Directed Evolution of a Surface-Displayed Artificial Allylic Deallylase Relying on a GFP Reporter Protein

Alain Baiyoumy; Jaicy Vallapurackal; Fabian Schwizer; Tillmann Heinisch; Tsvetan Kardashliev; Martin Held; Sven Panke; Thomas R Ward

doi:10.1021/acscatal.1c02405

Directed Evolution of a Surface-Displayed Artificial Allylic Deallylase Relying on a GFP Reporter Protein

ACS Catal. 2021 Sep 3;11(17):10705-10712. doi: 10.1021/acscatal.1c02405. Epub 2021 Aug 12.

Authors

Alain Baiyoumy^{1

2}, Jaicy Vallapurackal^{1

2}, Fabian Schwizer¹, Tillmann Heinisch¹, Tsvetan Kardashliev³, Martin Held³, Sven Panke^{3

2}, Thomas R Ward^{1

2}

Affiliations

¹ Department of Chemistry, University of Basel, Mattenstrasse 24a, 4058 Basel, Switzerland.
² Molecular Systems Engineering, National Competence Center in Research (NCCR), 4058 Basel, Switzerland.
³ ETH Zürich, D-BSSE, Mattenstrasse 26, 4058 Basel, Switzerland.

Abstract

Artificial metalloenzymes (ArMs) combine characteristics of both homogeneous catalysts and enzymes. Merging abiotic and biotic features allows for the implementation of new-to-nature reactions in living organisms. Here, we present the directed evolution of an artificial metalloenzyme based on Escherichia coli surface-displayed streptavidin (Sav^SD hereafter). Through the binding of a ruthenium-pianostool cofactor to Sav^SD, an artificial allylic deallylase (ADAse hereafter) is assembled, which displays catalytic activity toward the deprotection of alloc-protected 3-hydroxyaniline. The uncaged aminophenol acts as a gene switch and triggers the overexpression of a fluorescent green fluorescent protein (GFP) reporter protein. This straightforward readout of ADAse activity allowed the simultaneous saturation mutagenesis of two amino acid residues in Sav near the ruthenium cofactor, expediting the screening of 2762 individual clones. A 1.7-fold increase of in vivo activity was observed for Sav^SD S112T-K121G compared to the wild-type Sav^SD (wt-Sav^SD). Finally, the best performing Sav isoforms were purified and tested in vitro (Sav^PP hereafter). For Sav^PP S112M-K121A, a total turnover number of 372 was achieved, corresponding to a 5.9-fold increase vs wt-Sav^PP. To analyze the marked difference in activity observed between the surface-displayed and purified ArMs, the oligomeric state of Sav^SD was determined. For this purpose, crosslinking experiments of E. coli cells overexpressing Sav^SD were carried out, followed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and Western blot. The data suggest that Sav^SD is most likely displayed as a monomer on the surface of E. coli. We hypothesize that the difference between the in vivo and in vitro screening results may reflect the difference in the oligomeric state of Sav^SD vs soluble Sav^PP (monomeric vs tetrameric). Accordingly, care should be applied when evolving oligomeric proteins using E. coli surface display.