New 7-methylguanine derivatives targeting the influenza polymerase PB2 cap-binding domain

Article abstract of DOI:10.1021/jm401369y

The heterotrimeric influenza virus polymerase performs replication and transcription of viral RNA in the nucleus of infected cells. Transcription by "cap-snatching" requires that host-cell pre-mRNAs are bound via their 5′ cap to the PB2 subunit. Thus, the PB2 cap-binding site is potentially a good target for new antiviral drugs that will directly inhibit viral replication. Docking studies using the structure of the PB2 cap-binding domain suggested that 7-alkylguanine derivatives substituted at position N-9 and N-2 could be good candidates. Four series of 7,9-di- and 2,7,9-trialkyl guanine derivatives were synthesized and evaluated by an AlphaScreen assay in competition with a biotinylated cap analogue. Three synthesized compounds display potent in vitro activity with IC₅₀ values lower than 10 μM. High-resolution X-ray structures of three inhibitors in complex with the H5N1 PB2 cap-binding domain confirmed the binding mode and provide detailed information for further compound optimization.

Full text of DOI:10.1021/jm401369y

Subscriber access provided by BOSTON UNIV
Article
New 7-methyl-guanosine derivatives targeting the
influenza polymerase PB2 cap-binding domain
Stephane Pautus, Peter Sehr, Joe Lewis, Antoine Fortune, Andrea Wolkerstorfer, Oliver
Szolar, Delphine Guilligay, Thomas Lunardi, Jean-Luc Decout, and Stephen Cusack
J. Med. Chem., Just Accepted Manuscript • Publication Date (Web): 18 Oct 2013
Downloaded from http://pubs.acs.org on October 20, 2013
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 78
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
New 7ꢀmethylꢀguanosine derivatives targeting the influenza
polymerase PB2 capꢀbinding domain
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
1
,2#
3
3
1
Stéphane Pautus , Peter Sehr , Joe Lewis , Antoine Fortuné ,
4
4
2
2
Andrea Wolkerstorfer , Oliver Szolar , Delphine Gulligay , Thomas Lunardi ,
1
*
2*
JeanꢀLuc Décout and Stephen Cusack
1
Université de Grenoble I/CNRS, UMR 5063, Département de Pharmacochimie
Moléculaire, ICMG FR 2607, 470 rue de la Chimie, BP 53, Fꢀ38041 Grenoble, France.
2
Grenoble Outstation, European Molecular Biology Laboratory and Univ. Alpesꢀ
GrenobleꢀEMBLꢀCNRS Unit for Virus Host Cell Interactions, 6 rue Jules Horowitz,
BP181, Fꢀ38042 Grenoble Cedex 9, France.
3
Chemical Biology Core Facility, European Molecular Biology Laboratory,
Meyerhofstrasse 1, 69117 Heidelberg, Germany.
4
Savira pharmaceuticals GmbH, Vienna, Austria.
th
th
Submitted to J. Med. Chem, September 5 2013. Revised October 14 2013.
*
To whom correspondence should be addressed:
Eꢀmail : JeanꢀLuc.Decout@ujfꢀgrenoble.fr. Phone +33476635317; fax +33476635298.
Eꢀmail : cusack@embl.fr. Phone +33476207238
#
Current address: Laboratoire d'hématologie, Faculté de pharmacie, Université Paris Sud
XI, 5 rue JeanꢀBaptiste Clément, 92290 ChatenayꢀMalabry
ACS Paragon Plus Environment
1

Journal of Medicinal Chemistry
Page 2 of 78
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. The heterotrimeric influenza virus polymerase performs replication and
transcription of viral RNA in the nucleus of infected cells. Transcription by ‘capꢀ
snatching’ requires that hostꢀcell preꢀmRNAs are bound via their 5' cap to the PB2
subunit. Thus, the PB2 cap binding site is potentially a good target for new antiviral drugs
that will directly inhibit viral replication. Docking studies using the structure of the PB2
cap binding domain suggested that 7ꢀalkylguanine derivatives substituted at position Nꢀ9
and Nꢀ2 could be good candidates. Four series of 7,9ꢀdiꢀ and 2,7,9ꢀtriꢀalkyl guanine
derivatives were synthesized and evaluated by an AlphaScreen assay in competition with a
biotinylated cap analogue. Three synthesised compounds display potent in vitro activity
with IC50 lower than 10 ꢀM. Highꢀresolution Xꢀray structures of three inhibitors in
complex with the H5N1 PB2 capꢀbinding domain confirmed the binding mode and
provide detailed information for further compound optimization.
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
Keywords: influenza, polymerase, PB2, capꢀbinding domain, guanine, 7ꢀmethylguanine
derivatives, docking
ACS Paragon Plus Environment
2

Page 3 of 78
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Introduction
Influenza A viruses are negativeꢀstrand, segmented RNA viruses belonging to the
family Orthomyxoviridae. There are several subtypes labeled according to the type of
hemagglutinin (H) and neuraminidase (N). Influenza A viruses primarily infect wild
aquatic birds but domestic poultry and several mammals including humans can also be
infected. Avian strains do not readily infect humans but can be transmitted to humans and
other mammals. By reassortment highly infectious viruses can be generated leading to
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
1
potentially devastating pandemics . Since 1997, the emergence of highly virulent H5N1
avian strains has led to several hundred reported infections in human, with a very high
mortality
rate
of
about
60%
(http://www.who.int/csr/disease/avian_influenza/cases_table/). Very recently a new H7N9
strain has emerged in China which has caused severe disease and around 20% fatalities in
2
infected humans . The controversial demonstration that H5N1 strains require only a few
3, 4
mutations to acquire the ability to be transmitted between certain mammalian species
5
highlights the need for new treatments to counter severe influenza infections . Even for
less virulent strains such as the 2009 H1N1 pandemic strain or indeed seasonal influenza
cases, there are vulnerable risk groups potentially requiring intensive care that would
benefit from the availability of alternative, efficacious treatments.
The influenza A virus genome comprises eight separate RNA segments that code
for up till now eleven identified proteins (HA, NA, NP, M1, M2, NS1, NEP, PA, PB1,
PB1ꢀF2, PB2). These RNA segments are packaged into ribonucleoprotein complexes
(RNPs) containing the nucleoprotein (NP) and the viral RNAꢀdependent RNA
ACS Paragon Plus Environment
3

Journal of Medicinal Chemistry
Page 4 of 78
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
6
polymerase, which is a heterotrimer comprising subunits PA, PB1 and PB2 . The
replicative cycle offers several opportunities for chemotherapeutic interventions and thus
many antiꢀinfluenza agents targeting the influenza replicative cycle have been or are being
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
7
developed . Neuraminidase inhibitors (e.g. oseltamivir and zanamivir) and M2 proton
channel inhibitors (e.g. amantadine and rimantadine) are currently available in the clinic
as antiꢀinfluenza drugs, however resistance in both cases has been observed. Resistance to
8
neuraminidase inhibitors has been reported in the case of the 2013 H7N9 outbreak . There
9
is thus a consensus view that new antiꢀinfluenza treatments are required . Our strategy is
to target the viral RNA polymerase which is responsible for replication and transcription
1
0
of the viral RNA in the nucleus of infected cells .
For efficient translation and resistance to 5' exonucleases, mRNAs are generally 5'
capped. This coꢀtranscriptional enzymatic modification consists of a 7ꢀmethyl guanosine
7
(
m G) moiety connected to the mRNA via an unusual 5'ꢀ5' triphosphate linkage. Influenza
virus, which does not encode a capping enzyme, uses a unique capꢀsnatching mechanism
to ensure that its mRNAs are capped. The PB2 subunit binds to the 5'ꢀcap of host cell preꢀ
1
1
mRNAs , which are then cleaved at nucleotide 10ꢀ13 by the endonucleolytic activity of
1
2, 13
the PA subunit
. These capped oligomers are then used by the PB1 subunit as primers
for transcription of viral mRNAs.
An independently folded domain of PB2 has been identified, which has specific in
1
1
vitro cap binding activity . The Xꢀray structure of the domain with bound 7ꢀ
7
methylguanosine triphosphate (m GTP) has been obtained revealing a mode of binding
11
similar, but distinct from, other cap binding proteins . This combined with the fact that
ACS Paragon Plus Environment
4

Page 5 of 78
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
PB2 capꢀbinding site is essential for capꢀdependent transcription by viral RNPs makes it
potentially a very good target for the development of new antiviral drugs.
Here, we report the design and the synthesis of potent in vitro inhibitors of PB2
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
1
1
cap binding based on our previously reported Xꢀray crystal structure . Docking studies
were performed in order to select substituents on the guanine moiety that could improve
its affinity for the ligand binding site. The results of in vitro and cellꢀbased inhibition
assays are presented as well as the highꢀresolution Xꢀray structures of two of the inhibitors
in complex with the H5N1 capꢀbinding domain.
Results
Inhibitor design by docking.
Previous crystallographic studies allowed us to define the mode of recognition of
7
11
the m G by the PB2 capꢀbinding domain of an H3N2 strain . As for the competing hostꢀ
14
15
7
cell capꢀbinding proteins, eIF4E and CBC , the m G is sandwiched between two
aromatic residues forming strong πꢀstacking interactions. In PB2, those two aromatic
residues are Phe404 and His357. The guanine base is also recognized via hydrogen bonds
to key residues Glu361 (to Nꢀ1 and the Nꢀ2 of the guanine) and Lys376 (to Oꢀ6). On the
other hand, unlike other cap binding proteins, the ribose has no direct interactions with the
protein residues. The phosphate groups interact with PB2 residues His432, Asn429,
His357, Lys339 and Arg355. However it should be borne in mind that the triphosphate
might have a different conformation in the case of an extended capped
oligoribonucleotide.
ACS Paragon Plus Environment
5

Journal of Medicinal Chemistry
Page 6 of 78
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 order to analyze the interactions of some inhibitor candidates with the active site
and to rationalize the choice of ligands, molecular docking using the H3N2 capꢀbinding
16
7
domain structure was performed using the Autodock4 software . Docked m GTP
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
displayed a similar position to that observed in the Xꢀray structure (Figure 1A, RMSD =
1
.21 Å). This demonstrated that Autodock4 is a reliable tool that can be used to design
new compounds interacting with the PB2 cap binding domain. Indeed, for all coꢀcrystal
7
structures described below, the position of the m G heterocycle is essentially unchanged.
Therefore, for each docking, the cluster containing the guanine base in a similar position
to that of the Xꢀray structure was selected.
7
The strong interactions observed between the active site and m G suggest that it
will be difficult to improve the activity by modifying the structure of the guanine base.
Therefore, our studies focussed on replacing the ribose and the triphosphate groups at the
Nꢀ9 position with substituents having similar properties. Addition of substituents such as
carboxylates groups could mimic the negative charges of the triphosphate group and
interact strongly with the active site. Docking studies confirmed this possibility and
compounds that contained two carboxylate groups are generally predicted to display better
affinity.
Introduction of different hydrophobic groups (methyl, ethyl, allyl and benzyl) at
the Nꢀ7 position of the guanosine was also investigated. However the hydrophobic pocket
in that position is very small and the length of the alkyl chain should probably not exceed
three carbon atoms. Substitutions at Nꢀ2, Oꢀ6 and Cꢀ8 positions were also considered.
ACS Paragon Plus Environment
6

Page 7 of 78
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Docking suggested that no substituents were tolerated in Oꢀ6 position, while substitution
at Nꢀ2 and Cꢀ8 positions could results in higher affinity.
Overall, about 400 compounds were screened in docking studies on the PB2 cap
binding domain. About 40 target molecules divided in to four series were selected.
The first series of compounds was based on the compound RO0794238 (Figure 2)
which was previously reported to have an IC = 45 ꢁM (compared to IC = 113 ꢁM for
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
5
0
50
7
17
m GTP) against vRNPs . RO0794238 also showed good selectivity towards viral RNP
compared to the translation initiation factor eIF4E. Docking studies showed that
replacement of the butanol chain by an alkyl chain (n= 1ꢀ5) with a terminal carboxylate
function could mimic a phosphate group and improve the affinity (Figure 2). However,
docking of RO0794238 showed that the size of the benzyl group at the Nꢀ7 position would
prevent correct binding of the guanine base, whereas smaller substituents e.g. methyl and
ethyl groups could be accommodated. During the chemical synthesis of RO0794238,
dialkylated side products containing substituents both at Nꢀ2 and Nꢀ9 positions were
obtained (Scheme 1). Docking suggested that the carboxylate group of the chain at the Nꢀ2
position could form some additional hydrogen bonds with protein residues and thus
increase the affinity. Consequently, different 2,9ꢀdialkylated compounds were synthesized
and evaluated for PB2 cap binding inhibition.
The second series of compounds synthesized was based on the peptide nucleic
1
8
acid (PNA) structure (Figure 3) . PNAs are artificial polymers similar to DNA or RNA
which are resistant to enzymatic degradation and made of an aminoethylglycine
polyamide (peptide) backbone bearing nucleobase pendant groups. The number of bonds
ACS Paragon Plus Environment
7

Journal of Medicinal Chemistry
Page 8 of 78
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
along the backbone and connecting the backbone with the nucleobases is identical to those
of DNA or RNA. In PNA, the presence of amide functions allows the perfect orientation
of the nucleobase for hybridization with a complementary RNA or DNA strand.
Therefore, a carbonyl or an amide function introduced like in PNA in βꢀposition of the
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
7
guanine Nꢀ9 of the potential inhibitors could mimic a part of the ribose ring of the m GTP
cap. Consequently, we introduced an amino acid (D or L) in the guanine Nꢀ9 position in
order to reproduce an element of the PNA structure and also to keep carboxylate groups
(Figure 3). Docking showed that in all cases both enantiomers (D and L) fit in the active
site and that compounds with amino acids containing two carboxylate groups (Glu, Asp)
generally displayed higher predicted affinity for PB2.
Concerning the third series of analogues synthesized, a substituted ketomethylene
group was attached to the Nꢀ9 (Figure 4). In addition to the ketone function introduced in
the second series of compounds, an aromatic ring was added to improve the selectivity
and the πꢀstacking with the several aromatic residues involved in the base recognition and
particularly Pheꢀ323. Substitution of the aromatic ring with hydrophilic substituents
(hydroxy, carboxylic acid, amide) could also increase the binding through formation of
hydrogen bonds. Docking also suggested that filling the large hydrophobic pocket on top
of the active site could improve the affinity, so compounds bearing a methyl or a phenyl
substituent on the Nꢀ9 methylene group in αꢀposition of the carbonyl were also
synthesized.
The aromatic ring in the third series appeared to be slightly too far to stack with
Pheꢀ323 residue. Improvement of the stacking by linking the aromatic ring directly on the
ACS Paragon Plus Environment
8

Page 9 of 78
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
guanine Nꢀ9 could result in higher affinity. So, four Nꢀ9ꢀphenylguanine derivatives, three
of them bearing on the phenyl ring one or two substituents (methoxy and/or carboxylic
acid groups) were synthesized forming the fourth series.
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
Synthesis of new inhibitors.
The synthesis of the first three series followed the same scheme of synthesis
(Scheme 1). Because of the acidity of the Hꢀ1, it was first necessary to protect this
position before alkylating in Nꢀ9 position. In the literature, this function is usually
1
9, 20
protected with 6ꢀchloroguanine (6ꢀCG)
. 6ꢀCG is a commercial product and after
21
alkylation, the chlorine can be hydrolyzed to reform the guanine heterocycle . However,
6ꢀCG is an expensive reagent that does not allow production of many compounds in large
amounts. As a result, we chose to use a cheaper alternative, 2ꢀNꢀacetylꢀ6ꢀOꢀ
diphenylcarbamoylguanine 1. As well as protecting both Oꢀ6 and Nꢀ2 positions, the size
of the diphenylcarbamoyl group was able to sterically induce the selectivity for the
22
alkylation at Nꢀ9 position towards Nꢀ7 . Also the protected product could be obtained in
a large scale with very high yield (95% in our case) and protecting groups could both be
removed in the same step under mild conditions (NH /MeOH, 70 °C).
3
2ꢀNꢀacetylꢀ6ꢀOꢀdiphenylcarbamoylguanine 1 was prepared according to the
2
3
method developed by Robins et al. .
For the second series of compounds (the amino acid derivatives), the
chloroacetylamino acid derivatives 2a-h needed to be synthesized prior to alkylation of 1
(Scheme 2). They were formed in high yield by adding chloroacetylchloride to a mixture
ACS Paragon Plus Environment
9

Journal of Medicinal Chemistry
Page 10 of 78
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
containing the appropriate hydrochloride salt of amino acid ester, DIEA and toluene as a
2
4
solvent .
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
was then alkylated in Nꢀ9 position in the presence of a base (K CO or NaH) and
2 3
the appropriate halogenoalkylating reagent in dry DMF to give Nꢀ9ꢀalkylated guanine 3a-t
with variable yields (18ꢀ86%) (Scheme 1). During the alkylation with bromoaliphatic
esters in the presence of a large excess of K CO , dialkylation at the Nꢀ9 and Nꢀ2 positions
2
3
could be observed leading to a mixture of both the 9ꢀmonoalkylated (3aꢀt) and the 2,9ꢀ
dialkylated (4aꢀd) compounds. The alkylation at the Nꢀ2 position was confirmed by the
1
disappearance of the NH proton in H NMR. The bisꢀalkylation in Nꢀ2 position was not
observed with the ketomethylene derivatives and was minor with the amino acid
derivatives. Furthermore, it was possible to introduce different substituents in Nꢀ2 and Nꢀ9
positions in two steps. The Nꢀ9 position of 1 was first alkylated and then the Nꢀ2 position
was alkylated using a different halogeno alkylating reagent with the same method.
Concerning the deprotection step, it was possible to remove all the protecting
groups (the diphenylcarmoyl, acetyl and ester groups) using 1 M aq LiOH at 60 °C in high
yield. When the prepared compound did not contain any ester groups, the deprotection
was achieved with 2 M NH in MeOH at 70 °C. In both cases, 9ꢀalkylated and 2,9ꢀ
3
dialkylated guanines 5aꢀy were obtained in high yields without complex purification
methods. Concerning the tyrosine compound 3r, deprotection was performed as usual
with 1M aq LiOH and then the phenol tertꢀbutyl protection was removed using TFA in
CH Cl (1:2 v/v) to give compound 5z with 79% yield.
2
2
ACS Paragon Plus Environment
10

Page 11 of 78
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Docking studies showed us that keeping the acetyl protection could increase the
affinity between the ligand and the active site. A method was thus developed to cleave
selectively the carbamoyl and the ester groups (Scheme 3). The diphenylꢀcarbamoyl
protection could be removed selectively under acidic conditions using TFA in CH Cl (1:9
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
2
2
v/v) to give compounds 14a,b carrying the Nꢀacetyl group. On the other hand, the ester
protection could be removed selectively with a 0.1 M aqueous LiOH at room temperature.
Then, at the last step, an alkyl group (methyl, ethyl, allyl, benzyl) was introduced
at the Nꢀ7 position of guanine derivatives 5 and 14a,b using the appropriate alkylating
agent (dimethyl sulfate, 2ꢀiodoethane, allyl iodide, benzyl bromide) in dry DMF (Schemes
2
5, 26
1, 3, 4)
. During the purification process, the pH was adjusted to around pH 9 with a
few drop of a 20% aqueous ammonia solution. As Nꢀ7 alkylation decrease the pK of the
a
2
7
Nꢀ1 proton to about 7.5 products (6a-e, 7a-h, 8a-j, 9,10, 11, 12, 13, 15a,b, 18a-d) were
isolated after precipitation with acetone into a zwitterionic form containing both charges
7
on the m G moiety. Final products were then further purified by crystallization from
MeOH/H O 1:1 or reversed phase chromatography.
2
The synthesis of the fourth series was achieved in three steps from the 2,6ꢀ
diprotected guanine 1 (Scheme 4). The Nꢀ9 arylation of 1 was performed using conditions
2
8
set up by employing 2 equivalents each of aryl boronic acid, 1,10ꢀphenanthroline and
cupric acetate in DMF. After 3 days stirring at room temperature, the Nꢀ9ꢀarylpurines 16a-
d were produced with moderate yields (≈ 30%). Protecting groups were then removed
using 2M NH in MeOH at 70 °C or 1M aqueous LiOH at 40 °C to give compounds 17a-
3
ACS Paragon Plus Environment
11

Journal of Medicinal Chemistry
Page 12 of 78
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
d. Final compounds 18a-d were obtained by Nꢀ7 methylation using dimethylsulfate in
anhydrous DMF.
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
Crystal structures of inhibitor complexes
Since capꢀbinding domainꢀinhibitor complexes were not readily crystallized using
the original H3N2 human influenza A strain, we switched to the equivalent domain from
the avian strain A/duck/Shantou/4610/2003(H5N1). Coꢀcrystals of the H5N1 domain
7
diffracting to high resolution could be obtained with m GTP and three inhibitors,
compounds 8e, 8f and 11. Data collection and refinement statistics are given in Table 1.
The electron density for the four ligands is shown in Supplementary Figure S1. The
inhibitor bound structures are described in the next section.
The avian H5N1 capꢀbinding domain has ten amino acid changes from the human
H3N2 construct previously used for structure determination (Figure 5). Nevertheless the
structures are very similar (RMS deviation 0.75 Å for 155 aligned Cα), although the
flexible ‘424ꢀloop’ has a slightly different conformation (Figure 5A). The mutations do
not directly affect the capꢀbinding site, which is highly conserved in all influenza A strains
(Figure 5B). Three clear chloride ions from the crystallization medium (2 M NaCl) are
visible in the structure.
PB2 cap binding inhibition
7
The inhibition of binding of a biotinylated cap analogue (BiotinꢀEDAꢀm GpppA)
to a Hisꢀtagged PB2ꢀdomain of Influenza strain H3N2 by small molecules was measured
using the AlphaScreen assay technology. Nonꢀspecific effects of the compounds in the
ACS Paragon Plus Environment
12

Page 13 of 78
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
AlphaScreen were tested with a biotinylated GSTꢀHisꢀtag fusion protein. IC ꢀvalues are
50
indicated in Table 2 and Table 3.
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
17
Although described as an inhibitor of the viral RNP (IC = 45 ꢁM) , the
5
0
previously reported cap binding inhibitor RO0794238 appears to be inactive in our assay
using the isolated PB2 capꢀbinding domain. This is consistent with docking experiments,
7
which show that its Nꢀ7 benzyl group probably sterically prevents binding in the m G
recognition site. However, RO0794238 might interact with another part of the viral RNP.
Furthermore all monoalkylated compounds from the first series (6aꢀe) are also inactive
(
IC > 200 ꢁM).
50
On the other hand, dialkylated compounds 9, 10, 11 and 12 showed potency for
cap binding inhibition. This suggest that introduction of a chain carrying a carboxylate
group in Nꢀ2 position enhances the affinity for the active site. Compound 11 bearing two
carboxybutyl chains in Nꢀ2 and Nꢀ9 positions is the most active (IC = 5.3 ꢁM).
50
Compound 9 carrying two carboxymethyl groups in these positions is also interesting
IC = 44 ꢁM), however compound 12 with a carboxymethyl group in the Nꢀ9 position
(
5
0
and a carboxybutyl chain in the Nꢀ2 position showed less potency (IC = 95 ꢁM) than 9
5
0
and 12. Also dialkylated compounds 9, 10, 11 and 12 are very selective towards PB2.
The Xꢀray structure of 11 bound to the H5N1 PB2 cap binding domain was
determined at 1.9 Å resolution. There are two crystallographically independent ligand
complexes in the asymmetric unit, and in one (that shown in Figure 6), the carboxybutyl
chains are better ordered. This correlates with the better ordering of the Nꢀterminal strand
ACS Paragon Plus Environment
13

Journal of Medicinal Chemistry
Page 14 of 78
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
of the protein (visible from Ile319) to which the carboxybutyl chains are directed. The
presence of the carboxybutyl chain at Nꢀ2 does not interfere with binding of the guanine
7
moiety in a similar position to that of m GTP and the carboxylate group of this chain is
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
able to form a hydrogen bond with Ser321. The carboxybutyl chain at the Nꢀ9 position
partially stacks on Phe323 but has no other interactions with the active site.
Amino acid compounds 7a-g from the second series display moderate potency for
cap binding inhibition. As expected from the docking studies, compounds with two
carboxylate groups are generally more active and both enantiomers display potency. The
most active compound is the LꢀAsp derivative 7b, IC = 33 ꢁM. No conclusions can be
5
0
drawn about the best enantiomers as DꢀPhe derivative 7g (IC = 66 ꢁM) is more active
5
0
than its LꢀPhe enantiomer 7d (IC = 88 ꢁM) while DꢀAsp derivative 7f (IC = 91 ꢁM) is
5
0
50
less active than its enantiomer 7b. As expected from the results of the first series, addition
of a carboxybutyl chain in position Nꢀ2 of the DꢀPhe derivative resulted in slightly better
inhibition (13, IC = 54 ꢁM).
5
0
Evaluation of the third series revealed that introduction of a phenylꢀketomethylene
moiety enhances the activity. Compound 8a carrying an unsubstituted phenylꢀ
ketomethylene group display potent activity (IC = 14 ꢁM), though the naphthyl
5
0
derivative 8d is twofold more potent (IC = 7.6 ꢁM). This shows the importance of the πꢀ
50
stacking and the hydrophobic interactions with the active site. In addition, introduction of
hydrophilic substituents on the phenyl ring in comparison to compound 8a resulted in the
potent inhibitors 8e (IC = 0.6 ꢁM) and 8f (IC = 7.5 ꢁM) bearing a methyl group on the
50
50
Nꢀ7 nitrogen atom.
ACS Paragon Plus Environment
14

Page 15 of 78
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
The presence of a carboxylate group substituent on the phenyl ring in 8e enhances
the activity by a factor 10 in comparison to a carboxyamido substituent in 8f. This
confirms the important role of carboxylate group negative charges in cap binding
inhibition but the selectivity is also reduced. Compound 8f appeared to be very selective
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
(
IC GSTꢀHis biotin ≥ 200 ꢁM) in comparison to 8e ((IC GSTꢀHis biotin= 2.2 ꢁM).
50 50
However, as discussed in the conclusion, the carboxylate may interfere with the
AlphaScreen assay.
High resolution Xꢀray coꢀcrystal structures have been obtained with compounds 8e
(1.5 Å resolution) and 8f (1.85 Å resolution) bound in the H5N1 PB2 cap binding domain
7
(Figures 7ꢀ8). In these structures, the m G moiety is bound as expected and the substituted
phenylꢀketomethylene moiety projects towards the solvent, but with a different orientation
for each compound (~ 105º rotation about the linkage to the Nꢀ9 position) (Figure 7ꢀ8). In
the case of 8e, the plane of the phenylꢀketomethylene is perpendicular to the ring of
Phe323 and the keto oxygen makes a hydrogen bond with His357 (Figure 8B). In the case
of 8f, the phenylꢀketomethylene is at about 45º to Phe323 with the keto oxygen pointing
towards Phe325. Docking studies of 8e (Figure 1B) had suggested that the carboxylate
group of this compound would be in a similar position to that of the last phosphate group
7
7
of m GTP in the H3N2 coꢀcrystal structure (but not the H5N1ꢀm GTP structure) and it is
indeed observed to make a salt bridge with Lys339 (Figure 8B). For 8f there are no strong
interactions between the 3ꢀcarboxyamidoꢀ4ꢀhydroxyphenyl group and the active site.
Introduction of a methyl or a phenyl group on the methylene in the αꢀposition of
the ketone reduces the affinity with PB2 (8b, IC = 65 ꢁM; 8c, IC = 42 ꢁM). Introduction
5
0
50
ACS Paragon Plus Environment
15

Journal of Medicinal Chemistry
Page 16 of 78
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 carboxylate chain in Nꢀ2 position for this series of compounds would have probably
resulted in more potent inhibitors, unfortunately the corresponding synthesis was
unsuccessful.
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 influence of the substituent in Nꢀ7 position was then investigated. Different
substituent (ethyl, allyl, benzyl) were introduced in that position from the unsubstituted
hydroxybenzamide compound 5k (Table 2). Comparison between the unsubstituted
compounds 5k and 5q and the Nꢀ7 methylated compounds 8f and 8e shows that Nꢀ7
methylation enhances the affinity for the active site by a factor 10, but is not essential for
binding. Also the ethyl derivative 8g is about 2 fold more potent than 8f while addition of
a bigger substituent (allyl 8h or benzyl 8i) reduced the activity. So, only small substituent
e.g. methyl, ethyl are tolerated in the Nꢀ7 position.
Compounds of the fourth series displayed moderate potency. They are not as
active as the phenylketomethylene derivatives from the third series but more active than
the aliphatic equivalent from the first series. Both compounds containing carboxylate
groups 18c and 18d are more active than the methoxy derivative 18b and the unsubstituted
compound 18a.
Selected compounds were then tested for their antiviral activity in a cellꢀbased
assay using cell viability as surrogate endpoint for inhibition of virus induced cytopathic
2
9
effect as previously described . Unfortunately none of them displayed any potency under
the assay conditions. This might be due to insufficient affinity and/or the presence of
multiple charges that will not allow compounds to cross the biological membrane.
ACS Paragon Plus Environment
16

Page 17 of 78
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Conclusions
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
The present paper describes the design and synthesis of about forty novel Nꢀ7ꢀ
alkylguanine derivatives classified in four series and their evaluation in PB2 cap binding
inhibition assays. All potential inhibitors were prepared in three or four steps from 2ꢀNꢀ
acetylꢀ6ꢀOꢀdiphenylcarbamoylguanine 1.
First, we focused our approach keeping a methyl group at the Nꢀ7 position as in the
capped mRNA. After docking experiments from our previously described Xꢀray structure
of the PB2 cap binding domain with bound 7ꢀmethylguanosine triphosphate, various
substituents were introduced at the Nꢀ9 position: (i) substituents carrying one or two
carboxylate functions in order to mimic the phosphate charge(s) of the natural substrate
and (ii) hydrophobic or hydrophilic substituents incorporating a small aromatic nucleus
(phenyl, naphthyl) in order to increase the affinity through stacking with aromatic amino
acid residues, the aromatic nucleus being directly attached to the Nꢀ9 nitrogen atom or
attached via a ketomethylene linker in order to produce strong but restrained orientation
changes.
Analogues of one of the most efficient and selective inhibitor identified were also
prepared in which the 7ꢀmethyl group was replaced by a more hindered lipophilic ethyl,
allyl or benzyl group. This latter is found in the previously reported cap binding inhibitor
RO0794238 incorporating at Nꢀ9 position of the guanosine ring a 4ꢀhydroxybutyl chain.
This inhibitor appeared to be inactive in our tests as suggested by docking experiments
ACS Paragon Plus Environment
17

Journal of Medicinal Chemistry
Page 18 of 78
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
showing that the presence of the bulky Nꢀ7 benzyl group avoids binding to the 7MG
binding site.
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
Some 2,7,9ꢀtriakylguanine derivatives were also synthesized for the first time and
evaluated. Among the prepared guanine derivatives, the 7,9ꢀdialkyl derivatives
incorporating a ketomethylene group attached at Nꢀ9 nitrogen atom were the most active,
in particular, compound 8e with an IC = 0.6 ꢁM. However, the selectivity for the targeted
5
0
PB2 unit appeared to be low as shown by the IC = 2.2 ꢁM with the biotinylated Hisꢀ
50
tagged protein used. As discussed below, the lack of selectivity could result from the
presence of a carboxylate function attached to the phenylketomethylene group (Table 1)
since 8f in which the carboxylate function is replaced by the carboxyamido group showed
a higher IC value than 8e under 10 ꢁM with a high selectivity. 8d incorporating a nonꢀ
5
0
substituted 9ꢀnaphthylketomethylene group also displayed similar effects.
5q, analogue of 8e but not alkylated at the Nꢀ7 position, displayed an interesting IC50
value also under 10 ꢁM but with a weak selectivity. This result shows that the 3ꢀcarboxyꢀ
4
ꢀhydroxyphenyl group in 5q and 8e has an affinity for both PB2 and GSTHis biotin. The
3
0
salicylate group can complex nickel(II) and could be binding chelated Ni(II) used in our
AlphaScreen assay. The measured IC of 5q and 8e for GSTHis biotin appeared to be not
5
0
very different whereas 8e showed a much lower IC for PB2 than 5q. This confirms that
50
8e is a good ligand of PB2 as observed with the crystal structure and that it should have a
good selectivity for PB2 in the absence of Ni(II) and transition metal ions.
ACS Paragon Plus Environment
18

Page 19 of 78
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Among the analogues of 8f carrying different Nꢀ7 alkyl groups, only the ethyl
derivative 8g showed a similar behavior with an IC twice lower than the one of 8f. The
50
rigidity and steric hindrance introduced with the allyl (8h) and the benzyl (8i) group,
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
7
respectively probably decreases the affinity for the m G binding site. 8g appeared to be the
7
most interesting m G analogue identified in this study in regard to its lowest IC and high
5
0
selectivity.
Regarding the compounds in which more flexibility of the carboxylate group(s) is
possible, i.e. the 7,9ꢀdialkylated 6aꢀe and 7aꢀg and the 2,7,9ꢀdialkylated 9ꢀ13 carrying two
carboxylate groups, 11 carrying at Nꢀ7 and Nꢀ9 4ꢀcarboxybutyl substituents was the fifth
compound showing an IC value under 10 ꢁM with in addition a high selectivity. Its
50
parent derivative 6d carrying only one Nꢀ9 4ꢀcarboxybutyl substituent was found
completely inactive. Therefore, it is possible to introduce two carboxylate functions
without loss of selectivity. The presence of several negative charges well located as in the
RNA substrate should lead to a better inhibition.
None of the identified inhibitors showed any antiviral activity in infected cells. The
presence of negative charges in the inhibitors could explain their lack of antiviral activity
through a poor cellular uptake. Formulations with cationic lipophilic vectors or ester
prodrugs which should enter in cells more efficiently than the carboxylates could be
prepared for antiviral evaluation.
Three crystal structures of cap binding domainꢀinhibitor complexes obtained with the
7
active compounds 8e, 8f and 11, and confirm a unique binding mode of the m G scaffold
ACS Paragon Plus Environment
19

Journal of Medicinal Chemistry
Page 20 of 78
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
7
in the PB2 recognition site for these compounds and m GTP. However, docking
experiments predict a slightly different orientation of the guanine ring in the binding
pocket. This might be due to the rigidity of the active site used in the docking
experiments. Dynamics experiments should now be performed in order to improve our
understanding of the binding mode for the design of better inhibitors.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
7
In conclusion, the confirmation of the mode of binding of the m G scaffold by Xꢀray
7
crystallography, the low IC values obtained with four prepared m G derivatives (8d, 8f,
50
7
8
g and 11) and their concomitant high selectivity observed suggest that the m G scaffold
can be used to design strong inhibitors of mRNA binding to the PB2 cap binding domain.
ACS Paragon Plus Environment
20

Page 21 of 78
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
ꢀ EXPERIMENTAL SECTION
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
Docking. Ligands were built and geometrically optimized within Sybyl 7.3 (Tripos
International, USA) using Tripos force field and Gasteiger charges. MGL AutoDock Tool
3
1
(ADT) v1.5.4 was used to select rotatable bonds and write AutoDock formatted ligand
files. The coꢀcrystal structure of the PB2 capꢀbinding domain of influenza strain H3N2
7
with m GTP is available from the protein databank (pdb code: 2VQZ). All water
7
molecules and coꢀcrystallized compounds were removed from the protein and the m GTP
coordinates were kept as a reference ligand. Missing hydrogens and Gasteiger charges
were added using ADT and then non polar hydrogens and lone pairs were merged to the
7
corresponding heavy atoms. The docking region was defined as a box centered on m GTP
binding pocket with a grid spacing of 0,375 Å and 60x54x54 points. Docking was carried
1
6
out with Autodock 4.2 . The protein structure was considered as a rigid body while
ligands were fully flexible. For each ligand 100 docking solutions were computed using
the Lamarckian Genetic Algorithm (LGA), a maximal number of energy evaluations set to
7
10 and all other parameters set to default. The cluster with the lowest energy and best
RMSD with the reference guanidinium core was considered for further interaction studies.
Visual interaction analysis and pictures were made using ADT and PYMOL.
Crystallographic analysis.
The
capꢀbinding
domain
(residues
318ꢀ483)
from
the
avian
strain
A/duck/Shantou/4610/2003(H5N1) was expressed from a synthetic gene (Genart). Protein
ACS Paragon Plus Environment
21

Journal of Medicinal Chemistry
Page 22 of 78
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
1
1
expression and purification was as described .For crystallization of compound 8e, a 424ꢀ
loop deletion mutant was used with sequence …420FꢀGSGꢀL428…
Crystallization trials were performed using a Cartesian robot which set up 576 different
conditions. Cap binding protein was concentrated to 15 mg/mL in gel filtration buffer and
mixed with compound at 5mM final concentration in the drops, except compound 8f,
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
7
which was added directly as powder for crystallization. m GTP coꢀ crystals were grown in
0.1M sodium acetate pH 4.6, 2M NaCl. Compound 8e coꢀcrystals were grown in 0.2M
ammonium nitrate, pH 6.3, 20% PEG 3350. Compound 8f coꢀcrystals were grown in 0.1M
sodium acetate pH 4.2ꢀ4.4, 2.4M NaCl. Compound 11 coꢀ crystals were grown in 0.2 M
ammonium iodide, pH 6.3, 20% PEG 3350.
Crystals, cryoꢀprotected with 20ꢀ25% glycerol and flash frozen, were diffracted on
various beamlines at the European Synchrotron Radiation Facility (ESRF) (Table 1). Data
32
were integrated with XDS , scaled with XSCALE and further analysed with CCP4i. The
structure of the H5N1 capꢀbinding domain was solved by molecular replacement using
PHASER with search model the H3N2 domain (PDB entry 2VQZ). Structures were built
3
3
34
with COOT and refined with REFMAC5 .
AlphaScreen assay. All AlphaScreen experiments were performed in white 384ꢀwell
ProxiPlates (PerkinElmer, Boston, MA) using 25 mM Hepes pH 6.8, 100 mM NaCl, 1
mM DTT, 0.05% Tween and 2% DMSO as buffer. Compounds were 11ꢀfold 1:1ꢀserially
diluted starting from 200 ꢂM as highest concentration. Binding of the biotinylated capꢀ
7
analogue EDAꢀm GpppA (a gift from Edward Darzynkiewicz, Warsaw, Poland) to His ꢀ
6
tagged PB2ꢀ protein was detected with the 6ꢀhistidine kit (PerkinElmer) containing Nickel
ACS Paragon Plus Environment
22

Page 23 of 78
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Chelate coated acceptor beads and streptavidin coated donor beads. Equal final
concentrations of 7.5 ꢂg/ml acceptor and donor beads were always used. Nonꢀspecific
inhibition of the AlphaScreen by the compounds was tested in a parallel experiment with a
biotinylated GSTꢀHis fusion protein. This biotinylated GSTꢀHis fusion protein contains all
tags necessary for the interaction with the AlphaScreen detection beads within one
molecule and serves as control for inhibitors of the AlphaScreen detection technology
itself. All incubation steps with AlphaScreen beads were performed at room temperature
under subdued lighting conditions before the assay plates were read in an EnVision plate
reader (PerkinElmer).
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
Surface plasmon resonance measurements (SPR). SPR measurements of ligand binding
were performed on a Biacore X100 system equipped with CM7 sensor chips (GE
Healthcare). H5N1 PB2ꢀCBD was immobilized on the sensor surface by amine coupling
according to the manufacturer’s protocol using a protein concentration of 30 ꢂg mlꢀ1 and
5
mM m7GTP (Sigma Aldrich) in 10mM phosphate buffer pH 6.5 and HBSꢀEP buffer as
running buffer (10 mM HEPES, 150 mM NaCl, 3 mM EDTA, 0,005% surfactant p20). A
ꢀ1
contact time of 12 min and a flow rate of 10 ꢂL min results in an immobilization level of
approximately 7000 RU.
For compound screening a running buffer containing 10 mM TRIS, 3mM EDTA,
150 mM NaCl, 0.005% surfactant p20 (GE Healthcare/Biacore), 1 mM DTT, 0.5%
DMSO was used. DMSO stock solutions of each compound were diluted in sample buffer
without DMSO to a final compound concentration of 10 ꢂM and 0.5% DMSO and
injected over the immobilized PB2ꢀCBD and references flow cell surfaces at a flow rate of
ACS Paragon Plus Environment
23

Journal of Medicinal Chemistry
Page 24 of 78
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
1
0 ꢂL minꢀ1. Sensorgrams were processed using double referencing and solvent
correction for DMSO bulk effects. Affinity constants (K values) were determined by
D
measuring the binding affinity of the analyte to the ligand over a concentration range
ranging from 200 ꢂM to 1 nM and was calculated using a linear curve fit model of
Biacore X100 Evaluation Software.
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
Antiviral screening by CPE reduction. Performed as described .
General information concerning chemical synthesis. Thin layer chromatographies were
performed on silica gel (Alugram Sil G/UV ) from MachereyꢀNagel and spots were
2
54
detected either by UVꢀabsorption or by charring with ninhydrin. Flash columns
chromatographies were performed on silica gel (Acros Kieselgel 0.035ꢀ0.070 mm, 60 Å)
or on C reversed phase (MachereyꢀNagel polygoprep 60ꢀ50 C ). All starting materials
18
18
were obtained from suppliers and used without further purification unless otherwise noted.
1
13
H NMR (400 MHz) and C NMR (100 MHz) spectra were recorded with a BRUKER
ADVANCE 400 spectrometer using the residual solvent signal as internal standard.
HRMS were obtained from the Mass Spectrometry Service, CRMPO, at the University of
Rennes I, France, using a MICROMASS ZABSPECꢀTOF spectrometer and a VARIAN
MAT311 spectrometer. Melting points were determined with a BUCHI 510 apparatus.
Compound purity of all evaluated derivatives was confirmed to be ≥ 95% by HPLC using
a C18 reversed phase column eluting with gradients of water and methanol in the presence
of 0.3% trifluoroacetic acid (detection at 254 nm).
ACS Paragon Plus Environment
24

Page 25 of 78
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Experimental procedures for the synthesis and characterization of the prepared
compounds
General procedure for the preparation of chloroacetyl amino acids 2a-h. The
appropriate protected amino acid hydrochloride salt or dimethyl 5ꢀaminoisophthalate was
added to a solution of DIEA (3 equiv) and toluene (20 mL) that was cooled in an
acetone/liquid nitrogen bath to below ꢀ10 °C. A solution of chloroacetyl chloride (1.1
equiv) in toluene (3 mL) was added drop wise over 20 min maintaining the temperature
below 0 °C. The reaction was allowed to warm to room temperature and stirred for 2 h
under argon. The toluene was removed under reduced pressure and the remaining brown
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
liquid was taken into AcOEt and washed several times with H O and then once with 0.1
2
M HCl. The organic layer was dried (MgSO ) and the solvent removed under reduced
4
pressure to give the product without further purification.
General
procedure
for
the
synthesis
and
of
2-N-acetyl-9-N’-alkyl-6-O-
diphenylcarbamoylguanine
(3a-t)
2-N-acetyl-2,9-N,N’-dialkyl-6-O-
diphenylcarbamoyl-guanine (4a-g). Method A: A mixture of 1 (1 equiv), K CO (3
2
3
equiv) and the appropriate halogeno reagent (1.2 equiv) in anhydrous DMF (30 mL) was
stirred at room temperature for 1ꢀ2 days. The solvent was then evaporated and the
resulting mixture was partitioned between EtOAc (100 mL) and brine/H O (2:1 v/v) (75
2
mL). The organic layer was separated and washed with brine (2 × 50 mL), dried (MgSO4),
filtered and concentrated to dryness. The residue was then purified as indicated for each
compound.
ACS Paragon Plus Environment
25

Journal of Medicinal Chemistry
Page 26 of 78
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
Method B: To an ice cold solution of 1 (1 equiv) in anhydrous DMF (30 mL) was added
NaH (60% in mineral oil) (1.1 equiv). After 30 min, the appropriate halogen derivative
(
1.2 equiv) was added and the mixture was left to stir at rt overnight under argon. The
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
solvent was then evaporated under reduced pressure and the resulting mixture was
partitioned between EtOAc (50 mL) and H O (50 mL). The organic layer was separated
2
and washed with brine (2 × 50 mL) and dried over MgSO . The residue was then purified
4
as indicated for each compound.
General procedure for the deprotection of compounds 3a-t, 4a-g and 16a-d. Method
C: 2M NH in MeOH (20 mL) was added under argon to the 2,6ꢀdiprotected guanine. The
3
mixture was then left to stir for 3ꢀ5 h at 70 °C. The solvent was then removed under
reduced pressure and the white residue obtained was suspended in CH Cl (20 mL) and
2
2
stirred for 20 min. Then it was filtered and washed several times with CH Cl (3 × 5 mL)
2
2
to obtain the pure product without further purifications.
Method D: 1M aq LiOH (10 mL) was added to a mixture of the 2,6ꢀdiprotected guanine
in THF (15 mL) and it was stirred at 60 °C for 2ꢀ4 h. The solvent was then evaporated and
the mixture was poured in a solution of EtOAc (10 mL) and H O (10 mL). The aqueous
2
layer was washed with EtOAc (2 × 10 mL) and then acidified with 1M aq HCl until
formation of a precipitate (pH 2ꢀ3). The white solid formed was then filtrated and washed
with H O (20 mL).
2
Method E: 1M aq LiOH (16 mL) was added to a mixture of 3r (1.6 g, 2.36 mmol) in
THF (12 mL) and it was stirred at 40 °C for 4 h. The solvent was then evaporated and the
ACS Paragon Plus Environment
26

Page 27 of 78
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
mixture was poured in a solution of EtOAc (15 mL) and H O (10 mL). The aqueous layer
2
was washed with EtOAc (2 × 10 mL) and then acidified with a 1M aq HCl until formation
of a precipitate (pH 2ꢀ3). The white solid formed was then filtrated and washed with H O
2
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
(20 mL). The solid was then collected and CH Cl (10 mL) and triethylsilane (1.8 mL,
2
2
1
1.8 mmol) were added. The mixture was cooled to 0 °C and TFA (5 mL) was added
dropwise over a period of 5 min. after 30 min the reaction was allowed to warm to rt.
After 3 h stirring, the mixture was concentrated in vacuo to give a foam which was then
coevaporated with Et O (10 mL) and toluene (3 × 10 mL). The residue was then stirred in
2
CH Cl (20 mL) for 20 min, filtered and washed with CH Cl (3 × 5 mL) to give a white
2
2
2
2
solid (0.70 g, 79%).
1
9
-N-carboxymethylguanine 5a. Method D. White solid (82%). H NMR: δ (DMSOꢀd ):
6
1
0.70 (bs, 1H, NH), 7.77 (s, 1H, 8ꢀCH), 6.54 (bs, 2H, NH ), 4.79 (s, 2H, CH ).
2 2
1
2,9-N,N’-dicarboxymethylguanine 5b. Method D. White solid (54%). H NMR: δ
(
DMSOꢀd ): 10.92 (bs, 1H, NH), 7.68 (s, 1H, 8ꢀCH), 6.79 (bs, 1H, NH), 4.76 (s, 2H,
6
NCH ), 3.99 (s, 2H, NHCH ).
2
2
1
9
-N-2’-carboxyethylguanine 5c. Method D. White solid (53%). H NMR: δ (DMSOꢀ
d ): 11.38 (bs, 1H, NH), 7.62 (s, 1H, 8ꢀCH), 6.88 (bs, 2H, NH ), 4.06 (t, 2H, CH ), 2.50 (t,
6
2
2
2
H, CH2).
ACS Paragon Plus Environment
27

Journal of Medicinal Chemistry
Page 28 of 78
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
1
9
-N-3’-carboxypropylguanine 5d. Method D. White solid (88%). H NMR: δ (DMSOꢀ
d ): 10.58 (bs, 1H, NH), 7.63 (s, 1H, 8ꢀCH), 6.52 (bs, 2H, NH ), 3.95 (t, 2H, CH ), 2.24 (t,
6
2
2
2
H, CH ), 1.93 (m, 2H, CH ).
2
2
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
,9-N,N’-di-(3’-carboxypropyl)guanine 5e. Method D. White solid (42%). H NMR: δ
(
DMSOꢀd ): 10.50 (bs, 1H, NH), 7.70 (s, 1H, 8ꢀCH), 6.52 (bs, 1H, NH), 3.98 (t, 2H, CH ),
6
2
3
.31 (t, 2H, CH ), 2.36 (t, 2H, CH ), 2.28 (t, 2H, CH ), 1.97 (m, 2H, CH ), 1.77 (m, 2H,
2 2 2 2
1
3
CH ). C NMR: δ (DMSOꢀd ): 174.2 (C, COOH), 173.7 (C, COOH), 156.8 (C, Cꢀ6),
2
6
1
52.5 (C, Cꢀ2), 150.9 (C, Cꢀ4), 137.4 (CH, Cꢀ8), 116.4 (C, Cꢀ5), 41.9 (CH , NCH ), 40.8
2 2
(
CH , NCH ), 31.0 (CH , CH COOH), 30.6 (CH , CH COOH), 24.9 (CH ), 24.3 (CH ).
2 2 2 2 2 2 2 2
1
9
-N-4’-carboxybutylguanine 5f. Method D. White solid (56%). H NMR: δ (DMSOꢀ
d ): 10.53 (bs, 1H, NH), 7.71 (s, 1H, 8ꢀCH), 6.47 (bs, 2H, NH ), 3.95 (t, 2H, CH ), 2.24 (t,
6
2
2
2
H, CH ), 1.73 (m, 2H, CH ), 1.42 (m, 2H, CH ).
2
2
2
1
2
,9-N,N’-di(4’-carboxybutyl)guanine 5g. Method D. White solid (49%). H NMR: δ
(
DMSOꢀd ): 10.45 (bs, 1H, NH), 7.69 (s, 1H, 8ꢀCH), 6.42 (bs, 1H, NH), 3.98 (t, 2H, CH ),
6
2
3
.30 (t, 2H, CH ), 2.28 (m, 4H, 2 × CH ), 1.74 (m, 2H, CH ), 1.54 (m, 4H, 2 × CH ), 1.40
2 2 2 2
1
3
(
m, 2H, CH ). C NMR: δ (DMSOꢀd ): 174.4 (C, COOH), 174.2 (C, COOH), 156.8 (C,
2 6
Cꢀ6), 152.4 (C, Cꢀ2), 150.9 (C, Cꢀ4), 137.4 (CH, Cꢀ8), 116.4 (C, Cꢀ5), 42.1 (CH , NCH ),
2
2
4
0.1 (CH , NCH ), 33.3 (CH , CH COOH), 32.9 (CH , CH COOH), 28.8 (CH ), 28.3
2 2 2 2 2 2 2
(CH ), 21.8 (CH ), 21.4 (CH ).
2
2
2
ACS Paragon Plus Environment
28

Page 29 of 78
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
9
-N-5’-carboxypentylguanine 5h. Method D. White solid (58%). H NMR: δ (DMSOꢀ
d ): 10.74 (bs, 1H, NH), 7.68 (s, 1H, 8ꢀCH), 6.53 (bs, 2H, NH ), 3.95 (t, 2H, CH ), 2.19 (t,
6
2
2
2
H, CH ), 1.70 (m, 2H, CH ), 1.51 (m, 2H, CH ), 1.25 (m, 2H, CH ).
2 2 2 2
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
2
-N-(4’-carboxybutyl)-9-N’-carboxymethylguanine 5i. Method D. White solid (59%).
1
H NMR: δ (DMSOꢀd ): 10.51 (bs, 1H, NH), 7.66 (s, 1H, 8ꢀCH), 6.48 (bs, 1H, NH), 4.76
6
(
s, 2H, CH ), 3.27 (t, 2H, CH ), 2.26 (t, 2H, CH ), 1.52 (m, 4H, 2 × CH ).
2
2
2
2
9-N-4’-carboxybutyl-2-N’-carboxymethylguanine 5j. Method D. White solid (91%).
1
H NMR: δ (DMSOꢀd ): 10.84 (bs, 1H, NH), 7.72 (s, 1H, 8ꢀCH), 6.67 (bs, 1H, NH), 3.99
6
(
m, 4H, 2 × CH ), 2.23 (t, 2H, CH ), 1.77 (m, 2H, CH ), 1.44 (m, 2H, CH ).
2
2
2
2
9-N-(3’-carboxamido-4’-hydroxyphenyl)ketomethylguanine 5k. Method C. Bright
1
yellow solid (97%). H NMR: δ (DMSOꢀd ): 9.98 (bs, 1H, NH), 8.52 (s, 1H, Ar), 7.77 (d,
6
1
H, Ar), 7.59 (s, 1H, 8ꢀCH), 7.29 (bs, 1H, OH), 6.58 (d, 1H, Ar), 6.45 (bs, 2H, NH ), 5.46
2
1
3
(
s, 2H, CH ). C NMR: δ (DMSOꢀd ): 189.2 (C, CO), 170.1 (C, CONH ), 156.9 (C, Cꢀ6),
2 6 2
1
53.6 (C, Cꢀ2), 151.7 (C, Cꢀ4), 138.5 (CH, Cꢀ8), 132.5 (C, Ar), 131.2 (CH, Ar), 120.4
(
CH, Ar), 116.3 (C, Cꢀ5), 116.1 (CH, Ar), 48.4 (CH2).
1
9
-N-2’-naphthylketomethylguanine 5l. Method C. Yellow solid (58%). H NMR: δ
(DMSOꢀd ): 8.95 (bs, 1H, NH), 8.24ꢀ8.01 (m, 5H, Ar), 7.74 (m, 2H, Ar), 7.59 (s, 1H, 8ꢀ
6
13
CH), 7.03 (bs, 2H, NH ), 5.75 (s, 2H, CH ). C NMR: δ (DMSOꢀd ): 192.9 (C, CO),
2
2
6
1
56.4 (C, Cꢀ6), 152.1 (C, Cꢀ2), 135.4 (CH, Ar), 132.1 (C, Ar), 131.6 (C, Ar), 129.7 (CH,
Ar), 129.3 (CH, Ar), 128.7 (CH, Ar), 128.2 (CH, Ar), 127.7 (CH, Ar), 126.7 (CH, Ar),
23.2 (CH, Ar), 116.4 (C, Cꢀ5), 49.1 (CH2).
1
ACS Paragon Plus Environment
29