Mammalian Fox proteins are expressed in neuronal tissues and cell lines. (A) Human Fox genes are represented with first exons in yellow, untranslated regions in red, and coding exons in blue. Exons showing clear homology in the mouse are shaded. The exons are numbered to allow their identification in the table of genomic coordinates (see Table S1 in the supplemental material). This numbering does not necessarily correspond to previously reported exon numbers. The exon used for probing the Northern blot in panel B is marked with a green bar, and the peptides from the same exon used for antibody production are indicated. The exons encoding the RRM are marked with a gray bar. Both genes have multiple promoters and numerous alternative splicing events. The coding alternative exons shown all conform to splice site consensus rules and are conserved between human and mouse. Nearly all are predicted from at least two ESTs, and some were confirmed experimentally by RT-PCR (data not shown). Alternative splicing of exon 19 of Fox-1 and exon 13 of Fox-2 gives rise to two C-terminal peptide sequences. These two splice variants of Fox-1, Fox-2, and Fox-3 from human and mouse are aligned in Fig. S1 in the supplemental material, along with two worm homologues. (B) Northern blot analysis of mouse tissues. A blot (Ambion) carrying 2 μg of poly(A)⁺ RNA from each of the indicated mouse tissues was probed sequentially with Fox-1, Fox-2, and β-actin cRNAs. MWM, molecular weight markers; 14d, 14 days. (C) Western blot analysis of human and mouse cell lines. Protein (50 μg) from whole-cell lysates was separated by 4 to 20% gradient SDS-PAGE, transferred to nitrocellulose, and probed with the indicated antibodies. GAPDH is shown as a loading control. The top and middle panels are different exposures of the same blot. For the longer exposure, the GAPDH portion of the blot was cut off. The cell lines are HeLa, HEK293, 3T3, C2C12 without and with myotube differentiation (diff), mouse N2A and N1E-115 neuroblastomas without and with neuronal differentiation, human neuroblastoma LA-N-5, and human retinoblastoma WERI-1. Differentiated samples are marked with a red dash.

Fox-1 and Fox-2 are coexpressed in the nuclei of most neurons of the adult mouse brain but excluded from glia. (A) Immunostaining of Fox proteins in hippocampus. Fox-1 expression (red) overlaps completely with neuronal marker NeuN (green) but shows no overlap with glial marker GFAP (blue). Individual channels are shown in small panels on the right, and overlay images are shown in larger panels on the left. The boxed region in the left panel is shown in the right panels. (B) In the hippocampus, Fox-1 expression (red) also overlaps completely with Fox-2 expression (green). (C) High-power confocal microscopy in the cortex reveals that both Fox-1 (red) and Fox-2 (green) are expressed primarily in nuclei (blue). Some small nuclei (arrowheads), likely glia, contain no Fox labeling.

Fox proteins enhance neuron-specific splicing patterns in mouse N2A mRNAs. (A) EWS neuron-specific splicing is induced during neuronal differentiation of N2A cells. Cells were grown in standard media or differentiation media for 5 days, and RNA was harvested. EWS exon 4′ splicing was assayed by RT-PCR and quantified. The percent exon inclusion is indicated. std. dev., standard deviation. (B) Inclusion of the N1 exon in the endogenous c-src mRNA is enhanced by Fox overexpression. Each Fox plasmid (100 ng or 1 μg) was transfected into N2A cells, and total RNA was harvested after 60 h. RNA was assayed by RT-PCR and quantified. (C) Inclusion of the 4′ exon in the endogenous EWS mRNA is enhanced by Fox overexpression; similar to panel B, except assayed for EWS splicing by RT-PCR using an end-labeled primer in EWS exon 4. (D) Immunoblot confirming overexpression of Fox-1 and Fox-2 proteins in N2A cells. Whole-cell lysates (50 μg) were separated by 10% NuPage and subjected to immunoblotting with FLAG and GAPDH antibodies.

Fox enhances splicing through the c-src intronic splicing enhancer. (A) A chimeric minigene, pDUP4-5, with β-globin exons 1 and 2 flanking the c-src N1 exon and its intronic regulatory regions. This is completely repressed in HeLa cells. (B) Fox expression enhances N1 splicing. Splicing reporter (100 ng) was cotransfected with empty expression vector (lane 1) or 100 ng (lanes 2 and 5) or 900 ng (lanes 3 and 6) of Fox expression plasmid. Lanes 2 and 3 show increasing Fox-1, and lanes 5 and 6 are increasing Fox-2. RNA was harvested after 48 h, assayed by RT-PCR, and quantified. (C) A minigene carrying β-globin exons 1 and 2 with a 33-nt generic internal exon composed of β-globin sequence. The wild-type (WT) (pDUP4-28) or mutant (MUT) (pDUP4-28Mut) c-src intronic enhancer sequence is inserted in the second intron (38). With the wild-type enhancer, the internal exon is partially included without Fox expression. The putative Fox sites are shown below, with the mutations made for disrupting enhancer activity. (D) Fox expression enhances splicing through sequences in the c-src enhancer. Cotransfections and analysis were as for panel B, except with pDUP4-28 and pDUP4-28Mut. (E) Immunoblot confirming expression of the Fox proteins in HeLa cells. Whole-cell lysates (50 μg) were separated by 10% NuPAGE, transferred to nitrocellulose, and probed with FLAG and GADPH antibodies.

Fox-2 cross-links specifically to the UGCAUG element within the N1 exon intronic splicing enhancer. (A) Location of the labeled phosphate within the wild-type (WT) and mutant (MUT) BS27 RNAs. The labeled phosphates are indicated by asterisks. The wild-type and mutant UGCAUG elements are underlined. (B) Cross-linking to the wild-type UGCAUG element. (C) Cross-linking to the mutant UGACUG element (MUT). Lane 1 contains molecular weight markers and lane 2 the total cross-linking reaction (20% of the input to the immunoprecipitations in lanes 3 to 5). Lanes 3 to 5 are immunoprecipitations with the following antibodies: preimmune serum (lane 3); affinity purified anti-Fox-2 NT (Lane 4); anti-Fox-2 NT preblocked with the cognate Fox-2 NT peptide (lanes 5).

Fox enhances splicing through tandem UGCAUG sequences in HeLa cells. Test enhancers were inserted into the second intron of pDUP4-1 β-globin splicing reporter. (A) An enhancer construct with three copies of UGCAUG element separated by 10-nt globin intron spacers. pDUP4-108 (100 ng) was cotransfected with empty expression vector (lane 1) or 100 ng or 900 ng of Fox-1 or Fox-2 expression plasmid. Lanes 2 and 3 have increasing Fox-1, and lanes 5 and 6 have increasing Fox-2. RNA was harvested after 48 h, and inclusion was assayed by RT-PCR. (B) Conditions similar to those for panel A but using the pDUP4-108Rev reporter plasmid, which has the spacer sequences reversed. (C) Conditions similar to those for panel A but with the pDUP4-108Mut reporter plasmid carrying three copies of UGACUG. (D) Knockdown of Fox-2 in HeLa cells results in loss of enhancer activity. The pDUP4-183 reporter plasmid is similar to that in panel A but carries the 5′ half of the c-src N1 exon, a known ASF/SF2-dependent exonic splicing enhancer (ESE) (41a). HeLa cells were cotransfected with 100 ng of pDUP4-183, 900 ng pUC carrier DNA, and the indicated siRNAs. RNA was harvested after 60 h, and exon inclusion was assayed by RT-PCR.

RNAi of Fox-2 affects splicing of endogenous mRNAs in N2A cells. (A) Splicing of the c-src N1 exon after siRNA treatment. Cells were transfected with 1 μg of pUC plasmid and 20 pmols of the siRNA indicated at the top. After 60 h, RNA was harvested and assayed by RT-PCR. (B) Splicing of EWS 4′ exon after siRNA treatment. Similar to the conditions in panel A but assayed for EWS exon 4′. (C) Splicing of fibronectin EIIIB after siRNA treatment. The same procedure as in panel A, except with 5, 10, or 20 pmol of siRNA and with fibronectin primers. Note that an siRNA directed against Fox-1 had no effect (lanes 2 to 4), but the Fox-2 siRNA strongly increased skipping of fibronectin EIIIB (lanes 5 to 7; products labeled to the right). (D) Immunoblot of N2A whole-cell lysates after siRNA treatment. Carrier pUC DNA and siRNA were cotransfected, and whole-cell lysates were harvested after 60 h. Protein (50 μg) was separated by 10% NuPAGE, transferred to nitrocellulose, and probed with Fox-2 NT and GAPDH antibodies. The protein in lane 1 is from cells transfected with pUC plasmid only. Lanes 2 to 4 contain 1 μg pUC and 5 pmol, 10 pmol, or 20 pmol of Fox-2 siRNA, respectively. An immunoblot for Fox-1 showed no detectible expression in this experiment. GADPH is shown as a loading control.

Fox proteins bind to introns to influence the splicing of adjacent exons by multiple mechanisms. (A) Fox promotes exon skipping. In muscle cells, Fox proteins bind upstream of a non-muscle-specific exon (NM) to promote skipping (22). (B) Fox promotes exon inclusion. As described here, Fox proteins bind downstream of a neuron-specific exon (N) or the fibronectin EIIIB exon and activate splicing. (C) Fox promotes intron retention. Intron 6 of the C. elegans Xol-1 transcript is spliced poorly in the presence of C. elegans Fox-1 (43). The protein may bind to this intron to decrease its splicing efficiency.