Catalytic tin radical mediated tricyclisations. Part 1. Monocyclisation studies

Article abstract of DOI:10.1039/b000661k

A general strategy for catalytic tin radical mediated, radical cascade reactions is proposed in which three rings are constructed in a single step. The initial step in the tricyclisation process has been examined using 2,3-dideoxy-α-D-erythro-hex-2-enopyr

Full text of DOI:10.1039/b000661k

View Article Online / Journal Homepage / Table of Contents for this issue
Catalytic tin radical mediated tricyclisations. Part 1.
Monocyclisation studies†
David R. Kelly* and Mark R. Picton
1
Department of Chemistry, Cardiff University, PO Box 912, Cardiff, UK CF10 3TB
Received (in Cambridge, UK) 24th January 2000, Accepted 16th March 2000
Published on the Web 27th April 2000
A general strategy for catalytic tin radical mediated, radical cascade reactions is proposed in which three rings
are constructed in a single step. The initial step in the tricyclisation process has been examined using 2,3-dideoxy-
α--erythro-hex-2-enopyranosides bearing unsaturated substituents at the 1-O and/or 4-O-positions. Substrates
for cyclisation of substituents at the 1-O-position were prepared by a novel zinc chloride catalysed Ferrier
rearrangement of tri-O-acetyl--glucal with unsaturated alcohols, whereas substrates for cyclisation of substituents
at the 4-O-position were prepared by alkylation or acylation of ethyl 6-O-protected 2,3-dideoxy-α--erythro-hex-2-
enopyranosides. Propargyl substituents cyclise efficiently, but propenyl substituents less so. Propioloyl substituents
undergo hydrostannylation without cyclisation.
Tin hydride reagents are often rightly criticised because
reagents such as tri-n-butyltin hydride1 (TBTH; MW 291) and
triphenyltin hydride (TPTH; MW 351) have a low equivalence
and consequently a large amount of spent and excess reagent
has to be separated from the product. The higher equivalence
of trimethyltin hydride (MW 165) is an advantage, however
trimethyltin compounds are extremely toxic and hence they
must be handled with elaborate precautions. Tin hydride
reagents can be used catalytically if a co-reductant is present
(e.g. sodium borohydride2 or polymethyl hydrosiloxane
(PMHS), potassium fluoride3), but this reduces their inherent
chemoselectivity.
This situation has resulted in a wealth of prescriptions for
removing “tin residues”,4 and the development of alternative
reagents such as the fluorous5 and polar6 derivatives which can
be removed by extraction. The fluorous reagents have not been
widely adopted as yet, but the benefits for clean extraction of
the reagents are manifest. However, the molecular weights are
even higher than the hydrocarbon analogues, which reduces
their effectiveness on a weight for weight basis.
On the other hand, organotin radicals are excellent radical
abstractors and tin hydrides are excellent hydrogen radical
donors to alkyl radicals (vide infra), which results in reactions
of high selectivity and fidelity.7 It would be of great benefit if
the chemoselectivity of radicals generated using tin hydrides
could be utilised in a wholly catalytic process.
Results and discussion
This a priori requires a process in which the tin radical adds
to a system, effects some change and is then released. We envis-
aged a process in which a tin radical adds to an unsaturated
system 1 and initiates a series of addition reactions such that
the radical is translocated to a position at which it can cause
elimination of the stannyl radical (cf. 3). Since radicals do not
easily effect SH reactions at saturated centres8 this requires that
the stannyl group is temporarily attached to an unsaturated
moiety and substitution proceeds via radical addition followed
by β-elimination.9 Since the tin radical must undergo addition
to a moiety of the system which is unsaturated both before and
after the addition of the tin radical, an alkyne group is one
possibility. The next requirement is to construct a series of
unsaturated bonds capable of translocating the radical back to
the desired position. 5-exo-trig Radical cyclisations are an
† Full experimental and spectroscopic data for all compounds prepared
by methods other than free radical cyclisation are available as sup-
plementary data. For direct electronic access see http://www.rsc.org/
suppdata/p1/b0/b000661k/ (this includes the following compounds in
order of appearance: 7a, 8a, 7b, 7d, 10e, 7c, 8c, 7d, 7f, 7g, 7h, 10h, 7f, 9,
13a, 13b, 13c, 13d, 14, 18, 21, 22, 23, 25b, 25c, 26, 28, 29b, 31a, 31b, 33a,
33b, 35a, 35b, 35c, 36b).
Scheme 1
DOI: 10.1039/b000661k
J. Chem. Soc., Perkin Trans. 1, 2000, 1559–1569
This journal is © The Royal Society of Chemistry 2000
1559

View Article Online
obvious choice because of their predictable stereochemistry and
reliability.
that the proposal in Scheme 1 could be rapidly tested and
moreover the products would be single enantiomers. In prac-
tice, this sequence requires deprotection of the unsaturated
sugar 7a, selective derivatisation at the more accessible primary
6-hydroxy group, followed by 4-O-substitution. The alkynyl
alkene 7a could also be used to test the single step compon-
ents of the multiple cyclisation. Unsaturated sugars have been
heavily investigated by Fraser-Reid19 and others20 and
exploited in innumerable free radical cyclisations21 and total
syntheses.
It was anticipated that the propiolic ester moiety would react
preferentially with tri-n-butyltin radical due to the strongly elec-
tron withdrawing effect of the carboxy group. A propiolic ester
would also provide the correct side chain length to allow a
favourable initial 5-exo-trig cyclisation. Similar considerations
apply to a propargyl ether at the anomeric centre, but a
propiolic ester at this position would be too reactive to survive
the intermediate synthetic manipulations. Although it was
envisaged that attack of the radical should occur initially at the
propiolic ester, this is not a requirement for success. The reverse
sequence of radical translocations would also give the same
product.
Any number of radical translocations via additions to
alkene bonds will translocate the radical by an even number of
carbons. Hence the final cyclisation at the original site of
attack must always be via a cyclisation with an even number
of carbons between the radical and the original site of attack.
In the current case the final desired cyclisation is 6-endo-trig,
with the possibility of an undesirable 5-exo-trig cyclisation.10 It
was envisaged that the desired regiochemistry of cyclisation
could be enforced by the geometry of the vinyl radical and the
rigidity of the ring system created in the earlier steps. Radical
cascades are now commonly used in synthesis,11,12 but processes
in which the initiating radical is subsequently eliminated are
rare. Subsequent to our preliminary disclosure13 there have
been only three other examples as far as we are aware. Pancrazi
and co-workers14 achieved elimination of the initiating tri-n-
butyltin radical in an alkyne–alkene monocyclisation as did
Spino and Barriault with the tricyclisation of an acyclic
alkynone diene.15 Marco-Contelles accomplished tricyclisation
of an alkynic diene sugar16 closely resembling our original
report. Palladium catalysed cyclisation of similar substrates has
only yielded bisannulation products thus far.17
This publication describes our initial foray into this area and
some model studies with monocyclisations.
We conceived that the idea summarised in Scheme 1 could be
implemented by appending two unsaturated side chains to a
rigid ring system. This would hold the side chains in the correct
orientation to favour the formation of single stereoisomers. The
combination of a Ferrier rearrangement18 of an acylated glucal
e.g. 6a (Scheme 2), followed by attachment of an unsaturated
chain at the 4-O-position (to give e.g. 9) offered the prospect
Ferrier rearrangements are usually performed with boron
trifluoride–diethyl ether as the Lewis acid catalyst. Photo-
induced electron transfer,22 lithium tetrafluoroborate,23 ceric
ammonium nitrate,24 iodine,25 and NIS treatment of 3-O-
pentenyl glycals26 have been also been advocated, but only the
last two have found wide acceptance. In our experience, zinc
chloride gives cleaner products which require less purification,
than boron trifluoride–diethyl ether.27 Thus tri-O-acetyl--
glucal 6a was treated with propargyl alcohol and zinc chloride
under otherwise standard Ferrier reaction conditions. As with
the boron trifluoride–diethyl ether catalysed reaction, complete
conversion is signalled by the sudden appearance of a purple
colouration. Quenching and work up of the reaction at this
time afforded the propargyl α-glycoside 7a in good yield (87%).
TLC analysis of the product indicated a small amount of
1
the β-glycoside 8a which was confirmed by H-NMR of the
crude product to be present in the ratio 89:11, 7a:8a. Delay
in quenching the reaction resulted in the formation of an
approximately equal amount of α- and β-glycosides and a dis-
coloured crude product. The diacetates 7a, 8a were readily con-
verted to the corresponding diols 7b, 8b by reaction with
sodium methoxide in methanol. Selective protection of the
primary C-6 hydroxy group was then required to permit acyl-
ation at the secondary C-4 hydroxy group.
Reaction of the diols 7b, 8b with pivaloyl chloride, triethyl-
amine and DMAP28 at Ϫ30 ЊC gave a complex mixture, which
was separated by column chromatography to give the dipival-
ate 7d (25%), the furan 10e (22%), the desired monopivalate 7c
(6%), the monopivalate β-glycoside 8c (2%) and recovered
starting material 7b (2%). The collected unseparated frac-
tions contained an approximately 50:50 mixture of the
monopivalate anomers 7c, 8c (12%) to give a total recovery
of 69%.
1
The H-NMR spectrum of the furan 10e in deuteriochloro-
form showed coincident chemical shifts for the protons
attached to C-6, but these were resolved in benzene-d6. 1H–1H-
COSY experiments indicated two isolated spin systems each
containing three protons (δ 4.79, 4.41, 4.36; 7.11, 6.18, 6.12;
C6D6). The upfield spin system was readily identified as a
methine adjacent to a methylene group with vicinal and gem-
inal coupling constants (J 4.8, 6.5 and 11.4 respectively) typical
of those expected for 5-H and 6-H2 of a hexopyranose. The
downfield spin system had rather small vicinal coupling con-
3
stants (3J1,2 ca. 1 Hz, J2,3 3.1 Hz) which are typical of furans.
This material failed to react with tetracyanoethylene at room
temperature which excludes a number of other conceivable
Scheme 2
1560
J. Chem. Soc., Perkin Trans. 1, 2000, 1559–1569

View Article Online
diene structures. The furan 10e is a known compound which
was prepared by pivaloylation of the diol 10b.29 The reported
1H-NMR spectrum of this compound in deuteriochloroform
has couplings which are similar to those reported here, but the
chemical shifts for the ring protons are some 0.4–0.5 ppm
upfield, relative to those which we measured in the same sol-
vent. The signal for 5-H has a chemical shift (δ 4.86) which is
almost identical to that which we measured, but 6-H2 is
reported at the inconceivable value of δ 3.1 (δ 4.38, our work).
Comparison of our 1H-NMR and 13C-NMR data for the furan
10e with those reported for furfuryl alcohol30 reveals only
minor differences and hence we conclude that errors were made
in the cited work. The furan diol 10b has been prepared by
mercuric sulfate31 or copper sulfate32 catalysed rearrangement
of -glucal 6b and the 6-O-silyl furan 10j by treatment of the
pyranoside 7i with iodine.33 Presumably, triethylamine hydro-
chloride or an acylium species acts as a Lewis acid in a similar
way on the diol 7a or the monopivaloate 7c.
More practical results were obtained when the diol 7b was
reacted with pivaloyl chloride in pyridine at room temper-
ature34 to give a mixture of the monopivalate 7c (41%) and the
dipivalate 7d (22%) plus the β-anomer of the monopivalate
8c (0.3%).
We next turned to the tert-butyldimethylsilylation as a means
to achieve selective 6-O-protection. Treatment of the diol 7b
with tert-butyldimethylsilyl chloride, triethylamine, DMAP and
dichloromethane (Hernandez’s conditions35) gave a mixture
comparable to that produced by pivaloylation. Column chrom-
atography yielded the 6-O-silyl ether 7f (26%), the 4-O-silyl
ether 7g (4%), the 4,6-di-O-silyl ether 7h (8%), recovered start-
ing material 7b (4%) and a trace of the furan disilyl ether 10h.
However, by using tert-butyldimethylsilyl chloride and imid-
azole in DMF (Corey’s conditions36) at room temperature an
easily separated mixture of the 6-O-silyl ether 7f (75%) and
starting material 7b (7%) was produced in only 6 hours.
Scheme 3
trans-stannyl acrylates (δ 7.22, 6.65 J 16 and δ 7.73, 6.28, J 22
respectively).
This reaction although disappointing was instructive. The
regiochemistry of stannyl radical addition had proceeded as
predicted and the presence of cis- and trans-isomers indicates
that stannyl radicals had been generated efficiently. The lack of
cyclisation could be attributed to the intermediate vinyl radical
having an inappropriate conformation, however, given the poor
yield of the hydrostannylation product 12, there was also the
suspicion that the dialkynyl alkene 9 or a cyclisation product
might have decomposed before it could be isolated. We there-
fore turned our attention to a simpler analogue, which would
enable us to study the cyclisation of the 4-O-propiolate alone
and provide a route to α-methylene-γ-lactones.
Simple propiolic esters can be easily prepared by treatment of
propiolic acid with an excess of the alcohol,37 but this is not
applicable to valuable alcohols. Moreover propiolyl chloride is
highly unstable,38 and indeed we found that when propiolic acid
was refluxed with thionyl chloride, no acid chloride could be
detected by IR analysis. Consequently, we decided to investigate
the DMAP catalysed DCC esterification.39,40 Addition of the
alcohol 7f and propiolic acid to DMAP and DCC resulted in
polymerisation However dropwise addition of a solution of
DMAP and DCC to a mixture of a large excess of propiolic
acid and the alcohol 7f at 0 ЊC minimised polymerisation.41 The
reaction mixture containing the propiolate 9 was highly
unstable. Removal of solvent at >30 ЊC caused the yellow
tinged ester solution to rapidly change to dark green and the
single spot product became a multi-product mixture on analysis
by TLC. However by keeping the material below 30 ЊC, and
working quickly, a pure product could be isolated by column
chromatography, albeit in poor yield (10%). In other studies we
have found that the most efficient means for generating stannyl
radicals is TBTH with AIBN initiation. This of course means
that there is the potential for reduction before full cyclisation.
On the other hand use of hexabutylditin with various initiators
is much less efficient as has been noted by others.15
Monocyclisations of side chains at the 4-O-position
Reaction of tri-O-acetyl--glucal 6a with ethanol and zinc
chloride gave the glycoside 13a (67% yield; α:β ratio 89:11),
which upon treatment with sodium methoxide in methanol
readily underwent hydrolysis to give the diol 13b (91%
yield).18,22 Selective silylation of the primary hydroxy group
occurred without incident to give the 6-O-silyl ether 13c (49%
yield), plus starting material (28% yield) and a trace of the 4,6-
di-O-silyl ether 13d (3%). Treatment of the 6-O-silyl ether 13c
TBTH (1.3 equiv.) in benzene was added dropwise from a
syringe pump to a refluxing solution containing the enediyne 9
and AIBN (0.04 equiv.) under nitrogen.42 The crude reaction
mixture contained a complex mixture of products by TLC and
the 1H-NMR spectrum of the crude reaction mixture suggested
that one of the major reaction products was the hydrostannyl-
ation product 12. Column chromatography yielded an impure
fraction containing the major product plus tin residues. In the
1H-NMR spectrum the signal for the alkynic proton of the
propiolyl group (δ 2.9) was absent but that of the propargyl
group (δ 2.44) was present. Low field doublets with very
large coupling constants indicated the presence of cis- and
Scheme 4
with propiolic acid and a solution of DCC and DMAP gave a
complex mixture of products as before, from which the desired
product 14 was isolated in poor yield (10%) by column chrom-
atography. The presence of a propiolate ester was confirmed by
1H-NMR singlet signal for the alkynic proton (δ 2.92), the shift
of the signal for 4-H from δ 4.15 in the starting material 13c to
J. Chem. Soc., Perkin Trans. 1, 2000, 1559–1569
1561

View Article Online
1
δ 5.40 and a very strong acetylenic stretch at 2100 cmϪ1 in the
IR spectrum.
the H-NMR spectrum was a broad doublet (δ 5.77, J 1.9,
13-H) with coupling to tin (2JSn,H 57.2 Hz) although the reso-
lution was insufficient to distinguish 117Sn and 119Sn couplings.
The second furthest downfield signal was assigned to the ano-
meric proton, 4-H (δ 4.77, dd, J 5.9, 5.9), which was used as the
origin for the assignment of a 1H–1H-COSY NMR experiment.
1-H appeared as an apparent triplet (δ 3.89, dd, J 7.6, 7.6, with
a broadened centre line) and 6-H as a doublet of doublets of
doublets (δ 2.65, J 11.3, 6.3, 5.0). This indicates that the true
Treatment of the propiolate ester 14 with TBTH under
standard cyclisation conditions afforded a complex mixture of
products as determined by TLC intensities The 1H-NMR spec-
trum of the crude reaction mixture provided no evidence for the
presence of the desired cyclised product 15, however the dis-
tinctive alkenic signals of the vinyl stannane 16 were present.
Despite repeated chromatography, this product could not be
purified adequately and the only product actually isolated was
the silyl ether 13c (20% yield), albeit containing trace amounts
of impurities. Given that no starting material remained in the
crude reaction mixture, this seemingly indicates that the stannyl
ester 16 was cleaved during column chromatography. We attrib-
ute the lack of cyclisation products from both propiolates 9 and
14 to unfavourable conformational factors. α-Acrylate radicals
undergo reactions which are fairly typical of vinyl radicals and
organic radicals in general.43 However, allyl esters of propiolic
acid undergo radical addition induced cyclisation compar-
atively slowly44 compared to the corresponding ketones.45 This
observation and our results reflect the preference for the (Z)-
conformer over the (E)-conformer of the ester group (ca. 12 kJ
molϪ1) and the barrier to rotation of the corresponding radical
17.46 This conformational preference was circumvented by
3
value of J1,6 is about 6.3 Hz. Typical line broadening is about
1.5 Hz which is sufficient to meld couplings of 7.6 and 6.3 Hz
plus weak long range couplings into an apparent triplet. The
3
3
rough equivalence of J1,2 and J1,6 suggests at first sight that
both 1-H and 2-H and 1-H and 6-H might have a trans-diaxial
relationship (cf. 20) and this idea is reinforced by the large value
3
for Jax-5,6 (11.3 Hz). However molecular modelling and coup-
3
ling constant calculations with Haasnoot’s equation (19 J1,eq-6
3
5.9 Hz; 20 J1,ax-6 11 Hz), indicate cis-fusion with ax-5-H in a
position which is close to eclipsed with 6-H. The 6-O-trityl ana-
logue of compound 19 has been previously prepared by similar
means and was characterised by 1H-NMR and IR spectra. The
data is in satisfactory agreement with that reported here.47
With the propargyl cyclisation secured (albeit in low yield),
we turned to the corresponding allyl cyclisation. Allylation of
the silyl alcohol 13c under standard conditions (sodium
hydride, allyl iodide) gave a mixture of three products which
were separated by column chromatography (21, 8%; 22, 3%; 23,
using the corresponding propargyl ether 18, which was pre-
pared in good yield (57%) by treatment of the 6-O-silyl ether
13c with sodium hydride and propargyl bromide.
Initial attempts at the cyclisation gave mixtures of products
which were difficult to separate from “tin residues”. TLC moni-
toring indicated that the initial stages of the reaction occurred
with clean conversion, but in the later stages small amounts of a
host of by-products accumulated which made purification dif-
ficult. The intractable mixtures produced by prolonged reaction
times were avoided by running the reaction with incomplete
conversion, at the expense of reduced product yields, although
starting material could be recovered. This protocol was utilised
subsequently on many occasions. Cyclisation in toluene with
slow addition of TBTH gave a single product 19 plus starting
material 18 as determined by TLC. Purification by column
chromatography gave clean products. The lowest field signal in
21% yield). We had expected that the 4-O-tert-butyldimethyl-
silyl-6-O-allyl isomer of 23 might be formed due to silyl
migration, but this was not identified amongst the products. We
suspect that transsilylation occurred under the strongly basic
conditions to give the diol 13b plus the disilyl ether 13d and the
latter underwent retro-hydrosilylation to give the enone 22.
Treatment of the prop-2-enyl ether 23 with TBTH gave the
expected reaction mixture, and purification by column chrom-
atography afforded a single spot fraction containing equal
amounts of two components 24. The anomeric protons (δ 4.79,
0.5H, dd J 8.5, 5.6, 4-H and δ 4.74, 0.5H, dd, J 4.3, 4.3, 4Ј-H)
were clearly adjacent to two protons as required for the
cyclisation product and these could be located in a 1H-1H
J-COSY NMR spectrum. Rigorous analysis of 1H-NMR spec-
tra was not possible due to poor dispersion, nevertheless all of
the spectroscopic data was in agreement with the proposed
structure.
Monocyclisations of side chains at the 1-O-position
The compounds prepared in the course of the synthesis of the
dialkyne alkene 9 provided an opportunity to examine the
cyclisation of side chains attached to the anomeric position.
Treatment of the alkyne 7a with TBTH in refluxing toluene
under the standard conditions gave an excellent yield of the
1562
J. Chem. Soc., Perkin Trans. 1, 2000, 1559–1569

View Article Online
time was extended, (e.g. 1 hour) the β-anomer was detected. In
1
the H-NMR spectrum of the reaction mixture, signals for the
β- and α-anomeric protons appeared at δ 5.13, 5.08 (ratio
21:79). Attempted cyclisation of the α-anomer 28 with TBTH
gave a mixture of two products of very similar polarity. From
TLC intensities it appeared that the less polar component was
by far the major product and it was isolated by column chrom-
1
atography. H-NMR spectra showed signals characteristic of
the starting material except for the absence of the alkynic pro-
ton (δ 2.00), plus the presence of a tri-n-butylstannyl group and
two down field alkenic signals (δ 6.45, 5.86) with a common
coupling constant of 12 Hz.
This material was assigned structure 29b with a (Z)-stannyl-
alkene moiety, which results from kinetic hydrostannylation of
the alkyne group. The structure was verified by dissolving the
adduct 29a in wet ether containing silica gel (70–230 mesh) to
give the alkene 29b which was also prepared independently by
Ferrier rearrangement of tri-O-acetyl--glucal 6a with but-3-
en-1-ol. Conventional silica gel flash column chromatography
failed to separate the minor product 30 from the adduct 29a,
however chromatography using silica gel impregnated with
silver nitrate gave an impure fraction containing the minor
product (<5 mg). 1H-NMR analysis indicated the presence of a
single alkenic proton (δ 6.01, s, 11-H) consistent with the struc-
ture of the bicyclic adduct 30. The Chapleur group, isolated the
bicyclic stannane 30 (21%, Z:E, 3.5:1) and the vinyl stannanes
29a (59%, Z:E, 18:82).47 The thermodynamic ratio of vinyl-
stannane stereoisomers 29a obtained in this work is surprising,
because the equilibration of non-conjugated vinylstannanes is
usually only observed after protracted reaction times and/or
with high concentrations of stannyl radicals.52
vinyl stannane 25a (65%) as previously reported by Chapleur’s
group47 in refluxing benzene. Free radical induced cyclisation
with diphenylphosphine is reported to give a mixture of cyclic
(37%) and acyclic products (29%).48 The reaction was moni-
1
tored by H-NMR which indicated steady conversion to the
vinylstannane 25a by the disappearance of the characteristic
signals for the alkenic protons at C-2, C-3 (δ 5.92, 5.85) and the
alkyne (δ 2.48) and the appearance of a single alkenic signal at
δ 5.88. Iodo- and protio-destannylation with iodine and wet
silica gel gave the iodoalkene 25b and the alkene 25c respect-
ively. In each of these three compounds the 1H-NMR signal for
6-H was too broad to interpret, however 1-H appeared as a
clean doublet J 4.6, 4.6 and 6.2 Hz (for 25a, b, c respectively),
indicating cis-fusion. The alkene 25c was also prepared
independently by cyclisation of the vinyl chloride 26 in reflux-
ing toluene. This proceeded in good chemical yield (50%), con-
sidering that the abstraction step is not particularly facile, and
as expected starting material 26 was recovered (8%). Encour-
aged by these comparatively good yields we attempted to trap
the radical intermediate formed by the alkyne 7a with allyl tri-n-
butyltin (2.5 equiv.)49 using TBTH (10 mol%) as a propagation
catalyst and AIBN initiation (3 mol%). The allylation product
27 was not formed, but the monocyclisation product 25a was
produced (28%), plus starting material 7a (10%). Only part of
the reduced product can be accounted for by TBTH added to
the reaction. The most likely explanation for the remainder is
retro-hydrostannylation of allyl tri-n-butyltin to give allene and
TBTH; a process which is well precedented for crotyltri-n-
butyltin.50
We also probed the 6-exo-trig mode53 of cyclisation by the
attempted cyclisation of the bromopropyl glycoside 31a. Com-
parable cyclisations of bromoethyl glycosides give good
yields.49,54 Treatment of the bromopropyl glycoside 31a with
TBTH under a wide range of experimental conditions (tem-
perature, high dilution) gave exclusively the reduced product
31b (77% yield), which was also prepared independently by
Ferrier rearrangement. The reluctance of radicals to undergo
6-exo-trig cyclisation is well known and these results confirm
that such a cyclisation is unlikely to be a usable component
of the proposed tricyclisation on the structural framework
(Scheme 1).
Given the success of these cyclisations, we next attempted
to apply them to the next higher homologue. Zinc chloride
catalysed Ferrier reaction of tri-O-acetyl--glucal 6a with but-
3-yn-1-ol gave exclusively the butynyl α-glycoside 28 in 30 min-
utes in excellent yield, compared with the boron trifluoride–
diethyl ether catalysed reaction (31%).51 When the reaction
A propenyl glycoside would lack the necessary degree of
unsaturation required to both initiate and terminate the tandem
cyclisation process. However a propenyl group could readily be
used in the final cyclisation step if the tandem process was
initiated by an alkyne. The propenyl 33a and 2-methylpropenyl
glycosides 33b were prepared by Ferrier rearrangement as
before and subjected to cyclisation with TBTH. The propenyl
glycoside 33a yielded two products with similar mobilities on
TLC plus a plethora of minor products. Repeated chromato-
graphy eventually yielded material of adequate purity which
was assigned the structures 34a, b. Although the anomeric pro-
tons (δ 5.45, 5.35) could be easily assigned the remainder of the
spectrum was too complex to permit rigorous interpretation.
Similarly the 2-methylpropenyl glycoside 33b reacted with
TBTH to give a complex mixture of products. Separation by
column chromatography gave a fraction containing three close
running components in the ratio 50:35:15 as determined from
J. Chem. Soc., Perkin Trans. 1, 2000, 1559–1569
1563

View Article Online
as a clean pair of doublet of doublets with a large geminal
coupling (δ 2.78, dd, J 15.7, 5.7, ax-5-H; δ 2.63 dd, J 15.7, 2.7,
eq-5-H).
Conclusions
The results clearly demonstrate that 1-O and 4-O-propargyl
substituents act as good radical traps for tri-n-butyltin radicals
and the intermediate vinyl radicals undergo the anticipated
5-exo-trig cyclisation. Radical trapping by alkenyl substituents
is much less effective, but nevertheless cyclised products are also
obtained, albeit in low yields. In contrast, propiolyl substituents
exclusively undergo hydrostannylation, due to a preference for a
conformation which is unfavourable to cyclisation. The follow-
ing paper in this issue describes the successful exploitation of
these observations.
1
the integration of the anomeric protons in the H-NMR spec-
trum (δ 5.39, 5.31, 5.24 all d J 4.8). No alkenic signals were
present. A H–1H J-COSY NMR experiment enabled the 1-H
1
signals to be correlated with the 6-H signals and other tentative
assignments to be made. The 3J1,6 coupling constant is consist-
ant with cis-ring fusion as observed previously and the two
major components are assigned as the epimers 34c, d. The
structure of the third component remains unassigned.
We gratefully acknowledge sponsorship of this work by
Warner-Lambert, Parke-Davis and the provision of spectro-
meter time by the EPSRC at the National Mass Spectrometry
Service at the University of Swansea.
The work thus far established that both the propargyl glyco-
side 7a and the chloropropenyl glycoside 26 readily undergo 5-
exo-trig cyclisation, whereas alkenes 33a and 33b react poorly by
this cyclisation mode. Difficulties in isolating closely running
mixtures of epimers necessarily reduced the isolated yields. The
poor 6-exo-trig cyclisation of the butynyl glycoside 28 indicates
that cyclisations homologous to those originally proposed are
unlikely to be viable.
Experimental
General
Purified or dried solvents were freshly distilled under an argon
or nitrogen atmosphere from a suitable drying agent.
All reactions were monitored by thin layer chromatography
(TLC) using Merck aluminium backed precoated silica gel
plates (0.2 mm, 60, F254) with UV light or ethanolic phospho-
molybdic acid (3%) and heat for visualisation. Virtually all
products were purified by flash column chromatography using
Merck silica gel 60 (70–230 mesh), eluted with a gradient start-
ing with a low polarity solvent and then increasing amounts of
a more polar solvent. All products were homogeneous as
judged by TLC unless stated otherwise.
Unsubstituted organic radicals react more readily with elec-
tron deficient alkenes than electron rich alkenes. Therefore it
was anticipated that the enones 35 might cyclise even more
Infra red (IR) spectra were recorded on a Perkin-Elmer 1600
Series FTIR spectrophotometer, using sodium chloride cells.
Elemental analyses were performed on a Perkin-Elmer 240c.
Low resolution mass spectra were recorded on VG Trio 1 and
VG platform II spectrometers using electron impact (EI) or
chemical ionisation (CI, CH4). Some low resolution spectra,
CI–NH3 spectra and all accurate mass measurements were
recorded at the EPSRC Mass Spectrometry Centre at Swansea.
Mass spectra data for compounds with high molecular weights
(particularly those containing tin and halogens) were simulated
using the computer program HiMass.57 This calculates the
abundance of the ions in an ion cluster for a given elemental
formula. This is termed cluster analysis.
readily than the propargyl glycoside 7a which had given a 65%
yield of the bicycle 25a. This of course would require revision
of the original cyclisation plan, nevertheless it potentially
offered the opportunity to incorporate otherwise unfavourable
cyclisations into the cascade. Oxidation of the diol 7b with
manganese dioxide55 gave the desired enone 35a. However as
reported,56 treatment with the mildest bases or even storage
initiated a retro-aldol reaction and decomposition. This pre-
cluded protection at the 6-O-position. However manganese
dioxide oxidation of the 6-O-tert-butyldimethylsilyl 7b and
pivalate 7c derivatives gave the desired enones 35b, c in poor to
fair yields.
NMR spectra were recorded on Perkin-Elmer R12B, Varian
T-60, Bruker AMX-360, and Bruker Advance DPX-400 spec-
trometers. CDCl3 was used as solvent unless indicated. Tetra-
methylsilane or residual solvent peaks (e.g. CHCl3) were used as
frequency standards. 13C-NMR spectra were recorded with full
and partial proton decoupling and using the DEPT technique.
Coupling constants were determined using the computer
program Multiplet58 and are quoted in hertz (Hz). Multiplet
uses peak positions from peak listings to calculate line spacings
which are averaged to give putative couplings. These in turn are
permutated to give possible coupling patterns. Thus the calcu-
lated coupling constants have an accuracy which is only limited
by the digital resolution of the NMR machine and line broaden-
ing effects. Values are reported to 0.1 Hz, but have an
uncertainty of ca. 0.3 Hz (at 360 MHz), due to the digital
resolution of the FID accumulation and Fourier transform-
Treatment of the 6-O-pivalate 7c with the cyclisation stand-
ard conditions yielded a complex mixture from which the vinyl-
stannane 36a was isolated in low yield. To facilitate further
purification this was treated with iodine to give the vinyl iodide
36b. The small amount of material obtained only enabled the
1
compound to be characterised by NMR. The H-NMR spec-
trum of the vinyl stannane 36a showed an alkenic peak (δ 5.88,
dd, J 2.5, 2.5) at an identical chemical shift to the correspond-
ing proton (δ 5.88, m) in the di-O-acetyl analogue 25a. As
anticipated, the 5-H protons which were well separated (δ 2.23
and 1.70) in the di-O-acetyl analogue 25a were shifted down
field to give a single multiplet (δ 2.77) in the ketone 36b. The
1H-NMR spectrum of the iodide 29b was better dispersed
than that of the stannane 29a. The 5-H protons now appeared
1
ation. H-NMR spectra were simulated using RACCOON.59
Molecular modelling was performed with PC-Model60 on a
Compusys, 33 MHz, 80486 PC. The program implements
Allinger’s MM2 force field, version MM88 with several
enhancements. Structures were routinely optimised using the
1564
J. Chem. Soc., Perkin Trans. 1, 2000, 1559–1569

View Article Online
1
randomise option with typically 200–500 trials. The PMR
option in PC-Model was used to calculate vicinal coupling
constants. This implements modified versions of the Karplus
equation61 parameterised to take account of the effect of sub-
stituents on coupling constants. An average absolute error of 1
Hz for all the vicinal coupling constants of a given molecule,
with no single value with an error of > 2 Hz constitutes a
satisfactory fit between experimental data and calculated
values. Vinylic 1H- and 13C-NMR chemical shifts were predicted
using Shoolery’s rules.62
impurities as judged from TLC. It was identified by H-NMR
as the alcohol 13c (10 mg, 20% yield), resulting from cleavage of
the β-stannyl propenoate 16. A more pure sample of the alco-
hol 13c (3 mg, 7% yield) was obtained in another run.
(1S,2S,4S,6R)-4-endo-Ethoxy-2-exo-tert-butyldimethylsilyloxy-
methyl-7-exo-(tri-n-butylstannylmethylene)-3,9-dioxabicyclo-
[4.3.0]nonane 19
Electronic data
The full experimental and spectroscopic data for all compounds
prepared by free radical cyclisation are described in the Experi-
mental section which follows. Data for compounds prepared by
other means, are provided in the supplementary data for this
paper. This includes the following compounds in order of
appearance: 7a, 8a, 7b, 7d, 10e, 7c, 8c, 7d, 7f, 7g, 7h, 10h, 7f, 9,
13a, 13b, 13c, 13d, 14, 18, 21, 22, 23, 25b, 25c, 26, 28, 29b, 31a,
31b, 33a, 33b, 35a, 35b, 35c, 36b.
Ethyl 6-O-tert-butyldimethylsilyl-4-O-prop-2-ynyl-2,3-dideoxy-
α--erythro-hex-2-enopyranoside 18 (100 mg, 0.30 mmol), and
AIBN (3 mg, 0.011 mmol, 0.04 equiv.) were dissolved in dry
toluene (2 ml) and refluxed under nitrogen for 15 min. A solu-
tion of tri-n-butyltin hydride (123 mg, 0.43 mmol, 1.3 equiv.) in
toluene (4 ml) was added over 3 hours via a syringe pump. The
reaction solution was refluxed for 16 hours. TLC analysis indi-
cated reaction product and starting material. Purification by
column chromatography, eluent hexane to 10% diethyl ether in
hexane afforded the title compound 19 (23 mg, 12%) and start-
ing material 18 (24 mg, 24%), both as clear oils; δH 5.77 (1H, br
Attempted preparation of the tetracycle 11. Identification of
vinylstannanes 12
2
d, J 1.9, JSn,H 57.2, 117Sn, 119Sn not resolved, 13-H), 4.77 (1H,
dd, J 5.9, 5.9, 4-H), 4.26 (1H, br d, J 13.0, 8a-H), 4.16 (1H, dd,
J 12.9, 2.0, 8b-H), 3.89 (1H, dd, J 7.6, 7.6, 1-H), 3.75 (2H, m,
2-H, 11a-H), 3.63 (2H, m, 10-H2), 3.36 (1H, dq, J 9.5, 7.1, 11b-
H), 2.65 (1H, br ddd, J 11.3, 6.3, 5.0, 6-H), 1.96 (1H, ddd,
J 14.2, 5.7, 5.7, 5a-H), 1.66 (1H, br m, 5b-H), 1.41 (6H, m, 22-
H6, Sn(CH2)2CH2-), 1.21 (6H, m, 21-H6, SnCH2-CH2), 1.10 (3H,
dd, J 7.1, 7.1, 12-H3), 0.83 (24H, m, 17-H3, 18-H3, 19-H3, 20-H6,
23-H9, (CH3)3CSi, SnCH2(CH2)2CH3), 0.1 (6H, s, 14-H3, 15-H3,
Prop-2-ynyl 2,3-dideoxy-4-O-propiolyl-6-O-tert-butyldimethyl-
silyl-α--erythro-hex-2-enopyranoside 9 (50 mg, 0.14 mmol)
and AIBN (0.2 mg, 0.06 mmol, 0.04 equiv.) were dissolved in
dry benzene (2 ml) and warmed to reflux under nitrogen. A
solution of tri-n-butyltin hydride (55 mg, 0.19 mmol, 1.3 equiv.)
was added dropwise from a syringe pump. The reaction was
refluxed for 16 hours. TLC analysis indicated a very compli-
cated mixture of products from which, despite repeated
attempts at column chromatography, eluent hexane to 20%
ethyl acetate in hexane, no fraction free from tin residues was
isolated. 1H-NMR analysis of the major fraction indicated that
the characteristically sharp propiolic proton signal (δ 2.93, s,
12-H)) was absent, whereas, the propargyl alkynic proton
(δ 2.44, t, J 2.4, 9-H) was present. 1H-NMR spectra indicated a
50:50 mixture of vinyl stannanes 12; δH (90 MHz) 7.73 (0.5H,
d, J 22, 12-H), 7.22 (0.5H, d, J 16, 12-H), 6.65 (0.5H, d, J 16,
11-H), 6.28 (0.5H, d, J 22, 11-H), 5.82 (2H, br m, 2-H, 3-H),
5.30 (1H, m, 4-H), 5.18 (1H, br s, 1-H), 4.24 (2H, m, 7-H2), 3.70
(3H, m, 5-H, 6-H2), 2.44 (1H, dd, J 3, 3, 9-H), 1.32 (12H, m,
20-H6, 21-H6, SnCH2CH2CH2), 0.85 (24H, m, 16-H3, 17-H3,
18-H3, 19-H6, 22-H9, (CH3)3CSi, SnCH2(CH2)2CH3), 0.00 (6H,
s, 13-H3, 14-H3, (CH3)2Si).
1
(CH3)2Si); H–1H J-COSY NMR 1-H to 2-H to 10-H2, 1-H to
6-H, 4-H to 5a-H to 5b-H to 6-H, 2-H to 10-H2, 4-H to 5b-H to
6-H to 13-H (weak), 8a-H to 8b-H to 13-H, 8a-H to 13-H, 11a-
H to 11b-H to 12-H3, 11b-H to 12-H3, 20-H6 to 21-H6 to 22-H6
to 23-H9; δC 159.9 (C, 7-C), 117.3 (CH, 13-C), 97.3 (CH, 4-C),
77.4 (CH, 1-C), 72.9 (CH2, 8-C), 71.2 (CH, 2-C), 64.7 (CH2, 11-
C), 62.9 (CH2, 10-C), 42.1 (CH, 6-C), 31.9 (CH2, 5-C), 29.5
(CH2, 22-C, Sn-(CH2)2CH2-), 27.7 (CH2, 21-C, Sn-CH2CH2-),
26.3 (CH3, 17-C, 18-C, 19-C, (CH3)3CSi), 18.8 (C, 16-C), 15.5
(CH3, 12-C), 14.1 (CH3, 23-C, Sn(CH2)3CH3), 10.1 (CH2, 20-C,
SnCH2); m/z (EIϩ, peaks marked with * are tin cluster maxima)
614* (28%, C29H58O4SiSn, Mϩ), 574* (23%, M Ϫ EtOH), 557*
(51%, M Ϫ Bu), 513* (45%, M Ϫ EtOH Ϫ Bu), 329 (33%,
M Ϫ Bu3Sn), 288 (57%), 73 (100%); νmax (neat)/cmϪ1 2956, 2947,
2871, 2856, 1126, 1056, 837.
(1S,2S,4S,6R,7S)-4-endo-Ethoxy-2-exo-tert-butyldimethyl-
silyloxymethyl-7-exo-(tri-n-butylstannylmethyl)-3,9-
dioxabicyclo[4.3.0]nonane 24
Attempted tri-n-butyltin hydride mediated cyclisation of ethyl
6-O-tert-butyldimethylsilyl-4-O-propiolyl-2,3-dideoxy-ꢀ-D-
erythro-hex-2-enopyranoside 14
Ethyl 6-O-tert-butyldimethylsilyl-4-O-propiolyl-2,3-dideoxy-α-
-erythro-hex-2-enopyranoside 14 (60 mg, 0.18 mmol), and
AIBN (2 mg, 0.007 mmol, 0.04 equiv.) were dissolved in dry
toluene (2 ml) and refluxed under nitrogen for 15 min. A solu-
tion of tri-n-butyltin hydride (77 mg, 0.26 mmol, 1.5 equiv.) in
toluene (3 ml) was added dropwise over 3 hours via a syringe
pump. The reaction solution was refluxed for 18 hours. TLC
analysis indicated a complicated mixture. Column chromato-
graphy, eluent hexane to 10% ethyl acetate in hexane, afforded a
fraction containing one of the major products, with two trace
Ethyl 6-O-tert-butyldimethylsilyl-4-prop-2-enyl-2,3-dideoxy-α-
-erythro-hex-2-enopyranoside 23 (100 mg, 0.30 mmol) was
dissolved in dry toluene (3 ml) and the solution warmed to
reflux under nitrogen. AIBN (2 mg, 0.01 mmol, 0.04 equiv.) was
added and the solution refluxed for a further ten minutes. A
J. Chem. Soc., Perkin Trans. 1, 2000, 1559–1569
1565

View Article Online
solution of tri-n-butyltin hydride (135 mg, 0.46 mmol, 1.5
equiv.) in dry toluene (3 ml) was added over 3 hours from a
syringe pump. The solution was refluxed for 15 hours. TLC
analysis indicated partial reaction of the starting material.
Purification of the crude reaction mixture by flash column
chromatography, eluent hexane to 10% diethyl ether in hexane
afforded the title compound 24 as a mixture of isomers (14 mg,
7%) and recovered starting material 23 (38 mg, 38%). δH 4.79
(0.5H, dd, J 8.5, 5.6, 4-H), 4.74 (0.5H, t, J 4.3, 4-H), 3.94 (0.5H,
t, J 7.6, 8a-H), 3.90–3.47 (5H, m, 1-H, 2-H, 10-H2, 11a-H), 3.41
(0.5H, dq, J 8.5, 7.0, 11b-H), 3.32 (1H, m), 3.17 (0.5H, t, J 8.3,
8b-H), 1.90 (0.5H, dt, J 13.5, 5.3, 5-H), 1.80 (0.5H, dt, J 13.5,
5.2, 5-H), 1.71 (0.5H ϩ 0.5H, m, H-6), 1.38 (6H, m), 1.23 (6H,
m), 1.14 (3H, t, J 6.9, 12a-H3), 1.11 (3H, t, J 7.5, 12b-H3), 0.85–
0.70 (17-H3, 18-H3, 19-H3, 20-H6, 23-H9), 0.00 (6H, s, 14-H3,
6-H to eq-5-H to 4-H, 6-H to ax-5-H to 4-H, 4-H to 3-H to
10a-H to 10b-H, 3-H to 10b-H, 8a-H to 8b-H to 11-H, 8a-H to
15-H, 16-H6 to 17-H6 to 18-H6 to 19-H9; δC 171.3, 170.5 (2CO),
2
157.2 (C, 7-C, JC-Sn117/119 227), 119.8 (CH2, 15-C), 101.1 (CH,
1-C), 70.9 (CH2, 8-C), 70.1 (CH, 4-C), 66.6 (CH, 3-C), 63.4
(CH2, 10-C), 43.8 (CH, 6-C), 31.70 (CH2, 5-C), 29.5 (CH2, 18-
C, Sn(CH2)2CH2-), 27.7 (CH2, 17-C, SnCH2-CH2), 21.4 (2CH3,
12-C, 14-C), 14.1 (CH3, 19-C, Sn(CH2)3CH3), 10.2 (CH2, 16-C,
SnCH2); m/z (EIϩ), only the highest intensity ions are reported
for tin ion clusters 560 (0%, C25H44O6Sn), 500 (56%, M Ϫ
AcOH), 440 (43%, M Ϫ 2AcOH), 383 (76%, M Ϫ 2AcO-
H Ϫ Bu), 291 (71%, Bu3Sn), 235 (51%), 231 (93%), 57 (100%);
νmax (neat)/cmϪ1 3460 (weak), 2840, 1740 (str), 1625 (weak),
1460; νmax (CDCl3)/cmϪ1 2956, 2919, 2872, 2852, 1744 (str,
C᎐O), 1628 (weak), 1243; R 0.65 (hexane–EtOAc, 50:50).
᎐
f
1
15-H3, (CH3)2Si); H–1H J-COSY NMR 4-H to 5-H to 6-H,
8a-H to 8b-H, 11-H2 to 12a-H3, 11-H2 to 12b-H3; δC 98.1, 97.0
(peak height ratio 42:58, 4-C), 78.3, 77.6, 76.2, 75.5, 74.8, 72.7,
70.4, 65.0, 64.6, 63.0, 62.8, 44.7, 42.1, 41.5, 39.6, 29.9, 29.6,
28.1, 27.8, 26.33, 26.30, 18.6, 15.5, 14.1, 11.2, 9.6, 9.5; m/z (EI)
620 (C29H60O4120SnSi, M, absent), 576 (1%, tin cluster, M Ϫ O-
CH2CH3), 516 (100%, tin cluster, M Ϫ O-SitBuMe2), 289 (35%,
tin cluster), 175 (27%, tin cluster), 41 (56%); m/z (CIϩ, NH3) 638
(1S,3R,4S,6S)-4-endo-Acetoxy-3-exo-acetoxymethyl-7-
methylene-2,9-dioxabicyclo[4.3.0]nonane 25c by cyclisation of
26
1-(2-Chloroprop-2-enyl)
4,6-di-O-acetyl-2,3-dideoxy-α--
erythro-hex-2-enopyranoside 26 (100 mg, 0.31 mmol), was dis-
solved in dry toluene (3 ml) and heated to reflux under nitrogen.
A catalytic amount of AIBN (3 mg, 0.01 mmol, 0.04 equiv.) was
added and the solution refluxed for a further 10 min. A solution
of tri-n-butyltin hydride (131 mg, 0.15 mmol, 1.5 equiv.) in dry
toluene (3 ml) was added over 3 hours from a syringe pump and
the reaction solution allowed to reflux for a further 16 hours.
Purification by column chromatography, eluent hexane to 10%
diethyl ether in hexane, afforded the title compound 25c (27 mg,
30%) as a clear oil, with identical spectroscopy to previously
prepared product, and unconverted starting material 26 (8 mg,
8.0%) plus a slightly impure fraction containing the product 26
(23 mg, <26%).
ϩ
(6%, C29H64O4Si120SnN, M ϩ NH4 , cluster analysis correct),
592 (4%, M ϩ NH4 Ϫ EtOH), 575 (33%, tin cluster), 517 (24%,
tin cluster), 308 (100%, Sn cluster); νmax (CDCl3)/cmϪ1 2927,
1248.
(1S,3R,4S,6S)-4-endo-Acetoxy-3-exo-acetoxymethyl-7-(Z-tri-n-
butylstannylmethylene)-2,9-dioxabicyclo[4.3.0]nonane 25a
Attempted preparation of (1S,3R,4S,6S)-4-endo-acetoxy-3-exo-
acetoxymethyl-5-prop-2-enyl-7-(E-tri-n-butylstannylmethylene)-
2,9-dioxabicyclo[4.3.0]nonane 27
Prop-2-ynyl
4,6-di-O-acetyl-2,3-dideoxy-α--erythro-hex-2-
Prop-2-ynyl
4,6-di-O-acetyl-2,3-dideoxy-α--erythro-hex-2-
enopyranoside 7a (500 mg, 1.86 mmol) and AIBN (13 mg,
0.079 mmol, 0.04 equiv.) were dissolved in dry toluene (10 ml)
and warmed to reflux under argon. TBTH (597 mg, 2.05 mmol,
1.1 equiv.) was dissolved in dry toluene (1.5 ml) and added to
the reaction solution over 4 hours using a syringe pump. The
solution was allowed to reflux for a further 12 hours, cooled and
evaporated to dryness. Purification by flash column chromato-
graphy; eluent hexane to 20% ethyl acetate in hexane, afforded
the title compound 25a (677 mg, 65%). Repetition of this reac-
tion on 10 × scale: 7a (5 g, 18.6 mmol), AIBN (0.13 g, 0.79
mmol, 0.04 equiv.), in toluene (100 ml) plus TBTH (5.97 g, 20.5
mmol, 1.1 equiv.) in toluene (5 ml) afforded TBTH (1.13 g,
19%), the title compound 25a (2.75 g, 28%) and starting
material 7b (1.7 g, 34%). Combustion analysis: C25H44O6Sn
requires C 53.69, H 7.93; found C 53.47, H 7.78%; δH 5.88 (1H,
m, 2JH-Sn117/119 54, 15-H), 5.38 (1H, d, J 4.6, 1-H), 4.77 (1H, td,
J 9.3, 5.0, 4-H), 4.57 (1H, ddd, J 13.2, 2.4, 1.5, 8a-H), 4.27 (1H,
dd, J 12 0, 5.0, 10a-H), 4.22 (1H, dd, J 12.0, 3.0, 10b-H), 4.16
(1H, dd, J 12.1, 2.2, 8b-H), 4.01 (1H, ddd, J 9.1, 5.1, 2.3, 3-H),
2.77 (1H, br m, 6-H), 2.23 (1H, ddd, J 13.6, 6.7, 5.3, eq-5-H),
2.09 (3H, s, CH3CO), 2.02 (3H, s, CH3CO), 1.70 (1H, dt, J 13.4,
9.4, ax-5-H), 1.51 (6H, m, 18-H6, Sn(CH2)2CH2), 1.35 (12H, m,
16-H6, 17-H6, SnCH2CH2), 0.89 (9H, m, 19-H9, Sn(CH2)3CH3);
δH (C6D6) 5.83 (1H, d, J 1.2, 15-H), 5.44 (1H, d, J 4.4, 1-H), 5.08
(1H, ddd, J 9.5, 9.5, 4.8, 4-H), 4.72 (1H, br m, 8a-H), 4.55 (1H,
dd, J 12.1, 5.0, 10a-H), 4.35 (2H, m, 8b-H, 10b-H), 4.27 (1H,
ddd, J 9.5, 4.9, 2.4, 3-H), 2.43 (1H, m, 6-H), 2.21 (1H, ddd,
J 12.8, 7.2, 5.2, eq-5-H), 1.80 (3H, s, CH3CO), 1.75 (3H, s,
CH3CO), 1.72 (1H, m, ax-5-H), 1.64. (6H, m, 18-H6, Sn-
(CH2)2CH2), 1.42 (12H, m, 16-H6, 17-H6, SnCH2CH2), 1.04
enopyranoside 7a (2 g, 7.0 mmol), allyltri-n-butyltin (5.80 g,
17.5 mmol, 2.5 equiv.) and AIBN (51 mg, 0.2 mmol, 0.03 equiv.)
were dissolved in dry benzene (60 ml) and warmed to reflux.
Tri-n-butyltin hydride (210 mg, 0.72 mmol, 0.1 equiv.) was dis-
solved in dry benzene (5 ml) and added dropwise by a syringe
pump. The solution was refluxed for 14 hours, evaporated to
dryness, and purified by flash column chromatography, eluent
hexane to 10% ethyl acetate in hexane, to yield (1S,3R,4S,6S)-
4-endo-acetoxy-3-exo-acetoxymethyl-7-(Z-tri-n-butylstannyl-
methylene)-2,9-dioxabicyclo[4.3.0]nonane 25a (1.1 g, 28%),
starting material 7a (400 mg, 10%) and a mixture of 25a and 7a
(150 mg, 4%).
(Z)-(4-Tri-n-butylstannylbut-3-enyl) 4,6-di-O-acetyl-2,3-
dideoxy-ꢀ-D-erythro-hex-2-enopyranoside 29a and (1S,3R,4S,
6S)-4-endo-acetoxy-3-exo-acetoxymethyl-7-(E-tri-n-butyl-
stannylmethylene)-2,10-dioxabicyclo[4.4.0]decane 30
But-3-ynyl
4,6-di-O-acetyl-2,3-dideoxy-α--erythro-hex-2-
enopyranoside 28 (420 mg, 1.5 mmol) was dissolved in dry
toluene (3 ml). The solution was warmed to reflux under nitro-
gen and AIBN (10 mg, 0.06 mmol, 0.04 equiv.) added in one
1
(9H, m, 19-H9, Sn(CH2)3CH3); H-1H J-COSY NMR 1-H to
1566
J. Chem. Soc., Perkin Trans. 1, 2000, 1559–1569

View Article Online
portion. Tri-n-butyltin hydride (0.66 g, 2.3 mmol, 1.5 equiv.)
was dissolved in dry toluene (4 ml) and added dropwise from a
syringe pump over 4 hours. The reaction solution was refluxed
for 16 hours. TLC analysis of the reaction solution indicated a
complex mixture of starting material 28 and two products, both
of very similar polarity. The more polar product was present
only at a very small amount. The reaction solution was concen-
trated to approximately 1 ml and purified by repeated column
chromatography over silver nitrate impregnated silica gel, elu-
ent hexane to 10% diethyl ether in hexane, to afford the title
products 29a (270 mg, 32%), 30 (<5 mg), a mixture of 29a and
30 (50 mg, 6%) and starting material 28 (70 mg, 16%).
butyltin hydride (1.2 g, 4.2 mmol, 1.5 equiv.) in dry toluene
(3 ml) was added dropwise from a syringe pump over 4 hours
and the reaction solution refluxed for 12 hours. Analysis by
TLC indicated the presence of a single product and starting
material. Purification by column chromatography, eluent hex-
ane to 30% EtOAc in hexane, afforded the reduction product
31b (600 mg, 77%). None of the expected cyclised product was
detected. Combustion analysis: C13H20O6 requires C 57.34, H
7.40; found C 57.65, H 7.52%; δH 5.82 (1H, d, J 10.8, 3-H), 5.77
(1H, ddd, J 10.2, 1.9, 1.9, 2-H), 5.24 (1H, dd, J 9.6, 0.9, 4-H),
4.96 (1H, s, 1-H), 4.16 (2H, m, 6-H2), 4.05 (1H, ddd, J 9.6, 5.4,
2.4, 5-H), 3.66 (1H, ddd, J 9.5, 6.8, 6.8, 7a-H), 3.42 (1H, ddd, J
9.5, 6.6, 6.6, 7b-H), 2.04, 2.02 (3H, 3H, s, s, 11-H3, 13-H3,
2CH3CO), 1.56 (2H, m, 8-H2), 0.97 (3H, t, J 7.4, 9-H3); δC
171.2, 170.7 (10-C and 12-C, 2CO), 129.4 (3-C), 128.4 (2-C),
94.8 (1-C), 71.0 (4-C), 67.3 (7-C), 65.7 (5-H), 63.5 (6-C), 23.4
(8-C), 21.4, 21.2 (2CH3, 11-C, 13-C), 11.2 (CH3, 9-C); m/z (EIϩ)
272 (M, C13H20O6, absent), 213 (100%, M Ϫ OPr), 170 (68%,
M Ϫ OHCCH2OAc, retro Diels–Alder), 153 (73%, M Ϫ OPr Ϫ
AcOH), 128 (57%), 111 (54%, M Ϫ OHCCH2OAc–OPr), 86
(65%), 57 (42%); m/z (CIϩ, NH3) 290 (18%, M ϩ NH4,
C13H24NO6), 230 (28%, M ϩ NH4 Ϫ AcOH), 213 (100%,
M Ϫ OPr), 172 (15%), 153 (12%), 128 (11%); νmax (CDCl3)/cmϪ1
Spectroscopic data for (Z)-(4-tri-n-butylstannylbut-3-enyl)
4,6-di-O-acetyl-2,3-dideoxy-α-D-erythro-hex-2-enopyranoside
29a: combustion analysis: C26H46O6Sn requires C 54.47, H 8.09;
found C 54.42, H 8.28%; δH 6.45 (1H, ddd, J 12.6, 6.9, 6.9, 9-H),
5.87 (1H, d, J 12.4, 10-H), 5.82 (1H, br d, J 10.8, 3-H), 5.76
(1H, ddd, J 10.2, 2.1, 2.1, 2-H), 5.25 (1H, dd, J 9.7, 1.3, 4-H),
4.97 (1H, br s, 1-H), 4.20 (1H, dd, J 12.1, 5.2, 6a-H), 4.10 (1H,
dd, 12.0, 2.3, 6b-H), 4.05 (1H, ddd, J 9.6, 5.2, 2.2, 5-H), 3.75
(1H, ddd, J 9.4, 7.0, 7.0, 7a-H), 3.46 (1H, ddd, J 9.3, 6.9, 6.9,
7b-H), 2.29 (2H, app q, J 6.7, 8-H2), 2.04 (3H, s, 16-H3 or
18-H3, CH3CO), 2.02 (3H, s, 16-H3 or 18-H3, CH3CO), 1.43
(6H, m, 13-H6, Sn(CH2)2CH2), 1.25 (6H, m, 12-H6, SnCH2-
2980, 1740 (str, C᎐O), 1220; R 0.75 (ether).
᎐
f
1
CH2), 0.84 (15H, m, 11-H6, 14-H9, SnCH2(CH2)2CH3); H–1H
J-COSY NMR 1-H to 2-H to 3-H, 4-H to 5-H to 6a-H, 5-H to
6b-H, 6a-H to 6b-H, 7a-H to 7b-H, 7a-H to 8-H2, 7b-H to 8-H2,
8-H2 to 9-H (weak), 9-H to 10-H, 11-H6 to 12-H6 to 13-H6 to
14-H9; δH (C6D6) 6.61 (1H, dt, J 12.6, 6.9, 9-H), 6.08 (1H, d,
J 12.5, 10-H), 5.73 (1H, d, J 10.3, 3-H), 5.53 (1H, m, 2-H), 5.50
(1H, dd, J 9.6, 1.4, 4-H), 4.79 (1H, s, 1-H), 4.31 (2H, m, 6-H2),
4.23 (1H, ddd, J 9.5, 3.8, 3.8, 5-H), 3.80 (1H, dt, J 9.4, 6.9, 7a-
H), 3.37 (1H, dt, J 9.3, 6.6, 7b-H), 2.39 (2H, m, 8-H2), 1.67 (3H,
16-H3, or 18-H3, CH3CO), 1.56 (6H, m, 13-H6, Sn(CH2)2CH2),
1.54 (3H, 16-H3 or 18-H3, CH3CO), 1.35 (6H, m, 12-H6,
SnCH2-CH2), 1.01 (6H, m, 11-H6, SnCH2), 0.93 (9H, m, 14-H9,
Sn(CH2)3CH3); δC 170.7, 170.2 (s, s, 15-C, 17-C, CO), 144.9 (d,
9-C), 131.5 (d, 10-C), 129.4 (d, 3-C), 128.2 (d, 2-C), 94.8 (d,
1-C), 68.6 (t, 7-C), 67.3 (d, 4-C), 65.6 (d, 5-C), 63.4 (t, 6-C), 37.6
(t, 8-C), 29.6 (t, 13-C, Sn(CH2)2CH2), 27.6 (t, 12-C, SnCH2-
CH2), 21.4, 21.2 (q, q, 16-C, 18-C), 14.1 (q, 14-C, Sn(CH2)3-
CH3), 10.6 (t, 11-C, SnCH2); m/z (EI) 574 (C26H46O6120Sn, M,
absent), 517 (9%, tin cluster, M Ϫ Bu), 457 (4%, tin cluster,
M Ϫ Bu Ϫ AcOH), 415 (4%, tin cluster), 213 (100%, M Ϫ
(1S,3R,4S,6S)-4-endo-Acetoxy-3-exo-acetoxymethyl-7-(tri-n-
butylstannylmethyl)-2,9-dioxabicyclo[4.3.0]nonane 34a, b
Prop-2-enyl 2,3-dideoxy-α--erythro-hex-2-enopyranoside (100
mg, 0.54 mmol) 33a was dissolved in dry toluene (3 ml) and
warmed to reflux under nitrogen. AIBN (4 mg, 0.02 mmol, 0.04
equiv.) was added portionwise and the reaction refluxed for 15
min. A solution of tri-n-butyltin hydride (240 mg, 0.82 mmol,
1.5 equiv.) in dry toluene (3 ml) was added dropwise by a
syringe pump over 3 hours. The reaction solution was refluxed
for 16 hours. TLC analysis indicated a complex mixture con-
taining two major products. Repeated column chromatography,
eluent hexane to 10% ether in hexane, gave a mixture of two
products 34a, b, (10 mg, 5% yield), ratio 1:1 from analysis of
Bu Sn-CH᎐CH-CH -CH -OH),153(80%,M Ϫ Bu SnCH᎐CH-
᎐
᎐
3
2
2
3
CH2-CH2-OH Ϫ AcOH), 110 (74%), 43 (65%); νmax (Nujol)/
cmϪ1 1745 (C᎐O, str), 1595 (weak), 1230, 1040; R 0.55 (hexane–
1
᎐
f
the H-NMR spectrum of the crude reaction mixture. δH 5.45
EtOAc, 75:25).
(0.5H, br d, 1-H), 5.35 (0.5H, br d, 1-HЈ), 4.79 (1H, m, 4-H),
4.25 (2H, m, 8a-H, 10a-H), 4.13 (1H, m, 10b-H), 3.93 (1H, m,
8b-H), 3.55 (0.5H, m, 3-H), 3.45 (1H, m, 7-H), 3.36 (0.5H, m,
3-H), 2.63 (0.5H, m, 6-H), 2.23 (0.5H, m), 2.12 (0.5H, m), 2.05
(6H, s, CH3CO), 2.02 (3H, s, CH3CO), 1.99 (3H, s, CH3CO),
1.85 (1H, m, 5-H), 1.45 (6H, m, 18-H6, Sn(CH2)2CH2), 1.30
(12H, m, 16-H6, 17-H6, SnCH2(CH2)2), 0.92 (11H, m, 11-H2,
19-H9, CH2Sn(CH2)3CH3); m/z (EIϩ) 562 (M, C25H46O6120Sn,
absent), 505 (24%, Sn cluster, M Ϫ Bu), 445 (11%, Sn cluster,
M Ϫ AcOH), 385 (4%, M Ϫ Bu Ϫ AcOH Ϫ AcOH), 293 (53%,
Sn cluster, Bu3SnH), 179 (49%, BuSn), 57 (100%); νmax (CDCl3)/
Spectroscopic data for (1S,3R,4S,6S)-4-endo-acetoxy-3-exo-
acetoxymethyl-7-(E-tri-n-butylstannylmethylene)-2,10-dioxabi-
cyclo[4.4.0]decane 30: δH 6.01 (1H, s, 11-H), 5.15 (2H, m, 1-H,
4-H), 4.22, 3.95 (m, 3-H, 9-H2, 11-H2), 2.2 (m, H-6, 8-H2), 2.1
(6H, s). The remainder of the spectrum indicated the presence
of a large excess of non-stoichiometric tin residues; Rf 0.57
(hexane–EtOAc, 75:25).
Propyl 4,6-di-O-acetyl-2,3-dideoxy-ꢀ-D-erythro-hex-2-eno-
pyranoside 31b
cmϪ1 2927, 1739 (str, C᎐O), 1440, 1380, 1220, 1020.
᎐
(1S,3R,4S,6S)-4-endo-Acetoxy-3-exo-acetoxymethyl-7-methyl-
7-(tri-n-butylstannylmethyl)-2,9-dioxabicyclo[4.3.0]nonane
34c, d
2Ј-Methylprop-2-enyl 4,6-di-O-acetyl-2,3-dideoxy-α--erythro-
hex-2-enopyranoside 33b (500 mg, 1.76 mmol), was dissolved in
dry benzene (8 ml) and warmed to reflux under nitrogen. AIBN
(13 mg, 0.07 mmol, 0.04 equiv.) was added portionwise and the
solution refluxed for 10 min. A solution of tri-n-butyltin
hydride (560 mg, 1.9 mmol, 1.1 equiv.) was added dropwise over
3-Bromopropyl 4,6-O-di-acetyl-2,3-dideoxy-α--erythro-hex-2-
enopyranoside 31a (1.0 g, 2.8 mmol) was dissolved in dry ben-
zene (5 ml) and warmed to reflux under nitrogen. AIBN (20 mg,
1.2 mmol, 0.04 equiv.) was added in a single portion and the
reaction solution refluxed for 15 mins. A solution of tri-n-
J. Chem. Soc., Perkin Trans. 1, 2000, 1559–1569
1567

View Article Online
5 D. P. Curran, S. Hadida, S. Y. Kim and Z. Y. Luo, J. Am. Chem. Soc.,
1999, 121, 6607; J. H. Horner, F. N. Martinez, M. Newcomb,
S. Hadida and D. P. Curran, Tetrahedron Lett., 1997, 38, 2783.
6 R. Rai and D. B. Collum, Tetrahedron Lett., 1994, 35, 6221; D. L. J.
Clive and W. Yang, J. Org. Chem., 1995, 60, 2607; E. Vedejs, S. M.
Duncan and A. R. Haight, J. Org. Chem., 1993, 58, 3046.
7 For reviews of applications in natural product synthesis see M.
Ramaiah, Tetrahedron, 1987, 43, 3541; G. P. Jasperse, D. P. Curran
and T. L. Fevig, Chem. Rev., 1991, 91, 1237; B. Giese, B. Kopping,
T. Gobel, J. Dickhaut, G. Thoma, K. J. Kulicke and F. Trach,
Org. React., 1996, 48, 301.
8 For a rare exception see, D. P. Curran and A. E. Gabarda,
Tetrahedron, 1999, 55, 3327.
9 J. E. Baldwin, D. R. Kelly and C. B. Ziegler, J. Chem. Soc., Chem.
Commun., 1984, 133; D. R. Kelly and J. E. Baldwin, J. Chem. Soc.,
Chem. Commun., 1985, 682.
3 hours from a syringe pump. Reflux was maintained for 16
hours. TLC analysis indicated the presence of several products
and starting material 33b. Repeated column chromatography,
eluent hexane to 10% diethyl ether in hexane, afforded an
inseparable mixture (110 mg, 10%), and starting material 33b
(240 mg, 48%). Analysis of the mixture by 1H-NMR indicated a
mixture of three components 34c, d and an unassigned com-
pound. The ratio of components was 50:35:15, based on inte-
gration of signals at δ 5.39, 5.31 and 5.24 respectively; δH 5.39
(1H, d, J 4.9, 1a-H), 5.31 (1H, d, J 4.7, 1b-H), 5.24 (1H, d, J 4.8,
1c-H), 4.74 (2H, m), 4.24 (3H, m), 4.08 (3H, m), 3.87 (3H, m),
3.58 (1H, q, J 4.8, c-H), 3.49 (1H, dd, J 10.6, 8.2, b-H), 3.44
(1H, d, J 9.2, c-H), 3.32 (1H, dd, J 8.4, 4.7, a-H), 2.5 (2H, m,
c-H2), 2.18 (1H, m, c-H), 2.08 (2H, m, 2b-H, b-H), 2.02, 2.00,
1.997, 1.98 (3 × 6H, 4 × s, CH3CO), 1.85 (1H, m, 2a-H),
1.4 (6H, m, Bu3Sn(CH2)2CH2-), 1.25 (6H, m Bu3SnCH2CH2-),
0.78 (14H, m, 11-H2, 12-H3, Bu3SnCH2-, Bu3Sn(CH2)3-CH3);
δC 171.3, 170.4, 170.3, 103.6, 102.4, 101.1, 75.0, 73.1, 73.0, 70.0,
69.8, 67.5, 66.9, 63.5, 63.1, 45.1, 42.3, 40.6, 40.0, 29.6, 29.5,
27.8, 27.5, 26.0, 21.5, 21.2, 14.1, 14.0, 9.6, 9.56, 5.20; m/z (EIϩ)
588 (M, C27H48O6120Sn, absent), 505 (24%, Sn cluster), 445 (9%,
Sn cluster, 505 Ϫ AcOH), 385 (3%, Sn cluster, 505 Ϫ AcOH Ϫ
AcOH), 292 (Bu3SnH), 57 (100%); νmax (neat)/cmϪ1 2958, 1738
(str), 1464, 1376, 1248, 1024; Rf 0.7 (hexane–EtOAc, 50:50).
10 For a discussion of radical reactions in retrosynthetic planning see,
D. P. Curran, Synlett, 1991, 63.
11 Reviews: M. Malacria, Chem. Rev., 1996, 96, 289; R. Gigg
(ed.), Tetrahedron, symposium-in-Print No. 35, 1996, 52, 11385–
11664; A. Martinez-Graz and J. Marco-Contelles, Chem. Soc. Rev.,
1998, 27, 155; G. Pattenden and S. Handa, Contemp. Org. Synth.,
1997, 4, 196; F. Aldabbagh and W. R. Bowman, Contemp. Org.
Synth., 1997, 4, 261.
12 For radical induced polycyclisations see S. Handa and G. Pattenden,
J. Chem. Soc., Perkin Trans. 1, 1999, 843; K. Takasu, J. Kuroyanagi,
A. Katsumata and M. Ihara, Tetrahedron Lett., 1999, 40, 6277;
M. Breithor, U. Herden and H. M. R. Hoffmann, Tetrahedron,
1997, 53, 8401; S. Bogen, L. Fensterbank and M. Malacria, J. Org.
Chem., 1999, 64, 819; U. Jahn and D. P. Curran, Tetrahedron Lett.,
1995, 36, 8921; P. Devin, L. Fensterbank and M. Malacria, J. Org.
Chem., 1998, 63, 6764; W. R. Bowman, P. T. Stephenson and A. R.
Young, Tetrahedron, 1996, 52, 11445 and other radical cascade
reactions, E. Lee, C. U. Hur, Y. H. Rhee, Y. C. Park and S. Y. Kim,
J. Chem. Soc., Chem. Commun., 1993, 1466.
13 First presented in a lecture at ChemSpec Europe 94, Manchester,
19–20th April, 1994; D. R. Kelly, BACS Symp. Proc., 1994, 101;
D. R. Kelly, Spec. Chem., 1995, 68; D. R. Kelly, Agro Food Industry,
Hi-tech, 1995, 6, 33.
14 C. Anies, L. Billot, J.-Y. Lallemand and A. Pancrazi, Tetrahedron
Lett., 1995, 36, 7247.
(1S,3R,6S)-3-exo-Pivaloyloxymethyl-4-oxo-7-(Z-tri-n-
butylstannylmethylene)-2,9-dioxabicyclo[4.3.0]nonane 36a
15 C. Spino and N. Barriault, J. Org. Chem., 1999, 64, 5292.
16 J. Marco-Contelles, Chem. Commun., 1996, 2629. Structure 2 in this
paper is an incorrect regioisomer (corrigendum 1997, 725).
Structure 5 should be the 4-O-propenyl derivative rather than the
propynyl derivative. This was not corrected in the corrigendum;
A. Martinez-Graz and J. Marco-Contelles, Chem. Soc. Rev., 1998,
27, 155.
17 For related palladium catalysed bicyclisation reactions see, C. W.
Holzapfel and L. Marais, Tetrahedron Lett., 1997, 38, 8585;
J.-F. Nguefack, V. Bolitt and D. Sinou, Tetrahedron Lett., 1996, 37,
59; J.-F. Nguefack, V. Bolitt and D. Sinou, J. Org. Chem., 1997, 62,
6827.
18 R. J. Ferrier, J. Chem Soc., 1964, 5443-5449; R. J. Ferrier, N. Prasad
and G. H. Sankey, J. Chem Soc. (C), 1968, 974; R. J. Ferrier and
N. Prasad, J. Chem Soc. (C), 1969, 570.
19 B. Fraser-Reid, Acc. Chem. Res., 1975, 8, 192; B. Fraser-Reid, Acc.
Chem. Res., 1985, 18, 347; B. Fraser-Reid and R. C. Anderson, Prog.
Chem. Org. Nat. Prod., 1980, 39, 1; N. L. Holder, Chem. Rev., 1982,
82, 287.
20 S. Hanessian, Total Synthesis of Natural Products: The ‘Chiron’
Approach, Pergamon Press, Oxford, 1983.
21 E. Maudru, G. Singh and R. H. Wightman, Chem. Commun.,
1998, 1505; R. McCague, R. G. Pritchard, R. J. Stoodley and
D. S. Williamson, Chem. Commun., 1998, 2691.
22 K. Toshima, T. Ishizuka, G. Matsua, M. Nakata and M. Kinoshita,
J. Chem. Soc., Chem. Commun., 1993, 704.
23 B. S. Babu and K. K. Balasubramanian, Synlett, 1999, 1261;
B. S. Babu and K. K. Balasubramanian, Tetrahedron Lett., 1999, 40,
5777.
24 T. Linker, T. Sommermann, T. Gimisis and C. Chatgilialoglu,
Tetrahedron Lett., 1998, 39, 9637.
Prop-2-ynyl-2,3-dideoxy-6-O-pivaloyl-α--erythro-hex-2-eno-
pyran-4-uloside 35c (50 mg, 0.19 mmol) and AIBN (2 mg,
0.008 mmol, 0.04 equiv.) were dissolved in dry toluene (1 ml)
and warmed to reflux under argon. TBTH (66 mg, 0.23 mmol,
1.2 equiv.) was dissolved in dry toluene (2 ml) and added to the
reaction solution over 3 hours via a syringe pump. The solution
was allowed to reflux for a further 16 hours. TLC analysis of the
reaction solution indicated a complex mixture of products and
starting material. Purification by flash column chromatography,
eluent hexane to 20% ethyl acetate in hexane, afforded the title
compound 36a (8 mg, 8%). δH 5.88 (1H, dd, J 2.5, 2.5, 11-H),
5.82 (1H, d, J 7.1, 1-H), 4.45 (2H, m, 10-H2), 4.35 (2H, m,
8-H2), 4.05 (1H, dd, J 3.7, 3.0, 3-H), 3.43 (1H, br m, 6-H), 2.77
(2H, m, 5-H2), 1.46 (6H, m, 19-H6, Sn(CH2)2CH2-), 1.30 (6H,
m, 18-H6, SnCH2CH2), 1.18 (9H, s, 14-H3, 15-H3, 16-H3,
C(CH3)3), 0.92 (15H, m, 17-H6, 20-H6, SnCH2(CH2)2CH3); Rf
0.90 (hexane–EtOAc, 50:50).
25 F. W. Lichtenthaler and B. Werner, Carbohydr. Res., 1999, 319, 47;
B. K. Banik, O. Zegrocka, M. Manhas and A. K. Bose, Hetero-
cycles, 1997, 46, 173; M. Koreeda, T. A. Houston, B. K. Shull,
E. Klenke and R. J. Tuinman, Synlett, 1995, 90.
26 J. C. Lopez, A. M. Gomez, S. Valerde and B. Fraser-Reid, J. Org.
Chem., 1995, 60, 3851; J. C. Lopez and B. Fraser-Reid, J. Chem.
Soc., Chem. Commun., 1992, 94.
27 B. Fraser-Reid and S. Jarousz, unpublished method.
28 M. J. Robins, S. D. Hawrelak, T. Kanai, J. Siefert and R. Mengel,
J. Org. Chem., 1979, 44, 1317.
References
1 W. P. Neumann, Synthesis, 1987, 665.
2 E. J. Corey and J. W. Suggs, J. Org. Chem., 1975, 40, 2554.
3 I. Terstiege and R. E. Maleczka, J. Org. Chem., 1999, 64, 342.
4 D. Crich and S. X. Sun, J. Org. Chem., 1996, 61, 7200 and references
therein. In our experience the best of these procedures is
acetonitrile–hexane partition (J. M. Berge and S. M. Roberts,
Synthesis, 1979, 471), although the product must be fairly polar,
otherwise extraction into the acetonitrile phase is inefficient.
1568
J. Chem. Soc., Perkin Trans. 1, 2000, 1559–1569

View Article Online
29 M. P. Georgiadis and E. A. Couladouros, J. Heterocycl. Chem.,
1991, 28, 1325.
48 J. E. Brumwell, N. S. Simpkins and N. K. Terrett, Tetrahedron, 1994,
50, 13533.
30 C. J. Pouchert and J. Behnke, The Aldrich Library of 13C and 1H FT
NMR Spectra, Aldrich Chemical Company, Milwaukee, 1993,
furfuryl alcohol, vol. 3, 18A.
49 R. J. Ferrier, P. M. Petersen and M. A. Taylor, J. Chem. Soc., Chem.
Commun., 1989, 1247; R. J. Ferrier and P. M. Petersen, Tetrahedron,
1990, 46, 1.
31 F. M. Hauser, S. R. Ellenberger and W. P. Ellenberger, Tetrahedron
Lett., 1988, 29, 4939 and references therein.
32 G. Lauer and F. Oberdorfer, Angew. Chem., Int. Ed. Engl., 1993, 32,
272.
50 G. E. Keck and J. B. Yates, J. Organomet. Chem., 1983, 248,
C21.
51 W. E. Lindesell, P. N. Preston and A. B. Rettie, Carbohydr. Res., 1994,
254, 311.
33 J. Marco-Contelles and M. Gema, Synth. Commun., 1997, 27, 725.
34 B. Fraser-Reid, A. McLean and E. W. Usherwood, J. Am. Chem.
Soc., 1969, 91, 5392.
35 S. K. Chaudhary and O. Hernandez, Tetrahedron Lett., 1979, 99.
36 E. J. Corey and A. Ventateswarlu, J. Am. Chem. Soc., 1972, 94, 6190.
37 H. S. Rhinesmith, J. Org. Chem., 1975, 40, 1773.
38 W. J. Balfour, C. G. Grieg and S. Visaisouk, J. Org. Chem., 1974, 40,
725; M. E. Jung and K. R. Buszek, J. Am. Chem Soc., 1988, 110,
3965.
52 P. W. Collins, C. A. Jung, A. Gasiecki and R. Pappo, Tetrahedon
Lett., 1978, 3187; M. E. Jung and L. A. Light, Tetrahedron Lett.,
1982, 23, 3851; S.-M. L. Chen, R. E. Schaub and G. V. Grudzinskas,
J. Am. Chem. Soc., 1978, 43, 3450.
53 For 6-(π-exo)-endo-trig cyclisations, A. M. Gomez, G. O. Danelon,
S. Valverde and J. Christobal Lopez, J. Org. Chem., 1998, 63, 9626;
erratum, A. M. Gomez, G. O. Danelon, S. Valverde and J. Christobal
Lopez, J. Org. Chem., 1999, 64, 7280.
54 C. Lesueur, R. Nouguier, M. P. Bertrand, P. Hoffmann and A. De
Mesmaeker, Tetrahedron, 1994, 50, 5369; A. De Mesmaeker, P.
Hoffmann and B. Ernst, Tetrahedron Lett., 1989, 30, 57; A. De
Mesmaeker, P. Hoffmann, B. Ernst, P. Hug and T. Winkler,
Tetrahedron Lett., 1989, 30, 6307.
39 B. Neises and W. Steglich, Angew. Chem., Int. Ed. Engl., 1978, 17,
522.
40 E. Haslam, Tetrahedron, 1980, 36, 2409.
41 L. Balas, B. Jousseaume and B. Langwost, Tetrahedron Lett., 1989,
30, 4525.
42 G. Stork and R. Mook, Jnr., J. Am. Chem. Soc., 1987, 109, 2829.
43 S. Satoh, M. Sodeoka, H. Sasai and M. Shibasaki, J. Org. Chem.,
1991, 56, 2278.
44 E. Lee, S.-B. Go and K. W. Jung, Tetrahedron Lett., 1989, 30, 827.
45 E. Lee, C.-U. Hur and J.-H. Park, Tetrahedron Lett., 1989, 30,
7219.
46 E. Lee, C. H. Yoon, T. H. Lee, S. Y. Kim, T. J. Ha, Y.-S. Sung,
S.-H. Park and S. Lee, J. Am. Chem. Soc., 1998, 120, 7469;
cf. C. Andres, J. P. Duque-Soladana and R. Pedrosa, Chem.
Commun., 1999, 31.
47 Y. Chapleur and N. Moufid, J. Chem. Soc., Chem. Commun., 1989,
39; N. Moufid and Y. Chapleur, Tetrahedron Lett., 1991, 32, 1799;
N. Moufid, Y. Chapleur and P. Mayon, J. Chem. Soc., Perkin Trans.
1, 1992, 991; N. Moufid, Y. Chapleur and P. Mayon, J. Chem. Soc.,
Perkin Trans. 1, 1992, 999; Structure 14 in the last paper is the
1-O-(prop-2Ј-ynyl) derivative, whereas the correct structure is the
1-O-(but-3Ј-ynyl)-derivative.
55 A. J. Fatiadi, Synthesis, 1976, 65; J. S. Pizey, Synthetic Reagents,
vol. 2, pp. 143–174, Halstead Press, New York, 1974.
56 B. Fraser-Reid, A. McLean and E. W. Usherwood, J. Am. Chem.
Soc., 1969, 91, 5392; B. Fraser-Reid, A. McLean, E. W. Usherwood
and M. Yunker, Can. J. Chem., 1970, 48, 2877.
57 HiMass, D. R. Kelly, unpublished work, 1991.
58 Multiplet, release NMRUC51, D. R. Kelly, unpublished work, 1995.
59 RACCOON, P. F. Schatz, University of Wisconsin, 1988.
60 PC-Model, Version 3.2, Sierra Software, Bloomington, IN, 1989.
61 H–C(sp3)–C(sp3)–H; C. A. G. Haasnoot, F. A. A. M. De Leeuw and
C. Altona, Tetrahedron, 1981, 36, 2783; cf. C. Altona, R. Francke,
R. De Haan, J. H. Ippel, G. J. Daalmans, A. J. A. W. Hoekzema and
J. van Wijk, Magn. Reson. Chem., 1994, 32, 670; H–C(sp3)–C(sp2)–
H; E. W. Garbish, J. Am. Chem. Soc., 1964, 86, 5561.
62 E. Pretsch, W. Simon, J. Seibl and T. Clerc, Tables of Spectral Data
for Structure Determination of Organic Compounds, Springer-
Verlag, Berlin, 1989, p. H215. Implemented in the computer program
Shoolery’s Rules, S. D. Bull, unpublished work, 1989.
J. Chem. Soc., Perkin Trans. 1, 2000, 1559–1569
1569