Exaptation, the co-option of existing traits for new functions, is central to Darwinian evolution.
It has played a role in evolutionary innovations as different as the eye lens and antifreeze proteins. Exaptations typically require multiple small and poorly understood mutational steps. We usually do not know whether these steps are individually favored by natural selection or whether other evolutionary forces are needed to preserve them. Here, Andreas addresses this question in the context of gene regulation, a process underlying evolutionary innovations that range from new bacterial stress responses to new animal body plans. Gene regulation is mediated by transcription factor binding sites (TFBSs), short DNA words near a gene to which proteins called transcription factors (TFs) bind and from which they regulate gene expression. New forms of gene regulation require the evolution of new TFBSs, for example, throughthe exaptation of old TFBSs. I discuss a massively parallel experiment recently performed in my lab to investigate the potential of bacterial TFBSs to evolve exaptively for three Escherichia coli TFs. Using a massively parallel reporter assay, we mapped three adaptive landscapes that encompass all intermediate sequences between three pairs of strong binding sites for each TF. Our results revealed that these landscapes are smooth and navigable, with a monotonic relationship between mutations and their impact on gene regulation. Starting from a strong TFBS for one of our TFs, Darwinian evolution could thus create a strong binding site for another TF through a small number of individually adaptive mutations. Notably, most intermediate genotypes are prone to transcriptional crosstalk – gene regulation mediated by both TFs. Our study presents the first in vivo evidence that new TFBSs can evolve exaptively through multiple small and adaptive mutational steps. They also highlight the importance of regulatory crosstalk for the diversification of bacterial gene regulation.