The labeling pattern of mouse embryonic eye frozen sections incubated with radioiodinated brain acidic and basic fibroblasts growth factors (aFGF and bFGF) was investigated by autoradiography. Both growth factors bind to basement membranes in a dose-dependent way, with a higher affinity for bFGF. Similar data were obtained with eye-derived growth factors (EDGF), the retinal forms of FGF. There was a heterogeneity in the affinity of the various basement membranes toward these growth factors. The inner limiting membrane of the retina and the posterior part of the lens capsule have a higher binding capacity than the posterior part of the Bruch's membrane. The specificity of the growth factor-basement membrane interaction was demonstrated by the following experiments: (i) an excess of unlabeled growth factor displaced the labeling; (ii) unrelated proteins with different isoelectric points--gelatin, serum albumin, histones--did not modify the labeling; and (iii) iodinated EGF or PDGF did not label basement membrane. In order to get a better understanding of the nature of this binding, we performed the incubation of the frozen sections with iodinated FGFs preincubated with various compounds: (i) heparin which is known to have a strong affinity for aFGF and bFGF partially decreases the labeling, and (ii) chondroitin sulfate B and dextran sulfate at high concentrations were also partially effective. In addition, enzymatic treatment of the sections reveals that only heparitinase, not collagenase or chondroitinase ABC, completely prevents the labeling without destroying the overall structure of the basement membrane. An antibody against the proteic part of EHS mouse proteoheparan sulfate does not affect the signal. Esterification of the acidic groups cancelled the binding. These results demonstrate that FGFs bind specifically to basement membranes, probably on the polysaccharidic part of the proteoheparan sulfate, and suggest that this type of interaction may be a general feature of the mechanism of action of these growth factors.