bioRxiv preprint doi: https://doi.org/10.1101/313312; this version posted May 2, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

1 A Rev-CBP80-eIF4AI complex drives Gag synthesis from the HIV-1 unspliced mRNA 2 3 Daniela Toro-Ascuy1#; Bárbara Rojas-Araya1#; Francisco García-de-Gracia1#; Cecilia 4 Rojas-Fuentes1; Camila Pereira-Montecinos1; Aracelly Gaete-Argel1; Fernando Valiente- 5 Echeverría1; Théophile Ohlmann2,3 and Ricardo Soto-Rifo1* 6 7 1 Laboratory of Molecular and Cellular Virology, Virology Program, Institute of 8 Biomedical Sciences, Universidad de Chile Faculty of Medicine, Independencia 834100, 9 Santiago, Chile 10 2 INSERM U1111, CIRI, Lyon, F-69364 France. 11 3 Ecole Normale Supérieure de Lyon, Lyon, F-69364 France. 12 13 14 * To whom correspondence should be addressed. 15 RSR: Tel: (56) 2 978 68 69; Fax: (56) 2 978 61 24; Email: [email protected] 16 17 # The authors wish it to be known that, in their opinion, the first three authors should be 18 regarded as joint First Authors 19 20 21 Running title: Role of Rev in HIV-1 gene expression 22 23 24 Keywords: HIV-1 unspliced mRNA, Rev, CBP80, eIF4AI, gene expression 25 26 27 28 29 30 31 bioRxiv preprint doi: https://doi.org/10.1101/313312; this version posted May 2, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

32 33 Abstract 34 Gag synthesis from the full-length unspliced mRNA is critical for the production of the 35 viral progeny during human immunodeficiency virus type-1 (HIV-1) replication. While 36 most spliced mRNAs follow the canonical gene expression pathway in which the 37 recruitment of the nuclear cap-binding complex (CBC) and the exon junction complex 38 (EJC) largely stimulates the rates of nuclear export and translation, the unspliced mRNA 39 relies on the viral protein Rev to reach the cytoplasm and recruit the host translational 40 machinery. Here, we confirm that Rev ensures high levels of Gag synthesis by driving 41 nuclear export and translation of the unspliced mRNA. These functions of Rev are 42 supported by the CBC subunit CBP80, which binds Rev and the unspliced mRNA in the 43 nucleus and the cytoplasm. We also demonstrate that Rev interacts with the DEAD-box 44 RNA helicase eIF4AI, which translocates to the nucleus and cooperates with Rev to 45 promote Gag synthesis. Interestingly, molecular docking analyses revealed the assembly of 46 a Rev-CBP80-eIF4AI complex that is organized around the Rev response element (RRE). 47 Together, our results provide further evidence towards the understanding of the molecular 48 mechanisms by which Rev drives Gag synthesis from the unspliced mRNA during HIV-1 49 replication. 50 51 52 53 54 55 56 57 58 59 60 61 62 bioRxiv preprint doi: https://doi.org/10.1101/313312; this version posted May 2, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

63 Introduction 64 Human Immunodeficiency Virus type-1 (HIV-1) gene expression is a complex process that 65 leads to the synthesis of fifteen proteins from one single primary transcript (1,2). Once the 66 proviral DNA has been integrated into the host cell genome, the RNA polymerase II drives 67 the synthesis of a 9-kb, capped and polyadenylated pre-mRNA that undergoes alternative 68 splicing generating more than 100 different transcripts classified into three main 69 populations (3,4). The so-called 2-kb multiply spliced transcripts code for the key 70 regulatory proteins Tat and Rev and the accessory protein Nef and are the dominant viral 71 mRNA species at early stages of viral gene expression (1,2,5). Unlike cellular mRNAs or 72 the 2-kb transcripts, which are spliced to completion before they exit the nucleus, HIV-1 73 and other complex retroviruses produce an important fraction of viral transcripts that 74 remain incompletely spliced (2,6). These 4-kb transcripts are expressed during the 75 intermediate phase of gene expression and are used for the synthesis of the envelope 76 glycoprotein (Env) and the accessory proteins Vif, Vpr and Vpu (2,6). Finally, the full- 77 length 9-kb pre-mRNA in its unspliced form also reaches the cytoplasm to be used as an 78 mRNA template during the late stages of viral gene expression for the synthesis of the 79 major structural proteins Gag and Gag-Pol (1,2,6). 80 Gene expression in eukaryotic cells occurs through the intricate connection of different 81 processes including transcription, splicing, nuclear export, translation and mRNA decay 82 and is regulated by the specific recruitment of nuclear proteins that together form the 83 messenger ribonucleoprotein (mRNP) complex (7-9). As such, the early binding of the 84 nuclear cap-binding complex (CBC) to the 5´-end cap structure and the splicing-dependent 85 recruitment of nuclear proteins such as the exon-junction complex (EJC) onto the mRNA 86 has been shown to increase the rates of nuclear export and translation of spliced transcripts 87 (10-20). Eukaryotic cells have also evolved quality control mechanisms ensuring that only 88 properly processed mRNAs reach the cytoplasm and are decoded by the translational 89 machinery. These mechanisms include the EJC-dependent degradation of transcripts 90 containing premature stop codons through nonsense-mediated decay (NMD) or the NXF1- 91 dependent nuclear retention of unspliced transcripts mediated by the nucleoporin Tpr (21- 92 25). Consistent with these cellular quality control mechanisms, it has been widely reported 93 that viral intron-containing transcripts including the 4-kb and the 9-kb mRNA produced bioRxiv preprint doi: https://doi.org/10.1101/313312; this version posted May 2, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

94 during HIV-1 replication are retained and degraded in the host cell nucleus unless the viral 95 protein Rev is present (26-30). As such, Rev has been proposed to promote HIV-1 gene 96 expression from its target transcripts by i) avoiding mRNA degradation (26,31); ii) 97 promoting nuclear export (31-33) or by iii) promoting translation (34,35). In this study, we 98 developed Rev mutant proviruses and confirmed that Rev is required for both nuclear 99 export and translation of the HIV-1 unspliced mRNA. Interestingly, we show that the 100 nuclear cap-binding complex subunit CBP80 and the translation initiation factor eIF4AI 101 associate with Rev and the unspliced mRNA to promote Gag synthesis. Molecular docking 102 analyses suggest the assembly of a Rev-CBP80-eIF4AI complex that is reorganized when 103 Rev binds to the RRE. Together, our work provides further insights into the molecular 104 mechanism by which Rev drives Gag synthesis from the HIV-1 unspliced mRNA. 105 106 Materials and methods 107 DNA constructs: The pNL4.3 and pNL4.3R proviruses were previously described (36,37). 108 These vectors were digested with NheI and subjected to a 20 min polymerization reaction at 109 72 ºC using the Phusion® High-Fidelity DNA polymerase (New England Biolabs) in order 110 to create a frameshift that generates a premature stop codon within the env gene. The 111 resulting vectors were ligated with the T4 DNA ligase and transformed into E.coli DH5α. 112 To create the pNL4.3-ΔRev and pNL4.3R-ΔRev vectors, the above vectors were digested 113 with BamHI and subjected to the same polymerization/ligation reaction to create a 114 frameshift within the rev gene previously shown to abolish expression of a functional 115 protein (28). The pCMV-NL4.3R and pCMV-NL4.3R-ΔRev vectors were obtained by 116 replacing the FspAI/BssHII fragment of the corresponding vector by the CMV IE promoter 117 amplified from the pCIneo vector (Promega) as we previously reported (38). The pCDNA- 118 Flag-Rev vector was previously described (22). pCDNA-d2EGFP vector was generated by 119 inserting the d2EGFP ORF into pCDNA3.1 (Life Technologies). pCDNA HIV-1 5`-UTR 120 and pCDNA β-globin 5`-UTR were previously described (39). The dl HIV-1 IRES vector 121 was previously described (40). The pCIneo-HA-eIF4GI, -eIF4A and -eIF4E were 122 previously described (41). The pCMV-myc-eIF4E and CBP80 were previously described 123 (42). 124 bioRxiv preprint doi: https://doi.org/10.1101/313312; this version posted May 2, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

125 Cell culture and DNA transfection: HeLa cells and human microglia (C20 cells)(43) were 126 maintained in DMEM (Life Technologies) supplemented with 10% FBS (Hyclone) and

127 antibiotics (Hyclone) at 37 ºC and a 5% CO2 atmosphere. H9 T-lymphocytes (44) and

128 THP-1ATCC monocytes (45) were maintained in RPMI 1640 (Life Technologies)

129 supplemented with 10% FBS (Hyclone) and antibiotics (Hyclone) at 37 ºC and a 5% CO2 130 atmosphere. Cells were transfected using linear PEI ~25.000 Da (Polysciences) prepared as 131 described previously (46). Cells were transfected using a ratio µg DNA/µl PEI of 1/15. 132 133 Analysis of Renilla and firefly activities: Renilla activity was determined using the Renilla 134 Reporter Assay System (Promega) and Renilla/firefly activities were determined using the 135 Dual Luciferase Reporter Assay System (Promega) in a GloMax® 96 microplate 136 luminometer (Promega). 137 138 Western blot: Cells extracts from transfected cells were prepared by lysis with RIPA buffer 139 and 20 µg of total protein were subjected to 10% SDS-PAGE and transferred to an 140 Amersham Hybond™-P membrane (GE Healthcare). Membranes were incubated with an 141 HIV-1 p24 monoclonal antibody diluted to 1/1000 (47), a rabbit anti-Flag antibody (Sigma- 142 Aldrich) diluted to 1/1000 or and HRP-conjugated anti-actin antibody (Santa Cruz 143 Biotechnologies) diluted to 1/750. Upon incubation with the corresponding HRP- 144 conjugated secondary antibody (Santa Cruz Biotechnologies) diluted to 1/1000, membranes 145 were revealed with the ECL substrate (Cyanagene) using a C-Digit digital scanner (Li-Cor).

146 RNA extraction and RT-qPCR: Cytoplasmic RNA extraction and RT-qPCR from 147 cytoplasmic RNA were performed essentially as recently described (48). Briefly, cells were 148 washed intensively with PBS, recovered with PBS-EDTA 10 mM and lysed for 1-2 min at 149 room temperature with 200 µl of buffer [10 mM Tris-HCl pH=8.0, 10 mM NaCl, 3 mM 150 MgCl2, 1 mM DTT, 0.5% NP40 and 2 mM of vanadyl-ribonucleoside complex (VRC) 151 (New England Biolabs)]. Cell lysates were centrifuged at 5000 rpm for 5 min at 4 ºC and 152 supernatant containing the cytoplasmic fraction were recovered and RNA extraction was 153 carried out by adding 1 ml of TRIzol™ (Thermo Fisher) as indicated by the manufacturer. 154 Cytoplasmic RNAs (1 µg) were reverse-transcribed using the High Capacity RNA-to- bioRxiv preprint doi: https://doi.org/10.1101/313312; this version posted May 2, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

155 cDNA Master Mix (Life Technologies). For quantitative PCR, a 20 µl reaction mix was 156 prepared with 5 µl of template cDNAs (previously diluted to 1/10), 10 µl of FastStart 157 Universal SYBR Green Master (Rox) (Roche), 0.2 µM of sense and antisense primers and 158 subjected to amplification using the Rotorgen fluorescence thermocycler (Qiagen). The 159 GAPDH housekeeping gene was amplified in parallel to serve as a control reference. 160 Relative copy numbers of Renilla luciferase cDNAs were compared to GAPDH or 18S - Ct 161 rRNA using x Δ (where x correspond to the experimentally calculated amplification 162 efficiency of each primer couple).

163 164 Fluorescent in situ hybridization, immunofluorescence and confocal microscopy: RNA 165 FISH was carried out as we recently described (48). Briefly, HeLa cells were cultured in a 166 12 well plate with covers slide (Nunc™) and maintained and transfected with 1 µg of 167 pNL4.3 or 1 µg of the corresponding HA vectors as indicated above. At 24 hpt, cells were 168 washed twice with 1X PBS and fixed for 10 min at room temperature with 4% 169 paraformaldehyde. Cells were subsequently permeabilized for 5 min at room temperature 170 with 0.2% Triton X-100 and hybridized overnight at 37° C in 200 µl of hybridization mix 171 (10% dextran sulfate, 2 mM VRC, 0.02% RNase-free BSA, 50% formamide, 300 µg tRNA 172 and 120 ng of 11-digoxigenin-UTP probes) in a humid chamber. Cells were washed with 173 0.2X SSC/50% formamide during 30 min at 50° C and then incubated three times with 174 antibody dilution buffer (2X SSC, 8% formamide, 2 mM vanadyl-ribonucleoside complex, 175 0.02% RNase-free BSA). Mouse anti-digoxin and rabbit anti-HA (Sigma Aldrich) primary 176 antibodies diluted to 1/100 in antibody dilution buffer were added for 2 h at room 177 temperature. After three washes with antibody dilution buffer, cells were incubated for 90 178 min at room temperature with anti-mouse Alexa 488 and anti-rabbit Alexa 565 antibodies 179 (Molecular Probes) diluted at 1/1000. Cells were washed three times in wash buffer (2X 180 SSC, 8% formamide, 2 mM vanadyl-ribonucleoside complex), twice with 1X PBS, 181 incubated with Hoescht (Life Technologies) diluted to 1/10000, for 5 min at room 182 temperature, washed three times with 1X PBS, three times with water and mounted with 183 Fluoromount (Life Technologies). Images were obtained with a TCS SP8 Confocal 184 Microscope (Leica Microsystems) and images were processed using FIJI/ImageJ (NIH). 185 bioRxiv preprint doi: https://doi.org/10.1101/313312; this version posted May 2, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

186 Proximity ligation assay (PLA): PLA (49), was carried out using the DUOLINK II In Situ 187 kit (Sigma-Aldrich) and PLA probe anti-mouse minus and PLA probe anti-rabbit plus 188 (Sigma-Aldrich) as we have previously described (38,50). Briefly, PFA-fixed HeLa cells 189 were pre-incubated with blocking agent for 30 min at room temperature. Primary antibodies 190 were added at a dilution of 1:100 (mouse anti-HA, Santa Cruz Biotechnologies) and 1:200 191 (rabbit anti-Flag, Sigma-Aldrich) in 40 µl DUOLINK antibody diluent and incubated at 192 37°C for 1 h. Samples were washed three times with PBS for 5 min each and secondary 193 antibodies (DUOLINK anti-rabbit PLA-plus probe and DUOLINK anti-mouse PLA-minus 194 probe) were added and incubated at 37°C for 1 h. Ligation and amplification reactions were 195 performed following the same protocol described in (50). Samples were incubated with 196 DAPI (0.3 µg/ml in PBS) (Life Technologies) for 1 min at room temperature, washed three 197 times with PBS, three times with water and mounted with fluoromount (Sigma Aldrich). 198 Images were obtained with a TCS SP8 Confocal Microscope (Leica Microsystems) and 199 images were processed using FIJI/ImageJ (NIH). 200 201 In situ hybridization coupled to PLA (ISH-PLA): The ISH-PLA protocol was developed 202 by mixing the RNA-FISH and PLA protocols described above. Briefly, PFA-fixed Hela 203 cells growing on coverslips were permeabilized for 5 min at RT with 0.2% Triton X-100 204 and hybridized overnight at 37°C in 200 µl of hybridization mix (10% dextran sulphate, 205 2mM vanadyl–ribonucleoside complex, 0.02% RNase-free bovine serum albumin, 50% 206 formamide, 300 mg of tRNA and 120 ng of 11-digoxigenin-UTP probes) in a humid 207 chamber. Cells were washed with 0.2xSSC/50% formamide during 30 min at 50°C and 208 then incubated with blocking agent for 30 min to room temperature. Then incubated three 209 times with antibody dilution buffer (2xSSC, 8%formamide, 2mM vanadyl–ribonucleoside 210 complex and 0.02% RNase-free bovine serum albumin). Mouse anti-digoxin and rabbit 211 anti-protein of interest primary antibodies diluted to 1/100 in antibody dilution buffer were 212 added for 2 h at room temperature. After three washes with antibody dilution buffer and 213 two washes with PBS for 5 min each, then secondary antibodies (DUOLINK anti-rabbit 214 PLA-plus probe, DUOLINK anti-mouse PLA-minus probe) were added and incubated at 215 37°C for 1 h. Then, the ligation and amplification reaction were performed following the 216 same protocol described above. Thereafter, the covers were incubated with a solution of bioRxiv preprint doi: https://doi.org/10.1101/313312; this version posted May 2, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

217 DAPI (Life Technologies) (0.3 µg/ml in PBS) for 1 min at room temperature, washed three 218 times with PBS, three times with water and mounted with fluoromount aqueous mounting 219 medium (Sigma Aldrich). Images were obtained with a TCS SP8 Confocal Microscope 220 (Leica Microsystems) and the Images were processed using FIJI/ImageJ (NIH) 221 (Supplementary Fig. 2A). 222 223 Bioinformatic analyses: Molecular docking was performed using previously described 224 structures of CBP80 (PDB: 3FEY) (51), eIF4AI (PDB: 2ZU6) (52), Rev monomer (PDB: 225 2X7L) (53), Rev dimer (PDB: 3LPH) (54) and Rev dimer bound to the RRE (PDB: 226 4PMI)(55). Since the structure of the Rev monomer contains the L12S/L60R mutations 227 (54), we proceed with a structural analysis between the dimer and the monomer using the 228 SALIGN module of MODELLER, version 9.13 (56). Structure visualization was performed 229 with Visual Molecular Dynamics (VMD) (57). The local quality of the wild type Rev dimer 230 was estimated with ANOLEA (Anolea, atomic mean force potential) (58) for energy 231 evaluation and with PROCHECK for stereochemistry evaluation (59). Subsequently, an 232 energy minimization of 5000 steps was performed using the Steepest Descent (SD) method. 233 For protein-protein interactions, the crystallized structures were treated by the SD method 234 described above prior processing and refinement of the model. Macromolecular complexes 235 were obtained with PATCHDOCK (60), which analyses separately the surface of both 236 proteins and generates geometric patterns depending on the shape complementary of soft 237 molecular surfaces in order to generate the best starting candidate solution. Refinement of 238 the side-chain flexibility, rigid body optimization, scoring of docking candidates and 239 ranking were performed with FIREDOCK (61). Macromolecular complexes were selected 240 based on their energetic characteristics and the non-covalent interactions were determined 241 with the Discovery Studio Visualizer software. 242 243 Results 244 Rev promotes nuclear export and translation of the unspliced mRNA 245 The viral protein Rev promotes gene expression from the unspliced transcript by acting at 246 the post-transcriptional level but the precise mechanism by which this occurs remains 247 unclear (62-64). This prompted us to conduct a study aimed to gain further insights into the bioRxiv preprint doi: https://doi.org/10.1101/313312; this version posted May 2, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

248 function of Rev on HIV-1 gene expression during viral replication. Thus, we used the 249 pNL4.3 proviral DNA to introduce a frameshift within the rev gene previously shown to 250 abolish the expression of a functional protein (28). Consistent with a critical role of Rev in 251 the post-transcriptional regulation of the unspliced mRNA, we observed that Gag synthesis 252 was abolished in the absence of Rev and restored upon Rev expression in trans (Fig. 1A). 253 In agreement with several previous reports (31-33), we observed that most of the unspliced 254 mRNA is retained in the nucleus in the absence of Rev (Fig. 1B, compare wild type and 255 ΔRev). The cytoplasmic signal of the unspliced mRNA was recovered when Rev was 256 expressed in trans (Fig. 1B, see ΔRev + Flag-Rev), which is consistent with an important 257 role of Rev in nuclear export. However, it has been reported that Rev is also important for 258 translation of the unspliced mRNA during viral replication (34,35). Thus, in order to 259 quantify the impact of Rev on gene expression from the unspliced mRNA we used the 260 pNL4.3R reporter provirus to generate a ΔRev version as described above (see materials 261 and methods). Transfection of pNL4.3R-ΔRev in HeLa cells, T-cells (H9 cells), monocytes 262 (THP-1 cells) or human microglia (C20 cells) resulted in very low levels of Gag-Renilla 263 expression when compared to the wild type provirus indicating that our reporter proviruses 264 can be used to quantify the effects of Rev on gene expression from the unspliced mRNA 265 (Supplementary Fig. 1A). Transfection of the pNL4.3R-ΔRev together with a Flag-Rev 266 expressing vector restored Gag synthesis to the wild type levels indicating that the defects 267 in Gag-Renilla expression observed were exclusively due to the absence of Rev 268 (Supplementary Fig. 1B). Thus, we used the pNL4.3R-wt and -ΔRev proviruses to quantify 269 the effects of Rev on Gag synthesis, cytoplasmic levels of the unspliced mRNA and its 270 translational efficiency in HeLa cells as we have previously reported (37,38,41,48,65,66). 271 As observed above, Gag production was almost abolished in the absence of Rev (Fig. 1C, 272 left panel). Consistent with its previously described role in nuclear export of the unspliced 273 transcript (26,31-33), we observed that the cytoplasmic levels of the unspliced mRNA were 274 reduced by 3-fold in the absence of Rev (Fig. 1C, middle panel). Although these results 275 differ from those presented in Fig. 1B in which no genomic RNA could be detected in the 276 cytoplasm, it has been proposed that an important fraction of the unspliced transcript reach 277 the cytoplasm in the absence of Rev but remains trapped into a ribonucleoprotein complex 278 inaccessible to the probes used during in situ hybridization (31). It should be mentioned bioRxiv preprint doi: https://doi.org/10.1101/313312; this version posted May 2, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

279 that our cytoplasmic fractions were devoid of pre-GADPH mRNA discarding any 280 contamination with nuclear RNA (Supplementary Fig. 1C). However, this important 281 decrease in the cytoplasmic levels of the unspliced transcript does not account for the 282 dramatic reduction in Gag synthesis (>100-fold) indicating that most of the unspliced 283 mRNAs that reach the cytoplasm in the absence of Rev are not translated (Fig. 1C, right 284 panel). Together, these data confirm that the function of Rev during viral replication is not 285 restricted to nuclear export since it is also critical for translation. 286 While the role of Rev in nuclear export has been largely characterized (62), the molecular 287 mechanism by which Rev promotes translation of the unspliced transcript is not very well 288 understood (67). Interestingly, it was shown that Rev was able to promote translation of a 289 reporter RNA by an unknown mechanism involving the binding to an RNA motif (A-loop) 290 present within SL1 of the unspliced mRNA 5´-UTR (63,68). Consistent with this previous 291 work, we observed that expression of a Renilla luciferase-based monocistronic vector 292 harboring the 5´-UTR of the unspliced transcript (but not that of the human β-globin) was 293 stimulated up to 2-fold in the presence of Rev (Supplementary Fig. 1D). Such a stimulation 294 was not observed when IRES-driven translation was analyzed with a previously described 295 bicistronic vector (40), suggesting that the effect of Rev on the 5´-UTR is exerted at the 296 level of cap-dependent translation (Supplementary Fig. 1E). 297 Although the 2-fold stimulation observed with the reporter construct is consistent with the 298 previous report (63), it does not account for the strong dependence of Rev for unspliced 299 mRNA translation observed in the context of a full-length provirus (Fig. 1A and 1C), 300 suggesting that the 5´-UTR is not the only molecular determinant involved in Rev-mediated 301 translation of the unspliced mRNA. In order to confirm this hypothesis, we constructed a 302 ΔRev version of the CMV-pNL4.3R vector, a reporter provirus lacking the 5´-LTR and 303 most of the 5´-UTR but containing the CMV IE promoter (38). This proviral DNA is 304 expected to produce an unspliced mRNA that only contains the last 79 nucleotides of the 305 wild type 5´-UTR, therefore lacking the Rev binding site previously described in the A- 306 loop of SL1. We transfected the CMV-pNL4.3R and CMV-pNL4.3R-ΔRev vectors and 307 analyzed the role of Rev in Gag synthesis, cytoplasmic levels of the unspliced mRNA and 308 translation as described above. Interestingly, we observed a strong dependence for Rev in 309 translational efficiency of the unspliced transcript regardless of whether the entire 5´-UTR bioRxiv preprint doi: https://doi.org/10.1101/313312; this version posted May 2, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

310 was driving ribosome recruitment or not (Fig. 1D). Together, these data suggest that the 311 previously described Rev-binding site present within the 5´-UTR is not the major molecular 312 determinant involved in the translational stimulation mediated by Rev in the context of 313 viral replication. Given the fact that our CMV-pNL4.3R provirus also lacks major 314 determinants required for IRES-driven translation (40), these data also suggest that Rev 315 promotes cap-dependent translation. 316 317 The CBC subunit CBP80 interacts with Rev and promotes nuclear export and 318 translation of the unspliced mRNA 319 From data presented above, it seems that Rev promotes cap-dependent translation of the 320 unspliced mRNA. Interestingly, it was proposed that cap-dependent CBC-driven translation 321 could ensure Gag synthesis during an HIV-induced inhibition of eIF4E activity (69). In 322 agreement with this idea, we have previously shown that the cytoplasmic cap-binding 323 protein eIF4E is excluded from a translation initiation mRNP containing the HIV-1 324 unspliced mRNA together with the RNA helicase DDX3 and translation initiation factors 325 eIF4GI and PABPC1 (48). Thus, in order to determine whether the unspliced mRNA is 326 associated to the CBC or eIF4E, we developed a protocol based on in situ hybridization of 327 digoxin-labeled probes directed to the unspliced mRNA coupled to the proximity ligation 328 assay (ISH-PLA) in order to determine and quantify unspliced mRNA-protein interactions 329 (Supplementary Fig. 2A and materials and methods). By using our ISH-PLA protocol, we 330 observed that the unspliced mRNA preferentially associates with the CBC subunit CBP80 331 rather than eIF4E (Fig. 2A). Interestingly, despite the fact that most of the CBP80 signal 332 was observed in the nucleus in RNA FISH-IF experiments performed in parallel 333 (Supplementary Fig. 2B), we observed that the interaction between the unspliced mRNA 334 and the CBC subunit occurs predominantly in the cytoplasm suggesting that the unspliced 335 mRNA remains associated to the CBC upon nuclear export. Signal intensity quantifications 336 from the RNA FISH experiments performed in parallel revealed no differences in myc- 337 tagged protein expression (Supplementary Fig. 2C). To further investigate the role of 338 CBP80 and eIF4E on gene expression from the unspliced mRNA, we independently 339 overexpressed both proteins in HeLa cells and analyzed Gag synthesis, cytoplasmic 340 unspliced mRNA levels and translational efficiency as described above. Consistent with a bioRxiv preprint doi: https://doi.org/10.1101/313312; this version posted May 2, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

341 preferential association of the unspliced mRNA with CBP80, we observed that 342 overexpression of the CBC subunit but not eIF4E results in a marked increase in Gag 343 synthesis, which was due to an increase in both cytoplasmic accumulation and translation 344 of the unspliced mRNA (Fig. 2B). Interestingly, CBP80 overexpression only resulted in 345 marginal (2-fold) stimulation of the unspliced mRNA in the cytoplasm when the ΔRev 346 provirus was used suggesting that CBP80 cooperates with Rev during the post- 347 transcriptional control of the unspliced mRNA (Fig. 2C). Of note, Gag-Renilla activity 348 from the wild type provirus was largely higher than that observed from the ΔRev provirus 349 consistent with data presented in Fig. 1 (data not shown). Consistent with the dependence 350 on Rev for CBP80 function, we observed that CBP80 overexpression has marginal effects 351 on protein synthesis from the Rev-independent nef mRNA (Supplementary Fig. 2D). 352 Since it was previously shown that CBP80 interacts with Rev in vitro (70), we wanted to 353 evaluate whether this interaction also occurs in cells. Thus, we performed PLA and 354 observed that Flag-tagged Rev interacts with both endogenous and V5-tagged CBP80 355 (Supplementary Fig. 2E and Fig. 2D). 356 Together, these results suggest that the unspliced mRNA is preferentially associated to the 357 CBC and the CBC subunit CBP80 interacts and cooperates with Rev to promote nuclear 358 export and translation of this viral transcript. 359 360 DEAD-box helicase eIF4AI interacts with Rev and promotes Gag synthesis from the 361 unspliced mRNA 362 Having determined that the unspliced mRNA is preferentially associated with CBP80 and 363 that this CBC subunit interacts with Rev, we were interested in identify additional 364 translation initiation factors interacting with Rev that could be involved in Gag synthesis. 365 Although the CBP20/80-dependent translation initiation factor (CTIF) was shown to be 366 important for CBC-dependent translation (71), we observed that CTIF is rather a potent 367 inhibitor of Gag synthesis (García-de-Gracia et al, manuscript in preparation). Thus, we 368 reasoned that Rev and CBP80 form an mRNP different from the canonical CBC, which is 369 important for unspliced mRNA nuclear export and translation. Despite we showed that 370 eIF4E seems not relevant for unspliced mRNA translation (Fig. 2A), we looked whether 371 additional components of eIF4F such as eIF4GI and eIF4AI were associated with Rev. bioRxiv preprint doi: https://doi.org/10.1101/313312; this version posted May 2, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

372 Indeed, eIF4GI was shown to interact with CBP80 and thus, we supposed that it could be 373 recruited to the Rev-CBP80 complex (72). Interestingly, our PLA using HA-tagged 374 versions of eIF4E, 4GI and 4AI together with Flag-tagged Rev revealed that Rev forms 375 nuclear and cytoplasmic complexes with eIF4AI and at a much lesser extent with eIF4E 376 and eIF4GI (Figs. 3A and 3B). It should be mentioned that IF experiments performed in 377 parallel revealed no differences in the intensity signals amongst HA-tagged eIFs and Flag- 378 Rev indicating that the increased number of interactions observed between eIF4AI and Rev 379 were not due to differences in the ectopic expression levels of the proteins (Supplementary 380 Fig. 3A). From these data, it could be speculated that Rev recruits eIF4AI to the unspliced 381 mRNA in order to promote translation. Thus, in order to evaluate the involvement of 382 eIF4AI in Rev activity, we overexpressed the RNA helicase and analyzed its impact on Gag 383 synthesis, cytoplasmic unspliced mRNA and translation using our wild type and ΔRev 384 reporter proviruses (Fig. 3C). As expected, Gag-Renilla activity from the wild type provirus 385 was much higher than that observed from the ΔRev provirus (data not shown). Interestingly, 386 we observed that eIF4AI overexpression results in a 2- and 5-fold increase in Gag synthesis 387 from the wild type and ΔRev reporter proviruses, respectively (Fig. 3C, left panel). 388 Surprisingly, analysis of the cytoplasmic unspliced mRNA levels upon eIF4AI 389 overexpression revealed a 2- to 3-fold increase for the wild type provirus with no effects on 390 the ΔRev provirus (Fig. 3C, middle panel). More strikingly, we observed that translation 391 from the ΔRev provirus was stimulated up to 6-fold by eIF4AI overexpression with no 392 effects on translation from the wild type provirus (Fig. 3C, right panel). These results 393 suggest that the presence of Rev will determine the process (nuclear export or translation) 394 by which ectopically expressed eIF4AI will promote Gag synthesis from the unspliced 395 mRNA. 396 397 Rev regulates the association of CBP80 and eIF4AI to the unspliced mRNA 398 Giving the fact that the presence of Rev modulates the activity of CBP80 and eIF4AI on 399 Gag synthesis, we wanted to determine whether Rev was involved in the recruitment of 400 these cellular proteins to the unspliced mRNA. For this, we quantified the interaction 401 between the unspliced mRNA and CBP80 or eIF4AI in the presence and absence of Rev 402 using our ISH-PLA protocol. Interestingly, we observed that the interactions between bioRxiv preprint doi: https://doi.org/10.1101/313312; this version posted May 2, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

403 CBP80 and the unspliced mRNA were reduced in the ΔRev provirus but were restored 404 upon expression of Rev in trans suggesting that Rev favors and/or stabilizes the association 405 between CBP80 and the unspliced mRNA (Fig. 4A, left panel). Interestingly, we observed 406 that Rev is necessary to maintain the CBP80-unspliced mRNA interaction in the cytoplasm 407 (Fig. 4A, compare middle and right panels). Signal intensity analysis from RNA FISH 408 experiments performed in parallel revealed no differences in CBP80 expression between 409 each condition (Supplementary Fig. 4A and 4B). 410 We also observed that the eIF4AI-unspliced mRNA interactions were reduced in the 411 absence of Rev and restored when the viral protein was expressed in trans (Fig. 4B, left 412 panel). Interestingly, we noticed that while most of the interactions between the unspliced 413 mRNA and eIF4AI in the cytoplasm are independent of Rev, the viral protein favors the 414 interaction in the nucleus (Fig. 4B, middle and right panels). Signal intensity analysis from 415 RNA FISH experiments performed in parallel revealed no differences in eIF4A expression 416 between each condition (Supplementary Fig. 4C and 4D). This observation is consistent 417 with our data presented in Fig. 3C in which ectopic expression of eIF4AI favors the 418 accumulation of the unspliced mRNA in the cytoplasm in the presence of Rev and 419 promotes translation in the ΔRev provirus. Interestingly, we observed an increase in the 420 nuclear signal of HA-eIF4AI with the wild type provirus that was absent with the ΔRev 421 provirus suggesting that a fraction of the RNA helicase might translocate to the nucleus in 422 the presence of Rev (Supplementary Fig. 4C and 4E). 423 Together, these results suggest that Rev regulates the association of CBP80 and eIF4AI 424 with the unspliced mRNA in the cytoplasm and the nucleus, respectively. 425 426 Assembly of a Rev-CBP80-eIF4A complex onto the Rev response element 427 From our data presented above, it appears that Rev regulates Gag synthesis by associating 428 with CBP80 and eIF4AI and regulating the association of these cellular proteins with the 429 unspliced mRNA in the nucleus and the cytoplasm. Although it was proposed that eIF4AI 430 could be part of the CBC-associated mRNP (72), to our knowledge, the interaction between 431 CBP80 and eIF4AI has never been formally demonstrated. Thus, we finally sought to 432 determine whether CBP80 and eIF4AI interact in cells and whether Rev was influencing 433 such an interaction. As a first approach, we performed PLA to identify and characterize the bioRxiv preprint doi: https://doi.org/10.1101/313312; this version posted May 2, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

434 CBP80-eIF4A interaction and observed that both proteins interact at the nuclear periphery 435 and at a lesser extent in the nucleus (Fig. 5A, left panel). Interestingly, our PLA analysis 436 showed increased interactions in the nucleus when Rev is present (Fig. 5A, right panel). 437 Dots per cell quantification revealed that Rev indeed stimulates the association between 438 CBP80 and eIF4AI (Fig. 5B). 439 Having determined that Rev interacts with CBP80 and eIF4AI and that both cellular 440 proteins also interact in cells, we then wanted to gain insights into the potential assembly of 441 a trimeric Rev-CBP80-eIF4AI complex. We performed molecular docking in order to study 442 the non-covalent interactions formed between CBP80 and eIF4AI using previously 443 described structures (51,52). We selected the complex presenting the best binding energy 444 and observed fifteen non-covalent interactions between both cellular proteins (Fig. 5C and 445 Supplementary Table 1). Interestingly, we observed that the eIF4AI binding site on CBP80 446 overlaps with the CBP20-binding site suggesting that eIF4AI is recruited into a complex 447 alternative to the canonical CBC (Supplementary Fig. 5A and see discussion). 448 We then used a previously described structure of a Rev dimer in the RNA-free state (73), to 449 investigate whether the viral protein was recruited to the CBP80-eIF4AI complex. 450 Consistent with the lack of non-covalent interactions detected between eIF4AI and the Rev 451 dimer (data not shown), we observed that Rev binds preferentially to CBP80 in the context 452 of our modeled CBP80-eIF4AI complex (Fig. 5D). We mostly observed electrostatic and 453 hydrophobic interactions between positively charged residues in one of the monomers of 454 Rev and negative residues in CBP80 (Supplementary Table 2). Interestingly, residues R38, 455 R39 and R44 in one of the Rev monomers and R44 in the second Rev monomer, which we 456 detected as involved in CBP80 binding (Fig. 5D and Supplementary Table 2), are also 457 involved in RNA binding suggesting that the Rev dimer would not be able to bind the RRE 458 in the context of the Rev-CBP80-eIF4AI complex. However, it was also reported that the 459 interface of the Rev dimer is reorganized upon RNA binding with crossing angles of 120º 460 in the RNA-free state and 50º in the RNA-bound state (55,73). Thus, we finally sought to 461 evaluate whether RNA binding by Rev has an impact in the assembly of the Rev-CBP80- 462 eIF4AI complex. For this, we performed docking analyses using the structure of the 463 Rev/RRE complex and our modeled CBP80-eIF4AI complex. Interestingly, we observed 464 that the presence of the RRE results in the complete reorganization of the trimeric complex bioRxiv preprint doi: https://doi.org/10.1101/313312; this version posted May 2, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

465 (Fig. 5E and Supplementary Table 3). As such, we observed that both CBP80 and eIF4AI 466 establish non-covalent interactions with the RRE. In addition, while the interactions 467 between Rev and CBP80 were lost, we observed that eIF4AI was able to perform contacts 468 with the viral protein in its RNA-bound state (Fig. 5E and Supplementary Table 3). 469 Taking together, these data suggest that Rev forms a complex with CBP80 and eIF4AI that 470 reorganizes when the viral protein binds to the RRE. This complex drives Gag synthesis 471 from the unspliced mRNA. 472 473 Discussion 474 Gag synthesis from the unspliced mRNA is a critical step during HIV-1 replication 475 necessary for the efficient production of the viral progeny. Indeed, stoichiometric studies 476 revealed that up to 5000 molecules of Gag are required to build one single viral particle 477 (74), indicating that the unspliced mRNA needs to be efficiently expressed. However, since 478 this viral mRNA contains functional introns, it must overcome surveillance mechanisms 479 that induce its nuclear retention and degradation before it reaches the translational 480 machinery in the cytoplasm to produce Gag (75). As such, understanding how the unspliced 481 mRNA is efficiently exported from the nucleus and translated in the cytoplasm is not only 482 critical to improve our knowledge on the molecular mechanisms driving viral gene 483 expression but is also important to identify new pathways and/or interactions occurring 484 within the cell or induced by the virus that could be targeted by novel antiretroviral drugs. 485 Although the unspliced mRNA associated to Rev, CRM1 and other co-factors does not 486 resemble to a canonical mRNP that must be directed to the ribosomes upon nuclear export 487 (i.e., an mRNA associated to cellular components such as TREX, NXF1 and the EJC), 488 translation of the unspliced mRNA is highly efficient in cells and nuclear export across the 489 Rev/CRM1 pathway is critical for ribosome recruitment (37,76). In this study, we provide 490 evidence that the viral protein Rev acts as a nuclear imprint critical for nuclear export and 491 translation of the unspliced mRNA (Fig. 1). We reasoned that recruitment of Rev might 492 serve as a platform for the spatiotemporal recruitment of host factors required to 493 interconnect nuclear export and translation of the unspliced mRNA. In order to study the 494 interaction between the unspliced mRNA and host proteins, we developed the ISH-PLA 495 strategy, which allowed us to identify and quantify unspliced mRNA-protein interactions bioRxiv preprint doi: https://doi.org/10.1101/313312; this version posted May 2, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

496 but also to determine the cellular location in which such interactions occur. In agreement 497 with previous data (69), we observed that the HIV-1 unspliced mRNA is preferentially 498 associated to the CBC subunit CBP80 both in the nucleus and the cytoplasm (Fig. 2). 499 Interestingly, we also confirmed that CBP80 interacts with Rev and showed that the CBC 500 subunit supports the activity of Rev in nuclear export and translation. Despite the fact that 501 our in silico data suggest that Rev does not interfere with the interaction between CBP80 502 and CBP20 (data not shown), it is still unknown whether CBP80 alone, or in the context of 503 the CBC, is responsible of these functions. Indeed, we observed that eIF4AI, which also 504 interacts with CBP80 and Rev, would interfere with recruitment of CBP20. Moreover, our 505 unpublished data also shows that CTIF, the CBP80/20-dependent translation initiation 506 factor is a potent inhibitor of Gag synthesis (Garcia-de-Gracia et al. manuscript in 507 preparation). These observations strongly suggest that CBP80 might promote Gag synthesis 508 in the context of a non-canonical CBC. Interestingly, a recent study reported the existence 509 of an alternative CBC formed by CBP80 and NCBP3, a novel cap-binding protein 510 specifically associated to mRNA nuclear export (19). Thus, it would be of interest to 511 determine whether the translating unspliced mRNA is indeed associated to the CBC and 512 which of the cap-binding proteins, CBP20 or NCBP3, is bound to the cap structure of the 513 viral transcript during translation. In this sense, the HIV-1 unspliced mRNA was shown to 514 contain a m2,2,7GpppG trimethylated cap in a process catalyzed by the methyltransferase 515 PIMT and dependent on the presence of Rev (77). Interestingly, increased cap 516 trimethylation by PIMT overexpression was shown to promote Gag synthesis suggesting 517 that a trimethylated cap favors polysome association of the unspliced mRNA (77). Since 518 trimethylation reduces the affinity of CBP20 and eIF4E for the cap (78,79), it would be of 519 interest to test the affinity of NCBP3 for m2,2,7GTP and whether this new cap-binding 520 protein drives translation of the HIV-1 unspliced mRNA together with CBP80. It should be 521 mentioned that our ISH-PLA analyses do not discard an association of the unspliced 522 mRNA with the cap-binding protein eIF4E, which probably reflects the proportion of viral 523 transcripts that contain monomethylated caps. Previous reports have shown that HIV-1 Gag 524 synthesis and replication are maintained under inhibition of eIF4E-driven cap-dependent 525 translation (69,80,81). Whether our Rev-CBP80-eIF4AI complex or the IRES-driven bioRxiv preprint doi: https://doi.org/10.1101/313312; this version posted May 2, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

526 mechanism of ribosome recruitment are responsible of maintaining Gag synthesis under 527 unfavorable conditions needs to be further investigated. 528 We also identified the DEAD-box RNA helicase eIF4AI as an additional partner of Rev 529 (Fig. 3). Interestingly, we observed that ectopic expression of eIF4AI promoted both the 530 cytoplasmic accumulation of the unspliced mRNA or translation depending on whether Rev 531 was present or not. Consistent with these observations, we observed that Rev promotes the 532 interaction between eIF4AI and the unspliced mRNA in the nucleus (Fig. 4). Since 533 inhibition of eIF4AI/II function by hippuristanol treatment resulted in a strong inhibition of 534 unspliced mRNA translation from the wild type provirus (data not shown), we propose that 535 the RNA helicase plays a dual role during viral replication by assisting Rev during nuclear 536 export and by promoting unspliced mRNA translation independently of the viral protein. 537 Further work is necessary to decipher the mechanism by which eIF4AI cooperates with Rev 538 during nuclear export. We further showed that CBP80 associates with eIF4AI and this 539 interaction was stimulated in the presence of Rev (Fig. 5). Although this interaction was 540 proposed some time ago (72), to our knowledge this is the first experimental evidence for 541 the association between CBP80 and eIF4AI. Since we observed that the CBP20 and eIF4AI 542 might compete for binding CBP80, it would be of interest to determine the consequences of 543 the CBP80-eIF4AI interaction, for example, during the pioneer round of translation or 544 NMD. It would also be interesting to evaluate whether Rev is able to regulate these cellular 545 processes. Interestingly, our molecular docking analyses using the Rev dimer bound to the 546 RRE suggest that the viral RNA serves as a platform for the assembly of the trimeric Rev- 547 CBP80-eIF4AI complex. Thus, we propose a model in which Rev interconnects CRM1- 548 dependent nuclear export with ribosome recruitment of the unspliced mRNA by driving the 549 recruitment of host factors required for both processes (Fig. 6). The proper and timely 550 assembly of such a Rev-dependent mRNP will determine the efficient association of the 551 unspliced mRNA with the host machineries for nuclear export and translation initiation. 552 Last but not least, the small molecule ABX464, currently under a phase II clinical trial, was 553 shown to interfere with the Rev-CBP80 interaction (82,83). Therefore, results presented 554 here will be useful either for the better understanding of the mechanism of action of this 555 small molecule or for the rational design of new drugs targeting the Rev-CBP80-eIF4A 556 complex. bioRxiv preprint doi: https://doi.org/10.1101/313312; this version posted May 2, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

557 558 559 Acknowledgements 560 Authors wish thank to Dr. Richard Cerione (Cornell University, USA), Dr. Jonathan Karn 561 (Case Western Reserve University, USA), Dr. Yoon Ki Kim (University of Korea) and Dr. 562 Marcelo López-Lastra (PUC, Chile) for their generous provisions of plasmids and reagents. 563 Authors wish also thank to Alessandra Dellarossa for technical support. The following 564 reagents were obtained through the NIH AIDS Reagents Program, Division of AIDS, 565 NIAID, NIH: HIV-1 p24 Monoclonal Antibody (183-H12-5C) from Dr. Bruce Chesebro

566 and Kathy Wehrly, THP-1ATCC from Drs Li Wu and Vineet N. Kewal Ramani and H9 cells 567 from Dr. Robert Gallo. 568 569 Funding 570 Work at RSR laboratory is funded by grants from FONDECYT (Nº 1160176) and PCI- 571 CONICYT (DRI USA2013-0005). RSR and TO holds an ECOS-CONICYT Cooperation 572 Grant (Nº C15B03). DTA holds a postdoctoral fellowship from FONDECYT (Grant Nº 573 3160091). BRA, FGG, CPM and AGA are recipients of a National Doctorate fellowship 574 from CONICYT. 575 576 Conflict of interest 577 The authors declare there is no any competing financial interest. 578 579 References 580 581 1. Frankel, A.D. and Young, J.A. (1998) HIV-1: fifteen proteins and an RNA. Annu 582 Rev Biochem, 67, 1-25. 583 2. Karn, J. and Stoltzfus, C.M. (2012) Transcriptional and posttranscriptional 584 regulation of HIV-1 gene expression. Cold Spring Harbor perspectives in 585 medicine, 2, a006916. 586 3. Purcell, D.F. and Martin, M.A. (1993) Alternative splicing of human 587 immunodeficiency virus type 1 mRNA modulates viral protein expression, 588 replication, and infectivity. J Virol, 67, 6365-6378. 589 4. Ocwieja, K.E., Sherrill-Mix, S., Mukherjee, R., Custers-Allen, R., David, P., Brown, 590 M., Wang, S., Link, D.R., Olson, J., Travers, K. et al. (2012) Dynamic regulation of bioRxiv preprint doi: https://doi.org/10.1101/313312; this version posted May 2, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

591 HIV-1 mRNA populations analyzed by single-molecule enrichment and long- 592 read sequencing. Nucleic Acids Res, 40, 10345-10355. 593 5. Kim, S.Y., Byrn, R., Groopman, J. and Baltimore, D. (1989) Temporal aspects of 594 DNA and RNA synthesis during human immunodeficiency virus infection: 595 evidence for differential gene expression. J Virol, 63, 3708-3713. 596 6. Leblanc, J., Weil, J. and Beemon, K. (2013) Posttranscriptional regulation of 597 retroviral gene expression: primary RNA transcripts play three roles as pre- 598 mRNA, mRNA, and genomic RNA. Wiley Interdiscip Rev RNA, 4, 567-580. 599 7. Moore, M.J. (2005) From birth to death: the complex lives of eukaryotic mRNAs. 600 Science, 309, 1514-1518. 601 8. Moore, M.J. and Proudfoot, N.J. (2009) Pre-mRNA processing reaches back to 602 transcription and ahead to translation. Cell, 136, 688-700. 603 9. Haimovich, G., Choder, M., Singer, R.H. and Trcek, T. (2013) The fate of the 604 messenger is pre-determined: a new model for regulation of gene expression. 605 Biochim Biophys Acta, 1829, 643-653. 606 10. Nott, A., Le Hir, H. and Moore, M.J. (2004) Splicing enhances translation in 607 mammalian cells: an additional function of the exon junction complex. Genes & 608 development, 18, 210-222. 609 11. Nott, A., Meislin, S.H. and Moore, M.J. (2003) A quantitative analysis of intron 610 effects on mammalian gene expression. RNA, 9, 607-617. 611 12. Lu, S. and Cullen, B.R. (2003) Analysis of the stimulatory effect of splicing on 612 mRNA production and utilization in mammalian cells. RNA, 9, 618-630. 613 13. Le Hir, H., Gatfield, D., Izaurralde, E. and Moore, M.J. (2001) The exon-exon 614 junction complex provides a binding platform for factors involved in mRNA 615 export and nonsense-mediated mRNA decay. The EMBO journal, 20, 4987-4997. 616 14. Le Hir, H., Nott, A. and Moore, M.J. (2003) How introns influence and enhance 617 eukaryotic gene expression. Trends in biochemical sciences, 28, 215-220. 618 15. Wiegand, H.L., Lu, S. and Cullen, B.R. (2003) Exon junction complexes mediate 619 the enhancing effect of splicing on mRNA expression. Proceedings of the 620 National Academy of Sciences of the United States of America, 100, 11327- 621 11332. 622 16. Diem, M.D., Chan, C.C., Younis, I. and Dreyfuss, G. (2007) PYM binds the 623 cytoplasmic exon-junction complex and ribosomes to enhance translation of 624 spliced mRNAs. Nature structural & molecular biology, 14, 1173-1179. 625 17. Ma, X.M., Yoon, S.O., Richardson, C.J., Julich, K. and Blenis, J. (2008) SKAR links 626 pre-mRNA splicing to mTOR/S6K1-mediated enhanced translation efficiency of 627 spliced mRNAs. Cell, 133, 303-313. 628 18. Cheng, H., Dufu, K., Lee, C.S., Hsu, J.L., Dias, A. and Reed, R. (2006) Human mRNA 629 export machinery recruited to the 5' end of mRNA. Cell, 127, 1389-1400. 630 19. Gebhardt, A., Habjan, M., Benda, C., Meiler, A., Haas, D.A., Hein, M.Y., Mann, A., 631 Mann, M., Habermann, B. and Pichlmair, A. (2015) mRNA export through an 632 additional cap-binding complex consisting of NCBP1 and NCBP3. Nat Commun, 633 6, 8192. 634 20. Gonatopoulos-Pournatzis, T. and Cowling, V.H. (2014) Cap-binding complex 635 (CBC). Biochem J, 457, 231-242. bioRxiv preprint doi: https://doi.org/10.1101/313312; this version posted May 2, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

636 21. Porrua, O. and Libri, D. (2013) RNA quality control in the nucleus: the Angels' 637 share of RNA. Biochim Biophys Acta, 1829, 604-611. 638 22. Rajanala, K. and Nandicoori, V.K. (2012) Localization of nucleoporin Tpr to the 639 nuclear pore complex is essential for Tpr mediated regulation of the export of 640 unspliced RNA. PLoS One, 7, e29921. 641 23. Coyle, J.H., Bor, Y.C., Rekosh, D. and Hammarskjold, M.L. (2011) The Tpr protein 642 regulates export of mRNAs with retained introns that traffic through the Nxf1 643 pathway. RNA, 17, 1344-1356. 644 24. Bonnet, A., Bretes, H. and Palancade, B. (2015) Nuclear pore components affect 645 distinct stages of intron-containing gene expression. Nucleic Acids Res, 43, 646 4249-4261. 647 25. Bonnet, A. and Palancade, B. (2015) Intron or no intron: a matter for nuclear 648 pore complexes. Nucleus, 6, 455-461. 649 26. Chang, D.D. and Sharp, P.A. (1989) Regulation by HIV Rev depends upon 650 recognition of splice sites. Cell, 59, 789-795. 651 27. Legrain, P. and Rosbash, M. (1989) Some cis- and trans-acting mutants for 652 splicing target pre-mRNA to the cytoplasm. Cell, 57, 573-583. 653 28. Sodroski, J., Goh, W.C., Rosen, C., Dayton, A., Terwilliger, E. and Haseltine, W. 654 (1986) A second post-transcriptional trans-activator gene required for HTLV- 655 III replication. Nature, 321, 412-417. 656 29. Feinberg, M.B., Jarrett, R.F., Aldovini, A., Gallo, R.C. and Wong-Staal, F. (1986) 657 HTLV-III expression and production involve complex regulation at the levels of 658 splicing and translation of viral RNA. Cell, 46, 807-817. 659 30. Knight, D.M., Flomerfelt, F.A. and Ghrayeb, J. (1987) Expression of the art/trs 660 protein of HIV and study of its role in viral envelope synthesis. Science, 236, 661 837-840. 662 31. Felber, B.K., Hadzopoulou-Cladaras, M., Cladaras, C., Copeland, T. and Pavlakis, 663 G.N. (1989) rev protein of human immunodeficiency virus type 1 affects the 664 stability and transport of the viral mRNA. Proc Natl Acad Sci U S A, 86, 1495- 665 1499. 666 32. Malim, M.H., Hauber, J., Le, S.Y., Maizel, J.V. and Cullen, B.R. (1989) The HIV-1 667 rev trans-activator acts through a structured target sequence to activate 668 nuclear export of unspliced viral mRNA. Nature, 338, 254-257. 669 33. Hammarskjold, M.L., Heimer, J., Hammarskjold, B., Sangwan, I., Albert, L. and 670 Rekosh, D. (1989) Regulation of human immunodeficiency virus env 671 expression by the rev gene product. J Virol, 63, 1959-1966. 672 34. D'Agostino, D.M., Felber, B.K., Harrison, J.E. and Pavlakis, G.N. (1992) The Rev 673 protein of human immunodeficiency virus type 1 promotes polysomal 674 association and translation of gag/pol and vpu/env mRNAs. Mol Cell Biol, 12, 675 1375-1386. 676 35. Arrigo, S.J. and Chen, I.S. (1991) Rev is necessary for translation but not 677 cytoplasmic accumulation of HIV-1 vif, vpr, and env/vpu 2 RNAs. Genes Dev, 5, 678 808-819. 679 36. Adachi, A., Gendelman, H.E., Koenig, S., Folks, T., Willey, R., Rabson, A. and 680 Martin, M.A. (1986) Production of acquired immunodeficiency syndrome- bioRxiv preprint doi: https://doi.org/10.1101/313312; this version posted May 2, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

681 associated retrovirus in human and nonhuman cells transfected with an 682 infectious molecular clone. J Virol, 59, 284-291. 683 37. Soto-Rifo, R., Limousin, T., Rubilar, P.S., Ricci, E.P., Decimo, D., Moncorge, O., 684 Trabaud, M.A., Andre, P., Cimarelli, A. and Ohlmann, T. (2012) Different effects 685 of the TAR structure on HIV-1 and HIV-2 genomic RNA translation. Nucleic 686 Acids Res, Mar 1, 2653-2667. 687 38. Frohlich, A., Rojas-Araya, B., Pereira-Montecinos, C., Dellarossa, A., Toro-Ascuy, 688 D., Prades-Perez, Y., Garcia-de-Gracia, F., Garces-Alday, A., Rubilar, P.S., 689 Valiente-Echeverria, F. et al. (2016) DEAD-box RNA helicase DDX3 connects 690 CRM1-dependent nuclear export and translation of the HIV-1 unspliced mRNA 691 through its N-terminal domain. Biochim Biophys Acta. 692 39. Panthu, B., Terrier, O., Carron, C., Traversier, A., Corbin, A., Balvay, L., Lina, B., 693 Rosa-Calatrava, M. and Ohlmann, T. (2017) The NS1 Protein from Influenza 694 Virus Stimulates Translation Initiation by Enhancing Ribosome Recruitment to 695 mRNAs. J Mol Biol. 696 40. Brasey, A., Lopez-Lastra, M., Ohlmann, T., Beerens, N., Berkhout, B., Darlix, J.L. 697 and Sonenberg, N. (2003) The leader of human immunodeficiency virus type 1 698 genomic RNA harbors an internal ribosome entry segment that is active during 699 the G2/M phase of the cell cycle. J Virol, 77, 3939-3949. 700 41. Soto-Rifo, R., Rubilar, P.S., Limousin, T., de Breyne, S., Decimo, D. and Ohlmann, 701 T. (2012) DEAD-box protein DDX3 associates with eIF4F to promote 702 translation of selected mRNAs. The EMBO journal, 31, 3745-3756. 703 42. Choe, J., Oh, N., Park, S., Lee, Y.K., Song, O.K., Locker, N., Chi, S.G. and Kim, Y.K. 704 (2012) Translation initiation on mRNAs bound by nuclear cap-binding protein 705 complex CBP80/20 requires interaction between CBP80/20-dependent 706 translation initiation factor and eukaryotic translation initiation factor 3g. The 707 Journal of biological chemistry, 287, 18500-18509. 708 43. Garcia-Mesa, Y., Jay, T.R., Checkley, M.A., Luttge, B., Dobrowolski, C., Valadkhan, 709 S., Landreth, G.E., Karn, J. and Alvarez-Carbonell, D. (2017) Immortalization of 710 primary microglia: a new platform to study HIV regulation in the central 711 nervous system. J Neurovirol, 23, 47-66. 712 44. Mann, D.L., O'Brien, S.J., Gilbert, D.A., Reid, Y., Popovic, M., Read-Connole, E., 713 Gallo, R.C. and Gazdar, A.F. (1989) Origin of the HIV-susceptible human CD4+ 714 cell line H9. AIDS Res Hum Retroviruses, 5, 253-255. 715 45. Wu, L., Martin, T.D., Carrington, M. and KewalRamani, V.N. (2004) Raji B cells, 716 misidentified as THP-1 cells, stimulate DC-SIGN-mediated HIV transmission. 717 Virology, 318, 17-23. 718 46. Reed, S.E., Staley, E.M., Mayginnes, J.P., Pintel, D.J. and Tullis, G.E. (2006) 719 Transfection of mammalian cells using linear polyethylenimine is a simple and 720 effective means of producing recombinant adeno-associated virus vectors. J 721 Virol Methods, 138, 85-98. 722 47. Toohey, K., Wehrly, K., Nishio, J., Perryman, S. and Chesebro, B. (1995) Human 723 immunodeficiency virus envelope V1 and V2 regions influence replication 724 efficiency in macrophages by affecting virus spread. Virology, 213, 70-79. 725 48. Soto-Rifo, R., Rubilar, P.S. and Ohlmann, T. (2013) The DEAD-box helicase 726 DDX3 substitutes for the cap-binding protein eIF4E to promote bioRxiv preprint doi: https://doi.org/10.1101/313312; this version posted May 2, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

727 compartmentalized translation initiation of the HIV-1 genomic RNA. Nucleic 728 Acids Res, 41, 6286-6299. 729 49. Soderberg, O., Gullberg, M., Jarvius, M., Ridderstrale, K., Leuchowius, K.J., 730 Jarvius, J., Wester, K., Hydbring, P., Bahram, F., Larsson, L.G. et al. (2006) Direct 731 observation of individual endogenous protein complexes in situ by proximity 732 ligation. Nat Methods, 3, 995-1000. 733 50. Valiente-Echeverria, F., Melnychuk, L., Vyboh, K., Ajamian, L., Gallouzi, I.E., 734 Bernard, N. and Mouland, A.J. (2014) eEF2 and Ras-GAP SH3 domain-binding 735 protein (G3BP1) modulate stress granule assembly during HIV-1 infection. Nat 736 Commun, 5, 4819. 737 51. Dias, S.M., Wilson, K.F., Rojas, K.S., Ambrosio, A.L. and Cerione, R.A. (2009) The 738 molecular basis for the regulation of the cap-binding complex by the importins. 739 Nature structural & molecular biology, 16, 930-937. 740 52. Chang, J.H., Cho, Y.H., Sohn, S.Y., Choi, J.M., Kim, A., Kim, Y.C., Jang, S.K. and Cho, 741 Y. (2009) Crystal structure of the eIF4A-PDCD4 complex. Proc Natl Acad Sci U S 742 A, 106, 3148-3153. 743 53. DiMattia, M.A., Watts, N.R., Stahl, S.J., Rader, C., Wingfield, P.T., Stuart, D.I., 744 Steven, A.C. and Grimes, J.M. (2010) Implications of the HIV-1 Rev dimer 745 structure at 3.2 A resolution for multimeric binding to the Rev response 746 element. Proc Natl Acad Sci U S A, 107, 5810-5814. 747 54. Daugherty, M.D., Booth, D.S., Jayaraman, B., Cheng, Y. and Frankel, A.D. (2010) 748 HIV Rev response element (RRE) directs assembly of the Rev homooligomer 749 into discrete asymmetric complexes. Proc Natl Acad Sci U S A, 107, 12481- 750 12486. 751 55. Jayaraman, B., Crosby, D.C., Homer, C., Ribeiro, I., Mavor, D. and Frankel, A.D. 752 (2014) RNA-directed remodeling of the HIV-1 protein Rev orchestrates 753 assembly of the Rev-Rev response element complex. eLife, 4, e04120. 754 56. Eswar, N., Webb, B., Marti-Renom, M.A., Madhusudhan, M.S., Eramian, D., Shen, 755 M.Y., Pieper, U. and Sali, A. (2006) Comparative protein structure modeling 756 using Modeller. Curr Protoc Bioinformatics, Chapter 5, Unit-5 6. 757 57. Humphrey, W., Dalke, A. and Schulten, K. (1996) VMD: visual molecular 758 dynamics. J Mol Graph, 14, 33-38, 27-38. 759 58. Melo, F. and Feytmans, E. (1998) Assessing protein structures with a non-local 760 atomic interaction energy. J Mol Biol, 277, 1141-1152. 761 59. Laskowski, R.A., MacArthur, M.W., Moss, D.S. and Thornton, J.M. (1993) 762 PROCHECK: A program to check the stereochemical quality of protein 763 structures. Journal of Applied Crystalography, 26, 283-291. 764 60. Schneidman-Duhovny, D., Inbar, Y., Nussinov, R. and Wolfson, H.J. (2005) 765 PatchDock and SymmDock: servers for rigid and symmetric docking. Nucleic 766 Acids Res, 33, W363-367. 767 61. Mashiach, E., Schneidman-Duhovny, D., Andrusier, N., Nussinov, R. and Wolfson, 768 H.J. (2008) FireDock: a web server for fast interaction refinement in molecular 769 docking. Nucleic Acids Res, 36, W229-232. 770 62. Pollard, V.W. and Malim, M.H. (1998) The HIV-1 Rev protein. Annu Rev 771 Microbiol, 52, 491-532. bioRxiv preprint doi: https://doi.org/10.1101/313312; this version posted May 2, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

772 63. Groom, H.C., Anderson, E.C., Dangerfield, J.A. and Lever, A.M. (2009) Rev 773 regulates translation of human immunodeficiency virus type 1 RNAs. J Gen 774 Virol, 90, 1141-1147. 775 64. Blissenbach, M., Grewe, B., Hoffmann, B., Brandt, S. and Uberla, K. (2010) 776 Nuclear RNA export and packaging functions of HIV-1 Rev revisited. J Virol, 84, 777 6598-6604. 778 65. Soto-Rifo, R., Valiente-Echeverria, F., Rubilar, P.S., Garcia-de-Gracia, F., Ricci, 779 E.P., Limousin, T., Decimo, D., Mouland, A.J. and Ohlmann, T. (2014) HIV-2 780 genomic RNA accumulates in stress granules in the absence of active 781 translation. Nucleic Acids Res, 42, 12861-12875. 782 66. Alais, S., Soto-Rifo, R., Balter, V., Gruffat, H., Manet, E., Schaeffer, L., Darlix, J.L., 783 Cimarelli, A., Raposo, G., Ohlmann, T. et al. (2012) Functional mechanisms of 784 the cellular prion protein (PrP(C)) associated anti-HIV-1 properties. Cellular 785 and molecular life sciences : CMLS, 69, 1331-1352. 786 67. Groom, H.C., Anderson, E.C. and Lever, A.M. (2009) Rev: beyond nuclear export. 787 The Journal of general virology, 90, 1303-1318. 788 68. Gallego, J., Greatorex, J., Zhang, H., Yang, B., Arunachalam, S., Fang, J., Seamons, J., 789 Lea, S., Pomerantz, R.J. and Lever, A.M. (2003) Rev binds specifically to a purine 790 loop in the SL1 region of the HIV-1 leader RNA. J Biol Chem, 278, 40385-40391. 791 69. Sharma, A., Yilmaz, A., Marsh, K., Cochrane, A. and Boris-Lawrie, K. (2012) 792 Thriving under Stress: Selective Translation of HIV-1 Structural Protein mRNA 793 during Vpr-Mediated Impairment of eIF4E Translation Activity. PLoS pathogens, 794 8, e1002612. 795 70. Taniguchi, I., Mabuchi, N. and Ohno, M. (2014) HIV-1 Rev protein specifies the 796 viral RNA export pathway by suppressing TAP/NXF1 recruitment. Nucleic 797 Acids Res, 42, 6645-6658. 798 71. Kim, K.M., Cho, H., Choi, K., Kim, J., Kim, B.W., Ko, Y.G., Jang, S.K. and Kim, Y.K. 799 (2009) A new MIF4G domain-containing protein, CTIF, directs nuclear cap- 800 binding protein CBP80/20-dependent translation. Genes Dev, 23, 2033-2045. 801 72. Lejeune, F., Ranganathan, A.C. and Maquat, L.E. (2004) eIF4G is required for the 802 pioneer round of translation in mammalian cells. Nat Struct Mol Biol, 11, 992- 803 1000. 804 73. Daugherty, M.D., Liu, B. and Frankel, A.D. (2010) Structural basis for 805 cooperative RNA binding and export complex assembly by HIV Rev. Nat Struct 806 Mol Biol, 17, 1337-1342. 807 74. Briggs, J.A., Simon, M.N., Gross, I., Krausslich, H.G., Fuller, S.D., Vogt, V.M. and 808 Johnson, M.C. (2004) The stoichiometry of Gag protein in HIV-1. Nat Struct Mol 809 Biol, 11, 672-675. 810 75. Toro-Ascuy, D., Rojas-Araya, B., Valiente-Echeverria, F. and Soto-Rifo, R. (2016) 811 Interactions between the HIV-1 Unspliced mRNA and Host mRNA Decay 812 Machineries. Viruses, 8. 813 76. Coyle, J.H., Guzik, B.W., Bor, Y.C., Jin, L., Eisner-Smerage, L., Taylor, S.J., Rekosh, 814 D. and Hammarskjold, M.L. (2003) Sam68 enhances the cytoplasmic utilization 815 of intron-containing RNA and is functionally regulated by the nuclear kinase 816 Sik/BRK. Molecular and cellular biology, 23, 92-103. bioRxiv preprint doi: https://doi.org/10.1101/313312; this version posted May 2, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

817 77. Yedavalli, V.S. and Jeang, K.T. (2010) Trimethylguanosine capping selectively 818 promotes expression of Rev-dependent HIV-1 RNAs. Proceedings of the 819 National Academy of Sciences of the United States of America, 107, 14787- 820 14792. 821 78. Niedzwiecka, A., Marcotrigiano, J., Stepinski, J., Jankowska-Anyszka, M., 822 Wyslouch-Cieszynska, A., Dadlez, M., Gingras, A.C., Mak, P., Darzynkiewicz, E., 823 Sonenberg, N. et al. (2002) Biophysical studies of eIF4E cap-binding protein: 824 recognition of mRNA 5' cap structure and synthetic fragments of eIF4G and 4E- 825 BP1 proteins. Journal of molecular biology, 319, 615-635. 826 79. Worch, R., Niedzwiecka, A., Stepinski, J., Mazza, C., Jankowska-Anyszka, M., 827 Darzynkiewicz, E., Cusack, S. and Stolarski, R. (2005) Specificity of recognition 828 of mRNA 5' cap by human nuclear cap-binding complex. RNA, 11, 1355-1363. 829 80. Monette, A., Valiente-Echeverria, F., Rivero, M., Cohen, E.A., Lopez-Lastra, M. 830 and Mouland, A.J. (2013) Dual mechanisms of translation initiation of the full- 831 length HIV-1 mRNA contribute to gag synthesis. PLoS One, 8, e68108. 832 81. Amorim, R., Costa, S.M., Cavaleiro, N.P., da Silva, E.E. and da Costa, L.J. (2014) 833 HIV-1 transcripts use IRES-initiation under conditions where Cap-dependent 834 translation is restricted by poliovirus 2A protease. PLoS One, 9, e88619. 835 82. Campos, N., Myburgh, R., Garcel, A., Vautrin, A., Lapasset, L., Nadal, E.S., 836 Mahuteau-Betzer, F., Najman, R., Fornarelli, P., Tantale, K. et al. (2015) Long 837 lasting control of viral rebound with a new drug ABX464 targeting Rev - 838 mediated viral RNA biogenesis. Retrovirology, 12, 30. 839 83. Steens, J.M., Scherrer, D., Gineste, P., Barrett, P.N., Khuanchai, S., Winai, R., 840 Ruxrungtham, K., Tazi, J., Murphy, R. and Ehrlich, H. (2017) Safety, 841 Pharmacokinetics, and Antiviral Activity of a Novel HIV Antiviral, ABX464, in 842 Treatment-Naive HIV-Infected Subjects in a Phase 2 Randomized, Controlled 843 Study. Antimicrob Agents Chemother, 61. 844 845 846 Figures Legends 847 Figure 1: HIV-1 Rev promotes nuclear export and translation of the unspliced mRNA. 848 A) HeLa cells were transfected with 0.3 µg of pNL4.3-wt (wt), pNL4.3-ΔRev (ΔRev) or 849 pNL4.3-ΔRev together with 0.1 µg of the pCDNA-Flag-Rev vector as described in 850 materials and methods (pCDNA-d2EGFP was used as a control when Flag-Rev was not 851 included). At 24 hpt, cell extracts were used to detect Gag and Flag-Rev by Western blot. 852 Actin was used as a loading control. * Denotes an unspecific band detected with the anti- 853 actin-HRP antibody. 854 B) HeLa cells were transfected as above and were subjected to RNA FISH and laser scan 855 confocal microscopy analyzes as described in materials and methods. The unspliced mRNA 856 is showed in green and Flag-Rev in red. Scale bar 10 µm. bioRxiv preprint doi: https://doi.org/10.1101/313312; this version posted May 2, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

857 C) HeLa cells were transfected with 0.3 µg of pNL4.3R (wt) or pNL4.3R-ΔRev (ΔRev) 858 proviruses as described in materials and methods. At 24 hpt, cell extracts were prepared for 859 Gag-Renilla activity measurement and for cytoplasmic RNA extraction and RT-qPCR 860 analyzes. Results for Gag synthesis (left panel), cytoplasmic unspliced mRNA (middle 861 panel) and translational efficiency (right panel) were normalized to the wild type provirus 862 (arbitrary set to 100%) and presented as the mean +/- SD of three independent experiments 863 (**p<0.01; ****p<0.0001, t-test). 864 D) HeLa cells were transfected with 0.3 µg of pCMV-NL4.3R-wt (CMV-wt) or pCMV- 865 NL4.3R-ΔRev (CMV-ΔRev) proviruses as described in materials and methods. At 24 hpt, 866 cell extracts were prepared for Gag-Renilla activity measurement and for cytoplasmic RNA 867 extraction and RT-qPCR analyzes. Results for Gag synthesis (left panel), cytoplasmic 868 unspliced mRNA (middle panel) and translational efficiency (right panel) were normalized 869 to the wild type provirus (arbitrary set to 100%) and presented as the mean +/- SD of three 870 independent experiments (**p<0.01; ****p<0.0001, t-test). 871 872 Figure 2: CBP80 cooperates with the functions of Rev on Gag synthesis from the 873 unspliced mRNA. 874 A) HeLa cells were transfected with 1 µg pNL4.3-wt together with 1 µg pCMV-myc- 875 CBP80 or 1 µg pCMV-myc-eIF4E. At 24 hpt, the interaction between the unspliced mRNA 876 and the myc- tagged protein was analyzed by the ISH-PLA protocol described in material 877 and methods. Red dots indicate the interactions between the unspliced mRNA and the 878 corresponding myc-tagged protein. Scale bar 10 µm. A quantification of the dots/cell in 879 each condition is presented on the right (p<0.0001, Mann-Whitney test). 880 B) HeLa cells were transfected with 0.3 µg of pNL4.3R or pNL4.3R ∆Rev proviruses 881 together with 1 µg of pCMV-myc-CBP80 or pCMV-myc-eIF4E (pCVM-myc-d2EGFP was 882 used as control). At 24 hpt, cell extracts were prepared for Gag-Renilla activity 883 measurement and for cytoplasmic RNA extraction and RT-qPCR analyzes. Results for Gag 884 synthesis (left panel), cytoplasmic unspliced mRNA (middle panel) and translational 885 efficiency (right panel) were normalized to the wild type provirus (arbitrary set to 100%) 886 and presented as the mean +/- SD of three independent experiments (*p<0.05; ***p<0.001 887 and NS; non-significant, t-test). bioRxiv preprint doi: https://doi.org/10.1101/313312; this version posted May 2, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

888 C) HeLa cells were transfected with 0.3 µg of pNL4.3R or pNL4.3R ∆Rev proviruses 889 together with 1 µg of pCMV-myc-CBP80 (pCVM-myc-d2EGFP was used as control) as 890 described in materials and methods. At 24 hpt, cell extracts were prepared for Gag-Renilla 891 activity measurement and for cytoplasmic RNA extraction and RT-qPCR analyzes. Results 892 for Gag synthesis (left panel), cytoplasmic unspliced mRNA (middle panel) and 893 translational efficiency (right panel) were normalized to the wild type provirus (arbitrary set 894 to 100%) and presented as the mean +/- SD of three independent experiments. (*p<0.05; 895 **p<0.01 and NS; non-significant, t-test). 896 D) HeLa cells were transfected with 1 µg pCDNA-Flag-Rev and 1 µg pCDNA-V5-CBP80. 897 At 24 hpt, the interaction between Flag-Rev and V5-CBp80 was analyzed by PLA. Red 898 dots indicate interactions between both proteins. Scale bar 10 µm. 899 900 Figure 3: DEAD-box RNA helicase eIF4AI interacts with Rev and promotes Gag 901 synthesis from the unspliced mRNA 902 A) HeLa cells were transfected with 1 µg pCDNA-Flag-Rev together with 1 µg pCIneo- 903 HA-eIF4E, pCIneo-HA-eIF4G or pCIneo-HA-eIF4A. At 24 hpt, the interaction between 904 Flag- and HA- tagged eIFs was analyzed by PLA. Red dots indicate interactions between 905 Flag-Rev and the corresponding HA-tagged protein. Scale bar 10 µm 906 B) Dots/cell for Rev-4E, Rev-4G and Rev-4A were quantified using ImageJ (p<0.0001, 907 Mann-Whitney test). 908 C) HeLa cells were transfected with 0.3 µg of pNL4.3R-wt or pNL4.3R-ΔRev together 909 with 1 µg of the pCIneo-HA-eIF4A vector as described in materials and methods (pCIneo- 910 HA-d2EGFP was used as a control). At 24 hpt, cell extracts were prepared for Gag-Renilla 911 activity measurement and for cytoplasmic RNA extraction and RT-qPCR analyzes. Results 912 for Gag synthesis (left panel), cytoplasmic unspliced mRNA (middle panel) and 913 translational efficiency (right panel) were normalized to the wild type provirus (arbitrary set 914 to 100%) and presented as the mean +/- SD of three independent experiments (*p<0.05; 915 **p<0.01 and NS; non-significant, t-test). 916 917 Figure 4: Rev promotes the recruitment of CBP80 and eIF4A to the HIV-1 unspliced 918 mRNA bioRxiv preprint doi: https://doi.org/10.1101/313312; this version posted May 2, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

919 A) HeLa cells were transfected with 1 µg pNL4.3-wt, 1 µg pNL4.3 ∆Rev or 1 µg pNL4.3 920 ∆Rev + 0,3 µg pCDNA-Flag-Rev together with 1 µg pCMV-Myc-CBP80. At 24 hpt, the 921 interaction between unspliced mRNA and CBP80 was analyzed by ISH-PLA. Scale bar 10 922 µm (upper panel). Dots per cell quantifications for total unspliced mRNA-CBP80 923 interactions (left panel), cytoplasmic interactions (middle panel) and nuclear interactions 924 (right panel) are presented below. All interactions were quantified using ImageJ 925 (****p<0.0001 and NS; non-significant, Mann-Whitney test). 926 B) HeLa cells were transfected with 1 µg of pNL4.3-wt, pNL4.3-∆Rev or 1 µg pNL4.3- 927 ∆Rev + 0,3 µg pCDNA-Flag-Rev together with 1 µg pCIneo-HA-eIF4A. At 24 hpt, the 928 interaction between the unspliced mRNA and eIF4AI was analyzed by ISH-PLA. Scale bar 929 10 µm (upper panel). Dots per cell for total unspliced mRNA-eIF4A interactions (left 930 panel), cytoplasmic interactions (middle panel) and nuclear interactions (right panel) are 931 presented below. All interactions were quantified using ImageJ (**p<0.01; ****p<0.0001 932 and NS; non-significant, Mann-Whitney test) (lower panel) 933 934 Figure 5: A Rev-CBP80-eIF4AI assembles around the RRE 935 A) HeLa cells were transfected with 1 µg pCMV-myc-CBP80 and 1 µg pCIneo-HA-eIF4A 936 together 1 µg pEGFP-Rev (pEGFP was used as a control). At 24 hpt, the interaction 937 between HA-eIF4AI and myc-CBP80 was analyzed by PLA. Scale bar 10 µm. 938 B) Dots per cell quantification for eIF4A and CBP80 interactions using ImageJ 939 (****p<0.0001, Mann-Whitney test). 940 C) Model of the CBP80 (yellow) and eIF4AI (green) complex showing the interaction 941 interface and the residues involved. 942 D) Model of the Rev dimer (magenta), CBP80 (yellow) and eIF4AI (green) complex. The 943 interaction interface and the residues involved in the Rev-CBP80 interaction are presented. 944 E) Model of the Rev dimer (magenta)/RRE (pink), CBP80 (yellow) and eIF4AI (green) 945 complex showing the interaction interface and the residues involved (residues involved in 946 the Rev-RRE interaction were omitted for simplicity). 947 948 Figure 6: A model for the function of the Rev-CBP80-eIFAI complex on Gag synthesis bioRxiv preprint doi: https://doi.org/10.1101/313312; this version posted May 2, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

949 The Rev-CBP80-eIF4AI complex might assemble into the cytoplasm and be imported to 950 the nucleus to be recruited onto the RRE. Once Rev binds to CRM1 through its NES, this 951 nuclear export mRNP translocates to the cytoplasm. Upon dissociation of Rev and CRM1, 952 CBP80 remain associated to the gRNA to promote translation in a Rev-dependent manner. 953 In contrast, eIF4AI is either recruited or remain associated to the gRNA independently of 954 Rev in order to promote gRNA translation. The cap-binding protein associated to the HIV-1 955 gRNA is still unknown. 956 bioRxiv preprint doi: https://doi.org/10.1101/313312; this version posted May 2, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

A B Gag

CAp24

Flag-Rev

* Wild type Rev ΔRev + Flag-Rev AcCn Δ

wt + - -

ΔRev - + +

Flag-Rev - - +

C

120 **** 120 ** 120 ****

100 100 100

80 80 80 mRNA efficiency us 60 60 60

(% of control) 40 (% of control) 40 40 (% of control) Gag-Renilla RLU Cytoplasmic 20 20 Transla>onal 20

0 0 0 pNL4.3R wt pNL4.3R ΔRev pNL4.3R wt pNL4.3R ΔRev pNL4.3R wt pNL4.3R ΔRev DRev DRev DRev D

120 120 120 **** ** ****

100 100 100

80 80 80

60 60 60 (% of control) (% of control) (% of control)

Gag-Renilla RLU 40 40 40 Cytoplasmic us mRNA Transla>onal efficiency 20 20 20

0 0 0 CMV-wt CMV-∆Rev CMV-wt CMV-∆Rev CMV-wt CMV-∆Rev Figure 1 bioRxiv preprint doi: https://doi.org/10.1101/313312; this version posted May 2, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

PLA assay gRNA HIV-1 / protein A ISH-PLA CBP80- gRNA ISH-PLA eIF4E- gRNA 80 ****

60

40

DOTS/CELL 20

0

eIF4E-wt CBP80-wt B

1000 *** 350 * 350 *** 300 800 300 NS 250 NS 250 NS 600 200 200 400 150 150 (% of control) (% of control)

(% of control) 100

Gag-Renilla RLU 100 200 us mRNA transla>on 50 Cytoplasmic us mRNA 50 0 0 0 Control Control eIF4E CBP80 eIF4E CBP80 Control eIF4E CBP80 Control eIF4E CBP80

C ** NS ** * * NS 1800 700 350 1600 600 300 1400 500 250 1200 1000 400 200 800 300 150 600 (% of control) (% of control)

(% del control) 200 100 Gag-Renilla RLU 400 us mRNA transla>on 200 Cytoplasmic usmRNA 100 50 0 0 0 CBP80 CBP80 CBP80 CBP80 CBP80 CBP80 Control Control Control Control Control Control

wt ∆Rev wt ∆Rev wt ∆Rev

D PLA Rev-CBP80

Figure 2 bioRxiv preprint doi: https://doi.org/10.1101/313312; this version posted May 2, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

A PLA Rev-eIF4E PLA Rev-eIF4G PLA Rev-eIF4A

B **** 5 0 0 **** **** 4 0 0 l l

e 3 0 0 C / s t

o 2 0 0 D

1 0 0

0

C o n tr o l R e v -4 E R e v -4 G R e v -4 A

C

NS * NS * ** * 400 700 800

350 700 600 600 300 500 mRNA

RLU 250

500 us

400 transla>on 400 200 300 Renilla 300 150 (% of control) (% of control) mRNA

(% of control) 200

Gag- 200 100 us Cytoplasmic 100 50 100 0 0 0 eIF4A eIF4A eIF4A eIF4A eIF4A eIF4A Control Control Control Control Control Control

wt ∆Rev wt ∆Rev wt ∆Rev

Figure 3 bioRxiv preprint doi: https://doi.org/10.1101/313312; this version posted May 2, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. A ISH-PLA CBP80- gRNA ISH-PLA CBP80- gRNA ISH-PLA CBP80- gRNA

wt ΔRev ΔRev + Rev

ISH-PLA TotalCBP80-Unspliced interactions mRNA (Total) ISH-PLACytoplasmic CBP80-Unspliced interactions mRNA (cytoplasm) ISH-PLANuclear CBP80-Unspliced interactions mRNA (Nuclear) NS NS NS **** **** 80 **** 60 25

60 20 40 15 40 10 20 DOTS/CELL DOTS/CELL 20 DOTS/CELL 5

0 0 0

Rev Rev Rev Δ Δ Δ Rev+Rev Rev+Rev Rev+Rev CBP80-wt Δ CBP80-wt Δ CBP80-wt Δ CBP80- CBP80- CBP80-

CBP80- CBP80- CBP80- B

ISH-PLA eIF4A - gRNA ISH-PLA eIF4A - gRNA ISH-PLA eIF4A - gRNA

wt ΔRev ΔRev + Rev

ISH-PLA eIF4A-Total Unsplicedinteractions mRNA (Total) ISH-PLACytoplasmic eIF4A- Unspl interactions mRNA (cytoplasm) ISH-PLANuclear eIF4A- Unspliced interactions mRNA (Nuclear) NS NS NS 60 NS ** 40 40 ****

30 40 30

20 20 20 DOTS/CELL DOTS/CELL 10 DOTS/CELL 10

0 0 0

Rev Rev 4A-wt Δ Δ Rev 4A-wt 4A-wt Δ 4A- Rev+Rev Δ 4A- Rev+Rev 4A- Rev+Rev Δ Δ 4A- 4A- 4A- Figure 4 A B PLA CBP80-eIF4A

*** PLA CBP80-eIF4A PLA CBP80-eIF4A 100

80

60

40 DOTS/CELL 20

0 - Rev +Rev - Rev + Rev C

TYR70 MET19 eIF4AI LYS557 ILE23 TYR126

LYS511 GLU24 ARG329 HIS561 SER564 GLU317 LYS327 GLU397 VAL371 GLU396 LYS369 GLU658 LYS684 LEU270 LYS665 ASN401 CBP80 ASP685 ALA667 Cecilia Francís Rojas Fuentes Ingeniero en Bioinformática Licenciado en Ciencias Cecilia Francís Rojas Fuentes Ingeniero en Bioinformática Licenciado en Ciencias CBP80-4A-REV/RRE(sin los aa rev que anclan RRE) CBP80/D 4A-REV E

PRO31 ASN204 SER205 GLU578 ASN30

ARG44 ASN206 ARG38

ARG42 MET719 GLU723 GLU95 ARG39 THR722 GLU20

VAL726 LEU768 MET19 ARG38

PRO31 LYS20 LYS511 LYS509

Rev

RRE

eIF4AI

eIF4AI

CBP80 CBP80

Figure 5

4 5 Rev/RRE-CBP80-eIF4AI complex

CRM1

? AAA(N)A

Nuclear export mRNP

CRM1

Rev-CBP80-eIF4AI complex

? AAA(N)A

Transla7on ini7a7on mRNP

Figure 6