Letter https://doi.org/10.1038/s41586-019-0955-3

MHC class II proteins mediate cross-species entry of bat influenza viruses Umut Karakus1,17, Thiprampai Thamamongood2,3,4,5,17, Kevin Ciminski2,3, Wei Ran2,3, Sira C. Günther1, Marie O. Pohl1, Davide Eletto1, Csaba Jeney6, Donata Hoffmann7, Sven Reiche8, Jan Schinköthe8, Reiner Ulrich8, Julius Wiener9, Michael G. B. Hayes10, Max W. Chang10, Annika Hunziker1, Emilio Yángüez1, Teresa Aydillo11,12, Florian Krammer11, Josua Oderbolz13, Matthias Meier9, Annette Oxenius13, Anne Halenius2,3, Gert Zimmer14,15, Christopher Benner10, Benjamin G. Hale1, Adolfo García-Sastre11,12,16, Martin Beer7, Martin Schwemmle2,3,18* & Silke Stertz1,18*

Zoonotic influenza A viruses of avian origin can cause severe disease To identify receptor candidates for bat IAV, we performed tran- in individuals, or even global pandemics, and thus pose a threat to scriptional profiling on three cell lines that are susceptible to bat IAV human populations. Waterfowl and shorebirds are believed to be (Madin–Darby canine kidney II (MDCKII) clone no. 1, and human the reservoir for all influenza A viruses, but this has recently been glioblastoma (U-87MG) and lung cancer (Calu-3) cell lines), and on challenged by the identification of novel influenza A viruses in three cells lines that are not susceptible (MDCKII clone no. 2, and bats1,2. The major bat influenza A virus envelope glycoprotein, human adenocarcinomic alveolar basal epithelial (A549) and human haemagglutinin, does not bind the canonical influenza A virus glioblastoma (U-118MG) cell lines). The susceptibility of each cell receptor, sialic acid or any other glycan1,3,4, despite its high sequence line was characterized by two different H18- or H18N11-pseudotyped and structural homology with conventional haemagglutinins. virus-like particle (VLP) assays10,11 (Extended Data Fig. 1a, b). This functionally uncharacterized plasticity of the bat influenza A Enrichment filtering for transcripts that encode membrane-integral virus haemagglutinin means the tropism and zoonotic potential of or membrane-associated proteins identified ten genes that were highly these viruses has not been fully determined. Here we show, using expressed in all H18-susceptible cells (Fig. 1a, Extended Data Fig. 1c–e, transcriptomic profiling of susceptible versus non-susceptible cells Supplementary Table 1). Five of these ten candidates encode MHC-II- in combination with genome-wide CRISPR–Cas9 screening, that the associated proteins—namely, the α- and β-chains of both HLA-DR and major histocompatibility complex class II (MHC-II) human leukocyte HLA-DQ, and the MHC-II-associated invariant chain (CD74), which antigen DR isotype (HLA-DR) is an essential entry determinant for is required for HLA processing and transport12,13. We tested the effect bat influenza A viruses. Genetic ablation of the HLA-DR α-chain of downregulating expression of the candidates by RNA interference on rendered cells resistant to infection by bat influenza A virus, whereas H18-mediated VLP entry. We also included HLA-DP as an additional ectopic expression of the HLA-DR complex in non-susceptible MHC-II receptor. As functional cell-surface HLA-DR, HLA-DQ and cells conferred susceptibility. Expression of MHC-II from different HLA-DP complexes require both α- and β-chains to be expressed, we bat species, pigs, mice or chickens also conferred susceptibility to downregulated the α-chain of each complex. Knockdown of CD74 or infection. Notably, the infection of mice with bat influenza A virus HLA-DRA (the α-chain of dimeric HLA-DR)—but not of other HLA resulted in robust virus replication in the upper respiratory tract, genes or any other candidate—markedly reduced entry mediated by whereas mice deficient for MHC-II were resistant. Collectively, our H18 (Fig. 1b). In addition, cell-surface levels of HLA-DR correlated with data identify MHC-II as a crucial entry mediator for bat influenza A susceptibility (Extended Data Fig. 1f). To obtain independent evidence viruses in multiple species, which permits a broad vertebrate tropism. for genes involved in entry mediated by H18, we performed a genome- Influenza A viruses (IAVs) bind to host cells via haemagglutinin, a wide CRISPR–Cas9 screen (Fig. 1c). We introduced single-guide RNAs trimeric virion glycoprotein. Haemagglutinins of all non-bat-derived (sgRNAs) into U-87MG cells that express Cas9, and infected these cells IAVs recognize terminal sialic acids on glycoproteins or glycolipids, with vesicular stomatitis virus (VSV) that expresses H18 with a polyba- with no known exception. Avian IAVs bind sialic acids that display sic cleavage site (VSV-H18) in place of VSV-G (Extended Data Fig. 2). an α2,3-linkage to the penultimate sugar, whereas human IAVs pref- sgRNA sequences from surviving cells revealed several candidates, erentially recognize the α2,6 linkage5–7. Receptor binding is followed including four transcription factors that are known to regulate MHC-II by endocytic uptake and pH-mediated fusion of viral and endosomal expression (RFXANK, RFX5, CIITA and RFXAP) and HLA-DRB1 (the membranes8. The discovery of two novel IAV genomes (designated β-chain of HLA-DR) (Fig. 1c, Supplementary Table 2). These data sug- H17N10 and H18N11) in Central and South American bat species gest that HLA-DR is an entry determinant of H18N11 (Fig. 1d). expanded the known host range of IAVs, and raised questions about To specify the role of HLA-DR in bat IAV entry, we generated two their zoonotic potential and fundamental biology1,2. Although these clones of U-87MG HLA-DRA knockout cells (U-87MG KO no. 1 and bat-derived IAVs resemble classical IAVs in many ways, their haemag- no. 2). Upon knockout, HLA-DR surface expression was reduced to glutinins (H17 and H18) are unable to bind sialic acids, and thus chal- background (Fig. 2a). Notably, the HLA-DRA knockout cells were resist- lenge previous characterizations of this binding as central to IAVs1,3,4. ant to infection with H18- or H17-pseudotyped VLPs, VSV-H18, and The bat IAV receptor has remained unknown but is speculated to be VSV-H17 (Fig. 2b–d, Extended Data Fig. 3a, b), which suggests that the proteinaceous9. requirement for HLA-DR is shared between both subtypes of bat IAV,

1Institute of Medical Virology, University of Zurich, Zurich, Switzerland. 2Institute of Virology, Medical Center University of Freiburg, Freiburg, Germany. 3Faculty of Medicine, University of Freiburg, Freiburg, Germany. 4Spemann Graduate School of Biology and Medicine, University of Freiburg, Freiburg, Germany. 5Faculty of Biology, University of Freiburg, Freiburg, Germany. 6Department of Microsystems Engineering – IMTEK, University of Freiburg, Freiburg, Germany. 7Institute of Diagnostic Virology, Friedrich-Loeffler Institut, Greifswald-Insel Riems, Germany. 8Department of Experimental Animal Facilities and Biorisk Management, Friedrich-Loeffler Institut, Greifswald-Insel Riems, Germany. 9Helmholtz Pioneer Campus, Helmholtz Zentrum Munich, Neuherberg, Germany. 10Department of Medicine, University of California, San Diego, CA, USA. 11Department of Microbiology, Icahn School of Medicine at Mount Sinai, New York, NY, USA. 12Global Health and Emerging Pathogens Institute, Icahn School of Medicine at Mount Sinai, New York, NY, USA. 13Institute of Microbiology, ETH Zurich, Zurich, Switzerland. 14Division of Virology, Institute of Virology and Immunology, Mittelhäusern, Switzerland. 15Department of Infectious Diseases and Pathobiology, Vetsuisse Faculty, University of Bern, Bern, Switzerland. 16Department of Medicine, Division of Infectious Diseases, Icahn School of Medicine at Mount Sinai, New York, NY, USA. 17These authors contributed equally: Umut Karakus, Thiprampai Thamamongood. 18These authors jointly supervised this work: Martin Schwemmle, Silke Stertz. *e-mail: [email protected]; [email protected]

7 MARCH 2019 | VOL 567 | NATURE | 109 RESEARCH Letter

a Calu-3 b effect. To probe the interaction between HLA-DR and H18, we used a siRNA: polykaryon assay and observed that H18-mediated polykaryon forma- HLA-DRA CD74 HLA-DRAHLA-DPHLA-DQAA ITGA7 B3GALNT1B3GNT5GPRC5BSORBS2 MDCKII no. 1 1,093 HLA-DRB1 No. 1 HLA-DQA1 tion strictly depended on HLA-DR expression (Fig. 2h, i). Furthermore, 42 13 HLA-DQB1 CD74 No. 2 H18-dependent—but not H1-mediated—polykaryon formation could 8 10 B3GALNT1 206 B3GNT5 No. 3 be blocked by an MHC-II-specific antibody (Fig. 2h, i, Extended Data ITGA7 Fig. 6a), which strongly suggests that H18 interacts with HLA-DR. The 485 GPRC5B No. 4 SORBS2 interaction between H18 and HLA-DR was confirmed by proximity U-87MG % ligation assay (Extended Data Fig. 6b, c) and emulsion coupling (an c ޒ10 70 >130 10–15 RFXANK approach for protein–protein interactions based on digital-droplet RFX5 d CIITA PCR) (Fig. 2j, Extended Data Fig. 6d, e). 10–11 HLA-DRB1 RFXAP CIITA LAMTOR4 We aimed to render non-susceptible human HEK293T cells suscep- 10–7 EIF3H CBFB RFXANK NFYA tible to bat IAV by ectopic expression of HLA-DR. HEK293T cells were

RRA score RPS11 PRR21 NFYB transfected with plasmids that encode the α- and β-chains of HLA-DR –3

10 RFX5 RFXAP CREB NFYC HLA-DRB1 HLA-DRA (individually or in combination), or the transcriptional regulator CIITA. MHC-II gene locus 0 5,000 10,000 15,000 As expected, surface expression of HLA-DR was detected upon trans- Genes fection of both HLA-DRA and HLA-DRB1, and in CIITA-transfected Fig. 1 | Identification of HLA-DR as an entry determinant for bat IAV. cells (Fig. 3a). Only HEK293T cells that express ectopic HLA-DR (trans- a, Transcriptional profiles of MDCKII clone no. 1, Calu-3 and U-87MG fected with both chains of HLA-DR) or endogenous HLA-DR (induced cells, which are susceptible to H18-pseudotyped VLP infection, were by transfection with CIITA) were susceptible to H18-pseudotyped VLPs compared to those of non-susceptible MDCKII clone no. 2, A549 and (Fig. 3b) or VSV-H18 (Extended Data Fig. 7a). Because pigs and chickens U-118MG cells, respectively. The two-group comparison—each consisting have a central role in zoonotic transmission of conventional IAV, we over- of n = 4 independent samples—was performed using a gene-wise expressed HLA-DR homologues from these species and found that they negative binomial generalized linear model with quasi-likelihood tests, also support H18-mediated infection (Fig. 3c, d, Extended Data Fig. 7b). implemented in EdgeR. Adjustments for multiple comparisons were made As bats are likely to be the natural host of bat IAV, we aimed to con- by applying the Benjamini–Hochberg method. The numbers represent genes in the H18N11-susceptible cell lines that are significantly (P ≤ 0.01) fer susceptibility to H18-mediated infection by ectopic expression of HLA-DR homologues from different bat species. Because no MHC-II upregulated (fold change (expressed as log2(transcript level susceptible cell line/non-susceptible cell line)) ≥ 1), and filtered according to the Gene sequences from bat species that have tested positive for bat IAV RNA are Ontology annotations GO:0016020 (membrane), GO:0016021 (integral available, we generated expression constructs for HLA-DR homologues component of membrane) and GO:0005886 (plasma membrane). The ten published for three other bat species. Notably, expression of HLA-DR genes from MDCKII clone no. 1, Calu-3 and U-87MG cells that overlap homologues from Eptesicus fuscus (big brown bat), Myotis lucifugus (little are listed; the canid orthologues of HLA-DRA, HLA-DRB1, HLA-DQA1 brown bat) and Pteropus alecto (black flying fox) rendered cells suscep- and HLA-DQB1 are DLA-DRA, DLA-DRB1, DLA-DQA1 and DLA-DQB1, tible to H18-VLPs and VSV-H18 (Fig. 3c, d, Extended Data Fig. 7b). respectively. b, U-87MG cells were transfected with four different small HLA-DR homologues from bats and pigs (Sus scrofa), but not chickens interfering (si) RNAs (numbered 1 to 4) that target the indicated genes. At (Gallus gallus), also supported H17-mediated infection (Fig. 3e, Extended 48 h after siRNA transfection, cells were infected with luciferase-encoding VLPs, pseudotyped with H1N1 or H18, for 90 min at 37 °C. At 48 h p.i., Data Fig. 7c), which can be partially explained by the lower infection rate luciferase signals were measured and normalized to samples transfected of H17-VLPs relative to H18-VLPs (Fig. 2b, d). A bat kidney cell line with control siRNA. H18-dependent entry efficiency was further derived from Sturnira lilium (little yellow-shouldered bat), a putative host normalized to H1N1 entry efficiency, and relative values are shown as a of bat IAV1, has recently become available14. Genomic sequence infor- continuum of blue (low entry) to light red (high entry). c, Candidate genes mation has yet to be reported for this species, so we cloned the transcrip- identified by CRISPR–Cas9 forward screening. Cas9-expressing U-87MG tional regulator CIITA from kidney cells derived from P. al ec to (hereafter, cells transduced with a lentiviral library that encodes genome-wide PaCIITA), and expressed it in P. al ec to and S. lilium kidney cell lines. pools of sgRNAs were infected with VSV-H18. sgRNA sequences from PaCIITA overexpression specifically rendered both of these cell lines sus- surviving cells were sequenced. Data analysis was performed by model- ceptible to H18-pseudotyped VLPs, VSV-H18 and VSV-H17 (Fig. 3f–h, based analysis of genome-wide CRISPR–Cas9 knockout (MAGeCK) to identify enriched sgRNAs, and genes were rank-ordered by robust rank Extended Data Fig. 7d). Although we could not stain for surface MHC-II aggregation (RRA) scores. d, Schematic of candidate factors obtained in these bat cell lines, our results suggest that upregulation of bat MHC-II from comparative transcriptome analysis and CRISPR screening related by PaCIITA facilitates bat IAV infection mediated by haemagglutinin. to MHC-II. The identified factors (HLA-DRA, RFXANK, RFXAP, RFX5, Given the broad range of HLA-DR homologues that support the entry of CIITA and HLA-DRB1) on the MHC-II gene locus are highlighted in blue. bat IAV and the high sequence conservation between different MHC-II complexes (Extended Data Figs. 8, 9), we also tested other human surface consistent with the sequence similarity between H17 and H18 (Extended MHC-II complexes for their ability to mediate H18-dependent entry, Data Fig. 4). Of note, HLA-DRA knockout prevented infection with and found that HLA-DQ and HLA-DP also confer susceptibility to bat authentic H18N11 virus (Fig. 2e). By contrast, control infections with IAV (Fig. 3i). Thus, we conclude that surface MHC-II complexes from a H1N1-pseudotyped VLPs (Fig. 2b), conventional IAV or VSV (Fig. 2c, broad range of species mediate the entry of bat IAV. Extended Data Fig. 3a) were unaffected by knockout of HLA-DRA. The We tested whether mice are also susceptible to authentic H18N11 re-introduction of HLA-DRA in the knockout cells restored HLA-DR infection. First, we confirmed that ectopic expression of mouse MHC-II surface expression (Extended Data Fig. 3c), and made these cells suscep- confers susceptibility to H18-mediated entry in vitro (Extended Data tible to H18-pseudotyped VLPs (Fig. 2f) and VSV-H18 (Extended Data Fig. 10a, b). Next, we infected wild-type C57BL/6 mice (B6) with H18N11 Fig. 3d, e). HLA-DRA knockout in Calu-3 cells also conferred resistance virus (1 × 105 focus-forming units (ffu) in 40 μl) via the intranasal route. to H18-pseudotyped VLPs, as compared to controls (Extended Data Notably, we detected robust viral replication confined to the upper airways Fig. 3f, g). Moreover, ectopic expression of the HLA-DR complex in the (Fig. 4a). Immunohistochemistry and in situ hybridization of consecutive three non-susceptible cell lines used for transcriptomics rendered these sections revealed that viral matrix protein (Fig. 4b, panels 1 and 2) and viral cells susceptible to H18 (Extended Data Fig. 5a–d). We conclude that the H18 RNA (Fig. 4b, panels 3 and 4) were detectable only in the immediate HLA-DR complex is required for the entry of bat IAV. vicinity of the olfactory epithelium. We observed strong antigen accu- Next, we tested whether pre-incubation of cells with an antibody mulation at the apical ciliary border of the olfactory epithelium (Fig. 4b, against MHC-II could inhibit the entry of bat IAV. The anti-MHC-II Extended Data Fig. 10c), as compared to pan-cellular antigen accumula- antibody inhibited H18-mediated—but not H1-mediated—entry in a tion in the respiratory epithelium of mice infected with a conventional IAV dose-dependent manner (Fig. 2g), whereas the control antibody had no (strain A/Udorn (H3N2)) that is known to replicate efficiently in the upper

110 | NATURE | VOL 567 | 7 MARCH 2019 Letter RESEARCH

a Control KO no. 1 KO no. 2 b c VSV-H18 H7N7 VSV abH1N1 H18 100 H1N1 H18 100 100 HLA-DRA 100 80 +HLA-DRB1

Control 80 CIITA 60 HLA-DRA 10 60 HLA-DRB1 10 40 Control 40 20 KO no. 1 20 Relative cell count (%)

0 Entry-positive cells (%) Relative cell count (%)

1 Entry-positive cells (%) 2 3 4 5 0 10 10 10 10 Control KO KO 0 1 KO no. 2 2 3 4 5 A APC (HLA-DR) no. 1 no. 2 0 10 10 10 10 ol APC (HLA-DR) CIIT d e f 100 Contr HLA-DRA HLA-DRA H1N1 H17 c HLA-DRB1 + HLA-DRB1 100 DAPI NP H1N1 H18 100

Control 10 DAPI NP 10

KO no. 1 10 Entry-positive cells (%) 1 DAPI NP

Entry-positive cells (%) 1 Control Control KO no.KO 1 no. 2 KO no.KO 1 no. 2 Control KO KO KO no. 2 Entry-positive cells (%) no. 1 no. 2 LV control LV HLA-DRA 1 Control HLA-DR SLA-DR B-L E. fuscus DR M. lucifugus DR P. alecto DR g NS h d 1,000 NS H1N1 LV control LV HLA-DR * H18 No Ab No Ab ** 100 DAPI Control HLA-DR SLA-DR B-L E. fuscus DR M. lucifugus DR P. alecto DR mCherry H1N1 H18 10 LV HLA-DR LV HLA-DR e f 100 H17 100 + Anti-His + Anti-MHC-II Luciferase activity (%) (relative to untreated) 1 5 5 5 5 –1 20 20 20 20 μg ml : HEK293T mCherry + H18 1.25 1.25 1.25 1.25 10 0.3125 0.3125 0.3125 0.3125 10 Anti-MHC-II Anti-His i j ** Entry-positive cells (%) 125 H1 H18 *** MDCKII no. 2 LV control Entry-positive cells (%) 6,000 * MDCKII no. 2 LV HLA- 1 1 ol ol TA rol 100 5,000 DR–HA tag B-L DR DR WT WT 4,000 Contr HLA-DRSLA-DR PaCII PaCIITA 75 alecto DR LV contr LV LV cont LV 3,000 E. fuscus P. M. lucifugus SlK PaK 50 2,000 H1N1 H18 1,000 ghSlK PaK i 100 H17 100 25 0

Polykaryon formation (%) 0 –1,000 WT (relative to LV HLA-DR, no Ab) I –2,000 Count of detected complexes No Ab No Ab –3,000 l Anti-His 10 10 k Anti-MHC-I Moc Mock cont ro LV control LV HLA-DR VSV-H18 VSV-H18 LV

Anti-HA tag Anti-MHC-I GFP-positive cells (%) A Entry-positive cells (%) 1 1 l l A l Fig. 2 | HLA-DR mediates the entry of bat IAV into U-87MG cells. ro Pa CIIT PaCIITA Cont = LV HLA-DR HLA-DQ HLA-DP a, HLA-DR surface staining of indicated cells from a representative of n 3 LV contLVro LV contLVro PaCIIT independent experiments. APC, = allophycocyanin. b, Indicated cell lines SlK PaK were infected with H18- or H1N1-pseudotyped β-lactamase–M1 fusion- protein (BlaM1) VLPs, and entry-positive cells were quantified. KO no. 1, Fig. 3 | Homologues of HLA-DR from different species confer KO no. 2, HLA-DRA knockout clones of U-87MG cells. c, Cells from a were susceptibility to bat IAV. a, HEK293T cells were transfected as indicated infected with VSV-H18, H7N7 and VSV GFP-reporter viruses. Representative and HLA-DR surface expression was analysed at 48 h after transfection. Data = images of n = 3 independent experiments are shown. Scale bar, 100 μm. d, As are representative of n 2 independent experiments. b, HEK293T cells, in b, but for H1N1- or H17-pseudotyped BlaM1 VLPs. e, Indicated cells were transfected as in a for 48 h, were infected with BlaM1 VLPs pseudotyped infected with IAV H18N11 and stained for viral nucleoprotein (NP) in green. with H1N1 or H18. Relative numbers of entry-positive cells are shown. α Representative images of n = 3 independent experiments are shown. Scale c, HEK293T cells were co-transfected with plasmids encoding the - and β bar, 100 μm. f, Indicated cells were transduced with control lentivirus (LV) -chains of HLA-DR homologues from different species: S. scrofa (SLA-DR), or lentivirus encoding HLA-DRA, and then infected with H18-pseudotyped G. gallus (B-L), E. fuscus (E. fuscus DR), M. lucifugus (M. lucifugus DR) and BlaM1VLPs. b, d, f, Data are means ± s.d. from n = 3 independent P. al e c to (P. al e c to DR). At 48 h after transfection, cells were infected and experiments. Values below background are displayed on the x axis. g, U-87MG analysed as in b. d, HEK293T cells, transfected as in c, were infected with cells were incubated with indicated antibodies before infection with H18- or VSV-H18 (multiplicity of infection (MOI) of 10). Fluorescence microscopy μ H1N1-pseudotyped BlaM1VLPs. Mean values ± s.d. from n = 5 independent images were taken at 72 h p.i. Scale bar, 100 m. e, HEK293T cells were experiments are shown. Statistical significance was determined by Mann– transfected as indicated and infected with BlaM1 VLPs pseudotyped with H17. Relative numbers of entry-positive cells are shown. f, CIITA from Whitney test (one-tailed). *P = 0.0159, **P = 0.004. h, i, HEK293T cells transfected with mCherry and H1 or H18 were co-cultured with MDCKII P. al e c to (PaCIITA) was stably expressed in kidney cells of S. lilium (SlK) and clone no. 2 control or MDCKII clone no. 2 expressing HLA-DR, with or P. al e c to (PaK) before infection and analysis as in b. WT, wild type. g, Cells without the indicated antibodies (Ab). pH-induced polykaryon formation was described in f were infected with VSV-H18 at an MOI of 10. Fluorescence μ examined. Representative images are shown in h, means ± s.d. of normalized microscopy images were taken at 24 h p.i. Scale bar, 100 m. h, Cells polykaryon counts from n = 3 independent experiments in i. Scale bar, described in f were infected with VSV-H17 expressing GFP at an MOI of 1 25 μm. j, Indicated cells were mock-infected or infected with VSV-H18 and for 72 h. Relative numbers of GFP-positive cells are shown. i, HEK293T cells, processed for emulsion coupling. Normalized values are graphed as box plots; transfected as indicated for 48 h before infection and analysis as in b. b, c, ± = boxes represent interquartile range (IQR; first to third quartiles); whiskers e, f, h, i, Data are mean s.d. from n 3 independent experiments. Values are 1.5× IQR; horizontal mid-line, median; dot, mean; crosses denote below background levels of 1 are displayed on the x axes. d, g, Representative = extreme values. Anti-haemagglutinin (HA) tag samples (n = 8) are normally images of n 3 independent experiments are shown. distributed (one-sample Kolmogorov–Smirnov test, P = 0.02, 0.001, 0.001 < = and 0.001 in order of box plots), and anti-MHC-I samples (n 6) belong airways (Extended Data Fig. 10c, d). MHC-II-deficient mice were resist- to same distribution with zero mean (two-sample Kolmogorov–Smirnov test, P = 0.99, t-test against zero mean P = 0.64). Data are derived from at least two ant to H18N11, as no viral replication was detected in the upper or lower independent experiments; statistical significance is indicated; two-tailed t-test, airways at four days post infection (p.i.) (Fig. 4c). By contrast, infections heteroscedastic. ***P = 0.0006, **P = 0.01, *P = 0.05. of wild-type and MHC-II-deficient mice with the classical IAV strain PR8

7 MARCH 2019 | VOL 567 | NATURE | 111 RESEARCH Letter ab 8 Online content B6 1 7 Any methods, additional references, Nature Research reporting summaries, source 6 data, statements of data availability and associated accession codes are available at 5 fu per ml)) https://doi.org/10.1038/s41586-019-0955-3. (f

10 4 2

(log 3 Received: 8 February 2018; Accepted: 22 January 2019; 2 Tit re Published online 20 February 2019. 1

Lung Brain Upper Trachea airway 1. Tong, S. et al. New World bats harbor diverse infuenza A viruses. PLoS Pathog. c 3 8 B6 9, e1003657 (2013). 7 MHC-II de cient 2. Tong, S. et al. A distinct lineage of infuenza A virus from bats. Proc. Natl Acad. 6 Sci. USA 109, 4269–4274 (2012). 5 3. Zhu, X. et al. Hemagglutinin homologue from H17N10 bat infuenza virus fu per ml)) 4 (f

10 4 exhibits divergent receptor-binding and pH-dependent fusion activities. Proc. Natl Acad. Sci. USA 110, 1458–1463 (2013).

(log 3 2 4. Sun, X. et al. Bat-derived infuenza hemagglutinin H17 does not bind canonical Tit re 1 avian or human receptors and most likely uses a unique entry mechanism. Cell Rep. 3, 769–778 (2013). Lung Upper Lung Upper 5. Weis, W. et al. Structure of the infuenza virus haemagglutinin complexed with airway airway B6 its receptor, sialic acid. Nature 333, 426–431 (1988). d MHC-II de cient e 6. Gamblin, S. J. et al. The structure and receptor binding properties of the 1918 8 8 B6 infuenza hemagglutinin. Science 303, 1838–1842 (2004). 7 7 –/– pIV K14 CIITA Tg 7. Stevens, J. et al. Glycan microarray analysis of the hemagglutinins from modern 6 6 5 and pandemic infuenza viruses reveals diferent receptor specifcities. J. Mol.

fu per ml)) 5 (f

(pfu per ml)) 4 Biol. 355, 1143–1155 (2006). 10

10 4 3 8. Edinger, T. O., Pohl, M. O., Yángüez, E. & Stertz, S. Cathepsin W is required for

(log 3 (log 2 escape of infuenza A virus from late endosomes. MBio 6, e00297-15 (2015). 2 Tit re

Tit re 1 9. Wu, Y., Wu, Y., Tefsen, B., Shi, Y. & Gao, G. F. Bat-derived infuenza-like viruses 0 1 H17N10 and H18N11. Trends Microbiol. 22, 183–191 (2014).

Lung Lung 10. Pugach, P. et al. HIV-1 clones resistant to a small molecule CCR5 inhibitor use Upper rachea Upper Upper Upper airway T airway Trachea airway airway the inhibitor-bound form of CCR5 for entry. Virology 361, 212–228 (2007). 11. Tscherne, D. M., Manicassamy, B. & García-Sastre, A. An enzymatic virus-like Fig. 4 | MHC-II knockout mice are resistant to bat IAV infection. −1 particle assay for sensitive detection of virus entry. J. Virol. Methods 163, a, Viral organ titres (ffu ml ) of B6 mice (n = 5), intranasally infected 336–343 (2010). with H18N11 (1 × 105 ffu in 40 μl), determined by immunofluorescence 12. Schröder, B. The multifaceted roles of the invariant chain CD74—more than on day 4 p.i. Dashed line indicates detection limit. b, Detection of viral just a chaperone. Biochim. Biophys. Acta 1863, 1269–1281 (2016). antigen and RNA in paraffin-embedded tissue of B6 mice infected with 13. Stockinger, B. et al. A role of Ia-associated invariant chains in antigen processing = and presentation. Cell 56, 683–689 (1989). H18N11 (n 4) for 4 days as described in a. Intense matrix protein 14. Aydillo, T. et al. Specifc mutations in the PB2 protein of infuenza A virus immunoreactivity was found in an oligofocal and laminar pattern in the compensate for the lack of efcient interferon antagonism of the NS1 protein of apical ciliary border of the olfactory epithelium (1, arrows). The inset (2) bat infuenza A-like viruses. J. Virol. 92, e02021-17 (2018). is a higher magnification of the bottom right corner of panel 1. Panel 3 15. Irla, M. et al. Autoantigen-specifc interactions with CD4+ thymocytes control shows intense RNA specific to H18 haemagglutinin in an oligofocal and mature medullary thymic epithelial cell cellularity. Immunity 29, 451–463 pancellular pattern (arrows) within the olfactory epithelium. The inset (4) (2008). is a higher magnification of the bottom right corner of panel 3. Scale bars, 50 μm (main panels), 20 μm (insets). c, B6 (n = 10) and MHC-II-deficient Acknowledgements We thank C. Kastenholz, K. Flämig, J. Brandel, S. Schuparis, S. Sander and G. Czerwinski for assistance; A. Dudek, P. Staeheli, mice (n = 10) were infected with H18N11, as described in a. Viral titres −1 D. Schnepf and A. Trkola for discussions; and R. Zengerle, D. Szabó, P. Koltay, (ffu ml ) were determined by immunofluorescence at day 4 p.i. d, Viral A. Karsai and S. Zimmermann for their support. This work was supported by −1 organ titres (plaque-forming units (pfu) ml ) of B6 mice (n = 5) and grants from the Swiss National Science Foundation to S.S. (310030E-164065) MHC-II-deficient mice (n = 5), intranasally infected with mouse-adapted and B.G.H. (31003A_182464), a grant from the German Research Foundation IAV strain A/PR8 (H1N1) (1 × 103 pfu in 40 μl) were determined on day 4 to M.S. (SCHW 632/17-1) and M.B. (BE 5187/4-1) and the Excellence Initiative p.i. by plaque assay. Dashed line indicates the detection limit. e, B6 (n = 5) of the German Research Foundation (GSC-4, Spemann Graduate School) −/− to T.T. This work was also partly supported by CRIP (Center for Research on and pIV K14 CIITA Tg mice (n = 5) were infected with H18N11 as Influenza Pathogenesis) and NIAID funded Center of Excellence in Influenza −1 described in a. Viral titres (ffu ml ) in the upper airways were determined Research and Surveillance (CEIRS) to A.G.-S. (HHSN272201400008C). M.W.C. by immunofluorescence on day 4 p.i. Dashed line indicates the detection and C.B. were supported by NIAID grant U19AI135972. limit. a, c–e, Data are mean ± s.d. Reviewer information Nature thanks Michael Farzan, David Steinhauer and the (H1N1) showed no differences in viral titres, which confirms the specific other anonymous reviewer(s) for their contribution to the peer review of this work. restriction of bat IAV in MHC-II-deficient mice (Fig. 4d). To further characterize bat IAV tropism in mice, we infected mice in Author contributions K.C. and W.R. contributed equally to this work. U.K., T.T., which the IFNγ-inducible CIITA promoter pIV was knocked out, but K.C., W.R., S.C.G., M.O.P., D.E., C.J., D.H., S.R., J.S., R.U., J.W., M.G.B.H., M.W.C., transgenic CIITA expression under the control of the keratin-14 (K14) pro- A. Hunziker, E.Y. and A. Halenius performed experiments; U.K., T.T., K.C., M.S. −/− and S.S. designed experiments; T.A., F.K., J.O., A.O. and G.Z. contributed reagents moter was restored in the thymus (hereafter, pIV K14 CIITA Tg mice). and mice; M.M., C.B., B.G.H., A.G.-S., M.B., M.S. and S.S. supervised work and These animals specifically lack MHC-II expression induced by interfer- acquired funding; U.K., M.S. and S.S. wrote the manuscript. γ on- on cells that are not derived from bone marrow (such as endothelia Competing interests C.J. is the patentee of the patent covering the technology and epithelia, and with the exception of cortical thymic epithelial cells), of emulsion coupling (WO2016083793) and a shareholder in Actome GmbH but exhibit MHC-II expression on professional antigen-presenting cells15. (Germany), which holds the patent. We found that viral replication was undetectable in these mice (Fig. 4e), Additional information which—together with the histology—strongly suggests that the nasal epi- Extended data is available for this paper at https://doi.org/10.1038/s41586- thelium is the only site of productive bat IAV replication in mice. 019-0955-3. Our data uncover MHC-II as an entry mediator for bat IAVs, which Supplementary information is available for this paper at https://doi.org/ 10.1038/s41586-019-0955-3. identifies a critical molecular determinant of tropism for these newly Reprints and permissions information is available at http://www.nature.com/ identified viruses. Similar to conventional IAVs, bat IAVs are promis- reprints. cuous in that they use a widely expressed and highly conserved entry Correspondence and requests for materials should be addressed to M.S. or S.S. Publisher’s note: Springer Nature remains neutral with regard to jurisdictional factor found in many vertebrates—including humans, and livestock claims in published maps and institutional affiliations. routinely contacted by humans. Zoonotic potential for these bat viruses cannot, therefore, be excluded. © The Author(s), under exclusive licence to Springer Nature Limited 2019

112 | NATURE | VOL 567 | 7 MARCH 2019 Letter RESEARCH

Methods and cloned into pLVX-IRES-puro (Clontech no. 632183) using XhoI and BamHI No statistical methods were used to predetermine sample size. restriction sites. HLA-DRB1 cDNA was synthesized (IDT) and cloned into pLV- Cell lines. Human embryonic kidney cells 293T, human lung adenocarcinoma cell EF1a-IRES-Hygro using BamHI and EcoRI restriction sites. Lentiviruses encod- lines A549 and Calu-3 were purchased from the American Type Culture Collection ing either pLVX-HLA-DRA-IRES-puro or pLV-EF1a-HLA-DRB-IRES-Hygro, (ATCC). Immortalized bat kidney cell lines from P. alecto and S. lilium have pre- and empty vector control lentiviruses, were generated as previously described19. viously been described14,16. MDCKII clone no. 1 was obtained from G. Herrler Cells were simultaneously transduced with lentiviruses encoding pLVX-HLA- (University of Veterinary Medicine Hannover), and has previously been described17. DRA-IRES-puro and pLV-EF1a-HLA-DRB-IRES-Hygro, or empty vector control MDCKII clone no. 2 cells were obtained from ECACC. MDCKII clone no. 1 and lentiviruses. Successfully transduced A549 cells were selected with 1 μg/ml puro MDCKII clone no. 2 differ in the number of passages. Human glioblastoma cell and 300 μg/ml hygro, U-118MG cells with 0.5 μg/ml puro and 50 μg/ml hygro, lines U-87MG and U-118MG were originally from ATCC, provided by D. Holm- and MDCKII clone no. 2 cells with 2.5 μg/ml puro and 300 μg/ml hygro. Cell lines von Laer (University of Innsbruck) and have also previously been described17. stably expressing HLA-DR and control cell lines were designated LV HLA-DR and U-118MG cells are listed as one of the commonly misidentified cell lines, as LV control, respectively. some samples have been found to be contaminated with U-138MG cells (another To generate MDCKII clone no. 2 LV HLA-DR–HA tag, cDNA encoding hae- human glioblastoma cell line). As we use U-118MG as a control non-susceptible magglutinin-tagged HLA-DRB1 was synthesized (IDT) and cloned into pLV-EF1a- cell line in the transcriptome analysis, this contamination (if it existed) would not HLA-DRB-IRES-Hygro using BamHI and EcoRI restriction sites. Lentiviruses affect our results. A549, Calu-3 and MDCK cell lines were verified to be of human encoding HLA-DRB1–HA tag were generated as previously described before19, or canine origin, respectively, by sequencing the COX1 transcript. No additional cell and were used to transduce MDCKII clone no. 2 cells, simultaneously with lenti- line authentication was performed. Cell lines used in the Stertz and the Schwemmle viruses encoding pLVX-HLA-DRA-IRES-puro. Cells were selected with 2.5 μg/ml laboratory are routinely tested for mycoplasma contamination by sending repre- puro and 300 μg/ml hygro and tested for susceptibility by infection with VSV-18 sentative samples for mycoplasma testing at Eurofins. None of the cell lines used (Extended Data Fig. 6d) ever tested positive for mycoplasma. All cell lines—except S. lilium kidney cells— Plasmids. To clone PaCIITA, kidney cells of P. al ec to were treated were maintained in Dulbecco’s modified Eagle medium (DMEM) supplemented with 1,000 U IFN-α B/D (a gift of P. Staeheli) for 12 h. Total RNA was with fetal calf serum (FCS 10% v/v, Thermo Fisher Scientific) and penicillin–strep- extracted and reverse-transcribed by RevertAid First Strand cDNA syn- tomycin (100 U/ml, Thermo Fisher Scientific). S. lilium kidney cells were cultured thesis kit (Thermo Fisher Scientific) using a PaCIITA-specific primer in DMEM nutrient mixture F-12 (DMEM/F-12, Thermo Fisher Scientific), supple- (5′-CTGTTCCAGCTTGAGAAT-3′). PCR was performed with forward primer mented with FCS (15% v/v) and penicillin–streptomycin (100 U/ml). 5′-CGTGAATTCGCCGCCACCATGAACCACTTCCAGACC-3′ and reverse To generate the U-87MG–Cas9 cell line, lentiviruses expressing Cas9 (Addgene primer 5′-GCATCTAGATCATCTCAGGCTGATCC-3′ and the PCR product was plasmid 5296218, a kind gift of F. Zhang) were produced in HEK293T cells as pre- cloned into pLVX-puromycin vector (Clontech) using EcoRI and XbaI restriction viously described19. U-87MG cells were transduced with lentiviruses expressing sites. Cas9, and two days after transduction were selected with blasticidin (1 μg/ml) for For generation of the pET15-expression plasmid encoding His–NP seven days. Single cells were seeded into 96-well plates and 4 clones were selected of IAV H17N10, the corresponding NP open reading frame was amplified that express high levels of Cas9, as detected by western blot analysis using a mon- from pCAGGS-NP (H17N10)24 using the primer 5′-GATCCATATGA oclonal anti-Flag M2 antibody (Sigma-Aldrich). Cas9 activity was determined by CAACTCAAGGCCTTAAAC-3′ and 5′-GATCGGATCCTCAAATGTCGAACT transduction of U-87MG–Cas9 cell clones with a mixture of lentiviruses encoding CGTCTG-3′, and subsequently cloned into NdeI and BamHI of pET15 (Novagen). pLV–eGFP (Addgene plasmid 3608320, kindly provided by P. Tsoulfas) and one out Transcriptome analysis. Comparative transcriptome analyses were performed of three sgRNAs targeting eGFP (Addgene plasmids 80034, 80035 and 8003621, for the following pairs of H18N11-susceptible, versus non-susceptible, cell lines: kindly provided by J. Doench and D. Root) at an MOI of 1. Two days p.i., cells MDCKII clone no. 1 versus MDCKII clone no. 2, Calu-3 versus A549, and were selected with 1 μg/ml puromycin (puro) for ten days and the frequency of U-87MG versus U-118MG. For each cell line, total RNA was extracted using the eGFP-expressing cells was measured by flow cytometry. The cell clone with the RNeasy Mini Kit (Qiagen), according to manufacturer’s instructions. RNA samples highest reduction in GFP signal intensity, designated U-87MG–Cas9, was selected were processed as quadruplicates. for CRISPR screening. The quantity and quality of the isolated RNA was determined with a Qubit To generate HLA-DRA knockout cell lines, HLA-DRA knockout clones were (1.0) Fluorometer (Thermo Fisher Scientific) and a Tapestation 4200 (Agilent). generated by the means of CRISPR–Cas9-mediated genome editing using ribo- The TruSeq Stranded HT mRNA Sample Prep Kit (Illumina) was used in sub- nucleoproteins (RNP) consisting of Alt-R SpCas9 nuclease in complex with Alt-R sequent steps. In brief, total RNA samples (1 μg) were ribosome-depleted and CRISPR–Cas9 CRISPR RNA (crRNA) (HLA-DR_crRNA_1: (AltR1) rArG rCrUrG reverse-transcribed into double-stranded cDNA with actinomycin added during rUrGrC rUrGrA rUrGrA rGrCrG rCrUrC rGrUrU rUrUrA rGrArG rCrUrA rUrGr first-strand synthesis. The cDNA samples were fragmented, end-repaired and pol- C rU (AltR2) and HLA-DR_crRNA_2: (AltR1) rUrG rArUrG rArArA rArArU yadenylated before ligation to TruSeq adapters. The adapters contained the index rCrCrU rArGrC rArCrA rGrUrU rUrUrA rGrArG rCrUrA rUrGrC rU (AltR2)) for dual multiplexing. Fragments containing TruSeq adapters on both ends were (Integrated DNA technology) and trans-activating crRNA (Integrated DNA tech- selectively enriched with PCR. The quality and quantity of the enriched libraries nology no. 1075927). In brief, pre-assembled RNP complexes were delivered into were validated using Qubit (1.0) Fluorometer and the Tapestation 4200. The aver- U-87MG or Calu-3 cells by reverse transfection, using RNAiMax (Thermo Fisher age fragment size of the library was determined to be approximately 360 bp. The Scientific). At 48 h post transfection, cells were subjected to T7 endonuclease I libraries were normalized to 10 nM in Tris-Cl 10 mM, pH 8.5 with 0.1% Tween (New England Biolabs) assay to estimate the frequency of genome-editing events. 20. For cluster generation, the HiSeq 4000 SR Cluster Kit (Illumina) was used with Screening of individual clones, generated by limiting dilution, was performed 8 pM of pooled normalized libraries on the cBOT V2. Sequencing was performed by flow-assisted cell sorting (FACS) analysis of surface expression of HLA-DR, on the Illumina HiSeq with single-end 125-bp approach using the HiSeq 3000/4000 followed by genotyping22. The following PCR primers were used to amplify SBS Kit (Illumina). the region of interest: HLA-DRA_Fw: 5′-ACCCTTTGCAAGAACCCTTCC-3′, Reads were quality-checked with FastQC. Reads at least 20 bases long, and with HLA-DRA_Rv: 5′-ACCGACAGGATTTACACTCC-3′. an overall average phred quality score greater than 10, were aligned to the reference To generate HLA-DRA reconstituted U-87MG cell lines, a chemically synthe- genome and transcriptome (FASTA and GTF files, respectively, downloaded from sized HLA-DRA gBlock (IDT) was amplified by PCR with specific primers (for- Ensembl, genome build GRCh38) using STAR v,2.5.125 with default settings for ward: 5′-CGTGGATCCGCCGCCAACATGGCCATAAGTGGAGTCC-3′ and single-end reads. Distribution of the reads across genomic isoform expression reverse: 5′-GCAGAATTCTCAGCTCAGG-3′). The PCR product was cloned into was quantified using the R package GenomicRanges from Bioconductor v.3.026. pLV-EF1a-IRES-Hygro (Addgene plasmid 8513423, kindly donated by T. Meyer) Differentially expressed genes were identified using the R package edgeR from using BamHI and EcoRI restriction sites. Lentiviruses encoding HLA-DRA or Bioconductor v.3.027. To identify significantly upregulated genes overlapping in pLV-EF1a-IRES-Hygro vector were produced in HEK293T cells as previously all three H18-susceptible cell lines, transcripts were filtered as follows: P ≤ 0.01, 19 described . U-87MG HLA-DRA knockout cells (KO no. 1 and KO no. 2) were log2(ratio) ≥ 1. Additionally, significantly upregulated genes were filtered accord- transduced twice, with lentivirus either expressing HLA-DRA or empty vector. ing to their Gene Ontology annotations: GO: 0016020 (membrane), GO: 0016021 The second transduction was performed 48 h after the first transduction; then, (integral component of membrane) and GO: 0005886 (plasma membrane). cells were selected with 50 μg/ml hygromycin (hygro). RNA interference mini-screen on U-87MG cells. For gene silencing, four different To generate bat cell lines that stably express PaCIITA, kidney cells of P. al ec to siRNAs (Qiagen) per factor were used (Supplementary Table 3). As a control, a or S. lilium were transduced with lentiviruses expressing either PaCIITA or empty non-targeting siRNA (siSCR) was transfected in parallel. Cells were transfected in vector, and were selected with 2 μg/ml puro two days after infection. suspension with 30 nM siRNA diluted in Opti-MEM (Thermo Fisher Scientific) To generate of A549, U-118MG and MDCKII clone no. 2 cell lines that stably using RNAiMax (Thermo Fisher Scientific), according to manufacturer’s instruc- express HLA-DR, HLA-DRA cDNA was synthesized (Thermo Fisher Scientific) tions, and seeded onto poly-l-lysine-coated (Sigma-Aldrich) 96-well plates. Cells RESEARCH Letter were infected with luciferase-encoding VLPs pseudotyped with H1N1, H18 or and infection’. Luciferase signals from antibody-treated samples were normalized empty vector control, as described below. Fold differences in luciferase signals from to control samples without antibody. H1N1- or H18-infected samples to empty-vector-control infected samples were HLA-DR and DLA surface staining. Cells were detached with 0.25% trypsin- calculated and normalized to siSCR-transfected samples. To account for poten- EDTA (Thermo Fisher Scientific) and re-suspended in DMEM supplemented with tial cytotoxicity, RNA interference effects on H18-VLP entry were normalized to 10% FCS and penicillin–streptomycin to inactivate trypsin. Cells were washed H1N1-VLP entry. three times with FACS buffer (PBS supplemented with 1 mM EDTA (Thermo CRISPR screening. Lentiviruses encoding the human CRISPR-knockout pooled Fisher Scientific) and 2% bovine serum albumin (BSA, Sigma-Aldrich) by cen- sgRNA library Brunello (Addgene plasmid 7317921, a kind gift from D. Root and trifugation at 1,100 r.p.m. for 3 min at 4 °C, and resuspension of cell pellets. To J. Doench) were produced in HEK293T cells as previously described19,21. The stain for surface HLA-DR or DLA, cells were incubated for 1 h at 4 °C with APC library contains 76,441 sgRNAs targeting 19,114 human genes and 1,000 control anti-human HLA-DR antibody (BioLegend, catalogue no. 307610) or APC anti- non-targeting sgRNAs. Eighty million U-87MG–Cas9 cells were transduced with canine MHC-II (Thermo Fisher Scientific, catalogue no. 17-5909-41), respectively. the lentiviral sgRNA library at an MOI of 0.3. Two days after transduction, puromy- To exclude dead cells, the LIVE/DEAD Fixable Near-IR Dead Cell Stain Kit cin was added and transduced cells were selected for 10 days to achieve an average (Thermo Fisher Scientific) was used. Cells were washed three times with FACS 1,000-fold coverage of the library. The cells were infected with VSV-H18 at an MOI buffer as described above, and analysed on a FACSVerse System (BD). Typically, of 10. In parallel, mock-infected cells were collected. The survivors were expanded, 3–5 × 103 live cells were acquired and APC intensities were examined. reinfected with VSV-H18 under the same conditions for 4 rounds and collected on IAV-based VLP (BlaM1-VLP) production and infection. Expression plas- day 27 p.i. The susceptibility of the survivors to VSV-H18 infection was monitored mid coding for a BlaM1 fusion protein, consisting of the β-lactamase and by flow cytometry measuring the frequency of GFP-reporter virus-infected cells. the matrix protein (M1) from A/bat/Guat/164/09 (H17N10), was cloned Genomic DNA (gDNA) of cell pellets was isolated using Quick-DNA Miniprep using primers 5′-GACAACTAGTATGAGCATCTTAACAGAGG-3′ and Plus Kit (Zymo Research). 5′-GACAGGATCCTCATTTAAACCTCTGCATTTG-3′ and pCAGGS-A/bat/ For Illumina sequencing and screening analysis, PCR was performed on gDNA Guat/164/09-M124 as template. The PCR product was cloned into pCAGGS- to construct Illumina sequencing libraries, each containing 5 μg gDNA following A/WSN/33-BlaM111 to replace M1 of A/WSN/33, using the restriction sites the Broad Institute protocol PCR of sgRNAs for Illumina sequencing. In brief, SpeI and BamHI. To generate H1N1-, H17- or H18-pseudotyped BlaM1 VLPs, gDNA was aliquoted into PCR tubes in a total volume of 50 μl. A PCR mastermix HEK293T cells were co-transfected in 6-well plates with 2 μg of pCAGGS- containing ExTaq DNA polymerase (Clontech), ExTaq buffer, dNTP, P5 stagger A/WSN/33-BlaM1, 0.5 μg of pCAGGS-A/WSN/33-H1 and 0.5 μg pCAGGS- primers (Supplementary Table 4) and water was prepared. Forty microlitres of PCR A/WSN/33-N1, or with 2 μg of pCAGGS-A/bat/Guat/164/09-BlaM1 and 1 μg of mixture and 10 μl of a barcode P7 primer (Supplementary Table 4) were added to pCAGGS-A/bat/Peru/033/10-H18 or pCAGGS-bat/Guat/060/2010-H17. ViaFect each tube containing 50 μl of gDNA. PCR samples were amplified as follows: 95 °C was used as transfection reagent at a DNA-to-ViaFect ratio of 1:3. Medium was for 1 min, followed by 28 cycles of 95 °C for 30 s, 53 °C for 30 s, 72 °C for 30 s with exchanged 6 h after transfection with Opti-MEM containing penicillin–streptomycin. a final 10 min at 72 °C. The PCR products were then purified from 2% agarose VLPs were collected 72 h after transfection, and treated with 5 μg/ml TPCK-trypsin gels using a gel extraction kit (Zymo Research). PCR amplicons were purified via for efficient haemagglutinin cleavage. Trypsin was inactivated with 10 μg/ml double size-selection with Sera-Mag SpeedBeads (GE). Cleanup was verified by trypsin inhibitor from soybean. For infection, cells were washed with PBS and visualizing samples on 2% agarose gels. DNA concentration of each sample was incubated with 200 μl of BlaM1 VLPs in the presence of 0.1 μg/ml DEAE–dextran determined using a Qubit fluorometer with double-strand DNA high-sensitivity hydrochloride and 2% FCS for 4 h at 37 °C. Cells were collected by trypsinization assay reagents. Samples were then sequenced on an Illumina NextSeq 500. The and loaded with the fluorogenic β-lactamase substrate CCF2-AM (Thermo Fisher resulting sequencing reads were trimmed to 20-nt potential sgRNA sequences Scientific). Samples were analysed on a FACSVerse System (BD) and dead cells using cutadapt28. The MAGeCK software suite29 was used to assign these sequences were excluded by a live and dead staining. Entry-positive cells were quantified by to targets and genes. Comparisons between survivor and control samples were gating on events with cleaved CCF2-AM using FlowJo v.10 software. carried out using the robust rank aggregation method of MAGeCK for significance Virus infections. VSV-H18 GFP and VSV-H17 GFP reporter viruses contain- testing. Rankings from the significance test were also used for pathway enrichment ing a polybasic cleavage site in the haemagglutinin were generated as previ- analysis within MAGeCK. ously described17. Viral titres were determined on RIE1495 cells as previously HIV-based VLP production and infection. cDNAs encoding haemagglutinin H18 described17. For infection with VSV-H18 or VSV-H17, U-87MG, HEK293T, or and neuraminidase N11 of strain A/bat/Peru/033/10 or A/bat/Peru/033/10 were kidney cells of P. al ec to and S. lilium were seeded in monolayers and then infected synthesized (Thermo Fisher Scientific) and cloned into pCAGGS using restriction with the virus at indicated MOIs (ffu per cell) and maintained in culture medium as enzymes NheI and EcoRI. Luciferase-encoding, HIV-based VLPs—pseudotyped previously described17. Infections with VSV and the IAV (A/SC35M) (H7N7) GFP with either the IAV A/WSN/33 (H1N1) or A/bat/Peru/033/10 (H18N11) envelope reporter virus were performed as previously described30,31. Fluorescence images glycoproteins, or control particles lacking any surface glycoprotein—were generated. were acquired on a Zeiss Observer.Z1 inverted epifluorescence microscope (Carl HEK293T cells were co-transfected in 6-well plates with 2.8 μg of pNLLuc-AM10 Zeiss) equipped with an AxioCamMR3 camera using a 40× objective. and 0.5 μg of each of the envelope-encoding plasmids (pCAGGS-A/WSN/33-H1, Recombinant bat IAV H18N11 was generated as previously described17 and all pCAGGS-A/WSN/33-N1 or pCAGGS-A/bat/Peru/033/10-H18, pCAGGS-A/bat/ experiments were performed under biosafety level 3 conditions. Viral titres were Peru/033/10-N11) or 1 μg of the empty vector pCAGGS using a DNA to ViaFect determined by immunofluorescence on subconfluent MDCKII clone no. 1 cells (Promega) ratio of 1:3. At 6 h after transfection, medium was exchanged with Opti- using rabbit polyclonal anti-H18. For infection, U-87MG cells were seeded into MEM supplemented with penicillin–streptomycin (100 U/ml). VLPs were collected 48-well plates and kept in the incubator for 12 h. Cells were washed with PBS 72 h after transfection, and treated with 5 μg/ml TPCK-trypsin (Sigma-Aldrich) for containing 0.2% BSA and then inoculated with the indicated MOI in DMEM con- efficient haemagglutinin cleavage. Trypsin was inactivated with 10 μg/ml trypsin taining 0.2% BSA, penicillin (100 U/ml), streptomycin (100 mg/ml) and TPCK– inhibitor from soybean (Sigma-Aldrich) and VLPs were stored at −80 °C. For trypsin (1 μg/ml) for 1 h. Inoculum was removed and cells were washed with PBS infection, cells were seeded onto poly-l-lysine-coated, 96-well plates and washed containing 0.2% BSA and further incubated in MEM supplemented with FCS with phosphate-buffered saline (PBS) before infection with VLPs. After 90 min (15% v/v). For the immunofluorescence-based infection read-out, at 48 h p.i. cells of infection at 37 °C in the presence of 0.04 μg/ml DEAE–dextran hydrochloride were fixed using 4% paraformaldehyde (PFA) in PBS for 15 min, washed three (Sigma-Aldrich), the inoculum was replaced with DMEM supplemented with 10% times with PBS and permeabilized with 0.5% Triton X-100 in PBS for 5 min. Viral fetal calf serum and penicillin–streptomycin. VLP infectivity was measured at 48 proteins were stained with antibodies against bat IAV nucleoprotein for 1 h at room h p.i. by quantifying firefly luciferase activity using the ONE-Glo Luciferase Assay temperature. Then, cells were washed three times with PBS and stained using a System (Promega), according to manufacturer’s protocol. secondary anti-mouse IgG coupled to AlexaFluor 488 (Jackson ImmunoResearch, Antibody blocking assay. U-87MG cells were seeded into poly-l-lysine-coated, 115-546-062) for 1 h. Finally, cells were washed and nuclei were stained for 5 min 96-well plates and were washed once with PBS. Before infection, cells were incu- using 4′,6-diamidino-2-phenylindole (DAPI) at a dilution of 1:10,000 in PBS. bated for 1 h at 4 °C with an anti-HLA-DR, anti-HLA-DP, anti-HLA-DQ (anti- Samples were analysed by fluorescence microscopy. MHC-II, clone Tü39, BioLegend, catalogue no. 361702) or a control anti-6×His-tag For the flow cytometry-based infection read-out, to collect the infected and (Abcam, catalogue no. ab18184) antibody diluted in Opti-MEM at concentra- detached cells, supernatants were collected and centrifuged at 1,100 r.p.m. for tions ranging from 20 μg/ml to 0.3125 μg/ml. Supernatants were then replaced 3 min at 4 °C. Cells were detached with trypsin and re-suspended in DMEM con- with H1N1- or H18-pseudotyped VLPs containing 0.04 μg/ml DEAE–dextran taining FCS (10% v/v) to inactivate trypsin. Cells were washed twice with FACS hydrochloride (Sigma-Aldrich) and the above-mentioned antibodies at the cor- buffer (PBS containing 3% BSA (Sigma-Aldrich) by centrifugation at 1,100 r.p.m. responding concentrations. After 90 min at 37 °C, inoculum was replaced with for 3 min at 4 °C and fixed in 4% PFA in PBS for 15 min and washed with PBS DMEM supplemented with 10% fetal calf serum and penicillin–streptomycin. VLP 3 times. Samples were analysed on a FACS Canto II System (BD). The frequency infectivity was measured at 48 h p.i. as described in ‘HIV-based VLP production of GFP-positive cells was quantified using FlowJo software. Letter RESEARCH

Generation of a rabbit polyclonal anti-H18 serum. Recombinant VSV-H18 Cells were washed 3 times with PBS and permeabilized with 0.5% Triton X-100 in vector was generated and propagated in BHK-G43 helper cells as previously PBS for 5 min. Proximity ligation assay (PLA) was performed according to manu- described17. Two New Zealand rabbits (8 weeks old) were immunized intramus- facturer’s instructions (Sigma Duolink, DUO92001, DUO92005, DUO92007). In cularly with 108 ffu of VSV-H18 in the absence of adjuvant. Four weeks after the brief, cells were blocked using blocking solution for 1 h at room temperature. Cells primary immunization, the immune response of the animals was boosted by were then incubated with a mouse polyclonal H18 antibody and a rabbit monoclo- immunizing the animals a second time with the same vector vaccine. The animals nal HLA-DR antibody (Abcam, catalogue no. ab92511) in antibody diluent for 1 h were euthanized and bled four weeks after the second immunization. The animal at room temperature. Subsequently, samples were incubated with the PLA probes experiments were performed in compliance with the Swiss animal protection law anti-rabbit-IgG PLUS and anti-mouse-IgG MINUS at 37 °C for 1 h. Ligation was and approved by the animal welfare committee of the canton of Bern (authorization performed for 30 min at 37 °C, followed by PLA amplification reaction for 100 min number BE119/13). at 37 °C. Cells were washed and nuclei were stained with DAPI. Images were Generation of a mouse polyclonal anti-H18 serum. Mice were immunized with acquired with a Zeiss Observer Z1 epifluorescence microscope and μManager35. recombinant protein H18 (A/flat-faced bat/Peru/033/2010) expressed as previously All images were recorded with a Plan Apochromat 20× (NA 0.8) objective and an described32. In brief, H18 sequence was cloned into a modified pFastBacDual Orca 12-bit camera (Hamamatsu). To account for fluorescence intensity differences expression vector in-frame with a C-terminal T4 foldon and a hexahistidine tag. between PLA dots within different focal planes, z-stacks of 12–15 images with a Bacmids were generated by transfecting the plasmids into DH10Bac competent step-size of 0.5–0.75 μm were recorded for all samples. The images within a z-stack bacteria, and plasmids were prepared from white colonies. Bacmids were then were then combined to a maximum projection image (P) by transfected into Sf9 cells to generate recombinant baculovirus, recombinant protein expression was verified by western blot against the hexahistidine tag, and sequences ∀=xy,;PIxy,,max( xy()z ) were verified by Sanger sequencing. Baculovirus stocks were grown up in Sf9 cells, and these stocks were then used to infect High Five cells at an MOI of approxi- in which x and y are the row and column pixel positions, respectively, of image I. mately ten to express protein. Three days p.i., High Five culture supernatants were Within the resulting maximum projection images, all PLA dots were in focus. PLA collected and secreted recombinant haemagglutinin was purified using Ni-NTA dot detection was then performed by finding local maxima in conjunction with resin, followed by a buffer exchange step. Recombinant protein was analysed for a flood fill algorithm with a grey level tolerance T. T was set manually for images purity using sodium dodecyl sulfate polyacrylamide gel electrophoresis, and quan- from one experimental series in dependence of the fluorescence background. The tified using the Bradford method. dot count was evaluated per nucleus. To achieve this, the DAPI-stained nuclei were BALB/c female mice were intramuscularly injected with 10 μg of recombinant segmented by applying a median filter and using a previously described iterative protein in Freund’s incomplete adjuvant (Sigma). A subsequent boost was adminis- thresholding algorithm36. A Voronoi partitioning based on the Euclidean distance tered via intraperitoneal route three weeks after the primary injection. Four weeks map of the nuclei was then computed to segment regions. PLA dots were counted after booster immunization, mice were euthanized and blood was drawn. Serum per region. All image processing steps were carried out in ImageJ37. was separated from red blood cells by centrifugation. Animal experiments were Emulsion coupling. MDCKII clone no. 2 LV control or MDCKII clone no. 2 LV performed in accordance with protocols approved by the Icahn School of Medicine HLA-DR–HA-tag cells were seeded in 6-well plates and incubated with VSV-H18 at Mount Sinai Institutional Animal Care and Use Committee. The presence of at an MOI of 100 for 15 min at room temperature with continuous agitation. The specific H18 antibodies was assessed by enzyme-linked immunosorbent assay cells were washed once with PBS containing 3% FCS (v/v), twice with PBS and (ELISA) reactivity against H18 protein. collected by scraping the cells in 100 μl ice-cold PBS containing 20 mM bis(sul- Generation of a mouse monoclonal anti-NP antibody. His-tagged nucleop- fosuccinimidyl) suberate (BS3), a membrane impermeable cross-linker (Thermo rotein (H17N10) was purified from Rosetta DE3 (Novagen) using His-Trap-FF Fisher Scientific) and incubated for 1 h on ice. The cross-linker was removed (GE Healthcare). The His-tag was removed by thrombin digestion as described by washing twice with ice-cold PBS. The counted cells were re-suspended in by the manufacturer (Novagen), and uncleaved nucleoprotein was further cleaned ice-cold PLP buffer (1× Passive Lysis solution (Promega) and 1× Protease up by using HiTrap DEAE FF. Female BALB/c mice were immunized 5 times Inhibitor Cocktail (Roche) in PBS) at a concentration of 106 cells per 100 μl and intraperitoneally with 15 μg of recombinant nucleoprotein mixed with an equal lysed with continuous resuspension on ice for 3 h. The lysates were clarified by amount of GERBU Adjuvant MM (GERBU Biotechnik GmbH) in intervals of centrifugation at 16,100g for 10 min at 4 °C. For antibody labelling, anti-haemag- 4 weeks. Immunized mice were euthanized and the spleen was removed under glutinin tag antibody (Sigma-Aldrich H6908), mouse polyclonal H18 antibody aseptic conditions four days after the final boost. Monoclonal hybridoma cells were and HLA-ABC monoclonal antibody (Thermo Fisher Scientific W6/32) were generated by fusion of the isolated splenocytes with mouse myeloma SP2/0 cells purified and labelled using APEX Biotin-XX Antibody Labelling Kit (Thermo using a standard protocol and were screened by ELISA as previously described33,34. Fisher Scientific) following the manufacturer’s instructions, except that in the bio- Polykaryon formation assay. HEK293T cells were co-transfected in 6-well plates tin conjugation step, DBCO-PEG4-NHS ester (Jena Bioscience) was used and an with 2 μg of either pCAGGS-A/bat/Peru/033/10-H18 (pCAGGS-H18) or A/ additional step was introduced to conjugate the DBCO group with azide-single- WSN/33-H1 (pCAGGS-H1), and 2 μg of pmCherry-C1 (Clontech). For transfec- strand DNA (IDT) amplicons (Supplementary Table 5). The antibodies were eluted tions, ViaFect (Promega) was used at a DNA-to-ViaFect ratio of 1:3. At 24 h after and converted to Fab fragments by Fabricator cleavage (Genovis), according to the transfection, cells were treated with 10 μg/ml TPCK–trypsin (Sigma-Aldrich) for manufacturer’s instructions. 30 min at 37 °C to cleave and activate haemagglutinin. Transfected and trypsin- For emulsion coupling (Extended Data Fig. 6e), the labelled antibodies were treated HEK293T cells were seeded in the presence or absence of 20 μg/ml anti- mixed in PBS, and 2 μl of cross-linked cellular lysate and 2 μl of antibody mixture HLA-DR, anti-HLA-DQ, anti-HLA-DP (anti-MHC-II, clone Tü39, BioLegend, were reacted overnight to achieve equilibrium binding. The antibody mixture was catalogue no. 361702) or anti-6×His-tag (Abcam, catalogue no. ab18184) antibody also incubated with 2 μl 5% skimmed milk in PLP buffer as an antigen-negative, on top of MDCKII clone no. 2 LV HLA-DR or MDCKII clone no. 2 LV control antibody background control (HLA-ABC, detecting background at no interac- cells, which were previously seeded onto poly-l-lysine-coated (Sigma-Aldrich) tion). To ensure the formation of double-labelled complexes, high concentrations glass coverslips. The next day, cells were washed with PBS and exposed to low pH of antibodies were used (up to nM). The dilution was predetermined to have a (50 mM MES, 150 mM NaCl, pH 5.2) for 20 min in the presence or absence of the few thousand digital-droplet PCR (ddPCR) positive droplets for each antibody. above-mentioned antibodies at 20 μg/ml, washed twice with PBS and subsequently The overnight reaction was diluted in PBS ~100,000 times to dilute out any PCR incubated with DMEM supplemented with 10% FCS and penicillin–streptomycin inhibitory substances, to change the buffer to ddPCR master mix and to separate for 2 h at 37 °C, in the presence or absence of the above-mentioned antibodies at 20 the complexes into individual droplets (more than 20,000 droplets per 20 μl). Then, μg/ml. Cells were fixed with 3.7% (v/v) PFA for 10 min and nuclei were stained with 1 μl of diluted sample was used in ddPCR to get minimally overlapping partition- DAPI for 1h at room temperature. Coverslips were mounted onto glass microscopy ing of the labels. ddPCR was performed with QX200 Droplet Digital PCR System slides using ProLong Gold Antifade Mountant (Thermo Fisher Scientific). Images (Bio-Rad) using ddPCR Supermix for Probes (no dUTP) (Bio-Rad). Under these were acquired with a confocal laser scanning microscope (Leica SP5) using a 20× conditions, the standard ddPCR emulsification process separates the individual objective (HCX PL APO lambda blue 20.0 × 0.70 IMM UV). Polykaryons, defined complexes into separate droplets. The amplicon-label specific primers (IDT) and as multi-nucleated cells with more than two nuclei, from H18- and H1-transfected probes (with FAM–MGB and VIC–MGB fluorophores, Thermo Fisher Scientific; samples were counted from 3 images with approximately 3,000 nuclei and 6 images Supplementary Table 5) were used to amplify and detect single-stranded DNA with approximately 6,000 nuclei, respectively. labels, according to the manufacturer’s instructions (QX200 ddPCR Instrument, Proximity ligation assay. MDCKII clone no. 2 LV HLA-DR or MDCKII clone Bio-Rad). no. 2 LV control cells were seeded on collagen (Roche)-coated 384-well high-con- The evaluation of the emulsion coupling reaction is based on the partitioning tent-imaging glass-bottom microplates (Corning) and incubated with VSV-H18 of the labels in the ddPCR reactions (FAM and VIC labels measured in relative at an MOI of 100 for 1 h at 37 °C. Cells were washed once with PBS containing 3% fluorescence units (RFU)). The ordinary ddPCR evaluation determines the abso- FCS (v/v) and twice with PBS, then fixed using 4% (w/v) PFA in PBS for 15 min. lute number of labels (according to Bio-Rad protocol) in each reaction, which—in RESEARCH Letter turn—can be used to calculate the number of double-positive droplets (the overlap targeting base pairs 26 to 1132 of gene bank accession number CY125945.1, Fast Red of labels) based on the statistical probability of co-compartmentalization (Poisson as chromogen, and haematoxylin counterstain. background). The Poisson background was compared to the number of detected Reporting summary. Further information on research design is available in double-positive droplets and a statistical model was developed to calculate the the Nature Research Reporting Summary linked to this paper. number of molecular complexes explaining the number of double-positive droplets Code availability. The program code for emulsion coupling that supports the over the Poisson background (Python code available on request from C.J.). The findings of this study is available from Actome GmbH. Restrictions apply to the result was normalized against the no-interaction control (HLA-ABC) to gain the availability of this code, which was used under license for the current study (and absolute number of complexes with a zero-mean distribution. is therefore not publicly available). The code is available from the authors upon Ectopic expression of HLA-DR, HLA-DQ, HLA-DP and HLA-DR ortho- reasonable request, and with permission of Actome GmbH. logues on HEK293T cells. cDNA sequences of the following genes were synthe- sized (Thermo Fisher Scientific) and cloned into pmCherry-C1 (Clontech) using Data availability restriction sites NheI and EcoRI or NheI and BamHI, replacing the mCherry- The authors declare that the data supporting the findings of this study are availa- encoding sequence: human MHC-II DR-alpha (HLA-DRA, NCBI reference ble within the paper and its Supplementary Information files. The associated raw sequence NM_019111.4) and MHC-II DR-beta (HLA-DRB1, NM_001243965.1), data for Fig. 1a are provided in Supplementary Table 1, and raw data for Fig. 1c human MHC-II DQ-alpha (HLA-DQA1, NM_002122.3) and MHC-II DQ-beta in Supplementary Table 2. Any further relevant data are available from the corre- (HLA-DQB1; NM_002123.4), human MHC-II DP-alpha (HLA-DPA1; sponding authors upon reasonable request. NM_001242524.1) and MHC-II DP-beta (HLA-DPB1, NM_002121.5), swine MHC-II DR-alpha (SLA-DRA, NM_001113706) and MHC-II DR-beta 16. Crameri, G. et al. Establishment, immortalisation and characterisation of (SLA-DRB1, NM_001113695.1), chicken MHC-II alpha (BLA, NM_001245061.1) pteropid bat cell lines. PLoS ONE 4, e8266 (2009). and MHC-II beta (BLB2, NM_001318995.2), mouse MHC-II antigen A-alpha 17. Moreira, E. A. et al. Synthetically derived bat infuenza A-like viruses reveal a cell type- but not species-specifc tropism. Proc. Natl Acad. Sci. USA 113, (H2-Aa; NM_010378.2) and MHC-II antigen A-beta (H2-Ab1, NM_207105.3), 12797–12802 (2016). mouse MHC-II antigen E-alpha (H2-Ea (also known as H2-Ea-ps), NM_010381) 18. Sanjana, N. E., Shalem, O. & Zhang, F. Improved vectors and genome-wide and MHC-II antigen E-beta (H2-Eb1, NM_010382.2), E. fuscus MHC-II DR-alpha libraries for CRISPR screening. Nat. Methods 11, 783–784 (2014). (EfDRA, XP_008155535) and MHC-II DR-beta (EfDRB5, XP_008155622.1), 19. Shalem, O. et al. Genome-scale CRISPR-Cas9 knockout screening in human M. lucifugus MHC-II DR-alpha (MlDRA; XP_006105212) and MHC-II DR-beta cells. Science 343, 84–87 (2014). 38,39 20. Enomoto, M., Bunge, M. B. & Tsoulfas, P. A multifunctional neurotrophin with (MlDRB5; XP_006105637.1), P. al ec to MHC-II DR-alpha (PaDRA) and NTR 38,39 reduced afnity to p75 enhances transplanted Schwann cell survival and MHC-II DR-beta (PaDRB) . axon growth after spinal cord injury. Exp. Neurol. 248, 170–182 (2013). For transfection, HEK293T cells were seeded onto poly-l-lysine-coated plates 21. Doench, J. G. et al. Optimized sgRNA design to maximize activity and minimize and transfected with the above-mentioned constructs using Lipofectamine 2000 of-target efects of CRISPR–Cas9. Nat. Biotechnol. 34, 184–191 (2016). (Thermo Fisher Scientific) at a DNA-to-lipofectamine 2000 ratio of 1:4. At 48 h 22. Dehairs, J., Talebi, A., Cherif, Y. & Swinnen, J. V. CRISP-ID: decoding CRISPR post transfection, cells were either infected with BlaM1 VLPs, VSV-H18 or stained mediated indels by Sanger sequencing. Sci. Rep. 6, 28973 (2016). 23. Hayer, A. et al. Engulfed cadherin fngers are polarized junctional structures for HLA-DR surface expression. between collectively migrating endothelial cells. Nat. Cell Biol. 18, 1311–1323 Infection of mice. All animal experiments were performed in accordance with (2016). the guidelines of German animal protection law, and were approved by the 24. Juozapaitis, M. et al. An infectious bat-derived chimeric infuenza virus state of Baden-Württemberg (Regierungspräsidium Freiburg; reference num- harbouring the entry machinery of an infuenza A virus. Nat. Commun. 5, 4448 ber 35-9185.81/G-17/14). Mouse infection experiments were performed under (2014). 25. Dobin, A. et al. STAR: ultrafast universal RNA-seq aligner. Bioinformatics 29, BSL-3 conditions in accordance with local animal care committees. B6 mice 15–21 (2013). (29 female and 5 male) were obtained from Janvier and B6.Cg‐Thy1a‐H2‐Aatm1Blt 26. Lawrence, M. et al. Software for computing and annotating genomic ranges. mice (hereafter designated as MHC-II-deficient mice; 10 female and 5 male) PLOS Comput. Biol. 9, e1003118 (2013). were obtained from the Swiss Immunological Mouse Repository (SwImMR). 27. Robinson, M. D., McCarthy, D. J. & Smyth, G. K. edgeR: a Bioconductor package pIV−/− K14 CIITA Tg mice15 (in B6 background; 5 female) were kindly provided for diferential expression analysis of digital gene expression data. Bioinformatics 26, 139–140 (2010). by H. Acha-Orbea (University of Lausanne). No sample-size calculations were 28. Martin, M. Cutadapt removes adapter sequences from high-throughput performed. Allocation of mice to groups was random and investigators were not sequencing reads. EMBnet.journal 17, 10–12 (2011). blinded to group allocation. 29. Li, W. et al. MAGeCK enables robust identifcation of essential genes from Virus stocks were diluted in Opti-MEM medium containing 0.3% BSA (v/w). genome-scale CRISPR/Cas9 knockout screens. Genome Biol. 15, 554 (2014). For infection, 8-to-10-week-old B6 mice, MHC-II-deficient mice and pIV−/− K14 30. Reuther, P. et al. Generation of a variety of stable infuenza A reporter viruses by CIITA Tg mice were anaesthetized with a mixture of ketamine (100 mg per g genetic engineering of the NS gene segment. Sci. Rep. 5, 11346 (2015). 31. Zimmer, G., Locher, S., Berger Rentsch, M. & Halbherr, S. J. Pseudotyping of body weight) and xylazine (5 mg per g body weight) administered intraperito- vesicular stomatitis virus with the envelope glycoproteins of highly pathogenic neally, and subsequently inoculated intranasally with 40 μl of the indicated virus avian infuenza viruses. J. Gen. Virol. 95, 1634–1639 (2014). dose. Throughout the course of the experiment mice were monitored daily. At four 32. Krammer, F. et al. A carboxy-terminal trimerization domain stabilizes days p.i., mice were euthanized and indicated organs were collected. Organs were conformational epitopes on the stalk domain of soluble recombinant homogenized in 1 ml PBS by 3 subsequent rounds of mechanical treatment for hemagglutinin substrates. PLoS ONE 7, e43603 (2012). −1 33. Bussmann, B. M., Reiche, S., Jacob, L. H., Braun, J. M. & Jassoy, C. Antigenic and 25 s each at 6.5 ms . Tissue debris was removed by centrifuging homogenates for cellular localisation analysis of the severe acute respiratory syndrome 5 min at 5,000 r.p.m. at 4 °C. coronavirus nucleocapsid protein using monoclonal antibodies. Virus Res. 122, Histology. Tissues were fixed in 4% neutral phosphate buffered formaldehyde, 119–126 (2006). processed, embedded in paraffin wax, sectioned at 2–4 μm and stained with 34. Monaghan, S. J. et al. Expression of immunogenic structural proteins of cyprinid haematoxylin and eosin. Four standardized coronal sections of the nasal cavity herpesvirus 3 in vitro assessed using immunofuorescence. Vet. Res. 47, 8 (2016). 40 35. Edelstein, A., Amodaj, N., Hoover, K., Vale, R. & Stuurman, N. Computer control (according to the RITA recommendations ), as well as representative specimen of microscopes using μManager. Curr. Protoc. Mol. Biol. 92, 14.20.1–14.20.17 from the larynx, trachea, lung, heart, thymus, spleen, liver, pancreas, intestine, (2010). kidney and brain were evaluated for histopathologic lesions using an Axio Imager 36. Ridler, T. W. & Calvard, S. Picture thresholding using an iterative selection M2 microscope (Carl Zeiss Microscopy). method. IEEE Trans. Syst. Man Cybern. 8, 630–632 (1978). Immunohistochemistry was performed to detect IAV matrix protein using the 37. Rueden, C. T. et al. ImageJ2: ImageJ for the next generation of scientifc image avidin–biotin–peroxidase-complex method (Vectastain Elite ABC Kit Standard, data. BMC Bioinformatics 18, 529 (2017). 38. Papenfuss, A. T. et al. The immune gene repertoire of an important viral Vector Laboratories) with citric buffer (10 mM, pH 6) pre-treatment, mouse-on- reservoir, the Australian black fying fox. BMC Genomics 13, 261 (2012). mouse-kit (Vector M.O.M. Immunodetection Kit BASIC, Vector Laboratories), 39. Ng, J. H. J., Tachedjian, M., Wang, L. F. & Baker, M. L. Insights into the ancestral a monoclonal mouse anti-IAV–matrix protein immunoglobulin G1 containing organisation of the mammalian MHC class II region from the genome of the hybridoma supernatant (clone M2-1C6-4R341, American Type Culture Collection), pteropid bat, Pteropus alecto. BMC Genomics 18, 388 (2017). 3-amino-9-ethyl-carbazol as chromogen, and haematoxylin counterstain. 40. Kittel, B. et al. Revised guides for organ sampling and trimming in rats and mice–part 2. A joint publication of the RITA and NACAD groups. Exp. Toxicol. In-situ hybridization was performed to detect IAV (A/flat-faced bat/Peru/033/ Pathol. 55, 413–431 (2004). 2010 (H18N11)) haemagglutinin-specific RNA using an RNAscope 2.5 assay with target 41. Yewdell, J. W., Frank, E. & Gerhard, W. Expression of infuenza A virus internal retrieval and protease pretreatment, RNAscope 2.5 HD Reagent Kit - RED, the HybEZ antigens on the surface of infected P815 cells. J. Immunol. 126, 1814–1819 hybridization system with a 20ZZ probe named ‘V-bat-influenza-haemagglutinin’, (1981). Letter RESEARCH

Extended Data Fig. 1 | Analysis of transcriptional profiles from comparison—each of which consists of n = 4 independent samples—was H18N11-susceptible and non-susceptible cell lines. a, MDCKII clone performed using the gene-wise negative binomial generalized linear no. 1 (here labelled MDCKII #1), MDCKII clone no. 2 (here labelled model with quasi-likelihood tests, implemented in EdgeR. Adjustments MDCKII #2), U-87MG, U-118MG, Calu-3 and A549 cells were infected for multiple comparisons were made by applying the Benjamini– for 90 min at 37 °C with luciferase-encoding VLPs pseudotyped with Hochberg method. Significantly (P ≤ 0.01) upregulated (fold change H18N11 or H1N1, or control VLPs that lacked any surface glycoprotein (expressed as log2(transcript level susceptible cell line/non-susceptible cell (empty vector). At 48 h p.i., luciferase signals were measured and line) ≥ 1) and downregulated (fold change (expressed as log2(transcript normalized to control samples. Fold increase in luciferase signals relative level susceptible cell line/non-susceptible cell line)≤ −1) transcripts to samples infected with empty vector (EV) is plotted. b, MDCKII clone are shown in red and blue, respectively. Transcripts expressed at similar no. 1 (here labelled MDCKII #1), MDCKII clone no. 2 (here labelled levels (−1 ≤ fold change (expressed as log2(transcript level susceptible MDCKII #2), U-87MG, U-118MG and Calu-3 cells were infected with cell line/non-susceptible cell line) ≤ 1) are shown in grey. Black dots BlaM1 VLPs pseudotyped with H1N1 or H18, for 4 h at 37 °C. Entry- indicate significantly upregulated transcripts, filtered according to the positive cells were measured with the fluorogenic β-lactamase substrate Gene Ontology annotations GO:0016020 (membrane), GO:0016021 CCF2-AM, and quantified by flow cytometry. Relative numbers of entry- (integral component of membrane) and GO:0005886 (plasma membrane), positive cells are shown. **, signals from A549 were below the detection that overlap in the H18N11-susceptible cell lines. f, HLA-DR surface limit, owing to inefficient loading with the fluorogenic β-lactamase staining of U-87MG, U-118MG, Calu-3 and A549 cells, and DLA surface substrate. a, b, Data are mean ± s.d. from n = 3 independent experiments. staining of MDCKII clone no. 1 (here labelled MDCKII #1) and MDCKII Values below background levels of 1 are displayed on the x axes. clone no. 2 (here labelled MDCKII #2) cells, is shown in contour plots. c–e, Transcriptional profiles of the indicated H18N11-susceptible versus Representative plots of n = 3 independent experiments are shown. non-susceptible cell lines are shown as volcano plots. The two-group RESEARCH Letter

Extended Data Fig. 2 | Genome-wide CRISPR–Cas9 screen on U-87MG with VSV-H18 at an MOI of 10. Survivors were re-infected with VSV- cells to identify potential receptor candidates for bat influenza virus. H18 at an MOI of 10 for 4 additional rounds, and collected on day 27 p.i. a, Schematic of the genome-wide CRISPR–Cas9 screen. U-87MG cells Genomic DNA from cell pellets was isolated and analysed using next- stably expressing Cas9 (U-87MG–Cas9) were transduced with lentiviruses generation sequencing. b, Susceptibility of the survivors and the wild-type encoding a human CRISPR-knockout pooled sgRNA library at an MOI U-87MG cells was monitored by flow cytometry measuring GFP signals of 0.3. Two days after transduction, cells (U-87MG–Cas9–sgRNA) were from cells infected with H18-VSV reporter virus at different MOIs. selected with puromycin (1 μg/ml) for ten days, and subsequently infected Relative numbers of GFP-positive cells are shown. Letter RESEARCH

Extended Data Fig. 3 | HLA-DR is required for bat IAV infection in d, e, Indicated U-87MG cell lines were infected with VSV-H18 reporter U-87MG and Calu-3 cells. a) U-87MG control, U-87MG KO no. 1 (here virus expressing GFP at an MOI of 5 for 48 h. Representative images labelled KO #1) and U-87MG KO no. 2 (here labelled KO #2) cells were are shown in d. Frequency of GFP-positive cells was quantified by flow infected with recombinant VSV-H18 (red), VSV (light grey) and H7N7 cytometry and is depicted in e. a, e, Data are mean ± s.d. from n = 3 (dark grey) reporter viruses expressing GFP at an MOI of 1, 0.1 and 0.1, independent experiments. Values below background levels of 1 are respectively, for at least 24 h. The frequency of GFP-positive cells was displayed on the x axes. f, HLA-DR surface staining of Calu-3 HLA-DRA quantified by flow cytometry. b, U-87MG control, U-87MG KO no. 1 knockout cells (labelled ‘KO #1’ and ‘KO #2’) and Calu-3 cells (labelled and U-87MG KO no. 2 (labelled ‘KO #1’ and ‘KO #2’) were infected ‘Ctrl’) is shown from a representative of n = 2 independent experiments. with recombinant VSV-H17 expressing GFP at an MOI of 1 for 72 h. g, Calu-3 control, Calu-3 KO no. 1 and Calu-3 KO no. 2 cells (labelled Representative images of n = 3 independent experiments are shown. ‘KO #1’ and ‘KO #2’) were infected with BlaM1 VLPs pseudotyped with Scale bar, 100 μm. c, HLA-DR surface staining of U-87MG control, H1N1 or H18 for 4 h at 37 °C. Entry-positive cells were measured with U-87MG KO no. 1 and U-87MG KO no. 2 (labelled ‘KO #1’ and ‘KO the fluorogenic β-lactamase substrate CCF2-AM, and quantified by flow #2’), stably transduced with either a control lentivirus (labelled ‘LV-ctrl’) cytometry. Relative numbers of entry-positive cells are shown. Data are or a lentivirus encoding HLA-DRA (labelled ‘LV-HLA-DRA’) from a mean from n = 2 independent experiments. Values below background representative of n = 3 independent experiments is shown. levels of 1 are displayed on the x axis. RESEARCH Letter

Extended Data Fig. 4 | Alignment of protein sequences of receptor-binding site in conventional IAV are highlighted by light purple haemagglutinins from different subtypes of IAV. Residues conserved boxes, and the fusion peptide is marked by a yellow box. across all subtypes are highlighted in red. Residues known to be part of the Letter RESEARCH

Extended Data Fig. 5 | Ectopic expression of HLA-DR renders non- 24 h p.i. for A549 LV control and A549 LV HLA-DR (left), at 24 h p.i. for susceptible cell lines susceptible. a, HLA-DR surface staining of MDCKII clone no. 2 LV control and MDCKII clone no. 2 LV HLA-DR A549 LV control (LV-ctrl) and A549 LV HLA-DR (LV-HLA-DR) (left), (middle), and at 72 h p.i. for U-118MG LV control and U-118MG LV MDCKII clone no. 2 LV control and MDCKII clone no. 2 LV HLA-DR HLA-DR (right). b, c, Data are mean ± s.d. from n = 3 independent (middle), U-118MG LV control and U-118MG LV HLA-DR (right) from experiments. Values below background levels of 1 are displayed on the a representative of one experiment is shown. b, Indicated cell lines were x axes. d, MDCKII clone no. 2 LV control and MDCKII clone no. 2 LV infected with luciferase-encoding VLPs pseudotyped with H18 or H1N1, HLA-DR were infected with IAV H18N11 at MOI = 0.001. At indicated or control VLPs that lack any surface glycoprotein (empty vector). Fold time points p.i., supernatants were collected and viral titres were increase in luciferase signals relative to samples infected with empty determined by immunofluorescence on subconfluent MDCKII clone no. vector is plotted. c, Indicated cell lines were infected with VSV-H18 at an 1 cells, using rabbit polyclonal anti-H18. Data are mean ± s.d. from n = 3 MOI of 10 and GFP-positive cells were quantified by flow cytometry at independent experiments. RESEARCH Letter

Extended Data Fig. 6 | Polykaryon formation assay and PLAs detect highlighting the principal steps of the emulsion coupling assay. Emulsion interaction of H18 with HLA-DR complexes. a, HEK293T cells were coupling is a digital assay concept based on the detection of double- co-transfected with plasmids encoding mCherry and H1. Transfected labelled, individual molecular complexes in emulsion, which are identified HEK293T cells were treated with TPCK–trypsin to cleave haemagglutinin by ddPCR. In a first step, MDCKII clone no. 2 cells that express HLA-DR and seeded on top of control MDCKII clone no. 2 cells (labelled as LV- that contains a haemagglutinin tag (HLA-DR–HA tag) were incubated ctrl) or MDCKII clone no. 2 cells that stably express HLA-DR (labelled with VSV-H18 at an MOI of 100 for 15 min at room temperature (A). To as LV-HLA-DR), in the presence or absence of the indicated antibodies. stabilize and prevent the dissociation of complexes in the following steps, Polykaryon formation was examined after exposure of cells to low pH. the interaction was fixed with a membrane-impermeable cross-linker. Representative images of n = 3 independent experiments are shown. Cell Cells were lysed and the cell lysate was incubated with single-stranded- nuclei were stained by DAPI (blue). Scale bar, 25 μm. b, Visualization of DNA-labelled antibodies (B) that specifically recognize HLA-DR–HA H18 and HLA-DR complexes by PLA. MDCKII clone no. 2 cells (LV-ctrl) or H18, to achieve equilibrium binding (C). Before emulsification, the or MDCKII clone no. 2 cells that express HLA-DR (LV-HLA-DR) were reaction was highly diluted (~100,000 times) to achieve single-complex infected with VSV-H18 at an MOI of 100 at 37 °C. After 1 h, cells were separation. ddPCR was carried out using the standard ddPCR protocol, fixed and subjected to PLA using mouse anti-H18 and rabbit anti-MHC-II and the reactions were measured using a QX200 Droplet Digital PCR antibodies. Specific PLA dots obtained after amplification of the PLA instrument (D). The evaluation of the reaction was based on the fluorescence probe are indicated by red dots. Cell nuclei were stained partitioning of the labels in the ddPCR reactions using fluorescently by DAPI (blue colour). Representative images of three independent tagged PCR products (FAM- or VIC-labelled), and indicated as RFU (see experiments are shown. Scale bars, 20 μm. c, Scatter plot of the PLA dot scatter plot). According to the ddPCR standard evaluation, the number count per cell obtained from PLA images taken under the conditions given of labelled antibodies was determined in each reaction (counting all in c is shown from a representative of three independent experiments. The label-positive droplets for a given label, and using the same definition of cell number, counted at 37 °C, for LV-ctrl cells infected with VSV-H18, clusters for all reactions). Without complexes (interacting proteins with LV-HLA-DR mock-infected cells or LV-HLA-DR cells infected with VSV- two antibodies bound), the partitioning of the labelled antibodies follows H18 was 9,254, 10,747 and 14,966, respectively. Statistical significance a Poisson distribution, and results in a calculable number of double- − was assessed by Mann–Whitney U test (two-sided). ***P ≤ 2.2 × 10 16. coloured droplets. In the case of complexes present in the reaction, the d, MDCKII clone no. 2 cells (LV-ctrl) or MDCKII clone no. 2 cells that number of the detected double-coloured droplets is larger than would be express HLA-DR (LV-HLA-DR–HA-tag) were infected with VSV-H18 expected by Poisson distribution. We developed a Python code to calculate at an MOI of 10. GFP-positive cells were quantified by flow cytometry the number of complexes, which explains this difference on the basis of at 24 h p.i. Data are mean ± s.d. from n = 3 independent experiments. Poisson-based modelling of the partitioning of the molecular species of Values below background levels of 1 are displayed on the x axes. e, Cartoon the reaction; this results in the absolute counts for the detected complexes. Letter RESEARCH

Extended Data Fig. 7 | HLA-DR homologues from different species as a positive control; cells transfected with empty vector served as a confer susceptibility to bat IAV. a, HEK293T cells were transfected negative control. At 48 h after transfection, cells were infected with GFP with plasmids encoding CIITA, HLA-DRA or HLA-DRB1, and co- encoding VSV-H18 (b) or VSV-H17 (c) at an MOI of 1. At 48 h (b) or transfected with HLA-DRA and HLA-DRB1 or empty vector. At 48 h 72 h (c) p.i., the frequency of GFP-positive cells was quantified by flow after transfection, cells were infected with VSV-H18 at an MOI of 10. cytometry. d, CIITA from P. al e c to (PaCIITA) was stably expressed by Fluorescent microscopy images were taken at 72 h p.i. Representative lentiviral transduction in kidney-derived cell lines from bat species images of n = 3 independent experiments are shown. Scale bar, 100 μm. S. lilium and P. al e c to . Cells were infected with VSV-H18 at an MOI of b, c, HEK293T cells were co-transfected with plasmids encoding 10 for at least 24 h. The frequency of GFP-positive cells was quantified the α- and β-chains of HLA-DR homologues from different species: by flow cytometry. b–d, Data are mean ± s.d. from n = 3 independent S. scrofa (SLA-DR), G. gallus (B-L), E. fuscus (E. fuscus DR), M. lucifugus experiments. Values below background levels of 1 are displayed on the (M. lucifugus DR) and P. al e c to (P. al e c to DR). HLA-DR was included x axes. RESEARCH Letter

Extended Data Fig. 8 | Alignment of protein sequences of surface HLA-DQA1), S. scrofa (SLA-DRA), G. gallus (B-LA), E. fuscus (EfDRA), MHC-II α-chains from different species. Sequences included are the M. lucifugus (MlDRA) and P. al e c to (PaDRA). Residues conserved across α-chains of MHC-II from Homo sapiens (HLA-DRA, HLA-DPA1 and all species are highlighted in red. Letter RESEARCH

Extended Data Fig. 9 | Alignment of protein sequences of surface DQB1), S. scrofa (SLA-DRB1), G. gallus (B-LB2), E. fuscus (EfDRB5), MHC-II β-chains from different species. Sequences included are the M. lucifugus (MlDRB5) and P. al ec to (PaDRB). Residues conserved across β-chains of MHC-II from H. sapiens (HLA-DRB1, HLA-DPB1 and HLA- all species are highlighted in red. RESEARCH Letter

Extended Data Fig. 10 | See next page for caption. Letter RESEARCH

Extended Data Fig. 10 | MHC-II knockout mice are resistant to bat IAV inset (B) provides an overview of the epithelium (arrow) highlighted in infection. a, HEK293T cells were co-transfected with plasmids encoding panel A, and shows unspecific immunoreactivity in the lumen of a blood the α- and β-chains of mouse MHC-II H-2A or H-2E or transfected with vessel in the lamina propria (arrowhead). Panel C shows intense matrix control plasmid. At 48 h after transfection, cells were infected with BlaM1 protein immunoreactivity in a multifocal-to-coalescing cytoplasmic VLPs pseudotyped with H1N1 or H18 for 4 h at 37 °C. Entry-positive cells pattern, occasionally with an apically increased intensity (arrow). Some were measured with the fluorogenic β-lactamase substrate CCF2-AM nuclei within the immunoreactive cells are also labelled (arrowhead). The and quantified by flow cytometry. b, HEK293T cells were co-transfected inset (D) provides an overview of the localization of the immunoreactive with plasmids encoding the α- and β-chains of mouse MHC-II H-2A or respiratory epithelium (arrow) on the cranial nasal septum in the area H-2E, or transfected with control plasmid. At 48 h p.i., cells were infected of the transition from transitional (left side) to respiratory (right side) with VSV-H18 expressing GFP at an MOI of 10 for 48 h. Representative epithelium. In mock-infected animals (panel E), respiratory epithelium images are shown on the left, and a quantification of GFP-positive cells never displayed immunoreactivity. Only faint unspecific immunoreactivity by flow cytometry is depicted on the right. a, b, Data are mean ± s.d. was observed in the lumen of a blood vessel in the lamina propria from n = 3 independent experiments. Values below background levels (arrow) in panel E. The inset (F) provides an overview of the epithelium of 1 are displayed on the x axes. c, Detection of viral antigen in paraffin- highlighted in panel E. Scale bars, 10 μm (panels A, C, E), 50 μm (panels embedded tissue from B6 mice intranasally infected with H18N11 B, D, F). d, Detection of viral antigen and viral RNA in paraffin-embedded (n = 4, 1 × 105 ffu in 40 μl, panels A and B) or H3N2 (n = 3, 1 × 103 tissue of B6 mice intranasally infected with H3N2 (n = 3) for four days as pfu in 20 μl, panels C and D) or mock-infected (n = 2, panels E and F) described in c. A moderate amount of oligo- to multifocally distributed at 4 days p.i. IAV matrix protein M1 immunoreactivity was observed IAV matrix protein (panel A) but no H18-haemagglutinin-specific in an oligofocal and laminar pattern in the apical ciliary border of the RNA (panel B) was detectable in the respiratory epithelium in following epithelium (arrow), and in epithelial cells (arrowhead), which displayed sections of the nasal cavities by using immunohistochemistry and in situ equivocal faint cytoplasmic and nuclear immunoreactivity (panel A). The hybridization, respectively. nature research | life sciences reporting summary

Corresponding author(s): Martin Schwemmle & Silke Stertz

Life Sciences Reporting Summary Nature Research wishes to improve the reproducibility of the work that we publish. This form is intended for publication with all accepted life science papers and provides structure for consistency and transparency in reporting. Every life science submission will use this form; some list items might not apply to an individual manuscript, but all fields must be completed for clarity. For further information on the points included in this form, see Reporting Life Sciences Research. For further information on Nature Research policies, including our data availability policy, see Authors & Referees and the Editorial Policy Checklist.

Please do not complete any field with "not applicable" or n/a. Refer to the help text for what text to use if an item is not relevant to your study. For final submission: please carefully check your responses for accuracy; you will not be able to make changes later.

` Experimental design 1. Sample size Describe how sample size was determined. No sample-size calculations were performed. Sample size was determined to be adequate based on the magnitude and consistency of measurable differences between groups. 2. Data exclusions Describe any data exclusions. No data were excluded from the analysis. 3. Replication Describe the measures taken to verify the reproducibility For all major experiments at least three independent experiments were done and in all cases of the experimental findings. results could be reproduced. For data shown in fig. 2j, 3a, S3f, S3g two independent experiments were performed. The HLA-DR surface staining shown in fig S5a was only performed once. The number of repeats for each experiment is reported in the figure legends. 4. Randomization Describe how samples/organisms/participants were Allocation of samples and/or mice to groups was random. allocated into experimental groups. 5. Blinding Describe whether the investigators were blinded to Investigators were not blinded to group allocation in this study. Key experiments were group allocation during data collection and/or analysis. repeated independently by multiple members in two different laboratories. Note: all in vivo studies must report how sample size was determined and whether blinding and randomization were used.

6. Statistical parameters For all figures and tables that use statistical methods, confirm that the following items are present in relevant figure legends (or in the Methods section if additional space is needed). n/a Confirmed

The exact sample size (n) for each experimental group/condition, given as a discrete number and unit of measurement (animals, litters, cultures, etc.) A description of how samples were collected, noting whether measurements were taken from distinct samples or whether the same sample was measured repeatedly A statement indicating how many times each experiment was replicated The statistical test(s) used and whether they are one- or two-sided Only common tests should be described solely by name; describe more complex techniques in the Methods section.

A description of any assumptions or corrections, such as an adjustment for multiple comparisons November 2017 Test values indicating whether an effect is present Provide confidence intervals or give results of significance tests (e.g. P values) as exact values whenever appropriate and with effect sizes noted. A clear description of statistics including central tendency (e.g. median, mean) and variation (e.g. standard deviation, interquartile range) Clearly defined error bars in all relevant figure captions (with explicit mention of central tendency and variation)

See the web collection on statistics for biologists for further resources and guidance.

1 ` Software nature research | life sciences reporting summary Policy information about availability of computer code 7. Software Describe the software used to analyze the data in this STAR v2.5.1, R package GenomicRanges from Bioconductor Version 3.0, MAGeCK software, study. EnVision Manager Version 1.14.3049.528, GraphPad Prism 7.03, BD FACSuite software, FlowJo v10 software, MegAlign v12.2.0, Leica LAS AF Lite v. 2.6.3, Python version 2.7, ImageJ, Quantasoft v.1.7.4.0917, AxioVision Rel 4.8

For manuscripts utilizing custom algorithms or software that are central to the paper but not yet described in the published literature, software must be made available to editors and reviewers upon request. We strongly encourage code deposition in a community repository (e.g. GitHub). Nature Methods guidance for providing algorithms and software for publication provides further information on this topic.

` Materials and reagents Policy information about availability of materials 8. Materials availability Indicate whether there are restrictions on availability of All unique materials are available from the authors without restriction for non-commercial unique materials or if these materials are only available use. for distribution by a third party. 9. Antibodies Describe the antibodies used and how they were validated The following commercially available antibodies were used: for use in the system under study (i.e. assay and species). - APC-labelled anti-human HLA-DR antibody (L243, Biolegend, catalog# 307610, 1:50); validation by siRNA knockdown of HLA-DRA - AlexaFluor 488-labelled anti-mouse IgG (Jackson ImmunoResearch, catalog# 115-546-062, 1:500); validation by omission of the primary antibody - Monoclonal mouse anti IAV-matrix protein IgG1 (ATCC, clone M2-1C6-4R332, 1:100); validation on H18N11-infected and mock-infected cells - HA-tag antibody (Sigma; catalog#H6908, used at 100ug/ml); validation on MDCKII#2 cells expressing or lacking HLA-DR-HA-tag - HLA-ABC monoclonal (MHCI) antibody (W6/32; PMID: 87477, used at 100ug/ml); validation on cells expressing or lacking MHCI. - Rabbit monoclonal HLA-DR antibody (Abcam; catalog# ab92511, 1:500); validation on MDCKII#2 cells expressing or lacking HLA-DR. - Anti-HLA-DR ,-DP, -DQ antibody (anti-MHC-II, clone Tü39, Biolegend, catalog# 361702, used at 20ug/ml, 5, ug/ml, 1.25ug/ml, 0.3125 ug/ml); validation on HLA-DR-transfected 293T cells and HLA-DR, -DQ, -DP expressing U-87MG cells - Anti-6xHis-tag antibody (clone HIS.H8, Abcam, catalog# ab18184, used at 20ug/ml, 5, ug/ml, 1.25ug/ml, 0.3125 ug/ml); validation on 293T cells transfected with His-tagged proteins - APC-coupled anti-canine MHC Class II (ThermoFisher Scientific, catalog# 17-5909-41, 1:25); validation on canine peripheral blood mononuclear cells (PBMCs) -Purified anti-human HLA-DR Antibody (clone L243, Biolegend, catalog# 307602, 1:500 ); validation on HLA-DR-transfected 293T cells.

The following antibodies were generated within the frame of this study: rabbit polyclonal anti-H18 serum at 1:750; validation on H18N11-infected and mock-infected cells mouse monoclonal anti-NP antibody (supernatant); validation on H18N11-infected and mock- infected cells mouse polyclonal H18 antibody at 1:2500; validation on VSV-H18-infected MDCKII#2 cells expressing HLA-DR and mock-infected cells.

November 2017

2 10. Eukaryotic cell lines nature research | life sciences reporting summary a. State the source of each eukaryotic cell line used. Human embryonic kidney cells 293T, human lung adenocarcinoma cell lines A549 and Calu-3 were purchased from the American Type Culture Collection (ATCC). Bat cell lines PaKi and SliKi were described previously (ref. 14,16). MDCKII #1 were obtained from Georg Herrler, University of Veterinary Medicine Hannover, Hannover, Germany and previously described (ref. 17). MDCKII #2 cells were obtained from ECACC. MDCKII #1 and MDCKII #2 differ in the number of passages. Human glioblastoma cell lines U-87MG and U-118MG were originally from ATCC, provided by Dorothee Holm-von Laer, University of Innsbruck, Innsbruck, Austria and also previously described (ref. 17).

b. Describe the method of cell line authentication used. A549, Calu-3 and MDCK cell lines were verified to be of human or canine origin, respectively, by sequencing the COX1 transcript. No additional cell line authentication was performed.

c. Report whether the cell lines were tested for Cell lines used in the Stertz and the Schwemmle lab are routinely tested for mycoplasma mycoplasma contamination. contamination by sending representative samples for mycoplasma testing at GATC, Germany. None of the used cell lines ever tested positive for mycoplasma.

d. If any of the cell lines used are listed in the database U118-MG cells are listed as commonly misidentified cell line as some samples have been of commonly misidentified cell lines maintained by found to be contaminated with U138-MG cells, another human glioblastoma cell line. As we ICLAC, provide a scientific rationale for their use. only use U118-MG as a control non-susceptible cell line in the transcriptome analysis it would not impact our results.

` Animals and human research participants Policy information about studies involving animals; when reporting animal research, follow the ARRIVE guidelines 11. Description of research animals Provide all relevant details on animals and/or 8-10 weeks B6 (29 female and 5 male), B6.Cg-Thy1a-H2-Aatm1Blt (10 female and 5 male) and animal-derived materials used in the study. pIV-/- K14 CIITA Tg mice (5 female) were infected intranasally with the indicated virus strains.

Policy information about studies involving human research participants 12. Description of human research participants Describe the covariate-relevant population This study did not involve human research participants. characteristics of the human research participants. November 2017

3