A simple, but very reproducible, method for measuring the relative specificity of ribulosebisphosphate carboxylase-oxygenase for CO2, as opposed to O2, is described. The method uses [1-14C]ribulose bisphosphate as substrate and combines the advantages of supplying both gaseous substrates from the gas phase with HPLC separation of the labelled products. Volumetric or gravimetric accuracy is not required at any stage of the procedure and variations in ionic strength and pH have little effect on the measurements. This leads to excellent reproducibility without the need for normalisation. The average standard deviation was 1.3% of the measured CO2/O2 specificity. Use of very low ribulose bisphosphate concentrations ensures that the gaseous substrates cannot be depleted appreciably during the reaction and enhances the attractiveness of the procedure for measurements with crippled mutant enzymes. The procedure's ability to resolve small differences in relative specificity is demonstrated by its easy detection of the 5% increase in specificity that accompanies substitution of four residues at positions 338-341 of the cyanobacterial large subunit with the analogous higher-plant residues. This resolving power is essential for detecting small differences in the specificities of higher-plant ribulosebisphosphate carboxylases which may be the signature of continuing evolutionary refinement.

Life depends on a single enzyme, D-ribulose-1,5-bisphosphate carboxylase-oxygenase (Rubisco), for the acquisition of essentially all of its carbon. Applying Darwinian principles, one would expect this enzyme to have been rigorously selected for speed and accuracy, and it is a surprise to discover that, even in its most highly developed forms, it is both slow and confused. This review looks for clues about the causes of Rubisco's slow evolutionary refinement in its complex catalytic chemistry and in its tendency to catalyse abortive side reactions. We assess the possibilities for improving Rubisco artificially, either by random mutagenesis or by rational design, and imagine the consequences of an improved Rubisco for plant productivity and the global ecosystem.