Abstract: As the most abundant protein on Earth, ribulose-1,5-bisphosphate carboxylase/oxygenase (RuBisCO) is the key enzyme in carbon fixation of photosynthesis in all green plants. It is hexadecimal polymer, consisting of two kinds of subunits, large and small ones, which are encoded by different genomes. The large subunit (LSU) is encoded by the chloroplast genome; the small one by the nuclear genome. The function of RuBisCO relies on the two subunits, and only when they are combined to be hexadecimal, can it carry out its whole function. However, how the LSU and small subunits evolved, and if they coevolved, has remained mysterious. In this study, based on data in 21 species, phylogenies were analyzed on two subunits, respectively, and a mirror tree was reconstructed to reveal phylogenetic similarity. Furthermore, a comparative analysis method, using the codon usage bias and correlation calculation for RuBisCO subunits, was carried out. The relative synonymous codon usage (RSCU) values of LSU and small subunits were calculated. Finally, a novel parameter, RSCU distances-based correlation (RDC), for evaluating molecular coevolution was proposed based on Pearson's correlation coefficient between RSCU value matrices of two subunits. The results show that the topological similarity of the mirror tree gave positive evidence of coevolution between subunits. Similarly, the RDC was high, at 0.558, revealing that the RuBisCO subunits coevolved to an extent. The results confirmed that coevolution happened between the two subunits of RuBisCO, although they were encoded by different genomes. This study added new evidence that functional constraint plays an important role in the evolution of the two subunits.

Key words: codon usage, molecular coevolution, RSCU distances-based correlation, RuBisCO