Scaffold hopping in the rational design of novel HIV-1 non-nucleoside reverse transcriptase inhibitors

Article abstract of DOI:10.1016/j.bmcl.2007.03.097

High-throughput screening hit 1 was identified as a potent, broad-spectrum, non-nucleoside reverse transcriptase inhibitor (NNRTI) of HIV-1 replication. Analysis of the bound conformation of analogs of this inhibitor via molecular modeling and NMR contrib

Full text of DOI:10.1016/j.bmcl.2007.03.097

Bioorganic & Medicinal Chemistry Letters 17 (2007) 3362–3366
Scaffold hopping in the rational design of novel HIV-1
non-nucleoside reverse transcriptase inhibitors
*
´
Jeff A. O’Meara, Araz Jakalian, Steven LaPlante, Pierre R. Bonneau, Rene Coulombe,
Anne-Marie Faucher, Ingrid Guse, Serge Landry, Jennifer Racine, Bruno Simoneau,
Bounkham Thavonekham and Christiane Yoakim
Boehringer Ingelheim (Canada) Ltd, Research and Development, 2100 Cunard Street, Laval, Que., Canada H7S 2G5
Received 13 March 2007; revised 26 March 2007; accepted 29 March 2007
Available online 5 April 2007
Abstract—High-throughput screening hit 1 was identified as a potent, broad-spectrum, non-nucleoside reverse transcriptase inhib-
itor (NNRTI) of HIV-1 replication. Analysis of the bound conformation of analogs of this inhibitor via molecular modeling and
NMR contributed to the design of novel tertiary amide, carbamate, and thiocarbamate based NNRTIs.
Ó 2007 Elsevier Ltd. All rights reserved.
In the more than 20 years since AIDS was discovered it
has evolved into an unprecedented threat to global
health. Worldwide, the pandemic has infected 70 million
people, 25 million of whom have already succumbed to
the disease. Additionally, there were 4.3 million new
infections and 2.9 million deaths in 2006 alone.1
and closely related triazole derivatives are extremely po-
tent, broad-spectrum inhibitors of wild-type (WT) and
mutant HIV-1 reverse transcriptases. These compounds
however suffer from poor stability in the presence of hu-
man liver microsomes (HLM) which is predictive of ra-
pid clearance in humans. In order to facilitate patient
compliance, a demanding pharmacokinetic profile is
absolutely required in a contemporary NNRTI; we were
therefore challenged to address the metabolic weak-
nesses associated with these compounds.
In industrialized nations, an aggressive educational pro-
gram coupled with increasingly effective pharmaceutical
intervention has diminished the number of new infec-
tions and increased the life span of infected individuals
substantially. However, treatment failure as a result of
in vivo selection of multi-drug resistant mutants remains
a serious issue which must be continually addressed.
Organic sulfur compounds are known to be prone to
metabolic oxidation so we immediately considered the
thio-ether linkage as a potential metabolic liability.
However, SAR rapidly revealed the importance of the
sulfur atom in this lead series. Simply substituting the
sulfur with either oxygen (2), or carbon (3), led to 10-
and 27-fold decreases in intrinsic wild-type potency,
respectively (Table 1). It was interesting to note that
no improvements in HLM stability were observed by
these modifications. We therefore decided to continue
the exploration of the thio-tetrazole series in order to
As part of our effort to develop improved NNRTIs, we
were interested in evaluating a structurally distinct class
of inhibitors with good anti-viral activity versus clini-
cally relevant mutants and with DMPK properties con-
sistent with once per day oral administration. To this
end, a high-throughput screening campaign targeting
the prevalent NNRTI resistant double mutant K103N/
Y181C was initiated. The screen against this enzyme
identified thio-tetrazole derivative 1 as a lead structure.
We and others2a–i have revealed that these compounds
NO2
N N
H
N
N
S
N
O
Keywords: NNRTI; Scaffold hopping; Tertiary amide; Tertiary carba-
mate; Thio-tetrazole; Tetrazole thioacetanilide.
1.
*
Corresponding author. Tel.: +1 450 434 2381; e-mail:
jomeara@lav.boehringer-ingelheim.com
Figure 1. Hit identified in high-throughput screening.
0960-894X/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2007.03.097

J. A. O’Meara et al. / Bioorg. Med. Chem. Lett. 17 (2007) 3362–3366
Table 1. SAR of naphthyl tetrazole linkers
3363
R
N N
H
N
N
X
N
O
Compounds
X
R
IC50 (nM)
WT K103N/Y181C
t1/2 HLM
(min)
1
2
3
4
S
NO2 <20
NO2 200
849
11750
3
3
7
6
O
CH2
S
NO2 542 >10,000
Cl 20 1838
ascertain any structure–activity relationships while at
the same time monitoring the effects of different substi-
tutions on the metabolic stability. This exercise provided
significant improvements in potency, particularly versus
the enzyme containing the prevalent double mutant
K103N/Y181C. It was very quickly discovered that the
undesirable 2-nitro anilide could be replaced with a 2-
Cl anilide as in 4 (Table 1) with only a 2-fold decrease
in potency versus the double mutant. It was also
revealed that the tetrazole N-naphthyl could be replaced
by a 2-Cl phenyl derivative as in 5 (Table 2) resulting in
only a 2-fold loss in WT activity compared to 1. Fur-
thermore, it was discovered that the potency could be
significantly enhanced by the introduction of a substitu-
ent at the 4-position of this aromatic ring, para to the
tetrazole.
Figure 2. Model of 8 in the NNRTI binding pocket.
On the right-hand side of the molecule, further enhance-
ments in potency could be attained by substituting the 2-
Cl anilide at the 4-position. A wide array of functional
groups were tolerated at this position, with sulfone (9),
sulfonamide (10), and biphenyl derivatives 11 and 12
being optimal (Table 2). Replacement of the thio-ether
by an all carbon linker in these more potent analogs
provided 13 which was significantly less potent, losing
16-fold versus wild-type RT and nearly 80-fold versus
K103N/Y181C. Some of the structural changes
described above provided substantial gains in potency
and moderate improvements in the HLM stability
(i.e., compound 9) however, these compounds still did
not meet our target profile Figure 1.
The 4-methyl substituted analog 6 provided a 2-fold
improvement in potency versus K103N/Y181C RT rela-
tive to 5. Upon increasing the steric bulk of this substi-
tuent to cyclopropyl (7) or to tert-butyl (8), a further
8-fold gain versus the double mutant was realized and
compounds having IC50 values less than 100 nM versus
this enzyme were obtained. Interestingly, these improve-
ments were not observed versus the WT enzyme. In fact,
compounds 6, 7, and 8 were essentially equipotent
versus WT.
Molecular modeling provided a complex structure of
compound 8 bound in the WT NNRTI binding pocket
(Fig. 2). Notable features include: (1) a key H-bond
between the inhibitor’s amide carbonyl and the back-
bone N–H of K103; (2) an interesting CH–p interaction
in which the 4-substituent of the phenyl-tetrazole points
directly at, and perpendicular to, the indole side chain of
Trp229. Trp 229 is a highly conserved residue among all
Table 2. SAR of thio-tetrazole series
Y
H
N N
N
N
X
N
O
Cl
A
R
Compound
X
R
Y
A
IC50 (nM)
K103N/Y181C
t1/2 HLM (min)
WT
5
6
S
H
Me
NO2
Cl
H
39
9
1995
809
103
94
ND
2
S
H
7
S
c-Pr
t-Bu
t-Bu
t-Bu
t-Bu
t-Bu
t-Bu
Cl
Cl
H
13
8
ND
8
8
S
H
9
S
Cl
Cl
SO2Me
SO2NH2
3
27
36
45
23
7
10
11
12
13
S
6
S
Cl
Cl
Ph-4-OCH2CO2H
Ph-4-CH2CO2H
Ph-4-OCH2CO2H
7
23
29
S
CH2
8
115
23
ND
Cl
1777

3364
J. A. O’Meara et al. / Bioorg. Med. Chem. Lett. 17 (2007) 3362–3366
known RT sequences3 and has been implicated in tem-
plate primer utilization (3) the 4-position of the anilide
phenyl ring points directly toward a channel which is
lined by Pro236 and leads to solvent; (4) the inhibitor
binds in a ‘kinked’ conformation which involves a rota-
tion about the S–CH2–CO–N dihedral angle from 180°
in the fully extended free-state conformation to almost
0° in the bound state. This brings the two substituted
phenyl rings into close proximity and results in the
amide N–H pointing toward the sulfur.
amide conformation, was converted to the tertiary
amide 15. By 1H NMR in DMSO at room temperature,
15 exists as a 9:1 mixture of s-cis/s-trans amide rotamers.
Gratifyingly, a 22-fold improvement in potency was
observed relative to 14. However, tertiary amide 15
was 14-fold less potent compared to its reference thio-
tetrazole derivative 11 against K103N/Y181C and 96-
fold less potent versus the WT enzyme. It is interesting
to note that for the first time a compound was observed
to be more potent against the double mutant K103N/
Y181C than against WT reverse transcriptase. An
X-ray crystal structure was obtained for 16 bound to
the NNRTI binding pocket (data not shown) and in-
deed, it was found to bind as predicted with only minor
shifts in binding mode relative to that of the thio-tetra-
zole inhibitors. In order to further evaluate the potential
of the tertiary amide series, structure–activity relation-
ship studies were initiated.
Interestingly, however, no discernable interaction was
apparent for the thio-tetrazole motif which was vital
for potency. These data suggested that the tetrazole por-
tion of these inhibitors could simply be acting as a scaf-
fold which orients the pharmacophores into the proper
geometry for binding. Based on the above, we hypothe-
sized that alternate, potentially more stable scaffolds
could be designed.
Initially, the tertiary amide N-substituent was examined.
It was rapidly discovered that a small primary aliphatic
group was required. The N-ethyl analog 16 (Table 3)
was optimal with IC50‘s of 340 and 207 nM versus WT
and K103/Y181C. This was roughly 2-fold better than
N-methyl compound 15. The N-allyl derivative 17 was
equipotent to the N-ethyl compound 16 versus the dou-
ble mutant, however, it was less potent against wild-type
enzyme. The secondary aliphatic substituent N-cyclo-
propyl in 18 was not well tolerated, especially versus
the wild-type RT. Trifluoroethyl analog 19 represents
an attempt to subtly alter the electronic characteristics
of 16, while maintaining a similar steric volume. The
compound demonstrated a decreased affinity, losing
3.5-fold affinity against both enzymes.
Detailed examination of the bound conformation dem-
onstrated the obvious cis orientation of the tetrazole
appendages and that the N-2-chlorophenyl substituent
rests essentially perpendicular to the plane of the tetra-
zole. We therefore anticipated that we could mimic this
bound conformation with a variety of appropriately
substituted tertiary amides, carbamates or thiocarba-
mates. The structural rationale for this idea is as follows:
Tertiary amides such as 15 (Table 3) exist in either the
s-cis or s-trans conformation. The s-cis conformation
is preferred because of an unfavorable 1,5-steric interac-
tion between the N-methyl substituent and the protons
a to the carbonyl in the s-trans orientation. Further-
more, as in the bound conformation of the tetrazole
derivatives, the aromatic ring is forced to be perpendic-
ular to the plane of the amide in order to minimize a
similar destabilizing steric interaction between the aro-
matic o-substituents and the protons a to the tertiary
amide carbonyl.
We next focused on the influence of altering the nature
of the functional group of the alternate scaffold. The
general synthetic route utilized to obtain the desired
compounds was straightforward and is depicted in
Scheme 1.6 Chlorination of commercially available 4-
tert-butyl aniline (20) was accomplished in acetonitrile
using N-chlorosuccinimide. The 2-chloro aniline was
then treated with acetyl chloride and the resulting amide
was reduced using borane–dimethylsulfide complex to
give the secondary aniline 21 in good yield. Compound
21 was treated with 3-carbomethoxypropionyl chloride
to provide the amide intermediate. Alternatively, 21
was treated with phosgene to access the carbamoyl chlo-
ride which was in turn treated with the sodium salts of
methyl glycolate, methyl thioglycolate or with glycine
methyl ester hydrochloride to give the carbamate, thioc-
arbamate, and urea intermediates. Treatment of these
intermediates with lithium hydroxide generated the cor-
responding acids. Finally, amide formation with aniline
22 in pyridine using phosphorus trichloride followed by
treatment with sodium hydroxide in DMSO provided
the desired inhibitors 23, 24, 25, and 26.
The scaffold hopping5 hypothesis was evaluated when
secondary amide 14, a very weak inhibitor of K103N/
Y181C which exists almost exclusively in the s-trans
Table 3. SAR of tertiary amides
O
O
R
Cl
N
HN
Cl
OH
O
O
Compound
R
IC50 (nM)
t1/2 HLM
(min)
WT
K103N/Y181C
14
15
16
17
18
19
H
Me
6060
673
4620
328
207
237
506
718
ND
54
6
The amide 23 (Table 4) was prepared as a reference for
this study and it is noteworthy that the small change in
the biphenyl substitution from phenoxy-acetic acid
(compound 19) to the shorter acetic acid in compound
23 translated to 2.5-fold decrease in wild-type potency.
Et
340
520
Allyl
c-Pr
CH2CF3
9
1922
1117
15
ND

J. A. O’Meara et al. / Bioorg. Med. Chem. Lett. 17 (2007) 3362–3366
Cl
3365
O
H
N
NH2
HN
N
X
a,b, c
d, e, f, g ,
Cl
h
O
Cl
24 X = O
25 X = N
26 X = S
20
21
OH
O
i, g, h
Cl
Cl
H
O
H2N
N
N
O
Cl
22
O
O
OH
O
23
Scheme 1. Reagents and conditions: (a) N-chlorosuccinimide, CH3CN; (b) acetyl chloride, pyridine; (c) BH3–Me2S, toluene, reflux; (d) phosgene,
Et3N, THF; (e) methyl glycolate, NaH, THF or methyl thioglycolate, NaH, THF or glycine methylesterÆHCl, Et3N, THF; (f) LiOH, THF, MeOH;
(g) 22, PCl3, pyridine; (h) NaOH, DMSO; (i) 3-carbomethoxypropionyl chloride, pyridine, THF.
Table 4. SAR of tertiary amides, carbamates, and thiocarbamates
relative to both oxygen and CH2. This inhibitor is essen-
tially equipotent to the corresponding thio-tetrazole
O
derivative 12, versus the double mutant K103N/
Y181C. However, against wild-type enzyme the com-
pound is 4-fold less potent compared to 12. This sug-
gests that in the wild-type enzyme, the tetrazole is
more than simply a scaffold and is likely involved in rel-
atively weak associations with the side chains of Y181
and/or K103. It should be noted that none of these com-
pounds resulted in a substantial improvement in the
HLM stability.
O
X
Cl
N
HN
Cl
O
OH
Compound
X
IC50 (nM)
t1/2 HLM (min)
WT
K103N/Y181C
23
24
25
26
C
O
N
S
829
109
1530
34
305
98
3
9
In summary, a thorough understanding of the binding
mode of the thio-tetrazole based inhibitors permitted
the scaffold hop from tetrazoles to tertiary amides, car-
bamates, and thiocarbamates. Although these com-
pounds did not provide the desired improvement in
metabolic stability, they represent a novel,4 potent class
of NNRTIs with a broad spectrum of antiviral activity.
414
35
23
4
This decrease in affinity was not as pronounced versus
the double mutant K103N/Y181C.
Carbamate 24 exists as a 3:1 mixture of s-cis/s-trans
rotamers in DMSO (at room temperature) and is a po-
tent inhibitor of both WT and K103/Y181C reverse
transcriptases with IC50’s of 109 and 98 nM, respec-
tively. This represents an 8-fold improvement vs WT
and a 3-fold improvement versus the double mutant rel-
ative to the amide 23. The lower selectivity for the s-cis
conformation in 24 can be explained by the smaller vol-
ume occupied by oxygen and its two lone pairs relative
to CH2 in 23. The oxygen’s smaller volume compared
to that of CH2 results in a decreased repulsive steric
interaction with the N-ethyl substituent in the s-trans
conformation and therefore has a weaker influence on
overall molecular topography.
References and notes
1. Joint United Nations Programme on HIV/AIDS and
World Health Organization Aids Epidemic Update
2006.
2. (a) Girardet, J. L.; De la Rosa, M.; Kim, H. W.; Gunic, E.;
Jenket, C.; Boyle, U., et al. Bioorg. Med. Chem. Lett. 2006,
16, 4444; (b) Zhang, Z.; Xu, W.; Koh, Y. H.; Shim, J. H.;
Giradet, J. L.; Yeh, L. T.; Hamatake, R. K.; Hong, Z.
Antimicrob. Agents Chemother. 2007, 429; (c) He, Y.;
Wand, A.; Wu, B.; Kuhen, K. L.; Bursulaya, B.; Nguyen,
T. N.; Nguyen, D. G. Bioorg. Med. Chem. Lett. 2006, 16,
4174; (d) Muraglia, E.; Kinzel, O. D.; Laufer, R.; Miller,
M. D.; Moyer, G., et al. Bioorg. Med. Chem. Lett. 2006, 16,
2748; (e) Girardet, J. L.; Zhang, Z.; Hamatake, R.;
Hernandez M. De la Rosa; Gunic, E.; Hong, Z.; Kim, H.
W.; Koh, Y. H.; Nilar, S.; Shaw, S.; Yao, N. Patent WO
2004030611, 2004; (f) Simoneau, B.; Thavonekham, B.;
Landry, S.; O’Meara, J.; Yoakim, C.; Faucher, A.-M.
Patent WO 2004050643, 2004; (g) Shaw-Reid, C. A.; Miller,
M. D.; Hazuda, D. J.; Ferrer, M.; Sur, S. M.; Summa, V.;
Lyle, T. A.; Kinzel, O.; Pescatore, G.; Muraglia, E.;
The urea 25 was not well tolerated and was 2-fold less
potent than the corresponding amide 23. Thiocarbamate
26, on the other hand, is extremely potent showing
equivalent IC50 values of 35 nM versus both WT and
1
K103/Y181 enzymes. By H NMR compound 26 exists
in essentially only the s-cis conformation, an observa-
tion explained by the greater volume occupied by sulfur

3366
J. A. O’Meara et al. / Bioorg. Med. Chem. Lett. 17 (2007) 3362–3366
Orvieto, F.; Williams, P. D. Patent WO2005115147, 2005;
4. A patent covering closely related analogs was disclosed
during the preparation of this manuscript. Koch, U.;
Kinzel, O.; Muraglia, E.; Summa, V. Patent WO
2006037468, 2006.
5. Scaffold hopping has recently been reviewed and defined.
Brown, N.; Jacoby, E. Mini Rev. Med. Chem. 2006, 11, 1217.
6. The tetrazole based inhibitors were prepared in a similar
route as previously reported in Ref. 2d.
(h) Girardet, J. L.; Koh, Y. H.; Hernandez M. De la Rosa;
Gunic, E.; Kim, H. W.; Hong, W. Patent WO 2006026356,
2006; (i) Simoneau, B.; Thavonekham, B.; Landry, S.;
O’Meara, J.; Yoakim, C.; Faucher, A.-M. Patent WO
2005118575, 2005.
3. Pelemans, H.; Esnouf, R.; De Clercq, E.; Balzarini, J. Mol.
Pharmacol. 2000, 57, 954.