Phase behavior in a ternary lipid membrane estimated using a nonlinear response surface method and Kohonen’s self-organizing map
Graphical abstract
Our technique enables to distinguish ternary lipid mixture into several clusters with similar membrane properties. The results should contribute to elucidating lipid domain formation.
Introduction
A lipid raft is defined as a membrane microdomain rich in cholesterol (Ch) and sphingolipid located in the outer leaflet of the plasma membrane. Because numerous signaling molecules such as glycosylphosphatidylinositol (GPI)-anchored proteins [1], [2] and receptor- [3], [4] or nonreceptor-type tyrosine kinases [4], [5], [6] originate from these domains, the lipid raft is thought to act as a platform for protein segregation and signal transduction in the plasma membrane. Lipid-driven lateral separation of immiscible liquid phases is likely to be a crucial factor in the formation of lipid rafts in cell membranes [7]. Although the concept of a lipid raft has recently been widely accepted, its mode of formation in the cell membrane remains controversial.
Model membranes such as liposomes are effective tools for elucidating the formation of lipid rafts. Their lipid composition can be manipulated to suit the purpose of the experiment, and results obtained in this way are likely to be consistent. In addition, several membrane domains are known to coexist in model membranes as well as in cell membranes [8], [9], [10], [11], [12].
In general, lipid bilayers are classified into three different phases in order of increasing fluidity: a solid-ordered phase (Lβ), a liquid-ordered phase (lo), and a liquid-disordered phase (Lα) [7], [13]. The Lβ and Lα phases are also called gel and liquid crystalline phases, respectively. Ordered and tight packing are typical of the Lβ phase membrane, whereas fast axial rotation and high lateral mobility are observed in the Lα phase membrane. The Ch-rich membrane exists as an lo phase membrane; this phase is intermediate between the Lα and Lβ phases. Its ordered packing is similar to that of the Lβ phase, but its fast axial rotation and high lateral mobility are similar to that of the Lα state. The lipid raft is assumed to exist as an lo phase membrane in the plasma membrane.
A mixture of three different lipids, lipids with a high phase transition temperature (Tm) (e.g., lipid with saturated acyl chains), lipids with a low Tm (e.g., lipid with unsaturated acyl chains), and Ch, is required to generate membrane domains. Such membranes have been widely used to mimic plasma membranes to elucidate the formation or structures of lipid rafts [9], [11], [12], [14], [15]. Even though these membranes are much simpler than biological membranes, their phase behavior remains complicated. Despite numerous research studies, consensus on membrane phase behavior has yet to be reached. Veatch et al. generated a ternary phase diagram by observing the surface of giant unilamellar vesicles using confocal fluorescence microscopy (CFM) and described a condition in which phase separation occurs [11], [12]. Although microscopic observation can be used to directly detect phase separation on membrane surfaces, this method has limitations. First, because giant unilamellar vesicles are not obtained from all lipid compositions, whole-phase behavior can never be determined with this method. Second, because the border region between membrane domains is thought to be an ambiguous structure, it is difficult to distinguish membrane domains by subjective evaluation. Third, because the phase behavior of ternary lipids is very complicated and is substantially changed by slight differences in lipid composition, a bunch of model membranes differing in lipid composition should be examined to elucidate phase behavior.
Fluorescence analyses such as Förster resonance energy transfer (FRET) and fluorescence anisotropy are also employed to identify the lipid phase behavior of membranes [14], [16], [17], [18], [19]. Because these measurements do not require preparation of a giant unilamellar vesicle, a wider range of lipid compositions can be examined than with CFM. In addition, measurements based on fluorescence analysis enable objective and quantitative evaluation of domain formation, and the experimental procedure is not very complicated. However, large data sets are required to fully understand the relationship between lipid composition and phase behavior and, in most cases, the collection of such large data sets is impractical.
To overcome this problem and elucidate the phase behavior of ternary lipids, we applied a response surface method incorporating multivariate spline interpolation (RSM-S) and a data mining technique (Kohonen’s self-organizing maps; SOMs). Firstly, liposomes composed of sphingomyelin (SM), dioleoyl phosphatidylcholine (DOPC), and Ch were prepared and their fluorescence anisotropy was measured. Using data based on fluorescence anisotropy, we developed a scatter plot indicating the distribution of membranes with similar membrane properties. Response surface methods and a data mining technique were used to compensate for the lack of experimental data. Using these methods, we successfully resolved the membranes into several clusters. This is the first technical report on the phase behavior of lipids conducted using response surface and data mining methods.
Section snippets
Materials
Ch was purchased from Wako (Osaka, Japan). Chicken egg SM (Coatsome NM-10) and DOPC (Coatsome MC-8181) were purchased from Nippon Oil & Fat (Tokyo, Japan). More than 70% of the SM is 16:0 SM. 1,6-diphenyl-1,3,5-hexatriene (DPH) was purchased from Aldrich (Milwaukee, WI, USA). All other chemicals were of analytical grade and are commercially available.
Preparation of liposomes
Liposomes composed of SM, DOPC, and Ch were prepared as reported previously [19]. In brief, designated amounts of lipids dissolved in chloroform
Estimation of the distribution of membranes with similar membrane properties
Fluorescence anisotropy curves as functions of temperature obtained from 65 model membranes differed markedly in shape and value, indicating the diversity of the model membranes (data not shown). The experimental values were processed using RSM-S, and then response surfaces indicating the relationships between lipid composition and fluorescence anisotropy were generated. On behalf of all response surfaces generated using RSM-S, those at 25, 40, and 60 °C are shown in Fig. 3a–c. We also generated
Discussion
To clarify the phase behavior of ternary lipid membranes, SM, DOPC, and Ch were selected as components of the model membrane. SM, a high Tm lipid, is a component of ordered phase membranes, whereas DOPC, a low Tm lipid, is a component of disordered phase membranes, respectively. SM also forms a lo phase membrane with Ch. Fluorescence anisotropy was used to investigate the phase behavior of the ternary lipid mixture. The Lα, Lβ, and lo phases of lipid membranes can be distinguished according to
Conclusions
We successfully classified SM/DOPC/Ch membranes into six clusters with similar membrane properties. Our study, which encompassed a wide range of lipid compositions, contributes to the elucidation of the behavior of ternary lipid mixtures. Although an immense amount of experimental data would have been required to interpret the data using conventional means, we overcame this problem by using RSM-S and Kohonen’s self-organizing map. Furthermore, because the clusters were determined based on
Acknowledgments
The authors are grateful to Yamatake Corporation for providing us with dataNESIA version 3.0. We are also grateful to Ms. Eri Imajo at Hoshi University for her kind assistance with the experimental work.
References (28)
- et al.
Cell
(1992) - et al.
Immunol. Lett.
(2002) - et al.
J. Biol. Chem.
(2001) Biochim. Biophys. Acta
(2005)- et al.
Biophys. J.
(2001) - et al.
J. Biol. Chem.
(2003) - et al.
Biochim. Biophys. Acta
(1968) - et al.
Biophys. J.
(2003) - et al.
Biochim. Biophys. Acta
(2005) - et al.
Biochim. Biophys. Acta
(2007)