Discovery of a Small-Molecule Inhibitor of Interleukin 15: Pharmacophore-Based Virtual Screening and Hit Optimization

Article abstract of DOI:10.1021/acs.jmedchem.7b00485

Interleukin (IL)-15 is a pleiotropic cytokine, which is structurally close to IL-2 and shares with it the IL-2 β and γ receptor (R) subunits. By promoting the activation and proliferation of NK, NK-T, and CD8+ T cells, IL-15 plays important roles in innate and adaptative immunity. Moreover, the association of high levels of IL-15 expression with inflammatory and autoimmune diseases has led to the development of various antagonistic approaches targeting IL-15. This study is an original approach aimed at discovering small-molecule inhibitors impeding IL-15/IL-15R interaction. A pharmacophore and docking-based virtual screening of compound libraries led to the selection of 240 high-scoring compounds, 36 of which were found to bind IL-15, to inhibit the binding of IL-15 to the IL-2Rβ chain or the proliferation of IL-15-dependent cells or both. One of them was selected as a hit and optimized by a structure-activity relationship approach, leading to the first small-molecule IL-15 inhibitor with sub-micromolar activity.

Full text of DOI:10.1021/acs.jmedchem.7b00485

Subscriber access provided by NEW YORK UNIV
Article
Discovery of a Small-Molecule Inhibitor of Interleukin 15:
Pharmacophore-Based Virtual Screening and Hit Optimization
Agnès Quéméner, Mike Maillasson, Laurence Arzel, Benoit Sicard, Romy Vomiandry, Erwan
Mortier, Didier Dubreuil, Yannick Jacques, Jacques Lebreton, and Monique Mathé-Allainmat
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.7b00485 • Publication Date (Web): 28 Jun 2017
Downloaded from http://pubs.acs.org on June 29, 2017
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of Medicinal Chemistry is published by the American Chemical Society. 1155
Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 92
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Discovery of a SmallꢀMolecule Inhibitor of Interleukin 15:
PharmacophoreꢀBased Virtual Screening and Hit Optimization
†
†
‡
‡
† ‡
Agnès Quéméner,* Mike Maillasson, Laurence Arzel, Benoit Sicard, Romy Vomiandry,
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
†
‡
†
‡
Erwan Mortier, Didier Dubreuil, Yannick Jacques, Jacques Lebreton, Monique Mathéꢀ
‡
Allainmat*
(
†)
CRCINA, INSERM, CNRS, University of Nantes, Nantes 44007, France.
CEISAM, CNRS, University of Nantes, Faculty of Sciences, Nantes 44322, France.
(‡)
1
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 92
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
Interleukin (IL)ꢀ15 is a pleiotropic cytokine, which is structurally close to ILꢀ2 and shares
with it the ILꢀ2 β and γ receptor (R) subunits. By promoting the activation and proliferation of
NK, NKꢀT and CD8+ T cells, ILꢀ15 plays important roles in innate and adaptative immunity.
Moreover, the association of high levels of ILꢀ15 expression with inflammatory and autoꢀ
immune diseases has led to the development of various antagonistic approaches targeting ILꢀ
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
15. This study is an original approach aimed at discovering smallꢀmolecule inhibitors
impeding ILꢀ15/ILꢀ15R interaction. A pharmacophore and dockingꢀbased virtual screening of
compound libraries led to the selection of 240 highꢀscoring compounds, 36 of which were
found to bind ILꢀ15, to inhibit the binding of ILꢀ15 to the ILꢀ2Rβ chain and/or the
proliferation of ILꢀ15ꢀdependent cells. One of them was selected as a hit, optimized by a
structureꢀactivity relationship approach, leading to the first smallꢀmolecule ILꢀ15 inhibitor
with subꢀmicromolar activity.
Keywords: Cytokine, proteinꢀprotein interaction, inhibitor, interleukinꢀ15, inflammatory and
autoimmune diseases
2
ACS Paragon Plus Environment

Page 3 of 92
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1. INTRODUCTION
First identified as a T cell growth factor, Interleukin (IL)ꢀ15 is a 14ꢀ15 kDa cytokine
belonging to the ILꢀ2 cytokine family and involved in the differentiation and proliferation of
1
, 2
NK and T cells.
ILꢀ15 binds a trimeric ILꢀ15 receptor (ILꢀ15R) complex formed by a
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
specific receptor chain (ILꢀ15Rα) combined with the β and γ chains of the ILꢀ2R, acting as
signal transducing components. Binding of ILꢀ15 to ILꢀ2Rβ/γ induces Jak1 and Jak3
activation and subsequently phosphorylation of Stat3 and Stat5 (where Jak and Stat stand for
Janus kinase and signal transducer and activators of transcription, respectively). High levels
of ILꢀ15 expression have been associated with the pathogenesis of autoimmune and
3
4
5
inflammatory diseases like Crohn’s disease, psoriasis, leukemias, rheumatoid arthritis
6
7, 8
(
RA) and graft rejection, resulting in the ILꢀ15 system being a target of interest for the
treatment of these diseases. Several biological approaches have validated this concept in vivo,
including the use of soluble forms of ILꢀ15Rα, antagonist ILꢀ15 mutants, ILꢀ15ꢀ or ILꢀ15Rꢀ
directed monoclonal antibodies and fusion proteins that have proved effective in reducing RA
9ꢀ11
in animal models.
However, their therapeutic benefit in human diseases has not yet been
demonstrated and further studies and clinical trials are required in human systems to evaluate
1
2
their safety and efficacy. In this context, the design of smallꢀmolecule inhibitors of proteinꢀ
13, 14
protein interactions (PPIs) is an emerging and challenging research area
envisaged as an alternative approach to target the ILꢀ15/ILꢀ15R interfaces.
In fact, potent smallꢀmolecule inhibitors of the ILꢀ2/ILꢀ2Rα interface have already been
which could be
15
16
identified using fragmentꢀbased approaches or library screening combined with chemistry
optimization. Moreover, the targeting of the ILꢀ2Rβ chain by siRNA has been shown to be
17
effective in reducing disease severity in adjuvantꢀinduced arthritis in rats, highlighting the
interest of ILꢀ2Rβ as a therapeutic target for the treatment of RA. To date, no smallꢀmolecule
inhibitor of the ILꢀ15 system has been reported, nor has the targeting of the ILꢀ2Rβ interface
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 92
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
been explored by a PPI approach. This work is the first attempt to discover such low MW
nonꢀpeptidic inhibitors, by targeting the ILꢀ15/ILꢀ2Rβ interface. For this purpose, a
18, 19
pharmacophore model was built (Figure 1) based on ILꢀ15 specific residues
involved in
this interface, and was used to screen virtually commercial and academic chemical databases
(ca. 160,000 compounds). The resulting filtered libraries were subsequently docked to the
binding site of ILꢀ15 with ILꢀ2Rβ, and the bestꢀdocked compounds were retained. On the
basis of experimental in vitro properties such as the capacity to bind to ILꢀ15, to reduce the
binding of ILꢀ15 to ILꢀ2Rβ or to inhibit the proliferation of ILꢀ15ꢀdependent cells, one hit
was chosen among 240 purchased compounds. After chemical optimization of its structure,
this led to the development of a low MW inhibitor of the ILꢀ15/ILꢀ15R system with subꢀ
micromolar efficiency.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2. RESULTS AND DISCUSSION
2.1
Pharmacophoreꢀbased and dockingꢀbased virtual screening. In order to disrupt ILꢀ
15/ILꢀ15R recognition and impede ILꢀ15 signaling, our approach was to use small MW
molecules to target one of the ILꢀ15 domains that are interacting with the signal transducing
receptor chains ILꢀ2Rβ or ILꢀ2Rγ. The ILꢀ15/ILꢀ2Rγ interface was set aside to avoid
specificity problems due to the sharing of this receptor chain by numerous other cytokine
20
systems (ILꢀ15, ILꢀ2, ILꢀ4, ILꢀ7, ILꢀ9 and ILꢀ21). Thus, we chose to target the ILꢀ15/ILꢀ
2Rβ binding domain, a strategy that has not yet been described in the literature.
Our virtual screening strategy consisted of using pharmacophore searching and docking
21
methods. The 3D structure of ILꢀ15 (PDB code 2Z3Q) was used to generate a
pharmacophore model based on three residues Asp8, Asn65 and Leu69, previously shown by
1
8, 19
22, 23
directedꢀmutagenesis studies
or by homology to the ILꢀ2/ILꢀ2Rβ complex,
to be
involved in the binding of ILꢀ15 to ILꢀ2Rβ. Comparison of ILꢀ15 and ILꢀ2 structures after
4
ACS Paragon Plus Environment

Page 5 of 92
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
superimposition of ILꢀ15 onto ILꢀ2 in the tetrameric complex structure ILꢀ2:ILꢀ2Rα/β/γ (PDB
24
25
code 2B5I) (available at the time of this work unlike that of the ILꢀ15 tetrameric complex) ,
led us to consider (a) an additional orientation for the side chain of Asn65, similar to that of
its homologous Asn88 in ILꢀ2 (Figure 1A), and (b) the potential participation of two water
molecules close to residues Asp8 and Asn65 (Figure 1B). The final model consisted of three
main features (one projection of hydrogen bond donor (HBD) and one projection of hydrogen
bond donor or acceptor (HBDA) centered on Asp8, and one hydrophobic in front of Ile68 and
Leu69) completed by ten exclusion spheres centered on the main residues of the defined
binding site. In addition, three accessory features (one projection of HBD centered on Asn65,
considering two side chain orientations, and finally two projections of HBDA placed on both
water molecules) were created (Figure 1C). The pharmacophore model was used to search a
database of 160828 chemical compounds derived from commercial (Chembridge, Maybridge
and Preswick) or academic libraries (National Cancer Institute (NCI) and the French national
chemical library Chimiothèque Nationale). To be retained, a hit had to match at least the main
features whereas the presence of the accessory features was not mandatory. A total of 24115
molecules were returned by the pharmacophore searching and were subjected to dockingꢀ
based virtual screening using the LigandFit program implemented under Discovery Studio
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
(DS) 3.0 (Accelrys Software Inc, San Diego, CA). More than 30% of the filtered molecules
passed the docking stage and were ranked using 5 scoring functions. A consensus strategy
was adopted (see Experimental Section) and the 240 topꢀranked molecules were purchased
(Table S1). They were first tested at a concentration of 100 µM for their ability to bind to ILꢀ
15 using the surface plasmon resonance (SPR) technology (Figure 2A) or to inhibit the
binding of ILꢀ15 to soluble ILꢀ2Rβ in a Homogeneous Time Resolved Fluorescence (HTRF)
assay (Figure 2C). Thirtyꢀsix compounds were found to bind to ILꢀ15 as well as to our ILꢀ15
26
superagonist RLI (a fusion protein linking ILꢀ15 to the ILꢀ15Rα sushi domain) (Figure 2A
5
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 92
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
and B), thus discarding any binding of compounds to ILꢀ15 at its interface with ILꢀ15Rα.
Some were also shown to reduce the FRET signal induced by the interaction between
recombinant ILꢀ2Rβ and ILꢀ15 (Figure 2D). To evaluate their impact on a specific ILꢀ15
functional effect, we further tested these hits for their ability to inhibit ILꢀ15ꢀdependent cell
proliferation. Most of them were inhibitors and five of them showed IC50 values around or
below 10 µM (Figure S2). One of these, named compound 1 (Figure 3) was selected for
chemical optimization. Its binding mode to the ILꢀ15 structure was predicted by a molecular
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
27
docking technique using LigandFit and CꢀDocker programs. The interactions involved, as
predicted by the last fine docking step were πꢀalkyl hydrophobic interactions with residues
Ile68 and Leu69, an amideꢀπ stacked interaction with the Ser7 and Asp8 peptidic bond, a
hydrogen bond (HB) with Asp8 and a carbon HB with Asn65 (Figure 4A).
2.2
Synthesis. Among the identified hits, the chemical structure of compound 1 emerged
as a good platform to initiate synthetic modifications to improve its biological activity. The
chemical structure of this compound can be regarded as the association of three fragments: a
dihydroꢀphthalazinone tethered to a triazole moiety, which can be considered the main core of
the molecule, a terminal aromatic group and an alkyl linker chain to join both parts of the
molecule through SꢀC bond and amide bond formation (Figure 3). To the best of our
knowledge, the synthesis of compound 1 has never been described in the literature, and we
began our synthetic work to define the best strategy to access this hit. A good and
combinatorial approach consisted of condensing anilines with a carboxylic acid derived from
the Sꢀalkylation of the main dihydroꢀphthalazinone core, with bromoacetic acid (Figure 3,
route 1). Although this carboxylic acid derivative could easily be prepared with a 60% yield
28
in Williamsonꢀtype alkylation conditions, all attempts to access to the amide 1 in common
29, 30
coupling conditions with carbodiimide reagents
and 2,4ꢀdichloroaniline failed. We thus
6
ACS Paragon Plus Environment

Page 7 of 92
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
turned to a second strategy, exploring the Sꢀalkylation of the dihydroꢀphthalazinone main core
with a variety of halogenoalkylamides previously prepared (Figure 3, route 2).
The key dihydroꢀphthalazinone intermediates 4 and 5 were prepared from conjugated ester 2ꢀ
E
obtained
by
Wittig
olefination
of
phthalic
anhydride
with
the
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
1
(
carbethoxymethylidene)triphenylphosphorane. Successive additions of methylhydrazine
32
and hydrazine in a oneꢀpot twoꢀstep manner afforded the phthalazinone hydrazide 3 in good
3
3
yield. This latter intermediate was condensed with methyl isothiocyanate to give the
mercaptoꢀtriazole main intermediate 4 or with CS in the presence of KOH to afford
2
3
4
oxadiazole 5 (Scheme 1).
Next, halogenoalkylamide linkers 6ꢀ24 and 27ꢀ38 were obtained from commercially available
bromoalkyl carboxylic acids activated as their acid chloride and coupled to the selected
amines, mainly substituted anilines, but also hydrophobic amines such as 1ꢀadamantylamine
or 1ꢀnaphthylamine (Scheme 2). To substitute hydrophobic halogenes at the terminal aromatic
part of the molecule with hydrophilic substituents, dimethoxy compound 24 was subjected to
a demethylation step in the presence of BBr
Scheme 3). Because all attempts to prepare the Nꢀmethylamide analog 47 later from the hit 1
failed, we had to introduce the Nꢀmethyl group earlier by methylation of the alkylating agent
with an excess of methyl iodide. This alkylation step was also reluctant and iodo amide 39
was isolated in poor yield (Scheme 2). However, it should be noted that the amide compound
3 bearing a longer alkyl linker chain than 1, could easily be Nꢀmethylated in mild conditions
NaH, MeI, DMF, 0 °C) to give derivative 63 in 74 % yield.
3
to give the phenolic derivatives 25 and 26
(
6
4
(
In a final step, hit 1 and its analogs 41ꢀ54, 56, 64, 66ꢀ78 and 86ꢀ90 were obtained by Sꢀ
alkylation of the phathalazinone intermediate 4 with the appropriate bromoalkylamide
reagents (Scheme 4 and Schemes 5a and 5b). In some cases, a small amount of Nꢀalkylation
of the phthalazinone intermediate 4 could be observed due to the tautomeric properties of the
7
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 92
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
3
5
3ꢀmercaptoꢀ1,2,4ꢀtriazole ring. In similar conditions, oxadiazole analogs 85 and 91 (Scheme
5
c) were prepared starting from the thiol 5.
Variations at the terminal arylamide part of some compounds were achieved starting from aryl
ether or benzoate compounds. Following demethylation of compounds 51 and 74 in 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
3
presence of BBr , phenol intermediates 57 and 79 could react with suitable alkylating agents
to give the corresponding phenolic ether analogs 58ꢀ60 and 80ꢀ82 in good yields (Scheme 6).
On the other hand benzoic acid derivatives 61ꢀ62 and 83 were obtained from hydrolysis of the
corresponding esters 52ꢀ53 and 75 with LiOH (Scheme 7).
Because the 1,2,3ꢀtriazole ring system has received considerable interest in medicinal
chemistry due to its versatile potential to interact with biological systems through dipolar
3
6
interactions and hydrogen bonding, we opted for the formation of this heterocycle to replace
the terminal amide bond function or to be introduced into a long alkyl chain linker. For this
purpose, we prepared the azide 40 (Scheme 2) and the alkyne 55, which were involved in a
3
7
Huisgen 1,3ꢀdipolar cycloaddition under copper catalysis and gave access to compound 84
with 60% yield. Starting from commercially available 1ꢀazidoadamantane, the intermediate
5
5 reacted in the same conditions to form compound 56, the 1,2,3ꢀtriazole analog of Nꢀ1ꢀ
adamantyl amide 54 (Scheme 8).
2.3
Structureꢀactivity analyses. To date, to the best of our knowledge, no ILꢀ15 selective
inhibitor with low MW has been described in the literature. This work therefore opens the
way to the first in vitro SAR study to provide crucial information on the key binding
interactions of ILꢀ15 with its ILꢀ2Rβ binding site. Each fragment of compound 1 could be
subjected to pertinent modifications but, among the available options, we focused this
preliminary work on the study of the functional impact of the length of the alkyl linker and of
the structure of the terminal amidic group. The effectiveness of the novel compounds was
8
ACS Paragon Plus Environment

Page 9 of 92
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
evaluated using the ILꢀ15ꢀdependent cell proliferation assay and, for the best of them, by SPR
technology and the Stat5 phosphorylation assay.
The binding properties of the compounds were first monitored by SPR. Given that the binding
domains of ILꢀ15 and ILꢀ2 with ILꢀ2Rβ show high similarity, we chose to study compounds
activity systematically on both cytokine systems to provide the opportunity to detect potential
selective compounds rapidly. In the SPR assay, the compounds were flowed over a chip’s
surface, on which RLI or ILꢀ2 had previously been coupled. Binding of compounds induced a
change in mass on the chip, which was detected by the change in reflectance units (RU).
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
d
Direct interaction of compound 1 with immobilized RLI or ILꢀ2 gave similar K values in the
ten micromolar range (48 and 10 ꢀM, respectively). In the case of RLI, consisting of ILꢀ15
bound to ILꢀ15Rα, the binding of compound 1 could not have occured on the ILꢀ15 binding
domain to ILꢀ15Rα because the latter was no longer accessible.
For further classic SAR exploration, we preferred to perform a functional and cellular assay
with the aim of identifying a potential drug candidate. For this purpose, we first tested our hit
1
analogs, compounds 41ꢀ54 and 56ꢀ91, for their potency in inhibiting cell proliferation
induced by a fixed concentration (close to their respective EC50 values) of RLI (100 pM) or
ILꢀ2 (1.5 nM), on 32Dβ cells that express ILꢀ2Rβ/γ. RLI was preferred to ILꢀ15 because of its
38, 39
better efficacy (Figure S3).
of cytotoxicity (Figure S4).
In addition, the main compounds were checked for their lack
2.3.1 Probing the linker to interact with a lipophilic pocket. Based on our binding model
(Figure 4), it appeared that the main phthalazinoneꢀtriazole core of compound 1 was involved
in major interactions with specific residues of the binding site, such as πꢀalkyl hydrophobic
interactions with Ile68 and Leu69, and carbon HB with Asn65. Furthermore, the model
revealed a large hydrophobic patch formed by residues Val3, Ile6, Val104 and Ile111, close to
the predicted location of the terminal aromatic group of compound 1 and which could be
9
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 92
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
targeted for compound optimization. We reasoned that considering the bisꢀcyclic aromatic
part to be the main anchorage of the hit 1, lengthening the alkyl chain could increase the
affinity and activity through additive interactions with the lipophilic channel. We therefore
prepared compounds 41 to 46, which are analogs of compound 1 bearing an alkyl chain with 3
to 15 methylene groups.
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
In the proliferation cellular assay, the IC value followed an inverted bellꢀshaped curve when
50
the length of the alkyl chain was increased (Figure S5), decreasing from 17.3 ꢀM (n = 1) to
.9 ꢀM (n = 7) and increasing back to 7.3 µM (n = 15) (Table 1). Increasing the alkyl chain
therefore led to a gain of one order of magnitude in activity potency and identified compounds
3ꢀ45 as potential leads (Table 1). We prepared compound 46 with a long chain (n = 15) to
1
4
test the limit of such chemical modification. Surprisingly, compound 46 kept the activity,
suggesting a good adaptation of this long flexible lipophilic chain probably in a folded form.
However, no ILꢀ15 versus ILꢀ2 selectivity could be observed at this stage and the best IC50
values of 1.9 and 1.1 ꢀM were obtained with compound 43 for the inhibition of ILꢀ15ꢀ and
ILꢀ2ꢀinduced cell proliferation respectively.
Table 1. In vitro inhibition of 32Dβ cell proliferation with compound 1 and its homologs
41ꢀ46.
a
IC50 (ꢀM)
Cpd
1
n
b
c
RLI
ILꢀ2
1
17.3
11.6
10
ACS Paragon Plus Environment

Page 11 of 92
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
41
3
5
9.8
4.1
1.9
2.3
3.3
7.3
11.2
42
9.8
43
7
1.1
44
9
1.6
45
10
15
4.4
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
46
4.6
a
b
All values are the mean of two or more independent assays
.
Inhibition of 32Dβ cell proliferation
c
induced by RLI (0.1 nM). Inhibition of 32Dβ cell proliferation induced by ILꢀ2 (1.5 nM).
2.3.2 SAR study around the terminal amidic part of the molecules. The terminal amidic
part of the molecules, was predicted to form a stacking interaction between its substituted
terminal phenyl ring and the Ser7 and Asp8 peptidic bond and an H bond between its amide
function and the carboxyl function of Asp8 of ILꢀ15. Accordingly, the Nꢀmethylation of the
amide function (Table 2, compound 47) decreased the inhibitory effect. Unexpectedly, the
substitution of both 2,4ꢀchloro substituents with various halogens such as bromide or fluoride
(
Table 2, compounds 48 and 49) gave the same results. We next focused on the para position
of the terminal phenyl group, which appeared not far from lipophilic residues such as Ile6 and
Val104 in the structure of ILꢀ15 (Figure 4B). We therefore introduced hydrophobic
substituents such as iodide (Table 2 and Figure 5A, compound 50) or alkoxy substituents with
variable alkyl chains (Table 2, compounds 51, 58ꢀ60). To complete the series, we also
selected the adamantane moiety as an interesting lipophilic and steric system to substitute the
terminal phenyl group (Table 2, compound 54). Uniformly, no improvement in the activity
was observed compared to compound 1, and neither with compound 56 (Figure 5A), with a
triazole ring in place of the amide function in 54. Finally, to try to favor HB interaction with
Lys11, carboxylic and ester functions were introduced into the ortho (compounds 52 and 61)
or meta (compounds 53 and 62) positions of the aromatic phenyl cycle but such modifications
were detrimental for the activity because no more effect was observed (Figure 5A). In this
series, all attempts to improve the biological results failed. Because the IC50 value of the hit
11
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 92
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
compound 1 was close to the limit of sensitivity of the cell proliferation assay, the SAR study
of that series of compounds (n = 1) could not be refined up to 30 µM (Figure S6).
We then performed modulations on compounds 43 (n = 7) and 45 (n = 10) by varying the
terminal amidic part of the molecules. As both compounds had a longer aliphatic linker, it
could be expected that this lipophilic part of the molecule would interact more efficiently with
the hydrophobic channel formed by residues Val3, Ile111, Ile6 and Val104 of ILꢀ15. The
investigation first focused on the 2ꢀ and 4ꢀchloro substituents of the phenyl group as explored
with the hit 1. From the 2,4ꢀdichlorophenyl analog 45 (n = 10), deletion of one chloro
substituent in the para (compound 86) or ortho (compound 87) positions of the phenyl group
afforded compounds that were three times less potent, what was not observed with the
monochloro analog 71 of compound 43 (n = 7), which retained a good antiproliferative
activity (IC50 = 0.8 ꢀM, Figure 5B). Substitution of the phenyl ring of 45 with hydrophobic
and bulky groups, such as the naphthyl and adamantyl rings retained the activity (Table 2,
compounds 88 and 89, Figure 5C) but, surprisingly, deletion of this hydrophobic amidic part
of the molecule in compound 45 to give the Sꢀdecyl analog 90 did not dramatically affect the
results (IC50 around 10 ꢀM, Figure 5C).
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
We then turned our attention to the SAR on the terminal phenyl group of our best lead
compound 43 (n = 7). Replacement of one or both chloro substituents with bromide or
fluoride (compounds 64 and 66), with an apolar group such as methyl (compounds 67) or a
substituent such as methoxy (compounds 68 and 69) was uniformly tolerated by both ILꢀ15
and ILꢀ2 systems, with IC values in the micromolar range (Table 2). Deletion of the chloro
50
substituent at the 2ꢀposition of the phenyl ring to give the 4ꢀchloro phenyl compound 71 or
analogs bearing apolar or polar systems (compounds 72ꢀ74) also showed similar activity
(Table 2, Figure 5D). The azide analog 73 has been developed as a potential partner for a
crossꢀlinking study with the ILꢀ15 protein (data not shown). At this stage we could conclude
12
ACS Paragon Plus Environment

Page 13 of 92
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
that the terminal amidic part of the leads 43 and 45 bearing a long alkyl chain did not appear
to be crucial for anchorage on the binding site. Nevertheless, simple Nꢀmethylation of the
amide function of our lead compound 43 (compound 63) as well as 2ꢀ and 4ꢀ monoꢀ or diꢀ
substitution with a hydroxyl function (compounds 70, 75 and 79), caused an approximately
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
0ꢀfold decrease in the biological activity (Table 2). These results observed with phenolic
compounds suggested that binding to the protein was highly sensitive to acidic functions. This
was confirmed with the ester analog 76, whose acidic form 83 showed no effect. Moreover,
the ethyl ester 76 was identified as one of the best products tested with an IC50 value of 0.8
ꢀ
M (Table 2).
Starting with the inactive phenol analog 79, alkylation with short alkyl chains such as methyl,
propyl, isopropyl or isobutyl restored the activity (compounds 74, 80ꢀ82, Table 2 and Figure
5
E). Furthermore as observed with compound 45, substitution of the phenyl ring with a bulky
lipophilic substituent such as naphthyl (compounds 77 and 88) or adamantyl (compounds 78
and 89) preserved the activity and uniformly higher efficiency was observed with analogs
with the sevenꢀmembered chain (Table 2, Figure 5C).
2
.3.3 Modification of the mercapto triazole part of compounds 43 and 45. Suspecting
potential biological oxidative events in cellular assays, we reacted the dibromo compound 64
in a chemical oxidation step with metaꢀchloroperbenzoic acid to afford the sulfone derivative
65. Interestingly, the same results were obtained with the alkylsulfide parent 64 and its
sulfone 65, on ILꢀ15 as well as on ILꢀ2. As observed in our binding model (Figure 4), the
sulfur atom is surrounded by two asparagine residues, Asn65 and Asn4, which constitute a
favorable environment for the oxygen of 65 to make HB.
However, in relation to the model, a carbon HB with Asn65 and the methyl group of the Nꢀ
methyl triazole moiety was predicted. Consequently, we wanted to study the substitution of
this Nꢀmethyl system with an oxygen atom and so prepared the oxadiazole analogs 85 and 91
13
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 92
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 corresponding triazole compounds 43 and 45. No drastic decrease of the IC values
50
was observed and oxadiazole 85 appeared as one of the best compounds tested, with an IC50
of 0.8 ꢀM on both interleukines (Table 2, Figure 5F). The formation of a HB bond between
the amine function of Asn65 residue and the oxygen of the oxadiazole ring, could explain this
result. Because no ILꢀ15 versus ILꢀ2 selectivity could be observed, we didn’t go further with
the oxadiazole series.
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 2. SAR study with analogs of the selected compounds 1, 43 and 45.
a
a
a
IC50 (µM)
IC50 (µM)
IC50 (µM)
n = 1
n = 7
n = 10
b
c
b
c
b
c
RLI
ILꢀ2
RLI
ILꢀ2
RLI
ILꢀ2
X
A
R1
R2
Cpd
1
Cpd
43
Cpd
45
1
±
7.3
2.5
11.6
± 1.6
1.9
± 0.4
1.1
± 0.2
3.3
4.4
NMe
S
H
± 0.4
± 1.3
11.9
± 3.6
8.4
± 0.9
d
NMe
NMe
S
S
Me
H
47
48
> 30
NE
63
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
1
.5
1.3
± 0.5
> 30
ND
64
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
±
0.4
1
.4
1.6
± 1.1
e
NMe SO2
H
ꢀ
ꢀ
65
±
0.4
2.3
± 0.9
1.0
± 0.7
NMe
NMe
NMe
NMe
NMe
S
S
S
S
S
H
49
NE
ND
66
67
68
69
70
2
.3
1.7
± 0.5
H
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
±
±
1.0
1
.1
0.9
± 0.5
H
0.2
1.3
± 0.4
1.0
± 0.2
H
f
f
H
17
21
f
NMe
NMe
S
S
H
H
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
86
87
15
ND
ND
0
.8
0.8
± 0.2
f
71
20
±
0.2
14
ACS Paragon Plus Environment

Page 15 of 92
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
.3
0.8
± 0.4
1.1 ±
0.3
3.2
± 1.2
NMe
NMe
S
S
H
H
50
NE
ꢀ
NE
ꢀ
72
73
ꢀ
ꢀ
ꢀ
ꢀ
±
±
±
0.5
1
.8
0.6
3
1.8
.9
NMe
NMe
NMe
S
S
S
H
H
H
51
57
58
NE
NE
NE
NE
74
79
80
> 30
.6
0.8
1.1
> 30
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
0.9
± 0.7
0.6
± 0.2
3.0
> 30
> 30
±
NMe
NMe
S
S
H
H
59
60
> 30
≈ 30
> 30
≈ 30
81
82
ꢀ
ꢀ
ꢀ
ꢀ
± 0.5
4.8
± 1.9
± 1.9
f
f
NMe
NMe
NMe
NMe
NMe
S
S
S
S
S
H
H
H
H
H
ꢀ
ꢀ
75
27
23
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
52
53
61
62
≈ 30
NE
NE
NE
≈ 30
NE
NE
NE
ꢀ
ꢀ
0
.8
0.6
± 0.4
76
±
0.1
ꢀ
ꢀ
83
77
> 30
ND
1
.9
1.9
± 0.4
6.6
± 1.7
f
NMe
S
H
ꢀ
ꢀ
88
8.5
±
0.2
1
.6
1.8
± 0.3
3.8 ±
0.7
4.9 ±
1.4
NMe
S
S
H
H
54
NE
ꢀ
NE
ꢀ
78
85
89
91
±
0.2
0
.8
0.8
± 0.5
8.7
± 0.6
f
O
7.5
±
0.3
a
a
IC50 (µM)
IC50 (µM)
R
Cpd
b
c
RLI
ILꢀ2
56
≈ 30
≈ 30
> 30
84
> 30
f
90
7.6 ± 1.5
10
a
b
All values are the mean of at least two independent assays
.
Inhibition of 32Dβ cell proliferation
c
d
induced by RLI (100 pM). Inhibition of 32Dβ cell proliferation induced by ILꢀ2 (1.5 nM).
Compound 63 was prepared by Nꢀmethylation of amide 43 in DMF, in the presence of NaH (1 equiv)
and MeI (2 equiv). Compound 65 was prepared by Sꢀoxidation of thioether 64 in dichloromethane in
e
the presence of 5 equivalents of mCPBA. ND: not determined, NE: no effect.
15
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 92
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
2
.3.4 Measurement of direct interaction and pharmacological effect of selected
compounds on IL-15 and IL-2 systems. We further analyzed our best compounds using two
supplementary cytokine specific assays. We first compared them based on their direct binding
to RLI or ILꢀ2 using the SPR technology and then measured their potency in reducing the
phosphorylation of Stat5, a signal transducer that is activated when ILꢀ15 or ILꢀ2 binds to its
respective receptor expressed at the surface of NKꢀ92 cells (Figure S7).
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 advantage of the pꢀStat5 assay over the cell proliferation assay used in the first SAR
study, is that the readout is obtained over a short time (1 h), close to the initial cytokine
receptor binding event, and does not result from the sum of several downstream signaling
reactions, occurring in cascade and/or in parallel. Thus, the pꢀStat5 assay represents a specific
biological test more adapted to a fine study of SAR in a series of compounds.
The great majority of the selected compounds displayed K
d
values for RLI binding of around
1
0ꢀ20 µM, close to those for ILꢀ2 binding, when measured (Table 3). The compound
displaying a somewhat weaker affinity (78 µM) in this series (n = 7) was the adamantyl
derivative 78. This was probably due to the bulky nature of the tricyclic moiety, compared to
the aromatic phenyl one. Interestingly, lengthening the central alkyl linker to give the
d
adamantyl analog 89 (n = 10), decreased the K value to 16 µM, suggesting a better
hydrophobic interaction with the binding site on ILꢀ15, possibly near the hydrophobic patch
formed with residues Val3, Ile111, Ile6 and Val104.
Regarding the in vitro functional pꢀStat5 results, the IC50 values measured were mainly
around one order of magnitude lower than those found in the cell proliferation test (Table 3).
This could be explained by the shorter incubation time (1 h versus 2.5 days for pꢀStat5 and
cell proliferation, respectively), which could have a different impact on compound integrity
and efficiency. Overall, these biological results in accordance with those previously obtained
16
ACS Paragon Plus Environment

Page 17 of 92
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
in the cell proliferation assay, confirmed the gain in efficacy by lengthening the central alkyl
chain of compound 1 or 54, to afford the dichlorophenyl analogs 43 and 45 or the adamantyl
analogs 78 and 89, respectively (Figure 5G). Here too, SAR studies with both series (n = 7
and n = 10), did not reveal a unique lead compound (Table 3). However, the best results were
obtained with the phenoxy derivatives 80 and 82 and the ethyl ester analog 76 with pꢀStat5
inhibitory potencies of around 50 nM (Table 3). In the ether series, compounds with alkyl
chains having fewer than three extended carbons with or without ramification, such as
methoxy 74 and isopropoxy 81, were the least efficient (Figure 5H). The ester 76 could be
classified as the best compound in this study with pꢀStat5 inhibitory efficiencies of 57 and 43
nM with ILꢀ15 and ILꢀ2 respectively, and a cell proliferation IC50 of 0.8µM with both
cytokines.
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
Taken together, these results led to the discovery of the first smallꢀmolecule inhibitors
of the ILꢀ15 system showing efficacies in the fifty nanomolar range in a pꢀStat5 assay and in
the micromolar range in a cell proliferation assay as observed with the ester 76 and the alkoxy
derivatives 80 and 82. These compounds also proved to be effective on the ILꢀ2 system. It
should be noted that similar affinity or activity values were described in earlier studies for
nonꢀpeptidic smallꢀmolecule inhibitors of ILꢀ2 however, in that case, targeting specifically the
15, 16, 40, 41
ILꢀ2/ILꢀ2Rα interface.
Ongoing clinical trials are being conducted with humanized MikꢀBetaꢀ1 antibody
(HuꢀMikꢀBeta1) directed against ILꢀ2Rβ and which targets both ILꢀ15 and ILꢀ2 systems. This
antibody is currently tested in patients with arthritis, celiac disease, multiple sclerosis and
4
2,
43
HTLVꢀ1ꢀassociated
myelopathy/tropical
spastic
paraparesis
(HAM/TSP)
(clinicaltrials.gov), and has also been previously shown to prolong primate cardiac allograft
44
survival. Taking advantage of their combined profile, the kind of inhibitors described in this
17
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 92
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
study may be of therapeutic value in pathologies where both cytokines have been shown to be
17
45
involved such as RA, HAM/TSP, inflammatory bowel disease, or multiple sclerosis.
a
Table 3. Pharmacological data of selected potent 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
0
1
2
3
4
5
6
7
8
9
0
Inhibition of pꢀStat5 in NK92
Direct interaction
SPR (Kd µM)
cells
(IC50 nM)
R
n
Cpd
1
RLI
linked
ILꢀ2 linked
ILꢀ15 (50pM)
9110 + 1470
ILꢀ2 (250pM)
1
48.4
10.5
8.0
14070 + 2170
4
6
3
4
19.1
11.6
153 + 30
92 + 19
110 + 18
76 + 6
160 + 46
83 + 20
76 + 10
69 + 6
10.0
ND
b
66
ND
6
6
6
7
8
9
24.1
19.6
21.2
ND
7
27.6
12.5
205 + 10
360 + 6
162 + 24
226 + 17
7
8
4
0
26.0
19.0
ND
ND
201 + 12
40 + 5
204 + 28
53 + 1
18
ACS Paragon Plus Environment

Page 19 of 92
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
8
1
20.9
25.0
ND
ND
93 + 13
55 + 7
89 + 6
82
58 + 14
7
7
6
3
21.8
42.4
78
23.1
ND
ND
57 + 2
43 + 10
73 + 8
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
100 + 20
210 + 11
78
45
152 + 21
ND
ND
ND
142 + 10
124 + 8
159 + 16
59 + 5
10
1
8
9
16.0
54
ND
ND
>30000
>30000
a
b
All values are the mean of two or more independent assays. ND = not determined.
3
. CONCLUSION
In the present study, we report the development process of hit optimization from structureꢀ
based drug virtual screening to smallꢀmolecule compounds, blocking the interaction of ILꢀ15
to its receptor with subꢀmicromolar efficiency. This is the first and successful attempt to target
the ILꢀ15 system by a PPI inhibitor approach, which opens the way for the identification of
new classes of therapeutic drugs for ILꢀ15 related pathologies such as RA or graft rejection.
Binding and biological cell proliferation assays led to the selection of one hit among a number
of compounds initially selected by virtual screening. This hit has an inhibitory activity in the
ten micromolar range, and was used as a starting template for lead discovery.
Pharmacomodulation studies around the linker chain and the terminal amidic part of the hit
19
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 92
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
led to inhibitors with IC50 values close to 50 nM in cellꢀbased experiments, nevertheless
showing the same activity on the ILꢀ2 system. Having now identified the first class of smallꢀ
molecule inhibitors of the ILꢀ15/IL15R system, we are currently investigating the
development of analogs with higher ILꢀ15 selectivity, to rule out possible side effects to
specific ILꢀ2 blockade, such as the decrease of regulatory T cells.
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
4
. EXPERIMENTAL SECTION
4.1. Computational study
Pharmacophore modeling. At the time of this work, the structure of ILꢀ15 in complex with
ILꢀ2Rβ was not yet available. The Xꢀray crystal structure of ILꢀ15 extracted from the
2
1
structure of the ILꢀ15/ILꢀ15Rα complex (PDB code 2Z3Q) was thus used. The protein
structure was firstly prepared by adding the hydrogen atoms, removing the water molecules,
and inserting the missing loop regions using the Prepare Protein tool within DS3.0. After a
series of energy minimization carried out with CHARMM force field and Steepest Descent
algorithm implemented into DS3.0, the protein was used for generating a pharmacophore
model using the Structure Based Pharmacophore tool within DS3.0. Pharmacophore models
were created from the binding site of ILꢀ15 with ILꢀ15Rβ based on “hot spots” residues
(Asp8, Asn65 and Leu69). The superimposition of ILꢀ15 onto ILꢀ2 in the tetrameric complex
24
structure ILꢀ2:ILꢀ2Rα/β/γ (PDB code 2B5I) , had led us to consider an additional orientation
for the side chain of residue Asn65 of ILꢀ15, similar to that of its homologous ILꢀ2 residue
Asn88. Moreover, the Xꢀray crystal structure of ILꢀ2:ILꢀ2R complex shows two water
molecules involved in the hydrogen bonding networks between Asp20 and Asn88 (ILꢀ2
counterpart of Asp8 and Asn65 of ILꢀ15) and the side chains of ILꢀ2Rβ. Thus, two water
molecules close to residues Asp8 and Asn65 of ILꢀ15 were also taken into consideration. The
pharmacophore model was built using hydrogen bond donor (HBD), hydrogen bond donor
20
ACS Paragon Plus Environment

Page 21 of 92
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
and acceptor (HBDA) and HY (hydrophobic) features on the basis of the four ILꢀ15 residues
Asp8, Asn65, Ile68 and Leu69. Two HBD projections were centered on OD1 atom within
residues Asp8 and Asn65 (two orientations of its side chain considered) and a HBDA
projection on OD2 atom within residue Asp8, and all were located within a sphere of 2.2 Å
radius. The HY feature was manually created at distances of 5.062, 4.554 and 6.346 Å from
carbon alpha (CA) atoms of residues Ile68, Leu69 and Asn65 respectively. Exclusion spheres
of 1.6 Å radius were generated, centered on carbon beta (CB) atoms of residues Asn1, Asn4,
Ser7, Asp8, Thr62, Asn65, Ile68, Leu69 and Asn72 and on carbon gamma (CG) atom of
Lys11. Two HBDA projections were placed on water molecules 147 and 148. Finally, virtual
screening was done using as main features HBD and HBDA projections on Asp8, the HY
feature and the exclusion spheres and as accessory features the HBD projection on Asn65 and
HBDA projections on water molecules. A hit had to match at least the main features while the
presence of the accessory features was not mandatory.
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
Virtual screening. The chemical library for virtual screening was obtained by merging
commercial (Chembridge, Maybridge, Preswick) and academic collections (the Diversity set
II from the NCI and the French national chemical library Chimiothèque Nationale). Duplicate
structures were removed and 3D coordinates (using Catalyst) were generated using the
Prepare ligand tool within DS3.0. A multiꢀconformation ligand database was then created
using Catalyst within the Build 3D Database tool under DS3.0. The query was performed
using the Search 3D Database tool with the FAST search method under DS3.0. Hits were
screened for Fit values and those with a Fit value more than 1 were then taken for molecular
docking analysis.
Molecular docking. Molecular docking studies were performed using LigandFit option of
receptorꢀligand interactions protocol section available in DS3.0. Initially the protein was
prepared by adding the hydrogen atoms, removing the water molecules and inserting the
21
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 92
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
missing loop regions and then was minimized using CHARMM as described above. The
protein molecule thus prepared and minimized was defined as the total receptor. Four
docking conditions (4 targets: ILꢀ15 protein with residue Asn65 side chain in its initial
orientation (as referred to structure 2Z3Q) or in an orientation close to that of its ILꢀ2
homolog Asn88 (from structure 2B5I), as well as each of these targets with 2 water
molecules) were conducted in parallel. For each condition, the ligand molecules retained by
the pharmacophore model were docked into a binding site facing residues Asn1, Asn4, Ser7,
Asp8, Lys11, Thr62, Asn65, Ile68, Leu69 and Asn72 of ILꢀ15 structures. The volumes 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
3
binding site were 1396, 1538, 1481 and 1523 Å and contained 11171, 12304, 11851 and
12188 points respectively to the 4 targets described above. Docking was performed using CFF
as the energy grid. A penalty of 200 kcal/mol/atom was set up to reduce the dock score of
poses that occurred outside of the binding site. The conformational search of the ligand poses
was performed by the Monte Carlo trial method. Maximum internal energy was set at 10000
kcal/mol. A short rigid body minimization was then performed (steepest descent and Broyden
Fletcher Goldfarb Shanno (BFGS) minimizations). Twenty poses were saved for each ligand
after docking and 100 steps of BFGS rigid body minimization were then carried out. Scoring
46
47
was performed with five scoring functions: LigScore1, Ligscore2, PLP1, PLP2 and
4
8, 49
PMF.
CFF force field was used for LigScore calculations. Among the 24115 compounds
retained after the pharmacophoreꢀbased search, 7991 were actually docked to the targets. To
improve the screening accuracy, a consensus strategy was adopted after dividing compounds
in subsets based on MW (MW ≤ 600 and > 600 Da) and number of rotatable bonds (rotatable
bonds ≤ 5 and > 5). Compounds with a consensus score equal to 5 (i.e. in the TOP 40 % for
each scoring function) were retained within each subset for each docking condition. All
together, 1588 compounds were thereby retained. They were ranked by rescoring them
against each 4 targets and summing the four consensus score obtained (maximal score = 20).
22
ACS Paragon Plus Environment

Page 23 of 92
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
The topꢀranked 20 % were selected and represented 315 compounds, 240 of which were
actually available and purchased for biological testing.
Binding mode and molecular interactions analysis of compound 1. In order to determine the
potential binding mode of compound 1, we performed a refinement step on the top 50 poses
obtained from two more docking runs, one using LigandFit and the other using CꢀDocker,
another docking program implemented under DS3.0. This refinement procedure consisted in a
simulated annealing refinement (2000 steps at 700K, 5000 steps at 300K) followed by a final
full forcefield minimization using CHARMM (100 steps of steepest descent followed by
conjugate gradient with gradient tolerance of 0.01). The 100 poses retained with the two
docking runs were compared in term of interaction energy and those with the most favorable
binding were further visually validated for their proper docking. The top 10 poses belonged to
the same cluster and one of them has been selected.
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
Images for the figures 1 and 4 were captured from DS3.0 visualizer and treated with Adobe
Photoshop CS4 software.
4.2 Chemistry. General. Melting points were determined on a Stuart melting point SMP3
apparatus. All solvents used were reagent grade and TLC was performed on silicaꢀcovered
aluminum sheets (Kieselgel 60F , MERCK). Eluted TLC was revealed using UV radiation
2
54
(λ = 254 nm), or molybdate solution. Flash column chromatography was performed on silica
gel 60 ACC 40ꢀ63 µm (SDSꢀCarloErba). NMR spectra were recorded on a BRUKER AC300
1
13
1
(
300 MHz for H and 75 MHz for C) or on a BRUKER 400 (400 MHz for H and 100 MHz
13
for C) at room temperature, on samples dissolved in an appropriate deuterated solvent.
1
13
References of tetramethylsilane (TMS) for H and deuterated solvent signal for C were used.
Chemical displacement values ( ) are expressed in parts per million (ppm), and coupling
δ
constants (J) in Hertz (Hz). Clearly identified proton and carbon were specified (for
dihydrophthalazinone moiety as Phthal and for triazole moiety as Triaz). Lowꢀresolution
23
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 92
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
mass spectra (MS in Da unit) were recorded in the CEISAM laboratory on a ThermoꢀFinnigan
DSQII quadripolar at 70 eV (CI with NH gas) or on a Waters Xevo G2ꢀXS QTOF. Highꢀ
3
Resolution Mass Spectrometry (HRMS in Da unit) analyses were recorded on an LCꢀQꢀTOF
(SynaptꢀG2 HDMS, Waters) in the IRSꢀUN center (Mass Spectrometry platform, Nantes) or
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
on a MALDI ꢀTOFꢀTOF apparatus (Autoflex III from Bruker) in the INRA center (BIBS
platform, Nantes). The purity of the tested compounds was determined by analytical UPLCꢀ
MS (Acquity HꢀClassꢀ Xevo G2ꢀXS Qtof) using C18 reverseꢀphase column [BEH (1.7 ꢁm,
2.1 × 50 mm)] and was over 95%. The compounds described in Tables 1ꢀ3 were checked for
50
the presence of PAINS substructures as described by Baell and Holloway, using FAFꢀ
5
1
Drugs3 server and no PAINS substructures were identified except in 73 which was only
developed for the purpose of carrying out crossꢀlinking studies with ILꢀ15 protein .
Synthesis and chemical characteristics of the halogenoalkylamides 6ꢀ39 and the
azidoalkylamide 40 are detailed in the Supporting Information.
General Synthetic Procedure 1 for the Synthesis of Bromoalkylamides. The commercial
bromoacid reagent (1 eq) was dissolved in anhydrous DCM and oxalyl chloride (1.2 eq) was
added at room temperature. A catalytic amount of DMF was then added. The mixture was
stirred under argon atmosphere during 3 h. Carboxylic acid complete conversion to acyl
1
chloride was determined by H NMR in deuterated benzene after concentration in vacuum.
The acyl chloride was taken up in anhydrous THF and the amine (0.85 eq) was then added at
0
°C, followed by triethylamine (2.5 eq). After completion of the reaction, water was added
and the aqueous phase was extracted with Et O (the aqueous layer was acidified to pH = 2 for
2
compounds with phenol function). The organic phase was washed with brine, dried over
4
MgSO and concentrated under reduced pressure. The expected bromoalkylamide product
was then purified by flash column chromatography on silica gel.
24
ACS Paragon Plus Environment

Page 25 of 92
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
General Synthetic Procedure 2 for the S-alkylation of 4-((5-mercapto-4-methyl-4H-1,2,4-
triazol-3-yl)methyl)-2-methylphthalazin-1(2H)-one (4) and 4-((5-mercapto-1,3,4-oxadiazol-
2
-yl)methyl)-2-methylphthalazin-1(2H)-one (5) with halogenoalkyles or halogenoalkylamides
to give compounds 1, 41-54, 64, 66-78, 86-91. To a suspension of dry K CO (3 eq, 3 mmol)
in DMF (8 mL) was added thiol 4 (1 mmol). The mixture was stirred for 15 min and a
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
3
ꢀ1
solution of bromoalkylamide (1.5 eq) in CH Cl (1 mmol.mL ) was added. The mixture was
2
2
then stirred at room temperature until starting material disappeared (2 h to 18 h). After
completion of the reaction, a saturated solution of ammonium chloride was added. The
mixture was extracted twice with ethyl acetate. The organic layer was washed several times
with brine, dried over MgSO and concentrated under reduced pressure. The crude was
4
purified by flash chromatography on silica gel.
N-(2,4-dichlorophenyl)-2-((4-methyl-5-((3-methyl-4-oxo-3,4-dihydrophthalazin-1-
yl)methyl)-4H-1,2,4-triazol-3-yl)thio)acetamide (
from 4ꢀ((5ꢀmercaptoꢀ4ꢀmethylꢀ4 ꢀ1,2,4ꢀtriazolꢀ3ꢀyl)methyl)ꢀ2ꢀmethylphthalazinꢀ1(2
one (1 mmol, 0.29 g) and 2ꢀbromoꢀ ꢀ(2,4ꢀdichlorophenyl)acetamide (3 mmol, 0.84
1).
Following procedure 2 and starting
H
H
)ꢀ
4
N
6
g), purification by column chromatography on silica gel (DCM/MeOH; 90:10) afforded
1
compound
CDCl
Hz, 1H, H ), 8.19 (dd, J =7.8 and J = 1.5 Hz, 1H, HarꢀPhthal), 7.82 (ddd, J = 7.9, J = 7.8 and J
1
in 81% yield, as a white powder; mp 201.5 ꢀ 202 °C. H NMR (400 MHz,
3
) δ 10.0 (bs, 1H, NH), 8.45 (dd, J = 7.8 and J = 1.1 Hz, 1H, HarꢀPhthal), 8.23 (d, J = 8.8
ar
=
1.5 Hz, 1H, HarꢀPhthal), 7.77 ( ddd, J = 7.9, J = 7.8 and J = 1.1 Hz, 1H, HarꢀPhthal), 7.31 (d, J =
2.4 Hz, 1H, Har), 7.18 (dd, J = 8.8 and J = 2.4 Hz, 1H, Har), 4.49 (s, 2H, CH ), 4.05 (s, 2H,
) ) 167.1
2
13
CH
2
), 3.79 (s, 3H, NMePhthal), 3.53 (s, 3H, NMe). C NMR (100 MHz, CDCl
3
δ
(CONH), 159.5 (CONMe), 152.9 (CIVꢀC=N), 152.0 (CIVꢀC=N), 140.5 (CIVꢀC=N), 134.0 (CIV),
25
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 92
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
133.5 (CHarꢀPhthal), 132.1 (CHarꢀPhthal), 129.6 (C ), 129.1 (CH ), 129.0 (C ), 128.2 (C );
IV
ar
IV
IV
1
27.6 (CH ), 127.3 (CHarꢀPhthal), 125.1 (CHarꢀPhthal), 124.7 (C ), 123.2 (CH ), 39.5
ar IV ar
(NMePhthal), 36.0 (CH
2
2
), 30.8 (NMe), 30.02 (CH ). HRMS (MALDI: DHB, PEG 600): calcd
+
for C21 19Cl
H
2
N
6
O
2
S (M+H) 489.0662, found 489.0637.
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
3
1
2
-(3-Methyl-4-oxo-3,4-dihydrophthalazin-1-yl)acetohydrazide (3). To a solution of ethyl
[
(E)ꢀ3ꢀoxoꢀ1,3ꢀdihydroisobenzofuranꢀ1ꢀylidene]acetate 2ꢀE (13 mmol, 2.83 g,) in EtOH (15
mL), was added dropwise methylhydrazine (13 mmol, 0.52 mL). The mixture was then stirred
for 30 min and hydrazine monohydrate was added dropwise (6 eq, 78 mmol, 4.45 mL). The
solution was finally heated at reflux during 3h and the resulting suspension was cooled to 20
°
C and filtered. The white powder obtained was washed with Et O to give 94% of hydrazide 3
2
1
which was used in the next step without further purification. H NMR (400 MHz, DMSOꢀd
6
)
δ
9.36 (bs, 1H, NH), 8.28 (d, J = 7.6 Hz, 1H, HarꢀPhthal), 7.92 (m, 2H, 2 x HarꢀPhthal), 7.85 (m, J
=
4.3 Hz, 1H, HarꢀPhthal), 4.27 (bs, 2H, NH
2
2
), 3.77 (s, 2H, CH ), 3.70 (s, 3H, NMe). HRMS
+
(
MALDI: DHB, PEG 600) calcd for C H N O (M+H) 233.1033, found 233.1035.
11
13
4
2
4-((5-Mercapto-4-methyl-4H-1,2,4-triazol-3-yl)methyl)-2-methylphthalazin-1(2H)-one (4). To
a suspension of hydrazide 3 (12 mmol, 2.8 g) in EtOH (28 mL) was added
methylisothiocyanate (5 eq, 60 mmol, 4.54 g), triethylamine (5 eq, 60 mmol, 8 mL) and the
reaction mixture was heated to reflux during 8 h. The resulting suspension was cooled to 20
°C, filtered and washed with Et
2
O to give compound 4 with 99% yield as a yellowish powder.
1
This one was used in the next step without further purification. H NMR (400 MHz, DMSOꢀ
6
d ) δ 13.56 (bs, 1H, SH), 8.30 (m, 1H, HarꢀPhthal), 7.97ꢀ7.84 (m, 3H, 3 x HarꢀPhthal), 4.51 (s, 2H,
1
3
CH
2
), 3.67 (s, 3H, NMePhthal), 3.49 (s, 3H, NMe). C NMR (100 MHz, CDCl
3
)
δ
166.9 (CIVꢀ
26
ACS Paragon Plus Environment

Page 27 of 92
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
SH), 158.4 (CONMe), 149.6 (CIVꢀC=N), 140.3 (CIVꢀC=N), 133.2 (CHarꢀPhthal), 131.9 (CHarꢀPhthal),
128.7 (C ), 127.1 (C ), 126.0 (CHarꢀPhthal), 125.6 (CHarꢀPhthal), 39.0 (NMePhthal), 30.2 (NMe),
IV IV
+
+
2
2 14 5
8.3 (CH ). HRMS (ESI ): calcd for C13H N OS (M+H) 288.0914, found 288.0919.
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
4
-((5-Mercapto-1,3,4-oxadiazol-2-yl)methyl)-2-methylphthalazin-1(2H)-one
(5).
The
3
4
compound was prepared as earlier described for the Nꢀallyl analog. To a suspension of
hydrazide 3 (1 mmol, 0.23 g) in CS (10 mL) was added KOH (2.5 eq, 2.5 mmol) and the
2
reaction mixture was heated to reflux under vigorous stirring during 3 h. It was cooled to 10
2
°C and the precipitate formed was filtered, washed with Et O. The powder was stirred in hot
ethanol and filtered to give the compound 5 with 65% yield as a grey powder. This one was
1
used in the next step without further purification. H NMR (400 MHz, DMSOꢀd
6
)
δ
8.30 (m,
13
1
2
H, HarꢀPhthal), 7.98 (m, 3H, 3 x HarꢀPhthal), 4.49 (s, 2H, CH ), 3.70 (s, 3H, NMe). C NMR
(
100 MHz, CDCl3)
δ 178.2 (CIVꢀSH), 160.5 (CO), 158.3 (CIVꢀC=N), 139.7 (CIVꢀC=N), 133.4
(CHarꢀPhthal), 132.1 (CHarꢀPhthal), 128.5 (CIV), 127.1, (CIV) 126.2 (CHarꢀPhthal), 125.4 (CHarꢀPhthal),
ꢀ
ꢀ
45.4 (NMe), 29.1 (CH ). HRMS (ESI ): calcd for C H N O S (MꢀH) 273.0446, found
2 12 9 4 2
2
73.0445.
N-(2,4-dichlorophenyl)-4-((4-methyl-5-((3-methyl-4-oxo-3,4-dihydrophthalazin-1-yl)methyl)-
H-1,2,4-triazol-3-yl)thio)butanamide 41 Following general procedure 2 and starting
from 4ꢀ((5ꢀmercaptoꢀ4ꢀmethylꢀ4 ꢀ1,2,4ꢀtriazolꢀ3ꢀyl)methyl)ꢀ2ꢀmethylphthalazinꢀ1(2 )ꢀ
one (1 mmol, 0.29 g) and 4ꢀbromoꢀNꢀ(2,4ꢀdichlorophenyl)butanamide 7 (3 mmol)
4
(
).
H
H
4
,
purification by column chromatography on silica gel (DCM/MeOH; 95:5) afforded
1
compound 41 in 72% yield, as a white powder; mp 195 ꢀ 196 °C. H NMR (400 MHz,
3
CDCl ) δ 8.40 (m, 1H, HarꢀPhthal), 8.17 (m, 2H, HarꢀPhthal and Har), 8.01 (m, 1H, Har), 7.77 (ddd,J
=
8.2 and J = 7.3 and J = 1.5 Hz, 1H, HarꢀPhthal), 7.72 (ddd,J = 8.2, J = 7.8 and J = 1.3 Hz, 1H,
27
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 92
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
HarꢀPhthal), 7.31 (d, J = 2.3 Hz, 1H, H ), 7.19 (dd, J = 8.9 and J = 2.4 Hz, 1H, H ), 4.44 (s, 2H,
ar
ar
CH ), 3.77 (s, 3H, NMe
), 3.55 (s, 3H, NMe), 3.27 (t, J = 7.1 Hz, 2H, SCH ), 2.61 (t, J =
2
Phthal
2
13
7.1 Hz, 2H, CH
2
CO), 2.17 (tt, J = 7.1 Hz, J = 7.1 Hz, 2H, CH
2
). C NMR (100 MHz, CDCl
3
)
δ
170.5 (CONH), 159.5 (CONMe), 152.2 (CIVꢀC=N), 151.5 (CIVꢀC=N), 140.7 (CIVꢀC=N), 133.4
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
(CIV), 133.3 (CHarꢀPhthal), 131.9 (CHarꢀPhthal), 129.0 (CIV), 128.9 (CIV), 128.8 (CIV), 127.8
(
CH ), 127.1 (CHarꢀPhthal), 125.2 (CHarꢀPhthal), 124.1 (C ), 123.2 (CH ), 39.5 (NMePhthal), 35.8
ar
IV
ar
(
CH
2
), 32.2 (CH
2
), 30.7 (NMe), 30.0 (CH
2
), 25.6 (CH
2
). HRMS (MALDI: DHB, PEG 600):
+
calcd for C23
H
22Cl
2
N
6
O
2
SNa (M+Na) 539.0794, found 539.0819.
N-(2,4-dichlorophenyl)-6-((4-methyl-5-((3-methyl-4-oxo-3,4-dihydrophthalazin-1-yl)methyl)-
H-1,2,4-triazol-3-yl)thio)hexanamide 42 Following general procedure 2 and starting
from 4ꢀ((5ꢀmercaptoꢀ4ꢀmethylꢀ4 ꢀ1,2,4ꢀtriazolꢀ3ꢀyl)methyl)ꢀ2ꢀmethylphthalazinꢀ1(2 )ꢀ
one (1 mmol, 0.29 g) and 6ꢀbromoꢀNꢀ(2,4ꢀdichlorophenyl)hexanamide 9 (3 mmol)
4
(
).
H
H
4
,
purification by column chromatography on silica gel (DCM/MeOH; 90:10) afforded
1
compound 42 in 56% yield, as a white powder; mp 184 ꢀ 185 °C. H NMR (400 MHz,
CDCl
ddd,J = 7.8, J = 7.4 and J = 1.6 Hz, 1H, HarꢀPhthal), 7.74 (ddd,J = 7.8 and J = 8.0 Hz, J = 1.3,
1H, HarꢀPhthal), 7.67 (bs, 1H, NH), 7.33 (d, J = 2.4 Hz, 1H, Har), 7.21 (dd, J = 8.9 and J = 2.4
Hz, 1H, Har), 4.46 (s, 2H, CH ), 3.79 (s, 3H, NMePhthal), 3.53 (s, 3H, NMe), 3.18 (t, J = 7.3
3
) δ 8.42 (m, 1H, HarꢀPhthal), 8.27 (d, J = 8.9 Hz, 1H, Har), 8.20 (m, 1H, HarꢀPhthal), 7.78
(
2
Hz, 2H, SCH ), 2.41 (t, J = 7.3 Hz, 2H, CH CO), 1.77 (m, 4H, 2 x CH ), 1.51 (m, 2H, CH ).
2
2
2
2
1
3
3
C NMR (100 MHz, CDCl ) δ 171.2 (CONH), 159.5 (CONMe), 152.0 (CIVꢀC=N), 151.8 (CIVꢀ
C=N), 140.8 (CIVꢀC=N), 133.5 (CIV), 133.4 (CHarꢀPhthal), 131.9 (CHarꢀPhthal), 129.2 (CIV),
1
29.1(CIV), 128.8 (CHar), 128.1 (CIV), 127.9 (CHar), 127.2 (CHarꢀPhthal), 125.3 (CHarꢀPhthal),
2 2
23.6 (CIV), 122.8 (CHar), 39.5 (NMePhthal), 37.4 (CH ), 32.7 (CH ), 30.7 (NMe), 30.2 (CH ),
1
2
28
ACS Paragon Plus Environment

Page 29 of 92
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
29.2 (CH ), 28.0 (CH ), 24.8 (CH ). HRMS (MALDI: DHB, PEG 600): calcd for
2
2
2
+
C H Cl N O S (M+H) 545.1288, found 545.1281.
25
27
2
6
2
N-(2,4-dichlorophenyl)-8-((4-methyl-5-((3-methyl-4-oxo-3,4-dihydrophthalazin-1-yl)methyl)-
H-1,2,4-triazol-3-yl)thio)octanamide 43). Following general procedure 2 and starting
from 4ꢀ((5ꢀmercaptoꢀ4ꢀmethylꢀ4 ꢀ1,2,4ꢀtriazolꢀ3ꢀyl)methyl)ꢀ2ꢀmethylphthalazinꢀ1(2 )ꢀ
one (1 mmol, 0.29 g) and 8ꢀbromoꢀNꢀ(2,4ꢀdichlorophenyl)octanamide 10 (3 mmol, 1.09
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
4
(
H
H
4
g), purification by column chromatography on silica gel (DCM/MeOH; 98:2) afforded
1
compound 43 in 53% yield, as a white powder; mp 165 ꢀ 166 °C. H NMR (400 MHz,
3
CDCl ) δ 8.43 (m, 1H, HarꢀPhthal), 8.31 (d, J = 9.0 Hz, 1H, Har), 8.22 (m, 1H, HarꢀPhthal), 7.80
(m, 1H, HarꢀPhthal), 7.75 (m, 1H, HarꢀPhthal), 7.60 (bs, 1H, NH), 7.35 (d, J = 2.4 Hz, 1H, Har),
7
.23 (dd, J = 9.0 and J = 2.4 Hz, 1H, Har), 4.47 (s, 2H, CH
H, NMe), 3.18 (t, J = 7.3 Hz, 2H, SCH ), 2.40 (t, J = 7.5 Hz, 2H, CH
). C NMR (100 MHz, CDCl 171.5 (CONH), 159.6
2
), 3.79 (s, 3H, NMePhthal), 3.54 (s,
3
2
2
), 1.73 (m, 4H, 2 x
1
3
CH
2
), 1.48 – 1.29 (m, 6H, 3 x CH
2
3
) δ
(
CONMe), 152.1 (CIVꢀC=N), 152.0 (CIVꢀC=N), 140.9 (CIVꢀC=N), 133.6 (C ), 133.4 (CHarꢀPhthal),
IV
1
32.0 (CHarꢀPhthal), 129.1 (C ), 128.8 (CH ), 128.0 (C ), 127.2 (CHarꢀPhthal), 125.4 (CHarꢀ
IV ar IV
Phthal), 123.3 (CIV), 122.6 (CHar), 39.5 (NMePhthal), 37.9 (CH
2
), 33.1 (CH
2
), 30.7 (NMe), 30.2
(CH
2
), 29.5 (CH
2
), 29.0 (CH
2
), 28.8 (CH
2
), 28.4 (CH
2
), 25.4 (CH ). HRMS (MALDI: DHB,
2
+
PEG 600): calcd for C27
H
31Cl
2
N
6
O
2
S (M+H) 573.1601, found 573.1583.
N-(2,4-dichlorophenyl)-10-((4-methyl-5-((3-methyl-4-oxo-3,4-dihydrophthalazin-1-
yl)methyl)-4H-1,2,4-triazol-3-yl)thio)decanamide (44 . Following general procedure 2 and
starting from 4ꢀ((5ꢀmercaptoꢀ4ꢀmethylꢀ4 ꢀ1,2,4ꢀtriazolꢀ3ꢀyl)methyl)ꢀ2ꢀ
methylphthalazinꢀ1(2 (1 mmol, 0.29 g) and 10ꢀbromoꢀNꢀ(2,4ꢀ
)
H
H
)ꢀone
4
dichlorophenyl)decanamide 11 (3 mmol, 1.17 g), purification by column chromatography on
29
ACS Paragon Plus Environment