Granulation tissue was produced in rats by subcutaneous implantation of Visella sponges. D-penicillamine (D-pen) 100 or 500 mg/kg was administered daily for 42 days by gastric tubing. Pairfed, placebo treated animals were included as controls. Half of the groups were kept for additionally 28 days without medication. The inhibitory effect of D-pen on cross-link formation in newly synthesized collagen was readily reversible. By contrast, cross-link deficiency lasting beyond the observation period was observed in the higher polymeric collagen variants released by dilute acid, heat exposure or limited pepsin proteolysis as estimated by solubility, alpha/beta chain ratio and/or aldehyde content. By SDS-polyacrylamide gel electrophoresis on gels containing 3.6 M urea it was shown that purified dermal acid soluble collagen from treated animals consisted of a mixture of type I and III collagen, whereas only type I collagen was detected in controls. The band pattern was identical in reduced and unreduced collagen samples. Four weeks after D-pen discontinuance type III collagen had disappeared from the acid extract. Moreover, the ratio of type III to type I collagen in the pepsin digest from both granulation tissue and skin showed a persistent rise with D-pen. These observations indicate that D-pen destabilized type III collagen in particular by interference with its disulfide linkages. The amount of granulation tissue remained unaffected throughout the experiment, whereas the skin collagen content decreased at the higher dose level. The regeneration was not completed by the end of the observation period. Modulation of the molecular stability of granuloma collagens may be of relevance for the antirheumatoid effect of D-pen, but the sustained effect on normal tissues may imply a long standing impairment of their supportive capacity.