Synthesis and preliminary biochemical assessment of ethyl-terminated perfluoroalkylamine oxide surfactants

Article abstract of DOI:10.1016/S0960-894X(02)00242-1

The synthesis and usefulness in membrane biochemistry of a new class of surfactants have been investigated. 1-Ethyl-2-dimethylamine oxide polar heads were grafted onto a hydrocarbon, a fluorocarbon or an ethyl-capped fluorocarbon hydrophobic tail. The ability of the resulting surfactants to extract and/or to stabilize in aqueous solution a test membrane protein, cytochrome b₆ f, was evaluated. While it is not a detergent, the hemifluorinated derivative efficiently kept purified cytochrome b₆ f soluble, native and functional. The data suggest that alkyl-capped fluorocarbon surfactants provide an interesting alternative to classical detergents for handling membrane proteins in aqueous solutions under non-dissociating conditions.

Full text of DOI:10.1016/S0960-894X(02)00242-1

Bioorganic & Medicinal Chemistry Letters 12 (2002) 1587–1590
Synthesis and Preliminary Biochemical Assessment of
Ethyl-Terminated Perfluoroalkylamine Oxide Surfactants
Yann Chaudier,a Francesca Zito,b Philippe Barthelemy,a David Stroebel,b
Bruno Ameduri,c Jean-Luc Popotb,* and Bernard Puccia,*
a
` ´ ´ ´
Laboratoire de Chimie Bioorganique et des Systemes Moleculaires Vectoriels, Faculte des Sciences, Universite d’Avignon,
33 rue Louis Pasteur, F-84000 Avignon, France
´
b
´
Laboratoire de Physico-Chimie Moleculaire des Membranes Biologiques, CNRS and Universite Paris-7, UMR 7099,
Institut de Biologie Physico-Chimique, 13 rue Pierre et Marie Curie, F-75005 Paris, France
c
´
Laboratoire de Chimie Macromoleculaire, UMR CNRS 5076, ENSCM, 8 rue de l’Ecole Normale, F-34296 Montpellier
Cedex 5, France
Received 18 February 2002; accepted 11 April 2002
Abstract—The synthesis and usefulness in membrane biochemistry of a new class of surfactants have been investigated. 1-Ethyl-2-
dimethylamine oxide polar heads were grafted onto a hydrocarbon, a fluorocarbon or an ethyl-capped fluorocarbon hydrophobic
tail. The ability of the resulting surfactants to extract and/or to stabilize in aqueous solution a test membrane protein, cytochrome
b6 f, was evaluated. While it is not a detergent, the hemifluorinated derivative efficiently kept purified cytochrome b6 f soluble, native
and functional. The data suggest that alkyl-capped fluorocarbon surfactants provide an interesting alternative to classical deter-
gents for handling membrane proteins in aqueous solutions under non-dissociating conditions. # 2002 Elsevier Science Ltd. All
rights reserved.
Detergents are widely used for handling membrane
proteins in aqueous solutions.1 Their dissociative char-
acter, however, is responsible for many problems
encountered in membrane protein biochemistry.1,2 Sur-
factants with fluorocarbon hydrophobic tails, even
though they are highly tensioactive, are both hydro-
phobic and lipophobic: because they do not interact
favorably with the alkyl chains of natural lipids, their
partition coefficient into biological membranes is low,
which renders them non-cytolytic.3,4 This property
makes them precious in many applications such as tem-
porary blood substitutes, drug delivery systems or con-
trast agents.5 Because they are not detergents,
fluorinated surfactants have received little attention
from membrane biochemists. In a previous series of
experiments, we examined the behavior of membrane
proteins transferred to solutions of fluorinated surfac-
tants following their extraction by a classical, hydro-
genated detergent.6 Some of these molecules, obtained
by radical telomerization of Tris(hydroxymethyl)-
aminomethane (THAM) in the presence of o-per-
fluoroalkane thiol,7 proved able to prevent massive
protein precipitation provided they were used at suffi-
ciently high concentration.6 In order to try and improve
interactions between the protein transmembrane surface
and the hydrophobic chain of the surfactant without
making the latter lipophilic, we then added a hydro-
carbon tip to the fluorocarbon tail of a couple of
THAM-derived surfactants.8 In tests conducted using as
a model a membrane protein purified from photo-
synthetic membranes, cytochrome b6 f,2,9 hybrid surfac-
tants proved able to keep this fragile complex soluble in
aqueous solutions, native, and highly stable.8,10 While
the data appeared to support the usefulness of ethyl-
capped fluorocarbon chains, these first molecules pre-
sented practical drawbacks: on the one hand, the poly-
dispersity of THAM-derived telomeric surfactants 1
(Scheme 1) is not a desirable feature in membrane bio-
chemistry; on the other, monodisperse trigalactoside
THAM derivatives 2 were laborious to synthesize, and
they proved to be insufficiently stable in aqueous solu-
tions for routine biochemical use. Zwitterionic
detergents derived from amine oxide, such as dode-
cyldimethyl-N-amine oxide or lauroamido-N,N-di-
methyl-3-n-propylamine oxide have proven useful in
membrane biochemistry.1,11 Perfluorinated amine
*Corresponding authors. Tel.: +33-4-90-144442; fax: +33-4-90-
144449; e-mail: bernard.pucci@univ-avignon.fr (B. Pucci). Tel.: +33-
1-584-15014; fax: +33-1-584-15024; e-mail: jean-luc.popot@ibpc.fr
(J.L. Popot)
0960-894X/02/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(02)00242-1

1588
Y. Chaudier et al. / Bioorg. Med. Chem. Lett. 12 (2002) 1587–1590
Scheme 1. Chemical structure of compounds 1–2 and synthesis of amine oxide surfactants 5.
oxides have been previously synthesized for various
applications (fire extinguishing agents,12ꢀ17 blood sub-
stitutes or drug delivery systems5,18ꢀ20). In the present
work, we report on the synthesis and preliminary bio-
chemical assessment of fluorinated and hemifluorinated
surfactants bearing a 1-ethyl-2-dimethylamine oxide
polar head.
obtained in two steps. The first reaction involved an
activation of carboxylic acids by thionyl chloride. Crude
acyl chlorides were then reacted in cold methylene
chloride with two equivalents of diamine to provide the
corresponding tertiary amides 4a–d in high yield.
Because an excess of amine was used, all coupling reac-
tions were achieved without any additional base.
Finally, compounds 4a–d were oxidized into their amine
oxide analogues 5a–d. Oxidation was quantitatively
performed by an excess of hydrogen peroxide in hot
methanol. The amine oxide surfactants were isolated
with an overall yield ranging from 74 to 97%. They
were fully characterized by NMR spectroscopy and
mass spectrometry.23
Results and Discussion
Synthesis
We initially considered grafting the amine oxide to the
fluorocarbon tail via an ethyl spacer. However, given
the acidity of the methylene group linked to the per-
fluoroalkyl moiety, the introduction of an amine oxide
group in b position could favor a Cope elimination
reaction. The amine oxide synthesis recently developed
by Gellman et al.21 was therefore preferred. The amine
oxide compounds were obtained in two steps, namely (i)
condensation of N,N-dimethyl-N0-ethylethylene diamine
on the chosen carboxylic acid, followed by (ii) oxidation
of the tertiary amine by either hydrogen peroxide or
metachloroperbenzoic acid. This strategy was applied to
the different hydro-, fluoro- or hemifluoroacids 3.
Dodecanoic and o-perfluoroalkyl carboxylic acids are
commercially available. The hybrid ethyl-terminated
perfluoroalkyl chain was synthesized in the laboratory22
with 42% overall yield. The amine oxide hydrophilic
head was added according to Scheme 1. Amines were
Unlike previous non-ionic fluorinated surfactants,
which bore either polyhydroxyl, sulfoxyde or glycosidic
polar heads, and whose solubility in water was found to
be very sensitive to the nature and length of the per-
fluorocarbon chain,6,8,22 all amine oxide surfactants are
highly hygroscopic and exhibit a good water solubility,
in excess of 5 g Lꢀ1 (ꢁ10 mM). Preliminary experiments
with a Kruss K12 tensiometer using the Wilhelmy plate
method yielded the estimates of cmc values (in the Tris
buffer solutions used in biochemical assays) shown in
Table 1.
Regarding tensioactive properties, it is well-known that
surfactants containing a fluorocarbon chain as their
hydrophobic moiety display higher surface activity and
a lower cmc (critical micellar concentration) than their

Y. Chaudier et al. / Bioorg. Med. Chem. Lett. 12 (2002) 1587–1590
1589
experimental conditions, detrimental to the stability of
cytochrome b6 f: the complex released its associated
chlorophyll molecule, dissociated into a low-molecular
weight form (most likely a monomer), and lost its
enzymatic electron transfer activity. These observations
are typical of delipidation-induced inactivation.2 The
behavior of the b6 f complex in solutions of C6F13C2H4–
AO 5b was similar to that observed previously with
some other fluorinated surfactants:6 a high-molecular
weight (presumably dimeric), enzymatically active form
of the complex was predominant; however, a significant
tendency to aggregation indicated a relatively poor effi-
ciency of the surfactant to prevent protein–protein
hydrophobic interactions. With the hemifluorinated
compound, C2H5C6F12C2H4–AO 5d, no aggregation
was observed, most of the complex migrating as a sin-
gle, heavy species, most probably a dimer. This fraction
comprised all of the high molecular weight b6 f subunits,
including the Rieske Fe–S protein, as well as the b6 f-
associated chlorophyll. The latter two components are
known to be the first to be lost upon detergent-induced
dissociation of the complex.2 The electron transfer
activity of the complex was similar to that of control
preparations.
Table 1. Physico-chemical properties of amine oxide surfactants
Compd
cmc (mM)
gmin (mN mꢀ1
)
5a
5b
5c
5d
0.14
0.16
0.019
0.09
28
15
16
20
Measurements were carried out in 20 mM Tris–HCl buffer solution,
pH 8.0. Experimental uncertainties are estimated to be ca. ꢂ0.02 mM
on cmc (ca. ꢂ0.005 mM for compound 5c) and ꢂ1 mN mꢀ1 on gmin
values.
hydrocarbon counterparts of the same length. The phy-
sico-chemical data obtained with compounds 5a–d are
in good agreement with this rule (Table 1). Two points
are worth underlining: (i) the cmc values obtained are
very low and close to but slightly lower than those pre-
viously measured with compounds such as 2,9 and (ii)
once again, one notes that the cmc of the hybrid, hemi-
fluorinated compound C2H5C6F12CH2CH2–AO 5d
(0.09 mM) is abnormally high as compared to that of
C8F17CH2CH2–AO 5c (0.019 mM). This is probably due
to unfavorable intermolecular interactions lowering the
free energy of transfer of hybrid hydrophobic chains
from water to the core of the micelles.
Biochemical assays
Conclusion
Homologous hydrogenated, fluorinated and hemi-
fluorinated surfactants 5a–d were tested for their ability
(i) to solubilize integral membrane proteins from chloro-
plast thylakoid membranes and (ii) to maintain a pur-
ified membrane protein complex, cytochrome b6 f,
soluble in aqueous solutions while preserving its integrity,
enzymatic activity and monodispersity.
The results of biochemical assays on compounds 5a–d
can be summarized as follows.
1. C11H23–AO 5a is an efficient detergent. Under our
experimental conditions, it inactivated our test
membrane protein, cytochrome b6 f. This, however,
does not rule out its possible usefulness in biochem-
istry. The b6 f complex is a particularly fragile pro-
tein (which is one reason for our using it in these
experiments). At the high surfactant concentrations
chosen for our tests, dodecylmaltoside (C12M), one
of the most widely used detergents in membrane
biochemistry, has the same inactivating effect.2
2. C6F13C2H4–AO 5b and C8F17C2H4–AO 5c behave
in much the same way as some other surfactants
with perfluorinated alkyl chains that we character-
ized previously:6 neither of them solubilized thyl-
akoid membranes; C6F13C2H4–AO 5b was only
partially successful at keeping purified cytochrome
b6 f soluble when the latter was transferred into it
from a C12M solution, while C8F17C2H4–AO 5c was
inefficient. The usefulness of these two molecules as
tools for handling membrane proteins in aqueous
solutions is probably limited.
3. C2H5C6F12C2H4–AO 5d, on the contrary, exhibited
interesting properties. While it did not solubilize
thylakoids either, it efficiently kept purified cyto-
chrome b6 f soluble. The complex appeared struc-
turally and functionally intact, and did not show
any tendency to aggregation. In this respect, the
properties of this compound are similar to those
of two other hemifluorinated surfactants tested
previously, C2H5C6F12C2H4–SteloTHAM 1b and
C2H5C6F12C2H4–STHAMTriGal 2b.8,10 The gen-
erality of this observation seems to vindicate the
As previously observed for other fluorinated and hemi-
fluorinated surfactants,6,8,10 none of compounds 5b–d
was able to solubilize thylakoid membranes to any sig-
nificant degree even at high concentration (24 mM).
Their hydrogenated homologue, C11H23–AO 5a, is, on
the contrary, an efficient detergent, achieving significant
solubilization at 12 mM (corresponding to aꢁ1:1.2:2
weight ratio of surfactant/lipid/protein).
Surfactants 5a–d were then tested (at 5 mM) for their
ability to keep in a soluble state purified cytochrome b6 f
that had been transferred from a dodecylmaltoside
(C12M) solution. The hydrogenated 5a, short-chain
perfluorinated 5b and hemifluorinated 5d compounds
successfully passed this test, while the protein pre-
cipitated when transferred into C8F17C2H4–AO 5c
solutions.
The structural and functional state of cytochrome b6 f
complex transferred into surfactants 5a, 5b, or 5d was
then evaluated by determining its sedimentation coeffi-
cient, monodispersity, integrity and enzymatic activity
following ultracentrifugation on sucrose gradients con-
taining 5 mM of the molecule to be tested. Gradient
fractions were analyzed by SDS-Page, UV–visible spec-
trophotometry and electron transfer measurements.2,6,9
Exposure to C11H23–AO 5a proved to be, under our

1590
Y. Chaudier et al. / Bioorg. Med. Chem. Lett. 12 (2002) 1587–1590
predicament that led us to devise this family of
molecules, namely that they should be simul-
2. Breyton, C.; Tribet, C.; Olive, J.; Dubacq, J.-P.; Popot, J.-L.
J. Biol. Chem. 1997, 272, 21892.
3. Pucci, B.; Maurizis, J. C.; Pavia, A. A. Eur. Polym. J. 1991,
27, 1101.
4. Barthelemy, P.; Tomao, V.; Selb, J.; Chaudier, Y.; Pucci, B.
Langmuir. 2002, 18, 2557.
5. Kraft, M.-P. Adv. Drug Deliv. Rev. 2001, 47, 200.
6. Chabaud, E.; Barthelemy, P.; Mora, N.; Popot, J.-L.;
Pucci, B. Biochimie 1998, 80, 515.
7. Pavia, A. A.; Pucci, B.; Riess, J. G.; Zarif, L. Makromol.
Chem. 1992, 193, 2505.
8. Barthelemy, P.; Ameduri, B.; Chabaud, E.; Popot, J.-L.;
Pucci, B. Org. Lett. 1999, 1, 1689.
9. Pierre, Y.; Breyton, C.; Kramer, D.; Popot, J.-L. J. Biol.
Chem. 1995, 270, 29342.
10. Chabaud E. These de Doctorat de l’Universite Paris-VI,
2001.
11. Sardet, C.; Tardieu, A.; Luzzati, V. J. Mol. Biol. 1976,
105, 383.
12. Barry, J. M. (Procter & Gamble), US Patent 3194840,
1965.
13. Stern, R. M.; Fan, W. Q.; Blagev, L. P. PCT Int. Appl.,
1997.
14. Szonyi, S.; Vandamme, R.; Cambon, A. J. Fluor. Chem.
1985, 30, 37.
taneously less aggressive than detergents towards
membrane proteins (because of the physical proper-
ties of fluoroalkyl tails), while interacting more
favorably with them than surfactants with a per-
fluorinated alkyl chain (because of the ethyl tip).8,10
The present observations clearly establish the stabi-
lizing effect of the hemifluorinated chain, since the
homologous hydrocarbon surfactant, used under the
same conditions, totally inactivated our test protein.
This discrepancy was not as striking with the two
compounds studied earlier, which, probably thanks
to their large polar heads, were mild detergents in
their hydrocarbon version8,10 and inactivated cyto-
chrome b6 f only after protracted incubation.10 By
analogy with earlier experiments,24 we expect that
following fractionation on sucrose gradients con-
taining C2H5C6F12C2H4–AO, most of C12M will
have desorbed from the protein. On the other
hand, bound lipids ought to have been largely
retained. These points will be the object of direct
investigations.
15. Wirtz, P.; Broell, R. (Hoechst), Ger. Offen. DE 2829594,
1979.
From a practical point of view, C2H5C6F12C2H4–AO 5d
presents significant advantages over the molecules stud-
ied thus far : it is monodisperse and its synthesis is sim-
ple once the hemifluorinated acid has been prepared.
One of the two earlier molecules, C2H5C6F12C2H4–
STHAMTriGal 2b, was also monodisperse, but its
synthesis was particularly difficult and its stability in
aqueous solutions proved insufficient for biochemical
purposes. The second, C2H5C6F12C2H4–STeloTHAM
1b, was appropriately stable, easier if not simple to
produce, but featured a polydisperse, oligomeric polar
head. While detergents with heterogeneous polar heads,
such as Triton X-100, are still widely used in biochem-
istry, their polydispersity is a source of difficulties,
including problematic batch-to-batch reproducibility.
C2H5C6F12C2H4–AO 5d does not suffer from any of
these drawbacks. A full assessment of its potential use-
fulness in biochemistry will require further studies.
16. Clark, L. C., Jr.; Clark, E. W.; Moore, R. E.; Kinnett,
D. G.; Inscho, E. I., Jr. Prog. Clin. Biol. Res. 1983, 122, 169.
17. Corkill, J. M.; Goodman, J. F.; Ogden, C. P. Trans.
Faraday Soc. 1965, 61, 583.
18. Nivet, J. B.; Le Blanc, M.; Riess, J. G. J. Dispersion, Sci.
Technol. 1992, 13, 627.
19. Riess, J. G. New J. Chem. 1995, 19, 891.
20. Riess, J. G.; Krafft, M.-P. Biomaterials 1998, 19, 1529.
21. Mc Quade, D. T.; Quinn, M. A.; Yu, S. M.; Polans, A. S.;
Krebs, M. P.; Gellman, S. H. Angew. Chem., Int. Ed. 2000, 39,
758.
22. Chaudier, Y.; Barthelemy, P.; Pucci, B. Tet. Lett. 2001,
42, 3583.
23. 1H, 13C and 19F nuclear magnetic resonance (NMR)
spectra were recorded at 250 MHz on a Bruker AC250 spec-
trometer. Chemical shifts are expressed in parts per million
(ppm) relative to tetramethylsilane (1H and 13C spectra) or
CFCl3 (19F spectra). Mass spectra were recorded on a DX300
JEOLapparatus in Fab+. Example : compound 5d. 1H NMR
(DMSO-d6) d=1.09 (m, 6H, NCH2CH3 isomers cis–
3
trans+CH3CH2); 2.21 (m, 2H CH3CH2CF2 JHH=7.3 Hz,
Acknowledgements
3JHF=17 Hz, ), 2.65 (m 2H, CF2CH2); 3.16–3.48 (m, 12H,
CF2CH2CH2+N(CH2CH3)CH2CH2+NO(CH3)2); 3.71 and
3.81 (t, 2H, CH2NO(CH3)2, 3J=8 Hz isomers cis+trans).
19F NMR (CDCl3) d=ꢀ114.63 (CF2CH2CH2CO); ꢀ116.65
(CH3CH2CF2); ꢀ122.16 ((CF2)2CF2CH2CH2CO); ꢀ123.98
(CH3CH2CF2(CF2)2). 13C NMR (CDCl3) d=5.14 (CH3);
13.2 (NCH2CH3); 23.27 (CF2CH2CH2); 25.3 (CH3CH2);
25.73 (CF2CH2CH2); 40.43 (NO(CH3)2+(NCH2CH3); 42.18
(N(Et)CH2CH2); 58.14 (N(Et)CH2CH2); 106.35–122.56
(CF3+CF2); 169.54 (CO) FAB+ MS: m/z=517 (M+H)+.
HR MS m/z=517.1717 (M+H)+, calculated for
C17H25O2N2F12: 516.1646.
This research was supported in part by the inter-
disciplinary program Physique et Chimie du Vivant and
the Groupement de Recherche Colloıdes en Interaction
(GDR 692) of the Centre National de la Recherche
Scientifique, and by a Human Frontier Science Program
grant to J.-L.P., Y.C. and D.S. were supported by the
Ministere de la Recherche et de la Technologie.
References and Notes
24. Tribet, C.; Audebert, R.; Popot, J.-L. Langmuir 1997, 13,
5570.
1. Le Maire, M.; Champeil, P.; Møller, J. V. Biochim. Bio-
phys. Acta 2000, 1508, 86.