAVPA metabolism was evalu-

ated in DEX-induced microsomes at different concentra-

tions of CIM (12.5, 25, 50, 100 and 150 µ^M) dissolved in

ethanol (0.1% ﬁnal volume) and 33 µ^Mof HO-AAVPA.

The activity of the DEX-induced microsomes (CYP) in

these reactions was compared with that of the control

microsomes containing solvent only.

HO-AAVPA metabolism by rat liver

microsomal CYP enzymes

Rat liver microsomes (300 µg of protein in 0.1 M phos-

phate buffer, pH 7.4) for both DEX-induced and non-

induced microsomes were incubated in a 96-well plate at

different concentrations of HO-AAVPA dissolved in ACN

(5, 10, 20, 50, 100 and 250 µ^M). All samples were run in

triplicate. Plates were placed in a water bath at 37°C.

Diclofenac was used as the reference substrate for

CYP2C11.^[18]The reaction was initiated by the addition

of a solution of NADPH regenerating system (1.3 m^M

NADP⁺, 3.3 mM G6P, 0.4 U G6PDH and 3.3 mM

MgCl₂).^[19]With the purpose of excluding potential fac-

tors that could interfere in the reaction, the controls used

were (1) solution buffer only, (2) solution buffer contain-

ing microsomes only and (3) solution buffer containing

the reaction mix without the NADPH regenerating sys-

tem. The reactions were quenched with 0.6 ml of ice-cold

ACN after 0, 5, 15, 25, 35 or 45 min. Samples were vor-

texed and then centrifuged at 6940 g for 10 min in a

Hettich 320 R centrifuge (Hettich, Germany). Subse-

quently, the supernatant was removed and ﬁltered, and a

10 µL aliquot was injected into the HPLC instrument

(Agilent 1260 Inﬁnity Series; Agilent Technologies, Palo

Alto, CA, USA) for immediate analysis. The mobile phase

consisted of a mixture of A) 0.2% acetic acid in water

(v/v) and B) ACN, in a proportion of 40% A and 60%

B. HPLC analysis was performed with isocratic elution

and a ﬂow rate of 0.5 ml/min. The stationary phase was

a Zorbax SB-C18 column (5 µm, 4.6 9 150 mm) at a

temperature of 25°C. UV detection was measured at

242 nm.^[20]

The half-maximal inhibitory concentration (IC₅₀) val-

ues were determined by linear regression analysis of the

logarithm plot obtained from the per cent remaining

enzymatic activity after inhibition as a function of the

CIM concentration. The obtained values were converted

to Ki values using the Cheng–Prusoff equation:^[22]

K_i= IC₅₀/1+[S]/K_Mat different concentrations of CIM,

where K_iis the binding afﬁnity of the inhibitor, IC₅₀is

the functional strength of the inhibitor, [S] is the ﬁxed

substrate concentration, and K_Mis the concentration of

substrate at which the enzymatic activity is half of the

maximum.

HO-AAVPA metabolic proﬁle in rat liver

microsomes

Once the reaction was quenched, microsomal incubations

were processed as described in Section 2.4, and the super-

natants were analysed in triplicate by injection into an

ultrahigh-performance liquid chromatography instrument

coupled to an electrospray ionization mass spectrometer

(UPLC/ESI-MS; Waters Corporation, Milford, MA, USA).

The mobile phase consisted of A) a 0.2% triﬂuoroacetic

acid solution (v/v) at pH 3 and B) ACN 100%, in a propor-

tion of 40% A and 60% B. The stationary phase consisted

of a reversed-phase Zorbax SB-C18 column (5 µm,

4.6 9 150 mm).^[20]Analysis was carried out with an iso-

cratic ﬂux of 0.5 ml/min. The detector used was a Bruker

micrOTOF-Q II (Daltonics, Billerica, MA, USA). Analyses

were performed in positive ion mode, with the nebulizer

set to 5.8 psi and a scan range of 50–3000 m/z.

All graphs were constructed using GraphPad Prism ver-

sion 5.01 software (GraphPad Software, San Diego, CA,

3

Biotransformation of HO-AAVPA by CYP450

Jessica Elena Mendieta-Wejebe et al.

Scheme 2 Method of synthesis of the HO-AAVPA metabolites 3a (2,4) and 3b (2,5). Reaction conditions: (a) 0–5°C for 1 h; room temperature

for 3 h under a N₂atmosphere. (b) DCM anh., room temperature for 12 h. (c) DCM anh., ꢁ15°C for 24 h under a N₂atmosphere. The reaction

was quenched with ice water and neutralized with NaHCO₃.

was added under a nitrogen atmosphere. The mixture was

Substrate titrations

stirred at 0–5°C for 1 h and then allowed to warm to room

All experiments were carried out at room temperature in

matched quartz cuvettes (1 cm 9 1 cm internal) contain-

ing 2 mg/ml protein from the DEX-induced or non-

induced microsomes (v/v) in 0.1 ^Mpotassium phosphate

buffer, pH 7.4 with 20% glycerol.^[23]Analyses were per-

formed in a Lambda 25 UV/VIS spectrophotometer (Perkin

Elmer, Waltham, MA, USA). HO-AAVPA (20 µ^M) was dis-

solved in ACN (0.2% ﬁnal volume). Diclofenac and nico-

tine were used as substrate references, and both were

dissolved in water. To obtain the spectra, a baseline was

recorded with the microsomal solution added to both cuv-

ettes (sample and reference) followed by the titration of the

substrate into the sample cuvette and the titration of buffer

into the reference cuvette at the same volume. Spectra were

identiﬁed as type I (peak: 380–390 nm, trough: 415–

420 nm) or type II (peak: 421–435 nm, trough: 390–

410 nm) ligands.^[24]

temperature for 3 h.^[5]After the reaction time had elapsed,

the reaction mixture was neutralized with 1 ml of a 40%

trimethylamine (TEA) solution in anhydrous dichloro-

methane (DCM anh.). The mixture was again cooled to

0–5°C, maintaining the nitrogen atmosphere. Then, 3.0 eq

of the corresponding dimethoxyaniline dissolved in DCM

anh. was added slowly, and the reaction mixture was stirred

for 12 h at room temperature. The course of the reaction

was monitored by thin-layer chromatography (TLC) until

the maximum raw material consumption was observed.

Finally, the reaction was quenched by the addition of 10 ml

of water, and the reaction mixture was extracted with DCM

anh. (3 9 15 ml). The organic phase was washed with

brine (1 9 20 ml) and dried over anhydrous sodium sul-

fate (Na₂SO₄anh.). The solid was ﬁltered under reduced

pressure. The resulting crude product was puriﬁed by col-

umn chromatography using silica gel and a mixture of 95/5

hexane/ethyl acetate (EtOAc) (v/v) to give 2 as a white

solid.

In silico prediction and chemical synthesis

The chemical purity of intermediates 2a and 2b was anal-

ysed by HPLC with the following method: the mobile phase

consisted of a mixture of A) 40% ACN and B) 60% H₂O.

HPLC analysis was performed with isocratic elution and a

ﬂow rate of 0.5 ml/min. The stationary phase was a Zorbax

SB-C18 column (5 µm, 4.6 9 150 mm) at a temperature

of 25°C. UV detection was measured at 242 nm.

of HO-AAVPA metabolites

Before performing the synthesis (Scheme 2) of the HO-

AAVPA metabolites, HO-AAVPA was submitted for the

bioinformatic prediction of metabolic products by using

the SMARTCyp web server (https://smartcyp.sund.ku.dk/

mol_to_som). The most likely metabolites obtained from

CYP metabolism proved to be 2,4-dihydroxyphenyl-2-

propylpentanamide (3a) and 2,5-dihydroxyphenyl-2-

propylpentanamide (3b). Once these results were deter-

mined, the chemical synthesis was carried out as described

below to obtain standards of these two compounds.

General procedure for the synthesis of dihydroxy-

lated standards (3a, 3b)

The metabolites 3a and 3b were synthesized according to

the general procedure in Scheme 2. Dihydroxylated deriva-

tives were obtained by the deprotection of the methoxy

groups of the respective intermediates (2a and 2b).^[25]

Compounds 2a and 2b (1.0 eq) were dissolved in DCM

anh. and cooled to ꢁ15°C, and then, boron tribromide

(BBr₃, 10.0 eq) dissolved in DCM anh. was added dropwise

General procedure for the synthesis of methoxy-

lated intermediates (2a, 2b)

Valproic acid 1 (1.0 g, 7.50 mmol) was cooled to 0–5°C

and then 1.5 eq of oxalyl chloride (0.96 ml, 11.25 mmol)

4

Jessica Elena Mendieta-Wejebe et al.

Biotransformation of HO-AAVPA by CYP450

under a nitrogen atmosphere; the mixture stirred at ꢁ15°C

for 24 h. Finally, to quench the reaction, the mixture was

poured over ice water (50 ml) and stirred for 15 min. The

aqueous layer was neutralized with a saturated solution of

sodium bicarbonate (NaHCO₃) at pH 7 and then extracted

with DCM anh. (3 9 20 ml). The organic extracts were

washed with brine and dried over Na₂SO₄anh. The solid

was ﬁltered off, and the ﬁltrate was concentrated under

reduced pressure. The products were puriﬁed by recrystal-

lization using a mixture of DCM/hexane (5 : 1).

to the membrane, some parts of the CYP structure in addi-

tion to the N-terminal membrane anchor are immersed in

the POPC membrane, such as the F-G loop. These posi-

tions are in accordance with experimental reports for other

CYPs.^[27,28]

Ligand minimization

The 3D ligand minimum energies were achieved by means

of Gaussian 09 software at the AM1 level.^[29]The mini-

mized structure was used for the docking assays.

HPLC identiﬁcation of 3a and 3b

Docking procedure

HO-AAVPA and its previously synthesized metabolites 3a

and 3b were analysed by the HPLC method previously

reported by Silva-Trujillo et al.^[20]to verify their retention

times and for comparison to those obtained from the rat

hepatic microsomes treated with HO-AAVPA.

All possible rotatable bonds and partial atomic charges

(Gasteiger–Marsili formalism) of the ligands, as well as the

Kollman charges for all enzyme atoms of the enzyme, were

assigned using the AutoDock tool 1.5.4.^[30]Afterward, the

ligands were docked into CYP2C11 using AutoDock 4.2.0

with the hybrid Lamarckian genetic algorithm as the search

method, with an initial population of 100 randomly placed

individuals and a maximum energy of evaluation of

1.0 9 10⁷. All docking simulations were performed using a

Computational analyses

Protein model building

ꢁ³

The amino acid sequence of the CYP2C11 rat isoform was

retrieved from the Swiss-Prot entry code P08683. This

sequence was submitted to BLAST analysis to identify PDB

structures with the highest sequence identity. Subsequently,

the structures obtained were submitted to STRAP software

to conﬁrm their sequence homology with CYP2C11. Three

tridimensional (3D) structures of human CYP2C9 were

available from the PDB, one apo form (PDB ID: 1OG2)

and two ligand-bound complexes (PDB ID: 1OG5 and

1R9O). The crystal structure with the highest resolution

(1R9O) was selected as a template for building the model

of CYP2C11.

grid box on all proteins (120 A ) centred on the haem

ꢁ³

group with grid points separated by 0.375 A . All protein

visualizations were performed with PyMOL DeLano WL

(http://www.pymol.org).

Molecular dynamics (MD) simulations

Anchoring the complexes into a POPC membrane

The CYP2C11/HO-AAVPA complexes were oriented with

respect to the lipid membrane using the OPM (Orienta-

tions of Proteins in Membranes) server (http://opm.

phar.umich.edu/). A rectangular pre-equilibrated POPC

phospholipid bilayer was generated for each system

using the membrane-builder tool of CHARMM-GUI^[31]

Homology modelling

ꢁ

(111.416 9 111.416 9 125.962 A, XYZ). To place the

A homology model of CYP2C11 was constructed using the

I-TASSER hierarchical protein structure modelling

approach based on the secondary structure enhanced pro-

ﬁle–proﬁle threading alignment (PPA) and the iterative

implementation of the Threading ASSEmbly Reﬁnement

(TASSER) program. After model construction, the haem

group was anchored to the catalytic site using a standard

molecular docking protocol,^[9]establishing molecular

CYP2C11/HO-AAVPA complex into the bilayer, the

replacement method was used.^[32,33]The membrane con-

sisted of 338 POPC phospholipids with 161 on the top and

ꢁ

177 on the bottom. The size along the z-axis (125.962 A)

was determined by specifying the thickness of bulk water

from the protein extent along Z (Z centre = 24.356). The

membrane–receptor complexes obtained were thus solvated

and neutralized using the solvation and autoionized mod-

ules (ion placing method) of the charm-gui membrane

builder.^[31]The ionic strength was maintained at 0.15 M

NaCl using the TIP3 water model. The all-atom models

of each system were generated by using the CHARMM

force-ﬁeld parameters (http://mackerell.umaryland.edu/

CHARMM_ff_params.html). Then, these ﬁles were

recognition in proximity to the atoms Fe and C⁴³⁵

.

The transmembrane segments missing from the

CYP2C11 homology model were modelled as a-helices and

joined to the rest of the protein by a ﬂexible arm employing

the Modeller program.^[26]Then, the transmembrane helix

was rotated within the membrane to be perpendicular to

the membrane surface. In this CYP position, with respect

5

Biotransformation of HO-AAVPA by CYP450

Jessica Elena Mendieta-Wejebe et al.

converted into ﬁles that conformed to the Lipid11 naming

convention using the charmmlipid2amber.x script

included in the AMBER 12 package.^[34]Lipid11 is the

Amber lipid force ﬁeld that includes parameters from the

general Amber force ﬁeld (GAFF).^[35,36]Starting from the

CYP2C11/ligand complex obtained through docking pro-

cedures, the topologies to perform MD simulations were

constructed using the Leap module with the ff99SB and

Lipid11 force ﬁelds.^[37,38]The ligand topologies were built

with the antechamber module, based on the generalized

AMBER force ﬁeld (GAFF)^[36]followed by ab initio opti-

mization at the AM1-BCC level.^[37]The missing force-ﬁeld

parameters for the haem (all-atom) group were taken

from the AMBER parameter database at http://www.phar

macy.manchester.ac.uk/bryce/amber (Accessed March 13,

2013). Then, these structures were minimized and heated

with sander following three steps. First, the systems were

submitted for energy minimization during 10 000 steps

Analysis of MD trajectories

The ptraj module of Amber 12 was used to analyse the root-

mean-square deviation (RMSD) between structure pairs.

The area per lipid was calculated using the cpptraj module of

Amber 12 by dividing the area of our simulation box [(box

X length) 9 (box Y length)/(number of phospholipids per

layer)]. Before the RMSD calculations, the overall transla-

tional and rotational motions were removed, and the RMSD

was calculated for all heavy atoms. Average conformations

were obtained through a cluster analysis using the kclust

algorithm that belongs to the MMTSB toolset (http://

mmtsb.scripps.edu/software/mmtsbtoolset.html).

Then,

from the most populated cluster, the conformation with the

lowest RMSD to the cluster centre (centroid) was selected.

Images were prepared using PyMOL v0.99.

Calculation of absolute binding free energies

under the NVT ensemble, with position restraints (the

force constant was set to 500 kcal mol^ꢁ1

ꢁ^ꢁ2

Calculations of the absolute binding free energies were per-

formed using the MMGBSA method^[43–45]following the

single-trajectory approach. By using this methodology, the

molecular structures of the binding partners are taken from

the CYP2C11/HO-AAVPA complexes generated during the

MD simulation. Although the energetic contributions due

to the conformational changes are ignored, this approach

leads to a strong decrease in the statistical uncertainty of

the free energy components.^[43]

A ) on both

the protein and lipids to allow for the relaxation of the

water molecules. After this equilibration step, the system

was slowly heated in the NPT ensemble through two

sequential runs from 0 to 300 K while keeping the protein

atoms and lipids restrained (the force constant was set to

10 kcal mol^ꢁ1

ꢁ^ꢁ2

A ). First, the system was heated from 0

to 100 K using the Langevin thermostat, and the second

phase of heating consisted of slowly increasing the temper-

ature to the production temperature (300 K). The posi-

tions and velocities were read from the previous

temperature heating, and an anisotropic Berendsen weak-

coupling barostat^[38]was used to equilibrate the pressure,

whereas the Langevin thermostat was used to equilibrate

the temperature.

Before the calculations were performed, all counter ions

and water molecules were stripped from the snapshots for

each system, and 800 snapshots at time intervals of 100 ps

from the last 80 ns of the simulations runs were taken,

using a salt concentration of 0.1 ^Mand the Born implicit

solvent model of 2 (igb = 2). The analyses were performed

using the MMPBSA perl script^[44]provided in the Amber

12 package.^[34]The binding free energy (DG_bind) of each

The MD simulations were performed with pmemd.cuda

AMBER 12 executable,^[34]which allowed the acceleration

of explicit solvent particle mesh Ewald (PME) calcula-

tions^[39]through the use of GPUs.^[40,41]MD simulations of

150 ns in length were run with the input ﬁles without posi-

tion restraints under periodic boundary conditions (PBC)

and using an NPT ensemble at 300 K. The electrostatic

interactions were calculated via the particle mesh Ewald

complex can be calculated as DG_bind= G^complexꢁG^receptor

ꢁ

G^ligand= DE_MM+ DG_GB+ DG_SAꢁ TDS, where DE_MMis

the gas-phase interaction energy between the receptor and

the ligand, which includes the van der Waals (DE_vdw) and

the electrostatic (DE_ele) interaction energies. DG_GBand

DG_SA, respectively, are the electrostatic and non-polar con-

tributions to desolvation upon ligand binding, and ꢁTDS

is the entropy contribution resulting from the change in

the degrees of freedom of the solute molecules, which were

considered here; therefore, our values reported for the

MMGBSA calculations should be named absolute binding

free energies.

method,^[39]and a 10 A cut-off was used for the van der

ꢁ

Waals interactions. The bonds between the heavy atoms

and hydrogens were constrained with the SHAKE algo-

rithm.^[42]The temperature was controlled using Langevin

dynamics while the pressure was controlled using a semi-

isotropic constant surface tension to maintain a speciﬁc

area per lipid. The collision frequency was 1.0 ps^ꢁ1. The

pressure was maintained at 1 bar, and the pressure

coupling constant was set to 1 ps. The time step of the MD

simulations was set to 2.0 fs, and the coordinates were

saved for analysis every 1 ps.

Entropic contributions were evaluated using the

MMPBSA.py module implemented in Amber Tools 1.5,^[34]

and only 40 snapshots taken at time intervals of 2 ns were

chosen due to the large memory demand for this calcula-

tion.

6

Jessica Elena Mendieta-Wejebe et al.

Biotransformation of HO-AAVPA by CYP450

HO-AAVPA in the presence of either non-induced (Fig-

ure 1b) or DEX-induced microsomes (Figure 1d). For the

DEX-induced microsome incubations, only one signal

appeared at 5.3 min, in addition to the respective controls

(Figure 1a and 1c). These ﬁndings demonstrate that pre-

treatment of rats with DEX resulted in a 1.7-fold enhance-

ment in HO-AAVPA metabolic activity in the liver. CYP

activity was determined by quantiﬁcation of the HO-

AAVPA consumption (Figure 2a) and M1 production (Fig-

ure 2b). Screening for metabolic activity from 5 to 250 µ^M

HO-AAVPA allowed for determination of the K_Mand V_max

Statistical analysis

HO-AAVPA consumption rate and M1 formation data

were processed using the non-linear regression method,

and the effect of different concentrations on consumption

rate was analysed by the one-way ANOVA method and the

post hoc Tukey’s multiple comparison test. Data of inhibi-

tion of HO-AAVPA metabolism and the formation of M1

and M2 by CIM were analysed by the two-way ANOVA test

method. The interaction of HO-AAVPA with rat hepatic

non-induced and DEX-induced microsomes, measured as

the CYP optical difference spectra, was analysed by paired

Student’s t-test. In all cases, a value of P < 0.05 was consid-

ered to establish a statistically signiﬁcant difference. All

analyses and the corresponding graphs were performed in

GraphPad 5.01.

values,

which

were

26.43 ꢀ 1.75 µM

and

0.126 ꢀ 0.006 µ^M/min/nmol CYP, respectively (Table 1).

Effect of CIM inhibition on CYP-mediated

metabolism of HO-AAVPA

The presence of CIM in the rat liver microsomes incuba-

tions decreased the consumption of HO-AAVPA through

CYP2C11 inhibition. As shown in Figure 2c, when the

maximum concentration of CIM (150 µ^M) was used, the

consumption of HO-AAVPA was inhibited in rat liver

microsomes by 67.05% ꢀ 24.87% (P > 0.05), while the

production of M1 and M2 decreased by 52.07% ꢀ 9.89%

Results

Metabolic activity of microsomal CYP

enzymes on HO-AAVPA

Figure 1 shows the chromatograms of HO-AAVPA^[20]as

well as those corresponding to the incubation of

Figure 1 Representative UV-HPLC chromatograms of the HO-AAVPA metabolism by rat liver microsomes. (a) Control in the absence of NADPH

in non-induced microsomes. (b) Full reaction in non-induced microsomes. (c) Control in the absence of NADPH in DEX-induced microsomes. (d)

Full reaction in DEX-induced microsomes.

7

Biotransformation of HO-AAVPA by CYP450

Jessica Elena Mendieta-Wejebe et al.

Table 1

K

_M, V_maxand CL_intvalues determined using product forma-

tion measurements and substrate depletion measurements for rat

hepatic microsomes induced with DEX

Substrate depletion

Product formation

Parameter

(Mean ꢀ SD)

K_M(µM)

26.43 ꢀ 1.75

0.126 ꢀ 0.006

38.94 ꢀ 1.69

V_{max (}µM/min/

nmol CYP)

CL_int(µl/min/mg

protein)

0.0976 ꢀ 0.003

108.53 ꢀ 12.55

59.05 ꢀ 1.00

Metabolic proﬁle study in rat liver

microsomes

The chromatogram of non-induced microsomes incubated

with HO-AAVPA (Figure 1b) showed the same retention

time for HO-AAVPA at 10.1 min (Figure 3a). In addition,

two new peaks at 5.3 and 7.4 min were detected, suggesting

the presence of two metabolites (M1 and M2, respectively)

(Figure 1b and 1d). However, due to the low quantity of

M2, only metabolite M1 was further analysed by MS (Fig-

ure 4). For M1, one peak at 252.2 m/z was detected, which

corresponds to the hydroxylation of HO-AAVPA (236.2 m/

z; Figure 4). To support our hypothesis that M1 could be a

product of HO-AAVPA hydroxylation as predicted with the

SMARTCyp platform (Figure 5), we synthesized the corre-

sponding standards 3a and 3b as described in Section 2.8 to

be further characterized (Figures S1–S5). These compounds,

3a and 3b, show approximately the same HPLC retention

time (ﬃ5.4 min, Figure 3), and the MS molecular ions cor-

respond to 252.1595 and 252.1596 m/z, respectively (Fig-

ures 4 and S5).

Substrate titrations

Figure 2 (a) Depletion curve of HO-AAVPA in the range of

5–250 µ^Min DEX-induced microsomes and 30 min of incubation time.

Each point represents the mean of three independent replicates

(n = 3). P < 0.05: *[5] vs [40], [100] and [250]; ♦[10] vs [40], [100]

and [250]; °[20] vs [40], [100] and [250]; [40] vs [250]. (b) Michaelis–

Menten plot for M1 production. Reactions were performed with

5–250 µ^MHO-AAVPA in the presence of 0.3 mg/ml DEX-induced

microsomes and 30 min of incubation time. Each point is the mean of

three independent replicates (n = 3). (c) Plot of the inhibition of HO-

AAVPA consumption and formation of M1 and M2 by increasing the

concentration of CIM. Results are given as the mean values of ﬁve

independent replicates (n = 5). P = 0.0365 (among groups) and

P < 0.0001 (among concentrations): *12.5 µ^M(HO-AAVPA vs M1 and

M2); **25 µ^M(HO-AAVPA vs M1 and M2).

The UV spectra were used to determine the absorbance dif-

ferential (DA) exhibited by HO-AAVPA consumption, as

well as its binding mode with CYP enzymes present in rat

liver microsomes. In Table 2, it can be observed that the

DA value is higher in DEX-induced microsomes than that

in the non-induced microsomes. Furthermore, the peak

values (Figure 6 and Table 2) show that HO-AAVPA exhi-

bits a type I ligand binding mode with the microsomal CYP

enzymes.

Chemical synthesis of standards 3a and 3b

and 53.68% ꢀ 4.95% (P > 0.05), respectively. These

results revealed that CIM was able to decrease the forma-

tion of M1 and M2, suggesting involvement from CYP2C11

in the biotransformation of HO-AAVPA to yield these

metabolites.

Intermediate 2,4-dimethoxyaniline-2-propylpen-

tanamide (2a)

For 2a preparation, the general procedure was followed

using 2,4-dimethoxyaniline (2.8 g, 18.46 mmol) dissolved

8

Jessica Elena Mendieta-Wejebe et al.

Biotransformation of HO-AAVPA by CYP450

Figure 3 Liquid chromatograms of (a) HO-AAVPA, (b) metabolite 3a and (c) metabolite 3b.

Figure 4 HO-AAVPA submitted for LC-MS study after metabolism in rat liver microsomes.

in 14.0 ml of DCM anh. After puriﬁcation by column chro-

matography, the solid obtained was recrystallized from a

dimethoxyaniline-2-propylpentanamide (2a) as white crys-

tals (0.89 g, 48.0% yield), with a R_fof 0.46 (hexane/EtOAc,

8 : 2) and a melting point of 106–107°C. FTIR (ATR)

DCM/hexane

mixture

(1 : 5),

obtaining

2,4-

9

Biotransformation of HO-AAVPA by CYP450

Jessica Elena Mendieta-Wejebe et al.

1591, 1458, 1427, 1278, 1212, 1123, 1054, 866, 787,

1

718 cm^ꢁ1. H NMR (750 MHz, DMSO-d₆) d: 8.95 (s, 1H,

NH), 7.67 (d, J = 3.75 Hz, 1H, H-6), 6.93 (d, J = 9.00 Hz,

1H, H-3), 6.62 (dd, J = 9.00, 3.00 Hz, 1H, H-4), 3.76 (s,

3H, OMe), 3.67 (s, 3H, OMe), 2.61-2.58 (m, 1H, H-8),

1.53-1.48 (m, 4H, H-9, H-9⁰), 1.33-1.23 (m, 4H, H-10, H-

10⁰) 0.86 (t, J = 7.50 Hz, 6H, H-11, H11⁰) (Figure S2A).

¹³C NMR (188.5 MHz, DMSO-d₆) d: 175.00 (C-7), 153.37

(C-5), 144.22 (C-2), 128.67 (C-1), 112.40 (C-6, 109.01 (C-

4), 108.55 (C-3), 56.76 (OMe), 55.79 (OMe), 46.08 (C-8),

35.44 (C-9, C-9⁰), 20.61 (C-10, C10⁰), 14.52 (C-11, C-11⁰)

(Figure S2B). ESI-MS for C₁₆H₂₅NO₃. Calculated

[M + H]: 279.18. Found [M + H]+: 280.1909 (Fig-

ure S5B). HPLC purity: 98.4%. Retention time: 4.4 min.

Figure 5 Predicted HO-AAVPA sites for metabolism of by the

CYP2C9 isoform using SmartCyp. Red colour (1) means high possibil-

ity to be chemically modiﬁed from CYPs. Orange colour means 50%

to be chemically modiﬁed from CYPs. Blue colour means lowest possi-

bility (0) to be chemically modiﬁed from CYPs.

Metabolite 2,4-dihydroxyphenyl-2-propylpen-

tanamide (3a)

mmax: 3289, 2929, 1655, 1533, 1452, 1412, 1281, 1211,

1159, 1123, 1035, 825, 685 cm^ꢁ^{1 1}H NMR (750 MHz,

.

DMSO-d₆) d: 8.86 (s, 1H, NH), 7.57 (d, J = 8.25 Hz, 1H,

H-6), 6.60 (d, J = 2.25 Hz, 1H, H-3), 6.47 (dd, J = 6.75,

2.25 Hz, 1H, H-5), 3.79 (s, 3H, OMe), 3.74 (s, 3H, OMe),

2.51-2.47 (m, 1H, H-8), 1.55-1.49 (m, 4H, H-9, H-9⁰),

1.33-1.26 (m, 4H, H-10, H-10⁰) 0.88 (t, J = 6.75 Hz, 6H,

H-11, H11⁰) (Figure S1A). ¹³C NMR (188.5 MHz, DMSO-

d₆) d: 174.58 (C-7), 157.45 (C-4), 152.56 (C-2), 125.16 (C-

6), 120.77 (C-1), 104.43 (C-5), 99.27 (C-3), 56.15 (OMe),

55.73 (OMe), 45.85 (C-8), 35.50 (C-9, C-9⁰), 20.62 (C-10,

C10⁰), 14.53 (C-11, C-11⁰) (Figure S1B). ESI-MS for

C₁₆H₂₅NO₃. Calculated [M + H]: 279.18. Found

[M + H]+: 280.1930 (Figure S5A). HPLC purity: 99.6%.

Retention time: 5.4 min.

To a solution of 2a (0.5 g, 3.25 mmol) in DCM anh.

(5.0 ml), BBr3 (1.7 ml, 17 mmol) dissolved in anh. DMC

(2.0 ml). After recrystallization, 3a was obtained as white

crystals (0.201 g, 25% yield), with a R_fof 0.45 [5% TEA-

hexane, DCM/methanol (MeOH), 1 : 1] and a melting

point of 89-90 °C. FTIR (ATR) mmax: 3361, 3247, 2928,

2870, 1598, 1510, 1164, 1108, 975, 795 cm^ꢁ^{1 1}H NMR

.

(750 MHz, DMSO-d₆) d: 9.30 (s, 1H, OH-2), 8.92 (s, 1H,

NH), 8.78 (s, 1H, OH-4), 7.12 (d, J = 2.9 Hz, 1H, H-6),

6.66 (d, J = 8.6 Hz, 1H, H-5), 6.39 (dd, J = 8.6, 2.9 Hz,

1H, H-3), 2.56 (dd, J = 9.6, 4.8 Hz, 1H, H-8), 1.53 (tt,

J = 13.2, 9.4, 5.6 Hz, 4H, H-9, H-9⁰), 1.38–1.21 (m, 4H, H-

10,H-10⁰), 0.88 (t, J = 7.4 Hz, 6H, H-11, H-11⁰) (Fig-

ure S3A). ¹³C NMR (188.5 MHz, DMSO-d₆) d: 175.516

(C-7), 155.85 (C-4), 150.47 (C-2), 124.71 (C-1), 118.50 (C-

5), 106.50 (C-6), 104.19 (C-3), 45.80 (C-8), 35.44 (C-9, C-

9⁰), 20.66 (C-10, C-10⁰), 14.48 (C-11, C-11⁰) (Figure S3B).

ESI-MS for C₁₄H₂₁NO₃. Calculated [M + H]: 251.3260.

Found [M + H]+: 252.1595 (Figure S5C). HPLC purity:

99.0%. Retention time: 5.4 min.

Intermediate 2,5-dimethoxyaniline-2-propylpen-

tanamide (2b)

For 2b preparation, the general procedure was followed

using 2,5-dimethoxyaniline (2.8 g, 18.46 mmol) dissolved

in 14.0 ml of DCM anh. After puriﬁcation by column chro-

matography, the solid obtained was recrystallized from a

DCM/hexane mixture (5 : 1), obtaining 2,5-dimethoxyani-

line-2-propylpentanamide as white crystals (0.91 g, 49.0%

yield), with a R_fof 0.53 (hexane/EtOAc, 8 : 2) and a melt-

ing point of 70–71°C. FTIR (ATR) mmax: 3259, 2952, 1652,

Metabolite 2,5-dihydroxyphenyl-2-propylpen-

tanamide (3b)

To a solution of 2b (0.5 g, 3.25 mmol) in DCM anh.

(5 ml), BBr3 (1.7 ml, 17 mmol) dissolved in anh. DMC

(2 ml). After recrystallization, 3b was obtained as white

crystals (0.315 g, 39% yield), with a R_fof 0.51 (5% TEA/

hexane, DCM/MeOH, 1 : 1) and a melting point of 109–

111°C. FT IR (ATR) mmax: 3285, 3126, 2962, 2875, 1647,

1539, 1450, 1187, 971, 872, 736 cm^ꢁ1. ¹H NMR (750 MHz,

DMSO-d6) d: 9.29 (s, 1H, OH-2), 8.90 (s, 1H, NH), 8.76

(s, 1H, OH-5), 7.11 (d, J = 3.0 Hz, 1H, H-6), 6.65 (d,

J = 8.7 Hz, 1H, H-3), 6.38 (dd, J = 8.5, 3.0 Hz, 1H, H-4),

2.55 (dd, J = 9.6, 4.9 Hz, 1H, H-8), 1.56–1.48 (m, 4H,

Table 2 Rat hepatic microsomal CYP optical difference spectra in

the presence of HO-AAVPA (n = 5)

Wavelength (nm,

Mean ꢀ SD)

Microsomes

DA (Mean ꢀ SD)

Trough

Peak

Without treatment

DEX-induced

0.0022 ꢀ 0.0007

0.0040 ꢀ 0.0019

418 ꢀ 3.9

418 ꢀ 3.8

392 ꢀ 1.7

392 ꢀ 8.5

10

Jessica Elena Mendieta-Wejebe et al.

Biotransformation of HO-AAVPA by CYP450

Figure 6 Absolute spectra of rat microsomal CYP in 20% glycerol, 100 mM sodium phosphate, pH 7.4 and HO-AAVPA (20 µM) at room temper-

ature. Red lines show the basal line, and blue lines show the spectrum after HO-AAVPA addition in (a) untreated microsomes and (b) DEX-induced

microsomes.

H-10), 1.33 (tt, J = 12.8, 10.3, 5.5 Hz, 4H, H-9), 1.31–1.21

(m, 4H, H-10), 0.92–0.84 (m, 6H, H-11) (Figure S4A). ¹³

elements: helices I, F, and G; loop F-G; and loop B-C. HO-

AAVPA bound to the CYP2C11 substrate binding site

adopted a conformation in which most of the contacts

C

NMR (188.5 MHz, DMSO) d: 175.44 (C-7), 150.31 (C-5),

140.88 (C-2), 127.10 (C1), 117.33 (C-3), 111.83 (C-4),

109.62 (C-6), 46.09 (C-8), 35.44 (C-9, C-9⁰), 20.64 (C-10,

C-10⁰), 14.46 (C-11, C-11⁰) (Figure S4B). ESI-MS for

C₁₄H₂₁NO₃. Calculated [M + H]: 251.3260. Found

[M + H]+: 252.1596 (Figure S5D). HPLC purity: 99.0%.

Retention time: 5.3 min.

ꢁ

(residues within a distance of 4 A) are predominantly

hydrophobic (Figure 7b), as has been observed in some

modelling and crystallographic CYP complexes.^[9,47]In this

ligand conformation, the hydroxyphenyl ring of HO-

AAVPA (C-3 atom) is placed in close proximity to the

haem group and iron atom (Fe) (Figure 7b), with an aver-

ꢁ

age distance of approximately 4.7 A and an angle of 37.2°,

ꢁ

whereas C-4 is located 6.05 A from the haem iron atom.

Molecular docking

These atomic distances are in agreement with the hydroxy-

lation of steroid molecules^[48]and VPA.^[9]The calculated

docking energy was ꢁ6.77 kcal/mol, and the theoretical K_d

was 10.9 µM. This binding pose suggests that the active

high-valent iron-oxo could attack both C-3 and C-4 during

the oxidation reaction since these two carbons are close to

the oxo atom.

CYP2C9 exhibits high regioselectivity and catalytic efﬁ-

ciency for the oxidation of diclofenac.^[46]Since CYP2C11

and CYP2C9 share high-sequence identity (77%), it could

be expected that HO-AAVPA binds to CYP2C11. Figure 7a

depicts that HO-AAVPA reached the substrate binding site,

which is formed by the following secondary structure

11

Biotransformation of HO-AAVPA by CYP450

Jessica Elena Mendieta-Wejebe et al.

Figure 7 (a) Secondary structure elements forming the substrate binding site of the CYP2C11-HO-AAVPA complex. (b) Residues at a distance

ꢁ

≤4 A stabilizing the CYP2C11-HO-AAVPA complex predicted through docking calculations. The protein is shown as the green cartoon representa-

tion, the ligand is shown as the yellow ball-and-stick representation, the haem group is represented as the white ball-and-stick representation,

and the residues that stabilize the complex are shown as the green stick representation.

was conserved with some small conformational differences.

MD simulations

For the CYP2C11-HO-AAVPA complex, the cluster analysis

yielded three complex conformations, representing 31.60,

Convergence and equilibrium examination

23.70 and 15.6% of the entire present conformations, corre-

To verify the robustness and stability of the CYP2C11/HO-

AAVPA complex, long MD simulations were performed for

complexes anchored into a POPC membrane (Figure 8a)

considering the physiological environment for this sys-

tem.^[49,50]To evaluate whether the systems reached the equi-

librium stages, some structural properties during the MD

simulation were monitored, such as the root-mean-square

deviation (RMSD) of the backbone atoms with respect to the

initial structure and the area per lipid. Figure 8b shows that

the CYP2C11/HO-AAVPA complex exhibits a very high area

per lipid value at the beginning of the simulation and then

decreases towards a converged area per lipid on the order of

60–70 ns, obtaining an area per lipid value of 56.57 ꢀ 0.70

for CYP2C11/HO-AAVPA. These ﬁndings are in good agree-

ment with those found for other protein-POPC-membrane

systems.^[50,51]Similarly, RMSD analysis shows that the

CYP2C11/HO-AAVPA complex reached convergence on the

order of 60-70 ns (Figure 8c). Therefore, all structural analy-

ses and free energy calculations were performed on the last

80 ns of the 150 ns of the MD simulations.

sponding to the ﬁrst, second and third most populated con-

formations, respectively. Figure 9 shows the three most

populated complex conformations present during the last

80 ns of the MD simulation for CYP2C11/HO-AAVPA. Fig-

ure 9a shows that the ligand conformations are roughly sim-

ilar for the three HO-AAVPA conformations, with RMSD

ꢁ

values that oscillate between 0.28 and 1.6 A. However, some

differences can be observed when a more detailed analysis is

ꢁ

performed. Figure 9b shows the residues at a distance of 4 A

stabilizing the ﬁrst most populated ligand conformation,

where the ligand is stabilized through hydrophobic contacts,

with either the side chain or backbone atoms of V¹⁰², I¹¹³

,

N

²⁰⁴, F²⁰¹, F²⁰⁵, F²⁰⁸, E³⁰⁰, T³⁰¹, L³⁶¹and L³⁶⁶, positioning

the C-3 atom of the hydroxyphenyl ring of HO-AAVPA in

close proximity to the iron (Fe) atom of the haem group,

ꢁ

with an average distance of approximately 4.54 A, whereas

ꢁ

C-4 is located 5.27 A from the Fe atom of the haem. These

atomic distances are in agreement with the hydroxylation of

steroid molecules and are similar to those generated by the

docking procedures.^[10]The second most populated confor-

mation was stabilized by similar hydrophobic residues as

those observed for the ﬁrst conformation (Figure 9c). How-

ever, some interactions with other residues were present that

were not observed for the ﬁrst conformation (F¹¹⁴and V³⁶²).

In this ligand conformation, the hydroxyphenyl ring was also

close to the haem group with the C-5 atom in the closest

proximity to the Fe atom, with an average distance of

Interactional analysis and average properties

The detailed map of interactions was explored under physio-

logical environmental conditions generated through MD

simulations that incorporate water molecules, protein ﬂexi-

bility, and a membrane environment. The most populated

conformations for the two complexes were calculated

through a cluster analysis for the last 80 ns of the MD simu-

lations (see methods). The MD simulations showed that the

ligand conformation predicted by the docking procedures

ꢁ

approximately 4.28 A, whereas C-4 was located 5.27 A from

the Fe atom of the haem, an identical value to that observed

for the ﬁrst most populated conformation. The third most

populated conformation also shared a similar map of

12

Jessica Elena Mendieta-Wejebe et al.

Biotransformation of HO-AAVPA by CYP450

Figure 8 MD simulation box and equilibrium properties of the CYP2C11-HO-AAVPA complex. (a) MD simulation box of the extracellular and

transmembrane regions of the CYP2C11-HO-AAVPA complex embedded in the POPC membrane. (b) Area per lipid. (c) RMSD analysis. Secondary

structure elements of CYP2C11 are depicted in red (a-helix), yellow (b-sheets) and green (loops). The last conformation from the 150 ns long MD

simulation of the CYP2C11-HO-AAVPA complex was selected for representation.

interactions as the ﬁrst and second ligand conformations,

Free energy calculations

adding a new hydrophobic contact with F²³⁷, whereas the C-

3 atom showed a similar distance to the ﬁrst conformation,

ꢁ

with an average value of 4.38 A from the haem Fe atom (Fig-

To obtain the energetic contribution of the HO-AAVPA-

CYP2C11 complex, the electrostatic energies, non-polar solva-

tion, van der Waals interactions and total contribution from

the binding free energy values were calculated using the

MMGBSA method (see Materials and Methods). Our analysis

shows that the ligand–protein complex formation is mainly

driven by non-polar interactions (DE_vwd+ DG_npol,sol), with

the DE_vwdinteractions making the major contribution and the

polar (DE_ele+ DG_ele,sol) interactions being unfavourable to

the molecular recognition (Table 3). These energetic contri-

butions agree with the map of interactions observed for the

most populated conformations between CYP2C11 and HO-

AAVPA, for which all the complexes are mainly stabilized

ure 9d). Overall, these results suggest that HO-AAVPA

reaches a conformation that is quite stable in the substrate

binding site of CYP2C11, maintaining a similar map of

interactions as those predicted through the docking proce-

dures. Moreover, to obtain a more reﬁned simulation of the

protein–ligand interactions than the one obtained from

docking methods, MD simulations were performed. A clear

example of this was noticed between several phenylalanine

cluster elements and HO-AAVPA, for which the more ener-

getically stabilized p ꢁ p interactions were observed from

the MD simulation results (Figure 9b–d).

13

Biotransformation of HO-AAVPA by CYP450

Jessica Elena Mendieta-Wejebe et al.

Figure 9 The three most populated conformations of HO-AAVPA into the active site of CYP2C11. (a) Structural superposition of the three most

populated conformations of CYP2C11-HO-AAVPA. (b) The ﬁrst most populated conformation. (c) The second most populated conformation. (d)

ꢁ

The third most populated conformation. Residues at a distance of ≤4 A stabilizing the CYP2C11-HO-AAVPA complex are shown in the salmon

stick representation.

Table 3 Binding free energy components of the CYP–ligand complexes (in units of kcal/mol)

System

DE_vdw

DE_ele

DG_ele,sol

DG_npol,sol

DE_polarDE_npol

DG_mmgbsa

ꢁTDS

DG_bind

CYP2C11-HO-AAVPA ꢁ38.70 (0.16) 1.22 (0.11) 11.87 (0.11) ꢁ4.78 (0.02) 13.90

ꢁ43.48 ꢁ29.58 (5.6) ꢁ14.05 (2.65) ꢁ15.53

Polar (DE_ele+ DG_ele,sol) contributions (DE_polar) and non-polar (DE_vwd+ DG_npol,sol) contributions (DE_npol). All the energies were averaged over several

snapshots (see Methods) and are given in kcal/mol (ꢀstandard error of the mean).

through hydrophobic interactions (Figure 9b–d). The entro-

pic component indicates a decrease in the conformational

Discussion

mobility upon the formation of the complex, which is the pro-

duct of a reduction in the number of degrees of freedom, con-

tributing to an unfavourable entropy component in the

calculated binding free energy, bringing this calculated value

close to the experimental values. Overall, these DG_bindvalues

for HO-AAVPA are more energetically favourable than those

observed for VPA.^[9]

Reducing the toxic effects of a widely used therapeutic drug

is one of the main objectives in drug development. VPA

has been reported to be a hepatotoxic drug, and this effect

is attributed to its CYP-dependent metabolism.^[3,4]This

effect represents a high risk when VPA is administered con-

tinuously for long periods of time.^[52,53]For this reason, the

aryl VPA derivative HO-AAVPA has been proposed as a

14

Jessica Elena Mendieta-Wejebe et al.

Biotransformation of HO-AAVPA by CYP450

potential anticancer agent based on previous studies.^[5]

Hence, to contribute to the HO-AAVPA metabolic assess-

ment, we performed in vitro (Figures 1–4 and 6) and in sil-

ico (Figures 7–9) studies to evaluate the CYP-mediated

metabolism of HO-AAVPA.

AAVPA from the rat liver microsomes. The HPLC analysis

revealed that 3a and 3b showed approximately the same

retention time as M1 (ﬃ5.4 min, Figure 3). However, due

to the low quantities of M1 obtained from the microsomal

assays, complete characterization by NMR was not possible.

However, these results suggest that M1 corresponds to a

hydroxylated metabolite, either 3a or 3b.

The in vitro assays were developed with rat liver micro-

somes to determine the HO-AAVPA metabolism and

revealed a metabolite named M1 (Figure 1). Thus, experi-

mental assays show typical saturation kinetics, showing

concentration-dependent activity (Figure 2). Furthermore,

with the obtained kinetic parameters, it was possible to

demonstrate that the metabolic activity in DEX-induced

microsomes was higher than that in the untreated micro-

somes (Figures 1 and 2), proving that DEX is a potent

inducer of CYP enzymes.^[12]Previous studies have shown

that DEX levels play an important role in the upregulation

of CYP2C11 expression and, consequently, in its participa-

tion in microsomal oxidation reactions. Speciﬁcally, this

effect has been observed in vitro where activation of

CYP2C11 was detected when low concentrations of DEX

were added to rat hepatocyte cultures^[12,54]Likewise, this

effect has been observed in vivo using Sprague-Dawley rats

treated with DEX.^[55]Despite the fact that the kinetic

parameters obtained in our study could not be attributed

to the activity of a speciﬁc isoform, the obtained results

from the induction of rat liver microsomes with DEX and

their inhibition with CIM suggest that CYP2C11 is a main

isoform that participates in the HO-AAVPA biotransfor-

mation.^[13,14]

Interaction modes between substrates and the CYP2C11

isoform were characterized by the spectral changes pro-

duced in the absorption band of Soret’s peak. Type I

ligands bind to the hydrophobic cavity of CYP2C11, which

is located near the catalytic site. This interaction displaces a

water molecule coordinated to the ferric iron (Fe³⁺), leav-

ing this atom pentacoordinated due to the change from a

low- to a high-spin state.^[15]This state favours the reduc-

tion of the haem group by electrons supplied from the

NADPH via NADPH-CYP reductase, thus beginning the

CYP’s catalytic cycle of oxidative metabolism. Type II

ligands can directly interact with Fe³⁺as an axial ligand,^[23]

keeping this atom in a low-spin state. For this reason, it has

been established that this type of ligand diminishes the

oxidative capacity of the CYP, avoiding the reduction of

the Fe³⁺by the electrons donated by NADPH via NADPH-

CYP reductase to initiate the catalytic cycle.^[57]These data

suggest that HO-AAVPA has afﬁnity for the CYP2C11 iso-

form with an interaction mode corresponding to a type I

ligand (Figure 6).

This result means that HO-AAVPA could interact

directly with amino acid residues lining the cavity of the

catalytic site of this CYP isoform and not as an axial ligand

with the Fe atom of the haem group; thus, HO-AAVPA can

be considered a CYP substrate.

For in vitro studies, the results presented here suggest

that the K_Mand V_maxvalues obtained from the substrate

depletion method can be used to make inferences about the

metabolism of a compound (Table 1). This method is more

straightforward and does not require an authentic standard

of metabolites for the construction of calibration curves.

Additionally, the approximation in the calculation of the

kinetic parameters is signiﬁcantly better. Moreover, this

method is easier to implement in the early stages of the

drug discovery process.^[56]

In addition, through molecular docking studies, it was

possible to reproduce the ﬁrst part of the CYP catalytic

cycle, which implies substrate binding to the enzyme; thus,

it is possible to predict the afﬁnity of HO-AAVPA for

CYP2C11. Our results revealed that the binding mode of

HO-AAVPA to CYP2C11 consists of the aromatic ring of

HO-AAVPA oriented at an angle less than 90° to the haem

group (Figure 7a), a characteristic that is consistent with

other previously described aromatic substrates.^[58]The pre-

dicted interactions are mainly hydrophobic, in which HO-

AAVPA contacts various amino acid residues lining the

Furthermore, one of the metabolites detected (M1) by

LC-MS corresponded to a HO-AAVPA hydroxylation pro-

duct according to the observed molecular ion (252.2 m/z,

Figure 4). This result was consistent with the main metabo-

lites predicted by the SMARTCyp web server (Figure 5),

which suggested that the more liable sites to undergo

hydroxylation by the CYP2C9 isoform (which is homolo-

gous to rat CYP2C11) are the meta and para positions of

the aryl moiety. Hence, we decided to perform the chemical

synthesis of the predicted HO-AAVPA metabolites 3a and

3b (Scheme 2). Once the synthesis was complete, the com-

pounds were submitted for HPLC and MS analysis to use

as standards to be compared with the metabolites previ-

ously obtained by the biotransformation assay of HO-

haem group, and the interactions with V¹⁰², I¹¹³, N²⁰⁴, F²⁰¹

,

F

²⁰⁵, F²⁰⁸, F¹¹⁴E³⁰⁰, T³⁰¹, L³⁶¹, V³⁶²and L³⁶⁶are relevant

(Figure 7b).

,

However, despite the high-sequence homology between

the CYP2C11 and CYP2C9 isoforms and the literature data

suggesting that CYP2C11 could be used as a model to pre-

dict CYP2C9 activity,^[59-61]the obtained results showed

that there is a low probability that HO-AAVPA could be

metabolized by CYP2C9 under the same mechanism

observed in the CYP2C11 catalytic site.

15

Biotransformation of HO-AAVPA by CYP450

Jessica Elena Mendieta-Wejebe et al.

Therefore, we have proposed a possible mechanism of

HO-AAVPA biotransformation by CYP2C11, which con-

sists of aromatic hydroxylation, as molecular simulation

studies have shown that the most susceptible region for

oxidative metabolism is the aromatic ring (Figures 8 and

9). This is supported by the fact that the distal hydrox-

yphenyl ring is oriented towards the haem group, with C-4

being the atom located in close proximity to the haem iron

(Fe), so it could be expected that a hydroxyl group could be

added after CYP-mediated catalysis. These results are con-

sistent with those reported for other drugs that contain an

aromatic ring in their structure. For those drugs, the

formation of an arene oxide that subsequently produces a

r-cationic complex has been proposed, resulting in the

formation of an oxide that will be ﬁnally converted into a

hydroxyl group. In this sense, it has been shown that

diclofenac follows a catalytic mechanism by CYP that is

similar to the one described here, in which C-4 of the

dichlorophenyl ring is in closer proximity to the Fe atom,

thus producing a hydroxylated metabolite at this posi-

tion.^[46]

metabolites yielded from in vitro studies with rat liver

microsomes.

Declarations

Conﬂict of interest

The Author(s) declare(s) that they have no conﬂicts of

interest to disclose.

Acknowledgements

Arianna Silva-Trujillo wishes to thank CONACYT for her

M.Sc. scholarship (276824). The authors would like to

thank to American Journal Experts (www.aje.com) for their

English language revision.

Funding

This research was ﬁnancially supported by CONACYT

(CB254600, 782, SEP-CONACYT-ANUIES-ECOS Francia:

296636 and CB286653), ICyTDF (ICyTDF/45/2012), and

COFAA-SIP/IPN (20200104/20196684/20195089).

Conclusions

In this study, we combined in silico predictions and in vitro

studies to investigate the CYP-mediated metabolism of

HO-AAVPA. These results suggest that CYP2C11 could be

mainly involved in HO-AAVPA metabolism yielding two

metabolites 3a and 3b, whose chemical structures corre-

spond to the hydroxylation of HO-AAVPA as was con-

ﬁrmed by the LC-MS analysis of the predominant

Author’s contribution

JEMW, MCRH, JCB and AA: designed the project, wrote

the paper and made the discussion, AST, MB, HLMF,

NLGA, FTC and ARC: developed the project either experi-

mentally or in silico studies.

membrane leakiness and cellular pro-

teolysis. Toxicol In Vitro 2012; 26:

Expert Opin Drug Metab Toxicol 2006;

2: 875–894.

References

545–551.

5. Prestegui-Martel B et al. N-(2-hydrox-

9. Bello M et al. Structural and energetic

analysis to provide insight residues of

CYP2C9, 2C11 and 2E1 involved in

valproic acid dehydrogenation selec-

tivity. Biochem Pharmacol 2014; 90:

145–158.

1. Devlin EJ et al. Cancer treatment side

effects: a meta-analysis of the relation-

ship between response expectancies

and experience. J Pain Symptom Man-

age 2017; 54: 245–248.

yphenyl)-2-propylpentanamide,

a val-

proic acid aryl derivative designed in

silico with improved anti-proliferative

activity in HeLa, rhabdomyosarcoma

2. Farooq M et al. Biological screening

of novel derivatives of valproic acid

for anticancer and antiangiogenic

properties. Asian Pac J Cancer Prev

2014; 15: 7785–7792.

and breast cancer cells. J Enzyme Inhib 10. Wang H et al. Evaluation of the bind-

Med Chem 2016; 31(Suppl. 3): 140–149.

6. Guengerich FP. Mechanisms of cyto-

chrome P450-catalyzed oxidations.

ACS Catal 2018; 8: 10964–10976.

ing orientations of testosterone in the

active site of homology models for

CYP2C11 and CYP2C13. Biochem

Pharmacol 2009; 78: 406–413.

3. Kiang TKL et al. Glutathione deple-

tion by valproic acid in sandwich-cul-

tured rat hepatocytes: role of

biotransformation and temporal rela-

tionship with onset of toxicity. Toxicol

Appl Pharmacol 2011; 252: 318–324.

4. Pourahmad J et al. A new approach

on valproic acid induced hepatotoxic-

7. Sadeque AJM et al. Human CYP2C9 11. Chu V et al. In vitro and in vivo induc-

and CYP2A6 mediate formation of the

hepatotoxin 4-ene-valproic acid. J Phar-

macol Exp Ther 1997; 283: 698–703.

tion of cytochrome P450: a survey of

the current practices and recommenda-

tions:

a pharmaceutical research and

8. Martignoni M et al. Species differ-

ences between mouse, rat, dog, mon-

manufacturers of America perspective.

Drug Metab Dispos 2009; 37: 1339–1354.

key and human CYP-mediated drug 12. Iber H et al. Regulation of hepatic

metabolism, inhibition and induction. cytochrome P450 2C11 by

ity:

involvement

of

lysosomal

16

Jessica Elena Mendieta-Wejebe et al.

Biotransformation of HO-AAVPA by CYP450

glucocorticoids. Arch Biochem Biophys

1997; 345: 305–310.

13. Chang T et al. Selective inhibition of

equation, the Schild plot and related

power equations. J Pharmacol Toxicol

Methods 2002; 46: 61–71.

Amber Molecular Dynamics package.

J Chem Theory Comput 2013; 9: 4131–

4139.

ꢁ

rat hepatic microsomal cytochrome P- 23. Mailman RB et al. Effect of chemical 35. Skjevik AA et al. LIPID11: a modular

450. II. Effect of the in vitro adminis-

tration of cimetidine. J Pharmacol Exp

Ther 1992; 260: 1450–1455.

structure on type II spectra in mouse

hepatic microsomes. Drug Metab Dis-

pos 1974; 2: 301–308.

framework for lipid simulations using

amber. J Phys Chem B 2012; 116:

11124–11136.

14. Levine M et al. In vivo cimetidine 24. Schenkman JB et al. Spectral studies 36. Wang J et al. Development and test-

inhibits

hepatic

CYP2C6

and

of drug interaction with hepatic

microsomal cytochrome. Mol Pharma-

col 1967; 3: 113–123.

ing of a general Amber force ﬁeld. J

Comput Chem 2004; 25: 1157–1174.

37. Jakalian A et al. Fast, efﬁcient genera-

tion of high-quality atomic charges

AM1-BCC model: II. Parameterization

and validation. J Comput Chem 2002;

23: 1623–1641.

CYP2C11 but not CYP1A1 in adult

male rats. J Pharmacol Exp Ther 1998;

284: 493–499.

25. Kosak TM et al. Ether cleavage re-

investigated: elucidating the mechanism

of BBr 3-facilitated demethylation of

aryl methyl ethers. Eur J Org Chem

2015; 2015: 7460–7467.

15. Omura T, Sato R. The carbon monox-

ide-binding pigment of liver micro-

somes.

I.

Evidence

for

its

hemoprotein nature.

J

Biol Chem

38. Berendsen HJC et al. Molecular

dynamics with coupling to an external

bath. J Chem Phys 1984; 81: 3684–

3690.

1964; 239: 2370–2378.

26. Sali A, Blundell TL. Comparative pro-

tein modelling by satisfaction of spa-

tial restraints. J Mol Biol 1993; 234:

779–815.

16. Fabrizi L et al. Identiﬁcation of the

cytochrome P450 isoenzymes involved

in the metabolism of diazinon in the

39. Darden T et al. Particle mesh Ewald:

an NÁlog(N) method for Ewald sums

in large systems. J Chem Phys 1993;

98: 10089–10092.

rat liver. J Biochem Mol Toxicol 1999; 27. Williams PA et al. Mammalian micro-

13: 53–61.

somal cytochrome P450 monooxyge-

nase: structural adaptations for

17. Manno M et al. The mechanism of

the suicidal reductive inactivation of

microsomal cytochrome P-450 by

membrane binding and functional 40. Go€tz AW et al. Routine microsecond

diversity. Mol Cell 2000; 5: 121–131.

Molecular dynamics simulations with

AMBER on GPUs. 1. Generalized

born. J Chem Theory Comput 2012; 8:

1542–1555.

halothane. Arch Toxicol 1991; 65(3): 28. Scott EE et al. Substrate routes to the

191–198.

buried active site may vary among

cytochromes P450: mutagenesis of the

18. Melet A et al. Substrate selectivity of

human cytochrome P450 2C9: impor-

tance of residues 476, 365, and 114 in

F-G region in P450 2B1. Chem Res 41. Salomon-Ferrer

R

et al. Routine

Toxicol 2002; 15: 1407–1413.

microsecond Molecular dynamics sim-

ulations with AMBER on GPUs. 2.

Explicit Solvent Particle Mesh Ewald.

J Chem Theory Comput 2013; 9: 3878–

3888.

recognition of diclofenac and sul- 29. Frisch MJ et al. Gaussian 98, Revision

faphenazole and in mechanism-based

inactivation by tienilic acid. Arch Bio-

chem Biophys 2003; 409: 80–91.

A.7. Pittsburgh, PA: Gaussian, Inc,

1998.

30. Morris GM et al. Automated docking

19. Khmelnitsky YL et al. In vitro biosyn-

thesis, isolation, and identiﬁcation of

predominant metabolites of 2-(4-(2-

hydroxyethoxy)-3,5-dimethylphenyl)-

5,7-dimethoxyquinazolin-4(3H)-one

(RVX-208). Eur J Med Chem 2013; 64:

121–128.

using a Lamarckian genetic algorithm 42. van Gunsteren WF, Berendsen HJC.

and an empirical binding free energy

function. J Comput Chem 1998; 19:

1639–1662.

Algorithm

for

macromolecular

dynamics and constraint dynamics.

Mol Phys 1977; 34: 1311–1327.

31. Jo S et al. CHARMM-GUI Membrane 43. Gohlke H, Case DA. Converging free

Builder for mixed bilayers and its

application to yeast membranes. Bio-

phys J 2009; 97: 50–58.

energy estimates: MM-PB(GB)SA

studies on the protein-protein com-

plex Ras-Raf. J Comput Chem 2003;

25: 238–250.

20. Silva-Trujillo A et al. A simple validated

RP-HPLC bioanalytical method for the 32. Woolf TB, Roux B. Structure, energet-

quantitative determination of a novel

valproic acid arylamide derivative in rat

hepatic microsomes. Biomed Chro-

matogr 2015; 29: 523–528.

ics, and dynamics of lipid-protein 44. Kollman PA et al. Calculating struc-

interactions: molecular dynamics tures and free energies of complex

molecules: combining molecular

a

study of the gramicidin A channel in

a DMPC bilayer. Proteins 1996; 24:

92–114.

mechanics and continuum models.

21. Nath A, Atkins WM. A theoretical

Acc Chem Res 2000; 33: 889–897.

validation of the substrate depletion 33. Woolf TB, Roux B. Molecular dynam- 45. Miller BR III et al. MMPBSA.py: an

approach to determining kinetic

parameters. Drug Metab Dispos 2006;

34: 1433–1435.

ics simulation of the gramicidin chan-

nel in a phospholipid bilayer. Proc

Natl Acad Sci U S A. 1994; 91: 11631–

11635.

efﬁcient program for end-state free

energy calculations. J Chem Theory

Comput 2012; 8: 3314–3321.

22. Cheng HC. The power issue: determi-

46. Yan Z et al. Detection of a novel reac-

tive metabolite of diclofenac: evidence

for CYP2C9-mediated bioactivation

nation of KB or Ki from IC50. A clo- 34. Kaus J et al. Improving the efﬁciency

ser look at the Cheng-Prusoff of free energy calculations in the

17

Biotransformation of HO-AAVPA by CYP450

Jessica Elena Mendieta-Wejebe et al.

via arene oxides. Drug Metab Dispos 54. Thangavel

C

et al. Inducibility of 61. Wang X et al. Pharmacokinetic inter-

2005; 33: 706–713.

male-speciﬁc isoforms of cytochrome

P450 by sex-dependent growth

hormone proﬁles in hepatocyte

cultures from male but not female

rats. Drug Metab Dispos 2006; 34:

410–419.

action studies of tanshinones with

tolbutamide, a model CYP2C11 probe

substrate, using liver microsomes, pri-

mary hepatocytes and in vivo in the

rat. Phytomedicine 2010; 17: 203–211.

47. Wester MR et al. The structure of

human cytochrome P450 2C9 com-

ꢁ

plexed with ﬂurbiprofen at 2.0-A res-

olution. Biol Chem 2004; 279:

35630–35637.

J

48. Ali HI et al. Studies on 16-hydroxyla- 55. Lee DY et al. Effects of enzyme induc-

Supporting Information

tion of steroid molecules and regiose-

lective binding mode in homology-

modeled cytochrome P450–2C11. Int

J Med Chem 2011; 2011: 1–11.

ers and inhibitors on the pharmacoki-

netics of intravenous torasemide in

rats. Int J Pharm 2005; 298: 38–46.

Additional Supporting Information

may be found in the online version of

this article:

56. Henderson CJ et al. An extensively

humanized mouse model to predict

pathways of drug disposition and

drug/drug interactions, and to facili-

tate design of clinical trials. Drug

Metab Dispos 2019; 47: 601–615.

49. Bello

M et al. Predicting peptide

Figure S1. (A) 1H NMR spectrum

of 2a (750 MHz, DMSO-d6). (B) 13C

NMR spectrum of 2a (188.5 MHz,

DMSO-d6).

vaccine candidates against H1N1

inﬂuenza virus through theoretical

approaches. Immunol Res 2015; 62:

3–15.

50. Bello M, Correa-Basurto J. Molecular 57. Gan L et al. Role of NADPH-cy-

Figure S2. (A) 1H NMR spectrum

of 2b (750 MHz, DMSO-d6). (B) 13C

NMR spectrum of 2b (188.5 MHz,

DMSO-d6)d.

dynamics simulations to provide

insights into epitopes coupled to the

soluble and membrane-bound MHC-

II complexes. PLoS One 2013; 8:

e72575.

tochrome P450 reductase and cyto-

chrome-b₅/NADH-b₅reductase in

variability of CYP3A activity in

human liver microsomes. Drug Metab

Dispos 2009; 37: 90–96.

Figure S3. (A) 1H NMR spectrum

of 3a (750 MHz, DMSO-d6). (B) 13C

NMR spectrum of 3a (188.5 MHz,

DMSO-d6).

51. Wolf MG et al. g_membed: Efﬁcient 58. Masubuchi Y et al. Mechanism-based

insertion of a membrane protein into

an equilibrated lipid bilayer with min-

imal perturbation. J Comput Chem

2010; 31: 2169–2174.

inactivation of CYP2C11 by diclofe-

nac. Drug Metab Dispos 2001; 29:

1190–1195.

Figure S4. (A) 1H NMR spectrum

of 3b (750 MHz, DMSO-d6). (B) 13C

NMR spectrum of 3b (188.5 MHz,

DMSO-d6).

59. Lonsdale R et al. Quantum mechan-

ics/molecular mechanics modeling of

regioselectivity of drug metabolism in

cytochrome P450 2C9. J Am Chem

Soc 2013; 135: 8001–8015.

52. Gerstner T et al. Oral valproic acid

for epilepsy–long-term experience in

therapy and side effects. Expert Opin

Pharmacother 2008; 9: 285–292.

Figure S5. (A) ESI/MS spectrum of

2a. Molecular ion at 280.1913 m/z.

(B) ESI/MS spectrum of 2b. Molecu-

lar ion at 280.1903 m/z. (C) ESI/MS

spectrum of 3a. Molecular ion at

252.1595 m/z. (D) ESI/MS spectrum

of 3b. Molecular ion at 252.1596 m/z.

53. Hallas J et al. Cancer risk in long- 60. Barbosa-Sicard

E et al. Eicosapen-

term users of valproate: a population-

based case-control study. Cancer Epi-

demiol Biomarkers Prev 2009; 18:

1714–1719.

taenoic acid metabolism by cyto-

chrome P450 enzymes of the CYP2C

subfamily. Biochem Biophys Res Com-

mun 2005; 329: 1275–1281.

18

Article Doi

Exploring the biotransformation of N-(2-hydroxyphenyl)-2-propylpentanamide (an aryl valproic acid derivative) by CYP2C11, using in silico predictions and in vitro studies

DOI: 10.1111/jphp.13270

Source and publish data:

Authors:

Article abstract of DOI:10.1111/jphp.13270

Full text of DOI:10.1111/jphp.13270

Products guided by the article

R&D Labs maybe for 923556-77-8

Relevant to this article

Hot Product