Non-naturally Occurring Regio Isomer of Lysophosphatidylserine Exhibits Potent Agonistic Activity toward G Protein-Coupled Receptors

Article abstract of DOI:10.1021/acs.jmedchem.0c01126

Lysophosphatidylserine (LysoPS), an endogenous ligand of G protein-coupled receptors, consists of l-serine, glycerol, and fatty acid moieties connected by phosphodiester and ester linkages, respectively. An ester linkage of phosphatidylserine can be hydrolyzed at the 1-position or at the 2-position to give 2-acyl lysophospholipid or 1-acyl lysophospholipid, respectively. 2-Acyl lysophospholipid is in nonenzymatic equilibrium with 1-acyl lysophospholipid in vivo. On the other hand, 3-acyl lysophospholipid is not found, at least in mammals, raising the question of whether the reason for this might be that the 3-acyl isomer lacks the biological activities of the other isomers. Here, to test this idea, we designed and synthesized a series of new 3-acyl lysophospholipids. Structure-activity relationship studies of more than 100 "glycol surrogate"derivatives led to the identification of potent and selective agonists for LysoPS receptors GPR34 and P2Y10. Thus, the non-natural 3-acyl compounds are indeed active and appear to be biologically orthogonal with respect to the physiologically relevant 1-and 2-acyl lysophospholipids.

Full text of DOI:10.1021/acs.jmedchem.0c01126

Subscriber access provided by UNIV OF WESTERN ONTARIO
Article
Non-naturally-occurring Regio Isomer of Lysophosphatidylserine
Exhibits Potent Agonistic Activity Towards G Protein-coupled Receptors
Sho Nakamura, Misa Sayama, Akiharu Uwamizu, Sejin Jung, Masaya Ikubo, Yuko
Otani, kuniyuki kano, Jumpei Omi, Asuka Inoue, Junken Aoki, and Tomohiko Ohwada
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.0c01126 • Publication Date (Web): 28 Jul 2020
Downloaded from pubs.acs.org on August 2, 2020
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a service to the research community to expedite the dissemination
of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in
full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully
peer reviewed, but should not be considered the official version of record. They are citable by the
Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore,
the “Just Accepted” Web site may not include all articles that will be published in the journal. After
a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web
site and published as an ASAP article. Note that technical editing may introduce minor changes
to the manuscript text and/or graphics which could affect content, and all legal disclaimers and
ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or
consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W.,
Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the
course of their duties.

Page 1 of 90
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Non-naturally-occurring Regio Isomer of Lysophosphatidylserine Exhibits Potent
Agonistic Activity Towards G Protein-coupled Receptors
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Sho Nakamura,1,+ Misa Sayama, 1,+ Akiharu Uwamizu, +,2,3,5 Sejin Jung, Masaya Ikubo,
1
1
Yuko Otani, Kuniyuki Kano, Jumpei Omi, 2,3,5 Asuka Inoue, 3,4,5 Junken Aoki, * , 2,3,5,6
1
2,3,5
Tomohiko Ohwada*, 1
+
Equal contribution
1
Department of Organic and Medicinal Chemistry, Graduate School of Pharmaceutical
Sciences, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-0033, Japan
2
Department of Health Chemistry, Graduate School of Pharmaceutical Sciences, The University
of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-0033, Japan
3
Laboratory of Molecular and Cellular Biochemistry, Graduate School of Pharmaceutical
Sciences, Tohoku University, 6-3, Aoba, Aramaki, Aoba-ku, Sendai, Miyagi 980-8578, Japan
4
AMED-PRIME, Japan Science and Technology Corporation, Kawaguchi, 332-0012, Japan
5
AMED-LEAP, Japan Science and Technology Corporation, Kawaguchi, 332-0012, Japan
6
AMED-CREST, Japan Science and Technology Corporation, Kawaguchi, 332-0012, Japan
*To whom correspondence should be addressed.
Chemical and Computational studies:
For T.O.: phone, +81-3-5841-4730; fax, +81-3-5841-4735;
E-mail, ohwada@mol.f.u-tokyo.ac.jp
Biological study: For J.A.: phone, +81-3-5841-4720; fax, +81-3-3818-3173
E-mail, jaoki@m.tohoku.ac.jp, jaok@mol.f.u-tokyo.ac.jp
1
ꢀ
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 90
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Abstract
Lysophosphatidylserine (LysoPS), an endogenous ligand of G protein-coupled receptors,
consists of L-serine, glycerol and fatty acid moieties connected by phosphodiester and
ester linkages, respectively. An ester linkage of phosphatidylserine (PS) can be
hydrolyzed at the 1-position or at the 2-position, to give 2-acyl lysophospholipid or 1-
acyl lysophospholipid, respectively. 2-Acyl lysophospholipid is in non-enzymatic
equilibrium with 1-acyl lysophospholipid in vivo. On the other hand, 3-acyl
lysophospholipid is not found, at least in mammals, raising the question of whether the
reason for this might be that the 3-acyl isomer lacks the biological activities of the other
isomers. Here, to test this idea, we designed and synthesized a series of new 3-acyl
lysophospholipids. Structure-activity relationship studies of more than 100 “glycol
surrogates” derivatives led to the identification of potent and selective agonists for
LysoPS receptors GPR34 and P2Y10. Thus, the non-natural 3-acyl compounds are indeed
active, and appear to be biologically orthogonal with respect to the physiologically
relevant 1- and 2-acyl lysophospholipids.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
ꢀ
ACS Paragon Plus Environment

Page 3 of 90
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Introduction
Lysophospholipids (LPLs) and fatty acids are released by the hydrolysis of membrane
phospholipids. The fatty acids serve as precursors of biologically active prostaglandins
and leukotrienes, while the LPLs serve as regulators of multiple cellular responses.1
Among the LPLs, lysophosphatidylserine (LysoPS) is thought to have a
pathophysiological role in the immune system, including suppression of regulatory T
cells.2-5 Recently four G protein-coupled receptors (GPCRs), GPR34, P2Y10,
A630033H20 (a non-functional truncated pseudo-gene product in human) and GPR174,
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
have been identified as LysoPS cognate receptors and designated as LPS
1
, LPS , LPS2L
2
(LPS -like) and LPS
2
3
, respectively. 6, 7 The functional receptors are abundantly expressed
in the immune systems and several organs, and studies in vitro and in vivo indicate their
involvement in the control of the immune response. For example, Liebscher et al
concluded that GPR34 is involved in immune response to antigen and pathogen challenge,
based on a study of GPR34-deficient (KO) mice.8, 9 It seems likely that dysfunction of
these LysoPS receptors or dysregulation of LysoPS is linked to various diseases.
<
Figure 1>
LysoPS consists of L-serine, glycerol and fatty acid moieties connected by
phosphodiester and ester linkages, respectively. Physiologically, phosphatidylserine (PS)-
specific phospholipase A
1
(PS-PLA
1
) hydrolyzes PS regioselectively at the 1-position to
give 2-acyl lysophospholipid (Figure 1(a); Type II), from which 1-acyl lysophospholipid
1
0
(Type I) is formed by non-enzymatic migration of the ester. Thus, Types I and II are
both formed in mammals, and are biologically active.
(a)
A
P
O
OH
P
O
3 (R)
2
1
HO
O
O
O
O
NH2
OH
LysoPS (18:1)
3 (R)
1
3 (R)
1
3 (R)
1
P
A O
OH
P O
O A
O
OH
2
O A
2
O P
2
OH
I
II
III
(b)
3
ꢀ
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 90
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
3 (R)
1
3 (R)
1
P
A
A
P
A
O
OH
OH
O
O
O
2
2
OH
II O
I
A
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3 (R)
1
3 (S)
1
O
HO
2
O
2
O
III
III-enantiomer
P
P
Figure 1. Geometrical Isomers of Glycerol Substitution of LysoPS (a), and Possible
Mutual Transformations (b).
Types I and II are physiological, but Type III is not.
Indeed, we previously reported structure-activity-relationship (SAR) studies of synthetic
LysoPS analogues of Type I and Type II and identified GPR34-, P2Y10- and GPR174-
selective agonists.11-14 On the other hand, Type III has never been reported in mammals,
though it might be formed via multiple migrations of acyl and phosphodiester groups
(
Figure 1(b)). This raises the question of whether the reason why Type III is not produced
might be that it lacks biological activity. This possibility might be consistent with the
common view that when triacylglycerol (TG) is hydrolyzed in the body, only sn-1,2- and
1
5
sn-2,3-diacylglycerol (DAG), not sn-1,3 DAG, are formed. On the other hand, the
formation of sn-1,3 DAG was argued under specified conditions.16, 17 Therefore, Type III
LysoPS derivatives are imaginable, and their biological activities are worthy to be
studied.
Here, in order to answer this question, we designed and synthesized a series of new
Type III LysoPS derivatives and examined their GPCR activation activities. Interestingly
we found that these derivatives include potent and selective agonists for LysoPS receptors
GPR34 and P2Y10, and furthermore, conformational constraint of these analogues by
using a tetrahydropyran-3,4-diol framework enabled us to identify stable conformations
that may be responsible for the bioactivities. We discuss the reason why these compounds
are apparently not used physiologically, despite their activity.
Results and Discussion
Synthesis of LysoPS analogues
4
ꢀ
ACS Paragon Plus Environment

Page 5 of 90
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Type III LysoPS derivatives were synthesized according to established methods,11,13,14 by
the sequential buildup of protected fragments followed by global deprotection with TFA
in the final step. For example, the prototype C2-derivative 1a was synthesized as follows.
Ethylene glycol was mono-acylated with oleoyl chloride. Boc-L-serine tert-butyl ester
was phosphorylated with tert-butyl tetraisopropylphosphorodiamidite in the presence of
1H-tetrazole, followed by a second phosphorylation with the acylated ethylene glycol.
Subsequent in situ oxidation of the resultant phosphite triester intermediate with tert-butyl
hydroperoxide afforded the fully protected phosphate triester. Removal of all the
protecting groups with TFA furnished the LysoPS analogue as the TFA salt. For details,
see the Experimental Section.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Chain-length-dependence of the activities of LysoPS analogues
We previously showed that the effects of modification of the three modules or two
linkages of LysoPS, particularly Type I LysoPS (Figure 1), on the activating activity
towards GPR34, P2Y10 and GPR174 are different and rather orthogonal.11-14
For example, elimination of the hydroxyl group of the glycerol moiety, as in 1b (1-oleoyl-
2-deoxy LysoPS), diminished the GPR174 activation potency (Figure 2) without affecting
the activity towards GPR34 or P2Y10. Building on that work, we focused first on
synthesizing a series of Type III deoxy LysoPS analogues with different lengths of the
glycerol backbone (Figure 2), that is, 1a (deoxy-LysoPS C2 (18:1)), 1c (deoxy-LysoPS
C4 (18:1)) and 1d (deoxy-LysoPS C5 (18:1)), wherein the fatty acid moiety is an oleoyl
group (18:1). As expected, these deoxy analogues were not active towards GPR174,
enabling us to focus on the activities towards GPR34 and P2Y10, which were evaluated
by means of TGFa ectodomain shedding assay of alkaline phosphatase-tagged
TGFa (AP-TGFa shedding assay) 7, 18 (Table 1 and Figures 3 and 4).
<
Table 1> < Figures 2> <Figures 3 and 4 >
Among these 2-deoxyglycerol derivatives 1a-1d, the optimal carbon chain length
of the glycerol moiety was different for activities towards GPR34 and P2Y10. For GPR34,
all of the C2 (1a), C3 (1b), C4 (1c) and C5 (1d) derivatives showed very weak agonistic
activities compared with LysoPS (18:1) (EC50 = 230 nM). The EC50 values were as
follows: 1a EC50 > 3 µM, 1b EC50 = 540 nM; 1c EC50 = 320 nM; 1d EC50 = 470 nM. In
the case of P2Y10, the C2 (1a), C3 (1b) and C4 (1c) derivatives showed high potency (1a
EC50 = 66 nM; 1b EC50 = 20 nM; 1c EC50 = 53 nM), similar to that of LysoPS (18:1) (EC50
5
ꢀ
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 90
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
=
20 nM), while the C5 derivative 1d showed weaker activity (EC50 = 240 nM) (Table 1
and Figures 3 and 4, and see also Supporting Information Figures S1-S2).
O
OH
P
O
n
HO
O
O
O
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
O
NH2
1
a: n=1 deoxy-LysoPS C2 (18:1)
b: n=2 deoxy-LysoPS C3 (18:1)
1
1
c: n=3 deoxy-LysoPS C4 (18:1)
1d: n=4 deoxy-LysoPS C5 (18:1)
Figure 2 LysoPS Derivatives with Different Chain Lengths of the Glycerol Moiety
20
15
20
15
10
5
1
8:1-LysoPS
18:1-LysoPS
1a
4R
4S
5R
6
2R
1
0
5
0
2
S
R
3
3S
0
7
8
9
-
9
-8
-7
-6
-5
-9
-8
-7
-6
-5
-12
0
0
0
-10
-12
0
-10
1
0
10
10
10
10
10
10
10
10
10
1
1
1
0
10
10
10
20
20
15
10
5
1
8:1-LysoPS
18:1-LysoPS
1
1
5
0
5
0
10
11
12
18
19
20
21
13
14
0
15
1
-1
-9
-8
-7
-6
-5
-9
-8
-7
-6
-5
-12
-10
-12
0
-10
1
0
10
10
10
10
10
10 10 10 10 10
0
10
10
10
20
18:1-LysoPS
1
1
5
0
5
0
22
23
24
25
1
-1
-9
-8
-7
-6
-5
-12
-10
1
0
10 10 10 10
0
10 10
Compound (M)
Figure 3. Concentration-response curves of GPR34 for a panel of compounds
6
ꢀ
ACS Paragon Plus Environment

Page 7 of 90
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
HEK293 cells transiently transfected with expression vectors encoding alkaline
phosphatase (AP)-TGFa, chimeric Ga and mouse GPR34 were treated with test
compounds. Cells transfected with AP-TGFa, chimeric Ga and empty vector were used
as a negative control. Receptor-specific AP-TGFa release was determined by subtracting
the background responses in empty vector-transfected cells. Data are mean and SEM
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(standard error of the mean) for 3−8 independent experiments.
2
1
0
5
20
15
10
5
18:1-LysoPS
10
18:1-LysoPS
18
19
20
21
11
10
5
1
2
3
1
0
14
15
0
1
-1
-9
-8
-7
-6
-5
-
9
-8
-7
-6
-5
-120
-
12
0
-10
-10
10
10
10
10
10
10
10 10 10 10 10
1
0
10
10
10
2
1
1
0
5
0
5
0
1
8:1-LysoPS
22
23
24
25
-9
-8
-7
-6
-5
-12
0
-10
1
0
10
10
10
10
1
0
10
Compound (M)
Figure 4. Concentration-response curves of P2Y10 for some compounds
Activation of P2Y10 by various compounds. Experiments were done in the same manner
as in Figure 3, except that mouse P2Y10-encoding vector was used instead of mouse
GPR34-encoding vector. Receptor-specific AP-TGF α release responses were
determined by subtracting the background responses in empty vector-transfected cells.
Data are mean and SEM (standard error of the mean) for 3−8 independent experiments.
Modification of the C2 glycerol backbone, leading to GPR34-selective agonists
Methyl Scan: The activity of deoxy-LysoPS C2 (18:1) (1a) toward GPR34 was
exceptionally low, as compared with those of the C3 (1b), C4 (1c) and C5 (1d) derivatives
(Table 1 and Figures 3 and 4). This suggests that just a single carbon atom deletion has a
great impact on the GPR34 potency. Thus, we modified the “glycerol“ backbone of
deoxy-LysoPS C2 (18:1) (1a) by the addition of a single carbon methyl group, aiming to
make it equivalent to a “C3 glycerol” skeleton (Figure 5).
<
Figure 5>
7
ꢀ
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 90
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
There are four possible regio- and stereo-isomers (2R, 2S, 3R and 3S) of the methyl-
1
9
substituted derivative (Figure 5), of which one, 3R, was examined previously. The R-
configuration is the same as the original configuration of the glycerol moiety of naturally
occurring Type I LysoPS (such as LysoPS (18:1)).
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
O
OH
P
O
O
OH
P
O
3 (R)
(R)
1
O
OH
P
O
HO
O
O
HO
O
O 2
3
2
O
NH2
R
HO
H N
O
O
NH2
O
R
O
R
2
3R
Type II
O
2
R
O
Type III
O
1
a
1
OH
O
OH
P
O
3 (S)
1
(S)
R=
P
O
HO
O
O 2
HO
O
O
3
O
2
O
NH2
NH2
O
R
R
(
18:1)
2
S
3S
Type II-diasteromer
O
O
Type III-diasteromer
ꢀ
Figure 5 Methyl Scanning of the C2 Glycerol Backbone
Substitution of a methyl group at the geminal position with respect to the phosphodiester
group (2R and 2S) increased GPR34 potency, as compared with non-methyl C2 derivative
a (EC50 > 3 µM): 1(R)-2R EC50 = 24 nM and 1(S)-2S EC50 = 170 nM (Figure 6). On the
other hand, the P2Y10 potency was maintained or decreased (2R EC50 = 250 nM; 2S EC50
37 nM) as compared with LysoPS (18:1) (EC50 = 20 nM). Thus, the receptor subtype
selectivity of 2R and 2S between GPR34 and P2Y10 was moderate (Figure 6).
Figure 6>
1
=
<
A methyl group at the geminal position to the fatty acid ester group (3R and 3S) increased
the GPR34 potency of 3R (EC50 = 250 nM) to the same level as LysoPS (18:1) (GPR34
EC50 = 230 nM), though 3S was almost inactive (EC50 > 3 µM) (Figure 6). Both 3R and
3
S showed weaker P2Y10 potency (3R EC50 = 79 nM; 3S EC50 = 390 nM) than LysoPS
(18:1) (P2Y10 EC50 = 20 nM). Thus, the stereochemistry of the methyl group markedly
affects the activities towards GPR34 and P2Y10, and the compounds with the R-
configuration, i.e., the same as the original configuration of the glycerol moiety of Type
I LysoPS, are more potent than those with the S-configuration (Figure 6, Table 1 and
Figures 3 and 4).
8
ꢀ
ACS Paragon Plus Environment

Page 9 of 90
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
OH
P
OH
P
O
O
O
(R)
O
O
(S)
HO
H N
O
O
O R
O
O R
O
2
2R
2S
Type III
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Type III
GPR34: EC50
24 nM
170 nM
(37 nM)
(250 nM)
(P2Y10: EC )
5
0
R=
OH
OH
P
(S)
(R)
P
(
18:1)
O
O
O
O
O
O R
O
O
O R
O
LysoPS
3R
3S
Type II
Type II
230 nM
(20 nM)
2
50 nM
> 3 µM
(390 nM)
(79 nM)
Figure 6. Receptor Activation by Methylated Type III LysoPS Derivatives
Ether substitution Next, we examined the effect of substitution with a methoxymethyl
group (Figure 7). This substituent almost abrogated the P2Y10 potency of 4R and 4S
(both EC50 > 3 µM), though 5R was active (EC50 = 70 nM). As for GPR34 activation, 4R
exhibited slightly higher activity (EC50 = 11 nM) than the corresponding methyl-
substituted analogues (2R) (EC50 = 24 nM) (Figure 6), while the diastereomer 4S showed
reduced activity (EC50 = 620 nM) (Table 1 and Figures 3 and 4). The Type II analogue 5R
showed increased activity towards GPR34 (EC50 = 160 nM) as compared with the methyl
counterpart (3R: EC50 = 250 nM) (Table 1 and Figures 3 and 4). Based on the results for
the methyl- (2 and 3) and methoxymethyl-substituted analogues (4 and 5), R
configuration of the stereo center on the C2 skeleton appears to be crucial for high GPR34
potency.
<
Figure 7>
9
ꢀ
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 90
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
OMe
OMe
OH
OH
P
O
O
P
HO
O
O
(S)
HO
O
O
(R)
NH2
O
NH2
O
O R
O
O R
O
4
R
4S
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
O
Type III
Type III
HO
H N
GPR34: EC50
620 nM
(> 3 µM)
11 nM
2
(
P2Y10: EC ) (> 3 µM)
5
0
OH
P
O
(R)
R=
HO
O
O
OMe
NH2
O
O
R
(
18:1)
5R
O
Type II
160 nM
(70 nM)
Figure 7. Receptor Activation by Ether Type III LysoPS Derivatives
Other Ether Functionalities: To examine the effect of other ether functionalities of 1(R)-
methoxy-3-oleoyl LysoPS 4R on GPR34 potency, several derivatives bearing a variety of
carbon functional groups on the methoxy oxygen atom were synthesized (Figure 8).
Simple alkyl ethers such as ethyl ether (6 EC50 = 3.5 nM) and propyl ether (8 EC50 = 6.1
nM) were found to exhibit stronger GPR34 activation ability than the simple methyl
compound (4R EC50 = 24 nM) or unsaturated propargyl ether (7 EC50 = 12 nM). In the
case of the bulky benzyloxymethyl group (9), the GPR34 potency was significantly
decreased (9 EC50 = 880 nM) to a level lower than that of LysoPS (18:1). These results
may indicate that there is a hydrophobic pocket in the GPR34 ligand-binding site that can
accommodate glycerol backbone substituents up to a propyloxymethyl group, but not a
benzyloxymethyl group, in size (see the results of the docking study below). As regards
P2Y10 potency (Figure 8), simple alkyl ether derivatives such as an ethoxy and propyloxy
lacked activation ability, as in the case of the methoxy compound 4R (ethoxy 6 EC50 > 3
µM; propyloxy 8 EC50 > 3 µM). The unsaturated propargyl ether and benzyl ether also
showed weaker activity towards P2Y10 (propargyl ether 7: EC50 = 420 nM; benzyl ether
9
: EC50 = 1400 nM) as compared with LysoPS (18:1) (EC50 = 20 nM) (Table 1 and Figures
3 and 4). Therefore, the type III ethyl ether LysoPS derivative (6) is the most selective
GPR34 agonist among the compounds studied here.
<
Figure 8>
10
ꢀ
ACS Paragon Plus Environment

Page 11 of 90
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
OR
OH
O
P
O
O
HO
O
R
R=
(
18:1)
O
R
NH2
O
Type III
O
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
O
CH3
O
O
O
O
O
OH
OH
P
OH
P
O
CH OH
3
P
O
P
O
O *
O *
O *
O *
O
O
O
6
7
8
9
GPR34: EC50
3.5 nM
(P2Y10: EC ) (> 3 µM)
12 nM
(420 nM)
6.1 nM
(> 3 µM)
880 nM
(1400 nM)
5
0
Figure 8. Ether Derivatives of Type III LysoPS
Reproducibility of methyl scanning in fatty acid surrogates
Since Type III topology of the glycerol moiety was associated with selectivity and
potency for GPR34, we examined the general validity of this architecture. In our previous
report,11, 13 fatty acid surrogates including phenyl groups were found to increase the
potency and selectivity of LysoPS derivatives towards specified receptors. Here, we
introduced the non-fatty acid surrogates A and B (Figure 9) into the Type III
lysophospholipid arrangement (Figure 10 and Table 1).
<
Figure 9> <Figure 10>
HO
(18:1)
O
Oleic acid (18:1)
HO
HO
O
O
O
O
O
Fatty Acid Surrogate
Fatty Acid Surrogate
A
B
Figure 9. Fatty Acid Surrogates Used in This Work
11
ꢀ
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 90
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
OMe
O
OH
P
O
O
OH
P
O
HO
O
O
HO
O
O
NH2
O
NH2
O
O
O
O
O
220 nM
( 2300 nM)
5.7nM
( > 3 µM)
O
O
13
1
0
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
O
OH
P
OEt
O
O
OH
HO
O
O
O
P
O
NH2
O
HO
O
O
NH2
O
O
6
.3 nM
( 170 nM)
O
5
nM
> 3 µM)
O
O
O
(
1
1
14
O
OH
P
O
O
OH
P
HO
O
O
HO
O
O
O
NH
2
O
NH2
O
3
30 nM
( 5.7 nM )
O
O
O
O
9
nM
290 nM)
O
(
15
12
GPR34: EC50
P2Y10: EC )
(
5
0
Figure 10. Structures of Type III LysoPS Derivatives Bearing Fatty Acid
Surrogates
Non-lipid surrogates increased the GPR34 activity of all five derivatives 10-14
(10: EC50 = 220 nM; 11: EC50 = 6.3 nM; 12: EC50 = 9 nM; 13: EC50 = 5.7 nM; 14: EC50
nM) (Figure 10 and Table 1 and Figures 3 and 4). Among them, 14 was a selective
super-agonist for GPR34 (GPR34 EC50 = 5 nM; P2Y10 EC50 > 3 µM). The C2 derivative
0 was almost inactive towards P2Y10 (EC50 = 2300 nM), but the methyl-substituted
=
5
1
derivatives 11 and 12 showed activity towards P2Y10 (11 EC50 = 170 nM, 12 EC50 = 290
nM). Both ether-type derivatives (13 and 14) lacked activity towards P2Y10 (EC50 > 3
µM), indicating that these compounds are selective agonists for GPR34.
When we used fatty acid surrogate B (Figure 9), the potency of 15 towards P2Y10
was greatly increased (EC50 = 5.7 nM), and activity towards GPR34 also emerged (EC50
=
330 nM), so that 15 is a P2Y10-selective agonist.
Conformational characteristics of Type III glycerol analogues
While different conformations of the glycerol unit of Types I and II LysoPS derivatives
(
16, Figure 11) can arise from the rotation of the two geminal C-C single bonds along the
three-carbon chain of the glycerol moiety (C2-C3, C3-C4 in 16, Figure 11), the
conformations of the C2 glycerol surrogate of Type III LysoPS derivatives (17, Figure
12
ꢀ
ACS Paragon Plus Environment

Page 13 of 90
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
1), can be described in terms of corresponding rotations around the two vicinal bonds
along the two sets of four-atom chains, P-O-C-C (P) and C-C-O-C (FA), as shown in
Figure 11.
<
Figure 11>
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Here, we generated accessible conformers by means of accelerated molecular dynamics
2
0
calculation (replica exchange with solute tempering, REST) at 300 K in the presence of
explicit water (TIP3P) for 30 nsec. The obtained conformations (about 1000 snapshots
obtained every 300 fsec) are plotted in terms of the two dihedral angles in Figure 10.
Similar REST simulations were carried out in n-octanol, a hydrophobic environment
(Figure S3).
-
P
O
P
O
2
OH
-
O
O
O
1
3
O
+
FA
NH3
4
O
5
1
2
3
4
5
∠O C C C
2
3 4
∠
C C C O
R
O
1
6
-
O
P
ω
O
P
3
1
-
O
O
O
2
4
O
FA
+
NH3
O
R
5
1
2
3
4
6
∠
P O C C
O
3
4
5 6
∠
C C O C
17
ꢀ
Figure 11. Two Dihedral Angles in the Glycerol Moieties
We also generated possible conformations of C2 glycerol derivatives 17 by means of
accelerated molecular dynamics (MD) calculations with REST at 300 K, using the OPLS3
force field in a buffered water environment (0.15 M NaCl in TIP3P water). Sampled
snapshots of the conformations were plotted with respect to the dihedral angles around
the phosphate (P) side and fatty acid surrogate (FA) side. The results suggest that simple
methyl substitution greatly constrains the rotation of the geminal bond, affording different
distributions of conformations (Figure 12). The analysis, including cyclic derivatives (see
Figure 14), suggested that the conformations in terms of these two representative dihedral
angles are correlated at least to some degree with the receptor potency (vide infra).
13
ꢀ
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 90
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Similar distributions were obtained in n-octanol (Supporting Information, Figure S3(a)).
On the other hand, when we analyzed the distribution of the dihedral angle O2-C3-C4-
O5 (
correlation with the biological activities (see Figure S4).
Figure 12>
w
) of 17 (Figure 11) and related cyclic analogues (see Figure 15), we found no
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
<
P
3
O
O
P
FA
1
4
O
O
O
2
O
NH3
O
O
EC > 3 µM
1
a
50
1
80
1
20
60
0
-180
-120
-60
0
60
120
180
-60
120
180
-
-
Torsion P
14
ꢀ
ACS Paragon Plus Environment

Page 15 of 90
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
O
O
P
O
O
O
O
NH3
O
O
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2R
EC 24 nM
50
1
80
1
20
6
0
0
-
180
-120
-60
0
60
120
180
-60
-120
-180
Torsion P
O
O
P
O
O
O
O
NH3
O
O
2
S
EC 170 nM
50
1
80
1
20
6
0
0
-
180
-120
-60
0
60
120
180
-
60
-
120
180
-
Torsion P
ꢀ
15
ꢀ
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 90
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
O
O
P
O
O
O
O
NH3
O
EC 250 nM
O
50
3
R
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
80
20
60
0
1
-
180
-120
-60
0
60
120
180
-60
-120
-180
Torsion P
O
O
P
O
O
O
O
NH3
O
EC > 3 µM
5
0
O
3S
180
1
20
6
0
0
-180
-120
-60
0
60
120
180
-
60
-120
-180
Torsion P
ꢀ
16
ꢀ
ACS Paragon Plus Environment

Page 17 of 90
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Figure 12. Conformational distributions with respect to the dihedral angle around
the phosphodiester (P) functionality and around the ester linkage of the fatty acid
functionality (FA).
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
When the preferred conformations of the four possible isomers of methyl-substituted
derivatives (2R, 2S, 3R, 3S) were compared with each other in relation to the GPR34
potency (Figure 12), an intriguing relationship emerged: compounds exhibiting high
GPR34 potency (such as 2R) appear to have a preference for conformations with both
positive (P) and positive (FA) (Figure 12, yellow-circled quadrant). Compound 3S,
exhibiting weaker GPR34 potency, has a higher distribution of negative (FA)
conformations. This correlation between GPR34 activity and preferred conformation
suggested that certain values of dihedral angles, i.e., +(P) and +(FA) (Figure 12, yellow-
circled quadrant), are preferred for GPR34 potency (see also Figure 14). Interestingly, the
dihedral angle around the fatty acid moiety (FA) is more critical for the GPR34 activity
than that around the phosphodiester (P): while 2S takes -(P) conformation on the
phosphodiester side (P), and 3R takes both +(P) and -(P) conformations on the
phosphodiester side (P), the bioactivities are conserved, probably because 2S and 3R can
both take +(FA) conformation on the FA side (see yellow-circled quadrant in Figure 12).
<
Figure 12>
Conformationally constrained cyclic analogues of C2-glycerol surrogates of Type III
glycerol analogues
In order to constrain the conformation of open-chain C2-glycerol surrogates such as ether
6
we utilized a cyclic tetrahydropyran-3,4-diol framework (conserving R-configuration
at the C3 position) (Figure 13). We can define two stereoisomers of the diol as having
trans and cis stereochemistry. Among eight possible isomers of LysoPS derivatives
(Figure 13), we focused on four isomers of the Type III LysoPS derivatives.
<
Figure 13>
17
ꢀ
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 90
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1
1
4
O
O
2
2
R
R
HO
3
HO
3
4
OH
cis type
OH
trans type
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
H C
3
O
O
O
O
O
O
O
O
O
O
O
O
P
O
P O
R
O
P O
R
O
O
O
O
R
O
R
R
NH3
O
R
P
P
1
8 O
Type III
22
19
Type III
24
O
O
Type II
Type II
P
O
O
O
O
O
O
R=
P
O
R
O
P O
R
O
O
R
O
P
O
R
O
6
P
2
0
23
2
1 O
Type III
25
R=
O
Type II
Type II
Type III
O
Figure 13. Design of cyclic compounds
We calculated the accessible conformations of these cyclic derivatives (18-21) by
means of REST simulations in water (TIP3P, 300 K). The distributions of conformers
along the previously defined two dihedral angles around the phosphate (P) side and fatty
acid surrogate (FA) side were studied in a similar manner to the open-chain analogues
and the plots are shown in Figure 14. Only 20 has +(P) +(FA) configuration as a major
structure, while 21 has -(P) -(FA) conformation. Thus, we predicted that 20 would show
strong GPR34 agonistic activity, but 21 would be inactive towards GPR34. Also 18 has
+
(P) but -(FA), and 19 has +(FA) but -(P). Thus, we predicted that 18 would be inactive
because it has -(FA) conformation. Similar conformational distributions were found in n-
octanol (Figure S3(b)).
<
Figure 14>
18
ꢀ
ACS Paragon Plus Environment

Page 19 of 90
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
O
O
O
P
O
O
O
O
O
EC > 3 µM
NH3
50
O
O
1
8
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
1
80
20
6
0
0
-180
-120
-60
0
60
120
180
-
60
-
120
180
-
Torsion P
ꢀ
ꢀ
O
O
O
O
P
O
O
O
O
NH3
EC 220 nM
50
9 O
O
1
ꢀ
19
ꢀ
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 90
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
O
O
O
P
O
O
O
O
O
EC 0.83 nM
50
NH3
O
O
20
180
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
20
60
0
-
180
-120
-60
0
60
120
180
-
60
-
120
-
180
Torsion P
ꢀ
O
O
O
P
O
O
O
O
NH3
O
EC NA
50
O
O
2
1
ꢀ
Figure 14. Simulated Conformational Preferences of Type III Cyclic LysoPS
Derivatives
20
ꢀ
ACS Paragon Plus Environment

Page 21 of 90
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Type III analogs
Type II analogs
GPR34: EC50
GPR34: EC50
(P2Y10: EC50)
O
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(P2Y10: EC50
)
O
O
OH
O
OH
P
P
> 3 µM
(420 nM)
340 nM
(1300 nM)
HO
O
O
HO
O
O
O
O
O
O
NH2
NH2
O
O
O
O
O
18
O
22
O
OH
P
O
OH
P
6
3 nM
220 nM
(1300 nM)
HO
O
O
HO
O
O
(840 nM)
O
O
O
O
NH2
NH2
O
O
O
O
O
1
9
2
3
O
O
OH
O
OH
P
0
.83 nM
2.9 nM
P
HO
O
O
(1 nM)
HO
O
O
(26 nM)
O
O
O
O
NH2
NH2
O
O
O
O
O
2
0
24
O
O
OH
P
O
OH
P
NA
12 nM)
> 3 µM nM
(150 nM)
HO
O
O
HO
O
O
(
O
O
O
O
NH2
NH2
O
O
O
O
2
1
25
Figure 15. Cyclic Type III Analogs (18-21) and Cyclic Type II Analogs (19-25)
Eight selected conformationally constrained analogues (Figure 15), including Type III
(
18-21) and Type II (22-25), were synthesized as shown in Experimental Section.
<
Figure 15>
The Type II (22-25) analogues and Type III (18-21) analogues have different positions of
the oxygen atom in the tetrahydropyran ring, but the conformational preferences of Type
II (22-25) with respect to the aforementioned dihedral angles around the fatty acid moiety
(
FA) and around the phosphodiester (P) should be similar to those of Type III (18-21)
(
Figure 14) (see Figure S11).
The tetrahydropyran-3,4-diol frameworks were synthesized from readily available
monosaccharide as enantiomerically and configurationally pure carbohydrate isomers,
21
using reported procedures. The resulting diols were acylated with an aromatic non-lipid
surrogate of fatty acid C3-ph-o-O-C11 B (Figure 9), which was synthesized as described
11
in our previous report. The phosphotriester was constructed by the phosphoramidite
21
ꢀ
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 90
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
method and subsequent oxidation gave a mixture of regioisomers of LysoPS analogues
in protected forms. The mixture was separated by column chromatography, and the
structure of each isomer was established by means of several kinds of NMR spectroscopic
analysis.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Biological evaluations of newly synthesized analogues
The results of biological evaluations of the newly synthesized cyclic analogues (18-25)
are summarized in Figure 15 and Table 1 (see also Figures 3 and 4).
Type III analogues (18-21): Notably, the GPR34 potency predicted based on the
conformational analysis was consistent with the biological data: while 18 and 21 are
inactive towards GPR34 (18: EC50 > 3 µM; 21: NA (not available: too weak to evaluate)),
2
0 shows potent activity (EC50 = 0.83 nM) and 19 shows moderate activity (EC50 = 220
nM). Therefore, the +(P) and +(FA) conformation of 20 appears to be crucial for
activation of GPR34, as predicted. In contrast, the -(P) and -(FA) conformation as seen
in 21 is inappropriate for activation of GPR34, and indeed, 21 does not activate GPR34.
The dihedral angle around the fatty acid moiety (FA), +(FA) conformation, appears to be
a more crucial requirement than that around the phosphodiester (P), +(P) conformation,
because 19 (-(P), +(FA)) showed much stronger activity than 18 (+(P), -(FA)).
Intriguingly, while all the conformationally constrained Type III analogues 18-21 showed
increased P2Y10 activity, analogues (20 and 21) with the trans-tetrahydropyran-3,4-diol
framework showed higher activity than those with the cis-framework (18 and 19), and 18
and 21 are rather selective agonists for P2Y10: for P2Y10: 18 EC50 = 420 nM; 19 EC50
300 nM; 20 EC50 = 1 nM; 21 EC50 = 21 nM; for GPR34: 18 EC50 > 3 µM; 19 EC50 = 220
nM; 20 EC50 = 0.83 nM; 21 EC50 = NA.
=
1
1
3
In a previous work, we found that increased planarity of the glycerol C3 conformation
using another sugar structure) resulted in increased potency toward P2Y10 rather than
(
toward GPR34, increasing the P2Y10 selectivity. By analogy with that idea, it seems
reasonable that the critical difference between compounds 14 and 20 lies in the structure
of the C2 “glycerol” moiety: the increased conformational constraint in 20 may increase
the planarity and the potency toward P2Y10, so that 20 becomes a dual agonist of GPR34
and P2Y10.
22
ꢀ
ACS Paragon Plus Environment

Page 23 of 90
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Type II analogues (22-25): The cyclic Type II analogues (22-25) have rather similar
trends in biological activities (Figure 15 and Table 1 and Figures 3 and 4), depending on
the relevant dihedral angles (Figure S5), to those found for the Type III (18-21)
counterparts. The +(P) and +(FA) conformation is again crucial: 24, corresponding to
+(P) and +(FA) conformation (Figure S11), showed very potent activation of both GPR34
and P2Y10 (GPR34: EC50 = 2.9 nM; P2Y10: EC50 = 26 nM). In contrast, 25, which has
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(-(P), -(FA)) conformation (Figure S5), showed significantly reduced potency towards
GPR34 and a slightly weaker activity towards P2Y10 (GPR34: EC50 > 3 µM; P2Y10:
EC50 = 150 nM). Other conformations, (-(P), +(FA)) and (+(P), -(FA)) are seen in 22 and
2
3: 22 showed GPR34 EC50 = 340 nM and P2Y10 EC50 = 1300 nM, while 23 showed
GPR34 EC50 = 63 nM and P2Y10 EC50 = 840 nM. Therefore, again, +(FA) conformation
is crucial for biological activity, particularly for GPR34. Thus, the +(P) and +(FA)
conformation, or at least the +(FA) conformation, is crucial for GPR34 agonistic activity,
with the former conformation resulting in much greater potency than the latter. Among
the cyclic LysoPS derivatives, both Type II and Type III derivatives can take this
important +(P) and +(FA) conformation, and indeed, both show potent agonistic activities
towards GPR34.
2+
Ca mobilization assay LysoPS was reported to induce a rapid increase in intracellular
2
+
2+
Ca concentration ([Ca ]
i
) in HEK293 cells transiently expressing GPR34, which is
7
coupled with G-protein Gai. In order to confirm the potency of our LysoPS analogues
2
+
toward GPR34 observed in the TGFa-shedding assay, we also used this Ca mobilization
assay. The results are summarized in Figure 16. Although there was some difference in
2
+
the rank order of potency between the two assays, the results of the Ca mobilization
assay were broadly consistent with those of TGFa-shedding assay. Among the
2
+
conformationally constrained analogues, 20 expressed higher Ca mobilization activity
than cis-type analogue, while 21 showed almost no activity. Thus, the results of both
assays support our conclusions.
<
Figure 16>
23
ꢀ
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 90
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1
8:1-LysoPS
12
13
19
22
25
3
2
1
00
00
00
0
300
300
200
100
0
200
100
0
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
0
0
0
0
20
20
20
20
40
40
40
40
60
60
60
60
0
0
0
0
20
20
20
20
40
40
40
40
60
60
60
60
0
0
0
0
20
20
20
20
40
40
40
40
60
60
60
60
14
18
21
24
3
2
1
00
00
00
0
300
200
100
0
300
200
100
0
20
3
00
300
200
100
0
300
200
100
0
200
1
00
0
23
3
2
00
00
300
200
100
0
300
200
100
0
100
0
Time (sec)
Figure 16. Calcium mobility assay
Calcium influx upon addition of each compound in mouse GPR34-expressing HEK293
cells is shown. Each experiment was performed in triplicate, and solid and dashed lines
indicate the mean and the S.D., respectively. Black circles: 100 nM, gray circles: 10 nM.
Docking Study of Type III LysoPS derivative onto GPR34
A binding model of a Type I agonist to GPR34 was previously created by homology
modeling, and the binding pose was evaluated and validated by means of SAR studies.14
Since no crystallographic analysis of GPR34 has yet been reported, we used the reported
14
GPR34 model to carry out a ligand-docking study of a Type III agonist. The docking
site of the Type III agonist (14) is tentatively assumed to be similar to that of the Type I
agonist (see Figures S6 and S7).
24
ꢀ
ACS Paragon Plus Environment

Page 25 of 90
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
There seems to be sufficient space between the ligand glycerol moiety and the receptor
transmembrane helices (TMs) 4 and 5 for the Type III ligand to be accommodated at this
site (Figure 17). This is consistent with the experimental SAR finding that Type III
LysoPS derivatives, particularly those with short alkyl ether groups, show potent
biological activities, which should reflect increased binding affinity. This docking model
is also consistent with the existence of a small hydrophobic pocket in GPR34 that serves
as a ligand-binding site with sufficient space to accommodate a propyloxymethyl group
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(like 6), but not a benzyloxymethyl group (like 9) on the glycerol backbone (see Figure
1
7 and Figure 8).
<Figure 17> <Figure 18>
High-affinity compounds such as 14 are expected to show stable binding poses, that is, in
perturbation by means of REST calculations, the binding poses in terms of the two
dihedral angles will be well conserved (Figure 18). In the case of the potent agonist 14,
these dihedral angles are distributed mainly into the (+(P), +(FA)) and (+(P), -(FA))
conformational regions in water, implying that molecules in the (+(P), +(FA))
conformational region are indeed available for binding to GPR34 (Figure 12). Formation
of conformations in this region was also found in the cases of strong GPR34 agonists 2R
and 20 (see also Figure S6).
Top View
VI
VII
VI
VII
V
V
4
5
1
4
1
2
6
1
4
III
IV
I
III
II
IV
OEt
4
O
O
1
O
P
O
O
3
3
O
2
6
NH3
O
4
5
1
5
2
O
O
1
4
6
O
1
4
P -O -C -C (P): 148.6°
1
2
3
4
C -C -O -C (FA): 92.2°
3
4
5
6
Figure 17. Binding model of Type III agonist (14) with GPR34. Space-filling by the
ethoxy ether group results in tight contact with TM4 and TM5.
25
ꢀ
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 90
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
OEt
OEt
O
OH
P
O
O
OH
P
O
HO
O
O
HO
O
O
NH2
O
O
NH2
O
O
O
O
O
O
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
14 in water
180
14 Bound to GPR34
180
1
20
60
0
120
60
0
-180
-120
-60
0
60
120
180
-
180
-120
-60
0
60
120
180
-60
-60
120
180
-
120
180
-
-
-
Torsion P
Torsion P
Figure 18. Converging conformational distribution upon binding of potent agonist
4 to GPR34 receptor. In the REST calculations, the distribution of conformations
1
with respect to the two dihedral angles of 14 in water converged towards a single
region (+(P), +(FA)) upon binding to GPR34
ꢀ
Supporting evidence for the docking study: divergent responses of GPR34
homologues to compounds 14 and 20
Genes encoding GPR34 are conserved in a wide range of vertebrates, including
1
zebrafish. While GPR34 proteins show high levels of homology with > 90% amino acid
1
identity in mammals, only ~50% of the amino acids are conserved in zebrafish. When
we aligned the amino acid sequences of GPR34 from human, mouse and zebrafish (note
that there are two zGPR34, types a and b), we noticed that the amino acid identities in
TMs 4 and 5 are low, compared with other TMs (Table 2). Since GPR34 responds to
LysoPS (18:1) across species, amino acid residues important for the recognition of
LysoPS (18:1) should be conserved among species. We examined the agonistic activities
of Type III compounds 14 and 20 towards zGPR34 types a and b. Although compounds
1
4 and 20 showed potent agonism at mammalian GPR34, they were poor agonists for
zGPR34s (Table 3, Figure 19 and Figure S8). Since the amino acid sequences of
transmembrane domains 4 and 5 (TMs 4 and 5) are relatively poorly conserved among
26
ꢀ
ACS Paragon Plus Environment

Page 27 of 90
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
species (Figure S9: alignment), it seems plausible that the difference in activities between
Type I LysoPS (18:1) and Type III compounds (14 and 20) could be related to the
difference in TMs 4 and 5. TMs 4 and 5 are located close to the fatty acid moiety and the
glycerol moiety of the ligand, respectively, in the homology model, which would be
consistent with this idea. Thus, some amino acids in the TM4 and TM5 in mammalian
GPR34, which are not conserved in zGPR34 type a and type b, might be involved in
recognizing the fatty acid surrogates and C2-based glycerol moiety of 14 and 20. These
considerations provide some support for the validity of the present homology model.
ꢀ
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
<Table 2> <Table 3>< Figure 19>
Human
2
2
5
0
15
1
8:1-LysoPS
4
1
0
5
0
1
20
1
-1
-9
-8
-7
-6
-5
-12
0
-10
10 10 10 10 10
1
0
10 10
Zebrafish type a
Zebrafish type b
2
2
1
1
5
0
5
0
5
0
30
2
2
5
0
15
10
5
0
-9
-8
-7
-6
-5
-9
-8
-7
-6
-5
-12
0
-10
-12
0
-10
1
0
10
10
10
10
10
10
10
10
10
1
0
10
10
10
Compound (M)
Figure 19. Differences in the potency of 18:1-LysoPS, 14 and 20 towards human and
zebrafish GPR34 receptors The activation of human and zebrafish GPR34s by 18:1-
LysoPS and compounds 14 and 20 was evaluated. Experiments were done in the same
27
ꢀ
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 90
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
manner as in Figures 3 and 4. Data are mean and SEM (standard error of the mean) for 3
−
4 independent experiments.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Conclusion
In this work, we addressed the question of whether Type III LysoPS derivatives are
biologically active, even though they are not found in nature. Interestingly, the biological
activity exhibited a conformation dependency with respect to the C2-glycerol (“glycol
surrogates”) moiety that was similar to the dependency observed in physiological Type II
derivatives. Conformational analysis of such compounds by molecular dynamics led to
the synthesis of conformationally-constrained cyclic derivatives. The chemical work and
biological results are focused on the steric requirements of receptor binding and
optimization of the conformational equilibria, resulting in potent and selective ligands
agonists for LysoPS receptors GPR34 and P2Y10, even although mammalians at least do
not use them. This finding is intriguing in terms of molecular evolution. We suggest that
only Type I and Type II LysoPS derivatives, but not Type III derivatives, may have
emerged as the preferred species during evolution because Type I and II are readily inter-
transformable by a single favorable chemical rearrangement, acyloxy 1,2-migration,
whereas involvement of Type III would require at least two migrations, which would be
relatively unfavorable, particularly the 1,3-acyloxy migration (between III-enantiomer
and III, Figure 1(b)).22 In terms of medicinal chemistry, the importance and novelty of
new selective and non-selective ligands for GPCRs of lysoPS were identified, by
underlining the information that was gained by conformational analysis.
28
ꢀ
ACS Paragon Plus Environment

Page 29 of 90
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Table 1. Biological assay results of Type III LysoPS Analoguesa
GPR34
P2Y10
EC50
(LogEC50)
[E
max
]
<RIA>
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
n
Positive Control
110 nM
1.6 nM
(
[
-6.96 ± 0.11)
12.2 ± 0.6%]
(-8.81 ± 0.10)
[11.1 ± 1.1%]
<1 ± 0>
<1 ± 0>
8
8
18:1-LysoPS
230 nM
20 nM
(
[
<
-6.63 ± 0.06)
(-7.71 ± 0.09)
[13.6 ± 1.2%]
<0.083 ± 0.03>
8
8.3 ± 1.7%]
0.33 ± 0.07>
8
deoxy-LysoPS C2 (18:1)
> 3 μM
66 nM
1a
(> -5.5)
(-7.18 ± 0.05)
[10.6 ± 1.4%]
<0.02 ± 0.008>
3
[
NA]
NA>
3
<
deoxy-LysoPS C3 (18:1)
540 nM
(-6.27 ± 0.07)
6.2 ± 1.4%]
20 nM
1b
(-7.71 ± 0.10)
[13.7 ± 1.4%]
<0.11 ± 0.01>
3
[
<0.12 ± 0.05>
3
deoxy-LysoPS C4 (18:1)
320 nM
53 nM
1c
(-6.50 ± 0.06)
4.4 ± 0.7%]
(-7.28 ± 0.07)
[13.1 ± 2.2%]
[
<
0.12 ± 0.04>
<0.029 ± 0.011>
3
3
deoxy-LysoPS C5 (18:1)
470 nM
240 nM
1d
(-6.33 ± 0.07)
4.6 ± 1%]
0.076 ± 0.012>
(-6.62 ± 0.07)
[10.8 ± 0.8%]
<0.0061 ± 0.001>
3
[
<
3
2R
24 nM
250 nM
(
[
-7.62 ± 0.03)
12.8 ± 0.9%]
(-6.61 ± 0.08)
[12.4 ± 1.5%]
<0.0054 ± 0.004>
3
<4.8 ± 1.4>
3
2S
170 nM
37 nM
(
[
<
-6.76 ± 0.08)
11.2 ± 1.9%]
0.56 ± 0.17>
(-7.43 ± 0.06)
[11.3 ± 1.2%]
<0.034 ± 0.02>
3
3
3R
250 nM
79 nM
(
[
<
-6.59 ± 0.03)
(-7.10 ± 0.11)
[11.4 ± 2.5%]
<0.02 ± 0.003>
3
6.1 ± 1.2%]
0.22 ± 0.07>
3
3S
> 3 μM
390 nM
(
> -5.5)
(-6.41 ± 0.09)
[11.6 ± 2%]
<0.0034 ± 0.0016>
[NA]
<NA>
29
ꢀ
ACS Paragon Plus Environment