Synthesis and mesogenic properties of a Y-shaped glyco-glycero-lipid

Article abstract of DOI:10.1016/j.chemphyslip.2004.03.011

We synthesised a new glycoglycerolipid [1,3-di-O-(β-D-glucopyranosyl)- 2-deoxy-2-amino-N-palmitoyl-sn-glycerol] with a Y-shaped structure bearing two sugar head groups and only one fatty acid chain. Instead of an ester linkage between the glycerol and the fatty acid an amido function was introduced. The mesogenic properties were investigated using polarising microscopy and are discussed with respect to similar compounds. The lyotropism was measured using the contact preparation method and small-angle neutron-scattering.

Full text of DOI:10.1016/j.chemphyslip.2004.03.011

Chemistry and Physics of Lipids 131 (2004) 51–61
Synthesis and mesogenic properties of a Y-shaped
glyco-glycero-lipid
G. Milkereita, V.M. Garamusb, K. Veermansa, R. Willumeitb, V. Villa,∗
a
Institute of Organic Chemistry, University of Hamburg, Martin-Luther-King-Platz 6, 20146 Hamburg, Germany
b
GKSS Research Centre, Max Planck Str., Geesthacht 21502, Germany
Received 30 September 2003; received in revised form 24 March 2004; accepted 31 March 2004
Abstract
We synthesised a new glycoglycerolipid [1,3-di-O-(␤-d-glucopyranosyl)-2-deoxy-2-amino-N-palmitoyl-sn-glycerol] with a
Y-shaped structure bearing two sugar head groups and only one fatty acid chain. Instead of an ester linkage between the glycerol
and the fatty acid an amido function was introduced. The mesogenic properties were investigated using polarising microscopy
and are discussed with respect to similar compounds. The lyotropism was measured using the contact preparation method and
small-angle neutron-scattering.
© 2004 Elsevier Ireland Ltd. All rights reserved.
Keywords: Glyco-glycero-lipids; Amphiphiles; Lyotropic cubic phases; Small-angle neutron-scattering
1. Introduction
structure (Fig. 1b) were reported (Minden et al.,
2002a; Fischer et al., 1994).
Glyco-glycero-lipids are known to be impor-
tant constituent of many cell membranes. Natu-
rally occurring glyco-glycero-lipids normally have a
Y-shaped structure (Fig. 1a) with a polar carbohy-
drate head group linked to a 1,2- or 1,3-di-O-acyl
or alkyl-sn-glycerol (Ishizuka and Yamakawa, 1985;
Minden et al., 2002a,b; Mannock et al., 1987, 1990,
2001; Hinz et al., 1996; Six et al., 1983; Kuttenreich
et al., 1988; Nagatsu et al., 1994; Minamikawa et al.,
1994). But only a few compounds with an inverse
All of these amphotropic molecules form both
thermotropic liquid crystalline phases in their pure
state upon heating and lyotropic polymorphism can
be found on the addition of solvent. The driving force
of the mesophase formation in case of amphiphiles
is a micro phase separation leading to an aggregate
structure with separated regions for the lipophilic and
hydrophilic molecular parts.
The structures of the common lyotropic and ther-
motropic aggregates have been investigated exten-
sively (Curatolo, 1987; Prade et al., 1995; Blunk
et al., 1998; Jeffrey and Wingert, 1992).
The synthesis of glycolipids is often very tedious,
requires multistep synthesis, and the introduction of
other functional groups than alkoxy or esters (e.g.
∗
Corresponding author. Tel.: +49-40-42838-4269;
fax: +49-40-42838-4325.
E-mail address: vill@liqcryst.chemie.uni-hamburg.de (V. Vill).
0009-3084/$ – see front matter © 2004 Elsevier Ireland Ltd. All rights reserved.
doi:10.1016/j.chemphyslip.2004.03.011

52
G. Milkereit et al. / Chemistry and Physics of Lipids 131 (2004) 51–61
HO
O
HO
HO
R
O
HO
HO
O
O
O
HO
OH
O
HO
O
O
OR
R
OH
HO
HO
(a)
OH
(b)
Fig. 1. (a) Normal Y-shaped glyco-glycero-lipid and (b) inverted structure.
phosphate, nitrogen, sulphur) makes the synthesis of-
ten more complicated.
tion reaction. The desired product was obtained in a
yield of 15%, besides this an amount of ∼3% mono-
glycosylated product could be separated from the re-
action mixture, however, the reaction time exceeded
18 h, but no ␣-product was detected.
We report about the easy synthesis of a new
glyco-glycero-lipid with an inverted Y-shaped struc-
ture. An amido group substituted the oxygen at the
carbon-2 in order to make the molecule more polar.
The thermotropic and the lyotropic properties are dis-
cussed; the results from small-angle-neutron-scattering
are presented.
One could expect that it would make more sense
using other glycosylation procedures, e.g. for the syn-
thesis of the lipid 2-O-oleoyl-1,2-di-O-␤-d-glucopyra-
nosyl-sn-glycerol the key glycosylation step was
carried out using a variation of the Königs–Knorr pro-
cedure (Helferich and Ost, 1962), or using the trichlo-
racetimidate procedure (Schmidt and Stumpp, 1983)
for the synthesis of 1,3-diglycosylated 2-acyl-glycerol
esters (Milkereit and Vill, unpublished), due to higher
yields. But this holds not true for this glycosylation
reaction. The problem seems to be a lower reactivity
of the amino alcohol, because even using the trichlo-
racetimidate procedure, the yield was around 15–16%
and two more synthetic steps for the synthesis of the
imidate are necessary.
In the next step the cbz group was removed by
hydrogenation. The reaction was somehow tedious
and the product had to be purified by silica gel chro-
matography. The amino glycerol 3 was reacted with
palmitoyl chloride using the method described by
Masuda et al. (2000), however, dichlormethane was
used as solvent instead of N,N-dimethylformamide.
After chromatographic purification compound 4 was
obtained in 45% yield. In the last step the com-
pound was deacetylated using the Zemplen proce-
dure, leading to the desired lipid N-palmitoyl-1,3-di-
O-␤-d-glucopyranosyl-2-deoxy-2-amino-sn-glycerol
(5) in 94% yield.
2. Results and discussion
2.1. Synthesis
Fig. 2 shows the synthetic route used for the prepa-
ration of N-palmitoyl-1,3-di-O-(␤-d-glucopyranosyl)-
2-deoxy-2-amino-sn-glycerol.
In an earlier paper we reported the synthesis of
an 1,3-di-O-glucosyl-2-O-acyl-sn-glycerol (Minden
et al., 2002a) that required a multistep synthesis
with several protecting and deprotecting steps. The
synthesis of the title compound should afford less
synthetic steps. We started with commercial avail-
able 2-amino-1,3-propanediol (Serinol). For glyco-
sylation the amino function had to be protected.
We used in this case the benzyloxycarbonyl group
(cbz) as protective group. On treatment of 1 with
benzyl chloroformate according to the procedure
described by Harada et al. (1996) we obtained
2-benzyloxycarbonylamino-1,3-propanediol in 40%
yield. 1,2,3,4,6-penta-O-acetyl-␤-d-glucopyranoside
was prepared according to the literature (Fischer and
Helferich, 1911).
The glycosylation reaction was carried using the
method described by Vill et al. (1986), but instead of
tin tetrachloride boron trifluoride etherate, a more ef-
ficient Lewis acid for the synthesis of ␤-glycosides
(Ferrier and Furneaux, 1976) was used for glycosyla-
2.2. Thermotropic properties
Table 1 shows the thermotropic properties of the
synthesised compound, together with the data of some

G. Milkereit et al. / Chemistry and Physics of Lipids 131 (2004) 51–61
53
HO
NH2
O
HO
HO
OH
HO
OH
OH
CbzCl, K2CO3,
H2O
NaOAc,
Ac2O
AcO
HO
O
O
AcO
AcO
HO
OAc
N
O
H
OAc
1
.
BF3 Et2O,
CH2Cl2
AcO
OH
O
NHCbz
AcO
AcO
HO
OH
OH
O
O
O
OAc
2
H2, Pd/C
MeOH/THF
AcO
OH
O
NH2
AcO
AcO
HO
OH
OH
O
O
O
OAc
3
Palmitoyl chloride,
CH2Cl2, C5H5N
AcO
AcO
O
O
AcO
O
OAc
O
AcO
AcO
N
H
C15H31
AcO
O
OAc
O
4
NaOMe,
MeOH
HO
HO
O
HO
O
OH
O
HO
HO
N
H
C15H31
HO
O
OH
5
Fig. 2. Synthesis of N-palmitoyl-1,3-di-O-␤-d-glucopyranosyl-2-deoxy-2-amino-sn-glycerol.

54
G. Milkereit et al. / Chemistry and Physics of Lipids 131 (2004) 51–61
Table 1
Thermotropic properties of the synthesised lipid, together with some literature data for comparison (A, B: Schmidt and Jankowski, 1996;
Schmidt and Vill, unpublished results; C: Fischer et al., 1994; D: Minden et al., 2002a)
similar mesogenes for comparison (A, B: Schmidt and
Jankowski, 1996; Schmidt and Vill, unpublished re-
sults; C: Fischer et al., 1994; D: Minden et al., 2002a).
It depicts clearly the influence of the type of linkage
between the alkyl chain and the glycerol moiety, the
chain length and type of chain.
The type of chain has the strongest effects on the
melting point. An unsaturation in the hydrophobic part
of the molecule (D) leads to liquid crystallinity even
at ambient temperature, due to a disturbed packing
of the chains. Compounds A, B and C with saturated
alkyl chains show no liquid crystal phases at ambient
temperature. Hence the kind of linkage has for these
three compounds no remarkable influence on the melt-
ing point. For compound 5 the influence of the kind of
linkage is more significant. Compared to D one would
expect for a palmitoyl ester a much higher melting
temperature, but the palmitoyl amid 5 has a very un-
usual melting behaviour, due to a much more complex
hydrogen bond network, that differs from the ester.
The compound is in glass state at ambient temperature,
with a phase transitions temperature to the columnar

G. Milkereit et al. / Chemistry and Physics of Lipids 131 (2004) 51–61
55
phase that could not be clearly determined. On cool-
ing the phase transition from the columnar phase to
the glass state could only be seen in the “breaking”
of the glass at low temperatures. The length of the
hydrophobic chain can be seen in the different kind
of phases occurring with varying chain length. For
compound B and C, increasing the chain length with
two methylene groups leads to a second phase. For
compounds A, D and 5 the polymorphism increases
with increasing chain length (the type of linkage is
not considered), starting with only a columnar-discotic
phase for the short chain compound to a columnar,
cubic and lamellar phase diagram for the C18-chain.
Compound 5 forms a columnar phase of type I, the
hydrophilic/hydrophobic interface curved towards the
hydrocarbon chain region. As a fact of frustration, a
cubic phase is formed between 65 and 134 ◦C, since
the columnar phase is already destabilised due to the
increased volume of the fluid hydrocarbon chain re-
gion. A SA phase cannot be formed, because the hy-
drocarbon chain length is to short to accommodate the
hydrophilic headgroup, which is too broad compared
to the hydrocarbon chain. Compound D becomes at
higher temperatures rod-like, and is able to form a
SA phase, but the fluidity maximum of the hydropho-
bic chain of compound 5 is not enough to form a SA
phase.
In the next step we investigated the structure of
the lyotropic aggregates with small-angle neutron-
scattering. For the investigation of micellar structures
this method provides significant structural informa-
tion (He et al., 2002), and for the investigation of
discontinous cubic phases information of the size and
shape of the micelles in isotropic micellar phases it is
helpful to unravel the structure of the cubic phase.
2.4. Analysis of SANS data
2.4.1. Model-independent approach
The pair distribution function p(r) of scattering
length density gives in many cases information about
the overall shape and dimension of micelles (Glatter
and Kratky, 1982). Because the scattering data is only
measured in a finite range, a traditional Fourier trans-
formation cannot be performed, instead an indirect
Fourier transformation can be used (Glatter, 1977;
Pedersen, 1997). The p(r) function is approximated
by linear combination of a finite number of cubic
B-spline functions, giving the expression:
N
ꢀ
i=1
p(r) =
aiϕi(r)
(1)
After calculating the p(r) function we are able to calcu-
late integral parameters of the micelle, like dΣ(0)/dW
expressing the scattering at the so called “zero” angle
and Rg the radius of gyration (Glatter, 1977; Glatter
and Kratky, 1982), which is given by:
2.3. Lyotropic properties
The lyotropism was measured using the contact
preparation technique. Compound 5 displayed in the
contact with water the phase sequence:
ꢁ
2
D
max p(r)r2 dr
0
R2g =
(2)
ꢁ
D
max p(r) dr
0
(100% water)
cub
HI
(100% lipid)
with an upper limit for the maximum particle size of
Dmax
The occurrence of the hexagonal phase with the hy-
drophilic sugar moieties curved towards the water re-
gion was expected, more interesting is the cubic phase.
For compound D even two cubic phases were found
(Minden et al., 2002a), but the structure was not fur-
ther investigated.
The cubic phase can have two possible shapes,
discontinuous and bicontinous phases. Whereas the
discontinous cubic phases are based upon various
packing of spherical or slightly anisotropic micelles,
the bicontinous cubic phases with interwoven fluid
structures are based upon underlying infinite periodic
minimal surfaces (Fairhurst et al., 1998).
.
2.4.2. Modelling
To obtain specific structural information about the
micelle it is necessary to fit the experimental data by
the different models of micellar structures. Looking at
the results from the contact preparation, it seemed to
be enough to use a spherical and an ellipsoidal model
to fit the data.
The differential cross-sections of neutron-scattering
per unit volume by an ensemble of monodisperse
spherical-like objects with negligible interaction can

56
G. Milkereit et al. / Chemistry and Physics of Lipids 131 (2004) 51–61
be written as:
0,35
dΣ(q)
dΣ(0)
dΩF2(q) + Binc
=
(3)
0,30
0,25
0,20
0,15
0,10
0,05
0,00
-0,05
dΩ
where dΣ(0)/dW depends on the concentration of mi-
celles (n), the volume of micelle (V) and the difference
in the scattering length densities between micelles and
solvent leading to the equation:
dΣ(0)
= n(ꢀρꢁ − ρs)2V2
(4)
dΩ
ꢀρꢁ and ρs describe the average of the neutron-scattering
length densities of aggregates and solvent, respec-
tively, and Binc stands for the residual scattering
background.
0
20
40
60
80
r [Å]
Fig. 3. Pair distribution function (for c = 7.0 × 10−4 g mL−1).
The form factor F2(q) expresses the scattering
cross-section of one particle. Eq. (5) stands for the
form factor for spherical and elliptical micelles
The pair distance distribution function calculated by
the IFT method is shown in Fig. 3. It gives an indica-
tion that a spherical shape is more probable, due to its
nearly Gaussian form, therefore, the models for spher-
ical and ellipsoidal micelles were fitted to the data. The
fitting results are shown in Fig. 4, showing the model
fit for spherical micelles and the data, the fit for ellip-
soidal micelles is not shown. The application of ellip-
soidal model gives the value of parameter ε is equal
to 1, i.e. spherical shape. The structural results calcu-
lated from the model fitting results are summarised in
Table 2.
ꢂ
π/2
ꢃ
ꢄ2
F2(q) =
f(q, β) sin β dβ
(5)
0
In the case of an ellipsoid with three semiaxes a, a,
εa the function f(q,β) can be written as:
3[sin x − x cos x]
f(q, β) =
(6)
x3
where β is the angle between the vector q and the axis
of the particle and x = qb[sin2 β + ε2 cos2 β]1/2. For
ε = 1, Eq. (6) describes the scattering of spherical
aggregates with a = b = r.
The models described above were applied to fit the
experimental data.
104
103
102
101
2.5. Structural analysis
Two different concentrations were measured (0.07
and 0.1 wt.%). At both besides the smaller aggregates
some larger aggregates are formed, exceeding the mea-
surement range of SANS. On further increasing the
concentration the smaller aggregates stay nearly the
same size, therefore, the large aggregates have to grow,
but it is not observable on the present data due to lim-
ited q-range.
From this point of view, it seems to be useful to anal-
ysis only the data for the medium concentration, be-
cause the aggregates have for all concentrations nearly
the same size, and at medium concentration the smaller
aggregate can be investigated besides the larger aggre-
gate.
0.7 mg/mL
1 mg/mL
100
0,01
0,1
q, Å-1
Fig. 4. Scattering data for c = 7.0×10−4 g mL−1 (empty squares)
and c = 10.0 × 10−4 g mL−1 (solid triangles) together with the
model fit for spherical micelles (solid line).

G. Milkereit et al. / Chemistry and Physics of Lipids 131 (2004) 51–61
57
Table 2
the polar group, and therefore, the length of molecule.
From another point of view, the radius of the micelle
obtained from SANS analysis is the average value,
which is affected by the polydispersity of the aggre-
gates and by the fluctuation of the micellar surface.
That is why the obtained value of the micellar radius
(36 Å), which is higher due to the above mentioned
effects of hydration, polydispersity and fluctuation of
surface, is in good agreement with the length of the
“dry molecule” (31 Å).
Results of SANS data analysis by model independent approach
(IFT), Rg and dΣ(0)/dΩ/c, model fitting, r and calculated pa-
rameter of aggregates, i.e. radius of equivalent sphere
Radius of gyration, Rg (Å)
Radius of equivalent sphere (Å)
29.2
37.7
0.2
0.3
Radius of sphere from model fitting, r (Å) 36
Scattering at “zero”-angle, 172
dΣ(0)/dΩ/c (cm2 g−1
Apparent molar mass of micelle,
Mapp (g M−1
Radius of equivalent sphere (Å)
2
1
)
(8.4
0.4) × 104
)
29.6
1.5
On the other hand there is some disagreement
between the radius of equivalent sphere calculated
from radius of gyration (Rg) and the scattering at
“zero”-angle (dΣ(0)/dΩ/c). It is straight forward
to get the mass and volume of the micelle from
dΣ(0)/dΩ/c. In the present calculations we have
taken into account that eight H-atoms of hydroxyl
functions of the sugar head groups are under H–D
exchange. The value of equivalent sphere obtained
from dΣ(0)/dΩ/c is equal to 29.6 Å which is signif-
icant lower than the values obtained from Rg (IFT)
and model fitting. This difference can be explained
by the beginning of the phase separation, i.e. the co-
existence of two different kinds of aggregates (large
particles of the hexagonal phase and small spherical
micelles). The Scattering data were normalised by
total concentration of surfactant but in the interval of
the scattering vector q (0.01–0.25 Å−1) analysed by
IFT method only the scattering from the small spher-
ical aggregates is significant. That is why, actual val-
ues of the micellelar mass (volume) should be larger
and the corresponding radius of equivalent sphere
also. Taking into account these difficulties one can
conclude the observation of spherical micelles with
a radius of 36–38 Å and an aggregation number of
about 200.
The fitting of the data shows another interesting re-
sult of the experiments. It can be seen, that the fitted
curve is close to the experimental data. But it is re-
markable that at low q values, the curves of the model
fit and the experimental data diverge, and this discrep-
ancy is cannot be explained as an experimental er-
ror. We assume that in this experiment we observed
the formation of a non-micellar aggregate besides the
spherical micelles. The lowest part of experimental
data (q < 0.01 Å−1) was analysed by power law
dΣ(q)
∼ q−α
.
(7)
dΩ
The obtained value of α is equal to 2.7 0.1 which
points on the formation of aggregates with a relation
between the mass (M) and linear sizes (r) as M∼r2.7
for a length scale of r > 300 Å.
This change of micellar form is similar to a sec-
ond or higher order phase transition, but it is not
the same as the main liquid crystalline phase tran-
sition. However, from the neutron-scattering data
it was not possible to obtain information about the
space group of a possible cubic phase, this informa-
tion may be provided by X-ray diffraction experi-
ments.
The results in Table 2 describe the local structure
of the micelle in a solution of heavy water. The model
fits showed that the micelles are spherical with the ra-
dius in the range of 36–38 Å. The results of model
independent analysis and model fitting give the con-
sistent values. The radius of micelle corresponds to
the length of the molecule, which is equal to ∼31 Å.
This molecule length is the value for the length of the
“dry molecule”, under real conditions the hydration of
the carbohydrate headgroup increases the volume of
3. Summary
A new synthetic strategy was employed success-
ful to synthesise an unusual glyco-glycero-lipid. The
mesogenic properties are strongly influenced by the
amide linkage of the fatty acid chain to the hydrophilic
part of the molecule. The investigation of the lyotropic
structure by small-angle neutron-scattering revealed
an interesting lyotropism. The micellar structure was
determined and the size of the micelle was calculated

58
G. Milkereit et al. / Chemistry and Physics of Lipids 131 (2004) 51–61
by standard procedures. Another interesting observa-
tion is the phase transition that seems to occur in the
measured concentration range. The micelles stay the
same size, whereas with increasing concentration all
molecules are used for building up the new phase. In-
teresting is that many inverted Y-mesogenes exhibit
thermotropic and lyotropic cubic phases. The struc-
ture enables the formation of cubic structures, perhaps
this structure facilitates the formation of cubic phases
in biological processes.
conventional procedures (Wignall and Bates, 1986).
The two-dimensional isotropic scattering spectra were
azimuthally average, converted to an absolute scale,
and corrected for detector efficiency by dividing by the
incoherent scattering spectra of pure water (Wignall
and Bates, 1986) which was measured with a 1-mm
path-length quartz cell (Hellma).
The average excess scattering length density per unit
mass "ρm of the surfactant in deuterated water was
determined from the known chemical composition, it
was equal to "ρm = −4.297 × 1010 cm g−1 for the
full protonated and in the case of H–D exchange of
4. Experimental section
eight H atoms "ρm = −3.528 × 1010 cm g−1
.
4.1. Polarising microscopy
4.3. Materials
An Olympus BH optical polarising microscope
equipped with a Mettler FP 82 hot stage and a Mettler
FP 80 central processor was used to identify thermal
transitions and characterise anisotropic textures.
For the contact preparation a small amount of sam-
ple was placed on a microscope slide and covered with
a cover glass before heating to the SA phase. After-
wards a small amount of solvent was placed on the
slide at the edge of the cover glass. As soon as the sol-
vent had moved under the cover glass and the sample
was completely surrounded with the solvent, it was
placed again for some seconds on the hot stage at a
temperature of 100–120 ◦C and the phase behaviour
was investigated immediately afterwards by polarising
microscopy.
Dry solvents were obtained from Fluka or Merck,
all other solvents were freshly distilled before use.
Deuterium oxide was purchased from Deutero.
Thin-layer chromatography was performed on silica
gel (Merck GF254), and detection was effected by
spraying with a solution of: (a) ethanol–sulphuric
acid (9:1) or (b) ethanol–molybdic phosphoric acid
(99:1) followed by heating. Column chromatog-
raphy was performed using silica gel 60 (Merck,
0.063–0.200 nm, 230–400 mesh). NMR spectra
were recorded on a Bruker AMX 400 or a Bruker
DRX 5001 spectrometer (mc: centred multiplett; d:
doublet; t: triplet; dd: double doublet; dt: double
triplet).
4.2. Small-angle neutron-scattering
4.4. Synthesis
Small-angle neutron-scattering experiments were
made with the SANS-1 experiment at the FRG1 re-
search reactor at GKSS research Centre, Geesthacht,
Germany (Stuhrmann et al., 1995). Four sample-to-
detector distances (from 0.7 to 9.7 m) were employed
to cover the range of scattering vectors q from 0.005
to 0.25 Å−1. The neutron wavelength λ was 8.1 Å
with a wavelength resolution of 10% (full-width at
half-maximum value).
The samples were kept in quartz cells (Hellma) with
a path length of 5 mm. The samples were placed in a
thermostated sample holder, for isothermal conditions.
The raw spectra were corrected for the background
from the solvent, sample cell, and other sources, by
4.4.1. N-(Carbonylbenzyloxy)-2-amino-propane-
1,3-diol (1)
Benzyl chloroformate (8.4 mL; 10.2 g; 60 mmol)
was added dropwise to a stirred solution of 5 g
(54.8 mmol) Serinol (2-amino-1,3-propanediol) and
8.3 g (60 mmol) potassium carbonate in 40 mL of wa-
ter at 5 ◦C. The mixture was stirred for an additional
hour at 5 ◦C and then extracted four-times with 50 mL
ethyl acetate. The combined organic extracts were
dried over magnesium sulphate and solvent was re-
moved in vacuo. The residue was recrystallised from
acetone/n-hexane. Yield: 4.93 g (40%).
1H NMR (400 MHz, pyridine-d5): δ = 5.07–5.15
(m, 5H, H-arom.), 4.92 (s, 2H, CH2-benzyl.),

G. Milkereit et al. / Chemistry and Physics of Lipids 131 (2004) 51–61
59
4.15–4.22 (m, 2H, H-1a, H-3a), 4.06–4.14 (m, 2H,
4.4.3. 1,3-Di-O-(2,3,4,6-tetra-O-acetyl-β-d-
H-1b, H-3b), 3.92–4.02 (m, 1H, H-2).
glucopyranosyl)-2-deoxy-2-amino-sn-glycerol (3)
A mixture of 682 mg (0.77 mmol) 2 and 40 mg Pd/C
(10%) in 100 mL methanol/tetrahydrofurane (1:1 v/v)
was stirred under 1 atm of hydrogen. After 2 days
TLC showed the reaction to be complete. The catalyst
was removed by filtration. The mother liquor evapo-
rated in vacuo and the resulting residue purified by
silica gel chromatography (chloroform–methanol 9:1).
Yield: 567 mg (97%).
4.4.2. N-(Carbonylbenzyloxy)-1,3-di-O-
(2,3,4,6-tetra-O-acetyl-β-d-glucopyranosyl)-2-
deoxy-2-amino-sn-glycerol (2)
1,2,3,4,6-Penta-O-acetyl-␤-d-glucopyranosid (11.7
g; 30 mmol) and 2.3 g (10 mmol) N-carbonylben-
zyloxy-2-amino-1,3-propanediol were dissolved in
150 mL anhydrous dichlormethane under an atmo-
sphere of dry nitrogen. Boron trifluoride etherate
(3.81 mL; 4.3 g; 30 mmol) was added dropwise at
0 ◦C and the solution was stirred 18 h at ambient
temperature. After the reaction had been quenched
with a saturated solution of sodium hydrogen carbon-
ate, the organic layer was separated and the aqueous
layer extracted two-times with dichlormethane. The
combined organic phases were washed twice with
water (100 mL), dried over magnesium sulphate and
solvent was removed in vacuo. The residue was puri-
fied by silica gel chromatography (ethyl acetate–light
petroleum, bp 50–70 ◦C, 2:1). Yield: 1.30 g (15%).
1H NMR (500 MHz, CDCl3+TMS): δ = 7.28–7.37
(m, 5H, H-arom.), 5.19, 5.20 (je dd,1H, H-3ꢂ, H-3ꢂꢂ),
5.07, 5.08 (je dd, 1H, H-4ꢂ, H-4ꢂꢂ), 4.99, 5.01 (je dd,
1H, H-2ꢂ, H-2ꢂꢂ), 4.58, 4.62 (je d, 1H, CH2-benzyl.),
4.52, 4.55 (je d, 1H, H-1ꢂ, H-1ꢂꢂ), 4.24, 4.26 (je dd, 1H,
H-6aꢂ, H-6aꢂꢂ), 4.10, 4.14 (je dd, 1H, H-6bꢂ, H-6bꢂꢂ),
3.96, 3.88 (je dd, H-1a, H-3a), 3.59–3.70 (m, 4H,
H-1b, H-3b, H-5ꢂ, H-5ꢂꢂ), 3.41 (mc, 1H, H-2), 2.07,
2.06 (je s, 3H, OAc), 2.02 (s, 6H, OAc), 2.00 (s, 9H,
1H NMR (400 MHz, CDCl3 + TMS): δ = 5.19
(dd, 2H, H-3ꢂ, H-3ꢂꢂ), 5.07, 5.08 (je dd, 1H, H-4ꢂ,
H-4ꢂꢂ), 4.99 (mc, 2H, H-2ꢂ, H-2ꢂꢂ), 4.52, 4.54 (je d,
1H, H-1ꢂ, H-1ꢂꢂ), 4.23–4.26 (m, 2H, H-6aꢂ, H-6aꢂꢂ),
4.11, 4.13 (je dd, 1H, H-6bꢂ, H-6bꢂꢂ), 3.58–3.91
(m, 6H, H-1a, H-3b, H-1b, H-3b, H-5ꢂ, H-5ꢂꢂ), 3.09
(mc, 1H, H-2), 2.08, 2.05 (je s, 3H, OAc), 2.01 (s,
6H, OAc), 2.00, 1.98, 1.95, 1.93 (je s, 3H, OAc);
3
3
3
3
3
3
ꢂ
ꢂ
ꢂ
J1ꢂ,2 = 8.1 Hz, J3ꢂ,2 = 9.7 Hz, J3ꢂ,4 = 9.7 Hz,
J4ꢂ,5 = 9.7 Hz, 3J5ꢂ,6b = 2.6 Hz, 2J6a ,b = 12.2 Hz,
ꢂ
ꢂ
ꢂ
ꢂ
3
3
ꢂꢂ
ꢂꢂ ꢂꢂ
ꢂꢂ ꢂꢂ
J1ꢂꢂ,2 = 8.1 Hz, J3 ,2 = 8.1 Hz, J3 ,4 = 9.7 Hz,
3
ꢂꢂ ꢂꢂ
ꢂꢂ ꢂꢂ
J5 ,6 = 2.6 Hz, J6a ,b = 12.2 Hz.
4.4.4. N-(Palmitoyl)-1,3-di-O-(2,3,4,6-tetra-O-
acetyl-β-d-glucopyranosyl)-2-deoxy-2-amino-
sn-glycerol (4)
Compound 3 (560 mg; 0.74 mmol) and 75 ␮L
(73.2 mg; 0.9 mmol) pyridine were dissolved in 25 mL
anhydrous dichlormethane. At 0 ◦C a solution of
224 ␮L (203 mg; 0.74 mmol) palmitoyl chloride in
1 mL anhydrous dichlormethane was added slowly.
Afterwards the solution was stirred 30 min at 0 ◦C and
overnight at ambient temperature. The organic phase
was washed twice with a saturated solution of sodium
hydrogen carbonate (10 mL) and once with water
(5 mL), dried over magnesium sulphate, and evapo-
rated in vacuo. The residue was purified by silica gel
chromatography (ethyl acetate–light petroleum, bp
50–70 ◦C 2:1). Yield: 330 mg (45%).
3
3
ꢂ
ꢂ
OAc), 1.94 (s, 3H, OAc); J1ꢂ,2 = 8.1 Hz, J2ꢂ,3
=
3
3
3
ꢂ
ꢂ
ꢂ
9.7 Hz, J3ꢂ,4 = 9.7 Hz, J4ꢂ,5 = 9.7 Hz, J5ꢂ,6a
=
3
2
ꢂ
ꢂ
ꢂ
4.6 Hz, J5ꢂ,6b
=
2.6 Hz, J6a ,b
=
12.2 Hz,
3
3
3
ꢂꢂ
ꢂꢂ
ꢂꢂ
J1ꢂꢂ,2 = 8.1 Hz, J2ꢂꢂ,3 = 9.7 Hz, J3ꢂꢂ,4 = 9.7 Hz,
3
3
3
ꢂꢂ
ꢂꢂ
ꢂꢂ
J4ꢂꢂ,5 = 9.7 Hz, J5ꢂꢂ,6a = 4.6 Hz, J5ꢂꢂ,6b
=
=
2
3
2
ꢂꢂ ꢂꢂ
2.6 Hz, J6a ,b = 12.2 Hz, J1a,2 = 3.1 Hz, J1a,b
3
3
10.7 Hz, J3a,2 = 3.1 Hz, J3a,b = 10.7 Hz.
13C NMR (125 MHz, CDCl3 + TMS): δ = 172.63
=
(C O, amid), 170.64, 170.28, 170.24, 169.98, 169.35,
1H NMR (500 MHz, CDCl3 + TMS): δ = 5.19,
5.21 (je dd, 1H, H-3ꢂ, H-3ꢂꢂ), 5.07, 5.08 (je dd, 1H,
H-4ꢂ, H-4ꢂꢂ), 4.99, 5.02 (je dd, 1H, H-2ꢂ, H-2ꢂꢂ), 4.53,
4.56 (je d, 1H, H-1ꢂ, H-1ꢂꢂ), 4.24, 4.26 (je dd, 1H,
H-6aꢂ, H-6aꢂꢂ), 4.10, 4.13 (je dd, 1H, H-6bꢂ, H-6bꢂꢂ),
3.89, 3.96 (je dd, 1H, H-1a, H-3a), 3.59–3.69 (m, 4H,
H-1b, H-3b, H-5ꢂ, H-5ꢂꢂ), 3.41 (mc, 1H, H-2), 2.21
(m, 2H, Alkyl-␣-CH2), 2.07, 2.05 (je s, 3H, OAc),
=
169.26 (C O, OAc), 138.32 (Cquart-arom.), 128.42,
127.74, 127.59 (C-arom.), 101.21, 101.14 (C-1ꢂ,
C-1ꢂꢂ), 72.8, 72.70 (C-3ꢂ, C-3ꢂꢂ), 72.41 (C-benzyl.),
71.88 (C-5ꢂ, C-5ꢂꢂ), 71.38, 71.29 (C-2ꢂ, C-2ꢂꢂ), 70.35,
69.22 (C-1, C-3), 68.39 (C-4ꢂ, C-4ꢂꢂ), 61.88 (C-6ꢂ,
C-6ꢂꢂ), 52.03 (C-2), 20.74, 20.60, 20.18 (–CH3,
OAc).

60
G. Milkereit et al. / Chemistry and Physics of Lipids 131 (2004) 51–61
2.03 (s, 6H, OAc), 2.00 (s, 9H, OAc), 1.94 (s, 3H,
OAc), 1.49–1.59 (m, 2H, Alkyl-␤-CH2), 1.19–1.31
(m, 24H, Alkyl-CH2), 0.85 (t, 3H, Alkyl-CH3);
References
Blunk, D., Praefke, K., Vill, V., 1998. Amphotropic liquid crystals.
In: Demus, D., Goodby, J., Gray, G.W., Spiess, H.-W., Vill,
V. (Eds.), Handbook of Liquid Crystals, vol. 3. Wiley-VCH,
Weinheim, pp. 305–340.
Curatolo, W., 1987. The physical properties of glycolipids.
Biochim. Biophys. Acta 779, 381–401.
Fairhurst, C.E., Fuller, S., Gray, J., Holmes, M.C., Tiddy, G.J.,
1998. Lyotropic surfactant liquid crystals. In: Demus, D.,
Goodby, J., Gray, G.W., Spiess, H.-W., Vill, V. (Eds.),
Handbook of Liquid Crystals, vol. 3. Viley-VCH, Weinheim,
pp. 341–392.
Ferrier, R.J., Furneaux, R.H., 1976. Synthesis of 1,2-trans-related
1-thioglycoside esters. Carbohydr. Res. 52, 63–68.
Fischer, E., Helferich, B., 1911. Über neue synthetische glycoside.
Liebigs Ann. Chem. 383, 68–91.
Fischer, S., Fischer, H., Diele, S., Pelzl, G., Jankowski, K.,
Schmidt, R.R., Vill, V., 1994. On the structure of the
thermotropic cubic mesophase. Liq. Cryst. 17, 855–861.
Glatter, O., 1977. Data evaluation of small-angle scattering:
calculation of the radial electron density distribution by means
of indirect Fourier transformation. Acta Phys. Aust. 47, 83–
102.
3
3
2
3
3
ꢂ
ꢂ
ꢂ
J1ꢂ,2 = 8.1 Hz, J2ꢂ,3 = 9.7 Hz, J3ꢂ,4 = 9.7 Hz,
3
3
ꢂ
ꢂ
ꢂ
J4ꢂ,5 = 9.7 Hz, J5ꢂ,6a = 4.6 Hz, J5ꢂ,6b = 2.6 Hz,
3
3
ꢂ
ꢂ
ꢂꢂ ꢂꢂ
ꢂꢂ ꢂꢂ
J6a ,b = 12.2 Hz, J1 ,2 = 8.1 Hz, J2 ,3
=
=
9.7 Hz, 3J3 ,4 = 9.7 Hz, 3J4 ,5 = 9.7 Hz, 3J5ꢂꢂ,6a
ꢂꢂ ꢂꢂ
ꢂꢂ ꢂꢂ
ꢂꢂ
3
2
ꢂꢂ
ꢂꢂ ꢂꢂ
4.6 Hz, J5ꢂꢂ,6b = 2.6 Hz, J6a ,b = 12.2 Hz,
3J1a,2 = 3.1 Hz, J1a,b = 10.7 Hz, J3a,2 = 3.1 Hz,
2
3
3J3a,b = 10.7 Hz.
13C NMR (125 MHz, CDCl3 + TMS): δ = 172.51
=
(C O, amid), 170.63, 170.29, 170.18, 169.70, 169.35,
ꢂ
ꢂꢂ
=
169.27 (C O, OAc), 101.25, 101.19 (C-1 , C-1 ),
72.86, 72.72 (C-3ꢂ, C-3ꢂꢂ), 71.87 (C-5ꢂ, C-5ꢂꢂ), 71.38,
71.30 (C-2ꢂ, C-2ꢂꢂ), 70.35, 69.22 (C-1, C-3), 68.45
(C-4ꢂ, C-4ꢂꢂ), 61.84 (C-6ꢂ, C-6ꢂꢂ), 52.08 (C-2), 25.96,
24.83, 24.69, 24.17, 23.18, 22.70, 21.18 (Alkyl-CH2),
20.74, 20.61, 20.19 (–CH3, OAc), 14.08 (Alkyl-CH3).
4.4.5. N-(Palmitoyl)-1,3-di-O-(β-d-glucopyranosyl)-
2-deoxy-2-amino-sn-glycerol (5)
Glatter, O., Kratky, O. (Eds.), 1982. Small-Angle-X-ray Scattering.
Academic Press, London.
Compound 4 (325 mg; 0.33 mmol) was dissolved in
25 mL dry methanol and sodium methoxide was added
(pH 8–9). The solution was stirred at ambient temper-
ature until TLC revealed the reaction to be complete.
The reaction mixture was neutralised using Amberlyst
IR 120 ion exchange resign (protonated form), fil-
trated and evaporated in vacuo. The product was re-
crystallised from methanol. Yield: 201 mg (94%).
1H NMR (400 MHz, d4-methanol): δ = 4.33, 4.34
(je d, 1H, H-1ꢂ, H-1ꢂꢂ), 4.04, 4.09 (je d, 1H, H-1a,
H-3a), 3.81–3.94 (m, 4H, H-1b, H-3b, H-6aꢂ, H-6aꢂꢂ),
3.65–3.73 (m, 2H, H-6bꢂ, H-6bꢂꢂ), 3.26–3.42 (m, 7H,
H-2, H-3ꢂ, H-4ꢂ, H-5ꢂ, H-3ꢂꢂ, H-4ꢂꢂ, H-5ꢂꢂ), 3.22, 3.23
(je dd, 1H, H-2ꢂ, H-2ꢂꢂ), 2.39 (t, 2H, Alkyl-␣-CH2),
Harada, H., Morie, T., Hirokawa, Y., Kato, S., 1996. An efficient
synthesis of 6-substituted aminohexahydro-1H-1,4-diazepines
from 2-substituted aminopropenals. Chem. Pharm. Bull.
44 (12), 2205–2212.
He, L., Garamus, V.M., Funari, S.S., Malfois, M., Willumeit,
R., Niemeyer, B., 2002. Comparison of small-angle scattering
methods for the structural analysis of octyl-␤-maltopyranoside
micelles. J. Phys. Chem. B 106, 7596–7604.
Helferich, B., Ost, W., 1962. Synthese einiger ␤-d-xylopyranoside.
Chem. Ber. 95, 2612–2615.
Hinz, H.-J., Tenchova, R., Tenchov, B., Quinn, P.J.,
1996. Lamellar-non-lamellar phase transitions in synthetic
glycoglycerolipids studied by time-resolved X-ray diffraction.
Liq. Cryst. 20, 469–482.
Ishizuka, I., Yamakawa, T., 1985. Glycolipids. In: Wiegandt, H.
(Ed.), New Comprehensive Biochemistry, vol. 10. Elsevier,
pp. 101–197.
Jeffrey, G.A., Wingert, L.M., 1992. Carbohydrate liquid crystals.
Liq. Cryst. 12, 179–202.
1.59–1.67 (m, 2H, Alkyl-␤-CH2), 1.30–1.41 (m,
3
ꢂ
ꢂ
24H, Alkyl-CH2), 0.93 (t, 3H, Alkyl-CH3); J1 ,2
=
=
=
3
3
3
ꢂ
ꢂ
ꢂꢂ ꢂꢂ
8.1 Hz, J2 ,3 = 9.2 Hz, J1 ,2 = 8.1 Hz, J1a,2
Kuttenreich, H., Hinz, H.-J., Inczedy-Marcsek, M., Koynova, R.,
Technov, B., Laggner, P., 1988. Polymorphism of synthetic
2
3
3
4.1 Hz, J1a,b = 11.2 Hz, J3a,2 = 5.7 Hz, J3a,b
11.1 Hz.
1,2-di-O-alkyl-3-O-␤-d-galactopyranosyl-sn-glycerols
of
different alkyl chain length. Chem. Phys. Lipids 47, 245–260.
Mannock, D.A., Lewis, R.N.A.H., McElhaney, R.N., 1987. An
improved
procedure
for
the
preparation
of
1,2-di-O-acyl-3-O-(␤-d-glucopyranosyl)-sn-glycerols. Chem.
Phys. Lipids 43, 113–127.
Acknowledgements
Mannock, D.A., Lewis, R.N.A.H., McElhaney, R.N., 1990. The
chemical synthesis and physical characterisation of 1,2-di-O-
acyl-3-O-(␤-d-glucopyranosyl)-sn-glycerols, an important class
of membrane glycolipids. Chem. Phys. Lipids 55, 309–321.
We are grateful to the Deutsche Forschungsgemein-
schaft (SFB 470, Graduiertenkolleg 464) and the Gly-
coverein for financial support.

G. Milkereit et al. / Chemistry and Physics of Lipids 131 (2004) 51–61
61
Mannock, D.A., Harper, P.E., Gruner, S.M., McElhaney, R.N.,
2001. The physical properties of glycosyl diacylglycerols.
Calorimetric, X-ray diffraction and Fourier transform
spectroscopic studies of a homologous series of 12-di-O-acyl-
3-O-(␤-d-galactopyranosyl)-sn-glycerols. Chem. Phys. Lipids
111, 139–161.
Masuda, M., Vill, V., Shimizu, T., 2000. Conformational and
thermal phase behaviour of oligomethylene chains constrained
by carbohydrate hydrogen-bond networks. J. Am. Chem. Soc.
122, 12327–12333.
anti-tumor-promoting activity relationship of monogalactosyl
diacylglycerols. Bioorg. Med. Chem. Lett. 4, 1619–1622.
Pedersen, J.S., 1997. Analysis of small-angle scattering data from
colloids and polymer solutions: modelling and leastsquares
fitting. Adv. Colloid Interf. Sci. 70, 171–210.
Prade, H., Miethchen, R., Vill, V., 1995. Thermotrop flüssig-
kristalline Kohlenhydrat-amphiphile. J. Prakt. Chem. 337, 427–
440.
Schmidt, R.R., Stumpp, M., 1983. Synthese von 1-thioglycosiden.
Liebigs Ann. Chem. 1249–1256.
Minamikawa, H., Murakami, T., Hato, M., 1994. Synthesis of
1,3-di-O-alky-2-O-(␤-d-glycosyl)glycerols bearing oligosacc-
harides as hydrophilic groups. Chem. Phys. Lipids 72, 111–
118.
Minden, V.H.M., Morr, M., Milkereit, G., Heinz, E., Vill, V., 2002a.
Synthesis and mesogenic properties of glycosyl diacylglycerols.
Chem. Phys. Lipids 114, 55–80.
Minden, V.H.M., Milkereit, G., Vill, V., 2002b. Effects of
carbohydrate headgroups on the stability of induced cubic
phases in binary mixtures of glycolipids. Chem. Phys. Lipids
120, 45–56.
Schmidt, R.R., Jankowski, K., 1996. New types of nonionic
surfactants with sugar head groups. Liebigs Ann. 4, 867–879.
Six, L., Rueß, K.-P., Liefländer, M., 1983. An efficient synthesis of
12-O-dialkyl-3-O-␤-d-glycosyl-sn-glycerols. Tetrahedron Lett.
24, 1229–1232.
Stuhrmann, H.B., Burkhardt, N., Dietrich, G., Jünemann, R.,
Meerwinck, W., Schmitt, M., Wadzack, J., Willumeit, R., Zhao,
J., Nierhaus, K.H., 1995. Nucl. Instrum. Methods 356, 133–
137.
Vill, V., Böcker, T., Thiem, J., Fischer, F., 1986. Studies on liquid
crystalline glycosides. Liq. Cryst. 6, 349–356.
Nagatsu, A., Watanabe, M., Ikemoto, K., Hashimoto, M.,
Murakami, N., Sakakibara, J., Tokuda, H., Nishino, H.,
Iwashima, A., Yazawa, K., 1994. Synthesis and structure–
Wignall, G.D., Bates, F.S., 1986. Absolute calibration of
small-angle neutron-scattering data. J. Appl. Crystallogr. 20,
28–40.