Design and exploration of C-3 benzoic acid bioisosteres and alkyl replacements in the context of GSK3532795 (BMS-955176) that exhibit broad spectrum HIV-1 maturation inhibition

Article abstract of DOI:10.1016/j.bmcl.2021.127823

GSK3532795 (formerly BMS-955176) is a second-generation HIV-1 maturation inhibitor that has shown broad spectrum antiviral activity and preclinical PK predictive of once-daily dosing in humans. Although efficacy was confirmed in clinical trials, the observation of gastrointestinal intolerability and the emergence of drug resistant virus in a Phase 2b clinical study led to the discontinuation of GSK3532795. As part of the effort to further map the maturation inhibitor pharmacophore and provide additional structural options, the evaluation of alternates to the C-3 phenyl substituent in this chemotype was pursued. A cyclohexene carboxylic acid provided exceptional inhibition of wild-type, V370A and ΔV370 mutant viruses in addition to a suitable PK profile following oral dosing to rats. In addition, a novel spiro[3.3]hept-5-ene was designed to extend the carboxylic acid further from the triterpenoid core while reducing side chain flexibility compared to the other alkyl substituents. This modification was shown to closely emulate the C-3 benzoic acid moiety of GSK3532795 from both a potency and PK perspective, providing a non-traditional, sp³-rich bioisostere of benzene. Herein, we detail additional modifications to the C-3 position of the triterpenoid core that offer effective replacements for the benzoic acid of GSK3532795 and capture the interplay between these new C-3 elements and C-17 modifications that contribute to enhanced polymorph coverage.

Full text of DOI:10.1016/j.bmcl.2021.127823

Bioorg. Med. Chem. Lett. 36 (2021) 127823
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Design and exploration of C-3 benzoic acid bioisosteres and alkyl
replacements in the context of GSK3532795 (BMS-955176) that exhibit
broad spectrum HIV-1 maturation inhibition
,
,
,
,g
Jacob J. Swidorskia *, Susan Jenkinsb g, Umesh Hanumegowdab g, Dawn D. Parkerb
,
Brett R. Benoc , Tricia Protackd, Alicia Nga, Anuradha Guptae, Yoganand Shanmugame,
,
f
Ira B. Dickerd h, Mark Krystald g, Nicholas A. Meanwella , Alicia Regueiro-Rena
,
,
,
f
,f
a Department of Discovery Chemistry, Bristol Myers Squibb Research and Development, 5 Research Parkway, Wallingford, CT 06492, USA
b Department of Pharmaceutical Candidate Optimization, Bristol Myers Squibb Research and Development, 5 Research Parkway, Wallingford, CT 06492, USA
c Department of Computer-Assisted Drug Design, Bristol Myers Squibb Research and Development, 5 Research Parkway, Wallingford, CT 06492, USA
d Department of Virology, Bristol Myers Squibb Research and Development, 5 Research Parkway, Wallingford, CT 06492, USA
e Biocon Bristol Myers Squibb Research & Development Center, Bangalore, India
A R T I C L E I N F O
A B S T R A C T
Keywords:
GSK3532795 (formerly BMS-955176) is a second-generation HIV-1 maturation inhibitor that has shown broad
spectrum antiviral activity and preclinical PK predictive of once-daily dosing in humans. Although efficacy was
confirmed in clinical trials, the observation of gastrointestinal intolerability and the emergence of drug resistant
virus in a Phase 2b clinical study led to the discontinuation of GSK3532795. As part of the effort to further map
the maturation inhibitor pharmacophore and provide additional structural options, the evaluation of alternates
to the C-3 phenyl substituent in this chemotype was pursued. A cyclohexene carboxylic acid provided exceptional
inhibition of wild-type, V370A and ΔV370 mutant viruses in addition to a suitable PK profile following oral
dosing to rats. In addition, a novel spiro[3.3]hept-5-ene was designed to extend the carboxylic acid further from
the triterpenoid core while reducing side chain flexibility compared to the other alkyl substituents. This modi-
fication was shown to closely emulate the C-3 benzoic acid moiety of GSK3532795 from both a potency and PK
perspective, providing a non-traditional, sp3-rich bioisostere of benzene. Herein, we detail additional modifi-
cations to the C-3 position of the triterpenoid core that offer effective replacements for the benzoic acid of
GSK3532795 and capture the interplay between these new C-3 elements and C-17 modifications that contribute
to enhanced polymorph coverage.
HIV-1
Maturation inhibitors
Triterpene
Antiviral
GSK3532795
BMS-955176
Bioisostere
HIV-1 maturation inhibitors (MIs) are a new class of therapy tar-
geting the final step of five protease-mediated Gag protein cleavage
events that occur late in the HIV-1 replication cycle.1 Instead of binding
to the protease directly, maturation inhibitors bind to the Gag protein
near the junction between the capsid (CA) and spacer peptide 1 (SP1)
and disturb the final and rate-limiting step of the proteolytic process.
Disruption of this cleavage event interferes with the proper release of the
structural proteins essential for the carefully ordered assembly of the
conical core of the viral particle, resulting in virus that is unable to
complete the maturation process which ultimately compromises the
iterative cycle of replication.2–4
Bevirimat (1) was the first HIV-1 maturation inhibitor to be
advanced in clinical trials, establishing proof-of-concept for this mech-
anism of action (MOA).5 In an efficacy trial, 45% of patients adminis-
tered 1 responded to therapy (designated as >0.5 log10 viral load
reduction) experiencing a mean viral load reduction of 1.26 log10
copies/mL.6 However, the majority (55%) of patients experienced a
minimal effect on viremia, with an average viral load reduction of 0.05
* Corresponding author at: Small Molecule Drug Discovery, Bristol Myers Squibb Research and Early Development, PO Box 4000, Princeton, NJ 08543, USA.
E-mail address: jacob.swidorski@bms.com (J.J. Swidorski).
f
Small Molecule Drug Discovery, Bristol Myers Squibb Research and Early Development, PO Box 4000, Princeton, NJ 08543, USA.
g
ViiV Healthcare, 36 East Industrial Rd., Branford, CT 06405, USA.
h
Retired from ViiV Healthcare.
https://doi.org/10.1016/j.bmcl.2021.127823
Received 25 September 2020; Received in revised form 11 January 2021; Accepted 18 January 2021
Available online 26 January 2021
0960-894X/© 2021 Elsevier Ltd. All rights reserved.

J.J. Swidorski et al.
Bioorganic & Medicinal Chemistry Letters 36 (2021) 127823
Fig. 1. Previously-disclosed HIV-1 maturation inhibitors.
log10 copies/mL. Further analysis of the non-responder patient popula-
tion revealed a series of distinct, naturally-occurring polymorphisms in
the glutamine-valine-threonine (QVT) residues found proximal to the
cleavage site at the CA-SP1 junction of the Gag polyprotein that corre-
lated with reduced susceptibility to 1.6–8
of the Gag protein. A ΔV370/T371A mutant virus with two amino acid
substitutions was also used as a surrogate for the single mutant ΔV370
since it conferred a similar level of resistance to 3 but proved to be a
more robust virus to maintain. In addition, a select subset of the com-
pounds was evaluated against the V362I/V370A and A364V virus var-
iants to assess the potential to inhibit viruses that are reported to confer
resistance to 1.8
Recent studies have identified 2 (Fig. 1), which incorporates a 4-
substituted benzoic acid moiety in place of the 3′3′-dimethylsuccinate
present in 1, as an inhibitor of HIV-1 maturation that preserved the
antiviral potency of 1 while significantly improving the deleterious ef-
fect of the presence of serum associated with the antiviral activity of 1,
although this compound was insufficient to address the baseline poly-
morphism issue.9 Modifications to the C-17 position of the triterpenoid
core of 2 by the introduction of the (2-aminoethyl)thiomorpholine 1,1-
dioxide element in 3 significantly enhanced the in vitro antiviral profile
to encompass the most prevalent polymorphic viruses. Whereas the
presence of the polymorphisms V370A and ΔV370 in the Gag protein
reduced the potency of 1 compared to WT virus (55x for V370A and
>1000x for ΔV370), 3 exhibited equipotent antiviral activity toward
WT and V370A mutant virus while the potency shift for the ΔV370
mutant was a modest 7-fold.10,11 Moreover, human serum (HS) binding
for 3 was 86% using an ultracentrifugation method while 1 was found to
be >99% bound to HS.10,11 The improvements in antiviral potency and
serum binding for 3 combined with a comparable preclinical PK profile
to 1 supported its advancement into clinical trials.11,12 In the clinic,
once-daily doses of 10–120 mg of 3 administered as monotherapy for 10
days were associated with a median viral load decline of 1.05–1.75 log10
copies/mL in patients infected with WT HIV-1 and, importantly, similar
efficacy was observed toward virus variants with pre-existing Gag
polymorphisms (viral load declined by 0.97–1.98 log10 copies/mL in
this patient population).13 Unfortunately, development of 3 was sus-
pended after completion of a phase 2b clinical trial due the observation
of gastrointestinal intolerability and treatment-emergent drug
resistance.14
The design of target compounds focused on further defining the
spacing and geometry between the carboxylic acid extending from the C-
3 position and the triterpenoid core while probing the effects of
increased rigidity on the antiviral spectrum. Since the (2-aminoethyl)
thiomorpholine 1,1-dioxide element at C-17 conferred optimal antiviral
activity and PK properties in the context 3 and its congeners, this moiety
was typically held constant although alternate C-17 elements were oc-
casionally employed.17,18 It was clear from modeling studies that the C-3
modifications of 1 and 2 projected the CO2H moiety a similar distance
from the core (6.0 Å for 1 and 5.8 Å for 2) and an energy-minimized
model showed that the carboxylic acid elements extended into a topo-
graphically similar region (Fig. 2). Propiolic acid 5 allowed a similar
vector to be maintained but projected the CO2H moiety a substantially
shorter distance from the core (4.1 Å compared to 5.8 Å for 2). This
change corresponded with a 15-fold reduction in HIV-1 weight inhibi-
tory potency when compared with 2, presumably because of the shorter
spacer distance. In this particular example, a CH2OH substituent was
incorporated at C-17 in place of the COOH of 1 and 2 for reasons of
synthetic accessibility, a reasonable change since previous studies have
shown a similar inhibition profile for compounds containing both of
these variations.9
Conformationally mobile alkyl modifications extending from the C-3
position such as the pentanoic acid 6 maintained potent inhibition of WT
HIV-1 but this compound was a 50-fold less potent inhibitor of the
ΔV370/T371A variant. Although the distance from the core to the CO2H
was slightly increased compared to 3, the added flexibility of the
unsubstituted alkyl chain was detrimental to the polymorphic virus in-
hibition. Introduction of a gem-dimethyl substitution in either the 3 (7)
or 4 (8) positions of the pentane side chain, designed to influence
conformational mobility, fully restored potency toward ΔV370/T371A
mutant in the case of 7 and partially so with 8, reinforcing the impor-
tance of side chain topography and providing evidence that a modest
increase in the distance from C-3 of the core to the carboxylic acid could
be tolerated.19 In addition, within this aliphatic side chain series, 7
demonstrated the highest level of inhibition of the V362I/V370A virus
that emerges in response to HIV-1 maturation inhibitors, with the EC50
value reduced by almost 100-fold compared to 3. Inhibition of A364V
virus by 7 and the effects of serum on the inhibition of WT was com-
parable to 3. Termination of the alkyl chain with a bis-carboxylic acid
substituent, as in 9, more severely affected inhibition of both WT and
ΔV370/T371A viruses. Since the separation of the carboxylic acid
moieties from the core appears to be reasonably close to that of the other
potent compounds, the added electrostatic interactions or presumably
the physicochemical properties associated with a second carboxylic acid
moiety appear to be unfavorable. This was a surprising result since
initial study of C-3 benzoic acid derivatives had revealed that a C-3
phenyl ring incorporating both meta- and para-carboxylic acid sub-
stituents was found to be essentially equipotent with 2.9
In an effort to further understand and define the role of the C-3
substituent, additional motifs that could further enlighten the re-
quirements at this critical element of the pharmacophore were examined
in conjunction with select changes at C-17. Our efforts were focused on
preserving or improving the antiviral activity inherent to 3 while
differentiating the chemotype in the event of liabilities arising during
development. We specifically targeted inhibition of viruses containing
the key variations in the QVT region of Gag, V370A and ΔV370, in
addition to maintaining inhibition of wild-type (WT) virus. The V370A-
containing virus was chosen because it was the most prevalent naturally-
occurring polymorphic variant in subtype B patients and exhibited 50-
fold reduced susceptibility to 1 (Table 1).15 While only present in a
small percentage of the subtype B patient population (0.9%), ΔV370-
containing viruses showed a particularly high level of resistance (>1000
fold) to 1 and this is a common polymorphic variation in non-subtype B
patients.10,16 Consequently, this virus was selected as a high bar sentinel
for inclusion in the initial screening tier. In vitro potency was measured
utilizing a multiple cycle replication assay in MT-2 cells using an NL4-3-
derived virus expressing the Renilla luciferase gene which was incor-
porated as a marker for virus growth (NLRepRlucP373S). HIV-1 V370A
and ΔV370 viruses were identical to the WT virus except for a single
amino acid substitution (V370A) or deletion (ΔV370) at the 370 position
2

J.J. Swidorski et al.
Bioorganic & Medicinal Chemistry Letters 36 (2021) 127823
Table 1
Antiviral activity toward wild type HIV-1, V370A and ΔV370 or ΔV370/T371, V362I/V370A viruses, wild type HIV-1 with 40% human serum and 27 mg/mL human
serum albumin, cytotoxicity and calculated distances between C3 and the COOH of the side chain.
C-3
C-17
In vitro HIV-1 inhibition EC50 (nM)a
CC50
M)
Distance C3 to
COOH (Å)
(
μ
WT
V370A
ΔV370 or
ΔV370/
T371Ab
V362I/
V370A
A364V
WT, 40%
HS + 27 mg/mL
HSA-FB
1
2
10
16
552
233
>10,000
–
–
–
–
–
16.8
27
6.0
5.8
>6000
316
3
4
1.9
1.8
2.7
2.7
13/23b
12.0
148
1480
11.5
6.1
9.2
5.8
5.8
–
–
13.7
5
6
252
2.3
>3000
>3000
–
–
–
–
–
–
>10
4.1
6.5
–
118b
10.2
7
1.3
5.3
306
3.0
2.0
1.7
0.6
–
2.3b
29.7b
555b
5.6b
2.2b
19.7b
2.1
1.6
–
1290
6.8
–
12.3
8.2
6.1
8
–
–
6.5
9
–
–
–
–
>30
13.5
9.6
5.8, 6.5
6.9
10
11
12
13
–
12.4
9.7
–
1850
1040
–
–
–
–
6.5
–
–
13.6
12.2
6.5
0.8
–
–
8.9
5.9
13a
13b
14a
14b
15
0.9
1.3
2.7
1.4
99
1.3
2.0
–
3.1
4.1
2.7
4.2
40.1
–
–
5.7
5.0
5.0
6.6
–
4.1
5.9
–
–
9.0
5.9
14.7
9.6
113
1800
878
–
9.1
5.9
–
13.2
18.6
5.9
39.2
4.9, 5.9
16
17
28.3c
26.7
85.4
53.1
437
–
–
–
–
–
–
>10
>30
6.4, 7.1
5.8, 6.7
60.8
(continued on next page)
3

J.J. Swidorski et al.
Bioorganic & Medicinal Chemistry Letters 36 (2021) 127823
Table 1 (continued)
C-3
C-17
In vitro HIV-1 inhibition EC50 (nM)a
CC50
M)
Distance C3 to
COOH (Å)
(
μ
WT
13
V370A
28.8
ΔV370 or
ΔV370/
T371Ab
V362I/
V370A
A364V
WT, 40%
HS + 27 mg/mL
HSA-FB
18
61.1
–
–
–
>30
6.4, 7.1
7.1
19
a
1.5
1.6
5.9
49
980
7.0
11.9
The antiviral activities of the compounds were assessed in a multiple cycle assay in MT-2 cells in the presence of 10% fetal bovine serum (FBS). WT, V370A, ΔV370,
ΔV370/T371A EC50 and CC50 data are the mean of at least two experiments.
b
c
In vitro inhibition of ΔV370/T371A virus
All compounds were >90% pure based on 1H NMR integration with the exception of 14. The purity of 14 was limited to 86% because poor solubility interfered with
further purification attempts. 1H NMR was used to determine purity since most of the compounds lacked a UV absorbing chromophore typically used in HPLC analysis.
¨
Fig. 2. Superposition of the low energy conformers of 1, 2, 5, and 16 (OPLS3/MM-GBSA). Image created with PyMOL (v 2.3 Schrodinger, LLC).
Table 2
In vivo pharmacokinetic properties of select compounds in ratsa,†
.
#
Metabolic
F (%)
AUC0-6h,
AUCtotal,
tmax,
Cmax,
CL,
Vss,
C24h,
stability in LM,
PO (µM∙h)
PO (µM∙h)
PO (h)
PO (µM)
IV (mL/min/kg)
IV (L/kg)
PO (nM)
t1/2 (min)
Human/Rat
1b
>120/29.5
>120/>120
>120/>120
58/>120
54e
4
6.4d
13.8 ± 1.4f
–
14.2 ± 1.1
15.8 ± 5.2
13.5 ± 4.8
1.8 ± 2e
4.0g
2.6 ± 0.6e
1.3 (1.2, 1.5)
1.2 ± 0.2
5.7 ± 1e
1.0g
0.6 ± 0.2e
0.1g
6 ± 4e
2b,d
3c
5.7 (4.7, 6.8)
–
26
58
32
4.1 ± 0.7
6.0 ± 0.0
2.7 ± 0.6
6.3 ± 1.1
2.2 ± 0.3
4.3 ± 0.6
2.9 ± 0.5
1.1 ± 0.1
2.3 ± 0.3
1.4 ± 0.2
128 ± 27
196 ± 102
240 ± 90
13c
–
–
1.1 ± 0.5
19c
69/>120
0.9 ± 0.26
†
All studies were conducted in accordance with the GSK Policy on the Care, Welfare and Treatment of Laboratory Animals and were reviewed by the Institutional
Animal Care and Use Committee either at GSK or by an ethical review process at the institution where the work was performed.
a
b
Each pharmacokinetic property listed is an average of the measurement from 3 rats unless otherwise noted. Doses were 5 mg/kg PO and 1 mg/kg IV.
Vehicle: poly(ethylene glycol) 300 (PEG 300), ethanol, Tween 80 (TW80) (89:10:1 v/v).
c
Vehicle: poly(ethylene glycol) 300 (PEG 300), ethanol, 0.1 N NaOH, Tween 80 (TW80) (84.5:10:5:0.5 v/v).
d
e
n = 2.
F(%), tmax, Cmax, CL, Vss and C24 are reported from the full 24 h rat PK screen.
f
AUCtotal was calculated from a full 24 h rat PK run in addition to a 6 h rat PK screen.
g
The individual value for tmax, CL, IV and Vss, IV for 2 were within 0.1 (h, mL/min/kg and L/kg).
An element of rigidity was introduced by locking the alkyl chains
through a cyclization strategy, as exemplified by analogues 10–12.
Although it incorporated a longer spacer unit (6.9 Å compared to 5.8 Å
for 3), the cyclohexylmethyl homolog 10, isolated as a mixture of cis and
trans isomers, showed excellent antiviral properties, with WT HIV-1
inhibitory potency comparable to 3 and a ΔV370/T371A EC50 of 5.6
nM. In addition, inhibition of the V362I/V370A virus was improved
almost 12-fold compared to 3, although inhibitory activity toward the
A364V virus was not improved. Locking the pentanoic acid moiety into a
cyclohexane ring, as in 11 and 12, maintained inhibition of WT virus in
both isomers, but 11 was a several-fold more potent inhibitor of the
ΔV370/T371A virus and showed equipotent inhibition of the mutant
and the WT viruses. Inhibition of the V362I/V370A and the A364V vi-
ruses by 11 was similar to that of the 4-substituted analog 10. The
conformation of the carboxylic acid was further constrained by attach-
ing the ring directly to the C-3 position of the triterpenoid core via an
alkene connector, as depicted by 13–19. Cyclohexene 13 is a potent
inhibitor of WT, V370A and ΔV370 viruses, with a 3- to 6-fold advantage
4

J.J. Swidorski et al.
Bioorganic & Medicinal Chemistry Letters 36 (2021) 127823
over 3 and is subject to a similar potency shift in the presence of 10%
FBS, 40% HS and 27 mg/mL HSA. In 13, the spatial disposition of the
carboxylic acid is conserved when compared to 3 and there is only a
modest change in the acidity of the carboxylic acid; a cyclohexane
carboxylic acid has a pKa of 4.90 compared to a pKa of 4.20 for benzoic
acid which is lower because of conjugation of the acid moiety to the
benzene ring.20,21 Rat PK studies with 13 showed a similar AUCtotal
(15.8 µM∙h), Cmax and C24 with an improved oral bioavailability (F) of
58% when compared to 3 after doses of 5 mg/kg PO (Table 2). When
isolated individually, both diastereoisomers 13a and 13b showed
similar potency, although rat PK studies were not repeated with the
individual isomers. The modified C-17 substituent in the ethyl-linked
methylsulfonyl piperidine moiety in 14a and 14b was also evaluated
because the corresponding benzoic acid 4 showed nearly identical in
vitro potency to 3 while conferring additional structural differentiation
from the clinical candidate. While WT and ΔV370 virus inhibition for
14a and 14b were similar to 3, inhibition of V362I/V370A virus was
improved ~10 fold while activity toward the A364V variant was com-
parable. Consistent with the structure-activity relationships (SARs) of
the other C-3 alkyl compounds, the bis-carboxylic acid 15 was less
potent.
isolated by crystallization from an acidic solution at room temperature.
Isomer 17 confers additional rigidification to the ring structure that
constrains the placement of the two carboxylic acid moieties while
isomer 18 is less rigid and allows for more movement of the carboxylic
acid moieties. However, the difference in mutant virus inhibition be-
tween the two isomers was only modest. Consistent with the other series
discussed above, the mono-carboxylic acid derivative of the spiro[3.3]
hept-5-ene substituent provides improved inhibitory activity toward
all of the viruses in the screening tier. In fact, as a mixture of di-
astereomers 19 displayed equivalent inhibitory potency toward the
V370A and ΔV370 mutant viruses to the clinical candidate 3 even
though the carboxylic acid in 19 extends 1.3 Å further from the core
than in the benzoic acid-based prototype. Adding further to its potential
to function as an effective bioisostere of a benzoic acid, in a rat PK study
19 exhibited similar oral exposure (AUCtotal of 13.5 µM∙h, a tmax of 6.3
h, a C24 of 240 nM and F = 32%) to 3 (AUCtotal of 14.2 µM∙h, a tmax of 6
h, a C24 of 128 nM and F = 26%) following doses of 5 mg/kg (Table 2).
The mode of action of betulinic acid-derived HIV-1 maturation in-
hibitors is beginning to emerge along with a more complete picture of
the HIV-1 maturation process as the result of recent structural stud-
ies.23–31 X-Ray crystallography,23 cryogenic electron tomography (Cryo-
ET),24,25 cryogenic electron microscopy (Cryo-EM),26 solid-state NMR
(ssNMR)27–29 and computational modeling30,31 studies have individu-
ally and collectively provided a more detailed understanding of the
mechanisms of the proteolytic processing and assembly of the hexameric
capsid protein lattice. Proteolytic cleavage of the CA-SP1 peptide is a
carefully choreographed process that depends upon access of HIV-1
protease to the cleavage site. The assembly of the hexameric core ap-
pears to be catalyzed by inositol hexakisphosphate (IP6) which convenes
the immature lattice bundle by engaging in electrostatic interactions
with the basic residues Lys290 and Lys359 that are contributed by each
monomer of the CA-SP1 peptide.32–34 In cryo-EM studies, IP6 is located
at the center of the hexameric structure between the two rings of lysine
residues acting as a nidus for bundle assembly.32 How IP6 modulates the
availability of the CA-SP1 cleavage site in a conformation for proteolytic
processing is not clear but in the absence of cleavage the lattice forms
with the SP1 peptide still attached with the cleavage site on the inner
surface of the hexameric bundle and inaccessible to the protease. After
proteolytic removal of the SP1 peptide, IP6 appears to stabilize the
mature capsid protein by engaging with Arg18 residues from each of the
CA peptide monomers; presumably, IP6 removal is an important element
of the capsid disassembly process during infection. Thus, IP6 promotes
both the assembly, maturation and stabilization of the HIV-1 capsid
while also playing a role in capsid dissolution.33,34 Micro electron
diffraction (microED) mapping of frozen, hydrated 3D microcrystals of
the CA-SP1 with and without 1 bound revealed that the maturation
inhibitor binds close to the IP6 binding site. However, while the precise
orientation of 1 in the hexameric bundle was not discernible at the
resolution of the experiment, modeling studies suggested that the suc-
cinic carboxylic acid moiety engaged with Lys359 across the hexamer
interface.35 These observations suggest that maturation inhibitors
compete with IP6 to modulate hexamer bundle formation in a fashion
that usurps the normal choreography of protease-mediated CA-SP1
cleavage. While this scenario explains the preservation of antiviral ac-
tivity observed with some bis-carboxylic acid derivatives, additional
insights into the SARs associated with the C-3 substituent and the precise
relationship between the carboxylic acid moiety and the triterpenoid
core will require the combination of higher resolution structures and a
more detailed understanding of the biochemical and kinetic aspects of
hexameric bundle formation and processing.
To further expand the SARs, a novel spiro[3.3]hept-5-ene was
conceived as a rigid building block that could place the carboxylic acid
(s) further from the triterpenoid core while retaining the geometry that
appears consequential to mutant virus inhibition. In an effort to quickly
probe inhibitory effectiveness of this modification, the spiro[3.3]hept-5-
ene-2,2-dicarboxylic acid moiety was appended to the C-3 position in
the background of a free carboxylic acid substituent at C-17, providing
16. Intriguingly, while 16 exhibited a several-fold reduction in inhibi-
tory potency toward WT virus compared to 1 and 2, the only other
compounds compiled in Table 1 that are not modified at C-17, inhibition
of the V370A mutant virus was improved several-fold while the ΔV370
virus, which was insensitive to 1 and 2, was inhibited with an EC50 of
437 nM. Molecular models of 17 indicate that one of the carboxylic acid
moieties extends 6.4 Å from the core while the second carboxylic acid
moiety extends further at a distance of 7.1 Å. The less distant carboxylic
acid moiety projects at a downward angle, placing it adjacent to the
second ring of the spirocycle rather than extending it further from the
triterpenoid core.
In an effort to further improve the virological profile and the physical
properties of the molecule, introduction of the (2-aminoethyl)thio-
morpholine 1,1-dioxide moiety to the C-17 position was explored.
Interestingly, during the synthesis of this molecule, an isomerization
occurred at the diene-based junction of the C-3 spirocycle and the tri-
terpenoid core to give a mixture of the ester precursors to 17 and 18. The
1H NMR chemical shift and the coupling pattern of the diene at the ring
junction in 17 and 18 clearly differentiated the two isomers. Using
deuterated acetic acid as the solvent, the 1H NMR spectrum of isomer 17
was associated with two doublets with chemical shifts of 5.92 and 5.81
ppm and coupling constants of 9.8–9.9 Hz, consistent with the C-1–C-2
alkene moiety. Using the same solvent, the C-2–C-3 vinyl proton of
isomer 18 resonated as a doublet of doublets with a chemical shift of
5.61 ppm and coupling constants of 6.0 and 1.6 Hz, a signal that is
distinctive for compounds in this series. A singlet resonating at 5.98 ppm
is consistent with the C-1′–C-2′ alkene present in the spirocycle. The
same alkene chemical shifts and coupling constants were observed in the
1H NMR of the mixture of isomers that gave 20, a closely related analog
that is missing the C-17 (2-aminoethyl)thiomorpholine 1,1-dioxide
moiety. In addition to the 1H NMR data for 20, a single crystal X-ray
determination of 20 confirmed the chemical structure of this isomer.
The X-ray structure showed an elongated C-1′–C-2′ bond length (1.468
Å) and a shortened C-3–C-1′ bond length (1.366 Å). Additionally,
conjugation was apparent based on the C-1–C-2 (1.406 Å) and C-2–C-3
(1.401 Å) bond lengths. The isomerization to give 20 was observed
following an alternate route which heated the spirocycle with HCl
overnight; however, both 17 and 18 were sufficiently stable to be
With substantial effort to develop therapeutic treatments for HIV-1
infections, patient outcomes have improved since the introduction of
antiretroviral therapy. Between 1987 and 2014, more than 28 new
molecular entities were developed to treat HIV/AIDS infections.36
Although the pace of new discoveries in the field has slowed somewhat,
new classes and combination therapies are important to treat the
5

J.J. Swidorski et al.
Bioorganic & Medicinal Chemistry Letters 36 (2021) 127823
Scheme 1. Preparation of C-3 alkyne 5. Reagents and conditions: (a) TMS-acetylene, TEA, CuI, Pd(PPh3)4, DMF, rt, 3 h, 93%; (b) TBAF (75% in H2O), THF, rt, 2 h,
98%; (c) t-BuLi, ClCO2Et, THF, ꢀ 78 ℃, 2 h; (d) NaOH (1 N), 1,4-dioxane, rt, 19 h, 21% over 2 steps.
Scheme 2. β-Alkyl Suzuki-Miyaura coupling process to introduce C-3 alkyl substituents (examples 6–12). Reagents and conditions: (a) olefin ester, 9-BBN, THF, 0 ◦C-
rt, 2 h, then 1 M K3PO4, triflate, PdCl2(dppf)-toluene adduct, 85 ◦C, 16–19 h; (b) 4-(2-chloroethyl)thiomorpholine 1,1-dioxide∙HCl, K3PO4, KI, MeCN, 100 ◦C, 15–24
h; (c) 1 N NaOH, 1,4-dioxane, 75 ◦C, 3–72 h.
resistance that continues to emerge to existing therapies. Targeting the
final step of protease-mediated Gag protein processing to inhibit the
maturation of HIV-1 virions is of contemporary interest within industrial
groups focused on developing mechanistically differentiated drugs and
academic laboratories seeking to understand the maturation process
more deeply.
required in order to surmount the gastrointestinal side effects observed
with this compound after long-term dosing in the Phase 2b clinical
study. The compounds described in this study offer structural differen-
tiation at a pharmacophoric element that differs to that in 1, which was
not associated with gastrointestinal problems over long-term dosing, but
further study will be required in order to understand toxicity profiles.
The spiro[3.3]hept-5-ene moiety in 16–19 offers an effective, architec-
turally interesting and sp3-rich replacement for the phenyl ring of 3
while the cyclohexene derivatives 13–15 present a more topologically
The promising efficacy demonstrated by 3 in early clinical trials
suggests potential for HIV-1 maturation inhibitors to be a part of future
therapeutic options. However, additional structural refinement will be
Scheme 3. Preparation of boronate esters for C-3 Suzuki-Miyaura coupling. Reagents and conditions: (a) KHMDS (0.91 M in THF), 1,1,1-trifluoro-N-phenyl-N-
((trifluoromethyl)sulfonyl)methanesulfonamide, THF, ꢀ 78 ◦C, 5.5 h, 15–53%; (b) bis(pinacolato)diboron, KOAc, PdCl2(dppf)-CH2Cl2 adduct, 1,4-dioxane, 70 ◦C,
5–22 h.
6

J.J. Swidorski et al.
Bioorganic & Medicinal Chemistry Letters 36 (2021) 127823
Scheme 4. Preparation of spiro[3.3]hept-5-ene-2,2-dicarboxylic acid 16. Reagents and conditions: (a) 27, bis(pinacolato)diboron, Pd(PPh3)Cl2, PPh3, PhOK,
toluene, 50 ◦C then K3PO4, PdCl2(dppf)-CH2Cl2 adduct, DMF, 80 ◦C, 20%; (b) t-butyldimethylsilane, TEA, Pd(OAc)2, DCE, 60 ◦C; (c) aq. TBAF, THF, rt; (d) NaOH, 1,4-
dioxane, 75 ◦C, 44% over 3 steps.
faithful phenyl bioisostere that offers the potential for the introduction
of substituents that explore topographical space not available by sub-
stitution of the 2-dimensional phenyl ring of 3.
because of the very non-polar nature of the compounds. Although the
yield for the initial formation of the vinyl triflate was quite variable from
one ketone to the next, sufficient material was prepared using this
methodology to deliver these analogs.
The synthesis of alkyne 5 employed a Sonigashira coupling of triflate
intermediate 219 with TMS-acetylene as the coupling partner to give 22
(Scheme 1). The TMS-protected intermediate 22 was unmasked by
treating with aqueous tetrabutylammonium fluoride (TBAF) solution in
THF at room temperature to give 23 which was deprotonated with t-
BuLi and the acetylide treated with ethyl chloroformate to give the
alkynyl ester. Both the alkynyl carboxylic acid and the C-28 alcohol were
deprotected using aqueous NaOH to provide 5.
After difficulty using standard Pd(PPh3)4 conditions for the Suzuki-
Miyaura coupling of vinyl triflate 34 to boronate ester 30, the reverse
process in which a C-3 vinyl boronate ester (not shown) was coupled
with the vinyl triflate intermediate 29 was attempted but without suc-
cess. Consequently, a modification of the conditions developed by
Miyaura and co-workers was employed to accomplish the coupling
(Scheme 4).39 In a one-pot sequence, vinyl triflate 29 was converted to
the boronate ester to which the vinyl triflate 34 was added along with a
secondary catalyst and base to achieve the coupling and provide the C-3
spiro[3.3]hept-5-ene-2,2-dicarboxylic ester 35.9 Intermediate 35 was
carried into the next series of reactions in the presence of some chro-
matographically close impurities in anticipation that they would more
easily be removed after subsequent steps. A three-step sequence to
deprotect the carboxylic acids on both the spirocycle and the C-17 car-
boxylic acid was then used to give 16. The C-17 benzyl group was
converted to a TBDMS-protected carboxylic acid using Pd(OAc)2, TEA
and t-butyldimethylsilane. The TBDMS group was removed using
aqueous TBAF and the remaining carboxylic acids unmasked via hy-
drolysis of the carboxylic esters using NaOH to give 16. Although the
compound still had sub-optimal purity (86%) after the first purification
by preparative HPLC, further purification was complicated because of
the poor solubility in typical preparative HPLC or crystallization sol-
vents. This compound was tested in the antiviral assays in impure form
in order to obtain a measure of the potency and thus the potential of the
chemotype.
Installation of alkyl substituents at the C-3 position was achieved
using β-alkyl Suzuki-Miyaura conditions (Scheme 2).37 Terminal olefins
with pendent alkyl esters were converted to alkyl boranes via hydro-
boration using 9-BBN followed by palladium-catalyzed cross-coupling
with the vinyl triflate 24 to give the C-3 alkylated products. Alkylation of
the C-17 amine utilizing 4-(2-chloroethyl) thiomorpholine 1,1-diox-
ide,12 K3PO4 and KI in acetonitrile heated to 100 ◦C in a sealed tube
followed. The ester of the C-3 alkyl chain was hydrolyzed with aqueous
NaOH to give carboxylic acids 6–12.
The preparation of the C-3 modified analogs 13–19 employed a
strategy utilizing Suzuki-Miyaura coupling of a boronate ester with the
vinyl triflate intermediates 24 and 34. Boronate ester 27 was prepared
by a process that began with deprotonation of ketone 25 with potassium
bis(trimethylsilyl)amide (KHMDS) at ꢀ 78 ◦C followed by treatment
with
1,1,1-trifluoro-N-phenyl-N-((trifluoromethyl)sulfonyl)meth-
anesulfonamide to give the vinyl triflate 25 (Scheme 3). Conversion of
the vinyl triflate 26 to the boronate ester 27 was accomplished using bis
(pinacolato)diboron, KOAc and PdCl2(dppf)-CH2Cl2 adduct while heat-
ing to 70 ◦C in 1,4-dioxane. Preparation of the di-carboxylate analog 30
and the mono-carboxylate boronate ester 33 were accomplished using
the same two step sequence from ketones 28 and 31.38 It should be noted
that both 27 and 33 were used in the subsequent Suzuki-Miyaura
coupling steps with impurities present that were difficult to remove
The preparation of compounds 13, 15 and 17–19 were accomplished
in a straightforward manner using the Suzuki-Miyaura coupling of vinyl
triflate 249 with the corresponding boronate esters (Scheme 5). After the
Suzuki-Miyaura coupling, alkylation of the C-17 amine was accom-
plished utilizing 4-(2-chloroethyl) thiomorpholine 1,1-dioxide∙HCl,
Scheme 5. Preparation of cycloalkene analogs 13, 15 and 17–19. Reagents and conditions: (a) Boronate ester, Pd(PPh3)4, Na2CO3⋅H2O, 1,4-dioxane, H2O, 85 ◦C,
5–21 h or boronate ester, S-phos, Pd(OAc)2 K3PO4, 1,4-dioxane, H2O, 75 ◦C; (b) 4-(2-chloroethyl)thiomorpholine 1,1-dioxide∙HCl, K3PO4, KI, MeCN, 100–120 ◦C,
15–22 h; (c) aq. NaOH, 1,4-dioxane, 65–85 ◦C, 15–83 h.
7

J.J. Swidorski et al.
Bioorganic & Medicinal Chemistry Letters 36 (2021) 127823
Fig. 3. ORTEP depiction of the single crystal X-ray structure of 20.22
Scheme 6. Preparation of cycloalkene analogs 14a and 14b. Reagents and conditions: (a) BrCH2CH2Cl, K3PO4, MeCN, 120 ◦C, 23–24 h; (b) 4-(methylsulfonyl)
piperidine, 1,4-dioxane, DIPEA, 1,4-dioxane, 95–100 ◦C, 15–23 h; (c) NaOH (1 N), 1,4-dioxane, 60 ◦C, 4–5 h.
K3PO4, KI in acetonitrile heated to 100–120 ◦C in a sealed tube. The
alkylated intermediates were treated with aq. NaOH to hydrolyze the
carboxylate appendages of the cycles introduced from the Suzuki-
Miyaura coupling step to the carboxylic acids. Interestingly, two iso-
mers were formed during the alkylation step en route to 18. The two
isomers were separated by supercritical fluid chromatography (SFC)
after the final hydrolysis to give 17 and 18. Structural identification of
isomer 17 was confirmed by comparison of the spectroscopic data to the
data from the C-17 amine intermediate 20 (Fig. 3) formed from a
different synthetic route that encountered difficulties, but allowed for
the determination of a single crystal X-ray structure which confirmed the
assignment.
Ramasamy, Richard Rampulla, Arvind Mathur and the Bristol Myers
Squibb Department of Discovery Synthesis for providing key in-
termediates which were used to synthesize several of the compounds
described in this manuscript.
Appendix A. Supplementary data
Supplementary data to this article can be found online at https://doi.
org/10.1016/j.bmcl.2021.127823.
References
The preparation of compounds 14a and 14b employed alternate
Suzuki-Miyaura coupling conditions utilizing S-phos ligand and Pd
(OAc)2 to combine the chiral boronate esters40 with the vinyl triflate 24
(Scheme 5). The C-17 amine 36 was alkylated with 1-bromo-2-chloro-
ethane using K3PO4 in CH3CN at 120 ◦C in a sealed tube to provide
the aziridine 37 which was subsequently opened with 4-(methyl-
sulfonyl)piperidine to give the ethyl-linked methylsulfonyl piperidine
38. The cyclohexene esters were hydrolyzed to the corresponding car-
boxylic acids 14a and 14b using aqueous NaOH (Scheme 6).
1 (a) Lingappa JR, Reed JC, Tanaka M, Chutiraka K, Robinson BA. Virus Res. 2014;
193:89–107. https://doi.org/10.1016%2Fj.virusres.2014.07.001.(b) Bell, Neil M,
Lever AML. Trends Microbiol. 2013;21:136–144. https://doi.org/10.1016/j.
tim.2012.11.006.(c) Lee SK, Potempa M, Swanstrom R. J Biol Chem. 2012;287.
https://doi.org/10.1074/jbc.R112.399444, 867–840,874.
¨
2 Sundquist WI, Krausslich HG. Cold Spring Harb Perspect Med. 2012;7, a006924.
https://doi.org/10.1101/cshperspect.a006924.
3 Li F, Goila-Gaur R, Salzwedel K, et al. Proc Natl Acad Sci USA. 2003;100:
13555–13560. https://doi.org/10.1073/pnas.2234683100.
4 Zhou J, Yuan X, Dismuke D, et al. J Virol. 2004;78:922–929. https://doi.org/
10.1128/JVI.78.2.922-929.2004.
5 (a) Smith PF, Ogundele A, Forrest JW, et al. Antimicrob Agents Chemother. 2007;51:
3574–3581. https://doi.org/10.1128/AAC.00152-07.(b) Martin DE, Blum R,
Wilton J, et al. Antimicrob Agents Chemother. 2007;51:3063–3066. https://doi.org/
10.1128/AAC.01391-06.
Declaration of Competing Interest
6 McCallister S, Lalezari J, Richmond G, et al. Antiviral Ther. 2008;13:A10.
7 Van Baelen KV, Salzwedel K, Rondelez E, et al. Antimicrob Agents Chemother. 2009;53:
2185–2188. https://doi.org/10.1128/AAC.01391-0610.1128/AAC.01650-08.
8 (a) Margot NA, Gibbs CS, Miller MD. Antimicrob Agents Chemother. 2010;54:
2345–2353. https://doi.org/10.1128/AAC.01784-09.(b) Knapp DJHF, Harrigan RP,
Poon AFY, Brumme ZL, Brockman M, Cheung PK. J Clin Microbiol. 2011;49:201–208.
https://doi.org/10.1128/JCM.01868-10.(c) Dicker I, Zhang S, Ray N, et al. PLoS
One. 2019;14, e0224076. https://doi.org/10.1371/journal.pone.0224076.
9 Liu Z, Swidorski JJ, Nowicka-Sans B, et al. Bioorg Med Chem. 2016;24:1757–1770.
https://doi.org/10.1016/j.bmc.2016.03.001.
The authors declare that they have no known competing financial
interests or personal relationships that could have appeared to influence
the work reported in this paper.
Acknowledgement
The authors gratefully acknowledge the bioanalytical group, Qi Gao
and Amy Sarjeant for their X-ray crystallography support and Stella
Huang for her provision of spectroscopic data. In addition, we would like
to recognize Nanjundaswamy Kanikahalli Chikkananjunda, Duraisamy
10 Nowicka-Sans B, Protack T, Lin Z, et al. Antimicrob Agents Chemother. 2016;60:
3956–3969. https://doi.org/10.1128/AAC.02560-15.
11 Regueiro-Ren A, Liu Z, Chen Y, et al. ACS. Med Chem Lett. 2016;7:568–572. https://
doi.org/10.1021/acsmedchemlett.6b00010.
8

J.J. Swidorski et al.
Bioorganic & Medicinal Chemistry Letters 36 (2021) 127823
12 Regueiro-Ren A, Swidorski JJ, Liu Z, et al. J Med Chem. 2018;61:7289–7313. https://
doi.org/10.1021/acs.jmedchem.8b00854.
26 Bharat TAM, Menendez LRC, Hagen WJH, et al. Proc Natl Acad Sci USA. 2014;111:
8233–8238. https://doi.org/10.1073/pnas.1401455111.
13 Hwang C, Schurmann D, Sobotha C, et al. Clin Infect Dis. 2017;65:442–452. https://
doi.org/10.1093/cid/cix239.
27 Han Y, Hou G, Suiter CL, et al. J Am Chem Soc. 2013;135. https://doi.org/10.1021/
ja406907h, 793–717,803.
14 Morales-Ramirez J, Bogner JR, Molina J-M, et al. PLoS One. 2018;13. https://doi.
org/10.1371/journal.pone.0205368.
28 Bayro MJ, Ganser-Pornillos BK, Zadrozny KK, Yeager M, Tycko R. J Am Chem Soc.
2016;138:12029–12032. https://doi.org/10.1021/jacs.6b07259.
29 Gupta S, Louis JM, Tycko R. Proc Natl Acad Sci USA. 2020;117:10286–10293.
https://doi.org/10.1073/pnas.1917755117.
15 (a) Lu W, Salzwedel K, Wang D, et al. Antimicrob Agents Chemother. 2011;55:3324.
´
´
(b) Seclen E, Gonzalez M, Corral A, Mendoza C, Soriano V, Poveda E. AIDS. 2010;24:
467–469. https://doi.org/10.1097/QAD.0b013e328335ce07.
16 Dang Z, Ho P, Zhu L, et al. J Med Chem. 2013;56:2029–2037. https://doi.org/
10.1021/jm3016969.
30 Goh BC, Perilla JR, England MR, Heyrana KJ, Craven RC, Schulten K. Structures.
2015;23:1414–1425. https://doi.org/10.1016/j.str.2015.05.017.
31 Tomasini MD, Johnson DS, Mincer JS, Simon SM. PLoS One. 2018;13, e0196133.
https://doi.org/10.1371/journal.pone.0196133.
17 Swidorski JJ, Liu Z, Sit S-Y, et al. Bioorg Med Chem Lett. 2016;26:1925–1930. https://
doi.org/10.1016/j.bmcl.2016.03.019.
32 Dick RA, Zadrozny KK, Xu C, et al. Nature. 2018;560:509–512.
´
18 Chen Y, Sit S-Y, Chen J, et al. Bioorg Med Chem Lett. 2018;28:1550–1557. https://doi.
org/10.1016/j.bmcl.2018.03.067.
33 Mallery DL, Marquez CL, McEwan WA, Dickson CF, et al. eLife. 2018. https://doi.
org/10.7554/eLife.35335.
¨
19 Jung ME, Piizzi G. Chem Rev. 2005;105:1735–1766. https://doi.org/10.1021/
cr940337h.
34 (a) Obr M, Krausslich H-G. eLife. 2018. https://doi.org/10.7554/eLife.38895.(b)
Kleinpeter AB, Freed EO. Viruses. 2020;12:940. https://doi.org/10.3390/
v12090940.
20 University of Wisconsin pKa Data Compilation by R. Williams. https://www.chem.
wisc.edu/areas/reich/pkatable/pKa_compilation-1-Williams.pdf [Accessed online
May 08, 2020].
35 Purdy MD, Shi D, Chrustowicz J, Hattne J, Gonen T, Yeager M. Proc Natl Acad Sci
USA. 2018;115:13258–13263. https://doi.org/10.1073/pnas.1806806115.
36 Kinch MS, Patridge E. Drug Discov Today. 2014;19:1510–1513. https://doi.org/
10.1016/j.drudis.2014.05.012.
21 Gunaydin H, Bartberger MD. ACS Med Chem Lett. 2016;7:341–344. https://doi.org/
10.1021/acsmedchemlett.6b00099.
22 The X-ray crystal structure coordinates have been deposited in the or Cambridge
Crystallographic Data Centre as CCDC 2031147.
37 Miyaura N, Ishiyama T, Sasaki H, Ishikawa M, Sato M, Suzuki A. J Am Chem Soc.
1989;111:314–321. https://doi.org/10.1021/ja00183a048.
38 Radchenko DS, Grygorenko OO, Komarov IV. Tetrahedron Asymmetry. 2008;19:
2924–2930.
23 Wagner JM, Zadrozny KK, Chrustowicz J, et al. eLife. 2016. https://doi.org/10.7554/
eLife.17063.
24 Schur FKM, Obr M, Hagen WJH, et al. Science. 2016;353:506–508. https://doi.org/
10.1126/science.aaf9620.
39 Takagi J, Takahashi K, Ishiyama T, Miyaura N. J Am Chem Soc. 2002;124:
8001–8006. https://doi.org/10.1021/ja0202255.
´
25 Schur FKM, Hagen WJH, Rumlova M, et al. Nature. 2015;517:505–508. https://doi.
40 Swidorski J, Meanwell NA, Regueiro-Ren A, Sit S-Y, Chen J, Chen, Y. US 8,906,889
B2. Dec. 09, 2014.
org/10.1038/nature13838.
9