Ecological selection for small microbial genomes along a temperate-to-thermal soil gradient

Small bacterial and archaeal genomes provide insights into the minimal requirements for life1 and seem to be widespread on the microbial phylogenetic tree2. We know that evolutionary processes, mainly selection and drift, can result in microbial genome reduction 3,4. However, we do not know the precise environmental pressures that constrain genome size in free-living microorganisms. A study including isolates 5 has shown that bacteria with high optimum growth temperatures, including thermophiles, often have small genomes 6. It is unclear how well this relationship may extend generally to microorganisms in nature 7,8, and in particular to those microbes inhabiting complex and highly variable environments like soil 3,6,9. To understand the genomic traits of thermally-adapted microorganisms, here we investigated bacterial and archaeal metagenomes from a 45{degrees}C gradient of temperate-to-thermal soils overlying the ongoing Centralia, Pennsylvania (USA) coal seam fire. There was a strong relationship between average genome size and temperature: hot soils had small genomes relative to ambient soils (Pearsons r = -0.910, p < 0.001). There was also an inverse relationship between soil temperature and cell size (Pearsons r = -0.65, p = 0.021), providing evidence that cell and genome size in the wild are together constrained by temperature. Notably, hot soils had different community structures than ambient soils, implicating ecological selection for thermo-tolerant cells that had small genomes, rather than contemporary genome streamlining within the local populations. Hot soils notably lacked genes for described two-component regulatory systems and antimicrobial production and resistance. Our work provides field evidence for the inverse relationship between microbial genome size and temperature requirements in a diverse, free-living community over a wide range of temperatures that support microbial life. Our findings demonstrate that ecological selection for thermophiles and thermo-tolerant microorganisms can result in smaller average genome sizes in situ, possibly because they have small genomes reminiscent of a more ancestral state.