The preparation of liposomes bearing human (HLA) transplantation antigens

Model membranes bearing glycolipids and glycoproteins

The forces that hold cell membrane components together are non-covalent and thermodynamically favoured in aqueous media. Hence virtually any glycolipid or membrane glycoprotein might be expected to be incorporable into lipid bilayer membranes and this expectation has been borne out. In addition methods have been developed for linking lipid fragments to species that would not otherwise be expected to associate with bilayers. Techniques that have been successfully used to generate bilayer structures bearing glycolipids and glycoproteins include hydration of films dried down from non-aqueous solutions of the components, detergent removal from aqueous component solutions, exogenous addition to preformed membranes, and various organic solvent injection or reverse phase approaches. Bilayer association of glycolipids and membrane glycoproteins, with preservation of specific receptor function, seem easy to achieve--in fact difficult not to achieve. Optimization of receptor function to accurately mimic that of cell membranes and efficient preservation of functions such as transport or second messenger activation, are typically more demanding, although still feasible. A systematic approach can give considerable insight into the processes involved via identification of minimal necessary factors. Unfortunately, the actual relative arrangement of components, so critical to subtleties of glycolipid and glycoprotein function, remains almost totally unknown for lack of morphological information in the size range of individual macromolecules. The latter problem has come to be the most critical limitation to many studies.

Immunopurification and insertion into liposomes of native and mutant H-2Kb: quantification by solid phase radioimmunoassay

To study the interaction between T cells and isolated H-2Kb, we developed protocols for the immunopurification of the molecule from monoclonal anti-H-2Kb immunoadsorbent columns and for its insertion in lipid vesicles. Patterns of reactivity of two anti-H-2Kb monoclonal antibodies (mAb) (20-8-4 and Y3) on H-2 recombinant and H-2Kb mutant mice indicated that mAb Y3 reacted with all six mutant forms of H-2Kb tested. Binding competition studies indicated that Y3 and 20-8-4 recognized distinct epitopes of H-2Kb. A solid phase radioimmunoassay was established using these two mAb to monitor H-2Kb activity in detergent containing cell lysates, after immunopurification, and after insertion into liposomes. About 70% of H-2Kb activity could be eluted from anti-H-2Kb-immunoadsorbents in the presence of 3 M NH4SCN and octyl glucoside (pH 7.4). A procedure of liposome formation combining gel dilution and dialysis yielded liposomes bearing H-2Kb molecules which could inhibit antibody plus complement cytolysis and could stimulate in vivo primed T cells to generate cytotoxic T lymphocytes in vitro. The present protocol can be extended to immunopurify and obtain H-2Kbm1 bearing liposomes.

Immobilization of protein molecules on liposomes. Anchorage by artificially bound unsaturated hydrocarbon tails

A method for the immobilization of trypsin, a hydrophilic nonmembrane protein, on a liposomal surface has been developed. The technique consists of covalent coupling of linoleoyl residues to the protein globules and consequent binding of linoleoyl trypsin to liposomes by a detergent dilution method. The immobilized protein preserved its biological functions: specific esterolytic catalytic activity and ability to bind to a macromolecular trypsin protein inhibitor. Liposomes carrying immobilized trypsin were able to sequester glucose with the same efficiency as liposomes without trypsin.

Secondary cytolytic T lymphocyte stimulation by purified H-2Kk in liposomes

Purified H-2Kk incorporated into lipid vesicles induced a secondary allogeneic cytolytic T lymphocyte response. However, the level of the response was much less than that generated by using purified plasma membranes containing an equivalent amount of antigen. Similarly, reconstituted membranes stimulated less effectively than did intact plasma membranes. In both cases the stimulating activity of the antigen was increased by including a detergent-insoluble membrane matrix fraction during formation of the liposomes or reconstructed vesicles. Liposomes formed in the presence of the matrix were larger, were more irregular in shape, and had a higher density than those formed in its absence. Both the H-2 antigen and matrix proteins were incorporated into the same vesicles. The greater antigenicity of H-2 in vesicles containing the matrix protein might be due to either the larger size of the liposomes or interaction of the antigen with a component(s) of the matrix.

Lymphocyte recognition of H-2 antigen in liposomes

Liposomes containing purified H-2Kk will specifically stimulate generation of a secondary allogeneic cytolytic T lymphocyte response. Effective recognition was found to depend on the structure of the liposomes. Including detergent-insoluble plasma membrane matrix during formation resulted in liposomes having two- to fourfold more activity than those prepared using just lipid and H-2.

Association of gross virus-associated cell-surface antigen with liposomes

Gross Cell-Surface Antigen (GCSAa) was obtained from W/Fu (C58NT)D lymphoma cells by Nonidet P40(NP40) or 3M KCl extraction and further purified by Sephadex G200 filtration. GCSAa was associated with lipids (dipalmitoylphosphatidycholine, cholesterol and dicetylphosphate, in molar ratios of 7:2:1) to form multilamellar liposomes. The amount of protein associated with liposomes was found to be proportional to the protein concentration of the sensitizing cellular extract and to the amount of phospholipids used and, under defined conditions, 22-55% of the protein of the cellular extract could be associated with liposomes. Analysis of disrupted sensitized liposomes showed that the GCSAa-specific activity of the liposome-associated proteins was quite similar to that of the proteins of the sensitizing cellular extract. Ultracentrifugation of disrupted liposomes showed that about 75% of the liposome-associated GCSAa activity was firmly associated with lipids and that little GCSAa was trapped within aqueous compartments between lipidic lamellae. 1.8--8.0% of the liposome-associated GCSAa was expressed at the liposome surface. No striking differences in degree of GCSAa association were found between liposomes sensitized by NP40 or by 3M KCl extracts. Storage experiments at +4 degrees C showed that GCSAa-sensitized liposomes were fairly stable.

Insertion of Ia and H-2 alloantigens into model membranes

The study of immune phenomena dependent on the major histocompatibility complex (MHC) would be greatly simplified by the use of MHC antigen-containing liposomes in various functional systems. Towards this end, we have constructed unilamellar phosphatidylcholine liposomes containing H-2 and Ia antigens. These molecules were not simply trapped within the aqueous compartment of the liposome as assessed by their accessibility to papain digestion. They were shown to be integrally inserted in the liposome bilayer because they could not be dissociated from the liposome with high salt and EDTA concentrations but could be solubilized by detergent. A sensitive radioimmunoassay showed that the Ia molecules were antigenically active in the liposome environment. Both Ia and H-2 antigens could be immunoprecipitated from detergent-solubilized liposomes. By comparing liposome-associated Ia activity in the presence and absence of detergent and by showing accessibility of the Ia antigens to papain, it was concluded that the majority of Ia antigens were exposed on the external surface of the liposome. These results suggest that the orientation of MHC antigens in liposomes closely parallels their natural orientation in the cell membrane.

Reconstitution of purified detergent-soluble HLA-A and HLA-B antigens into phospholipid vesicles

Purified detergent-soluble human histocmpatibility antigens (HLA-A and HLA-B) were reconstituted into phospholipid vesicles by mixing the protein and lipid together in the presence of either octylglucoside (octyl-beta-D-glucopyranoside) or deoxycholate and removing the detergent by dialysis. The resulting preparation consisted of lipid vesicles containing all or most of the added protein. The protein in the vesicles was antigenically active, as demonstrated by specific binding to anti-beta2-microglobulin IgG-Sepharose beads and by specific inhibition of alloantibody and complement-mediated cytotoxicity. Protein incorporated into vesicles at a protein/phospholipid ratio of 1:10 showed an asymmetric distribution of the HLA-A and HLA-B molecules, with virtually all of the antigens oriented facing the external medium. Cleavage experiments with proteases showed that the molecule was attached to the vesicle membrane via the COOH terminus, consistent with its proposed structure in intact cellular plasma membranes. Electron micrographs of the vesicles showed 50-60 A knobs on the outer surface similar to structures observed for other membrane proteins. HLA-A and HLA-B could also be incoporated into vesicles together with Semliki Forest virus membrane proteins. The resulting preparations should be useful in defining the molecular interactions involving HLA-A and HLA-B antigens in the immune response.