Formation of Cell Membrane Component Domains in Artificial Lipid Bilayer

Ryugo Tero,Miyu Yoshida,Toshinori Motegi,Kohei Fukumoto,Ayumi Hirano-Iwata,Michio Niwano

doi:10.1038/s41598-017-18242-9

Abstract

The lipid bilayer environment around membrane proteins strongly affects their structure and functions. Here, we aimed to study the fusion of proteoliposomes (PLs) derived from cultured cells with an artificial lipid bilayer membrane and the distribution of the PL components after the fusion. PLs, which were extracted as a crude membrane fraction from Chinese hamster ovary (CHO) cells, formed isolated domains in a supported lipid bilayer (SLB), comprising phosphatidylcholine (PC), phosphatidylethanolamine (PE), and cholesterol (Chol), after the fusion. Observation with a fluorescence microscope and an atomic force microscope showed that the membrane fusion occurred selectively at microdomains in the PC + PE + Chol-SLB, and that almost all the components of the PL were retained in the domain. PLs derived from human embryonic kidney 293 (HEK) cells also formed isolated domains in the PC + PE + Chol-SLB, but their fusion kinetics was different from that of the CHO-PLs. We attempted to explain the mechanism of the PL-SLB fusion and the difference between CHO- and HEK-PLs, based on a kinetic model. The domains that contained the whole cell membrane components provided environments similar to that of natural cell membranes, and were thus effective for studying membrane proteins using artificial lipid bilayer membranes.

Highlights

108 domains with random area values with normal distribution were settled in a field of 600 px × 450 px
The average size of the domains used in this study was 10 px[2] corresponding to the area fraction of 4%
(1) Fusion of PL was detected for each pixel of each domain with a probability of Pf. (2) After the detection for all the domains, the area of each domain was increased by the area of PL (APL = 50 px2) by the frequency of fusion in (1). (3) The component of the domains (LD) was increased in every domain to which PL fused at least once, in proportion to the relative equilibrium constant K, the density of PL in the domain and the domain perimeter

Summary

Introduction

108 domains with random area values with normal distribution were settled in a field of 600 px × 450 px. The average size of the domains used in this study was 10 px[2] corresponding to the area fraction of 4%. The area of the domains were varied through the following steps. (1) Fusion of PL was detected for each pixel of each domain with a probability of Pf. (2) After the detection for all the domains, the area of each domain was increased by the area of PL (APL = 50 px2) by the frequency of fusion in (1). (4) The same amount as the sum of LD in (3) was removed from the domains to which PL had never fused, in proportion to their area. The domain was discarded if its area became less than 1 px[2]. The steps (1–4) were repeated, and the domains were redrawn with the new area after each cycle. Calculation and visualization were performed with Igor Pro software

Objectives

Methods

Results

Conclusion