AI-guided computational design of anti-virulence proteins as sustainable alternative antibacterials

From 01-10-2024 to 30-09-2028

Description

Antibiotics are some of the most essential drugs used in the clinic. The livestock industry also currently relies extensively on antibiotics for disease prevention and growth promotion. Significant ongoing concerns about the further emergence of antibiotic- resistant bacteria are thus a major medical, veterinary, and economic problem. In response to this pressing issue, we propose a groundbreaking project aimed at developing sustainable biologic alternatives to antibiotics. Our approach builds on the recently developed SymBody platform (Voet lab), which are ultrastable proteins that do not provoke an immune response and that can be produced sustainably. They can be administered orally through the gastrointestinal track without degradation, as well as intravenously.

Rather than relying on classical display methods to optimize SymBodies, in this project we will combine protein engineering with machine learning experience to enhance
the platform. This will result in an AI-based protein engineering method that can more rapidly design the desired protein for a given target. During the AI-design and evaluation phase, we will also take into account the optimization of the upscaling of the production through bacterial fermentation.

The Symbody-antibacterials (SymbAs) will be designed to be evolutionarily robust by following a double safety approach. On the one hand, the SymbAs will be iteratively designed by AI to interact with the target in such a way that it is more difficult
for resistant mutants to originate. On the other hand, by focusing on a virulence factor as target we aim to reduce the selective pressure on resistant mutants if they still would originate, and as such avoid their spread and enrichment within the pathogen population. Indeed, contrary to traditional antibiotics that target essential enzymes, anti-virulence drugs targeting specific types of virulence factors (such as public virulence factors or coincidental virulence factors) are predicted to be less prone to resistance selection. In this project, we will design SymbAs targeting the bacterial virulence factor Sortase A of both human pathogens and those relevant to animal husbandry. SortA is a validated target that is present in most Gram-positive bacteria. Over several iterative rounds, we will create and validate optimized proteins that can neutralize the virulence of pathogenic bacterial strains (e.g. MRSA, VRE, E. cecorum, S. suis,...) as well as validate the virulence targets as evolutionarily robust.

Hence, this project will provide a platform to further develop SymbAs targeting other virulence factors for different pathogens, as well as the expertise in AI-based protein design for different synthetic biology projects.

Team

• Yves Moreau, Co-promoter

Financing

Funding: KU Leuven - Internal Funding KU Leuven

Program/Grant Type: IDN - Interdisciplinary Networks