Designing Anti-Zika Virus Peptides Derived from Predicted Human-Zika Virus Protein-Protein Interactions

Thomas David Daniel Kazmirchuk, Kevin Dick, Daniel Burnside, Brad Barnes, Houman Moteshareie, Maryam Hajikarimlou, Katayoun Omidi, Duale Ahmed, Andrew Low, Clara Lettl, Mohsen Hooshyar, Andrew Schoenrock, Sylvain Pitre, Mohan Babu, Edana Cassol, Bahram Samanfar, Alex Wong, Frank Dehne, James R. Green, Ashkan GolshanibioRxiv 2017 · DOI

Abstract

The production of anti-Zika virus (ZIKV) therapeutics has become increasingly important as the propagation of the devastating virus continues largely unchecked. Notably, a causal relationship between ZIKV infection and neurodevelopmental abnormalities has been widely reported, yet a specific mechanism underlying impaired neurological development has not been identified. Here, we report on the design of several synthetic competitive inhibitory peptides against key pathogenic ZIKV proteins through the prediction of protein-protein interactions (PPIs). Often, PPIs between host and viral proteins are crucial for infection and pathogenesis, making them attractive targets for therapeutics. Using two complementary sequence-based PPI prediction tools, we first produced a comprehensive map of predicted human-ZIKV PPIs (involving 209 human protein candidates). We then designed several peptides intended to disrupt the corresponding host-pathogen interactions thereby acting as anti-ZIKV therapeutics. The data generated in this study constitute a foundational resource to aid in the multi-disciplinary effort to combat ZIKV infection, including the design of additional synthetic proteins.

Introduction

Despite the global health threat posed by the Zika virus—particularly its link to microcephaly and other neurodevelopmental abnormalities—the molecular mechanisms the virus uses to hijack human cells remain partially understood. Viral replication and pathogenesis depend heavily on Protein-Protein Interactions (PPIs) between the virus and the host. Mapping these interactions experimentally is slow and expensive, necessitating rapid, systematic computational approaches to identify these critical "handshakes."

The authors employed a dual-algorithm approach to predict the human-ZIKV interactome:

  1. PIPE (Protein-Protein Interaction Prediction Engine): A sequence-based tool identifying short, co-occurring polypeptide sequences.
  2. DeNovo: A tool designed for cross-species interactions that exploits the physicochemical properties of host proteins.

By screening the entire human proteome (20,515 proteins) against the 11 known ZIKV proteins, the researchers identified specific "PPI-Sites"—the exact amino acid sequences responsible for binding—which served as the basis for designing synthetic inhibitory peptides.

The Human-ZIKV Interactome

The study produced a comprehensive map of 871 predicted human-ZIKV interactions involving 209 human protein candidates. These proteins were classified based on their biological processes, cellular components, and molecular functions to better understand the virus's impact on the host cell.

Functional classification of predicted human protein candidates interacting with ZIKV.

Analysis revealed that ZIKV proteins, particularly NS3 and NS5, are highly "social," interacting with a wide array of human proteins involved in metabolic processes, cellular organization, and immune responses. A significant portion of these interactions occurs in the cytoplasm and nucleus, reflecting the virus's primary sites of replication and interference.

Targeting Pathogenesis and the "Decoy" Strategy

Rather than addressing all 209 predicted interactions, the authors prioritized 25 high-priority human proteins directly linked to ZIKV symptoms and neurodevelopmental phenotypes.

High-priority human proteins and their interactions with ZIKV proteins linked to specific disease phenotypes.

The core therapeutic strategy involves using synthetic peptides as "molecular decoys." By flooding a cell with synthetic versions of the interaction site, viral proteins bind to these decoys instead of the actual human proteins, effectively neutralizing the virus's ability to interfere with host cell function. For instance, targeting proteins like RNASET2 and CEP63, which are independently associated with microcephaly, provides a direct path to mitigating the most severe effects of infection.

Key Results: Lead Peptide Candidates

The researchers identified four specific peptide sequences (P1 through P4) as primary therapeutic candidates, each targeting different stages of the viral lifecycle or specific pathogenic pathways:

  • P1 (targets RNASET2): Prevents ZIKV NS3 from interfering with essential RNA catabolism in lysosomes.
  • P2 (targets ENO2): Protects a protein crucial for early brain development and neuronal maturation.
  • P3 (targets TRAF4): Blocks viral absorption and protects neural crest formation by disrupting interactions with the ZIKV Membrane (M) protein.
  • P4 (targets CEP63): Stops the virus from triggering p53-dependent microcephaly.

Limitations and Future Directions

This study is entirely in silico (computational). While based on robust, validated algorithms, these peptides require "wet lab" testing in cell cultures and animal models to confirm efficacy and safety. Future work must address experimental validation, peptide engineering to improve stability and half-life, and the development of effective delivery mechanisms to reach sensitive tissues like the brain or a developing fetus.

Summary updated

Abstract

Abstract The production of anti-Zika virus (ZIKV) therapeutics has become increasingly important as the propagation of the devastating virus continues largely unchecked. Notably, a causal relationship between ZIKV infection and neurodevelopmental abnormalities has been widely reported, yet a specific mechanism underlying impaired neurological development has not been identified. Here, we report on the design of several synthetic competitive inhibitory peptides against key pathogenic ZIKV proteins through the prediction of protein-protein interactions (PPIs). Often, PPIs between host and viral proteins are crucial for infection and pathogenesis, making them attractive targets for therapeutics. Using two complementary sequence-based PPI prediction tools, we first produced a comprehensive map of predicted human-ZIKV PPIs (involving 209 human protein candidates). We then designed several peptides intended to disrupt the corresponding host-pathogen interactions thereby acting as anti-ZIKV therapeutics. The data generated in this study constitute a foundational resource to aid in the multi-disciplinary effort to combat ZIKV infection, including the design of additional synthetic proteins.
BibTeX
@misc{kazmirchuk2017designing,
  author = {Thomas David Daniel Kazmirchuk and Kevin Dick and Daniel Burnside and Brad Barnes and Houman Moteshareie and Maryam Hajikarimlou and Katayoun Omidi and Duale Ahmed and Andrew Low and Clara Lettl and Mohsen Hooshyar and Andrew Schoenrock and Sylvain Pitre and Mohan Babu and Edana Cassol and Bahram Samanfar and Alex Wong and Frank Dehne and James R. Green and Ashkan Golshani},
  title = {{Designing Anti-Zika Virus Peptides Derived from Predicted Human-Zika Virus Protein-Protein Interactions}},
  year = {2017},
  doi = {10.1101/156695},
  howpublished = {bioRxiv},
}
Figure