Evolution-guided and Structure-based Approaches for Protein Engineering

Proteins enable the existence of life as we know it by performing diverse functions in all organisms, from structural support to biocatalysis. Their powerful properties have been harnessed by humans for a variety of biotechnological applications, in ecologically and economically important sectors ranging from food and biofuel production, to biomedicine and bioremediation.

In Nature, proteins have evolved their sequences and therefore structures and functions in natural environments, which are often quite different than the desired use of proteins for biotechnological purposes. Therefore, engineering is frequently necessary to generate useful properties relevant to industry.

To achieve this, directed evolution has been harnessed to generate and identify useful protein variants in a laboratory setting. Furthermore, solving the three-dimensional structure of a protein enables precise molecular information which can be useful for understanding functional implications. The work presented in this thesis approaches protein engineering from both the evolution-guided and structure-based approaches.

In the first section, evolution-guided approaches are explored, with a particular emphasis on optimizing transcription-factor based biosensors for in vivo screening applications. In the first study, a transcriptional activator BenM, which responds to muconic acid, a precursor for plastics, is targeted in the budding yeast Saccharomyces cerevisiae.

Through random mutagenesis of the effector-binding domain followed by various sorting regimes using Fluorescence-activated cell sorting (FACS), the biosensor response curve was engineered to optimize the performance for various applications. BenM variants were generated that exhibit increased dynamic range, a shift in operational range, an inverse of function, and a change in ligand specificity, thus exemplifying the flexibility of transcription-factors as biosensors and demonstrating the power of the FACS-based method.

In the second study, this established method is applied to engineering a transcriptional repressor, QdoR, in Escherichia coli. QdoR responds to flavonoids such as quercetin which is of interest due to potential anti-inflammatory and antioxidant properties, in addition to being an interesting target for investigating natural product glycosylation.

Through random mutagenesis of the QdoR-based biosensor, a variant with increased dynamic range was identified. Mutations in the promoter as well as the protein contributed to this altered response, which demonstrate the importance of investigating expression level and oligomerization for optimizing biosensor response.

In the second section, structure-based approaches are explored, with a particular emphasis on X-ray crystallography for uncovering structure-function relationships. In the first study, enzymes from the bacteria Bacteroides eggerthii are investigated to gain a better understanding of how alginate is digested in the human gut.

The crystal structure of an alginate lyase from a nearly unknown protein family is solved and compared to two homologs. This enzyme, BeKdgF, is solved with either a calcium atom or a zinc atom in a conserved metal-binding site. However, these atoms refine to lower occupancy and exhibit higher local B-factors, along with the coordinating side chains, than this site in the homologs, which is in alignment with the biochemical characterization that indicates a lack of metal dependency for activity, in contrast to the metal-dependent homologs.

In the second study, a recently identified glycoside hydrolase from the human gut bacteria Akkermansia muciniphila is investigated to understand the features than enable comparable α-retaining and β-inverting N-cetylgalactosaminidase (GalNAc) activities. The ligand-complex crystal structure is solved, which reveals a flexible loop that positions a histidine within hydrogen-bonding distance of the bound GalNAc.

This polar contact revealed the histidine as the probable acid/base catalyst, previously unidentified in glycoside hydrolase family 109, which is further confirmed through biochemical characterization and supported by molecular dynamics simulations. Taken together, these four studies exemplify the importance of approaching protein engineering from both evolution-guided and structure-based approaches, in order to modify protein properties and gain a better understanding of structure-function relationships.

Combining these approaches provides a holistic workflow for directed evolution endeavors and enables expedited engineering of novel functionalities.