Synthesis of new fluidity-enhanced amphiphilic compounds for soluble protein two-dimensional crystallization purpose

Article abstract of DOI:10.1016/S0009-3084(99)00088-2

The synthesis of new amphiphilic compounds is described. The structures are rationally designed for soluble protein two-dimensional (2D) crystallization purpose. Special attention is devoted to fluidity properties expected of resulting monolayers. A series of 13 compounds was prepared containing unsaturated, branched or fluorinated alkyl chains. Structures are either symmetrical or dissymmetrical and present a hydroxyl group as polar head, eventually complemented with two other 'secondary' hydrophilic functions. Copyright (C) 1999 Elsevier Science Ireland Ltd.

Full text of DOI:10.1016/S0009-3084(99)00088-2

Chemistry and Physics of Lipids
103 (1999) 21–35
www.elsevier.com/locate/chemphyslip
Synthesis of new fluidity-enhanced amphiphilic compounds
for soluble protein two-dimensional crystallization purpose
Simone Nuss, Charles Mioskowski 1, Luc Lebeau *
Laboratoire de Synthe`se, Bioorganique associe´ au CNRS, Uni6ersite´ Louis Pasteur, 74, route du Rhin-BP 24-67401,
Illkirch, France
Received 5 May 1999; accepted 25 June 1999
Abstract
The synthesis of new amphiphilic compounds is described. The structures are rationally designed for soluble protein
two-dimensional (2D) crystallization purpose. Special attention is devoted to fluidity properties expected of resulting
monolayers. A series of 13 compounds was prepared containing unsaturated, branched or fluorinated alkyl chains.
Structures are either symmetrical or dissymmetrical and present a hydroxyl group as polar head, eventually
complemented with two other ‘secondary’ hydrophilic functions. © 1999 Elsevier Science Ireland Ltd. All rights
reserved.
Keywords: Amphiphile synthesis; Ether lipids; Branched lipids; Fluorinated lipids; Monolayer fluidity; Protein two-dimensional
crystallization
tists as some can spontaneously organize and
form periodic arrays at interfaces (for recent re-
1. Introduction
views see: Ku¨hlbrandt, 1988; Boekema, 1990; Jap
Amphiphilic molecules are known to self-orga-
et al., 1992). A number of investigation means
nize either as layers at the air–water interface or
(electron microscopy, image processing, crystallo-
as vesicles or tubes in solution. Extensive studies
graphic analysis . . .) were elaborated to get struc-
characterized the fundamental physical character-
tural information from such macromolecular
istics of such surfaces delimiting volumes or
assemblies and so was born what is called now the
boundaries (Gaines, 1966; Ulman, 1991). Am-
electron crystallography technique (for recent re-
phiphilic macromolecules such as membrane
proteins especially focused the attention of scien-
views see: Glaeser, 1985; Chiu et al., 1993).
To extend the use of the electron crystallogra-
phy technique to the structural study of soluble
macromolecules, a protein–lipid interaction has
to be established in order to generate amphiphilic
complexes that can behave to some extent like
membrane proteins. Due to the development of
* Corresponding author. Tel.: +33-3-88-676857; fax: +33-
3-88-678891.
E-mail addresses: mioskowski@aspirine.u-strasbg.fr (C.
Mioskowski), lebeau@aspirine.u-strasbg.fr (L. Lebeau)
1 Also corresponding author.
0009-3084/99/$ - see front matter © 1999 Elsevier Science Ireland Ltd. All rights reserved.
PII: S0009-3084(99)00088-2

22
S. Nuss et al. / Chemistry and Physics of Lipids 103 (1999) 21–35
supramolecular chemistry and biophysical ap-
proaches, it became possible to design new am-
phiphilic molecules and to test experimentally
their mechanical properties when spread at the
air–water interface as well as their ability to bind
macromolecules. Since 1971, organized two-di-
mensional (2D) arrangements (Fromherz, 1971)
and later 2D crystals of water-soluble proteins
(for recent reviews see: Jap et al., 1992; Lebeau et
al., 1996) were obtained on layers of charged
lipids, naturally functionalized molecules or syn-
thetic lipid derivatives. Electron microscopy
through its more recent developments is the only
available method to propose three-dimensional
(3D) models from objects that are one molecule
thick and a few square micrometers large. That
novel method is versatile as a large number of
factors can be modified to optimize the various
interactions leading to protein crystallization and
to make the molecular environment suitable for
the stabilization of a particular functional state of
the protein. This kind of experiments provides
structural information in a wide range of resolu-
tions and may be optimized to require as little as
10−12 moles of biomolecules.
One of the major goals of the authors is to
improve the characteristics of the lipid derivatives
to bind, concentrate, orient, and crystallize water-
soluble proteins. Systematic trials are thus per-
formed to understand and optimize the different
steps of the process, taking advantage of the
design and synthesis of various lipid derivatives.
A part of that work focused on the fluidity of the
lipid films that are used and its relationship with
the protein 2D crystallization process. As can be
intuitively approached, once a protein is concen-
trated on a lipid layer, its organization into 2D
arrays results from displacements and reorienta-
tions at the interface. These motions induced onto
the film result from a balance between a driving
force, that originates from protein–protein inter-
actions, and a counter-force due to lipid–lipid
interactions. That counter-force is an unusually
strong attraction arising between the non-polar
moieties of the amphiphilic molecules and has
been termed the ‘hydrophobic interaction’
(Muller, 1990; Blokzjil and Engberts, 1993). Inter-
actions resulting from hydrophobic effect can be
partly decreased or enhanced through electrostatic
interactions between the polar heads of the lipids
and between these polar heads and the bulk sol-
vent (McDaniel et al., 1983; So¨derman et al.,
1983; Lebeau et al., 1988; Scherer and Seelig,
1989; Hauser, 1991; Lohner, 1991; Sanderson et
al., 1991). The balance between these forces can
be regarded in terms of fluidity of the lipid layer.
So according to the protein 2D crystallization
process, the more fluid the lipid film, the better
the crystallization should work. This is partly
acknowledged throughout the literature reporting
2D crystallization of proteins involving a strong
interaction between the macromolecule and the
lipid film. In all cases, protein crystals were only
obtained on lipid films in the fluid phase. When
lipids were used as monolayers and compressed to
the solid phase, close-packing of proteins were
obtained but no true 2D crystals could be ob-
served (Darst et al., 1991; Mosser and Brisson,
1991). In systems where protein–lipid interactions
are weaker, fluidity of the lipid film might be of
less importance. It can be argued that the protein
has only little difficulties to break a weak interac-
tion with a lipid and create another one with a
neighbour lipid. That dynamical equilibrium
should allow the protein to move laterally in the
plane and eventually compensates for a lack of
fluidity of the lipid layer. However in the case of
RNA polymerase holoenzyme it was shown that
non-fluid lipid layers could not afford protein 2D
crystals despite the weak electrostatic interaction
between the protein and the charged lipid film
(Darst et al., 1988).
In the course of our work, we designed and
synthesized
a series of original amphiphilic
molecules that were studied by monolayer and
NMR techniques in order to characterize some
physical parameters related to fluidity. Some of
these compounds were linked to a novobiocin
derivative in order to perform comparative 2D
crystallization experiments with the DNA gyrase
B-subunit. Design and synthesis of the am-
phiphilic structures are discussed herein. Analysis
of their physical properties and evaluation of their
performances in terms of soluble protein 2D crys-
tallization inducers is discussed in the following
article.

S. Nuss et al. / Chemistry and Physics of Lipids 103 (1999) 21–35
23
2. Design of lipid structures
crystallization of the protein cannot be achieved
(Lebeau et al., 1996). This is the reason why all
the 2D crystallization experiments are systemati-
cally worked out with an excess of lipid (5–10-
fold the theoretical amount needed to form one
monolayer on the trough). In such conditions, the
surface pressure obtained characterizes the physi-
cal state of the lipid layer and was previously
termed as pseudo-equilibrium spreading pressure
(Lebeau et al., 1990) or elsewhere maximal
spreading pressure (Mosser and Brisson, 1991).
That particular value of the surface pressure can
be measured on larger troughs and appears to be
slightly below the collapse pressure (yc). Conse-
quently it is absolutely necessary to perform the
crystallization experiments at a temperature above
the main phase transition temperature in order to
handle the lipid film formed by the lipid of inter-
est in the fluid phase. From these last remarks it is
clear that in crystallization experiments performed
on standard small troughs, the physical state and
the fluidity of the lipid film depend on the molec-
ular structure of the lipid itself (which assigns the
molecular packing of the fatty chains and, hence
on, the magnitude of the hydrophobic cohesion
inside the layer) and on the temperature at which
experiments are realized.
Investigating on the influence of the fluidity of
the lipid support by acting on the temperature in
a set of crystallization experiments meets two
major difficulties. First, most of the proteins
rapidly loose their functionalities when handled at
4°C and a fortiori faster at higher temperature.
Prolonged incubation times inevitably lead to
protein heterogeneity that is definitely detrimental
to crystal formation. Second, the diffusion rate of
a protein from the bulk solvent to the lipid film is
also temperature-dependent. So in order to realize
artefact-free experiments on the influence of the
mere fluidity on the protein 2D crystallization
process, it is valuable to perform the experiments
at a given temperature. Therefore, the only
parameter that can be varied is the molecular
structure of the lipid. This necessitates the design
of a series of amphiphilic compounds exhibiting,
at the same temperature, different fluidity proper-
ties one from each other. We chose to synthesize
a number of amphiphilic compounds which struc-
When in a crystalline phase, a monomolecular
thick lipid film is characterized by
a high
lipophilic cohesion between the fatty chains and a
low translational diffusion coefficient D (:
10−7 cm2 s−1). When entering the fluid phase, D
increases by at least an order of magnitude. Then
along the whole fluid phase it is generally not
more than doubled. We were interested in investi-
gating how small fluidity variations in a fluid lipid
monolayer can induce modifications if any in the
2D crystallization process of a soluble protein.
These fluidity variations can be induced a priori
by modifying the surface tension imposed by the
lipid at the interface. This is easily realized by
compressing the monolayer by the mean of the
well-known Langmuir trough or film balance
technique (Ulman, 1991). However that method-
ology imposes the use of a rather large trough
containing milliliters of a protein solution (i.e.
milligrams of protein) and, most of the time, is
not adapted. The poor availability of the biologi-
cal material imposes some technical constraints in
2D crystallization experiments that are in fact the
price for miniaturization of the process. In order
to work with as little as 0.1–1 mg of protein,
incubations are performed in small teflon® wells
(4 mm in diameter, 1 mm deep) which does not
allow direct monitoring of the lipid film physical
state by surface tension measurements. Important
‘border effects’ (easily evidenced when handling
chromophoric amphiphiles) make that when a
lipid is deposited at the interface it does not
spread into a homogeneous monolayer but forms
a lipid bulk reservoir at the edges of the trough.
The density of the lipid film at the center region of
the interface is likely to be far lower than that at
areas close to the edges. So, if the theoretical
amount of lipid needed to cover the interface with
a fluid condensed (or liquid) monolayer is spread
on a very small trough, there is in fact a lack of
lipid when moving off the edge and the formation
of a fluid expanded (or gaseous) lipid phase to-
ward the center. Binding of a water-soluble
macromolecule onto such a film results in pro-
gressive solubilization of the lipid into the
aqueous phase and subsequent concentration and

24
S. Nuss et al. / Chemistry and Physics of Lipids 103 (1999) 21–35
ture should allow to cover an extended fluidity
range and able to spread into stable monolayers
at the air–water interface. The design of these
molecules was based on the hypothesis we made
according to which the lipophilic cohesion be-
tween the fatty chains of lipids inside a monolayer
(related to the fluidity of that monolayer) corre-
lates with the probability the lipid chains of a
given molecule have to interact with the lipid
chains of neighbour molecules. This probability
should depend on the average surface A occupied
by a molecule of lipid at the interface. In the
particular case where a large excess of lipid is
spread at the interface, A is assumed to be equal
to the specific area As corresponding to the sur-
face occupied by one molecule in the most con-
to display more than one polar moiety (hydroxyl
group) towards the aqueous phase, and to vary
the distance between two adjacent polar residues
on the same molecule. The rational for that was
to compel the spreading out of the bottom part of
the fatty chains parallel to the water surface,
increasing As (when compared to 1 and 2). Com-
densed state of
a true monolayer at the
temperature of the experiment. The As values can
be extrapolated from the pressure-area curve as
previously described (Lebeau et al., 1988). We
thus settled structures with As values expected to
cover a wide surface range. This was achieved by
varying the structure of their fatty chains, in an
attempt to modulate the steric hindrance provided
by the molecules at an interface. In a first step we
prepared lipid precursors without biological lig-
and attached on, so as to study their mechanical
properties (Fig. 1). Compound 1 is defined as our
reference lipid as it was already involved in differ-
ent crystallization programs. In compounds 2 and
3 a ramification is introduced on the alkyl chain
instead of the cis-unsaturation in 1. The branched
fatty chains are expected to partly prevent a close
packing of the chains and to decrease the hydro-
phobic effect inside the monolayer (i.e. increasing
its fluidity). Length and position of the chain
substitution were chosen in order to give the
maximal expansion to the lipid at the interface.
Thus the choice of butyl and hexyl substituents on
the ninth carbon atom of the main chain was
extrapolated from results reported in the literature
(Weitzel et al., 1951; Izawa, 1952; Kannenberg et
al., 1983; Lewis et al., 1986, 1987; Mantsch et al.,
1987; Rice et al., 1987; Brezesinski et al., 1987;
Balthasar et al., 1988; Menger et al., 1988;
Brezesinski et al., 1989; Matuo and Cadenhead,
1989; Brezesinski and Rettig, 1990; Asgharian et
al., 1991). Compounds 4, 5, and 6 were designed
Fig. 1. Structure of the lipid precursors prepared in that work.

S. Nuss et al. / Chemistry and Physics of Lipids 103 (1999) 21–35
25
pound 6 simultaneously exhibits multiple polar
groups and branched chains. In order to avoid as
much as possible a regular packing of the fatty
chains, compounds 7–13 present non-symmetrical
structures. In addition, a major terminal portion
of one of the two fatty chains in compounds 8–13
is perfluorinated. Although hydrophobic, fluoro-
carbons are lipophobic too, and establish very few
if any cohesive interaction with hydrogenated ho-
mologous compounds. The non-ideality of mixing
of fluorocarbons and hydrocarbons and the only
partial miscibility of the two classes of com-
pounds were evidenced 20 years ago (Mukerjee
and Yang, 1976). The mechanical behavior of
some particular fluorinated amphiphiles has al-
ready been studied by forming liposomes (Elbert
et al., 1984; Santaella et al., 1991; Gaentzler and
Vierling, 1993; Groß and Ru¨diger, 1994) or
monolayers (Elbert et al., 1984; Barton et al.,
1992; Shin et al., 1992) but there is no result
concerning mixed chains compounds.
Compounds 1–13 were examined when spread
into monolayers at the air-water interface using a
Langmuir trough and the film balance technique.
Another set of experiments involving solid-state
NMR techniques was conducted on these lipid
structures. All the experiments were settled to
estimate different physical parameters related to
the fluidity of the compounds (Lebeau et al.,
manuscript in preparation). Molecules 1, 2, 3, 5,
and 10 were subsequently modified by anchoring
to a novobiocin derivative (Fig. 2). That antibi-
otic is specifically recognized by the B-subunit of
DNA gyrase, an enzyme which function is to
control and modify the topological state of duplex
DNA (for a review see: Reece and Maxwell,
1991). This enzyme plays a key role in all the
biological processes involving nucleic acids (e.g.
replication, transcription, sister chromatide ex-
change) and its structural analysis, by the means
of radiocrystallography (for moderate size frag-
ments) as well as electron crystallography, still
represents a challenge to understand a number of
DNA-dependent biological activities. The func-
tionalized lipids 14a–e were systematically evalu-
ated in 2D-crystallization experiments of the
B-subunit of the gyrase protein (Lebeau et al.,
1996).
Fig. 2. Structure of the novobiocin derivatives prepared from
compounds 1, 2, 3, 5, and 10.
3. Synthesis
Symmetrical compounds 1 to 6 were obtained
by reacting 2 equivalents of the adequate sodium
alkoxide with epibromohydrin following a previ-
ously described procedure (Fig. 3) (Altenburger et
al., 1992). Dissymmetrical compounds 7–13 were
prepared in a two-steps procedure. One equivalent
of the sodium salt of a first alcohol reacted with
epibromohydrin to form the corresponding gly-
cidyl ether that was isolated and purified. That
epoxide then reacted with a second alcohol in the
presence of a Lewis acid catalyst to yield the
dissymmetrical 1,3-dialkoxy-2-propanol.
Branched alcohols used in the preparation of
compounds 2, 3, 6, 7, and 11–13 were prepared
starting from decanoic acid following a five-steps
procedure (Fig. 4). Decanoic acid methoxy-
methyl-amide 15 was obtained by coupling de-
Fig. 3. General scheme for the synthesis of the lipid precursors
1–13.

26
S. Nuss et al. / Chemistry and Physics of Lipids 103 (1999) 21–35
canoic acid with N,O-dimethyl hydroxylamine
chlorhydrate and reacted with the Grignard
reagent prepared from 1-bromo-8-(tetrahydropy-
ran-2-yloxy)-octane (Degering and Boatright,
1950) to yield ketone 16 in high yield (Nahm and
Weinreb, 1981; Fehrentz and Castro, 1983). That
compound was then condensed with n-butyl
lithium (resp. n-hexyl lithium) to afford the corre-
sponding tertiary alcohol 17a (resp. 17b). Dehy-
dration of that compound was realized
simultaneously with deprotection of the primary
alcohol to give 18a (resp. 18b). That double trans-
formation required the successive utilization of
two different solvents with Amberlyst® H-15. In
methanol, removal of the primary hydroxyl group
protection was complete but dehydration was very
slow, whereas it was efficiently achieved in
dichloromethane. In this solvent however very few
deprotected primary alcohol was obtained. The
use of a mixture of the solvents did not give better
results. Unsaturated alcohol 18a (resp. 18b) was
obtained as a mixture of three regioisomers and
was saturated by catalytic hydrogenation to yield
branched fatty alcohol 19a (resp. 19b).
Precursors of compounds 4 and 5 were ob-
tained from oleyl bromide (Fig. 5). Oleyl magne-
sium bromide reacted with 3-benzyloxy-1-
propanal (resp. 4-benzyloxy-1-butanal)2 and the
subsequent secondary alcohol 21a (resp. 21b) was
protected as its tetrahydropyranyl ether 22a (resp.
22b). Primary alcohol 23a (resp. 23b) was ob-
tained by reaction of the latter compound with
sodium in liquid ammonia (Scho¨n, 1984). Fatty
alcohol 23c was obtained following the same syn-
thetic pathway starting from alkyl bromide 20,
unless debenzylation of the intermediate benzyl
ether 22c was achieved by catalytic hydrogena-
tion. The reaction of two equivalents of com-
pound 23a (resp. 23b, 23c) with epibromohydrin
afforded the bis-protected triol 24a (resp. 24b,
24c) and acidic treatment with Amberlyst® H-15
yielded compound 4 (resp. 5, 6) (Fig. 6).
Compounds 1, 2, 3, 10, and 24b were selected
to be functionalized with novobiocin in order to
2 3-Benzyloxy-1-propanal (resp. 4-benzyloxy-1-butanal) was
obtained from 1,3-propanediol monobenzyl ether (resp. 1,4-
butanediol monobenzyl ether) by Swern oxidation.
Fig. 4. Synthesis of the immediate branched fatty alcohols and
halides.

S. Nuss et al. / Chemistry and Physics of Lipids 103 (1999) 21–35
27
24b) was alkylated with methanesulfonyl ester 253
to give the azido compound 26a (resp. 26b, 26c,
26d, and 26e) that was reduced with lithium alu-
minium hydride into the primary amine 27a (resp.
27b, 27c, 27d, and 27e) and further condensed
with the carboxylic acid derivative of novobiocin
28 (Lebeau et al., 1992) to yield compound 14a
(resp. 14b, 14c, 14d, and 29). Compound 29 orig-
inating from 24b was treated with Amberlyst®
H-15 in a methanol/dichloromethane mixture to
restore the two THP-protected hydroxyl groups at
the bottom of the fatty chains, and afforded com-
pound 14e.
be appraised in crystallization experiments (Fig.
7). The secondary alcohol in 1 (resp. 2, 3, 10, and
Since the synthesis of a number of compounds
followed the same route, only one particular ex-
ample in each series will be given here. Isolated
yields for each step and for each compound are
recorded in Tables 1–3. Intermediates were in-
1
variably characterized by IR, H-NMR and 13C-
NMR. Final compounds 1–13 and 14a–e were
additionally characterized by HRMS. Complete
spectral data are available upon request4.
3.1. General
Full general statements were described else-
where (Lebeau et al., 1991). Mass spectra at
chemical ionization were recorded on a Finnigan-
4600 quadrupole instrument. High resolution
mass spectra at fast-atom bombardment ioniza-
tion were recorded on a Varian Mat 311 GC
coupled spectrometer (compounds 1–13) or on a
Fisons Autospec Q spectrometer (compounds
14a–e) using a triazine or m-nitrobenzylic alcohol
matrix.
3.2. 1,3-Bis-(9-butyl-octadecyloxy)-propan-2-ol
(2)
A sodium hydride suspension (60% in oil, 1.6 g,
40.0 mmol) was added by portions to alcohol 19a
(6.53 g, 20.0 mmol) and HMPA (6 ml) in 60 ml of
anhydrous THF at 0°C. The mixture was refluxed
Fig. 5. Preparation of the fatty alcohol precursors involved in
the synthesis of compounds 4, 5 and 6.
4 Spectra and tables of spectral data appear in: Arbogast-
Nuss, S. The`se de Doctorat de l’Universite´ Louis Pasteur,
Strasbourg, France, 1994.
3 Compound 25 was readily obtained from tetraethylene
glycol, by preparing its bis-methanesulfonyl ester that was
monosubstituted with sodium azide.

28
S. Nuss et al. / Chemistry and Physics of Lipids 103 (1999) 21–35
Fig. 6. Synthesis of compound 4.
for 1 h, cooled to 0°C before epibromohydrin
(0.85 ml, 10.0 mmol) was added and refluxed
again for 16 h. The solution was decomposed by
the addition of aqueous NH4Cl and extracted
with ether. The organic layer was reduced in
vacuo and the residue was purified on silica gel
(hexane/ether 8:2) to yield compound 2 (5.26 g,
36.3; 36.1; 32.6; 31.9; 29.7-29.3 (m); 27.2; 25.6;
22.7; 14.1. IR (neat): n 3429.9; 2925.1; 2853.7;
1464.9; 1102.1; 913.4; 744.1. HRMS (triazine):
calcd for C45H88O5 [M]+ 708.66314, found
708.6606.
74%) as
a
pale yellow oil (mixture of
3.4. 1-(3,3,4,4,5,5,6,6,7,7,8,8,9,9,10,10,10-Hep-
tadecafluoro-decyloxy)-3-(9-butyl-octadecyloxy)-
propan-2-ol (13)
diastereomers). TLC: Rf 0.4 (hexane/ether 7:3).
1H-NMR (200 MHz, CDCl3): d 3.94 (m, 1 H);
3.49 (m, 8 H); 1.59 (m, 4 H); 1.27 (m, 70 H); 0.89
(t, J=6.3 Hz, 12 H). 13C-NMR (50 MHz,
CDCl3): d 71.8; 71.6; 69.4; 37.3; 33.7; 33.3; 31.9;
30.1; 29.7–29.6 (m); 29.3; 28.9; 26.8; 26.1; 23.1;
22.6; 14.1; 14.0. IR (CHCl3): n 3462.5; 2923.7;
2854.4; 1465.7; 1377.3; 1119.2; 722.3. HRMS (tri-
azine): calcd for C47H92O3 [M]+ 704.70461, found
704.70295.
A sodium hydride suspension (60% in oil, 1.1 g,
27.6 mmol) was added by portions to alcohol 19a
(6.0 g, 18.4 mmol) and HMPA (5 ml) in 50 ml of
anhydrous THF at 0°C. The mixture was refluxed
for 1 h, cooled to 0°C before epibromohydrin (1.9
ml, 22.1 mmol) was added, and refluxed again for
16 h. The solution was decomposed by the addi-
tion of aqueous NH4Cl and extracted with ether.
The organic layer was reduced in vacuo and the
residue was purified on silica gel (hexane/ether
8:2) to yield the expected 2-(9-butyl-octadecy-
loxymethyl)-oxirane (5.5 g, 85%) as a colorless oil.
TLC: Rf 0.5 (hexane/ether 7:3). 1H-NMR (200
MHz, CDCl3): d 3.71 (dd, J=3.1, 11.4 Hz, 1 H);
3.53-3.44 (m, 2 H); 3.37 (dd, J=5.8, 11.4 Hz, 1
H); 3.18-3.14 (m, 1 H); 2.80 (dd, J=4.2, 5.0 Hz, 1
H); 2.61 (dd, J=2.6, 5.0 Hz, 1 H); 1.59 (m, 2 H);
1.28 (m, 35 H); 0.89 (t, J=6.3 Hz, 6 H). Zirco-
nium tetrachloride (153 mg, 0.65 mmol) was
added to a mixture of the previous glycidyl ether
(500 mg, 1.31 mmol) and 2-perfluorooctyl-1-
ethanol (608 mg, 1.31 mmol) in 5 ml of anhydrous
THF, and stirred at room temperature for 2 h.
The reaction mixture was decomposed by the
addition of a few drops of water and dried over
sodium sulfate. The solvent was removed in vacuo
and the residue was purified by chromatography
(hexane/ethyl acetate 9:1) to yield compound 13
3.3. 1,3-Bis-(3-hydroxy-heneicos-12-enyloxy)-pro-
pan-2-ol (4)
Alcohol 24a (106 mg, 0.12 mmol) and Am-
berlyst® H-15 (30 mg) were stirred in a mixture
methanol/ethanol 1:1 for 16 h at 40°C. The cata-
lyst was filtered off and the filtrate was evapo-
rated
and
purified
by
chromatography
(hexane/ethyl acetate 50:50 to 25:75) to afford
compound 4 (74 mg, 86%) as a colorless oil
(mixture of diastereomers). TLC: Rf 0.4 (hexane/
ethyl acetate 25:75). 1H-NMR (200 MHz, CDCl3):
d 5.34 (m, 4 H); 3.94 (m, 1 H); 3.53 (m, 10 H);
2.01 (m, 8 H); 1.76–1.40 (m, 8 H); 1.27 (m, 48 H);
0.89 (t, J=6.3 Hz, 6 H). 13C-NMR (50 MHz,
CDCl3): d 129.9; 78.3; 78.2; 72.4–69.6 (m); 37.5;
5 The two branched alkyl chains undergo H2 elimination
between the tertiary carbon atom and 1 of its 3 neighbours.

S. Nuss et al. / Chemistry and Physics of Lipids 103 (1999) 21–35
29
Fig. 7. Preparation of the reference compound 14a. Compounds 14b–e were obtained in a similar way (see in the body text).
3.5. No6obiocin lipid deri6ati6e (14b)
(266 mg, 24%) as a pale yellow oil (mixture of
diastereomers). TLC: Rf 0.4 (hexane/ethyl acetate
8:2). 1H-NMR (200 MHz, CDCl3): d 3.91 (m,
1 H); 3.58–3.38 (m, 8 H); 1.92–1.68 (m, 2 H); 1.55
(m, 2 H); 1.27 (m, 35 H); 0.89 (t, J=6.3 Hz, 6 H).
13C-NMR (50 MHz, CDCl3): d6 71.9; 71.7; 71.5;
70.4; 66.3; 66.3; 44.5; 37.3; 33.7; 33.3; 31.9; 30.1;
29.7-29.3 (m); 28.9; 26.7; 26.1; 23.1; 22.6; 14.1;
14.0. IR (CHCl3): n 3447.7; 2924.8; 2855.3;
2360.0; 1465.8; 1376.6; 1119.8; 727.0. HRMS (tri-
azine): calcd for C35H55F17O3 [M]+ 846.3880,
found 8467.
Novobiocin derivative 28 (53 mg, 79 mmol),
DCC (36 mg, 174 mmol) and HOBT (13 mg, 95
mmol) were stirred for 5 h in 2 ml of DMF/
dichloromethane 1:1 before the amine 27b (70 mg,
Table 1
Yields for the synthesis of fatty alcohols and precursors
Yield, %
Compound
a
b
c
17
18
19
21
22
23
76
86
98
69
85
96
76
89
98
73
90
96
–
–
–
6 Chemical shifts corresponding to the carbon atoms bearing
fluorine substituents were not determined due to the high
multiplicity of the related signals in fluorine non-decoupled
13C-NMR spectra.
66
89
98a
7 High resolution mass spectra could not be satisfactorily
measured for this compound as well as for the other members
of the fluorinated series.
a Compound 23c was obtained from 22c by catalytic hydro-
genation over Pd/C.

30
S. Nuss et al. / Chemistry and Physics of Lipids 103 (1999) 21–35
Table 2
1331.7; 1280.4; 1115.4; 747.4. HRMS (m-NBA):
calcd for C88H149N3O18Na [M+Na]+ 1559.0734,
found 1559.0743; calcd for C88H150N3O18 [M+
H]+ 1537.0914, found 1537.0913.
Yields for the synthesis of compounds 1–13
Com-
pound
Yield, %a HRMS
Formula
Calcd
Found
3.6. Decanoic acid methoxy-methyl-amide (15)
1
2
3
4
5
6
7
8
9
10
11
12
13
79
74
71
62
62
58
81
14
17
19
21
21
24
C39H76O3
C47H92O3
C51H100O3
C45H88O5
C47H92O5
C53H108O5
C43H86O3
592.5794 592.5799
704.7046 704.7029b
760.7672 760.7642b
708.6631 708.6606
736.6944 736.6950
824.8196 824.8241
650.6577 650.6542
Decanoic acid (17.2 g, 100 mmol) and HOBT
(14.9 g, 110 mmol) in 200 ml of dichloromethane
were added with DCC (22.7 g, 110 mmol) in 100
ml of dichloromethane, and stirred for 1 h at 0°C.
N,O-dimethyl hydroxylamine chlorhydrate (10.7
g, 110 mmol) and anhydrous triethylamine (15.3
ml, 110 mmol) were added and the reaction mix-
ture was stirred for 2 h. Precipitated DCU was
filtered off and the filtrate was reduced in vacuo.
Residual DCU was precipitated by adding 100 ml
of ether and filtered off once again. The crude
mixture was evaporated and purified by flash
chromatography on silica gel (hexane/ether 6:4) to
yield 15 (19.6 g, 91%) as a colorless oil. TLC: Rf
0.4 (hexane/ether 1:1). 1H-NMR (200 MHz,
CDCl3): d 3.68 (s, 3 H); 3.17 (s, 3 H); 2.41 (t,
J=7.4 Hz, 2 H); 1.56 (m, 2 H); 1.25 (m, 12 H);
0.87 (t, J=6.3 Hz, 3 H). 13C-NMR (50 MHz,
CDCl3): d 174.6; 61.0; 32.1; 31.7; 29.3–29.1 (m);
24.5; 22.5; 13.9. IR (neat): n 3494.0; 2932.2;
2854.4; 1731.9; 1668.4; 1464.6; 1413.9; 1383.6;
1317.6; 1177.7; 1119.3; 1002.6; 722.7. MS (CI/
NH3): 233.2 [M+NH4]+.
C27H45F9O3 588.3225 588c
C29H45F13O3 688.3161 688c
C31H45F17O3 788.3097 788c
C31H55F9O3 646.4007 646c
C33H55F13O3 746.3944 746c
C35H55F17O3 846.3880 846c
a Based on epibromohydrin (Fig. 3).
b The two branched alkyl chains undergo H2 elimination
between the tertiary carbon atom and one of its three neigh-
bors.
c High resolution mass spectra could not be measured satis-
factorily for that compound.
79 mmol) was added. The mixture was stirred for
40 h at room temperature, then the solvent was
removed in vacuo and the residue was purified by
chromatography (dichloromethane/ethanol 8:2) to
afford compound 14b (79 mg, 65%) as a glassy
solid (mixture of diastereomers). TLC: Rf 0.6
(dichloromethane/ethanol 9:1). 1H-NMR (200
MHz, CDCl3): d 7.75–7.65 (m, 3 H); 7.23 (d,
J=9.1 Hz, 1 H); 6.86 (d, J=8.4 Hz, 1 H); 5.63
(d, J=2.4 Hz, 1 H); 5.37–5.33 (m, 2 H); 4.83 (s,
2 H); 4.30 (m, 1 H); 3.80-3.37 (m, 31 H); 2.31 (s,
3 H); 1.85 (s, 6 H); 1.55 (m, 4 H); 1.38 (s, 3 H);
1.28 (m, 70 H); 1.17 (s, 3 H); 0.89 (t, J=6.3 Hz,
12 H). 13C-NMR (50 MHz, CDCl3): d 167.7;
167.4; 162.0; 158.9; 158.6; 157.5; 156.3; 155.8;
150.2; 133.7; 129.9; 128.2; 127.4; 124.0; 121.6;
115.4; 114.5; 110.7; 110.5; 106.9; 98.1; 81.3; 78.9;
78.3; 72.3; 71.7; 70.5–70.1 (m); 69.5–69.2 (m);
61.4; 38.9; 37.4; 33.7; 33.3; 31.9; 30.1; 29.7-29.3
(m); 28.9; 28.5; 26.7; 26.1; 25.8; 23.1; 22.8; 22.7;
17.9; 14.2; 14.1; 8.4. IR (CHCl3): n 3302.1; 2924.3;
2854.6; 1724.4; 1677.9; 1606.3; 1467.5; 1366.1;
3.7. 1-(Tetrahydro-pyran-2-yloxy)-octadecan-9-
one (16)
The Grignard reagent from 1-bromo-8-(tetrahy-
dropyran-2-yloxy)-octane was prepared by adding
dropwise the bromide (5.86 g, 20.0 mmol) in 15
ml of anhydrous THF on magnesium turnings
(0.53 g, 22.0 mmol). The reaction was initiated by
gentle warming and the mixture was stirred for 30
min at 35°C before amide 15 in 20 ml of THF was
added dropwise. The solution was stirred at 40°C
for 1 h and the reaction was quenched by addition
of aqueous NH4Cl. The crude mixture was ex-
tracted with ether and the organic layer was dried
over sodium sulfate and reduced in 6acuo. The
residue was purified by recrystallization in
methanol to yield 16 (5.89 g, 81 %) as a white

S. Nuss et al. / Chemistry and Physics of Lipids 103 (1999) 21–35
31
1
grease. TLC: Rf 0.2 (hexane/ether 8:2). H-NMR
(200 MHz, CDCl3): d 4.58 (m, 1 H); 3.90–3.30 (m,
4 H); 2.38 (t, J=7.3 Hz, 4 H); 2.90-1.50 (m, 6 H);
1.50–1.26 (m, 26 H); 0.88 (t, J=6.3 Hz, 3 H).
13C-NMR (50 MHz, CDCl3): d 211.2; 98.8; 67.6;
62.2; 42.8; 31.6; 30.8; 29.7–29.2 (m); 26.2; 25.5;
23.9; 22.7; 19.7; 13.9. IR (CHCl3): n 2917.9;
2849.1; 1697.9; 1466.0; 1200.7; 1121.8; 1065.6;
1032.4. MS (CI/NH3): 386.1 [M+NH4]+.
MHz, CDCl3): d 4.57 (m, 1 H); 3.67–3.35 (m, 4
H); 1.70-1.28 (m, 42 H); 0.88 (t, J=6.3 Hz, 3 H).
13C-NMR (50 MHz, CDCl3): d 98.8; 74.3; 67.6;
62.2; 39.4; 39.1; 31.9; 31.5; 30.8; 30.2; 29.8; 29.5–
29.3 (m); 26.9; 25.7; 25.6; 25.3; 23.5; 23.3; 22.6;
19.6; 13.9. IR (CHCl3): n 3476.7; 2928.3; 2854.6;
1465.9; 1378.0; 1352.1; 1260.7; 1200.4; 1137.3;
1120.1; 1078.7; 1024.1; 987.0; 907.0; 869.8. MS
(CI/NH3): 444.5 [M+NH4]+.
3.8. 9-Butyl-1-(tetrahydro-pyran-2-yloxy)-
octadecan-9-ol (17a)
3.9. 9-Butyl-octadec-8-en-1-ol (18a)
Alcohol 17a (4.0 g, 9.4 mmol) in 50 ml of
methanol was stirred with Amberlyst® H-15 (0.5
g) for 3 h at 40°C. The solid was filtered off and
the filtrate was evaporated to dryness. The residue
was stirred again with amberlyst® H-15 (0.5 g) in
50 ml of refluxing dichloromethane for 16 h. After
a second filtration and removal of the solvent in
vacuo, compound 18a (2.6 g, 86%) was obtained
as a colorless oil by purification on silica gel
(hexane/ether 75:25) (mixture of three regioiso-
mers). TLC: Rf 0.6 (hexane/ethyl acetate 7:3).
1H-NMR (200 MHz, CDCl3): d 5.09 (t, J=6.8
Hz, 1 H); 3.64 (t, J=6.4 Hz, 2 H); 1.95 (m, 6 H);
Ketone 16 (7.37 g, 20.0 mmol) in 50 ml of
anhydrous THF was treated dropwise with n-
butyllithium (1.5 M in hexane, 16.0 ml, 24.0
mmol) at 0°C. The mixture was stirred at room
temperature for 16 h and the reaction was
quenched by addition of aqueous NH4Cl. The
solution was extracted with ether and the organic
layer was dried over sodium sulfate, reduced in
vacuo and chromatographed (hexane/ethyl ac-
etate 95:5 to 85:15) to afford 17a (6.48 g, 76%) as
a colorless oil (mixture of diastereomers). TLC: Rf
0.4 (hexane/ethyl acetate 9:1). 1H-NMR (200
Table 3
Yields for the synthesis of novobiocin-lipids and precursors
Yield, %
HRMSa
Compound
26
68
73
74
27
98
98
98
14
62
65
58
Formula
Calcd
Found
a
b
c
C80H129N3O18Nab
1442.9170
1420.9351
1559.0734
1537.0914
1615.1360
1593.1540
1615.6574
1587.0318
1565.0500
1442.9173
1420.9352
1559.0743
1537.0917
1615.1350
1593.1536
1615e
c
C80H130N3O18
C88H149N3O18Nab
c
C
88H150N3O18
C92H157N3O18Nab
c
C92H158N3O18
d
e
51
76
*90d
98
48
C72H98F17N3O18
54f
C88H145N3O20Nab
1587.0322
1565.0500
c
C
88H146N3O18
a Relative to compounds 14a–e.
b Corresponds to the [M+Na]+ ion.
c Corresponds to the [M+H]+ ion.
d Due to the presence of degradation products, compound 27d was chromatographed on silica gel (dichloromethane/methanol
98:2 to 90:10).
e High resolution mass spectra could not be satisfactorily measured for that compound. The ion observed in mass spectrometry
was [M]+
.
f Combined yield (27e–29: 57%; 29–14e: 95%).

32
S. Nuss et al. / Chemistry and Physics of Lipids 103 (1999) 21–35
1.57 (m, 2 H); 1.27 (m, 26 H); 0.89 (t, J=6.3 Hz,
6 H). 13C-NMR (50 MHz, CDCl3): d 139.7; 139.6;
139.4; 124.7; 124.5; 124.4; 63.0; 36.9; 36.6; 32.8;
31.9; 30.7; 30.5; 30.2; 30.1; 30.0; 29.8; 29.7; 29.6;
29.5; 29.4; 29.3; 28.5; 28.3; 27.6; 25.7; 23.2; 22.8;
22.7; 22.5; 14.1; 13.8. IR (CHCl3): n 3317.5;
2926.4; 2855.1; 1465.8; 1377.1; 1057.3; 880.1;
722.5. MS (CI/NH3): 342.3 [M+NH4]+.
3.12. 1-Benzyloxy-heneicos-12-en-3-ol (21a)
A Grignard reagent was prepared by adding
dropwise oleyl bromide (8.3 g, 25.0 mmol) in 20
ml of anhydrous ether on magnesium turnings
(0.66 g, 27.5 mmol) in 3 ml of ether. The reaction
was initiated by the addition of one drop of
1,2-dibromoethane. The mixture was stirred for 1
h before 3-benzyloxy-1-propanal (3.6 g, 22.0
mmol) in 15 ml of ether was added. Stirring was
maintained for 16 h at room temperature and the
reaction was stopped by the addition of aqueous
NH4Cl. The crude mixture was extracted with
ether and the organic layer was reduced in vacuo
and purified by chromatography (hexane/ether
7:3) to yield compound 21a (6.31 g, 69%) as a
3.10. 9-Butyl-octadecan-1-ol (19a)
Compound 18a (2.5 g, 7.7 mmol) and Pd/C
10% (0.2 g) in 10 ml of methanol were stirred
under hydrogen atmosphere for 16 h. The reac-
tion mixture was filtered over a celite pad and
evaporated to dryness to yield alcohol 19a (2.46 g,
98%) as a colorless oil. TLC: Rf 0.6 (hexane/ethyl
1
colorless oil. TLC: Rf 0.3 (hexane/ether 7:3). H-
1
acetate 7:3). H-NMR (200 MHz, CDCl3): d 3.65
NMR (200 MHz, CDCl3): d 7.33 (m, 5 H); 5.36
(m, 2 H); 4.54 (s, 2 H); 3.86-3.63 (m, 3 H); 2.01
(m, 4 H); 1.74 (m, 2 H); 1.44 (m; 2 H); 1.29 (m, 24
H); 0.89 (t, J=6.3 Hz, 3 H). 13C-NMR (50 MHz,
CDCl3): 137.8; 129.8; 128.4; 127.6; 127.5; 73.3;
71.4; 69.3; 37.4; 36.4; 31.9; 29.6–29.3 (m); 27.2;
25.6; 22.6; 14.1. IR (CHCl3): n 3457.0; 2924.5;
2853.5; 1454.5; 1364.2; 1095.3; 1027.7; 735.1;
697.2. MS (CI/NH3): 417.3 [M+H]+; 434.3
[M+NH4]+.
(t, J=6.5 Hz, 2 H); 1.57 (m, 2 H); 1.26 (m, 35 H);
0.89 (t, J=6.3 Hz, 6 H). 13C-NMR (50 MHz,
CDCl3): d 63.1; 37.5; 33.8; 33.5; 32.9; 31.9; 30.2;
30.1; 29.7–29.1 (m); 26.8; 25.8; 23.1; 22.7; 14.1;
14.0. IR (CHCl3): n 3337.5; 2923.0; 2855.5;
1466.0; 1377.8; 1058.0; 722.8. MS (CI/NH3): 344.3
[M+NH4]+.
3.11. 1-Bromo-9-butyl-octadecan (20)
Carbon tetrabromide (3.97 g, 12.0 mmol) was
added to triphenyl phosphine (6.29 g, 24.0 mmol)
in 100 ml of dichloromethane at 0°C and stirred
for 5 min. Alcohol 19a (3.26 g, 10.0 mmol) in 25
ml of dichloromethane was added and the mixture
was allowed to warm at room temperature under
stirring for 2 h. Hexane was added, the precipitate
was removed by filtration and the filtrate was
reduced in vacuo. The residue was triturated in
hexane and filtered on a small pad of silica gel to
afford the alkyl bromide 20 (3.8 g, 97%) as a
3.13. 2-[1-(2-Benzyloxy-ethyl)-nonadec-10-
enyloxy]-tetrahydro-pyran (22a)
Dihydropyran (0.5 ml, 5.5 mmol) was added to
alcohol 21a (2.08 g, 5.0 mmol) and PPh3,HBr (50
mg, 0.15 mmol) in 10 ml of dichloromethane at
0°C. The mixture was stirred for 16 h at room
temperature, neutralized by the addition of solid
Na2CO3, filtered over a celite pad and the solvent
was removed in vacuum. The residue was purified
over silica gel (hexane/ether 9:1) to afford com-
pound 22a (2.13 g, 85%) as a colorless oil (mix-
ture of diastereomers). TLC: Rf 0.4 (hexane/ether
1
colorless oil. TLC: Rf 0.9 (hexane). H-NMR (200
MHz, CDCl3): d 3.42 (t, J=6.8 Hz, 2 H); 1.87
(m, 2 H); 1.27 (m, 35 H); 0.89 (t, J=6.3 Hz, 6 H).
13C-NMR (50 MHz, CDCl3): d 37.9; 33.8; 33.5;
32.9; 31.9; 30.2; 30.0; 29.7-29.1 (m); 28.8; 28.2;
26.8; 23.2; 22.7; 14.1; 14.0. IR (CHCl3): n 2924.8;
2854.0; 1465.2; 1377.3; 1253.4; 722.5. MS (CI/
CH4): 309 [M-Br]+; 389 [M]+.
1
8:2). H-NMR (200 MHz, CDCl3): d 7.34 (m, 5
H); 5.36 (m, 2 H); 4.60 (m, 1 H); 4.50 (m, 2 H);
3.69–3.44 (m, 5 H); 2.01 (m, 4 H); 1.88–1.51 (m,
8 H); 1.28 (m, 26 H); 0.87 (t, J=6.3 Hz, 3 H).
13C-NMR (50 MHz, CDCl3): 138.7; 138.4; 130.3;

S. Nuss et al. / Chemistry and Physics of Lipids 103 (1999) 21–35
33
129.8; 128.3; 128.2; 127.7; 127.6; 127.5; 127.4;
98.5; 97.5; 74.8; 73.9; 73.0; 72.9; 35.6; 35.1; 34.1;
33.9; 32.6; 31.9; 31.1; 29.8; 29.7; 29.6; 29.5; 29.3;
29.0; 27.2; 25.5; 25.4; 24.9; 22.7; 20.0; 14.1. IR
(CHCl3): n 2925.8; 2853.6; 1469.4; 1360.0; 1259.5;
1200.5; 1112.8; 1022.6; 811.8; 736.6; 697.9. MS
(CI/NH3): 518.6 [M+NH4]+.
silica gel (hexane/ether/triethylamine 78:20:2 to
68:30:2) to yield compound 24a (0.96 g, 72%) as a
pale yellow oil (mixture of diastereomers). TLC:
Rf 0.5 (hexane/ether 5:5). 1H-NMR (200 MHz,
CDCl3): d 5.35 (m, 4 H); 4.62 (m, 2 H); 3.91-3.47
(m, 15 H); 2.01 (m, 8 H); 1.81–1.53 (m, 16 H);
1.28 (m, 52 H); 0.88 (t, J=6.3 Hz, 6 H). 13C-
NMR (50 MHz, CDCl3): d 129.9; 98.2; 97.6; 79.2;
78.6; 74.6; 73.8; 72.7–70.9 (m); 68.6–68.1 (m);
63.1; 62.8; 35.6; 35.0; 34.0; 32.6; 31.9; 31.2; 29.7–
29.3 (m); 27.2; 25.5; 25.0; 22.7; 20.0; 14.1. IR
(CHCl3): n 3457.0; 2924.5; 2853.5; 1454.5; 1364.2;
1206.9; 1095.3; 1027.7; 735.1; 697.2. MS (CI/
NH3): 895.2 [M+NH4]+.
3.14. 3-(Tetrahydro-pyran-2-yloxy)-heneicos-
12-en-1-ol (23a)
Liquid ammonia (100 ml) was condensed at
−40°C in a flask containing benzyl ether 22a
(2.51 g, 5.0 mmol) and 4 ml of anhydrous THF.
Sodium (230 mg, 9.6 mmol) was introduced by
portions till a steady deep blue color of the solu-
tion. The mixture was stirred for 30 min before 2
ml of ethanol were added to decompose sodium in
excess. The solution was allowed to warm to
room temperature to remove ammonia and the
residue was washed with water and extracted with
ether. The organic layer was reduced in vacuo and
purified by chromatography (hexane/ether/triethy-
lamine 58:40:2) to afford primary alcohol 23a
(1.98 g, 96%) as a colorless oil. TLC: Rf 0.3
3.16. 2-[2-(2-{2-[Bis-(9-butyl-octadecyloxy)-
methoxy]-ethoxy}-ethoxy)-ethoxy]-ethylazide
(26b)
A sodium hydride suspension (60% in oil, 68
mg, 1.69 mmol) was added to alcohol 13 (400 mg,
0.56 mmol) and HMPA (0.2 ml) in 2 ml of
anhydrous THF at 0°C. The mixture was refluxed
for 1 h, cooled to 0°C before methanesulfonyl
ester 25 (336 mg, 1.13 mmol) was added, and
refluxed again for 16 h. The solution was decom-
posed by the addition of aqueous NH4Cl and
extracted with ethyl acetate. The organic layer
was reduced in vacuo and the residue was purified
on silica gel (hexane/ethyl acetate 8:2) to yield
compound 26b (387 mg, 73%) as a pale yellow oil
(mixture of diastereomers). TLC: Rf 0.5 (hexane/
1
(hexane/ether 6:4). H-NMR (200 MHz, CDCl3):
d 5.35 (m, 2 H); 4.48 (m, 1 H); 3.95–3.44 (m, 5 H);
2.00 (m, 4 H); 1.82–1.50 (m, 8 H); 1.28 (m, 26 H);
0.88 (t, J=6.3 Hz, 3 H). 13C-NMR (50 MHz,
CDCl3): d 129.7; 100.0; 75.1; 64.9; 59.2; 37.0; 35.2;
32.5; 31.8; 31.3; 29.7–29.1 (m); 27.1; 25.3; 25.0;
22.6; 21.3; 14.0. IR (CHCl3): n 3408.8; 2926.3;
2853.4; 1729.8; 1463.6; 1378.1; 1351.9; 1200.2;
1160.9; 1134.0; 1075.1; 1023.8; 999.6; 899.5; 868.7;
808.5; 722.9. MS (CI/NH3): 428.1 [M+NH4]+.
1
ethyl acetate 7:3). H-NMR (200 MHz, CDCl3): d
3.77–3.36 (m, 25 H); 1.55 (m, 4 H); 1.26 (m, 70
H); 0.89 (t, J=6.3 Hz, 12 H). 13C-NMR (50
MHz, CDCl3): d 78.4; 71.6; 70.8; 70.7; 70.5; 70.0;
69.7; 50.6; 37.3; 33.7; 33.3; 31.9; 30.1; 29.6; 29.5;
29.3; 28.9; 26.7; 26.1; 23.1; 22.7; 14.1; 14.0. IR
(CHCl3): n 2924.3; 2855.0; 2102.2; 1465.7; 1377.2;
1351.0; 1300.9; 1120.2; 722.5. MS (CI/NH3): 928.4
[M+NH4]+.
3.15. 1,3-Bis-[3-(tetrahydro-pyran-2-yloxy)-
heneicos-12-enyloxy]-propan-2-ol (24a)
A sodium hydride suspension (60% in oil, 0.24
g, 6.0 mmol) was added by portions to alcohol
23a (1.24 g, 3.0 mmol) and HMPA (1 ml) in 10 ml
of anhydrous THF at 0°C. The mixture was
refluxed for 1 h, cooled to 0°C before epibromo-
hydrin (0.13 ml, 1.5 mmol) was added, and
refluxed again for 16 h. The solution was decom-
posed by the addition of aqueous NH4Cl, and
extracted with ether. The organic layer was re-
duced in vacuo and the residue was purified on
3.17. 2-[2-(2-{2-[Bis-(9-butyl-octadecyloxy)-
methoxy]-ethoxy}-ethoxy)-ethoxy]-ethylamine
(27b)
Compound 26b (120 mg, 0.13 mmol) in 2 ml of
anhydrous ether was stirred with lithium alu-

34
S. Nuss et al. / Chemistry and Physics of Lipids 103 (1999) 21–35
Brezesinski, G., Dobner, B., Do¨rfler, H.D., Fischer, M., Haas,
S., Nuhn, P., 1987. Influence of a-branched fatty acid
chains on the thermotropic behaviour of 1-O-acyl-2-O-
hexadecyl glycerophosphocholines. Chem. Phys. Lipids 43,
257–264.
Brezesinski, G., Rettig, W., 1990. Vergleichende Unter-
sucherungen an Mono- und Bischichtsystemen von drei
methylverzweigten Phosphatidyl-cholinen. Z. Chem. 30,
373–374.
Brezesinski, G., Rettig, W., Grunewald, S., Kuschel, F., Hor-
vath, L.I., 1989. A mono- and bilayer study of homologous
branched-chain lecithins. Liquid Cryst. 5, 1677–1690.
Chiu, W., Schmid, M.F., Prasad, B.V.V., 1993. Teaching
electron diffraction and imaging of macromolecules. Bio-
phys. J. 64, 1610–1625.
Darst, S.A., Ahlers, M., Meller, P.H., Kubalek, E.W.,
Blankenburg, R., Ribi, H.O., Ringsdorf, H., Kornberg,
R.D., 1991. Two-dimensional crystals of streptavidin on
biotinylated lipid layers and their interaction with biotiny-
lated macromolecules. Biophys. J. 59, 387–396.
Darst, S.A., Ribi, H.O., Pierce, D.W., Kornberg, R.D., 1988.
Two-dimensional crystals of Escherichia coli RNA poly-
merase holoenzyme on positively charged lipid layers. J.
Mol. Biol. 203, 269–273.
minium hydride (15 mg, 0.39 mmol) for 30 min at
room temperature. A few drops of water were
added and the resulting mixture was dried over
sodium sulfate, filtered and reduced in vacuo to
yield amine 27b (114 mg, 98%) as a colorless oil
(mixture of diastereomers). TLC: Rf 0.2
(dichloromethane/methanol 9:1). 1H-NMR (200
MHz, CDCl3): d 3.78–3.43 (m, 23 H); 2.97 (t,
J=5.1 Hz, 2 H); 1.57 (m, 4 H); 1.27 (m, 70 H);
0.89 (t, J=6.3 Hz, 12 H). 13C-NMR (50 MHz,
CDCl3): d 78.4; 71.6; 70.7; 70.6; 70.5; 70.3; 69.9;
37.4; 33.7; 33.3; 31.9; 30.1; 29.7; 29.6; 29.5; 29.3;
28.9; 26.7; 26.1; 23.1; 22.7; 14.2; 14.1. IR (CHCl3):
n 3542.0; 2924.5; 2854.7; 1465.4; 1125.8. MS (CI/
NH3): 884.7 [M+H]+.
Acknowledgements
The authors are thankful to Elf-Atochem for
kindly providing starting perfluorinated chemi-
cals, and to Alain Valleix for running mass spec-
tra at chemical ionization. This work was
supported by Institut Scientifique Roussel and
Ministe`re de l’Education Nationale, de l’Enseigne-
ment Supe´rieur et de la Recherche (France).
Degering, E.F., Boatright, L.G., 1950. Studies on the synthesis
of lysine. J. Am. Chem. Soc. 72, 5137–5139.
Elbert, R., Folda, T., Ringsdorf, H., 1984. Saturated and
polymerizable amphiphiles with fluorocarbon chains. In-
vestigation in monolayers and liposomes. J. Am. Chem.
Soc. 106, 7687–7692.
Fehrentz, J.A., Castro, B., 1983. An efficient synthesis of
optically active a-(t-butoxycarbonylamino)-aldehydes from
a-amino acids. Synthesis, 676–678.
References
Fromherz, P., 1971. Electron microscopic studies of lipid
protein film. Nature 231, 267–268.
Altenburger, J.M., Lebeau, L., Mioskowski, C., Schirlin, D.,
1992. Synthesis of new glycerolipids linked to hydroxamate
derivatives designed for two-dimensional crystallization of
aminopeptidase M. Helv. Chim. Acta 75, 2538–2544.
Asgharian, B., Balthasar, D.M., Cadenhead, D.A., Rice, D.K.,
1991. Mixed monomolecular film behavior of methyl iso-
and anteiso-branched-chain and v-cyclohexyl phos-
phatidylcholines with cholesterol. Langmuir 7, 272–276.
Balthasar, D.M., Cadenhead, D.A., Lewis, R.N.A.H., McEl-
haney, R.N., 1988. Monomolecular film behavior of
Gaentzler, S., Vierling, P., 1993. Perfluoroalkylated phospho-
cholines, improved synthesis, surface activity, fluorocarbon
emulsifying capability and biological properties. New J.
Chem. 17, 337–344.
Gaines, G.L., Jr., 1966. Insoluble monolayers at liquid gas
–
interfaces. In: Prigogine, I. (Ed.), Interscience, New York,
London and Sydney.
Glaeser, R.M., 1985. Electron crystallography of biological
macromolecules. Annu. Rev. Phys. Chem. 36, 243–275.
Groß, U., Ru¨diger, S., 1994. Asymmetric lamellar phospho-
lipid aggregates bearing fluorocarbons. J. Fluorine Chem.
69, 31–34.
methyl
anteiso-branched-chain
phosphatidylcholines.
Langmuir 4, 180–186.
Barton, S.W., Goudot, A., Bouloussa, O., Rondelez, F., lin,
B., Novak, F., Acero, A., Rice, S.A., 1992. Structural
Hauser, H., 1991. Effect of inorganic cations on phase transi-
tions. Chem. Phys. Lipids 57, 309–325.
transitions in
a monolayer of fluorinated amphiphile
Izawa, M., 1952. Monolayers of branched-chain fatty acids. II.
a-Aromatic Substituted Fatty Acids 25, 314–318.
Jap, B.K., Zulauf, M., Scheybani, T., Hefti, A., Baumeister,
W., Aebi, U., Engel, A., 1992. 2D Crystallization: from art
to science. Ultramicroscopy 46, 45–84.
molecules. J. Chem. Phys. 96, 1343–1351.
Blokzjil, W., Engberts, B.F.N., 1993. Hydrophobic effects.
Opinion and facts. Angew Chem. Int. Ed. Engl. 32, 1545–
1579.
Boekema, E.J., 1990. The present state of two-dimensional
crystallization of membrane proteins. Electron Microsc.
Rev. 3, 87–96.
Kannenberg, K., Blume, A., McElhaney, R.N., Poralla, K.,
1983. Monolayer and calorimetric studies of phosphatidyl

S. Nuss et al. / Chemistry and Physics of Lipids 103 (1999) 21–35
35
cholines containing branched-chain fatty acids and of their
interactions with cholesterol and with a bacterial hopanoid
in model membrane. Biochim. Biophys. Acta 733, 111–
116.
methyloctadecanoic acid. Colloids Surf. 41, 287–301.
McDaniel, R.V., McIntosh, T.J., Simon, S.A., 1983. Nonelec-
trolyte substitution for water in phosphatidylcholine bilay-
ers. Biochim. Biophys. Acta 731, 97–108.
Ku¨hlbrandt, W., 1988. Three-dimensional crystallization of
membrane proteins. Q. Rev. Biophys. 21, 429–477.
Lebeau, L., Mioskowski, C., Oudet, P., 1988. Systematic study
of phospholipids linked to a steroid derivative, spread into
a monolayer at the air/water interface. Biochim. Biophys.
Acta 939, 417–429.
Lebeau, L., Olland, S., Oudet, P., Mioskowski, C., 1992.
Rational design and synthesis of phospholipids for the
two-dimensional crystallization of DNA gyrase, a key ele-
ment in chromosome organization. Chem. Phys. Lipids 62,
93–103.
Menger, F.M., Wood, M.G.J., Richardson, S., Zhou, Q.,
Elrington, A.R., Sherrod, M.J., 1988. Chain-substituted
lipids in monolayer films. A study of molecular packing. J.
Am. Chem. Soc. 110, 6797–6803.
Mosser, G., Brisson, A., 1991. Conditions of two-dimensional
crystallization of cholera toxin B-subunit on lipid films
containing ganglioside GM1. J. Struct. Biol. 106, 191–198.
Mukerjee, P., Yang, A.Y.S., 1976. Nonideality of mixing of
micelles of fluorocarbon and hydrocarbon surfactants and
evidence of partial miscibility from differential conduc-
tance data. J. Phys. Chem. 80, 1388–1390.
Lebeau, L., Oudet, P., Mioskowski, C., 1991. Synthesis of new
phospholipids linked to steroid-hormone derivatives de-
signed for two-dimensional crystallization of proteins.
Helv. Chim. Acta 74, 1697–1706.
Lebeau, L., Regnier, E., Schultz, P., Wang, J.C., Mioskowski,
C., Oudet, P., 1990. Two-dimensional crystallization of
DNA gyrase B-subunit on specifically designed lipid mono-
layers. FEBS Lett. 267, 38–42.
Lebeau, L., Schultz, P., Celia, H., Mesini, P., Nuss, S.,
Klinger, C., Olland, S., Oudet, P., Mioskowski, C., 1996.
Specifically designed lipid assemblies as tools for two-di-
mensional crystallization of soluble biological macro-
molecules. In: Barenholz, Y., Lasic, D.D. (Eds.),
Handbook of Non-Medical Applications of Liposomes,
vol. II. Boca Raton, CRC Press, pp. 153–186.
Muller, N., 1990. Search for a realistic view of hydrophobic
effects. Acc. Chem. Res. 23, 23–28.
Nahm, S., Weinreb, S.M., 1981. N-Methoxy-N-methylamides
as effective acylating agents. Tetrahedron Lett. 22, 3815–
3818.
Reece, R.J., Maxwell, A., 1991. DNA gyrase: Structure and
function. Crit. Rev. Biochem. Mol. Biol. 26, 335–375.
Rice, D.K., Cadenhead, D.A., Lewis, R.N.A.H., McElhaney,
R.N., 1987. A comparative monomolecular film study of a
straight-chain
phosphatidylcholine
(dipalmitoylphos-
phatidylcholine) with three iso-branched-chain phos-
phatidylcholines (diisoheptadecanoylphosphatidyl-choline,
diiso-octanoylphosphatidylcholine,
and
diisoei-
cosanoylphosphati-dylcholine). Biochemistry 26, 3205–
3210.
Lewis, E.N., Bittman, R., Levin, I.W., 1986. Methyl group
substitution at C(1), C(2) or C(3) of the glycerol backbone
of a diether phosphocholine: A comparative study of bi-
layer chain disorder in the gel and liquid-crystalline phases.
Biochim. Biophys. Acta 861, 44–52.
Lewis, R.N.A.H., Sykes, B.D., McElhaney, R.N., 1987. Ther-
motropic phase behavior of model membrane composed of
phosphatidylcholines containing dl-methyl anteiso-
branched fatty acids. 1. Differential scanning calorimetric
and 31P NMR spectroscopic studies. Biochemistry 26,
4036–4044.
Sanderson, P.W., Lis, L.J., Quinn, P.J., Williams, W.P., 1991.
The Hofmeister effect in relation to membrane lipid phase
stability. Biochim. Biophys. Acta 1067, 43–50.
Santaella, C., Vierling, P., Riess, J.G., 1991. Highly stable
liposomes derived from perfluoroalkylated glycerophos-
phocholines. Angew. Chem. Int. Ed. Engl. 30, 567–568.
Scherer, P.G., Seelig, J., 1989. Electric charge effects on phos-
pholipid headgroups. Phosphatidylcholine in mixtures with
cationic and anionic amphiphiles. Biochemistry 28, 7720–
7728.
Scho¨n, I., 1984. Sodium-liquid ammonia reduction in peptide
chemistry. Chem. Rev. 84, 287–297.
Lohner, K., 1991. Effect of small organic molecules on phos-
pholipid phase transitions. Chem. Phys. Lipids 57, 341–
362.
Shin, S., Collazo, N., Rice, S.A., 1992. A molecular dynamics
study of the packing structures in monolayers of partially
fluorinated amphiphiles. J. Chem. Phys. 96, 1352–1365.
So¨derman, O., Arvidson, G., Lindblom, G., Fontell, K., 1983.
The interactions between monovalent ions and phos-
phatidyl cholines in aqueous bilayers. Eur. J. Biochem.
134, 309–314.
Ulman, A., 1991. An Introduction to Ultrathin Organic Films
— From Langmuir–Blodgett to Self-Assembly. Academic
Press, San Diego.
Mantsch, H.H., Madec, C., Lewis, R.N.A.H., McElhaney,
R.N., 1987. Thermotropic phase behavior of model mem-
brane composed of phosphatidylcholines containing dl-
methyl anteisobranched fatty acids. 2. An infrared
spectroscopy study. Biochemistry 26, 4045–4049.
Matuo, H., Cadenhead, D.A., 1989. Mixed monolayers of
straight and branched chain fatty acids. Part I: Mixed
monolayers of n-tridecanoic, n-tetradecanoic and n-pen-
tadecanoic acids with the iso-fatty acids 13-methyl-tetrade-
canoic acid, 14-methylpentadecanoic acid, 15-methyl-
hexadecanoic acid, 16-methylheptadecanoic acid and 17-
Weitzel, G., Fretzdorff, A.M., Heller, S., 1951. Monomolecu-
lar film of several methylated stearic acids. Z. Physiol.
Chem. 288, 200–215.