Proteins play crucial roles as molecular machines in various biological processes. While they are abundant, naturally occurring proteins are only a small portion of the vast range of potential sequences. De novo proteins, which are synthesized in the lab and haven't been shaped by natural selection, provide a unique opportunity to explore new protein structures and functions. These proteins could have wide-ranging applications, from medical treatments to material engineering. One such protein, Syn-F4, was the first de novo enzyme shown to be active both in vitro and in vivo, and it successfully rescues auxotrophic bacteria. Rescuer 6 is another de novo protein, created by duplicating, fusing, and diversifying Syn-F4 according to Dayhoff’s hypothesis. Earlier studies indicated that Rescuer 6 exhibited greater biological activity than Syn-F4. However, this thesis examines the in vitro enzymatic activity of Rescuer 6 and finds it to be less effective than Syn-F4. To enhance its activity, directed evolution was used, and through random mutagenesis and selection, several truncated proteins with better rescuing abilities were identified. Three of these proteins were chosen for further analysis of their catalytic functions. Two of these proteins were found to be more efficient enzymes than both Rescuer 6 and Syn-F4. These experiments demonstrate that just a few rounds of directed evolution can lead to de novo enzymes with improved biological and catalytic functions. The fact that the best-performing proteins were truncated suggests that gene or protein duplication can serve as an intermediate step toward developing smaller, more efficient proteins.

The field of \textit{de novo} protein design carries immense potential for creating novel proteins guided by the principles of nature. Despite the challenges in designing proteins, recent research has leveraged the driving forces of protein folding, particularly the burial of hydrophobic residues, to computationally design new complex topologies using binary patterning.$\mathrm{<mrow\; data-mjx-texclass="ORD">1</mrow>}$ This work studies the structural tolerance of \textit{de novo} proteins to mutations within their hydrophobic cores by investigating variants from a combinatorial library. Differential scanning calorimetry (DSC), a technique that measures the melting temperature and enthalpy of a protein unfolding event, demonstrated high thermal stability of two mutant proteins. Hydrogen-deuterium exchange measured by mass spectrometry (HDX-MS) analyzed the conformational behavior of these mutants. While the mutants had varying degrees of isotopic exchange, only a fraction of the available hydrogens underwent exchange in the three most promising mutants. A consistent trend can be drawn from the results of both techniques: the most thermodynamically stable mutant of DSC incorporated the fewest deuterium atoms in HDX-MS experiments, suggesting a correlation between thermal and conformational stability. These findings demonstrate that \textit{de novo} proteins of a combinatorial library can be successfully designed with novel backbone structures, tolerating variations within their hydrophobic cores. This presents a promising avenue for true \textit{de novo} design, achieving complexity and stability approaching that of natural proteins.

One of the goals of synthetic biology is to isolate novel protein sequences that support essential biological functions. By using both rational and combinatorial design, the Hecht group has successfully discovered multiple synthetic sequences capable of sustaining life. Most of these functional de novo sequences form stable, well-folded protein structures, but the Rescuer 4 protein enables the survival of auxotrophic ∆metC E. coli cells in minimal media despite being disordered and insoluble. Preliminary experimental results suggested that the Rescuer 4 protein may rescue ∆metC through gene regulation, by co precipitating with MetJ, a transcriptional repressor protein. In order to investigate the rescue mechanism of Rescuer 4, multiple variants of the sequence were created through site-directed mutagenesis. These sequences were then transformed into ∆metC cells for life-or-death screening and growth analysis. Most variants exhibited reduced or abolished rescue function, verifying that the protein—rather than the RNA—is responsible for the rescue of ∆metC and confirming the interaction between MetJ and Rescuer 4 plays a key part in the mechanism. In addition, a variant sequence with enhanced rescue efficiency was discovered. Proteomics analysis of this strain points to an upregulation of chaperone proteins. These findings provide further insight into the mechanism of Rescuer 4 and demonstrate how de novo proteins can interact with natural biomolecules to support life-sustaining functions.