Supramolecular receptors from α-amino acid-derived lipids

Article abstract of DOI:10.1039/a903956b

Anionic lipids derived from L-, D-, DL-glutamic acids, L-aspartic acid, L-lysine, L-ornithine, and L-2,4-diaminobutyric acid were prepared and the way in which structurally related, solvatochromic cationic styryl dyes were incorporated into these lipid ag

Full text of DOI:10.1039/a903956b

View Article Online / Journal Homepage / Table of Contents for this issue
Supramolecular receptors from ꢀ-amino acid-derived lipids
Hiroshi Hachisako,*a Yutaka Murataa and Hirotaka Iharab
a Department of Applied Chemistry, Kumamoto Institute of Technology, 4-22-1 Ikeda,
Kumamoto 860-0082, Japan
b Department of Applied Chemistry & Biochemistry, Faculty of Engineering,
Kumamoto University, 2-39-1 Kurokami, Kumamoto 860-8555, Japan
Received (in Cambridge, UK) 8th May 1999, Accepted 2nd September 1999
Anionic lipids derived from -, -, -glutamic acids, -aspartic acid, -lysine, -ornithine, and -2,4-diaminobutyric
acid were prepared and the way in which structurally related, solvatochromic cationic styryl dyes were incorporated
into these lipid aggregates were investigated. The - or -glutamic acid-derived lipids with glutaric acid headgroups
aggregated to form specific hydrophobic cavities which exhibited inclusion ability for the styryl dyes mainly based
on the planarity recognition. The formation of such specific hydrophobic cavities can be achieved not only by
introducing amide groups capable of complementary hydrogen bondings between neighbouring lipids in the
aggregates but also by introducing appropriate spacer methylenes into, for example, glutarate headgroups.
Side-chain methylenes of the amino acid residue were also found to play a significant role in the formation of
specific hydrophobic cavities as well as the spacer methylenes. Such cavities were also formed for appropriately
designed -lysine- and -ornithine-derived lipids. These results indicate that the specific incorporation is not
peculiar to the glutamic acid-derived lipids but a general phenomenon for the aggregates from appropriately
designed -amino acid-derived lipids. A difference in the manner of assembly of -glutamic acid residues between
-, -isomers, and -mixtures in the lipid aggregates is suggested not only by the λmax shift of incorporated styryl
dyes but also by aggregate morphologies as evidenced by the TEM observations. These results suggest that the
two-dimensional assembly of pure enantiomeric α-amino acid residues in appropriately designed lipids can
produce specific hydrophobic cavities unless the -mixture phase-separates into two enantiomeric components.
ference in only one methylene group is crucial for the formation
of secondary structures.
Introduction
Considerable attention has been focused on inclusion com-
pounds and their molecular designs. Since the inclusion abilities
of the conventional inclusion compounds such as cyclodextrins,
calixarenes, cyclophanes, are restricted mostly by their primary
structures (covalently linked macrocyclic structure), the size
and the kind of guest molecule may be restricted to a consider-
able extent. Our interests have been focused on the construction
of inclusion compounds by using self-assembly of compon-
ent amphiphiles. In our previous papers,1–4 we have reported
that certain -glutamic acid-derived anionic lipids formed
specific hydrophobic cavities for cationic dyes. The cavities were
able to incorporate the cationic dyes due mainly to molecular
planarity of the dyes.3,4 However, the specific behaviors seemed
to be peculiar to the ester-type glutamate lipids such as -9. In
order to generalize the inclusion behaviour by a wide variety
of self-assembling compounds, it is necessary to extend the
molecular structure from the -glutamic acid-derived lipids to
other kinds of α-amino acid-derived lipids. It is also necessary
to gain insight into appropriate molecular designs of com-
ponent lipids for the supramolecular receptors.
On the other hand, it is known that the side-chains of
poly(amino acid)s are crucial for formation of their secondary
structures, e.g. α-helix and β-structure in water.5–11 For example,
poly(-lysine) (PLL) can adopt the α-helix and β-structure in
water, whereas the helix content of poly(-ornithine) (PLO),
which has a shorter side-chain by one methylene than PLL, is
considerably lower under the same conditions. Similar tenden-
cies were observed for the following pairs, e.g., poly(-glutamic
acid) (PLGA) and poly(-aspartic acid) (PLAA) which has a
shorter side-chain by one methylene than PLGA, and PLO and
poly(2,4-diaminobutyric acid) (PLDBA), which has a shorter
side-chain by one methylene than PLO. In these cases, the dif-
We are interested in how α-amino acid residues play import-
ant roles when assembled in specific ways such as head-to-
head orientation of the α-amino acid-containing lipids. The
effect of the side-chain on the intermolecular interaction
between assembled amino acid residues would be very signifi-
cant. Therefore, we have been investigating the construction of
specific hydrophobic cavities which exhibit inclusion behaviour
by means of self-assembly of appropriately designed lipids
containing single α-amino acid residues. The side-chain effect
of poly(amino acid)s, by which their secondary structures are
remarkably affected, would become apparent in some way by
two-dimensionally assembled single α-amino acid residues in
supramolecular lipid aggregates.
In this paper, we prepared various kinds of α-amino acid-
derived anionic lipids with plural amide groups per molecule
(-glutamic acid-derived lipids 1–7, -aspartic acid-derived
lipid 8, -lysine-derived lipids 11–13, -ornithine-derived lipids
14 and 15, and -2,4-diaminobutyric acid-derived lipid 16) in
order to investigate the effect of α-amino acid residues and
chemical structure requirements on formation of the specific
hydrophobic cavities in water, because it had previously been
confirmed that the -glutamic acid-derived lipids (such as
-1 and -2) with three amide groups per molecule can
form complementary hydrogen bonds between neigbouring
lipids when assembled with head-to-head orientation.12,13,17–21
The effect of the amino acid residues on the inclusion and
the molecular recognition were investigated using the structur-
ally related cationic styryl dyes (stilbazolium derivatives),
4-[4-(dimethylamino)styryl]-N-methylpyridinium iodide (abbre-
viated to St-4C1 hereafter), 2-[4-(dimethylamino)styryl]-N-
methylpyridinium iodide (abbreviated to St-2C1 hereafter), and
2-[4-(dimethylamino)styryl]-N-ethylpyridinium iodide (abbre-
J. Chem. Soc., Perkin Trans. 2, 1999, 2569–2577
This journal is © The Royal Society of Chemistry 1999
2569

View Article Online
When lipid -1 is added to the aqueous solution of St-4C1,
λmax shifted to 478 nm, comparable to λmax in methanol at 20 ЊC.
This indicates that St-4C1 is incorporated into the hydrophobic
microenvironment, that is, near the glutamate residue in the
lipid aggregates in cooperation with hydrophobic and electro-
static interactions.1 This conclusion is based on previous
results.1–4 When the temperature is raised, aggregates of -1
undergo a gel-to-liquid crystalline phase transition (peak-top
temperature, 43 ЊC by DSC) accompanied by a hypsochromic
shift of λmax of St-4C1. Further increases in temperature up to
65 ЊC, at which the -1 aggregates are in a liquid crystalline
state, lead to a λmax of 450 nm (Tables 1 and 2). This wavelength
is comparable to the λmax of St-4C1 alone in water. This indi-
cates that the microenvironment around the electrostatically
bound and incorporated St-4C1 becomes more hydrophilic. A
similar mechanism accounts for the hypsochromic shifts
observed for other lipids in Table 1 with an increase in temper-
ature (see footnotes of Tables 1 and 2). Note that, among the
lipids in Table 1, aggregates of -3 induced the largest
bathochromic shift in λmax to 510 nm at 20 ЊC. This indicates
that the most hydrophobic microenvironment was produced
around St-4C1 by self-assembly of -3. However, no similar
bathochromic shifts were induced by lipids 4–8. These results
indicate that not only ester-type -91,3,4 but also certain kinds of
amide-type lipids such as -1, -1, and -3 can produce specific
hydrophobic cavities by self-assembly. Also note that slight
changes in chemical structures considerably affect the polarity
of hydrophobic cavities produced by self-assembly of these
-glutamic acid-derived lipids.
Effect of complementary hydrogen bonds between
neighbouring lipids on formation of hydrophobic cavities
It is believed to be reasonable that the lipid -10 with two
dodecyl chains leads to relatively looser packing of component
lipids than that of -9 with two octadecyl chains.1,4 In fact, as
shown in Table 1, this resulted in no bathochromic shift of the
λmax of St-4C1 in -10 aggregates at 20 ЊC in contrast with a
remarkable bathochromic shift in the -9 aggregates system.
This indicates that the St-4C1 is located in the polar micro-
environment in -10 aggregates. However, it is noted that the
substitution of amide groups (-1) for the corresponding ester
groups (-10) resulted in the bathochromic shift of St-4C1. This
indicates that the St-4C1 is located in the less polar micro-
environment in -1 aggregates and that the amide groups can
stabilize the molecular packing by complementary inter-
molecular hydrogen bonding. The complementary hydrogen
bonds were previously examined by using the Corey–Pauling–
Koltun (CPK) model and molecular mechanics (MM2) calcu-
lation in the CAChe molecular modeling program12,13,19,20
(abbreviated to CAChe-MM2 hereafter). Three amide groups
in -glutamic acid-derived lipid -1 can collaborate between
neighbouring lipids in a similar manner to the related lipids
with different headgroups,12,13,19,20 assuming such conform-
ations are possible only under head-to-head orientation. On
the other hand, ester-type lipid -10 is incapable of forming
such complementary hydrogen bonds. Such complementary
hydrogen bonding by -1 occurs even in organic media, e.g.,
benzene.12,13,20,21 Owing to the remarkable difference in the
visible absorption spectral behaviour between -1 and -10, the
-1 aggregates are regarded as a model system showing cooper-
ation of plural weak intermolecular interactions (hydrogen
bonding and hydrophobic interactions in this case) essential in
supramolecular chemistry.
viated to St-2C2 hereafter) not only as a guest molecule but also
as microenvironmental probes due to their solvatochromic
properties.
Results and discussion
Spectral behaviour of St-4C1 in the presence of anionic lipid
assemblies
Effect of cooperation of complementary hydrogen bonding and
hydrophobic interactions between lipids on planarity recognition
As reported previously,1,4 the wavelengths at the ultraviolet–
visible absorption maximum (λmax) of St-4C1 in water and in
methanol are 450 and 475 nm, respectively, regardless of the
temperatures. This indicates that the St-4C1 has no thermo-
chromic properties but exhibits solvatochromic behaviour.
It is known that increasing the molar ratio of lipid to dye pro-
motes the incorporation of the dye into the hydrophobic region
and the λmax value reaches a constant value close to those in
2570
J. Chem. Soc., Perkin Trans. 2, 1999, 2569–2577

View Article Online
Table 1 Dispersion state of St-4C1 in the presence of various lipids in water ([St-4C1] = 1.5 × 10Ϫ4 mol dmϪ3 = const., pH 10.0)
λmax/nm of St-4C1
Aggregate
morphologies
[Lipid]/
[St-4C1]
Dispersion state of
St-4C1 at 20 ЊC
Lipid
20 ЊC
65 ЊC
None (in water)
None (in methanol)
-1
-1
—
—
0
0
450
475
478
480
448
462
510e
458
450
454
450
452
475
458
443
469
481
(525)g
456
471
448
450
475b
450
445
448
445
450
455
450
454
450
452
444
444
443
448
462
(525)g
447
443
448
Ma
Ma
Mc
Mc
Md
Mc
Mc
Md
Md
Md
Mf
Md
Mc
Mc,d
Md
Mc
Mc
Jg
Helices and tubules
Helices and tubules
Tubules
Helices and tubules
Tubules
Tubules
Tubules
Vesicular aggregates
No structure
Tubules
Vesicular aggregates
Vesicular aggregates
Tape-like aggregates
Vesicular aggregates
Tubules and vesicles
Tubules and vesicles
Vesicular aggregates
Tubules and vesicles
Vesicular aggregates
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
-1
-22
-3
-4
-5
-6
-7
-8
-91,3,4
-93,4
-101,3
-11
-12
-13
-14
-15
-16
Mc,d
Mc
Md
a M; molecularly dispersed monomer of St-4C1. b Measured at 60 ЊC because the boiling point of methanol is 65 ЊC. c M; incorporated monomer as
evidenced by the bathochromic shift of λmax comparable to that in methanol with an increase in lipid concentration. In general, increasing the molar
ratio of lipid to dye leads to an aggregate-to-monomer transition of the dye. d M; monomer of St-4C1 bound to the carboxylate and is not
incorporated. e The anomalous bathochromic shift is not due to the J-aggregates but due to the most hydrophobic microenvironment in which St-4C1
is incorporated. f M; molecularly dispersed monomer of St-4C1 because -7 shows no aggregation behaviour. g J; J-aggregates of St-4C1 bound to the
surface of lipid aggregates: this is supported by the fact that J-aggregates are converted to monomeric dye species when the molar ratio of lipid to dye
is increased as shown in Table 2.
molecular planarity, based on their molecular structures, is
considered to be St-4C1 > St-2C1 > St-2C2.1,3,4 A semiempirical
quantum mechanical calculation using MOPAC in the CAChe
molecular modeling program also supported the order of
molecular planarity in terms of the rotation angle of two aro-
matic rings in a molecule: 0Њ for St-4C1, ca. 20Њ for St-2C1, ca.
40Њ for St-2C2. It is also noted that the order of preferential
incorporation of the dyes in the -9 aggregate system was
St-4C1 > St-2C1 > St-2C2 as reported previously,1,3,4 indicating
that the incorporation is mainly based on planarity recognition
rather than the order of hydrophobicity (St-4C1 ≅ St-2C1 <
St-2C2). However, the order of incorporation of St-2C1 and St-
2C2 in the -9 system is inverted in the -1 system. In general,
the higher the hydrophobicity of the dye, the more easily the
dyes are incorporated at lower molar ratios of lipid to the dye.
In this respect St-2C2 is more hydrophobic than St-2C1 because
of the higher hydrophobicity of the N-ethyl group than the
N-methyl group of St-2C1. Therefore these contradictory
results between -1 and -9 strongly suggest that hydrophobic
interactions between longer alkyl chains (2C18H37-) of lipid -9
are preferred for the planarity recognition of the dyes than the
Fig. 1 Relationships between λmax of St-4C1 and molar ratios in the
presence of -1, -12, and -15. [St-4C1] = 1.5 × 10Ϫ4 mol dmϪ3 = const.,
pH 10, 20 ЊC. The arrows indicate break points of the complete
incorporation with increase in the molar ratio of lipid to St-4C1.
collaboration of hydrophobic interactions between relatively
shorter alkyl chains (2C12H25-) and complementary hydrogen
bonding between amide groups of lipid -1.
organic solvents.22 In general, it is believed that the lower the
critical molar ratio at which the dyes are incorporated com-
pletely, the more easily the dyes are incorporated into the lipid
aggregates. Therefore, the critical molar ratio is a measure of
preferential incorporation (inclusion). Table 2 shows the molar
ratio variations of -1 to St-4C1, St-2C1, and St-2C2 at fixed dye
concentrations (1.5 × 10Ϫ4 mol dmϪ3), respectively. Fig. 1 also
shows the molar ratio variations of -1 to St-4C1. The critical
molar ratios of complete incorporation of these three dyes
indicate that the order of preferential inclusion is St-4C1 >
St-2C2 > St-2C1 in the -1 aggregate systems because the critical
molar ratios estimated from the data in Table 2 are as follows:
5 for St-4C1, 10 for St-2C2, ca. 20 for St-2C1. The order of
Effect of amino acid residues on inclusion and planarity
recognition
It is noted that the specific hydrophobic cavities were
also formed from -lysine-derived lipids (-11 and -12) and the
-ornithine-derived lipid (-14 and -15) as shown in Tables 1
and 2, as evidenced by the bathochromic shift of λmax at 20 ЊC.
This indicates that the formation of such specific hydrophobic
cavities is not restricted to the - or -glutamic acid-derived
lipids (1–3, 9). However, a subtle difference in the chemical
structures between lipids -12 and -13 affected the incorpor-
ation of St-4C1 considerably. For example, lipid -12 with
urethane (carbamate) groups incorporated St-4C1, whereas
J. Chem. Soc., Perkin Trans. 2, 1999, 2569–2577
2571

View Article Online
Table 2 Dispersion state of various dyes in the presence of various lipids in water ([Dye] = 1.5 × 10Ϫ4 mol dmϪ3 = const., pH 10.0)
λmax of dye/nm
Dispersion state of
dye at 20 ЊC
Lipid
Dye
[Lipid]/[Dye]
20 ЊC
65 ЊC
None
None
None
-1
St-4C1
St-2C1
St-2C2
St-4C1
0
450
475
435
460
434
460
478
478
450
475b
435
460b
434
460b
450
452
452
448
450
446
445
444
440
443
445
445
445
447
445
447
468
469
465
458
450
456
454
446
444
438
450
448
445
445
438
445
447
445
445
443
Ma
0 (in MeOH)
0
0 (in MeOH)
0
Ma
Ma
Ma
Ma
0 (in MeOH)
Ma
20
10
5
4
2
20
10
4
Mc
Mc
477
Mc
465
Mc,d
Mc & Je
Mc
466 (550)e
463
-1
St-2C1
457
445
443
465
463
452
462
Mc,d
Md
2
Md
-1
-2
St-2C2
St-4C1
20
10
5
20
10
4
Mc
Mc
Mc,d
Mc
473 (545)e
474 (547)e
471 (563)e
483
Mc & Je
Mc & Je
Mc & Je
Mc
2
-12
-12
-12
-15
St-4C1
St-2C1
St-2C2
St-4C1
20
10
6
5
3
483
483
475
453
468
468
460
455
443
463
461
460
455
445
475
475
Mc
Mc
Mc,d
Mc,d
Mc
20
17.5
15
10
5
20
17.5
15
10
5
20
15
10
7.5
5
Mc
Mc
Mc,d
Mc,d
Mc
Mc
Mc
Mc,d
Mc,d
Mc
Mc
475
474
465
Mc
Mc
Mc,d
a M; molecularly dispersed monomer. b Measured at 60 ЊC because the boiling point of methanol is 65 ЊC. c M; incorporated monomer as evidenced
by the bathochromic shift of λmax comparable to that in methanol with increase in lipid concentrations. In general, increasing the molar ratio of
lipid to dye leads to aggregate-to-monomer transition of the dye.1–4,22 d M; monomer bound to the carboxylate and is not incorporated due to the low
molar ratio of lipid to dye. e J; J-aggregates bound to the surface of lipid aggregates: this is supported by the fact that, in general, aggregated dye
species are converted to monomeric dye species when the molar ratio of lipid to dye is increased. The anomalous bathochromic shifts shown in
parentheses do not result from dye incorporation.
lipid -13 with amide groups could not incorporate St-4C1
but induced J-aggregates.23 According to the CPK model and
the CAChe-MM2 calculation, no appreciable difference was
observed between aggregation modes of -12 and -13, shown
schematically in Fig. 2. However, the fact that -13 induced
J-aggregation23 of St-4C1 strongly suggests that -13 is more
densely packed than -12 and enough to lead to the non-
incorporation of St-4C1.† This may be related to the fact that a
carbamate group is more polar and/or flexible than a carb-
oxamide group.
Next, incorporation preference of St-4C1, St-2C1, and St-2C2
by aggregates of -12 was investigated. Table 2 shows the molar
ratio variations of -12 to St-4C1, St-2C1, and St-2C2 at fixed
dye concentrations (1.5 × 10Ϫ4 mol dmϪ3), respectively. Fig. 1
also shows the molar ratio variations of -12 to St-4C1. The
critical molar ratios of complete incorporation of these
three dyes indicate that the order of preferential inclusion is
St-4C1 > St-2C1 > St-2C2 in the -12 aggregate systems, because
the critical molar ratios estimated from the data in Table 2
are as follows: 6 for St-4C1, 17.5 for St-2C2, more than 20 for
St-2C1. This order of incorporation preference is in good
agreement with that for -9 aggregate systems,3,4 indicating that
the aggregates of -12 can recognize the molecular planarity
of these dyes more preferentially than their hydrophobicity as
well as in the systems of -9.3,4 Therefore, it is concluded that
the recognition of molecular planarity is not restricted to the
-glutamic acid-derived lipids (1–3, 9) and may be a general
phenomena for appropriately designed α-amino acid-derived
lipids such as -12.
Effect of side-chain methylene number of amino acid residue on
inclusion and planarity recognition
It is noted that the bathochromic shifts of St-4C1 are more
significant for the corresponding lipids with longer side-chains
(11 > 14 > 16, and 12 > 15) as shown in Table 1. These results
suggest that a longer side-chain is preferred for the formation
of more hydrophobic cavities. It is noted that St-4C1 are more
† It is considered that the J-aggregates are not incorporated into the
lipid aggregates because they are believed to be composed of 4–7 dye
molecules per dye aggregate.23
2572
J. Chem. Soc., Perkin Trans. 2, 1999, 2569–2577

View Article Online
Fig. 2 Schematic representation of complementary hydrogen bonding between amide groups in the main chain and side-chain of lipids of
carbamate-type -12 (a) and amide-type -13 (b).
equation suggested by Sepúlveda et al.24,25 are listed in Table 3.
Calculations were based on the values listed in Table 2, using
the following equation: f/(1 Ϫ f ) = Ks{[Dt] Ϫ [St]f } Ϫ Ks[cac],
in which f denotes the fraction [aggregate-incorporated dye]/
[total dye], Dt is the total lipid concentration, St is the total dye
concentration, and cac is the critical aggregation concentration
of the lipid, respectively. For the calculation of f, variation of
λmax was used instead of the variation of absorbance, because
absorbance was less reproducible for the dyes used in this study.
It is noted that the Ks values for the St-4C1 were remarkably
enhanced as compared with those for St-2C1 and St-2C2.
Although Ks values for St-4C1 are considered to contain con-
siderable experimental errors because only two data points were
used for the calculation of the slope (Ks) due to the drastic
change of λmax for St-4C1 at the narrow molar ratio range. How-
ever, these Ks values may be used as comparisons for systematic
interpretation.
Table 3 Binding constant (KS) of the cationic dyes to the anionic lipid
aggregates of -1, -9, and -12
Lipids
Dyes
KS/dm3 molϪ1
-1
St-4C1
St-2C1
St-2C2
St-4C1
St-2C1
St-2C2
St-4C1
St-2C1
St-2C2
St-4C1
3 × 105
2 × 103
2 × 104
1 × 105
1 × 103
2 × 103
2 × 104
2 × 103
2 × 103
1 × 105
-9
-12
-15
preferentially incorporated into aggregates of -12 (m = 4)
than -15 (m = 3) which has one less side-chain methylene
number than -12, as indicated by the arrows at the critical
molar ratios of complete incorporation in Fig. 1. It is also
noted that the λmax of St-4C1 in the presence of lipids -12
and -15 with β-alanine head groups are more bathochromic
than those for the corresponding lipids -11 and -14 without
β-alanine residues as shown in Table 1. This indicates that the
spacer methylene numbers between α-amino acid residues and
polar headgroups also play an important role for formation of
the specific hydrophobic cavities as well as the side-chain methyl-
ene numbers. These results support the previous conclusion1,3
that the St-4C1 is incorporated in between the polar headgroups
(including hydrophobic spacer methylenes) and α-amino acid
residues (including side-chain) with the cooperation of electro-
static interaction and hydrophobic interaction under aggre-
gation of the lipids with head-to-head orientation. It is, there-
fore, concluded that the specific hydrophobic cavities can be
produced by single α-amino acid residue-containing lipids,
based on the cooperation of hydrophobic interactions (by long
chain alkyl groups, side-chain methylenes, and spacer moieties)
and complementary hydrogen bonding between amide or
carbamate groups. The balance of the cooperation can be regu-
lated by appropriate molecular designs. Hydrophobicity of
side-chain methylenes of amino acids has a large influence on
the formation of the specific hydrophobic cavities as well as on
the secondary structures of the corresponding poly(α-amino
acid)s.
Formation of uniform tubular aggregates from racemic glutamic
acid-derived lipids
Fig. 3 shows the TEM images for lipids -1 (a) and -1 (b).
Aggregate morphologies for -1 in water are observed as a mix-
ture of helical and tubular aggregates. Similar aggregates were
observed for -1 with opposite helical sense (right-handed for
-1, left-handed for -1). However, only tubular aggregates
were observed and no helical aggregates were observed for the
-1. This suggests that -1 and -1 are completely miscible,
resulting in no helical aggregates, only tubular aggregates.
Remarkable bathochromic shift in the λmax of St-4C1 incorpor-
ated in aggregates of enantiomeric -1/D-1 is also consistent
with the difference in the polarity of microenvironments in
which the St-4C1 is incorporated since no λmax shift was
observed in the aggregates of -1 even at a high molar ratio as
shown in Fig. 1. These results indicate that the specific hydro-
phobic cavities are formed not from racemic but from enantio-
meric -glutamic acid-derived lipids.
Conclusions
It has been clarified that the formation of supramolecular
receptors from α-amino acid-derived lipids is possible as long as
the component lipids are appropriately designed. The molecu-
lar structure requirements clarified/suggested in this study are
as follows. i) Enantiomeric α-amino acids should be used
because the -glutamic acid-derived lipids were incapable of
inclusion. However, if the racemic lipids phase-separate into
Binding constants of cationic dyes to the anionic lipid aggregates
The binding constant (Ks) determined from the slope of the
J. Chem. Soc., Perkin Trans. 2, 1999, 2569–2577
2573

View Article Online
Experimental
Materials
-Glutamic acid, -lysine monohydrochloride, -2,4-diamino-
butyric acid dihydrochloride, palmitoyl chloride, and Pd black
were purchased from Wako Pure Chemicals Co. Ltd. (Japan)
and used as obtained. -Ornithine monohydrochloride and
diethyl cyanophosphonate (DECP) were purchased from
Aldrich Co. Ltd. and used as obtained. n-Hexadecyl chloro-
formate and n-butylamine were purchased from Tokyo Kasei
Co. Ltd. (Japan) and used as obtained. St-4C1, St-2C1, St-2C2
were purchased from Aldrich and used as obtained. Other
reagents were also used as obtained unless otherwise noted.
Lipid preparation
All the lipids except for -2,2 -9,1,3,4 -9,3,4 and -101,3 were
newly synthesized and identified by Fourier transform infrared
spectroscopy (FTIR) measurement, 1H-NMR measurement,
and elemental analyses.
NЈ,NЉ-Didodecyl-Nꢀ-[4-carboxybutanoyl]-L-glutamide (L-1).
NЈ,NЈ-Didodecyl--glutamide (-2014) (2.0 g, 4.2 × 10Ϫ3 mol)
and triethylamine (0.51 g, 5.1 × 10Ϫ3 mol) were dissolved in
tetrahydrofuran (THF) and stirred with cooling. Glutaric
anhydride (0.54 g, 4.7 × 10Ϫ3 mol) was added to the solution.
After being stirred for 1 day at room temperature, the solution
was concentrated in vacuo. The residue was recrystallized from
methanol and dried in vacuo to give a white solid (-1): yield
2.0 g (80%); mp 127–133 ЊC; νmax(KBr)/cmϪ1 3298, 2926, 2856,
1
1702, 1638 and 1560; H-NMR (CDCl3) δ 0.78–0.97 (m, 6H,
CH3 × 2), 1.03–1.50 (m, 44H, (CH2)10 × 2, CH2 × 2), 1.77–2.45
(m, 6H, CH C(᎐O)NH × 2, CH C(᎐O)), 2.90–3.40 (m, 4H,
᎐
᎐
2
2
CH NHC(᎐O) × 2), 3.60–4.25 (m, 1H, CH) (Anal. Found: C,
᎐
2
68.73; H, 12.14; N, 7.50. Calc. for C34H65N3O5: C, 68.53; H,
11.00; N, 7.05%).
-1 and -1 were prepared as described above: -1, yield
75%; mp 127–133 ЊC; spectral data were identical to those of
-1: -1, yield 80%; mp 97–100 ЊC; νmax(KBr)/cmϪ1 3294, 2922,
Fig. 3 TEM images of aqueous dispersions of -1 (a) and -1
(b) negatively stained by ammonium molybdate. pH 10.0. Scale bars
represent 2000 Å, respectively.
1
2854, 1702, 1638 and 1562; H-NMR (CDCl3) δ 0.68–0.99 (m,
6H, CH3 × 2), 1.00–1.76 (m, 44H, (CH2)10 × 2, CH2 × 2), 1.80–
two enantiomeric components in water, the racemic lipids
would show the same results as enantiomeric ones. ii) A
relatively longer spacer is preferred for the design of the lipid.
iii) Acidic amino acid (glutamic acid) and basic amino acids
(lysine, ornithine) with relatively longer side-chains can be used.
iv) For the -glutamic acid-derived lipids, it is possible to use
ester bondings instead of amide bondings when much longer
alkyl chains are used.1–4 In other words, when alkyl chains are
relatively shorter, such as dodecyl groups, amide groups should
be used instead of ester groups. v) An enantiomeric -glutamic
acid-derived lipid with at least three amide groups is a necess-
ary condition for the formation of supramolecular helical
aggregates. The corresponding racemic lipids led to only the
tubular aggregates but no helical aggregates. vi) Although the
assembling manner of α-amino acid-derived lipids in water
can be inferred to some extent using CPK molecular model,
CAChe-MM2, and/or CAChe-MOPAC, it cannot necessarily
be precisely predicted whether the lipid aggregates are capable
of specific inclusion of dyes, as is schematically seen for -12
and -13 (Fig. 2 and Table 1).
In conclusion, the two-dimensional arrangement of enantio-
meric single α-amino acid residues in anionic lipid assemblies
can produce specific hydrophobic cavities. Although the present
system is not necessarily superior to those of conventional
macrocyclic receptors, it may be meaningful that one method-
ology has been demonstrated without using the closed struc-
ture. Supramolecular receptors, e.g., helical aggregates from
lipid -1/D-1, possessing such specific hydrophobic cavities are
considered to have potential applications to many fields.
2.42 (m, 6H, CH C(᎐O)NH × 2, CH C(᎐O)), 2.90–3.35 (m, 4H,
᎐ ᎐
2 2
CH NHC(᎐O) × 2), 3.92–4.20 (m, 1H, CH) (Anal. Found: C,
᎐
2
66.49; H, 11.60; N, 6.90. Calc. for C34H65N3O5ؒ1.0H2O: C,
66.49; H, 11.00; N, 6.84%).
The related compounds were prepared as described above,
using the corresponding substituted glutaric anhydrides instead
of glutaric anhydride: 3-methylglutaric anhydride for -3, 3,3-
dimethylglutaric anhydride for -4, 3-ethyl-3-methylglutaric
anhydride for -5, 3,3-tetramethyleneglutaric anhydride for
-6, and 3-methylglutaric anhydride for -7. -3, yield 78%; mp
138–141 ЊC; νmax(KBr)/cmϪ1 3298, 2922, 2854, 1702, 1636 and
1
1562; H-NMR (CDCl3) δ 0.60–1.00 (m, 9H, CH3 × 3), 1.00–
1.84 (m, 43H, (CH2)10 × 2, C(᎐O)CH CH(CH ), CH ), 1.85–
᎐
2
3
2
2.60 (m, 6H, CH C(᎐O)NH × 2, CH C(᎐O)), 2.80–3.40 (m, 4H,
᎐
᎐
2
2
CH NHC(᎐O) × 2), 4.42–4.56 (m, 1H, CH) (Anal. Found: C,
᎐
2
67.88; H, 11.92; N, 6.80. Calc. for C35H67N3O5ؒ0.5H2O: C,
67.88; H, 11.07; N, 6.79%). -4, yield 79.9%; mp 100–102 ЊC;
1
νmax(KBr)/cmϪ1 3298, 2920, 2854, 1638 and 1560; H-NMR
(CDCl3) δ 0.73–1.00 (m, 12H, CH3 × 4), 1.00–1.67 (m, 42H,
(CH2)10 × 2, CH ), 1.90–2.45 (m, 6H, CH C(᎐O)NH × 2,
᎐
2
2
CH C(᎐O)), 2.90–3.43 (m, 4H, CH NHC(᎐O) × 2), 4.30–4.60
᎐ ᎐
2 2
(m, 1H, CH) (Anal. Found: C, 68.55; H, 11.32; N, 6.61. Calc.
for C36H69N3O5ؒ0.4H2O: C, 68.55; H, 11.15; N, 6.66%). -5,
yield 76%; mp 103–105 ЊC; νmax(KBr)/cmϪ1 3452, 1649 and
1
1562; H-NMR (CDCl3) δ 0.78–1.00 (m, 12H, CH3 × 4), 1.12–
1.49 (m, 44H, (CH2)10 × 2, *CHCH2, CH2), 1.90–2.50 (m, 6H,
CH C(᎐O)NH × 2, CH C(᎐O)), 2.97–3.39 (m, 4H, CH N-
᎐
᎐
2
2
2
HC(᎐O) × 2), 4.43–4.56 (m, 1H, CH) (Anal. Found: C, 68.37;
᎐
H, 12.54; N, 7.14. Calc. for C37H71N3O5ؒ0.7H2O: C, 68.37; H,
2574
J. Chem. Soc., Perkin Trans. 2, 1999, 2569–2577

View Article Online
1
11.22; N, 6.47%). -6, yield 96%; mp 45–50 ЊC; νmax(KBr)/
cmϪ1 3294, 2926, 2856, 1711, 1638 and 1560; 1H-NMR
(CDCl3) δ 0.72–1.00 (m, 6H, CH3 × 2), 1.01–1.80 (m, 50H,
2922, 1688, 1657 and 1543; H-NMR (CDCl3) δ 0.67–1.01 (t,
6H, CH3), 1.01–1.81 (m, 40H, (CH2)10 × 2), 2.26–2.87 (m, 2H,
*CHCH ), 2.87–3.52 (m, 4H, CH NHC(᎐O) × 2), 4.26–4.72
᎐
2
2
(CH2)10 × 2, (CH ) , *CHCH ), 1.86–2.61 (m, 6H, CH C(᎐O)-
(m, 1H, CH), 4.99–5.24 (s, 2H, CH2C6H5), 7.20–7.68 (s, 5H,
᎐
2
4
2
2
NH × 2, CH C(᎐O)), 3.00–3.43 (m, 4H, CH NHC(᎐O) × 2),
C6H5).
᎐
᎐
2
2
3.92–4.20 (m, 1H, *CH) (Anal. Found: C, 69.47; H, 11.18; N,
6.34. Calc. for C38H71N3O5ؒ0.7H2O: C, 69.47; H, 11.01; N,
6.40%).
NЈ,NЉ-Didodecyl-L-aspartamide (L-22). -19 (3.0 g, 5.0
× 10Ϫ3 mol) was dissolved in 300 cm3 of ethanol with heating
and Pd black (1 g) was added to the solution. H2 gas was bub-
bled slowly into the solution for 5 hours. After removal of the
benzyl group, Pd black was removed by filtration. The solution
was concentrated in vacuo. The residue was recrystallized from
methanol to give a white powder (-22): yield 2.0 g (85%);
mp 102 ЊC; νmax(KBr)/cmϪ1 3304, 2922, 2854, 1632 and 1545;
1H-NMR (CDCl3) δ 0.76–1.00 (t, 6H, CH3), 1.13–1.80 (m, 40H,
(CH2)10 × 2), 2.45–2.71 (m, 2H, *CHCH2), 3.02–3.40 (m, 4H,
CH NHC(᎐O) × 2), 3.51–3.82 (m, 1H, *CH).
NЈ,NЉ-Dibutyl-Nꢀ-benzyloxycarbonyl-L-glutamide (L-18). N-
Benzyloxycarbonyl--glutamic acid (-Glu(Z))15 (4.0 g, 1.4 ×
10Ϫ2 mol), n-butylamine (2.3 g, 3.1 × 10Ϫ2 mol), and triethyl-
amine (4.4 g, 4.3 × 10Ϫ2 mol) were dissolved in THF (50 cm3).
The solution was cooled to 0 ЊC, and DECP (5.8 g, 3.3 × 10Ϫ2
mol) was added to the solution. After being stirred for 1 day at
room temperature, the solution was concentrated in vacuo, and
the residue was dissolved in 200 cm3 of chloroform. The solu-
tion was washed with 1 M NaOH, 0.2 M HCl, and water. The
solution was dried over Na2SO4 and concentrated in vacuo to
give a waxy solid (-18): yield 5.5 g (98%); mp 172–174 ЊC;
᎐
2
NЈ,NЉ-Didodecyl-Nꢀ-(4-carboxybutanoyl)-L-aspartamide
(L-8). -8 was prepared similarly as -1 using -22 instead of
-2014 to give a white powder (-8): yield 59%; mp 112–118 ЊC;
νmax(KBr)/cmϪ1 3292, 2922, 2854, 1700, 1647 and 1545;
1H-NMR (CDCl3) δ 0.75–1.00 (m, 6H, CH3 × 2), 1.04–2.12
1
νmax(KBr)/cmϪ1 3296, 3094, 1690, 1644 and 1539; H-NMR
(CDCl3) δ 0.68–1.08 (t, 6H, CH3 × 2), 1.08–1.83 (m, 8H,
CH3(CH2)2 × 2), 1.83–2.29 (m, 2H, *CHCH2), 2.29–2.51 (m,
2H, *CHCH CH C(᎐O)), 3.01–3.60 (m, 4H, CH NHC(᎐O) ×
(m, 40H, (CH2)10 × 2), 2.06–2.70 (m, 6H, CH C(᎐O)NH × 2,
᎐
2
᎐
᎐
2
2
2
2), 4.00–4.40 (m, 1H, *CH), 4.88–5.27 (s, 2H, CH2C6H5), 7.20–
7.64 (s, 5H, C6H5).
CH C(᎐O)), 2.84–3.35 (m, 4H, CH NHC(᎐O) × 2), 3.67–4.08
᎐ ᎐
2 2
(m, 1H, *CH) (Anal. Found: C, 65.43; H, 11.29; N, 7.81. Calc.
for C33H63N3O5ؒ1.3H2O: C, 65.43; H, 10.93; N, 6.94%).
NЈ,NЉ-Dibutyl-L-glutamide (L-21). -18 (3.5 g, 8.9 × 10Ϫ3
mol) was dissolved in ethanol (300 cm3) with heating and
Pd black (1 g) was added to the solution. H2 gas was bubbled
slowly into the solution for 6 hours. After removal of the benzyl
group, Pd black was removed by filtration. The solution was
concentrated and dried in vacuo to give a waxy solid (-21):
yield 1.9 g (85%); νmax(KBr)/cmϪ1 3318, 2962, 2934, 2874, 1647
and 1553; 1H-NMR (CDCl3) δ 0.68–1.10 (t, 6H, CH3 ×
2), 1.10–1.78 (m, 8H, CH3(CH2)2 × 2), 1.85–2.20 (m, 2H,
Nꢀ,Nꢁ-Bis(hexadecyloxycarbonyl)-L-lysine (L-11). -Lysine
monohydrochloride (5.0 g, 2.7 × 10Ϫ2 mol) was dissolved in
water. The solution was cooled to 0 ЊC, and a 20 cm3 portion of
water containing 4.3 g (1.1 × 10Ϫ1 mol) of NaOH was added to
the solution at 0 ЊC. Then a 100 cm3 of acetone containing 12.5
g (4.10 × 10Ϫ2 mol) of n-hexadecyl chloroformate was added
dropwise to the solution and vigorously stirred for 2 days to
give a slurry-like suspension. The pH of the mixture was
adjusted to 2 and stirred for 3 h. After filtration, the residue was
recrystallized from ethanol and dried in vacuo to give a white
solid (-11): yield 6.9 g (62%); mp 71–75 ЊC; νmax(KBr)/cmϪ1
1694 and 1557; 1H-NMR (CDCl3) δ 0.70–1.01 (m, 6H,
CH3 × 2), 1.01–1.95 (m, 62H, (CH2)14 × 2, *CH(CH2)3), 3.04–
3.41 (m, 2H, NHCH2), 3.90–4.50 (m, 5H, CH2O × 2, CH)
(Anal. Found: C, 69.51; H, 11.44; N, 4.09. Calc. for C40H78-
N2O6ؒ0.2H2O: C, 69.51; H, 11.50; N, 4.05%).
*CHCH CH C(᎐O)), 2.20–2.71 (m, 2H, CH CH C(᎐O)NH),
᎐
᎐
2
2
2
2
2.93–3.50 (m, 4H, CH NHC(᎐O) × 2), 4.00–4.40 (m, 1H,
᎐
2
*CH).
NЈ,NЉ-Dibutyl-Nꢀ-[3-(methyl)-4-carboxybutanoyl]-L-glutam-
ide (L-7). -21 (1.0 g, 3.89 × 10Ϫ3 mol) was dissolved in 50 cm3
of THF. 3-Methylglutaric anhydride (0.60 g, 4.66 × 10Ϫ3 mol)
and triethylamine (0.47 g, 4.66 × 10Ϫ3 mol) were added to the
solution with cooling. After being stirrerd for 1 day, the solu-
tion was concentrated in vacuo. The residue was recrystallized
from methanol and dried in vacuo to give a waxy solid (-7):
yield 29%; mp 139–146 ЊC; νmax(KBr)/cmϪ1 1702, 1647 and
Nꢀ,Nꢂ-Bis(hexadecyloxycarbonyl)-L-ornithine (L-14). -14
was prepared as described above using -ornithine monohydro-
chloride instead of -lysine hydrochloride: yield 11.8 g (93.4%);
mp 59.5–63.0 ЊC; νmax(KBr)/cmϪ1 3318, 2962, 2934, 1690 and
1
1553; H-NMR (CDCl3) δ 0.75–1.10 (t, 9H, CH3 × 3), 1.10–
1
1.63 (m, 11H, CH (CH ) × 2, *CHCH CH C(᎐O), CH CH-
1545; H-NMR (CDCl3) δ 0.70–1.05 (m, 6H, CH3 × 2), 1.05–
᎐
3
2
2
2
2
2
(CH )CH ), 1.70–2.62 (m, 6H, CH C(᎐O)NH × 2, CH C(᎐O)),
1.95 (m, 60H, (CH2)14 × 2, *CH(CH2)2), 3.05–3.40 (m, 2H,
NHCH2), 3.90–4.20 (m, 5H, CH2O × 2, *CH) (Anal. Found: C,
68.94; H, 11.25; N, 4.05. Calc. for C39H76N2O6ؒ0.6H2O: C,
68.94; H, 11.45; N, 4.12%).
᎐
᎐
3
2
2
2
2.70–3.43 (m, 4H, CH NHC(᎐O) × 2), 3.80–4.40 (m, 1H, *CH)
᎐
2
(Anal. Found: C, 57.39; H, 9.39; N, 10.83. Calc. for C19H35-
N3O5ؒ1.6H2O: C, 57.39; H, 9.66; N, 10.57%).
N-(Benzyloxycarbonyl)-L-aspartic acid (L-Asp(Z)). -Asp(Z)
was prepared similarly according to the method reported previ-
ously:14 Yield 5.6 g (65%); mp 115–116 ЊC (lit.15 116 ЊC).
ꢃ-Alanine benzyl ester toluene-p-sulfonate (23). Benzyl
alcohol (17.0 g, 0.157 mol), β-alanine (10.0 g, 0.135 mol),
and toluene-p-sulfonic acid monohydrate (25.6 g, 0.135 mol)
were suspended in 400 cm3 of toluene and refluxed for 5 h,
removing water azeotropically. The reaction mixture was con-
centrated in vacuo and an excess amount of diethyl ether was
added. The precipitates formed were collected and dried in
vacuo to give white plates (23): yield 38.0 g (96.3%); mp 138–
139 ЊC (lit.16 138–139 ЊC); νmax(KBr)/ cmϪ1 3066, 1736, 1232
and 1156.
NЈ,NЉ-Didodecyl-Nꢀ-(benzyloxycarbonyl)-L-aspartamide (L-
19). -Asp(Z) (2.0 g, 7.5 × 10Ϫ3 mol), 1-aminododecane (3.0 g,
1.6 × 10Ϫ2 mol), and triethylamine (2.3 g, 2.3 × 10Ϫ2 mol) were
dissolved in THF (150 cm3). The solution was cooled to 0 ЊC,
and DECP (3.3 g, 1.9 × 10Ϫ2 mol) was added to the solution.
After being stirred for 1 day at room temperature, the solution
was concentrated in vacuo, and the residue was dissolved in
200 cm3 of chloroform. The solution was washed with 1 M
NaOH, 0.2 M HCl, and water. The solution was dried over
Na2SO4 and concentrated in vacuo to give a white powder
(-19): yield 3.8 g (85%); mp 150–162 ЊC; νmax(KBr)/cmϪ1 3300,
Nꢀ,Nꢁ-Bis(hexadecyloxycarbonyl)-N-[2-(benzyloxycarbonyl)-
ethyl]-L-lysinamide (L-24). -11 (3.0 g, 4.39 × 10Ϫ3 mol) was
dissolved in 100 cmϪ3 of chloroform. Then 30 cmϪ3 portion
of THF containing β-alanine benzyl ester hydrotoluene-p-
J. Chem. Soc., Perkin Trans. 2, 1999, 2569–2577
2575

View Article Online
sulfonate (23) (1.66 g, 4.89 × 10Ϫ3 mol) and triethylamine (1.33
g, 1.32 × 10Ϫ2 mol) was added to the solution and stirred with
cooling. Then DECP (1.07 g, 6.10 × 10Ϫ3 mol) was added to the
solution with cooling. After being stirred for 1 day at room
temperature, the solution was washed with 5% NaHCO3, 0.2 M
HCl, and water. The solution was dried over Na2SO4 and
concentrated in vacuo. The residue was recrystallized from
methanol and dried in vacuo to give a white powder (-24): yield
2.5 g (77%); mp 72–76 ЊC; νmax(KBr)/cmϪ1 3296, 2962, 2934,
1H-NMR (CDCl3) δ 0.70–1.00 (m, 6H, CH3 × 2), 1.00–1.40 (m,
58H, (CH2)13 × 2, *CH(CH ) ), 2.05–2.30 (m, 4H, CH C(᎐O)),
3.05–3.40 (m, 4H, NHCH2), 4.40–4.82 (m, 1H, *CH), 5.10–5.20
(s, 2H, CH2C6H5), 6.20–6.45 (m, 2H, NHCH2), 7.20–7.50 (m,
5H, C6H5).
᎐
2
3
2
Nꢀ,Nꢁ-Bis(hexadecanoyl)-L-lysine (L-29). -28 (1.68 g, 2.58 ×
10Ϫ3 mol) was dissolved in 400 cm3 of ethanol with heating and
Pd black (1 g) was added to the solution. H2 gas was bubbled
slowly into the solution for 5 h. After removal of benzyl group,
Pd black was removed by filtration. The solution was concen-
trated in vacuo. The residue was recrystallized from methanol
and dried in vacuo to give white powder (-29): yield 1.3 g
(88%); mp 120–125 ЊC; νmax(KBr)/cmϪ1 3318, 2962, 2934, 1717,
1644 and 1560; 1H-NMR (CDCl3) δ 0.70–1.00 (m, 6H,
CH3 × 2), 1.00–1.40 (m, 58H, (CH2)13 × 2, *CH(CH2)3), 2.05–
1
1738, 1698, 1651 and 1557; H-NMR (CDCl3) δ 0.70–1.00 (m,
6H, CH3 × 2), 1.00–1.90 (m, 62H, (CH2)14 × 2, *CH(CH2)3),
2.45–2.70 (m, 2H, CH C(᎐O)), 3.03–3.72 (m, 4H, NHCH × 2),
᎐
2
2
3.90–4.51 (m, 5H, CH2O × 2, *CH), 5.05–5.20 (m, 2H,
CH2C6H5), 7.20–7.50 (m, 5H, C6H5) (Anal. Found: C, 71.77;
H, 10.76; N, 5.57. Calc. for C50H89N3O7: C, 71.77; H, 10.63;
N, 5.02%).
2.30 (m, 4H, CH C(᎐O)), 3.01–3.42 (m, 4H, NHCH ), 4.42–
4.83 (m, 1H, *CH) (Anal. Found: C, 72.74; H, 11.87; N, 4.39.
Calc. for C38H74N2O4ؒ0.2H2O: C, 72.74; H, 11.98; N, 4.47%).
᎐
2
2
Nꢀ,Nꢂ-Bis(hexadecyloxycarbonyl)-N-[2-(benzyloxycarbonyl)-
ethyl]-L-ornithinamide (L-25). -25 was prepared as described
above using -14 instead of -11: yield 2.0 g (54%); mp 53.0–
57.9 ЊC; νmax(KBr)/cmϪ1 3308, 2962, 2934, 1738, 1692, 1651 and
1545; δ 0.70–1.05 (m, 6H, CH3 × 2), 1.05–1.95 (m, 60H,
Nꢀ,Nꢁ-Bis(hexadecanoyl)-N-[2-(benzyloxycarbonyl)ethyl]-L-
lysinamide (L-30). -29 (1.10 g, 1.83 × 10Ϫ3 mol) was dissolved
in 100 cm3 of chloroform. A 30 cm3 portion of THF containing
0.71 g (2.0 × 10Ϫ3 mol) of 23 and 0.56 g (5.5 × 10Ϫ3 mol) of
triethylamine. DECP (0.45 g, 2.6 × 10Ϫ3 mol) was added to the
mixture and stirred for 30 min with cooling and subsequently
for 1 day at room temperature. The solution was washed with
5% NaHCO3, 0.2 M HCl, and water. The solution was dried
over Na2SO4 and concentrated in vacuo. The residue was
recrystallized from methanol and dried in vacuo to give a white
powder (-30): yield 1.2 g (88%); mp 96–105 ЊC; νmax(KBr)/cmϪ1
3308, 2962, 2934, 1651 and 1551 (Anal. Found: C, 70.40; H,
10.71; N, 4.83. Calc. for C48H85N3O5ؒ2.2H2O: C, 70.40; H,
10.94; N, 5.11%).
(CH2)14 × 2, *CH(CH ) ), 2.45–2.71 (m, 2H, CH C(᎐O)O),
᎐
2
2
2
3.05–3.85 (m, 4H, NHCH2 × 2), 3.90–4.20 (m, 4H, CH2O × 2),
4.60–5.01 (m, 1H, *CH), 5.05–5.20 (m, 2H, CH2C6H5), 5.25–
5.50, 7.28–7.40 (m, 5H, C6H5) (Anal. Found: C, 69.97; H, 10.72;
N, 4.73. Calc. for C49H87N3O7ؒ0.6H2O: C, 69.97; H, 10.57; N,
5.00%).
Nꢀ,Nꢁ-Bis(hexadecyloxycarbonyl)-N-[2-carboxyethyl]-L-lysin-
amide (L-12). -24 (1.47 g, 1.73 × 10Ϫ3 mol) was dissolved in
ethanol (300 cm3) with heating and Pd black (1 g) was added to
the solution. H2 gas was bubbled slowly into the solution for 6
h. After removal of the benzyl group, Pd black was removed by
filtration. The solution was concentrated and dried in vacuo to
give a white powder (-12): yield 0.95 g (73%); mp 126–129 ЊC;
νmax(KBr)/cmϪ1 3326, 2962, 2934, 1698, 1655 and 1555;
1H-NMR (CDCl3) δ 0.70–1.00 (m, 6H, CH3 × 2), 1.00–1.40 (m,
Nꢀ,Nꢁ-Bis(hexadecanoyl)-N-(2-carboxyethyl)-L-lysinamide
(L-13). -30 (0.90 g, 1.19 × 10Ϫ3 mol) was dissolved in 400 cm3
of ethanol with heating and Pd black (1 g) was added to the
solution. H2 gas was bubbled slowly into the solution for 7 h.
After removal of the benzyl group, Pd black was removed by
filtration. The solution was concentrated in vacuo. The residue
was recrystallized from methanol to give white powder (-13):
yield 0.58 g (73%); mp 126–129 ЊC; νmax(KBr)/cmϪ1 3310, 2962,
2934, 1717, 1644 and 1560 (Anal. Found: C, 70.55; H, 11.48;
N, 5.91. Calc. for C41H79N3O5ؒ0.2H2O: C, 70.55; H, 11.50; N,
6.02%).
62H, (CH2)14 × 2, *CH(CH ) ), 2.45–2.70 (m, 2H, CH C(᎐O)),
᎐
2
3
2
3.00–3.70 (m, 4H, NHCH2 × 2), 3.90–4.15 (m, 5H, CH2O × 2,
*CH) (Anal. Found: C, 67.78; H, 10.95; N, 5.51. Calc. for
C43H83N3O7ؒ0.4H2O: C, 67.78; H, 11.10; N, 5.52%).
Nꢀ,Nꢂ-Bis(hexadecyloxycarbonyl)-N-[2-carboxyethyl]-L-
ornithinamide (L-15). -15 was prepared as described above
using -25 instead of -24: yield 1.1 g (84%); mp 45.0–52.0 ЊC;
νmax(KBr)/cmϪ1 3310, 2962, 2934, 1692, 1653 and 1547; δ 0.70–
1.05 (m, 6H, CH3 × 2), 1.05–1.95 (m, 60H, (CH2)14 × 2,
L-2,4-Diamonobutyric acid benzyl ester bis(toluene-p-sulfon-
ate) (L-27). -2,4-Diaminobutyric acid dihydrochloride (0.90 g,
4.71 × 10Ϫ3 mol) and toluene-p-sulfonic acid monohydrate
(2.15 g, 1.13 × 10Ϫ2 mol) were suspended in 60 cm3 of benzyl
alcohol and refluxed for 12 h. Excess amount of diethyl ether
was added to the solution. The precipitates formed were col-
lected and dried in vacuo to give a white powder (-27): yield
2.1 g (79%); mp 101–125 ЊC; νmax(KBr)/cmϪ1 3066, 1758, 1731,
1226 and 1178.
*CH(CH ) ), 2.45–2.71 (m, 2H, CH C(᎐O)OH), 2.95–3.75 (m,
᎐
2
2
2
4H, NHCH2 × 2), 3.80–4.50 (m, 5H, CH2O × 2, *CH) (Anal.
Found: C,67.39; H, 11.09; N, 5.33. Calc. for C41H79N3O7ؒ
0.3H2O: C, 67.39; H, 10.97; N, 5.75%).
L-Lysine benzyl ester bis(toluene-p-sulfonate) (L-26). -26 was
prepared according to the literature:16 white powder, yield 2.0 g
(13%); mp 148–151 ЊC; νmax(KBr)/cmϪ1 3436, 3034, 1752, 1526,
1218 and 1176.
Nꢀ,Nꢄ-Bis(hexadecyloxycarbonyl)-L-diaminobutyric
acid
Nꢀ,Nꢁ-Bis(hexadecanoyl)-L-lysine benzyl ester (L-28). -26
(1.90 g, 3.27 × 10Ϫ3 mol) was dissolved in 80 cm3 of chloroform.
The solution was cooled to 0 ЊC, and triethylamine (1.65 g,
1.64 × 10Ϫ2 mol) was added to the solution. A 20 cm3 portion
of chloroform containing 1.98 g (7.19 × 10Ϫ2 mol) of palmitoyl
chloride was added dropwise to the solution with cooling. After
being stirred for 1 day at room temperature, the solution was
washed with 5% NaHCO3, 0.2 M HCl, and water. The solution
was dried over Na2SO4 and concentrated in vacuo. The residue
was recrystallized from methanol and dried in vacuo to give
a white powder (-28): yield 1.7 g (70%); mp 92.0–94.5 ЊC;
νmax(KBr)/cmϪ1 3318, 2962, 2934, 1750, 1642 and 1555;
benzyl ester (L-31). -27 (1.00 g, 1.82 × 10Ϫ3 mol) was dissolved
in 80 cm3 of chloroform. The solution was cooled to 0 ЊC, and
triethylamine (0.92 g, 9.1 × 10Ϫ3 mol) was added to the solu-
tion. A 20 cm3 portion of chloroform containing 1.2 g
(4.0 × 10Ϫ3 mol) of n-hexadecyl chloroformate was added
dropwise to the solution with cooling. After being stirred for
1 day at room temperature, the solution was washed with 5%
NaHCO3, 0.2 M HCl, and water. The solution was dried over
Na2SO4 and concentrated in vacuo. The residue was recrystal-
lized from methanol and dried in vacuo to give a white powder
(-31): yield 0.77 g (56%); mp 38.0–42.6 ЊC; νmax(KBr)/cmϪ1
1
3310, 2962, 2934, 1738, 1690 and 1545; H-NMR (CDCl3)
2576
J. Chem. Soc., Perkin Trans. 2, 1999, 2569–2577

View Article Online
δ 0.70–1.00 (m, 6H, CH3 × 2), 1.00–1.80 (m, 58H, (CH2)14 × 2,
*CHCH2), 2.80–3.30 (m, 2H, NHCH2), 3.50–3.80 (m, 1H,
*CH), 3.85–4.20 (m, 4H, CH2O × 2), 5.15–5.20 (m, 2H,
CH2C6H5), 7.20–7.45 (s, 5H, C6H5).
using a heating rate of 2 ЊC minϪ1. Gel-to-liquid crystalline
phase transition temperatures are as follows: 43 ЊC for -1; 35,
42, and 45 ЊC for -1; 36, 44, and 46 ЊC for -1; 38 ЊC for -2;
40 ЊC for -3; no detection for -4; 47, 57, 54, 39, and 27 ЊC for
-5; 23 ЊC for -6; no detection for -7; 88 ЊC for -8; 46 ЊC for
-11; 68 ЊC for -12; 79 ЊC for -13; 52 ЊC for -14; 52 and 76 ЊC
for -15; 62 ЊC for -16. It was also observed using TEM that
most of the lipids (except for -7 with short alkyl chains)
formed highly-oriented lipid aggregates based on lipid bilayer
structures in water. Aggregate morphologies are summarized in
Table 1. It is noted that there is no particular relationship
between the different aggregate morphologies and the disper-
sion states of the dyes.
L-Nꢀ,Nꢄ-Bis(hexadecyloxycarbonyl)-2,4-diaminobutyric acid
(L-16). -31 (0.77 g, 1.0 × 10Ϫ3 mol) was dissolved in 250 cm3 of
ethanol with heating and Pd black (1 g) was added to the solu-
tion. H2 gas was bubbled slowly into the solution for 3 h. After
removal of the benzyl group, Pd black was removed by filtra-
tion. The solution was concentrated in vacuo. The residue was
recrystallized from methanol and dried in vacuo to give white
powder (-16): yield 0.60 g (87%); mp 57.7–58.8 ЊC; νmax(KBr)/
cmϪ1 3310, 2962, 2934, 1698 and 1562; 1H-NMR (CDCl3)
δ 0.70–1.05 (m, 6H, CH3 × 2), 1.00–1.80 (m, 58H, (CH2)14 × 2,
*CHCH2), 2.80–3.30 (m, 2H, NHCH2), 3.50–3.80 (m, 1H,
*CH), 3.85–4.20 (m, 4H, CH2O × 2).
Acknowledgements
We are grateful to Professor Hideto Shosenji of Kumamoto
University for elemental analyses, and Dr M. Nishida and
J. Sugise for taking transmission electron micrographs. We are
also grateful to T. Ishikawa, T. Mishiro, N. Saeki, S. Uchiyama,
N. Kanekura, K. Urashi and D. Goto for lipid preparations
and technical assistance.
Characterization of lipids
The chemical structures of all the compounds synthesized were
confirmed by Fourier transform infrared spectroscopy (FTIR)
1
measurement with a JASCO FT/IR-7000, by H NMR meas-
urement‡ with a JEOL JNM-EX-90, and by elemental analysis
with a Yanaco CHN Corder MT-3.
References
1 H. Hachisako, T. Yamazaki, H. Ihara, C. Hirayama and K.
Yamada, J. Chem. Soc., Perkin Trans. 2, 1994, 1671.
2 H. Hachisako, Y. Motozato, R. Murakami and K. Yamada, Chem.
Lett., 1992, 219.
3 H. Hachisako, T. Yamazaki, H. Ihara, C. Hirayama and K.
Yamada, J. Chem. Soc., Perkin Trans. 2, 1994, 1681.
4 H. Hachisako, T. Yamazaki, R. Murakami and K. Yamada,
Liq. Cryst., 1993, 15, 723.
5 M. J. Grourke and J. H. Gibbs, Biopolymers, 1971, 10, 795.
6 G. Blauer and Z. B. Alfassi, Biochim. Biophys. Acta, 1967, 133, 206.
7 M. J. Grourke and J. H. Gibbs, Biopolymers, 1967, 6, 586.
8 S. R. Chadhuri and J. T. Yang, Biochemistry, 1965, 4, 1249.
9 M. Hatano, M. Yoneyama, I. Ito, T. Nozawa and M. Nakai, J. Am.
Chem. Soc., 1969, 91, 2165.
Preparation of aqueous solutions of lipids
The lipids were suspended in water (pH 10) and quickly heated
in hot water prior to sonication. The suspension was then soni-
cated using Ultrasonic Generator with a 4280S type vibrator
produced by Kaijyo Denki Co. Ltd. Then the pH was adjusted
with hydrochloric acid and sodium hydroxide.
Preparation of an aqueous lipid–dye mixture
All solutions of the lipid and dye mixtures were prepared by
addition of the stock solution of the dyes to aqueous disper-
sions of the lipids and subsequent sonication. After adjustment
of the pH to 10.0 with sodium hydroxide and hydrochloric
acid, the solutions were used for visible absorption spectra
measurements.
10 I. Satake and J. T. Yang, Biochem. Biophys. Res. Commun., 1973, 54,
930.
11 H. Yamamoto and J. T. Yang, Biopolymers, 1974, 13, 1093.
12 H. Ihara, K. Shudo, C. Hirayama, H. Hachisako and K. Yamada,
Liq. Cryst., 1996, 20, 807.
13 H. Ihara, K. Shudo, M. Takafuji, C. Hirayama, H. Hachisako and
K. Yamada, Jpn. J. Polym. Sci. Tech., 1995, 52, 606.
14 H. Ihara, H. Hachisako, C. Hirayama and K. Yamada, Liq. Cryst.,
1987, 2, 215.
15 M. Bergmann and L. Zervas, Ber., 1932, 65, 1192.
16 N. Izumiya and S. Makisumi, Nippon Kagaku Zasshi, 1957, 78, 662.
17 H. Ihara, H. Hachisako, C. Hirayama and K. Yamada, J. Chem.
Soc., Chem. Commun., 1992, 1244.
Visible absorption spectra measurements
The samples in a 1 mm quartz cell were incubated for 15 min
at selected temperatures. The visible absorption spectra were
measured with a JASCO Ubest 35 spectrophotometer.
Characterization of lipid aggregates
18 H. Hachisako, H. Ihara, T. Kamiya, C. Hirayama and K. Yamada,
J. Chem. Soc., Chem. Commun., 1997, 19.
19 H. Hachisako, H. Ihara and K. Yamada, Recent Res. Dev. Pure
Applied Chem., 1998, 2, 59.
20 H. Ihara, M. Yoshitake, M. Takafuji, T. Yamada, T. Sagawa,
C. Hirayama and H. Hachisako, Liq. Cryst., 1999, 26, 1021.
21 M. Takafuji, H. Ihara, C. Hirayama, H. Hachisako and K. Yamada,
Liq. Cryst., 1995, 18, 97.
22 N. Nakashima, H. Fukushima and T. Kunitake, Chem. Lett., 1981,
1555.
23 A. H. Herz, Adv. Colloid Interface Sci., 1977, 8, 237.
24 C. Hirose and L. Sepúlveda, J. Phys. Chem., 1981, 85, 3689.
25 C. A. Bunton, F. Rivera and L. Sepúlveda, J. Org. Chem., 1978, 43,
1166.
Formation of highly ordered lipid aggregates in water was
confirmed by using transmission electron microscopy (TEM)
with a JEOL 2000FX transmission electron microscope. The
aqueous samples (6 × 10Ϫ4 mol dmϪ3, pH 10) were spotted onto
carbon-coated copper grids. The samples were air-dried
at room temperature, after which they were stained with 2 wt%
aqueous ammonium molybdenate. The phase transition tem-
perature was measured by differential scanning calorimetry
(DSC) with a SEIKO I&E DSC 120. The sample solution (20
mmol dmϪ3, pH 10) was sealed in an Ag capsule and scanned
‡ Most of the -glutamic acid-derived lipids with three amides per
molecule formed an organic gel in CDCl3.12,13,17,18 Therefore, 1H-NMR
signals were considerably broadened.
Paper 9/03956B
J. Chem. Soc., Perkin Trans. 2, 1999, 2569–2577
2577