Mycophenolic acid analogs with a modified metabolic profile

Article abstract of DOI:10.1016/j.bmc.2008.08.062

Mycophenolic acid (MPA), a clinically used immunosuppressant, is extensively metabolized into an inactive C7-glucuronide and removed from circulation. To circumvent the metabolic liability imposed by the C7-hydroxyl group, we have designed a series of hybrid MPA analogs based on the pharmacophores present in MPA and new generations of inosine monophosphate dehydrogenase (IMPDH) inhibitors. The synthesis of MPA analogs has been accomplished by an allylic substitution of a common lactone. Biological evaluations of these analogs and a preliminary structure-activity relationship (SAR) are presented.

Full text of DOI:10.1016/j.bmc.2008.08.062

Bioorganic & Medicinal Chemistry 16 (2008) 9340–9345
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
Mycophenolic acid analogs with a modified metabolic profile
a
a
b
a
Liqiang Chen a, , Daniel J. Wilson , Nicholas P. Labello , Hiremagalur N. Jayaram , Krzysztof W. Pankiewicz
*
a Center for Drug Design, Academic Health Center, University of Minnesota, 516 Delaware Street S.E., Minneapolis, MN 55455, USA
b Department of Biochemistry and Molecular Biology, Indiana University School of Medicine and Richard Roudebush Veterans Affairs Medical Center,
1481 West Tenth Street, Indianapolis, IN 46202, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Mycophenolic acid (MPA), a clinically used immunosuppressant, is extensively metabolized into an inac-
tive C7-glucuronide and removed from circulation. To circumvent the metabolic liability imposed by the
C7-hydroxyl group, we have designed a series of hybrid MPA analogs based on the pharmacophores pres-
ent in MPA and new generations of inosine monophosphate dehydrogenase (IMPDH) inhibitors. The syn-
thesis of MPA analogs has been accomplished by an allylic substitution of a common lactone. Biological
evaluations of these analogs and a preliminary structure–activity relationship (SAR) are presented.
Ó 2008 Elsevier Ltd. All rights reserved.
Received 14 July 2008
Revised 20 August 2008
Accepted 26 August 2008
Available online 29 August 2008
Keywords:
Mycophenolic acid
Inosine monophosphate dehydrogenase
Allylic substitution
1. Introduction
that mycophenolic acid and its close analogs behaved as agonists
of peroxisome proliferator activated receptor
c (PPARc
).10 Further-
Mycophenolic acid (Fig. 1, MPA), arguably the first antibiotic,
was isolated more than one hundred years ago.1 Subsequent stud-
ies revealed the role of MPA as a highly potent inhibitor of inosine
monophosphate dehydrogenase (IMPDH), a key enzyme in the de
novo synthesis of guanine nucleotide. Consequently, MPA exhibits
an impressive anti-proliferation activity and has been established
as an anti-cancer2 and anti-viral agent, as well as an immunosup-
pressant.3 Currently mycophenolate mofetil (Fig. 1, MMF, a pro-
drug of MPA) and mycophenolate sodium (MPS) are used in
combination with other immunosuppressants for the treatment
of organ rejection in heart, liver, and kidney transplants.
While it is widely used in transplantation, MPAs anti-cancer
properties are again of current interest. For instance, MPA has been
demonstrated to be an effective inducer of differentiation in a
number of cancer cell lines such as those of androgen-independent
prostate cancer,4,5 leukemia,6 and melanoma.7 A more recent study
showed that MPA exerted its anti-angiogenic effect through inhibi-
tion of human IMPDH type 1,8 an isoform long believed to be a
house keeping enzyme. Furthermore, MPA acts synergistically with
imatinib,9 a first line therapy in the treatment of chronic myolog-
enous leukemia (CML). These findings strongly support the role
of MPA as a potential anti-cancer agent.
more, we designed a hybrid compound, in which the carboxylic
acid of MPA was replaced with a hydroxamic acid group. This min-
or structural modification converted MPA into mycophenolic acid
hydroxamic acid (Fig. 1, MAHA),11 a dual inhibitor of both IMPDH
and histone deacetylases (HDAC). These studies confirmed versa-
tile biological properties of MPA and the significance of MPA core
structure as a potentially useful pharmacophore.
Unfortunately, mycophenolic acid suffers from a severe meta-
bolic drawback. The C7-phenolic hydroxyl group is highly suscep-
tible to glucuronidation, which abrogates its inhibitory activity
against IMPDH. This metabolic liability might explain the limited
clinical efficacy of MPA as demonstrated in a Phase I clinical trial
of mycophenolate mofetil in advanced multiple myeloma pa-
tients.12 Furthermore, subsequent cleavage of MPA glucuronide
in the gastrointestinal track causes side effects and also leads to
secondary absorption peaks through a process called enterohepatic
recirculation.13
Consequently, MPA analogs and mimics that lack MPAs meta-
bolic liability have been pursued by researchers in both academia
and industry. Significant efforts were focused on bio-isosteric
replacements, albeit with limited success. For instance, replace-
In addition to its anti-cancer activity, MPA has also been sub-
jected to structural modifications that elicit interesting biological
properties other than inhibition of IMPDH. A recent study revealed
OH
6
MPA, R = OH
O
1
7
O
R
MMF, R =
N
O
O
2
5
MPS, R = ONa
O
OMe
3
4
MAHA, R = NHOH
* Corresponding author. Tel.: +1 612 625 3264; fax: +1 612 624 8252.
E-mail address: chenx462@umn.edu (L. Chen).
Figure 1. Mycophenolic acid and its derivatives.
0968-0896/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2008.08.062

L. Chen et al. / Bioorg. Med. Chem. 16 (2008) 9340–9345
9341
ment of 7-hydroxyl group with a fluorine atom drastically reduced
O
N
MeO
O
OH
the inhibitory activity.14 A C7-amino analog of MPA was resistant
to glucuronidation, however with significantly reduced anti-
IMPDH inhibitory activity.14 In yet another isosteric replacement,
the phenolic hydroxyl group was replaced with an indole NH
group, resulting in a MPA analog with a low micromolar mean
GI50 in a NCI human tumor panel screen.15 However, further mod-
ification by introducing a methyl and methoxy group, both of
which were believed to favorably position the hexenoic side chain
in MPA, produced an inactive compound.16
OH
O
O
MeO
R2
N
R1
1
3
4
R1 = H, R2 = H
R1 = H, R2 = OMe
2
R1 = Me, R2 = OMe
Figure 3. Mycophenolic acid analogs as IMPDH inhibitors.
Our group designed MPA derivatives in which a truncated MPA
was connected to an adenosine moiety via a linker such as methyl-
enebis(phosphonate), leading to mycophenolic adenine dinucleo-
tides such as C2-MAD17 and its derivatives.18 Interestingly, C2-
MAD proved resistant to glucuronidation17 probably due to the
steric hindrance imposed by the additional adenosine moiety, thus
making C2-MAD a poor substrate of UDP-glucuronosyltransferases.
Furthermore, as exemplified by VX-49719 and BMS-33719720
(Fig. 2), structure-based drug design has yielded series of novel
IMPDH inhibitors based on a methoxy-(5-oxazolyl)-phenyl (MOP)
moiety, which was believed to bind in a fashion analogous to the
substituted benzofuranone present in MPA. It has been demon-
strated that the oxazolyl group can form key hydrogen bonds with
human IMPDH2 residues Gly326 and Thr333, both of which also
interact with the aromatic core of MPA. Nevertheless, lack of a phe-
nolic hydroxyl group prevents VX-497 engaging in a hydrogen
bonding with Thr333. Medicinal chemistry campaigns by a number
of pharmaceutical companies led to numerous and structurally di-
verse IMPDH inhibitors, which have been recently reviewed.3
R
I
R
OH
+
O
O
O
5
Figure 4. Retrosynthetic scheme of mycophenolic acid analogs.
duction of the side chain.23 For the synthesis of compounds 1–4,
we devised a convergent and efficient synthetic scheme (Fig. 4),
involving an allylic substitution reaction between cuprates derived
from aromatic iodides and lactone 5,24 which can be conveniently
prepared in one step from ethyl levulinate. In our approach, the
lactone serves as a leaving group and an allylic substitution reac-
tion leads to the requisite hexenoic carboxylic acid side chain in
one step. It was clear that our synthetic approach would yield a
mixture of E and Z isomers. However, we assumed that the E/Z ra-
tio would be similar in all the target compounds, allowing us to
perform a preliminary SAR analysis. In addition, subsequent sepa-
ration and evaluation individual isomers would enable us to inves-
tigate the role of olefin geometry on the inhibitory activity of pure
isomers.
The syntheses of various iodides are depicted in Scheme 1. Con-
densation of tosylmethyl isocyanide (TosMIC) with aldehyde 6
afforded 5-oxazolyl substituted bromide 9. The conversion of bro-
mide 9 into iodide 12 was accomplished by Buchwald’s copper-cat-
alyzed halogen exchange reaction25 in excellent yields. Preparation
of iodides 13 and 14 was accomplished in procedures almost iden-
tical to those described for iodide 12 from bromide 7 and 8, respec-
tively. Finally, iodonization of tetra-substituted aldehyde 15
yielded iodide 16, which was condensed with TosMIC to give io-
dide 17.
2. Results and discussion
In our continuing efforts on the design, synthesis and evaluation
of potential anti-cancer agents, we became interested in MPA
derivatives and analogs devoid of the metabolically labile phenolic
hydroxyl group. Here we report a series of hybrid compounds
(Fig. 3, 1–4) in which a MOP moiety is connected to a hexenoic side
chain. We expected that the MOP group would interact with
IMPDH as proposed for VX-497 while the hexenoic side chain
would behave like that of MPA. We were interested in investigat-
ing the structure–activity relationship (SAR) of compounds in
which the position of side chain is varied in relation to the MOP
group. In addition, we also hoped to assess the role of methoxy
and methyl groups which are placed adjacent to the side chain,
as proposed in compound 3 and 4.
These iodides were converted into Gringard reagents using pro-
cedures reported by Knochel (Scheme 2).26 Subsequent one-pot
transmetallation gave cuprates, which underwent an allylic substi-
tution reaction to give MPA analogs 1–4 in moderate yields.
2.1. Chemistry
Various synthetic approaches have been devised for the synthe-
sis of MPA and its analogs. Many of these syntheses involve forma-
tion of the aromatic moiety followed by introduction of the side
chain through various methodologies such as Stille coupling,21
ortho-ester Claisen rearrangement,15,16 and Heck carbonylation
and olefination.22 Nevertheless, a ring annulation sequence has
also been proposed for the formation of aromatic ring after intro-
O
O
O
N
N
R3
R2
R3
R2
R3
R2
H
b
a
MeO
MeO
MeO
R1
R1
R1
6
9
12
R
R1 = H, R2 = H, R3 = Br
7 R1 = H, R2 = Br, R3 = H
R1 = H, R2 = H, R3 = Br
10 R1 = H, R2 = Br, R3 = H
1 = H, R2 = H, R3 = I
13 R1 = H, R2 = I, R3 = H
8
11
14
R
1 = H, R2 = OMe, R3 = Br
R
1 = H, R2 = OMe, R3 = Br
R
1 = H, R2 = OMe, R3 = I
O
O
O
O
N
O
H
N
N
I
I
a
H
O
H
c
O
O
N
N
N
MeO
OMe
O
MeO
OMe
MeO
OMe
O
N
O
O
15
MeO
N
H
MeO
N
H
N
H
17
16
Scheme 1. Reagents and conditions: (a) TosMIC, K2CO3, MeOH, reflux; (b) NaI, CuI,
trans-N,N0-dimethylcyclohexane-1,2-diamine, 1,4-dioxane, 110 °C; (c) I2, silver
trifluoroacetate, CHCl3, rt.
VX-497 (Merimepodib)
BMS-337197
Figure 2. Representative IMPDH inhibitors.

9342
L. Chen et al. / Bioorg. Med. Chem. 16 (2008) 9340–9345
O
pound 1 led to inhibitor 3 (Ki = 4.3 lM) which exhibited slightly
O
N
N
I
OH
enhanced potency. An additional methyl group as shown in com-
pound 4 abrogated the inhibitory activity. This finding is clearly
in contrast to the trends observed for MPA.27
a
O
MeO
R2
MeO
R2
R1
R1
In order to further investigate the significance of olefin geome-
try, compound 3 was separated by HPLC to give compound 3-E and
3-Z. The olefin geometry assignment was based on the 1D nuclear
overhauser effect (NOE) experiment conducted on both compound
3-E and 3-Z. The selective NOE interactions are shown in Figure 5.
Among the NOE correlations, the most informative observation
was that the vinylic proton showed a correlation with the methyl
protons in 3-Z whereas there was no such correlation observed
in 3-E. The assignment of 3-E and 3-Z isomers also aided the deter-
mination of E/Z ratio of compound 3 as shown in Table 1. Both
compounds were also evaluated for their ability to inhibit human
12 R1 = H, R2 = H
14
1 R1 = H, R2 = H
3
R1 = H, R2 = OMe
17 R1 = Me, R2 = OMe
R1 = H, R2 = OMe
4 R1 = Me, R2 = OMe
MeO
O
I
MeO
O
OH
a
O
N
N
13
2
Scheme 2. Reagents and conditions: (a) i—iPrMgCl, LiCl, THF; ii—CuBrꢀMe2S,
ꢁ30 °C; iii—compound 5, ꢁ30 °C.
IMPDH enzymes. Isomer 3-E (Ki = 2.3 lM) was approximately 10-
fold more potent than 3-Z, indicating the key role played by the
olefin geometry in relation to IMPDH inhibition.
As we anticipated, MPA analogs 1–4 were obtained as a mixture
of E and Z isomers (Table 1). As detailed below, pure isomers of
compound 3 were separated by high-performance liquid chroma-
tography (HPLC) and hence the E/Z ratio was determined with E
isomer slightly preferred over Z isomer. Since the final iodides as
shown in Scheme 1 were structurally similar, it is reasonable to be-
lieve that compounds 1–4, which were prepared under nearly
identical reaction conditions, would show a similar selectivity to-
ward E isomers.
As indicated in Table 1, MPA analogs, like MPA itself, exhibited a
slight selectivity toward IMPDH2 over IMPDH1. This observation
appears to support our rationale that MPA analogs 1–4 interact
with IMPDH enzymes in a mode analogous to that MPA. Interest-
ingly, 3-E and 3-Z display a similar selectivity against IMPDH2,
suggesting that the selectivity mainly arises form the MOP moiety
whereas the geometry of side chain plays a minor role.
MPA analogs 1–3, 3-E, and 3-Z were also tested for their abili-
ties to inhibit the growth of K562 cells (Table 1) as previously de-
scribed.18 In line with their modest inhibitory activity in the
enzymatic assay, MPA analogs showed weak inhibition of K562
cells. Nevertheless, compounds 1–3, 3-E, and 3-Z do offer an
advantage considering the ratio of K562 cells proliferation IC50
over IMPDH2 Ki. For instance, the ratio for compound 3 was
approximately 10 and significantly lower than that of MPA (770).
Given the lack of C7-hydroxyl group in our MPA analogs, it is
reasonable to speculate that removal of the glucuronidation-prone
hydroxyl group diminishes their metabolic liability and therefore
enhances their cellular activity in comparison to their enzymatic
IC50 values. It was also interesting to note that compound 3-E
exhibited an inhibitory activity comparable to that of compound
3-Z even though the former compound was approximately 10-fold
more potent as an IMPDH inhibitor.
The E/Z ratios as shown in Table 1 were determined based on
the integrations of 1H NMR signals displayed by individual isomers,
assuming that E isomer was preferred. For compounds 1 and 2, the
methyl protons from the hexenoic side chain were used. However,
signals from the aromatic region were used for compounds 3 and 4
since signals from the side chain methyl group were identical for
both E and Z isomers. As listed in Table 1, compound 1 possessed
an E/Z ratio of 1.5:1 while compounds 2 and 3 displayed very sim-
ilar E/Z ratios, which would allow for a preliminary SAR analysis
without separation of individual isomers. Nevertheless, compound
4 exhibited a slightly higher E/Z ratio.
2.2. Biological evaluations
MPA analogs 1–4 were tested against human IMPDH type 1 and
type 2 (Table 1). Compound 1 showed modest activities against
both type 1 and type 2. Compound 2 was approximately twofold
less active than compound 1, indicating the importance of side
chain position.
2.3. Computational modeling
On the basis of structural information available in the litera-
ture,19 the oxazole ring of MOP moiety was initially positioned
within hydrogen bonding range of Gly326 and Thr333 and mini-
mized with the OPLS-2005 forcefield in a model of the IMPDH
binding site. While the measured distances and resulting geometry
were reasonable, the lowest energy conformations routinely fea-
tured a Thr333 hydrogen bond to Cys331 rather than the expected
interaction with the ligand. To determine the accuracy of the force-
field model, the structure was re-minimized using a hybrid QM/
MM method. That is, while most of the receptor was modeled with
the OPLS-2005 forcefield (all residues within five angstroms of the
Compounds 3 and 4 were designed to assess the roles played by
a methoxy or methyl group which were placed adjacent to the hex-
enoic side chain. In MPA, the methoxy and methyl groups play cru-
cial roles in the optimal orientation of the hexenoic side chain and
therefore the high potency of MPA. For instance, the introduction
of methoxy or methyl group resulted in a 280- and 20-fold increase
in potency, respectively.27 However, in our series of MPA analogs,
attachment of a methoxy group ortho to the side chain in com-
Table 1
aE/Z ratios and biological evaluations of MPA analogs
Compound E:Z
ratio
IMPDH1 Ki
M)
IMPDH2 Ki
M)
K562 cells proliferation
O
OH
(
l
(
l
IC50
(lM)
MPA
1
2
3
4
NA
0.04
0.01
7.7
O
O
1.5:1
1.5:1
1.9:1
2.8:1
NA
20.8 1.4
42.8 3.2
16.0 2.4
>100
10.3 1.2
97.3 33.9
8.7 1.0
14.3 1.8
4.3 0.5
>100
2.3 0.2
30.1 3.3
>100
>100
56
ND
88
N
N
OH
H
O
H
MeO
O
MeO
O
3-E
3-Z
3-
E
3-Z
NA
71
a
NA, not applicable; ND, not determined.
Figure 5. NOE correlations observed for compound 3-E and 3-Z.

L. Chen et al. / Bioorg. Med. Chem. 16 (2008) 9340–9345
9343
3. Conclusions
In summary, we have designed a series of MPA analogs devoid of
the metabolically labile C7-hydroxyl group based on the combina-
tion of structural elements from MPA and IMPDH inhibitors previ-
ously reported. We have devised a convergent synthetic scheme
which allowed a rapid assembly of our target compounds from
substituted aromatic iodides and lactone 5. Through these studies,
we have demonstrated that replacement of the benzofuranone moi-
ety of MPA by a MOP group retains biological activities. The result-
ing MPA analogs displayed modest activity against isoforms of
human IMPDH and weak activity against K562 cells. We have also
observed an enhancement of cellular activity in our MPA analogs
in comparison to their enzymatic potency, presumably because
these MPA analogs are not susceptible to glucuronidation which
has severely limited MPAs medical applications. Further studies
on metabolically stable MPA analogs are currently in progress.
Figure 6. Compound 3-E is overlayed with MPA in the IMPDH crystal structure
1JR1. The protein is presented in red ribbon while MPA and 3-E are green and
mauve sticks, respectively. Key residues are shown as orange sticks. The nucleotide
substrate is colored white.
4. Experimental
4.1. General methods
ligand allowed to move), the ligand and several important residues
(Thr333, Ser275, and Ser276) were treated at the B3LYP/6-31G* le-
vel of theory which yielded the conformation presented inFigure 6.
A hydrogen bond between Thr333 and the ring oxygen is present
(3.2 angstroms, 169 degrees) as expected. Moreover, MPA is super-
imposed on 3-E as shown in Figure 6. The reduced potency of com-
pound 1 relative to 2 can be explained by the position to which the
hexenoic acid chain is attached to MOP. Attachment at the meta-
position leads to an inhibitor that closely mimics MPA, superior
to attachment at the para-position.
However, a significant loss in activity is observed in compound
3-E with respect to MPA even though the acid tail is attached meta
to the oxazole ring. Furthermore, addition of a methyl group at the
position that is ortho to both methoxy groups in 3 results in 4, an
inactive compound. The complete lack of activity is initially sur-
prising as this region of the protein is apparently devoid of steric
and electrostatic clashes with the methyl group. Instead, the effect
appears to be to due to the methyl group’s influence on the ligand
itself and subtle changes in the conformation of the hexenoic chain
that reduce binding affinity. In both cases it seems that MOP, while
a reasonable replacement for the MPA benzofuranone head group,
is not the best choice for pairing with the hexenoic acid tail. Qual-
itatively this observation can be demonstrated by the relatively
different position and conformation of the chain in Figure 6. The
sensitivity of IMPDH to changes in the tail is well documented,28
and it has been suggested that the loss in potency due to modifica-
tions of the acid tail is related to a weakening of its interaction with
Ser276.19 The carboxylic acid forms at least two hydrogen bonds
(Ser276 side chain and backbone). Using the QM/MM treatment
described above interaction energies can be calculated for the li-
gand–receptor complexes. The ligand is optimized in both a recep-
tor and solvent environment and the relative interaction is
evaluated according to the following equation:
All commercial reagents (Sigma–Aldrich, Acros) were used as
provided unless otherwise indicated. An anhydrous solvent dis-
pensing system (J.C. Meyer) using 2 packed columns of neutral alu-
mina was used for drying THF, Et2O, and CH2Cl2, while 2 packed
columns of molecular sieves were used to dry DMF. Solvents were
dispensed under argon. Flash chromatography was performed with
Ultra Pure silica gel (Silicycle) with the indicated solvent system.
Nuclear magnetic resonance spectra were recorded on a Varian
600 MHz with Me4Si or signals from residual solvent as the inter-
nal standard for 1H or 13C. Chemical shifts are reported in ppm, and
signals are described as s (singlet), d (doublet), t (triplet), q (quar-
tet), m (multiplet), br s (broad singlet), and dd (double doublet).
Values given for coupling constants are first order. The E/Z ratios
were determined based on the integrations of 1H NMR signals dis-
played by individual isomers. High resolution mass spectra were
recorded on an Agilent TOF II TOF/MS instrument equipped with
either an ESI or APCI interface.
4.2. Chemical synthesis
4.2.1. Bromide 9
A mixture of commercially available 5-bromo-2-methoxybenz-
aldehyde (6, 2.28 g, 10.6 mmol), tosylmethyl isocyanide (TosMIC,
2.09 g, 10.6 mmol) and K2CO3 (3.72 g, 26.9 mmol) in anhydrous
MeOH (30 mL) was heated at reflux for 22 h. After cooling to rt,
the mixture was poured into 300 mL of ice-water and the solid pre-
cipitated was filtered. The solid was washed with water, air-dried,
collected and dried in vacuo to give bromide 9 as a yellow solid
(2.59 g, 96%). 1H NMR (CDCl3) d 7.91 (s, 1H), 7.89 (d, J = 2.4 Hz,
1H), 7.57 (s, 1H), 7.39 (dd, J = 8.4, 2.4 Hz, 1H), 6.86 (d, J = 8.4 Hz,
1H), 3.95 (s, 3H). 13C NMR (CDCl3) d 154.6, 149.8, 146.6, 131.6,
128.6, 126.4, 118.8, 113.2, 112.6, 55.8. MS (ESI) calcd for
C10H9BrNO2 254.0 (M+H)+, found 254.0.
EðIÞ ¼ EðcomplexÞ ꢁ ½EðreceptorÞ þ EðligandÞꢂ:
4.2.2. Bromide 10
While we stress that much of the thermodynamic cycle is missing
from a simple end point analysis, the general trend of MPA
(ꢁ208 kcal/mol) > 3-E (ꢁ181 kcal/mol) > 4-E (ꢁ131 kcal/mol) cor-
relates well with their inhibitory activities in enzymatic assays. As
the conformations examined of each molecule maintain the appro-
priate hydrogen bonding with Ser276, it further suggests that subtle
conformational shifts are indeed enough to significantly alter the
binding free energy and therefore provides an explanation in agree-
ment with the observed SAR.
Following procedures similar to those described for bromide 9,
commercially available 4-bromo-2-methoxybenzaldehyde (7,
1.40 g, 6.53 mmol) was converted into bromide 10 as a yellowish
solid (1.56 g, 94%). 1H NMR (CDCl3) d 7.90 (s, 1H), 7.63 (d, J = 8.4
Hz, 1H), 7.54 (s, 1H), 7.19 (d, J = 8.4 Hz, 1H), 7.12 (s, 1H), 3.96 (s,
3H). 13C NMR (CDCl3) d 156.0, 149.6, 147.2, 127.0, 125.8, 124.0,
122.6, 116.2, 114.6, 55.8. MS (ESI) calcd for C10H9BrNO2 254.0
(M+H)+, found 254.0.

9344
L. Chen et al. / Bioorg. Med. Chem. 16 (2008) 9340–9345
4.2.3. Bromide 11
collected and dried in vacuo to give iodide 17 as a pale solid
(3.74 g, 93%). 1H NMR (CDCl3) d 8.03 (s, 1H), 7.91 (s, 1H), 7.55 (s,
1H), 3.80 (s, 3H), 3.73 (s, 3H), 2.34 (s, 3H). 13C NMR (CDCl3) d
159.1, 156.5, 149.9, 146.6, 133.3, 126.9, 125.0, 119.9, 86.4, 60.4,
59.7, 10.5. MS (ESI) calcd for C12H13INO3 346.0 (M+H)+, found
346.1.
Following procedures similar to those described for bromide 9,
commercially available 5-bromo-2,4-dimethoxybenzaldehyde (8,
5.08 g, 20.7 mmol) was converted into bromide 11 as a yellow solid
(5.60 g, 95%). 1H NMR (CDCl3) d 7.84 (s, 1H), 7.79 (s, 1H), 7.36
(s, 1H), 6.46 (s, 1H), 3.90 (s, 3H), 3.87 (s, 3H). 13C NMR (CDCl3)
d 156.5, 156.2, 149.2, 146.8, 129.9, 124.3, 111.3, 102.5, 96.2, 56.4,
55.8. MS (ESI) calcd for C11H11BrNO2 284.0 (M+H)+, found 284.0.
4.2.9. Carboxylic acid 1
To dry THF (30 mL) at ꢁ30 °C were added a solution of iPrMgCl
(1.4 M in THF, 3.0 mL, 4.2 mmol) and LiCl (184 mg, 4.34 mmol)
and the mixture was allowed to stir for 30 min. After addition of io-
dide 12 (906 mg, 3.01 mmol), the resulting solution was allowed to
stir at ꢁ30 °C for 2 h and CuBrꢀMe2S (869 mg, 4.22 mmol) was
added. After stirring for 1 h, a solution of lactone 524 (255 mg,
2.02 mmol) in dry THF (3 mL) was added and the resulting mixture
was allowed to stir at ꢁ30 °C for 18 h. The reaction mixture was
poured into saturated NH4Cl (200 mL plus 10 mL of 1 N HCl) and
the mixture was extracted with EtOAc (200 mL). The organic layer
was washed with brine (2ꢃ 100 mL) and then dried over MgSO4.
After filtration, the filtrate was concentrated and the residue was
purified by column chromatography (1–9% MeOH/CH2Cl2) to give
carboxylic acid 1 (E/Z ratio ꢄ 1.5:1) as a greenish syrup (127 mg,
21%). 1H NMR (CDOD3) d 8.28 (s, 1H), 7.72–7.44 (m, 2H), 7.12 (d,
J = 6.6 Hz, 1H), 6.99 (d, J = 8.4 Hz, 1H), 5.42–5.32 (m, 1H), 3.92 (s,
3H), 3.38–3.32 (m, 2H), 2.47 (br s, 2H), 2.35 (br s, 2H), 1.75 (s,
1.2H), 1.75 (s, 1.8H). HRMS calcd for C17H18NO4 300.1235 (MꢁH)ꢁ,
found 300.1237.
4.2.4. Iodide 12
A
mixtre of bromide 9 (1.52 g, 6.00 mmol), CuI (59 mg,
0.31 mmol), and NaI (1.81 g, 12.1 mmol) in anhydrous 1,4-dioxane
(6.0 mL) was evacuated and backfilled with N2 for five times. After
addition of racemic trans-N,N0-dimethylcyclohexane-1,2-diamine
(95 lL, 0.60 mmol), the mixture was heated at 110 °C for 24 h
and then cooled to rt. The mixture was poured into 200 mL of
ice-water and the solid precipitated was filtered. The solid was
washed with water, air-dried, collected and then dried in vacuo
to give iodide 12 as a slightly tan solid (1.77 g, 98%). 1H NMR
(CDCl3) d 8.07 (d, J = 2.4 Hz, 1H), 7.91 (s, 1H), 7.62–7.50 (m, 2H),
6.75 (d, J = 9.0 Hz, 1H), 3.95 (s, 3H). 13C NMR (CDCl3) d 155.4,
137.7, 134.4, 119.2, 113.1, 82.9, 67.1, 55.6. MS (ESI) calcd for
C10H9INO2 302.0 (M+H)+, found 302.0.
4.2.5. Iodide 13
Following procedures similar to those described for iodide 12,
bromide 10 (1.36 g, 5.35 mmol) was converted into iodide 13 as
a slightly tan solid (1.56 g, 97%). 1H NMR (CDCl3) d 7.92 (s, 1H),
7.56 (s, 1H), 7.48 (d, J = 7.8 Hz, 1H), 7.40 (d, J = 8.4 Hz, 1H), 7.30
(s, 1H), 3.96 (s, 3H). 13C NMR (CDCl3) d 155.8, 130.1, 127.2, 120.3,
116.8, 93.9, 67.1, 55.8. MS (ESI) calcd for C10H9INO2 302.0
(M+H)+, found 302.0.
4.2.10. Carboxylic acid 2
Following procedures similar to those described for carboxylic
acid 1, iodide 13 (452 mg, 1.50 mmol) and lactone 5 (128 mg,
1.01 mmol) underwent an allylic substitution to give carboxylic
acid 2 (E/Z ratio ꢄ 1.5:1) as a greenish syrup (107 mg, 35%). 1H
NMR (CDOD3) d 8.32 (br s, 1H), 7.90–7.50 (m, 2H), 6.96–6.74 (m,
2H), 5.45–5.32 (m, 1H), 3.93 (s, 3H), 3.44–3.34 (m, 2H), 2.62–
2.32 (m, 4H), 1.74 (s, 1.2H), 1.73 (s, 1.8H). HRMS calcd for
C17H18NO4 300.1235 (MꢁH)ꢁ, found 300.1238.
4.2.6. Iodide 14
Following procedures similar to those described for iodide 12,
bromide 11 (3.41 g, 12.0 mmol) was converted into iodide 14 as
a slightly tan solid (3.64 g, 92%). 1H NMR (CDCl3) d 8.05 (s, 1H),
7.80 (s, 1H), 7.36 (s, 1H), 6.41 (s, 1H), 3.90 (s, 3H), 3.86 (s, 3H).
13C NMR (CDCl3) d 158.9, 157.3, 149.2, 146.7, 135.8, 124.2, 112.1,
95.4, 74.7, 56.5, 55.7. MS (ESI) calcd for C11H11INO3 332.0
(M+H)+, found 332.0.
4.2.11. Carboxylic acid 3
Following procedures similar to those described for carboxylic
acid 1, iodide 14 (1.50 g, 4.53 mmol) and lactone 5 (382 mg,
3.03 mmol) underwent an allylic substitution to give carboxylic
acid 3 (E/Z ratio ꢄ 1.9:1) as a greenish syrup (404 mg, 40%). 1H
NMR (CDOD3) d 8.18 (br s, 1H), 7.47 (s, 0.3H), 7.46 (s, 0.7H), 7.42
(s, 1H), 6.66 (s, 1H), 5.38–5.28 (m, 1H), 3.96 (s, 3H), 3.88 (s, 3H),
3.29–3.23 (m, 2H), 2.50–2.28 (m, 4H), 1.72 (s, 3H). MS (ESI) calcd
for C18H20NO5 330.1 (MꢁH)ꢁ, found 330.1.
4.2.7. Iodide 16
To a mixture of commercially available 2,4-dimethoxy-3-meth-
ylbenzaldehyde (15, 3.13 g, 17.4 mmol) and silver trifluoroacetate
(4.60 g, 20.8 mmol) was added
a solution of iodine (5.30 g,
20.9 mmol) in CHCl3 (200 mL). After the addition was complete,
the mixture was allowed to stir for 4 h. Additional silver trifluoro-
acetate (770 mg, 3.48 mmol) and iodine (880 mg, 3.47 mmol) were
added and the mixture was allowed to stir overnight. The reaction
mixture was filtered through a pad of Celite, and the filtrate was
washed with saturated NaHCO3 (200 mL), a solution of Na2S2O3
(200 mL), water (2ꢃ 200 mL) and brine (200 mL). The organic layer
was dried over Na2SO4 and then filtered. The filtrate was concen-
trated and dried in vacuo to give iodide 16 as a pale solid (5.10 g,
96%). 1H NMR (CDCl3) d 10.19 (s, 1H), 8.14 (s, 1H), 3.88 (s, 3H),
3.83 (s, 3H), 2.32 (s, 3H). 13C NMR (CDCl3) d 188.1, 164.0, 163.4,
136.6, 127.3, 127.2, 87.0, 63.4, 60.5, 10.1. MS (ESI) calcd for
C10H12IO3 307.0 (M+H)+, found 307.0.
4.2.12. Carboxylic acid 4
Following procedures similar to those described for carboxylic
acid 1, iodide 17 (1.03 g, 2.99 mmol) and lactone 5 (378 mg,
3.00 mmol) underwent an allylic substitution to give carboxylic
acid 4 (E/Z ratio ꢄ 2.8:1) as a clear syrup (478 mg, 46%). 1H NMR
(CDCl3) d 7.94 (s, 0.25H), 7.94 (s, 0.75H), 7.53 (s, 0.25H), 7.52
(s, 0.75H), 7.39 (s, 0.75H), 7.38 (s, 0.25H), 5.40–5.33 (m, 1H), 3.75
(s, 2.2H), 3.74 (s, 0.8H), 3.71 (s, 3H), 3.42–3.36 (m, 2H), 2.56–2.46
(m, 3H), 2.42–2.36 (m, 1H), 2.28 (s, 3H), 1.77 (s, 3H). HRMS calcd
for C19H22NO5 344.1503 (MꢁH)ꢁ, found 344.1516.
4.2.13. HPLC separation of E/Z isomers of carboxylic acid 3
4.2.8. Iodide 17
The analytical HPLC was performed on a Varian Microsorb col-
A mixture of iodide 16 (3.58 g, 11.7 mmol), TosMIC (2.33 g,
11.9 mmol) and K2CO3 (4.05 g, 29.3 mmol) in anhydrous MeOH
(60 mL) was heated at reflux for 24 h. After cooling to rt, the mix-
ture was poured into 500 mL of ice-water and the solid precipi-
tated was filtered. The solid was washed with water, air-dried,
umn (C18, 5
the preparative HPLC was performed on a Varian Dynamax column
(C18, 8
cratic (21%) gradient of CH3CN (with 0.1% Et3N) in water (with
1% Et3N) was used. Combined fractions were pooled and the organ-
l
, 4.6 ꢃ 250 mm) with a flow rate of 0.5 mL/min while
l
, 41.4 ꢃ 250 mm) with a flow rate of 40 mL/min. An iso-

L. Chen et al. / Bioorg. Med. Chem. 16 (2008) 9340–9345
9345
ic solvent was removed. The aqueous layer (with a reduce volume)
Acknowledgments
was acidified with 1 N HCl to pH ꢄ 3. The individual solid precipi-
tate was filtered, washed with water, air-dried and dried in vacuo
to give carboxylic acid 3-Z (retention time = 9.6 min) and 3-E
(retention time = 12 min).
These studies were founded by the Center for Drug Design, Aca-
demic Health Center, University of Minnesota. We thank Dr. Court-
ney Aldrich for his help with the biological assays.
4.2.13.1. Carboxylic acid 3-E. 1H NMR (CDCl3) d 7.86 (s, 1H), 7.48
(s, 1H), 7.41 (s, 1H), 6.50 (s, 1H), 5.37 (t, J = 7.2 Hz, 1H), 3.95 (s, 3H),
3.88 (s, 3H), 3.30 (d, J = 7.2 Hz, 2H), 2.49 (t, J = 7.8 Hz, 2H), 2.37 (t,
J = 7.5 Hz, 2H), 1.74 (s, 3H). 13C NMR (CD3OD/CDCl3) d 177.3, 160.0,
157.0, 150.8, 150.1, 135.7, 127.4, 124.4, 123.3, 123.1, 109.8, 96.2,
56.3, 56.2, 35.8, 34.0, 28.7, 16.3. HRMS calcd for C18H20NO5
330.1335 (MꢁH)ꢁ, found 330.1355.
References and notes
1. Bentley, R. Chem. Rev. 2000, 100, 3801.
2. Chen, L.; Pankiewicz, K. W. Curr. Opin. Drug Discov. Dev. 2007, 10, 403.
3. Ratcliffe, A. J. Curr. Opin. Drug Discov. Dev. 2006, 9, 595.
4. Floryk, D.; Huberman, E. Cancer Lett. 2006, 231, 20.
5. Floryk, D.; Tollaksen, S. L.; Giometti, C. S.; Huberman, E. Cancer Res. 2004, 64,
9049.
6. Yu, J.; Lemas, V.; Page, T.; Connor, J. D.; Yu, A. L. Cancer Res. 1989, 49,
5555.
4.2.13.2. Carboxylic acid 3-Z. 1H NMR (CDCl3) d 7.87 (s, 1H), 7.50
(s, 1H), 7.41 (s, 1H), 6.50 (s, 1H), 5.36 (t, J = 7.2 Hz, 1H), 3.95 (s, 3H),
3.88 (s, 3H), 3.32 (d, J = 7.2 Hz, 2H), 2.51 (t, J = 7.2 Hz, 2H), 2.47 (t,
J = 7.2 Hz, 2H), 1.75 (s, 3H). 13C NMR (CD3OD) d 177.5, 160.2, 157.2,
151.2, 150.3, 135.8, 127.6, 125.8, 123.4, 123.4, 110.0, 96.5, 56.3,
56.2, 34.0, 28.6, 28.4, 23.4. HRMS calcd for C18H20NO5 330.1335
(MꢁH)ꢁ, found 330.1354.
7. Kiguchi, K.; Collart, F. R.; Henning-Chubb, C.; Huberman, E. Cell Growth Differ.
1990, 1, 259.
8. Chong, C. R.; Qian, D. Z.; Pan, F.; Wei, Y.; Pili, R.; Sullivan, D. J., Jr.; Liu, J. O. J. Med.
Chem. 2006, 49, 2677.
9. Gu, J. J.; Santiago, L.; Mitchell, B. S. Blood 2005, 105, 3270.
10. Ubukata, M.; Takamori, H.; Ohashi, M.; Mitsuhashi, S.; Yamashita, K.; Asada, T.;
Nakajima, N.; Matsuura, N.; Tsuruga, M.; Taki, K.; Magae, J. Bioorg. Med. Chem.
Lett. 2007, 17, 4767.
11. Chen, L.; Wilson, D.; Jayaram, H. N.; Pankiewicz, K. W. J. Med. Chem. 2007, 50,
6685.
12. Takebe, N.; Cheng, X.; Wu, S.; Bauer, K.; Goloubeva, O. G.; Fenton, R. G.;
Heyman, M.; Rapoport, A. P.; Badros, A.; Shaughnessy, J.; Ross, D.; Meisenberg,
B.; Tricot, G. Clin. Cancer Res. 2004, 10, 8301.
4.3. Enzyme assays
13. Papageorgiou, C. Mini-Rev. Med. Chem. 2001, 1, 71.
IMPDH inhibition assays were performed essentially as de-
scribed.29 Briefly, assays were set up in duplicate and contained
either IMPDH type 1 (100 nM) or type 2 (30 nM) in reaction buffer
14. Franklin, T. J.; Jacobs, V. N.; Jones, G.; Ple, P. Drug Metab. Dispos. 1997, 25, 367.
15. Lai, G.; Anderson, W. K. Tetrahedron 2000, 56, 2583.
16. El-Araby, M. E.; Bernacki, R. J.; Makara, G. M.; Pera, P. J.; Anderson, W. K. Bioorg.
Med. Chem. 2004, 12, 2867.
17. Pankiewicz, K. W.; Lesiak-Watanabe, K. B.; Watanabe, K. A.; Patterson, S. E.;
Jayaram, H. N.; Yalowitz, J. A.; Miller, M. D.; Seidman, M.; Majumdar, A.;
Prehna, G.; Goldstein, B. M. J. Med. Chem. 2002, 45, 703.
18. Chen, L.; Gao, G.; Felczak, K.; Bonnac, L.; Patterson, S. E.; Wilson, D.; Bennett, E.
M.; Jayaram, H. N.; Hedstrom, L.; Pankiewicz, K. W. J. Med. Chem. 2007, 50,
5743.
19. Sintchak, M. D.; Nimmesgern, E. Immunopharmacology 2000, 47, 163.
20. Dhar, T. G. M.; Shen, Z. Q.; Guo, J. Q.; Liu, C. J.; Watterson, S. H.; Gu, H. H.; Pitts,
W. J.; Fleener, C. A.; Rouleau, K. A.; Sherbina, N. Z.; McIntyre, K. W.; Witmer, M.
R.; Tredup, J. A.; Chen, B. C.; Zhao, R. L.; Bednarz, M. S.; Cheney, D. L.;
MacMaster, J. F.; Miller, L. M.; Berry, K. K.; Harper, T. W.; Barrish, J. C.;
Hollenbaugh, D. L.; Iwanowicz, E. J. J. Med. Chem. 2002, 45, 2127.
21. Ple, P. A.; Hamon, A.; Jones, G. Tetrahedron 1997, 53, 3395.
22. Lee, Y. M.; Fujiwara, Y.; Ujita, K.; Nagatomo, M.; Ohta, H.; Shimizu, I. Bull. Chem.
Soc. Jpn. 2001, 74, 1437.
(50 mM Tris, pH 8.0, 100 mM KCl, 1 mM DTT, 100
lM IMP, and
100 M NAD). Inhibitors in DMSO were added to UV clear 96-well
l
half-area plates (Greiner) to give a final DMSO concentration of 1%.
One hundred microliters of the reaction mixture was dispensed
into the plates immediately after the addition of enzyme. Reactions
were run at 25 °C and the production of NADH was monitored by
following changes in absorbance at 340 nm on a Molecular Devices
M5e multimode plate reader. Steady state velocities were used to
determine IC50 and KiApp values by fitting the velocities versus
inhibitor concentration to the sigmoidal concentration-response
curve (variable slope) and the Morrison equation,30 respectively,
using GraphPad Prism. For more accurate KiApp values when the
Morrison equation was used, enzyme concentration was fixed
23. Covarrubias-Zuniga, A.; Gonzalez-Lucas, A.; Dominguez, M. M. Tetrahedron
2003, 59, 1989.
24. Pope, S. A.; Burtin, G. E.; Clayton, P. T.; Madge, D. J.; Muller, D. P. Free Radical
Biol. Med. 2002, 33, 807.
using the specific activities of IMPDH
0.35 mol/(min mg), respectively.
I and II, 0.12 and
l
25. Klapars, A.; Buchwald, S. L. J. Am. Chem. Soc. 2002, 124, 14844.
26. Krasovskiy, A.; Knochel, P. Angew. Chem., Int. Ed. 2004, 43, 3333.
27. Nelson, P. H.; Carr, S. F.; Devens, B. H.; Eugui, E. M.; Franco, F.; Gonzalez, C.;
Hawley, R. C.; Loughhead, D. G.; Milan, D. J.; Papp, E.; Patterson, J. W.;
Rouhafza, S.; Sjogren, E. B.; Smith, D. B.; Stephenson, R. A.; Talamas, F. X.;
Waltos, A. M.; Weikert, R. J.; Wu, J. C. J. Med. Chem. 1996, 39, 4181.
28. Nelson, P. H.; Eugui, E.; Wang, C. C.; Allison, A. C. J. Med. Chem. 1990, 33, 833.
29. Umejiego, N. N.; Li, C.; Riera, T.; Hedstrom, L.; Striepen, B. J. Biol. Chem. 2004,
279, 40320.
30. Morrison, J. F. Biochim. Biophys. Acta 1969, 185, 269.
31. MacroModel. Schrodinger, LLC: New York, NY, 2006.
32. Qsite, version 4.5, Schrodinger, LLC: New York, NY, 2007.
33. Sintchak, M. D.; Fleming, M. A.; Futer, O.; Raybuck, S. A.; Chambers, S. P.; Caron,
P. R.; Murcko, M. A.; Wilson, K. P. Cell 1996, 85, 921.
34. Kaminski, G. A.; Friesner, R. A.; Tirado-Rives, J.; Jorgensen, W. L. J. Phys. Chem. B
2001, 105, 6474.
35. Kaminski, G. A.; Stern, H. A.; Berne, B. J.; Friesner, R. A.; Cao, Y. X.; Murphy, R. B.;
Zhou, R.; Halgren, T. A. J. Comput. Chem. 2002, 23, 1515.
36. Lee, C. T.; Yang, W. T.; Parr, R. G. Phys. Rev. B 1988, 37, 785.
37. Becke, A. D. J. Chem. Phys. 1993, 98, 5648.
4.4. Molecular modeling
Model construction has been described elsewhere.18 All calcula-
tions were completed with the Schrodinger computational chemis-
try software which includes MacroModel31 for molecular
mechanics and Qsite32 for quantum mechanics. The ligands were
initially built utilizing structural information from multiple
sources19,33 and subjected to a conformational search using Macro-
model and the OPLS forcefield.34,35 After 10,000 steps that probed
translations of the ligand as well as sampled all rotatable bonds
the global minimum and vast majority of conformations within
5.0 kcal/mol maintained the hexenoic acid tail hydrogen bonding
with Ser276. The global minimum was further optimized at the
38
B3LYP36,37/6-31G* level of theory and subject to the interaction
energy calculation as described.
38. Ditchfield, R.; Hehre, W. J.; Pople, J. A. J. Chem. Phys. 1971, 54, 724.