Mapping viral DNA specificity to the central region of integrase by using functional human immunodeficiency virus type 1/visna virus chimeric proteins

Abstract

We previously described the construction and analysis of the first set of functional chimeric lentivirus integrases, involving exchange of the N-terminal, central, and C-terminal regions of the human immunodeficiency virus type 1 (HIV-1) and visna virus integrase (IN) proteins. Based on those results, additional HIV-1/visna virus chimeric integrases were designed and purified. Each of the chimeric enzymes was functional in at least one oligonucleotide-based IN assay. Of a total of 12 chimeric IN proteins, 3 exhibit specific viral DNA processing, 9 catalyze insertion of viral DNA ends, 12 can reverse that reaction, and 11 are active for nonspecific alcoholysis. Functional data obtained with the processing assay indicate that the central region of the protein is responsible for viral DNA specificity. Target site selection for nonspecific alcoholysis again mapped to the central domain of IN, confirming our previous data indicating that this region can position nonviral DNA for nucleophilic attack. However, the chimeric proteins created patterns of viral DNA insertion distinct from that of either wild-type IN, suggesting that interactions between regions of IN influence target site selection for viral DNA integration. The results support a new model for the functional organization of IN in which viral DNA initially binds nonspecifically to the C-terminal portion of IN but the catalytic central region of the enzyme has a prominent role both in specific recognition of viral DNA ends and in positioning the host DNA for viral DNA integration.

FIG. 1

HIV-1/visna virus chimeric IN proteins. A linear representation of the IN protein is shown at the top, with the relative positions of the conserved H-H-C-C and D,D-35-E motifs indicated. The numbers above and below the linear map denote positions in the HIV-1 IN and visna virus IN proteins, respectively, that define domains utilized to form the chimeric proteins; note that amino acids 186 to 190 of HIV-1 IN are identical to residues 188 to 192 of visna virus IN. Proteins are presented schematically, with solid bars denoting fragments derived from HIV-1 and stippled bars indicating visna virus sequences. Wild-type and chimeric proteins are grouped in pairs a to g, as indicated under “Description.” Sets b to d are our original six chimeras, named with three uppercase letters. Sets e to f are new chimeras, named with two uppercase letters and two lowercase letters; the second and third letters in this series can be considered an extended core region. The results of various oligonucleotide-based IN assays are summarized at the right, where “Preferred Viral U3 Substrate” refers to processing. −, inactive in that assay; +, active for strand transfer but at too low a level to display a reproducibly discernible pattern.

FIG. 2

SDS-PAGE demonstrating purification of new chimeric integrases. Three microliters of each purified IN (indicated above the lanes; nomenclature as in Results) were heated in sample buffer and separated by SDS-PAGE, and the gel was stained with Coomassie blue. Proteins are paired (a, e, f, and g) as in Fig. 1. Molecular mass markers are in lane 1 (sizes are shown in kilodaltons at the left) and represent 200 ng per band.

FIG. 3

Disintegration activity of new chimeric integrases. (Left) The substrate is a four-oligonucleotide complex representing the predicted immediate adduct of the HIV-1 U5 DNA end (thick lines) integrated into host DNA (thin lines). IN-mediated cleavage after the CA releases the viral DNA end, with concomitant joining of the juxtaposed 5′-radiolabeled 16-mer to the 15 nucleotides beyond the CA to yield a radiolabeled 31-mer product (asterisks denote ³²P labels). The substrate was incubated with protein buffer or purified integrases under standard conditions for 90 min and analyzed as described in Materials and Methods. (Right) An autoradiogram from a denaturing polyacrylamide gel is shown. The sizes of the labeled component of the substrate and the product are indicated (in nucleotides) at the right, aligned with the complexes at the left from which they were derived. Proteins are paired (a, e, f, and g) as in Fig. 1. The two wild-type integrases and all six new chimeric integrases were active in this assay.

FIG. 4

Processing activity of new chimeric integrases. (Top) A schematic of the site-specific 3′-end processing reaction is shown. Cleavage of blunt-ended viral DNA after the invariant CA (shown in boldface) converts a radiolabeled 18-mer to a labeled 16-mer (asterisks denote ³²P labels). Duplex oligonucleotide 18-mer substrates derived from the U5 or U3 termini of HIV-1 DNA or the U3 terminus of visna virus DNA and 5′ labeled on the plus or minus strand, as indicated in panels A to C, were incubated with protein buffer or purified integrases under standard conditions for 90 min and analyzed as described in Materials and Methods. The region of the autoradiogram extending down to 12-mers is shown to demonstrate the specificity of the reactions. Proteins are paired (a, e, f, and g) as in Fig. 1. Biologically relevant specific cleavage products two nucleotides shorter than the substrate are indicated by arrows. Sequence-specific markers are included in lanes 1, and their sizes (in nucleotides) are indicated at the left.

FIG. 5

Duplicate processing reactions for informative proteins. Duplex oligonucleotide 18-mer substrates derived from the U3 termini of HIV-1 (A) or visna virus (B) DNA, which best distinguish between the wild-type enzymes, were 5′ radiolabeled on the minus strand and incubated with protein buffer or selected integrases. Details are described in the legend to Fig. 4.

FIG. 6

Strand transfer activity of chimeric integrases. A duplex oligonucleotide derived from the visna virus U3 end, but preprocessed by omission of the final two nucleotides from the minus strand, was used as the substrate for the complete set of 14 purified integrases during 3-h incubations, and reactions were analyzed as described in Materials and Methods. An autoradiogram from the upper region of a denaturing polyacrylamide gel is shown. Insertion of the recessed 3′ terminus into various sites along other oligonucleotides (shown as thinner lines in the scheme on the left) yield longer radiolabeled products (asterisks denote ³²P labels). The positions of the labeled 16-mer component of the substrate and the labeled strands of longer products are indicated to the left of the gel. Proteins are paired (a to g) as in Fig. 1. The different patterns produced by HHH IN and VVV IN are demonstrated in lanes 2 and 3. Nine of the 12 chimeras had activity in this assay, but ∼5- to 20-fold more counts per minute were loaded in lanes 4, 6 to 9, and 11 to 15 to display the patterns produced by less-active chimeras.

FIG. 7

Nonspecific alcoholysis activity of the new chimeric integrases. A schematic of the reaction is shown at the top. IN catalyzes attack by nucleophilic OH groups of various alcohols (ROH) that nick and join to newly exposed 5′ phosphate groups at sites of DNA cleavage (asterisks denote ³²P labels). Nicks occur at every site except those close to DNA ends, with a reproducible pattern of preferential sites that is a function of the target DNA sequence and source of IN. A 5′-labeled 23-mer of nonviral sequence was annealed to a complementary oligonucleotide, incubated for 90 min with the wild-type proteins or the new chimeras in the presence of 40% ethylene glycol, and analyzed as described in Materials and Methods. A sequence-specific oligonucleotide ladder (as markers) is in lane 1, and sizes (in nucleotides) are indicated at the left. Proteins are paired (a, e, f, and g) as in Fig. 1. Equal volumes of reaction mixtures were loaded so that relative intensities reflect efficiencies of the different proteins. Five of the six chimeras created patterns that segregated clearly to the HHH pattern (chimeras VHhv, HHhv, and HHvh, in lanes 6, 7, and 9, respectively) or to the VVV pattern (chimeras HVvh and VVvh, in lanes 5 and 8, respectively). Chimera VVhv (lane 10) was repeatedly inactive in this assay.

FIG. 8

Model for the functional organization of the IN protein. Protein monomers A and B (shaded ovals) are shown as a dimer, in parallel orientations, with a linear map of amino acid positions at the extreme left, numbered from the N terminus to the C terminus. Initially (left), viral DNA is bound nonspecifically by the C-terminal region of IN. Amino acids in the central region (perhaps residues 191 to 212) recognize the terminal six positions of the viral DNA sequence (only the final four positions are shown) and position the CA bases for attack by a water molecule that is positioned by the C-terminal portion of the active site (defined by the three invariant acidic amino acids in the central region). The terminal 3′-OH of the viral DNA itself can be used instead of water to release cyclic (GT) dinucleotides (21). In the nucleus (right), host DNA is bound by other portions of the central region of IN. Following a conformational change in the protein, each processed viral DNA end is positioned by the C-terminal portion of the active site, with the aid of the residues responsible for viral DNA recognition and perhaps the C-terminal region, to attack the host DNA. During nonspecific alcoholysis, small alcohols can be used by the active site without requiring residues outside of numbers 50 to 190. The model would accommodate actions occurring in trans or involving multimers larger than dimers.