henhao
Sign in
Systems BiologyNature
Peer-reviewed journal article

A SARS-CoV-2-Human Protein-Protein Interaction Map Reveals Drug Targets and Potential Drug-Repurposing

David Ezra GordoniD, Gwendolyn Μ. JangiD, Mehdi BouhaddouiD, Jiewei XuiD, Kirsten ObernieriD, Kris M. WhiteiD, Matthew J. O’MearaiD, Veronica V. RezeljiD, Jeffrey Z. GuoiD, Danielle L. SwaneyiD, Tia A. TumminoiD, Ruth HüttenhainiD, Robyn M. KaakeiD, Alicia RichardsiD, Beril TutuncuogluiD, Helene FoussardiD, Jyoti BatraiD, Kelsey M. HaasiD, Maya ModakiD, Minkyu KimiD, Paige HaasiD, Benjamin J. PolaccoiD, Hannes BrabergiD, Jacqueline M. Fabius, Manon EckhardtiD, , Natalia JuraiD

Abstract

An outbreak of the novel coronavirus SARS-CoV-2, the causative agent of COVID-19 respiratory disease, has infected over 290,000 people since the end of 2019, killed over 12,000, and caused worldwide social and economic disruption1,2. There are currently no antiviral drugs with proven efficacy nor are there vaccines for its prevention. Unfortunately, the scientific community has little knowledge of the molecular details of SARS-CoV-2 infection. To illuminate this, we cloned, tagged and expressed 26 of the 29 viral proteins in human cells and identified the human proteins physically associated with each using affinity-purification mass spectrometry (AP-MS), which identified 332 high confidence SARS-CoV-2-human protein-protein interactions (PPIs). Among these, we identify 66 druggable human proteins or host factors targeted by 69 existing FDA-approved drugs, drugs in clinical trials and/or preclinical compounds, that we are currently evaluating for efficacy in live SARS-CoV-2 infection assays. The identification of host dependency factors mediating virus infection may provide key insights into effective molecular targets for developing broadly acting antiviral therapeutics against SARS-CoV-2 and other deadly coronavirus strains.

PDFDOICode
Full-text digest ↓

02 /Specification & Classification

Click any tag to browse matching papers
Citations
4,405
FWCI
117.80
Field percentile
top 1%
OpenAlex Topics

In-depth summary

✦ Multi-agent verified full-text synthesis

Claude Sonnet 4.6 · max effort · 8 Nov 2026 · from the full text of version 3 (27 Mar 2020) · CC BY 4.0

Affinity-purification mass spectrometry of 26 SARS-CoV-2 proteins in HEK293T cells yields 332 high-confidence host interactions and 67 druggable human proteins with known compounds.

Molecular knowledge of SARS-CoV-2 infection was minimal and no antivirals or vaccines with proven efficacy existed, which limited host-directed therapy design. This work maps virus-host physical interactions and overlays chemoinformatic drug searches, giving candidate compounds a mechanistic context.

Key results

  1. ShownSystematic AP-MS of tagged viral proteins in human cells produced a network of 332 high-confidence viral-human protein interactions, analysed from 27 baits.

    Step 1 filter: MiST ≥ 0.7, SAINTexpress BFDR ≤ 0.05, average spectral count ≥ 2; step 2 relaxed to MiST ≥ 0.6 for complex members. Orf7b excluded for high background; at least three biological replicates per bait. · Cloning and expression of predicted SARS-CoV-2…; Global analysis of SARS-CoV-2 host interacting…; Transfection; Mass spectrometry data acquisition and analysis

  2. ShownInteracting human proteins are preferentially expressed in lung, the typical site of infection, even though the AP-MS cells are not lung-derived. Interacting viral-human pairs also showed more correlated abundance changes during infection than other pairs.

    Lung ranked highest of 29 tissues; interactor median 25.52 TPM vs 3.198 TPM for overall lung RefSeq expression (p=0.0007, t-test). Correlation of interacting pairs vs other pairs: KS test p=4.8e−05. · Global analysis of SARS-CoV-2 host interacting…

  3. ShownChemoinformatic and expert-curated searches matched many interactors to approved, clinical, or preclinical compounds, linking drugs to specific viral-host interfaces such as BRD2/4, HDAC2, IMPDH2, Sigma receptors and mTORC1. Antiviral efficacy had not yet been tested.

    Chemoinformatics: 15 approved, four clinical, 18 preclinical; expert curation: 12 approved, 10 clinical, 10 preclinical. 63 of 332 interactors have modulators; 67 targets after relaxing the score threshold. ABBV-744 and CPI-0610 bind BRD2/4 at 2 nM and 39 nM. · Abstract; Identification of existing drugs targeting…; Chemoinformatic Analysis of SARS-CoV2…

  4. SuggestedThe map yields mechanistic hypotheses: Orf6 with the NUP98-RAE1 export complex, Orf10 with the CUL2-ZYG11B ligase, and E with BRD2/BRD4 via apparent histone H2A mimicry. The interactions come from AP-MS, while the proposed mechanisms rest on modelling and sequence similarity.

    Orf6 C-terminal motif matches VSV M protein features for NUP98-RAE1 binding; ZYG11B was the highest-scoring Orf10 interactor; E residues 25-44 superposed on H2A residues 49-60 and 63-70 at RMSD 2.8Å. · SARS-CoV-2 interacts with multiple innate immune…; The novel Orf10 of SARS-CoV-2 interacts with a…; SARS-CoV-2 envelope interacts with bromodomain…; Alignment of Protein E and Histone H2A

  5. SuggestedThe SARS-CoV-2 interactome overlaps most with those of West Nile Virus and Mycobacterium tuberculosis among 10 other pathogens. Its human interactors also carry reduced genetic variation in gnomAD.

    Pairwise overlap tested by hypergeometric test with background N=10,181 genes; the 332 interactors had depleted missense and premature stop variants relative to control genes in gnomAD. · Global analysis of SARS-CoV-2 host interacting…; Virus Interactome Similarity Analysis

Methods, data and reproducibility

Codon-optimised, 2xStrep-tagged viral proteins were expressed in HEK293T cells and affinity purified, then identified by LC-MS/MS. Interactors were scored with SAINTexpress and MiST and mapped to drugs by chemoinformatic and expert searches.

  • AP-MS of 2xStrep-tagged SARS-CoV-2 proteins in HEK293T cells (human cell line); wild-type constructs plus the Nsp5 C145A catalytic-dead mutant · human (HEK293T cells) · n = 27 baits analysed; at least three biological replicates · PXD018117
  • SARS-CoV-2 isolate 2019-nCoV/USA-WA1/2020 genome used for annotation of 29 open reading frames and mature proteins · SARS-CoV-2 · n = 29 ORFs/mature proteins · MN985325
  • In-house reference database of AP-MS baits used for MiST scoring of Orf8, Nsp8, Nsp13 and Orf9c · n = 87 baits
  • Protein abundance across human tissues and published infection time-course protein changes used to assess interactor relevance · human · n = 29 tissues; 4 timepoints
  • AP-MS interactomes of 10 other pathogens compared by shared human interactors · multiple pathogens · n = 10 pathogens; N=10,181 background genes
  • Drug and ligand databases: IUPHAR/BPS Guide to Pharmacology (2020-3-12) and ChEMBL25 · human targets · n = 2,472 and 10,883 molecules
  • Structural templates for modelling: VSV M-NUP98-RAE1 (4OWR), SARS-CoV E (2MM4), human nucleosome (6K1K) · PDB 4OWR, 2MM4, 6K1K
Methods & tools
2xStrep tag / MagStrep affinity purification, Q-Exactive Plus with Easy-nLC 1200, MaxQuant, SAINTexpress, MiST, CORUM, Cytoscape, clusterProfiler with MSigDB v6.1, PyRosetta 4 (v2020.02-dev61090), PyMOL 2.3.4, SwissModel and Modeller, IUPHAR/BPS Guide to Pharmacology, ChEMBL25, ZINC
Code
http://www.github.com/momeara/BioChemPantry/vignette/COVID19
Reagents & environment
HEK293T cells; 2xStrep-tagged constructs in a mammalian expression vector; MagStrep type3 beads; MaxQuant default settings against UniProt reviewed human sequences (28 Feb 2020); PyRosetta 4 v2020.02-dev61090; PyMOL 2.3.4; MSigDB v6.1.
Experimental conditions
Q-Exactive Plus with Easy-nLC 1200 (75 min acquisition); KingFisher Flex purification at 4°C; cells incubated for 40 hours before lysis; on-bead trypsin digestion at 37°C.
Reproducibility gaps
PRIDE accession PXD018117 and NDEx networks are cited without URLs. Supplementary tables are not in the text. Drug efficacy assays were still in progress.
Limitations (5)
  • HEK293T cells are permissive to infection but are not the primary physiological site of infection; the authors plan interaction maps in other human cell types and bat cells. · stated by the authors
  • Functional relevance of the interactions needs genetic validation. Whether human proteins aid or restrict the virus is unknown, so drugs against them could be detrimental or beneficial; antiviral testing was ongoing. · stated by the authors
  • Western blot did not confirm Nsp4, Nsp6, Nsp11 or Orf3b; the authors suggest Orf3b may not be a genuine protein-coding gene or may need other viral proteins for robust expression. · stated by the authors
  • Individually overexpressed, tagged viral proteins in the absence of infection may not reflect interactions in infected cells; no orthogonal validation of individual PPIs is reported in the text. · noted in this summary
  • The proposed Orf6, Orf10 and E-histone mechanisms rest on modelling and sequence similarity (RMSD 2.8Å for the E-H2A superposition), with no functional assays reported beyond AP-MS. · noted in this summary

Who should read:Read this if you want an early SARS-CoV-2 host interactome with drug-target mapping to prioritise host-directed antiviral candidates and mechanistic hypotheses.

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
THISChen '20Caillet-Sag '21Miller '25Labeau '21Guzzi '20Hoffmann '20
74% semantic matchMolecular Biology
Zhen Chen, Chao Wang, Xu Feng+7 more·bioRxiv·Dec 31, 2020· 15 citations