Cyclophilin A stabilizes the HIV-1 capsid through a novel non-canonical binding site

ARTICLE Received 4 Aug 2015 | Accepted 12 Jan 2016 | Published 4 Mar 2016 DOI: 10.1038/ncomms10714 OPEN Cyclophilin A stabilizes the HIV-1 capsid through a novel non-canonical binding site Chuang Liu1,2,*, Juan R. Perilla3,*, Jiying Ning1,2, Manman Lu2,4, Guangjin Hou2,4, Ruben Ramalho5, Benjamin A. Himes1, Gongpu Zhao1,2, Gregory J. Bedwell6, In-Ja Byeon1,2, Jinwoo Ahn1,2, Angela M. Gronenborn1,2, Peter E. Prevelige2,6, Itay Rousso5, Christopher Aiken2,7, Tatyana Polenova2,4, Klaus Schulten3 & Peijun Zhang1,2 The host cell factor cyclophilin A (CypA) interacts directly with the HIV-1 capsid and regulates viral infectivity. Although the crystal structure of CypA in complex with the N-terminal domain of the HIV-1 capsid protein (CA) has been known for nearly two decades, how CypA interacts with the viral capsid and modulates HIV-1 infectivity remains unclear. We determined the cryoEM structure of CypA in complex with the assembled HIV-1 capsid at 8-Å resolution. The structure exhibits a distinct CypA-binding pattern in which CypA selectively bridges the two CA hexamers along the direction of highest curvature. EM-guided all-atom molecular dynamics simulations and solid-state NMR further reveal that the CypA-binding pattern is achieved by single-CypA molecules simultaneously interacting with two CA subunits, in different hexamers, through a previously uncharacterized non-canonical interface. These results provide new insights into how CypA stabilizes the HIV-1 capsid and is recruited to facilitate HIV-1 infection. 1 Department of Structural Biology, University of Pittsburgh School of Medicine, Pittsburgh, Pennsylvania 15260, USA. 2 Pittsburgh Center for HIV Protein Interactions, University of Pittsburgh School of Medicine, Pittsburgh, Pennsylvania 15260, USA. 3 Department of Physics and Beckman Institute, University of Illinois at Urbana-Champaign, Urbana, Illinois 61801, USA. 4 Department of Chemistry and Biochemistry, University of Delaware, Newark, Delaware 19716, USA. 5 Department of Physiology and Cell Biology, Ben-Gurion University of the Negev, Be’er-Sheva, 84105, Israel. 6 Department of Microbiology, University of Alabama at Birmingham, Birmingham Alabama 35294, USA. 7 Department of Pathology, Microbiology and Immunology, Vanderbilt University School of Medicine, Nashville, Tennessee 37232, USA. * These authors contributed equally to this work. Correspondence and requests for materials should be addressed to K.S. (email: ) or to P.Z. (email: ). NATURE COMMUNICATIONS | 7:10714 | DOI: 10.1038/ncomms10714 | www.nature.com/naturecommunications 1 ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms10714 I nfection by retroviruses, including HIV-1, is critically dependent on the functions of the viral capsid1–3, which plays multiple roles during replication, including the prevention of innate sensor triggering4,5 and regulation of reverse transcription1,6 and nuclear import7–9. Such functions are highly dependent on interactions between the viral capsid and cellular factors. The host cell protein cyclophilin A (CypA) binds directly to the HIV-1 capsid and modulates capsid uncoating and viral infectivity10–12. Interference with CypA-binding inhibits HIV-1 replication in cell culture13–15. In addition, host cell proteins containing the CypA domain, including Trim-Cyp and Nup358, interact directly with the viral capsid and control HIV-1 infection8,16,17. CypA-binding appears to stabilize or destabilize the HIV-1 capsid, depending on the cell type6,18. In addition, the interaction between CypA and HIV-1 CA promotes HIV-1 infection of human cells19,20, yet, in non-human primates, the same interaction enhances the anti-HIV-1 restriction activity of Trim5a (refs 21–23). Furthermore, while CypA promotes HIV1 reverse transcription in human cells, its cell-type specific effect is best correlated with nuclear entry, and probably involves an unknown CypA-dependent host restriction factor24. Owing to this complexity, understanding the role of CypA in HIV-1 infection has been difficult, especially given the limited structural information that is available. The crystal structure of CypA in complex with the CA N-terminal domain (CANTD) shows that CA residues A88-G89P90 and the configuration of the CypA-binding loop are important for the binding interaction11. Binding of CypA to the CypA-binding loop of the viral capsid is essential for viral infectivity25,26, as substitutions in loop residues G89 or P90 in HIV-1 CA are deleterious to replication14,21,27. CypA binding to monomeric CA can be detected at high protein concentration11, but binding is enhanced by CA multimerization10, suggesting that CypA favours binding to an assembled capsid. Despite extensive studies on the interaction between CypA and HIV-1 CA, structural information on how CypA interacts with the assembled viral capsid, and an understanding of the consequences of the interaction for capsid function, are lacking. To elucidate the molecular interactions between CypA and the HIV-1 capsid, we determined the structure of CypA in complex with an HIV-1 CA tubular assembly at 8-Å resolution by cryoEM. The density map, combined with structure-guided molecular dynamics (MD) simulations, unexpectedly revealed a novel, noncanonical, second capsid-binding site on CypA that is vital for stabilizing the viral capsid. The finding was confirmed experimentally by solid-state NMR. Results Binding of CypA to wild-type (wt) CA assemblies. We began our studies by characterizing the binding of CypA to HIV-1 CA tubular assemblies at various CypA:CA molar ratios. Binding was measured in a high-speed centrifugation assay, after incubation of CypA with a fixed concentration of pre-assembled wild-type (wt) CA tubes. Incubation of CypA with CA tubes, up to a 6:6 molar ratio, resulted in co-sedimentation of CypA and CA (Fig. 1a, top), with the CypA:CA binding ratio increasing as the CypA concentration increased, up to 40 mM, beyond which point, binding was saturated (Fig. 1b, solid line). Interestingly, while a relatively low amount of CypA (CypA:CA input ratios r 2:6) slightly enhanced the amount of CA in the pelleted fraction (Fig. 1a top), CypA:CA ratios above 2:6 led to a reduction in the amount of pelleted CA (Fig. 1a, top). This finding is consistent with an earlier report that high molar ratios of CypA:CA alter CA assembly in vitro28. 2 We also tested the effect of CypA on the process of CA assembly by including it in the CA assembly reaction under the high-salt condition (2.25 M NaCl) (Fig. 1a, bottom). In this case, CypA binding was consistently higher compared with when it was added to pre-assembled CA tubes (Fig. 1b, dotted lines). However, the presence of CypA during CA assembly induced marked morphological changes in the resulting CA assemblies (Fig. 1c): in the absence of CypA, long CA tubes were formed; at an intermediate level of CypA, short tubes and cones were generated; and, at high CypA concentrations (6:6 CypA:CA molar ra (...truncated)