A suite of activity-based probes for human cytochrome P450 enzymes

Article abstract of DOI:10.1021/ja9037609

Cytochrome P450 (P450) enzymes regulate a variety of endogenous signaling molecules and play central roles in the metabolism of xenobiotics and drugs. We recently showed that an aryl alkyne serves as an effective activity-based probe for profiling mouse liver microsomal P450s in vitro and in vivo. However, individual P450s display distinct substrate and inhibitor specificities, indicating that multiple probe structures may be required to achieve comprehensive coverage of this large and diverse enzyme family. Here, we have synthesized a suite of P450-directed, activity-based protein profiling (ABPP) probes that contain: (1) varied chemical architectures validated as mechanism-based inhibitors of the P450 enzyme family, and (2) terminal alkyne groups for click chemistry conjugation of reporter tags. This set of probes was screened against a wide cross-section of human P450s, leading to the discovery of an optimal set of probes that provide broad coverage of this enzyme family. We used these probes to profile the effects on P450 activity of aromatase inhibitors in current clinical use for the treatment of breast cancer. We describe the surprising discovery that one of these aromatase inhibitors, anastrozole, significantly increases probe-labeling of P450 1A2, indicative of a heterotypic cooperativity effect on a central P450 isozyme involved in metabolizing numerous drugs and xenobiotics. The results presented herein greatly expand the suite of ABPP probes for profiling P450s and illuminate new applications for these tools to understand P450-drug interactions.

Full text of DOI:10.1021/ja9037609

Published on Web 07/07/2009
A Suite of Activity-Based Probes for Human Cytochrome P450
Enzymes
Aaron T. Wright, Joongyu D. Song, and Benjamin F. Cravatt*
The Skaggs Institute for Chemical Biology and Department of Chemical Physiology, The Scripps
Research Institute, 10550 North Torrey Pines Road, La Jolla, California 92037
Received May 11, 2009; E-mail: cravatt@scripps.edu
Abstract: Cytochrome P450 (P450) enzymes regulate a variety of endogenous signaling molecules and
play central roles in the metabolism of xenobiotics and drugs. We recently showed that an aryl alkyne
serves as an effective activity-based probe for profiling mouse liver microsomal P450s in vitro and in vivo.
However, individual P450s display distinct substrate and inhibitor specificities, indicating that multiple probe
structures may be required to achieve comprehensive coverage of this large and diverse enzyme family.
Here, we have synthesized a suite of P450-directed, activity-based protein profiling (ABPP) probes that
contain: (1) varied chemical architectures validated as mechanism-based inhibitors of the P450 enzyme
family, and (2) terminal alkyne groups for click chemistry conjugation of reporter tags. This set of probes
was screened against a wide cross-section of human P450s, leading to the discovery of an optimal set of
probes that provide broad coverage of this enzyme family. We used these probes to profile the effects on
P450 activity of aromatase inhibitors in current clinical use for the treatment of breast cancer. We describe
the surprising discovery that one of these aromatase inhibitors, anastrozole, significantly increases probe-
labeling of P450 1A2, indicative of a heterotypic cooperativity effect on a central P450 isozyme involved in
metabolizing numerous drugs and xenobiotics. The results presented herein greatly expand the suite of
ABPP probes for profiling P450s and illuminate new applications for these tools to understand P450-drug
interactions.
Introduction
genicity, as well as certifying them as targets for inhibition and
pro-drug activation.6-9
The human genome encodes 57 putatively functional cyto-
chrome P450 (P450) enzymes,1 which are membrane-bound
monooxygenases that catalyze the metabolism of a large and
diverse number of xenobiotics, drugs, and endogenous signaling
molecules.2 Despite sharing a conserved structural core with
an active site heme group, these enzymes are remarkably
diverse, exhibiting as little as 16% sequence homology.3 P450s
catalyze a multitude of reactions, including C-hydroxylation,
heteroatom oxygenation, dealkylation, epoxide formation, group
migration, and bond scission.4 Several P450 subfamilies have
dynamic active site architectures that permit broad substrate
specificity5 and account for their participation in catalyzing the
metabolism of 90% of all pharmaceuticals. Other P450 sub-
families are involved in the regulation of endogenous molecules
such as bioactive sterols, fatty acids, and eicosanoids. The
widespread expression and catalytic activity of the P450 family
has implicated these enzymes in cellular toxicity and carcino-
The expression, and overexpression, of P450s in tumor cells
is well-recognized.10-13 Cytochrome P450s are key enzymes
in cancer formation and cancer treatment, mediating the
metabolic activation of numerous procarcinogens and participat-
ing in the activation and inactivation of anticancer drugs. P450
1B1, a xenobiotic metabolizing enzyme, is overexpressed in a
number of tumor types, and therapeutic strategies have been
implemented to harness the catalytic reactivity of P450 1B1 to
activate pro-drugs, and to determine the extent of procarcinogen
activation.14,15 Multiple successful strategies to inhibit P450s,
including P450 19A1 (aromatase) to treat breast cancer and P450
17A1 to combat androgen-dependent prostate cancer, have also
been implemented.9,16,17 The catalytic activity of a number of
human P450s remains unannotated, but several are expressed
in tumors, suggesting a potentially broad role for this enzyme
family in cancer.18
(6) Johnson, W. W. Curr. Drug Metab. 2008, 9, 344–51.
(7) Guengerich, F. P. Chem. Res. Toxicol. 2008, 21, 70–83.
(8) Ioannides, C.; Lewis, D. F. Curr. Top. Med. Chem. 2004, 4, 1767–
88.
(1) Guengerich, F. P.; Wu, Z.-L.; Bartleson, C. J. Biochem. Biophys. Res.
Commun. 2005, 338, 465–69.
(2) Denisov, I. L.; Makris, T. M.; Sligar, S. G.; Schlichting, I. Chem.
ReV. 2005, 105, 2253–77.
(9) Bruno, R. D.; Njar, V. C. O. Bioorg. Med. Chem. 2007, 15, 5047–60.
(10) Agu´ndez, J. A. G. Curr. Drug Metab. 2004, 5, 211–24.
(11) McFadyen, M. C. E.; Melvin, W. T.; Murray, G. I. Mol. Cancer Ther.
2004, 3, 363–71.
(3) Meunier, B.; de Visser, S. P.; Shaik, S. Chem. ReV. 2004, 104, 3947–
3980.
(4) Lamb, D. C.; Waterman, M. R.; Kelly, S. L.; Guengerich, F. P. Curr.
Opin. Biotechnol. 2007, 18, 504–12.
(12) Modungo, F.; Knoll, C.; Kanbour-Shakir, A.; Romkes, M. Breast
Cancer Res. Treat. 2003, 82, 191–7.
(5) Negishi, M.; Uno, T.; Darden, T. A.; Sueyoshi, T.; Pedersen, L. G.
FASEB J. 1996, 10, 683–9.
(13) Patterson, L. H.; Murray, G. I. Curr. Pharm. Des. 2002, 8, 1335–47.
9
10692 J. AM. CHEM. SOC. 2009, 131, 10692–10700
10.1021/ja9037609 CCC: $40.75  2009 American Chemical Society

Activity-Based Probe Suite for Profiling Cytochrome P450s
A R T I C L E S
Figure 1. Structures of P450-directed activity-based probes (ABPs).
Figure 2. Modes of mechanism-based inactivation of cytochrome P450 enzymes. (A) Aryl alkynes are oxidized to reactive ketene intermediates that inactivate
P450s by covalent adduction to the apoprotein. (B) Propynyl-bearing inhibitors are oxidized to a reactive Michael acceptor and inactivate P450s by covalent
adduction to the heme or apoprotein. (C) Furanocoumarin inhibitors are oxidized to the reactive furan epoxide and form covalent adducts to the P450
apoprotein.
The study of P450 drug metabolism is a critical element
of the pharmaceutical development process. In vitro systems,
such as liver microsomes, hepatocyte cultures, and recom-
binant P450s, are used to evaluate the effects of new
pharmaceutical entities on P450 activity and potential
drug-drug interactions. These efforts rely largely on the
coadministration of drugs with P450 substrates and thus are
only applicable to relatively well-characterized members of
the P450 family. Moreover, in vitro assays may fail to account
for myriad post-translational mechanisms that are known to
regulate P450 activity in cells. Toward the goal of developing
a general strategy to profile P450 activities in native
biological systems, we generated a P450-directed activity-
based protein profiling (ABPP) probe, 1 (Figure 1, referred
to as 2EN-ABP in previous reference).19 This probe co-
valently labels P450s in an activity-dependent manner, and
these labeling events were detected by appending a fluores-
cent reporter group via “click chemistry” onto probe-P450
adducts.19,20 Notably, 1 proved capable of profiling P450
activities in both native liver proteomes and living mice, thus
adding P450s to the growing list of enzyme classes that can
be characterized by ABPP.21-23
Probe 1 was designed on the basis of the scaffold of
2-ethynylnaphthalene, a mechanism-based P450 inhibitor that
acts via P450-catalyzed oxidation of the aryl alkyne to a highly
reactive ketene, which subsequently acylates nucleophilic
residues within the P450 active site.24,25 Although 1 labeled
several P450 enzymes in mouse liver, it is unlikely that this
agent can serve as a universal ABPP probe for all mammalian
P450s, given the wide diversity of substrate selectivities
displayed by these enzymes. To expand the coverage of P450s
addressable by ABPP, we describe herein the synthesis and
functional characterization of a suite of activity-based probes
(ABPs) based on scaffolds of known mechanism-based inhibi-
tors of P450s (Figure 1) that operate by one or more of three
inhibition mechanisms (Figure 2).24,26-29 We show that these
probes have complementary reactivity profiles with a large set
(17) Yue, W.; Santen, R. J.; Wang, J.-P.; Hamilton, C. J.; Demers, L. M.
Endocr. Relat. Cancer 1999, 6, 157–64.
(18) Stark, K.; Guengerich, F. P. Drug Metab. ReV. 2007, 39, 627–37.
(19) Wright, A. T.; Cravatt, B. F. Chem. Biol. 2007, 14, 1043–51.
(20) Speers, A. E.; Cravatt, B. F. Chem. Biol. 2004, 11, 535–46.
(21) Cravatt, B. F.; Wright, A. T.; Kozarich, J. W. Annu. ReV. Biochem.
2008, 77, 383–414.
(14) McFadyen, M. C. E.; Cruickshank, M. E.; Miller, I. D.; McLeod, H. L.;
Melvin, W. T.; Haites, N. E.; Parkin, D.; Murray, G. I. Br. J. Cancer
2001, 85, 242–6.
(15) Rochat, B.; Morsman, J. M.; Murray, G. I.; Figg, W. D.; McLeod,
H. L. J. Pharmacol. Exp. Ther. 2001, 296, 537–41.
(16) Brueggemeier, R. W.; Hackett, J. C.; Diaz-Cruz, E. S. Endocr. ReV.
2005, 26, 331–45.
(22) Evans, M. J.; Cravatt, B. F. Chem. ReV. 2006, 106, 3279–3301.
(23) Jessani, N.; Cravatt, B. F. Curr. Opin. Chem. Biol. 2004, 8, 54–59.
(24) Kent, U. M.; Jushchyshyn, M. I.; Hollenberg, P. F. Curr. Drug Metab.
2001, 2, 215–43.
9
J. AM. CHEM. SOC. VOL. 131, NO. 30, 2009 10693

A R T I C L E S
Wright et al.
NADPH (1 mM). Samples were incubated at 37 °C for 15 min,
after which the strongest labeling probe for each P450 was added
(20 µM), and incubations proceeded at 37 °C for an additional 1 h.
Click chemistry and visualization were done as described above.
of human P450s and can be used to reveal unusual forms of
drug-drug interactions that affect P450 activity.
Experimental Section
Materials/Characterization. Unless otherwise noted, all chemi-
cals used for probe syntheses were purchased from Acros (Geel,
Belgium), Alfa Aesar (Ward Hill, MA), or Sigma-Aldrich (St.
Louis, MO). Tris(2-carboxyethyl) phosphine (TCEP) was purchased
from Fluka (St. Louis, MO). The click chemistry ligand, tris[(1-
benzyl-1H-1,2,3-triazol-4-yl)methyl]amine, was purchased from
Sigma-Aldrich. Aromatase inhibitor, anastrozole, was purchased
from Toronto Research Chemicals, Inc. (North York, Canada), and
formestane was purchased from Sigma-Aldrich. Probe 119 and
rhodamine-azide29 were prepared as described previously. Solvents
used for synthesis were obtained from a dry still system. NMR
spectra were obtained on a Varian-Innova 400 MHz instrument.
NMR chemical shifts are reported in ppm downfield relative to the
internal solvent peak, and J values are in Hz. High resolution-mass
spectrometry (HR-MS) was obtained at The Scripps Research
Institute Mass Spectrometry Core and recorded on an Agilent mass
spectrometer using ESI-TOF (electrospray ionization-time-of-flight).
Insect cell microsomes containing human P450 enzymes were
purchased from BD (San Jose, CA) as BD Supersomes. These are
recombinant cDNA-expressed P450s prepared from a baculovirus-
infected insect cell system. The insect cell microsomes contain the
necessary cytochrome P450 reductase. IC50 values were calculated
using Prism (GraphPad Software, Inc., La Jolla, CA) after obtaining
quantitative optical density data from SDS-PAGE gels. Quantitative
optical density data were collected from SDS-PAGE gel images
using ImageJ (NIH Freeware). Error values are calculated as
standard error.
Synthesis and Characterization of P450-Directed ABPs.
Details on the synthesis and analytical characterization of P450-
directed ABPs are provided in the Supporting Information.
Proteome Labeling and Analysis. Catalytic activity of P450s
is NADPH-dependent; therefore, control samples were prepared
without NADPH. BD Supersome microsomal proteomes (50 µL
of 1.0 mg/mL protein in PBS) were treated with a probe from the
suite (1-9) (20 µM; 0.5 µL of a stock solution in DMSO) in the
presence or absence of NADPH (1 mM; 0.5 µL of a stock solution
in PBS). Samples were incubated at 37 °C for 1 h. Following
incubation, probe-labeled proteomes were treated with rhodamine-
azide (100 µM; 6 mM stock solution in DMSO) followed by TCEP
(0.5 mM; 25 mM stock in water) and ligand (100 µM; 1.7 mM
stock in DMSO:t-butanol (1:4)). The samples were vortexed, and
cycloaddition was initiated by the addition of CuSO4 (1 mM; 50
mM stock in water). Samples were vortexed and left at room
temperature in the dark for 1 h at which time 2× SDS-PAGE
loading buffer (50 µL) was added. The samples were heated at 90
°C for 8 min, loaded onto SDS-PAGE gels (30 µL per well), and
visualized by in-gel fluorescent scanning using a Hitachi FMBio
Ile flatbed scanner (MiraiBio, Alameda, CA).
Results and Discussion
Probe Design and Synthesis. Click chemistry-compatible
ABPs are composed of three general components: (1) a reactive
group for targeting enzyme active sites, (2) a binding group for
directing the probe to a particular enzyme family, and (3) a
latent alkyne handle for click chemistry attachment of a reporter
group to visualize labeling events.20,21 Aryl alkynes satisfy the
first two criteria for many P450s, with the aromatic groups
providing positive binding interactions with hydrophobic P450
active sites and the aryl-conjugated alkyne promoting mecha-
nism-based enzyme inactivation via a highly reactive ketene
intermediate (Figure 2A). In addition to probe 1, we synthesized
four new aryl alkyne probes, 2, 3, 4, and 7. An aliphatic linker
group was added to each probe terminated by an alkyne for
use as a click chemistry handle to append reporter groups for
protein detection and/or enrichment. Through development of
a variant of 1 that possesses a linker group terminating in an
azide, we have determined that the aryl-alkyne functional unit
is likely the primary reactive moiety for inactivating P450s (see
Supporting Information Figure 1).
Probe 2 (Scheme 1A) was developed to provide a smaller
aryl-binding group as compared to 1. Beginning with com-
mercially available methyl-4-iodobenzoate (10), the aryl alkyne
moiety was installed via a Sonogashira coupling reaction to give
11 in moderate yield.30 The trimethylsilane protecting group
was removed and the methyl ester saponified in a single step
with NaOH giving the free acid 12.31 The aliphatic alkynyl
linker, 6-amino-1-hexyne, was coupled to the probe core using
1-hydroxybenzotriazole (HOBt) and 1-ethyl-3-(3′-dimethylami-
nopropyl)carbodiimide (EDCI) to give probe 2 in good overall
yield. We also synthesized an analogue of probe 1, probe 3,
where the aryl alkyne and linker groups were placed at the C1
and C2 positions on the naphthalene ring system, respectively.
Following our prior synthesis of 1, the methyl ester of 1-bromo-
2-naphthoic acid was synthesized with a coupling reaction giving
14 (Scheme 1B).32 As with all of our aryl alkyne probes, the
TMS-protected alkyne was installed onto the molecule with a
Sonogashira reaction (15)33 followed by one-pot deprotection
and saponification giving the free acid 16.33 The click chemistry
linker was installed with an HOBt/EDCI promoted coupling
reaction to 6-amino-1-hexyne providing probe 3. The biphenyl
probe 4 was synthesized (Scheme 1C) in the same manner as
3, starting with the esterification of 1734 to make 18,35
Profiling Inhibition of Probe 1 Labeling of P450 19A1
(Aromatase) by Anastrozole and Formestane. BD Supersome
P450 19A1 microsomal proteome (50 µL; 1 mg/mL protein in PBS)
was treated with an aromatase inhibitor (0-100 µM) and NADPH
(1 mM). Samples were incubated at 37 °C for 15 min, after which
1 (20 µM) was added, and incubations proceeded at 37 °C for an
additional 1 h. Click chemistry and visualization were done as
described above.
Global P450 Inhibition Profiles of Aromatase Inhibitors. BD
Supersome microsomal proteomes (50 µL of 1.0 mg/mL protein
in PBS) were treated with an aromatase inhibitor (2×, 10×, 20×,
and 40× the IC50 concentration for inhibition of P450 19A1) and
(27) Roberts, E. S.; Hopkins, N. E.; Foroozeh, M.; Alworth, W. L.; Halpert,
J. R.; Hollenberg, P. F. Drug Metab. Dispos. 1997, 10, 91–102.
(28) Ortiz de Montellano, P. R.; Mico, B. A. J. Biol. Chem. 1985, 260,
3330–6.
(29) Speers, A. E.; Adam, G. C.; Cravatt, B. F. J. Am. Chem. Soc. 2003,
125, 4686–7.
(30) Takahashi, S.; Kuroyama, Y.; Sonogashira, K.; Hagihara, N. Synthesis
1980, 8, 627–29.
(31) Yashima, E.; Matsushima, T.; Okamoto, Y. J. Am. Chem. Soc. 1997,
119, 6345–59.
(32) Masahiko, S.; Yamada, S.-I.; Kuroda, T.; Imashiro, R.; Shimizu, T.
Synthesis 2000, 12, 1677–80.
ˇ
(33) Stara´, I. G.; Stary´, I.; Kolla´rovicˇ, A.; Teply´, F.; Saman, D.; Fiedler,
P. Tetrahedron 1998, 54, 11209–34.
(25) Foroozesh, M.; Primrose, G.; Guo, Z.; Bell, L. C.; Alworth, W. L.;
Guengerich, F. P. Chem. Res. Toxicol. 1997, 10, 91–102.
(26) Hollenberg, P. F.; Kent, U. M.; Bumpus, N. N. Chem. Res. Toxicol.
2008, 21, 189–205.
(34) Simoni, D.; Giannini, G.; Roberti, M.; Rondanin, R.; Baruchello, R.;
Rossi, M.; Grisolia, G.; Invidiata, F. P.; Aiello, S.; Marino, S.;
Cavallini, S.; Siniscalchi, A.; et al. J. Med. Chem. 2005, 48, 4293–
99.
9
10694 J. AM. CHEM. SOC. VOL. 131, NO. 30, 2009

Activity-Based Probe Suite for Profiling Cytochrome P450s
A R T I C L E S
Scheme 1. P450 ABP Syntheses
9
J. AM. CHEM. SOC. VOL. 131, NO. 30, 2009 10695

A R T I C L E S
Wright et al.
Sonogashira coupling to obtain 19,34 base saponification and
deprotection of TMS to obtain the free acid 20, and EDCI/HOBt
coupling to the linker to obtain 4. Probe 4 delivers a biaryl ring
system with degrees of rotational freedom absent in the other
multi-ring aryl alkyne ABPs.
The final probe in the suite, 9, was synthesized from a
furanocoumarin core, 8-hydroxypsoralen. Furanocoumarins,
including 8-methoxypsoralen, are components of many foods
and inhibit the metabolism of several xenobiotics. 8-Methox-
ypsoralen is a potent NADPH-dependent mechanism-based
inactivator of human P450s 2A6, 2A13, and 2B1.44-46 The
mechanism of inactivation is ascribed to the covalent adduction
of the inhibitor to the apoprotein. To generate the furanocou-
marin reactive intermediate, a P450 oxidizes the furan alkene
to an epoxide, followed by subsequent nucleophilic attack to
form the dihydrofuranocoumarin product (Figure 2C). The
synthesis of probe 9 (Scheme 1H) commenced with nucleophilic
substitution of methylbromoacetate by 8-hydroxypsoralen (31)
yielding 32.47 The methyl ester was saponified with aqueous
LiOH to give 33,48 followed by HOBt/EDCI coupling of the
click chemistry linker yielding probe 9.
Probe Labeling of Human Cytochrome P450s. Our initial
studies with 1 had focused on profiling P450 targets in mouse
tissues.19 Although rodents are useful models for ascertaining
P450 function, their P450 repertoires are quite distinct from that
of humans; mice contain 102 putatively functional P450s as
compared to only 57 for humans.40,49,50 These differences create
complications in correlating function of P450s in humans and
rodents. The evaluation of human P450 activity, particularly
on new pharmaceutical entities, is therefore typically conducted
with human hepatocytes or recombinantly expressed human
P450s. Following this line of reasoning, we screened our suite
of ABPs against a panel of 14 human P450s coexpressed with
cytochrome P450 reductase in insect cell microsomes in the
presence or absence of NADPH. Following a 1 h incubation,
the proteomes were treated with rhodamine-azide under click
chemistry conditions and resolved by SDS-PAGE. Activity-
based labeling of P450 enzymes was defined as fluorescent
signals observed in the 48-55 kDa region of the gel in reactions
containing NADPH, but not in reactions lacking NADPH. No
labeling was observed in control insect proteomes lacking
recombinantly expressed human P450s (Supporting Information
Figure 2).
Two aliphatic ABPs, 5 and 6, were also synthesized. A prior
report demonstrated that undec-10-ynoic acid is a potent
NADPH-dependent inhibitor of P450s involved in ω-hydroxy-
lation and ω-1-hydroxylation of lauric acid.36 In these cases,
inhibition appeared to be primarily through inhibitor-heme
adduct formation. In addition to the mechanism of inhibition
via a reactive ketene intermediate (Figure 2A), a second
propynyl mechanism may also operate (Figure 2B). This
inhibition involves the P450-catalyzed dihydroxylation of the
carbon alpha to the alkyne group, followed by rearrangement
to a Michael acceptor.37,38 A nucleophilic amino acid residue
can add into the Michael acceptor leading to inhibition by
covalent adduction to the apo-P450. Probe 5 was synthesized
by a facile coupling of undec-10-ynoic acid (21) to the click
chemistry linker (Scheme 1D). It is possible that P450s could
react with either terminal alkyne on the probe molecule.
However, as will be shown later, significant P450 labeling
differences were observed between ABPs 5 and 6. Probe 6 was
synthesized with a conjugated terminal alkyne to determine the
effect on P450 labeling as compared to probe 5 (Scheme 1E).
The synthesis commenced with a Claisen-like condensation
yielding 23,39 followed by aqueous base saponification of the
ester groups to give 24,39 and copper promoted monodecar-
boxylation providing the free acid of the conjugated alkyne,
25.39 Probe 6 was obtained by coupling on the click chemistry
linker.
Probe 7 was developed from 7-ethynylcoumarin (7EC), a
known mechanism-based inhibitor of human P450 2B1.40
Compound 26 was converted to the 7-trifluoromethylsulfonate
coumarin 27, followed by a Pd-promoted coupling reaction to
give the TMS protected 7-ethynylcoumarin ester 28 (Scheme
1F). Consistent with the other probe syntheses, the methyl ester
was saponified and the TMS group cleaved with aqueous base
giving 29, followed by HOBt/EDCI coupling of the click
chemistry handle to give probe 7 in good yield.
The inactivation of human P450s 2B6, 3A4, and 3A5, and
rat P450 2B1 has been demonstrated by the oral contraceptive
17-R-ethynylestradiol (30, 17EE).41-43 Reactive intermediates
of 17EE were found to inactivate P450s in a NADPH-dependent
mechanism-based manner by a combination of heme alkylation
and apoprotein modification. We developed probe 8 as a variant
of 17EE by performing a bimolecular nucleophilic substitution
reaction with 30 and a tosylate variant of the click chemistry
linker to give probe 8 (Scheme 1G).
The panel of enzymes contained the following human P450s:
1A1, 1A2, 1B1, 2A6, 2B6, 2C9, 2C19, 2D6, 2E1, 2J2, 3A4,
4A11, 4F2, and 19A1 (aromatase). Eight of these enzymes are
directly involved in xenobiotic (pharmaceutical) metabolism,
1A1, 1A2, 2B6, 2C9, 2C19, 2D6, 2E1, and 3A4, being highly
expressed in liver and responsible for the catalytic conversion
of ∼75% of all pharmaceutical compounds51,52 (P450 3A4 alone
is involved in the catalytic metabolism of nearly 50% of known
drugs51). These drug-metabolizing P450s are remarkably flexible
and display broad substrate selectivity. Other enzymes analyzed
include: P450 1B1, an extrahepatic P450 often overexpressed
(35) Fletcher, D. I.; Ganellin, C. R.; Piergentili, A.; Dunn, P. M.; Jenkinson,
D. H. Bioorg. Med. Chem. 2007, 15, 5457–79.
(44) Koenigs, L. L.; Peter, R. M.; Thompson, S. J.; Rettie, A. E.; Trager,
W. F. Drug Metab. Dispos. 1997, 25, 1407–15.
(36) Ortiz de Montellano, P. R.; Reich, N. O. J. Biol. Chem. 1984, 259,
4136–41.
(45) Koenigs, L. L.; Trager, W. F. Biochemistry 1998, 37, 13184–93.
(46) von Weymarn, L. B.; Zhang, Q.-Y.; Ding, X.; Hollenberg, P. F.
Carcinogenesis 2005, 26, 621–9.
(37) Covey, D. F.; Hood, W. F.; Parikh, V. D. J. Biol. Chem. 1981, 256,
1076–79.
(47) Sharma, Y. N.; Zaman, A.; Kidwai, A. R. Tetrahedron 1964, 20, 87–
90.
(38) Metcalf, B. W.; Wright, C. L.; Burkhart, J. P.; Johnston, J. O. J. Am.
Chem. Soc. 1981, 103, 3221–2.
(48) Kim, K.-H.; Fan, X.-J.; Nielsen, P. E. Bioconjugate Chem. 2007, 18,
567–72.
(39) Jones, E. R. H.; Jones, J. B.; Skattebøl, L.; Whiting, M. C. J. Chem.
Soc. 1960, 3489–93.
(49) Guengerich, F. P. Chem. Biol. Interact. 1997, 106, 161–82.
(50) Nelson, D. R.; Zeldin, D. C.; Hoffman, S. M. G.; Maltais, L. J.; Wain,
H. M.; Nebert, D. W. Pharmacogenetics 2004, 14, 1–18.
(51) Williams, J. A.; Hyland, R.; Jones, B. C.; Smith, D. A.; Hurst, S.;
Goosen, T. C.; Peterkin, V.; Koup, J. R.; Ball, S. E. Drug Metab.
Dispos. 2004, 32, 1201–8.
(40) Regal, K. A.; Schrag, M. L.; Kent, U. M.; Wienkers, L. C.; Hollenberg,
P. F. Chem. Res. Toxicol. 2000, 13, 262–70.
(41) Guengerich, F. P. Mol. Pharmacol. 1988, 33, 500–08.
(42) Kent, U. M.; Mills, D. E.; Rajnarayanan, R. V.; Alworth, W. L.;
Hollenberg, P. F. J. Pharmacol. Exp. Ther. 2002, 300, 549–58.
(43) Lin, H.-L.; Kent, U. M.; Hollenberg, P. F. J. Pharmacol. Exp. Ther.
2002, 301, 160–7.
(52) Wienkers, L. C.; Heath, T. G. Nat. ReV. Drug DiscoVery 2005, 4,
825–833.
9
10696 J. AM. CHEM. SOC. VOL. 131, NO. 30, 2009

Activity-Based Probe Suite for Profiling Cytochrome P450s
A R T I C L E S
Figure 3. Probe suite labeling of four representative members of the 14-member P450 panel. Full SDS-PAGE separation was performed, but only the
45-55 kDa region is shown, where NADPH-dependent probe-labeled proteins were detected (consistent with the molecular masses of P450 enzymes). See
Supporting Information Figure 3 for labeling profiles of the entire 14-member P450 panel. Fluorescent gel shown in grayscale.
in cancer that is involved in estradiol and xenobiotic hydroxy-
lation; P450 2A6, a primarily hepatic P450 involved in
xenobiotic metabolism, most notably the mediation of nicotine’s
metabolic inactivation to cotinine;53 P450 2J2, an epoxygenase
predominantly expressed in liver and heart that metabolizes
arachidonic acid to biologically active eicosanoids; P450s 4A11
and 4F2, which are expressed primarily in liver and kidney and
catalyze fatty acid hydroxylation; and P450 19A1 (aromatase),
an extrahepatic P450 required for estrogen biosynthesis (includ-
ing conversion of testosterone to estradiol and androstenedione
to estrone) and the progression of estrogen receptor positive
[ER(+)] postmenopausal breast cancer.
The probe labeling profiles of recombinant insect cell-
expressed human P450s were diverse, with some enzymes
showing strong reactivity with many probes and others very
restricted reactivity with a select number of probes (Figure 3).
Intriguing structure-activity relationships were evident from
this library-versus-library analysis. For instance, the two
aliphatic alkyne probes (probes 5 and 6) showed very distinct
labeling profiles, with the former probe reacting strongly with
P450 1B1, but not P450 2D6, and the latter probe displaying
the opposite profile. Both probes contain the same click
chemistry linker chain; therefore, these reactivity differences
likely illustrate a preference for the long aliphatic chain of ABP
5 by P450 1B1, and a preference for the activated conjugated
alkyne of 6 by P450 2D6. Similarly profound reactivity
differences also occurred within the aryl alkyne subset of ABPs.
Indeed, although most P450 enzymes were labeled by multiple
probes, each enzyme showed a unique relative reactivity
signature (e.g., compare profiles of four P450s shown in Figure
3; complete labeling profiles for all P450s can be found in
Supporting Information Figure 3).
To more quantitatively assess differences in P450 reactivity
profiles, we measured the intensity of each P450-probe labeling
event by in-gel fluorescence scanning. To ensure that these
measurements reflected activity-based labeling of P450s, signals
were calculated as differences in fluorescent intensity of P450
bands in reactions with or without NADPH. Results of this
analysis are presented in Figure 4 for the probe labeling profiles
of the four P450s shown previously in Figure 3. A more
complete presentation of labeling profiles for the P450 panel is
provided as a heat-map in Figure 5, where the absolute (Figure
5A) and relative (Figure 5B) labeling signals are shown. The
relative signals were determined by normalizing data for each
P450 to the strongest labeling event for that enzyme. This
normalization facilitates identification of the optimal activity-
Figure 4. Extracted quantitative data from SDS-PAGE gels of probe
labeling, measured as fluorescence intensity (au, arbitrary units). The
reported fluorescent intensity values correspond to measurements of P450
bands observed in the gel lanes containing NADPH minus the equivalent
size regions of gel lanes lacking NADPH samples. These measurements
were restricted to well-defined boundaries around the fluorescence band
signals. Data represent the average ( standard error of three independent
experiments per P450 for each ABP. See Supporting Information Figure 4
for quantitative analysis of labeling profiles for the entire 14-member P450
panel.
based probe for each P450, even in cases where relatively low
overall signals were observed. Absent from this figure are the
labeling profiles of four P450s (2A6, 2B6, 2E1, 4A11) that failed
to show strong labeling with any members of the probe suite.
Evaluation of the quantitative labeling data revealed clear
differences in probe-P450 reactivity profiles, with some
enzymes showing strong preferential reactivity with a single
probe (e.g., 2D6 with 6; 2J2 with 1) and other enzymes
displaying broad reactivity with the probe library [e.g., 2C9,
3A4] (Figures 4 and 5). A brief synopsis of key findings is
provided below.
The P450 1A subfamily is critical to drug metabolism and
the oxidative catabolism of polycylic and heterocyclic hydro-
carbons, typically preferring planar substrates. As seen in the
normalized heat map, the aryl alkyne ABPs have the most
significant labeling of both P450 1A1 and 1A2. Like the P450
1A subfamily, P450 1B1 is able to activate a broad spectrum
of chemical carcinogens, including polycyclic hydrocarbons and
heterocyclic and aromatic amines. Interestingly, P450 1B1
showed a very broad probe labeling profiling, with every probe
reacting to an appreciable degree with this P450. This finding
supports the view that P450 1B1 is capable of activating many
different types of carcinogenic compounds and may serve as a
useful target for pro-drug activation,54 particularly for tumor
types that overexpress this P450.
(53) Mwenifumbo, J. C.; Tyndale, R. F. Pharmacogenomics 2007, 8, 1385–
402.
(54) McFadyen, M. C.; Murray, G. I. Future Oncol. 2005, 1, 259–263.
9
J. AM. CHEM. SOC. VOL. 131, NO. 30, 2009 10697

A R T I C L E S
Wright et al.
Figure 5. Heat maps illustrating probe labeling profiles for individual human P450 enzymes. (A) Absolute fluorescence signals of probe labeling events.
(B) Normalized fluorescence signals of probe labeling events, where data for each P450 enzyme are shown as a ratio of the strongest labeling signal for that
enzyme. A “1” is the strongest binding event for an individual P450. Negative values in both the raw data and the normalized data (reflecting less labeling
in the presence versus absence of NADPH) have been assigned “0”. Fluorescence intensity values are in arbitrary units and represent the mean of three
independent experiments per P450.
Figure 6. Blockade of probe 1 labeling of P450 19A1 (aromatase) by small-molecule inhibitors. (A) Type I (steroidal, formestane) and type II (nonsteroidal,
anastrozole) aromatase inhibitors. (B) Treatment of insect cell proteomes expressing P450 19A1 with increasing concentrations of aromatase inhibitors
(0-100 µM) resulted in concentration-dependent blockade of 1 labeling. The IC50 curves for inhibition of 1 labeling of P450 19A1 by both inhibitors are
shown. Data represent the average ( standard error of three independent experiments per inhibitor. Representative gel image of competitive blockade of
probe labeling shown for anastrozole.
Cytochrome P450 3A4 is the most abundant of all P450s in
the human body. It has been proposed, and crystal structures
and in silico studies have found, that P450 3A4 has an open
active site and that reactions are influenced largely by the
chemical lability of C-H bonds.55 Additionally, P450 3A4 is
known to exhibit non-Michaelis-Menten steady-state kinetics
for many substrates, indicating multiple substrates binding
simultaneously.56 Evaluating our raw data shows that P450 3A4
reacts with most all members of the probe suite, supporting the
known flexibility of this enzyme’s active site.
Cytochrome P450s involved in fatty acid metabolism, 2J2
and 4F2, were labeled by both aryl and aliphatic alkyne ABPs.
Interestingly, and somewhat counterintuitively, P450 2J2, an
epoxygenase that oxidizes arachidonic acid, was most strongly
labeled by 1. In contrast, P450 4F2 was most strongly labeled
by the fatty acid-derived probe 5. Yet, 1 labeled P450 4F2 with
the second greatest intensity. In both cases, the aliphatic
conjugated alkyne probe, 6, displayed minimal probe labeling,
likely reflecting a preference for interacting with extended,
hydrophobic small molecules. Finally, P450 19A1, which is
responsible for estrogen biosynthesis, was labeled moderately
by 1 and the two aliphatic probes. The probe labeling profiles
for P450s 2J2, 4A11, 4F2, and 19A1 were surprising. These
are enzymes with rather specific biological substrates, typically
regulating steroid or fatty acid metabolism. However, they all
react with ABPs of varying chemical composition, quite distinct
from their natural substrates. This highlights that the structural
composition of endogenous substrates and non-natural inhibitors/
probes can differ considerably for P450s.
A handful of P450s (2A6, 2E1, 4A11) failed to show
significant labeling with any of the ABPs. The most character-
istic and specific reaction of P450 2A6 is coumarin 7-hydroxy-
lation. Additionally, 8-methoxypsoralen is a potent inhibitor of
human P450 2A6.45 In the design of 9, we substituted the
8-methoxy group with an ester bond to the click chemistry
linker, which may have hindered probe entrance into the active
site of P450 2A6. Although little is known about the P450 2E1
active site, several low molecular weight solvents are potent
inhibitors. The only reported alkyne-bearing mechanism-based
inhibitor is tert-butyl acetylene (for rabbit P450 2E1).57 Other
mechanism-based inhibitors include disulfiram and its primary
(57) Blobaum, A. L.; Kent, U. M.; Alworth, W. L.; Hollenberg, P. F. Chem.
Res. Toxicol. 2002, 15, 1561–71.
(55) Smith, D. A.; Jones, B. C. Biochem. Pharmacol. 1992, 44, 2089–98.
(56) Kapelyukh, Y.; Paine, M. J.; Marechal, J. D.; Sutcliffe, M. J.; Wolf,
C. R.; Roberts, G. Drug Metab. Dispos. 2008, 36, 2136–44.
(58) Kharasch, E. D.; Hankins, D. C.; Jubert, C.; Thummel, K. E.; Taraday,
J. K. Drug Metab. Dispos. 1999, 27, 717–23.
9
10698 J. AM. CHEM. SOC. VOL. 131, NO. 30, 2009

Activity-Based Probe Suite for Profiling Cytochrome P450s
A R T I C L E S
Figure 7. Effects of aromatase inhibitors on P450 labeling. (A) Treatment of human P450s with formestane (2×, 10×, 20×, and 40× the IC50 of aromatase
inhibition), followed by probe (20 µM), resulted in inhibition (2C19, 3A4) or unchanged (1A2) P450 labeling. (B) Treatment of human P450s with anastrozole
(2×, 10×, 20×, and 40× the IC50 of aromatase inhibition) resulted in increased (1A2), decreased (2C9), or unchanged (4F2) P450 activity. Quantitative
values from the gels are normalized to the intensity value without inhibitor treatment. Data represent the average ( standard error of three independent
experiments per inhibitor.
metabolite diethyldithiocarbamate.58 P450 4A11, which was
weakly labeled by 7, is involved in ω-hydroxylation of lauric,
myristic, and palmitic acids. Surprisingly, neither of the aliphatic
ABPs, 5 or 6, labeled P450 4A11. It is possible that, in such
instances where probes would be predicted to bind to P450s,
labeling occurs predominantly on the heme molecule rather than
on the protein, which would preclude detection by ABPP.
Alternatively, we cannot exclude the possibility that these
enzymatic preparations contained low levels of expression and/
or activity of the corresponding P450s.
have therefore moved to a position of frontline therapy for ER+
breast cancer.59 Aromatase inhibitors are separated into two
classes, types I and II (Figure 6A). Type I inhibitors (formestane)
are irreversible steroidal mechanism-based inhibitors; they are
androgen substrate analogues that bind competitively but
irreversibly to the enzyme active site. Type II inhibitors
(anastrozole) are nonsteroidal triazoles; they bind reversibly to
the enzyme and orient within the aromatase active site such
that the azole nitrogens interact with the Fe of the prosthetic
heme group.60
Collectively, these data indicate that individual P450 enzymes
exhibit distinct probe-labeling signatures. These signatures
should facilitate the deduction of P450 identities from cell or
tissue proteome, especially in instances where insufficient
proteomic material is available for mass spectrometry analysis.
Preliminary studies where we have tested the concentration
dependence of P450 labeling suggest that labeling signatures
can be further refined by adjusting the concentrations of
individual probes. For instance, while P450 2B6 demonstrated
strong labeling with multiple probes tested at 20 µM, this
enzyme exhibited selective reactivity with probe 2 tested at 2
µM (see Supporting Information Figure 5).
Evaluation of Aromatase Inhibitors. As demonstrated in our
previous report,19 ABPs can be used to measure alterations in
P450 activity due to chemical compounds that act as inhibitors
and inducers. For many years, estrogen-receptor positive (ER+)
postmenopausal breast cancer was treated with tamoxifen, a
selective ER modulator. However, tamoxifen is an ER agonist
in some tissues, which has led to increased incidence of other
cancer types, such as endometrial cancer.59 Additionally,
resistance to tamoxifen therapy is a problem for many patients.9
Aromatase inhibitors, which block the biosynthesis of estrogen,
Our quantitative profiling experiments identified 1 as the most
potent probe for P450 19A1 (Figure 5). We therefore asked
whether probe 1 could report on the binding of aromatase
inhibitors to P450 19A1. Type I (formestane) and type II
(anastrozole) aromatase inhibitors were independently added to
insect cell expressed human P450 19A1, followed by addition
of 1, and subsequent click chemistry to rhodamine-azide (Figure
6B). We determined an IC50 ) 0.43 µM and 1.0 µM for
formestane and anastrozole, respectively (Figure 6B). These
values are in line with those reported previously in the
literature.61 These data demonstrate that 1 can monitor the
effects of inhibitory drugs on aromatase activity. We next
considered whether we could identify the consequences of
aromatase inhibitors on probe labeling of the other P450s in
our panel.
Profiling P450 Inhibition by Aromatase (P450 19A1) Inhibi-
tors. A major concern for all pharmaceuticals is their off-target
effects. Aromatase inhibitors, both types I and II, are considered
highly selective. There are no published reports of off-target
(60) Miller, W. R. Semin. Oncol. 2003, 30, 3–11.
(61) Njar, V. C. O.; Grun, G.; Hartmann, R. W. J. Enzyme Inhib. Med.
Chem. 1995, 9, 195–202.
(59) Fornander, T.; Hellstrom, A. C.; Moberger, B. J. Natl. Cancer Inst.
1993, 85, 1850–5.
(62) Grimm, S. W.; Dyroff, M. C. Drug Metab. Dispos. 1997, 25, 598–
602.
9
J. AM. CHEM. SOC. VOL. 131, NO. 30, 2009 10699

A R T I C L E S
Wright et al.
changes to P450 activity due to formestane; however, inhibition
of P450s 1A2, 2C9, and 3A4 has been reported for anastrozole
at concentrations of 8, 10, and 10 µM, respectively.62 These
inhibitory effects were determined by coincubating human liver
microsomes with anastrozole and P450-specific substrates. We
treated each of the 14 human P450s in our panel with inhibitors
at approximately 2×, 10×, 20×, and 40× the IC50 concentration
of aromatase inhibition for the respective inhibitor, followed
by probe labeling with the specific ABP that most strongly
labeled each P450 at 20 µM (Figure 5), and visualization by
click chemistry to rhodamine-azide.
were all synthesized in “clickable” form, empowering them
with the potential to profile P450 enzymes in vitro and in
vivo.19 Screening of the probes against a panel of human
P450 enzymes identified optimal probes for individual P450s.
Promiscuous probes that target many P450s (e.g., probes 1-3
and 5) could be readily distinguished from more selective
probes that showed preferred labeling with individual P450s
(e.g., probes 6-8). Conversely, some P450s (e.g., 2C9, 2D6,
and 3A4) exhibited broad reactivity with the suite of ABPs,
likely reflecting malleable active site structures that can
accommodate a range of small molecules. Other P450s,
however, showed highly restricted or, in certain cases, no
reactivity with members of our probe set. We further
demonstrated that ABPP probes can detect not only inhibition,
but also stimulation of P450 activities by small molecules,
uncovering a previously unappreciated activation of P450
1A2 by the anticancer drug anastrozole. Considering the
central role that P450s play in metabolizing drugs, xenobi-
otics, and endogenous bioactive molecules, we anticipate that
the ABPs suite described herein should prove of value for
globally assessing P450 function in a wide range of biological
systems. For instance, the probes could be used to profile
P450s in human liver biopsies isolated from different disease
conditions, as well as to characterize these enzymes and their
drug interactions in living cells and animals. On this note,
we have previously shown that probe 1 is capable of profiling
P450 activities and drug interactions in living mice.19 Future
work may also include the optimization of probe selectivity
for individual P450s, such that these enzymes can be
independently monitored in primary human specimens.
Each aromatase inhibitor affected the probe labeling of two
P450s (in addition to their effects on P450 19A1). Formestane
decreased probe 5 labeling of P450 2C19 and 1 labeling of
P450 3A4 (Figure 7A), while anastrozole inhibited labeling
of P450 2C9 by probe 5 and, interestingly, increased labeling
of P450 1A2 by probe 3 up to 175% of control values (Figure
7B). This last finding indicates that anastrozole and probe 3
may show heterotypic cooperativity with P450 1A2. Heterotypic
cooperativity, where the metabolism of one substrate is increased
in the presence of a second compound, is known to occur with
many P450 isoforms, and this form of drug-drug interaction
can have a major impact on the rate of in vivo clearance for
specific small molecules.63 Cooperativity of human P450 1A2
has previously been observed with polycyclic hydrocarbons.64
The heterotropic cooperative effects associated with human P450
1A2 are ascribed to a more productive rearrangement of the
substrate for P450 catalysis because of either steric packing or
more specific P450-ligand interactions due to the presence of
a second compound.64 To our knowledge, these data provide
the first evidence of heterotypic cooperativity for anastrozole
and P450 1A2 (or any other P450) and suggest that this
aromatase inhibitor could produce unanticipated drug-drug
interactions due to activation of P450 1A2. Because previous
studies have also reported an inhibitory effect of anastrozole
on P450 1A2, we speculate that the (positive or negative) impact
of this aromatase inhibitor on P450 1A2 activity will depend
on the specific nature of the test substrate. More generally, our
results underscore the versatility of ABPP for characterizing
both increases and decreases in P450 activity that occur in
response to drug treatment.
Acknowledgment. We thank the Cravatt laboratory for helpful
discussions and critical reading of the manuscript. This work was
supported by the National Institutes of Health (CA087660), the
California Breast Cancer Research Program (A.T.W.), and the
Skaggs Institute for Chemical Biology.
Supporting Information Available: Probe syntheses and
characterization. Gel analysis of an azido-variant of 1, probe
labeling gel profiles for the entire P450 panel; gel analysis of
probe labeling of insect cell control proteomes, charts showing
quantitative intensity of P450 probe labeling; and concentration
dependence of P450 probe labeling. This material is available
free of charge via the Internet at http://pubs.acs.org.
Conclusions
Here, we have synthesized and tested a new set of ABPs
directed toward the P450 family of enzymes. These probes
(63) Tracy, T. S. Curr. Drug Metab. 2003, 4, 341–6.
(64) Sohl, C. D.; Isin, E. M.; Eoff, R. L.; Marsch, G. A.; Stec, D. F.;
Guengerich, F. P. J. Biol. Chem. 2008, 283, 7293–7308.
JA9037609
9
10700 J. AM. CHEM. SOC. VOL. 131, NO. 30, 2009