Metabolic activation and major protein target of a 1-benzyl-3- carboxyazetidine sphingosine-1-phosphate-1 receptor agonist

Article abstract of DOI:10.1021/tx300017s

1-{4-[(4-Phenyl-5-trifluoromethyl-2-thienyl)methoxy]benzyl} azetidine-3-carboxylic acid (MRL-A) is a potent sphingosine-1-phosphate-1 receptor agonist, with potential application as an immunosuppressant in organ transplantation or for the treatment of autoimmune diseases. When administered orally to rats, radiolabeled MRL-A was found to undergo metabolism to several reactive intermediates, and in this study, we have investigated its potential for protein modification in vivo and in vitro. MRL-A irreversibly modified liver and kidney proteins in vivo, in a dose- and time-dependent manner. The binding was found to occur selectively to microsomal and mitochondrial subcellular fractions. Following a nonspecific proteolytic digestion of liver and kidney proteins, a single major amino acid adduct was observed. This adduct was characterized with LC/MS/UV and NMR spectroscopy and was found to be the product of an unprecedented metabolic activation of the azetidine moiety leading to the formation of a ring-opened α,β-unsaturated imine conjugated to the ε-amino group of a lysine residue. The formation of this adduct was not inhibited when rats were pretreated with 1-aminobenzotriazole, indicating that P450 enzymes were not involved in the metabolic activation of MRL-A. Rather, our findings suggested that MRL-A underwent bioactivation via a β-oxidation pathway. Several other minor adducts were identified from protein hydrolysates and included lysine, serine, and cysteine conjugates of MRL-A. These minor adducts were also detected in microsomal incubations fortified with the cofactors for acyl-CoA synthesis and in hepatocytes. Trypsin digestion of crude liver homogenates from rats treated with radiolabeled MRL-A led to the identification of a single radioactive peptide. Its sequence, determined by LC/MS analysis, revealed that the target of the major reactive species of MRL-A in vivo is Lys676 of long chain acyl-CoA synthetase-1 (ACSL1). This lysine residue has been found to be critical for ACSL1 activity, and its modification has the potential to lead to biological consequences such as cardiac hypertrophy or thermogenesis dysregulation.

Full text of DOI:10.1021/tx300017s

Article
pubs.acs.org/crt
Metabolic Activation and Major Protein Target of a 1-Benzyl-3-
carboxyazetidine Sphingosine-1-phosphate-1 Receptor Agonist
Herve Aloysius, Vincent W. Tong, Jocelyn Yabut, Scott A. Bradley,§ Jackie Shang, Yan Zou,
and Richard A. Tschirret-Guth*
Department of Pharmacokinetics, Pharmacodynamics, and Drug Metabolism, Merck Research Laboratories, Rahway, New Jersey
07065, United States
S
* Supporting Information
ABSTRACT: 1-{4-[(4-Phenyl-5-trifluoromethyl-2-thienyl)-
methoxy]benzyl}azetidine-3-carboxylic acid (MRL-A) is a
potent sphingosine-1-phosphate-1 receptor agonist, with
potential application as an immunosuppressant in organ
transplantation or for the treatment of autoimmune diseases.
When administered orally to rats, radiolabeled MRL-A was
found to undergo metabolism to several reactive intermediates, and in this study, we have investigated its potential for protein
modification in vivo and in vitro. MRL-A irreversibly modified liver and kidney proteins in vivo, in a dose- and time-dependent
manner. The binding was found to occur selectively to microsomal and mitochondrial subcellular fractions. Following a
nonspecific proteolytic digestion of liver and kidney proteins, a single major amino acid adduct was observed. This adduct was
characterized with LC/MS/UV and NMR spectroscopy and was found to be the product of an unprecedented metabolic
activation of the azetidine moiety leading to the formation of a ring-opened α,β-unsaturated imine conjugated to the ε-amino
group of a lysine residue. The formation of this adduct was not inhibited when rats were pretreated with 1-aminobenzotriazole,
indicating that P450 enzymes were not involved in the metabolic activation of MRL-A. Rather, our findings suggested that MRL-
A underwent bioactivation via a β-oxidation pathway. Several other minor adducts were identified from protein hydrolysates and
included lysine, serine, and cysteine conjugates of MRL-A. These minor adducts were also detected in microsomal incubations
fortified with the cofactors for acyl-CoA synthesis and in hepatocytes. Trypsin digestion of crude liver homogenates from rats
treated with radiolabeled MRL-A led to the identification of a single radioactive peptide. Its sequence, determined by LC/MS
analysis, revealed that the target of the major reactive species of MRL-A in vivo is Lys676 of long chain acyl-CoA synthetase-1
(ACSL1). This lysine residue has been found to be critical for ACSL1 activity, and its modification has the potential to lead to
biological consequences such as cardiac hypertrophy or thermogenesis dysregulation.
detected, it is often not possible a priori to determine which
one is involved in protein modification, and more detailed
studies are required to identify the most reactive species.
1-{4-[(4-Phenyl-5-trifluoromethyl-2-thienyl)methoxy]-
benzyl}azetidine-3-carboxylic acid [MRL-A (Figure 1)] is a
potent sphingosine-1-phosphate-1 receptor agonist that dis-
played immunomodulatory properties in preclinical species.7 As
part of its evaluation as a potential drug development candidate,
its metabolic fate was investigated in BDC rats.8 The
metabolism of MRL-A in this species was extensive (Figure
1), and several metabolites indicative of potentially reactive
species were identified in bile. These metabolites included two
acyl glucuronides (M3 and M8), a glutathione conjugate (M7),
a taurine conjugate (M5) that was most likely formed via a
reactive acyl-CoA thioester intermediate, and a β-lactam (M11)
and its acyl glucuronide conjugate (M9). In a subsequent in
vivo covalent binding study in rats, it was determined that
MRL-A-related material was irreversibly bound to liver and
kidney proteins after oral administration of MRL-A (vide infra).
INTRODUCTION
■
It is now well accepted that the metabolic conversion of drugs
to chemically reactive species is an undesired feature as it has
the potential to cause damage to cellular macromolecules and
alteration of critical cellular pathways. These biochemical
changes can lead to clinical adverse events such as drug-
induced liver injury (DILI) or drug hypersensitivity.1,2 Early
assessment of the bioactivation potential of drug candidates has
thus become quite common in the pharmaceutical industry
because of the risk of, and cost associated with, drug
development termination or market withdrawal.3−5 The
formation and structure of reactive metabolites are often
determined via in vitro experiments using trapping agents such
as GSH or cyanide.6 The structural information obtained from
these studies can be used by medicinal chemists in their lead
optimization efforts to eliminate potential sites of bioactivation.
One of the limitations of this approach is that the in vivo
relevance of the pathways identified in vitro often is not known.
Alternatively, metabolic activation can be inferred from
metabolites identified during excretion studies in bile duct-
cannulated (BDC) animals.6 However, when more than one
potentially reactive metabolite or stable products thereof are
Received: January 9, 2012
Published: July 5, 2012
© 2012 American Chemical Society
1412
dx.doi.org/10.1021/tx300017s | Chem. Res. Toxicol. 2012, 25, 1412−1422

Chemical Research in Toxicology
Article
neutralized with HCl, and the radioactivity content was assessed by
liquid scintillation counting using a TRI-CARB 1900TR liquid
scintillation counter (Packard Instrument, Meriden, CT). The protein
content was determined using the bicinchoninic acid (BCA) protein
assay (Pierce, Rockford, IL).
Microsomal Incubations. [14C]MRL-A (10 μM) was incubated
with liver microsomes (1 mg/mL) at 37 °C for 60 min in 100 mM
phosphate buffer containing 5 mM MgCl2 with no cofactors or with 1
mM NADPH in total incubation volumes of 1 mL. Conditions for
CoA thioester formation included 2 mM ATP, 0.6 mM CoA, 0.05%
Triton X-100, and 1 mM DTT. The reactions were stopped by the
addition of 4 volumes of acetonitrile containing 2% acetic acid and the
mixtures processed as described above. For the identification of
adducts, the starting MRL-A concentration was increased to 50 μM
and the volume to 5 mL.
Hepatocyte Suspensions. [14C]MRL-A (10 or 100 μM) was
incubated with rat and human hepatocytes (1 × 106 cells/mL)
suspended in Krebs-Henseleit buffer for up to 4 h at 37 °C in total
incubation volumes of 1 mL. The reactions were stopped by the
addition of 4 volumes of acetonitrile containing 2% acetic acid and
mixtures processed as described above. For the identification of
adducts, the MRL-A concentration was 50 μM and the reaction
volume 5 mL.
Cultured Hepatocytes. Cryopreserved rat or human hepatocytes
suitable for cell culture were plated in 24-well collagen I plates at a
density of 0.35 × 106 cells/well and maintained in In Vitro Gro
Hepatocyte plating medium at 37 °C in 5% CO2. Plates were
incubated overnight, and before the treatment of hepatocytes with test
compound, the medium was replaced with In Vitro Gro incubation
medium at 37 °C in 5% CO2. Dosing solutions were prepared in
culture media to give a final [14C]MRL-A concentration of 50 μM.
Incubations were terminated at 48 h by removing the medium and
washing with cold phosphate-buffered saline. Cells were lysed in ice-
cold acetonitrile containing 1% formic acid and centrifuged at 3000g
for 10 min. The protein pellets were processed as described above.
Metabolite Identification. Metabolite and protein adduct
identification were performed using either an LCQ classic or an
LXQ ion-trap mass spectrometer (Thermo Fisher, Waltham, MA)
equipped with electrospray sources operated in positive ionization
mode interfaced with a Shimadzu HPLC system equipped with a
photodiode array detector. Accurate mass measurements were
performed on an LTQ-Orbitrap mass spectrometer (Thermo Fisher).
Metabolites and/or amino acid adducts were separated using a
Synergy Hydro-RP column (4.6 mm × 150 mm, 4.0 μm)
(Phenomenex, Torrance, CA) operated at a flow rate of 1 mL/min
using a gradient of solvent B [100:0.1 (v/v) acetonitrile:acetic acid]
into solvent A [100:0.1 (v/v) water:acetic acid] using the following
program: 30% B from 0 to 2 min, 30 to 45% B from 2 to 17 min, 45%
B from 17 to 20 min, 45 to 100% B from 20 to 23 min, 100% B from
23 to 26 min, 100 to 30% B from 26 to 27 min, and 30% B from 27 to
30 min. The radioactivity in the eluent was monitored online using a
Flow Scintillation Analyzer (PerkinElmer, Waltham, MA) or by
collecting fractions in 96-well Wallac Scintiplates (PerkinElmer) and
counting radioactivity in a Wallac MicroBeta TriLux counter
(PerkinElmer) after samples had been dried under a stream of
nitrogen.
Identification of Protein Adducts. Tissue homogenates and in
vitro samples were washed to remove unbound radioactive material
using acetonitrile containing 2% acetic acid, as described above. The
final pellets were resuspended in 100 mM Tris-HCl buffer (pH 7.5)
containing 5 mM CaCl2 and 0.1% SDS. For amino acid adduct
identification, Pronase [protease XIV (Sigma-Aldrich, St. Louis, MO)]
was added to a final concentration of 2 mg/mL and the sample
incubated for 24−72 h at 37 °C. For peptide adduct identification,
trypsin was added to a final concentration of 0.5 mg/mL and the
sample incubated for 16 h at 37 °C. After digestion, the samples were
extracted using HLB OASIS solid phase extraction (SPE) cartridges
(Waters Corp., Milford, MA) eluted with acetonitrile and/or
methanol. Extraction recoveries were typically around 80%. Eluants
were dried and redissolved in 30% acetonitrile prior to analysis. For
Figure 1. Structure of MRL-A and its major metabolites excreted in rat
bile. The asterisk indicates the position of mono-3H or 14C labeling.
GS, glutathione; Gluc, glucuronic acid.
In the investigation presented here, we have sought to define
the nature of the bound adduct, its mechanism of formation,
and the identity of the major protein target. We describe herein
our findings of a novel bioactivation pathway of the 3-
carboxyazetidine moiety and its selective modification of a
single protein in vivo.
MATERIALS AND METHODS
■
Chemicals. MRL-A (including [3H]MRL-A and [14C]MRL-A) and
its lysine and serine conjugates were prepared at Merck Research
Laboratories (Rahway, NJ). The radiotracers had a radiochemical
purity of >98%. The specific activity of [3H]MRL-A was 13 mCi/
μmol, and that of [14C]MRL-A was 54 μCi/μmol. All other chemicals
were purchased from standard vendors and were of the highest
analytical purity available. Rat liver microsomes were prepared
according to a standard protocol.9 Cryopreserved rat and human
hepatocytes were purchased from In Vitro Technologies, Inc.
(Baltimore, MD). Edman sequencing was conducted at Argo
BioAnalytica (Kenilworth, NJ).
Covalent Binding Studies. In Vivo. All animal studies were
conducted in accordance with federal guidelines for the care and use of
laboratory animals and were approved by the Merck & Co., Inc.,
Institutional Animal Care and Use Committee. Rats were treated
orally with either [3H]- or [14C]MRL-A (4, 10, 25, 50, and 100 mg/kg
at ∼50 μCi/rat) as a suspension in 0.5% methylcellulose. 1-
Aminobenzotriazole (ABT) was administered at 100 mg/kg in 0.5%
methylcellulose 16 h prior to administration of MRL-A. Plasma, liver,
and kidneys were harvested 2, 6, and 24 h postdose. Livers and kidneys
were homogenized with 4 volumes of water. Subcellular fractions of
liver samples were obtained after differential centrifugation using a
published method.10 Aliquots of plasma, liver, and kidney homoge-
nates or subcellular fractions were precipitated with 4 volumes of
acetonitrile containing 2% acetic acid, held at 4 °C for 1 h, and
centrifuged at 3000g for 10 min. For removal of residual unbound
radioactive material, the protein pellets were resuspended in 1 mL of
water, treated with an additional 4 volumes of acetonitrile containing
2% acetic acid, and centrifuged. These steps were repeated until the
radioactivity in the supernatant had fallen to background levels. The
final protein pellets were redisolved in 0.25 N NaOH overnight and
1413
dx.doi.org/10.1021/tx300017s | Chem. Res. Toxicol. 2012, 25, 1412−1422

Chemical Research in Toxicology
Article
Table 1. Total Tissue Concentrations and Radioactivity Irreversibly Bound to Plasma, Liver, and Kidney Proteins following
Oral Administration of Radiolabeled MRL-A to Male Rats
total concentration in tissue (μM equiv or nmol
irreversible binding
(pmol equiv/mg of protein)
equiv/g)
radiolabel
dose (mg/kg) time (h)
plasma
liver
kidney
plasma
liver
16
42 10
93
kidney
[3H]MRL-A
[14C]MRL-A
10
2
6
3.7 1.8
7.7 5.0
118 31
83 45
78 15
<1
<1
<1
2
9
6
14
30
33
44
50
71
10
11
5
2
5
4
4
1
5
1
1
84
31
7
6
24
24
24
24
24
24
24
1.3 0.3
23
27
67
8
3
2
6
10
25
50
100
4
1.0 0.1
58 19
1
3
2
1
118 17
169 21
3.4 0.5
126
6
5
11 1.0
284 35
549 138
143 21
304 65
2.8 0.1
2.7 0.4
7
253
8
21 6.2
9
329 13
[14C]MRL-A
0.051 0.002
0.058 0.004
10
11
1
1
<1
<1
54
70
2
4
[14C]MRL-A with ABT (100 mg/kg)
4
large scale preparations, the final samples obtained after SPE extraction
were fractionated on a semipreparative Vydac C4 column (10 mm ×
250 mm) eluted at a rate of 3 or 4 mL/min using a gradient of solvent
B [100:0.1 (v/v) acetonitrile:trifluoroacetic acid] into solvent A
[100:0.1 (v/v) water:trifluoroacetic acid] using the following program:
10% B from 0 to 2 min, 10 to 70% B from 2 to 70 min, 70 to 90% B
from 70 to 80 min, and 90% B from 80 to 90 min. Final peptide
analysis was performed on a Vydak C18 column (2 mm × 250 mm)
eluted at a rate of 200 μL/min using a gradient of solvent B [100:0.1
(v/v) acetonitrile:acetic acid] into solvent A [100:0.1 (v/v) water:-
acetic acid] using the following program: 30% B from 0 to 2 min, 30 to
60% B from 2 to 52 min, 60 to 90% B from 52 to 55 min, and 90% B
from 55 to 60 min.
NMR Analysis. NMR spectra of the parent drug and adducts
thereof were obtained in a [2H3]acetonitrile/2H2O mixture [8:1 (v/v)]
at room temperature using a Varian Inova 600 MHz NMR
spectrometer (Agilent, Santa Clara, CA) equipped with a 3 mm probe.
Figure 2. Irreversible binding (“IB”) of drug-related material to liver
subcellular fractions in rats treated orally with 10 mg of [14C]MRL-A/
kg (n = 3; 24 h).
modified amino acids. The resulting radiochromatographic
profiles obtained from liver at the 10 and 100 mg/kg doses are
shown in panels A and B of Figure 3, respectively. The
RESULTS
■
In Vivo Covalent Binding. The potential of MRL-A to
undergo metabolic activation and irreversible binding to
proteins in vivo was examined in male rats following oral
administration of either [3H]MRL-A or [14C]MRL-A at doses
ranging from 10 to 100 mg/kg (Table 1). There was a time-
and dose-dependent increase in the levels of irreversible
binding to liver and kidney proteins up to ca. 330 and 70
pmol equiv/mg, respectively, at the 100 mg/kg dose. In
contrast, the levels of binding to plasma proteins were low at all
time points. Of note, at 24 h, the levels of irreversible binding
following a 10 mg/kg dose of either radiotracer were closely
similar, suggesting that the benzylic position was not a major
site of bioactivation. In liver, the irreversible binding was found
to be selective for microsomal and mitochondrial proteins
relative to cytosolic proteins (Figure 2). At a [14C]MRL-A dose
of 4 mg/kg, pretreatment of rats with 1-aminobenzotriazole
(ABT, 100 mg/kg), a nonspecific P450 inhibitor,11 did not
abolish binding to liver and kidney proteins (Table 1). Instead,
a small increase in the level of irreversible binding to liver
proteins was observed.
Extraction and LC/MS analysis of the unbound radioactive
material from liver and kidney homogenates indicated that
unchanged MRL-A was the major drug-related component in
these tissues [>90% (data not shown)]. Small amounts of the
taurine conjugate (M5) and the lactam derivative (M11) were
also observed.
Identification of in Vivo Adducts. To identify the nature
of the adduct(s) formed in rat liver and kidneys, the protein
pellets obtained after removal of unbound material were
digested with Pronase, a nonspecific protease, to liberate single
Figure 3. Radiochromatograms of Pronase digests of liver proteins of
rats treated with [14C]MRL-A at (A) 10 or (B) 100 mg/kg.
radiochromatograms obtained from kidney homogenates were
essentially identical to those observed from liver at both doses,
albeit with lower overall radioactivity levels (Figure S1 of the
Supporting Information). There was a dose dependence in the
profiles obtained, and a greater number of adducts were
observed at the higher dose. Altogether, a total of seven MRL-
a
A-related products (A1−6 and M2) were detected at 24 h.
1414
dx.doi.org/10.1021/tx300017s | Chem. Res. Toxicol. 2012, 25, 1412−1422

Chemical Research in Toxicology
Article
Pretreatment with ABT did not result in significant changes in
the profile, and A1 remained the major peak observed (Figure
S2 of the Supporting Information).
Adduct A1 had a protonated molecular ion at m/z 590 that,
upon collision-induced dissociation (CID), yielded fragment
ions at m/z 546 and 241, the latter being a characteristic
fragment of MRL-A8 (Figure 4A,B and Table 2). The exact
characterized by UV/vis and NMR spectroscopy. The UV
spectrum of A1 had a characteristic absorption maximum at
298 nm that was not present in the spectrum of the parent
molecule (Figure 4C), which suggested the presence of a newly
formed conjugated system. Under mild base treatment (pH
∼9), A1 underwent hydrolysis to a single product with a
molecular ion at m/z 364 that was identified as primary amine
M2 (Figure S3 of the Supporting Information). Product M2
had also been detected in the Pronase digest (Figure 3) and
possibly resulted from limited decomposition of MRL-A during
proteolysis and/or sample processing. Finally, a large scale
isolation of A1 was performed to obtain sufficient material for
NMR characterization. The 1H and 13C NMR chemical shifts of
MRL-A and A1 are listed in Table 3, and the low field region of
1
the H NMR sprectum and the heteronuclear multiple-bond
correlation (HMBC) and nuclear Overhauser effect spectros-
copy (NOESY) correlations of A1 are presented in the Figures
S4 and S5, respectively, of the Supporting Information. The
signals for the protons of the azetidine ring at 4.17 ppm (4H,
m, H-f) and 3.62 ppm (1H, pent, J = 8.4 Hz, H-g) in the parent
spectrum were absent in the spectrum of the adduct. In their
place were two new signals at 7.62 ppm (1H, s, H-f) and 7.56
ppm (1H, s, H-h) in the 1H NMR spectrum of the adduct. The
two-dimensional (2D) 1H−13C HMQC spectrum correlated H-
f to a carbon at 164.8 ppm and H-h to a carbon at 165.3 ppm.
1
The H and 13C chemical shifts of these two units are best
explained by enamine/imine functionalities. The location of
1
these two units was deduced from the 2D H−13C HMBC
spectra. Proton f showed long-range coupling to the benzylic
methylene (52.4 ppm, C-e) and a carbonyl carbon at 172.5
ppm (C-i). Proton h exhibited long-range coupling to this same
carbonyl carbon (C-i) and to the carbon of the ε-methylene in
the lysine residue (49.2 ppm, C-e′). This connectivity was also
1
supported by the 2D H−1H NOESY spectrum, in which
correlations were observed between H-e and H-f and between
H-h and H-e′. These data led to the proposed structure of A1
(Figure 5 and Table 3), which links the lysine to the parent
compound through an enamine/imine moiety.
Figure 4. Characterization of adduct A1: (A) MS spectrum, (B) MS2
spectrum at m/z 590, and (C) UV/vis spectrum.
Adducts A2, A3, and A5 had protonated molecular ions at
m/z 576, 535, and 549 (Table 2), respectively, consistent with
amido or ester conjugates of MRL-A with lysine, serine, and
threonine, respectively. Accordingly, the CID spectra of all
three adducts had characteristic fragment ions at m/z 241 and
347, indicating that modification had occurred on the 3-
carboxyazetidine moiety of MRL-A (Table 2). Definitive
structural assignment for A2 was achieved by comparison to
an authentic standard that confirmed that the carboxylic acid
mass of the protonated molecular ion of A1 was consistent with
a molecular ion formula of C29H31N3O5F3S+ [Δ 1.02 ppm
(Table 2)] and, on the basis of the numbers of added carbon
and nitrogen atoms, indicated the addition of a lysine
(C6H14N2O2) moiety to a product of oxidation of MRL-A. A
definitive structure could not be unambiguously assigned using
the mass data alone (vide infra), and A1 was thus further
Table 2. MS Characterization of MRL-A and Its Amino Acid Adducts
a
b
adduct
observed mass (m/z)
formula
calcd mass (m/z) Δppm
MS2 (m/z)
MS3 (m/z)
MRL-A
A1
448.1196
590.1943
576.2155
535.1524
551.1837
549.168
C23H21F3NO3S+
C29H31F3N3O5S+
C29H33F3N3O4S+
C26H26F3N2O5S+
C27H30F3N2O5S+
C27H28F3N2O5S+
448.1189
590.1937
576.2138
535.1509
551.1822
549.1666
670.1322
462.1345
755.2027
1.56
1.02
2.95
2.80
2.72
2.55
4.92
2.38
4.47
241, 347
241, 347, 484, 501, 546, 572
241, 347, 452, 470, 531, 558
241, 347, 429, 475, 517
241, 347, 448, 504, 533
241, 347
241, 305, 347, 400, 501, 529
A2
A3
A4
A5
+
A6
670.1355
462.1356
755.2059
C29H31F3N3O6S3
C24H23F3NO3S+
C33H38F3N4O9S2
241, 347, 551, 564
A7
241, 347
+
A8
241, 347, 374, 482, 520, 609, 626
a
b
MS2 indicates the ions formed following CID of the molecular ion. The predominant peak is underlined and shown in bold. MS3 indicates the ions
formed following CID of the predominant obtained via MS2.
1415
dx.doi.org/10.1021/tx300017s | Chem. Res. Toxicol. 2012, 25, 1412−1422

Chemical Research in Toxicology
Article
Table 3. NMR Proton and Carbon Chemical Shifts of MRL-A and Adduct A1
MRL-A
A1
atom
1H
7.22 (1H, s)
13C
1H
7.21 (1H, s)
5.31 (2H, s)
13C
a
130.7
64.6
115.8
132.2
57.6
55.2
32.6
−
130.6
64.7
b
5.33 (2H, s)
c
7.07 (2H, d, J = 8.8 Hz)
7.03 (2H, d, J = 8.6 Hz)
115.7
129.6
52.4
d
7.39 (2H, d, J = 8.8 Hz)
7.28 (2H, d, J = 8.6 Hz)
e
4.25 (2H, s)
4.54 (2H, s)
f
4.17 (4H, m)
7.62 (1H, s)
164.8
b
g
3.62 (1H, p, J = 8.5 Hz)
−
nd
h
i
−
−
−
−
7.56 (1H, s)
165.3
172.5
130.0
157.9
142.7
128.8
55.0
b
nd
−
−
−
−
j
122.7
159.3
142.4
128.8
−
−
−
−
−
k
l
−
Ph
a′
b′1
b′2
c′
d′
e′
7.43 (5H, m)
7.43 (5H, m)
a
−
−
−
−
−
−
3.54 (na, m)
a
b
1.76 (na, m)
nd
a
b
1.83 (na, m)
nd
a
b
1.40 (na, m)
nd
a
1.63 (na, m)
29.2
49.2
a
−
3.43 (na, m)
a
b
Not applicable. Integration was not possible because of the overlap with impurities. Not detected.
Figure 5. Structures of adducts of MRL-A identified in vivo and in vitro. The configuration of adduct A1 shown is arbitrary, and the stereochemistry
is not fixed.
group was linked to the ε-NH2 group of a lysine. Adduct A3
was stable under mild base treatment and thus could not be an
ester derivative [A3a (Figure 5)]. Most likely, it was the
product of intramolecular rearrangement of the A3a adduct to
1416
dx.doi.org/10.1021/tx300017s | Chem. Res. Toxicol. 2012, 25, 1412−1422

Chemical Research in Toxicology
Article
Figure 6. Isolation and characterization of the major tryptic peptide adduct, P1, from the liver of rats treated orally with 10 mg of [14C]MRL-A/kg.
(A) Fractionation of the crude tryptic digest. (B) LC/UV chromatogram of P1. (C) UV/vis spectrum of P1. (D) Full scan mass spectrum of P1.
form the more stable amide conjugate following release of the
free α-amino group during proteolysis. Comparison with an
authentic standard of A3 confirmed that the serine was linked
via the α-NH2 group of serine. The structure of the threonine
adduct, A5, was inferred from that of A3. Adduct A4 had a
protonated molecular ion at m/z 551.1837, consistent with a
molecular formula of C27H30N2O5F3S+ [Δ 2.72 ppm (Table
2)]. Its fragmentation pattern included major ions at m/z
533.1719 (loss of H2O) and m/z 448.1192 (formula identical
to that of MRL-A). In addition, A4 was readily hydrolyzed to
MRL-A under mild basic conditions (data not shown). In spite
of these data, a satisfactory structure could not be assigned to
A4. Adduct A6 had a protonated molecular ion at m/z
670.1355 (Table 2) and fragment ions at m/z 241 and 347
indicating that modification had also occurred on the 3-
carboxyazetidine moiety. An additional fragment ion at m/z
551.1291 was consistent with the loss of cysteine (Δ 1.8 ppm).
A6 was resistant to mildly acidic or basic treatments but slowly
degraded when exposed to air to form two products with parent
ions at m/z 583 (A6c) and m/z 599 (A6d) (Figure S6 of the
Supporting Information). The only structure consistent with
the available data was that of an N-linked cystine adduct
(Figure 5). It was most likely formed upon intramolecular
rearrangement of a cysteine thioester A6a to the amide A6b
that either reacted directly with free cystine or was first oxidized
to a sulfenic acid and reacted with cysteine to form A6. Under
aerobic storage conditions, A6 was oxidized to the sulfinic and
sulfonic acid derivatives A6c and A6d, respectively. The last
MRL-A-related compound identified was A7, a methyl ester of
the parent molecule, which most likely was an artifact of sample
processing that included a brief methanol wash of the protein
pellet prior to Pronase digestion and might have resulted in
limited methanolysis of some of the ester or thioester adducts.
Protein Target of MRL-A in Vivo. The intriguing structure
of adduct A1 led us to attempt to identify the protein from
which it was derived. The protein pellet of the liver
homogenate from the 10 mg/kg dose was treated with trypsin
and the resulting peptide mixture fractionated using a
semipreparative column to yield one major radiolabeled peak
(Figure 6A). Subsequent LC/UV/MS analysis of this peak
resulted in the identification of a single peptide (P1) bearing
the characteristic UV/vis spectrum of adduct A1 (Figure 6B,C).
Peptide P1 had a singly protonated molecular ion at m/z
1913.9302 and doubly and triply protonated ions at m/z
957.9704 and 638.9830, respectively (Figure 6D). The product
ion spectra of the latter two ions (Figure S7 of the Supporting
Information) were consistent with the sequence
SIDNGLLTPTLK*AK (Figure 7). Edman sequencing un-
equivocally confirmed the seven first amino acids of the
sequence (SIDNGLL). A search of the rat genomic database
indicated that this sequence uniquely belonged to rat long chain
acyl-CoA-synthase 1 (ACSL1) and that Lys676 was the site of
modification by the reactive intermediate of MRL-A.
In Vitro Covalent Binding in Rat Liver Microsomes.
When incubated with rat liver microsomal proteins in the
presence of NADPH, [14C]MRL-A yielded low levels of
irreversible binding (Table 4). The binding was time- and
NADPH-dependent and was effectively prevented by the
addition of GSH to the incubation medium. LC/MS analysis
of the incubation performed in the presence of GSH indicated
1417
dx.doi.org/10.1021/tx300017s | Chem. Res. Toxicol. 2012, 25, 1412−1422

Chemical Research in Toxicology
Article
Figure 7. Amino acid sequence and fragmentation patterns of P1.
Figure 8. Radiochromatograms of (A) Pronase and (B) trypsin digests
of liver microsomes incubated with [14C]MRL-A (50 μM) and
NADPH. The differences in retention times are due to a different LC
method used for each type of digestion (see Materials and Methods).
The peak at 13 min in panel A was also present in control incubations
in the absence of NADPH.
Table 4. Radioactivity Irreversibly Bound to Proteins after
Incubation of [14C]MRL-A with Rat and Human Liver
a
Microsomes and Hepatocyte Suspensions
[MRL-
A]
irreversible binding
(pmol equiv/mg of
protein)
time
In Vitro Covalent Binding in Hepatocytes. Incubation of
[14C]MRL-A with rat hepatocyte suspensions led to a time- and
concentration-dependent increase in the level of irreversible
binding of MRL-A-related radioactivity to proteins (Table 4).
The metabolites formed during these incubations included the
two isomeric N-oxides (M4 and M6) and taurine conjugate M5
(not shown). The protein adducts were characterized by LC/
MS following Pronase digestion of the protein extracts, and a
representative radiochromatogram is shown in Figure 9A. All
adducts previously identified in liver samples in vivo (Figure 3)
were also detected, albeit at different relative levels. Of note,
adduct A1, which was a major adduct in rat liver in vivo, was
now only found in trace amounts. Interestingly, in cultured rat
hepatocytes incubated for 48 h with [14C]MRL-A, A1 became
the predominant adduct formed (Figure 9B). Irreversible
binding to proteins was also observed in human hepatocytes
incubated with [14C]MRL-A (Table 4); however, the major
adduct identified in both suspended and plated hepatocytes was
A6, the product of conjugation to cysteine residues (Figure
9C,D). Adduct A1 was also observed in plated human
hepatocytes, albeit at levels much lower than those observed
in rat hepatocytes. In both suspended and plated hepatocyte
incubations, unchanged MRL-A was also detected in significant
amounts in the final protein adduct profiles. Its relative levels
were consistent when assays were repeated, and thus, it did not
appear to arise from incomplete removal of unbound material
prior to proteolytic digestion but rather from decomposition of
one or more adducts during sample workup.
liver preparation
rat microsomes
(μM)
cofactor
(min)
10
none
60
0
1.9
1.4
21
NADPH
30
60
60
34
NADPH,
GSH
3.6
ATP, CoA
60
0
32
1.8
9.1
16
rat hepatocytes
10
100
10
−
30
60
120
0
20
−
17
30
60
120
0
68
99
120
1.0
74
human hepatocytes
120
a
Data are averages of two or three measurements.
the formation of a single GSH adduct [A8 (Figure S8 of the
Supporting Information)] with a protonated molecular ion at
m/z 755.2059 (Table 2). Its fragmentation ions were consistent
with addition of GSH to a ring-opened azetidine moiety
(Figure 5 and Figure S8 of the Supporting Information).
Pronase or trypsin digestion of the protein pellets obtained
from incubation of MRL-A in the presence of NADPH did not
yield any detectable amino acid or peptide adducts. In both
cases, the only MRL-A-related peak identified was primary
amine M2 (Figure 8A,B). When [14C]MRL-A was incubated in
the presence of CoA and ATP instead of NADPH, the levels of
irreversible binding were likewise relatively low (Table 4). The
acyl-CoA thioester of MRL-A was not observed in the
supernatant of these incubations; however, after Pronase
digestion of the protein pellet, adducts A2−A4 and A6 were
detected by selected reaction monitoring LC/MS analysis
(Figure S9 of the Supporting Information). In contrast, adduct
A1 was not observed in these samples.
DISCUSSION
■
In this study, we have investigated the potential of MRL-A, a
sphingosine-1-phosphate-1 receptor agonist, to form reactive
intermediates and bind to cellular proteins in vivo and in vitro.
Because most of the metabolites of MRL-A identified in rat bile
were indicative of potentially reactive intermediates, it was not
unexpected to see significant levels of irreversible binding to rat
liver and kidney proteins in vivo. Identification of the
pathway(s) involved in protein modification could not be
inferred from these metabolites alone and required more
1418
dx.doi.org/10.1021/tx300017s | Chem. Res. Toxicol. 2012, 25, 1412−1422

Chemical Research in Toxicology
Article
Figure 9. Radiochromatograms of Pronase digests of incubations of MRL-A (50 μM) with (A) suspended and (B) cultured rat hepatocytes and (C)
suspended and (D) cultured human hepatocytes. No MRL-A-related ions could be identified under the peak eluting at 25 min in panel B.
detailed investigations. As a first step, we attempted to digest
tissue homogenates to single amino acids using 6 N HCl, but
MRL-A degraded rapidly under these conditions. Instead, we
used a nonspecific protease that allowed us to use milder
conditions. The large side chain modifications did not hamper
proteolysis, and single modified amino acids were readily
obtained.
The structure of the major adduct (A1) required detailed
analysis before its identity could be defined unequivocally. Our
initial structural proposal [A1a (Figure 10)] was based solely
spectrum of the A1 peak exhibited an absorbance maximum at
298 nm that the assignment was deemed incorrect because the
proposed structure did not contain any conjugated system
consistent with an absorption band at this wavelength.
Furthermore, we noticed that A1 degraded to the benzylic
amine M2 under mild base treatment, which was also
inconsistent with structure A1a. To unequivocally assign the
structure of A1, a larger portion of the liver homogenate was
then digested to isolate enough material for NMR analysis. We
estimated that at 100 pmol/mg of irreversible binding, 1 g of
liver would yield roughly 1−2 μg of adduct.
The final structure of A1, which was unexpected because
bioactivation of an azetidine moiety has yet to be reported, is
most likely the product of the reaction of a ring-opened α,β-
unsaturated aldehyde [I2 (Figure 11 or 12)] with a protein
lysine side chain. We initially hypothesized that I2 was formed
via a P450-mediated pathway (Figure 11) as supported by our
findings in liver microsomes. (1) The identification of
glutathione conjugate A8 suggested a P450-mediated ring
opening of the azetidine to aldehyde intermediate I1 via a
carbinolamine. (2) Benzylamine M2 was released from the
microsomal protein pellet after proteolysis, which was also
observed from the liver samples in vivo. However, these two
observations could ultimately not be reconciled with formation
of A1, which would require further oxidation of I1 to α,β-
unsaturated aldehyde I2 (Figure 11, path a). Rather, what
seemed to be occurring in microsomes was reduction of the
aldehyde moiety of I1, possibly via a carbonyl reductase (path
b), to form alcohol I3, which upon dehydration yielded α,β-
unsaturated acid I4. In the presence of GSH, I4 is trapped to
form adduct A8. The extent to which I4 is also able to react
with proteins is unclear as we were unable to identify the
corresponding cysteine adduct after Pronase digestions, despite
multiple attempts. The fact that M2 was observed instead
suggests that I4 is further oxidized to α,β-unsaturated imine I5,
which is most likely the species that binds to proteins in
microsomes to form adducts such as A9. The imine bond in A9
is expected to be less stable after hydrolysis than the one
contained in A1 because of the lack of conjugation and the
presence of the adjacent carboxylate group that is well
positioned to assist in hydrolysis at neutral pH. The highly
Figure 10. Alternate structure and mechanism of formation of adduct
A1.
on unit-resolution LC/MS data. The molecular ion of A1 (m/z
590) was consistent with a lysine adduct of M11, a β-lactam
metabolite of MRL-A previously identified in rat bile. Its
formation was easily rationalized because the acyl glucuronide
conjugate of M11, M9, was the second most abundant
metabolite identified in rat bile. Acyl glucuronides have been
implicated in the bioactivation of carboxylic acids and in the
covalent modification of proteins,12−14 and thus, we believed
that we had identified the major bioactivation pathway of MRL-
A in rats (Figure 10). Our hypothesis was further supported by
subsequent high-resolution mass spectroscopy (HRMS)
analysis, which yielded a molecular ion consistent with the
molecular formula of the proposed structure (C29H31F3N3O5S+,
Δ 1.02 ppm). It was only after noticing that the UV/vis
1419
dx.doi.org/10.1021/tx300017s | Chem. Res. Toxicol. 2012, 25, 1412−1422

Chemical Research in Toxicology
Article
because M2, the final product obtained after hydrolysis, was
observed in the Pronase digests of liver and kidney
homogenates.
A second pathway by which MRL-A could be oxidized to I2
involves a β-oxidation mechanism initiated by formation of an
acyl-CoA intermediate [I6 (Figure 12)]. This mechanism is, in
part, supported by the presence of taurine conjugate M5 in rat
bile, which represented the major route of elimination of MRL-
A (path a). The first step in the formation of taurine conjugates
is thioesterification with CoA, a common reaction with many
carboxylic acid-containing xenobiotics.12,14 Although acyl
thioester I6 was not observed when MRL-A was incubated in
liver microsomes in the presence of ATP and CoA, we did
observe the formation of amino acid adducts A2−6, which are
consistent with its formation as an intermediate (path b). β-
Oxidation of acyl-CoA thioesters can occur in mitochondria
and peroxisomes, but their entry into the mitondria involves a
carnitine-palmitoyltransferase shuttle that is sensitive to chain
length and branching at the α position. Thioester I6 is thus
more likely to be further metabolized in the peroxisome where
branched and/or bulky acyl-CoA thioesters can be further
degraded.15,16 The simplest and most direct conversion of I6 to
the final reactive α,β-unsaturated aldehyde I2 would involve
dehydrogenation to I7, followed by hydration and ring opening
to aldehyde I8, a second dehydrogenation step to form I9, the
acyl-CoA thioester of I2, and a final hydrolytic step. These
steps have been extensively characterized in the context of
mitochondrial or peroxisomal fatty acid β-oxidation and are all
known to occur enzymatically.15,17,18 However, to the best of
our knowledge, oxidation of an azacyclic carboxylic acid is
unprecedented, and further work to identify the organelle and
enzymes involved in the oxidation of MRL-A is ongoing.
Our initial attempt to identify the target protein(s) of MRL-
A by SDS−PAGE was unsuccessful as we did not observe any
specific labeled protein band. This was most likely due to the
Figure 11. Metabolic scheme of the bioactivation of MRL-A in rat
liver microsomes.
conjugated system in A1 keeps the carboxylate coplanar with
the α,β-unsaturated imine and prevents it from being involved
in a similar fashion. The involvement of P450 enzymes in the
formation of A1 in vivo was also ruled out by predosing the
nonspecific P450 inhibitor ABT with MRL-A that prevented
neither irreversible binding of MRL-A to proteins in vivo nor
formation of adduct A1. However, the bioactivation pathway
observed in microsomes cannot be entirely ruled out in vivo
Figure 12. Proposed pathway for the metabolic activation of MRL-A via β-oxidation, and formation of protein adduct A1.
1420
dx.doi.org/10.1021/tx300017s | Chem. Res. Toxicol. 2012, 25, 1412−1422

Chemical Research in Toxicology
Article
instability of adduct A1 at the higher pH (8.8) of the resolving
gel. Direct trypsinization of liver homogenate protein pellets
and fractionation of the resulting peptides by HPLC were more
successful, and only one major radioactivity-bearing peptide
was detected. The target protein, ACSL1, belongs to a subclass
of acyl-CoA synthetases consisting of five different isoforms
(ACSL1 and -3−6).19,20 They primarily catalyze the ATP-
dependent two-step activation of saturated and unsaturated
fatty acids of 14−22 carbons and are necessary for important
biochemical processes such as β-oxidation and the synthesis of
triacylglycerol and phospholipids. Of note, ACSL1 has also
been shown to catalyze the formation of CoA thioesters of
carboxylic acid-containing drugs such as tolmetin, nafenopine,
and ibuprofen.21,22 ACSL1 is the most abundant isoform
expressed in liver and has been detected in the cytosol,
microsomes, mitochondria, and mitochondrion-associated
membranes.23,24 The cytosolic protein, however, was found to
be poorly active compared to the microsomal enzyme, and this
fact could explain the relatively weaker covalent binding of
radiolabeled MRL-A observed in the cytosol. It has been
demonstrated recently that Lys676 is essential for the catalytic
activity of ACSL125 and is part of a highly conserved KxxK/R
motif that is present in all long chain acyl CoA synthetases20
(Figure S10 of the Supporting Information). This motif has
been proposed to be equivalent to the A10 motif found in other
members of the acyladenylate synthetase superfamily.26 The
A10 motif contains an essential lysine that is involved in
substrate binding during the first adenylation step and is in
contact with both the ATP ribose ring oxygen and one of the
carboxylate oxygens of the substrate. Collectively, these
observations would indicate that ACSL1 recognizes the reactive
α,β-unsaturated aldehyde I2 as a substrate and that while
Lys676 is positioning ATP and the acyl group for nucleophilic
attack, it is also in the proximity of the aldehyde moiety and
reacts with it to form adduct A1. Mutation of Lys676 to alanine,
glutamine, or aspartic acid abolishes enzymatic activity,25 and it
is thus expected that formation of A1 inactivates ACSL1.
On the basis of mRNA levels, ACSL1 appears to be strongly
expressed in liver, heart, and adipose tissues24 of rodents, and
recent studies have highlighted the possible biological
consequences of ACSL1 inactivation in these tissues. Liver
specific deletion of ACSL1 in mice led to a 50% decrease in
hepatic ACSL activity and impairment of triacylglycerol
synthesis and β-oxidation.27 However, there were no changes
in liver morphology and weights or plasma lipids, suggesting
the possible induction of compensatory mechanisms. In whole-
body Acsl1-null mice, the ACSL activity in the heart was
reduced 90%, which resulted in a marked impairment of fatty
acid oxidation and in ventricular hypertrophy.28 Cardiac
function was unchanged during the 2 month period of the
study, but the long-term effect of loss of ACSL1 in the heart
remains to be more fully assessed. Finally, mice deficient in
ACSL1 in adipose tissues had an 80% drop in total ACSL
activity and an accompanying loss of β-oxidation.29 Because of
the regulation of thermogenesis by fatty acid oxidation, these
mice displayed pronounced cold intolerance. Within 4 h of
exposure to a temperature of 4 °C, the body temperature of the
knockout mice decreased to 30 °C. In contrast, the body
temperature of control mice remained unchanged up to 24 h.
Human and rat ACSL1 are highly homologous (86%
identical and 94% similar), and the sequence of the homologue
of peptide P1 in humans differs only by a methionine at
position 675. There is therefore a potential for MRL-A to
covalently modify and inactivate ACSL1 in humans. We have
found that the protein adduct profile obtained in cultured rat
hepatocytes is qualitatively similar to the one observed in rat
liver in vivo, and this observation suggests that cultured
hepatocytes could be used to assess the formation of A1 in
humans. Interestingly, in both suspended and cultured human
hepatocytes, A6, the product of cysteine thioesterification by
MRL-A, was the major adduct detected. In contrast, A1 was
observed in only minor amounts, suggesting that, in humans,
modification of ACSL1 may not occur to the extent observed in
rats. However, it would appear that, instead, the CoA thioester
of MRL-A (I6) preferentially modifies cysteine residues, a
pathway potentially not adequately covered in rats for
toxicological assessment. Attempts to identify the protein
target of MRL-A in human hepatocytes have thus far been
unsuccessful.
In conclusion, the work presented here describes a novel
bioactivation pathway of the 3-caboxyazetidine moiety,
previously not known to undergo metabolic activation. The
reaction is most likely initiated by ACSL1, and although the
subsequent steps are catalyzed by enzymatic systems that are
located in different cellular sites, the final reactive intermediate
appears to again be recognized by ACSL1, which is alkylated
through binding to a specific lysine residue. The identification
of the bioactivation mechanism and protein target in vivo
provides an understanding of the potential safety issues
associated with chronic dosing of MRL-A and affords guidance
to the discovery team in its effort to synthesize new analogues
devoid of such risks.
ASSOCIATED CONTENT
■
S
* Supporting Information
Additional figures. This material is available free of charge via
the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
■
Corresponding Author
*Department of Pharmacokinetics, Pharmacodynamics, and
Drug Metabolism, Merck Research Laboratories, P.O. Box
2000, Rahway, NJ 07065. Telephone: (732) 594-6363. E-mail:
richard_tschirretguth@merck.com.
Present Address
§Eli Lilly and Co., Indianapolis, IN 46285.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We thank Dr. Conrad Raab and Dr. Ashok Chaudhary from
Merck Research Laboratories for the synthesis of radiolabeled
MRL-A and Prof. Thomas Baillie (University of Washington,
Seattle, WA), Prof. M. W. Anders (University of Rochester,
Rochester, NY), Prof. Paul Van Veldhoven (Katholieke
Universiteit Leuven, Leuven, Belgium), and Prof. Rosalind
Coleman (University of North Carolina, Chapel Hill, NC) for
stimulating discussions about this research.
ABBREVIATIONS
■
ABT, 1-aminobenzotriazole; ACSL, long chain acyl-CoA
synthetase; BDC, bile duct-cannulated; BCA, bicinchoninic
acid; CID, collision-induced dissociation; DILI, drug-induced
liver injury; HMBC, heteronuclear multiple-bond correlation;
1421
dx.doi.org/10.1021/tx300017s | Chem. Res. Toxicol. 2012, 25, 1412−1422

Chemical Research in Toxicology
Article
(17) Hiltunen, J. K., and Qin, Y. M. (2000) β-Oxidation: Strategies
for the metabolism of a wide variety of acyl-CoA esters. Biochim.
Biophys. Acta 1484, 117−128.
HRMS, high-resolution mass spectroscopy; NOESY, nuclear
Overhauser effect spectroscopy; SPE, solid phase extraction.
(18) Houten, S. M., and Wanders, R. J. (2010) A general
introduction to the biochemistry of mitochondrial fatty acid β-
oxidation. J. Inherited Metab. Dis. 33, 469−477.
(19) Soupene, E., and Kuypers, F. A. (2008) Mammalian long-chain
acyl-CoA synthetases. Exp. Biol. Med. 233, 507−521.
ADDITIONAL NOTE
■
a
The following nomenclatiure is used: M, metabolite; A, amino
acid adduct; I, reaction intermediate; P, peptide adduct.
(20) Watkins, P. A., Maiguel, D., Jia, Z. Z., and Pevsner, J. (2007)
Evidence for 26 distinct acyl-coenzyme A synthetase genes in the
human genome. J. Lipid Res. 48, 2736−2750.
(21) Brugger, R., Reichel, C., Alia, B. G., Brune, K., Yamamoto, T.,
Tegeder, I., and Geissinger, G. (2001) Expression of rat liver long-
chain acyl-CoA synthetase and characterization of its role in the
metabolism of R-ibuprofen and other fatty acid-like xenobiotics.
Biochem. Pharmacol. 61, 651−656.
(22) Sevoz, C., Benoit, E., and Buronfosse, T. (2000) Thioester-
ification of 2-arylpropionic acids by recombinant acyl-coenzyme A
synthetases (ACS1 and ACS2). Drug Metab. Dispos. 28, 398−402.
(23) Lewin, T. M., Kim, J. H., Granger, D. A., Vance, J. E., and
Coleman, R. A. (2001) Acyl-CoA synthetase isoforms 1, 4, and 5 are
present in different subcellular membranes in rat liver and can be
inhibited independently. J. Biol. Chem. 276, 24674−24679.
(24) Mashek, D. G., Li, L. O., and Coleman, R. A. (2006) Rat long
chain acyl-CoA synthetase mRNA, protein, and activity vary in tissue
distribution and in response to diet. J. Lipid Res. 47, 2004−2010.
(25) Frahm, R. A., Lei, O. L., Grevengoed, T. J., and Coleman, R. A.
(2011) Phosphorylation and acetylation of acyl-CoA synthetase-1. J.
Proteomics Bioinf. 4, 129−137.
(26) Gulick, A. M. (2009) Conformational dynamics in the acyl-CoA
synthetases, adenylation domains of non-ribosomal peptide synthe-
tases, and firefly luciferase. Chem. Biol. 4, 811−827.
(27) Li, L. O., Ellis, J. M., Paich, H. A., Wang, S. L., Gong, N.,
Altshuller, G., Thresher, R. J., Koves, T. R., Watkins, S. M., Muoio, D.
M., Cline, G. W., Shulman, G. I., and Coleman, R. A. (2009) Liver-
specific loss of long chain acyl-CoA synthetase-1 decreases
triacylglycerol synthesis and β-oxidation and alters phospholipid fatty
acid composition. J. Biol. Chem. 284, 27816−27826.
(28) Ellis, J. M., Mentock, S. M., DePetrillo, M. A., Koves, T. R., Sen,
S., Watkins, S. M., Muoio, D. M., Cline, G. W., Taegtmeyer, H.,
Shulman, G. I., Willis, M. S., and Coleman, R. A. (2011) Mouse
cardiac acyl coenzyme A synthetase-1 deficiency impairs fatty acid
oxidation and induces cardiac hypertrophy. Mol. Cell. Biol. 31, 1252−
1262.
(29) Ellis, J. M., Li, L. O., Wu, P. C., Koves, T. R., Ilkayeva, O.,
Stevens, R. D., Watkins, S. M., Muoio, D. M., and Coleman, R. A.
(2010) Adipose acyl-CoA synthetase-1 directs fatty acids toward β-
oxidation and is required for cold thermogenesis. Cell Metab. 12, 53−
64.
REFERENCES
■
(1) Liebler, D. C. (2008) Protein damage by reactive electrophiles:
Targets and consequences. Chem. Res. Toxicol. 21, 117−128.
(2) Park, B. K., Kitteringham, N. R., Maggs, J. L., Pirmohamed, M.,
and Williams, D. P. (2005) The role of metabolic activation in drug-
induced hepatotoxicity. Annu. Rev. Pharmacol. Toxicol. 45, 177−202.
(3) Kumar, S., Kassahun, K., Tschirret-Guth, R. A., Mitra, K., and
Baillie, T. A. (2008) Minimizing metabolic activation during
pharmaceutical lead optimization: Progress, knowledge gaps and
future directions. Curr. Opin. Drug Discovery Dev. 11, 43−52.
(4) Park, B. K., Boobis, A., Clarke, S., Goldring, C. E. P., Jones, D.,
Kenna, J. G., Lambert, C., Laverty, H. G., Naisbitt, D. J., Nelson, S.,
Nicoll-Griffith, D. A., Obach, R. S., Routledge, P., Smith, D. A.,
Tweedie, D. J., Vermeulen, N., Williams, D. P., Wilson, I. D., and
Baillie, T. A. (2011) Managing the challenge of chemically reactive
metabolites in drug development. Nat. Rev. Drug Discovery 10, 292−
306.
(5) Walsh, J. S., and Miwa, G. T. (2011) Bioactivation of Drugs: Risk
and drug design. Annu. Rev. Pharmacol. Toxicol. 51, 145−167.
(6) Kumar, S., Mitra, K., Kassahun, K., and Baillie, T. A. (2010)
Approaches for minimizing metabolic activation of new drug
candidates in drug discovery. Handb. Exp. Pharmacol. 196, 511−544.
(7) Hale, J. J., Lynch, C. L., Neway, W., Mills, S. G., Hajdu, R.,
Keohane, C. A., Rosenbach, M. J., Milligan, J. A., Shei, G. J., Parent, S.
A., Chrebet, G., Bergstrom, J., Card, D., Ferrer, M., Hodder, P.,
Strulovici, B., Rosen, H., and Mandala, S. (2004) A rational utilization
of high-throughput screening affords selective, orally bioavailable 1-
benzyl-3-carboxyazetidine sphingosine-1-phosphate-1 receptor ago-
nists. J. Med. Chem. 47, 6662−6665.
(8) Anari, M. R., Creighton, M. D., Ngui, J. S., Tschirret-Guth, R. A.,
Teffera, Y., Doss, G. A., Tang, W., Yu, N. X., Ciccotto, S. L., Hobra, D.
F., Coleman, J. B., Vincent, S. H., and Evans, D. C. (2006) Species
differences in metabolism and pharmacokinetics of a sphingosine-1-
phosphate receptor agonist in rats and dogs: Formation of a unique
glutathione adduct in the rat. Drug Metab. Dispos. 34, 1367−1375.
(9) Raucy, J. L., and Lasker, J. M. (1991) Isolation of P450-enzymes
from human liver. Methods Enzymol. 206, 577−586.
(10) Tirmenstein, M. A., and Nelson, S. D. (1989) Subcellular
binding and effects on calcium homeostasis produced by acetamino-
phen and a nonhepatotoxic regioisomer, 3′-hydroxyacetanilide, in
mouse liver. J. Biol. Chem. 264, 9814−9819.
(11) Strelevitz, T. J., Foti, R. S., and Fisher, M. B. (2006) In vivo use
of the P450 inactivator 1-aminobenzotriazole in the rat: Varied dosing
route to elucidate gut and liver contributions to first-pass and systemic
clearance. J. Pharm. Sci. 95, 1334−1341.
(12) Boelsterli, U. A. (2002) Xenobiotic acyl glucuronides and acyl
CoA thioesters as protein-reactive metabolites with the potential to
cause idiosyncratic drug reactions. Curr. Drug Metab. 3, 439−450.
(13) Regan, S. L., Maggs, J. L., Hammond, T. G., Lambert, C.,
Williams, D. P., and Park, B. K. (2010) Acyl Glucuronides: The good,
the bad and the ugly. Biopharm. Drug Dispos. 31, 367−395.
(14) Skonberg, C., Olsen, J., Madsen, K. G., Hansen, S. H., and
Grillo, M. P. (2008) Metabolic activation of carboxylic acids. Expert
Opin. Drug Metab. Toxicol. 4, 425−438.
NOTE ADDED AFTER ASAP PUBLICATION
■
This paper was published on the Web on July 5, 2012. The
caption of Figure 4 has been updated. The corrected version
was reposted on July 16, 2012.
(15) Van Veldhoven, P. P., and Mannaerts, G. P. (1999) Role and
organization of peroxisomal β-oxidation. In Current Views of Fatty Acid
Oxidation and Ketogenesis (Quant, P. A. and Eaton, S., Eds.) pp 261−
272, Kluwer Academic/Plenum, New York.
(16) Poirier, Y., Antonenkov, V. D., Glumoff, T., and Hiltunen, J. K.
(2006) Peroxisomal β-oxidation: A metabolic pathway with multiple
functions. Biochim. Biophys. Acta 1763, 1413−1426.
1422
dx.doi.org/10.1021/tx300017s | Chem. Res. Toxicol. 2012, 25, 1412−1422