Probing the Molecular Interactions between CXC Chemokine Receptor 4 (CXCR4) and an Arginine-Based Tripeptidomimetic Antagonist (KRH-1636)

Article abstract of DOI:10.1021/acs.jmedchem.5b00987

We here report an experimentally verified binding mode for the known tripeptidomimetic CXCR4 antagonist KRH-1636 (1). A limited SAR study based on the three functionalities of 1 was first conducted, followed by site-directed mutagenesis studies. The recep

Full text of DOI:10.1021/acs.jmedchem.5b00987

Subscriber access provided by NEW YORK MED COLL
Article
Probing the Molecular Interactions Between CXC Chemokine Receptor 4
(CXCR4) and an Arginine-Based Tripeptidomimetic Antagonist (KRH-1636)
Zack Georg Zachariassen, Stefanie Karlshøj, Bengt Erik Haug, Mette M. Rosenkilde, and Jon Våbenø
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.5b00987 • Publication Date (Web): 23 Sep 2015
Downloaded from http://pubs.acs.org on September 29, 2015
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 45
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Probing the Molecular Interactions Between CXC
Chemokine Receptor 4 (CXCR4) and an Arginineꢀ
Based Tripeptidomimetic Antagonist (KRHꢀ1636)
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
Zack G. Zachariassen,† Stefanie Karlshøj,§ Bengt Erik Haug,‡ Mette M. Rosenkilde,§ and Jon
Våbenø*,†
†Department of Pharmacy, Faculty of Health Sciences, UiT The Arctic University of Norway,
Breivika, NOꢀ9037 Tromsø, Norway. §Laboratory for Molecular Pharmacology, Department of
Neuroscience and Pharmacology, Faculty of Health and Medical Sciences, The Panum Institute,
University of Copenhagen, Blegdamsvej 3, DKꢀ2200 Copenhagen, Denmark. ‡Department of
Chemistry and Centre for Pharmacy, University of Bergen, Allégaten 41, NOꢀ5007 Bergen,
Norway.
ACS Paragon Plus Environment
1

Journal of Medicinal Chemistry
Page 2 of 45
1
2
3
4
5
6
7
8
9
Abstract: We here report an experimentally verified binding mode for the known
tripeptidomimetic CXCR4 antagonist KRHꢀ1636 (1). A limited SAR study was first conducted
based on the three functionalities of 1, followed by siteꢀdirected mutagenesis studies. The
receptor mapping showed that both the potency and affinity of 1 were dependent on the
transmembrane residues His113, Asp171, Asp262, and His281, and also suggested the involvement of
Tyr45 and Gln200 (potency) and Tyr116 and Glu288 (affinity). Molecular docking of 1 to an Xꢀray
structure of CXCR4 showed that the LꢀArg guanidino group of 1 forms polar interactions with
His113 and Asp171, the (pyridinꢀ2ꢀylmethyl)amino moiety is anchored by Asp262 and His281, while
the naphthalene ring is tightly packed in a hydrophobic subpocket formed by the aromatic side
chains of Trp94, Tyr45, and Tyr116. The detailed picture of ligandꢀreceptor interactions provided
here will assist in structureꢀbased design and further development of smallꢀmolecule
peptidomimetic CXCR4 antagonists.
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
INTRODUCTION
Chemokines are small (8ꢀ14 kDa) proteins from the cytokine superfamily that exert their
biological effects through interaction with GPCRs.1 The involvement and physiological role of
the chemokine ligandꢀreceptor system extends from control of chemotaxis (directional migration
of leukocytes)2 and hematopoiesis3 to embryonic development, angiogenesis and cancer
metastasis.4, 5 The human CXC chemokine receptor 4 (CXCR4) is activated by CXC chemokine
ligand 12 (CXCL12), which is a 68ꢀresidue polypeptide.6, 7 The CXCL12:CXCR4 interaction is
involved in development of the hematopoietic, nervous, and gastrointestinal systems8, 9 and plays
an essential role in immune defense and inflammation.10, 11 In addition to these physiological
roles, CXCR4 is the coꢀreceptor for cellular entry of Tꢀtropic HIV strains;12, 13 thus, inhibition of
ACS Paragon Plus Environment
2

Page 3 of 45
Journal of Medicinal Chemistry
1
2
3
4
CXCR4 has emerged as a promising strategy for antiꢀHIV therapy. CXCR4 has also been shown
5
6
4
to play an important role in angiogenesis, metastasis and stemꢀcell mobilization.2,
7
8
9
Consequently, development of CXCR4 antagonists for therapeutic applications remains an
important goal.
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
Significant advances in the development of novel CXCR4 antagonists have been made and
several different classes of CXCR4 antagonists, both peptides and nonꢀpeptides, have been
reported.14ꢀ16 So far, the hematopoietic stem cell mobilizing agent plerixafor (AMD3100, Figure
1A)17 is the only CXCR4 antagonist that has reached the market.18 Due to its high positive
charge at physiological pH, AMD3100 exhibits poor oral bioavailability, and is therefore
administered by subcutaneous injection.
Extensive studies by Fujii and coꢀworkers on polyphemusin IIꢀderived peptide CXCR4
antagonists culminated in the discovery of the potent cyclopentapeptide antagonist FC131
(Figure 1B).19 Successive SAR studies of FC131 have since shown that the Arg1ꢀArg2ꢀ2ꢀNal3
tripeptide fragment of the cyclopentapeptide serves as an essential recognition motif for peptideꢀ
based CXCR4 antagonists,20ꢀ24 providing impetus for development of tripeptidomimetic ligands
that block CXCR4. Indeed, a series of tripeptideꢀlike CXCR4 antagonists containing a central
arginine residue has been reported by Kureha Chemical Industries25ꢀ27 with the potent
29
tripeptidomimetic 1 (KRHꢀ1636, Figure 1C)28,
as a prototype compound. These smallꢀ
molecule peptidomimetics represent a particularly interesting class of CXCR4 antagonists due to
their potentially improved pharmacokinetic properties compared to their peptide counterparts.30,
31
Based on structural comparison of lowꢀenergy conformations of 1 with a 3Dꢀpharmacophore
model for the cyclopentapeptide CXCR4 antagonists, 1 has been suggested to mimic the Arg1ꢀ
ACS Paragon Plus Environment
3

Journal of Medicinal Chemistry
Page 4 of 45
1
2
Arg2ꢀ2ꢀNal3 fragment of FC131.32 However, it has not been shown experimentally that the two
ligands bind to CXCR4 in the same way.
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
B
A
Gly4
NH HN
NH HN
O
O
OH
NH HN
D
-Tyr5
O
N
HN
H
H
NH
N
2-Nal3
N
N
NH2
NH
O
NH HN
O
Arg1
AMD3100
Arg2
HN
C
H2N
NH
FC131
O
N
H
H
N
N
N
D
F
H
Arg1-Arg2-(1-Nal)3-Cys4-Tyr5-Gln6-Lys7- -Pro8
D
O
D
-Pro16-Gly15-Arg14-Cys13-Cit12-Arg11-Tyr10-Pro9
HN
2 (CVX15)
HN
NH2
1 (KRH-1636)
E
H
N
N
N
N
S
N
NH2
N
NH
S
N
3 (IT1t)
4 (AMD11070)
Figure 1. Structures of selected CXCR4 antagonists: (A) bicyclam AMD3100, (B)
cyclopentapeptide FC131, (C) prototype arginineꢀbased tripeptidomimetic 1, (D) 16ꢀmer peptide
2, (E) isothiourea derivative 3, and (F) tetrahydroquinoline derivative 4.
The first Xꢀray structures of CXCR4 were reported in 2010, and are coꢀcrystallized complexes
containing the 16ꢀmer peptide antagonist 2 (CVX15, Figure 1D) and the smallꢀmolecule
antagonist 3 (IT1t, Figure 1E).33 Very recently, an Xꢀray structure of a covalent complex
between CXCR4 and the viral chemokine vMIPꢀII was also released.34 The CXCR4:2 complex
provided the first detailed insight into binding of a peptide antagonist to CXCR4, and based on
this structure we have recently reported an experimentally verified binding mode for the
ACS Paragon Plus Environment
4

Page 5 of 45
Journal of Medicinal Chemistry
1
2
cyclopentapeptide FC131 (Figure 1B).35 Prior to the publication of the Xꢀray structures of
CXCR4, extensive receptor mapping studies were performed for the AMDꢀcompound series,
which resulted in the identification of plausible binding modes for the bicyclam AMD3100,36ꢀ38
its monocyclam analogs,39, 40 and the nonꢀcyclam 4 (AMD11070, Figure 1F).40 However, while
the molecular pharmacology of the prototype smallꢀmolecule peptide and nonꢀpeptide CXCR4
antagonists has been extensively characterized, experimental binding mode studies for the more
drugꢀlike tripeptidomimetic CXCR4 antagonists have not been reported. Thus, in order to
facilitate rational design of novel peptidomimetic CXCR4 antagonists, we here report the
binding mode of the known tripeptidomimetic CXCR4 antagonist 1 using a combination of SAR
studies, receptor mapping by siteꢀdirected mutagenesis, and molecular docking.
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
RESULTS AND DISCUSSION
SAR Study. Design. The lead compound 1 was developed from a hit compound identified
through screening of a chemical library,28, 29 and although recent patent literature discloses
compounds related to 1,25ꢀ27 no SAR data has been reported for this compound class. In order to
confirm the importance of the functionalities in 1 that could be assumed to represent
pharmacophoric groups, we first carried out a limited SAR study focusing on the two positively
charged groups (R1 and R2) and the aromatic group (R3) (Figure 2). In order to probe the R1
position, the pyridine ring was replaced with a phenyl ring (5), the pyridinꢀ2ꢀylmethyl substituent
was removed to give the primary amine (6), and the charged (pyridinꢀ2ꢀylmethyl)amino moiety
was removed altogether (7). The importance of the charged guanidino group (R2) was
investigated by replacing it with a urea group (8), while the aromatic position (R3) was probed by
removal of the naphthalene unit (9).
ACS Paragon Plus Environment
5

Journal of Medicinal Chemistry
Page 6 of 45
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 2. Structures of compounds 1 and 5–9; the modified groups (R1, R2, and R3) in the lead
compound 1 are highlighted.
Chemistry. Compounds 1 and 5–9 (Figure 2) were all prepared by adopting a procedure
previously reported for the synthesis of 1 (Scheme 1).29 For compound 1, the rightꢀhand side was
prepared by Nꢀalkylation of Bocꢀprotected pyridinꢀ2ꢀylmethanamine (10) with methyl 4ꢀ
(bromomethyl)benzoate followed by hydrolysis of the methyl ester to give carboxylic acid 11.
The leftꢀhand side was assembled by coupling of FmocꢀLꢀArg(Pbf)ꢀOH with (S)ꢀ1ꢀ(naphthalenꢀ
1ꢀyl)ethanamine using PyBOP as the coupling reagent, followed by Fmocꢀdeprotection with
diethylamine in DMF to afford amine 12. Coupling of the two halves (11 and 12) was facilitated
by EDCI/HOBt, and subsequent removal of the Boc and Pbf protecting groups by treatment with
a TFA cocktail gave 1. For the preparation of 6 and 7, the left hand side (12) was coupled with
NꢀBoc 4ꢀ(aminomethyl)benzoic acid and 4ꢀmethylbenzoic acid, respectively. Analog 5 was
prepared by Nꢀalkylation of 6 with benzyl bromide using K2CO3 as base and DMF as solvent.
ACS Paragon Plus Environment
6

Page 7 of 45
Journal of Medicinal Chemistry
1
2
The R2 analog 8 was prepared using Fmocꢀprotected citrulline as the starting material while the
R3 analog 9 was prepared by use of ethanamine in the first coupling step (Scheme 1).
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
Scheme 1. Synthesis of compound 1.
Reagents and conditions: (a) i) NaH, DMF, 0 °C; ii) Methyl 4ꢀ(bromomethyl)benzoate, DMF,
0 °C to rt; (b) 1M NaOH/THF/MeOH (1:1:1); (c) (S)ꢀ1ꢀ(naphthalenꢀ1ꢀyl)ethanamine, PyBOP,
DIPEA, CH2Cl2 0 °C to rt.; (d) Et2HN, DMF; (e) 11, EDCI, HOBt, DMF; (f) TFA/TIS/H2O
(95:2.5:2.5).
Biology and SAR. The antagonistic potency of compounds 1 and 5–9 (Table 1) was assessed in
a functional assay measuring the inhibition of CXCL12ꢀinduced activation of human CXCR4,
transiently expressed in COSꢀ7 cells. The known lead compound 1 was originally reported by
Ichiyama et al. to efficiently inhibit binding of [125I]CXCL12 (IC50 = 0.013
functional data validates 1 as a potent CXCR4 antagonist (EC50 = 0.50 M, Table 1) with respect
to the endogenous ligand CXCL12. The phenylꢀanalog 5 (EC50= 8.5 M) displayed 17ꢀfold
µ
M),28 and our own
µ
µ
lower potency than 1, demonstrating an important role of the pyridine nitrogen in 1. The potency
ACS Paragon Plus Environment
7

Journal of Medicinal Chemistry
Page 8 of 45
1
2
3
4
of the primary amine 6 (EC50= 12 µM) was 24ꢀfold lower than for 1, representing the overall
5
contribution from the pyridinꢀ2ꢀylmethyl substituent to the antagonistic potency, while the loss
M) confirms the positively charged aminoꢀgroup in R1
6
7
8
9
of potency for compound 7 (EC50 > 100
µ
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
as an essential pharmacophoric element. Compound 8, containing a neutral urea group that partly
retains the Hꢀbonding properties of the original guanidino group, failed to show antagonistic
potency (EC50 > 100 µ
M), demonstrating the importance of the positive charge in R2 position.
Similarly, the loss of potency for the R3 analog 9 (EC50 > 100
µM) confirms an important role
for the naphthyl group.
Table 1. Antagonistic potency of 1 and 5–9 on human CXCR4.
compd log EC50 ± SEMa
EC50 (ꢁM)
0.50
1b
5
ꢀ6.30 ± 0.07
ꢀ5.12 ± 0.08
ꢀ4.93 ± 0.08
> ꢀ4
8.5
12
6
> 100
> 100
> 100
7
> ꢀ4
8
> ꢀ4
9
aEC50 values represent the mean of at least three independent experiments performed in
duplicates. bKnown compound.
Earlier attempts to improve the pharmacokinetic profile of the bicyclam CXCR4 antagonists,
i.e. AMD3100 (Figure 1A), led to the discovery of the monocyclam AMD3465 (Figure 3),41
which was shown to be 22ꢀfold more potent than its precursor AMD310039 and contains the
same (pyridinꢀ2ꢀylmethyl)amino moiety as 1. The importance of this group for the antagonistic
ACS Paragon Plus Environment
8

Page 9 of 45
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
activity of AMD3465 has been investigated previously using the analogs 13 (AMD3529) and 14
(AMD3389) (Figure 3),39 which are structural counterparts of the KRHꢀanalogs 5 and 7 in the
present study (Figure 2). In line with the present findings for 5 and 7 (Table 1), 13 and 14 were
shown to display significantly lower potencies for CXCR4 than AMD3465 as determined by a
functional assay inhibiting CXCL12ꢀinduced receptor activation.39
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. Structures and reported potencies of the monocyclam AMD3465 and the two analogs
13 and 14,39 also showing the structural analogy of these monocyclams with the R1 substituent in
compounds 1, 5, and 7 in the present study.
Regarding the central Arg residue in 1 (R2), it has been suggested that this position could
correlate to Arg2 in the cyclopentapeptide antagonist FC131 (Figure 1).32 SAR studies of the
cyclopentapeptides have shown that the Arg2 position is very sensitive to modifications,20, 23 and
that replacement of the guanidino group with a urea group (Cit2) leads to complete loss of
antagonistic potency,23 which is in agreement with the present data for the KRHꢀanalog 8.
Collectively, the findings from this initial SAR study confirmed that the secondary amine, the
guanidino group, and the aromatic naphthyl group are essential for the antagonistic potency of 1,
while the pyridine ring in R1 position contributes to increased potency.
Identification of Binding Site for 1: Receptor Mutagenesis. Next, we carried out
mutagenesis studies to map key binding interactions between 1 and CXCR4 using a library of 25
ACS Paragon Plus Environment
9

Journal of Medicinal Chemistry
Page 10 of 45
1
2
3
4
5
6
7
8
9
CXCR4ꢀmutants (Table 2). Most of the mutations were located in the transmembrane helices
(TMs), but some were also introduced in extracellular loop 2 (ECL2) that connects TM4 and
TM5. In most cases, the mutations included replacement with Ala; however, certain acidic Asp
residues were alternatively substituted with the structurally similar and uncharged Asn residue.
Selective substitutions with Trp were also done in order to introduce more bulk. We have
recently utilized the same receptor library to map the binding of the cyclopentapeptide CXCR4
antagonist FC131, where we reported the surface expression (using ELISA) and functional
response to CXCL12 of WTꢀCXCR4 and all mutant receptors35 (the same data is provided in
Table 2). While there was some variation in surface expression and ability to be activated by
CXCL12, the receptors in the mutant library were considered suitable for mapping of the binding
site, as previously discussed.35
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
Inhibition of CXCL12ꢀinduced Activation of CXCR4. We first assessed the antagonistic
potency of 1 on the entire CXCR4 mutant library (Table 2). The H113A, D262N, and H281A
mutations were found to have the largest impact (>25ꢀfold reduction) on the antagonistic potency
of 1 relative to WTꢀCXCR4 (the position of residues based on the generic numbering system
proposed by Baldwin42 and modified by Schwartz43 followed by the Ballesteros/Weinstein
numbering system44 is given in Tables 2 and 3). The potency of 1 was also reduced (by 5ꢀ25
fold) by mutations of Tyr45 (Y45A), Asp171 (D171N), Gln200 (Q200A and Q200W), Trp252
(W252A), Ile259 (I259A), and Ile284 (I284A). In contrast, the W94A, D97A, T117A, and F189A
mutants led to increased potency (>2ꢀfold).
As the mutagenesis data for 1 alone did not provide information about which parts of the
ligand that interact with the different parts of the receptor, the R1 analogs 5 and 6 (Figure 2),
which showed decent activity on WTꢀCXCR4 (Table 1), were also included in the mutagenesis
ACS Paragon Plus Environment
10

Page 11 of 45
Journal of Medicinal Chemistry
1
2
study to further probe the molecular interactions of the R1ꢀside chain with CXCR4. The
collective functional data analysis for 1, 5, and 6 (Table 2) showed that the mutations in TM1
(Y45A), TM2 (W94A, D97A), TM3 (H113A, T117A), TM4 (D171N), and ECL2 affected the
three ligands in a similar manner. However, several mutations in TMs 5ꢀ7 affected the three
ligands in a distinctive manner. Specifically, the H281A mutation in TM7 was the mutation with
largest effect on the potency of 1 (61ꢀfold reduction), but affected analog 5 to a much lesser
extent (6ꢀfold). Thus, the overall trend identified in the phosphatidylinositol (PI) turnover
experiments for 1, 5, and 6 (Table 2) suggests that the structurally modified R1 group contacts
TMs 5ꢀ7; this trend is further illustrated in Figure 4.
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
The absolute majority of the mutations had no or little effect on the potency of CXCL12 (Table
2), demonstrating that most of the mutants were functional. However, the W94A, D97A, and
D187A mutations reduced the agonistic potency of CXCL12 by 8ꢀ14 fold (Table 2).
Furthermore, the Y116A and E288A mutants were not activated by CXCL12 in spite of
expression levels of 69 and 77%, respectively (Table 2); consequently, no data on the
antagonistic action of 1, 5, and 6 on these mutants are available. In line with the findings in the
present study, the D97A, D187A, and E288A mutations have previously been found to affect
both the binding and signaling of CXCL12.37, 45ꢀ49 Moreover, Trp94 and Tyr116, together with the
three aforementioned residues, have been implicated as part of “site two” in the “twoꢀsite model”
for binding of CXCL12 to CXCR4.33, 45, 50, 51 Additionally, the direct interaction of Tyr116 with
agonists has been suggested in activation of GPCRs.34, 52 Thus, for these residues there are
certain limitations in using CXCL12 as a probe.
ACS Paragon Plus Environment
11

Journal of Medicinal Chemistry
Page 12 of 45
1
2
Table 2. Potency of the lead compound 1 and the R1ꢀanalogs 5 and 6 on WTꢀCXCR4 and 25
mutant receptors, measured as inhibition of CXCL12ꢀinduced activation in COSꢀ7 cells
transiently coꢀtransfected with CXCR4 receptor mutants and chimeric G protein Gqi4myr
(functional assay).
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
b
aData for CXCL12 activation are from ref. 35. The position of each residue is given based on
the generic numbering system proposed by Baldwin and modified by Schwartz, followed by the
Ballesteros/Weinstein numbering system. Fmut is the ratio of the mutant and WTꢀCXCR4
c
potencies. Red: Fmut >25; orange: Fmut from 10ꢀ25; yellow: Fmut from 5ꢀ10; green: Fmut <0.5.
EC50 values >100 ꢁM are underlined and also colored red. P<0.001***, P<0.01**, P<0.05*.
dThe number of independent experiments is shown in parentheses (n). These mutant receptors
e
were also tested in binding assay (Table 3).
ACS Paragon Plus Environment
12

Page 13 of 45
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 4. (A-C) Barꢀgraph representations of foldꢀreduction in potency (Fmut) for compounds 1,
5, and 6 for each mutant with respect to WTꢀCXCR4 (functional data, Table 2). Grey bars:
mutants in TMs 1ꢀ4 and ECL2; black bars: mutants in TMs 5ꢀ7. (D-F) Residues identified to be
important for the potency of compounds 1, 5, and 6 (functional data, Table 2), shown in a helical
ACS Paragon Plus Environment
13

Journal of Medicinal Chemistry
Page 14 of 45
1
2
3
4
5
6
wheel diagram of CXCR4. Red background color: >25 fold decrease; orange: 10ꢀ25 fold
decrease; yellow: 5ꢀ10 fold decrease.
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 5. (A) Ability of the lead compound 1 and analogs 5 and 6 to inhibit CXCL12 induced
activation of WTꢀCXCR4, and (B) to displace 125Iꢀ12G5 from WTꢀCXCR4.
Displacement of Radiolabeled Monoclonal Antibody (125Iꢀ12G5). To address the issue of
reduced activity of CXCL12 in certain mutants (W94A, D97A, Y116A, D187A, and E288A) a
competitive binding assay using the radiolabeled monoclonal antibody 12G5 (125Iꢀ12G5) was
employed to probe ligand affinity (Table 3). A good correlation between antiꢀHIV potency and
binding affinity measured as displacement of 125Iꢀ12G5 has previously been demonstrated for the
38, 53
bicyclam antagonists,36,
which represents an additional advantage of using 12G5 as
radioligand. In addition to the five mutants mentioned above, the binding experiments were
extended to also include four of the mutants that were identified as important for the potency of 1
in the functional assay (H113A, D171N, D262N, and H281A). All nine mutants had similar
affinities for 12G5 as the WT receptor (Table 3), meaning that the mutations did not significantly
affect receptor folding.
ACS Paragon Plus Environment
14

Page 15 of 45
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Compared to the strong antagonistic potency (EC50 = 0.50 ꢁM) measured in the functional
assay (Figure 5A), 1 displayed lower affinity to WTꢀCXCR4 (IC50 = 8.0 ꢁM) measured against
125Iꢀ12G5 (Figure 5B), suggesting that this compound binds allosterically with respect to 12G5.
This finding clearly illustrates the differences between displacing 125Iꢀ12G5 (Table 3) and
inhibiting the action of CXCL12 (Tables 1 and 2). The two R1 analogs 5 and 6 did not
completely displace 12G5 (Figure 5B) – which also indicates an allosteric binding mode – and
were therefore excluded from further binding experiments. The affinity of 1 determined against
125Iꢀ12G5 was strongly (8ꢀ to >12ꢀfold) negatively affected by the TM3 mutants H113A, the
TM4 mutant D171N, the TM5 mutant D262N, and the TM7 mutants H281A and E288A (Table
3). Moreover, 1 could not displace 12G5 from the Y116Aꢀmutant. Overall, the binding data
show that His113, Tyr116, and Asp171 at the interface between TM3 and TM4, and Asp262, His281,
and Glu288 at the interface between TM6 and TM7 in the main binding pocket are important for
the displacement of 12G5 by 1.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
To summarize, the functional studies (Tables 1 and 2) and the binding studies (Table 3) show
that Tyr45 (TM1), His113 and Tyr116 (TM3), Asp171 (TM4), Gln200 (TM5), Asp262 (TM6), and
His281 and Glu288 (TM7) are important for the function and/or binding of 1.
ACS Paragon Plus Environment
15

Journal of Medicinal Chemistry
Page 16 of 45
1
2
3
4
Table 3. Affinity of 1 for WTꢀCXCR4 and nine mutant receptors measured as displacement of
5
6
125Iꢀlabeled antibody 12G5 in transiently transfected COSꢀ7 cells (competition binding assay).
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
b
aData for 12G5 binding are from ref. 35. Fmut is the ratio of the mutant and WTꢀCXCR4
c
d
affinities. The number of independent experiments is shown in parentheses (n). P<0.001***,
P<0.01**, P<0.05*. The position of each residue is given based on the generic numbering
system proposed by Baldwin and modified by Schwartz, followed by the Ballesteros/Weinstein
numbering system.
e
Molecular Docking and Derived Binding Model. Binding Model: Key Interactions. In order
to rationalize the experimental data, 1 was docked to the Xꢀray structure of CXCR4 (PDB code
3OE0)33 using Schrödinger’s inducedꢀfit docking protocol,54 which models the conformational
changes induced by ligand binding. In order to avoid generation of irrelevant poses, a Hꢀbond
constraint was set on His113 (Nδ1), which was shown to be highly important for both the affinity
and potency of 1. Visual inspection of the 15 generated ligandꢀreceptor complexes resulted in the
identification of a pose that was in agreement with the experimental data. In our proposed
binding model (Figure 6), the (pyridinꢀ2ꢀylmethyl)amino moiety (R1 side chain) of 1 is oriented
towards the TM6 and TM7 region, and the R1 side chain adopts a bent conformation around the
ACS Paragon Plus Environment
16

Page 17 of 45
Journal of Medicinal Chemistry
1
2
secondary amino group, where the pyridine and phenyl rings lie in almost parallel planes. His281
in TM7 forms an aromatic πꢀπ stacking interaction with the pyridine ring, and Asp262 in TM6 is
involved in a bidentate interaction with the charged secondary amine in the R1 side chain of 1.
The guanidino group of Arg (R2ꢀposition) forms polar interactions with His113 and Thr117 (TM3)
and Asp171 (TM4). The aromatic naphthyl ring (R3ꢀposition) is embedded in a hydrophobic
pocket formed by the aromatic sideꢀchains of Tyr45 and Trp94, and further stabilized by the Tyr116
sideꢀchain located directly below Trp94. Interestingly, previous studies indicate that the three
aromatic residues Tyr45, Trp94 and Tyr116 of CXCR4, which are also found in the CCR5
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
chemokine receptor as Tyr37, Trp86, and Tyr108, respectively, might form a hydrophobic pocket
56
that is involved in the binding of smallꢀmolecule inhibitors.55,
Thus, mutations of the
aforementioned three residues could destabilize the hydrophobic subpocket that interacts with
the naphthyl ring of 1. Furthermore, the observation that Tyr116 is involved in a direct interaction
with Glu288 (see Supporting Information, Figure S1) implies that the Y116A mutation can affect
the overall geometry of the main binding crevice, specifically the orientations of TM3 and TM7.
Thus, the Y116A mutation may indirectly affect the potential interactions between the ligand and
His113, Thr117, His281, and Glu288.
Binding Model: Indirect Effects. The docking pose for 1 also reveals the involvement of
Glu288, which is a key residue as shown in the binding assay (Table 3). A Glu residue at this
position (VII:06/7.39) is highly conserved among chemokine receptors and acts as an anchor
point for several small molecule ligands.57 Glu288 in CXCR4 has been consistently implicated in
binding modes of several different classes of CXCR4 antagonists, i.e. AMD3100, AMD3465, 4,
FC131, and also the CXCR4 coꢀcrystallized ligands 2 and 3,33, 35, 37, 39, 40 either through direct or
ACS Paragon Plus Environment
17

Journal of Medicinal Chemistry
Page 18 of 45
1
2
3
4
5
6
waterꢀmediated interactions with the ligand. Similarly, our binding model shows a waterꢀ
mediated interaction between Glu288 and the ligand backbone (Figure 6).
7
8
9
Substitution of Gln200 with Ala and Trp (Q200A and Q200W) caused a 7ꢀ and 21ꢀfold
reduction in potency of 1 with respect to WTꢀCXCR4; the involvement of this residue was also
seen in previous binding mode studies of the monocyclam antagonist AMD3465.39 In our
binding model for 1 (Figure 6), an intraꢀreceptor Hꢀbond between Gln200 and Asp262 was
identified, suggesting that substitution of Gln200 can affect the interaction between Asp262 and the
ligand.
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 model further shows an interaction between the guanidino group of Arg188 and the ligand
backbone (Figure 6), which was also previously identified in the binding modes of the peptide
antagonists FC131 and 2.23, 33, 58 However, the R188A mutation did not have any effect on the
antagonistic potency of 1 or the two analogs (Table 2); interestingly, the R188Aꢀmutant was also
reported to have no effect on the potency of FC131.35 A possible explanation is that both 1 and
FC131 are doubly positively charged, which results in electrostatic repulsion by Arg188.
Substitution of Arg188 with Ala (R188A) therefore results in two opposing effects: an
unfavorable removal of the Hꢀbond interaction with the ligand, and a favorable removal of the
electrostatic repulsion. Thus, it can be hypothesized that the net result is no overall effect of the
R188A mutation on 1 and FC131.
ACS Paragon Plus Environment
18

Page 19 of 45
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 6. (A) Effects of the H113A and D171N mutations on the inhibition of CXCL12ꢀinduced
PI turnover by 1. (B) Proposed binding model for 1, showing the key interacting residues in
CXCR4; the ligand is shown in orange sticks, the receptor in greyꢀwhite sticks and ribbons, and
the water molecule as a red sphere; Hꢀbonds are shown in magenta, and πꢀπ interactions in green.
(C) Effects of the H281A and D262N mutations on the inhibition of CXCL12ꢀinduced PI
turnover by 1. (D) 2Dꢀrepresentation of the docking pose for 1.
ACS Paragon Plus Environment
19

Journal of Medicinal Chemistry
Page 20 of 45
1
2
3
4
5
6
7
8
9
While the effect of most mutations (Tables 2 and 3) can be explained by the binding model, it
is not possible to rationalize every single one. For example, the W94A and D97A mutations both
led to increased potencies for all three analogs (Table 2). Earlier studies have suggested the
involvement of these residues in the second binding step (“site two”) for the endogenous ligand
CXCL12,33, 45 and the present mutagenesis data showed a reduced ability (13ꢀ14 fold) of
CXCL12 to induce signaling for the W94A and D97A mutants. Moreover, the Xꢀray structure of
CXCR4 suggests that Asp97 constitutes a gateꢀentry position that controls ligand access to the
major binding pocket;33, 35 thus, a preꢀinteraction with Asp97 is implicated in the passage of the
ligands to the TM receptor region. Taken together, the existing data supported by our
mutagenesis suggest that the TM2 residues (Trp94 and Asp97) have a less distinctive involvement
in the binding of these ligands to the receptor. First, mutation of Trp94 and Asp97 will weaken the
binding of CXCL12, thereby increasing the antagonistic potency of the three analogs as reflected
in Table 2. Second, the potential preꢀinteraction with Asp97 is removed in the D97A mutant,
which allows easier access of the ligands to the TM binding pocket. Third, the involvement of
the Trp94 in the hydrophobic pocket surrounding the naphthalene ring (R3ꢀposition) in 1 (Figure
6), suggests that removal of the bulky side chain of Trp (W94Aꢀmutant) allows a more
energetically favorable position of the naphthyl group in the binding pocket.
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
Conversely, mutation of hydrophobic residues in TM6 and TM7 reduced the potency of 1
(W252A, I259A and I284A) and 6 (W252A and I284A). However, residues Trp252, Ile259, and
Ile284 are not directly involved in ligand binding, meaning that the observed effects of the
mutations most likely are secondary, i.e. implemented through local or global changes in TM6
and TM7.
ACS Paragon Plus Environment
20

Page 21 of 45
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Despite the lower WT potency of 6 (Table 2), the effects of mutants Y45A and D171N were
more pronounced for this analog. Our SAR data showed that gradual removal of the R1 group led
to reduced potency, while the receptor mapping (Table 2) suggested that the R1 group contacts
TMs 5ꢀ7. It could be reasoned that the weakened interaction of 6 with TMs 5ꢀ7 (R1 group
interaction) makes the remaining interactions more important, which would explain the stronger
dependence on Tyr45 (R3 group interaction) and Asp171 (R2 group interaction) for 6 (Table 2).
Comparison with Experimentally Determined Binding Modes for Other Small-molecule
CXCR4 Antagonists. Isothiourea Derivatives. Gebhard et al. have reported a series of smallꢀ
molecule isothiourea derivatives that selectively block CXCL12 and HIV gp120 binding to
CXCR4.59, 60 The Xꢀray structures of CXCR4 in complex with the most potent compound 3
(Figure 1E)33 later showed that 3 exhibits a unique binding mode by occupying the soꢀcalled
“minor binding pocket” between TMs 1, 2, 3, and 7,61 i.e. without interacting with TMs 4ꢀ6.
Consequently, the binding mode of 3 differs from other reported binding modes of CXCR4
antagonists (see below), and is distinct from the present binding mode for 1.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Peptide Antagonists. Prior to the publication of the Xꢀray structure of CXCR4, computational
binding mode studies had been performed for the cyclopentapeptide CXCR4 antagonist FC13162,
63
and 163 based on molecular docking to homology models of CXCR4 that were built on the
crystal structure of bovine rhodopsin. These studies were purely theoretical, i.e. not based on any
experimentally verified ligandꢀreceptor interactions. Moreover, when the Xꢀray structure of
CXCR433 was revealed in 2010 it showed that bovine rhodopsin was a poor template for
CXCR4,64 questioning the validity of the previously proposed binding models for the small
molecule peptideꢀbased antagonists.
ACS Paragon Plus Environment
21

Journal of Medicinal Chemistry
Page 22 of 45
1
2
3
4
5
6
7
8
9
However, we have recently proposed an experimentally verified binding mode for FC131
based on ligand modifications, receptor mutagenesis, and molecular docking to the Xꢀray
structure of CXCR4.35 The binding model for FC131 was generally consistent with the binding
mode for the Arg1ꢀArg2ꢀ1ꢀNal3 fragment of the 16ꢀmer peptide antagonist 2 in complex with
CXCR4 (PDB code 3OE0),33 i.e. the Arg1 sideꢀchain in FC131 projected towards ECL2 of the
receptor to interact with Asp187, the Arg2 guanidino group interacted with His113, Thr117, and
Asp171 in TMs 3ꢀ4, while the aromatic 2ꢀNal3 sideꢀchain was accommodated in a hydrophobic
pocket around TM5.
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
Thus, the Arg2 guanidino group in FC131 and the guanidino group (R2 position) in 1 both
interact with His113, Thr117, and Asp171 (Figure 6); structural comparison of the reported binding
mode for FC13135 and the present binding mode for 1 (Supporting Information, Figure S2) also
shows that the LꢀArg2 side chain of FC131 overlaps relatively well with the central arginine
residue of 1. However, the orientation of the naphthyl groups in FC131 and 1 are different; the 2ꢀ
naphthyl group in FC131 is oriented towards TM5, while the 1ꢀnaphthyl group in 1 is facing
TM2.
AMDꢀcompound Series. Based on siteꢀdirected mutagenesis studies of CXCR4 and binding
site transfer to CXCR3, the critical CXCR4 receptor residues for binding of the bicyclam
AMD3100 (Figure 1) have been shown to be Asp171, Asp262, and Glu288.36, 37 Based on the fact
that AMD3100 and 1 inhibit binding of the same set of monoclonal antibodies, directed against
the extracellular loops of CXCR4, it has been suggested that they have similar binding sites and
molecular mode of action.28, 65 As Asp171, Asp262, and indirectly Glu288 are also involved in
binding of 1 (Figure 6), the present study confirms that these two compounds bind to the same
receptor region.
ACS Paragon Plus Environment
22

Page 23 of 45
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
In addition to the key interactions seen for AMD3100, the monocyclam analog AMD3465
(Figure 3) has been shown to interact with His113 in TM3 and His281 in TM7,39, 40 which were
also identified as important for 1 in the present study. The reduced potency observed for 1 in the
D262N and H281A mutants are of particular interest as the same mutations were shown to have
selective importance for the receptor interactions of AMD3465,39 which contains the same
(pyridinꢀ2ꢀylmethyl)amino moiety as 1. Thus, in addition to the key interactions with His113 and
Asp171 (R2 guanidino group) and Asp262 (R1 amino group), the present study suggests that the
pyridine ring in the R1 side chain of 1 picks up an additional interaction with His281 in TM7, in
the same way as AMD3465.
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 interaction pattern for the nonꢀcyclam 4 (Figure 1F) is more unclear, but it has been
suggested that it binds in a similar fashion as AMD3100 and AMD3465; however, Asp97 and
Asp171 appeared to have a unique importance for the binding of this analog.40, 66
CONCLUSIONS
The present study describes the binding of the prototype tripeptidomimetic antagonist 1 to
CXCR4 based on SAR studies, receptor mapping, and molecular docking. This ternary approach
resulted in the identification of a binding model for 1 in which the secondary amino group is
anchored by interaction with Asp262 in TM6, with the distal pyridine ring further involved in a πꢀ
π stacking interaction with His281 in TM7; the Argꢀguanidino group engages in polar interactions
mainly with His113 in TM3 and Asp171 in TM4, while the naphthyl group is tightly packed in a
hydrophobic subpocket formed by residues Trp94, Tyr45 and Tyr116. Our findings contradict
earlier suggestions that 1 mimics the binding of the Arg1ꢀArg1ꢀ2ꢀNal3 fragment in the
cyclopentapeptide antagonist FC131. Instead, we propose a “hybrid” binding mode for 1 where
ACS Paragon Plus Environment
23

Journal of Medicinal Chemistry
Page 24 of 45
1
2
the guanidino group of the central arginine residue (R2 side chain) overlaps with the binding
mode of the Arg2 residue in the peptide antagonist FC131 and 2, while the (pyridinꢀ2ꢀ
ylmethyl)amino moiety (R1 side chain) overlaps with the binding mode for the same moiety in
the monocyclam AMD3465. This experimentally verified binding model for 1 provides
important insight for future structureꢀbased design of smallꢀmolecule CXCR4 antagonists.
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
EXPERIMENTAL SECTION
Chemistry. General. All reagents and solvents were purchased from SigmaꢀAldrich and used
1
as received. H and 13C NMR spectra were recorded on a 400 MHz Varian spectrometer.
13
Chemical shifts are expressed in ppm relative methanol (1H 3.31 ppm, C 49.0 ppm). Coupling
constants are given in Hertz (Hz). Mass spectra were recorded on a Micromass Quattro LC
(Micromass, Manchester, UK) and highꢀresolution mass spectra were recorded on an LTQ
Orbitrap XL (Thermo Scientific, Bremen, Germany) using positiveꢀmode ESI. Preparative RPꢀ
HPLC was performed on a Waters 600 Semi Prep System equipped with an XBridge™ C18
column (250 mm × 19 mm, 10 µm particle size) using 15 mL/min flow rate or a Waters 2695
system equipped with an XBridge™ C18 column (250 mm × 10 mm, 5 µm particle size) using 10
mL/min flow rate with mixtures of acetonitrile and water (both containing 0.1% TFA) as the
eluent in both cases Analytical RPꢀHPLC was performed on a Waters 2695 system equipped
with an XBridge™ C18 column (250 mm × 4.6 mm, 5 µm particle size,) using 1 mL/min flow
rate with detection at 214 nm and 254 nm facilitated by a PDA detector (210 – 310 nm). In some
cases, analyses were performed on a Waters ACQUITY UHPLC HꢀClass equipped with a
Waters ACQUITY UHPLC BEH C18 column (1 × 150 mm, 1.75 µm particle size) with a 0.120
ACS Paragon Plus Environment
24

Page 25 of 45
Journal of Medicinal Chemistry
1
2
3
4
5
6
mL/min flow rate. All compounds undergoing biological evaluation were found to be of >95%
purity as judged by analytical RPꢀHPLC with PDA detection (210 – 310 nm).
7
8
9
N-((S)-5-guanidino-1-(((S)-1-(naphthalen-1-yl)ethyl)amino)-1-oxopentan-2-yl)-4-
(((pyridin-2-ylmethyl)amino)methyl)benzamide (1). The title compound was prepared
following an earlier reported procedure.29 Purification by RPꢀHPLC afforded after lyophilization
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
the title compound as a white fluffy material. H NMR (400 MHz, CD3OD) δ 8.12 (d, J = 8.3,
1H), 7.96 (d, J = 8.0, 2H), 7.91 – 7.85 (m, 2H), 7.81 (d, J = 8.2, 1H), 7.64 (d, J = 8.0, 2H), 7.60 –
7.37 (m, 7H), 5.94 – 5.82 (m, 1H), 4.62 (app t, J = 8.0, 1H), 4.41 (s, 2H), 4.39 (s, 2H), 3.19 –
13
3.05 (m, 2H), 1.94 – 1.74 (m, 2H), 1.66 (d, J = 6.9, 3H), 1.64 – 1.50 (m, 2H); C NMR (100
MHz, CD3OD) δ 172.7, 169.2, 158.5, 152.3, 150.6, 140.1, 138.8, 136.3, 136.1, 135.4, 132.2,
131.2, 129.9, 129.3, 129.0, 127.2, 126.7, 126.4, 125.1, 124.1, 124.1, 123.7, 54.9, 51.5, 51.3, 46.2,
41.9, 30.5, 26.4, 21.2; HRMS (ESI): m/z [M + H]+ calcd for C32H38N7O2: 552.3081, found
552.3081. The purity of the title compound was found to be >99% using analytical RPꢀHPLC
(diode array detection, 210 nm – 310 nm).
4-(aminomethyl)-N-((S)-5-guanidino-1-(((S)-1-(naphthalen-1-yl)ethyl)amino)-1-
oxopentan-2-yl)benzamide (6). The title compound was prepared adopting a literature
procedure29 where NꢀBoc 4ꢀ(methylamino)benzoic acid was used instead of 11 (Scheme 1).
Purification by RPꢀHPLC afforded after lyophilization the title compound as a white fluffy
material. 1H NMR (400 MHz, CD3OD) δ 7.89 (d, J = 8.3, 1H), 7.72 (d, J = 8.3, 2H), 7.66 (app t,
J = 8.2, 1H), 7.58 (d, J = 8.2, 1H), 7.41 – 7.23 (m, 6H), 5.69 – 5.60 (m, 1H), 4.41 (app dd, J =
8.3, 6.0, 1H), 3.98 (s, 2H), 3.00 – 2.82 (m, 2H), 1.76 – 1.55 (m, 2H), 1.52 – 1.45 (m, 1H), 1.44
(d, J = 6.9, 3H), 1.40 – 1.30 (m, 1H); 13C NMR (100 MHz, CD3OD) δ 172.7, 169.1, 158.3,
139.9, 138.0, 135.6, 135.2, 132.0, 129.8, 129.7, 129.1, 128.9, 127.1, 126.6, 126.2, 123.9, 123.5,
ACS Paragon Plus Environment
25

Journal of Medicinal Chemistry
Page 26 of 45
1
2
54.8, 46.0, 43.6, 41.7, 30.3, 26.3, 21.1; HRMS (ESI) : m/z [M + H]+ calcd for C26H33N6O2:
461.2660, found 461.2664. The purity of the title compound was found to be >99% using
analytical RPꢀHPLC (diode array detection, 210 nm – 310 nm).
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
4-((benzylamino)methyl)-N-((S)-5-guanidino-1-(((S)-1-(naphthalen-1-yl)ethyl)amino)-1-
oxopentan-2-yl)benzamide (5). A solution of 6 (crude product following TFA mediated
deprotection, 0.09 mmol) in DMF (2.0 mL) was cooled to 0 °C and K2CO3 (24 mg, 0.18 mmol)
was added. To this suspension, benzyl bromide (10 µL, 0.09 mmol) was added slowly, and the
reaction mixture was allowed to stir at rt for 24 h. Water (1 mL) was added and the mixture was
extracted with diethyl ether (10 mL), washed with water (2 x 5 mL) and a saturated solution of
NaCl (10 mL). The organic layer was dried over anhydrous MgSO4, filtered and the solvents
were removed under reduced pressure. The crude product was purified using RPꢀHPLC followed
1
by lyophilization to give the title compound as a fluffy white material. H NMR (400 MHz,
CD3OD) δ 8.12 (d, J = 8.2, 1H), 7.95 (d, J = 8.2, 2H), 7.90 (d, J = 7.8, 1H), 7.81 (d, J = 8.2, 1H),
7.60 (t, J = 7.6, 3H), 7.57 – 7.45 (m, 8H), 5.95 – 5.82 (m, 1H), 4.62 (app dd, J = 7.9, 6.3, 1H),
4.33 (s, 2H), 4.28 (s, 2H), 3.21 – 3.06 (m, 2H), 1.97 – 1.73 (m, 2H), 1.66 (d, J = 6.9, 3H), 1.64 –
1.48 (m, 2H); 13C NMR (100 MHz, CD3OD) 173.0, 169.5, 158.7, 140.3, 136.5, 136.3, 135.7,
132.5, 132.4, 131.4, 131.3, 131.0, 130.6, 130.1, 129.6, 129.3, 127.5, 127.0, 126.6, 124.3, 123.9,
55.2, 52.5, 51.7, 46.5, 42.1, 30.7; HRMS (ESI): m/z [M + H]+ calcd for C33H39N6O2: 551.3129,
found 551.3131. The purity of the title compound was found to be >99% using analytical RPꢀ
HPLC (diode array detection, 210 nm – 310 nm).
N-((S)-5-guanidino-1-(((S)-1-(naphthalen-1-yl)ethyl)amino)-1-oxopentan-2-yl)-4-
methylbenzamide (7). The title compound was prepared adopting a literature procedure29 where
4ꢀmethylbenzoic acid was used instead of 11 (Scheme 1). Purification by RPꢀHPLC afforded
ACS Paragon Plus Environment
26

Page 27 of 45
Journal of Medicinal Chemistry
1
2
after lyophilization the title compound as a white fluffy material. 1H NMR (400 MHz, CD3OD) δ
8.12 (d, J = 8.3, 1H), 7.90 (d, J = 7.7, 1H), 7.81 (d, J = 8.2, 1H), 7.76 (d, J = 8.2, 2H), 7.60 –
7.45 (m, 4H), 7.30 (d, J = 8.2, 2H), 5.88 (m, 1H), 4.70 – 4.59 (m, 1H), 3.20 – 3.04 (m, 2H), 2.41
(s, 3H), 1.93 – 1.72 (m, 2H), 1.66 (d, J = 6.9, 3H), 1.64 – 1.48 (m, 2H); 13C NMR (100 MHz,
CDCl3) 171.5, 168.9, 157.0, 143.1, 138.7, 133.8, 130.5, 129.8, 129.4, 128.9, 128.0, 127.2, 126.3,
125.8, 125.4, 122.6, 122.4, 53.2, 45.5, 40.4, 29.6, 24.5, 21.4, 21.1; HRMS (ESI): m/z [M + H]+
calcd for C26H32N5O2: 446.2551, found 446.2550. The purity of the title compound was found to
be >99% using analytical RPꢀHPLC (diode array detection, 210 nm – 310 nm).
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
N-((S)-1-(((S)-1-(naphthalen-1-yl)ethyl)amino)-1-oxo-5-ureidopentan-2-yl)-4-(((pyridin-2-
ylmethyl)amino)methyl)benzamide (8). The title compound was prepared adopting a literature
procedure29 where FmocꢀCitꢀOH was used instead of FmocꢀArg(Pbf)ꢀOH. Purification by RPꢀ
1
HPLC afforded after lyophilization the title compound as a white fluffy material. H NMR (400
MHz, CD3OD) δ 8.38 (d, J = 4.3, 1H), 7.81 (d, J = 8.3, 1H), 7.67 (d, J = 8.3, 2H), 7.63 – 7.57
(m, 2H), 7.51 (d, J = 8.2, 1H), 7.32 (m, 3H), 7.27 – 7.13 (m, 5H), 5.56 (app q, J = 6.8, 1H), 4.37
(app dd, J = 8.7, 5.6, 1H), 4.12 (s, 2H), 4.09 (s, 2H), 2.89 – 2.73 (m, 2H), 1.65 – 1.43 (m, 2H),
1.36 (d, J = 6.9, 3H), 1.34 – 1.16 (m, 2H); 13C NMR (100 MHz, CD3OD) δ 173.3, 169.3, 162.3,
152.1, 150.4, 140.2, 139.1, 136.3, 135.9, 135.4, 132.1, 131.2, 129.9, 129.4, 129.0, 127.3, 126.7,
126.4, 125.2, 124.4, 124.1, 123.6, 55.2, 51.5, 51.2, 46.3, 40.4, 30.5, 27.7, 21.4; HRMS (ESI): m/z
[M + H]+ calcd for C32H37N6O3: 553.2922, found 553.2921. The purity of the title compound
was found to be >99% using analytical RPꢀHPLC (diode array detection, 210 nm – 310 nm).
(S)-N-(1-(ethylamino)-5-guanidino-1-oxopentan-2-yl)-4-(((pyridin-2-
ylmethyl)amino)methyl)benzamide (9). The title compound was prepared adopting a literature
procedure29 where ethanamine was used instead of (S)ꢀ1ꢀ(naphthaleneꢀ1ꢀyl)ethanamine.
ACS Paragon Plus Environment
27

Journal of Medicinal Chemistry
Page 28 of 45
1
2
3
4
5
6
7
8
9
Purification by RPꢀHPLC afforded after lyophilization the title compound as a white fluffy
material. 1H NMR (400 MHz, CD3OD) δ 8.66 (d, J = 4.8, 1H), 7.97 (d, J = 8.2, 2H), 7.87 (td, J =
7.7, 1.7, 1H), 7.64 (d, J = 8.2, 2H), 7.46 – 7.39 (m, 2H), 4.55 (dd, J = 8.6, 5.9 Hz, 1H), 4.40 (s,
2H), 4.39 (s, 2H), 3.29 – 3.19 (m, 4H), 1.97 – 1.61 (m, 4H), 1.14 (t, J = 7.3, 3H); 13C NMR (100
MHz, CD3OD) δ 173.6, 169.3, 158.8, 152.3, 150.7, 138.8, 136.3, 136.1, 131.2, 129.4, 125.1,
124.2, 55.0, 51.5, 51.3, 42.0, 35.4, 30.4, 26.5, 14.8; HRMS (ESI): m/z [M + H]+ calcd for
C22H32N7O2: 426.2612, found 426.2611. The purity of the title compound was found to be >99%
using analytical RPꢀHPLC (diode array detection, 210 nm – 310 nm).
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
Biology. Materials. The human CXCR4 receptor cDNA was kindly provided by Tim Wells
(GSK, UK). The human chemokine CXCL12 was purchased from Peprotech (NJ, USA). The
monoclonal antibody 12G5 was kindly provided by Jim Hoxie (University of Pennsylvania,
Philadelphia, PA, USA), and the promiscuous chimeric G protein Gαꢂ6qi4myr (abbreviated
Gαqi4myr) was kindly provided by Evi Kostenis (University of Bonn, Germany)67.
Myo[3H]inositol (PT6ꢀ271), scintillation proximity assay (SPA) beads, and BoltonꢀHunter
reagent for 12G5 iodination (inꢀhouse procedure)37 were purchased from Perkin Elmer (MA,
USA). PrimerꢀcDNAs for mutagenesis were purchased from TAG Copenhagen (Copenhagen,
Denmark), and AG1ꢀX8 anion exchange resin from BioꢀRad (CA, USA). Stock solutions of the
ligand antagonists were prepared in different mixtures of DMSO:water, and dilutions were made
in water. For dilutions of CXCL12, a buffer with 1 mM acetic acid + 0.1% BSA was used.
Site-directed Mutagenesis. Point mutations were introduced in the receptor by the polymerase
chain reaction overlap extension technique68 or the QuikChangeTM technique (Agilent
Technologies, CA, USA), using WTꢀCXCR4 as a template. All reactions were carried out using
Pfu polymerase (Stratagene, CA, USA) under conditions recommended by the manufacturer. The
ACS Paragon Plus Environment
28

Page 29 of 45
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
mutations were cloned into the eukaryotic expression vector pcDNA3.1+ (Invitrogen, UK), and
were verified by restriction endonuclease digestion and DNA sequencing (Eurofins MWG
Operon, Germany).
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
Transfections and Tissue Culture. COSꢀ7 cells were grown at 10% CO2 and 37 °C in
Dulbecco's modified Eagle's medium with glutamax (Invitrogen, U.K.) adjusted with 10% fetal
bovine serum, 180 units/mL penicillin, and 45 ꢁg/mL penicillin/streptomycin (PenStrep).
Transfection of the COSꢀ7 cells was performed by the calcium phosphate precipitation method.69
Phosphatidylinositol (PI) Turnover Assays. Transfection of COSꢀ7 cells was carried out
according to the abovementioned procedure. Shortly, 6 × 106 cells were transfected with 20 ꢁg of
receptor cDNA along with 30 ꢁg of the promiscuous chimeric G protein, Gαqi4myr, which turns
the Gαiꢀcoupled signal, the most common pathway for endogenous chemokine receptors, into the
Gαq pathway (phospholipase C activation measured as IP3 turnover). Two alternative methods,
which were found to give the same result, were used:
Anionꢀexchange Separation of Phosphoinositides. In the classic IP3ꢀassay which includes an
anionꢀexchange chromatography step, one day after transfection COSꢀ7 cells (1.5 × 105
cells/well) were incubated for 24 h with 2 ꢁCi of myo[3H]inositol in 0.3 mL of growth medium
per well in a 24ꢀwell plate. Next day, the cells were washed twice in PBS, and were incubated
(90 min) in 0.2 mL of Hank's balanced salt solution (Invitrogen, U.K.) supplemented with 10
mM LiCl at 37 °C in the presence of various concentrations of the tested compounds. Cells were
then extracted by addition of 1 mL of 10 mM formic acid to each well followed by incubation on
ice for 30−60 min. The generated [3H]inositol phosphates were purified on an AG 1ꢀX8 anion
exchange resin. Following the addition of multipurpose liquid scintillation cocktail (Gold Star,
TriskemꢀInternational, France), γꢀradiation was counted in a Beckman Coulter counter LS6500.
ACS Paragon Plus Environment
29