Discovery of nanomolar desmuramylpeptide agonists of the innate immune receptor nucleotide-binding oligomerization domain-containing protein 2 (NOD2) possessing immunostimulatory properties

Article abstract of DOI:10.1021/acs.jmedchem.7b01052

Muramyl dipeptide (MDP), a fragment of bacterial peptidoglycan, has long been known as the smallest fragment possessing adjuvant activity, on the basis of its agonistic action on the nucleotide-binding oligomerization domain-containing protein 2 (NOD2). There is a pressing need for novel adjuvants, and NOD2 agonists provide an untapped source of potential candidates. Here, we report the design, synthesis, and characterization of a series of novel acyl tripeptides. A pivotal structural element for molecular recognition by NOD2 has been identified, culminating in the discovery of compound 9, the most potent desmuramylpeptide NOD2 agonist to date. Compound 9 augmented pro-inflammatory cytokine release from human peripheral blood mononuclear cells in synergy with lipopolysaccharide. Furthermore, it was able to induce ovalbumin-specific IgG titers in a mouse model of adjuvancy. These findings provide deeper insights into the structural requirements of desmuramylpeptides for NOD2-activation and highlight the potential use of NOD2 agonists as adjuvants for vaccines.

Full text of DOI:10.1021/acs.jmedchem.7b01052

Subscriber access provided by - Access paid by the | UCSB Libraries
Discovery of nanomolar desmuramylpeptide agonists of the innate
immune receptor nucleotide-binding oligomerization domain-
containing protein 2 (NOD2) possessing immunostimulatory properties
Martina Gobec, Tihomir Tomaši#, Adela Štimac, Ruža Frkanec, Jurij
Trontelj, Marko Anderluh, Irena Mlinaric-Rascan, and Žiga Jakopin
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.7b01052 • Publication Date (Web): 15 Mar 2018
Downloaded from http://pubs.acs.org on March 15, 2018
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 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 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.
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 67
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Discovery of nanomolar desmuramylpeptide agonists of the
innate immune receptor nucleotideꢀbinding oligomerization
domainꢀcontaining protein 2 (NOD2) possessing
immunostimulatory properties
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
Martina Gobec1, Tihomir Tomašič1, Adela Štimac2, Ruža Frkanec2, Jurij Trontelj1, Marko
Anderluh1, Irena MlinaričꢀRaščan1 and Žiga Jakopin1,*
1Faculty of Pharmacy, University of Ljubljana, Aškerčeva 7, SI – 1000 Ljubljana, Slovenia.
2University of Zagreb, Centre for Research and Knowledge Transfer in Biotechnology,
Rockefellerova 10, 10000 Zagreb, Croatia.
*Corresponding author:
Žiga Jakopin
Phone: +386 1 4769 646
Fax: + 386 1 4258 031
Eꢀmail address: ziga.jakopin@ffa.uniꢀlj.si
1
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 67
1
2
3
Abstract:
4
5
6
7
8
9
Muramyl dipeptide (MDP), a fragment of bacterial peptidoglycan, has long been known as the
smallest fragment possessing adjuvant activity, on the basis of its agonistic action on the
nucleotideꢀbinding oligomerization domainꢀcontaining protein 2 (NOD2). There is a pressing
need for novel adjuvants and NOD2 agonists provide an untapped source of potential
candidates. Here, we report the design, synthesis and characterization of a series of novel acyl
tripeptides. A pivotal structural element for molecular recognition by NOD2 has been
identified, culminating in the discovery of compound 9, the most potent desmuramylpeptide
NOD2 agonist to date. Compound 9 augmented proꢀinflammatory cytokine release from
human peripheral blood mononuclear cells in synergy with lipopolysaccharide. Furthermore,
it was able to induce ovalbuminꢀspecific IgG titers in a mouse model of adjuvancy. These
findings provide deeper insights into the structural requirements of desmuramylpeptides for
NOD2ꢀactivation and highlight the potential use of NOD2 agonists as adjuvants for vaccines.
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
Keywords:
Muramyl
dipeptide,
immunomodulator,
desmuramylpeptide,
NOD2
agonist,
immunostimulant, adjuvant
2
ACS Paragon Plus Environment

Page 3 of 67
Journal of Medicinal Chemistry
1
2
3
1. INTRODUCTION
4
5
6
7
8
9
Innate immunity, once assumed to be the redundant remnant of an ancient antimicrobial
system, is in fact the cornerstone of the body’s ability to combat infection.1 Pattern
recognition receptors (PRRs) comprise a constellation of evolutionarily conserved immune
sensors that possess the ability, not only to sense invading pathogens via the detection of their
specific molecular signatures, but also to orchestrate the initial defense response as part of the
innate immune process.1 Nucleotideꢀbinding oligomerization domainꢀcontaining protein 2
(NOD2), that belongs to a major cytosolic class of PRRs, the NODꢀlike receptors (NLR), is of
utmost importance since it has an important task within the innate immune response.2 It is
predominantly expressed in immune cells and, to a lesser degree, in Paneth cells and certain
epithelial cells.2,3 The structure of NOD2 protein is assembled from the following structural
units: two caspase activation and recruitment domains (CARDs), a nucleotideꢀbinding
oligomerization domain (NOD) and multiple leucineꢀrich repeats (LRRs). It has been
suggested that it typically exists in an autoinhibited state.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
Muramyl dipeptide (MDP), a conserved constituent of the cell wall of the majority of types of
bacteria, is the smallest structural subunit of bacterial peptidoglycan capable of eliciting
NOD2 activation5,6; direct contact between MDP and NOD2 has recently been proven by both
biophysical and biochemical assays.7,8 On recognition of MDP, NOD2 undergoes selfꢀ
oligomerization and activates the serineꢀthreonine kinase receptorꢀinteracting protein 2
(RIP2). Subsequently, the nuclear factor κB (NFꢀκB) and mitogenꢀactivated protein kinase
(MAPK) signaling pathways are initiated, resulting in the production of cytokines, certain
antimicrobial peptides and induction of additional processes, which characterize a
proinflammatory/antimicrobial response. 3,7,9 Of note, a number of regulatory proteins closely
control the downstream signaling pathways underlying NOD2 activation.9
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 67
1
2
3
4
5
6
7
8
9
Activation of NOD2 itself is sufficient to shape adaptive immune responses10,11 but, more
importantly, NOD2 agonist synergistically augments the adjuvant capacity of other PRR
ligands such as Tollꢀlike receptor (TLR) ligands. This NLRꢀTLR crosstalk has been proved to
be essential for adaptive immunity and underlines the importance of NLRs.12ꢀ15 Interestingly,
in spite of innate immunity not being known to possess memory, NOD2 ligands have been
shown to induce epigenetic reprogramming of monocytes, resulting in nonspecific protection
against reinfection; this phenomenon has been termed “trained immunity”.16,17 Recently
gained insight into the crosstalk between innate and adaptive immune responses is of
significant clinical importance and has opened up new avenues for development of novel
vaccines.18 Curently, only four adjuvants are approved for human vaccines in the US and
Europe. Although MDP was long ago identified as the smallest component underlying the
efficacy of Freund’s complete adjuvant, thus opening up a new era in immunology, it has
proved to be too pyrogenic for clinical application, while it also suffers from rapid
elimination.19,20 Therefore, MDP has been limited to veterinary vaccines.20 Besides their
adjuvant properties, NOD2 agonists have been shown to possess indirect antimicrobial
properties as well as the ability to act synergistically in combination with antibiotics and to
stimulate the nonꢀspecific resistance against infections and tumors.18,21,22
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
Recent decades have witnessed the discovery of diverse nonpyrogenic MDP analogs, either
chemically modified MDP derivatives or conjugated with carriers, with preserved
adjuvant/immunostimulatory activity.23ꢀ32 Certain lipophilic derivatives of MDP have
exhibited stronger adjuvant activity than that of the parent structure, showing that MDP can
be tailored to increase the adjuvant activity.20, 30, 31 Thus mifamurtide has been tested as part of
an influenza vaccine in a number of clinical trials.21
4
ACS Paragon Plus Environment

Page 5 of 67
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
The term “desmuramylpeptides“ refers to MDP derivatives lacking the sugar Nꢀ
acetylmuramyl moiety (MurNAc). It has been shown unequivocally that the presence of
MurNAc is not mandatory for NOD2 recognition, as evidenced by recently developed
desmuramylpeptides possessing NOD2 agonistic activity.24,25,33,34 Here, we describe the
design, synthesis and characterizarion of a series of novel acyl tripeptides as a continuation of
our efforts to generate desmuramylpeptides with potent NOD2 agonistic activity. The NOD2ꢀ
dependent NFꢀκB activation capacity of the synthesized library of fourth generation
desmuramylpeptides has been determined on HEK293 cells overexpressing the NOD2 gene.
The pivotal structural elements of acyl tripeptides for molecular recognition by NOD2 were
identified. Compound 9, in which the carbohydrate moiety was replaced by a transꢀferuloyl
moiety, has been identified as the strongest desmuramylpeptide NOD2 agonist to date, being
twice as potent as MDP. Its immunostimulatory properties have been determined in in vitro
experiments that revealed the capacity of 9 to augment proꢀinflammatory cytokine release
from human peripheral blood mononuclear cells (PBMC), by itself and in synergy with
lipopolysaccharide (LPS). Furthermore, when coꢀadministered subcutaneously with the model
antigen, ovalbumin (OVA), it induced OVAꢀspecific IgG titers in a mouse model of
adjuvancy. These findings provide deeper insights into the structural requirements of
desmuramylpeptides for NOD2ꢀactivation and point out the potential utility of NOD2 agonists
as adjuvants for vaccines.
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. RESULTS AND DISCUSSION
2.1 Design
The SLC15 family of peptide transporters, outer membrane vesicles and endocytosis have all
been suggested as a potential entry route of MDP into the cytoplasm,35ꢀ38 while
desmuramylpeptides were shown to cross the membrane via passive absorption.39 In
5
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 67
1
2
3
4
5
6
7
8
9
accordance with this, the introduction of lipophilic groups into the structures of NOD
agonists, as well as the use of lipophilic delivery carrier systems, strongly enhanced the
cellular immunity and strikingly increased the adjuvant activity of compounds, further
reinforcing this suggestion.31,40,41 Similarly, previous studies of the structureꢀactivity
relationships (SAR) of MDP derivatives established that introduction of a lipophilic moiety
into MDP can increase its adjuvant activity.42ꢀ44 Although the molecular nature of the
interaction between MDP and NOD2 is not known, in our previous research, cinnamoylꢀGly
(i.e. compound II) and 6ꢀPhꢀindoleꢀGly (i.e. compound I) moieties were highlighted as
appropriate surrogates for the MurNAc group. Leads I and II were identified as effective
NOD2 agonists and were selected for further optimisation (Fig. 1).24,25,34 The structural
requirements thus elucidated for NOD2 agonist activity also include the presence of ethyl
ester groups at the DꢀGlu moiety.
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
Here, we explored the chemical space of our leads I and II (shown in Fig. 1) and investigated
how modifications of the amino acid type and the acyl moiety affect their NOD2 agonist
activity. The underlying principle that guided our design was the previously reported
promising immunostimulatory effect of lipophilic derivatives. In the light of this, we decided
to modify the tripeptide and carbohydrate moiety so that the analogs bear the features of leads
I and II, while differing from each other in subtle substitutions. The first group of
modifications involved replacement of the LꢀAla residue with its more lipophilic counterparts,
LꢀVal and
LꢀPhe, since only limited variations of the essential LꢀAlaꢀDꢀisoGln pharmacophore
are allowed.42,43 Also, a previous study has suggested that, by designing MDP derivatives
harboring LꢀVal, it should be posssible to decouple the inflammationꢀinducing capacity from
the ability to induce adaptive immunity.42 The second group of modifications includes
replacement of the aforementioned acyl moieties by a careful selection of their closely related
6
ACS Paragon Plus Environment

Page 7 of 67
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
mimetics. As the analogs needed to be synthetically accessible, we focused more intensively
on modifications of the cinnamoyl rather than the 6ꢀphenylindole fragment. Unsubstituted and
substituted cinnamoyl moieties are present in many natural products and drugs. The feruloyl
(4ꢀhydroxyꢀ3ꢀmethoxycinnamoyl) moiety, for example, is considered to be a privileged
fragment and has found application as a component of the drug cyclovalone.45 In addition, the
cyclopropyl ring system has been introduced into the cinnamoyl moieties as a bioisosteric
replacement of the alkene bond. It is woth noting that cyclopropane has increasingly been
featured in a wide array of drugs with the aim of preventing the isomerization of the E/Z
alkene bond and/or facilitating the projection of the attached groups in a desirable transꢀ
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
geometry.46 Finally, minor modifications were also introduced into the
DꢀGlu residue of the
tripeptides by equipping this moiety with rigid mimetics. The previously reported
ꢀglutamic acid47 were incorporated into the
conformationally constrained mimetics of
D
structure of 9, used as the starting point, with the aim of potentiating its NOD2ꢀagonistic
activity. This optimization approach has been wellꢀknown among medicinal chemists for
decades. These combined efforts afforded several novel desmuramyldipeptides which were
tested for their NOD2ꢀagonist properties on the commercially available HEKꢀBlue NOD2 cell
line.
7
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 67
1
2
3
4
O
O
H
N
N
H
HN
5
6
7
NH
O
O
I
lead
O
OEt
8
9
OEt
O
O
O
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
F
HO
F
OMe
HN
HN
F
O
O
O
O
F
lipophilic L-Ala substitutes
proposed new MurNAc surrogates
MurNAc surrogate
O
O
N
H
HN
L-Ala-D-Glu diethyl ester
NH
O
O
lead II
O
OEt
OEt
NH
NH
NH
O
O
O
O
EtO
MeO
OMe
OMe
OH
O
O
OMe
NH
NH
N
OEt
O
O
EtO
EtO
Proposed D-Glu(OEt)2 mimetics
Figure 1. Design of acylated tripeptide analogs and structural modifications of leads I and II.
2.2 Chemistry
Preparation of tripeptides was guided by the desire to establish a reproducible and convenient
route suitable for multigram production. Thus, a four step synthesis of the key intermediates
BocꢀGlyꢀLꢀValꢀDꢀGlu(OEt)2 (3a) and BocꢀGlyꢀLꢀPheꢀDꢀGlu(OEt)2 (3b) was executed as
shown in Scheme 1. The starting Dꢀglutamic acid was first esterified via acidꢀcatalyzed
esterification, using SOCl2 in ethanol, to yield the corresponding diethyl ester 1a. The crude
diethyl ester was then coupled with BocꢀLꢀVal or BocꢀLꢀPhe, using dicyclohexylcarbodiimide
(DCC)/1ꢀhydroxybenzotriazole (HOBt) methodology, thus giving the corresponding
8
ACS Paragon Plus Environment

Page 9 of 67
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
dipeptides BocꢀLꢀvalylꢀDꢀglutamic acid diethyl ester (2a) and BocꢀLꢀphenylalanylꢀDꢀglutamic
acid diethyl ester (2b). The Bocꢀprotecting group was then removed from 2a/2b under
acidolytic conditions using trifluoroacetic acid (TFA), followed instantly by coupling with
BocꢀGly, thus affording the Bocꢀprotected tripeptides BocꢀGlyꢀLꢀValꢀDꢀGlu(OEt)2 (3a) and
BocꢀGlyꢀLꢀPheꢀDꢀGlu(OEt)2 (3b). The Bocꢀprotecting group was removed from tripeptides 3a
and 3b, using TFA, giving the corresponding trifluoroacetate salts of the tripeptides. Having
prepared intermediates that enable acylation of the free amino group of Gly at a lateꢀstage ,
we generated several compounds with diverse functionalities in order to ascertain how
minimal modifications of the cinnamoyl moiety affect the NOD2ꢀagonistic activities of the
compounds. By coupling of the commercially available acids with the key intermediates 3a
and 3b, using standard 1ꢀethylꢀ3ꢀ(3ꢀdimethylaminopropyl)carbodiimide (EDC)/HOBt
coupling methodology, corresponding Nꢀacyl GlyꢀLꢀValꢀDꢀGlu(OEt)2 analogs 4ꢀ9 and Nꢀacyl
GlyꢀLꢀPheꢀDꢀGlu(OEt)2 analogs 10ꢀ15 (see Scheme 1) were afforded.
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
Scheme 1.a Synthesis of acylated tripeptide analogs.
a Reagents and conditions: (i) BocꢀLꢀVal or BocꢀLꢀPhe, DCC, HOBt, DIPEA, DMAP, EtOAc;
(ii) TFA/CH2Cl2 (1:5); (iii) BocꢀGly, DCC, HOBt, DIPEA, DMAP, EtOAc; (iv) R'COOH,
EDC, HOBt, DIPEA, DMAP, DMF. * a mixture of two diastereomers; ** a mixture of four
diastereomers.
9
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 67
1
2
3
4
5
6
7
8
Similarly, an LꢀAla carrying congener of 9, compound 16, was prepared by coupling transꢀ
ferulic acid with BocꢀGlyꢀLꢀAlaꢀDꢀGlu(OEt)2 prepared as described [34]. Further, compound
9 was converted to its free acid analog 17 via alkaline hydrolysis (Scheme 2).
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
Scheme 2.a Synthesis of derivatives of compound 9.
aReagents and conditions: (i) transꢀferulic acid, EDC, HOBt, DIPEA, DMAP, DMF; (ii) 1M
NaOH, EtOH, rt.
It has been reported that only minuscule variations of the second amino acid of the essential
L
ꢀAlaꢀ
DꢀGlu mimetics affects the NOD2ꢀagonistic activity, derivatives harboring modified DꢀGlu
moiety were prepared by introducing previously reported constrained mimetics of ꢀglutamic
D
ꢀGlu moiety are allowed.42,43 In order to determine how the incorporation of rigidified
D
acid47 into the structure of compound 9 (Scheme 3). The glycine was first was first esterified
via acidꢀcatalyzed esterification, using SOCl2 in ethanol, to yield the corresponding ethyl ester
1c. The crude ester was then coupled with transꢀferulic acid, using EDC/HOBt methodology,
affording the corresponding Nꢀferuloyl derivative 18. In the succeeding synthetic step, the
ethyl ester was hydrolyzed under alkaline conditions, furnishing the free acid 19 which was
coupled with HCl×LꢀValꢀOEt (1b) to generate Nꢀferuloyl dipeptide 20. On alkaline hydrolysis
of the latter, the intermediate free acid 21 was coupled with the commercially available
amines using the established EDC/HOBt procedure, resulting in the desired rigidified
mimetics of 9, the compounds 22ꢀ24 (Scheme 3). Finally in order to investigate in what
10
ACS Paragon Plus Environment

Page 11 of 67
Journal of Medicinal Chemistry
1
2
3
manner the forced planarity − a direct consequence of the introduced aromatic ring − affects
4
5
NOD2 recognition, Nꢀferuloyl glycine 19 was also linked to aromatic mimetics of the
LꢀValꢀ
6
7
8
DꢀGlu dipeptide 25b, 26b and 27b (prepared as described in Supporting Information)
9
affording compounds 28ꢀ30.
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
Scheme 3.a Exploration of the chemical space of
Dꢀglutamic acid.
aReagents and conditions: (i) transꢀferulic acid, EDC, HOBt, DIPEA, DMAP, DMF; (ii) 1M
NaOH, EtOH, rt; (iii) 1b, EDC, HOBt, DIPEA, DMAP, DMF; (iv) ethyl isonipecotate or
ethyl (1R,2S)ꢀ2ꢀaminocyclohexaneꢀ1ꢀcarboxylate or ethyl (1S,2S)ꢀ2ꢀaminocyclohexaneꢀ1ꢀ
carboxylate, EDC, HOBt, DIPEA, DMAP, DMF; (v) 25b or 26b or 27b, EDC, HOBt,
DIPEA, DMAP, DMF. Constrained mimetics of Dꢀglutamic acid are highlighted in dashed red
rectangles. Core structure of 9 is highlighted in dashed blue rectangles.
11
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 67
1
2
3
2.3 Biological studies
4
5
2.3.1 Evaluation of the cytotoxicity of the synthesized compounds
6
7
8
9
The cytotoxic potential of MDP and the novel desmuramylpeptides was first established by
measuring the proliferation rates of HEKꢀBlue NOD2 cells treated with these compounds,
using the standard metabolic activity assay. Cells were treated for 18 h with MDP and
desmuramylpeptides at 20 µM concentration and their measured metabolic activities
compared with that of the untreated control. The compounds proved to be nonꢀcytotoxic to
HEKꢀBlue NOD2 cells, since none of the residual metabolic activities were less than 80% at
the tested concentration (Fig. S12).
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
2.3.2 Evaluation of NOD2ꢀagonistic activity of the synthesized compounds
The ability of the synthesized desmuramylpeptides to activate NOD2 at the cell level was
investigated employing a validated, commercially available HEKꢀBlue NOD2ꢀdependent NFꢀ
κBꢀinducible SEAP (secreted embryonic alkaline phosphatase) reporter gene assay. NOD2
agonistic activity of a compound is directly related to the elicited NFꢀκB transcriptional
activity and, in turn, the production of SEAP. HEKꢀBlue NOD2 cells were treated with
desmuramylpeptides, and with MDP as the reference compound, for 18 h at 2 or 20 µM. The
results were normalized to that of the untreated control (NT). Since all tested compounds
initially exhibited maximal NOD2 activation at 20 µM (data not shown), the biological
evaluation was repeated at the 10ꢀfold lower concentration of 2 µM (Table S1). Of the 20
synthesized MDPꢀderived compounds, most belong to two parallel series of acyl tripeptides
harboring modifications at the second amino acid and carrying slightly different acyl moieties.
The ability of the synthesized acyl tripeptides 4 to 15 (the LꢀVal and the LꢀPhe series), with
MDP as reference compound, to activate NOD2 at the level of the cell led to several
interesting observations (Fig. 2). The majority of the compounds feature a cinnamoyl
12
ACS Paragon Plus Environment

Page 13 of 67
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
structural motif, previously employed successfully in the preparation of lead II34 and,
surprisingly, in the recently disclosed MDPꢀderived NOD2 antagonists48,49a and dual
NOD1/NOD2 antagonists revealed in a paper by Wang et al.49b The SAR analysis offered
some explanation on how slight structural variations of the cinnamoyl moiety linked to the
glycine amino group affect NOD2 activation, reflected by NFꢀκB transcriptional activity, with
respect to untreated cells (1.0ꢀfold) and MDP (5.6ꢀfold NOD2 activation). It has been
suggested that the 6ꢀphenylindole fragment, as a surrogate of MurNAc, is a pivotal structural
element for NOD2 recognition, as evidenced by our lead I.34 Surprisingly, the 6ꢀphenylindoleꢀ
incorporating compound 4 was in fact a weaker NOD2 activator (3.9ꢀfold) than all the
cinnamoyl moietyꢀbearing analogs. All the tested variations of the substituted cinnamoyl
derivatives were tolerated. Introduction of two fluorine atoms on the phenyl ring of 5 (5.7ꢀ
fold), yielding the 3,4ꢀdifluorosubstituted 7 (5.6ꢀfold), did not enhance the activity of the
parent compound. Analogously, the 3ꢀmethoxyꢀ4ꢀhydroxy substituted derivative 9 proved to
be as strong as the unsubstituted derivative 5, with a 5.6ꢀfold NOD2 activation.
Cyclopropanation of the alkene moieties of 5 and 7 resulted in the corresponding
cyclopropaneꢀcarrying derivatives 6 and 8 with NOD2 agonist activities of 4.4ꢀfold and 5.8ꢀ
fold. The minor drop in NOD2 activation capacity, as compared to that of the parent
compound, is probably a result of the different spatial orientations of the diastereomers
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
composing 6. A similar trend was observed with the LꢀPhe series. The 3,4ꢀdifluorosubstituted
13 (5.4ꢀfold) and the 3ꢀmethoxyꢀ4ꢀhydroxy 15 (5.7ꢀfold) exhibited activities comparable to
that of 11 (5.3ꢀfold), harboring the unsubstituted cinnamoyl moiety. Compounds 12 and 14,
that represent the cyclopropanated congeners of 11 and 13, retained the NOD2 agonistic
activity of the parent compounds, exhibiting 5.2ꢀfold and 5.3ꢀfold activation. Again, the fact
that 12 and 14 are mixtures of diastereomers has to be taken into account. Finally, in
13
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 67
1
2
3
agreement with the
LꢀVal series, the 6ꢀphenylindoleꢀincorporating 10 proved to be a weaker
4
5
NOD2 activator (4.8ꢀfold) than any of the cinnamoyl moietyꢀfeaturing analogs.
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 2. (A) Effects of desmuramylpeptides 4ꢀ15 on NFꢀκB transcriptional activity. (B)
Effects of desmuramylpeptides 16, 17, 22ꢀ24 and 28ꢀ30 on NFꢀκB transcriptional activity.
SEAP activities were measured in both NOD2ꢀspecific and NOD1ꢀspecific HEKꢀBlue cells
after incubation for 18 h with compounds of interest (2 µM), MDP (2 µM) or C12ꢀiEꢀDAP
(100 nM). The data are shown relative to that of the control (NT). Columns represent the
means of duplicates of three independent experiments. Error bars indicate ± S.E.M.
14
ACS Paragon Plus Environment

Page 15 of 67
Journal of Medicinal Chemistry
1
2
3
In an effort to improve the SAR of synthesized desmuramylpeptides and to enable comparison
4
5
with the previously reported compounds, the
LꢀAlaꢀincorporating congener of 9, compound
6
7
8
16, was prepared. Introduction of the ꢀAla moiety still allowed for NOD2 activation,
L
9
although to a lesser extent, as evidenced by the 4.5ꢀfold activation. Previous experiments
suggested that the increased NOD2 activation was a result of facilitated entry of
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
desmuramylpeptide esters into the cytosol.25,34 The influence of modifications of the
DꢀGlu
moiety was therefore further addressed by investigating the effect of removing the diethyl
ester moieties on NOD2 stimulatory activity (Fig. 2b). Replacement of the diethyl ester
groups of 9 by free carboxyl groups gave compound 17 that exhibited significantly
diminished (2.7ꢀfold) NOD2 stimulatory capacity. This observation suggests that the presence
of ester moieties is a critical factor contributing to the NOD2ꢀagonist activity of the
compound.
Replacement of the
additional information concerning the ability of these compounds to activate NOD2. It has
been speculated that both the carboxylate groups of ꢀGlu form important interactions with
Dꢀglutamic acid moiety by its customꢀmade mimetics provided some
D
the receptor.50,51 We thus anticipated that amines carrying only one carboxylate group
(incorporated into 22ꢀ24) would act as poor mimetics of glutamic acid. Nevertheless, a
slightly better activity was expected of compounds 29 and 30 which carry two COOMe
groups at various positions on the aromatic ring. Unfortunately, the mimetic carrying the 4ꢀ
aminoꢀ1,3ꢀdicarboxyl substitution pattern on the phenyl ring could not be prepared due to
difficulties in synthesis. Nevertheless, the results of the activation assay substantiate the
importance of the Dꢀglutamic acid moiety for NOD2 recognition, since its replacements with
rigidified mimetics completely abolished the NOD2 agonistic capacity of the compounds.
15
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 67
1
2
3
4
5
This shows that the
recognition.
DꢀGlu fragment is an important structural characteristic for NOD2
6
7
8
9
The selectivity against NOD1 of all synthesized desmuramylpeptides was also established,
employing commercially available HEKꢀBlue NOD1ꢀdependent NFꢀκBꢀinducible SEAP
reporter gene assay. These cells were incubated with synthesized desmuramylpeptides (20
µM) for 18 h, followed by the SEAP activity determination in the supernatant. The results
which were normalized to that of the control (NT), have shown that none of the
desmuramylpeptides nor MDP induced any NFꢀκB activity relative to the untreated cells (data
shown in Fig. 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
The specificities of all the synthesized compounds on NOD2 were evaluated in parallel, using
HEKꢀBlue NOD2 cells in the presence or absence of GSKꢀ669, a known NOD2 antagonist.52
HEKꢀBlue NOD2 cells were preꢀincubated with NOD2 antagonist (10 µM) and then
stimulated with synthesized desmuramylpeptides or MDP (2 µM) for 18 h, followed by the
SEAP activity determination in the supernatant. The control cells were stimulated only with
synthesized desmuramylpeptides and MDP (2 µM). The results which were normalized to
those of the control cells, indicate that the induced NFꢀκB activity is indeed a consequence of
the NOD2 agonistic activity of desmuramylpeptides and MDP (Fig. S13).
The marginal differences in NOD2 activation observed for most compounds belonging to the
LꢀVal and LꢀPhe series in the singleꢀpoint assay, however, make it impossible to draw any
clear conclusions regarding SAR. For this reason, the doseꢀdependent effect of selected
compounds was investigated. The selection of desmuramylpeptides was based on the results
obtained from the preliminary screening on HEKꢀBlue NOD2 cells at a concentration of 2
16
ACS Paragon Plus Environment

Page 17 of 67
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
ꢁM. At this concentration several of the synthesized desmuramylpeptides exhibited NOD2
agonistic capacities slightly better than, or comparable to, that of MDP, thus proving that all
compounds are full agonists of NOD2. MDP and the selected compounds were then tested
over a range of concentrations in order to guarantee that the activation experiments were
conducted at an effective concentration and to determine their EC50 values.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 3. Concentrationꢀdependent activation of NOD2 by selected desmuramylpeptides and
by MDP. SEAP activity was measured in NOD2ꢀspecific HEKꢀBlue cells after incubation for
18 h with MDP or with the compound of interest at concentrations ranging from 0.001 µM to
20 µM. Data points represent means of duplicates of three independent experiments; error
bars indicate ± S.E.M.
NFꢀκB transcriptional activity was induced using eleven concentrations of compounds and
MDP (ranging from 0.001 ꢁM to 20 ꢁM) allowing for a precise calculation of their EC50
values (Table S2). As expected, the results indicate that increasing concentrations of
desmuramylpeptides and MDP have an impact on the activation of NOD2. All the tested
desmuramylpeptides, as well as MDP (EC50 = 92 nM), induced a doseꢀdependent stimulatory
17
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 67
1
2
3
4
5
6
7
8
effect on NOD2 activation (Fig. 3). Several conclusions can be drawn in terms of structural
requirements for stimulation of the receptor. Comparison of the EC50 values of three
compounds harboring only modifications of the second amino acid residue of the tripeptide
9
moiety, compounds 9 (
= 2.79 ꢁM), revealed an important contribution of the bulkiness and lipophilicity of the amino
acid side chain to NOD2 agonism. The methyl side chain of ꢀAla (16) lacks the ability to
LꢀVal; EC50 = 46 nM), 15 (LꢀPhe; EC50 = 121 nM) and 16 (LꢀAla; EC50
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
L
form strong hydrophobic interactions, in contrast to the isopropyl sidechain (9) which resulted
in approximately 60ꢀfold loss in activity. The benzyl side chain of 15, on the other hand,
possesses the potential to form hydrophobic interactions but is possibly too large to fit
appropriately inside the pocket of the NOD2 receptor. Compound 15 was indeed less active,
presumably due to steric hindrance. This finding suggests that a hydrophobic moiety of the
appropriate size serves to greatly increase NOD2 stimulation activity. In line with these
results, comparison of compound 10 (EC50 = 777 nM), carrying the LꢀPhe moiety, with its Lꢀ
Ala incorporating counterpart lead I (EC50 = 2.88 ꢁM) revealed the importance of the amino
acid side chain. Compounds 5 (EC50 = 223 nM) and 11 (EC50 = 388 nM), incorporating the
cinnamoyl moiety, exhibited potent NOD2 agonist activity, so we examined the chemical
space around the phenyl core. Introduction of two fluorine atoms at positions 3 and 4 resulted
in compounds 7 and 13 (EC50 = 172 nM). The 3,4ꢀdifluorosubstitution slightly increased (2.3ꢀ
fold) the activity relative to its unsubstituted counterpart. The transꢀferuloyl acyl moiety
featured in 9 and 15 proved to be the most advantageous in both series of compounds (
LꢀVal
and ꢀPhe), thus highlighting this fragment as the most effective mimetic of the carbohydrate
L
portion of the MDP known to date. Replacement of the cinnamoyl moiety (compound 11;
EC50 = 388 nM) with its cyclopropaneꢀincorporating bioisostere 12 (EC50 = 472 nM) did not
affect the NOD2 agonistic activity very much. However, it still has to be borne in mind that
12 is, in fact, a mixture of two diastereomers, indicating that one of them could adopt a
18
ACS Paragon Plus Environment

Page 19 of 67
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
conformation with the most favorable interactions with the receptor. Interestingly,
cyclopropanation of the 3,4ꢀdifluoro derivative 13 (EC50 = 172 nM) diminished the capacity
for NOD2 activation, as shown in the case of compound 14 (EC50 = 276 nM). However, in
9
analogy to 8 (EC50 = 247 nM) from the LꢀVal series, 14 is a mixture consisting of four
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
diastereomers, of which only one is likely to bind ideally to NOD2. These compounds were
isolated as mixtures of four diastereomers, as confirmed by HPLC chromatography. The
stereochemistry on the cyclopropane ring presumably has a role in mimicking the preferred
spatial orientation of the appended groups. This makes compounds 14 and 8 interesting for
further investigations to identify the most potent diastereomer. Evidently,
desmuramylpeptides carrying the
LꢀVal moiety are more effective, in terms of NOD2
recognition, than their ꢀPhe counterparts. Finally, conversion of the diethyl ester
L
functionalities of the most potent compound 9 into free carboxylic acid groups resulted in
compound 17 that displays an EC50 value of 6.98 ꢁM, indicating a significant 152ꢀfold drop in
activity.
2.3.3 Determination of stability and cell uptake of compound 9 and its hydrolyzed analog 17
The higher activity of the diester may be related to improved passive uptake of the ligands,
which would result in greater availability of the ligand for the NOD2 receptor. The
incorporation of ester linkages has been shown to result in increased internalization of the
compounds.53 Our preliminary results suggest that diethyl ester derivatives are merely
prodrugs, since introduction of two 1,2,4ꢀoxadiazole moieties as nonꢀhydrolyzable,
bioisosteric replacements of the two diethyl ester moieties54 resulted in complete loss of
NOD2 agonist activity.34 Compounds incorporating ester functionalities are most often
cleaved with carboxylic ester hydrolases, particularly carboxylesterases. Most mammalian
carboxylesterases belong to the CES1 and CES2 families, which are differentiated on the
19
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 67
1
2
3
4
5
6
7
8
9
basis of substrate specificity and tissue distribution. CES1, which preferentially catalyzes
hydrolysis of substrates carrying a small alcohol group and a large acyl moiety, is expressed
predominantly in human liver and monocytes/macrophages, including the THPꢀ1 cell line.55,56
For example, the ethyl ester, once absorbed, is rapidly and predominantly hydrolyzed by the
human carboxylesterase 1 (CES1/hCEꢀ1) in the liver.57,58 In contrast, the CES2 (hCEꢀ2)
isozyme is the dominant component in the small intestine, where it catalyzes primarily the
hydrolysis of compounds carrying a small acyl group and a large alcohol group.59 This raises
important questions about the fate of compound 9, namely, which is the precise active moiety
that is recognized by the receptor and which are the diethyl esters hydrolyzed under the
conditions employed, thus converting 9 into 17. HEK293 cells are not particularly known to
express carboxylesterase enzymes,60,61 so we speculated that the diester moieties of 9 remain
intact, suggesting that the diester is not a prodrug per se, but in fact contributes to NOD2
recognition. Use of a highly sensitive LCꢀMS method enabled us to shed some light on the
fates of compounds 9 and 17 under the conditions employed. First, both compounds, at a
concentration of 2 µM, remained stable in culture medium for at least 18 hours (Fig. S11).
Next, we addressed the stability of compound 9 towards the cell enzymatic system. We
collected the extracellular media from an HEK293ꢀNOD2 cell culture treated by incubation
with 2 µM concentrations of both compounds for 18 h. Cell lysates obtained from the same
experiment were also collected in order to quantify the amount of the tested compound that
moved into the cytosol and, also, to examine the extent of compound 9 intracellular ester
hydrolysis. First, the extracellular concentrations measured indicate that circa 40% of
compound 9 entered the cytosol, while 17 presumably crossed the membrane to a much lesser
extent. The lysate concentrations, which were found to be 0.13 (compound 9) and 0.40 µM
(compound 17), further confirming that both species enter the cytosol (Fig. 4). In light of the
strikingly high lysate concentration of compound 17, it should be noted that the determined
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
ACS Paragon Plus Environment

Page 21 of 67
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
lysate concentrations are not as accurate as the determined extracellular concentrations.
Further, compound 9 has proven to be susceptible to hydrolysis by the cells’ enzyme
machinery, shown by approximately 65% of internalized compound 9 being converted to
compound 17 after an 18 h incubation (Fig. 4), thus actually supporting the hypothesis that the
diester moieties predominantly allow for successful internalization and most likely contribute
to binding to NOD2 only to a minor extent.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 4. The concentrations of compounds 9 and 17 in cell supernatant (CS) and cell lysate
(LYS) determined in the cell uptake and stability experiment performed by 18 h incubations
in the presence of 2 µM compound 9 (diagram A) and 2 µM compound 17 (diagram B) in cell
incubation media. Dark grey bars represent compound 9 and light grey bars compound 17.
Columns represent means of two independent experiments performed in duplicates. Error bars
represent S.E.M.
2.3.4 Docking studies of compounds 9, 17 and MDP into the crystal structure of NOD2
During the course of our study, the crystal structure of rabbit NOD2 was determined in the
inactive ADPꢀbound state. A hydrophobic pocket on the concave surface of the LRR domain
has been proposed as a potential MDPꢀbinding site.51 Recent experiments involving docking
MDP to a homology model of zebrafish NOD2ꢀLRR have shown that the MurNAc moiety is
responsible for holding the ligand in its correct orientation, suggesting that MDP requires a
21
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 67
1
2
3
hydrophobic pocket for its MurNAc moiety and a positively charged pocket for its LꢀAlaꢀDꢀ
4
5
6
7
8
9
iGln part.62 Further, using surface plasmon resonance methodology, Lauro et al. recently
established that the molecular recognition of MDP indeed occurs in the LRR domain and they
defined the critical residues responsible for NOD2 recognition. In addition, they showed that
both the dipeptide and the carbohydrate moiety contribute to the binding.50 The putative
binding residues for the NOD2ꢀMDP interaction, namely Arg857, Trp911, and Ser913, were
highlighted in all models and are thus presumed to be critical for binding. Aromatic residues
such as Trp911 are markedly involved in carbohydrateꢀbinding sites, in line with recent
findings showing that aromatic rings of the receptor often contribute to carbohydrate
recognition in spite of the highly hydrophilic characters of most sugars.63,64 Compound 9,
incorporating the transꢀferuloyl mimetic of MurNAc, displayed the most potent NOD2
agonistic activity, suggesting that the 4ꢀhydroxyꢀ3ꢀmethoxycinnamoyl fragment is a pivotal
structural feature for NOD2 recognition. However, stability and cell uptake studies suggest
that its carboxylic acid derivative 17, which is the result of hydrolysis of compound 9 after
uptake, could be the active compound. Therefore, a proposed model for the orientation of
MDP and compounds 9 and 17 in the putative binding pocket was developed using the crystal
structure of rabbit NOD2 (PDB entry: 5IRN) and GOLD software. The recognition is
predicted to take place in the center of the LRR horseshoe fold. This model shows that
Arg857 makes hydrogen bonding interactions with both the carbohydrate portion of MDP and
the dipeptide. (Fig. 5a), as observed previously.50 However, there are certain differences in the
predicted MDP binding modes to NOD2, which can be attributed to flexibility of the proposed
binding pocket. This was partially addressed by making selected sidechains flexible in the
study of Lauro et al, while in our docking protocol, the protein was kept rigid. Nevertheless,
the MDP sugar moiety was found to occupy the same hydrophobic pocket and the dipeptide
was predicted to be in contact with the Arg857 side chain in both studies. The stability of 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
22
ACS Paragon Plus Environment

Page 23 of 67
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
NOD2ꢀMDP complex was evaluated through molecular dynamics simulation in which the
orientation of MDP slightly changed, but it still remained in contact with relevant residues
(i.e. Arg857 and Trp911, Fig. S1).
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
a)
b)
Figure 7. Predicted binding modes of a) MDP (magenta sticks) and b) 17 (green sticks) in the
putative binding pocket on the concave surface of the LRR domain of NOD2 (PDB entry:
5IRN). The protein is shown in surface presentation and, for clarity, only amino acid side
chains interacting with the ligands are shown as grey sticks. Hydrogen bonds are shown as
black dashed lines.
23
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 67
1
2
3
4
5
6
7
8
9
Docking of compound 17 suggests its similar orientation to that of MDP (Fig. 5b) in the
putative binding pocket on the concave surface of the LRR domain of NOD2. The 4ꢀhydroxyꢀ
3ꢀmethoxycinnamoyl moiety of 17 was suggested to mimic the sugar moiety of MDP, since it
binds to the same hydrophobic pocket, where it forms πꢀπ stacking with Trp911 side chain
and cationꢀπ interaction with Arg857, while its methoxy group interacts with Ser913 side
chain. Additional interaction is predicted between the hydroxy group of the 4ꢀhydroxyꢀ3ꢀ
methoxycinnamoyl moiety and backbone NH of Trp887. Tripeptide moiety of 17 forms
several hydrogen bonds with NOD2 residues. In detail, Gly carbonyl oxygen of 17 is
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
hydrogen bonded to Arg803 side chain, while αꢀcarboxylate of the DꢀGlu of 17 interacts with
the Tyr801 side chain and γꢀcarboxylate with the Lys827 and Arg857 side chains. The
stability of the proposed binding mode of 17 was tested with molecular dynamics simulation,
in which only slight reorientation of the tripeptide portion of 17 was observed (Fig. S2).
Compound 9 was predicted to bind in two different orientations to NOD2 – the first similar to
that of compound 17 and the second in a different orientation in relation to MDP (Fig. S3).
However, when NOD2 in complex with 9 was studied by molecular dynamics simulations
neither of these poses proved stable (Figs. S4 and S5), which is in agreement with the docking
studies that predicted more extensive hydrogen bonding network of the carboxylate 17 than its
ester derivative 9 (Figs. 5b and S3).
Taken together, our docking model thus suggests good complementarity of MDP and
compound 17 and the binding pocket in the NOD2 LRR domain. Additional structural
insights have thus been provided into the pharmacophores required for NOD2 activation.
24
ACS Paragon Plus Environment

Page 25 of 67
Journal of Medicinal Chemistry
1
2
3
2.3.5 Immunostimulatory properties of desmuramylpeptide 9 and MDP on PBMCs
4
5
6
7
8
9
The immunostimulatory activity of the best compound 9 was investigated on human
peripheral blood mononuclear cells (PBMC) using the Human Th1/Th2/Th17 cytokine kit, in
order to extend the analysis beyond reporter gene assays (Fig. 6). Desmuramylpeptide 9 and
MDP were evaluated for their ability to modulate proꢀ and antiꢀinflammatory cytokine
production in PBMCs, alone or in combination with LPS. PBMCs constitute a physiologically
relevant model system for investigating the immunomodulatory properties of a broad range of
innate immune stimulants.65 The immunostimulatory capacity of NOD2 agonists is grounded
mainly on stimulation of monocytes/macrophages. Similarly, LPS acts as an effective
stimulus for the induction of Th1ꢀtype cytokines and mainly stimulates monocytes.66
Unstimulated PBMCs produce few or no cytokines spontaneously and, as expected,
stimulation of PBMCs with MDP and 9 alone (2 ꢁM concentration) produced modest
amounts of cytokines. Nevertheless, the release of TNFꢀα, ILꢀ6 and ILꢀ10 was increased
significantly by both compounds (compared to untreated cells).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 6. Effect of MDP and compound 9 on cytokine release in PBMCs. Cells were treated
with 2 ꢁM MDP or 9 in the presence or absence of 10 ng/mL LPS. Cytokine release was
determined after 18 h. Columns represent means of duplicates ± S.E.M. of two independent
experiments (data are representative of two independent donors); * p < 0.05 vs untreated (NT)
or LPSꢀtreated cells.
25
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 67
1
2
3
4
5
6
7
8
9
On the other hand, stimulation with LPS brought about a considerable increase in cytokine
release. This effect was evidently potentiated by the presence of MDP or compound 9 for all
cytokines tested. The LPSꢀevoked release of the antiꢀinflammatory cytokine ILꢀ10 was
significantly amplified 5ꢀfold by MDP or compound 9. This effect was also observed in the
case of the release of proinflammatory cytokines. Namely, the TNFꢀα release was increased
3ꢀfold by compound 9 and 3.6ꢀfold by MDP, while ILꢀ6 release was increased 4ꢀfold by both
compounds.
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 results are in agreement with those in studies that identified a powerful amplification
effect of MDP on TLR4ꢀinduced cytokine production from isolated PBMCs while, on its
own, inducing only low levels cytokines.14,67,68 The capacity of MDP to synergize with LPS
and other TLR ligands, resulting in an amplified proinflammatory cytokine response, is well
documented.69 Such a synergy can be of great physiological importance since, during
bacterial infection, MDP and LPS are likely to be present in the same tissues.
2.3.6 Adjuvant properties of liposomeꢀencapsulated 9 and MDP in OVAꢀinduced antibody
response
An evident correlation between NOD2ꢀstimulating capacities as well as adjuvant and
antimicrobial properties of MDP derivatives has been shown,42 so desmuramylpeptide 9 was
investigated further to evaluate its usefulness as an adjuvant. First, we established that 9 was
devoid of any pyrogenic activity due to potential endotoxin contamination using a standard
bacterial endotoxin test (data not shown).
26
ACS Paragon Plus Environment

Page 27 of 67
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 7. Comparison of the adjuvant effects of MDP and 9 on the production of total antiꢀ
OVA IgGs in the sera of three experimental groups of NIH/OlaHsd female mice (n=6) after
the second booster. 1. OVA in liposomes (50 ꢁg of OVA in sterile PBS), 2. OVA+MDP in
liposomes (50 ꢁg of OVA with 200 ꢁg/mouse of MDP); 3. OVA+9 in liposomes (50 ꢁg of
OVA with 200 ꢁg/mouse of 9). ○ Denotes each mouse separately, ––– denotes group mean
value; * p<0.05 (Kruskal–Wallis ANOVA, followed by multiple Mann–Whitney Uꢀ
nonparametric tests) in comparison to the group not treated by adjuvant or in comparison with
the group connected with dashed lines.
NOD2 induces a predominant Th2 polarized response, while a combination of PRR ligands
such as NLR/TLR evokes a Th1 and Th17 response. Also, Tukhvatulin et al showed that, by
employing a multiꢀPRR activation approach as part of a standard alumꢀbased vaccine
formulation, activation of both cellular and humoral immunity was significantly increased. In
fact, it proved to be superior to alum or a combination of alum and an individual PRR
ligand.70 The adjuvanticity of MDP was also shown to depend on the mode of administration;
thus, MDP evoked a potent response when delivered with lipophilic carrier systems, such as
liposomes.71 Liposomes have long been used as carriers of antigens and adjuvants to enhance
27
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 67
1
2
3
4
5
6
7
8
9
the immune response.72 The aim of this experiment was to evaluate liposomeꢀencapsulated
MDP and 9 for their immunoadjuvant activity in a well defined mouse model. Specifically,
their effect on the induction of a humoral immune response specific for protein antigen
ovalbumin (OVA) as well as their capacity to skew the immune response towards Th1/Th2
were investigated. The group treated with OVA alone acted as a negative control. Both MDP
and 9 increased antiꢀOVA IgG levels in sera relative to those in the control (Fig. 7).
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
Table 1. The ratio of antiꢀOVA IgG1 and antiꢀOVA IgG2a levels following a second booster
Experimental group
log10 (antiꢀOVA IgG1/antiꢀOVA IgG2a)
1.69 ± 0.28
1. OVA in liposomes
2. (OVA + MDP) in liposomes
3. (OVA + 9) in liposomes
2.37 ± 0.39*
2.62 ±0.55*
The IgG1/IgG2a ratio was calculated after the second booster for each mouse serum. The results for each
experimental group (n=6; females) are reported as the means± SD; *p ≤ 0.01 in comparison to the first group.
It is well known that vaccine adjuvants can modulate the Th1/Th2ꢀbias of an induced immune
response. In our study, the nature of the immune response produced was estimated indirectly
by quantification of antigenꢀspecific IgG1 (Th2ꢀassociated isotype) and IgG2a (Th1ꢀ
associated isotype) and calculation of their ratio (Table 1). The immune response specific for
OVA was Th2ꢀbiased with an IgG1/IgG2a ratio of 44.3 (log IgG1/IgG2a = 1.69) as
demonstrated by Habjanec,73,74 while the adjuvanticity of MDP depends upon the
administration context. Administration in saline increased the humoral immunity, while its
incorporation into liposomes resulted in cellꢀmediated immunity.75 In addition, hydrophilic
MDP derivatives are known to induce a Th2ꢀtype response as opposed to lipophilic
derivatives which augment the Thꢀ1 type immune response. Of note, it was shown that coꢀ
28
ACS Paragon Plus Environment

Page 29 of 67
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
administration of MDP and alum skews the immune response towards the Th1 type. IgG1 was
the predominant IgG subclass for all groups while only MDP, albeit weakly, induced an IgG2
response, resulting in an IgG1/IgG2a ratio of 133 (log IgG1/IgG2a = 2.37). In contrast, 9
displayed an IgG1/IgG2a ratio of 258 (log IgG1/IgG2a = 2.62) indicating a slight shift of the
immune reaction towards the Th2 type of response. The result of this preliminary experiment
sheds some light on the immunoadjuvant activity of both MDP and 9. Compound 9 could
function as an adjuvant, as evidenced by the OVAꢀspecific serum IgG1 antibodies detected
after a primeꢀboost immunization regimen with OVA+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
CONCLUSIONS
A series of fourthꢀgeneration desmuramylpeptides have been designed and prepared. Diverse
variations were incorporated into the parent molecules providing further insights into the
structureꢀactivity relationship. The obtained results significantly increase our understanding of
the structural requirements of desmuramylpeptides for NOD2ꢀactivation. In particular,
compound 9 was identified as the, so far, most potent NOD2 agonist of the
desmuramylpeptide structural class. Compound 9 also possesses the ability to augment LPSꢀ
induced proꢀinflammatory cytokine release from PBMCs as well as to increase OVAꢀspecific
IgG titers in a mouse model of adjuvancy. Together, our results indicate that 9 can induce
both innate and adaptive immune responses and thus demonstrate the potential utility of
NOD2 agonists as adjuvants for vaccines.
29
ACS Paragon Plus Environment