Poster #26: Cameron Liggett

A Multi-Evidence Computational Pipeline Prioritizes Candidate HSV-1 Us3 Kinase Substrates Identified by Proximity-based Phosphoproteomics

Cameron Liggett¹; Pierce Longmire, PhD¹˒²; Katharina E. Hayer, MS¹˒²˒³; Matthew D. Weitzman, PhD¹˒²; ¹ Division of Protective Immunity, Children’s Hospital of Philadelphia, Philadelphia, PA, USA; ² Department of Pathology and Laboratory Medicine, University of Pennsylvania Perelman School of Medicine, Philadelphia, PA, USA; ³ Department of Biomedical and Health Informatics, Children’s Hospital of Philadelphia, Philadelphia, PA, USA

Herpes simplex virus type 1 (HSV-1) is a human pathogen associated with severe disease outcomes for immunocompromised individuals and newborns. Virus replication occurs in the host cell nucleus, where viral progeny must navigate the complex network of chromatin and nuclear lamina to egress during packaging of the viral genome. The viral serine/threonine kinase Us3 phosphorylates nuclear lamina proteins to disrupt the lamina meshwork to facilitate nuclear escape. However, we lack a complete understanding of the role of Us3 in this process. Using a biotin identification-based phosphoproteomics approach to characterize Us3 substrates, we identified 1,022 candidate proteins, many of which may reflect spatial proximity to Us3 rather than direct phosphorylation. Literature curation of such lists is time-consuming and can be biased toward better-studied proteins. We wished to develop a computational pipeline which could rank likely “true” Us3 substrates higher than physically near proteins. We developed the Phosphoproteomics Prioritization Pipeline (PPP), which scores candidates across five channels: 1) STRING v12.0 functional-association PageRank proximity to 18 curated lamina proteins; 2) Us3 motif compatibility using the published consensus sequence and a bootstrapped null; 3) literature co-occurrence using PubTator3; 4) Gene Ontology overlap with lamina and nuclear-envelope terms; and 5) unsupervised LDA topic similarity to lamina literature. Channel ranks are fused using Robust Rank Aggregation, which prioritizes candidates that rank consistently well across channels beyond random expectation. A mechanistic sub-track incorporating only network and motif evidence reduces dependence on literature-derived information. Three known Us3 substrates ranked among the top 10 candidates. Further prioritization on the mechanistic sub-track included chromatin and nuclear-regulatory proteins AHCTF1, NUMA1, and SMARCA4. The PPP pipeline is intended to prioritize candidate kinase substrates rather than infer direct phosphorylation. All predictions require biochemical confirmation. Our methodology is broadly applicable to other kinases with known substrate motifs and relevant interaction context.