STRUCTURE OF ALPHA-2-MACROGLOBULIN PROTEASE COMPLEXES - METHYLAMINE COMPETITION SHOWS THAT PROTEASES BRIDGE 2 DISULFIDE-BONDED HALF-MOLECULES

Alpha-2-macroglobulin (alpha-2M) forms several different covalent complexes with proteases. These include unusual forms in which more than one of the four identical subunits of alpha-2M are cross-linked by amide bonds to more than one lysyl amino group of the bound protease. The structure of these complexes and the question of how the identical subunits are arranged to form two protease binding sites are matters of current controversy. The 185-kDa subunits are arranged into two disulfide-bonded half-molecules which are, in turn, noncovalently associated. We have provided evidence that, in the major multivalent cross-linked form, proteases can span the two half-molecules, forming a covalently bonded tetramer [Wang, D., Yuan, A.I., & Feinman, R.D. (1984) Biochemistry 23, 2807-2811]. An alternative theory has recently been proposed in which the major high molecular weight form has two bonds to protease that are within half-molecules-a multivalent cross-linked dimer [Sottrup-Jensen, L., Hansen, H. F., Pedersen, H. S., & Kristensen, L. (1990) J. Biol. Chem. 265,17727-17737]. To resolve this conflict, experiments were carried out to determine the structure of one of the high molecular weight bands (band 3) seen on SDS-PAGE. Band 3 has anomalous migration, corresponding to markers of apparent molecular mass of 550 kDa (between the tetramer and dimer). In the experiments described here, reactions of thrombin with alpha-2M were run in the presence of methylamine, which competes for one of the two thrombin-alpha-2M covalent bonds. The logic of the experiment was that the location of labeled methylamine would indicate the original covalent bridging site. The results showed that methylamine was only found in half-molecules that did not contain a bound thrombin molecule. The results can only be explained by a mechanism whereby (in the absence of amines) proteases must bridge two half-molecules. The results also support, but do not prove, the theory that two subunits from different half-molecules constitute the minimum protease binding site.