Cross-linking of the fingers subdomain of human immunodeficiency virus type 1 reverse transcriptase to template-primer

Abstract

Cross-linking experiments were performed with human immunodeficiency virus type 1 (HIV-1) reverse transcriptase (RT) mutants with unique cysteine residues at several positions (positions 65, 67, 70, and 74) in the fingers subdomain of the p66 subunit. Two approaches were used--photoaffinity cross-linking and disulfide chemical cross-linking (using an oligonucleotide that contained an N(2)-modified dG with a reactive thiol group). In the former case, cross-linking can occur to any nucleotide in either DNA strand, and in the latter case, a specific cross-link is produced between the template and the enzyme. Neither the introduction of the unique cysteine residues into the fingers nor the modification of these residues with photocross-linking reagents caused a significant decrease in the enzymatic activities of RT. We were able to use this model system to investigate interactions between specific points on the fingers domain of RT and double-stranded DNA (dsDNA). Photoaffinity cross-linking of the template to the modified RTs with Cys residues in positions 65, 67, 70, and 74 of the fingers domain of the p66 subunit was relatively efficient. Azide-modified Cys residues produced 10 to 25% cross-linking, whereas diazirine modified residues produced 5 to 8% cross-linking. Disulfide cross-linking yields were up to 90%. All of the modified RTs preferentially photocross-linked to the 5' extended template strand of the dsDNA template-primer substrate. The preferred sites of interactions were on the extended template, 5 to 7 bases beyond the polymerase active site. HIV-1 RT is quite flexible. There are conformational changes associated with substrate binding. Cross-linking was used to detect intramolecular movements associated with binding of the incoming deoxynucleoside triphosphate (dNTP). Binding an incoming dNTP at the polymerase active site decreases the efficiency of cross-linking, but causes only modest changes in the preferred positions of cross-linking. This suggests that the interactions between the fingers of p66 and the extended template involve the "open" configuration of the enzyme with the fingers away from the active site rather than the closed configuration with the fingers in direct contact with the incoming dNTP. This experimental approach can be used to measure distances between any site on the surface of the protein and an interacting molecule.

FIG. 1

Thiol-selective biotinylation of single cysteine residues in mutant RTs before and after modification with photocross-linkers. Biotinylation was visualized by Western blotting. Equal amounts of proteins were loaded on a gel in each lane. 65, K65C; 67, D67C; 70, K70C; 74, L74C; a, APTP; b, BATDHP.

FIG. 2

Photoaffinity cross-linking and chemical cross-linking. (A) Appropriate relationship of the β3-β4 loop that carries the cysteine substitutions at positions 65, 67, 70, and 74 and the template-primer. In the diagram, the end of the extended template is ³²P labeled (∗). (B) In the photocross-linking experiments, templates with different extensions (0 to 15) were each annealed to the same primer. Modified RTs were incubated with template-primer with an 11-nt extension (TP11) and cross-linked (366 nm). The cross-linked samples were fractionated on an SDS-polyacrylamide gel (4 to 20% polyacrylamide). Lanes: 1, K65C-APTP; 2, K65C-BATDHP; 3, D67C-APTP; 4, D67C-BATDHP; 5, L74C-BATDHP; 6, L74C-APTP; 7, K65C-APTP with TP11 without UV treatment; 8, K65C-APTP cross-linked to TP11 and subsequently cleaved with DTT; 9, K65C-BATDHP cross-linked to TP11 and subsequently cleaved with 10 mM NaIO₄. (C) For chemical cross-linking, a single template containing modified G (G-S-S-R) was ³²P labeled (∗). These template-primers were allowed to react with mutant HIV-1 RTs containing a cysteine residue and were fractionated by nonreducing PAGE. Lanes: 1, K65C with template-primer 2 (TP2) and subsequently cleaved with DTT; 2, K65C+TP4; 3, K65C+TP3; 4, K65C+TP2; 5, K65C+TP1.

FIG. 3

Photocross-linking efficiency of template and primer strands. The ratio was calculated as follows: % crosslinking to template/% cross-linking to primer. Tris-borate-EDTA (6%)–urea gels were used to fractionate the reaction products. The average of five independent experiments is plotted, and error is calculated as standard deviation. □, −dNTP; ■, +dNTP.

FIG. 4

Yield of photocross-linking as a function of template extension length. The average of seven independent experiments is plotted, and error is calculated as standard deviation. 65, K65C; 67, D67C; 70, K70C; 74, L74C. ◊, APTP; ●, BATDHP.

FIG. 5

Relative percent change of the yield of photocross-linking in binary and ternary complexes of RT with various template extension lengths. The relative yield was calculated as follows: [(cross-linking in binary complex − cross-linking in ternary complex)/cross-linking in binary complex)] × 100. The average of multiple independent experiments is plotted, and error is calculated as standard deviation. 65, K65C; 67, D67C; 70, K70C; 74, L74C. □, APTP; ■, BATDHP.

FIG. 6

Yield of SH-cross-linking as a function of template extension length in binary and ternary complexes of RT. The average of five independent experiments is plotted, and error is calculated as standard deviation. 65, K65C; 67, D67C; 70, K70C; 74, L74C. □, binary complex; ▴, ternary complex.

FIG. 7

Models for the extended template in the binary and ternary complexes. This figure (A) is based on the structure determined by Ding et al. (9). It shows the fingers of p66 in yellow with the side chains of L74, K65, D67, and K70 marked. The modeling of the position of the extended template is similar to that of Boyer et al. (5). In the binary complex, the fingers are in the open configuration, giving the extended template the opportunity to interact with amino acids in the β3-β4 loop. (B) Closed configuration (ternary complex) with bound dNTP. The figure is based on the structure determined by Huang et al. (25). In the closed configuration, the β3-β4 loop moves closer to the polymerase active site and closes the dNTP binding pocket. In the structure shown by Huang et al., the extended template passes over the fingers near L74. In the figure, the template was extended based on the portion of the template in the crystal structure. In the figure, the template extension is not in a position where it can easily interact with amino acids at position 65, 67, or 70.