Retroviral DNA integration: reaction pathway and critical intermediates

Abstract

The key DNA cutting and joining steps of retroviral DNA integration are carried out by the viral integrase protein. Structures of the individual domains of integrase have been determined, but their organization in the active complex with viral DNA is unknown. We show that HIV-1 integrase forms stable synaptic complexes in which a tetramer of integrase is stably associated with a pair of viral DNA ends. The viral DNA is processed within these complexes, which go on to capture the target DNA and integrate the viral DNA ends. The joining of the two viral DNA ends to target DNA occurs sequentially, with a stable intermediate complex in which only one DNA end is joined. The integration product also remains stably associated with integrase and likely requires disassembly before completion of the integration process by cellular enzymes. The results define the series of stable nucleoprotein complexes that mediate retroviral DNA integration.

Figure 1

Concerted integration products are stably associated with integrase. (A) Schematic showing the expected integration products after deproteinization. (B) The products of an integration reaction were subjected to electrophoresis in an agarose gel after deproteinization (lane 1), after the addition of EDTA (lane 2), EDTA and 600 mM NaCl (lane 3), EDTA and 10 μg/ml heparin (lane 4) and EDTA and 0.1% NP-40 (lane 5). The migration position of linear size markers is shown on the left. (C, D) Kinetics of STC formation parallels formation of concerted integration product. Reaction mixtures were incubated for the indicated time and subjected to electrophoresis with (C) or without (D) deproteinization. The band just above the 6 kb marker in panel B, lane 1 and in panel C results from a half-site insertion of another donor DNA molecule into the concerted integration product. (E, F) show quantitation of the data in panels C and D, respectively. This quantitation slightly under-represents the abundance of the SSC in panel F because some of the complex runs as a smear. Similarly, in the absence of deproteinization, some of the half-site product in panel F runs as a smear and is slightly under-represented.

Figure 2

Analysis of the stable complexes by two-dimensional gel electrophoresis. (A) The products of a 1 h reaction were subjected to electrophoresis in a native dimension (left to right). The gel was then soaked in SDS to dissociate protein–DNA complexes and then subjected to electrophoresis in a second dimension (top to bottom). (B) Time course of two-dimensional gel analysis. Each panel is the same as in (A), except that reactions were incubated for 0.5, 1 or 2 h. Note that the complex with only one viral DNA end joined is chased to concerted product by the 2 h time point. The smear resulting from integration of pairs of viral ends into another viral DNA end is indicated by the arrow. Quantitation of the data shows that the percentage of STC with both viral DNA ends covalently joined at 0.5, 1.0 and 2.0 h is 39, 60 and 89%, respectively.

Figure 3

The SSC accumulates in the presence of an inhibitor of DNA strand transfer. Reactions, which were incubated for the indicated times, included a 984 bp viral DNA substrate, which is more efficient for SSC formation than the 355 bp substrate. Products were subjected to electrophoresis in an agarose gel without deproteinization. (A), No inhibitor and (B) reactions were incubated in the presence of 0.5 mM inhibitor. The band above the 6 kb marker in panel B varies in intensity between experiments and we suspect it represents two SSCs associated with one another. (C) Complexes are not formed with a 970 bp substrate that lacks the HIV-1 LTR sequence. The substrate was incubated with integrase and subjected to electrophoresis as for panel A.

Figure 4

Complex B contains a pair of viral DNA ends. Complexes were assembled with a mixture of 984 and 1513 bp viral DNA ends for 30 min in the presence of inhibitor and subjected to electrophoresis without (A) or with (B) deproteinization. The ratio of short:long DNA was, lane 1, 2:1; lane 2, 1:1 and lane 3, 1:2.

Figure 5

The SSC contains both processed and unprocessed viral DNA. (A) In order to enable the base resolution required to monitor 3′ processing, the viral DNA substrate was internally labeled with ³²P such that cleavage of unprocessed DNA by EcoRI gives a 35-base-labeled fragment and cleavage of processed DNA gives a labeled 33 base fragment upon denaturing gel electrophoresis. (B) Lane 1, oligonucleotide markers for the 35 and 33 base cleavage products; lane 2, unreacted viral DNA excised from the gel and processed identically to DNA from the excised SSC and cut with EcoRI and lane 3, DNA extracted from SSCs and cut with EcoRI. The band migrating at the 34 base position results from a contaminating nuclease activity.

Figure 6

(A) Footprint of integrase in the SSC and STC. Complexes were assembled in the presence of inhibitor and treated with limited digestion with DNase I before separation on an agarose gel. After excision of the bands, DNA was extracted and subjected to electrophoresis in a denaturing polyacrylamide gel. Lanes 1, 2 and 3 are G, G+A and C+T markers, respectively, made by Maxam–Gilbert chemical cleavage. The samples shown in lanes 4–6 were processed identically. Lane 4, free DNA; lane 5, STC; and lane 6, SSC. (B) Footprint of integrase in SSC complex assembled with processed viral DNA ends with a 3′ ddA that blocks strand transfer. Lanes 1, 2, 3 and 4 are A+C, G, G+A and C+T chemical cleavage markers, respectively. Lanes 5 and 6 are free DNA and SSC, respectively.

Figure 7

The SSC contains a tetramer of integrase. Complexes were assembled in the presence of inhibitor and crosslinked with DSS before separation in an agarose gel. Bands were excised and extracted protein was subjected to electrophoresis in a denaturing polyacrylamide gel. The gel was transferred to a membrane and probed for integrase by Western blotting. Lane 1, integrase without crosslinking; lane 2, integrase after crosslinking in solution; lane 3, size marker; and lane 4, integrase from crosslinked SSCs. We note that the electrophoretic mobility of the crosslinked integrase subunits is slightly faster than expected based on their molecular weight relative to the non-crosslinked size markers. This is owing to conformational constraints resulting from crosslinking. However, comparison of the ladder of integrase subunits crosslinked in solution with the crosslinked species present in the SSC unambiguously establishes that the major form is tetrameric.