Aniline-Based Inhibitors of Influenza H1N1 Virus Acting on Hemagglutinin-Mediated Fusion

Article abstract of DOI:10.1021/acs.jmedchem.7b00908

Two series of easily accessible anilines were identified as inhibitors of influenza A virus subtype H1N1, and extensive chemical synthesis and analysis of the structure-activity relationship were performed. The compounds were shown to interfere with low pH-induced membrane fusion mediated by the H1 and H5 (group 1) hemagglutinin (HA) subtypes. A combination of virus resistance, HA interaction, and molecular dynamics simulation studies elucidated the binding site of these aniline-based influenza fusion inhibitors, which significantly overlaps with the pocket occupied by some H3 HA-specific inhibitors, indicating the high relevance of this cavity for drug design.

Full text of DOI:10.1021/acs.jmedchem.7b00908

Subscriber access provided by READING UNIV
Article
Aniline-based inhibitors of influenza H1N1
virus acting on hemagglutinin-mediated fusion
Rosana Leiva, Marta Barniol-Xicota, Sandra Codony, Tiziana Ginex, Evelien Vanderlinden,
Marta Montes, Michael Caffrey, F. Javier Luque, Lieve Naesens, and Santiago Vazquez
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.7b00908 • Publication Date (Web): 08 Dec 2017
Downloaded from http://pubs.acs.org on December 9, 2017
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 free 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 accessible to all
readers and 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.
Journal of Medicinal Chemistry 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 83
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Anilineꢀbased inhibitors of influenza H1N1
virus acting on hemagglutininꢀmediated fusion
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
∫
∫,ǁ
∫,ǁ
# ,ǁ
Rosana Leiva, Marta BarniolꢀXicota, Sandra Codony, Tiziana Ginex, Evelien
§
∫
†
#
,§
Vanderlinden, Marta Montes, Michael Caffrey, F. Javier Luque, Lieve Naesens,*
,∫
and Santiago Vázquez*
∫
Laboratori de Química Farmacèutica (Unitat Associada al CSIC), Facultat de Farmàcia
i Ciències de l’Alimentació, and Institute of Biomedicine (IBUB), Universitat de
Barcelona, Av. Joan XXIII, 27ꢀ31, Barcelona, Eꢀ08028, Spain
#
Department of Nutrition, Food Science and Gastronomy, Faculty of Pharmacy and
Food Sciences, and Institute of Biomedicine (IBUB), Universitat de Barcelona, Av. Prat
de la Riba 171, Santa Coloma de Gramanet, Eꢀ08921, Spain
§
Rega Institute for Medical Research, KU Leuven, Bꢀ3000 Leuven, Belgium
†
Department of Biochemistry & Molecular Genetics, University of Illinois at Chicago,
900 South Ashland Avenue, Chicago, Illinois 60607, United States
KEYWORDS. Influenza A virus, antiviral drug, hemagglutinin, fusion, molecular
dynamics.
1
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 83
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
ABSTRACT. Two series of easily accessible anilines were identified as inhibitors of
influenza A virus subtype H1N1, and extensive chemical synthesis and analysis of the
structureꢀactivity relationship were performed. The compounds were shown to interfere
with low pHꢀinduced membrane fusion mediated by the H1 and H5 (group 1)
hemagglutinin (HA) subtypes. A combination of virus resistance, HA interaction and
molecular dynamics simulation studies elucidated the binding site of these anilineꢀbased
influenza fusion inhibitors, which significantly overlaps with the pocket occupied by
some H3 HAꢀspecific inhibitors, indicating the high relevance of this cavity for drug
design.
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
Influenza A and B viruses are highly contagious respiratory pathogens and the cause
1
of annual epidemics with high medical and economic burden. Besides, influenza A
causes occasional pandemics due to gene reassortment between human and zoonotic
viruses. The notorious 1918 Spanish influenza infected one third of the world
2
population with >40 million deaths. The recent pandemic of 2009 was caused by a
swineꢀorigin H1N1 virus that killed ~300,000 people within only 18 months,
3
particularly in younger populations. Hence, costꢀeffective antiviral strategies must be
developed for preventing and eradicating emerging influenza pandemics.
Influenza therapy currently relies on two classes of antiviral drugs: the M2 proton
4
channel blockers (restricted to influenza A) and neuraminidase inhibitors (NAIs). The
currently approved M2 inhibitors suffer from global virus resistance, making it
5
necessary to develop new chemical scaffolds. Resistance is also a growing issue for the
6
NAIs, particularly oseltamivir. Hence, there is an urgent need for new drugs having
superior efficacy, high barrier for resistance, and eventually targeting other proteins
2
ACS Paragon Plus Environment

Page 3 of 83
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
7
ꢀ10
relevant for the viral life cycle. The most extensively studied targets are the influenza
11
12
13
hemagglutinin (HA) , RNA polymerase complex, and NS1 protein.
The homotrimeric HA, abundantly present on the viral envelope, has a crucial and
14
dual role in virus entry into the host cell. First, the HA globular head mediates virus
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
binding via its receptorꢀbinding site for sialylated cell surface glycans. The second
function lies in the HA stem domain. After endosomal uptake, this domain undergoes
low pHꢀinduced and drastic refolding, which leads to release of the hydrophobic fusion
15
peptide and fusion of the viral envelope with the endosomal membrane. Once the viral
genome segments are released into the cytoplasm and transferred into the cell nucleus,
1
2
viral RNA replication begins.
One notable difficulty in targeting HA is related to its sequence diversity. The
eighteen HA subtypes of influenza A fall into two phylogenetic groups, for instance H1
and H5 belong to group 1 whereas H3 and H7 are in group 2. In the last years, several
broadly neutralizing antibodies against HA group 1, 2, or both, have advanced towards
16
clinical trials. They target conserved head or stem epitopes which were revealed by
17
cocrystallographic analyses. In contrast, broadlyꢀacting small molecule inhibitors of
HA are challenging since the only example is arbidol (umifenovir, Chart 1), which is
18
commercially available in Russia and in clinical trials elsewhere. Besides other
antiviral mechanisms of action, arbidol was shown to prevent fusion by inhibiting the
19
conformational change of HA at low pH. Only recently, cocrystallization of arbidol
with H3 and H7 HAs (PDB codes: 5T6N and 5T6S, respectively) revealed that it binds
20
in a hydrophobic cavity within the HA trimer stem, clamping the protomers together.
21
This site partially overlaps with that of tertꢀbutylhydroquinone (TBHQ), another
influenza fusion inhibitor for which HA cocrystallization data are available (PDB codes:
2
2
3
EYK and 3EYM). The structureꢀactivity relationship (SAR) of TBHQ derivatives
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 83
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
2
3
was recently revisited. In addition, the literature contains many molecules (some of
which are shown in Chart 1) which, based on mechanistic and virus resistance data,
9
classify as HA stemꢀbinding fusion inhibitors. Although the HA subtypeꢀdependent
antiviral activity of these molecules forms an obstacle for clinical development, they are
suitable scaffolds to design inhibitors that potentially cover a broader range of HA
subtypes.
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
Chart 1. Structures of some reported inhibitors of influenza A HAꢀmediated fusion.
2
4
TBHQ and a substituted azaspiro compound, 1, which were predicted to target the
same HA stem pocket, are restricted towards H3 (a group 2) HA. The structurally
25ꢀ27
28
29ꢀ31
related compounds 2 (BMYꢀ27709),
Chart 1) proved limited to HA subtypes H1, H2 and/or H5 (group 1). The latter activity
spectrum was also reported for 5 (RO5464466) and its 2ꢀchloro derivative, 6
3 (BMSꢀ199945), and 4 (CL385319)
(
32ꢀ33
34
35ꢀ36
(
RO5487624),
and the entirely different inhibitors 7 (S20) and 8 (MBX2546).
4
ACS Paragon Plus Environment

Page 5 of 83
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
All these HA fusion inhibitors potently suppress replication of some influenza A virus
subtypes in cellꢀbased assays, and for 6 activity was confirmed in a lethal mouse model
33
with influenza H1N1.
We here describe the synthesis, antiviral activity and mechanism of action of two
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
series of novel aniline derivatives and related compounds of general structures I and II
(Scheme 1). Considering that the benzamide group of 3 could be replaced by an aniline
group as in 5 and 6, we reasoned that, starting from 4, compounds of general structure I
may also display antiꢀinfluenza activity. On the other hand, the similar physicochemical
properties between the 1,3,3ꢀtrimethylcyclohexꢀ1ꢀyl moiety present in 3 and the
adamantyl unit, suggest that they may behave as bioisosteric groups (see Table S1 in
Supporting Information). This inspired us to take advantage of the wellꢀknown antiviral
37
activity of adamantane derivatives, and the large experience of our group in the
38ꢀ43
synthesis of adamantaneꢀcontaining and adamantaneꢀrelated polycyclic compounds,
to explore compounds in which the cyclohexyl moiety of 5 is replaced by a bulkier
adamantane. Specifically, we developed a series of analogs to investigate: a) the optimal
substitution for the aromatic moiety in series I and II; b) the effects of the heteroatom in
the linker of I; and c) the size of the hydrophobic moiety. Antiꢀinfluenza A activity was
evaluated in cellꢀbased assays with H1N1 and H3N2 strains. The fusion inhibiting effect
was demonstrated in an influenza polykaryon assay, and resistant virus was selected and
characterized. Finally, NMR experiments and molecular dynamics simulations were
conducted to predict the plausible binding site of the lead compounds within the HA
protein.
5
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 83
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
Scheme 1. General structures of the newly designed HA inhibitors. Y = NH, O, S; m =
1 or 2; n = 0, 1 or 2; Z = H or CH3.
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
Results and discussion
Chemistry. The synthesis of piperidines with general structure I (n = 1) and related
ringꢀexpanded (n = 2) and ringꢀcontracted (n = 0) analogs is shown in Scheme 2. These
compounds could be accessed in just one facile synthetic step from commercially
available starting materials. Anilines, phenols or thiophenols bearing the desired
substitution were used for the aromatic moiety, as well as a range of diverse chloroalkyl
compounds. The attack of the nucleophilic group in the aromatic moiety to the
chloroalkyl derivative upon heating furnished the desired products in low (for the less
reactive anilines) to very good (for the phenols and thiophenols) yields. While a few of
the alkylated anilines were obtained in high yields (ca 80%) by heating a neat mixture
44
of the aniline and Nꢀ(2ꢀchloroethyl)piperidine hydrochloride at 120 ºC overnight, the
less reactive anilines (e.g. bearing electronꢀwithdrawing groups) needed the addition of
2 3
base, either K CO or NaH, KI catalyst and extended (30 hours) heating, furnishing the
products in only low to moderate yields. The corresponding alkylation of phenols and
thiophenols proceeded in medium to very high yields by using NaOH in refluxing
ethanol.
6
ACS Paragon Plus Environment

Page 7 of 83
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
A different synthetic strategy was undertaken for the preparation of the
(adamantylmethyl)anilines of general structure II (Scheme 3). Compounds of structure
1
9 were prepared via a threeꢀstep route, starting from the commercially available 1ꢀ
adamantanecarboxylic acid, 16a. Using thionyl chloride, the corresponding acyl
chloride was readily obtained and added to a solution of the corresponding aniline in
dry acetone. Upon continuous heating, the desired amides 18a-w did precipitate in good
yields. Without further purification, this precipitate was treated with sodium bisꢀ(2ꢀ
methoxyethoxy)aluminum hydride to furnish the corresponding anilines in moderate to
good yields. Aniline 19x was synthesized in a similar way, starting from commercially
available 3,5ꢀdimethylꢀ1ꢀadamantanecarboxylic acid, 16b.
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
The preparation of the 3ꢀhydroxyadamantane derivatives 22a-b was first envisaged
as a direct step, through the direct lowꢀtemperature hydroxylation with HNO /H SO of
3
2
4
anilines 19. Unfortunately, this procedure led to complex mixtures of products,
probably arising from competitive oxidation of the aniline. Alternatively, derivatives
45
22a-b were accessed from 3ꢀhydroxyadamantanecarboxylic acid, 20, which reacted
with the required aniline using EDC as a coupling agent to furnish amides 21a-b, and
finally to products 22a-b after reduction with sodium bisꢀ(2ꢀmethoxyethoxy)aluminum
hydride. Finally, aniline 25, featuring an additional methylene unit in the linker, was
prepared in an analogous manner to the 19 series (Scheme 3).
7
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 83
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
Scheme 2. Synthesis of the general scaffold I derivatives 9ꢀ15.
—HCl
H
—
HCl
1
) a or b or c
N
+
N
m
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
R
N
n = 0, 1, 2
m = 1, 2
R
Cl
m
n
2) HCl / Et O
NH2
2
n
9
ꢀ12
H
N
—
HCl
—HCl
N
R
R
N
N
N
N
—
HCl
H
H
9
10
11
H
9a: H
9b: 2ꢀCH3
9g: 4ꢀCl
9h: 2ꢀBr
9m: 3,5ꢀdiF
9n: 3,5ꢀdiCF3
10a: 2ꢀCF3
10b: 2ꢀCH CH
N
N
2
3
9
9
9
9
c: 2ꢀCH CH3
9i: 3ꢀCF3
9j: 4ꢀCF3
9o: 3ꢀCl, 5ꢀCF3
9p: 3ꢀCl, 5ꢀF
9k: 2ꢀCF3 9q: 3ꢀF, 5ꢀCF3
9l: 3,5ꢀdiCl
10c: 2ꢀCH(CH3)2
2
d: 2ꢀCH(CH3)
e: 2ꢀC(CH3)3
f: 2ꢀOCH3
2
—HCl
12
—
HCl
—HCl
1
) NaOH, EtOH,
Y
+
N
n = 1, 2
Y = O, S
R
N
R
Cl
O
n
YH
2) HCl/dioxane
n
13ꢀ15
O
S
HCl
15
15c
N
R
N
R
N
R
—
HCl
14
14a: 2ꢀCH CH
—
HCl
—
13
1
1
3a: H
2
13c: 2ꢀCH CH3
2
3
15a
15b
: H
: 4ꢀCl
: 2ꢀCH2CH3
3b: 4ꢀCl 13d: 2ꢀCH(CH3)
2
14b: 2ꢀCH(CH3)2
a
Reagents and conditions: (a) K
2 3
CO , KI, DMF, heat; (b) heat; (c) NaH, KI, dry DMF,
heat. See experimental for further details.
8
ACS Paragon Plus Environment

Page 9 of 83
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
a
Scheme 3. Synthesis of anilines 19a-x, 22a-b and 25.
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
a
2 3
Reagents and conditions: (a) SOCl , 80 ºC, 2 h; (b) Et N, acetone, reflux, 3 h; (c)
EDC·HCl, DMAP, dichloromethane, rt, overnight; (d) NaAlH (OCH CH OCH ) ,
2
2
2
3 2
toluene, 95 ºC, 24 h; (e) BH
3
·THF, 0ºC to rt, 24 h; (f) HCl/Et
2
O; (g) HNO
3
, H
2
SO
4
, rt,
overnight, 95%. See experimental and Supporting Information for further details.
Of note, while the reduction of 18j, featuring a trifluoromethyl group in the meta
position, proceeded as expected to furnish aniline 19j, the reduction of its ortho and
para isomers, 18i and 18k, did not furnish the expected anilines, but 19c and 19b,
respectively (Scheme 4). The formation of 19b and 19c can be easily explained taking
into account the strong basicity of the hydride reagent, which triggers a dehalogenation
reaction that competes with the reduction process. In order to avoid this problem, we
switched from the hydride to borane, an electrophilic reductor agent that smoothly
furnished anilines 19i and 19k.
9
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 83
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
Scheme 4. Reduction of amides 18iꢀk with sodium bisꢀ(2ꢀmethoxyethoxy)aluminum
a
hydride.
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
a
Reagents and conditions: (a) NaAlH
2 2 2 3 2
(OCH CH OCH ) , toluene, 95 ºC, 24 h. See
experimental for further details.
Overall, from the aforementioned synthetic work, fifty eight new amines were
synthesized, fully characterized as the corresponding hydrochloride derivatives and
subsequently tested for antiviral activity.
Anti-influenza virus activity and cytotoxicity in cell culture. The novel compounds
were tested in MDCK cells using three influenza A strains [A/Hong Kong/7/87 (H3N2
subtype); A/PR/8/34 (H1N1 subtype); and A/Virginia/ATCC3/2009 (H1N1 subtype)] as
well as an influenza B strain (B/Hong Kong/5/1972). The A/Hong Kong/7/87 strain
carries an amantadineꢀsensitive wildꢀtype M2 channel, while the other influenza A
strains carry mutations in M2 that confer resistance to amantadine, i.e. V27T plus S31N
in A/PR/8/34 and S31N in A/Virginia/ATCC3/2009. The inhibitory effect of the
compounds on virus replication as well as their cytotoxicity, were monitored by
microscopic examination of the viral cytopathic effect (CPE) at three days post
24
infection, and confirmed by the colorimetric MTS cell viability assay (Table 1).
All compounds were found to be inactive against influenza A/H3N2 and influenza B
results not shown). On the other hand, several ones displayed favorable activity against
(
1
0
ACS Paragon Plus Environment

Page 11 of 83
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
the influenza A/H1N1 strains revealing some interesting SAR trends. Regarding the
aromatic moiety, ringꢀsubstitution is required for activity, as the unsubstituted 9a, 13a,
15a and 19a were inactive (data not shown). Moreover, the ortho position is the most
appropriate substitution site for antiꢀinfluenza activity, whereas paraꢀ, metaꢀ or full ring
substitution patterns, e.g. 9g, 9i, 19w respectively, were deleterious. More specifically,
the isopropyl group in ortho position yielded very potent compounds (9d, 13d and 19e).
Of note, compounds 9lꢀq, featuring two electron withdrawing groups in the meta
positions, were inactive, in sharp contrast with 4 and several analogs of 6 that feature
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
29,32
one or two electron withdrawing groups in the meta positions (Chart 1).
Within the
series of 2ꢀalkylꢀNꢀ[2ꢀ(piperidinꢀ1ꢀyl)ethyl]anilines 9a-d, progressive chain increase
from hydrogen (9a), methyl (9b), ethyl (9c), to isopropyl (9d), increased the antiviral
activity against both A/H1N1 strains. In addition, the drop in activity for the tertꢀbutylꢀ
substituted compound 9e, sets the length limit for this structural part.
Regarding the linker between the aromatic ring and the piperidine, a length of three
atoms was found to be optimal for the inhibitory activity (e.g. 9d vs 11). Concerning the
nature of the linker’s heteroatom, a nitrogen was similar or slightly better than oxygen
(compare 9c vs 13c, 9d vs 13d, 10b vs 14a and 10c vs 14b), while the sulfur analogs
were always inactive (15a-c).
Focusing our attention on the heterocyclic ring, piperidine seemed to give the best
activity. Ringꢀcontraction from piperidine to pyrrolidine reduced the antiviral activity
for both A/H1N1 strains (e.g. 9d vs 12). Ringꢀexpansion from piperidine to the azepane
ring was deleterious for the activity against the A/PR/8/34 strain and also increased the
cytotoxicity (e.g. 9c vs 10b).
1
1
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 83
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
Table 1. Antiviral activity in influenza virus H1N1ꢀinfected MDCK cells. For the sake
of clarity, only active compounds are shown.
b,c
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
Compound
Antiviral EC (ꢀM)
Cytotoxicity (ꢀM)
5
0
Influenza A/H1N1
A/PR/8/34
A/Virginia/ATCC3/2009
d
e
CC50
MCC
CPE
21
MTS
21
CPE
13
MTS
9.6
9
9
9
9
9
9
1
10b
10c
11
c
d
e
>100
>100
>100
>100
>100
61
11
14
52
74
100
100
>100
>100
>100
100
20
20
100
60
>100
100
≥100
20
20
>100
2.0
4.6
5.5
1.7
36
1.5
26
29
32
f
>100
45
>100
48
32
8.3
h
k
0a
45
47
15
10
5.4
≤0.8
1.7
1.6
18
2.6
<0.8
1.5
1.3
7.9
>100
>100
8.4
>100
>100
8.9
>100
23
>100
39
1
1
1
1
1
1
1
1
1
2
17
8.0
>100
>100
>100
11
3c
3d
4a
4b
8f
9d
9e
9q
8.9
11
8.0
0.89
4.0
2.6
52
5.7
1.0
<0.8
2.0
39
30
15
>100
>100
>100
>100
>100
>100
5.6
>100
>100
>100
>100
>100
>100
5.2
26
>100
0.87
3.4
1.9
23
1.6
>500
>100
0.19
0.8
0.21
>100
<0.8
>500
12
0.15
<0.8
0.12
>100
<0.8
>500
9.8
4.0
2.0
20
4
>500
≥100
22a
25
>100
Amantadine 85
Ribavirin
16
>100
178
45
a
b
MDCK: MadinꢀDarby canine kidney cells. 50% Effective concentration, or
concentration producing 50% inhibition of viral cytopathicity, as determined by
microscopic scoring of the cytopathic effect (CPE) or by measuring the cell viability
with the colorimetric formazanꢀbased MTS assay. All compounds were inactive
highest tested concentration: 100 µM) against influenza A/H3N2 (A/HK/7/87) and
influenza B (B/HongKong/5/1972). 50% Cytotoxic concentration, as determined by
measuring the cell viability with the colorimetric formazanꢀbased MTS assay.
Minimum compound concentration that causes a microscopically detectable alteration
c
(
d
e
of normal cell morphology. Values shown are the mean of 2 or 3 determinations.
1
2
ACS Paragon Plus Environment

Page 13 of 83
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Remarkably, 18f and 22a seemed not to follow the previous activity trends. The amide
18f displayed antiviral activity against A/Virginia/ATCC3/2009 and stood out for two
reasons: it is the only amide from series 18a-x endowed with antiꢀinfluenza activity, and
it bears a paraꢀbenzyloxy group in the aromatic moiety, being the only active
compound with an electron donating group. Interestingly, 22a, which bears a hydroxyl
group in the adamantane moiety, is not only the sole active analog with an unsubstituted
aromatic moiety, but also the only one to inhibit A/PR/8/34 but not
A/Virginia/ATCC3/2009. Taking into account the unique antiviral profile of
adamantanolꢀcontaining compound, 22a, and that 5 and 6 also feature a hydroxyl group,
further derivatives of 22a should be explored.
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
Activity in the influenza polykaryon assay. While this work was in progress, Basu et
al. reported that compound 14a (MBXꢀ2329 in their paper) inhibits influenza HAꢀ
mediated membrane fusion, as demonstrated in a hemolysis assay with influenza
3
5
A/PR/8/34. Hence, we determined the inhibitory activity of 9d and 14a in the
polykaryon assay which is based on cellꢀcell fusion when HAꢀtransfected cells are
24
exposed to low pH. We included not only H1 HA (from A/PR/8/34), but also H5 HA
derived from the avian H5N1 virus A/duck/Hunan/795/2002) and H3 HA (from the
(
A/X31 virus). As shown in Figure 1, compounds 9d and 14a, at a concentration of 100
µM, provided 100% inhibition of H1 and H5 HAꢀmediated polykaryon formation. The
two compounds were equipotent for H1 HA (EC50: 10.1 and 9.6 µM for 9d and 14a,
respectively), whereas for H5 HA, 14a was twoꢀfold more active than 9d (EC : 16.7
5
0
and 39.6 µM, respectively). This means that 14a produces a similar effect on H1 HA as
d. The lack of activity of 14a in the CPE reduction assay with A/PR/8/34 thus appears
9
related to its relatively high cytotoxicity, which most likely obscures an antiviral effect.
No inhibition of H3 HAꢀinduced polykaryon formation was observed (data not shown),
1
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 83
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
consistent with the lack of antiviral activity of 9d and 14a against H3N2 virus. Finally,
the possibility was excluded that the different activity against the H1N1 and H3N2
viruses could be related to an action of these molecules on the neuraminidase
component, since 9d and 14a produced no inhibition in an enzymatic neuraminidase
assay with A/PR/8/34 virus (IC50 values of 9d and 14a and zanamivir: >100 µM; 100
µM and 1.4 nM, respectively; method adapted from reference 46).
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 1. Inhibition of HAꢀinduced polykaryon formation at low pH. HeLa cells
expressing HA subtype H1 (top row) or H5 (middle row) were treated with trypsin to
render the HA fusionꢀcompetent. After preꢀincubation with compound 9d or 14a at 100
µM, the cells were briefly exposed to pH 5.2 to induce membrane fusion, then incubated
with medium during 3 h to allow polykaryon formation. The panels on the right show
HAꢀtransfected controls exposed to pH 7, and the bottom row are untransfected control
cells. Representative fields are shown (original magnification x200). The table
summarizes the EC50 values (mean ± SEM; n = 3), defined as the compound
concentration at which the number of polykaryons was 50% compared to the no
compound control.
1
4
ACS Paragon Plus Environment

Page 15 of 83
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Selection and characterization of 9d-resistant virus mutants. Having confirmed that
9d represents a class of influenza fusion inhibitors with group 1 specificity, we next
characterized its binding interaction with the HA protein. First, we selected resistant
mutants by serial virus passaging under compound 9d. As shown in Table 2, virus
clones obtained after three passages under up to 50 µM of compound 9d were resistant
to 9d and its ringꢀexpanded analogue 10c, whereas the control virus passaged in the
absence of compound was equally sensitive as the parent A/PR/8/34 virus used at the
start. Sequencing revealed that both 9d-resistant virus clones contained four
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 2
substitutions in HA (Table 2). The changes S186 P and I10 V were considered less
relevant since they were linked to cell culture adaptation in a previous A/PR/8/34
39
passage study. The two remaining HA changes are both located in the HA stem, with
2 2
T107 I lying at a more membraneꢀdistal site than R153 I. The resistance to 9d was not
related to an increase in the viral fusion pH (which indicates that the HA has become
less stable allowing its refolding at less acidic pH), since the pH to achieve 50%
hemolysis was 5.4 for the virus passaged under 9d compared to 5.3 for the virus
passaged in the absence of compound (data not shown).
We thus concluded that either Thr107
in inhibitor binding. Support for the relevance of Thr107
an azide analogue of 3 localized to this region in HA
2
or Arg153
2
lies at or close to the region involved
comes from the finding that
in photoaffinity labelling
experiments. Also, substitution of Phe110 was observed in H1N1 influenza viruses
2
2
28
2
2
6
with resistance to 2.
1
5
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 83
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
Table 2. Phenotypic and genotypic characterization of H1N1 virus that was passaged
a
under 9d.
res
res
No cpd- Parent
Compound 9d -cl1 9d -cl2
Cytotoxicity (µM)
cl1
virus
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
b
Substitutions in HA
T107
R1532I
2
I
T107
R1532I
2
I
T54
2
A
ꢀ
c
d
e
Antiviral EC50
(ꢀM)
CC50
MCC
9
1
d
0c
>100
>100
11
>100
>100
11
2.3
1.5
15
4.6
8.4
16
>100
52
>100
100
100
100
Ribavirin
a
A/PR/8/34 virus was passaged three times under compound 9d (at 12.5, 25 and 50
µM, successively); a control was included that was passaged in the absence of
compound. The harvested viruses were plaqueꢀpurified (cl1 and cl2: clone 1 and 2,
respectively) prior to HA sequencing. The parent virus means the virus stock that was
b
used to start passage #1. The residues are numbered per HA polypeptide part, i.e. HA
1
res
res
and HA
2
. Clones 1d ꢀcl1 and 1d ꢀcl2 contained the indicated substitutions besides two
P and I10 V) previously associated to cellꢀculture
additional changes (S186
1
2
39 c
adaptation. 50% Effective concentration determined by cytopathic effect (CPE) assay;
similar EC50 values were obtained by MTS assay (data not shown). 50% Cytotoxic
d
concentration, as determined by measuring the cell viability with the colorimetric
e
formazanꢀbased MTS assay. Minimum compound concentration that causes a
microscopically detectable alteration of normal cell morphology. Values shown are the
mean of 2 or 3 determinations.
NMR experiments. The saturation transfer difference (STD) NMR experiment has
been shown to be useful to characterize the interaction between small molecules and
large protein targets, including interactions between an inhibitor and influenza
23,47
1
hemagglutinin.
In this experiment, the H atoms of the protein are selectively
1
saturated and the observation of magnetization transfer from the protein H to the ligand
1
1
H is observed, which enables identification of ligand H in closest proximity to the
48
protein surface. Taking into account that the sequence identity between H1 and H5
HA in the HA2 region is very high (see Table S2 and Figure S1 in the Supporting
Information), we assayed the binding of 9d and 10c to recombinant H5 HA, a group 1
HA, by STD NMR. As shown in Figure 2A, the signals of 9d and 10c clearly suggest
1
6
ACS Paragon Plus Environment

Page 17 of 83
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
binding to H5 HA. Interestingly, in matched experiments, 10c exhibited significantly
1
stronger STD signals than 9d, which suggests a better interaction with H in closer
proximity to the H5 HA surface. As shown in Figure 2B, quantitation of the individual
STD signals suggests that the aromatic moiety of both molecules is in closest relative
contact to the H5 HA surface and that the piperidine and azepane moieties of 9d and
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
10c, respectively, are more distant. In addition, the isopropyl group of 9d appears to
make better relative contact with the H5 HA surface.
Figure 2. STD NMR experiments (A) and relative STD signals (B) of compounds 9d
and 10c in the presence of recombinant H5 HA. The experimental conditions were 100
4
µM compound and 1 µM H5 HA in 20 mM PO /pH 7.5, 150 mM NaCl, and 100%
2
2
H O performed on a Bruker 900 MHz Avance spectrometer equipped with a cryoprobe
at 25˚C.
Molecular modeling studies.
Homology modeling. A 3D model of H1 HA in the open form was built up from the Xꢀ
ray crystal structure of TBHQ bound to X31 HA (see Experimental Section and Figure
S2 in the Supporting Information). Although the residues that shape the TBHQ pocket
1
7
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 83
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
in H3 HA appeared to be generally conserved in the H1 protein, two important
2 2
differences were noted. First, Arg54 in H3 HA, which forms a salt bridge with Glu97
in the Xꢀray structure, is replaced by Ser in A/PR/8/34, and by Thr in
A/Virginia/ATCC3/2009, respectively. Second, the electrostatic potential in the binding
pocket shows a more pronounced negative potential in H1 HA compared to the H3
protein (Figure 3), which can stabilize positively charged ligands such as 9d, whereas in
H3 HA the site is less polar, and hence better suited to accommodate hydrophobic
ligands.
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
8
ACS Paragon Plus Environment

Page 19 of 83
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Figure 3. Comparison of (left) the binding sites for the H1 HA subtypes (A) A/PR/8/34
and (B) A/Virginia/ATCC3/2009, and (C) the H3 HA subtype A/HK/7/87 obtained by
homology modelling. HA2 residues that differ among the three strains are shown in red
font. (Right) Representation of the solventꢀscreened electrostatic potential for (A)
A/PR/8/34, (B) A/Virginia/ATCC3/2009, and (C) A/HK/7/87. The isosurface
corresponds to electrostatic potential values ranging from ꢀ9 (red) to +3 (blue) kcal/mol.
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
Docking and Molecular Dynamics calculations. Choice of the starting structures of
the ligandꢀprotein complexes used in MD simulations relied on the analysis of both
docking score and population of poses, as well as on the comparison with the available
21
Xꢀray structural information, specifically regarding the binding mode of THBQ.
Clusterization was performed to identify poses with similar arrangement in the binding
pocket (poses were grouped into a single cluster if the RMSD was less than 2 Å). This
process led to two and four major binding poses in the binding cavity of A/PR/8/34 and
A/Virginia/ATCC3/2009, respectively (see Figure S3 in the Supporting Information).
With regard to A/PR/8/34, the most populated pose of compound 9d revealed that
the backbone oxygens of Lys58 and Thr54 act as hydrogenꢀbond acceptors of the
2
2
protonated piperidine and aniline nitrogen atoms (see Figure S3 in the Supporting
Information for the docking scores). In the second cluster, the protonated piperidine
2
interacts with the backbone of Val55 . A solution pertaining to the first, bestꢀscored
cluster was selected for MD simulations (see below). With regard to
A/Virginia/ATCC3/2009, clusters 1 and 4 show a similar orientation of compound 9d in
the binding pocket, which is reversed in clusters 2 and 3. Keeping in mind the docking
scores reported in Figure S3, and the resemblance with the position occupied by TBHQ
in the Xꢀray structure, clusters 1 and 4 were considered for subsequent MD simulations.
However, only the simulation starting from cluster 4 led to a stable complex along the
1
9
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 83
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
trajectory (see below). We note that the representative poses of cluster 1 in A/PR/8/34
and cluster 4 in A/Virginia/ATCC3/2009 adopt a similar arrangement in the binding
pocket (see Figure S3 in the Supporting Information). Finally, attempts to dock
compound 9d within the TBHQ binding site of H3 HA (A/HK/7/87) models were
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
unsuccessful, likely due to the reduced accessibility caused by the Arg54 side chain
and the electrostatic repulsion with the positive charge of 9d, in agreement with its H1
selectivity in biological experiments.
Next, the structural integrity of the selected poses of compound 9d bound to
A/PR/8/34 and A/Virginia/ATCC3/2009 was examined by MD simulations. In both
cases a stable accommodation of the ligand in the binding pocket was found after the
first 25 ns of simulation (see Figure S4 in the Supporting Information). The 2ꢀ
isopropylaniline moiety is placed in a subpocket delimited by Leu98 , Leu99 , and
2
2
Leu101
piperidine ring would occupy the inner part of the cavity, facing Tyr308
and Phe309 in A/Virginia/ATCC3/2009 (Figure 4). According to this arrangement, it
2
, matching most of the region filled by TBHQ in H3 HA, whereas the
1
in A/PR/8/34
1
can be expected that relatively small hydrophobic substituents bound to the aniline
moiety, such as ethyl (9c), tꢀbutyl (9e) and bromine (9h), would be tolerated, while
substitutions in meta (9i and 9l-q) or para (9g, 9j, 13b or 15b) may be expected to
cause steric clashes with residues in the binding pocket, thus explaining the reduced
antiviral activity. On the other hand, binding of 9d would be assisted by stable
hydrogenꢀbonding interactions between the piperidine and aniline nitrogen atoms with
the backbone carbonyl groups of Glu57 and Thr54 in A/PR/8/34, and of Val55 and
2
2
2
Ser54
2
in A/Virginia/ATCC3/2009 (Figure 4). Among the two hydrogenꢀbond
interactions, the former can be expected to be more stabilizing, since it involves a
positively charged hydrogenꢀbond donor (i.e., the protonated piperidine). This is
2
0
ACS Paragon Plus Environment

Page 21 of 83
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
reflected in average distances close to 2.8 Å (piperidine N) and 3.5 Å (aniline N), and
…
O HꢀN angles in the range of 130°ꢀ160°. It was further quantified by means of QM
calculations performed at the M062X/6ꢀ311++G(d,p) level for model systems
composed of formaldehyde and either protonated trimethylamine or methylaniline (see
Table S3 in the Supporting Information). These calculations were performed for two
distinct orientations, which correspond to a linear hydrogenꢀbond arrangement and a
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
…
geometry shifted 45º from linearity (chosen according to the range of O HꢀN angles
sampled in simulations; see above). The results indicate that the interaction of the
protonated piperidine with the carbonyl group of Glu57 (in A/PR/8/34) / Val55 (in
2
2
A/Virginia/ATCC3/2009) is intrinsically around 13 kcal/mol more stabilizing than the
interaction formed by the aniline nitrogen atom. Hence, it can be concluded that the
hydrogen bond formed by the aniline moiety is less stabilizing for the binding of
compound 9d.
Figure 4. Proteinꢀinhibitor complexes for compound 9d within the HA2 polypeptide of
(A) A/PR/8/34 and (B) A/Virginia/ATCC3/2009, obtained after 50 ns of MD
simulations.
2
1
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 83
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
Hence, along the MD simulations, we observed a stable interaction pattern that
resulted from ligand anchoring due to hydrogen bonding, especially involving the
piperidine moiety (see Figure S4 in the Supporting Information). Nevertheless, from a
wider perspective, MM/PBSA calculations performed for the set of snapshots collected
along the last 15 ns of MD trajectories, showed that the calculated HA binding affinity
of compound 9d for A/Virginia/ATCC3/2009 exceeds that of A/PR/8/34 by around 3.5
kcal/mol (see Table S4 in the Supporting Information). This preference can be mainly
ascribed to the nonꢀpolar contribution (ꢀ16.0 and ꢀ19.5 kcal/mol, respectively for
A/PR/8/34 and A/Virginia/ATCC3/2009). In this context, we noted that nonꢀ
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
electrostatic interactions were identified to have a major contribution in the binding of
49
1
4a to HA. Taking together the lower impact of the hydrogen bond formed by the
aniline moiety on the interaction energy, and the significant contribution of the nonꢀ
electrostatic terms, it can be expected that the replacement of the aniline moiety (in 9d)
by a phenoxyꢀ group (in 13d) should not be a critical modification in terms of binding
affinity, since loss of the weaker hydrogenꢀbond may be counterbalanced by nonꢀ
electrostatic interactions in the binding pocket.
The arrangement of compound 9d in the binding pocket shows a significant overlap
with both TBHQ and arbidol (see Figure S5 in the Supporting Information). Moreover,
this pocket is also close to the residues (Glu103
2 2 2
, Asn104 , and Glu105 ) in HA2 that
were primarily implicated in an A/PR/8/34 HA photoaffinity labelling study with an
28
azide analog of fusion inhibitor 3 (see Figure S5 in the Supporting Information). Very
recently, this HA binding pocket of A/PR/8/34 was proposed to accommodate the
36
fusion inhibitor 8.
2
2
ACS Paragon Plus Environment

Page 23 of 83
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Impact of 9dꢀresistance mutation T107 I. As explained above, serial passaging of
2
A/PR/8/34 in the presence of compound 9d resulted in a resistant virus carrying a
Thr107
peptide portion and helps to stabilize the orientation of the three
TBHQ pocket” (Figure 5) through a waterꢀassisted network of polar interactions
involving Thr107 , Arg106 , Glu103 , Lys51 and Thr54 . The Thr107 Ile change
could destabilize these interactions, particularly the electrostatic interaction between
Lys51 and Glu103 , which in turn interact with Arg106 and water molecules below
→
Ile change in HA2 (Table 2). Residue Thr107
2
is located near the fusion
αꢀhelices that form the
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
2
2
2
2
→
2
2
2
the ligand binding site. Such a destabilization could weaken the binding affinity for 9d,
thus explaining why the mutant virus is insensitive to 9d.
Finally, we noted that most of the resistance mutations obtained under the fusion
3
4
36
2
inhibitors 7 and 8 coꢀlocalize with the position of Thr107 , which is placed close to
the proposed binding pocket for compound 9d (see Figure S6 in the Supporting
Information).
2
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 83
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. (A) Putative effects of the 9dꢀassociated Thr107→Ile mutation on waterꢀ
assisted polar network among residues pertaining to HA2
1 2
and HA2 helices in
A/PR/8/34 HA. Detailed view of specific interactions in (B) HA2 , and (C) HA2 and
1
1
HA2
the hydroxyl group of Thr107
interaction between Lys51 and Glu103
water molecules below the ligand binding site.
2
helices. The Thr107→Ile change not only would break the interaction between
and Lys51 , but would also destabilize the electrostatic
, which in turn interacts with Arg106 and
2
2
2
2
2
Binding mode for the adamantane analogues. To evaluate the effect of substitution of
the piperidine by the adamantane moiety, the docking and MD simulations protocol
described for compound 9d was also performed for compound 25, which contains
aniline and adamantane moieties linked by two methylene units. This compound was
found to be potent against the A/Virginia/ATCC3/2009 strain, yet poorly active against
A/PR/8/34.
2
4
ACS Paragon Plus Environment

Page 25 of 83
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
The different activity profile of compound 25 can be explained by its distinct
accommodation in the binding site (see Figure 6). Compound 25 nicely fills the binding
site of A/Virginia/ATCC3/2009, where the adamantane moiety occupies the
2 2 1 1
hydrophobic cavity formed by Trp92 , Tyr94 , Phe309 and Pro308 , forming van der
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
Waals interactions with these residues (see Figure 6C). On the opposite site of the
binding pocket, the aniline nitrogen atom is engaged in a hydrogen bond with the
backbone oxygen of Ser54 , while the aromatic ring fills the hydrophobic region
2
occupied by TBHQ in the reference structure 3EYM.
The worse binding of compound 25 to A/PR/8/34 is reflected in the shift of the
adamantane unit toward the entrance of the cavity, where it is surrounded by polar
residues (Asp90 , Lys321 , Asn60 , and Arg324 ) and becoming more exposed to the
2
1
2
1
aqueous solvent (see Figure 6A). This agrees with the worse fit of the adamantane unit
in the interior of the cavity, as the presence of Tyr308 in A/PR/8/34 (replaced by
Phe309 in A/Virginia/ATCC3/2009) tends to reduce the hydrophobicity of the pocket,
1
1
but possibly more importantly changes the shape of the inner pocket in
A/Virginia/ATCC3/2009 due to the hydrogen bond formed by Tyr308 with Tyr94 in
1
2
A/PR/8/34 (see Figure 6B, D).
2
5
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 83
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 6. Front (A, C) and back (B, D) view of the MDꢀderived proteinꢀligand complex for
compound 25 in the HAs from A/PR/8/34 (A, B) and A/Virginia/ATCC3/2009 (C, D). Key
residues in the binding site are shown in red. Solid contours reflect the shape of the binding site.
Note the distinct shape of the pockets due to the replacement of Tyr308 in A/PR/8/34 by
Phe309
1
in A/Virginia/ATCC3/2009, which breaks the hydrogen bond formed between Tyr94
in the former.
2
and Tyr308
1
Evaluation against other viruses. In addition to influenza virus testing, all compounds
underwent broad evaluation in CPE reduction assays with a panel of DNA and RNA
viruses, i.e. herpes simplex virus type 1 and type 2; vaccinia virus; feline coronavirus;
2
6
ACS Paragon Plus Environment

Page 27 of 83
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
feline herpesvirus; vesicular stomatitis virus (VSV); Coxsackie B4 virus; respiratory
syncytium virus; paraꢀinfluenzaꢀ3 virus; reovirusꢀ1; Sindbis virus; and Punta Toro
5
0
virus. Three derivatives carrying the adamantyl scaffold proved active against the
veterinary virus VSV, with antiviral EC50 values of 45 µM (14b); 12 µM (19t) and 12
µM (19u) (data not shown). Compound 19i was a potent inhibitor of parainfluenzaꢀ3
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
virus (EC50: 9 ꢀM) whereas 19f displayed activity against Coxsackie B4 virus (EC50
:
8.9 ꢀM). Based on these results, the general scaffolds I and II appear relevant for the
design of novel antiviral compounds against diverse viruses with medical importance.
Conclusions
Two series of novel anilines have been synthesized, characterized and evaluated as
antiviral compounds. Several ones displayed activity against A/H1N1 influenza with
EC50 values in the low micromolar range. Influenza fusion, virus resistance as well as
NMR experiments with the lead molecule 9d demonstrated that it interferes with HAꢀ
mediated fusion by binding to the HA stem and preventing its refolding at low pH. MD
simulations suggest that ligand 9d is able to fill the “TBHQ pocket” in the HAs of
A/PR/8/34 and A/Virginia/ATCC3/2009. This implies that the “TBHQ pocket”
represents a common and particularly relevant site for smallꢀmolecule HA fusion
inhibitors, although distinct chemotypes are required to address the different polarity of
this cavity in groupꢀ1 versus groupꢀ2 HA subtypes.
Experimental Section
Chemical Synthesis. General Methods. Melting points were determined in open
1
13
capillary tubes with a MFB 595010M Gallenkamp. 400 MHz H/100.6 MHz C NMR
1
spectra, and 500 MHz H NMR spectra were recorded on Varian Mercury 400, and
2
7
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 83
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
Varian Inova 500 spectrometers, respectively. The chemical shifts are reported in ppm
(
δ
scale) relative to internal tetramethylsilane, and coupling constants are reported in
Hertz (Hz). Assignments given for the NMR spectra of the new compounds have been
1
1
carried out on the basis of DEPT, COSY H/ H (standard procedures), and COSY
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
H/ C (gHSQC and gHMBC sequences) experiments. IR spectra were run on Perkinꢀ
Elmer Spectrum RX I spectrophotometer. Absorption values are expressed as waveꢀ
−1
numbers (cm ); only significant absorption bands are given. Highꢀresolution mass
spectrometry (HRMS) analyses were performed with an LC/MSD TOF Agilent
Technologies spectrometer. The elemental analyses were carried out in a Flash 1112
series Thermofinnigan elemental microanalyzator (A5) to determine C, H and N.
Column chromatography was performed on silica gel 60 AC.C (35−70 mesh, SDS, ref
2000027). Thinꢀlayer chromatography was performed with aluminumꢀbacked sheets
with silica gel 60 F254 (Merck, ref 1.05554), and spots were visualized with UV light
4
and 1% aqueous solution of KMnO . The analytical samples of all of the new
compounds which were subjected to pharmacological evaluation possessed purity ≥95%
as evidenced by their elemental analyses.
General Procedure A. To a solution of the required aniline (2 eq) in dry DMF at
2 3
room temperature, the chloroalkyl derivative (1 eq), K CO (2 eq) and KI (0.1 eq) were
sequentially added. The reaction mixture was stirred and heated at 90 ºC during 30 h.
After cooling down to room temperature, the resulting crude material was dissolved in
2 4
DCM. The solution was washed with water, dried over anh. Na SO , filtered and
concentrated in vacuo to obtain the desired product and some recovered starting
materials. Column chromatography (Hexane/Ethyl acetate mixtures) gave the expected
product. An analytical sample of its hydrochloride salt was prepared by adding an
2
8
ACS Paragon Plus Environment

Page 29 of 83
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
excess of HCl either in diethyl ether or 1,4ꢀdioxane to a solution of the base in ethyl
acetate.
General Procedure B. A solution of the required aniline (2 eq) and 1ꢀ(2ꢀ
chloroethyl)piperidine hydrochloride (1 eq) was stirred and heated at 120 ºC overnight.
After cooling down to room temperature, the resulting mixture was dissolved in water
and solid sodium acetate was added to obtain pH 5.5. The excess of aniline was
extracted with diethyl ether. The aqueous layer was alkalized with 2 N NaOH aqueous
solution and extracted with diethyl ether. The combined organic layers were dried over
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 4
anh. Na SO , filtered and concentrated in vacuo to obtain the desired product. An
analytical sample of its hydrochloride salt was prepared by adding an excess of HCl
either in diethyl ether or 1,4ꢀdioxane to a solution of the base in ethyl acetate.
General Procedure C. Sodium hydroxide (0.800 g, 20 mmol, 2 eq) was refluxed in
ethanol (150 mL) for 30 min. To the resulting solution was added the corresponding
phenol or tiophenol (10 mmol, 1 eq) and further refluxed for 1 h. A solution in ethanol
(35 mL) of either 1ꢀ(2ꢀchloroethyl)piperidine hydrochloride or 1ꢀ(2ꢀchloroethyl)azepane
hydrochloride (10 mmol, 1 eq) was added to the basic solution and the reaction mixture
refluxed for further 3 h. After pouring into cold water (125 mL), the product was
extracted with DCM (4 x 125 mL) and the joined extracts were washed with water (4 x
2 4
150 mL), dried over anh. Na SO , filtered and concentrated in vacuo to obtain the
desired product. Its hydrochloride salt was prepared by adding an excess of HCl either
in diethyl ether or 1,4ꢀdioxane to a solution of the base in ethyl acetate.
5
1
N
-[2-(piperidin-1-yl)ethyl]aniline hydrochloride, 9a.
procedure B, aniline (1.82 mL, 20 mmol) and 1ꢀ(2ꢀchloroethyl)piperidine hydrochloride
1.83 g, 10 mmol) gave the product as an orange oil (2.06 g, 86% yield), that formed its
hydrochloride salt as a white solid (904 mg). The analytical sample was obtained by
Following general
(
2
9
ACS Paragon Plus Environment