Cycloaddition and one-carbon homologation studies in the synthesis of advanced iridoid precursors

Article abstract of DOI:10.1039/b902452b

A Diels-Alder cycloaddition approach to the sweroside aglycone intermediate of iridoids was explored using silylated butenolides and levoglucosenone as dienophiles under both Lewis acid and thermal conditions. Results of this study reveal no evidence that

Full text of DOI:10.1039/b902452b

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/obc | Organic & Biomolecular Chemistry
Cycloaddition and one-carbon homologation studies in the synthesis of
advanced iridoid precursors†
Anne T. Stevens,a Mino R. Caira,a James R. Bulla and Kelly Chibale*a,b
Received 5th February 2009, Accepted 4th June 2009
First published as an Advance Article on the web 8th July 2009
DOI: 10.1039/b902452b
A Diels–Alder cycloaddition approach to the sweroside aglycone intermediate of iridoids was explored
using silylated butenolides and levoglucosenone as dienophiles under both Lewis acid and thermal
conditions. Results of this study reveal no evidence that using less sterically demanding derivatives
compromise the diastereofacial selectivity of the cycloaddition using silylated butenolides. Further
chemistry performed on cycloadducts concentrated on the identification and management of
methodologies suitable for its conversion into sweroside aglycone. During the course of these studies, a
dehydrative cyclisation onto a preformed tetrahydrofuran ring to a bis-tetrahydrofuranoid moiety was
unravelled. In addition studies on levoglucosenone-derived cycloadducts provide extensive insight into
the conformational behaviour and reactivity. Further, the X-ray crystal structure of an alcohol
intermediate from one-carbon homologation studies provided the first structural evidence confirming
the diastereoselectivity of the cycloaddition procedure.
Key structures identified during the retrosynthetic analysis
are depicted in Scheme 1. Deprotection of the thioacetal with
Introduction
Chiral a,b-butenolides or 2–5(H)-furanones have frequently been
utilised as chiral synthons in the enantioselective synthesis of
natural products. Examples include the synthesis of an aflatoxin1
and the preparation of the ABC ring of paclitaxel (TaxolTM).2
Although the use of acrylic acid derivatives as dienophiles in
enantioselective Diels–Alder cycloadditions has been extensively
explored,3 that of butenolides (as cyclic acrylate equivalents) is
limited. We reasoned that the high degree of stereoselectivity
imparted by Diels–Alder methodology combined with the high
density of functionality associated with butenolides make this
an attractive option for the synthesis of sweroside aglycone 1
(Fig. 1) as part of our continuing efforts in iridoid synthesis.4,5
For this synthesis, the cycloadduct obtained from the reaction of
4-tert-butyldiphenylsilyloxymethyl substituted butenolide 2 with
butadiene was identified as a suitable chiral intermediate.
concomitant ring closure to form the dihydropyran moiety [step
(a)] would deliver the target, sweroside aglycone 1. The one-carbon
homologation represented in step (b) is an overall conversion
of hydroxyl functionality into a formyl group. Ring opening
of the butenolide residue, by reducing the lactone to a lactol,
followed by aldehyde trapping to give the thioacetal-alcohol
intermediate is depicted in step (c). The cycloaddition reaction
of a silylated butenolide followed by a series of functional
group interconversions involving oxidation and chemoselective
lactonization is depicted in step (d).
Fig. 1 Chemical structures of sweroside aglycone, butenolide, levoglu-
cosenone ent-levoglucosenone and cycloadduct.
Scheme 1 Retrosynthetic steps identified for the synthesis of 1 from 2.
Many enantioselective syntheses of the hydroxymethyl buteno-
lides, from which 2 can be accessed in a simple silyla-
tion step, are known. These are based on (i) transformation
of intermediates from the chiral pool, and (ii) asymmetric
transformations on achiral starting materials. From the chiral
pool, the desired (4S)-enantiomer has been synthesised from
(S)-glutamic acid,6 D-ribonolactone,7 D-mannitol8 and serine-
derived isopropylidineglycelaldehyde.9 The latter procedure was
selected for this work because it is well described and the starting
material is inexpensive.
aDepartment of Chemistry, University of Cape Town, Rondebosch, 7701,
South Africa. E-mail: Kelly.Chibale@uct.ac.za; Fax: 27 21 689 7499; Tel: 27
21 650 2553
bInstitute of Infectious Disease and Molecular Medicine, University of Cape
Town Medical School, Cape Town, South Africa. E-mail: Kelly.Chibale@
uct.ac.za; Fax: 27 21 689 7499; Tel: 27 21 650 2553
† Electronic supplementary information (ESI) available: Experimental
section, spectroscopic and crystallographic data. CCDC reference number
731873. For ESI and crystallographic data in CIF or other electronic
format see DOI: 10.1039/b902452b
This journal is
The Royal Society of Chemistry 2009
Org. Biomol. Chem., 2009, 7, 3527–3536 | 3527
©

View Article Online
The Diels–Alder adduct arising from the reaction between 2
and butadiene was first reported by Mann et al.8 who performed
a Lewis-acid catalysed cycloaddition to give the corresponding
cycloadduct in 76% yield. The product was then used in the
enantioselective synthesis of a prostaglandin analogue.10,11 The
Mann report prompted a publication by Ortun˜o et al.12 in which
thermal cycloadditions with non-silylated analogues of 2 are
reported. Further work by this group includes studies on the regio-
and endo/exo-selectivities of the reactions of these butenolides
with isoprene and cyclopentadiene.13 The latter diene results in
polyfunctional norbornene-type derivatives that have been used
in synthesis of several biologically active compounds.13
On the other hand levoglucosenone 3 has been used as a
chiral template in natural product synthesis and its chemistry
has been extensively explored and documented.14 More recently
levoglucosenone has been employed in the synthesis of chiral
cyclopropanecarboxylic acids.15,16 For our work we selected lev-
oglucosenone from the chiral pool as a useful dienophile for
the synthesis at hand because the cycloadduct 4, obtained from
the Diels–Alder reaction with butadiene, is highly functionalised.
If the internal acetal moiety is unravelled, the conversion of 4
into the enantiomer of the desired hydrocarbon skeleton can be
envisaged. Levoglucosenone could thus provide access to the ent-
series of secoiridoid intermediates, thus expanding the scope for
biological testing, whilst also providing a model for the synthesis
of the natural secoiridoid series, starting from ent-levoglucosenone
5 which is available synthetically.
A retrosynthetic analysis of the proposed model synthesis using
levoglucosenone is illustrated in Scheme 2. Deprotection of the
thioacetal in step (a) would provide the enolic participant for the
formation of the dihydropyran moiety in ent-sweroside aglycone,
whilst the electrophilic carbonyl participant would be expected to
arise from oxidative cleavage of the vicinal diol as shown in step
(b). Access to the glycol moiety of levoglucosenone derivatives by
trapping of the carbonyl group as a cyclic thioacetal [step (c)] has
been described.17 In step (d) oxidative cleavage of the cyclohexenyl
olefin followed by reduction of the termini to give hydroxyethyl
groups was envisaged. Hydrolysis of the nitrile moiety to give a
carboxyl group, followed by chemoselective d-lactone formation
would allow selective functionalisation of the remaining hydroxyl
group to give the required terminal olefin. Step (e) represents a
one-carbon homologation at C-8 on 4. The presence of carbonyl
functionality provides wide scope for one-carbon homologation.
The scope for direct carboxylation is limited so the introduction
of a nitrile group, from which the carboxyl functionality could be
revealed by acid hydrolysis at an appropriate stage, was planned.
Results and discussion
Cycloaddition of silylated butenolides
Of the hydroxymethyl protecting groups utilised in the cycload-
dition reactions cited above, a silylated version was selected
as a starting point for this synthesis. The facility with which
chemoselective desilylation can be achieved using the fluoride
ion under conditions that are compatible with most functional
groups was an important consideration. In order to increase
the scope for applying chemoselective differentiation of similar
functionality should this be required later, an investigation of
the Diels–Alder reaction of butadiene with butenolides bearing
alternative silyl groups was planned. The more robust examples
of commonly used silyl protecting groups, triisopropylsilyl (TIPS)
and tert-butyldimethylsilyl (TBDMS) were thus included with tert-
butyldiphenylsilyl (TBDPS) in the investigation.
The hydroxy butenolide precursor was routinely prepared on a
10 g scale from D-mannitol8,9 Silylation under standard conditions
afforded the butenolides 6, 7 and 8. Earlier attempts to reproduce
the catalytic efficiency of the aluminium trichloride catalysed
cycloaddition between 2 and butadiene as described by Mann8
provided erratic results. An alternative aluminium-based Lewis-
acid catalyst, ethylaluminium dichloride was used in the reaction
with 2a to deliver 4, Scheme 3. The successful outcome of this
reaction was reproducible, but was however followed by less
encouraging results with 2a and 2b.
Scheme 3 Reagents and conditions: (a) butadiene, EtAlCl2 (0.3 mol
equiv.), CH2Cl2, 55 ◦C, 7 d, 70% (6), 41% (7), 14% (8)/15% (9); (b)
butadiene, 210 ◦C, 63% (6), 68% (7), 65% (8); (c) TBAF–HOAc, THF,
10 ◦C, 78% or HF–MeCN, MeCN, 25 ◦C, 75%.
The use of TBDPS as a protecting group in the original work8
was based on the desire to maximise facial differentiation on
the dienophile during the cycloaddition. Although the yields of
cycloadducts 7 and 8 obtained in our work are poor, they were the
only diastereomers detected. This study has thus given no evidence
that using less sterically demanding derivatives compromised the
diastereofacial selectivity of the cycloaddition. Although the yield
of 6 was as expected, the lability of the TIPS and TBDMS ethers
under these conditions was confirmed by the poor recoveries of
7 and 8 as well as the presence of a spirolactone and deprotected
cycloadduct 9. The spirolactone has been isolated previously from
the thermal cycloaddition of the hydroxybutenolide precursor and
its acetate to butadiene.12
The complexity of the product mixtures containing 7 and
8 render this option unattractive for the preparation of these
cycloadducts as starting materials under Lewis acid conditions.
Scheme 2 Retrosynthetic plan for the synthesis of ent-1 from the known
cycloadduct 4.
3528 | Org. Biomol. Chem., 2009, 7, 3527–3536
This journal is
The Royal Society of Chemistry 2009
©

View Article Online
The thermal cycloaddition of these butenolides with butadiene
was thus investigated. These reactions provided the desired
cycloadducts exclusively and no decomposition products were
detected. However, isolation of the products was difficult due
to extensive butadiene polymerisation under thermal conditions
and careful chromatography using large proportions of silica
gel was required in order to isolate clean cycloadducts. The
addition of hydroquinone, as described by Ortun˜o12 did not
inhibit polymerisation. This is probably attributable to the higher
dilutions of the butenolide in butadiene used in that study.
This dilution was considered excessive for the preparative scale
conditions sought. Owing to the problems encountered during the
isolation of products and the safety problems associated with the
use of butadiene under thermal conditions, this approach was not
considered suitable for the production of multigram quantities of
the cycloadducts. The modified Mann procedure was thus used
to react 5 g quantities of the TBDPS protected parent butenolide
and further work towards the synthesis of sweroside aglycone 1
was made on the cycloadduct 4.
Scheme 4 Reagents and conditions: (a) DIBAH, toluene, -78 ◦C, 91%;
(b) ethanedithiol, TiCl4, CH2Cl2, -78 ◦C, 80%; (c) LiBH4, THF, RT, 85 h,
63%; (d) TIPSCl, imidazole, MeCN, 96%; (e) (i) LAH, Et2O, reflux, 1 h,
(ii) acetone, p-TsOH, CuSO4 (anhydrous), reflux, 28% (over two steps); (f)
MeONHMe·HCl, Me2AlCl, CH2Cl2.
carboxyl moiety in sweroside aglycone. Once again, no reaction
occurred at the free hydroxyl group under standard tosylation
conditions.
Treatment of 6 with tetrabutylammonium fluoride (TBAF) at
room temperature gave an inseparable 85 : 15 (by NMR) mixture
of 9, and an unidentified by-product. Based on the duplication of
signals in the NMR spectra of the mixture, the by-product was
assumed to be the trans diastereomer of 9, which had resulted
from epimerisation a to the carbonyl group, via enolisation under
the basic reaction conditions. Although this observation was
unexpected, literature precedent for the abstraction of a-acidic
protons in this reaction medium does exist.18
When the reduction of 6 to 12 was attempted with NaBH4,
DIBAH, or LAH, silyl deprotection accompanied reduction. An
excess of LAH in refluxing diethyl ether for 60 min provided
complete conversion into a single product that was assumed to be
the triol. Because the extraction and purification of the triol were
hampered by its polarity, the 1,2-diol moiety in the crude reduction
product was subjected to acetonide formation conditions22 to give
14, Scheme 4. However, the yield of this compound was too low to
be useful in a total synthesis and, despite numerous attempts using
different work-up and extraction procedures for the reduction, it
could not be optimised.
Although the reduction–aldehyde trapping sequence applied
above was successful, its timing in the synthetic route is prob-
lematic owing to the susceptibility of sulfur to oxidation under the
conditions for oxidative cleavage of the olefin. Another possibility
examined was the opening of the lactone to the Weinreb amide 15.
Chemoselective reduction of the Weinreb amide to an aldehyde (as
was targeted) in an excess of DIBAH is well known.23 Treatment
of 6 with Weinreb’s salt and dimethylaluminium chloride gave
complete reaction to a single product. The amide appeared to
relactonise during work up, chromatography, and in CDCl3, and
could not be isolated and characterised, Scheme 4.
Reaction at 0 and -15 ◦C gave a similar diastereomeric mixture.
Maintaining the temperature of the reaction mixture at -20 to
-30 ◦C provided a single diastereomer, but the reaction was slow.
In order to temper the basicity of TBAF, buffering of the reaction
medium with acetic acid has been reported.19 The reaction of
6 with equimolar quantities of TBAF and acetic acid at 10 ◦C
provided a single diastereomer, but 9 could only be recovered in
78% yield. Hydrogen fluoride was also used as a fluoride source
for deprotection, although large excesses of HF were required in
order to achieve complete reaction. After 20 min at 40 ◦C, 55% of
the desired product was recovered. This was improved to 75% by
decreasing the reaction temperature to 25 ◦C for 24 h. At 0 ◦C, no
reaction was observed.
The relative ease with which C-7a was deprotonated during
deprotection of 6 indicated that reduction of the lactone moiety
was required prior to deprotection. The formyl level of oxidation
was required in any event, so 6 was reduced with DIBAH at low
temperature to the diastereomeric lactol mixture 10, Scheme 4.
A radical deoxygenation process on 11 was envisaged, which
after appropriate transformation would allow an enolate-mediated
one-carbon homologation step. All attempts to produce a radical
deoxygenation precursor (xanthate ester, thiocarbonyl imidazole
or phenoxythiocarbonyl)20 failed, possibly due to limited reagent
access to the hydroxyl which was a to a tertiary carbon and b to
the bulky TBDPS moiety.
Ozonolysis of 6 followed by a reductive work up with NaBH4
proceeded efficiently to give diol 16. The analytical and spectral
data gathered could not preclude the possibility of formation of
the alternative g-lactone or the d-lactone. The diol 16 was thus
bis-esterified with 3,5-dinitrobenzoyl chloride in pyridine to give
17, Scheme 5.
With the oxidative cleavage product in hand, the reduction–
aldehyde trapping sequence was attempted. The unprotected diol
16 in toluene at -78 ◦C was treated with 3 equivalents of DIBAH.
The product 19 was extremely polar and full characterisation
was hampered by a minor co-eluting impurity. Acetylation of
product 19 provided internal acetal 20. The formation of 19 was
rationalised as a dehydration of the reduced product under the
acidic work up conditions. This bis-tetrahydrofuranoid moiety
is present in a number of natural products, in particular the
aflatoxins. Although dehydrative cyclisation onto a preformed
tetrahydrofuran ring has been used in the total synthesis of
aflatoxins,24,25 this result produces a useful synthon, easily derived
8
The exhaustive reduction of 6 to 12 with LiBH4 was followed
by the selective protection of the primary hydroxyl group as
a triisopropyl silyl ether 13. In this instance, tosylation of the
secondary hydroxyl group, followed by an SN2 substitution with
CN-,21 to give the corresponding nitrile was planned. The nitrile,
after hydrolysis at an appropriate point, would provide the desired
This journal is
The Royal Society of Chemistry 2009
Org. Biomol. Chem., 2009, 7, 3527–3536 | 3529
©

View Article Online
material 4. It has been shown that the oxacyclohexanone ring
adopts a chair-like conformation. Ignoring pendant functionality,
stereoelectronic control analogous to that demonstrated in cyclo-
hexanone could be expected, where the hydride is delivered from
the exo face leading to an equatorial hydroxy substituent on the
endo face of the oxacyclohexanoid ring,29 Fig. 2.
Scheme 5 Reagents and conditions: (a) O3, MeOH, -78 ◦C, then NaBH4,
RT, 88%; (b) ArC(O)Cl, pyr, 80%; (c) MEMCl, iPr2NEt, CH2Cl2, 92%;
(d) DIBAH, toluene, -78 ◦C, (ii) Ac2O, DMAP, pyr; (e) DIBAH, toluene,
-78 ◦C, 93%; (f) ethanedithiol, TiCl4, CH2Cl2, -84 ◦C, 32%.
Fig. 2 Hydride approach to the 8-keto group of 4.
On this basis, 24 should have been the favoured alcohol.
However, the presence of the cyclohexenoid ring introduces a steric
factor to this reduction. The hydride was preferentially delivered
from the a-face, which is less sterically hindered, thus resulting
in the axially substituted hydroxyl group on the b-face that was
observed in the major product, 23. This argument is supported by
the fact that the stereoselectivity improved when hydride delivery
took place from the more sterically demanding reducing agent
L-Selectride.
from the chiral pool for the enantioselective synthesis of similar
systems. This route was deemed unsuitable for the synthesis at
hand and was accordingly abandoned. Nevertheless it proved use-
ful (albeit serendipitously) for unravelling a bis-tetrahydrofuran
moiety
In order to prevent the observed dehydration, the free hydroxyl
groups in 16 were protected as methoxyethoxymethyl (MEM)
ethers to give 18. The DIBAH reduction of 18 produced a
diastereomeric mixture of the hemiacetals 21. Treatment of 21 with
ethanedithiol and TiCl4 gave a very poor yield of the dithiolane
22, Scheme 5. The MEM protection was selected because it would
allow chemoselective TBAF-mediated deprotection of the TBDPS
moiety in a more advanced intermediate although the labile nature
of MEM acetals in Lewis acidic conditions was recognised.26
One-carbon homologation
Although the carbonyl functionality at C-8 in 4 provided a func-
tional handle for one-carbon homologation, a suitable method for
the direct conversion of the carbonyl group into a carboxylic acid
was not obvious. However, the one-step conversion of ketones
into nitriles in a reductive cyanation reaction with tosylmethyl
isocyanide (TosMIC) is known and was envisaged to lead to the
corresponding carboxylic acid on pyrolysis.30,31
The conditions described32 were applied to the reductive cyana-
tion of 4, and produced a mixture so complex that neither starting
material nor any single product could be isolated from it. A similar
scenario was encountered when the traditional reaction conditions
of van Leusen30,33 were used. The reaction outcome could not be
improved despite extensive efforts using varying rates of reagent
addition and reaction temperatures.
In the light of the failure of the reductive cyanation reaction,
displacement chemistry was considered for the introduction of the
nitrile group. The alcohol 23 was available in high yield from the
stereoselective reduction of 4. Although the hydroxyl moiety on
the oxacyclohexane in 23 ring was axial, which is the preferred
orientation for leaving groups in SN2 reactions, it was recognised
that the approach trajectory of the nucleophile would be syn to the
oxymethylene bridge and steric hindrance could compromise the
substitution reactivity. Several procedures utilising SN2 method-
ology have been developed for the introduction of cyanide.34 The
most common of these is the displacement of sulfonic acid esters
by the cyanide anion. A non-aqueous cyanation procedure using
lithium cyanide has been described to be more reactive than the
traditional sodium and potassium variants that frequently require
the use of phase transfer catalysts and extreme temperatures to
achieve reaction. The hydroxyl group in 23 was converted into the
tosylate 25, to act as a more electrophilic partner. Heating of 25 at
80 ◦C in the presence of LiCN in DMF produced a single product.
Model studies using levoglucosenone
Printed newspaper was used as the source of cellulose for pro-
ducing levoglucosenone.3 The paper was treated as described in a
literature procedure.27 The recovery of 3 was variable and averaged
0.8%. Following literature procedures,28 the thermal cycloaddition
of levoglucosenone was performed at 160 ◦C for 3 h to give
after chromatography a product (72%) that was assigned as the
expected cycloadduct 4 (Scheme 6). To prepare intermediates for
possible 8-homologation studies, hydride-mediated reduction of 4
was attempted. The reaction with sodium borohydride afforded
the diastereomeric alcohols 23 and 24 in 65 and 19% yields
respectively. A more sterically demanding reducing agent was used
in an attempt to improve the diastereoselectivity of the reduction.
Reduction with L-Selectride in THF at -78 ◦C afforded only 23 in
93% yield.
Scheme 6 Reagents and conditions: (i) butadiene, 160 ◦C; (ii) butadiene,
EtAlCl2 (0.3 mol equiv.), CH2Cl2, 0 ◦C; (ii) NaBH4, MeOH, 0 ◦C,
(iiii) L-Selectride, THF, -78 ◦C.
The rationalisation of the stereochemical outcome of this
reaction requires insight into the conformation of the starting
3530 | Org. Biomol. Chem., 2009, 7, 3527–3536
This journal is
The Royal Society of Chemistry 2009
©

View Article Online
Instead of substitution, the competing elimination reaction had
occurred exclusively to give the olefin 26 (Scheme 7).
Scheme 8 Mechanism for the SmI2-mediated reductive elimination in
cyanophosphates.
The crystalline nature of alcohol 30 provided an opportunity
for crystal structure determination (see ESI†). A search of the
Cambridge Crystal Database41 showed that no compounds bear-
ing the levoglucosenone cycloadduct motif had been registered.
The crystal structure obtained as shown in Fig. 3 confirmed the
relative stereochemistry at the bridgehead positions C-2 and C-7,
thus confirming the diastereoselectivity that had been assumed
for the cycloaddition reaction. Calculated bond lengths and
angles are consistent with the structural assignment. Puckering
analysis confirmed that the six-membered ring C2→C7 adopts
the expected half-chair conformation (puckering parameters Q =
Scheme
7
Reagents and conditions: (i) pTsCl, DMAP, pyr, 97%;
(ii) LiCN, DMF, 80 ◦C, 70%; (iii) Dess–Martin periodinane; (iv) LiCN,
(EtO)2P(O)Cl, DMF–THF, (v) SmI2, HMPA, tBuOH, THF, 69%;
(vi) Ph3PCH3I, nBuLi, THF, 0–25 ◦C, 85%; (vii) 9-BBN, THF, then 1 M
NaOH, H2O2; (iv) 8 M CrO3, acetone, -4 ◦C, 32%; (viii) (COCl)2, DMSO,
CH2Cl2, -78 ◦C, then Et3N, -78 to 25 ◦C.
◦
◦
˚
0.398(2) A, q = 130.7(3) , f = 138.4(4) ) while the six-membered
Alternative procedures for this substitution, which make use of
NaCN and KCN in the presence of crown ethers or phase-transfer
catalysts are used for this transformation,34 but in the light of the
facility with which elimination appears to have occurred in 25,
these were not considered. In the light of these factors, alternative
homologation procedures were investigated preferentially.
The addition of cyanide to a carbonyl group to give a cyanohy-
drin was first reported in 183235 and numerous efficient procedures
for this transformation have since been developed.36 Reaction
conditions whereby the cyanohydrin hydroxyl group is trapped as a
phosphate ester37,38 to give a cyanohydrin O,O¢-diethyl phosphate
(cyanophosphate) were attractive here because there is literature
precedent for the reductive elimination of the phosphate group
to give a nitrile.39 The reaction of 4 with diethylchlorophosphate
and lithium cyanide38 afforded cyanophosphate 27 as a single
diastereomer. Surprisingly, the treatment of 27 with SmI2 in the
presence of tert-butyl alcohol gave 28 as the major product instead
of the expected nitrile.
ring containing atom O10 assumes a twist-chair form (pu◦ckering
◦
˚
parameters Q = 0.635(2) A, q = 18.4(2) , f = 2.8(6) ). The
five-membered ring has an envelope conformation (flap at O10).
An intermolecular hydrogen bond O2¢-H2¢ ◊ ◊ ◊ O11a (a = 1 - x,
1
2
1
2
-
+ y, - z) links molecules into infinite spirals parallel to
the crystallographic b-axis. The configuration at C-8, where the
hydroxymethyl substituent was axially orientated on the a-face
of the oxacyclohexane ring, confirmed the deduction (argued for
the hydride reduction of 4) that reactions at C-8 were under steric
rather than stereoelectronic control.
The conversion of 27 into 28 appears to be one of overall
b-elimination as opposed to reductive elimination. However, the
accepted mechanism40 for the reductive elimination of cyanophos-
phates does not allow for this outcome.
The proposed mechanism is as depicted in Scheme 8. There
is initial dissociative electron transfer to an easily reducible
a-substituent. Subsequent reduction by a second equivalent of
SmI2 generates an enolate that becomes protonated to give the
carbonyl product. It is speculated here that when 27 interacts with
SmI2, the reaction proceeds along a similar path to produce the
product of b-elimination as illustrated in Scheme 8.
As an alternative, the obvious option of Wittig olefination
methodology was applied to the conversion of the carbonyl group
in 4 into a chain-extended carboxyl moiety. A stepwise approach of
methylenation, followed by further functionalisation was adopted.
The methylenated product 29 was afforded by the reaction of 4
with the methyltriphenylphosphorane (Scheme 7).
Fig. 3 X-Ray crystal structure of 30 (thermal ellipsoids are drawn at the
50% probability level).
In order to proceed to the required carboxyl level of oxidation
at C-1¢, Jones oxidation of 30 was investigated. In a reaction
carried out at -4 ◦C the aldehyde 31 proved to be a discrete
and isolable intermediate. However, under the prolonged reaction
This journal is
The Royal Society of Chemistry 2009
Org. Biomol. Chem., 2009, 7, 3527–3536 | 3531
©

View Article Online
times required to achieve complete consumption of the alde-
hyde, extensive decomposition intervened, as evidenced by TLC
monitoring, and yields of 32 were poor. This behaviour was
presumed to be due to the acid-labile nature of the internal
acetal functionality. Other methods of achieving direct oxidation
of the primary alcohol, including reactions with pyridinium
dichromate and ruthenium tetroxide were attempted, but in both
cases chemoselective reaction could not be achieved owing to
the presence of the acetal and olefin moieties respectively. The
conversion of 30 into the aldehyde 31 was most efficiently achieved
using Swern oxidation conditions. However, the oxidation of 31
with sodium chlorite gave a complex mixture from which the
carboxylic acid could not be isolated.
The hydroxymethyl group in 30 was pivaloylated, to give the
ester 33 (Scheme 9). The ozonolysis and in situ reduction followed
by silylation gave 34 in excellent yield. The acetal trapping was
however beset by a deprotection problem. The poor yield of
thioacetal 35 was accompanied by the recovery of TBDPSOH.
This indicated that, although TBDPS ethers are reported to be
stable to BF3·OEt2 in dichloromethane at room temperature,26 silyl
cleavage had been induced under the reaction conditions applied
here. A chemoselective transformation could not be achieved by
varying the temperature or concentration of the reaction.
achieved in a stepwise, but efficient process. The opening of the
internal acetal to reveal the vicinal diol, and the oxidative cleavage
of that diol to give the required aldehyde functionality have also
been successfully demonstrated. Considerable progress has been
made in determining the sequence in which these methodologies
should be applied in the final synthetic route, as well as in
the identification of protecting groups that complement these
chemoselective processes.
In addition, the reduction studies have provided extensive
insight into the conformational behaviour and reactivity of the
tricyclic cycloadduct 4, and allowed the identification of an
artifact from the cycloaddition reaction as a trans epimer of
the primary cycloadduct. Further, the X-ray crystal structure
of 30 has provided the first structural evidence confirming the
diastereoselectivity of the cycloaddition procedure, and of further
transformations at C-8 in 30. Although 28 was not the expected
product, it does contain a dihydropyran ring as well as the masked
carboxyl group in the form of a nitrile. Hydrolytic cleavage of the
enol ether in 28, thereby releasing the vicinal diol for oxidative
cleavage, followed by reclosure of the dihydropyran moiety via
hemiacetal formation, could be envisaged.
Experimental
Hydride reduction of 16
Diisobutylaluminium hydride (1.5 M in toluene, 1.66 cm3,
2.49 mmol) was added to a stirred solution of 16 (275 mg,
◦
0.62 mmol) in toluene (20 cm3) at -78 C. After 5 min 1 M HCl
was added until pH 1 was reached and the mixture was warmed
to 25 ◦C. The toluene was removed under reduced pressure
and the mixture was extracted with ethyl acetate. The organic
phase was dried (MgSO4) and the solvent was removed under
reduced pressure to give a residue (274 mg) which was further
purified by column chromatography on silica gel (12 g) using ethyl
acetate as eluent to give crude 19 (242 mg, ~88%) which required
derivatisation to allow complete characterisation.
Scheme 9 Reagents and conditions: (i) PvCl, Et3N, CH2Cl2, 95%;
(ii) O3, MeOH, -78 ◦C, then NaBH4, 25 ◦C followed by TBDPSCl,
imidazole, DMF, CH2Cl2,97%; (iii) ethanedithiol, BF3·OEt2, CH2Cl2,
0 ◦C; (iv) Pb(OAc)2, 69%; (v) LAH, THF, 0 ◦C, (vi) RuO2 (cat.), NaIO4,
CCl4–MeCN–H2O, 0 ◦C, 93%; (vii) MeI, K2CO3, DMF, 0 ◦C, 99%.
(3aR,6aS)-2-[(S)-tert-Butyldiphenylsilyloxymethyl]-3-[(S)-2-
acetoxyethyl]hexahydrofuro[2,3-b]furan 20. Acetic anhydride
(0.41 cm3, 443 mg, 4.34 mmol) and 4-dimethylaminopyridine
(20 mg, 0.16 mmol) were added sequentially to a stirred solution
of 19 (480 mg, 1.08 mmol) in dry pyridine (10 cm3). After stirring
Despite the poor yield of the thioacetal, and the attendant
necessity to improve these steps for a practical synthesis, the
available material sufficed to demonstrate the oxidative cleavage of
the glycol moiety. Treatment of 35 with lead tetraacetate afforded
a cleavage product that was purified by flash chromatography to
give the aldehyde 36 in excellent yield. No epimerisation a to the
aldehyde carbonyl group could be detected in the NMR spectra
of 36.
In the light of the observed instability of the TBDPS ethers
in 34, an alternative approach to this portion of the synthesis
was embarked on. The pivaloate protection on 34 was removed
by reduction with LAH to give the alcohol 37. The alcohol was
treated with ruthenium tetroxide to give the carboxylic acid 38,
which was characterised as its methyl ester 39 (Scheme 9).
◦
for 1 h at 25 C toluene was added and the volume was reduced
under reduced pressure. The resultant slurry was dissolved in ethyl
acetate, washed with water and brine, dried (MgSO4) and the
solvent was removed under reduced pressure. The residue (497 mg)
was purified by column chromatography on silica gel (50 g) using
ethyl acetate–hexane (3 : 7) an eluent to give the acetate 20 (456 mg,
90%) as an oil, [a]D +16.9 (c 2.0 in CHCl3); nmax(CHCl3)/cm-1
1733 (CO); dH (400 MHz, CDCl3) 1.06 [9H, s, C(CH3)3], 1.59–
1.74 (1H, m, 1¢¢-HA), 1.76–1.98 (3H, m, 4-H2 and 1¢¢-HB), 2.03
[3H, s, C(O)CH3], 2.39–2.50 (1H, m, 3-H), 2.87–2.96 (1H, m, 3a-
H), 3.63–3.78 (2H, m, 2-H, 1¢-HA), 3.80–3.98 (2H, m, 5-H2 and
1¢-HB), 4.03–4.19 (2H, m, 2¢¢-H2), 5.75 (1H, d, J 5.0 Hz, 6a-H),
7.35–7.43 (6H, m, Ar-H), 7.62–7.75 (4H, m, Ar-H); dC (100 MHz,
CDCl3) 19.5 [C(CH3)3], 21.1 [C(O)CH3], 25.4 (C-4), 26.8 (C-1¢¢),
27.0 [C(CH3)3], 39.9 (C-3), 46.1 (C-3a), 63.4 (C-2¢¢), 64.2 (C-1¢),
68.9 (C-5), 83.4 (C-2), 108.7 (C-6a), 127.9(0) and 127.9(3), 129.8(9)
and 129.9(1), 133.4(8) and 133.7(0), 135.8(1) and 135.9(0) (Ar-C)
Conclusion
The oxidative cleavage and functionalisation of the termini to
hydroxyl groups has been achieved in high yields. One-carbon
homologation and functionalisation to a carboxylic acid were
3532 | Org. Biomol. Chem., 2009, 7, 3527–3536
This journal is
The Royal Society of Chemistry 2009
©

View Article Online
and 171.2 [C(O)CH3] (Found: M+ - C4H9, 411.1642. Calc. for
C23H27O5Si: M, 411.1628)
solvent was removed under reduced pressure to give a solid residue
(1.02 g). Chromatography on silica gel (100 g) using ethyl acetate–
hexane (1 : 1) as eluent afforded the cyanophosphate 27 (1.17 g,
61%), mp 85–86 ◦C (from ethyl acetate–hexane); [a]D -37.8 (c 1.3
in CHCl3); nmax(CHCl3)/cm-1 1273 (PO); dH (400 MHz, CDCl3),
1.30–1.37 (6H, m, 2 ¥ OCH2CH3), 1.84–1.93 (1H, m, 2-H), 1.97–
2.07 (1H, m, 3-HA), 2.24–2.34 (1H, m, 6-HA), 2.41–2.54 (2H, m,
3-HB and 6-HB), 2.60–2.68 (1H, m, 7-H), 3.92 (1H, dd, J 7.6 and
5.0 Hz, 11-HA), 4.01 (1H, dd, J 7.6 and 0.8 Hz, 11-HB), 4.09–
4.20 (4H, m, 2 ¥ OCH2CH3), 4.42 (1H, dd, J 4.4 and 0.8 Hz,
1-H), 5.52–5.55 (2H, m, 4-H and 5-H) and 5.95 (1H, s, 9-H);
dC (100 MHz, CDCl3), 15.9 (d, J 6.0 Hz, OCH2CH3), 16.0 (d, J
9.2 Hz, OCH2CH3), 24.8 and 25.1 (C-3 and C-6), 32.7 (C-7), 34.1
(C-2), 64.5 (d, J 24 Hz, OCH2CH3), 64.8 (d, J 24 Hz, OCH2CH3),
67.8 (C-11), 75.8 (C-8), 76.7 (C-1), 100.0 (C-9), 116.8 (CN), 124.0
and 125.1 (C-4 and C-5) (Found: C, 52.6, H, 6.6, N, 4.0%, M+,
343. Calc. for C15H22NO6P: C, 52.5, H, 6.5, N, 4.1%, M, 343).
Reduction of cycloadduct 4
(a) L-Selectride (1.0 M in tetrahydrofuran, 2.7 cm3) was added to
a stirred solution of 4 (400 mg, 2.22 mmol) in toluene (100 cm3)
at -78 ◦C. After 2 h at -78 ◦C, saturated aqueous ammonium
◦
chloride was added and the mixture was warmed to 25 C with
stirring. Sodium hydroxide (1 M, 50 cm3) was added, followed by
hydrogen peroxide (30%, 50 cm3) and the mixture was stirred for
60 min. The aqueous phase was extracted with dichloromethane,
the organic extract was washed with aqueous saturated sodium
hydrogen carbonate and brine, dried (MgSO4) and the solvent
was removed under reduced pressure to give a residue (507 mg)
which was chromatographed on silica gel (50 g) using ethyl
acetate–hexane (3 : 17) as eluent, to give (1S,2S,7R,8R,9R)-10,12-
dioxatricyclo[7.2.1.02,7]dodec-4-en-8-ol 23 (376 mg, 93%), [a]D
-9.3 (c 1.0 in CHCl3); nmax(CHCl3)/cm-1 3525 (OH); dH (400 MHz,
CDCl3), 1.78–1.89 (2H, m, OH and 2-H), 2.01–2.10 (1H, m, 6-HA),
2.11–2.21 (1H, m, 3-HA), 2.27–2.43 (2H, m, 6-HB and 7-H), 2.43–
2.55 (1H, m, 3-HB), 3.40 (1H, br s, 8-H), 3.85 (1H, dd, J 7.2 and
5.2 Hz, 11-HA), 3.93 (1H, dd, J 7.2 and 1.2 Hz, 11-HB), 4.33 (1H,
m, 1-H), 5.33 (1H, d, J 2.2 Hz, 9-H) and 5.78 (2H, m, 4-H and
5-H); dC (100 MHz, CDCl3), 26.0 and 26.1 (C-3 and C-6), 26.9
(C-7), 34.4 (C-2), 67.0 (C-11), 72.6 (C-8), 76.5 (C-1), 103.1 (C-9),
125.7 and 126.3 (C-4 and C-5) (Found: M+ 182.0964. Calc. for
C10H14O3: M, 182.0943).
(1S,4aR,8aS)-1-Hydroxymethyl-4a,5,8,8a-tetrahydro-1H-iso-
chromene-4-carbonitrile 28. 1,2-Diiodoethane (564 mg,
2.0 mmol) in tetrahydrofuran (20 cm3) was slowly added to
samarium (451 mg, 3.0 mmol) with rapid stirring. The solution
was refluxed until a deep blue colour developed (~60 min) after
which it was cooled and hexamethylphosphoramide (0.01 cm3,
0.06 mmol) was added. To this was added a solution of 27
(200 mg, 0.58 mmol) and tert-butyl alcohol (43 mg, 0.58 mmol) in
tetrahydrofuran (6 cm3). After stirring at 25 ◦C for 4 h, 1 M HCl
was added (till pH 2 was reached). The mixture was diluted with
ethyl acetate, washed with aqueous sodium thiosulfate (1 M),
dried (MgSO4), and the solvent was removed under reduced
pressure to give an oily residue (140 mg). Chromatography on
silica gel (15 g) using ethyl acetate–hexane (2 : 3) as eluent yielded
the carbonitrile 28 as a colourless oil (76 mg, 69%), [a]D +18.6
(c 1.3 in CHCl3); nmax(CHCl3)/cm-1 3425 (OH), 2212 (CN); dH
(400 MHz, CDCl3), 1.80 (1H, br s, OH), 1.99–2.13 (2H, m, 5-HA.
and 8-HA), 2.16–2.31 (2H, m, 8a-H and 8-HB), 2.42–2.51 (1H, m,
5-HB), 2.52–2.59 (1H. m, 4a-H), 3.74 (1H, dd, J 12.4 and 5.6 Hz,
1¢-HA), 3.80 (1H, dd, J 12.4 and 3.4 Hz, 1¢-HB), 4.08 (1H, ddd, J
7.6, 5.6 and 3.4 Hz, 1-H), 5.59–5.62 (2H, m, 6-H and 7-H) and
7.10 (1H, s, 3-H); dC (100 MHz, CDCl3), 24.4 (C-8), 27.9 (C-4a),
28.3 (C-5), 29.0 (C-8a), 62.7 (C-1¢), 78.6 (C-1), 92.5 (C-4), 118.4
(CN), 123.2 and 124.3 (C-6 and C-7) and 155.6 (C-3) (Found: M+,
191.0948. Calc. for C11H13NO2: M, 191.0946)
(b) Sodium borohydride (90 mg, 2.44 mmol) was added
to a stirred solution of 4 (400 mg, 2.22 mmol) in methanol
(10 cm3) at 0 ◦C. The mixture was stirred at 0 ◦C for 1 h after
which saturated ammonium chloride was added and the volatile
material was removed under reduced pressure. The aqueous
residue was extracted with ethyl acetate, dried (MgSO4) and
the solvent was removed under reduced pressure to give a
mixture of the alcohols (500 mg). Chromatography on silica gel
(50 g) using ethyl acetate–hexane (3 : 17) as eluent afforded 23
(262 mg, 65%), followed by (1S,2S,7R,8S,9R)-8-hydroxy-10,12-
dioxatricyclo[7.2.1.02,7]dodec-4-ene 24 (78 mg, 19%), mp 92–
93 ◦C (from ethyl acetate–hexane); [a]D -82.4 (c 0.4 in CHCl3);
n
max(CHCl3)/cm-1 3574 (OH); dH (400 MHz, CDCl3), 1.61 (1H,
br d, OH), 1.87–2.60 (3H, m, 2-H, 3-HA, 7-H), 2.09–2.20 (1H, m,
6-HA), 2.30–2.45 (2H, m, 3-HB, 6-HB), 3.38 (1H, dd, J 9.7 and
1.3 Hz, 8-H), 3.85 (1H, dd, J 7.2 and 4.8 Hz, 11-HA), 3.90 (1H,
dd, J 7.2 and 0.8 Hz, 11-HB), 4.29 (1H, d, J 5.2 Hz, 1-H), 5.32
(1H, d, J 1.3 Hz, 9-H), 5.55–5.63 (1H, m, 5-H) and 5.65–5.75 (1H.
m. 4-H); dC (100 MHz, CDCl3), 23.8 (C-6), 24.2 (C-3), 33.7 (C-7),
35.7 (C-2), 67.9 (C-11), 70.1 (C-8), 77.3 (C-1), 103.0 (C-9), 123.8
(C-5) and 125.0 (C-4) (Found: C, 65.9, H, 7.7%, M+ 182. Calc. for
C10H14O3: C, 65.5, H, 7.8%, M, 182).
(1S,2S,7R,9R)-8-Methylene-10,12-dioxatricyclo[7.2.1.02,7 ]-
dodec-4-ene 29. n-Butyllithium (10 M solution in hexane,
2.40 cm3) was added to a stirred slurry of methyltriphenylphos-
phonium iodide (9.42 g, 23.3 mmol) in tetrahydrofuran (50 cm3) at
0 ◦C. The resulting solution was warmed to 25 ◦C and stirred for
2 h. The solution was cooled to 0 ◦C and a solution of 4 (2.10 g,
11.7 mmol) in tetrahydrofuran (20 cm3) was slowly added. The
◦
(1S,2S,7R,8S,9R)-8-Hydroxy-10,12-dioxatricyclo[7.2.1.02,7 ]-
dodec-4-en-8-carbonitrile diethylphosphate 27. Diethyl chloro-
phosphate (1.12 cm3, 1.34 g, 7.75 mmol) and lithium cyanide
(0.5 M in N,N -dimethylformamide, 22.0 cm3) were added to
a stirred solution of 4 (1.00 g, 5.56 mmol) in tetrahydrofuran
(30 cm3). After stirring at 25 ◦C for 60 min, water was added
and the mixture was concentrated under reduced pressure. The
aqueous residue was extracted with ethyl acetate, the organic
extract was washed with water and brine, dried (MgSO4), and the
reaction was warmed to 25 C and stirred for 18 h. The mixture
was acidified with 1 M HCl and concentrated. The aqueous residue
was extracted with dichloromethane, the organic extract was dried
(MgSO4) and the solvent was removed under reduced pressure.
The yellow residue (10.90 g) was adsorbed onto silica gel (25 g)
and then chromatographed on silica gel (200 g) using ethyl acetate–
hexane (1 : 9) as eluent to yield olefin 29 (1.77 g, 85%), mp 88–90 ◦C
(from hexane); [a]D 32.9 (c 1.0 in CHCl3); dH (400 MHz, CDCl3),
1.92–1.96 (1H, m, 3-HA), 1.97–2.06 (1H, m, 2-H), 2.22–2.33 (2H,
This journal is
The Royal Society of Chemistry 2009
Org. Biomol. Chem., 2009, 7, 3527–3536 | 3533
©

View Article Online
m, 6-H2), 2.34–2.49 (1H, m, 3-HB), 2.93–3.00 (1H, m, 7-H), 3.89
(1H, dd, J 7.2 and 4.9 Hz, 11-HA), 4.06 (1H, dd, J 7.2 and 0.8 Hz,
11-HB), 4.35 (1H, dd, J 4.9 and 2.1 Hz, 1-H), 4.70 (1H, d, J 2.3 Hz,
1¢-HA), 4.96 (1H, d, J 2.3 Hz, 1¢-HB), 5.53 (1H, s, 9-H) and 5.58–
5.66 (2H, m, 4-H and 5-H); dC (100 MHz, CDCl3), 23.7 (C-3),
24.4 (C-6), 30.2 (C-7), 38.4 (C-2), 67.7 (C-11), 77.2 (C-1), 105.0
(C-9), 107.6 (C-1¢), 123.2 and 125.0 (C-4 and C-5) and 146.4 (C-8)
(Found: C, 73.9, H, 8.0%, M+ 178. Calc. for C11H14O2: C, 65.5, H,
7.8%, M, 178).
ozone was replaced by oxygen which until the solution became
colourless after which sodium borohydride (95 mg, 2.5 mmol)
was added. The reaction mixture was warmed to 25 ◦C and
then quenched with aqueous saturated ammonium chloride.
The volatile material was removed under reduced pressure
and the remaining mixture was extracted with chloroform,
dried (MgSO4) and the solvent was removed to give the diol
(161 mg) which was dissolved in acetonitrile (10 cm3). Imidazole
(102 mg, 1.5 mmol) and tert-butyldiphenylsilyl chloride (0.29 cm3,
1.1 mmol) were added and the solution was stirred for 60 min.
The mixture was concentrated under reduced pressure and
water was added to the residue. The aqueous mixture was
extracted with dichloromethane, the organic extract was dried
(MgSO4) and the solvent was removed under reduced pressure.
The residue (550 mg) was chromatographed on silica gel (50 g)
using ethyl acetate–hexane (1 : 9) as eluent to give the bissilyl
ether 34 (385 mg, 97%) as an oil, [a]D -10.4 (c 1.0 in CHCl3);
(1S,2S,7R,8S,9R)-8-Hydroxymethyl-10,12-dioxatricyclo[7.2.
1.02,7]dodec-4-ene 30. 9-Borabicyclo[3.3.1]nonane (4.93 g,
40.4 mmol) was added to a stirred solution of 29 (3.60 g,
20.2 mmol) in tetrahydrofuran (100 cm3). After 6 h of stirring
◦
at 25 C 1 M NaOH (50 cm3) was slowly added, followed by
hydrogen peroxide (30%, 50 cm3). The mixture was diluted with
water and extracted with ethyl acetate. The extract was dried
(MgSO4), and the solvent was removed under reduced pressure.
The residue (6.72 g) was chromatographed on silica gel (100 g)
using ethyl acetate–hexane (4 : 1) as eluent to yield the alcohol 30
(3.83 g, 97%), mp 80–82 ◦C (from ethyl acetate–hexane); [a]D 67.4
(c 1.0 in CHCl3); nmax(CHCl3)/cm-1 3626 (OH); dH (400 MHz,
CDCl3), 1.65–1.95 (4H, m, 2-H, 6-HA, 8-H and OH), 2.00–2.12
(1H, m, 3-HA), 2.24–2.39 (2H, m, 3-HB and 6-HB), 2.50–2.59 (1H,
m, 7-H), 3.64 (1H, dd, J 10.5 and 4.4 Hz, 1¢-HA), 3.76 (1H, t, J 2 ¥
10.5 Hz, 1¢-HB), 3.83 (1H, dd, J 7.0 and 5.3 Hz, 11-HA), 3.97 (1H,
dd, J 7.0 and 0.7 Hz, 11-HB), 4.30 (1H, dd, J 5.3 and 0.7 Hz, 1-H)
and 5.62–5.75 (3H, m, 4-H, 5-H and 9-H); dC (100 MHz, CDCl3),
24.7 (C-7), 26.2 (C-3), 27.0 (C-6), 35.1 (C-2), 48.3 (C-8), 61.3
(C-1¢), 67.8 (C-11), 77.1 (C-1), 103.2 (C-9), 125.5 and 126.2 (C-4
and C-5) (Found: C, 67.1, H, 8.3%, M+ 196. Calc. for C11H16O3:
C, 67.3, H, 8.2%, M, 196).
n
max(CHCl3)/cm-1 1719 (CO); dH (300 MHz, CDCl3), 1.03
[9H, s, TPS-C(CH3)3], 1.05 [9H, s, TPS-C(CH3)3], 1.16 [9H, s,
Pv-C(CH3)3], 1.35–1.73 (5H, m, 2-H, 1¢-H2 and 1¢¢-H2), 1.80–1.91
(1H, m, 4-H), 2.50–2.64 (1H, m, 3-H), 3.55–3.82 (6H, m, 7-H2,
2¢-H2 and 2¢¢-H2), 3.92–4.17 (2H, m, 1¢¢¢-H2), 4.28 (1H, d, J
5.0 Hz, 1-H), 5.46 (1H, d, J 1.2 Hz, 5-H), 7.30–7.46 (12H, m,
Ar-H) and 7.58–7.64 (8H, m, Ar-H); dC (75 MHz, CDCl3), 19.1
and 19.2 [2 ¥ TPS-C(CH3)3], 26.8 and 26.9 [2 ¥ TPS-C(CH3)3],
27.2 [Pv-C(CH3)3], 27.3 (C-3), 29.4 and 32.0 (C-1¢ and C-1¢¢), 37.5
(C-2), 38.7 [Pv-C(CH3)3], 42.7 (C-4), 61.0 and 62.2 (C-2¢ and C-
2¢¢), 62.4 (C-1¢¢¢), 68.1 (C-7), 75.5 (C-1), 102.1 (C-5), 127.6(7) and
127.6(9), 129.6(2) and 129.6(9), 133.6(5), 133.6(9), 133.7(2) and
=
133.7(6), 135.5(3), 135.5(6) and 135.5(8) (Ar-C) and 178.2 (C O)
(Found (FAB): M+ + Rb, 877.3369. Calc. for C48H64O6Si2Rb: M,
877.3360).
(1S,2S,7R,8S,9R)-8-Pivaloyloxymethyl-10,12-dioxatricyclo-
[7.2.1.02,7]dodec-4-ene 33. Triethylamine (5.0 cm3, 3.63 g,
35.9 mmol) and pivaloyl chloride (1.0 cm3, 0.98 g, 8.1 mmol)
were added to a solution of the alcohol 30 in dichloromethane
(20 cm3). The resulting mixture was stirred for 3 h after which it
was acidified with 1 M HCl and extracted with dichloromethane.
The organic phase was washed with brine, dried (MgSO4) and the
solvent was removed under reduced pressure. The residue (2.09
g) was chromatographed on silica gel (100 g) using ethyl acetate–
hexane (1 : 9) as eluent to give the pivaloate 33 (1.79 g, 95%) as
an oil, [a]D +52.3 (c 1.7 in CHCl3); nmax(CHCl3)/cm-1 1719 (CO);
dH (400 MHz, CDCl3), 1.10 (9H, s, C(CH3)3), 1.67–1.79 (1H, m,
2-H), 1.81–1.95 (2H, m, 6-HA and 8-H), 1.96–2.10 (1H, m, 3-HA),
2.20–2.37 (2H, m, 3-HB and 6-HB), 2.47–2.56 (1H, m, 7-H), 3.76
(1H, dd, J 7.0 and 5.2 Hz, 11-HA), 3.91 (1H, d, J 7.0 Hz, 11-HB),
3.98 (1H, t, J 3 ¥ 11.2 Hz, 1¢-HA), 4.12 (1H, dd, J 11.2 and 4.1 Hz,
1¢-HB), 4.24 (1H, d, J 5.1 Hz, 1-H), 5.44 (1H, s, 9-H) and 5.57–
5.67 (2H, m, 4-H and 5-H); dC (100 MHz, CDCl3), 24.7 (C-7), 26.0
(C-6), 26.8 (C-3), 27.1 [C(CH3)3], 34.9 (C-2), 38.7 [C(CH3)3], 45.2
(2S,3S,4R,5S)-3,4-Bis[2-(tert-butyldiphenylsilanyloxy)ethyl]-5-
(1,3-dithiolan-2-yl)-6-pivaloyloxyhexane-1,2-diol
35. Ethane-
dithiol (0.40 cm3, 4.8 mmol) followed by titanium tetrachloride
(1.0 M in dichloromethane, 2.8 cm3, 2.8 mmol) were added to
a stirred solu◦tion of 34 (2.80 g, 3.5 mmol) in dichloromethane
(60 cm3) at 0 C. The reaction was stirred at this temperature for
4 h and then quenched with aqueous saturated sodium hydrogen
carbonate, extracted with dichloromethane, dried (MgSO4) and
the solvent was removed under reduced pressure. The residue
(3.80 g) was chromatographed on silica (150 g) using ethyl
acetate–hexane (1 : 9 to 3 : 7) to give the thioacetal 37, (2.17 g,
69%), [a]D -11.6 (c 1.6 in CHCl3); nmax(CHCl3)/cm-1 1718 (CO),
3566 (OH); dH (300 MHz, CDCl3), 1.02 [9H, s, TPS-C(CH3)3],
1.03 [9H, s, TPS-C(CH3)3], 1.16 [9H, s, Pv-C(CH3)3], 1.49–1.66
(3H, m, 1¢-H2 and 1¢¢-HA), 1.73–1.90 (2H, m, 3-H and 1¢¢-HB),
1.94–2.12 (2H, m, 4-H and OH), 2.47–2.57 (1H, m, 5-H),
2.62–2.70 (1H, br d, OH), 2.98–3.27 (4H, m, 4¢¢¢-H2 and 5¢¢¢-H2),
3.45–3.75 (7H, m, 1-H2, 2-H, 2¢-H2 and 2¢¢-H2), 4.21 (1H, dd,
J 11.7 and 7.0 Hz, 6-HA), 4.36 (1H, dd, J 11.7 and 3.6 Hz,
6-HB), 4.73 (1H, d, J 3.8 Hz, 2¢¢¢-H), 7.30–7.45 (12H, m, Ar-H)
and 7.60–7.70 (8H, m, Ar-H); dC (75 MHz, CDCl3), 19.0 [2 ¥
TPS-C(CH3)3], 26.8 [2 ¥ TPS-C(CH3)3], 27.1 [Pv-C(CH3)3], 31.2
and 31.3 (C-1¢ and C-1¢¢), 37.0 (C-3), 38.0 and 39.3 (C-4¢¢¢ and
C-5¢¢¢), 38.6 (C-4), 44.4 (C-5), 54.2 (C-2¢¢¢), 62.4 and 62.9 (C-2¢ and
C-2¢¢), 65.1 (C-1), 65.2 (C-6), 73.1 (C-2), 127.5(8) and 127.6(4),
129.5(3) and 129.6(4), 133.2(9), 135.4(7) and 135.5(0) (Ar-C),
(C-8), 63.1 (C-1¢), 67.9 (C-11), 77.1 (C-1), 102.9 (C-9), 125.4 and
+
=
126.0 (C-4 and C-5) and 178.2 (C O) (Found (FAB): M + Rb,
365.0812. Calc. for C16H24O4Rb: M, 365.0793).
(1S,2S,3R,4S,5R)-2,3-Bis(2-tert-butyldiphenylsilanyloxyethyl)-
4-pivaloyloxymethyl-6,8-dioxabicyclo[3.2.1]octane
34. Ozone
was bubbled through a solution of 33 (140 mg, 0.50 mmol) in
◦
methanol (8 cm3), at -78 C until the solution turned blue. The
3534 | Org. Biomol. Chem., 2009, 7, 3527–3536
This journal is
The Royal Society of Chemistry 2009
©

View Article Online
+
=
and 178.2 (C O) (Found (FAB): M + Rb, 971.3257. Calc. for
C50H70O6S2Si2Rb: M, 971.3270).
was dissolved in N,N -dimethylformamide (10 cm3). Potassium
carbonate (390 mg, 2.82 mmol) and iodomethane (0.11 cm3,
1.83 mmol) were added and the mixture was stirred at 0 ◦C for 2 h.
The mixture was diluted with ethyl acetate (200 cm3). The organic
phase was washed twice with water, dried (MgSO4) and the solvent
removed under reduced pressure to give a residue (2.30 g) which
was purified by chromatography on silica gel (50 g) using ethyl
acetate–hexane (1 : 9) as eluent, to yield the ester 39 (1.03 g, 99%)
as a gum, [a]D -11.3 (c 1.5 in CHCl3); nmax(CHCl3)/cm-1 1740
(CO); dH (400 MHz, CDCl3), 1.04 [9H, s, C(CH3)3], 1.06 [9H, s,
C(CH3)3], 1.55 (1H, s, 2-H), 1.59–1.92 (4H, m, 1¢-H2, and 1¢¢-H2),
2.59–2.69 (2H, m, 3-H and 4-H), 3.61 (3H, s, OCH3), 3.63–3.78
(5H, m, 7-HA, 2¢-H2 and 2¢¢-H2), 3.81 (1H, d, J 7.1 Hz, 7-HB), 4.37
(1H, d, J 5.7 Hz, 1-H), 5.67–5.72 (1H, br s, 5-H), 7.32–7.45 (12H,
m, Ar-H) and 7.60–7.67 (8H, m, Ar-H); dC (100 MHz, CDCl3),
19.1 [2 ¥ C(CH3)3], 26.9 and 26.9 [2 ¥ C(CH3)3], 27.5 (C-3), 28.7
and 32.4 (C-1¢ and C-1¢¢), 37.6 (C-2), 48.6 (C-4), 51.2 (OCH3),
61.3 and 62.1 (C-2¢ and C-2¢¢), 68.4 (C-7), 75.6 (C-1), 100.5 (C-5),
127.6, 129.6, 133.8(0) and 133.8(2), 135.5(0), 135.5(3) and 135.5(6)
(Ar-C) and 170.8 (C-1¢¢¢) (Found: M+ - CH3, 721.3387. Calc. for
C43H53O6Si2: M, 721.3381).
(2S,3R,4S)-2,3-Bis[2-(tert-butyldiphenylsilanyloxy)ethyl]-4-(1,
3-dithiolan-2-yl)-5-pivaloyloxypentanal 36. Lead tetraacetate
(0.90 g, 2.03 mmol) was added to a solution of 35 (1.50 g,
1.70 mmol) in toluene (100 cm3) at 25 ◦C. After 15 min, ethylene
glycol (5 drops) was added and the mixture was stirred for 10 min,
then filtered through Celite and MgSO4 layers and concentrated
to give an oil (1.97 g). Flash chromatography on silica gel (100 g)
using ethyl acetate–hexane (1 : 9) as eluent afforded the aldehyde
36 (1.37 g, 94%), [a]D 4.4 (c 1.5 in CHCl3); nmax(CHCl3)/cm-1 1722
(CO); dH (300 MHz, CDCl3), 1.04 [18H, s, 2 ¥ TPS-C(CH3)3], 1.17
[9H, s, Pv-C(CH3)3], 1.47–1.63 (1H, m, 1¢-HA), 1.64–1.86 (2H, m,
1¢¢-H2), 1.96–2.12 (1H, m, 1¢-HB), 2.12–2.24 (1H, m, 4-H), 2.30–
2.43 (1H, m, 3-H), 2.89–2.99 (1H, m, 2-H), 3.04–3.25 (4H, m,
4¢¢¢-H2 and 5¢¢¢-H2), 3.54–3.80 (4H, m, 2¢-H2 and 2¢¢-H2), 4.27 (2H,
d, J 4.7 Hz, 5-H2), 4.73 (1H, d, J 7.4 Hz, 2¢¢¢-H), 7.28–7.45 (12H,
m, Ar-H), 7.57–7.70 (8H, m, Ar-H) and 9.66 (1H, d, J 1.9 Hz,
1-H); dC (75 MHz, CDCl3), 19.0 [2 ¥ TPS-C(CH3)3], 26.8 [2 ¥
TPS-C(CH3)3], 27.1 [Pv-C(CH3)3], 31.4 and 32.0 (C-1¢ and C-1¢¢),
38.1 and 38.8 (C-4¢¢¢ and C-5¢¢¢), 38.4 (C-4), 38.6 [Pv-C(CH3)3],
46.1 (C-3), 48.1 (C-2), 54.3 (C-2¢¢¢), 61.5 and 61.8 (C-2¢ and C-2¢¢¢),
Acknowledgements
63.9 (C-5), 127.5, 129.5, 133.5, 135.4(4) and 135.4(8) (Ar-C), 178.7
(C O) and 204.2 (C-1) (Found (FAB): M + Rb, 939.3021. Calc.
+
=
We thank the South African National Research Foundation for
financial support and Drs Gareth Arnott and Seanette Wilson for
preliminary studies.
for C49H66O5S2Si2Rb: M, 939.3008).
(1S,2S,3R,4S,5R)-2,3-Bis(2-tert-butyldiphenylsilanyloxyethyl)-
4-hydroxymethyl-6,8-dioxabicyclo[3.2.1]octane 37. Lithium alu-
minium hydride (134 mg, 3.5 mmol) was added to a solution
of 34 (2.53 g, 3.2 mmol) in tetrahydrofuran (250 cm3) at 0 ◦C.
After 10 min, the reaction was quenched with water and the
solvent removed under reduced pressure. The aqueous slurry was
diluted with 1 M sodium hydroxide, extracted with chloroform, the
organic extract was dried (MgSO4) and the solvent was removed in
vacuo to give the an oily residue (2.70 g). Chromatography on silica
gel (250 g) using ethyl acetate–hexane (2 : 3) as eluent afforded the
alcohol 37 (2.09 g, 93%) as an oil, [a]D -18.7 (c 1.0 in CHCl3);
References
1 B. M. Trost and F. D. Toste, J. Am. Chem. Soc., 1999, 121, 3543.
2 G. Stork, K. Manabe and L. Liu, J. Am. Chem. Soc., 1998, 120,
1337.
3 W. Oppolzer, Angew. Chem., Int. Ed., 1984, 23, 876.
4 A. T. Stevens, J. R. Bull and K. Chibale, Synlett, 2007, 3175–3179.
5 A. T. Stevens, J. R. Bull and K. Chibale, Org. Biomol. Chem., 2008, 6,
586–595.
6 K. Tomioka, T. Ishiguro, Y. Iitaka and K. Koga, Tetrahedron, 1984, 40,
1303.
7 P. Camps, J. Cardellach, J. Font, R. M. Ortun˜o and O. Ponsat´ı,
Tetrahedron, 1982, 38, 2395.
n
max(CHCl3)/cm-1 3450 (OH); dH (300 MHz, CDCl3), 1.04 [9H, s,
C(CH3)3], 1.06 [9H, s, C(CH3)3], 1.32–1.66 (6H, m, 2-H, 1¢-H2,
1¢¢-H2 and OH), 1.66–1.76 (1H, m, 4-H), 2.43–2.64 (1H, m, 3-H),
3.50–3.74 (7H, m, 7-HA, 2¢-H2, 2¢¢-H2 and 1¢¢¢-H2), 3.79 (1H, d, J
7.6 Hz, 7-HB), 4.27 (1H, d, J 5.2 Hz, 1-H), 5.56 (1H, d, J 1.2 Hz,
5-H), 7.31–7.46 (12H, m, Ar-H) and 7.59–7.68 (8H, m, Ar-H); dC
(75 MHz, CDCl3), 19.2 [2 ¥ C(CH3)3], 26.9 and 26.9 [2 ¥ C(CH3)3],
27.6 (C-3), 29.5 and 32.1 (C-1¢ and C-1¢¢), 37.7 (C-2), 45.8 (C-4),
60.4 (C-1¢¢¢), 61.5 and 62.3 (C-2¢ and C-2¢¢), 68.0 (C-7), 75.5 (C-1),
102.5 (C-5), 127.7, 129.7, 133.7 and 135.6 (Ar-C) (Found: M+ -
C4H9, 651.2975. Calc. for C39H47O5Si2: M, 651.2962).
8 M. G. B. Drew, J. Mann and A. Thomas, J. Chem. Soc., Perkin Trans.
1, 1986, 2279.
9 S. Takano, A. Kurotaki, M. Takahashi and K. Ogasawara, Synthesis,
1986, 403.
10 C. R. Schmid and J. D. Bryant, Org. Synth., 1995, 72, 6.
11 J. Mann and A. Thomas, J. Chem. Soc. Perkin Trans. 1, 1986, 2287.
12 R. M. Ortun˜o, J. Font and J. Corbera, Tetrahedron Lett., 1986, 27,
1081.
13 R. Batllori, J. Font, M. Monsalvatje, R. M. Ortun˜o and F. Sanchez-
Ferrando, Tetrahedron, 1989, 45, 1833.
14 Levoglucosenone and Levoglocosans: Chemistry and Applications, ed.
Z. J. Witczak, A T L Pr. Scientific Pub., Washington DC, 1996 .
15 A. V. Samet, A.M. Shestopalov, D. N. Lutov, L. A. Rodinovskaya,
A. A. Shestopalov and V. V. Semenov, Tetrahedron: Asymmetry, 2007,
18, 1986.
16 A. V. Samet, D. N. Lutov, L. D. Konyushkin, Y. A. Strelenko and V. V.
Semenov, Tetrahedron: Asymmetry, 2008, 19, 691.
17 K. Mori, M. Mori, T. Chuman and K. Kato, Tetrahedron Lett., 1982,
23, 4593.
18 L. A. Paquette, A. M. Doherty and C. M. Rayner, J. Am. Chem. Soc.,
1992, 114, 3910.
19 D. L. Boger, R. M. Borzilleri, S. Nukui and R. T. Beresis, J. Org. Chem.,
1997, 62, 4721.
20 W. Hartwig, Tetrahedron, 1983, 39, 2609.
21 S. Harusawa, R. Yoneda, Y. Omori and T. Kurihara, Tetrahedron Lett.,
1987, 28, 4189.
Methyl (1S,2S,3R,4R,5R)-2,3-bis(2-tert-butyldiphenylsilanylo-
xyethyl)-6,8-dioxabicyclo[3.2.1]octane-4-carboxylate 39. Sodium
metaperiodate (10% solution in water, 12 cm3) and ruthenium
dioxide (19 mg, 0.14 mmol) were added sequentially to a vigorously
stirred solution of 37 (1.00 g, 1.41 mmol) in carbon tetrachloride–
acetonitrile (1 : 1) (100 cm3) at 0 ◦C. Stirring was continued at
this temperature for 18 h. The resulting mixture was poured into
ethyl acetate–water and the aqueous layer was extracted with ethyl
acetate. The organic phase was dried (MgSO4) and the solvent was
removed under reduced pressure to give the acid (1.10 g) which
This journal is
The Royal Society of Chemistry 2009
Org. Biomol. Chem., 2009, 7, 3527–3536 | 3535
©

View Article Online
22 P. Kocienski, Protecting Groups, Georg Thieme Verlag, Stuttgart, 1994,
p. 104.
23 S. Nahm and S. M. Weinreb, Tetrahedron Lett., 1981, 22, 3815.
24 T. W. Greene and P. G. M. Wuts, Protective Groups in Organic Synthesis,
John Wiley & Sons, New York, 1999, p. 41. .
33 O. H. Oldenziel, D. van Leusen and A. M. van Leusen, J. Org. Chem.,
1977, 42, 3114.
34 M. North, in Comprehensive Organic Functional Group Transforma-
tions, ed. A. R. Katritzky, O. Meth-Cohn and C. W. Rees, Elsevier,
Oxford, 1995, p. 612.
25 A. J. Castellino and H. Rapoport, J. Org. Chem., 1986, 51,
35 F. W. Winkler, Liebigs. Ann. Chem., 1832, 4, 246.
36 M. North, in Comprehensive Organic Functional Group Transforma-
tions, ed. A. R. Katritzky, O. Meth-Cohn and C. W. Rees, Elsevier,
Oxford, 1995, p. 615.
37 R. Yoneda, S. Harusawa, T. Kurihara, Y. Hamada and T. Shioiri,
Tetrahedron Lett., 1984, 25, 427.
38 I. Mico´ and C. Na´jera, Tetrahedron, 1993, 49, 4327.
39 R. Yoneda, S. Harusawa and T. Kurihara, J. Org. Chem., 1991, 56,
1827.
1006.
26 E. R. Civitello and H. Rapoport, J. Org. Chem., 1994, 59, 3775.
27 J. S. Swenton, J. N. Freskos, P. Dalidowicz and M. L. Kerns, J. Org.
Chem., 1996, 61, 459.
28 D. D. Ward and F. Shafizadeh, Carbohydr. Res., 1981, 95, 155.
29 J. Seyden-Penne, Reductions by the Alumino- and Borohydrides in
Organic Synthesis, VCH Publishers, New York, 1991, p. 47.
30 O. H. Oldenziel and A. M. van Leusen, Synth. Commun., 1972, 2,
281.
40 G. A. Molander, Org. React., 1994, 46, 211.
31 J. March, Advanced Organic Chemistry, John Wiley & Sons, New York,
41 Cambridge Structural Database and Cambridge Structural Database
System, Version 5.21, Cambridge Crystallographic Data Centre, Uni-
versity Chemical Laboratory, Cambridge, April 2001.
1992, p. 887.
32 J. R. Bull and A. Tuinman, Tetrahedron, 1975, 31, 2151.
3536 | Org. Biomol. Chem., 2009, 7, 3527–3536
This journal is
The Royal Society of Chemistry 2009
©