The effect of unsaturation on the formation of self-assembled gels from fatty acid L-serine amides and their cytotoxicity towards caco-2 cancer cells

Article abstract of DOI:10.1071/CH09211

A series of saturated and unsaturated fatty acid l-serines 3 were synthesized and their ability to form self-assembled gels was investigated. The saturated (lauroyl 3a and steraoyl 3b) and monounsaturated (oleoyl 3c) fatty acid l-serines form gels in both water and organic solvent, whereas the diunsaturated linoleyl-l-serine 3d does not form gels in these solvents, indicating that unsaturation adversely affects the gelation process. Cytotoxicity studies on these compounds with Caco-2 cancer cells in vitro show that these gels are only moderately cytotoxic at concentrations up to 0.5 mM, making them a promising candidate for applications such as drug delivery. CSIRO 2009.

Full text of DOI:10.1071/CH09211

RESEARCH FRONT
Rapid Communication
CSIRO PUBLISHING
www.publish.csiro.au/journals/ajc
Aust. J. Chem. 2009, 62, 653–656
The Effect of Unsaturation on the Formation of
Self-Assembled Gels from Fatty Acid L-Serine Amides
and their Cytotoxicity Towards Caco-2 Cancer Cells
Li Yun Grace Lim,A Yingying Su,B Filip Braet,B and Pall ThordarsonA,C
ASchool of Chemistry, The University of New South Wales, Sydney, NSW 2052, Australia.
BAustralian Key Centre for Microscopy and Microanalysis, The University of Sydney,
NSW 2006, Australia.
CCorresponding author. Email: p.thordarson@unsw.edu.au
A series of saturated and unsaturated fatty acid l-serines 3 were synthesized and their ability to form self-assembled gels
was investigated. The saturated (lauroyl 3a and steraoyl 3b) and monounsaturated (oleoyl 3c) fatty acid l-serines form
gels in both water and organic solvent, whereas the diunsaturated linoleyl-l-serine 3d does not form gels in these solvents,
indicating that unsaturation adversely affects the gelation process. Cytotoxicity studies on these compounds with Caco-2
cancer cells in vitro show that these gels are only moderately cytotoxic at concentrations up to 0.5 mM, making them a
promising candidate for applications such as drug delivery.
Manuscript received: 10 April 2009.
Final version: 27 May 2009.
OH
OH
Introduction
O
O
O
O
O
NHS
L-Serine
O
Self-assembled or molecular gels are formed by small low-
molecular weight organic gelators (LMOG) that self-assemble
into a supramolecular polymer network that encapsulates the sol-
vent in a solid-like state.[1] Self-assembled gels can be classified
as smart materials owing to the reversible nature of the bonds
that hold them together, as shown in our recent work on a novel
series of anion-sensitive pyromellitamide-based self-assembled
gels.[2,3] As the related and more commonly known (covalent)
polymeric gels, self-assembled gels can be further subdivided
into organogelsor hydrogels depending onthesolventused. Both
polymeric and self-assembled gels are showing great promise for
various medical applications, and their reversible nature makes
them a particularly attractive target for tissue regeneration[4,5]
and targeted drug delivery.[6,7] Leroux and coworkers have
shown that fatty acid amide methyl ester-based organogels can
be used for the controlled delivery of rivastigmine, which is
used to treat Alzheimer’s disease.[8] The fatty acid–amino acid
is a well-known motif in LMOG, including the dodecyl-l-serine
hydrogelator 3a[9] but little is known about the effect of unsat-
uration on the gelation properties of these compounds. To this
end, we synthesized a series of C18-fatty acid amides and com-
pared their gelation properties with the known hydrogelator 3a.
Additionally, the toxicity of these compounds was tested, but
biological compatibility issues are often overlooked in papers
describing developments in this field.
R
R
N
H
NaHCO3
THF/H2O
R
OH EtOAc
O
1a–d
2a–d
3a–d
a
b
c
d
R ϭ
R ϭ
R ϭ
R ϭ
C12:0
C18:0
C18:1
C18:2
Scheme 1. Synthesis of fatty acid l-serine amides.
with l-serine under Schotten–Baumann conditions[11,12] to yield
the fatty acid l-serines 3a–d (Scheme 1).
The gelation of the synthesized fatty acid l-serine amides 3a–
d was studied in pure water (pH 5–6), chloroform, and hexane
(Table 1).
The dodecanoyl-l-serine 3a (C12:0) has been reported to gel
water (pH 4.9–6.4) and toluene,[9] a finding that was confirmed
in the present work. The saturated 3b (C18:0) and the monoun-
saturated 3c (C18:1) fatty acid-l-serines synthesized here were
also able to form opaque white gels in water. However, the diun-
saturated linoleyl-l-serine 3d (C18:2) remained insoluble even
after heating to the boiling point of the solvent. It was also noted
The fatty acid amides described here are derived from lau-
royl (C12:0) 1a, stearoyl (C18:0) 1b, oleoyl (C18:1) 1c, and
linoleyl (C18:2) 1d fatty acids, the latter two mono- and di-
unsaturated, respectively. UsingthemethoddescribedbyLapidot
and coworkers,[10] these were converted to the corresponding N-
hydroxysuccinimide (NHS) esters 2a–d, followed by a reaction
© CSIRO 2009
10.1071/CH09211
0004-9425/09/070653

RESEARCH FRONT
654
L. Y. G. Lim et al.
100
80
60
40
20
0
Table 1. Gelation properties of fatty acid l-serine amides 3 in various
solvents
OG, opaque gel; TG, transparent gel; PG, partial gel (opaque); IS, insoluble
Gelator
WaterA
CHCl3
Hexane
3a (C12:0
3b (C18:0
)
)
OG
OG
OG
IS
TG
IS
TG
IS
PG
PG
TG
IS
3a
3b
3c
3d
3c (C18:1
)
3d (C18:2
)
AUnbuffered deionized water, final pH = 5–6.
that the gels formed from oleoyl-l-serine 3c in water started
to collapse within 30 min. These results, and the inability of
the diunsaturated 3d to form gels, indicate that restrictions on
the flexibility of the chains of these fatty acid l-serine amides
compared with the saturated steraoyl derivative 3b significantly
impact on their ability to form gels owing to their diminished
solubility in water.
0
0.1
0.2
0.3
0.4
0.5
Concentration [mM]
Fig. 1. The average percentage viability of Caco-2 cells in dodecanoyl-
l-serine 3a, stearoyl-l-serine 3b, oleoyl-l-serine 3c, and linoleyl-
l-serine 3d.
A similar but slightly more complex trend is observed for the
ability of these fatty acid l-serines 3 to form gels in chloroform
and hexane. The diunsaturated linoleyl-l-serine 3d is insoluble
in these solvents, whereas the monounsaturated oleoyl-l-serine
3c forms transparent gels. The two saturated gelators are a little
different; the shorter dodecanoyl-l-serine 3a (C12:0) forms a gel
in chloroform and partial gels hexane, whereas the stearoyl-l-
serine 3b (C18:0) is not soluble enough in chloroform to form gels
while it does form partial gels in hexane. These results suggest
that, as for water, the solubility of these gelators is the key factor
in determining their ability to form gels in organic solvents.
In order to access the suitability of these fatty acid l-serine
amides 3 for applications in biological systems, including drug
delivery, cytotoxicity studies were undertaken on Caco-2 cells.
The viability of these human colorectal adenocarcinoma cells
was assessed using a colourimetric assay based on the capacity
of cells to bind to the trophenylmethane dye crystal violet (CV),
which is used to directly quantify the cell numbers in vitro in
response to various experimental conditions. The results from
this assay are therefore proportional to the absorbance of the
dye taken up by the cells.[13] As it was essential that the hydro-
gelators were in a solution in the assay, they were tested for
dissolution in various solvents. Dodecanoyl-l-serine 3a, oleoyl-
l-serine 3c, and linoleoyl-l-serine 3d could be dissolved up to
1 mM in 4% ethanol with 96% aqueous phosphate buffer saline
(PBS), whereas to dissolve stearoyl-l-serine 3b, it was found
necessary to use 50% ethanol in 50% PBS, making the results
obtained for this gelator difficult to compare with the others.
The results obtained from these assays show that cells in both
dodecanoyl-l-serine 3a and oleoyl-l-serine 3c had similar via-
bilities of close to 95% after 24 h at 0.1 mM concentration, which
gradually declined to 86–92% for 0.2–0.5 mM concentrations of
3a and 3c, with the results for the less reliable stearoyl-l-serine
3b (in 50% ethanol) following a similar trend (Fig. 1).
this compound (50% ethanol), which may cause the Caco-2 cells
to be fixed to the 96-well plate and therefore give false-positive
results in the CV assays, these results should be interpreted with
caution.[14]
In conclusion, we have shown that introducing unsaturation
in the family of fatty acid l-serine gelators 3 does seem to
adversely affect their ability to aggregate into gel structures in
both water and organic solvents. The solubility of these com-
pounds seems to be the dominating factor in determining their
ability to form stable gels. We have also shown that this family
of gelators shows only mild toxic effects (<15–20% reduction
in viability) towards Caco-2 even at a fairly high concentration
(0.5 mM). These results indicate that these compounds could be
useful as carriers in localized drug delivery.
Experimental Details
Materials
The syntheses of the N-succinimidyl dodecanoate 2a,[9,10,15]
dodecanoyl-l-serine 3a,[10,16] N-succinimidyl octadecanoate
2b,[10,17] and octadecanoyl-l-serine 3b[10,17] have been reported
previously. All synthetic reactions were carried out in an inert
environment containing nitrogen, unless otherwise specified.
All reagents used in all reactions were commercially available
reagent-grade chemicals. Dichloromethane, hexane, methanol
were purified by distillation before use. Dry solvents including
tetrahydrofuran and ethyl acetate were distilled over appropriate
drying agents or obtained from a Pure Solv dry solvent sys-
tem (Innovative Technology Inc., model PS-MD-7). Deuterated
solvents for NMR were purchased from Merck and Aldrich.
Equipment
Melting points were measured with a melting point apparatus
and are uncorrected. Infrared spectra were recorded on a Nico-
let Avatar 360 Fourier-transform (FT)-IR or Nicolet Avatar 370
Of the four compounds tested, the linoleyl-l-serine 3d seems
to be slightly more toxic than the others, with observed cell
viabilities of 93% after 24 h at 0.1 mM concentration, declining
to less than 80% at 0.5 mM. It is noteworthy, though, that when
this compound was tested at 1 mM, cells did also show ∼80%
viability (not shown), indicating that although it does adversely
affect Caco-2 cells, they remain viable even in the presence of
a relatively high concentration of this compound. The results in
Fig. 1 also indicate that stearoyl-l-serine 3b was the least toxic
of these gelators; however, given the unusual conditions used for
1
FTIR using EZ OMNIC ESP 5.1 software. H Nuclear mag-
netic resonance spectra were recorded on a Bruker DPX 300
spectrometer at a frequency of 300.17 MHz. 13C NMR spec-
tra were recorded on a Bruker DPX 300 spectrometer at a
frequency of 75.48 MHz at 300 K. Low-resolution electrospray
ionization (ESI) was carried out on a Finnigan LCQ-DECA elec-
trospray instrument at the University of Sydney or a Waters
Micromass ZQ electrospray instrument at the University of

RESEARCH FRONT
The Effect of Unsaturation on the Formation of Self-Assembled Gels
655
New South Wales. High-resolution ESI (HR-ESI) was carried
out on a Thermo LTG FT instrument at the Bioanalytical Mass
Spectrometry Facilities (BMSF), Analytical Centre, University
of New South Wales.
and the filtrate was evaporated. It was left overnight under
nitrogen to precipitate out and further cooled in an ice bath to
give N-succinimidyl cis,cis-9,12-octadecadienoate (NHS ester
of linoleic acid) 2d as colourless liquid with sediments (3.00 g,
79%). νmax(MeOH)/cm−1 3363w, 2856–3008s, 1785–1815m,
1740s, 1465m. δH (300 MHz, [D6]DMSO) 0.86 (t, J 7.14,
3H, –CH3), 1.15–1.64 (m, 14H, –(CH2)7–), 1.59–1.64 (m, 2H,
–CH2CH2C=O(ON)), 1.98–2.03 (m, 4H, –CH2C=C–CH2),
2.51–2.61 (t, J 7.1, 2H, –CH2C=O(ON)), 2.66–2.75 (t, J 7.1, 4H,
–CH2C=O(ON)–), 5.26–5.39 (m, 4H, –HC=CH–). m/z (ESI)
376.12; [M]+ requires 376.53.
Synthesis of N-Succinimidyl cis-9-Octadecenoate
(NHS Ester of Oleic Acid) 2c
Oleic acid 1c (2.82 g, 3.17 mL, 10.0 mmol) was added to a solu-
tion of NHS (1.17 g, 10.2 mmol) in dry ethyl acetate (40 mL).
A solution of dicyclohexylcarbodiimide (2.07 g, 10.0 mmol) in
dry ethyl acetate (10 mL) was added and the reaction mixture
was stirred overnight at room temperature and under nitrogen.
The side product dicyclohexylurea was removed by filtration
and the filtrate was evaporated. It was left overnight under
nitrogen to precipitate out and dried to give N-succinimidyl
cis-9-octadecenoate (NHS ester of oleic acid) 2c as pale yel-
low gel-like paste (3.43 g, 90%). νmax(MeOH)/cm−1 2924s,
2853s, 1781s, 1741s, 1464s. δH (300 MHz, [D6]DMSO) 0.85
(t, J 6.8, 3H, –CH3), 1.24–1.47 (m, 20H, –(CH2)10–), 1.61
(m, 2H, –CH2CH2C=O(ON)–), 1.97–1.98 (m, 4H, –CH2C=C–
CH2), 2.14–2.19 (t, 2H, J 7.2, –CH2C=O(NO)–), 2.61–2.67 (t,
J 7.1, 4H, –CH2C(N)=O), 5.27–5.37 (m, 2H, –HC=CH–). δC
(75 MHz, [D6]DMSO) 13.81, 21.96, 24.14, 24.38, 25.30, 26.44,
27.87, 28.46, 28.56, 28.70, 28.89, 28.97, 30.05, 31.14, 33.57,
129.48, 129.53, 168.83, 170.10 (2C), 174.35. m/z (ESI) 402.21;
[M + Na]+ requires C22H37NO4Na 402.52.
Synthesis of cis,cis-9,12-Octadecadienoyl-L-serine
(Linoleyl-L-serine) 3d
Based on the method reported by Lapidot,[10] a solution of NHS
ester of linoleic acid 2d (0.76 g, 2.02 mmol) in distilled tetra-
hydrofuran (10 mL) was added to a solution of l-serine (0.23 g,
2.17 mmol) and sodium bicarbonate (0.18 g, 12.17 mmol) in
water (10 mL). It was stirred under nitrogen for 24 h and acidified
to pH 2 with aqueous hydrochloric acid (5 M, 1–2 mL). Tetra-
hydrofuran was then removed under vacuum. Water (50 mL)
was then added to the remaining suspension which was then
filtered to give a sticky colourless crude product. This crude
was dissolved in a minimum of diethyl ether (20 mL) and
precipitated into cold hexane (150 mL). The solution was cen-
trifuged and dried to give cis,cis-9,12-octadecadienoyl-l-serine
(linoleyl-l-serine) 3d as pale yellow gel-like substance (0.33 g,
44%). νmax(MeOH)/cm−1 3327w, 3011w, 2856–2928s, 2364w,
1727m, 1648m, 1544m. δH (300 MHz, [D6]DMSO) 0.94 (t,
J 7.1, 3H, –CH3), 1.33 (m, 14H, –(CH2)7–), 1.55 (s, 2H,
–CH2CH2C=O), 2.06–2.11 (m, 2H, –CH2C=O(N)–), 2.18–
2.23 (t, J 6.7, 4H, –CH2C=C–CH2), 3.63–3.76 (m, 2H, 2CH2),
4.29–4.34 (m, 1H, 1CH), 5.34–5.47 (m, 4H, –HC=CH–), 7.94–
7.97 (d, J 7.5, 1H, –NH–). δC (75 MHz, [D6]DMSO) 13.88,
21.93 (2C), 25.19 (2C), 26.56, 26.61, 28.56, 28.60, 28.68, 29.01,
30.86, 35.01, 54.49, 61.43, 127.72 (2C), 129.72 (2C), 172.12,
172.24. m/z (ESI) 366.30; [M]− requires 366.53. m/z (HR-MS
ESI) 366.2667; [M − H]− C21H36NO4 requires 366.2644.
Synthesis of cis-9-Octadecenoyl-L-serine
(Oleoyl-L-serine) 3c
Based on the method reported by Lapidot,[10] a solution of N-
hydroxysuccinimide ester of oleic acid 2c (0.50 g, 1.33 mmol)
in distilled tetrahydrofuran (10 mL) was added to a solution of
l-serine (0.14 g, 1.31 mmol) and sodium bicarbonate (0.13 g,
1.52 mmol) in water (10 mL). It was stirred under nitrogen for
43 h and acidified to pH 2 with aqueous hydrochloric acid
(5 M, 1–2 mL). Tetrahydrofuran was then removed under vac-
uum. Water (50 mL) was then added to the remaining suspension
which was then filtered and the precipitate dried to give a
white solid. This crude product was dissolved in a minimum
of ether (10 mL) and precipitated into cold hexane (150 mL).
The solution was centrifuged and precipitate was dried to give
cis-9-octadecenoyl-l-serine (oleoyl-l-serine) 3c as white pow-
der (0.09 g, 19%). Mp 50–60◦C. νmax(KBr)/cm−1 3339s, 3000s,
2919s, 2852s, 1739s, 1614s. δH (300 MHz, [D6]DMSO) 0.85 (t,
J 6.8, 3H, –CH3), 1.24 (m, 20H, –(CH2)10–), 1.44–1.49 (t, J
6.0, 2H, –CH2CH2C=O), 1.97–1.99 (m, 4H, –CH2C=C–CH2),
2.10–2.15 (t, J 7.5, 2H, –CH2C=O(N)–), 3.53–3.67 (m, 2H,
2CH2), 4.18–4.24 (m, 1H, 1CH), 5.27–5.37 (m, 2H, –HC=CH–),
7.80–7.83 (d, J 7.9, 1H, –NH–). δC (75 MHz, [D6]DMSO) 13.93,
22.20, 26.55, 26.60, 28.56, 28.62, 28.65, 28.70, 28.80, 29.07,
29.11, 30.67(2C), 31.25, 35.01, 54.51, 61.42, 129.63(2), 172.13,
172.27. m/z (ESI) 368.43; [M]− requires 368.52. m/z (HR-MS
ESI) 368.2803; [M − H]− C21H38NO4 requires 368.2801.
Gelation
In a typical experiment, 1% gel/water (w/v) was made by mixing
gelator 3 (5 mg) and MilliQ water (0.5 mL) in a sealed vial. The
vial was heated until a clear solution was obtained before the
pH was adjusted to 5–6 using dilute hydrochloric acid or sodium
hydroxide.The solution was allowed to cool to room temperature
before the gelation was checked by the inversion test. If there
was no solvent flow, gelation was achieved.
Cell Culture
Caco-2 cells were cultured to 80% confluency in 75-cm2 cell
culture flasks with an aliquot of Advanced Dulbecco’s medium
from Dulbecco’s Modified Eagle Medium (40 mL; DMEM from
GIBCO BRL 12491), supplemented with 5% (2 mL) heat-
inactivated foetal calf serum (Gibco 10100–147), 0.5 mL of
antibiotic-antimycotic (Gibco 15240) and 0.5 mL of Glutamax
(Gibco 35050, 200 mM).
Synthesis of N-Succinimidyl cis,cis-9,12-Octadecadienoate
(NHS Ester of Linoleic Acid) 2d
Linoleic acid 1d (2.80 g, 3.10 mL, 10 mmol) was added to a solu-
tion of NHS (1.16 g, 10.1 mmol) in dry ethyl acetate (40 mL).
A solution of dicyclohexylcarbodiimide (2.09 g, 10.1 mmol) in
dry ethyl acetate (10 mL) was added and the reaction mixture
was stirred overnight at room temperature and under nitrogen.
The side product dicyclohexylurea was removed by filtration
The used medium was then removed and the cells were
rinsed in PBS (5 mL; Gibco 10010) before trypsin–EDTA (2 mL;
SAFC Biosciences 59430C) was added to the cell monolayer
at 37◦C for 5–10 min for trypsinization, which detaches the
cells from the culture flask. The enzymatic reaction was inacti-
vated by the addition of fresh serum-containing DMEM media

RESEARCH FRONT
656
L. Y. G. Lim et al.
(5 mL) and the cells were centrifuged in a Labofuge 300 Her-
aeus for 2 at 1800 rpm. The cells were then resuspended in fresh
DMEM media (2 mL, 37◦C) to produce a suspension of cells
suitable for testing. All cell culture reagents were obtained from
Invitrogen Australia, except trypsin, which was obtained from
Sigma–Aldrich.
[2] J. E. A. Webb, M. J. Crossley, P. Turner, P. Thordarson, J. Am. Chem.
Soc. 2007, 129, 7155. doi:10.1021/JA0713781
[3] K. W. K. Tong, S. Dehn, J. E. A. Webb, K. Nakamura, F. Braet,
P. Thordarson, Langmuir 2009, in press.
[4] G. A. Silva, C. Czeisler, K. L. Niece, E. Beniash, D. A. Harrington,
J. A. Kessler, S. I. Stupp, Science 2004, 303, 1352. doi:10.1126/
SCIENCE.1093783
[5] K. Rajangam, H. A. Behanna, M. J. Hui, X. Han, J. F. Hulvat,
J. W. Lomasney, S. I. Stupp, Nano Lett. 2006, 6, 2086. doi:10.1021/
NL0613555
[6] A. Friggeri, B. L. Feringa, J. van Esch, J. Control. Release 2004, 97,
241. doi:10.1016/J.JCONREL.2004.03.012
[7] A. Vintiloiu, J.-C. Leroux, J. Control. Release 2008, 125, 179.
doi:10.1016/J.JCONREL.2007.09.014
[8] A. Vintiloiu, M. Lafleur, G. Bastiat, J.-C. Leroux, Pharm. Res. 2008,
25, 845. doi:10.1007/S11095-007-9384-3
[9] C. Boettcher, B. Schade, J.-H. Fuhrhop, Langmuir 2001, 17, 873.
doi:10.1021/LA001054P
[10] Y. Lapidot, S. Rappopor, Y. Wolman, J. Lipid Res. 1967, 8, 142.
[11] C. Schotten, Chem. Ber. 1884, 17, 2544.
Crystal Violet Assay
Typically, 10 000 cells in DMEM (100 µL) were seeded per 96-
multiwell plate and allowed to adhere to the bottom of the well
for 24 h at 37◦C. The cell cultures were then exposed to the toxic
agent (100 µL) for a further of 24 h at 37◦C.The supernatant was
removed and crystal violet solution (100 µL; CVS) was added.
The cells were then incubated at 37◦C for 10 min CVS before
it was removed and washed thoroughly with distilled water. The
stained cells were then air-dried and 33.3% acetic acid solution
(100 µL) was added to lyse the cells. The multiwall plate was
gently shaken for 10 min before the absorbance was recorded on
a Perkin–Elmer 1420 Victor3 V Multilabel Counter at 531 nm.
The data were processed with Wallac 1420 Manager software.
[12] E. Baumann, Chem. Ber. 1886, 19, 3218.
[13] B. Bonnekoh, A. Wevers, F. Jugert, H. Merk, G. Mahrle, Arch.
Dermatol. Res. 1989, 281, 487. doi:10.1007/BF00510085
[14] K. F. Foley, S. De Frutos, K. E. Laskovski, W. Tegge, W. R. Dostmann,
Front. Biosci. 2005, 10, 1302. doi:10.2741/1620
Acknowledgements
The authors acknowledge the facilities as well as technical assistance
from staff at the Australian Key Centre for Microscopy and Microanaly-
sis (AKCMM) of the University of Sydney. We thank Dr Frankie Stevens
of the AKCMM for excellent technical assistance with cell cultures and the
cytotoxicity assays. We would also like to thank the Australian Research
Council (ARC) for a Discovery Project Grant (DP0985059) to P.T. and F.B.
and the University of New South Wales Centre for Nanomedicine initiative
for financial support to P.T. for the present project. We also thank the ARC
for an Australian Research Fellowship (DP0666325) to P.T.
[15] T. Azzam, H. Eliyahu, A. Makovitzki, M. Linial, A. J. Domb,
J. Control. Release 2004, 96, 309. doi:10.1016/J.JCONREL.2004.
01.022
[16] C. W. Phoon, B. Somanadhan, S. C. H. Heng, A. Ngo, S. B. Ng,
M. S. Butler, A. D. Buss, M. M. Sim, Tetrahedron 2004, 60, 11619.
doi:10.1016/J.TET.2004.09.046
[17] T. Menahem, Y. Mastai, N. J. Chem. 2008, 32, 925. doi:10.1039/
B800211H
References
[1] R. G. Weiss, P. Terech, in Molecular Gels: Materials with Self-
Assembled Fibrillar Networks 2006 (Springer: Dordrecht).