Computational design of proteins and protein crystals

Protein design is motivated by the desire to study, understand, and exploit the versatile structures and functions capable with proteins. Nature leverages the physico-chemical properties of the amino acids to arrive at highly functional sequences that spontaneously fold, where structural and functional properties are fine-tuned during the course of evolution. Proteins comprise tens to thousands of amino acids, and backbone and side-chain degrees of freedom result in an immense number of possible configurations for a single sequence. Enumerating these myriad sequences is prohibitively expensive even with modern computational power, thus, protein designers have sought methods and/or approximations to reduce the computational burden. One such approximation is the inverse design\u27\u27 approach. In essence, rather than attempting to design a sequence outright that will fold into a particular structure, the designer specifies a single scaffold structure having the particular fold of interest. A sequence is then designed with the goal of stabilizing that scaffold. Removing the scaffold flexibility is often seen as an overly-stringent constraint, leading protein designers to seek new methodologies that incorporate some scaffold flexibility. Throughout this work we take advantage of both the inverted approach and a unique sequence design methodology to arrive at energy landscapes that provide a birds-eye\u27\u27 view of the relationship between sequence and structure. These landscapes are particularly useful in helping to identify structures that are designable\u27\u27 and thus stand a better chance of being experimentally realized. Central to the approach is the recent increase in computational power, allowing us to design entire ensembles (many thousands) of scaffold structures with closely related geometries providing for many possible sequences that fit our design objectives. This is in marked contrast to other design efforts where only a single structure is identified consistent with a single or handful of scaffold structures, providing for little or no recourse when a design fails in the laboratory. Here, we focus on two types of sequence-structure relationships: (i) that of coiled-coil secondary structure where we take advantage of the Crick parameterization to design ensembles of coiled-coil dimers that reversibly fold in the presence of zinc and (ii) that of protein-protein interfaces and quaternary structures, were we optimize the protein sequence in a periodic crystalline\u27\u27 environment consistent with ensembles of unit cells with varying geometries. The sequence-structure approach has proven to be invaluable; we\u27ve successfully designed the first ever protein crystal and have taken that success to the next level by proposing designs that bind and orient non-biological non-linear optical chromophores in high density arrangements. We\u27ve also designed a coiled-coil dimer that folds reversibly dependent on the presence of coordinating zinc ions and has particular advantages for reversible cell-tethering, fluorescence labeling and nanoparticle assembly