Autopoietic self-reproduction of chiral fatty acid vesicles

Article abstract of DOI:10.1021/ja961728b

The self-reproduction of vesicles formed by (S)- and (R)-2-methyldodecanoic acid (4) was investigated in order to relate the autocatalytic increase of the vesicle concentration with enantioselectivity. 4(R) and 4(S) were synthesized with an enantiomeric excess greater than 98%. 4 forms vesicles in aqueous solution in the pH region between 8.8 and 7.5. Chiral properties of the vesicles were studied by differential scanning calorimetry (DSC) and circular dichroism (CD). For self-reproduction studies, the hydrolysis of the water-insoluble 2-methyldodecanoic anhydride (8) was investigated in a biphasic system consisting of an aqueous solution and 8. The reaction rates of 8(RR) and 8(SS) catalyzed by 4(R) or 4(S) vesicles were the same within experimental errors, indicating that the chiral vesicles cannot induce significant enantioselectivity. However, a clear effect was observed at 10°C: racemic vesicles destabilized during hydrolysis, causing phase separation, whereas homochiral vesicles remained stable and continued to self-reproduce.

Full text of DOI:10.1021/ja961728b

292
J. Am. Chem. Soc. 1997, 119, 292-301
Autopoietic Self-Reproduction of Chiral Fatty Acid Vesicles
Kenichi Morigaki,† Sabrina Dallavalle,‡ Peter Walde,† Stefano Colonna,‡ and
Pier Luigi Luisi*,†
Contribution from the Institut fu¨r Polymere, ETH-Zentrum, UniVersita¨tstrasse 6,
CH-8092 Zu¨rich, Switzerland, and Istituto di Chimica Organica, Facolta` di Farmacia,
UniVersita` degli Studi di Milano, Via Veneziana 21, I-20133 Milano, Italy
ReceiVed May 22, 1996. ReVised Manuscript ReceiVed October 29, 1996X
Abstract: The self-reproduction of vesicles formed by (S)- and (R)-2-methyldodecanoic acid (4) was investigated in
order to relate the autocatalytic increase of the vesicle concentration with enantioselectivity. 4(R) and 4(S) were
synthesized with an enantiomeric excess greater than 98%. 4 forms vesicles in aqueous solution in the pH region
between 8.8 and 7.5. Chiral properties of the vesicles were studied by differential scanning calorimetry (DSC) and
circular dichroism (CD). For self-reproduction studies, the hydrolysis of the water-insoluble 2-methyldodecanoic
anhydride (8) was investigated in a biphasic system consisting of an aqueous solution and 8. The reaction rates of
8(RR) and 8(SS) catalyzed by 4(R) or 4(S) vesicles were the same within experimental errors, indicating that the
chiral vesicles cannot induce significant enantioselectivity. However, a clear effect was observed at 10 °C: racemic
vesicles destabilized during hydrolysis, causing phase separation, whereas homochiral vesicles remained stable and
continued to self-reproduce.
Introduction
leading to self-reproduction may also lead to an increase of the
population of one enantiomer over the other.
In the last couple of years vesicular systems have been
described which are able to self-reproduce, i.e. able to increase
their population number owing to an autocatalytic process which
takes place within their own boundary.1,2 The surfactant was a
fatty acid (e.g. caprylic or oleic acid).3 The reaction was the
vesicle-catalyzed hydrolysis of the corresponding water-
insoluble anhydride which takes place within the boundary of
the vesicles, yielding the same fatty acid surfactant which then
leads to the spontaneous increase of the vesicle number.1 Self-
reproduction is a key process in the mechanism of life, and in
fact this particular vesicular reaction has been viewed as a simple
model for the transition to autopoiesis and for mimicking cellular
life.4,5 The present work focusses on the self-reproduction of
chiral vesicles (i.e. vesicles made of chiral surfactants), with
the aim of establishing a connection between self-reproduction
and chirality. More specifically, we have addressed the question
whether and to what extent the exponential autocatalytic process
For this study a surfactant system based on the chemistry
developed earlier with linear fatty acids has been chosen,1
preparing fatty acids and fatty anhydrides bearing an asymmetric
carbon atom in the R-position to the carboxyl group.
In the following, we are reporting in the first part on the
chemical synthesis of 2-methyldodecanoic acid (4) and on the
preparation and physicochemical characterization of vesicles of
4. In the second part, the self-reproduction of these supramo-
lecular aggregates will be described. The hydrolysis of 2-me-
thyldodecanoic anhydride (8) was studied by comparing first
the behaviors of homochiral and racemic vesicles.6 Secondly,
we will investigate whether homochiral vesicles induce selection
of one enantiomer over the other in the autocatalytic self-
reproduction process.
Results and Discussion
Synthesis of 4 and 8. (a) Synthesis of the Enantiomers of
2-Methyldodecanoic Acid (4). The title compound was
prepared according to Evans et al. (Scheme 1).7 Reaction of
lithiated (R)-(+)-4-benzyl-2-oxazolidinone (1) with dodecanoyl
chloride in anhydrous THF afforded (4R)-3-(dodecanoyl)-4-
benzyl-2-oxazolidinone (2(R)) in 91% yield, [R]D -71.6° (c 1,
CH3COCH3). This compound was added to a -78 °C solution
of sodium hexamethyldisilylamide in tetrahydrofuran to give
the corresponding enolate, which was alkylated with methyl
iodide affording (4R)-3-((2′R)-2′-methyldodecanoyl)-4-benzyl-
2-oxazolidinone (3(RR)) in 61% yield, [R]D -88.7° (c 1, CH3-
COCH3). The deacylation of oxazolidinone 3(RR) was carried
out with lithium hydroperoxide at 0 °C according to the
literature,8 giving (R)-(-)-2-methyldodecanoic acid (4(R)) in
90% yield, [R]D -15.6° (c 1, CH3COCH3). The (R) absolute
configuration of compound 4 has been made by comparison of
* To whom to address correspondence.
† Institut fu¨r Polymere, ETH-Zu¨rich.
‡ Istituto di Chimica Organica, Universita` degli Studi di Milano.
X Abstract published in AdVance ACS Abstracts, December 15, 1996.
(1) (a) Walde, P.; Wick, R.; Fresta, M.; Mangone, A.; Luisi, P. L. J.
Am. Chem. Soc. 1994, 116, 11649-11654. (b) Wick, R.; Walde, P.; Luisi,
P. L. J. Am. Chem. Soc. 1995, 117, 1435-1436. (c) Bachmann, P. A.;
Luisi, P. L.; Lang, J. Nature 1992, 357, 57-59.
(2) We use in this paper the terms self-replication and self-reproduction
as described previously.1a The term self-replication is limited to linear
structures whose replication is based on the template chemistry of nucleic
acids; the term self-reproduction is a more general one, valid for all
structures including micelles and vesicles; and the term autopoietic self-
reproduction is for the particular case in which reactions leading to self-
reproduction take place within the boundary of the reproductive unit (e.g.,
within the boundary of a vesicle).
(3) In analogy to the previous work,1a we use for simplicity the term
fatty acid independent of the protonation degree. Therefore, a “fatty acid
vesicle” is composed of a mixture of fatty acid molecules and soap
molecules.
(4) Maturana, H. R.; Varela, F. J. Autopoiesis and Cognition: The
Realization of the LiVing; D. Reidel Publishing Co.: Boston, MA, 1980.
(5) (a) Walde, P.; Goto, A.; Monnard, P.-A.; Wessicken, M.; Luisi, P.
L. J. Am. Chem. Soc. 1994, 116, 7541-7547. (b) Oberholzer, T.; Wick,
R.; Luisi, P. L.; Biebricher, C. K. Biochem. Biophys. Res. Commun. 1995,
207, 250-257. (c) Oberholzer, T.; Albrizio, M.; Luisi, P. L. Chem. Biol.
1995, 2, 677-682.
(6) With homochiral Vesicles we mean vesicles formed by one pure
enantiomer; with racemic Vesicles we mean vesicles composed of the
racemic mixture of the two enantiomers.
(7) Gage, J. R.; Evans, D. A. Org. Synth. 1990, 68, 83-91.
(8) Evans, D. A.; Britton, T. C.; Eliman, J. A. Tetrahedron Lett. 1987,
28, 6141-6144.
S0002-7863(96)01728-3 CCC: $14.00 © 1997 American Chemical Society

Autopoietic Self-Reproduction of Chiral Fatty Acid Vesicles
J. Am. Chem. Soc., Vol. 119, No. 2, 1997 293
Scheme 1
Scheme 3
DMPU was added as a cosolvent and the reaction mixture was
treated with 1.1 molar equiv of methyl iodide, to give the title
compound in 90% yield13 (Scheme 3).
(c) DSC Analysis. The melting points of n-dodecanoic acid
7 and of all 2-methyldodecanoic acids 4(R), 4(S), and 4(rac)
as determined by DSC measurements are 42.7 (for 7), 19.9 (for
4(rac)), 20.7 (for 4(R)), and 21.3 °C (for 4(S)). Although the
melting points of the two enantiomers 4(R) and 4(S) were very
similar to the melting point of the racemate 4(rac), differences
were seen in the DSC traces. The DSC (heating) trace of 4(R)
or 4(S) showed a small exothermal peak at about 2 °C, indicating
a transition between two different types of crystals. This is
probably due to polymorphism and it indicates that 4(R) or 4-
(S) can form two types of crystals depending on the temperature.
In contrast, 4(rac) forms only one type of crystal in the
temperature range measured. We have also measured the
melting points of various compositions of the two opposite
enantiomers 4(R) and 4(S) (data not shown). Analysis of the
corresponding DSC traces indicated that the crystalline structure
of the racemic mixture is a racemic compound.14 In other
words, two enantiomers are coexisting in the same unit cell and
they do not separate into domains of enantiomeric crystals
spontaneously.
Scheme 2
(d) Synthesis of 2-Methyldodecanoic Anhydride (8). One
equivalent of 4 and anhydrous pyridine were mixed in anhydrous
Et2O and 0.5 equiv of thionyl chloride was added at -10 °C.
The title compound was obtained in 80% yield after purification.
8(RR), 8(SS), and 8(mix) were obtained starting from 4(R),
4(S), and 4(rac), respectively. For 8(RR) and 8(SS) DSC
heating measurements gave almost identical endothermic peaks
at around 12 °C.
Preparation and Characterization of Vesicles of 4. (a)
Spontaneous Formation of Vesicles. After having described
the chemical synthesis work, we will now report on the
aggregation properties of 4, in particular on the formation of
vesicles of 4.
In aqueous medium, dispersed fatty acids form different types
of aggregates, depending on the ionization degree of the
carboxyl group:15-18 while at alkaline pH micelles are formed,19
the fatty acid molecules assemble into bilayers (vesicles) at
conditions where about half of the molecules are protonated.
In analogy to the previous work on linear fatty acids,16,17 we
have studied the aggregation behavior of 4 at various degrees
of ionization by a controlled addition of HCl to a micellar
the sign of its optical rotation with that reported in the literature.9
The enantiomeric excess of (R)-(-)-2-methyldodecanoic acid
1
(4(R)) (99% e.e.) was determined via H-NMR spectroscopy
in the presence of Eu(hfc)3 as chiral shift reagent, and confirmed
through the conversion of compound 4(R) into the correspond-
ing diastereoisomeric (RR) and (RS) amides 5 via reaction of
its acyl chloride with enantiomerically pure R-phenylethylamine
(see Experimental Section).10 The overall process is highly
stereoselective and its stereochemical course is in line with the
formation of chelated (Z)-enolate such as 6 (Scheme 2), in the
creation of a diastereofacial bias in the acylation reaction.11 It
is worth mentioning that the stereoselectivity of the acylation
is very high in spite of the small steric requirement of methyl
iodide.
In agreement with Evans et al.11 one must employ alkylating
agents reacting at convenient rate at temperature <20 °C in order
to avoid the decomposition of sodium enolate 6. Satisfactory
chemical yields are obtained by using allyl bromide and benzyl
bromide as alkylating agents whereas with ethyl iodide no
alkylation occurs (unpublished result). A possible explanation
is that nucleophiles may attack hindered acyloxazolidinones
more rapidly at the ring carbonyl than at the exocyclic N-acyl
group, as recently found by Fleming and co-workers.12
The (S)-2-methyldodecanoic acid 4(S) (99% e.e.), [R]D
+15.6° (c 1, CH3COCH3), has been obtained by the same
procedure starting from (S)-(-)-4-benzyl-2-oxazolidinone.
(b) Synthesis of Racemic 2-Methyldodecanoic Acid (4(rac)).
Dodecanoic acid 7 reacted with 2 molar equiv of LDA in
anhydrous THF at 0 °C to give the corresponding dianion;
(13) Pfeffer, P. E.; Sibert, L. S.; Chirinko, J. M. J. Org. Chem. 1972,
37, 451-458.
(14) Jacques, J.; Collet, A.; Wilen, S. H. Enantiomers, Racemates and
Resolutions; Wiley-Interscience: New York, 1980.
(15) (a) Gebicki, J. M.; Hicks, M. Nature 1973, 243, 232-234. (b)
Gebicki, J. M.; Hicks, M. Chem. Phys. Lipids 1976, 16, 142-160. (c)
Hargreaves, W. R.; Deamer, D. W. Biochemistry 1978, 17, 3759-3768.
(16) (a) Cistola, D. P.; Atkinson, D.; Hamilton, J. A.; Small, D. M.
Biochemistry 1986, 25, 2804-2812. (b) Cistola, D. P.; Hamilton, J. A.;
Jackson, D.; Small, D. M. Biochemistry 1988, 27, 1881-1888. (c) Cistola,
D. P.; Small, D. M. J. Am. Chem. Soc. 1990, 112, 3214-3215.
(17) Rosano, H. L.; Chistodoulou, A. P.; Feinstein, M. E. J. Colloid
Interface Sci. 1969, 29, 335-344.
(18) (a) Haines, T. H. Proc. Natl. Acad. Sci. U.S.A. 1983, 80, 160-164.
(b) Li, W.; Haines, T. H. Biochemistry 1986, 25, 7477-7483.
(19) The cmc values for 7, 4(R), and 4(rac) are listed in Table 1. The
cmc values of 4 are lower than the cmc value of the linear n-dodecanoic
acid 7. This can be understood on the basis of the fact that 2-methyldode-
canoic acid has a larger hydrophobic part than n-dodecanoic acid. The
cmc value was the same for 4(R), 4(S), and 4(rac).
(9) Meyers, A. I.; Poindexter, G. S.; Brich, Z. J. Org. Chem. 1978, 43,
892-898.
(10) Sonnet, P. E.; Baillangeon, M. W. Lipids 1991, 26, 295-300.
(11) Evans, D. A.; Ennis, M. D.; Mathre, D. J. J. Am. Chem. Soc. 1982,
104, 1737-1739.
(12) Archibald, S. C.; Fleming, I. J. Chem. Soc., Perkin Trans. 1 1993,
751-754.

294 J. Am. Chem. Soc., Vol. 119, No. 2, 1997
Morigaki et al.
B
Figure 1. Equilibrium titration curve of 4(R). Various amounts of 1
M HCl were added to samples of 0.5 mL of 80 mM 4(R), 88 mM
NaOH, and the pH was measured at 25 °C 6 days after preparation.
The horizontal arrows indicate the region where micelles, vesicles
(lamellar bilayer), and/or oil emulsion are present.
Table 1. Chain-Melting Transition Temperature (Tc) of Hydrated
Bilayers (50% Ionized) of n-Dodecanoic Acid (7) and 2-Methyldo-
decanoic Acid (4) and cmc Values of the Corresponding Sodium
Salts (T ) 25 °C)
cmc (mM)
fatty acid
Tc (°C)a
conductometry
spectrometry
7b
4(rac)
4(R)
4(S)
30.0
12.2
7.6
25
21
21
21
22
16
19
7.4
n.d.c
a Tc were determined by extrapolating the slope of the onset of the
endothermal peak to the base line. b Literature cmc values: 24.4 mM;30b
23 mM;30c 22.5 mM.30d c n.d.: not determined.
Figure 2. (A) Freeze fracture electron micrograph of spontaneously
formed vesicles of 4(rac) in 0.05 M Tris/HCl buffer solution at pH
8.0. The length of the bar corresponds to 200 nm. (B) Size distribution
of the vesicles formed upon hydrolysis of 8(SS) in a two-phase system
consisting initially of an aqueous phase (10 mL of 0.05 M bicine buffer,
pH 8.2) and 0.15 mmol of 8(SS) at 25 °C. Freeze fracture electron
micrographs were analyzed after the reaction had finished.
solution of fully ionized 4. Figure 1 shows the room temper-
ature titration curve for 4(R) which will now be discussed on
the basis of the Gibbs phase rule.16b,20
Before the titration was started, 4(R) was dissolved at a
concentration of 0.08 M (>cmc, Table 1) in water by adding a
10 mol % excess of NaOH. This solution was optically clear,
containing micelles in equilibrium with monomers (Figure 1,
point A). Addition of HCl induced slight turbidity at point B.
Between B and C the pH remained constant at 8.8 and the
turbidity of the system increased, indicating the coexistence of
micelles, vesicles, and monomers. Upon further addition of
HCl, the pH value decreased again from C to D (vesicles and
monomers are coexisting). At point D the system became milky
white and between D and E the pH remained constant at 7.5
(vesicles, oil emulsion, and monomers are coexisting) and then
dropped abruptly at point E and the sample contained a separate
oil phase floating on top of a clear aqueous phase (point F).
Half of the carboxylate groups were protonated at pH 7.7,21
a value which is much higher than the pKa values of the
monomeric fatty acids in aqueous solution (typically 4.8). This
bilayer effect is known from literature for the case of linear
carboxylic acids.15-18 The same results as in Figure 1 were
obtained with 4(rac).
The observation described above shows that vesicles of 4
are formed spontaneously by adjusting pH. We could also
obtain vesicles in the same pH region starting from acidic
solutions by adding the appropriate amount of NaOH.
Figure 2A shows a freeze fracture electron micrograph of
4(rac) suspension which was spontaneously formed (0.05 M
Tris/HCl, pH 8.0; pH region between point C and point D in
Figure 1): polydisperse multilamellar vesicles are present with
(21) The degree of ionization has been calculated from the amount of
HCl added and the amount of HCl needed for full protonation of the carboxyl
group, and the aggregation morphology could be correlated to the degree
of ionization from the titration curve. Point B represents the minimum
ionization degree for micelles (ca. 0.8). Oil emulsions were formed at a
very low degree of ionization (<0.25). Vesicles were present at intermediate
ionization degrees (0.4-0.7, point C-point D). It should be noted that
the ionization degree is not restricted to 0.5 for the formation of lamellar
phase above Tc. It is in sharp contrast to the crystalline phases of 1:1 acid-
soaps (below Tc) which require a fixed 1:1 stoichiometry of fatty acid to
soap (e.g., the case of n-dodecanoic acid at room temperature).22b
(20) Since the chain-melting phase transition temperature (Tc) of
anhydrous and hydrated samples of 4 (acid-soap 1:1 complexes) are both
below room temperature (Table 1), all lipids during titration were in the
fluid-analogue state (above Tc). It is in clear contrast with lauric acid (7)
whose Tc is higher than room temperature. It has been generally noted
that chain substitution of lipids changes their thermal properties. (Menger,
F. M.; Wood, M. G., Jr.; Zhou, Q. Z.; Hopkins, H. P.; Fumero, J. J. Am.
Chem. Soc. 1988, 110, 6804-6810).

Autopoietic Self-Reproduction of Chiral Fatty Acid Vesicles
J. Am. Chem. Soc., Vol. 119, No. 2, 1997 295
vesicle sizes ranging from 10 nm to at least 1 µm. The presence
of giant vesicles (with diameters greater than 1 µm) has been
confirmed by light microscopy (data not shown). After extru-
sion through membranes with 50-nm pores large vesicles
disappeared and the size became more homogeneous. Spon-
taneously formed and extruded vesicle suspensions were fairly
stable, as estimated by monitoring the optical density (turbidity)
at 450 nm as a function of time, which remained constant over
1 week at room temperature (data not shown).
The internal volume of spontaneously formed 4(rac) vesicles
was in the range of 2.5-4.0 µL/µmol of lipid, somewhat higher
than in the case of spontaneously formed oleic acid vesicles
(1.0-1.6 µL/µmol of lipid).1a The larger internal volume of
4(rac) vesicles is presumably due to the presence of a larger
fraction of bigger vesicles.
As in the case of vesicles obtained from linear fatty
acids,1,15-18 several observations indicate that we are dealing
with equilibrium systems. First of all, these vesicles are formed
spontaneously by simply mixing the surfactant in water at
appropriate pH and concentrationssee also the examples in the
literature.22 Once formed, they are very sensitive to changes
of the external medium, e.g. changes by dilution or changes in
pH.
Figure 3. Dependence of the phase transition temperature (Tc) of
vesicles on the compositions of 4(R) and 4(S) (binary phase diagram).
Total concentration of 4 was 100 mM. Half of the carboxyl groups
were ionized.
It is worthwhile recalling that a high monomer concentration
and the rapid equilibrium between monomers and aggregates
are characteristic features of the fatty acid vesicles, clearly
different from the case of the double chain phospholipids. The
concentration of 4(rac) free monomer in equilibrium with the
vesicles at various pH was determined by ultrafiltration. The
monomer concentration obtained by this means cannot be strictly
considered as the cac (the critical concentration for aggregate
formation), but the fact that it is sensitive to pH certainly reflects
the dynamic character of the vesicles. At pH 7.7, this
concentration was 2.7 mM, at pH 8.0 it was 4.4 mM, at pH 8.2
it was 6.2 mM, and at pH 8.5 it was 9.1 mM. These values are
considerably lower than the cmc at alkaline pH (Table 1: 21
mM). In other words, the lower the pH, the lower is the
monomer concentration, reflecting the lower mole fraction of
ionized 4 in the bilayer.23
(b) Thermal Analysis of the Vesicles. We have measured
thermal phase transitions of vesicles of 4 as a function of the
composition of the (R) and (S) isomers, and a binary phase
diagram has been constructed (Figure 3). Enantiomeric 4(R)
or 4(S) bilayers show a transition peak at 9.5 °C, whereas the
endothermic peak of the 4(rac) bilayer is at 15 °C. There were
two minimal temperatures at the 4(R) mole fractions of ca. 0.2
and 0.8. These temperatures correspond to the two eutectic
points of enantiomeric and racemic crystals. The shape of the
binary phase diagram indicates that the crystal of the racemate
is a racemic compound, as in the case of the bulk compound
(see above). The racemic bilayer has a higher Tc than the
enantiomeric bilayer which is in agreement with the macroscopic
observation that 4(rac) vesicles crystallize more easily than
enantiomeric vesicles at low temperature (see below).
In the second series of measurements, the kinetics of fatty
acid exchange between vesicles was followed qualitatively by
Figure 4. Change in the DSC heating traces upon mixing equal
amounts of 4(R) vesicles and 4(S) vesicles. Concentration of 4 was
100 mM. (a) 4(R) vesicles, before mixing; (b) 3 min after mixing; (c)
30 min after mixing; (d) 4(rac) vesicles (equilibrated for several days).
DSC. Figure 4 shows the DSC traces after mixing separately
prepared vesicle suspensions of 4(R) and 4(S). Both of the
enantiomeric vesicles initially showed an endothermic peak at
ca. 9.5 °C (trace (a) in Figure 4 for 4(R) vesicles). The DSC
measurements carried out 3 min after mixing (trace (b) in Figure
4) showed that a large portion of the molecules have been
exchanged between vesicles and an endothermic peak has
appeared at a temperature which has been obtained for racemic
vesicles. The measurement which was carried out 30 min after
mixing (trace (c) in Figure 4) revealed that the molecules have
been almost completely exchanged and the DSC trace is
superimposable to that of 4(rac) vesicles (trace (d) in Figure
4). Although precise kinetic data are not available from these
measurements, they indicate that the molecular exchange
between vesicles is rapid.
(22) (a) Talmon, Y.; Evans, D. F.; Ninham, B. W. Science 1983, 221,
1047-1048. (b) Hauser, H.; Gains, N.; Mu¨ller, M. Biochemistry 1983,
22, 4775-4781. (c) Kaler, E. W.; Murthy, A. K.; Rodriguez, B. E.;
Zasadzinski, J. A. N. Science 1989, 245, 1371-1374. (d) Safran, S. A.;
Pincus, P.; Andelman, D. Science 1990, 248, 354-356. (e) Cantu, L.; Corti,
M.; Musolino, M.; Salina, P. Prog. Colloid Polym. Sci. 1991, 84, 21-23.
(23) The carboxyl group of monomers in the aqueous phase is thought
to be mostly ionized, since the solubility of protonated fatty acid in water
is much lower than ionized fatty acid. For example, the solubility of
n-dodecanoic acid in water is 12 µM (Bell, G. H. Chem. Phys. Lipids 1973,
10, 1-10), whereas the monomer solubility of sodium n-dodecanoate is
approximately 25 mM (cmc, Table 1).
(c) Optical Activity of the Vesicles. Figure 5A compares
the CD spectrum of 4(R) dissolved in methanol with the CD
spectrum of vesicles from 4(R). In both cases a negative peak
was observed at about 210 nm. The chromophore is the
carboxyl group (λmax ∼ 210 nm, n f π* transition) next to the

296 J. Am. Chem. Soc., Vol. 119, No. 2, 1997
Morigaki et al.
Figure 6. Dependence of the CD signal intensity at 220 nm on the
degree of protonation of the carboxyl group in 4(R). (a) Vesicle
suspensions in water and (b) methanol solution. Concentration of 4-
(R) was 80 mM. The vertical lines indicate the boundaries between
regions where micelles and/or vesicles (lamellar bilayer) are present
in the aqueous solution.
CD spectra. The CD signal intensity decreases as the pH is
raised, and practically no CD signal is observed in a micellar
solution at pH 12.1. A plot of the CD signal intensity at 220
nm versus the mole fraction of the protonated species in solution
is shown in Figure 6. The signal intensity is proportional to
the mole fraction of protonated molecules up to half-protonation.
Above that ratio, measurements became difficult due to
increased turbidity. The linearity in Figure 6 is a clear indication
that only protonated molecules are contributing to the CD signal
around 210 nm; the type of aggregates (micelles or vesicles)
did not affect the intensity.
However, as shown by Figure 6, a linear behavior is also
observed for the methanol solution, and in this case the intensity
of the dichroic signal is a factor 3.5 lower than that of the water
solution. In other words, the aggregation formed in water brings
about an enhancement of the dichroic signal.25 It is due to a
restriction of conformational equilibrium in the aggregated form.
A similar effect has been reported in the case of lecithin reverse
micelles,26 and actually also in the case of phosphatidylcholine
liposomes (unpublished result). Furthermore, we would like
to draw attention to the fact that monomers coexisting with the
vesicles are mostly ionized and therefore the CD signal of
vesicular solution comes exclusively from the protonated species
in vesicles. Formation of the aggregates is accompanied by
protonation of the carboxyl groups due to their higher pKa and
induces the specific CD signal of the aggregates.
Figure 5. CD spectra of 4(R). (A) Comparison of (a) an aqueous
solution (vesicles, concentration of 4(R): 80 mM, pH 7.8) and (b) a
methanol solution (monomers, concentration of 4(R): 88 mM). (B)
pH dependence of the CD spectrum in water. Concentration of 4(R)
was 80 mM: (a) pH 12.1; (b) pH 9.1; (c) pH 8.8; (d) pH 8.6; (e) pH
7.8. All spectra were measured at 25 °C. We have measured CD
spectra at various temperatures and observed that they are identical, if
the bilayers are in the fluid analogue state.
chiral R-carbon. The CD signal of methanolic 4(R) arises from
the molecular optical activity. The vesicles of 4(R) (80 mM,
pH 7.8), on the other hand, showed a significantly more intense
negative CD band at 210 nm, being about 1.8 times larger than
that of the methanolic solution. Since, as we shall discuss later,
only protonated species show a CD signal, and since half of
the molecules are protonated in the bilayer, the normalized CD
signal of the protonated species in the bilayer is about 3.5 times
stronger than the signal observed in methanol solution.
Since the CD signal arises from the carboxyl group which is
localized on the surface of the bilayer, it is expected to be
sensitive to protonation and ionization of the carboxyl group.24
We have therefore studied the pH dependence of the CD signal
of aggregated 4(R), and Figure 5B shows the corresponding
Self-Reproduction of the Vesicles. (a) Catalytic Properties
of the Vesicles. Basically two set of hydrolysis experiments
were performed (Figure 7A).
(i) In the first case, the hydrolysis of anhydride 8 was studied
in a two-phase system, the aqueous phase being a buffer solution
(24) Similar dependence of optical activity on the protonation of carboxyl
groups has been studied for amino acids and hydroxy acids. (a) Jirgensons,
B. Optical ActiVity of Proteins and Other Macromolecules; Springer-
Verlag: Berlin, 1973. (b) Katzin, L. I.; Gulyas, E. J. Am. Chem. Soc. 1968,
90, 247-251.
(25) It is important to notice that this effect is not due to solvent
perturbation effects on the chromophore, as the UV absorption spectrum
in methanol (data not shown) is the same as that in water solution.
(26) Colombo, L. M.; Nastruzzi, C.; Luisi, P. L.; Thomas, R. M. Chirality
1991, 3, 495-502.

Autopoietic Self-Reproduction of Chiral Fatty Acid Vesicles
J. Am. Chem. Soc., Vol. 119, No. 2, 1997 297
Figure 8. Hydrolysis of 8(SS) (circles) of 8(rac) (diamonds) in a two-
phase system consisting initially of an aqueous phase (10 mL of 0.05
M bicine buffer, pH 8.2) and 0.15 mmol of 8 (corresponding to 30
mM of 4 in the aqueous phase after hydrolysis). The concentration of
4 in the aqueous phase was determined as a function of time at a
reaction temperature of 25 (filled symbols) and 10 °C (open symbols).
vesicles (initially 20 mM 4(rac)) about 23 mM 4(rac) was newly
formed in the same period of time. The result clearly shows
that vesicles acted as catalyst, in analogy to the systems based
on normal fatty acids described before.1
(b) Self-Reproduction of the Homochiral and Racemic
Vesicles. We have studied the autocatalytic hydrolysis in the
presence of homochiral and racemic vesicles. The hydrolyses
of 8(SS) and 8(rac) (the 1:1 mixture of 8(RR) and 8(SS)) were
compared starting from the aqueous/anhydride two-phase sys-
tems ((i) in Figure 7A). The results are shown in Figure 8. In
all cases, the reaction initially was rather slow until the rate of
hydrolysis drastically increased due to the formation of ag-
gregates (vesicles). At 25 °C we obtained the same behavior
for the hydrolysis of 8(SS) and 8(rac). The length of the slow
initial phase necessary for the formation of aggregates was the
same for 8(SS) and 8(rac). In both cases, formation of
aggregates led to the enhancement of hydrolysis rates. The
autocatalytic hydrolysis rates of the enantiomer and the racemate
were also the same. It indicates that the enantiomeric composi-
tion of the bilayer does not affect the aggregation property of
4 and the catalytic ability of the vesicles at this temperature.
The mean size and size distribution of the vesicles formed
during the hydrolysis of 8(SS) has been determined by analyzing
freeze fracture electron micrographs. Figure 2B shows the
results analyzed after 8 days reaction at 25 °C. Most of the
vesicles formed during the reaction were unilamellar and smaller
than 200 nm with a mean diameter of 60 nm.27
Figure 7. (A) Schematic illustration of the hydrolysis experiments,
(i) starting from a buffer solution without vesicles and (ii) with
preformed vesicles. (B) Hydrolysis of 8(mix) (0.2 mmol) at 25 °C in
a two-phase system containing an aqueous phase (10 mL, 0.05 M bicine
buffer, pH 8.7) initially either (a) with vesicles of 20 mM 4(rac) or
(b) without vesicles. The increase in concentration of 4 in the aqueous
phase was determined as a function of time. In the case of (a), slight
decrease of the concentration was observed for the initial 2 h, due to
the uptake of 4 molecules from the vesicles into the oil phase. We
have plotted in the graph the concentration increase that followed this
initial process to make the true comparison of hydrolysis rates.
at pH ∼8.5 and the oil phase of the anhydride on the surface of
the aqueous phase. In this case the anhydride is hydrolyzed
and fatty acids are formed which partition into the aqueous phase
where aggregates (vesicles) are formed as soon as the critical
concentration for aggregate formation is reached. The hydroly-
sis of the anhydride is accelerated autocatalytically at the onset
of the newly formed vesicles.
(ii) In the second series of experiments, the rate of anhydride
hydrolysis was studied in the presence of vesicles from the
beginning. In other words, the two-phase system was composed
of an aqueous buffer system containing preformed vesicles and
a supernatant anhydride phase. In this case the rate of hydrolysis
is accelerated by the presence of vesicles from the very
beginning.
The concentration increase of 4 in the aqueous phase due to
the hydrolysis is shown in Figure 7B. Without vesicles, the
reaction initiallysbefore vesicles are formedsis slow, while
in contrast, the presence of vesicles catalyzes the hydrolysis of
8(mix). After 30 h, for example, about 3 mM 4(rac) was
formed in the absence of vesicles, while in the presence of the
Let us compare now the hydrolysis of 8(SS) and 8(rac) at
10 °C. This temperature is higher than the Tc of vesicles of
4(S), but lower than the Tc of vesicles of 4(rac). We have
observed quite different behaviors in the two cases. First, the
lag time before the onset of the autocatalytic hydrolysis was
considerably shorter for the hydrolysis of 8(SS) compared with
that of 8(rac). Secondly, the hydrolysis of 8(rac) gave rise to
(27) Some giant vesicles with diameters considerably larger than 1 µm
were also present after the reaction (data not shown).

298 J. Am. Chem. Soc., Vol. 119, No. 2, 1997
Morigaki et al.
still contained a considerable amount of aqueous solution, the
real composition of the solid phase is most likely very close to
that of the racemate.
The observation just described can be interpreted as follows.
As 8(SS) is hydrolyzed at 10 °C in the presence of 4(R) vesicles,
the enantiomeric composition of the bilayer changes continu-
ously toward the racemate. Tc of the bilayers changes accord-
ingly, since the racemic bilayer of 4 has a higher Tc than the
enantiomers. When the amount of 4(S) in the bilayer exceeds
a critical value, Tc becomes higher than the reaction temperature
(10 °C) and formation of the precipitates starts. The solid phase
is the racemate that has the highest Tc value. Formation of the
racemic precipitates leaves vesicles in the aqueous phase with
a higher enantiomeric excess of 4(R). All this is in sharp
contrast to the hydrolysis of 8(RR) in the presence of 4(R)
vesicles, as in this case the enantiomeric composition of bilayers
does not change and the vesicle suspension remains stable.
(c) Stereoselectivity of the Vesicle Catalysis. One interest-
ing question is whether the self-reproduction process occurs with
some degree of enantioselectivity. In order to answer the
question, we compared the hydrolysis rates of 8(RR) and 8-
(SS) in the presence of enantiomerically pure vesicles (either
4(R) or 4(S)). Typically 0.1 mmol of 8(RR) or 8(SS)
(corresponding to 20 mM of 4(R) or 4(S) in the aqueous phase
after hydrolysis) was overlaid on 10 mL vesicle suspensions of
4(R) or 4(S) (40 mM, pH 8.3) and the hydrolysis reaction was
followed by FT-IR and CD spectroscopies. The reactions were
conducted at both 25 and 10 °C repeatedly. In the limit of error
(which in this type of experiment is relatively large, ca. 10%),
anhydrides with the same chirality as the vesicles were
hydrolyzed at the same rate as anhydrides with the opposite
chirality with respect to the vesicles, indicating that the chiral
vesicles of 4 cannot induce significant enantioselectivity.
Conversely, a significant difference was observed when we
compared the hydrolysis of 8(RR) and the diastereomeric
mixture 8(mix). The appropriate amount of 8 was overlaid on
vesicle suspensions of 4(R) and the concentration increase of 4
during the hydrolysis reaction was monitored. Figure 10 shows
a typical result. The release of 4 was slightly faster during the
hydrolysis of 8(RR) compared with 8(mix). We observed the
same trend also in the presence of 4(rac) vesicles (8(RR) was
hydrolyzed faster than 8(mix)). The difference of hydrolysis
rates between 8(RR) and 8(mix) is due to differences in the
reactivity of the mesoform 8(RS) in comparison with the two
enantiomers 8(RR) and 8(SS). This observation demonstrates
diastereomeric effects in catalytic vesicles.
Figure 9. Optical density (turbidity) change at 800 nm during the
hydrolysis of (a) 8(SS) and (b) 8(RR) in the presence of 4(R) vesicles.
Either 8(RR) or 8(SS) (0.03 mmol; corresponding to 20 mM of 4(R)
or 4(S) after hydrolysis, respectively) was overlaid on the aqueous phase
(3 mL) containing the vesicle suspension of 40 mM 4(R) and was
hydrolyzed at 10 °C.
phase separation: a white, gel-like material was formed after
ca. 150 h (concentration of ca. 20 mM) and precipitated out.
At this point, the hydrolysis rate was considerably slowed down.
This precipitate is due to the fact that at the given temperature
(10 °C) the racemic vesicles are below their Tc and are not stable.
On the other hand, the aggregates formed by 4(S)swhich were
regular vesicles as confirmed by electron micrographssare at
this temperature in a fluid-liquid analogue state and stable. The
hydrolysis reaction continued until 8(SS) was exhausted.
In order to evidence even more clearly this effect, the
experiment illustrated in Figure 9 was carried out, in which we
started from already formed vesicles of 4(R). The process was
monitored by following the optical density (turbidity) of the
vesicle dispersion at the arbitrary chosen wavelength of 800
nm. Hydrolysis of 8(RR) in the presence of 4(RR) vesicles
changed the turbidity of vesicles only slightly and gradually.28
In contrast, in the case of 8(SS) hydrolysis, an abrupt increase
of the turbidity due to the phase separation was observed, a
process which can easily be seen by the eyes. The macroscopic
feature of the condensed phase was a heterogeneous mixture
of white granulates and a gel-like substance swollen with the
aqueous solution. The aqueous phase was separated from the
precipitated phase by centrifugation at 4 °C. The supernatant
aqueous phase was turbid, indicating the presence of vesicles.
To obtain the enantiomeric composition of the separated phases,
4 was extracted from each phase into isooctane and the
concentration and optical activity were determined by FT-IR
and CD measurements, respectively. The precipitates contained
58 ( 2% 4(R) and 42 ( 2% 4(S), whereas the aqueous phase
contained 78 ( 2% 4(R) and 22 ( 2% 4(S). In other words,
the enantiomeric excess in the aqueous phase is considerably
higher than that in the precipitated phase, and the latter has an
enantiomeric composition that is close to that of the racemate.
Considering the quantitative separation of the two phases by
centrifugation was not very efficient and that the precipitates
Concluding Remarks
The vesicles from 2-methyldodecanoic acid (4) share with
the previously described normal fatty acids1 two important
features: the spontaneous vesiculation, namely the capability
of building vesicles without additional energy input, and the
capability of undergoing an autocatalytic self-reproduction
process. The main idea was to possibly combine exponential
autocatalysis and enantioselectivity, i.e., to use the power of an
exponential growth process to kinetically discriminate between
two stereoisomers. It did not work the way we wanted at room
temperature. However, a marked difference in the behaviors
of homochiral and racemic vesicles was observed at 10 °C. The
fact that “racemic vesicles” separate out in a gel-like form
potentially provides a way to enrich the enantiomeric purity of
the starting mixture in vesicles.
(28) The gradual change of the absorbance in Figure 9 is presumably
due to the constraints of experimental setup. We conducted the experiment
in the spectrophotometer using a round glass tube in order to make efficient
mixing of the aqueous phase possible. Consequently, we could obtain only
poor transmission of light through the cell. Furthermore, the rather vigorous
stirring of the aqueous phase during the reaction might have caused
formation of oil emulsions of 8 in the aqueous phase during the reaction,
which would have influenced the turbidity.
One may ask why the difference between homochiral and
racemic vesicles observed at 10 °C is not reflected in an
observable difference in the self-reproduction rate at room
(29) Gerrand, W.; Thrush, A. M. J. Chem. Soc. 1953, 2117-2120.

Autopoietic Self-Reproduction of Chiral Fatty Acid Vesicles
J. Am. Chem. Soc., Vol. 119, No. 2, 1997 299
solvent system on EM Reagents silica gel 60 (230-400 mesh). All
reactions were carried out under an atmosphere of nitrogen in glassware
that had been flame dried under a stream of nitrogen.
Synthesis of (4R)-3-(Dodecanoyl)-4-benzyl-2-oxazolidinone (2(R)).
To a magnetically stirred solution of 3.5 g (20.0 mmol) of (R)-4-benzyl-
2-oxazolidinone (1) in 60 mL of anhydrous THF under nitrogen at -78
°C was added dropwise 12.6 mL (1.6 M in hexane, 20.2 mmol) of
n-butyllithium over a 15-min period. Freshly distilled dodecanoyl
chloride (5.08 mL, 22.0 mmol) was added in one portion after
completion of the addition of butyllithium. The resulting clear, yellow
solution was warmed to 0 °C over a 1-h period and then stirred for an
additional hour. Excess dodecanoyl chloride was quenched by the
addition of 20 mL of saturated aqueous ammonium chloride. Volatiles
were removed in vacuo and the product was extracted with three 20-
mL portions of dichloromethane. The combined organic extracts were
successively washed with 10 mL of 1 M aqueous potassium carbonate
and 10 mL of brine (saturated solution of NaCl), dried over anhydrous
sodium sulfate, and filtered. The solvent was removed in vacuo to
leave 10.0 g of a yellow solid, which was purified by flash chroma-
tography (9:1 hexane/ethyl acetate as eluant) to afford 6.54 g (91%) of
(4R)-3-(dodecanoyl)-4-benzyl-2-oxazolidinone (2(R)) as a white crys-
talline solid: mp 46-47 °C; 1H-NMR (300MHz, CDCl3) δ ppm 0.85
(t, 3H, CH3, J ) 7.2 Hz), 1.1-1.5 (m 18H, CH2), 2.8 (dd, 1H, J )
13.3, 9.6 Hz, CH2C6H5), 2.9 (m, 2H, CH2CO), 3.3 (dd, 1H, J ) 13.4,
3.3 Hz, CH2C6H5), 4.1 (m, 2H, CH2O), 4.65 (m, 1H, CHN), 7.1-7.4
(m, 5H, ArH); [R]D -71.6° (c 1, CH3COCH3). Elemental Anal. Calcd
for C22H33NO3: C, 73.49; H, 9.25; N, 3.90. Found: C, 73.50; H, 9.23;
N, 3.80.
Synthesis of (4R)-3-((2′R)-2′-Methyldodecanoyl)-4-benzyl-2-ox-
azolidinone (3(RR)). To a -78 °C solution of 14.7 mL (14.7 mmol,
1.1 equiv, 1.0 M in tetrahydrofuran) of sodium hexamethyldisilylamide
was added dropwise a 0 °C solution of 4.8 g (13.4 mmol, 1 equiv) of
(4R)-3-(dodecanoyl)-4-benzyl-2-oxazolidinone (2(R)) in 14.4 mL of
THF. After the reaction mixture was stirred at -78 °C for 1 h, 9.5 g
(4.2 mL, 67.0 mmol, 5 equiv) of iodomethane in 1.4 mL of THF at
-78 °C was added. The solution was stirred 4 h at -78 °C and then
quenched by the addition of 20 mL of aqueous saturated ammonium
chloride solution. Volatiles were removed by rotary evaporation, and
the resultant slurry was extracted with three 20-mL portions of
methylene chloride. The combined organic fractions were washed with
100 mL of aqueous 1 M sodium sulfite solution, dried over anhydrous
sodium sulfate, and concentrated in vacuo to give 5.0 g of a yellow
oil. Purification by flash chromatography (9:1 hexane/ethyl acetate as
eluant) gave 3.05 g (61%) of the title compound as a clear oil: 1H-
NMR (300 MHz, CDCl3) δ ppm 0.85 (t, 3H, CH3), 1.1-1.45 (d + m,
19H, CH3, CH2), 1.3-1.45 (m, 1H, CHH), 1.65-1.8 (m, 1H, CHH),
2.7 (dd, 1H, J ) 13.3, 9.6 Hz, CH2C6H5), 3.25 (dd, 1H, J ) 13.4, 3.3
Hz, CH2C6H5), 3.65-3.75 (m, 1H, CHCO), 4.1 (m, 2H, CH2O), 4.65
(m, 1H, CHN), 7.1-7.4 (m, 5H, ArH), [R]D -88.7 (c 1, CH3COCH3).
Elemental Anal. Calcd for C23H35NO3: C, 73.95; H, 9.44; N, 3.75.
Found: C, 73.90; H, 9.40; N, 3.73.
Figure 10. Hydrolysis of 0.2 mmol of 8(RR) (diamonds) or 0.2 mmol
of 8(mix) (circles) (corresponding to 40 mM of 4(R) and 4(rac) after
hydrolysis, respectively) in a two-phase system containing an aqueous
phase (10 mL, 50 mM bicine buffer, pH 8.7) initially with vesicles of
20 mM 4(R). The increase in concentration of 4 in the aqueous phase
was determined as a function of time at a reaction temperature of 25
°C.
temperature where vesicles are stable and do not precipitate.
Most likely, above Tc the bilayer are too fluid to allow for a
discrimination of the pair interactions R-R with respect to R-S.
This points out an interesting difficulty for further studies on
this field: above Tc the interaction difference may be too small
to give a rate effect, however below Tc it may be so large that
phase separation with destruction of the vesicles may occur.
Materials and Methods
Reagents. All the following reagents used were purchased from
Fluka (Buchs, Switzerland): (R)-4-benzyl-2-oxazolidinone (1(R)), (S)-
4-benzyl-2-oxazolidinone (1(S)), n-butyllithium (n-BuLi), methyl iodide
(MeI), dodecanoyl chloride (C11H23COCl), dichloromethane, tetrahy-
drofuran (THF), sodium hexamethyldisilylamide (NaN(SiMe3)2), eu-
ropium(III) tris[3-(heptafluoropropylhydroxymethylene)-d-camphorat]
(Eu(hfc)3), (R)-(+)-R-phenylethylamine, thionyl chloride (SOCl2), N,N-
dimethylformamide (DMF), diisopropylamine, pyridine (C5H5N), 1,3-
dimethyl-2-oxohexahydropyrimidine (DMPU), triethylamine, n-dode-
canoic acid (7) (g99%), oleic acid (g99%), and arsenazo III.
Pinacyanol chloride was from Sigma (St. Louis, MO, USA), and
sepharose 4B from Pharmacia (Uppsala, Sweden). Standard 1 M
solutions of NaOH and HCl were purchased from Merck (Darmstadt,
Germany). Lithium hydroperoxide (LiOOH) was generated in situ from
H2O2 and LiOH.
Synthesis of (R)-2-Methyldodecanoic Acid (4(R)). A solution of
3.0 g (8 mmol) of (4R)-3-((2′R)-2′-methyldodecanoyl)-4-benzyl-2-
oxazolidinone (3(RR)) in 160 mL of 3:1 THF/distilled water solution
(0.05 M) was treated at 0 °C with 6.5 mL of 30% H2O2 (8 equiv)
followed by 680 mg of 98% LiOH‚H2O (2 equiv). After the solution
was stirred at 0 °C for 3 h, the excess peroxide was quenched at 0 °C
with a 10% excess of 1.5 N aqueous sodium sulfite. The bulk of the
THF was removed in vacuo and the resulting mixture was acidified to
pH 1 by the addition of an aqueous 10% hydrochloric acid solution.
The aqueous layer was extracted with three 20-mL portions of
methylene chloride. The combined organic extracts, containing a
mixture of the title compound 4(R) and of the recovered oxazolidinone
1, were dried over anhydrous sodium sulfate, filtered, and concentrated
in vacuo to give 2.8 g of a yellow solid. This slurry was chromato-
graphed (flash silica gel, gradient from 8:2 hexane/ethyl acetate to 100%
ethyl acetate) to give 1.3 g of the recovered (R)-4-benzyl-2-oxazoli-
dinone (1) as a white crystalline solid and 1.5 g (90%) of pure (R)-2-
methyldodecanoic acid (4(R)) as a colorless liquid: bp 120 °C (5 ×
When necessary solvents and reagents were distilled under nitrogen
prior to use. Dichloromethane and diisopropylamine were distilled from
calcium hydride. THF, diethyl ether, and triethylamine were distilled
from sodium metal/benzophenone ketyl.
General Methods for Synthesis and Product Characterization.
Infrared spectra were recorded on a Perkin Elmer FT-IR 16PC
spectrometer. 1H-NMR spectra were recorded on a Bruker AC300
spectrometer at ambient temperature. Chemical shifts are reported in
ppm from tetramethylsilane on the δ scale, with the solvent resonance
employed as the internal standard (deuteriochloroform at 7.24 ppm).
Optical rotations were measured on a Jasco DIP-181 digital polarimeter.
Melting points are uncorrected. Analytical thin-layer chromatography
was performed on EM Reagent 0.25 mm silica gel 60-F plates.
Purification of reaction products was carried out by liquid chromatog-
raphy using a forced flow (flash chromatography) of the indicated
1
10-3 mmHg); H-NMR (300 MHz, CDCl3) δ ppm 0.85 (t, 3H, CH3),
1.17 (d, 3H), 1.18-1.35 (m, 16H, CH2), 1.35-1.45 (m, 1H, CHH),
1.55-1.75 (m, 1H, CHH), 2.45 (m, 1H, CHCO); [R]D -15.6° (c 1,

300 J. Am. Chem. Soc., Vol. 119, No. 2, 1997
Morigaki et al.
three anhydrides was prepared by the same procedure used for the
synthesis of 8(RR), starting from racemic 2-methyldodecanoic acid (4-
(rac)).
Scheme 4
Differential Scanning Calorimetry Measurements. The thermal
phase transitions of anhydrous and hydrated fatty acid samples of 4, 7,
and 8 were determined by differential scanning calorimetry (DSC) using
a TA-4000 system from Mettler (Na¨nikon-Uster, Switzerland). For
anhydrous samples, a scan rate of 1 K/min was used. In the case of
hydrated fatty acids, the samples were first prepared in glass tubes by
dispersing appropriate amounts of lipid in water. NaOH corresponding
to half the equivalent of the lipid was added to the water in advance to
form a 1:1 acid-soap mixture. Equilibration was ensured by leaving
the sealed tubes at room temperature for 5 days. The samples were
cooled slowly (typical rate: 0.3-1 K/min) down to a temperature
substantially below the phase transition temperature (Tc), kept at that
temperature for 1 or 2 h for equilibration, and then heated (5 K/min)
to get the DSC traces. We used a relatively high scan rate to obtain a
reasonable signal-to-noise ratio from a limited amount of sample.
Cooling and heating processes were repeated for the same sample and
it was confirmed that DSC traces are reproducible.
In order to get a qualitative idea on the kinetics of exchange of
molecules between vesicles, DSC measurements were carried out after
mixing equal amounts of 4(R) and 4(S) vesicle suspensions. The
change in the thermal phase transition temperature (Tc) was observed
as a function of time after mixing.
Cmc Determinations. The critical concentration for micelle forma-
tion (cmc) of alkaline fatty acid solutions was determined either by
conductivity measurements using a CDM 83 conductometer (Radiom-
eter, Copenhagen, Denmark) or spectrophotometrically using pinacyanol
chloride.30,31 In the latter method, 8 µL of a 1.3 mM solution of
pinacyanol chloride in methanol was added to 1 mL of the aqueous
soap solution and the absorbance was measured at 605 nm with an
Uvikon 820 spectrometer (Kontron, Zu¨rich, Switzerland). Concentra-
tions where abrupt changes in the concentration dependence of either
conductivity or absorption of pinacyanol chloride occurred were used
as cmc values.
Titration of Alkaline Fatty Acid Solutions with HCl. Ten
milliliters of a micellar fatty acid solution were first prepared by
dissolving 0.08 M fatty acid in 0.088 M NaOH (10 mol % excess with
respect to the fatty acid). Aliquots of 0.5 mL of this solution were
pipeted into small glass vials and various amounts of 1 M HCl were
added to each vial. The glass vials were sealed and the samples were
equilibrated at room temperature for at least 5 days. Then pH
measurements were conducted using a PHM 82 pH meter (Radiometer,
Copenhagen, Denmark) with a micro pH-electrode (Ingold, Steinbach,
Germany).
Preparation of Fatty Acid Vesicles. The fatty acid was weighed
into a glass flask and an appropriate amount of buffer solution was
added. NaOH was then added to readjust the pH. This procedure led
to the formation of a turbid suspension containing spontaneously formed
vesicles. Homogenization of the suspension was achieved by vortexing.
The samples were left for a few days to equilibrate. In certain cases,
the mean size of the vesicles was decreased by extrusion through
Nucleopore polycarbonate membranes filters with pores of defined
size,32 using “The Extruder” from Lipex Biomembranes Inc. (Vancou-
ver, Canada). This process reduced the maximum size of the vesicles
to approximately the size of the pores and made the suspensions more
homogeneous.
CH3COCH3). Elemental Anal. Calcd for C13H26O2: C, 72.84; H,
12.23. Found: C, 72.85; H, 12.20.
(R)-2-methyldodecanoic acid (4(R)) was quantitatively converted into
the corresponding methyl ester by reaction with an ethereal solution
of diazomethane. Chiral shift studies on this methyl ester using Eu-
(hfc)3 indicated that the product 4(R) had an enantiomeric excess g98%.
The same value (d.e. g98%) was obtained by GLC analysis (Carbowax
20 M/230 °C) on the diastereomeric mixture of amides 5a and 5b
obtained by reaction of the acid 4(R) with (R)-(+)-R-phenylethylamine.
Synthesis of the Diastereomeric Amides 5a and 5b. To a
magnetically stirred solution of 140 mg (0.65 mmol) of 2-methyldode-
canoic acid (4) in anhydrous diethyl ether were added dropwise 71.4
µL (0.98 mmol) of thionyl chloride and 4.6 µL of DMF. After standing
at 25 °C for 2 h, the mixture was decanted, concentrated, dissolved in
2 mL of CH2Cl2, and added to a CH2Cl2 solution of (R)-(+)-R-
phenylethylamine (126.3 mL, 0.98 mmol) and triethylamine (136 mL,
0.98 mmol). The mixture was stirred for 30 min at 25 °C, then washed
with 2 N HCl and water, dried over anhydrous Na2SO4, and filtered.
The solvent was removed in vacuo to give 186 mg (90%) of the
corresponding (RR)- and (RS)-R-phenylethylamides as white solid: mp
60-62 °C. 1H-NMR (300 MHz, CDCl3) δ 0.85 (t, 3H, CH3), 1.15 (d,
3H, CH3), 1.15-1.4 (m, 17H, CH2CHH), 1.5 (d, 3H, CH3), 1.5-1.7
(m, 1H, CHH), 2.15 (m, 1H, CHCO), 5.15 (m, 1H, CHNH), 5.65 (d,
1H, NH), 7.2-7.4 (m, 5H, C6H5). GLC analysis (Carbowax 20m, 230
°C) of the amides gave the relative amounts of the diastereomers.
Synthesis of Racemic 2-Methyldodecanoic Acid (4(rac)).
A
solution of diisopropylamine (4.12 mL, 29 mmol) in anhydrous THF
(25 mL) was charged in a flask under nitrogen atmosphere. After
cooling at 0 °C, 18.1 mL (29 mmol) of n-butyllithium (1.6 M in hexane)
was slowly added in order to keep the temperature at 0 °C. After
standing at this temperature for an additional 20 min, n-dodecanoic
acid (7) (2.8 g, 14 mmol) was added dropwise while maintaining the
reaction temperature below 0 °C. A milky white solution formed and
after 30 min DMPU (1.7 mL, 14 mmol) was added and the mixture
was stirred at room temperature for 1 h. Then it was cooled again at
0 °C and CH3I (0.9 mL, 15 mmol) was rapidly added. The reaction
was complete by stirring the mixture at room temperature overnight.
The product was recovered by neutralization with ice cooled 10% HCl,
followed by extraction with Et2O (3 × 30 mL). The combined organic
layers were washed with water and brine, dried over anhydrous Na2-
SO4, and concentrated in vacuo to obtain 2.9 g of a yellow liquid. Pure
4(rac) (2.7 g, 90% yield) was obtained by distillation. Bp 120 °C (5
1
× 10-3 mmHg); H-NMR (300 MHz, CDCl3) δ 0.85 (t, 3H, CH3),
1.17 (d, 3H); 1.18-1.35 (m, 16H, CH2), 1.35-1.45 (m, 1H, CHH),
1.55-1.75 (m, 1H, CHH), 2.45 (m, 1H, CHCO).
Synthesis of (RR)-(-)-2-Methyldodecanoic Anhydride (8(RR))
(Scheme 4). A solution of 2 g (9.3 mmol) of 4(R) and 0.75 mL (9.3
mmol) of anhydrous pyridine in 9.5 mL of anhydrous Et2O was cooled
at -10 °C. Dropwise addition of 0.34 mL (4.65 mmol) of thionyl
chloride afforded immediately a white precipitate. The mixture was
stirred at -10 °C for 15 min, then filtered, and the residue was rapidly
washed with anhydrous Et2O. The solvent was removed in vacuo to
leave 1.9 g of yellow oil. Purification by distillation (bp 165 °C, 5 ×
10-3 mmHg) gave 1.5 g (80%) of the title compound as a colorless
oil. 1H-NMR (300MHz, CDCl3) δ 0.85 (t, 6H, CH3), 1.17 (d 6H, CH3),
1.18-1.35 (m, 32H, CH2), 1.35-1.45 (m, 2H, CHH), 1.55-1.75 (m,
2H, CHH), 2.51 (m, 2H, CHCO). IR: 1748 and 1815 cm-1 (CdO).
Elemental Anal. Calcd for C26H50O5: C, 69.04; H, 13.37. Found: C,
69.00; H, 13.37. [R]D -17.2° (c 1, CH3COCH3).29 DSC heating
measurements gave an endothermic peak at around 12 °C.
The formation of vesicles was confirmed mainly by microscopic
methods. Electron micrographs were taken by using the freeze-fracture
technique as described before.1a The size of vesicles was determined
from the electron micrographs by taking into account nonequatorial
fracturing.33 Vesicle dispersions were also analyzed by light microscopy
(30) (a) Mukerjee, P.; Mysels, K. J. Critical Micelle Concentrations of
Aqueous Surfactant Systems; NSRDS-NBS, 1971. (b) Campbell, A. N.;
Lakshminarayana, G. R. Can. J. Chem. 1965, 43, 1729-1737. (c) Harva,
O. Recl. TraV. Chim. 1956, 75, 112-116. (d) Markina, Z. N.; Tsikurina,
N. N.; Kostova, N. Z.; Rebinder, P. A. Kolloid Zh. 1964, 26, 76-82.
(31) (a) Corrin, M. L.; Klevens, H. B.; Harkins, D. J. Chem. Phys. 1946,
14, 480-486. (b) Menger, F. M.; Portnoy, C. E. J. Am. Chem. Soc. 1967,
89, 4698-4703.
Synthesis of (SS)-(+)-2-Methyldodecanoic Anhydride (8(SS)). 8-
(SS) was synthesized in the same manner starting from 4(S). DSC
heating measurements gave an endothermic peak around 12 °C, almost
identical to 8(RR).
Synthesis of a Mixture of (RR)-, (SS)- and (RS)-(-)-2-Methyl-
dodecanoic Anhydride (8(mix)). The diastereomeric mixture of the
(32) Mayer, L. D.; Hope, M. J.; Cullis, P. R. Biochim. Biophys. Acta
1986, 858, 161-168.

Autopoietic Self-Reproduction of Chiral Fatty Acid Vesicles
J. Am. Chem. Soc., Vol. 119, No. 2, 1997 301
for the presence of giant vesicles, using an Axioplan light microscope
from Zeiss (Germany), connected to a Sony XC-75CE video camera.
cm-1 by FT-IR. The typical procedure was as follows: 50 µL of the
vesicle suspension was acidified by adding 100 µL of 1 N HCl. The
protonated fatty acid was then extracted into 200 µL of isooctane and
FT-IR was measured by a Nicolet 5SXC FT-IR spectrometer using a
CaF2 cell with a 0.2 mm optical path length.
Determination of the Internal Volume of the Vesicles. The
internal aqueous volume of 4(rac) vesicles was determined with
arsenazo III34 as water-soluble dye marker. The vesicle suspension
was first prepared in the presence of arsenazo III (4.9 mM), and vesicles
with entrapped dye molecules were separated from non-entrapped
arsenazo III by size exclusion chromatography (sepharose 4B, length
5 cm, diameter 0.6 cm). The internal aqueous volume of the vesicles
was determined by assuming equal dye concentration both inside and
outside of the vesicles. The elution was performed with an eluent
saturated with monomeric 4(rac) (3 mM in 50 mM Tris/HCl, pH 7.7)
to avoid dissolution of vesicles during the filtration. The concentration
of the fatty acid was determined by FT-IR spectroscopy (see below).
For the calculation of the internal volume, the free monomer
concentrationsas estimated by ultrafiltration experiments (see below)shas
been taken into account: [fatty acid]in vesicles ) [fatty acid]total - [fatty
Kinetic Measurements. Typical experimental conditions were as
follows:1a
(a) Hydrolysis of 8 in the Absence of Vesicles. Inside a flat-bottom
test tube (diameter 2.5 cm) 0.15 mmol (6.15 mg) of 8 were added onto
10 mL of bicine buffer (0.05 M, pH 8.3), containing a magnetic stirrer
bar (length 2.4 cm, thickness 0.5 cm; 100 rpm), at 25 °C. From time
to time 50 µL of the aqueous phase were withdrawn and the
concentration of 4 was determined by FT-IR.
(b) Hydrolysis of 8 in the Presence of Vesicles of 4. All the
conditions were the same as those just described, with the only
exception that the aqueous phase contained spontaneously formed
vesicles of 4, typically at a concentration of 40 mM.
Some of the hydrolysis experiments were performed in a diode array
spectrophotometer (Hewlett Packard, Model 8452A). In these experi-
ments, the turbidity of the reaction solution was monitored continuously
under the following experimental configuration. A flat-bottom test tube
(diameter 1.0 cm) was filled with 3 mL of 4 vesicle solution (40 mM
of 4 in 0.05 M bicine buffer at pH 8.3). The temperature of the cell
was controlled by a thermostat. The solution was stirred constantly
using a magnetic stirrer. The appropriate amount of 8 (typically 0.03
mmol, corresponding to 20 mM of 4 after hydrolysis) was put onto
the surface of the solution and the hydrolysis reaction was started.
Centrifugation. Hydrolysis experiments which were carried out at
10 °C using vesicles and anhydride molecules of opposite chirality led
to the formation of a gel-like solid phase (see below). This solid phase
was separated from the aqueous phase by centrifugation at 4 °C (6.7
× 103 × g, 70 min) using a ALC micro CENTRIFUGETTE 4214
centrifugator.
acid]monomer
.
Ultrafiltration of Fatty Acid Vesicle Samples. The experiments
for the determination of the amount of non-aggregated monomeric fatty
acid at a particular pH value were carried out in the following way:
80 mM 4(rac) vesicles were first prepared in 50 mM Tris/HCl buffer
(pH ranging from 7.7 to 8.5) and equilibrated for 3 days at room
temperature. The vesicle sample was then subjected to ultrafiltration
by using a Centricon-30 membrane filter device (molecular weight
cutoff 30 000, Amicon Inc.: Beverly, MA) at a centrifugation speed
of 2000 rpm for 90 min. The concentration of fatty acids appearing in
the ultrafiltrate was determined by FT-IR spectroscopy.
Circular Dichroism Measurements. The optical activity of vesicles
of 4 was determined by circular dichroism (CD) measurements using
a Jasco J-600 instrument. The cells used for the measurement had
path lengths of either 0.02 or 0.05 cm depending on the turbidity of
the samples.
Infrared Spectroscopy (FT-IR). The concentrations of 4 in vesicle
suspensions were determined through the intensity of CdO(st) at 1711
Acknowledgment. We would like to thank Ernst Wehrli and
Michaela Wessicken (ETH-Zu¨rich) for performing the electron
microscopy analysis. Support from the Swiss National Science
Foundation (CHiral 2, 20-44′277.95) is gratefully acknowledged.
(33) Egelhaaf, S. U.; Wehrli, E.; Mu¨ller, M.; Adrian, M.; Schurtenberger,
P. J. Microsc. In press.
(34) Weissmann, G.; Collins, T.; Evers, A.; Dunham, P. Proc. Natl. Acad.
Sci. U.S.A. 1976, 73, 510-514.
JA961728B