Enhanced Suppression of a Protein-Protein Interaction in Cells Using Small-Molecule Covalent Inhibitors Based on an N-Acyl- N-alkyl Sulfonamide Warhead

Article abstract of DOI:10.1021/jacs.1c00703

Protein-protein interactions (PPIs) intimately govern various biological processes and disease states and therefore have been identified as attractive therapeutic targets for small-molecule drug discovery. However, the development of highly potent inhibitors for PPIs has proven to be extremely challenging with limited clinical success stories. Herein, we report irreversible inhibitors of the human double minute 2 (HDM2)/p53 PPI, which employ a reactive N-acyl-N-alkyl sulfonamide (NASA) group as a warhead. Mass-based analysis successfully revealed the kinetics of covalent inhibition and the modification sites on HDM2 to be the N-terminal α-amine and Tyr67, both rarely seen in traditional covalent inhibitors. Finally, we demonstrated prolonged p53-pathway activation and more effective induction of the p53-mediated cell death in comparison to a noncovalent inhibitor. This study highlights the potential of the NASA warhead as a versatile electrophile for the covalent inhibition of PPIs and opens new avenues for the rational design of potent covalent PPI inhibitors.

Full text of DOI:10.1021/jacs.1c00703

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
pubs.acs.org/JACS
Article
Enhanced Suppression of a Protein−Protein Interaction in Cells
Using Small-Molecule Covalent Inhibitors Based on an
N‑Acyl‑N‑alkyl Sulfonamide Warhead
Tsuyoshi Ueda, Tomonori Tamura,* Masaharu Kawano, Keiya Shiono, Fruzsina Hobor,
Andrew J. Wilson, and Itaru Hamachi*
Cite This: J. Am. Chem. Soc. 2021, 143, 4766−4774
Read Online
sı
*
Metrics & More
Article Recommendations
Supporting Information
ACCESS
ABSTRACT: Protein−protein interactions (PPIs) intimately govern
various biological processes and disease states and therefore have
been identified as attractive therapeutic targets for small-molecule
drug discovery. However, the development of highly potent
inhibitors for PPIs has proven to be extremely challenging with
limited clinical success stories. Herein, we report irreversible
inhibitors of the human double minute 2 (HDM2)/p53 PPI,
which employ a reactive N-acyl-N-alkyl sulfonamide (NASA) group
as a warhead. Mass-based analysis successfully revealed the kinetics of
covalent inhibition and the modification sites on HDM2 to be the N-terminal α-amine and Tyr67, both rarely seen in traditional
covalent inhibitors. Finally, we demonstrated prolonged p53-pathway activation and more effective induction of the p53-mediated
cell death in comparison to a noncovalent inhibitor. This study highlights the potential of the NASA warhead as a versatile
electrophile for the covalent inhibition of PPIs and opens new avenues for the rational design of potent covalent PPI inhibitors.
appropriate warheads that covalently target residues other than
cysteine are highly desirable for the extension of the TCI
approach to PPI inhibition.
INTRODUCTION
■
Protein−protein interactions (PPIs) are key to all signaling
pathways in cells. Thus, controlling PPIs with small molecules is
an attractive therapeutic strategy for a wide range of diseases.1,2
Although there are several successful examples such as Nutlin-
3a3 and its derivative Idasanutlin,4,5 reversible inhibitors of a
human double minute 2 (HDM2)/p53 PPI; the development of
highly potent inhibitors for PPIs has been recognized as one of
the most challenging tasks in drug discovery.1,2 This arises
because PPI interfaces are generally large (∼1000−2000 Å2 per
side) and shallow, making them less susceptible to high affinity
and selective competitive inhibition with small-molecule
ligands.1,6 To overcome the unfavorable thermodynamics in
blocking PPIs, the targeted covalent inhibition of PPIs has
recently been considered as a promising strategy.7 Targeted
covalent inhibitors (TCIs) contain weakly electrophilic war-
heads in the protein ligand scaffold.8 Upon ligand binding, TCIs
can typically react with an amino acid side chain of the target
protein as a consequence of the high effective molarity,
reinforcing the noncovalent affinity and resulting in irreversible
inhibition. Compared to noncovalent inhibitors, TCIs offer
several advantages, including increased potency and prolonged
duration of inhibition. To date, the design strategy for the
established TCIs has focused on targeting cysteine residues near
the ligand binding site because of its high reactivity toward
modest electrophiles, such as acrylamides and chloroaceta-
mides.8−13 Unfortunately, however, PPI interfaces rarely
contain free cysteine residues in an ideal position. Therefore,
Although still challenging, recent efforts identified several
promising warheads for the rapid and selective covalent
modification of noncysteine amino acids.9−11,14 For example,
aryl sulfonyl fluorides have been used for the covalent inhibition
of PPIs targeting lysine, tyrosine, or histidine residues.15 Our
group also recently demonstrated that ligand-tethered N-acyl-N-
alkyl sulfonamides (NASAs) efficiently react with the amino
group of a noncatalytic lysine residue of Hsp90 and are
applicable to the TCI approach.16−18 This encouraged us that
the NASA warhead would facilitate the development of effective
TCIs for PPIs exploiting distinct reactivity.
Here, we report NASA-based small-molecule covalent
inhibitors for the HDM2/p53 PPI, a promising therapeutic
target for several types of cancers (Figure 1a).3,19 While two
covalent inhibitors based on stapled peptides for the HDM2/
p53 PPI were recently reported, their modification sites,
mechanisms of action, and target selectivity in a cellular context
Received: January 20, 2021
Published: March 18, 2021
https://doi.org/10.1021/jacs.1c00703
J. Am. Chem. Soc. 2021, 143, 4766−4774
© 2021 American Chemical Society
4766

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Journal of the American Chemical Society
pubs.acs.org/JACS
Article
Figure 1. Irreversible inhibition of the HDM2/p53 PPI using NASA-appended Nutlin-3. (a) Schematic illustration of the reaction mechanism of the
NASA-based covalent inhibitor. (b) Molecular structures of NASA-based covalent inhibitors 1−3 and acrylamide-based compound 4.
Figure 2. Synthesis scheme of covalent HDM2 inhibitor 1.
were not fully evaluated.20,21 In this work, we clearly showed that
our optimal covalent inhibitor reacts with the N-terminal amino
group and Tyr67 of HDM2 with rapid kinetics and selectively
disrupts the HDM2/p53 interaction in an irreversible fashion in
cells.
and -3b, Figure S1),3 a first-in-class reversible HDM2 inhibitor,
as a scaffold for our covalent inhibitor because of its abundant
information on structure−activity relationships and sufficient
affinity (Kd = 0.263 × 10−6 M for the racemic mixture)22
according to a guideline for the ligand-directed covalent
modification of proteins with NASA chemistry.16,17 The
reported cocrystal structure of HDM217−108 with Nutlin-3a
(the more potent enantiomer) (protein data bank (PDB):
4HG7) indicated to us that a targetable lysine (Lys51) for a
RESULTS AND DISCUSSION
■
Design and Synthesis of Covalent Inhibitors for
HDM2. We employed ( )-Nutlin-3 (a mixture of Nutlin-3a
4767
https://doi.org/10.1021/jacs.1c00703
J. Am. Chem. Soc. 2021, 143, 4766−4774

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Journal of the American Chemical Society
pubs.acs.org/JACS
Article
Figure 3. In vitro characterization of covalent HDM2 inhibition by 1. (a) MALDI-TOF MS analysis of the covalent modification of the recombinant
○
N-terminal domain of HDM2 (HDM21−125). Reaction conditions: 1 μM HDM21−125, 3 μM 1, PBS buffer, pH 7.4, 37 °C. , native HDM21−125 (Mw =
●
14 955); , covalent adduct of the HDM21−125 with 1 (Mw = 15 703). (b) Kinetic analysis of the covalent modification of HDM21−125 with increasing
concentrations of 1. kapp was determined by the plots shown in Figure S4. Error bars represent the standard deviation, n = 3. (c) Superimposed model of
the HDM2 (17−108 a.a.) (gray)−Nutlin-3a (cyan) complex (PDB: 4HG7) with HDM2 (1−118 a.a.) (orange; PDB: 1Z1M). (d) Fluorescence
anisotropy changes in FL-labeled p53 peptide28 (100 nM) upon addition of reversibly or irreversibly inhibited HDM21−125 (M1G) (0.2 to 2000 nM).
⧫
■
●
Red , 1-treated HDM21−125 (M1G) after SEC; blue , ( )-Nutlin-3-treated HDM21−125 (M1G) after SEC; green , ( )-Nutlin-3-treated
▲
HDM21−125 (M1G) without SEC; , nontreated HDM21−125 (M1G). Error bars represent the standard deviation, n = 3.
NASA warhead is located at a distance of approximately 11 Å
from the 2-oxopiperazine moiety of Nutlin-3a (Figure S1).
Therefore, we designed covalent HDM2 inhibitor 1 appended
with a NASA group and an ∼12 Å linker (Figure 1b). The 2-
oxopiperazine moiety of ( )-Nutlin-3 was replaced with a
piperazine moiety to connect the NASA warhead. To examine
structure−activity relationships, we also prepared inhibitors 2
and 3 containing a shorter (∼6 Å) and a longer (∼17 Å) linker,
respectively, relative to 1 (Figure 1b). Acrylamide-type inhibitor
4 was also synthesized as a control compound that is mostly used
for targeting cysteine. The synthesis of 1 was carried out as
shown in Figure 2. The detailed procedures and synthesis of
other compounds are described in the Supporting Information
(SI). All of the final compounds were fully characterized by
NMR and high-resolution mass spectrometry.
Characterization of the Covalent Modification of
HDM2. We initially tested the covalent modification of
HDM2 with 1−4 in vitro. Recombinant HDM2 N-terminal
domain (1−125 a.a.) with an L33E mutation, for the
improvement of protein stability, was obtained by a standard
bacterial expression protocol (Figure S2). This protein (here-
after referred to as HDM21−125, 1 μM) was incubated with each
covalent inhibitor (3 μM) at 37 °C in PBS buffer (pH 7.4). The
matrix-assisted desorption/ionization time-of-flight mass spec-
trometry (MALDI-TOF MS) analysis of HDM21−125 treated
with 1 and 3 gave a new peak corresponding to the addition of
one molecule of the ligand moiety coupled with the loss of N-
cyanomethylbenzenesulfonamide, and 2 afforded negligible
conversion under this reaction condition (Figure 3a and Figure
S3). It was also found that control compound 4 failed to modify
the recombinant HDM21−125 (Figure S3). The time course
analysis showed that the reaction with 1 and 3 is completed
within 2 h, and the covalent modification was significantly
suppressed in the presence of ( )-Nutlin-3 as a competitive
ligand, indicating that the reaction was based on an affinity-
mediated proximity effect (Figure S3).
As for compound 1, we determined the affinity (Kd) and the
rate constant of covalent modification (kinact) using a previously
reported MS-based method (Figure 3b and Figure S4).16 The
obtained Kd value of 1 was 5.7 × 10−7 M, which indicated that
the incorporation of the linker and NASA warhead into the
parent scaffold of ( )-Nutlin-3 (Kd = 0.263 × 10−6 M for a
racemic mixture)22 slightly reduced the binding affinity. The
k
inact value of 1 was determined to be 4.8 × 10−4 s−1. Then, the
4768
https://doi.org/10.1021/jacs.1c00703
J. Am. Chem. Soc. 2021, 143, 4766−4774

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Journal of the American Chemical Society
pubs.acs.org/JACS
Article
Figure 4. Evaluation of p53 activation in cells by irreversible HDM2 inhibition after compound washout. (a) Illustration of the HDM2/p53 pathway.
(b) Western blotting analysis of p53 activation and up-regulation of HDM2 and p21. Cells were treated with ( )-Nutlin-3 or 1 for 1 h, followed by
washing with media and further incubation for 1.5 or 3 h. Normalized band intensities of (c) p53, (d) HDM2, and (e) p21. Error bars represent the
standard deviation, n = 3.
second-order rate constant kinact/Kd value for 1 was estimated to
be 8.4 × 102 M−1 s−1, indicating that compound 1 would possess
sufficient reactivity for effective covalent inhibition.
Lys51 is also positioned in proximity to the ligand, the lower pKa
value of the N-terminal α-amino group (pKa ≈ 8) compared to
the ε-amino group of a Lys side chain (pKa ≈ 10) may contribute
to the regioselectivity of the reaction with the NASA
warhead.25,26 The modification on Tyr67 is also reasonable,
given its distance from Nutlin-3a (∼10 Å) (Figure 3c) and our
recent study on the amino acid selectivity of the NASA
warhead.27
The reaction of NASA-tethered covalent inhibitors toward
the N-terminus and Tyr67 was also evaluated in live HEK293T
cells transiently expressing HDM21−125 or its variants with
probes 5−7 equipped with a clickable alkyne tag in the
( )-Nutlin-3 scaffold of 1−3, respectively. (See Figure S11 for
their molecular structures and reaction profiles with
HDM21−125.) The cell-based evaluation of modification sites
with 5 and 7 was consistent with in vitro experiments using 1 and
3, respectively, that is, these modified the N-terminus and Tyr67
(Figure S12). On the other hand, unlike the case of in vitro
reaction, short linker type compound 6 (an analog of 2) can
covalently bind to the HDM2 N-terminal domain in the live cell
context, and the modification site is exclusively the N-terminus.
(See the Figure S12 caption for a more detailed discussion.)
Although these results suggest that 2 and 3 are also promising for
the covalent inhibition of HDM2, they showed serious p53-
independent cytotoxicity (Figure S13) while 1 did not. Since
this study focuses on the development of covalent inhibitors
capable of blocking HDM2−p53 interaction and the evaluation
of how efficient these are for selectively activating the p53
The modification sites on HDM21−125 after reaction with
NASA-based covalent inhibitors were identified by MS-based
peptide mapping analysis (Figure S5). Although 1 was designed
to target Lys51, to our surprise the modification reaction
occurred on the α-amino group of the N-terminus and Tyr67 of
HDM21−125 (Figure S6). The N-terminal modification was also
verified by top-down sequencing of HDM21−125 with in-source
decay MALDI-TOF/TOF MS analysis (Figure S8). We also
confirmed that 3 reacts with the N-terminal amine and Tyr67 as
well as with 1 (Figures S5, S7, and S8). A Y67F mutation in
HDM21−125 resulted in the disappearance of the corresponding
Y67-labeled peptide with the remainder of the N-terminally
labeled peptide in the peptide mapping analysis (Figures S9 and
S10), which clearly indicated the modification of Tyr67.23
The N-terminal region of HDM2 is an intrinsically disordered
region that modulates the function of HDM2, and it was
truncated in the cocrystal structural analysis of the HDM2 (17−
108 a.a.)/Nutlin-3 complex.19 Indeed, the N-terminal region of
HDM2 has rarely been observed24 in cocrystal structures, and
the first five residues were never crystallographically resolved.19
We thus made a superimposed model of the HDM2 (17−108
a.a.)/Nutlin-3a complex (PDB: 4HG7) with HDM2 (1−118
a.a.) (PDB: 1Z1M) determined by an NMR study and
confirmed that the N-terminal amino group is likely located in
the vicinity of Nutlin-3a (within ∼12 Å) (Figure 3c). While
4769
https://doi.org/10.1021/jacs.1c00703
J. Am. Chem. Soc. 2021, 143, 4766−4774