Membrane localization of membrane type 5 matrix metalloproteinase by AMPA receptor binding protein and cleavage of cadherins.

Matrix metalloproteinases (MMPs) have been proposed to remodel the extracellular environment of neurons. Here, we report that the metalloproteinase membrane-type 5 MMP (MT5-MMP) binds to AMPA receptor binding protein (ABP) and GRIP (glutamate receptor interaction protein), two related postsynaptic density (PSD) PDZ (postsynaptic density-95/Discs large/zona occludens-1) domain proteins that target AMPA receptors to synapses. The MT5-MMP C terminus binds ABP PDZ5 and the two proteins coimmunoprecipitated and colocalized in heterologous cells and neurons. MT5-MMP localized in filopodia at the tips of growth cones in young [2-5 d in vitro (DIV)] cultured embryonic hippocampal neurons, and at synapses in mature (21 DIV) neurons. Its enrichment in synaptosomes also indicated a synaptic localization in the mature brain. Deletion of the PDZ binding site impaired membrane trafficking of MT5-MMP, whereas exogenous ABP splice forms that are associated either with the plasma membrane or with the cytosol, respectively, colocalized with MT5-MMP in synaptic spines or recruited MT5-MMP to intracellular compartments. We show that endogenous MT5-MMP is found in cultured neurons and brain lysates in a proenzyme form that is activated by furin and degraded by auto-proteolysis. We also identify cadherins as MT5-MMP substrates. These results suggest that ABP directs MT5-MMP proteolytic activity to growth cones and synaptic sites in neurons, where it may regulate axon pathfinding or synapse remodeling through proteolysis of cadherins or other ECM or cell adhesion molecules.

Plasmin activates pro-matrix metalloproteinase-2 with a membrane-type 1 matrix metalloproteinase-dependent mechanism.

Membrane-type 1 matrix metalloproteinase (MT1-MMP) has been implicated as a physiological activator of progelatinase A (MMP-2). We previously reported that plasmin treatment of cells results in proMMP-2 activation and increased type IV collagen degradation. Here, we analyzed the role of MT1-MMP in plasmin activation of MMP-2 using HT-1080 cells transfected with MT1-MMP sense or antisense cDNA. Control, vector-transfected cells that expressed endogenous MT1-MMP, and antisense cDNA transfectants with very low levels of MT1-MMP did not activate proMMP-2. Conversely, cells transfected with sense MT1-MMP cDNA expressed high MT1-MMP levels and processed proMMP-2 to 68/66-kDa intermediate activation products. Control cells and MT1-MMP transfectants had much higher levels of cell-associated MMP-2 than antisense cDNA transfectants. Addition of plasmin(ogen) to control or MT1-MMP-transfected cells generated active, 62-kDa MMP-2, but was ineffective with antisense cDNA transfectants. The effect of plasmin(ogen) was prevented by inhibitors of plasmin, but not by metalloproteinase inhibitors, implicating plasmin as a mechanism for proMMP-2 activation independent of the activity of MT1-MMP or other MMPs. Plasmin-mediated activation of proMMP-2 did not result from processing of proMT1-MMP and did not correlate with alpha(v)beta(3) integrin or TIMP-2 levels. Thus, plasmin can activate proMMP-2 only in the presence of MT1-MMP; however, this process does not require the catalytic activity of MT1-MMP.