Respiratory syncytial virus fusion inhibitors. Part 5: Optimization of benzimidazole substitution patterns towards derivatives with improved activity

Article abstract of DOI:10.1016/j.bmcl.2007.05.102

Extensive SAR studies and optimization of ADME properties of benzimidazol-2-one derivatives led to the identification of BMS-433771 (3) as an orally active RSV fusion inhibitor. In order to extend the structure-activity relationships for this compound ser

Full text of DOI:10.1016/j.bmcl.2007.05.102

Bioorganic & Medicinal Chemistry Letters 17 (2007) 4592–4598
Respiratory syncytial virus fusion inhibitors. Part 5:
Optimization of benzimidazole substitution patterns
towards derivatives with improved activityq
Xiangdong Alan Wang,a,* Christopher W. Cianci,b Kuo-Long Yu,a,
Keith D. Combrink,a,à Jan W. Thuring,a,§ Yi Zhang,a Rita L. Civiello,a
Kathleen F. Kadow,b Julia Roach,b Zhufang Li,b David R. Langley,c
Mark Krystalb and Nicholas A. Meanwella
aDepartment of Chemistry, Bristol-Myers Squibb Research and Development, 5 Research Parkway, Wallingford, CT 06492, USA
bDepartment of Virology, Bristol-Myers Squibb Research and Development, 5 Research Parkway, Wallingford, CT 06492, USA
cDepartment of Computational-Aided Drug Design, Bristol-Myers Squibb Research and Development,
5 Research Parkway, Wallingford, CT 06492, USA
Received 9 April 2007; revised 23 May 2007; accepted 25 May 2007
Available online 8 June 2007
Abstract—Extensive SAR studies and optimization of ADME properties of benzimidazol-2-one derivatives led to the identification
of BMS-433771 (3) as an orally active RSV fusion inhibitor. In order to extend the structure–activity relationships for this com-
pound series, substitution of the benzimidazole ring was examined with a view to establishing additional productive interactions
between the inhibitor and functionality present in the proposed binding pocket. Amongst the compounds synthesized, the 5-ami-
nomethyl analogue 10aa demonstrated potent antiviral activity towards wild-type RSV and retained excellent inhibitory activity
towards a virus that had been developed to express resistance to BMS-433771 (3), data consistent with an additional productive
interaction between the inhibitor and the fusion protein target.
Ó 2007 Elsevier Ltd. All rights reserved.
Respiratory syncytial virus (RSV) is a leading cause of
lower respiratory tract infection in infants and children.5
Moreover, the recurrence of RSV infection in later life, a
consequence of an immune response of poor durability,
poses a serious health threat to the elderly and immuno-
compromised.6 Despite the significant morbidity and
mortality associated with RSV infections, the only ther-
apeutic option currently available is the nucleoside ana-
logue ribavirin, typically administered as an aerosol.7,8
However, ribavirin is neither a particularly potent nor
specific inhibitor of RSV and its teratogenic properties
complicate administration.8 Consequently, there is an
unmet medical need for effective and specific therapies
to treat this disease.
We have recently described the discovery of a class of
benzimidazole-based inhibitors of RSV fusion, generi-
cally represented by 1,1 from which evolved a series of
more potent RSV inhibitors based on the benz-
imidazol-2-one template 2.2 This chemotype tolerated
a broad range of substituents appended to both the
benzimidazole and benzimidazol-2-one moieties that
encompassed a wide variation of size and functionality.2
In vivo efficacy towards RSV infection was demonstrated
in the cotton rat model with water-soluble derivatives
delivered by small particle aerosol.3 Further optimization
for antiviral potency, membrane permeability and
metabolic stability in human liver microsomes led to the
identification of BMS-433771 (3) as a potent RSV inhib-
itor with oral bioavailability in the mouse, rat, dog and
cynomolgus monkey that demonstrated antiviral activity
Keywords: RSV fusion inhibitor; Substitution of the benzimidazole;
Binding pocket; 5-Aminomethyl analogue; Retained activity; Resistant
virus; Asp200.
q
For parts 1–4 of this series, see Refs. 1–4.
Corresponding author. Tel.: +1 203 677 6540; fax: +1 203 677
7702; e-mail: xiangdong.wang@bms.com
*
Present address: Lilly Research Laboratories, Eli Lilly and Com-
pany, Indianapolis, IN 46285, USA.
à
Present address: Cumbre Inc., 1502 Viceroy Drive, Dallas, TX 75235,
USA.
§
Present address: Johnson & Johnson Pharmaceutical Research &
Development, Turnhoutsebaan 30, B-2340 Beerse, Belgium.
0960-894X/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2007.05.102

X. A. Wang et al. / Bioorg. Med. Chem. Lett. 17 (2007) 4592–4598
4593
O
O
N
N
N
N
R2
N
N
N
N
N
N
N
N
N
R1
R1
N
OH
1
2
3: BMS-433771
in the BALB/c and cotton rat models of infection follow-
ing oral administration.4,9,10
Mechanistic and photo-affinity labelling studies indi-
cated that this class of RSV inhibitors interfered with
the assembly of the F protein 6-helix bundle, a critical
step in the fusion of host cell and virus membranes.11,12
The 6-helix bundle structure is thought to be formed in a
stepwise fashion as part of the RSV F1 protein rear-
rangement in which the lipophilic amino terminus, the
fusion peptide, is inserted into the host cell membrane
and a proximal heptad repeat region associates into a
trimeric structure. This trimer element subsequently
interacts with a complementary heptad repeat present
in the carboxy terminal that binds to it in an anti-paral-
lel fashion, creating the 6-helix bundle that is thought to
mediate the coalescence of host and viral membranes.
Modelling studies suggested that 1–3 and related com-
pounds bind into a pocket created by the association
of the amino terminus heptad repeat into the trimeric
structure.11,12 The favoured binding pose placed the
benzimidazole heterocycle in the cavity occupied by
Phe483 of the carboxy terminus heptad repeat in the
fusogenic state whilst the benzimidazol-2-one element
extended into the pocket accommodating Phe488 and
Leu492 of the C-terminus.11 The proposed binding
mode provided an opportunity for structure-based
inhibitor design involving an examination of substitu-
tion patterns associated with the benzimidazole hetero-
cycle and also suggested potential avenues towards
further enhancing the antiviral activity of the chemo-
type. More specifically, the binding mode hypothesis
suggested that substitution at C-4 and C-7 of the benz-
imidazole would be poorly tolerated, since these sites
are in close proximity to elements of the protein, whilst
C-5 and C-6 would be more accommodating. More
intriguingly, the vectors associated with C-5 and C-6
project towards Asp200, suggesting that the introduc-
tion of basic substituents in this region may lead to
the establishment of a productive salt bridge or hydro-
gen bonding interaction (Fig. 1). In this communication,
we describe the results of this initiative which led to the
identification of potent inhibitors of wild-type RSV
in vitro that demonstrated improved activity towards a
resistant strain, consistent with the notion that an addi-
tional interaction between the inhibitor and the viral fu-
sion protein had been established.13
Figure 1. Engaging Asp200 in the proposed binding site.
nitrobenzene 4 as the starting material, as depicted in
Scheme 1. The introduction of the side chain ultimately
attached to the benzimidazole ring was accomplished by
displacement of the chlorine or fluorine atom of 4 by iso-
amylamine to give the nitroaniline 5 in good yield.
Hydrogenation of the nitro group in 5, catalyzed either
by Pd/C (when X = CN) or Raney nickel (when X = Br),
afforded the diamine 6 quantitatively. Acylation and
ring closure to the benzimidazole 7 was effected by reac-
tion with 2-(benzyloxy)acetyl chloride. The benzyl pro-
tecting group of 7 was removed by treating with BBr3
and the resultant alcohol converted to the chloride 8
using SOCl2. Coupling of the chloride 8 with a mono-
substituted benzimidazol-2-one 9, readily obtained by
several complementary pathways,14–19 was accom-
plished using either Cs2CO3/acetone, NAH/DMF, or
BTPP/THF5 to afford bromo- or cyano-substituted
derivatives 10. The bromide moiety of 10i was converted
to the 5-ethyl analogue 10q by a Pd-catalyzed Stille20
coupling with tributylvinyltin, to afford the vinyl substi-
tuted compound 10p, followed by hydrogenation. Ami-
nation of 10i with diphenylmethanimine under
Buchwald–Hartwig conditions,21 followed by acid
hydrolysis, provided the aniline 10z. The nitrile-substi-
tuted derivatives of 10 were particularly versatile inter-
mediates, readily elaborated into a series of substituted
analogs, as summarized in Scheme 2. Hydrogenation
of the nitrile moiety of 10j and 10aq provided the amino-
methyl derivatives 10aa and 10ar, respectively (Scheme
2, step b1), which were readily acylated or sulfonylated
(Scheme 2, steps c and d). Heating 10j with NaN3 in
DMF provided the tetrazole 10v (Scheme 2, step p)
whilst reaction with NH2OH in EtOH at reflux provided
hydroxyamidines 10al–an (Scheme 2, step q). The disub-
stituted aminomethyl compound 10ah was obtained by
the reaction of methylcerium with the nitrile 10j (Scheme
2, step a).22 Hydrolysis of the nitrile moiety provided the
structurally analogous carboxylic acid derivatives
The target molecules 10 were assembled by the pathways
outlined in Schemes 1 and 2. Substituents were intro-
duced to the benzimidazole ring by employing either a
bromo- or cyano-substituted-2-chloro- or 2-fluoro-

4594
X. A. Wang et al. / Bioorg. Med. Chem. Lett. 17 (2007) 4592–4598
O
NO2
NO2
Cl,F
NH2
NH
X
1) BBr3
H2 Pd/C
OBn
N
N
X
Cl
X
X
2) SOCl2
NH
or Raney
Nickel
when X=Br
OBn
K2CO3
CH3CN
H2N
X=CN or Br
4
5
6
7
R2
N
O
N
N
Et
N
N
N
H
1) tributylvinytin
O
X
O
X
Pd(PPh3)4/toluene/10i
9
R2
N
N
N
N
Cl
N
R2
N
2)H2/Pd/C/MeOH
Cs2CO3
or NaH
or BTPP
8
10q
1) 10i/Ph2C=NH//NaOtBu
BINAP/Pd2(dba)3/toluene
2) HCl/THF
10: X = 5-CN or Br
10z: X = 5-NH2
Scheme 1.
HO
H2N
q
N
O
O
N
N
S
O
N
H3C
N
H
N
H
N
H
H3C
p
d
O
c
N
N
N
N
b1
a
H2N
H2N
NC
e
O
O
O
j
i
g
h
H3C
F
N
HO
HO
H3CO
CH3
f
H2N
O
H3C
k
N
H
H2N
m
b2
l
o
MsO
N
H
NC
n
H3C
N
CH3
Scheme 2. Reagents and conditions: (a) CeCl3, MeLi, THF; (b1 and b2) H2, Pd/C, MeOH; (c) acetyl chloride, TEA, THF; (d) MsCl, TEA, THF; (e)
6 M NaOH, refluxing; (f) ammonia, EDC, DCM; (g) dimethylamine, EDC, DCM; (h) MeOH, H2SO4; (i) LAH, THF; (j) DAST, DCM; (k) MsCl,
TEA, DCM; (l) KCN, DMF; (m) methylamine, THF; (n) dimethylamine, THF; (o) cyclopropylamine, THF; (p) NaN3, DMF; (q) NH2OH, refluxing
EtOH.
(Scheme 2, step e) which were coupled with ammonia or
dimethylamine, to afford the corresponding amides
(Scheme 2, steps f and g), or heated in acidic MeOH
to provide the methyl ester derivatives (Scheme 2, step
h). Reduction of the methyl ester using LAH in THF
afforded the corresponding benzyl alcohol (Scheme 2,
step i), which was converted to a fluoride by exposure
to DAST (Scheme 2, step j). Alternatively, mesylation
of the benzyl alcohol (Scheme 2, step k) and subsequent
displacement with KCN yielded the benzyl cyanide
(Scheme 2, step l) which, in turn, was hydrogenated to
give the aminoethyl homologue (Scheme 2, step b2).
Reaction of the mesylate with methylamine, dimethyl-
amine and cyclopropylamine afforded a series of benzyl-
amines in good yield (Scheme 2, steps m–o). The
compounds prepared as part of this survey are compiled
in Table 1.
Inhibition of the Long A strain of RSV replicating in the
HEp-2 human lung epithelial carcinoma cell line was
used to evaluate the antiviral activity of target com-
pounds by monitoring the RSV-induced cytopathic
effect (CPE).9,10 Long strain virus resistant to BMS-
433771 (3) that expressed a K394R mutation in the F1
protein was generated in HEp-2 cells by passage of virus
in the presence of increasing concentrations of 2-(2-((3-

X. A. Wang et al. / Bioorg. Med. Chem. Lett. 17 (2007) 4592–4598
4595
Table 1. Structure, RSV inhibitory activity and cytotoxicity associated with a series of 5-substituted benzimidazole derivatives
O
4
N
R2
5
1
N
N
N
R
6
7
R1
R2
EC50 (lM)
CC50 (lM)
Therapeutic indexb
a
Compound
2a
H
Isopropenyl
i-Pr
i-Pr
0.004 (0.005, 0.003)
0.715 (0.395, 1.03)
0.230 (0.375, 0.084)
0.82 (1.34, 0.30)
24.6 (37.6, 11.7)
7.00 2.50 (n = 3)
3.92 (3.71, 4.12)
22.0 (14.1, 29.8)
26.7 (20.5, 32.9)
49.2 (7.39, 91)
6162
9.78
17.0
26.8
3.56
0.213
0.482
0.096
0.011
1006
16.6
26.0
12.6
114
10a
10b
10c
10d
10e
10f
10g
10h
10i
4-CH2NH2
4-CH2CH2NH2
4-CH2OH
4-CH2OCH3
4-CO2CH3
4-CH2CN
4-CH2CH2CN
4-CH2CH2CH2NH2
5-Br
i-Pr
i-Pr
7.50 (7.50, 7.51)
230 (230, 230)
i-Pr
i-Pr
241 (241, 241)
234 (234, 234)
213 (213, 213)
116 (182, 49.4)
2.24 (1.38, 3.10)
2.34 (1.54, 3.13)
18.1 (15.4, 20.8)
7.93 (10.9, 4.96)
1.48 (1.04, 1.96)
32.2 17.8 (n = 3)
13.3 (19.8, 6.81)
0.74 (0.50, 0.98)
5.20 2.00 (n = 3)
6.52 (4.45, 8.58)
18.5 (33.7, 6.37)
i-Pr
i-Pr
i-Pr
i-Pr
0.018 (0.012, 0.025)
0.476 (0.298, 0.655)
0.057 (0.026, 0.088)
2.55 2.22 (n = 3)
0.117 (0.153, 0.081)
0.179 (0.264, 0.094)
0.193 0.178 (n = 3)
0.630 (0.490, 0.770)
0.692 (0.420, 0.963)
10j
5-CN
5-CN
5-CN
10k
10l
Isopropenyl
t
CH2CO2 Bu
10m
10n
10o
10p
10q
5-CH2CN
5-CH3C(O)
5-CH3C(O)
5-CH2=CH
5-CH3CH2
i-Pr
Isopropenyl
Me
i-Pr
4.1
39.8
10.3
26.8
i-Pr
5
N
O
10r
10s
i-Pr
21.6 20.2 (n = 3)
239 (n = 1)
223 2.9 (n = 3)
2.97 (n = 1)
10.3
N
5
N
N
N
Me
0.012
N
10t
10u
5-CO2H
5-CO2H
Isopropenyl
i-Pr
264 (229, 299)
238 (n = 1)
30.0 (26.8, 33.2)
17.3 (n = 1)
0.11
0.07
5
N
N
10v
i-Pr
225 (225, 225)
2.00 ( 1.60, 2.41)
0.009
HN
N
10w
10x
5-CO2Me
5-C(O)NH2
i-Pr
i-Pr
230 (230, 230)
33.7 (15.5, 51.9)
23.4 (21.1, 25.6)
25.6 (12.5, 38.6)
32.2 (23.2, 41.3)
3.92 (3.23, 4.62)
11.0 (10.8, 11.2)
1.48 (1.04, 1.96)
17.2 (26.7, 7.62)
5.65 (3.30, 8.00)
178 (178, 178)
0.15
141
0.166 (0.146, 0.187)
24.8 (20.4, 29.1)
0.125 (0.206, 0.044)
0.002 (0.002, 0.002)
4.33 (1.16, 7.5)
10y
10z
5-C(O)N(CH3)2
5-NH2
5-CH2NH2
i-Pr
i-Pr
1.03
258
10aa
10ab
10ac
10ad
10ae
10af
10ag
10ah
10ai
10aj
10ak
10al
10am
10an
10ao
10ap
10aq
10ar
10as
10at
10au
10av
10aw
10ax
10ay
i-Pr
i-Pr
1962
2.54
5-CH2CH2NH2
5-CH2NH.CH3
5-CH2NH.c-Pr
5-CH2N(CH3)2
5-CH2NHC(O)CH3
5-CH2NHS(O)2Me
5-C(CH3)2NH2
5-C(NH)NH2
5-C(NH)NH2
5-C(NH)NH2
5-C(NOH)NH2
5-C(NOH)NH2
5-C(NOH)NH2
5-CH2OH
i-Pr
i-Pr
0.057 (0.026, 0.088)
0.363 (0.313, 0.414)
183 (183, 183)
26.0
47.3
i-Pr
i-Pr
0.031
282
0.631 (0.357, 0.905)
0.725
i-Pr
i-Pr
30.8 (21.4, 40.1)
3.27 (0.50, 0.98)
10.8 (10.2, 11.4)
7.90 (7.40, 8.40)
173 5.2 (n = 3)
3.47 (2.22, 4.72)
9.14 (11.2, 7.07)
19.8 (10.6, 29.1)
31.6 17.7 (n = 3)
37.3 (29.2, 80.6)
10.0 (17.5, 2.53)
28.1 (16.6, 40.2)
81.6 (119, 44.3)
4.81 3.41 (n = 3)
26.7 15.4 (n = 3)
57.0 87.4 (n = 3)
125
42.4
4.71
0.694 (0.689, 0.698)
0.004 (0.004, 0.004)
0.009 (0.006, 0.012)
0.031 0.011 (n = 3)
0.005 (0.006, 0.005)
0.111 (0.194, 0.028)
0.016 (0.014, 0.018)
0.050 0.044 (n = 3)
0.047 (0.037, 0.058)
0.044 (0.026, 0.063)
0.095 (0.118, 0.071)
0.134 (0.220, 0.047)
187
Isopropenyl
t
CH2CO2 Bu
2700
878
CH2CO2H
Isopropenyl
i-Pr
5574
694
82.2
1240
t
CH2CO2 Bu
i-Pr
i-Pr
i-Pr
i-Pr
i-Pr
i-Pr
i-Pr
i-Pr
i-Pr
i-Pr
i-Pr
632
793
5-CH2F
6-CN
228
296
6-CH2NH2
6-CH2OH
609
0.026
6-CH2NHCH3
6-CH2N(CH3)2
6-CH2CH2NH2
6-CO2H
182
0.026
8.02
7.11 6.89 (n = 3)
238
123 (238, 8.04)
0.525
0.406
6-C(O)NH2
7-CH2NH2
49.9 (42.6, 57.2)
11.0 (13.6, 8.50)
192 (192, 192)
0.058
(continued on next page)

4596
X. A. Wang et al. / Bioorg. Med. Chem. Lett. 17 (2007) 4592–4598
Table 1 (continued)
a
Compound
R1
R2
EC50 (lM)
CC50 (lM)
Therapeutic indexb
10az
10aaa
10aab
7-CO2CH3
7-CH2OH
7-N3
i-Pr
i-Pr
i-Pr
230 (230, 230)
246 (246, 246)
3.70 (6.45, 0.94)
25.3 (15.5, 35.1)
246
57.9 (52.9, 62.9)
0.11
1.00
15.6
a Values are means of two or more experiments performed on consecutive weeks with the data from individual experiments shown in parentheses.
b Therapeutic index = CC50/EC50
.
iodo-1H-indazol-1-yl)methyl)-1H-benzo[d]imidazol-1-yl)-
N,N-dimethylethanamine (BMS-243458), as described
previously.9 The EC50 values presented in Table 1 repre-
sent the concentration of test compound preserving 50%
of infected cells whilst the CC50 data are the concentra-
tion of drug at which cytotoxicity towards 50% of
uninfected HEp-2 cells was observed. The therapeutic
index reported in Table 1 is the ratio of CC50 to EC50.
Compounds were typically tested in two experiments
conducted on consecutive weeks with additional evalua-
tions performed if there was a significant discrepancy
between the initial data sets. Where the data reported
are the average of two experiments, the individual
results are provided as a measure of assay variability.
The introduction of substituents at C-5 proved to be a
considerably more productive exercise, with the conse-
quence that the survey conducted at this position of
the benzimidazole heterocycle is more extensive, exam-
ined in the context of the series of derivatives 10i–ap.
The structure–activity relationships enumerated by
10i–ap are both interesting and informative, providing
critical insights into aspects of the pharmacophore and
the proposed mode of binding of this series of fusion
inhibitors to the RSV F1 protein trimer. Small aliphatic
elements, bromine, cyanide, cyanomethyl, and acetyl
substituents afford compounds 10i–q that are quite po-
tent antiviral agents when compared to the progenitor
2a, with the weaker activity associated with 10l attribut-
able to the acetic ester moiety appended to the benz-
imidazol-2-one.2 The poor activity associated with
oxadiazole 10r is suggestive of limited tolerance for ste-
ric bulk at C-5, an observation further explored with the
methylated tetrazole 10s, which is not only poorly ac-
tive, but also cytotoxic. The two C-5 carboxylates 10t
and 10u and the isosteric23 tetrazole 10v are also cyto-
toxic at micromolar concentrations. The ester 10w dem-
onstrates cytotoxicity comparable to the acid 10u,
possibly a consequence of esterolytic cleavage under
the assay conditions. Stabilization of this moiety by con-
version to the primary amide 10x revealed the inherent
potency of a non-negatively charged element at this site
but increasing the steric bulk by dimethylation led to a
100-fold reduction in potency. The C-5 amino derivative
10z proved to be modestly potent but was substantially
improved by homologation to the more basic amino-
methyl derivative 10aa, a compound with antiviral activ-
ity comparable to the parent 2a. Informatively,
extension of the amino moiety of 10aa by a single car-
bon atom, to give 10ab, led to a precipitous 2000-fold
decline in potency, indicative of a precise disposition
for the basic amine relative to the heterocyclic nucleus.
Whilst 10aa and 2a demonstrate comparable activity to-
wards wild-type virus, a more effective indicator of the
intrinsic antiviral properties of the aminomethyl deriva-
tive 10aa is revealed by the EC50 for inhibition of the
K394R resistant virus.9 The K394R mutation in the
RSV F1 protein confers resistance ranging from 35- to
>1250-fold towards representatives of this fusion inhib-
itor chemotype but remains highly sensitive to inhibition
by 10aa, which displays an EC50 of 20 nM, as assessed
by the reduction in RSV-specific protein synthesis.9 In
contrast, the EC50 for BMS-433771 (3) and related com-
pounds towards the K394R-containing virus is >20 lM.
These data suggest the presence of an additional positive
and specific interaction between 10aa and the target pro-
tein and are consistent with the formation of a salt
bridge interaction between the amine and Asp200 in
The proposed model of the binding of this class of RSV
fusion inhibitors to the assembled trimer of the N-termi-
nal heptad repeat elements suggests that the 4- and 7-
positions of the benzimidazole ring are proximal to the
wall of the binding pocket. Consequently, antiviral
activity would be anticipated to be quite sensitive to sub-
stitution at either of these sites of the core heterocycle. It
is clearly apparent from the data presented in Table 1
for the series of 4-substituted analogues 10a–h and the
7-substituted derivatives 10ay–aab that the introduction
of substituents at these positions of the heterocyclic nu-
cleus results in a significant reduction in potency com-
pared to the parent compound 2a.2 Both an
aminomethyl (10a) and a hydroxymethyl (10c) substitu-
ent at C-4 result in over a 150-fold abrogation of po-
tency, a circumstance marginally improved by the
aminoethyl homologue 10b but considerably degraded
by the larger or more rigid polar substituents exempli-
fied by 10d–h. The 7-position is particularly sensitive
to modification and, with the exception of azide
10aab, all of the compounds examined in this series,
10ay–aaa, are at least 30,000-fold less potent than the
unsubstituted prototype 2a. Only the azide 10aab
demonstrates measurable RSV inhibitory activity but
this compound is, nevertheless, still 900-fold weaker
than 2a.
The C-6 position proved to be more tolerant of substitu-
tion with cyanide (10aq), aminomethyl (10ar) and
hydroxymethyl (10as) derivatives potent inhibitors of
RSV, although they are inferior to the parent compound
2a by an order of magnitude. Homologation of the ami-
nomethyl derivative 10ar to the aminoethyl substituted
compound 10av erodes potency by 75-fold whilst the
mono- and di-methylated amines 10at and 10au are
inactive. The final compound in this series, the carbox-
ylic acid 10aw, is cytotoxic without demonstrating sig-
nificant antiviral activity.

X. A. Wang et al. / Bioorg. Med. Chem. Lett. 17 (2007) 4592–4598
4597
the active site model proposed for these molecules. A
series of derivatives of 10aa prepared to further probe
the requirements of this interaction lend support to this
concept. Alkylation of 10aa leads to reduced potency,
with the methyl analogue 10ac and the less basic24 cyclo-
propylamine 10ad significantly weaker antiviral agents,
whilst the dimethyl derivative 10ae is essentially inactive.
Similarly, acylation (10af) and sulfonylation (10ag) lead
to substantially diminished antiviral activity whilst
gem-dimethylation of the benzylic carbon (10ah), a
modification that adds steric bulk and is anticipated to
affect conformational preference,25 also erodes activity.
Substitution at the benzylic methylene atom in the
context of an amidine, compounds 10ai–k, provided a
symmetrical and more basic element that conferred po-
tent RSV inhibition. However, in this series, antiviral
activity is not absolutely dependent on basicity since
the hydroxyamidines26 10al–n retain activity, possibly
a reflection of productive hydrogen bonding interactions
with Asp200. The survey was completed with the evalu-
ation of the hydroxymethyl (10ao) and fluoromethyl
(10ap) analogues, both of which are more potent than
the 5-ethyl derivative 10q.
Figure 2. A model of 10aa in the binding site of the RSV F1 protein
amino terminus heptad repeats showing the proposed intermolecular
interactions with Asp200. The model depicts 10aa bound into a pocket
created by two of the a-helixes, designated A (green) and B (light blue).
All N atoms are depicted in blue, O atoms in red and H atoms in white
with only polar H atoms depicted for simplicity. This model is based
on the 1G2C.pdb file.12
tein that showed cross-resistance to BMS-433771 (3),
strongly suggesting that these two compounds share a
similar mode of action.9,32 Whilst BABIM (11) is a topo-
logically symmetrical molecule in contrast to the asym-
metric silhouette presented by BMS-433771 (3) and
closely related compounds, the structure–activity
relationships associated with the benzotriazole series 1
with which activity was originally discovered indicated
that a pseudo-symmetrical silhouette for the two hetero-
cycle elements is compatible with antiviral activity.1 In
the benzotriazole series, those matched pairs of N-1
and N-2 isomers examined proved to be essentially
equipotent.1 However, BABIM (11) clearly lacks the
hydroxybutyl side chain found in BMS-433771 (3), an
element of the pharmacophore tolerant of structural var-
iation but of critical importance for potency within this
series of fusion inhibitors.1–4 Taken together, these data
suggest that the absence of a side chain in BABIM (11)
may be offset by the positive effect of the amidine moiety
which, presumably, can effectively engage Asp200.
The pattern of structure–activity relationships summa-
rized by the data presented in Table 1 are compatible
with the basic tenets of the model of the binding of
RSV fusion inhibitors to the pocket originally identified
by Zhao and colleagues as a potential target for small
molecule drugs.12 The generally poor activity associated
with the introduction of substituents at C-4 and C-7 of
the benzimidazole nucleus and the greater, but still lim-
ited, tolerance for substituents at C-6 are predicted quite
effectively by the proposed binding mode.11 Most grati-
fyingly based on the original premise, the introduction
of basic functionality at C-5 provided compounds
retaining full activity towards wild-type virus that also
demonstrated potent antiviral activity towards virus
developed to be resistant to representatives of this class
of fusion inhibitors. Virus expressing the K394R muta-
tion is potently inhibited by the benzylamine 10aa with
an EC50 comparable to that observed towards wild-type
virus. A model of 10aa bound in the proposed binding
site and engaging Asp200 is depicted in Figure 2, where
the vector for projecting the basic amine or amidine
moiety is optimal. However, modest alterations in the
binding pose may allow access to Asp200 from C-4 or
C-6, providing a potential explanation for the significant
antiviral activity retained by the aminomethyl deriva-
tives 10a and 10ar, respectively.
N
H2N
HN
NH2
NH
N
NH
HN
11 (BABIM)
In summary, the structure–activity relationships sum-
marized in Table 1 are consistent with the binding pose
that was proposed based on an analysis of photo-affinity
labelling experiments. This model suggests limited toler-
ance for substitution at C-4 and C-7 of the benzimid-
azole moiety and the potential to establish a salt
bridge interaction with Asp200 in the pocket created
by assembly of the N-terminal heptad repeat into a tri-
mer. The installation of basic moieties at C-5, particu-
larly aminomethyl and amidine, resulted in potent
RSV inhibitors that fully retain activity towards virus
demonstrating over 1000-fold resistance to molecules
devoid of substitution.
The important effect of a basic functional group at C-5
of these RSV fusion inhibitors provides an interesting
convergence of the silhouette, particularly for the ami-
dine derivatives 10ai–h, with bis(5-amidino-2-benzimi-
dazolyl)methane (BABIM, 11), a compound with
broad spectrum, trypsin-like serine protease inhibition
that has also been reported to be an inhibitor of RSV fu-
sion.27–31 In our hands, BABIM (11) displays somewhat
modest potency with an EC50 of 100 nM towards the
Long A strain of virus. More importantly, serial passage
of RSV in the presence of BABIM (11) selected for resis-
tant virus incorporating a F140I mutation in the F1 pro-

4598
X. A. Wang et al. / Bioorg. Med. Chem. Lett. 17 (2007) 4592–4598
Colonno, R.; Meanwell, N. A.; Krystal, M. Proc. Natl.
Acad. Sci. U.S.A. 2004, 101, 15046.
References and notes
12. Zhao, X.; Singh, M.; Malashkevich, V. N.; Kim, P. S.
Proc. Natl. Acad. Sci. U.S.A. 2000, 97, 14172.
13. Wang, X. A.; Yu, K.-L.; Kadow, K. F.; Cianci, C. W.;
Roach, J.; Langley, D. R; Krystal, M; Meanwell, N.
Abstracts of Papers, The 230th ACS National Meeting,
Washington, DC, Aug 28–Sept 1, 2005.
1. For Part 1 of this series, see: Yu, K.-L.; Zhang, Y.;
Civiello, R. L.; Kadow, K. F.; Cianci, C.; Krystal, M.;
Meanwell, N. A. Bioorg. Med. Chem. Lett. 2003, 13, 2141.
2. For Part 2 of this series, see: Yu, K.-L.; Zhang, Y.;
Civiello, R. L.; Trehan, A. K.; Pearce, B. C.; Yin, Z.;
Combrink, K. D.; Gulgeze, H. B.; Wang, X. A.; Kadow,
K. F.; Cianci, C.; Krystal, M.; Meanwell, N. A. Bioorg.
Med. Chem. Lett. 2004, 14, 1133.
14. Vernin, G.; Domlog, H.; Siv, C.; Metzger, J.; El-Shafei, A.
K. J. Het. Chem. 1981, 18, 85.
15. Almeida, P. S.; Lobo, A. M.; Prabhakar, S. Heterocycles
1989, 28, 653.
16. Meanwell, N. A.; Sit, S.-Y.; Gao, J.; Wong, H. S.; Gao,
Q.; Laurent, D. R. St.; Balasubramanian, N. J. Org.
Chem. 1995, 60, 1565.
17. Bell, I. M.; Gallicchio, S. N.; Theberge, C. R.; Zhang, X.-F.;
Stump, C.; Zartman, C. B. PCT Int. Appl., 2004, WO-2004/
082605.
18. Yu, K.-L.; Wang, X.; Sun, Y.; Cianci, C.; Thuring, J. W.;
Combrink, K.; Meanwell, N.; Zhang, Y.; Civiello, R. L.
PCT Int. Appl., 2003, WO-2003/053344.
3. For Part 3 of this series, see: Yu, K.-L.; Wang, X. A.;
Civiello, R. L.; Trehan, A. K.; Pearce, B. C.; Yin, Z.;
Combrink, K. D.; Gulgeze, H. B.; Zhang, Y.; Kadow, K.
F.; Cianci, C. W.; Clark, J.; Genovesi, E. V.; Medina, I.;
Lamb, L.; Wyde, P. R.; Krystal, M.; Meanwell, N. A.
Bioorg. Med. Chem. Lett. 2006, 16, 1115.
4. For Part 4 of this series, see: Yu, K.-L.; Sin, N.; Civiello,
R. L.; Wang, X. A.; Combrink, K. D.; Gulgeze, H. B.;
Venables, B. L.; Wright, J. K.; Dalterio, R. A.; Zadjura,
L.; Marino, A.; Dando, S.; D’Arienzo, C.; Kadow, K. F.;
Cianci, C. W.; Clark, J.; Genovesi, E. V.; Medina, I.;
Lamb, L.; Colonno, R. J.; Krystal, M.; Meanwell, N. A.
Bioorg. Med. Chem. Lett. 2007, 17, 895.
19. (a) Davoll, J. J. Chem. Soc. 1960, 308; (b) Davoll, J.;
Laney, D. H. J. Chem. Soc. 1960, 314.
20. Stille, J. K. Angew. Chem. Int. Ed. Engl. 1986, 25, 508.
21. (a) Wolfe, J. P.; Buchwald, S. L. Tetrahedron Lett. 1997,
38, 6367; (b) Wolfe, J. P.; Wagaw, S.; Marcoux, J.-F.;
Buchwald, S. L. Acc. Chem. Res. 1998, 31, 805; (c)
Hartwig, J. F. Angew. Chem., Int. Ed. Engl. 1998, 37, 2046.
22. Ciganek, E. J. Org. Chem. 1992, 57, 4521.
23. Herr, R. J. Bioorg. Med. Chem. 2002, 10, 3379.
24. Gillaspy, M. L.; Lefker, B. A.; Hada, W. A.; Hoover, D. J.
Tetrahedron Lett. 1995, 36, 7399.
25. Jung, M. E.; Piizzi, G. Chem. Rev. 2005, 105, 1735.
26. Charton, M. J. Org. Chem. 1965, 30, 969.
27. Tidwell, R. R.; Geratz, J. D.; Dubovi, E. J. J. Med. Chem.
1983, 26, 294.
28. Dubovi, E. J.; Geratz, J. D.; Tidwell, R. R. Virology 1980,
103, 502.
29. Dubovi, E. J.; Geratz, J. D.; Shaver, S. R.; Tidwell, R. R.
Antimicrob. Agents Chemother. 1981, 19, 649.
30. Dubovi, E. J.; Geratz, J. D.; Tidwell, R. R. Infect. Immun.
1983, 40, 351.
5. Thompson, W. W.; Shay, D. K.; Weintraub, E.; Brammer,
L.; Cox, N.; Anderson, L. J.; Fukuda, K. JAMA 2003, 289,
179.
6. Falsey, A. R.; Hennessey, P. A.; Formica, M. A.; Cox, C.;
Walsh, E. E. N. Engl. J. Med. 2005, 352, 1749.
7. Gilbert, B. E.; Knight, V. Antimicrob. Agents Chemother.
1986, 30, 201.
8. Ventre, K.; Randolph, A. G. Cochrane Database Syst.
Rev. 2007. doi:DOI: 10.1002/14651858.CD000181.pub3
(www3.interscience.wiley.com/cgi-bin/mrwhome/106568753/
HOME; http://www.thecochranelibrary.com).
9. Cianci, C.; Yu, K.-L.; Combrink, K.; Sin, N.; Pearce, B.;
Wang, A.; Civiello, R.; Voss, S.; Luo, G.; Kadow, K.;
Genovesi, E. V.; Venables, B.; Gulgeze, H.; Trehan, A.;
James, J.; Lamb, L.; Medina, I.; Roach, J.; Yang, . Z.;
Zadjura, L.; Colonno, R.; Meanwell, N.; Krystal, M.
Antimicrob. Agents Chemother. 2004, 48, 413.
10. Cianci, C.; Genovesi, E. V.; Lamb, L.; Medina, I.; Yang, Z.;
Zadjura, L.; Yang, H.; D’Arienzo, C.; Sin, N.; Yu, K.-L.;
Combrink, K.; Li, Z.; Colonno, R.; Meanwell, N.; Clark, J.;
Krystal, M. Antimicrob. Agents Chemother. 2004, 48, 2448.
11. Cianci, C.; Langley, D. R.; Dischino, D. D.; Sun, Y.; Yu,
K.-L.; Stanley, A.; Roach, J.; Li, Z.; Dalterio, R.;
31. Katz, B. A.; Clark, J. M.; Finer-Moore, J. S.; Jenkins, T.
E.; Johnson, C. R.; Ross, M. J.; Luong, C.; Moore, W. R.;
Stroud, R. M. Nature 1998, 391, 608.
32. Cianci, C.; Krystal, M. Unpublished results.