Discovery of benzodihydroisofurans as novel, potent, bioavailable and brain-penetrant prolylcarboxypeptidase inhibitors

Article abstract of DOI:10.1016/j.bmcl.2012.01.002

A series of benzodihydroisofurans were discovered as novel, potent, bioavailable and brain-penetrant prolylcarboxypeptidase (PrCP) inhibitors. The structure-activity relationship (SAR) is focused on improving PrCP activity and metabolic stability, and red

Full text of DOI:10.1016/j.bmcl.2012.01.002

Bioorganic & Medicinal Chemistry Letters 22 (2012) 1550–1556
Contents lists available at SciVerse ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Discovery of benzodihydroisofurans as novel, potent, bioavailable and
brain-penetrant prolylcarboxypeptidase inhibitors
Hong C. Shen a, , Fa-Xiang Ding a, Jinlong Jiang a, Andreas Verras a, Renee M. Chabin b, Suoyu Xu c,
⇑
Xinchun Tong c, Qing Chen c, Dan Xie d, Mike E. Lassman d, Urmi R. Bhatt d, Margarita M. Garcia-Calvo d,
Wayne Geissler d, Zhu Shen e, Beth Ann Murphy f, Judith N. Gorski f, Judyann Wiltsie g, Ranabir SinhaRoy d,
Jeffrey J. Hale a, Shirly Pinto e, Dong-Ming Shen a
a Department of Medicinal Chemistry, Merck Research Laboratories, PO Box 2000, Rahway, NJ 07065-0900, USA
b Department of Research Operations, Merck Research Laboratories, PO Box 2000, Rahway, NJ 07065-0900, USA
c Department of Drug Metabolism, Merck Research Laboratories, PO Box 2000, Rahway, NJ 07065-0900, USA
d Department of Diabetes, Merck Research Laboratories, PO Box 2000, Rahway, NJ 07065-0900, USA
e Department of Obesity, Merck Research Laboratories, PO Box 2000, Rahway, NJ 07065-0900, USA
f Department of Central Pharmacology, Merck Research Laboratories, PO Box 2000, Rahway, NJ 07065-0900, USA
g Department of In Vitro Sciences, Merck Research Laboratories, PO Box 2000, Rahway, NJ 07065-0900, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
A series of benzodihydroisofurans were discovered as novel, potent, bioavailable and brain-penetrant
prolylcarboxypeptidase (PrCP) inhibitors. The structure–activity relationship (SAR) is focused on improv-
ing PrCP activity and metabolic stability, and reducing plasma protein binding. In the established diet-
induced obese (eDIO) mouse model, compound ent-3a displayed target engagement both in plasma
and in brain. However, this compound failed to induce significant body weight loss in eDIO mice in a
five-day study.
Received 29 November 2011
Revised 2 January 2012
Accepted 3 January 2012
Available online 10 January 2012
Keywords:
Prolylcarboxypeptidase
PrCP
Ó 2012 Elsevier Ltd. All rights reserved.
Brain-penetrant
Inhibitors
Prolylcarboxypeptidase (PrCP) is a member of the single chain
serine peptidase S28 family,1 showing activity in lysosomes, the
membrane of human alveolar macrophages, plasma and urine.2
This enzyme is involved in inflammation and hypertension patho-
physiology. Regarding its roles in inflammation, the transient
abnormalities of PrCP are related to rheumatoid arthritis and ton-
sillitis.3 With respect to its roles related to hypertension, PrCP is
known to convert angiotensin II (AngII) to Ang 1–7 and angiotensin
III (AngIII) to Ang 2–7, both indicating PrCP’s potential role in mod-
ulating blood pressure and electrolyte balance.4 In addition, the
human gene for PrCP (11q14) may be associated with essential
hypertension.5 PrCP has also been identified as an endothelial cell
prekallikrein activator which may contribute to the basal plasma
prekallikrein activation and bradykinin liberation thereby regulat-
ing blood pressure.6 The strong biochemical and genetic evidence
of PrCP expression, inflammation, and hypertension suggests that
PrCP could be a drug target for the anti-inflammatory and anti-
hypertensive indications.
Besides the aforementioned therapeutic opportunities associ-
ated with PrCP, the burgeoning potential of PrCP inhibitors to treat
obesity, diabetes and related conditions has also drawn our atten-
tion.7 In particular, we employed a small molecule approach to in-
hibit PrCP and then examined target-dependent weight loss in
obesity models. Furthermore, we attempted to understand
whether central and/or peripheral PrCP inhibition is relevant to
any observed efficacy, as this enzyme is also located in brain.8
Different from the previous approach of substrate-based inhibi-
tor design,9 a high-throughput screening of the Merck sample col-
lection led to a variety of PrCP inhibitors.10 One hit (1a) that
caught our attention is shown in Figure 1. The ensuing structure–
activity relationship (SAR) study of each of the four different re-
gions (Ar1, fused heteroaryl, Ar2, and G, colored in red, orange, blue,
and purple, respectively) of the general scaffold (1) was conducted
in order to identify highly potent, bioavailable, and brain penetrant
PrCP inhibitors. Compounds with good PrCP inhibition in plasma
and brain would serve as tools to understand whether the PrCP
peripheral and central inhibition would affect food intake and/or
body weight in a rodent model of obesity. In all single enantiomer
cases, the listed data are for the eutomers, which are typically
⇑
Corresponding author.
E-mail address: hongshen@stanfordalumni.org (H.C. Shen).
0960-894X/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2012.01.002

H. C. Shen et al. / Bioorg. Med. Chem. Lett. 22 (2012) 1550–1556
1551
Table 1
In Vitro SAR on human and mouse PrCP inhibition and mouse AngIII assaysa
Ar1
G
Ar1
O
N
O
N
O
N
O
A
D
E
1a
G
Mouse AngIII IC50
(nM)
Compounds
PrCP IC50 (h, m) (nM)
Ar2
Cl
1
Cl
Cl
Cl
Ar1 =
Figure 1. The pharmacophore of benzodihydroisofuran PrCP inhibitors.
23 15 (2), 2.5
>1000
O
1a
several hundred to 1000-fold more potent than the corresponding
distomers.11
The in vitro PrCP inhibitory activity of compounds can be first
measured in the human and mouse PrCP enzyme inhibition assays,
and the mouse whole serum AngIII assay. In the AngIII assay,
mouse plasma was pre-treated with protease cocktail inhibitors
and ethylenediaminetetraacetic acid (EDTA) before the addition
of compounds and AngIII. The protease inhibitors and EDTA were
utilized to eradicate the background AngIII cleavage by other pro-
teases and metalloproteases. In a control study, it was found that
the protease cocktail inhibitors and EDTA had no influence on
the efficiency of AngIII cleavage by PrCP. As such, the mouse whole
serum AngIII assay was used to evaluate the serum shift of inhib-
itors in order to guide the selection of compounds for pharmacody-
namic (PD) studies.
OCF3
F
1.9 0.5 (5), 1.0 0.6
(3)
>1000
—
ent-1b
11 2 (2), 2.5 0.1
(2)
N
ent-1c
CF3
N
3.2 1.5, 0.71 0.10
41 7 (2), 17
474, —
144
—
N
1d
N
1e
The SAR study began with the optimization of the Ar1 group
(Table 1). Of numerous compounds prepared, 4-substituted benzyl
(ent-1b) and 4-substituted 2-pyridyl (ent-1c, 1d) analogs appeared
to be the most active inhibitors of PrCP. Benzyls with 2- or 3-sub-
stitutions and other pyridyl regioisomers resulted in a significant
loss of inhibitory activity for PrCP (data not shown). In addition,
4-substituents that provided the best potency against PrCP were
trifluoromethoxy (ent-1b), 4-fluoro (ent-1c), and trifluoromethyl
(1d) groups.
Despite the excellent PrCP inhibitory activity, one major issue of
these compounds was the extremely high serum shift, which was
manifest by the poor IC50’s of 1a and ent-1b in the mouse whole
serum AngIII assay. The introduction of the 2-pyridyl group (1d)
provided similar potency as ent-1b in the enzyme inhibition assay,
but the increased polarity of 1d apparently rendered improved
activity in the mouse serum AngIII assay (IC50 = 144 nM) relative
to ent-1b. To further reduce the serum shift, the attempts of incor-
porating a polar amine or a heteroaryl as the 4-substituent, as
exemplified by analogs 1e and 1f, typically led to a substantial
reduction of activity against PrCP.
N
N
—
N
1f
a
The number of experiments is shown in parenthesis and the listed values are
the average of multiple experiments, or from a single experiment if no parenthesis
is provided after the value.
Table 2
In Vitro SAR on human and mouse PrCP inhibition and mouse AngIII assaysa
OCF3
N
A
D
E
PrCP IC50 (h,
m) (nM)
Mouse AngIII
IC50 (nM)
Compounds
O
G
Cl
Cl
We then modified the aromatic group fused with dihydroisofu-
ran (Table 2). Three of the four possible pyridine regioisomers (2a–
2c) had single digit nanomolar IC50’s in the human and mouse PrCP
inhibition assays. Compounds 2b and 2c were significantly more
potent than 1a and ent-1b in the mouse serum AngIII assay, which
was consistent with the anticipation that higher compound polar-
ity led to less serum shift due to higher unbound fraction in
plasma.
A@N; D, E, G@CH: ent-2a
D@N; A, E, G@CH: 2b
E@N; A, D, G@CH: 2c
G@N; A, D, E@CH: 2d
3.2 2.0 (2),
1.4 0.1 (2)
3.4 0.6 (3),
1.3 0.1 (2)
4.5 1.8 (3),
5.2 1.0 (2)
169, —
—
380
427
—
a
The number of experiments is shown in parenthesis and the listed values are
the average of multiple experiments, or from a single experiment if no parenthesis
is provided after the value.
Various Ar2 groups were also examined (Table 3). The 5-chloro-
2-pyridyl analog ent-3a gave a subnanomolar IC50 in both human
and mouse PrCP inhibition assays. The good potency for enzyme
was translated to good mouse serum activity (IC50 = 45 nM). In
comparison, the relocation or the absence of the chloride substitu-
tion (3b or 3c) led to reduced potency against PrCP. One strategy to
further reduce the serum shift was to introduce N-oxide in the case
of ent-3d, which also showed excellent inhibition activity against
PrCP. Indeed, ent-3d had 0.6% and 0.7% free fraction in human
and mouse plasma, respectively, while ent-3a had <0.06% free frac-
tion in either human or mouse plasma. Finally, the combination of
two pyridyl groups led to analog 3e which had reasonable inhibi-
tory activity against human and mouse PrCP.

1552
H. C. Shen et al. / Bioorg. Med. Chem. Lett. 22 (2012) 1550–1556
Table 3
In Vitro SAR on human and mouse PrCP inhibition and mouse AngIII assaysa
OCF3
N
Compounds
PrCP IC50 (h, m) (nM)
Mouse AngIII IC50 (nM)
O
Ar2
Ar2 =
N
0.44 0.13 (12), 0.18 0.05 (11)
23 13 (2)
ent-3a
Cl
N
3b
54 32 (2), 25 8 (2)
—
Cl
84 44 (2), 20 5 (2)
—
—
N
3c
O-
ent-3d
N+
1.2 0.8 (5), 0.38 0.07 (2)
Cl
OCF3
N
N
O
15.6 0.5 (2), 5.8 0.3 (2)
—
3e
N
Cl
a
The number of experiments is shown in parenthesis and the listed values are the average of multiple experiments, or from a single
experiment if no parenthesis is provided after the value.
Table 4
In Vitro SAR on human and mouse PrCP inhibition and mouse AngIII assaysa
Compounds
PrCP IC50 (h, m) (nM)
Mouse AngIII IC50 (nM)
F
N
N
O
4a
1.0 0.4 (2), 1.0 0.3 (2)
10
N
Cl
O
S
O
OCF3
N
N
4b
1.6 0.8 (2), 1.8 0.9 (2)
712
O
Cl
Cl

H. C. Shen et al. / Bioorg. Med. Chem. Lett. 22 (2012) 1550–1556
1553
Table 4 (continued)
Compounds
PrCP IC50 (h, m) (nM)
Mouse AngIII IC50 (nM)
F
F
F
N
N
O
N
4c
9.5 2.2 (2), 10.6 0.5 (2)
1.9 1.5 (2), 1.1 0.1 (2)
0.73 0.28 (2), 0.31
19
N
Cl
N
N
O
N
ent-4d
14
Cl
Cl
N
N
O
4e
90
Cl
Cl
HO
N
O
N
4f
200 69 (2), 110 27 (2)
—
Cl
Cl
OCF3
N
O
N
4g
0.49 0.20 (2), 0.39
351
N
Cl
HO
N
O
N
4h
8.3 1.9 (2), 5.2
—
Cl
Cl
2 isomers
(continued on next page)

1554
H. C. Shen et al. / Bioorg. Med. Chem. Lett. 22 (2012) 1550–1556
Table 4 (continued)
Compounds
O
PrCP IC50 (h, m) (nM)
Mouse AngIII IC50 (nM)
O
S
CF3
N
N
O
4i
0.78 0.51 (2), 0.11
—
O-
N+
Cl
The number of experiments is shown in parenthesis and the listed values are the average of multiple experiments, or from a single
a
experiment if no parenthesis is provided after the value.
The last region subject to SAR study was linker G in Figure 1
(Table 4). It appeared that one methylene linker between the ter-
tiary amine nitrogen and Ar1 was optimal for good potency (data
not shown). A methyl substituent at the methylene linker boosted
the potency 10-fold (4e vs ent-1c). While tertiary amides were gen-
erally weak PrCP inhibitors (data not shown), aryl sulfonamides
(4b, 4i) were potent. Three approaches were adopted to improve
compound polarity in order to achieve better potency against PrCP
in serum. The first approach was to combine the pyridyl motifs
which were present in Ar1, Ar2 and the ring fused with dihydroi-
sofuran (4a, 4c, ent-4d, 4g, 4i). Indeed several analogs were potent
PrCP inhibitors in mouse serum (4a, 4c, ent-4d) further corroborat-
ing with the correlation of compound polarity and the activity
against enzyme in serum. Polar hydroxyl group was introduced
as the second approach (4f, 4h). While the primary alcohol 4f gave
inferior PrCP inhibition, the potency was largely restored by intro-
ducing the gem-dimethyl group (4h). The third strategy to increase
compound activity in serum was to introduce pyridine-N-oxide. To
our delight, compound 4i was 2- and 10-fold more active than 4b
in the human and mouse PrCP enzyme inhibition assays, respec-
tively. Furthermore, 4i had improved unbound concentration –
5% in human plasma and 2.5% in mouse plasma. In contrast, 4b
had <0.06% free fractions in both human and mouse plasma. To
block the potentially metabolically unstable benzylic position, a
gem-dimethyl moiety was incorporated in analog 4g, which pro-
vided excellent inhibitory activity against PrCP. However, the good
enzyme activity was not translated to satisfactory activity in
mouse serum as a result of 700-fold serum shift (0.5–351 nM).
Selected compounds were then submitted for mouse pharma-
cokinetics (PK) studies (Table 5). With the exception of compounds
ent-1b, 1e, ent-2a, and 3b, most compounds displayed high total
clearance (>30 mL/min/kg). Although analogs ent-1b, 1e, and 3b
had relatively low total clearance, the poor serum PrCP inhibition
of ent-1b and the suboptimal enzyme activity of 1e and 3b prohib-
ited them from further PD studies. Unfortunately, compounds 4a,
4c, and ent-4d did not possess sufficient oral bioavailability
(<10%) and exposure (<0.03), despite their excellent potency in
plasma. Analog ent-3a and two pyridine regioisomers ent-2a and
2b had similar PK profiles characterized by moderate to high clear-
ance, reasonable oral bioavailability (22–35%), oral exposure, and a
good half life (4–6.7 h). When compound ent-3a was tested in rats
and dogs, superior PK profiles were obtained as shown by the
excellent half life, oral exposure, oral bioavailability and low clear-
ance. Interestingly, the N-oxide analog ent-3d had higher oral bio-
availability and exposure than ent-3a in mice. Considering higher
unbound fraction of compound in mouse plasma for ent-3d
(0.7%) than ent-3a (<0.06%), the intrinsic clearance of ent-3d, which
could be estimated by total clearance divided by unbound fraction,
would be significantly lower than that of ent-3a.
Table 5
PK profiles of selected PrCP inhibitorsa
Compounds
F%
Cl
Vdss
(L/kg)
T1/2
(h)
AUCNpo
M h kg/mg)
(mL/min/kg)
(l
ent-1bd
ent-1c
1d
53
22
16
23
22
29
35
60
88
35
25
72
47
7.8
36
2.6
5.4
22
17
41
16
23
8.1
28
51
7.6
28
35
45
11
5.9
30
70
57
74
96
33
127
69
31
126
ND
38
87
4.4
16
11
4.7
8.2
15
20
5.9
5.5
14
26
27
28
13
9.5
17
19
14
20
ND
7.3
12
7.0
7.0
4.1
8.5
4.0
6.1
6.7
8.9
14
6.5
5.0
5.6
5.0
1.7
4.8
2.1
4.0
5.9
2.3
>8
2.16
1.36
0.11
1.22
0.26
0.28
0.28
1.8
1e
ent-2a
2b
ent-3ad
ent-3ab,d
ent-3ac,d
3b
5.0
0.46
0.14
0.43
0.23
0.032
0.33
0.008
0.028
0.27
0.05
0.41
0.11
0.10
3c
ent-3dd
3ed
4a
4b
4c
ent-4d
4e
4f
4g
4h
4i
3.4
2.1
a
PK studies were conducted in C57BL/6 mice, and with codosing of 4–5 different
compounds unless otherwise noted. Formulation: 0.1 mg/mL in ethanol/PEG300:-
water (15:60:25, v/v/v). IV dose: 0.5 mg/kg (mpk). PO dose: 1 mpk. Blood concen-
tration was determined by LC/MS/MS following protein precipitation with
acetonitrile. ND, not determined due to very slow clearance of compound in plasma.
b
Rat PK data.
c
Dog PK data. The IV dose at 1 mpk was formulated in EtOH/PEG400:Water
(10:40:50, v/v/v). The oral dose at 2 mpk was formulated in 0.5% methylcellulose.
d
The PK study was done with a single compound at 1 and 2 mpk IV and PO,
respectively. Formulation: 0.2 mg/mL in ethanol/PEG300/water (15:60:25, v/v/v).
In the human and mouse p-glycoprotein (pgp) transporter as-
says, a subset of benzodihydroisofurans (ent-3a, 3c, ent-3d, 4a,
4c, 4f, 4i) were tested and none exhibited BA/AB ratio of more than
2. These data suggested that this series were not pgp substrates
and thus they might exhibit good brain exposure. Compound ent-
3a was further tested in a CF-1 mouse PK study. Four hours post
a 3 mg/kg (mpk) dose (po), ent-3a gave a brain/plasma concentra-
tion ratio of 0.94 in the CF-1 wild-type mice and a ratio of 1.90 in
the pgpÀ/À mice. This result indicated that pgp was likely still in-
volved in the removal of ent-3a from the brain but this effect was
quite modest.
The reasonable PrCP inhibitory activity in mouse plasma and
acceptable mouse PK of ent-3a prompted us to examine its
ex vivo target engagement at different oral doses in established
diet-induced obese (eDIO) C57BL6 mice (3, 10, and 30 mpk, 2 days,
qd; n = 3; body weight: 52 g; age: 22 weeks). The plasma was ob-
tained from mice at the end of the study. The resulting plasma

H. C. Shen et al. / Bioorg. Med. Chem. Lett. 22 (2012) 1550–1556
1555
1200
1000
800
600
400
200
0
WT vehicle
KO vehicle
1
0
WT 30 mpk ent-3a
KO 30 mpk ent-3a
-1
Figure 3. The weight change of eDIO and PrCP knock out mice upon the treatment
of PrCP inhibitors (5 days, qd, po, KO@PrCP knock out).
3mpk
10mpk
30mpk
vehicle
Figure 2. Target engagement of compound ent-3a in eDIO mice (n = 3, po).
vested to measure the drug level and target engagement. At trough
Table 6
PK/PD relationship of ent-3a at different doses in eDIO mice (n = 3).
the drug level of ent-3a was 1.42
brain homogenate for the wild-type eDIO mice, and 1.57
plasma and 0.82
l
M in plasma and 0.69
l
l
M in
M in
Dose (
lM)
Blood
Plasma
Brain
PrCP %inh. in plasma
l
M in brain homogenate for the PrCP knock out
3 mpk
10 mpk
30 mpk
0.08
0.28
0.59
0.05
0.24
0.52
0.03
0.11
0.52
53
82
93
mice. Compound ent-3a demonstrated 88% ex vivo target engage-
ment in the plasma and 61% target engagement in brain at trough
in wild-type mice. The latter was measured by the cleavage of the
PrCP substrate in heat inactivated brain homogenate in a similar
manner as described for the plasma assay above.
The synthesis of analog ent-3a commenced with the addition of
ortho-lithium benzoic carboxylate to ketone 6 followed by a cycli-
zation to furnish lactone 7 (Scheme 1).12 The ensuing addition of
in situ-generated 2-pyridyl lithium to lactone 7 afforded hemiketal
9, which was then reduced employing trifluoroacetic acid and tri-
ethyl silane to provide intermediate 10. Finally a reductive amina-
tion and chiral separation yielded ent-3a in good overall yield.
In conclusion, a series of benzodihydroisofurans were discov-
ered as novel, potent, orally bioavailable and brain-penetrant PrCP
inhibitors. A representative compound ent-3a displayed an ade-
quate PK profile and acceptable brain exposure. Furthermore, this
analog exhibited good ex vivo target engagements in plasma and
brain of eDIO mice. However, ent-3a failed to induce significant
body weight loss in eDIO mice in a five-day study. This study, tak-
ing together with earlier results,9a,10 suggested that almost com-
plete inhibition of PrCP in plasma is required to show weight loss
in the eDIO model. Clarifying the role of PrCP inhibition in body
weight homeostasis would require even more potent inhibitors
in vivo which shall be the subject of future communications.
homogenate, diluted four fold, were separately heated to inactivate
the endogenous enzymes that could hydrolyze the substrate. To
the heat inactivated plasma were then added mouse PrCP and
the same PrCP substrate used in the in vitro assay. The PrCP activity
was then measured based on the conversion of the substrate. Fig-
ure ure2 showed that ent-3a effectively inhibited PrCP in a dose-
dependant manner after 2 days of 30 mpk oral dose. Table 6
showed that PrCP activity was inhibited by 53%, 82%, and 93%, at
3, 10, and 30 mpk doses, respectively. In addition, the drug levels
in plasma and brain were about the same at 30 mpk dose for two
days.
To ascertain the efficacy of compound ent-3a in reducing body
weight, a corhort of eDIO mice and a cohort of PrCP knock out mice
of similar body weight (n = 9) were orally dosed with ent-3a at
30 mpk for five days. Over the whole time course, there was no sta-
tistically significant difference in body weight between the treated
mice and vehicle in either the eDIO group or the PrCP knock out
group (Fig. 3). On the sixth day, both plasma and brain were har-
Boc
Boc
N
N
Br
c
a
b
O
O
OH
Br
CO2H
O
N Boc
O
N
N
5
6
7
Cl
Cl
8
9
H
N
OCF3
OCF3
N
O
N
O
d
f
O
OHC
OCF3
N
N
N
11
Cl
Cl
Cl
10
ent-3a
12
Scheme 1. (a) n-BuMgCl, n-BuLi, toluene, À10 °C, 1 h; AcOH, H2O, 40 °C, 16 h, 49%; (b) n-BuLi, toluene, À78 °C to rt, 16 h, 70%; (c) trifluoroacetic acid, triethylsilane, 72 °C, 3 h,
100%; (d) Na(OAc)3BH, THF, acetic acid, rt, 55%; (f) ChiralPak OD column, 70% isopropanol/heptane, the fast-eluting enantiomer, 44%.

1556
H. C. Shen et al. / Bioorg. Med. Chem. Lett. 22 (2012) 1550–1556
added. After stirring at À5 °C for 1 h, to the above solution was added acetic
References and notes
acid (15 mL) and water (40 mL). The resulting solution was heated at 40 °C for
16 h. The reaction mixture was extracted with methyl tert-butylether
(200 mL Â 2), and the combined organic phase was washed with NaOH (1 N,
400 mL), saturated K2CO3 solution (400 mL), brine (400 mL), and dried over
Na2SO4. After evaporating the volatile, the residue was treated with heptane
(100 mL) and the precipitate was collected by filtration to give 7 (7.3 g, 49%) as
white solid. To the solution of 7 (10 g, 52 mmol) in toluene (150 mL) was added
n-BuLi (22 mL, 2.5 M, 54 mmol) at À78 °C drop-wise. The resulting solution
was stirred at À78 °C for 0.5 h before the addition of 8 (7.9 g, 26 mmol) in one
portion. The reaction mixture was stirred from À78 °C to rt for 16 h, and was
then quenched by saturated NaHCO3 solution (300 mL). The mixture was
extracted with EtOAc (200 mL Â 3). The combined organic phase was dried
over Na2SO4 and concentrated. The residue was purified on silica gel using 10–
40% EtOAc/hexane to give 9 (7.5 g, 69%) as an oil. The solution of 9 (7.5 g,
18.0 mmol), triethylsilane (5.75 mL, 36.0 mmol) in trifluoroacetic acid (70 mL)
was heated at reflux for 3 h. After removing the volatile, the residue was
dissolved in CH2Cl2 (300 mL) and extracted with aqueous HCl (1 N,
250 mL Â 4). The acidic aqueous phase was then basified by NaOH (5 N) to
pH 10, and then extracted with CH2Cl2 (500 mL Â 3). The combined organic
phase was dried over Na2SO4, and concentrated to give 10 (5.54 g, 100%) as an
oil. To the solution of 10 (5.54 g, 18.4 mmol) in THF/acetic acid (4/1, 250 mL)
was added 11 (4.20 g, 22.10 mmol) and sodium triacetoxyborohydride (11.7 g,
55.3 mmol). The resulting mixture was stirred at rt for 3 h before being
quenched by dropwise addition of water (50 mL). After removing the volatile,
the residue was diluted with CH2Cl2 (300 mL), washed with NaOH (1 N,
300 mL) and saturated NaHCO3 solution (300 mL), and then dried over Na2SO4.
After removing the solvent in vacuo, the residue was purified on silica gel using
20–40% EtOAc/hexane to afford racemate 12 (4.8 g, 55%) as an oil. The
racemate 12 (4.8 g, 10.11 mmol) was separated on a chiral OD column using
70% isopropanol/heptane to provide ent-3a (2.09 g, 44%, the fast eluting peak is
the more potent isomer against PrCP) as a colorless oil. 1H NMR (acetonitrile-
d3, 500 MHz) = 8.54–8.53 (d, J = 5.4 Hz, 1 H), 7.50–7.48 (m, 3 H), 7.35–7.25 (m,
7 H), 6.21 (s, 1 H), 3.62 (s, 2 H), 2.87–2.80 (m, 2 H), 2.59–2.46 (m, 2 H), 2.22–
2.16 (m, 2 H), 1.96–1.90 (m, 2 H); m/z: calcd for C25H22ClF3N2O2: 474.13; found
[M+H+]: 475.22 (100%), 477.18 (33%).
1. Tan, F.; Morris, P. W.; Skidgel, R. A.; Erdos, E. G. J. Biol. Chem. 1993, 268, 16631.
2. Jackman, H. L.; Tan, F.; Schraufnagel, D.; Dragovic, T.; Dezso, B.; Becker, R. P.;
Drdos, E. G. Am. J. Respir. Cell. Mol. Biol. 1995, 13, 196.
3. (a) Kumamoto, K.; Stewart, T. A.; Johnson, A. R.; Erdos, E. G. J. Clin. Invest. 1981,
67, 210; (b) Palmeri, D.; Zuo, F. R.; Rosen, S. D.; Hemmerich, S. J. Leukoc. Biol.
2004, 75, 910; (c) Wang, L.; Feng, Y.; Zhang, Y.; Zhou, H.; Jiang, S.; Niu, T.; Wei,
L.-J.; Xu, X.; Xu, X.; Wang, X. Am. J. Obstet. Gynecol. 2006, 195, 162.
4. (a) Odya, C. E.; Marinkovic, D. V.; Hammon, K. J.; Stewart, T. A.; Erdos, E. G. J.
Biol. Chem. 1978, 253, 5927; (b) Mallela, J.; Yang, J.; Shariat-Madar, Z. Int. J.
Biochem. Cell Biol. 2009, 41, 477.
5. Watson, B.; Nowak, N. J.; Myracle, A. D.; Shows, T. B.; Warnock, D. G. Genomics
1997, 44, 365.
6. (a) Shariat-Madar, Z.; Mhdi, F.; Schmaier, A. H. J. Bioorg. Chem. 2002, 277,
17962; (b) Moreira, C. R.; Schmaier, A. H.; Mahdi, F.; da Motta, G.; Nader, H. B.;
Shariat-Madar, Z. FEBS Lett. 2002, 523, 167.
7. Horvath, T.; Diano, S.; Gao, Q.; Warden, C. H. WO2005115446.
8. Palmiter, R. D. J. Clin. Invest. 2009, 119, 2130.
9. (a) Zhou, C.; Garcia-Calvo, M.; Pinto, S.; Lombardo, M.; Feng, Z.; Bender, K.;
Pryor, K. D.; Bhatt, U. R.; Chabin, R. M.; Geissler, W. M.; Shen, Z.; Tong, X.;
Zhang, X.; Wong, K. K.; Roy, S. S.; Chapman, K. T.; Yang, L.; Xiong, Y. J. Med.
Chem. 2010, 53, 7251; (b) Shen, H. C.; Ding, F.-X.; Verras, A.; Chabin, R. M.; Xu,
S.; Tong, X.; Xie, D.; Lassmand, M. E.; Bhattd, U. R.; Garcia-Calvo, M. M.;
Geissler, W.; Shen, Z.; Chen, D.; Sinha-Roy, R.; Hale, J. J.; Tata, J. R.; Pinto, S.;
Shen, D.-M.; Colletti, S. L. Bioorg. Med. Chem. Lett. 2011, 21, 1299.
10. Grahama, T. H.; Shen, H. C.; Liu, W.; Xiong, Y.; Verras, A.; Bleasby, K.; Bhattd, U.
M.; Chabin, R. M.; Chen, D.; Chen, Q.; Garcia-Calvo, M. M.; Geissler, W. M.; He,
H.; Lassman, M. E.; Shen, Z.; Tong, X.; Tung, E. C.; Xie, D.; Xu, S.; Colletti, S. L.;
Tata, J. R.; Hale, J. J.; Pinto, S.; Shen, D.-M. Bioorg. Med. Chem. Lett. 2011, 22, 658.
11. It should be noted that throughout the paper compounds are racemic unless
‘ent’ is designated before the compound number.
12. Experimental Procedures: To the solution of 5 (9.79 g, 48.7 mmol) in THF
(50 mL) at À10 °C was added BuMgCl (21.9 mL, 2.0 M, 43.8 mmol) followed by
drop-wise addition of n-BuLi (25.3 mL, 2.5 M, 63.3 mmol). The resulting
solution was stirred at À10 °C for 30 min before 6 (10.67 g, 53.6 mmol) was