Splicing Mechanisms That Generate Distinct Isoforms of Protein 4.1R During Terminal Erythroid Differentiation.

Abstract 4036

Poster Board III-972

The protein 4.1R gene is a large complex gene with two translation initiation sites (AUG1 and AUG2) that encode protein isoforms with distinct N-terminal structure and function. Expression of these isoforms is regulated by alternative splicing at either of two splice acceptor sites that flank exon 2' (E2'), in which AUG1 is located. In late erythroblasts E2' is excluded, ensuring translation at AUG2 and synthesis of 80kDa protein 4.1R isoforms. Our earlier studies (EMBO J. 27:122-31, 2008) described a two-step intrasplicing pathway that enforces this splicing outcome exclusively in 4.1R pre-mRNA initiated at exon 1A (E1A), the major transcription start site in late erythropoiesis. The downstream exon 1B (E1B) first splices to the proximal acceptor site at E2' to generate an intermediate structure, which is then re-processed by splicing of E1A to the internal acceptor at E2, removing E1B as well as E2'/AUG1. Experiments with minigenes suggested that intrasplicing is likely independent of specific promoter elements at E1A, but absolutely requires 5' splice site and branch point motifs associated with E1B. Here we sought evidence for functionality of these latter elements in the more physiological context of the endogenous 4.1R gene. The intrasplicing model predicts that morpholino oligonucleotides complementary to key regulatory motifs will block the first step of the pathway in natural 4.1R pre-mRNA transcripts, and yield inappropriate splicing of E1A to the first acceptor at E2'. Antisense morpholinos directed against the E1B branchpoint or E1B 5'splice site were transfected into cells and 4.1R splicing was examined 48hrs later by RT-PCR analysis. Both anti-4.1 morpholinos, but not a control morpholino, resulted in a concentration-dependent shift of E1A splicing to the proximal E2' acceptor site. In other studies we explored whether intrasplicing could occur internally within a gene, using a model pre-mRNA in which a constitutive exon was engineered between E1A and E1B. Analysis of this experiment suggested that any exon upstream of active E1B would follow the intrasplicing pathway and delete E2'. We speculate that internal E1B-like elements in other genes could be selectively activated or silenced by splicing regulatory motifs in order to control downstream splice acceptor choice. Preliminary experiments indicate that weakening the E1B 5' splice site and its upstream pyrimidine tract might permit such regulation as observed for other alternative exons. Together these results provide additional support for the intrasplicing model and suggest that it could function more widely in human genes to coordinate splicing events and generate multiple protein isoforms with distinct functions.