Positive selection in dNTPase SAMHD1 throughout mammalian evolution

Significance Animals defend themselves from viral infection using innate immunity proteins that disrupt various stages of the virus life cycle. In response, viruses produce proteins that bind these host factors and compromise their activity, resulting in evolutionary conflict as immunity and virus proteins adapt to prevent and restore binding, respectively. We report that evolutionary conflict involving the host innate immunity protein SAMHD1 has occurred throughout mammalian evolution. We observe adaptation in a region of SAMHD1 that regulates its activity, and we demonstrate how mutations here influence its enzymatic properties, suggesting that evolutionary conflict has involved modulation of SAMHD1 regulation and function. This correlates with reduced restriction of HIV-1, indicating that positive selection has influenced both SAMHD1’s dNTPase and antiviral activities. The vertebrate protein SAMHD1 is highly unusual in having roles in cellular metabolic regulation, antiviral restriction, and regulation of innate immunity. Its deoxynucleoside triphosphohydrolase activity regulates cellular dNTP concentration, reducing levels below those required by lentiviruses and other viruses to replicate. To counter this threat, some primate lentiviruses encode accessory proteins that bind SAMHD1 and induce its degradation; in turn, positive diversifying selection has been observed in regions bound by these lentiviral proteins, suggesting that primate SAMHD1 has coevolved to evade these countermeasures. Moreover, deleterious polymorphisms in human SAMHD1 are associated with autoimmune disease linked to uncontrolled DNA synthesis of endogenous retroelements. Little is known about how evolutionary pressures affect these different SAMHD1 functions. Here, we examine the deeper history of these interactions by testing whether evolutionary signatures in SAMHD1 extend to other mammalian groups and exploring the molecular basis of this coevolution. Using codon-based likelihood models, we find positive selection in SAMHD1 within each mammal lineage for which sequence data are available. We observe positive selection at sites clustered around T592, a residue that is phosphorylated to regulate SAMHD1 activity. We verify experimentally that mutations within this cluster affect catalytic rate and lentiviral restriction, suggesting that virus–host coevolution has required adaptations of enzymatic function. Thus, persistent positive selection may have involved the adaptation of SAMHD1 regulation to balance antiviral, metabolic, and innate immunity functions.