Formation of specific dipolar microenvironments complementary to dipolar betaine dye by nonionic peptide lipids in nonpolar medium

Article abstract of DOI:10.1039/b818218c

This paper describes the host-guest interaction between nonionic peptide lipids and solvatochromic dipolar betaine dyes in nonpolar aprotic organic solvent. We have serendipitously found that the colour of Reichardt's Dye (referred to as Et(30) hereafter, although the term Et(30) has been used as a polarity parameter) in chlorobenzene unusually blue-shifted in the presence of L-glutamic acid-derived peptide lipid 1 with a benzyloxycarbonylated Gly headgroup. Since it is widely accepted that Et(30) shows negative solvatochromism, i.e., the visible absorption band of this dye blue-shifts as the solvent polarity increases, the blue-shift indicates that Et(30) was in contact with the more polar microenvironment produced by the peptide lipid 1 rather than chlorobenzene under aggregate-free conditions. The binding site was assumed to be N-H^δ+ and CO^δ- attached to both sides of the Gly residue, respectively, i.e., the O^- and N ⁺ of Et(30) complementarily bound to N-H^δ+ and CO^δ- through hydrogen bonding and ion-dipole interaction, respectively. Since Et(30) is practically non-fluorescent, it was not feasible to use fluorescence spectrometry, which is a powerful method for the study of host-guest interactions, in order to specify the binding mode of Et(30). Therefore, a synthetic approach, although very laborious but reliable, has been used in conjunction with solvatochromic probing using visible absorption spectroscopy to specify the binding site on peptide lipid 1. The binding site has been found to be located on two dipoles, i.e., N-H^δ+ and CO^δ- attached to both sides of the Gly residue, respectively, because introducing steric hindrance into the Gly moiety using several L-α-amino acids with bulky α-substituents interfered with the binding of Et(30). Similar specific binding behaviour of Et(30) was observed by replacing the Gly residue of the lipid 1 with sarcosine (Sar). It was found that self-assembly of the peptide lipid was necessary for effective capture of Et(30). The molecular structural requirements of the peptide lipids that form such specific polar microenvironments complementary to dipolar betaine dyes have also been investigated. The Royal Society of Chemistry 2009.

Full text of DOI:10.1039/b818218c

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/obc | Organic & Biomolecular Chemistry
Formation of specific dipolar microenvironments complementary to dipolar
betaine dye by nonionic peptide lipids in nonpolar medium†
Hiroshi Hachisako,*a,b Naoya Ryu,b Hiromi Hashimotob and Ryoichi Murakamia,b
Received 15th October 2008, Accepted 9th February 2009
First published as an Advance Article on the web 18th March 2009
DOI: 10.1039/b818218c
This paper describes the host–guest interaction between nonionic peptide lipids and solvatochromic
dipolar betaine dyes in nonpolar aprotic organic solvent. We have serendipitously found that the colour
of Reichardt’s Dye (referred to as ET(30) hereafter, although the term ET(30) has been used as a
polarity parameter) in chlorobenzene unusually blue-shifted in the presence of L-glutamic acid-derived
peptide lipid 1 with a benzyloxycarbonylated Gly headgroup. Since it is widely accepted that ET(30)
shows negative solvatochromism, i.e., the visible absorption band of this dye blue-shifts as the solvent
polarity increases, the blue-shift indicates that ET(30) was in contact with the more polar
microenvironment produced by the peptide lipid 1 rather than chlorobenzene under aggregate-free
d-
conditions. The binding site was assumed to be N-Hd+ and C O attached to both sides of the Gly
=
residue, respectively, i.e., the O- and N+ of ET(30) complementarily bound to N-Hd+ and C Od- through
=
hydrogen bonding and ion-dipole interaction, respectively. Since ET(30) is practically non-fluorescent, it
was not feasible to use fluorescence spectrometry, which is a powerful method for the study of
host–guest interactions, in order to specify the binding mode of ET(30). Therefore, a synthetic
approach, although very laborious but reliable, has been used in conjunction with solvatochromic
probing using visible absorption spectroscopy to specify the binding site on peptide lipid 1. The binding
d-
site has been found to be located on two dipoles, i.e., N-Hd+ and C O attached to both sides of the
=
Gly residue, respectively, because introducing steric hindrance into the Gly moiety using several
L-a-amino acids with bulky a-substituents interfered with the binding of ET(30). Similar specific
binding behaviour of ET(30) was observed by replacing the Gly residue of the lipid 1 with sarcosine
(Sar). It was found that self-assembly of the peptide lipid was necessary for effective capture of ET(30).
The molecular structural requirements of the peptide lipids that form such specific polar
microenvironments complementary to dipolar betaine dyes have also been investigated.
appropriate self-assembling acyclic molecules in nonpolar organic
Introduction
systems has been highly restricted. Since double-chained lipophilic
peptide lipids possessing, as dipolar functional groups, at least
three amide groups per molecule are soluble in nonpolar aromatic
media such as benzene and self-assemble into fibrillar aggregates
responsible for gelation,2 appropriately designed nonionic peptide
lipids are regarded as promising candidates for host compounds
for specific polar guests soluble in nonpolar organic media. It is
widely accepted that polar solutes are hardly solvated in nonpolar
solvent, and therefore, dipolar functional groups such as amide
groups exist almost bare in such a solvent. This results in much
stronger electrostatic interactions between oppositely charged
dipolar moieties in nonpolar solvent than in water, because both
positive and negative solutes in water are surrounded by the
oppositely polarized moiety of water molecules, respectively, and
therefore, net charges are considerably reduced. In this respect,
appropriately positioned plural amide groups in a peptide lipid
molecule would cooperatively act not only as self-assembly driving
forces for the lipid but also as receptors towards e.g., lipophilic
dipolar betaine molecules through complementary intermolecular
interactions such as hydrogen bonding and dipole–dipole inter-
actions in noncompetitive nonpolar media. Fortunately, ET(30),
one of the representative solvatochromic dipolar betaine dyes, was
found to be soluble in chlorobenzene which did not interfere
Host–guest chemistry through a bottom-up approach both in
water and in organic solvents using self-assembling molecules
is indispensable to fundamental study and applications in the
field of nanoscience and nanotechnology. In the accompanying
paper,1 we have demonstrated that a highly hydrophobic local
microenvironment could be produced in water using newly de-
signed self-assembling peptide amphiphiles. On the other hand, it
is also of great significance to produce specific polar local microen-
vironments in nonaqueous and nonpolar media using nonionic
and lipophilic self-assembing molecules. Polar guest molecules
would prefer the local polar microenvironment rather than the
nonpolar organic media. This would enable the construction of
host–guest systems using polar molecules in the nonpolar solvent.
In general, however, it is difficult to dissolve highly polar molecules
in nonpolar organic media without using e.g., lipophilic macro-
cyclic compounds. Therefore, studying host–guest chemistry using
aDepartment of Nanoscience, Faculty of Engineering, Sojo University, 4-22-
1 Ikeda, Kumamoto 860-0082, Japan. E-mail: hatisako@nano.sojo-u.ac.jp
bDivision of Applied Chemistry, Graduate School of Engineering, Sojo
University, 4-22-1 Ikeda, Kumamoto 860-0082, Japan
† Electronic supplementary information (ESI) available: (1) Experimental
details; (2) Supplementary figures. See DOI: 10.1039/b818218c
2338 | Org. Biomol. Chem., 2009, 7, 2338–2346
This journal is
The Royal Society of Chemistry 2009
©

View Article Online
with the self-assembly of the lipophilic peptide lipids. Since
ET(30) was sparingly soluble in nonpolar organic solvents such
as benzene and toluene that have been found suitable for the
self-assembly of peptide lipids,2 chlorobenzene was employed
in this study. We herein report on novel supramolecular hosts
formed from self-assembled L-glutamic acid-derived nonionic
peptide lipids that contain Gly and Sar residues. These lipids
were found to capture ET(30) especially when they self-assembled
in chlorobenzene as evidenced using the solvatochromic nature
of ET(30) under its aggregate-free condition. Furthermore, the
methodology of the receptor design has been potentially extended
to an ethylenediamine derivative.
is raised (Fig. S2†). When 3.0 mM of 1 was added to the
chlorobenzene solution of ET(30) at 20 ◦C, the lmax remarkably
blue-shifted to 672 nm (Fig. 1). Since ET(30) is virtually aggregate-
free at 0.15 mM in chlorobenzene (Fig. S2),† this blue-shift
(hypsochromic shift) of the lmax indicates that ET(30) was in
contact with a more polar microenvironment than chlorobenzene.
Therefore, the visually observed colour change from green (lmax
;
752 nm) to blue (lmax; 672 nm) in chlorobenzene could be ascribed
to the solvatochromism of monomeric ET(30) species bound to
the polar moieties of lipid 1. Molecular modelling suggested that
the compound 1 (see Scheme 1) and ET(30) formed a host–guest
complex as shown in Figs. 2 (a) and S3 (a).†
Results and discussion
Visible absorption spectral behaviour of ET(30) in chlorobenzene
It is widely accepted that ET(30) exhibits negative solva-
tochromism,3 i.e., the visible absorption maximum wavelength
(lmax) blue-shifts as the solvent polarity is increased. In fact,
its lmax in chlorobenzene, 1-butanol, and methanol (0.15 mM,
◦
20 C) were 752 nm (Fig. 1), 575 nm, and 515 nm, respectively.
Also, ET(30) exhibits slight thermochromism, i.e., the molar
extinction coefficient gradually increases as the temperature
Fig. 2 Schematic representation of complementary intermolecular in-
teractions between ET(30) and the spacer moiety of the dipeptide lipids:
(a) dipeptide lipid 1, (b) dipeptide lipid 7 with an N-methylated carbamate
group, (c) dipeptide lipids 2–6 with bulky side chains (R) on the spacer
moiety.
Specifying binding site, effect of self-assembly, and dye specificity
It was assumed that the receptor moiety on lipid 1 was Na-H of
=
the Gln residue and C O of the amino-protecting group (Z-group
or carbamate) attached to the Gly residue as shown in Figs. 2 (a)
and S3 (a). Although, in general, fluorescence spectrophotometry
is a powerful tool for the study of host–guest chemistry, ET(30)
was found to be practically non-fluorescent. Therefore, it was not
feasible to use fluorescence measurement to specify the binding
site of ET(30). Thus, a synthetic approach was used, i.e., replacing
the original Gly residue of 1 at the receptor moiety with chiral
a-amino acid residues. If our assumption were right, various
chiral a-amino acid residues with bulky a-substituents would
Fig. 1 Visible absorption spectra of ET(30) in the presence of lipids 1–7 in
chlorobenzene. Temp., 20 ◦C, path length of quartz cell: 1.0 cm, [lipid] =
3.0 mM, [ET(30)] = 0.15 mM, [lipid]/[ET(30)] = 20. Receptor ability is
in the order of Sar > Gly > Leu > Ala > Ile > Phe > Val under these
conditions as evaluated from the extent of the blue-shift.
This journal is
The Royal Society of Chemistry 2009
Org. Biomol. Chem., 2009, 7, 2338–2346 | 2339
©

View Article Online
Scheme 1
interfere with the complementary binding of ET(30) to the Na-
to be right as schematically shown in Fig. 2 (c). It may safely
d-
H and C O connected with these a-amino acid residues. Since
be concluded that the N–Hd+ and C O are complementarily
=
=
this approach is very laborious but reliable, we have prepared
peptide lipids 2–6 (see Scheme 1) with L-Ala-Z, L-Val-Z, L-Leu-
Z, L-Ile-Z, and L-Phe-Z headgroups, respectively. Fig. 1 shows
the visible absorption spectra of ET(30) in the presence of lipids
2–6 in chlorobenzene. As expected, bulky substituents interfered
with the blue-shift of absorption lmax of ET(30) at 20 ◦C, i.e.,
the binding of ET(30).‡ Therefore, our assumption was proved
bound to the O- and N+ of ET(30) through hydrogen bonding and
ion–dipole interaction, respectively. This conclusion was further
confirmed by using peptide lipid 7 (see Scheme 1) containing a
Sar residue as shown in Fig. 2 (b). The N-methyl group of the
Sar residue did not interfere with the binding of ET(30) at all.
Similarly, lipid 10 with a benzyl succinamate-headgroup exhibited
the most blue-shifted lmax as shown in Fig. 3. This indicates that
the N-H of the Gly residue of 1 has nothing to do with the binding
of ET(30). This is consistent with the conclusion described above.
It is noted that lipid 12 with a GABA residue as the spacer moiety
did not act as a receptor as evidenced by the result in Fig. 3,
although, judging from its molecular structure, the GABA residue
can complementarily bind to ET(30) as shown in Scheme 1. This
‡ As shown in Table 1, lipids 2, 4, and 6 with relatively lower Tc exhibited
induced-fit at elevated temperatures below their Tc. This indicates that
even some lipids with bulky substituents responsible for steric hindrance
at the receptor moiety can undergo induced-fit to some extent when their
self-assemblies are loosened at temperatures near their Tc.
2340 | Org. Biomol. Chem., 2009, 7, 2338–2346
This journal is
The Royal Society of Chemistry 2009
©

View Article Online
multiple interactions cooperating in the supramolecular cavity
formed between self-assembled lipids 1, 7, 8 and 9. With respect
to dye specificity, similar host–guest interactions were observed
when structurally related betaine dyes ET(1) and ET(33) were used
instead of ET(30) as listed in Table 1. These results suggest that
the host functions of the peptide lipids are not peculiar to ET(30)
and are a general phenomena towards dipolar betaine dyes.
Binding constants
The binding constant (Ks) was tentatively determined as reported
previously4b based on the equation suggested by Sepu¨lveda et al.5
Calculations were based on the following equation: f /(1 - f ) =
Ks{[Dt] - [St]f } - Ks[cac], in which f denotes the fraction
[aggregate-incorporated dye]/[total dye],§ Dt is the total lipid
concentration, St is the total ET(30) concentration, and cac is
the critical aggregation concentration of the lipid, respectively.
For the calculation of f , variation of lmax was used instead of the
variation of absorbance, because absorbance was less reproducible
for the dyes used in this study. The Ks values of ET(30) for 1, 7,
Fig. 3 Temperature dependence of the lmax of ET(30) in the presence of
lipids L-1, L-10, L-11, L-12 and L-13 in chlorobenzene: , L-1;+, L-10;
, L-11; ¥, L-12; ꢀ, L-13; ᭹, dye alone; [lipid] = 3.0 mM, [ET(30)] =
0.15 mM, [lipid]/[ET(30)] = 20.
◦
8, and 9 at 20 C were 5 ¥ 104 M-1, 3 ¥ 104 M-1, 1 ¥ 104 M-1,
and 1 ¥ 104 M-1, respectively. Although these values may contain
considerable errors, the Ks values could be used as comparisons
for systematic interpretation. It is concluded that separation of
the Gly or Sar residue from the L-Gln residue by insertion of a
b-Ala residue as a spacer lowered the Ks values. This indicates that
indicates that ET(30) is not incorporated into the lipid assembly
from the opposite side of the phenolate anion due to considerable
steric hindrance by the side chain. Furthermore, lipids 1, 7, 8 and
9 captured ET(30) with an increasing molar ratio of lipid to dye
at the constant dye concentration (0.15mM). The representative
visible absorption spectra of 8 and 9 are shown in Fig. 4. The
temperature dependences of the visible absorption spectra at the
respective molar ratios in the heating process (data not shown)
are very similar to the spectral changes with decreasing molar
ratio at 20 ◦C in Fig. 4. These results indicate that aggregation
of the lipid promotes the binding of ET(30), probably due to
§ Fraction f was determined according to the following equation f = (l -
la)/(lm - la), where l, la and lm denote the lmax value of ET(30) in the
presence of varying concentrations of lipid in chlorobenzene, the lmax value
of ET(30) alone in chlorobenzene, and the lmax value of ET(30) completely
bound to the lipid assembly in chlorobenzene, respectively. The slope (Ks)
could be obtained by plotting f /(1 - f ) against {[Dt] - [St]f }.
Fig. 4 Visible absorption spectra of ET(30) in the presence of lipids 8 (a) and 9 (b) in chlorobenzene at 20 ◦C. [ET(30)] = 0.15 mM = const.,
[lipid]/[ET(30)] = 0, 0.5, 1.0, 2.0, 2.9, 5.0, 6.7, 13.3, 20.0. Path length, 1.0 cm.
This journal is
The Royal Society of Chemistry 2009
Org. Biomol. Chem., 2009, 7, 2338–2346 | 2341
©

View Article Online
Table 1 The lmax of ET(30) in the presence of various peptide lipids in
is maintained. These results indicate that ET(30) interacts with 11
to give a change in colour corresponding to that in more polar
solvents in the heating process below Tc. Molecular modelling
suggested that ET(30) complementarily bound to Na-H of the Gln
chlorobenzene; [ET(30)] = 0.15 mM, [lipid] = 3.0 mM, [lipid]/[ET(30)] =
20
lmax at
Minimum
lmax/nm
20 ◦C/nm
Dlmax/nmb
residue and C O of the Z-group attached to the b-Ala residue of
=
Peptide lipid
Dye
11 due to a drastic conformational change as schematically shown
in Fig. 5. Lipids 12 and 13 with an increased number (m = 3 for 12
and m = 5 for 13) of methylene spacers than in 11 (m = 2) exhibited
less effective binding of ET(30) as shown in Fig. 3. Furthermore,
the minimum lmax values in the heating processes are closely related
to the spacer lengths as listed in Table 1. In other words, the shorter
the spacer length is, the more remarkably the lmax blue-shifted as
evidenced by the minimum lmax values of lipids 1 (m = 1), 11
(m = 2), 12 (m = 3), and 13 (m = 5). From these results, it is
concluded that the induced-fit of the b-Ala residue of lipid 11 is
responsible for the blue-shift of lmax with a drastic conformational
change from anti to non-anti conformation. Similar behaviours
were observed for the structurally related lipids 14 and 15 (see
Scheme 1 and Fig. S4†), although detailed behaviours in terms
of the temperature dependence of lmax were different depending
on the side-chain length that affected Tc values. The gel-to-sol
phase transition of self-assembled 11 by heating above Tc led to
destruction of the self-assembly of 11 and the host–guest complex
of 11 and ET(30), accompanied by restoratoion of the original lmax
of ET(30) alone in chlorobenzene. A schematic illustration of the
behaviour is shown in Fig. S5 (b).† Similar disassembly behaviours
of host–guest complexes at elevated temperatures above their Tc
were observed for all the lipids that captured ET(30) in their self-
assembled states.
1
ET(30)
ET(30)
ET(1)
672
652a
572
569
720
757
708
753
746
652
649
546
552
655
549
553
666
748d
718e
678e
704
728
752
736e
718e
752
647
633
513
659 (44 ◦C)c
652 (20 ◦C)a
–
13
0
–
–
32‡
0
11‡
0
30‡
0
1
1
1
ET(33)
ET(30)
ET(30)
ET(30)
ET(30)
ET(30)
ET(30)
ET(30)
ET(1)
–
2
688 (44 ◦C)c
757 (20 ◦C)
697 (41 ◦C)c
753 (20 ◦C)
716 (59 ◦C)c
652 (20 ◦C)
649 (20 ◦C)
–
3
4
5
6
7
8
0
8
–
8
ET(33)
ET(30)
ET(1)
–
–
9
655 (20 ◦C)
–
0
9
–
–
9
ET(33)
ET(30)
ET(30)
ET(30)
ET(30)
ET(30)
ET(30)
ET(30)
ET(30)
ET(30)
ET(30)
ET(1)
–
10
635 (50 ◦C)c
671 (53 ◦C)f
674 (56 ◦C)f
668 (41 ◦C)f
691 (50 ◦C)c
721 (53 ◦C)c
701 (59 ◦C)f
696 (59 ◦C)f
675 (50 ◦C)f
–
31
77g
44g
10
13
7
L-11
L-11
DL-11
12
13
14
51g
40g
43g
–
15
16
None
None
None
None (in
MeOH)
–
–
–
–
–
–
ET(33)
ET(30)
a [lipid]/[ET(30)] = 6.7, [ET(30)] = 0.15 mM. b Dlmax is defined as
subtraction of the minimum lmax in the heating process from lmax at
20 ◦C. c This is ascribed to slight induced-fit. d The lipid–dye mixture was
prepared without heat treatment. It was found that the lmax value of lipid 11
depended on the heat treatment prior to UV–visible absorption spectral
measurement. e The blue-shifted lmax from 752 nm is ascribed to initial
induced-fit caused by conformational change of the b-Ala residue. f The
remarkably blue-shifted lmax is ascribed to induced-fit caused by drastic
conformational change of the b-Ala residue. g The large Dlmax value could
be regarded as a measure of the drastic conformational change of the b-Ala
residue
self-assembled L-Gln residue and its neighbours contribute to the
enhancement of Ks values cooperatively. This is consistent with
the result that binding is promoted as the lipid concentration is
increased, i.e., self-assembly of lipid is promoted.
Fig. 5 Schematic representation of complementary intermolecular in-
teractions between ET(30) and the spacer moiety of peptide lipid L-11.
Adjacent lipids and bilayer structure are omitted for clarity.
Induced-fit and gel-to-sol phase transition in the heating process
Next, the host–guest interaction between 11 and ET(30) was
investigated. ET(30) alone in chlorobenzene (0.15 mM) exhibited
a green colour and lmax of 752 nm at 20 ◦C. When 3.0 mM of 11
was dissolved into the solution of ET(30) by sonication, the lmax
Molecular structural requirements of peptide lipids capable of
capturing dipolar dyes
Molecular structural requirements of the peptide lipid for specific
incorporation of ET(30) are summarized in Table 2. It is noted that
N¢,N¢¢-didodecyl-Na -[(3-carboxy)propanoyl]-L-glutamide (2C12-
L-Gln-suc)4b,4c with a carboxylic acid headgroup, capable of
complementary binding with ET(30) judging by its molecular
structure, did not induce a blue-shift in ET(30), indicating that
ET(30) was blocked at the surface of the lipid assembly probably
due to hydrogen bonding between –COOH and ET(30). This
suggests that a dipolar headgroup acting as a hydrogen bonding
◦
slightly blue-shifted to 748 nm at 20 C as shown in Fig. 3 and
Table 1.¶ The lmax blue-shifted to a minimum value of 671 nm
with increasing temperature as long as self-assembly of the lipid
¶ The initial values of lmax of ET(30) at 20 ◦C were found to be remarkably
affected by the heat treatment accompanied by aging at 20 ◦C prior to
visible absorption spectral measurements. This is due to the difference in
the extent of the initial induced-fit arising from the difference in the heat
treatment procedure. Table 1 shows two kinds of lmax data for L-11.
2342 | Org. Biomol. Chem., 2009, 7, 2338–2346
This journal is
The Royal Society of Chemistry 2009
©

View Article Online
Table 2 Structural requirements of acidic a-amino acid-derived peptide lipids and conditions for effective host–guest interaction with ET(30)
donor is not suitable as a headgroup. Racemic lipid (DL-11)
corresponding to L-11 exhibited a more hypsochromic shift of
the visible absorption band of ET(30) than L-11 under the same
conditions. It was reported that the molecular packing of glutamic
acid-derived enantiomeric lipids is tighter than the corresponding
racemates as long as they do not completely phase-separate
into respective enantiomeric components.4b DSC measurements
revealed that the DSC thermograms of L-11 and DL-11 are
not identical (data are given in the Experimental section: Char-
acterization of lipid aggregates), indicating that DL-11 is not
phase-separated into respective pure enantiomeric components. In
general, enantiomeric amphiphiles in aqueous systems are densely
packed and favorable to the specific incorporation of organic
guests due to the major driving force of hydrophobic interactions
between the hydrophobic moiety of the amphiphiles and the
incorporated guest. It was found that ET(30) was more effectively
captured by DL-11 than the corresponding enantiomeric L-11
as evidenc◦ed by the remarkable hypsochromic shift of the initial
◦
lmax at 20 C and the minimum lmax at 44 C of ET(30) (Fig. 6).
Since the molecular packing of DL-11 is assumed to be looser
tha◦n that of L-11 because of the lower Tc and smaller DH (Tc,
◦
◦
66 C and 76 C (DH; 10.1 kcal mol-1) for L-11; Tc, 54 C and
◦
67 C (DH; 6.4 kcal mol-1) for DL-11), this opposite behaviour
compared to conventional aqueous systems may be ascribed to the
This journal is
The Royal Society of Chemistry 2009
Org. Biomol. Chem., 2009, 7, 2338–2346 | 2343
©

View Article Online
d-
the binding site was assumed to be polarized N-Hd+ and C O
=
attached to the Gly residue as a spacer of the lipid. The binding
site was specified by means of introducing steric hindrance arising
from bulky a-substituents into the spacer moiety using several
L-a-amino acids instead of Gly. As expected, bulky substituents
interfered with the binding of ET(30). Therefore, it may safely be
concluded that the O- and N+ of ET(30) complementarily bind to
d-
N-Hd+ and the C O attached to the Gly residue of the lipid
=
through hydrogen bonding and ion–dipole interaction, respec-
tively. This conclusion was further confirmed by using Sar residue
instead of Gly because the N-methyl group of Sar residue did not
interfere with the binding of ET(30). It is noted that aggregation of
the lipid promoted the binding of ET(30), indicating that ET(30)
was cooperatively incorporated into the supramolecular cavity of
the self-assembled lipids through multiple interactions operating
in the supramolecular cavity. Similar host–guest interactions were
observed when the structurally related dipolar betaine dyes, ET(1)
and ET(33), were used instead of ET(30). Induced-fit was observed
when a b-Ala residue was used instead of the Gly residue. This
is presumably caused by both optimum spacer length and odd-
even effect of spacer methylenes. It was found that the present
findings, in terms of the design of a receptor moiety using Gly
or Sar, were potentially applicable to diacylated ethylenediamine
derivatives.
Fig. 6 Temperature dependence of the visible absorption maximum
wavelength (lmax) of ET(30) in the presence of lipids L-11 and DL-11
in chlorobenzene: [lipid] = 3.0 mM, [ET(30)] = 0.15 mM, [lipid]/
[ET(30)] = 20.
absence of hydrophobic interactions in the present non-aqueous
non-polar system. The same is true for the molecular structure of
the guest dyes, i.e., ET(30) is incorporated into the lipid assembly
from the bulky anionic side as schematically shown in Figs. 2, 5,
S3, and S5,† and Table 2. This direction in terms of incorporation
of ET(30) is supported by the fact that lipid 12 with a GABA
residue as the spacer moiety did not capture ET(30) effectively
as demonstrated by the data in Fig. 3. This is probably due
to steric hindrance between ET(30) and the side chain of lipid
12 as inferred from its chemical structure in Scheme 1. These
results appear to be peculiar to non-aqueous systems. Elongation
of the double-chain alkyl groups from didodecyl to dihexadecyl
was found to effect less influence on the blue-shift of lmax of
ET(30) (Fig. S6).† The present findings could be applied to related
dipolar molecules other than ET(30), ET(1), and ET(33) as long
as they are soluble in the nonpolar organic media in which the
peptide lipid can self-assemble. The present findings were extended
to a nonionic receptor, N,N¢-diacetylethylenediamine (Ac-ED-
Ac). Although Ac-ED-Ac formed precipitates in chlorobenzene at
room temperature due to aggregation via intermolecular hydrogen
bonding, this compound induced a lmax of 669 nm (absorbance;
0.93) in ET(30) under similar conditions. This indicates that diacy-
lated ethylenediamine would also behave as an effective receptor
towards ET(30) and its analogues in chlorobenzene, provided that
it is attached to an appropriate self-assembling lipophilic lipid
capable of complementary intermolecular hydrogen bonding, such
as those derived from L-lysine,1 to serve not only as a lipophilic
moiety but also as a support for arrangement of the diacylated
ethylenediamine residues in a nonpolar solvent.
In conclusion, we have demonstrated that self-assembled non-
ionic peptide lipids could act as receptors for dipolar dyes in
=
nonpolar aprotic solvent. The dipolar N-H and C O attached
to both sides of the Gly or Sar could complementarily bind to
not only dipolar ET(30) but also its analogues (ET(1) and ET(33))
in chlorobenzene. Molecular structural requirements for effective
capture of ET(30) have been clarified. Some results different
from aqueous systems were found: (i) The self-assembled racemic
mixture of 11 was rather superior at the capture of ET(30). This
is probably because ET(30) was incorporated more easily into the
lipid assembly due to less steric hindrance arising from looser
molecular packing of the racemic lipid. (ii) The phenolate anion
side of ET(30) molecule was preferentially incorporated rather
than the pyridinium moiety in the middle of ET(30), although the
phenolate anion side of ET(30) is rather bulky. These results may
be caused by the absence of hydrophobic interactions between
adjacent lipids. An ethylenediamine derivative was also found
to be potentially applicable to the receptor moiety. Induced-fit
behaviour typical of aqueous natural systems was found even in
the artificial chlorobenzene system. In view of these results, it
=
is suggested that the positioning of the N-H and C O groups
by a combination of amide and amide, amide and carbamate,
amide and ester, or carbamate and ester, respectively, via an
appropriate spacer would be a promising methodology for the
design of a binding site for dipolar molecules in nonpolar
aprotic organic media, provided that the molecules containing
these dipolar functional groups are soluble in such media. The
present findings could be applied to, e.g., normal phase liquid
chromatography for the separation of dipolar molecules when
appropriate peptide microparticles composed of not only a-amino
acids but also various w-amino acids could be prepared as the
stationary phase. It may also be possible to detect glycine residue
in peptides and proteins using solvatochromic dipolar dyes if
appropriate microenvironments for preferential binding could be
produced.
Conclusions
It has been found that ET(30) binds to L-glutamic acid-derived
double-chain lipids with benzyloxycarbonylated Gly headgroups
and changes colour in chlorobenzene as if ET(30) were dissolved
in more a polar organic solvent than chlorobenzene. This colour
change was ascribed to the formation of a complementary host–
guest complex between the peptide lipid and the dipolar ET(30) in
nonpolar aprotic chlorobenzene. From the molecular modelling,
2344 | Org. Biomol. Chem., 2009, 7, 2338–2346
This journal is
The Royal Society of Chemistry 2009
©

View Article Online
Characterization of peptide lipid aggregates
Experimental
Formation of highly ordered peptide lipid aggregates in chloroben-
zene was confirmed by using transmission electron microscopy
(TEM) with a JEOL 2000FX transmission electron microscope
or Phillips TECNAI F20 S-TWIN. The viscous chlorobenzene
solutions (fragmentary gels) of 1.0 mmol dm-3 peptide lipids were
spotted onto carbon-coated copper grids. The samples were air-
dried at room temperature, after which they were post-stained with
2 wt% aqueous ammonium molybdate. Aggregate morphologies
of 1–16 in chlorobenzene were observed using TEM. It has been
found that all the peptide lipids (1–16) formed fibrillar aggregates
based on bilayer structures after dissolution in chlorobenzene at
Materials and methods
N-Benzyloxycarbonylated a-amino acids (Z-a-amino acids)
and b-amino acid used in this study were as follows:
N-benzyloxycarbonylglysine (Gly-Z), N-benzyloxycarbonyl-L-
alanine (L-Ala-Z), N-benzyloxycarbonyl-L-valine (L-Val-Z), N-
benzyloxycarbonyl-L-leucine (L-Leu-Z), N-benzyloxycarbonyl-
L-isoleucine (L-Ile-Z), N-benzyloxycarbonyl-L-phenylalanine
(L-Phe-Z), N-benzyloxycarbonyl-sarcosine (Sar-Z), and N-
benzyloxycarbonyl-b-alanine (b-Ala-Z). These Z-amino acids
were purchased from Kokusan Chemicals. Co., Japan and used
as received. N,N¢-Diacetylethylenediamine (Ac-ED-Ac) was pur-
chased from Tokyo Kasei Kogyo Co., Ltd, Japan and used
as received. All the peptide compounds containing the N-
benzyloxycarbonylated a-amino acids were prepared according
to ordinary peptide synthesis procedures as described in the
preceding paper.1 All the a-amino acid derivatives denote L-
isomers unless otherwise specified. Reichardt’s Dye (2,6-diphenyl-
4-(2,4,6-triphenyl-1-pyridinio)phenolate) (referred to as ET(30)
in this paper), 1-(4-hydroxyphenyl)-2,4,6-triphenylpyridinium hy-
droxide, inner salt hydrate (referred to as ET(1) hereafter), and
2,6-dichloro-4-(2,4,6-triphenyl-1-pyridinio)phenolate (referred to
as ET(33) hereafter) were commercially available and used as
received.
◦
85 C in water bath and subsequent aging at room temperature.
The representative aggregate morphologies of L-1 and L-11 are
shown in Fig. S1.† Differential scanning calorimetry (DSC)
measurements of 20 mM chlorobenzene solutions also supported
the existence of self-assembled peptide lipids. The phase transition
temperature was measured by DSC with a SEIKO I & E DSC
120. The sample solution (20 mmol dm-3) was◦sealed in an Ag
capsule and scanned using a heating rate of 2.0 C min-1. Gel-to-
sol phase transition temperature (Tc) and transition enthalpy (DH)
of bilayer-based aggregates estimated by DSC measurements were
◦
◦
as follows: Tc, 57 C (DH; 9.3 kcal mol-1) for 1; Tc, 66 C (DH;
4.8 kcal mol-1) for 2; Tc, 88 ◦C (DH; 18 kcal mol-1) for 3; Tc, 51 ◦C
(DH; 14 kcal mol-1) for 4; Tc, 81 ◦C (DH; 6.9 kcal mol-1) for 5; Tc,
63 ◦C (DH; 14 kcal mol-1) for 6; Tc, 52 ◦C (DH; 10 kcal mol-1) for
Characterization of peptide lipids
◦
◦
7; Tc, 33 C (DH; 4.2 kcal mol-1) for 8; Tc, 25 C (DH; 2.4 kcal
mo◦l-1) for 9; Tc, 91 ◦C (DH; 1.8 kcal mol-1) for 10; Tc, 66 ◦C and
The chemical structures of all the compounds synthesized were
confirmed by Fourier transform infrared spectroscopy (FTIR)
measurement with a JASCO FT/IR-7000, by 1H NMR mea-
surement with a JEOL JNM-EX-270, and by elemental analysis
with a Yanaco CHN Corder MT-3. It is noted that most of the
peptide lipids and their precursors with more than three amide
bonds per molecule formed organogels in CDCl3, and therefore, 1H
NMR signals were considerably broadened. Therefore, 1H NMR
assignment was restricted to main signals, and no appreciable
existence of impurities was confirmed to ensure the reliability of
elemental analysis.
◦
◦
76 C (DH; 10.1 kcal mol-1) for L-11; Tc, 54 C and 67 C (DH;
6.4 kcal mol-1) for DL-11; Tc, 51 ◦C and 71 ◦C (DH; 13 kcal mol-1)
for 12; Tc, 65 ◦C (DH; 5.7 kcal mol-1) for 13; Tc, 77 ◦C and 83 ◦C
(DH; 4.0 kcal mol-1) for 14; Tc, 42 ◦C and 80 ◦C, (DH; 11 kcal
mol-1) for 15, and Tc, 71 ◦C (DH; 6.7 kcal mol-1) for 16.
Molecular modelling
Molecular modelling was performed using CAChe WorkSystem
Ver. 4.1.1 for Power Macintosh (1999 Oxford Molecular Ltd.)
installed on Macintosh G3 OS 8.6. To begin with, molecular
structure was produced using Editor, then transferred to Me-
chanics (settings: Parameter File: MM2 Parameters; Force Field
Type: MM2) to optimize conformation, and finally the results
were transferred to Visualizer. The graphics finally obtained were
further transferred to software for drawing and then appropriately
processed. It is noted that the modelling results were not repro-
ducible, and therefore, the figures produced may be regarded as
schematic to some extent.
Preparation of chlorobenzene solutions of lipid–dye mixtures
Peptide lipid was dissolved in chlorobenzene solution of 0.15 mM
ET(30) by sonication and heating in 85 ◦C water unless otherwise
◦
noted.ꢀ After being allowed to stand at 20 C for several hours
in the dark, the solutions were subjected to visible absorption
spectral measurements.
Visible absorption spectra measurements
The samples in a 1.0 cm◦quartz cell were incubated in a sample
holder for 10 min at 20 C. The visible absorption spectra were
measured with a JASCO Ubest V-530 spectrophotometer.
Acknowledgements
This work was supported in part by the Ministry of Education,
Culture, Sports, Science and Technology, Japan (S0801085). We
are grateful to Dr. H. Ihara and Dr. M. Takafuji of Kumamoto
University for elemental analyses, and Dr. M. Nishida of Kyushu
University for taking transmission electron micrographs using
JEOL 2000FX and Dr. R. Tomoshige of the Research Centre
for Advances in Impact Engineering, Faculty of Engineering,
Sojo University for taking transmission electron micrographs
Preparation of peptide lipids
All the peptide lipids 1–16 including intermediates 17–20 were
prepared according to previously reported methods4 (see ESI†).
ꢀ In the case of L-11, two kinds of samples with and without heating,
respectively, were prepared and the results are shown in Table 1.
This journal is
The Royal Society of Chemistry 2009
Org. Biomol. Chem., 2009, 7, 2338–2346 | 2345
©

View Article Online
using Phillips TECNAI F20 S-TWIN. We are also grateful to
N. Saeki, Y. Dan, S. Yamaura, S. Hirai, T. Sakaguchi, H. Momoi,
N. Kanekura, and Y. Umenaga for peptide lipid preparations and
technical assistance.
2358; (c) C. Reichardt, S. Asharin-Fard and G. Scha¨fer, Chem. Ber.,
1993, 126, 143–147; (d) N. O. Mchedlov-Petrossyan, N. A. Vodolazkaya,
A. A. Kornienko, E. L. Karyakina and C. Reichardt, Langmuir, 2005,
21, 7090–7096; (e) L. P. Novaki and O. A. E. Seoud, Langmuir, 2000,
16, 35–41; (f) M. J. Muldoon, C. M. Gordon and I. R. Dunkin, J. Chem.
Soc., Perkin Trans. 2, 2001, 433–435; (g) M. F. Vitha, J. Chem. Ed., 2001,
78, 370–372; (h) D. W. Allen and X. Li, J. Chem. Soc., Perkin Trans. 2,
1997, 1099–1103.
4 (a) H. Ihara, H. Hachisako, C. Hirayama and K. Yamada, Liq.
Cryst., 1987, 2, 215–221; (b) H. Hachisako, Y. Murata and H. Ihara,
J. Chem. Soc., Perkin Trans. 2, 1999, 2569–2577; (c) H. Hachisako,
Y. Motozato, R. Murakami and K. Yamada, Chem. Lett., 1992, 219–
222.
References
1 Preceding paper: H. Hachisako, N. Ryu and R. Murakami, Org. Biomol.
Chem., 2008, DOI: 10.1039/b818206j.
2 (a) H. Ihara, H. Hachisako, C. Hirayama and K. Yamada, J. Chem.
Soc., Chem. Commun., 1992, 1244–1245; (b) H. Hachisako, H. Ihara, T.
Kamiya, C. Hirayama and K. Yamada, Chem. Commun., 1997, 19–20.
3 (a) C. Reichardt, Solvents and Solvent Effects in Organic Chemistry,
VCH, Weinheim, 2004; (b) C. Reichardt, Chem. Rev., 1994, 94, 2319–
5 (a) C. Hirose and L. Sepu¨lveda, J. Phys. Chem., 1981, 85, 3689–3694;
(b) C. A. Bunton, F. Rivera and L. Sepu¨lveda, J. Org. Chem., 1978, 43,
1166–1173.
2346 | Org. Biomol. Chem., 2009, 7, 2338–2346
This journal is
The Royal Society of Chemistry 2009
©