MHC-I

Dynamics of MHC-I Patches on Cell Plasma Membranes
In collaboration with Prof. Michael Edidin, Johns Hopkins University Hwang J, Gheber LA, Margolis L, Edidin M. 1998. Domains in cell plasma membranes investigated by near-field scanning optical microscopy. Biophysical Journal 74(5):2184-2190. Gheber LA, Edidin M. 1999. A model for membrane patchiness: Lateral diffusion in the presence of barriers and vesicle traffic. Biophysical Journal 77(6):3163-3175. Lavi Y, Edidin M, Gheber LA 2007. Dynamic Patches of Membrane Proteins. Biophysical Journal: Biophysical Letters 93: L35.
Our view of cell membranes as two-dimensional (2D) fluids, all the constituents of which move freely, has been changed in recent years by demonstrations of lateral heterogeneities, patches, and domains in cell surface membranes. Membrane domains have been defined mainly by measurements of the lateral diffusion of membrane proteins and lipids, although other methods have also been used. ( The measurements suggest that the size of many membrane domains is in the range of hundreds of nanometers, at or below the resolution of far-field microscopy. Near-field scanning optical microscopy (NSOM) uses the contrast techniques of far-field microscopy but achieves resolution of surface structures well beyond the diffraction limit of far-field microscopy. We therefore imaged human fibroblasts labeled with with fluorescent antibodies to membrane proteins, HLA class I (HLA-I) molecules.			Near-field images of a fixed human skin fibroblast acquired in liquid, labeled with Cy3-Fab KE-2 against HLA-I molecules. The 3D contours are from shear-force data and the color scale represents the fluorescence intensity. Topography height varies over a range of 165 nm (A) and 215 nm (B). Autocorrelation profiles and Gaussian fits for the images in A and B yield characteristic lengths of 315 nm (D), 580 nm (E).
The reason for the existence of patches has been inferred to be a network of actin filaments, underlying the plasma membrane, with which the cytosolic tail of the proteins interacts. However, the persistence of these patches at steady state is not consistent with a membrane in which proteins and lipids are free to diffuse, and in which barriers to this lateral diffusion open every few seconds. If patches persist, they must either be stabilized by specific molecular interactions or they must reflect some other aspect of membrane physiology at steady state. We have made a quantitative model of the cell surface that includes random walks, dynamic barriers to lateral diffusion, and vesicle traffic. Analyzing this model, we find that vesicle traffic, together with dynamic barriers to lateral diffusion, can create and maintain patches on a scale of hundreds of nanometers, apparent membrane domains. The barrier spacing scales the size of the patches. Vesicle traffic, the delivery and removal of membrane components, determines the persistence of the population of patches and their concentration relative to the average for the entire membrane.
Simulated free diffusion Simulated obstructed diffusion (barriers)		Simulated membrane, with barriers but no vesicle traffic: no patches	Simulated membrane, with barriers AND vesicle traffic: patches are formed and maintained	Simulated membrane, with barriers and stopped vesicle traffic: patches disappear
Scaling to real world parameters: a randomly organized membrane is expected to create patches of the size dictated by the barriers within 2 minutes.			The model predicts the following: 1. Patches should be dynamic in nature. They should be born, decay and disappear, then re-appear with the next vesicle delivered. 2. Patches should disappear if vesicle traffic is stopped. 3. Patches should disappear if the actin mesh is depolymerized. In order to check these predictions, we are using cells transfected with GFP-tagged MHC-I, and image them live with Total Internal Reflection Fluorescence Microscopy (TIRFM).
Epifluorescence	TIRFM (just the membrane)	Click image to see movie