5-Aminothiophene-2,4-dicarboxamide analogues as hepatitis B virus capsid assembly effectors

Article abstract of DOI:10.1016/j.ejmech.2018.12.047

Chronic hepatitis B virus (HBV) infection represents a major health threat. Current FDA-approved drugs do not cure HBV. Targeting HBV core protein (Cp) provides an attractive approach toward HBV inhibition and possibly infection cure. We have previously identified and characterized a 5-amino-3-methylthiophene-2,4-dicarboxamide (ATDC) compound as a structurally novel hit for capsid assembly effectors (CAEs). We report herein hit validation through studies on absorption, distribution, metabolism and excretion (ADME) properties and pharmacokinetics (PK), and hit optimization via analogue synthesis aiming to probe the structure-activity relationship (SAR) and structure-property relationship (SPR). In the end, these medicinal chemistry efforts led to the identification of multiple analogues strongly binding to Cp, potently inhibiting HBV replication in nanomolar range without cytotoxicity, and exhibiting good oral bioavailability (F). Two of our analogues, 19o (EC₅₀ = 0.11 μM, CC₅₀ > 100 μM, F = 25%) and 19k (EC₅₀ = 0.31 μM, CC₅₀ > 100 μM, F = 46%), displayed overall lead profiles superior to reported CAEs 7–10 used in our studies.

Full text of DOI:10.1016/j.ejmech.2018.12.047

Accepted Manuscript
5-Aminothiophene-2,4-dicarboxamide analogues as hepatitis B virus capsid assembly
effectors
Jing Tang, Andrew D. Huber, Dallas L. Pineda, Kelsey N. Boschert, Jennifer J. Wolf,
Jayakanth Kankanala, Jiashu Xie, Stefan G. Sarafianos, Zhengqiang Wang
PII:
S0223-5234(18)31084-5
DOI:
https://doi.org/10.1016/j.ejmech.2018.12.047
EJMECH 10981
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 5 November 2018
Revised Date: 6 December 2018
Accepted Date: 19 December 2018
Please cite this article as: J. Tang, A.D. Huber, D.L. Pineda, K.N. Boschert, J.J. Wolf, J. Kankanala, J.
Xie, S.G. Sarafianos, Z. Wang, 5-Aminothiophene-2,4-dicarboxamide analogues as hepatitis B virus
capsid assembly effectors, European Journal of Medicinal Chemistry (2019), doi: https://doi.org/10.1016/
j.ejmech.2018.12.047.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT
Graphical abstract

ACCEPTED MANUSCRIPT
5
-Aminothiophene-2,4-dicarboxamide Analogues as Hepatitis B
Virus Capsid Assembly Effectors
#
1
# 2,3
2,4
2,5
Jing Tang , Andrew D. Huber
, Dallas L. Pineda , Kelsey N. Boschert , Jennifer J. Wolf
2
,6
1
1
2,4,6
1
,
Jayakanth Kankanala , Jiashu Xie , Stefan G. Sarafianos
and Zhengqiang Wang*
1
Center for Drug Design, College of Pharmacy, University of Minnesota, Minneapolis, MN 55455, USA
2
Christopher S. Bond Life Sciences Center, University of Missouri, Columbia, Missouri, 65211, USA
3
Department of Veterinary Pathobiology, College of Veterinary Medicine, University of Missouri, Columbia, Missouri, 65211, USA
4
Department of Biochemistry, University of Missouri, Columbia, Missouri, 65211, USA
5
Department of Nutrition and Exercise Physiology, University of Missouri, Columbia, Missouri, 65211, USA
6
Department of Molecular Microbiology & Immunology, School of Medicine, University of Missouri, Columbia, Missouri, 65211, USA
Abstract
Chronic hepatitis B virus (HBV) infection represents a major health threat. Current FDA-
approved drugs do not cure HBV. Targeting HBV core protein (Cp) provides an attractive
approach toward HBV inhibition and possibly infection cure. We have previously identified and
characterized a 5-amino-3-methylthiophene-2,4-dicarboxamide (ATDC) compound as a
structurally novel hit for capsid assembly effectors (CAEs). We report herein hit validation
through studies on absorption, distribution, metabolism and excretion (ADME) properties and
pharmacokinetics (PK), and hit optimization via analogue synthesis aiming to probe the
structure-activity relationship (SAR) and structure-property relationship (SPR). In the end, these
medicinal chemistry efforts led to the identification of multiple analogues strongly binding to Cp,
potently inhibiting HBV replication in nanomolar range without cytotoxicity, and exhibiting
good oral bioavailability (F). Two of our analogues, 19o (EC = 0.11 µM, CC >100 µM, F =
5
0
50
2

ACCEPTED MANUSCRIPT
2
5%) and 19k (EC = 0.31 µM, CC >100 µM, F = 46%), displayed overall lead profiles
5
0
50
superior to reported CAEs 7–10 used in our studies.
Introduction
HBV chronically infects an estimated 250 million people worldwide and remains a major health
threat [1]. Despite a successful vaccine, there are still around 1 million people newly infected
each year. Chronic HBV infection is often associated with severe liver diseases which typically
progress through fibrosis, cirrhosis and eventually develop into hepatocellular carcinoma (HCC)
[2]. It is estimated that HBV-associated liver diseases result in approximately 660,000 deaths
annually. Current FDA-approved HBV drugs include the immunomodulatory agent pegylated
interferon alpha (IFN-α) [3] and direct acting nucleos(t)ide analogues (NAs) which target
reverse transcriptase (RT) of the HBV P protein complex [4]. These NAs are based on all four
endogenous nucleosides with distinct approaches with respect to mimicking the ribose moiety
(
Figure 1), including β-L-nucleosides [5-6] lamivudine (1, 3TC, analogue of dC) and telbivudine
2, LdT, analogue of T), the carbocyclic nucleoside entecavir (3, ETV, analogue of dG) [7], the
(
acyclic nucleoside phosphonate (ANPs) [8] adefovir (4, ADV, analogue of dA) and two forms of
tenofovir (TFV, analogue of dA): tenofovir disoproxil fumarate (5, TDF) and tenofovir
alafenamide (6, TAF). While NAs are generally well tolerated [9] and largely effective [10-11]
in suppressing viral load, they do not cure HBV infection presumably because they do not
eliminate HBV covalently closed circular DNA (cccDNA) which is the main viral reservoir for
persistent infection [12]. HBV cure calls for novel antiviral approaches to completely eliminate
or functionally inactivate cccDNA [13-15]. Otherwise, clinical management of chronic HBV
infection requires lifetime treatment with current drugs.
3

ACCEPTED MANUSCRIPT
Figure 1. FDA-approved NAs for treating chronic HBV infection. 1: lamivudine (3TC); 2:
telbivudine (LdT); 3: entecavir (ETV); 4: adefovir (ADV); 5: tenofovir disoproxil fumarate
(TDF); and 6: tenofovir alafenamide (TAF).
The HBV replication cycle [16] entails a critical step where Cp dimers are assembled into
functional viral capsids which encapsulate both viral pregenomic RNA (pgRNA) and polymerase
complex for active reverse transcription to generate genomic, partially double-stranded relaxed
circular DNA (RC-DNA). Interestingly, RC-DNA-containing mature nucleocapsids are either
enveloped then secreted as new infectious virions, or enter the nucleus to replenish the cccDNA
pool, constituting the intracellular amplification pathway of cccDNA. Disrupting the proper
assembly of capsid can therefore inhibit HBV by blocking the production of infectious viruses
and depleting cccDNA replenishment, and potentially contribute to HBV cure [17]. Therefore,
targeting HBV Cp represents an attractive approach for treating chronic HBV infection. A few
chemotypes have been reported as potent CAEs (Figure 2) [14, 18-19]. Amongst these, the
heteroaryldihydropyrimidine (HAP) [20-24] chemotype as represented by compound 7 and the
4

ACCEPTED MANUSCRIPT
sulfamoylbenzamide (SBA) chemotype [25-26] as represented by compound 8 are particularly
well studied with analogues in clinical development [27-29]. Mechanistically, all known CAEs
accelerate capsid assembly, albeit with different assembled products. The HAP series constitutes
a class of its own by inducing misassembly to form aberrant nonfunctional capsid particles,
whereas SBA and other chemotypes, such as phenylpropenamide (9) [30] and NZ-4 (10) [31],
promote the formation of empty, morphologically normal capsids [32].
Figure 2. Major CAE chemotypes reported. 7 (BAY-7690) represents the HAP chemotype; 8
(DVR-56) represents the SBA chemotype.
Recently we conducted a high-throughput screening (HTS) of commercial libraries in a thermal
shift assay (TSA) that measures binding to HBV Cp and identified compound 11 as a strong
binder (Figure 3) [33]. In the subsequent antiviral assay 11 inhibited HBV total DNA production
(
EC = 3.4 µM) with no cytotoxicity observed (CC >100 µM). Mechanistically, compound 11
50 50
promoted the formation of large Cp aggregates and prevented Cp from entering nucleus [33].
These results confirmed 11 as a valid CAE hit. We report herein the optimization of CAE hit 11
through analogue synthesis, SAR as well as ADME and PK studies.
5

ACCEPTED MANUSCRIPT
Figure 3. Our CAE hit. Compound 11 strongly bound to Cp, inhibited viral DNA production
without cytotoxicity, and disrupted capsid assembly. 11 also demonstrated favorable ADME
properties in vitro and decent oral bioavailability in vivo. These properties, combined with its
low MW and high LE, render 11 a high quality lead.
Results and Discussion
In the present work, compound 11 was first evaluated in various in vitro assays for its
physicochemical and ADME properties, including aqueous solubility and stability, plasma
stability, plasma protein binding and microsomal stability. Overall, compound 11 exhibited
favorable ADME properties (Figure 3) predicting good oral bioavailability. This was confirmed
through in vivo PK studies in mice from which the oral bioavailability of 11 was determined to
be 31%. In addition, compound 11 has relatively low molecular weight (MW) and a good ligand
efficiency (LE), two highly desired attributes that bode well with lead optimization. Based on
these, 11 was considered a high quality CAE lead.
Analogue design and synthesis. To optimize hit 11, analogues were designed and synthesized
to explore the SAR around the central thiophene ring (Figure 4). Specifically, analogue synthesis
includes subtype 12 which aims to study the SAR of the C5 amino group; subtypes 13—16
which concern the SAR around the C4 carboxamide moiety; and subtype 19 which probes the
6

ACCEPTED MANUSCRIPT
phenyl ring of the C2 carboxamide. In addition, 4,5-cyclized analogues (subtype 17) and 2,3-
cyclized analogues (subtype 18) were also synthesized (Figure 4) to explore the impact of
molecular rigidity.
Figure 4. SAR design for hit 11. Sites for structural variation include the C5 amino group
(
subtype 12), the C4 carboxamide moiety (subtypes 13—16), and the C2 carboxamide phenyl
ring (subtype 19). In addition, 4,5-cyclized analogues (subtype 17) and 2,3-cyclized analogues
subtype 18) are also designed.
(
Synthetically, all these analogues can be easily accessed via different versions of the Gewald
synthesis [34]. This synthesis features a three-component one-pot reaction to construct the
thiophene ring with the desired substitution pattern (Scheme 1) [35-36].
a
Scheme 1 Analogue synthesis for hit 11
7

ACCEPTED MANUSCRIPT
a
Reagents and conditions: i) toluene, reflux, t-BuOK, 40-63%; ii) morpholine, S , EtOH, reflux,
8
8
9-53%; iii) substituted carboxylic acid, PyAOP, DIEA, DMF, rt, overnight; or substituted acid
3
chloride, DMAP, Et N, DMF, rt, overnight, 54% -81%; iv) appropriate amine, EtOH, reflux, 45-
7
0 %; v) RCN, dioxane, HCl, refluxing, overnight, 56-85%.
Biology and SAR. All analogues were first evaluated in our recently developed TSA [33] which
is not quantitative but does provide effective monitoring of HBV capsid assemble states. From
this assay, hits with significant binding, as observed from the TSA curve (see Figure 5A for an
example), were then tested in an antiviral assay which measures total intracellular HBV DNA
8

ACCEPTED MANUSCRIPT
production. Compounds with significant antiviral activity were further evaluated in a cytotoxicity
assay.
SAR around C-5 amino group. This domain of SAR concerns mainly the effect of derivatizing
the C-5 amino group. Toward this end, the amino group was converted into urea, aliphatic
amides, aromatic amides and tetrazole. In the binding assay, the urea analogue (12a) and the
tetrazole compound (12k) were completely inactive and aliphatic amides (12b-c) showed weak
Cp binding (Table 1). By contrast, strong binding was preserved for all aromatic amides (12d-j),
with the exception of compound 12h (Table 1). In addition, all aromatic amides (12d-j) remained
potent in the antiviral assay with EC50 values (1.4-4.7 µM) comparable to that of the lead
compound 11 (EC50 = 3.4 µM). These results suggest that acylating the C-5 amino group is not
tolerated with an aliphatic acyl group, and tolerated or even moderately beneficial with an
aromatic acyl group. Particularly interesting is compound 12e which exhibited improved
antiviral potency (EC = 1.4 µM). However, this moderate potency improvement was conferred
5
0
by the addition of a relatively large group, and hence may reflect a reduced ligand efficiency.
Table 1. SAR around C-5 amino group.
a
b
c
Compound
Structure
TSA
EC50 (µM) CC50 (µM)
1
1
+++
3.4 ± 0.6
>100
9

ACCEPTED MANUSCRIPT
d
1
2a
-
NT
NT
1
2b
+
+
~10
>10
>100
>100
1
2c
1
2d
+++
2.7 ± 1.9
49.5 ± 0.1
1
2e
+++
+++
1.4 ± 0.02
4.7 ± 0.4
>100
>100
1
2f
1
2g
+++
+
4.1 ± 2.2
2.9 ± 0.7
>100
>100
1
2h
1
0

ACCEPTED MANUSCRIPT
12i
+++
2.1 ± 0.3
>100
1
2j
+++
3.5 ± 2.1
>100
NT
1
2k
-
NT
a
Thermal shift assay that measures changes in the thermal stability of HBV Cp upon binding of a
small molecule: “−” denotes no binding, “+” weak binding, “++” moderate binding and “+++”
strong binding.
b
Concentration of a compound inhibiting HBV total DNA production by 50%, expressed as the
mean ± standard deviation from at least two independent experiments.
c
Concentration of a compound causing 50% cytotoxicity, expressed as the mean ± standard
deviation from at least two independent experiments.
d
NT = not tested
SAR around C-4 carboxamide. In this SAR domain, the C-4 carboxamide group was first
changed to three different bioisosteres: the thiocarboxamide, cyano and ester (Table 2). Strong
binding to Cp was retained with the thiocarboxamide (13) and the esters (15a-b), though no
antiviral activity was observed with either 13 or 15a at concentrations up to 10 µM, and 15b
exhibited reduced potency (EC = 6.1 µM). In the meantime, replacing carboxamide with cyano
5
0
also led to the loss of binding affinity to Cp (14a and 14d). The binding was restored by
introducing substituents at the para and meta positions of the phenyl ring on the left (14b-c),
with analogue 14c demonstrating slightly improved antiviral potency compared to lead 11. These
results suggest that thiocarboxamide, cyano and ester groups at C-4 are either not tolerated or
1
1

ACCEPTED MANUSCRIPT
inferior to the carboxamide group. Further SAR in this domain then focused on the substitution
of amide through the synthesis of analogues 16a-m (Table 2). When the amide is substituted
with a phenyl ring, the resulting analogues all showed strong Cp binding (16a-f) with the single
exception of compound 16g. Most of these analogues (16a-d) also inhibited HBV DNA
production with potency similar to lead 11. Of particular interest is analogue 16f, which
exhibited antiviral potency in the nanomolar range (EC50 = 0.68 µM), a major potency
improvement over lead 11. However, phenyl substitution on the amide also resulted in moderate
cytotoxicity (CC50 = 17-40 µM), rendering these analogues less attractive than lead 11.
Interestingly, when a methylene group is inserted between the phenyl substituent and the amide,
the resulting analogues (16h-i) were rendered completely inactive in the antiviral assay,
suggesting that alkyl substitution may not be tolerated. This is confirmed through the synthesis
of alkyl and cycloalkyl substituted analogues (16j-m) which showed no binding affinity to Cp
and no antiviral activity at concentrations up to 10 µM. The only exception is 16m, which
retained strong Cp binding without conferring antiviral activity. Based on all these results,
unsubstituted carboxamide as seen in lead 11 is preferred over all other variants.
Table
2.
SAR
around C-4 carboxamide.
a
b
c
Compound
Structure
TSA
EC50
(µM)
CC50 (µM)
1
1
+++
3.4 ± 0.6
>100
1
2

ACCEPTED MANUSCRIPT
1
3
+++
-
>10
>100
NT
d
1
4a
NT
1
4b
+++
~10
>100
1
4c
+++
2.0 ± 0.2
96 ± 10
1
4d
-
NT
>10
NT
1
5a
+++
+++
76 ± 1.8
>100
1
5b
6.1 ± 2.8
1
6a
+++
+++
+++
3.2 ± 0.3
2.4 ± 0.2
3.4 ± 0.3
>100
1
6b
40 ± 5.1
26 ± 2.7
1
6c
1
3

ACCEPTED MANUSCRIPT
1
6d
++
+++
+++
-
3.6 ± 1.6
>10
17 ± 1.3
NT
1
6e
1
6f
0.68 ± 0.01
40 ± 2.6
NT
1
6g
NT
1
6h
+++
>10
>10
>10
>10
>10
NT
NT
NT
NT
NT
O
NH
O
N
1
6i
+++
H
OMe
S
NH2
1
6j
-
-
-
1
6k
1
6l
16m
+++
>10
NT
1
4

ACCEPTED MANUSCRIPT
a
Thermal shift assay that measures changes in the thermal stability of HBV Cp upon binding of a
small molecule: “−” denotes no binding, “+” weak binding, “++” moderate binding and “+++”
strong binding.
b
Concentration of a compound inhibiting HBV total DNA production by 50%, expressed as the
mean ± standard deviation from at least two independent experiments.
c
Concentration of a compound causing 50% cytotoxicity, expressed as the mean ± standard
deviation from at least two independent experiments.
d
NT = not tested
SAR with C-4, C-5 cyclization and C-2, C-3 cyclization. This part of the SAR mainly probes the
impact of added rigidity to the chemotype. Since binding events typically result in entropy loss
that needs to be compensated, rigidifying a molecule could potentially benefit target binding by
reducing the required entropy compensation. Toward this end, a few C-4, C-5 cyclized analogues
(17a-d) and C-2, C-3 cyclized analogues (18a-c) were synthesized (Table 3). Unfortunately,
none of these analogues showed any binding affinity toward Cp (Table 3), suggesting that the
attempted cyclization is not tolerated.
Table 3. SAR with C-4, C-5-cyclized analogues and C-2, C-3-cyclized analogues.
a
b
c
Compound
Structure
TSA
EC50
(µM)
CC50 (µM)
1
1
+++
3.4 ± 0.6
>100
NT
d
1
7a
-
NT
1
5

ACCEPTED MANUSCRIPT
1
7b
-
-
NT
NT
NT
NT
1
7c
1
7d
-
-
NT
NT
NT
NT
1
8a
1
8b
-
-
NT
NT
NT
NT
1
8c
a
Thermal shift assay that measures changes in the thermal stability of HBV Cp upon binding of a
small molecule: “−” denotes no binding, “+” weak binding, “++” moderate binding and “+++”
strong binding.
b
Concentration of a compound inhibiting HBV total DNA production by 50%, expressed as the
mean ± standard deviation from at least two independent experiments.
c
Concentration of a compound causing 50% cytotoxicity, expressed as the mean ± standard
deviation from at least two independent experiments.
d
NT = not tested
SAR around the phenyl ring of C-2 carboxamide. This domain of SAR began with the synthesis
of analogue 19a. With the addition of a flexible methylene group between the amide and the
phenyl ring, strong Cp binding was preserved while antiviral potency dropped considerably
1
6

ACCEPTED MANUSCRIPT
(Table 4), suggesting that direct phenyl substitution is preferred at the C-2 carboxamide. The
remaining SAR then focused on exploring the effects of substituents at the para and meta
positions of the phenyl ring. Toward this end, compounds with a mono substituent at the para
position (19b-f) were first synthesized. Among them, only the analogue with a para F group
(19e) retained both strong Cp binding and antiviral potency when compared to lead 11. Cp
binding was preserved also with para OMe substituted compound (19b), albeit with much
reduced antiviral activity. None of the other para substituted analogues showed binding affinity
toward Cp (Table 4). These SAR results suggest that mono substitution at the para position is
generally undesired, with only the F group tolerated. The next series of compounds consists of
analogues with a para F group and a meta substituent (19g-h, 19k-m and 19o). The addition of a
meta substituent drastically improved the activity profile, as all analogues showed strong Cp
binding, markedly improved antiviral potency (EC50 = 0.11-1.2 µM), and no cytotoxicity at
concentrations up to 100 µM (Table 4). Interestingly, when the para F group is replaced with a
Cl or Br group, the resulting analogues were much less potent (19i-j vs 19h, 19n vs 19m). In
addition, when a second meta position is also substituted, the resulting tri-substituted analogue
(19p) exhibited significantly reduced antiviral potency compared to the di-substituted analogues.
These results indicate that a para F group in combination with one meta substituent on the
phenyl ring confers the optimal antiviral potency.
Table 4. SAR around C-2 carboxamide phenyl ring.
a
b
c
Compound
Structure
TSA
EC50
(µM)
CC50 (µM)
1
7

ACCEPTED MANUSCRIPT
1
1
+++
+++
3.4 ± 0.6
~10
>100
d
1
9a
NT
1
9b
+++
~10
>10
NT
NT
1
9c
-
1
9d
-
+++
-
NT
NT
1
9e
3.0 ± 0.7
NT
>100
NT
1
9f
1
9g
+++
0.76 ± 0.22
0.21 ± 0.09
1.7 ± 0.4
>100
>100
>100
1
9h
+++
O
Cl
NH
NH2
1
9i
+++
O
S
NH2
1
8

ACCEPTED MANUSCRIPT
O
Br
NH
NH2
1
9j
+++
+++
7.1 ± 1.7
107 ± 7.2
>100
O
S
NH2
1
9k
0.31 ± 0.07
1
9l
+++
+++
0.47 ± 0.25
1.2 ± 0.03
>100
>100
1
9m
1
9n
+++
+++
+++
4.6 ± 0.4
0.11 ± 0.01
4.6 ± 1.1
72 ± 6.4
>100
1
9o
1
9p
72 ± 4.1
a
Thermal shift assay that measures changes in the thermal stability of HBV Cp upon binding of a
small molecule: “−” denotes no binding, “+” weak binding, “++” moderate binding and “+++”
strong binding.
b
Concentration of a compound inhibiting HBV total DNA production by 50%, expressed as the
mean ± standard deviation from at least two independent experiments.
c
Concentration of a compound causing 50% cytotoxicity, expressed as the mean ± standard
deviation from at least two independent experiments.
d
NT = not tested
Overall, our SAR efforts led to the identification of five analogues with strong Cp binding,
nanomolar antiviral potency and no cytotoxicity (Table 5). To benchmark our CAEs, we also
1
9

ACCEPTED MANUSCRIPT
resynthesized representative compounds (7-10) of all four known CAE chemotypes (Figure 2)
and tested them in our assays. As shown in Table 5, AT-61 (9) and NZ-4 (10) were at least one
order of magnitude less potent than our CAEs, while HAP analogue 7 and SBA analogue 8
exhibited antiviral potency largely comparable to that of our nanomolar CAEs (Table 5). In
addition, significant cytotoxicity was observed with all but one control compounds (8-10, CC =
5
0
1
7-61 µM), whereas our CAEs of subtype 19 showed no cytotoxicity at concentrations up to 100
µM. Direct comparison between HAP analogue 7 and our nanomolar CAEs revealed that our
best compound 19o was about four times as potent as 7.
Table 5. Summary of nanomolar CAEs of the ATDC series
Compound
9g
TSA
EC (
µ
M)
CC (µM)
50
50
1
+++
0.76 ± 0.22
0.21 ± 0.09
0.31 ± 0.07
0.47 ± 0.25
0.11 ± 0.01
0.41 ± 0.05
0.25 ± 0.07
4.8 ± 0.6
>100
>100
1
1
9h
9k
+++
+++
+++
+++
+++
+++
+++
+++
>100
1
9l
9o
7
>100
1
>100
>100
8
17 ± 4.4
61 ± 25
29 ± 10
9
1
0
10 ± 0.5
To confirm the mechanism of action our novel CAEs, a selected compound 19h was further
investigated using TEM for its effect on capsid assembly. As shown in Figure 5C, comparing to
the DMSO control, 19h promoted the aggregation of HBV capsid, an effect similar to that
2
0

ACCEPTED MANUSCRIPT
observed with hit molecule 11 [33]. The impact of 19h on capsid assembly is also reflected in the
TSA curve (Figure 5A). The TSA was conducted under conditions to allow two denaturation
peaks corresponding to the Cp dimer and the assembled capsid, respectively (DMSO control).
However, in the presence of 19h, only the denaturation peak corresponding to fully assemble
capsid was observed, corroborating the mechanism that 19h confers potent antiviral activity
(
Figure 5B) by promoting capsid assembly.
Figure 5. Characterization of compound 19h. (A) TSA curve (dotted line). Solid line depicts the
curve of DMSO control; (B) dose-response antiviral testing; (C) TEM images of HBV capsid.
HBV capsid aggregation observed in the presence of compound 19h as compared to the DMSO
control.
Physicochemical properties, in vitro ADME and in vivo PK. To assess drug-like properties of
our best CAE subtype 19, analogues with nanomolar antiviral activity (19g-h, 19k-l and 19o)
were tested in various physicochemical and in vitro ADME studies (Table 6). First, aqueous
2
1

ACCEPTED MANUSCRIPT
stability and solubility were evaluated in Dulbecco's Phosphate-Buffered Saline (DPBS). While
all tested compounds were highly stable in DPBS, a few of them (19g, 19l and 19o) exhibited
poor aqueous solubility. Second, all tested compounds were stable in both human and mouse
plasma. Third, plasma protein binding was in a reasonable range for all compounds with higher
binding observed in human plasma (80-94%) than in mouse plasma (50-90%). Fourth, all
compounds exhibited excellent phase I and phase II microsomal metabolic stability, with the
exception of 19h which demonstrated greater susceptibility to phase I NADPH-dependent
metabolism particularly in mouse liver microsomes. Finally, two compounds were also tested in
Caco-2 cell line for permeability, and the obtained high apparent permeability coefficient (Papp >
−6
1
0×10
cm/s) suggested a good in vitro drug absorption with minimal efflux
(
Papp[B→A]/Papp[A→B < 2). Taken together, except for the poor solubility of a few analogues, our
]
potent CAEs possess favorable physicochemical and ADME properties.
a
Table 6. Physicochemical, in vitro ADME , and oral bioavailability profile of selected analogues
Plasma Protein
Microsomal Stability
d
−6
Plasma Stability
t1/2 (h)
c
Caco-2 Papp (10
Aqueous
g
Binding
CLint
F
cm/s)
Compd Solubility
b
(
%)
(Phase I / Phase II)
a
(%)
(
µM)
e
f
A to B
B to A
-
Human Mouse Human Mouse
Human
Mouse
1
1
19
> 24
> 24
86
50
5.0 / 2.0
18 / 4.4
-
34
19g
19h
19k
11
31
31
> 24
> 24
> 24
> 24
> 24
> 24
86
91
94
80
82
90
3.4 / <0.1
14 / <0.1
7.6 / < 0.1
25 / 5.2
92 / 5.2
28 / 1.6
24.8
18.6
-
27.6
26.9
-
-
11
46
1
9l
6.7
9.5
> 24
> 24
> 24
> 24
80
89
69
82
2.4 / 0.1
5.8 / 3.2
4.2 / 2.0
22 / 10
-
-
-
-
-
1
9o
25
a
Aqueous stability of selected analogues were determined in Dulbecco's Phosphate-Buffered Saline (DPBS). All
analogues showed excellent stability with remaining percentage at 24 h > 90%. Percent of fraction bound. CLint
intrinsic clearance, µl/min/mg protein. Papp: The apparent permeability coefficient. A: apical side. B: basolateral
b
c
:
d
e
f
g
side
. F: Oral bioavailability in mice.
Based on the relatively high solubility and good cell permeability as well as the superb antiviral
potency, analogue 19h was initially selected for in vivo PK studies. Unfortunately, the oral
2
2

ACCEPTED MANUSCRIPT
bioavailability was found to be quite low (F=11%), which likely reflects the poor metabolic
stability of 19h. Upon further examination, the main structural difference between compound
1
9h and all other potent CAEs is the additional methyl group at the meta position of the phenyl
ring (Figure 6). We hypothesized that this metabolically labile methyl group could be the culprit
for low bioavailability. Therefore, we decided to replace this methyl group with a metabolically
more stable -CF bioisostere (19m, Figure 6). Unfortunately, this replacement resulted in a six
3
fold drop in antiviral potency (Table 4, 19m vs 19h), presumably due to the complete loss of the
C-H bond which can hyperconjugate using its σ* orbital with a π system in the protein. Such a
hyperconjugation forms the molecular basis for the well-known magic methyl effect [37]. To
restore the ability for hyperconjugation, the -CF group was changed to -CHF group and the
3
2
resulting analogue 19o showed not only substantially improved oral bioavailability (F = 25%),
but also enhanced antiviral potency (EC50 = 0.11 µM, Figure 6). The improved bioavailability
correlated well with increased phase I metabolic stability as the intrinsic clearance for 19o falls
within the low to medium range of clearance according to literature[38] in both human and
mouse liver microsomes (Table 6 and Figure 6).
Figure 6. Lead optimization from 19h to 19o. The highlighted -CH group in 19h is a metabolic
3
3 3
liability which likely accounts for the low F. Replacing the -CH group with a -CF group
2
3

ACCEPTED MANUSCRIPT
substantially reduced antiviral potency (19m), whereas a -CHF substituent enhanced antiviral
2
activity while improving metabolic stability (19o).
Lastly, animal PK studies also included analogue 19k, which displayed good balance between
antiviral potency (EC50 = 0.31 µM) and ADME profile, particularly solubility and microsomal
stability. The oral bioavailability of 19k was determined to be 46% (Table 6), the highest among
analogues tested. This observation further confirms that solubility and metabolic stability are key
factors in achieving good oral bioavailability.
Figure 7. Molecular modeling of 19o. (A) Predicted binding mode of 19o within the crystal
structure of HBV Cp (PDB code: 5T2P). (B) Structures of SBA analogue (29) and HAP
analogue (30) in reported co-crystal structures. (C) Structural overlay of predicted binding mode
2
4

ACCEPTED MANUSCRIPT
of 19o and 29 within the HBV capsid. (D) Structural overlay of predicted binding mode of 19o
and 30 within the HBV capsid. Key residues are highlighted in yellow sticks. H-bond
interactions are depicted as black dotted lines. Alkyl-π interactions is represented as double
headed arrow in black.
Molecular modeling. To gain understanding on the binding mode of our new CAEs, docking
analysis was performed using Glide XP (version 6.9) [39-40]. The predicted binding mode of
compound 19o within HBV Cp suggests a key interaction between the carbonyl oxygen of
benzamide group in 19o and W102. The halogenated benzamide group of 19o extends into a
deep pocket lined by several hydrophobic residues; P25, L30, T33, W102 and I105 and makes
CH-π interaction between the phenyl of benzamide group and P25 (Figure 7A). A similar
interaction was observed and found to be critical for the reported SBA inhibitor of HBV Cp [41].
An overlay of ligand 19o and SBA analog (29) is shown in Figure 7C. The aminothiophene
carboxamide within 19o was predicted to occupy another large hydrophobic sub pocket lined by
several residues F23, F110, Y118 and F122. These residues are located at the dimer-dimer
interface and are known to be critical for icosahedral capsid assembly [41-43]. The amine of
carboxamide at 3 position of thiophene makes hydrogen bond interaction with backbone
carbonyl of P138. The methyl group at 4 position of thiophene core makes hydrophobic
interactions within the pocket with F23, Y118 and F122. The interactions associated with
thiophene group of 19o within this large hydrophobic sub pocket was observed to be similar to
that of SBA (Figure 7C). The bis(fluoranyl)pyrrolidine-2-carboxylic acid within HAP (30)
occupies this sub pocket and makes interactions identical to the thiophene group of 19o (Figure
7
D).
2
5

ACCEPTED MANUSCRIPT
Conclusion. ATDC analogue 11, previously identified as a CAE hit, was confirmed as a CAE
lead via further ADME and PK studies. Extensive SAR revealed that a para F substitution along
with a meta substitution on the phenyl ring confers the best antiviral activity, with five analogues
of subtype 19 exhibiting nanomolar antiviral activity and no cytotoxicity. Interestingly, analogue
1
9o was identified via a combination of SAR and SPR from a highly potent compound 19h. In
the end, our SAR and SPR efforts identified two analogues, 19o (EC50 = 0.11 M, CC50 >100
M, F = 25%) and 19k (EC = 0.31 M, CC >100 M, F = 46%), with potent antiviral activity,
µ
µ
µ
µ
50
50
no cytotoxicity and good oral bioavailability. Such lead profiles render 19o and 19k discernibly
superior to the reported CAEs 7-10 used in our studies and excellent candidates for preclinical
development.
Experimental
Chemistry
General Procedures. All commercial chemicals were used as supplied unless otherwise
indicated. Dry solvents were either purchased (toluene and dioxane, EtOH) or dispensed under
argon from an anhydrous solvent system with two packed columns of neutral alumina or
molecular sieves. Flash chromatography was performed on a Teledyne Combiflash RF-200 with
RediSep columns (silica) and indicated mobile phase. All moisture sensitive reactions were
1
13
performed under an inert atmosphere of ultra-pure argon with oven-dried glassware. H and C
NMR spectra were recorded on a Varian 600 MHz spectrometer or Bruker 400 MHz
spectrometer. Mass data were acquired on an Agilent TOF II TOS/MS spectrometer capable of
ESI and APCI ion sources. Melting points were determined in open glass capillaries using a
MEL-TEMP melting-point apparatus. Analysis of sample purity was performed on a Varian
2
6

ACCEPTED MANUSCRIPT
Prepstar SD-1 HPLC system with a Phenomenex Gemini, 5 micron C18 column (250mm x 4.6
mm). HPLC conditions: solvent A = H O, solvent B = MeCN; flow rate = 1.0 mL/min;
2
compounds were eluted with a gradient of 20% MeCN/H O to 100% MeCN for 30 min. Purity
2
was determined by total absorbance at 254 nm. All tested compounds have a purity ≥ 96%.
Characterization data for final compounds other than the representative analogues are shown in
Supplementary Data.
5
-Amino-3-methyl-N2-phenylthiophene-2,4-dicarboxamide (11). To a solution of 2-
cyanoacetamide (59 mg, 0.70 mmol) in anhydrous EtOH (3mL) was added acetoacetanilide (124
mg, 0.70 mmol) and sulfur (23 mg, 0.70 mmol), followed by the addition of morpholine (121
mg, 1.40 mmol) dropwise. The reaction mixture was stirred at 80°C until no starting material
left as shown by TLC. Solvent was removed and the residue was purified by flash
chromatography on silica gel using hexanes:EtOAc (1:2) to obtain compound 11 (170 mg, 89
1
%
). Mp 234−235°C H NMR (600 MHz, DMSO- d ) δ 9.57 (s, 1H), 7.67 (d, J = 8.4 Hz, 2H),
6
1
3
7
1
2
.36 (m, 4H), 7.12 (m, 1H), 7.05 (s, 2H), 2.41 (s, 3H); C NMR (100 MHz, DMSO-d ) δ 167.9,
6
+
62.0, 161.3, 140.2, 139.7, 128.9, 123.5, 120.6, 112.4, 112.0, 16.3; C H N O S [M+H]
13
14
3
2
76.0803, found 276.0801.
General procedure for the synthesis of compounds 12a-k. To a solution of a carboxylic acid
derivative (0.4 mmol) in anhydrous DMF (1mL) was added PyAOP (0.42 mmol) and DIEA (0.8
mmol). The resulting mixture was stirred at rt for 30 min before 5-amino-3-methyl-N2-
phenylthiophene-2,4-dicarboxamide (0.4 mmol) was added. This reaction mixture was stirred at
rt overnight, and quenched by adding H O (5 mL). The aqueous was extracted with EtOAc (4 ×
2
1
5 mL, and the combined organics were washed with brine (2 × 40 mL) and dried over MgSO4.
2
7

ACCEPTED MANUSCRIPT
Solvent was removed and the residue was purified by flash chromatography on silica gel using
hexanes: EtOAc (3:2) to obtain 12a-k.
3
-Methyl-5-(morpholine-4-carboxamido)-N2-phenylthiophene-2,4-dicarboxamide
(12a).
1
Yield 57 %; Mp 270−271°C; H NMR (600 MHz, DMSO- d ) δ 12.68 (s, 1H), 10.32 (s, 1H),
6
8
.28 (s, 2H), 7.78 (d, J = 7.8 Hz, 1H), 7.45 (m, 2H), 7.22 (m, 2H), 3.41 (s, 4H), 2.82 (s, 4H), 2.62
1
3
(
s, 3H); C NMR (150 MHz, DMSO-d
6
) δ 167.9, 162.1, 161.4, 140.1, 139.7, 129.0, 123.6,
+
1
3
5
20.7, 112.4, 112.2, 63.7, 43.1, 16.3; HRMS (ESI+) calcd. for for C H N O S [M+H]
18 21 4 4
89.1282, found 389.1280.
-Amino-4-carbamothioyl-N-(3,4-difluorophenyl)-3-methylthiophene-2-carboxamide (13)
[44]. To a solution of 2-cyanoacetamide (0.3 mmol) in anhydrous EtOH (1 mL) was added 2-
cyanoethanethioamide (0.3 mmol), followed by the addition of morpholine (0.6 mmol) dropwise.
The resulting mixture was stirred at 80 °C until no starting material left as shown by TLC.
Solvent was removed and the residue was purified by flash chromatography on silica gel using
1
hexanes:EtOAc (1:2 ) to give compound 13 (61%). Mp 175−176°C; H NMR (600 MHz,
CD OD) δ 7.64 (d, J = 7.8 Hz, 2H), 7.44 (t, J = 7.8 Hz, 2H), 7.23 (t, J = 7.2 Hz, 1H), 2.56 (s,
3
1
3
3
1
H); C NMR (151 MHz, DMSO-d ) δ 167.9, 162.0, 161.3, 140.1, 139.7, 128.9, 123.5, 120.6,
6
+
12.4, 112.0, 16.2; HRMS (ESI+) calcd. for C H N OS [M+H] 292.0575, found 292.0573.
1
3
14
3
2
General procedure for the synthesis of compounds 14a-d [45]. To a solution of malononitrile
0.7 mmol) in anhydrous EtOH (3mL) was added an acetoacetanilide derivative (0.7 mmol) and
(
sulfur (0.7 mmol), followed by the addition of morpholine (1.4 mmol) dropwise. The resulting
mixture was stirred at 80 °C until no starting material left as shown by TLC. Solvent was
removed and the residue was purified by flash chromatography on silica gel using
hexanes:EtOAc (1:2) to give compounds 14a-d.
2
8

ACCEPTED MANUSCRIPT
5
-Amino-4-cyano-N-(4-fluoro-3-methylphenyl)-3-methylthiophene-2-carboxamide(14c).
1
Yield 47 %; Mp 212−213°C; H NMR (600 MHz, CD OD) δ 7.67 (d, J = 6.6 Hz, 1H), 7.62 (m,
3
1
3
1
1
H), 7.27 (t, J = 9.0 Hz, 1H), 2.71 (s, 3H), 2.53 (s, 3H); C NMR (150 MHz, DMSO) δ 165.5,
60.8, 158.1, 156.6, 140.8, 135.3, 124.4, 123.4, 120.1, 115.9, 113.3, 88.4, 15.3, 14.8; HRMS
+
(
ESI+) calcd. for C H FN OS [M+H] 290.0758, found 290.0760.
1
4
13
3
General procedure for the synthesis of compounds 15a-b [46-47]. To a solution of ethyl 2-
cyanoacetate (0.5 mmol) (0.5 mmol) in anhydrous EtOH (2.0 mL) was added was added
acetoacetanilide (0.5 mmol) and sulfur (16 mg, 0.5 mmol), followed by the addition of
morpholine (1.0 mmol) dropwise. The resulting mixture was stirred at 80 °C until no starting
material left as shown by TLC. Solvent was removed and the residue was purified by flash
chromatography on silica gel using hexanes:EtOAc (1:2) to give compounds 15a-b.
Ethyl 2-amino-4-methyl-5-(phenylcarbamoyl)thiophene-3-carboxylate (15a). Yield 83 %;
1
Mp 180−181°C; H NMR (600 MHz, DMSO-d ) δ 9.66 (s, 1H), 7.72 (s, 2H), 7.59 (d, J = 7.2
6
Hz, 2H), 7.30 (t, J = 7.8 Hz, 2H), 7.04 (t, J = 7.2 Hz, 1H), 4.23 (q, J = 6.6 Hz, 2H), 2.48 (s, 2H),
1
3
1
1
3
.29 (t, J = 6.6 Hz, 3H); C NMR (100 MHz, DMSO-d
6
) δ 165.5, 165.4, 161.8, 140.7, 139.6,
-
29.0, 123.7, 120.5, 113.4, 105.9, 59.7, 17.0, 14.8; HRMS (ESI-) calcd. for C15
03.0798, found 303.0802.
15 2 3
H N O S [M-H]
General procedure for the synthesis of compounds 16a-m. To a solution of a cyanoacetamide
derivative (0.4 mmol) in anhydrous EtOH (2mL) was added an acetoacetanilide derivative (0.4
mmol) and sulfur (0.4 mmol), followed by the addition of morpholine (0.8 mmol) dropwise. The
resulting mixture was stirred at 80 °C until no starting material left as shown by TLC. Solvent
was removed and the residue was purified by flash chromatography on silica gel using
hexanes:EtOAc (1:2 ) to give compounds 16a-m. (55-83 %).
2
9