An erythroid differentiation stage-specific alternative splicing switch involving activation of protein 4.1R exon 16 splicing is critical for the mechanical stability of the erythrocyte plasma membrane. We have previously shown that inclusion of E16 can be negatively regulated by binding of hnRNP A/B proteins to splicing silencer element(s) in E16 and that strongly decreased expression of hnRNP A/B proteins is temporally correlated with exon 16 activation. Moreover, our earlier unpublished data showed that Fox-2 is a candidate activator protein for this splicing switch, based on observations that Fox-2 binds to an intron enhancer containing three copies of UGCAUG located 96–144nt downstream of exon 16; that Fox-2 enhances exon 16 splicing in HeLa cell co-transfection assays; that mutations blocking Fox-2 binding abrogate its stimulation of exon 16 splicing; and that Fox-2 is expressed in erythroblasts. New experiments reinforce these findings by showing that knockdown of Fox-2 expression, using two different siRNA sequences, strongly inhibits exon 16 splicing efficiency. Together these results indicate that A1 and Fox-2 have antagonistic splicing activities on exon 16. To test whether antagonism involves competitive binding to the intron enhancer region, in vitro binding studies were performed using a biotinylated 39-mer RNA containing two UGCAUG elements from the intron 16 enhancer. This RNA bound strongly to in vitro-synthesized Fox-2 protein, as shown by pull down assays followed by Western blot analysis. hnRNP A1 was also bound to this intronic enhancer region. UGCAUG mutations eliminated Fox-2 binding to the RNA, but did not block A1 binding; such mutants exhibited significantly lower exon 16 splicing efficiency. Most importantly, a competitive binding experiment showed that Fox-2 protein reduces A1 binding to the enhancer RNA in a concentration-dependent manner. Together with previous findings by ourselves and others, these results suggest that exon 16 splicing is governed by two pairs of antagonist interactions: (1) in exon 16, the silencer protein hnRNP A1 antagonizes SF2/ASF activity; and (2) in the downstream intron, hnRNP A1 antagonizes the Fox-2 activator protein. Exon 16 is known to be expressed in brain and muscle in addition to late erythroid cells. We propose that this regulatory network may provide independent mechanisms for exon 16 activation in different cell types by altering the relative abundance of these activators and inhibitors, and thereby the relative efficiency of spliceosome recruitment as the first step in the exon inclusion pathway. To test the mechanism of Fox-2 activity, we are attempting to isolate physiologically relevant co-factors that interact with Fox-2. A recombinant Fox-2 protein containing a biotinylation tag has been expressed in HeLa cells. In transfected cells, the biotin-tagged Fox-2 is properly localized to the nucleus, and retains the ability to enhance 4.1 exon 16 splicing in standard splicing assays as judged by RT/PCR analysis of mRNA products that include or exclude exon 16. Streptavidin pull-down assays should facilitate isolation of Fox-2 complexes and ultimately provide novel insights into the mechanism of this critical splicing switch during erythroid differentiation.

Author notes

Corresponding author

Sign in via your Institution