Directed -in vitro- evolution of Precambrian and extant Rubiscos

Abstract

Rubisco is an ancient, catalytically conserved yet slow enzyme, which plays a central role in the biosphere's carbon cycle. The design of Rubiscos to increase agricultural productivity has hitherto relied on the use of in vivo selection systems, precluding the exploration of biochemical traits that are not wired to cell survival. We present a directed -in vitro- evolution platform that extracts the enzyme from its biological context to provide a new avenue for Rubisco engineering. Precambrian and extant form II Rubiscos were subjected to an ensemble of directed evolution strategies aimed at improving thermostability. The most recent ancestor of proteobacteria -dating back 2.4 billion years- was uniquely tolerant to mutagenic loading. Adaptive evolution, focused evolution and genetic drift revealed a panel of thermostable mutants, some deviating from the characteristic trade-offs in CO₂-fixing speed and specificity. Our findings provide a novel approach for identifying Rubisco variants with improved catalytic evolution potential.

Figure 1

Dual HTS assay. (a) NADH-linked enzymatic assay of Rubisco activity. The reaction monitors the rate 3 PGA production from the carboxylation of Ribulose-1,5-bisphosphate (RuBP) by Rubisco. PGK, phosphoglycerokinase; BPGA, 1,3-bisphosphoglycerate; GAPDH, glyceraldehyde-3P dehydrogenase; GAP, glyceraldehyde-3P; TPI, triose-P isomerase; DHAP, dihydroxyacetone-P; GPDH, glycerol-3P dehydrogenase; G3P, glycerol-3P. (b) HPLC-ELSD chromatogram for the detection of 3PGA production/RuBP depletion by Rubisco (+and - represent cationic and anionic compounds in the sample, respectively) and (c), the variation in the concentration of these two compounds during the reaction as measured by HPLC-ELSD. (d) Fitness landscapes for RubRr mutant libraries that are adjusted by varying the MnCl₂ concentrations (in the case of Taq libraries) or by the amount of the DNA template (for mutazyme libraries). The relative RubRr activity of the clones is plotted in descending order and the dashed line shows the activity of the parental type in the assay: red circles, library I (mutazyme, 750 ng DNA template); blue circles, library II (Taq, 0.02 mM MnCl₂); black circles, library III (mutazyme, 100 ng DNA template); white circles, library IV (Taq, 0.03 mM MnCl₂); green circles, library V (Taq, 0.05 mM MnCl₂). (e) Directed evolution landscape for thermostability and activity relative to the parental RubRr enzyme (dashed lines). The mutants selected for further re-screening are contained in the blue rectangle (using thresholds of 0.5- and 1.3-fold over RubRr activity and thermostability, respectively).

Figure 2

Directed -in vitro- evolution campaign. (a) General overview of the directed evolution campaigns carried out on Precambrian and modern (RubRr) form II Rubisco. (b) Neutral mutations identified in the amino acid sequence of RubRr (left) and mapped to its crystal structure (right). Mutations identified once are in yellow and those found in several clones are in red.

Figure 3

Phylogenetic tree generated from the amino acid sequences of 47 different form II Rubisco sequences. The colored areas represent the different bacterial phyla and classes as follows: Firmicutes (Green); α and δ proteobacteria (Blue); γ and β proteobacteria (orange). The brown squares highlight the nodes whose sequences were selected for resurrection and the reference R. rubrum Rubisco (RubRr) sequence is framed. See also Supplementary Figs 10–12.

Figure 4

Comparison between modern RubRr and the ancestral -resurrected- Rubisco. (a) Amino acid differences (in surface mode) in MRProFir (green), MRPro (yellow) and MRβ/γPro (red) relative to RubRr. (b) 3D motifs in RubRr and ancestral Rubisco. (c) Multiple sequence alignment of RubRr and ancestral Rubiscos. Expected motifs are highlighted in colors. (pink, beta sheets; grey, alpha helices; blue, loops). The asterisk indicates positions with a fully conserved residue; colon indicates conservation between groups of strongly similar properties; period indicates conservation between groups of weakly similar properties. (d) Frequency distribution of the pairwise sequence distances for form II Rubisco sequences obtained from a database search using the RubRr as the query. The sequence distances are represented on the X axis and the number of sequences is reflected on the Y axis. (e) Number of amino acids different from RubRr and billion years old between the ancestral nodes. The multiple sequence alignment was performed with Clustal Omega V1.2.4 available at https://www.ebi.ac.uk/Tools/msa/clustalo/. The structures were modeled using the PDB code 9RUB and the Phyre2 server (Protein Homology/analogY Recognition Engine V 2.0: Kelley and Sternberg, 2009) available at www.sbg.bio.ic.ac.uk/phyre2.

Figure 5

Expression of RubRr, ancestral Rubisco and their mutant offspring in E. coli. (a) Rubisco content as a percentage of the total soluble E. coli protein (TSP, %) determined by ¹⁴C-CABP binding and confirmed by (b) SDS-PAGE. Lane 1, RubRr; Lane 2, clone 11; Lane 3, clone 27; Lane 4, clone 25; Lane 5, clone 9; Lane 6, MRPro; Lane 7, clone B2; Lane 8, clone B3; Lane 9, clone B9; Lane 10, MW protein standard. The Rubisco RbcL subunit corresponds to the prominent 50 kDa band. Refer to Fig. 2a for the origins and mutations of each clone.