henhaobio
Sign in

NatureSystems BiologyPeer-reviewed journal article

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

Alexey StukaloviD, Virginie GiraultiD, Vincent GrassiD, Özge KarayeliD, Valter BergantiD, Christian UrbaniD, Darya A. HaasiD, Yiqi HuangiD, Lila OubrahamiD, Anqi WangiD, M. Sabri Hamad, Antonio PirasiD, Fynn M. HanseniD, Maria C. TanzeriD, Igor Paron, Luca ZinzulaiD, Thomas EngleitneriD, Maria ReineckeiD, Teresa M. LavaccaiD, Rosina K. EhmanniD, Roman WölfeliD, Joerg JoresiD, Bernhard KüsteriD, Ulrike ProtzeriD, Roland RadiD, , Andreas PichlmairiD

PDFDOIPreprintCodeData
In-depth full-text digest available · Jump to digest ↓

Abstract

The global emergence of SARS-CoV-2 urgently requires an in-depth understanding of molecular functions of viral proteins and their interactions with the host proteome. Several individual omics studies have extended our knowledge of COVID-19 pathophysiology1-10. Integration of such datasets to obtain a holistic view of virus-host interactions and to define the pathogenic properties of SARS-CoV-2 is limited by the heterogeneity of the experimental systems. We therefore conducted a concurrent multi-omics study of SARS-CoV-2 and SARS-CoV. Using state-of-the-art proteomics, we profiled the interactome of both viruses, as well as their influence on transcriptome, proteome, ubiquitinome and phosphoproteome in a lung-derived human cell line. Projecting these data onto the global network of cellular interactions revealed crosstalk between the perturbations taking place upon SARS-CoV-2 and SARS-CoV infections at different layers and identified unique and common molecular mechanisms of these closely related coronaviruses. The TGF-{beta} pathway, known for its involvement in tissue fibrosis, was specifically dysregulated by SARS-CoV-2 ORF8 and autophagy by SARS-CoV-2 ORF3. The extensive dataset (available at https://covinet.innatelab.org) highlights many hotspots that can be targeted by existing drugs and it can guide rational design of virus- and host-directed therapies, which we exemplify by identifying kinase and MMPs inhibitors with potent antiviral effects against SARS-CoV-2.

In-depth summary

Claude Sonnet 4.6 · xhigh effort · 8 Nov 2026 · from the full text of version 2 (15 Mar 2021) · CC BY 4.0

Concurrent multi-omics of SARS-CoV-2 and SARS-CoV in A549 cells, integrated by network diffusion, maps shared and virus-specific host perturbations and nominates antiviral drug hotspots.

Earlier omics studies of coronavirus infection used heterogeneous experimental systems, which prevented a single integrated view of virus-host interactions. This work profiles both viruses in one lung-derived cell system across interactome, effectome, transcriptome, proteome, ubiquitinome and phosphoproteome.

Key results

  1. ShownAP-MS of individually expressed viral proteins defined a large virus-host interactome, including a SARS-CoV-2-exclusive ORF8 interaction with the TGFB1-LTBP1 complex that points to TGF-β pathway targeting.

    Host interactions identified for 24 SARS-CoV-2 and 27 SARS-CoV baits (four replicate purifications per bait); ORF8 binding to TGFB1-LTBP1 seen only for SARS-CoV-2; endogenous TGF-β co-precipitation tested in n=2. · Comparative SARS-CoV-2 and SARS-CoV virus-host…; Affinity purification and mass spectrometric…; Figure legends

  2. ShownTime-resolved infection profiling showed extensive, partly virus-specific regulation of ubiquitination and phosphorylation, including SARS-CoV-2-only activation of the antiviral kinase PKR at S33.

    Large sets of ubiquitination and phosphorylation sites changed after infection; more than half of significant ubiquitination sites were regulated similarly by both viruses; RAB7A S72 higher in SARS-CoV-2 than SARS-CoV or mock. · Multi-omics profiling of SARS-CoV-2 and SARS-CoV…

  3. SuggestedNetwork diffusion linking interactors to proteome effects of each viral protein suggested that SARS-CoV-2 ORF3 blocks autophagic flux, with APOB accumulation as a downstream consequence.

    Diffusion-derived connections were significantly shorter than in 1000 permutations of vertex weights; MAP1LC3B accumulation by western blot on ORF3 expression (n=3) and infection (n=3); APOB abundance increased (n=4). · Integrative analysis highlights the perturbation…; Network diffusion analysis; Figure legends

  4. ShownA data-guided screen of pathway-targeting drugs found compounds that suppressed SARS-CoV-2 growth, with MMP inhibitors selective for SARS-CoV-2 over SARS-CoV, while B-RAF and JAK1/2 inhibitors increased growth.

    48 drugs tested (SARS-CoV-2-GFP, MOI 3, n=4, Wilcoxon); Gilteritinib, Ipatasertib, Prinomastat, Marimastat strongest; Prinomastat and Marimastat lowered SARS-CoV-2 but not SARS-CoV N RNA by qRT-PCR (n=3). · Data-guided drug testing reveals hotspots for…; Viral inhibitor assay; Figure legends

  5. ShownMost detected viral proteins carried ubiquitin or phosphate modifications, including spike ubiquitination at conserved sites and new nucleocapsid sites mapped onto the C-terminal domain structure.

    21 of 27 detected viral proteins ubiquitinated; 5 SARS-CoV-2 and 8 SARS-CoV proteins phosphorylated; spike sites K97, K528, K825, K835, K921, K947; SARS-CoV-2 N K338 ubiquitination; N CTD superposition r.m.s.d. 0.492 Å. · Phosphorylation and ubiquitination of functional…; Figure legends

Methods, data and reproducibility

HA-tagged viral proteins were expressed in A549 cells for AP-MS and DIA proteomics, and ACE2-expressing A549 cells were infected for time-course RNA-seq, proteome, phospho- and diGly proteomics. Bayesian linear models and HierarchicalHotNet diffusion on ReactomeFI integrated the layers.

  • AP-MS interactome of SARS-CoV-2 and SARS-CoV baits plus HCoV-229E/NL63 ORF3 homologs in A549 cells · Human cells (A549); SARS-CoV-2, SARS-CoV · n = 24 SARS-CoV-2 and 27 SARS-CoV baits; four replicates each · PXD022282, PXD020461, PXD020222; IntAct IM-28109
  • DIA total proteome (effectome) of A549 cells expressing individual viral proteins · Human cells (A549) · n = 54 viral proteins
  • Time-course of SARS-CoV-2 and SARS-CoV infection in A549-ACE2 cells (MOI 2): proteome, phosphoproteome, ubiquitinome · Human cells (A549-ACE2); SARS-CoV-2, SARS-CoV · n = n=4 proteome and phospho; n=3 ubiquitination · PXD022282, PXD020461, PXD020222
  • Transcriptome of infected A549-ACE2 cells (mock, SARS-CoV, SARS-CoV-2) over time · Human cells (A549-ACE2) · n = n=3 independent experiments · PRJEB38744
  • Live-cell imaging antiviral screen with SARS-CoV-2-GFP reporter in A549-ACE2 cells · Human cells (A549-ACE2); SARS-CoV-2 · n = 48 drugs; n=4
  • Nucleocapsid C-terminal domain structures used for site mapping · SARS-CoV-2, SARS-CoV · PDB 6YUN, 2CJR
Methods & tools
MaxQuant 1.6.14, Spectronaut 13/14, msglm with rstan 2.19, Limma 3.46.0, Dropseq tool v1.12, HierarchicalHotNet.jl, ReactomeFI network (2019), OptEnrichedSetCover.jl, Ingenuity Pathway Analysis, PyMOL, Chimera, PDBePISA, IncuCyte S3 imaging, Cytoscape v.3.8.1
Software & environment
R 3.6, Julia 1.5, Python 3.8; rstan 2.19; MaxQuant 1.6.14; Spectronaut 13 and 14; Limma 3.46.0; in-house R and Julia packages.
Reproducibility gaps
qRT-PCR primer sequences only on request; the MCMC models were fit per protein group at high computational cost, with no hardware stated.
Limitations (2)
  • Omics data come mainly from one lung-derived line (A549, with ectopic ACE2 for infection) and rely on viral protein overexpression; some validations used HEK293-derived cells, and drug effects were tested only in vitro, so physiological relevance is untested. · noted in this summary
  • Most MS significance calls use unadjusted p-value cutoffs with no multiple-testing correction (model priors are relied on instead), and the co-IP validations have only n=2, so false positives are possible. · noted in this summary

Read this if you want a matched multi-omics resource comparing SARS-CoV-2 and SARS-CoV, or Bayesian MS modeling and network diffusion for drug target prioritization.

Citations
686
FWCI
45.25
Field percentile
top 1%

Reproducibility ledger

Nobody has reported trying to reproduce this paper yet. Did you replicate an experiment or run the code? Your report saves the next lab a week.

Sign in to file a reproducibility report.

Related papers

Similarity Map
SelectedSame fieldCross-field
THISLaurent '20Chen '20May '22Liang '20Kim '22Messina '20
72% semantic matchSystems Biology
Estelle M. N. Laurent, Yorgos Sofianatos, Anastassia V. Komarova+28 more·bioRxiv·Aug 28, 2020· 71 citations