JVI Accepted Manuscript Posted Online 8 June 2016 J. Virol. doi:10.1128/JVI.00458-16 Copyright © 2016 Bulli et al. This is an open-access article distributed under the terms of the Creative Commons Attribution 4.0 International license. 1 Complex interplay between HIV-1 Capsid and MX2-independent IFNα- 2 induced antiviral factors 3 4 Lorenzo Bulli1,2,*, Luis Apolonia3,*, Juliane Kutzner1, Darja Pollpeter3, Caroline 5 Goujon3,¶, Nikolas Herold2,4, Sarah-Marie Schwarz2, Yannick Giernat2, Oliver 6 T. Keppler2,†, Michael H. Malim3,§ and Torsten Schaller1,2,3,§ 7 D 8 1Department of Infectious Diseases, Virology, Heidelberg University Hospital, o w n 9 Im Neuenheimer Feld 324, 69120, Heidelberg, Germany lo 10 2Institute for Medical Virology, University Hospital Frankfurt/Main, Paul Ehrlich ad e 11 Str. 40, 60596 Frankfurt, Germany d f 12 3Department of Infectious Diseases, King's College London, Great Maze ro m 13 Pond, SE1 9RT London, United Kingdom h t 14 4Childhood Cancer Research Unit, Astrid Lindgrens Children’s Hospital, tp : / 15 Karolinska Hospital, Q6:05, SE-171 76, Stockholm, Sweden /jv 16 i.a s 17 ¶Present address: Centre d’études d’agents Pathogènes et Biotechnologies m . o 18 pour la Santé CPBS - FRE 3689 / CNRS – UM, 1919 route de Mende, 34 293 r g / 19 MONTPELLIER Cedex 5, France o n 20 †Present address: Max von Pettenkofer-Institute, Virology, Ludwig-Maximilians M a 21 University Munich, 81377, Munich, Germany r c h 22 2 23 *Equal contributions 8, 2 24 §Co-corresponding authors 0 1 25 Emails: [email protected], [email protected] 9 b 26 y g 27 Key words: HIV-1, interferons, post-entry blocks, MX2, CPSF6, CypA, u e s 28 cyclosporine, capsid t 1 29 Abstract 30 31 Type I interferons (IFNs), including IFNα upregulate an array of interferon- 32 stimulated genes (ISGs) and potently suppress HIV-1 infectivity in CD4+ T 33 cells, monocyte-derived macrophages (MDMs) and dendritic cells (MDDCs). 34 Recently, we and others identified the ISG myxovirus resistance 2 (MX2) to 35 block HIV-1 nuclear entry. However, additional antiviral blocks exist upstream D 36 of nuclear import, but the ISGs that suppress infection e.g. prior to (or during) o w n 37 reverse transcription remain to be defined. Here we show that HIV-1 CA lo a 38 mutants such as N74D or A105T, which both allow escape from inhibition by d e 39 MX2 and the truncated version of the cleavage and polyadenylation specific d f 40 factor 6 (CPSF6), as well as the cyclophilin A (CypA)-binding loop mutant ro m 41 P90A, each exhibit increased sensitivity to IFNα-mediated inhibition. Using h t 42 CRISPR/Cas9 technology, we demonstrate that the IFNα-hypersensitivity of tp : / 43 these mutants in THP-1 cells is independent of MX2 or CPSF6. As expected, /jv 44 CypA depletion had no additional effect on the behavior of the P90A mutant, i.a s m 45 but modestly increased the IFNα-sensitivity of wild type virus. Interestingly, . o 46 the infectivity of wild type or P90A virus could be rescued from the MX2- r g / 47 independent IFNα-induced blocks in THP-1 cells by treatment with o n 48 cyclosporine (Cs), or its non-immunosuppressive analogue SDZ-NIM811, M a 49 indicating that Cs-sensitive host cell cyclophilins other than CypA contribute to r c h 50 the activity of IFNα-induced blocks. We propose that cellular interactions with 2 8 51 incoming HIV-1 capsids help shielding the virus from recognition by antiviral , 2 52 effector mechanisms. Thus, the CA protein is a fulcrum for the dynamic 0 1 53 interplay between cell-encoded functions that inhibit or promote HIV-1 9 b 54 infection. y g 55 u e s 56 Importance t 57 Human immunodeficiency virus type-1 (HIV-1) is the causative of the acquired 58 immunodeficiency syndrome (AIDS). During acute HIV-1 infection, numerous 59 pro-inflammatory cytokines are produced, including type I interferons (IFNs). 60 IFNs can limit HIV-1 replication by inducing the expression of a set of antiviral 61 genes that inhibit HIV-1 at multiple steps in its life cycle, including the post- 62 entry steps of reverse transcription and nuclear import. This is observed in 2 63 cultured cell systems, as well as in clinical trials in HIV-1 infected patients. 64 The identities of the cellular antiviral factors, their viral targets and the 65 underpinning mechanisms are largely unknown. We show here that the HIV-1 66 Capsid protein plays a central role in protecting the virus from IFN-induced 67 inhibitors that block early post-entry steps of infection. We further show that 68 host cell cyclophilins play an important role in regulating these processes, 69 thus highlighting the complex interplay between antiviral effector mechanisms D 70 and viral survival. o w n 71 lo a d e d f r o m h t t p : / / jv i. a s m . o r g / o n M a r c h 2 8 , 2 0 1 9 b y g u e s t 3 72 Introduction 73 Acute human immunodeficiency virus type 1 (HIV-1) infection presents with a 74 dramatic loss of CD4+ T cells, which is accompanied by the production of 75 large quantities of cytokines (1, 2). Studies on simian immunodeficiency virus 76 (SIV) infection of macaques suggest that this cytokine production contributes 77 to initial limitation of viral spread, lowering the viral burden to a level defining 78 the virological set point, and facilitating the partial recovery of CD4+ T cell D 79 counts (3). o w n 80 lo a 81 Type I interferons (IFNs), a group of cytokines mainly released by d e 82 plasmacytoid dendritic cells during acute virus infection (4), include thirteen d f 83 different subtypes of IFNα as well as IFNβ, IFNε, IFNκ and IFNω (5) and have ro m 84 long been known to potently suppress HIV-1 replication in certain types of h t 85 natural target cells (6-19). In addition to treating infections by other human tp : / 86 pathogens (e.g. hepatitis C virus, HCV), recombinant IFNα therapy has also /jv 87 been investigated as a treatment strategy for HIV-1 infection. Although a i.a s m 88 substantial reduction in viral load was observed in chronic infection, viral . o 89 rebound over time suggests that HIV-1 in-patient evolution may overcome r g / 90 IFNα-induced antiviral host factors (20, 21). It is therefore likely that different o n 91 HIV-1 strains have different sensitivities to type I IFNs. Comparison of diverse M a 92 HIV-1 strains suggested that transmitted founder (T/F) viruses of subtype B, r c h 93 but not subtype C, show a relative resistance to IFNα-induced blocks, arguing 2 8 94 that type I IFNs may play an important role in limiting transmission in a , 2 95 subtype-defined context (22-24). The viral determinants for partially 0 1 96 overcoming the IFNα-induced blocks to HIV-1 are unknown. It is therefore 9 b 97 important to identify the host cell effectors induced by type I IFNs, and to y g 98 understand the molecular interplay between the host and the virus after IFNα u e s 99 treatment. t 100 101 Addition of type I IFNs to cultured CD4+ T cells or monocyte-derived 102 macrophages (MDM) changes the expression profile of thousands of host 103 genes (25) and induces the production of many antiviral proteins of which only 104 a few have been characterized in detail (reviewed in (26, 27)). Pre-incubation 105 of susceptible cells with type I IFNs blocks HIV-1 infection at an early step 4 106 prior to or during reverse transcription (17, 28-31). The cellular host factors 107 mediating this effect are unknown. One recently discovered type I IFN- 108 induced factor that inhibits HIV-1 is the GTPase myxovirus resistance 2 (MX2, 109 also called MXB) (32-34). MX2 blocks HIV-1 after reverse transcription at the 110 level of nuclear entry, arguing that IFN-induced host cell barriers are likely to 111 interfere with HIV-1 infection at multiple early steps (32). Intriguingly, MX2 112 restriction of HIV-1 appears to be sensitive to changes in the HIV-1 capsid D 113 protein (Gagp24/CA), reminiscent to restriction by the bona fide CA binding o w n 114 factor TRIM5α (32-35). Consistent with these observations, in vitro studies lo a 115 suggest direct binding of MX2 to CA (36, 37). In addition, certain T/F viruses d e 116 show some degree of resistance to the antiviral activity of MX2, suggesting a d f 117 functional role of MX2 in limiting HIV-1 transmission (38). ro m 118 h t 119 The HIV-1 CA protein is essential for efficient infection and is genetically tp : / 120 fragile, i.e. many amino acid residues are highly conserved and many single /jv 121 amino acid substitutions can abrogate viral infectivity (39, 40). Most likely this i.a s m 122 is due to the central function of CA in the assembly and architecture of the . o 123 viral core, a complex conical fullerene-like structure consisting of hexameric r g / 124 and pentameric CA-sub-complexes (41-45). Amino acid substitutions in CA o n 125 can affect capsid assembly, stability and uncoating (46, 47), reverse M a 126 transcription (48), usage of a productive nuclear import pathway (47, 49), r c h 127 infection of non-dividing cells (50), as well as integration site selection (51) – 2 8 128 processes that are all interlinked. , 2 129 0 1 130 Some HIV-1 mutants carrying changes in CA, including N74D or P90A, have 9 b 131 been shown to induce IFNβ secretion in MDMs or dendritic cells (DCs), as a y g 132 result of cyclic GMP-AMP synthase (cGAS)-mediated recognition of reverse u e s 133 transcription intermediates, arguing that the capsid shell may ordinarily shield t 134 viral DNA from recognition by innate immune sensors (52-54). The inability of 135 these CA mutant viruses to replicate efficiently in MDMs (51, 55) may 136 therefore be partially explained by a cascade of innate sensing, type I IFN 137 secretion and induction of interferon stimulated genes (ISGs). 138 5 139 The HIV-1 CA mutant N74D escapes inhibition by a truncated version of the 140 cleavage and polyadenylation specific factor 6 (CPSF6-358) and does not 141 bind a CPSF6 derived peptide in vitro (49, 56, 57). Similarly, CA mutant 142 A105T is not restricted by CPSF6-358 (58). Cyclophilin-binding loop mutant 143 CA P90A binds with reduced affinity to the peptidylprolyl cis-trans isomerase 144 A (PPIA), also known as cyclophilin A (CypA), and is resistant to CypA- 145 mediated isomerization of the G89-P90 peptide bond in CA (59-61). In D 146 addition to CypA (62, 63) and CPSF6 (56, 57), HIV-1 CA has also been o w n 147 proposed to bind the nuclear pore complex (NPC) protein NUP153 (64), as lo a 148 well as to the cyclophilin domain of NUP358 (51, 65), and binding-deficient d e 149 CA mutants have been characterized. All of the aforementioned host proteins d f 150 have been invoked as co-factors acting in early HIV-1 infection. (49, 66-69). ro m 151 h t 152 While HIV-1 CA mutants can escape inhibition by ectopically expressed MX2, tp : / 153 we now demonstrate that the infectivities of HIV-1 CA mutants N74D, A105T /jv 154 or P90A, are hypersensitive to IFNα induced suppression when compared to i.a s m 155 wild type virus. This suggests that the susceptibility of viral determinants that . o 156 are targeted by IFNα-induced MX2-independent blocks is increased when r g / 157 capsid functionality is compromised. In addition, this MX2-independent IFNα- o n 158 induced post-entry block can be relieved by pharmacological inhibition of M a 159 cyclophilins, suggesting a role of host cell cyclophilins in the early type I IFN r c h 160 induced suppression of HIV-1 infection. We suggest that the CA protein and 2 8 161 the capsid core may therefore protect incoming HIV-1 nucleic acids from , 2 162 detection by innate pattern recognition receptors (52, 53) as well as IFNα- 0 1 163 induced effectors, thereby providing dual protection against host defense. 9 b 164 y g 165 Material and Methods u e s 166 Cells t 167 THP-1 cells were grown in RPMI-1640 Glutamax (Gibco) supplemented with 168 10% heat-inactivated fetal calf serum (FCS) and 100 U/ml penicillin and 100 169 µg/ml streptomycin. THP-1 cells were differentiated with 25 ng/ml phorbol 12- 170 myristate 13-acetate (PMA) (Sigma Aldrich) for 24 h. Purification of primary 171 blood mononuclear cells (PBMC) has been described before (70). Primary 172 CD4+ T cells or CD14+ monocytes were derived from 50 ml whole blood or 6 173 from buffy coats and grown in RPMI-1640 Glutamax with 10% heat- 174 inactivated FCS and penicillin/streptomycin. CD4+ T cells were isolated using 175 the CD4+ T Cell Isolation Kit II (Miltenyi Biotec) or the RosetteSep Human 176 CD4+ T Cells Enrichment Cocktail (StemCell Technologies), activated with 177 100 IU/ml IL-2 (Biomol) and 2 µg/ml PHA (Sigma Aldrich) for 3 days. 178 Monocyte-derived macrophages were differentiated for 7 days using 100 179 ng/ml granulocyte-macrophage colony stimulating factor (GM-CSF; R&D D 180 Systems). 293T were grown in Dulbecco’ s modified Eagle medium (DMEM o w n 181 Glutamax) (Gibco) with 10% heat-inactivated FCS and penicillin/streptomycin. lo a 182 d e 183 Plasmids and viral vectors d f 184 VSV-G pseudotyped wild type or CA mutant GFP encoding HIV-1 vectors ro m 185 were produced using the GagPol encoding plasmid pCMV-ΔR8.91, the GFP h t 186 reporter vector pCSGW and the VSV-G encoding plasmid pMD.G which have tp : / 187 been described before (51). pCMV-ΔR8.91 derived HIV-1 GagPol plasmid /jv 188 encoding SIV CA has been described before (71). Full-length wild type i.a MAC s m 189 and CA mutant HIV-1 GFP reporter virus were generated from pNLENG- . o 190 IRES-Nef (51, 72). THP-1 CRISPR/Cas cells were generated by transduction r g / 191 with VSV-G pseudotyped HIV-1 lentiviral particles produced using pCMV- o n 192 ΔR8.91, pMD.G and plentiCRISPRv2 (Addgene) (73, 74). Guide RNA M a 193 encoding oligos (MWG/Eurofins) were annealed and cloned into BsmBI r c h 194 linearized plentiCRISPRv2 according to the manufacturer’s guidelines 2 8 195 (Addgene). Oligos for MX2g1 were fwd/ rev , 2 196 caccgAATTGACTTCTCCTCCGGTA/ aaacTACCGGAGGAGAAGTCAATTc, 0 1 197 for MX2g2 were fwd/ rev caccgACAAGCCTTGGCCCTACCGG/ 9 b 198 aaacCCGGTAGGGCCAAGGCTTGTc, for CPSF6g1 were fwd/ rev y g 199 caccgATAGACATTTACGCGGATGT/ aaacACATCCGCGTAAATGTCTATc, u e s 200 for CPSF6g2 were fwd/ rev caccgCATCCGCGTAAATGTCTATG/ t 201 aaacCATAGACATTTACGCGGATGc, for CPSF6g3 were fwd/ rev 202 caccgGGACCACATAGACATTTACG/ aaacCGTAAATGTCTATGTGGTCCc, 203 for CPSF6g4 were fwd/ rev caccgTCCATGTAATCTCGGTCTTC/ 204 aaacGAAGACCGAGATTACATGGAc, for CypAg1 were fwd/rev 205 caccgGCCCGACCTCAAAGGAGACG/ aaacCGTCTCCTTTGAGGTCGGGCc. 206 7 207 Viral vector and HIV-1 production 208 293 T cells grown in 10 cm plates were transfected at a confluence of ~70 to 209 80% with 4.5 μg viral vector plasmid, 3 μg pCMVΔR8.91 and 3 μg pMD.G 210 using 4 μg polyethylenimine (PEI) per μg DNA in 1 ml of OptiMEM (Gibco) per 211 plate. For VSV-G pseudotyped full-length HIV-1 production 8 µg HIV-1 GFP 212 reporter plasmid and 2 µg pMD.G were co-transfected per plate. The medium 213 was changed 24 h post transfection, the viruses were harvested at 48 h and D 214 72 h post transfection, passed through a 0.45 μm filter and collections were o w n 215 pooled. Depending on the experiment, viral supernatants were subjected to lo a 216 sucrose purification as described before (70). d e 217 d f r 218 Generation of CRISPR/Cas9 THP-1 cell clones o m 219 THP-1 cells were transduced with VSV-G pseudotyped HIV-1 LV delivering h t 220 plentiCRISPRv2 at an estimated multiplicity of infection (MOI) of 1. tp : / 221 Transduced cell populations were selected with 1 µg/ml puromycin for two /jv 222 weeks. Single-cell clones were generated by limiting dilution and grown in 96- i.a s m 223 well plates for at least two weeks in the absence of puromycin. MX2 gene . o 224 disruption was validated by PCR amplification of the targeted genomic region r g / 225 using oligonucleotides fwd/rev AGCAAAGGAACATTGAGACTCTACTG/ o n 226 TTATTGTGGTGGGCTTACATGACAGC. M a 227 r c h 228 Infections 2 229 5 x 105 THP-1 or CD4+ T cells were plated in 100 µl media per well in 96-well 8 , 2 230 plates and treated for 24 h with IFNα. Then 100 µl supernatant containing 0 1 231 VSV-G pseudotyped GFP-reporter lentiviral vectors or NL4.3GFP-reporter 9 b 232 virus were added. To compare different CA mutants with wild type GFP y g 233 reporter vector/ virus we normalized viral input by 293T infectious titers or by u e s 234 units of reverse transcriptase in the supernatant as determined by a SYBR t 235 Green PCR-enhanced reverse transcriptase assay (SG-PERT) previously 236 described (75). Cells were fixed in 4% paraformaldehyde (PFA) two to three 237 days later, infectivities were determined from the percentage of GFP+ cells by 238 flow cytometry using a FACSVerse (BD Biosciences) and infectious titers 239 were determined on at least three different virus doses. Average infectious 240 titers were calculated with standard deviations as error bars. Experiments with 8 241 MDMs were performed in 48-well plates seeding 5 x 105 monocytes per well 242 prior to differentiation. For analysis by flow cytometry MDMs were trypsinized 243 for at least 30 minutes, resuspended and fixed in 4% PFA. Virus titrations 244 were usually performed by 3-fold serial dilutions of the viral supernatant. In 245 case of drug titration experiments (IFNα, or Cs) a single dose of supernatant 246 containing reporter virus was used aiming for a MOI ≤ 1. For experiments 247 using Cs or SDZ-NIM811, drugs were added at the time of infection with the D 248 reporter virus. o w n 249 lo a 250 Immunoblotting and antibodies d e 251 Proteins were separated in Mini-PROTEAN® TGX Stain-Free™ Precast Gels d f 252 (Biorad) at 120 V for 1 h and subsequently subjected to UV-activation and ro m 253 quantitative gel-imaging to compare amounts of loaded protein. Proteins were h t 254 transferred to nitrocellulose membranes using Trans-Blot® Turbo™ Transfer tp : / 255 System (Biorad). Primary antibodies used were: rabbit anti-MX1 (1:1.000, /jv 256 Proteintech), rabbit anti-MX2 (1:1.000, Novusbio), rabbit anti-CypA SA296 i.a s m 257 (1:3.000; Biomol), mouse anti-alpha-Tubulin (1:3.000; Sigma Aldrich), rabbit . o 258 anti-CPSF6 (1:3.000; Abcam), rabbit anti-MAPK (Erk1/2) (1:1.000; Cell r g / 259 Signaling). Goat anti-mouse or anti-rabbit IgG secondary antibodies were o n 260 coupled to horseradish peroxidase (Cell Signaling Technology) and proteins M a 261 were detected using Pierce ECL plus western blotting substrate (Thermo r c h 262 Scientific). 2 8 263 , 2 264 Drugs 0 1 265 SDZ-NIM811 was a kind gift from R. Bartenschlager (Heidelberg University 9 b 266 Hospital, Germany). Cyclosporin (Sandoz) was diluted in DMSO to a stock of y g 267 1.0 mM. IFNα2 (Roferon-A, IFNα-2a) stock concentration was 6.0 x 106 IU/ ml u e s 268 and was used at indicated concentrations. t 269 270 Taqman qPCR 271 THP-1 cells were seeded at 1 x 106 cells per well in 6-well plates, treated or 272 not with 500 IU/ml IFNα for 24 h and infected the next day with the same dose 273 of wild type or CA mutant viruses. Total DNA was isolated from the cells 4 or 274 24 h post infection using the QiaAmp extraction kit (Qiagen). Taqman qPCR 9 275 was performed using GFP fwd/ rev primers 276 CAACAGCCACAACGTCTATATCAT/ ATGTTGTGGCGGATCTTGAAG and 277 probe FAM-CCGACAAGCAGAAGAACGGCATCAA-TAMRA, 2-LTR circle 278 fwd/ rev primers AACTAGAGATCCCTCAGACCCTTTT/ 279 CTTGTCTTCGTTGGGAGTGAATT and probe FAM- 280 CTAGAGATTTTCCACACTGAC-TAMRA (76). Samples were normalized 281 either to total DNA concentration or by Taqman qPCR for GAPDH with fwd/ D 282 rev primers GGCTGAGAACGGGAAGCTT/ AGGGATCTCGCTCCTGGAA o w n 283 and probe FAM-TCATCAATGGAAATCCCATCACCA-TAMRA. Taqman lo a 284 qPCRs were performed using the Applied Biosystems 7500 Real Time PCR d e 285 system (Applied Biosystems). d f 286 ro m 287 Results h t 288 HIV-1 CA mutants N74D or P90A display enhanced sensitivity to IFNα- tp : / 289 induced blocks /jv 290 Overexpression of the type I IFN-induced protein MX2 blocks HIV-1 infection i.a s m 291 in a CA-sensitive manner, and CA amino acid substitutions N74D or P90A . o 292 reduce the sensitivity of HIV-1 to ectopic MX2-mediated repression (32, 33). r g / 293 We reasoned that N74D or P90A changes in HIV-1 CA might therefore reduce o n 294 the sensitivity of HIV-1 to IFNα-induced post-entry blocks. To test this M a 295 hypothesis, we treated the myeloid cell line THP-1 with increasing amounts of r c h 296 IFNα and challenged with equal doses of wild type VSV-G pseudotyped HIV-1 2 8 297 GFP lentiviral vector (LV) or viruses with the CA mutants N74D or P90A, as , 2 298 judged by their infectious titers on 293T cells, and determined the 0 1 299 percentages of infected cells two days later. In untreated THP-1 cells, no 9 b 300 substantial differences in infectious titers were detected between wild type or y g 301 CA mutants (Fig.1A, left panel). Likewise, VSV-G pseudotyped full length u e s 302 NL4.3GFP reporter virus bearing wild type or N74D or P90A mutant CA, t 303 showed similar titers in THP-1 cells (Fig.1B, left panel). In contrast, CA 304 mutants N74D or P90A were suppressed up to 10-fold more efficiently by 305 pretreatment with different concentrations of IFNα (Fig.1A and B, right panel). 306 307 We next examined these effects in primary monocyte-derived macrophages 308 (MDMs) or IL-2/PHA-activated CD4+ T cells. We confirmed that CA mutants 10