Rearrangements of haloalkynol derivatives of glucofuranose

Article abstract of DOI:10.1021/ol026422q

(matrix presented) Bromoalkynol derivatives of diacetone glucose undergo rearrangements to dihaloenol ethers contained in furo[3,4-b]furan cores when treated with halonium-producing reagents.

Full text of DOI:10.1021/ol026422q

ORGANIC
LETTERS
2002
Vol. 4, No. 20
3387-3390
Rearrangements of Haloalkynol
Derivatives of Glucofuranose
Maureen Blandino and Edward McNelis*
Department of Chemistry, New York UniVersity, New York, New York 10003
edward.mcnelis@nyu.edu
Received June 25, 2002
ABSTRACT
Bromoalkynol derivatives of diacetone glucose undergo rearrangements to dihaloenol ethers contained in furo[3,4-b]furan cores when treated
with halonium-producing reagents.
Halonium-producing reagents such as N-iodosuccinimide
(NIS) and catalytic amounts of protic acids such as
p-toluenesulfonic acid (TsOH) or iodine and Lewis acids
such as [hydroxy(tosyloxy)iodo]benzene (Koser’s reagent)
have been used to convert haloalkynols to â,â-dihaloenones.1
Group shifts with significant stereospecificity are involved
and have been applied to ring expansions of representative
cyclopentyl haloalkynols to â,â-dihalocyclohexanones.2
Efforts to extend ring expansions to the conversions of
furanoses to pyranoses were undertaken but led to products
of haloetherification.3 For example, iodination of 3-iodo-
ethynyl-1,2-O-(1-methylethylidene)-5-O-methyl-R-D-pento-
furanose led not to expansion but to the formation of a furo-
[3,4-b]furan system. It arose by the unexpected attack by a
neighboring ether on the vinyl cation formed during the
reaction. Such ether cleavages are uncommon.4 The products,
with their exocyclic dihaloglycals, represent attractive starting
points for novel tamoxifen-like compounds after aryl ex-
changes of the halogens.5
and C-6 was expected to block the formation of haloethers
and lead to ring expansions to ketopyranoses. This group,
however, did not survive the halogenation conditions, and
the reactions’ main products were those of haloetherification
at the C-5 site. (Scheme 1)
Scheme 1. Preference for Ether Formation vs Ring Expansion
We report now the halogenation of haloalkynol derivatives
of diacetone glucose (1,2,5,6-di-O-(1-methylethylidene)-R-
D-glucofuranose) 1, whose acetonide protecting group at C-5
(1) Koser, G. F. Aldrichim. Acta 2001, 34, 89.
(2) Herault, X.; McNelis, E. Tetrahedron 1996, 52, 10267.
(3) Djuardi, E.; McNelis, E. Tetrahedron Lett. 1999, 40, 7193.
(4) Sonada, T.; Kobayashi, S.; Taniguchi, H. Bull. Chem. Soc. Jpn. 1976,
49, 2560.
(5) (a) Stille, J. K. Angew. Chem., Int. Ed. Engl. 1986, 25, 508. (b) Suzuki,
A. Pur. Appl. Chem. 1991, 63, 419.
Commercially available diacetone glucose was oxidized
at the C-3 position with the Dess-Martin reagent in dry
methylene chloride to afford 1,2,5,6-di-O-(1-methyleth-
10.1021/ol026422q CCC: $22.00 © 2002 American Chemical Society
Published on Web 08/31/2002

ylidene)-R-D-hex-3-ulofuranose 2 along with its hydrate in
99% yield (Scheme 2). The IR spectrum of 2 had a carbonyl
Bromoalkynyl 4 was treated with several methods for
producing iodonium ion to induce ring expansion. The
expected ketopyranose would give an M+ - 15 peak at m/z
473/475 in the mass spectrum as well as a peak at m/z 372/
374 (M+ - 116), and perhaps the molecular ion peak at m/z
488/490. After iodination with 1 equiv of NIS and a catalytic
amount of TsOH monohydrate in refluxing 15% aqueous
acetonitrile, compound 4 was converted to a white crystalline
solid with a melting point of 202-204 °C as the major
isolable product 5 in 48% yield. Loss of one isopropylidene
Scheme 2. Synthesis of 5
1
group from this compound was shown by H NMR and by
the mass spectrum with a molecular ion peak at m/z 448/
450. A large peak at m/z 332/334 in the mass spectrum
indicated the loss of 116 from m/z 448/450. Thus, the
isopropylidene group at C-1 and C-2 was still intact. In
addition, there was no peak at m/z 101, confirming the loss
of the C-5 and C-6 isopropylidene group. The complete NMR
1
spectra for 5 were as follows: H NMR (CD3OD, 300 MHz)
δ 5.91 (d, 1H, J ) 3.9 Hz), 5.10 (d, 1H, J ) 3.9 Hz), 4.65
(s, 1H), 4.39 (t, 1H, J ) 5.5 Hz), 3.72 (d, 2H, J ) 5.3 Hz),
1.58 (s, 3H), 1.41 (s, 3H) ppm; 13C NMR (CD3OD, 300
MHz) δ 159.5, 114.5, 107.5, 89.6, 87.9, 86.9, 84.0, 62.4,
27.6, 27.4, 25.5 ppm.
peak at 1725 cm-1 and a hydrate peak at 3400 cm-1. The
mass spectrum had the typical fragmentation peaks at m/z
243 (M+ - 15, from the loss of a methyl of an isopropylidene
group), 142 (M+ - 116, from the splitting off of the C-1/
C-2 isopropylidene group and the furanose oxygen), 101
(from the C5H9O2 group at C-4), and 157 (M+ - 101).
Five different structures seemed plausible for compound
5 (structures A-E, Figure 1). Structures A and B, the two
The reaction of 2 with a nucleophile is known to give rise
to a product with high stereoselectivity via attack of the
nucleophile from the top face of the ring due to the bulkiness
of the 1,2-isopropylidene group.6 Upon treatment of ketose
2 with ethynylmagnesium chloride (0.5 M in THF) in
anhydrous ether, one stereoisomer of 3-C-ethynyl-1,2,5,6-
di-O-(1-methylethylidene)-R-D-allofuranose 3 was produced
in 79% isolated yield as a crystalline solid with a melting
point of 106-107 °C. The mass spectrum of 3 had key
fragmentation peaks at m/z 269 (M+ - 15), 183 (M+ - 101),
168 (M+ - 116), and 101. The IR spectrum had a hydroxyl
peak at 3500 cm-1, an alkyne C-H stretch at 3230 cm-1,
and an alkyne C-C stretch at 2100 cm-1. A single proton
1
with a signal at 2.68 ppm in the H NMR spectrum was
indicative of the alkynyl hydrogen. All data were consistent
with those reported in the literature.7
Reaction of alkyne 3 with N-bromosuccinimide (NBS) and
silver nitrate in acetone gave 3-bromoethynyl-1,2,5,6-di-O-
(1-methylethylidene)-R-D-allofuranose 4 as a white solid with
a melting point of 101-102 °C in 84% crude yield.8 GCMS
data confirmed the addition of a bromine to and loss of a
proton from 3. Loss of the alkynyl hydrogen was noted in
Figure 1. Possible structures for compound 5.
1
the H NMR with the absence of the singlet at 2.68 ppm
and in the IR with the absence of the peak at 3230 cm-1.9
regioisomers of the ring-expanded ketopyranose with loss
of the 5,6-protecting group, were ruled out due to the lack
of carbonyl signals in the IR and 13C NMR spectra. This
confirmed that ring expansion to a ketopyranose did not
occur. Since ring expansion did not take place and the 5,6-
protecting group was lost, the possibility of ether formation
as seen with the xylose compound in Scheme 1 seemed
likely. New peaks in the 13C NMR at 159.53 ppm (meth-
ylidene carbon) and 25.5 ppm (methylidene carbon with
(6) Baker, D. C.; Brown, D. K.; Horton, D.; Nickol, R. Carbohydr. Res.
1974, 32, 299.
(7) Kakinuma, K.; Iihama, Y.; Takagi, I.; Ozawa, K.; Yamauchi, N.;
Imamura, N.; Esumi, Y.; Uramoto, M. Tetrahedron 1992, 48, 3763.
(8) Hofmeister, H.; Annen, K.; Laurent, H.; Wiechert, R. Angew. Chem.,
Int. Ed. Engl. 1984, 23, 727.
(9) Structures of all new compounds described herein agreed with
elemental analyses provided by Schwarzkopf Microanalytical Laboratories,
Woodside, NY.
3388
Org. Lett., Vol. 4, No. 20, 2002

iodine and bromine) supported either structures D or E as
the product and ruled out compound C.10
Analysis of the H NMR data along with examination of
iodine. Thus, 5 is assigned as structure D, 4-(E)-bro-
moiodomethylene-(cis)-3a-hydroxy-6-(cis)-hydroxy methyl-
2,3-O-(1-methylethylidene)-2,3,3a,4,6,6a-hexahydrofuro[3,4-
b]furan.
The dibromo analogue of 5, namely, 6, was prepared
directly from 3 in 50% crude yield by using 2.4 equiv of
NBS with catalytic AgNO3 in acetone (Scheme 4). Com-
1
molecular models supported structure D over E. The chiral
center at C-5 had an absolute configuration of R since it was
R in starting material 1, and there was no evidence for bond
breaking at C-5 during the intermediate steps. With C-5 as
R, molecular models showed that if the hydroxyl group at
C-6 attacked the vinyl cation to form the six-membered ring
(structure E), hydrogens H-4 and H-5 would be syn. (Scheme
3) If the hydroxyl group at C-5 attacked the vinyl cation to
Scheme 4. Synthesis of 6
Scheme 3. Preference for Structure D
pound 6 was also observed as a side-product in the formation
of 4 (Scheme 2). Compound 6 contained two bromine atoms
with the loss of one protecting group as shown by key
fragments in the mass spectrum at m/z 400/402/404 (M+),
385/387/389 (M+ - 15), 284/286/288 (M+ - 116), and 253/
255/257 (M+ - 116 - 31). The proton NMR for 6 was
nearly identical to that of 5, as was the carbon NMR with
the major difference being the dibromo methylidene carbon’s
signal at 68.2 ppm vs the bromoiodo methylidene carbon of
5’s signal at 25.5 ppm. The NMR spectra for 6 were as
follows: 1H NMR (CD3OD, 300 MHz) δ 5.91 (d, 1H, J )
3.9 Hz), 5.09 (d, 1H, J ) 3.9 Hz), 4.62 (s, 1H), 4.43 (t, 1H,
J ) 5.6 Hz), 3.73 (d, 2H, J ) 5.3 Hz), 1.57 (s, 3H), 1.40 (s,
3H) ppm; 13C NMR (CD3OD, 300 MHz) δ 158.9, 114.7,
107.7, 89.5, 88.1, 87.7, 83.1, 68.2, 62.5, 27.7, 27.5 ppm.
Dibromo 6 was also derived from 7 (3 without its 5,6-
isopropylidene group) in 74% yield with 2.2 equiv of NBS.
Compound 7, mp 97-98 °C, was prepared from 3 by acetic
acid treatment in 94% yield.13 In addition to a peak at m/z
229 (M+ - 15), the mass spectrum of 7 displayed a major
peak at 98 corresponding to losses of 116 and 30. The
reaction from 7 to 6 confirms that the loss of the protecting
group at C-5 and C-6 of 3 is required to form 6 as well as
5.
form the five-membered ring (structure D), H-4 and H-5
would be anti to one another and C-6 would be pseudoequa-
torial. In the H NMR of 5, H-4 appeared as a singlet. H-4
1
could only appear as a singlet if H-5 were anti to it, as in
structure D. This effect was noted in the pentofuranose case3
and by the Tsuchiya group for this type of heterocycle, as
well as by the Yoshimura and Baker groups.6,11
Further support for structure D came from the fragmenta-
tion pattern in the mass spectrum. The base peak of m/z 301/
303 occurs from a splitting off of 116 from the parent m/z
448/450 and then further cleavage of 31. The loss of 31 likely
comes from protonated formaldehyde, which can be cleaved
from structure D, but not structure E.
On the basis of iodonium-induced shifts, it is most likely
that the stereochemistry of bromoiodoenol 5 is E.12 Those
findings were consistent with the shift anti to an iodonium
bridging effect on the alkynyl groups. By extension, attack
by an alcohol to form an enol ether should be anti to the
(10) An alkylidene carbon bearing two halogens has an upfield peak in
the 13C NMR spectrum. (See ref 12 below and: Brunel, Y.; Rousseau, G.
Tetrahedron Lett. 1995, 36, 2619.)
(11) (a) Tsuchiya, T.; Aito, K.; Umezawa, S.; Ikeda, A. Carbohydr. Res.
1984, 126, 45. (b) Yoshimura, J.; Kobayashi, K.; Sato, K.; Funabashi, M.
Bull. Chem. Soc. Jpn. 1972, 45, 1806.
(12) (a) Bovonsombat, P.; McNelis, E. Tetrahedron Lett. 1994, 35, 6431;
(b) Tetrahedron 1993, 49, 1525; (c) Tetrahedron Lett. 1993, 34, 4277. (d)
Djuardi, E.; Bovonsombat, P.; McNelis, E. Tetrahedron 1994, 50, 11793.
(e) Bovonsombat, P.; McNelis, E. Synth. Commun. 1995, 25, 1223.
(13) Qureshi, S.; Shaw, G. J. Chem. Soc., Perkin Trans. 1 1985, 875.
NMR data for 7: 1H NMR (CD3OD, 300 MHz) δ 5.76 (d, 1H, J ) 3.58
Hz), 4.56 (d, 1H, J ) 3.58 Hz), 3.95 (dd, 1H, J ) 5.63 Hz, J ) 2.60 Hz),
3.90 (d, 1H, J ) 8.47 Hz), 3.78 (dd, 1H, J ) 11.58 Hz, J ) 2.29 Hz), 3.59
(dd, 1H, J ) 11.62 Hz, J ) 5.60 Hz), 3.16 (s, 1H), 1.54 (s, 3H), 1.35 (s,
3H) ppm; 13C NMR (CD3OD, 300 MHz) δ 114.8, 105.4, 86.4, 83.0, 80.8,
78.53, 77.0, 73.4, 65.5, 27.4, 27.3 ppm.
Org. Lett., Vol. 4, No. 20, 2002
3389

Thus, the ether formation does occur with preference for
a possible ring expansion. The latter reaction may be
inhibited by the poor migrating aptitude of carbons bearing
oxygen atoms, although there are reports of ring expansions
of glycosides and xylofuranoses with diazomethane in a
methanol/ether (3:1, v/v) mixture.14 Since the shift reactions
of haloalkynols proceed best in acidic aqueous solvents,
hydrolysis of the 5,6-isopropylidene masking group proceeds
prior to haloetherification. Efforts to minimize or eliminate
aqueous acid in the reaction to promote ring expansion in
these systems is being tested, as well as examination of acid
insensitive protecting groups at the C-5 position.
Acknowledgment. M.B. thanks the Graduate School of
Arts and Sciences of New York University for a Kramer
Fellowship.
Supporting Information Available: Experimental pro-
cedures for the preparations of compounds 4-6. This
material is available free of charge via the Internet at
http://pubs.acs.org.
(14) (a) Flaherty, B.; Overend, W. G.; Willams, N. R. Chem. Ind. 1966,
434. (b) Nahar, S.; Overend, W. G.; Williams, N. R. Chem. Ind. 1967,
2114.
OL026422Q
3390
Org. Lett., Vol. 4, No. 20, 2002