|Home | About | Journals | Submit | Contact Us | Français|
This chapter describes a method to generate plasma membrane sheets that are large enough to visualize the membrane architecture and perform quantitative analyses of protein distributions. This procedure places the sheets on electron microscopy grids, parallel to the imaging plane of the microscope, where they can be characterized by transmission electron microscopy. The basic principle of the technique is that cells are broken open (“ripped”) through mechanical forces applied by the separation of two opposing surfaces sandwiching the cell, with one of the surfaces coated onto an EM grid. The exposed inner membrane surfaces can then be visualized with electron dense stains and specific proteins can be detected with gold conjugated probes.
The plasma membrane is the outer border of a cell and physically separates its interior from the surrounding environment. However, the plasma membrane is not an inert shell. Rather, it is utilized in many cellular processes, and therefore its composition and structure are of great interest to many researchers. One of the early attempts to describe the plasma membrane was the “fluid mosaic” model by Singer and Nicholson (1). In this model, the plasma membrane is a homogeneous two-dimensional lipid bilayer in which proteins can diffuse freely. Over the last decades, this view has been revised, as it became clear that diffusion in the plasma membrane is slower than expected, molecules are not evenly distributed over the cell surface, and molecules are confined to membrane domains and/or surface areas with dimensions of less than one micrometer. These findings led to more complex models, including the “lipid raft” (2), “picket fence” (3), and “protein island” models (4). However, the actual organization of the plasma membrane and its associated molecules remains controversial.
Many biological processes in immune cells utilize and reorganize the plasma membrane. Most immune cells are activated via cell surface receptors, which in some cases is accompanied by the reorganizations of the plasma membrane, most dramatically seen in the formation of the immunological synapse (5). The systematic functions of immune cells often involve the plasma membrane, e.g., endocytosis, phagocytosis, and secretion. Here, we describe a method based on Sanan et al. (6) that has been successfully used in the analyses of the two-dimensional architecture of the plasma membrane in T cells (4, 7), B cells (8), mast cells (9, 10), and other cell types (6, 11, 12). It is based on the generation of plasma membrane sheets attached to EM grids (Fig. 1), by breaking (“ripping”) cells open through forces applied by separating two opposing surfaces sandwiching the cells of interest (Fig. 2). This procedure has been used with a variety of modifications, each optimized for the study of a particular question. In this chapter, the method is divided into four steps: (1) Surface choices and preparations, (2) Cell binding to primary surface, (3) Generation of plasma membrane sheets, and (4) Labeling of plasma membrane sheets. For each step, up to three possible protocols are shown, which can be “mixed and matched,” and further adapted, to suit any specific study aim.
The generation of plasma membrane sheets requires the simultaneous interaction of cells with two opposing surfaces (Fig. 2). Therefore, for each study suitable surfaces have to be found or developed. The “primary” surface is initially used to capture and adhere cells as efficiently as possible. In the case of adherent cells, this can be achieved through simply cultivating the cells on surfaces. Alternatively, immobilized ligands to surface receptors, or affinity reagents, like antibodies or streptavidin, can capture and adhere cells. The latter is generally faster and occurs with higher affinities. These primary surfaces can also be used to manipulate (activate, polarize, etc.) the cells of interest. The “secondary” surface is added so that the forces necessary to “rip” the cells can be applied, and in most cases, binds the cells nonspecifically, quickly, and only for a short period of time. In the methods described here, the secondary surfaces interact with the cells for 10–20 s, but up to 15 min have been reported (12). Depending on which surface is used to coat the EM grid, plasma membrane sheets originating from the adherent (Fig. 2a) or solvent exposed side (Fig. 2b) can be analyzed while attached to a surface. This section describes the preparation of glass surfaces (Subheading 3.1.1) or EM grids (Subheading 3.1.2), which can then be coated with two different methods (PLL, Subheading 3.1.3 and immobilized proteins, Subheading 3.1.4). Alternatively, polyvinylidene fluoride (PVDF) membranes (Subheading 3.1.5) can be used instead of the above.
PLL surfaces bind cells nonspecifically due to the ion bonds between the positively charged lysines and negative charges on the cell surface.
These surfaces bind specific receptors or other molecules on the cell surface and can be used to activate the signaling pathway downstream of a specific receptor. In the case of T cells, specific major histocompatibility complex II molecules together with the co-stimulatory receptor B7.1 have been used successfully (4, 7). In this case, T cell signaling was initiated by the surfaces, and the T cells polarized toward the adhesion site. In different studies, other endogenous ligands or antibodies against surface molecules can be used in similar fashion.
After steps 4, 5, 6, or 7, the surfaces can be stored for up to several days under HEPES buffer, with or without BSA, in a humiditybox at 4°C. However, some recombinant ligands might not be stable under these conditions for an extended period of time.
PVDF membranes have high protein binding capacity and bind well to cells. Due to their flexibility, they are an easy material for the generation of membrane sheets. PVDF membranes do not allow TEM analysis of the plasma membrane attached to them, and are only suitable as secondary surface.
Plasma membrane sheets are bound to EM grids via the extracellular leaflet and the cytosolic leaflet is exposed to solvent. Thus, they can only be sufficiently labeled with reagents that recognize the cytosolic portion of molecules. Proteins or other markers, fused to intracellular tags (HA-tag, Myc-tag, GFP, etc.), can be ectopically expressed in cells and detected in plasma membrane sheets with the appropriate antibodies. In studies that require thedetection of extracellular molecules, gold-conjugated probes have to be used prior to the ripping procedure. Live cells can be labeled when bound to the primary surface if a study focuses on the solvent exposed cell side, or alternatively in suspension for either side. However, in this latter case, the label may inhibit binding to the primary surface. Multivalent probes can multimerize their targets and/or induce endocytosis in live cells, which will affect the results of the TEM analyses. In studies, where cells are activated by receptor cross-linking this can be a desired effect. Multimerization and activation can be prevented by fixation of the cells prior to labeling. However, the degree of fixation has to be optimized to avoid multimerization, but still allow binding to the surfaces and ripping of the cell. Here, three possible conditions that have been successfully used to bind cells to primary surfaces are described.
This procedure is similar to the binding of cells to PLL surfaces (Subheading 3.2.1). However, some adjustments have to be considered.
Depending on what surface is used to coat the EM grid, the plasma membrane attached to either the primary surface (adherent cell side; Fig. 2a) or secondary surface (solvent exposed cell side; Fig. 2b) can be analyzed. In general, the primary surfaces yield larger, more intact and higher numbers of plasma membrane sheets. Therefore, if the study permits it, the EM grid is ideally coated with the primary surface. The ripping procedure can take place on ice or at RT. If cold conditions are required, place all beakers with buffers in and the ripping surface on compressed ice in a large insulated tray. Under any conditions, excessive amounts of drying that would destroy the membrane sheets or condensation that affects buffer concentrations should be avoided (see Note 4). Here, two procedures are described that were successfully used to generate plasma membrane sheets from the adherent (Subheading 3.3.1) and the solvent exposed side (Subheading 3.3.2) of T cells.
Multiple membrane-associated molecules of interest can be labeled with different sized gold-conjugated detection reagents (Fig. 1b). However, more than two different sizes can make the identification of the gold species difficult during analyses. Probes specific to certain molecules can be directly conjugated to colloidal gold, thus no secondary label is required.
The specimens are imaged using a transmission electron microscope (TEM), and the gold distribution in the resulting images can be analyzed for clustering, sizes of clusters, co-localization, and other parameters by a multitude of statistical methods (e.g., (13, 14)).
The authors thank Dr. Bridget S. Wilson for advice on TEM and plasma membrane sheet preparation, and Dr. Fleur E. Tynan for comments on the manuscript.
1EM grids are best handled with nonmagnetic, fine tipped and curved type 7 forceps, and coverslips are most conveniently manipulated with reverse action type N5 forceps.
2The choice of EM grids is crucial for the preparation of plasma membrane sheets. The EM grid should distribute the force during the attachment of the secondary surface equally onto the carbon-coated formvar sheet. Hexagonal nickel EM grids have wide metal bands forming the mesh, which is ideal for this procedure. EM grids made from wire cut the formvar/carbon sheet and are not suited for this procedure.
3Due to surface tension of liquid trapped between the two tines, EM grids easily become attached to the surface of forceps during their release. Therefore, when removing liquid by touching the edge of the EM grid with an ashless filter paper, touch the gap between the tines simultaneously, which will remove any liquid between them. When floating EM grids on liquid, make sure that the EM grid is in contact with liquid surface during the opening of the tines, which will ensure that it floats onto the liquid.
4The ripping conditions can be optimized by TEM analyses of plasma membrane sheets labeled with the electron dense stains only or by the analyses of cells labeled with a fluorescent membrane marker (e.g., DiO, DiI, or DiD) on an inverted fluorescence microscope. For the latter, submerge the EM grid with the plasma membrane sheet side down in a microscopy chamber.
5The efficiency of the ripping procedure can be increased for some cells, by incubating them in hypotonic buffer (25–75 mM salt) for a short period of time prior and during the ripping procedure. However, this potentially induces changes in the cell morphology and may activate stress-related responses.
6If PVDF membranes are used as secondary surfaces, it is easier to pick the EM grids with forceps, when the membrane is bend between two fingers with the EM grid on the outside of the arch. If glass coverslips are used, place them on top of the EM grids with the edge of the EM grid slightly extended past the surface of the coverslip. This makes the removal of the EM grids from the coverslips easier.
7Fixation and binding condition for the labeling can be optimized by preliminary experiments using fluorescence-activated cell sorting (FACS). This enables many conditions to be examined quickly.