Highly Expressed Genes Are Preferentially Co-Opted for C4 Photosynthesis

Novel adaptations are generally assembled by co-opting pre-existing genetic components, but the factors dictating the suitability of genes for new functions remain poorly known. In this work, we used comparative transcriptomics to determine the attributes that increased the likelihood of some genes being co-opted for C4 photosynthesis, a convergent complex trait that boosts productivity in tropical conditions. We show that independent lineages of grasses repeatedly co-opted the gene lineages that were the most highly expressed in non-C4 ancestors to produce their C4 pathway. Although ancestral abundance in leaves explains which genes were used for the emergence of a C4 pathway, the tissue specificity has surprisingly no effect. Our results suggest that levels of key genes were elevated during the early diversification of grasses and subsequently repeatedly used to trigger a weak C4 cycle via relatively few mutations. The abundance of C4-suitable transcripts therefore facilitated physiological innovation, but the transition to a strong C4 pathway still involved consequent changes in expression levels, leaf specificity, and coding sequences. The direction and amount of changes required for the strong C4 pathway depended on the identity of the genes co-opted, so that ancestral gene expression both facilitates adaptive transitions and constrains subsequent evolutionary trajectories.