The inhibitory properties of TIMP-4 for matrix metalloproteinases (MMPs) were compared to those of TIMP-1 and TIMP-2. Full-length human TIMP-4 was expressed in E. coli, folded from inclusion bodies, and the active component was purified by MMP-1 affinity chromatography. Progress curve analysis of MMP inhibition by TIMP-4 indicated that association rate constants (k(on)) and inhibition constants (K(i)) were similar to those for other TIMPs ( approximately 10(5) M(-)(1) s(-)(1) and 10(-)(9)-10(-)(12) M, respectively). Dissociation rate constants (k(off)) for MMP-1 and MMP-3 determined using alpha(2)-macroglobulin to capture MMP dissociating from MMP-TIMP complexes were in good agreement with values deduced from progress curves ( approximately 10(-)(4) s(-)(1)). K(i) and k(on) for the interactions of TIMP-1, -2, and -4 with MMP-1 and -3 were shown to be pH dependent. TIMP-4 retained higher reactivity with MMPs at more acidic conditions than either TIMP-1 or TIMP-2. Molecular interactions of TIMPs and MMPs investigated by IAsys biosensor analysis highlighted different modes of interaction between proMMP-2-TIMP-2 (or TIMP-4) and active MMP-2-TIMP-2 (or TIMP-4) complexes. The observation that both active MMP-2 and inactive MMP-2 (with the active site blocked either by the propeptide or a hydroxamate inhibitor) have essentially identical affinities for TIMP-2 suggests that there are two TIMP binding sites on the hemopexin domain of MMP-2: one with high affinity that is involved in proMMP-2 or hydroxamate-inhibited MMP-2; and the other with low affinity involved in formation of the complex of active MMP-2 and TIMP-2. Similar models of interaction may apply to TIMP-4. The latter low-affinity site functions in conjunction with the active site of MMP-2 to generate a tight enzyme-inhibitor complex.