W. Worawalai et al. / Bioorg. Med. Chem. Lett. 24 (2014) 5530–5533
5531
OH
O
OH
the idea of N-linked dicyclitols was introduced by Hudlicky and co-
workers since 2002;11 however, all synthesized targets (Fig. 2)
showed weak inhibition (IC50 370–2,000 M). In the recent study,
we hope that the replacement of the above cyclitols by (+)-proto-
quercitol would promote inhibitory effect more potent than other
N-linked dicyclitols previously reported.
OH
OH
a
O
O
O
O
e
1
5
3
l
75%
HO
54%
O
O
OH
O
O
1
4
2
The starting material (+)-proto-quercitol (1) utilized in this
investigation was obtained from the stems of Arfeuillea arborescens
using the procedure described elsewhere.6a Crucial to the success-
ful synthesis of amine-linked diquercitols (7 or 8) was the use of
reductive amination of aminoquercitols 3 or 6 and ketoquercitol
4 (Scheme 1), in which high diastereoselectivity was observed in
each step.
b, 78%; c, 79%; d, 71%
f, 93%
OH
NH2
O
O
O
O
O
O
O
O
5
All key coupling intermediates 3, 4 and 6 were synthesized from
(+)-proto-quercitol (1) (Scheme 2), using our previous methodol-
ogy.6a Initially, hydroxyl groups in 1 were protected by reaction
with Me2C(COMe)2 in the presence of p-TsOH as a catalyst, yielding
exclusively bisacetonide 2 in 75% yield. This observation could be
rationalized by initially favored formation of cis-acetonide at C-2
and C-3 followed by inevitable trans-acetonide formation at C-4
and C-5, resulting in the free hydroxyl group at C-1 for further
functionalization. The remaining hydroxyl group in 2 was con-
verted to a better leaving group (a mesyl moiety), which was fur-
ther substituted by azide group (NaN3). The desired
aminoquercitol 3 was eventually obtained by hydride reduction
using LiAlH4. Noticeably, transformation of 2–3 proceeded stereo-
specifically with inversion of configuration.
3
b, 82%; c, 84%; d, 43%
NH2
O
O
O
O
6
Scheme 2. Reagents and conditions: (a) Me2C(OMe)2, DMF, p-TsOH; (b) MeSO2Cl,
Et3N, DMAP; (c) NaN3, DMF, 15-crown-5-ether, 100 °C; (d) LiAlH4; (e) Ac2O, DMSO;
(f) LiAlH4.
Having aminoquercitol 3 in hand, we turned our plan to gener-
ate another coupling motif, ketoquercitol 4. Starting from 2, the
target ketone 4 was synthesized in moderate yield (54%) by Alb-
right-Goldman oxidation12 using Ac2O-DMSO. At this step, we also
have an idea to produce another aminoquercitol that is epimeric to
3, which would be coupled to ketone 4 to produce diastereomeric
N-linked diquercitol. Ketone 4 was first reduced by LiAlH4, afford-
OH
OH
OH
OH
H
N
1) NaBH3CN, AcOH,
MeOH, 31%
HO
HO
OH
OH
3
4
2) HCl/MeOH, Dowex
50W-X8 (H+), 70%
7
OH
OH
OH
1) NaBH3CN, AcOH,
MeOH, 21%
OH
H
HO
HO
N
OH
OH
6
4
RN
OH
2) HCl/MeOH, Dowex
50W-X8 (H+), 35%
HO
HO
OH
OH
8
OH
R
IC50 ( M)
µ
Scheme 3.
amine-linked diconduritol
H
1,400
1,350
acetamide-linked diconduritol Ac
ing the hydroxybisacetonide 5 with excellent yield (93%) as sole
product. The desired aminoquercitol 6 was thus generated using
the similar conditions applied for 3.
HO
HO
RN
OH
OH
Having aminoquercitols 3 and 6 together with ketone 4 in hand,
the required N-linked diquercitols were synthesized by reductive
amination13 using NaBH3CN14 (Scheme 3). Coupling of 3 and 4 fol-
lowed by deprotection under acid condition afforded amine-linked
diquercitol 7 while compound 8 was generated from 6 and 4.
Although amine-linked diquercitols 7 and 8 were obtained as sin-
gle product in each synthetic route, the configuration of newly gen-
erated chiral center (C-1) remained unclear.
R'
HO
HO
OH
OH
OH
R
R'
IC50 (µM)
acetamide-linked inositol
amine-linked inositol
diaminoinositol dimer
Ac OH
H
H
370
>2,000
>2,000
OH
NH2
The severely overlapped 1H NMR signals, particularly in diag-
nostic region (d 3.0–4.0 ppm), made impossible determination of
configuration. We therefore turned our attempt to inspect 1H
NMR data of the bisacetonides (7a and 8a) of 7 and 8 because they
displayed well-separated spectra. We have demonstrated, in our
previous report,6a that 1H NMR pattern and coupling constants of
methylene protons in 3 and 6 are distinct enough to apply for
predicting configuration of amino-connected chiral carbon as
Figure 2.
O
OH
O
OH
1'
OH
H
N
O
O
O
OH
OH
O
O
NH2
HO 3'
HO
1
OH
OH
*
3
*
+
HO
5'
5
O
OH
O
OH
OH
a
- or b-oriented. Apparently, bisacetonide 7a revealed a single set
7
8
3
6
4
1
or
or
of methylene protons (H2-6/and H2-60) identical to those of 3 whereas
bisacetonide 8a contained two different patterns of 3 and 6 (Fig. 3).
Scheme 1.