Structural basis for the activity of the RSK-specific inhibitor, SL0101

Article abstract of DOI:10.1016/j.bmc.2007.03.087

Inappropriate activity of p90 ribosomal S6 kinase (RSK) has been implicated in various human cancers as well as other pathologies. We previously reported the isolation, characterization, and synthesis of the natural product kaempferol 3-O-(3″,4″-di-O-acetyl-α-l-rhamnopyranoside), termed SL0101 [Smith, J. A.; Poteet-Smith, C. E.; Xu, Y.; Errington, T. M.; Hecht, S. M.; Lannigan, D. A. Cancer Res., 2005, 65, 1027-1034: Xu, Y.-M; Smith, J. A.; Lannigan, D. A.; Hecht, S. M. Bioorg. Med. Chem., 2006, 14, 3974-3977: Maloney, D. J.; Hecht, S. M. Org. Lett., 2005, 7, 1097-1099]. SL0101 is a potent and specific inhibitor of RSK; therefore, we performed an analysis of the structural basis for the inhibitory activity of this lead compound. In in vitro kinase assays we found that acylation of the rhamnose moiety and the 4′, 5, and 7-hydroxyl groups are responsible for maintaining a high affinity interaction of RSK with SL0101. It is likely that the hydroxyl groups facilitate RSK binding through their ability to form hydrogen bonds. To determine whether the SL0101 derivatives were specific for inhibition of RSK we analyzed their ability to preferentially inhibit the growth of the human breast cancer line, MCF-7, compared to the normal human breast line, MCF-10A. We have previously validated this differential growth assay as a convenient readout for analyzing the specificity of RSK inhibitors [Smith, J. A.; Maloney, D. J.; Clark, D. E.; Xu, Y.-M.; Hecht, S. M.; Lannigan, D. A. Bioorg. Med. Chem., 2006, 14, 6034-6042]. We found that acylation of the rhamnose moiety was essential for maintaining the selectivity for RSK inhibition in intact cells. Further, the efficacy of SL0101 in intact cells is limited by cellular uptake as well as possible hydrolysis of the acetyl groups on the rhamnose moiety by ubiquitous intracellular esterases. These studies should facilitate the development of a RSK inhibitor, based on the SL0101 pharmacophore, as an anti-cancer chemotherapeutic agent.

Full text of DOI:10.1016/j.bmc.2007.03.087

Bioorganic & Medicinal Chemistry 15 (2007) 5018–5034
Structural basis for the activity of the RSK-specific inhibitor,
SL0101
Jeffrey A. Smith,a,b David J. Maloney,c Sidney M. Hechtc,d,* and Deborah A. Lannigana,e,
*
aCenter for Cell Signaling, University of Virginia, Charlottesville, VA 22908, USA
bDepartment of Pathology, University of Virginia, Charlottesville, VA 22908, USA
cDepartment of Chemistry, University of Virginia, Charlottesville, VA 22904, USA
dDepartment of Biology, University of Virginia, Charlottesville, VA 22904, USA
eDepartment of Microbiology, University of Virginia, Charlottesville, VA 22908, USA
Received 6 March 2007; revised 29 March 2007; accepted 29 March 2007
Available online 2 April 2007
Abstract—Inappropriate activity of p90 ribosomal S6 kinase (RSK) has been implicated in various human cancers as well as other
pathologies. We previously reported the isolation, characterization, and synthesis of the natural product kaempferol 3-O-(300,400-di-
O-acetyl-a-L-rhamnopyranoside), termed SL0101 [Smith, J. A.; Poteet-Smith, C. E.; Xu, Y.; Errington, T. M.; Hecht, S. M.; Lann-
igan, D. A. Cancer Res., 2005, 65, 1027–1034: Xu, Y.-M; Smith, J. A.; Lannigan, D. A.; Hecht, S. M. Bioorg. Med. Chem., 2006, 14,
3974–3977: Maloney, D. J.; Hecht, S. M. Org. Lett., 2005, 7, 1097–1099]. SL0101 is a potent and specific inhibitor of RSK; there-
fore, we performed an analysis of the structural basis for the inhibitory activity of this lead compound. In in vitro kinase assays we
found that acylation of the rhamnose moiety and the 40, 5, and 7-hydroxyl groups are responsible for maintaining a high affinity
interaction of RSK with SL0101. It is likely that the hydroxyl groups facilitate RSK binding through their ability to form hydrogen
bonds. To determine whether the SL0101 derivatives were specific for inhibition of RSK we analyzed their ability to preferentially
inhibit the growth of the human breast cancer line, MCF-7, compared to the normal human breast line, MCF-10A. We have
previously validated this differential growth assay as a convenient readout for analyzing the specificity of RSK inhibitors [Smith,
J. A.; Maloney, D. J.; Clark, D. E.; Xu, Y.-M.; Hecht, S. M.; Lannigan, D. A. Bioorg. Med. Chem., 2006, 14, 6034–6042]. We found
that acylation of the rhamnose moiety was essential for maintaining the selectivity for RSK inhibition in intact cells. Further, the
efficacy of SL0101 in intact cells is limited by cellular uptake as well as possible hydrolysis of the acetyl groups on the rhamnose
moiety by ubiquitous intracellular esterases. These studies should facilitate the development of a RSK inhibitor, based on the
SL0101 pharmacophore, as an anti-cancer chemotherapeutic agent.
ꢀ 2007 Elsevier Ltd. All rights reserved.
1. Introduction
SL0101 (1) inhibits the proliferation of the human breast
cancer cell line, MCF-7.2 Remarkably, however, SL0101
(1) does not alter the growth of the normal human
breast line MCF-10A, even though SL0101 (1) inhibits
RSK activity in MCF-7 and MCF-10A cells.2 Taken
together, these results suggest that RSK may be an
important new anti-cancer target.
The members of the p90 ribosomal S6 kinase family are
downstream effectors of p42/p44 mitogen activated pro-
tein kinase (MAPK).1 There is accumulating evidence
that RSK plays a role in a number of different can-
cers.2–6 These studies are partly based on our discovery
of the RSK-specific inhibitor, SL0101 (kaempferol
3-O-(300,400-di-O-acetyl-a-L-rhamnopyranoside))
which was initially isolated from an extract from Forste-
ronia refracta, a member of the Apocynaceae (dogbane)
family found in the South American rainforest.2,7
(1),
Protein kinases have been shown to be excellent drug
targets and a number of kinase inhibitors are being suc-
cessfully used as anti-cancer agents in the clinic.8 Protein
kinases form a large superfamily of proteins that contain
a conserved overall folding pattern for the kinase cata-
lytic domain. The dual-lobed structure contains a smal-
ler lobe for binding of MgÆATP and a larger lobe for
association with the substrate. Another conserved fea-
ture is an ‘activation loop’ near the catalytic center that
differs in size and sequence among various kinases.
Keywords: Kaempferol 3-O-(300,400-di-O-acetyl-a-L-rhamnopyrano-
side); SL0101; RSK; Kinase inhibitor; Structure–activity relationship.
*
Corresponding authors. Tel.: +1 434 924 3906; fax: +1 434 924 7856
(S.M.H.); tel.: +1 434 924 1144; fax: +1 434 924 1236 (D.A.L.);
e-mail addresses: sidhecht@virginia.edu; dal5f@virginia.edu
0968-0896/$ - see front matter ꢀ 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2007.03.087

J. A. Smith et al. / Bioorg. Med. Chem. 15 (2007) 5018–5034
5019
However, in spite of the high degree of similarity
between the active sites of protein kinases there are
now a number of examples of very specific inhibitors
that utilize the ATP binding pocket and compete with
ATP binding.9 RSK is an unusual kinase in that it con-
tains two non-identical functional kinase domains, an
N-terminal kinase domain (NTKD) and a C-terminal
kinase domain (CTKD).1 The NTKD is responsible
for phosphorylating exogenous substrates; the only
known function of the CTKD is autophosphorylation.
The natural product, SL0101 (1), is a kaempferol glyco-
side that belongs to the class of kinase inhibitors that are
ATP competitors.2 In vitro and cell-based assays have
demonstrated that SL0101 (1) is potent and remarkably
specific for inhibition of RSK NTKD activity.2 Thus,
SL0101 (1) is useful as a lead compound.
appears that the structure of the RSK2 NTKD ATP
binding pocket is reasonable because we are able to clo-
sely superimpose the conformations of staurosporine
and its derivative, Ro31-8820, bound to the different
kinases (Fig. 1A). Therefore, we used this model to
To further develop SL0101 (1) as an anti-cancer agent
we analyzed the affinity and specificity of various
SL0101 derivatives for inhibition of RSK activity. We
have determined that acylation of the rhamnose moiety
is essential for the specific inhibition of RSK activity,
whereas the 40, 5, and 7-hydroxyl groups on the kaempf-
erol backbone of SL0101 are required for affinity. It is
likely that the phenolic hydroxyl groups facilitate RSK
binding by virtue of their ability to participate in hydro-
gen bonding. We have also found that increasing the
hydrophobic character of SL0101 increases its potency
of inhibition of RSK activity in cell-based assays, pre-
sumably by facilitating cellular uptake. Thus, this infor-
mation should be helpful in designing the next
generation of RSK inhibitors based on the SL0101
pharmacophore.
2. Results
2.1. The ATP binding pocket of the RSK NTKD is unique
Recently an atomic model of the RSK2 NTKD was de-
scribed, which docks the distinct molecular scaffolds,
SL0101, GF109203X, and the staurosporine derivative,
Ro31-8820.10 We compared this model to the X-ray
crystal structures of the archetypal kinases, protein ki-
nase A11 and protein kinase C (PKC),12 both of which
were solved as complexes with staurosporine. We fo-
cused on those residues that comprise the ATP binding
pocket for each kinase. Initially, we superimposed these
structures and removed the amino acid residues contrib-
uted by each of the kinases. Based on this comparison it
c
Figure 1. A model illustrating the unique features of the RSK2 NTKD
ATP binding pocket. (A) The crystal structures of PKAa (1STC) and
PKCh (1XJD), both bound to staurosporine, were superimposed on the
atomic model for RSK2 NTKD interacting with the staurosporine
derivative, Ro31-8820. The superimposition of the compounds is
shown. (B) The residues present in the adenosine-interacting loop (AIL)
and the glycine-rich loop (GRL) for human RSK2 (U08316), PKAa
(X07767), and PKCh (X52479) are shown. The
indicates those
*
residues whose alteration abrogates SL0101 binding. (C) As in A except
the backbone of the polypeptide chains for PKA, PKC, and RSK2
comprising the AIL and the GRL are shown. (D) As in C except the side
chains are also shown and the indicates residues that were mutated.
*

5020
J. A. Smith et al. / Bioorg. Med. Chem. 15 (2007) 5018–5034
study how the RSK NTKD interacts with SL0101. We
demonstrated previously that a major determinant of
specificity of RSK for SL0101 is provided by a unique
sequence present in the ATP binding pocket in the
NTKD of RSK.2 This sequence is predicted to contact
the adenosine moiety of ATP and is referred to as the
adenosine-interacting loop (AIL). The AIL and the gly-
cine-rich loop (GRL) are not contiguous in the primary
sequence (Fig. 1B). The GRL interacts with the phos-
phates of ATP and is a major determinant in the affinity
of interaction of kinases with ATP. The conformation of
the AIL and the GRL present in the ATP binding pock-
et of RSK2 was compared with that of PKA and PKC
by superimposing the structures. One view shows the
backbone of the polypeptides (Fig. 1C) and the other
view also shows the side chains (Fig. 1D). Based on
our analysis the AIL in RSK2 would be farther away
from the adenosine of ATP than in PKA and PKC.
Thus, the ATP binding pocket of RSK2 has a different
overall conformation than those of PKA and PKC. It
is likely that both the differences in size of the ATP bind-
ing pocket as well as the residues that line the pocket are
important in the specific interaction of the NTKD with
SL0101 (1). To further understand the interaction of the
NTKD with SL0101(1) we designed and synthesized
various SL0101 derivatives and determined their affinity
and specificity for inhibition of RSK activity.
in a facile manner. The first analogue synthesized was
triOH-SL0101 (4), which was obtained via deacetylation
and debenzylation of fully protected SL0101 derivative
2,13 as shown in Scheme 1.
The synthesis of the ethylated analogue, Et-SL0101 (6),
was carried out in a similar manner by ethylation (ethyl
iodide, NaH) of the 300 and 400-OH groups of compound
3, followed by hydrogenolysis of the benzyl ethers to
afford the Et-SL0101 derivative (6) in 48% yield over
two steps (Scheme 2).
The synthesis of 5-deoxy-SL0101 (11) was achieved uti-
lizing the Algar–Flynn–Oyamada14–16 reaction of the
known chalcone 7 to afford 5-deoxykaempferol 817 in
one step (Scheme 3). Notably, this reaction only pro-
ceeds in good to moderate yield when the chalcone lacks
functionality at C-5. Glycosylation with previously
reported glycosyl bromide 913 gave the fully protected
5-deoxy-SL0101 (10) in 62% yield. Debenzylation affor-
ded 5-deoxy-SL0101 (11) in 96% yield.
The syntheses of 40, 5, 7-trimethoxy-SL0101 (16), 40, 7-
dideoxy-SL0101 (22), and 7-deoxy-SL0101 (27) utilized
the same methodology described for the previous ana-
logues (vide supra); these are outlined in Scheme 4. Per-
acetylated-SL0101 (28) was obtained via peracetylation
of tri-O-Ac-SL0101 (29)18 in 95% yield (Scheme 5).
2.2. Synthesis of SL0101 derivatives
2.3. Acylation of the rhamnose moiety regulates affinity
and specificity of SL0101 for RSK
The synthesis of the SL0101 analogues was generally
carried out in a manner analogous to that reported for
(1),13 in which the requisite chalcone intermediate was
oxidatively cyclized to afford the protected kaempferol
moiety. Heterogeneous glycosylation conditions
(Ag2O) with the appropriate glycosyl bromide, followed
by deprotection of the phenolic and carbohydrate
hydroxyl groups, afforded the desired SL0101 analogues
To determine the importance of the rhamnose acetyl
groups in the interaction with RSK we synthesized tri-
hydroxy-SL0101 (4) and determined its IC50 for inhibi-
tion of RSK. The IC50 was identified by measuring the
ability of various concentrations of trihydroxy-SL0101
(4) to inhibit RSK activity in an in vitro kinase assay
OBn
OBn
b
OH
BnO
O
O
BnO
O
O
HO
O
O
a
O
O
O
BnO
BnO
HO
O
O
O
2
3
4
AcO
OBn
HO
OBn
HO
OH
OAc
OH
OH
Scheme 1. Reagents: (a) K2CO3, 1:1 THF–MeOH (90%); (b) Pd(OH)2/C, H2, 1:1 THF–MeOH (80%).
OBn
OH
BnO
O
O
HO
O
O
a
b
3
O
O
BnO
HO
6
O
O
5
EtO
OBn
EtO
OH
OEt
OEt
Scheme 2. Reagents and condition: (a) NaH, iodoethane, THF, reflux (48%); (b) Pd(OH)2/C, H2, 1:1 THF–MeOH (100%).

J. A. Smith et al. / Bioorg. Med. Chem. 15 (2007) 5018–5034
5021
R
R
R
O
O
OH
O
H2O2, OH-
OH
R
5
Algar-Flynn-Oyamada reaction
OBn
OBn
BnO
O
O
BnO
OH
OBn
BnO
O
a
b
O
O
Br
OBn
OBn
OAc
OH
O
O
O
8
AcO
10
7
OAc
(62%) 9
OAc
OH
HO
O
c
O
O
OH
OAc
O
OAc
11
˚
Scheme 3. Reagents and conditions: (a) H2O2, 5% aq. NaOH, EtOH, 1,4-dioxane (46%); (b) Ag2O, 9, 4 A mol. sieves, CH2Cl2 (62%); (c) Pd(OH)2/C,
H2, 1:1 THF–MeOH (96%).
R3
4'
R2
OH
O
R3
R2
OH
O
R3
R2
O
O
b
a
7
+
H
R1
O
R1
R1
R1 = OMe, R2 = OMe, R3 = OMe (12)
R1 = OMe, R2 = OMe, R3 = OMe (13)
R1 = OBn, R2 = H (17)
1
2
3
1
2
3
18
19
)
R = OBn, R = H, R = H (
)
R = OBn, R = H, R = H (
R1 = OBn, R2 = H, R3 = OBn (23)
R1 = OBn, R2 = H, R3 = OBn (24)
R3
R3
R3
R2
O
O
R2
O
O
R2
O
O
c
d
e
OH
O
O
R1
R1
OBn
OAc
R1
OH
O
O
OAc
OAc
OAc
R1 = OMe, R2 = OMe, R3 = OMe (14)
R1 = OMe, R2 = OMe, R3 = OMe (15)
R1 = OMe, R2 = OMe, R3 = OMe (16)
1
2
3
1
2
3
1
2
3
20
R = OBn, R = H, R = H (
21
R = OBn, R = H, R = H (
22
R = OH, R = H, R = H ( )
)
)
R1 = OBn, R2 = H, R3 = OBn (25)
R1 = OBn, R2 = H, R3 = OBn (26)
R1 = OH, R2 = H, R3 = OH (27)
Scheme 4. Reagents and conditions: (a) 40% w/v KOH in MeOH, MeOH, reflux; (b) I2, DMSO, 140 ꢁC; (c) i—DMDO (0.05–0.1 M in acetone),
˚
CH2Cl2, 4 ꢁC; ii—pTsOH, CHCl3; (d) Ag2O, 9, 4A mol. sieves, CH2Cl2; (e) Pd(OH)2/C, H2, 1:1 THF–MeOH.
OH
OAc
HO
O
O
AcO
O
OAc O
28
a
O
O
OH
OAc
OAc
OAc
OAc
O
O
OAc
29
OAc
Scheme 5. Reagents and conditions: (a) Ac2O, Et3N, DMAP, CH2Cl2 (95%).

5022
J. A. Smith et al. / Bioorg. Med. Chem. 15 (2007) 5018–5034
using recombinant, constitutively active RSK2.2 The
data were fit using nonlinear regression analysis; the
IC50 value for trihydroxy-SL0101 (4) was >10-fold high-
er than that of SL0101 (1) (Fig. 2). To permit the direct
comparison of the various derivatives with SL0101 we
performed in vitro kinase assays using SL0101 (1) in
parallel. The IC50 value for SL0101 (1) was in agreement
with our previously published result.18 Thus, these
results clearly demonstrate that appropriate regioselec-
tive acetylation of the rhamnose moiety increases the
affinity of RSK for SL0101 (1).
than that of SL0101 (1) (Fig. 2). Previously, we found
that replacement of the acetyl groups on the rhamnose
with butyryl groups, Bu-SL0101 (30), did not substan-
tially alter the IC50.18 However, Bu-SL0101 (30) was
not specific for inhibition of RSK.18 To assess whether
Et-SL0101 (6) was specific for inhibition of RSK activity
we examined whether Et-SL0101 (6) preferentially
inhibited the growth of MCF-7 compared to MCF-
10A cells. We have reported previously that MCF-7 cells
have become dependent on RSK activity for their
growth, whereas inhibition of RSK activity in MCF-
10A cells does not decrease their growth rate.2 There-
fore, this differential dependence on RSK activity for
growth provides a convenient assay to assess the speci-
ficity of the various SL0101 derivatives for RSK. Previ-
ously, we found that at 100 lM Bu-SL0101 (30)
To investigate whether the nature of the O-substituent
on the rhamnose moiety influenced the affinity of RSK
for SL0101 (1) we synthesized Et-SL0101 (6) and found
that its IC50 value for RSK inhibition was ꢀ3-fold lower
Figure 2. Alkylation of the rhamnose regulates the affinity of RSK for SL0101 in an in vitro kinase assay. The potencies of kaempferol, SL0101, and
its derivatives in inhibiting RSK catalytic activity in vitro were measured. Kinase assays were performed using immobilized substrate. The reactions
were initiated by the addition of 10 lM ATP (final concentration). Reactions were terminated after 30 min. All assays measured the initial reaction
velocity. The extent of phosphorylation was determined using phosphospecific antibodies in combination with HRP-conjugated secondary
antibodies. HRP activity was measured as described in the Experimental. Maximum and minimum activity is the relative luminescence detected in the
presence of vehicle and 200 mM EDTA, respectively. Points, mean (n = 2 in quadruplicate); bars = SD. aIC50 values for compounds 1 and 31 were as
reported.2 bIC50 values for compounds 29, 30, 32, and 33 were as reported.18 The log P for each compound is indicated in unbolded parentheses.

J. A. Smith et al. / Bioorg. Med. Chem. 15 (2007) 5018–5034
5023
inhibited both MCF-7 and MCF-10A cell prolifera-
tion.18 Analysis of phosphoprotein profiles in the intact
cell demonstrated that Bu-SL0101 (30), unlike SL0101
(1), inhibited the phosphorylation of a ꢀ25 kDa protein,
as detected by the anti-phospho AKT motif antibody.18
These results demonstrated that Bu-SL0101 (30) is not
specific for RSK, and also validated the differential
growth assay as a measure of the specific inhibition of
RSK activity. At all concentrations tested Et-SL0101
(6) was only slightly better at inhibiting the growth of
MCF-7 compared to MCF-10A cells (Fig. 3). In con-
trast, triOH-SL0101 (4) was completely ineffective at
inhibiting MCF-7 cell growth. To permit the direct com-
parison of the various derivatives with SL0101 we also
performed proliferation assays in MCF-7 and MCF-
10A cells using SL0101 (1). The results were similar to
those reported previously in which concentrations of
SL0101 (1) up to 200 lM did not affect MCF-10A cell
growth, whereas MCF-7 cell growth was completely
inhibited at doses >100 lM. Therefore, Et-SL0101 (6)
is not specific for RSK even though its affinity for
RSK is improved. Taken together, these results demon-
strate that the acetyl groups on the rhamnose moiety
regulate the specificity of SL0101 (1) for RSK.
synthesized 5-deoxy-SL0101 (11), 40, 7-dideoxy-SL0101
(22), and 7-deoxy-SL0101 (27). The IC50 value for
5-deoxy-SL0101 (11) was 24.6 lM; thus elimination of
the 5-hydroxyl group resulted in a >65-fold reduction
in affinity compared to SL0101 (1) (Fig. 4). These data
demonstrate that the 5-hydroxyl group of SL0101 is
essential for RSK binding with high affinity. Elimination
2.4. The 40, 5, and 7 hydroxyl groups on the kaempferol
nucleus influence the affinity of RSK for SL0101
To determine the importance of the phenolic hydroxyl
groups of the kaempferol moiety in RSK binding, we
Figure 3. The nature of O-substitution of the rhamnose moiety
regulates the specificity of SL0101 for inhibition of RSK activity in the
intact cell. The potency of SL0101 derivatives as inhibitors of MCF-7
and MCF-10A cell proliferation was determined. The cell number was
measured after 48 h of treatment. Values given are the fold prolifer-
ation as a percentage of that observed with vehicle-treated cells. Points,
mean (n = 2 in quadruplicate); bars = SD.
Figure 4. The hydroxyl groups on the kaempferol nucleus regulate the
affinity of RSK for SL0101. The potency of the SL0101 derivatives as
inhibitors of RSK catalytic activity in vitro was measured as described
in Figure 2. The log P for each compound is indicated in unbolded
parentheses. ND, not determined.

5024
J. A. Smith et al. / Bioorg. Med. Chem. 15 (2007) 5018–5034
of the 7-hydroxyl group (27) increased the IC50 ꢀ 4-fold
compared to that obtained with SL0101 (1) (Fig. 4). Re-
moval of the 40-hydroxyl group in addition to that of the
7-hydroxyl group (22) resulted in an IC50 of 13.2 lM
(Fig. 4). The substantial decrease in affinity can be
attributed mostly to the absence of the 40-hydroxyl
group (cf. 22 vs 27). Thus, the 40-, 5-, and 7-hydroxyl
groups all contribute to the affinity of SL0101 for RSK.
myristate acetate (PMA), for 20 min.2 In agreement with
our previous results the phosphorylation of p140 is stim-
ulated by PMA and inhibited by SL0101 (1) and U0126,
an inhibitor of mitogen activated kinase kinase, an up-
stream activator of RSK (Fig. 5B). A 2-h pretreatment
with 75 lM 7-deoxy-SL0101 (27) abolished the mito-
gen-induced phosphorylation of p140, whereas 40, 7-
dideoxy-SL0101 (22) inhibited it by ꢀ50%. It is likely
that 7-deoxy-SL0101 (27) is more effective at inhibiting
p140 phosphorylation than 40, 7-dideoxy-SL0101 (22)
due to its higher affinity. At 75 lM both 40,7-dideoxy-
SL0101 (22) and 7-deoxy-SL0101 (27) specifically inhibit
RSK activity, which argues that the 40 and 7 hydroxyl
groups do not substantially contribute to the specificity
of SL0101 (1) for RSK.
It seems likely that loss of hydrogen bonding causes the
decrease in affinity when the phenolic hydroxyl groups
at the 40, 5, and 7 positions of SL0101 are eliminated.
To study this point further, the hydroxyl groups were
O-methylated to disrupt any possible donor hydrogen
bonding. 40, 5, 7-Trimethoxy-SL0101 (16) was completely
ineffective at inhibiting RSK activity (Fig. 4). It could be
argued that the steric effect of the methoxyl group, rather
than disruption of H-bonding, was responsible for the
absence of RSK binding by 40, 5, 7-trimethoxy-SL0101 (16).
Another species studied was peracetylated SL0101 deriv-
ative 28. From a steric perspective, the substituents on the
40, 5, and 7-positions in 28 are larger than those in 16, but
the substantial dipole associated with the ester carbonyl
moieties may provide an opportunity for new protein–
ligand interactions not accessible with 16, thus compen-
sating for the putative loss of H-bonding interactions. It
has already been shown that acetylation of the hydroxyl
groups on the rhamnose moiety does not substantially
alter the affinity of RSK for tri-O-Ac-SL0101 (29).18
Thus, if RSK does not bind 40, 5, 7-trimethoxy-SL0101
(16) purely due to steric constraints, it would be predicted
that peracetylated-SL0101 (28) would also not bind to
RSK. However, we found that RSK activity could be
inhibited by ꢀ50% using 10 lM peracetylated-SL0101
(28) (Fig. 4). It was not possible to perform a complete
dose response curve with peracetylated-SL0101 (28)
because of its limited solubility in aqueous solution.
Nonetheless these results argue that residues in RSK
hydrogen bond to the 5-hydroxyl group and the 40- or
7-hydroxyl group, or both of these groups, in SL0101 (1).
3. Discussion
We have shown that the ability of RSK to interact with
SL0101 depends on the sugar moiety because RSK has a
40-fold greater affinity for SL0101 (1) compared to
kaempferol (31) (Fig. 2).2 During the isolation of
SL0101 we also identified compounds that differ from
SL0101 (1) only in the number and position of the acetyl
groups on the rhamnose moiety ((32) and (33)) and we
found that RSK was able to bind these compounds with
similar in vitro affinities as SL0101 (1) (Fig. 2).18 Fur-
thermore, acylation of all the hydroxy groups on the
rhamnose to generate tri-O-Ac-SL0101 (29) did not sub-
stantially alter the inhibitory potency or specificity for
inhibiting RSK activity in vitro (Fig. 2).18 However, tri-
hydroxy-SL0101 (4) was found to be much less potent at
inhibiting RSK than SL0101 (1), indicating that some of
the sugar acyl groups must be required for enzyme
binding.
Our studies further suggest that the efficacy of SL0101 in
intact cells could be improved by increasing its hydro-
phobic character. This conclusion is based on our
observations that 7-deoxy-SL0101 (27) and 40, 7-dide-
oxy-SL0101 (22) are more potent RSK inhibitors than
SL0101 in the intact cell despite their lower in vitro affin-
ity for RSK. It is highly likely that 7-deoxy-SL0101 (27)
and 40, 7-dideoxy-SL0101 (22) are more potent than
SL0101 (1) in the proliferation assay, which occurs over
a longer time period, because their increased hydropho-
bic character facilitates cellular uptake, which compen-
sates for their lower affinity. Additionally, it was
observed that 15 lM peracetylated SL0101 (28) inhib-
ited MCF-7 cell proliferation by ꢀ50% while having
no effect on MCF-10A cell proliferation (data not
shown). We attempted to use Cremophor to aid in solu-
bilizing the peracetylated-SL0101 (28) in the prolifera-
tion assays but nonspecific toxicity was observed with
percentages of Cremophor >1%, which were needed to
keep the peracetylated-SL0101 (28) soluble at concen-
trations >15 lM. Thus even though peracetylated-
SL0101 (28) is less effective than SL0101 (1) in the
in vitro kinase assay, its greater hydrophobic character
makes it more potent in the intact cell. Taken together,
these results suggest that the efficacy of SL0101 (1) is
limited in the intact cell by cellular uptake. It is of inter-
To determine whether the 40-, 5- or 7-hydroxyl groups
also contributed to the specific inhibition of RSK by
SL0101 we examined the ability of the derivatives to
inhibit MCF-7 and MCF-10A cell growth. The 5-
deoxy-SL0101 (11) was a poor inhibitor of MCF-7 pro-
liferation causing only a 50% reduction at 100 lM
(Fig. 5A), which was expected based on the low affinity
of RSK for 5-deoxy-SL0101 (11). The specificity of
7-deoxy-SL0101 (27) and 40, 7-dideoxy-SL0101 (22) for
RSK at 100 lM was decreased in comparison to
SL0101 (1) (Fig. 5A). However, at 75 lM both deriva-
tives completely inhibited the growth of MCF-7 cells,
whereas MCF-10A growth was decreased by only
ꢀ25–40%. Thus, both of these derivatives are more
potent at 75 lM for inhibition of RSK activity in the
proliferation assay than SL0101 (1). To directly assess
the efficacy of these derivatives for inhibition of RSK
activity in the intact cell, their ability to decrease the
mitogen-induced phosphorylation of the RSK substrate,
p140, was analyzed. In these assays MCF-7 cells were
pre-treated for 2 h with the indicated concentration of
inhibitor and then stimulated with the mitogen, phorbol

J. A. Smith et al. / Bioorg. Med. Chem. 15 (2007) 5018–5034
5025
Figure 5. Increasing hydrophobic character of SL0101 improves potency in the intact cell. (A) The efficacy of the SL0101 derivatives as inhibitors of
MCF-7 and MCF-10A cell proliferation was determined as in Figure 3. (B) The potency of the SL0101 derivatives as inhibitors of RSK activity in
intact cells was analyzed. Serum-deprived MCF-7 cells were pre-incubated with vehicle, or the indicated concentration of inhibitor for 2 h. Cells were
treated with 500 nM PMA or vehicle for 20 min prior to lysis. Protein concentration of lysates was measured and lysates were electrophoresed,
transferred, and immunoblotted. Equal loading of lysate is demonstrated by the anti-Ran immunoblot.
est to note that the acetyl groups present in SL0101 (1)
and peracetylated-SL0101 (28) may well be hydrolyzed
by ubiquitous intracellular esterases. Thus, both com-
pounds could potentially be metabolized to triOH-
SL0101 (4), which is a less effective inhibitor of RSK
activity than SL0101 (1) and is ineffective at inhibiting
MCF-7 cell growth. Therefore, the efficacy of SL0101
(1) in the intact cell is most likely limited both by cellular
uptake and metabolism.
In the RSK2 NTKD atomic model described by Nguyen
et al.10 SL0101 (1) is positioned so that the kaempferol
nucleus is located near the AIL and the rhamnose moi-
ety is near the GRL. Based on our structural analysis, it
is likely that residues within the AIL are involved in
hydrogen bonding with the hydroxyl groups on the
kaempferol nucleus of SL0101 (1). Furthermore, it
appears that alkylation of the rhamnose moiety can
facilitate the interaction with the GRL of a number of
different kinases. However, the acetylation of the rham-
nose moiety confers specificity for RSK.
The hydroxyl groups, in particular the 5-OH and 40-OH
groups on the kaempferol nucleus, are important deter-
minants of the affinity of SL0101 for RSK. It is likely
that these groups are involved in hydrogen bonding
because methylating the hydroxy groups eliminated
inhibition of RSK activity. Interestingly, the 40- and
7-OH groups do not appear to contribute critically to
the specific interaction of RSK with SL0101 because
both 7-deoxy-SL0101 (27) and 40,7-dideoxy-SL0101
(22) were able to specifically inhibit RSK activity in
the proliferation assay at 75 lM concentration.
These studies should aid in the design of SL0101 (1)
derivatives having optimized pharmacokinetic and phar-
macodynamic properties. A dihydropteridinone deriva-
tive, BI-D1870, has recently been identified that also
inhibits RSK activity.19 SL0101 and BI-D1870 are
ATP competitors that directly inhibit the phosphoryla-
tion of exogenous substrates through inhibition of
RSK NTKD activity. Pyrrolo[2,3-d]pyrimidine deriva-
tives that are irreversible inhibitors of the RSK CTKD

5026
J. A. Smith et al. / Bioorg. Med. Chem. 15 (2007) 5018–5034
have also been reported.20 These inhibitors are thought
to inhibit RSK NTKD activity indirectly by preventing
its activation by the CTKD. The in vivo potencies of
BI-D1870 and the pyrrolo[2,3-d]pyrimidine derivatives
as inhibitors of RSK activity have not yet been reported.
performed under a nitrogen or argon atmosphere. Flash
chromatography was performed using Silicycle 40–60
mesh silica gel. Analytical TLC was performed using
0.25 mm EM silica gel 60 F250 plates that were visualized
by irradiation (254 nm) or by staining with Hanessian’s
stain (cerium molybdate). Optical rotations were
1
In addition to its involvement in various cancers, inap-
propriate RSK activity has been implicated in the etiol-
obtained using a Jasco digital polarimeter. H and 13C
NMR spectra were obtained using 300 and 500 MHz
Varian instruments. Chemical shifts are reported in
parts per million (ppm, d) referenced to the residual
1H resonance of the solvent (CDCl3, 7.26 ppm;
DMSO-d6, 2.49 ppm). 13C spectra were referenced to
the residual 13C resonance of the solvent (CDCl3,
77.3 ppm; DMSO-d6, 39.5 ppm). Splitting patterns are
designated as follows: s, singlet; br, broad; d, doublet;
dd, doublet of doublets; t, triplet; q, quartet; m,
multiplet. High resolution mass spectra were obtained
at the Michigan State University-NIH Mass Spectrome-
try Facility. Log P values were calculated using
Molinspiration Property Calculation Service (www.
molinspiration.com).
ogy of
a
number of other diseases including
cardiomyopathy, infection, and lead poisoning.21–23
Further, elevated activity of MAPK, the immediate
upstream RSK activator, has also been shown in various
neural and heart pathologies, inflammation, and Borna
disease.24–35 It is possible that RSK is also involved in
these disorders through activation of its kinase activity
by MAPK. Therefore, it is possible to envision a number
of potential clinical uses for RSK inhibitors. However,
RSK2 is important in normal development and has been
associated with Coffin-Lowry syndrome.36 In this syn-
drome patients have an X-linked mental retardation
condition and skeletal abnormalities. Some of the
pathology associated with this syndrome may be due
to the fact that the patients express truncated mutants
of RSK2. These mutants may have functions that differ
from wild-type RSK2. In a RSK2 mouse knockout
model MAPK activity is elevated and it is thought that
RSK2 may act as a negative regulator of MAPK
activity.37 Increased MAPK activity would have wide
ranging biological effects and therefore might be delete-
rious.1 However, it is not clear whether this feedback
loop exists only as a consequence of the loss of the
RSK2 protein during development. Interestingly, in
Drosophila, flies that have a kinase-dead RSK mutant
instead of wild-type RSK develop normally.38 In com-
parison, flies that are null for RSK have increased
MAPK activity and are defective. These results suggest
that it may be possible to inhibit RSK activity without
deleterious consequences to a developing organism. Fur-
thermore, our data suggest that RSK activity is not
required for the maintenance of normal homeostasis in
nontumorigenic cells.2 Thus, it is possible that an inhib-
itor of RSK activity may be well tolerated in vivo.
OBn
BnO
O
O
OBn O
OBn
OH
O
OH
4.2.1. 40,5,7-Tri-O-benzyl-kaempferol 3-O-(200-O-benzyl-
a-L-rhamnopyranoside) (3). To a solution containing
0.08 g (0.09 mmol) of 2 in 5 mL of 1:1 THF–MeOH
was added catalytic K2CO3 and the reaction mixture
was stirred at room temperature under a N2 atmo-
sphere for 3 h. The reaction mixture was then filtered
though a Celiteꢂ pad and washed with ethyl acetate,
and the filtrate was concentrated under diminished
pressure. The residue was purified by flash chromato-
graphy on a silica gel column (25 · 3 cm). Elution with
1:1 hexanes–ethyl acetate gave 3 as a colorless foam:
4. Experimental
4.1. Modeling
yield 0.061 g (90%); silica gel TLC Rf 0.19 (1:1 hex-
20
D
anes–ethyl acetate); ½aꢁ ꢂ40.1 (c 1.0, CHCl3); 1H
NMR (CDCl3) d 0.95 (d, 3H, J = 6.0 Hz), 2.33 (br s,
2H), 3.25 (m, 1H), 3.36 (t, 1H, J = 9.3 Hz), 3.74 (m,
1H), 4.38 (d, 1H, J = 1.5 Hz), 4.71 (d, 1H,
J = 11.7 Hz), 4.88 (d, 1H, J = 11.7 Hz), 5.09 (s, 2H),
5.14 (s, 2H), 5.26 (s, 2H), 5.78 (s, 1H), 6.49 (s, 1H),
6.58 (s, 1H), 7.10 (d, 1H, J = 7.5 Hz), 7.30–7.43 (m,
19H), 7.62 (d, 2H, J = 7.5 Hz) and 7.81 (d, 2H,
J = 8.4 Hz); 13C NMR (CDCl3) d 17.36, 70.00, 70.32,
70.73, 70.99, 71.43, 72.73, 73.54, 77.50, 77.85, 94.16,
98.44, 98.76, 110.25, 114.93, 123.44, 126.83, 127.68,
127.86, 128.21, 128.37, 128.48, 128.70, 128.80, 128.91,
129.01, 130.71, 135.84, 136.60, 138.07, 138.30, 154.56,
159.12, 159.99, 160.67, 163.10 and 173.63; mass
spectrum (FAB), m/z 793.3010 (M+H)+ (C49H45O10 re-
quires 793.3013).
The models were of generated by Swiss PdbViewer
3.70b0 and are based on the coordinates under Protein
Data Bank entry codes 1STC for PKAa11 and 1XJD
for PKCh,12 both bound to staurosporine. The models
for PKAa and PKCh were superimposed on the atomic
model for RSK2 interacting with the staurosporine
derivative, Ro31-8820.10
4.2. Chemistry
Reagents and solvents were reagent grade and were used
without further purification. Methylene chloride was
distilled from calcium hydride. Anhydrous grade THF
was purchased from VWR. All reactions involving air-
or moisture- sensitive reagents or intermediates were

J. A. Smith et al. / Bioorg. Med. Chem. 15 (2007) 5018–5034
5027
1
0.46 (2:1 hexanes–ethyl acetate); H NMR (CDCl3) d
0.94 (d, 3H, J = 5.4 Hz), 1.17 (m, 6H), 3.26 (m, 2H),
3.46 (q, 2H, J = 6.9 Hz); 3.60 (m, 2H), 3.83 (m, 1H),
4.41 (m, 1H), 4.81 (m, 2H), 5.08 (s, 2H), 5.14 (s, 2H),
5.25 (s, 2H), 5.52 (s, 1H), 6.49 (d, 1H, J = 1.8 Hz),
6.55 (d, 1H, J = 2.1 Hz), 7.07 (d, 2H, J = 9.0 Hz),
7.23–7.47 (m, 16H), 7.50 (d, 2 H, J = 7.2 Hz), 7.64 (d,
2H, J = 7.5 Hz) and 7.80 (d, 2H, J = 8.7 Hz); 13C
NMR (CDCl3) d 15.88, 16.08, 17.73, 65.33, 68.31,
70.17, 70.37, 70.72, 71.00, 72.53, 75.23, 79.25, 79.64,
94.12, 98.38, 100.33, 110.29, 114.90, 123.62, 126.87,
127.58, 127.65, 127.87, 128.14, 128.35, 128.47, 128.69,
128.80, 128.93, 129.01, 130.74, 135.87, 136.65, 136.70,
138.82, 138.93, 154.54, 159.11, 160.01, 160.59, 163.04
and 173.68.
OH
HO
O
O
O
OH
OH
OH
O
OH
4.2.2. Kaempferol 3-O-a-L-rhamnopyranoside (trihy-
droxy-SL0101) (4). A suspension containing 0.06 g
(0.07 mmol) of 3 and 0.010 g of Pd(OH)2/C in 5 mL of
1:1 THF–MeOH was purged three times with H2. The
reaction vessel was kept under a H2 atmosphere for
1 h, then filtered through a Celiteꢂ pad and washed with
ethyl acetate. The filtrate was concentrated under dimin-
ished pressure to afford a solid. Product 4 was pure and
further purification was not necessary: yield 0.024 g
(80%); silica gel TLC Rf 0.15 (1:1:0.1 hexanes–ethyl ace-
OH
HO
O
O
20
D
O
1
tate–methanol); ½aꢁ ꢂ111.4 (c 0.8, CHCl3); H NMR
OH
OH
(CD3OD) d 0.89 (d, 3H, J = 5.4 Hz), 2.43 (m, 1H),
2.46 (t, 1H, J = 7.8 Hz), 3.69 (m, 1H), 4.19 (m, 1H),
4.32 (t, 1H, J = 6.9 Hz), 5.34 (d, 1H, J = 1.5 Hz), 6.16
(d, 1H, J = 1.8 Hz), 6.33 (d, 1H, J = 1.8 Hz), 6.91 (d,
2H, J = 8.7 Hz) and 7.72 (d, 2H, J = 8.7 Hz); 13C
NMR (CD3OD) d 17.62, 71.94, 72.01, 72.16, 73.28,
95.22, 100.40, 103.47, 104.88, 116.03, 122.98, 132.34,
135.06, 158.91, 159.47, 161.50, 163.07, 166.42 and
178.56; mass spectrum (FAB), m/z 433.1136 (M+H)+
(C21H21O10 requires 433.1134).
O
OEt
OEt
4.2.4. Kaempferol 3-O-(300,400-di-O-ethyl-a-L-rhamnopyr-
anoside) (Et-SL0101) (6). A suspension containing
0.04 g (0.05 mmol) of 5 and 0.02 g of Pd(OH)2/C in
6 mL of 1:1 THF–MeOH was purged three times with
H2. The reaction vessel was kept under a H2 atmo-
sphere for 1 h, then filtered through a Celiteꢂ pad
and washed with ethyl acetate. The filtrate was con-
centrated under diminished pressure. The residue was
purified by flash chromatography on a silica gel col-
umn (20 · 2 cm). Elution with 1:1:0.1 hexanes–ethyl
acetate–methanol gave 6 as a colorless oil: yield
OBn
BnO
O
O
OBn O
OBn
OEt
O
0.02 g (100%); silica gel TLC Rf 0.39 (1:1:0.1 hex-
20
D
anes–ethyl acetate–methanol); ½aꢁ ꢂ115.8 (c 1.1, ethyl
acetate); 1H NMR (acetone-d6)
d 0.99 (d, 3H,
OEt
J = 6.3 Hz), 1.16 (t, 3H, J = 6.9 Hz), 1.28 (t, 3H,
J = 6.9 Hz), 3.23 (t, 1H, J = 9.3 Hz), 3.47 (m, 1H),
3.63 (m, 3H), 3.82 (m, 2H), 4.04 (br s, 1H), 4.40
(m, 1H), 5.58 (d, 1H, J = 1.8 Hz), 6.35 (d, 1H,
J = 1.8 Hz), 6.56 (d, 1H, J = 2.1 Hz), 7.10 (d, 2H,
J = 8.7 Hz), 7.94 (d, 2H, J = 8.7 Hz), 9.47 (br s, 2H)
and 12.80 (s, 1H); 13C NMR (acetone-d6) d 15.95,
16.08, 17.93, 65.47, 68.08, 68.93, 70.45, 79.81, 80.12,
94.54, 99.53, 102.53, 105.81, 116.26, 122.52, 131.71,
135.69, 157.97, 158.37, 160.87, 163.20, 164.89 and
179.26; mass spectrum (FAB), m/z 489.1762 (M+H)+
(C25H29O10 requires 489.1760).
4.2.3. 40,5,7-Tri-O-benzyl-kaempferol 3-O-(200-O-benzyl-
300,400-di-O-ethyl-a-L-rhamnopyranoside) (5). To a solu-
tion containing 0.10 g (0.13 mmol) of 3 in 5 mL of anhy-
drous THF at 0 ꢁC was added 0.02 g (0.51 mmol) of a
60% dispersion of NaH in mineral oil. The reaction mix-
ture was stirred at this temperature for 15 min, then
0.12 g (0.06 mL, 0.76 mmol) of iodoethane was added
and the reaction mixture was heated to reflux for 20 h.
The cooled reaction mixture was quenched with satu-
rated aq NH4Cl and diluted with ether. The aqueous
layer was extracted with three 10-mL portions of ether.
The combined organic layer was washed with two 15-
mL portions of brine, dried (MgSO4), and concentrated
under diminished pressure. The resulting residue was
purified by flash chromatography on a silica gel column
(25 · 3 cm). Elution with 6:1 hexanes–ethyl acetate gave
5 as a colorless oil: yield 0.053 g (48%); silica gel TLC Rf
OBn
BnO
O
O
OH

5028
J. A. Smith et al. / Bioorg. Med. Chem. 15 (2007) 5018–5034
4.2.5. 40,7-Di-O-benzyl-5-deoxykaempferol (8).17 To a
stirred suspension containing 2.00 g (4.58 mmol) of 1-(4-
benzoxy-2-hydroxyphenyl)-3-(4-benzoxyphenyl) prope-
none (7) in 60 mL of EtOH was added 18 mL of dioxane,
followed by 2 mL of H2O2 and 16 mL of 5% aqueous
NaOH. The reaction mixture was stirred at room temper-
ature overnight, then heated to 50 ꢁC for 3 h. The cooled
reaction mixture was poured into 200 mL of H2O and
acidified with 1 N HCl. The yellow precipitate was col-
lected by filtration and washed with hot EtOH. The insol-
uble solid was collected and dried to afford 8 as a light
yellow solid: yield 0.85 g (46%); mp 208–210 ꢁC, lit.17
mp 210–212 ꢁC; silica gel TLC Rf 0.65 (1:1 hexanes–ethyl
acetate); 1H NMR (DMSO-d6) d 5.22 (s, 2H), 5.28 (s, 2H),
7.12 (dd, 1H, J = 8.7 and 1.8 Hz), 7.22 (d, 2H, J = 9.0 Hz),
7.36–7.50 (m, 11H), 8.01 (d, 1H, J = 8.7 Hz), 8.18 (d, 2H,
J = 8.7 Hz) and 9.37 (br s, 1H).
OH
HO
O
O
O
O
AcO
OH
OAc
4.2.7. 5-Deoxykaempferol 3-O-(300,400-di-O-acetyl-a-L-
rhamnopyranoside) (5-deoxy-SL0101) (11). A suspension
containing 0.10 g (0.14 mmol) of 10 and 0.08 g of
Pd(OH)2/C in 5 mL of 1:1 THF–MeOH was purged
with H2 and kept under a H2 atmosphere for 1 h.
The reaction mixture was then diluted with ethyl ace-
tate and filtered through a Celiteꢂ pad. The filtrate
was concentrated under diminished pressure to give
11 as a light brown solid: yield 0.065 g (96%); silica
gel TLC Rf 0.24 (1:1:0.1 hexanes–ethyl acetate–metha-
OBn
BnO
O
O
23
D
1
nol); ½aꢁ ꢂ130.2 (c 0.6, MeOH); H NMR (DMSO-
d6) d 0.67 (d, 3H, J = 12.9 Hz), 1.95 (s, 3H), 1.99 (s,
3H), 3.27 (m, 1H), 4.22 (br s, 1H), 4.94 (m, 1H),
4.98 (dd, 1H, J = 10.2 and 2.7 Hz), 5.38 (d, 1H,
J = 1.5 Hz), 5.65 (br s, 1H), 6.89 (d, 2H, J = 3.3 Hz),
6.91 (d, 2H, J = 8.7 Hz), 7.71 (d, 2H, J = 8.4 Hz),
7.90 (d, 1H, J = 8.7 Hz) and 10.50 (br s, 2H); 13C
NMR (DMSO-d6) d 17.03, 20.54, 20.78, 67.37, 67.76,
69.77, 71.05, 100.45, 102.19, 115.12, 115.40, 116.19,
120.85, 126.85, 126.74, 130.44, 135.41, 156.36,
156.62, 159.89, 162.76, 169.64, 170.01, 170.36 and
172.75; mass spectrum (FAB), m/z 501.1395 (M+H)+
(C25H25O11 requires 501.1397).
O
O
AcO
OBn
OAc
4.2.6. 40,7-Di-O-benzyl-5-deoxykaempferol 3-O-(200-O-
benzyl-300,400-di-O-acetyl-a-L-rhamnopyranoside) (10). To
a stirred suspension containing 0.10 g (0.24 mmol) of
˚
8, 0.11 g (0.48 mmol) of Ag2O, and 4A molecular
sieves in 5 mL of CH2Cl2 was added 0.12 g
(0.31 mmol) of 3,4-di-O-acetyl-2-O-benzyl-a-L-rhamno-
pyranosyl bromide (9)13 as a solution in 5 mL of
CH2Cl2. The reaction mixture was allowed to stir at
room temperature for 4 h, then diluted with 20 mL
of CH2Cl2 and filtered through a Celiteꢂ pad. The fil-
trate was concentrated under diminished pressure and
the residue was purified by flash chromatography on a
silica gel column (25 · 3 cm). Elution with 3:1 hex-
anes–ethyl acetate gave 10 as a colorless solid: yield
0.11 g (62%); silica gel TLC Rf 0.39 (2:1 hexanes–ethyl
OMe
MeO
OH
OMe O
4.2.8. E-3-(4-Methoxyphenyl)-1-(2,4-bismethoxy-6-hydro-
xyphenyl)propenone (12).39 To a solution containing
3.0 g (1.03 mmol) of 40,5,7-trihydroxyflavanone in
50 mL of acetone was added 4.31 g (34.2 mmol) of di-
methyl sulfate followed by 6.18 g (44.1 mmol) of
K2CO3. The reaction mixture was heated to reflux under
N2 overnight then the solvent was concentrated under
diminished pressure and the residue was redissolved in
ethyl acetate. The organic layer was then washed with
two 100-mL portions of brine, one 100-mL portion of
1 N HCl, dried (MgSO4), and concentrated under
diminished pressure. Purification by flash chromatogra-
phy on a silica gel column (25 · 4 cm) and elution with
2:1 hexanes–ethyl acetate gave 12 as a yellow solid: yield
1.72 g (50%); mp 113–114 ꢁC, lit.39 mp 113–114 ꢁC; sil-
ica gel TLC Rf 0.45 (2:1 hexanes–ethyl acetate); 1H
NMR (CDCl3) d 3.80 (s, 3H), 3.82 (s, 3H), 3.88 (s,
3H), 5.92 (d, 1H, J = 2.4 Hz), 6.07 (d, 1H, J = 2.4 Hz),
22
1
acetate); ½aꢁ ꢂ81.3 (c 0.1, CHCl3); H NMR (CDCl3)
D
d 0.87 (d, 3H, J = 6.3 Hz), 1.96 (s, 3H), 1.98 (s, 3H),
3.44 (m, 1H), 4.31 (d, 1H, J = 1.5 Hz), 4.57 (d, 1H,
J = 12.3 Hz), 4.78 (d, 1H, J = 12.3 Hz), 5.14 (m, 6H),
5.28 (dd, 1H, J = 10.2 and 3.3 Hz), 5.73 (s, 1H),
6.97 (br s, 1H), 7.05 (d, 1H, J = 8.7 Hz), 7.12 (d,
2H, J = 8.4 Hz), 7.40 (m, 14H), 7.87 (d, 2H,
J = 8.4 Hz) and 8.13 (d, 1H, J = 9.0 Hz); 13C NMR
(CDCl3) d 17.42, 21.04, 21.11, 68.50, 70.35, 70.78,
71.26, 72.90, 75.78, 77.78, 98.80, 101.32, 115.05,
115.24, 118.43, 123.52, 127.42, 127.53, 127.72,
127.92, 128.21, 128.28, 128.45, 128.64, 128.93,
129.00, 130.84, 135.91, 136.53, 137.34, 138.09,
156.75, 157.21, 160.85, 163.33, 170.03, 170.40 and
173.99; mass spectrum (FAB), m/z 771.2803 (M+H)+
(C46H43O11 requires 771.2806).

J. A. Smith et al. / Bioorg. Med. Chem. 15 (2007) 5018–5034
5029
6.90 (d, 2H, J = 9.0 Hz), 7.54 (d, 2H, J = 9.0 Hz), 7.77
(s, 2H) and 14.45 (s, 1H); 13C NMR (CDCl3) d 55.60,
55.76, 56.02, 91.37, 93.99, 114.55, 125.27, 128.46,
130.32, 142.66, 161.56, 162.65, 166.23, 168.59 and
192.75.
(CDCl3) d 55.3, 55.7, 56.3, 92.3, 95.6, 106.1, 113.9,
123.5, 128.8, 137.4, 142.2, 158.8, 160.4, 164.2 and
171.8.
OMe
MeO
O
O
OMe
O
MeO
O
OMe
O
AcO
AcO
OMe O
OBn
4.2.9. 5,7-Dimethoxy-2-(4-(methoxy)phenyl)-4H-chromen-
4-one (13).40 To a solution containing 1.40 g (4.53 mmol)
of 12 in 50 mL of DMSO was added 0.12 g (0.45 mmol) of
I2. The reaction mixture was heated to reflux overnight
under N2, then cooled to room temperature, diluted with
200 mL ethyl acetate, and washed with two 100-mL por-
tions of 1 N HCl. The organic layer was separated, dried
(MgSO4), and concentrated under diminished pressure.
The residue was purified by flash chromatography on a
silica gel column (28 · 4 cm). Elution with 1:2 hexanes–
ethyl acetate gave 13 as an off-white solid: yield 1.21 g
(86%); mp 153–155 ꢁC, lit.40 mp 158–159 ꢁC; silica gel
TLC Rf 0.10 (1:2 hexanes–ethyl acetate); 1H NMR
(CDCl3) d 3.83 (s, 3H), 3.86 (s, 3H), 3.91 (s, 3H), 6.31
(d, 1H, J = 2.1 Hz), 6.50 (d, 1H, J = 2.1 Hz), 6.54 (s,
1H), 6.93 (d, 2H, J = 8.7 Hz) and 7.75 (d, 2H,
J = 8.7 Hz); 13C NMR (CDCl3) d 55.67, 55.96, 56.61,
92.99, 96.24, 107.76, 109.32, 114.51, 123.92, 127.75,
159.97, 160.82, 160.99, 162.21, 164.07 and 177.84.
4.2.11. 40,5,7-Tri-O-methylkaempferol 3-O-(200-O-benzyl-
300,400-di-O-acetyl-a-L-rhamnopyranoside) (15). To a stir-
red suspension containing 0.10 g (0.31 mmol) of 14,
˚
0.14 g (0.61 mmol) of Ag2O, and 4 A molecular sieves
in 5 mL of CH2Cl2 was added 0.25 g (0.61 mmol) of
913 as a solution in 5 mL of CH2Cl2. The reaction mix-
ture was stirred at room temperature for 4 h, then di-
luted with 20 mL of CH2Cl2 and filtered through a
Celiteꢂ pad. The filtrate was concentrated under dimin-
ished pressure and the residue was purified by flash
chromatography on a silica gel column (25 · 3 cm). Elu-
tion with 2% MeOH in CHCl3 gave 15 as a colorless
solid: yield 0.09 g (43%); silica gel TLC Rf 0.58 (5%
22
D
MeOH in CHCl3); ½aꢁ ꢂ64.0 (c 0.3, CHCl3); 1H
NMR (CDCl3) d 0.82 (d, 3H, J = 6.0 Hz), 1.93 (s, 3H),
1.96 (s, 3H), 3.25 (m, 1H), 3.86 (s, 3H), 3.87 (s, 3H),
3.95 (s, 3H), 4.31 (m, 1H), 4.59 (d, 1H, J = 12.0 Hz),
4.80 (d, 1H, J = 12.0 Hz), 5.06 (t, 1H, J = 9.9 Hz), 5.23
(dd, 1H, J = 10.2 and 3.3 Hz), 5.79 (d, 1H,
J = 0.9 Hz), 6.34 (d, 1H, J = 2.1 Hz), 6.47 (d, 1H,
J = 2.1 Hz), 7.03 (d, 2H, J = 9.0 Hz), 7.24–7.40 (m,
5H) and 7.81 (d, 2H, J = 8.7 Hz); 13C NMR (CDCl3) d
17.39, 20.99, 21.09, 55.70, 56.01, 56.71, 68.33, 70.84,
71.34, 73.06, 76.06, 92.72, 96.20, 98.41, 109.61, 114.11,
123.19, 127.77, 128.13, 128.44, 130.67, 137.37, 138.28,
154.61, 159.22, 161.22, 161.51, 164.21, 170.02, 170.41
and 173.65; mass spectrum (FAB), m/z 649.2288
(M+H)+ (C35H37O12 requires 649.2285).
OMe
MeO
O
OH
OMe O
4.2.10. 40,5,7-Tri-O-methylkaempferol (14).41 To a solu-
tion containing 0.30 g (0.96 mmol) of 13 in 30 mL of
CH2Cl2 at 0 ꢁC was added 30 mL of a 0.9–0.11 M solu-
tion of dimethyl dioxirane (DMDO) in acetone. The
reaction mixture was allowed to stir at 4 ꢁC under N2
overnight. The solvent was concentrated under dimin-
ished pressure and the residue was then dissolved in
50 mL of CHCl3. Catalytic p-TsOH was added and the
reaction mixture was stirred at 0 ꢁC for 30 min. The sol-
vent was concentrated under diminished pressure, and
the resulting residue was purified by flash chromatogra-
phy on a silica gel column (25 · 3 cm). Elution with 2%
MeOH in CHCl3 gave 14 as a light brown solid: yield
0.16 g along with 0.09 g of unreacted 13 (63% based
on consumed starting material); silica gel TLC Rf 0.42
OMe
MeO
O
O
OMe O
O
AcO
AcO
OH
4.2.12. 40,5,7-Tri-O-methylkaempferol 3-O-(300,400-di-O-
acetyl-a-L-rhamnopyranoside) (40,5,7-trimethoxy-SL0101)
(16). A suspension containing 0.09 g (0.13 mmol) of 15
and 0.05 g of Pd(OH)2/C in 6 mL of 1:1 THF–MeOH
was purged with H2 and kept under a H2 atmosphere
for 1 h. The reaction mixture was then diluted with ethyl
acetate and filtered through a Celiteꢂ pad. The filtrate
1
(2% MeOH in CHCl3); H NMR (CDCl3) d 3.86 (s,
3H), 3.90 (s, 3H), 3.97 (s, 3H), 6.33 (d, 1H,
J = 2.1 Hz), 6.54 (d, 1H, J = 2.1 Hz), 7.02 (d, 2H,
J = 8.6 Hz) and 8.16 (d, 2H, J = 8.7 Hz); 13C NMR

5030
J. A. Smith et al. / Bioorg. Med. Chem. 15 (2007) 5018–5034
was concentrated under diminished pressure to give 16 as
a colorless solid: yield 0.043 g (77%); silica gel TLC Rf 0.16
gave 18 as an orange-yellow solid: yield 2.95 g (72%);
mp 128–130 ꢁC, lit.45 mp 130–131 ꢁC; silica gel TLC Rf
22
1
(1:1:0.1 hexanes–ethyl acetate–methanol); ½aꢁ ꢂ146.6 (c
0.51 (3:1 hexanes–ethyl acetate); H NMR (CDCl3) d
D
0.6, CHCl3); 1H NMR (CDCl3) d 0.79 (d, 3H, J = 6.0 Hz),
1.98 (s, 3H), 2.01 (s, 3 H), 3.32 (m, 1H), 3.88 (s, 6H), 3.94
(s, 3H), 4.17 (br s, 1H), 4.69 (br s, 1H), 5.08 (t, 1H,
J = 9.6 Hz), 5.30 (dd, 1H, J = 9.6 and 3.0 Hz), 5.54 (d,
1H, J = 1.8 Hz), 6.31 (d, 1H, J = 1.5 Hz), 6.45 (d, 1H,
J = 1.8 Hz), 7.03 (d, 2H, J = 8.7 Hz) and 7.84 (d, 2H,
J = 8.7 Hz); 13C NMR (CDCl3) d 17.39, 21.04, 21.18,
55.72, 56.02, 56.44, 68.19, 69.26, 71.58, 71.76, 92.58,
96.01, 101.54, 109.33, 114.06, 123.21, 130.79, 137.77,
154.96, 159.23, 161.22, 161.54, 164.38, 170.18, 170.39
and 173.75; mass spectrum (FAB), m/z 559.1813
(M+H)+ (C28H31O12 requires 559.1815).
5.11 (s, 2H), 6.53 (dd, 1H, J = 8.4 and 0.6 Hz), 6.65
(dd, 1H, J = 8.4 and 0.9 Hz), 7.10 (m, 2H), 7.21 (m,
2H), 7.38–7.42 (m, 5H), 7.49 (m, 2H), 7.74 (m, 1H),
7.86 (m, 1H) and 13.57 (s, 1H); 13C NMR (CDCl3) d
71.64, 102.51, 111.60, 127.88, 128.74, 128.78, 128.84,
128.93, 129.15, 130.33, 135.28, 135.82, 136.32, 143.61,
160.51, 165.76 and 194.73.
O
OBn O
OH
4.2.15. 5-(Benzoxy)-2-phenyl-4H-chromen-4-one (19).46
To a solution containing 1.0 g (3.03 mmol) of 18 in
25 mL of DMSO was added 0.09 g (0.35 mmol) of I2.
The reaction mixture was heated to 140 ꢁC and stirred at
this temperature for 1 h. The cooled reaction mixture
was then poured into 150 mL of 1 N HCl and extracted
with three 100-mL portions of ethyl acetate. The com-
bined organic layer was washed with 200 mL of brine,
dried (MgSO4), and concentrated under diminished pres-
sure. The resulting residue was then purified by flash chro-
matography on a silica gel column (20 · 4 cm). Elution
with 2:1 hexanes–ethyl acetate gave 19 as an off-white so-
lid: yield 0.75 g (75%); mp 146–148 ꢁC; silica gel TLC Rf
OBn O
4.2.13. 2-Benzoxy-6-hydroxyacetophenone (17).42–44 To a
solution containing 10.0 g (66.0 mmol) of 2,6-dihydroxy-
acetophenone in 100 mL of acetone were added 13.3 g
(9.2 mL, 80.0 mmol) of BnBr, 18.0 g (106 mmol) of KI,
and 30.0 g (212 mmol) of K2CO3. The reaction mixture
was heated to reflux overnight under a N2 atmosphere,
then cooled and filtered. The filtrate was concentrated un-
der diminished pressure, then the residue was redissolved
in ethyl acetate, washed with 250 mL of H2O, dried
(MgSO4), and concentrated under diminished pressure.
The residue was purified by flash chromatography on a
silica gel column (25 · 5 cm). Elution with 2:1 hexanes–
ethyl acetate gave the product 17 as a light yellow solid:
yield 12.5 g (78%); mp 109–111 ꢁC, lit.42–44 mp 110–
111 ꢁC; silica gel TLC Rf 0.52 (3:1 hexanes–ethyl acetate);
1H NMR (CDCl3) d 2.62 (s, 3H), 5.13 (s, 2H), 6.46 (d, 1H,
J = 8.4 Hz), 6.59 (d, 1H, J = 8.4 Hz), 7.37 (m, 6H) and
13.29 (s, 1H); 13C NMR (CDCl3) d 34.31, 71.34, 102.45,
111.20, 128.18, 128.68, 128.98, 135.99, 136.30, 160.84,
164.94 and 205.39.
1
0.22 (3:1 hexanes–ethyl acetate); H NMR (CDCl3) d
5.27 (s, 2H), 6.71 (s, 1H), 6.81 (d, 1H, J = 8.4 Hz), 7.10
(d, 1H, J = 8.4 Hz), 7.23–7.51 (m, 7H), 7.60 (d, 2H,
J = 7.2 Hz) and 7.86 (m, 2H); 13C NMR (CDCl3) d
71.64, 76.88, 102.51, 111.60, 127.88, 128.74, 128.78,
128.84, 128.93, 129.12, 130.33, 135.28, 135.82, 136.32,
143.61, 160.51,165.76 and 194.72; mass spectrum (FAB),
m/z 329.1177 (M+H)+ (C22H17O3 requires 329.1178).
O
OH
OH
OBn O
OBn O
4.2.16. 5-O-Benzyl-40,7-dideoxykaempferol (20). To a
solution containing 0.35 g (1.1 mmol) of 19 in 20 mL of
CH2Cl2 at 0 ꢁC was added 20 mL of a 0.5–0.9 M solution
of DMDO in acetone. The reaction mixture was stirred at
4 ꢁC overnight, then concentrated under diminished pres-
sure. The residue was then dissolved in 20 mL of CHCl3
and cat. p-TsOH was added. The reaction mixture was
stirred at 0 ꢁC for 30 min and then concentrated under
diminished pressure. The residue was purified by flash
chromatography on a silica gel column (20 · 4 cm). Elu-
tion with 1:1 hexanes–ethyl acetate gave 20 as a light
brown solid: yield 0.29 g (78%); silica gel TLC Rf 0.46
(2:1 hexanes–ethyl acetate); 1H NMR (CDCl3) d 5.29 (s,
4.2.14. E-3-Phenyl-1-(2-benzoxy-6-hydroxyphenyl)prope-
45
none (18).
To
a
suspension containing 3.0 g
(12.3 mmol) of 17 and 1.31 g (1.26 mL, 12.4 mmol) of
benzaldehyde in 50 mL of MeOH was added 50 mL of
40% w/v KOH in MeOH. The reaction mixture was then
heated to reflux overnight, and the cooled reaction mix-
ture was diluted with 300 mL of water. The aqueous
layer was then extracted with three 100-mL portions
of ethyl acetate, washed with 75 mL of 1 N HCl, dried
(MgSO4), and concentrated under diminished pressure.
Purification by flash chromatography on a silica gel col-
umn (25 · 4 cm), elution with 4:1 hexanes–ethyl acetate,

J. A. Smith et al. / Bioorg. Med. Chem. 15 (2007) 5018–5034
5031
2H), 6.80 (d, 1H, J = 8.1 Hz), 7.13 (d, 1H, J = 8.4 Hz),
7.29–7.54 (m, 8H), 7.62 (d, 2H, J = 7.2 Hz) and 8.23 (d,
2H, J = 7.2 Hz); 13C NMR (CDCl3) d 71.04, 106.89,
110.84, 112.11, 126.92, 127.70, 128.07, 128.82, 128.92,
130.13, 131.26, 134.07, 136.59, 139.00, 142.66, 157.57,
158.62 and 173.05; mass spectrum (FAB), m/z 345.1126
(M+H)+ (C22H17O4 requires 345.1127).
purged three times with H2. The reaction vessel was then
maintained under a H2 atmosphere for 1 h, then filtered
through a Celiteꢂ pad and washed with ethyl acetate.
The filtrate was concentrated under diminished pressure
to afford a solid. The product 22 was pure enough so
that further purification was unnecessary: yield 0.08 g
(95%); silica gel TLC Rf 0.62 (1:1:0.1 hexanes–ethyl ace-
21
1
tate–methanol); ½aꢁ ꢂ143.2 (c 0.9, CHCl3); H NMR
D
(CDCl3) d 0.80 (d, 3H, J = 6.5 Hz), 1.98 (s, 3H), 2.08
(s, 3H), 3.13 (br s, 1H), 3.32 (m, 1H), 4.48 (s, 1H),
5.00 (t, 1H, J = 9.5 Hz), 5.22 (dd, 1H, J = 9.5 and
3.0 Hz), 5.52 (d, 1H, J = 1.0 Hz), 6.77 (d, 1H,
J = 8.0 Hz), 6.94 (d, 1H, J = 8.5 Hz), 5.55 (m, 4H),
7.90 (m, 2H) and 12.23 (s, 1H); 13C NMR (CDCl3) d
17.23, 20.96, 21.15, 68.44, 69.28, 71.03, 71.40, 101.18,
107.42, 111.10, 111.41, 128.87, 129.24, 130.36, 131.43,
135.73, 136.12, 155.83, 158.55, 160.86, 170.26, 170.37
and 179.50; mass spectrum (FAB), m/z 485.1450
(M+H)+ (C25H25O10 requires 485.1448).
O
O
OBn
O
OBn
OAc
O
OAc
4.2.17. 5-O-Benzyl-40,7-dideoxykaempferol 3-O-(200-O-
benzyl-300,400-di-O-acetyl-a-L-rhamnopyranoside) (21). To
a stirred suspension containing 0.10 g (0.29 mmol) of
OH
OBn
˚
20, 0.14 g (0.58 mmol) of Ag2O, and 4A molecular sieves
in 5 mL of CH2Cl2 was added 0.23 g (0.58 mmol) of 913
as a solution in 3 mL of CH2Cl2. The reaction mixture
was stirred at room temperature for 4 h, then diluted
with 20 mL of CH2Cl2 and filtered through a Celiteꢂ
pad. The filtrate was concentrated under diminished
pressure and purified by flash chromatography on a sil-
ica gel column (25 · 3 cm). Elution with 3:1 hexanes–
ethyl acetate gave 21 as a colorless foam: yield 0.17 g
(86%); silica gel TLC Rf 0.26 (2:1 hexanes–ethyl acetate);
OBn O
4.2.19. E-3-(4-Benzoxyphenyl)-1-(2-benzoxy-6-hydroxy-
phenyl)propenone (23). To a solution containing 2.00 g
(8.26 mmol) of 17 and 1.93 g (9.10 mmol) of 4-(benz-
oxy)benzaldehyde in 50 mL of MeOH was added
50 mL of 40% w/v KOH in MeOH. The reaction mix-
ture was heated to reflux overnight, cooled to room tem-
perature, and poured into 300 mL of H2O. The aqueous
layer was extracted with three 100-mL portions of ethyl
acetate. The combined organic layer was dried (MgSO4)
and concentrated under diminished pressure. The resi-
due was purified by flash chromatography on a silica
gel column (24 · 4 cm). Elution with 3:1 hexanes–ethyl
acetate gave 23 as a bright yellow solid: yield 2.70 g
(75%); mp 120–122 ꢁC; silica gel TLC Rf 0.39 (6:1 hex-
22
1
½aꢁ ꢂ89.5 (c 1.0, CHCl3); H NMR (CDCl3) d 0.80 (d,
D
3H, J = 6.3 Hz), 1.92 (s, 3H), 1.97 (s, 3H), 3.29 (m, 1H),
4.38 (m, 1H), 4.63 (d, 1H, J = 12.3 Hz), 4.78 (d, 1H,
J = 12.3 Hz), 5.08 (t, 1H, J = 9.9 Hz), 5.22 (dd, 1H,
J = 10.2 and 3.3 Hz), 5.30 (d, 2H), 5.70 (d, 1H,
J = 1.5 Hz), 6.80 (d, 1H, J = 8.1 Hz), 7.05 (d, 1H,
J = 8.7 Hz), 7.23–7.55 (m, 12H), 7.61 (d, 2H,
J = 7.5 Hz) and 7.88 (m, 2H); 13C NMR (CDCl3) d
17.24, 20.98, 21.02, 68.41, 70.69, 71.06, 71.13, 72.83,
75.63, 98.89, 108.13, 110.65, 115.45, 126.83, 127.86,
128.41, 128.44, 128.64, 128.73, 129.13, 130.72, 130.85,
133.83, 136.69, 138.11, 138.51, 155.03, 157.64, 158.91,
170.00, 170.30 and 174.08; mass spectrum (FAB), m/z
665.2389 (M+H)+ (C39H37O10 requires 665.2387).
1
anes–ethyl acetate); H NMR (CDCl3) d 5.08 (s, 4H),
6.51 (d, 1H, J = 8.1 Hz), 6.65 (d, 1H, J = 8.1 Hz), 6.79
(d, 2H, J = 8.7 Hz), 7.03 (d, 2H, J = 8.7 Hz), 7.33–7.51
(m, 11H), 7.77 (m, 2H), and 13.78 (s, 1H); 13C NMR
(CDCl3) d 70.22, 71.53, 102.40, 111.51, 111.91, 115.22,
125.57, 127.68, 128.27, 128.39, 128.70, 128.85, 128.89,
129.09, 130.53, 135.86, 136.02, 136.67, 143.61, 160.41,
160.67, 165.79 and 194.46; mass spectrum (FAB), m/z
437.1754 (M+H)+ (C29H25O4 requires 437.1753).
O
O
OBn
OH
O
OH
O
O
OAc
OAc
OBn
O
4.2.18. 40,7-Dideoxykaempferol 3-O-(300,400-di-O-acetyl-
a-L-rhamnopyranoside (40,7-dideoxy-SL0101)) (22). A
suspension containing 0.11 g (0.17 mmol) of 21 and
0.03 g of Pd(OH)2/C in 6 mL of 1:1 THF–MeOH was
4.2.20. 5-(Benzoxy)-2-(4-(benzoxy)phenyl)-4H-chromen-
4-one (24). To a solution containing 1.00 g (2.29 mmol)
of 23 in 20 mL of DMSO at 140 ꢁC was added 0.06 g

5032
J. A. Smith et al. / Bioorg. Med. Chem. 15 (2007) 5018–5034
(0.23 mmol) of I2. The reaction mixture was stirred at
140 ꢁC for 30 min, then cooled to room temperature
and poured into 200 mL of H2O. The aqueous layer
was extracted with three 75-mL portions of ethyl acetate.
The combined organic layer was washed with 200 mL of
brine, dried (MgSO4), and concentrated under dimin-
ished pressure. The residue was purified by flash chroma-
tography on a silica gel column (20 · 4 cm). Elution with
2:1 hexanes–ethyl acetate gave 24 as a colorless foam:
yield 0.76 g (76%); mp 145–147 ꢁC; silica gel TLC Rf
4.2.22. 7-Deoxy-40,5-di-O-benzylkaempferol 3-O-(200-O-
Benzyl-300,400-di-O-acetyl-a-L-rhamnopyranoside) (26). To
a stirred suspension containing 0.11 g (0.24 mmol) of
˚
25, 0.11 g (0.49 mmol) of Ag2O, and 4A molecular sieves
in 10 mL of CH2Cl2 was added 0.20 g (0.49 mmol) of 913
as a solution in 3 mL of CH2Cl2. The reaction mixture was
stirred at room temperature for 4 h, then diluted with
20 mL of CH2Cl2 and filtered through a Celiteꢂ pad.
The filtrate was concentrated under diminished pressure
and purified by flash chromatography on a silica gel col-
umn (20 · 3 cm). Elution with 3:1 hexanes–ethyl acetate
1
0.10 (3:1 hexanes–ethyl acetate); H NMR (CDCl3) d
5.12 (s, 2H), 5.28 (s, 2H), 6.81 (s, 1H), 6.82 (d, 1H,
J = 8.1 Hz), 7.06 (d, 2H, J = 8.7 Hz), 7.08 (d, 1H,
J = 8.1 Hz), 7.40 (m, 9H), 7.64 (d, 2H, J = 7.5 Hz) and
7.82 (d, 2H, J = 8.7 Hz); 13C NMR (CDCl3) d 70.32,
71.02, 107.98, 108.62, 110.62, 115.43, 124.13, 126.79,
127.67, 127.80, 127.96, 128.43, 128.76, 128.90, 133.61,
136.44, 136.85, 158.40, 158.67, 161.20, 161.49 and
178.24; mass spectrum (FAB), m/z 435.1598 (M+H)+
(C29H23O4 requires 435.1597).
gave 26 as a colorless foam: yield 0.13 g (67%); silica gel
21
TLC Rf0.17 (2:1 hexanes–ethyl acetate); ½aꢁ ꢂ86.5 (c
D
0.9, CHCl3); 1H NMR (CDCl3) d 0.87 (d, 3H,
J = 6.3 Hz), 1.95 (s, 3H), 1.99 (s, 3H), 3.44 (m, 1H), 4.39
(dd, 1H, J = 3.3 and 1.8 Hz), 4.70 (d, 1H, J = 12.3 Hz),
4.83 (d, 1H, J = 12.3 Hz), 5.10 (d, 1H, J = 9.9 Hz), 5.15
(s, 2H), 5.28 (dd, 1H, J = 10.2 and 3.3 Hz), 5.32 (s, 2H),
5.73 (d, 1H, J = 1.5 Hz), 6.81 (d, 1H, J = 8.1 Hz), 7.28–
7.48 (m, 17H), 7.64 (d, 2H, J = 7.2 Hz) and 7.86 (d, 2H,
J = 8.7 Hz); 13C NMR (CDCl3) d 17.35, 21.08, 68.47,
70.31, 70.75, 71.11, 71.31, 72.87, 75.76, 98.92, 108.15,
110.64, 114.98, 115.47, 123.37, 126.88, 127.57, 127.87,
128.41, 128.47, 128.77, 128.90, 130.79, 133.79, 136.53,
136.77, 138.12, 138.18, 154.91, 157.58, 158.92, 160.75,
170.04, 170.35 and 174.07; mass spectrum (FAB), m/z
771.2808 (M+H)+ (C46H43O11 requires 771.2805).
OBn
O
OH
OBn O
4.2.21. 7-Deoxy-40,5-di-O-benzylkaempferol (25). To a
solution containing 0.30 g (0.69 mmol) of 24 in
20 mL of CH2Cl2 at 0 ꢁC was added 20 mL of a
0.5–0.9 M solution of DMDO in acetone. The reac-
tion mixture was stirred at 4 ꢁC overnight and then
concentrated under diminished pressure. The residue
was then dissolved in 20 mL of CHCl3 and cat.
p-TsOH was added. The reaction mixture was stirred
at 0 ꢁC for 30 min and concentrated under diminished
pressure. The residue was purified by flash chromatog-
raphy on a silica gel column (26 · 3 cm). Elution with
3:1 hexanes–ethyl acetate gave 25 as a light brown so-
lid: yield 0.22 g (72%); silica gel TLC Rf 0.26 (3:1 hex-
OH
O
O
OH
O
O
OH
O
O
O
O
1
anes–ethyl acetate); H NMR (CDCl3) d 5.14 (s, 2H),
5.29 (s, 2H), 7.10 (d, 2H, J = 8.4 Hz), 7.35–7.52 (m,
11H), 7.65 (d, 2H, J = 7.5 Hz) and 8.20 (d, 2H,
4.2.23. 7-Deoxykaempferol 3-O-(300,400-di-O-acetyl-a-L-
rhamnopyranoside) (7-deoxy-SL0101) (27). A suspension
containing 0.30 g (0.39 mmol) of 26 and 0.07 g of
Pd(OH)2/C in 6 mL of 1:1 THF–MeOH was purged
three times with H2. The reaction vessel was then kept
under a H2 atmosphere for 1 h, filtered through a Celiteꢂ
pad, and washed with ethyl acetate. The filtrate was con-
centrated under diminished pressure to afford a solid.
Product 27 was pure and further purification was not
J = 8.4 Hz); 13C NMR (CDCl3)
d 70.25, 70.99,
106.86, 110.75, 112.10, 115.13, 123.91, 126.89,
127.71, 128.01, 128.36, 128.88, 129.40, 133.75,
136.64, 136.71, 138.18, 142.98, 157.37, 158.51, 160.18
and 172.74; mass spectrum (FAB), m/z 451.1544
(M+H)+ (C29H23O5 requires 451.1546).
necessary: yield 0.17 g (86%); silica gel TLC Rf 0.43
21
D
(1:1:0.1 hexanes–ethyl acetate–methanol); ½aꢁ ꢂ103.3
OBn
1
(c 0.8, MeOH); H NMR (acetone-d6) d 0.83 (d, 3H,
J = 6.0 Hz), 1.97 (s, 3H), 2.03 (s, 3H), 3.00 (br s, 1H),
3.55 (m, 1H), 4.46 (m, 1H), 4.82 (br s, 1H), 5.09 (t, 1H,
J = 9.6 Hz), 5.22 (d, 1H, J = 9.3 Hz), 5.59 (s, 1H), 6.77
(d, 1H, J = 7.8 Hz), 7.07 (d, 2H, J = 7.8 Hz), 7.65 (t,
1H, J = 7.8 Hz), 7.92 (d, 2H, J = 7.2 Hz), 9.26 (br s,
1H) and 12.48 (s, 1H); 13C NMR (acetone-d6) d 17.35,
21.05, 21.25, 68.52, 69.35, 71.32, 71.72, 100.99, 107.47,
111.10, 111.31, 116.07, 122.00, 131.17, 135.26, 135.63,
O
O
OBn O
O
OBn
O
O
O
O

J. A. Smith et al. / Bioorg. Med. Chem. 15 (2007) 5018–5034
5033
155.72, 158.87, 159.50, 160.77, 170.82, 171.09 and
179.43; mass spectrum (FAB), m/z 501.1395 (M+H)+
(C25H25O11 requires 501.1397).
polyclonal phosphospecific antibody developed against
the phosphopeptide, CGLASTND, and HRP-conju-
gated anti-rabbit antibody (211-035-109, Jackson
ImmunoResearch Laboratories, West Grove, Pennsyl-
vania) were used to detect serine phosphorylation of
the substrate. HRP activity was measured using Wes-
tern Lightning Chemiluminescence Reagent (NEL102,
PerkinElmer Life Sciences) according to the manufac-
turer’s protocol. Maximum and minimum activity is
the relative luminescence detected in the presence of
vehicle and 200 mM EDTA, respectively. His-tagged
active RSK was expressed in Sf9 cells and purified
using NiNTA resin (Qiagen, Valencia, California). Bac-
ulovirus was prepared using the Bac-to-Bacꢂ baculovi-
rus
expression
system
(Invitrogen,
Carlsbad,
California). Maximum responses and the concentra-
tions at half the inhibitory response (IC50) were deter-
mined by performing a best-fit analysis of the data
(GraphPad Prism).
4.2.24. 40,5,7-Tri-O-acetylkaempferol 3-O-(200,300,400-tri-
O-acetyl-a-L-rhamnopyranoside) (Peracetylated-SL0101)
(28). To a solution containing 0.038 g (0.068 mmol) of
kaempferol
3-O-(200,300,400-triacetoxy-a-L-rhamnopyr-
4.4. Cell culture
anoside) (29)18 in 4 mL of CH2Cl2 were added 0.035 g
(0.034 mmol, 0.03 mL) of acetic anhydride, 0.048 g
(0.48 mmol, 0.07 mL) of triethylamine, and cat. DMAP.
The reaction mixture was stirred at room temperature
under N2 for 4 h, then quenched with 15 mL of brine
and extracted with three 15-mL portions of ethyl ace-
tate. The combined organic layer was dried (MgSO4)
and concentrated under diminished pressure. Purifica-
tion by flash chromatography on a silica gel column
(2 · 30 cm) using 1:1 ethyl acetate–hexanes gave 28 as
For proliferation studies cells were seeded at 5000 cells
per well in 96-well tissue culture plates in the appropri-
ate medium as described by American Type Culture
Collection. After 24 h, the medium was replaced with
medium containing compound or vehicle as indicated.
Cell viability was measured 48 h later using CellTiter-
GloTM assay reagent (Promega, Madison, Wisconsin)
according to the manufacturer’s protocol. Maximum
responses and the concentrations of half the effective
response (EC50) were determined by performing a
best-fit analysis of the data (GraphPad Prism). For
specificity studies, cells were seeded at 3 · 105 cells/well
in 6-well plates. After 24 h, the cells were serum-starved
for 24 h and then incubated with compound or vehicle
for 4 h prior to a 20-min treatment with phorbol myr-
istate acetate. Cells were lysed with boiling SDS-sample
buffer without dithiothreitol (DTT). The lysates were
normalized for total protein, and DTT was added to
an aliquot, which was electrophoresed and immuno-
blotted. Antibodies used on cell lysates included anti-
pan-MAPK (anti-ERK1/2) (610124) from BD Trans-
duction Laboratories; anti-phospho-MAPK (V8031)
from Promega; anti-eEF2 (2332), anti-phospho-eEF2
(2331), anti-phospho-Akt Substrate (9611), anti-phos-
pho-PKA Substrate (9621), and anti-phospho-PKC
Substrate (2261) from Cell Signaling Technology.
Anti-Ran was a generous gift from Ian Macara (Uni-
versity of Virginia).
1
a colorless oil: yield 0.044 g (95%); H NMR (CDCl3)
d 0.85 (d, 3H, J = 6.3 Hz), 1.96 (s, 3H), 1.97 (s, 3H),
2.11 (s, 3H), 2.31 (s, 3H), 2.33 (s, 3H), 2.43 (s, 3H),
3.24 (m, 1H), 4.91 (t, 1H, J = 9.9 Hz), 5.25 (dd, 1H,
J = 10.2 and 3.3 Hz), 5.56 (d, 1H, J = 1.5 Hz), 5.63 (m,
1H), 6.82 (d, 1H, J = 2.4 Hz), 7.27 (m, 3H) and 7.89
(d, 2H, J = 8.7 Hz); 13C NMR (CDCl3) d 16.95, 20.62,
20.69, 20.85, 21.05, 21.08, 21.13, 68.37, 68.77, 69.04,
70.23, 98.01, 108.92, 113.56, 115.11, 122.17, 127.42,
130.13, 136.94, 150.32, 152.70, 153.90, 155.25, 156.64,
156.68, 167.96, 168.65, 169.37, 169.60, 169.91, 169.98
and 172.21; mass spectrum (FAB), m/z 685.1766
(M+H)+ (C33H33O16 requires 685.1769).
4.3. Kinase assays
Glutathione-S-transferase (GST)-fusion protein (1 lg)
containing the sequence—RRRLASTNDKG (for ser-
ine/threonine kinase assays) was adsorbed in the wells
of LumiNunc 96-well polystyrene plates (MaxiSorp
surface treatment). The wells were blocked with sterile
3% tryptone in phosphate-buffered saline. Kinase
(5 nM) in 70 lL of kinase buffer (5 mM b-glycerophos-
phate, pH 7.4, 25 mM HEPES, pH 7.4, 1.5 mM DTT,
30 mM MgCl2, 0.15 M NaCl) was dispensed into each
well. Compound at the indicated concentrations or
vehicle was added and reactions were initiated by the
addition of 30 lL of ATP to a final ATP concentration
of 10 lM. Reactions were terminated after 30 min by
addition of 75 lL of 500 mM EDTA, pH 7.5. All as-
says measured the initial velocity of reaction. After
extensive washing of wells, anti-p-p140 antibody, a
Acknowledgment
This work was supported by the Department of Defense
#DAMD17-03-1-0366 USAMRMC. We thank Ian
Macara for a sample of anti-Ran.
References and notes
1. Roux, P. P.; Blenis, J. Microbiol. Mol. Biol. Rev. 2004, 68,
320–344.

5034
J. A. Smith et al. / Bioorg. Med. Chem. 15 (2007) 5018–5034
2. Smith, J. A.; Poteet-Smith, C. E.; Xu, Y.; Errington, T.
M.; Hecht, S. M.; Lannigan, D. A. Cancer Res. 2005, 65,
1027–1034.
24. Zhu, X.; Castellani, R. J.; Takeda, A.; Nunomura, A.;
Atwood, C. S.; Perry, G.; Smith, M. A. Mech. Ageing Dev.
2001, 123, 39–46.
3. Clark, D. E.; Errington, T. M.; Smith, J. A.; Frierson, H.
F., Jr.; Weber, M. J.; Lannigan, D. A. Cancer Res. 2005,
65, 3108–3116.
25. Valjent, E.; Corvol, J. C.; Pages, C.; Besson, M. J.;
Maldonado, R.; Caboche, J. J. Neurosci. 2000, 20, 8701–
8709.
4. Hurbin, A.; Coll, J. L.; Dubrez-Daloz, L.; Mari, B.;
Auberger, P.; Brambilla, C.; Favrot, M. C. J. Biol. Chem.
2005, 280, 19757–19767.
5. David, J. P.; Mehic, D.; Bakiri, L.; Schilling, A. F.;
Mandic, V.; Priemel, M.; Idarraga, M. H.; Reschke, M.
O.; Hoffmann, O.; Amling, M.; Wagner, E. F. J. Clin.
Invest. 2005, 115, 664–672.
6. Hayashi, M.; Fearns, C.; Eliceiri, B.; Yang, Y.; Lee, J. D.
Cancer Res. 2005, 65, 7699–7706.
7. Xu, Y.-M.; Smith, J. A.; Lannigan, D. A.; Hecht, S. M.
Bioorg. Med. Chem. 2006, 14, 3974–3977.
8. Vieth, M.; Sutherland, J. J.; Robertson, D. H.; Campbell,
R. M. Drug Discov. Today 2005, 10, 839–846.
9. Noble, M. E.; Endicott, J. A.; Johnson, L. N. Science
2004, 303, 1800–1805.
10. Nguyen, T. L.; Gussio, R.; Smith, J. A.; Lannigan, D. A.;
Hecht, S. M.; Scudiero, D. A.; Shoemaker, R. H.;
Zaharevitz, D. W. Bioorg. Med. Chem. 2006, 14, 6097–
6105.
11. Prade, L.; Engh, R. A.; Girod, A.; Kinzel, V.; Huber, R.;
Bossemeyer, D. Structure 1997, 5, 1627–1637.
12. Xu, Z. B.; Chaudhary, D.; Olland, S.; Wolfrom, S.;
Czerwinski, R.; Malakian, K.; Lin, L.; Stahl, M. L.;
Joseph-McCarthy, D.; Benander, C.; Fitz, L.; Greco, R.;
Somers, W. S.; Mosyak, L. J. Biol. Chem. 2004, 279,
50401–50409.
26. Ji, R. R.; Befort, K.; Brenner, G. J.; Woolf, C. J.
J. Neurosci. 2002, 22, 478–485.
27. Kulich, S. M.; Chu, C. T. J. Neurochem. 2001, 77, 1058–
1066.
28. Gomez-Santos, C.; Ferrer, I.; Reiriz, J.; Vinals, F.;
Barrachina, M.; Ambrosio, S. Brain Res. 2002, 935, 32–39.
29. Hu, Y.; Dietrich, H.; Metzler, B.; Wick, G.; Xu, Q.
Arterioscler. Thromb. Vasc. Biol. 2000, 20, 18–26.
30. Bokemeyer, D.; Panek, D.; Kramer, H. J.; Lindemann,
M.; Kitahara, M.; Boor, P.; Kerjaschki, D.; Trzaskos, J.
M.; Floege, J.; Ostendorf, T. J. Am. Soc. Nephrol. 2002,
13, 1473–1480.
31. Pahl, A.; Zhang, M.; Kuss, H.; Szelenyi, I.; Brune, K. Br.
J. Pharmacol. 2002, 135, 1915–1926.
32. Slomiany, B. L.; Slomiany, A. Biochem. Biophys. Res.
Commun. 2002, 294, 220–224.
33. Opavsky, M. A.; Martino, T.; Rabinovitch, M.; Pennin-
ger, J.; Richardson, C.; Petric, M.; Trinidad, C.; Butcher,
L.; Chan, J.; Liu, P. P. J. Clin. Invest. 2002, 109, 1561–
1569.
34. Planz, O.; Pleschka, S.; Ludwig, S. J. Virol. 2001, 75,
4871–4877.
35. Dineley, K. T.; Westerman, M.; Bui, D.; Bell, K.; Ashe, K.
H.; Sweatt, J. D. J. Neurosci. 2001, 21, 4125–4133.
36. Delaunoy, J. P.; Dubos, A.; Marques Pereira, P.; Han-
auer, A. Clin. Genet. 2006, 70, 161–166.
13. Maloney, D. J.; Hecht, S. M. Org. Lett. 2005, 7, 1097–
1099.
37. Dufresne, S. D.; Bjorbaek, C.; El-Haschimi, K.; Xhao, Y.;
Aschenbach, W. G.; Moller, D. E.; Goodyear, L. J. Mol.
Cell. Biol. 2001, 8, 81–87.
38. Kim, M.; Lee, J. H.; Koh, H.; Lee, S. Y.; Jang, C.; Chung,
C. J.; Sung, J. H.; Blenis, J.; Chung, J. EMBO J. 2006, 25,
3056–3067.
14. Algar, J.; Flynn, J. P. Proc. R. Irish Acad. 1934, 42b, 1–8.
15. Oyamada, T. Nippon Kagaku Kaishi 1934, 55, 1256–1261.
16. Oyamada, T. Bull. Chem. Soc. Jpn. 1935, 10, 182–186.
17. De Meyer, N.; Haemers, A.; Mishra, L.; Pandey, H. K.;
Pieters, L. A. C.; Vanden Berghe, D. A.; Vlietinck, A.
J. Med. Chem. 1992, 34, 736–746.
39. Geissman, T. A.; Clinton, R. O. J. Am. Chem. Soc. 1946,
68, 697–700.
18. Smith, J. A.; Maloney, D. J.; Clark, D. E.; Xu, Y.; Hecht,
S. M.; Lannigan, D. A. Bioorg. Med. Chem. 2006, 14,
6034–6042.
19. Sapkota, G. P.; Newell, F. S.; Armstrong, C.; Bain, J.;
Frodin, M.; Grauert, M.; Hoffmann, M.; Schnapp, G.;
Steegmaier, M.; Cohen, P.; Alessi, D. R. Biochem. J. 2007,
401, 29–38.
20. Cohen, M. S.; Zhang, C.; Shokat, K. M.; Taunton, J.
Science 2005, 308, 1318–1321.
21. Takeishi, Y.; Huang, Q.; Abe, J.; Che, W.; Lee, J. D.;
Kawakatsu, H.; Hoit, B. D.; Berk, B. C.; Walsh, R. A.
Cardiovasc. Res. 2002, 53, 131–137.
40. Machida, K.; Osawa, K. Chem. Pharm. Bull. 1989, 37,
1092–1094.
41. Jain, N.; Alam, M. S.; Kamil, M.; Ilyas, M.; Niwa, M.;
Sakae, A. Phytochemistry 1990, 29, 3988–3991.
42. Murata, T.; Shimada, M.; Sakakibera, S.; Yoshino, T.;
Masuda, T.; Shintani, T.; Sato, H.; Koriyama, Y.;
Fukushima, K.; Nunami, N.; Yamauchi, M. Bioorg.
Med. Chem. Lett. 2004, 14, 4019–4022.
43. Pinto, D. G.; Silva, A. M. S.; Cavaleiro, J. A.
J. Heterocycl. Chem. 2000, 37, 1629–1634.
44. Patil, A. D.; Deshpande, V. H. Indian J. Chem. Sect. B
1983, 22, 109–113.
22. McDonald, C.; Vacratsis, P. O.; Bliska, J. B.; Dixon, J. E.
J. Biol. Chem. 2003, 278, 18514–18523.
45. Pinto, D. G.; Silva, A. M. S.; Cavaleiro, J. A.
J. Heterocycl. Chem. 1996, 1887–1893.
23. Lu, H.; Guizzetti, M.; Costa, L. G. J. Pharmacol. Exp.
Ther. 2002, 300, 818–823.
46. Pinto, D. G.; Silva, A. M. S.; Cavaleiro, J. A. Tetrahedron
Lett. 1994, 35, 5899–5902.