Chemoproteomics-Enabled De Novo Discovery of Photoswitchable Carboxylesterase Inhibitors for Optically Controlled Drug Metabolism

Article abstract of DOI:10.1002/anie.202011163

Herein, we report arylazopyrazole ureas and sulfones as a novel class of photoswitchable serine hydrolase inhibitors and present a chemoproteomic platform for rapid discovery of optically controlled serine hydrolase targets in complex proteomes. Specifica

Full text of DOI:10.1002/anie.202011163

A Journal of the Gesellschaft Deutscher Chemiker
Angewandte
Chemie
International Edition
www.angewandte.org
Accepted Article
Title: Chemoproteomics-Enabled De Novo Discovery of
Photoswitchable Carboxylesterase Inhibitors for Optically
Controlled Drug Metabolism
Authors: Brendan Dwyer, Chao Wang, Daniel Abegg, Brittney
Racioppo, Nan Qiu, Zhensheng Zhao, Dany Pechalrieu,
Anton Shuster, Dominic Hoch, and Alexander Adibekian
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Angew. Chem. Int. Ed. 10.1002/anie.202011163
Link to VoR: https://doi.org/10.1002/anie.202011163

10.1002/anie.202011163
Angewandte Chemie International Edition
RESEARCH ARTICLE
Chemoproteomics-Enabled De Novo Discovery of
Photoswitchable Carboxylesterase Inhibitors for Optically
Controlled Drug Metabolism
Brendan G. Dwyer,+[a] Chao Wang,+[a, b] Daniel Abegg,[a] Brittney Racioppo,[a] Nan Qiu,[a] Zhensheng
Zhao,[a] Dany Pechalrieu,[a] Anton Shuster,[a] Dominic G. Hoch,[a, c] and Alexander Adibekian*[a]
2]
release.[1a,
More recently, however, small molecule
Abstract: Herein, we report arylazopyrazole ureas and sulfones as a
novel class of photoswitchable serine hydrolase inhibitors and present
a chemoproteomic platform for rapid discovery of optically controlled
serine hydrolase targets in complex proteomes. Specifically, we
identify highly potent and selective photoswitchable inhibitors of the
photoswitches that enable repeated isomerization between
the active and inactive state have emerged as a powerful tool
to optically and reversibly control receptors, hydrolases[3] and
other enzymes,[4] and even the cytoskeleton.[5] Most
commonly, photoswitchable probes are developed through
“azologization”, replacement of a moiety such as a stilbene,
benzyl aniline, or N-aryl benazamide with a structurally similar
azobenzene group.[1a, 1b, 6] An exemplifying application of this
strategy is the recent development of photostatins based on
the natural product combretastain A-4 to modulate
microtubule dynamics.[5e] While azologization has proved
tremendously successful over the past decade, this strategy
is limited to established small molecules with known mode of
action.
drug-metabolizing enzymes carboxylesterases
1 and 2 and
demonstrate the pharmacological application of our compounds by
optically controlling the metabolism of the immunosuppressant drug
mycophenolate mofetil. Collectively, this proof-of-concept study
provides a first example of photopharmacological tools to optically
control drug metabolism by modulating the activity of a metabolizing
enzyme. Our arylazopyrazole ureas and sulfones offer synthetically
accessible scaffolds that can be expanded to a larger library with
hundreds of compounds to identify specific photoswitchable inhibitors
for other serine hydrolases, including lipases, peptidases, and
proteases and our chemoproteomic platform can be expanded to
other photoswitches and scaffolds to achieve optical control over
diverse protein classes.
Despite the advances in photopharmacology in the past
few decades, de novo discovery of photoswitchable protein
modulators remains a valuable yet underexplored area.
Rather than replacing structural mimics for azologization,
incorporating a photoswitch into the original design of small
molecule ligands would rapidly expand the biochemical
application repertoire of photoswitchable probes and
ultimately facilitate the development of optically-controlled
therapeutics. In this proof-of-concept study, we report
arylazopyrazole ureas and sulfones as a novel class of
photoswitchable serine hydrolase inhibitors. Further, we
Introduction
Photopharmacology is a promising therapeutic strategy
that utilizes light to achieve fast spatiotemporal control over
the biological activities of small molecules.[1] The most
commonly used photopharmacological tools to date are the
“caged” bioactive ligands where light activates their
present
a chemoproteomic platform that allows rapid
discovery of optically controlled serine hydrolase targets in
complex cellular lysates. Rapid testing of a small collection of
arylpyrazole ureas identifies highly potent and selective,
photoswitchable inhibitors of human carboxylesterases 1
(CES1) and 2 (CES2), enzymes responsible for ester-
containing drug metabolism in human body and promising
clinical targets for drug metabolism modulation.[7] We
demonstrate the utility of our photoswitchable CES inhibitors
by optically controlling the CES1-mediated hydrolysis of the
immunosuppressant drug mycophenolate mofetil in cellular
lysates. Our de novo discovered photoswitchable CES1 and
CES2 inhibitors are the first reported photopharmacological
tools to optically control drug metabolism by modulating the
activity of a metabolizing enzyme and represent valuable
pharmacological tools to understand CES1/2-mediated drug
and xenobiotic metabolism and beyond.
[a]
B. G. Dwyer, Dr. C. Wang, Dr. D. Abegg, N. Qiu, B. Racioppo, Dr. Z
Zhao, Dr. D. Pechalrieu, A. Shuster, Dr. D. G. Hoch, Prof. Dr. A.
Adibekian
Department of Chemistry, The Scripps Research Institute
130 Scripps Way, Jupiter, FL, 33458 (United States)
E-mail: aadibeki@scripps.edu
[b]
Dr. C. Wang (current address)
Department of Molecular Medicine, The Scripps Research Institute
10550 North Torrey Pines Road, La Jolla, CA, 92037 (United
States)
[c]
[+]
Dr. D. G. Hoch (current address)
Laboratory of Organic Chemistry, ETH Zürich
8093 Zürich (Switzerland)
These authors contributed to this work equally.
Supporting information for this article (synthesis of compounds,
cell culture, bioassays, and sample preparation for proteomics) is
available on the WWW under http://
This article is protected by copyright. All rights reserved.

10.1002/anie.202011163
Angewandte Chemie International Edition
RESEARCH ARTICLE
expectedly long[9a] half-lives (t1/2) determined for 4-9 (Figures
2C, 2D, S2, and S3).
Results and Discussion
Inspired by our previous work on triazole ureas as active
site serine-directed, covalent serine hydrolase inhibitors,[8] we
sought to design photoswitchable serine hydrolase inhibitors
by replacing the triazole leaving group with the
arylasopyrazole photoswitch. We selected arylazopyrazole as
the photoswitchable moiety because of its excellent
photochemical properties such as quantitative switching and
high thermal stability of the Z isomer (Figure 1).[9] A collection
of six arylazopyrazole urea inhibitors with varying head groups
was synthesized in three synthetic steps with excellent yields
(Scheme 1). First, aniline 1 was converted into a diazonium
salt using sodium nitrate and hydrochloric acid, which was
then added to a suspension of 2,4-pentanedione and sodium
acetate to afford the hydrazone 2 by enolate addition. The
arylazopyrazole 3 was then obtained upon refluxing with
hydrazine monohydrate in ethanol. The arylazopyrazole ureas
4-9 were then generated by reacting the corresponding
carbamoyl chlorides with arylazopyrazole 3 under basic
conditions.
We were able to obtain and analyze the crystal
structures of E- and Z-4. As expected, arylazopyrazole urea 4
undergoes a pronounced conformational switch upon E-Z
conversion, changing the dihedral angles from 6.1° to 34.6°
(ϕNNhet) and 25.6° to 41.5° (ϕpheNN, Figure 2E). We speculated
at this point that such a significant difference in conformation
of our E- and Z-ureas may lead to sufficient disparity between
the binding affinities of both isomers towards a given target
serine hydrolase.
Scheme 1. Synthesis of arylazopyrazole urea inhibitors; a) 12 M HCl,
NaNO2 (1.2 eq.), AcOH/H2O, 0 °C, 1 h then acetylacetone (1.17 eq.), NaOAc
(3.0 eq.), EtOH/H2O, 0 °C to r.t., 1 h, 95 %; b) hydrazine monohydrate (1.0 eq),
EtOH, reflux, 3 h, quantitative yield; c) corresponding carbamoyl chloride, NaH
(2.0 eq.), THF, r.t., 1 h. See Supporting Information for synthetic methods and
characterization.
We next proceeded with the identification of potential
serine hydrolase targets of the E- and Z- isomers our ureas 4-
9 in complex proteomes using a gel-based competitive
proteomic profiling approach.[10] Here we sought to identify
differences in the gel band pattern of lysates pretreated with
the E-isomers of our compounds as competitiors, treated with
light at 365 nm or not, and labeled with a serine hydrolase-
specific
activity-based
probe.
Human colorectal adenocarcinoma Caco-2 cell lysate was
first treated with 100 nM of the indicated E-isomer, with or
without concomitant UV irradiation for 5 minutes, for one hour
at 37 °C and then with our recently published serine
hydrolase-directed ABPP probe 10 (Figure 3A) for one hour at
r.t.[11] A carboxytetramethylrhodamine (TMR) fluorescent dye
was then installed on 10 using copper(I)-catalyzed azide-
alkyne cycloaddition (CuAAC), the probe-labeled proteins
were separated by SDS-PAGE, and the resulting fluorescent
bands were visualized using in-gel fluorescence scanning.
Indeed, we observed disappearence of a single band at
approximately 60 kDa with the E-isomers of 4-6 showing the
greatest competition (Figure S4). Further evaluation using 30
nM inhibitors showed that E-4 and E-5 completely competed
this band (Figure 3A), while the UV irradiated E-isomers
showed no significant competition. Intriguingly, no other
serine hydrolase band on the gel appeared to be competed by
our compounds.
Figure 1. Our concept of photoswitchable arylazopyrazole urea serine
hydrolase inhibitors inspired by previous work. PSS = photostationary state.
CES1 structure was derived from PDB: 5A7G.
We next proceeded to determine the photochemical
properties of the obtained ureas, and compare them with the
original arylazopyrazole photoswitch.[9a] First, we used UV-Vis
spectroscopy to determine the optimal E-Z isomerization
wavelengths. Due to the π-acceptor properties of the
pyrazole-attached carbonyl group, our urea compounds 4-9
were slightly blue-shifted in their π-π* λmax values (308-312
nm for E-isomer) compared to the original N-methyl
arylazopyrazole (335 nm; Figures 2A and S1).[9a] We
observed a significant difference in absorbance between the
E and Z isomers at 365 nm, assuring high yielding E-Z
conversion at this wavelength. Indeed, urea 4 exhibited
excellent conversion after just five minutes of UV irradiation
(365 nm) at room temperature with a photostationary state
(PSS) of 95% Z isomer (Figure 2B). Likewise, all other
arylazopyrazole ureas exhibited greater than 94% E-Z
conversion after UV irradiation (Figure S2). Further, the
thermal relaxation of Z arylazopyrazole urea conversion to E
was measured over time at 37 °C via NMR spectroscopy with
To identify the 60 kDa serine hydrolase, we performed a
competitive mass spectrometry-based experiment in Caco-2
lysate again using probe 10.[11] The lysate was treated with E-
4 or DMSO for one hour at 37 °C and then with 10 µM probe
10 followed by streptavidin enrichment, digestion, and MS-
based analysis. Our data analysis showed complete
competition of two targets, carboxylesterases 1 (CES1) and 2
(CES2), both ~60 kDa enzymes, with ratios of 0.04 and 0
respectively. No other serine hydrolase was competed in this
experiment, thus confirming the excellent specificity of E-4
(Figure S5; Table S1).
This article is protected by copyright. All rights reserved.

10.1002/anie.202011163
Angewandte Chemie International Edition
RESEARCH ARTICLE
Figure 2. Photoisomerization properties of the arylazopyrazole urea inhibitors. A) UV-Vis spectrum of E-4 and Z-4 in DMSO. B) 1H-NMR spectra of E-4 (top), 95%
Z-4 after UV irradiation (5 min) at 365 nm of E-4 (middle), and 100% Z-4 after purification via silica column (bottom). C) 1H NMR spectra of Z-4 converting back to
E-4 over time at 37 °C and (D) quantification of the conversion. E) X-ray crystallographic structures of E-4 and Z-4 with annotated dihedral angles.
experiment, live Hep G2 cells were treated with 14 for 4 hours
followed by cell lysis, CuAAC with a TMR dye, and in-gel
fluorescence visualization (Figure S7). Collectively, these
experiments clearly demonstrate the outstanding and
unprecedented selectivity of our arylazopyrazole ureas
towards CES1 and CES2. Indeed, in contrast to previously
discovered CES inhibitors with limited or unreported
selectivity, our arylazopyrazole urea compounds selectively
inhibit CES1/2 in the presence of other serine hydrolases.[7a,
Carboxylesterases are serine esterases from the
alpha/beta fold hydrolase family. Because they participate in
phase I metabolism by hydrolyzing ester- and amide bonds of
xenobiotics, carboxylesterases can activate or deactivate
various
therapeutics
and
drastically
affect
the
pharmacokinetics and the pharmacodynamics of the
metabolized drugs. Consequently, CES enzymes have been
predominantly explored as drug targets for controlling the
metabolism of ester-containing drugs to optimize their
12]
14]
pharmacological properties.[7a,
Since the majority of
.
carboxylesterases as well as other drug hydrolyzing serine
hydrolases are expressed in the liver,[13] we performed
another competitive MS experiment in the liver cancer cell line
Hep G2, this time with both E-4 or E-5. Once again, E-4 only
competed CES1 (ratio 0.35) and CES2 (ratio 0.06), while E-5
competed CES1 (ratio 0.25) specifically (Figure 3B; Table S2),
while none of the other 36 serine hydrolases detected in this
MS experiment were significantly competed.
Competitive ABPP with the fluorophosphonate probe
can only reveal serine hydrolases as potential targets.
Therefore, to exclude the possibility that our ureas 4 and 5
may also target other classes of proteins, we synthesized the
alkyne probe 14 derived from inhibitor 5, by replacing an allyl
arm with a ‘clickable’ alkyne handle (Scheme 2). Allylamine 11
was treated with propargyl bromide to afford the secondary
amine 12, which was then activated using carbonyldiimidazole
under basic conditions to yield the imidazole urea 13. Finally,
coupling with arylazopyrazole 3 yielded probe 14.
Figure 3. Competitive ABPP reveals unique optically controlled serine
hydrolases. A) Gel-based competitive profiling using the fluorophosphonate
probe 10 (10 µM; 1 hour at r.t.) in Caco-2 lysate pretreated with the E-isomers
or irradiated E-isomers of 6, 5, or 4 (30 nM) for 1 hour at 37 °C. CBB =
Coomassie Brilliant Blue staining. B) Competitive MS-based identification of 30
nM E-4 (top) and E-5 (bottom) protein targets using the fluorophosphonate 10
(10 µM) in Hep G2 lysate (n = 6). Error bars are SEM.
We then performed a gel-based labeling experiment with
lysates from 13 cell lines of different tissue origins and thus
with different protein expression profiles. Remarkably,
treatment with 100 nM probe 14 only labeled two bands in Hep
G2 and one in Caco-2 lysates (Figure 4A), and no band was
labeled in any of the remaining cell lines. These proteins were
again confirmed to be CES1 and CES2 by enrichment with 14
from Hep G2 lysate and identification by MS (Figure S6; Table
S3). As further validation, probe 14 labeled overexpressed
CES1 and CES2 in HEK 293T cell lysates (Figure 4B) and
native CES1 and CES2 in live Hep G2 cells. For the latter
Due to the high specificity and covalent engagement of CES,
probe 14 can also be used for fluorescence imaging of CES
enzymes (Figure 4C). Caco-2 cells were first subjected to
CES1 siRNA knockdown or a non-targeting control for 48
hours before re-seeding the cells on coverslips and treating
them with 14 (100 nM) or DMSO for 4 hours followed by
methanol fixation, conjugation to a TMR dye via a CuAAC
reaction, and fluorescence microscopy imaging. We observed
This article is protected by copyright. All rights reserved.

10.1002/anie.202011163
Angewandte Chemie International Edition
RESEARCH ARTICLE
for 4 hours. Cells were lysed, treated with 14, conjugated to
TMR azide via a CuAAC reaction, and visualized by in-gel
fluorescence scanning. Indeed, also the in situ
treatment/isomerization displayed a 5.1-fold difference in IC50
values of E-4 and irradiated E-4 (5.0 nM and 25.4 nM
respectively; Figure S8). The difference in IC50 values
demonstrated by our inhibitors is at least similar, if not better,
than the difference reported for other bioactive
photoswitchable inhibitors.[1c, 18]
an ER-like staining pattern consistent with previous reports of
15]
CES localization.[13,
As expected, an incomplete CES1
siRNA knockdown proportionally abolished the fluorescence
signal from treatment with 14. A similar degree of decrease in
labeling was observed by fluorescence gel in lysates of CES1
knockdown cells treated with 14 (Figures 4C and 4D). Taken
together, these results confirm that 14 can be used as a
specific imaging reagent for native CES1.
Scheme 2. Synthesis of the activity-based probe 14; a) allylamine (6.0 eq.),
propargyl bromide (1.0 eq., 80 % w/v in toluene), 0 °C to r.t., overnight, 32 %;
b) carbonyldiimidazole (3.0 eq.), NaH (1.2 eq), THF, 0 °C to r.t., 12 h, 93 %; c)
3 (1.0 eq.), DBU (1.5 eq.), dry MeCN, r.t., 24 h, 68 %.
Figure 5. Optical modulation of CES1 inhibition. A) Competitive gel-based
assay comparing the binding of E-4 and irradiated E-4 to endogenous CES1 in
Hep G2 lysate using 14 (300 nM) as probe and (B) assay quantification. C)
Competitive gel-based assay comparing the binding of E-5 and irradiated E-5
and (D) assay quantification. Quantifications are shown as normalized band
intensity ± SD (n = 3).
Figure 4. Evaluation of the CES1/2-selectve activity-based probe 14. A) Gel-
based profiling of 14 (100 nM; 1 hour at 37 °C) in various cell lysates (2 mg/mL).
B) Fluorescence SDS-PAGE and corresponding Western blot images of mock
and overexpressed CES1-V5 and FLAG-CES2 labeled with 14 (300 nM). C)
Microscopy images of Caco-2 cells treated with 14 (3 biological replicates with
>57 cells counted) and CES1 siRNA and fluorescence SDS-PAGE documenting
the siRNA-mediated CES1 knockdown, and (D) quantification of the microscopy
and SDS-PAGE results. Scale bar represents 30 µm. Error bars are SEM. ****
indicates p < 0.0001; ns indicates p > 0.05.
Scheme 3. Synthesis of arylazopyrazole sulfone 16; a) NaH (1.5 eq.) in dry THF
at 0 °C, 3 (1.0 eq.), and then 15 (2.5 eq.), r.t., 2 h, 85 %.
Having reversible photoswitchable CES inhibitors would
allow repeated cycling between active and inactive states and
that may be desirable in some basic research applications
such as pharmacokinetic parameter calculations. We
hypothesized that replacing the carbonyl group in our covalent
inhibitor 4 with a structurally similar yet less electrophilic
transition state mimic such as a sulfonyl group may produce a
non-covalent CES inhibitor. To synthesize the sulfone
analogue of inhibitor 4, the sulfamoyl chloride 15 was coupled
to the arylazopyrazole 3 under basic conditions yielding the
sulfone-based arylazopyrazole 16 (Scheme 3). UV-Vis
spectroscopy of 16 showed a slight red-shift of the λmax to 329
nm compared to the arylazopyrazole ureas (Figure 6A).[9a] The
arylpyrazole sulfone 16 could be switched between E and Z
isomers using 365 nm and 525 nm light respectively. 1H-NMR
analysis showed PSS conversion of 87% from E- to Z-isoform
(Figure S9). Further, the arylazopyrazole sulfone 16 had
excellent thermal stability with a t1/2 of 105 hours via 1H-NMR
analysis measuring conversion of Z- to E-16 over time at 37
°C (Figures 6B and 6C). Finally, 16 also demonstrated great
photostability as its isomerization could be cycled at least 8
CES1 is the best studied xenobiotic ester hydrolytic
enzyme in the human carboxylesterase family.[16] For
example, CES1 is responsible for the activation of many
prodrugs such as angiotensin-converting enzyme (ACE)
inhibitors, oseltamivir, and many others.[17] We therefore
further focused on CES1 and proceeded to evaluate and
quantify the disparate CES1 inhibitory activity of the E- and Z-
isomers of ureas 4 and 5 using gel-based competition
experiments with probe 14. Hep G2 cell lysate was treated
with different concentrations of 4 or 5 followed by treatment
with probe 14, CES1 band fluorescence signal was quantified,
and IC50 values were calculated (Figures 5A-D). E-4 and E-5
had similar IC50 values (13 and 11 nM respectively; Figures
5B and 5D). Intriguingly, we found that the E- and irradiated
E-isomers of 4 had a 7.4-fold difference in IC50 values, while
the isomers of 5 had a 5.7-fold difference. We therefore
decided to proceed with inhibitor 4. Further, we performed the
isomerization experiment on live Hep G2 cells. Cells were
treated with E-4, then left in the dark at 4 °C or irradiated with
365 nm light for 5 minutes at 4 °C, and then incubated at 37 °C
times between
Z
and E-isomers without noticeable
decomposition (Figures 6D and 6E). Furthermore, we
performed a series of photoisomerization experiments to
ascertain that our compounds remain stable and switch
This article is protected by copyright. All rights reserved.

10.1002/anie.202011163
Angewandte Chemie International Edition
RESEARCH ARTICLE
Figure 6. Photoisomerization properties of arylazopyrazole sulfone 16. A) UV-Vis spectrum of 16 in DMSO. B) 1H NMR spectra and (C) quantification of Z-16
conversion to E-16 over time at 37 °C. E-16 was converted to Z-16 by 10 mins irradiation at 365 nm UV light. D) 1H NMR spectra and (E) plot representation of 16
repeated cycling between Z and E-isomers.
We then proceeded to evaluate the selectivity of 16
towards binding to other serine hydrolases. We first optimized
the labeling condition using various broad activity-based
probes to observe most optimal competition of CES1. In this
context, our recently reported para-nitrophenyl phosphonate
activity-based probe 17 (30 nM) provided the best
visualization of 16 binding to CES1 in Hep G2 lysate (Figure
S14).[11] Under the same conditions, we performed an MS
experiment to identify proteins competed by sulfone 16 in Hep
G2 lysate. Because of the reversible nature of inhibitor 16,
probe 17 was again used at low concentration (30 nM). Thus
only a modest number of 12 serine hydrolases was detected
in this experiment. Indeed at 80 μM of E-16, only CES1 and
CES2 had ratios below 0.5 relative to the DMSO control
(Figure 7D; Table S4). Thus, based on our gel- and MS-based
competitive proteomics experiments, we conclude that our
sulfone 16 retained the remarkable specificity towards CES1
and CES2.
reliably in aqueous buffers. To this end, we determined the
thermal relaxation half-lives (Figure S10) and measured the
Z-E photoisomerization of compounds Z-4, Z-5, Z-14 and Z-
16 in PBS (Figure S11), and evaluated the stability of 14 and
16 upon repeated switching in PBS containing 10 mM
glutathione (Figure S12). The obtained results confirm
previous observations[19] of compatibility of arylazopyrazole
photoswitches with aqueous buffers.
We first tested whether 16 still binds to CES1 via gel-
based competition assays with the probe 14. Visualizing
competition of a covalently binding probe with a potentially
non-covalent inhibitor requires careful optimization of the time,
temperature, and probe concentration, because the
irreversible back exchange between the enzyme-bound
reversible inhibitor and the irreversible probe must be
minimized. Reversible inhibitors will only compete irreversible
probe labeling of a given enzyme for a limited period of time,
depending on the inhibitor affinity and the probe reactivity.[10]
Hep G2 lysate was treated with E-16 for one hour at 37 °C and
then rested on ice for 5 min before the addition of 14 to a final
concentration of 30 nM. The samples were incubated 15 min
on ice followed by conjugation to TMR azide and visualized by
in-gel fluorescence as described above. Using these
conditions, we observed a 3-fold difference in IC50 values
between E-16 (37µM) and irradiated E-16 (112 µM; Figures
7A and 7B). To further confirm that 16 binds CES1 reversibly,
we performed gel filtration experiments followed by gel-based
ABPP.[8b] Hep G2 lysate was incubated with 16 for 1 hour at
37 °C and a part of the mixture was subjected to Sephadex
gel filtration to remove the non-covalently bound inhibitor. The
filtered and unfiltered samples were normalized for protein
concentration using the Bradford assay and then were
subjected to treatment with probe 14 (30 nM). Indeed, we
observed competition of CES1 labeling by sulfone 16 without
filtration, while filtration almost completely restored CES1
labeling (Figures 7C and S13). In contrast, filtration of urea 4
did not result in recovery of CES1 labeling (Figures 7C and
S13). These results strongly suggest that 16 binds reversibly
and 4 binds covalently as we predicted. While sulfone-based
serine hydrolase inhibitors have been reported previously, [20]
designing reversible inhibitors from a covalent scaffold by
Figure 7. Arylazopyrazole sulfone 16 reversibly binds to CES1. A) Competitive
gel-based assay and (B) quantification comparing the binding of E-16 and
irradiated E-16 to endogenous CES1 in Hep G2 lysate using 14 (30 nM) as
probe. C) Quantification of the competitive gel-based assay where gel filtration
reduces competition by E-16 but not E-4 of CES1 labeling by probe 14 in Hep
G2 lysate (see Figure S13 for gel images). D) Chemical structure of the activity-
based serine hydrolase probe 17[11] and competitive MS-based identification of
E-16 (80 μM) protein targets using 17 (30 nM) in Hep G2 lysate (n = 6). Error
bars are SD *** indicates p < 0.001 by Student’s t-test.
replacing
a carbonyl with a less electrophilic sulfone
represents a novel and very simple approach for discovery of
serine hydrolase reversible inhibitors.
This article is protected by copyright. All rights reserved.

10.1002/anie.202011163
Angewandte Chemie International Edition
RESEARCH ARTICLE
16 in the presence of large, non-limiting excess MMF
(150 µM) and irradiated the sample repetitively for two
complete rounds of isomerization (Figure 8D). Indeed,
photoswitching between the E- and Z- isoforms conferred
significant changes in MMF hydrolysis rate consistent with our
experimental data shown above. These experiments,
collectively, show that our photoswitchable inhibitors can
temporally control the half-life and metabolism of drugs such
as mycophenolate mofetil via optical control over the key
metabolizing enzyme CES1.
Since CES1 plays an important role in the metabolism of
many ester-containing drugs, light-mediated control of CES1
activity using our inhibitors would provide valuable information
over the CES1-mediated metabolism and pharmacokinetics of
these drugs in complex proteomes such as cellular and tissue
lysates. We employed an MS-based assay to measure the
CES-catalyzed hydrolysis of the ester-containing prodrug
mycophenolate mofetil 18 (MMF) to mycophenolic acid 19
(MPA) and 2-morpholinoethan-1-ol (Figure 8A).[21] MMF is a
known substrate of CES1 and an immunosuppressant that is
widely applied to prevent organ transplant rejection and to
treat Crohn's disease.[21-22] The morpholino ethyl ester in MMF
masks the polar carboxylate group to increase the drug
lipophilicity. We treated native Caco-2 lysate with the E-
configured inhibitor 4, with or without concomitant
photoisomerization, for one hour at 37 °C before the addition
of MMF (30 μM). At the indicated time, extraction with ethyl
acetate was performed and quantification of the remaining
MMF was done via LC-MS. As expected, 30 nM E-4 but not
irradiated E-4 inhibited CES1-mediated hydrolysis of 18
(Figure 8B). Indeed, E-4 increased the half-life of MMF from
43 to 79 min (Figure 8B). The half-life of MMF in irradiated E-
4 treated lysate was nearly identical to the value obtained in
the DMSO treated samples. Remarkably, also treatment with
our reversible sulfone inhibitor E-16 led to a 2.1-fold increase
in the half-life of MMF relative to irradiated E-16 (Δt1/2 = 29
min; Figure 8C). Importantly, in our genetic control
experiment, siRNA-mediated knockdown of CES1 also
doubled the half-life (Figure S15), thus precisely recapitulating
the results obtained with E-4 and E-16.
Conclusion
In summary, we have developed arylazopyrazole ureas
and sulfones as a novel class of photoswitchable serine
hydrolase inhibitors and established a chemoproteomic
platform that allowed rapid screening of our inhibitors to
discover optically controlled serine hydrolase targets in
complex cellular lysates. We identified highly potent and
selective, photoswitchable inhibitors of drug-metabolizing
enzymes carboxylesterase 1 (CES1) and 2 (CES2) and
demonstrated the utility of these photoswitchable inhibitors by
optically
controlling
the
metabolism
of
the
immunosuppressant drug mycophenolate mofetil in cell
lysates. To the best of our knowledge, this is the first study to
demonstrate optical control of drug metabolism via activity
modulation of a drug metabolizing enzyme.
Aside from their photoswitchability, the herein reported
inhibitors offer several advantages over the existing probes of
the human CES enzymes.[7a, 12] First, our urea- and sulfone-
based compounds inhibit both CES1 and CES2, two closely
related enzymes that share ~46% sequence identity and
therefore have largely overlapping xenobiotic substrates.[21]
Thus, dual CES inhibition with our compounds should lead to
a decisively longer (pro)drug half-life compared to isoform-
selective inhibitors. Additionally, as shown by our gel- and MS-
based proteomics experiments, our inhibitors showed
extraordinary specificity for CES1 and 2, without any
detectable off-target activity against other serine hydrolases
or other classes of proteins, thus allowing direct quantification
of CES1/2-dependent drug hydrolysis in complex cell lysates
with dozens of other natively expressed serine hydrolases.
Owing the exceptional potency and specificity of our
CES inhibitors, we anticipate that these compounds will have
broad applications as photopharmacological tools to study
CES1/2-mediated drug metabolism and prodrug activation, for
example in hepatocyte stability or microsomal stability assays.
The results presented herein demonstrate that such assays
can directly be performed in complex cellular lysates with
endogenous carboxylesterase expression levels. A large
variety of clinical drugs are deactivated through hydrolysis or
activated through prodrug cleavage by CES1/2, including the
herein studied immunosuppressant prodrug mycophenolate
Figure 8. Optical control of CES1-mediated drug hydrolysis. A) CES1 catalyzes
hydrolysis of 18 (MMF) to the immunosuppressant drug 19 (MPA). B-C) MS-
based hydrolysis assay measuring the CES1-catalyzed hydrolysis of 18 in
presence of B) E-4 and irradiated E-4 (30 nM) or C) E-16 and irradiated E-16
(80 μM). D) Reversible optical control of CES1-mediated MMF hydrolysis by
repeated photoswitching of the reversible inhibitor 16 (80 μM). Dashed lines
represent linear extrapolation of MMF hydrolysis. Green lines represent
irradiation time periods. Error bars are SD (n = 3).
mofetil,
the
anticancer
prodrug
irinotecan,
the
antihypertensive prodrug imidapril, or the popular anaesthetic
drug lidocaine.[23] They may also serve as a blueprint for future
development of precision medicines under optical control.
Apart from their well-established function as major xenobiotic
hydrolases in human cells, growing evidence links CES1/2 to
hypertriglyceridaemia, diabetes, and obesity,[24] implicating
Finally, we explored the possibility of reversible optical control
of MMF hydrolysis in lysates by repeated photoisomerization
of the reversible inhibitor 16. We treated Caco-2 lysate with E-
This article is protected by copyright. All rights reserved.

10.1002/anie.202011163
Angewandte Chemie International Edition
RESEARCH ARTICLE
[2]
[3]
G. C. R. Ellis-Davies, Nat Methods 2007, 4, 619-628.
CES1 and CES2 as possible targets for pharmaceutical
intervention beyond drug metabolism.
a) X. Chen, S. Wehle, N. Kuzmanovic, B. Merget, U. Holzgrabe, B.
Konig, C. A. Sotriffer, M. Decker, ACS Chem Neurosci 2014, 5,
377-389; b) J. Broichhagen, I. Jurastow, K. Iwan, W. Kummer, D.
Trauner, Angew Chem Int Edit 2014, 53, 7657-7660; c) P. R.
Westmark, J. P. Kelly, B. D. Smith, J Am Chem Soc 1993, 115,
3416-3419; d) A. J. Harvey, A. D. Abell, Tetrahedron 2000, 56,
9763-9771; e) D. Pearson, A. D. Abell, Org Biomol Chem 2006, 4,
3618-3625; f) D. Pearson, N. Alexander, A. D. Abell, Chemistry
2008, 14, 7358-7365; g) A. D. Abell, M. A. Jones, A. T. Neffe, S.
G. Aitken, T. P. Cain, R. J. Payne, S. B. McNabb, J. M. Coxon, B.
G. Stuart, D. Pearson, H. Y. Y. Lee, J. D. Morton, J Med Chem
2007, 50, 2916-2920.
a) W. Szymanski, M. E. Ourailidou, W. A. Velema, F. J. Dekker, B.
L. Feringa, Chemistry 2015, 21, 16517-16524; b) L. Albert, J. Xu,
R. W. Wan, V. Srinivasan, Y. L. Dou, O. Vazquez, Chemical
Science 2017, 8, 4612-4618; c) R. Mogaki, K. Okuro, T. Aida, J
Am Chem Soc 2017, 139, 10072-10078.
a) A. Acosta-Ruiz, V. A. Gutzeit, M. J. Skelly, S. Meadows, J. Lee,
P. Parekh, A. G. Orr, C. Liston, K. E. Pleil, J. Broichhagen, J.
Levitz, Neuron 2020, 105, 446-463 e413; b) W. C. Lin, M. C. Tsai,
R. Rajappa, R. H. Kramer, J Am Chem Soc 2018, 140, 7445-7448;
c) S. Berlin, S. Szobota, A. Reiner, E. C. Carroll, M. A. Kienzler, A.
Guyon, T. Xiao, D. Trauner, E. Y. Isacoff, Elife 2016, 5; d) R.
Ferreira, J. R. Nilsson, C. Solano, J. Andreasson, M. Grotli, Sci
Rep 2015, 5, 9769; e) M. Borowiak, W. Nahaboo, M. Reynders, K.
Nekolla, P. Jalinot, J. Hasserodt, M. Rehberg, M. Delattre, S.
Zahler, A. Vollmar, D. Trauner, O. Thorn-Seshold, Cell 2015, 162,
403-411; f) M. Borowiak, F. Kullmer, F. Gegenfurtner, S. Peil, V.
Nasufovic, S. Zahler, O. Thorn-Seshold, D. Trauner, H. D. Arndt, J
Am Chem Soc 2020, 142, 9240-9249.
a) A. Damijonaitis, J. Broichhagen, T. Urushima, K. Hull, J. Nagpal,
L. Laprell, M. Schonberger, D. H. Woodmansee, A. Rafiq, M. P.
Sumser, W. Kummer, A. Gottschalk, D. Trauner, ACS Chem.
Neurosci. 2015, 6, 701-707; b) M. Schoenberger, A. Damijonaitis,
Z. Zhang, D. Nagel, D. Trauner, ACS Chem. Neurosci. 2014, 5,
514-518.
a) D. Wang, L. Zou, Q. Jin, J. Hou, G. Ge, L. Yang, Acta Pharm
Sin B 2018, 8, 699-712; b) S. C. Laizure, V. Herring, Z. Hu, K.
Witbrodt, R. B. Parker, Pharmacotherapy 2013, 33, 210-222; c) M.
J. Hatfield, R. A. Umans, J. L. Hyatt, C. C. Edwards, M. Wierdl, L.
Tsurkan, M. R. Taylor, P. M. Potter, Chem Biol Interact 2016, 259,
327-331.
a) A. Adibekian, B. R. Martin, C. Wang, K. L. Hsu, D. A.
Bachovchin, S. Niessen, H. Hoover, B. F. Cravatt, Nature
Chemical Biology 2011, 7, 469-478; b) A. Adibekian, B. R. Martin,
J. W. Chang, K. L. Hsu, K. Tsuboi, D. A. Bachovchin, A. E. Speers,
S. J. Brown, T. Spicer, V. Fernandez-Vega, J. Ferguson, P. S.
Hodder, H. Rosen, B. F. Cravatt, J Am Chem Soc 2012, 134,
10345-10348.
a) C. E. Weston, R. D. Richardson, P. R. Haycock, A. J. White, M.
J. Fuchter, J Am Chem Soc 2014, 136, 11878-11881; b) L.
Stricker, E. C. Fritz, M. Peterlechner, N. L. Doltsinis, B. J. Ravoo,
J. Am. Chem. Soc. 2016, 138, 4547-4554.
D. Leung, C. Hardouin, D. L. Boger, B. F. Cravatt, Nat Biotechnol
2003, 21, 687-691.
Using light to spatially or temporally deactivate the
administered E-isomers of our CES inhibitors may create
exciting opportunities of spatiotemporal control over prodrug
activation and/or drug metabolism in clinical settings. Many
medication side effects are caused by high circulating
concentrations of the drug in the human body. For example,
high concentrations of SN-38, the active, hydrolyzed form of
irinotecan, can cause life-threatening enterocolitis and
diarrhea in cancer patients.[25] Stepwise activation of CES
activity in the liver and colon through E-Z conversion or, in
case of acute onset of gastrointestinal symptoms, rapid on-
demand deactivation of CES by Z-E photoisomerization of our
inhibitor may reduce the circulating concentration and
accumulation of SN-38 in colon and prevent these severe side
effects.
[4]
[5]
Covalent inhibitors offer several advantages over
reversible drugs[26] and there has been a resurgence in the
development of covalent drugs.[26a] Yet having reversible CES
inhibitors with optical control to allow repeated cycling
between active and inactive states may be desirable in basic
drug metabolism research. Herein, we successfully obtained
a reversible CES inhibitor by replacing the carbonyl group in
inhibitor 4 with a less electrophilic sulfone, demonstrating a
novel strategy for the design of reversible serine hydrolase
inhibitors starting from covalent compounds. Indeed, the
sulfone-based inhibitor 16 enabled repeated cycling between
the active and inactive conformations in our MMF metabolism
experiment.
[6]
[7]
[8]
Serine hydrolases are a large and functionally very
diverse class of enzymes with more than 200 different
members expressed in the human proteome.[27] Our
arylazopyrazole ureas and sulfones offer synthetically
accessible scaffolds that can be expanded to a larger library
with hundreds of compounds to identify specific
photoswitchable inhibitors for other serine hydrolases,
including lipases, peptidases, and proteases. As a result, it is
conceivable that our approach may potentiate optical control
over proteolysis, lipid homeostasis, peptide hormone
signaling, and many other serine hydrolase-regulated
metabolic processes of fundamental significance to the cell.
[9]
[10]
[11]
[12]
[13]
C. Wang, D. Abegg, B. G. Dwyer, A. Adibekian, Chembiochem
2019, 20, 2212-2216.
D. D. Wang, L. W. Zou, Q. Jin, J. Hou, G. B. Ge, L. Yang,
Fitoterapia 2017, 117, 84-95.
Finally, our chemoproteomic platform can be expanded
to other scaffolds and photoswitches to enable optical control
of other protein classes. These studies are currently underway
and will be reported in due course. Projecting forward, we
strongly believe that the herein presented platform will be a
valuable tool for future development of photopharmaceuticals.
M. Uhlen, L. Fagerberg, B. M. Hallstrom, C. Lindskog, P. Oksvold,
A. Mardinoglu, A. Sivertsson, C. Kampf, E. Sjostedt, A. Asplund, I.
Olsson, K. Edlund, E. Lundberg, S. Navani, C. A. Szigyarto, J.
Odeberg, D. Djureinovic, J. O. Takanen, S. Hober, T. Alm, P. H.
Edqvist, H. Berling, H. Tegel, J. Mulder, J. Rockberg, P. Nilsson, J.
M. Schwenk, M. Hamsten, K. von Feilitzen, M. Forsberg, L.
Persson, F. Johansson, M. Zwahlen, G. von Heijne, J. Nielsen, F.
Ponten, Science 2015, 347, 1260419, www.proteinatlas.org.
a) M. J. Hatfield, P. M. Potter, Expert Opin Ther Pat 2011, 21,
1159-1171; b) L. W. Zou, Q. Jin, D. D. Wang, Q. K. Qian, D. C.
Hao, G. B. Ge, L. Yang, Curr Med Chem 2018, 25, 1627-1649; c)
D. A. Bachovchin, T. Y. Ji, W. W. Li, G. M. Simon, J. L. Blankman,
A. Adibekian, H. Hoover, S. Niessen, B. F. Cravatt, P Natl Acad
Sci USA 2010, 107, 20941-20946.
Acknowledgements
[14]
[15]
We thank The Scripps Research Institute for financial support.
Keywords: carboxylesterases • drug discovery • inhibitors •
photopharmacology • proteomics
S. Macintyre, D. Samols, P. Dailey, J. Biol. Chem. 1994, 269,
24496-24503.
[16]
[17]
T. Imai, K. Ohura, Curr Drug Metab 2010, 11, 793-805.
a) H. J. Zhu, J. S. Markowitz, Drug Metab Dispos 2009, 37, 264-
267; b) H. J. Zhu, X. W. Wang, B. E. Gawronski, B. J. Brinda, D. J.
Angiolillo, J. S. Markowitz, J Pharmacol Exp Ther 2013, 344, 665-
672; c) K. Nishi, H. Z. Huang, S. G. Kamita, I. H. Kim, C.
Morisseau, B. D. Hammock, Arch Biochem Biophys 2006, 445,
[1]
a) K. Hüll, J. Morstein, D. Trauner, in Chemical Reviews, Vol. 118,
2018, pp. 10710-10747; b) J. Broichhagen, J. A. Frank, D.
Trauner, Acc Chem Res 2015, 48, 1947-1960; c) M. M. Lerch, M.
J. Hansen, G. M. van Dam, W. Szymanski, B. L. Feringa, Angew.
Chem. Int. Ed. Engl. 2016, 55, 10978-10999.
This article is protected by copyright. All rights reserved.

10.1002/anie.202011163
Angewandte Chemie International Edition
RESEARCH ARTICLE
115-123; d) P. M. Potter, J. S. Wolverton, C. L. Morton, M. Wierdl,
M. K. Danks, Cancer Res 1998, 58, 3627-3632.
[18]
a) F. Bonardi, G. London, N. Nouwen, B. L. Feringa, A. J.
Driessen, Angew Chem Int Ed Engl 2010, 49, 7234-7238; b) M. J.
Hansen, W. A. Velema, G. de Bruin, H. S. Overkleeft, W.
Szymanski, B. L. Feringa, Chembiochem 2014, 15, 2053-2057; c)
M. Wegener, M. J. Hansen, A. J. M. Driessen, W. Szymanski, B. L.
Feringa, J Am Chem Soc 2017, 139, 17979-17986; d) M. M. Lerch,
M. J. Hansen, G. M. van Dam, W. Szymanski, B. L. Feringa,
Angew Chem Int Edit 2016, 55, 10978-10999.
[19]
[20]
[21]
[22]
L. Stricker, M. Bockmann, T. M. Kirse, N. L. Doltsinis, B. J. Ravoo,
Chem-Eur J 2018, 24, 8639-8647.
W. Du, C. Hardouin, H. Cheng, I. Hwang, D. L. Boger, Bioorg.
Med. Chem. Lett. 2005, 15, 103-106.
T. Fukami, M. Kariya, T. Kurokawa, A. Iida, M. Nakajima, Eur J
Pharm Sci 2015, 78, 47-53.
a) D. Cattaneo, M. Cortinovis, S. Baldelli, A. Bitto, E. Gotti, G.
Remuzzi, N. Perico, Clin J Am Soc Nephro 2007, 2, 1147-1155; b)
S. E. Hughes, S. A. Gruber, J Clin Pharmacol 1996, 36, 1081-
1092; c) T. R. Srinivas, B. Kaplan, H. U. Meier-Kriesche, Expert
Opin Pharmaco 2003, 4, 2325-2345; d) M. F. Neurath, R.
Wanitschke, M. Peters, F. Krummenauer, K. H. Meyer zum
Buschenfelde, J. F. Schlaak, Gut 1999, 44, 625-628; e) T. Tan, I.
C. Lawrance, World J Gastroentero 2009, 15, 1594-1599.
T. Fukami, M. Kariya, T. Kurokawa, A. Iida, M. Nakajima, Eur J
Pharm Sci 2015, 78, 47-53.
a) E. Dominguez, A. Galmozzi, J. W. Chang, K. L. Hsu, J. Pawlak,
W. Li, C. Godio, J. Thomas, D. Partida, S. Niessen, P. E. O'Brien,
A. P. Russell, M. J. Watt, D. K. Nomura, B. F. Cravatt, E. Saez,
Nature chemical biology 2014, 10, 113-121; b) M. Jernas, B.
Olsson, P. Arner, P. Jacobson, L. Sjostrom, A. Walley, P. Froguel,
P. G. McTernan, J. Hoffstedt, L. M. Carlsson, Biochem. Biophys.
Res. Commun. 2009, 383, 63-67; c) G. R. Steinberg, B. E. Kemp,
M. J. Watt, Am. J. Physiol. Endocrinol. Metab. 2007, 293, E958-
964; d) E. Wei, Y. Ben Ali, J. Lyon, H. Wang, R. Nelson, V. W.
Dolinsky, J. R. Dyck, G. Mitchell, G. S. Korbutt, R. Lehner, Cell
Metab. 2010, 11, 183-193; e) V. W. Dolinsky, D. Gilham, M. Alam,
D. E. Vance, R. Lehner, Cell. Mol. Life Sci. 2004, 61, 1633-1651.
E. Araki, M. Ishikawa, M. Iigo, T. Koide, M. Itabashi, A. Hoshi, Jpn
J Cancer Res 1993, 84, 697-702.
[23]
[24]
[25]
[26]
[27]
a) R. A. Bauer, Drug Discov Today 2015, 20, 1061-1073; b) A.
Tuley, W. Fast, Biochemistry 2018, 57, 3326-3337.
D. A. Bachovchin, B. F. Cravatt, Nat Rev Drug Discov 2012, 11,
52-68.
This article is protected by copyright. All rights reserved.

10.1002/anie.202011163
Angewandte Chemie International Edition
RESEARCH ARTICLE
RESEARCH ARTICLE
Brendan G. Dwyer, Chao Wang,
Daniel Abegg, Brittney Racioppo, Nan
Qiu, Zhensheng Zhao, Dany
Pechalrieu, Anton Shuster, Dominic
G. Hoch, and Alexander Adibekian*
Page No. – Page No.
Chemoproteomics-Enabled De Novo
Discovery of Photoswitchable
Carboxylesterase Inhibitors for
Optically Controlled Drug Metabolism
Here, we present a new chemoproteomic platform for the de novo discovery of photoswitchable serine hydrolase inhibitors. Using this
platform, we identified highly potent and selective photoswitchable inhibitors of the drug-metabolizing enzyme carboxylesterase and
demonstrated the pharmacological application of our compounds by optically controlling hydrolysis of an immunosuppressant drug.
This article is protected by copyright. All rights reserved.