Efficient gap repair catalyzed in vitro by an intrinsic DNA polymerase activity of human immunodeficiency virus type 1 integrase

Abstract

Cleavage and DNA joining reactions, carried out by human immunodeficiency virus type 1 (HIV-1) integrase, are necessary to effect the covalent insertion of HIV-1 DNA into the host genome. For the integration of HIV-1 DNA into the cellular genome to be completed, short gaps flanking the integrated proviral DNA must be repaired. It has been widely assumed that host cell DNA repair enzymes are involved. Here we report that HIV-1 integrase multimers possess an intrinsic DNA-dependent DNA polymerase activity. The activity was characterized by its dependence on Mg2+, resistance to N-ethylmaleimide, and inhibition by 3'-azido-2',3'-dideoxythymidine-5'-triphosphate, coumermycin A1, and pyridoxal 5'-phosphate. The enzyme efficiently utilized poly(dA)-oligo(dT) or self-annealing oligonucleotides as a template primer but displayed relatively low activity with gapped calf thymus DNA and no activity with poly(dA) or poly(rA)-oligo(dT). A monoclonal antibody binding specifically to an epitope comprised of amino acids 264 to 273 near the C terminus of HIV-1 integrase severely inhibited the DNA polymerase activity. A deletion of 50 amino acids at the C terminus of integrase drastically altered the gel filtration properties of the DNA polymerase, although the level of activity was unaffected by this mutation. The DNA polymerase efficiently extended a hairpin DNA primer up to 19 nucleotides on a T20 DNA template, although addition of the last nucleotide occurred infrequently or not at all. The ability of integrase to repair gaps in DNA was also investigated. We designed a series of gapped molecules containing a single-stranded region flanked by a duplex U5 viral arm on one side and by a duplex nonviral arm on the other side. Molecules varied structurally depending on the size of the gap (one, two, five, or seven nucleotides), their content of T's or C's in the single-stranded region, whether the CA dinucleotide in the viral arm had been replaced with a nonviral sequence, or whether they contained 5' AC dinucleotides as unpaired tails. The results indicated that the integrase DNA polymerase is specifically designed to repair gaps efficiently and completely, regardless of gap size, base composition, or structural features such as the internal CA dinucleotide or unpaired 5'-terminal AC dinucleotides. When the U5 arm of the gapped DNA substrate was removed, leaving a nongapped DNA template-primer, the integrase DNA polymerase failed to repair the last nucleotide in the DNA template effectively. A post-gap repair reaction did depend on the CA dinucleotide. This secondary reaction was highly regulated. Only two nucleotides beyond the gap were synthesized, and these were complementary to and dependent for their synthesis on the CA dinucleotide. We were also able to identify a specific requirement for the C terminus of integrase in the post-gap repair reaction. The results are consistent with a direct role for a heretofore unsuspected DNA polymerase function of HIV-1 integrase in the repair of short gaps flanking proviral DNA integration intermediates that arise during virus infection.

FIG. 1

Analysis of DNA polymerase activity by molecular exclusion chromatography. Integrase was purified by elution from a nickel chelate resin and applied to a Sephacryl S-300 column as described in Materials and Methods. The top panel shows the DNA polymerase activity profile obtained with DISPOL 17 DNA and 1 μl of each column fraction in reaction mixtures. The elution positions of standard proteins (ADH, alcohol dehydrogenase; BSA, bovine serum albumin) used to calibrate the column are indicated by the arrows. The bottom panel shows an SDS-PAGE analysis of proteins in individual fractions eluted from the S-300 column. Prior to analysis, samples were concentrated 10-fold by centrifugation in a Microcon 30 unit. Proteins were detected by silver staining. The position of integrase is indicated at the left.

FIG. 2

SDS-PAGE and Western blot analysis of wild-type and deletion mutant Δ713 integrase. The panel on the left shows a silver-stained SDS-PAGE profile of wild-type (WT) and mutant (Δ713) integrase purified on a Ni²⁺-NTA resin; the panel on the right shows an immunoblot analysis of Δ713 and WT integrase. Nitrocellulose filters were incubated with antibody directed at the amino terminus of integrase (N-Term) or the carboxy terminus of integrase (C-Term) as described in Materials and Methods. The lane marked M contained Rainbow colored protein standards.

FIG. 3

Sephacryl S-300 DNA polymerase activity elution profile of wild-type (•) and deletion mutant Δ713 (◊) integrase. The mutant elution profile was superimposed on the wild-type profile given in Fig. 1. Fractions of 1 ml were collected, and aliquots of 1 μl were assayed for DNA polymerase activity as described in Materials and Methods.

FIG. 4

Effect of AZT-TP on DNA polymerase activity. DNA polymerase reactions were conducted in the presence of various concentrations of AZT-TP, using either DISPOL 17 DNA (▪) or poly(dA)-oligo(dT) (⧫) as a template primer. The DNA polymerase assays were done in triplicate.

FIG. 5

Neutralization of DNA polymerase activity by an anti-integrase MAb. Dilutions of a mouse ascites fluid containing MAb 35 were prepared in phosphate-buffered saline, and 1 μl of the diluted antibody was added to each 25-μl reaction mixture. Half-filled triangles represent reactions with integrase; open triangles represent reactions with E. coli DNA polymerase I (Klenow fragment).

FIG. 6

Schematic representation of the gapped DNA molecules used in this study. Torsional strain in the region of the hairpin termini probably prevents base pairing at the ends of the molecule. The molecules feature a U5 hairpin contiguous with the CA dinucleotide to the left of the gap (except for G₁H, where five nucleotides adjacent to the gapped region were changed to a nonviral sequence), a 5′ unpaired AC tail (except for G₂), and a gapped region which varies in length and base composition adjacent to the CA dinucleotide. The right hairpin in each molecule is composed of a nonviral sequence which contains an NdeII cleavage site 10 nucleotides from the 3′ end of the DNA.

FIG. 7

Repair products obtained in an integrase DNA polymerase reaction using gapped and nongapped DNAs as template primers. The diagrams on the left and right illustrate polymerase reactions with gapped (G₁) and nongapped (T₅H) DNA substrates, respectively. Following digestion with NdeII, samples were analyzed on a denaturing DNA sequencing gel as described in Materials and Methods. The radiolabeled DNA fragments seen in the autoradiogram range from 11 to 15 nucleotides in length as determined by using 5′-end labeled oligonucleotides of known length as markers (oligonucleotides of 8, 12, 14, and 17 nucleotides were used; see Fig. 11 for an in-gel comparison). A_1-5 refers to the number of dAMP residues polymerized at the 3′ end of the respective DNA template-primers. The inclined planes lying over the photo of the autoradiogram represent increases in the amount of DNA added to reaction mixtures ranging from 10 to 20 ng (lanes 1 and 2) or 10, 20, and 50 ng (lanes 3 to 5).

FIG. 8

Quantitative analysis of repair reactions conducted with gapped (G₁; top) and nongapped (T₅H; bottom) DNA. The radiolabeled DNA fragments seen by autoradiography were imaged by using a detection screen. The data were scanned into a Bio-Rad phosphorimaging unit and quantified by computer analysis of the scanned images. The numbers 1 to 5 on the abscissa of each graph refer to the number of dAMP residues added to the 3′ end of the DNA template-primer. The data were corrected for cumulative increases in band intensity caused by differences in fragment size. □, 5 ng; ░⃞, 10 ng; , 20 ng.

FIG. 9

Influence of base composition on gap repair. Integrase DNA polymerase reactions were conducted in the presence of dGTP with either G₁ (lane 1) or G₃C₅ (lane 2) DNA. The repair products released from the G₃C₅ DNA substrate by NdeII digestion are indicated by G_1-5 at the lower right. The zone marked by the asterisk marks the position of the products of incomplete digestion with NdeII. A 36-mer marker oligonucleotide, of the structure expected if G₃ molecules participating in repair underwent 5′ joining followed by cleavage at the CA dinucleotide adjacent to the C₅ gap, comigrated with the band marked 5′-joined and processed. The structure of this repair product was not investigated further. The arrow marks the position of repaired DNA that is not treated with NdeII.

FIG. 10

Utilization of ddATP by the integrase DNA polymerase. The DNA repair products obtained with a G₁ gapped DNA substrate were compared for reactions conducted either with dATP (lane 1) or ddATP (lane 2) as the nucleotide precursor. The repair products released from the G₁ DNA substrate by NdeII digestion are indicated by A_1-5 and ddA₁ for the dATP and ddATP reactions, respectively. The radiolabeled DNA repair products ranged in size from 11 to 15 nucleotides.

FIG. 11

Effect of a 5′ AC unpaired dinucleotide on the DNA repair reaction. Repair reactions were conducted as described in the text, using either tailed (G₁; lane 2) or untailed (G₂; lane 3) gapped DNA as a template-primer. The lanes marked M₄ (lanes 1 and 4) contain 5′-³²P-labeled oligonucleotides whose lengths (between 8 and 17 nucleotides) are indicated at the sides. The significance of the asterisk and the arrow is explained in the legend to Fig. 9.

FIG. 12

Effect of gap size on DNA repair by the integrase DNA polymerase. The major repair products range in size from 11 to 13, 11 to 14, 11 to 17, and 11 to 19 nucleotides in the cases of G₁T₁ (lane 2), G₁T₂ (lane 3), G₁ (lane 1), and G₁T₇ (lane 4), respectively. Fragments of DNA corresponding to the complete fill-in of each gap are 11, 12, 15, and 17 in the cases of G₁T₁ (lane 2), G₁T₂ (lane 3), G₁ (lane 1), and G₁T₇ (lane 4), respectively. The size of each DNA fragment was determined in comparison with oligonucleotide size markers (M₄ in Fig. 11). Except for lane 3, the presumed nucleotide composition of the added nucleotides is shown opposite the repair products. In lane 3, the four bands at the bottom represent the addition of A₁, A₂, TA₂, or GTA₂. The sequences (read from right to left) represent the addition of nucleotides to the gapped DNA substrates in a 5′-3′ direction. The position of a 48-mer oligonucleotide corresponding to the structure expected for a 5′-joined product after NdeII cleavage is shown near the top. The significance of the asterisk and the identity of the bands near the top of the photograph are as described above.