Aging is thought to occur through the accumulation of biochemical damage affecting DNA, proteins, and lipids. The major source of cellular damage involves the generation of reactive oxygen species produced during mitochondrial respiratory activity of the electron transport chain. Energetic metabolism, antioxidative processes, genome maintenance, and cell cycle are the cellular functions most commonly associated with aging, from experimental studies of model organisms. The significance of these experiments with respect to longevity-related selective constraints in nature remains unclear. Here we took a phylogenomic approach to identify the genetic targets of natural selection for elongated life span in mammals. By comparing the nonsynonymous and synonymous evolution of approximately 5.7 million codon sites across 25 species, we identify codons and genes showing a stronger level of amino acid conservation in long-lived than in short-lived lineages. We show that genes involved in lipid composition and (collagen associated) vitamin C binding have collectively undergone increased selective pressure in long-lived species, whereas genes involved in DNA replication/repair or antioxidation have not. Most of the candidate genes experimentally associated with aging (e.g., PolG, Sod, Foxo) have played no detectable role in the evolution of longevity in mammals. A large body of current medical research aims at discovering how to increase longevity in human. In this study, we uncovered the way natural selection has completed this task during mammalian evolution. Cellular membrane and extracellular collagen composition, not genome integrity, have apparently been the optimized features.