(534f) Combinatorial Chimeragenesis of Novel Lysins to Treat Antibiotic Resistant Infections

Bacterial pathogens are known to quickly evolve or acquire resistance to new antibiotics, and the global spread of multidrug-resistant bacteria represents a critical threat to public health. Additionally, the decline in approval of new antibacterial chemotherapies is prompting a search for innovative treatment paradigms. The use of bacteriolytic enzymes as biotherapies presents a novel approach to combatting bacterial infections. Antibacterial enzymes, or ‘lysins’, actively degrade the peptidoglycan of pathogenic bacteria, resulting in rapid cell lysis and death. As a result of this unique mechanism of action, lysins can effectively kill contemporary drug-resistant strains while suppressing, or at least slowing, the development of new resistance phenotypes.

Lysin structures are highly modular, and typically consist of a catalytic domain that cleaves some moiety of the peptidoglycan, and a cell wall binding domain that targets the enzyme to the cell wall. Engineering lysins as an antibacterial therapy typically involves synthetically fusing catalytic domains to known cell wall binding domains to generate recombinant protein chimeras with enhanced activity. Thus far, the approach has been limited to the rational design of a small group of chimeric proteins, resulting in a slow trickle of putative antimicrobial enzymes. A generalizable approach to rapidly screen different combinations of lysin catalytic domains and cell wall binding domains would greatly enhance the throughput of candidate development.

We have developed a pipeline that combines both computational and experimental methodologies to discover, engineer, and screen for novel synthetic lysins with enhanced therapeutic efficacy. While our initial focus is on the critical pathogen Staphylococcus aureus (S. aureus), our approach is broadly generalizable to the discovery and engineering of bacteriolytic enzymes against any gram-positive bacteria.

We first performed an extensive bioinformatics search using known lysin sequences as seeds to generate a database of approximately 20,000 putative lytic enzymes, comprised of 80,000 unique domains, derived from a diverse collection of over 4,500 bacterial organisms and bacteriophage. We next performed phylogenetic and sequence analysis to identify forty-nine diverse catalytic domains for experimental testing. These catalytic domains represent seven classes of lytic enzymes that target different moieties of the peptidoglycan. We subsequently developed a modular combinatorial chimeragenesis approach to shuffle individual catalytic domain sequences together with cell wall binding domain sequences to generate thousands of unique synthetic lysin sequences. We combined the forty-nine catalytic domains with four linkers and thirteen high-performance cell wall binding domains to generate a 2,548-member library. While our initial library was designed for a simple two-domain fusion, our approach is inherently scalable and capable of generating larger and more diverse multi-domain fusions. We subsequently identified lysin variants with activity against S. aureus via a halo-forming screen of our recombinant Escherichia coli expression library. Using this approach, we tested over 15,000 individual clones and identified six novel lysins with anti-staphylococcal activity (Figure). We are currently characterizing these lysins in vitro, and believe they could represent new therapeutic candidates against antimicrobial-resistant infections. Future directions will focus on further enhancing the therapeutic potential of these enzymes through both computational design and directed evolution, and expanding our discovery pipeline to other important bacterial pathogens. We believe our pipeline, and the engineered lysins we have created will provide valuable new tools in the fight against drug-resistant bacterial pathogens.