Particulate glycoproteins lacking sialic acid, such as desialylated enveloped viruses, readily activate complement through the alternative pathway. Human immunodeficiency virus type 1 (HIV-1) contains two heavily glycosylated and partially sialylated envelope glycoproteins: a surface gp120 and a transmembrane gp41. The abilities of naturally glycosylated HIV-1 and glycosylation-modified HIV-1 to interact with the complement system were examined with a biological assay which measured the binding of whole virus particles to cells expressing CR2 (CD21), the complement receptor found naturally in abundance on follicular dendritic cells and immature B cells. HIV-1 IIIB was synthesized in the presence or absence of the mannosidase II inhibitor, swainsonine, to give rise to high-mannose-type, nonsialylated, nonfucosylated carbohydrate moieties. The virus also was treated with neuraminidase or endo-beta-galactosidase to remove terminal sialic acids. An enzyme immunoassay specific for HIV-1 p24 core protein was used to quantitate the amount of virus bound to cell surfaces. Virus particles incubated with 1:3-diluted, fresh HIV-1-negative human serum as a source of complement readily bound to MT-2 (CD4+ CR2+) and Raji-3 (CD4- CR2+) cells but not to CEM (CD4+ CR2-) cells, suggesting that the virus bound to CR2 independently of CD4. Compared with heat-inactivated or C3-deficient sera, fresh complement increased binding by as much as 62 times for naturally glycosylated virus, and 5 times more than this for glycosylation-modified virus. Similar observations were made with freshly isolated, non-mitogen-stimulated peripheral blood mononuclear cells. Additional evidence that HIV-1 bound to CR2 independently of CD4 was provided by the fact that binding was blocked by monoclonal antibody OKB7 (anti-CR2) but not by OKT4a (anti-CD4). Also, the virus bound to transfected K562 cells (CD4-) which expressed recombinant human CR2 but did not bind to untransfected K562 cells. Results obtained with complement component-deficient sera indicated that binding required the alternative complement pathway. Raji-3 and transfected K562 cells could not be infected with HIV-1 in the presence of complement, suggesting that utilization of CR2 as a receptor in the absence of CD4 does not allow virus entry. The demonstration of CR2 as a receptor for HIV-1 in the presence of complement, together with the ability to enhance binding by desialylation, provides new insights into mechanisms of HIV-1-induced immunity and immunopathogenesis.