key: cord-254957-jqp1gto6 authors: Klann, Kevin; Bojkova, Denisa; Tascher, Georg; Ciesek, Sandra; Münch, Christian; Cinatl, Jindrich title: Growth factor receptor signaling inhibition prevents SARS-CoV-2 replication date: 2020-08-11 journal: Mol Cell DOI: 10.1016/j.molcel.2020.08.006 sha: doc_id: 254957 cord_uid: jqp1gto6 Summary SARS-CoV-2 infections are rapidly spreading around the globe. The rapid development of therapies is of major importance. However, our lack of understanding of the molecular processes and host cell signaling events underlying SARS-CoV-2 infection hinder therapy development. We employed a SARS-CoV-2 infection system in permissible human cells to study signaling changes by phospho-proteomics. We identified viral protein phosphorylation and defined phosphorylation-driven host cell signaling changes upon infection. Growth factor receptor (GFR) signaling and downstream pathways were activated. Drug-protein network analyses revealed GFR signaling as key pathway targetable by approved drugs. Inhibition of GFR downstream signaling by five compounds prevented SARS-CoV-2 replication in cells, assessed by cytopathic effect, viral dsRNA production, and viral RNA release into the supernatant. This study describes host cell signaling events upon SARS-CoV-2 infection and reveals GFR signaling as central pathway essential for SARS-CoV-2 replication. It provides with novel strategies for COVID-19 treatment. Co-expression clustering Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), a novel coronavirus, has been rapidly spreading around the globe since the beginning of 2020. In people, it causes coronavirus disease 2019 (COVID-19) often accompanied by severe respiratory syndrome (Chen et al., 35 2020) . To conquer the global health crisis triggered by COVID-19, rapidly establishing drugs is required to dampen the disease course and relieve healthcare institutions. Thus, repurposing of already available and (ideally) approved drugs might be essential to rapidly treat COVID-19. Many studies for proposing repurposing of specific drugs have been conducted in the last months, but mostly remain computational without tests in infection models (Smith and Smith, 40 2020; Wang, 2020) . In addition, they are hindered by the lack of knowledge about the molecular mechanisms of SARS-CoV-2 infection and the resulting host-cell responses required to allow 2 viral replication. To rationally repurpose drugs, a molecular understanding of the infection and the changes within the host cell pathways is essential. Experimentally identifying viral targets in the cell allows candidate drugs to be selected with high confidence for further testing in the 45 clinics to reduce the risks for patients resulting from tests with drugs lacking in vitro validation. Growth factor receptor (GFR) signaling plays important roles in cancer pathogenesis and has also been reported to be crucial for infection with some viruses (Beerli et al., 2019; Kung et al., 2011; Zhu et al., 2009) . GFR activation leads to the modulation of a wide range of cellular processes, including proliferation, adhesion, or differentiation (Yarden, 2001) . Various viruses, 50 such as Epstein-Barr virus, influenza, or hepatitis C, have been shown to use the epidermal growth factor receptor (EGFR) as an entry receptor (Eierhoff et al., 2010; Kung et al., 2011; Lupberger et al., 2011) . In addition, EGFR activation can suppress interferon signaling and thus the antiviral response elicited in respiratory virus diseases, for instance influenza A and rhinovirus (Ueki et al., 2013) . Activation of GFR signaling might play an important role also in 55 other respiratory viruses, such as SARS-CoV-2. In the last years, it has been shown for many viruses that modulation of host cell signaling is crucial for viral replication and might exhibit strong therapeutic potential (Beerli et al., 2019; Pleschka et al., 2001) . However, how SARS-CoV-2 infection changes host cell signaling has remained unclear. We recently established an in vitro cell culture model of SARS-CoV-2 60 infection using the colon epithelial cell line Caco-2, which is highly permissive for the virus and commonly used for the study of coronaviruses (Herzog et al., 2008; Ren et al., 2006) . Here, we determine changes in the cellular phospho-protein networks upon infection with SARS-CoV-2 to gain insight into infection-induced signaling events. We found extensive rearrangements of cellular signaling pathways, particularly of GFR signaling. Strikingly, inhibiting GFR signaling 65 using prominent (anti-cancer) drugs -namely pictilisib, omipalisib, RO5126766, lonafarnib, and sorafenib -prevented SARS-CoV-2 replication in vitro, assessed by cytopathic effect and viral RNA replication and release. These compounds prevented replication at clinically achievable concentrations. Due to their clinical availability, these drugs could be rapidly transitioned towards clinical trials to test their feasibility as COVID-19 treatment option. 70 In a previous study, we analyzed the effect of SARS-CoV-2 infection on the host cell translatome and proteome (Bojkova et al., 2020) . This study found the effects 24 hours after SARS-CoV-2 infection especially useful for identifying druggable host pathways. To evaluate changes in 75 intracellular signaling networks brought about by SARS-CoV-2 infection, we quantified phosphoproteome changes 24 hours after infection ( Figure 1A ). Caco-2 cells were mock-infected or infected with SARS-CoV-2 patient isolates (in five biological replicates at an MOI of 1) for one hour, washed, and incubated for 24 hours before cell harvest. Extracted proteins were digested and split to 1) carry out whole-cell proteomics of a tandem mass tag (TMT) 10-plex samples 80 using liquid chromatography synchronous precursor selection mass spectromety (LC-SPS-MS 3 ), or 2) use Fe-NTA phosphopeptide enrichment (achieving 98% enrichment) for phospho proteome analyses of a TMT 10-plex analyzed by LC-MS 2 , due to the higher precision and identification rates of MS 2 based methods during phosphopeptide measurements (Hogrebe et al., 2018) . We identified and quantified 7,150 proteins and 16,715 different phosphopeptides for 85 a total of 15,093 different modification sites ( Figure 1B , C, S1, and Table S1 , S2). The main fraction of phosphopeptides were modified serines (86.4%), followed by threonine (13.4%), and tyrosine (0.2%) ( Figure 1D ). Upon infection, 2,197 and 799 phosphopeptides significantly increased or decreased respectively (log 2 FC > 1, p value < 0.05). 3 Viral proteins are produced in the host cell and underlie (and often require) post-translational 90 modification (PTM) by host cell enzymes (Wu et al., 2009) . Accordingly, we assessed viral proteins phosphorylated in the host cell. We identified 33 modification sites on 6 different viral proteins ( Figure 1E -J). Possible functions of the observed modifications largely remain unclear due to a lack of understanding of their molecular function and regulation. SARS-CoV-2 protein 3a was phosphorylated on the luminal side of this transmembrane protein ( Figure 1E ) Membrane 95 protein M was phosphorylated at three serines in close proximity, at the C-terminal, cytoplasmic region of the protein ( Figure 1F ), suggesting a high-activity modification surface. SARS-CoV-1 protein 6 was described to accelerates infections in murine systems (Tangudu et al., 2007) . We found a single phosphorylation of the SARS-CoV-2 protein homologue non-structural protein 6 in host cells ( Figure 1G ) Protein 9b was modified at two sites ( Figure 1H ). However, its function in 100 SARS-CoV-1 or SARS-CoV-2 remains unknown. Polyprotein 1b is a large 7,096 amino acid protein heavily processed to generate distinct proteins in SARS-CoV-1 (Tangudu et al., 2007) . We found polyprotein 1b to be modified at three residues, two in a region of unknown function and one in the non-structural protein 11 (NSP11) part of the protein ( Figure 1I ). Our data cannot distinguish whether phosphorylation occurred before or after cleavage and whether 105 phosphorylation may affect processing. SARS-CoV-2 nucleoprotein was heavily phosphorylated ( Figure 1J ). Mapping phosphosites to the structure (residues 47 to 173, PDB: 6vyo) revealed a small surface region, suggesting specific regulation and interaction changes ( Figure 1K ). To reveal host kinases potentially phosphorylating viral proteins, we bioinformatically assessed identified phosphorylation motifs using NetPhos 3.1 and GPS5 (Blom et al., 2004; Wang et al., 110 2020) (Table S3) . Some motifs present in nucleoprotein were predicted to be modified by CMGC kinases. Among several others, casein kinase II (CK2) kinases are part of the CMGC family and have been independently identified as interaction partners of the nucleoprotein, when expressed in cells (Gordon et al., 2020) . Inhibition of CK2 kinases, could be employed to study possible functional interactions between kinase and viral protein. Taken together, we identified extensive changes in phosphorylation of host and viral proteins after SARS-CoV-2 infection. The roles of viral protein modifications remain unclear. However, targeting the corresponding host kinases may offer new treatment strategies. To identify the key host signaling pathway networks modulated by infection, we carried out 120 protein-protein co-regulation analysis on all proteins quantified in phosphorylation and total protein level. We first standardized phosphorylation and total protein levels by individual Zscoring to compare the different datasets. Subsequently, to merge phosphorylation and proteome data, we collapsed all phospho-sites for each protein into one average profile and calculated combined Z-scores. Patterns of co-regulation were identified using protein-protein 125 correlation and hierarchical clustering ( Figure 2A ). This generalized approach allows to study large scale patterns of dependencies of protein and phosphorylation levels, that then can be dissected into individual phosphorylation sites and protein levels for downstream analysis. The dynamic landscape of the proteome revealed three main clusters of co-regulated proteins, each one representing different sets of pathways (discussed in detail below): 130 The first cluster was mainly comprised of receptor signaling and endocytic pathways ( Figure 2B ). Prominent among these pathways were platelet derived growth factor receptor (PDGFR), ErbB1 (EGFR) signaling, metabolism, and various pathways associated with vesicle trafficking (Table S4 ). As changes in phospho-peptide abundance can represent different ratios in phosphorylated versus non-phosphorylated peptide or a change in protein abundance (with the same ratio of 135 protein being phosphorylated), we integrated our phospho-proteome dataset with total proteome data ( Figure 2C ). When comparing abundances of individual phosphopeptides and their protein levels, extensive changes were observed in the phospho-proteome; however, no general J o u r n a l P r e -p r o o f 4 changes were seen for the total proteome ( Figure 2C , Table S2 ). Thus, phosphorylation changes were induced by signaling activity alteration resulting in increased phosphorylation and not due 140 to protein abundance differences. The second cluster was mainly comprised of proteins decreased in phosphorylation and was highly connected to cell cycle and translation initiation ( Figure 2D and Table S4 ). We reported recently that inhibition of cellular translation prevented SARS-CoV-2 replication in cells (Bojkova et al., 2020) , consistent with regulation of translation by altering phosphorylation patterns. To 145 further distinguish the regulations within this cluster, we correlated protein levels with differential phosphorylation abundance ( Figure 2E ) and found two groups of proteins: The first was contained translation related pathways (identified in Figure 2E ) and was predominantly regulated by decreased modification. The second set of proteins was decreased in phosphorylation and on total protein level. The majority of proteins found in the second cluster belonged to diverse cell 150 cycle pathways. Consistent with these findings, cell cycle pathways were also enriched in the set of proteins significantly decreased on protein level ( Figure 2F , S2, and Table S4 and S5). Translation pathways were not regulated on protein level to this extent. Analysis of the third cluster revealed signaling events of the splicing machinery (Table S4) possibly explaining previously observed changes in splicing machinery abundance upon SARS-155 CoV-2 infection (Bojkova et al., 2020) . Consistent with previous literature (Grimmler et al., 2005; Ilan et al., 2017; Mathew et al., 2008; Mermoud et al., 1994) , we therefore hypothesized that the host splicing machinery is extensively reshaped during viral infection. This finding further supports splicing as a potential therapeutic target, in agreement with decreased SARS-CoV-2 pathogenic effects when inhibiting splicing by pladeinolide B. Additionally, we found carbon 160 metabolism among the pathways showing significantly increased phosphorylation upon SARS-CoV-2 infection (Table S4 ) in addition to previously described changes of total protein levels of enzymes part of glycolysis and carbon metabolism (Bojkova et al., 2020 )( Figure S3 ). Taken together, we showed that, during SARS-CoV-2 infection, specific rearrangements of signaling pathways were elicited in the cellular proteome. Regulation was mainly comprised of 165 cellular signaling and translational pathways as well as proteins regulated not only by phosphorylation, but also in total protein abundance. Proteins exhibiting decreased protein levels were significantly enriched in cell cycle proteins. We observed over 2,000 phospho-peptides to be increased in abundance while their protein 170 levels stay constant upon infection ( Figure 2C and S4). This reveals differential modification activity (e.g. signaling events) for these phospho-proteins. For many kinases in cellular signaling pathways there are already approved drugs available. Hence, we investigated the potential to repurpose drugs to treat COVID-19 by mapping already available drugs via ReactomeFI to the set of proteins increased in phosphorylation. We filtered the network for drugs and direct targets 175 and found EGFR as one of the central hits, including a number of regulated proteins in the downstream signaling pathway of EGFR ( Figure 3A ). These downstream targets are also regulated by other GFRs and could thus also be explained by their observed activation upon SARS-CoV-2 infection ( Figure 2 ). 28 clinically approved drugs (largely used in cancer therapy) are already available to target EGFR or downstream targets. Indeed, we found a subnetwork of 180 GFR signaling components remodeled ( Figure 3B ). We mapped identified members of GFR signaling and their respective phosphorylation differences upon SARS-CoV-2 infection ( Figure 3C ) revealing an extensive overall increase in phosphorylation of the whole pathway, including related components for cytoskeleton remodeling and receptor endocytosis. How GFR signaling might regulate SARS-CoV-2 infection is still matter for speculation. However, GFR signaling 185 J o u r n a l P r e -p r o o f 5 inhibition might provide a useful approach already implicated in SARS-CoV induced fibrosis therapy and might be a viable strategy to treat COVID-19. Since GFR signaling seemed to be central during SARS-CoV-2 infection, we examined the use of inhibitors as antiviral agents. Since there are several GFRs integrating their signaling and 190 regulating a number of processes inside the cell, directly targeting downstream signaling components is likely to be more successful to prevent signaling of different GFRs and to avoid mixed effects of multiple pathways. GFR signaling, amongst others, results in activation of 1) the RAF/MEK/ERK MAPK signaling cascade and 2) integrates (via phosphoinositide 3 kinase [PI3K] and protein kinase B [AKT]) into mTORC1 signaling to regulating proliferation ( Figure 4A ). To 195 explore the antiviral efficiency of targeting proteins downstream of GFRs, we first tested the PI3K inhibitors pictilisib and omipalisib (Ippolito et al., 2016; Sarker et al., 2015; Schmid et al., 2016) . Both compounds inhibited viral replication, based on their propensity to prevent cytopathogenic e ect (CPE) and viral RNA production in cells ( Figure 4B -D, S5, and S6). Our drug-target analyses identified mitogen activated protein kinase kinase (MAP2K2, better known as MEK) 200 and the RAF inhibitor sorafenib (Wilhelm et al., 2006) as promising targets inhibiting downstream signaling of GFRs ( Figure 4A ). Thus, we tested sorafenib and the dual RAF/MEK inhibitor RO5126766 in our viral replication assays. Both compounds inhibited cytopathic effects during infection and the viral replication ( Figure 4B -D, S5, and S6). To validate our findings in another cell line, we repeated the treatments in UKF-RC-2 cells infected with SARS-CoV-2. Quantifying 205 the viral RNA copies in the supernatant, we observed that the compounds efficiently inhibited virus replication ( Figure 4E and S7A) at non-toxic conditions ( Figure S7B , except for sorafenib). Overall, five compounds, inhibiting downstream signaling of GFRs, prevented SARS-CoV-2 replication at clinically achievable concentrations ( Figure 4B and 5) (Eskens et al., 2001; Fucile et al., 2014; Martinez-Garcia et al., 2012; Munster et al., 2016; Sarker et al., 2015) , emphasizing 210 the importance of GFR signaling during SARS-CoV-2 infection and revealing clinically available treatment options as drug candidates for COVID-19. With the rapid spreading of the COVID-19 pandemic, investigating the molecular mechanisms underlying SARS-CoV-2 infection are of high importance. In particular, the processes underlying 215 infection and host-cell response remain unclear. These would offer potential avenues for pharmacological treatment of COVID-19. Here, we report global, differential phosphorylation analysis of host cells after infection with intact SARS-CoV-2 virus. We could identify phosphorylation sites on numerous viral proteins in cells, showing that they can undergo efficient modification in infected cells. Until now, we can only speculate about the host kinases involved 220 and the functions driven by PTMs, which will be an important topic for follow-up studies. For SARS-CoV-1, it was shown that modification of viral proteins can lead to regulation of RNA binding of the nucleoprotein (Wu et al., 2009 ) and is needed for viral replication. Although similar effects in SARS-CoV-2 are likely, this remains to be studied in this novel virus. A recent paper analyzed the interaction profile of SARS-CoV-2 proteins expressed in HEK293T cells (Gordon et 225 al., 2020) . For the heavily phosphorylated nucleoprotein they could identify interactions the host casein kinases, which might indicate possible modification events by the latter. Also for the ORF9b/protein 9b that we found modified in cells, interaction mapping identified MARK kinases as interaction partners. By exploring the signaling changes inside the host cell, we could gain important insights into host 230 cell signaling during infection. We found essential GFR signaling pathways activated such as EGFR or PDGFR, together with a plethora of RhoGTPase associated signaling molecules. We could furthermore show modulation of the splicing machinery, in line with previous results indicating dependency of viral in vitro pathology on the host spliceosome (Bojkova et al., 2020) . The same is true for metabolic reprogramming, for which we found differential post-translational 235 modification of most members of the carbon metabolic pathways, namely glycolysis, pentose phosphate and TCA cycle. These metabolic pathways were significantly up-regulated on total protein levels in the presented dataset, consistent with our previous study (Bojkova et al., 2020) , suggesting that these key pathways are regulated on multiple levels. A number of drugs to treat COVID-19 have been suggested, largely based on bioinformatics 240 analyses of genetics or cellular data (Gordon et al., 2020; Li et al., 2020; Wang, 2020) . However, for many of these compounds, studies explaining their working mechanisms in the context of SARS-CoV-2 or viral assays to determine their efficacy of blocking viral replication in cell models of SARS-CoV-2 infection, are missing. While monitoring signaling changes in host cells, we observed activation of GFR signaling cascades after infection, consistent with other viruses 245 relying on the receptors themselves or elicited signal transduction (Eierhoff et al., 2010; Kung et al., 2011; Lupberger et al., 2011; Ueki et al., 2013; Wu et al., 2017; Zhu et al., 2009) . From our data we could not clearly conclude which GFR might be activated and thus tested whether GFR downstream signaling inhibition can prevent SARS-CoV-2 replication, as reported for some other viruses (Baturcam et al., 2019; Pleschka et al., 2001) . Previously, temporal kinome analysis 250 identified antiviral potential of RAS/RAF/MEK and PI3K/AKT/ for MERS-CoV (Kindrachuk et al., 2015) . By targeting the RAS/RAF/MEK and PI3K/AKT/mTOR downstream axes of GFR signaling, we found efficient inhibition of viral replication in two different cell lines derived from different tissues (Figure 4) . GFR signaling was shown to play a role in diverse virus infections as well as in fibrosis induction by SARS-CoV-1 (Beerli et al., 2019; Kung et al., 2011; Lupberger et 255 al., 2011; Pleschka et al., 2001; Ueki et al., 2013; . Thus, our results in cytopathic effects might indeed indicate cytoprotective roles for GFR signaling axes during SARS-CoV-2 infection and possible development of fibrosis (Luo et al., 2020) . Notably, some inhibitors used in our study such as omipalisib were shown to suppress fibrosis progression in patients with idiopathic pulmonary fibrosis, which may share deregulation of signaling pathways 260 involved in lung fibrosis of coronavirus patients . These findings suggest that inhibitors of GFR downstream signaling may bring benefit to COVID-19 patients independently of their antiviral activity. Taken together, this study provides new insights into molecular mechanisms elicited by SARS-CoV-2 infection. Proteomic analyses revealed several pathways that are rearranged during 265 infection and showed that targeting of those pathways is a valid strategy to inhibit cytopathic effects triggered by infection. In this study, cancer cell lines were used to assess the effect of SARS-CoV-2 on host cells during infections. We chose the experimental time-point, based on previous analysis of the 270 infection course in these cells (Bojkova et al., 2020) . Notably, the kinetics of infection are likely to be different in other cell lines or primary material, since we also observed a different MOI needed for UKF-RC-2 cells. In addition, we tested the efficiency of the presented drugs only in the context of in vitro cell line experiments. Thus, the results do not represent direct evidence for the use of these therapeutics in patients, as the effects might differ in primary tissue. The results 275 presented indicate potential anti-viral effects that have to be further validated in other models and clinical trials to assess their usefulness for the treatment of COVID-19. (A) Experimental scheme. Caco-2 cells were infected with SARS-CoV-2 for one hour (MOI: 1), washed and incubated for additional 24 hours. Proteins were extracted and prepared for bottomup proteomics. All ten conditions were multiplexed using TMT10 reagents. 250 µg of pooled samples were used for whole cell proteomics (24 Fractions) and the remainder (~1 mg) enriched for phosphopeptides by Fe-NTA. Phosphopeptides were fractionated into 8 fractions and 305 concatenated into 4 fractions. All samples were measured on an Orbitrap Fusion Lumos. (B) Volcano plot showing fold changes of infected versus mock cells for all quantified phosphopeptides. P values were calculated using an unpaired, two-sided student's t-test with equal variance assumed and adjusted using the Benjamini Hochberg FDR method (N = 5 biological replicates). Orange or blue points indicate significantly increased or decreased 310 phosphopeptides, respectively. (C) Volcano plot showing differences between SARS-CoV-2 and mock infected cells in total protein levels for all quantified proteins. P values were calculated using an unpaired, two-sided student's t-test with equal variance assumed and adjusted using the Benjamini Hochberg FDR method (N = 5). Orange or blue points indicate significantly increased or decreased 315 phosphopeptides, respectively. (D) Distribution of phosphorylation sites identified across modified amino acids. See also Figure S1 and Tables S1 and S2. See also Table S1 , S2, S3 and Figure S1 . (E) Scatter plot showing correlation between fold changes of phosphopeptides compared to fold changes of total proteins levels. Two subsets of phosphopeptides were detected: one was 340 J o u r n a l P r e -p r o o f 9 mainly regulated by differential modification (indicated in yellow), the second by changes in protein abundance. (F) STRING network analysis of proteins decreased in total protein levels ( Figure 1C) . Inserts indicate pathways found in the network. See also Table S1 , S2, S4, S7, Figure S2 and S3. 345 See also Figure S4 . See also Figure S5 . Resource Availability Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact: Christian Münch: ch.muench@em.uni-frankfurt.de (C.M.) 400 UKF-RC-2 human kidney cells derived from renal carcinoma are available upon request. The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE (Perez-Riverol et al., 2019) Human Caco-2 cells, derived from colon carcinoma, was obtained from the Deutsche Sammlung von Mikroorganismen und Zellkulturen (DSMZ; Braunschweig, Germany). Cells were grown at 37°C in Minimal Essential Medium (MEM) supplemented with 10% fetal bovine serum (FBS) and containing 100 IU/ml penicillin and 100 µg/ml streptomycin. All culture reagents were purchased from Sigma. 415 Cell line designated UKF-RC-2 was established from a tumor sample of a patient with a diagnosis of renal carcinoma hospitalized at Department of Urology, University Hospital Frankfurt. Tumor tissue was cut in pieces and dissociated using 0.2% trypsin solution. Primary tumor cells and passaging of cell line was performed using IMDM medium supplemented with 10% FCS and antibiotics. UKF-RC-2 cells between passages 15 and 20 were used for antiviral 420 experiments. All culture reagents were purchased from Sigma. SARS-CoV-2 was isolated from samples of travelers returning from Wuhan (China) to Frankfurt (Germany) using human colon carcinoma cell line CaCo-2 as described previously 12 . SARS-CoV-2 stocks used in the experiments had undergone one passage on CaCo-2 cells and were 425 stored at -80° C. Virus titers were determined as TCID50/ml in confluent cells in 96-well microtiter plates. Confluent layers of CaCo-2 cells in 96-well plates were infected with SARS-CoV-2 at MOI 0.01. Virus was added together with drugs and incubated in MEM supplemented with 2% FBS with different drug dilutions. Cytopathogenic e ect (CPE) was assessed visually 48 h after infection. To assess effects of drugs on Caco-2 cell viability, confluent cell layers were treated with different drug concentration in 96-well plates. The viability was measured using the Rotitest Vital (Roth) according to manufacturer's instructions. Data for each condition was collected for at 435 least three biological replicates. For dose response curves, data was fitted with all replicates using OriginPro 2020 with the following equation: IC50 values were generated by OriginPro 2020 together with metrics for curve fits. For UKF-RC-2 cells, the assay was performed as described above, except for usage of a MOI of 440 0.1 as staining experiments for SARS-CoV-2 infection in UKF-RC-2 cells revealed the need of a higher MOI to achieve comparable effects to CaCo-2 cells. SARS-CoV-2 RNA from cell culture supernatant samples was isolated using AVL buffer and the QIAamp Viral RNA Kit (Qiagen) according to the manufacturer's instructions. Absorbance-based 445 quantification of the RNA yield was performed using the Genesys 10S UV-Vis Spectrophotometer (Thermo Scientific). RNA was subjected to OneStep qRT-PCR analysis using the Luna Universal One-Step RT-qPCR Kit (New England Biolabs) and a CFX96 Real-Time System, C1000 Touch Thermal Cycler. Primers were adapted from the WHO protocol (Corman et al., 2020) targeting the open reading frame for RNA-dependent RNA polymerase 450 (RdRp): RdRP_SARSr-F2 (GTG ARA TGG TCA TGT GTG GCG G) and RdRP_SARSr-R1 (CAR ATG TTA AAS ACA CTA TTA GCA TA) using 0.4 µM per reaction. Standard curves were created using plasmid DNA (pEX-A128-RdRP) harboring the corresponding amplicon regions for RdRP target sequence according to GenBank Accession number NC_045512. All quantification experiments have been carried out with biological replicates. 455 Effect of selected compounds on viral replication was assessed by staining of double-stranded RNA, which has been shown to be sufficient for measurement of SARS-CoV-1 replication (Weber et al., 2006) . Briefly, cells were fixed with acetone/methanol (40:60) solution 48 h post infection. Immunostaining was performed using a monoclonal antibody directed against dsRNA (1:150 460 dilution, SCICONS J2, mouse, IgG2a, kappa chain, English & Scientific Consulting Kft., Szirák, Hungary), which was detected with biotin-conjugated secondary antibody (1:1000 dilution, Jackson ImmunoResearch) followed by application streptavidin, peroxidase conjugate (1:3000 dilution, Sigma Aldrich). Lastly, the dsRNA positive cells were visualized by addition of AEC substrate. Wells were imaged at different areas to visualize a larger area (presented in 465 Supplementary figures). For all proteomics analysis, Caco-2 cells were infected at an MOI of 1 and the sample preparation was performed as described previously . Briefly, lysates were precipitated by methanol/chloroform and proteins resuspended in 8 M Urea/10 mM EPPS pH 470 8.2. Concentration of proteins was determined by Bradford assay and 300 µg of protein per samples was used for digestion. For digestion, the samples were diluted to 1 M Urea with 10mM EPPS pH 8.2 and incubated overnight with 1:50 LysC (Wako Chemicals) and 1:100 Sequencing grade trypsin (Promega). Digests were acidified using TFA and tryptic peptideswere purified by tC18 SepPak (50 mg, Waters). 125 µg peptides per sample were TMT labelled and the mixing 475 was normalized after a single injection measurement by LC-MS/MS to equimolar ratios for each channel. 250 µg of pooled peptides were dried for offline High pH Reverse phase fractionation by HPLC (whole cell proteome) and remaining 1 mg of multiplexed peptides were used for phospho-peptide enrichment by High-Select Fe-NTA Phosphopeptide enrichment kit (Thermo Fisher) after manufacturer`s instructions. After enrichment, peptides were dried and 480 resuspended in 70% acetonitrile/0.1% TFA and filtered through a C8 stage tip to remove contaminating Fe-NTA particles. Dried phospho-peptides then were fractionated on C18 (Empore) stage-tip. For fractionation C18 stagetips were washed with 100% acetonitrile twice, followed by equilibration with 0.1% TFA solution. Peptides were loaded in 0.1% TFA solution and washed with water. Elution was performed stepwise with different acetonitrile concentrations in 485 0.1% Triethylamine solution (5%, 7.5%, 10%, 12.5%, 15%, 17.5%, 20%, 50%). The eight fractions were concatenated into four fractions and dried for LC-MS. Peptides were fractionated using a Dionex Ultimate 3000 analytical HPLC. 250 µg of pooled and purified TMT-labeled samples were resuspended in 10 mM ammonium-bicarbonate (ABC), 5% 490 ACN, and separated on a 250 mm long C18 column (X-Bridge, 4.6 mm ID, 3.5 µm particle size; Waters) using a multistep gradient from 100% Solvent A (5% ACN, 10 mM ABC in water) to 60% Solvent B (90% ACN, 10 mM ABC in water) over 70 min. Eluting peptides were collected every 45 s into a total of 96 fractions, which were cross-concatenated into 24 fractions and dried for further processing. 495 All mass spectrometry data was acquired in centroid mode on an Orbitrap Fusion Lumos mass spectrometer hyphenated to an easy-nLC 1200 nano HPLC system using a nanoFlex ion source (ThermoFisher Scientific) applying a spray voltage of 2.6 kV with the transfer tube heated to 300°C and a funnel RF of 30%. Internal mass calibration was enabled (lock mass 445.12003 500 m/z). Peptides were separated on a self-made, 32 cm long, 75µm ID fused-silica column, packed in house with 1.9 µm C18 particles (ReproSil-Pur, Dr. Maisch) and heated to 50°C using an integrated column oven (Sonation). HPLC solvents consisted of 0.1% Formic acid in water (Buffer A) and 0.1% Formic acid, 80% acetonitrile in water (Buffer B). For total proteome analysis, a synchronous precursor selection (SPS) multi-notch MS3 method 505 was used in order to minimize ratio compression as previously described (McAlister et al., 2014) . Individual peptide fractions were eluted by a non-linear gradient from 7 to 40% B over 90 minutes followed by a step-wise increase to 95% B in 6 minutes which was held for another 9 minutes. Full scan MS spectra (350-1400 m/z) were acquired with a resolution of 120,000 at m/z 200, maximum injection time of 100 ms and AGC target value of 4 x 10 5 . The 20 most intense 510 precursors with a charge state between 2 and 6 per full scan were selected for fragmentation ("Top 20") and isolated with a quadrupole isolation window of 0.7 Th. MS2 scans were performed in the Ion trap (Turbo) using a maximum injection time of 50ms, AGC target value of 1.5 x 10 4 and fragmented using CID with a normalized collision energy (NCE) of 35%. SPS-MS3 scans for quantification were performed on the 10 most intense MS2 fragment ions with an 515 isolation window of 0.7 Th (MS) and 2 m/z (MS2). Ions were fragmented using HCD with an NCE of 65% and analyzed in the Orbitrap with a resolution of 50,000 at m/z 200, scan range of 110-500 m/z, AGC target value of 1.5 x10 5 and a maximum injection time of 120ms. Repeated sequencing of already acquired precursors was limited by setting a dynamic exclusion of 45 seconds and 7 ppm and advanced peak determination was deactivated. 520 For phosphopeptide analysis, each peptide fraction was eluted by a linear gradient from 5 to 32% B over 120 minutes followed by a step-wise increase to 95% B in 8 minutes which was held for another 7 minutes. Full scan MS spectra (350-1400 m/z) were acquired with a resolution of 120,000 at m/z 200, maximum injection time of 100 ms and AGC target value of 4 x 10 5 . The 20 most intense precursors per full scan with a charge state between 2 and 5 were selected for 525 fragmentation ("Top 20"), isolated with a quadrupole isolation window of 0.7 Th and fragmented via HCD applying an NCE of 38%. MS2 scans were performed in the Orbitrap using a resolution of 50,000 at m/z 200, maximum injection time of 86ms and AGC target value of 1 x 10 5 . Repeated sequencing of already acquired precursors was limited by setting a dynamic exclusion of 60 seconds and 7 ppm and advanced peak determination was deactivated. An MS 2 based 530 method was chosen, because of higher precision and identification rates (Hogrebe et al., 2018) . Phospho-peptide fractions intrinsically exhibit lower complexity, rendering them less prone to ratio compression by isolation interference. Raw files were analyzed using Proteome Discoverer (PD) 2.4 software (ThermoFisher 535 Scientific). Spectra were selected using default settings and database searches performed using SequestHT node in PD. Database searches were performed against trypsin digested Homo Sapiens SwissProt database, SARS-CoV-2 database (Uniprot pre-release). Static modifications were set as TMT6 at the N-terminus and lysines and carbamidomethyl at cysteine residues. Search was performed using Sequest HT taking the following dynamic modifications into 540 account: Oxidation (M), Phospho (S, T, Y), Met-loss (Protein N-terminus), Acetyl (Protein Nterminus) and Met-loss acetyl (Protein N-terminus). For whole cell proteomics, the same settings were used except phosphorylation was not allowed as dynamic modification. For phosphoproteomics all peptide groups were normalized by summed intensity normalization and then analyzed on peptide level. For whole cell proteomics normalized PSMs were summed for each 545 accession and data exported for further use. Peptide and protein identifications were validated using a concatenated target-decoy strategy and FDR was estimated using q-values calculated by Percolator applying 1% and 5% cut-offs for high and medium confidence hits, while only high confident proteins are reported. Phosphosite localization probabilities were calculated using the ptmRS-node working in "PhosphoRS mode" and using default settings. 550 Unless otherwise stated significance was tested by unpaired, two-sided students t-tests with equal variance assumed. Resulting P values were corrected using the Benjamini-Hochberg FDR 555 procedure. Adjusted P values smaller/equal 0.05 were considered significant. For phosphoproteomics an additional fold change cutoff was applied (log2 > |1|), while for total protein levels, due to different dynamic range, a fold change cutoff of log 2 > |0.5| was applied. Kinase motifs of phosphopeptides from SARS-CoV-2 proteins were predicted using NetPhos 3.1 560 (Blom et al., 1999) and GPS 5.0 (stand-alone version) using the fasta-file of the Uniprot prerelease which was also used for the proteomics data analysis. (Blom et al., 2004; Wang et al., 2020) . For NetPhos, only Kinases with a score above 0.5 were considered as positive hits. For GPS 5.0, sequences were submitted separately for S/T-and Y-kinases and the score threshold was set to "high". For the final list in Supplementary Table 3 , only the top hits with the highest 565 scores were considered. Z-scores were calculated for each phospho-site and the total protein levels individually. Phosphosites were collapsed by average. For merging phosphorylation and total protein levels Z-scores for collapsed phosphorylation and protein level were added for each condition and 570 replicate. Thus, both negative Z-scores (downregulation) will produce a lower combined Z-score and vice versa two positive Z-scores will produce a larger combined Z-score. Next, Euclidean distance correlation for all possible protein-protein pairs were calculated, taking all conditions and replicates individually into account. A heatmap was then build by Euclidean distance hierarchical clustering of the correlation matrix. 575 Pathway enrichment analysis was performed by ReactomeFI cytoscape plugin or by STRING functional enrichment analysis. Both analysis used Reactome database for pathway annotations. All proteins were loaded into ReactomeFI cytoscape plugin to visualize protein-protein functional 580 interaction network. Next, drugs were overlaid by ReactomeFI and network was filtered for the drugs and the first interacting partners. Layout was calculated by yFilesLayout algorithm. All proteins showing significant regulation were loaded by OmicsVisualizer cytoscape plugin and STRING interaction network was retrieved with a confidence cutoff of 0.9. For EGFR 585 subnetwork, EGFR was selected with first interacting neighbors. Highlights: • Phosphoproteomics of SARS-CoV-2 infected cells reveal the signaling landscape • SARS-CoV-2 proteins are extensively phosphorylated in host cells • Infection leads to activation of growth factor receptor signaling • Drugs inhibiting growth factor receptor signaling prevent viral replication In this study, Klann et al. dissected the host cell signaling landscape upon infection with SARS-CoV-2. Mapping differential signaling networks identified a number of pathways activated during infection. Drug-target network analysis revealed potential therapeutic targets. Growth factor receptor signaling was highly activated upon infection and its inhibition prevented SARS-CoV-2 replication in cells. J o u r n a l P r e -p r o o f Growth factor receptor Growth factor receptor Growth factor receptor MEK inhibition drives anti-viral defence in RV but not RSV challenged human airway epithelial cells through AKT/p70S6K/4E-BP1 signalling Vaccinia virus hijacks EGFR signalling to enhance virus spread 620 through rapid and directed infected cell motility Sequence and structure-based prediction of eukaryotic protein phosphorylation sites1 1Edited by Prediction of post-translational glycosylation and phosphorylation of proteins from the amino acid sequence Proteomics of SARS-CoV-2-infected host cells reveals therapy targets Epidemiological and clinical characteristics of 99 cases of 2019 novel coronavirus 630 pneumonia in Wuhan, China: a descriptive study Detection of 2019 novel coronavirus (2019-nCoV) by real-time RT-PCR The epidermal 635 growth factor receptor (EGFR) promotes uptake of influenza A viruses (IAV) into host cells Phase I and Pharmacokinetic Study of the Oral Farnesyl Transferase Inhibitor SCH 66336 Given Twice Daily 640 to Patients With Advanced Solid Tumors Measurement of sorafenib plasma concentration by highperformance liquid chromatography in patients with advanced hepatocellular carcinoma: is it useful the application in clinical practice? A pilot study A SARS-CoV-2-Human Protein-Protein Interaction Map Reveals Drug Targets and Potential Drug-Repurposing Phosphorylation regulates the activity of the SMN complex during assembly of spliceosomal U snRNPs Plaque assay for human coronavirus NL63 using human colon carcinoma cells Benchmarking common quantification strategies for large-scale phosphoproteomics PKR activation and eIF2α phosphorylation mediate human globin mRNA splicing at spliceosome assembly Omipalisib (GSK458), a Novel Pan-PI3K/mTOR Inhibitor, Exhibits In Vitro Anti-Lymphoma Activity in Chemotherapy-Sensitive and -Resistant Models of Burkitt Lymphoma Antiviral Potential of ERK/MAPK and PI3K/AKT/mTOR Signaling Modulation for Middle East Respiratory Syndrome Coronavirus Infection as Identified by Temporal Kinome Analysis Functional Translatome Proteomics Reveal 670 Converging and Dose-Dependent Regulation by mTORC1 and eIF2α Epstein-Barr Virus LMP1 Activates EGFR, STAT3, and ERK through Effects on PKCδ Network Bioinformatics Analysis Provides Insight into Drug Repurposing for Clinical Pathology of Critical Patient with Novel Coronavirus Pneumonia EGFR and EphA2 are host factors for hepatitis C virus entry 680 and possible targets for antiviral therapy First-in-Human, Phase I Dose-Escalation Study of the Safety, Pharmacokinetics, and Pharmacodynamics of RO5126766, a First-in-Class Dual MEK/RAF Inhibitor in Patients with Solid Tumors Phosphorylation of human PRP28 by SRPK2 is required for integration of the U4/U6-U5 tri-snRNP into the spliceosome MultiNotch MS3 enables accurate, sensitive, and 690 multiplexed detection of differential expression across cancer cell line proteomes Regulation of mammalian spliceosome assembly by a protein phosphorylation mechanism First-in-Human Phase I Study of 695 GSK2126458, an Oral Pan-Class I Phosphatidylinositol-3-Kinase Inhibitor, in Patients with Advanced Solid Tumor Malignancies The PRIDE database and related tools and resources in 2019: improving support for quantification data Influenza virus propagation is impaired by inhibition of the Raf/MEK/ERK signalling cascade Analysis of ACE2 in polarized epithelial cells: surface expression and function as receptor for severe acute respiratory syndrome-associated coronavirus First-in-Human Phase I Study of Pictilisib (GDC-0941), a 710 Potent Pan-Class I Phosphatidylinositol-3-Kinase (PI3K) Inhibitor, in Patients with Advanced Solid Tumors Phase II Randomized Preoperative Window-of-Opportunity Study of the PI3K Inhibitor Pictilisib Plus Anastrozole Compared With Anastrozole Alone in 715 Patients With Estrogen Receptor-Positive Breast Cancer Repurposing Therapeutics for COVID-19: Supercomputer-Based Docking to the SARS-CoV-2 Viral Spike Protein and Viral Spike Protein-Human ACE2 Interface Severe acute 720 respiratory syndrome coronavirus protein 6 accelerates murine coronavirus infections Respiratory virus-induced EGFR activation suppresses IRF1-dependent interferon λ and antiviral defense in airway epithelium The role of epidermal growth factor receptor (EGFR) signaling in SARS coronavirus-induced pulmonary fibrosis Overactive EGFR Signaling Leads to Increased Fibrosis After SARS-CoV Infection Fast Identification of Possible Drug Treatment of Coronavirus Disease -19 730 (COVID-19) Through Computational Drug Repurposing Study 0: An Update on the Prediction of Kinase-specific Phosphorylation Sites in Proteins Double-735 stranded RNA is produced by positive-strand RNA viruses and DNA viruses but not in detectable amounts by negative-strand RNA viruses Discovery and development of sorafenib: a multikinase inhibitor for treating cancer Glycogen synthase kinase-3 regulates the phosphorylation of severe acute respiratory syndrome coronavirus nucleocapsid protein and viral replication Human cytomegalovirus glycoprotein complex gH/gL/gO uses PDGFR-α as a key for entry The EGFR family and its ligands in human cancer: signalling mechanisms and therapeutic opportunities Airway Mucin Production Involves a Novel TLR3-EGFR-Dependent Pathway We thank the Quantitative Proteomics Unit (IBC2, Goethe University Frankfurt) for support and expertise on LC-MS instrumentation and data analysis and Christiane Pallas and Lena 280Stegmann for support of with experimental work.