key: cord-0729148-32qvqril
authors: Lesage, Sarah; Chazal, Maxime; Beauclair, Guillaume; Batalie, Damien; Cerboni, Silvia; Couderc, Elodie; Lescure, Aurianne; Del Nery, Elaine; Tangy, Frédéric; Martin, Annette; Manel, Nicolas; Jouvenet, Nolwenn
title: Discovery of genes that modulate flavivirus replication in an interferon-dependent manner
date: 2021-09-29
journal: J Mol Biol
DOI: 10.1016/j.jmb.2021.167277
sha: c0b0a7b55530a1a5c91dad8a1d0f8d262420060a
doc_id: 729148
cord_uid: 32qvqril

Establishment of the interferon (IFN)-mediated antiviral state provides a crucial initial line of defense against viral infection. Numerous genes that contribute to this antiviral state remain to be identified. Using a loss-of-function strategy, we screened an original library of 1156 siRNAs targeting 386 individual curated human genes in stimulated microglial cells infected with Zika virus (ZIKV), an emerging RNA virus that belongs to the flavivirus genus. The screen recovered twenty-one potential host proteins that modulate ZIKV replication in an IFN-dependent manner, including the previously known IFITM3 and LY6E. Further characterization contributed to delineate the spectrum of action of these genes towards other pathogenic RNA viruses, including Hepatitis C virus and SARS-CoV-2. Our data revealed that APOL3 acts as a proviral factor for ZIKV and several other related and unrelated RNA viruses. In addition, we showed that MTA2, a chromatin remodeling factor, possesses potent flavivirus-specific antiviral functions induced by IFN. Our work identified previously unrecognized genes that modulate the replication of RNA viruses in an IFN-dependent manner, opening new perspectives to target weakness points in the life cycle of these viruses.

Establishment of the interferon (IFN)-mediated antiviral state provides a crucial initial line of defense 25 against viral infection. Numerous genes that contribute to this antiviral state remain to be identified.

Using a loss-of-function strategy, we screened an original library of 1156 siRNAs targeting 386 27 individual curated human genes in stimulated microglial cells infected with Zika virus (ZIKV), an 28 emerging RNA virus that belongs to the flavivirus genus. The screen recovered twenty-one potential 29 host proteins that modulate ZIKV replication in an IFN-dependent manner, including the previously 30 known IFITM3 and LY6E. Further characterization contributed to delineate the spectrum of action of 31 these genes towards other pathogenic RNA viruses, including Hepatitis C virus and SARS-CoV-2. Our 32 data revealed that APOL3 acts as a proviral factor for ZIKV and several other related and unrelated 33 RNA viruses. In addition, we showed that MTA2, a chromatin remodeling factor, possesses potent 34 flavivirus-specific antiviral functions induced by IFN. Our work identified previously unrecognized 35 genes that modulate the replication of RNA viruses in an IFN-dependent manner, opening new 36 perspectives to target weakness points in the life cycle of these viruses.

Viruses are high on the list of global public health concerns, as illustrated by recent epidemics 42 caused by Ebola, Zika (ZIKV) and Nipah viruses, as well as by the ongoing SARS-CoV-2 pandemic.

The vast majority of these emerging RNA viruses have zoonotic origins and have recently crossed host 44 species barrier [1] . In order to establish itself in a host species, one of the first and most restrictive 45 barriers that a virus needs to overcome is the antiviral innate immune system. This response has evolved 46 to rapidly control viral replication and limit virus spread via detection of viral nucleic acids by pathogen 47 recognition receptors (PRRs) [2] . These PRRs can be membrane-associated, such as Toll-like receptor 48 (TLRs), or cytosolic, such as retinoic acid inducible gene I (RIG-I)-like receptor (RLRs). Upon binding 49 to viral nucleic acids, these PRRs interact with adaptor proteins and recruit signaling complexes. These 50 events lead to the expression of type I and type III interferons (IFNs). Secreted type I and type III IFNs 51 will then bind to their heterodimeric receptor, IFNAR1/IFNAR2 and IFN-λR1/IL-10R2, respectively, 52 and activate the canonical JAK/STAT pathway in infected and surrounding cells [3] . This activation 53 triggers the assembly of the interferon-stimulated gene 3 (ISGF3) complex (composed of STAT1, 54 STAT2 and IRF-9 proteins), which subsequently induces the expression of up to approximately 2000 55 IFN-stimulated genes (ISGs) [4, 5] , effectively establishing the antiviral state. ISGs comprise a core of 56 genes that are induced at high levels essentially in all cell types, as well as cell-type specific genes that 57 are the result of transcriptome remodeling [6, 7] , highlighting the importance of studying ISGs in 58 relevant cell types. Some of these ISGs have been well characterized. They directly block the viral life 59 cycle by targeting specific stages of virus replication, including entry into host cells, protein translation, 60 replication or assembly of new viral particles [4, 8] . Some ISGs are specific to a virus or a viral family, 61 while others are broad-spectrum. They can also be negative or positive regulators of IFN signaling and 62 thus facilitate, or not, the return to cellular homeostasis. However, the contribution of most ISGs to the 63 antiviral state remains poorly understood. has been linked to several neurological disorders, including Guillain-Barré syndrome (GBS), 78 meningoencephalitis, myelitis and congenital microcephaly, fetal demise and abortion [11] . Children 79 exposed to ZIKV in utero may present neurocognitive deficits, regardless of head size at birth. ZIKV 80 infection is now identified as a sexually-transmitted illness as well [12] . As all flaviviruses, ZIKV is an 81 enveloped virus containing a positive-stranded RNA genome of ~11 kb. Upon viral entry, the viral 82 genome is released and translated by the host cell machinery into a large polyprotein precursor. The 83 latter is processed by host and viral proteases into three structural proteins, including C (core), prM 84 (precursor of the M protein) and E (envelope) glycoproteins, and seven non-structural proteins (NS) 85 called NS1, NS2A, NS2B, NS3, NS4A, NS4B and NS5 [9] . The structural proteins constitute the viral 86 particle, while NS proteins coordinate RNA replication, viral assembly and modulate innate immune 87 responses.

The importance of IFN signaling in mediating host restriction of ZIKV is illustrated by the severe 89 pathogenicity in IFNAR1-/-and STAT2-/-but not in immunocompetent mice [13] [14] [15] . Moreover, the 90 Zika strain that is responsible for the recent epidemics has accumulated mutations that increase 91 neurovirulence via the ability to evade the immune response [16] . Microglial cells, which are the resident 92 macrophages of the brain, represent ZIKV targets and potential reservoirs for viral persistence [17] . 93 Moreover, they may play a role in ZIKV transmission from mother to fetal brain [18] and affect the 94 proliferation and differentiation of neuronal progenitor cells [19] . In order to comprehend the molecular 95 bases behind the efficacy of the IFN response to ZIKV replication, we set up a high throughput assay to 96 identify genes that are modulating viral replication in human microglial cells (HMC3) stimulated with 97 IFN.

A loss-of-function screen identified genes modulating ZIKV replication in IFN-stimulated 101 human microglial cells

We selected 386 human genes based on a gene signature defined by clustering and correlation of 103 expression with MX1, a well-described ISG, in a dataset of gene expression in primary human CD4 + T 104 cells (Cerboni et al., in preparation) (Table S1 ). 36% of the identified genes overlapped with previous 105 ISG libraries [20, 21] , ensuring that the screen would be simultaneously capable of identifying expected 106 hits and new genes of interest.

Pilot experiments were conducted to optimize IFN2 concentration and viral multiplicity of infection 108 (MOI) (Fig. S1A ). SiRNAs targeting IFNAR1 and IFNAR2 were used as positive controls. SiRNAs 109 against IFITM3, an ISG known to potently inhibit ZIKV replication in several human cell lines and 110 primary fibroblasts [23, 24] , were used as additional positive controls in these experiments (Fig. S1A ).

In cells transfected with siRNAs targeting IFNAR1, the level of infection was almost rescued to the 112 level of non-treated cells, in contrast to cells transfected with non-targeting siRNAs (Fig. S1A ). In cells 113 silenced for IFITM3 expression, the number of infected cells was partly restored to the level of non-114 treated cells (Fig. S1A) , which was expected since the antiviral state requires the concerted action of 115 numerous ISGs [20] .

To conduct the screens, cells were transfected with individual siRNA 6 hours post-seeding, treated 117 with IFN2 at day 2, infected with ZIKV at day 3 and fixed 24 hours later (Fig. 1A) . Percentages of 118 infected cells were determined by confocal analysis by measuring the number of cells expressing the 119 viral E protein (Fig. 1A) . Quality control, performed as described in the method section, confirmed 120 siRNA transfection efficiency by using siRNAs against KIF11, a protein essential for cell survival [22] , 121 and the efficacy of positive IFNAR1-and IFNAR2-siRNA targeting controls ( Fig. S1B-E) . The screen 122 was conducted in duplicate and analyzed using 2 approaches. The first approach took into consideration 123 the intensity of the E signal per cell while the second one identified infected cells based on the E 124 presence, independently of the intensity of the signal. The 2 analysis identified similar number of 125 infected cells (Fig. S1F ). Results were expressed as robust Z-scores for each siRNA (Fig. 1B, C) . Genes 126 were defined as hits when at least two over three of their robust Z scores had an absolute value superior 127 to 2 in the two replicates, in at least one of the analysis. The screen identified 9 antiviral genes and 12 128 proviral ones (Fig. 1D ). Some hits were previously described as modulators of ZIKV replication, such 129 as IFITM3 [23, 24] and LY6E [25] , thus validating our loss-of-function screening approach. These 130 twenty-one hits were selected for further validation. to histone deacetylase activity [28] . Ten times more viral RNA copies were recovered in cells silenced 146 for MTA2 expression than in control cells (Fig. 1E ) and four times more cells were positive for the viral 147 E protein (Fig. 1F ). These effects were comparable to the ones induced by IFNAR1 silencing (Fig. 1E 148 and F). Reducing the expression of the 3 other antiviral candidates (PXK, NMI and IFI16) had no 149 significant effect on ZIKV replication in these assays ( Fig. 1E and 1F ), suggesting that they may be 150 false positive candidates. Reducing the expression of LY6E, ISG15 and APOL3 significantly decreased 151 both viral RNA production and the number of cells positive for the E protein ( Fig. 1G and 1H) , validating 152 the pro-viral activities of these 3 candidates. The pro-viral function of USP18 and NAPA were also 153 validated since reducing their expression led to a significant reduction of the number of infected cells 154 as compared to control cells (Fig. 1H ). Reduced expression of ISG20 or CCND3 had no significant 155 effect on ZIKV replication ( Fig. 1G and 1H ). IRF2, which was identified as a pro-viral hit by the screen, in stimulated Huh-7.5 cells (Fig. S2A ). Since LY6E and APOL3 mRNA levels were under the limit of 172 detection of the assays in IFN2-treated Huh-7.5 cells (Fig. S2A ), they were excluded from further 173 analysis. As expected, reduced expression of IFNAR and IRF9 significantly enhanced HCV RNA yield 174 and the production of infectious particles in IFN2-treated Huh-7.5 cells, as compared to control cells 175 ( Fig. 2A, B) . Reduced expression of IFITM3, MTA2 or GPD2, which significantly enhanced ZIKV 176 replication in HCM3 cells (Fig. 1E , F), did not affect HCV RNA production ( Fig. 2A) . However, 177 surprisingly, their reduced expression triggered a significant decrease in the release of infectious HCV 178 particles as compared to control cells (Fig. 2B ). This suggests that they might favor a late stage of HCV 179 replication cycle. RT-qPCR analysis and titration assays identified USP18 and ISG15 as pro-HCV 180 factors in IFN2-treated Huh-7.5 cells (Fig. 2C, D) , validating previous results [32] [33] [34] [35] . Of note, HCV 181 RNA production and infectious particle release were significantly increased in cells with reduced NAPA 182 levels ( Fig. 2C, D) , suggesting that NAPA may exert an antiviral effect on HCV, while it was not the 183 case for ZIKV (Fig. 1G, H) . 184 SARS-CoV-2 replication was assessed in A549 alveolar epithelial cells expressing the viral receptor 185 ACE2 (A549-ACE2) by RT-qPCR and flow cytometry analysis using an antibody against the viral 186 protein Spike (S). Of note, silencing GBP3 in A549-ACE2 cells triggered cell death. RT-qPCR analyses 187 showed that all siRNA pools were reducing the expression of their respective targets in stimulated A549-188 ACE2 cells (Fig. S2B ). These analyses revealed the ability of IRF9 to act as an anti-SARS-CoV-2 gene 189 ( Fig. 2E, F) . Unexpectedly, GPD2 and IFITM3, which we identified as genes possessing anti-ZIKV 190 activities ( Fig. 1) , tended to behave like pro-viral genes in the context of SARS-CoV-2 infection ( (Table S1 ). Among these 5 APOLs, only APOL3 was identified as a facilitator of 214 ZIKV infection by our screen (Fig. 1) . We decided to test the ability of APOL1 to modulate ZIKV 215 replication since it was previously identified in a high-throughput overexpression screen as an ISG able 216 to increase YFV infection in STAT1 −/− fibroblasts and Huh-7cells [20].

Analysis of mRNA levels of APOL1 and APOL3 revealed that the genes were upregulated by 218 IFN2 treatment in HMC3 cells (Fig. 3A) . Both genes thus qualify as genuine ISGs in these cells. The 219 implication of APOL1 and APOL3 in ZIKV replication was investigated using loss-of-function 220 approaches. siRNA-silencing reduced the levels of APOL1 and APOL3 mRNAs by ~80% and ~85%, 221 respectively, when compared to cells expressing scrambled control siRNAs (Fig. 3B) . USP18, which is 222 known to negatively regulates the JAK-STAT pathway, and, as such, is a broad-spectrum pro-viral 223 factor [39] , was identified during our screen as a pro-ZIKV candidate in HMC3 cells (Fig. 1D ). Since 224 its pro-ZIKV function was validated in our system (Fig. 1H ), siRNAs specific for USP18 were used as 225 positive controls. siRNA-silencing reduced the abundance of USP18 mRNAs by ~80% when compared 226 to cells expressing control siRNAs (Fig. 3B ). Viral replication was assessed by flow cytometry by 227 measuring the number of cells positive for the viral protein E in cells silenced for APOL1, APOL3 or 228 USP18, treated or not with IFN2. Higher MOIs were used in IFN2-treated cells than in untreated 229 ones to compensate for its antiviral effects. As expected ( of the viral proteins NS5 and E were slightly decreased in IFN2-cells expressing reduced levels of 237 APOL1 or APOL3, compared to control cells ( Fig 3D) . Together, these results suggest that ZIKV 238 requires the expression of APOL1 and APOL3 for optimal replication in HMC3 cells. By contrast to 239 APOL1, the pro-viral action of APOL3 was dependent on IFN2-treatment.

To ensure that the APOL1-and APOL3-mediated modulation of viral replication was not restricted 

We tested whether APOL1 and APOL3 were active against DENV-2 or WNV, which are flaviviruses 252 closely related to ZIKV. HMC3 cells were treated or not with IFN2 and the MOIs were adapted to the 253 IFN2 treatment. Flow cytometry analysis using anti-E antibodies revealed that both DENV-2 and 254 WNV replication were significantly decreased in IFN2-treated cells silenced for APOL1 or APOL3 255 expression (Fig. 3I) . Reducing APOL1 and APOL3 expression in non-treated cells also significantly 256 reduced WNV replication (Fig. 3I) . Thus, APOL1/3 may well have flavivirus-specific proviral activities 257 since they seems to contribute to ZIKV, WNV and DENV replication (Fig. 3C , D, G, H and I) but not 258 to SARS-CoV-2 replication (Fig. 2G, H) . To further delineate the spectrum of proviral action of these 259 two genes, we tested the effect of APOL1/3 silencing on the replication of Vesicular Stomatitis virus 260 (VSV) and Measles virus (MeV), which are negative-strand RNA viruses belonging to the 261 Rhabdoviridae and Paramyxoviridae families, respectively. Experiments were performed with a MeV 262 strain modified to express GFP [41] . We also included in the analysis Modified Vaccinia Ankara virus 263 (MVA), a DNA virus belonging to the poxviridae family, that was engineered to express GFP (MVA-264 GFP). Flow cytometry analysis using an antibody against the viral protein G revealed that VSV was 265 highly dependent on APOL1 and APOL3 expression for efficient replication in IFN2-treated HMC3 266 cells (Fig. 3I ). Optimal replication of MeV-GFP in stimulated HMC3 cells also required APOL1 and 267 APOL3 expression (Fig. 3I ). APOL1 proviral activity was also observed in unstimulated cells (Fig. 3I ). 

We then performed experiments with a well-characterized PI4KB kinase inhibitor that decreases 294 PI(4)P expression [47] . We first analyzed by immunofluorescence the intensity of the PI(4)P signal in 295 cells treated for 24 h with different concentrations of the drug in HCM3 cells. The presence of the PI4KB 296 inhibitor triggered a dose-dependent decrease of the PI(4)P signal (Fig. 4C) . We then infected cells with 297 ZIKV in the presence of different concentration of the inhibitor. Since the effect of APOL3 on ZIKV 298 replication is dependent on IFN2 (Fig. 3) , the analysis was also performed in stimulated cells.

Coxsackie B3 virus (CVB3), an enterovirus that replicates in a PI(4)P-dependent manner, was used as 300 a positive control since its replication is sensitive to the drug [48] . As negative controls, we used cells 301 infected with WNV, whose replication is not affected by the PI4KB inhibitor [49] . As previously shown 302 in HeLa cells [48], a dose-dependent reduction of the number of cells positive for the CVB3 viral protein 303 1 (VP1) was triggered by the inhibitor treatment (Fig. 4D ). As shown previously in monkey cells [49], 304 WNV replication was unaffected by the PI4KB inhibitor in HCM3 cells (Fig. 4D ). ZIKV protein 305 production was not sensitive to the treatment with the PI4KB kinase inhibitor, independently of the 306 presence of IFN2 (Fig. 4D) . These experiments suggest that the pro-viral activities of APOL1 and 307 APOL3 are not related to their interaction with PI4KB or PI(4)P in microglial cells. 308 309 MTA2 restricts ZIKV replication in IFN2-stimulated human cells.

MTA2 was identified in our screen as a gene with potent anti-ZIKV activities (Fig. 1) 

Assessment of viral replication by RT-qPCR revealed that cell-associated viral RNA yields were 325 significantly higher in IFN-treated HMC3 cells silenced for MTA2 expression, as compared to controls 326 cells (Fig. 5E ). This is in line with previous results (Fig. 1E) . Cytometry analysis using anti-E antibodies 327 confirmed that MTA2 anti-ZIKV activities were dependent on the presence of IFN in HCM3 cells (Fig.   328   5F ). Since MTA2 is not an ISG in HMC3 cells (Fig. 5A) , these results suggest that MTA2 may require 329 an active IFN signaling to exert its anti-ZIKV activities in these cells. As in HMC3 cells, reduced 330 expression of MTA2 triggered a significant increase of intracellular viral RNA production in stimulated 331 Huh-7 cells (Fig. 5G ). Reducing MTA2 expression had a more pronounced effect on the percentage of 332 infected cells that reducing IFNAR1 expression in stimulated Huh-7 cells (Fig. 5H ). Albeit to a lesser 333 extent than in stimulated cells, MTA2 anti-ZIKV activity was also observed in non-stimulated Huh-7 334 cells by flow cytometry and RT-qPCR analysis ( Fig. 5G and H) .

The effect of MTA2 on viral protein production was further assessed by Western blot analysis using 336 anti-E and anti-NS5 antibodies in stimulated and unstimulated HMC3 and Huh-7 cells. These 337 experiments validated further the efficacy of the siRNAs against MTA2 in both cell lines ( Fig. 5I and   338 5J). Expression of the viral proteins NS5 and E were increased in stimulated HMC3 and Huh-7 cells 339 expressing reduced levels of MTA2 or IFNAR1, as compared to control cells ( Fig. 5I and 5J) . In 340 agreement with the flow cytometry analysis (Fig. 5H) , MTA2 anti-ZIKV activity was less dependent of 341 IFN-treatment in Huh-7 cells than in HMC3 cells ( Fig. 5I and 5J ). As observed in flow cytometry 342 analysis (Fig. 5H ), MTA2 effect on viral protein production was more potent than the one of IFNAR1 343 in stimulated Huh-7 cells (Fig. 5J ).

These results represent the first evidence of the ability of MTA2 to restrict the replication of any 345 virus. As a first hint towards deciphering the function of MTA2 in ZIKV infection, we studied its cellular 346 localization by indirect immunofluorescence in uninfected and infected Huh-7 cells. Besides its 347 chromatin-remodeling activities, MTA2 has been reported to associate with the centrosome in non-348 stimulated human cells [50] . These experiments revealed that MTA2 localized in the nucleus of non-349 infected cells (Fig. S4) . The specificity of MAT2 labeling was confirmed in Huh-7 cells silenced for 350 MTA2 expression (Fig. S4A) . No evident change in MTA2 localization was observed in ZIKV-infected 351 cells (Fig. S4B) , which were marked with antibodies specific for the viral NS5 protein, which is known 352 to localize in the nucleus [51, 52] . These results suggest that MTA2-mediated effect on ZIKV replication 353 may be linked to its chromatin-remodeling activity. 354 355 MTA2 restricts YFV and WNV replication in IFN2-stimulated Huh-7 cells.

We tested whether MTA2 was active against WNV and YFV in Huh-7 cells, which are permissive 357 to these 2 flaviviruses. As in previous experiments, higher MOIs were used in the presence of IFN2.

Cytometry analysis revealed that MTA2 silencing significantly enhanced the replication of these 2 359 flaviviruses in an-IFN dependent manner ( Fig. 6A and 6B) , indicating that MTA2 antiviral activity is 360 broader that ZIKV. We then tested the effect of MTA2 silencing on the replication of VSV and MeV in 361 Huh-7 cells. Reduced expression of MTA2 decreased the number of cells infected with VSV and MeV 362 ( Fig. 6C and 6D) , independently of the IFN stimulation. This is consistent with the pro-HCV activity of 363 MTA2, as measured by titration in stimulated Huh-7.5 cells (Fig. 2B) 

We validated the role of 5 hits as genes contributing to an optimal ZIKV replication in stimulated 395 HMC3 cells using RT-qPCR and flow cytometry analysis: LY6E, USP18, ISG15, APOL3 and NAPA. [69, 70] . Hijacking individual ISG for promoting their replication is one of them. This hypothesis may 419 explain why, to our surprise, our screen recovered more pro-viral genes that antiviral ones.

Our results identified APOL3 and APOL1 as ISGs required for optimal ZIKV replication in HMC3 

We formulated the hypothesis that APOL1 and APOL3 pro-viral activities could be linked to their 435 ability to bind to anionic phospholipids, including several phosphoinositides, in particular PI (4) 

For quality control purposes, the number of cells in each condition of the 2 replicates were analysed.

We observed an expected distribution of the number of cells in 3 fields with a median close to 1000 555 cells per well for the two replicates (Fig. S1B ). The number of cells per condition was slightly higher in 556 the first replicate than in the second one. However, the R² coefficient of determination of the linear 557 regression was close to 0.7 (Fig. S1C) , indicating that the reproducibility of the experiment was correct.

As expected [22] , siRNAs against KIF11 were lethal, validating the transfection protocol (Fig. S1B, C) .

The number of cells expressing the viral protein E distributed as predicted, with a median close to 15% 560 for the 2 screens (Fig. S1D ). As expected from pilot experiments (Fig. S1A) , siRNAs targeting IFNAR-561 1 and -2 rescued ZIKV replication in IFN-treated cells (Fig. S1D, E) . The reproducibility of the infection 562 status of the cells between the 2 screens, with a R² greater than 0.8, was satisfactory (Fig. S1E ).

In the first analysis, data were processed using a software developed internally at the Biophenics 566 platform. For hit identification, the robust Z-score method was used under the assumption that most 567 siRNAs are inactive against ZIKV and can serve as controls [87, 88] . Raw values were log-transformed 568 for cell count only to make the data more symmetric and close to normal distribution. In order to correct Antiviral  effect  IRF9  IFITM3  MTA2  PXK  GPD2  C1R  XCL1  NMI  IFI16   Proviral  effect  LY6E  USP18  ISG15  APOL3  GBP3  NAPA  NADK  ISG20  IRF2 

UK) for generously providing the podocytes

cells; Cinzia Traboni (IRBM

Institut Pasteur) 663 for the CVB3 Nancy strain; N. Escriou (Institut Pasteur) for the VSV Indiana strain and anti-VSV-G 664 antibodies

French Polynesia) 665 for the ZIKV-PF13 strain; L. Hermida and G. Enrique Guillen Nieto from the Centro de Ingeniería 666

Cuba, for the DENV-2 strain Malaysia SB8553

Japan) for pJFH1 HCV cDNA

Germany) for 668 pJFH1-2EI3-adapt cDNA; the French National Reference Centre for Respiratory Viruses

Institut Pasteur (France) and headed by S. van der Werf for providing the historical SARS-CoV-2 viral 670 strains

Estonia) for anti-ZIKV NS5 antibodies; P. Desprès (Université 671 de la Réunion, PIMIT) for 4G2 hybridoma cells; H. Mouquet (Institut Pasteur) for anti-SARS-CoV-2 S 672 antibodies

USA) for the plasmid 673 encoding the full-length Zika MR766 genome

France) for advice concerning the siRNA screening work

Belgium) for APOL1 and APOL3 plasmids and for stimulating APOL-focused 677 discussions. We are grateful to the members of our laboratories for helpful discussions and technical 678 advice. Finally, we thank Emeline Perthame (Bioinformatics and Biostatistics HUB, Institut Pasteur) 679 for her help in statistical analysis

Urgence COVID-19' 681 fundraising campaign of Institut Pasteur

EMBO Young Investigator programm bringing fund and the Agence Nationale de la Recherche sur le 684 sida et les hépatites virales (ANRS-CSS12-2019-2). The BioPhenics laboratory is supported by Institut 685

PICT-IBiSA) and part of ChemBioFRance and France-BioImaging infrastructure 686 supported by the French National Research Agency (ANR-10-INSB-04, «Investments for the future

The funders had no role in study design, 688 data collection and analysis

Declaration of interests. The authors declare no competing interests

Human viruses: 695 discovery and emergence

Stimulation of Innate Immunity by Host and Viral RNAs

Decoding type I and III interferon signalling 700 during viral infection

Interferome 704 v2.0: an updated database of annotated interferon-regulated genes

Combinatorial flexibility of cytokine function during human T helper cell 708 differentiation

Parsing the 710 interferon transcriptional network and its disease associations

Interferon-Stimulated Genes: A Complex 713 Web of Host Defenses

The continued threat of emerging flaviviruses

Usutu virus: an emerging flavivirus in 718

The emergence of Zika virus and its new clinical 720 syndromes

Zika virus 722 infects human testicular tissue and germ cells

725 Characterization of a Novel Murine Model to Study Zika Virus

A 731 novel Zika virus mouse model reveals strain specific differences in virus pathogenesis and 732 host inflammatory immune responses

An 735 evolutionary NS1 mutation enhances Zika virus evasion of host interferon induction

Axl Mediates ZIKA Virus Entry in Human Glial Cells and Modulates Innate Immune 739 Responses

Role of microglia in the 741 dissemination of Zika virus from mother to fetal brain

A 747 diverse range of gene products are effectors of the type I interferon antiviral response

Systematic identification of type I 750 and type II interferon-induced antiviral factors

Identification of cytoskeleton-associated proteins essential for lysosomal stability and 753 survival of human cancer cells

Inhibit Zika Virus Replication

Zika 758 virus induces massive cytoplasmic vacuolization and paraptosis-like death in infected cells

Flavivirus internalization is regulated by a size-dependent 761 endocytic pathway

Interferon Regulatory Factor 9 Structure and Regulation

IFITM proteins-cellular inhibitors of viral 766 entry

Highly permissive cell lines for subgenomic and 770 genomic hepatitis C virus RNA replication

Regulating 772 intracellular antiviral defense and permissiveness to hepatitis C virus RNA replication 773 through a cellular RNA helicase, RIG-I

Hepatitis C virus-mediated enhancement of 776 microRNA miR-373 impairs the JAK/STAT signaling pathway

Protease 18 Promotes Hepatitis C Virus Production by Increasing Viral Infectivity. Mediators 780 of Inflammation

Silencing of 782 USP18 Potentiates the Antiviral Activity of Interferon Against Hepatitis C Virus Infection

Hepatitis C virus reveals a novel early control in acute immune response

ISG15, a 788 ubiquitin-like interferon-stimulated gene, promotes hepatitis C virus production in vitro: 789 implications for chronic infection and response to treatment

LY6E 792 impairs coronavirus fusion and confers immune control of viral disease

The function of apolipoproteins L

The apolipoprotein L gene 797 cluster has emerged recently in evolution and is expressed in human vascular tissue

UBP43 is a 800 novel regulator of interferon signaling independent of its ISG15 isopeptidase activity

Zika Virus Infection of the Human Glomerular Cells: Implications for 803 Viral Reservoirs and Renal Pathogenesis

A 806 Molecularly Cloned Schwarz Strain of Measles Virus Vaccine Induces Strong Immune 807 Responses in Macaques and Transgenic Mice

Terminal Variants May Trigger Kidney Disease through Interference with APOL3 811 Control of Actomyosin

The Great Escape: how phosphatidylinositol 4-kinases and PI4P promote 813 vesicle exit from the Golgi (and drive cancer)

Strating JR, van Kuppeveld FJ. Viral rewiring of cellular lipid metabolism to create 819 membranous replication compartments

Conserved role for Gga proteins 822 in phosphatidylinositol 4-kinase localization to the trans-Golgi network

Design and Structural Characterization of Potent and Selective Inhibitors of 826 Phosphatidylinositol 4 Kinase IIIβ

Viral 829 reorganization of the secretory pathway generates distinct organelles for RNA replication

virus replication requires fatty acid synthesis but is independent on 833 phosphatidylinositol-4-phosphate lipids

Chromatin Remodeling 836 Proteins Interact with Pericentrin to Regulate Centrosome Integrity

Nuclear localization of Zika 839 virus NS5 contributes to suppression of type I interferon production and response

Zika virus NS5 nuclear accumulation is protective of protein degradation 842 and is required for viral RNA replication

Identification of five 845 interferon-induced cellular proteins that inhibit west nile virus and dengue virus infections

Expressional screening of interferon-stimulated genes for antiviral activity against hepatitis C 849 virus replication

A 854 CRISPR screen identifies IFI6 as an ER-resident interferon effector that blocks flavivirus 855 replication

PARP12 suppresses Zika virus 857 infection through PARP-dependent degradation of NS1 and NS3 viral proteins

C19orf66 interrupts Zika virus 860 replication by inducing lysosomal degradation of viral NS3. PLOS Neglected Tropical 861 Diseases

A Viperin Mutant Bearing the K358R 863 Substitution Lost its Anti-ZIKA Virus Activity

Cell-type-and region-specific 866 restriction of neurotropic flavivirus infection by viperin

ISG15 in antiviral immunity and beyond

Multiple functions of USP18

Mechanistic insights into the 873 SNARE complex disassembly

UVRAG is required for 878 virus entry through combinatorial interaction with the class C-Vps complex and SNAREs

Glycerol 881 phosphate shuttle enzyme GPD2 regulates macrophage inflammatory responses

Interference with the production of infectious viral particles and bimodal inhibition of 887 replication are broadly conserved antiviral properties of IFITMs

Viral Evasion Strategies in Type I IFN Signaling -A 892 Summary of Recent Developments

The innate immune factor apolipoprotein L1 897 restricts HIV-1 infection

Basal expression of interferon regulatory factor 1 drives intrinsic hepatocyte resistance to 900 multiple RNA viruses

Defining how viruses 902 manipulate lipid phosphoinositides through activation of PI4P kinases to mediate viral 903 replication

The Endoplasmic Reticulum 905 Provides the Membrane Platform for Biogenesis of the Flavivirus Replication Complex

Toxoplasma Effector 908 Recruits the Mi-2/NuRD Complex to Repress STAT1 Transcription and Block IFN-γ-909

Nusinzon I, Horvath CM. Interferon-stimulated transcription and innate antiviral 912 immunity require deacetylase activity and histone deacetylase 1

Histone 915 deacetylase inhibitors suppress IFNα-induced up-regulation of promyelocytic leukemia 916 protein

A 918 conditionally immortalized human podocyte cell line demonstrating nephrin and podocin 919 expression

Growth of human hepatoma 921 cells lines with differentiated functions in chemically defined medium

Syncytia 924 formation by SARS-CoV-2-infected cells

RIG-I 927 Recognizes the 5' Region of Dengue and Zika Virus Genomes

NS2 930 proteins of GB virus B and hepatitis C virus share common protease activities and membrane 931 topologies

Production of 933 infectious hepatitis C virus in tissue culture from a cloned viral genome

Cell culture 936 adaptation of hepatitis C virus and in vivo viability of an adapted variant

Differential 939 regulation of the Wnt/beta-catenin pathway by hepatitis C virus recombinants expressing core 940 from various genotypes

Statistical practice in high-942 throughput screening data analysis

Statistical Methods for Analysis of High-Throughput RNA Interference Screens

CellProfiler 3.0: Next-generation image processing for biology

Reverse 950 genetic system, genetically stable reporter viruses and packaged subgenomic replicon based 951 on a Brazilian Zika virus isolate

953 Rescue of the 1947 Zika Virus Prototype Strain with a Cytomegalovirus Promoter-Driven 954 cDNA

A, E) APOL1 mRNA and APOL3 mRNA abundance were 993 quantified by RT-qPCR analysis in HMC3 cells or podocytes treated or not with IFN2 (200U/mL) for 994 24 hours and expressed as copy numbers per g of total cellular RNA. HMC3 cells (B) or podocytes (F) 995 were transfected with pool of 3 siRNAs targeting APOL1, APOL3 and USP18 mRNAs or with non-996 targeting (NT) control siRNAs. The relative amounts of APOL1, APOL3 and USP18 mRNAs were 997 determined by RT-qPCR analysis and were normalized to that of GAPDH mRNA

HMC3 cells were infected at a MOI of 2 and podocytes 1001 at a MOI of 1. The percentages of cells that were positive for the viral E proteins were determined by 1002 flow cytometric analysis. Data are expressed relatively to the siRNA NT control of each experiment. 1003 HMC3 cells (D) or podocytes (H) were treated with IFN2 (200U/mL), transfected with the indicated 1004 siRNAs pools and subjected to Western blotting analysis with antibodies against the indicated proteins. 1005 (I) HMC3 cells were transfected with the indicated siRNAs pools, treated with IFNa2 (200U/mL) for 1006 24 hours and infected with the indicated viruses for 18 to 24 hours, at the MOI indicated in the MM 1007 section. The percentages of the cells positive for viral proteins or GFP were determined by flow 1008 cytometric analysis

APOL1 and APOL3 promote viral replication independently of their interaction with 1012 phosphoinositides. HMC3 cells were transfected with GFP-tagged versions of APOL1 and APOL3

Thirty hours later, they were stained with antibodies recognizing GM130 (A) or PI4KB (B) and with 1014

Images are representative of numerous observations over 2 independent 1015 experiments. The white arrow shows an APOL1-GFP-positive vesicle. (C) HCM3 cells were treated 1016 with different concentrations of the PI4KB inhibitor and were stained for PI(4)P. (D) HMC3 cells treated 1017 with different doses of PI4KB inhibitor were infected with CVB3, WNV or ZIKV, in the presence or 1018 absence of IFN2 (200U/mL). The percentages of the cells positive for viral proteins were determined 1019 by flow cytometric analysis

01,***p<0.001, one-way ANOVA

Figure 5. MTA2 restricts ZIKV replication in IFN2-stimulated cells. HMC3

transfected with pool of 3 siRNAs targeting MTA2 or IFNAR1 mRNAs 1024 or non-targeting (NT) control siRNAs, treated or not with IFN2 (100U/mL) for 24 hours, and infected 1025 with ZIKV (MOI of 1 for HMC3 cells, MOI of 5 for Huh-7 cells) for 24 hours. (A-D) The relative 1026 amounts of MTA2 and IFNAR1 mRNAs were determined by RT-qPCR analysis and normalized to that of GAPDH mRNA and siRNA-NT without IFN

RT-qPCR and expressed as genome equivalents (GE) per µg of total cellular RNA. (F, H) Number of 1029 infected cells was assessed by staining of viral protein E and flow cytometry analysis. (I, J) Cells were 1030 treated with IFN2 (200U/mL) or left untreated, transfected with the indicated siRNAs pools and 1031 subjected to Western blotting analysis with antibodies against the indicated proteins. Data are means ± 1032 SD of three independent experiments

Effect of reduced expression of MTA2 on the replication of YFV, WNV, VSV and MeV-1035

200U/mL) for 24 hours and infected with WNV (A) or YFV (B) for 24 hours, at the MOIs indicted in 1037 the MM section. The percentages of the cells positive for viral protein Env was determined by flow 1038 cytometric analysis. HMC3 cells were transfected with the indicated siRNAs pool

1040 at the MOIs indicted in the MM section. The percentages of the cells positive for viral protein G or GFP 1041 were determined by flow cytometric analysis. Data are expressed relatively to the siRNA NT control of 1042 each experiment. Data are means ± SD of three or four independent experiments

Quality control and reproducibility of the screens. (A) HMC3 cells were transfected with 1048 either pool of 3 siRNAs against IFNAR or IFITM3 or non-targeting (NT) siRNAs, treated with IFN2 1049 (1000U/mL) for 24 hours and infected with ZIKV PF13 at a MOI of 7 PFU/cell for 24 hours

B) Distribution of the "number of cells per 3 fields" parameter for each screen

The values of the control wells (cells transfected with siRNA targeting KIF11) are shown in dark gray

Representation of the number of cells per 3 fields of screen 1 as a function of the screen 2. The green 1054 line represents the linear regression as compared to the expected perfect correlation

The 1056 values of the control conditions (cells transfected with siRNAs targeting IFNAR1 or IFNAR2) are 1057 shown in dark grey. (E) Representation of the percentage of infected cells per well of screen 1 as a 1058 function of screen 2, as identified by the first analysis. The green line represents the linear regression as 1059 compared to the expected perfect correlation (dotted black line). (F) Representation of the percentage 1060 of infected cells per well in the analysis 1 as a function of analysis 2. The green line represents the linear 1061 regression. (G) HMC3 cells were transfected with pools of 3 siRNAs targeting the indicated genes or Declaration of interests ☒ The authors declare that they have no known competing financial interests or personal relationships

☐The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Nolwenn Jouvenet

Declaration of Interest Sample CRediT author statement

Sarah Lesage: Investigation, Conceptualization, Methodology, Validation, Formal analysis, Writing -Review & Editing, Visualization. Maxime Chazal: Investigation, Conceptualization, Methodology, Validation, Formal analysis, Writing -Review & Editing, Visualization. Guillaume Beauclair: Conceptualization, Software, Writing -Review & Editing, Visualization. Damien Batalie: Investigation, Conceptualization. Silvia Cerboni: Conceptualization, Methodology, Resources. Elodie Couderc: Investigation, Writing -Review & Editing. Aurianne Lescure: Investigation, Conceptualization, Methodology, Resources, Writing -Review & Editing. Elaine Del Nery

Annette Martin: Conceptualization, Methodology, Validation, Formal analysis, Writing -Review & Editing, Visualization, Supervision, Funding acquisition

with non-targeting (NT) control siRNAs. The relative abundances of the mRNAs of the candidate genes 1063 were determined by RT-qPCR analysis and were normalized with respect to GAPDH mRNA level.

They are expressed relatively to abundance in cells transfected with NT siRNAs set at 1. Data are means 1065 ± SD of three or four independent experiments. ND: not determined due to mRNA levels below assay 1066 threshold. The samples are the same than in Fig. 1E