Plant defensins are a diverse family of small, disulphide-rich, innate immunity proteins. They are part of a larger superfamily that includes fungal and invertebrate antimicrobial proteins, as well as plant signalling proteins and arthropod toxins. They also achieve these varied functions via an array of mechanisms including membrane disruption, enzyme and ion channel inhibition, and receptor binding. However, the relationship between sequence, structure and function is currently poorly understood and orders of magnitude more sequences are known than have been characterised.
Design and engineering of defensins has been hampered by our limited understanding of the relationship between their sequence, structure and function. We address this shortfall by mining existing diversity and using evolution to inform engineering. In order to do this we have had to develop new methods of cysteine-rich protein sequence alignment and analysis, as traditional phylogenetic methods are inadequate for such divergent sequence. We adapt some of the principals of peptide quantities structure-activity relationship analysis (QSAR) to proteins with more complex and variable structures and mechanisms. Multivariate analysis of protein sequence space allows us to categorise defensins into naturally occurring clusters.
We are investigating both the clusters in defensin sequence space, as well as the voids between them. To map the most poorly described regions of the sequence space we are biochemically characterising the most divergent plant defensins. We have found that defensins with even the most diverse sequence properties can have antimicrobial functions. We are currently investigating whether this variation represents neofunctionalisation to different targets or mechanism of action, or whether it is reflective of extreme genetic drift for the same function. Complementing this, we are investigating the densely-populated sequence space clusters by designing sequences that represent archetypes of that cluster. These archetypes may be a viable alternative to the highly successful technique of ancestral sequence reconstruction for engineering increased activity, promiscuity and stability.