Evidence on the structure of coyolosa. Synthesis of 6,6′-ether linked hexoses

Article abstract of DOI:10.1016/j.tetlet.2003.11.033

The structure previously reported for coyolosa, a hypoglycemic substance isolated from natural sources, as a 6,6′-ether linked pseudo-disaccharide derived from allose, is incorrect since 6-O-(6-deoxy-D-allos-6-yl)-D-allose 10, synthesised by an unequivoca

Full text of DOI:10.1016/j.tetlet.2003.11.033

Tetrahedron
Letters
Tetrahedron Letters 45 (2004) 835–837
Evidence on the structure of coyolosa. Synthesis of 6,60-ether
linked hexoses
Alan H. Haines*
Centre for Carbohydrate Chemistry, School of Chemical Sciences and Pharmacy, University of East Anglia, Norwich NR4 7TJ, UK
Received 29 September 2003; revised 21 October 2003; accepted 6 November 2003
Abstract—The structure previously reported for coyolosa, a hypoglycemic substance isolated from natural sources, as a 6,60-ether
linked pseudo-disaccharide derived from allose, is incorrect since 6-O-(6-deoxy-D-allos-6-yl)-D-allose 10, synthesised by an
unequivocal route, and its derived peracetate 11, both have spectral properties, which differ from those reported for coyolosa and its
peracetate. A versatile synthesis has been developed for this type of compound, containing two carbohydrate moieties joined
through an ether linkage at their primary centres, which involves the displacement of a primary carbohydrate triflate with a car-
bohydrate alkoxide, and it allows the preparation of both symmetrical and unsymmetrical representatives of this class of com-
pound.
ꢀ 2003 Elsevier Ltd. All rights reserved.
Considerable interest has been raised by the report1 of a
new hypoglycemic compound that was isolated by
methanol extraction of the Acrocomia mexicana root.
The active compound was termed ÔcoyolosaÕ and on the
basis of elemental and spectral analysis a structure was
suggested as comprising two hexopyranose units joined
through an ether link at their 6-positions, which can be
represented by the general formula 1. Only one set of
signals for H-1 to H-6 and C-1 to C-6 were observed in
spectrum. Therefore, in order to confirm the proposed
structure of coyolosa, or to throw light on alternatives, a
synthesis of the 6,60-ether linked allo-isomer was
undertaken; the availability of the synthetic compound
should allow a direct comparison with the natural
material.2
HO
OH
OH
1
the H and 13C NMR spectra, respectively, indicating
O
OH
O
H2C
O
either identical or enantiomeric related sub-units. The
structure 2 drawn in the original publication depicts the
O CH2
O
HO
HO
OH
HO
ether-linked pseudo-disaccharide with
pyranose moieties, an unusual combination, but a
- and - or an - and -combination would also satisfy
D- and L-allo-
OH
2
HO
OH
HO
D
D
L
L
1
2
the NMR data in a nonchiral solvent. Unfortunately, no
optical rotation data were given for coyolosa, mea-
surements which would have ruled out the
L
D- and
The most direct method for forging the 6,60-linkage is
direct displacement of a suitably protected 6-triflate of
-combination if optical activity had been observed.
Furthermore, there is some confusion in the report1
regarding the solvent used for measurement of the NMR
spectrum of coyolosa (DMSO being given in the text
and CDCl3 in Table 1), and in the reporting of the
D
-allose with the alkoxide of the corresponding alcohol.
3-O-Benzyl-1,2-di-O-isopropylidene-a- -allofuranose
(Scheme 1, 3)3 was prepared from 1,2:5,6-di-O-isopro-
pylidene-a-
-allofuranose4 by sequential O-benzylation
D
1
D
chemical shifts and coupling constants in the H NMR
and partial hydrolysis. Treatment of 3 with tert-butyl-
diphenylsilyl chloride gave the 6-O-silyl derivative 4,5
benzylation of which gave the 3,5-di-O-benzyl derivative
5. Removal of the silyl protecting group in 5 afforded the
6-hydroxy compound 6, which gave the 6-triflate 7 on
treatment with triflic anhydride. Reaction of the sodium
Keywords: Coyolosa; Acrocomia mexicana; Ether-linked sugars; Hypo-
glycemic.
* Fax: +44-1603-592-015; e-mail: a.haines@uea.ac.uk
0040-4039/$ - see front matter ꢀ 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2003.11.033

836
A. H. Haines / Tetrahedron Letters 45 (2004) 835–837
R1O
BnO
for the carbohydrate carbons, with the value for the
anomeric signal (dC 90.06) agreeing well with that
reported6 for peracetylated b- -allopyranose (dC 90.1).7
R1O
R2O
O
(c)
O
D
The reported1 values for the 13C NMR signals of the
carbohydrate carbons of coyolosa peracetate are dC
91.77, 89.13, 72.82, 69.90, 67.88 and 61.54 whereas
values for compound 11 in the same solvent are at 90.06,
72.90, 70.22, 68.28, 68.03 and 65.90. Thus, it is clear that
coyolosa does not possess the structure suggested1 and
in an attempt to identify a possible alternative structure,
O
O
O
BnO
CMe2
O
BnO
CMe2
3 R1 = R2 = H
4 R1 = t-BuSiPh2Si, R2 = H
5 R1 = t-BuSiPh2Si, R2 = Bn
6 R1 = H
(a)
(b)
(d)
7 R1 = CF3SO2-
the 6,60-ether linked pseudo-disaccharides from
cose and -galactose have also been similarly prepared
D-glu-
Scheme 1. Reagents and conditions: (a) t-BuSiPh2Cl/C5H5N, 94%; (b)
NaH/PhCH2Br/DME, 68%; (c) Bu4NF/THF, 91%; (d) Tf2O/Et3N/
CH2Cl2, 80%.
D
and full details of their synthesis will be reported else-
where. However, the spectral characteristics of neither
of these compounds, most noticeably the 13C NMR
spectra of the peracetates, are in agreement with those
reported for coyolosa peracetate and therefore the
structure of coyolosa is still an open question.
alkoxide of 6 with triflate 7 in THF (Scheme 2) gave the
dimeric structure 8, hydrogenation of which gave the
6,60-ether 9, which on acidic hydrolysis gave the 6,60-
ether linked
D-allose 10 as a hygroscopic solid that
could exist, presumably, in the di-furanose, di-pyranose,
or a furanose–pyranose form. Coyolosa has a reported1
mp of 170–172 ꢁC. The 1H NMR spectrum of 10 in D2O
suggested that the b,b-pyranose form was predominant
(ꢀ80%, dH 4.88, J1;2 8.2 Hz, H-1) with a minor signal
(13% of the total anomeric signal) for an a-pyranose
anomeric form (dH 5.13, J1;2 3.4 Hz) and two small sig-
nals (7% of total) at dH 5.23 and 5.36. These values for
the anomer ratios agree well with those reported6 for the
Despite a report9 in 1969 that an ether-linked pseudo-
disaccharide was a constituent of the exotoxin from
Bacillus thuringiensis, which has inhibitory action on the
de novo synthesis of RNA and the DNA dependent
RNA polymerase, this class of carbohydrates has been
given relatively little attention to date. Synthesis of the
sugar fragment of this endotoxin, which contains a
5 fi 4
D-ribose to D
-glucose ether link, was reported10 in
1971, but the first synthesis of this type of compound
was by Whistler and Frowein11 in 1961 who heated
together equimolar amounts of 1,2-O-isopropylidene-a-
composition of D-allose in aqueous solution of 77.5%
b-pyranose, 14% a-pyranose and 8.5% furanose forms.
Although individual hydroxy proton resonances could
D
-glucofuranose and 5,6-anhydro-1,2-O-isopropylidene-
a- -glucofuranose at 150 ꢁC, which yielded, after hydro-
lysis of the product and extensive purification by column
1
be observed in the H NMR spectrum of 10 in DMSO-
D
d6 as reported,1 the separate C–H proton resonances
were generally uninformative due to signal overlap and
increased complexity due to coupling with the hydroxy
protons. The 13C NMR spectrum of 10 showed two
anomeric signals at dC 94.19 (major) and 93.57 (minor),
which are in close agreement with those for b- and a-
chromatography, 6,60-di-
D-glucose anhydride. A related
approach has been used by Villa and co-workers12 to
construct an amphiphile containing 6,60-ether linked
D
-glucose residues. An attempt to use this type of reaction
to prepare 10 in the present work, by fusing together 3,5-
di-O-benzyl-1,2-O-isopropylidene-a- -allofuranose and
-allo-
D
-allose at 94.3 and 93.7, respectively.7
D
5,6-anhydro-3-O-benzyl-1,2-O-isopropylidene-a-
D
Acetylation of 10 gave, after recrystallisation from
EtOAc–hexane, a crystalline peracetate 11, mp 198–
199 ꢁC, which contrasts8 with that for coyolosa perace-
tate, prepared1 by acetylation of coyolosa by a similar
procedure, which had mp 132–134 ꢁC. The H NMR
spectrum of 11 confirmed it as a single isomer with a
b-configuration at the anomeric centres (dH 5.97, J1;2
8.7 Hz, H-1) and its 13C NMR spectrum had six signals
furanose13 was unsuccessful, the reactants showing
remarkable stability even at elevated temperatures. A
patent14 records the preparation of the 6,60-ether linked
derivative from 1,2:3,4-di-O-isopropylidene-D-galactose,
Paulsen and von Deyn15 have linked some cyclitol
derivatives to carbohydrates to produce pseudo-disac-
charides with ether linkages using triflate methodology,
Schmidt and co-workers16 introduced a 2 fi 6 ether link
1
O
O
O
R1O
H2C
HO
CH2
HO
(c)
(a)
O
O
O
O
HO
HO
6
OH
OH
HO
HO
OH
OH
O
HO
R1O
CMe2
O
HO
(d)
2
2
8 R1 = Bn
9 R1 = H
10
11
(b)
Scheme 2. Reagents and conditions: (a) NaH/THF then 7, 89%; (b) H2/Pd–C/EtOH–EtOAc, 85%; (c) CF3CO2H/H2O (9:1), 96%; (d) Ac2O/C5H5N,
62%.

A. H. Haines / Tetrahedron Letters 45 (2004) 835–837
837
79.45 (4-C), 77.68 (2-C), 77.43 (5-C), 77.13 (3-C), 73.52
and 72.04 (2 · CH2Ph), 71.41 (6-C), 26.84 and 26.54
(CMe2); m=z (CI): 800.5 [M+NH4]þ. (Found: [M+NH4]þ
800.3995. C46H58NO11 requires m=z 800.4004.)
between
D-gluconolactone and D-glucose residues, and
17
ꢀ
Hodosi and Kovac, in investigating the use of 1,2-
stannylene acetals for glycoside synthesis, isolated
compounds with 3 fi 6, 2 fi 4 and 2 fi 6 ether links be-
tween two carbohydrate residues, in one case as the
major product. It would seem that ether-linked pseudo-
disaccharides merit further attention, both from the
synthetic standpoint and because of their potentially
interesting biological properties.
9 (oil): ½aꢁ +56.7ꢁ (c 1.33, CHCl3); dH (400 MHz, CDCl3)
D
5.71 (1H, d, J1;2 3.7 Hz, 1-H), 4.58 (1H, dd, J2;3 4.3 Hz, 2-
H), 4.09 (1H, dd, J3;4 8.6 Hz, 3-H), 4.01 (1H, m, 5-H), 3.85
(1H, dd, J4;5 3.5 Hz, 4-H), 3.68 (1H, dd, J5;6 3.4, J6a;6b
10 Hz, 6a-H), 3.60 (1H, dd, J5;6b 3.9 Hz, 6b-H), 1.51 and
1.29 (each 3H, CMe2); dC (100 MHz, CDCl3) 112.78
(CMe2), 103.42 (1-C), 81.08 (4-C), 79.75 (2-C), 71.30 (6-
C), 69.94 (3-C), 68.88 (5-C), 26.64 and 26.31 (CMe2); m=z
(CI): 440.3 [M+NH4]þ. (Found: [M+NH4]þ 440.2129.
C18H34NO11 requires m=z 440.2126.)
Acknowledgements
10 (oil): ½aꢁ +19.1ꢁ (c 0.83, H2O); dH (400 MHz, D2O)
D
(major isomer) 4.88 (d, J1;2 8.2 Hz, 1-H), 4.16 (br s, 3-H),
3.95–3.86 (complex, 5-H), 3.82 (br d, J6a;6b 11 Hz, 6a-H),
3.77–3.61 (complex, 4- and 6b-H), 3.41 (br d, 2-H)); dC
(100 MHz, H2O) 94.19 (major anomer 1-C), 93.57 (minor
anomer 1-C), 73.10, 71.86, 71.81, 71.30, 67.55 (2-C to 6-
C); m=z (ES): 365.2 [M+Na]þ. (Found: [M+Na]þ
365.1059. C12H22O11Na requires m=z 365.1054.)
I thank the EPSRC Mass Spectrometry Service Centre,
Swansea for determination of the high-resolution mass
spectra, and Dr. Balaram Mukhopadhyay for helpful
discussion and for his measurement of the NMR spectra
without which this work would not have been possible.
11, mp 132–134 ꢁC: ½aꢁ )4ꢁ (c 0.45, CHCl3); dH
D
(400 MHz, CDCl3) 5.97 (1H, d, J1;2 8.7 Hz, 1-H), 5.70
(1H, dd, J2;3 3, J3;4 3 Hz, 3-H), 4.98 (1H, dd, J4;5 8.4 Hz, 4-
H), 4.97 (1H, dd, 2-H), 4.10 (1H, ddd, J5;6a 3, J5;6b 4.8 Hz,
5-H), 3.67 (1H, dd, J6a;6b 11.8 Hz, 6a-H), 3.56 (1H, dd, 6b-
H), 2.15, 2.11, 2.01, 2.00 (each 3H, CH3CO); dC (100 MHz,
References and notes
ꢀ
ꢀ
ꢀ
1. Perez, G. S.; Perez, G. R. M.; Perez, G. C.; Zavala, S. M.
A.; Vargas, S. R. Pharm. Acta Helv. 1997, 72, 105–111.
2. Unfortunately, the author has not been able to obtain a
sample of coyolosa, its peracetate, or original spectra for
comparison purposes.
CDCl3)
169.95,
169.34,
169.13,
169.09
(4 ·
–CO–), 90.06 (1-C), 72.90 (5-C), 70.22 (6-C), 68.28 (3-C),
68.03 (2- or 4-C), 65.90 (4- or 2-C), 20.76, 20.52, 20.38 (·2)
(CH3CO); m=z (CI): 696.3 [M+NH4]þ. (Found:
[M+NH4]þ 696.2348. C28H42NO19 requires m=z 696.2346.)
6. Collins, P. M.; Ferrier, R. J. Monosaccharides; Wiley:
Chichester, 1995. p 41.
3. Roberts, S. M.; Shoberu, K. A. J. Chem. Soc., Perkin
Trans. 1 1992, 2625–2632.
4. Baker, D. C.; Horton, D.; Tindall, C. G. In Methods in
Carbohydrate Chemistry; Whistler, R. L., BeMiller, J. N.,
Eds.; Academic Press: New York, 1976; Vol. 7, pp 3–6.
5. All new compounds gave consistent spectral and high
resolution mass spectrometric data. Data are given below
for selected compounds; optical rotations were measured
at 20 ꢁC and for compounds 8, 9, 10 and 11, in which the
two carbohydrate rings are related by symmetry, NMR
data are recorded for one ring moiety only.
7. Collins, P. M.; Ferrier, R. J. Monosaccharides; Wiley:
Chichester, 1995; p 535.
8. Disagreement in the mp of the two peracetates could be
explained by the seemingly unlikely occurrence of the
natural material as the 6,60-ether derived from
D
- and
allose, but the NMR spectra of such an unsymmetrical
dimer would be identical to that of the symmetrical
or -compound with the very reasonable assumption
that there is no internuclear coupling across the ether link.
L-
D,D-
L,L
4 (oil): ½aꢁ +39ꢁ (c 0.6, CHCl3); dH (300 MHz, CDCl3)
D
ꢁ
ꢁ
ꢀ
9. Farkas, J.; Sebesta, K.; Horska, K.; Samek, Z.; Dolejs, L.;
Sorm, F. Collect. Czech. Chem. Commun. 1969, 34, 1118–
ꢁ
7.72–7.62, 7.48–7.33, 7.30–7.26 (15H, 3 · m, Ar-H), 5.72
(1H, d, J1;2 3.6 Hz, 1-H), 4.67 and 4.50 (each 1H and d, JAB
12 Hz, OCH2Ph), 4.52 (1H, dd, J2;3 4.2 Hz, 2-H), 4.10 (1H,
dd, J3;4 8.7, J4;5 3.9 Hz, 4-H), 4.05 (1H, m, 5-H), 3.93 (1H,
dd, 3-H), 3.82–3.72 (2H, complex, 6a- and 6b-H), 1.57 and
1.35 (each 3H, CMe2), 1.07 (9H, CMe3); dC (75 MHz,
CDCl3) 137.55–127.81 (10 · s, Ar-C), 112.95 (CMe2),
104.10 (1-C), 77.88 (4-C), 77.74 (2-C), 77.52 (3-C), 72.05
(CH2Ph), 71.88 (5-C), 64.47 (6-C), 26.69 (CMe3, CMeMe),
26.51 (CMeMe), 19.11 (SiCMe3); m=z (CI): 566.3
[M+NH4]þ. (Found: [M+NH4]þ 566.2935. C32H44NO6Si
requires m=z 566.2932.)
ꢁ
1120.
10. Prystas, M.; Sorm, F. Collect. Czech. Chem. Commun.
ꢁ
ꢁ
1971, 36, 1448–1471.
11. Whistler, R. L.; Frowein, A. J. Org. Chem. 1961, 26, 3947–
3948.
ꢀ
12. Vanbaelinghem, L.; Gode, P.; Goethals, G.; Martin,
P.; Ronoco, G.; Villa, P. Carbohydr. Res. 1998, 311, 89–
94.
13. Rauter, A. P.; Figueiredo, J.; Ismael, M.; Canda, T.; Font,
J.; Figueredo, M. Tetrahedron: Asymmetry 2001, 12, 1131–
1146.
14. Rkagaku Konkjuse. JP 50128024, 1977; RU 9N112P,
1978.
15. Paulsen, H.; von Deyn, W. Liebigs Ann. Chem. 1987, 141–
152.
16. Vlahov, I. R.; Valhova, P. I.; Schmidt, R. R. Tetrahedron
Lett. 1992, 33, 7503–7506.
8 (oil): ½aꢁ +85.7ꢁ (c 0.43, CHCl3); dH (400 MHz, CDCl3)
D
7.38–7.17 (10H, m, Ar-H), 5.65 (1H, d, J1;2 3.3 Hz, 1-H),
4.68 and 4.51 (each 1H and d, JAB 11.8 Hz, OCH2Ph), 4.63
(2H, s, OCH2Ph), 4.44 (1H, dd, J2;3 3.7 Hz, 2-H), 4.19 (1H,
dd, J3;4 8.5, J4;5 1.6 Hz, 4-H), 4.00 (1H, dd, 3-H), 3.91 (1H,
m, 5-H), 3.54 (2H, d, J5;6a and J5;6b 6 Hz, 6a- and 6b-H),
1.56 and 1.33 (each 3H, CMe2); dC (75 MHz, CDCl3)
139.04–127.35 (8 · s, Ar-C), 112.98 (CMe2), 104.15 (1-C),
ꢀ
17. Hodosi, G.; Kovac, P. Carbohydr. Res. 1998, 308, 63–75.