The single-stranded nucleocapsid protein that coats the RNA genome of human immunodeficiency virus within the virion core has been produced in Escherichia coli and purified to homogeneity. The mature 55-amino acid protein, normally generated from the gag polyprotein precursor by HIV protease-catalyzed processing of both its amino and carboxyl termini, was produced in E. coli with authentic termini directly, without the need for processing. The protein was purified 30-fold to apparent homogeneity, as determined by both amino acid analysis and SDS-polyacrylamide gel electrophoresis. Sequencing of each terminus of the purified protein indicated that no proteolytic degradation occurred. A molar extinction coefficient (epsilon 280 = 8350 cm-1 M-1) was determined. The purified nucleocapsid protein binds tightly to single-stranded RNA as judged by a nitrocellulose filter binding assay. A binding constant (Kw) of 1 x 10(8) M-1 was calculated. Using fluorescence quenching of nucleocapsid protein upon RNA binding as an assay, a binding site size of seven nucleotides was determined. These results contrast to a larger 15-nucleotide site measured by others for a larger form of nucleocapsid protein-containing sequences from its immature precursor. The possible relevance of these findings are discussed.