Sequential desymmetrization-fluorination: Enantioselective synthesis of fluorinated cyclitols

Article abstract of DOI:10.1002/chem.200600861

An asymmetric synthesis of enantioenriched, highly functionalized fluorinated carbocycles has been developed based on an enantioselective Sharpless dihydroxylation of cyclohexa-dienylsilanes, combined with a diastereoselective electrophilic fluorodesilyla

Full text of DOI:10.1002/chem.200600861

DOI: 10.1002/chem.200600861
Sequential Desymmetrization–Fluorination: Enantioselective Synthesis of
Fluorinated Cyclitols
Sophie Purser,[a] Barbara Odell,[a] Timothy D. W. Claridge,[a] Peter R. Moore,[b] and
VØronique Gouverneur*[a]
Abstract: An asymmetric synthesis of
enantioenriched, highly functionalized
fluorinated carbocycles has been devel-
oped based on an enantioselective
Sharpless dihydroxylation of cyclohexa-
dienylsilanes, combined with a diaste-
reoselective electrophilic fluorodesily-
lation. Several parameters define the
level of diastereocontrol for the fluori-
nation step. These include the relative
stereochemistry of the starting endocy-
clic allylsilanes and the structural fea-
tures of the reactants. As expected for
an SE2’ mechanism, the fluorodesilyla-
tion occurred with clean transposition
of the double bond, with the attack of
the electrophilic fluorinating agent
taking place preferentially anti to the
silyl group. For the fluorination step,
the best selectivities were observed for
the monocyclic anti,syn benzyl-protect-
ed precursors and for the syn,syn start-
ing allylsilane. Full NMR spectroscopic
analysis of the fluorinated monocyclic
compounds, which possess an endocy-
clic double bond flanked by two elec-
tronegative groups (the methoxy group
and the fluorine substituent) revealed
that a subtle combination of steric (1,3-
axial/pseudoaxial interaction) and ste-
reoelectronic effects (p–s* interaction)
favours the preferential conformers
featuring the methoxy group in a pseu-
doaxial position.
Keywords: carbocycles · cyclitols ·
dihydroxylation
·
electrophilic
substitution · fluorine
Introduction
lations, thereby broadening the range of accessible fluoro-
compounds. In this paper, we wish to report the diastereo-
and enantioselective synthesis of highly functionalized dihy-
droxylated fluorinated carbocycles and a detailed study of
their structural assignment. The chosen strategy combines a
catalytic asymmetric route to enantioenriched organosilanes
followed by an electrophilic fluorodesilylation process. This
concise approach delivers enantioenriched fluorinated car-
bocycles featuring endocyclic allylic fluorides, a class of
compounds rarely explored in synthesis. Noteworthy excep-
tions are the fluorinated analogues of shikimic acids, which
are of relevance as potential antifungal, antibacterial and
antiparasitic agents.[4]
The ongoing interest for fluorine chemistry stems from the
[1]
À
unique properties of the C F bond. As a result, a lot of in-
terest has focused on this research area and more recently
on the development of asymmetric fluorination reactions.[2]
These fluorinations were mostly applied to the preparation
of enantioenriched a-fluorinated carbonyl derivatives featur-
ing one, or more rarely, two stereogenic centres.[3] Conse-
quently, a relatively restricted range of fluorinated targets is
accessible when using this chemistry. The quest for efficient
methods for the preparation of more complex fluorinated
targets is still actively ongoing. We are interested in devel-
oping novel routes to enantioenriched fluorinated building
blocks other than fluorinated carbonyl derivatives. These
compounds could offer alternative functional-group manipu-
On the basis of relevant precedent from our laboratory,[5]
we reasoned that compounds of general structure
I
(Scheme 1) could be prepared by diastereoselective electro-
philic fluorodesilylation of the corresponding enantioen-
riched allylsilanes. A known catalytic asymmetric desymmet-
rization process will enable the construction of the starting
enantioenriched dihydroxylated cyclic allylsilanes. In this
strategy, it was anticipated that the silyl group could control
the diastereoselectivity of the fluorination process. The pres-
ence of the two additional nonsilylated stereogenic carbon
atoms can influence the stereochemical outcome of the fluo-
rination and will enable us to determine easily whether the
[a] S. Purser, Dr. B. Odell, Dr. T. D. W. Claridge, Dr. V. Gouverneur
Chemistry Research laboratory, University of Oxford
12Mansfield Road, OX1 3TA Oxford (UK)
Fax : (+44)1865-275-644
[b] Dr. P. R. Moore
AstraZeneca Discovery Chemistry, Alderley Park
Macclesfield, Cheshire, SK10 4TG (UK)
Supporting information for this article is available on the WWW
under http://www.chemeurj.org/ or from the author.
9176
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 9176 – 9185

FULL PAPER
4,6-pyrimidinediyl
diether).
Under these conditions, the de-
sired dihydroxylated cyclic ally-
silanes were obtained in good
yields with complete diastereo-
control.
The
enantiotopic
Scheme 1.
double bonds were relatively
well differentiated, as products
(+)-2a and (+)-2b were
approach of the fluorinating reagent takes place preferen-
tially anti or syn to the silyl group. The cyclic allylic fluo-
rides described herein are versatile building blocks. In this
paper, we also report that their derivatisation upon dihy-
droxylation gives orthogonally protected fluorinated cycli-
tols. These compounds fall within the category of carbohy-
drate mimics, which are frequently used as powerful molecu-
lar probes for studying the biological function of oligosac-
charides present on cell membranes.[6]
formed with enantiomeric excesses of 87 and 69%, respec-
tively. The enantioselective dihydroxylation proceeded, as
expected, anti to the trimethylsilyl or the tert-butyldimethyl-
silyl groups. We subsequently protected the diols with the
objective of accessing a series of monocyclic and bicyclic al-
lylsilanes. It was anticipated that the preferentially reactive
conformations of these compounds might differ and could
give fluorinated compounds with different levels of diaste-
reocontrol upon electrophilic fluorodesilylation. Compounds
3–7 were therefore prepared by using standard procedures,
with their 1,2-diols protected as ketals or as ethers.[8]
With these compounds in hand, initially we studied the
feasibility of the fluorination process. Preliminary attempts
on the unprotected diol 2a proved unsuccessful.[9] We there-
fore examined the reactivity of the protected diols 3–7. The
fluorination of the protected trimethylsilylated precursor
(+)-3 was carried out in aceto-
Results and Discussion
Synthesis of fluorinated carbocycles: The synthesis of repre-
sentative cyclic allylsilanes was carried out following a reac-
tion sequence developed by Landais et al (Scheme 2,
nitrile at room temperature in
the presence of 1.1 equivalents
of Selectfluor and 1.2equiva-
lents of NaHCO3 (Scheme 3).
This reaction delivered the
allylic fluorides (À)-8a and
(À)-8b, resulting from a clean
transposition of the double
bond in 61% overall isolated
yield (Scheme 3). Up to three
days were required for the
starting material to be fully
consumed. Attempts to reduce
the reaction time by heating the
reaction mixture were unsuc-
cessful as decomposition was
Scheme 2.
TMEDA=tetramethylethylene-
diamine).[7] Two cyclohexadie-
nylsilanes were conveniently
prepared by metalation of cy-
clohexa-1,4-diene with sBuLi
followed by silylation of the re-
sulting carbanion with trime-
thylsilane chloride or with tert-
Scheme 3.
butyldimethylsilane
chloride.
The desymmetrization of the
silyl-2,5-cyclohexadienes 1a and 1b was carried out by using
an asymmetric Sharpless dihydroxylation in the presence of
(DHQ)2PYR ((DHQ)2PYR=hydroquinine 2,5-diphenyl-
observed under these conditions. The product was formed as
a mixture of two diastereomers (dr=69:31). These diaste-
reomers displayed different physicochemical properties and
Chem. Eur. J. 2006, 12, 9176 – 9185
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
9177

V. Gouverneur et al.
proved to be remarkably easy to separate by simple silica-
gel chromatography, enabling full characterization. The data
revealed that the major product syn,syn-(À)-8a resulted
from addition of the fluorinating reagent anti to the trime-
thylsilyl group, thereby delivering the all-syn-fluorinated car-
bocycle. We subsequently studied the influence of the silyl
group and the protecting group on the yield and stereocon-
trol of the fluorination process (Table 1).
corresponding fluorinated products (À)-10a–b and (À)-11a–
b were obtained in 52and 82% yield, respectively. The level
of diastereocontrol was higher for these transformations.
The two diastereomers (À)-10a and (À)-10b were formed
with a diastereomeric ratio of 80:20. Very similar results
were obtained for the fluorination of (+)-7, as the two prod-
ucts (À)-11a and (À)-11b were formed with a dr of 82:18.
For compounds (À)-10 and (À)-11, the diastereomers could
not be separated by silica-gel
chromatography. Careful NMR
Table 1. Electrophilic fluorination of (+)-4–7.
spectroscopic analysis of the
Entry
Starting materials
Products
dr[a]
Yield [%][b]
mixture of diastereomers re-
vealed that these two transfor-
mations delivered the major all-
syn stereoisomers syn,syn-(À)-
10a and syn,syn-(À)-11a, re-
sulting from a preferential ap-
proach of the fluorinating agent
anti with respect to the trime-
thylsilyl group.
With the pure stereoisomer
syn,syn-(À)-8a in hand, we ex-
amined the possibility of fur-
ther functionalisation of the re-
1
67:33
48
2
3
4
55:45
60
52
82
80:20[c]
82:18[c]
maining
double
bond
(Scheme 4). A dihydroxylation
was performed by using a cata-
lytic amount of osmium tetrox-
ide and stoichiometric amounts
of NMO (the “Upjohn” pro-
cess; NMO=N-methylmorpho-
1
[a] dr measured on crude H and 19F NMR spectra. [b] Isolated yields. [c] Inseparable diastereomers
The fluorodesilylation of (+)-
4, featuring
a tert-butyldime-
thylsilyl group instead of a tri-
methylsilyl group took place
with a lower chemical yield of
48%, but with similar diaste-
reocontrol (dr=67:33, entry 1,
Table 1). The use of allylsilanes
with bulkier silyl groups is
therefore not beneficial. The
effect of the diol protecting
group on both the yield and the
diastereomeric ratio was investigated next. The choice of
the ketal proved critical as the level of diastereocontrol
dropped significantly when the spiro compound (+)-5 de-
rived from cyclohexanone was reacted with Selectfluor in
acetonitrile (entry 2, Table 1). For this electrophilic fluoro-
desilylation, the two separable diastereomers syn,syn-(À)-9a
and anti,syn-(À)-9b were formed in roughly equimolar
amounts, with a diastereomeric ratio of 55:45 (entry 2,
Table 1). The monocyclic allylsilanes (+)-6 and (+)-7, fea-
turing two methoxy or two benzyloxy groups were also fluo-
rinated successfully under our standard conditions (entries 3
and 4, Table 1). Upon electrophilic fluorodesilylation, the
Scheme 4.
line-N-oxide).[10] This transformation took place with excel-
lent diastereocontrol and afforded compound (+)-12, a pro-
tected fluorinated analogue of 2-deoxy-allo-inositol, in 73%
yield and with a dr greater than 98:2. The relative stereo-
chemistry of this compound, assigned by NMR spectroscopy
and unambiguously confirmed by single crystal X-ray analy-
sis, revealed that the dihydroxylation had occurred solely
anti with respect to the three existing stereocentres of
syn,syn-(À)-8a. The fluorinated compound (+)-12 was ob-
tained with an enantiomeric excess of 86%, suggesting that
no racemisation occurred when the silylated precursor (+)-
2a was protected, fluorinated and subsequently dihydroxy-
9178
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 9176 – 9185

Sequential Desymmetrization–Fluorination
FULL PAPER
lated. As compound (+)-12
could not be separated by
chiral HPLC, its enantiomeric
excess was determined by 1H
and 19F NMR spectroscopy
after quantitative derivatisation
with (R)-Mosherꢁs acid chlo-
ride.[11] Similarly, the minor
isomer anti,syn-(À)-8b was sub-
mitted to a dihydroxylation and
this process delivered in 62%
yield the desired product (+)-
13 as
a single diastereomer
(dr=>98:2), resulting from an
approach of the osmium tetrox-
ide anti to the acetonide group.
These results suggested that the
fluorinated stereogenic carbon
atom has no influence on the
stereochemical outcome of the
dihydroxylation process.
Figure 1. Selected steady-state NOE enhancements, 1H–1H and 1H–19F coupling constants for isomers (À)-10a
(C6D6) and (À)-10b (CDCl3). The prime notation indicates the H4 of lowest chemical shift. *: values could not
be determined reliably due to the H1 and H2 protons of the minor isomer being hidden under those of the
major form.
Structural and conformational
analysis: Only major diol (+)-12 was available in the crystal-
line form and was characterized by X-ray diffraction. Com-
pounds 8–13 underwent extensive NMR spectroscopic anal-
ysis to determine both the configuration at the position of
fluorine substitution and the resulting conformations of the
fluorinated products in solution. These analyses were based
on equilibrium (steady-state) NOE measurements to enable
quantitative intramolecular NOE comparisons to be made,
demonstrating the fluorine substituent was itself syn to the
protected diol functionality. The large H4–H5 coupling of
11.7 Hz indicates H5 adopts an axial position antiperiplanar
to the axial H4 (throughout the H4 protons will be de-
syn
scribed as being syn or anti with respect to the protected
diol moiety) indicative of the half-chair conformation as
shown. In this conformation, H3 adopts the pseudo-axial po-
sition and shares a distinctively small H3–H2 coupling of
2.5 Hz due to the dihedral angle approaching 90o, with a
and on the magnitudes of J
N
G
1
stants. Due to the complexity of many H multiplet struc-
tures, the values of coupling constants derived from direct
interpretation of multiplet patterns were confirmed, and
where necessary modified, by direct spectrum simulation.
Furthermore, to assist in this process, the magnitudes of
coupling of 8.5 Hz to H4 being consistent with the pseu-
syn
doaxial-axial arrangement. The near eclipsed relationship
between H4 and the fluorine is likewise consistent with
syn
this proton exhibiting a larger fluorine coupling (16.1 Hz)
than H4’anti (6.6 Hz).
1
some heteronuclear H–19F coupling constants were first es-
The NOE data for the minor isomer (À)-10b (Figure 1)
now indicated H3 to be syn relative to the methyl ethers
and, as expected, the fluorine substitution to have occurred
anti to the protected diol. The measured coupling constants
suggested a similar chair conformation as defined for the
major isomer, but with the fluorine now adopting the
pseudo-axial position. Thus, H5 still displayed a relatively
timated from multiplet peak displacements within 1H–1H
COSY crosspeaks and subsequently verified through simula-
tion. Specifically, these J(H,F) displacements could be most
E
easily observed within the H3–H4 and H3–H2 crosspeaks due
2
to the inherently large J
A
cyclic fluorinated compounds (À)-8a–b and (À)-9a–b, the
two diastereomers in each case could be separated and the
NMR analyses were performed on pure stereoisomers. For
compounds (À)-10a–b and (À)-11a–b, NMR spectroscopic
experiments were performed on mixtures of the two insepa-
rable diastereomers in both cases.
The data for the monocyclic compounds (À)-10a–b and
(À)-11a–b could be interpreted as adopting conventional
half-chair conformations and for clarity of subsequent dis-
cussions will be described first. The major isomer (À)-10a
was analyzed in deuterobenzene, as the H4 protons were co-
incident in deuterochloroform which did, however, provide
suitable dispersion for the minor isomer (À)-10b. The NOE
results for (À)-10a (Figure 1) indicated H3 to sit syn to H5,
large coupling to H4 (9.3 Hz) which can only be accounted
syn
3
for with axial geometry, whilst the large J
(F,H4syn) value of
24.7 Hz is likewise consistent with a H4syn–fluorine axial-
pseudo-axial relationship. The H3 proton consequently dis-
plays coupling of similar magnitudes with the H4syn, H4’
anti
and H2 protons; the increase in the coupling with H2 relative
to that observed in the major isomer (5.0 versus 2.5 Hz) is
consistent with H3 now occupying a pseudo-equatorial posi-
tion.
The NOE and coupling (J) data for the major and minor
[8]
isomers (À)-11a and (À)-11b collected in CDCl3 were
very similar to those of the major and minor forms respec-
tively of (À)-10 and indicate similar substitution configura-
Chem. Eur. J. 2006, 12, 9176 – 9185
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
9179

V. Gouverneur et al.
tions and conformations in both the methyl and benzyl
ethers, and as such these demand no further discussion. For
compounds (À)-10a–b and (À)-11a–b, it appears that the
major determinant for their conformation in solution is the
preferential location of the allylic methoxy group in the
pseudo-axial position.
suggests that there were no conformational changes ob-
served between these solvents. The NOE data indicate H3
was syn, and, hence, the fluorine atom anti to the acetonide.
From consideration of the coupling data, it is striking that
H5 no longer exhibited a large diaxial coupling, whereas H3
3
and H4’ did share a large value of 9.1 Hz. In addition, J-
anti
Analysis of the NOE data for the bicyclic major isomer
(À)-8a (Figure 2) clearly indicated that the fluorine substitu-
(H3,H2) is only 2.5 Hz (similar to the major forms of 10 and
11) all of which suggested that H3 occupies the pseudo-axial
position. That the fluorine atom
must, therefore, be pseudo-
equatorial is borne out by the
fluorine coupling to the H4 pro-
tons whereby no large axial-
pseudoaxial coupling is ob-
served. Furthermore, there
exists a NOE between H4’
anti
and H6. These data are consis-
tent with (À)-8b adopting a
half-chair conformation that is
inverted with respect to that
observed for the isomers of 10
and 11, as indicated in Figure 2.
Presumably this is again, in
some part at least, attributable
to the influence of the aceto-
nide ring but may also be possi-
bly favoured by the fluorine
atom occupying a pseudo-equa-
torial position, so releasing 1,3-
Figure 2. Selected steady-state NOE enhancements, 1H–1H and 1H–19F coupling constants for isomers (À)-8a
(CDCl3) and (À)-8b (C6D6). The prime notation indicates the H4 of lowest chemical shift. For H5 and H6, the
numerous small, long-range couplings meant that a precise measurement of these could not be established.
pseudoaxial/axial strain.
The proton spectra collected
for the cyclohexylidene protect-
ed systems (À)-9a and (À)-9b
bore close similarities with
tion had taken place syn to the acetonide functionality as
for the major isomers above. However, consideration of the
coupling constant data suggests that, while the fluorine re-
mains predominantly pseudo-equatorial, slight deviation
from the idealized half-chair conformation occurs for this
those of the corresponding acetonide-protected isomers, sug-
gesting that the configuration and conformational behaviour
of these systems is the same as for the (À)-8 diastereomers.
The relative stereochemistry of the diol (+)-12 was deter-
mined unambiguously by single-crystal X-ray analysis and
showed that the dihydroxylation of the fluorinated aceto-
nide (À)-8a had occurred solely in an anti fashion with re-
spect to the existing stereocentres (Figure 3).
compound. Notably the H5–H4’ coupling has reduced to
syn
8.5 Hz (from approximately 11 Hz) and the H3–H4’ cou-
syn
pling has also reduced, although less markedly. Furthermore,
the increase in the H5–H6 coupling to 5.8 Hz (from approxi-
mately 3 Hz) suggests that these protons occupy a more
eclipsed relationship than in (À)-10a or (À)-11a, whilst the
The solid-state conformation is predominantly that of a
chair, with the fluorine atom in an axial position. Some dis-
tortion is observed at C5, however, with C4–C5–C6 occupying
increase in the F–H4 coupling to 16.3 Hz (from approxi-
a slightly more planar geometry, moving H4 and H6 fur-
anti
anti
mately 8 Hz) suggests a move toward a more antiperiplanar
relationship for these. Taken together, these data indicate a
flattening of the cyclohexene half-chair conformation, such
that the C4–C5–C6 carbon atoms sit in an almost coplanar ge-
ometry (as indicated schematically in Figure 2) and we con-
sider that this may be attributed to the additional conforma-
tional constraint imposed by the acetonide protection. The
conformational differences appear even more marked for
the minor isomer (À)-8b. This was studied in both CDCl3
and C6D6, the latter removing the troublesome coincidence
of H5 and H6 observed in CDCl3, although the NMR data
ther apart (Figure 3b) presumably due to the additional con-
formational constraint imposed by the acetonide. The NMR
data (Figure 4) are largely consistent with a distorted chair
conformation in acetone but are suggestive of greater flat-
tening of the chair conformation around C5, as indicated
schematically in Figure 4. Thus H6 shares only a 4.2Hz cou-
pling with H1 and displays no 1,3-diaxial NOE to H4, sug-
gesting this is somewhat removed from the axial position of
a cyclohexane chair. In contrast, the fluorine–H4 couplings
are consistent with the fluorine atom holding the axial posi-
tion and thus it appears that the presence of the acetonide
9180
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 9176 – 9185

Sequential Desymmetrization–Fluorination
FULL PAPER
lecular interactions may impart differing conformational
constraints in the crystal.
The NOE data of the minor isomer (+)-13 showed that
dihydroxylation of the minor fluorinated acetonide (À)-8b
had occurred solely anti to the acetonide and syn with re-
spect to the fluorine atom. Moreover, there existed a signifi-
cant H1–H3 NOE, suggesting the presence of the chair con-
formation with H3 axial, also consistent with the sizeable
3J(H3,H4 ) of 11.3 Hz. The H1–H6 coupling of 7.8 Hz again
anti
suggests some distortion from an idealized chair with some
flattening imposed by the acetonide, but it appears less pro-
nounced than for (+)-12 above. This suggests that the pres-
ence of the axial fluorine atom in (+)-12 may contribute to
greater distortion of the chair conformation in solution due
to unfavourable 1,3-diaxial interactions.
Stereochemistry: Allylsilanes are known to react in the pres-
ence of electrophiles according to an SE2’ mechanism and
are highly selective in the anti sense with respect to the silyl
group.[12] Although the mechanism of electrophilic fluorina-
tion of various nucleophiles with Selectfluor has been the
subject of many debates,[13] it is helpful in understanding the
site and stereochemical outcome of the fluorinations of
compounds (+)-3–7 to postulate the formation of carbocat-
ionic intermediates (Scheme 5).
Figure 3. X-ray structure of (+)-12. The conformation is predominantly
that of a chair (view (a)) with slight distortion bringing C5 down toward
C4–C6 (view (b); see text for discussions).
With allylsilanes, the b cations
are highly stabilized through
À
overlap with the Si C bonding
orbital, thereby defining the
site of attack of the electro-
phile. Accordingly, the fluorina-
tions of (+)-3–7 all proceed
with complete regiocontrol,
leading to clean double-bond
transposition upon desilylation.
With an SE2’ mechanism oper-
ating, the relatively low reactiv-
ity of (+)-3–7 in comparison
with other cyclic and acyclic al-
lylsilanes[5] is due to the imper-
À
fect alignment of the C Si
bond with the p system, result-
ing in late stabilization of the
developing positive charge to-
wards the carbocationic inter-
mediate. In terms of stereo-
chemical outcome, the best dia-
stereoselectivities were ob-
Figure 4. Selected steady-state NOE enhancements, 1H–1H and 1H–19F coupling constants for isomers (+)-12
((CD3)2CO) and (+)-13 ((CD3)2CO/C6D6). The prime notation indicates the H4 of lowest chemical shift.
results in a net flattening of the chair around C4–C5–C6 as
suggested for the acetonide-protected cyclohexene (À)-8a.
It appears there may be greater conformational distortion in
tained for the fluorination of allylsilanes (+)-6 and (+)-7,
which feature two methoxy or benzyloxy groups. The major
isomers (À)-10a and (À)-11a obtained upon fluorodesilyla-
tion result from an approach of the fluorinating agent anti
to the trimethylsilyl group. For cyclohexene (+)-6 in its half-
chair conformation (A) with the silyl group on the pseudo-
equatorial position, this preferential stereochemical out-
come is the result of an axial addition of Selectfluor taking
place anti to the silyl group. This approach is hampered by
solution than in the solid state, but we note that in the crys-
1
À
tal structure, the hydroxyl group adjacent to the ketal (C
6
À
OH) forms a hydrogen bond to a ketal oxygen atom (C O)
in a neighbouring molecule, while the other hydroxyl group
(C OH) forms a hydrogen bond to a hydroxyl oxygen
atom (C OH) in a third molecule and that these intermo-
2
À
1
À
Chem. Eur. J. 2006, 12, 9176 – 9185
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
9181

V. Gouverneur et al.
Scheme 6.
an endo-selective Diels–Alder cycloaddition from 1-trime-
thylsilylbutadiene, a procedure developed by Fleming.[16]
Upon fluorination of the monocyclic all-syn precursor (Æ)-
14, the desired product (Æ)-15 was isolated in 36% yield
and with a dr superior to 98:2. The major diastereomer ob-
served for this fluorination resulted from an anti axial ap-
proach of the fluorinating reagent. This approach is clearly
favoured as it is free from unfavourable steric interactions
and this is reflected in the very high level of diastereocon-
trol observed for this electrophilic fluorination (Scheme 7).
Scheme 5.
an 1,3-diaxial interaction with the methoxy group, leading to
erosion of anti stereospecificity. If the ring adopts the alter-
native inverted conformation, which features a pseudo-axial
silyl group (B), with the reagent approaching the double
bond axially, the attack is syn to the large silyl group with
concomitant unfavourable steric interactions. Not surprising-
ly, this approach leads to the minor diastereomer. We did
not consider the less favourable equatorial attacks for the
electrophile as these approaches would lead to twisted inter-
mediates presumably of higher energy.[14] The lower level of
diastereocontrol observed for the fluorinations of com-
pounds (+)-3–5 can be rationalized by the additional steric
interactions brought about by the ketal group when the ap-
proach of Selectfluor is takes place anti to the silyl group.
For the major isomer (À)-10a with the electronegative
fluoro group and methoxy substituent on the allylic position,
the structural assignment in solution revealed that the pref-
erential conformer features the methoxy group pseudo-axial
and the fluorine atom pseudo-equatorial. This preferential
conformer enabled the nonallylic methoxy group to adopt
the more favourable equatorial position. For the minor
isomer (À)-10b, these two substituents are pseudoaxial, ena-
bling favourable interactions with the p-system (p–s*).[15]
Noteworthy is that the level of diastereocontrol observed
for the fluorodesilylations of (+)-3–7 is lower than for elec-
trophilic substitutions, other than fluorinations, reported in
the literature on structurally related cyclic all-syn allylsi-
lanes.[16] For example, the reaction of phenylsulfenyl chlo-
ride on the all-syn diester (Æ)-14 gave preferentially the de-
sired sulfide resulting from an anti addition of the electro-
phile with respect to the silyl group (dr=9:1) (Scheme 6).
To compare these results with our data, we carried out
the electrophilic fluorodesilylation of allylsilane (Æ)-14,
which was prepared as a racemate by using, as the key step,
Scheme 7.
Conclusion
We have studied the electrophilic fluorodesilylation of vari-
ous endocyclic allylsilanes. These reactions afforded fluori-
nated carbocycles featuring allylic fluorides, which can be
easily further functionalized. The fluorinations are regiose-
lective and gave products resulting from clean transposition
of the double bond. The data revealed that both anti,syn
and all-syn allylsilanes are suitable starting materials and de-
livered the corresponding fluorinated carbocycles in good
yields with the level of diastereocontrol dependent on the
relative stereochemistry and the geometrical constraints im-
posed by the starting allylsilanes. The best diastereoselectiv-
ity was observed for the fluorination of the all-syn allylsi-
lane. The approach of the fluorinating reagent takes place
preferentially anti to the silyl group as expected for these bi-
molecular electrophilic substitutions. We have also demon-
strated that the dihydroxylation of endocyclic allylic fluo-
rides is feasible and leads to novel highly functionalised, par-
tially protected fluorinated analogues of 2-deoxy-allo-inosi-
tol. These fluorocompounds, which feature five stereocen-
ters are accessible by using an operationally simple five-step
synthesis from 1,4-cyclohexadiene combining
Sharpless dihydroxylation with an electrophilic fluorodesily-
lation as the two key steps. The fluorinations presented
a known
9182
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 9176 – 9185

Sequential Desymmetrization–Fluorination
FULL PAPER
brine, dried over MgSO4 and the solvent removed in vacuo. Column
chromatography (hexane/Et2O 90:10, Rf =0.3) furnished (+)-5 as a col-
ourless oil (0.181 g, 0.8 mmol, 80% yield). [a]2D98 =+85.1 (c=0.255 in
CHCl3); 1H NMR (400 MHz, CDCl3): d=5.71 (m, 1H), 5.60 (dt, J=4.4,
10.0 Hz, 1H), 3.85 (m, H), 4.32 (m, 1H), 2.23 (m, 1H), 2.23–2.17 (m,
2H), 1.91 (m, 1H), 1.67–1.55 (m, 8H), 1.42–1.35 (m, 2H), 0.06 ppm (s,
9H); 13C NMR (100 MHz, CDCl3): d=126.1, 121.0, 108.2, 73.4, 72.0, 37.3,
herein are the first examples of electrophilic fluorodesilyla-
tion of endocyclic allylsilanes. Combined with its robustness,
this new synthetic route to fluorinated carbocycles comple-
ments other approaches that typically involve nucleophilic
fluorination rather than electrophilic fluorination.[17] The
products are versatile fluorinated building blocks offering
multiple possibilities for further functional-group manipula-
tions of the double bond. A detailed study of the synthetic
scope of these new building blocks and related compounds
is ongoing in our laboratory.
35.0, 31.5, 28.8, 25.3, 24.1, 23.8, À2.5 ppm; IR (neat): n˜ =2936, 1675 cmÀ1
;
HRMS (CI): m/z: calcd for C15H26O2Si: 267.1780 [M+H]+; found:
267.1788.
[(1S,5S,6S)-5,6-Dimethoxy-2-en-1-yl](trimethyl)silane ((+)-6): Diol (+)-
2a (0.153 g, 0.82mmol) in THF (1 mL) was added dropwise to a stirring
suspension of NaH (0.1 g, 2.5 mmol) in THF (2 mL) at 08C under argon
and the mixture was stirred at 08C for 30 min. MeI (0.13 mL, 2mmol)
was added and the reaction mixture was stirred for a further 2h at room
temperature. The reaction was quenched with H2O at 08C and extracted
with Et2O (3”20 mL). The combined organics were washed with brine
(40 mL), dried over MgSO4 and the solvent removed in vacuo. Column
chromatography (hexane/Et2O 80:20, Rf =0.25) furnished (+)-6 as a col-
ourless oil (0.1 g, 0.53 mmol, 66% yield). [a]2D98 =+88.2( c=0.2in
CHCl3); 1H NMR (400 MHz, CDCl3): d=5.57–5.51 (m, 2H), 3.64 (dd,
J=2.0, 3.2 Hz, 1H), 3.43 (s, 3H), 3.42 (s, 3H), 2.36–2.26 (m, 2H), 2.05–
2.03 (m, 1H), 0.08 ppm (s, 9H); 13C NMR (100 MHz, MeOD): d=125.4,
121.3, 77.8, 76.8, 55.9, 55.6, 32.5, 27.2, À3.2ppm; IR (neat): n˜ =2950,
1639 cmÀ1; HRMS (ESI): m/z: calcd for C11H22O2Si: 237.1283 [M+Na]+;
found: 237.1281.
Experimental Section
General information: 1H NMR spectra were recorded in deuterated sol-
vents by using Bruker DPX200, DPX400, AV400 and AV500 spectrome-
ters, calibrated by using residual protonated solvent as an internal refer-
ence. 13C NMR spectra were recorded in deuterated solvents by using
Bruker DPX200, DPX400, AV400 and AV500 spectrometers. 19F spectra
were recorded on a Bruker AV400 spectrometer. NOE difference experi-
ments were performed at 500 MHz on nondegassed solutions with satura-
tion times totalling 5 s. These were performed with frequency cycling be-
tween the individual lines within each multiplet to ensure more even sup-
pression of the wide multiplet structures. NMR spectra were processed in
ACD/SpecManager and IUPAC names were obtained by using the ACD/
I-lab service. Proton spectrum simulations were performed with the
Mestre-C NMR software. Chemical shifts (d) are quoted in parts per mil-
lion (ppm) and coupling constants (J) are measured in hertz (Hz). The
following abbreviations are used to describe multiplicities s=singlet, d=
doublet, t=triplet, q=quartet, br=broad, m=multiplet. Mass spectra
were recorded on Micromass GCT (CI), Autospec-oaTof instruments.
Optical rotations were determined on a Perkin–Elmer 241 polarimeter in
a 1 dm cell. [a]D values are given in 10À1 degcm2 gÀ1. IR spectra were re-
corded as thin films on NaCl plates in CHCl3 on a Bruker Tensor 27
FTIR spectrometer. Absorptions are measured in wavenumbers and only
peaks of interest are reported. All reactions requiring anhydrous condi-
tions were conducted in dried apparatus under an inert atmosphere of
argon or nitrogen. Solvents were dried and purified before use according
to standard procedures. All reactions were monitored by TLC using
Merck Kiesegel 60 F254 plates. Visualisation of the reaction components
was achieved by using UV fluorescence (254 nm) and KMnO4 stain.
Column chromatography was carried out over Merck silica gel C60 (40–
60 mm).
A
(+)-2a (0.327 g, 1.75 mmol) in THF (2 mL) was added dropwise to a sus-
pension of NaH (0.2g, 5.25 mmol) at 0 8C under argon, and the reaction
mixture was stirred at 08C for 30 min. BnBr (0.52mL, 4.2mmol) was
added and the reaction was stirred for 3 h at room temperature. The re-
action was quenched with H2O at 08C and extracted with Et2O (3”
20 mL). The combined organics were washed with brine (50 mL), dried
over MgSO4 and the solvent removed in vacuo. Column chromatography
(hexane/Et2O 95:5, Rf =0.25) furnished (+)-7 as a colourless oil (0.4 g,
1.11 mmol, 63% yield). [a]2D98 =+62.1 (c=0.69 in CHCl3); 1H NMR
(400 MHz, CDCl3): d=7.41–7.27 (m, 10H), 5.57 (ddd, J=2.0, 4.4, 10 Hz,
1H), 5.51 (ddd, J=5.2, 5.5, 10 Hz, 1H), 4.76 (d, J=12.4 Hz, 1H), 4.70 (d,
J=12.4 Hz, 1H), 4.58 (d, J=12.8 Hz, 1H), 4.55 (d, J=12.8 Hz, 1H), 3.88
(m, 1H), 3.59 (ddd, J=1.6, 5.5, 9.6 Hz, 1H), 2.50 (dddd, J=2.0, 5.2, 9.6,
16.6 Hz, 1H), 2.30 (ddd, J=5.2, 5.5, 16.6 Hz, 1H), 2.01 (m, 1H),
À0.09 ppm (s, 9H); 13C NMR (100 MHz, CDCl3): d=139.2, 138.8, 128.3–
121.6, 76.7, 74.9, 71.2, 70.2, 34.6, 27.6, À2.4 ppm; IR (neat): n˜ =3026,
2952, 1642 cmÀ1
; HRMS (ESI): m/z: calcd for C23H30O2Si: 389.1906
[M+Na]+; found: 389.1907.
A
General procedure for the electrophilic fluorodesilylation of allylsilanes:
A solution of the allylsilane (1 equiv) and NaHCO3 (1.2equiv) in anhy-
drous CH3CN (c=0.1m) was treated with Selectfluor (1.1 equiv) and the
mixture was stirred at room temperature for 3 d. Water was added and
the mixture extracted three times with Et2O. The combined organic
layers were washed with brine, dried over MgSO4 and the solvent re-
moved in vacuo.
silane ((+)-3): Diol (+)-2a (0.325 g, 1.75 mmol) was dissolved in acetone
(5 mL) and DMP (DMP=dimethoxypropane, 4 mL). A catalytic amount
of pTsOH (0.011 g, 0.0525 mmol) was added and the reaction was stirred
at room temperature for 2h. The solvents were evaporated in vacuo and
a saturated solution of aq NaCO3 was added. The aqueous layer was ex-
tracted with Et2O (3”30 mL) and the combined organic layers washed
with brine, dried over MgSO4 and the solvent removed in vacuo. Column
chromatography (hexane/Et2O 75:25, Rf =0.5) furnished (+)-3 as a pale
yellow oil (0.337 g, 1.49 mmol, 85% yield). [a]2D98 =+1.1 (c=0.15 in
CHCl3); 1H NMR (400 MHz, CDCl3): d=5.73 (dt, J=5.6, 10 Hz, 1H),
5.62(dtd, J=0.8, 5.0, 10.0 Hz, 1H), 4.35 (m, 2H), 2.23 (m, 2H), 1.90 (m,
1H), 1.44 (s, 3H), 1.36 (s, 3H), 0.06 ppm (s, 9H); 13C NMR (100 MHz,
CDCl3): d=126.2, 120.8, 107.6, 73.8, 72.5, 31.6, 28.3, 27.5, 25.5, À2.5 ppm;
IR (neat): n˜ =3031, 1684 cmÀ1; HRMS (CI): m/z: calcd for C12H22O2Si:
226.1389 [M]+; found: 226.1387.
A
(À)-8b and (3S,5R,7R)-5-fluoro-2,2-dimethyl-3, 4,5,7-tetrahydro-1,3-ben-
zodioxole (À)-8a: A mixture of diastereomers in a ratio 1:2.3 was ob-
tained. Column chromatography (hexane/Et2O 90:10) furnished both the
minor compound (À)-8b and the major compound (À)-8a with an over-
all 64% yield.
Minor anti isomer (À)-8b: Yield: 0.124 g, 0.72 mmol; Rf =0.38 (hexane/
Et2O 90:10); [a]2D98 =À11.25 (c=0.24 in CHCl3); 1H NMR (400 MHz,
CDCl3): d=6.02(ddddd, J=1.0, 1.3, 2.3, 10.4, 11.5, Hz, 1H), 5.80
(ddddd, J=1.5, 1.5, 1.6, 3.2, 10.4 Hz, 1H), 5.25 (dddddd, J=1.6, 2.0, 2.3,
5.6, 9.1, 49.4 Hz, 1H), 4.51 (m, 2H), 2.60 (ddddd, J=1.5, 4.4, 5.6, 5.6,
13.6 Hz, 1H), 1.91 (dddd, J=2.9, 9.1, 13.6, 14.8 Hz, 1H), 1.37 ppm (s,
6H); 13C NMR (100 MHz, CDCl3): d=129.6 (d, 21.6 Hz), 128.8 (d, J=
9.6 Hz), 108.8, 84.9 (d, J=162.2 Hz), 72.2 (d, J=11.2Hz), 71.1 (d, J=
2.8 Hz), 32.3 (d, J=19.2 Hz), 27.7, 26.3 ppm; 19F {1H} NMR (376 MHz,
Trimethyl-(3S,4S,7S)-tetrahydrospiro[1,3-benzodioxole-2,1’-cyclohexan]-
4-yl)silane ((+)-5): Diol (+)-2a (0.186 g, 1 mmol) was dissolved in cyclo-
hexanone (3 mL) and 1,2-dimethoxyketal (3 mL). A catalytic amount of
pTsOH (0.006 g, 0.03 mmol) was added and the reaction was stirred at
room temperature for 2h. The solvents were removed in vacuo and satu-
rated aq Na2CO3 (10 mL) was added. The aqueous layer was extracted
with Et2O (3”20 mL) and the combined organic layers were washed with
Chem. Eur. J. 2006, 12, 9176 – 9185
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
9183

V. Gouverneur et al.
CDCl3): d=À179.5 ppm; IR (neat): n˜ =2987, 1648 cmÀ1; HRMS (EI):
m/z: calcd for C9H13O2F: 173.0978 [M+H]+; found: 173.0975.
syn and anti), 6.04–5.94 (m, 2H; syn and anti), 5.23 (ddddd, J=0.5, 5.0,
5.0, 6.0, 48.5 Hz, 1H; anti), 5.02(ddddd, J=1.3, 2.5, 6.4, 8.4, 48.2 Hz, 1H;
syn), 4.76 (d, J=12.4 Hz, 1H; syn), 4.75 (d, J=12.0 Hz, 1H; anti), 4.71
(d, J=12.4 Hz, 1H; syn), 4.67–4.66 (m, 2H; anti), 4.68 (d, J=12.4 Hz,
1H; syn), 4.61 (d, J=12.4 Hz, 1H; syn), 4.09 (t, J=3.5 Hz, 1H; anti), 4.03
(dd, J=1.0, 3.0 Hz, 1H; syn), 3.99 (dddd, J=1.0, 2.7, 3.5, 9.4 Hz, 1H;
anti), 3.53 (dddd, J=1.3, 3.0, 3.0, 11.0 Hz, 1H; syn), 2.48 (dddd, J=5.0,
9.4, 14.0, 25.3 Hz, 1H; anti), 2.38 (dddd, J=8.4, 11.0, 11.7, 16.0 Hz, 1H;
syn), 2.27 (ddddd, J=1.0, 3.0, 6.4, 7.5, 11.7 Hz, 1H; syn), 2.02 ppm (dddd,
J=2.7, 6.0, 14.0, 18.2 Hz, 1H; anti); 13C NMR (125 MHz, CDCl3): d=
138.7 (syn), 138.5 (anti), 138.4 (anti), 138.3 (syn), 130.9 (d, J=9.9 Hz;
anti), 130.1 (d, J=20.8 Hz; syn), 128.9 (d, J=10.0 Hz; syn), 128.4 (syn),
128.3 (d, J=1.4 Hz; anti), 128.33 (anti), 128.0 (syn), 127.9 (syn), 127.88
(anti), 127.81 (syn), 127.7 (anti), 127.6 (anti), 127.6 (anti), 127.6 (syn),
127.6 (anti), 127.6 (syn), 86.6 (d, J=160 Hz; anti), 86.6 ppm (d, J=
166 Hz; syn); 19F {1H} NMR (376 MHz, CDCl3): d=À171.8 (syn),
À174.0 ppm (anti); IR (neat): n˜ =2984, 1454, 1027 cmÀ1; HRMS (ESI): m/
z: calcd for C20H21FO2: 335.1416 [M+Na]+; found: 335.1418.
Major syn isomer (À)-8a: Yield: 0.232 g, 1.35 mmol; Rf =0.25 (hexane/
Et2O 90:10); [a]2D98 =À14.3 (c=0.175 in CHCl3); 1H NMR (400 MHz,
CDCl3): d=6.07 (dddd, J=1.4, 2.0, 4.5, 10.2 Hz, 1H), 6.00 (dddd, J=0.5,
2.5, 9.8, 10.2 Hz, 1H), 5.03 (dddddd, J=1.4, 1.5, 2.5, 4.8, 7.5, 48.2 Hz,
1H), 4.44 (dddd, J=1.0, 1.5, 2.0, 5.8 Hz, 1H), 4.32 (ddd, J=4.5, 5.8,
8.5 Hz, 1H), 2.25 (dddd, J=4.5, 4.8, 13.2, 16.3 Hz, 1H), 2.05 (dddddd, J=
0.5, 1.0, 7.5, 8.5, 13.2, 13.8 Hz, 1H), 1.50 (s, 3H), 1.39 ppm (s, 3H);
13C NMR (125 MHz, CDCl3): d=130.4 (d, J=21.6 Hz), 127.6 (d, J=
9.6 Hz), 84.9 (d, J=166.2Hz), 71.1 (d, J=9.6 Hz), 70.7 (d, J=2.0 Hz),
32.0 (d, J=18.8 Hz), 28.2, 26.2 ppm; 19F {1H} NMR (376 MHz, CDCl3):
d=À176.5 ppm; IR (neat): n˜ =2987, 1648 cmÀ1; HRMS (CI): m/z: calcd
for C9H13O2F: 173.0978 [M+H]+; found: 173.0970.
A
hexane] ((À)-9b) and (3S,5R,7R)-5-fluoro-3,4,5,7-tetrahydrospriro[1,3-
benzodiozole-2,1’-cyclohexane] ((À)-9a): A mixture of diastereomers in
a ratio of 1:1.3 was obtained. Column chromatography (hexane/Et2O
95:5) furnished the minor compound (À)-9b and the major compound
(À)-9a with an overall yield of 60%.
Minor anti isomer (À)-9b: Yield: 0.047 g, 0.22 mmol; Rf =0.14 (hexane/
Et2O 95:5); [a]2D98 =À15.8 (c=2.3 in CHCl3); 1H NMR (400 MHz,
CDCl3): d=6.02–5.96 (m, 1H), 5.83–5.79 (m, 1H), 5.28–5.12 (m, 1H),
4.52–4.47 (m, 2H), 2.64–2.57 (m, 1H), 1.90 (m, 1H), 1.60–1.52 (m, 8H),
1.41–1.34 ppm (m, 2H); 13C NMR (100 MHz, CDCl3): d=129.5 (d, J=
21.5 Hz), 129.1 (d, J=9.5 Hz), 109.4, 85.1 (d, J=161.6 Hz), 71.9, 70.7,
37.4, 35.7, 32.4, 25.0, 24.0 ppm; 19F {1H} NMR (376 MHz, CDCl3): d=
À179.5 ppm; IR (neat): n˜ =2934, 1655 cmÀ1; HRMS (CI): m/z: calcd for
C12H17O2F: 213.1291 [M+H]+; found: 213.1288.
(3R,4S,5R,6R,7S)-6-Fluoro-2,2-dimethylhexahydro-1,3-benzodioxole-4,5-
diol ((+)-12):[18] The fluorinated acetonide (À)-8a (0.23 g, 1.3 mmol) was
added to a stirring solution of OsO4 (0.02g, 0.065 mmol), NMO (0.45 g,
3.9 mmol), acetone (55 mL) and water (55 mL), and the reaction was stir-
red at room temperature for 3 d. NaSO3 (1 g) was added and the reaction
was stirred for 10 min, before the acetone was removed in vacuo. The re-
maining mixture was extracted with EtOAc (3”30 mL). The combined
organics were washed with brine (50 mL), dried over MgSO4 and the sol-
vent was removed in vacuo. Column chromatography (hexane/EtOAc
50:50, Rf =0.2) furnished (+)-12 as a white crystalline solid (0.195 g,
0.95 mmol, 73% yield). 86% ee (measured by using Mosherꢁs ester analy-
sis).[11]; m.p. 100–1028C; [a]2D98 =+0.13 (c=1.49 in MeOH); 1H NMR
(400 MHz, C3D6O): d=4.70 (dddd, J=5.4, 5.4, 6.2, 50.4 Hz, 1H), 4.38 (q,
J=5.4 Hz, 1H), 4.33 (d, J=6.0 Hz, 1H), 4.21 (d, J=4.7 Hz, 1H), 4.16
(dd, J=4.2, 5.4 Hz, 1H), 4.01 (ddd, J=2.8, 6.2, 14.9 Hz, 1H), 3.92 (ddd,
J=2.8, 4.2, 14.9 Hz, 1H), 2.34 (dddd, J=5.4, 5.4, 14.9, 25.9 Hz, 1H), 2.01
(dddd, J=5.4, 5.4, 14.9, 18.7 Hz, 1H), 1.40 (s, 3H), 1.29 ppm (s, 3H);
13C NMR (125 MHz, C3D6O): d=108.9, 92.1 (d, J=172.6 Hz), 78.6, 72.8
(J=5.3 Hz), 72.0 (d, J=21.9 Hz), 71.7 (d, J=8.1 Hz), 32.1 (d, J=
20.0 Hz), 28.2, 25.5 ppm; 19F {1H} NMR (376 MHz, CDCl3): d=
À188.63 ppm; IR (neat): n˜ =3423, 3012, 1639 cmÀ1; HRMS (ESI): m/z:
calcd for C9H15FO4: 229.0847 [M+Na]+; found: 229.0845.
Major syn isomer (À)-9a: Yield: 0.036 g, 0.17 mmol; Rf =0.24; [a]D298
=
À18.6 (c=0.7 in CHCl3); 1H NMR (400 MHz, CDCl3): d=6.06 (td, J=
2.5, 10.0 Hz, 1H), 5.98–5.94 (m, 1H), 5.09–4.94 (m, 1H), 4.48–4.44 (m,
1H), 4.31 (ddd, J=4.7, 5.4, 8.4 Hz, 1H), 2.31–2.22, (m, 1H), 2.04 (tt, J=
8.4, 13.5 Hz, 1H), 171–1.55 (m, 8H), 1.44–1.35 ppm (m, 2H); 13C NMR
(100 MHz, CDCl3): d=130.7, 127.5, 110.8, 85.2 (d, J=133 Hz), 70.7, 70 2,
37.9, 35.6, 32.3, 25.0, 24.0 ppm; 19F {1H} NMR (376 MHz, CDCl3): d=
À177.12ppm; IR (neat): n˜ =2936, 1654 cmÀ1; HRMS (CI): m/z: calcd for
C12H17O2F: 213.1291 [M+H]+; found: 213.1289.
A
((À)-10a)
and
(3R,4S,6S)-6-fluoro-3,4-dimethoxycyclohexene ((À)-10b): An inseparable
mixture of diastereoisomers in a ratio of 1:4.2 anti/syn was obtained.
Column chromatography (hexane/Et2O 75:25, Rf =0.34) afforded a mix-
ture of (À)-10a and (À)-10b as a colourless oil (0.058 g, 0.36 mmol, 52%
yield). [a]2D98 =À76.9 (c=0.36 in CHCl3); 1H NMR (400 MHz, CDCl3):
d=6.08–6.04 (m, 1H; syn and anti), 6.20–5.97 (m, 1H; syn and anti), 5.19
(dddd, J=1.0, 5.0, 6.0, 48.4 Hz, 1H; anti), 5.03 (dddd, J=1.5, 6.4, 8.5,
48.2Hz, 1H; syn), 3.90 (dddd, J=1.0, 1.1, 3.3, 4.6 Hz, 1H; anti), 3.85 (dq,
J=1.0, 3.5 Hz, 1H; syn), 3.89 (dddd, J=1.0, 2.8, 3.5, 9.3 Hz, 1H; anti),
3.50 (s, 3H; syn), 4.37 (s, 3H; anti), 3.46 (s, 2H; anti), 3.44 (s, 3H; syn),
3.42(dddd, J=1.5, 3.3, 3.3, 11.7 Hz, 1H; syn), 2.38 (dddd, J=5.0, 9.3,
14.0. 24.7 Hz, 1H; anti), 2.26 (dddd, 8.5, 11.7, 11.7, 16.1 Hz, 1H; syn),
2.17 (ddddd, J=1.0, 6.4, 6.6, 11.7, 11.7 Hz, 1H; syn), 1.97 ppm (dddd, J=
2.8, 6.0, 14.0, 18.0 Hz, 1H; anti); 13C NMR (100 MHz, CDCl3): d=130.3
(d, J=9.9 Hz; anti), 130.2(d, J=20.0 Hz; syn), 128.5 (d, J=9.9 Hz; anti),
128.2 (d, J=17.9 Hz; syn), 86.3 (d, J=166.3 Hz; syn), 86.3 (d, J=160.7;
anti), 75.2(d, J=14.4 Hz; anti), 74.9 (d, J=9.1 Hz; syn), 73.7 (d, J=
2.4 Hz; anti), 72.9 (d, J=1.6 Hz; syn), 57.5 (syn), 57.4 (anti), 56.9 (anti),
56.4 (syn), 30.2(d, J=19.9 Hz; anti), 29.4 ppm (d, J=18.3; syn); 19F
{1H} NMR (376 MHz, CDCl3): d=À171.3 (syn), À174.2( anti); IR (neat):
n˜ =2092, 1642 cmÀ1; HRMS (ESI): m/z: calcd for C8H13FO2: 183.0790
[M+Na]+; found: 183.0792.
(3R,4S,5R,6S,7S)-6-Fluoro-2,2-dimethylhexahydro-1,3-benzodioxole-4,5-
diol ((+)-13): The fluorinated acetonide (À)-8b (0.4 g, 2.37 mmol) was
added to a stirring solution of NMO (0.82g, 7.11 mmol), OsO 4 (0.038 g,
0.118 mmol), acetone (90 mL) and water (30 mL), and the reaction was
stirred at room temperature for 7 d. NaSO3 (1 g) was added and the reac-
tion was stirred for 10 min, before the acetone was removed in vacuo.
The remaining mixture was extracted with EtOAc (3”100 mL). The com-
bined organics were washed with brine (100 mL), dried over MgSO4 and
the solvent was removed in vacuo. Column chromatography (hexane/
EtOAc 50:50, Rf =0.1) furnished (+)-13 (0.3 g, 1.47 mmol, 62% yield).
M.p. 105–1068C; [a]2D98 =+26.5 (c=0.2in CHCl 3); 1H NMR (500 MHz,
CDCl3): d=4.84 (dddd, J=2.0, 5.4, 11.3, 47.64 Hz, 1H), 4.38 (dddd, J=
2.3, 4.6, 4.7, 5.6 Hz, 1H), 4.27 (dddd, J=1.0, 2.0. 2.8, 11.8 Hz, 1H), 4.13
(dd, J=5.6, 7.8 Hz, 1H), 3.65 (ddd, J=2.0, 2.8, 7.8 Hz, 1H), 3.03 (s, 1H),
2.79 (s, 1H), 2.37 (dddd, J=4.6, 11.3, 11.5, 13.8 Hz, 1H), 2.30 (ddddd, J=
1.0, 2.0, 4.7, 5.4, 13.8 Hz, 1H), 1.47 (s, 3H), 1.36 ppm (s, 2H); 13C NMR
(125 MHz, CDCl3): d=109.1, 85.0 (d, J=175.5 Hz), 78.5, 72.7 (d, J=
15.3 Hz), 72.5 (d, J=10.0 Hz), 71.6 (d, J=17.6 Hz), 28.4, 26.7 (d, J=
22.4 Hz), 26.0 ppm; 19F {1H} NMR (235 MHz, CDCl3): d=À194.5 ppm;
IR (CHCl3): n˜ =3424, 1640, 1215 cmÀ1; HRMS (ESI): m/z: calcd for
C9H15O4F: 229.0847 [M+Na]+; found: 229.0847.
Dimethyl-5-fluorocyclohex-3-ene-1,2-dicarboxylate ((Æ)-15):
A stirring
1,1-[[(1S,2R,5R)-5-Fluorocyclohex-3-ene-1,2-diyl]bis(oxymethylene)]di-
benzene ((À)-11a) and 1,1-[[(1S,2R,5S)-5-fluorocyclohex-3-ene-1,2-diyl]-
bis(oxymethylene)]dibenzene ((À)-11b): An inseparable mixture of dia-
stereomers in a ratio 1:3.72 syn/anti was obtained. Column chromatogra-
phy (hexane/Et2O 90:10, Rf =0.23) afforded a mixture of (À)-11a and
(À)-11b as a colourless oil (0.35 g, 1.125 mmol, 82% yield). [a]2D98 =À60.4
(c=0.63 in CHCl3); 1H NMR (400 MHz, CDCl3): d=7.41–7.30 (m, 10H;
solution of the allylsilane (0.08 g, 0.35 mmol) in CH3CN (4 mL) was treat-
ed with Selectfluor (0.13 g, 0.38 mmol) and stirred at room temperature
for 24 h. H2O (10 mL) was added and the aqueous layer was extracted
with EtOAc (3”10 mL). The combined organics were washed with brine
(20 mL), dried over MgSO4 and the solvent was removed in vacuo.
Column chromatography (30–408C petroleum ether/Et2O 90:10, Rf =0.2)
9184
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 9176 – 9185

Sequential Desymmetrization–Fluorination
FULL PAPER
furnished the anti compound (Æ)-15 (0.027 g, 0.13 mmol, 36% yield) as a
colourless oil. 1H NMR (500 MHz, C3D6O) d=6.07 (ddd, J=3.8, 5.4,
9.7 Hz, 1H), 5.92(m, 1H), 4.96 (dq, J=4.0, 48.0 Hz, 1H), 3.58 (m, 1H),
3.56 (s, 3H), 3.54 (s, 3H), 2.91 (dddd, J=1.0, 5.4, 5.4, 11.5 Hz, 1H), 2.28–
2.13 ppm (m, 2H); 13C NMR (100 MHz, C3D6O): d=173.4, 173.6, 131.1
(d, J=10 Hz), 127.3 (d, J=16 Hz), 84.5 (d, J=161 Hz), 52.3, 52.1, 42.7 (d,
J=3 Hz), 37.3, 28.9 ppm; 19F {1H} NMR (376 MHz, C3D6O): d=
c) M. Tredwell, K. Tenza, M. C. Pacheco, V. Gouverneur, Org. Lett.
2005, 7, 4495–4497; d) M. Tredwell, V. Gouverneur, Org. Biomol.
Chem. 2006, 4, 26–32.
[6] a) For a recent review of fluorinated carbohydrates see a special
issue of Carbohydr. Res. 2000, 327; b) A. P. Kozikowski, A. H. Fauq,
G. Powis, D. C. Melder, J. Am. Chem. Soc. 1990, 112, 4528–4531;
c) R. Miethchen, J. Fluorine Chem. 2004, 125, 895–901; d) J. Faw-
cett, G. A. Griffiths, J. M. Percy, S. Pintat, C. A. Smith, N. S. Spenc-
er, E. Uneyama, Chem. Commun. 2004, 302–303; e) A. Deleuze, C.
Menozzi, M. M. Sollogoub, P. Sinay, Angew. Chem. 2004, 116, 6848–
6851; Angew. Chem. Int. Ed. 2004, 43, 6680–6683; f) J. L. Offer, J. C.
Metcalfe, G. A. Smith, J. Chem. Soc. Chem. Commun. 1990, 1312–
1313; g) H-X. Zhai, P-S. Lei, J. C. Morris, K. Mensa-Wilmot, T. Y.
Shen, Tetrahedron Lett. 1995, 36, 7403–7406; h) D. J. A. Schedler,
D. C. Baker, Carbohydr. Res. 2004, 339, 1585–1595.
[7] a) L. Chabaud, P. James, Y. Landais, Eur. J. Org. Chem. 2004, 3173–
3199; b) Y. Landais, E. Zekri, Eur. J. Org. Chem. 2002, 4037–4053;
c) R. Angelaud, Y. Landais, J. Org. Chem. 1996, 61, 5202–5203;
d) R. Angelaud, O. Babot, T. Charvat, Y. Landais, J. Org. Chem.
1999, 64, 9613–9624.
[8] For further details see the Supporting Information.
[9] Consumption of starting materials was observed; however, no dis-
cernable products could be isolated.
[10] V. VanRheenen, R. C. Kelly, D. Y. Cha, Tetrahedron Lett. 1976, 17,
1973–1976.
[11] J. A. Dale, D. L. Dull, H. S. Mosher, J. Org. Chem. 1969, 34, 2543–
2549.
[12] M. J. C. Buckle, I. Fleming, S. Gil, K. L. C. Pang, Org. Biomol.
Chem. 2004, 2, 749–769.
[13] P. T. Nyffeler, S. G. Duron, M. D. Burkart, S. P. Vincent, C.-H.
Wong, Angew. Chem. 2005, 117, 196–217; Angew. Chem. Int. Ed.
2005, 44, 192–212.
[14] The twist form resulting from an equatorial electrophilic attack is
over 5 kcalmolÀ1 less stable than the chair, resulting from an axial
attack: G. M. Kellie, F. G. Ridell, Top. Stereochem. 1974, 8, 225.
[15] Y. Senda, S. Imaizumi, Tetrahedron 1974, 30, 3813–3815.
[16] a) M. J. Carter, I. Fleming, A. Percival, J. Chem. Soc. Perkin Trans. 1
1981, 2415–2434; b) M. J. Carter, I. Fleming, J. Chem. Soc. Chem.
Commun. 1976, 679–680.
[17] G. Haufe, J. Fluorine Chem. 2004, 125, 875–894.
[18] CCDC-611642contains the supplementary crystallographic data for
this paper. These data can be obtained free of charge from the Cam-
bridge Crystallographic Data Centre via www.ccdc.cam.ac.uk/data_
request/cif.
À167.5 ppm; IR (CDCl3): n˜ =2955, 1740, 1437 cmÀ1
; HRMS (CI):
m/z: calcd for C10H14O4F: 217.0876 [M+H]+; found: 217.0871.
Acknowledgements
We thank the EPSRC and AstraZeneca for supporting this project. We
also thank Professor Y. Landais for helpful advise on the synthesis of
compounds 2a–b.
[1] a) P. Kirsch, Modern Fluoroorganic Chemistry, Wiley, Weinheim,
2004; b) R. D. Chambers, Fluorine in Organic Chemistry, Blackwell
Publishing, Oxford, 2004.
[2] For recent reviews on catalytic asymmetric fluorinations see:
a) P. M. Pihko, Angew. Chem. 2006, 118, 558–561; Angew. Chem.
Int. Ed. 2006, 45, 544–547; b) J.-A. Ma, D. Cahard, Chem. Rev.
2004, 104, 6119–6146; c) H. Ibrahim, A. Togni, Chem. Commun.
2004, 1147–1155; d) C. Bobbio, V. Gouverneur, Org. Biomol. Chem.
2006, 4, 2065–2075.
[3] Y. Huang, A. M. Walji, C. H. Larsen, D. W. C. MacMillan, J. Am.
Chem. Soc. 2005, 127, 15051—15053.
[4] a) R. Brettle, R. Cross, M. Frederickson, E. Haslam, F. S. MacBeath,
G. M. Davies, Bioorg. Med. Chem. Lett. 1996, 6, 1275–1278; b) R.
Brettle, R. Cross, M. Frederickson, E. Haslam, F. S. MacBeath,
G. M. Davies, Tetrahedron 1996, 52, 10547–10556; c) S. Jiang, G.
Singh, D. J. Boam, J. R. Coggins, Tetrahedron: Asymmetry 1999, 10,
4087–4090; d) L. Begum, M. G. B. Drew, J. L. Humphreys, D. J.
Lowes, P. R. Russi, H. L. Whitby, R. C. Whitehead, Tetrahedron
Lett. 2004, 45, 6249–6253; e) E. J. Parker, J. R. Coggins, C. Abell, J.
Org. Chem. 1997, 62, 8582–8585; f) E. J. Parker, C. G Bello, J. R.
Coggins, A. R. Hawkins, C. Abell, Bioorg. Med. Chem. Lett. 2000,
10, 231–234; g) M. N. Ramjee, S. Balasubramanian, C. Abell, J. R.
Coggins, G. M. Davies, T. R. Hawkes, D. J. Lowe, R. N. F. Thorneley,
J. Am. Chem. Soc. 1992, 114, 3151–3153.
[5] a) B. Greedy, J.-M. Paris, T. Vidal, V. Gouverneur, Angew. Chem.
2003, 115, 3413–3416; Angew. Chem. Int. Ed. 2003, 42, 3291–3294;
b) S. Thibaudeau, V. Gouverneur, Org. Lett. 2003, 5, 4891–4893;
Received: June 20, 2006
Published online: October 25, 2006
Chem. Eur. J. 2006, 12, 9176 – 9185
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
9185