---
title: SARS-CoV-2 Host-Viral Interactome
url: https://www.emergentmind.com/topics/host-viral-interactome-of-sars-cov-2
type: topic
---

# SARS-CoV-2 Host-Viral Interactome

The host-viral interactome of SARS-CoV-2 comprises the network of molecular interactions between viral proteins and host (predominantly human) proteins, forming a systems-level landscape that underlies viral entry, replication, immune evasion, and COVID-19 pathogenesis. Large-scale affinity-purification mass spectrometry (AP-MS), computational network propagation, and ontology-driven knowledge graphs have mapped a repertoire of direct and indirect interactions. These analyses identify core “hubs,” functional modules, and disease subnetworks, providing mechanistic bases for therapeutic targeting and drug repurposing.

## 1. Experimental and Computational Mapping of the SARS-CoV-2 Host-Viral Interactome

Reference human-SARS-CoV-2 interactome mapping began with high-throughput AP-MS experiments, notably by Gordon et al. (2020), which identified 332 high-confidence physical interactions between 27 SARS-CoV-2 viral “baits” and human proteins [2006.01968][2004.12420]. This seed set serves as the substrate for computational expansion and systems analysis.

Network propagation methods such as the Regularized Laplacian (RL) kernel allow diffusion of initial experimentally validated interactions across a weighted human protein–protein interaction (PPI) network. For a graph $G = (V,E,w)$ (nodes = proteins; edges = interactions; $w_{uv}$ = confidence weights), an initial binary label vector $y$ is set ($y_u = 1$ if $u$ is a direct viral interactor). The diffusion yields $s = (I+\alpha \tilde{L})^{-1} y$, where $\tilde{L}$ is the normalized Laplacian and $\alpha$ is tuned so the expected walk length matches the empirical median shortest-path distance from viral seeds (typically $\ell\approx3$ steps for $\alpha\approx3.4$) [2006.01968]. 

These approaches can be extended with network-based provenance tracing, which decomposes the influence of each seed interactor on candidate protein predictions, yielding a transparent “audit trail” from prediction to experimental evidence [2006.01968].

Table 1: Major Experimentally Validated and Predicted PPI Databases

| Database/Method         | Description                                       |
|------------------------|---------------------------------------------------|
| AP-MS/IntAct           | Experimental virus–human PPI (332 nodes/edges)    |
| STRING                 | Human PPI confidence network for propagation      |
| CoVex                  | Platform integrating experimental, predicted PPI, drug data [2004.12420] |

## 2. Network Structure, Topological Features, and Key Interactors

Global network analysis situates the viral–host interactome within a human PPI network that exhibits scale-free topology (degree distribution $P(k)\sim k^{-\gamma}$, $\gamma\approx2.5$), high connectivity, and modular community structure [2004.12420][2009.06035].

Centrality-based ranking of human proteins (“hubs”) identifies frequent interactors and system bottlenecks:

- UBC (polyubiquitin-C): highest eigenvector and degree centrality in host-host subnetworks, suggests targeting of the ubiquitin–proteasome system for viral proteostasis [2009.06035].
- HNRNPA1, DDX5: high-degree RNA-binding proteins, co-opted for viral RNA processing [2009.06035][2102.03253].
- CDK1, PPP2R1A, RBX1: top-degree nodes implicated in cell cycle, phosphoregulation, and ubiquitin ligase activity [2102.03253].

Viral proteins such as ORF7A (deg=60; apoptosis and virion tethering), NSP13 (helicase; high vertex energy), and the RNA polymerase complex NSP7–NSP8–NSP12, are consistently key hubs, controlling replication and immune evasion [2009.06035].

## 3. Functional Module and Disease Subnetwork Analysis

Centrality and modular analyses, coupled with Gene Ontology (GO) and KEGG pathway enrichment, reveal that viral targeting is highly non-uniform across the host interactome.

Enriched functional modules in high-confidence expansions include [2006.01968][2102.03253]:

- Endoplasmic reticulum (ER) stress and protein folding (HSPA5/GRP78, CALNEXIN)
- Organelle organization, cilium assembly (tubulins)
- Vesicle trafficking, endocytosis (RAB GTPases, CAPZB)
- Ubiquitin–proteasome pathway (RBX1, ITCH, PSMC5)
- Immune signaling (STAT1, NFKB1)

Disease subnetworks derived from host factors demonstrate extensive overlap with COVID-19 comorbidities. For example, the Type 2 Diabetes module incorporates 56% of virus–host network nodes and 60% of edges, suggesting viral modulation of diabetic pathophysiology. Coagulopathy subnetworks (e.g., involving ORF7A–VKORC1) rationalize the prevalence of clotting complications in severe disease [2009.06035].

Table 2: Selected Enriched Pathways in Host Interactome

| Pathway                    | Genes | Category                |
|----------------------------|-------|-------------------------|
| Endocytosis                | 22    | Cellular                |
| ER protein folding         | >10   | Stress response         |
| Ubiquitin–proteasome       | 7     | Degradation/proteostasis|
| Type 2 Diabetes subnetwork | >40   | Metabolism/comorbidity  |

## 4. Molecular Mechanisms of Host Entry, Viral Manipulation, and Intracellular Trafficking

The foundational event in host–viral interaction is Spike-ACE2 binding, engaging clathrin-mediated endocytosis and additional co-receptors. Analysis of short linear motifs (SLiMs) in ACE2 and integrin cytosolic tails exposes combinatorial regulatory switches that control cell entry, cytoskeletal remodeling, and autophagy recruitment [2004.10274].

- ACE2 tail motifs: YxxΦ (AP2 μ₂ binding and clathrin endocytosis), LIR (LC3/GABARAP binding for autophagy), NPxY (PTB domain binding), and PDZ-binding C-terminal motif (scaffold assembly).
- Spike RGD motif and ACE2 MIDAS-like motif confer potential for integrin co-receptor usage, supporting viral entry and membrane fusion.
- Integrin tails (β₁/β₃): NPxY and LIR motifs, facilitating direct recruitment of endocytic and autophagic machinery.

Dynamic propagation analysis shows that the largest contributions to indirect host protein perturbation originate with direct neighbours in the physical interactome [2006.01968].

## 5. Information-Theoretic and Multiscale Network Approaches

Beyond node-level and topological statistics, information-theoretic and statistical physics models assess the macroscopic impact of viral perturbations.

A communication network model representing the immune system as a directed signaling graph quantifies viral “information transfer” via Kullback–Leibler divergence in node state distributions. The total information transferred by all SARS-CoV-2 PPIs corresponds to $\approx$65.4 bits, with core immune modules such as cytokine signaling, apoptosis, and mitochondrial dynamics most susceptible [2208.06355].

Spectral entropy analysis, leveraging the graph Laplacian’s Gibbsian state, clusters viruses by the magnitude and scale of network perturbation. SARS-CoV-2 is most similar to respiratory viruses (SARS-CoV, Influenza) at local scales, but clusters with retroviruses (HIV-1, HTLV-1) at global scales, explaining drug efficacy observed for HIV protease inhibitors in early clinical trials [2008.09649].

## 6. Therapeutic and Drug Repurposing Insights

Integration of interactome, pathway, and network proximity analyses yields rational targets for host-directed therapy:

- ER chaperones HSPA5 and HSPA1A: identified as direct interactors, upregulated in viral infection, and targets for ATPase inhibitors (e.g., aspirin) [2006.01968][2102.03253].
- Vesicular trafficking and RAB GTPases: GTPase inhibitors disrupt viral entry and maturation [2102.03253].
- Ubiquitin–proteasome system: inhibitors (e.g., bortezomib) shown to blunt inflammation by targeting central hubs like RBX1 [2102.03253].
- Endocytosis and kinase inhibitors: imatinib, saracatinib, and AP2/clathrin inhibitors affect uptake and viral trafficking [2004.10274].
- Repurposing of anticoagulants (tenecteplase, lanoteplase, alteplase) is supported by the link between ER-stress (HSPA5) and coagulation factor chaperoning [2006.01968].
- BTK inhibitors are top-ranked by in silico reduction of viral information control in immune signaling [2208.06355].
- Antiviral strategies targeting entry modules (e.g., camostat mesylate for TMPRSS2 activation) are rationalized by network hub analysis [2006.00639], and the frequent inclusion of UBC, ACE2, ABCB1, and CYP3A4 among FDA-approved drug targets supports druggability [2005.04108].

Mechanistic subnetworks associated with SARS-CoV-2–human interactome modules enable systematic prioritization of therapeutics in platforms such as CoVex [2004.12420].

## 7. Generalization, Ontology-Based Reasoning, and Future Directions

The methodological advances applied to SARS-CoV-2 interactome analysis—such as regularized diffusion, provenance tracing, multitask transfer learning, and ontological modeling—are generically applicable to the host interactomes of emerging pathogens. Ontology-driven frameworks (e.g., CIDO) formalize mechanistic knowledge as axioms, facilitating automated reasoning over molecular effects, clinical phenotypes, and intervention strategies [2006.00639].

By integrating layers of experimental, computational, and knowledge-based predictions, the SARS-CoV-2 host–viral interactome constitutes a reference “periodic table” for COVID-19 molecular pathogenesis and therapeutic development. Future directions include direct functional perturbation studies, deeper mapping of tissue and cell-type interactomes, and dynamic modeling of temporal evolution under therapeutic pressure.

Source: https://www.emergentmind.com/topics/host-viral-interactome-of-sars-cov-2