(A) PTBP1 downregulation is necessary for the miR-124 mediated splicing of NS-specific 3′ alternative exons. Transfected CAD cells were prepared as in Figure 1F and analyzed by RT-qPCR. (Top) Structure of the alternatively spliced regions of Mtap4, Rufy3, and Cdc42 genes. Blue, NS-specific alternative 3′ exons induced by miR-124; peach, other exons. Black arrowheads, polyadenylation signals. (Bottom) RT-qPCR analysis of the N splicing (primers F1/R1) and G splicing (primers F1/R2). The color codes are as shown to the left of the first histogram.
(B) RNAi depletion of PTBP1 (blue bars) stimulates N splicing and represses G splicing in Mtap4, Rufy3, and Cdc42 pre-mRNAs. Yellow bars, samples transfected with a control siRNA. In all RT-qPCR analyses, data are means of three amplifications ± SD, normalized to RI vector + pcDNA3 in (A) or control siRNA transfection in (B).

(A) CAD cells were transfected with siRNAs against PTBP1 and/or PTBP2 as indicated, and the protein levels were analyzed by immunoblotting 72 hr posttransfection. Knockdown of both PTBP1 and PTBP2 induced a more complete derepression of (B) cassette NS exons and (C) 3′-terminal alternative exons than knockdown of PTBP1 alone. RT-qPCR in (C) was carried out as in Figure 2. Data are means of three amplifications ± SD, normalized to control siRNA transfection. For multiplex RT-PCR, we used mixtures of the corresponding F1/R1/R2 primers (see Figure 2A and Table S1).

(A) At E14.5, the mutant forebrain (bottom row) is substantially reduced in size and enveloped by a thin lamina of Foxg1-positive cells expressing high levels of PTBP1 mRNA and the Mtap4 G splice form but low levels of the Mtap4 N form. Wild-type control sections are shown on the top. In each panel, dashed lines mark the lateral and diencephalic boundaries of the telencephalon. LV, lateral ventricle; VZ, ventricular zone; di, diencephalon; and me, mesenchyme. Rostral, left; lateral, top.
(B) Quantitation of PTBP1 expression from (A). To quantify, PTBP1 mRNA levels in the lateral telencephalic wall of Dicer mutant and the lateral layer of wild-type VZ were normalized to corresponding levels in mesenchymal cells and the normalized values were averaged from six mutant and six wild-type forebrain sections (±SD).
(C–E) miR-124 reduces PTBP1 expression in cortical neurons. (C) Primary cortical cultures from E15.5 embryos contain ~95% Tuj1-positive cells. (D) Transfection of these cells with 2′-OMe-RNA against miR-124, but not miR-1, inactivates the endogenous miR-124. (E) RT-qPCR data showing that miR-124 knockdown induces an increase in the PTBP1 and Vamp3 mRNA levels, but not the levels of Rps17 mRNA. Hprt mRNA was used as a normalization control, and the corresponding expression levels in neurons transfected with anti-miR-1 were set to 1. Results are averaged from two independent transfection experiments each done in triplicate ± SD.

(A) CAD cells were transfected with the RI or RIPmiR-124 plasmids and FACS sorted to enrich dsRed-positive cells and postcultured for 12 hr or 60 hr. The expression of miR-124 was analyzed by northern blotting (2.5 mg/lane total RNA) along with RNA from newborn (P0) mouse brain (2.5 mg/lane) and a synthetic miR-124 22-mer RNA (Dharmacon; 5 fmol/lane). The mature miR-124 species (21–24 nt) and the hairpin-like miR-124 precursor (pre) (Kim, 2005) are indicated on the left. Bottom panel, U6 snRNA loading control.
(B) Table based on microarray data listing genes upregulated by miR-124 >2-fold. NS enrichment of exon-included and exon-skipped forms was calculated as a fraction of corresponding cDNA/EST clones originating from the NS.
(C) Luciferase reporter constructs containing PTBP1 and PTBP2 3′UTRs. Positions of the miR-124 MTSs are indicated.
(D) Interactions of miR-124 with the conserved 8- and 9-mers within the PTBP1 and PTBP2 3′UTRs.
(E) N2a cells were cotransfected with RIPmiR-124 luciferase reporter constructs containing the shown 3′UTRs, and the luciferase activity was measured 24 hr posttransfection. For each experiment, luciferase activity was normalized to the corresponding RI vector control. Data are the means of three independent transfections, each done in triplicate, ± SD.
(F) Immunoblot showing that miR-124 induces a decrease in PTBP1 protein and an increase in PTBP2 protein. CAD cells were transfected with RIPmiR or RI in the presence of pcDNA3 or pCMV-fPTBP, sorted, and postcultured for 60 hr. Tubulin (Tub) is a loading control. Graphs on the right show quantitation of the corresponding band intensities performed by detecting chemiluminescence with a CCD camera.
(G) PTBP1 downregulation is required for the miR-124-induced neuronal differentiation of CAD cells. Cells were prepared as in (F) and imaged using dsRed fluorescence.

(A) Ptbp2 gene diagram showing the alternatively spliced exon 10 (blue). The correct stop codon and the premature termination codon (PTC) generated as a result of exon 10 skipping are indicated by red and purple arrowheads, respectively.
(B) Both miR-124 and PTBP1-specific siRNA reduce PTBP1 mRNA levels and induce a reciprocal increase in PTBP2 mRNA. The RT-qPCR analysis was carried out using primers specific to constitutively spliced regions of the two genes (Ptbp1 F1/R1 and Ptbp2 F2/R2, respectively; Table S1). Data are means of three amplifications ± SD. The mean of the RI transfection experiment was set to 1.
(C and D) (C) Time course of PTBP1 protein downregulation by RNAi. Only PTBP1 knockdown leads to the accumulation of PTBP2 protein. (D) RT-PCR showing the effects of PTBP1 and Rent1 knockdowns on the distribution of exon 10 included (i) and skipped (s) forms of PTBP2 mRNA.
(E) Rent1 knockdown consistently increased the fraction of the (s) form. Quantifications are averages of four independent experiments ± SD.

(A) Mature miR-124 expression in an E12.0 embryo visualized by ISH with an miR-124-specific LNA probe.
(B) ISH of E10.5 transverse sections at the cervical level showing that PTBP1 mRNA and miR-124 are expressed in a reciprocal pattern, while the expression of miR-124 and PTBP2 mRNA overlap. The exception to this is in the NL, where both PTBP1 and PTBP2 mRNA are present. However, at the protein level PTBP1 and PTBP2 are not coexpressed in the NL, suggesting that PTBP1 is translationally repressed.
(C) Immunofluorescence analysis shows that PTBP1 protein (green) is largely excluded from the spinal cord, whereas PTBP2 protein (red) is expressed throughout the spinal cord but excluded from cells expressing high levels of PTBP1 protein.
(D) ISH of horizontal sections of diencephalon at E13.5 showing miR-124 expression in differentiating neurons, opposite to PTBP1 expression in the VZ containing neural precursor cells, both dorsally (Neurog2-positive region) and ventrally (Mash1). PTBP2 expression in the VZ is relatively low, but it is higher in the subventricular zone (SVZ), which contains cells committed to differentiation. The pattern is characterized by the SVZ marker Dlx5.
(E) ISH of horizontal sections of forebrain at E13.5 showing PTBP1 is expressed in the VZ of the telencephalon, where little or no miR-124 is detected. As expected, the N splicing patterns of Cdc42 and Rufy3 are opposite to that of PTBP1 expression, and identical to that of miR-124 expression. By contrast, the G forms of Cdc42 and Rufy3 are enriched or at least not excluded in the regions expressing PTBP1. sc, spinal cord; rh, rhombencephalon; me, mesencephalon; di, diencephalon; te, telencephalon; NL, neuroepithelial layer; ML, mantle layer; DRG, dorsal root ganglia; VZ, ventricular zone; SVZ, subventricular zone; CP, cortical plate; and LV, lateral ventricle.

(A) miR-124 stimulates neuronal differentiation of P19 cells. Cells were transfected with RIPmiR-124 plasmid or RI vector, induced with RA for 4 days, and plated on polylysine/laminin-coated coverslips. Two days after plating, the cells were fixed and stained with antibodies against MAP2. Arrows, dsRed-positive cells, which are also MAP2 positive (neurons); arrowheads, dsRed-positive cells, which do not express MAP2 (immature neurons and glia).
(B) PTBP1 downregulation is required for an optimal stimulatory effect of miR-124. The experiment was carried out as in (A), but cells were cotransfected with RIPmiR-124 plasmid or RI vector and pCIG vector or pCIG-PTBP1 expression plasmid prior to RA treatment. Both pCIG and pCIG-PTBP1 plasmids also encode EGFP protein that localizes to the nucleus. The fraction of MAP2-positive cells among cells expressing both dsRed and EGFP was averaged over seven randomly selected fields ± SD.
(C) Regulatory networks controlled by miR-124. In nonneuronal cells, high levels of PTBP1 repress NS-specific alternative splicing, which helps maintain a nonneuronal state. During neuronal differentiation, high levels of miR-124 downregulate PTBP1, leading to a switch to NS-specific alternative splicing and neuronal differentiation. A decrease in the PTBP1 levels also promotes exon inclusion in PTBP2 mRNA and accumulation of PTBP2 protein. PTBP2 fine-tunes the splicing pattern by weakly repressing the inclusion of a subset of NS-specific exons. In addition to PTBP1, miR-124 also downregulates several other nonneuronal mRNAs, including those of SCP1, laminin γ1, and integrin β1 (Cao et al., 2007; Visvanathan et al., 2007), as well as proteins involved in cell proliferation (Figure S5; Wang, 2006).