Organic Letters
Letter
2
8
−
76.8 to −77.0 ppm) and negative-ion ESI-MS (m/z: 113,
and particularly those with the D-DMDP configuration
((+)-broussonetine E, IC50 = 0.0020 μM). While not tested
2
2
49) spectra of all final products, and as such their isolated
yields and all measurements dependent thereof were calculated
as TFA salts.
against this specific enzyme, natural glyphaeaside C displayed
only mild inhibition of Aspergiullus oryzae β-galactosidase (82%
at 1000 μM), providing further evidence of its L-DMDP-
configuration. Regarding almond β-glucosidase inhibition, 2b
was also the strongest inhibitor, though it was 5.1× less potent
than natural glyphaeaside C (IC50 = 0.15 μM). The difference
between the synthesized compounds and the targeted natural
product is further exemplified by their inhibition of snail β-
mannosidase, specifically that the synthesized compounds were
weaker by approximately an order of magnitude.
The H and 13C NMR spectroscopic data of 2a and 2b in
1
14
CD OD matched perfectly with those reported for natural
3
glyphaeaside C in the same solvent (Δδ = 0.00−0.01 ppm
H
and Δδ = 0.0−0.1 ppm). While the diastereoisomers 2a and
C
2
b were identical by RP-HPLC and NMR analysis in CD OD,
3
they were identified as different diastereomers, each with a
high diastereomeric purity (>98%), from their NMR analysis
in C D N (see Figures S78 and S79). Thus, a re-examination
5
5
of NMR profile of the natural product using C D N may be
In conclusion, a comparison of the NMR spectroscopic data
of glyphaeaside C with similar reported compounds led to the
determination that the originally proposed structure 1 was
incorrect and more likely a stereoisomer of general structure 2.
As part of a total synthesis investigation, the 5-C-hydroxyalkyl
moiety of 2 was accessed via the ring-opening of epoxide 10a, a
novel synthesis route that should find future useful applications
in azasugar synthesis. The four 8′,9′-diol diastereomers of 2
were prepared utilizing the complementary ADH reactions of
(E/Z)-5 and the eventual separation by semipreparative HPLC
after global deprotection. The biological activity profiles of the
prepared compounds are relatively conserved with respect to
their side-chain configurations and consistent with similar D-
DMDP-derived alkaloids. However, they are notably different
from the natural product, which is now believed to possess the
L-DMDP configuration. A comparison of specific rotation data
suggests that natural glyphaeaside C is the enantiomer of 2a,
although further studies, such as a reisolation of the natural
product, are needed to unambiguously confirm this assertion.
5
5
29
necessary to determine its absolute 8′,9′-diol configuration.
All four final products displayed positive specific rotations,
the first indication that the natural product did, in fact, have a
L-arabinose configuration. Although none of the specific
rotations perfectly matched the magnitude of the natural
2
0
25
product ([α] −0.8 (c 1.18, MeOH)), that of 2a ([α] +3.0
D
D
(
c 0.46, MeOH)) was much closer in magnitude than that of
b {[α] +19.9 (c 0.42, MeOH)}, which would posit that the
2
5
2
D
structure of natural glyphaeaside C is ent-2a. While some
2
naturally occurring L-arabino-iminosugars have been observed,
this revised structure would represent the first example of an L-
DMDP derivative to be found in nature and is certainly
opposite to that found in the broussonetines (Figure 1, for
example).
The four final products were tested against a panel of
glycosidases to ascertain their corresponding inhibitory effects
(Table 2). In general, glycosidase inhibition was independent
Table 2. Concentrations of Glyphaeaside C Stereoisomers
Giving a 50% Inhibition of the Selected Glycosidases
a
ASSOCIATED CONTENT
sı Supporting Information
■
*
IC50 (μM)
1
enzyme
natural
2a
2b
2a′
2b′
α-glucosidase
rice
β-glucosidase
almond
bovine liver
human lysosome
β-galactosidase
bovine liver
A. oryzae
b
c
c
c
c
13%
21.3%
5.7%
3.0%
4.2%
FAIR data, including the primary NMR FID files, for
compounds 2a, 2a′, 2b, 2b′, (E/Z)-5, 6a, 6b, 7, 9, 10a,
0.15
0.92
0.037
33
0.77
0.019
195
0.99
0.060
36
0.82
0.038
55
d
-
d
1
0b, 11a, 11b, 13a, 13a′, 13b, and 13b′ (ZIP)
-
d
-
0.031
0.020
d
0.043
d
0.019
d
Corresponding Authors
■
b
d
82%
-
-
-
-
β-mannosidase
snail
Stephen G. Pyne − School of Chemistry and Molecular
4.5
48
30
56
35
a
b
See the Supporting Information for complete assay data. Inhibition
percet at 1000 μM. Inhibition percent at 100 μM. Not tested.
c
d
Brendan J. Byatt − School of Chemistry and Molecular
of the side-chain configuration; in cases where the IC50 was
determined (compounds that showed >50% inhibition at a 100
μM inhibitor concentration), the enzyme that experienced the
greatest variation of inhibition was human lysosomal β-
glucosidase, where the IC50 values of 2a (33 μM) and 2b
Author
(
195 μM) differed by a factor of 5.9. Concerning mammalian
β-glucosidase activity, the compounds showed a far greater
inhibitory activity toward bovine liver, where 2b was the
strongest inhibitor (IC = 0.019 μM) and showed an increase
50
in potency of four orders of magnitude compared to that of
human lysosomal. All compounds displayed the nanomolar
inhibition of bovine liver β-galactosidase (IC50 = 0.019−0.043
μM), a trait shared with many of the broussonetine alkaloids
Notes
The authors declare no competing financial interest.
D
Org. Lett. XXXX, XXX, XXX−XXX