4,6-Diphenylpyridines as Promising Novel Anti-Influenza Agents Targeting the PA-PB1 Protein-Protein Interaction: Structure-Activity Relationships Exploration with the Aid of Molecular Modeling

Article abstract of DOI:10.1021/acs.jmedchem.5b01935

Influenza is an infectious disease that represents an important public health burden, with high impact on the global morbidity, mortality, and economy. The poor protection and the need of annual updating of the anti-influenza vaccine, added to the rapid emergence of viral strains resistant to current therapy make the need for antiviral drugs with novel mechanisms of action compelling. In this regard, the viral RNA polymerase is an attractive target that allows the design of selective compounds with reduced risk of resistance. In previous studies we showed that the inhibition of the polymerase acidic protein-basic protein 1 (PA-PB1) interaction is a promising strategy for the development of anti-influenza agents. Starting from the previously identified 3-cyano-4,6-diphenyl-pyridines, we chemically modified this scaffold and explored its structure-activity relationships. Noncytotoxic compounds with both the ability of disrupting the PA-PB1 interaction and antiviral activity were identified, and their mechanism of target binding was clarified with molecular modeling simulations.

Full text of DOI:10.1021/acs.jmedchem.5b01935

Subscriber access provided by NEW YORK UNIV
Article
4,6-Diphenyl Pyridines as Promising Novel Anti-Influenza Agents
Targeting the PA-PB1 Protein-Protein Interaction: Structure-Activity
Relationships Exploration With the Aid of Molecular Modeling
Iuni Margaret Laura Trist, Giulio Nannetti, Cristina Tintori, Anna Lucia Fallacara, Davide
Deodato, Beatrice Mercorelli, Giorgio Palù, Maikel Wijtmans, Tzveta Gospodova, Ewald
Edink, Mark Verheij, Iwan J. P. De Esch, Lilia Viteva, Arianna Loregian, and Maurizio Botta
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.5b01935 • Publication Date (Web): 29 Feb 2016
Downloaded from http://pubs.acs.org on February 29, 2016
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 69
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
4
,6ꢀDiphenyl Pyridines as Promising Novel Antiꢀ
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Influenza Agents Targeting the PAꢀPB1 Proteinꢀ
Protein Interaction: StructureꢀActivity
Relationships Exploration With the Aid of
Molecular Modeling
1
‡
2‡
1
1
1
Iuni M. L. Trist , Giulio Nannetti , Cristina Tintori , Anna Lucia Fallacara , Davide Deodato ,
2
2
3
4
3
Beatrice Mercorelli , Giorgio Palù , Maikel Wijtmans , Tzveta Gospodova , Ewald Edink , Mark
3
3
4
2
ς
1,5ς
Verheij , Iwan de Esch , Lilia Viteva , Arianna Loregian * and Maurizio Botta
*
1
Dipartimento di Biotecnologie, Chimica e Farmacia, Università degli Studi di Siena, Via A.
2
Moro, Iꢀ53100 Siena, Italy. Dipartimento di Medicina Molecolare, Università degli Studi di
3
Padova, Via A. Gabelli 63, Iꢀ35121 Padova, Italy. Division of Medicinal Chemistry, Amsterdam
Institute for Molecules, Medicines and Systems, VU University Amsterdam, De Boelelaan 1083,
4
1
081 HV Amsterdam, The Netherlands. Institute of Organic Chemistry with Centre of
Phytochemistry, Bulgarian Academy of Sciences, Acad. Georgy Bonchev str. BI. 9, 1113 Sofia,
5
Bulgaria. Sbarro Institute for Cancer Research and Molecular Medicine, Temple University,
th
BioLife Science Building, Suite 333, 1900 N 12 Street, Philadelphia, Pennsylvania 19122,
United States
ACS Paragon Plus Environment
1

Journal of Medicinal Chemistry
Page 2 of 69
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
KEYWORDS
Influenza A virus, Antiviral agents, Viral RNA polymerase, PAꢀPB1 interaction, Docking,
Molecular dynamics simulation
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
ACS Paragon Plus Environment
2

Page 3 of 69
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
ABSTRACT
Influenza is an infectious disease that represents an important public health burden, with high
impact on the global morbidity, mortality and economy. The poor protection and the need of
annual updating of the antiꢀinfluenza vaccine, added to the rapid emergence of viral strains
resistant to current therapy make the need for antiviral drugs with novel mechanisms of action
compelling. In this regard, the viral RNA polymerase is an attractive target that allows the design
of selective compounds with reduced risk of resistance. In previous studies we showed that the
inhibition of the polymerase acidic proteinꢀbasic protein 1 (PAꢀPB1) interaction is a promising
strategy for the development of antiꢀinfluenza agents. Starting from the previously identified 3ꢀ
cyanoꢀ4,6ꢀdiphenylꢀpyridines, we chemically modified this scaffold, and explored its structureꢀ
activity relationships. Nonꢀcytotoxic compounds with both the ability of disrupting the PAꢀPB1
interaction and antiviral activity were identified and their mechanism of target binding was
clarified with molecular modelling simulations.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment
3

Journal of Medicinal Chemistry
Page 4 of 69
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
INTRODUCTION
Despite the efforts in antiviral research, influenza (Flu) is an infectious disease that still
represents an important public health burden affecting 5ꢀ10% of the World’s population,
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
resulting in 3ꢀ5 million severe cases and in 250ꢀ500 thousands deaths each year. It is caused by
the ribonucleic acid (RNA) viruses of the Orthomyxoviridae family and they are classified in A,
2
ꢀ4
B and C according to their antigenic differences. Among these, influenza A virus (FluA)
infects a variety of animals as well as humans. Hemagglutinin (HA) and neuraminidase (NA) are
the two major viral antigenic glycoproteins and the 18 HA (H1ꢀH18) and 11 NA (N1ꢀN11)
5
identified in FluA virus result in the generation of distinct viral subtypes. Some of them are
responsible for severe upper respiratory diseases in humans that occur in both seasonal
epidemics and sporadic pandemics associated with high morbidity and mortality. Millions of
6,7
deaths were recorded in the most important pandemics of the last Century. In chronological
8
,9
order, these were: the Spanish Flu (H1N1 strain, in 1918), the Asian Flu (H2N2 strain, in
9
9
1
957), the Hong Kong Flu (H3N2 strain, in 1968) and the “swine” Flu (H1N1 strain, in
10
2
009). Recently, the zoonotic infection caused by the “avian” FluA subtype H7N9 registered a
fatality rate of 25% of the confirmed cases. In this outbreak, it was proven that most of the
patients were directly infected from poultry, although the existence of limited humanꢀtoꢀhuman
spread could not be excluded. However, the possibility of sustained acquisition of humanꢀtoꢀ
human transmission capabilities by this pathogen could result in a devastating worldwide health
11
problem, raising much concern for its pandemic potential.
Reꢀassortments of the genetic material between different strains allow the virus to avoid host
detection systems and the emergence of the extremely aggressive viruses responsible for the
pandemics. These, antigenic drift in particular, are complicating factors hampering the
ACS Paragon Plus Environment
4

Page 5 of 69
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
development of an effective and universal antiꢀFlu vaccine and, thus, the development of antiꢀ
1
2,13
influenza drugs is essential for the fight of future Flu outbreaks.
Clinical therapy relies only
on two classes of drugs that inhibit either the M2 ion channel (amantadine and rimantadine) or
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
14,15
NA (e.g. oseltamivir and zanamivir).
M2 ion channel inhibitors were the first antiꢀinfluenza
drugs approved. However, they are no longer recommended due to their serious side effects on
the Central Nervous System (CNS) and the very rapid emergence of drugꢀresistant influenza
1
4ꢀ16
strains.
Furthermore, drugꢀresistant FluA viruses are also emerging for the current first line
1
4,15,17ꢀ18
therapy (NA inhibitors), making the discovery of novel antiviral strategies essential.
A new molecular target that was identified in the last few years is FluA RNAꢀdependent RNA
polymerase (RdRp), a protein complex essential for the viral replication and pathogenesis.
Indeed, it is responsible for viral RNA replication and an endonuclease activity for the “capꢀ
19
snatching”, which is necessary for both transcription initiation and translation. Importantly,
FluA RdRp is significantly different from human polymerases and it is highly conserved,
facilitating the design of selective compounds with reduced probability of virus resistance
development. RdRp is a complex of three virusꢀencoded subunits: the polymerase acidic protein
(PA), the polymerase basic protein 1 (PB1) and the polymerase basic protein (PB2). Only the
1
9ꢀ23
correct assembly of the three subunits allows the correct functioning of the enzyme.
Thus, an
attractive mechanism of FluA RdRp inhibition is the disruption of the PAꢀPB1 proteinꢀprotein
interaction. This is a very promising strategy for antiꢀinfluenza drug discovery, as demonstrated
by the emergence of proteinꢀprotein interaction (PPI) smallꢀmolecule inhibitors in recent
1
4,15,24ꢀ36
years.
Among these, a hit compound disrupting the PAꢀPB1 interaction was identified
through a highꢀthroughput docking as a potential FluA polymerase inhibitor and from this
encouraging result a small library of purchased analogues was evaluated. With IC values in the
5
0
ACS Paragon Plus Environment
5

Journal of Medicinal Chemistry
Page 6 of 69
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
micromolar range, the activity related to the 3ꢀcyanoꢀ4,6ꢀdiphenylꢀpyridine scaffold was
confirmed as being responsible for the disruption of the PAꢀPB1 complex through the interaction
of these molecules with the Cꢀterminal domain of PA, mimicking the hydrophilic Nꢀterminal
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
35
portion of PB1 (residues 1ꢀ4). In order to obtain a better understanding of the structureꢀactivity
relationship (SAR) of this promising class of compounds and aiming the improvement of their
activity, a hit optimization process was performed in the study presented here.
ACS Paragon Plus Environment
6

Page 7 of 69
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
RESULTS AND DISCUSSION
Hit exploration.
Starting from a simplified structure (compound 1) of the novel class of inhibitors identified in
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
35
the previous virtual screening campaign, a series of analogues was designed for a better
understanding of the underlying structureꢀactivity relationships. The 3ꢀcyanoꢀ4,6ꢀdiphenylꢀ
pyridine scaffold was systematically modified by either removing or replacing molecule
portions. Furthermore, attempts to increase compound solubility (e.g. 2), to address the
possibility of a covalent interaction between the protein and the thioether (e.g. 5 and 6) and to
increase the synthetic feasibility (e.g. 14 and 15) were considered. All the synthesized
compounds can be found in Table 1.
Chemistry.
Due to the removing and replacing strategy adopted in the design of the molecules, several
synthetic routes were applied. However, 2ꢀchloroꢀ4,6ꢀdiphenylnicotinonitrile (16) was used as a
common precursor for the synthesis of several compounds (1, 3ꢀ8, and 10), as shown in Scheme
1. Of these, 1, 3, 4, 7 and 8 were obtained by nucleophilic aromatic substitution using the
appropriate nucleophile (compounds 17aꢀe). Reaction conditions differed according to the
reactant properties and in several cases temperature control was important in order to avoid the
formation of a highly fluorescent ringꢀclosed side product. Extensive analysis was performed to
confirm the regiochemistry of these reactions. An efficient Heck coupling reaction between Nꢀ
(
mꢀtolyl)acrylamide (18) and precursor 16 was performed for the synthesis of compound 6,
which has a trans stereochemistry as indicated by the large J constant between the two vinylic
ACS Paragon Plus Environment
7

Journal of Medicinal Chemistry
Page 8 of 69
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
protons (14.8 Hz). This alkene was then reduced to the saturated analogue 5. Compound 10 was
synthesized by hydrogenolysis of the CꢀCl bond of compound 16.
a
Scheme 1. Synthesis of compounds 1, from 3 to 7, and 10.
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: (i) DMF, K CO , rt, 1h (for 1); MeCN, CuI, Cs CO , MWO 120°C,
2
3
2
3
1
.5h (for 3); DMF, K CO , 45°C, 18h (for 4); MeCN, K CO , 0°C, 0.5h then: rt, 3h (for 7 and 8);
2 3 2 3
(
ii) DMF, Pd(OAc) , TBAB, (oꢀtol) P, NaOAc, 145°C, 18h; (iii) EtOAc, H , Pd/C, rt, 11h; (iv)
2 3 2
NaOAc, MeOH, H , PdCl , rt, 36h
2
2
Precursors 17d and 17e were prepared through the neat acylation of 3ꢀmethylꢀaniline (19) with
either mercaptoꢀpropionic acid (20) or dihydrothiophenꢀ2(3H)ꢀone (21) respectively (Scheme 2).
Intermediate 18 was obtained through the acylation of 3ꢀmethylꢀaniline (19) with acryloyl
chloride (22, Scheme 2).
ACS Paragon Plus Environment
8

Page 9 of 69
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
a
Scheme 2. Synthesis of precursors 17d, 17e and 18.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
a
Reagents and conditions: (i) rt, 1h, then MWO, 180°C, 1h; (ii) MWO 120°C → 240°C, 28h;
(
iii) TEA, THF, 0°C, 2.5h
Final products 12, 13, 14 and 15 were synthesized by aromatic nucleophilic substitution on the
appropriate aromatic electrophile 23aꢀd with the nucleophile 2ꢀmercaptoꢀNꢀ(mꢀtolyl)acetamide
17a) in acetonitrile, as shown in Scheme 3.
(
a
Scheme 3. Synthesis of compounds 12, 13, 14 and 15.
a
Reagents and conditions: (i) MeCN, K CO , rt, 30min (for 12); MeCN, NaOAc, rt, 18h then
2
3
40°C, 5d (for 13); MeCN, K CO , rt, 5h (for 14); MeCN, K CO , 0°C to rt, 2.5h (for 15)
2
3
2
3
A different synthesis approach was required for compound 11. As shown in Scheme 4, the
pyridone 24 was chlorinated with POCl to yield 2ꢀchloroꢀ4,6ꢀdiphenylpyridine (25), which then
3
reacted with thiol (17a) through a Cuꢀcatalysed aromatic nucleophilic substitution, giving the
final product (11).
ACS Paragon Plus Environment
9

Journal of Medicinal Chemistry
Page 10 of 69
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 4. Synthesis of compound 11.
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: (i) POCl , reflux, 4d; (ii) DME, CuI, Lꢀproline, K CO , reflux in
3
2
3
sealed ampule, 30h
Scheme 5 reports the synthetic strategy used for compounds 2 and 9. The precursor 2ꢀchloroꢀ1ꢀ
morpholinoethanone (28) was synthesized through the reaction of morpholine (26) with
chloroacetylchloride (27). Then, 4,6ꢀdiphenylpyridineꢀ2ꢀthiol (29) was used as a nucleophile in
the nucleophilic substitution on 28 or iodomethane for the synthesis of the final products 2 and 9
respectively.
a
Scheme 5. Synthesis of compounds 2 and 9.
a
Reagents and conditions: (i) K CO , THF, 0°C to rt, 1h; (iiꢀa) DMF, Cs CO , TBAI, rt, 1h; (iiꢀ
2
3
2
3
b) 0°C to rt, 2h
ACS Paragon Plus Environment
10

Page 11 of 69
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Biological evaluation of the compounds.
All the synthesized molecules were evaluated for their inhibitory activity against the PAꢀPB1
interaction. The abilities of the test compounds to inhibit the physical interaction between PA
and PB1 and to block Flu virus replication were tested in EnzymeꢀLinked Immunosorbent
Assays (ELISA) and in plaque reduction assays (PRA) in MardinꢀDarby Canine Kidney
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(
MDCK) cells infected with FluA virus (A/PR/8/34 strain), respectively. In order to exclude the
possibility that the observed antiviral activity might be due to toxic effects on the infected cells,
compound cytotoxicity was also assessed in two cell lines (MDCK and human embryonic
37
kidney, HEK 293T) by the MTT assay. In these studies, the PB1(1ꢀ15)ꢀTat peptide was used as a
positive control for inhibition in the ELISA, while the broadꢀspectrum inhibitor of RNA viruses
3
8
polymerases Ribavirin (RBV) was used as a positive control in the PRA as it inhibits viral
polymerases, even though its mechanism of action that does not involve the inhibition of PAꢀ
3
9
PB1 formation. Moreover, the first line therapeutic NA inhibitor Oseltamivir (OSV) was
included as a positive control in the PRA. The results of the experiments are shown in Table 1.
Most importantly, data revealed for the first time not only that the chemical scaffold studied in
this hit optimization process is able to inhibit the formation of the FluA PAꢀPB1 complex with
comparable IC with respect to the control (PB1(1ꢀ15)ꢀTat peptide), but also that it exhibits a good
5
0
cytotoxicity profile and a low micromolar antiviral activity, comparable to RBV. Despite less
efficient antiviral activity was recorded for our compound family with respect to OSV, the
family of compounds presented here inhibits a target that is lessꢀprone to mutations than NA, for
which resistance has frequently already emerged.
ACS Paragon Plus Environment
11

Journal of Medicinal Chemistry
Page 12 of 69
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
Table 1. Structure and biological activity of the synthesized compounds.
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
Cytotoxicity
c
(MTT assay)
ELISA PAꢀ
PB1
Interaction MDCK cells
Assay
PRA in
HEK 293T
cells
f
Compd
R1
R2
R3
Z
MDCK cells
SI
a
b
d
e
IC (µM)
EC (µM)
CC (µM)
CC (µM)
5
0
50
50
50
1
2
CCN
CCN
175 ± 22
177 ± 20
9.2 ± 0.7
100 ± 9
>250
>250
>250
>27.2
>2.5
184 ± 14
3
4
5
CCN
CCN
CCN
>200
>200
>200
93.9 ± 7.8
28.6 ± 4.9
>100
>250
>250
>250
>250
>250
>2.7
>8.7
NC
140.0 ± 7.8
ACS Paragon Plus Environment
1
2

Page 13 of 69
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
6
7
8
9
CCN
190 ± 31
70.7 ± 5.7
>250
>250
>3.5
g
CCN
CCN
CCN
CCN
CH
122 ± 19
140 ± 22
195 ± 15
116 ± 26
52.6 ± 9
>10
150 ± 14
95.6 ± 15.8
144 ± 22
>250
13.9 ± 2.0
41.5 ± 5.4
125 ± 16
105 ± 7
NC
5.1
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
8.2 ± 1.1
>50
NC
2.3
10
H
45.1 ± 3.0
7.3 ± 1.3
11
>250
150 ± 19
20.5
g
12
13
14
H
CCN
CCN
N
>200
>200
>10
32.0 ± 3.6
180 ± 5
22.3 ± 1.8
>250
7.1 ± 1.1
41.5 ± 4.2
10.1 ± 1.7
>250
NC
NC
2.9
g
H
>10
165 ± 18
>200
3.5 ± 0.5
15
N
26.5 ± 4.3
>9.4
RBV
OSV
PB1(
ND
ND
12.8 ± 2.1
>250
>250
>250
>250
>19.5
0.015 ± 0.006
>16000
1–
35.9 ± 5.6
49.7 ± 5.1
>250
>250
>5.0
15)–Tat
ACS Paragon Plus Environment
1
3

Journal of Medicinal Chemistry
Page 14 of 69
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
peptide
a
Activity of the compounds in the ELISA PAꢀPB1 interaction assays. Values represent the compound concentration (in µM) that
b
reduces the interaction between PA and PB1 by 50% (IC ). Antiviral activity of the compounds in plaque reduction assays against
5
0
the Flu A/PR/8/34 strain. Values represent the compound concentration (in µM) that inhibits 50% of plaque formation (EC ).
5
0
c
Cytotoxicity of the compounds exhibited in the MTT assays. Values represent the compound concentration (in µM) that causes a 50%
d
e
f
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
decrease in cell viability (CC ). CC assessed in HEK 293T cells. CC assessed in MDCK cells. Selectivity index, defined as the
ratio between the CC value assessed in MDCK cells and the EC value. Compound concentrations higher than those reported could
not be tested in PRA due to cytotoxicity. All values represent the averages ± SD of data derived from at least three independent
experiments in duplicate. ND: not determined. NC: not calculated because of cytotoxicity or absence of antiviral activity at tested
concentrations.
50
50
50
g
5
0
50
ACS Paragon Plus Environment
1
4

Page 15 of 69
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Lower activity in the ELISA experiments compared to the PRA could be observed, in
particular for compounds 1, 4, 14 and 15. Compounds 1 and 14 exhibited a lowꢀmicromolar EC50
value in PRA (9.2 and 3.5 µM, respectively) and a 19ꢀ47−fold higher IC value in ELISA (175
5
0
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
and 165 µM, respectively), while 4 and 15 presented an EC around 30 µM in PRA (28.6 and
50
26.5 µM, respectively) and complete loss of the ability to disrupt the PAꢀPB1 interaction in
ELISA. In order to investigate whether this could be due to the assay conditions, the ELISA was
repeated for all molecules with EC values below 30 µM (1, 4, 8, 11, 14, 15). In this assay
5
0
(henceforth mentioned as “Modified ELISA PAꢀPB1 Interaction Assay”), compounds were
dissolved in a similar medium to the one used in the cellꢀbased assays instead of the PBS buffer
(see Experimental Section). Similarly to the previous ELISA PAꢀPB1 Interaction Assay, the
37
PB1(1ꢀ15)ꢀTat peptide was used as a positive control for inhibition. As reported in Table 2, a
significant improvement in the inhibitory activity was obtained mainly for 1 and 15, which were
more soluble in the DMEM serumꢀfree buffer with respect to the PBS one, while more modest
improvements were obtained for 11 and 14. The same IC values were registered for compounds
5
0
4
and 8. As expected, comparable IC values were also obtained for the PB1(1ꢀ15)ꢀTat peptide in
50
the two assay conditions, suggesting that the medium was affecting the behaviour of some of our
compounds rather than the accessibility of the target site.
Table 2. Inhibitory activity of selected compounds on the PAꢀPB1 interaction in the modified
ELISA.
Modified ELISA PAꢀ
PB1 Interaction Assay
Compd
a
IC (µM)
5
0
1
4
35.2 ± 5.1
>200
ACS Paragon Plus Environment
15

Journal of Medicinal Chemistry
Page 16 of 69
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
8
145 ± 24.0
1
1
4
5
40.5 ± 15.6
113 ± 15
1
1
64.5 ± 5.4
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
PB1(1–15)–Tat
peptide
31.7 ± 10.8
a
Ability of the compounds to inhibit the PAꢀPB1 interaction in the ELISA performed with the
DMEM serumꢀfree medium in place of the PBS buffer. Values represent the compound
concentration (in µM) that reduces the interaction between PA and PB1 by 50% (IC ). Values
represent the averages ± SD of data derived from at least two independent experiments in
duplicate.
5
0
Data presented in both Table 1 and Table 2 are of great value for the understanding of the
general mechanism of action of the 3ꢀcyanoꢀ4,6ꢀdiphenylꢀpyridine scaffold. A significant
decrease in activity was observed in the attempts of substituting the sulphur atom (3, 4, 5 and 6),
indicating the importance of this heteroatom. Only for compound 4 the antiviral activity was
partially retained. The whole sulphurꢀcontaining side chain appears to be also essential as its
complete (10) or its partial (9) removal resulted in a reduction or a complete loss of activity.
Nevertheless, a mild antiviral activity was observed for compound 10. However, considering its
significant cytotoxicity in MDCK cells and the high IC value in ELISA, its antiviral activity in
5
0
PRA might be related to cell toxicity rather than to the inhibition of PAꢀPB1 interaction. This
hypothesis is in keeping with the observation that the compound 10 lacks of one of the
interaction with a relevant hydrophobic area within the PA binding region. Only one methylene
group should be placed between the sulphur atom and the carbonyl moiety, as the extension of
the chain (7 and 8) increased cytotoxicity. Furthermore, the 3ꢀmethyl substituted phenyl ring
cannot be replaced by a more hydrophilic group, as indicated by the increased EC value of
5
0
compound 2. Given the inactivity and the higher cytotoxicity of 12 and 13, the two phenyl rings
directly attached to the pyridine ring (in positions 4 and 6) appear to be essential in the molecular
ACS Paragon Plus Environment
16

Page 17 of 69
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
scaffold. Interestingly, the removal of the cyano moiety linked to the same central core (11) did
not affect activity. Good antiviral activity was observed also with the substitution of the pyridine
ring with a pyrimidine one (14), but a lower ability to dissociate the PAꢀPB1 complex and a
higher cytotoxicity were detected as well. However, this toxicity decreased with the introduction
of two electron withdrawing CF groups in the structure (15). For this compound, a better IC
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
3
50
value was obtained in the modified ELISA, but a worse antiviral activity profile in respect to
compounds 1, 8, 11, and 14 was also detected.
From these results, compounds 1, 11, and 15 emerged as the most promising molecules since
they display micromolar inhibition of the PAꢀPB1 interaction (with IC values ranging from
5
0
35.2 to 64.5 µM), interesting antiviral activity and a good cytotoxicity profile with a selectivity
index (SI) higher than 9.4. Among the other compounds, 4 showed an antiviral activity
comparable to 15 and a good cytotoxicity profile (SI > 8.7), but no inhibition of the PAꢀPB1
interaction. Thus, aiming to better understand the mechanism of action of these compounds, the
ability of 1, 4, 11 and 15 to inhibit FluA RNAꢀdependent RNA polymerase (RdRp) enzymatic
activity in a cellular context was assessed by minireplicon assays in transfected HEK 293T cells,
3
8
using RBV as a reference inhibitor. The results are presented in Table 3. All of the tested
compounds were found to inhibit the viral polymerase activity, with encouraging EC values for
5
0
compounds 11 and 15 (22.0 to 30.8 µM) in particular. Furthermore, from the results it emerges
that probably compound 1, in addition to the inhibition of the PAꢀPB1 interaction, might be
involved in other mechanisms of action not related to the viral polymerase activity because its
low EC50 value in the PRA assay (9.2 µM) does not reflect completely the activity in the
minireplicon assay (EC50 86.6 µM). Compound 4 inhibited the catalytic activity of the viral
RdRp in the 293T cells (EC 46.3 µM), indicating that this molecule might exert its antiviral
5
0
ACS Paragon Plus Environment
17

Journal of Medicinal Chemistry
Page 18 of 69
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
effects by inhibiting the the viral RdRp activity, but through a mechanism of action that likely
does not involve the PAꢀPB1 complex disruption.
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
Table 3. Activity of the selected compounds in the minireplicon assay.
Minireplicon Assay
Compd
a
EC (µM)
50
1
4
86.6 ± 15.0
46.3 ± 8.7
30.8 ± 5.3
22.0 ± 4.1
23.8 ± 4.5
11
15
RBV
a
The EC values represent the compound concentration that reduces by 50% the activity of
50
FluA virus RNA polymerase in HEK 293T cells. All values represent the averages ± SD of data
derived from at least two independent experiments in duplicate.
The ability of these compounds to inhibit the polymerase activity in cells as well as the PAꢀ
PB1 interaction and the formation of viral plaques is a substantial indication that the most
promising 4,6ꢀdiphenyl pyridine (or pyrimidine) molecules exert their antiꢀinfluenza activity
through the disruption of the PAꢀPB1 complex.
Molecular modelling.
In order to rationalize the SARs, all the synthesized compounds were examined by means of
docking simulations within the PB1ꢀbinding site of PA in order to produce their putative binding
poses.
The 3ꢀcyanoꢀ4,6ꢀdiphenyl pyridine series was discovered with a previous highꢀthroughput
docking study performed on a homology model built using the FluA H1N1 sequence and two
2
0
21 35
PAꢀPB1 complexes templates (PDB IDs: 2ZNL and 3CM8 ). Subsequent to the publication
ACS Paragon Plus Environment
18

Page 19 of 69
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
35
of the manuscript by Tintori et al., a novel Xꢀray structure of FluA H17N10 RdRp was
2
3
published (PDB ID: 4WSB) and comprised the whole PA in complex with PB1, PB2 and a
viral RNA promoter. Since the Cꢀterminal portion of PA (PAC) in 4WSB is more complete than
the structures used to build the PAꢀPB1 complex in silico (2ZNL and 3CM8), the two PAC were
compared in order to assess the quality of the homology model. The two structures have 71.21%
identity in sequence alignment and they superpose very well (RMSD 1.09Å). Remarkably, the
superposition is even better in the PB1ꢀbinding site region (RMSD of 0.537Å), validating the
reliability of the homology model. The backbone superposition of the homology model and PDB
ID 4WSB is shown in Figure S1 of the Supplementary Information. Since biological evaluations
were performed on the FluA H1N1 strain, and since the homology model was assessed of good
quality, for consistency purposes the homology model was used for the docking simulations.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
35
Similarly to the procedure applied in the previous study, docking simulations were carried
4
0
41
out with two widely used programs (Glide and GOLD ) and the binding poses were then
compared. Most importantly, only the active compounds reproduced the already reported binding
35
model of the 3ꢀcyanoꢀ4,6ꢀdiphenylꢀpyridine derivatives, confirming the proposed mode of
binding of this class of inhibitors. The pose obtained for compound 1 is shown in Figure 1 as an
example. In this model, two hydrogen bonds between the amide portion of the ligands and
residues Gln408 and Val621 resemble PB1ꢀPA interactions. When present in the ligand, a third
hydrogen bond occurs between the CN nitrogen of the compound and Lys643 side chain amine.
The three phenyl rings are accommodated in three hydrophobic areas indicated as I (Val621,
Phe411 and Cys415), II (Pro620, Val628 and His713) and III (Phe658, Leu666, Phe707 and
Phe710) in the figure. A further hydrophobic interaction involves Trp706 and the pyridine (or
pyrimidine) ring of the small molecule.
ACS Paragon Plus Environment
19

Journal of Medicinal Chemistry
Page 20 of 69
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 1. Binding pose for compound 1 (orange sticks).
Two examples of bad ligand placements within the binding site are shown in Figure 2, where
the conformations of 12 and 3 (cyan sticks) are compared with that of compound 1 (orange
sticks). As previously discussed, compounds lacking of important portions of compound 1 were
not able to reproduce fundamental proteinꢀligand interactions in the docking simulations. In
particular, the increase of IC values of molecules lacking of either one of the phenyl groups
5
0
that interact with the protein regions II and III could be explained by the impossibility to form
any of the important interactions, most of the times presenting a completely different binding
mode, as exemplified in panel A in Figure 2. With the substitution of the sulphur atom
(compounds 3ꢀ5) and with the introduction of a double bond (6) the effective interactions were
sometimes partially reproduced (exemplified in panel B in Figure 2) or completely lost. The
different geometry properties of the atoms substituting the sulphur atom and of the CꢀC double
bond do not allow the small molecule to assume a suitable conformation for correct binding,
resulting in a bad docking pose and explaining the inactivity of these molecules.
ACS Paragon Plus Environment
20

Page 21 of 69
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 2. Examples of bad binding of inactive compounds. In both images compound 1 (orange
sticks) is used as a reference. A) Binding pose of compound 12 (cyan sticks). B) Binding pose of
compound 3 (cyan sticks).
Compounds 11, 14 and 15, all lacking the nitrile group, were able to reproduce all the
important interactions, except the hydrogen bond with Lys643, thus overlapping very well with
the binding pose of 1 (Figure 3). The overlapping binding poses of compounds 11, 14 and 15 can
be found in Figure S1 of the Supplementary Information. From the binding model, we expected
that the lack of the hydrogen bond with Lys643 would result in a less favourable binding of the
small molecules to PA, with a decrease in the ability to disrupt the PAꢀPB1 interaction, but this
was not confirmed by the biological data. Thus, the PAꢀcompound complexes for the most
ACS Paragon Plus Environment
21

Journal of Medicinal Chemistry
Page 22 of 69
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
promising compounds (1, 11 and 15) were further investigated by means of molecular dynamics
4
2
(MD) using Schrödinger Desmond.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 3. Binding pose of compounds 1 (orange sticks) and compound 11 (cyan sticks) from
docking simulations.
After extensive equilibration, a 14ns MD simulation was performed for each complex starting
form the most representative docking pose. The evolution in time of the rootꢀmeanꢀsquare
deviation (RMSD) calculated on the backbone atoms of the protein in respect to the starting
structure is presented in panel A of Figure 4. After an initial increase due to further equilibration
of the complex, the RMSD values were less variable in both simulations, in particular for the
compound 11 in complex with PA, whose RMSD stabilised around 3.19Å after 3ns of
simulation. For the molecular system containing compound 1 the average RMSD calculated
between 2 and 8ns was 2.61Å, while it increased up to 3.44Å (at 12ns) after this time period.
RMSD average values for the PAꢀ15 complex were 2.15Å and 2.77Å between 3 and 8ns and
after 8ns respectively. The rootꢀmeanꢀsquare fluctuations (RMSFs) of the backbone atoms were
then evaluated on the whole trajectories. As shown in panel B of Figure 4, the RMSF trends were
the same for the three systems. Predictably, high fluctuations were recorded for the protein
ACS Paragon Plus Environment
22

Page 23 of 69
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
portions (residues 349ꢀ353, 372ꢀ388, 389ꢀ396 and 550ꢀ557) that were not solved in the original
k
crystal structure (PDB ID 2ZNL ). Higher values generally corresponded to random coils and did
not include residues interacting with the ligands (highlighted in purple), suggesting that the
structure of PB1ꢀbinding site within PA was stable during the production MD. This was further
supported by cluster analysis that indicated high similarity of the binding site conformation
between the clusters. Figure 5 shows the superposition of the average structures of the ten
clusters obtained for compound 1, 11 and 15 complexes (panels A, B and C respectively). In all
panels, one structure of the ligand (in sticks) is shown for a better identification of the binding
site.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment
23

Journal of Medicinal Chemistry
Page 24 of 69
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 4. A) Time evolution of the backbone atoms RMSD values for the complexes between
PA and compounds 1 (blue line), 11 (red line) and 15 (green line). B) RMSF evaluation of the
backbone atoms for the complexes between PA and compounds 1 (blue line), 11 (red line) and
15 (green line). Ligand contacts are highlighted in purple.
ACS Paragon Plus Environment
24

Page 25 of 69
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 5. Superposition of the average structures of the 10 clusters obtained by cluster analysis.
Protein backbone is shown as ribbon and the ligand structure in sticks. A) Clusters for the PAꢀ
compound 1 complex. B) Clusters for the PAꢀcompound 11 complex. C) Clusters for the PAꢀ
compound 15 complex.
For a more detailed analysis of the stability of the PAꢀligand complexes, the interactions
between the protein and the small molecules were investigated on the whole 14ns simulation
time. Occupancy percentages were also calculated and are listed in Table 4. Hydrogen bonds and
ionic interactions were the most stable contacts between PA and the ligand, with occupancy
higher than 60%, whereas hydrophobic interactions were generally weaker (occupancy less than
30%). An exception was the hydrophobic interaction with Trp706 that resulted to be stable and
important, particularly for compound 11 (65.90% occupancy) and secondly for compound 15
(50.72% occupancy). Predictably, the ionic contact between Lys643 side chain amine group and
the compounds was much more important in the complex between the protein and 1, where the
interaction involved the CN group of the small molecule. Differently, negligible or weak ionic
contacts occurred between Lys643 and the sulphur atom of 11 and 15 respectively. Not predicted
by the docking simulation, a πꢀcation contact between Lys643 side chain amine and the phenyl
moiety in position 4 in the pyridine core of the ligands was detected and it appeared more stable
for compound 1 (74.54% versus 31.30% occupancy for 11). This interaction was however
ACS Paragon Plus Environment
25

Journal of Medicinal Chemistry
Page 26 of 69
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
negligible for compound 15 (3.54% occupancy), where the πꢀcation contact was compensated by
a stronger ionic interaction of Lys643 with the sulphur atom (25.88% occupancy) and by a weak
hydrogen bond between Lys643 and the carbonyl oxygen of the ligand (15.78% occupancy).
Importantly, Lys643 was also involved in a waterꢀbridge contact with the pyrimidine nitrogen in
position Z (according to Table 1; henceforth named N1) of compound 15 with 46.09%
occupancy, while this type of interaction is not possible for compounds 1 or 11.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
ACS Paragon Plus Environment
26

Page 27 of 69
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Table 4. Ligand interactions occupancies for 1 and 11 in complex with PA.
c
Occupancy (%)
Comp 11
98.36
97.05
6.38
Ligand
atom/group
Type of
a
b
Residue(atom)
Val621(O)
Gln408(HE)
Lys643(NZ)
Lys643(HZ)
Lys643
contact
bbHB
scHB
I
Comp 1
25.10
22.04
56.35
0.00
Comp 15
24.97
14.68
25.88
15.78
3.54
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
NH
O
CN or S
O
scHB
Pcat
Hyd
Hyd
PP
0.00
4ꢀphenyl
pyridine
3ꢀMeꢀphenyl
3ꢀMeꢀphenyl
3ꢀMeꢀphenyl
6ꢀphenyl
6ꢀphenyl
4ꢀphenyl
4ꢀphenyl
4ꢀphenyl
4ꢀphenyl
4ꢀphenyl
74.54
35.41
16.80
30.29
5.93
31.30
65.90
15.63
1.23
Trp706
50.72
27.31
33.20
3.60
Phe411
Phe411
Cys415
Hyd
Hyd
Hyd
Hyd
Hyd
Hyd
PP
6.72
Pro620
4.15
0.24
0.00
Val628
3.22
0.79
3.05
Leu666
2.88
2.47
0.79
Phe707
23.69
46.97
25.77
10.04
NA
12.61
24.16
29.66
14.20
NA
30.02
19.01
17.75
0.00
Phe710
Phe710
His713
PP
c
Lys643
N1
WB
46.09
a
Only for hydrogen bonds and ionic contacts, the interacting atom is reported in brackets using
the following abbreviations: O=backbone oxygen; HE=hydrogen of the side chain amide;
b
NZ=nitrogen of the side chain amine; HZ=hydrogen side chain. Type of interaction between the
protein and the ligand. Abbreviations used: bbHB=hydrogen bond with residue backbone;
scHB=hydrogen bond with residue side chain; I=ionic interaction; Pcat=πꢀcation interaction;
c
Hyd=Hydrophobic interaction; PP=πꢀπ stacking; WB=water bridge. N1= pyrimidine nitrogen in
d
position Z in the compound, according to Table 1 numeration. Percentage of trajectory frames in
which the interaction was recorded; NA=not applicable.
ACS Paragon Plus Environment
27

Journal of Medicinal Chemistry
Page 28 of 69
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
Remarkably, hydrogen bonds between the amide portion of compound 11 and Val621
backbone oxygen and Gln408 side chain amide hydrogen presented occupancy values close to
1
00%; while the same interactions in the other PAꢀcompound complexes were much weaker,
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
with occupancy lower than 30%. From the visualization of the trajectories a change of
conformation of 1 and of 15 were observed. In the PAꢀcompound 1 trajectory, while the
interaction between the ligand’s CN group and Lys643 occurred during the whole duration of the
simulation, the hydrogen bonds with Val621 and Gln408 broke after approximately 4ns. An
approximately 1.7Å shift of the pyridine core of 1 weakened the interaction of Trp706, placed
the phenyl ring in position 4 in the pyridine in a more suitable area for the πꢀcation contact with
Lys643 and enhanced the hydrophobic interactions with the areas II and III, explaining the
higher occupancy values for these contacts (Lys643 πꢀcation contact and Phe707 and Phe710
hydrophobic interaction in particular). Furthermore, lacking the constraints given by the
hydrogen bonds involving the ligand’s amide moiety, the methylꢀsubstituted phenyl ring was
able to give weak Tꢀshaped πꢀπ stacking with Phe411 in the hydrophobic area I. As an example
of this change in conformation, panel A of Figure 6 shows the superposition of a frame saved
from the trajectory after 4.8ns of simulation with the starting structure of the PAꢀcompound 1
complex. Similarly to what observed for the system containing compound 1, after approximately
3
.7ns, a 2.3Å shift of the central aromatic ring occurred in the PAꢀ15 simulation. Also in this
case, the change in binding mode weakened the interaction with Trp706, increased the
hydrophobic interactions with the areas II and III, with residues Phe411 and Phe707 in particular,
and allowed the establishment of a weak Tꢀshaped πꢀπ stacking with Phe411. However, in this
simulation Lys643 did not give a πꢀcation interaction with the phenyl moiety, but it interacted
with compound 15 through a direct hydrogen bond (with the carbonyl oxygen) and a water
ACS Paragon Plus Environment
28

Page 29 of 69
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
bridge (with the pyrimidine N1 nitrogen), mainly after the compound repositioning.
Interestingly, the water molecule that bridges the interaction between Lys643 and N1 overlaps
very well the CN nitrogen of 1, when the structures of the two complexes extracted after 4ns
simulation are superposed (panel B of Figure 6). Thus, in the case of 15, the water molecule had
the same role of the CN group of 1 in the interactions with the protein and this explains why the
MD simulation for the PAꢀ15 system was more similar to the one containing 1 than to the one
containing 11.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 6. A) Change in proteinꢀligand interactions during the MD simulation of 1 in complex
with PA. The ligands and the protein residues discussed in the text are shown as sticks and as
lines respectively. Carbon atoms of the starting complex structure are coloured in orange; while
ACS Paragon Plus Environment
29