Alternative splicing allows more than one protein product to be generated from a single gene by selectively including or excluding particular exons in the mature mRNA transcripts. This is a prevalent mechanism of gene regulation with as many as 94% of human genes predicted to undergo the process [1, 2]. Alternative splicing is important in development, in the establishment of tissue specificity and sex differences, and in human disease etiology and progression [3–7].
Alternative splicing is a tightly regulated process involving cis-sequences on the RNA and protein factors that can either promote the inclusion or the skipping of a particular alternative exon in the mature mRNA. Regulatory mechanisms that govern alternative splicing have been extensively studied, and a number of splicing regulatory proteins have been identified and the cis-sequences to which they bind have been characterized [5, 6, 8–15]. More recently other means of splicing regulation have been demonstrated including chromatin remodeling and involvement of the C-terminal domain of RNA Polymerase II as a staging platform for splicing factors during coupled transcription and splicing [16–18].
Two families of well-characterized splicing regulators are the CUG-binding protein (C UG-BP) and embryonic lethal abnormal vision (E LAV) l ike f amily (CELF) and the Muscleblind-like (MBNL) proteins. CELF and MBNL proteins play important roles in the human neuromuscular disease myotonic dystrophy (DM), where their mis-regulation causes alterations in splicing patterns of their target mRNAs. In DM1, CELF protein activity is up regulated, while MBNL protein activity is lost. Interestingly, while these two groups of RNA-binding proteins are known to have distinct mRNA targets, it is also well established that they function antagonistically in the regulation of several alternative exons. The well-characterized common pre-mRNA targets that are antagonistically regulated by CELF and MBNL proteins include cardiac troponin T (cTNT) exon 5, insulin receptor (IR) exon 11, chloride channel 1 (CLCN1) exon 7a, and tau exon 6 [19–21]. Alternative splicing of these exons is mis-regulated in myotonic dystrophy. In these well-studied targets, CELF and MBNL proteins bind to distinct cis-elements. For example, Ho and colleagues utilized cTNT exon 5 minigene reporters in which the potential CELF or MBNL motifs were disrupted to demonstrate that the loss of one family’s binding site does not impact regulation of cTNT exon 5 by the other protein family . In the case of cTNT exon 5, it has been established that MBNL proteins compete with the essential basal splicing factor U2AF65 for binding of the 3’ end of the cTNT intron, and when MBNL prevails it is bound to and possibly stabilizes a secondary structure that prevents U2AF65 binding . An additional six antagonistically regulated targets were identified in a microarray analysis in the developing heart by Kalsotra and colleagues . In DM1 disease, the antagonistically regulated CELF and MBNL protein splicing targets are especially adversely affected, since MBNL function is lost and CELF function is dramatically increased. For this reason, it is important to identify additional antagonistically regulated targets of these two families of regulatory proteins.
Our laboratory has identified one of the alternative exons of the neurofibromatosis type I (NF1) pre-mRNA, exon 23a, as a target of complex splicing regulation. Exon 23a is a particularly attractive exon to study because its coded amino acid sequences are located within the best-characterized domain of the NF1 protein known as the GTPase activating protein-related domain (GRD). The GRD allows the NF1 protein to mediate the conversion of active guanosine-triphosphate bound Ras (Ras-GTP) to inactive guanosine-diphosphate bound Ras (Ras-GDP) (reviewed by ). Interestingly, the type II isoform which includes exon 23a is ten times weaker at regulating the conversion of Ras-GTP to Ras-GDP than the type I isoform in which exon 23a is skipped [25, 26]. Previously, our laboratory has shown that this exon is regulated by at least three different splicing factor protein families: CELF, Hu, and TIA-1 and TIAR [27–29].
Recently we have identified two potential MBNL binding sites, both containing UGCUGU, in the intronic region upstream of exon 23a. In this report we provide evidence to support that the MBNL family of splicing regulators act as positive regulators of NF1 exon 23a inclusion. MBNL1, 2, and 3 all promote exon 23a inclusion when over-expressed in a low MBNL protein-expressing, neuron-like cell line along with an NF1 minigene reporter. Simultaneous siRNA-mediated knockdown of endogenous MBNL1 and MBNL2 proteins in HeLa cells promotes NF1 exon 23a skipping. Our UV cross-linking assays demonstrate that recombinant MBNL1 binds to wild-type RNA oligonucleotides, but not to mutant RNA oligonucleotides in which the potential MBNL sites have been disrupted by mutation to AUAAUA. We show in cells that the relative levels of MBNL and CELF proteins govern whether or not exon 23a will be included, thus showing that CELF and MBNL proteins antagonistically regulate NF1 exon 23a. These results add NF1 exon 23a to a short list of alternative exons that are under complex control by these two families of RNA-binding proteins.