NUCLEOTIDE DISTRIBUTION IN GYMNOSPERM NUCLEAR SEQUENCES SUGGESTS A MODEL FOR GC-CONTENT CHANGE IN LAND-PLANT NUCLEAR GENOMES

Nuclear protein coding sequences from gymnosperms are currently scarce . We have determined 4 kb of nuclear protein coding sequences from gym nosperms and have collected and analyzed >60 kb of nuclear sequences f rom gymnosperms and nonspermatophytes in order to better understand pr ocesses influencing genome evolution in plants. We show that conifers possess both biased and nonbiased genes with respect to GC content, as found in monocots, suggesting that the common ancestor of conifers an d monocots may have possessed both biased and nonbiased genes. The lac k of biased genes in dicots is suggested to be a derived character for this lineage. We present a simple but speculative model of land-plant genome evolution which considers changes in GC bias and CpG frequency , respectively, as independent processes and which can account for sev eral puzzling aspects of observed nucleotide frequencies in plant gene s.