The primary protein product of the human T-cell leukemia virus type 1 (HTLV-1) env gene, gp61, is cleaved to produce both the exterior (gp46) and the transmembrane (gp21) portions of the HTLV-1 envelope protein. To compare the reactivity with human antibodies of different regions of this gp61 protein, five plasmids (A, B, B1, C, and D) were constructed to express recombinant proteins (RPs) in Escherichia coli. RP-A, RP-B, RP-B1, and RP-C contain amino acid residues 26 to 165, 166 to 229, 166 to 201, and 229 to 308, respectively, of the exterior envelope protein gp46. Serum samples from HTLV-1-seropositive subjects were assayed for reactivity with these RPs by Western immunoblotting. The percentages of positive reactivity with each of the RPs were as follows: 18.9% (23 of 122) for RP-A, 89.6% (112 of 125) for RP-B, 70.2% (85 of 121) for RP-B1, and 92.9% (117 of 126) for RP-C. These results indicate that the C-terminal half of gp46 (RP-B plus RP-C) can detect 97.6% (123 of 126) of positive samples, while the N-terminal half of gp46 (RP-A) can only detect 18.9% of the HTLV-1-positive sera (P less than 0.005). Furthermore, RP-A, -B, and -C, which together span the entire length of gp46 except the first five amino acids at the N terminus and the last four amino acids at the C- terminus, detected 99.2% (125 of 126) of the HTLV-1-positive subjects. In contrast, RP-D, which contains the HTLV-1 transmembrane envelope protein gp21 minus the first amino acid at the N terminus, had a lower rate of antibody reactivity at 73.7% (84 of 114) (P less than 0.005). The difference in seropositive rates for RP-D between HTLV-1 carriers (55.6%) and adult T-cell leukemia patients (85.5%) is statistically significant (P less than 0.01). This study therefore indicates that the C-terminal half of gp46, especially the amino acid sequence from 200 to 308, contains the most reactive epitopes of the HTLV-1 gp61 envelope glycoprotein.