The structure-activity relationship of the series of non-peptide small antagonists for p56lck SH2 domain

Article abstract of DOI:10.1016/j.bmc.2006.06.059

The antagonists for the SH2 domain are regarded as novel therapeutic candidates for cancer, autoimmune disease, and chronic inflammatory disease. Previously, we identified rosmarinic acid (α-o-caffeoyl-3,4-dihydroxyphenyl-lactic acid; RosA) from Prunella vulgaris as an antagonist for the p56lck SH2 domain by screening natural products. RosA not containing phosphotyrosine surrogate had a considerable inhibitory activity for T-cell antigen receptor (TCR)-induced interleukin (IL)-2 expression, and subsequent T-cell proliferation in vitro cell assay. To investigate the structure-activity relationship of RosA and to identify a novel p56lck SH2 antagonist with more potent in vitro T-cell inhibitory activity, we synthesized several analogs of RosA by using rational design. All synthesized compounds were tested in vitro binding activity for the SH2 domain and in vitro T-cell inhibitory activity. All four hydroxyl groups of RosA were essential for binding with the p56lck SH2 domain and T-cell inhibitory activity. Unexpectedly, conformationally less constrained analogs 4 and 9 showed a more potent binding affinity for the SH2 domain than that of RosA, and chirality of the analog did not play an important role in protein binding. We successfully identified several RosA analogs with a more potent T-cell inhibitory activity than that of RosA. Overall results revealed important structural requirements of the p56lck SH2 antagonists for in vitro T-cell inhibitory activity and in vitro protein binding activity.

Full text of DOI:10.1016/j.bmc.2006.06.059

Bioorganic & Medicinal Chemistry 15 (2007) 3938–3950
The structure–activity relationship of the series of non-peptide small
antagonists for p56lck SH2 domain
See-Hyoung Park,a, Hyun-Sik Oh,b Mi-Ae Kang,a,à Hyeongjin Cho,b
Joshi Bishnu Prasad,b Jonghwa Wona,* and Keun-Hyeung Leeb,*
aSignal Transduction Laboratory, Mogam Biotechnology Research Institute, 341 Pojung-Ri, Koosung-Myun,
Yongin-City, Kyunggi-Do, 449-910, Republic of Korea
bDepartment of Chemistry, Inha University, 253 Younghyong-Dong, Nam-Gu, Inchon-City, 402-751, Republic of Korea
Received 18 May 2006; revised 23 June 2006; accepted 24 June 2006
Available online 5 April 2007
Abstract—The antagonists for the SH2 domain are regarded as novel therapeutic candidates for cancer, autoimmune disease, and
chronic inflammatory disease. Previously, we identified rosmarinic acid (a-o-caffeoyl-3,4-dihydroxyphenyl-lactic acid; RosA) from
Prunella vulgaris as an antagonist for the p56lck SH2 domain by screening natural products. RosA not containing phosphotyrosine
surrogate had a considerable inhibitory activity for T-cell antigen receptor (TCR)-induced interleukin (IL)-2 expression, and sub-
sequent T-cell proliferation in vitro cell assay. To investigate the structure–activity relationship of RosA and to identify a novel
p56lck SH2 antagonist with more potent in vitro T-cell inhibitory activity, we synthesized several analogs of RosA by using rational
design. All synthesized compounds were tested in vitro binding activity for the SH2 domain and in vitro T-cell inhibitory activity.
All four hydroxyl groups of RosA were essential for binding with the p56lck SH2 domain and T-cell inhibitory activity. Unexpect-
edly, conformationally less constrained analogs 4 and 9 showed a more potent binding affinity for the SH2 domain than that of
RosA, and chirality of the analog did not play an important role in protein binding. We successfully identified several RosA analogs
with a more potent T-cell inhibitory activity than that of RosA. Overall results revealed important structural requirements of the
p56lck SH2 antagonists for in vitro T-cell inhibitory activity and in vitro protein binding activity.
ꢀ 2006 Elsevier Ltd. All rights reserved.
1. Introduction
osteoporosis, and chronic inflammatory disease.4–7
The p56lck SH2 domain, found in T cells and NK
cells, is indispensable for efficient initiation of T-cell
antigen receptor (TCR)-signaling as well as activa-
tion-induced intracellular Ca2+ mobilization, and pro-
duction of cytokine such as interleukin-2 (IL-2).8,9
Thus, the antagonists of the p56lck SH2 domain have
therapeutic potential for immune disorders such as
organ transplantation rejection, graft-versus-host dis-
eases, autoimmune diseases, and chronic inflammatory
diseases. A peptide library screen revealed the specific-
ity of phosphopeptides for various SH2 domains and
identified a lead peptide, AcpYEEIE, that showed a
potent binding affinity for the p56lck SH2 domain
(IC50 = 0.1–1.8 lM).3 X-ray structural studies about
p56lck/Src SH2 domains in complex with this phos-
phopeptide revealed that the peptide was anchored
by insertion of the pY and pY + 3 side chains into
two pockets of SH2 domains (‘‘two-pronged plug
two-holed socket’’ model).10,11 Moderately high-affini-
ty peptide-based ligands for the p56lck SH2 domain
have been recently and successfully synthesized from
AcpYEEIE by using rational design.12–14 Even though
The src homology-2 (SH2) domain is a highly con-
served non-catalytic module consisting of 100 amino-
acid residues that is found in many intracellular
signal-transduction proteins.1 The SH2 domains specif-
ically recognize phosphotyrosine (pY)-containing pro-
teins with high affinity and each SH2 domain
appears to have a selectivity toward particular phos-
photyrosyl peptide sequences.2,3 Specific antagonists
of the SH2 domains, which inactivate inappropriately
hyperstimulated cell signaling, can be developed as
novel therapeutic agents to treat a broad range of
human diseases such as cancer, autoimmune disease,
Keywords: T-cell inhibition; Non-peptide; Antagonist; Inhibitor; SH2;
Binding affinity.
*
Corresponding authors. Tel.: +82 32 860 7674; fax: +82 32 867
5604; e-mail: leekh@inha.ac.kr
Present address: Department of Comparative Pathology, One Shields
Avenue, University of California, Davis, CA 95616, USA.
Present address: Department of Immunology, One Shields Avenue,
University of California, Davis, CA 95616, USA.
à
0968-0896/$ - see front matter ꢀ 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2006.06.059

S.-H. Park et al. / Bioorg. Med. Chem. 15 (2007) 3938–3950
3939
high-affinity peptide-based antagonists for the SH2
domain (IC50 = 10 nM to 1 lM) were successfully syn-
thesized, most did not show in vitro cell inhibitory
activity for several reasons. First, phosphotyrosine,
essential for SH2 domain recognition, is hydrolytically
unstable in the presence of protein tyrosine phospha-
tase.15 The peptide antagonists could not pass through
cell lipid membrane due to high negative charge or
peptidic character of the peptide antagonists. Recently,
several non-hydrolyzable pY surrogates were applied
into SH2 peptide antagonists to develop novel SH2
antagonists with less peptidic character and with a re-
duced number of negative charges.14,16,17 However,
still most of peptide-based antagonists containing even
pY surrogate did not exhibit activity in vitro cell
assay.
ture–activity relationships and identifying new lead
compounds for the p56lck SH2 domain which will
exhibit more potent activity in in vitro cell assay.
All synthesized compounds were tested for in vitro
binding affinity for the p56lck SH2 domain as well
as in vitro T-cell inhibitory activity by using IL-2
gene expression assay. We successfully characterized
structural requirements of this small chemical antago-
nist for interacting with p56lck SH2. All four
hydroxyl groups of the compounds were essential
for the binding with the SH2 domain. Interestingly,
4 and 9 that were conformationally flexible inhibitors,
exhibited more potent binding affinity than did RosA,
and chirality of the analog did not play an important
role in the protein binding. We successfully identified
some analogs of RosA which showed more potent
inhibitory activity for T-cell antigen receptor (TCR)-
induced interleukin (IL)-2 expression than that of
RosA. Overall results revealed important structural
requirements of the small chemical antagonists for
the SH2 domain binding activity and T-cell inhibitory
activity.
Even though screening of natural products must be a
useful and conceptually straightforward strategy for
the development of therapeutic agents, there is little
research about chemical antagonist for SH2 do-
mains.18,19 Recently, we successfully identified a small
chemical antagonist for the p56lck SH2 domain by
screening natural products from Prunella vulgaris.19
The small chemical antagonist was characterized to be
rosmarinic acid (a-o-caffeoyl-3,4-dihydroxyphenyl-lactic
acid; RosA) which was frequently found in medicinal
plant, herbs, and spices.20 Interestingly, RosA that does
not contain pTyr had considerable inhibitory activity on
the interactions between p56lck SH2 and AcpYEEIE in
ELISA and had inhibitory activity for T-cell antigen
receptor (TCR)-induced interleukin (IL)-2 expression
and subsequent T-cell proliferation in vitro.19,21 Fur-
thermore, the compound marginally improved allograft
survival rate in the animal model and the combination
of RosA and Rapamycin prolonged synergistically sur-
vival in the allograft mouse model.22
2. Results and discussion
2.1. Design and chemical synthesis of the RosA analogs
To design the analogs of RosA with a potent binding
affinity, information about the binding mode of RosA
with p56lck SH2 domain is needed. However, X-ray
crystallographic information is not available for the
binding mode of RosA with the p56lck SH2 domain.
The previous research in which caffeic acid-pYEEIE
was identified as the most potent p56lck SH2 antagonist
through the screening of unnatural amino acid at the
pY-1 position of -pYEEIE may provide a clue for the
binding mode of RosA to the p56lck SH2 domain.23
As shown in Figure 1, RosA (I) shared the similar moi-
ety of the peptide inhibitors (II-IV) for the p56lck SH2
In this study, we designed and synthesized RosA-
based analogs with the aim of investigating the struc-
OH
HO
H2O3PO
COOH
COOH
O
H
O
H
N
N
HN
OH
O
N
N
COH
O
O
H
O
O
H
O
O
HO2C
HO
HO
HO
I (compound 6)
HO
II
H2O3PO
O
H2O3PO
CH3
N
N
O
H
H
H
N
N
N
HN
HN
N
O
O
O
H
O
O
O
OMe
HO2C
HO
O
III
IV
Figure 1. The structures of rosmarnic acid (I) and p56lck SH2 peptide inhibitors (II–IV).

3940
S.-H. Park et al. / Bioorg. Med. Chem. 15 (2007) 3938–3950
domain previously reported by us as well as the other
researchers.23–25 Comparison of the structures of the
RosA and the peptide inhibitor revealed two character-
istics. The caffeic acid moiety of RosA may correspond
to the amino acid at the pY-1 position of the peptide
inhibitors. And the dihydroxyphenylacetic acid (DOPL)
moiety may correspond to the pY of the peptide inhib-
itors. Interestingly, the stereochemistry of the DOPL
(R form) moiety of RosA is the reverse of that of the
pY (S form) of the peptide inhibitors. Thus, we synthe-
sized each enantiomer of the RosA analog containing
amide bond by coupling caffeic acid with D-DOPA
(3,4-dihydroxyphenyl-L-alanine) and L-DOPA, respec-
tively, as shown in Scheme 1. There is some possibility
that RosA interacts with the pY+3 binding pocket,
but this possibility is likely low because RosA has a con-
siderably different structure compared to small chemical
inhibitors screened for the pY+3 binding site of the Lck
SH2 domain.26 According to the molecular model of
RosA, there is a considerable conformational constraint
in the caffeic acid moiety, and the phenyl and vinyl
chains of this moiety are locked into a coplanar orienta-
tion. To find out the requirement of the coplanar orien-
tation of this moiety for its binding affinity, the vinyl
group of caffeic acid was reduced or the caffeic acid moi-
ety was replaced with 3,4-dihydroxybenzoic acid. Even
though RosA containing a carboxylic group exhibited
considerable in vitro cell activity, considering the
requirement of cell penetration for in vitro cell activity,
the acid group of the compound was modified into var-
ious aliphatic ester groups, as shown in Scheme 2. As the
hydroxyl group of the catechol moiety of RosA was par-
tially negatively charged under physiological condition,
the hydroxyl group might play an important role in
the interaction with the SH2 domain as well as in cell pe-
netrating ability. To explore the importance of the cate-
chol hydroxyl groups for binding activity and in vitro
cell activity, each hydroxyl group of RosA was
methylated.
The proposed analogs of RosA were synthesized as out-
lined in Schemes 1 and 2. As shown in Scheme 1, the ser-
ies of the analog 2 containing amide bond were
synthesized by a coupling reaction of (R) or (S) 3,4-
dihydroxyphenyl-alanine (DOPA) methylester 1 with
various dihydroxyphenyl acid compounds including caf-
feic acid as the similar method previously described.27
Hydrolysis of the ester of 2 by the treatment of LiOH
provided the corresponding acid 3. RosA analogs con-
taining ester bond were synthesized as outlined in
Scheme 2. Reduction of RosA (I) by H2 in the presence
of Pd–C provided compound 4 in a satisfactory yield.
O
O
O
HO
HO
HO
a
OMe
HO
HO
OMe
R
b
OH
R
NH2
HN
HO
HN
O
O
1
2
3
Scheme 1. Reagents and condition: (a) RCOOH, PyBOP, TEA, DMF, DCM; (b) LiOH, 0 ꢁC, DMF/H2O.
O
HO
HO
OH
OH
OH
O
O
4
a
O
O
O
O
HO
HO
OH
OH
HO
HO
OH
OH
OH
b
OR
O
O
RosA (I)
14-17
c
O
O
HO
HO
OH
OH
O
O
5
Scheme 2. Reagents and conditions: (a) H2, Pd/C; (b) SOCl2, ROH, R = CH3, CH2CH3, CH2CH2CH3, CH(CH3)2; (c) isobutylene, dioxane/À10 ꢁC.

S.-H. Park et al. / Bioorg. Med. Chem. 15 (2007) 3938–3950
3941
Activation of the carboxyl group of RosA, coupled with
various alcohols provided ester analogs 14–17 in appre-
ciable yields. The reaction of RosA (I) with isobutylene
in the presence of acid provided compound 5. All syn-
thesized analogs employed in this study were further
purified by preparative HPLC with a Vydac C18 column
(22 mm · 250 mm) using a water (0.1% TFA)-acetoni-
trile (0.1% TFA) gradient. The purity (generally >95%
by RP analytical HPLC, UV214nm) and the retention
time of the compound were analyzed by using analytical
HPLC (5–45% linear gradient of acetonitrile during
40 min). logP of each compound was measured accord-
ing to a literature method and compared with the reten-
tion time on a Vydac C18 column in HPLC to investigate
the hydrophobicity of the compounds.28
RosA. This is an encouraging result because compound
7 has a different scaffold and was easily synthesized by
coupling caffeic acid with DOPA. Furthermore, we com-
pared the binding affinity between compound 7 and
RosA by using BIAcore. As shown in Figure 3, com-
pound 7 showed almost similar binding affinity to that
of RosA in this competition assay. We concluded that
replacement of ester bond with amide bond of RosA
did not affect the binding affinity for the SH2 domain.
Unexpectedly, compound 8 that was an enantiomer of
compound 7 showed the similar binding affinity to those
of compound 7 and RosA, which strongly suggested
that the stereochemistry of the amino a carbon of the
antagonist was not a critical factor for the interaction
with the SH2 domain. Interestingly, compounds 4 and
9, which have a less conformationally constrained moi-
ety than RosA, exhibited a more potent binding affinity
than did RosA. Compounds 12 and 13 which have
dihydroxylphenyl acetic acid and dihydroxylbenzyl acid
moiety instead of caffeic acid moiety showed more po-
tent binding affinity than did RosA and compound 7.
The overall results indicated that planarity of caffeic
acid moiety of RosA was not essential for binding with
the p56lck SH2 domain. Compounds 14–17 which have
a different size of alkyl ester group showed a less potent
binding affinity than did RosA. Particularly, compound
16 containing n-propylester showed the lowest binding
affinity among the active ester analogs, and compound
18 containing long aliphatic ester did not show binding
affinity even at the concentration of 500 lM. The bind-
ing affinity of the ester analog series decreased with
increasing carbon number of the alkyl chain of the ester.
Compounds 19–22 which contained at least one methyl-
ated hydroxyl group at the phenyl ring moiety did not
show any binding affinity, which strongly suggested that
all hydroxyl groups of RosA were critically required for
the interaction with the p56lck SH2 domain. As catechol
compounds were reported to have covalent binding to
Src family SH2 domains,29 we investigated the binding
affinity of DOPL (23) and ( )-DOPA (24). Both com-
pounds did not show binding affinity for the p56lck
SH2 domain.
2.2. In vitro binding activity of RosA and its analogs
The binding affinity of the compounds for the p56lck
SH2 domain was measured by using a competitive assay
as described.23 As shown in Figure 2, RosA and the ac-
tive analogs inhibited the binding of EPQpYEEIPIYL
with the p56lck SH2 domain in a concentration-depen-
dent manner and a control phosphopeptide, Ac-
pYEEIE, exhibited approximately 1.8 lM of IC50
value, which is highly consistent with the previously
reported value.23,24 The measured IC50 values of each
compound are summarized in Table 1. Compound 7
bound to the SH2 domain with IC50 = 20 lM, that is,
just 10-fold less potent than the phosphopeptide,
AcpYEEIE. Interestingly, compound 7 containing amide
bond showed a more potent binding activity than did
2.3. In vitro T-cell inhibitory activity of RosA and its
analogs
All employed compounds were evaluated for their pro-
tein binding affinity as well as for their ability to inhibit
T-cell activation. It was demonstrated that point muta-
tion in the p56lck SH2 domain, to disrupt possible inter-
actions with phosphotyrosine-containing molecules,
abrogated TCR-induced Ca2+ flux and subsequent
IL-2 promoter activation.8,9,30 Thus, in vitro T-cell
inhibitory activity was measured by using the blocking
of TCR-induced IL-2 gene activation as described in
Section 3. The commercially available immune suppres-
sive drug, Cyclosporin A (CsA), was used as a positive
control in this assay.31 As shown in Figure 4, in vitro cell
inhibition activity was measured at two different concen-
trations (10 lM and 30 lM). A positive control, CsA,
inhibited T-cell activation almost 100% at 10 lM,
whereas, RosA inhibited about 50% under the same
condition. AcpYEEIE with a potent binding affinity
Figure 2. The inhibition of compounds for p56lck SH2-phosphopep-
tide interaction in ELISA. Each of the compounds is added to a
solution containing GST-p56lck-SH2 domain and incubated and then
the mixture is added to biotinyl-e-aminocaproyl-EPQpYEEIPIYL
coated wells in duplicate. After washing, the degree of interaction
between SH2 domain and biotinyl-e-aminocaproyl-EPQpYEEIPIYL
is detected by incubation with an anti-GST polyclonal antibody,
followed by incubation with a peroxidase-conjugated anti-rabbit
antibody and peroxidase substrate solution. The color development
is monitored at 450 nm. Each value was calculated from three
independent experiments performed in duplicate, which provided a
standard deviation below 20%.

3942
S.-H. Park et al. / Bioorg. Med. Chem. 15 (2007) 3938–3950
Table 1. logP, Lck SH2 binding affinity, and T-cell inhibitory activity of the compounds
d
tR (min)
Compound
Structure
IC50 value (lM)
T-cell Inhibitionb (%)
logPc
ELISAa
O
HO
HO
OH
OH
OH
OH
OH
OH
OH
6(I)
24
20
21
15
13
2
2
2
1
1
40.4 1.17
0.834
27
17
17
13
13
O
HN
HN
O
O
O
HO
HO
OH
OH
7
8
9
4
10.3 0.92
0.758
0.731
0.503
0.493
O
O
HO
HO
OH
OH
11 1.1
O
O
HO
HO
OH
OH
7.6 0.81
O
O
HO
HO
OH
OH
0
4.52
HN
O
O
HO
HO
OH
OH
OMe
OMe
10
11
12
13
14
15
71
32
21
19
80
91
2
3
2
1
4
3
61.6 3.23
7.5 0.49
23 1.53
1.386
0.344
0.138
0.125
1.479
1.499
22
20
18
17
30
32
HN
O
O
HO
HO
OH
OH
HN
HN
O
O
HO
HO
OMe
OMe
OH
OH
O
O
HO
HO
0.4 0.06
HN
OH
OH
O
O
HO
HO
OH
OH
OMe
54.7 5.09
O
O
O
O
HO
HO
OH
OH
O
22.7
1
O
O
HO
HO
OH
OH
O
16
145
4
87.5 4.1
1.520
34
O
O

S.-H. Park et al. / Bioorg. Med. Chem. 15 (2007) 3938–3950
3943
Table 1 (continued)
d
tR (min)
Compound
Structure
IC50 value (lM)
T-cell Inhibitionb (%)
logPc
ELISAa
O
HO
HO
OH
OH
O
O
17
110
127
5
3
86.1 5.86
80.7 4.5
29.4 0.56
1.833
33
O
O
O
O
HO
HO
OH
OH
5
1.995
1.683
32
29
O
O
O
OMe
OH
18
>500
HO
HO
N
H
O
OH
O
O
HO
HO
OH
O
OMe
OMe
OMe
OMe
19
20
21
22
>500
>500
>500
>500
2.2 0.2
5.4 0.4
3.5 0.28
9.4 0.43
1.313
1.371
1.522
1.535
25
26
29
28
HN
HN
HN
HN
O
O
HO
HO
O
OH
O
O
HO
HO
O
O
O
O
O
O
OH
OH
O
OH
OH
23
24
>500
>500
2.0 0.2
0.6 0.2
—
—
—
—
HO
O
COOH
HO
*
NH2
HO
a The average IC50 values were calculated from three independent experiments performed in duplicate, which provided a standard deviation below
20% and the IC50 value of AcpYEEIE was 1.8 lM.
b T-cell Inhibition (%) values were obtained at 10 lM and from three independent experiments performed in duplicate, which provided a standard
deviation below 10%. Cyclosporin A as a control inhibited T-cell activation completely at 10 lM, whereas RosA inhibited T-cell activation by
about 40% at 10 lM.
c LogP values were calculated from three independent experiments performed in duplicate, which provided a standard deviation below 5%.
d Retention times were obtained with analytical HPLC system (5–45% linear gradient of acetonitrile during 40 min).
(IC50 = 1.8 lM) did not show T-cell inhibition activity
even at 100 lM, which can be attributed to its low cell
penetrating activity as well as low stability against phos-
phatase. As in vitro cell inhibition activity of RosA ana-
logs was more clearly differentiated at the concentration
of 10 lM, the percent inhibition of T-cell activation at
this concentration is summarized in Table 1. Both com-
pounds 7 and 8 (IC50 = 20 lM) were not as effective as
RosA (IC50 = 24 lM), for inhibiting IL-2 expression
in vitro cell assay; both compounds are four times less
active than RosA in in vitro cell assay. Considering their
low logP value and the short retention time on the C18
reverse phase HPLC column, the low in vitro cell activ-
ity of both compounds must be due to their low cell pen-
etration. This suggestion was supported by the fact that
compound 10 (methyl ester form of compound 7)
showed increased inhibition activity than did RosA in
in vitro cell assay even though the binding affinity of

3944
S.-H. Park et al. / Bioorg. Med. Chem. 15 (2007) 3938–3950
gated the stability of the compounds in buffer solution at
a
pH 7.4 by using LC–mass spectrometer; the ester com-
pounds were rapidly converted into the corresponding
acid or unidentified small molecule maybe due to the
proximity effect of hydroxyl group of catechol moiety
and the half lives of the compounds were around 3–
4 h (data not shown). Unexpected low logP values of
these compounds can be explained by the rapid hydroly-
sis of the ester group of the compounds. We assumed
that low in vitro T-cell inhibitory activities of the com-
pounds might be due to the low cell penetrating ability
and low stability in the buffer solution.
The acid group of RosA was modified into various alkyl
esters and in vitro cell activity was compared. Methyla-
tion of RosA (compound 14) resulted in the decrease of
protein binding affinity three times (IC50 = 80 lM) but
an increase of the inhibition activity in in vitro cell
assay. Interestingly, IL-2 gene expression was totally
inhibited at 30 lM concentration of compound 14,
whereas only 50% inhibition was observed at 30 lM
concentration of RosA as shown in Figure 4. The com-
parison between protein binding affinity and in vitro cell
activity of compounds 6 and 14–18 showed the follow-
ing trend; as the chain length of C-terminal ester group
was increased, their binding activity was steadily
reduced, whereas in vitro cell inhibition activity was im-
proved by an increase of the hydrophobicity if the
antagonist had still considerable binding affinity. This
result indicates that there are two factors, the protein
binding activity and cell penetrating activity of the
antagonist, which contribute to determining in vitro cell
activity. As the chain length of the C-terminal ester
group of the antagonist increased for the cell penetrating
activity, the binding activity for the protein was reduced.
Thus, the analog with potent in vitro cell inhibition
activity must have optimal balance between binding
affinity and hydrophobicity.
b
Figure 3. The inhibition of the RosA (a) and compound 7 (b) for GST-
p56lck-SH2 domain by using BIAcore. The BIAcore instrument used
in this study was manufactured by BIAcore AB (BIAcore 3000,
Uppsala, Sweden). First, biotinyl-e-aminocaproyl-EPQpYEEIPIYL at
10 lg/ml in the running buffer (50 mM Tris, 150 mM NaCl, pH 7.5)
was injected at a flow rate 5 ll/min for 5 min onto streptavidin coated
chip surface. For the competitive binding assay, the GST-p56lck-SH2
domain was diluted to 100 nM in the running buffer. Samples were
titrated into GST-p56lck-SH2 domain from 100 to 0 lM and allowed
to come to equilibrium in the running buffer; 10 ll of each reaction
mixture containing samples and SH2 protein was passed over the
surface with looped injections at a flow rate of 5 ll/min, with 5 ll of
100 mM NaCl and the resonance unit was measured in the running
period.
Compound 15 having a considerable protein binding
affinity (IC50 = 90 lM), exhibited the lowest T-cell
inhibitory activity among the ester analog series. As
the ethylester analog had an advantage over the methy-
lester analog because the ethylester analog was hydro-
lyzed to release less toxic ethanol, we measured
in vitro T-cell inhibitory activity of compound 15 several
times. Compound 15 exhibited low T-cell inhibitory
activity in several independent experiments, but we
could not explain this low in vitro cell inhibitory activity
of compound 15 by its hydrophobicity. Compound 18
with no protein binding affinity showed marginal cell
inhibition activity, which suggested the possibility that
hydrophobic compounds may show in vitro cell activity
through non-specific effects in in vitro cell assay. Meth-
ylation of any hydroxyl group of catechol moiety of the
inhibitors (19–22) resulted in the loss of binding affinity
as well as in vitro cell inhibition activity.
compound 10 (IC50 = 70 lM) was lower than that of
compound 7. Compounds 4 and 9, with a potent binding
affinity (IC50 = 13–15 lM) showed almost no inhibition
activity in vitro cell activity. Both compounds showed a
relatively low logP value compared to those of the other
analogs with in vitro cell activity and no in vitro cell
activity of compounds 4 and 9 must be due to their
low cell penetration ability. Compounds 11–13 with a
high protein binding affinity exhibited a low T-cell inhib-
itory activity. Considering the low logP, we assumed
that low cell inhibitory activity of the analogs might
be due to low cell penetration. However, we observed
the discrepancy between logP and HPLC retention time
of compounds 11–13.
As SH2 domains recognize the proteins containing pY,
which is dephosphorylated by protein tyrosine phospha-
tase (PTPase), the SH2 inhibitors may interact with
PTPase. Compounds 6–8 that showed potent binding
affinity for p56lck SH2 domain were selected and
We assumed that an inconsistency between logP value
and the retention time must be due to the low stability
of these compounds in buffer solution. Thus, we investi-

S.-H. Park et al. / Bioorg. Med. Chem. 15 (2007) 3938–3950
3945
Figure 4. The intracellular inhibition of the rosmarinic acid and its derivatives (IL-2 Luc assay). Jurkat cells were transfected with the IL-2 Luc
reporter, treated with 10 lM and 30 lM of Cyclosporin A, rosmarinic acid, and it derivatives for 2 h and activated with immobilized anti-CD3
antibody or PMA plus ionomycin for 16 h. After stimulation, the cells were washed, lysed, and assayed for luciferase activity with microplate
luminometer. Each value was calculated from three independent experiments performed in duplicate, which provided a standard deviation below 10%.
examined against human PTPase 1B. Enzyme assay
proved that all tested compounds did not show inhibi-
tion activity even at 100 lM (data not shown). RosA,
a lead compound, was proven to have no inhibition
activity for Lck, a protein tyrosin kinase.19 These results
indicate that the small chemical antagonists employed in
this study might have a selectivity for the p56lck SH2
domain over other proteins of T-cells.
methyl alcohol, dicyclohexylcarbodiimide (DCC),
N-methylpyrrolidone (NMP), and N-hydroxybenzo-
triazole (HOBt) were purchased from Applied Biosys-
tems, Inc. (Foster city, CA, USA). Trifluoroacetic acid
(TFA), (+)-biotin, e-aminocaproic acid, 3,4-dihydroxy-
cinnamic acid, 3,4-dihydroxy-L-phenylalanine, 3,4-dihy-
droxy-D-phenylalanine, and N,N-dimethylformamide
(DMF) were purchased from Aldrich (Milwaukee, WI,
USA). SH2-GST fusion Lck (120–226) and polyclonal
rabbit anti-GST antibody were purchased from Santa
Cruz Biotechnology (Santa Cruz, CA, USA). Horserad-
ish peroxidase-conjugated mouse anti-rabbit antibody,
peroxidase substrate (1-step Turbo TMB-ELISA, trim-
ethylbenzidine), and streptavidin coated 96-well plates
were purchased from Pierce (Rockford, IL, USA). RosA
was obtained from Indofine Chemical Company
(Somerville, NJ, USA). CyclosporinA was kindly pro-
vided by Dr. G. Lee (Hanmi Pharm. Co., Korea). All
chemicals were of reagent grade and used without fur-
ther purification.
In conclusion, we synthesized several analogs of RosA
and evaluated their binding affinity for the p56lck SH2
domain and inhibition activity for T-cell. Four hydroxyl
groups in catechol moieties of the antagonists were essen-
tial for the binding with the p56lck SH2 domain. Replace-
ment of the ester bond of RosA with an amide bond
resulted in retention of binding affinity but a decrease of
in vitro cell inhibition activity. Interestingly, a conforma-
tionally less constrained analog showed more improved
binding affinity for the SH2 domain, while R and S form
analogs showed similar binding affinity for the protein
and in vitro cell inhibition activity. The chemical antago-
nists with a potent cell activity required an optimal bal-
ance between the protein binding affinity and cell
penetrating ability. Overall results revealed a number of
important features regarding the design of novel p56lck
SH2 antagonists with more improved in vitro binding
affinity as well in vitro cell activity.
3.1. General chemistry
1H and 13C NMR spectra were recorded on a Bruker
AC 400 MHz spectrometer with DMSO-d6 as solvent,
d values in ppm (TMS as internal standard). Coupling
constants (J) were recorded in hertz. Electron spray ion-
ization (ESI)-mass spectra were recorded on VG plat-
form II mass spectrometer. Flash chromatography was
performed on silica gel 60H (Merck) using the different
solvent systems as indicated below. Analytical HPLC
was performed on Gilson HPLC system equipped with
UV detector and reversed phase C18 column, Delta
PAK at a flow rate of 1 ml/min. Retention times were
3. Materials
PyBOP and DCC for the solution phase amide coupling
were purchased from Calbiochem-Novabiochem Corp.
(San Diego, CA, USA). Piperidine, acetic anhydride,

3946
S.-H. Park et al. / Bioorg. Med. Chem. 15 (2007) 3938–3950
obtained by using 5–45% linear gradient of acetonitrile
during 40 min.
1H NMR (DMSO-d6, 400 MHz): 2.21–2.37 (m, 2H),
3.93–3.95 (m, 1H), 5.85–5.94 (m, 2H), 6.07–6.10(m,
2H), 6.21–6.23(m, 1H), 6.30–6.32 (m, 1H), 6.42 (s,
1H), 6.70 (d, J = 16Hz, 1H), 7.85 (d, J = 8 Hz, 1H),
8.28 (br, 2H), 8.72 (br, 1H), 8.92 (br, 1H).
3.1.1. 3-(3.4-Dihydroxy phenyl)-2-[3-(3,4-dihydroxy
phenyl)-acryloylamino]-propionic acid methyl ester (10).
DOPA (D form 2.0 g, 10.14 mmol, 1 equiv) was sus-
pended in methanol (40 ml) and to the solution was
added thionyl chloride (7.4 ml, 101.4 mmol, 10 equiv)
at 0 ꢁC ice bath. The reaction mixture was stirred for
18 h under N2 gas and then excess methanol and thionyl
chloride were evaporated. The crude product was puri-
fied by crystallization (methanol/ethyl acetate) and the
yield was 93%. DOPA methyl ester (2.0 g, 8.07 mmol,
1 equiv) was dissolved in DMF, (10 ml) and to the reac-
tion mixture was added caffeic acid (1.45 g, 8.07 mmol,
1 equiv) and it was diluted with dichloromethane
(DCM, 20 ml). To the reaction mixture were added
PyBOP (4.2 g, 8.07 mmol, 1 equiv) and tri-ethylamine
(TEA, 3.4 ml, 24.21 mmol, 3 equiv) at 0 ꢁC ice bath.
The reaction mixture was stirred for 18 h under N2 gas
and excess dichloromethane was evaporated in vacuo.
Then the reaction mixture was diluted with ethyl acetate
(10 ml) and sequentially washed with 1 N aqueous HCl,
10% aqueous NaHCO3, distilled water, and brine. The
organic layer was dried over MgSO4 and the crude prod-
uct was purified by silica gel column chromatography
(normal phase, eluting sequentially with 5:4:1, n-hex-
ane/ethyl acetate/methanol) to afford the 85% yield of
the title compound.
13C NMR (DMSO-d6, 400 MHz): 25.9, 51.8, 59.7, 112.9,
113.9, 116.4, 119.9, 122.9, 126.2, 127.8, 133.9, 140.0,
143.9, 144.9, 145.5, 147.4, 148.8, 165.3, 172.3.
ESI MS: (M+H)+ calcd m/e 360.10, obsd m/e 359.70.
3.1.3. 3-(3,4-Dihydroxy-phenyl)-2-[3-(3,4-dihydroxy-phe-
nyl)-propionylamino]-propionic acid (4). This was synthe-
sized by the catalytic hydrogenation of compound 7
over Pd/C catalyst and the yield was 95%.
1H NMR (DMSO-d6, 400 MHz): 2.21–2.25 (m, 2H),
2.45–2.47 (m, 2H), 2.64–2.75 (m, 2H), 4.27–4.30
(m, 1H), 6.33–6.37 (m, 2H), 6.50–6.57 (m, 4H), 8.23
(d, J = 8 Hz, 1H), 8.64 (br, 1H), 8.73 (br, 1H), 8.80
(br, 1H).
13C NMR (DMSO-d6, 400 MHz): 30.5, 36.4, 37.2, 53.9,
115.3, 115.6, 116.3, 118.6, 119.8, 127.8, 132.0, 143.3,
143.9, 145.0, 171.6, 172.3.
ESI MS: (M+H)+ calcd m/e 362.12, obsd m/e 361.75.
3.1.4. 3-(3,4-Dihydroxy-phenyl)-2-[3-(3,4-dihydroxy-phe-
nyl)-propionyloxy]-propionic acid (9). This was synthe-
sized by the catalytic hydrogenation of D form of
RosA over the Pd/C catalyst and the yield was 94%.
1H NMR (DMSO-d6, 400 MHz): 2.21–2.37 (m, 2H),
3.08 (s, 3H), 3.93–3.95 (m, 1H), 5.85–5.94 (m, 2H),
6.07–6.10 (m, 2H), 6.21–6.23 (m, 1H), 6.30–6.32
(m, 1H), 6.42 (s, 1H), 6.70 (d, J = 16 Hz, 1H), 7.85
(d, J = 8 Hz, 1H), 8.28 (br, 2H), 8.72 (br, 1H), 8.92
(br, 1H).
1H NMR (DMSO-d6, 400 MHz): 2.39–2.41 (m, 2H),
2.68–2.72 (m, 2H), 2.86–2.94 (m, 2H), 4.76–4.78
(m, 1H), 6.28–6.30 (m, 1H), 6.38–6.40 (m,1H), 6.51–
6.57 (m, 4H), 8.23 (d, J = 8 Hz, 1H), 8.64 (br, 1H),
8.73 (br, 1H), 8.80 (br, 1H).
13C NMR (DMSO-d6, 400 MHz): 25.9, 51.8, 59.7, 112.9,
113.9, 116.4, 119.9, 122.9, 126.2, 127.8, 133.9, 140.0,
143.9, 144.9, 145.5, 147.4, 148.8, 165.3, 172.3.
13C NMR (DMSO-d6, 400 MHz): 29.6, 35.7, 74.4, 115.3,
115.5, 115.6, 116.6, 118.6, 119.8, 128.4, 131.2, 143.4,
143.7, 144.8, 145.0, 171.2, 171.7.
ESI MS: (M+H)+ calcd m/e 374.12, obsd m/e 373.90.
3.1.2. 3-(3,4-Dihydroxy-phenyl)-2-[3-(3,4-dihydroxy-phe-
nyl)-acryloylamino]-propionic acid (7,8). Compound 10
(1.0 g, 2.67 mmol, 1 equiv) was dissolved in acetone
(30 ml) and to the reaction mixture was added drop-
wise distilled water (20 ml) and concentrated HCl
solution (10 ml) in 0 ꢁC ice bath so as to reduce the
precipitation of the dissolved compound and then it
was refluxed under drying tube (CaCl2) for 20 h at
80 ꢁC. After the excess acetone was evaporated, the
reaction mixture was washed with n-hexane (20 ml,
·3) and extracted with ethyl acetate (20 ml, ·3). The
collected organic layer was washed sequentially with
distilled water, brine and dried over anhydrous
MgSO4 and evaporated in vacuo. The crude product
was purified by flash silica gel column chromatogra-
phy (normal phase, elution with 4:5:1 n-hexane/ethyl
acetate/methanol) to afford the 83% yield of the title
compound 7. Compound 8 was synthesized from the
enantiomer of 10 by the same procedure and the yield
was 81%.
ESI MS: (M+H)+ calcd m/e 363.10, obsd m/e 361.68.
3.1.5. 3-(3,4-Dihydroxy-phenyl)-2-[3-(3,4-dihydroxy-phe-
nyl)-propionylamino]-propionic acid methyl ester (11).
This was also synthesized by the catalytic hydrogenation
of compound 10 over palladium/carbon catalyst and the
yield was quantitative.
1H NMR (DMSO-d6, 400 MHz): 2.29–2.36 (m, 2H),
2.71–2.77 (m, 2H), 3.01–3.02 (m, 2H), 3.57 (s, 3H),
4.34–4.32(m, 1H), 6.40–6.41 (m, 2H), 6.56–6.60
(m, 4H), 8.25(d, J = 8 Hz, 1H), 8.64 (br s, 1H), 8.73
(br s, 1H), 8.8 (br s, 1H), 8.84 (br s,1H).
13C NMR (DMSO-d6, 400 MHz): 30.5, 36.4, 37.1, 51.7,
53.9, 115.4, 115.6, 116.3, 118.6, 119.8, 127.8, 132.03,
143.3, 143.9, 145.03, 171.7, 172.3.
ESI MS: (M+H)+ calcd m/e 376.13, obsd m/e 375.29.

S.-H. Park et al. / Bioorg. Med. Chem. 15 (2007) 3938–3950
3947
3.1.6. 3-(3,4-Dihydroxy-phenyl)-2-[2-(3,4-dihydroxy-phe-
nyl)-acetylamino]-propionic acid methyl ester (12). (3,4-
Dihydroxy-phenyl)-acetic acid (0.2 g, 1.18 mmol,
1 equiv) was dissolved in DMF (10 ml) and to the reac-
tion mixture was added DOPA methyl ester (0.2 g,
1.01 mmol, 1 equiv) and PyBop (4.2 g, 8.07 mmol,
1 equiv) and tri-ethylamine (TEA, 3.4 ml, 24.21 mmol,
3 equiv) at 0 ꢁC ice bath. The reaction mixture was stir-
red for 18 h under N2 gas and excess dichloromethane
was evaporated. And then the reaction mixture was fur-
ther proceeded as in compound 10. The crude product
was purified by preparative HPLC using the solvent sys-
tem (0.1% TFA H2O/acetonitrile) to afford the title com-
pound and the yield was 45%.
chloride (0.2 ml, 2.77 mmol, 10 equiv) at 0 ꢁC ice bath.
Each reaction mixture was equally stirred for 18 h under
N2 gas and further proceeded as previously discussed in
compound 10. The crude product was purified by pre-
parative HPLC by using same solvent system as in com-
pound 12. The yields for 14, 15, 16, and 17 were 93%,
84%, 52%, and 47%, respectively.
3.1.8.1. Compound 14. 1H NMR (DMSO-d6/TMS,
400 MHz): 2.96 (br, 2H), 3.64 (s, 3H), 5.11–5.12
(m, 1H), 6.28 (d, J = 16 Hz, 1H), 6.51 (d, J = 8 Hz,
1H), 6.63–6.65 (m, 2H), 6.78 (d, J = 8 Hz, 1H), 7.02–
7.07 (m, 2H), 7.50 (d, J = 16 Hz, 1H), 8.76 (s, 1H), 8.8
(s, 1H), 9.17 (s, 1H), 9.68 (s, 1H).
1H NMR (DMSO-d6/TMS, 400 MHz): 2.65–2.84 (m,
2H), 3.22 (s, 2H), 3.56 (s, 3H), 4.25–4.34 (m, 1H),
6.38–6.42 (m, 2H), 6.58–6.60 (m, 4H), 8.75 (br, 4H).
13C NMR (DMSO-d6/TMS, 400 MHz): 30.7, 52.0, 72.8,
112.8, 114.9, 115.4, 115.7, 116.7, 121.7, 125.3, 126.6,
134.9, 144.1, 145.0, 145.6, 146.4, 148.7, 165.9, 169.9.
13C NMR (DMSO-d6/TMS, 400 MHz): 36.4, 41.2, 51.7,
54.0, 115.3, 116.4, 116.7, 119.7, 126.7, 126.8, 127.7,
128.1, 143.8, 143.9, 144.8, 150.0, 170.7, 173.2.
ESI MS: (M + H)+ calcd m/e 375.10, obsd m/e 374.79.
3.1.8.2. Compound 15. 1H NMR (DMSO-d6,
400 MHz): 1.66(t, J = 6 Hz, 3H), 3.40 (br, 2H),
4.50–4.52 (m, 2H), 5.56 (m, 1H), 6.72 (d, J = 15.6 Hz),
6.93–6.94 (m, 1H), 7.07–7.09 (m, 2H), 7.20–7.21
(m, 1H), 7.44–7.50 (m, 2H), 7.94 (d, J = 15.6 Hz, 1H),
9.19 (br, 2H), 9.58 (br, 1H), 10.08 (br, 1H).
ESI MS: (M+H)+ calcd m/e 362.12, obsd m/e 361.75.
3.1.7. 2-(3,4-Dihydroxy-benzoyl amino)-3-(3,4-dihy-
droxy-phenyl)-propionic acid methyl ester (13). (3,4-
Dihydroxy)-benzoic acid (0.2 g, 0.78 mmol, 1.2 equiv)
was dissolved in DMF (10 ml) and to the reaction mix-
ture were added DCC (0.4 g, 1.94 mmol, 3 equiv) and
HOBt (0.26 g, 1.94 mmol, 3 equiv). After activating
3,4-dihydroxy benzoic acid for 30 min, to the reaction
mixture were added DOPA methyl ester (0.2 g,
0.78 mmol, 1.2 equiv) dissolved in DMF (5 ml) and
DMAP (0.24 g, 1.94 mmol, 3 equiv) and then it was stir-
red for 18 h. The crude product was further proceeded
as mentioned earlier in compound 10 and purified by
preparative HPLC by using the same solvent system as
in compound 12 and the yield was 48%.
13C NMR (DMSO-d6/TMS, 400 MHz): 13.8, 36.1, 60.6,
72.7, 112.8, 114.8, 115.3, 115.6, 116.6, 120.0, 121.5,
125.2, 126.5, 144.0, 144.8, 145.5, 146.1, 148.6, 165.8,
169.3.
ESI MS: (M+H)+ calcd m/e 389.12, obsd m/e 388.77.
3.1.8.3. Compound 16. 1H NMR (DMSO-d6,
400 MHz): 0.83(t, J = 7.2 Hz, J = 7.6 Hz, 3H),
1.51–1.56 (m, 2H), 2.97 (d, J = 6.8 Hz, 2H), 4.0
(t, d = 6.4 Hz, 6.4 Hz, 2H), 5.09 (t, J = 7.2 Hz, 6.0 Hz,
1H), 6.29 (d, J = 16 Hz, 1H), 6.50–6.52 (m, 1H), 6.78
(d, J = 8 Hz, 1H), 7.01–7.07 (m, 2H), 7.51 (d, J =
16 Hz, 1H).
1H NMR (DMSO-d6, 400 MHz): 3.30–3.33 (m, 2H), 4.02
(s, 3H), 4.87–4.90 (m, 1H), 6.91–6.93 (m, 1H), 7.01–7.03
(m, 1H), 7.05–7.06 (m, 1H), 7.15–7.19 (m, 1H),
7.59–7.61 (m, 1H), 7.66–7.67 (m, 1H), 8.81 (d, J = 7.4 Hz).
13C NMR (DMSO-d6/TMS, 400 MHz): 10.1, 21.4,
36.2,66.1, 72.9, 112.9, 114.9, 115.4, 115.7, 116.7, 120.1,
121.7, 125.3, 126.6, 144.1, 145.0, 145.6, 146.3, 148.7,
165.9, 169.5.
13C NMR (DMSO-d6, 400 MHz): 36.4, 52.4, 55.3, 115.4,
115.8, 115.9, 116.9, 120.0, 120.4, 125.5, 129.1, 144.4,
145.4, 145.5, 149.2, 172.7, 173.3.
ESI MS: (M+H)+ calcd m/e 403.13, obsd m/e 402.56.
ESI MS: (M+H)+ calcd m/e 348.10 obsd m/e 347.32.
3.1.8.4. Compound 17. 1H NMR (DMSO-d6/TMS,
400 MHz): 1.08–1.17 (m, 6H), 2.95 (d, J = 6.4Hz, 2H),
4.84–4.90 (m, 1H), 5.02 (t, J = 6.4 Hz, 1H), 6.30
(d, J = 16Hz, 1H), 6.50–6.53 (m, 1H), 6.63–6.66
(m, 2H), 6.76–6.78 (m, 1H), 7.02–7.07 (m, 2H), 7.51
(d, J = 16 Hz, 1H).
3.1.8. 3-(3,4-Dihydroxy-phenyl)-acrylic acid 2-(3,4-dihy-
droxy-phenyl)-1-methoxy carbonyl-ethyl ester, 3-(3,4
dihydroxy-phenyl)-acrylic acid 2-(3,4-dihydroxy-phenyl)-
1-ethoxy carbonyl-ethyl ester, 3-(3,4 dihydroxy-phenyl)-
acrylic acid 2-(3,4-dihydroxy-phenyl)-1-propoxy carbonyl-
ethyl ester, and 3-(3,4 dihydroxy-phenyl)-acrylicacid 2-
(3,4-dihydroxy-phenyl)-1-isopropoxycarbonyl-ethyl ester,
(14, 15, 16, and 17). Rosmarinic acid (0.1 g, 0.277 mmol,
1 equiv) was, respectively, suspended in methanol for
compound 14, ethanol for compound 15, 1-propanol
for compound 16, and 2-propanol for compound
17, and to the solution was added dropwisely thionyl
13C NMR (DMSO-d6/TMS, 400 MHz): 21.4, 21.5, 25.5,
36.2, 62.0, 68.4, 73.0, 113.5, 115.0, 115.4, 115.8, 116.8,
120.2, 121.7, 125.3, 126.5, 144.1, 145.0, 145.6, 146.3,
148.7, 165.9, 168.9.
ESI MS: (M+H)+ calcd m/e 403.13, obsd m/e 402.56.

3948
S.-H. Park et al. / Bioorg. Med. Chem. 15 (2007) 3938–3950
3.1.9. 3-(3,4 Dihydroxy-phenyl)-acrylicacid 1-tert-butoxy-
carbonyl-2-(3,4-dihydroxyphenyl) ethyl ester (6). RosA
was dissolved in dioxane containing 1% sulfuric acid
and then the resulting solution was cooled to À10 ꢁC. Li-
quid isobutylene was added into the solution and stirred
at 0 ꢁC for 48 h. The excess solvent was removed by re-
duced pressure and the crude product was further purified
by preparative HPLC and the yield was 38%.
acid for compound 19, 4-methoxy caffeic acid for com-
pound 20, and 3,4 dimethoxy caffeic acid for compound
21, respectively, and added 1.2 equiv of PyBop and
3 equiv of triethyl amine. Then each reaction mixture
was equally stirred for 18 h under nitrogen environment
and further proceeded as in compound 10, and the crude
product was purified by HPLC by using the same sol-
vent system as mentioned in compound 12. The yields
for 19, 20, and 21 were 65%, 67%, 72%, respectively.
1H NMR (DMSO-d6/TMS, 400 MHz): 1.33 (s, 9H), 2.93
(d, J = 6.4 Hz, 2H), 4.95 (t, J = 6.0 Hz, J = 6.8Hz, 1H),
6.29 (d, J = 16 Hz, 1H), 6.51–6.53 (m, 1H), 6.63–6.67
(m, 2H), 6.78 (d, J = 8 Hz, 1H), 7.01–7.07 (m, 2H),
7.51 (d, J = 16Hz, 1H), 8.79 (br, 2H), 9.18 (br, 1H),
9.68 (br, 1H).
1
3.1.11.1. Compound 19. H NMR (DMSO-d6/TMS,
400 MHz): 2.73–2.90 (m, 2H), 3.61 (s, 1H), 3.80
(s, 1H), 4.45–4.50 (m, 1H), 6.44–6.47 (m, 1H), 6.97–
6.99 (m, 1H), 7.12 (d, J = 2Hz, 1H), 7.31 (d,
J = 15.6 Hz, 1H), 8.31 (d, J = 7.6 Hz, 1H).
13C NMR (DMSO-d6/TMS, 400 MHz): 27.5, 36.1, 73.1,
81.2, 113.0, 114.9, 115.3, 115.7, 116.8, 120.2, 121.5,
125.3, 126.7, 144.0, 144.9, 145.5, 146.0, 148.6, 165.8,
168.4.
13C NMR (DMSO-d6/TMS, 400 MHz): 36.8, 51.8, 54.1,
55.5, 110.6, 115.4, 115.6, 116.3, 118.2, 119.8, 121.8,
126.2, 127.8, 139.8, 143.9, 145.0, 147.8, 148.4, 165.4,
172.4.
ESI MS: (M+H)+ calcd m/e 417.15, obsd m/e 417.60.
ESI MS: (M+H)+ calcd m/e 388.13, obsd m/e 387.50.
1
3.1.11.2. Compound 20. H NMR (DMSO-d6/TMS,
3.1.10. 3-(3,4 Dihydroxy-phenyl)-acrylic acid 2-(3,4-dihy-
droxy-phenyl)-1-(5-methoxy carbonyl-pentylcarbamoyl)-
ethyl ester (18). 6-Aminocaproic acid methyl ester (1 g,
6.8 mmol, 1.2 equiv) was dissolved in DMF (5 ml) and
to the reaction mixture was added RosA (2.45 g, 6.8
mmol, 1 equiv) and diluted in dichloromethane
(20 ml). To the reaction mixture were added PyBop
(1.2 equiv) and tri-ethylamine (3 equiv) at 0 ꢁC ice bath.
The reaction mixture was stirred for 18 h under N2 gas.
The reaction was further proceeded as mentioned earlier
in compound 10 and the crude product was purified by
preparative HPLC using the same solvent system as pre-
viously mentioned in compound 12 and the yield was
52%.
400 MHz): 2.73–2.89 (m, 2H), 3.61 (s, 3H), 3.79
(s, 3H), 4.46–4.48 (m, 1H), 6.44–6.49 (m, 2H), 6.59–
6.63 (m, 2H), 6.92–6.97 (m, 3H), 7.26 (d, J = 15.6 Hz,
1H), 8.4 (d, J = 7.2 Hz, 1H).
13C NMR (DMSO-d6/TMS, 400 MHz): 36.4, 51,8, 55.6,
112.1, 113.4, 115.4, 116.3, 118.7, 119.8, 120.4, 127.6,
127.8, 139.7, 144.9, 146.7, 149.3, 165.2, 172.4.
ESI MS: (M+H)+ calcd m/e 388.13, obsd m/e 387.45.
1
3.1.11.3. Compound 21. H NMR (DMSO-d6/TMS,
400 MHz): 2.74–2.90 (m, 2H), 3.61 (s, 3H), 3.78
(s, 3H), 3.79 (s, 3H), 4.51 (q, J = 2.8 Hz, J = 5.4Hz,
J = 2.8Hz, 1H), 6.45–6.47 (m, 1H), 6.56–6.63 (m, 3H),
6.99 (d, J = 8.4 Hz, 1H), 7.09–7.15 (m, 2H), 7.35
(d, J = 15.6 Hz, 1H), 8.34 (d, J = 7.6Hz, 1H), 8.73
(br, 1H), 8.78 (br, 1H).
1H NMR (DMSO-d6/TMS, 400 MHz): 1.19–1.23
(m, 2H), 1.34–1.37 (m, 2H), 1.45–1.51 (m, 2H), 2.18–
2.29 (m, 2H), 2.78–2.89 (m, 2H), 3.02–3.03 (m, 2H),
3.57 (s, 3H), 5.01–5.04 (m, 1H), 6.25 (d, J = 16Hz,
1H), 6.48 (d, J = 8Hz, 1H), 6.59–6.63 (m, 2H), 6.77
(d, J = 8 Hz, 1H), 7.00 (d, J = 8 Hz, 1H), 7.03 (s, 1H),
7.46 (d, J = 16 Hz, 1H), 7.99 (br, 1H), 8.69 (s, 1H),
8.75 (s, 1H), 9.16 (br, 1H).
13C NMR (DMSO-d6/TMS, 400 MHz): 31.4, 52.5, 56.0,
56.2, 110.5, 112.3, 116.0, 116.9, 119.8, 120.4, 122.2,
128.1, 128.4, 140.1, 144.6, 145.6, 149.5, 150.8, 165.9,
172.9.
13C NMR (DMSO-d6/TMS, 400 MHz): 24.1, 25.7, 28.6,
33.2, 37.1, 38.2, 51.2, 113.6, 114.7, 115.3, 115.7, 116.7,
120.0, 121.5, 125.4, 127.5, 143.8, 144.9, 145.6, 148.5,
165.8, 168.8, 173.4.
ESI MS: (M+H)+ calcd m/e 402.15, obsd m/e 401.80.
3.1.12.
2-[3-(3,4-dihydroxy-phenyl)-acryloylamino]-3-
(3,4-dimethoxy-phenyl)-propionic acid methyl ester (22).
3,4 Dimethoxy phenylalanine (0.2 g, 0.887 mmol,
1 equiv) was suspended in methanol (5 ml) and add 4
equiv of thionyl chloride at 0 ꢁC ice bath. The reaction
mixture was stirred under nitrogen atmosphere for
18 h. The excess methanol was evaporated in vacuo
and further proceeded as in compound 10.
ESI MS: (M+H)+ calcd m/e 488.18, obsd m/e 487.30.
3.1.11.
3-(3,4-Dihydroxy-phenyl)-2-[3-(4-hydroxy-3-
methoxy-phenyl)-acryloylamino]-propionic acid methyl
ester, 3-(3,4 dihydroxy-phenyl)-2-[3-(3-hydroxy-4-meth-
oxy-phenyl)-acryloylamino]-propionic acid methyl ester,
and 3-(3,4 dihydroxy-phenyl)-2-[3-(3,4-dimethoxy-phe-
nyl)-acryloylamino]-propionic acid methyl ester (19, 20,
and 21). D-DOPA methyl ester (0.2 g, 0.94 mmol,
1 equiv) was dissolved in DMF/DCM and to the mix-
ture was added the same equivalent of 3-methoxycaffeic
Then, add 1.1 equiv of caffeic acid, 1.2 equivalent of Py-
Bop, and 4 equiv of tri-ethylamine were added to the
above product and stirred under nitrogen atmosphere
for 18 h. The reaction was further proceeded as in

S.-H. Park et al. / Bioorg. Med. Chem. 15 (2007) 3938–3950
3949
compound 10 and purified by flash column chromatog-
raphy by using 5:4:1 (Hx/EtOAc/MeOH) to afford the
yield of 83% of the title compound.
absence of rosmarinic acid or its derivative (10 lg/ml)
for 2 h before stimulation. Cells were activated either
with anti-CD3 mAb (5 lg/ml, UCHT1, IgG2a isotype)
coated on the plate or with PMA (5 lg/ml) and ionomy-
cin (0.5 lg/ml) for 16 h. In the rosmarinic acid or its
derivative treated group, they were present through the
whole 16 h incubation process. After stimulation, the
cells were washed, lysed, and assayed for luciferase
activity according to the manufacturer’s instructions
(Luciferase Assay System kit, Promega, Madison, WI)
with Microplate luminometer LB96V (Perkin-Elmer,
Foster City, CA).
1H NMR (DMSO-d6/TMS, 400 MHz): 2.96 (d, 2H),
3.57 (s, 3H) 3.65 (s, 6H), 4.48–4.49 (m, 1H), 6.35 (d,
J = 15.6 Hz, 1H), 6.45–6.47 (m, 1H), 6.67–6.69 (d, J =
8 Hz, 2H), 6.78 (m, 3H), 6.88 (s, 1H), 7.17 (d,
J = 15.6 Hz, 1H), 8.35 (d, J = 7.6 Hz, 1H), 9.2 (br,
1H), 9.4 (br, 1H).
13CNMR(DMSO-d6/TMS,400 MHz): 36.4, 51.8, 54.9,
55.4, 111.6, 112.9, 113.8, 115.7, 117.5, 120.5, 121.0,
126.1, 129.5, 140.0, 145.6, 147.5, 147.6, 148.4, 165.4,
172.3.
3.4. Estimation of logP
The relative hydrophobicity was assessed for each of the
rosmarinic acid and its derivatives using a typical
method involving measuring the partitioning of the
compound between 1-octanol and distilled water.
HPLC-grade 1-octanol was pre-saturated with distilled
water, and aqueous phase distilled water was saturated
with 1-octanol before use. The derivatives were each dis-
solved in 1-octanol/water phase at final concentration of
1 mM, and an equal volume of 1-octanol/water was
added. The tubes were then vortexed continuously for
30 min. The final concentration of compound in both
1-octanol and aqueous phases was measured by HPLC
with C18 column. The partition coefficient, P, was deter-
mined by dividing the concentration of the derivative in
1-octanol by the concentration in the aqueous phase.
ESI MS: (M+H)+ calcd m/e 402.15, obsd m/e 401.33.
3.2. ELISA binding assay
The assays were performed as we described previ-
ously.19,23 Briefly, an enzyme-linked immunosorbant as-
say was employed to measure the binding affinity of
synthesized derivatives for each SH2 domains. 100 ll
of biotinyl-e-aminocaproyl-EPQpYEEIPIYL (10 ng/ml
in 50 mM Tris, 150 mM NaCl, pH 7.5) was added to
each well of streptavidin-coated 96-well micro titer
plates. The plates were shaken overnight at 4 ꢁC and
rinsed with TBS (50 mM Tris, 150 mM NaCl, pH 7.5,
3· 200 ll). Each well was then blocked with 100 ll of
a solution containing 2% BSA and 0.2% Tween 20 in
TBS (1 h at 37 ꢁC). The wells were then rinsed with 4·
200 ll of a standard BSA-T-TBS solution (0.2% BSA,
0.1% Tween 20, TBS). A 50-ll solution of samples
(1 mM, in BSA-T-TBS) and a 50-ll solution of the
SH2-GST fusion proteins (100 ng/ml, in BSA-T-TBS)
were added in each well and the plate was shaken for
1 h at room temperature. The solutions were removed
and each well rinsed with 4· 200 ll BSA-T-TBS.
100 ll of polyclonal rabbit anti-GST antibody (1 lg/ml
in BSA-T-TBS) was then added to each well and incu-
bated for 1 h at room temperature. Following subse-
quent washing steps with BSA-T-TBS (4· 200 ll),
100 ll of horseradish peroxidase-conjugated mouse
anti-rabbit antibody (2 lg/ml in BSA-T-TBS) was added
to each well and subsequently incubated for 1 h at room
temperature. After a series of final wash steps (4· 200 ll
BSA-T-TBS; 2· 300 ll and TBS), 100 ll of peroxidase
substrate (1-Step Turbo TMB-ELISA, trimethylbenzi-
dine) was added to each well and incubated for
5–15 min. 100 ll of 1 M sulfuric acid solution was intro-
duced to stop the peroxidase reaction and absorbance
was measured at 450 nm with a plate reader.
Acknowledgment
This work was supported by the grant (R01-2006-000-
10956-0) from the Basic Research Program of the Korea
Science & Engineering Foundation. B.P. Joshi was
recipients of BK21(II) fellowship.
References and notes
1. Pawson, T.; Gish, G. D.; Nash, P. Trends Cell Biol. 2001,
11, 504.
2. Eck, M. J.; Atwell, S. K.; Shoelson, S. E.; Arrison, S. S.
Nature 1994, 368, 764.
3. (a) Songyang, Z.; Shoelson, S. E.; Chaudhuri, M.;
Gish, G.; Pawson, T.; Haser, W. G.; King, F.; Roberts,
T.; Ratnofsky, S.; Lechleider, R. J.; Neel, B. G.; Birge,
R. B.; Fajardo, J. E.; Chou, M. M.; Hanafusa, H.;
Schaffhausen, B.; Cantley, L. C. Cell 1993, 72, 767; (b)
Eck, M. J.; Shoelson, S. E.; Harrison, S. C. Nature
1993, 362, 87; (c) Songyang, Z.; Cantley, L. C. Trends
Biochem. Sci. 1995, 20, 470.
4. (a) Broadbridge, R. J.; Sharma, R. P. Curr. Drug Targets
2000, 1, 365; (b) Proudfoot, J. R.; Betageri, R.;
Cardozo, M.; Gilmore, T. A.; Glynn, S.; Hickey, E. R.;
Jakes, S.; Kabcenell, A.; Kirrane, T. M.; Tibolla, A. K.;
Lukas, S.; Patel, U. R.; Sharma, R.; Yazdanian, M.;
Moss, N.; Beaulieu, P. L.; Cameron, D. R.; Ferland,
J.-M.; Gauthier, J.; Gillard, J.; Gorys, V.; Poirier, M.;
Rancourt, J.; Wernic, D.; Llinas-Brunet, M. J. Med.
Chem. 2001, 44, 2421; (c) Shi, Z. D.; Lee, K.; Wei, C. Q.;
Roberts, L. R.; Worthy, K. M.; Fisher, R. J.; Burke, T. R.,
Jr. J. Med. Chem. 2004, 47, 788.
3.3. Luciferase assay
The assays were performed as we described previously.19
Briefly, a total of 2 · 106 Jurkat T cells were transfected
with 2 lg of IL-2-Luciferase (gifts of Gerald Crabtree,
Stanford University, Stanford, CA) using Lipofectamine
according to manufacturer’s protocol (GibcoBRL).
After incubation with DNA-Lipofectamine mixtures
for 24 h, cells were pre-incubated in the presence or

3950
S.-H. Park et al. / Bioorg. Med. Chem. 15 (2007) 3938–3950
5. Sawyer, T. K. Biopolymers 1998, 47, 243.
6. Cody, W. L.; Lin, Z. W.; Panek, R. L.; Rose, D. W.;
17. Burke, T. R., Jr.; Yao, Z. J.; Liu, D. G.; Voigt, J.; Gao, Y.
Biopolymers 2001, 60, 32.
Rubin, J. R. Curr. Pharm. Des. 2000, 6, 59.
7. Garcia-Echeverria, C. Curr. Med. Chem. 2001, 8, 1589.
8. Lewis, L. A.; Chung, C. D.; Chen, J.; Parnes, J. R.;
Moran, M.; Patel, V. P.; Miceli, M. C. J. Immunol. 1997,
159, 2292.
9. Straus, D. B.; Chan, A. C.; Patai, B.; Weiss, A. J. Biol.
Chem. 1996, 271, 9976.
10. Tong, L.; Warren, T. C.; King, J.; Betageri, R.; Rose, J.
J. Mol. Biol. 1996, 256, 601.
18. Nam, J. Y.; Kim, H. K.; Kwon, J. Y.; Han, M. Y.; Son, K.
H.; Lee, U. C.; Choi, J. D.; Kwon, B. M. J. Nat. Prod.
2000, 63, 1303.
19. Won, J.; Hur, Y. G.; Hur, E. M.; Park, S. H.; Kang, M.
A.; Choi, Y.; Park, C.; Lee, K. H.; Yun, Y. Eur. J.
Immunol. 2003, 33, 870.
20. Petersen, M.; Simmonds, M. S. Phytochemistry 2003, 62,
121.
21. Kang, M. A.; Yun, S. Y.; Won, J. Blood 2003, 101, 3534.
22. Yun, S. Y.; Hur, Y. G.; Kang, M. A.; Lee, J.; Ahn, C.;
Won, J. Transplantation 2003, 75, 1758.
11. Waksman, G.; Shoelson, S. E.; Pant, N.; Cowburn, D.;
Kuriyan, J. Cell 1993, 72, 779.
12. Lunney, E. A.; Para, K. S.; Rubin, J. R.; Humblet, C.;
Fergus, J. H.; Marks, J. S.; Sawyer, T. K. J. Am. Chem.
Soc. 1997, 119, 12471.
13. Nguyen, H. N.; Guofeng, Y.; Sun, G.; Parang, K. J. Med.
Chem. 2004, 47, 3131.
23. Park, S. H.; Won, J.; Lee, K. H. Bioorg. Med. Chem. Lett.
2002, 12, 2711.
24. Lee, T. R.; Lawrence, D. S. J. Med. Chem. 2000, 43, 1173.
25. Beaulieu, P. L.; Cameron, D. R.; Ferland, J.-M.; Gauthi-
er, J.; Ghiro, E.; Gillard, J.; Gorys, V.; Poirier, M.;
Rancourt, J.; Wernic, D.; Llinas-Brunet, M.; Betageri, R.;
Cardozo, M.; Hickey, E. R.; Ingraham, R.; Jakes, S.;
Kabcenell, A.; Kirrane, T.; Lukas, S.; Patel, U.; Proud-
foot, J.; Sharma, R.; Tong, L.; Moss, N. J. Med. Chem.
1999, 42, 1757.
26. Huang, N.; Nagarsekar, A.; Guanjun, X.; Jun, H.;
Mackerell, A. D., Jr. J. Med. Chem. 2004, 47, 3502.
27. Park, S. H.; Kang, S. H.; Lim, S. H.; Oh, H. S.; Lee, K. H.
Bioorg. Med. Chem. Lett. 2003, 13, 3455.
28. Valko, K.; My, Du. C.; Bevan, C.; Reynolds, D. P.;
Abraham, M. H. Curr. Med. Chem. 2001, 8, 1137.
29. Christopher, T. J.; Scott, E. K.; Jongming, L. Bioorg.
Med. Chem. Lett. 1997, 7, 113.
30. Xu, H.; Littman, D. R. Cell 1993, 74, 633.
31. Kiani, A.; Rao, A.; Aramburu, J. Immunity 2000, 12, 359.
14. Bohacek, R. S.; Dalgarno, D. C.; Hatada, M.; Jacobsen,
V. A.; Lynch, B. A.; Macek, K. J.; Merry, T.; Metcalf, C.
A., III; Narula, S. S.; Sawyer, T. K.; Shakespeare, W. C.;
Violette, S. M.; Weigele, M. J. Med. Chem. 2001, 44, 660.
15. Mustelin, T.; Brockdorff, J.; Rudbeck, L.; Gjorloff-Win-
gren, A.; Han, S.; Wang, X.; Tailor, P.; Saxena, M. Cell
Signal. 1999, 11, 637.
16. Shakespeare, W.; Yang, M.; Bohacek, R.; Cerasoli, F.;
Stebbins, K.; Sundaramoorthi, R.; Azimioara, M.; Vu, C.;
Pradeepan, S.; Metcalf, C., III; Haraldson, C.; Merry, T.;
Dalgarno, D.; Narula, S.; Hatada, M.; Lu, X.; van
Schravendijk, M. R.; Adams, S.; Violette, S.; Smith, J.;
Guan, W.; Bartlett, C.; Herson, J.; Iuliucci, J.; Weigele,
M.; Sawyer, T. Proc. Natl. Acad. Sci. U.S.A. 2000, 97,
9373.