From rigid cyclic templates to conformationally stabilized acyclic scaffolds. Part II: Acyclic replacements for the (3S)-3-benzylpiperidine in a series of potent CCR3 antagonists

Article abstract of DOI:10.1016/j.bmcl.2007.11.087

Conformational analysis of the 3-benzylpiperidine in CCR3 antagonist clinical candidate 1 (BMS-639623) predicts that the benzylpiperidine may be replaced by acyclic, conformationally stabilized, anti-1,2-disubstituted phenethyl- and phenpropylamines. Ab i

Full text of DOI:10.1016/j.bmcl.2007.11.087

Available online at www.sciencedirect.com
Bioorganic & Medicinal Chemistry Letters 18 (2008) 586–595
From rigid cyclic templates to conformationally stabilized
acyclic scaffolds. Part II: Acyclic replacements for the
(3S)-3-benzylpiperidine in a series of potent CCR3 antagonistsq
Daniel S. Gardner, Joseph B. Santella, III, Andrew J. Tebben, Douglas G. Batt,
Soo S. Ko, Sarah C. Traeger, Patricia K. Welch, Eric A. Wadman, Paul Davies,
Percy H. Carter and John V. Duncia*
Bristol-Myers Squibb Company, R & D, PO Box 4000, Princeton, NJ 08543-4000, USA
Received 20 July 2007; revised 16 November 2007; accepted 20 November 2007
Available online 28 November 2007
Abstract—Conformational analysis of the 3-benzylpiperidine in CCR3 antagonist clinical candidate 1 (BMS-639623) predicts that
the benzylpiperidine may be replaced by acyclic, conformationally stabilized, anti-1,2-disubstituted phenethyl- and phenpropylam-
ines. Ab initio calculations, enantioselective syntheses, and evaluation in CCR3 binding and chemotaxis assays of anti-1-methyl-2-
hydroxyphenethyl- and phenpropylamine-containing CCR3 antagonists support this conformational correlation.
Ó 2007 Elsevier Ltd. All rights reserved.
In our previous letter, we disclosed the discovery of clin-
ical candidate BMS-639623 (1), a small molecule CCR3
F
N
N
receptor antagonist indicated for the treatment of asth-
ma.1 It is believed that antagonism of the CCR3 recep-
tor on eosinophils should prevent their chemotaxis into
the lungs where they cause bronchial mucosal damage
by releasing major basic protein, membrane-derived li-
pid mediators, and a host of other proteins and toxic
substances. This damage is thought to give rise to the
clinical features of asthma—bronchial hyperresponsive-
ness and airway obstruction.2
N
H
N
N
HN
OH
N
2
: CCR3 IC50 =1.0 nM3,4
Chemotaxis IC50 = 0.010 nM4
O
F
N
N
H
N
H
N
N
N
N
O
1
:
BMS-639623
Given our previous success1 in predicting the stereochem-
istry of 1 from the conformational analysis of the cyclo-
hexane ring of lead molecule 2 (Fig. 1), we wanted to see
if we could extend that analysis to the 3-benzyl-substi-
tuted piperidine ring. We envisioned that by introducing
R1 and R2 onto phenpropylamine 3 (Fig. 2), we could dis-
cover a conformationally rigid acyclic isostere for the
CCR3 IC50 = 0.3 nM5 Chemotaxis IC50 =0.038 nM
Figure 1. The discovery of 1 via conformational analysis of 2.
F
N
OH
N
H
N
H
N
N
N
N
2
1
1
3
O
?
Keywords: CCR3 antagonist; Eosinophil chemotaxis; Acyclic scaffold;
Conformational analysis; Ab initio; Asthma; 3-Benzylpiperidine.
q
F
R2 RN OH
N
N
N
H
N
H
N
This work was originally presented in part by J.V.D. at Balticum
Organicum Syntheticum 2006, BOS06, a Biennial International
N
3
N
2
1
3
R1
Conference on Organic Synthesis, June 26, 2006, Tallinn, Estonia:
www.BOS06.ttu.ee, Poster 28.
O
Figure 2. Can the R1- and R2-substituted phenpropylamine of 3 mimic
the 3-benzylpiperidine of 1?
*
Corresponding author. Tel.: +1 609 252 3123; e-mail:
john.duncia@bms.com
0960-894X/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2007.11.087

D. S. Gardner et al. / Bioorg. Med. Chem. Lett. 18 (2008) 586–595
587
3-benzylpiperidine. Conformational analysis would dic-
tate the stereochemistry of the R1 and R2 groups.
basicity on the nitrogen, such as CF3CH2, Ac, and
MeSO2, decrease binding potency. As described sepa-
rately in greater detail, cyclic linker analogs were also
investigated (see representative examples 16–18).6 These
seem to be much more potent than their acyclic counter-
parts when we compare the chemotaxis potencies of 17
and 4a. For all of the more potent compounds, we no-
tice a disconnect between binding IC50 and chemotaxis
IC50 potencies which has previously been observed in
the acyclic1 and cyclic linker4 3-benzylpiperidine series
of CCR3 antagonists.
Before we introduce our R1 and R2 substituents onto a
phenpropylamine, we first must investigate the SAR of
the unsubstituted phenpropylamine itself (3: R1,
R2 = H). We find in Table 1 that all of the compounds
with only one tetrazole substituent on the phenylurea
(R00 = H) exhibit weaker chemotaxis IC50 values
(compounds 4a–8a). However, in the presence of a 3,5-
disubstituted phenylurea (in this case a [3-ethyl-5-(1-
methyltetrazol-5-yl)phenyl]urea) we observe an increase
in potency, as seen in other series of CCR3 antagonists.1
In this 3,5-disubstituted phenylurea series, substituents
on the basic nitrogen larger than R = methyl (4b), such
as R = ethyl (5b) and R = cyclopropylmethyl (8b), yield
compounds exhibiting single-digit picomolar chemotaxis
IC50 values.
In Table 1, we see a dramatic increase in chemotaxis
inhibition potency of about 3 orders of magnitude going
from R = methyl to R = ethyl or cyclopropylmethyl (cf.
4a and 5a; 4b and 5b, 8b). This large difference in po-
tency is unlikely due to a simple increase in van der
Waals interactions of ethyl versus methyl with the recep-
tor site. Ab initio calculations on fragments 4aa and 5aa
show that for R = Et, there is an increase in energy rel-
ative to R = Me of about 2 kcal/mol when bond angle w
Substituents larger than a cyclopropylmethyl on the ba-
sic nitrogen are less potent. Substituents lowering the
Table 1. Binding and eosinophil chemotaxis inhibition activities of the unsubstituted phenpropylamines in both the ‘acyclic linker’ (A) and ‘cyclic
linker’ series (B)
F
N
R
N
OH
H
N
H
N
N
N
N
A
O
O
F
N
R"
N
H
N
R
N
H
N
N
N
B
R"
a
CCR3 IC50 (nM)
b
Chemotaxis IC50 (nM)
Compound
Structure A/B
R
R00
4a
5a
A
A
A
A
A
A
A
A
A
Me
Et
H
H
H
H
H
H
H
H
Et
0.4
0.2
0.2
0.4
0.4
2.5
59
0% at 30 nM
34% at 0.3 nM
6a
7a
Pr
c-Pr
43% at 3 nM
—
8a
9a
c-Pr–CH2
c-Hex
CF3CH2
44% at 3 nM
—
10a
11a
12
—
—
c
AdCH2
H
421
0.3
55% at 30 nM
42% at 3 nM
4b
5b
6b
A
A
A
Me
Et
Et
Et
Et
0.3
0.4
0.5
2.6
0.004
Pr
62% at 0.3 nM
42% at 0.03 nM
7b
8b
9b
A
A
A
c-Pr
c-Pr–CH2
c-Hex
Et
Et
Et
0.6
0.6
3.2
—
0.005
58% at 30 nM
0% at 3 nM
10b
11b
13
A
A
A
CF3CH2
AdCH2
PhCH2
Et
Et
Et
9.0
211
2.5
—
—
61% at 0.3 nM
9% at 0.03 nM
14
15
16
17
18
A
A
B
B
B
Ac
Et
Et
H
H
H
6.0
—
—
SO2Me
H
Me
21.6
3.7
—
0.6 0.8
—
0.6 0.3
60
CF3CH2
a See Ref. 5 for CCR3/CCR2 chimera binding assay.
b See Ref. 4 for chemotaxis assay. Values without standard deviations represent a single determination.
c Ad = adamant-1-yl.

588
D. S. Gardner et al. / Bioorg. Med. Chem. Lett. 18 (2008) 586–595
cyclohexyl, and benzyl are less potent, most likely due to
increased rotational degrees of freedom and/or increased
steric hindrance with the receptor site.
Having found that N-ethyl is optimal on phenpropyl-
amine 3 where R1 and R2 = H, we are now ready to
make conformationally rigid acyclic isosteres for the 3-
benzylpiperidine of 1. Conformational analysis of the
3-benzylpiperidine and its fragments will allow us to
determine the stereochemistry of R1 and R2 in phenpro-
pylamine 3 that is needed to impart conformational
rigidity while simultaneously mimicking the binding
conformation of the piperidine. The piperidine, like a
cyclohexane, is a six-membered ring and therefore exists
in a chair in its lowest energy conformation, which we
assume is the binding conformation. Varnes et al. had
earlier discovered via the synthesis of rigidized piperi-
dine-containing CCR3 antagonists, that an equatorial
benzyl group imparts more potency than an axial
one.7 Thus, we can draw a (3S)-3-benzylpiperidine in a
chair conformation as shown in Figure 4 (20a or 20e)
with the benzyl in the equatorial position. Before we
break up the piperidine ring and convert it into a disub-
stituted phenpropylamine, we must add a substituent at
the 2-position which will eventually become R1. The
substituent R1 may be substituted either axially (20a)
or equatorially (20e). Taking 20e and cutting out a
methylene group at the piperidine 5-position yields an
acyclic 1,2-disubstituted phenylpropylamine which can
reside in three low energy conformers, 21, 22, and 23.
Conformer 21 is preferred since it has only 2 non-bond-
F
R
OH
H
H
N
H
N
H
H
H
N
ψ
H
CH3
H
O
H
H H
H H H
4aa: R = Me 5aa: R = Et
Figure 3. Dihedral drive of 4aa (green) and 5aa (red). Shown is the
relevant diastereomer which was used in the ab initio calculations.
is rotated to eclipse R and OH (w = 130°, Fig. 3). Thus
R = Et results in a more restricted bond angle w com-
pared to R = Me. Apparently, this restricted bond angle
permits better alignment of functional groups leading to
increased chemotaxis inhibition potency. R = cyclopro-
pylmethyl is similar enough to R = Et to share the same
enhancement in potency. Other groups such as n-propyl,
H
*
R2
F
R1
R2
NR'R''
H
NR'R''
o
o
*
24: mimics 20e
R1
25
F
cut here
H
5
6
H
H
H
*
R1
Bn
NR'R''
R''R'N
Bn
H
H
R1
R2
1
R1
4
N
2
N
*
3
H
F
R2
R2
Bn
H
R1
23
20e
NR'R''
22
F
19
21
preferred
o
o
conformer
cut here
N
H
H
H
H
H
NR'R''
R2
R''R'N
Bn
R1
R2
R1
H
H
X
R1
R2
Bn
Bn
X
R1
H
28
X
20a
NR'R''
F
preferred
26
conformer
27
H
R2
F
R2
R1
R''R'N
NR'R''
H
R1
30
F
29: does not mimic 20a
Figure 4. Conformational analysis of S-3-benzylpiperidine 19 leading to structures 25 and 30. ‘X’ denotes unfavorable gauche interactions.
‘ ’ denotes the position in the acyclic phenpropylamine which mimics the 3-position of 3S-3-benzylpiperidine 19.
*

D. S. Gardner et al. / Bioorg. Med. Chem. Lett. 18 (2008) 586–595
589
ing gauche interactions, while conformers 22 and 23
have three. Redrawing conformer 21 leads to structure
24 which clearly shows the identical relative stereochem-
istry as in benzylpiperidine 20e between the benzyl and
amine groups. Further redrawing 24 yields anti-1,
2-disubstituted 3-phenylpropylamine 25 which theoreti-
cally should mimic the equatorial conformer of 3-ben-
zylpiperidine. Since S-3-benzylpiperidine 19 has been
shown3 to be the more potent enantiomer, the enantio-
mer drawn for isostere 25 should be the more potent
one also. On the other hand, if we substitute an axial
R1 group as in 20a, the preferred conformer is 28. When
we redraw 28 as 29, we clearly see that the relative ori-
entation of the benzyl and amino groups is different
from that found in 3-benzylpiperidine 20a. Redrawing
29 as 30, we obtain a syn-1,2-disubstituted 3-phenylpro-
pylamine. Thus, only the anti- and not the syn-1,2-disub-
stituted-3-phenylpropylamine mimics the equatorial
conformer of 3-benzylpiperidine.
1–1.5 kcal/mol. For syn-diastereomer 30 when
R1 = R2 = Me (fragment 30a), the most preferred con-
former is 28 and it is favored over 26 and 27 by approx-
imately the same amount. We will now see that in
addition to being easier to synthesize, substitution of
R2 = OH instead of Me has a second benefit: the energy
gap widens even further between the preferred and non-
preferred conformers! Thus, for anti-diastereomer 25
when R1 = Me, R2 = OH (fragment 35a), the most pre-
ferred conformer is again 21, but it is now favored over
22 and 23 by about 5–7 kcal/mol. A similar energetic
profile is seen with pseudoephedrine fragment 31a where
preferred conformer 21 is more favored over 22 and 23
by 3–6 kcal/mol. Thus, substitution of R2 = OH seems
to enhance the preference for conformer 21 in both the
phenpropylamine and phenethylamine series. The rea-
son for this is that when R2 = OH in lowest energy con-
former 21, there is a stabilizing electrostatic interaction
between the partially positive NH proton and the par-
tially negative oxygen of the OH (a positive dipole–di-
pole interaction). In higher energy conformers 22 and
23, there is a destabilizing steric interaction and repul-
sion between the OH and methyl groups (a negative di-
pole–dipole interaction) (see Supplemental information
where ab initio calculation results are summarized for
the positive and negative dipole–dipole interactions).
The question now arises what should R1 and R2 be in our
acyclic S-3-benzylpiperidine replacement 25? In our pre-
vious letter1 we showed that the cyclohexane group of 2
can be mimicked by a syn-1,2-disubstituted propyl chain
as found in 1 where R1 = Me and R2 = OH. Can we also
substitute R1 = Me and R2 = OH in structure 25? The
synthesis would be much simpler than when R1,
R2 = Me. The 1-methyl-2-hydroxy substitution pattern
is also shared by the well-known phenethylamine, pseu-
doephedrine. Although pseudoephedrine is not a phen-
propylamine, we felt it was sufficiently close in structure
to merit investigation as a 3-benzylpiperidine replace-
ment, especially when all four diastereomers are commer-
cially available to test our conformational analysis. It is
also known that conformer 21 of pseudoephedrine (re-
place Bn with Ph) predominates for both the non-proton-
ated and protonated forms in D2O with intramolecular
H-bonding not occurring.8
Using what we learned from Table 1, namely that the N-
ethyl group together with a 3-ethyl-5-(1-methyltetrazol-
5-yl)phenylurea is optimal for chemotaxis inhibition, we
synthesized the 1-methyl-2-hydroxyphenpropylamine
and pseudoephedrine/ephedrine analogs in Table 2. As
predicted from the conformational analysis in Figure
4, incorporation of (1S,2S)-(+)-pseudoephedrine leads
to 31H, the most potent of all of the isomers possessing
a (1-methyltetrazol-5-yl)phenylurea tail piece. It con-
tains anti-stereochemistry, as does 25, which we pre-
dicted to best mimic a 3-benzylpiperidine. We also
expected the (1S,2S)-(+)-enantiomer of pseudoephed-
rine (31H) and not the (1R,2R)-(ꢀ)-enantiomer (32H)
to lead to the more potent diastereomer since it embod-
ies the same absolute stereochemistry as in 25. Interest-
ingly, syn-diastereomer 33H is only 10-fold less potent
Ab initio calculations7 confirm our conformational
analysis (Fig. 5). For anti-diastereomer 25 when
R1 = R2 = Me (fragment 25a), the most preferred con-
former is 21 and it is favored over 22 and 23 by about
F
F
H
H
N+
φ
N+
φ
25a—green
30a—red
31a—magenta
30a
25a
35a--cyan
1a—blue
F
OH
φ
OH
H
N+
H
N+
φ
31a
35a
F
H
N+
φ
(S)-3-(4-fluorobenzyl)-1-methylpiperidine, 1a
Figure 5. Energetic profiles obtained from ab initio calculations for fragments 25a (green), 30a (red), 31a (magenta), 35a (cyan), and (S)-3-(4-
fluorobenzyl)-1-methylpiperidine 1a (blue). Numbers 21/26, 22/27, 23/28 correspond to the conformers shown in Figure 3.

590
D. S. Gardner et al. / Bioorg. Med. Chem. Lett. 18 (2008) 586–595
Table 2. Replacement of S-3-benzylpiperidine of 1 with 1-methyl-2-hydroxyphenethyl- and phenpropylamines, together with their corresponding
CCR3 binding and eosinophil chemotaxis IC50 values
N
OH
H
N
H
N
N
N
R
N
O
R"
a
CCR3 IC50 (nM)
Compound
R
R00
Chemotaxisb (nM)
OH
2
1
31H
31Et
N
H
Et
0.4
0.5
22, 67% at 30 nM
78% at 0.03 nM
(1S,2S)-(+)-pseudoephedrine
OH
1
N
2
32H
33H
34H
H
H
H
577
—
—
—
(1R,2R)-(-)-pseudoephedrine
OH
1
N
2
5.7
(1R,2S)-(-)-ephedrine
OH
1
N
2
238
(1S,2R)-(+)-ephedrine
F
35H
35Et
H
19
OH
N
Et
1.2
F
36H
36Et
H
Et
88
3.4
—
—
OH
N
F
37H
37Et
H
Et
2.5
0.2
N
F
5H
5Et
H
Et
0.2
0.4
55% at 0.3 lM
0.004
N
F
1H
1Et
H
Et
0.3
1.8 0.6
0.04 0.01
0.013
N
a See Ref. 5 for CCR3/CCR2 chimera binding assay.
b See Ref. 4 for chemotaxis assay. Values without standard deviations represent a single determination.
than anti-31H, and it is more potent than the anti-32H.
In addition, syn-isomer 34H contains the correct abso-
lute chirality at C-2 as in 25, but it is greater than 100-
fold less potent than syn-33H, which does not. We
may infer from this that the chirality about C-1 is a
more important determinant for potency than is the chi-
rality about C-2, the chiral center that is positionally
equivalent to the chiral center of S-3-benzylpiperidine
(see Fig. 4 and follow the asterisks). Once the chirality
about C-1 is set, then additional increases in potency

D. S. Gardner et al. / Bioorg. Med. Chem. Lett. 18 (2008) 586–595
591
arise from anti-substitution at C-2. Although sharing
similar binding potencies, 31H and 31Et differ consider-
ably in chemotaxis inhibition potency. Here again we see
in 31Et the increase in potency from 3,5-disubstitution
on the phenyl urea where the chemotaxis IC50 is less
than 30 pM, similar to that of 3-benzylpiperidine 1H.
It is not clear whether 31Et is as potent as unsubstituted
phenpropylamine 5Et or the 3,5-disubstituted benzylpi-
peridine 1Et,9 but it is close. Thus the (1S,2S)-1-
methyl-2-hydroxyphenethylamine moiety is an excellent
isostere for the (3S)-3-benzylpiperidine.
37Et. Not only does a single OH group have a negative
impact on binding, but so does a single methyl group
(compare 37H with 5H)—although the ‘3,5-disubstitu-
tion effect’ seems to make up the difference in compound
37Et. Unfortunately, we do not have a chemotaxis IC50
value for 37Et to prove or disprove that.9 We would ex-
pect 37H to be more potent than 5H since conforma-
tional analysis predicts the correct rotamer 21a to be
populated roughly 50% of the time (Fig. 7), the other
50% being populated by rotamer 22a. Monosubstituted
propylamine 37H is also less potent than pseudoephed-
rine 31H although 37Et and 31Et have equal and potent
binding affinities due to the ‘3,5-disubstitution effect’.
We were a bit surprised at the nanomolar binding po-
tency of syn-33H, expecting it to be much less potent.
Seeking an explanation, we performed ab initio calcula-
tions on syn-fragment 33a (Fig. 6) and found that both
conformers 26 and 28 (replace Bn with Ph in Fig. 4) are
actually of equal energy unlike what we had predicted.
Thus 33H not only exists in the inactive conformation
28, but also in the active binding conformation 26.
One can rationalize the observed energetics as follows.
Even though conformer 26 has 3 gauche interactions,
only one of them involves the bulky phenyl and there
is a positive dipole–dipole interaction between the par-
tially positive NH proton and the partially negative oxy-
gen of the OH group. In our previously predicted
‘favored conformer’ 28, there is one gauche interaction
with the phenyl and one repulsive dipole–dipole interac-
tion between the R1 = Me and the OH oxygen. In con-
former 27, there are two gauche interactions involving
the phenyl and one repulsive methyl/oxygen interaction
to make it the highest in energy.
To explain the decreased binding potency of the substi-
tuted phenpropylamines we once again performed ab
initio calculations shown in Figure 8. Unsubstituted
propylamine fragment 5aaa’s lowest energy conformers
occur at h = 60° (60–80°), 140°, and 180°. Pseudoephed-
rine fragment 31aaa shares a low energy conformer with
5aaa at 60°, thereby possibly making this low energy
conformer the binding conformer. Monosubstituted
and disubstituted phenpropylamine fragments 37aaa
and 35aaa, respectively, do not share an energy mini-
mum at this dihedral angle. This explains the near equiv-
alence in potency of 31H with 5H, 31Et with 5Et and
their greater potency than 35H, 35Et, 37H, and possibly
37Et.9 At 60°, both 35aaa and 37aaa have a gauche
interaction between the a-methyl group and the N-ethyl
group which apparently is severe enough to disallow an
energy minima to occur. On the other hand, 5aaa at 60°
has a gauche interaction between a small hydrogen and
the N-ethyl group, while 31aaa between the a-methyl
group and the N-methyl group. These are apparently
less severe than the N-ethyl interactions in 35aaa and
37aaa thus allowing for low energy minima to occur.
Looking at the 1,2-disubstituted phenpropylamines in
Table 2, we see that the anti-substituted 35H and 35Et
are again more potent than their syn-counterparts 36H
and 36Et as predicted by conformational analysis. How-
ever, compounds 35H, 35Et, 36H, and 36Et are all for
the most part less potent than their phenethylamine
pseudoephedrine/ephedrine counterparts. They are also
less potent than their des-OH counterparts 37H and
Another interaction favors the increased potency found
for the pseudoephedrines 31H and 31Et. The h = 180°
conformer represents the ‘fully extended’ conformation
where the longest chains are opposite one another in a
Newman projection (not drawn). Both unsubstituted
31a – magenta
33a - cyan
H
F
H
R1
NR'R''
NR'R''
H
R1
24a: mimics 20e
25a
F
22/27
21/26
H
H
H
R1
NR'R''
H
H
R1
H
R''R'N
Bn
H
H
Bn
Bn
R1
23a
H
NR'R''
22a
OH
φ
OH
φ
21a
H
H
N+
N+
preferred
conformer
Figure 7. Conformational analysis of 25 when R2 = H (25a). ‘X’
denotes unfavorable gauche interactions. There is only one non-
bonding gauche interaction in conformers 21a and 22a. There are two
in conformer 23a. Therefore both conformers 21a and 22a are favored
nearly equally.
31a
33a
Figure 6. Energetic profile of compounds 31a and 33a. The inlaid
numbers indicate the conformation depicted in the Newman projec-
tions from Figure 4.

592
D. S. Gardner et al. / Bioorg. Med. Chem. Lett. 18 (2008) 586–595
31aaa in Figure 8, and angle w (ꢀ75 to ꢀ90°) in 5aaa
in Figure 3. Ab initio calculations on the entire mole-
cule of 31Et also show an intramolecular H-bond that
we did not anticipate: an H-bond between the two
OH groups. Thus, not only does the pseudoephedrine
OH group stabilize conformer 21 more than a methyl
group as discussed previously, but it also ‘locks up’ half
of the molecule in a stable conformation via an intramo-
lecular H-bond. Since the a-methyl group to the urea has
been shown to control the conformation of the other half
of the molecule,1 one can say that 31Et’s entire scaffold is
essentially rigid and favors the one conformation
shown.10 Overlap of 31Et with the minimized structure
of BMS-639623 (2) in Figure 9b shows that the ben-
zylpiperidine of 2 provides a scaffold in which h is
locked at an angle of 180°, yielding a fully extended
‘phenylpropylamine’ backbone which is different from
that of 31Et. Fortunately, pseudoephedrine’s phenyl
group in 31Et resides in the same general area as the
phenyl in the benzylpiperidine of 2 resulting in out-
standing binding affinity and chemotaxis inhibition po-
tency. When we overlap a minimized structure of 5H as
in Figure 9c, we see that the phenylpropylamine moiety
can follow the same extended backbone conformation
(h = 180°) as found in the benzylpiperidine of 2. How-
ever, we can also make h = 80° in 5H and induce the
phenpropylamine’s carbon backbone to align itself with
that of 31Et (Fig. 9d) and this conformation also turns
out to be an energy minimum. Thus for phenpropyl-
amine 5H, the phenpropylamine portion appears to
have at least two different binding conformations:
either that of pseudoephedrine 31H or that of benzylpi-
peridine 2 thereby making it very potent, even though it
is not as conformationally rigid as 2 or 31Et. As men-
tioned earlier, since both 35H and 37H cannot have
h = 60–80° or 180°, they are less potent.
F
5aaa: X=Y=H
X
OH
OH
H
N
H
N
H
37aaa: X=H, Y=Me
35aaa: X=OH, Y=Me
+
θ N
Y
O
OHθ N
H
N
H
N
H
+
31aaa
O
Figure 8. Dihedral drive of fragments 5aaa (red), 37aaa (blue), 35aaa
(green), and 31aaa (magenta).
propylamine 5aaa and monosubstituted phenpropyl-
amine fragment 37aaa display a low energy conformer
in the ‘fully extended’ conformation when h = 180°. What
happens to 31aaa and 35aaa at 180°? When h = 180°, the
X = OH and the N-alkyl groups in this fully extended con-
formation become eclipsed resulting in higher energy
(negative dipole–dipole interaction) similar to when the
N-ethyl group eclipsed the other OH in the molecule
(Fig. 3, w = 130°). Recall that the N-ethyl group enhanced
potency in the unsubstituted phenpropylamines by favor-
ing a restricted bond angle w. For 31aaa and 35aaa, a sim-
ilar situation also occurs withbond angle h. The N-alkyl—
OH negative dipole–dipole interaction disfavors the bond
angle h = 180°. Thus a restricted bond angle is induced,
favoring a low energy conformer where h = 60° for
31aaa and h = 100° for 35aaa. Since unsubstituted propyl-
amines 5H and 5Et also share the low energy conformer
with h = 60°, compounds 31H and 5H, 31Et and 5Et are
similar in binding potency and 31Et and 5Et are similar
in chemotaxis inhibition potency. Unfortunately for
35H, 35Et, 37H, and 37Et, the N-ethyl gauche interaction
with the a-methyl group disallows a low energy confor-
mation at h = 60° and thus 35H, 35Et, 37H, and possibly
37Et9 are all less potent.
Compounds in Tables 1 and 2 were synthesized by the
methods shown in Scheme 1. Epoxide 41 was synthe-
sized by the method of Beaulieu.11 Subsequent epoxide
opening with pseudoephedrine or a phenpropylamine
48 yields 42. Hydrogenolysis of the benzyl groups fol-
lowed by reaction with carbamate 441 yields urea 45.
Compounds in Table 1, structure B were synthesized
according to the method of De Lucca4 using the appro-
priate phenpropylamine. Compounds in Table 2 were
made by the methods disclosed in Scheme 2. Wittig reac-
tion of BOC-L-alaninal12 yields chiral styrene 53. Asym-
metric Sharpless dihyroxylation13 with AD-mix-b yields
glycol 54. Absolute stereochemistry was assigned based
on precedent with a similar substrate where the 1-CH3
group in 53 is replaced by a 1-CH2OH.14 The chiral cen-
ter at C-1 is reported not to play a role in determining
the outcome of the asymmetric dihydroxylation.14 Ben-
zylic hydroxyl removal via hydrogenation, followed by
deprotection, benzylation, reductive amination with
acetaldehyde, and debenzylation yields, amine 57. Reac-
tion with epoxide 58,15 followed by deprotection and
reaction with carbamate 44,1 yields ureas 35H and
35Et. Compounds 36H and 36Et may be synthesized
by repeating the sequence using AD-mix-a.13 Mono-
methyl-substituted compounds 37H and 37Et may be
synthesized by hydrogenating styrene 53 and completing
Incorporating what we have learned about angles /, h,
and w, we performed ab initio calculations on the entire
molecules of 31Et, 5H, and BMS-639623 (2) (Fig. 9). In
Figure 9a, the a-methyl-b-hydroxypropylurea of 31Et is
in a conformation which mimics that of the cyclohexyl
linker of 2 and which was calculated in our previous
letter1 (see also Fig. 9b). Angles /, h, and w are 161°,
80°, and ꢀ79°, respectively. These angles are all close
to the ones found earlier when the calculations were
done for discrete bond angles only: angle / in con-
former 21 (160° to 170°—the benzyl and amino groups
are nearly opposite one another), angle h (60–80°) in

D. S. Gardner et al. / Bioorg. Med. Chem. Lett. 18 (2008) 586–595
593
Figure 9. Ab initio minimized structures of (a) 31Et (green); (b) 31Et and 2 (gray); (c) 31Et, 2, and 5H (yellow) where 5H is in the fully extended
conformation (h = 180°); (d) 31Et, 2, and 5H where in 5H, h = 80°.
O
O
c
O
a
b
N
N
N
N
41
H
NH2
HO
HO
40
39
38
d
H(F)
e
R2
n
R
N
R1
OH
H(F)
R2
n
R
N
R1
OH
NH2
n=0,1
n=0,1
43
42
N
N
N
COOH
H
N
f
N
PhO
46
44
F
F
F
g,h
O
R"
R2
n
H(F)
NH2
N
N
R
N
OH
H
47
H
N
N
N
i,j
k
N
O
R1
NHR
n=0,1
48
49
45
R"
CHO
F
Scheme 1. Reagents and conditions: (a) i—BnBr, K2CO3, EtOH, rt; ii—LAH, reflux, 24 h, 72%; (b) DMSO, Py–SO3, 10–15 °C, 2 h, 96% crude; (c)
CH2Br2, THF, n-BuLi, ꢀ55 °C, 95%, 6:1 mixture of diastereomers; (d) pseudoephedrine or 48, EtOH, reflux, 20 h; (e) Pd(OH)2, H2, MeOH, HOAc;
(f) 44 (R00 = H, Et), acetonitrile, 25 °C; (g) i—SOCl2, CHCl3; ii—NH4OH; (h) LAH, THF 67% for two steps; (i) R0COCl, TEA, CH2Cl2; (j) LAH,
THF; (k) RNH2, STAB, CH2Cl2.
the reaction sequence in Scheme 2 omitting steps d
and e.
dine. In this unsubstituted phenpropylamine series, we
found that substitution of groups on the basic nitrogen
larger than methyl, such as ethyl, cyclopropyl, and
cyclopropylmethyl, leads to molecules with chemotaxis
IC50 values in the single-digit picomolar range.
In summation, we have found that unsubstituted phen-
propylamines can effectively mimic a (S)-3-benzylpiperi-

594
D. S. Gardner et al. / Bioorg. Med. Chem. Lett. 18 (2008) 586–595
Cl
PPh3
F
Boc
NH
O
O
Boc
NH
a,b
NH2
NH
H
N
c
HO
Boc
F
51
53
52
50
d
F
F
F
OH Boc
NH
OH Boc
NH
OH
f
e
OH
54
56
55
g,h,i
F
F
OH
OH
OH
j,k
O
N
NH2
CBZ
59
N
H
57
58
l
F
N
N
OH
OH
H
H
N
N
N
N
N
O
35H, 35Et
R"
Scheme 2. Reagents and conditions: (a) EtOCOCl, Hunig’s base, EtSH, CH2Cl2; (b) Et3SiH, 10% Pd–C, acetone; (c) KHMDS, THF/toluene, 18%
for three steps; (d) AD-mix-b, MeSO2NH2, t-BuOH/H2O, quant; (e) Pd(OH)2, H2, MeOH, 62%; (f) HCl/dioxane, 94%; (g) PhCHO, STAB,
ClCH2CH2Cl, AcOH, 67%; (h) CH3CHO, STAB, ClCH2CH2Cl, AcOH, quant; (i) H2, Pd(OH)2, MeOH, quant; (j) 58, EtOH, reflux, 43%; (k) H2,
Pd(OH)2, MeOH, quant; (l) 44 (R00 = H, Et), CH3CN, 32%.
Through the use of conformational analysis of the 3-
benzylpiperidine moiety of CCR3 antagonist 1, to-
gether with ab initio calculations, we were able to pre-
dict that the anti-1,2-disubstituted phenethyl- and
phenpropylamine isosteres should be more potent than
their syn-counterparts. We could quickly test and
validate our conformational analysis through the
incorporation of the four commercially available pseu-
doephedrine diastereomers. It was found that (1S,2S)-
(+)-pseudoephedrine is an excellent acyclic mimic of
cyclic (S)-3-benzylpiperidine. The pseudoephedrine
portion of 31Et binds in a slightly different conforma-
tion than does the benzylpiperidine of 2. However,
pseudoephedrine’s phenyl and benzylpiperidine’s phe-
nyl end up occupying the same space which most
likely accounts for 31Et’s outstanding binding and
chemotaxis inhibition potencies. The presence of an
intramolecular H-bond between the two OH groups
in 31Et essentially ‘locks up’ half of the molecule into
a single conformation. Since the rest of the molecule
has been previously shown to be rigid,1 the entire
scaffold of 31Et is therefore rigid. We also found that
anti-disubstituted phenpropylamines were more potent
than syn but that both were less potent than the pseu-
doephedrine and unsubstituted phenpropylamines. We
were able to explain the greater potency of the latter
compounds using ab initio calculations.
In this letter and in our previous letter,1 we have per-
formed conformational analyses on both the cyclohexyl
and the benzylpiperidine of lead 2. This permitted us to
discover potent CCR3 receptor antagonists such as 31Et
containing a totally acyclic but conformationally rigid
scaffold as shown in Figure 10. The opposite strategy is
usually the norm in medicinal chemistry: acyclic scaffolds
are rigidized into more potent cyclic ones. We hope that
F
H
N
OH
H
OH
N+
N
Conformational
Analysis
H
N
H
N
N+
N
N
N
N
N
N
CH3
O
NH
CH3
N
H
O
31Et
2
Acyclic
Cyclic
Figure 10. Via conformational analysis of the cyclohexane and benzylpiperidine moieties of lead molecule 2, we eventually were able to design potent
CCR3 antagonist 31Et.

D. S. Gardner et al. / Bioorg. Med. Chem. Lett. 18 (2008) 586–595
595
Welch, P.; Covington, M.; Stowell, N.; Wadman, E.;
Davies, P.; Solomon, K.; Newton, R. C.; Trainor, G. L.;
Decicco, C. P.; Wacker, D. Bioorg. Med. Chem. Lett.
2004, 14, 1645; (b) Ab initio calculations were performed
as disclosed in the previous letter in this series (Ref. 1, and
then see Ref. 6 therein).
our two exercises in conformational analysis will inspire
medicinal chemists to design conformationally stable
acyclic alternatives to their heterocyclic and carbocyclic
drug targets.16
8. Tsai, H.; Roberts, J. D. Magn. Reson. Chem. 1992, 30,
828.
Supplementary data
9. It was at this point when the CCR3 discovery research
program was concluded. Thus there are no chemotaxis
inhibition data.
Supplementary data associated with this article can be
found, in the online version, at doi:10.1016/j.bmcl.
2007.11.087.
10. An interpretable 1H NMR spectrum of the TFA salt of
31Et could only be obtained in CD3OD. Other NMR
solvents yielded spectra with broadened peaks. It can be
envisaged, that H-bonding by CD3OD could prevent the
optimal conformer (Fig. 9a) from occurring. Surprisingly,
the pseudoephedrine portion of 31Et yielded coupling
constants that matched those that were calculated from
the conformation shown in Figure 9a (see Table below).
However, observed coupling constants for He–Hf differed
from those calculated. 1H NMR spectra and selective
1H–1H decoupled NMR spectra were obtained at both 500
References and notes
1. Santella, J. B., III; Gardner, D. S.; Yao, W.; Shi, C.;
Reddy, P.; Tebben, A. J.; Delucca, G. V.; Wacker, D. S.;
Watson, P. S.; Welch, P. K.; Wadman, E. S.; Davies, P.;
Solomon, K. A.; Graden, D. M.; Yeleswaram, S.; Man-
dlekar, S.; Kariv, I.; Decicco, C. P.; Ko, S. S.; Carter, P.
H.; Duncia, J. V. Bioorg. Med. Chem. Lett. 2008, 18, 576,
Available online 22 November, 2007.
2. For an abridged etiology of asthma and its eosinophil-
contributing component, see Supplement in Ref. 1.
3. The binding assay was carried out using 125I-labeled
eotaxin and CHO cells stably transfected with a gene
encoding human CCR3 as described in Ref. 4.
4. De Lucca, G. V.; Kim, U.-T.; Vargo, B. J.; Duncia, J. V.;
Santella, J. B., III; Gardner, D. S.; Zheng, C.; Liauw, A.;
Wang, Z.; Emmett, G.; Wacker, D. A.; Welch, P. K.;
Covington, M.; Stowell, N. C.; Wadman, E. A.; Das, A.
M.; Davies, P.; Yeleswaram, S.; Graden, D. M.; Solomon,
K. A.; Newton, R. C.; Trainor, G. L.; Decicco, C. P.; Ko,
S. S. J. Med. Chem. 2005, 48, 2194.
1
1
and 700 MHz. In addition, H–1H and H–13C 2D NMR
spectra were obtained at 500 MHz.
He
Hd
H
Ha
OH
OH
N
N
H
N
H
N
N
N +
N
-
CF3CO2
CH3
O
Hb
Hc
Hf
Hx-Hy Calculated Jx-y
,
Measured Jx-y
,
(Hz)
(Hz)
5. The binding assay was carried out using 125I-labeled
eotaxin and CHO cells stably transfected with a gene
encoding a chimeric receptor, consisting of the intracellu-
lar domain of human CCR2 with the extracellular and
transmembrane domains of human CCR3. This variation
was found to give results nearly identical to those obtained
using the native CCR3 receptor as shown below:
Ha-Hb
Hc-He
Hd-He
He-Hf
9.6
10
2.5
10
5
3.2
11.0
10.7
11. Beaulieu, P. L.; Wernic, D. J. Org. Chem. 1996, 61, 3635.
12. Miyazaki, T.; Han-ya, Y.; Tokuyama, H.; Fukuyama, T.
Synlett 2004, 477.
13. Sharpless, K. B.; Amberg, W.; Bennani, Y. L.; Crispino,
G. A.; Hartung, J.; Jeong, K.-S.; Kwong, H.-L.; Morik-
awa, K.; Wang, Z.-M.; Xu, D.; Zhang, X.-L. J. Org.
Chem. 1992, 57, 2768.
14. Imashiro, T. R.; Sakurai, O.; Yamashita, T.; Horikawa,
H. Tetrahedron 1998, 54, 10657.
15. Prepared by a slight modification of the procedure found
in Chen, P.; Cheng, P. T. W.; Spergel, S. H.; Zahler, R.;
Wang, X.; Thottathil, J.; Barrish, J. C.; Polniaszek, R. P.
Tetrahedron Lett. 1997, 38, 3175.
Inhibitory effects
of 1 on 125I-labelled
100
eotaxin on
hCCR3/CCR2
50
chimera transfected
CHO cells. IC50
=
0
0.4 nM, n=8 Each
point represents
mean SEM.
-7 -6 -5 -4 -3 -2 -1
Log cin,uM
0
Taken from U.S. 2005/0123972, published June 9, 2005.
6. Ko, S. S.; Wang, J.; Clark, C. M.; Srivastava, A. S.;
Chaudhari, A.; Batt, D. G.; Houghton, G.; DeLucca, G.;
Duncia, J. V.; Welch, P. K.; Wadman, E. A.; Davies, P.;
Solomon, K. A.; Graden, D. M.; Yeleswaram, S.; Decicco,
C. P.; Carter, P. H., manuscript in preparation.
16. For approaches to designing conformationally rigid
acyclic scaffolds which mimic conformationally preorga-
nized acyclic natural products, see (a) Hoffmann, R. W.
Angew. Chem., Int. Ed. Engl. 1992, 31, 1124; (b)
Hoffmann, R. W. Angew. Chem., Int. Ed. Engl. 2000,
39, 2054; (c) Schopfer, U.; Stahl, M.; Brandl, T.;
Hoffmann, R. W. Angew. Chem., Int. Ed. Engl. 1997,
36, 1745.
7. (a) Varnes, J. G.; Gardner, D. S.; Santella, J. B.; Duncia,
J. V.; Estrella, M.; Watson, P. S.; Clark, C. M.; Ko, S. S.;