Results and conclusions concerning the ability of HIV glycoprotein (gp) 120 to stimulate monokine secretion have been equivocal, based on observations using natural gp120 derived from infected human cells and a Chinese hamster ovary (CHO) cell-derived recombinant fusion protein. Current studies were designed to determine whether differences in recombinant gp120 proteins could result in failure to trigger monokine production. We found that natural gp120 could stimulate monocytes to release TNF-alpha, IL-1 beta, IL-6, and granulocyte-macrophage-CSF, and this effect could be blocked with soluble CD4. Full-length rgp120 either expressed from an adenovirus vector and purified from infected human cells, or derived from CHO cells, could function similarly. In contrast, full-length recombinant envelope protein expressed in a baculovirus system and a CHO cell-derived recombinant fusion protein tested previously, consistently failed to stimulate monokine production. The stimulatory capacity of both natural and full-length CHO cell-derived gp120 was eliminated by heating at 100 degrees C, and could be blocked with excess CHO cell-derived gp120 fusion protein. Inasmuch as the baculovirus-expressed gp120 and the CHO cell-derived recombinant fusion protein can bind to CD4, these results suggest that HIV gp120 binding to CD4 on the monocyte surface may of itself be insufficient for stimulation of monokine secretion. Therefore, primary protein structure, as well as posttranslational protein modifications, may determine this activity.