Date Published: June 21, 2016
Publisher: Public Library of Science
Author(s): Adarsh Dharan, Sarah Talley, Abhishek Tripathi, João I. Mamede, Matthias Majetschak, Thomas J. Hope, Edward M. Campbell, Bryan R. Cullen.
Following envelope mediated fusion, the HIV-1 core is released into the cytoplasm of the target cell and undergoes a series of trafficking and replicative steps that result in the nuclear import of the viral genome, which ultimately leads to the integration of the proviral DNA into the host cell genome. Previous studies have found that disruption of microtubules, or depletion of dynein or kinesin motors, perturb the normal uncoating and trafficking of the viral genome. Here, we show that the Kinesin-1 motor, KIF5B, induces a relocalization of the nuclear pore component Nup358 into the cytoplasm during HIV-1 infection. This relocalization of NUP358 is dependent on HIV-1 capsid, and NUP358 directly associates with viral cores following cytoplasmic translocation. This interaction between NUP358 and the HIV-1 core is dependent on multiple capsid binding surfaces, as this association is not observed following infection with capsid mutants in which a conserved hydrophobic binding pocket (N74D) or the cyclophilin A binding loop (P90A) is disrupted. KIF5B knockdown also prevents the nuclear entry and infection by HIV-1, but does not exert a similar effect on the N74D or P90A capsid mutants which do not rely on Nup358 for nuclear import. Finally, we observe that the relocalization of Nup358 in response to CA is dependent on cleavage protein and polyadenylation factor 6 (CPSF6), but independent of cyclophilin A. Collectively, these observations identify a previously unappreciated role for KIF5B in mediating the Nup358 dependent nuclear import of the viral genome during infection.
Human Immunodeficiency Virus Type-1 (HIV-1), like all primate lentiviruses, possesses the ability to infect non-dividing cells. The ability to infect non-dividing cells is conveyed by the viral capsid (CA) protein which makes up the viral core that houses the viral genome [1,2,3]. CA has important functions during the early stages of HIV infection. Specifically, it acts to shield the viral genome from cytoplasmic sensors capable of inhibiting infection and activating innate immune signaling pathways[4,5,6,7]. The ability to protect the viral genome from host factors in the cytoplasm and also mediate the nuclear import of the viral genome is complicated by the dimensions of the viral core, which at ~120nm x 60 nm [8,9], significantly exceeds the size limitation of nuclear pore cargoes, which is ~39 nm [10,11]. These findings collectively suggest that core disassembly, known as uncoating, must be properly regulated so that the viral genome can be delivered to the nucleus while keeping the genome shielded from host factors in the cytoplasm. CA must therefore interact with numerous host factors to ensure that these functions are performed in a spatiotemporally appropriate fashion.
In this study, we demonstrate that HIV-1 infection induces the KIF5B dependent relocalization of Nup358. This relocalization of Nup358, and cytoplasmic association with HIV-1, was dependent on entry of HIV-1 into the target cell cytoplasm, as this relocalization was not observed using WT/ΔEnv HIV-1. Although the role of Nup358 in facilitating HIV-1 infection has been observed by others [12,14,16,17,18,26,27], a clear understanding of the molecular interactions occurring between CA and Nup358 during infection has not emerged. As a large, multi-domain protein associated with the cytoplasmic side of the NPC, numerous functions have been ascribed to Nup358 with respect to its role in HIV-1 infection[16,18,26,27]. Two studies have supported a role for Nup358 driving HIV-1 uncoating via interaction with the Cyp homology domain in Nup358 [18,26]. However, another study which demonstrated that the Cyp homology domain is not required for the Nup358 dependent enhancement of HIV-1 infection suggests that the role of Nup358 in HIV-1 infection cannot be fully explained by the Cyp homology domain . To understand the determinants in Nup358 required for the interaction with HIV-1 cores, we exploited the measurable association between Nup358 and HIV-1 cores in the cytoplasm, and CA mutants which disrupt putative Nup358 binding surfaces. Specifically, we observed that mutations which disrupt the hydrophobic binding pocket in assembled CA (N74D) and mutations which disrupt the conserved Cyp binding loop on CA (P90A) both perturb the relocalization of Nup358 induced by HIV-1 infection (Fig 4) and the association of CA and Nup358 during infection (Figs 5 and 6). The association of CA and Nup358 in the cytoplasm was measured in two distinct but related methods. When the intensity of Nup358 staining associated with individual viral cores was measured, neither the N74D or P90A mutant exhibited significantly more colocalization with Nup358 than WT/ΔEnv control virus (Fig 5), suggesting that both CA epitopes are required for Nup358 association. However, when the interaction between CA and Nup358 was measured by PLA, we observed a significant degree of interaction between P90A CA and Nup358, while the N74D mutant did not induce significantly more puncta than was observed following WT/ΔEnv or mock infection (Fig 6). This suggests that although the association between CA and Nup358 may be facilitated by both the hydrophobic binding pocket and Cyp binding loop of CA, the hydrophobic pocket is required for this interaction. Of note, the PLA, by its nature, detects the frequency but not the magnitude of interactions between the viral core and Nup358, as the puncta identified are generated by enzymatic amplification of nucleotides when individual secondary antibodies are in close enough proximity to enable the reaction. As such, the PLA sensitively detects the interaction between two proteins, rather than the degree of association. Despite these methodological differences, both methods demonstrated that the N74D mutation prevented the association of Nup358 with viral cores. Differences in the colocalization analysis and PLA analysis suggest that the P90A mutation reduces, but does not eliminate, the binding of Nup358 to CA. Multivalent binding of Nup358 and the viral core may help to reconcile previously discordant observations regarding the role of the Cyp homology domain in infection [18,26,27]. We also observe that CPSF6 depletion similarly prevents the redistribution of NUP358 to the cytoplasm and HIV-1 association. This suggests that CPSF6 somehow facilitates the interaction between HIV-1 and NUP358, consistent with a role for CPSF6 in HIV-1 nuclear import observed in other imaging based approaches [48,49]. CPSF6 is known to bind to the hydrophobic pocket, formed between adjacent CA monomers, and this pocket is known to be disrupted by the N74D mutation [22,23,24,25,50]. The similarity in the results obtained following CPSF6 depletion and using the N74D mutant suggest a common mechanism. However, given the large number of hydrophobic pockets present in the viral core, it remains unclear if CPSF6 acts as an adaptor linking the viral core and NUP358, or if CPSF6 engagement somehow facilitates the ability of NUP358 to bind the core in similar, adjacent hydrophobic pockets present on the core.