Directed Laboratory Evolution of Biomineralizing Enzymes

Biological systems have achieved the capacity to synthesize sophisticated mineral components under ambient conditions via natural genetic evolution. Identification of the silicateins, enzymes that exhibit unique catalytic and templating activities in silica biogenesis, has opened the possibility of applying enzyme engineering methodologies to develop new mineralizing systems. This dissertation describes a novel strategy to recapitulate, in a laboratory setting, the essential evolutionary processes responsible for the development of biosilica-forming silicatein systems, and to direct this evolution to engineer new mineral-synthesizing enzymes with specifically tailored activities. Millions of variant silicatein genes are produced by DNA shuffling and mutagenesis of parent silicateins identified from marine sponge biosilica. The gene library is either linked to polystyrene mircobeads or prepared as a dilute suspension, and then added to bacterial extracts in water-in-oil emulsions for in vitro protein expression. Reacting the droplet-partitioned proteins with mineral precursors creates millions of unique artificial biomineralization vesicles. High-throughput screening for mineralizing activity is conducted using fluorescence-activated cell sorting, permitting beads or vesicles with functional enzyme variants to be identified and sorted, either based on light-scattering signals indicating mineral-oxide growth from bead surfaces, or based on photoluminescent signals indicating quantum dot synthesis within oil-membrane-bound vesicles. Based on the applied screening conditions, protein-silicate and protein-titania inorganic composites were genetically evolved in vitro to exhibit dispersed nanoparticlate morphologies and to synthesize unique crystalline oxide products at room temperature. Further, silicatein genes coding for CdS nanoparticle synthesis (selected by their characteristic photoluminescence) were isolated, thus demonstrating genetic tuning of mineral properties. To illustrate how future improvements may be made to these in vitro evolution efforts, ultra-high-throughput sequencing is harnessed to accelerate discovery of specific single-stranded DNA sequences with high ZnO-binding affinity. The overall strategies developed here should be generally applicable for generating novel mineral forming enzymes from a variety of recombinant parent enzymes. The results of this work will help elucidate native biological mechanisms of mineralization, provide new pseudo-synthetic models for extending biomimetic materials engineering strategies, and establish substantial groundwork towards the future development of gene-based synthetic materials.