In vertebrates, there are over 15 different myosin II isoforms, each of which contains different myosin II heavy chains (MHC IIs). MHC II isoform diversity is generated by multiple genes as well as by alternative splicing of pre-mRNA. Previous studies have demonstrated cell type-specific expression of MHC II isoforms as well as changes in MHC II isoforms during the course of muscle and neural tissue development. This research program has investigated the regulatory mechanisms responsible for the expression of three nonmuscle MHC II (NMHC II) genes, NMHC II-A, NMHC II-B, and NMHC II-C. We have been studying the transcriptional regulation of NMHC II-A and II-C genes as well as tissue-dependent regulation of alternative splicing of NMHC II-B and C genes. In this report, we focus on regulation of alternative splicing of NMHC II-B. The gene encoding NMHC II-B generates alternatively spliced isoforms, which include or exclude a cassette of amino acids (aa) near the ATP-binding domain. Inclusion of alternative exon B1 (also called exon N30) encoding 10 aa in NMHC II-B mRNAs is restricted to some types of neural cells. We have previously reported that the intronic region downstream of B1, called the intronic distal downstream enhancer (IDDE), is required for activation of B1 splicing. Two copies of an RNA element UGCAUG within the IDDE are essential for B1 inclusion in neural cells. Recently a vertebrate homolog of C. elegans Fox-1 was reported to bind to UGCAUG in a highly sequence-specific manner. A database search revealed that there are three genes for Fox-1 homologs (Fox family) in mammals, Fox-1, Fox-2 and Fox-3. First, we characterized two members of this family, Fox-1 (also called A2BP1) and Fox-2 (also called Fxh and Rbm9). Fox-1 is expressed in brain and striated muscles whereas Fox-2 is expressed in various tissues including brain and muscles. We found that both the Fox-1 and Fox-2 genes generate multiple tissue-dependent isoforms by differential usage of alternative promoters and alternative splicing of internal exons. The isoforms differ in sequences of the N-terminal, internal and C-terminal segments but share an identical RNA recognition motif (RRM). The isoforms of Fox-1 or Fox-2 which are preferentially expressed in brain and contain particular internal and C-terminal segment sequences were found to enhance B1 inclusion more efficiently, compared to the muscle-specific isoforms of Fox-1 and 2. Next, we extended our analysis to the third member of the Fox family proteins, Fox-3, which has been poorly characterized to date. Fox-3 contains an RRM highly homologous to those of Fox-1 and 2 and binds to the UGCAUG element. A number of cDNA variants for Fox-3 were obtained, and the internal and C-terminal sequences of all these variants showed high similarity to the sequences of the Fox-1 and 2 isoforms which are active for B1 splicing, but not those of the less active isoforms. Fox-3 is also capable of activating B1 inclusion. Notably, Fox-3 expression was found to be restricted to neural tissues. We raised antibodies specific to each of the Fox proteins and analyzed distribution of Fox proteins in mouse brain and spinal cord using histological techniques. Expressions of Fox-1, 2 and 3 overlap in many kinds of neural cells, but we also observed differences in their expression in certain types of neural cells. Comparison of these Fox expression patterns with expression of the B1-included NMHC II-B mRNA detected by in situ hybridization suggested that the Fox-3 expressing cells tend to include the B1 insert. This notion was further supported by the RT-PCR analysis of B1 splicing patterns of brain and spinal cord cells which were dissociated and sorted according to Fox-3 expression. Fox-3 positive cells from cerebellum, brainstem and spinal cord include the B1 insert in a Fox-3 concentration-dependent manner. Fox-3 negative cells almost completely exclude the B1 insert, despite expression of Fox-1 and 2. Therefore there is a better correlation of the extent of B1 inclusion with the level of Fox-3 expression rather than with that of Fox-1 or 2 expression in intact tissues, although Fox-3 and some of the isoforms of Fox-1 and 2 are similarly capable of enhancing B1 splicing when they are over-expressed in cultured cells. During the course of characterization of Fox-3, we identified Fox-3 as an antigen of a monoclonal antibody, anti-NeuN. Anti-NeuN has been widely used as a reliable tool to detect most post-mitotic neuronal cell types in neuroscience, developmental biology, and stem cell research fields as well as diagnostic histopathology. However, the molecular identity of the antigen NeuN was previously unknown. Anti-Fox-3 and anti-NeuN cross-react with proteins with the same molecular masses, and immunofluorescence microscopy analysis shows exact overlap staining with two antibodies throughout brain. Anti-NeuN recognizes the recombinant Fox-3 and siRNA targeted to Fox-3 eliminates the NeuN antigen in cultured cells. Identification of NeuN as Fox-3 clarified an important element of neurobiology research and should promote future study on neural development and disease in relation to Fox-3 function.
