Back to the OMM Exhibits

HIV-1 Integrase
The Catalytic Core Domain & Minimal DNA Binding Domain

Matthew Bok Walsh (1) and David Marcey (2)
© David Marcey, 2001


 

I. Introduction
II. Structural Features
A. The Domains
B. Catalytic Core Domains
C. DNA Binding Domain
D. Interaction of Catalytic Core and DNA Binding Domain
E. Problems with a Dimeric Model
III. References

Note:  This exhibit is best viewed if the cue buttons ( ) are pressed in sequence and if the viewer does not independently manipulate the molecule on the left.


I. Introduction

Human immunodeficiency virus type 1 (HIV-1), like all retroviruses, depends upon the integration of a DNA copy of its viral genome into host cell chromosomes as part of its infection cycle. Multiple steps in this integration process are catalyzed by HIV-1 integrase, as shown schematically at left. The integration of HIV-1 DNA into the host chromosome is achieved by the integrase protein performing a series of DNA cutting and joining reactions (A-C).

(A) The first step in the integration process is 3' processing. This step requires linear double-stranded viral DNA with sequence specific 3' ends, synthesized by reverse transcription from the viral RNA genome. The integrase protein removes two nucleotides from each 3' end of this viral DNA, leaving recessed CA OH's at the 3' ends.

(B) In a second step, termed strand transfer, the integrase protein joins the previously processed 3' ends to the 5' ends of strands of target DNA at the site of integration. The 5' ends are produced by integrase-catalyzed staggered cuts, 5 bp apart. A Y-shaped, gapped, recombination intermediate results, with the 5' ends of the viral DNA strands and the 3' ends of target DNA strands remaining unjoined, flanking a gap of 5 bp (C).

(C) Integrase may catalyze the excision of viral DNA, termed disintegration. Alternatively, integration may ensue. This involves host DNA repair synthesis in which the 5 bp gaps between the unjoined strands (see above) are filled in and then ligated. Since this process occurs at both cuts flanking the HIV genome (only 1 is detailed at left), a 5 bp duplication of host DNA is produced at the ends of HIV-1 integration.

 


II. Structural Features

A. The Domains

The full length HIV-1 integrase (288 amino acids) has three domains; the catalytic core, the C-terminal, and the N-terminal domains. Although all three domains are required for integration, it is thought that the catalytic core domain (residues 50-212, shown at left) contains the active site responsible for catalysis of all the reactions of integration/disintegration. The C-terminal domain confers the capacity to bind both viral and host DNA. The structures of the catalytic core and C-terminal domains have been determined separately. The structure and function of the N-terminal domain are presently unknown, but it contains a His2Cys2 zinc binding motif suggesting interaction with nucleic acid.

  B. Catalytic core domain

The structure of the catalytic core domain (CCD) of HIV-1 integrase consists of a central five-stranded b-sheet with six surrounding helices . Three amino acids in the CCD are highly conserved among retrotransposon and retroviral integrases. Mutation of these residues generally leads to a loss of all catalytic activities of these proteins, and they are therefore thought to be essential components of the integrase active site. Two of these in HIV-1 integrase are Asp64 and Asp116 . The third conserved residue, Glu152, lies near the other two in a 13-residue disordered region (not visualized by x-ray crystallography) bordered by residues 140 and 154 .

Studies in vitro have indicated that functional HIV-1 integrase is multimeric protein, but the number of monomers in a functional enzyme has not been determined with certainty. Examination of the crystal structure shows two contacts between CCD monomers, one of which suggests a dimeric model for functional integrase. In this contact, the two monomers are related by a dyad axis, with a large, solvent-excluded surface (1300 Å2/monomer). {The rough dimeric model shown at left was constructed by the authors by orienting two monomers with Hyperchem, creating a dyad axis and approximating an appropriate interface.} The dimer is stabilized by salt bridges and hydrogen bonds involving b-strand 3 and a-helices 1, 3, 5, and 6 . The interaction between core domain monomers can be compared to the interactions involved in antibody binding of protein antigens, which generate solvent excluded interfaces of ~800 Å2 for each molecule.

Note that the putative active sites, one on both monomers (each containing Asp64 and Asp116), are separated by ~35 Å in the dimeric model . This feature is discussed further below, in the Problems with a Dimeric Model section.

C. DNA binding domain

The solution structure of the HIV-1 integrase minimal binding domain is a dimer. Each monomer is constructed of five b-strands that are arranged in an antiparallel orientation to form a five-stranded b-barrel .

Interactions between two, three-stranded b-sheets occur at the interface between monomers . These interactions are predominantly hydrophobic (Trp243, Ala248 and Val250 from one monomer interacting with Leu242, Val250, Ile257 and Val259 of the other) . Further stabilization of the dimer is achieved by hydrogen bonding between the carboxylate of Glu246 of one monomer to the carboxamide of Gln252 of the other .

The dimeric DNA binding domain possesses a large, saddle-shaped groove bounded by loops connecting b-strands 1 and 2 in each monomer . This groove has the appropriate dimensions to fit B-form DNA (24 X 23 X 12 Å). A model of DNA within the binding domain's groove, produced by positioning B-DNA into the the DNA binding domain with Hyperchem, is shown at left. This model approximates that of Lodi, et al.(1995). It can be observed that the side chains of Lys264 from each monomer can interact with the sugar-PO4 backbone of the DNA. Lys264 is a key residue in DNA interactions as revealed by mutagenesis expeiments. Further important interactions may include those between bases in the major groove of DNA and Arg231 of each monomer, located at the tips of the loops that clasp the DNA.

D. Interaction of Catalytic Core and DNA Binding Domain

The catalytic core and the DNA binding domain are both modeled (at left) as dimers. The separation between the C-terminal residues (Ile208's) of each monomer of the catalytic domain is 21 Å. The separation between the N-terminal residues (Ile220's) of the two monomers of the DNA binding domain is ~17 Å. With such dimensions, the dimerized catalytic core can be envisioned to fit directly on top of the dimerized DNA binding domain in the intact protein, the two domains being connected by an 11 residue stretch missing in both models.

E. Problems with a Dimeric Model

Despite the strong structural evidence supporting dimeric models for both the catalytic core and DNA binding domains, there are spatial problems to be considered. Any model of the mechanism with which a multimer carries out integration must be consistent with the nucleotide spacing between the phosphates on the two ends of each half of target DNA (see Introduction). This five nucleotide spacing requires a pair of active sites separated by approximately ~15 Å. However, as discussed above, the active sites on a dimeric model of the catalytic core domain are separated by ~ 35 Å .

Perhaps considerable distortion of the core domain and/or of the DNA substrate occurs upon binding. Alternatively, it may be that dimers associate into functional tetramers. A tetrameric model formed by a pair of dimers could position two of the active sites within the required 15 Å (Dyda, et al., 1994). In this model, two active sites from separate dimers would catalyze the 3' processing and strand transfer. The other two active sites would not carry out any catalysis, and would possibly serve only as structural components.


III. References

Dyda, F., Hickman, A.B., Jenkins, T.M., Engelman, A., Craigie, F., and D.R. Davies (1994). Crystal Structure of the Catalytic Domain of HIV-1 Integrase: Similarity to Other Polynucleotidyl Transferases. Science 266: 1981-1986.

Lee, S.P., Hyung, G.K., Censullo, M.L., and Han, M.K. (1995). Characterization of Mg2+-Dependent 3' Processing Activity for Human Immunodeficiency Virus Type 1 Integrase in vitro: Real-Time Kinetic Studies Using Fluorescence Resonance Energy Transfer. Biochemistry 34: 10205-10214.

Lodi, P.J., Ernst, J.A., Kuszewski, J., Hickman, A.B., Engelman, A., Craigie, R., Marius Clore, G., and A.M. Gronenborn (1995). Solution Structure of the DNA Binding Domain of HIV-1 Integrase. Biochemistry 34: 9826-9833.

 


1, Kenyon College, Gambier, Ohio. A first draft of this exhibit was created for D. Marcey's Molecular Biology class, Biology 63.


2, Kenyon College, Gambier, Ohio. Present address: California Lutheran University. Address correspondence to this author (see below).


Back to Introduction

Back to the OMM Exhibits

Feedback to David Marcey: marcey@clunet.edu