Probing remote residues important for catalysis in Escherichia coli ornithine transcarbamoylase

Revisiting the Roles of Catalytic Residues in Human Ornithine Transcarbamylase

Human ornithine transcarbamylase (hOTC) is a mitochondrial transferase protein involved in the urea cycle and is crucial for the conversion of toxic ammonia to urea. Structural analysis coupled with kinetic studies of Escherichia coli, rat, bovine, and other transferase proteins has identified residues that play key roles in substrate recognition and conformational changes but has not provided direct evidence for all of the active residues involved in OTC function. Here, computational methods were used to predict the likely active residues of hOTC; the function of these residues was then probed with site-directed mutagenesis and biochemical characterization. This process identified previously reported active residues, as well as distal residues that contribute to activity. Mutation of active site residue D263 resulted in a substantial loss of activity without a decrease in protein stability, suggesting a key catalytic role for this residue. Mutation of predicted second-layer residues H302, K307, and E310 resulted in significant decreases in enzymatic activity relative to that of wild-type (WT) hOTC with respect to l-ornithine. The mutation of fourth-layer residue H107 to produce the hOTC H107N variant resulted in a 66-fold decrease in catalytic efficiency relative to that of WT hOTC with respect to carbamoyl phosphate and a substantial loss of thermal stability. Further investigation identified H107 and to a lesser extent E98Q as key residues involved in maintaining the hOTC quaternary structure. This work biochemically demonstrates the importance of D263 in hOTC catalytic activity and shows that residues remote from the active site also play key roles in activity.

MAHOMES II: A webserver for predicting if a metal binding site is enzymatic

Recent advances have enabled high-quality computationally generated structures for proteins with no solved crystal structures. However, protein function data remains largely limited to experimental methods and homology mapping. Since structure determines function, it is natural that methods capable of using computationally generated structures for functional annotations need to be advanced. Our laboratory recently developed a method to distinguish between metalloenzyme and nonenzyme sites. Here we report improvements to this method by upgrading our physicochemical features to alleviate the need for structures with sub-angstrom precision and using machine learning to reduce training data labeling error. Our improved classifier identifies protein bound metal sites as enzymatic or nonenzymatic with 94% precision and 92% recall. We demonstrate that both adjustments increased predictive performance and reliability on sites with sub-angstrom variations. We constructed a set of predicted metalloprotein structures with no solved crystal structures and no detectable homology to our training data. Our model had an accuracy of 90%-97.5% depending on the quality of the predicted structures included in our test. Finally, we found the physicochemical trends that drove this model's successful performance were local protein density, second shell ionizable residue burial, and the pocket's accessibility to the site. We anticipate that our model's ability to correctly identify catalytic metal sites could enable identification of new enzymatic mechanisms and improve de novo metalloenzyme design success rates.

MAHOMES II: A webserver for predicting if a metal binding site is enzymatic

Recent advances have enabled high-quality computationally generated structures for proteins with no solved crystal structures. However, protein function data remains largely limited to experimental methods and homology mapping. Since structure determines function, it is natural that methods capable of using computationally generated structures for functional annotations need to be advanced. Our laboratory recently developed a method to distinguish between metalloenzyme and non-enzyme sites. Here we report improvements to this method by upgrading our physicochemical features to alleviate the need for structures with sub-angstrom precision and using machine learning to reduce training data labeling error. Our improved classifier identifies protein bound metal sites as enzymatic or non-enzymatic with 94% precision and 92% recall. We demonstrate that both adjustments increased predictive performance and reliability on sites with sub-angstrom variations. We constructed a set of predicted metalloprotein structures with no solved crystal structures and no detectable homology to our training data. Our model had an accuracy of 90 - 97.5% depending on the quality of the predicted structures included in our test. Finally, we found the physicochemical trends that drove this model's successful performance were local protein density, second shell ionizable residue burial, and the pocket's accessibility to the site. We anticipate that our model's ability to correctly identify catalytic metal sites could enable identification of new enzymatic mechanisms and improve de novo metalloenzyme design success rates.

Significance statement

Identification of enzyme active sites on proteins with unsolved crystallographic structures can accelerate discovery of novel biochemical reactions, which can impact healthcare, industrial processes, and environmental remediation. Our lab has developed an ML tool for predicting sites on computationally generated protein structures as enzymatic and non-enzymatic. We have made our tool available on a webserver, allowing the scientific community to rapidly search previously unknown protein function space.

Enhancing the Catalytic Activity of Type II L-Asparaginase from Bacillus licheniformis through Semi-Rational Design

Low catalytic activity is a key factor limiting the widespread application of type II L-asparaginase (ASNase) in the food and pharmaceutical industries. In this study, smart libraries were constructed by semi-rational design to improve the catalytic activity of type II ASNase from Bacillus licheniformis. Mutants with greatly enhanced catalytic efficiency were screened by saturation mutations and combinatorial mutations. A quintuple mutant ILRAC was ultimately obtained with specific activity of 841.62 IU/mg and kcat/Km of 537.15 min−1·mM−1, which were 4.24-fold and 6.32-fold more than those of wild-type ASNase. The highest specific activity and kcat/Km were firstly reported in type II ASNase from Bacillus licheniformis. Additionally, enhanced pH stability and superior thermostability were both achieved in mutant ILRAC. Meanwhile, structural alignment and molecular dynamic simulation demonstrated that high structure stability and strong substrate binding were beneficial for the improved thermal stability and enzymatic activity of mutant ILRAC. This is the first time that enzymatic activity of type II ASNase from Bacillus licheniformis has been enhanced by the semi-rational approach, and results provide new insights into enzymatic modification of L-asparaginase for industrial applications.

Analysis of electrostatic coupling throughout the laboratory evolution of a designed retroaldolase

The roles of local interactions in the laboratory evolution of a highly active, computationally designed retroaldolase (RA) are examined. Partial Order Optimum Likelihood (POOL) is used to identify catalytically important amino acid interactions in several RA95 enzyme variants. The series RA95.5, RA95.5–5, RA95.5–8, and RA95.5–8F, representing progress along an evolutionary trajectory with increasing activity, is examined. Computed measures of coupling between charged states of residues show that, as evolution proceeds and higher activities are achieved, electrostatic coupling between the biochemically active amino acids and other residues is increased. In silico residue scanning suggests multiple coupling partners for the catalytic lysine K83. The effects of two predicted partners, Y51 and E85, are tested using site‐directed mutagenesis and kinetic analysis of the variants Y51F and E85Q. The Y51F variants show decreases in k_cat relative to wild type, with the greatest losses observed for the more evolved constructs; they also exhibit significant decreases in k_cat/K_M across the series. Only modest decreases in k_cat/K_M are observed for the E85Q variants with little effect on k_cat. Computed metrics of the degree of coupling between protonation states rise significantly as evolution proceeds and catalytic turnover rate increases. Specifically, the charge state of the catalytic lysine K83 becomes more strongly coupled to those of other amino acids as the enzyme evolves to a better catalyst.