A new structural class of S-adenosylhomocysteine hydrolase inhibitors

Article abstract of DOI:10.1016/j.bmc.2009.07.061

Effective inhibitors of S-adenosylhomocysteine hydrolase hold promise towards becoming useful therapeutic agents. Since most efforts have focused on the development of nucleoside analog inhibitors, issues regarding bioavailability and selectivity have bee

Full text of DOI:10.1016/j.bmc.2009.07.061

Bioorganic & Medicinal Chemistry 17 (2009) 6707–6714
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
A new structural class of S-adenosylhomocysteine hydrolase inhibitors
Byung Gyu Kim a, Tae Gyu Chun b, Hee-Yoon Lee b, Marc L. Snapper a,
*
a Merkert Chemistry Center, Department of Chemistry, Boston College, Chestnut Hill, MA 02467, USA
b Center for Molecular Design & Synthesis, Department of Chemistry, Korea Advanced Institute of Science & Technology, Daejon 305-701, Republic of Korea
a r t i c l e i n f o
a b s t r a c t
Article history:
Effective inhibitors of S-adenosylhomocysteine hydrolase hold promise towards becoming useful thera-
peutic agents. Since most efforts have focused on the development of nucleoside analog inhibitors, issues
regarding bioavailability and selectivity have been major challenges. Considering the marine sponge
metabolite ilimaquinone was found to be a competitive inhibitor of S-adenosylhomocysteine hydrolase,
new opportunities for developing selective new inhibitors of this enzyme have become available. Based
on the activities of various hybrid analogs, SAR studies, pharmacophore modeling, and computer docking
studies have lead to a predictive understanding of ilimaquinone’s S-adenosylhomocysteine hydrolase
inhibitory activities. These studies have allowed for the design and preparation of simplified structural
variants possessing new furanoside bioisosteres with 100-fold greater inhibitory activities than that of
the natural product.
Received 17 July 2008
Revised 21 July 2009
Accepted 23 July 2009
Available online 28 July 2009
Keywords:
Ilimaquinone
S-Adenosylhomocysteine hydrolase
Inhibitor
Pharmacophore modeling
Ó 2009 Published by Elsevier Ltd.
1. Introduction
versible fashion, or are insufficiently selective, either of which
can lead to undesired consequences and represent drawbacks for
S-Adenosylhomocysteine hydrolase (AdoHcy hydrolase; EC
3.3.1.1),1 which reversibly hydrolyzes S-adenosylhomocysteine
(AdoHcy) to homocysteine and adenosine (Eq. 1),2 has been con-
sidered an attractive target for the design of antiviral, antipara-
sitic, antiarthritic, and immunosuppressive agents.3 Recent
elucidation of the crystal structure of AdoHcy hydrolase isoforms
from human and rat has contributed to our understanding of this
enzyme.4 Inhibition of AdoHcy hydrolase results in the accumula-
tion of AdoHcy, which leads to feedback inhibition of S-adenosyl-
methionine (AdoMet) dependent methylations, a process that can
impede viral replication,5 and can lower serum cholesterol lev-
els.6 Furthermore, the corresponding reduction of homocysteine
levels is believed to reduce the risk of developing coronary artery
disease.5
their development into clinically useful drugs.7 Moreover, amongst
the common nucleoside analogs, in addition to issues regarding
potency and selectivity, there are also concerns of bioavailability.
Eritadenine, for example, is a potent inhibitor of AdoHcy hydrolase,
however, it is rapidly cleared by the liver. As a result, there is need
for reversible selective inhibitors of AdoHcy hydrolase with im-
proved bioavailability.
NH2
N
NH2
N
NH2
N
N
N
N
N
N
N
N
N
N
HO
O
O
HO
OH
HO
HO
OH
HO
OH
H
H
NH2
N
NH2
N
eritadenine
noraristeromycin
neplanocin A
IC50 (0.03 µM)
Schanche, J.S.; Schanche, T.;
Ueland, P. M.; Holy, A.; Votruba ,
I. Mol Pharmacol 1984, 26, 553.
IC50 (1.1 µM)
Steere, Honek Bioorg. Med.
Chem. Lett. 2003, 13, 3963
IC50 (0.9 µM)
Kim, et al. Bioorg. Mec.
Chem. Lett. 2004, 14, 2091.
N
N
N
N
O2C
HO
+
N
N
NH3
S
O
O
AdoHcy
hydrolase
O2C
NH3 SH
HO
HO
OH
OH
S-adenosylhomocysteine
(AdoHcy)
homocysteine
adenosine
Our interest in this enzyme arose from the intriguing biological
activities of (À)-ilimaquinone,8 a natural product that was first iso-
lated in 1979 from the Red Sea sponge Smenospongia sp. and over
the years has been reported to possess antiviral, antimitotic, anti-
microbial, and anti-HIV activities.9 In addition, this marine sponge
metabolite has also been shown to vesiculate the Golgi appara-
tus,10 as well as to protect cells against the toxic effects of ricin
and diphtheria.11 In light of these myriad biological effects, we
Numerous inhibitors of AdoHcy hydrolase have been identified
from natural and synthetic sources, but many function in an irre-
* Corresponding author. Tel.: +1 617 552 8096; fax: +1 617 552 1442.
E-mail address: marc.snapper@BC.edu (M.L. Snapper).
0968-0896/$ - see front matter Ó 2009 Published by Elsevier Ltd.
doi:10.1016/j.bmc.2009.07.061

6708
B. G. Kim et al. / Bioorg. Med. Chem. 17 (2009) 6707–6714
sought to use the natural product and structural variants to better
understand the molecular basis of its activities.
could facilitate the development of new and non-toxic inhibitors
of this enzyme. Our preliminary efforts toward this objective are
described herein.
Me
2. Results and discussion
O
OMe
O
H
Given ilimaquinone is a competitive inhibitor of AdoHcy hydro-
lase, and therefore competes with the nucleoside derived AdoHcy
for binding to the enzyme’s active site, perhaps the natural product
is also functioning in a non-obvious manner as a nucleoside analog.
Along these lines, we imagined several possibilities.16 One scenario
is the quinone moiety of the natural product, which is likely depro-
tonated under physiological conditions and therefore rendered
non-electrophilic, is substituting for the adenine region of AdoHcy.
This idea stemmed from the existence of natural products that pos-
sess adenine-like groups attached to a decalin ring system. Natural
products such as kolavenol,17 C-4–14,18 asmarine A, and asmarine
B,19 all possess adenine-functionality in place of ilimaquinone’s
quinone group. An overlay of these two ring systems is shown in
Figure 1 (Model 1). Unfortunately, our early evaluation of C-4–14
in a cellular secretion assay did not show much promise for these
naturally occurring structural variants.13
Me
Me
HO
(-)-ilimaquinone
IC50 (40 µM)
Modifications to the total synthesis of (À)-ilimaquinone12 al-
lowed for the preparation, screening, and identification of active
variants that could be used to report on the cellular interactions
of the natural product.13 Whole cell and cytosolic photoaffinity
labeling and affinity chromatography studies with these reagents
converged on AdoHcy hydrolase as a cellular target of the natural
product.14 Subsequent studies with isolated enzyme indicated that
ilimaquinone is a competitive inhibitor of AdoHcy hydrolase.
Moreover, new methylations, particularly amongst membrane-
bound targets, were shown to be significantly reduced upon cellu-
lar incubation with ilimaquinone.14a This effect is consistent with
inhibition of AdoHcy hydrolase elevating the concentration of
AdoHcy, which leads to feedback inhibition of methyltransferases.
Likewise, it was shown that the addition of AdoMet rescues the cell
from the effects of ilimaquinone. In this case, the exogenous
AdoMet serves to overcome AdoHcy’s inhibition of the methyl-
transferases.15
H2N
H2N
N
N
N
N
N
N
N
N
N
N
N
N
N
Cl
N
Cl
Me
H
Me
H
N
OH
N
OH
N
N
Me
Me
Me
Me
Me
Me
Me
Me
Me
H
Me
Me
Me
Given these observations, ilimaquinone’s inhibition of AdoHcy
hydrolase and indirect inhibition of cellular methylations appears
to account for many of ilimaquinone’s biological activities. More-
over, from a structural perspective, (À)-ilimaquinone represents
a unique inhibitor of AdoHcy hydrolase. Considering the biomedi-
cal potential for novel, non-toxic inhibitors of AdoHcy hydrolase in
antiviral, antiparasitic, and antiarthritic therapies, a better under-
standing of (À)-ilimaquinone’s inhibition of AdoHcy hydrolase
H
Me
Me
Me
Me Me
Me
C-4-14
kolavenol
asmarine A
asmarine B
Alternatively, it is also possible, especially considering the oxi-
dized intermediates formed in AdoHcy hydrolase’s enzymatic cy-
cle, that the quinone group could serve as a replacement for the
Figure 1. Pharmacophore modeling of (À)-ilimaquinone.

B. G. Kim et al. / Bioorg. Med. Chem. 17 (2009) 6707–6714
6709
AdoHcy hydrolase inhibiton assay
100
90
80
70
60
50
40
30
20
10
0
Graph 1. Comparison (À)-ilimaquinone (IQ) and analogs (50
lM) in a AdoHcy hydrolase inhibition assay.
ribose portion of AdoHcy. A similar situation was reported for the
inhibition of Hsp90 by geldanamycin. Geldanamycin possesses a
quinone moiety that binds to the active site of Hsp90 where the ri-
bose portion of ATP binds.20 This possibility for ilimaquinone is
illustrated in Model 2. In either case, both pharmacophore models
serve as inspiration for the design of new AdoHcy hydrolase
inhibitors.
To understand ilimaquinone’s interaction with AdoHcy hydro-
lase while identifying structural variants with improved inhibitory
activities, we introduced simple modifications to the quinone moi-
ety as suggested by the adenine-functionality in Model 1. We
hoped that the inclusion of nitrogens (and other heteroatom-con-
taining substituents) on the quinone would favorably influence
binding with AdoHcy hydrolase. Synthesis of these variants started
from the known chloroquinone 1.12,13 Replacement of chloride
with fluoride and ethanethiol gave thioquinone 2a and fluoroqui-
none 2b, respectively. Treatment of 1 with methylamine and so-
dium bicarbonate afforded 2c under short reaction times, but
produced diaminated compound 2d with longer reactions at ele-
vated temperatures. Similarly, ethylamine, allylamine, i-propyl-
amine, n-butylamine, and t-butylamine were installed on the
quinone to provide analogs 2f–2k.
illustrated in Figure 2. Given our pharmacophore modeling, we
were particularly interested in evaluating the relative activities of
hybrids 3 and 4. Coupling of alternative fragments such as in 5
and 6 were also of relevance. For example, hybrid analog 5 is rem-
iniscent of the known adenosinyl containing natural products
mentioned above.
In any case, we prepared the ilimaquinone/adenosine hybrid
structures 3 and 4 where the quinone moiety was substituted for
either the adenine or the ribose portion of adenosine. As summa-
rized in Scheme 2, coupling of adenine with 2-chloro-3,5-dime-
thoxybenzyl bromide followed by Boc protection afforded 7 in
48% yield, and subsequent CAN oxidation gave the quinone 8 in
79% (57% conversion). Deprotection of 8 using formic acid provided
hybrid 4 in 83% yield. Synthesis of compound 3, the quinone/ribose
hybrid started from b-ribose-furanose 1-acetate 2,3,5-tribenzoate,
which in the presence of a Lewis acid was coupled with 1,2,4,5-tet-
ramethoxybenzene to give intermediate 9. Deprotection of 9 with
sodium methoxide and subsequent CAN oxidation afforded 3 in
25% yield for these final two steps.
As before, compounds 3 and 4 were evaluated at 50
lM for inhi-
bition of AdoHcy hydrolase. Their performance in this assay is also
The effects of adding other heteroatom substituents on the qui-
none moiety of ilimaquinone in an AdoHcy hydrolase inhibition as-
say were promising. As illustrated in Graph 1, with the exception of
fluoroquinone 2b, all other quinone variants (2a, 2c–2k) displayed
increased AdoHcy hydrolase inhibitory activities relative to ilim-
aquinone. While encouraged by these favorable results, their abil-
ity to provide insight into active site interactions and differentiate
between binding models was limited.
MeO
O
MeO
O
MeO
O
O
O
O
SEt
Cl
F
a
b
Me
Me
Me
H
H
H
Me
Me
Me
Me
Me
Me
1
2a
2b
d
A more telling approach, we believed, would be to build hybrid
structures that paired portions of ilimaquinone with components
of typical adenosyl inhibitors.21 Four such compounds (3–6) are
c
MeO
O
RHN
RHN
O
O
H
O
O
O
NHt-Bu
Cl
NHR
NH2
NH2
Me
Me
Me
H
H
MeO
Me
N
Me
Me
Me
2d (R = Me)
2f (R = allyl)
2g (R = H)
N
N
N
N
O
O
H
Me
N
N
O
O
X
N
O
Me
Me
Me
Me
2j
2h (R = Et)
2i (R = i-Pr)
2k (R = n-butyl)
X
MeO
X
H
O
Me
Me
2c (R = Me)
2e (R = allyl)
HO OH
X
HO OH
Me
3
4
Scheme 1. Preparation of quinone analogs of (À)-ilimaquinone. (a) EtSH, NaHCO3,
MeOH, 84%; (b) KF (excess), Bu4NBr (cat.), 47%; (c) RNH2, NaHCO3, MeOH, rt,
10 min, 87%; (d) RNH2, NaHCO3, MeOH, 12 h, 40 °C, 81%.
5
6
Figure 2. (À)-Ilimaquinone/adenosine hybrid analogs and natural products.

6710
B. G. Kim et al. / Bioorg. Med. Chem. 17 (2009) 6707–6714
NBoc2
NBoc2
NH2
NBoc2
Hybrid 10
NH2
N
N
N
N
N
N
N
N
N
NH2
N
N
N
OH
N
RO
N
HO
CO2Me
N
N
N
N
a
c
N
b
N
O
N
O
N
N
a
b-d
N
H
O
O
MeO
Cl MeO
Cl MeO
Cl
RO
O
O
RO
EtO
EtO
OMe
MeO
O
O
7
8
O
O
4
OH
HO
12
13a-e
14a-e
10a-e
MeO
OEt
MeO
OMe
O
O
Hybrid 11
NBoc2
NH2
N
BzO
OAc
BzO
HO
O
O
O
N
N
N
d
e
N
N
OMe
OMe
HO
CO2Me RO
e,f
CO2Me
g,h
N
N
i
BzO OBz
BzO OBz
HO OH
9
3
CHO
OBn
Scheme 2. Preparation of hybrids 3 and 4. (a) NaH, DMF, then 2-chloro-3,5-
dimethoxybenzyl bromide; Boc2O, DMAP, 48% (two steps); (b) CAN, CH3CN/H2O,
79% (57% conv.); (c) formic acid, 83%; (d) AlCl3, 1,2,4,5-tetramethoxybenzene, 30%;
(e) NaOMe, MeOH; CAN, CH3CN/H2O, 25% (two steps).
OH
OBn
RO
OBn RO
OH
OBn
OH
11a-e
(11a R = H)
15a-e
16a-e
R = (a) benzyl, (b) pentyl, (c) cyclohexylmethyl,
(d) cyclobutylmethyl, (e) cyclopentylmethyl
summarized in Graph 1. Interestingly, hybrid 3 showed little inhib-
itory activity against the enzyme, whereas compound 4 displayed
high activity at this concentration. Further evaluation indicated
Scheme 3. Synthesis simplified aromatic hybrid analogs. (a) RBr, K2CO3, acetone;
LAH, THF; (b) SOCl2, pyridine, benzene, 70–80%; (c) bis-Boc-adenine, Bu4NI, K2CO3,
DMF; (d) HCO2H; (e) RX, Cs2CO3, acetone; (f) H2O2, NaOH, 1,4-dioxane; BnBr,
Cs2CO3, Bu4NI, acetone; (g) LAH, THF; (h) SOCl2, pyridine, benzene; bis-Boc-adenine,
K2CO3, DMF, Bu4NI; (i) HCO2H; H2, Pd/C (10% on carbon), EtOAc/MeOH, 1 atm.
that hybrid 4 possesses an IC50 of 4.3 lM ( 1.2 lM) against AdoHcy
hydrolase, a 10-fold improvement over (À)-ilimaquinone’s activity.
These results provide support for binding Model 2 (Fig. 1), where
ilimaquinone’s quinone moiety may interact with AdoHcy hydro-
lase in a fashion similar to the ribose 2,3-dihydroxyl group of
AdoHcy.22
Based on this binding model, we wondered whether the qui-
none ring could be simplified even further. For example, perhaps
the quinone could be replaced by an aromatic ring. This system
should still provide the same key hydrogen-bonding interactions
with the enzyme target, while also interfering with the crucial re-
dox chemistry that is indispensable for AdoHcy hydrolase activity.
Along these lines, two simplified adenine/polyhydroxylated arene
hybrids 10 and 11 were designed and prepared (Fig. 3).
was followed by a Bayer–Villager reaction25 and protection of the
resulting diols with benzyl groups to afford compounds 15a–e in
approx. 60–65% yield (three steps). Reduction of ester 15a–e with
LAH and subsequent chlorination of the resulting benzyl alcohol
using thionyl chloride and pyridine provided benzyl chloride,
which was directly coupled with bis-Boc-adenine to provide
Table 1
Evaluation of compounds 10a–e as AdoHcy hydrolase inhibitors
% Inhibition
100
l
M
50
lM
10
l
M
2
lM
1
lM
0.1 lM
As described in Scheme 3, the synthesis of hybrid 10 com-
menced with orthoformate protection of commercially available
methyl 3,4,5-trihydroxybenzoate to afford compound 12 in 83%
yield.23 Coupling of 12 with various aliphatic R groups produced
ethers, which were converted to the corresponding benzyl alcohols
13a–e by reduction with LAH in 85–90% yield (two steps). At-
tempts to prepare the corresponding benzyl bromides resulted in
unsatisfactory yields (ca. 30%), presumably due to instability of
these reactive alkylating agents. Reaction of 13a–e with thionyl
chloride and pyridine in dichloromethane gave the requisite benzyl
chlorides in highly improved yields, which were directly coupled
with bis-Boc-adenine21 to provide analogs 14a–e in 71–80% yield
for these two steps. Deprotection of 14a–e with formic acid fur-
nished the desired hybrids 10a–e in greater than 90% yields.
The preparation of the hybrid 11 analogs is also described in
Scheme 3. Selective alkylation of methyl 3,5-dihydroxy-2-form-
ylbenzoate24 with various alkyl halides under the basic condition
10a
10b
10c
10d
10e
93 ( 7)
72 ( 4)
100 ( 1)
92 ( 1)
89 ( 2)
70 ( 2)
47 ( 8)
74 ( 4)
nd
53 ( 2)
26 ( 1)
64 ( 6)
72 ( 1)
62 ( 2)
26 ( 5)
12 ( 2)
17 ( 1)
nd
27 ( 2)
nd
15 ( 4)
29 ( 4)
31 ( 1)
nd
nd
nd
7 ( 1)
6 ( 1)
nd
nd
Table 2
Evaluation of compounds 11a–e as AdoHcy hydrolase inhibitors
% Inhibition
100
l
M
10
l
M
1
l
M
0.1
l
M
0.01
lM
11a
11b
11c
11d
11e
89 ( 3)
95 ( 2)
98 ( 2)
95 ( 1)
104 ( 4)
85 ( 3)
78 ( 4)
91 ( 2)
86 ( 2)
97 ( 1)
67 ( 14)
48 ( 11)
69 ( 4)
62 ( 12)
74 ( 3)
31 ( 12)
32 ( 13)
37 ( 15)
22 ( 2)
10 ( 4)
13 ( 10)
4 ( 6)
6 ( 1)
20 ( 1)
32 ( 1)
NH2
N
NH2
N
NH2
N
NH2
N
N
N
N
N
N
N
N
N
X
O
O
N
MeO
N
N
N
O
RO
HO
RO
or
HO OH
Cl
OH
10
11
OH
OH
4
AdoHcy
(X = homocysteine)
Figure 3. Simplified (À)-ilimaquinone/adenosine hybrids.

B. G. Kim et al. / Bioorg. Med. Chem. 17 (2009) 6707–6714
6711
aliphatic
region
ribose
region
base
O
X
O
MeO
O
MeO
O
region
N
X
NH2
(X = Cl, OH)
Me
Me
S
-adenosylhomocysteine
O2C
N
S
H
H
N
(AdoHcy)
Me
Me
O
N
O
NH3
3
4
BocHN
O
BocHN
O
HO
OH
Me
Me
HO
O
O
17
(> 250 µM IC50
18
(-)-ilimaquinone
IC50 (40 µM)
H
)
(< 20 µM IC50)
Me
Me
Me
O
O
OR
N
NH2
11c
IC50 (0.4 µM)
N
N
N
OH
HO
Figure 4. Refinement of the pharmocophore model.
16a–e in 50–60% yield for the three steps. Successive Boc deprotec-
tion of 16a–e with formic acid and hydrogenolysis of the resulting
products furnished simplified adenine hybrids 11a–e in approxi-
mately 50% yield.
The inhibitory activities of compounds 10a–e and 11a–e against
AdoHcy hydrolase were investigated and an approximate IC50
value for each compound was obtained. Gratifyingly, as shown in
Tables 1 and 2, most of these simplified analogs displayed at low
lM range IC50. In general, the hybrid 11 series were more potent
inhibitors of AdoHcy hydrolase than the 10 series. Moreover, some
of the hybrid 11 compounds possessed submicromolar IC50 activi-
ties, levels indicating 100-fold more potent inhibitory activities
than that of (À)-ilimaquinone.
These favorable results encouraged further refinement of our
pharmacophore model. Figure 4 illustrates the three key sub-
strate/enzyme interacting regions and how they may correlate to
ilimaquinone and simplified variants. Based on this model and crys-
tallographic information of known inhibitors bound to AdoHcy
hydrolase, computational studies of our simplified inhibitors
docked with the enzyme were pursued. First, ilimaquinone was
docked using the Discovery/Insight II into the catalytic domain of
the enzyme with the quinone moiety positioned where the ribose
unit of AdoHcy was located in the X-ray structure.26 The hydrogens
on the enzyme’s residues and inhibitors were allowed to relax dur-
ing energy minimization steps. As implied in Figure 4, and shown in
Figure 5a and b, a binding comparison of ilimaquinone and AdoHcy
indicates that the decalin portion of ilimaquinone can replace the
homocysteine region of AdoHcy. Moreover, the ketomethoxy func-
tionality of the quinone can effectively replace the ribose unit of
AdoHcy with good overlap of the hydroxyl groups of the ribose.
Of particular note is the ability of this binding model to account
for the inhibitory activities of previously examined ilimaquinone
variants. For example, the functionalized reagents 17 and 18 were
prepared and screened in a secretory assay designed to probe Golgi
function.13 Interestingly, analog 17, which was generated from the
natural product, showed no antisecretory activities at concentra-
tions tested, while the synthetic analog 18 displayed activities
comparable to the natural product. This notable difference in activ-
ity indicated that functional affinity and fluorescent probe reagents
should be prepared from analog 18. Now our AdoHcy hydrolase
binding model can account for the activity differences in these
two ilimaquinone-derived reagents.
As illustrated in Figure 5c, there is a narrow channel that ex-
tends from the enzyme’s active site that can accommodate the
C4 side chain found in ilimaquinone analog 18. When the side
chain emanates from C3 of the decalin ring system, however, this
active site tunnel is not accessible to the substrate.
Figure 5. Comparison of binding of (a) AdoHcy and (b) ilimaquinone in AdoHcy
hydrolase (c) modeling of ilimaquinone analog 18 made for affinity studies, and (d)
the bonding mode of 11c.

6712
B. G. Kim et al. / Bioorg. Med. Chem. 17 (2009) 6707–6714
In a likewise gratifyingly manner, the binding model is also con-
OH
O
sistent with structure–activity relationships for compounds noted
in Scheme 1. The methoxy group on the quinone, which functions
as a hydrogen-bond acceptor, is replaceable by substituted nitro-
gens (2c, 2e). The quinone hydroxyl group can also be exchanged
with substituted nitrogens to provide improved activity (2d-2k)
as the additional alkyl chain provides favorable hydrophobic
interactions with the enzyme. The molecular modeling also can ac-
count for subtle differences in the inhibitory activities for the ana-
logs 10a–e and 11a–e. Modeling of analogs 11a–e show that the
aromatic dihydroxyl groups, as well as the adenine group, are well
positioned in the hydrophilic region typically occupied by the ri-
bose and adenine groups of known inhibitors, while the pendent
aliphatic functionality provides additional hydrophobic stabiliza-
tion to the enzyme/ligand complex (Figs. 5d and 6). When com-
pared with the analogs 11a–e, the aromatic hydroxyl groups in
analogs 10a–e, on the other hand, have differing hydrogen-bond-
ing interactions with the enzyme. Analogs 11a–e hydrogen bond
with Asp-109, while analogs 10a–e have hydrogen-bonding inter-
actions with Thr-57. This, in part, may account for the 10-fold
greater activities of the 11a–e analog series. More importantly,
the molecular model provides guidance for the design of new ilim-
aquinone-based inhibitors.
With a predictive binding model in place, our strategy for
optimizing interactions with the enzyme was to reexamine the
three regions for optimal cooperative enzyme binding. In that re-
gard, systematic changes to the aliphatic, ribose, and base re-
gions were planned. As indicated in Figure 6, initial
optimization studies focused on the placement of hydroxyl
groups on the aromatic ring. Specifically, an alternative to the
unattractive 1,2-catechol functionality would be preferred for
further development of this series. Accordingly, 1,3-dihydroxya-
renes were investigated as this functionality benefits from both
a hydrogen bond with Asp-109, as in the 11 analog series, and
a favorable interaction with Thr-57 as enjoyed by the 10 analog
series (Fig. 6).
HO
O
a,b
c,d
OBn
BnO
BnO
OBn
HO
OH
NH2
N
20
21
N
N
N
e,f
O
HO
OH
19
Scheme 4. Synthesis of analog 19. (a) BnBr, K2CO3, acetone, 95%; (b) m-CPBA,
CH2Cl2, then K2CO3, MeOH, 92%; (c) cyclohexylmethyl bromide, Cs2CO3, Bu4NI,
acetone, 85%; (d) DMF, POCl3; NaBH4, MeOH, 93%; (e) PPh3, bis-Boc-adenine, DEAD,
THF/benzene (1:1), 27%; (f) HCO2H, 96%; Pd/C (10%), H2, MeOH/EtOAc, 88%.
reaction, Boc deprotection, and hydrogenolysis furnished com-
pound 19 in 23% overall yield for three steps.
Likewise, an adenine surrogate study was also pursued. It was
envisioned that a late stage coupling of adenine variants with
benzyl chloride 25 would be an efficient entry into these com-
pounds (Scheme 5). Reduction of 22 with LAH and subsequent
TBS protection of the resulting benzyl alcohol afforded com-
pound 23. Lithium anion of 23 generated by metal halogen ex-
change with n-BuLi, was reacted with triisopropyl borate
followed by oxidative workup to provide a phenol, which was
then reacted with bromomethyl cyclohexane to afford 24 in
three steps. Successive TBS deprotection and chlorination of 24
then furnished the desired benzyl chloride 25. With compound
25 in hand, we substituted adenine for various secondary
amines. The coupling of benzyl chloride 25 and different amines
were straightforward and gave us 6 analogs of the chimeric hy-
brid 26a–f (Fig. 7).
In a similar fashion, we prepared an ester analog of inhibitor
11c; also using a late stage coupling strategy (29, Scheme 5).
The synthesis started from 5-bromobenzyl alcohol, which was
converted to nitrile 27. Hydrolysis of 27 followed by benzylation
of the resulting acid afforded an ester, which was transformed
to the corresponding benzyl chloride 28. Coupling of 28 with
bis-Boc-adenine and subsequent hydrogenolysis, EDC coupling,
and deprotection of the resulting adduct furnished 29 in four
steps.
Toward this end, 1,3-dihydroxy analog 19 was prepared from
commercially available 2,4-dihydroxybenzaldehyde, which was
converted to phenol 20 through dibenzylation followed by basic
Bayer–Villager reaction in 87% yield in two steps (Scheme 4). Reac-
tion of phenol 20 with cyclohexylmethyl bromide and subsequent
Vilsmeier formylation afforded the corresponding benzaldehyde,
which was reduced to the benzyl alcohol 21. Successive Mitsunobu
Preliminary screening of inhibitory activity of compounds 26a–
f against AdoHcy hydrolase was carried out and percent inhibition
Figure 6. Design for new AdoHcy hydrolase inhibitors.

B. G. Kim et al. / Bioorg. Med. Chem. 17 (2009) 6707–6714
6713
OTBS
O
AdoHcy hydrolase inhibition assay
OTBS
O
O
Br
O
Br
OMe
100
90
80
70
60
50
40
30
20
10
0
c,d
a,b
O
O
23
g
O
O
22
24
e,f
Cl
O
OH
O
O
NC
26a-f
O
O
27
25
NH2
N
h,i
N
O
Cl
O
O
N
BnO
11c
26a
26b
26c
26d
26f
29
19
j-l
N
O
Graph 2. Relative inhibition of AdoHcy hydrolase at 50
l
M.
O
OH
OH
28
29
3. Conclusions
Scheme 5. Adenine and linker alternatives in hybrid 11 series. (a) LAH, THF, 99%;
(b) TBSCl, Et3N, CH2Cl2, 93%; (c) n-BuLi; B(OiPr)3; H2O2, NaOH, 86%; (d) Cs2CO3,
cyclohexylmethyl bromide, Bu4NI, DMF, 82%; (e) TBAF, THF, quant.; (f) SOCl2, pyr,
CH2Cl2, À50 °C, 82%; (g) CuCN, DMF, 150 °C, 61%; (h) LiOH, MeOH/H2O, 70 °C; BnBr,
K2CO3, DMF, 86%; (i) SOCl2, CH2Cl2, pyr, 79%; (j) K2CO3, bis-Boc-adenine, K2CO3,
DMF, Bu4NI, rt, 72%; (k) Pd/C (10% on carbon), MeOH, 93%; (l) EDC, DMAP,
cyclohexyl alcohol, CH2Cl2; formic acid 88%.
Several new ilimaquinone analogs were prepared and screened
for their ability to inhibit AdoHcy hydrolase. The design of these
inhibitors were based on the assumption that ilimaquinone is
functioning as a nucleoside analog. These assumptions combined
with molecular modeling have lead to the preparation of simplified
analogs with improved inhibitory activities over that of ilimaqui-
none. Specifically, the preparation and evaluation of ilimaqui-
none/adenosine hybrids indicated that, when interacting with the
enzyme, the quinone moiety of ilimaquinone is serving as a ribose
mimic. These insights may lead to a new class of AdoHcy hydrolase
inhibitors with improved specificity, activity and bioavailability.
of AdoHcy hydrolase at 50
marized in Graph 2. At 50
l
M concentration of inhibitor is sum-
lM, none of these compounds showed
improved inhibition of AdoHcy hydrolase over analog 11c. Molec-
ular modeling suggests that the adenine amine enjoys hydrogen-
bonding interactions with Asp-141 and Thr-216 of the enzyme.
While compounds 26a and 26f are capable of maintaining this
hydrogen bonding interaction with AdoHcy hydrolase, the other
analogs are devoid of such functional groups. This difference may
be responsible for the inhibitory trends observed for compounds
26b–d.
4. Experimental section
4.1. S-Adenosylhomocysteine hydrolase assay
The ilimaquinone structural analogs were added to assay at fi-
nal concentrations ranging between 0.01 and 100
Adenosylhomocysteine (9.4 g, 25.4 nM, 21.8 Ci) was incubated
in the S-Adenosylhomocysteine hydrolase (0.35 g) at 37 °C for
10 min in a total reaction volume of 250 L. The incubation buffer
l
M. [8-14C]-S-
The inhibitory activity of compounds 19 and 29 were also inves-
tigated against AdoHcy hydrolase. As predicted from the molecular
modeling study, ester 29 possesses an inhibitory activity similar to
l
l
l
l
that of hybrid analog 11c (IC50 = 0.4
lM
0.2 lM), and the meta-
was 15 mM HEPES, 5 mM Mg(OAc)2, 150 mM KCl, 2 mM 2-mercap-
toethanol, 0.25% bovine serum albumin, and pH 7.0 including
adenosine deaminase (50 units/mL). The incubations were stopped
by the addition of formic acid (5 N, 50 mL) and 250 mL from the
solution was poured onto a SP-Sephadex-C25 column (10 mm Â
30 mm) pre-equilibrated with 0.1 N formic acid. [14C]-Inosine
was eluted with formic acid (0.1 N, 6 mL) and collected in a scintil-
lation vial with scintillation fluid (5 mL).
substituted dihydroxyl analog 19 also shows comparable inhibi-
tory activities in the nanomolar range.
NH2
N
NHMe
N
SH
N
N
N
N
N
N
N
N
N
N
Acknowledgments
N
N
N
N
R
R
R
R
11c
26a
26f
26b
We thank the National Institutes of Health (CA-66617) for
financial support. TGC and HYL acknowledge support by a grant
from Marine Biotechnology Program funded by Ministry of Land,
Transport and Maritime Affairs, Republic of Korea. We also thank
Professor Mark P. Wentland for his critical reading and comments.
SMe
N
SMe
N
N
N
N
N
N
R
N
N
N
N
N
26c
26d
26e
R
R
Supplementary data
O
Supplementary data (experimental procedures, as well as ana-
lytical and spectral characterization data for new compounds)
associated with this article can be found, in the online version, at
doi:10.1016/j.bmc.2009.07.061.
R =
OH
OH
Figure 7. Adenine variants 11c and 26a–f.

6714
B. G. Kim et al. / Bioorg. Med. Chem. 17 (2009) 6707–6714
V. J. Biol. Chem. 1993, 122, 11976; (c) Jamora, C.; Takizawa, P. A.; Zaarour, R. F.;
References and notes
Denesvre, C.; Faulkner, D. J.; Malhotra, V. Cell 1997, 91, 617.
11. Nambiar, M. P.; Wu, H. C. Exp. Cell Res. 1995, 219, 671.
12. Bruner, S. D.; Radeke, H. S.; Tallarico, J. A.; Snapper, M. L. J. Org. Chem. 1995, 60,
1114.
13. Radeke, H. S.; Digits, C. A.; Bruner, S. D.; Snapper, M. L. J. Org. Chem. 1997, 62,
2823.
14. (a) Radeke, H. S.; Digits, C. A.; Casaubon, R. L.; Snapper, M. L. Chem. Biol. 1999, 6,
639; (b) Casaubon, R. L.; Snapper, M. L. Bioorg. Med. Chem. Lett. 2001, 14, 133.
15. Cools, M.; De Clercq, E. Biochem. Pharmacol. 1990, 40, 2259.
16. (a) Romines, K. R.; Chrusciel, R. A. Curr. Med. Chem. 1995, 2, 825; (b) Wender, P.
A.; Cribbs, C. M. Adv. Med. Chem. 1992, 1, 1; (c) Wender, P. A.; Cribbs, C. M.;
Koehler, K. F. Proc. Natl. Acad. Sci. U.S.A. 1988, 85, 7197.
1. Ueland, P. M. Pharmacol. Rev. 1982, 34, 223.
2. For initial mechanistic insight into AdoHcy, see: Palmer, J. L.; Abeles, R. H. J. Biol.
Chem. 1979, 254, 1217.
3. (a) Yuan, C. S.; Saso, Y.; Lazarides, E.; Borchardt, R. T.; Robins, M. J. Exp. Opin.
Therap. Pat. 1999, 9, 1197; Yuan, C. S. et al Adv. Antiviral Drug Design 1996, 2, 41;
(b) Chiang, P. K. Pharmacol. Ther. 1998, 77, 115; (c) Daelemans, D.; Este, J. A.;
Witvrouw, M.; Pannecouque, C.; Jonckheere, H.; Aquaro, S.; Perno, C.-F.; Clercq,
E. D.; Vandamme, A.-M. Pharmacology 1997, 52, 1157; (d) De Clercq, E.
Nucleosides, Nucleotides, Nucleic Acids 2005, 24, 1395.
4. (a) Turner, M. A.; Yuan, C.-S.; Borchardt, R. T.; Hershfield, M. S.; Smith, G. D.;
Howell, P. L. Nat. Struct. Biol. 1998, 5, 369; (b) Hu, Y.; Komoto, J.; Huang, Y.;
Gomi, T.; Ogawa, H.; Takata, Y.; Fujioka, M.; Takusagawa, F. Biochemistry 1999,
38, 8323.
5. (a) Wnuk, S. F. Mini-Rev. Med. Chem. 2001, 1, 307; (b) De Clercq, E.;
Montgomery, J. A. Antiviral Res. 1983, 3, 17; (c) De Clercq, E. Nucleosides,
Nucleotides 1998, 17, 625.
6. Yamada, T.; Komoto, J.; Lou, K.; Ueki, A.; Hua, D. H.; Sugiyama, K.; Takata, Y.;
Ogawa, H.; Takusagawa, F. Biochem. Pharmacol. 2007, 73, 981.
7. (a) Wolfe, M. S.; Borchardt, R. T. J. Med. Chem. 1991, 34, 1521; For representative
examples, also see: (b) Jarvi, E. T.; McCarthy, J. R.; Mehdi, S.; Matthews, D. P.;
Edwards, M. L.; Prakash, N. J.; Bowlin, T. L.; Sunkara, P. S.; Bey, P. J. Med. Chem.
1991, 34, 647. and references cited therein; (c) Wnuk, S. F.; Mao, Y.; Yuan, C.-S.;
Borchardt, R. T.; Andrei, G.; Balzarini, J.; De Clercq, E.; Robins, M. J. J. Med. Chem.
1998, 41, 3078; (d) Parry, R. J.; Muscate, A.; Askonas, L. J. Biochemistry 1991, 30,
9988; (e) Guillerm, G.; Muzard, M.; Glapski, C.; Pilard, S.; De Clercq, E. J. Med.
Chem. 2006, 49, 1223; (f) Glazer, R. I.; Knode, M. C. J. Biol. Chem. 1984, 259, 12964.
8. (a) Luibrand, R. T.; Erdman, T. R.; Vollmer, J. J.;Scheuer, P. J.;Clardy, J. C. Tetrahedron
1979, 35, 609; (b) Capon, R. J.; MacLeod, J. K. J. Org. Chem. 1987, 52, 5059.
9. (a) Loya, S.; Tal, R.; Kashman, Y.; Hizi, A. Antimicrob. Agents Chemother. 1990, 34,
2009; (b) Kushlan, D. M.; Faulkner, D. J.; Parkanyi, L.; Clardy, J. C. Tetrahedron
1989, 45, 3307; (c) Bourguet-Kondracki, M.-L.; Longeon, A.; Morel, E.; Guyot, M.
Int. J. Immunopharm. 1991, 13, 393; (d) Schroder, H. C.; Wenger, R.; Gerner, H.;
Reuter, P.; Kuchino, Y.; Sladic, D.; Muller, W. E. G. Cancer Res. 1989, 49, 2069; (e)
Sarin, P. S.; Sun, D.; Thornton, A.; Muller, W. E. G. J. Natl. Cancer. Inst. 1987, 78,
663; (f) Belisario, M. A.; Maturo, M.; Pecce, R.; De, R. S.; Villani, G. R. D.
Toxicology 1992, 72, 221; (g) Muller, W. E. G.; Dogovic, N.; Zahn, R. K.; Maidhof,
A.; Diehl, S. B.; Bedker, C.; Sachsse, W.; Gasic, M. J.; Schroeder, H. C. Basic Appl.
Histochem. 1985, 29, 321; (h) Rangel, H. R.; Dagger, F.; Compagnone, R. S. Cell
Biol. Int. 1997, 21, 337.
17. Total syntheses of the diterpenoids (À)-kolavenol and (À)-agelasine B: Piers,
E.; Roberge, J. Y. Tetrahedron Lett. 1992, 33, 6923.
18. A sample was generously provided by Professor Sidney M. Hecht, University of
Virginia.
19. Asmarines A–C; three novel cytotoxic metabolites from the marine sponge
Raspailia sp.: Yosief, T.; Rudi, A.; Stein, Z.; Goldberg, I.; Gravalos, G. M. D.;
Schleyer, M.; Kashman, Y. Tetrahedron Lett. 1998, 39, 3323.
20. (a) Stebbins, C. E.; Russo, A. A.; Schneider, C.; Rosen, N.; Hartl, F. U.; Pavletich,
N. P. Cell 1997, 89, 239; (b) Prodromou, C.; Roe, S. M.; O’Brien, R.; Ladbury, J. E.;
Piper, P. W.; Pearl, L. H. Cell 1997, 90, 65.
21. For recent AdoHcy hydrolase inhibitors, see: (a) Yin, X. Q.; Schneller, S. W.
Tetrahedron 2005, 61, 1839; (b) Yang, M.; Schneller, S. W. Bioorg. Med. Chem.
Lett. 2005, 15, 149; (c) Guillerm, G.; Muzard, M.; Glapski, C.; Pilard, S. Bioorg.
Med. Chem. Lett. 2004, 14, 5803; (d) Guillerm, G.; Muzard, M.; Glapski, C. Bioorg.
Med. Chem. Lett. 2004, 14, 5799; (e) Wnuk, S. F.; Lewandowska, E.; Sacasa, P. R.;
Crain, L. N.; Zhang, J.; Borchardt, R. T.; De Clercq, E. J. Med. Chem. 2004, 47,
5251; (f) Yin, X.; Schneller, S. W. Nucleosides, Nucleotides, Nucleic Acids 2004, 23,
67; (g) Kumamoto, H.; Onuma, S.; Tanaka, H. J. Org. Chem. 2004, 69, 72; (h)
Moon, H. R.; Lee, K. M.; Lee, H. J.; Lee, S. K.; Park, S. B.; Chun, M. W.; Jeong, L. S.
Nucleosides, Nucleotides, Nucleic Acids 2005, 24, 707; (i) Kitade, Y.; Ando, T.;
Yamaguchi, T.; Hori, A.; Nakanishi, M.; Ueno, Y. Bioorg. Med. Chem. 2006, 14,
5578.
22. This binding model implies that these compounds should function as
competitive inhibitors of AdoHcy, however, we have not yet experimentally
confirmed this prediction.
23. Alam, A.; Takaguchi, Y.; Ito, H.; Yoshida, T.; Tsuboi, S. Tetrahedron 2005, 61,
1909.
24. Dey, S.; Garner, P. J. Org. Chem. 2000, 65, 7697.
25. Witiak, D. T.; Patel, D. B.; Lin, Y. J. Am. Chem. Soc. 1967, 89, 1908.
26. Hagler, A. T.; Lifson, S.; Dauber, P. J. Am. Chem. Soc. 1979, 101, 5122.
10. (a) Talizawa, P. A.; Yucel, J. K.; Veit, B.; Faulkner, D. J.; Deerinck, T.; Soto, G.;
Ellisman, M.; Malhotra, V. Cell 1993, 73, 1079; (b) Veit, B.; Yucel, J. K.; Malhotra,