Nontemplated nucleotide addition by HIV-1 reverse transcriptase.

We studied the kinetics of nontemplated nucleotide addition by the reverse transcriptase (RT) of human immunodeficiency virus type 1 (HIV-1) using model substrates derived from the 3' end of HIV-1 minus-strand strong-stop DNA. The addition of a nontemplated nucleotide was highly dependent on the nature of the base (fastest addition with dATP), type of nucleoside, and pH of the reaction buffer. The salt concentration, presence or absence of nucleocapsid protein, and nature of the blunt-ended duplex (DNA/DNA versus RNA/DNA) had only limited effects. The efficiency and base specificity were strongly affected by the sequence at the 3' end of the blunt-ended duplex. In every case, nontemplated nucleotide addition was much slower than templated polymerization. The K(d) for the incoming dNTP with an RT bound to a blunt-ended duplex was at least 1000-fold higher than with a duplex with a template overhang. At concentrations normally found in vivo, ATP can compete with dNTPs for binding to the polymerase active site and reduce the efficiency of nontemplated nucleotide addition. Although a stable ternary complex RT/DNA/dNTP could be readily detected by gel retardation assays if the DNA had a template overhang, stable ternary complexes were not observed with a blunt-ended duplex substrate. At in vivo concentrations of dNTPs (5-10 microM), nontemplated nucleotide addition occurred, but it was very inefficient and the rate of nontemplated polymerization is at least 10000-fold slower than the rate of templated polymerization. We could conclude that, in vivo, the unfavorable binding of the incoming dNTP, low concentration of dNTPs, the presence of a large concentration of ATP, and the inability to form a stable ternary complex prior to the polymerization step collaborate to reduce the efficiency of nontemplated nucleotide addition.