Proteome partitioning constraints on long-term laboratory evolution

Abstract

Adaptive laboratory evolution experiments provide a controlled context in which the dynamics of selection and adaptation can be followed in real-time at the single-nucleotide level1. And yet this precision introduces hundreds of degrees-of-freedom as genetic changes accrue in parallel lineages over generations2. On short timescales, physiological constraints have been leveraged to provide a coarse-grained view of bacterial gene expression characterized by a small set of phenomenological parameters3-5. Here, we ask whether this same framework, operating at a level between genotype and fitness, informs physiological changes that occur on evolutionary timescales. Using Lenskis Ara-1 lineage adapted to growth in glucose minimal medium6, we find that the proteome is substantially remodeled over 40 000 generations. We apply our existing quantitative proteomics analysis to partition hundreds of expressed proteins into six sectors with shared metabolic function and physiological response4. To accommodate the increased growth rates in the evolved strains, expression of metabolic enzymes undergoes sector-specific adaptation to enable increased fluxes. We find that catabolic proteins adapt by increasing the total enzyme abundance, whereas anabolic and glycolytic proteins exhibit decreased free-enzyme pools. We propose that flux-dependent regulation7 and substrate saturation8 can account for the sector-specific remodeling.