Electrophilicity of pyridazine-3-carbonitrile, pyrimidine-2-carbonitrile, and pyridine-carbonitrile derivatives: A chemical model to describe the formation of thiazoline derivatives in human liver microsomes

Article abstract of DOI:10.1021/tx500256j

(Chemical Equation Presented) Certain aromatic nitriles are well-known inhibitors of cysteine proteases. The mode of action of these compounds involves the formation of a reversible or irreversible covalent bond between the nitrile and a thiol group in th

Full text of DOI:10.1021/tx500256j

Article
pubs.acs.org/crt
Electrophilicity of Pyridazine-3-carbonitrile, Pyrimidine-2-carbonitrile,
and Pyridine-carbonitrile Derivatives: A Chemical Model To Describe
the Formation of Thiazoline Derivatives in Human Liver Microsomes
Sarmistha Sinha,† Deepak Ahire,† Santosh Wagh,† Dibakar Mullick,‡ Ramesh Sistla,§ Kumaravel Selvakumar,‡
Janet Caceres Cortes,∥ Siva Prasad Putlur,† Sandhya Mandlekar,⊥ and Benjamin M. Johnson*,#
‡
§
†Pharmaceutical Candidate Optimization, Medicinal Chemistry Department, and Advanced Biotechnology Department,
Biocon Bristol-Myers Squibb R&D Center (BBRC), Syngene International Ltd, Plot No. 2 & 3, Bommasandra IV Phase,
Jigani Link Road, Bangalore 560100, India
∥Department of Pharmaceutical Candidate Optimization, Bristol-Myers Squibb Co., Route 206 & Province Line Road, Princeton,
New Jersey 08543, United States
⊥Pharmaceutical Candidate Optimization, Bristol-Myers Squibb India Ltd. BBRC, Bangalore 560100, India
#Department of Pharmaceutical Candidate Optimization, Bristol-Myers Squibb Co., 5 Research Parkway, Wallingford,
Connecticut 06492, United States
S
* Supporting Information
ABSTRACT: Certain aromatic nitriles are well-known inhibitors of cysteine proteases. The mode of action of these compounds
involves the formation of a reversible or irreversible covalent bond between the nitrile and a thiol group in the active site of the
enzyme. However, the reactivity of these aromatic nitrile-substituted heterocycles may lead inadvertently to nonspecific
interactions with DNA, protein, glutathione, and other endogenous components, resulting in toxicity and complicating the use
of these compounds as therapeutic agents. In the present study, the intrinsic reactivity and associated structure−property
relationships of cathepsin K inhibitors featuring substituted pyridazines [6-phenylpyridazine-3-carbonitrile, 6-(4-fluorophenyl)-
pyridazine-3-carbonitrile, 6-(4-methoxyphenyl)pyridazine-3-carbonitrile, 6-p-tolylpyridazine-3-carbonitrile], pyrimidines [5-p-
tolylpyrimidine-2-carbonitrile, 5-(4-fluorophenyl)pyrimidine-2-carbonitrile], and pyridines [5-p-tolylpicolinonitrile and 5-(4-
fluorophenyl)picolinonitrile] were evaluated using a combination of computational and analytical approaches to establish correla-
tions between electrophilicity and levels of metabolites that were formed in glutathione- and N-acetylcysteine-supplemented
human liver microsomes. Metabolites that were characterized in this study featured substituted thiazolines that were formed
following rearrangements of transient glutathione and N-acetylcysteine conjugates. Peptidases including γ-glutamyltranspeptidase
were shown to catalyze the formation of these products, which were formed to lesser extents in the presence of the selective
γ-glutamyltranspeptidase inhibitor acivicin and the nonspecific peptidase inhibitors phenylmethylsulfonyl fluoride and aprotinin.
Of the chemical series mentioned above, the pyrimidine series was the most susceptible to metabolism to thiazoline-containing
products, followed, in order, by the pyridazine and pyridine series. This trend was in keeping with the diminishing electrophilicity
across these series, as demonstrated by in silico modeling. Hence, mechanistic insights gained from this study could be used to assist
a medicinal chemistry campaign to design cysteine protease inhibitors that were less prone to the formation of covalent adducts.
been revealed by studies of the bioactivation of xenobiotics and
endogenous compounds.1−7,9 Such compounds may represent a
INTRODUCTION
■
Compounds that contain electrophilic functional groups, like
alkyl halides and Michael acceptors, are prone to react in vivo and
are usually avoided in drug design for safety reasons. Much of
what is known about the effects of such reactive compounds has
Received: June 30, 2014
Published: November 5, 2014
© 2014 American Chemical Society
2052
dx.doi.org/10.1021/tx500256j | Chem. Res. Toxicol. 2014, 27, 2052−2061

Chemical Research in Toxicology
Article
Figure 1. Nitrile-substituted heterocyclic compounds evaluated in this study.
burden for living organisms because of the potential for covalent
modification of essential cellular molecules like DNA and
proteins, and, in certain cases, the formation of these com-
plexes can lead to adverse events. Glutathione (GSH), a thiol-
containing tripeptide that is present in cells at concentrations of
1−10 mM, plays an important role in protecting cells by trapping
reactive electrophiles in the form of stable metabolites that can be
detected readily using LC-MS.4−6 For this reason, screening for
GSH conjugates of new chemical entities in vitro is done regularly
to help assess the suitability of compounds as drug candidates.7−9
Despite this practice, the pharmaceutical industry maintains
a considerable interest in the discovery and development of
covalent inhibitors: drugs that contain an electrophilic group that
can react either reversibly or irreversibly with a nucleophile in the
active site of a target protein. This approach has a long history
that extends back to the discovery of aspirin and includes several
of the world’s top-selling medicines such as omeprazole and
clopidogrel.8 Additionally, chemotypes that feature electrophilic
moieties have been explored as therapeutic agents in the areas of
osteoporosis and osteoarthritis, where they act as inhibitors
of cysteine proteases. For example, several electrophilic nitrile-
containing compounds have been identified as inhibitors of
cathepsin K, and, among these, three chemotypes have been
explored extensively: cyanamides, nitrile-appended aromatic
heterocycles, and aminoacetonitriles.11,16 The mode of action
of these compounds involves the formation of a reversible or
irreversible covalent bond between the nitrile and a cysteine
residue within the catalytic site.10−17 A computational approach
was devised by Oballa et al. to calculate the theoretical reactivity
of such compounds with a cysteine residue, demonstrating that
reactivity was governed by the intrinsic electrophilicity of the
aromatic nitrile-substituted heterocycles.11 However, such
compounds that exhibit biological activity against cysteine
proteases may also have the potential to react with endogenous
thiols and other nucleophiles like proteins, peptides, and
DNA.18,19 Therefore, in vitro risk assessment tools are essential
for nitriles that might have the potential for reactivity as a result
of their connectivity to neighboring chemical groups.
To this end, Macfaul et al.12 described a simple method to
assess potential drug-related liabilities by measuring the chemical
reactivity of nitriles in phosphate buffer, using glutathione and
cysteine as trapping agents. This method did not account for
the contributions of drug metabolizing enzymes, which might
influence the rate of conjugate formation and improve (or detract
from) the utility of the model as a risk-assessment tool. The
present study represents an extension of Macfaul et al.’s work,
where nitrile-substituted pyridazines [6-phenylpyridazine-3-
carbonitrile (1), 6-(4-fluorophenyl)pyridazine-3-carbonitrile
(2), 6-(4-methoxyphenyl)pyridazine-3-carbonitrile (3), 6-p-
tolylpyridazine-3-carbonitrile (4)], pyrimidines [5-p-tolylpyr-
imidine-2-carbonitrile (5), 5-(4-fluorophenyl)pyrimidine-2-car-
bonitrile (6)], and pyridines [5-p-tolylpicolinonitrile (7) and
5-(4-fluorophenyl)picolinonitrile (8)] (Figure 1) were incubated
in GSH- and N-acetylecysteine (NAC)-supplemented human
liver microsomes, and the roles of hepatic enzymes in the forma-
tion of metabolites were characterized. Furthermore, the reacti-
vity toward GSH and NAC and associated structure−property
relationships of these substituted heterocycles were explored in
the context of a physiologically relevant system, and the utility of
the model as a risk-assessment tool was evaluated and interpreted
in the context of computational assessments of substrate
electrophilicity. The substrates that were selected for this study
included 2-cyanopyrimidines, a series that has been explored in
the design of cathepsin K inhibitors,10−17 and two similar series
based on cyano-substituted pyridazine and pyridine motifs.11,16
It was hypothesized that the formation of GSH- and
NAC-derived conjugates would depend on the electrophilicity
of the corresponding substrates and might also be influenced by
hepatic enzymes. This hypothesis was evaluated by estimating
the electrophilicity of the cyano carbon atom of different ligands
and by calculating the free energy of formation of GSH-related
conjugates using quantum chemical methods and then
2053
dx.doi.org/10.1021/tx500256j | Chem. Res. Toxicol. 2014, 27, 2052−2061

Chemical Research in Toxicology
Article
tube was sealed with a Teflon screw cap and stirred at 80 °C for 6 h.
The reaction mixture was cooled to room temperature, water (10 mL)
was added, and the mixture was extracted with ethyl acetate (2 ×
10 mL). The combined organic layer was washed with brine (10 mL),
dried over Na2SO4, filtered, and concentrated. The crude product was
purified by flash chromatography to afford the desired compound as a
white solid with 77% yield. The MS and 1H NMR (400 MHz, CDCl3)
values of 2 were as follows: [M + H] = 202; H NMR (400 MHz,
CDCl3): δ 8.16 (m, 2H), 7.96 (d, J = 8.8 Hz, 1H), 7.87 (d, J = 8.8 Hz,
1H), 7.27 (m, 2H). The rest of the compounds, 1 and 3−8, were pre-
pared in a similar fashion (Scheme 2). Analytical data of these com-
pounds are provided as Supporting Information.
comparing the computational and experimental results to each
other. Additional experiments to characterize the role of enzymes
in these reactions, to measure disappearance kinetics, and to
define the origin of the GSH- and NAC-derived moieties were
also carried out. As a complicating factor in such studies, aliphatic
and aromatic nitriles are known to react chemically with cysteine
and then rearrange to form thiazole derivatives (Scheme 1).11
+
1
Scheme 1. Reaction between a Nitrile-Containing Compound
and Cysteine To Form a Thiazoline Derivative
General in Vitro Incubations. Each compound, 1−8 and
4-methoxybenzonitrile (negative control, 30 μM), was incubated with
human liver microsomes (1 mg/mL) and GSH (3 mM) in sodium
phosphate buffer (100 mM; pH 7.4). The total reaction volume was
500 μL. The mixture was incubated for 60 min at 37 °C at 60 rpm in an
orbital incubator shaker (Innova 40, New Brunswick Scientific, CT,
USA). Similarly, each compound, 1−8 and 4-methoxybenzonitrile
(negative control), was incubated with GSH (3 mM) in sodium
phosphate buffer (100 mM; pH 7.4) without human liver microsomes
for 60 min at 37 °C at 60 rpm. Control experiments were performed in a
similar fashion without the compounds. At the 0 and 60 min time points,
the reaction was terminated by adding a 200 μL aliquot from the
incubation to an equal volume of acetonitrile. The reaction mixture was
vortex-mixed and centrifuged at 20 827g for 10 min. The resulting
supernatants were analyzed by LC-MS/MS. Additionally, each
compound, 1−8 (30 μM), was incubated with NAC (3 mM) with
and without human liver microsomes (1 mg/mL) in a similar fashion as
described above for subsequent LC-MS/MS analysis.
Compound 6 (30 μM) was incubated with stable-labeled GSH
(glycine-13C2,15N; 3 mM) in sodium phosphate buffer (100 mM;
pH 7.4) with human liver microsomes (1 mg/mL) in a total reaction vol-
ume of 500 μL. The incubations and sample processing were carried out
in a similar fashion as described above, and samples were analyzed by
LC-MS/MS.
As an additional control for enzyme activity, human liver microsomes
were heated at 80 °C for 5 min and then incubated with 6 (30 μM) and
GSH (3 mM). The incubation conditions, time, and sample processing
methods were identical to those used in the microsomal incubations
described above. Samples were analyzed using LC-MS/MS.
For example, the formation of thiazole derivatives in vivo from
aliphatic nitriles such as odanacatib, a cathepsin K inhibitor,20 has
been reported. To help characterize the reactivity of these nitrile-
substituted heterocycles further, the mechanism of formation of
GSH-derived thiazoline derivatives was investigated using 6 as a
model substrate.
EXPERIMENTAL PROCEDURES
■
Materials. Arylboronic acids, 4-methoxybenzonitrile, and Cs2CO3
were purchased from Aldrich Chemical Co. (Bangalore, India).
5-Bromopicolinonitrile and PdCl2(dppf)−CH2Cl2 complex were
purchased from Alfa Aesar (Hyderabad, India). 5-Bromopyrimidine-2-
carbonitrile was purchased from OChem Incorporation (Chicago, IL,
USA), 8-oxobenzyl guanine was purchased from Toronto Research
Chemicals (Toronto, Canada), and 6-chloropyridazine-3-carbonitrile
was prepared using a reported procedure.21−23 Human liver microsomes
were purchased from Invitrogen Bioservices India Pvt. Ltd. Dimethyl-
formamide (DMF), arylboronic acid, cesium carbonate (Cs2CO3), glu-
tathione (GSH), stable-isotope-labeled glutathione (glycine-13C2,15N),
cysteine, acivicin, phenylmethanesulfonyl fluoride (PMSF), and
aprotinin were purchased from Sigma-Aldrich (Bommasandra Industrial
Area, Bangalore, India). HPLC grade acetonitrile and formic acid were
purchased from Merck Specialties Private Ltd. (Worli, Mumbai, India).
Preparation of 6-(4-Fluorophenyl)pyridazine-3-carbonitrile.
Compound 2 was synthesized by a coupling reaction between
6-chloropyridazine-3-carbonitrile and (4-fluorophenyl)boronic acid.
An oven-dried Schlenk tube under argon was charged with 6-chloro-
pyridazine-3-carbonitrile (80.0 mg, 0.573 mmol), cesium carbonate
(374 mg, 1.15 mmol), (4-fluorophenyl)boronic acid (88.0 mg, 0.631
mmol), PdCl2(dppf)CH2Cl2 (41.9 mg, 0.0570 mmol), DMF (1 mL),
and water (0.1 mL). The solution was purged with argon, and the
In Vitro Incubations with Inhibitors. To a microsomal suspension
(1 mg/mL) in sodium phosphate buffer (100 mM; pH 7.4), either
acivicin24−26 or a mixture of PMSF and aprotinin27 was added to give
a final concentration of 5 mM of each, and the suspensions were
preincubated for 20 min. Incubations were initiated with the addition of
6 and GSH to a final concentration of 30 μM. Control experiments were
performed in a similar fashion without the inhibitors. Incubations and
sample processing were conducted as described under General in Vitro
Incubations, and samples were analyzed using LC-MS/MS.
Kinetics Studies. Compounds 2, 6, and 8 were selected as repre-
sentatives from each series for an assessment of kinetics. Each com-
pound (1 μM) was incubated with cysteine (3 mM) in 100 mM
Scheme 2. Synthetic Scheme Illustrating the Preparation of Carbonitriles Featuring Pyridazine, Pyrimidine, and Pyridine Moieties
2054
dx.doi.org/10.1021/tx500256j | Chem. Res. Toxicol. 2014, 27, 2052−2061

Chemical Research in Toxicology
Article
phosphate buffer (pH 7.4) in duplicate. The total reaction volume was
1 mL. The mixture was incubated for 60 min at 37 °C at 60 rpm. At each
time point (0, 2.5, 5, 15, 30, 45, and 60 min), a 50 μL aliquot was
withdrawn from the reaction mixture and treated with 150 μL
acetonitrile containing 8-oxobenzyl guanine (1 μM) as an internal
standard. The reaction mixture was vortex mixed and centrifuged at
20 827g for 10 min. The resulting supernatant was analyzed by LC-MS.
LC-MS/MS Analysis. Liquid chromatography-tandem mass spec-
trometry (LC-MS/MS) was carried out using an Agilent 1100 HPLC
(Agilent Technologies Inc., Wilmington, DE, USA) connected in-line to
an Agilent photodiode array detector and an LTQ Velos Orbitrap mass
spectrometer (Thermo Fisher Scientific, Bremen, Germany) fitted with
an electrospray ionization source and operated in positive-ion mode.
Chromatographic separations were performed on a Thermo Hypersil
Gold C18 (150 × 3 mm; 5 μm) column maintained at 40 °C. The
mobile phase consisted of solvent A (0.1% formic acid in water) and
solvent B (0.1% formic acid in acetonitrile) delivered at a flow rate of
0.4 mL/min with a linear gradient as follows: 2 to 60% B over 10 min,
60% B isocratic for 20 min, 98% B isocratic for 2 min, and 2% isocratic
for 8 min. For electrospray mass spectrometry, nitrogen was used as
both the sheath gas and auxiliary gas, and helium was used for collision-
induced dissociation. The capillary temperature and source voltage were
set to 300 °C and 5 kV, respectively. The Orbitrap was operated in
Fourier transform mass spectrometry mode to enable the acquisition of
both full-scan mass spectra and product-ion MS/MS spectra with mass
accuracies within 5 ppm. A typical data-dependent MS method included
three scan events: (1) full-scan MS over the range of m/z 100−800, (2)
MS2 of the most intense ions exceeding 1000 cps in the previous full-
scan mass spectrum, and (3) MS3 of the most intense product ions from
the MS2 scan. Dynamic exclusion criteria included mass widths of
5 ppm, a repeat count of 2, and a repeat duration of 6 s. The MS2 and
MS3 normalized collision energies were ramped over the range of
35−45%. Data acquisition and processing were carried out using
Xcalibur software (version 1.3, Thermo Fisher Scientific).
Isolation of NAC Conjugate of 6 for NMR. Compound 6
(∼150 μg) was incubated in rat liver microsomes (1 mg/mL) supple-
mented with NAC (3 mM) for 60 min at 37 °C at 60 rpm in an orbital
shaker. The total reaction volume was 15 mL. The reaction was termi-
nated by adding an equal volume of cold acetonitrile. Then, the reaction
mixture was vortexed and centrifuged at 20 827g for 10 min. The result-
ing supernatants were concentrated to 1−1.5 mL under nitrogen using a
TurboVap (Caliper Life Sciences, Hopkinton, MA, USA) maintained at
25 °C. During evaporation, acetonitrile was added three to four times
(1:2/1:3 v/v) to remove excess protein. Finally, the concentrated
sample was subjected to HPLC as described in the LC-MS section, and
the NAC conjugate was isolated using a fraction collector (Gilson Inc.,
Middleton, WI, USA). All fractions containing the desired NAC adduct
were pooled, and the solvent was evaporated to dryness under nitrogen
in a TurboVap. The purified sample containing the NAC adduct was
analyzed by NMR.
was low and it represented a closer approximation to GSH. The free
energies of both the reactants and the products were calculated in the gas
phase with B3LYP/6-31G(**) basis sets. The Jaguar 8.0 module from
the Schrodinger suite was used for these calculations with the accuracy
̈
level of calculations set to fully analytic. The approximate free energy of
formation (ΔG) was calculated from the gas-phase energies (G) as
follows: ΔG = Gproduct − G(cys+ligand). The electrophilicity of the carbon
atom in the cyano group on different ligands was also estimated by
quantum chemical optimization of the ligands using density functional
theory with B3LYP/6-31G(**) basis sets. The Jaguar 8.0 module from
the Schrodinger suite was used for these calculations. Subsequent to the
̈
optimization, the Fukui functions were calculated. Fukui functions,
f+ and f−, indicate the electrophilicity and nucleophility potential of
atoms, respectively.28
RESULTS
■
Detection of Conjugates in Human Liver Microsomes
and Structure Elucidation of 6A and 6B Based on HRMS.
In human liver microsomes, two GSH-derived conjugates of 1−8
(denoted 1A, 1B, 2A, etc., where the number corresponds to the
substrates) were detected by UV and MS (Table 1 and Figures 2
and 3A for 6; Supporting Information for the remaining
compounds). The masses of all GSH-derived conjugates A and
B were 161.0146 and 103.9932 u higher than their respective
parent compounds 1−8, respectively. Moreover, the retention
time, exact mass, and product-ion spectrum of each NAC-derived
conjugate was identical to that of its corresponding GSH-derived
metabolite B, suggesting that one of the metabolites produced
during reactions with both GSH and NAC was the same. In
phosphate buffer that was supplemented with either GSH or
NAC, but not human liver microsomes, all compounds were still
converted to their respective metabolites A and B, respectively;
however, the abundance of these products was lower in this
system. No conjugates were found in controls prepared using
4-methoxybenzonitrile as a substrate or in incubations where
no substrates were present. Trace quantities of these conju-
gates were detected in some of the samples that were withdrawn
at the 0 min time point. In liver microsomes, the formation
of these conjugates was independent of the P450 cofactor
NADP(H).
The MS/MS product-ion spectra of 6 and its metabolites 6A
and 6B are shown in Figure 3A (extracted-ion chromatograms
and MS/MS spectra of the remaining compounds and their
metabolites are shown in the Supporting Information). The
product ions of m/z 343 and 315 were produced by water loss
and decarboxylation from the glycine moiety, respectively. The
fragment ion of m/z 258 resulted from the cleavage of carbo-
xamide bond and corresponded to fluorophenylpyrimidine-
thiazoline moiety. The product ion of m/z 200 resulted from
cleavage of the thiazoline ring and loss of acetamidoacetic acid
moiety, a fragmentation pathway that was also observed during
an MS/MS product-ion scan of the parent compound, 6. The
product ion spectra of metabolites A and B of other compounds
(Supporting Information) exhibited similar fragmentation
patterns to those of 6A and 6B. To characterize its structure
further, 6B was prepared on a larger scale, isolated, and analyzed
by 1H, 13C, and COSY NMR (Figure 3B and Table 2).
Nuclear Magnetic Resonance (NMR). The isolated metabolite
and parent samples were dissolved in DMSO-d6 (50 μL) and placed in a
1.7 mm NMR tube. The samples were analyzed using a Bruker Avance
600 MHz NMR spectrometer equipped with a 1.7 mm TCI triple
resonance cryoprobe (Bruker, Billerica, MA, USA) at 27 °C. Proton
spectra were recorded using the single pulse (zg30) sequence and
multiple solvent suppression (lc1pncwps). 13C chemical shift data were
obtained from a one-dimensional carbon spectrum of the parent and
heteronuclear single quantum coherence (HSQC) and multiple-bond
correlation spectroscopy (HMBC) of the metabolite. Peak assignments
1
1
were made based on H, H−1H gradient correlation spectroscopy
(COSY), 1H−13C-edited DEPT HSQC, and HMBC spectroscopic anal-
ysis. ACD/NMR prediction and processing software from Advanced
Chemistry Development Inc. (Toronto, ON, Canada) was used to help
interpret the NMR data.
Characterization of 6B by NMR. The 1H and 13C chemical
shifts of both 6 and 6B are listed in Table 2. Comparison of these
spectra revealed small chemical shift differences for the protons
of the fluorophenyl pyrimidine ring, with the largest change of
0.09 ppm being observed for protons 2 and 4. New proton reso-
nances, labeled 17′, 17″, and 18, were observed between 3 and
5 ppm. These protons exhibit COSY correlations (Figure 3B)
Free Energy Calculations. Density functional theory calculations
were carried out to estimate the reaction energies between the reactants
(thiol and the nitrile moieties) and the conjugates. GSH was represented
by a cysteine residue to simplify the calculations. Cysteine was selected
as an alternative to ethanethiol because the marginal computational cost
2055
dx.doi.org/10.1021/tx500256j | Chem. Res. Toxicol. 2014, 27, 2052−2061

Chemical Research in Toxicology
Article
In Vitro Incubation with Stable-Labeled GSH (Glyci-
ne-13C2,15N). The involvement of endogenous cysteinylglycine
in the formation of 6A was assessed by incubating 6 with stable-
labeled GSH in human liver microsomes. The incubations
resulted in the formation of [13C2,15N]-6A, a metabolite whose
mass-to-charge ratio was greater, by m/z 3, than that of the
conjugate generated in the incubations with unlabeled GSH in
human liver microsomes (Figure 3A). Since all three labels were
present on the glycine moiety of GSH, the measured mass of 6B
was observed to be the same after each incubation.
Incubations in Heat-Inactivated Human Liver Micro-
somes. In order to investigate the contribution of microsomal
enzymes in the formation of thiazoline derivatives, 6 was incu-
bated with heat-inactivated human liver microsomes supple-
mented with GSH (Table 3). Both 6A and 6B were formed in
this incubation. However, the amounts that were observed were
more than 80% lower than those observed in active human liver
microsomes.
Incubations with Inhibitors and Human Liver Micro-
somes. Compound 6 was incubated with acivicin (5 mM), a
selective inhibitor of γ-glutamyl transpeptidase,24−26 together
with GSH in human liver microsomes in order to investigate the
role of γ-glutamyl transpeptidase in the formation of 6A and 6B
(Table 4). The amount of 6A that was formed was more than
50% lower in the presence of acivicin. However, no effect of
acivicin was noted in the formation of 6B. Similarly 6 was incu-
bated with a mixture of aprotinine (5 mM) and PMSF (5 mM),
nonspecific peptidase inhibitors,27 to assess the role of peptidases
in the formation of 6A and 6B (Table 4). It was found that
formation of both cyclized adducts, 6A and 6B, were almost
completely inhibited in the presence of these inhibitors.
Kinetics Studies. Compounds 2, 6, and 8 were selected as
representatives of the pyridazine, pyrimidine, and pyridine series,
respectively, for use in a kinetics assessment. The half-lives of
these compounds were measured in the presence of aqueous
cysteine (pH 7.4). In this experiment, the half-life of 2 was 11
0.09 min, the half-life of 6 was 8 0.38 min, and the half-life of
8 was >1 h.
Modeling Studies. Nucleophilic attack by GSH on the
carbon of the nitrile group is a key step in the formation of the
thiazoline derivatives shown in Scheme 3. Hence, predicting
the energy of formation of the nitrile with GSH was the focus of
the effort. Since quantum chemical calculations are computa-
tionally intensive, cysteine was used as a surrogate for GSH in
order to reduce the number of atoms in these calculations. The
order of adduct formation was predicted according to the free
energy of the reaction: the lower the free energy, the more
favored the adduct formation. Figure 4 shows that the predicted
free energies of adduct formation of 1−8 correlate well with
the experimental measurements with an R2 of 0.79. Gas-phase
energies were calculated in each case and, since the effort was
aimed at rank-ordering compounds, it was assumed that the
hydration effects would not alter the correlation dramatically.
In order to estimate the electrophilicity of the cyano carbon
atoms of the different compounds, the Fukui functions were
calculated after the quantum chemical optimization. Fukui func-
tions, f+ and f−, for each atom provide useful reactivity indices
for nucleophilic and electrophilic attack, respectively.28 Using
this approach, the nitrile groups of pyrimidine compounds 5 and
6 were found to be the most electrophilic followed by those of the
pyridines (7 and 8) and pyridazines (1−4).
Table 1. Positive-Ion MS and MS/MS Data and Relative UV
Abundance of Metabolites of 1−8 Formed Following
Reaction with GSH and NAC in Human Liver Microsomes
observed precursor
theoretical precursor
MS/MS
fragments (m/z)
metabolite
ion (m/z)
ion (m/z)
1A
343.0852
343.0859
325, 297, 251,
240, 182
1B
2A
286.0636
361.0757
286.0645
361.0765
242, 196, 182
343, 315, 269,
258, 200
2B
3A
304.0545
373.0963
304.0551
373.0965
260, 214, 200
355, 327, 281,
276, 212
3B
4A
316.0745
357.1001
316.0750
357.1010
272, 226, 212
339, 311, 265,
254, 196
4B
5A
300.0789
357.1006
300.0801
357.1010
256, 210, 196
339, 311, 265,
254, 196
5B
6A
300.0800
361.0760
300.0801
361.0765
254, 196
343, 315, 269,
258, 200
6B
7A
304.0547
356.1055
304.0551
356.1063
258, 200
338, 310, 264,
253, 195
7B
8A
299.0840
360.0804
299.0849
360.0813
255, 195
342, 314, 268,
257, 199
8B
303.0590
303.0598
259, 199
relative abundance in the
presence of GSH (%)
relative abundance in the
presence of NAC (%)
metabolite
1A
1B
2A
2B
3A
3B
4A
4B
5A
5B
6A
6B
7A
7B
8A
8B
31
not observed by UV
36
65
45
27
69
72
3
47
not observed by UV
19
5
18
5
67
11
78
10
3
not observed by UV
9
not observed by UV
6
among one another, consistent with the presence of the
thiazoline carboxylic acid moiety. Additional key data supporting
the structure of 6B includes changes in the chemical shifts
associated with the pyrimidine moiety (Table 2 and Figure 3B).
N1 and N5 of the conjugate exhibited an upfield chemical shift of
5.5 ppm compared to the parent. Similarly, C6 and C7 of the con-
jugate exhibited downfield chemical shifts of 15.2 and 48.3 ppm
compared to parent, respectively. Additionally, new peaks were
1
observed in the H−13C DEPT-HSQC and HMBC spectra of
6B, consistent with the presence of the thiazoline carboxylic
acid moiety. The NMR data confirmed that the structure of 6B
was 2-(5-(4-fluorophenyl)pyrimidin-2-yl)-4,5-dihydrothiazole-4-
carboxylic acid. Additionally, this data also suggested that the
structure of 6A was 2-(5-(4-fluorophenyl)pyrimidin-2-yl)-4,5-
dihydrothiazole-4-carboxamido acetic acid (Figure 3A,B), as the
mass difference between 6A and 6B was equal to the mass of
glycine.
2056
dx.doi.org/10.1021/tx500256j | Chem. Res. Toxicol. 2014, 27, 2052−2061

Chemical Research in Toxicology
Article
Figure 2. LC-UV Analysis (284 nm) of supernatants that were collected following incubations of 6 in human liver microsomes supplemented with either
GSH (A) or N-acetylcysteine (B), showing the formation of conjugates (A) 6A and 6B and (B) 6B.
Figure 3. (A) MS/MS product-ion spectra of 6A and 6B and structural interpretation of fragment ions. The measured mass-to-charge ratios of all ions
are within 5 ppm of the corresponding theoretical values. (B) COSY spectrum of 6B showing the presence of new resonances in the 3−6 ppm region.
Correlations are observed between protons 17′, 17″, and 18, consistent with the presence of the thiazoline carboxylic acid moiety.
functional groups that are prone to bioactivation are usually
avoided in new chemical entities for safety reasons.
DISCUSSION
■
Much of what is known about the biological effects of electro-
philes comes from the bioactivation literature. The formation of
reactive intermediates during the metabolism of xenobiotics is
usually considered to be an undesirable event in drug discovery,
since such intermediates can form covalent adducts with DNA
or proteins. Alternatively, electrophiles can undergo covalent
adduction with GSH, leading to depletion of GSH levels and
leaving the liver vulnerable to hepatotoxicity.29−34 Therefore,
However, some functional groups can participate in reactions
with GSH without bioactivation by cytochromes P450.35 The
extent of such reactions may depend on the electronic architec-
ture of the substrate. Such a trend was observed in the present
study, where the formation of GSH- and NAC-derived conju-
gates was demonstrated and correlated with the structures of
multiple N-heterocycles in a risk-assessment exercise. To this
end, HRMS and NMR data was used to elucidate the structures
2057
dx.doi.org/10.1021/tx500256j | Chem. Res. Toxicol. 2014, 27, 2052−2061

Chemical Research in Toxicology
Article
suggested that this metabolite formed via an oxime and/or a
nitrile oxide intermediate.36 Formation of this intermediate was
proposed to require P450, although it should be noted that addi-
tional steps were necessary in this sequence, since the start-
ing material was a primary amine instead of a nitrile. In the
present study, formation of thiazoline metabolites was NADP-
(H)-independent, suggesting that no such intermediate was
involved. The mechanism of formation of the thiazoline meta-
bolites was postulated to entail intramolecular cyclization of
the S-linked GSH conjugate and concomitant elimination of
pyroglutamic acid and ammonia, resulting in a dihydrothiazole
carboxamido intermediate that subsequently underwent loss of
glycine to yield a thiazoline carboxyl derivative.37 However, the
role of enzymes in the formation of the thiazoline metabolites
was not demonstrated. In the present study, the mechanism of
formation of these thiazoline metabolites was investigated using
6 as a model substrate. The metabolism study of 6 using labeled
GSH, which demonstrated the formation of a product that was
greater in mass by 3 u than the product of a study using unlabeled
GSH, established that the formation of 6A in human liver
microsomes was GSH-mediated and was not dependent on some
other endogenous cysteine-related substrate such as cyste-
inylglycine. Moreover, the thiazoline metabolites were also de-
tected in lower quantities following reactions in phosphate
buffer; however, their abundance was increased by the addition of
human liver microsomes and decreased by the addition of heat-
inactivated human liver microsomes. Therefore, these results de-
monstrated the role of microsomal enzymes like γ-glutamyl-
transpeptidase in catalyzing the formation of these thiazoline
metabolites. Indeed, the formation of 6A was inhibited in the
presence of acivicin, suggesting a role for γ-glutamyltranspepti-
dase in catalyzing the removal of the γ-glutamyl moiety from the
GSH conjugate. It is unlikely that this enzyme was further
involved in the cyclization of the final product, since the amine
intermediate would have intrinsically improved nucleophilicity vs
the amide, and intramolecular reaction to form 6A would be
favored entropically. Conversely, the formation of 6B in human
liver microsomes was not inhibited by acivicin and therefore
occurred independently of γ-glutamyltranspeptidase. Other
peptidases also appeared to assist in the formation of both 6A
and 6B, since a cocktail of PMSF and aprotinin inhibited these
reactions. Therefore, the formation of the thiazoline metabolites
was dependent on both chemical and enzymatic factors as de-
picted in Scheme 3. In this interpretation, the reaction was
initiated by nucleophilic attack of GSH on the nitrile carbon to
produce an intact GSH conjugate as an intermediate step, likely
in reversible fashion. Subsequently, γ-glutamyltranspeptidase
and another unspecified peptidase catalyzed the removal of the
glutamic acid moiety from the GSH conjugate,36,37,39 resulting in
the formation of a stable product. Again, the formation of the
thiazoline ring is proposed to occur spontaneously for the
reasons cited above. In the final step, 6B was produced from 6A
by loss of glycine, a reaction that was subject to inhibition by
PMSF/aprotenine and therefore was likely peptidase-dependent.
A similar mechanism could be drawn to explain the formation of
this metabolite from NAC, where removal of the N-acetyl moiety
could be catalyzed by an amidase, esterase, or N-deacetylase.
The formation of these thiazoline metabolites differed across
the series, decreasing in the order pyrimidine > pyridazine ≫
pyridine (Table 1), consistent with decreasing electrophilicity.
The general trend is consistent with that reported by Oballa
et al.11 It is also consistent with our own data on the kinetics of
disappearance of representative compounds in the presence of
Table 2. Proton, Carbon, and Nitrogen Chemical Shift
Assignments for 6B
metabolite
position
group
δ (1H)
9.26 (s)
δ (13C)
δ (15N)
1
N
292.7
2
CH
C
155.1
132.0
155.1
3
4
CH
N
9.26 (s)
5
292.7
6
C
157.6
164.6
7
C
8
N
9
C
129.6
129.0
116.0
162.8
116.0
129.0
10
11
12
13
14
15
16
17
CH
CH
C
7.96 (m)
7.41 (t, J = 8.6 Hz)
CH
CH
F
7.41 (t, J = 8.6 Hz)
7.96 (m)
S
CH2
3.48 (overlap),
35.3
3.64 (t, J = 8.2 Hz)
5.06 (t, J = 8.2 Hz)
18
19
CH
83.6
CO
172.0
Table 3. Comparison of Formation of 6A and 6B in Human
Liver Microsomes, Heated Human Liver Microsomes, and
Buffer
buffer denatured human liver microsomes human liver microsomes
6A
6B
28%
8%
11%
70%
12%
0.9%
Table 4. Formation of 6A and 6B in the Presence of Acivicin,
an Inhibitor of γ-Glutamyltranspepetidase, and a Mixture
Containing the Peptidase Inhibitors PMSF and Aprotinin
human liver
microsomes +
acivicin
human liver
microsomes only
human liver microsomes +
aprotinin + PMSF
6
28%
55%
8%
47%
23%
9%
98%
2%
6A
6B
0%
of thiazoline-containing metabolites that were formed by reac-
tions between these nitrile-substituted heterocycles and GSH
(Figure 3 and Scheme-3) in human liver microsomes. These
metabolites exhibited 2-(4,5-dihydrothiazole-4-carboxamido)-
acetic acid (A) and 4,5-dihydrothiazole-4-carboxylic acid (B)
moieties. Additional information was obtained by carrying out
incubations of these heterocycles with NAC, which provided a
similar but simplified route to the B group of metabolties that
represented the final products of these reactions. Similar
metabolites have been reported on multiple occasions. For
example, DPC423 (1-[3-(aminomethyl)phenyl]-N-[3-fluoro-2′-
(methylsulfonyl)[1,1′-biphenyl]-4-yl]-3-(trifluoromethyl)-1H-
pyrazole-5-carboxamide), a potent inhibitor of blood coagulation
factor Xa, produced a thiazoline metabolite in the rat, and it was
2058
dx.doi.org/10.1021/tx500256j | Chem. Res. Toxicol. 2014, 27, 2052−2061

Chemical Research in Toxicology
Article
Scheme 3. Proposed Mechanism of Formation of 6A and 6B on Reaction of 6 with GSH in Human Liver Microsomes
Figure 4. Predicted free energies of reactions between cysteine and compounds 1−8 and correlation with yield of metabolites A, as measured
experimentally.
2059
dx.doi.org/10.1021/tx500256j | Chem. Res. Toxicol. 2014, 27, 2052−2061

Chemical Research in Toxicology
Article
cysteine, where half-lives increased in the order 6 (pyrimidine) <
2 (pyridazine) < 8 (pyridine). On the basis of these results,
2-cyanopyrimidine was interpreted to be the most electrophilic
functionality due to the presence of its two electron-withdrawing
nitrogen atoms adjacent to the 2-carbon, suggesting that the
electrophilicity of the nitrile was sensitive to inductive effects. By
comparison, the pyridazine series was less electrophilic than that
of the pyrimidine series due to the altered position of the second
nitrogen atom distal from the 2-carbon, diminishing this effect.
By extension, the pyridine series was less succeptible to nucleo-
philic attack owing to fewer electron-withdrawing nitrogen
atoms as ring constituents compared to pyrimidine and an
accordingly diminished inductive effect on the nitrile group. The
functional groups at the para position of the phenyl ring also
appeared to regulate the electrophilicity of the nitrile and influ-
ence the rates of reaction with GSH. As a general trend, p-fluoro
substitution of the phenyl moiety increased electrophilicity of the
cyano group via an electron-withdrawing resonance effect,
resulting in higher amounts of adduct, whereas p-methyl and
p-methoxyl substitution lowered the electrophilicity of the cyano
group by hyperconjugation, resulting in lower amounts of
adduct. The calculated gas-phase reaction energies, which ac-
counted for both the makeup of the heterocycles and the pendant
substituents on the phenyl group, were consistent with the trend
in nitrile electrophilicity, as evidenced by relative extents of
reaction observed among the pyrimidine, pyridazine, and
pyridine compounds. Solvation effects were not accounted for
these calculations, since the aim was to rank-order compounds
(not necessarily to calculate energies with a high degree of
accuracy), and it was inferred that solvation effects would not
influence this outcome. In each series, electron-donating
substituents on the phenyl ring were found to reduce conjugate
formation, whereas electron-withdrawing substituents facilitated
it. The reaction energies predicted this trend with reasonable
accuracy. The computed energies are shown in Figure 4. As
mentioned in Oballa et al.,11 the reaction energies would not be
expected to correspond to empirical reaction rates in a
quantitative sense but may be used to rank-order compounds
within a series, an expectation that is consistent with the results of
this study. It is also noteworthy that the present computational
study aimed to predict the extent of product formation, which is
a thermodynamic quantity, as opposed estimating reactivity via
electrophilicity, a kinetic parameter. Time-dependent depletion
of compounds from the pyrimidine, pyridine, and pyridazine
series in buffer and human liver microsomes was performed.
Higher turnover of 6 in human liver microsomes indicated that
liver enzymes indeed catalyzed the formation of the thiazoline
metabolites. Compound 6 exhibited the highest turnover
and formed the highest amounts of thiazoline metabolites in
microsomes, demonstrating that the nitrile on the pyrimidine
core was the most electrophilic and that reaction with GSH was
catalyzed by microsomal enzymes. The Fukui indices for these
three compounds agreed with the observed high turnover of the
cyano pyrimidine. The cyano carbon that forms the adduct had
a value of 0.000435 in the context of the pyrimidine ring and
scored lower as a substituent of the pyridine (0.000387) and
pyridazine (0.000355) rings.
would correlate better with free energy of formation.38 This is
in keeping with the approach of Oballa et al. that recognized the
role of thermodynamic factors in governing the equilibrium
of these reversible reactions. However, the results of the pre-
sent study are also consistent with those of Berteotti et al., which
showed that the activation energies en route to thio adducts were
lower for cyano pyrimidines than for cyano pyridines.
In conclusion, the mechanism of formation of thiazoline meta-
bolites from nitrile-substituted N-heterocycles and GSH was
proposed following computational studies and experiments using
human liver microsomes. Specifically, the structure−property
relationships of these compounds were investigated by modeling
the electrophilicity of the nitrile groups using a computational
approach, and then output from that model was correlated
with metabolite formation, as estimated experimentally using
LC-MS. This approach demonstrated that electrophilicity of the
N-heterocycles played a key role in determining the extent of
formation of the thiazoline metabolites. The lowering of elec-
trophilicity of such N-heterocycles, either by replacement of the
pyrimidine ring with pyridine or by para-phenyl substitution with
electron-donating groups, could be used as a strategy to minimize
the conversion of these compounds to thiazoline metabolites.
Hence, the mechanistic insights gained from this study could be
used to assist a medicinal chemistry effort to design a refined
chemotype with improved stability toward glutathione and other
biological thiols.
ASSOCIATED CONTENT
* Supporting Information
■
S
LC-MS chromatograms and MS/MS product-ion spectra of all
metabolites as well as NMR spectra of 6 and 6B. This material is
available free of charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
*E-mail: benjamin.m.johnson@bms.com.
■
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
The authors thank Ramaswamy A. Iyer, Yue-Zhong Shu,
Jonathan L. Josephs, and William Griffith Humphreys for helpful
suggestions and discussion. The authors also acknowledge Silvi
Ann Chacko and Senthil Kumar Murugesan for helping to
conduct experiments and LC-MS method development.
ABBREVIATION
■
NAC, N-acetylcysteine; PMSF, phenylmethanesulfonyl fluoride;
DMF, dimethylformamide; Cs2CO3, cesium carbonate; HSQC,
heteronuclear single quantum coherence; HMBC, heteronuclear
multiple bond corelation spectroscopy; COSY, correlated
spectroscopy; DEPT, distortionless enhancement by polar-
ization transfer; HRMS, high-resolution mass spectroscopy
REFERENCES
■
(1) Guengerich, F. P. (2005) Principles of covalent binding of reactive
metabolites and examples of activation of bis-electrophiles by
conjugation. Arch. Biochem. Biophys. 433, 369−378.
Recently, Berteotti et al.28 reported a computational study on
the reactivity of nitriles. In this study, activation energies were
estimated via modeling of the transition state of the nitrile−
cysteine reaction and correlated with a kinetic measure of nitrile
reactivity. In contrast, the adduct yields that were measured in the
present study were interpreted as thermodynamic quantities that
(2) Liebler, D. C. (2007) Protein damage by reactive electrophiles:
targets and consequences. Chem. Res. Toxicol. 21, 117−128.
(3) Anders, M. (2004) Glutathione-dependent bioactivation of
haloalkanes and haloalkenes. Drug Metab. Rev. 36, 583−594.
2060
dx.doi.org/10.1021/tx500256j | Chem. Res. Toxicol. 2014, 27, 2052−2061

Chemical Research in Toxicology
Article
(4) Anders, M. (2007) Chemical toxicology of reactive intermediates
formed by the glutathione-dependent bioactivation of halogen-
containing compounds. Chem. Res. Toxicol. 21, 145−159.
(5) Marnett, L. (1998) Chemistry and biology of DNA damage by
malondialdehyde. IARC Sci. Publ., 17−27.
conjugates of 1-[3-(aminomethyl) phenyl]-N-[3-fluoro-2′-(methylsul-
fonyl)-[1,1′-biphenyl]-4-yl]-3-(trifluoromethyl)-1H-pyrazole-5-carbox-
amide (DPC 423) and its analogs: the role of γ-glutamyltranspeptidase
in the biotransformation of benzylamines. Drug Metab. Dispos. 29,
1296−1306.
(25) Antoine, B., Rahimi-Pour, A., Siest, G., Magdalou, J., and Galteau,
M. M. (1987) Differential time-course of induction of rat liver gamma-
glutamyltransferase and drug-metabolizing enzymes in the endoplasmic
reticulum, golgi and plasma membranes after a single phenobarbital
injection. Evaluation of protein variations by two-dimensional electro-
phoresis. Cell Biochem. Funct. 5, 217−231.
(26) Raulf, M., Stuning, M., and Konig, W. (1985) Metabolism of
leukotrienes by L-gamma-glutamyl-transpeptidase and dipeptidase from
human polymorphonuclear granulocytes. Immunology 55, 135−147.
(27) Miles, K. K., Kessler, F. K., Smith, P. C., and Ritter, J. K. (2006)
Characterization of rat intestinal microsomal UDP-glucuronosyltrans-
ferase activity toward mycophenolic acid. Drug Metab. Dispos. 34, 1632−
1639.
(28) Berteotti, A., Vacondio, F., Lodola, A., Bassi, M., Silva, C., Mor,
M., and Cavalli, A. (2014) Predicting the reactivity of nitrile-carrying
compounds with cysteine: a combined computational and experimental
study. ACS Med. Chem. Lett. 5, 501−505.
(29) Ju, C., and Uetrecht, J. (2002) Mechanism of idiosyncratic drug
reactions: reactive metabolite formation, protein binding and the
regulation of the immune system. Curr. Drug Metab. 3, 367−377.
(30) Liebler, D. C., and Guengerich, F. P. (2005) Elucidating
mechanisms of drug-induced toxicity. Nat. Rev. Drug Discovery 4, 410−
420.
(31) Uetrecht, J. P. (1999) New concepts in immunology relevant to
idiosyncratic drug reactions: the “danger hypothesis” and innate
immune system. Chem. Res. Toxicol. 12, 387−395.
(32) Cohen, S. D., Pumford, N. R., Khairallah, E. A., Boekelheide, K.,
Pohl, L. R., Amouzadeh, H., and Hinson, J. A. (1997) Selective protein
covalent binding and target organ toxicity. Toxicol. Appl. Pharmacol. 143,
1−12.
(33) Mitchell, J., Jollow, D., Potter, W., Davis, D., Gillette, J., and
Brodie, B. (1973) Acetaminophen-induced hepatic necrosis. I. Role of
drug metabolism. J. Pharmacol. Exp. Ther. 187, 185−194.
(34) Mitchell, J., Jollow, D., Potter, W., Gillette, J., and Brodie, B.
(1973) Acetaminophen-induced hepatic necrosis. IV. Protective role of
glutathione. J. Pharmacol. Exp. Ther. 187, 211−217.
(6) Guengerich, F. P. (2005) Activation of alkyl halides by glutathione
transferases. Methods Enzymol. 401, 342−353.
(7) Kalgutkar, A. S., Gardner, I., Obach, R. S., Shaffer, C. L., Callegari,
E., Henne, K. R., Mutlib, A. E., Dalvie, D. K., Lee, J. S., Nakai, Y.,
O’Donnell, J. P., Boer, J., and Harriman, S. P. (2005) A comprehensive
listing of bioactivation pathways of organic functional groups. Curr. Drug
Metabol. 6, 161−225.
(8) Singh, J., Petter, R. C., Baillie, T. A., and Whitty, A. (2011) The
resurgence of covalent drugs. Nat. Rev. Drug Discovery 10, 307−317.
(9) Park, B. K., Boobis, A., Clarke, S., Goldring, C. E., 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., and Baillie, T. A. (2011) Managing the
challenge of chemically reactive metabolites in drug development. Nat.
Rev. Drug Discovery 10, 292−306.
(10) Morley, A. D., Kenny, P. W., Burton, B., Heald, R. A., Macfaul, P.
A., Mullett, J., Page, K., Porres, S. S., Ribeiro, L. R., and Smith, P. (2009)
5-Aminopyrimidin-2-ylnitriles as cathepsin K inhibitors. Bioorg. Med.
Chem. Lett. 19, 1658−1661.
(11) Oballa, R. M., Truchon, J.-F., Bayly, C. I., Chauret, N., Day, S.,
Crane, S., and Berthelette, C. (2007) A generally applicable method for
assessing the electrophilicity and reactivity of diverse nitrile-containing
compounds. Bioorg. Med. Chem. Lett. 17, 998−1002.
(12) Macfaul, P. A., Morley, A. D., and Crawford, J. J. (2009) A simple
in vitro assay for assessing the reactivity of nitrile containing compounds.
Bioorg. Med. Chem. Lett. 19, 1136−1138.
(13) Deaton, D. N., and Tavares, F. X. (2005) Design of cathepsin K
inhibitors for osteoporosis. Curr. Top. Med. Chem. 5, 1639−1675.
(14) Grabowskal, U., Chambers, T. J., and Shiroo, M. (2005) Recent
developments in cathepsin K inhibitor design. Curr. Opin. Drug
Discovery Dev. 8, 619−630.
(15) Altmann, E., Cowan-Jacob, S. W., and Missbach, M. (2004) Novel
purine nitrile derived inhibitors of the cysteine protease cathepsin K. J.
Med. Chem. 47, 5833−5836.
(16) Rankovic, Z., Cai, J., Kerr, J., Fradera, X., Robinson, J., Mistry, A.,
Hamilton, E., Mcgarry, G., Andrews, F., and Caulfield, W. (2010) Design
and optimization of a series of novel 2-cyano-pyrimidines as cathepsin K
inhibitors. Bioorg. Med. Chem. Lett. 20, 1524−1527.
(17) Ehmke, V., Quinsaat, J. E., Rivera-Fuentes, P., Heindl, C.,
Freymond, C., Rottmann, M., Brun, R., Schirmeister, T., and Diederich,
F. (2012) Tuning and predicting biological affinity: aryl nitriles as
cysteine protease inhibitors. Org. Biomol. Chem. 10, 5764−5768.
(18) 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.
(19) Shin, N. Y., Liu, Q., Stamer, S. L., and Liebler, D. C. (2007)
Protein targets of reactive electrophiles in human liver microsomes.
Chem. Res. Toxicol. 20, 859−867.
(35) Kalgutkar, A. S., Mascitti, V., Sharma, R., Walker, G. W., Ryder, T.,
McDonald, T. S., Chen, Y., Preville, C., Basak, A., McClure, K. F., Kohrt,
J. T., Robinson, R. P., Munchhof, M. J., and Cornelius, P. (2011)
Intrinsic electrophilicity of a 4-substituted-5-cyano-6-(2-methylpyridin-
3-yloxy)pyrimidine derivative: structural characterization of glutathione
conjugates in vitro. Chem. Res. Toxicol. 24, 269−278.
(36) Mutlib, A., Chen, S., Espina, R., Shockcor, J., Prakash, S., and Gan,
L. (2002) P450-mediated metabolism of 1-[3-(aminomethyl)phenyl]-
N-[3-fluoro-2′-(methylsulfonyl)-[1,1′-biphenyl]-4-yl]-3 (trifluoro-
methyl)-1H-pyrazole-5-carboxamide (DPC 423) and its analogues to
aldoximes. Characterization of glutathione conjugates of postulated
intermediates derived from aldoximes. Chem. Res. Toxicol. 15, 63−75.
(37) Kera, Y., Kiriyama, T., and Komura, S. (1985) Conjugation of
acetaldehyde with cysteinylglycine, the first metabolite in glutathione
breakdown by gamma-glutamyltranspeptidase. Agents Actions 17, 48−
52.
(20) Kassahun, K., Black, W. C., Nicoll-Griffith, D., Mcintosh, I.,
Chauret, N., Day, S., Rosenberg, E., and Koeplinger, K. (2011)
Pharmacokinetics and metabolism in rats, dogs, and monkeys of the
cathepsin K inhibitor odanacatib: demethylation of a methylsulfonyl
moiety as a major metabolic pathway. Drug Metab. Dispos. 39, 1079−
1087.
(38) Baidya, M., and Mayr, H. (2008) Nucleophilicities and carbon
basicities of DBU and DBN. Chem. Commun. 15, 1792−1794.
(39) Chandra, A. K., and Nguyen, M. T. (2002) Use of local softness
for the interpretation of reaction mechanisms. Int. J. Mol. Sci. 3, 310−
323.
(21) Leclerc, R., and Uguen, D. (1994) A convenient preparation of
optically pure 3-hydroxyglutaric acid derivatives. Tetrahedron Lett. 35,
1999−2002.
(22) White, E. H., and Worther, H. (1966) Analogs of firefly luciferin.
III. J. Org. Chem. 31, 1484−1488.
(23) Goodman, A. J., Stanforth, S. P., and Tarbit, B. (1999)
Desymmetrization of dichloroazaheterocycles. Tetrahedron 55,
15067−15070.
(24) Mutlib, A., Shockcor, J., Chen, S.-Y., Espina, R., Lin, J., Graciani,
N., Prakash, S., and Gan, L.-S. (2001) Formation of unusual glutamate
2061
dx.doi.org/10.1021/tx500256j | Chem. Res. Toxicol. 2014, 27, 2052−2061