Antagonists of the human CCR5 receptor as anti-HIV-1 agents. Part 3: A proposed pharmacophore model for 1-[N-(methyl)-N-(phenylsulfonyl)amino]-2-(phenyl)-4-[4-(substituted) piperidin-1-yl]butanes

Article abstract of DOI:10.1016/S0960-894X(01)00491-7

Structure-activity relationship studies directed toward the optimization of (2S)-2-(3-chlorophenyl)-1-[N-(methyl)-N-(phenylsulfonyl)amino]-4-[4- (substituted)piperidin-1-yl]butanes as CCR5 antagonists resulted in the synthesis of the spiro-indanone derivative 8c (IC₅₀ = 5 nM). These and previous results are summarized in a proposed pharmacophore model for this class of CCR5 antagonist.

Full text of DOI:10.1016/S0960-894X(01)00491-7

Bioorganic & Medicinal Chemistry Letters 11 (2001) 2469–2473
Antagonists of the Human CCR5 Receptor as Anti-HIV-1 Agents.
Part 3: A Proposed Pharmacophore Model for 1-[N-(Methyl)-N-
(phenylsulfonyl)amino]-2-(phenyl)-4-[4-(substituted)piperidin-1-
yl]butanes
Paul E. Finke,a,* Laura C. Meurer,a Bryan Oates,a Shrenik K. Shah,a
Jennifer L. Loebach,a Sander G. Mills,a Malcolm MacCoss,a Laurie Castonguay,b
Lorraine Malkowitz,c Martin S. Springer,c Sandra L. Gouldc and Julie A. DeMartinoc
aDepartment of Medicinal Chemistry, Merck Research Laboratories, PO Box 2000, Rahway, NJ 07065, USA
bDepartment of Molecular Systems, Merck Research Laboratories, PO Box 2000, Rahway, NJ 07065, USA
cDepartment of Immunology Research, Merck Research Laboratories, PO Box 2000, Rahway, NJ 07065, USA
Received 9 April 2001; accepted 2 July 2001
Abstract—Structure–activity relationship studies directed toward the optimization of (2S)-2-(3-chlorophenyl)-1-[N-(methyl)-N-
(phenylsulfonyl)amino]-4-[4-(substituted)piperidin-1-yl]butanes as CCR5 antagonists resulted in the synthesis of the spiro-indanone
derivative 8c (IC50=5 nM). These and previous results are summarized in a proposed pharmacophore model for this class of CCR5
antagonist. # 2001 Elsevier Science Ltd. All rights reserved.
The chemokine receptor CCR5, a member of the seven-
transmembrane G-protein coupled family of receptors,1
has been identified as a primary co-receptor with CD4
by which macrophage tropic HIV-1 virus strains infect
their host cells.2 These CCR5 utilizing HIV-1 strains,
now called R5 variants, have been associated with the
initial and early phases of HIV-1 infection, although
they are generally present throughout the course of the
disease AIDS. Individuals homozygous for a 32-base
pair deletion in the gene for CCR5, which prevents
expression of functional receptor on the cell surface,
have been identified as being highly resistant to HIV-1
infection,3 while infected heterozygous individuals
showed significantly delayed progression to AIDS.4
Given the importance of CCR5 for the establishment,
and possible maintenance, of HIV-1 infection in vivo,
and the lack of an overt detrimental phenotype in
humans that do not express functional CCR5, numer-
ous efforts have been initiated to identify suitable CCR5
antagonists for use as potential therapeutic agents for
the treatment of HIV-1 infection.5ꢀ8
In our first manuscript in this series,9 the discovery of
(2S)-2-(3,4-dichlorophenyl)-1-[(N-methyl-N-phenylsul-
fonyl)amino]-4-[spiro(2,3-dihydrobenzthiophene-3,40 -
piperidin-10-yl)]butane S-oxide (1, mixture of R- and S-
sulfoxides) as our initial key lead structure was de-
scribed, having an IC50=35 nM for inhibition of [125I]-
MIP-1a binding to CCR5. Subsequent investigation of
the C-2 phenyl revealed that the 3-chlorophenyl deriva-
tive 3 had improved CCR5 binding affinity of 10 nM.10
In addition, the structurally simplified 4-phenylpiper-
idine derivatives (e.g., 5, IC50=30 nM) were found to
have binding affinities within 3- to 4-fold of the corre-
sponding spiro compounds. Herein, further structure–
activity relationship (SAR) studies for the spiro and 4-
phenyl series are described and a preliminary pharma-
cophore model for this class of CCR5 antagonist is
proposed.
The synthesis of these modified piperidine derivatives
was based on our previously described routes10ꢀ12
(Scheme 1) and utilized (S)-(3-(3-chlorophenyl))-4-(N-
methyl-N-phenylsulfonylamino)butanal (7). The reduc-
tive amination of appropriately substituted piperidines
(6a) or piperazines (6b) using sodium triacetoxyborohy-
dride in dichloroethane (DCE)13 afforded the coupled
products 8a–bb (see Tables 1 and 2) in 55–90% yields.
*Corresponding author. Fax: +1-732-594-5790; e-mail: paul_finke@
merck.com
0960-894X/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(01)00491-7

2470
P. E. Finke et al. / Bioorg. Med. Chem. Lett. 11 (2001) 2469–2473
Scheme 2. Reagents: (a) LiN(SiMe3)2, THF, ꢀ70 ꢁC, then 9, ꢀ70 ꢁC
to rt; (b) oxalyl chloride, DMF (cat), DCM, rt; (c) MeNH2 (40% aq),
THF, 0 ꢁC; (d) DIBAL, DCM, 0 ꢁC to rt; (e) PhSO2Cl, DIPEA, DCM,
rt; (f) HOAc, NaBH(OAc)3, DCE, rt.
The hydroxy compound 8i (1:1 mixture of alcohol dia-
stereomers) was prepared from the ketone 8c by reduction
with sodium borohydride in methanol.
sulfone 4 had only slightly less affinity (IC50=270, 10,
and 15 nM for 2, 3, and 4, respectively).10 An initial
survey of other known spiro piperidines was undertaken
to understand the beneficial binding affinities of the
sulfoxide and sulfone moieties, whether due to a hydro-
gen bond acceptor interaction or a simple polar effect.
As anticipated, the spiro-indane 8a (IC50=180 nM) was
about the same as the sulfide 2. The observation that 8a
was significantly poorer than the simple 4-phenyl com-
pound 5 indicated that the phenyl ring is not optimally
disposed in 8a and that the sulfoxide and sulfone moi-
eties have a substantial interaction with the receptor
which more than compensates for this deficiency. While
the spiro-indoline 8b did show somewhat improved
binding over 8a (IC50=50 and 180 nM), the best spiro
piperidine found was the ketone 8c (IC50=5 nM), which
now directs the carbonyl oxygen into the plane of the
phenyl ring rather than out of the plane as with 3 and 4.
This improved binding was also reflected in a PBMC-
based antiviral assay in which 8c was 8-fold better than
3 in a side-by-side assay17 (IC95=1500 vs 12,500 nM18).
Surprisingly, the lactams 8d and 8e, in which the oxygen
is similarly disposed as 8c, but would be expected to be
a better hydrogen bond acceptor, were poorer inhibitors
(see below), as were both orientations of the six-member
lactams 8f and 8g. Extension of the sulfonyl as in the
sulfonamide 8h was detrimental (IC50=12% I@1000
nM). Reduction of the ketone of 8c to the alcohol deri-
vative 8i (1:1 mixture of alcohol isomers) indicated that
a hydroxy at this position was at best only marginally
beneficial compared to 8a (IC50=100 vs 180 nM).
The requisite piperidines and piperazines that were uti-
lized in Scheme 1 were available commercially (5 and
8l–z,bb) or were prepared by literature procedures (2–
4,10,11 8a, 8c, 8j–k,14 8b,d–h,11 and 8aa15).
Two non-piperidine derivatives were also prepared by a
modified procedure (Scheme 2) in the original 3,4-
dichlorophenyl series. Alkylation of 3,4-dichlorophenyl-
acetic acid (10) with bromides 9 afforded 11. Conversion
to the amides 12, reduction and acylation with phe-
nylsulfonyl chloride afforded 13 and 14. Several open
chain compounds of varying length (e.g., 16 and 17)
were prepared by reductive amination of the
corresponding amines 15 with 7 as in Scheme 1.
These compounds were then evaluated in a [125I]-MIP-
1a based binding assay of CCR5 stably expressed on
Chinese hamster ovary (CHO) cells9,16 (Tables 1 and 2).
As previously reported, the sulfide 2 displayed only
moderate binding compared to the sulfoxide 3, while the
Other arrangements of the spiro linkage were not
advantageous, for example, the 2-spiro-benzofuran 8j
and the spiro-pyran 8k (Table 1), although the phenyl
substituted spiro derivative 8l and the non-spiro urea
8m (Fig. 1) did have moderate activity (IC50=300 and
90 nM). The importance of the substitution being
exclusively at the 4-postion was verified with the 3,4-
Scheme 1. Reagents: (a) HOAc (or DIPEA if HCl salt of 6 is used),
NaBH(OAc)3, DCE, rt.

P. E. Finke et al. / Bioorg. Med. Chem. Lett. 11 (2001) 2469–2473
2471
In search of a minimum structure for the piperidine,
both the 4-unsubstituted compound 8s and the 4-t-butyl
8t were found to be inactive (27% and 40% I@1000
nM), thus, substitution on the phenyl ring of 5 was
undertaken (Table 2). The 2-methyl 8u was a poorer
inhibitor, again indicating that the phenyl perpendicular
to the plane of the piperidine was not optimal, although
this was mostly overcome with the 2-MeO derivative 8v.
The 3-trifluoromethyl 8w and 4-chloro 8x derivatives
were also less active, although the 4-fluoro 8y was at
least equipotent to 5. Thus, it appeared that substitution
on this ring is limited and that there is an apparent steric
interaction with 4-piperidine substituents which linearly
extend much further than a phenyl. Extension of the
phenyl ring as a benzyl or phenylethyl (8z and 8aa)
again reduced activity, although further homologation
to the less rigid 3-phenylpropyl 8bb restored binding
potency comparable to the best spiro derivative (8bb
and 8c, IC50=5 nM). However, the antiviral activity of
8bb was not improved over 3 (IC95=50,000 vs 25,000
nM for 3 in the same assay18).
Table 1. Structures and activities for spiro piperidine derivatives
Compd
X
Y-Z
CCR5a
IC50 (nM)b
c
2
3
4
—
—
–SCH2–
–S(O)CH2–
–S(O)2CH2–
–CH2CH2–
–NHCH2–
270
10d
15
—
—
—
—
—
—
—
—
—
—
–CH2–
–O–
8a
8b
8c
8d
8e
8f
8g
8h
8i
8j
8k
180d
50
–C(O)CH2–
–C(O)NH–
5d,e
45d
100d
35
–C(O)N(Me)–
–C(O)NHCH2–
–NHC(O)CH2–
–N(MeSO2)CH2–
–CH(OH)CH2–
–O–
35
12%f
100
260
24%f
–C(O)CH2–
aSee refs 9 and 14 for a description of the binding assay.
Table 2. Structures and activities for non-spiro piperidine derivatives
bThe IC50 values are an average of three independent titrations having
calculated standard errors usually less than 15%. The assay-to-assay
variation was generally ꢂ2-fold or less based on the results for the
standard compound 3.
cThe X in these structures is a single bond.
dThe average of two or more separate experiments.
eThe IC50 for compound 8c was at least 2-fold better than 3 in two
separate head-to-head experiments.
fPercent inhibition at 1000 nM; this was highest concentration used.
Compd
R
Y
CCR5a
IC50 (nM)b
fused piperidines 8n and 8o and the 3-phenyl isomer 8p
(IC50=250 nM, 32% and 40%@1000 nM, respec-
tively). Thus, in light of these findings and the fact that
the simple 4-phenyl derivative 5 was within 3-fold in
activity to 3, alternative non-spiro structures were fur-
ther investigated.
5
8q
8r
8s
8t
8u
8v
8w
8x
8y
8z
8aa
8bb
C6H5–
C6H5–
2-MeC6H4–
H
t-Bu
–CH–
–N–
–N–
30
700
2400
27%c
40%c
400
70
120
200
25
–CH–
–CH–
–CH–
–CH–
–CH–
–CH–
–CH–
–CH–
–CH–
–CH–
2-MeC6H4–
2-MeOC6H4–
3-CF3C6H4–
4-ClC6H4–
4-FC6H4–
C6H5CH2–
C6H5CH2CH2–
C6H5CH2CH2CH2–
The role of the basic piperidine was first investigated
with the corresponding less basic piperazines19 8q and 8r
which both showed greatly diminished binding com-
pared to 5 (IC50=700 and 2400 vs 30 nM, Table 2). The
biphenyl 13 and 4-phenylpyridine 14 (see Scheme 2)
were completely inactive (IC50’s >10,000 nM). Several
basic, but non-piperidine structures, as exemplified by
the open chain compounds 16 and 17, were also all
inactive (IC50’s >10,000 nM). Thus, with this scaffold
the piperidine subunit seemed to be required in terms of
a critical N+–H salt interaction with the receptor as
well as proper orientation of the 4-phenyl substituent.
250
65
5
aSee refs 9 and 14 for a description of the binding assay.
bThe IC50 values are an average of three independent titrations having
calculated standard errors usually less than 15%. The assay-to-assay
variation was generally ꢂ2-fold or less based on the results for the
standard compound 3.
cPercent inhibition at 1000 nM.
Based on the above results, a tentative pharmacophore
model for inhibition of MIP-1a binding to CCR5 was
developed (Fig. 2). In our first report,9 the critical role
of the N-methyl-N-phenylsulfonylamino had been
established in that the N–H phenylsulfonamide and
corresponding benzamide derivatives were much less
active. Thus, the conformation of this portion must be
limited to conformations not readily accessible by the
later derivatives. The sulfonamide oxygens may also
directly interact with the receptor or point towards a
hydrophilic area since the corresponding ether deriva-
tive was much less active (41% I@1000 nM). Since the
Figure 1. Structures of alternative piperidine analogues.

2472
P. E. Finke et al. / Bioorg. Med. Chem. Lett. 11 (2001) 2469–2473
4-position on the phenylsulfonyl moiety was shown to
be compatible with a wide range of substitution (but not
the 2- and 3-positions), the phenyl appears to lay in an
extended channel. The stereochemistry of the C-2
phenyl had also been determined to be stereospecific for
the (S) configuration. The requirement for the C-2
phenyl and its substituent preferences were previously
discussed10 wherein substitution was shown to be lim-
ited to small, nonpolar moieties at the 3-position (H, Cl,
F, and Me), although 4-methyl was not detrimental.
Thus, the phenyl appears to be in a small, lipophilic
area. The need for the piperidine nitrogen, assumed to
be protonated at the pH of the testing medium, implies
an important salt interaction with the receptor. The
piperidine is also required for proper orientation of the
4-substitution.
resulted from a more optimal hydrogen bonding
orientation for the oxygen.
However, the results from the indane 8a and the 2-
methylphenyl analogue 8r compared to 5, as well as the
poor activity of the lactams, seemed to contradict this
simple assumption and led to consideration of other
possible motifs (Fig. 3). Conformer C was initially of
interest since the phenyl would still occupy the same
space as the equatorial 4-phenyl of 5 and might explain
the poorer activity of the lactams. Conformational ana-
lysis of the piperidine moiety21 (as the protonated N-Me
derivative) yielded structures A–C and afforded the
indicated relative conformational energies. However,
this analysis showed that C was significantly higher in
energy than either A or B (5–7 kcal/mol) and thus might
not be energetically accessible. Nevertheless, structures
A and B were found to have little (<2 kcal) conforma-
tional energy difference. Similar analysis for the lactam
8d also shed some light on its poorer activity in that
there is a larger energy difference between its corre-
sponding conformers A and B (+3.2 kcal), thus sig-
nificantly deceasing its relative abundance. The
observation that 8a and 8r were significantly less active
than 5 might also be rationalized by a beneficial inter-
action of the phenyl with the same receptor residue
when not forced perpendicular to the piperidine (see
Fig. 2, hashed line phenyl ring). Also noteworthy is the
observation that the more flexible 3-phenylpropyl 8bb
might advantageously place the phenyl in a similar beta
orientation, but further removed from the piperidine
ring. Thus, as indicated in Figure 2, conformer B is the
preferred model at this time and was the subject of fur-
ther investigation.22
Figure 2. Proposed CCR5 pharmacophore model for 8c and 8y
(hashed lines). The orientation of the three substituents on the central
methine carbon is a low energy conformation based on the SAR and
modeling of 8c and several other compounds in this series.20
Figure 3 shows three possible conformations of the
spiro piperidine subunit. The role of the spiro structure
in combination with the activity of the simple 4-phenyl
compound 5 initially seemed to imply that structure A,
and not B, was the active conformation and that the
sulfoxide provided an additional beneficial receptor
interaction, thus the observed 3-fold improvement in
binding activity (IC50=10 vs 30 nM for 3 and 5). The
effect of the sulfoxide stereochemistry, and thus the
absolute stereochemistry for the sulfoxide interaction,
was not determined since the two diastereomers were
not separated. The observed improvement with ketone
8c was again consistent with this model and might have
Thus, the results presented here offer a detailed
description of a pharmacophore model for inhibition of
MIP-1a binding to CCR5 with this type of antagonist
scaffold, which has been demonstrated to be capable of
preventing the infection of PBMC cells by R5 strains of
HIV-1. A more complete description of a CCR5 binding
model incorporating these and related compounds will
be described in the near future.
Acknowledgements
We would like to thank Elizabeth Roth and William
Schleif for performing the PBMC assays on compounds
8c and 8bb.
References and Notes
1. For a review of the chemokines and their receptors, see
Baggiolini, M.; Dewald, B.; Moser, B. Annu. Rev. Immunol.
1997, 15, 675.
2. For a review of the co-receptor search, see Fauci, A. S.
Nature 1996, 384, 529.
3. Liu, R.; Paxton, W. A.; Choe, S.; Ceradini, D.; Martin,
S. R.; Horuk, R.; MacDonald, M. E.; Stuhlmann, H.; Koup,
R. A.; Landau, N. R. Cell 1996, 86, 367.
4. Michael, N. L.; Chang, G.; Louie, L. G.; Mascola, J. R.;
Dondero, D.; Birx, D. L.; Sheppard, H. W. Nat. Med. 1997, 3,
338.
Figure 3. Structures and relative energies of alternative piperidine
conformations. A and B are the two chair conformations and C is a
twist boat conformation.

P. E. Finke et al. / Bioorg. Med. Chem. Lett. 11 (2001) 2469–2473
2473
5. For a review of HIV-1 entry mechanisms and current inhi-
bitors, see: (a) Blair, W. S.; Lin, P.-F.; Meanwell, N. A.; Wal-
lace, O. B. Drug Discovery Today 2000, 5, 183. (b) Horuk, R.;
Ng, H. P.; Med. Res. Rev. 2000, 2, 155.
6. Shiraishi, M.; Aramaki, Y.; Seto, M.; Imoto, H.; Nishi-
kawa, Y.; Kanzaki, N.; Okamoto, M.; Sawada, H.; Nishi-
mura, O.; Baba, M.; Fujino, M. J. Med. Chem. 2000, 43, 2049.
7. Armour, D. R.; Price, D. A.; Stammen, B. L. C.; Wood, A.;
Perros, M.; Edwards, M. P. EPO Patent 1,013,276 A1, 2000
(Pfizer); Chem. Abstr. 2000, 133, 74,024x.
8. Baroudy, B. M.; Clader, J. W.; Josien, H. B.; McCombie,
S. W.; McKittrick, B. A.; Miller, M. W.; Neustadt, B. R.;
Palani, A.; Smith, E. M.; Steensma, R.; Tagat, J. R.; Vice, S.
F.; Laughlin, M. A.; Gilbert, E.; Labroli, M. A. WO 00/66558,
2000 (Schering); Chem. Abstr. 2000, 133, 350, 248c.
15. Elderfield, R. C.; Fischer, B.; Lagowski, J. M. J. Org.
Chem. 1957, 22, 1376.
16. For a description of the binding assay, see ref 9, footnote
25. The CCR5 binding titrations were usually done in tripli-
cate with standard errors of <15%. Assay-to-assay variability
was generally less than ꢂ2-fold based on the standard com-
pound 3.
17. Condra, J. H.; Schleif, W. A.; Blahy, O. M.; Gabryelski,
L. J.; Graham, D. J.; Quintero, J. C.; Rhodes, A.; Robbins,
H. L.; Roth, E.; Shivaprakash, M.; Titus, D.; Yang, T.; Tep-
pler, H.; Squires, K. E.; Deutsch, P. J.; Emini, E. A. Nature
1995, 374, 569.
18. The results of the PBMC assay were highly variable with
IC95 values for 3 (used as the standard in each assay) ranging
from 1500 to 25,000 nM.
9. Dorn, C. P.; Finke, P. E.; Oates, B.; Budhu, R. J.; Mills,
S. G.; MacCoss, M.; Malkowitz, L.; Springer, M. S.; Daugh-
erty, B. L.; Gould, S. L.; DeMartino, J. A.; Siciliano, S. J.;
Carella, A.; Carver, G.; Holmes, K.; Danzeisen, R.; Hazuda,
D.; Kessler, J.; Lineberger, J.; Miller, M.; Schleif, W. A.;
Emini, E. A. Bioorg. Med. Chem. Lett. 2001, 11, 259.
10. Finke, P. E.; Meurer, L. C.; Oates, B.; Mills, S. G.; Mac-
Coss, M.; Malkowitz, L.; Springer, M. S.; Daugherty, B. L.;
Gould, S. L.; DeMartino, J. A.; Siciliano, S. J.; Carella, A.;
Carver, G.; Holmes, K.; Danzeisen, R.; Hazuda, D.; Kessler,
J.; Lineberger, J.; Miller, M.; Schleif, W. A.; Emini, E. A.
Bioorg. Med. Chem. Lett. 2001, 11, 265.
11. MacCoss, M.; Mills, S. G.; Shah, S. K.; Chiang, Y.-C. P.;
Dunn, P. T.; Koyama, H. US Patent 6,013,652, 2000; Chem.
Abstr. 2000, 132, 78470s.
12. Hale, J. J.; Finke, P. E.; MacCoss, M. Bioorg. Med. Chem.
Lett. 1993, 3, 319.
19. Piperidines are generally more basic than the correspond-
ing piperazines as evident from the pKa’s for N-methylpiper-
idine and N,N-dimethylpiperazine (10.1 and 8.1, respectively).
20. Conformations for 8c were generated using an in-house
distance geometry program (Kearsley, S. K., unpublished
results) and energy minimized using a distance-dependent
dielectric constant of 78 and the MMFFs forcefield.23 The
pharmacophore model was created by generating a consensus
overlay of several compounds in this series using MEGA-SQ.24
21. For the piperidine moiety, the conformational analysis
was performed starting from both chair structures as depicted
by A and B. Different ring conformations were generated and
minimized as described above20 to afford conformers A–C
with their indicated relative energies.
22. Finke, P. E.; Oates, B.; Mills, S. G.; MacCoss, M.; Mal-
kowitz, L.; Springer, M. S.; Gould, S. L.; DeMartino, J. A.;
Carella, A.; Carver, G.; Holmes, K.; Danzeisen, R.; Hazuda,
D.; Kessler, J.; Lineberger, J.; Miller, M.; Schleif, W. A.;
Emini, E. A. Bioorg. Med. Chem. Lett 2001, 11, 2475.
23. Halgren, T. A. J. Comp. Chem. 1999, 20, 730.
13. Abdel-Magid, A. F.; Carson, K. G.; Harris, B. D.; Mar-
yanoff, C. A.; Shah, R. D. J. Org. Chem. 1996, 61, 3849.
14. Mills, S. G.; Springer, M. S.; MacCoss, M. US Patent
6,013,644, 2000; Chem. Abstr. 2000, 132, 78,469y.
24. Miller, M. M. Med. Chem. Res. 1999, 9, 513.