Using normal human serum and EDTA-plasma as the two sources of S-protein (vitronectin) in an enzyme-linked immunosorbent assay, we determined that heparin pretreatment of immobilized rgp120 or of immobilized CD4 caused the serum form of S-protein to deposit in a dose-dependent manner. Interestingly, the EDTA-plasma form of S-protein (native form) had little or no interaction with either of the heparin-treated surfaces. Several other sulfated polysaccharides such as dextran sulfate, pentosan polysulfate, heparan sulfate, and fucoidan, likewise mediated the deposition of the serum form S-protein on immobilized rgp120 and CD4. These findings may explain why certain glycosaminoglycans are effective against HIV infectivity in cell culture where the serum form of S-protein is present, yet ineffective in vivo where the native form of S-protein is predominant. The elevated glycosaminoglycan levels in gingival crevicular exudates, coupled with the effects of the serum form of S-protein and salivary-mediated neutralization mechanisms may explain the reduced rates of salivary HIV transmission.