Discovery of potent, selective 4-fluoroproline-based thrombin inhibitors with improved metabolic stability

Article abstract of DOI:10.1016/j.bmc.2006.06.040

Previous reports from our laboratories described potent tripeptide thrombin inhibitors which incorporate heterocycle-substituted chlorophenyl groups in the P1 position. Using these as lead compounds for further optimization, we identified sites of metabolism and designed analogs with 4-fluoroproline in P2 and cyclopropane-containing side chains in P3 as an approach to reducing metabolism and improving their oral pharmacokinetic performance. The large (300-fold) difference in potency between analogs containing (4R)- and (4S)-4-fluoroproline was rationalized by analyzing inhibitor-enzyme interactions in crystal structures of related compounds and by molecular modeling which indicated that the more potent (4R)-4-fluoroproline isomer stabilizes a proline ring conformation that is preferred for binding to the enzyme. An optimal compound from this work, 41, exhibits high potency in a coagulation assay in human plasma (2×APTT = 190 nM), excellent selectivity versus the digestive enzyme trypsin (K_i = 3300 nM), and excellent oral bioavailability in dogs with moderate clearance (F = 100%, CL = 12 mL/min/kg).

Full text of DOI:10.1016/j.bmc.2006.06.040

Bioorganic & Medicinal Chemistry 14 (2006) 6900–6916
Discovery of potent, selective 4-fluoroproline-based
thrombin inhibitors with improved metabolic stability
Donnette D. Staas,a,* Kelly L. Savage,a Vanessa L. Sherman,a Heidi L. Shimp,a
Terry A. Lyle,a Lekhanh O. Tran,a Catherine M. Wiscount,a Daniel R. McMasters,a
Philip E. J. Sanderson,a Peter D. Williams,a Bobby J. Lucas, Jr.,b Julie A. Krueger,b
S. Dale Lewis,b Rebecca B. White,c Sean Yu,c Bradley K. Wong,c
Christopher J. Kochansky,c M. Reza Anari,c Youwei Yand and Joseph P. Vaccaa
aDepartment of Medicinal Chemistry, Merck Research Laboratories, West Point, PA 19486, USA
bDepartment of Biological Chemistry, Merck Research Laboratories, West Point, PA 19486, USA
cDepartment of Drug Metabolism, Merck Research Laboratories, West Point, PA 19486, USA
dDepartment of Structural Biology, Merck Research Laboratories, West Point, PA 19486, USA
Received 4 May 2006; revised 10 June 2006; accepted 19 June 2006
Available online 25 July 2006
Abstract—Previous reports from our laboratories described potent tripeptide thrombin inhibitors which incorporate heterocycle-
substituted chlorophenyl groups in the P1 position. Using these as lead compounds for further optimization, we identified sites
of metabolism and designed analogs with 4-fluoroproline in P2 and cyclopropane-containing side chains in P3 as an approach to
reducing metabolism and improving their oral pharmacokinetic performance. The large (300-fold) difference in potency between
analogs containing (4R)- and (4S)-4-fluoroproline was rationalized by analyzing inhibitor–enzyme interactions in crystal structures
of related compounds and by molecular modeling which indicated that the more potent (4R)-4-fluoroproline isomer stabilizes a pro-
line ring conformation that is preferred for binding to the enzyme. An optimal compound from this work, 41, exhibits high potency
in a coagulation assay in human plasma (2·APTT = 190 nM), excellent selectivity versus the digestive enzyme trypsin
(Ki = 3300 nM), and excellent oral bioavailability in dogs with moderate clearance (F = 100%, CL = 12 mL/min/kg).
ꢀ 2006 Elsevier Ltd. All rights reserved.
1. Introduction
of novel heterocycle-substituted chlorophenyl P1 groups
and their incorporation into this classical tripeptide scaf-
fold, giving rise to potent inhibitors such as 1 and 2.2
The serine protease thrombin is a critical trypsin-like
enzyme in the blood coagulation cascade. Thrombin
mediates the conversion of fibrinogen to fibrin and is also
the most potent stimulator of platelet aggregation known.
A direct thrombin inhibitor would therefore be expected
to have therapeutic applications in the treatment of
disease states involving undesirable clot formation,
including the treatment of deep vein thrombosis and
pulmonary embolism. We and others have previously
described the development of potent thrombin inhibitors
based on a D-Phe-Pro-Arg tripeptide motif.1 Recent
reports from these laboratories described the discovery
Cl
P2
H
N
N
O
N
N
X
O
N
NH2
P3
P1
1 (X = CH): Ki (thrombin) = 1.8 nM
2 (X = N): Ki (thrombin) = 0.33 nM
When dosed orally in dogs at 1 mg/kg, 1 and 2 exhibited
modest plasma levels (AUC = 0.6 and1 lM h,respectively)
with short half-lives (1.7 and 2.5 h, respectively). Incuba-
tion of these compounds with microsomes obtained from
rat, dog, monkey, and human liver tissues indicated that
Keywords: Thrombin inhibitor; Fluoroproline; Metabolic stability;
Anticoagulant.
*
Corresponding author. Tel.: +1 215 652 5452; fax: +1 215 652
3971; e-mail: donnette_staas@merck.com
0968-0896/$ - see front matter ꢀ 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2006.06.040

D. D. Staas et al. / Bioorg. Med. Chem. 14 (2006) 6900–6916
6901
no metabolism had occurred on the triazole moiety in P1;
however, extensive metabolic oxidation of the central P2
proline and the P3 side chain had occurred in all species.
As an approach to improving the pharmacokinetic prop-
erties of 1 and 2, we therefore focused our efforts on
designing analogs which had the potential for greater sta-
bility toward oxidative metabolism in the P2–P3 portion
of the molecule. Fluorination has been shown to be an
effective strategy for reducing oxidative metabolism of
susceptible functional groups,3 and therefore we sought
to improve the metabolic stability in P2 by incorporating
4-fluoroproline, the preparation of which has been de-
scribed in the literature.4 Additionally, we sought to
incorporate novel amino acids in P3 which were
designed to improve the metabolic stability of the side
chain. This manuscript describes the synthesis, conforma-
tional properties, and biological profiles of analogs of 1
and 2 containing such modifications.
appropriately functionalized Boc-protected prolines.
Deprotection of the intermediate prolinamides followed
by a second amino acid coupling to the protected P3
amino acids and removal of the protecting group affor-
ded the desired targets. The 4-fluoroproline P2 amino
acids were prepared according to procedures previously
described in the literature.4 The P2 amino acid Boc-L-
proline and the P3 amino acid Boc-tert-butyl-D-alanine
are commercially available, and the remaining amino
acids incorporated into our inhibitor scaffold were
synthesized from readily available starting materials as
described below.
The syntheses of the novel Boc-3-(1-methylcyclopropyl)-
D-alanine 6, (2R)-2-azido-3-(1-chlorocyclopropyl)propa-
noic acid (13, precursor to the chlorocyclopropyl alanine
P3 side chain), and Boc-cyclopropylvaline 20 have not
been previously reported in the literature. We devised
a three-step route to 6 starting with alkylation of the
Williams chiral auxiliary 35 with methallyl bromide.
Cyclopropanation of the resulting olefin 4 and reductive
removal of the auxiliary afforded acid 6 (Scheme 1).
2. Chemistry
Synthesis of the compounds listed in Tables 1 and 2
began with assembly of the P1–P2 cores via standard
amino acid coupling of the P1 benzylamines2 to the
Asymmetric synthesis of (2R)-2-azido-3-(1-chlorocyclo-
propyl)propanoic acid 13 was accomplished in several
Table 1. P3 amino acid analogs
Cl
H
N
N
O
N
N
R
O
X N
NH2
Compound
R
X
Thrombin
Ki (nM)
2·APTT
(nM)
Trypsin
Ki (nM)
Trypsin/2·
APTT ratio
Dog Oral
AUC (lM h)*
Dog Oral
t1/2 (h)
1
2
CH
N
1.8
430
230
210
140
170
100
4100
590
9.5
2.6
45
0.57a
0.99b
2.1c
1.8d
2.8e
1.1
1.7
2.5
2.1
2.6
3.4
3.6
0.31
0.61
0.09
0.60
0.08
37
38
39
40
CH
N
9500
1400
6200
850
10
CH
N
36
8.5
* Values based on 1 mpk dose; see notes below.
a Data normalized to 1 mpk dose. Raw data: dose = 0.65 mpk, AUC = 0.37 lM h.
b Data normalized to 1 mpk dose. Raw data: dose = 0.75 mpk, AUC = 0.74 lM h.
c Data normalized to 1 mpk dose. Raw data: dose = 0.85 mpk, AUC = 1.78 lM h.
d Data normalized to 1 mpk dose. Raw data: dose = 0.85 mpk, AUC = 1.51 lM h.
e Data normalized to 1 mpk dose. Raw data: dose = 0.90 mpk, AUC = 2.55 lM h.

6902
D. D. Staas et al. / Bioorg. Med. Chem. 14 (2006) 6900–6916
Table 2. P2 fluoroproline analogs
Cl
Y
Z
H
N
N
O
N
R
N
O
X N
NH2
Compound
R
X
Y
H
Z
Thrombin
Ki (nM)
2·APTT
(nM)
Trypsin
Ki (nM)
Trypsin/2·
APTT ratio
Dog Oral
AUC (lM h)*
Dog Oral
t1/2 (h)
39
CH
H
0.60
0.08
0.37
110
170
6200
36
8.5
17
—
—
2.7
24
14
2.8a
1.1
5.0
—
3.4
3.6
2.8
—
40
N
H
H
F
H
F
H
F
F
F
F
100
190
—
850
3300
41
CH
CH
CH
N
42
26,000
4600
43
F
3.6
—
—
—
44
F
0.55
1.3
190
540
260
520
2.8b
4.5
1.5
4.0
2.0
1.4
45a13
46a13
CH
CH
H
H
13,000
3600
0.32
* Values based on 1 mpk dose; see notes below.
a Data normalized to 1 mpk dose. Raw data: dose = 0.90 mpk, AUC = 2.55 lM h.
b Data normalized to 1 mpk dose. Raw data: dose = 0.85 mpk, AUC = 2.39 lM h.
O
O
O
O
O
O
Br
b
O
N
O
N
O
N
Ph
Ph
Ph
a
O
Ph
3
O
Ph
4
O
Ph
5
O
c
OH
O
NH
O
6
Scheme 1. Reagents and conditions: (a) NaHMDS, THF, À70 ꢁC; (b) CH2N2, Pd(OAc)2, THF/Et2O, 0 ꢁC; (c) Li/NH3(l), THF.
steps starting from commercially available 2-chloroacry-
lonitrile 7 (Scheme 2). Treatment of 7 with an ethereal
solution of diazomethane at 0 ꢁC afforded the intermedi-
ate dihydropyrazole adduct 8 which was thermally
decomposed to yield 1-chlorocyclopropanecarbonitrile
9. DIBAL reduction of nitrile 9 followed by Wittig
olefination of the resulting aldehyde with (carbometh-
oxymethylene)-triphenylphosphorane afforded the a,

D. D. Staas et al. / Bioorg. Med. Chem. 14 (2006) 6900–6916
6903
Cl
7
Cl
CN
Cl
CN
Cl
c
b
a
OMe
CN
N
N
O
8
9
10
Ph
N
Cl
Cl
N3
Cl
f
d
e
OH
OH
O
O
O
O
O
11
13
12
Scheme 2. Reagents and conditions: (a) MNNG, 40% aq KOH, Et2O, 0 ꢁC; (b) benzene, reflux; (c) i—DIBAL, toluene À78–0 ꢁC;
ii—Ph3P@CHCO2Me, toluene, rt; (d) i—H2 (balloon), 10% Pd–C, MeOH, rt; ii—LiOHÆH2O, 1:1:1 THF/MeOH/H2O, rt; (e) i—(CH3)3CC(O)Cl,
Et3N, THF, À20 ꢁC; ii—(4R)-4-benzyl-1,3-oxazolidin-2-one, LiCl, rt; (f) i—KHMDS, toluene, À78 ꢁC; ii—trisyl azide, warm to 30 ꢁC;
iii—LiOHÆH2O, 2:2:1 THF/MeOH/H2O, 0 ꢁC to rt.
b-unsaturated ester 10. Catalytic hydrogenation of the
double bond of 10 followed by ester hydrolysis with
lithium hydroxide in an aqueous THF–methanol
mixture afforded 3-(1-chlorocyclopropyl)propanoic acid
11. Acid 11 was coupled to the chiral auxiliary (4R)-4-
benzyl-1,3-oxazolidin-2-one to produce 12 which
underwent stereoselective azidation with potassium
bis(trimethylsilyl)amide and triisopropylbenzene sulfo-
nyl azide.6 Removal of the chiral auxiliary under basic
conditions afforded the azido acid 13, which was subse-
quently coupled to the P1–P2 core prior to reduction of
the azide to the desired P3 amine.
We initially designed an asymmetric synthesis of Boc-
cyclopropylvaline 20 via the known chiral sultam oxime
precursor 15 (Scheme 3).7 Removal of the chiral auxilia-
ry of 15 and cyclopropanation of the resulting interme-
diate 16 would then have afforded the desired amino
acid core. However, attempts to cleave the chiral auxil-
iary of 15 under basic,8 Lewis acidic9 or phase transfer
conditions10 were unsuccessful in our hands. We consis-
tently observed either no reaction or the formation of
products which appeared to arise from the reaction of
nucleophiles (e.g., hydroxide or peroxide) at the sulfur
atom of 15 resulting in ring opening of the sultam,
O
O
O
Br
OH
X
N
N
NH
S
S
a
BnO
O2
O2
N
NH
BnO
BnO
O
15
14
O
16
O
O
NHBoc
c
b
d
O
OH
NHBoc
OMe
OH
NHBoc
NHBoc
17
18
19
20
c, d
O
O
O
O
g, h
e, f
H
Br
OH
OCH3
OMe
OMe
a
O
N
NH
23
NHBoc
24
BnO
BnO
21
22
Scheme 3. Reagents and conditions: (a) Zn, satd aq NH4Cl, THF, rt; (b) LDA, TMSCl, THF, À78 ꢁC then 60 ꢁC; (c) CH2N2, Et2O, Pd(OAc)2, 0 ꢁC
to rt; (d) LiOHÆH2O, THF–H2O, 40 ꢁC; (e) BnONH2ÆHCl, NaOAc, 1:1 MeOH–H2O, rt; (f) DCC, MeOH, DMAP, CH2Cl2, 0 ꢁC to rt; (g) Mo(CO)6,
CH3CN, H2O, reflux; (h) Boc2O, CH3CN, 0 ꢁC to rt.

6904
D. D. Staas et al. / Bioorg. Med. Chem. 14 (2006) 6900–6916
rather than from the desired reaction at the carbonyl
center of 15 to liberate the free acid 16 and the sultam
auxiliary. We attributed these results to steric shielding
imposed on the carbonyl carbon of 15 by both the adja-
cent neopentyl-like a-carbon center and the bulky
sultam.
by basic hydrolysis afforded the desired protected cyclo-
propylvaline 20.
Standard EDC coupling of Boc-L-proline 25 or the
Boc-protected 4-fluoroprolines 26,4 27,4 or 284 to the
benzylamines 29 or 30 (prepared as described in Young
et al.2) followed by HCl-mediated cleavage of the Boc
group afforded the P1–P2 cores 31–36 as hydrochloride
salts (Scheme 4). EDC-mediated coupling of the protect-
ed P3 amino acids Boc-tert-butyl-D-alanine, 6 or 20 to
intermediates 31–36 followed by TFA or HCl-mediated
removal of the Boc group afforded the target
compounds 37–44 and 46. The diastereomers of 46 were
resolved by chiral chromatography to afford compound
46a (the more potent diastereomer) and its isomer 46b.13
Compound 45 was prepared via EDC coupling of the
azido acid 13 to the P1–P2 intermediate 33 followed
by triphenylphosphine-mediated reduction of the azide
to the amine. Some epimerization occurred during the
reduction and the diastereomeric amines were separated
by chiral chromatography to afford 45a (the more
potent diastereomer) and its isomer 45b.13
We thus resorted to an achiral synthesis of 20 via an
ester enolate Claisen rearrangement11 of 17 (prepared
by DCC-mediated coupling of Boc-glycine and 3-meth-
ylbut-2-en-1-ol) to give 2-[(tert-butoxycarbonyl)amino]-
3,3-dimethylpent-4-enoic acid 18 (Scheme 3). Treatment
of 18 with excess diazomethane in the presence of
Pd(OAc)2 afforded methyl ester 19. Hydrolysis of ester
19 to give the desired racemic Boc-cyclopropylvaline
20 required heating, which was consistent with our
earlier hypothesis of reduced reactivity at the carbonyl
carbon of this scaffold due to steric shielding by the
adjacent neopentyl-like a-carbon.
Alternatively, acid 20 could be prepared via an achiral
variant of the oxime chemistry described above7 in much
higher overall yield than the Claisen route. Thus,
glyoxylic acid 21 was condensed with O-benzylhydroxyl-
amine and the resulting benzyloxyiminoacetic acid was
esterified by treatment with DCC and methanol to give
methyl ester 22. Imino ester 22 was then treated with
methallyl bromide in the presence of zinc and saturated
aqueous ammonium chloride solution in THF to obtain
key olefin intermediate 23, which underwent molybde-
num hexacarbonyl-mediated removal of the benzyloxy
group12 followed by Boc protection of the intermediate
amino ester to give 24. Cyclopropanation of 24 followed
3. Biological methods
Inhibition constants (Ki values) of test compounds were
determined in an in vitro assay using human thrombin
and trypsin as previously described.14,15 The 2·APTT
value is defined as the concentration of test compound
required to double the activated partial thromboplastin
time of human plasma, and was determined using
previously described methods.15 Pharmacokinetic
Cl
Cl
Y
Y
Z
Z
H
a, b
H2N
N
+
OH
N
H
N
N
O
N
Boc
O
N
N
X
N
X N
25 Y = Z = H
26 Y = H; Z = F
27 Y = F; Z = H
28 Y = Z = F
29 X = CH
30 X = N
31 X = CH; Y = Z = H
32 X = N; Y = Z = H
33 X = CH; Y = H; Z = F
34 X = CH; Y = F; Z = H
35 X = CH; Y = Z = F
36 X = N; Y = Z = F
Cl
N
Y
Z
c; b or d or e
H
N
N
O
R
O
N
NH2
X N
Compounds 37-46;
see Tables 1 and 2
Scheme 4. Reagents and conditions: (a) 25 or 26 or 27 or 28, EDC, HOAT, DMF, rt; (b) 4 M HCl–dioxane, EtOAc, 0 ꢁC; (c) 6 or 13 or 20 or
Boc-tert-butyl-D-alanine, EDC, HOAT, DIEA, DMF, rt; (d) TFA, CH2Cl2, rt; (e) (Compound 45 only) i—Ph3P, THF, 0 ꢁC; ii—H2O, 40 ꢁC, 2 h
then rt, 66 h.

D. D. Staas et al. / Bioorg. Med. Chem. 14 (2006) 6900–6916
6905
studies were carried out in conscious male beagle dogs
(n = 2) weighing 10–12 kg. Test compounds were dosed
orally as amorphous trifluoroacetate or hydrochloride
salts or free bases, either as a solution or as a suspension
in 0.5% methocel, 1% methocel or 5% dextrose-PEG200;
data were subsequently normalized to a 1 mg/kg dose.
Blood samples were collected via the jugular vein at reg-
ular intervals over an 8 h time period, with a final sam-
ple taken 24 h post-dosing. Plasma samples were
assayed by solid-phase extraction followed by quantita-
tive LCMS/MS analysis against a standard curve.
Metabolite identification studies were carried out by
incubation of rat, dog, monkey or human liver micro-
somes (4 mg/mL) with 10 lM solutions of test com-
pound, 100 mM potassium phosphate buffer, 6 mM
magnesium chloride, and 1 mM EDTA in the presence
of either 1 mM NADPH or 1 mM NADPH + 1 mM
UDPGA at 37 ꢁC for 2 h with monitoring via LCMS/
MS analysis. Standard error in the measurement of
thrombin and trypsin Ki values, 2·APTT values, and
oral AUC values is 20%.
amino acid methylcyclopropylalanine in P3. The potency
and selectivity of methylcyclopropanes 39 and 40 were
the same as the corresponding tert-butyl analogs 37
and 38, and there was no significant change in the oral
PK properties in dogs.
Exposure of 39 to microsomes from rat, dog, monkey,
and human liver tissues indicated that oxidative metab-
olism was still occurring on the P3 side chain (including
oxidation of the methyl group to a carboxylic acid) and
on the P2 proline ring (Fig. 1). To follow up on these
observations, we focused first on P2 metabolism and
prepared analogs of 39 containing one or two fluorines
at the proline ring 4-position (41–43, Table 2).
We were intrigued to find a nearly 300-fold difference in
potency for inhibiting thrombin when comparing the
two diastereomeric monofluoroproline derivatives 41
and 42, with the (4S)-4-fluoroproline-derived analog
42 being almost 200-fold less potent than the des-fluoro
analog 39. A possible explanation for the loss in potency
of 42 compared to 39 can be gleaned from the crystal
structure of the related P2 proline-based inhibitor 4719
bound in the active site of thrombin (Fig. 2). In this
structure, it is seen that the pro-(S) hydrogen at the pro-
line ring 4-position of 47 is near to the aromatic side
chains of residues Y60A and W60D, well within distance
4. Results and discussion
As previously described,2 thrombin inhibitors contain-
ing triazole- and tetrazole-substituted chlorophenyl P1
groups such as 1 and 2 exhibited high levels of selectivity
(>1000-fold) for inhibiting thrombin versus the digestive
serine protease, trypsin, in enzyme assays. In looking at
the ratio of the Ki for inhibiting trypsin and the func-
tional potency for inhibiting thrombin, that is, the
2·APTT value, however, the levels of selectivity are
much lower. For example, the ratios of Ki values for
inhibiting trypsin versus thrombin for 1 and 2 are
2300 and 1900, respectively; whereas the ratios using
the 2·APTT values are 9.5 and 2.6 (see Table 1). Since
the 2·APTT concentration approximates the plasma
concentration required for therapeutic efficacy, we were
interested in maintaining as large a window as possible
between the 2·APTT potency and the Ki for inhibiting
trypsin, and therefore express selectivity in this way.
20
˚
(3.4–3.7 A) to be engaged in a CH-p-type interaction.
Substitution of fluorine for this hydrogen would then
replace the weakly attractive CH-p interaction with a
repulsive CF-p interaction.21
Cl
H
N
O
N
O
O
NHEt
O
47
NH2
We were also interested in the effect of fluorination
on proline ring conformation. The proline ring can
adopt two main puckered conformations as depicted in
Table 3, namely a ‘syn’ ring pucker, in which C4 points
in the same direction as the substituent at C2, and an ‘an-
ti’ ring pucker, in which C4 is oriented in the opposite
direction to that of the C2 substituent. The anti pucker
is the conformation observed in the majority of crystal
structures of proline-based inhibitors bound in the
thrombin active site (e.g., see the crystal structure in
Fig. 2).22 Energy calculations on model compounds
48–51 using MMFF and ab initio methods revealed
Previous reports from these laboratories described the
incorporation of P3 groups containing small, nonaro-
matic side chains as a means to obtain potent P2
proline-based thrombin inhibitors with excellent phar-
macokinetic properties.16 We adopted a similar strategy
and screened several analogs of 1 and 2 containing
commercially available nonaromatic amino acids in
P3. From this screen we identified tert-butylalanine
analogs 37 and 38, which were 2-fold more potent than
1 and 2 in the 2·APTT assay and additionally exhibited
improved selectivity versus trypsin. Both 37 and 38
showed a slight improvement in oral absorption
compared to 1 and 2, but no improvement in oral
half-life was observed. We speculated that the methyl
groups of the tert-butylalanine side chain in 37 and 38
could be potential sites for oxidative metabolism. It is
known that cyclopropyl groups are more resistant to
hydrogen abstraction (the initial step in oxidative
metabolism) than alkyl groups,17,18 and therefore we
prepared analogs 39 and 40 which contain the novel
Cl
oxidation
H
N
N
oxidation, including
oxidation of the methyl
substituent to the
carboxylic acid
O
N
N
O
N
NH2
39
Figure 1. Observed metabolic profile upon incubation of compound 39
in rat, dog, monkey, and human liver microsomes.

6906
D. D. Staas et al. / Bioorg. Med. Chem. 14 (2006) 6900–6916
Figure 2. Crystal structure of compound 47 bound in the thrombin active site, from PDB structure code 1TA6.19 The pro-(S) hydrogen is colored
orange, and the aromatic rings of residues Y60A and W60D are shown in green. Also note the ‘anti’ pucker of the proline ring as described in the
molecular modeling discussion above.
Table 3. Calculated energies of proline conformers of 4-fluoro and des-fluoro prolinamides
Y
R
R
2
4
Z
2
O
N
4
Y
N
H
H
O
H3C
Z
H3C
"syn"
"anti"
binding conformer
Model compound
Eanti À Esyn (MMFF)a
Eanti À Esyn (ab initio)b
Corresponding
compound in Table 2
Ki (nM)
(kcal/mol)
(kcal/mol)
Y = Z = H (48)
Y = H, Z = F (49)
Y = F, Z = H (50)
Y = Z = F (51)
+0.98
À1.2
+3.3
+1.1
+0.29
À0.90
+1.8
39
41
42
43
0.60
0.37
110
+0.69
3.6
a R =
O
N
H
Cl
for MMFF method.
b R = Me for ab initio method; data generated using density functional theory (DFT) calculations (B3LYP method).
significant effects of fluorination on the relative energies
of the syn and anti conformers (Table 3). Results from
both computational methods indicate that relative to
the unfluorinated system 48, the equilibrium with the
(4R) monofluoro diastereomer 49 is shifted toward the
anti conformer and the equilibria with both the (4S)
monofluoro diastereomer 50 and the difluoro compound
51 are shifted toward the syn conformer.23 Thus, using
the crystallographic evidence to infer that the anti
conformer is the preferred conformer for binding to the
enzyme, both sets of calculations on model compounds
48–51 correctly predict the rank order of potencies
observed for compounds 39 and 41–43.
Excellent potency in the 2·APTT assay was obtained
with (4R)-4-fluoroprolyl triazole 41 and 4,4-difluoro-
prolyl tetrazole 44, although trypsin selectivity for the
latter is quite low. Compared to their non-fluorinated

D. D. Staas et al. / Bioorg. Med. Chem. 14 (2006) 6900–6916
6907
proline analogs 39 and 40, however, no significant
change in the oral PK properties in dogs was realized.
Incubation of (4R)-4-fluoroprolyl triazole 41 with
microsomes obtained from rat, dog, monkey, and hu-
man liver tissues revealed that the fluorine substituent
on the proline ring had effectively suppressed metabolic
transformations in P2—all metabolites arose exclusively
from oxidative transformation of the P3 methylcyclo-
propyl alanine side chain. Based on this result and the
P3 metabolites identified with 39, we synthesized the
analog of 41 in which the methyl group of the methylcy-
clopropylalanine P3 side chain is replaced with a chlo-
rine atom. The resulting compound, 45a, lost ꢀ3-fold
potency in the 2·APTT assay compared to 41, and there
was no improvement in the oral dog PK profile. Intro-
duction of a gem-dimethyl substituent at the b-carbon
of the P3 side chain to potentially block oxidation at
that position gave rise to compound 46a. Excellent
2·APTT potency resulted with this modification, and
further, the compound showed good stability in the
presence of hepatic microsomes across species. Perfor-
mance in the oral dog PK assay, however, was unre-
markable, with lower exposure and shorter half-life
compared to 41. Because at best only modest beneficial
effects were observed in the PK assay with compounds
containing several types of metabolism-reducing modifi-
cations, we concluded that compounds in this series are
being cleared predominantly by non-oxidative and/or
extrahepatic mechanisms.
(12 mL/min/kg), and a low volume of distribution
(1.7 L/kg). Compared to melagatran,24 a thrombin
inhibitor which has seen clinical use, 41 exhibited similar
functional potency with greatly improved selectivity ver-
sus trypsin. Continued pharmacokinetic optimization of
41 holds the potential for delivering a therapeutically
useful thrombin inhibitor suitable for once-daily oral
dosing.
Cl
F
H
N
N
O
N
N
O
N
NH2
41
Ki (thrombin) = 0.37 nM
Ki (trypsin) = 3300 nM
2xAPTT = 190 nM
NH
NH2
H
N
N
N
H
HO
O
O
O
Melagatran24
Ki (thrombin) = 1.1 nM
Ki (trypsin) = 7.5 nM
2xAPTT = 280 nM
6. Experimental
6.1. General
5. Conclusion
All starting materials were obtained from commercial
sources and were used without further purification.
1-Methyl-3-nitro-1-nitrosoguanidine, used for the gener-
ation of diazomethane in situ, is abbreviated as MNNG
throughout the text. 1-Hydroxy-7-azabenzotriazole is
abbreviated as HOAT throughout the text. 1-(3-Dim-
ethylaminopropyl)-3-ethylcarbodiimide hydrochloride
is abbreviated as EDC throughout the text. Anhydrous
solvents were obtained from Aldrich in Sure-seal bottles.
Commercial diethyl ether was used without further dis-
tillation or drying. MP-carbonate resin was purchased
from Argonaut Technologies, which is now owned by
Biotage AB. All reactions were carried out under nitro-
gen atmosphere unless otherwise noted. Normal-phase
chromatography was carried out using silica gel 60F
(230–400 mesh) or RediSepꢂ silica cartridges (purchased
from Teledyne Isco, Inc.). Reverse-phase preparative
HPLC was carried out on a Gilson 215 HPLC unit
using a YMC-Pack Pro C18 column (dimensions
150 · 20 mm, 5 lm particle size). NMR spectra were
recorded in CDCl3, d6-DMSO or CD3OD solution on
Varian Unity Inova spectrometers as follows: 1H
NMR spectra were recorded at 300, 400 or 500 MHz
as noted. Chemical shifts are expressed in ppm relative
to tetramethylsilane and the 1H coupling constants (J)
are quoted in Hertz (Hz). Low resolution mass measure-
ment data (MS) were generally recorded on a Waters
LC–MS using electrospray ionization (ESI) for the mass
detector and the following HPLC conditions: Stationary
In summary, structural modifications targeted at
improving the pharmacokinetic properties of lead com-
pounds 1 and 2 via reduction of metabolic oxidation in
P2 and P3 led to the identification of several novel, po-
tent inhibitors containing 4-fluoroproline in P2. The
300-fold difference in potency between the (4R)- and
(4S)-4-fluoroproline diastereomers was rationalized by
examining the crystal structure of a related compound
to infer a destabilizing interaction between the fluorine
of the less potent (4S)-4-fluoroproline diastereomer
and the p-electrons of nearby aromatic amino acid side
chains of the enzyme, and also by computational
methods which indicated that the fluorine in the more
potent (4R)-4-fluoroproline isomer stabilizes the proline
ring conformation observed in enzyme–inhibitor crystal
structures. Fluorination of the P2 proline was shown to
effectively suppress oxidative metabolism of the proline
ring in hepatic microsomes. Syntheses of several novel
a-amino acids containing cyclopropyl side chains were
developed with the goal of improving the metabolic sta-
bility of the P3 side chain. Compared to lead compound
1, inhibitor 41 which contains (4R)-4-fluoroproline in P2
and methylcyclopropylalanine in P3 showed a 2-fold
improvement in functional potency and trypsin selectiv-
ity, and a 10-fold improvement in oral plasma concen-
trations after a 1 mg/kg oral dose in dogs. Full PK
characterization of 41 in dogs after a 1 mg/kg oral dose
and a 1 mg/kg intravenous dose indicated that the com-
pound was completely absorbed (F = 100%) with a short
half-life (intravenous t1/2 = 2.0 h), moderate clearance
˚
phase: YMC Pro C18 5 lm 120 A 3.0 · 50 mm column;

6908
D. D. Staas et al. / Bioorg. Med. Chem. 14 (2006) 6900–6916
Mobile phase: A = H2O containing 0.05% TFA by vol-
ume, B = CH3CN containing 0.0425% TFA by volume;
Gradient: 8:92 B/A to 100:0 B/A over 3.6 min; Flow
Rate: 1.5–2.0 mL/min over 3.6 min; UV detection at
215 and/or 254 nm. High resolution accurate mass mea-
surement data (HRMS) were obtained as follows: Elec-
trospray (ESI) and atmospheric pressure chemical
ionization (APCI) mass spectral data were acquired on
a Bruker Daltonics 3T Fourier transform ion cyclotron
resonance mass spectrometer (FT/ICR/MS). External/
internal calibration was accomplished with polypropyl-
ene glycol.
as described above. The reaction mixture was allowed
to stir overnight under Ar and was then quenched by
the addition of a few drops of HOAc, then filtered as
above to give 900 mg of an almost colorless solid that
was chromatographed on a 35 g RediSepꢂ column
(loaded in CHCl3, eluted with 0–50% EtOAc–hexane).
The pure fractions were combined to give tert-butyl-
(3R,5R,6S)-3-[(1-methylcyclopropyl)methyl]-2-oxo-5,6-
diphenylmorpholine-4- carboxylate 5 as a colorless solid
(701 mg, 86%). 1H NMR (300 MHz, DMSO-d6) mixture
of rotamers: d 7.35–7.00 (m, 8H), 6.52 (t, J = 6.3 Hz,
2H), 6.35–6.28 (m, 1H), 5.18–4.85 (m, 2H), 2.5–1.95
(m, 2H), 1.40 (s, 4H), 1.22 (s, 3H), 1.01 (s, 9H).
6.2. tert-Butyl (3R,5R,6S)-3-(2-methylprop-2-enyl)-2-
oxo-5,6-diphenyl-morpholine-4-carboxylate (4)
6.4. N-(tert-Butoxycarbonyl)-3-(1-methylcyclopropyl)-D-
alanine (6)
To a stirred solution of tert-butyl (5R,6S)-2-oxo-5,6-
diphenylmorpholine-4-carboxylate 3 (1.5 g, 4.24 mmol)
in dry THF (30 mL) was added dropwise a 1 M solution
of NaHMDS in THF (4.67 mL, 4.67 mmol) at À78 ꢁC
under Ar. The mixture was stirred at this temperature
for 30 min and was then treated with 1-bromo-2-meth-
yl-3-propene (0.43 mL, 4.24 mmol) dropwise via syringe.
The mixture was stirred at À78 ꢁC for 30 min and the
bath was then removed and the mixture stirred at room
temperature overnight. The reaction was quenched by
the addition of 20% aqueous NH4Cl and the THF was
removed at reduced pressure. The residue was parti-
tioned between EtOAc and H2O, and the aqueous layer
extracted with EtOAc. The combined organic layers
were washed with brine, dried over Na2SO4, and con-
centrated at reduced pressure to give a residue that
was dissolved in CHCl3 and chromatographed on a
35 g RediSepꢂ column using a gradient elution of
0–50% EtOAc–hexane, 40 min at a flow rate of 40 mL/
min. The pure fractions were combined and concentrat-
ed to give tert-butyl (3R,5R,6S)-3-(2-methylprop-2-
To a stirred solution containing ethanol (0.96 mL,
16.6 mmol), liquid ammonia (ꢀ40 mL), dry THF
(20 mL), and tert-butyl (3R,5R,6S)-3-[(1-methylcyclo-
propyl) methyl]-2-oxo-5,6-diphenylmorpholine-4-car-
boxylate 5 (701 mg, 1.66 mmol) cooled to À78 ꢁC was
added lithium wire (230 mg, 33.3 mg-atom) in small
pieces until a deep blue color persisted for 10 min. The
reaction was quenched by careful addition of 20% aque-
ous NH4Cl, the cold bath removed, and the ammonia
allowed to evaporate overnight. The reaction mixture
was diluted with water and extracted with three portions
of ether. The aqueous phase was cooled in ice, acidified
with 2 M HCl, and immediately extracted with three
portions of EtOAc. The combined EtOAc layers were
washed with brine, dried over Na2SO4, and concentrat-
ed at reduced pressure to give N-(tert-butoxycarbonyl)-
3-(1-methylcyclopropyl)-D-alanine 6 as a colorless solid
(347 mg, 86%). 1H NMR (300 MHz, CDCl3): d 4.99
(br d, J = 6.1 Hz, 1H), 4.45–4.35 (m, 1H), 1.84 (dd,
J = 5.6, 14.2 Hz, 1H), 1.55 (dd, J = 8.5, 14.2 Hz, 1H),
1.46 (s, 9H), 1.11 (s, 3H), 0.40–0.25 (m, 4H).
enyl)-2-oxo-5,6-diphenyl-morpholine-4-carboxylate
as
4
a
colorless solid (785 mg, 46%). 1H NMR
(300 MHz, DMSO-d6) mixture of rotamers: d 7.35–
7.00 (m, 8H), 6.52 (t, J = 6.3 Hz, 2H), 6.35–6.28 (m,
1H), 5.15–4.82 (m, 4H), 2.95–2.70 (m, 2H), 1.87 (s,
3H), 1.41 (s, 3H), 1.03 (s, 6H).
6.5. 1-Chlorocyclopropanecarbonitrile (9)
MNNG (18.39 g, 125 mmol) was added in portions to a
stirred two phase mixture of ether (180 mL) and 40%
potassium hydroxide solution (55 mL) at 0 ꢁC. After
15 min, the ether layer was decanted into a stirred solution
of 2-chloroacrylonitrile (7) (7.98 mL, 100 mmol) in ether
(300 mL) at 0 ꢁC. After 15 min, acetic acid (1.50 mL)
was added dropwise and the solution was dried over
potassium hydroxide pellets, filtered, and evaporated in
vacuo (200 mm Hg) to give 3-chloro-4,5-dihydro-3H-pyr-
azole-3-carbonitrile 8 as a clear oil. A solution of 8 in ben-
zene (100 mL) was heated at reflux for 16 h and was then
distilled, collecting the fraction boiling at 92 ꢁC (200 mm
Hg) to give 1-chlorocyclopropanecarbonitrile 9 (3.14 g,
31%) containing 22% by mass of 2-chlorobut-2-enenitrile
(as a 20:1 mixture of isomers) as an oil. 1H NMR
(400 MHz, CDCl3): d 1.67 (m, 2H), 1.49 (m, 2H).
6.3. tert-Butyl-(3R,5R,6S)-3-[(1-methylcyclopropyl)
methyl]-2-oxo-5,6-diphenylmorpholine-4-carboxylate (5)
To a stirred mixture of ether (8 mL) and 25% aqueous
KOH (2.15 mL) at 0 ꢁC was added MNNG (637 mg,
4.33 mmol) in small portions. After 5 min, the aqueous
layer was removed via pipette, and the yellow organic
layer was added to a solution of tert-butyl (3R,5R,6S)-
3-(2-methylprop-2-enyl)-2-oxo-5,6-diphenylmorpholine-
4-carboxylate 4 (785 mg, 1.93 mmol) in THF (10 mL) at
0 ꢁC under Ar. To this stirred solution was added a very
small portion of Pd(OAc)2, resulting in mild gas evolu-
tion. After 10 min, a second portion of Pd(OAc)2 was
added. After 20 min, the reaction mixture was filtered
through a pad of SiO2 eluting with THF and the sol-
vents removed at reduced pressure to give a white solid.
NMR analysis indicated that this material contained
approximately 50% of 4. This material was treated with
double the amount of diazomethane solution prepared
6.6. Methyl (2E)-3-(1-chlorocyclopropyl)prop-2-enoate
(10)
DIBAL (33 mL of a 1 M solution in toluene) was added
to a stirred solution of 1-chlorocyclopropanecarbonitrile

D. D. Staas et al. / Bioorg. Med. Chem. 14 (2006) 6900–6916
6909
9 (3.02 g, 29.7 mmol, containing 22% by mass of 2-chlo-
robut-2-enenitrile) in toluene (30 mL) at À78 ꢁC under
nitrogen. The solution was warmed to 0 ꢁC and after
30 min was quenched with 1 M hydrochloric acid
(50 mL). Concentrated hydrochloric acid (6 mL) was
added to give a cloudy mixture and sodium chloride
was added to saturate the aqueous layer. The layers were
separated and the organic layer was dried (Na2SO4) and
filtered, washing the solids with a minimal volume of tol-
uene. (Carbomethoxymethylene)triphenylphosphorane
(9.94 g, 29.7 mmol) was added to the filtrate and the
solution was stirred for 1 h. The volatiles were removed
at reduced pressure (12 mm Hg, rt) and hexanes were
added to the residue. The solids were removed by filtra-
tion and the filtrate was evaporated in vacuo. The resi-
due was purified by flash column chromatography on
silica (1:1 methylene chloride/hexanes) to give methyl
(2E)-3-(1-chlorocyclopropyl)prop-2-enoate 10 (3.58 g,
75%) containing 11% by mass of methyl 4-chlorohexa-
2,4-dienoate (as a mixture of isomers) as a crystalline
product was purified by flash column chromatography
on silica (ethyl acetate/hexanes gradient, 20–30% ethyl
acetate) to give (4R)-4-benzyl-3-[3-(1-chlorocyclopro-
pyl)propanoyl]-1,3-oxazolidin-2-one 12 (1.36 g, 40%) as
1
a heavy oil. H NMR (400 MHz, CDCl3): d 7.20–7.36
(m, 5H), 4.68 (m, 1H), 4.20 (m, 2H), 3.33–3.21 (m, 3H),
2.77 (dd, J = 9.7, 13.4 Hz, 1H), 2.07 (t, J = 7.7 Hz, 2H),
1.09 (m, 2H), 0.85 (m, 2H).
6.9. (2R)-2-Azido-3-(1-chlorocyclopropyl)propanoic acid
(13)
A solution of (4R)-4-benzyl-3-[3-(1-chlorocyclopropyl)-
propanoyl]-1,3-oxazolidin-2-one 12 (1.01 g, 3.29 mmol)
in THF (10 mL) was added to a stirred solution of potas-
sium bis(trimethylsilyl)amide (7.25 mL of a 0.5 M solu-
tion in toluene, 3.62 mmol) in THF (10 mL) at À78 ꢁC
under nitrogen. After 30 min, a solution of 2,4,6-tri-
isopropylbenzenesulfonyl azide (1.27 g, 4.11 mmol) in
THF (10 mL) was added and followed after 1 min by ace-
tic acid (0.866 mL, 15.1 mmol). The resulting mixture was
warmed to 30 ꢁC and after 1 h was diluted with ethyl ace-
tate and washed with brine which was basified with sodi-
um hydrogen carbonate. The organic layer was dried
(Na2SO4) and evaporated to an oil. Lithium hydroxide
monohydrate (276 mg, 6.58 mmol) was added to a stirred
solution of this crude material in 2:2:1 THF/methanol/
water (25 mL) at 0 ꢁC and the reaction was warmed to
room temperature. After 16 h, excess saturated sodium
hydrogen carbonate solution was added and the volatiles
were evaporated in vacuo. Water was added to the mix-
ture to dissolve the salts and the solution was washed with
methylene chloride (4 times). The aqueous solution was
acidified with concentrated hydrochloric acid and extract-
ed with ether. The ether extract was dried (Na2SO4) and
evaporated in vacuo to an oil which was purified by flash
column chromatography on silica (1% acetic acid/hex-
anes/ethyl acetate gradient, 9–29% ethyl acetate) to give
(2R)-2-azido-3-(1-chlorocyclopropyl)propanoic acid 13
(190 mg, 30%) containing 20% by mass of 3-(1-chlorocy-
clopropyl)propanoic acid, as an oil. 1H NMR (400 MHz,
CDCl3): d 8.05 (br s, 1H), 4.43 (dd, J = 4.4, 9.5 Hz, 1H),
2.55 (dd, J = 4.2, 14.8 Hz, 1H), 1.76 (dd, J = 9.5,
14.8 Hz, 1H), 1.24–0.80 (m, 4H).
solid. 1H NMR (400 MHz, CDCl3):
d 6.53 (d,
J = 14.9 Hz, 1H), 6.17 (d, J = 14.9 Hz, 1H), 3.75 (s,
3H), 1.51 (m, 2H), 1.27 (m, 2H).
6.7. 3-(1-Chlorocyclopropyl)propanoic acid (11)
A suspension of 10% palladium on carbon (180 mg) in a
solution of methyl (2E)-3-(1-chlorocyclopropyl)prop-2-
enoate 10 (3.50 g, 21.8 mmol, containing 11% by mass
of methyl 4-chlorohexa-2,4-dienoate) in methanol
(60 mL) was stirred under an atmosphere of hydrogen
(balloon). After 16 h, the mixture was filtered through
Celite washing with ethyl acetate, and evaporated in vac-
uo. Lithium hydroxide (1.03 g, 24.5 mmol) was added to a
solution of the residual oil in 1:1:1 THF/methanol/water
(60 mL). After 30 min, the volatiles were evaporated in
vacuo and the remaining solution was washed with ether.
The aqueous solution was acidified with 1 M hydrochloric
acid and was extracted with ether. The extract was dried
(Na2SO4) and evaporated in vacuo to give 3-(1-chlorocy-
clopropyl)propanoic acid 11 (1.66 g, 50%) containing
20% by mass of hexanoic acid, as an oil. 1H NMR
(400 MHz, CDCl3): d 2.69 (t, J = 7.8 Hz, 2H), 1.99
(t, J = 7.8 Hz, 2H), 1.08 (m, 2H), 0.81 (m, 2H).
6.8. (4R)-4-Benzyl-3-[3-(1-chlorocyclopropyl)propanoyl]-
1,3-oxazolidin-2-one (12)
6.10. 3-Methylbut-2-enyl N-(tert-butoxycarbonyl)
glycinate (17)
Triethylamine (3.89 mL, 27.9 mmol) was added to a stir-
red solution of 3-(1-chlorocyclopropyl)propanoic acid 11
(1.66 g, 11.2 mmol, 20% by mass of hexanoic acid) in
THF (130 mL) and the solution was cooled to À20 ꢁC
(carbon tetrachloride/dry ice bath). Pivaloyl chloride
(1.38 mL, 11.2 mmol) was added to give a precipitate.
After 2 h, lithium chloride (0.525 g, 12.3 mmol) and
(4R)-4-benzyl-1,3-oxazolidin-2-one (1.98 g, 11.2 mmol)
were added and the mixture was warmed to room temper-
ature. After 16 h, the mixture was filtered through a glass
frit and the filtrate was evaporated in vacuo. The residual
oil was partitioned between ethyl acetate and dilute
hydrochloric acid. The organic layer was washed with
water, sodium hydrogen carbonate solution, and brine,
dried (Na2SO4), and evaporated to an oil. The crude
DCC (1 M in CH2Cl2, 26.2 mL, 26.2 mmol) was added
to a stirred solution of Boc-glycine (4.24 g, 24.2 mmol),
3-methyl-2-buten-1-ol (2.15 g, 25.0 mmol), and DMAP
(152 mg, 1.25 mmol) in anhydrous CH2Cl2 (155 mL) at
0 ꢁC. The mixture was then warmed to room tempera-
ture while stirring overnight. The mixture was filtered
twice to remove the precipitated dicyclohexylurea,
washing the solids well with CH2Cl2. The filtrate was
concentrated to a slightly colored oil. Silica gel
chromatography (15% EtOAc–hexanes) afforded
3-methylbut-2-enyl N-(tert-butoxycarbonyl) glycinate
17 as
a
colorless oil (5.52 g, 91%). 1H NMR
(400 MHz, CDCl3): d 5.34 (t, J = 7.2 Hz, 1H), 5.00 (br
s, 1H), 4.65 (d, J = 7.2 Hz, 2H), 3.91 (d, J = 5.2 Hz,
2H), 1.76 (s, 3H), 1.72 (s, 3H), 1.45 (s, 9H).

6910
D. D. Staas et al. / Bioorg. Med. Chem. 14 (2006) 6900–6916
6.11. 2-[(tert-Butoxycarbonyl)amino]-3,3-dimethyl pent-
4-enoic acid (18)
CH3CN (15 mL) was added LiOHÆH2O (77 mg,
1.82 mmol, 2.0 equiv) in water (5 mL). The reaction mix-
ture was stirred at 25 ꢁC for 3 h then was warmed to
70 ꢁC for 2 h. Acetonitrile was removed in vacuo, and
the remaining aqueous solution was cooled to 0 ꢁC and
acidified to pH 3 by dropwise addition of 1 N HCl solu-
tion. Following saturation with NaCl, the solution was
extracted into EtOAc three times. The organic layers
were combined, dried over Na2SO4, and concentrated
in vacuo to afford N-(tert-butoxycarbonyl)-3-cyclopro-
pyl-DL-valine 20 as a clear oil. 1H NMR (400 MHz,
CDCl3): d 6.41 and 5.64 (rotamers, s, 1H), 5.23 (d,
J = 8.8 Hz, 1H), 4.19 (d, J = 9.2 Hz, 1H), 1.45 (s, 9H),
0.91 (s, 3H), 0.88–0.84 (m, 1H), 0.82 (s, 3H), 0.36–0.34
(m, 2H), 0.26–0.23 (m, 2H). MS (ESI): 258.1 (M+H)+.
LDA was prepared by the dropwise addition of BuLi
(2.5 M in hexanes, 7.73 mL, 19.3 mmol) to a solution
of anhydrous diisopropylamine (2.8 mL, 20.3 mmol) in
anhydrous THF (65 mL) at 0 ꢁC under nitrogen. The
solution was stirred for 15 min at 0 ꢁC and was then
cooled to À78 ꢁC. A solution of 3-methylbut-2-enyl N-
(tert-butoxycarbonyl)glycinate 17 (2.24 g, 9.2 mmol) in
anhydrous THF (5 mL) was added dropwise over
2 min. The mixture was stirred for 15 min at À78 ꢁC
and then treated with TMSCl (2.5 mL, 19.3 mmol).
The mixture was then stirred at À78 ꢁC for 5 min, then
at room temperature for 30 min, then at 60 ꢁC for 1 h.
TLC after 1 h (50% EtOAc/hexanes, KMnO4 stain)
showed that most of the starting material had been con-
sumed. After 1 h and 45 min, the reaction mixture was
cooled to room temperature and quenched with MeOH
(5 mL). The THF was removed in vacuo and the aque-
ous residue was extracted once with ether to remove
any remaining starting material. The aqueous layer
was cooled to 0 ꢁC and adjusted to pH 3 by the dropwise
addition of 1 N HCl. The mixture was saturated with
solid NaCl and extracted immediately with EtOAc three
times. The combined organic extracts were dried
(Na2SO4), filtered and concentrated in vacuo to give 2-
[(tert-butoxycarbonyl)amino]-3,3-dimethylpent-4-enoic
acid 18 as a beige, oily solid, which was further dried
overnight under high vacuum (420 mg, 19%). 1H
NMR (400 MHz, CDCl3): d 5.90–5.84 (m, 1H), 5.14–
4.92 (m, 2H), 4.12 (br s, 1H), 1.42 (s, 9H), 1.10 (s, 6H).
6.14. Methyl (2E)-[(benzyloxy)imino]acetate (22)
Sodium acetate (32.8 g, 400 mmol) was added to a
stirred solution of glyoxylic acid hydrate 21 (14.8 g,
200 mmol) and O-benzylhydroxylamine hydrochloride
(31.9 g, 200 mmol) in 1:1 MeOH/water (300 mL).
The mixture was stirred at room temperature over-
night. The methanol was then removed in vacuo
and the residue partitioned between EtOAc and water.
The layers were separated and the aqueous layer was
saturated with NaCl and exhaustively extracted with
EtOAc. The combined organic extracts were dried
(anhydrous MgSO4), filtered, and concentrated in
vacuo to afford (2E)-[(benzyloxy)imino]acetic acid as
an oily colorless solid (32.2 g, 90%). 1H NMR
(400 MHz, CDCl3): d 7.55 (s, 1H), 7.39–7.35 (m,
5H), 5.32 (s, 2H). DCC (6.33 g, 30.7 mmol), was added
to a stirred mixture of the crude (2E)-[(benzyloxy)im-
ino]acetic acid (5.00 g, 27.9 mmol), anhydrous MeOH
6.12. Methyl N-(tert-butoxycarbonyl)-3-cyclopropyl vali-
nate (19)
(1.24 mL,
30.7 mmol)
and
DMAP
(170 mg,
To a mixture of 5 N NaOH solution (24 mL) and Et2O
(30 mL) at 0 ꢁC was added MNNG (2.29 g,
15.6 mmol), portionwise over 15 min. The mixture
was stirred vigorously for 30 min and was then allowed
to partition into two layers. The organic layer was
transferred by pipet to a solution of 2-[(tert-butoxy-
1.40 mmol) in anhydrous CH2Cl2 (180 mL) at 0 ꢁC.
The mixture was stirred overnight while warming to
room temperature. The precipitated dicyclohexyl urea
was removed by filtration and the filtrate was concen-
trated in vacuo. Purification of the residual oil by
silica gel chromatography (20% EtOAc–hexanes) affor-
ded methyl (2E)-[(benzyloxy)imino] acetate 22 as a
clear, colorless oil (4.78 g, 89%). 1H NMR
(400 MHz, CDCl3): d 7.56 (s, 1H), 7.38–7.33 (m,
5H), 5.30 (s, 2H), 3.85 (s, 3H).
carbonyl)amino]-3,3-dimethylpent-4-enoic
(316 mg, 1.30 mmol) in Et2O (30 mL) at 0 ꢁC. Immedi-
ately following transfer, catalytic amount of
acid
18
a
Pd(OAc)2 was added. The reaction mixture was stirred
at 25 ꢁC for 18 h. Acetic acid was added dropwise until
the yellow reaction color diminished and bubbling
ceased, and the solvent was then removed in vacuo.
Purification by silica gel chromatography (hexanes to
20% EtOAc/hexanes) afforded methyl N-(tert-butoxy-
carbonyl)-3-cyclopropylvalinate 19 as a pale yellow
oil (264 mg, 75%). 1H NMR (400 MHz, CDCl3): d
5.25–5.21 (br m, 1H), 4.21 (d, J = 9.6 Hz, 1H), 3.73
(s, 3H), 1.44 (s, 9H), 0.85 (s, 3H), 0.81–0.77 (m, 1H),
0.75 (s, 3H), 0.34–0.32 (m, 2H), 0.21–0.19 (m, 2H).
MS (ESI): 272.2 (M+H)+.
6.15. Methyl 2-[(benzyloxy)amino]-3,3-dimethylpent-4-
enoate (23)
Zinc dust (3.03 g, 46.6 mmol, 1.8 equiv) was added slowly
in portions over 10 min to a stirred suspension
of methyl (2E)-[(benzyloxy)imino]acetate 22 (4.98 g,
25.6 mmol, 1 equiv) and prenyl bromide (5.38 g,
4.16 mL, 36.1 mmol, 1.4 equiv) in THF (29 mL) and sat-
urated aqueous ammonium chloride (131 mL) at room
temperature. Upon complete addition of zinc, the
mixture was stirred vigorously for 30 min and was then
extracted twice with ether. The combined organic
extracts were washed with water and brine, dried
(anhydrous Na2SO4), filtered, and concentrated in
vacuo. Further drying under high vacuum afforded
methyl 2-[(benzyloxy)amino]-3,3-dimethylpent-4-enoate
6.13. N-(tert-Butoxycarbonyl)-3-cyclopropyl-DL-valine
(20)
To a stirred solution of methyl N-(tert-butoxycarbonyl)-
3-cyclopropylvalinate 19 (248 mg, 0.91 mmol) in

D. D. Staas et al. / Bioorg. Med. Chem. 14 (2006) 6900–6916
6911
1
23 as a clear, yellow oil (6.28 g, 92% yield). H NMR
6.19. N-[5-Chloro-2-(1H-1,2,4-triazol-1-yl)benzyl] proli-
namide (31)
(400 MHz, CDCl3): d 7.37–7.28 (m, 5H), 5.99 (d,
J = 12.8 Hz, 1H), 5.77 (dd, J = 10.8, 17.2 Hz, 1H),
5.01–4.94 (m, 2H), 4.64 (s, 2H), 3.72 (s, 3H), 3.39 (d,
J = 11.6 Hz, 1H), 1.03 (s, 3H), 1.01 (s, 3H).
A mixture of 1-(tert-butoxycarbonyl)proline 25 (1.03 g,
4.79 mmol, equiv), 5-chloro-2-(1H-1,2,4-triazol-1-
1
yl)benzylamine 292 (998 mg, 4.79 mmol), EDC
(1.38 g, 7.18 mmol, 1.5 equiv), and HOAT (326 mg,
2.39 mmol, 0.5 equiv) in DMF (5 mL) was stirred at
room temperature for 18 h. The reaction mixture
was diluted with water and extracted with EtOAc
(·3). The combined organic extracts were dried (anhy-
drous Na2SO4) and concentrated in vacuo. The resi-
due was purified by normal-phase medium pressure
chromatography (gradient elution with 20–80%
EtOAc–hexanes) to afford N-[5-chloro-2-(1H-1,2,4-tria-
zol-1-yl)benzyl]-[(1-tert-butoxy)carbonyl] prolinamide
(1.66 g, 85% yield). HCl gas was bubbled through a
cold (0 ꢁC) solution of this material (1.66 g,
4.07 mmol) in EtOAc (45 mL; a few drops of MeOH
were added to aid dissolution) for 5 min. The mixture
was then stirred at room temperature for 1 h. The
solvent was removed in vacuo to afford the bis-hydro-
chloride salt of N-[5-chloro-2-(1H-1,2,4-triazol-1-
yl)benzyl]prolinamide 31 as an off-white, hygroscopic
powder (1.66 g; 97% yield based on CHN analysis be-
low). 1H NMR (400 MHz, CD3OD): d 9.79 (s, 1H),
8.87 (m, 1H), 8.83 (s, 1H), 7.70 (d, J = 1.6 Hz, 1H),
7.61–7.60 (m, 2H), 4.44–4.41 (m, 2H), 4.30 (app t,
J = 7.6 Hz, 1H), 3.39–3.31 (m, 2H), 2.46–2.40 (m,
1H), 2.18–1.92 (m, 3H). MS (ESI): 306.1 (M+H)+.
Elemental analysis: calcd for (C14H16ClN5O + 2.60
equiv HCl + 0.40 H2O + 0.15 equiv Et2O): C,
41.86%; H, 5.03%; N, 16.72%; found: C, 41.86%; H,
5.04%; N, 16.82%.
6.16. Methyl 2-[(tert-butoxycarbonyl)amino]-3,3-dimeth-
ylpent-4-enoate (24)
Molybdenum hexacarbonyl (4.40 g, 16.7 mmol) was
added to a stirred solution of methyl 2-[(benzyloxy)ami-
no]-3,3-dimethylpent-4-enoate 23 (6.27 g, 23.8 mmol) in
acetonitrile (355 mL) and water (23.5 mL). The mixture
was heated to reflux for 2.5 h. The dark mixture was
cooled to room temperature and then immersed in an
ice bath. di-tert-butyl dicarbonate (1 M in THF,
28.6 mL, 28.6 mmol) was then added dropwise and the
mixture was stirred overnight while warming to room
temperature. The solvent was removed in vacuo and
the resulting black tar was purified by silica gel
chromatography (20% EtOAc–hexanes). The product
fractions were combined and concentrated to afford
methyl
pent-4-enoate 24 as a pale yellow oil (3.49 g, 57%). H
NMR (400 MHz, CDCl3): 5.86–5.79 (m, 1H),
2-[(tert-butoxycarbonyl)amino]-3,3-dimethyl-
1
d
5.11–5.00 (m, 3H), 4.15 (d, J = 9.6 Hz, 1H), 3.71 (s,
3H), 1.43 (s, 9H), 1.09 and 1.08 (overlapping s,
6H).
6.17. Methyl N-(tert-butoxycarbonyl)-3-cyclopropylvali-
nate (19)
A biphasic mixture of 5 N aqueous sodium hydroxide
(70 mL) and diethyl ether (90 mL) cooled to 0 ꢁC was
treated portionwise with MNNG (7.09 g, 48.2 mmol).
The mixture was stirred for 30 min and was then
allowed to partition into two layers. The ether layer
was transferred via pipet to a stirred solution of methyl
2-[(tert-butoxycarbonyl)amino]-3,3-dimethylpent-4-eno-
ate 24 (1.55 g, 6.02 mmol) in diethyl ether (90 mL) at
0 ꢁC. After 5 min, the stirred mixture was treated with
a catalytic amount of palladium (II) acetate (added
spatula tip at a time until faint bubbling was observed)
and the mixture stirred overnight at room temperature.
The reaction was then quenched by dropwise addition
of glacial acetic acid until bubbling ceased and the mix-
ture was concentrated in vacuo. The residue was puri-
fied by silica gel chromatography (gradient elution
100% hexanes to 20% EtOAc–hexanes) to afford meth-
yl N-(tert-butoxycarbonyl)-3-cyclopropylvalinate 19 as
6.20. N-[5-Chloro-2-(1H-tetrazol-1-yl)benzyl] prolina-
mide (32)
A
mixture of 1-(tert-butoxycarbonyl)proline 25
(513 mg, 2.4 mmol), 5-chloro-2-(1H-tetrazol-1-yl)ben-
zylamine 302 (500 mg, 2.4 mmol), EDC (686 mg,
3.6 mmol), and HOAT (162 mg, 1.2 mmol) in DMF
(12 mL) was stirred at room temperature for 18 h.
The reaction mixture was diluted with water and satu-
rated aqueous potassium carbonate solution and was
then extracted into EtOAc. The organic layer was
washed with brine, and the combined aqueous washes
were extracted once with CH2Cl2. The combined
organic layers were dried (anhydrous Na2SO4) and
concentrated. Silica gel chromatography (50–80%
EtOAc–hexanes) afforded a yellow foam (935 mg, 96%
yield). This material was then dissolved in EtOAc
(46 mL) and gaseous HCl was bubbled through the
solution for 10 min. The solvent was removed in vacuo
and the resulting residue triturated with hexanes to af-
ford the bis-hydrochloride salt of N-[5-chloro-2-(1H-
tetrazol-1-yl)benzyl]prolinamide 32 as a hygroscopic
pale yellow solid (622 mg, 71% yield). 1H NMR
(400 MHz, CD3OD): d 9.59 (s, 1H), 8.83 (br s, 1H),
7.69 (d, J = 2.0 Hz, 1H), 7.59 (dd, J = 2.0, 8.4 Hz,
1H), 7.52 (d, J = 8.4 Hz, 1H), 4.32 (m, 2H), 4.27–4.24
(m, 1H), 3.40–3.32 (m, 2H), 2.44–2.36 (m, 1H),
2.09–1.89 (m, 3H). MS (ESI): 307.1 (M+H)+.
1
a yellow oil (879 mg, 54% yield). The H NMR spec-
trum of this material is identical to that reported in
6.12 above.
6.18. N-(tert-Butoxycarbonyl)-3-cyclopropyl-DL-valine
(20)
Methyl N-(tert-butoxycarbonyl)-3-cyclopropylvalinate
19 was hydrolyzed as described in 6.13 above to
afford N-(tert-butoxycarbonyl)-3-cyclopropyl-DL-valine
20. Spectral data are identical to those of the material
prepared via the route described in Sections 6.10
through 6.13 above.

6912
D. D. Staas et al. / Bioorg. Med. Chem. 14 (2006) 6900–6916
6.21. (4R)-N-[5-Chloro-2-(1H-1,2,4-triazol-1-yl)benzyl]-
4-fluoroprolinamide (33)
1H), 2.51–2.45 (m, 1H). HRMS (ESI): calcd for
(C14H16ClFN5O)+, 324.1022; found: 324.1030.
A mixture of (4R)-1-(tert-butoxycarbonyl)-4-fluoropro-
line 264 (1.12 g, 4.79 mmol, 1 equiv), 5-chloro-2-(1H-
1,2,4-triazol-1-yl)benzylamine 292 (1.00, 4.79 mmol, 1
equiv), EDC (1.38 g, 7.19 mmol, 1.5 equiv), and
HOAT (326 mg, 2.40 mmol, 0.5 equiv) in anhydrous
DMF (14 mL) was stirred at room temperature for
2.5 h. Water and solid potassium carbonate were add-
ed and the mixture was extracted with EtOAc (·3).
The combined organic extracts were dried (Na2SO4)
and concentrated in vacuo to give a yellow oil. Purifi-
cation by silica gel chromatography (EtOAc eluent)
6.23. N-[5-Chloro-2-(1H-1,2,4-triazol-1-yl)benzyl]-4,4-
difluoroprolinamide (35)
A mixture of 1-(tert-butoxycarbonyl)-4,4-difluoro pro-
line 284 (1.00 g, 3.98 mmol), 5-chloro-2-(1H-1,2,4-tria-
zol-1-yl)benzylamine 292 (831 mg, 3.98 mmol), EDC
(1.14 g, 5.97 mmol), and HOAT (271 mg, 1.99 mmol)
in DMF (3 mL) was brought to pH 8 by dropwise addi-
tion of Hunig’s base and stirred at room temperature for
¨
18 h. The solvent was removed in vacuo. Purification by
reverse-phase chromatography [95:5 water (+0.1%
TFA)/CH3CN (+0.1% TFA) to 5:95 water (+0.1%
TFA)/CH3CN (+0.1% TFA) over 30 min] afforded a sol-
id. The solid was dissolved in EtOAc, and the solution
was washed with saturated aqueous K2CO3 solution
(·3). The organic layer was then washed with water
and brine, dried (anhydrous Na2SO4), and concentrated
in vacuo to afford Boc-N-[5-chloro-2-(1H-1,2,4-triazol-
1-yl)benzyl]-4,4-difluoroprolinamide. This material was
dissolved in CH2Cl2 (10 mL), and the solution was sat-
urated with hydrogen chloride gas. After stirring for
2 h, the solvent was removed in vacuo to afford the
bis-hydrochloride salt of N-[5-chloro-2-(1H-1,2,4-tria-
zol-1-yl)benzyl]-4,4-difluoroprolinamide 35 as a white
solid. 1H NMR (400 MHz, CD3OD): d 9.48 (s, 1H),
8.64 (s, 1H), 7.68–7.67 (m, 1H), 7.58–7.55 (m, 2H),
4.67–4.62 (m, 1H), 4.45–4.42 (m, 2H), 3.86–3.79 (m,
2H), 3.02–2.95 (m, 1H), 2.63–2.57 (m, 1H). HRMS
(ESI): calcd for (C14H15ClF2N5O)+, 342.0928; found:
342.0937.
afforded
yl)benzyl]-4-fluoroprolinamide as
Boc-(4R)-N-[5-chloro-2-(1H-1,2,4-triazol-1-
white foam
a
(1.97 g, 97% yield). A stirred solution of this material
(1.96 g, 4.62 mmol) in EtOAc (51 mL) was treated
with HCl–dioxane solution (4 M, 21 mL, excess) at
room temperature for 1 h, at which point LCMS indi-
cated 97% conversion. Additional HCl–dioxane (4 M,
6 mL, 5 equiv) was added to drive the reaction to
completion. After 30 min, the mixture was concentrat-
ed to dryness. Repeated trituration with ether afforded
the bis-hydrochloride salt of (4R)-N-[5-chloro-2-(1H-
1,2,4-triazol-1-yl)benzyl]-4-fluoroprolinamide 33 as a
1
hygroscopic white solid (2.0 g, quantitative yield). H
NMR (400 MHz, CD3OD): d 9.37 (s, 1H), 8.84 (m,
1H), 8.57 (s, 1H), 7.66 (d, J = 2.0 Hz, 1H), 7.58–7.52
(m, 2H), 5.47 (app dt, J = 3.2, 51.6 Hz, 1H), 4.52–
4.45 (m, 1H), 4.42–4.37 (m, 2H), 3.71–3.57 (m, 2H),
2.80–2.69 (m, 1H), 2.26–2.09 (m, 1H). HRMS (ESI):
calcd for (C14H16ClFN5O)+, 324.1022; found:
324.1040.
6.24. N-[5-Chloro-2-(1H-tetrazol-1-yl)benzyl]-4,4-difluo-
roprolinamide (36)
6.22. (4S)-N-[5-Chloro-2-(1H-1,2,4-triazol-1-yl)benzyl]-
4-fluoroprolinamide (34)
A mixture of 1-(tert-butoxycarbonyl)-4,4-difluoropro-
line 284 (200 mg, 0.80 mmol), 1-[5-chloro-2-(1H-tetra-
zol-1-yl)phenyl]methanamine 302 (179 mg, 0.85 mmol),
EDC (229 mg, 1.19 mmol), and HOAT (54 mg,
0.40 mmol) in DMF (4.4 mL) was stirred at room tem-
perature for 2 h. The DMF was removed in vacuo and
the residue was partitioned between EtOAc and
saturated aqueous K2CO3. The layers were separated
and the organic layer was washed with brine. The com-
bined aqueous layers were extracted once with CH2Cl2
and the combined organic extract was then dried (anhy-
drous Na2SO4) and concentrated to an orange oil. Silica
gel chromatography (60% EtOAc–hexanes) afforded 1-
(tert-butoxycarbonyl)-N-[5-chloro-2-(1H-tetrazol-1-yl)ben-
zyl]-4,4-difluoroprolinamide as a white foam (309 mg,
88%). 1H NMR (300 MHz, CDCl3): d 8.98 (s, 1H),
7.65 (br s, 1H), 7.48–7.26 (m, 3H), 4.45 (br m, 1H),
4.32–4.20 (m, 2H), 3.88–3.69 (m, 2H), 2.90 (br m, 1H),
2.55 (br m, 1H), 1.54 (br s, 9H). MS (ESI): 443.1
(M+H)+.
A mixture of (4S)-1-(tert-butoxycarbonyl)-4-fluoro-L-
proline 274 (995 mg, 4.27 mmol), 5-chloro-2-(1H-1,2,4-
triazol-1-yl)benzylamine 292 (890 mg, 4.27 mmol),
EDC (1.23 g, 6.40 mmol), and HOAT (290 mg,
2.13 mmol) in DMF (3 mL) was brought to pH 8 by
dropwise addition of Hunig’s base and stirred at room
¨
temperature for 18 h. The solvent was removed in vac-
uo. Purification by reverse-phase chromatography
[95:5 water (+0.1% TFA)/CH3CN (+0.1% TFA) to
5:95 water (+0.1% TFA)/CH3CN (+0.1% TFA) over
30 min] afforded a solid. The solid was dissolved in
EtOAc, and the solution was washed with saturated
aqueous K2CO3 solution (·3). The organic layer was
then washed with water and brine, dried over Na2SO4,
and concentrated in vacuo to afford Boc-(4S)-N-[5-chlo-
ro-2-(1H-1,2,4-triazol-1-yl)benzyl]-4-fluoroprolinamide.
This material was dissolved in CH2Cl2 (10 mL), and the
solution was saturated with hydrogen chloride gas.
After stirring for 2 h, the solvent was removed in vacuo
to afford the bis-hydrochloride salt of (4S)-N-[5-chloro-
2-(1H-1,2,4-triazol-1-yl)benzyl]-4-fluoroprolinamide 34
A solution of 1-(tert-butoxycarbonyl)-N-[5-chloro-2-
(1H-tetrazol-1-yl)benzyl]-4,4-difluoroprolinamide (460 mg,
1.04 mmol) in EtOAc (5.8 mL) at room temperature
was treated with a 4 M solution of HCl in dioxane
(1 mL, excess). After 1.5 h, additional 4 M HCl–dioxane
1
as a white solid. H NMR (400 MHz, CD3OD): d 9.64
(s, 1H), 8.74 (s, 1H), 7.66–7.65 (m, 1H), 7.57–7.56 (m,
2H), 5.42 (br d, J = 51.2 Hz, 1H), 4.89–4.38 (m, 3H),
3.78–3.70 (m, 1H), 3.62–3.50 (m, 1H), 2.80–2.65 (m,

D. D. Staas et al. / Bioorg. Med. Chem. 14 (2006) 6900–6916
6913
(1 mL) was added and the mixture stirred at room tem-
perature overnight. The solvent was removed and the
residue dried in vacuo to afford the bis-hydrochloride
salt of N-[5-chloro-2-(1H-tetrazol-1-yl)benzyl]-4,4-diflu-
3H), 0.96 (s, 9H). HRMS (APCI): calcd for
(C21H29ClN6O2)+, 433.2113; found: 433.2104.
6.27. 4-Methyl-D-leucyl-N-[5-chloro-2-(1H-tetrazol-1-
yl)benzyl]prolinamide (38)
1
oroprolinamide 36 as a yellow gum (425 mg, 98%). H
NMR (400 MHz, CD3OD): d 9.56 (s, 1H), 8.85 (s,
1H), 7.70 (d, J = 2.3 Hz, 1H), 7.61 (dd, J = 2.3, 8.5 Hz,
1H), 7.54 (d, J = 8.5 Hz, 1H), 4.57 (app t, J = 8.7 Hz,
1H), 4.34 (br s, 2H), 3.88–3.74 (m, 2H), 3.14–2.88 (m,
1H), 2.65–2.51 (m, 1H). MS (ESI): 343.0 (M+H)+.
Compound 38 was prepared from Boc-tert-butyl-D-ala-
nine (32 mg, 0.13 mmol, 1 equiv) and N-[5-chloro-2-
(1H-tetrazol-1-yl)benzyl]prolinamide
32
(50 mg,
0.13 mmol, 1 equiv) according to general procedure
6.25 and was isolated as the trifluoroacetate salt
(56 mg, 77% yield). LCMS tR = 1.15 min. 1H NMR
(400 MHz, CDCl3): d 9.18 (s, 1H), 8.45 (br s, 2H),
8.00 (t, J = 5.2 Hz, 1H), 7.73 (d, J = 2.0 Hz, 1H), 7.41
(dd, J = 2.0, 8.4 Hz, 1H), 7.19 (d, J = 8.4 Hz, 1H),
4.41–4.34 (m, 2H), 4.25 (dd, J = 4.8, 14.8 Hz, 1H),
4.00–3.94 (m, 1H), 3.89–3.85 (m, 1H), 3.53–3.46 (m,
1H), 2.25–2.22 (m, 1H), 2.09–1.85 (m, 5H), 1.00 (s,
9H). HRMS (APCI): calcd for (C20H28N7O2Cl)+,
434.2066; found: 434.2065. Elemental analysis: calculat-
ed for (C20H28N7O2Cl + 1.15 equiv TFA + 0.70 equiv
H2O + 0.10 equiv Et2O): C, 46.60%; H, 5.44%; N,
16.76%; found: C, 46.59%; H, 5.35%; N, 16.76%.
6.25. General EDC coupling and deprotection procedure
for the preparation of compounds 37–44 and 46
The following is a representative procedure for the
coupling of compounds 31–36 with the Boc-protected
P3 amino acids and subsequent deprotection to afford
compounds 37–44 and 46 unless otherwise noted: A
0.2 M solution of the Boc-protected P3 amino acid
(1 equiv) in DMF was treated with one of prolina-
mides 31–36 (1 equiv, compounds 31–36 used as their
bis-hydrochloride salts unless otherwise noted), EDC
(1.5 equiv), and HOAT (0.5 equiv). The pH of the
mixture was adjusted to 6–8 by addition of Hunig’s
¨
base and the reaction mixture was then stirred at room
temperature for 18 h. The solvent was removed in vac-
uo and the resulting residue was dissolved in 2:1
CH2Cl2/TFA (0.08 M based on starting acid) and stir-
red for 1 h. The solvent was again removed in vacuo
and purification by reverse-phase chromatography
[gradient elution on a YMC Pro C18 150 · 20 mm,
5 lm column with 95:5 water (+0.1% TFA)/CH3CN
(+0.1% TFA) to 50:50 water (+0.1% TFA)/CH3CN
(+0.1% TFA) unless otherwise noted] followed by con-
centration of product fractions and trituration of the
resulting residues with ether afforded the title com-
pounds as the trifluoroacetate salts. Alternatively, the
crude coupling product could be purified by reverse-
phase chromatography then deprotected by stirring
with HCl–dioxane solution (4 M, 10 equiv) at room
temperature for 1 h. The reaction progress was moni-
tored by LCMS and additional HCl–dioxane was add-
ed as needed to drive the reaction to completion.
Upon complete conversion the solvent was removed
in vacuo and the residue stripped twice from methanol
then triturated with ether to afford the hydrochloride
salts of the title compounds.
6.28. 3-(1-Methylcyclopropyl)-D-alanyl-N-[5-chloro-2-
(1H-1,2,4-triazol-1-yl)benzyl]prolinamide (39)
Compound 39 was prepared from Boc-3-(1-methylcy-
clopropyl)-D-alanine 6 (75 mg, 0.31 mmol) and N-[5-
chloro-2-(1H-1,2,4-triazol-1-yl)benzyl] prolinamide 31
(94 mg of the free base, 0.31 mmol) according to general
procedure 6.25 with the following changes: the interme-
diate coupling product was purified by normal-phase sil-
ica gel chromatography on an ISCO Combiflash system
(10 g RediSepꢂ column eluting with 93:7 CHCl3/saturat-
ed NH3–MeOH). This purified product (124 mg,
0.23 mmol, 74% yield from the coupling step) was then
deprotected and purified as described in procedure
6.25 and was isolated as the trifluoroacetate salt (white
solid, 55 mg, 43% yield). LCMS tR = 0.95 min. 1H
NMR (400 MHz, CD3OD): d 8.80, (s, 1H), 8.22 (s,
1H), 7.67 (d, J = 2.4 Hz, 1H), 7.50 (dd, J = 2.4, 8.4 Hz,
1H), 7.45 (d, J = 8.4 Hz, 1H), 4.39 (d, J = 15.6 Hz,
1H), 4.36–4.30 (m, 2H), 4.22 (d, J = 16.0 Hz, 1H),
3.83–3.78 (m, 1H), 3.68–3.62 (m, 1H), 2.28–2.19 (m,
1H), 2.14–1.88 (m, 4H), 1.48 (dd, J = 8.2, 14.6 Hz,
1H), 1.16 (s, 3H), 0.49–0.28 (m, 4H). HRMS (ESI):
calcd for (C21H28ClN6O2)+, 431.1957; found: 431.2004.
Elemental analysis: calcd for (C21H7ClN6O2 + 1.70
equiv TFA): C, 46.91%; H, 4.63%; N, 13.45%; found:
C, 46.74%; H, 4.96%; N, 13.68%.
6.26. 4-Methyl-D-leucyl-N-[5-chloro-2-(1H-1,2,4-triazol-
1-yl)benzyl]prolinamide (37)
Compound 37 was prepared from Boc-tert-butyl-D-ala-
nine (32 mg, 0.13 mmol, 1 equiv) and N-[5-chloro-2-
(1H-1,2,4-triazol-1-yl)benzyl]prolinamide 31 (50 mg,
0.13 mmol, 1 equiv) according to general procedure
6.25 and was isolated as the trifluoroacetate salt
(55 mg, 77% yield). LCMS tR = 1.12 min. 1H NMR
(400 MHz, CDCl3): d 8.81 (s, 1H), 8.35 (br s, 2H),
8.23 (s, 1H), 8.09 (t, J = 5.6 Hz, 1H), 7.50 (d,
J = 2.0 Hz, 1H), 7.35 (dd, J = 2.0, 8.4 Hz, 1H), 7.24 (d,
J = 8.4 Hz, 1H), 4.41–4.38 (m, 1H), 4.29 (dd, J = 5.6,
15.6 Hz, 1H), 4.20–4.12 (m, 2H), 3.87–3.83 (m, 1H),
3.50–3.45 (m, 1H), 2.12–2.02 (m, 3H), 1.85–1.72 (m,
6.29. 3-(1-Methylcyclopropyl)-D-alanyl-N-[5-chloro-2-
(1H-tetrazol-1-yl)benzyl]prolinamide (40)
Compound 40 was prepared from Boc-3-(1-methyl-
cyclopropyl)-D-alanine
6
(75 mg, 0.31 mmol) and
N-[5-chloro-2-(1H-tetrazol-1-yl)benzyl]-L-prolinamide
32 (117 mg, 0.31 mmol) according to general procedure
6.25 (coupling step allowed to proceed for two days at
room temperature) and was isolated as the bis hydro-
chloride salt (83 mg, 53%). LCMS tR = 1.10 min. 1H
NMR (500 MHz, CD3OD): d 9.54 (s, 1H), 7.73 (d,

6914
D. D. Staas et al. / Bioorg. Med. Chem. 14 (2006) 6900–6916
J = 2.5 Hz, 1H), 7.57 (dd, J = 2.0, 8.5 Hz, 1H), 7.50 (d,
J = 8.5 Hz, 1H), 4.33–4.30 (m, 3H), 4.20 (d,
J = 15.5 Hz, 1H), 3.83–3.78 (m, 1H), 3.66–3.61 (m,
1H), 2.25–2.18 (m, 1H), 2.11–1.99 (m, 3H), 1.93–1.87
(m, 1H), 1.49 (dd, J = 8.2, 14.8 Hz, 1H), 1.16 (s, 3H),
0.49–0.42 (m, 2H), 0.39–0.35 (m, 1H), 0.32–0.29 (m,
1H). HRMS (APCI): calcd for (C20H27ClN7O2)+,
432.1909; found: 432.1905.
(s, 3H), 0.48–0.29 (m, 4H). HRMS (ESI): calcd for
(C21H26ClF2N6O2)+, 467.1769; found: 467.1760.
6.33. 3-(1-Methylcyclopropyl)-D-alanyl-N-[5-chloro-2-
(1H-tetrazol-1-yl)benzyl]-4,4-difluoroprolinamide (44)
Compound 44 was prepared from Boc-3-(1-methylcy-
clopropyl)-D-alanine 6 (61 mg, 0.15 mmol) and N-[5-
chloro-2-(1H-tetrazol-1-yl)benzyl]-4,4-difluoroprolinamide
36 (91 mg, 0.24 mmol) according to general procedure
6.25 and was isolated as the bis-trifluoroacetate salt,
a fluffy white solid after lyophilization (71 mg, 70%).
LCMS tR = 1.25 min. 1H NMR (400 MHz, CD3OD):
d 9.52 (s, 1H), 8.76–8.75 (m, 1H), 7.73 (d, J = 2.0 Hz,
1H), 7.57 (dd, J = 2.0, 8.4 Hz, 1H), 7.51 (d,
J = 8.4 Hz, 1H), 4.54 (dd, J = 6.4, 9.2 Hz, 1H), 4.39–
4.33 (m, 1H), 4.28–4.03 (br m, 4H), 2.86–2.74 (m,
1H), 2.48–2.36 (m, 1H), 2.04 (dd, J = 6.0, 14.4 Hz,
1H), 1.50 (dd, J = 8.2, 14.4 Hz, 1H), 1.16 (s, 3H),
0.51–0.29 (br m, 4H). HRMS (APCI): calcd for
(C20H25ClF2N7O2)+, 468.1721; found: 468.1723.
6.30. 3-(1-Methylcyclopropyl)-D-alanyl-(4R)-N-[5-chloro-
2-(1H-1,2,4-triazol-1-yl)benzyl]-4-fluoro prolinamide
(41)
Compound 41 was prepared from Boc-3-(1-methylcy-
clopropyl)-D-alanine 6 (33 mg, 0.14 mmol) and (4R)-
N-[5-chloro-2-(1H-1,2,4-triazol-1-yl)benzyl]-4-fluoropro-
linamide 33 (60 mg, 0.14 mmol) according to general
procedure 6.25 and was isolated as the trifluoroace-
tate salt (42 mg, 55% yield). LCMS tR = 1.12 min.
1H NMR (400 MHz, CD3OD): d 8.79 (s, 1H), 8.21
(s, 1H), 7.68 (d, J = 2.4 Hz, 1H), 7.49 (dd, J = 2.4,
8.4 Hz, 1H), 7.45 (d, J = 8.4 Hz, 1H), 5.38 (br d,
J = 52.4 Hz, 1H), 4.47–4.38 (m, 2H), 4.26–4.20 (m,
2H), 4.03–3.83 (m, 2H), 2.59–2.51 (m, 1H), 2.18–
1.97 (m, 2H), 1.52 (dd, J = 8.0, 14.8, 1H), 1.14 (s,
3H), 0.49–0.29 (m, 4H). HRMS (ESI): calcd for
(C21H27ClFN6O2)+, 449.1868; found: 449.1855.
6.34. 3-(1-Chlorocyclopropyl)alanyl-(4R)-N-[5-chloro-2-
(1H-1,2,4-triazol-1-yl)benzyl]-4-fluoro prolinamides (45a
and 45b)
A
mixture
of
(2R)-2-azido-3-(1-chlorocyclopro-
pyl)propanoic acid 13 (105 mg, 0.56 mmol), (4R)-N-
[5-chloro-2-(1H-1,2,4-triazol-1-yl)benzyl]-4-fluoroprolina-
mide 33 (200 mg, 0.56 mmol, 1.0 equiv), EDC
(160 mg, 0.83 mmol, 1.5 equiv), and HOAT (38 mg,
0.28 mmol, 0.5 equiv) in DMF (1 mL) was brought
6.31. 3-(1-Methylcyclopropyl)-D-alanyl-(4S)-N-[5-chloro-
2-(1H-1,2, 4-triazol-1-yl)benzyl]-4-fluoro-L-prolinamide
(42)
Compound 42 was prepared from Boc-3-(1-methylcy-
clopropyl)-D-alanine 6 (30 mg, 0.12 mmol) and (4S)-
N-[5-chloro-2-(1H-1,2,4-triazol-1-yl)benzyl]-4-fluoropro-
linamide 34 (49 mg, 0.12 mmol) according to general
procedure 6.25 and was isolated as the bis-trifluoroac-
etate salt (white powder, 28 mg, 34% yield). LCMS
to pH 8 by dropwise addition of Hunig’s base and
¨
stirred at room temperature for 18 h. The solvent
was removed in vacuo. Purification by reverse-phase
chromatography [95:5 water (+0.1% TFA)/CH3CN
(+0.1% TFA) to 5:95 water (+0.1% TFA)/CH3CN
(+0.1% TFA)] afforded an oil. The product was dis-
solved in EtOAc, and the solution was washed with
saturated aqueous NaHCO3 solution. The aqueous
layer was extracted twice into EtOAc. The organic
layers were combined, dried over Na2SO4, and con-
centrated in vacuo to afford (4R)-1-[(2R)-2-azido-3-
(1-chlorocyclopropyl)propanoyl]-N-[5-chloro-2-(1H-1,2,
4-triazol-1-yl)benzyl]-4-fluoro-L-prolinamide as the free
base (156 mg, 57%). LCMS tR = 1.91 min. 1H NMR
(400 MHz, CDCl3): d 8.74 (s, 1H), 8.32 (s, 1H), 7.60
(d, J = 2 Hz, 1H), 7.43 (dd, J = 2.0, 8.4 Hz, 1H),
7.29 (d, J = 8.8 Hz, 1H), 5.34 (br d, J = 52.4 Hz,
1H), 4.64 (app t, J = 8.0 Hz, 1H), 4.32 (d,
J = 6.0 Hz, 2H), 4.25 (dd, J = 6.0, 7.6 Hz, 1H), 4.16–
4.08 (m, 1H), 3.83–3.71 (m, 1H), 2.66–2.06 (m, 3H),
2.05–1.98 (m, 1H), 1.20–0.77 (m, 4H). MS (ESI):
495.3 (M+H)+. A solution of this material (156 mg,
0.32 mmol) in THF (2 mL) at 0 ꢁC was treated with
triphenylphosphine (87 mg, 0.33 mmol, 1.05 equiv).
After stirring for 1 h at 0 ꢁC, water (114 lL,
6.30 mmol, 20 equiv) was added. The reaction mixture
was warmed to 40 ꢁC for 2 h then allowed to stir at
25 ꢁC for 66 h. The solvent was removed in vacuo,
and the residue was purified by reverse-phase chroma-
tography [95:5 water (+0.1% TFA)/CH3CN (+0.1%
TFA) to 50:50 water (+0.1% TFA)/CH3CN (+0.1%
tR = 1.13 min. 1H NMR (400 MHz, CD3OD):
d
8.77(s, 1H), 8.21 (s, 1H), 7.61 (d, J = 2.0 Hz, 1H),
7.47 (dd, J = 2.0, 8.4 Hz, 1H), 7.42 (d, J = 8.4 Hz,
1H), 5.37 (br d, J = 53.4 Hz, 1H), 4.59 (d,
J = 10.0 Hz, 1H), 4.38–4.10 (m, 4H), 3.99–3.87 (m,
1H), 2.60–2.33 (m, 2H), 2.11–2.06 (m, 1H), 1.50–1.41
(m, 1H), 1.16 (s, 3H), 0.50–0.27 (m, 4H). HRMS
(APCI): calcd for (C21H27ClFN6O2)+, 449.1868; found:
449.1857.
6.32. 3-(1-Methylcyclopropyl)-D-alanyl-N-[5-chloro-2-
(1H-1,2,4-triazol-1-yl)benzyl]-4,4-difluoro prolinamide
(43)
Compound 43 was prepared from Boc-3-(1-methylcy-
clopropyl)-D-alanine 6 (30 mg, 0.12 mmol) and N-[5-
chloro-2-(1H-1,2,4-triazol-1-yl)benzyl]-4,4-difluoropro-
linamide 35 (51 mg, 0.12 mmol) according to general
procedure 6.25 and was isolated as the bis-trifluoroace-
tate salt (white powder, 31 mg, 37% yield). LCMS
tR = 1.26 min. 1H NMR (400 MHz, CD3OD): d 8.77
(s, 1H), 8.21 (s, 1H), 7.64 (s, 1H), 7.50–7.43 (m, 2H),
4.85–4.55 (m, 1H), 4.43–4.37 (m, 1H), 4.28–4.15 (m,
3H), 4.08–4.03 (m, 1H), 2.83–2.77 (m, 1H), 2.45–2.40
(m, 1H), 2.06–2.01 (m, 1H), 1.53–1.47 (m, 1H), 1.14

D. D. Staas et al. / Bioorg. Med. Chem. 14 (2006) 6900–6916
6915
TFA)] to afford 130 mg (59%) of a diastereomeric
mixture of bis-trifluoroacetate salts. LCMS
1 mL/L of diethylamine). The fractions of each
diastereomer were concentrated and dried under high
vacuum overnight to remove residual diethylamine.
The hydrochloride salts were prepared by dissolution
of each diastereomer in MeOH (1 mL) followed by addi-
tion of an HCl–ether solution (2 M, 2 mL) with stirring.
The solvents were removed in vacuo to give the title
compound 46a (free base tR = 4.7 min on an analytical
Chiralpak AD 4.6 · 250 mm 10 lm column, isocratic
elution at 1 mL/min with 40:10:50 hexanes/MeOH/
iPrOH containing 1 mL/L of diethylamine) as a white
powder (42% yield) and its diastereomer 46b (free base
tR = 8.2 min on an analytical Chiralpak AD
4.6 · 250 mm 10 lm column, isocratic elution at 1 mL/
min with 40:10:50 hexanes/MeOH/iPrOH containing
1 mL/L of diethylamine) as a white powder (47% yield;
total yield of both diastereomers from deprotection
step = 89%). Diastereomer 46a: 1H NMR (400 MHz,
CD3OD): d 8.91 (s, 1H), 8.29 (s, 1H), 7.72 (d,
J = 2.0 Hz, 1H), 7.53 (dd, J = 2.4, 8.8 Hz, 1H), 7.48 (d,
J = 8.8 Hz, 1H), 5.36 (br d, J = 51.6 Hz, 1H), 4.50
(app t, J = 8.8 Hz, 1H), 4.42 (d, J = 15.2 Hz, 1H),
4.28–4.22 (m, 2H), 4.10 (s, 1H), 3.88–3.76 (m, 1H),
2.52–2.20 (m, 1H), 2.18–1.97 (m, 1H), 0.99 (s, 3H),
0.91 (s, 3H), 0.90–0.86 (m, 1H), 0.46–0.41 (m, 2H),
0.34–0.31 (m, 2H). HRMS (ESI): calcd for
(C22H29ClFN6O2)+, 463.2019; found: 463.2023. Elemen-
tal analysis: calcd for (C22H28ClFN6 O2 + 2.10 equiv
HCl + 0.15 equiv ether): C, 49.30%; H, 5.78%; N,
15.26%; found: C, 49.33%; H, 6.01%; N, 15.33%. Diaste-
tR = 1.08 min. The components were separated by chi-
ral HPLC (Chiralcel OD 5 · 50 cm column, eluting at
75 mL/min with 1:1 A/B, where A = 0.1% diethyl-
amine/hexanes and B = 2-propanol) to afford com-
pound 45a (earlier eluting diastereomer; free base
tR = 5.21 min on Chiralcel OD 250 · 4.6 mm column,
eluting at 1.0 mL/min with 1:1 A/B as defined above)
as a white solid (58 mg, 39%). The later eluting diaste-
reomer (free base tR = 7.38 min on Chiralcel OD
250 · 4.6 mm column, eluting at 1.0 mL/min with 1:1
A/B as above; 14 mg, 9.5%) contained residual diethyl-
amine and was further purified by reverse-phase chro-
matography [95:5 water (+0.1% TFA)/CH3CN (+0.1%
TFA) to 50:50 water (+0.1% TFA)/CH3CN (+0.1%
TFA)] to give the bis-trifluoroacetate salt as an oil,
45b (1.4 mg, 0.6%). Diastereomer 45a: 1H NMR
(400 MHz, CD3OD): d 8.80 (s, 1H), 8.21 (s, 1H),
7.72 (d, J = 1.6 Hz, 1H), 7.48 (dd, J = 2.0, 8.4 Hz,
1H), 7.43 (d, J = 8.4 Hz, 1H), 5.37 (br d,
J = 52.4 Hz, 1H), 4.50–4.45 (m, 1H), 4.38–4.27 (m,
3H), 4.03–3.85 (m, 2H), 2.61–2.51 (m, 1H), 2.20–2.03
(m, 2H), 1.94–1.88 (m, 1H), 1.09–0.88 (m, 3H),
0.78–0.73 (m, 1H). HRMS (ESI): calcd for
(C20H24Cl2FN6O2)+, 469.1317; found: 469.1288.
Diastereomer 45b: 1H NMR (300 MHz, CD3OD): d
8.82 (s, 1H), 8.23 (s, 1H), 7.69 (d, J = 2.4 Hz, 1H),
7.51 (dd, J = 2.1, 8.4 Hz, 1H), 7.46 (d, J = 8.7 Hz,
1H), 5.41 (br d, J = 52.2 Hz, 1H), 4.61–4.56 (m, 2H),
4.44–4.37 (m, 1H), 4.25–4.10 (m, 2H), 3.99–3.82
(m, 1H), 2.64–2.41 (m, 2H), 2.25–1.99 (m, 2H),
1.21–1.11 (m, 2H), 1.01–0.96 (m, 2H). MS (ESI):
469.2 (M+H)+.
1
reomer 46b: H NMR (400 MHz, CD3OD): d 9.17 (s,
1H), 8.44 (s, 1H), 7.74 (s, 1H), 7.51–7.46 (m, 2H), 5.36
(br d, J = 52.0 Hz, 1H), 4.63 (dd, J = 7.2, 9.6 Hz, 1H),
4.42 (d, J = 16.0 Hz, 1H), 4.24–4.15 (m, 3H), 3.88–3.75
(m, 1H), 2.60–2.49 (m, 1H), 2.21–2.01 (m, 1H), 0.99 (s,
3H), 0.97–0.91 (m, 1H), 0.84 (s, 3H), 0.45–0.41 (m,
2H), 0.37–0.33 (m, 2H). HRMS (ESI): calcd for
(C22H29ClFN6O2)+, 463.2019; found: 463.2025. Elemen-
tal analysis: calcd for (C22H28ClFN6O2 + 2.50 equiv
HCl + 0.80 equiv H2O + 0.30 equiv ether): C, 47.17%;
H, 5.99%; N, 14.23%; found: C, 47.15%; H, 5.99%; N,
14.26%.
6.35. 3-Cyclopropylvalyl-(4R)-N-[5-chloro-2-(1H-1,2,4-
triazol-1-yl)benzyl]-4-fluoroprolinamides (46a and 46b)
N-(tert-butoxycarbonyl)-3-cyclopropylvaline 20 (519 mg,
2.02 mmol 1.6 equiv) and (4R)-N-[5-chloro-2-(1H-1,2,4-
triazol-1-yl)benzyl]-4-fluoroprolinamide 33 (500 mg,
1.26 mmol, 1 equiv) were coupled according to general
procedure 6.25. The crude mixture of diastereomers
was purified by reverse-phase chromatography [gradient
elution on a YMC Pro C18 150 · 20 mm, 5 lm column
with 95:5 water (+0.1% TFA)/CH3CN (+0.1% TFA) to
5:95 water (+0.1% TFA)/CH3CN (+0.1% TFA)] to
afford an oily white foam which was taken up in MeOH
(5 mL) and shaken with MP-carbonate resin
(2.73 mmol/g, 1.35 g, 4 equiv) for 1 h. The resin was
filtered off, washing well with MeOH and CH2Cl2, and
the filtrate concentrated in vacuo to give a foamy solid
(634 mg, 89% yield). A portion of this purified mixture
of diastereomers (61 mg, 0.11 mmol) was dissolved in
dioxane (2 mL) and the solution cooled to 0 ꢁC. Excess
HCl–dioxane (2 mL of a 4 M solution) was added and
the mixture stirred at room temperature. After 4 h, addi-
tional HCl–dioxane (0.2 mL) was added and the mixture
stirred for 1 h. The solvent was removed in vacuo to
afford a white solid (65 mg). The diastereomers were
separated by chiral chromatography (Chiralpak AD
5 · 50 cm 20 lm column, isocratic elution at 50 mL/
min with 40:10:50 hexanes/MeOH/iPrOH containing
6.36. Molecular modeling
Molecular mechanics optimizations were performed
with MacroModel25 using the MMFFs force field
and a distance-dependent dielectric of 4r. Ab initio
calculations were performed with Gaussian 03.26 Start-
ing geometries from MMFFs calculations were opti-
mized at the B3LYP/6-31G** level. Frequency
calculations were performed at the same level to deter-
mine zero-point energy corrections; all frequencies
were positive.
Acknowledgments
The authors thank Mr. Carl Homnick for chiral HPLC
purification of compounds 45 and 46, Dr. Charles Ross
III for high resolution mass spectral data, and Dr. John
Wai for helpful discussions.

6916
D. D. Staas et al. / Bioorg. Med. Chem. 14 (2006) 6900–6916
Mao, S. S.; Veber, D. F.; Nutt, R. F.; Lynch, J. J.; Cook,
References and notes
J. J.; Gardell, S. J.; Shafer, J. A. Thromb. Haemost. 1995,
74, 1107–1112.
15. Lewis, S. D.; Ng, A. S.; Baldwin, J. J.; Fusetani, N.;
Naylor, A. M.; Shafer, J. A. Thromb. Res. 1993, 70, 173–
190.
16. Morrissette, M. M.; Stauffer, K. J.; Williams, P. D.; Lyle,
T. A.; Vacca, J. P.; Krueger, J. A.; Lewis, S. D.; Lucas, B.
J.; Wong, B. K.; White, R. B.; Miller-Stein, C.; Lyle, E.
A.; Wallace, A. A.; Leonard, Y. M.; Welsh, D. C.; Lynch,
J. J.; McMasters, D. R. Bioorg. Med. Chem. Lett. 2004, 14,
4161–4164.
1. (a) Bajusz, S.; Szell, E.; Bagdy, D.; Barabas, E.; Horvath,
G.; Dioszegi, M.; Fittler, Z.; Szabo, G.; Juhasz, A.;
Tomori, E.; Szilagyi, G. J. Med. Chem. 1990, 33, 1729–
1735; (b) Wiley, M. R.; Chirgadze, N. Y.; Clawson, D. K.;
Craft, T. J.; Gifford-Moore, D. S.; Jones, N. D.; Olkowski,
J. L.; Weir, L. C.; Smith, G. F. Bioorg. Med. Chem. Lett.
1996, 6, 2387–2392; (c) Dahlgren, A.; Johansson, P.-O.;
Kvarnstro¨m, D. M.; Nilsson, I.; Samuelsson, B. Bioorg.
Med. Chem. 2002, 10, 1829–1839.
2. Young, M. B.; Barrow, J. C.; Glass, K. L.; Lundell, G. F.;
Newton, C. L.; Pellicore, J. M.; Rittle, K. E.; Selnick, H.
G.; Stauffer, K. J.; Vacca, J. P.; Williams, P. D.; Bohn, D.;
Clayton, F. C.; Cook, J. J.; Krueger, J. A.; Lewis, S. D.;
Lucas, B. J.; McMasters, D. R.; Miller-Stein, C.; Pietrak,
B. L.; Wallace, A. A.; White, R. B.; Wong, B.; Nantermet,
P. G. J. Med. Chem. 2004, 47, 2995–3008.
17. Smith, D. A.; Jones, B. C.; Walker, D. K. Med. Res. Rev.
1996, 16, 243–266.
18. Manoury, P. M.; Binet, J. L.; Rousseau, J.; Lefevre-Borg,
F. M.; Cavero, I. G. J. Med. Chem. 1987, 30, 1003–1011.
19. Tucker, T. J.; Brady, S. F.; Lumma, W. C.; Lewis, S.
D.; Gardell, S. J.; Naylor-Olsen, A. M.; Yan, Y.; Sisko,
J. T.; Stauffer, K. J.; Lucas, B. J.; Lynch, J. J.; Cook, J.
J.; Stranieri, M. T.; Holahan, M. A.; Lyle, E. A.;
Baskin, E. P.; Chen, I.-W.; Dancheck, K. B.; Krueger,
J. A.; Cooper, C. M.; Vacca, J. P. J. Med. Chem. 1998,
41, 3210–3219.
3. For reviews on the effect of fluorination on the metabolism
of compounds, see (a) Park, B. K.; Kitteringham, N. R.;
O’Neill, P. M. Annu. Rev. Pharmacol. Toxicol. 2001, 41,
443–470; (b) Bo¨hm, H.-J.; Banner, D.; Bendels, S.; Kansy,
M.; Kuhn, B.; Muller, K.; Obst-Stander, U.; Stahl, M.
¨
Chem. Biochem. 2004, 5, 637–643.
20. Brandl, M.; Weiss, M. S.; Jabs, A.; Suhnel, J.; Hilgenfeld,
¨
R. J. Mol. Biol. 2001, 307, 357–377.
4. Demange, L.; Menez, A.; Dugave, C. Tetrahedron Lett.
1998, 39, 1169–1172.
5. Williams, R. M.; Im, M.-N. J. Am. Chem. Soc. 1991, 113,
9276–9286.
6. Evans, D. A.; Britton, T. C. J. Am. Chem. Soc. 1987, 109,
6881–6883.
7. Hanessian, S.; Yang, R.-Y. Tetrahedron Lett. 1996, 37,
5273–5276.
21. During preparation of this manuscript, Lange, U.E.W.
et al. published a paper describing a similar observation of
loss in potency of a P2 4,4-difluoroproline-based thrombin
inhibitor relative to the prolyl analog which the authors
attributed to an unfavorable interaction between the
difluoromethylene group at C4 and the p-electrons of
Tyr60A. See Lange, U. E. W.; Baucke, D.; Hornberger,
W.; Mack, H.; Seitz, W.; Hoffken, H. W. Bioorg. Med.
Chem. Lett. 2006, 16, 2648–2653.
22. There are 58 thrombin–inhibitor complexes in the
Protein Data Bank featuring a P2 proline binding in
the insertion loop as in Figure 2. Of these, 26 show the
proline with an anti pucker, compared with only 4 with
a syn pucker; in 28 structures the proline is presented as
flat as defined by a pseudorotational amplitude s < 20
8. Reactions carried out with lithium hydroxide in THF,
both in the presence and absence of hydrogen peroxide.
Only the hydrate and/or the peroxy adduct (in addition to
the major component of unreacted starting material) were
observed by LCMS-monitoring of the reaction.
9. Titanium isopropoxide and ethanol at reflux for >24 h
gave no reaction with 15. The relevant chemistry is
discussed in Oppolzer, W.; Tamura, O.; Deerberg, J. Helv.
Chim. Acta 1992, 75, 2572.
10. Treatment of 15 with lithium hydroxide monohydrate,
lithium bromide, and tetra-n-butylammonium bromide in
acetonitrile afforded quantitative recovery of unreacted
starting material. These conditions are described in
Karoyan, P.; Sagan, S.; Clodic, G.; Lavielle, S.; Chassa-
ing, G. Bioorg. Med. Chem. Lett. 1998, 8, 1369–1374.
11. Kazmaier, U. Synlett 1995, 11, 1138–1140.
12. For related chemistry involving the reductive cleavage of
the N–O bond in isoxazolidines with molybdenum hexa-
carbonyl, see Cicchi, S.; Goti, A.; Brandi, A.; Guarna, A.;
De Sarlo, F. Tetrahedron Lett. 1990, 31, 3351–3354.
13. The more potent diastereomers of compounds 45 and 46
(i.e., compounds 45a and 46a) were assigned R stereo-
chemistry at the P3 amine center based on previous SAR
of other P3 amino acid-based thrombin inhibitors in which
the P3 R-diastereomer is usually more potent versus
thrombin than the S-diastereomer.
degrees. Furthermore, among the structures with
˚
a
resolution better than 2.0 A, 12 display an anti pucker,
3 are flat, and none are syn.
23. Our results with the monofluoro diastereomers are in
agreement with computations on a related 4-fluoroproline
peptide reported in the literature: see DeRider, M. L.;
Wilkens, S. J.; Waddell, M. J.; Bretscher, L. E.; Weinhold,
F.; Raines, R. T.; Markley, J. L. J. Am. Chem. Soc. 2002,
124, 2497–2505.
24. The prodrug of melagatran, ximelagatran, was until
recently marketed as ExantaTM. The compound was
withdrawn from the market in early 2006. For
a
recent review of ximelagatran, see Crowther, M. A.;
Weitz, J. I. Expert. Opin. Investig. Drugs 2004, 13,
403–413.
25. Mohamadi, F.; Richard, N. G. J.; Guida, W. C.; Liskamp,
R.; Lipton, M.; Caufield, C.; Chang, G.; Hendrickson, T.;
Still, W. C. J. Comput. Chem. 1990, 11, 440–467.
26. Gaussian 03, Revision C.01, Gaussian, Inc., Pittsburgh
PA, 2003.
14. Lewis, S. D.; Ng, A. S.; Lyle, E. A.; Mellott, M. J.;
Appleby, S. D.; Brady, S. F.; Stauffer, K. S.; Sisko, J. T.;