Amadori ketoses with calcium hydroxide and the Kiliani reaction on 1-deoxy ketoses: Two approaches to the synthesis of saccharinic acids

Article abstract of DOI:10.1016/j.tetlet.2005.11.018

Saccharinic acids (2-C-methyl aldonic acids) may be formed by treatment of Amadori ketoses with calcium hydroxide or by the Kiliani reaction of 1-deoxy ketoses with cyanide. Thus (i) N,N-dibenzyl or N,N-dimethyl-1-amino-1-deoxy-D- fructose with aqueous ca

Full text of DOI:10.1016/j.tetlet.2005.11.018

Tetrahedron
Letters
Tetrahedron Letters 47 (2006) 315–318
Amadori ketoses with calcium hydroxide and the
Kiliani reaction on 1-deoxy ketoses: two approaches to the
synthesis of saccharinic acids
David J. Hotchkiss,a Sarah F. Jenkinson,a Richard Storer,b Thomas Heinzc and
George W. J. Fleeta,*
aChemistry Research Laboratory, Department of Chemistry, University of Oxford, Mansfield Road, Oxford OX1 3TA, UK
b
`
Laboratoires Idenix Sarl, Parc Eurome´decine, Rue de la Valsiere, 34090 Montpellier, France
cChemical and Analytical Development, Novartis Pharma AG, Postfach, CH-4002 Basel, Switzerland
Received 27 October 2005; accepted 3 November 2005
Available online 22 November 2005
Abstract—Saccharinic acids (2-C-methyl aldonic acids) may be formed by treatment of Amadori ketoses with calcium hydroxide or
by the Kiliani reaction of 1-deoxy ketoses with cyanide. Thus (i) N,N-dibenzyl or N,N-dimethyl-1-amino-1-deoxy-D-fructose with
aqueous calcium hydroxide afforded 2-C-methyl-D-ribono-1,4-lactone under green conditions and (ii) reaction of methyl magnesium
bromide with 2,3-O-isopropylidene-D-erythronolactone gave 1-deoxy-3,4-O-isopropylidene-D-ribulose, which on subsequent treat-
ment with aqueous sodium cyanide and hydrolysis, formed 2-C-methyl-D-arabinono-1,4-lactone. Such branched sugar lactones are
likely to be of value as chirons containing branched carbon chains.
Ó 2005 Published by Elsevier Ltd.
1. Introduction
hydroxymethyl group. This letter exemplifies two hith-
erto unexplored approaches to carbohydrate lactones
bearing a methyl branch at C-2 to form saccharinic acids
(2-C-methyl aldonic acids). The strategies are most
clearly seen using Fischer projection formulae (Scheme
1).
Although almost all readily available carbohydrate scaf-
folds contain linear carbon chains,1 the Kiliani ascen-
sion2 on unprotected ketoses3 provides access to
branched carbohydrate lactones under environmentally
friendly conditions; the convenient preparations of
branched carbon chain diacetonides from D-fructose,
L-sorbose,4 D-tagatose and D-psicose5 suggest that a
new family of carbohydrate chirons containing
branched carbon chains will be useful for the efficient
synthesis of complex homochiral targets.6 The Kiliani
reaction on ketohexoses limits the branch at C-2 to a
The first approach involved the reaction of D-glucose 1
with a secondary amine to give the corresponding fruc-
tosamine 2 by the Amadori reaction; treatment of the
Amadori ketohexose 2 with calcium hydroxide in water
gave the branched ribonic acid 3, isolated as its lactone.
The overall reaction is simply an isomerisation between
CHO
OH
CH2NR2
O
CH3
H
HO
H
CO2H
OH
CO2H
Me
O
H
OH
HO
H
H
Me
H
HO
H
CO2H
OH
H
H
OH
OH
OH
OH
OH
H
H
H
OH
H
OH
H
OH
H
OH
OH
CH2OH
CH2OH
CH2OH
CH2OH
CH2OH
CH2OH
5
1
2
3
4
6
Scheme 1. Approaches to 2-C-methyl aldonic acids.
*
Corresponding author. E-mail: george.fleet@chem.ox.ac.uk
0040-4039/$ - see front matter Ó 2005 Published by Elsevier Ltd.
doi:10.1016/j.tetlet.2005.11.018

316
D. J. Hotchkiss et al. / Tetrahedron Letters 47 (2006) 315–318
the six carbon sugar glucose 1 and the saccharinic acid 3.
No protection is needed during the transformation.
Alternatively two sequential one carbon nucleophilic
additions to a four carbon sugar provided access to a
six carbon saccharinic acid. Thus addition of methyl
magnesium bromide to a lactone derived from 6 affor-
ded a protected 1-deoxy-D-ribulose 5; subsequent addi-
tion of cyanide resulted in a Kiliani chain extension to
give 2-C-methyl-D-arabinonic acid 4.
groups at C-1 other than the hydroxyl function is
lengthy and cumbersome.
The Amadori rearrangement is the principal mechanism
for non-enzymatic conjugation of proteins to carbohy-
drates; this directly introduces an amino function as
an alternative leaving group to the hydroxide ion at
C-1 of the ketose.13 The reaction of dibenzylamine with
glucose 1 in ethanol in the presence of acetic acid gave
the Amadori ketose 11 as a crystalline material14 in 86%
yield.15 Subsequent treatment of the dibenzyl ketose-
amine 11 with calcium hydroxide in water afforded, after
acidic work-up, the branched ribono-1,4-lactone 10 in
16% yield. This represented a significant increase in yield
in comparison to that obtained in the base treatment of
fructose and, more importantly, reduced the time and
effort in preparing 10 (Scheme 3).
2. Reaction of Amadori ketoses with calcium hydroxide
The reaction of glucose with aqueous base to give sac-
charinic acids7 is one of the oldest reactions known; it
is also one of the most complicated.8 More than 50 com-
pounds have been identified from the reaction of glucose
with aqueous calcium hydroxide with only a trace of the
saccharinic acid 3 being formed.9 A plausible pathway
for the isomerisation of 1 to 3 is illustrated in Scheme
2;10 an initial Lobry de Bruyn rearrangement11 of glu-
cose 1 to fructose 7 is followed by formation of ene-diol
8. Base catalysed dehydration of 8 to an enol, followed
by tautomerism, affords the diketone 9. A subsequent
benzilic acid rearrangement of 9 leads to the branched
ribonic acid 3, invariably isolated as its crystalline lac-
tone 10.
The insolubility of the dibenzylamine 11 in water pre-
sented some practical problems in the reaction with cal-
cium hydroxide. Thus attempts were made to synthesise
a water-soluble Amadori compound using dimethyl-
amine as the amine component for the rearrangement.
The isolation of the water-soluble dimethylamine Ama-
dori ketose 12 proved challenging; however, both 13C
NMR spectroscopy (in the formation of a new quater-
nary C and of an additional methylene group) and mass
spectrometry provided evidence for the formation of 12
together with the absence of glucose. Subsequent cal-
cium hydroxide treatment of crude 12 afforded 2-C-
methyl-D-ribono-1,4-lactone 10 in a 19% overall yield
from D-glucose.
Although the yield of the branched ribonolactone 10
from reaction of glucose with calcium hydroxide was
only 0.4%,9 treatment of fructose with calcium hydrox-
ide under carefully optimised conditions of temperature
and time allowed the isolation of 10 in a yield of around
11%.12 The time of the reaction, together with the need
to remove large volumes of water, make the production
of 10 a troublesome procedure. The key step for the gen-
eration of diketone 9 is the base-catalysed elimination of
water from 8. However, the introduction of leaving
A practical method for the synthesis of lactone 10 is as
follows: D-Glucose 1 (3.17 g, 17.6 mmol) suspended in
ethanol (5 mL) and glacial acetic acid (1 mL) was trea-
ted with a solution of dimethylamine (33% solution in
methylated spirits, 3.2 mL, 18.0 mmol) and the reaction
CH2OH
OH
Me
CH2OH
OH
HOH2C
HO
CO2H
Me
HOH2C
HO
HOH2C
HO
HOH2C
HO
O
OH
OH
O
O
HOH2C
O
O
Me
HO
OH
OH
O
HO
OH
OH
OH
OH
OH
OH
OH
1
7
8
9
3
10
Scheme 2. Base catalysed isomerisations of glucose and fructose.
CH2N(CH2Ph)2
O
OH
(ii)
(i)
HO
OH
OH
HOH2C
HO
O
OH
OH
O
HOH2C
O
HOH2C
O
11
O
OH
Me
HO OH
HO
Me
OH
CH2NMe2
OH
(iii)
O
13
10
1
none isolated
HO
OH
OH
12
Scheme 3. Reagents and conditions: (i) (PhCH2)2NH, AcOH, EtOH; (ii) Ca(OH)2, H2O; then H+; (iii) Me2NH in EtOH/MeOH, AcOH; then
Ca(OH)2, H2O; then H+.

D. J. Hotchkiss et al. / Tetrahedron Letters 47 (2006) 315–318
317
O
O
O
O
O
O
HOH C
(i)
OH
Me
HOH C
(ii)
2
2
O
O
O
+
OH
OH
Me
O
Me
HO
Me
HO OH
O
O
O
O
O
16
13
10
(iii)
minor isomer
14
15
Scheme 4. Reagents and conditions: (i) MeMgBr, THF, ꢁ78 °C; (ii) NaCN, H2O; then Amberlite IR 120 H+; (iii) CF3COOH, H2O.
mixture stirred at 80 °C for 1.5 h. The resulting dark
orange solution was concentrated in vacuo to afford a
dark oil (6.55 g), which was dissolved in water (150
mL). The aqueous solution was stirred at 70 °C with cal-
cium oxide (5.61 g, 100 mmol) for 24 h. Oxalic acid
dihydrate (6.74 g, 53.5 mmol) was added to the suspen-
sion; the reaction mixture was filtered through CeliteÒ.
The filtrate was passed through AmberliteÒ IR 120 ion
exchange resin. The residue obtained after removal of
water was re-dissolved in water (50 mL) and heated at
40 °C for 15 min. The water was removed and the crude
product purified by chromatography and subsequent
crystallisation to afford the lactone 10 (534 mg, 19%).16
introduced at C-2 of the lactone. In particular, the over-
all isomerisation of glucose 1 to 2-C-methyl-D-ribonic
acid 3 by a green, environmentally friendly procedure
indicates that a new class of chiron may be readily avail-
able. The scope and generality of the reaction of Ama-
dori ketoses with calcium hydroxide is presently under
investigation.25
Acknowledgements
A Postdoctoral Fellowship for S.F.J. was provided by
Idenix Pharmaceuticals and Novartis Pharma AG.
References and notes
None of the epimeric arabinonolactone 13 was isolated
during any of the reactions reported; a possible rationale
for this exclusive diastereoselectivity is under investiga-
tion. Nonetheless this modification of the treatment of
a ketose with base provides a much more convenient
synthesis of the lactone 10 than has been hitherto
available.
1. Lichtenthaler, F. W.; Peter, S. Compt. Rend. Chim. 2004,
7, 65–90.
2. Hudson, C. S. Adv. Carbohydr. Chem. 1945, 1, 2–36;
Hudson, C. S. J. Am. Chem. Soc. 1951, 73, 4498–4499;
Pratt, J. W.; Richtmyer, N. K.; Hudson, C. S. J. Am.
Chem. Soc. 1953, 75, 4503–4545.
3. Ferrier, R. J. J. Chem. Soc. 1962, 3544–3549; Kiliani, H.
Ber. Dtsch. Chem. Ges. 1885, 18, 3066–3072; Kiliani, H.
Ber. Dtsch. Chem. Ges. 1886, 19, 221–227; Kiliani, H. Ber.
Dtsch. Chem. Ges. 1928, 61, 1155–1169; Woods, R. J.;
Neish, A. C. Can. J. Chem. 1953, 31, 471–475; Woods, R.
J.; Neish, A. C. Can. J. Chem. 1954, 32, 404–414; Gorin, P.
A. J.; Perlin, A. S. Can. J. Chem. 1956, 36, 480–485.
4. Hotchkiss, D. J.; Soengas, R.; Simone, M. I.; van Ameijde,
J.; Hunter, S.; Cowley, A. R.; Fleet, G. W. J. Tetrahedron
Lett. 2004, 45, 9461–9464.
5. Soengas, R.; Izumori, K.; Simone, M. I.; Watkin, D. J.;
Skytte, U. P.; Soetart, W.; Fleet, G. W. J. Tetrahedron
Lett. 2005, 46, 5755–5759.
6. Simone, M. I.; Soengas, R.; Newton, C. R.; Watkin, D. J.;
Fleet, G. W. J. Tetrahedron Lett. 2005, 46, 5761–5765.
7. Sowden, J. C. Adv. Carbohydr. Chem. 1957, 12, 35–79.
8. Peligot, E. Compt. Rend. 1838, 7, 106–108.
9. Yang, B. Y.; Montgomery, R. Carbohydr. Res. 1996, 280,
27–45.
3. Kiliani reaction of 1-deoxy ketoses
An alternative approach to the synthesis of saccharinic
acids is illustrated in Scheme 4 in which a six carbon sac-
charinic acid is prepared by two sequential additions of
nucleophilic carbon to a four carbon sugar. Reaction of
the protected D-erythronolactone 1417 with methyl mag-
nesium bromide in THF afforded the lactol 15, mp 82–
85 °C (lit.18 87–89 °C), as a mixture of anomers in 99%
yield. Treatment of the protected 1-deoxy ribulose 15
with aqueous sodium cyanide, followed by work-up
with acid, gave the protected d-lactone 1619 in 18% yield
and a mixture of the deprotected c-lactones 1320–22 and
10 in an approximate ratio of 5 to 1 and a combined
yield of 45%. Recrystallisation of the mixture of c-lac-
tones afforded the pure arabinonolactone 13.
10. Bols, M. Carbohydrate Building Blocks; John Wiley &
Sons: New York, 1996.
The isopropylidene protecting group in 16 was removed
in quantitative yield by treatment with aqueous trifluo-
roacetic acid to give 2-C-methyl-D-arabinono-1,4-lac-
tone 13. The structures of the protected23 and
unprotected24 arabinonolactones were firmly established
by X-ray crystallographic analysis.
11. Ekeberg, D.; Morgenlie, S.; Stentstrom, Y. Carbohydr.
Res. 2005, 340, 373–377; Angyal, S. J. Top. Curr. Chem.
2001, 215, 1–14; de Bruyn, C. A. L.; van Ekenstein, W. A.
Rec. Trav. Chim. 1897, 16, 265–278.
12. Whistler, R. L.; BeMiller, J. N. Methods Carbohydr.
Chem. 1963, 2, 484–485.
13. Hodge, J. E. Adv. Carbohydr. Chem. 1955, 10, 169–205;
Funcke, W.; Klemer, A. Carbohydr. Res. 1976, 50, 9–13;
Funcke, W. Annalen 1978, 2099–2104.
14. Hou, Y.; Wu, X.; Xie, W.; Braunschweiger, P. G.; Wang,
P. G. Tetrahedron Lett. 2001, 42, 825–829.
4. Summary
The two complementary procedures provide practical
access to the epimeric 2-C-methyl aldonolactones 10
and 13, and the Kiliani reaction on 1-deoxy ketoses
could potentially allow groups other than methyl to be
15. Hodge, J. E.; Fisher, B. E. Methods Carbohydr. Chem.
1963, 2, 99–107.
16. Selected data for 10: mp 158–159 °C (lit.12 160–161 °C);
21
20
½aꢀD +87.5 (c 0.76, water) [lit.12 ½aꢀD +93 (water)]; mmax

318
D. J. Hotchkiss et al. / Tetrahedron Letters 47 (2006) 315–318
20. Aparicio, F. J. L.; Cubero, I. I.; Olea, M. D. P. Carbohydr.
(film): 3356 (s, br, O–H), 1773 (s, C@O) cmꢁ1; dH
(CD3OD, 400 MHz): 1.40 (3H, s, CH3), 3.71 (1H, dd,
Res. 1984, 129, 99–109.
0
H-5, J5,4 4.5, J5;5 12.8), 3.90 (1H, d, H-3, J3,4 7.8), 3.94
21. Novak, J. J. K. Collect. Czech. Chem. Commun. 1974, 39,
869–882.
(1H, dd, H-50, J5 ;4 2.4, J5 ;5 12.8), 4.29 (1H, ddd, H-4, J4,5
4.5, J4;5 2.4, J4,3 7.8); dC (CD3OD, 100.6 MHz): 20.02
(CH3), 60.03 (C-5), 72.55 (C-2), 72.64 (C-3), 83.43 (C-4),
176.97 (C-1); dH (DMSO-d6, 400 MHz): 1.24 (3H, s, CH3),
0
0
0
22. Selected data for 13: mp 65–68 °C (from ethyl acetate/
cyclohexane) [lit.20 89–91 °C; lit.21 63–65 °C (moist ethyl
21
acetate)]; ½aꢀD +78.7 (c 1.02, H2O) [lit.20 [a]D +69 (c 1.6,
20
3.47–3.54 (1H, ddd, H-5, J5,4 5.2, J5,OH 5.9, J5;5 12.7), 3.72
H2O); lit.21 ½aꢀD +82.5 (c 0.9, H2O)]; mmax (thin film): 3383
0
(1H, dd, H-3, J3,OH 7.4, J3,4 7.8), 3.72–3.77 (1H, ddd, H-50,
(s, br, O–H), 1771 (s, C@O); dH (CD3OD, 500 MHz): 1.34
0
0
0
0
J5 ;4 2.3, J5 ;OH 5.2, J5 ;5 12.7), 4.13–4.17 (1H, m, H-4),
5.00–5.03 (1H, dd, OH-5, JOH;5 5.4, JOH,5 5.9), 5.43 (1H,
(3H, s, CH3), 3.70 (1H, dd, H-5, J5,4 4.7, J5;5 12.8), 3.90
(1H, dd, H-50, J5 ;4 2.5, J5 ;5 12.8), 4.04–4.08 (1H, m, H-4),
4.16 (1H, d, H-3, J3,4 7.8); dC (CD3OD, 100.6 MHz): 18.0
(CH3), 60.9 (C-5), 75.13 (C-3), 76.9 (C-2), 82.1 (C-4), 179.3
(C-1); dH (DMSO-d6, 400 MHz): 1.18 (3H, s, CH3), 3.46–
0
0
0
d, OH-3, JOH,3 7.3), 5.71 (1H, s, OH-2); dC (DMSO-d6,
100.6 MHz): 20.9 (CH3), 59.5 (C-5), 72.0 (C-2), 72.1 (C-3),
82.8 (C-4), 176.2 (C-1); m/z (ESIꢁ): 161 (MꢁH+, 100%).
0
17. Cohen, N.; Banner, B. L.; Laurenzano, A. J.; Carozza, L.
Org. Synth. 1984, 63, 127–135.
3.52 (1H, ddd, H-5, J5,4 4.8, J5,OH 5.8, J5;5 12.7), 3.69–3.74
(1H, ddd, H-50, J50;4 2.3, J5 ;OH 5.3, J5 ;5 12.7), 3.91–3.95
0
0
0
18. Cubero, I. I.; Lopez-Espinosa, M. T. P. Carbohydr. Res.
1986, 154, 71–80.
(1H, ddd, H-4, J4;5 2.3, J4,5 4.7, J4,3 8.0), 3.99–4.02 (1H,
dd, H-3, J3,OH 5.3, J3,4 8.0), 5.07 (1H, a-t, OH-5, JOH,5
5.6), 5.71 (1H, d, OH-3, JOH,3 5.2), 5.79 (1H, s, OH-2); dC
(DMSO-d6, 100.6 MHz): 18.0 (CH3), 59.6 (C-5), 73.5
(C-3), 75.1 (C-2), 81.8 (C-4), 178.0 (C-1); m/z (ESIꢁ): 161
(MꢁH+, 100%).
21
19. Selected data for 16: mp 106–109 °C; ½aꢀD ꢁ134.8 (c 1.0,
CHCl3); mmax (thin film): 3449 (s, br, OH), 1742 (s, C@O);
dH (CD3CN, 500 MHz): 1.34 (3H, s, CCH3), 1.36 (3H, s,
CCH3), 1.45 (3H, s, CH3), 4.03 (1H, br, s, O–H), 4.25–4.28
0
(1H, dd, H-5, J5,4 0.9, J5;5 12.3), 4.31–4.33 (1H, d, H-3,
23. Punzo, F.; Watkin, D. J.; Jenkinson, S. F.; Fleet, G. W. J.
Acta Cryst. 2005, E61, o127–o129.
J3,4 7.5), 4.57–4.59 (1H, m, H-4), 4.85–4.88 (1H, dd, H-50,
0
0
J5 ;4 2.0, J5 ;5 12.3); dC (CD3CN, 125.7 MHz): 22.3 (CH3),
24.2 (CCH3), 26.5 (CCH3), 69.3 (C-5), 72.7 (C-2), 73.2 (C-
4), 80.0 (C-3), 110.0 (C(CH3)2), 171.6 (C-1); m/z (ESIꢁ):
247 (M þ HCOꢁ2 , 100%).
24. Punzo, F.; Watkin, D. J.; Jenkinson, S. F.; Fleet, G. W. J.
Acta Cryst. 2005, E61, o326–o327.
25. Soengas, R.; Hotchkiss, D. J.; Eastwick-Field, V.; Watkin,
D. J.; Fleet, G. W. J., in preparation.