Addition of electrophilic radicals to 2-benzyloxyglycals: Synthesis and functionalization of fluorinated α-C-glycosides and derivatives

Article abstract of DOI:10.1002/chem.201302070

A new method for the synthesis of fluorinated α-C-glycosides is described. The reactions between highly electrophilic radicals (fluorinated or unfluorinated) and a 2-benzyloxyglucal or galactal provide 2-keto-D-arabino- or 2-keto-D-lyxo-hexopyranosides th

Full text of DOI:10.1002/chem.201302070

DOI: 10.1002/chem.201302070
Addition of Electrophilic Radicals to 2-Benzyloxyglycals: Synthesis and
Functionalization of Fluorinated a-C-Glycosides and Derivatives
Sophie Colombel,[a] Nathalie Van Hijfte,[a] Thomas Poisson,[a] Eric Leclerc,*[b] and
Xavier Pannecoucke*[a]
Abstract: A new method for the syn-
thesis of fluorinated a-C-glycosides is
described. The reactions between
highly electrophilic radicals (fluorinat-
ed or unfluorinated) and a 2-benzylox-
yglucal or galactal provide 2-keto-d-
arabino- or 2-keto-d-lyxo-hexopyrano-
sides through an addition/fragmenta-
tion process. Sodium borohydride
mediated or Meerwein–Ponndorf–
Verley (MPV) reduction of these com-
pounds provides a-C-glycosides that
feature appropriate anchoring groups
for further synthetic elaboration. The
presence of CF2CO2iPr or CF2Br
groups at the pseudo-anomeric position
allows efficient reduction/olefination or
Br/Li-exchange/nucleophilic-addition
sequences. These transformations open
the way for the synthesis of fluorinated
C-glycosidic analogues of glycoconju-
gates.
Keywords: fluorine · glycomimet-
ics · olefination · radicals · reduc-
tion
Introduction
of C-glycosidic analogues of immunoregulative a-galactosyl-
ceramides by the Franck group. These molecules exhibited a
100-to-1000-fold increase in activity against malaria or mela-
noma cells compared to their parent O-glycolipid.[3] Al-
though isolated, this example demonstrates the relevance of
preparing C-glycosidic surrogates of biologically active O-
glycoconjugates. A further step in the development of C-gly-
cosides as glycomimetics was to replace the anomeric
oxygen atom by a CHF or CF2 group instead of a nonpolar
CH2 group.[4] Thanks to the physical and electronic proper-
ties of fluorine and fluoroalkyl groups (atom size, electrone-
gativity, bond energy, length, etc.), this replacement was ex-
pected to provide compounds with better mimicking abili-
ties.[5–7] Of course, this hypothesis remains to be challenged,
but first requires the development of efficient and stereose-
lective methods that allow the preparation of analogues of
complex O-glycoconjugates. However, despite the work that
has been performed by our group and others since the semi-
nal publications of Motherwell and co-workers, the chemist’s
toolbox is far less complete than that for the synthesis of
classical CH2-glycosides.[4,8,9] Therefore, we wanted to devel-
op strategies that afford selective access to CF2-glycosidic
building blocks for a given carbohydrate series and a given
configuration of the pseudo-anomeric center. Moreover,
these intermediates should feature a functional group that
would allow the introduction of an aglycon chain, for an ef-
ficient synthesis of glycoconjugate analogues. In two previ-
ous communications, we described a new method based on
the addition of difluoromethyl radicals to 2-benzyloxygly-
cals.[10] In addition to these preliminary results, herein, we
report a full study on this topic, including the extension of
the addition reaction to other electrophilic radicals, as well
as additional results for the reduction step and for the func-
tionalization reactions.
Among the various biomolecules, carbohydrates and their
derivatives, such as glycopeptides and glycolipids, have re-
ceived careful attention in recent years for the purpose of
drug research.[1] Indeed, such structures often play a crucial
role in various biological processes, in particular in protein
structure modulation or cell–cell recognition. Therefore,
their use as drugs would be promising if their efficiency
were not undermined by their low metabolic stability, owing
to the in vivo cleavage of the anomeric bond by glycosidas-
es. As such, the resulting low bioavailability of carbohy-
drate-based drugs is a severe drawback that could be over-
come by the creation of analogues with similar bioactivity
and increased in vivo stability. C-glycosides, in which the
anomeric oxygen atom is replaced by a CH2 group, were the
first structures to be developed for this purpose.[2] Surpris-
ingly, their synthesis has been intensively investigated but
scarcely applied to the preparation of analogues of bioactive
glycoconjugates. One striking exception is the preparation
[a] Dr. S. Colombel, Dr. N. Van Hijfte, Dr. T. Poisson,
Prof. Dr. X. Pannecoucke
Normandie Universitꢀ
COBRA, UMR 6014 et FR 3038
Universitꢀ Rouen
INSA Rouen, CNRS, 1 rue Tesniꢁre
76821 Mont Saint-Aignan Cedex (France)
E-mail: xavier.pannecoucke@insa-rouen.fr
[b] Dr. E. Leclerc
Institut Charles Gerhardt
UMR 5253 CNRS-UM2-UM1-ENSCM 8
rue de l’Ecole Normale
34296 Montpellier Cedex 5 (France)
E-mail: eric.leclerc@enscm.fr
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201302070.
12778
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 12778 – 12787

FULL PAPER
Results and Discussion
tions led to the isolation of two diastereomers in the crude
mixture, that is, compounds 4a and 4b, in a 3:1 ratio (by
19F NMR spectroscopy), which were readily separated by
column chromatography on silica gel (Scheme 2). Then,
The addition of perfluoroalkyl or difluoromethyl radicals to
double bonds is a well-known and efficient method for the
synthesis of fluorine-containing molecules.[11] The addition
of difluoromethyl and other electrophilic radicals to stand-
ard glycals has also been studied.[12,13] This addition reaction
was regioselective and exclusively took place at the C2
carbon atom, although the introduction of an alkoxy sub-
stituent at this position was expected to direct the addition
to the less-hindered C1 carbon atom (Scheme 1). This as-
Scheme 2. Addition of ethyl bromodifluoroacetate to 2-benzyloxy-d-
glucal 3.
these optimized conditions were tested on benzylated d-gal-
actal 5 and the appreciable a selectivity for compound 3 im-
proved to complete selectivity. Indeed, 2-ketohexopyrano-
side 6 was present as a single diastereomer in the crude mix-
ture (19F NMR spectroscopy) and could be isolated in 58%
yield (Table 1, entry 1). Thus, this method appeared to be a
good route to a-C-galactosides and, hence, the scope of
alkyl radicals that could undergo an addition to compound 5
Scheme 1. Synthesis of CF2-glycosides through radical addition to 2-al-
koxyglycals.
sumption was confirmed by the group of Miethchen, who re-
ported the sodium dithionite mediated addition of bromo-
chlorodifluoromethane to compound 2 during the early
stages of our study.[9e] The corresponding a-CF2-glucoside
analogue (1, Y=Cl) was obtained with good selectivity, but
with low conversion (Scheme 1).
We were interested in the introduction of fluorinated syn-
thons that were more suitable for further synthetic elabora-
tion than the moderately reactive CF2Cl group.[9f] As such,
ethyl bromodifluoroacetate, a commercially available and
easy-to-handle fluorinated synthon, or dibromodifluorome-
thane seemed more appropriate for such a purpose. The ad-
dition of the CF2CO2Et radical to double bonds generally
involves the use of the less-common iodide or requires the
painstaking synthesis of the corresponding selenide.[14]
Indeed, the use of ethyl bromodifluoroacetate has scarcely
been reported; nonetheless, the reaction was tested by using
this reagent[15] with acetylated
À
was explored next. As anticipated, the addition of R X to
this electron-rich double bond only occurred efficiently with
electrophilic radicals. Dibromodifluoromethane underwent a
clean and fast addition, because 2-oxogalactoside 7 was the
sole reaction product, according to TLC and NMR analysis
of the crude mixture. However, the moderate stability of
compound 7 on silica gel led to a disappointing 41% yield
after chromatography (entry 2). The addition of fluorotri-
bromomethane was less efficient, because compound 8 was
isolated in poor yield from a complex reaction mixture
(entry 3). The xanthate that was derived from ethyl bromo-
fluoroacetate underwent a dilauroyl peroxide (DLP)-pro-
moted addition to afford compound 9 in 59% yield and as a
1:1 mixture of C1’ epimers (entry 4).[17,18] The same reaction
conditions allowed the addition of diethyl bromomalonate
to occur (entry 5). It remains unclear why the DLP-mediat-
or benzylated d-glucal (2 or 3)
as the starting material.[16]
Table 1. Addition of electrophilic radicals to 2-benzyloxy-d-galactal 5.
A short survey of various in-
itiators and conditions led to a
first positive result: 2-Ketohex-
opyranoside 4 was isolated, al-
though in trace amounts, by
using benzylated d-glucal 3 as
the substrate and triethylbor-
ane as the initiator under non-
reductive conditions in CH2Cl2.
Pleasingly, polar solvents great-
ly promoted the reaction and
the yield was increased to 21%
in a 2:1 THF/water mixture
and to 51% in DMF, by using
an excess of the reagents.
These latter two sets of condi-
À
Entry
R X (equiv)
Initiator (equiv)
Conditions
Product
Yield [%]
À
1
2
3
EtO2CCF2 Br (5)
BrCF2 Br (10)
Br2CF Br (5)
BEt3 (2.5)
BEt3 (2)
BEt3 (1)
DMF, air, RT
DMF, air, RT
DMF, air, RT
6
7
8
58
41
16
À
À
4
5
DLP (1)
DLP (1)
DCE/tBuOH (1:1), 958C
DCE/tBuOH (1:1), 958C
9
59
54
10
6
7
Mn
N
MeCN, 858C
10
11
52
38
À
PhSO2CF2 Br (3)
BEt3 (3)
DMF, air, RT
Chem. Eur. J. 2013, 19, 12778 – 12787
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
12779

E. Leclerc, X. Pannecoucke et al.
ed conditions were better suited to these last two examples,
for which lower yields were obtained with BEt3. Thus, both
sets of conditions were examined for each substrate to deter-
mine the appropriate procedure.[19] Compound 10 could also
be obtained in similar yield directly from diethyl malonate
by using MnIII-mediated oxidative conditions (entry 6).[13,20]
Finally, bromodifluoromethyl phenyl sulfone was also exam-
ined and the reaction afforded compound 11 in modest yield
(entry 7).[21] In all cases, only the a diastereomer was detect-
ed and isolated.
To our disappointment, no reactions occurred with dibro-
momethane, ethyl bromoacetate, di-isopropyl bromomethyl-
phosphonate, or bromomethyl phenyl sulfone, under any
conditions. We first suspected that the electrophilicity of the
corresponding radicals was too low. However, this explana-
tion was surprising and unsatisfactory for explaining the re-
sults that were obtained with dibromodifluoromethane.
Indeed, the electrophilicity of simple trifluoromethyl or di-
fluoromethyl radicals is often overemphasized, as demon-
strated by several experimental and theoretical studies.[22,23]
For example, the trifluoromethyl radical has recently been
classified as less electrophilic than the tert-butoxycarbonyl-
methyl radical and ranked as a weak nucleophile.[23] Al-
Scheme 3. Rationale for the stereoselectivity of the radical addition.
Next, the reduction of 2-ketohexopyranosides 4 and 6–10
into their corresponding C-glycosides was explored. First,
the simple sodium borohydride mediated reductions of glu-
cose derivatives 4a and 4b were examined and afforded the
desired C-glycosides. Notably, low temperatures were re-
quired to avoid the simultaneous reduction of the difluor-
oester group: This undesired transformation occurred at
room temperature and compound 12 could be obtained in
68% yield if a three-fold excess of sodium borohydride was
used (Scheme 4). The selective reduction of the ketone
group was achieved by performing the reduction at À788C.
Then, a-d-glucoside 13 and b-d-mannoside 14 were obtained
from compounds 4a and 4b, respectively, in fair yields.
Moreover, in each case, only one diastereomer could be de-
tected from the 19F NMR spectrum of the crude mixtures
(Scheme 4). As described in a preliminary communication,
the configurations of compounds 12, 13, and 14 were con-
firmed by careful NMR spectroscopic experiments and an
X-ray diffraction study on crystalline compound 13.[10a] As
reported in the study of Jimꢀnez-Barbero, Vogel, and co-
workers on CF2 analogues of an a-galactoside, compound 13
C
though no data are available for the CF2Br radical, its be-
C
havior should be close to that of CF3 and, thus, it should
C
hardly be considered as more electrophilic than CH2CO2Et.
However, the generally higher reactivity of fluoro- and di-
fluoromethyl radicals relative to their hydrogenated counter-
parts should certainly account for these results.[11a] The posi-
tive effect of a polar solvent on BEt3-mediated reactions is
also in agreement with the general behavior of fluoroalkyl
radicals. Indeed, the addition reactions of such radicals to
electron-rich double bonds generally gives rise to polar tran-
sition states, which are stabilized in polar media.[11a] The for-
mation of 2-ketohexopyranosides in these reactions results
from a fragmentation of the radical that is obtained after
C
the addition of CF2CO2Et. The driving force for this frag-
mentation process was probably the departure of a stabi-
lized tolyl radical, even if no byproducts that could account
for the formation of this radical (benzyl bromide, 1,2-diphe-
nylethane, etc.) were isolated (Scheme 3).[24] In addition,
monitoring of these reactions by 19F NMR spectroscopy
never showed the presence of other addition products than
compounds 6–11, thus ruling out a putative bromohydrin-
type intermediate that would collapse upon hydrolysis of
the crude reaction mixture. Finally, the rationale for the
strong a selectivity that is observed in these reactions relies
on the same line of reasoning as the interpretation of Le Bel
et al. and Beckwith regarding the addition of sulfuryl radi-
cals to cyclohexene.[25] The attack on the pro-equatorial face
of glucal 3 is indeed disfavored because it would lead to a
twist intermediate through a high-energy transition state. If
the same explanation also stands for galactal 5, then the
attack on the pro-equatorial face would be even more disfa-
vored, owing to steric repulsion between the axial C4 sub-
stituent and the incoming radical.
4
adopted a classical C1 chair conformation in the solid state,
whereas the vicinal proton–proton coupling constants clearly
indicated that this conformation was not the only one in sol-
ution.[9k] The observed stereoselectivities were in agreement
with the numerous reports on the reduction of a- and b-2-
keto-d-arabino-hexopyranosides.[26]
A similar reduction of galactose derivatives 6–10 provided
CF2-glycosides 15–18 in good yields and with complete dia-
stereoselectivity (Scheme 4). In contrast to the glucose de-
rivatives, the determination of the relative configuration of
these compounds was not straightforward. HOESY
19F/1H NMR spectroscopic experiments showed no signifi-
cant correlations and the coupling constants that were ex-
1
tracted from the H NMR spectra of compounds 15 and 16
were not compatible with a 4C1 conformation. Moreover,
1
the H NMR spectrum of carboxylic acid 19, which was de-
rived from compound 15, revealed that the J
ACHTUGNERTN(NUGN H1,H2), J-
A
ACHTUNGTRENNUNG
12780
www.chemeurj.org
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 12778 – 12787

Addition of Electrophilic Radicals to 2-Benzyloxyglycals
FULL PAPER
Scheme 4. Reduction of 2-ketohexopyranosides 4 and 6–10 with sodium borohydride.
with an axial–axial–equatorial disposition of atoms H1/H2/
conditions. A two-step/single-purification procedure for the
radical-addition/reduction sequence allowed us to signifi-
cantly improve the global yields (Scheme 5). The case of
compound 7 is notable because avoiding the purification of
this sensitive compound led to an appreciable 53% global
yield of compound 21 from galactal 5. However, this state-
ment should be tempered by the slight decrease in yield that
was observed when up-scaling the reaction (41% starting
from 2 g of compound 5).
H3. Thus, compounds 15–18 were assigned to have an a-d-
1
taloside configuration, with a C4 conformation for the most-
populated conformer in each case (Scheme 4). Although dis-
appointing, this stereochemical outcome has previously been
reported for other 2-oxogalactosides.[27] Other hydride-medi-
ated reductions were tested (diisobutylaluminum hydride
(DIBAH), l-selectride, Et3BHLi, etc.), but led to the same
stereoselectivity. Substitution reactions with inversion of
configuration at the C2-position were also attempted from
compound 15, but met with failure. For instance, the Mitsu-
nobu-type reactions only afforded unreacted starting materi-
al and substitution reactions of the triflate that was derived
from compound 15 did not afford the desired a-d-galacto-
side, but rather yielded an elimination product.
To overcome the issue of diastereoselectivity in the metal
hydride mediated reduction of 2-oxogalactosides 6–10, we
followed a very simple line of reasoning: Because these re-
actions are presumably under kinetic control, that is, they
proceed through an attack of the hydride on the least-hin-
dered face, the use of a thermodynamically controlled re-
duction reaction might reverse the selectivity. As a matter
of fact, compound 15 was not expected to be the most-
stable diastereomer because of its presumably strong 1,3-di-
axial interaction between the substituents at the C3 and C5-
positions. Thus, the Meerwein–Ponndorf–Verley (MPV) re-
action, a well-known hydride-transfer and reversible reduc-
tion reaction, was considered.[28] By using standard condi-
tions (aluminum isopropoxide at reflux in isopropanol),
CF2-glycosides 20–22 were obtained from compounds 6,
7,and 9, respectively, in good yields, each time as a single di-
astereomer. In contrast to the use of sodium borohydride, a
direct reduction of the crude compound that was obtained
from the radical addition was even possible under the MPV
Because the NMR spectroscopic data of compounds 16
and 21 were clearly different, we concluded that they were
diastereomers and, therefore, that compound 21 was the de-
sired a-CF2-d-galactoside. Ethyl esters were converted into
isopropyl esters during the MPV reduction but the saponifi-
cation of compounds 15 and 20 led to carboxylic acids with
different NMR spectroscopic data, thus leading to the same
conclusion. The saponification of monofluoroesters 17 and
22 also led to different sets of diastereomers. The coupling
constants between atoms H1, H2, and H3 in compound 20
were now within the same range as for a-C-glucose deriva-
4
tives 12 and 13, thus indicating that, once again, the C1 con-
formation did not seem to be the most-populated one in sol-
ution. However, the H1/H2, H3/H5 and H1/H6 NOESY cor-
relations could fit with an a-d-galactoside configuration and
1
1
with contributions from the S3 skew-boat and C4 chair con-
formations.[10k] Finally, an X-ray diffraction study of com-
pound 21 confirmed the configuration at the C2 position.
However, crystallization of this benzyl-protected compound
occurred in a 1C4 chair conformation, in contrast with the re-
sults that were obtained by Jimꢀnez-Barbero, Vogel, and co-
workers. Indeed, their fully deprotected a-CF2-galactoside
4
adopted a C1 conformation in the solid state. However, this
time in agreement with the above-mentioned study, our
compound appeared to be flexible in solution. For example,
Chem. Eur. J. 2013, 19, 12778 – 12787
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
12781

E. Leclerc, X. Pannecoucke et al.
Scheme 6. Experiments to challenge the reversibility of the reaction.
probably arises from a 1,3-diaxial interaction between the
C4 benzyloxy group and the incoming hydride, which would
be stronger than the 1,2-gauche interaction. As suggested in
our first communication, a different conformational behav-
ior between compounds 4a and 6–10 might also favor this
orientation.[10a] Finally, if the relative energies of the two
possible diastereomers are responsible of the observed se-
lectivity in the MPV reduction of compounds 6–10, the
strong 1,3-diaxial interactions in structures B and B’ would
certainly account for the preferential formation of structure
A and compounds 20–22. However, given the observed con-
formational flexibility of compounds 15–22, these transition
states should be merely considered as a working hypothesis.
Thanks to this radical-addition/reduction process, we had
several functionalized fluorinated a-C-glycosides in hand
and we wanted to explore their further synthetic elabora-
tion. Our goal was to develop a method that could allow us
to convert these advanced intermediates into their glycocon-
jugate analogues. The CFXCO2R and CF2Br groups in de-
rivatives 13, 15, and 20–22 appeared to be appropriate an-
choring groups for introducing aglycon moieties through
olefination reactions or Br/Li exchange, followed by nucleo-
philic addition. Thus, the reduction/olefination sequence
that was reported in the preliminary communication for
compound 24 was also applied to tallose and galactose de-
rivatives 26, 28, and 30, thereby affording the corresponding
products in satisfactory yields (Scheme 8). As reported in
the earlier communication, the double bond in compounds
25 and 27 was successfully hydrogenated in the presence of
the benzyl ether protecting groups by using Wilkinson’s cat-
alyst.[10a] If successful, such a reaction sequence with a-ami-
Scheme 5. MPV reduction of compounds 6–9.
a significant H3/H5 NOESY correlation, which was only
1
4
compatible with S3 skew-boat or C1 chair conformations,
1
was observed, thus demonstrating that the C4 chair confor-
mation, if populated, was not exclusive.
To gain deeper insight into the mechanism of the MPV re-
action, the reversibility of the reaction was tested though
several experiments. Compound 15 (the “kinetic diastereom-
er”) was converted into its isopropyl ester (23) and placed
under the MPV reduction conditions by using acetone or 3-
pentanone as a co-oxidant. No equilibration towards the
“thermodynamic diastereomer” (20) occurred under these
conditions. Initially, we thought that the oxidation potentials
of acetone and 3-pentanone were too low to reverse the re-
action. Thus, these co-solvents were replaced with 1.2 equiv-
alents of starting 2-ketopyranoside 6 to mimic the exact re-
action conditions (Scheme 6). A 1:1.2 mixture of compounds
23 and 20, was obtained, with the latter product arising from
the reduction of compound 6. Clearly, no equilibration took
place during the process and, therefore, the MPV reduction
of compound 20 was not reversible. However, as demon-
strated by Burke et al., irreversible MPV reductions are
characterized by a late transition state. Consequently, they
typically afford the thermodynamically more-stable product,
even if the reaction is not reversible.[29] A possible rationale
for the stereoselectivity of the different reduction reactions
is described in Scheme 7. The sodium borohydride mediated
reduction of glucose derivatives 4a and 4b follows the liter-
ature reports, that is, an attack of the hydride on the oppo-
site side of the CF2Y group to avoid 1,2-gauche interactions.
The inversion of selectivity for galactose derivatives 6–10
nophosphonoacetates
as
Horner–Wadsworth–Emmons
(HWE) reagents would have provided compounds 32, which
featured the required functional groups for a synthesis of a-
C-galactosylserine or a-C-galactosylceramide. Disappoint-
ingly, despite our numerous attempts, the HWE reactions
only met with failure. The hemiketal nature of the inter-
mediate might explain this lack of reactivity, even though
similar reactions have been reported on less functionalized
fluorinated substrates.[30]
An alternative to this HWE strategy was to explore the
derivatization of the CF2Br group of compound 21. Bro-
mine/lithium exchange, followed by addition onto an appro-
priate electrophile, would provide an interesting route to
12782
www.chemeurj.org
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 12778 – 12787

Addition of Electrophilic Radicals to 2-Benzyloxyglycals
FULL PAPER
Scheme 7. Rationale for the diastereoselectivities that were observed in the reduction reactions.
(34) was obtained in 21–45% yield, if at all, along
with various amounts of compound 35, which resulted
from
a
Fritsch–Buttenberg–Wiechell
rearrangement
(Scheme 9).[31,32] Of course, this typical carbenoid rearrange-
ment also occurs in the absence of an electrophile, but the
yield of compound 35 remained, at best, moderate.
Scheme 9. Br/Li exchange and intermolecular addition to electrophiles.
Thus, this strategy appeared to be compromised until a
Br/Li exchange on OTBS-derivative 36 (TBS=tert-butyldi-
methylsilyl) was performed, which resulted in a migration of
the TBS group from the O2 to C1’ atoms to provide the cor-
responding CF2TBS derivative (37) in a nonoptimized 44%
yield (Scheme 10). Despite this moderate yield, the reaction
seemed to proceed cleanly and no degradation product was
detected. Thus, intramolecular trapping of the lithium spe-
cies emerged as the method of choice to circumvent the po-
tential problem of stability. As a matter of fact, Br/Li ex-
change on compound 38 resulted in the immediate trapping
of the lithium species by the neighboring acetate.[33] Stable
hemiketal species 39 was isolated in 75% yield as an insepa-
rable mixture of diasteromers. Reduction of this intermedi-
ate provided alcohol 40 in 68% yield. Therefore, the global
transformation can be considered to be a formal addition of
the lithiated anion of compound 21 to acetaldehyde. The
two diastereomers of compound 40 were only partially sepa-
rated and their ratio could only be estimated to be approxi-
mately 2:1.
Scheme 8. Side chain elongation by HWE olefination. Boc=tert-butoxy-
carbonyl; Cbz=carbobenzyloxy.
glycoconjugate analogues. However, the carbenoid nature of
the corresponding RCF2Li species was worrisome because it
would probably compromise its thermal stability.[31] Our first
attempts at generating this lithium compound confirmed
this prediction because performing a Br/Li exchange on
compound 33, followed by trapping with Garner’s aldehyde,
led to poor and irreproducible results. The addition product
Chem. Eur. J. 2013, 19, 12778 – 12787
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
12783

E. Leclerc, X. Pannecoucke et al.
Scheme 10. Br/Li exchange and intramolecular addition to electrophiles.
DMAP=4-dimethylaminopyridine.
This reductive migration of an ester from the O2 to C1’-
positions appeared as a nice entry to O-glycoconjugate ana-
logues, provided that a highly functionalized ester was
placed at the 2-position of compound 21. Thus, two carbox-
ylic acids of increasing complexity were prepared and cou-
pled with compound 21 (Scheme 11). The acetonide that
was derived from N-Boc-d-serine methyl ester, a precursor
of Garner’s aldehyde,[34b] was converted into acid 41 and
then into ester 42 through a N,N’-diisopropylcarbodiimide
(DIPC)-mediated coupling reaction with compound 21. A
more complex substrate was prepared by using phytosphin-
gosine-derived carboxylic acid 44. Savage’s procedure was
used to prepare compound 43 from Garner’s aldehyde.[35]
Benzylation, acetonide hydrolysis, and a two-step procedure
that involved a 2-iodoxybenzoic acid (IBX)-mediated Pin-
nick oxidation yielded 44. The same esterification method as
described above was used to prepare compound 45 from 21
and 44. With two functionalized substrates (42 and 45) in
hand, the reductive migration sequence was attempted next.
Starting from compound 42, the sequence was uneventful:
Br/Li exchange with 1.2 equivalents of nBuLi at À788C pro-
vided the corresponding hemiketal (46) in 57% yield as a
mixture of diastereomers (Scheme 12). Reduction of the
latter compound afforded the desired functionalized a-C-
galactoside (47) in 57% yield, once again as an inseparable
mixture of diastereomers; however, the evaluation of the di-
astereomeric ratio was impossible, owing to the presence of
rotamers. In contrast, the reaction with phytosphingosine-
derived substrate 45 required some optimization. The use of
Scheme 12. Reductive acyl chain migration sequence with functionalized
substrates.
1.2 equivalents of nBuLi afforded the expected hemiketal
(48) in low yield (<15%), along with recovery of large
amounts of the starting material (50%) and of its reduced
CF2H derivative (35%). Performing the reaction with
2.2 equivalents of nBuLi allowed us to obtain complete con-
version, but the results were rather irreproducible and a
maximum yield of 50% was only obtained once. Whereas,
tBuLi only led to degradation, MeLi was a much better me-
diator because the addition of 2.2 equivalents of this reagent
yielded compound 48 in a reproducible 70% yield. The re-
duction of this hemiketal was performed as above and com-
pound 49 was obtained in 74% yield. Surprisingly, this com-
pound was isolated as a single diastereomer. The conversion
of this compound into oxazolidinone 50 under basic condi-
tions was an opportunity to determine the configuration at
Scheme 11. Synthesis of functionalized carboxylic acids.
12784
www.chemeurj.org
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 12778 – 12787

Addition of Electrophilic Radicals to 2-Benzyloxyglycals
FULL PAPER
1
4.91 (d, J=11.4 Hz, 1H), 4.59 (d, J=10.3 Hz, 1H), 4.55 (d, J=9.7 Hz,
1H), 4.53–4.42 (m, 5H), 4.30 (app t, JACTHNUGRTNEUNG(app)=2.4 Hz, 1H), 3.68–3.59 ppm
the C3’-position; however, the H NMR spectrum of com-
pound 50 was crowded with the signals of the benzyl groups
and, thus, useless for NOESY experiments or for the extrac-
tion of coupling constants. Importantly, a standard hydroge-
nolysis reaction provided compound 51 and a solution to
(m, 2H); 19F NMR (282.5 MHz; CDCl3): d=À52.4 (dd, J
ACHTUNGTREN(NGNU F,F)=168.8, J-
ACHUTNGTRENNUG(F,F)=168.8, JAHCTUNGTREN(NUNG F,H)=14.0 Hz, 1F);
(F,H)=8.0 Hz, 1F), À54.4 ppm (dd, J
13C NMR (75.5 MHz; CDCl3): d=199.4, 137.8, 137.7, 137.1, 128.7, 128.6,
128.5, 128.4, 128.3, 128.1, 128.0, 127.9, 119.4 (t, J=313.1 Hz), 83.5 (t, J=
23.6 Hz), 83.5, 77.5, 77.0, 75.0, 74.0, 73.6, 68.0 ppm; IR: n˜ =3090, 3064,
1
this problem, because, this time, the H NMR spectrum of
3031 (nCHAr), 2925, 2858 (nCHAliph), 1760 cmÀ1 (nC O); MS (ESI+): m/z:
=
this compound was much clearer. The strong NOESY corre-
lation and the 7.4 Hz scalar coupling constant between the
H2’ and H3’ atoms were, beyond doubt, diagnostic of a cis
relationship between these two protons.[36] Thus, this O2-to-
C1’ acyl-transfer sequence was extremely efficient, even on
highly functionalized substrates. Provided that a simple solu-
tion can be found to remove the hydroxy group at the 2’-po-
sition, this method could allow us to prepare CF2-glycosidic
analogues of galactosylaminoacids or galactosylceramides.
585.1, 583.1 [M+Na]+; elemental analysis calcd (%) for C28H27BrF2O5:
C 59.24, H 4.60; found: C 59.27, H 4.58.
Ethyl fluoro-2-(3,4,6-tri-O-benzyl-a-d-lyxo-hexopyranosid-2-ulosyl)ace-
tate (9):
A degassed solution of 2,3,4,6-tetra-O-benzyl-d-galactal (5,
0.783 g, 1.5 mmol) and ethyl bromodifluoroacetate derived xanthate
(0.679 g, 3.0 mmol) in 1,2-dichloroethane (3 mL) and tBuOH (3 mL) was
heated at reflux (958C).
A degassed solution of dilauroyl peroxide
(0.598 g, 1.5 mmol) in 1,2-dichloroethane (9 mL) and tBuOH (9 mL) was
added to the refluxing solution over 6 h with a syringe pump. The mix-
ture was heated at reflux for a further 6 h (conversion was checked by
TLC; cyclohexane/EtOAc, 8:2) and evaporated. Purification by chroma-
tography on a flash purification systemflash column chromatography on
silica gel (EtOAc/cyclohexane, 2–20%) afforded compound 9 (mixture of
diastereomers) as a colorless oil (0.115 g, 59% yield). Analytical TLC
(silica gel 60; EtOAc/cyclohexane, 20%): Rf =0.28; 1H NMR (300 MHz;
CDCl3): d=7.38–7.17 (m, 15H and 15H’), 5.36 (dd, J=46.7, J=1.9 Hz,
1H), 5.33 (dd, J=47.1, J=1.8 Hz, 1H’), 5.02 (d, J=11.8 Hz, 1H’), 5.01
(d, J=11.9 Hz, 1H), 4.92 (d, J=11.9 Hz, 1H), 4.91 (d, J=11.8 Hz, 1H’),
4.74–4.48 (m, 5H and 5H’), 4.46–4.33 (m, 2H and 2H’), 4.32–4.18 (m, 3H
and 3H’), 3.68–3.52 (m, 2H and 2H’), 1.25 ppm (t, J=7.2 Hz, 3H and
3H’); 19F NMR (282.5 MHz; CDCl3): d=À200.9 (dd, J=47.4, J=
21.7 Hz, 1F’), À204.2 ppm (dd, J=46.4, J=33.0 Hz, 1F); 13C NMR
(75.5 MHz; CDCl3): d=204.7, 204.4 (d, J=6.0 Hz), 166.6 (d, J=24.2 Hz),
166.4 (d, J=24.2 Hz), 137.9, 137.8, 137.5, 137.2, 128.7, 128.6, 128.5,
128.45, 128.4, 128.3, 128.2, 128.1, 128.0, 127.9, 127.8, 127.7, 127.6, 88.9 (d,
J=194.8 Hz), 88.1 (d, J=192.5 Hz), 83.4, 83.1 (d, J=3.0 Hz), 79.6 (d, J=
20.4 Hz), 79.5 (d, J=20.4 Hz), 76.9, 76.5, 76.1, 76.0, 74.7, 74.5, 73.6, 73.5,
73.3, 73.1, 67.5, 67.3, 62.2, 62.1, 14.2, 14.1 ppm; IR: n˜ =3031.4 (nCH Ar),
Conclusion
Overall, the different strategies that have been described in
this article provide efficient access to fluorinated a-C-glyco-
sides. The core of the method involves an addition of di-
fluoromethyl radicals to 2-benzyloxyglycals, which can be
extended to other electrophilic radicals. Reduction of the re-
sulting 2-ketohexopyranosides afforded the desired a-C-gly-
cosides in satisfactory overall yields with complete diaster-
eoselectivity for each step. The synthetic elaboration of
these compounds has also been studied and useful methods
have been disclosed. In particular, the reactions that in-
volved a Br/Li exchange on derivatives of bromide 21 ap-
peared to be a method of choice for the future synthesis of
O-glycoconjugate analogues. Indeed, the intramolecular ad-
dition of the lithiated species to a neighboring ester group
allows the introduction of various aglycon moieties at the
pseudo-anomeric position through a simple and mild proc-
ess. We are currently investigating the synthesis of a whole
family of difluorinated a-C-galactosylceramides, as well as
the biological evaluation of these compounds.
2918, 2850 (nCHAliph), 1755 cmÀ1 (nC O); MS (ESI+): m/z: 554.1
=
[M+H2O]+; elemental analysis calcd (%) for C31H33FO7: C 69.39, H 6.20;
found: C 69.41, H 6.13.
Bromodifluoro-(3,4,6-tri-O-benzyl-a-d-galactopyranosyl)methane
(21):
Al(OiPr)3 (0.286 g, 1.4 mmol) was added to a solution of bromodifluoro-
AHCTUNGTRENNUNG
(3,4,6-tri-O-benzyl-a-d-lyxo-hexopyranosid-2-ulosyl)methane (7, 0.310 g,
0.56 mmol) in isopropanol. The resulting suspension was heated at reflux
and the reaction was complete within 4–6 h, as monitored by TLC (cyclo-
hexane/EtOAc, 8:2). Then, the mixture was neutralized with a 1m aque-
ous solution of HCl (5 mL) and extracted with EtOAc (4ꢃ10 mL). The
combined organic extracts were washed with water (2ꢃ10 mL), a saturat-
ed aqueous solution of NaHCO3 (2ꢃ10 mL), and brine (10 mL), dried
over Na2SO4, and evaporated. Purification by chromatography on a flash
purification system (EtOAc/cyclohexane, 5–40%) afforded compound 21
as a white crystalline solid (0.212 g, 68% yield). The application of this
procedure directly from crude bromodifluoro-(3,4,6-tri-O-benzyl-a-d-
lyxo-hexopyranosid-2-ulosyl)methane (7) afforded compound 21 in 55%
overall yield from compound 5. Analytical TLC (silica gel 60; EtOAc/cy-
clohexane, 20%): Rf =0.17; [a]2D0 =+37.4 cm3 gÀ1 dmÀ1 (c=0.76, CHCl3);
1H NMR (300 MHz, CD3OD): d=7.34–7.25 (m, 15H), 4.68 (d, J=
11.9 Hz, 1H), 4.61 (d, J=11.9 Hz, 1H), 4.60 (d, J=11.7 Hz, 1H), 4.55
(app s, 2H), 4.54 (d, J=11.7 Hz, 1H), 4.35 (ddd, J=8.5, J=6.0, J=
2.8 Hz, 1H), 4.24 (ddd, J=13.6, J=8.5, J=2.3 Hz, 1H), 4.21–4.17 (m,
1H), 4.13 (dd, J=6.0, J=2.9 Hz, 1H), 4.00 (dd, J=11.8, J=8.5 Hz, 1H),
3.83 (dd, J=4.9, J=2.9 Hz, 1H), 3.74 ppm (dd, J=11.8, J=2.8 Hz, 1H);
19F NMR (282.5 MHz; CD3OD): d=À53.0 (dd, J=164.0, J=14.2 Hz,
1F), À55.3 ppm (d, J=164.0 Hz, 1F); 13C NMR (75.5 MHz; CD3OD):
d=139.6, 139.5, 139.4, 129.4, 129.3, 129.0, 128.9, 128.8, 128.7, 128.6, 123.3
(dd, J=312.6 Hz, J=308.8 Hz), 78.2, 77.0, 75.3 (dd, J=25.0 Hz, J=
20.2 Hz), 74.3, 74.0, 73.9, 73.0, 68.2, 66.7 ppm; IR: n˜ =3410 (nOH), 3031
(nCHAr), 2877 cmÀ1 (nCHAliph); MS (ESI+): m/z: 582.0, 580.0 [M+H2O]+; el-
emental analysis calcd (%) for C28H29BrF205: C 59.69, H 5.19; found:
C 59.94, H 5.15.
Experimental Section
For the sake of brevity, only the most important procedures are described
in the Experimental Section. For full experimental details, see the Sup-
porting Information.
Bromodifluoro-(3,4,6-tri-O-benzyl-a-d-lyxo-hexopyranosid-2-ulosyl)me-
thane (7): Dibromodifluoromethane (0.229 mL, 2.5 mmol, 5 equiv) and
triethylborane (1m in THF, 0.500 mL, 0.5 mmol, 1 equiv) were added to a
solution of 2,3,4,6-tetra-O-benzyl-d-galactal (5, 0.260 g, 0.5 mmol) in aer-
ated DMF (4 mL). The mixture was stirred at RT for 3 h and the conver-
sion was checked by TLC (cyclohexane/EtOAc, 8:2). The addition of
CF2Br2 (5 equiv) and BEt3 (1 equiv) and stirring for a further 3 h were
typically required to reach completion. Then, a saturated aqueous solu-
tion of NH4Cl (50 mL) was added and the mixture was extracted with
Et2O (3ꢃ30 mL). The organic layer was washed with water (5ꢃ30 mL),
dried over MgSO4, and evaporated. Purification by column chromatogra-
phy on silica gel (cyclohexane/EtOAc, 97:3) afforded 7 as a colorless oil
(0.115 g, 41% yield). Analytical TLC (silica gel 60; EtOAc/cyclohexane,
20%): Rf =0.50; [a]D20 =+311.5 cm3 gÀ1 dmÀ1 (c=0.48, CHCl3); 1H NMR
(300 MHz, CDCl3): d=7.40–7.20 (m, 15H), 5.00 (d, J=11.8 Hz, 1H),
Chem. Eur. J. 2013, 19, 12778 – 12787
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
12785

E. Leclerc, X. Pannecoucke et al.
Ethyl 2-fluoro-2-(3,4,6-tri-O-benzyl-a-d-galactopyranosyl)acetate (22):
The same procedure as for the synthesis of compound 21 was applied to
2-fluoro-2-(3,4,6-tri-O-benzyl-a-d-lyxo-hexopyranosid-2-ulosyl)acetate (9,
0.500 g, 0.93 mmol). The reaction was complete within 4–6 h, as moni-
tored by TLC (cyclohexane/EtOAc, 8:2). Purification by chromatography
on a flash purification system (EtOAc/cyclohexane, 3–30%) afforded
compound 22 as a colorless oil (0.298 g, 58% yield). Careful purification
allowed us to obtain fractions that contained each diastereomer as the
major compound. The application of this procedure directly from crude
2-fluoro-2-(3,4,6-tri-O-benzyl-a-d-lyxo-hexopyranosid-2-ulosyl)acetate
(9) afforded compound 22 in 41% overall yield from compound 5.
ane, 7–60%) afforded compound 40 as a colorless oil (0.098 g, 58%
yield). Each diastereomer could be obtained in its almost pure form, with
only a slight amount of the other diastereomer.
Diastereomer 1: Analytical TLC (silica gel 60; 30% EtOAc in cyclohex-
ane): Rf =0.22; 1H NMR (300 MHz; CD3OD): d=7.34–7.23 (m, 15H),
4.70–4.45 (m, 6H), 4.41–4.25 (m, 2H), 4.22–4.02 (m, 4H), 3.87 (dd, J=
4.6, J=2.9 Hz, 1H), 3.71 (dd, J=11.7, J=2.5 Hz, 1H), 1.22 ppm (d, J=
6.8 Hz, 3H); 19F NMR (282.5 MHz; CD3OD): d=À120.7 (dd, J
ACHTUNGTRENNUNG
259.9, J
(F,H)=20.6 Hz, 1F), À127.0 ppm (dd, J
ACHUTGTNREN(NUG F,F)=261.0, JACHTUNGTRENNUNG
20.6 Hz, 1F); 13C NMR (75.5 MHz; CD3OD): d=139.8, 139.7, 139.4,
129.5, 129.4, 129.2, 129.1, 129.0, 128.9, 128.8, 128.7, 123.6 (dd, J=253.7,
J=247.1 Hz), 78.0, 76.3, 74.3, 74.2, 74.1, 72.9, 67.6 (dd, J=24.7, J=
8.2 Hz), 67.5, 67.0 (dd, J=22.0, J=12.1 Hz), 66.9, 15.0 ppm.
Diastereomer 1: Analytical TLC (silica gel 60; EtOAc/cyclohexane,
20%): Rf =0.18; 1H NMR (300 MHz; CDCl3): d=7.36–7.23 (m, 15H),
5.38 (dd, J=47.9, J=1.7 Hz, 1H), 5.13 (hept, J=6.2 Hz, 1H), 4.82 (d, J=
11.5 Hz, 1H), 4.74 (d, J=11.5 Hz, 1H), 4.60–4.39 (m, 6H), 4.25 (app t,
J=6.5 Hz, 1H), 4.08 (brs, 1H), 3.99–3.93 (m, 1H), 2.50 (brs, 1H),
1.23 ppm (d, J=6.2 Hz, 6H); 19F NMR (282.5 MHz; CDCl3): d=
À196.7 ppm (appt, J=41.7 Hz, 1F); 13C NMR (75.5 MHz; CDCl3): d=
167.7 (d, J=25.3 Hz), 138.5, 138.1, 137.9, 128.7, 128.6, 128.5, 128.4, 128.2,
128.1, 128.0, 127.9, 127.8, 127.7, 91.0 (d, J=192.7 Hz), 80.3 (d, J=
25.3 Hz), 74.7, 74.6, 74.4 (d, J=9.9 Hz), 73.4, 73.0, 72.0, 69.7, 68.7, 68.5,
21.8 ppm.
Diastereomer 2: Analytical TLC (silica gel 60; 30% EtOAc in cyclohex-
ane): Rf =0.16; 1H NMR (300 MHz; CD3OD): d=7.32–7.22 (m, 15H),
4.65 (d, J=11.7 Hz, 1H), 4.60 (d, J=11.7 Hz, 1H), 4.57 (d, J=11.7 Hz,
1H), 4.55–4.43 (m, 3H), 4.37–4.20 (m, 2H), 4.19–4.02 (m, 4H), 3.86 (dd,
J=4.4 Hz, J=3.1 Hz, 1H), 3.71 (dd, J=11.9 Hz, J=2.3 Hz, 1H),
1.25 ppm (d, J=6.2 Hz, 3H); 19F NMR (282.5 MHz; CD3OD): d=À114.5
(appdt, J
260.9, J
G
E
ACHTUNGTRENNUNG(F,F)=
AHCTUNGTRENNUNG
139.7, 139.6, 129.5, 129.45, 129.4, 129.2, 129.0, 128.9, 128.8, 128.7, 128.6,
123.1 (dd, J=252.0, J=247.6 Hz), 77.6, 76.8, 74.2, 74.1, 72.8, 65.6 (appt,
J=27.2 Hz), 67.7, 67.6 (appt, J=26.4 Hz), 66.7, 16.0 ppm; IR: n˜ =3339
(nOH), 3028 (nCHAr), 2925 cmÀ1 (nCHAliph); MS (ESI+): m/z: 1079.0 [2M<
M+ >Na]+, 551.4 [M+Na]+; elemental analysis calcd (%) for C30H34F206:
C 68.17, H 6.48; found: C 68.28, H 6.44.
Diastereomer 2: Analytical TLC (silica gel 60; EtOAc/cyclohexane,
20%): Rf =0.14; 1H NMR (300 MHz; CDCl3): d=7.38–7.23 (m, 15H),
5.17 (dd, J=48.9, J=2.9 Hz, 1H), 5.12 (hept, J=6.2 Hz, 1H), 4.73 (d, J=
11.7 Hz, 1H), 4.72 (d, J=11.5 Hz, 1H), 4.62–4.44 (m, 5H), 4.32–4.22 (m,
2H), 4.06 (appt, J=2.8 Hz, 1H), 3.94 (dd, J=7.7, J=2.3 Hz, 1H), 3.80
(dd, J=10.2, J=7.0 Hz, 1H), 3.64 (dd, J=10.2, J=7.0 Hz, 1H), 1.25 ppm
(d, J=6.2 Hz, 6H); 19F NMR (282.5 MHz; CDCl3): d=À196.2 (dd, J=
47.4, J=30.9 Hz, 1F); 13C NMR (75.5 MHz; CDCl3): d=167.5 (d, J=
23.1 Hz), 138.3, 138.1, 138.0, 128.7, 128.6, 128.5, 128.4, 128.3, 128.1, 128.0,
127.9, 127.8, 127.75, 127.7, 127.6, 91.5 (d, J=191.1 Hz), 79.0, 74.9 (d, J=
2.2 Hz), 73.5, 73.4, 73.0, 72.5, 69.8, 67.9, 67.8, 21.6 ppm; IR: n˜ =3480
Acknowledgements
This work was supported by the INSA-ROUEN, University of Rouen,
Rꢀgion Haute-Normandie, Centre National de la Recherche Scientifique
(CNRS), European Fund for Research Development (EFRD), Agence
Nationale de la Recherche (ANR), and LABEX SynOrg. The ANR is
gratefully acknowledged for a doctoral fellowship to S.C. The EFRD
(FEDER 43209) is gratefully acknowledged for a postdoctoral fellowship
to N.V.H. We also thank Dr. P. Lameiras and Prof. H. Oulyadi for per-
forming the NOESY and HOESY experiments.
(nOH), 3028 (nCHAr), 2928 (nCHAliph), 1758 cmÀ1 (nC O); MS (ESI+): m/z:
=
575.2 [M+Na]+; 570.1 [M+H2O]+; 553.1 [M+H]+; elemental analysis
calcd (%) for C32H37F07: C 69.55, H 6.75; found: C 69.51, H 6.76.
1,1-Difluoro-1-(3,4,6-tri-O-benzyl-a-d-galactopyranosyl)propan-2-ol (40):
nBuLi (1.53m in hexanes, 0.337 mL, 0.52 mmol) was added to a solution
of acetate 38 (0.260 g, 0.43 mmol), which was obtained by the acetylation
of compound 21 by using Ac2O, NEt3, and DMAP (cat.), in THF (5 mL)
at À788C under an argon atmosphere. The mixture was stirred for 1 h at
À788C (TLC analysis showed complete conversion) and then quenched
with 1m HCl (5 mL) and extracted with EtOAc (3ꢃ10 mL). The com-
bined organic compounds were washed with water (10 mL) and brine
(10 mL), dried over Na2SO4, and evaporated. Purification by chromatog-
raphy on a flash purification system (EtOAc/cyclohexane, 3–30%) af-
forded intermediate 39 as a colorless oil (0.171 g, 75% yield) and as a
mixture of diastereomers. Analytical TLC (silica gel 60; EtOAc/cyclohex-
ane, 15%): Rf =0.16; 1H NMR (300 MHz; CD3OD): d=7.42–7.20 (m,
15H), 4.84–4.64 (m, 4H), 4.57–4.17 (m, 5H), 4.10–3.94 (m, 2H), 3.68–
3.54 (m, 3H), 1.47–1.37 ppm (m, 3H); 19F NMR (282.5 MHz; CD3OD):
[1] Carbohydrate-Based Drug Discovery (Ed.: C.-H. Wong), Wiley-
VCH, Weinheim, 2003.
[2] For reviews on C-glycoside synthesis, see: a) T. Nishikawa, M.
Adachi, M. Isobe in Glycoscience: Chemistry and Chemical Biology,
Vol. 3, 2nd ed. (Eds.: B. Fraser Reid, K. Tatstuta, J. Thiem, G. L.
Cotꢀ, S. Flitsch, Y. Ito, H. Kondo, S. Nishimura, B. Yu), Springer,
Heidelberg, 2008, pp. 755–811; b) B. Vauzeilles, D. Urban, G. Dois-
neau, J. M. Beau in Glycoscience: Chemistry and Chemical Biology,
Vol. 9, 2nd ed. (Eds.: B. Fraser Reid, K. Tatstuta, J. Thiem, G. L.
Cotꢀ, S. Flitsch, Y. Ito, H. Kondo, S. Nishimura, B. Yu), Springer,
Heidelberg, 2008, pp. 2021–2077, and the references therein.
[3] a) G. Chen, M. Chien, M. Tsuji, R. W. Franck, ChemBioChem 2006,
7, 1017–1022; b) R. W. Franck, M. Tsuji, Acc. Chem. Res. 2006, 39,
692–701; c) G. Yang, J. Schmieg, M. Tsuji, R. W. Franck, Angew.
Chem. 2004, 116, 3906–3910; Angew. Chem. Int. Ed. 2004, 43, 3818–
3822.
[4] For a review on nonhydrolyzable fluorinated glycomimetics, see: E.
Leclerc, X. Pannecoucke, M. Ethꢁve-Quelquejeu, M. Sollogoub,
Chem. Soc. Rev. 2013, 42, 4270–4283.
[5] a) P. Kirsch, Modern Fluoroorganic Chemistry, Wiley-VCH, Wein-
heim, 2004; b) H. J. Bçhm, D. Banner, S. Bendels, M. Kansy, B.
Kuhn, K. Mꢄller, U. Obst-Sander, M. Stahl, ChemBioChem 2004, 5,
637–643; c) J.-P. Bꢀguꢀ, D. Bonnet-Delpon, Bioorganic and Medici-
nal Chemistry of Fluorine, Wiley, Hoboken, 2008; d) S. Purser, P. R.
Moore, S. Swallow, V. Gouverneur, Chem. Soc. Rev. 2008, 37, 320–
330.
d=À117.0 (dd, J
232.0, J
(F,H)=19.1 Hz, 1F’), À129.6 (dd, J
1F’), À130.3 ppm (dd, (F,F)=243.4,
(F,H)=3.1 Hz, 1F); 13C NMR
A
N
ACHTUNGTRNE(NUNG F,F)=
A
R
ACHTUNGTRENNUNG
J
N
JACHTUNGTRENNUNG
(75.5 MHz; CD3OD): d= 139.9, 139.8, 139.7, 139.4, 139.3, 129.5, 129.4,
129.3, 129.3, 129.2, 129.1, 129.05, 129.0, 128.9, 128.85, 128.8, 128.75, 128.7,
128.6, 124.5 (dd, J=262.5, J=255.9 Hz), 101.9 (dd, J=31.8, J=23.1 Hz),
100.3 (dd, J=29.7, J=23.6 Hz), 81.5, 81.4, 77.1, 76.9 (d, J=2.7 Hz), 75.4,
75.3, 75.2, 75.1, 74.8 (t, J=25.3 Hz), 74.5 (t, J=16.5 Hz), 74.3, 74.2, 73.9,
73.3, 72.7, 69.6, 68.8, 20.6, 20.55 ppm; MS (ESI+): m/z: 544.1 [M+H2O]+.
NaBH4 (0.019 g, 0.50 mmol) was added to a solution of compound 39
(0.171 g, 0.32 mmol) in MeOH (5 mL) at 08C. The mixture was stirred
for 3 h at 08C, at which point TLC analysis (cyclohexane/EtOAc, 70:30)
showed the complete conversion of compound 39. The mixture was
warmed to RT, quenched with a saturated aqueous solution of NH4Cl,
and extracted with EtOAc (3ꢃ10 mL). The combined organics were
washed with brine (10 mL), dried over Na2SO4, and evaporated. Purifica-
tion by chromatography on a flash purification system (EtOAc/cyclohex-
12786
www.chemeurj.org
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 12778 – 12787

Addition of Electrophilic Radicals to 2-Benzyloxyglycals
FULL PAPER
[6] a) A. Bondi, J. Phys. Chem. 1964, 68, 441–451; b) J. C. Biffinger,
H. W. Kim, S. G. DiMagno, ChemBioChem 2004, 5, 622–627, and
references therein; c) D. OꢅHagan, Chem. Soc. Rev. 2008, 37, 308–
319.
[7] For potent nonhydrolyzable CF2 analogues of bioactive molecules,
see: a) T. R. Burke, Jr, K. Lee, Acc. Chem. Res. 2003, 36, 426–433;
b) T. R. Burke, Jr, M. S. Smyth, A. Otaka, M. Nomizu, P. P. Roller,
G. Wolf, R. Case, S. E. Shoelson, Biochemistry 1994, 33, 6490–6494;
c) G. M. Blackburn, D. L. Jakeman, A. J. Ivory, M. P. Williamson,
Bioorg. Med. Chem. Lett. 1994, 4, 2573–2578; d) T. Nakano, M.
Makino, Y. Morizawa, Y. Matsumura, Angew. Chem. 1996, 108,
1082–1084; Angew. Chem. Int. Ed. Eng. 1996, 35, 1019–1021, and
the references therein.
d) X. Yang, W. Yuan, S. Gu, X. Yang, F. Xiao, Q. Shen, F. Wu, J.
Fluorine Chem. 2007, 128, 540–544; e) E. Godineau, K. Schenk, Y.
Landais, J. Org. Chem. 2008, 73, 6983–6993.
[16] Glycals 2, 3, and 5 were prepared according to a very efficient litera-
ture procedure; see: D. J. Chambers, G. R. Evans, A. J. Fairbanks,
Tetrahedron 2004, 60, 8411–8419.
[17] L. Jean-Baptiste, S. Yemets, R. Legay, T. Lequeux, J. Org. Chem.
2006, 71, 2352–2359.
[18] B. Quiclet-Sire, S. Z. Zard, Chem. Eur. J. 2006, 12, 6002–6016.
[19] The use of DLP at reflux in DCE/tBuOH was not possible with
BrCF2CO2Et and CF2Br2 because of their volatility.
[20] B. B. Snider, Chem. Rev. 1996, 96, 339–363.
[21] a) D. J. Burton, D. M. Wiemers, J. Fluorine Chem. 1981, 18, 573–
582; b) G. K. Surya Prakash, J. Hu, G. A. Olah, J. Org. Chem. 2003,
68, 4457–4463; c) Y. Li, H. Li, J. Hu, Tetrahedron 2009, 65, 478–
483.
[8] For an account of our work, see: E. Leclerc, X. Pannecoucke, C. R.
Chim. 2012, 15, 57–67.
[9] a) J. S. Houlton, W. B. Motherwell, B. C. Ross, M. J. Tozer, D. J. Wil-
liams, A. M. Z. Slawin, Tetrahedron 1993, 49, 8087–8106; b) T. Brig-
aud, O. Lefebvre, R. Plantier-Royon, C. Portella, Tetrahedron Lett.
1996, 37, 6115–6116; c) T. F. Herpin, W. B. Motherwell, J.-M.
Weibel, Chem. Commun. 1997, 923–924; d) H. Berber, T. Brigaud,
O. Lefebvre, R. Plantier-Royon, C. Portella, Chem. Eur. J. 2001, 7,
903–909; e) A. Wegert, R. Miethchen, M. Hein, H. Reinke, Synthe-
sis 2005, 1850–1858; f) A. Wegert, M. Hein, H. Reinke, N. Hoff-
mann, R. Miethchen, Carbohydr. Res. 2006, 341, 2641–2652; g) J.
Picard, N. Lubin-Germain, J. Uziel, J. Augꢀ, Synthesis 2006, 979–
982; h) G. Hirai, T. Watanabe, K. Yamaguchi, T. Miyagi, M. Sodeo-
ka, J. Am. Chem. Soc. 2007, 129, 15420–15421; i) K. A. Tony, R. W.
Denton, A. Dilhas, J. Jimꢀnez-Barbero, D. R. Mootoo, Org. Lett.
2007, 9, 1441–1444; j) R. W. Denton, K. A. Tony, J. J. Hernandez-
Gay, F. J. Canada, J. Jimꢀnez-Barbero, D. R. Mootoo, Carbohydr.
Res. 2007, 342, 1624–1635; k) M. Kolympadi, M. Fontanella, C. Ven-
turi, S. Andrꢀ, H.-J. Gabius, J. Jimenez-Barbero, P. Vogel, Chem.
Eur. J. 2009, 15, 2861–2873; l) J. L. Chaytor, R. N. Ben, Bioorg.
Med. Chem. Lett. 2010, 20, 5251–5254.
[22] L. A. Buttle, W. B. Motherwell, Tetrahedron Lett. 1994, 35, 3995–
3998.
[23] a) F. De Vleeschouwer, V. Van Speybroeck, M. Waroquier, P. Geerl-
ings, F. De Proft, Org. Lett. 2007, 9, 1441–1444; b) F. De Vlee-
schouwer, V. Van Speybroeck, M. Waroquier, P. Geerlings, F. De
Proft, J. Org. Chem. 2008, 73, 9109–9120.
[24] M. G. Roepel, Tetrahedron Lett. 2002, 43, 1973–1976.
[25] a) N. A. LeBel, R. F. Czaja, A. DeBoer, J. Org. Chem. 1969, 34,
3112–3126; b) A. L. J. Beckwith, Tetrahedron 1981, 37, 3073–3100.
[26] a) F. W. Lichtenthaler, M. Lergenmꢄller, S. Peters, Z. Varga, Tetra-
hedron: Asymmetry 2003, 14, 727–736; b) F. W. Lichtenthaler, M.
Lergenmꢄller, S. Schwidetzky, Eur. J. Org. Chem. 2003, 3094–3103.
[27] L. Cipolla, B. La Ferla, L. Lay, F. Peri, F. Nicotra, Tetrahedron:
Asymmetry 2000, 11, 295–303.
[28] C. F. de Graauw, J. A. Peters, H. van Bekkum, J. Huskens, Synthesis
1994, 1007–1017.
[29] A. K. Dilger, V. Gopalsamuthiram, S. D. Burke, J. Am. Chem. Soc.
2007, 129, 16273–16277.
[10] a) B. Moreno, C. Quehen, M. Rose-Hꢀlꢁne, E. Leclerc, J.-C. Quiri-
on, Org. Lett. 2007, 9, 2477–2480; b) S. Colombel, M. Sanselme, E.
Leclerc, J.-C. Quirion, X. Pannecoucke, Chem. Eur. J. 2011, 17,
5238–5241.
[30] a) J. Lamar, J. Hu, A. B. Bueno, H.-C. Yang, D. Guo, J. D. Copp, J.
McGee, B. Gitter, D. Timm, P. May, J. McCarthy, S.-H. Chen,
Bioorg. Med. Chem. Lett. 2004, 14, 239–243; b) Z. Hu, W. Han, Tet-
rahedron Lett. 2008, 49, 901–902.
[11] a) W. R. Dolbier, Jr, Chem. Rev. 1996, 96, 1557–1584; b) P. Kirsch
in Modern Fluoroorganic Chemistry, Wiley-VCH, Weinheim, 2004,
pp. 91–101 c) J. Hu, W. Zhang, F. Wang, Chem. Commun. 2009,
7465–7478.
[12] a) R. Miethchen, M. Hein, H. Reinke, Eur. J. Org. Chem. 1998,
919–923; b) A. Wegert, H. Reinke, R. Miethchen, Carbohydr. Res.
2004, 339, 1833–1837.
[31] For a recent review on lithium carbenoids, see: V. Capriati, S.
Florio, Chem. Eur. J. 2010, 16, 4152–4162.
[32] This widely described and useful rearrangement for sp2 carbenoids
is much less common in the sp3 series; see: a) M. Braun in The
Chemistry of Organolithium Compounds (Eds.: Z. Rappoport, I.
Marek), Wiley, Chichester, 2004, pp. 829–900; b) E. Jahnke, R. R.
Tykwinski, Chem. Commun. 2010, 46, 3235–3249.
[13] a) T. Linker, T. Sommermann, F. Kahlenberg, J. Am. Chem. Soc.
1997, 119, 9377–9384; b) T. Linker, D. Schanzenbach, E. Elamparu-
thi, T. Sommermann, W. Fudickar, V. Gyꢆllai, L. Somsꢇk, W.
Demuth, M. Schmittel, J. Am. Chem. Soc. 2008, 130, 16003–16010.
[14] a) Z.-Y. Yang, D. J. Burton, J. Org. Chem. 1991, 56, 5125–5132;
b) K. Iseki, D. Asada, M. Takahashi, T. Nagai, Y. Kobayashi, Tetra-
hedron Lett. 1994, 35, 7399–7400; c) F. Xiao, F. Wu, Y. Shen, L.
Zhou, J. Fluorine Chem. 2005, 126, 63–67; d) S. Murakami, H. Ishii,
T. Tajima, T. Fuchigami, Tetrahedron 2006, 62, 3761–3769; e) H.
Nagura, S. Murakami, T. Fuchigami, Tetrahedron 2007, 63, 10237–
10245; f) L. Leung, B. Linclau, J. Fluorine Chem. 2008, 129, 986–
990.
[33] For a related cyclization of a difluoromethyllithium species, see:
a) R. S. Timofte, B. Linclau, Org. Lett. 2008, 10, 3673–3676; b) B.
Linclau, A. J. Boydell, R. S. Timofte, K. J. Brown, V. Vinader, A. C.
Weymouth-Wilson, Org. Biomol. Chem. 2009, 7, 803–814.
[34] a) P. Garner, J. M. Park, Org. Synth. 1992, 70, 18–24; b) A. Dondo-
ni, D. Perrone, Org. Synth. 2000, 77, 64–77.
[35] R. Goff, Y. Gao, J. Mattner, D. Zhou, N. Yin, C. Cantu, L. Teyton,
A. Bendelac, P. B. Savage, J. Am. Chem. Soc. 2004, 126, 13602–
13603.
[36] For the relationship between the coupling constant and the relative
configuration of fluoroalkyl-substituted oxazolidinone, see: M.
Omote, Y. Eto, A. Tarui, K. Sato, A. Ando, Tetrahedron: Asymmetry
2007, 18, 2768–2772.
[15] For other radical additions of BrCF2CO2Et, see: a) D. Morel, F.
Dawans, Tetrahedron 1977, 33, 1445–1447; b) K. Sato, M. Omote,
A. Ando, I. Kumadaki, J. Fluorine Chem. 2004, 125, 509–515; c) E.
Godineau, C. Schꢈfer, Y. Landais, Org. Lett. 2006, 8, 4871–4874;
Received: May 30, 2013
Published online: August 16, 2013
Chem. Eur. J. 2013, 19, 12778 – 12787
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
12787