Irreversible binding of an anticancer compound (BI-94) to plasma proteins

Article abstract of DOI:10.3109/00498254.2015.1025250

1. We investigated the mechanisms responsible for the in vivo instability of a benzofurazan compound BI-94 (NSC228148) with potent anti-cancer activity.2. BI-94 was stable in MeOH, water, and in various buffers at pHs 2.5-5, regardless of the buffer compo

Full text of DOI:10.3109/00498254.2015.1025250

http://informahealthcare.com/xen
ISSN: 0049-8254 (print), 1366-5928 (electronic)
Xenobiotica, Early Online: 1–16
2015 Informa UK Ltd. DOI: 10.3109/00498254.2015.1025250
!
RESEARCH ARTICLE
Irreversible binding of an anticancer compound (BI-94) to plasma
proteins
Nagsen Gautam1, Rhishikesh Thakare1, Sandeep Rana2, Amarnath Natarajan2, and Yazen Alnouti1
2
1Department of Pharmaceutical Sciences, College of Pharmacy, University of Nebraska Medical Center, Omaha, NE, USA and Eppley Institute for
Cancer Research, University of Nebraska Medical Center, Omaha, NE, USA
Abstract
Keywords
1. We investigated the mechanisms responsible for the in vivo instability of a benzofurazan
compound BI-94 (NSC228148) with potent anti-cancer activity.
Glutathione, irreversible protein binding,
isotopic filtering, LC-MS/MS,
N-acetylcysteine
2. BI-94 was stable in MeOH, water, and in various buffers at pHs 2.5–5, regardless of the buffer
composition. In contrast, BI-94 was unstable in NaOH and at pHs 7–9, regardless of the buffer
composition. BI-94 disappeared immediately after spiking into mice, rat, monkey, and
human plasma. BI-94 stability in plasma can be only partially restored by acidifying it, which
indicated other mechanisms in addition to pH for BI-94 instability in plasma.
3. BI-94 formed adducts with the trapping agents, glutathione (GSH) and N-acetylcysteine
(NAC), in vivo and in vitro via nucleophilic aromatic substitution reaction. The kinetics of
adduct formation showed that neutral or physiological pHs enhanced and accelerated GSH
and NAC adduct formation with BI-94, whereas acidic pHs prevented it. Therefore,
physiological pHs not only altered BI-94 chemical stability but also enhanced adduct
formation with endogenous nucleophiles. In addition, adduct formation with human serum
albumin-peptide 3 (HSA-T3) at the Cys34 position was demonstrated.
History
Received 13 January 2015
Revised 26 February 2015
Accepted 28 February 2015
Published online 14 April 2015
4. In conclusion, BI-94 was unstable at physiological conditions due to chemical instability and
irreversible binding to plasma proteins.
Introduction
Drugs may bind to various macromolecular components in
plasma, including albumin, a1-acid glycoproteins, lipopro-
The effects of binding to plasma proteins on the pharmaco-
kinetic, pharmacological, and toxicological profiles of thera-
peutic molecules are well established. Only the free
(unbound) fractions of drugs are available for distribution,
metabolism, excretion, and to exert pharmacological and
toxicological effects (Schmidt et al., 2010). In addition,
plasma protein binding plays a role in drug–drug interactions,
non-linear and stereoselective pharmacokinetics, and inter-
and intra-individual variabilities in pharmacological effects
(Ito et al., 1998; Paliwal et al., 1993). Therefore, the kinetics
of plasma protein binding is determined at early stages in drug
discovery and an increasing number of drugs are monitored
by measuring the unbound, rather than the total, plasma
concentrations as a part of their therapeutic drug monitoring
(Dasgupta, 2007; Schmidt et al., 2010).
teins, and immunoglobulins (Chan
& Gerson, 1987).
In addition, significant portions of drugs may bind to
erythrocytes. Typically, drug binding to plasma proteins
takes place via the formation of reversible bonds (hydrogen
bonds or van der Waals forces). However, reactive species
resulting from metabolic activation (Jenkins et al., 2009;
Jinno et al., 2011; Kalgutkar & Didiuk, 2009; Meng et al.,
2011) or spontaneous chemical activation (Meng et al.,
2015) of drugs may bind strongly to proteins or other
macromolecules via covalent/irreversible chemical bonds
(Dubois et al., 1993; Kalgutkar & Didiuk, 2009). The
modification of endogenous macromolecules by reactive
metabolites can generate ‘‘foreign’’ macromolecules (hap-
tens) that can lead to immune-mediated idiosyncratic adverse
reactions (Nassar & Lopez-Anaya, 2004). Therefore, the
formation of irreversible complexes between reactive metab-
olites and macromolecules is often associated with drug
toxicity (Kalgutkar & Didiuk, 2009) including skin sensitiza-
tion, respiratory sensitization, liver toxicity, chromosomal
aberration, and a wide range of idiosyncratic toxicities (Enoch
et al., 2011; Kalgutkar & Didiuk, 2009). Examples of
compounds that demonstrate irreversible protein binding
and toxicity due to metabolic activation include tolmetin,
Address for correspondence: Yazen Alnouti, Department of
Pharmaceutical Sciences, College of Pharmacy, University of Nebraska
Medical Center, 986025 Nebraska Medical Center, Omaha, NE 68198-
6025, USA. Tel: +1 402 559 4631. Fax: +1 402 559 9543. E-mail:
yalnouti@unmc.edu

2
N. Gautam et al.
Xenobiotica, Early Online: 1–16
zomepirac, troglitazone, practolol, benoxaprofen, ticrynafen,
nomifensine, and p-phenylenediamine (PPD) (Hyneck et al.,
1988; Jenkinson et al., 2009; Smith et al., 1986). However,
not all drugs metabolized into reactive metabolites that bind
to macromolecules are associated with any toxicity
(Nakayama et al., 2009). In general, reactive metabolites
are either electrophiles (electron deficient) or free radicals
(Freed et al., 2008; Uetrecht, 2003). Electrophiles can also be
classified into hard (localized charge) or soft electrophiles
(delocalized charge). Soft electrophiles include epoxides,
a,b-unsaturated carbonyl compounds, quinone, quinone
imines, quinone methides, imine methide, isocynate, iso-
thiocynates, aziridinium, and episulphonium, whereas alde-
hydes and iminium ions belong to the group of hard
electrophiles (Prakash et al., 2008; Tang & Lu, 2010). Soft
electrophiles react with cysteine residues in proteins and with
thiol groups of glutathione or N-acetylcysteine (soft nucleo-
philes), while hard electrophiles react with pyrimidines and
purines in DNA and with lysine residues in proteins (hard
nucleophiles) (Li et al., 2011a; Prakash et al., 2008).
analogues such as NAC are soft trapping agents used to trap
soft electrophiles; whereas hard nucleophiles such as
semicarbazide, methoxylamine, N-acetyl-lysine, and cyanide
ions are used to trap hard electrophiles (Kalgutkar & Didiuk,
2009; Kerksick & Willoughby, 2005; Li et al., 2011a). The
downsides of such approach include the lack of single
trapping agent that can serve as a universal surrogate for
protein adducts and the failure to trap very reactive metab-
olites that are unable to escape from their site of formation
because they react rapidly with the surrounding environment.
In addition to the experimental determination, in silico
models are available to screen compounds early in drug
discovery to predict whether they can undergo metabolic
activation to generate reactive metabolites (Liu et al., 2013;
Ma & Subramanian, 2006; Uetrecht, 2003).
BI-94 is a synthetic sulfonyl group-containing compound
(4-(benzylsulfonyl)-7-(hydroxy(oxido)amino)-2,1,3-benzoxa-
diazole) with an average GI50 value of 2.29 ꢀ 10ꢁ5 M in the
NCI 60 cell line panel. The carboxy-terminus domains of the
early onset of breast cancer gene 1 (BRCT-BRCA1) recognize
and bind phosphorylated proteins (Manke et al., 2003;
Yu et al., 2003). These protein–protein interactions play a
critical role in the DNA damage response pathway (Cantor
et al., 2001; Kim et al., 2007; Liu et al., 2007; Sobhian
et al., 2007; Wang et al., 2007). It was shown that
tetrapeptides bind BRCT-BRCA1 with nanomolar affinities
(Joseph et al., 2010; Yuan et al., 2011a,b). A poly-arginine-
tagged tetrapeptide inhibitor blocked the BRCA1–phospho-
protein interaction and sensitized cells to PARP inhibition
(Pessetto et al., 2012). The relatively short half-life of the
peptide inhibitor led us to screen the NCI diversity set for
small molecule inhibitors using a BRCT–BRCA1 fluores-
cence polarization assay (Lokesh et al., 2006; Simeonov
et al., 2008). This led to the identification of the benzofur-
azan compound NSC228148 (BI-94), developed originally by
the Sanford-Burnham Medical Research Institute as a com-
pound-regulating apoptosis, as a hit compound. Other high-
throughput screening campaigns have identified substituted
benzofurazans as inhibitors of cyclic adenosine monopho-
sphate (cAMP) response-element (CRE) binding protein
(CREB)-mediated gene transcription (Xie et al., 2013),
influenza A virus (Kessler et al., 2013), and as antifungal
agents (Wang et al., 2014).
Several techniques are used for the determination of
molecules binding to plasma proteins. These techniques can
be divided into methods that determine unbound concentra-
tion of drugs from which the bound concentration is indirectly
estimated, and methods that directly determine bound
concentrations of drugs (Yan
&
Caldwell, 2004).
Commonly used methods in the first category include
ultrafiltration, blood partitioning, equilibrium dialysis, chro-
matography using protein immobilized columns, ultracentri-
fugation, charcoal adsorption, high performance frontal
analysis, solid phase microextraction, fluorescence, surface
plasmon resonance, and solid supported lipid membranes
(Ballard & Rowland, 2011; Banker & Clark, 2008; Barre
et al., 1985; Chuang et al., 2009; Gautam et al., 2013;
Howard et al., 2010; Schuhmacher et al., 2004; Shibukawa
et al., 1999; Yuan et al., 1995).
Direct quantification of bound concentration, which is
used to quantify irreversible binding to plasma proteins,
involves the application of radioactive compounds, isolation
of plasma proteins, washing non-covalently bound fractions
using strong-organic solvents, followed by quantification of
the remaining covalently bound radioactivity (Hyneck et al.,
1988; Kappus et al., 1973; Smith et al., 1986). In addition,
specific covalent protein adducts can be isolated and
quantified using immunoaffinity assays. Protein modifica-
tions due to adduct formation can also be detected using mass
spectrometry (MS) (Rappaport et al., 2012; Uetrecht, 2003).
These studies can be performed in various in vitro tissue
systems such as hepatic microsomes or hepatocytes or in vivo
(Wen & Fitch, 2009).
During the analytical method development for BI-94, as a
prelude to assess the in vivo efficacy of BI-94 in tumor animal
models, we were not able to recover BI-94 from mouse, rat,
monkey, or human plasma. Consequently, we investigated
several approaches to improve the extraction recovery of
BI-94 from mouse plasma and tissues, which led to the
identification of irreversible covalent binding between BI-94
and plasma proteins. This could be explained by the chemical
structure of BI-94, which is considered a benzofurazan (BFZ)
derivative. BFZ derivatives are highly reactive compounds
that can react with many functional groups including amines,
thiols, hydroxyls, carbonyls, and carboxyls, which lead to the
formation of adducts with macromolecules in vivo (Federici
et al., 2009; Ricci et al., 2005; Xie et al., 2013; Yang &
Guan, 2015). The presence of electron-withdrawing substitu-
ents, like the ones in BI-94, considerably increases BFZ
reactivity (Ghosh et al., 1972), which correlate with their
However, many drug-macromolecule adducts cannot be
directly detected due to uncertainty of the binding position,
complex structure, or further degradation after binding.
Therefore, indirect methods including trapping agents such
as glutathione (GSH) and N-acetylcysteine (NAC) are used to
capture reactive drug metabolites (Dubois et al., 1993; Jahn
et al., 2012a,b; Masubuchi et al., 2007; Nakayama et al.,
2011). The resulting more stable complexes of the reactive
metabolites with these trapping agents are then quantified as
surrogates for the overall protein adducts. GSH and its

DOI: 10.3109/00498254.2015.1025250
Irreversible binding of BI-94 to plasma proteins
3
toxicity (Patridge et al., 2012). In addition, BFZs promote
oxidative stress in vivo via the generation of reactive oxygen
species (Patridge et al., 2012). In this manuscript, we
characterized BI-94 stability and its irreversible binding to
mouse plasma proteins.
meta-chloroperoxybenzoic acid (2.54 g, 17.4 mmol) was
added and the reaction mixture was stirred at room tempera-
ture for 3 h. Reaction was monitored by thin layer chroma-
tography using 5:1 hexane:ethyl acetate eluent. The reaction
mixture was filtered and the supernatant was washed with
saturated sodium bicarbonate, dried over magnesium sulfate,
and column chromatographed to obtain the desired product
Materials and methods
1
(510 mg, 46%) as off-white solid. H NMR (CDCl3) ꢀ 8.35
Chemicals and reagents
(m, J ¼ 8.0 Hz, 1H), 7.46–7.32 (m, 5H), 7.18 (d, J ¼ 8.0 Hz,
1H), 4.52 (s, 2H); 13C NMR ꢀ 149.07, 142.37, 140.83, 133.61,
132.81, 130.59, 129.06, 128.70, 128.33, 121.24, 36.53.
BI-94 was synthesized as described below. 4-Chloro-7-
nitrobenzo[c][1,2,5]oxadiazole, glutathione, glutathione-
(glycine-13C2, 15N), and N-acetylcysteine were obtained
from Sigma-Aldrich (St. Louis, MO). Synthetic human
serum albumin (HSA)-T3 peptide (third largest peptide
forms after tryptic digestion) was purchased from Biomatik
(Wilmington, DE). HPLC-grade methanol, acetonitrile,
ammonium acetate, ammonium formate, ammonium hydrox-
ide, formic acid, and acetic acid were obtained from Fisher
Scientific (Fair Lawn, NJ). Mouse, rat, monkey, and human
plasma were purchased from Equitech-Bio (Kerrville, TX).
Direct MS/MS and LC-MS/MS analysis
A 4000 Q TRAPÕ quadrupole linear ion trap hybrid mass
spectrometer (Agilent Technologies, Palo Alto, CA) with
electrospray ionization (ESI) source was used throughout
(Applied Biosystems, MDS Sciex, Foster City, CA). The
qualitative detection of adduct formation of BI-94 was
performed via direct infusion into the MS using 55-2222
infusion syringe pump (Harvard Apparatus, Holliston, MA) at
a flow rate of 10 mL/min. Enhanced MS (EMS) scans were
obtained in both positive and negative ESI modes with the
following parameters: ion spray voltage, 4500 V; source
temperature, 500 ꢂC; curtain gas (nitrogen), 12 arbitrary
units; and collision gas (nitrogen), ‘‘High’’, range 100–700
(m/z) with 1000 Da/s scan rate. Adducts were further
characterized using product ion scans of precursor ions with
ramped collision energy (CE) over the range of ꢁ130 to ꢁ10
eV to identify their fragmentation patterns.
Synthesis of BI-94
BI-94 was synthesized in two steps from 4-chloro-7-
nitrobenzo[c][1,2,5]oxadiazole and purified (> 98%) using
flash chromatography Figure 1A). Flash chromatography was
carried out on silica gel (200–400 mesh). Thin layer chroma-
tography (TLC) was run on pre-coated EMD silica gel 60 F254
plates and observed under UV light at 254 nm and with basic
potassium permanganate dip. Column chromatography was
performed with silica gel (230–400 mesh, grade 60, Fisher
Scientific, Fair Lawn, NJ). 1H NMR (500 MHz) and 13C
NMR (125 MHz) spectra were recorded in chloroform-d on a
Varian-500 and Varian-600 spectrometer (Varian Medical
The LC-MS/MS system comprised a Waters ACQUITY
ultra-performance liquid chromatography (UPLC) system
(Waters, Milford, MA) coupled to the same MS system
described above. All chromatographic separations were
performed with an ACQUITY UPLCÕ BEH Phenyl
(2.1ꢀ150 mm, 1.7 mm; Waters, Milford, MA) analytical
column equipped with an ACQUITY UPLC C18 guard
column (Waters, Milford, MA). Mobile phase A consisted of
ammonium acetate (7.5 mM, pH 4.0) and mobile phase B
consisted of 50% acetonitrile (ACN) and 50% methanol
(MeOH). The initial mobile phase composition was 52.5% B
for the first 7 min, gradually decreased to 30% B in 0.25 min
and held constant for 2 min, then gradually increased to 90%
B in 0.25 min and held constant for 4 min. Mobile phase B
was then reset to 52.5% in 0.25 min and the column was
equilibrated for 3 min before the next injection. A flow rate of
0.25 mL/min was used and the injection volume of all samples
was 10 lL.
1
Systems, Palo Alto, CA) (CDCl3 was 7.27 ppm for H and
77.23 ppm for 13C)_. Proton and carbon chemical shifts were
reported in ppm relative to the signal from residual solvent
proton and carbon. LC-MS for the compounds was generated
on an Agilent 1200 series system (Agilent Technologies, Palo
Alto, CA) with UV detector (214 nm and 254 nm) and an
Agilent 6130 quadrupole mass detector (Agilent
Technologies, Palo Alto, CA).
4-(Benzylthio)-7-nitrobenzo[c][1,2,5]oxadiazole
To a solution of 4-chloro-7-nitrobenzo[c][1,2,5]oxadiazole
(3.0 g, 15.03 mmol) in ethanol (25 mL), dry pyrdine (1.5 mL,
15.03 mmol) was added and the reaction mixture was stirred
at room temperature for 10 min. Phenylmethanethiol
(3.55 mL, 30.06 mmol) was added dropwise at room tem-
perature (color change from yellow to brown) followed by
heating the reaction mixture at 80 ꢂC for 16 h. Reaction was
monitored by thin layer chromatography using 10:1 hex-
ane:ethyl acetate eluent. Cold hexane (30 mL) was added to
the reaction mixture and filtered to obtain solid as a pure
product (1.38 g, 32%), which was used as such for the next
step without further purification.
Screening for GSH and NAC adducts
Qualitative screening for adducts formation with NAC
(method i only) and GSH (methods i–vi) were performed
using various approaches:
(i) Enhanced
MS-information-dependent
acquisition-
enhanced product ion scan (EMS-IDA-EPI) in both
positive and negative ionization modes. The EMS survey
scan was conducted in the mass range from 160 to
750 Da, at a 1000 Da/s scan rate followed by an
enhanced resolution (ER) scan at 250 Da/s. The IDA
(information dependent acquisition) threshold was set at
4-(Benzylsulfonyl)-7-nitrobenzo[c][1,2,5]oxadiazole
To a solution of 4-(benzylthio)-7-nitrobenzo[c][1,2,5]oxadia-
zole (1.0 g, 3.4 mmol) in methylene chloride (30 mL),

4
N. Gautam et al.
Xenobiotica, Early Online: 1–16
Figure 1. (A) Chemical structure of BI-94, and plausible mechanisms for the instability of BI-94 under basic conditions. (B) Product ion spectra of BI-
94 degradation product, under basic condition, in the negative ESI mode (m/z ꢁ 180).
5000 counts per second (cps), above which enhanced
product ion (EPI) spectrum collection is triggered
from the precursor mass of that particular channel.
The EPI scan rate was 4000 Da/s and the scan range
was 70–750 Da.
Isotopic filtering
A mixture of 1:1 glutathione:stable isotopically labeled
glutathione (glycine-13C2, 15N) was used in the GSH screen-
ing incubations. The total ion chromatograms (TIC), obtained
from the screening scans mentioned above, were filtered for
the isotopic pattern of 1:1 abundance of ion pairs with a mass
difference of 3 Da, using the elemental targeting feature in
Analyst software (AB Sciex, Framingham, MA) (Liao et al.,
2012; Mutlib et al., 2005).
(ii) Neutral loss scan of m/z 129 in the positive mode, over a
range of 130–800 Da/1.5 s.
(iii) Neutral loss scan of m/z 307 in the positive ion mode,
over a range of 308–800 Da/1.5 s.
(iv) Precursor ion scan of m/z 130 in the positive mode, over
a range of 150–800 Da/1.5 s.
Quantification of BI-94, GSH, NAC, and their adducts
(v) Precursor ion scan of m/z 254 in the negative mode, over
a range of 250–800 Da/1.5 s.
LC-MS/MS quantification of BI-94 was performed by
monitoring the transition m/z 319.1!179.8. Quantitative
detections of adducts were performed by monitoring the
(vi) Precursor ion scan of m/z 272 in the negative mode, over
a range of 250–800 Da/1.5 s.

DOI: 10.3109/00498254.2015.1025250
Irreversible binding of BI-94 to plasma proteins
5
transitions of m/z 468.99!196 and m/z 578.0!305 for A1
and A2 adducts of GSH, respectively, in the negative mode.
The m/z 195.88!46, 304.91!150, and 454.81!196.1 were
used to quantify the A1, A2, and A3 adducts of NAC,
respectively, in the negative mode. GSH and NAC themselves
were monitored in the negative mode using m/z 306.2!128
and m/z 161.79!83.9, respectively. There was no need to use
absolute concentrations in these analyses; therefore, absolute
peak areas were used in a semi-quantitative manner.
analyzed immediately by LC-MS/MS. Separate incubations of
BI-94 (final concentration of 10 lg/mL) for 30 min in water
were utilized for the structural identification of NAC and
GSH adducts using direct LC-MS/MS and MS/MS analysis.
HSA-T3 peptide adduct formation in solution
About 10 mL of 1 mg/mL HSA-T3 peptide in 50% ACN and
10 mL of BI-94 in 70% MeOH in DMSO were added to
480 mL of H2O and were incubated for 30 min. Then, the
incubation mixture was mixed with 500 mL acetonitrile
(a final concentration of 10 mg/mL). Qualitative screening
for adducts formation of BI-94 with HSA-T3 was performed
using EMS scans in both positive and negative modes over a
range of 100–2500 Da. Adducts were further characterized
using enhanced product ion (EPI) scans of precursor ions to
identify their fragmentation patterns.
Stability of BI-94 in solution, plasma, serum, and
blood
The stability of BI-94 was tested in various buffers and at
different pH levels including pH 2.5 and pH 3 ammonium
formate (7.5 mM), pH 3 and 5 ammonium acetate (7.5 mM),
pH 7 ammonium bicarbonate (7.5 mM), pH 7 phosphate
buffer saline (PBS), pH 9 ammonium carbonate (7.5 mM),
NaOH (0.75 mM), and deionized water at room temperature.
About 10 mL of a 1 mg/mL working standard solution of
BI-94 in 70% MeOH in DMSO was added into 990 mL
volume of these buffers to make a final BI-94 concentration
of 10 mg/mL. About 10 mL aliquots were collected in 90 mL of
55.0% methanol at 1, 20, and 60 min, and then analyzed
immediately by LC-MS/MS.
GSH and NAC adduct formation in vivo and in spiked
plasma samples
Eight-week-old male Balb/c mice (n ¼ 3) were purchased
from Charles River Laboratories (Wilmington, MA).
Sterilized 7012 Teklad diet (Harlan, Madison, WI) was used
for mice, and water was provided ad libitum. Mice were
housed in the University of Nebraska Medical Center
(UNMC) laboratory animal facility according to the
American Animal Association and Laboratory Animal Care
guidelines. All procedures were approved by the Institutional
Animal Care and Use Committee at UNMC as set forth by
the National Institutes of Health (NIH). The 10-mg/kg
Stability studies of BI-94 were performed in mouse plasma
(stored in ꢁ80 ꢂC), diluted plasma, and in plasma extracted by
protein precipitation with methanol (post-extracted) plasma at
10 mg/mL concentration at room temperature. One- and 10-ꢀ
diluted plasma were prepared by the addition of one or nine
volumes of water or 7.5 mM ammonium formate buffer (pH
2.5) to one volume of plasma. Diluted and undiluted
plasma samples were extracted by adding 400 lL of methanol
to a 100 lL of blank plasma samples pre-spiked with 10 lL of
a 100 lg/mL BI-94 stock solution, after 1, 20, and 60 min of
incubation time. Samples were then vortexed and centrifuged
at 16 000 ꢀ g for 10 min. A 100 lL of the supernatant was
aspirated, and mixed with a 100 lL of 20% methanol in
7.5 mM ammonium formate buffer (pH 2.5) for LC-MS/MS
analysis.
intravenous dose was administered as
a 100 mL of
2.5 mg/mL BI-94 solution in a dimethyl sulfoxide (DMSO)-
ethanol-polyethylene glycol (PEG) 400-propylene glycol (PG)
mixture (20/20/40/20 v/v). A 100 mL of blood samples were
collected at 5 and 60 min post-dose and 50 mL were used to
obtain plasma, while the other 50 mL were used as whole
blood. Plasma and whole blood samples were immediately
extracted with four volumes of MeOH, centrifuged at
16 000 ꢀ g for 10 min. Supernatants from both plasma and
blood samples were mixed with equal volumes of 20%
methanol in 7.5 mM ammonium formate (pH 2.5) buffer for
LC-MS/MS analysis.
For post-extracted plasma, blank plasma samples were
extracted with 400 mL MeOH, centrifuged at 16 000 ꢀ g, and
the supernatants were collected. A 100 lL of the supernatant
was aspirated, mixed with a 100 lL of 20 % methanol in H2O
or in 7.5 mM ammonium formate buffer (pH 2.5), spiked with
10 lL of a 20 mg/mL BI-94, and analyzed by LC-MS/MS. In
addition to mouse plasma, stability of BI-94 was also
investigated in mouse blood and serum, and in rat, monkey,
and human plasma.
Adduct formation with NAC and GSH was also monitored
after spiking BI-94 in freshly collected mouse blood/plasma
and in frozen plasma. Freshly collected blank blood/plasma
and frozen plasma were spiked with BI-94 at 10 mg/mL
concentration and incubated for 5 and 20 min at room
temperature. Samples were then prepared as described
above for LC-MS/MS analysis. An additional set of frozen
plasma samples was fortified with exogenous GSH at 9.8 mM
before spiking with BI-94.
GSH and NAC adduct formation in solution
The kinetics of adduct formation of BI-94 with GSH and
NAC was also investigated in H2O, PBS, and in 7.5 mM
ammonium formate buffer (pH 2.5). As a control, BI-94
without NAC or GSH was also incubated under the same
conditions. About 10 mL of 1 mg/mL stock solutions of NAC
or GSH and 10 mL BI-94 were added to 980 mL of H2O, PBS,
or buffer to start the reaction. Aliquots of 10 mL were
collected in 90 mL of 55.0% methanol in H2O (1 lg/mL final
BI-94 concentration) at 1, 5, 15, 30, 45, and 60 min, and then
Results and discussion
BI-94 immediately disappeared after spiking in mouse
plasma. To understand the mechanism of its instability in
plasma, the stability of BI-94 was first tested in various
buffers and at different pH levels (Figure 2A). BI-94 was
found to be stable in water and in various buffers at pHs
2.5–5, regardless of the buffer composition. In addition, BI-94

6
N. Gautam et al.
Xenobiotica, Early Online: 1–16
Figure 2. Percentage recovery versus incuba-
tion time profiles of (A) BI-94 in NaOH,
deionized water and in various buffers, and
(B) BI-94 in untreated mouse plasma, water-
diluted plasma, plasma pre-treated with
ammonium formate buffer (pH 2.5), post-
extracted plasma (reconstituted with 50%
MeOH in water) and post-extracted plasma
reconstituted with ammonium formate buffer
(pH 2.5).
was stable in the autosampler for 48 h in the 50 % methanol in
H2O reconstitution solution (data not shown). In contrast, BI-
94 was not stable in NaOH or in various buffers at pHs 7–9,
regardless of the buffer composition. These data indicate that
BI-94 is only stable at acidic pHs and not stable at
physiological or basic pHs. Therefore, we reexamined BI-94
stability in plasma to look for conditions that may increase it.
Recovery of BI-94 from mouse plasma up to 1 h of incubation
time, at various conditions, is shown in Figure 2(B). BI-94
was below detection limit immediately after spiking in mouse
plasma and in plasma diluted with one or nine volumes
of H2O. However, one volume dilution with 7.5 mM
NH4-formate buffer (pH 2.5) increased immediate recovery
from mouse plasma to 62%, which declined to 10% over
60 min of incubation, whereas nine volume dilution with the
same buffer completely recovered BI-94 from plasma.
proposed structure for this 181-MW peak as shown in Figure
1(B). Furthermore, this peak was not detected under acidic
conditions, and its area under the curve was 100-fold higher,
when BI-94 was incubated with ammonium carbonate (pH 9)
buffer compared with H2O.
In addition, the nitro and sulfone groups on the 4- and 7-
positions, respectively, make the 5- and 6-positions on BI-94
susceptible to nucleophillic attacks. For example, the addition
of hydroxide ions following an oxa-Michael mechanism may
take place (Figure 1A). However, we have not detected such
degradation products under our LC-MS conditions.
BI-94 stability was also characterized in post-extracted
plasma, in which plasma samples were initially pre-extracted
with MeOH only or with a combination of MeOH and formate
buffer before spiking with BI-94 (Figure 2B). Recovery of BI-
94 from plasma pretreated with MeOH was a 100% imme-
diately after spiking, however, the recovery decreased to 20%
after a 60-min incubation. In contrast, recovery of BI-94 from
blank plasma pre-treated with a combination of MeOH and
formate buffer remained at a 100% after a 60-min incubation
(Figure 2B). Therefore, pre-treating plasma with MeOH or
acidifying plasma samples with formate buffer enhance BI-94
stability, while a combination of acid and MeOH pretreatment
further enhances it resulting in 100% stability in plasma.
Similar to mouse plasma, BI-94 disappeared immediately
after spiking into monkey, rat, and human, plasma with a
recovery 51% (data not shown).
BI-94 instability in neutral and basic pH conditions can be
attributed to the electron withdrawing functional groups on
BI-94. The strong electron withdrawing nature of nitro and
sulfone functional groups on the 4- and 7-positions makes
these positions susceptible to nucleophillic substitution with
ammonia, carbonate, bicarbonate, and hydroxide ions from
the neutral and basic conditions described above, possibly via
an ipso attack following a Boulton–Katritzky mechanism
(Crampton et al., 2003) (Figure 1A). We were able to detect
an extra peak after BI-94 incubation under various neutral and
basic conditions with a retention time of 2 min (RT of BI-94
was 9 min). MS analysis revealed a MW of 181 for this extra
peak, which matches the MW of the 7-hydroxy-4-nitrobenzo-
furazan degradation product shown in Figure 1. Further
MS/MS analysis supported a fragmentation pattern of the
Because acidifying plasma alone did not result in complete
recovery of BI-94, additional mechanisms beyond pH were
responsible for BI-94 instability in plasma. Therefore, we
evaluated the metabolic stability of BI-94 in mouse plasma by

DOI: 10.3109/00498254.2015.1025250
Irreversible binding of BI-94 to plasma proteins
7
Figure 3. EMS spectra of (A) parent I-94, (B) adducts of BI-94 with GSH (A1 and A2), and (C) adducts of BI-94 with NAC (A1, A2, and A3) after 30-
min incubation in water followed by direct infusion into the MS system.
LC-MS and LC-UV. No additional peaks including peaks
specific to predicted phase I and phase II metabolites were
detected, by UV or by MS, after incubation of BI-94 with
plasma or with post-extracted plasma. Together these studies
suggested irreversible plasma protein binding as an additional
mechanism for the disappearance of BI-94 upon incubation
with plasma.
fragment of BI-94, when incubated in deionized water for
30 min. Similarly three NAC adducts were detected, which
were formed with fragments of NAC and BI-94. EMS spectra,
from direct MS analysis, of parent BI-94, adducts of BI-94
with GSH (A1: 469.4 and A2: 578.4), and adducts of BI-94
with NAC (A1: 196.1, A2: 305.3, and A3: 455.1) are shown in
Figure 3. Chemical structures for these adducts were proposed
based on their fragmentation pattern in MS/MS. The GSH
adduct A1 (m/z 469) produced two product ions of m/z 195.9
and 150.0 (Figure 4A), and adduct A2 (m/z 578.5) produced
two product ions of m/z 304.8 and 150.1 (Figure 4B). The
NAC adduct A1 (m/z 196) produced three product ions of m/z
Adduct formation of BI-94 with the common trapping
agents, GSH and NAC, was used as a surrogate to predict
irreversible binding to plasma proteins. We were not able to
detect intact adducts of GSH and NAC with BI-94. However,
we detected two adducts with GSH, which were formed with a

8
N. Gautam et al.
Xenobiotica, Early Online: 1–16
Figure 4. Product ion spectra of BI-94 with GSH adducts: (A) adduct A1 (m/z – 469) and (B) adduct A2 (m/z – 578).
150, 122.5, and 105.9 (Figure 5A), and adduct A2 (m/z 304.7)
produced four product ions of m/z 155, 150, 105.9, and 63.8
(Figure 5B). The NAC adduct A3 (m/z 455) did not produce
any specific fragmentation pattern (data not shown). During
the synthesis of BI-94, the chlorine atom at the 7-position of
4-nitro-7-chloro benzofurazan was displaced by the sulfhydryl
group in phenylmethanethiol (SNAr type reaction). Therefore,
plausible mechanisms for the formation of these adducts with
GSH and NAC are the addition of sulfhydryl groups to the
5-position via a Michael type reaction and SNAr type
substitution of NO2 group and/or the substituted sulfone.
In addition to the above-mentioned direct MS analysis, the
formation of GSH adducts with BI-94 in deionized water was
further investigated by LC-MS/MS using various scans and
with the isotope filtering technique. EMS scans in both
positive and negative ionization modes failed to show the
formation of BI-94-GSH adducts due to the lack of sensitivity
and specificity (Figure 6A and B). More specific scans
previously reported for the detection of GSH adducts
(Li et al., 2011a; Ma & Subramanian, 2006) including
precursor scans of 130 (positive ionization), 254, and 272
(negative ionization), m/z as well as neutral loss scans of 129
and 307 (positive ionization), were also investigated
(Figure 6). The precursor ion scan of 254 and 272 m/z
failed to detect any adducts due to the lack of sensitivity
(Figure 6C and D). The precursor ion scan of 130 m/z and the
neutral loss scans of 129 and 307 m/z were able to detect GSH
adducts (Figure 6E–G). As expected, the MRM scans
demonstrated the highest sensitivity and selectivity in detect-
ing GSH adducts (Figure 6H). However, MRM scans can only
detect pre-identified adducts specific for the compounds of
interest, for which the MS was specifically tuned and
optimized. In contrast all other scans are generic scans
designed to detect GSH adducts regardless what compounds
they are conjugated with; therefore, are more likely to detect
unpredicted adducts.
The same scans described above were used in combination
with isotopic filtering. In isotope filtering, a mixture of stable

DOI: 10.3109/00498254.2015.1025250
Irreversible binding of BI-94 to plasma proteins
9
Figure 5. Product ion spectra of BI-94 with NAC adducts: (A) adduct A1 (m/z – 196) and (B) adduct A2 (m/z – 305).
isotope-labeled compound and its un-labeled counterpart in a
pre-determined ratio is used to tag all ions related to this
compound with a signature isotopic pattern. In this context,
we utilized a 1:1 mixture of labeled and unlabeled GSH with
a 3-Da mass difference. Therefore, all adducts resulting from
conjugation of substrates with GSH can be filtered out from
the TIC chromatogram, as ion pairs with a 3-Da mass
difference and a 1:1 abundance ratio. Figure 7 shows
representative chromatograms from the 129 m/z neutral loss
scan with and without isotopic filtering. Isotopic filtering has
clearly increased sensitivity and selectivity (Figure 7B). More
importantly isotopic filtering only extracted ion pairs with the
pre-identified isotopic pattern (1:1 of M:M+3 isotopes),
which resulted in the unambiguous identification of any ions
related to GSH including the GSH-BI-94 adducts (A1 and
A2), GSH itself, and a new GSH-BI-94 adduct (A3) that was
not detected using all approaches described earlier.
characterize the kinetics of GSH and NAC adduct formation
with BI-94 at physiological pH using PBS. BI-94 and GSH
were gradually depleted while A1 and A2 adducts were
gradually formed over the 1-h course of incubation in water
(Figure 8A). After 60 min of incubation, the peak areas of BI-
94 and GSH decreased by 64%, whereas there was a 6-fold
increase in the peak area of A1 and A2 adducts. When
incubation was performed in PBS, there was an abrupt
depletion (94.5% decrease in peak area) of BI-94, and the
maximum formation of GSH adducts was observed within
5 min of incubation (Figure 8B). However, when the reaction
was performed in NH4-formate buffer of pH 2.5, only 14% of
BI-94 was depleted over 1 h, and both A1 and A2 were barely
detected (Figure 8C). A similar trend was observed from the
kinetics of adduct formation of BI-94 with NAC, in water,
PBS, and ammonium formate buffer (pH 2.5) (data not
shown). Therefore, neutral/physiological pHs enhanced and
accelerated the GSH and NAC adduct formation with BI-94,
whereas acidic pHs completely prevented it. This can
After optimizing and establishing the methods to detect
and quantify adduct formation, we used these methods to

10 N. Gautam et al.
Xenobiotica, Early Online: 1–16
Figure 6. Representative LC-MS chromatograms of GSH adducts with BI-94 after 30-min incubation in water: (A) enhanced MS scan in the positive
mode, (B) enhanced MS scan in the negative mode, (C) precursor ion scan of m/z 254 in the negative mode, (D) precursor ion scan of m/z 272 in
the negative mode, (E) precursor ion scan of m/z 130 in the positive mode, (F) neutral loss scan of m/z 129 in the positive mode, (G) neutral loss scan for
m/z 307 in the positive mode, and (H) MRM scans (A1 m/z 469; A2 m/z 578) in the negative mode.
contribute to the instability of BI-94 in plasma, where the
physiological pH not only triggered the chemical instability of
BI-94, as discussed earlier (Figure 2A) but also enhanced
adduct formation with endogenous nucleophiles.
(Figure 9). In addition, BI-94 adducts with GSH (A1 and A2),
NAC (A1 and A2), and GSH itself were detected in vivo from
plasma samples collected 5 and 60 min after 10 mg/kg
intravenous dose administration to mice (Figure 9G and H).
These NAC adducts which have only thiol addition could
come from any endogenous thiol donor, including NAC.
Furthermore, all three mice died within 2 h after BI-94
When BI-94 was incubated with blank plasma freshly
collected from mice, the GSH adducts (A1 and A2), a NAC
adduct (A1), and endogenous GSH itself were detected

DOI: 10.3109/00498254.2015.1025250
Irreversible binding of BI-94 to plasma proteins 11
Figure 7. Representative LC-MS chromatograms and spectrums of BI-94 adducts with GSH for neutral loss of m/z 129 after 30-min incubation
in water, (A) total ion chromatogram (TIC), (B) the same TIC with a 3-Da isotopic filter, and (C-1, C-2, and C-3) extracted mass spectrums from
peaks in B.
administration. Collectively, these data provide strong evi-
dence suggesting that irreversible binding of BI-94 to plasma
proteins and/or other nucleophiles takes place in vivo.
levels of GSH in human plasma in vivo). Surely, we detected
the same GSH adducts in these GSH-fortified samples at
similar levels to the ones in fresh plasma samples (data not
shown). Therefore, the use of frozen blank plasmas to make
conclusions about the formation of GSH and NAC adducts
can be misleading due to GSH instability and rather fresh
plasma should be used instead. The instability of GSH itself
and/or GSH adducts after their formation may preclude their
detection in vitro and in vivo. Similar results were obtained
when whole in vivo (mice dosed with BI-94) and spiked blood
samples rather than plasma were analyzed (data not shown).
The peak area of GSH adducts was up to nine-fold higher in
blood compared with plasma, because the endogenous level
of GSH in mouse blood was higher compared with plasma
(GSH peak area was 57-fold higher in mouse blood compared
with plasma). Similarly, it was previously reported that the
In contrast, none of the adducts or the endogenous GSH or
NAC themselves, which we detected in fresh plasma, were
detected when BI-94 was spiked into frozen blank mouse
plasma samples. This may be because GSH is not stable for
long periods of time and can undergo non-enzymatic
oxidation in biological samples (Squellerio et al., 2012).
The reported endogenous levels of GSH in human blood and
plasma were 849 63 mM and 3.39 1.04 mM, respectively
(Michelet et al., 1995). Whereas the reported NAC endogen-
ous concentration in human plasma was 0.08 mM (Gabard &
Mascher, 1991). Consequently, we monitored the formation of
BI-94 adducts in frozen blank mouse plasma after fortifying
them with 9.8 mM of exogenous GSH (near physiological

12 N. Gautam et al.
Xenobiotica, Early Online: 1–16
Figure 8. Peak area-time profile for BI-94,
GSH, and their A1 and A2 adducts in (A)
H2O, (B) PBS and, (C) NH4-formate buffer
(pH 2.5), in binding study for 1 h.
endogenous level of GSH in human blood was 250-fold
higher compared with plasma (Michelet et al., 1995).
EMS scans, when BI-94 incubated with HSA-T3 peptide in
deionized water for 30 min. These adducts were further
analyzed for their fragmentation pattern in MS/MS to
determine the site of adduct formation. As shown in Table 1
and Figure 10, mass shifts of 164 (A1) and 273 (A2), as a
result of adduct formation with BI-94, were only observed
with HSA-T3 fragments that contain the Cys34. For example,
the HSA-T3 fragment y7 did not show any mass shift after
incubation with BI-94. Whereas the y8 fragment, which is the
same as y7 with the addition of Cys34, showed both 164 (A1)
and 273 (A2) mass shifts. This provides an evidence that
BI-94 forms adducts with albumin at the Cys34 position.
In conclusion, we investigated the mechanisms responsible
for the instability of BI-94 in vivo and in vitro. Studies with
various buffers at different pHs suggested that BI-94 was not
stable in solution at neutral or basic pHs as well as in plasma,
while it was stable at acidic pHs. The recovery of BI-94 from
plasma can be increased through acidification. Our data also
show that BI-94 undergoes attack by nucleophiles including
GSH, NAC, and T3-HAS which suggests that BI-94 binds
Albumin is the most abundant plasma protein. Therefore,
detection of a compound–albumin complex would provide a
stronger evidence for irreversible binding with plasma
proteins compared with using surrogate trapping agents
such as GSH and NAC. However, detection of such large
complexes is not always feasible (Jahn et al., 2012a).
Therefore, large peptides resulting from albumin digestion
and containing the free Cys34, which is thought as the
primary albumin binding site, are used as surrogates for
binding to intact albumin. An example of such peptides is
HSA-T3, which was shown to be the only peptide resulting
from albumin digestion that bind covalently to a variety of
electrophiles (aldehydes, nitrogen mustards, oxiranes, quin-
ones, metal ions, and small molecules) in a similar way as
intact albumin (Li et al., 2011b; Rubino et al., 2009). In this
study similar to GSH, adducts with fragments of BI-94, A1
(m/z+3 ¼ 866.8 ¼ HSA-T3+164) and A2 (m/z+3 ¼ 903.1 ¼
HSA-T3+273), rather than intact BI-94 were detected in

DOI: 10.3109/00498254.2015.1025250
Irreversible binding of BI-94 to plasma proteins 13
Figure 9. Representative LC-MS/MS chromatograms of (A) GSH adducts with BI-94 after 30-min incubation in water, (B) NAC adducts with BI-94
after 30-min incubation in water, (C) endogenous GSH in fresh plasma, (D) endogenous NAC in fresh plasma, (E) GSH adducts with BI-94 5 min after
spiking in fresh plasma, and (F) NAC adducts with BI-94 5 min after spiking in fresh plasma, (G) GSH adducts in plasma samples after 5 min of
intravenous dose administration, and (H) NAC adducts in plasma samples after 5 min of intravenous dose administration.
Table 1. Observed MS/MS fragments for HSA-T3 peptide and its BI-94 adducts.
HSA-T3
fragments
m/z
Charge (HSA-T3)
HSA-T3
fragments
m/z
Charge (HSA-T3) (BI-94 + HSA-T3)
m/za
Mass
m/za (BI-94 + HSA-T3)
Mass
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
y1
y2
y3
y4
y5
y6
y7
y8
145.1
244.2
381.2
496.2
625.3
772.4
869.4
972.4
1100.5
1228.5
1341.6
1504.7
1632.7
1703.8
b20
b19
b18
b17
b16
b15
b14
b13
b12
b11
b10
b9
2287.2
2188.1
2051.1
1936.0
1806.9
1659.9
1562.8
1459.8
1331.8
1203.7
1090.6
927.6
+
+
+
+
+
+
+
+
+
383.3
498.2
627.4
774.2
871.4
974.4
1102.5
1230.6
1343.5
No modification
No modification
No modification
No modification
No modification
+
+
+
+
+
+
+
+
1562.8
1459.6
1331.7
1203.6
1090.6
927.4
1137.4 (+164), 1247.5 (+273)
b
y9
y10
y11
y12
y13
y14
1394.6 (+164), 1503.6 (+273)
No modification
b
b
b
b
b
+
1506.6
b
b
b8
b7
799.5
728.5
799.5
728.7
No modification
No modification
++
853.4
936.5 (+164/2), 989.9 (+273/2)
(continued )

14 N. Gautam et al.
Xenobiotica, Early Online: 1–16
Table 1. Continued
HSA-T3
fragments
m/z
Charge (HSA-T3)
HSA-T3
fragments
m/z
m/za
Charge (HSA-T3) (BI-94 + HSA-T3)
Mass
m/za (BI-94 + HSA-T3)
Mass
y15
y16
y17
y18
y19
y20
1850.8
1921.9
2034.9
2148.1
2247.1
2360.2
++
++
++
++
927.4
963.0
1009.4 (+164/2), 1063.4 (+273/2)
1044.5 (+164/2), 1099.5 (+273/2)
1101.4 (+164/2), 1155.5 (+273/2)
1158.6 (+164/2), 1112.1 (+273/2)
b6
b5
b4
b3
b2
b1
581.4
510.4
397.3
284.2
185.1
72.1
+
+
+
+
581.4
510.4
397.4
284.2
No modification
No modification
No modification
No modification
1019.5
1076.0
1125.5
++
1208.6 (+164/2), 1261.7 (+273/2)
+
185.1
No modification
b
b
b
b
b
b
aObserved m/z after mass shift resulting from adduct formation with BI-94 [A1 (+164) and A2 (+273)]; no modification: mass of HSA-T3 fragments
were not changed after incubation with BI-94.
bNot detected.
Figure 10. (A) Theoretical fragmentation pattern for HSA-T3 peptide, EPI spectra of (B-I) HSA-T3 ¼ 812.5+3 m/z, (B-II) adduct A1 ¼ 866.8+3 m/z
(HSA-T3 +164), and (B-III) adduct A2 ¼ 903.1+3 m/z (HSA-T3 + 273). Mass shifts were detected for circled fragments, as a result of adduct formation
with BI-94 at the cys34 amino acid position.

DOI: 10.3109/00498254.2015.1025250
Irreversible binding of BI-94 to plasma proteins 15
Ito K, Iwatsubo T, Kanamitsu S, et al. (1998). Prediction of
pharmacokinetic alterations caused by drug–drug interactions: meta-
bolic interaction in the liver. Pharmacol Rev 50:387–412.
Jahn S, Faber H, Zazzeroni R, Karst U. (2012a). Electrochemistry/liquid
chromatography/mass spectrometry to demonstrate irreversible bind-
ing of the skin allergen p-phenylenediamine to proteins. Rapid
Commun Mass Spectrom 26:1415–25.
Jahn S, Faber H, Zazzeroni R, Karst U. (2012b). Electrochemistry/mass
spectrometry as a tool in the investigation of the potent skin sensitizer
p-phenylenediamine and its reactivity toward nucleophiles. Rapid
Commun Mass Spectrom 26:1453–64.
Jenkins RE, Meng X, Elliott VL, et al. (2009). Characterisation of
flucloxacillin and 5-hydroxymethyl flucloxacillin haptenated HSA in
vitro and in vivo. Proteomics Clin Appl 3:720–9.
Jenkinson C, Jenkins RE, Maggs JL, et al. (2009). A mechanistic
investigation into the irreversible protein binding and antigenicity of
p-phenylenediamine. Chem Res Toxicol 22:1172–80.
irreversibly to plasma proteins including albumin. Therefore,
the instability of BI-94 in plasma is likely due to a
combination of chemical instability and irreversible binding
to plasma proteins and other nucleophiles. Irreversible protein
binding of BI-94 may be the reason behind its strong toxicity
in vivo. Further structural modifications of BI-94 are required
to improve its stability and safety.
Declaration of interest
This work was supported National Institutes of Health (Grants
DA028555-01 and CA127239). The authors report no
conflicts of interest. The authors alone are responsible for
the content and writing of the paper.
Jinno F, Yoneyama T, Morohashi A, et al. (2011). Chemical reactivity of
ethyl (6R)-6-[N-(2-chloro-4-fluorophenyl)sulfamoyl]cyclohex-1-ene-
1-carboxylate (TAK-242) in vitro. Biopharm Drug Dispos 32:408–25.
Joseph PR, Yuan Z, Kumar EA, et al. (2010). Structural characterization
of BRCT-tetrapeptide binding interactions. Biochem Biophys Res
Commun 393:207–10.
Kalgutkar AS, Didiuk MT. (2009). Structural alerts, reactive metabolites,
and protein covalent binding: how reliable are these attributes as
predictors of drug toxicity? Chem Biodivers 6:2115–37.
Kappus H, Bolt HM, Remmer H. (1973). Irreversible protein binding of
metabolites of ethynylestradiol in vivo and in vitro. Steroids 22:
203–25.
Kerksick C, Willoughby D. (2005). The antioxidant role of glutathione
and N-acetyl-cysteine supplements and exercise-induced oxidative
stress. J Int Soc Sports Nutr 2:38–44.
Kessler U, Castagnolo D, Pagano M, et al. (2013). Discovery and
synthesis of novel benzofurazan derivatives as inhibitors of influenza
A virus. Bioorg Med Chem Lett 23:5575–7.
References
Ballard P, Rowland M. (2011). Correction for nonspecific binding to
various components of ultrafiltration apparatus and impact on
estimating in vivo rat clearance for a congeneric series of 5-ethyl,5-
n-alkyl barbituric acids. Drug Metab Dispos 39:2165–8.
Banker MJ, Clark TH. (2008). Plasma/serum protein binding determin-
ations. Curr Drug Metab 9:854–9.
Barre J, Chamouard JM, Houin G, Tillement JP. (1985). Equilibrium
dialysis, ultrafiltration, and ultracentrifugation compared for deter-
mining the plasma-protein-binding characteristics of valproic acid.
Clin Chem 31:60–4.
Cantor SB, Bell DW, Ganesan S, et al. (2001). BACH1, a novel helicase-
like protein, interacts directly with BRCA1 and contributes to its DNA
repair function. Cell 105:149–60.
Chan S, Gerson B. (1987). Free drug monitoring. Clin Lab Med 7:
279–87.
Kim H, Huang J, Chen J. (2007). CCDC98 is a BRCA1-BRCT domain-
binding protein involved in the DNA damage response. Nat Struct Mol
Biol 14:710–15.
Li F, Lu J, Ma X. (2011a). Profiling the reactive metabolites of
xenobiotics using metabolomic technologies. Chem Res Toxicol 24:
744–51.
Chuang VT, Maruyama T, Otagiri M. (2009). Updates on contemporary
protein binding techniques. Drug Metab Pharmacokinet 24:358–64.
Crampton MR, Lunn RE, Lucas D. (2003). Sigma-adduct formation and
oxidative substitution in the reactions of 4-nitrobenzofurazan and
some derivatives with hydroxide ions in water. Org Biomol Chem 1:
3438–43.
Li H, Grigoryan H, Funk WE, et al. (2011b). Profiling Cys34 adducts of
human serum albumin by fixed-step selected reaction monitoring. Mol
Cell Proteomics 10: M110 004606.
Dasgupta A. (2007). Usefulness of monitoring free (unbound) concen-
trations of therapeutic drugs in patient management. Clin Chim Acta
377:1–13.
Liao S, Ewing NP, Boucher B, et al. (2012). High-throughput screening
for glutathione conjugates using stable-isotope labeling and negative
electrospray ionization precursor-ion mass spectrometry. Rapid
Commun Mass Spectrom 26:659–69.
Liu X, Shen Q, Li J, et al. (2013). In silico prediction of cytochrome
P450-mediated site of metabolism (SOM). Protein Pept Lett 20:
279–89.
Liu Z, Wu J, Yu X. (2007). CCDC98 targets BRCA1 to DNA damage
sites. Nat Struct Mol Biol 14:716–20.
Lokesh GL, Rachamallu A, Kumar GD, Natarajan A. (2006). High-
throughput fluorescence polarization assay to identify small molecule
inhibitors of BRCT domains of breast cancer gene 1. Anal Biochem
352:135–41.
Dubois N, Lapicque F, Maurice MH, et al. (1993). In vitro irreversible
binding of ketoprofen glucuronide to plasma proteins. Drug Metab
Dispos 21:617–23.
Enoch SJ, Ellison CM, Schultz TW, Cronin MT. (2011). A review of the
electrophilic reaction chemistry involved in covalent protein binding
relevant to toxicity. Crit Rev Toxicol 41:783–802.
Federici L, Lo Sterzo C, Pezzola S, et al. (2009). Structural basis for the
binding of the anticancer compound 6-(7-nitro-2,1,3-benzoxadiazol-4-
ylthio)hexanol to human glutathione s-transferases. Cancer Res 69:
8025–34.
Freed AL, Strohmeyer HE, Mahjour M, et al. (2008). pH control of
nucleophilic/electrophilic oxidation. Int J Pharm 357:180–8.
Gabard B, Mascher H. (1991). Endogenous plasma N-acetylcysteine and
single dose oral bioavailability from two different formulations as
determined by a new analytical method. Biopharm Drug Dispos 12:
343–53.
Ma S, Subramanian R. (2006). Detecting and characterizing reactive
metabolites by liquid chromatography/tandem mass spectrometry. J
Mass Spectrom 41:1121–39.
Manke IA, Lowery DM, Nguyen A, Yaffe MB. (2003). BRCT repeats as
phosphopeptide-binding modules involved in protein targeting.
Science 302:636–9.
Masubuchi N, Makino C, Murayama N. (2007). Prediction of in vivo
potential for metabolic activation of drugs into chemically reactive
intermediate: correlation of in vitro and in vivo generation of reactive
intermediates and in vitro glutathione conjugate formation in rats and
humans. Chem Res Toxicol 20:455–64.
Meng X, Jenkins RE, Berry NG, et al. (2011). Direct evidence for the
formation of diastereoisomeric benzylpenicilloyl haptens from
benzylpenicillin and benzylpenicillenic acid in patients. J Pharmacol
Exp Ther 338:841–9.
Gautam N, Bathena SP, Chen Q, et al. (2013). Pharmacokinetics, protein
binding and metabolism of a quinoxaline urea analog as an NF-
kappaB inhibitor in mice and rats by LC-MS/MS. Biomed
Chromatogr 27:900–9.
Ghosh PB, Ternai B, Whitehouse MW. (1972). Potential antileukemic
and immunosuppressive drugs. 3. Effects of homocyclic ring substi-
tution on the in vitro drug activity of 4-nitrobenzo-2,1,3-oxadiazoles
(4-nitrobenzofurazans) and their N-oxides (4-nitrobenzofuroxans).
J Med Chem 15:255–60.
Howard ML, Hill JJ, Galluppi GR, McLean MA. (2010). Plasma protein
binding in drug discovery and development. Comb Chem High
Throughput Screen 13:170–87.
Meng X, Maggs JL, Usui T, et al. (2015). Auto-oxidation of isoniazid
leads to isonicotinic-lysine adducts on human serum albumin. Chem
Res Toxicol 28:51–8.
Hyneck ML, Smith PC, Munafo A, et al. (1988). Disposition and
irreversible plasma protein binding of tolmetin in humans. Clin
Pharmacol Ther 44:107–14.

16 N. Gautam et al.
Xenobiotica, Early Online: 1–16
Michelet F, Gueguen R, Leroy P, et al. (1995). Blood and plasma
glutathione measured in healthy subjects by HPLC: relation to sex,
aging, biological variables, and life habits. Clin Chem 41:1509–17.
Mutlib A, Lam W, Atherton J, et al. (2005). Application of stable isotope
labeled glutathione and rapid scanning mass spectrometers in
detecting and characterizing reactive metabolites. Rapid Commun
Mass Spectrom 19:3482–92.
Nakayama S, Atsumi R, Takakusa H, et al. (2009). A zone classification
system for risk assessment of idiosyncratic drug toxicity using daily
dose and covalent binding. Drug Metab Dispos 37:1970–7.
Nakayama S, Takakusa H, Watanabe A, et al. (2011). Combination of
GSH trapping and time-dependent inhibition assays as a predictive
method of drugs generating highly reactive metabolites. Drug Metab
Dispos 39:1247–54.
Nassar AE, Lopez-Anaya A. (2004). Strategies for dealing with reactive
intermediates in drug discovery and development. Curr Opin Drug
Discov Dev 7:126–36.
Paliwal JK, Smith DE, Cox SR, et al. (1993). Stereoselective, competi-
tive, and nonlinear plasma protein binding of ibuprofen enantiomers
as determined in vivo in healthy subjects. J Pharmacokinet Biopharm
21:145–61.
Shibukawa A, Kuroda Y, Nakagawa T. (1999). High-performance frontal
analysis for drug–protein binding study. J Pharm Biomed Anal 18:
1047–55.
Simeonov A, Yasgar A, Jadhav A, et al. (2008). Dual-fluorophore
quantitative high-throughput screen for inhibitors of BRCT–phospho-
protein interaction. Anal Biochem 375:60–70.
Smith PC, McDonagh AF, Benet LZ. (1986). Irreversible binding of
zomepirac to plasma protein in vitro and in vivo. J Clin Invest 77:
934–9.
Sobhian B, Shao G, Lilli DR, et al. (2007). RAP80 targets BRCA1 to
specific ubiquitin structures at DNA damage sites. Science 316:
1198–202.
Squellerio I, Caruso D, Porro B, et al. (2012). Direct glutathione
quantification in human blood by LC-MS/MS: comparison with
HPLC with electrochemical detection. J Pharm Biomed Anal 71:
111–18.
Tang W, Lu AY. (2010). Metabolic bioactivation and drug-related
adverse effects: current status and future directions from a pharma-
ceutical research perspective. Drug Metab Rev 42:225–49.
Uetrecht J. (2003). Screening for the potential of a drug candidate to
cause idiosyncratic drug reactions. Drug Discov Today 8:832–7.
Wang B, Matsuoka S, Ballif BA, et al. (2007). Abraxas and RAP80 form
a BRCA1 protein complex required for the DNA damage response.
Science 316:1194–8.
Wang L, Zhang YY, Liu FY, et al. (2014). Benzofurazan derivatives as
antifungal agents against phytopathogenic fungi. Eur J Med Chem 80:
535–42.
Wen B, Fitch WL. (2009). Analytical strategies for the screening and
evaluation of chemically reactive drug metabolites. Expert Opin Drug
Metab Toxicol 5:39–55.
Patridge EV, Eriksson ES, Penketh PG, et al. (2012). 7-Nitro-4-
(phenylthio)benzofurazan is a potent generator of superoxide and
hydrogen peroxide. Arch Toxicol 86:1613–25.
Pessetto ZY, Yan Y, Bessho T, Natarajan A. (2012). Inhibition of
BRCT(BRCA1)-phosphoprotein interaction enhances the cytotoxic
effect of olaparib in breast cancer cells: a proof of concept study for
synthetic lethal therapeutic option. Breast Cancer Res Treat 134:
511–17.
Prakash C, Sharma R, Gleave M, Nedderman A. (2008). In vitro
screening techniques for reactive metabolites for minimizing bioacti-
vation potential in drug discovery. Curr Drug Metab 9:952–64.
Rappaport SM, Li H, Grigoryan H, et al. (2012). Adductomics:
characterizing exposures to reactive electrophiles. Toxicol Lett 213:
83–90.
Ricci G, De Maria F, Antonini G, et al. (2005). 7-Nitro-2,1,3-
benzoxadiazole derivatives, a new class of suicide inhibitors for
glutathione S-transferases. Mechanism of action of potential antic-
ancer drugs. J Biol Chem 280:26397–405.
Xie F, Li BX, Broussard C, Xiao X. (2013). Identification, synthesis and
evaluation of substituted benzofurazans as inhibitors of CREB-
mediated gene transcription. Bioorg Med Chem Lett 23:5371–5.
Yan Z, Caldwell G. (2004). Optimization in drug discovery: in vitro
methods. Totowa, NJ: Humana Press.
Yang Y, Guan X. (2015). Rapid and thiol-specific high-throughput assay
for simultaneous relative quantification of total thiols, protein thiols,
and nonprotein thiols in cells. Anal Chem 87:649–55.
Yu X, Chini CC, He M, et al. (2003). The BRCT domain is a phospho-
protein binding domain. Science 302:639–42.
Rubino FM, Pitton M, Di Fabio D, Colombi A. (2009). Toward an
‘‘omic’’ physiopathology of reactive chemicals: thirty years of mass
spectrometric study of the protein adducts with endogenous and
xenobiotic compounds. Mass Spectrom Rev 28:725–84.
Schmidt S, Gonzalez D, Derendorf H. (2010). Significance of protein
binding in pharmacokinetics and pharmacodynamics. J Pharm Sci 99:
1107–22.
Schuhmacher J, Kohlsdorfer C, Buhner K, et al. (2004). High-throughput
determination of the free fraction of drugs strongly bound to plasma
proteins. J Pharm Sci 93:816–30.
Yuan J, Yang DC, Birkmeier J, Stolzenbach J. (1995). Determination of
protein binding by in vitro charcoal adsorption. J Pharmacokinet
Biopharm 23:41–55.
Yuan Z, Kumar EA, Campbell SJ, et al. (2011a). Exploiting the P-1
pocket of BRCT domains toward a structure guided inhibitor design.
ACS Med Chem Lett 2:764–7.
Yuan Z, Kumar EA, Kizhake S, Natarajan A. (2011b). Structure-activity
relationship studies to probe the phosphoprotein binding site on the
carboxy terminal domains of the breast cancer susceptibility gene 1.
J Med Chem 54:4264–8.