Reported in vivo splice-site mutations in the factor IX gene: severity of splicing defects and a hypothesis for...

Small consensus sequences have been defined for RNA splicing, but questions about splicing in humans remain unanswered. Analysis of germline mutations in the factor IX gene offers a highly advantageous system for studying the mutational process in humans. In a sample of 860 families with hemophilia B, 9% of independent mutations are likely to disrupt splicing as their primary mode of action. This includes 26 splicing mutations reported herein. When combined with the factor IX splice mutations reported by others, at least 104 independent mutations have been observed, 80 of which are single base substitutions within the splice donor and splice acceptor consensus sequences. After analysis of these mutations, the following inferences emerge: (1) the susceptibility of a splice donor sequence to deleterious mutation depends on the degree of similarity with the donor consensus sequence, suggesting a simple "5-6 hypothesis" for predicting deleterious vs. neutral mutations; (2) the great majority of mutations that disrupt the splice donor or splice acceptor sequences result in at least a 100-fold decrement in factor IX coagulant activity, indicating that the mutations at these sites generally function as an on/off switch; (3) mutations that create cryptic splice junctions or that shorten but do not interrupt the polypyrimidine tract in the splice acceptor sequence can reduce splicing by a variable amount; and (4) there are thousands of potential donor-acceptor consensus sequence combinations in the 38-kb factor IX gene region apparently not reduced by evolutionary selective pressure, presenting an apparent paradox; i.e., mutations in the donor and acceptor consensus sequences at intron/exon splice junctions can dramatically alter normal splicing, yet, appropriately spaced, good matches to the consensus sequences do not predispose to significant amounts of alternative splicing.