M. Zhong et al. / Bioorg. Med. Chem. Lett. xxx (2014) xxx–xxx
3
Scheme 3. Reagents and conditions: (a) LDA, 1,3-dimethyl-3,4,5,6-tetrahydro-2(1H)-pyrimidinone (DMPU), THF, ꢀ78 °C, 40 min then 1-bromo-2-chloroethane, ꢀ78 °C to rt,
1 h; (b) LDA, ꢀ78 °C to rt, overnight, 52% yield two steps from 23; (c) KOH, MeOH/H2O, 90 °C, overnight, 95% yield; (d) HgO, CH2Br2, Br2, 80 °C, 95% yield; (e) AlCl3, benzene,
DCM, ꢀ20 to 0 °C, overnight, 98% yield; (f) AcCl, AlCl3, DCM, ꢀ20 °C to rt, overnight, 95% yield; (g) (COCl)2, cat. DMF, DCM, rt; (h) CH2N2 in ether, 0 °C to rt, overnight, 82% yield
three steps from 28; (i) HBr (33% in AcOH), AcOH, rt, 2 h, 88% yield; (j) N-Boc- or N-Cbz-L-Pro-OH, MeCN, DIPEA, rt; (k) TMSOTf, DIPEA, DCM, 0 °C, then
phenyltrimethylammonium tribromide (PTAT), 0 °C to rt; (l) NH4OAc, DIPEA, o-xylene, reflux; (m) 4.0 N HCl in dioxane, rt; (n) 10% Pd(OH)2 on carbon, EtOH, AcOH, H2
(1 atm); (o) functionalized amino acid(s), HATU, DIPEA, DMF, rt.
acetyl chloride, cyclohexene and benzene in the presence of AlCl3
gave a mixture of cis- and trans-18,18 from which the trans- isomer
was readily separated by column chromatography. Further acyla-
of the gt-1a potency with an EC50 of 3.8 nM. It is worth noting that
the N-Moc- -Val analog 2d (entry 6) has no comparable potency
relative to its N-Moc- -Val counterpart 2c. Also, substitution at
D
L
tion of trans-18, followed by
a
-bromination and ester formation
the C4-position of the piperidyl residue as shown in 3 (entry 7)
is not tolerated with respect to the gt-1a potency. Substituting
the piperidyl residue in 2c by a 1,4-substituted cyclohexyl moiety
afforded a more active analog 4b (entry 9) with an EC50 of 2.0 nM
in the gt-1a replicon. Similar to the structure and activity relation-
with N-Cbz- or N-Boc- -proline afforded 20a–d. Cyclization of
L
20a–d with NH4OAc generated 21a–d. Subsequent Cbz- and/or
Boc-deprotection(s) and amide formation(s) gave 4a–k.
Scheme 3 shows a general approach to functionalized bicy-
clo[2,2,2]octylphenyl bisimidazoles. A two-step alkylation protocol
of commercially available dimethyl cyclohexane-1,4-dicarboxylate
(23) gave dimethyl bicyclo[2,2,2]octane-1,4-dicarboxylate (24).19
Selective saponification of 24, followed by HgO mediated decar-
boxylation and bromination afforded 26. Next, an AlCl3 catalyzed
Friedel–Crafts reaction of 26 with benzene produced 27, which
was subsequently followed by another Friedel–Crafts reaction with
AcCl to form 28. The methyl ester in 28 was readily converted to
ship (SAR) observed in piperidylphenyl series, N-Moc-
4a (entry 8) is less favorable than the corresponding N-Moc-
D
-Phg analog
-Val
L
analog 4b. Not surprisingly, substitution at C4-position of the
cyclohexyl residue in 4b (5, entry 10) is also detrimental to the
potency. Furthermore, replacement of one of the phenyl residues
in daclatasvir with linear and rigid bicyclo[2,2,2]octyl led to 6
(entry 11). In terms of intrinsic solubility of these new compounds,
2c, 4b, and
6 exhibited significant improvement relative to
a
-bromoacetyl through a typical four-step sequence. Esterification
of 30 with N-Boc or -Cbz- -proline gave mono-ester 31a or 31b,
which was -brominated, followed by esterification with N-Boc
or N-Cbz- -proline to give 33a, –b, –c, or –d. Subsequently, imidaz-
daclatasvir at pH 7.4 (entries 5, 9 and 11 vs 1).
L
To understand the SAR on the regions of W and W0 (Table 2), a
set of compounds with pyrrolidine replacements was synthesized
and evaluated. It turned out that either ring expansion (2e–f,
entries 3 and 4) or disconnection (2g–i, entries 5–7) reduces the
gt-1a potency, while such structural modifications are less sensi-
tive to the gt-1b potency. Moreover, W and W0 residues could be
exchanged without significantly decreasing potency (2f vs 2k,
entries 4 and 9; and 2h vs 2j, entries 6 and 8), which is aligned with
the dimeric feature of the HCV NS5A protein.20 In general, pyrrol-
idyl residue as W or W0 seems to be optimal, providing higher
potency relative to the modifications made in Table 2.
a
L
ole formation of 33a–d in the presence of NH4OAc afforded 34a–d.
These key intermediates were readily converted to 6 or its analogs
through typical sequence(s) of deprotection(s) and amide forma-
tion(s) with substituted amino acid(s).
Both HCV genotype (gt) 1a and 1b replicons were used as the
primary cell-based functional assays to evaluate the potency of
all of the bisimidazoles. As depicted in Table 1, replacement of
the biphenyl residue with a mono-fluoro substituted piperidylphe-
nyl motif gave 2a (entry 3), which has an EC50 of 38 nM in the gt-1a
replicon and an EC50 of 0.095 nM in the gt-1b replicon. The potency
was enhanced by the introduction of a second fluoro atom onto the
central phenyl residue of 2a (entry 4). It was observed that, in the
biphenyl series, the N-Moc-
was about twice as potent as the corresponding N-Moc-
log, Daclatasvir (entry 1) in both the gt-1a and gt-1b replicon. Inter-
estingly, when N-Moc- -Phg in 2b was replaced with N-Moc- -Val,
the resulting compound 2c (entry 5) showed a 3-fold improvement
In parallel, a set of analogs (4c–k, Table 3) with a variety of Z
and Z0 was prepared to establish the SAR of these two regions.
When N-Moc-L-Val was replaced with other capping groups (4c–f,
entries 3–6) derived from natural amino acids bearing various alkyl
side chains, no significant shift (<3 folds) of the gt-1a potency was
observed; however, the gt-1b potency decreased by up to 17-fold
(4d vs 4b, entries 4 and 2). Interestingly, more soluble N,N0-
D-Phg substituted analog 1b (entry 2)
L
-Val ana-
D
L
dimethyl-
D-Phg can serve as a replacement of N-Moc-D-Phg (4g
vs 4j, entries 7 and 10; and 4h vs 4a, entries 8 and 1) without a