Ring-closing double reductive amination route to aza-heteroannulated sugars

Article abstract of DOI:10.1016/j.tetlet.2004.11.032

1,4-Dicarbonyl derivatives of glycosides are produced by ozonolysis or Wacker oxidation. A stable ozonide is isolated and a carbonyl group reduced whilst maintaining the ozonide functionality. The 1,4-dicarbonyl compounds are converted to various N-substi

Full text of DOI:10.1016/j.tetlet.2004.11.032

Tetrahedron
Letters
Tetrahedron Letters 46 (2005) 307–310
Ring-closing double reductive amination route to
aza-heteroannulated sugars
Dominic M. Laventine,a Michelle Davies,a Emma L. Evinson,b Paul R. Jenkins,a,*
Paul M. Cullisa and John Fawcetta
aDepartment of Chemistry, University of Leicester, Leicester UK LE1 7RH, UK
bMolecularNature Ltd, Plas Gogerddan, Aberystwyth, Ceredigion SY23 3EB, UK
Received 8 October 2004; revised 3 November 2004; accepted 8 November 2004
Available online 25 November 2004
Abstract—1,4-Dicarbonyl derivatives of glycosides are produced by ozonolysis or Wacker oxidation. A stable ozonide is isolated
and a carbonyl group reduced whilst maintaining the ozonide functionality. The 1,4-dicarbonyl compounds are converted to various
N-substituted pyrrolidines by diastereoselective double reductive amination The resulting aza-heteroannulated sugars no significant
inhibition of any glycosidase, with the exception of compound 12g, which is a weak inhibitor of b-galactosidase.
Ó 2004 Elsevier Ltd. All rights reserved.
Polyhydroxylated aza-heterocycles and aminosugars are
of interest because they evince a wide range of biological
activity.1–3 Particular attention has been paid to their
ability to inhibit glycoside-processing enzymes, with
resulting potential therapeutic applications in treatment
of cancer, HIV and diabetes.4,5
HO
HO
HO
HO
+
NH
HO
NH
HO
HO
OH
OH
HO
2
3
The mechanism of glycosidases has been extensively
studied, and is now thought to proceed via protonation
of the exo-cyclic oxygen, which after bond cleavage
gives rise to an oxo-carbonium ion intermediate such
as 1.1
Other potent inhibitors, such as castanospermine (4) and
(2R,5R,3R,4R)-2,5-bis(hydroxymethyl)-3,4-dihydroxy-
pyrrolidine (5), lack an hydroxyl group adjacent to the
endo-cyclic heteroatom, and therefore cannot form
an iminium ion analogous to 3.6 In these compounds
the presence of a nitrogen that can be protonated is
sufficient to give high glycosidase inhibitory activity.1
Mannostatin A (6) is a potent inhibitor without
obvious resemblance to the proposed transition state.
This illustrates the importance of a basic amino group
being present.7a Acetylation of this amino group has
been found to destroy inhibitory activity.7b
HO
HO
HO
HO
+
O
O
HO
HO
+
HO
HO
1
Design of inhibitors has subsequently focused upon
mimicking the shape and charge of the transition state
that leads to intermediate 1. Potent naturally-occurring
inhibitors such as nojirimycin (2) are thought to work
via formation of an iminium ion (3), which is an aza-
analogue of oxonium ion 1.
OH
OH
OH
OH
H
N
HO
HO
HO
HO
HO
OH
NH
2
HO
N
Keywords: Pyrrolidines; Carbohydrate annulation; Amino sugars;
Stable ozonide decomposition; Glycosidase inhibitors.
H
SCH
3
4
*
5
6
Corresponding author. Tel.: +44 0116 252 2124; e-mail: kin@le.ac.uk
0040-4039/$ - see front matter Ó 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2004.11.032

308
D. M. Laventine et al. / Tetrahedron Letters 46 (2005) 307–310
The synthesis of such compounds has therefore
attracted considerable interest, with routes to enantio-
pure materials of special importance.
O
OMe
O
O
OMe
O
O
O
i
Ph
O
Ph
O
O
O
Previous work within this group has focused on the use
of carbohydrates as scaffolds for stereoselective con-
struction of various cyclic compounds.8,9 Extension of
this approach to include aza-heterocycles was expected
to provide a facile route to substituted pyrrolidines. Pro-
duction of the desired compounds by double reductive
amination of a 1,4-dicarbonyl precursor was envisaged.
Several examples of highly stereoselective double reduc-
tive amination ring-closures indicated the broad applica-
bility of the intended approach.10,11 The potential to
create a diverse range of derivatives by altering the sub-
stituent on the nitrogen of the resulting pyrrolidine was
also attractive, as it allows for possible optimisation of
any biological activity, which is found.
7
8
O
v
ii
O
OMe
O
OMe
O
O
O
Ph
O
O
O
Ph
O
9
10
O
iii
iv
iii , vi
The transformation of the known alkene 78 to a 1,4
dicarbonyl compound can be achieved by various meth-
ods in which oxidation of the alkene bond is a common
feature. The simplest route was expected to be the ozon-
olysis of the double bond.
O
OMe
O
OMe
H
1
5
6
14
O
HO
H
2
8P-1h3
O
HO
3
R1
R1
4
7
N
N
H
H
H
16
15α
R2
R2
H
15β
R3
R3
12a-g
11a-e,g
Formation of ozonide 8 in 85% yield was achieved by
bubbling O3 through a solution of alkene 7 in DCM
at À78°C.12,13 Decomposition of ozonide 8 was initially
attempted by treatment with excess DMS at rt over-
night.12 Removal of the DMS and DCM resulted in a
clear syrup that was identified as a mixture of two dia-
Scheme 1. Reagents and conditions: (i) O3, DCM, rt, 1h, 85%; (ii)
polymer-bound PPh3, PhMe, 90°C, 1h, 62%; (iii) R1NH2, AcOH,
THF, then NaCNBH3, rt, 2h (Table 1); (iv) 80% AcOH aq, reflux, 4h
or 80% AcOH in EtOH, reflux, 30h (Table 1); (v) PdCl2, CuCl2, O2, 1:1
DMF–H2O, rt, 4h, 85%; (vi) chromatographic separation.
1
stereoisomers of ozonide 8 by H NMR (Scheme 1).13
Further attempts at reduction of the ozonide were made
using thiourea, which also failed. Reduction using pow-
dered zinc in acetic acid, which has been reported in the
literature as reducing ozonides impervious to reduction
with DMS14 gave a complex mixture of products that
O
OMe
O
O
OMe
OH
O
O
i
Ph
O
Ph
O
O
O
O
O
13
1
8
did not have the Ph or OMe H NMR signals of the
O
O
starting material, indicating loss of the protecting
groups. Direct conversion of the ozonide to an amine
was also attempted without success.15 Reaction of ozon-
ide 8 with NaCNBH3 in 2:1 THF–DCM at rt resulted in
the reduction of the ketone functional group without
decomposition of the ozonide moiety to afford ozonide
13 (Scheme 2); indicating an unusually stable ozonide.16
Scheme 2. Reagents and conditions: (i) NaCNBH3, THF–DCM, rt,
1h, 83%.
addition of NaCNBH3 (Scheme 1). We were encouraged
by the initial success with aliphatic amines, in which
only one diastereoisomer was formed, in moderate
yields (Table 1).21 The configuration and stereochemis-
try of the ring junction was confirmed by single-crystal
Decomposition of ozonide 8 was eventually achieved by
treatment with 1equiv of PPh3 in DCM at rt to afford
the dicarbonyl compound 9 in 62% yield.17,18 Difficulties
in removal of the Ph3PO from 9 were overcome by use
of a polymer-bound reagent (Scheme 1).19
1
X-ray crystallography (Fig. 1)22 and H NMR NOESY
experiments. For example, NOE interactions were ob-
served between H-3 and H-5 for 11d, and between H-
4, H-15b and H-16b and H-2, H-15a and H-16a for 12a.
An alternative approach for the conversion of ketone 7
to a 1,4-dicarbonyl compound makes use of the Wacker
oxidation. This involves the oxidation of a terminal
alkene to a methyl ketone using PdCl2.20 The desired
diketone 10 was produced according to the literature
procedure in 85% yield as a colourless syrup and identi-
Table 1. Structures and yields of compounds 11a–e, g and 12a–g
R1
R2
R3
% Yield % Yield
of 11
of 12
11a, 12a Pri
11b, 12b Pri
H
H
H
49
29
78
83
67
94
65
80
77
Me
H
1
13
Pri
Me 15
fied by H and C NMR.8
11c, 12c
11d, 12d CH2CH2CH2OH
H
H
H
H
H
45
42
—
57
11e, 12e CH2CO2Et
12f CH2CO2H
11g, 12g OH
H
Ring closure of the dicarbonyl compound 9 was
achieved by treatment with an excess of amine with a
catalytic amount of AcOH in THF at rt, followed by
H
H

D. M. Laventine et al. / Tetrahedron Letters 46 (2005) 307–310
309
for both 11b and 11c. However interactions between
H-3, H-15a and H-16 and between H-15b and the methyl
attached at C-16 were found for 11b and not 11c,
whereas for 11c interactions between H-4, H-15b and
H-16, and between H-15a and the methyl attached at
C-16 were seen.
Removal of the benzylidene protecting group was
achieved in good yields by heating in 80% aq AcOH
for 4h (Table 1).23 Pyrrolidine 11e was also heated to re-
flux for 30h in 80% ethanolic AcOH to afford deprotec-
tion of the diol whilst maintaining the ester functionality
of 12e (Scheme 1).
Figure 1. X-ray crystal structure of 12g. One of the unique molecules
of the asymmetric unit with atom label scheme and 50% displacement
ellipsoids. The unique molecules have different orientations of the
phenyl group. H atoms except those on C2 and C5 are omitted for
clarity.22
Measurement of the inhibitory activity of compounds
12a–g against various glycoside processing enzymes
was undertaken. The results indicate that these com-
pounds are not significant inhibitors of a range of glyco-
sidases. Compound 12g was found to be a weak
inhibitor of b-D-galactosidase (Table 2).
The stereochemistry at the C-2 ring junction is thought
to arise due to an initial reaction at C-16, followed by
ring closure, resulting in the less strained cis ring junc-
tion. Nonring-closing reductive amination at C-2 of
analogous compounds was found to give the opposite
stereochemistry (i.e., a-amines) due to the preference
of the hydride to attack from the opposite direction to
the C-1 substituent.
In conclusion, we have developed a versatile and effi-
cient method for the construction of pyrrolidine rings
on carbohydrates. Work is in progress to synthesise fur-
ther analogues of 12, and to alter the regio- and stereo-
chemistry of the ring junction. Further evaluation of the
glycosidase inhibitory activity of these compounds will
also be undertaken.
The applicability of this reaction to a number of amines
containing functional groups was studied, including aro-
matic rings, alcohols and esters (Table 1). In all cases
the reaction was stereoselective, with yields in the range
40–55% after purification by column chromatography.
Acknowledgements
We would like to acknowledge the ESPRC and the Uni-
versity of Leicester for financial support of this research.
Diketone 10 was found to react under the same condi-
tions. It was hoped that high diastereoselectivity would
be observed due to the preference for a cis ring junction
and for reduction at C-16 from the less hindered face.
However, it was found that diastereomers 11b and 11c
were produced in a 2:1 ratio when diketone 10 was re-
acted with iso-propylamine (Table 1). Separation by
flash chromatography allowed assignment of stereo-
References and notes
1. Lillelund, V. H.; Jensen, H. H.; Liang, X.; Bols, M. Chem.
Rev. 2002, 102, 515.
2. Heightman, T. D.; Vasella, A. T. Angew. Chem., Int. Ed.
1999, 38, 750.
1
chemistry by H NMR and NOESY experiments, con-
firming that the products differed in configuration at
the centre bearing the methyl group. For example,
NOE interactions were observed between H-3 and H-5
3. Popowycz, F.; Gerber-Lemaire, S.; Demange, R.; Rodri-
guez-Garcia, E.; Asenjo, A. T. C.; Robina, I.; Vogel, P.
Bioorg. Med. Chem. Lett. 2001, 11, 2489.
Table 2. % Inhibition of 12a–g (0.14mg/mL solutions) against various glycoside-processing enzymes
Enzyme
12a
12b
12c
NI
12d
12e
12f
NI
12g
a-D-Glucosidase (yeast)
a-D-Glucosidase (bacillus)
a-D-Glucosidase (rice)
b-D-Glucosidase
a-D-Galactosidase
b-D-Galactosidase
a-L-Fucosidase
a-D-Mannosidase
Naringinase
NI
NI
2.1
6.8
12.4
—
NI
NI
6.9
NI
2.9
5.9
NI
5.3
NI
5.7
15.4
NI
1.3
NI
NI
1.3
NI
NI
4.7
NI
—
NI
NI
NI
NI
NI
41.3
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
14.7
NI
2.6
NI
8.6
3.9
NI
4.9
9.1
NI
NI
NI
—
2.8
NI
NI
—
NI
NI
3.7
3.2
NI
3.3
NI
14.2
12.4
NI
7.9
NI
NI
NI
NI
NI
NI
7.7
10.1
NI
3.0
NI
NI
NI
N-Acetyl-b-D-glucosaminidase (bovine kidney)
N-Acetyl-b-D-glucosaminidase (jack bean)
N-Acetyl-b-D-hexosaminidase
Amyloglucosidase
7.3
NI
NI
NI
NI indicates no inhibition was observed.

310
D. M. Laventine et al. / Tetrahedron Letters 46 (2005) 307–310
4. Asano, N. Glycobiology 2003, 13, 93R.
5. Watson, A. A.; Fleet, G. W. J.; Asano, N.; Molyneux, R.
J.; Nash, R. J. Phytochemistry 2001, 56, 265.
6. Pearson, W. H.; Hembre, E. J. J. Org. Chem. 1996, 61,
5546.
7. (a) Winkler, D. A. J. Med. Chem. 1996, 39, 4332; (b)
Tropea, J. E.; Kaushal, G. P.; Pastuszak, I.; Mitchel, M.;
Aoyagi, T.; Molyneux, R. J.; Elbein, A. D. Biochemistry
1990, 29, 10062.
8. Wood, A. J.; Holt, D. J.; Dominguez, M.; Jenkins, P. R.
J. Org. Chem. 1998, 63, 8522.
(1H, dd, J = 10.5 5.0, H-6eq), 4.70 (1H, s, H-1), 5.52 (1H, s,
H-7), 7.35-7.52 (5H, m, Ph), 9.82 (1H, t, J = 1.0, H-16). dC
(75.5MHz, CDCl3) 38.0 (CH2, C-15), 46.6 (CH, C-3), 55.8
(CH3, C-14), 64.4 (CH, C-5), 68.9 (CH2, C-6), 80.2 (CH,
C-4), 100.6 (CH, C-1), 101.5 (CH, C-7), 126.1 (CH, o-Ph),
128.4 (CH, m-Ph), 129.3 (CH, p-Ph), 136.7 (C, Ph), 198.4
(C, C-2), 199.3 (CH, C-16). FAB HRMS calcd for
C16H19O6 (M+H)+ 307.11816, found 307.11814.
19. Ferraboschi, P.; Gambero, C.; Azadani, M. N.; Santani-
ello, E. Synth. Commun. 1986, 16, 667.
20. Mereyala, B.; Lingannagaru, S. R. Tetrahedron 1997, 53,
17501.
9. Holt, D. J.; Barker, W. D.; Jenkins, P. R. J. Org. Chem.
2000, 65, 482.
21. For 11a dH (300MHz, CDCl3) 0.86 (3H, d, J = 6.4, H-
18a), 1.05 (3H, d, J = 6.7, H-18b), 1.56 (1H, dddd,
J = 12.9, 10.1, 9.0, 6.7, H-15a), 1.84 (1H, ddd, J = 12.9,
7.4, 2.8, H-15b), 2.38 (1H, dt, J = 10.2, 6.1, H-3), 2.70 (1H,
td, J = 10.2, 3.0, H-16a), 2.90 (1H, d, J = 5.8, H-2), 2.95
(1H, obs. m, H-16b), 2.96 (1H, sept, J = 6.6, H-17), 3.31
(3H, s, H-14), 3.54 (1H, ddd, J = 9.9, 7.2, 1.3, H-4), 3.70
(1H, obs. m, H-5), 3.70 (1H, ddd, J = 17.3, 10.6, 1.5, H-
6ax) , 4.18 (1H, dd, J = 16.2, 10.8, H-6eq), 4.57 (1H, s, H-
1), 5.48 (1H, s, H-7), 7.22–7.32 (3H, m, m-Ph, p-Ph), 7.38-
7.45 (2H, m, o-Ph). dC (75.5MHz, CDCl3) 13.3 (CH3, C-
18a), 22.4 (CH3, C-18b), 26.1 (CH2, C-15), 39.9 (CH, C-3),
43.2 (CH, C-16), 49.1 (CH3, C-17), 54.9 (CH3, C-14), 62.5
(CH, C-5), 63.5 (CH, C-2), 69.3 (CH2, C-6), 77.8 (CH, C-
4), 98.9 (CH, C-1), 102.0 (CH, C-7), 126.2 (CH, o-Ph),
128.3 (CH, m-Ph), 128.9 (CH, p-Ph), 137.9 (C, Ph). FAB
HRMS calcd. for C19H27O4N (M+H)+ 334.20183, found
334.20184.
10. Baxter, E. W.; Reitz, A. B. J. Org. Chem. 1994, 59, 3175.
11. Cheng, X.; Kumaran, G.; Mootoo, D. R. Chem. Commun.
2001, 9, 811.
12. Meyers, A. I.; Bienz, S.; Kwon, H.; Wallace, R. H. Helv.
Chim. Acta 1996, 79, 1026.
13. For 8, two diastereoisomers, white solid; dH (300MHz,
CDCl3) 1.92–1.98 (1H, overlapping AB systems, dt,
J = 14.3, 5.0 and dt, J = 14.1, 4.6, H-15a), 2.17–2.25 (1H,
overlapping AB systems, ddd, J = 15.2, 7.2, 5.6 and ddd,
J = 14.4, 7.0, 5.4, H-15b), 3.18–3.22 (1H, overlapping AB
systems, dt, 11.7, 4.5 and dt, J = 11.6, 4.6, H-3), 3.42 (3H,
s, OMe), 3.48 (1H, dd, J = 11.5, 9.4, H-4), 3.69 (1H, t,
J = 10.3, H-6ax), 4.27 (1H, td, J = 9.6, 5.1, H-5), 4.31 (1H,
dd, J = 10.4, 4.9, H-6eq), 4.58 (1H, s, H-1), 5.01 (2H, m, H-
17a,b), 5.39–5.47 (2H, m, H-16, H-7), 7.28–7.44 (5H, m,
Ph). dC (75.5MHz, CDCl3), 27.0/27.2 (CH2, C-15), 47.0/
47.1 (CH, C-3), 55.7 (CH3, C-14), 64.4 (CH, C-5), 69.0
(CH2, C-6), 81.3/81.4 (CH, C-4), 93.9 (CH, C-17), 100.8
(CH, C-1), 101.5 (CH, C-7), 101.7 (CH, C-16), 126.1 (CH,
o-Ph), 128.3 (CH, m-Ph), 129.2 (CH, p-Ph), 136.8 (C, Ph),
199.0 (C, C-2). ESI MS found for C17H24O8 (M+H)+ 353,
(MÀOMe)+ 321. Mp = 118–121°C.
For 12a, colourless oil; dH (400MHz, CDCl3) 0.87 (3H, d,
J = 6.3, H-11a), 1.03 (3H, d, J = 6.7, H-11b), 1.71 (1H, dtd,
J = 12.7, 9.4, 6.7, H-8b), 1.90 (1H, m, H-8a), 2.25 (1H, dt,
J = 8.8, 5.9, H-3), 2.75 (1H, td, J = 10.1, 3.0, H-9a), 2.90
(1H, d, J = 5.9, H-2), 2.94 (1H, obscured m, H-9b ), 2.99
(1H, sept, J = 6.5, H-10), 3.23 (2H, br s, 4-OH, 6-OH),
3.38 (3H, s, H-7), 3.57 (2H, overlapping m, H-4, H-5), 3.84
(1H, dd, J = 11.6, 3.3, H-6a), 4.18 (1H, dd, J = 11.6, 3.9,
H-6b), 4.64 (1H, s, H-1). dC (75.5MHz, CDCl3) 14.1 (CH3,
C-11a), 22.1 (CH3, C-11b), 27.2 (CH2, C-8), 43.0 (CH, C-
3), 43.9 (CH, C-9), 49.7 (CH3, C-10), 54.9 (CH3, C-7), 63.0
(CH, C-2), 63.1 (CH2, C-6), 67.1 (CH, C-5), 70.9 (CH, C-
4), 98.8 (CH, C-1). FAB HRMS calcd. for C12H23O4N
(M+H)+ 246.17053, found 246.17047.
14. Chen, L.; Wiemer, D. F. J. Org. Chem. 2002, 67, 7561.
15. Hon, Y.; Lu, L. Tetrahedron Lett. 1993, 34, 5309.
16. For 13, major diastereoisomer, white solid; dH (300MHz,
CDCl3) 1.52 (1H, br s, 2-OH), 2.01 (2H, dd, J = 11.2, 5.9,
H-15a,b), 2.14 (1H, tt, J = 16.4, 5.5, H-3), 3.24 (1H, dd,
J = 10.6, 9.1, H-4), 3.40 (3H, s, H-14), 3.42 (1H, d, J = 4.9,
H-2), 3.62 (1H, t, J = 10.1, H-6ax), 3.71 (1H, td, J = 10.0,
4.7. H-5), 4.20 (1H, dd, J = 9.8, 4.4, H-6eq), 4.63 (1H, d,
J = 3.7, H-1), 5.05 (2H, m, H-17a,b), 5.42 (1H, s, H-7), 5.49
(1H, dd, J = 11.9, 6.6, H-16), 7.27–7.43 (5H, m, Ph). dC
(75.5MHz, CDCl3) 29.1 (CH2, C-15), 39.1 (CH, C-3), 55.4
(CH3, C-14), 63.7 (CH, C-5), 69.2 (CH2, C-6), 71.6 (CH,
C-2), 80.2 (CH, C-4), 93.9 (CH2, C-17), 99.0 (CH, C-1),
101.6 (CH, C-7), 102.6 (CH, C-16), 126.1 (CH, o-Ph),
128.3 (CH, m-Ph), 129.0 (CH, p-Ph), 137.3 (C, Ph).
ESI MS found for C17H22O8 (M+H)+ 355, (MÀOMe)+
323.
22. Crystal data for C16H21NO5 (12g), M = 307.34, mono-
clinic, space group P21, a = 11.8679(15), b = 8.1933(13),
3
˚
˚
c = 16.417(2) A, b = 96.296(11)°, V = 1586.74(4) A , T =
210(2)K, Z = 4, Dc = 1.287gcmÀ3
,
l
(Mo-Ka) =
0.096mmÀ1. Final, R1 = 0.0336 (for 2624 reflections with
I > 2r(I)) and wR2 = 0.0911 for all data. Crystallographic
data for 12g has been deposited with the Cambridge
Crystallographic Data Centre as supplementary publica-
tion number CCDC 252127. Copies of the data can be
obtained, free of charge, on application to CCDC, 12
Union Road, Cambridge CB2 1EZ, UK [fax: +44(0)-1223-
336033 or e-mail: deposit@ccdc.cam.ac.uk].
17. Hon, Y.; Yan, J. Tetrahedron Lett. 1993, 41, 6591.
18. For 9, colourless oil; dH (400MHz, CDCl3 ) 2.79 (1H, ddd,
J = 18.0, 4.3, 1.1, H-15a), 2.92 (1H, ddd, J = 18.0, 6.8, 0.9,
H-15b), 3.54 (3H, s, OMe), 3.59 (1H, ddd, J = 11.2, 6.8,
4.3, H-3), 3.67 (1H, dd, J = 11.7, 9.0, H-4), 3.80 (1H, t,
J = 10.4, H-6ax), 4.27 (1H, ddd, J = 9.7, 9.2, 4.8, H-5), 4.42
23. Sakakibara, T.; Shindo, T.; Hirai, H. Carbohydr. Res.
2002, 337, 2061.