Benchmark Analysis of Native and Artificial NAD⁺-Dependent Enzymes Generated by a Sequence-Based Design Method with or without Phylogenetic Data

Abstract

The expansion of protein sequence databases has enabled us to design artificial proteins by sequence-based design methods, such as full-consensus design (FCD) and ancestral-sequence reconstruction (ASR). Artificial proteins with enhanced activity levels compared with native ones can potentially be generated by such methods, but successful design is rare because preparing a sequence library by curating the database and selecting a method is difficult. Utilizing a curated library prepared by reducing conservation energies, we successfully designed two artificial l-threonine 3-dehydrogenases (SDR-TDH) with higher activity levels than native SDR-TDH, FcTDH-N1, and AncTDH, using FCD and ASR, respectively. The artificial SDR-TDHs had excellent thermal stability and NAD recognition compared to native SDR-TDH from Cupriavidus necator (CnTDH); the melting temperatures of FcTDH-N1 and AncTDH were about 10 and 5 °C higher than that of CnTDH, respectively, and the dissociation constants toward NAD of FcTDH-N1 and AncTDH were 2- and 7-fold lower than that of CnTDH, respectively. Enzymatic efficiency of the artificial SDR-TDHs were comparable to that of CnTDH. Crystal structures of FcTDH-N1 and AncTDH were determined at 2.8 and 2.1 Å resolution, respectively. Structural and MD simulation analysis of the SDR-TDHs indicated that only the flexibility at specific regions was changed, suggesting that multiple mutations introduced in the artificial SDR-TDHs altered their flexibility and thereby affected their enzymatic properties. Benchmark analysis of the SDR-TDHs indicated that both FCD and ASR can generate highly functional proteins if a curated library is prepared appropriately.

Catalytic mechanism of ancestral L-lysine oxidase assigned by sequence data mining

“…In this study, we attempted to assign new LAAOs from the database utilizing one of the previously designed AncLAAO sequences (28) as a template. Through combinational approaches using paralog searches, the application of previously reported original protein sequence selection methods (27,(34)(35)(36), and biochemical assays, we newly characterized LAAOs exhibiting high specificity toward L-Lys. These included L-Lysine oxidase (LLysO) from the Chryseobacterium, Flavobacterium, and Pedobacter species.…”

Following the Evolutionary Track of a Highly Specific l -Arginine Oxidase by Reconstruction and Biochemical Analysis of Ancestral and Native Enzymes

“…The "key residues" can be assigned utilizing only primary sequence data and can reduce the sum of the conservation energy (E c ), which was defined by Jäckel et al (28). A full consensus protein generated by utilizing a curated library exhibited high solubility, thermal stability, and reactivity (27). The merit of curation is that sequences bearing low identity compared with the target protein can be omitted by using key residues (27).…”

“…A full consensus protein generated by utilizing a curated library exhibited high solubility, thermal stability, and reactivity (27). The merit of curation is that sequences bearing low identity compared with the target protein can be omitted by using key residues (27). Excluding these sequences can improve the accuracy of multiple-sequence alignment (MSA), which is crucial to the successful design of full-length consensus proteins (29).…”

Faculty Opinions recommendation of AirID, a novel proximity biotinylation enzyme, for analysis of protein-protein interactions.

“…A comprehensive and curated sequence library was prepared querying the Blastp web server and using a custom Python script (Source code 1), which exhibited more than 30% sequence identity with E. coli BirA (EU08004.1). Next, further curation approaches were applied to the library as shown in previous studies (Nakano et al, 2018;Nakano et al, 2019). The procedure consists of the following steps: 1) preparation of sequence pairs consisting of one of the submitted BirA sequences and one sequence (total 1275 genes) in the library, 2) sequence alignment of the all pairs, and 3) selection of sequences bearing 'key residues' Figure 1.…”

“…Evolutionary protein engineering using metagenome data have recently been used to improve enzymes (Nakano and Asano, 2015;Nakano et al, 2018;Nakano et al, 2019). Here, we newly designed five ancestral BirA enzymes using an ancestral enzyme reconstruction algorithm and a large genome dataset.…”