Glucose-based surfactants with hydrogenated, fluorinated, or hemifluorinated tails: Synthesis and comparative physical-chemical characterization

Article abstract of DOI:10.1021/jo801379e

(Chemical Equation Presented) (Hemi)fluorinated hydrophobic chains have been found to minimize the denaturating propensity of surfactants toward membrane proteins. The work reported herein deals with the synthesis of a new series of non-ionic glucose-based surfactants endowed with a hybrid hemifluorocarbon chain. The convergent synthesis is based on a one-pot reduction/alkylation of hemifluorinated thioacetate and glucosylated trishydroxymethyl acrylamidomethane using NaBH₄ in methanol. This "mild" alkylation was studied in order to improve yields and to pass up the use of an excess of commercially unavailable hemifluorinated thiols. The physical-chemical properties in aqueous solution of this novel series were studied by surface tension measurement and dynamic light scattering (DLS), as well as their behavior upon reverse-phase chromatography, and were compared with those of their hydrogenated and perfluorinated analogues. The atypical effect of the additional ethyl tip to the fluorinated chain was demonstrated by higher critical micellar concentration values and abnormal hydrophobicities measured by reverse-phase chromatography. Moreover, according to Israelachvili's concept, DLS studies showed that surfactants bearing bulkier polar head self-assemble into small and well-defined aggregates, suggesting the formation of spherical micelles rather than the cylindrical ones usually observed with classical fluorinated surfactants.

Full text of DOI:10.1021/jo801379e

Glucose-Based Surfactants with Hydrogenated, Fluorinated, or
Hemifluorinated Tails: Synthesis and Comparative
Physical-Chemical Characterization
Maher Abla, Gre´gory Durand,* and Bernard Pucci*
Laboratoire de Chimie Bioorganique et des Syste`mes Mole´culaires Vectoriels, Faculte´ des Sciences,
UniVersite´ d’AVignon et des Pays de Vaucluse, 33 rue Louis Pasteur, F-84000 AVignon, France
gregory.durand@uniV-aVignon.fr; bernard.pucci@uniV-aVignon.fr
ReceiVed July 1, 2008
(Hemi)fluorinated hydrophobic chains have been found to minimize the denaturating propensity of
surfactants toward membrane proteins. The work reported herein deals with the synthesis of a new series
of non-ionic glucose-based surfactants endowed with a hybrid hemifluorocarbon chain. The convergent
synthesis is based on a one-pot reduction/alkylation of hemifluorinated thioacetate and glucosylated
trishydroxymethyl acrylamidomethane using NaBH4 in methanol. This “mild” alkylation was studied in
order to improve yields and to pass up the use of an excess of commercially unavailable hemifluorinated
thiols. The physical-chemical properties in aqueous solution of this novel series were studied by surface
tension measurement and dynamic light scattering (DLS), as well as their behavior upon reverse-phase
chromatography, and were compared with those of their hydrogenated and perfluorinated analogues. The
atypical effect of the additional ethyl tip to the fluorinated chain was demonstrated by higher critical
micellar concentration values and abnormal hydrophobicities measured by reverse-phase chromatography.
Moreover, according to Israelachvili’s concept, DLS studies showed that surfactants bearing bulkier polar
head self-assemble into small and well-defined aggregates, suggesting the formation of spherical micelles
rather than the cylindrical ones usually observed with classical fluorinated surfactants.
Introduction
dissociating effect of detergents can be difficult to control, and
this results in the destabilization and irreversible inactivation
of the protein. Specific interactions between TM segments
belonging to the same or to different polypeptides are essential
to the folding and assembly of MPs, and in many cases, bound
lipids or other hydrophobic cofactors are required for the
stability and function of MPs. It is therefore a delicate task to
stop dissociation at the desired stage, namely, that of a discrete
entity that needs to be purified and studied in vitro while
retaining its biological activity and a modicum of stability. 2-4
The two likely mechanisms leading to inactivation are the
intrusion of the detergent into the TM region of the protein and/
or the dissociation of stabilizing lipids, cofactors, or subunits.2,5
Obtaining membrane proteins (MPs) that are soluble, active,
and stable so as to be able to study in vitro their function and
structure is of major scientific and biomedical importance, given
that MPs represent more than half of the targets of pharmaceuti-
cal drugs and 20-40% of proteins encoded by genomes.1
However, there is currently a dearth of information on their
function and structure (less than 1% of the entries in the Protein
Data Bank are of MPs). MPs, which may comprise either one
or several polypeptide chains (subunits) and hydrophobic
cofactors, need to be extracted from the membrane for most in
vitro studies. This is commonly done by using detergents, which
bind, in place of the lipids, to the transmembrane (TM) surface
of the proteins, making them soluble in water. Detergents are
dissociating compounds, a specificity that allows them to
solubilize membranes, to compete with lipid-lipid and lipid-
protein interactions, and therefore to extract MPs. However, the
(2) Bowie, J. U. Curr. Opin. Struct. Biol. 2001, 11, 397–402.
(3) Garavito, R. M.; Ferguson-Miller, S. J. Biol. Chem. 2001, 276, 32403–
32406.
(4) Gohon, Y.; Popot, J.-L. Curr. Opin. Colloid Interface Sci. 2003, 8, 15–
22.
(5) Breyton, C.; Tribet, C.; Olive, J.; Dubacq, J.-P.; Popot, J.-L. J. Biol. Chem.
1997, 272, 21892–21900.
(1) Wallin, E.; Von-Heijne, G. Protein Sci. 1998, 7, 1029–1038.
8142 J. Org. Chem. 2008, 73, 8142–8153
10.1021/jo801379e CCC: $40.75 2008 American Chemical Society
Published on Web 09/30/2008

Glucose-Based Surfactants
biochemical stabilities.13 Recently, it has been demonstrated that
HFSs can be used for in vitro synthesis of MPs or their insertion
into a lipid bilayer.14
However, because the synthesis of the polar head of HF-
TAC involves radical polymerization and leads to inevitable
polydispersity, batch-to-batch variations can be observed. In
order to circumvent this problem, we have thus synthesized
hemifluorinated surfactants with chemically defined polar heads
derived either from aminoxide15 or lactose.16 (Figure 1) The
hemifluorinated zwitterionic aminoxide (HF-AO) was found to
significantly decrease MPs stability, while hemifluorinated
lactobionamide derivative (HF-Lac) was found to be much more
efficient in keeping several MPs water-soluble and active.
Interestingly, we underlined the inactivation character of the
hydrogenated dodecyl homologue of HF-Lac, H12-Lac, com-
pared to dodecyl maltoside (DDM). This shows that the
lactobionamide head and thus its galactose moiety are not
particularly favorable to protein stability and leaves open the
perspective of optimizing tail/head combinations so as to obtain
yet milder fluorinated or hemifluorinated surfactants.
Thus we synthesized and examined the hemifluorinated (HF-
Malt) analogue of this widely used detergent DDM.17 The
integrity of bacteriorhodopsine and cytochrom b6f complex, as
protein tests solubilized in this fluorinated surfactant, was
investigated, and these two MPs were found in their native state
and as active as in DDM. Their stability over time was either
comparable to or higher than that in DDM. However, HF-Malt
form large and polydisperse aggregates, which could be rod-
like, disk, or cylindrical assemblies. Such aggregate formation
seems to affect the nature of their complexes with MPs, causing
a serious drawback for further studies. Indeed, having prepara-
tion homogeneous in size is a prerequisite for any functional
and structural study of the protein: manipulating membrane
proteins can lead to aggregation, dissociation, oligomerization,
etc., and it is thus important to be able to check the homogeneity
of the protein. If the heterogeneity of the sample is induced by
the surfactant, we lose a precious means of monitoring the
oligomeric state of the protein. Furthermore, no structural studies
can be undertaken in these conditions.
Therefore, these first results show that the use of fluorinated
tails holds for various types of polar heads and also stress the
impact of the volumetric ratio between the polar head and the
hydrophobic tail of the FS on the micellar structure formed.
Indeed, when comparing the chemical structures and physical-
chemical properties of compounds of the HF-TAC, HF-Lac, and
HF-Malt series and keeping in mind Israelachvili’s concept18
linking the nature of the aggregate formed by a surfactant to
the volumetric ratio between its polar head and its hydrophobic
tail, one is led to conclude that the introduction of a bulkier
polar head for (hemi)fluorinated surfactants should induce an
increase of the interfacial curvature of the aggregates and thus
favor the formation of small and monodisperse micelles rather
FIGURE 1. Chemical structures of previously developed hemifluori-
nated surfactants.
In an attempt at overcoming these problems, we have
undertaken the charge to study the potentiality of fluorocarbon
surfactants (FS).6,7 These molecules have the same general
structure as classical detergents (i.e., a hydrophilic headgroup
and a hydrophobic tail), but the hydrophobic tail, rather than
being a hydrogenated aliphatic chain, is a fluorinated chain. Our
rationale was the following: alkanes and perfluorinated alkanes,
while they are both hydrophobic, are poorly miscible; for this
reason,surfactantswithfluorinatedalkylchainsarelyophobic-they
do not partition well into biological membranes and therefore
have little cytolytic effect.7-10 For the same reason, they are
poor solvents for lipids and hydrophobic cofactors and can be
expected to be less delipidating. Their hydrophobic moieties
being lyophobic, bulkier, and more rigid than their hydrogenated
counterparts, they may also be expected to intrude less easily
into the protein structure itself. However, by the same token,
perfluorinated chains have little affinity for the hydrogenated
TM surface of MPs. To improve the interaction between the
hydrophobic, hydrogenated domain of integral membrane
proteins and the surfactant, an ethyl hydrogenated tip has been
grafted at the extremity of the surfactant tail, leading to
hemifluorinated surfactants (HFSs).11
Among them, surfactants with a polymeric head derived from
trishydroxymethyl acrylamidomethane (THAM), the HF-TAC
family (Figure 1), proved to be particularly mild toward MPs.12
MPs, initially extracted using hydrogenated detergents, remained
soluble and active when transferred into HFSs with increased
(6) Pucci, B.; Maurizis, J. C.; Pavia, A. A. Eur. Polym. J. 1991, 27, 1101–
1106.
(7) Chabaud, E.; Barthelemy, P.; Mora, N.; Popot, J. L.; Pucci, B. Biochimie
1998, 80, 515–530.
(8) Barthelemy, P.; Tomao, V.; Selb, J.; Chaudier, Y.; Pucci, B. Langmuir
2002, 18, 2557–2563.
(9) Mukerjee, P. Colloids Surf., A 1994, 84, 1–10.
(10) Nakano, T. Y.; Sugihara, G.; Nakashima, T.; Yu, S. C. Langmuir 2002,
18, 8777–8785.
(13) Park, K. H.; Berrier, C.; Lebaupain, F.; Pucci, B.; Popot, J.-L.; Ghazi,
A.; Zito, F. Biochem. J. 2007, 403, 183–187.
(14) Palchevskyy, S. S.; Posokhov, Y. O.; Olivier, B.; Popot, J. L.; Pucci,
B.; Ladokhin, A. S. Biochemistry 2006, 45, 2629–2635.
(15) Chaudier, Y.; Zito, F.; Barthelemy, P.; Stroebel, D.; Ameduri, B.; Popot,
J.-L.; Pucci, B. Bioorg. Med. Chem. Lett. 2002, 12, 1587–1590.
(16) Lebaupain, F.; Salvay, A. G.; Olivier, B.; Durand, G.; Fabiano, A. S.;
Michel, N.; Popot, J. L.; Ebel, C.; Breyton, C.; Pucci, B. Langmuir 2006, 22,
8881–8890.
(11) Barthelemy, P.; Ameduri, B.; Chabaud, E.; Popot, J. L.; Pucci, B. Org.
Lett. 1999, 1, 1689–1692.
(12) Breyton, C.; Chabaud, E.; Chaudier, Y.; Pucci, B.; Popot, J. L. FEBS
Lett. 2004, 564, 312–318.
(17) Polidori, A.; Presset, M.; Lebaupain, F.; Ameduri, B.; Popot, J.-L.;
Breyton, C.; Pucci, B. Bioorg. Med. Chem. Lett. 2006, 16, 5827–5831.
(18) Israelachvili, J. M., D.; Ninham, B. J. Biochim. Biophys. Acta 1977,
470, 185–201.
J. Org. Chem. Vol. 73, No. 21, 2008 8143

Abla et al.
SCHEME 1. Retrosynthetic Routes for Hemifluorinated
Surfactants
SCHEME 2. Synthesis of THAM Derivatives
than rod-like or cylindrical micelles. Smaller and well-defined
surfactant aggregates should lead to monodisperse MP/surfactant
complexes, while preserving if not improving the useful
biochemical properties exhibited thus far by fluorinated surfactants.
This led us to design a new class of surfactants that would
make it possible to easily tune the polar head/hydrophobic tail
volumetric ratio of the surfactant in order to study its effect on
the physical-chemical properties. We chose THAM as the basic
structure for the polar head synthesis where one, two, or three
glucose moieties, respectively, were grafted on the hydroxyl
groups. The hemifluorinated hydrophobic chain was synthesized
from 1,10-diiodo-3,3,4,4,5,5,6,6,7,7,8,8-dodefluorodecane and
was linked to the polar head through a thioether bond leading
to a new series of hemifluorinated surfactants having different
volumetric ratios. For the sake of comparison, perfluorinated
and hydrogenated analogues were also synthesized following
the same synthetic route. The comparative physical-chemical
properties in aqueous solution were studied by surface tension
measurement and dynamic light scattering (DLS), as well as
their behavior upon reverse-phase chromatography.
selective protections of the hydroxyl groups. Two hydroxyl
groups of THAM1 were protected by reaction with the 2,2-
dimethoxypropane as previously published20 to lead after
purification to compound 1. Grafting of the ꢀ-D-glucose was
performed following the Helferich’s method21 using tetra-O-
acetyl-D-glucopyranosyl bromide, previously prepared according
to the Rosevear’s method.22 Then, hydrolysis of the 1,3-
acetonide in the presence of Montmorillonite K10 led to
monoglucosylated compound 2a. Acetylation of 1 followed by
removal of 1,3-acetonide under acidic condition led to com-
pound 3. Diglucosylation of compound 3 was performed in the
presence of an excess of tetra-O-acetyl-D-glucopyranosyl bro-
mide under ultrasonic activation leading to compound 4a. We
previously reported that the synthesis of THAM galactosylated
derivatives using the Helferich method was improved under
ultrasounds conditions.23 Furthermore, the application of ultra-
sound was found to increase the solubilization of the reagents
and the reaction was almost immediate. Finally, THAM bearing
three ꢀ-D-glucose moieties 5a was obtained by a direct gluco-
sylation of THAM under ultrasounds activation. As it was
expected, the synthesis of the tetra-O-acetyl-glucosylated tris
polar head was achieved in good overall yields from THAM of
38%, 27%, and 50% for compounds 2a, 4a, and 5a, respectively.
Finally, removal of the acetyl protective groups was performed
under Zemple´n24 conditions and afforded compounds 2, 4, and
5 in quantitative yields.
Results
Synthesis. Our convergent synthetic strategy relies on the
formation of the C-S bond of the surfactant. This thioether
bond links the polar headgroup to the hydrophobic chain, and
its formation is considered as the key step in this synthesis.
Retrosynthetically, such a bond can be disconnected at two
positions (Scheme 1). The first strategy involves the condensa-
tion of a hemifluorinated thiol onto an acrylamide-type polar
head, which is easily obtained from THAM. Hemifluorinated
thiol, structurally made of a fluorinated core end-capped with
an ethyl and a thiol ethyl groups, is prepared from the
monoiodide hemifluorinated chain. On the other hand, the
second strategy involves the condensation of a thiol-derived
polar head on the hemifluorinated monoiodide compound.
Condensation of mercaptopropionic acid on tris(hydroxymethyl)-
aminomethane (TRIS) followed by selective glucosylation led
to the thiol-derived polar head.
Synthesis of Hemifluorinated Thioacetate. The second part
of the synthesis represents the major task. It consists of the
preparation of the nonsymmetrical hemifluorinated thiol made
of a fluorinated core and bordered by an ethyl and a thiol ethyl
groups. As previously described, the bis monoethylenation of
diiodoperfluorohexane25 led to 1,10-diiodo 3,3,4,4,5,5,6,6,7,7,8,8
(20) Polidori, A.; Braun, O.; Mora, N.; Pucci, B. Tetrahedron Lett. 1997,
38, 2475–2478.
(21) Helferich, B.; Zimmer, J. Chem. Ber. 1962, 95, 2604–2611.
(22) Rosvear, P.; Vanken, T.; Baxter, J.; Miller, S. F. Biochemistry 1980,
19, 4108–4115.
(23) Polidori, A.; Pucci, B.; Maurizis, J. C.; Pavia, A. A. New J. Chem. 1994,
18, 839–848.
(1) First Strategy. Synthesis of THAM-Derived Polar
Head. The synthesis of THAM glucosylated derivatives required
(24) Zemplen, G. Chem. Ber. 1927, 60, 1555–1564.
(25) Manseri, A.; Ameduri, B.; Boutevin, B.; Kotora, M.; Hajek, M.;
Caporiccio, G. J. Fluorine Chem. 1995, 73, 151–158.
(19) Jedlinsky, J.; Paprotivy, J. Rocz. Chem. 1966, 40, 1487–1491. Pucci,
B.; Maurizis, J. C.; Pavia, A. A. Eur. Polym. J. 1991, 27, 1101–1106.
8144 J. Org. Chem. Vol. 73, No. 21, 2008

Glucose-Based Surfactants
SCHEME 3. Synthesis of Hemifluorinated Thioacetate
radical condensation of the hemifluorinated thiol and monoglu-
cosylated THAM 2a in equimolar conditions. Unfortunately only
a low yield was obtained in such condition (9%). So instead of
using thiol consuming radical condensation, we chose to use
the Michael acceptor character of acrylamide to react with
nucleophilic hemifluorinated thiol.
Michael reaction is commonly carried out using acid cata-
lysts32 or Lewis acids33 or under alkaline conditions.34 While
the acidic mediums were found to give satisfactory results, they
cannot be applied in the presence of acetal functions such as
O-glucosyl. On the other hand, thiols are known to be sensitive
to basic media as a result of the formation of disulfide bonds
leading to a decrease of the yield. Moreover, as stated above,
due to the acidic character of the hydrogen vicinal to the
fluorinated core, fluorinated thiols are highly vulnerable to bases.
Thus, performing the Michael reaction in basic medium seemed
to be a limited solution. Fall et al.35 pointed out the ability of
the one-step reduction/alkylation of thioacetate compounds,
using NaOH and NaBH4 in EtOH, to afford the corresponding
thioether in very good yield.
Despite the fact that our work is essentially devoted to the
synthesis of the hemifluorinated surfactants and taking into
consideration the difficulties concerning the synthesis of the
hemifluorinated thioacetate, we performed the first attempts
using the perfluorinated one (C6F13CH2CH2SCOCH3), which
was prepared from the commercially available C6F13CH2CH2I
in one step with 75% yield (see Supporting Information for
further details). A first attempt using 1 equiv of perfluorinated
thioacetate and compound 2 using the method of Fall et al. led
to the desired compound 9 but in a low yield (9%). Compared
to the high yields reported by these authors, we assumed that
basic conditions (NaOH) might be responsible to the fluorinated
thiol degradation. Furthermore the use of other bases such as
TEA or KOH did not improve significantly the yield (20-24%).
Therefore we applied the same experimental conditions as
previously reported, including some modifications in the process:
(i) the use of NaBH4 without any additional base and (ii) the
use of methanol as solvent. In these conditions the yield was
improved and compound 9 was isolated with 60% yield (Scheme
4).
Thus, the condensation of perfluorothioacetate on mono-, di-,
or triglucosylated THAM derivatives 2, 4, or 5 provided the
desired compounds 9-11 with satisfactory yields ranging
between 52% and 60%. In comparison to the yield obtained
using Fall’s method and those obtained with different bases,
these results encouraged us to apply this method to the
hemifluorinated series. Even with hemifluorinated compounds
the yields were in the same range (44% and 60%, respectively,
for di- and triglucosylated THAM). Finally, to extend our
physical-chemical investigation, hydrogenated analogues were
also synthesized following the same synthetic route. However
the synthesis of the hydrogenated analogues was directly carried
out using the commercially available decanethiol and dode-
canethiol, demonstrating the extent of this procedure.
dodecafluorodecane 6, which was chosen as starting material
for the synthesis of the commercially unavailable hemifluori-
nated thiol11 (Scheme 3).
The first step involves a selective monoreduction of one iodide
using HSi(SiMe3)3, in the presence of a catalytic amount of
AIBN as radical initiator, affording after purification on silica
gel pure 1-iodo 3,3,4,4,5,5,6,6,7,7,8,8 dodecafluorodecane 7
(30%), pure unreacted diiodo compound 6 (40%), and a small
amount of bis reduced compound (10%). The remaining diiodo
compound 6 can be recovered and reused to complete the
reaction. It should be noted that the use of HSi(SiMe3)3 was
preferred to Bu3SnH. The use of the latter reducing agent suffers
from several drawbacks such as the well-known toxicity26 of
the tributyltin-containing compounds, as well as the difficulties
concerning the purification of such tin derivatives.
The second step in the synthesis of the hemifluorinated part
consists in the conversion of the iodide into a thiol. A rapid
search in the literature shows that monohalide compounds react
with thiourea27 to afford after hydrolysis the corresponding thiol;
however, some previous experiments have revealed the weak
reproducibility of this method (yields ranging between 17% and
90%).28,29 Moreover, some other alternative routes to generate
perhydrogenated or perfluorinated alkanethiols from their cor-
responding halides, using either thiolacetic acid30 or potassium
thioacetate,31 have been largely described in the literature. In
our case, monoiodide compound reacted with potassium thio-
acetate (1.5 equiv) in anhydrous DMF and afforded the desired
compound 8 (56%). It has to be underlined that the use of a
larger excess of potassium thioacetate led to secondary reactions
(possibly elimination reactions due to the acidity of the
methylene linked to perfluoroalkyl moiety) and subsequently
decreased the yield to 30%. In addition to the good yield of
formation of the thioacetate derivatives, the presence of the
acetyl as protecting group of thiol is at the same time appreciated
since thiols are extremely vulnerable to oxygen and can be easily
oxidized in sulfoxide in the presence of very low traces of O2.
Furthermore, thioacetate can be quickly converted into corre-
sponding thiol under the Zemple´n24 conditions.
Thioether Bond Formation. The synthesis of hemifluori-
nated thioether galactosylated surfactants using radical conden-
sation has already been published.23 However, radical conden-
sations of alkane thiol onto acrylamide require a large excess
of thiol (3 equiv) to get satisfactory yields. Taking into account
the difficulties to synthesize hemifluorinated thiol, the radical
condensation presents some limitations concerning large scale
synthesis. In a first attempt at overcoming this problem, we used
All of the amphiphilic compounds were purified by Sephadex
LH-20 size exclusion chromatography and most of them by C18
(26) Boyer, I. J. Toxicology 1989, 55, 253–298.
(27) Brace, N. O. U.S. Patent 3,172,910, 1965.
(28) Chaudier, Y., Thesis, Universite´ d’Avignon et des Pays de Vaucluse,
France, 2002.
(29) Olivier, B., Thesis, Universite´ d’Avignon et des Pays de Vaucluse,
France, 2004.
(30) Naud, C.; Calas, P.; Blancou, H.; Commeyras, A. J. Fluorine Chem.
2000, 104, 173–183.
(31) Hong, H. G.; Park, W. Langmuir 2001, 17, 2485–2492.
(32) Firouzabadi, H.; Iranpoor, N.; Jafarpour, M.; Ghaderi, A. J. Mol. Catal.,
A 2006, 249, 98–102.
(33) Alam, M. M.; Varala, R.; Adapa, S. R. Tetrahedron Lett. 2003, 44,
5115–5119.
(34) Martinelli, M. J.; Vaidyanathan, R.; Khau, V. V.; Staszak, M. A.
Tetrahedron Lett. 2002, 43, 3365–3367.
(35) Fall, Y.; Diouf, O.; Gomez, G.; Bolano, T. Tetrahedron Lett. 2003, 44,
6069–6072.
J. Org. Chem. Vol. 73, No. 21, 2008 8145

Abla et al.
SCHEME 5. Synthesis of Thiol-Derived Polar Head
SCHEME 4. Thioether Bond Formation
activation we observed that the thioether bond was partially
cleaved, suggesting that even mild Lewis acid catalysis such
as Hg(CN)2 are able to deprotect trityl groups. Furthermore, in
addition to the consumption of Hg(CN)2 in the deprotection
reaction, the formation of the thiol at this step leads to other
secondary thioglucosylation reactions. Currently, the use of
acetyl group as protective agent of mercapto propionic acid
derivatives and their one-pot mild reduction/alkylation with
1-iodo-3,3,4,4,5,5,6,6,7,7,8,8-dodecafluorodecane 7 in the pres-
ence of NaBH4 is under investigation as an alternative synthetic
route.
Physical-Chemical Characterization. All fluorinated sur-
factants were labeled by F6, according to the number of
fluorinated carbons of the hydrophobic chain followed by
Mono-, Di-, or Triglu according to the nature and number of
sugar moieties grafted onto the polar head. In order to dif-
ferentiate perfluorinated surfactants from their hemifluorinated
analogues, “H2” was added to the given name of these latter
surfactants, indicating the presence of an ethyl tip. Perhydro-
genated surfactants were labeled by Hn, where n indicates the
number of the hydrogenated carbons in the corresponding thiol
starting material, followed by the nature of the polar head as
described for fluorinated surfactants. All surfactants were very
soluble in water up to at least 4-5 mM at room temperature,
and the solutions were very clear (4-5 mM is the maximum
concentration used for the physical-chemical assays; however,
it does not reveal the maximum solubility of the surfactants).
Only H12-Diglu and H12-Triglu at 5 mM were subjected to
heating at 50 °C for 10-20 min in order to complete solubi-
lization and then kept at room temperature at least 12 h before
any physical-chemical measurement. In these conditions, both
H12-surfactant solutions remain clear after cooling at room
temperature and no precipitation was observed.
a Reaction performed following Fall et al.35 experimental conditions.
b See the physical-chemical characterization section below for further details.
reverse-phase HPLC and then lyophilized to give pure surfac-
1
tants as white powder. They were fully characterized by H,
13C, 19F, DEPT, and HMQC NMR experiments as well as high
resolution mass spectrometry, allowing the observation of
characteristic adducts (see Supporting Information). The purity
of the samples was checked by RP-HPLC and was higher than
98%.
In the aim to study the role of the methanol in such one-pot
alkylation, two experiments using fluorinated thioacetate
(C6F13CH2CH2SCOCH3) and commercially available benzyl
bromide in either methanol or acetonitrile as solvent were
performed in the presence of NaBH4 (1.5 equiv). Whereas the
reaction in methanol afforded the desired compound in a very
good yield (91%), only unreacted starting materials were
recovered after 48 h of reflux in acetonitrile. Furthermore, we
found that the reaction of C6F13CH2CH2SH with benzyl bromide
in acetonitrile afforded the desired compound (86%). Finally,
the reaction of C6F13CH2CH2SH with benzyl bromide in
methanol without using NaBH4 did not lead to the thioether
bond formation. These latter experiments clearly demonstrate
that the methanol in the presence of NaBH4 is capable of
transesterification of the thioacetate group, yielding therefore
to the thiol, which is subsequently converted in thiolate due to
the excess of NaBH4.
(2) Second Strategy. Despite the success of the first strategy,
we also investigated an alternative synthetic route using
3-mercaptopropionic acid and TRIS as starting materials. The
thiol group was first protected by reaction with triphenylmeth-
ylchloride in dichloromethane, and then TRIS was grafted
through an amide bond onto the protected 3-mercaptopropionic
acid in the presence of 2-ethoxy-1-ethoxycarbonyl-1,2-dihyd-
roquinoleine (EEDQ) in refluxing ethanol.36 Nevertheless,
glycosylation of the hydroxyl groups under the same conditions
used in the first strategy using Hg(CN)2 was not possible. The
triphenylmethyl group as well as most of the commonly used
thiol protecting groups are usually cleavable in the presence of
strong acids. During the glucosylation reaction under ultrasound
(a) Surface Tension Measurement. The surface tension data
are summarized in Table 1, and the curves for diglucosylated
surfactants are represented in Figure 2. From this technique we
determined different parameters: (i) the critical micellar con-
centration (CMC); (ii) the surface tension attained at the CMC
(γCMC); (iii) the concentration of surfactant in the bulk phase
that produces a reduction of 20 mN/m in the surface tension of
the solvent (C20) and its corresponding logarithm pC20, which
measure the adsorption efficiency; (iv) the surface excess
concentration at the CMC (Γmax), which measures the maximum
concentration of adsorbed species; (v) the minimum area
occupied by a surfactant at the air/water interface at the CMC
when the surface is saturated (Amin); and (vi) the ratio CMC/
C20 measuring the relative effect of structural or microenvi-
ronemental factor on micellization and on adsorption.
As frequently reported,37,38 for a given polar head the addition
of two methylene to the hydrocarbon chain of a surfactant
usually decreases 5-15 times the CMC. Such a rule is followed
(36) Sharma, K. S.; Durand, G.; Giusti, F.; Olivier, B.; Fabiano, A.-S.;
Bazzacco, P.; Dahmane, T.; Ebel, C.; Popot, J.-L.; Pucci, B. LangmuirAccepted
for publication.
8146 J. Org. Chem. Vol. 73, No. 21, 2008

Glucose-Based Surfactants
TABLE 1. Surface Active Dataa
b
d
surfactant
CMC (mM)
γCMC (mN/m)
Amin (Å2)
Γmax 10-12 (mol/mm2)b
C20 10-3 (mM)c
pC20
CMC/C20
log k′W
F6-Monoglu
F6-Diglu
F6-Triglu
H2F6-Diglu
H2F6-Triglu
H12-Diglu
H12-Triglu
H10-Diglu
H10-Triglu
0.113 ( 0.006
0.233 ( 0.012
0.947 ( 0.042
0.347 ( 0.029
0.747 ( 0.084
0.039 ( 0.001
0.102 ( 0.007
0.555 ( 0.005
1.175 ( 0.037
18.3 ( 0.6
28.3 ( 0.6
32.3 ( 0.6
36.0 ( 1.0
36.0 ( 3.6
40.7 ( 2.3
41.5 ( 0.7
36.0 ( 1.4
38.5 ( 2.1
52.7 ( 1.5
97.5 ( 2.1
114.0e
104.5 ( 2.5
131.5 ( 14.8
102.7 ( 10.1
125.0 ( 17.0
92.0 ( 17.0
107.7 ( 6.4
3.13 ( 0.07
1.79 ( 0.14
1.46e
1.50 ( 0.13
1.47 ( 0.57
1.61 ( 0.17
1.34 ( 0.18
1.84 ( 0.33
1.55 ( 0.09
1.37 ( 0.06
1.54 ( 0.37
4.60e
5.4 ( 2.7
6.6 ( 1.6
3.4 ( 0.4
5.00 ( 0.17
17.5 ( 6.7
22.5 ( 15.4
2.86
2.81
2.34
2.27
2.18
2.47
2.30
1.76
1.65
82.8
151.8
205.9
64.3
112.8
11.5
20.4
31.8
52.3
5.2
4.8
4.7
4.9
4.7
5.6
5.4
4.4
4.2
b
a Data presented are the average of two or three experiments except when noted otherwise. Amin and Γmax were estimated from the slope of the
c
surface tension curve. C20 values were obtained by extrapolation of the surface tension versus log C to 52 mN/m. d pC20 ) -log C20. e Data from one
experiment only.
ethyl-ended fluorinated surfactants. It has also to be underlined
that an increase of the polar head size (demonstrated by an
evolution of the surface area per molecule at the air/water
interface, Amin) induces a significant increase of the limit surface
tension for F6 compounds, as observed with F6-Monoglu and
F6-Diglu. Such a phenomenon could be explained by an increase
of the gap between fluorocarbon chains at the air/water interface.
The results for Γmax and the surface area per molecule Amin
were estimated from the slope of the surface tension curves.
Amin gives information about the space that every molecule needs
to accommodate at the air/water interface. For a given polar
head (Di- or Triglu, for instance), the length and the nature of
the hydrophobic chain have almost no effect on Γmax, whereas
for a given chain the surface concentration Γmax decreases as
the size of the headgroup increases. For instance, F6-Monoglu
has the highest surface concentration and this value significantly
decreases as the number of glucose increases. An opposite trend
is obviously observed regarding Amin. When comparing the
values obtained for the hemifluorinated derivatives with those
of the perfluorinated ones, it appears that the occupied area per
molecule is slightly higher when adding the ethyl tip, as we
previously observed,16,17 and subsequently the surface concen-
tration is slightly lower.
FIGURE 2. Surface tensions versus log CMC plot for H10-Diglu (]),
H12-Diglu (0), H2F6-Diglu (×), and F6-Diglu (∆).
by hydrogenated di- and triglucosylated surfactants. However,
hemifluorinated surfactants do not follow the same rule. H2F6-
Diglu, despite the additional ethyl tail, exhibits a CMC (0.347
mM) close to that of the F6-Diglu (0.233 mM). A similar trend
was observed with F6-Triglu and H2F6-Triglu (0.947 and 0.730
mM, respectively), which is in perfect agreement with our
previous observations.11,16 Fluorocarbon surfactants are known
to exhibit CMCs lower than those of their hydrocarbon
analogues.39,40 On an other hand, we noted herein that increasing
the number of glucose moieties induces an increase of the CMC.
Thus, the CMC of F6-Triglu is ∼8 times higher than that for
F6-Monoglu; nevertheless the CMC of F6-Diglu is only 2 times
higher than that measured for F6-Monoglu. Finally, whatever
the nature of the hydrophobic chain, the general trend is that
the CMC for Triglu derivatives is 2-4 times higher than that
of the Diglu compounds.
Limit surface tensions (γCMC) of surfactants are known to
decrease with increasing fluorine content of the hydrophobic
chain, the result of a higher order of packing of the fluorocarbon
layer at the air/water interface.40 In agreement with this
observation H10- and H12-Diglu and Triglu compounds exhibited
γCMC values higher than those of F6 derivatives. For a given
polar head, γCMC for the hemifluorinated surfactants is between
those of the hydrogenated and perfluorinated surfactants as
previously observed16 confirming the particular properties of
The efficiency of the surfactant is defined by the value of
the negative logarithm of the bulk concentration necessary to
reduce its surface tension by 20 mN/m and is designed by pC20.
The values listed in Table 1 clearly show that the efficiency of
a hydrocarbon or fluorocarbon surfactant increases by increasing
the length of the hydrophobic chain and decreases by increasing
the volume of the polar head. However, the efficiency of H2F6-
surfactants do not follow the same trend as we can observe with
Diglu-surfactants where the value of pC20 of the F6-surfactant
drops from 2.81 to 2.27 with the addition of the ethyl tip (similar
behavior was observed for Triglu-surfactants). Finally, we
determined the CMC/C20 ratio.41 This ratio gives information
about which of the micellization or the adsorption phenomenon
at the air/water interface is favored. Usually, an increase in the
CMC/C20 ratio indicates that adsorption phenomenon is more
facilitated than micellization (or micellization more inhibited
than adsorption). Table 1 shows that CMC/C20 ratio increases
when replacing the hydrogen chain by a fluorinated chain and/
or with the introduction of a bulkier hydrophilic head. However,
for a given polar head, it decreases with increasing the
hydrogenated chain length.
(b) Dynamic Light Scattering. All measurements were taken
at a concentration at least 4 times above the CMC of the
(37) Zhang, T.; Marchant, R. J. Colloid Interface Sci. 1996, 177, 419–426.
(38) Shinoda, K.; Hato, M.; Hayashi, T. J. Phys. Chem. 1972, 76, 909–914.
(39) Sadtler, V. M.; Guilleri, F.; Krafft, M. P.; Riess, J. G. Chem. Eur. J.
1998, 4, 1952–1956.
(40) Kissa, E. Fluorinated Surfactants: Synthesis, Properties, Applications;
Dekker: New York, 1994; Vol. 7.
(41) Rosen, M. J. Surfactants and Interfacial Phenomenon; John-Wiley: New
York, 1989.
J. Org. Chem. Vol. 73, No. 21, 2008 8147

Abla et al.
TABLE 2. Size of Aggregates in Aqueous Solutions
of a hemifluorinated thioacetate (using NaBH4 in methanol) and
THAM derivatives. The interest of this procedure and the
modifications we brought about lie in three major points: (i)
the mildness of this Michael addition that avoids using bases
that can be aggressive toward thiol such as fluorinated ones;
(ii) the role of NaBH4/MeOH, which can avoid secondary
reactions such as oxidation or disulfide bond formation and lead
to the formation of the thiolate in situ; (iii) finally, the efficient
thioether bond formation that can be performed with a slight
excess (1.2-1.5 equiv) of acryloyl monomer, allowing us to
circumvent the necessity of a large excess (3 equiv) of thiol as
required in radical condensations.23 Following the same strategy,
the syntheses of the perfluorinated and hydrogenated analogues
were also achieved in satisfactory yields.
a
surfactant
DH (nm)
% volb
HHWc (nm)
concn (mM)
F6-Monoglu
F6-Diglu
F6-Triglu
H2F6-Diglu
H2F6-Triglu
H12-Diglu
H12-Triglu
H10-Diglu
H10-Triglu
25.3
5.8
5.1
6.5
5.1
5.9
5.2
5.4
4.5
99.8
99.8
99.7
99.9
99.3
99.9
99.9
100
34.8
1.3
1.2
1.6
1.4
1.4
1.2
1.3
1.2
4
4
4
5
5
5
5
5
5
100
a DH: hydrodynamic diameter of particles of the main peak. The
values reported are the average of 10 runs. b Volume particle size
distribution. c HHW, the width of the peak at half-height, an indication
of the degree of polydispersity of the aggregates.
Physical-Chemical Characterization. Three different pa-
rameters were considered for the study of the physical-chemical
properties of this novel series: determination of the surface
activity (CMC, the limit surface tension γCMC, and the occupied
area per surfactant molecule Amin); determination of log k′W,
which is correlated to the hydrophobic character of the surfac-
tant, and finally, determination of the size of the supra-molecular
assemblies formed in aqueous solution. The physical-chemical
properties recorded were in agreement to previous observations
on other series of hydrogenated, fluorinated, and hemifluorinated
surfactants.16,17
Effect of the Sulfur Atom. It is important to note that a sulfur
atom inserted within the hydrophobic chain of a surfactant is
usually considered as a hydrophobic unit. Evidence for this
assertion comes from the unfavorable affinity toward water of
the sulfur atom owing to its poor hydrogen bond acceptor
properties45 and the Hansch46 partition constant between octanol
and water (i.e., π ) 0.45 and -0.47 for the SCH3 and OCH3
group, respectively). The sulfide group might be considered as
a methylene regarding its behavior in aqueous media. For
instance, the CMC of n-octyl-ꢀ-D-glucopyranoside is in the
range of 20-25 mM, whereas its thio analogue, the n-octyl-ꢀ-
D-thioglucopyranoside, exhibits a CMC value of 9 mM that is
corresponding surfactant. Stock solutions were stored at room
temperature overnight before measurements. The values reca-
pitulated in Table 2 show that all surfactants bearing two or
three glucose moieties, despite the nature of the hydrophobic
chain or its length, self-organize in water into well-defined
monodisperse particles (see Figures S38-S40, Supporting
Information) with apparent hydrodynamic diameters ∼5-6 nm
as expected for globular micelles.42 Nevertheless, the diameter
of the particles formed by the Triglu compounds is 12-22%
lower than that found for the Diglu compounds. Only F6-
Monoglu self-aggregates into larger and polydisperse micelles
(diameter ∼25 nm). Since F6-Monoglu gives large and poly-
disperse micelles, there was no need for the synthesis of its
hemifluorinated analogue.
(c) log k′W Determination. Table 1 reports values of log k′w,
a parameter closely related to the molecule’s water/octanol
partition coefficient, which can be obtained from reverse-phase
HPLC.43 This parameter is usually considered to reflect the
hydrophobic character of the surfactant, and we previously found
that the volume and/or the nature of the polar group have only
a low impact on this parameter.44 Indeed, H10-Diglu and H10-
Triglu exhibit similar log k′w values, whereas those of the H12
are significantly higher. The addition of the ethyl tip at the end
of the fluorinated chain does not affect significantly the
hydrophobicity of the surfactant, as we can observe in the case
of F6-Diglu and H2F6-Diglu (log k′w ) 4.8 and 4.9, respectively).
Furthermore, whereas H10-Diglu and H2F6-Diglu can be con-
sidered as having the same hydrophobic chain length (10 C up
to the sulfur group), the hemifluorinated surfactant exhibits a
higher log k′w value and therefore appears to be more hydro-
phobic than its hydrogenated analogue (4.4 and 4.9, respec-
tively). Similar observation was noted with the TriGlu surfac-
tants (4.2 and 4.7, respectively, for H10-Triglu and H2F6-Triglu).
By the same token, we observed that to a given polar head,
even though perfluorinated F6 surfactants have a shorter
hydrophobic chain, they exhibit a higher log k′W, such a behavior
of the fluorinated surfactants is conformable to previous
observations on other series of surfactants.16
close to the CMC of the n-nonyl-ꢀ-D-glucopyranoside (6.5
48,49
mM).47 Furthermore, Menger and co-workers
found that
the insertion of a sulfide group causes an increase in the CMC,
the magnitude of which depends on the position of the sulfur
atom. In our case, as the location of the sulfur atom remains
constant for the whole series, distant by only two methylenes
from the polar head, we consider the thioether unit as a part of
the hydrophobic entity. Thus, the hydrogenated (H10 and H12),
the perfluorinated, and the hemifluorinated surfactants presented
in this article have to be considered bearing a hydrophobic chain
made of the alkyl group as well as the sulfur atom and the two
methylene from the THAM.
Comparison of Contributive Effects of the Hydrogenated
and the Fluorinated Chains. While the addition of an ethyl tail
to a hydrogenated surfactant decreases (∼10 times) the CMC
value without affecting the γCMC, the substitution of hydrogen
to fluorine results in lowering significantly both the CMC and
the γCMC at a constant chain length. This has been explained
by the larger absolute value of the free energy of micellization
for a fluorinated compound.16,50 Moreover, as the energy of
Discussion
Synthesis. We described herein the synthesis of a new class
of hemifluorinated glucose-based surfactants derived from
THAM. The synthesis is based on a one-pot reduction/alkylation
(45) Allen, F. H.; Bird, C. M.; Rowland, R. S.; Raithby, P. R. Acta
Crystallogr., Sect. B 1997, B53, 696–701.
(46) Leo, A.; Hansch, C.; Elkins, D. Chem. ReV. 1971, 71, 525–616.
(47) Wenk, M. R.; Seelig, J. Biophys. J. 1997, 73, 2565–2574.
(48) Menger, F. M.; Shi, L. J. Am. Chem. Soc. 2006, 128, 9338–9339.
(49) Lundberg, D.; Shi, L.; Menger, F. M. Langmuir 2008, 24, 4530–4536.
(42) Zana, R. Colloids Surf. A 1997, 123-124, 27–35.
(43) Braumann, T. J. Chromatogr., A 1986, 373, 191–225.
(44) Durand, G.; Poeggeler, B.; Boeker, J.; Raynal, S.; Polidori, A.; Pappolla,
M. A.; Hardeland, R.; Pucci, B. J. Med. Chem. 2007, 50, 3976–3979.
8148 J. Org. Chem. Vol. 73, No. 21, 2008

Glucose-Based Surfactants
transfer to the air/water interface becomes more favorable with
fluorinated derivatives, the pressure of the film at the CMC
becomes higher. The effect of the polar head size on Amin and
Γmax is obviously demonstrated when we compare the Diglu
and Triglu hydrogenated and fluorinated compounds. The
efficiency parameter (pC20) of the fluorinated series increases
with the order F6-Triglu < F6-Diglu < F6-Monoglu, demonstrat-
ing that the proneness to adsorb at the air/water interface
increases with the hydrophobic character of the molecule. This
is dramatically confirmed with the values recorded for the Diglu
compounds, 2.81, 2.47, and 1.76 for the F6, H12, and H10 chains,
respectively. As an evaluation of the hydrophobic character of
surfactants, the log k′W parameter has also been used.16,51,44,52
Logically, the values of log k′W increase with the length of the
chain, and thus H10 derivatives exhibit values of log k′W lower
than those of H12 ones, while as previously reported, the volume
and/or the nature of the polar group have only a low impact.
This is confirmed by the much closed values exhibited for H10-
Diglu and H10-Triglu (the same observation was made for H12
and F6 derivatives).
Atypical Behavior of Hemifluorinated Surfactants. The
hemifluorinated surfactants exhibit atypical behavior when
compared to the H- or F-surfactants in several respects: (i)
Whereas an increase of the hydrophobic chain length induces a
decrease of the CMC, we observed that hemifluorinated sur-
factants display CMC values very similar to those of their
fluorocarbon analogues devoid of ethyl tip. Similar observations
have already been made for the TAC, the Lac, and the Malt
series.11,16,17 (ii) The limit surface tension values are between
the expected much lower values of the fluorinated compounds
and the higher values of the hydrogenated ones. (iii) Whereas
the efficiency parameter would be expected to dramatically
increase as the fluorine content of the chain increase or as the
chain length increase, the H2F6 compounds exhibit intermediate
values. (iv) Finally, despite the addition of an ethyl tip, the log
k′W values are closer to that of the parent fluorinated compound.
However, it has to be underlined that the surface area per
molecule was slightly affected by the additional ethyl tip, as
has been already reported for the Lac and Malt series.16,17
Among the contributions that lead to the unexpected hydro-
phobicity and the paradoxical behavior of the surface activity
of such hemifluorinated surfactants, two are likely to be16 (i)
poorer packing induced by either unfavorable van der Waals
interactions between fluorinated and hydrogenated atoms and/
or steric hindrance in the core of H2F6-surfactants micelles and
(ii) the acidic character of the methylene group vicinal to the
fluorocarbon core, a consequence of the electron-withdrawing
effect of fluorine atoms, which favors hydrogen bonding with
the aqueous phase. Such an effect would increase the energy
of transfer of the hemifluorinated chain of H2F6-surfactants to
a hydrophobic phase and shift its partition coefficient toward
water.
FIGURE 3. Correlation of the hydrophobic characters (log k′W)
determined by reverse-phase HPLC with (a) the efficiency parameter
(pC20) (a ) 0.63, R2 ) 0.45) and (b) the critical micellar concentration
(CMC) (a ) -0.99, R2 ) 0.86).
6.5).53 This was explained by an inhibition of the adsorption at
the interface of the latter surfactant, due to its greater hydro-
philicity, as it has also been reported for glucocationic surfac-
tants54 and gemini surfactants.55 We found opposite results as
shown with the F6 series with the following order: F6-Triglu >
F6-Diglu > F6-Monoglu. Confirmation of this observation was
made with the H10 and H12 series, while no significant difference
was observed for the H2F6 series. This suggests that both the
steric effect and the hydrophilicity of the polar head have to be
considered. Indeed, in the case of linear glycosidic polar head,
the number of the sugar moieties has limited effect on the
volume of the polar head while the hydrophilicity is obviously
increased with the number of sugars. Thus, the most contribu-
tively effect of the number of sugar on the CMC/C20 ratio is
the hydrophilicity. On the contrary, with ramified or branched
polar head, the steric hindrance appears to be more contributive
than the hydrophilicity as observed in this work. The nature of
the tail is also a determinant parameter toward the tendency to
adsorb or to micellize. For a given glucosylated polar head,
fluorinated surfactants exhibit highest values while intermediate
values are found for the hemifluorinated and the lowest for the
hydrogenated surfactants. Such a trend occurs likely because
the substitution of hydrogen by fluorine atoms within the tails
favors the adsorption at the interface as previously reported41
leading to a higher order of packing and thus a compact film.
Hydrophobic and Hydrophilic Effects toward the Micel-
lization. The CMC/C20 ratio is a measure of the tendency of the
surfactants to adsorb at the air/water interface relative to its tendency
to form micelles. Rosen and Shulthana reported that the decyl-ꢀ-
glucoside (CMC/C20 ) 11.1) has a greater tendency to adsorb at
the air/water interface than its maltoside analogue (CMC/C20
)
(50) Ravey, J. C.; Ste´be´, M. J. J. Coll. Surf., A 1994, 84, 11–31.
(51) Durand, G.; Polidori, A.; Ouari, O.; Tordo, P.; Geromel, V.; Rustin, P.;
Pucci, B. J. Med. Chem. 2003, 46, 5230–5237.
(52) Ortial, S.; Durand, G.; Poeggeler, B.; Polidori, A.; Pappolla, M. A.;
Boker, J.; Hardeland, R.; Pucci, B. J. Med. Chem. 2006, 49, 2812–2820.
(53) Rosen, M. J.; Sulthana, S. B. J. Colloid Interface Sci. 2001, 239, 528–
534.
(54) Quagliotto, P.; Viscardi, G.; Barolo, C.; D’Angelo, D.; Barni, E.;
Compari, C.; Duce, E.; Fisicaro, E. J. Org. Chem. 2005, 70, 9857–9866.
(55) Li, F.; Rosen, M. J.; Sulthana, S. B. Langmuir 2001, 17, 1037–1042.
J. Org. Chem. Vol. 73, No. 21, 2008 8149

Abla et al.
The peculiar behavior of the hemifluorinated chain, i.e.,
unfavorable interactions between the ethyl tip and the fluoro-
carbon core as well as the acidity of the methylene close to the
fluorinated core16 could explain the lower propensity of these
compounds to adsorb at the air/water interface comparatively
to their fluorinated analogues.
A relatively marked correlation between the log k′W and the
log CMC or the pC20 is shown in Figure 3. A significant
correlation was observed in the plot of the hydrophobic character
with the CMC (R2 ) 0.86). As regards their efficiency (which
measures in fact the ability of a surfactant to decrease the surface
tension by 20 mN/m), the adsorption phenomenon increases with
the hydrophobicity. Similarly, Zhang and Marchant demon-
strated that the efficiency of N-alkylmaltosamides increases
linearly with the length of the chain.37 However, a less
significant correlation (R2 ) 0.45) was found with pC20,
suggesting that this parameter does not depend only on the
nature of the hydrophobic part but also on the hydrophilicity
of the polar head while the log k′W is unaffected by the nature
and by the volume of the polar head.
Aggregation Behavior. The results of DLS experiments were
in perfect agreements with Israelachvili’s concept18 and con-
firmed the impact of the volumetric ratio on the nature and the
size of the aggregates formed in aqueous solutions. The
interfacial curvature of the aggregates increased with the size
of the polar head, leading to the formation of smaller and well-
defined aggregates (∼5-6 nm diameter). Moreover, for a given
type of polar head, fluorocarbon surfactants usually exhibit a
propensity to form structures with less curvature than hydro-
carbon surfactants and thus larger aggregates; this was verified
with lactobionamide16 or maltoside17 surfactants, for instance.
Thus, whereas F6-Monoglu self-aggregated into large particles
that might possibly be rod-like or cylindrical micelles, F6-
Diglu and F6-Triglu formed smaller and well-defined mono-
disperse particles as expected for globular micelles.42 Finally,
the interaction between the ethyl tip and the fluorinated core
of the HF-surfactants did not significantly affect the size of
the aggregates as we previously found for hemifluorinated
surfactants with smaller polar head.16,17 Again, one can expect
that an increase of the polar head size should induce the
formation of globular micelles.
Materials
Determination of log k′w Values. The compounds were dis-
solved in MeOH at 1 g L-1 and injected onto a reverse-phase
column (C18, 5 µm granulometry, 250 × 4.6 mm) at room
temperature. Compounds were eluted with various MeOH/H2O
mixtures (from 6:4 to 9:1 v/v). Linear regression analyses (r2 >
0.996) were performed on three points for H12-Diglu, H12-Triglu,
and H10-Triglu (MeOH/H2O from 7:3 to 9:1) and four points for
the other compounds (MeOH/H2O from 6:4 to 9:1). The measure-
ments were performed at a flow rate of 0.8 mL min-1 and detected
at 205 nm. The value of log k′ was calculated as log k′ ) log((t -
t0)/t0), where t is the retention time of the surfactant and t0 is the
elution time of MeOH, which is not retained on the column. log
k′w values were obtained by extrapolation of the linear regression
to 0% MeOH.43,51,56
Surface Tension Measurements. The surface activity of sur-
factants in solution at the air/water interface was determined by
the Wilhelmy plate technique. The surfactant solution was prepared
12 h prior to the measurements using Milli-Q water (resistivity of
18.2 MΩ cm, surface tension of 72.0 mN m-1 at 20 °C). Twenty
milliliters initial volume of the surfactant solution was taken in a
glass trough, and surface tensions were determined by dilution
technique. The platinum plate was cleaned by flaming before
experiments. All measurements were made at 25 ( 0.5 °C. Other
conditions were identical as reported else where.16,52 All measure-
ments were repeated 2 or 3 times. An estimate of the area per
molecule, Amin, at the interface is also given, as derived from the
surface excess calculated using the Gibbs adsorption isotherm, Γmax
) -(1/RT)(dγ/d(ln C)) where Γmax is the surface excess (moles
per unit area), R is the universal gas constant, T is the absolute
temperature, γ is the surface tension, and C is the surfactant
concentration. The Amin can be calculated as Amin ) 1/(NAΓmax),
where NA is the Avogadro number. C20 values were obtained by
extrapolation of the slope of the surface tension curve to 52 mN
m-1. The efficiency, pC20, was calculated as the negative logarithm
of C20.
Dynamic Light Scattering. The hydrodynamic particle size
distributions and polydisersity of surfactants at different concentra-
tions and temperatures were determined by using a He-Ne laser
(λ ) 633 nm, 4.0 mW). In a typical experiment, stock solutions
were made and stored at room temperature overnight before
measurements. On the day of the experiment aqueous solutions of
samples were passed through a 0.45 µm filter directly into
scintillation vials of 1.0 cm diameter. A closed plastic cuvette was
filled with 1.5 mL samples, and the size of the particles was
measured 1 h after filtration. The experimental run time was 10.0
min. The time-dependent correlation function of the scattered light
intensity was measured at a scattering angle of 173° relative to the
laser source (backscattering detection). The hydrodynamic radius
(R) of the particles was estimated from their diffusion coefficient
(D) using the Stokes-Einstein equation, D ) kBT/6πηR, where kB
is Boltzmann’s constant, T is the absolute temperature, and η is
the viscosity of the solvent.
Conclusion
A novel class of hemifluorinated glucose-based surfactants was
synthesized, as well as their hydrogenated and perfluorinated
analogues. The value of these syntheses lies in the efficient thioether
bond formation in mild conditions using thioacetate (or thiol) and
a slight excess of THAM-derived polar head. The surface
properties and related parameters were found to be influenced
by both the nature of the tail and the number of the glucose
moieties, demonstrating once again the particular behavior of
hemifluorinated surfactants. Finally, dynamic light scattering
investigations confirmed the impact of the volumetric ratio on
the nature and the size of the aggregates formed in aqueous
solutions. Whatever the nature of the tail, Diglu and Triglu
surfactants led to the formation of small and well-defined
aggregates that may be globular micelles. Considering that such
micelles should lead to monodisperse membrane proteins/
surfactant complexes, these novel surfactants show promising
structural features. Confirmation by neutron diffusion of the
globular shape of these aggregates and their potency to maintain
MP in their native form in an aqueous solution are underway.
Experimental Section
N-1-Acetoxymethyl-1-[(2′,3′,4′,6′-tetra-O-acetyl-ꢀ-D-glucopyra-
nosyl)oxymethyl]acetoxyethyl Acrylamide (2a). First, 2.4 g (11.35
mmol, 1 equiv) of 1, 7.0 g (17.03 mmol, 1.5 equiv) of 2,3,4,6-
tetra-O-acetyl-D-glucopyranosyl bromide, and 4.3 g of HgCN2
(17.03 mmol, 1.5 equiv) were dissolved in CH3CN in the presence
of drierite. The reaction was stirred 24 h, filtered over a pad of
Celite, and then concentrated under reduced pressure. The syrup
residue was redissolved in EtOAc. The organic layer was succes-
sively washed with NaHCO3, KI (10%), and saturated Na2S2O3
solutions, dried over Na2SO4, and then concentrated under reduced
(56) Hseih, M. M.; Dorsey, J. G. Anal. Chem. 1995, 67, 48–57.
8150 J. Org. Chem. Vol. 73, No. 21, 2008

Glucose-Based Surfactants
pressure. To this, 8.2 g of Montmorillonite K10 in CH2Cl2 was
added, and the reaction was stirred overnight. The crude mixture
was filtered over a pad of Celite and then concentrated under
reduced pressure. The resulting crude compound was purified by
flash chromatography (EtOAc/cyclohexane 90:10 v/v) to give 3 g
(5.94 mmol, 52%) of 2a as a white powder. Rf ) 0.24 (EtOAc/
165.4, 131.0, 126.7, 101.2, 72.6, 71.8, 71.4, 68.6, 68.2, 61.7, 61.7,
59.3, 20.7, 20.6.
Deacetylation of compound 5a was performed following the
same procedure as described for 2a, affording 5 in a quantitative
yield, which was used without further purification.
3,3,4,4,5,5,6,6,7,7,8,8-Dodecafluoro-1-iododecane (7). First, 3 g
(4.92 mmol, 1 equiv) of 1,10-diiodo-3,3,4,4,5,5,6,6,7,7,8,8-dode-
cafluorodecane25 and 1.23 g (4.92 mmol, 1 equiv) of tris(trimeth-
ylsilyl)silane were dissolved in 15 mL of toluene, under argon
atmosphere, and then heated at 70 °C in the presence of a catalytic
amount of AIBN. After 24 h of being heated, the mixture was
cooled at room temperature, and 25 mL of CH2Cl2 was added. The
organic layer was washed with Na2S2O3 saturated solution and
distilled water. The aqueous layer was extracted with CH2Cl2. The
combined organic layers were then dried over Na2SO4 and
concentrated under reduced pressure. The crude compound was
purified by silica gel chromatography (cyclohexane) to give 0.7 g
(1.44mmol, 30%) of pure compound 7 as a translucent liquid. Rf
cyclohexane 9:1 v/v). Mp ) 42.9-44.3 °C. [R]25 ) -10.95° (c
D
1, MeOH). HRMS (ESI+) calcd for C21H31NO13 ([M + H]+):
1
506.1868, found 506.1870. H NMR (CDCl3) δ 6.61 (br s, 1H),
2
3
3
6.34 (dd, J ) 1.66 Hz, Jtrans ) 16.86 Hz, 1H), 6.21 (dd, Jcis
)
)
3
2
3
9.87 Hz, Jtrans ) 16.86 Hz, 1H), 5.74 (dd, J ) 1.65 Hz, Jcis
9.75 Hz, 1H), 5.31-.95 (m, 3H, H2, H3, H4), 4.52 (d, J ) 7.93 Hz,
′
1H, H1), 4.43-3.45 (m, 9H, CH2OGlu, H5, H6, H6 , CH2OH),
2.11-2.02 (4s, 12H, COCH3). 13C NMR (CDCl3) δ 170.7, 170.1,
170.0, 169.5, 166.7, 130.4, 127.8, 100.9, 72.1, 72.0, 68.2, 71.3,
70.0, 63.8, 63.2, 61.7, 60.9, 20.7.
A 0.72 g (1.42 mmol) portion of 2a was dissolved in methanol
in the presence of a catalytic amount of MeONa and stirred
overnight at room temperature. Two spatulas of IRC 50 resin were
added, and the solution was filtered then concentrated under vacuum
to afford 2 in a quantitative yield, which was used without further
purification.
1
) 0.35 (cyclohexane). H NMR (CDCl3) δ 3.27 (m, 2H, CH2I),
2.84-2.63 (m, 2H, CH2Rf), 2.22-2.05 (m, 2H, CH3CH2Rf), 1.17
(t, J ) 7.47 Hz, 3H, CH3). 13C NMR (CDCl3) δ 36.5, 24.5, 11.5,
4.4. 19F NMR (CDCl3) -114.8 (2F), -116.3 (2F), -121.8 (4F),
-123.5 (4F).
N-1,1-Di[(2′,3′,4′,6′-tetra-O-acetyl-ꢀ-D-glucopyranosyl)oxyme-
thyl]acetoxyethyl Acrylamide (4a). First, 1 g (4.6 mmol, 1 equiv)
of 3, 7.27 g (17.48 mmol, 3.8 equiv) of 2,3,4,6-tetra-O-acetyl-D-
glucopyranosyl bromide and 3.48 g (13.8 mmol, 3 equiv) of HgCN2
were dissolved in CH3CN in the presence of drierite and sonicated
for 1 h. The reaction mixture was filtered over a pad of Celite, and
the solvent was removed under vacuum. Then, the residue was
dissolved in EtOAc. The organic layer was successively washed
with NaHCO3, KI (10%), and saturated Na2S2O3 solutions, dried
over Na2SO4 and then concentrated under reduced pressure. The
crude product was purified by flash chromatography (EtOAc/
cyclohexane 6:4 v/v) to give 1.8 g (2.05 mmol, 45%) of 4a as a
white powder. Rf ) 0.45 (EtOAc/cyclohexane 6:4 v/v). Mp )
In addition, 1.2 g (1.97mmol, 40%) of 6 and 0.17 g (0.49mmol,
10%) of bis reduced compound were isolated.
3,3,4,4,5,5,6,6,7,7,8,8-Dodecafluorodecane Thioacetate (8). First,
2.0 g (4.13 mmol, 1 equiv) of 7 was dissolved in 10 mL of
anhydrous DMF under argon atmosphere. Next, 0.57 g (4.96 mmol,
1.2 equiv) of potassium thioacetate was added, and the mixture
was stirred for 1 h at room temperature. Then 30 mL of water was
added, and the aqueous layer was extracted with EtOAc. The
organic layer was washed with distilled water, dried over Na2SO4,
and then concentrated under reduced pressure. The crude compound
was purified by flash chromatography (cyclohexane/CH2Cl2 95:5
v/v) to give 1.0 g (2.31 mmol, 56%) of 8 as a translucent liquid. Rf
) 0.23 (cyclohexane/CH2Cl2 95:5 v/v).1H NMR (CDCl3) δ 3.12
(t, J ) 8 Hz, 2H, CH2S), 2.64-2.37 (m, 2H, CF2CH2), 2.37 (s,
3H, CH3CO), 2.11-2.06 (m, 2H, CH3CH2), 1.16 (t, J ) 7.4 Hz,
3H, CH3 CH2). 13C NMR (CDCl3) δ 194.8, 31.7, 30.5, 24.6, 20.3,
4.5. 19F NMR (CDCl3) δ -114.8 (2F), -116.6 (2F), -122.2 (4F),
-123.8 (4F).
N-1-Acetoxymethyl-1-[(2′,3′,4′,6′-tetra-O-acetyl-ꢀ-D-glucopyra-
nosyl)oxymethyl]acetoxyethyl-4-thia-7,7,8,8,9,9,10,10,11,11,12,12,12-
tridecafluorododecanamide (9, F6-Monoglu). First, 500 mg (1.18
mmol, 1 equiv) of 1H,1H,2H,2H-perfluorooctylthioacetate (see
Supporting Information for further details) and 67.4 mg (1.78 mmol,
1.5 equiv) of NaBH4 were dissolved in 3 mL of MeOH under argon
atmosphere. After 15 min of being refluxed, a solution of 600 mg
(1.78 mmol, 1.5 equiv) of 2, in 9 mL of MeOH was added, and
the reaction was stirred for 3 h. The reaction mixture was cooled
at room temperature and then concentrated under reduced pressure.
The crude mixture was purified by flash chromatography (EtOAc/
MeOH 9:1 v/v) and by size-exclusion chromatography (MeOH),
to give 510 mg (0.72 mmol, 60%) of 9 as a white powder. Rf )
0.7 (EtOAc/MeOH/H2O 7:2:1 v/v/v). Mp ) 153.7. [R]25D ) -0.6°
(c 1, MeOH). HRMS (ESI+) calcd for C21H28NO9SF13 ([M + H]+):
718.1349, found 718.1351. 1H NMR (CD3OD) δ 4.32 (d, J ) 7.69
Hz, 1H, H1), 4.07 (d, J ) 10.02 Hz, 1H, CH2OGlu), 3.92-3.65
(m, 7H, CH2OGlu, CH2OH, H6, H6′), 3.42-3.20 (m, 4H, H5, H4,
H3, H2,), 2.89-2.79 (m, 4H, CH2SCH2), 2.61-2.41 (m, 2H, CH2Rf),
2.58 (t, J ) 7.1 Hz, 2H, CH2CO). 13C NMR (CD3OD) δ 173.4,
103.3, 76.6, 70.2, 73.6, 67.9, 61.7, 61.3, 60.9, 36.2, 31.6, 27.2, 21.9.
19F NMR (CD3OD) δ -82.4 (3F), -115.2 (2F), -122.9 (2F),
-123.8 (2F), -124.3 (2F), -127.3 (2F).
52.7-54.5 °C. [R]25 ) -25.10° (c 1, MeOH). HRMS (ESI+)
D
calcd for C37H52NO23 ([M + H]+): 878.2925, found 878.2925. 1H
NMR (CDCl3) δ 6.25 (dd, 2J ) 1.63 Hz, 3Jtrans ) 17 Hz, 1H, CH2),
3
3
6.17 (br s, 1H, NH), 6.21 (dd, Jcis ) 10.02 Hz, Jtrans ) 17.5 Hz,
1H, CH), 5.63 (dd, 2J ) 1.62 Hz, 3Jcis ) 10 Hz, 1H, CH2), 5.25-.94
(m, 6H, H2, H3, H4), 4.50 (d, J ) 9.5 Hz, 1H, H1), 4.49 (d, J )
7.88 Hz, 1H, H1), 4.41 (d, J ) 13.56 Hz, 2H, CH2OAc), 4.33-4.25
′
(m, 2H, H6), 4.17-3.92 (m, 6H, H6 , CH2Oglu), 3.75-3.69 (m,
2H, H5), 2.19-1.93 (9s, 27H, COCH3). 13C NMR (CDCl3) δ 170.9,
170.7, 170.6, 170.2, 170.1, 169.5, 169.4, 169.3 165.5, 130.9, 126.9,
101.1, 101.0, 72.6, 72.4, 71.9, 71.8, 68.8, 68.6, 68.2, 68.1, 68.2,
68.1, 63.0, 61.7, 58.8, 20.8, 20.7, 20.6.
Deacetylation of compound 4a was performed following the
same procedure as described for 2a, affording 4 in a quantitative
yield, which was used without further purification.
N-Tris[(2′,3′,4′,6′-tetra-O-acetyl-ꢀ-D-glucopyranosyl)oxymethyl]-
methyl Acrylamide (5a). The synthetic procedure was essentially
the same as for compound 4a; 0.75 g (4.27 mmol, 1 equiv) of
THAM, 10.54 g (25.66 mmol, 6 equiv) of 2,3,4,6-tetra-O-acetyl-
D-glucopyranosyl bromide, and 6.48 g (25.66, 6 equiv) of HgCN2
were used as starting materials. Purification by flash chromatography
(EtOAc/cyclohexane 7:3 v/v) led to 2.5 g (2.16 mmol, 50%) of 5a
as a white powder. Rf ) 0.47 (EtOAc/cyclohexane 8:2 v/v). Mp )
72.5-74.3 °C. [R]25 ) -28.15° (c 1, MeOH). HRMS (ESI+)
D
calcd for C49H68NO31 ([M + H]+): 1166.3770, found 1166.3772.
2
3
1H NMR (CDCl3) δ 6.25 (dd, J ) 1.67 Hz, Jtrans ) 15 Hz, 1H,
CH2), 6.14 (br s, 1H, NH), 6.07 (dd, 3Jcis ) 9.99 Hz, 3Jtrans ) 17.5
2
3
Hz, 1H, CH), 5.60 (dd, J ) 1.69 Hz, Jcis ) 12.5 Hz, 1H, CH2),
5.26-4.94 (m, 9H, H2, H3, H4), 4.48 (d, J ) 7.9 Hz, 3H, H1), 4.31
(dd, J ) 4.54 Hz, J ) 12.5 Hz, 3H), 4.19 (d, CH2OGlu, J ) 10.55
Hz, 3H), 4.11 (dd, J ) 2.3 Hz, J ) 12.5 Hz, 3H), 3.83 (d, J )
10.41 Hz, 3H, CH2OGlu), 3.75-3.69 (m, 2H, H5), 2.15-2.01 (12s,
36H, COCH3). 13C NMR (CDCl3) δ 170.6, 170.2, 169.5, 169.4,
N-1,1-Di[(2′,3′,4′,6′-tetra-O-acetyl-ꢀ-D-glucopyranosyl)oxymeh-
tyl]acetoxyethyl-4-thia-7,7,8,8,9,9,10,10,11,11,12,12,12-tridecafluo-
rododecanamide (10, F6-Diglu). The synthetic procedure was
essentially the same as for 9. 250 mg (0.59 mmol, 1 equiv) of
1H,1H,2H,2H-perfluorooctylthioacetate, 34.5 mg (0.91 mmol, 1.5
J. Org. Chem. Vol. 73, No. 21, 2008 8151

Abla et al.
equiv) of NaBH4 and 420 mg (0.86 mmol, 1.45 equiv) of 4 were
used as starting materials. Purification by flash chromatography
(EtOAc/MeOH 8:2 v/v), by size-exclusion chromatography (MeOH),
and RP-HPLC (H2O/CH3CN 65:35 v/v) followed by lyophilization
led to 285 mg (0.32 mmol, 55%) of 10 as a white powder. Rf )
3.42-3.18 (m, 12H, H5, H4, H3, H2); 2.86-2.78 (m, 4H,
CH2SCH2); 2.58-2.45 (m, 2H, CH2Rf; 2.55 (t, J ) 6.96 Hz, 2H,
CH2CO); 2.26-2.06 (m, 2H, CH3CH2Rf); 1.16 (t, J ) 7.43 Hz,
3H, CH3CH2Rf). 13C NMR (CD3OD) δ 172.9, 103.5, 76.6, 76.6,
70.3, 73.3, 67.8, 61.3, 59.9, 36.4, 31.6, 27.07, 24.10, 22.04, 3.3.
19F NMR (CD3OD) δ -115.32 (2F), -117.53 (2F), -122.89 (4F),
-124.42 (2F), -124.70 (2F).
0.46 (EtOAc/MeOH/ H2O 7:2:1 v/v/v). Mp ) 169.4 °C. [R]25
)
D
-10.25° (c 1, MeOH). HRMS (ESI+) calcd for C27H38NO14SF13
([M + H]+): 880.1878, found 880.1878. 1H NMR (CD3OD) δ 4.35
(d, J ) 7.72 Hz, 1H, H1), 4.34 (d, J ) 7.73 Hz, 1H, H1), 4.21-4.13
(m, 2H, CH2OGlu), 3.96-3.84 (m, 6H, CH2OGlu, CH2OH, H6,),
3.72-3.65 (m, 2H, H6′), 3.38-3.19 (m, 8H, H5, H4, H3, H2),
2.85-2.79 (m, 4H, CH2SCH2), 2.60-2.44 (m, 2H, CH2Rf,), 2.57
(t, J ) 7.21 Hz, 2H, CH2CO). 13C NMR (CD3OD) δ 173.2, 103.4,
103.3, 76.6, 70.3, 73.3, 67.8, 67.7, 61.3, 61.0, 60.9, 36.23, 31.5,
27.1-22.0. 19F NMR (CD3OD) δ -82.4 (3F), -115.3 (2F), -122.9
(2F), -123.9 (2F), -124.3 (2F), -127.3 (2F).
N-1,1-Di[(2′,3′,4′,6′-tetra-O-acetyl-ꢀ-D-glucopyranosyl)oxymeth-
yl]acetoxyethyl-4-thia-tetradecanamide (14, H10-Diglu). The syn-
thetic procedure was essentially the same as for compound 9; 118
mg (0.68mmol, 1 equiv) of 1-decanethiol, 33.3 mg (0.88 mmol,
1.3 equiv) of NaBH4, and 440 mg (0.88 mmol, 1.3 equiv) of 4
were used as starting materials. Purification by size-exclusion
chromatography (MeOH) and RP-HPLC (H2O/CH3CN 60:40 v/v)
followed by lyophilization led to 232 mg (0.34mmol, 51%) of 14
as a white powder. Rf ) 0.45 (EtOAc/MeOH/ H2O 7:2:1 v/v/v).
Mp ) 129.9 °C, [R]25 ) -9.55° (c 1, MeOH), HRMS (ESI+)
N-Tris[(2′,3′,4′,6′-tetra-O-acetyl-ꢀ-D-glucopyranosyl)oxymethyl]-
methyl-4-thia-7,7,8,8,9,9,10,10,11,11,12,12,12-tridecafluorodode-
canamide (11, F6-Triglu). The synthetic procedure was essentially
the same as for compound 9; 150 mg (0.4 mmol, 1 equiv) of
1H,1H,2H,2H-perfluorooctylthioacetate, 19.7 mg (0.52 mmol, 1.3
equiv) of NaBH4, and 344 mg (0.52 mmol, 1.3 equiv) of 5 were
used as starting materials. Purification by size-exclusion chroma-
tography (MeOH) and RP-HPLC (H2O/CH3CN 65:35 v/v) followed
by lyophilization led to 217 mg (0.21mmol, 52%) of 11 as a white
powder. Rf ) 0.35 (EtOAc/MeOH/ H2O 7:2:1 v/v/v). Mp ) 180.2
D
calcd for C29H55NO14S ([M + H]+): 674.3416, found 674.3422.
1H NMR (CD3OD) δ 4.35 (d, J ) 7.72 Hz, 1H, H1,), 4.34 (d, J )
7.64 Hz, 1H, H1), 4.21-4.12 (m, 2H, CH2OGlu), 3.94-3.83 (m,
6H, CH2OGlu, CH2OH, H6), 3.72-3.66 (m, 2H, H6′), 3.39-3.19
(m, 8H, H5, H4, H3, H2,), 2.77 (t, J ) 7.29 Hz, 2H, CH2CO),
2.59-2.51 (m, 4H, CH2SCH2), 1.63-1.44 (m, 2H, SCH2CH2,), 1.32
(s, 14H, CH2), 0.93 (m, J ) 5.66 Hz, 3H, CH3). 13C NMR (CD3OD)
δ 173.6, 103.4, 103.3, 76.6, 70.2, 73.6, 67.8, 67.7, 61.3, 61.0, 60.9,
36.6, 31.6, 31.4, 29.3, 29.1, 29.0, 28.5, 27.2, 22.3, 13.1.
°C. [R]25 ) -19.75° (c 1, MeOH). HRMS (ESI+) calcd for
N-Tris[(2′,3′,4′,6′-tetra-O-acetyl-ꢀ-D-glucopyranosyl)oxymethyl]-
methyl-4-thia-tetradecanamide (15, H10-Triglu). The synthetic
procedure was essentially the same as for compound 9; 175 mg (1
mmol, 1 equiv) of 1-decanethiol, 56.9 mg (1.5 mmol, 1.5 equiv)
of NaBH4 and 995 mg (1.5 mmol, 1.5 equiv) of 5 were used as
starting materials. Purification by size-exclusion chromatography
(MeOH) and RP-HPLC (H2O/CH3CN 65:35 v/v) followed by
lyophilization led to 480 mg (0.57mmol, 57%) of 15 as a white
powder. Rf ) 0.20 (EtOAc/MeOH/ H2O 7:2:1 v/v/v). Mp ) 167
D
C33H48NO19SF13 ([M + H]+): 1042.2406, found 1042.2411. H
1
NMR (CD3OD) δ 4.14- 4.07 (m, 6H, H1, CH2OGlu), 3.74-3.64
(m, 6H, CH2Oglu, H6), 3.47-3.42 (m, 3H, H6′), 3.19-2.95 (m,
12H, H5, H4, H3, H2), 2.64-2.61 (m, 4H, CH2SCH2), 2.32 (t, J )
7.46 Hz, 2H, CH2CO), 2.35-2.29 (m, 2H, CH2Rf). 13C NMR
(CD3OD) δ 172.9, 103.5, 76.6, 70.2, 73.7, 67.8, 61.3, 59.9, 36.4,
31.5, 27.1, 22.0. 19F NMR (CD3OD) δ -82.4 (2F), -115.3 (2F),
-122.9 (2F), -123.9 (2F), -124.3 (2F), -127.3.
°C. [R]25 ) -19.55° (c 1, MeOH). HRMS (ESI+) calcd for
N-1,1-Di[(2′,3′,4′,6′-tetra-O-acetyl-ꢀ-D-glucopyranosyl)oxymeth-
yl]acetoxyethyl-4-thia-7,7,8,8,9,9,10,10,11,11,12,12-dodecafluoro-
tetradecanamide (12, H2F6-Diglu). The synthetic procedure was
essentially the same as for 9; 450 mg (1.04 mmol, 1 equiv) of 8,
53.8 mg (1.56 mmol, 1.5 equiv) of NaBH4, and 727 mg (1.45 mmol,
1.4 equiv) of 4 were used as starting materials. Purification by size-
exclusion chromatography (MeOH) and RP-HPLC (H2O/CH3CN
60:40 v/v) followed by lyophilization led to 399.9 mg (0.45 mmol,
44%) of 12 as a white powder. Rf ) 0.54 (EtOAc/MeOH/H2O 7:2:1
v/v/v). Mp ) 160 °C. [R]25D ) -3.95° (c 1, MeOH). HRMS (ESI+)
calcd for C29H43NO14SF12 ([M + H]+): 890.2285, found 890.2293.
1H NMR (CD3OD) δ 4.36 (d, J ) 7.71 Hz, 1H, H1), 4.34 (d, J )
7.77 Hz, 1H, H1), 4.20-4.13 (m, 2H, CH2OGlu), 3.94-3.84 (m,
6H, CH2OGlu, CH2OH, H6), 3.70-3.67 (m, 2H, H6′), 3.40-3.23
(m, 8H, H5, H4, H3, H2), 2.87-2.78 (m, 4H, CH2SCH2,), 2.61-2.45
(m, 2H, CH2Rf), 2.58 (t, J ) 7.08 Hz, 2H, CH2CO), 2.21-2.11
(m, 2H, CH3CH2Rf), 1.16 (t, J ) 7.41 Hz, 2H, CH3CH2). 13C NMR
(CD3OD) δ 173.3, 103.4, 103.3, 76.5, 70.1, 73.3, 67.9, 67.7, 61.2,
60.9, 36.2, 31.6, 27.1, 24.1, 22.0, 3.3. 19F NMR (CD3OD) δ -115.2
(2F), -117.5 (2F), -122.9 (4F), -124.4 (2F), -127.7 (2F).
N-Tris[(2′,3′,4′,6′-tetra-O-acetyl-ꢀ-D-glucopyranosyl)oxymethyl]-
methyl-4-thia-7,7,8,8,9,9,10,10,11,11,12,12-dodecafluorotetrade-
canamide (13, H2F6-Triglu). The synthetic procedure was essentially
the same as for 9; 200 mg (0.46 mmol, 1 equiv) of 8, 23.9 mg
(0.69 mmol, 1.5 equiv) of NaBH4, and 458 mg (0.69 mmol, 1.5
equiv) of 5 were used as starting materials. Purification by size-
exclusion chromatography (MeOH) and RP-HPLC (H2O/CH3CN
65:35 v/v) followed by lyophilization led to 291 mg (0.28 mmol,
60%) of 13 as a white powder. Rf ) 0.33 (EtOAc/MeOH/H2O 7:2:1
D
C35H65NO19S ([M + H]+): 836.3944, found 836.3947.1H NMR
(CD3OD) δ 4.22 (d, J ) 7.71 Hz, 3H, H1), 4.19 (d, J ) 10.11 Hz,
3H, CH2OGlu), 3.83 (d, J ) 10.26 Hz, 3H, CH2OGlu), 3.79-3.74
(m, 3H, H6), 3.59-3.57 (m, 3H, H6′), 3.26-3.06 (m, 12H, H5, H4,
H3, H2), 2.60 (t, J ) 7.27 Hz, 2H, CH2CO), 2.47-2.36 (m, 4H,
CH2SCH2,), 1.51-1.45 (m, 2H, SCH2CH2), 1.19 (s, 14H, CH2),
0.80 (t, J ) 6.32 Hz, 3H, CH3). 13C NMR (CD3OD) δ 173.0, 103.4,
76.6, 76.6, 70.2, 73.3, 67.9, 61.4, 59.9, 36.8, 31.4, 29.3, 29.1, 29.1,
28.6, 27.1, 22.4, 13.1.
N-1,1-Di[(2′,3′,4′,6′-tetra-O-acetyl-ꢀ-D-glucopyranosyl)oxymeth-
yl]acetoxyethyl-4-thia-hexadecanamide (16, H12-Diglu). The syn-
thetic procedure was the same as for 9; 150 mg (0.57 mmol, 1
equiv) of 1-dodecanethiol, 27.6 mg (0.80 mmol, 1.4 equiv) of
NaBH4, and 0.4 g (0.80 mmol, 1.4 equiv) of 4 were used as starting
materials. Purification by size exclusion chromatography (MeOH)
followed by lyophilization led to 0.21 g (0.29 mmol, 53%) of 16
as a white powder. Rf ) 0.43 (EtOAc/MeOH/H2O 7:2:1 v/v/v).
Mp ) 125.4 °C. [R]25D ) -1.65° (c,1,MeOH). HRMS (ESI+) calcd
for C31H59NO14S ([M + H]+): 702.3729, found 702.3727.1H NMR
(CD3OD) δ 4.34 (d, J ) 7.69 Hz, 1H, H1), 4.33 (d, J ) 7.71 Hz,
1H, H1), 4.21-4.13 (m, 2H, CH2OGlu), 3.96-3.83 (m, 6H,
CH2OGlu, CH2OH, H6), 3.72-3.65 (m, 2H, H6′), 3.39-3.19 (m,
8H, H5, H4, H3, H2), 2.71 (t, J ) 7.01 H, 2H, CH2CO), 2.59-2.51
(m, 4H, CH2SCH2), 1.63-1.55 (m, 2H, SCH2CH2), 1.32 (s, 18H,
CH2), 0.92 (t, J ) 6.25 Hz, 3H, CH3). 13C NMR (CD3OD) δ 173.6,
103.4, 103.3, 76.6, 70.2, 73.6, 67.8, 67.7, 61.3, 61.0, 60.9, 36.6,
31.4, 29.4, 29.3, 29.1, 29.0, 28.6, 27.1, 22.4, 13.1.
N-Tris[(2′,3′,4′,6′-tetra-O-acetyl-ꢀ-D-glucopyranosyl)oxymethyl]-
methyl-4-thia-hexadecanamide (17, H12-Triglu). The synthetic
procedure was essentially the same as for compound 9; 100 mg
(0.49 mmol, 1 equiv) of 1-dodecanethiol, 24.29 mg (0.64 mmol,
1.3 equiv) of NaBH4, and 424 mg (0.64 mmol, 11.3 equiv) of 5
were used as starting materials. Purification by size-exclusion
v/v/v). Mp ) 177.5 °C. [R]25 ) -18.25° (c 1, MeOH). HRMS
D
(ESI+) calcd for C35H53NO19SF12 ([M + H]+): 1052.2813, found
1052.2826. -115.2 (2F), -122.8 (2F), -123.8 (2F), -124.2 (2F),
-127.3. 1H NMR (CD3OD) δ 4.36-4.30 (m, 6H, H1, CH2OGlu);
3.97-3.87 (m, 6H, CH2OGlu, H6); 3.76-3.65 (H6’, m, 3H);
8152 J. Org. Chem. Vol. 73, No. 21, 2008

Glucose-Based Surfactants
chromatography (MeOH) and RP-HPLC (H2O/CH3CN 55:45 v/v)
followed by lyophilization led to 185 mg (0.21mmol, 44%) of 17
as white powder. Rf ) 0.29 (EtOAc/MeOH/ H2O 7:2:1 v/v/v). Mp
) 164 °C. [R]25D ) -6.7° (c 1, MeOH). HRMS (ESI+) calcd for
sample of 1,10-diiodo-3,3,4,4,5,5,6,6,7,7,8,8-dodecafluorode-
cane. We also thank Dr. A.-S. Fabiano-Tixier, Dr. K. S. Sharma,
and Dr. C. Breyton (IBPC-Paris) for their helpful discussions
and suggestions.
1
C37H69NO19S ([M + H]+): 864.4257, found 864.4257. H NMR
(CD3OD) δ 4.22 (d, J ) 7.64 Hz, 3H, H1), 4.16 (d, J ) 9.89 Hz,
3H, CH2OGlu), 3.79 (d, J ) 10.27 Hz, 3H, CH2OGlu,), 3.79-3.74
(m, 3H, H6), 3.59-3.53 (m, 3H, H6′), 3.26-3.06 (m, 12H, H5, H4,
H3, H2), 2.63 (t, J ) 7.16 Hz, 2H, CH2CO), 2.47-2.36 (m, 4H,
CH2SCH2), 1.51-1.43 (m, 2H, SCH2CH2), 1.19 (s, 18H, CH2), 0.80
(t, J ) 6.27 Hz, 3H, CH3). 13C NMR (CD3OD) δ 173, 103.5, 76.6,
76.6, 70.2, 73.3, 67.9, 61.4, 59.9, 36.8, 31.7, 31.5, 29.4, 29.3, 29.1,
29.0, 28.6, 27.1, 22.4, 13.1.
Supporting Information Available: General experimental
methods;experimentalprocedureforthesynthesisof1H,1H,2H,2H-
1
perfluorooctylthioacetate; H, 13C, 19F, DEPT 135, DEPT 90,
DEPT 45, COSY and HMQC NMR spectra of F6-Monoglu
(compound 9); 1H and 13C spectra of compounds 2, 4, 5, 8, and
10-17 and 19F spectra of compound 8 and 10-13; analytical
RP-HPLC chromatogram and high resolution mass spectrum
of F6-Monoglu; hydrodynamic distribution and surface tension
curves of all the surfactants. This material is available free of
charge via the Internet at http://pubs.acs.org.
Acknowledgment. This work was supported by the Univer-
site´ d’Avignon et des Pays de Vaucluse and by the European
Community Specific Targeted Research Project “Innovative
tools for membrane structural proteomics”. We are indebted to
Dr. B. Ame´duri (Montpellier) who provided us with a generous
JO801379E
J. Org. Chem. Vol. 73, No. 21, 2008 8153