Great progress has been made the past several years on the discovery (cloning, expression, purification) of lysins and initial evaluation of their therapeutic potential in animal models of colonization. The Nelson group is now interested in studying a small subset of these lysins in detail to understand their basic structures, mechanisms, functions, and evolution. Based on our accumulated data and that of others, we can make a few generalized statements about these enzymes. Lysins from bacteriophage that infect Gram-positive organisms are modular enzymes composed of one or more conserved catalytic domains and a cell wall binding domain (CBD). The catalytic domain is represented by one of four families of peptidoglycan hydrolases: N-acetylglucosaminidases, N-acetylmuramidases (lysozymes), N-acetylmuramoyl-L-alanine amidases, and endopeptidases. In contrast, the CBDs are notably divergent and can distinguish discrete epitopes present within the cell wall, typically carbohydrates, giving rise to the species- or strain-specific activity of a particular lysin.

Our research focuses on four inter-related projects:

1. Identification of lysin binding receptors. The specificity of a lysin is directed by the CBD, which recognizes an epitope unique to susceptible species or serovars, much like a fingerprint. For example, the pneumococcal cell wall is characterized by a high choline content, which is not observed in other bacterial species. Not surprisingly, pneumococcal-specific lysins have been shown to utilize a choline binding domain for recognition and attachment to the pneumococcal cell wall. However, this is the only known receptor described for a lysin and is exclusive to pneumococcal lysins. Therefore, we have been developing chemical and HPLC methods to extract, fractionate, and characterize (composition and linkage analysis) the components of cell walls from several bacterial species in order to map these epitopes against a repertoire of lysins.

2. Lysin structure/function/mechanism. To date, there are only two known crystal structures of lysins from Gram-positive infecting bacteriophage and neither one is of a full-length molecule. Because these enzymes contain at least two distinct domains (catalytic and CBD), they are often attached through a flexible linker sequence which makes crystallography of the full length protein challenging. In order to stabilize the full-length structure of lysins, we are genetically modulating the linker length as well as attempting to secure the overall structure with antibodies or crosslinkers. In an alternate approach, we are investigating directed evolution techniques to map critical residues for binding or activity. Based on these approaches, we plan to make mutations, insertions, and deletions to better understand lysin function and mechanism.

3. Lysin kinetics. Lysins presumably act in a multi-step kinetic manner. First, the CBD binds to the target receptor or epitope on the cell wall through ionic interactions, and then the catalytic domain cleaves the labile bond within the peptidoglycan. The catalytic domain is released, and then the CBD dissociates from its receptor. Unfortunately, simple synthetic substrates mimicking only the peptidoglycan typically do not work for lysins due to the required interaction between catalytic domain and CBD. Likewise, custom substrates that incorporate CBD binding epitopes do not exist due to the lack of known CBD receptors. Therefore, no detailed study of lysin kinetics has been carried out to date. We are using a combination of isothermal titration calorimetry (ITC) and surface plasmon resonance (SPR) in conjunction with active site mutants and deletion constructs to measure lysin binding constants and catalytic efficiency.

4. Lysin diagnostics. Direct administration of lysins to susceptible bacteria results in lysis, liberating cytoplamic ATP, which can be measured in a luciferin/luciferase based assay. Using this premise, we can discern as few as 10 streptococcal colony forming units in a background of non-target organisms using a commercial hand-held luminometer. In an alternative diagnostic approach, we are using fluorogenic tags crosslinked to CBDs in order to visualize and quantitate target organisms as well as the ability to employ flow cytometry for bacterial detection. Finally, we have started to crosslink quantum dots to CBDs, which offers perhaps the greatest sensitivity in diagnostic research.