Historical Contingency Limits Adaptive Diversification in a Spatially Structured Environment

Understanding how genotype-by-environment (GxE) interactions influence evolutionary trajectories and contribute to historical contingency is key to predicting evolution. In spatially structured, heterogeneous environments, populations often diversify into ecotypes, resulting in niche specialization. However, the ability to specialize depends not only on ecological opportunity but also on whether genetic variation permits access to novel niches, as genetic disruptions may inhibit adaptation unless alternative trajectories or compensatory mutations are subsequently accessible. Previously, we demonstrated that Escherichia coli populations rapidly diversify into two co-existing ecotypes in a nutrient-rich, spatially structured environment. The adaptation of both ecotypes results in benefits only perceived in the spatially structured culture tube environment. Here, diversification is initiated by first-step mutations associated with the overexpression of genes encoding the type 1 fimbria, the major attachment pilus involved in biofilm development, enabling range expansion and allowing E. coli to occupy the surface-air interface of the culture tube.

To investigate how first-step mutations shape evolutionary trajectories, we experimentally evolved wild-type and fimbrial-deficient (Δ fimA ) E. coli for 91 days in both structured (tube) and unstructured (flask) environments. While a fimA deletion initially confers a fitness benefit by avoiding the cost of insufficient biofilm formation, it ultimately prevents range expansion in structured environments and is not compensated by expression of cryptic fimbriae by the end of experimental evolution. As a result, Δ fimA populations show constrained adaptation in tubes compared to wild-type. Alternatively, both genotypes perform similarly in flasks, where biofilm formation is not advantageous and whole population sequencing reveals that flask-evolved populations similar early mutational trajectories. Our results highlight the ruggedness of the adaptive landscape in structured environments and show how an initially beneficial mutation can trap a lineage on a local fitness peak, underscoring the importance of G×E interactions and early mutational events in shaping the predictability and contingency of evolutionary outcomes.