Factors that mold the nuclear landscape of HIV-1 integration

Abstract

The integration of retroviral reverse transcripts into the chromatin of the cells that they infect is required for virus replication. Retroviral integration has far-reaching consequences, from perpetuating deadly human diseases to molding metazoan evolution. The lentivirus human immunodeficiency virus 1 (HIV-1), which is the causative agent of the AIDS pandemic, efficiently infects interphase cells due to the active nuclear import of its preintegration complex (PIC). To enable integration, the PIC must navigate the densely-packed nuclear environment where the genome is organized into different chromatin states of varying accessibility in accordance with cellular needs. The HIV-1 capsid protein interacts with specific host factors to facilitate PIC nuclear import, while additional interactions of viral integrase, the enzyme responsible for viral DNA integration, with cellular nuclear proteins and nucleobases guide integration to specific chromosomal sites. HIV-1 integration favors transcriptionally active chromatin such as speckle-associated domains and disfavors heterochromatin including lamina-associated domains. In this review, we describe virus-host interactions that facilitate HIV-1 PIC nuclear import and integration site targeting, highlighting commonalities among factors that participate in both of these steps. We moreover discuss how the nuclear landscape influences HIV-1 integration site selection as well as the establishment of active versus latent virus infection.

Figure 1.

Overview of HIV-1 cellular ingress and principle determinants of integration targeting. Infection is initiated by receptor binding and membrane fusion, which releases the viral core into the cell cytoplasm where reverse transcription begins. During reverse transcription the core is trafficked to the nuclear pore where it is transported into the nucleus via interactions between HIV-1 CA and several nucleoporins, including Nup358 and Nup153. Following translocation, CPSF6 frees the core from the nuclear pore complex and facilitates progression of the PIC beyond the nuclear periphery and into the nuclear interior. Integration is highly biased away from lamina-associated domains and towards speckle-associated domains, which are characterized by active transcription and high gene density. The interaction of PIC-borne IN with LEDGF/p75 directs integration into the interior of gene bodies. HIV-1 proviruses are typically well expressed by cellular RNA polymerase following integration. However, a small population of proviruses (marked by lollipops) are not expressed and become latent. These latent proviruses can be activated upon stimulation years after the initial infection.

Figure 2.

Nuclear pore complex organization and CA-interacting components. (A) Cartoon depiction of the nuclear pore complex. Structural elements were derived from entry 3103 in the Electron Microscopy Data Bank (EMDB) and PDB entries 5a9q and 5ijn in the Protein Data Bank (PDB). The diagram depicts the nuclear pore complex as a vertical cross-section through the 8-fold symmetric architecture, revealing Y-complex Nups and several inner ring Nups. The locations of cytoplasmic filament Nups and nuclear basket Nups were approximated manually. POMs are not depicted. The identities of individual nucleoporins that are depicted in the cartoon are labeled in matching colors and bold-face font. Nups previously shown to facilitate PIC nuclear import and/or interact with HIV-1 CA are labeled in italicized bold-face font. (B, C) Different perspectives of the intact 8-fold symmetrical nuclear pore. (B) Top-view and (C) view tilted 40° from the top clearly highlight the overall toroidal architecture of the complex.

Figure 3.

HIV-1 capsomeres and the capsid lattice. (A) The capsid lattice is comprised of exactly 12 CA pentamers and ∼200 hexamers, which are shown in top and side views. In both capsomeres the CA CTD is shown in dark blue. While the CA NTD within the hexamer is light blue, it is shown in orange within the pentamer. In each capsomere, a single CA subunit is depicted as a ribbon diagram with matched coloring scheme. (B) An all-atom model of the assembled capsid shell derived from PDB entry 3j3y using the color scheme defined in panel A.

Figure 4.

Capsid interactions in HIV-1 integration targeting. (A) Organization of principle CA-binding proteins including CypA, Nup358, Nup153 and CPSF6. Locations of CA-binding portions of Nup358, Nup153 and CPSF6 are colored as in subsequent panels. Domain labels are as follows: LRR – leucine-rich region; roman numerals I-IV – Ran binding domains I–IV; ZF – zinc finger; E3 – E3 ligase domain; CHD – cyclophilin homology domain; NTD – N-terminal domain; FG – phenylalanine/glycine repeat domain; RRM – RNA recognition motif; PRD – proline-rich domain; RSLD – arginine/serine-like domain. (B) Interactions of Cyp-like protein domains with the CA NTD. CypA and the C-terminus CHD of Nup358 both interact with the conserved CypA-binding loop, one of two principle binding sites within HIV-1 capsomeres. The CypA and Nup358 structures were derived from PDB entries 1ak4 and 4lqw, respectively. Secondary structural elements of CA are noted on the leftward image. (C) Structures of hexameric capsomeres with peptides derived from Nup153 (top) and CPSF6 (bottom) from PDB entries 4u0c and 4wym, respectively. These peptides lie in a pocket formed between two individual CA subunits. The binding orientations of the respective peptides are non-identical, highlighting the promiscuity of this binding pocket for mediating CA-host factor interactions. (D) Detailed superposition of Nup153 and CPSF6 peptides in complex with CA. Nup153-bound CA molecules are shown in blue while the CPSF6 CA pair is shown in red. For both pairs of CA molecules, individual monomers are differentiated by light and dark coloring. The interaction of each host factor with respective background CAs (light coloring) is anchored by a phenylalanine residue [F284 in CPSF6 (isoform 1 numbering scheme) and F1417 in Nup153]. Docking of this phenylalanine facilitates main chain hydrogen bonding of each host factor with CA residue N57. N74 in CA by contrast preferentially interacts with CPSF6 (green sticks) and not Nup153 (orange sticks). The most pronounced differences in binding modes are derived from interactions with the foreground (darker) CA monomer. CPSF6 adopts a nearly cyclic conformation and interacts primarily with the CTD of the second monomer. In contrast, Nup153 is more linear and interacts with the NTD of the foreground CA subunit.

Figure 5.

The integrase-LEDGF/p75 interaction. (A) Domain organization of HIV-1 IN and LEDGF/p75. Domain annotations are as follows: NTD – N-terminal domain; CCD – catalytic core domain; CTD – C-terminal domain; PWWP – Pro-Trp-Trp-Pro domain; CR – charged region; AT-hooks – adenosine/thymine DNA binding motif; IBD – integrase binding domain. The key interacting domains, the IN CCD and LEDGF/p75 IBD, are colored blue and dark red, respectively. (B) Depiction of the core tetramer of the HIV-1 strand transfer complex intasome (PDB 5u1c) bound by the LEDGF/p75 IBD, which was created by superimposing the CCDs of the IBD–HIV-1 IN CCD structure (PDB 2b4j) with the CCDs of the strand transfer complex. Because LEDGF/p75 interacts with integrase at the interface between two CCD dimers, a single intasome contains multiple potential LEDGF/p75 binding sites. Whether all or just a fraction of bound LEDGF/p75 molecules participates in HIV-1/lentiviral integration targeting is not presently known. The IBD color in panel B matches panel A; one of the two CCD dimers in B also matches the panel A coloring.