Spontaneous membrane fusion induced by chemical formation of ceramides in a lipid bilayer

Article abstract of DOI:10.1021/ja0652969

Mere chemical generation of ceramide and related double-chain lipids in the membrane of small unilamellar vesicles (SUVs) induces fusion of the vesicles. The lipids can be successfully prepared by dehydrocondensation between single-chain lipids (fatty acids and sphingosine or its analogues) in a lipid bilayer of the SUV by using a combination of 2-chloro-4,6-dimethoxy-1,3,5-triazine and amphiphilic tertiary amine catalysts, a process that can be compared to a successive enzyme model system for a fatty acyl-CoA synthetase followed by acyltransferase. The SUV spontaneously undergoes membrane fusion upon this internal chemical stimulation by the artificial enzyme system. Copyright

Full text of DOI:10.1021/ja0652969

Published on Web 10/20/2006
Spontaneous Membrane Fusion Induced by Chemical Formation of
Ceramides in a Lipid Bilayer
Munetaka Kunishima,*,†,‡ Masafumi Tokaji,† Keisuke Matsuoka,‡ Jin Nishida,‡ Masanori Kanamori,†
Kazuhito Hioki,† and Shohei Tani†
Faculty of Pharmaceutical Sciences, Kobe Gakuin UniVersity, and PRESTO, JST, Nishi-ku, Kobe 651-2180, Japan
Received July 24, 2006; E-mail: kunisima@pharm.kobegakuin.ac.jp
Although membrane fusion is an essential and fundamental event
in living organisms, the molecular mechanism of the process is
not yet fully understood. As structures of biomembranes are
asymmetric and heterogeneous, lipid composition appears to play
a significant role in morphological changes of the membranes.1-3
However, the direct relationship between dynamic alterations of
the lipid composition of a membrane and dynamic morphological
transformation of membrane structures has not been elucidated
previously. We report here, for the first time, that mere chemical
generation of ceramide and related double-chain lipids in the
membrane of small unilamellar vesicles (SUVs) induces fusion of
the vesicles.
Morphological changes in cellular membranes actually occur
through the cooperation of multiple factors,2,3 most likely with
fusogenic proteins controlling these events.4 Because the process
involves such a complex biological system, it has been difficult to
investigate the function of individual factors, particularly the role
of lipids, on fusion or fission events. To elucidate lipid function, it
is necessary to employ a simple model system that excludes any
extra biological factors, even identified factors such as enzymes,
in which composition of membrane lipids can be altered by
chemical reactions with minimal perturbation. Thus, we have
successfully developed an artificial system in which the synthesis
of double-chain lipids of the ceramide family occurs by dehydro-
Figure 1. Concept of the membrane fusion by ceramide synthesis.
condensation of fatty acids and a sphingosine or its analogues at
the surface of a liposome membrane.
To that end, we employed 1,2-dihydroxy-4-azahexadecane (2p,
pseudo-sphingosine with a dodecyl group at the nitrogen atom)
instead of sphingosine 2. With its two hydroxyl groups and one
amino group, the chemical properties of 2p would be similar to
those of 2. Dimethylglycine hexadecyl ester 3a was used as a
tertiary amine catalyst. Liposome fusion was monitored by a probe
mixing assay or a probe dilution assay based on fluorescence
resonance energy transfer between NBD-PE and Rh-PE.6
The results of the probe mixing assay are shown in Table 1.
Fusion was estimated from the efficiency of energy transfer value,
E (%),7 which should increase upon fusion. As we expected, a large
increase in E value (65%) was observed, strongly suggesting the
induction of membrane fusion when liposomes that include phos-
phatidylcholine (PC), laurate 1a, 2p, and 3a are treated with a large
excess of CDMT for 3 h (run 1).8 In a control experiment without
3a, no increase in E value (<1%) was observed. Similar results
were obtained with stearate 1b and oleate 1c with longer alkyl
chains (runs 2 and 3). The larger E value with 1c (78%) over that
with 1b (41%) could be due to a higher fluidity of the membrane
composed of 1c, with a cis-olefin, than 1b, with a saturated alkyl
group. No significant further increase in E values with 1a and 1c
was observed by extending the reaction time to 24 h, meaning that
the membrane fusion event with these fatty acids could almost finish
within 3 h.
As shown in Figure 1, we designed the reaction at the membrane
surface based on 1,3,5-triazine dehydrocondensing agents, which
enable us to conduct coupling of carboxylic acids and amines in
micelles as well as in an aqueous solvent by catalysis of a tertiary
amine.5 When SUVs containing fatty acid salts 1, sphingosine 2,
and an amphiphilic tertiary amine catalyst 3 are treated with
2-chloro-4,6-dimethoxy-1,3,5-triazine (CDMT), ceramide 6 can be
efficiently produced via a three-step reaction, because all of the
reactants and intermediates (1-4) are amphiphilic, with the reacting
site at the polar headgroup. Only the polar headgroups of lipids
change significantly; the composition of the hydrophobic moieties
is maintained during the reaction. Thus, inverted cone- or cylindri-
cal-shaped lipids (1 and 2) are converted into cone-shaped lipids
(6). We hypothesized that the reduced surface charge of the
liposomes may cause their aggregation by desolvation or elimination
of counterions at the surface. Furthermore, the presence of cone-
shaped lipids in the outer monolayer of liposomes is believed to
promote membrane fusion, whereas inverted cone-shaped lipids
inhibit it.3 We were therefore curious whether the resulting
liposome, with a ceramide-enriched outer monolayer, would
undergo spontaneous fusion.
† Kobe Gakuin University.
‡ PRESTO, JST.
9
14452
J. AM. CHEM. SOC. 2006, 128, 14452-14453
10.1021/ja0652969 CCC: $33.50 © 2006 American Chemical Society

C O M M U N I C A T I O N S
Table 1. E Values after Definite Times in Probe Mixing Assaya
after 24 h), despite the formation of a large amount of 6a (87%),
indicating, as one would expect, that membrane stability is very
sensitive to the chemical structure of the composite lipids. Several
reports concerning dynamic morphological changes of a molecular
assembly induced by chemical reaction of lipids have appeared in
the literature;9 however, to the best of our knowledge, there has
been no report of successful spontaneous membrane fusion of
liposomes by synthetic chemical transformation (dehydroconden-
sation) of naturally occurring lipids.
E (%)
fatty acid
run
salt
product
1 h
2 h
3 h
24 h
1
2
3
1a
1b
1c
6ap
6bp
6cp
1
0
7
39
13
63
65
41
78
68
66
73
a The reaction was performed with 1, 2p, 3a, and CDMT. 1a, sodium
laurate; 1b, sodium stearate; 1c, sodium oleate.
It is well known that membrane fusion of liposomes is readily
induced by the addition of various fusogenic molecules, like
polyethyleneglycol and alkali earth ions.1c These methods can be
classified as external chemical stimulation, in which the fusogenic
molecules force the membrane to fuse by interaction with a lipid
headgroup or the surface of vesicles. In marked contrast, in our
system the SUV spontaneously undergoes membrane fusion upon
internal chemical stimulation by the product of our artificial enzyme
system.
Ceramide has been reported to serve a variety of functions in
inducing morphological changes of membranes; for example, it acts
as a signal molecule and a constituent of a microdomain (a lipid
raft).10 The present work introduces the possibility that endogenous
mere formation of ceramides from single-chain lipids (sphingosine
and fatty acids) may induce membrane fusion, without requiring
fusogenic proteins or other biological factors. Our system should
be very useful for elucidating the dynamic functions of a variety
of lipids in membranes.
Figure 2. Freeze-fracture transmission electron microscopy (TEM) of
liposome. Scale bar ) 200 nm. 1.3 mM each of PC, 1a, and 2p, and 0.26
mM of 3a. (a) Liposomes are present with vesicle size ranging from 50 to
100 nm before addition of CDMT. (b) Large to giant vesicles with diameters
up to about 1 µm are formed at 3 h after addition of CDMT (6.5 mM). (c)
A control experiment during the same period without CDMT.
When a water-soluble, dehydrocondensing agent 4-(4,6-dimethoxy-
1,3,5-triazin-2-yl)-4-methylmorpholinium chloride (DMT-MM),5a
available in water, was employed instead of a combination of
catalyst 3a and CDMT in the reaction using 1a, the E value was
very low, even after 72 h (22%).
Since the probe mixing assay is known to be sensitive to
aggregation of vesicles, a probe dilution assay that is generally
insensitive to the mere aggregation of vesicles was also employed.6
This assay confirmed that, for experiments using 1a, membrane
fusion involving lipid mixing did indeed occur (see Supporting
Information).
The formation of pseudo-ceramide 6p appears to be essential for
inducing membrane fusion in this system, in that it apparently occurs
prior to the increase in the E value. Thus, the yields of 6ap, which
were determined by ESI-MS, were 48% and 77% at 1 and 2 h,
respectively, in the reaction under the same conditions as in Table
1, run 1. On the other hand, no significant formation of 6ap (2%)
was observed in the above system using DMT-MM after 24 h, and
under these conditions the E value was only 12%.
The change in size of the liposomes during the reaction was
measured by dynamic light scattering. The mean volume diameter
of the liposome, including 1a with CDMT, increased from 46 to
166 nm over the course of 3 h, whereas no change in the diameter
was observed without addition of CDMT. We also observed
liposomes prepared from lipids of a higher concentration (20 times)
by freeze-fracture transmission electron microscopy (TEM). The
liposomes containing 1a were in the range of 50-100 nm before
reaction (Figure 2a). When CDMT was added to start dehydro-
condensation, large to giant vesicles with diameters up to about 1
µm were formed after 3 h (Figure 2b). No change in the size of
liposomes was observed without CDMT during the same period
(Figure 2c). The increase in the diameter (up to 20 times the original
in the largest examples) indicates that hundreds of SUVs fused one
after another.
Acknowledgment. This work was supported partially by a
Grant-in-Aid for Science Research (No. 18659006) from the
Ministry of Education, Culture, Sports, Science and Technology,
Japan.
Supporting Information Available: Synthetic methods, experi-
mental details, and further discussion. This material is available free
of charge via the Internet at http://pubs.acs.org.
References
(1) (a) Devaux, P. F. Biochemistry 1991, 30, 1163-1173. (b) Lipowsky, R.
Biophys. J. 1993, 64, 1133-1138. (c) Blumenthal, R.; Clague, M. J.;
Durell, S. R.; Epand, R. M. Chem. ReV. 2003, 103, 53-69. (d) Cevc, G.;
Richardsen, H. AdV. Drug DeliVery ReV. 1999, 38, 207-232.
(2) McMahon, H. T.; Gallop, J. L. Nature 2005, 438, 590-596.
(3) Chernomordik, L.; Kozlov, M. M.; Zimmerberg, J. J. Membr. Biol. 1995,
146, 1-14.
(4) (a) White, J. M. Science 1992, 258, 917-924. (b) Su¨dhof, T. C. Annu.
ReV. Neurosci. 2004, 27, 509-547. (c) Chen, E. H.; Olson, E. N. Science
2005, 308, 369-373. (d) Bonifacino, J. S.; Glick, B. S. Cell 2004, 116,
153-166.
(5) (a) Kunishima, M.; Kawachi, C.; Hioki, K.; Terao, K.; Tani, S. Tetrahedron
2001, 57, 1551-1558. (b) Kunishima, M.; Yoshimura, K.; Morigaki, H.;
Kawamata, R.; Terao, K.; Tani, S. J. Am. Chem. Soc. 2001, 123, 10760-
10761. (c) Kunishima, M.; Imada, H.; Kikuchi, K.; Hioki, K.; Nishida,
J.; Tani, S. Angew. Chem., Int. Ed. 2005, 44, 7254-7257.
(6) NBD-PE, N-(7-nitro-2,1,3-benzoxadiazol-4-yl)dimyristoylphosphatidyl-
ethanolamine; Rh-PE, N-(lissamine rhodamine B sulfonyl)dimyristoylphos-
phatidylethanolamine. Du¨zgu¨nes, N.; Allen, T. M.; Fedor, J.; Papahad-
jopoulos, D. Biochemistry 1987, 26, 8435-8442.
(7) Struck, D. K.; Hoekstra, D.; Pagano, R. E. Biochemistry 1981, 20, 4093-
4099.
(8) On the basis of the E value, it can be estimated that approximately 70%
fusion occurred after 3 h.
(9) (a) Bachmann, P. A.; Luisi, P. L.; Lang, J. Nature 1992, 357, 57-59. (b)
Menger, F. M.; Gabrielson, K. D. Angew. Chem., Int. Ed. Engl. 1995, 34,
2091-2106. (c) Jaeger, D. A.; Clark, T., Jr. Langmuir 2002, 18, 3495-
3499. (d) Takakura, K.; Toyota, T.; Sugawara, T. J. Am. Chem. Soc. 2003,
125, 8134-8140.
(10) (a) Kro¨nke, M. Chem. Phys. Lipids 1999, 101, 109-121. (b) van
Blitterswijk, W. J.; van der Luit, A. H.; Veldman, R. J.; Verheij, M.;
Borst, J. Biochem. J. 2003, 369, 199-211.
Finally, we have found that the membrane fusion proceeded with
an SUV comprised of naturally occurring lipids, in which sphin-
gosine 2 was used instead of 2p, leading to the formation of
ceramide 6a. Interestingly, under the same conditions as those
shown in Table 1, run 1, the E value increased moderately (36%
JA0652969
9
J. AM. CHEM. SOC. VOL. 128, NO. 45, 2006 14453