From Inhibition to Degradation: Targeting the Antiapoptotic Protein Myeloid Cell Leukemia 1 (MCL1)

Article abstract of DOI:10.1021/acs.jmedchem.9b00455

Protein-protein interactions (PPIs) have emerged as significant targets for therapeutic development, owing to their critical nature in diverse biological processes. An ideal PPI-based target is the protein myeloid cell leukemia 1 (MCL1), a critical prosurvival factor in cancers such as multiple myeloma where MCL1 levels directly correlate to disease progression. Current strategies for halting the antiapoptotic properties of MCL1 revolve around inhibiting its sequestration of proapoptotic factors. Existing inhibitors disrupt endogenous regulatory proteins; however, this strategy actually leads to an increase of MCL1 protein levels. Here, we show the development of hetero-bifunctional small molecules capable of selectively targeting MCL1 using a proteolysis targeting chimera (PROTAC) methodology leading to successful degradation. We have confirmed the involvement of the E3 ligase CUL4A-DDB1 cereblon ubiquitination pathway, making these PROTACs a first step toward a new class of antiapoptotic B-cell lymphoma 2 family protein degraders.

Full text of DOI:10.1021/acs.jmedchem.9b00455

Subscriber access provided by UNIV AUTONOMA DE COAHUILA UADEC
Article
From Inhibition to Degradation: Targeting the Anti-
apoptotic Protein Myeloid Cell Leukemia 1 (MCL1)
James William Papatzimas, Evgueni Gorobets, Ranjan Maity, Mir Ishruna Muniyat,
Justin L. MacCallum, Paola Neri, Nizar Jacques Bahlis, and Darren Jason Derksen
J. Med. Chem., Just Accepted Manuscript • Publication Date (Web): 22 May 2019
Downloaded from http://pubs.acs.org on May 25, 2019
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a service to the research community to expedite the dissemination
of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in
full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully
peer reviewed, but should not be considered the official version of record. They are citable by the
Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore,
the “Just Accepted” Web site may not include all articles that will be published in the journal. After
a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web
site and published as an ASAP article. Note that technical editing may introduce minor changes
to the manuscript text and/or graphics which could affect content, and all legal disclaimers and
ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or
consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W.,
Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
From Inhibition to Degradation: Targeting the Anti-
apoptotic Protein Myeloid Cell Leukemia 1 (MCL1)
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
James W. Papatzimas, ‡† § Evgueni Gorobets, ‡†§ Ranjan Maity, ‡§ Mir Ishruna Muniyat,║ Justin
L. MacCallum, Paola Neri, Nizar J. Bahlis, Darren J. Derksen.* †§‡
║
§
* §‡
†
Department of Chemistry, University of Calgary, 2500 University Dr. NW, T2N 1N4, Calgary,
Alberta, Canada.
§
Arnie Charbonneau Cancer Institute, University of Calgary, 3280 Hospital Dr. NW, T2N 4Z6,
Calgary, Alberta, Canada.
║
Department of Chemistry and Centre for Molecular Simulation, University of Calgary, 2500
University Dr. NW, T2N 1N4, Calgary, Alberta, Canada.
ABSTRACT
Protein-protein interactions (PPIs) have emerged as significant targets for therapeutic
development, owing to their critical nature in diverse biological processes. An ideal PPI-based
target is the protein myeloid cell leukemia 1 (MCL1), a critical pro-survival factor in cancers such
as multiple myeloma where MCL1 levels directly correlate to disease progression. Current
strategies for halting the anti-apoptotic properties of MCL1 revolve around inhibiting its
ACS Paragon Plus Environment
1

Journal of Medicinal Chemistry
Page 2 of 71
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
sequestration of pro-apoptotic factors. Existing inhibitors disrupt endogenous regulatory proteins,
however this strategy actually leads to an increase of MCL1 protein levels. Here we show the
development of heterobifunctional small molecules capable of selectively targeting MCL1 using
a Proteolysis Targeting Chimera (PROTAC) methodology leading to successful degradation. We
have confirmed the involvement of the E3 ligase CUL4A-DDB1 cereblon (CRBN) ubiquitination
pathway, making these PROTACs a first step toward a new class of anti-apoptotic BCL-2 family
protein degraders.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
INTRODUCTION
Myeloid cell leukemia 1 (MCL1) is a pro-survival protein overexpressed in a variety of
1
,2
different cancers and is of tremendous therapeutic interest. MCL1 is involved in complex
3
protein-protein interactions (PPIs) involving pro-apoptotic factors Bim, Bak, and Bax. These
anti-apoptotic interactions prevent the activation of caspase cascades, promoting cell survival.
Due to this anti-apoptotic nature, MCL1 has been recognized as a vital survival factor in human
cancers such as lymphoma, leukemia, breast cancer, and multiple myeloma (MM) – where levels
4
of MCL1 directly correlate to disease progression. The ability of MCL1 to silence apoptotic
pathways allows it to circumvent the typical clearance mechanisms of cells, and is therefore
often overexpressed by tumor cells to gain a survival advantage.3
MCL1 is a challenging drug target as it is amongst the 85% of proteins in the human genome
which have been deemed ‘undruggable’ – a term used to describe protein targets which are
5
,6
currently chemically intractable, making them onerous to drug. Despite this label, there have
been some successes, and MCL1 inhibitors have been developed with a variety of structural
ACS Paragon Plus Environment
2

Page 3 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
geometries, including peptides, macrocycles, and boronic acids, although the majority of
progress has been experienced with small molecule leads.7-10 These compounds modulate MCL1
through competitive inhibition of interactions with its pro-apoptotic targets by disrupting “hot
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
spots” of the PPI interfaces.11
PPIs are difficult to target with small molecule therapeutics due to their shallow and relatively
1
2
featureless binding regions. The majority of compounds that target MCL1 either occupy the flat
binding site of the BH3 groove,7,8,12,13 an allosteric binding site, or indirectly influence MCL1
14
via downstream effects through tyrosine kinase and HDAC inhibition15-16 or CDK9
1
7
degradation. Nonetheless, MCL1 has gained much attention as a target for anti-cancer
therapeutics,1,2,18 and several organizations have acute programs targeting MCL1 with
compounds AMG 176, AMG 397 (Amgen), AZD 5991 (Astra Zeneca), and S 64315 (Servier)
currently undergoing clinical trials.7,8,12,13,19-23
One inherent pitfall of inhibition techniques is the dependence on the on/off rates of these
compounds within binding pockets. Inhibitors require very small dissociation constants and slow
1
1
metabolic clearance to ensure maximum efficacy. These challenges make discovering
effectively potent inhibitors increasingly difficult for new, unexplored drug targets. A new
paradigm in the field of drug discovery aims to perturb PPIs via proximity mediated
24
manipulation. The literature has seen a recent surge in applications towards selective protein
degradation, especially since seminal reports utilizing Proteolysis Targeting Chimera (PROTAC)
technology.25-29 Numerous pharmaceutical companies now have protein degrader programs, with
some compounds beginning to enter clinical trials. 30-31
ACS Paragon Plus Environment
3

Journal of Medicinal Chemistry
Page 4 of 71
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
PROTACs are small molecule conjugates which tether target proteins and E3 ligases through
heterobifunctional poles, inducing ubiquitination and labeling proteins for proteasomal
degradation (Figure 1).24,26,32 The intermolecular recruitment of E3 ligases facilitates a gain-of-
function response in the local environment by covalently linking ubiquitin subunits to lysine
residues of the target protein. Examples of suitable E3 ligases are the von Hippel-Lindau (VHL)
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3
3
34
protein with ligands developed by the Crews lab, as well as cereblon (CRBN) which is
targeted by thalidomide and related analogues.35
Figure 1. Ubiquitin-Proteasome Pathway (UPP) schematic. (i) PROTAC poles coordinate the
protein target and an E3 ubiquitin ligase to form a ternary complex. (ii) E3 ligase transfers
ubiquitin onto the protein target. (iii) Ubiquitination occurs several times, polyubiquitinating the
target protein. (iv) Ternary complex dissociates and PROTAC is recycled. Polyubiquitinated
protein undergoes degradation via 26S proteasomes.
ACS Paragon Plus Environment
4

Page 5 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
According to reports in the literature, PROTAC-mediated degradation of proteins is beneficial
over direct inhibition as it only requires the ternary complex to exist long enough to ubiquitinate
3
6
the target protein. Upon ubiquitination of its target, the PROTAC and E3 ligase can then
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
dissociate from the ternary complex and be recycled to take part in future ubiquitination events,
3
7
enabling catalytic degradation cycles. Herein, we demonstrate the development of novel
PROTACs capable of inducing degradation of the anti-apoptotic protein MCL1, through a
proteasome-mediated pathway.
RESULTS AND DISCUSSION
The MCL1-binding pole of the PROTACs was motivated by the MCL1 inhibitor A-1210477
7
(Figure 2), which has an inhibition constant (K
i
) of 0.45 nM. The CRBN-binding portion was
inspired by thalidomide analogs which bind to CRBN via the thalidomide binding domain
TBD).35,38 CRBN is capable of recruiting DDB1, forming the requisite complex for
(
ubiquitination to occur. Notably, thalidomide, lenalidomide, and pomalidomide all recruit
CRBN, and are existing clinical therapies for MM. These immunomodulatory imide drugs
(
IMiDs) have similar affinities for CRBN, with K
D
values of 250 nM, 178 nM, and 157 nM
respectively.39
The linker length and pole composition of PROTACs affect characteristics such as cell
40
41
42
permeability, solubility, and physical distance. The distance between targeting molecules is
vital for effective recruitment and orientation of target proteins and E3 ligases. We chose 4-
hydroxythalidomide as the CRBN pole and synthesized biotinylated affinity probes in order to
confirm whether this would still be an appropriate CRBN-targeting ligand once conjugated to
various linkers. In order to facilitate rapid syntheses of numerous biotin probes we employed
ACS Paragon Plus Environment
5

Journal of Medicinal Chemistry
Page 6 of 71
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
pentafluorophenyl (Pfp) esters of biotin. The 4-hydroxythalidomide motif was conjugated to
biotin through various terminal diamine linkers, and these probes were utilized in
immunoprecipitation (IP) experiments and visualized for CRBN binding. These affinity
experiments allow for the qualitative visualization of binary target engagement and are a
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
4
3
powerful diagnostic tool for linker length validation. Binding to endogenous CRBN was
accommodated when the hydroxythalidomide pole was conjugated to biotin through the
polyethylene glycol (PEG) linker 4,7,10-trioxa-1,13-tridecanediamine (Supplementary Figure 1).
A recent report showed that the truncated glutarimide portion of IMiDs is sufficient for CRBN
3
5
binding to occur. We applied this biotinylated affinity probe strategy to glutarimide to
determine whether this would be a suitable, minimalized CRBN pole. Two glutarimide-biotin
conjugates were synthesized (2-3) to study which would afford better binding. In our IP
experiments, neither glutarimide probe expressed effective binding, suggesting that the entire
thalidomide ligand is required to accommodate potential linkers required for PROTAC synthesis.
We investigated MCL1 ligand binding using the same biotinylation methodology. Previously
reported SAR studies suggested that certain substitutions at the aryl piperazine of A-1210477
7
could still afford binding to the BH3 groove of MCL1. Our IP experiments revealed that
conjugation of biotin through the aryl piperazine (8)yielded modest MCL1 binding (Figure 2b).
Substitution of the morpholine moiety to create an isosteric, piperazine-linked biotin probe (17)
significantly improved MCL1 binding (Figure 2b). This interaction provided evidence that this
newly implemented motif occupied an appropriate exit vector for conjugating linkers. It is
noteworthy that N-alkylation of the indole twists the carboxylate in the 2-position out of the
plane of the indole ring, thus improving the hydrogen bond interaction of the ligand with Arg
4
4
263 of MCL1, and affording more efficient binding within the BH3 groove. Surface plasmon
ACS Paragon Plus Environment
6

Page 7 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
resonance (SPR) studies were performed using His-tagged human MCL1 and A-1210477 or
dMCL1-2 to determine if the structural changes between the two compounds compromised
MCL1 binding. The inhibitor A-1210477 was found to have a K
D
of 19 nM, whereas dMCL1-2
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
had a K of 30 nM (Supplementary Table 1). As the K of the MCL1 degrader does not differ
D
D
significantly from that of the inhibitor, it validates the position of indole substitution and the
piperazine moiety as a suitable linker to occupy an effective exit vector from the BH3 groove
without affecting MCL1 binding.
Figure 2. Biotinylation affinity studies for MCL1 ligands. (a) Chemical structures of MCL1
inhibitor A-1210477 (AbbVie) and biotinylated affinity probes 8 and 17. (b) Streptavidin
immunoprecipitation with 8 and 17 followed by immunoblot for MCL1 to visualize the
formation of the streptavidin-MCL1 complex. The input shows endogenous MCL1 long and
short isoform expression in OPM2 cell lysate.
ACS Paragon Plus Environment
7

Journal of Medicinal Chemistry
Page 8 of 71
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
A published docking study demonstrated that the free carboxylic acid at the 2-position of the
7
indole of A-1210477 was responsible for anchoring the molecule into the BH3 groove. To
confirm this claim, and due to synthetic accessibility, our initial PROTAC had the CRBN pole
coupled to A-1210477 through a diaminobutane linker at the carboxylic acid (dMCL1-3). Our
experimental findings corroborated the literature results as no MCL1 degradation was observed.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 3. Chemical structures of MCL1 degraders synthesized in this work and MCL1 inhibitor
A-1210477 (AbbVie).
ACS Paragon Plus Environment
8

Page 9 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Our PROTAC iterations which were linked directly through the aryl piperazine (dMCL1-
/5/6, Figure 3) provided underwhelming degradation. In vitro degradation studies in the MM
4
cell line OPM2 revealed that conjugates linked through the aryl piperazine ring accommodated
MCL1 degradation, but with modest to poor efficacy. Our results indicated that coupling the
CRBN pole through this locale made the ligand sterically inaccessible for effective binding
within the BH3 groove, consistent with our biotin-probe results (Figure 2). Due to the aryl
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
7
piperazine ring’s contribution to the subnanomolar MCL1 affinity, this motif was maintained
for the multi-gram scale intermediate synthesis required for extending our own SAR study
(
Scheme 1).
Influenced by our previous results, we next investigated substitution of the morpholine ring of
A-1210477 into a point of conjugation. Crystallographic data shows that the morpholine ring
extends out of the BH3 groove into the solvent space, providing a greater effect on solubility
7
than MCL1 binding affinity. Coupling this information with the knowledge of the other
suboptimal conjugation sites, we redesigned our MCL1 ligand synthesis. The terminal sulfamide
moiety was maintained, and installed prior to a convergent Suzuki cross coupling, establishing
the main skeleton of the ligand (Scheme 1). Indole N-alkylation was performed with a benzyl
protected piperazine derivative, affording an appropriate linker to occupy an effective exit vector
from the BH3 groove. This installed a protected secondary amine protruding from the core of the
BH3 groove, available for subsequent couplings upon deprotection. The benzyl group was
liberated through hydrogenation, followed by saponification of the ethyl ester with lithium
hydroxide to provide 15. While Wurz and coworkers use click-chemistry to rapidly form
4
5
bispecific conjugates, we utilized a pentafluorophenyl activated ester of 4-hydroxythalidomide,
to form the requisite amide bonds. Notably, this advanced coupling intermediate was simply
ACS Paragon Plus Environment
9

Journal of Medicinal Chemistry
Page 10 of 71
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
purified via trituration with diethyl ether and was rapidly prepared for similar final coupling
steps. This Pfp coupling approach was used to conjugate the MCL1 ligand to the CRBN pole
4
6
through a known PEG extension (dMCL1-1), as well as directly to the newly installed
piperazine linker (dMCL1-2) (Figure 3). The distance between the targeting poles was varied to
examine the role of linker length on ternary complex formation, ubiquitination efficacy, and
ultimately degradation properties.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Scheme 1. Synthetic routes for MCL1 biotinylated affinity probes and PROTACs.a
aReagents and conditions: (i) Pd(dppf)Cl
chloroethyl)piperazine dichloride, Cs CO
, Cs
CO
, DMF, 82%; (ii) 1-benzyl-4-(2-
, Pd/C, DMF, quantitative; (iv)
2
2
3
2
3
, DMF, 42%; (iii) H
2
LiOH, H
2
O, MeOH, 99%; (v) pentafluorophenyl ester of R, DIPEA, DMF, 62-76%.
ACS Paragon Plus Environment
1
0

Page 11 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
To elucidate the effect physical distance and linker conformation have on ternary complex
formation, dMCL1-2 was employed in IP experiments using purified His-tagged MCL1.
Formation of the ternary complex [CRBN-dMCL1-2-MCL1] was confirmed by IP, followed by
an immunoblot for CRBN (Figure 4a). Upon immunoprecipitation of MCL1-His, CRBN was
only successfully recruited when cells were treated with dMCL1-2, implying that this conjugate
is capable of bringing MCL1 and CRBN within close proximity.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
In order to gain an understanding of possible structures of this key ternary complex, we
employed atomistic simulations using the Modeling Employing Limited Data (MELD)
4
7
approach. We simulated the formation of the ternary complex of CRBN and MCL1 with either
dMCL1-1 or dMCL1-2. We added restraints to the simulations in order to: (1) keep the
structures of CRBN and MCL1 close to their respective X-ray crystallographic structures; and
(2) guide the ligand poles to their respective binding pockets on CRBN or MCL1, ensuring that
7
the bound conformations are identical to previous crystallographic studies, so that the major
factors being compared are the conformation of the linker and the relative orientation of CRBN
and MCL1. MELD replica exchange simulations were carried out in triplicate for 200 ns
(Supplementary Information). Analysis of our simulations revealed several trends consistent
across replicates. Both dMCL1-1 and dMCL1-2 are found to bridge between CRBN and MCL1,
bringing them into close proximity (Figure 4b). For both ligands, there is substantial flexibility in
the relative orientation between CRBN and MCL1 (Figure 4c). Additionally, the ligand poles are
deeply bound within their respective binding sites and there is extensive contact between CRBN
and MCL1 (Figure 4d). Although these trends are similar for both ligands, we observed a
number of differences between the simulations with dMCL1-1 and dMCL1-2.
ACS Paragon Plus Environment
1
1

Journal of Medicinal Chemistry
Page 12 of 71
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 4. Ternary complex formation and overview of MELD simulation results. (a)
Immunoblot for CRBN after a 4 hour pre-treatment of MCL1-His with DMSO or dMCL1-2,
subsequent exposure to OPM2 cell lysate, followed by IP of MCL1-His and bound ternary
complexes. (b) PROTAC dMCL1-2 (orange) binds to MCL1 (white) and CRBN (green),
bridging the two proteins in close proximity. (c) Sample of spatial orientations of CRBN (yellow,
light green, dark green) relative to MCL1 (white). (d) PROTAC dMCL1-2 is deeply bound to
both the BH3 groove of MCL1 (white) and the thalidomide binding domain of CRBN (green).
PDB ID entries 4tzc was used for CRBN, and 5vkc was used for MCL1.
We investigated the degree of ligand extension in our simulations by measuring the distance
between select atoms on each of the PROTAC poles (CRBN pole: glutarimide nitrogen atom,
MCL1 pole: C5 on the naphthalene ring). In the dMCL1-1 ternary complex, the ligand extension
varies from 24–33 Å with a mean distance of 28 Å, whereas the dMCL1-2 complex varies
between from 19–26 Å with a mean distance of 23 Å. Although the larger extension of dMCL1-
1
is consistent with the longer linker, the difference of 5 Å is considerably less than expected if
the linker structures were fully extended (Figure 3) due to the tendency of dMCL1-1 to coil
between the two proteins (Supplementary Figure 4). Previous studies have also observed that
shorter linkers can promote more efficient degradation by maximizing interactions between
ACS Paragon Plus Environment
1
2

Page 13 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
29
protein targets and E3 ligases. Our results are consistent with previous findings that linker
length plays a significant role in degradation efficiency.48
There is substantial flexibility in the relative orientation between CRBN and MCL1 (Figure 5).
We observe large variations in the twist and tilt angles for both linkers, with somewhat more
variability for dMCL1-1 compared to dMCL1-2 (particularly for the tilt angle), consistent with
the longer linker. Despite some overlap, the orientational distributions for dMCL1-1 and
dMCL1-2 are different, which would potentially present different faces of MCL1 towards the
ubiquitin ligase, in turn affecting ubiquitylation kinetics that may explain the difference in
efficacy between the two ligands. A recent study by Nowak and co-workers successfully used
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
orientational preferences to design a PROTAC specific for the Bromodomain and Extra Terminal
49
(
BET) domain of BRD4 over the homologous domains from BRD2 and BRD3. In that same
work, the authors also report several crystal structures of unusually low resolution (ca. 4–6 Å)
for some linker-target combinations, which might be consistent with substantial conformational
heterogeneity.
ACS Paragon Plus Environment
1
3

Journal of Medicinal Chemistry
Page 14 of 71
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 5. Graphical representation of relative orientations of CRBN with respect to MCL1
observed for either dMCL1-1 (blue) or dMCL1-2 (orange) in MELD binding simulations.
Vectors were defined between the center of mass of each protein and a selected amino acid on
the surface for CRBN (green arrows) and MCL1 (grey arrows). The distribution of twist (top)
and tilt (bottom) angles were visualized using kernel density estimation using a von Mises
circular kernel. PDB ID entry 4tzc was used for CRBN, and 5vkc was used for MCL1.
Taken together, these observations indicate that structural heterogeneity and geometry may
affect the ubiquitination and degradation of MCL1, although detailed mechanistic understanding
must be improved before general predictive models suitable for rational design can be made.
ACS Paragon Plus Environment
1
4

Page 15 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Consistent with previous work, these results underscore the sensitivity of binding, ubiquitination,
and degradation to the nature of the linker and its site of attachment.29,50
To observe the degree of MCL1 polyubiquitination upon formation of the ternary complex, we
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
performed in vitro ubiquitination assays of OPM2 cell lysate incubated with dMCL1-1 or
5
1
dMCL1-2 (Figure 6a). MCL1 contains 5 lysine residues which participate in ubiquitination. E3
ligases transfer ubiquitin subunits to residues 5/40/136/194/197K, forming various polyubiquitin
chains which allow MCL1 to be recognized by the 26S proteasome for degradation.
Ubiquitinylation assays were performed using a commercial kit (Abcam) to confirm
ubiquitination via CRBN recruitment by visualizing MCL1 ubiquitin units in the immunoblots.
These assays suggest that both linker lengths accommodate MCL1 ubiquitination, but to
differing degrees. The shorter distance between poles of dMCL1-2 yielded a greater extent of
MCL1 ubiquitination versus the extended spacer of dMCL1-1 (Figure 6a).
These finding support that modifications in linker length and composition affect ubiquitination
efficacy. These results also provided evidence that our optimized conjugation site was suitable
for positioning MCL1 within an appropriate proximity to CRBN for ubiquitination to occur, as
predicted by our biotinylation experiments. We employed the established PROTAC dBET1,
2
8
developed by Bradner for the degradation of BRD4, as a positive control for visualization of
ubiquitination. The level of MCL1 ubiquitination observed with dMCL1-1 and dMCL1-2
(Figure 6a) was qualitatively comparable to the ubiquitination of BRD4 by dBET1 (Figure 6b).
Upon confirming ternary complex formation and MCL1 ubiquitination, degradation
capabilities of dMCL1-2 were assessed in human MM OPM2 cells. Marked decreases in MCL1
levels were observed at 100 nM in OPM2 cells compared to DMSO controls, providing evidence
ACS Paragon Plus Environment
1
5

Journal of Medicinal Chemistry
Page 16 of 71
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
of successful UPP mediated degradation in vitro (Figure 6c). Endogenous MCL1 degradation
was reduced at treatments of 1 μM, compared to 100-500 nM. This observation is a characteristic
pharmacological property of PROTACs known as the ‘hook effect’ where higher concentrations
of bifunctional degraders prevent degradation as binary complex formation outcompetes against
ternary complex formation due to saturation of protein binding sites by the large number of
available ligands.24
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 6. Ubiquitination and in vitro degradation studies of MCL1 degraders. (a)
Immunoprecipitation (IP) of MCL1 from in vitro ubiquitinylation assays with immunoblotting for
ubiquitin. (b) IP of BRD4 from in vitro ubiquitinylation assays and immunoblotting for ubiquitin.
(c) Immunoblot for MCL1 after 48 hours of in vitro degradation conditions with DMSO or
dMCL1-2 in OPM2 cells. Degradation is reported as a percentage of protein abundance relative
to the DMSO control.
We next investigated the mechanistic dependency on proteasome function and CRBN using
biochemical and gene editing techniques, respectively (Figure 7). First, we rescued dMCL1-2
induced MCL1 degradation using the FDA approved proteasome inhibitor bortezomib. The
ACS Paragon Plus Environment
1
6

Page 17 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
increased MCL1 levels in the combination treatment of cells with dMCL1-2 and Bortezomib
confirm the necessity for proteasome activity in this mode of action (Figure 7a). It is noteworthy
that the loss of proteasome activity prevents natural MCL1 ubiquitination events from occurring
via the endogenous MCL1 ubiquitin ligase E3 (Mule), thereby increasing baseline MCL1
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
5
2
levels. To definitively establish the CRBN dependence of this degradation mechanism, a
CRBN knockout cell line, OPM2CRBN-/-, was generated using CRISPR Cas9 gene editing and this
CRBN deficient MM cell line was treated with A-1210477 or dMCL1-2. No degradation of
MCL1 was observed with A-1210477 independent of cell line (Figure 7c), consistent with its
known inhibitory mode of action. Instead, A-1210477 binds tightly in the BH3 pocket,
competitively inhibiting Mule and resulting in increased MCL1 levels in vitro (lanes 2-5 and 11-
53
1
4), a prevalent pitfall of A-1210477. Treatment of OPM2WT cells with dMCL1-2 afforded
MCL1 degradation at nanomolar concentrations (lanes 6-9). Considering dMCL1-2 also
occupies the BH3 groove, this degrader will displace Mule from the BH3 groove as well,
preventing some level of natural ubiquitination from occurring. Therefore, there is marked
degradation of both isoforms of MCL1 with dMCL1-2 relative to the stabilization of MCL1 that
occurs from occupation of the BH3 groove, as is seen in lanes 2-5.
ACS Paragon Plus Environment
1
7

Journal of Medicinal Chemistry
Page 18 of 71
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 7. Cellular degradation assays of dMCL1-2. (a) Immunoblot for MCL1 after 3 hours of
treatment with DMSO, dMCL1-2, and/or bortezomib in OPM2 cells. (b) Chemical structure of
dMCL1-2. (c) Biochemical analysis of dMCL1-2 mediated MCL1 degradation. Immunoblot for
MCL1 long and short isoforms after 48 hours of treatment with DMSO, dMCL1-2, or A-
210477 in OPM2WT and OPM2CRBN-/- cells.
1
In contrast, the same treatment of OPM2CRBN-/- cells with dMCL1-2 was ineffective (Figure
7c). MCL1 levels experienced insignificant changes in abundance when CRBN is absent (lanes
5-18). Consistent with our proposed mode of action, OPM2CRBN-/- cells no longer possess the
1
capability to form the CUL4A-DDB1 complex and ubiquitinate MCL1, preventing any
degradation from occurring, and supporting the UPP degradation mechanism of dMCL1-2. This
CRBN knockout study is consistent with degradation events being selective for a CRBN
-/-
mediated mode of action. The persistence of MCL1 in our CRBN cell lines confirms that
ACS Paragon Plus Environment
1
8

Page 19 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
MCL1 degradation is CRBN dependent. Coupling our studies with the work previously reported
7
on the MCL1 ligand showing selective affinity for MCL1 over other BCL2 family members as
well as multiple kinases and GPCRs, strongly suggests that the MCL1 degradation observed with
our compounds is not due to biochemical downstream events rooting from off-target inhibition
of other targets such as epigenetic regulators or tyrosine kinases. Even whilst MCL1 was not
degraded in OPM2CRBN-/- cells, some cell death still occurred (Supplementary Figure 6),
consistent with the activity of A-1210477 alone. Moreover, dMCL1-2 induces apoptosis at 250
and 500 nM after a 24 hour treatment with 1% FBS in OPM2WT cells as revealed by cleavage of
Caspase-3 (Figure 8). In preliminary metabolic stability studies using human liver microsomes,
both A-1210477 and dMCL1-2 have similar T1/2 values of 21.7 and 20.6 minutes respectively
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(Supplementary Table 2).
Figure 8. Apoptosis induction factor assay. Cleaved Caspase-3 was probed by immunoblot of
OPM2 cells treated with dMCL1-2 for 24 hours with 1% FBS.
ACS Paragon Plus Environment
1
9

Journal of Medicinal Chemistry
Page 20 of 71
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
MCL1 has been recognized as chemically intractable due to its flat and featureless BH3
groove. MCL1 is also considered a very difficult protein to target due to its rapid turnover rate,
5
3
under two hours in most cell lines and conditions. To further comprehend the degradation
effect of this class of compounds against natural MCL1 degradation, we performed control
treatments of OPM2 and MM1.S cells with dMCL1-2 and cycloheximide (Figure 9a).
Cycloheximide (CHX) inhibits protein translation, providing insight into the natural turnover of
MCL1. This also clarifies the extent of chemically induced degradation occurring by following
MCL1 abundance in the presence and absence of dMCL1-2. Upon treatment of OPM2 and
MM1.S whole cells with CHX, MCL1 levels were depleted much more rapidly in the presence
of dMCL1-2 than in its absence in both cell lines (Figure 9a). Decreased protein expression was
observed at timepoints as early as 30 minutes after treatment with CHX. This finding does not
coincide with the normal depletion of MCL1 expression of cells treated with CHX alone,
indicating that dMCL1-2 is indeed influencing chemically induced degradation beyond that of
the natural MCL1 turnover. The same experiment was performed on the genetically engineered
OPM2CRBN-/- cell line (Figure 9b). Consistent with the former experiment, there were marked
differences in MCL1 levels in OPM2WT cells treated with CHX in the presence versus absence of
dMCL1-2. Levels of MCL1 in OPM2CRBN-/- cells degraded consistently over all timepoints of
CHX treatment regardless of the presence of dMCL1-2, again confirming the necessity for
CRBN in this proteasome-mediated degradation mode of action.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
ACS Paragon Plus Environment
2
0

Page 21 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 9. Cyclohexamide (CHX) chase assays to compare MCL1 degradation to normal protein
turnover. (a) Immunoblot for MCL1 after 72 hour treatment with DMSO or dMCL1-2 and then
treated for the last 30, 60, or 120 minutes with CHX in OPM2 or MM1.S cells. (b) Immunoblot
for MCL1 after 72 hour treatment with DMSO or dMCL1-2 and then treated for the last 60 or
20 minutes with CHX in OPM2WT or OPM2CRBN-/- cells.
1
CONCLUSION
Most proximity mediated protein degradation efforts in the literature to date have focused on
1
7
54
targeting tumour driving proteins such as CDK9, kinases, and bromodomain and extra
terminal (BET) proteins.27-30a Our SAR studies have yielded dMCL1-2, the first demonstration
of a PROTAC which effectively enhances proximity between MCL1 and the E3 ligase CRBN,
inducing direct ubiquitination of MCL1 and labeling it for proteasomal degradation at nanomolar
ACS Paragon Plus Environment
2
1

Journal of Medicinal Chemistry
Page 22 of 71
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
concentrations (Figure 7c). Through biotinylated affinity studies, we qualitatively assessed
binding capabilities in a rapid manner, guiding the synthetic design. Our investigations have
shown that these compounds can form ternary complexes between CRBN and MCL1 (Figures 4
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
&
5), necessary for PROTAC mediated degradation. The CRBN and proteasome dependency of
this mechanism have also been confirmed through a CRISPR knockout of CRBN as well as
biochemical studies with bortezomib. Our pharmacokinetic studies show that there is cellular
uptake of dMCL1-2 in OPM2 cells sufficient for degradation to occur at nanomolar
concentrations. Future clinical development of this compound will require detailed quantitative
pharmacokinetic analysis. Computational studies provide a dynamic prediction of the intimate
interactions between MCL1 and CRBN when tethered through these degraders. These
simulations are consistent with experimental observations of MCL1 ubiquitination and
degradation, revealing extensive contacts between MCL1 and CRBN, which corroborates the
sensitivity of the linker and its attachment location.
Through our systematic design strategy, computational modelling, and molecular biology
guided SAR studies, we have developed novel PROTACs which target the anti-apoptotic protein
MCL1. These compounds effectively recruit the CUL4A-DDB1 CRBN E3 ligase pathway to
degrade MCL1 at nanomolar concentrations, activating the cellular apoptosis machinery. As new
MCL1 ligands continue to be developed and optimized, this chemical probe technology will
provide a firm foundation for the discovery of future iterations of MCL1 PROTACs. This report
is a fundamental first step toward the development of a new class of anti-apoptotic BCL-2 family
protein degraders which will be powerful tools for studying this family of anti-apoptotic proteins
and provide a more in-depth understanding into their mechanisms of action.
EXPERIMENTAL SECTION
ACS Paragon Plus Environment
2
2

Page 23 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
All chemical reactions were carried out under a nitrogen atmosphere with dry solvents under
anhydrous conditions, unless otherwise noted. Reagents were purchased at the highest
commercial quality and used without further purification, unless otherwise stated.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Tetrahydrofuran (THF) and toluene were distilled immediately before use from sodium-
benzophenone. “Dri-Solv” EMD Millipore grade DMF was used. A kdScientific KDS-210
syringe pump was used for dropwise additions of reagents. Triturations were performed using a
VWR Model 75T Ultrasonic Cleaner. Solvents were removed in vacuo using either a Buchi R-
3
0
00 rotavapor (equipped with an I-300 Pro Interface, B-300 Base Heating Bath, Welch 2037B-
1 DryFast pump and VWR AD15R-40-V11B Circulating Bath), or Biotage V-10 evaporator, or
Kugelrohr short path distillation apparatus. Reactions were monitored by thin layer
chromatography (TLC) carried out on Merck glass silica gel plates (60F254) using UV light as a
visualizing agent and iodine and/or phosphomolybdic acid stain as developing agents. Manual
flash chromatography was performed using Silicycle SiliaFlash F60 silica gel (particle size
0
.040–0.063 mm, 230-400 mesh) as well as for automated flash chromatography. Solvents for
silica gel chromatography were used as supplied by Sigma-Aldrich. Automated flash
chromatography was performed on a Biotage Isolera instrument, equipped with a UV detector
and Biotage Dalton mass detector. Chromatograms were recorded at 254 and 280 nm. High-
resolution mass spectra (HRMS) and low-resolution mass spectra (LRMS) were obtained using
Agilent 6520 Accurate-Mass QTOF LC/MS or Bruker Maldi-TOF Autoflex III and GenTech
5890 Series II SSQ 7000 instruments respectively. Final compounds tested in experiments were
evaluated after analytical high performance liquid chromatography (HPLC) performed on an
Agilent 1260 Infinity LC equipped with an Agilent 1260 autosampler, an Agilent 1260 multi-
wavelength UV detector, and an Agilent 1260 automated fraction collector with a Poroshell 120
ACS Paragon Plus Environment
2
3

Journal of Medicinal Chemistry
Page 24 of 71
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
EC-C18 4.6 x 50 mm 2.7 µm column coupled with a Poroshell 120 EC-C18 4.6x5 mm 2.7 µm
UHPLC guard column. Purification was run with a flow rate of 1.5 mL/min. Solvents (H O,
2
acetonitrile, isopropanol) containing 0.1% trifluoroacetic acid (TFA) were used. The following
gradient was used at 60ºC: Method A: 5-95% MeCN in water, 0-20 min. The purity of final
compounds was evaluated using the analytical HPLC system described above, characterized by
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
13
MS and NMR and compound purity was >95%. H and C Nuclear Magnetic Resonance (NMR)
spectra were recorded on Bruker Avance III 400 MHz (BBFO probe), Bruker DRX 400 MHz
(BBO probe), Bruker Avance 400 MHz (BBO probe), or Bruker Avance III 600 MHz (BBO
probe) spectrometers. The following abbreviations are used to designate multiplicities: s =
singlet, d = doublet, t = triplet, q = quartet, p = pentet, m = multiplet, br = broad.
Materials. Compounds 9, tert-butyl 4-(4-((4-bromo-1,3-dimethyl-1H-pyrazol-5-
yl)methoxy)phenyl)piperazine-1-carboxylate, and ethyl 7-(1,3-dimethyl-5-((4-(piperazin-1-
yl)phenoxy)methyl)-1H-pyrazol-4-yl)-1-(2-morpholinoethyl)-3-(3-(naphthalen-1-yloxy)propyl)-
7
1
1
H-indole-2-carboxylate were prepared according to the procedure of Souers. Perfluorophenyl
-((2-(2,6-dioxopiperidin-3-yl)-1,3-dioxoisoindolin-4-yl)oxy)-2,17-dioxo-7,10,13-trioxa-3,16-
4
6
diazaicosan-20-oate was prepared by a known procedure. A-1210477 was purchased from
Selleckchem and used without further purification.
N-(1-((2-(2,6-dioxopiperidin-3-yl)-1,3-dioxoisoindolin-4-yl)oxy)-2-oxo-7,10,13-trioxa-3-
azahexadecan-16-yl)-5-((3aS,4S,6aR)-2-oxohexahydro-1H-thieno[3,4-d]imidazol-4-
yl)pentanamide (1). To a mixture of 4,7,10-trioxa-1,13-tridecanediamine (0.81 g, 3.7 mmol) and
o
Et
over 1 hour. The mixture was allowed to stir at ambient temperature for 1 hour more and was
then concentrated in vacuo. The residue was triturated with Et O (2 x 7 ml x 2 hours) furnishing
3
N (0.20 ml, 1.4 mmol) was added a solution of 16 (0.15 g, 0.37 mmol) in DMF (6 ml) at 0 C
2
ACS Paragon Plus Environment
2
4

Page 25 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
0
.17 g (99%) of the corresponding mono-protected amine as a white powder. H NMR (400
MHz, CD OD) δ 4.53-4.48 (m, 1H), 4.34-4.30 (m, 1H), 3.70-3.58 (m, 13H), 3.54 (t, J = 6.0 Hz,
2H), 3.32 (t, J = 7.1 Hz, 2H), 3.30-3.20 (m, 1H), 2.98-2.83 (m, 4H), 2.71 (d, J = 12.7 Hz, 1H),
.22 (t, J = 7.3 Hz, 1H), 1.88-1.57 (m, 9H), 1.50-1.42 (m, 2H). All other data was in agreement
3
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
with literature.55
A solution of this mono-protected amine (45 mg, 0.10 mmol), DIPEA (0.035 ml, 0.20 mmol) and
2
1 (0.035 g, 0.089 mmol) in DMF (1.5 ml) was stirred at ambient temperature for 2 hours. The
solvent was removed in vacuo and the residue was purified by silica gel flash chromatography
(10-20% MeOH in CHCl
3
) to afford 0.029 g (43%) of the title compound as a 1:1 mixture of
) δ 10.32 (d, J = 24.0 Hz, 1H), 7.74 (t, J = 7.9 Hz,
1
diastereomers. H NMR (400 MHz, CDCl
3
1
H), 7.60 (t, J = 5.8 Hz, 1H), 7.54 (d, J = 7.3 Hz, 1H), 7.22 (d, J = 8.4 Hz, 1H), 6.67 (t, J = 5.6
Hz, 1H), 6.34 (d, J = 5.7 Hz, 1H), 5.87 – 5.74 (m, 1H), 5.06 – 4.94 (m, 1H), 4.66 (s, 2H), 4.47
(dd, J = 7.8, 4.9 Hz, 1H), 4.28 (dd, J = 8.0, 4.6 Hz, 1H), 3.68 – 3.53 (m, 12H), 3.51 – 3.41 (m,
3
H), 3.31 (q, J = 6.1 Hz, 2H), 3.12 (td, J = 7.3, 4.5 Hz, 1H), 2.82 (dtd, J = 18.5, 11.6, 10.8, 4.1
Hz, 4H), 2.72 (d, J = 12.9 Hz, 1H), 2.20 – 2.07 (m, 3H), 1.86 (p, J = 6.5 Hz, 2H), 1.75 (p, J = 6.0
1
3
Hz, 2H), 1.67 – 1.55 (m, 3H), 1.39 (p, J = 7.4 Hz, 2H). C NMR (101 MHz, CDCl3) δ 173.0,
1
1
3
72.1, 172.0, 172.0, 169.0, 169.0, 166.9, 166.7, 166.0, 164.0, 154.6, 137.0, 133.5, 119.8, 119.8,
18.1, 117.3, 70.4, 70.1, 69.9, 69.8, 68.7, 68.2, 68.2, 61.8, 61.8, 60.2, 55.6, 50.7, 49.3, 40.5, 40.5,
7.7, 37.6, 36.5, 35.9, 35.9, 31.5, 29.3, 28.8, 28.8, 28.2, 28.2, 28.1, 28.0, 25.5, 25.5, 22.6. HRMS
+
(ESI) m/z calc. for [C35
H
48
N
6
O
11S + H] = 783.2994, found 738.3005.
N-(2,6-dioxopiperidin-3-yl)-5-((3aS,4S,6aR)-2-oxohexahydro-1H-thieno[3,4-d]imidazol-4-
yl)pentanamide (2). A solution of 3-amino-piperidine-2,6-dione hydrochloride (0.066 g, 0.40
mmol), DIPEA (0.21 ml, 0.16 g, 1.20 mmol) and 16 (0.165 g, 0.040 mmol) in DMF (3.0 ml) was
ACS Paragon Plus Environment
2
5

Journal of Medicinal Chemistry
Page 26 of 71
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
stirred at ambient temperature overnight. The solvent was removed in vacuo and the residue was
redissolved in DMF (3 ml) and stirred with K
2
CO (0.15 g, 1.1 mmol) for 2 hours. The inorganic
3
salts were filtered off, and the filtrate was concentrated in vacuo. The residue was purified by
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
silica gel flash chromatography (20% MeOH in CHCl ) to afford 0.125 g (88%) of the title
3
1
compound. H NMR (400 MHz, DMSO-d
6
) δ 10.77 (s, 1H), 8.14 (d, J = 8.3 Hz, 1H), 6.38 (d, J =
2
4.5 Hz, 2H), 4.53 (q, J = 8.4 Hz, 1H), 4.30 (dd, J = 7.8, 5.0 Hz, 1H), 4.13 (ddd, J = 7.4, 4.4, 1.8
Hz, 1H), 3.10 (ddd, J = 8.6, 6.1, 4.3 Hz, 1H), 2.82 (dd, J = 12.4, 5.0 Hz, 1H), 2.71 (ddd, J = 18.0,
1
4
0.6, 8.2 Hz, 1H), 2.57 (d, J = 12.4 Hz, 1H), 2.13 (td, J = 7.3, 1.9 Hz, 2H), 1.90 (td, J = 10.6, 9.4,
1
3
.9 Hz, 2H), 1.67 – 1.41 (m, 4H), 1.40 – 1.26 (m, 2H). C NMR (101 MHz, CD OD) δ 174.8,
3
173.4, 172.3, 164.8, 61.8, 60.3, 55.5, 49.6, 39.6, 35.1, 30.7, 28.0, 28.0, 25.2, 24.2. HRMS (ESI)
+
m/z calc. for [C15
H
22
N
4
O
4
S + H] = 355.1435, found 355.1437.
N1-(2,6-dioxopiperidin-3-yl)-N4-(15-oxo-19-((3aS,4S,6aR)-2-oxohexahydro-1H-
thieno[3,4-d]imidazol-4-yl)-4,7,10-trioxa-14-azanonadecyl)succinamide (3). To a mixture of
4
,7,10-trioxa-1,13-tridecanediamine (0.81g, 3.7 mmol) and Et N (0.20 ml, 1.4 mmol) was added
3
o
a solution of 16 (0.15 g, 0.37 mmol) in DMF (6 ml) at 0 C over 1 hour. The mixture was
allowed to stir at ambient temperature for 1 hour and was then concentrated in vacuo. The
residue was triturated with Et
2
O (2 x 7 ml x 2 hours) furnishing 0.17 g of the corresponding
1
mono-protected amine (99%) as a white powder. H NMR (400 MHz, CD
3
OD) δ 4.53-4.48 (m,
1
H), 4.34-4.30 (m, 1H), 3.70-3.58 (m, 13H), 3.54 (t, J = 6.0 Hz, 2H), 3.32 (t, J = 7.1 Hz, 2H),
3.30-3.20 (m, 1H), 2.98-2.83 (m, 4H), 2.71 (d, J = 12.7 Hz, 1H), 2.22 (t, J = 7.3 Hz, 1H), 1.88-
.57 (m, 9H), 1.50-1.42 (m, 2H). All other data was in agreement with literature.55
1
A solution of this mono-protected amine (0.040 g, 0.09 mmol), DIPEA (0.23 ml, 0.168 g, 1.30
mmol) and succinic anhydride (0.036 g, 0.089 mmol) in DMF (3.0 ml) was stirred at ambient
ACS Paragon Plus Environment
2
6

Page 27 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
temperature overnight. The solvent was removed in vacuo and the residue was triturated with a
1
:1 CHCl
3
/Et O mixture (2 x 10ml) for 15 minutes, the product was then pelleted through
2
centrifugation of the resultant gel. The wet solid was dried in vacuo to furnish 0.15 g (76%) of
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
the corresponding acid. H NMR (400 MHz, DMSO-d
6
) δ 12.00 (br s, 1H), 7.79 (t, J = 5.9 Hz,
1
3
2
H), 7.73 (t, J = 5.8 Hz, 1H), 6.40 (s, 1H), 6.34 (s, 1H), 4.35-4.28 (m, 1H), 4.16-4.10 (m, 1H),
.58-3.48 (m, 8H), 3.46 (t, J = 6.3 Hz, 4H), 3.15-3.03 (m, 5H), 2.84 (dd, J = 12.5, 5.1 Hz, 1H),
.31 (t, J = 6.7 Hz, 2H), 2.05 (t, J = 6.9 Hz, 2H), 1.65-1.58 (m, 5H), 1.55-1.25 (m, 6H).
To a solution of the free acid (0.15 g, 0.27 mmol) and DIPEA (0.17 ml, 0.129 g, 1.0 mmol) in
o
DMF (1.5 ml) cooled to 0 C was added pentafluorophenyl trifluoroacetate (0.11 g, 0.070 ml,
o
0.41 mmol) over 1 minute and the mixture was stirred at 0 C for an additional 30 minutes. The
reaction mixture was allowed to warm to ambient temperature and after 2 hours the solvent was
removed in vacuo and the residue was triturated with Et O (2 x 7 ml) for 30 minutes, the product
2
was then pelleted through centrifugation of the resultant gel. The wet solid was dried in vacuo to
afford 0.12 g (54%) of the corresponding pentfluorophenyl-ester. This intermediate was used
1
without further purification. H NMR (400 MHz, CDCl
3
) δ 6.71 (t, J = 5.9 Hz, 1H), 6.44 (t, J =
5.8 Hz, 1H), 5.63 (s, 1H), 4.83 (s, 1H), 4.55-4.48 (m, 1H), 4.36-4.30 (m, 1H), 3.70-3.55 (m,
1
1
2H), 3.43-3.35 (m, 4H), 3.20-3.15 (m, 1H), 3.05 (t, J = 6.4 Hz, 2H), 2.93 (dd, J=12.7, 4.8 Hz,
H), 2.72 (d, J = 12.7 Hz, 1H), 2.63 (t, J = 6.8 Hz, 2H), 2.22-2.17 (m, 2H), 1.95-1.65 (m, 10H).
A solution of this activated ester (0.12 g, 0.17 mmol), DIPEA (0.10 ml, 0.748 g, 0.58 mmol) and
3-amino-piperidine-2,6-dione hydrochloride (0.03 g, 0.18 mmol) in DMF (2.0 ml) was stirred at
ambient temperature for 2 hours. The solvent was removed in vacuo and the residue was
triturated with Et
2
O (2 x 5 ml) for 30 minutes. The white powder (0.13 g) was redissolved in
CO (0.050 g, 0.36 mmol) for 1 hour. The inorganic salts were
DMF (1.5 ml) and stirred with K
2
3
ACS Paragon Plus Environment
2
7

Journal of Medicinal Chemistry
Page 28 of 71
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
removed by filteration, the filtrate was concentrated in vacuo and the residue was triturated with
a 1:1 CHCl
3
/Et O mixture (2 x 5ml) for 30 minutes. The product was then pelleted through
2
centrifugation of the resultant gel. The wet solid was dried in vacuo to afford 0.098 g (83%) of
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
the title compound. H NMR (400 MHz, DMSO-d
6
) δ 10.77 (s, 1H), 8.20 (d, J = 8.2 Hz, 1H),
7
5
3
7
4
.77 (dt, J = 23.8, 5.5 Hz, 2H), 6.49 – 6.28 (m, 2H), 4.53 (q, J = 8.4 Hz, 1H), 4.31 (dd, J = 7.7,
.0 Hz, 1H), 4.14 (ddd, J = 7.7, 4.4, 1.9 Hz, 1H), 3.57 – 3.45 (m, 8H), 3.39 (t, J = 6.4 Hz, 4H),
.32 (s, 2H), 3.08 (p, J = 6.7 Hz, 5H), 2.83 (dd, J = 12.4, 5.1 Hz, 1H), 2.72 (ddd, J = 18.0, 9.5,
.6 Hz, 1H), 2.62 – 2.56 (m, 1H), 2.42 – 2.27 (m, 4H), 2.05 (t, J = 7.4 Hz, 2H), 1.90 (tt, J = 7.3,
1
3
.1 Hz, 2H), 1.62 (q, J = 6.7 Hz, 4H), 1.55 – 1.43 (m, 3H), 1.37 – 1.24 (m, 2H). C NMR (151
MHz, DMSO-d
6
) δ 173.4, 172.7, 172.3, 171.9, 171.5, 163.1, 70.2, 70.0, 68.5, 68.5, 61.5, 59.6,
5
5.9, 49.4, 36.2, 36.1, 35.6, 31.3, 31.2, 31.2, 29.8, 29.8, 28.7, 28.5, 25.7, 24.8. HRMS (ESI) m/z
+
calc. for [C29
H
48
N
6
O S + H] = 657.3276, found 657.3295.
9
N-(3-(2-(2-(3-aminopropoxy)ethoxy)ethoxy)propyl)-5-((3aS,4S,6aR)-2-oxohexahydro-1H-
thieno[3,4-d]imidazol-4-yl)pentanamide (4). To a mixture of 4,7,10-trioxa-1,13-
tridecanediamine (0.81g, 3.7 mmol) and Et N (0.20 ml, 1.4 mmol) was added a solution of 16
3
o
(0.15 g, 0.37 mmol) in DMF (6 ml) at 0 C over 1 hour. The mixture was allowed to stir at
ambient temperature for 1 hour more and was then concentrated in vacuo. The residue was
triturated with Et
2
O (2 x 7 ml x 2 hours) furnishing 0.17 g of the corresponding mono-protected
1
amine (99%) as a white powder. H NMR (400 MHz, CD
3
OD) δ 4.53-4.48 (m, 1H), 4.34-4.30
(m, 1H), 3.70-3.58 (m, 13H), 3.54 (t, J = 6.0 Hz, 2H), 3.32 (t, J = 7.1 Hz, 2H), 3.30-3.20 (m,
1
1
H), 2.98-2.83 (m, 4H), 2.71 (d, J = 12.7 Hz, 1H), 2.22 (t, J = 7.3 Hz, 1H), 1.88-1.57 (m, 9H),
.50-1.42 (m, 2H). All other data was in agreement with literature.55
ACS Paragon Plus Environment
2
8

Page 29 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
4
,20-dioxo-24-((3aS,4S,6aR)-2-oxohexahydro-1H-thieno[3,4-d]imidazol-4-yl)-9,12,15-
trioxa-5,19-diazatetracosan-1-oic acid (5). A solution of 4 (0.16 g, 0.36 mmol), DIPEA (0.23
ml, 0.168 g, 1.30 mmol) and succinic anhydride (0.036 g, 0.089 mmol) in DMF (3.0 ml) was
stirred at ambient temperature overnight. The solvent was removed in vacuo and the residue was
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
triturated with a 1:1 CHCl
3
/Et O mixture (2 x 10ml) for 15 min, the product was then pelleted
2
through centrifugation of the resultant gel. The wet solid was dried in vacuo to furnish 0.15 g
1
(
76%) of the corresponding acid. H NMR (400 MHz, d-DMSO) δ 12.00 (br s, 1H), 7.79 (t, J =
5
1
.9 Hz, 1H), 7.73 (t, J = 5.8 Hz, 1H), 6.40 (s, 1H), 6.34 (s, 1H), 4.35-4.28 (m, 1H), 4.16-4.10 (m,
H), 3.58-3.48 (m, 8H), 3.46 (t, J = 6.3 Hz, 4H), 3.15-3.03 (m, 5H), 2.84 (dd, J = 12.5, 5.1 Hz,
1H), 2.31 (t, J = 6.7 Hz, 2H), 2.05 (t, J = 6.9 Hz, 2H), 1.65-1.58 (m, 5H), 1.55-1.25 (m, 6H).
Perfluorophenyl 4,20-dioxo-24-((3aS,4S,6aR)-2-oxohexahydro-1H-thieno[3,4-d]imidazol-
4
-yl)-9,12,15-trioxa-5,19-diazatetracosan-1-oate (6). A solution of the free acid 5 (0.15 g, 0.27
o
mmol) and DIPEA (0.17 ml, 0.129 g, 1.0 mmol) in DMF (1.5 ml) was cooled to 0 C.
Pentafluorophenyl trifluoroacetate (0.11 g, 0.070 ml, 0.41 mmol) was added over 1 minute and
o
the mixture was stirred at 0 C for an additional 30 minutes. The reaction mixture was allowed to
warm to ambient temperature and after 2 hours the solvent was removed in vacuo and the residue
was triturated with Et O (2 x 7 ml) for 30 min. The product was then pelleted through
2
centrifugation of the resultant gel. The wet solid was dried in vacuo to afford 0.12 g (54%) of the
corresponding pentfluorophenyl-ester. This intermediate was used without further purification.
1H NMR (400 MHz, CDCl
) δ 6.71 (t, J = 5.9 Hz, 1H), 6.44 (t, J = 5.8 Hz, 1H), 5.63 (s, 1H),
.83 (s, 1H), 4.55-4.48 (m, 1H), 4.36-4.30 (m, 1H), 3.70-3.55 (m, 12H), 3.43-3.35 (m, 4H), 3.20-
.15 (m, 1H), 3.05 (t, J = 6.4 Hz, 2H), 2.93 (dd, J = 12.7, 4.8 Hz, 1H), 2.72 (d, J = 12.7 Hz, 1H),
3
4
3
2.63 (t, J = 6.8 Hz, 2H), 2.22-2.17 (m, 2H), 1.95-1.65 (m, 10H).
ACS Paragon Plus Environment
2
9