key: cord-0909789-efctb6hg
authors: Grossegesse, Marica; Bourquain, Daniel; Neumann, Markus; Schaade, Lars; Schulze, Jessica; Mache, Christin; Wolff, Thorsten; Nitsche, Andreas; Doellinger, Joerg
title: Deep Time Course Proteomics of SARS-CoV- and SARS-CoV-2-Infected Human Lung Epithelial Cells (Calu-3) Reveals Strong Induction of Interferon-Stimulated Gene Expression by SARS-CoV-2 in Contrast to SARS-CoV
date: 2022-01-04
journal: J Proteome Res
DOI: 10.1021/acs.jproteome.1c00783
sha: e203e2b61ff027c7fb3439ae776239156cd0c496
doc_id: 909789
cord_uid: efctb6hg

[Image: see text] Severe acute respiratory syndrome (SARS)-CoV and SARS-CoV-2 infections are characterized by remarkable differences, including infectivity and case fatality rate. The underlying mechanisms are not well understood, illustrating major knowledge gaps of coronavirus biology. In this study, protein expression of the SARS-CoV- and SARS-CoV-2-infected human lung epithelial cell line Calu-3 was analyzed using data-independent acquisition–mass spectrometry. This resulted in a comprehensive map of infection-related proteome-wide expression changes in human cells covering the quantification of 7478 proteins across four time points. Most notably, the activation of interferon type-I response was observed, which is surprisingly absent in several proteome studies. The data reveal that SARS-CoV-2 triggers interferon-stimulated gene expression much stronger than SARS-CoV, which reflects the already described differences in interferon sensitivity. Potentially, this may be caused by the enhanced abundance of the viral M protein of SARS-CoV in comparison to SARS-CoV-2, which is a known inhibitor of type I interferon expression. This study expands the knowledge on the host response to SARS-CoV-2 infections on a global scale using an infection model, which seems to be well suited to analyze the innate immunity.

In late 2019, the first cases of severe pneumonia of unknown origin were reported in Wuhan, China. Shortly afterward, a new coronavirus was discovered as the causative agent and named severe acute respiratory syndrome (SARS)-CoV-2 and the related disease COVID-19. The virus turned out to be highly infectious and caused a world-wide pandemic, which is still ongoing and has led to the death of >4,500,000 humans worldwide by September 2021. Already in 2002, another coronavirus, SARS-CoV, was discovered in China which caused an outbreak with about 780 deaths. 1 However, at this time, the outbreak could be controlled probably due to the lower infectivity of SARS-CoV compared to SARS-CoV-2. 2 SARS-CoV and SARS-CoV-2 share about 80% of their genome sequence and protein homology ranges between 40 and 94%. 3, 4 Although both viruses mainly lead to respiratory tract infections and can cause severe pneumonia, they are characterized by remarkable differences, including infectivity and case fatality rate. 5 As the respiratory tract is the first and main target of SARS-CoV and SARS-CoV-2 infections, it seems conclusive to use airway epithelia cells to study the differences between these two viruses. However, no comparative proteomics study has been published using Calu-3 cells, which is the only permissive lung cell line available for SARS-CoV and SARS-CoV-2. 6 Other human lung cell lines, such as A549, are only susceptible to SARS-CoV-2 infection upon overexpression of the SARS-CoV receptor ACE2, 6 which was recently found to be an interferonstimulated gene (ISG). 7 In the present study, we used dataindependent acquisition−mass spectrometry (DIA−MS) to analyze the protein expression in Calu-3 cells infected with SARS-CoV and SARS-CoV-2 over the time course of 24 h. In total, 8391 proteins were identified, 7478 of which could be reliably quantified across the experiment. This results in a deep and comprehensive proteome map, which reflects timedependent protein expression changes during SARS-CoV and SARS-CoV-2 infections and provides deep insights into the virus-specific immunomodulation of human lung cells.

Calu-3 cells (ATCC HTB-55) were cultivated in Eagle's minimum essential medium containing 10% fetal calf serum (FCS), 2 mM L-glutamine, and non-essential amino acids. A total of 5 × 10 5 cells per well were seeded in six-well plates and incubated overnight at 37°C and 5% CO 2 in a humified atmosphere. The medium was removed, and the cells were infected with SARS-CoV (strain Hong Kong) or SARS-CoV-2 (hCoV-19/Italy/INMI1-isl/2020 (National Institute for Infectious Diseases, Rome, Italy, GISAID Accession EPI_-ISL_410545) at a multiplicity of infection (MOI) of 5. Mock samples were treated with medium only. One hour post infection (p.i.), the cells were washed with phosphate-buffered saline (PBS), and fresh medium was added. After 2, 6, 8, 10, and 24 h p.i., the medium was removed and stored at −80°C. The cells were washed with PBS and prepared for proteomics as described below. For each time point and virus, triplicate samples were taken. Additionally, triplicate mock samples per time point were taken.

Calu-3 cells were propagated in DMEM containing 15% FCS, 2 mM L-glutamine, 100 U/ml penicillin, 100 μg/ml streptomycin, 1× non-essential amino acids, and 1 mM sodium pyruvate and incubated at 37°C with 5% CO 2 in a humidified atmosphere. For differentiation, the cells were seeded in ThinCert tissue culture inserts (0.4 μm pore size) and were cultivated under an air−liquid-Interface (ALI) for 14 days prior to infection.

The cells were infected with SARS-CoV-2 [hCoV-19/ Germany/BY-ChVir-929/2020, lineage B.1.153 (GISAID accession: EPI_ISL_406862)] or SARS-CoV (Frankfurt-1) at an MOI of 0.1 in D-PBS containing 0.3% BSA for 1 h at 37°C. Afterward, the cells were washed apically with D-PBS, and fresh medium was added to the apical and basolateral chambers. To quantify the infectious virus particles, apical and basolateral supernatants were harvested at indicated time points and titrated on Vero E6 cells by the standard plaque titration assay.

The amount of SARS-CoV and SARS-CoV-2 RNA in the supernatant was analyzed by qPCR at 2, 6, 8, 10, and 24 h p.i. Supernatants were extracted using the QIAamp Viral RNA Mini Kit (Qiagen, Hilden, Germany) according to manufacturer's recommendations and eluted in 60 μL of RNase-free water. Real-time polymerase chain reaction (PCR) targeting the viral E gene was carried out as described by Michel et al. 8 using the primers and probe used in the study published by Corman et al. 9 Quantification of viral genome equivalents was done using the SARS-CoV-2 E gene WHO reference PCR standard.

Supernatants of infected polarized Calu-3 cells were analyzed using the R&D DuoSet ELISA Kits for human IFNα (DY9345-05), IFNβ (DY814-05), IFNλ (DY1598B), IP-10 (DY266), and IL-6 (DY206) according to the manufacturer's instructions.

Calu-3 cell monolayers were washed once with prechilled PBS and lysed in 100 μL of ice cold lysis buffer (10 mM Tris/HCl, pH 7.5, 150 mM NaCl, 0.5 mM EDTA, and 1% NP40). After at least 30 min of incubation on ice, cell lysates were centrifuged at 15,000g and 4°C for 10 min. The supernatants were supplemented with 20 μL of 6× Laemmli sample buffer containing 5% β-mercaptoethanol and boiled for 10 min. Afterward, protein samples were separated using reducing sodium dodecyl sulfate−polyacrylamide gel electrophoresis under denaturing conditions and transferred onto nitrocellulose membranes. The membranes were incubated with antibodies detecting the SARS-CoV-2 Spike protein (Genetex, GTX632604), human Stat 1 (Santa Cruz Biotechnology, Inc., sc-464), human IFIT 2 (Abcam, ab113112), or human GAPDH (Cell Signaling Technology, Inc., #2118), respectively. Incubation with a suitable secondary horseradish-peroxidaseconjugated antibody (Agilent Technologies Inc., P0260/ P0217) allows development on X-ray films using the Super-Signal TM West Dura Extended Duration Substrate (Thermo Fischer Scientific).

ACE2-A549-Dual cells were seeded into 96-well plates at 4 × 10 4 cells per well and incubated overnight at 37°C and 5% CO 2 in a humified atmosphere. The cells were infected with either SARS-CoV or SARS-CoV-2 at an MOI of 1.0. At 2 days p.i., the interferon regulatory factor (IRF)-activity was assayed using the QUANTI-Luc luminescence reagent (InvivoGen, San Diego, CA, USA) and an INFINITE 200 PRO microplate reader (Tecan, Mannedorf, Switzerland).

Samples were prepared for proteomics using sample preparation by easy extraction and digestion. 10 At first, the medium was removed and the cells were washed using PBS. Afterward, 200 μL of trifluoroacetic acid (TFA) (Thermo Fisher Scientific, Waltham, MA, USA) was added, and the cells were incubated at room temperature for 3 min. The samples were neutralized by transferring TFA to prepared reaction tubes containing 1.4 mL of 2 M TrisBase. After adding tris(2-carboxyethyl)phosphine to a final concentration of 10 mM and 2-chloroacetamide to a final concentration of 40 mM, the samples were incubated at 95°C for 5 min. 200 μL of the resulting solutions was diluted 1:5 with water and subsequently digested for 20 h at 37°C using 1 μg of Trypsin Gold, MS Grade (Promega, Fitchburg, WI, USA). The resulting peptides were desalted using 200 μL StageTips packed with three Empore SPE Disks C18 (3 M Purification Inc., Lexington, USA) and concentrated using a vacuum concentrator. 11, 12 Dried peptides were suspended in 20 μL of 0.1% TFA and quantified by measuring the absorbance at 280 nm using an Implen NP80 spectrophotometer (Implen, Munich, Germany). for 6 min and were subsequently separated on a 200 cm μPAC column (PharmaFluidics) using a stepped 160 min gradient of 80% acetonitrile (solvent B) in 0.1% formic acid (solvent A) at a 300 nL/min flow rate: 3−10% B for 22 min, 10−33% B for 95 min, 33−49% B for 23 min, 49−80% B for 10 min, and 80% B for 10 min. Column temperature was kept at 50°C using a butterfly heater (Phoenix S&T, Chester, PA, USA). The Q Exactive HF was operated in a data-independent (DIA) manner in the m/z range of 350−1150. Full scan spectra were recorded with a resolution of 120,000 using an automatic gain control (AGC) target value of 3 × 10 6 with a maximum injection time of 100 ms. The full scans were followed by 84 DIA scans of dynamic window widths using an overlap of 0.5 Th (Supporting Information Table S1 ). For the correction of the predicted peptide spectral libraries, a pooled sample was measured using gas-phase separation (8 × 100 Th) with 25 × 4 Th windows in each fraction using a shift of 2 Th for subsequent cycles. Window placement was optimized using Skyline (Version 4.2.0). 12 DIA spectra were recorded at a resolution of 30,000 using an AGC target value of 3 × 10 6 with a maximum injection time of 55 ms and a first fixed mass of 200 Th. The normalized collision energy was set to 25%, and the default charge state was set to 3. Peptides were ionized using electrospray with a stainless-steel emitter, I.D. 30 μm (Proxeon, Odense, Denmark), at a spray voltage of 2.0 kV, and a heated capillary temperature of 275°C.

Protein sequences of Homo sapiens (UP000005640, 95,915 sequences, downloaded 23/5/19), SARS-CoV (UP000000354, 15 sequences, downloaded 21/9/20), and SARS-CoV-2 (UP000464024, 14 sequences, downloaded 21/9/20) were obtained from UniProt. 13 A combined spectral library was predicted for all possible peptides with strict trypsin specificity (KR not P) in the m/z range of 350−1150, with charge states of 2−4 and allowing up to one missed cleavage site using Prosit. 14 Input files for library prediction were generated using EncyclopeDIA (Version 0.9.5). 15 The in silico library was corrected using the data of the gas-phase fractionated pooled sample in DIA-NN (Version 1.7.10). 16 Mass tolerances were set to 10 ppm for MS1 and 20 ppm for MS 2 spectra, and the "unrelated run" option was enabled with the false discovery rate (FDR) being set to 0.01. The single-run data were analyzed using the corrected library with fixed mass tolerances of 10 ppm for MS1 and 20 ppm for MS 2 spectra with enabled "RT profiling" using the "robust LC (high accuracy)" quantification strategy. The FDR was set to 0.01 for precursor identifications, and proteins were grouped according to their respective genes. The resulting identification file was filtered using R (Version 3.6) in order to keep only proteotypic peptides and proteins with protein q-values < 0.01. Visualization and further analysis were done in Perseus (Version 1.6.5). 17 Relative protein quantification was done based on log (2)-transformed and Z-score normalized "MaxLFQ" intensities. Proteins which were not Journal of Proteome Research pubs.acs.org/jpr Article and the replication between SARS-CoV and SARS-CoV-2 was comparable over the period of the experiment ( Figure 1A ). This is consistent with the expression of viral proteins, which was detectable from 6 h p.i. as well. The majority of viral proteins including nucleoprotein, spike glycoprotein, ORF3a, and

ORF9a are not differentially expressed between SARS-CoVand SARS-CoV-2-infected cells. An exception is the membrane protein (M) whose abundance is enhanced in SARS-CoVinfected cells compared to SARS-CoV-2-infected ones ( Figure  1B ,D, Supporting Information Tables S2 and S10). The expression of 2642 human proteins differed significantly between the sample groups (ANOVA, FDR = 0.05), which was reduced to 261 proteins using a post-hoc test (FDR = 0.05) when only proteins with at least one significant pairwise difference in an infected cell with its time-matched mock control were kept (Supporting Information Table S3 ). This large (Figure 2 and Supporting Information Tables S4−S6). Out of the five clusters, two clusters (up-regulated 2 h p.i. and down-regulated 6 h p.i.) revealed no significantly enriched GO terms and among others contained several proteins related to immune response such as OAS1, INAVA, and NFΚBIB. Another cluster consisting of proteins with virus-specific time-course-dependent upregulation was found to be related to mitochondrial translation (adjusted pvalue: 2.5 * 10 −4 , MRPL17, MRPL27, MRPL47, MRPL50, and MRPS7). The other two main clusters included upregulated proteins 24 h p.i. and are related to either the regulation of complement activation (adjusted p-value: 7.9 * 10 −3 , C3 and C5) or interferon alpha/beta signaling (adjusted p-value: 7.8 * 10 −20 , e.g., MX1, MX2, DDX58, STAT1, OAS2, OAS3, and IFIT3). Strikingly, the main difference between SARS-CoV-and SARS-CoV-2-infected cells was observed for proteins derived from ISGs, whose expression is enhanced in SARS-CoV-2infected cells in comparison to SARS-CoV infection.

The induction of the type I IFN response was further validated in differentiated Calu-3 cell ALI cultures, which mimic the human respiratory tract more closely in comparison to nonpolarized cell systems. SARS-CoV and SARS-CoV-2 replicated in differentiated Calu-3 cells to similar high viral titers of about 10 8 plaque-forming units per milliliter peaking at 72 h p.i ( Figure  3A ). IFN β and -λ and the IFN-induced chemokine IP-10 were detected at late time points specifically in supernatants of cells following infection with SARS-CoV-2 but not SARS-CoV ( Figure 3B ), which is in line with the selective upregulation of ISGs following SARS-CoV-2 infection observed in the proteome analysis ( Figure 2 ). In contrast to IFNs, pro-inflammatory cytokines, as exemplified by IL-6, were induced by both viruses. RNA viruses such as SARS-CoV-2 are recognized by the innate immune system via the cytoplasmic double-stranded RNA (dsRNA) sensors RIG-I and MDA5, which signal via the adaptor protein MAVS to induce the expression of IFNs and proinflammatory cytokines. 21 Activation of MAVS leads to the recruitment and activation of the downstream kinases TBK1/ IKKε, which in turn regulate the expression of IFNs via phosphorylation and activation of IRF 3. Interestingly, treatment of SARS-CoV-2-infected cells with BX-795, an inhibitor of IKKε and TBK1, strongly reduced the expression of IFNβ and IFNλ and of the ISGs IP-10, IFIT 2, and STAT 1 ( Figure 3C ), indicating an IKKε-or TBK1-dependency of the IFN response induced following SARS-CoV-2 infection. At the same time, induction of IL-6 remained unchanged. IFN-induction by SARS-CoV-2 was further analyzed in ACE2-A549 reporter cells, confirming a higher IRF activity in SARS-CoV-2-infected cells compared to no detectable IRF activity upon infection with SARS-CoV (Supporting Information Figure S1 ).

As the type I IFN response is one of the most important responses of the innate immune system to RNA viruses, we compared the expression data of related proteins from this study with other proteome studies of SARS-CoV-2-infected human cells. The selection was based on data availability (Supporting Information Table S11 ). For this purpose, all identified proteins annotated with the GO term "type I interferon signaling pathway" (GO:0060337) were extracted from the data of 

Innate immunity is the host's first line of defence to fight infections. One of the most important mechanisms to combat replication of RNA viruses is the interferon response. It is based on the recognition of pathogen-associated molecular patterns, especially dsRNA, which in the end results in the secretion of type I IFNs, which in turn induce the expression of ISGs including multiple antiviral proteins. 27 SARS-CoV-2 is more susceptible to both IFN-α and IFN-β treatment in cultured cells than SARS-CoV, 28−31 which is why type I IFNs could be a possible treatment for COVID-19. 32 Data from SARS-CoV-2infected patients report low or absent levels of IFN-I in serum but induction of ISG expression. 3, 33 In this study, protein expression of SARS-CoV-and SARS-CoV-2-infected Calu-3 cells was analyzed, which should be well suited to uncover the modulation of the type I IFN response during infection. This analysis resulted in a comprehensive proteome map of SARS-CoV-and SARS-CoV-2-infected Calu-3 cells covering ∼7400 proteins across four time points. Expression of 261 proteins changed during the course of infection, which cluster into five main groups. One of those clusters reveals a strong induction of ISG expression 24 h p.i. in SARS-CoV-2-infected cells. Strikingly, this induction was observed at a much lower level in SARS-CoV-infected cells. Among those ISG proteins is, for example, the interferon-induced GTP-binding protein MX1, which is known for its antiviral activity against a wide range of mainly RNA viruses. MX1 expression is increased in SARS-CoV-2-infected patients and correlates well with viral load. 34 Furthermore, ISG expression is induced in SARS-CoV-2infected patients in general and the increase of ISG expression, including MX1, negatively correlates with disease severity. 33 Surprisingly, these findings are not well reflected in the current literature of large-scale proteome analysis of infected human cells and are completely absent in two studies. 22 Recently, it was proposed that SARS-CoV-2 ORF6 interferes less efficiently with human IFN induction and IFN signaling than SARS-CoV ORF6, which could explain the virus-specific induction of ISG expression and the varying IFN sensitivity. 37 The proteome data from this study point toward an additional mechanism. The abundance of viral proteins was highly similar between SARS-CoV and SARS-CoV-2 except for the M protein whose abundance is enhanced in SARS-CoV. This protein is a component of the viral envelope, but its functions beyond are not well characterized. It is known that the homologous M proteins of MERS and SARS-CoV inhibit type I IFN expression. 38, 39 Overexpression of the M protein from SARS-CoV-2 in human cells inhibits the production of type I and III IFNs induced by dsRNA-sensing via direct interaction with RIG-I (DDX58) and reduces the induction of ISGs after Sendai virus infection and poly (I/C) transfection. 36, 40 Additionally, the M protein of SARS-CoV inhibits the formation of the TRAF3− TANK−TBK1/IKKϵ complex, resulting in the inhibition of IFN transcription. 38 In line with these findings, we were able to show that the induction of type I IFN induced following SARS-CoV-2 infection is dependent on the TBKI pathway as well (Figure 3c ). We therefore hypothesize that the enhanced abundance of the M protein of SARS-CoV reduces the induction of ISG expression in infected cells in comparison to SARS-CoV-2 and thereby contributes to the varying IFN sensitivity of both viruses. However, it should be noted that also sequence differences in the M protein of both viruses (amino acid identity = 90.5%) could lead to differences in the IFN-antagonizing capacity, which is not known so far.

In summary, this study presents the so far most comprehensive comparative quantitative proteomics data set of SARS-CoV-and SARS-CoV-2-infected Calu-3 cells, which are the only permissive human lung cell line for SARS-CoV-2. 6 By showing a diverse regulation of ISG expression upon infection, we conclude that Calu-3 cells present a good model system for studying differences in IFN sensitivity of SARS-CoV and SARS-CoV-2. 

The continuing 2019-nCoV epidemic threat of novel coronaviruses to global health -The latest 2019 novel coronavirus outbreak in Wuhan, China

Comparing SARS-CoV-2 with SARS-CoV and influenza pandemics

Genomic characterization of the 2019 novel humanpathogenic coronavirus isolated from a patient with atypical pneumonia after visiting Wuhan

A Sequence Homology and Bioinformatic Approach Can Predict Candidate Targets for Immune Responses to SARS-CoV-2

Differences and similarities between SARS-CoV and SARS-CoV-2: spike receptor-binding domain recognition and host cell infection with support of cellular serine proteases

SARS-CoV-2 cellular tropism

Resource-efficient internally controlled in-house realtime PCR detection of SARS-CoV-2

Detection of 2019 novel coronavirus (2019-nCoV) by real-time RT-PC R

Sample Preparation by Easy Extraction and Digestion (SPEED) -A Universal, Rapid, and Detergent-free Protocol for Proteomics Based on Acid Extraction

Protocol for micropurification, enrichment, pre-fractionation and storage of peptides for proteomics using StageTips

The Skyline ecosystem: Informatics for quantitative mass spectrometry proteomics

UniProt: a worldwide hub of protein knowledge

Prosit: proteome-wide prediction of peptide tandem mass spectra by deep learning

Chromatogram libraries improve peptide detection and quantification by data independent acquisition mass spectrometry

DIA-NN: neural networks and interference correction enable deep proteome coverage in high throughput

The Perseus computational platform for comprehensive analysis of (prote)omics data

ClueGO: a Cytoscape plug-in to decipher functionally grouped gene ontology and pathway annotation networks

Cytoscape: a software environment for integrated models of biomolecular interaction networks

Isolation Window Optimization of Data-Independent Acquisition Using Predicted Libraries for Deep and Accurate Proteome Profiling

SARS-CoV-2 sensing by RIG-I and MDA5 links epithelial infection to macrophage inflammation

Multilevel proteomics reveals host perturbations by SARS-CoV-2 and SARS-CoV

Proteomics of SARS-CoV-2-infected host cells reveals therapy targets

Systematic analysis of SARS-CoV-2 infection

Delayed induction of type I and III interferons mediates nasal epithelial cell permissiveness to SARS-CoV-2

STRING 8-a global view on proteins and their functional interactions in 630 organisms

Type I Interferon Susceptibility Distinguishes SARS-CoV-2 from SARS-CoV

Antiviral activities of type I interferons to SARS-CoV-2 infection

Differential immune activation profile of SARS-CoV-2 and SARS-CoV infection in human lung and intestinal cells: Implications for treatment with IFN-beta and IFN inducer

Interplay between SARS-CoV-2 and the type I interferon response

Impaired type I interferon activity and inflammatory responses in severe COVID-19 patients

SARS-CoV-2 Infection Boosts MX1 Antiviral Effector in COVID-19 Patients

Imbalanced Host Response to SARS-CoV-2 Drives Development of COVID-19

Activation and evasion of type I interferon responses by SARS-CoV-2

Interferon antagonism by SARS-CoV-2: a functional study using reverse genetics

Severe acute respiratory syndrome coronavirus M protein inhibits type I interferon production by impeding the formation of TRAF3.TANK.TBK1/IKKepsilon complex

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) membrane (M) protein inhibits type I and III interferon production by targeting RIG-I/ MDA-5 signaling