Evolutionary escape from the Rubisco activase requirement

Abstract

Ribulose 1,5-bisphosphate carboxylase/oxygenase (Rubisco) catalyzes the rate-limiting carbon fixation step in photosynthesis. Despite its essential role, Rubisco is inefficient and possesses sluggish kinetics. In addition, Rubisco from many species (including plants) are prone to inhibition by sugar phosphates and require assistance from Rubisco activase to facilitate inhibitor release. Thus, Rubisco has been a target for improvement using directed evolution. Here, we performed directed evolution of Rubisco from Rhodobacter sphaeroides and Acidithiobacillus ferrooxidans using a Rubisco-dependent Escherichia coli system (RDE). Similar to plant Rubisco, R. sphaeroides and A. ferrooxidans Rubiscos form inhibited complexes with RuBP and require an activase. In addition, Rs-Rubisco was also evolved in presence of its activase using Rubisco and activase dependent E. coli (RADE) system where the presence of functional and compatible Rubisco and activase pair are required for survival. Rubisco variants with reoccurring mutated residues were obtained from screening and 22 of these were biochemically characterized using radioactive and spectrophotometric Rubisco assays. Many variants isolated during RDE screening no longer form stable inhibited complexes with RuBP and exhibited impaired kinetics. However, the CO2/O2 specificity of these variants was largely unchanged compared to the wild type. In contrast, a Rubisco variant obtained from RADE screening still formed stable inhibited complexes while having a modest reduction in kinetics. Structural analysis of the variants with increased rates of inhibitor release revealed that common mutations were clustered around elements involved in substrate binding (60s loop) and active site closure (C-terminus). The observed mutations likely altered the local conformation such that the active site becomes less rigid and this facilitated rapid inhibitor release. Simultaneously, plant/cyanobacteria (Oryza sativa/ Synechococcus) chimeric Rubiscos were successfully produced in E. coli. Unlike plant Rubisco, cyanobacterial Rubisco does not form inhibited complexes with sugar phosphates and the organisms do not generally encode an activase. One chimeric Rubisco exhibited a catalytic fallover effect and was also able to form inhibited complexes with RuBP. In addition, the inhibited chimeric Rubisco could be recognized and remodeled by plant activase. The directed evolution of this plant-like chimeric Rubisco in the RDE screening system yielded variants that no longer form stable inhibited complexes. Therefore, relief from the activase dependency can easily be discovered using directed evolution methodologies. Using a combination of rational and evolutionary methods to modify Rubisco, we obtained insights into the underlying principles governing the shortcomings of the enzyme and its protein-protein interactions with accessory proteins.