Globin Gene Sequences and Patterns of Single Nucleotide Substitutions (SNS) as a Result of Selective Processes

Human α, β and γ coding sequences (CS) and the distribution of SNS found in a random sample of 160 (75 α and 85 β) globin variants show, among others, the following characteristics: (a) All 61 non-termination codons are translatable. (b) Non-obliga¬tory preterm codons are never used in α and β and avoided in γ genes. (c) Substitutions involving U are much less frequent than expected (p « 0.001). The complementary strands (CCS) of β, γ (and ε), if read in the same frame of globin CS, contain coding, leader and poly A tailing sequences as well as four splicing sequences, with the appropriate polarity, thus showing the required properties for coding a protein. Points (a) and (b) hold true also for CCS gene. Moreover, none of the described β variants (168) removes initiation or introduces termination codons in the CCS. According to these observations it seems conceivable that fixation of the actual globin gene sequences, among the ~ 1070 synonymous, and of the pattern of SNS are the result of selective processes that avoid drastic consequences for hemoglobin function. The chance of early termination of the translation process is in fact reduced by 90% because of (a) and (b) and because non-U→U substitutions are required to produce term codons, from preterm obligatory codons, in about 4/5 of the cases. Non-polar hydrophobic↔hydrophilic aminoacid substitutions, since they require a non¬ U↔U substitution, are reduced as well (by 50%). Further increase of the selective constraint would come from the existence of proteins coded by CCS and from the fact that they also obey conditions (a) and (b).