Interaction between cartilage proteoglycan and the collagen(s) composed of 1 alpha, 2 alpha, and 3 alpha chains was studied in vitro. Most of the collagen was insoluble under the conditions of assay (0.15 M NaCl, 0.008 M phosphate buffer, pH 7.4; 4 degrees C) and was in the form of fibrils 20 nm in diameter or thinner. The larger fibrils had 60-70 nm periodicity, characteristic of native collagens. Proteoglycan monomers which had been labeled by incubating cartilage slices in vitro with Na2 35SO4 were used to assay the interaction. The insoluble collagen fraction bound proteoglycan from solution. At proteoglycan:collagen ratios lower than 1:2, binding was rapid and linear, and the dissociation constant was 1.7 X 10(-9) M. At higher proteoglycan:collagen ratios, more proteoglycan was bound, but at a slower rate. Binding of proteoglycan to collagen did not require fibrils, since soluble 1 alpha, 2 alpha, and 3 alpha containing collagen also bound to proteoglycan and formed an insoluble complex. Denatured collagens did not bind proteoglycan or compete for binding with normal collagen. Optimum binding occurred with intact proteoglycan, but proteoglycan which had been treated with protease was also bound at low levels. Both protease-treated proteoglycan and free chondroitin sulfate competed with intact proteoglycan in the binding assays, but neither chondroitinase ABC-treated proteoglycan nor the oligosaccharides produced by digestion of chondroitin sulfate with testicular hyaluronidase altered the binding of proteoglycan to collagen. Hyaluronic acid did not compete with radioactive proteoglycan, but heparin and dextran sulfate were extremely effective inhibitors of binding. These data suggest a relatively nonspecific interaction between sulfated polyanions and 1 alpha, 2 alpha, and 3 alpha containing collagens. However, given the location of these collagens near the chondrocyte surface, the interaction of fibrillar 1 alpha, 2 alpha, 3 alpha collagen with proteoglycan is likely to occur and to be of biological importance.