To establish successful infection, a retrovirus must insert a DNA replica of its genome into host cell chromosomal DNA. This process is orchestrated by integrase, a viral enzyme that belongs to the DDE(D) nucleotidyltransferase/transposase superfamily. A tetramer of integrase assembles on viral DNA ends forming a highly stable nucleoprotein complex termed intasome 1,2. The intasome engages chromosomal DNA within a target capture complex to carry out strand transfer, irreversibly joining the viral and cellular DNA molecules. Using data collected at Diamond Light Source’s MX beamline I02, we have determined crystal structures of a retroviral intasome in complex with target DNA, elucidating the entire functional synaptic complex in its pre- and post-catalytic states. These fundamental results reveal the structural basis for retroviral DNA integration and moreover provide a framework for the design of integrases with altered target sequences.
To carry out its function, retroviral integrase must act upon two types of DNA substrates: a pair of viral DNA ends and chromosomal target DNA. Initially, a tetramer of integrase assembles on viral DNA ends, forming a long-lived nucleoprotein complex termed intasome. We recently reported a series of crystal structures that elucidated the intasome in its active and drug-inhibited states 1,2(reviewed in 3). In this work we took advantage of favourable biochemical properties of the integrase encoded by the prototype foamy virus (PFV). Although this retrovirus is harmless, it serves an excellent model for studies of retroviral DNA integration. Crucially, the conservation of the retroviral integrase active site allowed us to use PFV integrase as a convenient proxy to study the mechanism of action of HIV integrase strand transfer inhibitors and to explain the most important drug resistance mutations 1,2.
Figure 1: Target capture complex (TCC), viewed along (top) or normal (bottom) to the principal two-fold axis. The target DNA strands are shown in black and pink; the viral DNA strands are magenta and orange. Adapted from 3.
To elucidate how the retroviral integration machinery engages chromosomal DNA, we now co-crystallized the PFV intasome with a model target DNA construct, which was designed on the basis of the PFV integration site consensus. Inclusion of Mg2+ allowed strand transfer to occur during crystallization experiments, resulting in crystals of the post-catalytic strand transfer complex (STC), while pre-catalytic target capture complex (TCC) crystals were obtained in the absence of the essential catalytic metal (TCCApo). Furthermore, using a viral DNA mimic lacking the reactive 3´-hydroxyl group enabled us to grow crystals of the catalytically trapped complex (TCCddA) in the presence of Mg2+, which considerably extended their diffraction limit. The data were collected at the Diamond Light Source’s beamline I02, and the STC, TCCApo, and TCCddA structures were refined to 2.81, 3.32, and 2.97 Å resolution, respectively.
Figure 2: Conformation of the synapsed viral and target DNA molecules prior to (left) and following (right) strand transfer. Adapted from 3.
The structures revealed that target DNA binds along the groove dividing the symmetric intasome (Fig. 1). A severe deformation of the target DNA, accompanied by a dramatic expansion of the major groove, allows insertion of the scissile phosphodiesters into the intasome active sites separated by ~27 Å. By contrast, the intasome itself does not undergo significant changes upon target DNA capture. The bending of target DNA is focused on a single dinucleotide step at the centre of the integration site, leading to a complete unstacking of the consecutive base pairs (Fig. 2). Concordantly, PFV integration sites are naturally enriched with more flexible pyrimidine - purine dinucleotides at the central positions. The PFV TCC and STC crystal structures explained the classical early observation that pre-bending of target DNA strongly stimulates the strand transfer activity of retroviral integrases 4,5.
Figure 3: Ejection of the phosphodiester bond from the active site following strand transfer. Dotted red line indicates direction of the nucleophilic attack by the viral 3’-hydroxyl group on the target phosphodiester; orange arrow indicates the displacement of the target phosphodiester in the STC relative to its position in the TCC. Adapted from 3.
Although the overall conformations of the DNA chains within TCC and STC structures are quite similar (Fig. 2), the positions of the scissile phosphodiester in the TCC and the phosphodiester joining viral and tDNA strands in the STC do not match. While the former is placed into the integrase active site, the latter is shifted by ~2.3 Å and thus removed from the active site (Fig. 3). The configuration of the STC cannot support a reverse reaction, which would involve a nucleophilic attack by the 3´ hydroxyl of the target DNA on the newly formed phosphodiester. The conformational change following strand transfer makes the ligation of the viral and chromosomal DNAs irreversible.
The TCC and STC structures revealed a set of contacts between integrase and bases of host DNA, which explain the weak target sequence consensi observed for retroviral integration sites. Mutations at these positions strongly influenced integration site sequence preference. Among many remaining open questions are how the retroviral DNA integration machinery engages the nucleosome, how the resulting STC is disassembled, and how this process is coupled to post-integration DNA repair. We are also optimistic that PFV TCC and STC structures will guide the design of target DNA sequence-specific integrases, which would be extremely helpful in reducing the chance of side effects in gene therapy applications.
Maertens, G.N., Hare, S. and Cherepanov, P. The mechanism of retroviral integration from X-ray structures of its key intermediates. Nature.468, 326-329 (2010)
Medical Research Council grants G0900116 and G1000917.
Diamond Light Source is the UK's national synchrotron science facility, located at the Harwell Science and Innovation Campus in Oxfordshire.
Copyright © 2018 Diamond Light Source
Diamond Light Source Ltd
Harwell Science & Innovation Campus