Origin and Evolution of RAS Membrane Targeting

KRAS, HRAS and NRAS proto-oncogenes belong to a family of 40 highly homologous genes, which in turn are a subset of a superfamily of >160 genes encoding small GTPases. RAS proteins consist of a globular G-domain (aa1-166) and a 22-23 aa unstructured hypervariable region (HVR) that mediates membrane targeting. The evolutionary origins of the RAS isoforms, their HVRs and alternative splicing of the KRAS locus has not been explored. We found that KRAS is basal to the RAS proto-oncogene family and its duplication generated HRAS in the common ancestor of vertebrates. In a second round of duplication HRAS generated NRAS and KRAS generated an additional RAS gene we have designated KRASBL, absent in mammals and birds. KRAS4A arose through a duplication and insertion of the 4th exon of NRAS into the 3rd intron of KRAS. We found evolutionary conservation of a short polybasic region (PBR1) in HRAS, NRAS and KRAS4A, a second polybasic region (PBR2) in KRAS4A, two neutralized basic residues (NB) and a serine in KRAS4B and KRASBL, and a modification of the CaaX motif in vertebrates with farnesyl rather than geranylgeranyl polyisoprene lipids, suggesting that a less hydrophobic membrane anchor is critical to RAS protein function. The persistence of four RAS isoforms through >400 million years of evolution argues strongly for differential function.