Quantifying drug-target engagement in live cells using sulfonyl fluoride chemical probes

Article abstract of DOI:10.1016/bs.mie.2019.02.010

Phenotypic screening in disease-relevant models identifies small molecule hits with desirable efficacy but often with unknown modes of action. Target identification and validation are integral to successful biomedical research. Technologies are required to validate the biological target (or targets) through which a pharmacological agent is proposed to exert its effects. This work details the rational structure-based design, synthetic preparation and cell-based application of a clickable sulfonyl fluoride chemical probe to directly report on the mechanism of a series of compounds previously discovered in a reporter gene assay. Quantification of drug-target occupancy in living human primary cells enabled a deeper understanding of the molecular pharmacology of the chemotype. The technology described herein should be of broad interest to those involved in chemical biology research and the drug discovery endeavor.

Full text of DOI:10.1016/bs.mie.2019.02.010

ARTICLE IN PRESS
Quantifying drug-target
engagement in live cells
using sulfonyl fluoride chemical
probes
a,
b
c
Lyn H. Jones *, Hua Xu , Olugbeminiyi O. Fadeyi
a
Jnana Therapeutics, Boston, MA, United States
Pfizer Inc, Medicine Design, Cambridge, MA, United States
Exploratory Science Center, Merck & Co, Cambridge, MA, United States
b
c
*
Corresponding author: e-mail address: ljones@jnanatx.com
Contents
1
2
3
4
. Introduction
2
5
6
. Target validation of the decapping scavenger enzyme (DcpS)
. Structure-based chemical probe design
. Preparation of sulfonyl fluoride chemical probes
8
4
4
4
4
.1 Equipment
.2 Chemicals
.3 Synthesis of SF-p1
.4 Synthesis of SF-p1-yne
8
8
8
9
5
6
. Protocols
13
13
15
16
17
18
18
5
5
5
.1 Covalent labeling of recombinant DcpS
.2 Labeling of DcpS in live cells
.3 Competition labeling and quantification of target engagement
. Conclusions
Acknowledgments
References
Abstract
Phenotypic screening in disease-relevant models identifies small molecule hits with
desirable efficacy but often with unknown modes of action. Target identification and
validation are integral to successful biomedical research. Technologies are required
to validate the biological target (or targets) through which a pharmacological agent
is proposed to exert its effects. This work details the rational structure-based design, syn-
thetic preparation and cell-based application of a clickable sulfonyl fluoride chemical
probe to directly report on the mechanism of a series of compounds previously discov-
ered in a reporter gene assay. Quantification of drug-target occupancy in living human
Methods in Enzymology
ISSN 0076-6879
https://doi.org/10.1016/bs.mie.2019.02.010
# 2019 Elsevier Inc.
All rights reserved.
1

ARTICLE IN PRESS
2
Lyn H. Jones et al.
primary cells enabled a deeper understanding of the molecular pharmacology of the
chemotype. The technology described herein should be of broad interest to those
involved in chemical biology research and the drug discovery endeavor.
1. Introduction
Successful drug discovery relies heavily on rigorous, and unbiased,
therapeutic target validation ( Jones, 2016). Productivity in the pharma-
ceutical industry is plagued by attrition in Phase II where the efficacy of
the candidate drug is usually tested in patients for the first time. A key rea-
son for this derives from an incomplete picture of the pharmacokinetic-
pharmacodynamic (PK-PD) relationships of the drug candidate and its
target under investigation. To build confidence in translational pharmacol-
ogy we have previously defined a framework that attempts to establish confi-
dence in target validation using chemical probes, such that high quality
pharmacological modulators of pathophysiologically-relevant targets are
progressed through preclinical and clinical R&D (Bunnage, Chekler, &
Jones, 2013). The four pillars of target validation can be framed as questions:
(1) Does the compound achieve the desired concentration at the site of
action? (2) Does the compound engage the therapeutic target? (3) Does
engagement of the target result in the desired functional effects? (4) Does this
functional pharmacology have the desired phenotypic consequences in a
disease-relevant context? It is important not only to confirm that the drug
under investigation engages the nominated target in the relevant location/
compartment but also to quantify the relationship between drug-target
occupancy and its functional consequences (i.e., between pillar 2 and 3) as this
builds translational PK-PD understanding (Morgan et al., 2012).
Functional chemical probes have found a myriad of uses in chemical
biology and medicinal chemistry. For example, reactivity- and activity-
based probes are small molecules that possess electrophilic motifs that are
designed to engage their target proteins resulting in the formation of a cova-
lent bond (Cravatt, Wright, & Kozarich, 2008; Sanman & Bogyo, 2014).
The probes also incorporate a reporter group such as a fluorescent dye that
enables visualization of the protein using Western blot or imaging analysis
(Fig. 1). Alternatively, an appended biotin group provides an opportunity
to streptavidin-enrich low expressing adducted proteins from complex
proteomes. Analytical techniques employing quantitative mass spectrometry

Fig. 1 General chemical probe-based labeling strategy to validate target engagement in complex proteomes.

ARTICLE IN PRESS
4
Lyn H. Jones et al.
(MS) proteomics have considerably advanced the area as they provide a
simultaneous readout of multiple proteins, enabling assessments of drug
selectivity in complex proteomes (Drewes & Knapp, 2018). Importantly,
labeling and subsequent enrichment of the isolated protein (or proteins)
can be competed by pre-incubation of the proteome with the parent drug
or compound under investigation. By performing these experiments with
different doses of the drug, a quantitative assessment of target occupancy
is achieved (Fig. 1).
Unfortunately, theconjugationofa dye or biotin functionalitytotheprobe
perturbs its physicochemical properties that often hinders cellular permeation
(significantly increasing molecular weight, hydrogen bonding capacity and
electric charge) and as a result, these probes are only suitable for use in cell
lysate. A more physiologically-relevant assessment of drug-target engagement
is essential to understand the “true” nature of on-target occupancy and off-
target selectivity. It is not unusual to observe differences in target engagement
between cell lysate and intact cells andtissues. Reasons for variations in binding
site occupancy include context-dependent protein-protein interactions, post-
translational modifications, substrate competition and subcellular compart-
mentalization of the target protein, or the chemical probe itself. Technologies
that assess drug-target engagement in live cells are therefore of considerable
interest and are an active area of drug discovery research ( Jones, 2015).
The advent of “click chemistry” has considerably advanced the area of
functional chemical probe development. The concept of click chemistry
was developed by Sharpless and coworkers in 2001 that refers to highly effi-
cient, modular and often biocompatible reactions (Kolb, Finn, & Sharpless,
2
001). A commonly used click reaction is the copper(I)-catalyzed azide-
alkyne cycloaddition (CuAAC) that is orthogonal to other reactive biomo-
lecular functionalities (Rostovtsev, Green, Fokin, & Sharpless, 2002;
Tornøe, Christensen, & Meldal, 2002). Incorporation of one of the reaction
partners (azide or alkyne) into a chemical probe does not significantly per-
turb its molecular weight or physicochemical properties. Reactivity-based
probes that switch a dye or biotin for a “silent” click reporter such as a ter-
minal alkyne are thus enabled for cellular permeation and use in live cells. In
this workflow, cells are treated with the chemical probe, lysed and then the
azide-tagged biotin or dye (or both if a bifunctional linker is used) is
“clicked” onto the probe-protein adducts to mediate enrichment and anal-
ysis. As before, pre-incubation of the proteome with the drug being inves-
tigated at different concentrations and competition labeling allows target
occupancy to be quantified in whole cells.

ARTICLE IN PRESS
Quantifying drug-target engagement in live cells
5
Regarding selection of the electrophilic warhead, most approaches
target the highly nucleophilic cysteine residue using a variety of reactive
functionalities (Liu et al., 2013). However, there are issues associated with
targeting cysteine—it is difficult to achieve selectivity due to its intrinsically
high nucleophilicity, and the residue is under-represented in protein func-
tional sites. Therefore, new synthetic labeling technologies are required to
target residues beyond cysteine. Sulfonyl fluoride warheads have received
considerable attention in the design of chemical probes due to their ability
to react with a variety of nucleophilic residues beyond cysteine (tyrosine,
lysine, histidine, serine and threonine) in a context-dependent manner,
i.e., residue engagement is templated (Narayanan & Jones, 2015). This is
due to a need to activate the fluorine leaving group through hydrogen
bonding in the microenvironment of the binding site (Dong, Krasnova,
Finn, & Sharpless, 2014).
2
. Target validation of the decapping scavenger
enzyme (DcpS)
Spinal muscular atrophy is a rare genetic neuromuscular disease that
causes muscle wasting and often leads to early death. It is caused by mutations
in SMN1 that codes for survival motor neuron (SMN) protein. Another
gene, SMN2 also codes for the SMN protein and small molecules that
upregulate SMN2 expression have therefore been of interest for the treat-
ment of SMA. Previously, cell-based screening in a reporter gene assay iden-
tified a series of diaminoquinazoline (DAQ) derivatives that potently
upregulated SMN2 ( Jarecki et al., 2005; Thurmond et al., 2008). Medicinal
chemistry optimization yielded a clinical candidate RG3039, yet the biolog-
ical target had not been determined at that time. Protein microarrays, that
utilize glutathione S-transferase (GST)-tagged and immobilized recombi-
nant proteins, were screened using a radiolabeled DAQ derivative and bind-
ing to a single protein, the mRNA decapping scavenger enzyme (DcpS), was
identified (Singh et al., 2008). A crystal structure of related DAQ derivatives
provided further evidence that the putative target was the DcpS enzyme due
to their location in the mRNA cap binding pocket of the enzyme (Singh
et al., 2008). However, there was no direct validation of DcpS as the target
of the DAQ series in the physiologically-relevant context of human live
cells. Moreover, the dibasic and lipophilic nature of RG3039 translates to
lysosomal accumulation (Gopalsamy et al., 2017), a feature that has been
reported to be linked to perturbation of autophagy (Ashoor, Yafawi,

ARTICLE IN PRESS
6
Lyn H. Jones et al.
Jessen, & Lu, 2013), and autophagic dysregulation is a characteristic of SMA
Garcera, Bahi, Periyakaruppiah, Arumugam, & Soler, 2013). It was there-
(
fore important to validate the in vivo efficacy of RG3039 derives from a
DcpS-dependent mechanism, and not from its physicochemical properties.
Here we detail the design, synthesis and application of a clickable sulfonyl
fluoride chemical probe that directly reports on DAQ-DcpS engagement in
primary human peripheral blood mononuclear cells (PBMCs) (Hett et al.,
2
015; Xu et al., 2016). This work is representative of a chemical labeling
strategy to quantify drug-target occupancy in live cells that should be of
broad interest to those working in chemical biology and drug discovery
research. The impact these studies had on the progression of DAQ inhibitors
for the treatment of SMA is also discussed.
3. Structure-based chemical probe design
As mentioned in the Introduction, most functional chemical probes
target reactive cysteine thiols, but it is an amino acid that is infrequently pre-
sent in protein functional sites. In the DAQ-DcpS crystal structures there are
no obviously targetable cysteine residues in the binding site (Singh et al.,
2
008). However, there are two tyrosine residues (Tyr113 and Tyr143)
clearly proximal to DAQ ligands. Specifically, a sulfonyl fluoride motif at
the para-position of the benzene ring would be positioned perfectly to attack
Tyr143 (Fig. 2). An analysis of tyrosine residues in proteins that were found
previously to fortuitously engage sulfonyl fluoride electrophiles identified a
preponderance of neighboring basic residues that presumably reduce the pKa
and enhance the nucleophilicity of the phenol hydroxyl group (Hett et al.,
2
015). It was noticeable that in the microenvironment of the DcpS cap bind-
ing site there is a proximal lysine (Lys142) that may enhance Tyr143 reac-
tivity (Fig. 2).
An important reason for the increased use of sulfonyl fluoride probes in
chemical biology research has been the recent advances in their synthetic
preparation and the availability of sulfonyl fluoride-containing reagents
(Dong et al., 2014; Narayanan & Jones, 2015). Benzylation of a DAQ-
phenol with para-fluorosulfonyl benzyl bromide readily delivered the
desired probe following protecting group removal. As predicted, SF-p1
was found to be a covalent DcpS inhibitor that reacted with Tyr143 (con-
firmed using X-ray crystallography and peptide mapping experiments).
Positioning of the click handle in the DAQ template was also enabled using
structure-based design, and a terminal alkyne ortho- to the sulfonyl fluoride

ARTICLE IN PRESS
Quantifying drug-target engagement in live cells
7
Fig. 2 Rational design of a clickable covalent chemical probe to directly measure in-cell
DcpS-DAQ target engagement. The crystal structure of a reversible DAQ inhibitor with
DcpS (e.g., PDB code 3BLA) shows the proximity of the para-position of the benzene ring
to the Tyr143 hydroxy group (dotted line). The SF-p1-DcpS crystal structure (PDB code
4
QDV) highlights the area of the binding pocket that can accommodate the terminal
alkyne click handle (white arrow).
electrophile would appear to occupy a hydrophobic pocket in the enzyme
(Fig. 2). Incorporation of the terminal alkyne relied on a palladium-
mediated Stille coupling reaction with an aryl bromide derivative (per-
formed in the presence of the sulfonyl fluoride functionality—see
Section 4.4). The physicochemical properties of the resulting probe
SF-p1-yne were commensurate with those of the original DAQ reversible
inhibitor, thus enabling subsequent cell-based occupancy experiments.

ARTICLE IN PRESS
8
Lyn H. Jones et al.
4. Preparation of sulfonyl fluoride chemical probes
4
. Silica gel flash chromatography was performed using a medium pressure
.1 Equipment
1
2
Biotage or ISCO system
1
. H, C[ H], and F[ H] NMR characterization data were collected at
13
1
19
1
3
1
00K on a Varian 400MHz (AS400) spectrometer operating at 400.1,
00.5, and 376MHz (respectively), or a Varian 500MHz (AS500) spec-
trometer operating at 500, 126, and 470MHz (respectively)
. IR spectra were recorded on a Nicolet Avatar 360 FT-IR spectrometer
. High resolution mass spectroscopy (HRMS) was performed on an
Agilent (6220) LC-MS TOF using an Xbridge C18 2.5μm 3.0ꢀ5.0mm
. LC-MS was performed using an Agilent 1100 HPLC system connected
to a Waters LCT Premier XE mass spectrometer
3
4
5
6
. Peptide mapping was performed using an LTQ Orbitrap XL mass
spectrometer
4.2 Chemicals
All chemicals were purchased from Sigma-Aldrich unless otherwise stated.
4.3 Synthesis of SF-p1
The synthesis of fluoride inhibitor SF-p1 is described in Fig. 3. Cyano-
difluorobenzene 1 was subjected to a nucleophilic aromatic substitution
(SNAr) reaction with benzyl alcohol to afford cyano-fluorobenzene derivative
Fig. 3 Synthetic scheme for the preparation of SF-p1, a sulfonyl fluoride covalent inhib-
itor of DcpS

ARTICLE IN PRESS
Quantifying drug-target engagement in live cells
9
2, which was then subjected to guanidine cyclization to provide benzyl ether
3. Boc-protection and benzyl deprotection gave the desired phenol 4.
O-alkylationwithpSO FC H CH BrandBoc-deprotectionprovidedSF-p1.
2
6
4
2
1
.1. Compound 4 was prepared according to the published procedure
as shown in Fig. 3 (Hett et al., 2015). The key synthetic steps to
provide SF-p1 are described below.
1.2. A mixture of compound 4 (80mg, 0.288mmol) and potassium car-
bonate (80mg, 0.288mmol) in acetone (5mL) was allowed to stir at
50°C for 1h. 4-(bromomethyl)-benzenesulfonyl fluoride (188mg,
0.6mmol) was added to the mixture, which was then allowed to stir
for 16h at 50°C. LC-MS showed the reaction to be complete. The
mixture was filtered and purified by prep-HPLC to give tert-butyl
2
-amino-5-(2-(fluorosulfonyl)benzyloxy)quinazolin-4-ylcarbamate
as a light yellow solid (also readily purified on silica gel, 0–100%
EtOAc/heptane).
1
.3. This material was then dissolved in dichloromethane (5mL) and the
resulting solution was cooled to 0°C and treated dropwise with 4M
HCl/EtOAc (0.5mL), and allowed to stir at room temperature over-
night. LC-MS showed the reaction to be complete. The mixture was
concentrated and the residue was purified by prep-HPLC to give the
title compound Synthesis of 2-((2,4-diaminoquinazolin-5-yloxy)
methyl)benzene-5-sulfonyl fluoride (SF-p1) (15mg, 15%) as a gray
1
solid. H NMR (400MHz, CD OD) δ 8.12 (2H, d, J ¼8.59Hz),
3
7
.85 (2H, d, J ¼8.20Hz), 7.67 (1H, t, J ¼8.39Hz), 6.99 (2H, dd,
19
J ¼8.39, 2.54Hz), 5.63 (2H, s) ppm. F NMR (376MHz, CD OD)
3
13
δ 64.00ppm. C NMR (100MHz, CD OD) δ 164.6, 158.1, 156.1,
3
1
1
45.9, 142.9, 137.6, 134.2 (1C, d, J ¼24.9Hz, C-SO -F), 130.1,
2
30.1, 110.6, 108.9, 101.7, 71.1ppm. IR (neat) 3323 (w, br), 1650
ꢁ
1
s), 1609 (s), 1404 (m), 1213 (m), 772 (w), 597 (w) cm . HRMS
(
+
ESI+) calcd for C H FN O S [M+H] : 352.0770, found:
(
15
14
4
3
352.0772.
4.4 Synthesis of SF-p1-yne
The synthesis of SF-p1-yne (Fig. 4) begins with the treatment of commer-
cially available sulfonyl chloride 5 with potassium fluoride to give the desired
sulfonyl fluoride. This was subjected to carboxylic acid reduction with
borane to furnish benzene sulfonyl fluoride intermediate 6 in good yield.
The benzyl alcohol was subsequently converted to benzyl bromide 7 and

ARTICLE IN PRESS
10
Lyn H. Jones et al.
Fig. 4 Synthetic scheme for the preparation of SF-p1-yne, a clickable functional probe
of DcpS occupancy.
coupled with phenol 4 to give sulfonyl fluoride intermediate 8. The terminal
alkyne was installed via Stille coupling, followed by Boc-deprotection that
yielded the desired sulfonyl fluoride probe, SF-p1-yne.
2
.1. To a mixture of the commercially available compound 3-bromo-4-
(
(
chlorosulfonyl)benzoic acid (5, 9.88g, 33mmol) and potassium fluoride
9.59g, 165mmol) was added acetone (82mL) and water (82mL). The
reactionwasstirredatroomtemperatureovernight.Thesolventwascon-
centrated under vacuum and the residue dissolved in ethyl acetate. The
organic layer was washed with brine, dried over magnesium sulfate, fil-
tered and concentrated. The residue was purified by chromatography on
silica gel eluting with a gradient of (0–50% EtOAc/heptane) to give
3
-bromo-4-(fluorosulfonyl)benzoic acid (7.5g, 80%) as a white solid.
1
H NMR (400MHz, (CD ) SO) δ 8.42 (1H, d, J ¼1.56Hz), 8.36
3
2
19
(
1H, d, J ¼8.59Hz), 8.21 (1H, d, J ¼8.20Hz) ppm. F NMR
13
(376MHz, (CD ) SO) δ 58.44 (s, 1F) ppm. C NMR (126MHz,
3
2
(
CD ) SO) δ 165.0, 139.2, 136.5, 135.9 (d, 2 JCF¼23.9Hz), 133.3,
3
2
1
7
29.9, 120.8ppm. IR (neat): 1702, 1510, 1417, 1305, 1281, 1214,
ꢁ
1
88, 763, 732, 618cm . HRMS (ESI-) m/z: calcd for C H BrFO S
7
4
4
ꢁ
[
M-H] : 280.8928; found: 280.8925.
.2. To a solution of 3-bromo-4-(fluorosulfonyl)benzoic acid (1.2g,
.24mmol) in THF (17mL) at 0 °C was added borane tetrahydrofuran
2
4
complex (1.0M in THF, 9.3mL, 9.3mmol) dropwise over 15min.
The reaction was gradually warmed to room temperature overnight
and quenched using saturated aqueous ammonium chloride solution.
The aqueous phase was extracted with EtOAc (2ꢀ100mL) and the

ARTICLE IN PRESS
Quantifying drug-target engagement in live cells
11
combined organic layers were washed with brine, dried over magne-
sium sulfate, filtered and concentrated. The residue was purified by
chromatography on silica gel (0–50% EtOAc/heptane) to give com-
1
pound 6 (0.85g, 75%) as a white solid. H NMR (400MHz,
(
CD ) SO) δ 8.21 (1H, d, J ¼8.20Hz), 7.99 (1H, s), 7.69 (1H, d,
3
2
J ¼8.20Hz), 5.66 (1H, t, J ¼5.85Hz), 4.66 (2H, d, J ¼5.85Hz)
1
9
13
ppm. F NMR (376MHz, (CD ) SO) δ 59.07 (s, 1F) ppm.
C
3
2
NMR (126MHz, (CD ) SO) δ 154.2, 133.6, 132.8, 130.6 (d, 2
3
2
JCF¼23.94Hz), 126.6, 120.3, 61.8ppm. IR (neat): 3331–3247,
1
5
2
591, 1413, 1350, 1211, 1049, 1029, 987, 875, 821, 773, 680, 586,
ꢁ
1
48cm . HRMS (ESI-) m/z: calcd for C H BrFO S [M-H] :
ꢁ
7
6
3
66.9131; found: 266.9132.
2.3. To a solution of 2-bromo-4-(hydroxymethyl)benzenesulfonyl fluo-
ride (6) (2.69g, 10.0mmol) in CH Cl (100mL) and DMF (77μL,
2
2
0
.1mmol) was added PBr (1M in CH Cl , 25.0mL, 2.5mmol) at
3 2 2
room temperature. The reaction was stirred at room temperature
overnight and then concentrated under reduced pressure. The residue
was diluted with CH Cl and quenched over ice. The organic phase
2
2
was separated, washed with saturated aqueous sodium bicarbonate,
brine, dried over magnesium sulfate, filtered and concentrated to
afford the crude product as oil. The residue was purified by silica
gel chromatography (0–50% EtOAc/heptane) to give compound 7
1
(
1.85g, 56%) as a white solid. H NMR (400MHz, (CD ) SO) δ
3
2
8
4
.24 (1H, d, J ¼8.20Hz), 8.20 (1H, s), 7.83 (1H, d, J ¼8.20Hz),
1
9
.81 (2H, s) ppm. F NMR (376MHz, (CD ) SO) δ 58.89 (s, 1F)
3
2
1
3
ppm. C NMR (126MHz, (CD ) SO) δ 148.8, 136.9, 133.3,
3
2
1
3
8
32.2 (d, 2 JCF¼30.24Hz), 130.1, 120.6, 31.2ppm. IR (neat):
210, 1584, 1557, 1406, 1269, 1203, 1150, 1118, 1030, 302, 885,
ꢁ
1
41, 768, 713, 663, 637, 570, 535cm . GC–MS: m/z calcd for
C H Br FO S [M]: 329.8361; found: 329.8.
7
5
2
2
2
.4. To a mixture of tert-butyl (2-amino-5-hydroxyquinazolin-4-yl)
carbamate (4) (1.1g, 4.0mmol), and 2-bromo-4-(bromomethyl)
benzenesulfonyl fluoride (7) (1.46g, 4.4mmol) in DMF (20mL)
was added Et N (1.62g, 2.23mL, 16.0mmol) at room temperature.
3
The reaction was heated at 60°C for 16h. After cooling to room
temperature, the reaction was diluted with EtOAc. The organic layer
was washed with water, saturated aqueous ammonium chloride, and
brine, dried over magnesium sulfate, filtered and concentrated. The
resultingresiduewas purified twiceby silica gel (0–20%MeOH/EtOAc)

ARTICLE IN PRESS
12
Lyn H. Jones et al.
1
to give the title compound (8) (0.52g, 25%) as a yellow solid. H NMR
(
500MHz, (CD ) SO): δ 9.61 (1H, s), 8.38 (1H, d, J ¼1.47Hz), 8.35
3
2
(
1H, d, J ¼8.07Hz), 8.00(1H, d, J ¼8.07Hz),7.55(1H, t, J ¼8.31Hz),
7
.01 (1H, dd, J ¼8.56, 0.73Hz), 6.83 (1H, d, J ¼7.58Hz), 6.58 (2H, s),
1
9
5
.38(2H,s),1.27(9H,s)ppm. FNMR(470MHz,(CD ) SO):δ58.54
3
2
13
(
s, 1F) ppm. C NMR (126MHz, (CD ) SO): δ 160.3, 156.9, 155.7,
3
2
1
1
1
1
55.3, 149.1, 146.2, 136.6, 133.8, 133.2, 132.7 (d, 2 JCF¼23.94Hz),
29.5, 120.9, 119.1, 103.8, 102.5, 80.4, 69.9, 28.0ppm. IR (neat): 3332,
749, 1612, 1571, 1510, 1437, 1421, 1365, 1250, 1213, 1143, 1103,
ꢁ
1
031, 994, 810, 778, 695, 665, 625, 592, 562, 544, 531cm . HRMS
+
ESI+) m/z: calcd for C H BrFN O S [M+H] : 527.0395; found:
(
20 20 4 5
527.0387.
2
.5. A microwave vial equipped with a stir bar was charged with tert-butyl
2-amino-5-((3-bromo-4-(fluorosulfonyl)benzyl)oxy)quinazolin-4-
yl)carbamate (8) (0.24g, 0.45mmol), tributyl(ethnyl)stannate (0.43g,
.37mmol), anhydrous toluene (4.0mL) and anhydrous DMSO
(
1
(
4.55mL). The reaction was degassed by sparging with argon gas for
0
1
0min. [1,1 -bis(di-tert-butylphosphino)ferrocene]dichloropalladium(II)
(0.059g, 0.09mmol) was added and reaction was heated in the microwave
at 110 °C for 45min. The reaction was diluted with EtOAc and filtered
throughcelitetwice.Thesolventwasconcentratedunderreducedpressure
to give crude tert-butyl (2-amino-5-((2-ethynyl-4-(fluorosulfonyl)-
benzyl)oxy)quinazolin-4-yl)carbamate which was used directly in the
next step.
2
.6. The residue was diluted with hydrochloric acid (4M in dioxane,
1.5mL, 6mmol) and stirred at room temperature for 1h. The residue
was diluted in DMSO (1mL) and purified by reverse-phase HPLC
using a Kinetex pentafluorophenyl column 21.2ꢀ100mm, 5μm;
Mobile phase A: 0.05% TFA in water (v/v); Mobile phase B: 0.05%
TFA in acetonitrile (v/v); gradient (75:25 to 65:35), flow 25mL/min
to give title compound 4-(((2,4-Diaminoquinazolin-5-yl)oxy)methyl)-
3
-ethynylbenzenesulfonyl fluoride 2,2,2-trifluoroacetate (SF-p1-yne,
1
Fig. 4) (12mg, 5%) as white solid. H NMR (500MHz, (CD ) SO):
3
2
δ 12.48 (1H, s, br, TFA), 8.97 (1H, s), 8.63 (1H, s), 8.20 (1H, d,
J ¼8.31Hz), 8.04 (1H, d, J ¼1.47Hz), 7.98–7.64 (2H, s, br), 7.83
(
1H, d, J ¼8.31Hz), 7.63(1H, t, J ¼8.31Hz), 6.98(1H, d, J ¼8.07Hz),
6
.91(1H,d,J ¼8.31Hz),6.54(0.25H,s,TFA),5.63(2H,s),5.01(1H,s),
1
9
2
.54 (s, DMSO) ppm. F NMR (470MHz, (CD ) SO): δ 60.52 (s, 1F),
3
2
13
ꢁ
75.33 (s, 3.74F, 1.25 equivalent of TFA) ppm. C NMR (126MHz,

ARTICLE IN PRESS
Quantifying drug-target engagement in live cells
13
(
CD ) SO):δ162.2,155.9,154.2,145.1,141.8,136.0,134.3,132.7(d,2
3
2
JCF¼22.68Hz), 130.7, 128.6, 121.2, 109.4, 107.11, 100.1, 90.6, 77.8,
1
6
7
8.3ppm. H NMR (500MHz, CD OD) δ 8.18 (1H, d, J ¼8.59Hz),
3
.99 (1H, s), 7.81 (1H, d, J ¼8.20Hz), 7.70 (1H, t, J ¼8.39Hz), 7.01
(
1
2H, t, J ¼8.20Hz), 5.62 (2H, s), 4.29 (1H, s), 2.55 (s, DMSO) ppm.
9
F NMR (470MHz, (CD OD): δ 58.20 (s, 1F), ꢁ76.95 (s, 3.74F,
3
13
1
1
1
.25 equivalent of TFA) ppm. C NMR (126MHz, CD OD) δ
3
64.7, 158.1, 156.3, 145.6, 143.1, 137.7, 135.9 (d, 2 JCF¼23.94Hz),
35.6, 132.0, 129.4, 124.0, 110.8, 108.9, 101.9, 89.6, 79.0, 70.7, 40.6
(
trace DMSO) ppm. IR (neat): 3291, 3105, 2361, 1682, 1613, 1401,
ꢁ
1
253, 1210, 835, 801, 722, 687, 579, 561, 542, 528cm . HRMS
1
+
ESI+) m/z: calcd for C H FN O S [M+H] : 373.0765; found:
(
17 13 4 3
373.0767.
5. Protocols
5
.1 Covalent labeling of recombinant DcpS
Now with the sulfonyl fluoride probe SF-p1 in-hand, we initially wanted
to confirm covalent labeling of recombinant DcpS in vitro.
3
.1. Incubate 0.5μM recombinant DcpS with 1.5μM SF-p1 in 50mM Tris
buffer with 1mM TCEP at room temperature. Take samples at 15, 30,
45, and 75min for intact mass measurement.
3
.2. Determine protein masses by LC-MS using an Agilent 1100 HPLC
system connected to a Waters LCT Premier XE mass spectrometer.
Solvents for HPLC: A, 0.1% formic acid in water; B, 0.1% formic acid
in acetonitrile. LC conditions: 0–1min, isocratic at 2% solvent B;
1
–22.5min, 2–75%B running on a Higgins Analytical PROTO 300
C4 column (5 μM particle: 100mmꢀ0.5mm: p/n RS-10M5-
W045) with a flow rate of 0.02mL/min. The retention time of the
DcpS protein is 17.5min.
3.3. Rate of adduct formation between DcpS and SF-p1 is plotted using
the MS intensity from the previous step using the following equation:
%
DcpSꢁprobe adduct ¼
adduct intensity=ðadduct intensity + apoDcpS intensityÞꢀ100%
As shown in Fig. 5A, SF-p1 achieves approximately 50% DcpS labeling
after 20min incubation.

ARTICLE IN PRESS
14
Lyn H. Jones et al.
A
B
C
Fig. 5 (A) Rate of DcpS-probe adduct formation detected by LC-MS. (B) Tryptic peptide
Tyr143-Leu144-Arg145 modified at Tyr143 (m/z 779.33). (C) MS/MS of the tryptic pep-
tide, m/z 604.2 resulting from benzyl ether cleavage.
3
.4. For peptide mapping experiments, incubate 0.15mL of 0.5μM
DcpS, that was labeled by 1.5μM SF-p1, with 0.60mL ethanol at
ꢁ
80°C for 48h. Then centrifuge at 14,000rpm for 10min. Remove
the supernatant, and dry the sample in a centrifugal concentrator.
Resuspend the dried sample in 0.02mL of 8M urea and 10mM
iodoacetamide in phosphate-buffered saline solution (PBS) and incu-
bate at room temperature for 30min. Following a 5-h incubation
with 0.5μg trypsin, samples are then analyzed on an LTQ Orbitrap
XL mass spectrometer. A unique tryptic peptide of DcpS Tyr143-
Leu144-Arg145 modified at Tyr143 was identified (Fig. 5B, m/z

ARTICLE IN PRESS
Quantifying drug-target engagement in live cells
15
7
79.33). MS/MS of this species supported the structural assignment
(
Fig. 5C, neutral loss of the diaminoquinazoline fragment resulting
from benzyl ether cleavage, m/z 604.2)
Tip: we skipped the standard DTT reduction step, since DcpS contains
no disulfide bonds and milder sample processing conditions are required
due to the chemical lability of the sulfonyl ester linkage.
These preliminary experiments confirmed that the location of the sulfo-
nyl fluoride electrophile within the DAQ template is appropriate for protein
labeling at the Tyr143 residue in the mRNA cap-binding site as predicted.
Having established sulfonyl fluoride probe engagement of the functional
site of the target enzyme, we next progressed the clickable, covalent
DAQ (SF-p1-yne) to live cell experiments, where the terminal alkyne moi-
ety enabled cycloaddition with biotin-azide and subsequent enrichment and
analysis of the probe-DcpS adduct.
5.2 Labeling of DcpS in live cells
4
.1. Thaw frozen human peripheral blood mononuclear cells (StemCell
Technologies), transfer to 10mL RPMI medium with 10% FBS,
and centrifuge at 300g for 5min. Following aspiration of the medium,
resuspend cells in 3mL 100% FBS and incubate at 37°C for 1h. Wash
cells with 30mL PBS, and then centrifuge at 300ꢀg for 5min.
Resuspend cells in 10mL RPMI with 0.1% FBS.
Tip: 25 million cells are required for a sufficient signal after enrich-
ment. In addition, cell clumps need to be removed by a BD Falcon
Cell Strainer (40 μm Nylon) prior to the probe or inhibitor treatment
(otherwise significant cell loss is observed).
4
4
.2. Seed 1mL cells per well in a 6-well plate with 25 million cells per sam-
ple. Add SF-p1-yne to the cells and incubate at room temperature for
2
1
0min. Then spin the cells at 1000ꢀg for 1min, and wash twice with
mL PBS to remove excess probe.
.3. Lyse cells with 50 μL lysis buffer (25mM Tris-HCl pH 7.4, 150mM
NaCl, 1mM EDTA, 1% NP-40 and 5% glycerol) with brief sonica-
tion to disrupt the nuclei. Remove cell debris by centrifugation for
1
min at 10,000ꢀg, and then transfer the supernatant to a 1.5mL
Eppendorf tube.
®
4
.4. Perform click chemistry with cell lysate using the Click-iT kit from
Life Technologies. Specifically, add 100 μL of 80 μM biotin-azide in
buffer A, briefly vortex, and then add 10 μL of 40mM CuSO , and
4

ARTICLE IN PRESS
16
Lyn H. Jones et al.
vortex. Add 10 μL additive 1 solution and vortex. After 2–3min, add
2
0 μL of additive 2, vortex, and then rotate end-over-end at room
temperature for 1h.
Tip: click reagents need to be stored at ꢁ20°C after reconstitution
in water.
4
4
.5. Add 500 μL of 6M urea made in the lysis buffer to the click reaction
mixture and then add 150 μL high capacity streptavidin agarose beads
Thermo Fisher) and incubate at room temperature for 1h. Centrifuge
(
at 1000ꢀg for 1min and wash three times with 1mL 3M urea in lysis
buffer. Elute proteins off the beads by adding 100 μL of 2XLDS sample
dye with reducing agent and heat to 95°C for 5min.
.6. Separate proteins on 4–12% Bis-Tris protein gels, transfer to a PVDF
membrane, and then detect DcpS by immunoblot (anti-DcpS anti-
body from Sigma HPA039632) using chemiluminescent SuperSignal
West Femto substrate (Thermo Fisher).
Tip: Femto substrate needs to be used here due to the relatively
weak DcpS signals following enrichment.
5
.3 Competition labeling and quantification of target
engagement
.1. Incubate human PBMCs with varying concentrations of RG3039
5
(
0.3nM–1μM) for 1.5h, followed by a 20min incubation with
1
μM SF-p1-yne. Then lyse the cells, perform click chemistry and
streptavidin enrichment as described in steps 4.3–4.6.
5
.2. Calculate DcpS occupancy signals by comparing DcpS signals from
RG3039 treated samples to the signal from the “no inhibitor” sample.
Fit the data to the following equation to obtain OC :
50
Y ¼ Bottom + ðTopꢁBottomÞ=ð1 + 10^ððLogOC ꢁXÞ∗Hill SlopeÞÞ
Where Y is the percentage of DcpS signals after normalizing to the “no
50
inhibitor” samples, X is the logarithm of the inhibitor concentration, the
Hill Slope describes the ”steepness” of the curve, ”Top” and ”Bottom”
are plateaus in the units of the Y axis, and OC is the concentration at which
50
the inhibitor achieves 50% occupancy of the protein target.
Western blot images and the occupancy curve for RG3039 are shown in
Fig. 6. RG3039 displayed an OC of 11nM for DcpS. This work validates
50
RG3039 as a potent ligand of the mRNA cap-binding site of DcpS in living
human primary cells.

ARTICLE IN PRESS
Quantifying drug-target engagement in live cells
17
A
B
Fig. 6 (A) Western blot images of dose-dependent RG3039 inhibition of DcpS enrich-
ment using SF-p1-yne in human PBMCs. (B) RG3039-DcpS occupancy curve in PBMCs.
6. Conclusions
Here, we have described a conceptual basis for target validation using
reactive chemical probes and exemplified a broadly applicable strategy to
assess drug-target engagement in living cells. We strongly urge the drug dis-
covery community to adopt these techniques to build the necessary confi-
dence in translational pharmacology required for successful target validation.
Sulfur(VI) fluoride exchange chemistry is finding a renaissance in medicinal
chemistry and chemical biology research that is underpinned by its magical
reactivity (aqueous stability combined with specific and context-dependent
amino acid engagement). The chemical probe-based technology detailed
here leverages the emerging innovations in sulfonyl fluoride chemical biol-
ogy and reports the first example of rationally targeting a specific tyrosine
residue in a protein through the development of a covalent inhibitor.
The DAQ inhibitors of DcpS, such as the clinical candidate RG3039,
were recently confirmed as lysosomotropic compounds, which complicates
the assessment of their pharmacology (Gopalsamy et al., 2017). The results
described here validate the annotated target as relevant to this chemotype in
human primary cells. Potent DcpS inhibitors with improved physicochem-
ical properties that lack lysosomal accumulation were shown to possess
in vivo efficacy in a mouse model of SMA, thus providing further confi-
dence in the proposed molecular mode of action (Cherry et al., 2017). How-
ever, more research is required to further elucidate the link between DcpS
inhibition and amelioration of disease-relevant phenotypes in the model.
Recently, DcpS was identified as a target of interest for the treatment of
acute myeloid leukemia (using RG3039 as a pharmacological inhibitor) and
the technology described here provides an opportunity to build PK-PD for
this indication (Yamauchi et al., 2018).

ARTICLE IN PRESS
18
Lyn H. Jones et al.
We expect functional chemical probes will continue to be widely used to
ensure that novel therapeutic targets are appropriately validated. However,
the approach should be used in conjunction with orthogonal technologies
such as the label-free cellular thermal shift assay (CETSA) to provide further
confidence in the results (Savitski et al., 2014). Importantly for the research
described here, CETSA was also used to measure DAQ-DcpS occupancy in
live cells and the values obtained were almost identical to those using the
sulfonyl fluoride-based method (Xu et al., 2016). Nevertheless, it should
be noted that although CETSA continues to make considerable technical
advances, a significant change in melting temperature upon ligand binding
is not always observed (false negatives), and membrane proteins are particu-
larly difficult proteins to assess. Chemical labeling methods directly measure
drug-target engagement for practically any protein of interest that reacts with
the probe. The functional chemical probe approach also enables binding site
identification using MS peptide mapping experiments, as highlighted here.
A related use of sulfonyl fluoride covalent chemical probes includes
the stabilization of difficult-to-crystallize proteins, resulting in useful struc-
tural information that informs rational drug design (Glukhova et al., 2017;
Liu et al., 2016). A future application of the approach described may include
target identification of small molecule hits in phenotypic screens when using
quantitative MS analytical techniques. One can foresee even wider applica-
tions of this chemistry in biomedical research, including bioconjugation
(such as the development of cleavable linkers to polyethylene glycol that
improve the PK of protein therapeutics) and molecular biology (to enhance
crosslinking-based sequencing technologies for instance). Sulfur(VI) fluoride
chemical biology will continue to drive a deeper understanding of molecular
pharmacology by augmenting the toolkit of protein labeling chemistry.
Acknowledgments
We thank Kieran Geoghegan, Lee Roberts, Erik Hett, Robert Kyne, Ariamala Gopalsamy,
Arjun Narayanan, Mihir Parikh and Ralph Robinson for their essential contributions to this
research and the development of sulfonyl fluoride chemical biology.
References
Ashoor, R., Yafawi, R., Jessen, B., & Lu, S. (2013). The contribution of lysosomotropism
to autophagy perturbation. PLoS One, 8(11), e82481. https://doi.org/10.1371/journal.
pone.0082481.
Bunnage, M. E., Chekler, E. L., & Jones, L. H. (2013). Target validation using chemical pro-
bes. Nature Chemical Biology, 9, 195–199.
Cherry, J. J., DiDonato, C. J., Androphy, E. J., Calo, A., Potter, K., Custer, S. K., et al.
(2017). In vitro and in vivo effects of 2,4 diaminoquinazoline inhibitors of the decapping

ARTICLE IN PRESS
Quantifying drug-target engagement in live cells
19
scavenger enzyme DcpS: Context-specific modulation of SMN transcript levels.
PLoS One, 12(9), e0185079. https://doi.org/10.1371/journal.pone.0185079.
Cravatt, B. F., Wright, A. T., & Kozarich, J. W. (2008). Activity-based protein profiling:
From enzyme chemistry to proteomic chemistry. Annual Review of Biochemistry, 77,
3
83–414. https://doi.org/10.1146/annurev.biochem.75.101304.124125.
Dong, J., Krasnova, L., Finn, M. G., & Sharpless, K. B. (2014). Sulfur(VI) fluoride exchange
SuFEx): Another good reaction for click chemistry. Angewandte Chemie (International
(
Edition in English), 53(36), 9430–9448. https://doi.org/10.1002/anie.201309399.
Drewes, G., & Knapp, S. (2018). Chemoproteomics and chemical probes for target discov-
ery. Trends in Biotechnology. https://doi.org/10.1016/j.tibtech.2018.06.008.
Garcera, A., Bahi, N., Periyakaruppiah, A., Arumugam, S., & Soler, R. M. (2013). Survival
motor neuron protein reduction deregulates autophagy in spinal cord motoneurons
in vitro. Cell Death & Disease, 4, e686. https://doi.org/10.1038/cddis.2013.209.
Glukhova, A., Thal, D. M., Nguyen, A. T., Vecchio, E. A., J o€ rg, M., Scammells, P. J., et al.
(
1
2017). Structure of the adenosine A. Cell, 168(5), 867–877.e813. https://doi.org/
0.1016/j.cell.2017.01.042.
Gopalsamy, A., Narayanan, A., Liu, S., Parikh, M. D., Kyne, R. E., Fadeyi, O., et al. (2017).
Design of potent mRNA decapping scavenger enzyme (DcpS) inhibitors with improved
physicochemical properties to investigate the mechanism of therapeutic benefit in spinal
muscular atrophy (SMA). Journal of Medicinal Chemistry, 60(7), 3094–3108. https://doi.
org/10.1021/acs.jmedchem.7b00124.
Hett, E. C., Xu, H., Geoghegan, K. F., Gopalsamy, A., Kyne, R. E., Menard, C. A., et al.
2015). Rational targeting of active-site tyrosine residues using sulfonyl fluoride probes.
(
ACS Chemical Biology, 10(4), 1094–1098. https://doi.org/10.1021/cb5009475.
Jarecki, J., Chen, X., Bernardino, A., Coovert, D. D., Whitney, M., Burghes, A., et al.
(
2005). Diverse small-molecule modulators of SMN expression found by high-
throughput compound screening: Early leads towards a therapeutic for spinal muscular
atrophy. Human Molecular Genetics, 14(14), 2003–2018. https://doi.org/10.1093/hmg/
ddi205.
Jones, L. H. (2015). Cell permeable affinity- and activity-based probes. Future Medicinal
Chemistry, 7(16), 2131–2141. https://doi.org/10.4155/fmc.15.100.
Jones, L. H. (2016). An industry perspective on drug target validation. Expert Opinion on Drug
Discovery, 11(7), 623–625. https://doi.org/10.1080/17460441.2016.1182484.
Kolb, H. C., Finn, M. G., & Sharpless, K. B. (2001). Click chemistry: Diverse chemical func-
tion from a few good reactions. Angewandte Chemie (International Edition in English),
4
0(11), 2004–2021.
Liu, S., Dakin, L. A., Xing, L., Withka, J. M., Sahasrabudhe, P. V., Li, W., et al. (2016).
Binding site elucidation and structure guided design of macrocyclic IL-17A antagonists.
Scientific Reports, 6, 30859. https://doi.org/10.1038/srep30859.
Liu, Q., Sabnis, Y., Zhao, Z., Zhang, T., Buhrlage, S. J., Jones, L. H., et al. (2013). Devel-
oping irreversible inhibitors of the protein kinase cysteinome. Chemistry & Biology, 20(2),
1
46–159. https://doi.org/10.1016/j.chembiol.2012.12.006.
Morgan, P., Van Der Graaf, P. H., Arrowsmith, J., Feltner, D. E., Drummond, K. S.,
Wegner, C. D., et al. (2012). Can the flow of medicines be improved? Fundamental
pharmacokinetic and pharmacological principles toward improving phase II survival. Drug
Discovery Today, 17(9–10), 419–424. https://doi.org/10.1016/j.drudis.2011.12.020.
Narayanan, A., & Jones, L. H. (2015). Sulfonyl fluorides as privileged warheads in chemical
biology. Chemical Science, 6, 2650–2659.
Rostovtsev, V. V., Green, L. G., Fokin, V. V., & Sharpless, K. B. (2002). A stepwise huisgen
cycloaddition process: Copper(I)-catalyzed regioselective “ligation” of azides and terminal
alkynes. Angewandte Chemie (International Edition in English), 41(14), 2596–2599. https://
doi.org/10.1002/1521-3773(20020715)41:14<2596::AID-ANIE2596>3.0.CO;2-4.

ARTICLE IN PRESS
20
Lyn H. Jones et al.
Sanman, L. E., & Bogyo, M. (2014). Activity-based profiling of proteases. Annual Review of
Biochemistry, 83, 249–273. https://doi.org/10.1146/annurev-biochem-060713-035352.
Savitski, M. M., Reinhard, F. B., Franken, H., Werner, T., Savitski, M. F., Eberhard, D.,
et al. (2014). Tracking cancer drugs in living cells by thermal profiling of the proteome.
Science, 346, 1255784.
Singh, J., Salcius, M., Liu, S. W., Staker, B. L., Mishra, R., Thurmond, J., et al. (2008). DcpS
as a therapeutic target for spinal muscular atrophy. ACS Chemical Biology, 3(11),
7
11–722. https://doi.org/10.1021/cb800120t.
Thurmond, J., Butchbach, M. E., Palomo, M., Pease, B., Rao, M., Bedell, L., et al. (2008).
Synthesis and biological evaluation of novel 2,4-diaminoquinazoline derivatives as
SMN2 promoter activators for the potential treatment of spinal muscular atrophy. Journal
of Medicinal Chemistry, 51(3), 449–469. https://doi.org/10.1021/jm061475p.
Tornøe, C. W., Christensen, C., & Meldal, M. (2002). Peptidotriazoles on solid phase:
[1,2,3]-triazoles by regiospecific copper(i)-catalyzed 1,3-dipolar cycloadditions of
terminal alkynes to azides. The Journal of Organic Chemistry, 67(9), 3057–3064.
Xu, H., Gopalsamy, A., Hett, E. C., Salter, S., Aulabaugh, A., Kyne, R. E., et al. (2016).
Cellular thermal shift and clickable chemical probe assays for the determination of
drug-target engagement in live cells. Organic & Biomolecular Chemistry, 14(26),
6
179–6183. https://doi.org/10.1039/c6ob01078d.
Yamauchi, T., Masuda, T., Canver, M. C., Seiler, M., Semba, Y., Shboul, M., et al. (2018).
Genome-wide CRISPR-Cas9 screen identifies leukemia-specific dependence on a pre-
mRNA metabolic pathway regulated by DCPS. Cancer Cell, 33(3), 386–400.e385.
https://doi.org/10.1016/j.ccell.2018.01.012.