Studies on Aryl-Substituted Phenylalanines: Synthesis, Activity, and Different Binding Modes at AMPA Receptors

Article abstract of DOI:10.1021/acs.jmedchem.5b01666

A series of racemic aryl-substituted phenylalanines was synthesized and evaluated in vitro at recombinant rat GluA1-3, at GluK1-3, and at native AMPA receptors. The individual enantiomers of two target compounds, (RS)-2-amino-3-(3,4-dichloro-5-(5-hydroxyp

Full text of DOI:10.1021/acs.jmedchem.5b01666

Subscriber access provided by Mount Allison University | Libraries and Archives
Article
Studies on aryl-substituted phenylalanines - synthesis,
activity and different binding modes at AMPA receptors
Ewa Szymanska, Karla Frydenvang, Darryl S Pickering, Christian Krintel, Birgitte Nielsen, Ayesheh
Kooshki, Linda G. Zachariassen, Lars Olsen, Jette Sandholm Kastrup, and Tommy N. Johansen
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.5b01666 • Publication Date (Web): 12 Dec 2015
Downloaded from http://pubs.acs.org on December 15, 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 49
Journal of Medicinal Chemistry
1
Studies on aryl-substituted phenylalanines - synthesis, activity and different
binding modes at AMPA receptors
2
3
4
5
6
7
8
9
Ewa Szymanska,a,b,c Karla Frydenvang,a,c* Darryl S. Pickering,a Christian Krintel,a Birgitte Nielsen,a
Ayesheh Kooshki,a Linda G. Zachariassen,a Lars Olsen,a Jette S. Kastrup,a and Tommy N.
Johansena*
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
aDepartment of Drug Design and Pharmacology, Faculty of Health and Medical Sciences, University
of Copenhagen, 2100 Copenhagen Ø, Denmark; bpresent affiliation: Department of Technology and
Biotechnology of Drugs, Jagiellonian University Medical College, Medyczna 9, PL 30ꢀ688 Kraków,
Poland.
Keywords
ionotropic glutamate receptors, GluA2, competitive antagonist, chiral resolution, Xꢀray structure,
molecular modeling
1
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 49
1
2
Abstract
3
4
5
6
7
8
9
A series of racemic arylꢀsubstituted phenylalanines was synthesized and evaluated in vitro at
recombinant rat GluA1–3, at GluK1–3 and at native AMPA receptors. The individual enantiomers of
two target compounds, (RS)ꢀ2ꢀaminoꢀ3ꢀ(3,4ꢀdichloroꢀ5ꢀ(5ꢀhydroxypyridinꢀ3ꢀyl)phenyl)propanoic
acid 37 and (RS)ꢀ2ꢀaminoꢀ3ꢀ(3'ꢀhydroxybiphenylꢀ3ꢀyl)propanoic acid 38 were characterized. (S)ꢀ37
and (R)ꢀ38 were identified as the only biologically active isomers, both being antagonists at GluA2
receptors with Kb of 1.80 and 3.90 µM, respectively. To address this difference in
enantiopharmacology, not previously seen for amino acidꢀbased AMPA receptor antagonists, Xꢀray
crystal structures of both eutomers in complex with the GluA2 ligand binding domain were solved.
The coꢀcrystal structures of (S)ꢀ37 and (R)ꢀ38 showed similar interactions of the amino acid parts,
but unexpected and different orientations and interactions of the biꢀaromatic parts of the ligands
inside the binding site, with (R)ꢀ38 having a binding mode not previously identified for amino acidꢀ
based 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
2
ACS Paragon Plus Environment

Page 3 of 49
Journal of Medicinal Chemistry
1
2
Introduction
3
4
5
6
7
8
9
Most excitatory neurotransmission within the mammalian central nervous system is mediated by the
ionotropic glutamate receptors (iGluRs), which include three functionally distinct subclasses named
according to their selective affinity for the specific agonists: NꢀmethylꢀDꢀaspartic acid (NMDA),
(RS)ꢀ2ꢀaminoꢀ3ꢀ(3ꢀhydroxyꢀ5ꢀmethylꢀ4ꢀisoxazolyl)propionic acid (AMPA) and kainate (KA).1, 2 All
three subclasses of iGluRs play an important role in controlling synaptic plasticity, underlying
learning and memory processes. On the other hand, the excessive activation of calciumꢀpermeable
NMDA receptors (and in some circumstances AMPA receptors) leads to abnormal Ca2+ influx that
can initiate neurotoxic intracellular signaling cascades and neuronal cell death observed in several
neurological diseases, including stroke, Alzheimer, Parkinson, Huntington and neuropathic pain.3, 4
Structurally, iGluRs are tetrameric assemblies of subunits encoded by separate genes. Native AMPA
and kainate receptors are homoꢀ or heterooligomers formed by subunits GluA1–GluA4 and GluK1–
GluK5, respectively. In case of the NMDA subclass seven subunits denoted GluN1, GluN2A–
GluN2D and GluN3A–GluN3B have been cloned and characterized.1 Each individual subunit
consists of an extracellular aminoꢀterminal domain (ATD), an extracellular ligand binding domain
(LBD), a transmembrane domain that forms the cationꢀselective pore and a cytoplasmatic Cꢀterminal
domain involved in receptor localization and modulation.
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
Taking into account the role of iGluRs in a variety of neurological disorders, the development of
subtypeꢀspecific competitive antagonists as neuroprotective agents is of considerable therapeutic
interest. Over the past decades a wide group of competitive antagonists of nonꢀNMDA iGluRs has
been developed, however, due to a highly conserved nature of the LBD only few of them show
subtype selectivity.2, 5ꢀ7 Structurally, most of the known competitive antagonists belong to one of two
major classes: 1) expanded versions of agonists and 2) quinoxalines (Figure 1).2 The first class
includes compounds with an αꢀamino acid motif of agonists and a distal negatively charged group,
most often a carboxylate, phosphonate or tetrazolyl ring. Typical representatives of this category are
isoxazoleꢀbased compounds8, 9 [e.g. (S)ꢀATPO (1)], willardiine derivatives10 (e.g. UBP310 (2)) or
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 49
1
2
3
4
5
6
7
8
cisꢀdecahydroisoquinolines11, 12 (e.g. tezampanel (3)). Furthermore, it has been shown that a series of
phenylalanine derivatives (such as compound 4) also belong to this class.13, 14 The second class of
nonꢀNMDA iGluRs competitive antagonists comprises a variety of quinoxalines,15, 16 represented by
NBQX (5) or ZKꢀ200775 (6). Some structurally dissimilar competitive AMPA receptor antagonists,
such as the isatine oxime NS1209 (7), have also been described.17, 18
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 advances in crystallography of soluble constructs of LBDs of iGluRs in complex with agonists
and competitive antagonists have provided information on receptor motifs for recognition of
structurally diverse ligands.10, 19ꢀ21 Comparison of crystal structures of the GluA2 or GluK1 subtype
LBD with bound αꢀamino acidꢀbased compounds reveals the preserved binding mode of the αꢀamino
acid fragment for both agonists and competitive antagonists, while the distal fragment of antagonist
structures interacts with residues of domain D2 with a footꢀinꢀtheꢀdoor mechanism, stabilizing the
open conformation of the receptor. Furthermore, a predominant amount of these compounds bind with
their αꢀamino acid part in the Sꢀconfiguration,20 although few examples of Rꢀconfigured amino acidꢀ
based ligands are known.22ꢀ25
In a previous study, we have reported a new series of phenylalanines with a rigid or a flexible
biaromatic ring system (Figure 2A).26 Among these compounds, the dichloroꢀsubstituted biphenyl
analogue substituted with a hydroxyl group in position 3 of the distal phenyl ring (9) showed
selective affinity to AMPA receptors (IC50 = 4.6 ꢁM), while the similar compound with a 3ꢀ
carboxylic substituent turned out to bind selectively to GluK1 receptors (Ki = 2.8 ꢁM). Encouraged
by these results we report here the synthesis and chiral chromatography resolution of a new series of
aryl substituted phenylalanines (Figure 2B) as well as their biological evaluation at native and
selected recombinant homomeric iGluRs. Furthermore, crystal structures were solved of selected
eutomers bound to the LBD of the GluA2 subunit to investigate the binding modes at AMPA
receptors.
4
ACS Paragon Plus Environment

Page 5 of 49
Journal of Medicinal Chemistry
1
2
Results
3
4
5
6
7
8
9
In the series of the target amino acids we kept the biaryl core of compounds with an additional
phenyl or pyridine ring introduced into the 5ꢀposition of the phenylalanine ring (Figure 2B). The
structural modifications were focused mainly on the distal ring, namely: 1) a change of the position
and the number of hydroxyl groups, 2) replacement of the hydroxyl substituent of (9) and (10) with
either an amino or sulfonamide group, 3) blockade of the hydrogen donor availability of the hydroxyl
group by introducing an alkyl fragment. Taking into account the influence of the presence of two
chlorines on AMPA receptor affinity,26 in most of new compounds we preserved the
dichlorophenylalanine substitution pattern.
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
Chemistry
The synthesis of new amino acids was based on previously developed convergent synthesis
strategies26 (Schemes 1 and 2). Treatment of 3ꢀiodotoluene precursors 11 and 12 with NBS under
free radical conditions, followed by substitution with the sodium salt of diethyl acetamidomalonate
afforded the key intermediates 13 and 14. The final amino acids 27ꢀ38 were prepared as a result of a
number of Suzuki coupling reactions of aryl iodides 13 and 14 with boronic acids or esters, followed
by deprotection in 48% aqueous hydrobromic acid (or, in case of 30 and 31, in aqueous hydrochloric
acid and acetic acid) and purified by reverseꢀphase HPLC.
Synthesis of Sꢀ and Rꢀenantiomers of 38 was performed in a similar way (Scheme 2). Ethyl esters of
commercially available (S)ꢀ and (R)ꢀNꢀBocꢀ3ꢀbromophenylalanine were reacted with
3-methoxyphenylboronic acid under Suzuki conditions. Deprotection and purification by reverseꢀ
phase HPLC gave the desired amino acids (S)ꢀ38 (ee = 98.3%) and (R)ꢀ38 (ee = 98.0%).
Chromatographic resolution of racemic 37 was carried out by chiral HPLC using the Chirobiotic T
column.27, 28 The firstꢀeluting (+)ꢀisomer as well as the lateꢀeluting (−)ꢀisomer were isolated in good
to high stereochemical purity (ee > 97% and 99.7%, respectively). In order to address the
configurational assignment of (+)ꢀ and (−)ꢀ37, the elution order of (+)ꢀ and (−)ꢀ37 as well as (R)ꢀ(−)ꢀ
5
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 49
1
2
and (S)ꢀ(+)ꢀ38 were studied using the Chirobiotic T column and using a ligandꢀexchange HPLC
3
4
5
6
7
8
9
column which contains (R)ꢀPro as a chiral selector.29 As expected for αꢀamino acids,28, 29 (S)ꢀ(+)ꢀ38
is the first eluting enantiomer both on the Chirobiotic T and on the (R)ꢀPro columns. When analyzed
on the (R)ꢀPro column, (+)ꢀ37 turned out to be the first eluting enantiomer. Thus, the elution order of
(+)ꢀ and (−)ꢀ37 on two columns with different chiral selectors indicate that (+)ꢀ37 has Sꢀ
configuration. The configurational assignment of (S)ꢀ(+)ꢀ37 was later supported by Xꢀray
crystallography (See Structure Determination for further details).
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
Binding Pharmacology
The results of in vitro pharmacological activity at native AMPA receptors in rat cortical membranes
for compounds 27–38 (and, for comparison, for previously described 8–1026) are given in Table 1.
Among the tested dichlorosubstituted phenylalanines, only amino acids possessing a 3ꢀhydroxyl
substituent in the distal aromatic ring showed measurable affinity to AMPA receptors (27, 28, 37,
38), with IC50 values in the range of 1.5–16 ꢁM. The introduction of an additional hydroxyl group
into the distal phenyl did not enhance binding at AMPA receptors; however, a 2–3ꢀfold decrease in
affinity for 27 and 28 was observed relative to the reference compound 9. The exchange of the
hydroxyl for an ether group (29, 30), a sulfonamide (34, 35) or even a hydroxymethyl group (31, 32)
significantly decreased or completely removed the affinity. No affinity (IC50 > 100 µM) was
observed for 33, in which the hydroxyl group in the 3ꢀposition was replaced with an amino group.
Modification of the distal aryl and exchange of a phenyl for a pyridine ring, keeping the hydroxyl
substitution pattern of 9, yielded the most potent AMPA receptor ligand, (RS)ꢀ2ꢀaminoꢀ3ꢀ(3,4ꢀ
dichloroꢀ5ꢀ(5ꢀhydroxypyridinꢀ3ꢀyl)phenyl)propanoic acid (37) with affinity determined to 1.5 ꢁM,
whereas the regional isomer 36 showed no AMPA receptor affinity (IC50 > 100 µM). Furthermore,
compound 38, which was included within the present series as a close structural analogue of 9
without the dichloro pattern in the phenylalanine ring, showed an AMPA receptor affinity
comparable with that of 9. In AMPA receptor binding studies using enantiomerically pure 37 and 38,
6
ACS Paragon Plus Environment

Page 7 of 49
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
(S)ꢀ37 and (R)ꢀ38 were clearly identified as the respective eutomers (Table 1). In the homogenate
binding studies to native KA and NMDA receptors as well as binding studies to recombinant rat
GluK1 homomeric kainate receptors expressed in Sf9 insect cell membranes, all target compounds
showed either very weak micromolar (50 – 100 ꢁM) or no affinity (> 100 ꢁM).
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 individual enantiomers of the two most active compounds among the present series of
phenylalanines, 37 and 38, were subsequently characterized at the cloned rat GluA1–3 and GluK1–3
receptors (Table 2). In these radioligand binding assays the Sꢀenantiomer of 37 also was found to be
responsible for the biological activity of the racemic compound at GluA1–3 homomeric receptors
with the strongest binding to the GluA2 receptor (Ki = 1.7 ꢁM), and at the same time showing no
affinity (Ki > 100 µM) to any of the homomeric GluK1–3 receptors. A 4ꢀfold higher affinity at
GluA2 relative to GluA1 and GluA3 (P < 0.0001, oneꢀway ANOVA with HolmꢀSidak postꢀtest) was
observed for this compound; however, no statistically significant difference was found between
affinity to GluA1 and GluA3 receptors. The Rꢀenantiomer of 37 was inactive at all tested
recombinant receptors (Ki > 100 ꢁM).
In contrast to 37 and most other αꢀamino acidꢀbased ligands of AMPA receptors, but in agreement
with the stereoselectivity observed in AMPA receptor binding to native receptors (Table 1), only the
Rꢀenantiomer of 38 was found to show biological activity, while (S)ꢀ38 lacked affinity to both
homomeric AMPA and kainate receptors (Ki > 100 ꢁM; Table 2). (R)ꢀ38 presented micromolar
binding to both recombinant GluA1ꢀ3 receptors as well as to kainate recombinant GluK3 receptors
(Table 2).
Functional Pharmacology
Using two electrode voltage clamp (TEVC) electrophysiology the antagonist properties of (S)ꢀ37 and
(R)ꢀ38 were examined at homomeric GluA2(Q)i receptors expressed in X. laevis oocytes. As shown
in Figure 3, (S)ꢀ37 doseꢀdependently antagonized the responses to (S)ꢀGlu. The functional Kb value
of (S)ꢀ37 at GluA2(Q)i was determined to be 1.80 ꢀM, which is not statistically significantly
7
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 49
1
2
3
4
5
6
7
different from its respective Ki value (1.74 ꢀM) determined from radioligand binding experiments at
full length GluA2(R)o (P = 0.949, tꢀtest) (Table 2). (R)ꢀ38 antagonized (S)ꢀGlu responses with a
calculated Kb value of 3.90 ꢀM, which was more potent than the observed GluA2(R)o binding
8
9
affinity (Ki = 24 ꢀM). Application of either 0.60 mM (S)ꢀ37 or 1 mM (R)ꢀ38, in the presence of 100
ꢀM cyclothiazide, did not produce any detectable agonist responses at GluA2(Q)i, indicating that
neither (S)ꢀ37 nor (R)ꢀ38 show partial agonist effects at these concentrations under our experimental
conditions. However, it cannot be completely ruled out that partial agonism may be seen at much
higher concentrations of the 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
Structure determination
The GluA2 LBD (GluR2ꢀS1S2J) was coꢀcrystallized with (S)ꢀ37 and (R)ꢀ38, respectively, and
formed GluA2 dimeric complexes in both cases. In the coꢀcrystal structures (S)ꢀ37 as well as (R)ꢀ38
were located in the (S)ꢀGlu binding cleft of the LBD, but found to adopt very different binding
modes.
The structure of GluA2 LBD with (S)-37. The structure of GluA2 LBD in complex with (S)ꢀ37
was solved at 2.0 Å resolution, and the complex crystallized in space group P21212 with one
molecule in the asymmetric unit of the crystal. A twoꢀfold symmetrical dimer of GluA2 LBD is
formed by interaction with a symmetryꢀrelated molecule as previously seen.19 For further
information on data collection and refinement statistics, see Table 3. (S)ꢀ37 induces partial LBD
domain opening (rigid body movements of lobes D1 and D2) of 13.5° of the GluA2 LBD, relative to
the structure with Glu (PDB ID 1FTJ, molB19) (Figure 4A). The Ile633ꢀIle633(sym) linkerꢀlinker
distance of 32.6 Å is intermediate among the linkerꢀlinker distances observed for full agonists and
antagonists, in accordance with the partial domain opening.20
Generally, the electron density is well defined and (S)ꢀ37 could unambiguously be modelled into the
density (Figure 4B), thus confirming the absolute configuration indicated by the chiral HPLC
analyses. The binding mode of (S)ꢀ37 at GluA2 LBD resembles that of other amino acidꢀbased
8
ACS Paragon Plus Environment

Page 9 of 49
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
competitive antagonists.9, 10, 13, 23 Compared to these antagonists, the αꢀcarboxyl and αꢀamino groups
of (S)ꢀ37 are observed at the same position in the GluA2 binding cleft and show the same type of
hydrogen bonding interactions to surrounding residues (Pro478, Thr480, Arg485 and Glu705; full
length numbering without signal peptide).
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 hydroxyꢀpyridine moiety of (S)ꢀ37 is accommodated in a cavity formed by the residues Thr649,
Leu650, Thr655, Thr686, Tyr702, Leu703, Leu704 and Glu705 (Figure 4C). Two hydrogen bonds
are formed between the hydroxyl group of (S)ꢀ37 and the sideꢀchain hydroxyl groups of Thr686 and
Tyr702, as well as one hydrogen bond between the pyridine Nꢀatom of (S)ꢀ37 and the side chain
hydroxyl group of Thr655 (Figure 4B). The aromatic ring with the two chlorine atoms is located in
an area close to residues Glu402, Tyr405, Pro478, Met708 and Tyr732 (Figure 4B). The residue
Met708 is found in a conformation with the side chain pointing away from the binding cleft to avoid
steric repulsion. The presence of the two aromatic rings prohibits the interlobe hydrogen bond seen
in structures with agonists,19 between residues Glu402 in D1 and Thr686 in D2 (distance is 4.5 Å). A
molecule of glycerol is found in this area where it is involved in a hydrogenꢀbond bridge between
these two residues (Figure 4C).
The structure of GluA2 LBD with (R)-38. The structure of GluA2 LBD in complex with (R)ꢀ38
was solved at 1.8 Å resolution, and the complex crystallized in space group P212121 with four
molecules in the asymmetric unit of the crystal. In this structure, two nonꢀcrystallographic twoꢀfold
dimers are formed (molA/molC and molB/molD, respectively). For further information on data
collection and refinement statistics, see Table 3. (R)ꢀ38 induces full domain opening (20.2°, 20.4°,
19.2° and 18.8° for molA, molB, molC and molD, respectively) of the GluA2 LBD, relative to the
structure with Glu (PDB ID 1FTJ, molB19) (Figure 4D). The domain opening is similar to the
opening of the apo structure of GluA2 LBD (molA, 21.1°, PDB ID 1FTO19). The linkerꢀlinker
distances are 26.9 Å (dimer A/C) and 26.7 Å (dimer B/D) in accordance with full domain opening.20
Generally, the electron density is well defined and (R)ꢀ38 could unambiguously be modelled into the
density (Figure 4E). The aminoꢀacid part of (R)ꢀ38 forms contacts to bindingꢀsite residues in a
9
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 49
1
2
3
4
5
6
7
8
comparable manner to that of (S)ꢀ37 (Figures 4E and 4B). The biꢀcyclic aromatic system of (R)ꢀ38 is
accommodated in a new cavity close to the hinge region formed by the residues Thr480, Thr482,
Leu498, Ile500, Ser654, Glu705, Asp728 and Lys730 (Figure 4F). The hydroxyl group of (R)ꢀ38
makes two hydrogen bonds; one to the side chain carboxylate of Asp728 and one to a water
molecule. The presence of the two aromatic rings in the hinge region prevents domain closure by
steric hindrance and thereby the formation of a hydrogen bond between Glu402 in D1 and Thr686 in
D2 (distance is 6.5 Å).
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
Molecular docking studies
In order to analyze the observed structureꢀactivity relationship SiteMaps of ligand binding sites were
generated and docking studies carried out. The SiteMaps show that the amino acid parts of (S)ꢀ37
and (R)ꢀ38 fits well into the acceptor and donor regions (Figure 5). The chlorine atom in the 5ꢀ
position of the phenylalanine ring fits into a hydrophobic region. Moreover, there are small acceptor
and donor regions close to the hydroxyꢀpyridinyl part of (S)ꢀ37, representing hydrogen bonds
between Thr655, Thr686 and Tyr702 and the heteroatoms of the hydroxyꢀpyridinyl group. These
small acceptor and donor regions might explain why (S)ꢀ37 fits well into this pocket. In the SiteMap
of the ligand binding site of (R)ꢀ38, it is of notice that the phenoxy group of (R)ꢀ38 fits into a
hydrophobic region with no acceptor surface (Figure 5), indicating that (R)ꢀ37 would lack an
interaction with a hydrogen bonding donor for the hydroxyꢀpyridinyl nitrogen atom.
Next, all compounds listed in Table 1 were docked into the GluA2 LBD in complex with (S)ꢀ37 as
their Sꢀenantiomer (Table 4). Generally, compounds with measurable affinity have XPꢀscores that
are better than −9 kcal/mol. The inactive analogues 30, 34 and 35, with IC50 > 100 µM, show only
limited XPꢀscoring in docking experiments, most likely reflecting steric hindrance due to large
substituents (Table 1). The inactive 3ꢀmethoxyꢀsubstituted analogue 29, having only hydrogen bond
acceptor possibilities, shows medium XPꢀscores but high strain. The two hydroxymethyl substituted
analogues 31 and 32 are both inactive and display a relatively good score (–9.3 kcal/mol) and a very
10
ACS Paragon Plus Environment

Page 11 of 49
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
poor XPꢀscore (–1.3 kcal/mol), respectively. Considering the position of the hydroxyl group of 31
compared to 37, this compound will only be able to form one hydrogen bond since it is positioned
too far away from Tyr702. In addition, in order for the hydroxymethyl substituent to interact
favorably with Thr655 and/or Tyr702, the 5ꢀchloro substituent of 31 might need to be positioned less
favorably as compared to that of compound 37. Whereas the 3ꢀamino substituted analogue 33 in
principle should be able to interact with Thr655 and/or Tyr702 in a manner similar to that of
compound 37, differences in proteolytic properties of the aminoꢀphenyl substituent compared to that
of the hydroxypyridyl substituent might explain the inactivity of compound 33. The inactive pyridylꢀ
substituted analogue 36 seems to be able to form hydrogen bonding interactions with Thr655 but
only weak contacts to the side chain of Thr649 as the hydroxyl group of Thr649 already donates a
hydrogen bond to the backbone of Leu703. Finally, the two diꢀhydroxyl substituted analogues 27 and
28 were the best scoring compounds analyzed but showed no improved affinity compared to lead
compound 9. As indicated by the calculated strain energy (Table 4), this might be explained by a
strained hydrogen bonding geometry of the hydroxyl groups.
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
Discussion
In an effort to obtain competitive AMPA receptor antagonists capable of differentiating between the
individual AMPA receptor subtypes, we have synthesized a series of αꢀamino acidꢀbased
antagonists. As expected from previously published αꢀamino acidꢀbased antagonists,20 the
biologically active enantiomer of 37 was found to be in the Sꢀconfiguration, but in the case of the
structurally similar analogue 38, the antagonist activity exclusively was in the Rꢀenantiomer. To
address this difference in enantiopharmacology we determined the Xꢀray crystal structures of GluA2
LBD in complex with antagonists (S)ꢀ37 and (R)ꢀ38, respectively.
During recent years, a number of ligands have been coꢀcrystallized with the LBD of different
iGluRs, revealing that αꢀamino acidꢀbased ligands adopt similar binding mode of the αꢀamino acid
moiety for both agonists and competitive antagonists. A predominant number of these compounds
11
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 49
1
2
3
4
5
6
7
bind with their αꢀamino acid part in the Sꢀconfiguration,20 although a few examples of Rꢀconfigured
amino acidꢀbased ligands have been described, e.g. the agonists TDPA22 and 5ꢀHPCA24 as well as
the antagonists kaitocephalin23 and IKMꢀ159.25
8
9
Whereas (S)ꢀ37 binds in the expected manner with the aromatic part positioned in an area of the
GluA2 binding site often occupied by competitive antagonists (Figure 6A), (R)ꢀ38 was found to
adopt a new binding mode, in which the aromatic part of the ligand reaches into a cavity not
previously occupied by antagonists (Figure 6A). Interestingly, only two structural differences
between 37 and 38 (two hydrogen atoms in the phenylalanine ring of 38 instead of the dichloro
substitution in 37 as well as a change of a pyridine ring in 37 to phenyl in 38; Figure 2) led to the
opposite stereochemistry of the biologically active enantiomer and very different binding modes. The
pyridine nitrogen of 37 allows (S)ꢀ37 to form a hydrogen bond to the sideꢀchain hydroxyl group of
Thr655, whereas (S)ꢀ38 lacks the possibility to form this polar interaction. This might partly explain
why (S)ꢀ37 shows better affinity than (S)ꢀ38 in the (S)ꢀ37 binding mode. In addition, the phenyl
substituent of (S)ꢀ38 would not fit as well into a region of more polar character (Figure 5). On the
other hand, in the (R)ꢀ38 binding mode the two phenyl rings of (R)ꢀ38 fit well into the hydrophobic
region, whereas the pyridine ring of (R)ꢀ37 would loose a hydrogen bond. The other difference
between 37 and 38 is the presence of chlorine atoms in the phenylalanine ring (Figure 2). The
chlorine atom of (S)ꢀ37 in the 4ꢀposition of the phenylalanine ring forces the hydroxyꢀpyridinyl
substituent outꢀofꢀtheꢀplane, which has a marked influence on the orientation of the hydroxyl group.
Based on the Xꢀray structures it seems that this outꢀofꢀtheꢀplane twist is favorable for optimal
hydrogen bonding of the hydroxyꢀpyridinyl substituent of (S)ꢀ37. However, in the (R)ꢀ38 binding
mode, it seems likely that a corresponding outꢀofꢀtheꢀplane twist may force the hydroxyꢀaromatic
part of the target compound to come too close to the receptor backbone or place the key aromatic
hydroxyꢀsubstituent in an unfavorable position for interaction with Asp728. Moreover, the chlorine
atom in the 5ꢀposition of (S)ꢀ37 is involved in several favorable contacts to the nonꢀpolar part of the
(S)ꢀ37 binding site (Figure 5). The proteolytic properties of the phenolic hydroxyl groups may also
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
12
ACS Paragon Plus Environment

Page 13 of 49
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
contribute to the observed pharmacology of 37 and 38. Since the pKaꢀvalue of the hydroxyl group of
3ꢀhydroxypyridine is lower than that of phenol,30 the pKaꢀvalue of the hydroxyl group of 37 is
likewise lower than that of 38, with the two electronꢀwithdrawing chlorine atoms of 37 lowering the
pKa even more. Thus, 37 would be stabilized in the (S)ꢀ37 binding mode by hydrogen bonding
formation to the hydroxyl groups of Thr686 and Tyr702, whereas the interaction from (R)ꢀ37 to
Asp728 in the (R)ꢀ38 binding mode would be repulsive.
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
Design of subtype selective ligands has been pursued for decades, but despite that a large number of
AMPA receptor ligands have been synthesized and characterized pharmacologically, only few of
them have shown high AMPA receptor subunit selectivity. Previously, it has been discussed that
selectivity of agonists at AMPA receptors might be achieved through interaction of the agonist with
the sideꢀchain hydroxyl group of Tyr702 in GluA2 (Tyr698 in GluA1) directly or through water
molecules.31, 32 In GluA3 and GluA4 this residue is Phe, which cannot form sideꢀchain hydrogen
bonds. However, for antagonists the same trend has not been observed. The previously described
phenylalanineꢀbased antagonist, (S)ꢀ2ꢀaminoꢀ3ꢀ(2ꢀ(2ꢀcarboxyethyl)ꢀ5ꢀchloroꢀ4ꢀnitrophenyl)propioꢀ
nic acid, forms a waterꢀmediated contact to Tyr702 but the compound was not selective on AMPA
receptor subunits.13 The present antagonist (S)ꢀ37 interacts directly with Tyr702, but again no
subunit selectivity is observed for this compound (Table 2). Thus, whereas Tyr702 has been shown
to be important for agonist selectivity, it seems to be of minor importance in gaining subunit
selectivity of antagonists.
One of the most surprising observed differences between the two resolved crystal structures is the
degree of domain opening. Whereas (R)ꢀ38 leads to full domain opening (18.8ꢀ20.4°), the structure
of GluA2 LBD with (S)ꢀ37 shows the smallest antagonist induced domain opening of the GluA2
LBD (13.5°) seen to date. As (S)ꢀ37 is a functional antagonist, it suggests that compounds may
switch from partial agonists to antagonists when domain openings are in the range 11ꢀ13°.
13
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 49
1
2
Conclusion
3
4
5
6
7
8
9
In an effort to obtain competitive AMPA receptor antagonists capable of differentiating between the
individual AMPA receptor subtypes and to study the relationship between structure and function at
the molecular level, we have presented a new series of arylꢀsubstituted phenylalanines, and among
these identified two compounds with opposite stereochemistry, (S)ꢀ37 and (R)ꢀ38, as eutomers and
selective AMPA receptor antagonists with Kb of 1.80 ꢀM and 3.90 ꢀM at recombinant GluA2
AMPA receptors, respectively. Coꢀcrystallization of the two compounds with the GluA2 LBD,
followed by Xꢀray structure determination, revealed that (S)ꢀ37 and (R)ꢀ38 interact with GluA2 LBD
in quite different modes. (R)ꢀ38 interacts with the receptor in a new binding mode with the two
aromatic rings positioned close to the D1ꢀD2 hinge region. The new binding mode for amino acid
based antagonists might be a starting point for the future design of subtypeꢀselective AMPA receptor
antagonists. Whereas (S)ꢀ37 binds in a conventional manner in GluA2 LBD compared to most other
antagonists, it shows the smallest antagonist induced domain opening seen to date. Thus, our results
suggest that compounds may switch from partial agonists to antagonists at GluA2 receptors when
domain openings are in the range 11ꢀ13°.
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 Procedures. Analytical thin-layer chromatography (TLC) was performed on silica gel 60
F254 aluminium plates (Merck). All compounds were detected using UV light and a KMnO4 spraying
reagent, and the target amino acids were also visualized using a ninhydrin spraying reagent. Flash
column chromatography (FC) was performed using the CombiFlash Companion System (Teledyne
Isco, Inc.) on RediSep columns with silica gel (average particle size 35 to 70 ꢀm). A mixture of
1
heptane and ethyl acetate (both HPLC purity) was used as eluents, unless otherwise stated. H NMR
spectra were recorded on a 300 MHz Varian Mercury 300BB or 300 MHz Varian Gemini 2000BB
13
spectrometer. C NMR spectra were recorded on the Varian Gemini spectrometer. CDCl3, D2O and
14
ACS Paragon Plus Environment

Page 15 of 49
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
dimethyl-d6-sulfoxide (DMSO-d6) were used as solvents. Chemical shifts (δ) are given in parts per
million (ppm), and coupling constants (J) are given in Hertz (Hz). GC-MS was recorded on a
Shimadzu QP 5050A mounted with a Supelco MDN-5S column and chemical ionization was used.
Elemental analyses were performed by J. Theiner, Microanalytical Laboratory, Department of
Physical Chemistry, University of Vienna, Austria. All target compounds possessed a purity of ≥95%
as verified by elemental analyses by comparison with the theoretical values or by UPLC analysis
(compound 33).
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
Unless otherwise stated, preparative chromatography was performed using HPLC system consisting
of a Jasco 880-PU pump, a Rheodyne 7125 injector equipped with a 5.0 mL sample loop, a
Shimadzu SPD-6A detector (220 nm or 250 nm) and a Merck-Hitachi D-2000 Chromato-Integrator.
Purification of the target racemic amino acids was performed on the reverse-phase XTerra MS C18
column (10 × 300 mm, 10 ꢁm, Waters). The column was eluted at 8 mL/min with 15 mM acetic acid
containing methanol in varying concentrations depending on the structure of the target amino acid.
UPLC analysis was performed using a Waters Acquity UPLC system equipped with a dual
wavelength UV detector (observing at 215 and 254 nm) and a QDa single quadropole MS detector
operated in positive electro spray mode (ES+) using a Waters Acquity UPLC BEH C18 column (1.7
µm; 2.1 × 50 mm). The column was eluted at 0.3 mL/min with a linear gradient starting at 5%
MeCN (aq) to 100% MeCN over 2 minutes followed by one minute of 100% MeCN (total run time 3
min). The mobile phases were supplemented with 0.1 % formic acid.
Preparative chiral separation was performed using Chirobiotic T column (10 × 500 mm, ASTEC)
equipped with a Chirobiotic T guard column (4.6 × 50 mm, ASTEC) and connected to the HPLC
system described above (detection at 250 nm). Ethanol:water (80:20) was used as mobile phase at
1.5 mL/min. The enantiomeric purity expressed by the enantiomeric excess (ee) was determined by
chiral HPLC using an analytical Chirobiotic T column (4.6 × 150 mm, ASTEC) equipped with a
Chirobiotic T guard column (4.6 × 50 mm, ASTEC) and connected to the HPLC system described
above (detection at 250 nm). Elution was performed with 60% (v/v) ethanol in water adjusted with
15
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 49
1
2
3
4
5
6
7
8
acetic acid to pH = 4 at 0.5 mL/min. Analyses of elution orders were carried out using a chiral
ligandꢀexchange column containing (R)ꢀPro as chiral selector and connected to the HPLC system
described above (detection at 215 nm). The column was thermostated at 50 °C and eluted with an
aqueous solution of KH2PO4 (50 mM, pH = 4.5) at 1.0 mL/min. Optical rotation was measured on a
Perkin-Elmer 241 polarimeter. CD spectra were recorded at ambient temperature in 1.0 cm cuvettes
on an OLIS DSM (Digital Subtractive Method) 10 CD spectrophotometer. Compounds 13 and 14
were prepared as described previously.26
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
General procedure for diethyl 2-acetamido-2-(biphenyl-3-ylmethyl)malonates (15-23, 26) and
diethyl 2-acetamido-2-(3-(pyridin-3-yl)benzyl)malonates (24, 25). Compounds 15−26 were
prepared in analogy to the known literature method.33 Suspension of 13 (502 mg, 1 mmol) or 14 (433
mg, 1 mmol) and bis(triphenylphosphine)palladium(II) dichloride PdCl2(PPh3)2 (35 mg, 0.05 mmol)
was stirred in DME (25 mL) under nitrogen atmosphere at room temperature for 15 min. The
adequate boronic acid (or boronic acid pinacol ester) (1.5 mmol), triethylamine (5.58 mL, 40 mmol)
and water (25 mL) were added and the mixture was stirred under nitrogen at room temperature for
several hours, controlled by GC-MS. If the starting material was not converted after 24 h of stirring,
the temperature was raised to 50°C. When the reaction was finished, the mixture was filtrated and
water (30 mL) was added. After extraction with ethyl acetate (3×50 mL) the combined extracts were
washed with 2M NaOH (50 mL) following water (50 mL), dried (MgSO4) end evaporated. The raw
product was purified by flash column chromatography.
Diethyl 2ꢀacetamidoꢀ2ꢀ((5,6ꢀdichloroꢀ3',4'ꢀdimethoxybiphenylꢀ3ꢀyl)methyl)malonate (15). Yield
1
0.372 g, 73%; H NMR (300 MHz, CDCl3): δ 1.28 (t, J = 7.2 Hz, 6H, OCH2CH3), 2.03 (s, 3H,
NHCOCH3), 3.63 (s, 2H, ArCH2), 3.88 (s, 3H, OCH3), 3.92 (s, 3H, OCH3), 4.27 (q, J = 7.2 Hz, 4H,
OCH2CH3), 6.57 (s, 1H, NH), 6.85 (d, J = 2.2 Hz, 1H, Ar), 6.87ꢀ6.91 (m, 3H, Ar), 7.08 (d, J =
2.2 Hz, 1H, Ar); 13C NMR (75 MHz, CDCl3): δ 14.5, 23.5, 37.3, 56.2, 56.3, 63.3, 67.3, 111.0, 112.7,
121.7, 130.1, 130.7, 131.3, 131.9, 133.6, 135.0, 142.5, 148.6, 149.0, 167.4, 169.4.
16
ACS Paragon Plus Environment

Page 17 of 49
Journal of Medicinal Chemistry
1
2
Diethyl 2ꢀacetamidoꢀ2ꢀ((5,6ꢀdichloroꢀ3',5'ꢀdimethoxybiphenylꢀ3ꢀyl)methyl)malonate (16). Yield
3
4
5
1
0.423 g, 83%; H NMR (300 MHz, CDCl3): δ 1.28 (t, J = 7.2 Hz, 6H, OCH2CH3), 2.03 (s, 3H,
6
7
8
9
NHCOCH3), 3.63 (s, 2H, ArCH2), 3.80 (s, 3H, OCH3), 3.83 (s, 3H, OCH3), 4.26 (q, J = 7.2 Hz, 4H,
OCH2CH3), 6.44 (d, J = 2.2 Hz, 2H, Ar), 6.47 (t, J = 2.2 Hz, 1H, Ar), 6.58 (s, 1H, NH), 6.85 (d, J =
2.2 Hz, 1H, Ar), 7.09 (d, J = 1.9 Hz, 1H, Ar); 13C NMR (75 MHz, CDCl3): δ 14.5, 23.5, 37.3, 55.8,
63.3, 67.3, 100.3, 107.6, 121.7, 130.2, 130.9, 131.0, 133.5, 135.0, 141.1, 142.7, 160.6, 167.4, 169.5.
Diethyl 2ꢀacetamidoꢀ2ꢀ((5,6ꢀdichloroꢀ3'ꢀmethoxybiphenylꢀ3ꢀyl)methyl)malonate (17). Yield 0.357 g,
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
74%; H NMR (300 MHz, CDCl3): δ 1.28 (t, J = 7.2 Hz, 6H, OCH2CH3), 2.03 (s, 3H, NHCOCH3),
3.63 (s, 2H, ArCH2), 3.82 (s, 3H, OCH3), 4.26 (dq, J1 = 7.2 Hz, J2 = 1.4 Hz, 4H, OCH2CH3), 6.58 (s,
1H, NH), 6.85-6.93 (m, 3H, Ar), 7.10 (d, J = 1.9 Hz, 1H, Ar), 7.24 (s, 1H, Ar), 7.32 (t, J = 7.7 Hz,
1H, Ar); 13C NMR (75 MHz, CDCl3): δ 14.3, 23.3, 37.1, 55.5, 63.1, 67.1, 113.6, 114.9, 121.5, 129.3,
130.1, 130.7, 130.9, 133.5, 134.9, 140.3, 142.4, 159.2, 167.2, 169.3.
Diethyl 2ꢀacetamidoꢀ2ꢀ(3,4ꢀdichloroꢀ5ꢀ(2,3ꢀdihydrobenzo[b][1,4]dioxinꢀ6ꢀyl)benzyl)malonate (18).
Yield 0.416 g, 81%; 1H NMR (300 MHz, CDCl3): δ 1.29 (t, J = 7.2 Hz, 6H, OCH2CH3), 2.04 (s, 3H,
NHCOCH3), 3.63 (s, 2H, ArCH2), 4.23-4.34 (m, 8H, OCH2CH3, CH2 dioxane), 6.58 (s, 1H, NH),
6.79 (dd, J1 = 8.3 Hz, J2 = 1.9 Hz, 1H, Ar), 6.83 (d, J = 2.2 Hz, 1H, Ar), 7.85 (d, J = 1.9 Hz, 1H,
Ar), 6.89 (d, J = 8.3 Hz, 1H, Ar), 7.07 (d, J = 1.9 Hz, 2H, Ar); 13C NMR (75 MHz, CDCl3): δ 14.5,
23.5, 37.3, 63.3, 64.7, 64.8, 67.3, 117.3, 118.4, 122.5, 130.3, 130.6, 131.3, 132.5, 133.5, 135.0,
142.2, 143.3, 143.7, 167.4, 169.5.
Diethyl 2ꢀacetamidoꢀ2ꢀ((5,6ꢀdichloroꢀ3'ꢀ(hydroxymethyl)biphenylꢀ3ꢀyl)methyl)malonate (19). Yield
1
0.366 g, 76%; H NMR (300 MHz, CDCl3): δ 1.28 (t, J = 7.1 Hz, 6H, OCH2CH3), 1.81 (t, J = 5.9
Hz, 1H, CH2OH), 2.03 (s, 3H, NHCOCH3), 3.64 (s, 2H, ArCH2), 4.27 (q, J = 7.1 Hz, 4H,
OCH2CH3), 4.75 (d, J = 5.9 Hz, 2H, CH2OH), 6.61 (s, 1H, NH), 6.87 (d, J = 2.1 Hz, 1H, Ar), 7.12
(d, J = 2.1 Hz, 1H, Ar), 7.26ꢀ7.29 (m, 1H, Ar), 7.34ꢀ7.45 (m, 3H, Ar); 13C NMR (75 MHz, CDCl3):
δ 14.5, 23.5, 37.3, 63.4, 65.5, 67.3, 126.2, 126.8, 127.9, 128.6, 129.3, 130.9, 131.3, 133.6, 135.2,
139.5, 141.2, 142.6, 167.4, 169.5.
17
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 49
1
2
Diethyl 2ꢀacetamidoꢀ2ꢀ((5,6ꢀdichloroꢀ4'ꢀ(methoxymethyl)biphenylꢀ3ꢀyl)methyl)malonate (20). Yield
3
4
5
1
0.433 g, 87%; H NMR (300 MHz, CDCl3): δ 1.28 (t, J = 7.2 Hz, 6H, OCH2CH3), 2.04 (s, 3H,
6
7
8
9
NHCOCH3), 3.46 (s, 3H, OCH3), 3.65 (s, 2H, ArCH2), 4.27 (dq, J1 = 7.2 Hz, J2 = 2.2 Hz, 4H,
OCH2CH3), 4.51 (s, 2H, CH2OCH3), 6.59 (s, 1H, NH), 6.85 (d, J = 2.2 Hz, 1H, Ar), 7.12 (d, J = 1.9
Hz, 1H, Ar), 7.31 (d, J = 8.3 Hz, 1H, Ar), 7.39 (d, J = 8.3 Hz, 1H, Ar); 13C NMR (75 MHz, CDCl3):
δ 14.3, 23.3, 37.1, 58.6, 63.1, 67.1, 74.5, 127.5, 129.2, 130.1, 130.7, 131.0, 133.3, 134.9, 138.1,
138.3, 142.3, 167.1, 169.3.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Diethyl 2ꢀacetamidoꢀ2ꢀ((3'ꢀaminoꢀ5,6ꢀdichlorobiphenylꢀ3ꢀyl)methyl)malonate (21). Yield 0.381 g,
1
82%; H NMR (300 MHz, CDCl3): δ 1.27 (t, J = 7.2 Hz, 6H, OCH2CH3), 2.02 (s, 3H, NHCOCH3),
3.62 (s, 2H, ArCH2), 3.79 (br.s, 2H, NH2), 4.26 (dq, J1 = 7.2 Hz, J2 = 1.2 Hz, 4H, OCH2CH3), 6.64ꢀ
6.70 (m, 4H, NH, Ar), 6.86 (d, J = 2.1 Hz, 1H, Ar), 7.10 (d, J = 2.1 Hz, 1H, Ar), 7.18 (t, J = 7.8 Hz,
1H, Ar); 13C NMR (75 MHz, CDCl3): δ 14.4, 23.3, 37.2, 63.2, 67.2, 114.9, 115.9, 119.4, 129.3,
130.1, 130.6, 131.0, 133.3, 134.9, 140.2, 142.9, 146.3, 167.1, 169.3.
Diethyl 2ꢀacetamidoꢀ2ꢀ((5,6ꢀdichloroꢀ3'ꢀsulfamoylbiphenylꢀ3ꢀyl)methyl)malonate (22). Yield
1
0.281 g, 53%; H NMR (300 MHz, CDCl3): δ 1.28 (t, J = 7.1 Hz, 6H, OCH2CH3), 2.05 (s, 3H,
NHCOCH3), 3.66 (s, 2H, ArCH2), 4.28 (q, J = 7.1 Hz, 4H, OCH2CH3), 4.94 (br.s, 2H, SO2NH2),
6.63 (s, 1H, NH), 6.88 (d, J = 1.8 Hz, 1H, Ar), 7.17 (d, J = 1.8 Hz, 1H, Ar), 7.57ꢀ7.61 (m, 2H, Ar),
7.89 (s, 1H, Ar), 7.96 (d, J = 6.5 Hz, 1H, Ar); 13C NMR (75 MHz, CDCl3): δ 14.3, 23.3, 37.1, 63.3,
67.1, 126.1, 127.2, 129.1, 129.9, 131.0, 131.5, 133.6, 133.7, 135.4, 140.0, 140.6, 142.2, 167.1, 169.6.
Diethyl 2ꢀacetamidoꢀ2ꢀ((5,6ꢀdichloroꢀ4'ꢀsulfamoylbiphenylꢀ3ꢀyl)methyl)malonate (23). Yield
1
0.336 g, 63%; H NMR (300 MHz, CDCl3): δ 1.28 (t, J = 7.2 Hz, 6H, OCH2CH3), 2.03 (s, 3H,
NHCOCH3), 3.66 (s, 2H, ArCH2), 4.27 (q, J = 7.2 Hz, 4H, OCH2CH3), 4.92 (br.s, 2H, SO2NH2),
6.60 (s, 1H, NH), 6.83 (d, J = 1.8 Hz, 1H, Ar), 7.18 (d, J = 1.8 Hz, 1H, Ar), 7.49 (d, J = 8.2 Hz, 2H,
Ar), 7.99 (d, J = 8.2 Hz, 2H, Ar); 13C NMR (75 MHz, CDCl3): δ 14.3, 23.3, 37.1, 63.2, 67.1, 126.5,
129.9, 130.0, 130.7, 131.7, 133.8, 135.4, 140.7, 141.5, 143.5, 167.1, 169.4.
18
ACS Paragon Plus Environment

Page 19 of 49
Journal of Medicinal Chemistry
1
2
Diethyl 2ꢀacetamidoꢀ2ꢀ(3,4ꢀdichloroꢀ5ꢀ(6ꢀmethoxypyridinꢀ3ꢀyl)benzyl)malonate (24). Yield 0.426 g,
3
4
5
1
88%; H NMR (300 MHz, CDCl3): δ 1.30 (t, J = 7.2 Hz, 6H, OCH2CH3), 2.06 (s, 3H, NHCOCH3),
6
7
8
9
3.66 (s, 2H, ArCH2), 4.01 (s, 3H, OCH3), 4.28 (dq, J1 = 7.2 Hz, J2 = 2.5 Hz, 4H, OCH2CH3), 6.00 (s,
1H, NH), 6.81ꢀ6.85 (m, 2H, Ar, pyridine), 7.14 (d, J = 1.7 Hz, 1H, Ar), 7.62 (dd, J1 = 8.5 Hz, J2 =
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
13
2.5 Hz, 1H, pyridine), 8.13 (d, J = 2.2 Hz, 1H, pyridine); C NMR (75 MHz, CDCl3): δ 14.3, 23.3,
37.1, 53.8, 63.2, 67.1, 110.4, 128.0, 130.5, 130.9, 131.1, 133.6, 135.2, 139.1, 139.5, 146.6, 163.7,
167.1, 169.3.
Diethyl 2ꢀacetamidoꢀ2ꢀ(3,4ꢀdichloroꢀ5ꢀ(5ꢀmethoxypyridinꢀ3ꢀyl)benzyl)malonate (25). Yield 0.362 g,
1
75%; H NMR (300 MHz, CDCl3): δ 1.30 (t, J = 7.2 Hz, 6H, OCH2CH3), 2.06 (s, 3H, NHCOCH3),
3.68 (s, 2H, ArCH2), 3.92 (s, 3H, OCH3), 4.29 (dq, J1 = 7.2 Hz, J2 = 2.5 Hz, 4H, OCH2CH3), 6.61 (s,
1H, NH), 6.87 (d, J = 1.9 Hz, 1H, Ar), 7.18 (d, J = 2.2 Hz, 1H, Ar), 7.53ꢀ7.57 (m, 1H, pyridine),
8.18 (d, J = 1.7 Hz, 1H, pyridine), 8.34 (d, J = 2.8 Hz, 1H, pyridine); 13C NMR (75 MHz, CDCl3): δ
14.5, 23.5, 37.3, 56.1, 63.4, 67.3, 121.5, 130.6, 131.1, 131.8, 133.9, 135.4, 135.6, 137.3, 138.9,
142.0, 155.2, 167.3, 169.5.
1
Diethyl 2ꢀacetamidoꢀ2ꢀ((3'ꢀmethoxybiphenylꢀ3ꢀyl)methyl)malonate (26). Yield 0.367 g, 89%; H
NMR (300 MHz, CDCl3): δ 1.31 (t, J = 7.2 Hz, 6H, OCH2CH3), 2.04 (s, 3H, NHCOCH3), 3.72 (s,
2H, ArCH2), 3.86 (s, 3H, OCH3), 4.29 (dq, J1 = 7.2 Hz, J2 = 1.7 Hz, 4H, OCH2CH3), 6.58 (s, 1H,
NH), 6.88 (dd, J1 = 8.0 Hz, J2 = 2.2 Hz, 1H, Ar), 7.00 (d, J = 7.4 Hz, 1H, Ar), 7.05 (t, J = 2.2 Hz,
1H, Ar), 7.10 (d, J = 7.7 Hz, 1H, Ar), 7.21 (s, 1H, Ar), 7.29ꢀ7.36 (m, 2H, Ar), 7.46 (d, J = 7.7 Hz,
1H, Ar); 13C NMR (75 MHz, CDCl3): δ 14.3, 23.3, 38.0, 55.4, 62.9, 67.3, 112.7, 112.8, 119.5, 126.1,
128.7, 128.8, 128.9, 129.8, 135.7, 141.1, 142.4, 159.9, 167.5, 169.1.
General procedure for preparation of compounds 2-amino-3-(biphenyl-3-yl)propanoic acids (27-
35, 38) and 2-amino-3-(3-(pyridin-3-yl)phenyl)propanoic acids (36, 37). A mixture of 15, 16, 19-26
(0.5 mmol) in 48% aq HBr (6 mL) was refluxed for 2 h. The reaction mixture was evaporated under
reduced pressure and purified by reverse-phase HPLC (mobile phase: a mixture of methanol in 15
mM acetic acid) to give white crystals of the final amino acid.
19
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 49
1
2
3
4
5
6
7
8
Compounds 17 and 18 were refluxed in a mixture of HCl:water:acetic acid (1:1:1), deprotection was
controlled by TLC. After evaporation of solvents the amino acids were purified by reverse-phase
HPLC (mobile phase: a mixture of methanol in 15 mM acetic acid) to give white crystals of 29 and
30, respectively.
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
(RS)ꢀ2ꢀAminoꢀ3ꢀ(5,6ꢀdichloroꢀ3',4'ꢀdihydroxybiphenylꢀ3ꢀyl)propanoic acid (27). Yield 88 mg, 51%;
1H NMR (300 MHz, DMSO-d6 + 50 ꢁL D2O): δ 2.92 (dd, J1 = 14.4 Hz, J2 = 7.4 Hz, 1H, ArCH2),
3.10 (dd, J1 = 14.4 Hz, J2 = 4.1 Hz, 1H, ArCH2), 3.50-3.54 (m, 1H, ArCH2CH), 6.69 (d, J = 7.9 Hz,
1H, Ar), 6.80 (dd, J1 = 8.2 Hz, J2 = 1.8 Hz, 1H, Ar), 6.83ꢀ6.84 (m, 1H, Ar), 7.18 (s, 1H, Ar), 7.44 (s,
1H, Ar).
(RS)ꢀ2ꢀAminoꢀ3ꢀ(5,6ꢀdichloroꢀ3',5'ꢀdihydroxybiphenylꢀ3ꢀyl)propanoic acid (28). Yield 82 mg, 48%;
1H NMR (300 MHz, DMSO-d6 + 50 ꢁL D2O): δ 2.92 (dd, J1 = 14.3 Hz, J2 = 7.7 Hz, 1H, ArCH2),
3.10 (dd, J1 = 14.4 Hz, J2 = 4.4 Hz, 1H, ArCH2), 3.51 (dd, J1 = 7.7 Hz, J2 = 4.4 Hz, 1H, ArCH2CH),
6.22ꢀ6.24 (m, 3H, Ar), 7.16 (d, J = 1.9 Hz, 1H, Ar), 7.46 (d, J = 1.9 Hz, 1H, Ar).
(RS)ꢀ2ꢀAminoꢀ3ꢀ(5,6ꢀdichloroꢀ3'ꢀmethoxybiphenylꢀ3ꢀyl)propanoic acid (29). Yield 129 mg, 76%; 1H
NMR (300 MHz, DMSO-d6 + 50 ꢁL D2O): δ 2.96 (dd, J1 = 14.4 Hz, J2 = 7.3 Hz, 1H, ArCH2), 3.14
(dd, J1 = 14.4 Hz, J2 = 4.4 Hz, 1H, ArCH2), 3.51-3.58 (m, 1H, ArCH2CH), 3.74 (s, 3H, OCH3), 6.90ꢀ
7.00 (m, 3H, Ar), 7.20 (d, J = 2.1 Hz, 1H, Ar), 7.35 (t, J = 7.8 Hz, 1H, Ar), 7.48 (d, J = 1.8 Hz, 1H,
Ar).
(RS)ꢀ2ꢀAminoꢀ3ꢀ(3,4ꢀdichloroꢀ5ꢀ(2,3ꢀdihydrobenzo[b][1,4]dioxinꢀ6ꢀyl)phenyl)propanoic acid (30).
1
Yield 138 mg, 75%; H NMR (300 MHz, DMSO-d6 + 50 ꢁL D2O): δ 2.92 (dd, J1 = 14.3 Hz, J2 =
7.7 Hz, 1H, ArCH2), 3.11 (dd, J1 = 14.3 Hz, J2 = 3.9 Hz, 1H, ArCH2), 3.54 (dd, J1 = 7.7 Hz, J2 = 4.4
Hz, 1H, ArCH2CH), 6.85ꢀ6.92 (m, 3H, Ar), 7.17 (d, J = 1.9 Hz, 1H, Ar), 7.45 (d, J = 1.9 Hz, 1H,
Ar).
(RS)ꢀ2ꢀAminoꢀ3ꢀ(5,6ꢀdichloroꢀ3'ꢀ(hydroxymethyl)biphenylꢀ3ꢀyl)propanoic acid (31). Yield 102 mg,
1
60%; H NMR (300 MHz, DMSO-d6 + 50 ꢁL D2O): δ 2.94 (dd, J1 = 15.3 Hz, J2 = 7.3 Hz, 1H,
20
ACS Paragon Plus Environment

Page 21 of 49
Journal of Medicinal Chemistry
1
2
3
4
ArCH2), 3.10ꢀ3.14 (m, 1H, ArCH2), 3.54-3.56 (m, 1H, ArCH2CH), 4.52 (s, 2H, CH2OH), 7.21 (s,
1H, Ar), 7.27ꢀ7.43 (m, 4H, Ar), 7.50 (s, 1H, Ar).
5
6
7
(RS)ꢀ2ꢀAminoꢀ3ꢀ(5,6ꢀdichloroꢀ4'ꢀ(hydroxymethyl)biphenylꢀ3ꢀyl)propanoic acid (32). Yield 140 mg,
8
9
1
82%; H NMR (300 MHz, DMSO-d6 + 50 ꢁL D2O): δ 2.94 (dd, J1 = 14.2 Hz, J2 = 7.8 Hz, 1H,
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
ArCH2), 3.11ꢀ3.15 (m, 1H, ArCH2), 3.49-3.52 (m, 1H, ArCH2CH), 4.53 (s, 2H, CH2OH), 7.24 (s,
1H, Ar), 7.39ꢀ7.40 (m, 4H, Ar), 7.52 (s, 1H, Ar).
1
(RS)ꢀ2ꢀAminoꢀ3ꢀ(3'ꢀaminoꢀ5,6ꢀdichlorobiphenylꢀ3ꢀyl)propanoic acid (33). Yield 60 mg, 37%; H
NMR (300 MHz, DMSO-d6 + 50 ꢁL D2O): δ 2.92 (dd, J1 = 14.3 Hz, J2 = 7.7 Hz, 1H, ArCH2), 3.11
(dd, J1 = 14.3 Hz, J2 = 4.4 Hz, 1H, ArCH2), 3.50 (dd, J1 = 7.7 Hz, J2 = 4.4 Hz, 1H, ArCH2CH), 6.52
(d, J = 7.7 Hz, 1H, Ar), 6.57ꢀ6.59 (m, 2H, Ar), 7.07 (t, J = 7.7 Hz, 1H, Ar), 7.16 (d, J = 1.9 Hz, 1H,
Ar), 7.46 (d, J = 1.9 Hz, 1H, Ar).
(RS)ꢀ2ꢀAminoꢀ3ꢀ(5,6ꢀdichloroꢀ3'ꢀsulfamoylbiphenylꢀ3ꢀyl)propanoic acid (34). Yield 139 mg, 71%;
1H NMR (300 MHz, DMSO-d6 + 50 ꢁL D2O): δ 2.96 (dd, J1 = 14.1 Hz, J2 = 7.3 Hz, 1H, ArCH2),
3.13 (dd, J1 = 14.4 Hz, J2 = 4.4 Hz, 1H, ArCH2), 3.50-3.54 (m, 1H, ArCH2CH), 7.28 (s, 1H, Ar),
7.59 (s, 1H, Ar), 7.66ꢀ7.69 (m, 2H, Ar), 7.86ꢀ7.88 (m, 2H, Ar).
(RS)ꢀ2ꢀAminoꢀ3ꢀ(5,6ꢀdichloroꢀ4'ꢀsulfamoylbiphenylꢀ3ꢀyl)propanoic acid (35). Yield 169 mg, 87%;
1H NMR (300 MHz, DMSO-d6 + 50 ꢁL D2O): δ 2.96 (dd, J1 = 14.4 Hz, J2 = 7.3 Hz, 1H, ArCH2),
3.14 (dd, J1 = 14.4 Hz, J2 = 4.4 Hz, 1H, ArCH2), 3.510-3.58 (m, 1H, ArCH2CH), 7.28 (s, 1H, Ar),
7.57 (s, 1H, Ar), 7.63 (dd, J1 = 8.2 Hz, J2 = 1.8 Hz, 2H, Ar), 7.90 (dd, J1 = 8.2 Hz, J2 = 1.8 Hz, 2H,
Ar).
(RS)ꢀ2ꢀAminoꢀ3ꢀ(3,4ꢀdichloroꢀ5ꢀ(6ꢀhydroxypyridinꢀ3ꢀyl)phenyl)propanoic acid (36). Yield 143 mg,
1
88%; H NMR (300 MHz, DMSO-d6 + 50 ꢁL D2O): δ 2.93 (dd, J1 = 14.3 Hz, J2 = 7.7 Hz, 1H,
ArCH2), 3.11 (dd, J1 = 14.3 Hz, J2 = 4.4 Hz, 1H, ArCH2), 3.53 (dd, J1 = 7.7 Hz, J2 = 4.4 Hz, 1H,
ArCH2CH), 6.44 (d, J = 9.4 Hz, 1H, pyridine), 7.23 (d, J = 2.2 Hz, 1H, Ar), 7.47 (d, J = 1.9 Hz, 1H,
Ar), 7.50 (d, J = 2.2 Hz, 1H, pyridine), 7.60 (dd, J1 = 9.6 Hz, J2 = 2.8 Hz, 1H, pyridine).
21
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 49
1
2
(RS)ꢀ2ꢀAminoꢀ3ꢀ(3,4ꢀdichloroꢀ5ꢀ(5ꢀhydroxypyridinꢀ3ꢀyl)phenyl)propanoic acid (37). Yield 143 mg,
3
4
5
1
88%; H NMR (300 MHz, DMSO-d6 + 50 ꢁL D2O): δ 2.94 (dd, J1 = 14.1 Hz, J2 = 7.5 Hz, 1H,
6
7
ArCH2), 3.12 (dd, J1 = 14.1 Hz, J2 = 4.2 Hz, 1H, ArCH2), 3.55ꢀ3.58 (m, 1H, ArCH2CH), 7.21ꢀ7.25
8
9
(m, 2H, Ar, pyridine), 7.53 (s, 1H, Ar), 8.04 (s, 1H, pyridine), 8.12 (s 1H, pyridine).
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
(S)ꢀ(+)ꢀ and (R)ꢀ(
−)ꢀ2ꢀaminoꢀ3ꢀ(3,4ꢀdichloroꢀ5ꢀ(5ꢀhydroxypyridinꢀ3ꢀyl)phenyl)propanoic acid [(S)ꢀ
(+)ꢀ and (R)ꢀ( )ꢀ(37)]. Racemic 37 dissolved in EtOH: H2O (1:1) (110 mg; 10 mg/mL) was
−
separated by chiral HPLC using a semiꢀpreparative Chirobiotic T column in 10 mg injections.
Fractions containing the firstꢀeluting (+)ꢀ37 isomer were pooled, evaporated and recrystallized from
water to give (+)ꢀ37 (27 mg, 49%); mp 234 – 249 °C (dec.); [α]D20 +19.3 [c 0.22, EtOH:HCl(aq) (0.25
1
M) (1:1)]; ee > 97%. H NMR (300 MHz, DMSO-d6): δ 2.94 (dd, J1 = 14.1 Hz, J2 = 7.2 Hz, 1H,
ArCH2), 3.08 (dd, J1 = 14.1 Hz, J2 = 4.5 Hz, 1H, ArCH2), 3.51 (dd, J1 = 7.2 Hz, J2 = 4.5 Hz, 1H,
ArCH2CH), 7.22 (dd, J1 = 2.7 Hz, J2 = 1.8 Hz, 1H, Ar), 7.27 (d, J = 1.8 Hz, 1H, Ar), 7.56 (d, J =
1.8 Hz, 1H, Ar), 8.06 (d, J = 1.5 Hz, 1H, Ar), 8.15 (d, J = 2.4 Hz, 1H, Ar).
Fractions containing the secondꢀeluting (−)ꢀ37 isomer were pooled, evaporated and recrystallized
from water to give (−)ꢀ37 (34 mg, 62%); mp 236 – 242 °C (dec.); [α]D22 −19.2 [c 0.22, EtOH:HCl(aq)
(0.25M) (1:1)]; ee = 99.7%. 1H NMR (300 MHz, DMSO-d6) was identical with that of (+)ꢀ37.
1
(RS)ꢀ2ꢀAminoꢀ3ꢀ(3'ꢀhydroxybiphenylꢀ3ꢀyl)propanoic acid (38). Yield 92 mg, 52%; H NMR (300
MHz, DMSO-d6 + 50 ꢁL D2O): δ 2.93 (dd, J1 = 14.3 Hz, J2 = 8.3 Hz, 1H, ArCH2), 3.19 (dd, J1 =
14.2 Hz, J2 = 3.9 Hz, 1H, ArCH2), 3.52 (dd, J1 = 8.3 Hz, J2 = 4.1 Hz, 1H, ArCH2CH), 6.73 (dd, J1 =
8.0 Hz, J2 = 1.4 Hz, 1H, Ar), 7.00ꢀ7.01 (m, 1H, Ar), 7.06 (d, J = 8.0 Hz, 1H, Ar), 7.20ꢀ7.25 (m, 2H,
Ar), 7.34 (t, J = 7.7 Hz, 1H, Ar), 7.43 (d, J = 8.0 Hz, 1H, Ar), 7.47 (s, 1H, Ar).
General procedure for preparation of (S)- and (R)-ethyl 2-(tert-butoxycarbonylamino)-3-(3'-
methoxybiphenyl-3-yl)propanoates (41, 42). The compounds 41 and 42 were prepared from Sꢀ and
Rꢀisomers of ethyl 3ꢀ(3ꢀbromophenyl)ꢀ2ꢀ(tertꢀbutoxycarbonylamino)propanoate (242 mg,
0.65 mmol) and 3-methoxyphenylboronic acid according to the procedure described for compounds
15-26.
22
ACS Paragon Plus Environment

Page 23 of 49
Journal of Medicinal Chemistry
1
2
(S)ꢀEthyl 2ꢀ(tertꢀbutoxycarbonylamino)ꢀ3ꢀ(3'ꢀmethoxybiphenylꢀ3ꢀyl)propanoate (41). Yield 228 mg,
3
4
5
1
89%; H NMR (300 MHz, CDCl3): δ 1.22 (t, J = 7.2 Hz, 3H, OCH2CH3), 1.41 (s, 9H, CH3), 3.15-
6
7
8
9
3.17 (m, 2H, ArCH2), 3.87 (s, 3H, OCH3) 4.16 (q, J = 7.1 Hz, 2H, OCH2CH3), 4.61 (d, J = 7.1 Hz,
1H, CHNH), 5.03 (d, J = 7.1 Hz, NH), 6.89ꢀ6.91 (m, 1H, Ar), 7.10ꢀ7.16 (m, 3H, Ar), 7.32ꢀ7.38 (m,
3H, Ar), 7.46 (d, J = 7.6 Hz, 1H, Ar); 13C NMR (75 MHz, CDCl3): δ 14.4, 28.5, 38.6, 54.6, 55.5,
61.6, 80.1, 112.8, 112.9, 119.9, 125.9, 128.4, 128.4, 128.9, 129.8, 136.6, 141.8, 142.5, 155.5, 159.9,
171.8.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(R)ꢀEthyl 2ꢀ(tertꢀbutoxycarbonylamino)ꢀ3ꢀ(3'ꢀmethoxybiphenylꢀ3ꢀyl)propanoate (42). Yield 221 mg,
1
86%; H NMR (300 MHz, CDCl3): δ 1.22 (t, J = 7.2 Hz, 3H, OCH2CH3), 1.41 (s, 9H, CH3), 3.13-
3.17 (m, 2H, ArCH2), 3.87 (s, 3H, OCH3), 4.16 (q, J = 7.1 Hz, 2H, OCH2CH3), 4.60 (d, J = 7.1 Hz,
1H, CHNH), 5.03 (d, J = 7.1 Hz, NH), 6.88ꢀ6.91 (m, 1H, Ar), 7.10ꢀ7.16 (m, 3H, Ar), 7.32ꢀ7.38 (m,
3H, Ar), 7.46 (d, J = 7.6 Hz, 1H, Ar); 13C NMR (75 MHz, CDCl3): δ 14.4, 28.5, 38.6, 54.6, 55.5,
61.6, 80.1, 112.8, 112.9, 119.7, 125.9, 128.4, 128.4, 128.9, 129.8, 136.6, 141.8, 142.5, 155.5, 159.9,
171.8.
General procedure for preparation of (S)- and (R)-2-amino-3-(3'-hydroxybiphenyl-3-yl)propanoic
acids. The target amino acids (S)ꢀ38 and (R)ꢀ38 were obtained from 41 and 42, respectively (197 mg,
0.5 mmol) according to the procedure described for compounds 27, 28 and 31-38, followed by twoꢀ
time recrystallization from water.
(S)ꢀ2ꢀAminoꢀ3ꢀ(3'ꢀhydroxybiphenylꢀ3ꢀyl)propanoic acid ((S)ꢀ38). Yield 53 mg, 41%; ee = 98.3%;
22
1
[α]D +25.4° [c 0.014, 1M HCl:EtOH (1:1)]; ꢂε (214 nm) = +0.044 m2/mol; H NMR (300 MHz,
DMSO-d6): δ 2.94ꢀ2.97 (m, 1H, ArCH2), 3.20 (d, J = 14.9 Hz, 1H, ArCH2), 3.45 (m, 1H, ArCH2CH),
6.75 (d, J = 7.3 Hz, 1H, Ar), 7.05ꢀ7.08 (m, 2H, Ar), 7.20ꢀ7.26 (m, 2H, Ar), 7.34 (t, J = 7.6 Hz, 1H,
Ar), 7.43 (d, J = 8.2 Hz, 1H, Ar), 7.51 (s, 1H, Ar).
(R)ꢀ2ꢀAminoꢀ3ꢀ(3'ꢀhydroxybiphenylꢀ3ꢀyl)propanoic acid ((R)ꢀ38). Yield 34 mg, 26%; ee = 98.0%;
22
1
[α]D −24.5° [c 0.014, 1M HCl:EtOH (1:1)]; ꢂε (212 nm) = −0.038 m2/mol; H NMR (300 MHz,
DMSO-d6 + 50 ꢁL D2O): δ 2.90ꢀ2.97 (m, 1H, ArCH2), 3.17ꢀ3.22 (m, 1H, ArCH2), 3.53 (dd, J1 = 7.9
23
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 49
1
2
3
4
Hz, J2 = 4.1 Hz, 1H, ArCH2CH), 6.73 (d, J = 8.2 Hz, 1H, Ar), 7.02 (s, 1H, Ar), 7.07 (d, J = 7.6 Hz,
1H, Ar), 7.22ꢀ7.25 (m, 2H, Ar), 7.33ꢀ7.37 (m, 1H, Ar), 7.44 (d, J = 7.6 Hz, 1H, Ar), 7.48 (s, 1H, Ar).
5
6
7
8
9
Pharmacology
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
Receptor binding assays using native rat brain homogenates. Rat brain membrane preparations
used in the native receptor binding experiments were prepared according to the method described by
Ransom and Stec.34 Affinity for AMPA,35 KA36 and NMDA37 receptor sites was determined using 5
nM [3H]AMPA, 5 nM [3H]KA, and 2 nM [3H]CGP 39653 (Perkin Elmer) with some modifications
as previously described.38
Recombinant rat AMPA and KA receptor binding assays. GluA1o, GluA2(R)o, GluA3o,
GluK1(Q)1b, GluK2(V,C,R)A and GluK3A were inserted into recombinant baculoviruses, receptors
expressed by infection of Sf9 insect cells and infected Sf9 cell membranes utilized for radioligand
binding assays. GluA1o–4o were assayed using [3H]ꢀAMPA radioligand while GluK1(Q)1b,
GluK2(V,C,R) and GluK3A receptor binding assays were carried out using [3H]ꢀSYM 2081
radioligand as previously detailed.38, 39
GluA2 LBD binding assay. Purified soluble GluA2 LBD (GluR2ꢀS1S2J; 50 ng) was incubated with
1ꢀ3 nM [3H]ꢀAMPA in the presence of competing ligand for 2 h at 4°C in 0.25 mL assay buffer (50
mM TrisꢀHCl, 100 mM KSCN, 2.5 mM CaCl2, 10% (v/v) glycerol, pH 7.2 at 4°C). Samples were
diluted with 1 mL iceꢀcold assay buffer, immediately filtered onto 25 mm 0.2 ꢀM mixed cellulose
ester filters (Tisch Scientific, Cleves, OH) and washed once with 1 mL iceꢀcold assay buffer. Filters
were dried 1 h at 70°C and then dissolved in 2 mL FilterꢀCount (PerkinElmer) and radioactivity
determined by liquid scintillation counting (TriCarb 2900, PerkinElmer). Data were analyzed using
GraphPad Prism v6 (GraphPad Software, San Diego, CA). The Kd value of [3H]AMPA at GluA2
LBD (12.8 nM) was previously determined.31
TEVC pharmacology. Surgical procedures were conducted under the approval of the Danish
Ministry of Justice Animal Experiments Inspectorate. Mature female Xenopus lævis were
24
ACS Paragon Plus Environment

Page 25 of 49
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
anaesthetized using 0.1% ethyl 3ꢀaminobenzoate methanesulfonate (tricaine) and ovaries were
surgically removed. The ovarian tissue was dissected and treated with 2 mg/mL collagenase in
nominally Ca2+ꢀfree Barth’s medium (in mM: 88 NaCl, 1 KCl, 0.33 Ca(NO3)2, 0.41 CaCl2, 0.82
MgSO4, 2.4 NaHCO3, 10 HEPES, pH 7.4) for 2 h at room temperature. On the second day, oocytes
were injected with 25ꢀ50 nL of (≈ 1 mg/mL) rat GluA2(Q)i cRNA and incubated in Barth’s medium
with 0.10 mg/mL gentamicin (Sigma Chemical) and 1% penicillinꢀstreptomycin (Life Technologies,
Paisley, UK) at 17°C. Oocytes were typically used for recordings from 3 to 10 days postꢀinjection
and were voltageꢀclamped with the use of a twoꢀelectrode voltage clamp (GeneClamp 500B, Axon
Instruments, Union City, CA) with both microelectrodes filled with 3 M KCl. Recordings were made
while the oocytes were continuously superfused with frog Ringer’s solution (in mM: 115 NaCl, 2
KCl, 1.8 BaCl2, 5 HEPES, pH 7.6). Drugs were dissolved in frog Ringer’s solution and added by
bath application. Recordings were made at room temperature. Efficacy measurements were made at
GluA2(Q)i in the presence of 100 ꢀM cyclothiazide in order to block receptor desensitization
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
(cyclothiazide EC50 at GluA2(Q)i = 7.6 ꢀM.)40 To determine the maximum response, oocytes were
stimulated with 1 mM (S)ꢀglutamate plus 100 ꢀM cyclothiazide. Concentrationꢀresponse curves of
antagonists made in the presence of 10 ꢀM (S)ꢀGlu were analyzed using GraphPad Prism v6
(Graphpad Software Inc., San Diego, CA) to determine the IC50 and Hill value (nH), using a four
parameter logistic equation. Kb was calculated from the IC50 value by the ChengꢀPrusoff equation.
Homomeric GluA2(Q)i was expressed in Xenopus laevis oocytes and channel activity measured as
previously described.41 Concentrationꢀresponse curves of (S)ꢀ37 inhibition of 10 ꢀM (S)ꢀGlu
responses were fit to a logistic equation to determine slope (nH) and IC50, from which Kb was
calculated using the ChengꢀPrusoff equation.
Structure determination
Crystallization of GluA2 LBD in complex with (S)-37. The rat GluA2 LBD (GluR2ꢀS1S2J19) was
expressed and purified essentially as previously described.42, 43 GluA2 LBD in complex with (S)ꢀ37
25
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 49
1
2
3
4
5
6
7
8
was crystallized using the hanging drop vapor diffusion method at 7 °C. The drop contained 1 ꢁL of
the complex solution (5.5 mg/mL GluA2 LBD in 10 mM HEPES pH 7.0, 20 mM NaCl and 1 mM
EDTA equilibrated with the ligand (S)ꢀ37 as solid compound) and 1 ꢁL of reservoir solution of 20%
PEG4000, 0.3 M Li2SO4 and 0.1 M phosphateꢀcitrate pH 4.5. Reservoir volume was 0.5 mL. The
crystals appeared within one week and were flash cooled in liquid nitrogen after soaking in cryo
buffer consisting of the reservoir solution with 20% glycerol added.
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
Crystallization of GluA2 LBD in complex with (R)-38. Crystallization of GluA2 LBD with (R)ꢀ38
was carried out by the hanging drop vapor diffusion method at 7 °C. The complex solution contained
8.5 mg/mL GluA2 LBD (in 10 mM HEPES, 20 mM NaCl and 1 mM EDTA, pH 7.0) equilibrated
with (R)ꢀ38 as solid compound. Crystals were grown in drops of 1 µL complex solution and 1 µL
reservoir solution of 16% PEG2000, 0.1 M Li2SO4 and 0.1 M phosphateꢀcitrate buffer pH 4.5. The
reservoir volume was 0.5 mL. The crystals appeared within one week, and were flash cooled in
liquid nitrogen after soaking in cryo buffer consisting of the reservoir solution including 8%
glycerol, 8% ethylene glycol, 9% sucrose and 2% glucose.
Data collection and processing. Xꢀray data of the GluA2 LBD in complex with (S)ꢀ37 were
collected at the ID23ꢀ2 beamline (ESRF, Grenoble, France) at a wavelength of 0.8726 Å to 2.0 Å
resolution. Data processing was performed with the CCP4i suite of programs.44 The data of the
GluA2 LBD in complex with (R)ꢀ38 were collected at the I911ꢀ3 beamline (MAXꢀLab, Lund,
Sweden) at a wavelength of 1.0000 Å to 1.8 Å resolution. Data processing was performed with
XDS45 and scaling using the CCP4 suite of programs.
Both structure determinations were carried out by molecular replacement using the program
PHASER46 implemented in CCP4i. The GluA2 LBD with (S)ꢀATPO (PDB ID 1N0T, molA9) was
used as a search model, including protein atoms only. Clear solutions were found, showing one
molecule in the asymmetric unit for the complex with (S)ꢀ37 and four molecules in the asymmetric
unit of the complex with (R)ꢀ38. Subsequently, the amino acid residues of the GluA2 complex were
automatically built using ARP/wARP within CCP4i except for a few residues, which were manually
26
ACS Paragon Plus Environment

Page 27 of 49
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
built using COOT.47 The ligand coordinates were created in Maestro48 and fitted into the electron
density. Topology and parameter files for (S)ꢀ37 and (R)ꢀ38 were obtained using eLBOW49, keeping
geometry after ligand optimization in MacroModel (MMFF94s).50 The two structures were refined
using PHENIX,51 and between refinements the model was inspected and corrected in COOT. (S)ꢀ37,
(R)ꢀ38, water, ethylene glycol and glycerol molecules, as well as chloride and sulfate ions, were
manually modelled into the electron densities. Both structures were validated using tools in PHENIX
and COOT. For statistics on data collection and refinements, see Table 3.
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
Structure Analysis. Domain opening of GluA2 LBD in complex with (S)ꢀ37 and (R)ꢀ38, both
relative to the structure of GluA2 LBD with Glu (PDB ID 1FTJ, molB19) was calculated using the
DynDom Server.52 Figures were prepared in PyMOL.53
Molecular docking studies
The two proteinꢀligand complexes (GluA2 LBD in complex with (S)ꢀ37 and (R)ꢀ38, respectively)
were prepared with the Protein Preparation Wizard in Maestro from Schrodinger (Maestro, version
9.5, Schrödinger, LLC, New York, NY, 2013). The A chain and the ligand were kept for each of the
proteinꢀligand complexes. Default settings were applied for the preparation, except that only the
positions of the hydrogen atoms were optimized in the final refinement step. Prior to docking the
new ligands, they were optimized with Macromodel (MacroModel, version 10.1, Schrödinger, LLC,
New York, NY, 2013) using default settings (i.e. with the OPLSꢀ2005 force field in water).
A conformational search was carried out using the Monte Carlo multiple minimization (MCMM)
approach implemented in Macromodel. Default settings were applied for the conformational search.
A constrained minimization of the ligand, in the conformation as when bound to the protein, was
done. All nonꢀpolar hydrogen atoms were free in the minimization while the heavy atoms and the
polar hydrogen atoms were constrained by using a force constant of 100.0 kJ/mol/Å2 and wellꢀsize of
0.3 Å in a flatꢀbottom potential. The strain of the docked ligands was determined as the energy
27
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 49
1
2
3
4
5
6
7
8
difference from the constrained minimization and the MCMM calculation. The amino acid moiety of
the ligands was not considered in the calculations. All docking were performed with the XP version
of Glide (Glide, version 6.0, Schrödinger, LLC, New York, NY, 2013).54ꢀ56 Compounds were docked
in the tautomer form and protonation state shown in Table 1, except for 36 that was docked in its
pyridone tautomer form. All water molecules were removed. The receptor grid was generated using
default settings, except that the size of the enclosing box was increased to 16 Å x 16 Å x 16 Å
around the ligand and the hydroxyl groups of Thr655, Thr686 and Tyr702 were allowed to be
flexible. Sitemaps of the two proteins were generated using Sitemap (SiteMap, version 2.9,
Schrödinger, LLC, New York, NY, 2013). The ligands were used as center in the calculations.
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
28
ACS Paragon Plus Environment

Page 29 of 49
Journal of Medicinal Chemistry
1
2
Figure legends
3
4
5
6
7
Figure 1. Structures of selected competitive antagonists of AMPA and kainate receptors.
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) Structures of the most active previously obtained phenylalanines. (B) Structures of the
present target compounds.
Figure 3. Antagonism of 10 ꢁM (S)ꢀGlu responses at GluA2(Q)i receptors recorded by TEVC in X.
lævis oocytes. Shown are means ± SEM of pooled data from eight normalized experiments, conducted
in duplicate. (S)-37 IC50 = 3.7 ꢁM. Inset: Sample trace from one experiment at Vh = ꢀ60 mV with
duplicate stimulations of 10 ꢁM (S)ꢀGlu + (S)ꢀ37 (in ꢀM): zero, 0.01, 0.03, 0.1, 0.3, 1, 3, 10, 30, 100.
Scale bars are 250 nA and 5 min. (R)-38 IC50 = 6.3 ꢁM. Inset: Sample trace from one experiment at
Vh = ꢀ70 mV with duplicate stimulations of 10 ꢁM (S)ꢀGlu + (R)-38 (in ꢀM): zero, 0.01, 0.1, 0.3, 1, 3,
10, 30, 100, 300. Scale bars are 100 nA and 10 min.
Figure 4. Structures of GluA2 LBD in complex with (S)ꢀ37 (A-C) and (R)ꢀ38 (D-F), respectively. (A)
Cartoon representation of the GluA2 LBD dimer (in dark salmon) in complex with (S)ꢀ37 (in yellow).
The degree of domain opening induced by (S)ꢀ37 as well as the Ile633ꢀIle633 linkerꢀlinker distance is
indicated on the figure. (B) Zoom on the ligand binding site of GluA2 LBD with (S)ꢀ37 bound. A 2Foꢀ
Fc omit map at 1 σ carved around the ligand at 1.6 Å radius is shown. Potential hydrogen bonds
within 3.2 Å to water molecules (red spheres) and surrounding residues are shown as black, dashed
lines. Residues within 4.0 Å distance from chlorine atoms have been included. (C) Zoom on the
cavity accommodating the pyridine ring system of (S)ꢀ37. (GOL, glycerol) (D) Cartoon representation
of the GluA2 LBD dimer (in dark salmon) in complex with (R)ꢀ38 (in cyan). The degree of domain
opening induced by (R)ꢀ38 as well as the Ile633ꢀIle633 linkerꢀlinker distance is indicated on the
figure. (E) Zoom on the ligand binding site of GluA2 LBD with (R)ꢀ38 bound. A 2FoꢀFc omit map at
29
ACS Paragon Plus Environment