Aldol reactions on 1-deoxy-3,4:5,6-di-O-isopropylidene-L-fructose as a route to higher-carbon carbohydrates

Article abstract of DOI:10.1016/S0008-6215(99)00191-3

With a view to preparing higher-carbon carbohydrates, crossed-aldol reactions of the methyl ketone 1-deoxy-3,4:5,6-di-O-isopropylidene-L-fructose with a representative series of aldehydes have been investigated, and the feasibility has been demonstrated o

Full text of DOI:10.1016/S0008-6215(99)00191-3

www.elsevier.nl/locate/carres
Carbohydrate Research 321 (1999) 197–213
Aldol reactions on
1
-deoxy-3,4:5,6-di-O-isopropylidene- -fructose as a route
L
to higher-carbon carbohydrates
Alan H. Haines *, Andrew J. Lamb
School of Chemical Sciences, Uni6ersity of East Anglia, Norwich NR4 7TJ, UK
Received 28 May 1999; accepted 24 July 1999
Abstract
With a view to preparing higher-carbon carbohydrates, crossed-aldol reactions of the methyl ketone 1-deoxy-
,4:5,6-di-O-isopropylidene- -fructose with a representative series of aldehydes have been investigated, and the
3
L
feasibility has been demonstrated of constructing a C-11 unit containing some of the key functionality found in the
carbohydrate component of the herbicidins. © 1999 Elsevier Science Ltd. All rights reserved.
Keywords: Aldol reaction; Higher-carbon sugars; Herbicidins
1
. Introduction
sugar portion. Because of their potential as
selective herbicides, compounds of this class
are interesting synthetic targets and several
approaches towards their synthesis have been
reported [4]. Studies into the synthesis of the
carbohydrate moiety have largely centred on
the C-6 plus C-5 approach, and Newcombe
and co-workers [4a] have successfully con-
structed the 11-carbon skeleton 2 with correct
stereochemistry in an elegant approach based
on a C-6 carbohydrate-derived enolate as the
donor, and a C-5 aldehyde as the acceptor.
However, construction of the enolate precur-
sor, a bicyclic ketone, required considerable
manipulation and we were led to consider an
alternative C-6 plus C-5 strategy based on the
retrosynthesis shown in Scheme 1, involving a
straight-chain C-6 carbohydrate-derived ke-
tone 3 containing the functionality MeCOꢀR
and the same aldehyde 4 as used by New-
combe and co-workers, with formation of the
C-ring occurring at a late stage in the synthe-
sis through an alkoxide-based ring closure.
The ketone 3 of our retrosynthesis is a pro-
The synthesis of monosaccharides possess-
ing more than six carbon atoms, usually re-
ferred to as higher-carbon sugars, continues to
be of considerable interest. In addition to
widely distributed members of this class such
as the octulosonic acid KDO and the
aminononulosonic acids (sialic acids), there
are others which are constituents of natural
products possessing interesting biological
properties, such as the mildiomycins [1] and
tunicamycins [2], nucleoside antibiotics con-
taining carbohydrate moieties of 10 and 11
carbon atoms, respectively. Also of particular
interest are the herbicidins [3], of which herbi-
cidin C (1) is a particular example, a group of
nucleoside antibiotics exhibiting herbicidal ac-
tivity that were isolated from Streptomyces
saganonensis and which contain an 11-carbon
*
Corresponding author. Tel.: +44-1603-593-133; fax: +
4
4-1603-592-015.
E-mail address: a.haines@uea.ac.uk (A.H. Haines)
0008-6215/99/$ - see front matter © 1999 Elsevier Science Ltd. All rights reserved.
PII: S0008-6215(99)00191-3

198
A.H. Haines, A.J. Lamb / Carbohydrate Research 321 (1999) 197–213
Scheme 1. A retrosynthetic analysis for the herbicidin skeleton.
tected acyclic form of 1-deoxy-
D
-fructose and
ulose as an oil in 30% yield with a 32%
recovery of starting ketone 5. The yield is not
surprising in view of the known low concen-
tration of dimeric product formed in aldol
reactions performed with two molecules of the
same ketone.
indeed subsequent to the start of our work,
Narkunan and Nagarajan [5] published their
results, which employed a related approach to
the synthesis of higher sugars including a Wit-
tig–Horner reaction between a protected 1-
phosphonate of 1-deoxy-
D
-fructose, prepared
-arabinose,
To develop suitable conditions for forma-
tion of crossed-aldol products, reaction of 5
with benzaldehyde 8 was investigated under
three sets of conditions. Addition of freshly
distilled benzaldehyde to the lithium enolate
(from LDA) of 5 in THF solution led to
complete reaction from TLC evidence, but
product isolation by column chromatography
gave a significant amount (21%) of 5, pre-
sumably due to retro-aldol chemistry, a very
small amount of material tentatively identified
as the self-addition product (6 or 7), and an
oil that crystallised on standing and was re-
crystallised to give a sharp-melting mixture
of the diastereoisomeric crossed-aldol prod-
ucts 1-(R)- and 1-(S)-2-deoxy-4,5:6,7-di-O-
from 2,3:4,5-di-O-isopropylidene-
D
and aldehyde 4. Because of the ready
availability in three simple steps of the enan-
tiomeric 1-deoxy- -fructose in a suitable pro-
L
tected form 5 from commercially available
-rhamnose, we based our feasibility studies
on this -isomer and now report results of
L
L
aldol reactions (Claisen–Schmidt condensa-
tions) between this ketone and a series of
aldehydes including 4 (Scheme 2).
2
. Results and discussion
Because in crossed-aldol reactions of the
type envisaged there is the likelihood of self-
addition of components, a self-aldol addition
on 5 was first attempted, which was expected
to give a mixture of undec-5-uloses 6 and 7.
Aqueous sodium hydroxide, potassium t-bu-
toxide in tetrahydrofuran (THF) and 1,8-diaz-
abicyclo[5.4.0]undec-7-ene (DBU) in toluene
were unsuccessful in bringing about the de-
sired reaction, but reaction was achieved with
lithium diisopropylamide (LDA) in THF. Re-
action of 1 molar equivalent of LDA with 2
molar equivalents of 5 gave one stereoisomer,
based on NMR spectroscopy, of the undec-5-
isopropylidene-1-C-phenyl- -arabino-hept-3-ul-
L
ose (9 and 10, respectively, 10%) in an approx-
imately 1:1 ratio. This allocation of overall
1
structure was supported by the H NMR spec-
trum of the product, which showed signals for
protons in isopropylidene methyl groups and
the phenyl groups in the ratio of 12:5 and
signals for the isopropylidene methyl groups
1
13
in both H and C NMR spectra that were
attributable to the expected eight groups. In
addition, the AB portions of two ABX sys-
tems could be identified centred on l 3.08 and
l 3.09, arising in each case from the

A.H. Haines, A.J. Lamb / Carbohydrate Research 321 (1999) 197–213
199
Scheme 2.
diastereoisotopic methylene protons at C-2
coupled to the proton at the adjacent chiral
centre at C-1. Analytical HPLC confirmed the
presence of two isomers, though lack of a
baseline separation prevented determination
of an accurate isomer ratio. Use of lithium
hexamethyldisilazide (LiHMDS) as the base in
THF resulted in an improved yield (67%) of
the combined stereoisomers 9 and 10, again in
an approximate 1:1 ratio, with only 5% recov-
ered starting ketone 5 and no self-reaction
product. In the third variation, the procedure
involving the use of boron enolates [6] for
aldol reactions was followed. Thus, ketone 5
in THF was added to dibutylboron triflate
and triethylamine in dichloromethane fol-
lowed by benzaldehyde. Chromatography af-
forded the combined diastereoisomers 9 and
10 in 43% yield, with an unassigned isomer
1
ratio, estimated by H NMR spectroscopy in
the presence of a europium shift reagent, of
7:3.
The crossed-aldol reaction between ketone 5
and
2,3-O-isopropylidene- -glyceraldehyde
D
(13) was performed using enolates prepared
from 5 with LiHMDS and with dibutylboron
triflate. From reaction of the lithium enolate
with a slight excess of aldehyde 13, following
chromatography, starting ketone 5 (12%), the
self-addition product 6 or 7 (19%) and an oily

200
A.H. Haines, A.J. Lamb / Carbohydrate Research 321 (1999) 197–213
cross-reaction product as a mixture of the
5 with aldehyde 4 gave after chromatography
a mixture of the two diastereoisomers in 19%
yield but also a recovery of starting ketone in
29% yield. The ratio of isomers in the conden-
diasteroisomeric non-5-uloses 14 and 15 (48%)
1
were isolated. The H NMR spectrum of the
mixture of diastereoisomers, with a consider-
ation of relative intensities because of some
signal overlap, showed 12 methyl signals and
the AB parts (H-4 and H-4%) of two ABX
systems (H-3, H-4 and H-4%) were discernible,
centred on l 2.90 and l 2.91. Comparative
integration of the isopropylidene methyl
groups indicated the ratio of the two unas-
signed diastereoisomers to be approximately
1
sation product was 1.5:1 from H NMR inte-
gration measurements, and results from a
subsequent reaction via the sodium enolate
suggested that the predominant stereoisomer
in the mixture of 23 and 24 was the 5-(R)-iso-
mer 24 (Scheme 3).
Reaction of aldehyde 4 with the sodium
enolate of ketone 5, prepared in THF solution
by reaction of the ketone with sodium hexam-
ethyldisilazide (NaHMDS), gave recovered
ketone 5 in 22% yield and, surprisingly, a
single diastereoisomer (based on NMR spec-
troscopic and HPLC evidence) in 53% yield,
which was tentatively identified on mechanis-
1
3
1
.5:1. The C NMR spectrum of the mixture
contained peaks that could readily be assigned
to the 36 distinct carbon atoms expected for
the mixture of isomers. Reaction of the ketone
5
and aldehyde 13 using the boron enolate
procedure gave the same diastereoisomers in
an approximately 1:1 ratio in 35% yield with
tic argument as the
L
-arabino-a- -gluco- iso-
D
1
0% starting ketone 5 and 9% of the self-addi-
mer. Thus, on the reasonable assumption that
a sodium ion might favour a metal ion-coordi-
nated transition state involving the oxygen of
the 3-O-benzyl group, the oxygen atom of the
aldehydic group at C-5 in 4, and the enolate
oxygen atom, two possible intermediates A
and B may be envisaged (Fig. 1), which differ
in disposition of the di-O-isopropylidene-con-
taining residue involving C-3 to C-6 in the
parent ketone 5. In structure A, this residue
occupies a position lying over the furanose
ring of the reactant aldehyde 4, whereas in
structure B, this residue is directed away from
this region in space and thus may be favoured.
If this is so, then the preferred stereoisomer
will have the R-configuration at the new chiral
centre and the predominant isomer will be
3-O-benzyl-6-deoxy-1,2:8,9:10,11-tri-O-iso-
tion products 6 and 7.
Reaction between ketone 5 and 2,3:4,5-di-
O-isopropylidene- -arabinose (18) via the
D
lithium enolate of 5 gave, after chromatogra-
phy, 12% of 5, 13% of a mixture of 6 and 7
and the desired crossed-aldol product in 49%
yield as a mixture of the diastereoisomeric
undec-5-uloses 19 and 20. Analytical HPLC
indicated an unassigned isomer ratio of 1.2:1.
However, separation of the two stereoisomers
1
on an analytical HPLC column enabled an H
NMR spectrum to be obtained on each one
and allowed analysis of the individual ABX
systems. Performing the crossed-aldol reaction
on 5 and 18 via the boron enolate procedure
gave 19 and 20 in a combined yield of 80%
but with a reversed isomer ratio of 1:1.6,
indicating a selectivity for the isomer that had
been the minor constituent in the alternative
preparation.
propylidene-
L
-arabino-a- -gluco-undecofura-
D
nos-7-ulose (24). This procedure via the
sodium enolate was used in all subsequent
reactions between 4 and 5.
Reaction between ketone 5 and the C-5
aldehyde 3-O-benzyl-1,2-O-isopropylidene-a-
The route we envisaged for the transforma-
tion of 23 or 24 to the tricyclic skeleton of the
herbicidins involved first, elimination of the
hydroxy group at the newly formed chiral
centre to give an alkene 30 or 31 with subse-
quent O-10 to C-6 ring closure leading to
formation of ring C of the herbicidins. Such a
closure might be achieved through alkene
epoxidation followed by selective removal of
the acetal group at O-10 and O-11 and then
D-xylo-pentodialdo-1,4-furanose (4) via the
lithium enolate of 5 gave the self-addition
products 6 and 7 (15%) and a 25% yield of the
crossed-aldol products, the undecofuranos-7-
uloses 23 and 24, in a 1:1 ratio as indicated by
integration of signals for the anomeric hydro-
gens. Careful column chromatography of the
mixture afforded a sample of one diastereoiso-
mer. Reaction of the boron enolate of ketone

A.H. Haines, A.J. Lamb / Carbohydrate Research 321 (1999) 197–213
201
Scheme 3.
an alkoxide-mediated attack on the epoxide
involving O-10. Alternatively, in a more ambi-
tious approach, electrophilic attack at the
alkene might lead directly to formation of the
six-membered ring by nucleophilic attack of
O-10 at C-6 in which electrophilic character
had been generated, as for example in an
iodonium intermediate resulting from reaction
of the alkene with N-iodosuccinimide. In this
approach, prior cleavage of the 10,11-acetal is
not required; examples are known [7] where
just such functionality is involved in ring clo-
sure. Reliable methods exist for the removal
of the group introduced at C-5 during ring
closure, for example, hydroxy or iodo, and
control of the stereochemistry at C-6 may be
possible in view of the carbonyl group in the
a-position (C-7). Deprotection at O-3 should
then lead to spontaneous formation of ring B
and completion of the tricylic skeleton found
in herbicidins. In view of the central position
of alkene 30 or 31 in the proposed synthetic
sequence, elimination reactions were investi-
gated on the mixed-aldol products 9 and 10,
14 and 15, 19 and 20, 23 and 24, and the
single stereoisomer thought to be 24, in order
to develop the most favourable reaction
conditions.
Treatment of the isomer mixture 9 and 10
with acetic anhydride–pyridine led to produc-
tion of the (E)-hept-1-en-3-ulose 27 (50%) in
addition to the expected acetates 11 and 12
Fig. 1. Possible intermediates in the aldol reaction between
sodium enolate of ketone 5 and aldehyde 4.

202
A.H. Haines, A.J. Lamb / Carbohydrate Research 321 (1999) 197–213
(
identified spectroscopically), the proportion
Our synthetic strategy involving elaboration
of 30 or 31 to form the C-ring of the herbi-
cidins via alkene epoxidation was thwarted
by an alternative sequence of reactions under-
gone by the epoxide. Thus, treatment of the
enone 30 with alkaline hydrogen peroxide
reagents in methanol gave a mixture of five
products, as indicated by methoxy resonances
of 27 increasing with time. The structure of
the alkene was supported by elemental analy-
sis and its NMR parameters, with the E-
configuration indicated by the 16.2 Hz
coupling constant between the alkene protons.
Acetylation of the mixed stereoisomers 14
and 15 under the same conditions gave the
mixed 3-acetates 16 and 17 (6%) and an elimi-
nation product, the (E)-non-3-en-5-ulose (28)
in 39% yield, the E-configuration being confi-
rmed by the large coupling constant for the
alkenic protons of 15.5 Hz.
1
in the H NMR spectrum, and these pro-
ducts were only partially separable by chro-
matography. NMR spectroscopic data sug-
gested that the products arose from the
required epoxide (or epoxides), four being
stereoisomers (see 32 and 33) resulting from
methoxide cleavage of epoxides (MeO reso-
nances at l 3.22, 3.24, 3.27, and 3.28) and one
being a methyl ester (see 34 and 35) (MeO
resonance at l 3.71) resulting from a Fa-
vorskii rearrangement of the epoxide. Exam-
ples of the latter type of rearrangement in
such systems have been documented previ-
ously [8]. Use of a milder base (K CO ), af-
Acetylation of the mixed stereoisomers 19
and 20 followed a similar path affording a
mixture of the expected 7-acetates 21 and 22
(
19%), as identified by their spectroscopic
properties, and an elimination product, the
E)-undec-6-en-5-ulose 29 (13%) (J6,7 15.5
(
Hz), as well as a considerable amount of
recovered starting materials. An elimination
reaction on the mixture of 19 and 20 using
p-tolylsulphonyl chloride in pyridine gave the
enone 29 in an improved yield of 27% and
recovered starting materials in 29% yield.
Dehydration of the mixed aldol adducts 23
and 24 was investigated under a variety of
conditions in view of the likely importance of
the elimination product for further transfor-
mations. Reaction of the mixed alcohols with
acetic anhydride–pyridine at room tempera-
ture gave after 6 days and chromatographic
separation the starting alcohols (16%), the
diastereoisomeric acetates 25 and 26 (10%) as
identified by NMR spectroscopy, and the (E)-
undec-5-enofuranos-7-ulose (30) (J5,6 15.8 Hz)
in 46% yield. TLC analysis of the progress of
the reaction indicated that the alkene 30 was
formed via the acetates, verification being ob-
tained when a sample of the mixed acetates
was converted to the alkene 30 when subject
to the original acetylation conditions. The
yield of 30 was raised to 61% by conducting
the acetylation at 45 °C for 8 h, and to 72%
when 4-dimethylaminopyridine (DMAP) was
present in the acetylation medium with reac-
tion at room temperature for 70 h, but in the
latter case the chromatographically recovered
alkene was found to contain approximately
2
3
forded the methoxy derivatives and no
rearranged product. Reaction in THF with
H O –NaHCO , a reportedly mild procedure
2
2
3
for the epoxidation of enones [9], gave a
1
product that was tentatively identified by H
NMR and IR spectroscopy was tentatively as
a carboxylic acid (rather than a methyl ester),
resulting presumably from a Favorskii rear-
rangement of an epoxide.
An attempt to bring about ring closure on
0 by an iodonium-ion-mediated cyclisation
3
was not successful; only starting material was
recovered on prolonged treatment of the
alkene with N-iodosuccinimide in dichloro-
methane or with molecular iodine in THF.
In view of the difficulties encountered in
forming ring C at an early stage, prior forma-
tion of ring B was investigated. Elaboration of
a fused furano–pyrano ring system from 23 or
2
4 to mimic the A/B fused ring system found
in the herbicidins requires removal of the ben-
zyl protecting group, which should lead to
spontaneous hemiacetal formation and thus
the pyrano ring B. Hydrogenolysis of the ben-
zyl group in 24 was readily achieved using
palladium black as catalyst in acidic methanol
to give 36 as a crystalline solid in high yield.
No carbonyl absorption was apparent in the
IR spectrum of this product, in contrast to the
5
% of the Z-isomer 31 (J5,6 11.9 Hz).

A.H. Haines, A.J. Lamb / Carbohydrate Research 321 (1999) 197–213
203
−
1
absorption at 1720 cm found in the starting
used on a Spherisorb S5W normal phase
column. Thin-layer chromatography (TLC)
was performed on pre-coated plates of silica
gel with fluorescent indicator (Machery–
Nagel SIL G UV . Detection was either by
13
material. A signal in the C NMR spectrum
at 96.2 ppm was assigned to the carbon at the
newly created hemiacetal centre (C-7), the re-
action evidently affording a single isomer at
this centre. Thus, the number of peaks in the
2
54
viewing under UV light (254 nm), or by spray-
ing with a 10% H SO , 1.5% molybdic acid,
1
13
H and C NMR spectra was consistent with
2
4
only one isomer. The axial orientation of the
hydroxy group at C-7 and therefore also the
equatorial orientation of the C-8 to C-11 side
chain might reasonably be expected to be
more favoured rather than the reverse
arrangement.
Although the eventual aim of this project
has still to be achieved, current results indicate
that our approach leads to facile formation of
an undecose in a usefully functionalised form
and that closure to give ring B is not problem-
atical. The crucial step of ring C formation is
being pursued by reaction of enone 30 with
alternative more powerful electrophiles and by
a strategy involving cyclisation of the 10,11-
diol formed by the partial deprotection of 30.
1% ceric sulfate spray followed by heating to
150 °C. Column chromatography was per-
formed on Matrex Silica 60 (70–200 mm mesh,
Fisons) or Kieselgel 60 (70–230 mm mesh, E.
Merck). Solvents for reactions and column
chromatography were SLR grade or better
and were dried appropriately when required
anhydrous. Thus, THF was dried by storage
over and distillation from sodium wire;
CH
hydrous by distilling from CaH
molecular sieves; MeOH was dried
Cl
, pyridine and Et
N were obtained an-
2
2
3
and storage
2
over 4 A
by distillation from magnesium methoxide
formed by reaction with activated magne-
sium) and stored over powdered 3 A molecu-
,
(
,
lar sieves. Light petroleum refers to the
fraction with a boiling range of 40–60 °C,
unless stated otherwise. When mixed solvents
were used, the ratios given are v/v except if
otherwise stated. Solvent A is 6:1 light
petroleum–EtOAc, and solvent B is 4:1 light
petroleum–EtOAc. Organic solutions were
3
. Experimental
General methods.— H NMR spectra were
1
recorded on a Jeol EX90 FT spectrometer, a
Jeol EX270 FT spectrometer or a Varian
dried over anhydrous Na SO .
L
-Rhamnitol
2
4
was prepared from
L
-rhamnose by adapting
Gemini 2000 FT spectrometer in CDCl with
3
the procedure described by Wolfrom and
Thompson for the preparation of galactitol
Me Si as internal standard. Where appropri-
4
ate, signal assignments were deduced by
DEPT, COSY and HETCOR NMR experi-
from
into 1,2:3,4-di-O-isopropylidene-
as described by Bukhari and co-workers [11].
2,3-O-Isopropylidene- -glyceraldehyde (13)
was prepared from 1,2:5,6-di-O-isopropyli-
dene- -mannitol [12,13] by periodate cleavage
according to the procedure of Jackson [14].
D
-galactose [10] and was then converted
L
-rhamnitol
1
ments. In H NMR spectra of products con-
taining two diastereoisomers, protons at
related centres in the isomers that give rise to
resolved and assignable signals are distin-
guished by subscripts ‘a’ and ‘b’, otherwise no
distinction is made. Optical rotations were
measured at 20 °C with a Perkin–Elmer 141
polarimeter. High-resolution mass spectra
were recorded by the EPSRC Mass Spec-
trometry Service at the University College of
Swansea. Low-resolution EI mass spectra and
elemental analyses were performed by A.W.R.
Saunders at the University of East Anglia.
HPLC was performed on Pye Unicam appara-
tus using a PU 4015 pump connected to a PU
D
D
2,3:4,5-Di-O-isopropylidene- -arabinose (18)
D
was synthesised by the method of Inch and
co-workers [15] from 1,2:3,4-di-O-isopropyli-
dene-
D
-mannitol [16]. 3-O-Benzyl-1,2-O-iso-
propylidene-a-
D
-xylo-pentodialdo-1,4-fura-
nose (4) was prepared either by oxidative
cleavage of the 5,6-diol in 3-O-benzyl-1,2-O-
isopropylidene-a- -glucofuranose [17], as de-
D
scribed by Anderson and Fraser-Reid [18], or
by the method of Lemieux and Howard [19],
directly from 3-O-benzyl-1,2:5,6-di-O-iso-
4
025 UV detector operating at 300 nm unless
otherwise stated. HPLC grade solvents were
propylidene-a- -glucofuranose [20].
D

204
A.H. Haines, A.J. Lamb / Carbohydrate Research 321 (1999) 197–213
1
-Deoxy-3,4:5,6-di-O-isopropylidene-
L
-fruc-
(b) Method 2: using dibutylboron triflate/
3
tose (5).—To a stirred soln of oxalyl chloride
Et N. The procedure for the crossed-aldol re-
(
2 mL, 20.8 mmol) in anhyd CH Cl (40 mL)
action via the boron enolate was essentially
that described by Paterson and co-workers [6]
but using dibutylboron triflate in place of
dicyclohexylboron chloride. A soln of 5 (1
molar equiv) in anhyd THF was added, drop-
wise, to a stirred soln of dibutylboron triflate
(1.0 M soln in CH Cl ; 1.5 molar equiv) and
2
2
at −50 °C under N , was added an anhyd
2
soln of Me SO (3.5 mL, 48.9 mmol) in CH Cl
2
2
2
(
−
10 mL) and the solution was then cooled to
78 °C. A soln of 1,2:3,4-di-O-isopropyli-
dene- -rhamnitol (4.88 g, 20.0 mmol) in an-
hyd CH Cl (20 mL) was then added dropwise
L
2
2
2
2
over 10 min. After 15 min, Et N (14.0 mL,
.10 mol) was added and after a further 5 min,
3
freshly distilled Et N (8–10 molar equiv) in
3
0
anhyd THF (3 mL/mmol of boron reagent) at
the stirred reaction mixture was allowed to
warm to room temperature (rt) when water
−
78 °C under N . The soln was allowed to
2
warm to 0 °C, stirred at this temperature for
5 min and then re-cooled to −78 °C. Freshly
(
100 mL) was added. The organic and
4
aqueous layers were separated, the aqueous
distilled aldehyde (1.2–3 molar equiv) was
added either neat or as a concd soln in anhy-
drous THF. After a further 30 min at
layer was extracted with CH Cl (2×100 mL)
2
2
and the combined organic layers were washed
consecutively with satd aq NaCl (200 mL), 0.1
M HCl (200 mL), aq Na CO (5% w/v; 200
−
78 °C, the mixture was stored at −30 °C
2
3
for 14 h then brought to rt and partitioned
mL) and water (50 mL). The dried organic
solution was concentrated to give a pale yel-
low oil (5.0 g). Column chromatography (sol-
vent A) gave ketone 5 as a mobile liquid (4.17
g, 85%); [h] −0.10° (c 1.2, CHCl ), lit. ꢀ0°
between equal volumes (10 mL/mmol of 5) of
pH 7 buffer solution and Et O. The aq layer
2
was extracted further with Et O and the com-
2
bined organic layers were concd to give a
crude product which was dissolved in equal
volumes (6 mL/mmol of 5) of MeOH and pH
D
3
(
[
(
(
1
c 4.2, CHCl ) [15] and lit. 0.0° (c 1.5, CHCl )
3
3
21]; IR (film): w 1730 (CꢁO), 1370 and 1385
7
buffer soln. To this soln was cautiously
−1
C(CH ) ) , no absorption near 3300 cm
1
3
2
added 30% aq H O (3 mL/mmol of 5). The
2
2
OH); H NMR (90 MHz): l 1.34, 1.36, 1.40,
.45 (4×s, 4×3 H, 2×C(CH ) ), 2.29 (s, 3
resulting mixture was stirred until TLC (light
petroleum–EtOAc mixture) indicated the ab-
sence of the boron-coordinated products,
which remained on or near the baseline on
TLC in such solvent systems employed; typi-
cally between 1 and 3 h were required. Extrac-
tion of the solution with CH Cl (3×8 mL
3
2
H, H C-1), 3.92–4.38 (complex, 5 H, H-3, 4,
3
1
3
5
2
6
, 6, 6%); C NMR (67.9 MHz): l 24.9 (C-1),
6.0, 26.3, 26.5, 26.9 (2×C(CH ) ), 66.4 (C-
3
2
), 76.3, 77.9 (C-4, 5), 83.1 (C-3), 109.7, 111.1
(
2×C(CH ) ), 207.2 (CꢁO). Anal. Calcd for
3
2
2
2
C H O : C, 59.0; H, 8.25. Found: C, 59.2; H,
1
2
20
5
for 1 mmol of 5) and concentration of the
dried organic soln gave crude material that
was purified by column chromatography (light
petroleum–EtOAc mixture).
Where alternative procedures to (a) or (b)
have been used, the methods are described in
full.
8
.4.
General procedures for aldol reactions on
5
.—(a) Method 1: using LiHMDS. A solution
of 5 (1 molar equivalent) in anhyd THF was
added to LiHMDS (1.0 M solution in THF;
1
.3 molar equiv) at −78 °C under N , and the
2
reaction mixture was stirred at −78 °C for 20
min. Freshly distilled aldehyde (1.1–1.3 molar
equiv) was added either neat, in one portion,
or as a concd soln in anhyd THF. After 20
min at −78 °C, the mixture was allowed to
Self-addition of ketone 5 to gi6e 6-deoxy-
,2:3,4:8,9:10,11-tetra-O-isopropylidene-7-C-
1
methyl-
and 6-deoxy-1,2:3,4:8,9:10,11-tetra-O-isopro-
pylidene-7-C-methyl- -arabino- -gulo-undec-
L
-arabino-
L
-manno-undec-5-ulose (6)
warm to rt, satd aq NaHCO was added, and
L
D
3
the solution was extracted repeatedly with
5-ulose (7).—To a stirred soln of diisopropyl-
amine (0.17 mL, 0.12 g, 1.2 mmol) in anhyd
THF (5 mL) at 0 °C under nitrogen was
added dropwise n-butyllithium (2.5 M soln in
hexane; 0.5 mL, 1.25 mmol). After 10
Et O. The combined extracts were dried, con-
2
centrated, and the resulting oil was purified by
column chromatography (light petroleum–
EtOAc mixture).

A.H. Haines, A.J. Lamb / Carbohydrate Research 321 (1999) 197–213
205
min at 0 °C, the soln was cooled to −78 °C
and a soln of 5 (0.50 g, 2.05 mmol) in anhyd
THF (5 mL) was added dropwise in two ap-
proximately equal portions, at such a rate that
the temperature remained below −65 °C. Af-
ter a further 30 min stirring at −78 °C, satd
and 0.11). The reaction was quenched with
water (25 mL), extracted with CH Cl (3×25
2
2
mL) and the combined organic extracts were
dried and concentrated to give an oil (1.37 g),
which on column chromatography (solvent A)
gave three components. The first, ketone 5
(0.21 g, 21%) presumably arose by a retro-al-
dol reaction on the silica column. The second
aq NaHCO (10 mL) was added and the
3
mixture was allowed to warm to rt. Extraction
of the aq soln with Et O (3×15 mL) and
material (R 0.20; 4 mg) was tentatively iden-
2
f
concentration of the combined, dried organic
solutions gave a pale yellow oil, which was
shown by TLC (solvent B) to contain starting
material (R 0.46) and another component (R
tified as a mixture of 6 and 7. The least mobile
component (R 0.11; 0.28 g), isolated as an oil,
f
crystallised on storage under high vacuum and
was recrystallised from light petroleum–
EtOAc to give a product shown to be an
approximately equimolar mixture of the two
diastereoisomeric crossed-aldol products, 9
and 10 (0.14 g, 10%), mp 62–63 °C; IR (Nu-
jol): w 3480 (OH), 1720 (CꢁO), 1380 and 1370
f
f
0
.3). Column chromatography (solvent B)
yielded 5 (0.16 g, 32%) and the less mobile
component which, on spectroscopic evidence,
was a single diastereoisomer, the aldol
product 6 or 7 as an oil (0.15 g, 30%); IR
−
1
1
(
film): w 3480 (OH), 1720 (CꢁO), 1380 and
cm (C(CH ) ); H NMR (270 MHz): l 1.33,
3
2
−
1
1
1
1
1
1
3
9
370 cm (C(CH ) ); H NMR (270 MHz): l
1.34 (×2), 1.37, 1.40, 1.42, 1.44 (×2) (8×s,
8×3 H, 4×C(CH ) ), 3.00 (dd, 1 H, J
3
2
.32, 1.33, 1.35 (×2), 1.37, 1.39, 1.43 (×2),
.45 (9×s, 9×3 H, 4×C(CH ) and H C-7),
3
2
1a,2a
3.3, J2a,2%a 17.8 Hz, H-2a), 3.03 (dd, 1 H, J1b,2b
4.0, J2b,2%b 18.5 Hz, H-2b), 3.12 (dd, 1 H, J1b,2%b
8.6 Hz, H-2%b), 3.13 (br s, 2×1 H, 2×OH),
3.17 (dd, 1 H, J1a,2%a 9.2, H-2%a), 3.96 (complex,
2×1 H, H-7), 4.08–4.22 (complex, 2×3 H,
H-5, 6, 7%), 4.37 and 4.39 (2×d, 2×1 H, J4a,5a
2.6, J4b,5b 3.0 Hz, H-4a and H-4b), 5.20 (com-
plex, 2×1 H, H-1), 7.26–7.37 (complex, 2×5
3
2
3
.92 (br s, 1 H, OH), 2.97 (s, 2 H, H-6, 6%),
.60–4.42 (complex, 10 H, H-1, 1%, 2, 3, 4, 8,
1
3
, 10, 11, 11%); C NMR (67.9 MHz): l 23.2
(
(
(
(
H C-7), 25.2, 25.3, 26.3, 26.4, 26.6, 27.1, 27.2
3
×2) (4×C(CH ) ), 47.0 (C-6), 66.7, 67.5
3
2
C-1, 11), 71.4 (C-7), 76.6, 76.8, 77.6, 77.9
C-2, 3, 9, 10), 83.7, 84.2 (C-4, 8), 109.3, 109.9,
1
3
1
10.0, 111.4 (4×C(CH ) ), 208.9 (CꢁO);
H, 2×C H ); C NMR (67.9 MHz): l 25.1
3
2
6
5
EIMS: m/z 473 (0.7, M−15). Anal. Calcd for
(×2), 26.1, 26.2, 26.5 (×2) and 27.0 (×2)
(4×C(CH ) ), 48.1 (2×C-2), 66.7, 66.8 (2×
C H O : C, 59.0; H, 8.25. Found: C, 58.7;
H, 8.2.
2
4
40 10
3 2
C-7), 69.6, 69.8, (2×C-6), 76.4, 76.5 (2×C-
1), 78.1, 78.2, 83, 83.2 (2×C-4, 5), 110.0
(×2), 111.5 and 111.6 (4×C(CH ) ), 125.6,
Reaction of ketone 5 with benzaldehyde.
Preparation of 1-(R)- and 1-(S)-2-deoxy-
,5:6,7-di-O-isopropylidene-1-C-phenyl- -ara-
bino-hept-3-ulose (9) and (10), respecti6ely.—
a) Using LDA: to a stirred soln of diiso-
3
2
4
L
125.7, 127.7, 127.8, 128.6, 142.8 (C-aromatic),
209.2 and 209.5 (2×CꢁO). Anal. Calcd for
C H O : C, 65.1; H, 7.5. Found: C, 65.1; H,
(
1
9
26
6
propylamine (0.6 mL, 0.43 g, 4.3 mmol) in
anhyd THF (10 mL) at −78 °C under N2,
was added n-butyllithium (2.5 M soln in hex-
ane; 1.8 mL, 4.5 mmol). The mixture was
stirred for 15 min and then a soln of 5 (1.00 g,
7.4.
Analytical HPLC (detection at 254 nm) in-
dicated the presence of two isomers, but a
baseline separation was not obtained in a
1
variety of solvent systems. However, H and
1
3
4
.09 mmol) in anhyd THF (5 mL) was added
C NMR spectroscopy showed that the ratio
dropwise, over 10 min. After 15 min, freshly
distilled benzaldehyde (0.42 mL, 4.09 mmol)
was added neat in one portion. The mixture
was stirred at −78 °C for 30 min and then
allowed to warm to rt when TLC (solvent A)
of the isomers was approximately 1:1.
(b) Using LiHMDS: reaction of ketone 5
(0.75 g, 3.07 mmol) with benzaldehyde (0.35
mL, 3.43 mmol) under Method 1 conditions
gave, after column chromatography (solvent
B), starting ketone 5 (0.04 g, 5%) followed by
the mixed crossed-aldol products 9 and 10
confirmed the disappearance of 5 (R 0.34)
f
and formation of two components (R 0.20
f

206
A.H. Haines, A.J. Lamb / Carbohydrate Research 321 (1999) 197–213
(
0.72 g, 67%). Recrystallisation of the mixture
two diastereoisomers from light
65.5, 66.7 (×3), 67.6, 68.7 (2×C-1, 3, 9),
76.4, 76.5, 77.6, 77.7, 78.0, 78.3 (2×C-2, 7, 8),
83.1, 83.3 (2×C-6), 109.5, 109.6, 110.0 (×2),
111.5 and 111.6 (6×C(CH ) ), 208.5, 209.8
of
petroleum–EtOAc gave material with mp 60–
1
6
4 °C. The product was identical by IR, H
3
2
13
+
and C NMR spectroscopy to the sample of
the crossed-aldol product, prepared as in (a).
(2×CꢁO); EIMS: m/z 374 (0.8, M ), 359
(7.6, M−15). Anal. Calcd for C H O : C,
1
8
30
8
1
H NMR spectroscopy indicated the isomer
57.7; H, 8.1. Found: C, 57.5; H, 8.1.
ratio as approximately 1:1.
By comparison of the integration values of
1
(
c) Using dibutylboron triflate–Et N: reac-
certain peaks in the H NMR spectrum in the
3
tion of ketone 5 (0.15 g, 0.61 mmol) with
benzaldehyde (0.20 mL, 1.79 mmol) under
Method 2 conditions gave, after column chro-
matography (solvent B), the crossed-aldol
product as a pale yellow oil (0.12 g), which
solidified on storage. Recrystallisation from
light petroleum–EtOAc, gave a mixture of 9
and 10 (92 mg, 43%), mp 57–61 °C. The
isomeric ratio (unassigned) in this preparation
as determined by H NMR spectroscopy in
the presence of Eu(fod) was 7:3. No self-con-
densation aldol products 6 or 7, or ketone 5
were obtained, although unreacted benzalde-
hyde was isolated (0.11 g, 58% recovery).
Reaction of ketone 5 with 2,3-O-isopropyli-
presence of Eu(fod) , the ratio of isomers
3
(unassigned) was estimated as approximately
1.5:1.
(b) Using dibutylboron triflate–Et N: reac-
3
tion of ketone 5 (0.61 g, 2.50 mmol) with
aldehyde 13 (0.36 g, 2.77 mmol) under
Method 2 conditions gave, after column chro-
matography (solvent B), 5 (62 mg, 10% recov-
ery), a small amount of the self-aldol product
6 and/or 7 (54 mg, 9%) and, as an oil, a
mixture of the crossed-aldol reaction products
14 and 15 (0.33 g, 35%). Spectroscopic data
1
3
1
13
(IR, and H and C NMR) on this material
agreed with that for the mixture of 14 and 15
prepared in (a), but differing peak intensities
reflected a different ratio of diastereoisomers.
Analytical HPLC in hexane-2-propanol (24:1)
gave two peaks at R 11.62 min and R 13.19
dene-
deoxy-1,2:6,7:8,9-tri-O-isopropylidene-
cero- -galacto-non-5-ulose (14) and 4-deoxy-
,2:6,7:8,9-tri-O-isopropylidene- -glycero-
D
-glyceraldehyde (13). Preparation of 4-
L
-gly-
L
t
t
1
L
L
-
min in a ratio of 13:12.
gulo-non-5-ulose (15).—(a) Using LiHMDS:
reaction of ketone 5 (2.0 g, 8.19 mmol) and
aldehyde 13 (1.17 g, 9.0 mmol) under Method
Reaction between ketone 5 and 2,3:4,5-di-O-
-arabinose (18). Preparation of
6-deoxy-1,2:3,4:8,9:10,11-tetra-O-isopropyli-
dene- -arabino- -manno-undec-5-ulose (19)
and 6-deoxy-1,2:3,4:8,9:10,11-tetra-O-isopro-
pylidene - - arabino - - gulo - undec - 5 - ulose
isopropylidene-
D
1
conditions gave, after column chromatogra-
phy (solvent B), starting ketone 5 (0.24 g,
2%), a product shown by comparison with a
D
L
1
D
D
known sample (TLC, and IR and NMR spec-
troscopy) to be 6 and/or 7 (0.38 g, 19%) and
finally, as an oil, a mixture of the title com-
pounds 14 and 15 (1.48 g, 48%); IR (film): w
(20).—(a) Using LiHMDS: reaction of ketone
5 (0.40 g, 1.64 mmol) with aldehyde 18 (0.41
g, 1.78 mmol) under Method 1 conditions
gave, after column chromatography (solvent
B), ketone 5 (48 mg, 12% recovery), the self-
aldol product 6 and/or 7 (52 mg, 13%) and, as
a colourless oil, a mixture of the title com-
pounds 19 and 20 (0.38 g, 49%); IR (film): w
−
1
3
470 (OH), 1720 (CꢁO), 1380 and 1370 cm
1
(
C(CH ) ); H NMR (270 MHz): l 1.33 (×2),
3
2
1
(
.35 (×4), 1.38, 1.39, 1.41, 1.42, 1.43 and 1.46
12×s, 12×3 H, 6×C(CH ) ), 2.74 (dd, 1 H,
3
2
−
1
J3a,4a 2.3, J4a,4%a 16.2 Hz, H-4a), 2.82 (dd, 1 H,
J3b,4b 8.3, J4b,4%b 18.2 Hz, H-4b), 2.99 (dd, 1 H,
J3b,4%b 9.2 Hz, H-4%b), 3.00 (br d, 21 H, J3,OH
3470 (OH), 1720 (CꢁO), 1385 and 1375 cm
1
(C(CH ) ); H NMR (270 MHz): l 1.34–1.48
(complex, 48 H, 8×C(CH ) ), 2.88 (dd, 1 H,
3 2
3 2
3
.6 Hz, 2×OH), 3.05 (dd, 1 H, J3a,4%a 2.6 Hz,
H-4%a), 3.82–4.23 (complex, 2×8 H, H-1, 1%,
, 3, 7, 8, 9, 9%), 4.34–4.42 (complex, 2×1 H,
J6a,6%a 17.5, J6a,7a 8.9 Hz, H-6a), 2.92 (br s,
2×1 H, 2×OH), 2.97 (2×d, 2 H, J6b,7b 6.3,
J6%b,7b 4.6 Hz, H-6b and H-6%b), 3.06 (dd, 1 H,
J6%a,7a 3.6 Hz, H-6%a), 3.60–4.41 (complex, 2×
2
1
3
H-6); C NMR (67.9 MHz): l 25.1 (×4),
6.2, 26.3, 26.4, 26.5 (×2), 26.6 and 27.0
×2) (6×C(CH ) ), 42.5, 42.9 (2×C-4),
1
3
2
(
11 H, H-1, 1%, 2, 3, 4, 7, 8, 9, 10, 11, 11%); C
NMR (67.9 MHz): l 25.1–27.1 (8×
3
2

A.H. Haines, A.J. Lamb / Carbohydrate Research 321 (1999) 197–213
207
C(CH ) ), 42.9, 43.6 (2×C-6), 66.4, 66.6,
(0.50 g, 1.80 mmol) under Method 1 condi-
tions gave, after column chromatography (sol-
vent B), first the self-aldol reaction product
6 and/or 7 (60 mg, 15%) and secondly, as an
oil, a mixture of the title compounds 23 and
24 (0.20 g, 25%). By re-chromatography of the
mixture of crossed-aldol isomers 23 and 24,
3
2
6
8
1
7.7, 67.8 (2×C-1, 11), 68.9 (2×C-7), 76.4–
3.4 (2×C-2, 3, 4, 8, 9, 10), 109.5, 109.6,
09.7, 109.8, 109.9, 110.2, 111.4, 111.5 (8×
C(CH ) ), 208.5, 208.6 (2×CꢁO); EIMS: m/z
3
2
+
4
74 (0.3, M ), 459 (2.7, M−15). Anal. Calcd
for C H O : C, 58.2; H, 8.1. Found: C, 57.9;
2
3
38 10
1
H, 8.1.
(solvent A), the faster running isomer could
Analytical HPLC (49:1 hexane–2-propanol)
be obtained pure (61 mg, 7%); IR (film):
showed two isomers (R 12.5 min and R 13.6
w 3470 (OH), 1720 (CꢁO), 1610 (weak, C H ),
t
t
6
5
−
1
1
min) in a ratio of 1.2:1, respectively. Prepara-
tive HPLC was used to isolate pure samples of
1380 and 1370 cm
(C(CH ) ); H NMR
3
2
(270 MHz): l 1.31, 1.33, 1.34, 1.42, 1.44 and
1
both stereoisomers in order to obtain an H
1
1
.48 (6×s, 6×3 H, 3×C(CH ) ), 2.89 (br s,
H, OH), 2.93 (dd, 1 H, J5,6 8.9, J6,6% 18.1 Hz,
3
2
NMR spectrum of each one.
1
Data for isomer with R 12.5 min; H NMR
t
H-6), 3.12 (dd, 1 H, J5,6% 2.6 Hz, H-6%), 3.94–
.73 (complex, 9 H, H-2, 3, 4, 5, 8, 9, 10,
(
270 MHz): l 1.35, 1.36, 1.38, 1.39, 1.41, 1.43
4
(
×2) and 1.46 (8×s, 8×3 H, 4×C(CH ) ),
3
2
11, 11%), 4.59 (d, 1 H, JAB 11.6 Hz,
C H CH H ), 4.71 (d, 1 H, C H CH H ),
2
.82 (br d, 1 H, J7,OH 6.3 Hz, OH), 2.88 (dd, 1
6
5
A
B
6
5
A
B
H, J6,6% 17.5, J6,7 8.9 Hz, H-6), 3.06 (dd, 1 H,
J6%,7 3.6 Hz, H-6%), 3.89–4.58 (complex, 10 H,
H-1, 1%, 2, 3, 7, 8, 9, 10, 11, 11%), 4.41 (d, 1 H,
5.89 (d, 1 H, J1,2
4.0 Hz, H-1), 7.16–7.36
5
1
3
(complex, 5 H, C H ); C NMR (67.9 MHz):
6
l 25.2, 26.2, 26.3, 26.5, 26.8 and 27.0 (3×
C(CH ) ), 43.5 (C-6), 64.9 (C-11), 72.4
J3,4 3.3 Hz, H-4).
3
2
1
Data for isomer with R 13.6 min; H NMR
t
(C H CH ), 66.7, 76.5, 78.0, 81.3, 81.9, 82.5,
6
5
2
(
270 MHz): l 1.36 (×4), 1.37, 1.44 (×2),
8
3.3 (C-2, 3, 4, 5, 8, 9, 10), 105.1 (C-1), 109.9,
1
2
1
1
1
.46 (8×s, 8×3 H, 4×C(CH ) ), 2.97 (2×d,
3
2
111.5, 111.8 (3×C(CH ) ), 127.9, 128.1,
3
2
H, J6,7 6.3, J6%,7 4.6 Hz, H-6, 6%), 3.55 (br d,
H, J7,OH 1.6 Hz, OH), 3.73–4.30 (complex,
0 H, H-1, 1%, 2, 3, 7, 8, 9, 10, 11, 11%), 4.41 (d,
H, J3,4 5.6 Hz, H-4).
128.6, 137.4 (C-aromatic), 210.2 (CꢁO);
EIMS: m/z 273 (0.8, M−249), 249 (1.3, M−
2
7
73). Anal. Calcd for C H O : C, 62.1; H,
.3. Found: C, 61.7; H, 7.15.
Integration of the signals for H-1 in the
2
7
38 10
(b) Using dibutylboron triflate–Et N: reac-
3
tion of ketone 5 (1.00 g, 4.09 mmol) with
aldehyde 18 (1.04 g, 4.52 mmol) under
Method 2 conditions gave, after column chro-
matography (solvent B), only the crossed-al-
dol reaction products 19 and 20 (1.57 g, 80%).
The spectroscopic data for a mixed-isomer
sample agreed with those for the previous
sample, differing only because of its different
diastereoisomeric ratio.
NMR spectrum of the original isomeric mix-
ture of 23 and 24, gave an approximate
diastereoisomeric ratio of 1:1.
(
b) Using dibutylboron triflate–Et N: reac-
3
tion of ketone 5 (1.00 g, 4.09 mmol) with
aldehyde 4 (1.23 g, 4.42 mmol) under Method
2
(
5
conditions gave, on column chromatography
solvent B) as the first eluted material, ketone
(0.29 g, 29%). Further elution gave, as a
mixture of isomers, the crossed-aldol products
3 and 24 (0.41 g, 19%). The spectroscopic
Analytical HPLC (49:1 hexane-2-propanol)
showed two isomers (R 12.7 min and R 13.7
t
t
2
min) in a ratio of 1:1.6, respectively.
data for this mixture of isomers were identical,
except for differences resulting from a differ-
ent diastereoisomeric ratio, to those described
above in (a) for the aldol product obtained
from the reaction using LiHMDS.
Reaction of ketone 5 with 3-O-benzyl-1,2-O-
-xylo-pentodialdo-1,4-fura-
nose (4). Preparation of 3-O-benzyl-6-deoxy-
,2:8,9:10,11-tri-O-isopropylidene- -arabino-i-
-ido-undecofuranos-7-ulose (23) and 3-O-
benzyl-6-deoxy-1,2:8,9:10,11-tri-O-isopropyli-
dene- -arabino-h- -gluco-undecofuranos-7-
isopropylidene-h-
D
1
L
L
L
D
1
By comparison with the single-isomer sample obtained via
ulose (24).—(a) Using LiHMDS: reaction of
ketone 5 (0.42 g, 1.72 mmol) with aldehyde 4
the sodium enolate of 5, which is tentatively assigned as 24,
this material obtained from chromatography is also 24.

208
A.H. Haines, A.J. Lamb / Carbohydrate Research 321 (1999) 197–213
Integration of the signals for H-1 in the
satd aq NaHCO (5 mL) and water (5 mL).
3
NMR spectrum of the original isomeric mix-
ture of 23 and 24 gave an approximate
diastereoisomeric ratio of 1.5:1. The major
isomer was identical to the single isomer sam-
ple prepared as described in (c) below, which
is tentatively assigned as 24.
The dried organic soln was co-evaporated
with toluene (2×15 mL) to remove residual
pyridine and then concentrated to give a crude
syrup (87 mg) containing three components.
Column chromatography (solvent A) yielded,
initially, as an oil, alkene 27 (62 mg, 50%);
[h] −17.0° (c 5.3, CHCl ); IR (film): w 1700
(c) Using NaHMDS: to a stirred soln of
D
3
NaHMDS (1.0 M in THF; 25 mL, 25 mmol)
under argon at −78 °C was added dropwise a
soln of ketone 5 (5.00 g, 20.5 mmol) in anhyd
THF (50 mL) such that the temperature of the
reaction mixture remained below −65 °C.
The soln was stirred at −78 °C for 30 min
and then a concd soln of aldehyde 4 (6.56 g,
(CꢁO), 1630 (CꢁC), 1595 (C H ), 1380 and
6
5
−1 1
1370 cm (C(CH ) ); H NMR (270 MHz): l
3
2
1.37, 1.38, 1.43 and 1.51 (4×s, 4×3 H, 2×
C(CH ) ), 4.01 (dd, 1 H, J 5.0, J7,7% 8.6 Hz,
3
2
6,7
H-7), 4.14 (dd, 1 H, J6,7% 6.3 Hz, H-7%), 4.25
(ddd, 1 H, J5,6 7.0 Hz, H-6), 4.35 (dd, 1 H, J4,5
5.3 Hz, H-5), 4.62 (d, 1 H, H-4), 7.20 (d, 1 H,
J1,2 16.2 Hz, H-2), 7.39–7.42 and 7.58–7.62
2
3.6 mmol) in anhyd THF (10 mL) was added
1
3
in three rapidly consecutive portions. After 20
min at −78 °C, the reaction mixture was
allowed to warm to −20 °C and after a fur-
ther 10 min, it was quenched by the addition
(complex, 5 H, C H ), 7.76 (d, 1 H, H-1); C
6
5
NMR (67.9 MHz): l 25.2, 26.3, 26.6 and 27.2
(2×C(CH ) ), 66.7 (C-7), 76.6 (C-6), 78.5 (C-
3
2
5), 82.6 (C-4), 109.9 and 111.4 (2×C(CH ) ),
3
2
of satd aq NaHCO (100 mL). Separation of
121.6 (C-2), 128.6, 128.9, 130.8, 134.5 (C-aro-
3
the two phases, extraction of the aq phase
matic), 144.4 (C-1), 197.6 (CꢁO); EIMS: m/z
+
with Et O (3×100 mL) and concentration of
332 (1.9, M ), 317 (7.3, M−15). Anal. Calcd
2
the combined, dried organic solutions gave a
syrup (10.86 g). This material mainly con-
tained the desired aldol products 23 and 24
plus some of the starting materials 4 and 5, as
indicated by TLC (solvent B); for further
transformations it could be used in this crude
form or purified by column chromatography
in solvent B. Such purification on a portion of
the crude syrup (1.0 g) yielded as the first-
eluted material ketone 5 (0.10 g; 22%), then a
single isomer (as indicated by NMR spec-
troscopy and analytical HPLC) of the crossed-
aldol product, tentatively assigned (see Section
for C H O : C, 68.7; H, 7.3. Found: C, 68.4;
1
9
24
5
H, 7.3.
Further elution gave a mixture of the 1-(R)-
and 1-(S)-acetates, 11 and 12, respectively,
combined with alkene 27, from which, by
re-chromatography, could be obtained as an
oily material identified by NMR spectroscopy
as one of the acetates 11 or 12 (6.5 mg, 5%);
1
H NMR (270 MHz): l 1.41, 1.55 and 1.56
and 1.57 (4×s, 4×3 H, 2×C(CH ) ), 1.99–
3
2
2.17 (complex, 2 H, and H-2, 2%), 2.03 (s, 3 H,
CH CO), 3.67–4.14 (complex, 6 H, H-1, 4, 5,
3
1
3
6, 7, 7%), 7.26–7.47 (complex, 5 H, C H ); C
6
5
2
) as 24 (0.52 g, 53%); [h] −18.7° (c 6.3,
NMR (75.4 MHz): l 21.0 (COCH ), 25.1,
D
3
CHCl ). Other spectroscopic data were identi-
25.9, 26.5 and 26.7 (2×C(CH ) ), 60.0 (C-2),
3
3 2
cal to those for the related component in the
previously separated sample of mixed isomers.
67.8 (C-7), 73.1, 76.5 (C-5, 6), 77.1 (C-4), 80.8
(C-1), 110.3 and 111.0 (2×C(CH ) ), 128.1,
3
2
(
E) - 1,2-Dideoxy-4,5:6,7-di-O-isopropyli-
dene-1-C-phenyl- -arabino-hept-1-en-3-ulose
27).—A solution of aldol adducts 9 and 10
130 mg, 0.37 mmol) and Ac O (0.15 mL, 0.11
129.8, 129.0, 135.7 (C-aromatic), 169.5
L
(CH CꢁO), 209.4 (CꢁO).
3
(
(
Finally, a mixture of starting materials 9
and 10 (2.7 mg, 2%) was recovered.
2
g, 1.08 mmol) in pyridine (2 mL) was stored at
rt (18 °C) for 45 h and water (0.3 mL) and
3-O-Acetyl-4-deoxy-1,2:6,7:8,9-tri-O-iso-
propylidene-
L
-glycero- -galacto-non-5-ulose
L
satd aq NaHCO (2.5 mL) were then added
(16), 3-O-acetyl-4-deoxy-1,2:6,7:8,9-tri-O-iso-
propylidene- -glycero- -gulo-non-5-ulose (17),
and (E)-3,4-dideoxy-1,2:6,7:8,9-tri-O-isopro-
pylidene- -gluco-non-3-en-5-ulose (28).—To a
soln of a mixture of 14 and 15 (0.30 g,
3
sequentially. After being stirred for 30 min,
L
L
the soln was extracted with CH Cl (2×10
2
2
mL) and the combined organic solutions were
washed with 0.1 M HCl (5 mL), water (5 mL),
L

A.H. Haines, A.J. Lamb / Carbohydrate Research 321 (1999) 197–213
209
1
3
0
.80 mmol) in pyridine (3 mL) was added
2
8.0. From the C NMR spectrum the two
isomers were in a ratio of 2:1.
Ac O (0.15 mL, 0.11 g, 1.09 mmol) and the
reaction mixture was stored at rt (18 °C) for
Eluted last as an inseparable mixture were
starting materials 14 and 15 (30 mg, 10%).
(E)-6,7-Dideoxy-1,2:3,4:8,9:10,11-tetra-O-
2
h. Product isolation, as described for
the preparation of 27, gave a syrup (0.20 g)
containing [TLC (solvent B)] three compo-
nents, R 0.34 (major), R 0.23 (minor), and R
isopropylidene-
-ulose (29).—(a) A solution of 19 and 20,
0.15 g, 0.32 mmol) and Ac O (0.10 mL, 0.72
D
-erythro-
L
-manno-undec-6-en-
5
f
f
f
(
0
.11 (14 and 15). On column chromatography
solvent B), first eluted as an oil was alkene
8 (0.11 g, 39%); [h] +13.6° (c 0.93, CHCl );
2
mmol) in pyridine (2 mL) was stored at rt for
8 h, and the reaction mixture then worked-
(
4
2
D
3
up as in the preparation of 27 to yield a crude
oil (0.12 g), which was subjected to column
chromatography (solvent B). First eluted, as a
IR (film): w 1700 (CꢁO), 1630 (CꢁC), 1380
−
1
1
and 1370 cm
MHz): l 1.35 (×2), 1.42 (×2) 1.46 and 1.47
6×s, 6×3 H, 3×C(CH ) ), 3.68 (dd, 1 H,
(C(CH ) ); H NMR (270
3
2
colourless oil (R 0.60), was the alkene 29 (19
(
f
3
2
mg, 13%); [h] −15.1° (c 1.65, CHCl ); IR
J1,1% 7.6, J1,2 7.6 Hz, H-1), 3.96 (dd, 1 H, J8,9
.6, J9,9% 8.6 Hz, H-9), 4.09–4.30 (complex, 4
D
3
(
1
film): w 1695 (CꢁO), 1630 (CꢁC), 1380 and
4
−
1
1
370 cm
(C(CH ) ); H NMR (270 MHz):
3
2
H, H-1%, 7, 8, 9%), 4.53 (d, 1 H, J6,7 5.3 Hz,
H-6), 4.71 (m, 1 H, H-2), 6.81 (dd, 1 H, J
l 1.33, 1.34, 1.35, 1.39, 1.41, 1.43 (×2), 1.47
(8×s, 8×3 H, 4×C(CH ) ), 3.65–4.30
2
,3
4
3
2
5
.6, J3,4 15.5 Hz, H-3), 6.97 (dd, 1 H, J 5.0
2,4
13
(complex, 8 H, H-1, 1%, 2, 3, 9, 10, 11, 11%),
.53 (d, 1 H, J3,4 5.0 Hz, H-4), 4.58 (ddd, 1 H,
J6,8 1.7, J7,8 4.3, J8,9 7.9 Hz, H-8), 6.89 (dd, 1
H, J6,7 15.5 Hz, H-6), 7.09 (dd, 1 H, H-7); C
NMR (67.9 MHz): l 25.2 (×2), 26.3, 26.5,
6.7, 26.8, 27.0 and 27.2 (4×C(CH ) ), 66.7,
Hz, H-4); C NMR (67.9 MHz): l 25.1, 25.6,
4
4
2
6.2, 26.4, 26.5 and 27.1 (3×C(CH ) ), 66.6
3
2
(
(
C-9), 68.6 (C-1), 75.1, 76.4 (C-7, 8), 78.2
1
3
C-2), 82.2 (C-6), 109.8, 110.2 and 111.4 (3×
C(CH ) ), 125.5 (C-4), 144.6 (C-3), 197.1
3
2
2
6
1
+
3 2
(
CꢁO); EIMS: m/z 356 (1.9, M ), 341 (18.1,
7.5 (C-1, 11), 76.5, 77.0, 78.3, 79.4 (C-2, 3, 9,
0), 81.2, 82.4 (C-4, 8), 109.9, 110.0, 110.3
M−15), 255 (1.2, M−101), 155 (4.7, M−
2
01); HRMS: calcd for C H O : 356.1835;
18
28
7
and 111.5 (4×C(CH ) ), 124.7 (C-6), 145.2
3
2
found: m/z 356.1835. Anal. Calcd for
(
C-7), 197.5 (CꢁO); EIMS: m/z 456 (3.5,
C H O : C, 60.7; H, 7.9. Found: C, 60.25;
+
1
8
28
7
M ), 441 (23.4, M−15). Anal. Calcd
H, 7.8.
for C H O : C, 60.5; H, 7.95. Found: C,
2
3
36
9
Further elution gave, as an oil, the mixture
6
0.5; H, 8.1.
of stereoisomers, 16 and 17, (21 mg, 6%); IR
Further elution gave an unseparated mix-
ture of two slower-running components (Rf
(film): w 1760–1700 (br, CꢁO), 1380 and 1370
cm
−1 1
(C(CH ) ); H NMR (270 MHz): l
3
2
0
.25 and 0.28) as an oil, shown by spec-
1
1
4
9
.24–1.44 (complex, 18 H, 3×C(CH ) ),
3
2
troscopy to be an approximately equimolar
.98, 2.00 (2×s, 3 H in total, COCH ), 3.67–
3
mixture of the two isomers 21 and 22 (32 mg,
.74 (complex, 10 H, H-1, 1%, 2, 4, 4%, 6, 7, 8,
1
9%); IR (film): w 1710 (br, CꢁO), 1380 and
13
, 9%), 5.31–5.47 (complex, 1 H, H-3);
C
1
cm
375 (C(CH ) ), no absorption near 3400
3
2
NMR (67.9 MHz): l major isomer: 20.8
COCH ), 25.1 (×2), 26.1, 26.2, 26.3 and
−1
1
(OH); H NMR (270 MHz): l 1.32–
(
3
1.47 (complex, 16×3 H, 8×C(CH ) ), 2.04,
3
2
2
6.4 (3×C(CH ) ), 48.8 (C-4), 65.8, 66.4 (C-
3
2
2.06 (2×s, 2×3 H, 2×COCH ), 2.88 (dd, 1
3
1
, 9), 69.3, 71.7, 74.5, 77.9 (C-2, 3, 7, 8), 82.6
H, J6a,6%a 17.1, J6a,7a 3.4 Hz, H-6a), 2.97 (d, 2
H, J6b,7b 6.2, J6%b,7b 6.2 Hz, H-6b, 6%b), 3.07
(dd, 1 H, J6a,6%a 8.8, J6%a,7a 8.8 Hz, H-6%a),
3.66–4.87 (complex, 2×11 H, H-1, 1%, 2, 3, 4,
(
C-6), 109.7, 109.8 and 110.0 (3×C(CH ) ),
3
2
1
2
2
1
69.7 (COCH ), 207.5 (CꢁO); minor isomer:
3
0.9 (COCH ), 25.1, 25.2, 25.3, 26.0, 27.0 and
3
1
3
7.1 (3×C(CH ) ), 48.2 (C-4), 65.6, 67.0 (C-
7, 8, 9, 10, 11, 11%); C NMR (22.4 MHz)): l
3
2
, 9), 69.0, 74.1, 75.9, 78.0 (C-2, 3, 7, 8), 84.2
25.2 and 25.4 (2×COCH ), 26.4–27.5 (8×
3
(
C-6), 109.8, 109.9, 112.4 (3×C(CH ) ), 170.0
C(CH ) ), 52.0, 54.7 (2×C-6), 65.8–84.8
3
2
3 2
(
CH CꢁO), 205.1 (CꢁO); EIMS: m/z 417 (1.8,
(2×C-1, 2, 3, 4, 7, 8, 9, 10, 11), 109.4–112.0
(8×C(CH ) ), 175.3 and 175.4 (2×
3
+
+
M +1), 401 (2.7, M −15). Anal. Calcd for
3
2
C H O : C, 57.7; H, 7.7. Found: C, 58.0; H,
CH CꢁO), 204.9 and 209.7 (2×CꢁO).
2
0
32
9
3

210
A.H. Haines, A.J. Lamb / Carbohydrate Research 321 (1999) 197–213
1
3
Further elution with 2:1 light petroleum–
7.24–7.36 (complex, 5 H, C H ); C NMR
6
5
EtOAc gave a mixture of starting materials 19
and 20 (33 mg, 22%).
(75.4 MHz): l 25.2, 26.2, 26.3, 26.6, 26.9
and 27.2 (3×C(CH ) ), 66.9 (C H CH ), 72.3
3
2
6
5
2
(b) Treatment of 19 and 20 (0.207 g, 0.44
(C-11), 76.8, 78.4, 80.0, 82.5, 83.0, 83.2 (C-2,
3, 4, 8, 9, 10), 105.2 (C-1), 110.1, 111.8 and
112.1 (3×C(CH ) ), 126.9 (C-6), 128.0, 128.8,
mmol) in pyridine (2 mL) with p-tolylsulfonyl
chloride (0.102 g, 0.54 mmol) at rt for 5 days
and conventional work-up gave a crude oil
3
2
128.3, 137.3 (C-aromatic), 141.9 (C-5), 197.3
(
0.13 g) which, when subjected to column
(CꢁO); HRMS: calcd for C H O : m/z,
2
7
36
9
chromatography (solvent B), gave as an oil
alkene 29 (54 mg, 27%).
504.2359; found: 504.2360.
Further elution gave, according to NMR
spectroscopy, a mixture of the isomeric ac-
etates 25 and 26 containing a small amount of
the elimination product 30. Re-chromatogra-
phy of the mixture gave a sample mostly
comprising one of the acetates (presumably
isomer 26, as shown by comparison with the
single isomeric O-acetate isolated in a similar
reaction conducted in the presence of DMAP
on the compound tentatively identified as 24,
see later) containing approximately 10 mol%
of the second isomer 25 (15 mg, 10%); IR
Further elution (2:1 light petroleum–
EtOAc) gave a mixture of 19 and 20 (61 mg,
2
9% recovery). No products of O-p-tolylsul-
fonylation could be identified.
E)-3-O-Benzyl-5,6-dideoxy-1,2:8,9:10,11-
tri-O-isopropylidene- -erythro-h- -gluco-un-
(
L
D
dec-5-enofuranos-7-ulose (30).—Acetylation/
elimination reactions on the mixed isomer
sample of the aldol adducts 23 and 24 to give
the a,b-unsaturated ketone and an isomeric
mixture of the corresponding acetates 25 and
2
6 were carried out under four different reac-
(film): w 1745 (CH CꢁO), 1720 (CꢁO), 1375
3
−
1
1
tion conditions to optimise the yield and to
increase the rate of the reaction. That per-
formed in the presence of a catalytic quantity
of DMAP was adopted for the bulk prepara-
tion of product and was also used to acetylate
and eliminate the single-isomer sample of the
aldol adduct (i.e., the isomer presumed to be
and 1365 cm
(C(CH ) ); H NMR (300
3
2
MHz) major isomer: l 1.24, 1.31, 1.34, 1.42
(×2) and 1.47 (6×s, 6×3 H, 3×C(CH ) ),
3
2
1.93 (s, 3 H, CH CꢁO), 3.12 (dd, 1 H, J 7.0,
3
5,6
J
18.0 Hz, H-6), 3.21 (dd, 1 H, J 4.2, J6,6%
6
,6%
5,6%
Hz, H-6%), 3.92–4.37 (complex, 7 H, H-2, 3, 4,
9, 10, 11, H-11%), 4.47 (d, 1 H, JAB 11.5 Hz,
C H CH H ), 4.58 (d, 1 H, J 3.6 Hz, H-8),
2
4) to give 26 and (E)-enone 30.
6
5
A
B
8,9
Acetylation of a solution of 23 and 24 (137
4.60 (d, 1 H, C H CH H ), 5.60 (ddd, 1 H,
6
5
A
B
mg, 0.26 mmol) with Ac O (0.12 mL, 0.87
J4,5 5.6 Hz, H-5), 5.91 (d, 1 H, J1,2 3.7 Hz,
2
mmol) in pyridine (2 mL) for 6 days, and
product isolation, as described in the pre-
paration of 27, gave a crude syrup (119 mg),
which on column chromatography (solvent A)
yielded initially, as an oil, the enone 30
H-1), 7.27–7.37 (complex, 5 H, C H ); minor
6
5
isomer: 1.25 (×2), 1.34, 1.39 and 1.45 (×2)
(6×s, 6×3 H, 3×C(CH ) ), 1.95 (s, 3 H,
3
2
CH CꢁO), 2.74 (dd, 1 H, J 8.2, J6,6% 16.2 Hz,
3
5,6
H-6), 2.98 (dd, 1 H, J 3.3 Hz, H-6%), 3.90–
5
,6%
(
62 mg, 46%); [h] −44.9° (c 4.5, CHCl ); IR
4.34 (complex, 7 H, H-2, 3, 4, 9, 10, 11, 11%),
4.48 (d, 1 H, JAB 11.8 Hz, C H CH H ), 4.54
D
3
(
film): w 1700 (CꢁO), 1640 (CꢁC), 1620 (C H ),
6
5
6
5
A
B
−
1
1
1
385 and 1370 cm
(C(CH ) ); H NMR
(d, 1 H, J8,9 5.3 Hz, H-8), 4.62 (d, 1 H,
C H CH H ), 5.53 (ddd, 1 H, J 6.3 Hz,
3
2
(
300 MHz) l 1.32, 1.33, 1.35, 1.42, 1.45 and
.49 (6×s, 6×3 H, 3×C(CH ) ), 3.97 (dd, 1
6
5
A
B
4,5
1
H-5), 6.01 (d, 1 H, J 3.9 Hz, H-1), 7.27–7.37
3
2
1,2
13
H, J10,11 4.8, J11,11% 8.5 Hz, H-11), 4.00 (d, 1
H, J3,4 3.4 Hz, H-3), 4.12 (dd, 1 H, J10,11% 6.2
Hz, H-11%), 4.19 (ddd, 1 H, J9,10 6.9 Hz, H-10),
(complex, 5 H, C H ); C NMR (75.4 MHz)
6
5
(for 26): l 21.0 (CH CO), 25.2, 26.1, 26.3,
3
26.5, 26.9, 27.1 (3×C(CH ) ), 40.3 (C-6),
3 2
4
.29 (dd, 1 H, J8,9 5.4 Hz, H-9), 4.47 (d, 1
66.7, 67.6, 77.3, 78.0, 80.7, 81.7, 82.1 and 83.1
H, JAB 12.0 Hz, C H H H ), 4.54 (d, 1 H,
(C-2, 3, 4, 5, 8, 9, 10, 11), 72.3 (C H CH ),
6
5
A
B
6
5
2
H-8), 4.60 (d, 1 H, C H H H ), 4.65 (d, 1 H,
J1,2 3.8 Hz, H-2), 4.84 (ddd, 1 H, J4,5 4.8, J
105.2 (C-1), 110.1, 111.6, 112.1 (3×C(CH3)2),
6
5
A
B
4
128.3, 128.7, 128.8, 137.2 (C-aromatic), 170.3
4,6
1
1
.6 Hz, H-4), 6.01 (d, 1 H, H-1), 6.88 (dd,
H, J5,6 15.8 Hz, H-5), 7.06 (dd, 1 H, H-6),
(CH CꢁO), 206.5 (CꢁO); EIMS: m/z 473 (2.2,
3
M−91).

A.H. Haines, A.J. Lamb / Carbohydrate Research 321 (1999) 197–213
211
Finally, the starting alcohols 23 and 24 were
recovered (22 mg, 16%).
new, closely-running materials were formed,
all with lower R values than 30. The reaction
f
The isomeric mixture of acetates 25 and 26
was subjected to the same acetylation condi-
tions used in their preparation. Monitoring of
the reaction by TLC (solvent A) revealed slow
conversion of acetates 26 and 26 to an elimi-
nation product. After 3 days, the concentra-
tion of the starting acetate was negligible, and
the major product was confirmed as (E)-enone
mixture was then poured into cold water (5
mL), the aqueous solution was extracted with
Et O (3×5 mL) and the combined organic
2
solutions were washed with water (10 mL),
dried and then concentrated to give a syrup
(81 mg). Column chromatography (solvent A)
allowed only mixtures of the components to
be isolated.
3
0, by TLC and NMR spectroscopic compari-
A fast-running fraction appeared to contain
three distinct but inseparable isomeric compo-
nents in a ratio of approximately 1:1:1. Evi-
dence as to the structure of these components
son with an authentic sample.
Repetition of the reaction on the mixture of
2
3 and 24 at 45 °C for 8 h gave (E)-enone 30
1
(
0.17 g, 61%) and the mixed acetates 25 and
6 (55 mg, 18%). Acetylation at 60 °C gave a
negligible amount of 25 and 26, and (E)-enone
was based solely on the H NMR spectrum of
2
the mixture. Two of these isomers appeared to
result from opening of an epoxide by methox-
ide to yield two of the eight possible stereoiso-
mers depicted by 32 and 33. The third
component in the faster-running mixture, on
3
0 in 40% yield. Reaction of the single
stereoisomer, presumed to be 24, with Ac O–
2
pyridine containing DMAP for 70 h at 18 °C
gave a 20:1 mixture of the (E)- and (Z)-
enones 30 and 31 in 72% yield. Thus, the 300
1
the basis of its H NMR spectrum, was a 6- or
7-C-methoxycarbonyl
derivative
of
a
1
MHz H NMR spectrum of the product
stereoisomer of 3-O-benzyl-6-deoxy-1,2:7,8:9,-
10-tri-O-isopropylidene- -decafuranose, 34
showed alkenic couplings of 11.9 and 15.8 Hz.
Additional peaks that could be assigned to the
L
and 35, originating from Favorskii rearrange-
(
Z)-enone 31 are: l 5.56 (ddd, 1 H, J 3.5,
ment of the intermediate epoxide [8] (11 mg,
H
3,4
4
1
J4,5 6.7, J 1.6 Hz, H-4), 5.98 (d, 1 H, J1,2 3.8
11%); H NMR (300 MHz): l 1.31–1.52
4,6
Hz, H-1), 6.48 (dd, 1 H, J5,6 11.9 Hz, H-5),
6
(complex, 18 H, 3×C(CH ) ), 1.60 (br s, 1 H,
3 2
.77 (dd, 1 H, H-6). Negligible quantities of
OH), 2.20–3.33 (complex), 3.27, 3.28 and 3.71
both the acetate 26 and starting material 24
were obtained from the column separation.
However, if this reaction was quenched after
shorter reaction times then, in addition to the
enones 30 and 31, a single acetate 26 was
(3s, 3 H, 2×CH O (32, 33) and CO CH (34,
3
2
3
35), respectively), 3.85–4.75 (complex), 5.86,
5.88 and 5.91 (3×d, 1 H, J1,2 3.9, 3.7 and 3.7,
respectively, H-1), 7.27–7.37 (complex, 5 H,
C H ).
6
5
1
1
obtained; [h] −26.3° (c 4.75, CHCl ). The H
A slower-running fraction was shown by H
NMR spectroscopy to contain two further
components in a ratio of 2:1, tentatively iden-
D
3
1
3
and C NMR spectra of this material were
identical to those of the major isomer ob-
tained by column chromatography of the
product of acetylation of the mixture of 23
1
tified from the H NMR spectrum to be two
other stereoisomers of 32, 33 (19 mg, 17%);
1
and 24 with Ac O–pyridine.
major isomer: H NMR (270 MHz): l 1.32,
2
Attempted epoxidation of (E)-3-O-benzyl-
1.33, 1.35, 1.47 (×2) and 1.49 (6×s, 6×3 H,
3×C(CH ) ), 1.65 (br s, 1 H, OH), 2.69–3.19
5
,6- dideoxy - 1,2:8,9:10,11 - tri - O - isopropyli -
3
2
dene-
-ulose (30).—(a) Using hydrogen peroxide
and NaOH in MeOH: to a stirred solution of
0 (101 mg, 0.20 mmol) and aq 30% H O (0.1
L
-erythro-h-D-gluco-undec-5-enofuranos-
(complex, 2 H), 3.24 (s, 3 H, CH O), 3.99–
3
7
4.18 (complex, 6 H), 4.34 (ddd, 1 H, J9,10 5.6,
J10,11 5.6, J10,11% 5.6 Hz, H-10), 4.49 (d, 1 H,
JAB 11.6 Hz, C H CH H ), 4.65, (d, 1 H, J
3
2
2
6
5
A
B
8,9
mL) in MeOH (5 mL), at 0 °C was added aq
.5 M NaOH (0.1 mL, 0.25 mmol). The mix-
ture was then allowed to warm to rt and TLC
solvent B) indicated that after 2 h, all the
starting material had been consumed and five
4.0 Hz, H-8), 4.73 (d, 1 H, C H CH H ), 6.01
6
5
A
B
2
(d, 1 H, J1,2 3.6 Hz, H-1), 7.26–7.38 (complex,
1
5 H, C H ); minor isomer: H NMR (270
6
5
(
MHz): l 1.25, 1.33, 1.39, 1.47, 1.50 and 1.51
(6s, 6×3 H, 3×C(CH ) ), 1.65 (br s, 1 H,
3
2

212
A.H. Haines, A.J. Lamb / Carbohydrate Research 321 (1999) 197–213
OH), 2.69–3.19 (complex, 2 H), 3.22 (s, 3 H,
gluco-undec-5-enofuranos-7-ulose (30).—(a)
Using N-iodosuccinimide (NIS): a soln of 30
(78 mg, 0.15 mmol) in anhyd CH Cl (5 mL)
CH O), 3.84–4.18 complex 6 H), 4.30 (ddd, 1
3
H, J9,10 5.6, J10,11 5.6, J10,11% 5.6 Hz, H-10), 4.52
2
2
(
d, 1 H, JAB 11.9 Hz, C H CH H ), 4.65 (d, 1
containing NIS (129 mg, 0.57 mmol) was
6
5
A
B
H, J8,9 4.0 Hz, H-8), 4.74 (d, 1 H,
C H CH H ), 6.02 (d, 1 H, J 2.5 Hz, 1-H),
stirred under N at rt. TLC (solvent B) indi-
2
cated that after 2 weeks, only starting material
was present; the mixture was filtered through
Kieselguhr and then concentrated to give 30
(59 mg, 76%).
6
5
A
B
1,2
7
5
.26–7.38 (complex, 5 H, C H ); EIMS: m/z
6
5
37 (0.3, M−15,), 534 (0.3, M−18).
b) Using hydrogen peroxide and potassium
carbonate in MeOH: treatment of 30 (41 mg,
(
(b) Using iodine: a soln of 30 (54 mg, 0.11
mmol) in anhyd THF (2 mL) containing re-
sublimed iodine (0.14 g, 0.55 mmol) and
8
1 mmol) with aq 30% H O (0.05 mL) in
2
2
MeOH (2 mL) and water (1 mL) at 0 °C
containing K CO led to loss of starting mate-
NaHCO (0.10 g, 1.19 mmol) was stirred at rt.
2
3
3
rial, (TLC). Product isolation gave crude ma-
TLC (solvent B) indicated that only starting
material was apparent after 6 days. The mix-
ture was filtered through Kieselguhr and parti-
tioned between water (5 mL) and CH Cl (5
1
terial which by H NMR spectroscopy was
consistent with it containing four possible iso-
1
mers of 32, 33: H NMR (300 MHz): l 1.26–
2
2
1
1
3
5
4
7
.55 (complex, 18 H, 3×C(CH ) ), 1.60 (br s,
mL). The dried organic solution was concen-
trated to yield 30 (21.6 mg, 40%).
3
2
H, OH), 2.77–5.29 (complex 12 H,), 3.22,
.24, 3.28 and 3.29 (4×s, 3 H total, CH O),
6-Deoxy-1,2:8,9:10,11-tri-O-isopropylidene-
3
.86, 5.90, 6.01, and 6.22 (4×d, 1 H total, J1,2
.1, 3.9, 3.9, and 3.7, respectively, H-1), 7.20–
.45 (complex, 5 H, C H ). No products due
L
-arabino-h- -gluco-undecofuranos-7-ulo-7,3-
D
pyranose (36).—(a) Using H and palladium
2
black, catalysed by a trace amount of acid: to
a soln of 24 (307 mg, 0.59 mmol) in anhyd
MeOH (5 mL) containing palladium black (50
mg) was added a 0.014 M soln of HCl in
anhyd MeOH (1.5 mL). The reaction mixture
6
5
to Favorskii rearrangement were observed in
the crude mixture.
(c) Using hydrogen peroxide and NaHCO3
in THF [9]: treatment of 30 (195 mg, 0.39
mmol) with aq 30% H O (0.1 mL) in a THF
was stirred under H at rt and atmospheric
2
2
2
(
0
5 mL)/satd aq NaHCO (0.1 mL) mixture at
pressure for 4 h after which time TLC (1:1
light petroleum–EtOAc) indicated that all
3
°C gave [TLC (solvent B)] a product running
as a single component in an approximately
equal proportion to remaining starting mate-
rial. Column chromatography (14:1 toluene–
acetone) gave initially 30 (86 mg, 44%) and a
new component (11 mg, 5%), tentatively iden-
starting material (R 0.72) had reacted to give
f
a single product (R 0.47). The suspension was
f
filtered through a pad of Kieselguhr and the
filtrate, which was slightly acidic to moist
indicator paper, was stirred with Amberlite
1
−
tified, on the basis of its IR and H NMR
IRA-400 ion-exchange resin (HO form; 5
spectra, as a mixture of 6- and/or 7-C-carboxy
derivatives of a stereoisomer of 3-O-benzyl-6-
mL) for 1 h. After removal of the resin by
filtration, the solution was concentrated to a
deoxy-1,2:7,8:9,10-tri-O-isopropylidene-
L
-
colourless oil. Dissolution of this oil in Et O
2
deca-1,4-furanose arising from the Favorskii
rearrangement of the intermediate epoxide [8];
and subsequent removal of the solvent under
vacuum resulted in crystallisation to give a
single anomer of the title product 36 (201 mg,
IR (film): w 3400 (br, CO H), 1750–1600
2
−
1
1
cm (CꢁO and C H ); H NMR (300 MHz):
79%) mp 145.5–148°C; [h] +11.9° (c 1.0,
6
5
D
l 1.22–1.54 (complex, 18 H, 3×C(CH ) ),
CHCl ); IR (Nujol): w 3400 (br, OH), 1380
3
2
3
−
1
1
1
.59 (br s, 1 H, OH), 1.70–2.02 (complex, 2
and 1375 cm
(C(CH ) ); H NMR (270
3
2
H, H-6, 6%), 3.78–4.79 (complex, 10 H), 5.52–
MHz): l 1.33, 1.35, 1.36, 1.38, 1.48 and 1.52
(6×s, 6×3 H, 3×C(CH ) ), 1.84 (dd, 1 H,
5
5
.58 (complex, 1 H, H-1), 7.10–7.38 (complex,
H, C H ), 8.55 (CO H).
3
2
J5,6 4.6 and J 12.6 Hz, H-6), 2.02 (ddd, 1 H,
6
5
2
6,6%
4
Attempted iodonium-induced cyclisation re-
action on (E)-3-O-benzyl-5,6-dideoxy-1,2:8,9:
0,11 - tri - O - isopropylidene - - erythro - h -
J5,6% 11.9, J
2.0 Hz, H-6%), 2.13 (d, 1 H,
6%,8
J5,OH 10.6 Hz, 5-OH), 3.76 (d, 1 H, J ꢀ0,
8,9
1
L
D-
J9,10 7.6 Hz, H-9), 4.02–4.30 (complex, 5 H,

A.H. Haines, A.J. Lamb / Carbohydrate Research 321 (1999) 197–213
213
Antibiot., 35 (1982) 1711–1714. (f) H. Yoshikawa, Y.
Takiguchi, M. Terao, J. Antibiot., 36 (1983) 30–35.
[4] (a) N.J. Newcombe, M.F. Mahon, K.C. Molloy, D.
Alker, T. Gallagher, J. Am. Chem. Soc., 115 (1993)
H-5, 10 11, 11%, 7-OH), 4.31 (d, 1 H, J ꢀ0,
,4
2
,3
J3 2.0 Hz, H-3), 4.40 (dd, 1 H, J3,4 and J4,5
3
4
.0 Hz, H-4), 4.56 (d, 1 H, J1,2 3.6 Hz, H-2),
13
6430–6431. (b) P.J. Cox, A.M. Griffin, N.J. Newcombe,
.81 (d, 1 H, H-8), 5.87 (d, 1 H, H-1);
C
S. Lister, M.V.J. Ramsay, D. Alker, T. Gallagher, J.
Chem. Soc., Perkin Trans. 1, (1995) 1443–1447. (c) H.M.
Binch, T. Gallagher, J. Chem. Soc., Perkin Trans. 1,
(1995) 401–402. (d) H.M. Binch, A.M. Griffin, S.
Schwidetzky, M.V.J. Ramsay, T. Gallagher, F.W. Licht-
enthaler, J. Chem. Soc., Chem. Commun., (1995) 967–
968. (e) J.R. Bearder, M.L. Dewis, D.A. Whiting,
Synlett, (1993) 805–806. (f) J.R. Bearder, M.L. Dewis,
D.A. Whiting, J. Chem. Soc., Perkin Trans. 1, (1995)
NMR (67.9 MHz): l 25.0, 26.3 (×2), 26.4,
6.8 and 27.0 (3×C(CH ) ), 35.2 (C-6), 63.8
2
3
2
(
(
1
4
C-11), 67.9, 74.9, 75.7, 76.5, 77.4, 84.4, 84.8
C-2, 3, 4, 5, 8, 9, 10), 96.2 (C-7), 105.1 (C-1),
09.6, 110.4 and 112.2 (3×C(CH ) ); EIMS:
3
2
17 (2.6, M−15), 399 (0.7, M−33), 289 (1.2,
M−143), 143 (17.2, M−289); HRMS: calcd
for C H O (M−15): 417.1761; found: m/z
227–233. (g) F. Emery, P. Vogel, Tetrahedron Lett., 26
1
9
29 10
(1993) 4209–4212. (h) F. Emery, P. Vogel, J. Org.
Chem., 60 (1995) 5843–5854. (i) A.J. Fairbanks, E. Per-
rin, P. Sinay, Synlett, (1996) 679–681.
4
17.1760. Anal. Calcd for C H O : C, 55.55;
2
0
32 10
H, 7.5. Found: C, 56.0; H, 7.85.
[
5] (a) K. Narkunan, M. Nagarajan, Indian J. Chem., Sect.
B, 32 (1993) 607–608. (b) K. Narkunan, M. Nagarajan,
J. Org. Chem., 59 (1994) 6386–6390.
[
6] (a) I. Paterson, R.D. Tillyer, J. Org. Chem., 58 (1993)
Acknowledgements
4182–4184. (b) A.S. Franklin, I. Paterson, Cont. Org.
Synth., 1 (1994) 317–338. (c) C.J. Cowden, I. Paterson,
Org. React., 51 (1997) 1–200.
7] (a) J.G. Buchanan, M. Chac o´ n-Fuertes, A Stobie, R.H.
A.J.L. thanks the University of East Anglia
for financial support in the form of a stu-
dentship. The authors thank the EPSRC Mass
Spectrometry Service Centre, Swansea, for de-
termination of the high-resolution mass
spectra.
[
Wightman, J. Chem. Soc., Perkin Trans. 1, (1980) 2561–
2
566. (b) G. Aslani-Shotorbani, J.G. Buchanan, A.R.
Edgar, P.K. Shahidi, Carbohydr. Res., 136 (1985) 37–52.
c) M. Seepersaud, M. Blumenstein, D.R. Mootoo, Te-
(
trahedron, 53 (1997) 5711–5724.
[
8] (a) H.O. House, W.F. Gilmore, J. Am. Chem. Soc., 83
(
1961) 3972–3980. (b) R.W. Mouk, K.M. Patel, W.
Reusch, Tetrahedron, 31 (1975) 13–19.
[
9] S. Danishefsky, R. Zamboni, M. Kahn, S.J. Etheredge,
J. Am. Chem. Soc., 103 (1981) 3460–3467.
References
[
10] M.L. Wolfrom, A. Thompson, Methods Carbohydr.
[
1] (a) T. Iwasa, T. Kusaka, K. Suetomi, J. Antibiot., 31
Chem., 2 (1962) 65–68.
[11] M.A. Bukhari, A.B. Foster, J. Lehmann, J.M. Webber,
J. Chem. Soc., (1963) 2287–2290.
(
1978) 511–518. (b) S. Harada, T. Kishi, J. Antibiot., 31
1978) 519–524. (c) S. Harada, E. Mizuta, T. Kishi,
(
Tetrahedron, 37 (1981) 1317–1327.
2] (a) A. Takatsuki, K. Arima, G. Tamura, J. Antibiot., 24
[12] G.J.F. Chittenden, Carbohydr. Res., 84 (1980) 350–352.
[13] G.J.F. Chittenden, Carbohydr. Res., 222 (1991) 283–287.
[14] D.Y. Jackson, Synth. Commun., 18 (1988) 337–341.
[15] T.D. Inch, R.V. Ley, P. Rich, J. Chem. Soc., (C), (1968)
1683–1692.
[16] L.F. Wiggins, J. Chem. Soc., (1946) 13–14.
[17] A.S. Meyer, T. Reichstein, Hel6. Chim. Acta, 29 (1946)
152–162.
[
(
1971) 215–223. (b) A. Takatsuki, G. Tamura, J. An-
tibiot., 24 (1971) 224–231. (c) A. Takatsuki, G. Tamura,
J. Antibiot., 24 (1971) 232–238. (d) A. Takatsuki, K.
Kawamura, M. Okina, Y. Kodama, T. Ito, G. Tamura,
Agric. Biol. Chem., 41 (1977) 2307–2309.
[
3] (a) M. Arai, T. Haneishi, N. Kitahara, R. Enokita, K.
Kawakubo, Y. Kondo, J. Antibiot., 29 (1976) 863–869.
[18] R.C. Anderson, B. Fraser-Reid, J. Org. Chem., 50 (1985)
4781–4786.
(
b) T. Haneishi, A. Terahara, H. Kayamori, J. Yabe, M.
Arai, J. Antibiot., 29 (1976) 870–875. (c) Y. Takiguchi,
H. Yoshikawa, A. Terahara, A. Torikata, M. Terao, J.
Antibiot., 32 (1979) 857–861. (d) Y. Takiguchi, H.
Yoshikawa, A. Terahara, A. Torikata, M. Terao, J.
Antibiot., 32 (1979) 862–867. (e) A. Terahara, T.
Haneishi, M. Arai, T. Hata, H. Kuwano, C. Tamura, J.
[19] R.U. Lemieux, J. Howard, Can. J. Chem., 41 (1963)
308–316.
[20] T. Iwashige, H. Saeki, Chem. Pharm. Bull., 15 (1967)
1803–1806.
[21] T. Anthonsen, S. Hagen, W. Lwande, Acta Chem.
Scand., Ser. B, 34 (1980) 41–45.
.