Involvement of the Mouse Prp19 Gene in Neuronal/Astroglial Cell Fate Decisions*

The molecular mechanisms involved in neuronal/astroglial cell fate decisions during the development of the mammalian central nervous system are poorly understood. Here, we report that PRP19β, a splice variant of mouse PRP19α corresponding to the yeast PRP19 protein, can function as a neuron-astroglial switch during the retinoic acid-primed neural differentiation of P19 cells. The β-variant possesses an additional 19 amino acid residues inframe in the N-terminal region of the α-variant. The forced expression of the α-variant RNA caused the down-regulation of oct-3/4 and nanog mRNA expression during the 12-48 h of the late-early stages of neural differentiation and was sufficient to convert P19 cells into neurons (but not glial cells) when the cells were cultured in aggregated form without retinoic acid. In contrast, the forced expression of the β-variant RNA suppressed neuronal differentiation and conversely stimulated astroglial cell differentiation in retinoic acid-primed P19 cells. Based on yeast two-hybrid screening, cyclophilin A was identified as a specific binding partner of the β-variant. Luciferase reporter assay mediated by the oct-3/4 promoter revealed that cyclophilin A could act as a transcriptional activator and that its activity was suppressed by the β-variant, suggesting that cyclophilin A takes part in the induction of oct-3/4 gene expression, which might lead to neuroectodermal otx2 expression within 12 h of the immediate-early stages of retinoic acid-primed neural differentiation. These results show that the α-variant gene plays a pivotal role in neural differentiation and that the β-variant participates in neuronal/astroglial cell fate decisions.

In mammals, the neural tube consists of a morphologically identical population of neuroepithelial cells from which all the diverse neuronal and glial subtypes of the central nervous system are generated. As neurogenesis begins, proliferating neural progenitors become restricted to the ventricular zone at the inner surface of the neural tube. Stem cells within the ventricular zone cease to produce neurons at around midgestation and instead start to produce glia. What causes these stem cells to switch from neurogenesis to gliogenesis is one of the most important issues in neural differentiation.
In Drosophila, neuronal and glial fates are determined by the gcm ( glial cell missing) gene encoding a transcription factor as a binary switch (1). In a gcm mutant, glial cells are converted to neurons, whereas ectopic expression of gcm in neurons causes neuron-glial transforma-tion. Two mammalian gcm homologs, Gcm1 and Gcm2, have been identified in mice, rats, and humans (2)(3)(4)(5)(6); however, the sites of Gcm1 and Gcm2 expression are largely restricted to the placenta (6, 7) and parathyroid glands (6), respectively. Targeted disruption of the Gcm1 gene in mice results in a severe defect in labyrinth formation in the placenta (8,9). Introduction of the Gcm1 gene into cultured embryonic brain cells results in the induction of an astrocyte lineage. Nonetheless, cultures from Gcm1-deficient mouse brain do not exhibit a significant reduction in the number of astrocytes, suggesting that mouse Gcm1 exhibits the potential to induce gliogenesis, but may function in the generation of a minor subpopulation of glial cells (10). Furthermore, Gcm2-targeted mice exhibit a selective loss of parathyroid glands, but no abnormalities were reported in the nervous system (11). Thus, in mammals, although there is evidence that neurons and glia share a common progenitor (12), the genes that might control a GCM-like switching mechanism have not yet been identified.
Recent studies have shown that the Sox9 transcription factor (13), the QKI RNA-binding protein (14), and extrinsic cues such as bone morphogenetic protein (BMP) 2 and ciliary neurotrophic factor (CNTF) (15) promote the differentiation of neural progenitors into glia. Nevertheless, these limited factors may be insufficient for the regulation of glial cell differentiation because mammalian astrocytes exhibit a large heterogeneity, differing in morphology, distribution, molecular types expressed, function, and cell lineage.
In this context, to clarify the molecular mechanisms in neuronal/ astroglial fate decisions in mammals, we applied a subtractive cDNA cloning strategy for the retinoic acid (RA)-induced neural differentiation system of mouse P19 embryonic carcinoma cells as described previously (16). By this approach, we isolated PRP19␣ (precursor RNA processing-19␣) and its splice variant PRP19␤, with an additional 19-amino acid residues in-frame. PRP19␣ and PRP19␤ are homologs of yeast Prp19, which has been identified in a screen of temperature-sensitive mutants defective in precursor RNA processing of Saccharomyces cerevisiae and shown to be essential for pre-mRNA splicing (17). Nonetheless, there is no report concerning the participation of PRP19␣ and PRP19␤ in neural differentiation so far.
In this study, we show that P19 cells modified to express PRP19␣ constitutively in large quantities no longer required RA for neuronal differentiation. When these cells were cultured in aggregated form, the mRNA expression of neurogenic basic helix-loop-helix (bHLH) transcription factors, including Mash-1, Ngn-1 (neurogenin-1), and Neu-roD, was promoted to sufficient levels for the induction of ␤-tubulin III-positive neurons, but not glial fibrillary acidic protein (GFAP)-positive astroglia, accompanied by the down-regulation of oct-3/4 and nanog mRNA expression. On the other hand, the forced expression of Prp19␤ RNA suppressed neuronal differentiation through the inhibition of CypA (cyclophilin A), which acts as a transcriptional activator, and inversely promoted glial cell differentiation in RA-primed P19 cells. The down-regulation of Prp19␤ RNA by the antisense 15-mer oligodeoxynucleotides (ODNs) specific for the insertional sequence of Prp19␤ stimulated neuronal differentiation and suppressed astroglial cell differentiation in RA-primed P19 cells. These results imply that the regulation of alternative splicing of the mouse Prp19 gene plays an important role in neuronal and astroglial cell fate decisions.

MATERIALS AND METHODS
Cell Cultures and Experimental Animals-Mouse P19 embryonic carcinoma cells were obtained from American Type Culture Collection (Manassas, VA). To induce neural differentiation, 1 ϫ 10 6 aggregated P19 cells were cultured in 10-cm bacteriological grade dishes in 10 ml of ␣-minimal essential medium containing 10% fetal calf serum and 5 ϫ 10 Ϫ7 M all-trans-RA (Sigma) for 4 days as described previously (16). The cell aggregates were suspended by mild pipetting and transferred to tissue culture dishes. The cells were cultured in RA-free ␣-minimal essential medium containing 10% fetal calf serum for an additional 4 days to induce ␤-tubulin III-positive neurons and for 7 days to induce GFAP-positive astroglial cells. ICR mice were purchased from Charles River Japan (Kanagawa, Japan).
Culture of Mouse Neural Stem (NS) Cells-NS cells from hippocampi of ICR mice at postnatal day 7 were cultured in serum-free Dulbecco's modified Eagle's medium/F-12 nutrient mixture containing 20 ng/ml mouse epidermal growth factor (EGF; Sigma) as described previously (18). To induce neural differentiation, the spheres formed by NS cells were plated onto tissue culture dishes precoated with 20 mg/ml poly-L-lysine and cultured in EGF-free medium containing 20 ng/ml recombinant rat CNTF (R&D Systems, Minneapolis, MN) for 2 weeks.
Subtractive Cloning-Tester poly(A) ϩ RNA was prepared from aggregated P19 cells treated with 5 ϫ 10 Ϫ7 M RA for 24 h, and firststrand cDNA was hybridized with excess driver RNA prepared from untreated P19 cells, and then a subtractive cDNA probe was prepared by the addition of the cross-linking agent 2,5-diaziridinyl-1,4-benzoquinone and used to screen 9 ϫ 10 5 plaques of rat brain 5Ј-stretch plus a gt11 cDNA library (Clontech) as described previously (16). The cDNAs inserted into the positive phage were then amplified by PCR using primers based on the phage DNA sequence: 5Ј-primer, 5Ј-GCC ACG ACT CCT GGA GCC CG-3Ј; and 3Ј-primer, 5Ј-CAC CAG ACC AAC TGG TAA TG-3Ј. The amplified cDNAs were digested with EcoRI and subcloned into the EcoRI site of pGEM-7zf(ϩ) (Promega Corp., Madison, WI).
A rat full-length Prp19␣ cDNA clone was isolated from the second screening of 4.5 ϫ 10 5 plaques from a rat brain gt10 cDNA library (Clontech) using the 1.1-kb EcoRI cDNA fragment obtained from the first screening as a probe and then subcloned into the EcoRI site of pGEM-7zf(ϩ). The plasmid was designated as pGEM-Prp19␣.
The Prp19␤ PCR product was subcloned into the EcoRV site of pGEM-5zf(ϩ) and designated as pGEM-Prp19␤. The CypA PCR product was digested with NsiI and inserted into the EcoRI-XhoI site of pACT2 (Clontech) and named pACT2-CypA.
Real-time PCR Analysis of PRP19␣ and PRP19␤ Expression during Neural Differentiation-To generate cDNA from P19 cell culture samples, 5 g of total RNAs was reverse-transcribed with random primer. Real-time PCR was performed using SYBR Green PCR Master Mix (Eurogentec, Seraine, Belgium) according to the manufacturer's protocol. The specific primers used were as follows: Prp19␣, 5Ј-ATC AAC AAC CAG CCT CTC TCA GAG-3Ј (5Ј-primer) and 5Ј-TGC AGC ATG ACT GCA TCC C-3Ј (3Ј-primer); and Prp19␤, 5Ј-CAG AGG AGC AGC TCA TCG ACA-3Ј (5Ј-primer) and 5Ј-AGG GTA GGC CAC TGT GAG GAC T-3Ј (3Ј-primer). Fluorescence was detected at the end of each cycle to monitor the amount of PCR product. The quantities of specific mRNA in the sample were measured according to the corresponding gene-specific standard curve. Amplification of specific transcripts was confirmed by a final melting curve profile and gel electrophoresis. The levels of Prp19␣ and Prp19␤ transcripts at each time point were calibrated as compared with those of ribosomal phosphoprotein transcripts as an internal control (31).
Northern Blot Analysis-Aliquots of 20 g of total RNAs were electrophoresed on 6% formaldehyde-containing 1% agarose gel, transferred to Hybond-N ϩ nylon membrane (Amersham Biosciences, Buckinghamshire, UK), and hybridized with 32 P-labeled PCR-amplified probe as described above. Developed x-ray films were scanned into a Macintosh G4, and the expression levels of these RNAs were quantified using the NIH Image 1.62 ppc program.
Western Blot Analysis-Antibodies against PRP19␣ and PRP19␤ were prepared using glutathione S-transferase-fused rat PRP19␣ and synthetic peptides specific for PRP19␤ (SPALPQSSQWPTLSQ), respectively. Cells were lysed in SDS sample buffer (62.5 mM Tris-HCl (pH 6.8), 2% SDS, 10% glycerol, and 5% 2-mercaptoethanol). Aliquots of 30 g of cell lysate were heated at 95°C for 5 min and subjected to 10% SDS-PAGE. The proteins were transferred to a clear blot membrane  Cruz Biotechnology, Inc.) was used as a secondary antibody as indicated. Signals were visualized with the ECL system (Amersham Biosciences).
Immunocytochemistry-P19 cells were cultured in a Lab-Tek II chamber slide (Nalge Nunc International, Naperville, IL) and fixed with 4% paraformaldehyde for immunocytochemical analysis. The cell samples were incubated in Ca 2ϩ /Mg 2ϩ -free phosphate-buffered saline containing 10% normal rabbit serum for 20 min at 37°C. The samples were then incubated for 1 h at 37°C with antibody against ␤-tubulin III (1:400) or GFAP (1:400) under the same conditions described above. After washing with Ca 2ϩ /Mg 2ϩ -free phosphate-buffered saline, the samples were incubated with biotin-conjugated rabbit anti-mouse IgG ϩ IgA ϩ IgM (Nichirei, Tokyo) as a secondary antibody, followed by incubation with peroxidase-conjugated streptavidin (Nichirei). Visualization was carried out using 3,3Ј-diaminobenzidine. The samples were counterstained with hematoxylin.  ). B, preferential induction of Prp19␤ mRNA during the CNTF-primed neural differentiation of NS cells. Total RNAs were extracted from untreated P19 cells and NS cells (NSCs) cultured with EGF or CNTF for 2 weeks and then analyzed by reverse transcription-PCR. The glyceraldehyde-3-phosphate dehydrogenase (GAPDH) transcript was also analyzed by reverse transcription-PCR as an internal control. C, nucleotide and predicted amino acid sequences of mouse Prp19␤ cDNA. The putative ATG start site and termination codon are indicated at positions 5-7 and 1574 -1576, respectively. The Prp19␤-specific sequence is highlighted in black. The modified RING finger domain, nuclear localization signal, and WD40 repeat are boxed, underlined, and shaded, respectively. The nucleotide sequence of mouse Prp19␤ cDNA was deposited in the DDBJ/GenBank TM /EBI Data Bank under accession number AB222602.

Construction of Enhanced Green Fluorescent Protein (EGFP)-fused PRP19␣ and PRP19␤ Expression
Vectors-EGFP-fused PRP19␣ and PRP19␤ expression vectors were constructed by insertion of the entire coding regions of PRP19␣ and PRP19␤, respectively, downstream of the EGFP-coding sequence in-frame into the pEGFP-C1 vector (Clontech) and designated as pEGFP-PRP19␣ and pEGFP-PRP19␤, respectively. P19 cells were transfected with the pEGFP-PRP19␣ and pEGFP-PRP19␤ expression vectors using Lipofectamine PLUS TM (Invitrogen) and cultured in the presence of 400 g/ml G418 for selection. To see the subcellular localization of PRP19␣ and PRP19␤, the cells were fixed with 4% paraformaldehyde and washed with Ca 2ϩ /Mg 2ϩ -free phosphatebuffered saline. The nuclei were stained with 1 g/ml Hoechst 33258. The cells were observed under a fluorescence microscope (Axioplan 2, Carl Zeiss, Oberkochen, Germany).
Specific Down-regulation of Prp19␤ RNA by Antisense ODNs-The sequence of antisense ODNs used was 5Ј-TAGGCCACTGTGAGG-3Ј, which corresponds to nucleotides 282-296 of the specific region of mouse Prp19␤ cDNA; and the sequence of sense ODNs used for the control experiment was 5Ј-CCTCACAGTGGCCTA-3Ј. P19 cells were treated with various concentration of ODNs as described previously (33).
Immunoprecipitation-The hemagglutinin (HA)-fused CypA expression vector was constructed by insertion of the blunt-ended BglII open reading frame fragment of pACT2-CypA into the EcoRV site of pcDNA3-EF1-␣. P19 cells (3 ϫ 10 5 ) were transfected with 2 g of DNA using Lipofectamine PLUS TM . The cells were harvested 48 h after transfection and lysed with immunoprecipitation buffer (20 mM Tris-HCl (pH 7.5), 150 mM NaCl, 5 mM MgCl 2 , 0.1% Nonidet P-40, and 1 mM phenylmethylsulfonyl fluoride). For immunoprecipitations, equal amounts of cell lysate proteins were incubated overnight with protein A-Sepharose beads charged with anti-PRP19␣ or anti-HA antibody at 4°C. The beads were washed twice with immunoprecipitation buffer. Immunoprecipitated proteins or cell lysates were mixed with 3ϫ SDS sample buffer and separated by SDS-PAGE. Western blot analysis was performed as described above with the appropriate antibodies.
Yeast Two-hybrid Screening-Mouse Prp19␤-specific region cDNA (amino acids 83-101) was inserted into pAS2-1 (Clontech) in-frame with the Gal4 DNA-binding domain as bait. An RA-treated P19 cell cDNA library was constructed using pACT2. Yeast two-hybrid screening was performed as described previously (34).
Luciferase Reporter Assay-The PCR-amplified mouse oct-3/4 promoter (Ϫ412 to ϩ39) was subcloned into the pTRE2-Luc vector (Clontech). The reporter plasmid was termed pLuc-RAREoct. For each transfection, P19 cells (1 ϫ 10 5 /35-mm dish) were transfected with 1 g of pLuc-RAREoct and 0.5 g of the ␤-galactosidase expression vector pcDNA3.1/Myc-His/LacZ (Invitrogen) together with a total of 1 g of pcDNA3-EF1-␣-sense CypA, pcDNA3-EF1-␣-sense Prp19␤, and vacant pcDNA3-EF1-␣ expression vectors in various combinations using Lipofectamine PLUS TM . After 24 h of transfection, the cells were treated with RA for an additional 24 h and lysed with the lysis buffer from the Promega luciferase assay kit, and luciferase activities were determined according to the manufacturer's instructions using a Luminous CT9000D luminometer (DiaIatron, Tokyo). Reporter gene activities were normalized using ␤-galactosidase activity as an internal control.

RESULTS
Isolation of Prp19␣ cDNA-To isolate the early response genes that participate in the RA-primed neural differentiation of mouse P19 cells, we employed a subtractive cDNA cloning strategy to enrich genes that are highly expressed in P19 cells after 24 h of RA treatment as described previously (16). Using the subtractive probes obtained, ϳ9 ϫ 10 5 plaques from the rat brain gt11 cDNA library were screened, and nine clones whose expression levels were stimulated by Ͼ2-fold over the untreated control cells were identified. A homology search of the nine clones in the DDBJ, EBI, and GenBank TM Data Banks showed that five were novel cDNA clones, and the remaining four clones were identified as acyl-CoA synthetase III, cytochrome c oxidase, Trip15/CSN2, and fibroblast growth factor receptor-2. In this study, we focused on one of the five novel clones (30-1-2) because the increment in its expression level was highest of the five novel clones during the RA-primed neural differentiation of P19 cells. To isolate the full-length 30-1-2 cDNA, we screened the rat brain gt10 cDNA library using the 1.1-kb EcoRI cDNA fragment obtained from the first screening as a probe and isolated the 1853-bp full-length cDNA clone. The cDNA encodes 504 amino acid residues, and its molecular mass is 55,238 Da. The cDNA sequence was deposited in DDBJ/GenBank TM /EBI Data Bank under accession number AB020022 (Fig. 1A). A conserved domain search identified five WD40 repeat domains in the C-terminal region and a modified RING finger domain in the N-terminal region. A protein BLAST search in the GenBank TM Data Bank revealed that the protein is identical or closely related to PRP19␣ in Mus musculus (100% identity), Homo sapiens (99.4%), and S. cerevisiae (36.0%) (Fig. 1B). Therefore, we designated the full-length 30-1-2 cDNA clone isolated from the rat brain cDNA library as rat Prp19␣. The deduced amino acid sequences of the rat and mouse PRP19␣ proteins are completely identical, suggesting that they could function in a similar manner. Thus, we used rat Prp19␣ cDNA throughout the experiment, even though P19 cells were originally established in a mouse embryo (36).
Detection of the PRP19␣ Splice Variant PRP19␤-Because Prp19␣ mRNA expression was up-regulated in the RA-primed neural differentiation of P19 cells, we analyzed the expression pattern of the Prp19␣ gene during the differentiation of mouse hippocampus-derived NS cells (Fig. 2). The NS cells formed neurospheres in medium containing 20 ng/ml EGF and actively incorporated bromodeoxyuridine ( Fig. 2A, panel a) as described previously (18). When EGF-generated spheres were plated onto poly-L-lysine-coated tissue culture dishes after the removal of EGF and cultured in medium containing 20 ng/ml CNTF for 2 weeks, the NS cells differentiated into ␤-tubulin III-positive neurons ( Fig. 2A, panel b) and myelin basic protein-positive oligodendrocytes (panel c). The expression level of Prp19␣ mRNA during the differentiation of NS cells was analyzed by semiquantitative real-time PCR. In the presence of EGF, the expression level of Prp19␣ mRNA in NS cells was similar to that in untreated P19 cells, whereas the expression level in CNTF-treated NS cells was 1.5-fold higher than that in NS cells cultured with EGF (Fig. 2B). Furthermore, an additional PCR product termed Prp19␤ was detected in cDNA derived from CNTF-treated NS cells.
Sequence analysis revealed that this Prp19␤ cDNA possesses an additional 57 bp encoding 19 amino acid residues in-frame (DDBJ/Gen-Bank TM /EBI accession number AB222602), located 245 bp downstream of the initiation codon of Prp19␣, and that its molecular mass is 57,301 Da (Fig. 2C). The chromosomal organization of exons encoding Prp19␣ was determined by alignment of the Prp19␣ cDNA in the mouse genome data base. The Prp19␣ gene was deduced to be composed of 16 exons with lengths ranging from 59 to 174 bp (Fig. 3A). Mouse genome sequence analysis revealed that the inserted sequence of Prp19␤ was derived from intron 3 by alternative splicing at a 3Ј-AG splice site and is located in-frame just before exon 4 (Fig. 3B). Both PRP19␣ and PRP19␤ possess three domains, including a modified RING finger domain, a nuclear localization signal, and WD40 repeat domains (Fig. 3C). To analyze the intracellular localization of proteins, pEGFP-PRP19␣ and pEGFP-PRP19␤ expression vectors were constructed and introduced into P19 cells (Fig. 3D). EGFP-fused PRP19␣ was detected in the nuclei, whereas EGFP-fused PRP19␤ was observed in the nuclei and cytoplasm. The same results were obtained by immunocytochemical analysis with anti-PRP19␣ and anti-PRP19␤ antibodies (data not shown).
Expression Pattern of Prp19␤ mRNA during Neural Differentiation-To examine the expression levels of Prp19␣ and Prp19␤ mRNAs during the neural differentiation of P19 cells, total RNAs were extracted from P19 cells treated with RA for various times and analyzed by real-time PCR. The primers were designed to amplify the specific fragments of two splice variants. The expression of Prp19␣ mRNA was induced shortly after the addition of RA. Maximal expression was observed after 12 h of treatment and was estimated to be 3.0-fold of the original level at 0 h (Fig. 4A, upper panel). The maximal expression level of Prp19␤ was estimated to be 2.5-fold of the expression level of the untreated cells, was observed 1-2 days after RA treatment, and decreased steeply thereafter to the basal level (Fig. 4A, lower panel). The expression levels of PRP19␣ and PRP19␤ proteins were also analyzed by Western blotting using specific antibodies (Fig. 4B). Within 3 h of RA treatment, the expression level of PRP19␣ was augmented, as observed for its mRNA expression level (Fig. 4B, upper panel). In contrast, the expression level of PRP19␤ increased steeply after 1 day of RA treatment and was estimated to be 4.2-fold of the original level at 0 h (Fig. 4B, lower panel). Thus, these results indicated that Prp19␣ gene expression was induced prior to Prp19␤ gene expression.
Using these transfectants, we analyzed the effects of exogenous expression of PRP19␣ and PRP19␤ on the neural differentiation of P19 cells by immunocytochemical analysis (Fig. 5B). ␣-S1 cells could differentiate into ␤-tubulin III-positive neurons after the simple aggregation culture without the addition of RA (Fig. 5, B, panel i; and C). In contrast, V1, ␣-A1, and ␤-S1 cells did not differentiate into neurons in the absence of RA (Fig. 5, B, panels a, e, and m; and C). In the presence of RA, all the transfectants differentiated into neurons (Fig. 5, B, panels b, f, j, and n; and D). Nonetheless, the neuronal differentiation of ␣-S1 cells was significantly stimulated compared with that of V1 cells (Fig. 5D). The neuronal differentiation of ␣-A1 and ␤-S1 cells was decreased conversely.
The effects of PRP19␣ and PRP19␤ on GFAP-positive astroglial cell differentiation were also examined. Although the differentiation of all transfectants to GFAP-positive astroglial cells was not observed without the addition of RA (Fig. 5B, panels c, g, k, and o), all transfectants differentiated into astroglial cells in the presence of RA (panels d, h, l, and p). The differentiation of ␤-S1 cells to astroglial cells induced by RA was significantly increased (Fig. 5, B, panel p; and E), whereas the astroglial differentiation of ␣-A1 cells was strikingly suppressed (Fig. 5, B, panel h; and E).
We further analyzed the effects of exogenous expression of PRP19␣ and PRP19␤ on the expression of neuronal and astroglial cell marker proteins by Western blotting with anti-␤-tubulin III, anti-N-CAM, and anti-GFAP antibodies (Fig. 6). In aggregated ␣-S1 cells, the expression levels of the neuronal markers ␤-tubulin III and 180-kDa neuron-specific N-CAM were markedly induced without any addition of RA (Fig.  6A). The stimulatory effects of the forced expression of PRP19␣ on the neuronal markers in RA-primed ␣-S1 cells were also observed. On the other hand, the expression of neuronal markers in ␤-S1 and ␣-A1 cells was reduced in the presence of RA. In all transfectants, the expression of GFAP, a marker of astroglial cells, was not observed without RA (Fig.  6B). In the presence of RA, the expression levels of GFAP in ␣-S1 and ␤-S1 cells were significantly enhanced. Conversely, the expression of GFAP in RA-primed ␣-A1 cells was decreased.
These results were consistent with the data from immunocytochemical analysis. Thus, it seems that PRP19␣ plays a crucial role in the early stages of neural differentiation of P19 cells, whereas PRP19␤ acts as a stimulator of astroglial cell differentiation and conversely exhibits an inhibitory effect on neuronal differentiation.
Effect of Prp19␤-specific Antisense ODNs on Neural Differentiation-To further confirm the function of Prp19␤ in neural differentiation, we examined the effect of Prp19␤-specific antisense 15-mer ODNs on the neural differentiation of RA-primed P19 cells. When P19 cells were treated with 5 and 10 M antisense Prp19␤ ODNs, the expression level of PRP19␤ protein was reduced to 40 and 30% of the level in untreated control cells, respectively (Fig. 7A). On the other hand, PRP19␣ expression levels in antisense ODN-treated cells were not affected. Immunocytochemical analysis revealed that, in the presence of the antisense ODNs, the neuronal differentiation of RA-primed P19 cells was dosedependently stimulated, whereas astroglial cell differentiation was strikingly suppressed (Fig. 7, B-D). The inhibition observed in the presence of sense ODNs was perhaps due to the nonspecific inhibition of protein synthesis, as observed previously (33).
We also examined the effect of antisense ODNs on the expression levels of neuronal and astroglial cell markers in RA-primed P19 cells by Western blotting. The expression of the neuron-specific markers ␤-tubulin III and 180-kDa neuron-specific N-CAM was increased in antisense ODN-treated cells compared with sense ODN-treated cells (Fig.  7E). In contrast, the expression level of GFAP in antisense ODN-treated cells was down-regulated (Fig. 7F). Thus, these results were consistent with the data from the forced expression of Prp19␤ RNA as described above.
Effect of PRP19␣ and PRP19␤ on Neurogenic Gene Expression-The data described above suggested that the forced expression of Prp19␣ RNA promoted both neurogenesis and gliogenesis, whereas the forced expression of Prp19␤ RNA promoted astroglial (but not neuronal) differentiation. To further explore the molecular mechanisms underlying the neural differentiation regulated by PRP19␣ and PRP19␤, we investigated the expression patterns of neural differentiation-related genes such as oct-3/4, nanog, Mash-1, and GFAP by Northern blotting. The transcription factor Oct-3/4 maintains the undifferentiated state of P19 cells as well as embryonic stem (ES) cells (37). The down-regulation of oct-3/4 mRNA is required for the onset of neural differentiation (16). On the basis of these facts, we analyzed the expression pattern of oct-3/4 mRNA during the neural differentiation of RA-primed V1, ␣-A1, ␣-S1, and ␤-S1 cells (Fig. 8, A, upper panels; and F). The expression level of oct-3/4 mRNA in V1 cells was reduced to 50% of the original level after 28 h of RA treatment and became undetectable after 30 h. In contrast, the expression level of oct-3/4 mRNA in ␣-S1 cells was lower than that in V1 cells and decreased to Ͻ50% of the original level after 20 h of RA treatment. Moreover, in aggregated ␣-S1 cells (but not in aggregated V1 cells), the down-regulation of oct-3/4 mRNA expression occurred within 48 h in the absence of RA (Fig. 8, A, lower panels; and F). The decrease in the expression level of oct-3/4 mRNA was slowed in RAprimed ␣-A1 and ␤-S1 cells, and oct-3/4 mRNA expression was continued by at least 40 h. Aggregated V1, ␣-A1, ␣-S1, and ␤-S1 cells were treated with or without RA for 4 days. A, the expression levels of ␤-tubulin III and N-CAM 4 days after replating were analyzed by Western blotting with the respective antibodies. N-CAMs detected at 140, 160, and 180 kDa are specific markers of astrocytes, astrocytes/neurons, and neurons, respectively (71). B, the expression levels of GFAP 7 days after replating were analyzed by Western blotting with anti-GFAP antibody. In the absence of RA, the expression of GFAP was barely detected in any transfectants.
Nanog is a recently discovered homeodomain transcription factor and is both sufficient and necessary for the maintenance of pluripotency of mouse ES cells and mouse epiblasts, independent of the leukemia inhibitory factor (LIF)/STAT3 (signal transducer and activator of transcription) signaling pathway (22,23). We therefore analyzed the expression level of nanog mRNA during neural differentiation. Northern blot analysis revealed that nanog mRNA expression was induced in all aggregated transfectants regardless of treatment with RA, and its high level of expression was maintained until the onset of neural differentiation (Fig.  8B, upper panels). In RA-primed ␣-S1 cells, the enhanced expression was lowered to a negligible level within 20 h. In contrast, in ␤-S1 cells, substantial expression was detected even at 28 h after RA addition and became undetectable at 32 h. In V1 and ␣-A1 cells, the enhanced expression disappeared after 24 h. In ␣-S1 cells, the down-regulation of aggregation culture-induced nanog mRNA expression began after 20 h and reached a negligible level at 48 h without the addition of RA (Fig. 8, B, lower panels; and J). On the other hand, in the absence of RA, the down-regulation of nanog transcription was not observed in aggregated V1 cells. From these results, it seems likely that PRP19␣ participates in the down-regulation of oct-3/4 and nanog gene expression in the early stages of RA-primed neural differentiation of P19 cells, whereas PRP19␤ suppresses the down-regulation of the expression of these genes. We further analyzed the expression level of neurogenic bHLH genes such as Mash-1, Ngn-1, and NeuroD, which lead to RA-induced neuronal differentiation of P19 cells after the down-regulation of the oct-3/4 and nanog genes (16, 31, 38 -40). Mash-1 mRNA expression in RAprimed ␣-S1 cells was substantially detected after 28 h, when significant down-regulation of oct-3/4 and nanog mRNA expression was observed ( Fig. 8, C, upper panel; and G). In contrast, the induction of Mash-1 mRNA expression was observed after 60 h in RA-primed V1 cells, in which oct-3/4 and nanog gene expression was sustained for a longer time compared with ␣-S1 cells treated with RA. Similar to V1 cells, the induction of Mash-1 mRNA expression in RA-primed ␣-A1 and ␤-S1 cells occurred later than in ␣-S1 cells. The induction of Mash-1 gene expression was observed in aggregated ␣-S1 cells in the absence of RA, corresponding to the down-regulation of oct-3/4 and nanog gene expression (Fig. 8, C, lower panel; and K). The expression of Ngn-1 and NeuroD mRNAs was also transiently induced in all transfectants after 3-5 days of RA treatment (data not shown). However, the period of time required for the maximal expression of both mRNAs was different among these transfectants. Maximal expression of the Ngn-1 and Neu-roD genes in ␣-S1 cells was observed after 3 days of RA treatment and occurred faster than in any other transfectant. Thus, the results support the idea that PRP19␣ participates in the induction of neurogenic gene expression through the down-regulation of oct-3/4 and nanog mRNA expression.
We also analyzed the expression level of GFAP mRNA, a typical astroglial marker, during neural differentiation (Fig. 8, D and H). The substantial expression of GFAP mRNA in RA-primed ␤-S1 cells was already observed after 6 days of RA treatment, and its expression level was much more robust than that of any other transfectant throughout neural differentiation. In RA-primed ␣-A1 cells, the induction of GFAP mRNA expression was detected at 8 days, 1-2 days later than that observed in other cell lines. This delay coincided with a lower level of astroglial differentiation determined by immunocytochemical and Western blot analyses using anti-GFAP antibody as shown in Figs. 5 and 6. Under these conditions, the level of ribosomal phosphoprotein mRNA analyzed as an internal control was not significantly altered, although some bias was observed (Fig. 8E).
Involvement of PRP19␣ in pre-mRNA Splicing-Pre-mRNA splicing is a prerequisite for the expression of most eukaryotic genes. PRP19, CDC5L, and PSF are major components of the spliceosome (32,42). To determine whether PRP19␣ and PRP19␤ participate in neural differentiation through their pre-mRNA splicing functions, we analyzed the spliceosome prepared from RA-primed P19 cells by immunoprecipitation using anti-PRP19␣ and anti-PRP19␤ antibodies. As shown in Fig.  9A, PSF and CDC5L were efficiently precipitated by anti-PRP19␣ (but not anti-PRP19␤) antibody, even though their quantitative changes were not obvious after 24 h of RA treatment. Sepharose 4B gel filtration analysis of the spliceosome also revealed the partial coexistence of PRP19␣ and CDC5L in the higher molecular mass fractions corresponding to 930 kDa (Fig. 9B). After 24 h of RA treatment, the localization of both PRP19␣ and CDC5L was shifted to the lower molecular mass region, showing the possibility that the qualitative change in the spliceosome occurs during neural differentiation. Yeast two-hybrid analysis using PRP19␣ as bait revealed that PRP19␣ could directly bind to CDC5L (amino acids 203-803) via its N-terminal region (amino acids 1-166) (Fig. 9C). Thus, it appears that PRP19␣ (but not PRP19␤) takes part in pre-mRNA splicing during neural differentiation as one of its major functions.
Identification of PRP19␤-specific Binding Proteins-To further analyze the role of PRP19␤ in the process of neural differentiation, cDNA clones encoding mouse PRP19␤-binding proteins were searched using a yeast two-hybrid screening system. Yeast strain Y153 (Leu Ϫ /Trp Ϫ / His Ϫ ) was cotransfected with a bait vector expressing a fusion protein composed of the yeast Gal4 DNA-binding domain and sequence 83-101 of mouse PRP19␤ and with an RA-treated P19 cell cDNA library directing the synthesis of fusion proteins composed of cDNA-encoded proteins and the Gal4 transcriptional activation domain (Fig. 10A). Strain Y153 contains the HIS3 and lacZ genes linked to the GAL4 promoter. Upon transfection of 1.3 ϫ 10 6 Y153 cells, 38 His ϩ clones were developed, and 10 of the 38 clones were ␤-galactosidase-positive. Sequence analysis of the clones showed that three of the 10 clones were novel cDNA clones, and six were identified as known cDNA clones, such as CypA; prothymosin-␣; endometrial bleeding-associated factor; proteasome activator subunit-4; and ribosomal proteins S7, S15a, and S26. Among them, CypA is specifically expressed in neurons and participates in the neuronal (but not astroglial) differentiation of P19 cells (44,45). Therefore, in this study, we focused on the functional analysis of CypA in neural differentiation regulated by PRP19␣ and PRP19␤. Yeast two-hybrid analysis showed that the isolated CypA cDNA encoding amino acids 61-163 specifically bound to the N-terminal region (amino acids 83-101) of PRP19␤, although association with full-length PRP19␤ was not detected (Fig. 10, A and B).
P19 cells were transfected with the pcDNA3-EF1-␣-HA-CypA plasmid vector to express HA-tagged full-length CypA and cultured for 24 h. The cells were then treated with RA for an additional 24 h and subjected to Western blot analysis. Significant expression of HA-tagged CypA and PRP19␤ was observed (Fig. 10C). Immunoprecipitation of the extract from the cells with anti-HA antibody revealed that CypA coprecipitated with PRP19␤. In the absence of RA, coprecipitation was not detected, as in the vacant vector-transfected cells. These results imply that the conformational change in PRP19␤ after RA treatment contributes to the association with CypA.
Involvement of CypA in Neurogenesis-To confirm the participation of CypA in neuronal differentiation, we constructed pcDNA3-EF1-␣-sense CypA and pcDNA3-EF1-␣-antisense CypA expression vectors and introduced them into P19 cells. The resulting stable transfectants were designated CypA-S1 and CypA-A1 cells, respectively. The expression levels of CypA-S1 and CypA-A1 cells were estimated to be 1.32-and 0.68-fold of that of empty vector-transfected V1 cells, respectively (Fig.  11A). In the absence of RA, CypA-S1 cells differentiated into ␤-tubulin A, coexistence of PRP19␣ with the major spliceosome components. Cell lysates were prepared from P19 cells treated with RA for 0 and 24 h and then immunoprecipitated (IP) with anti-PRP19␣ or anti-PRP19␤ antibody. The resulting immunoprecipitates were subjected to 10% SDS-PAGE and analyzed by Western blotting (WB) with anti-PSF and anti-CDC5L antibodies. B, gel filtration analysis of the spliceosome. Cell lysates prepared from P19 cells treated with RA for 0 and 24 h were fractionated by Sepharose 4B column chromatography. Each fraction was analyzed by Western blotting with anti-PRP19␣, anti-CDC5L, and anti-PRP19␤ antibodies to determine the elution profiles of the respective proteins. Blue dextran 2000 (2000 kDa), thyroglobulin (670 kDa), apoferritin (480 kDa), ␥-globulin (160 kDa), and bovine serum albumin (66 kDa) were used as size markers. C, yeast two-hybrid analysis of the interaction between PRP19␣ and CDC5L. Prp19␣ deletion constructs were used to identify the binding region of PRP19␣ with CDC5L by yeast two-hybrid analysis. CDC5L bound to the N-terminal region (N) of PRP19␣ (amino acids 1-166), but not the middle region (M; amino acids 166 -368), the C-terminal region (C; amino acids 368 -504), and the region from M to C (MϩC; amino acids 166 -504). NLS indicates nuclear localization signal. ϩϩ, strong interaction as scored by growth on Leu Ϫ /Trp Ϫ /His Ϫ medium and quantitative ␤-galactosidase assays; ϩ, weak interaction; Ϫ, no interaction.
III-positive neurons after the simple aggregation culture, as observed in ␣-S1 cells (Fig. 11, B, panel d; and C). In the presence of RA, the neuronal differentiation of CypA-S1 cells was enhanced, whereas the neurogenesis of CypA-A1 cells was vigorously suppressed compared with V1 cells (Fig. 11, B, panels e and h; and D). Although the effect of sense CypA on astroglial cell differentiation was not evident in the presence of RA, the appearance of GFAP-positive astroglia in CypA-A1 cells was strongly suppressed (Fig. 11 B, panels f and i; and E). Thus, these results are consistent with a previous report except for the suppression of gliogenesis (45), suggesting that CypA participates in the early stages of neural differentiation.
CypA mRNA Expression during Neural Differentiation-To elucidate the role of CypA in neural differentiation, we investigated the mutual relation of expression patterns among CypA, oct-3/4, otx2, chicken ovalbumin upstream transcription factor I (COUP-TFI), RAR␣, and RAR␤ mRNAs in RA-primed P19 cells (Fig. 12). CypA mRNA expression was maintained at a nearly constant level until 5 days after RA addition and thereafter diminished to half of the original level (Fig. 12A). In contrast, oct-3/4 and otx2 (a neuroectodermal marker) were stimulated within 1 h of RA treatment and reached their maximal levels after 6 h (Fig. 12B). Subsequently, oct-3/4 expression was reduced to the original level at ϳ24 h and was undetectable after 72 h. otx2 expression vanished completely by 24 h of RA treatment. COUP-TFI, RAR␣, and RAR␤ were also analyzed because COUP-TFI represses oct-3/4 promoter activity through an RARE (the RAREoct site), whereas three different RAR/retinoid X receptor (RXR) heterodimers (RAR␣/RXR␣, RAR␤/RXR␣, and RAR␤/RXR␤) specifically bind and activate the oct-3/4 promoter via the RAREoct site in a ligand-dependent manner (46). COUP-TFI expression was gradually stimulated depending upon the reduction of oct-3/4 expression and was enhanced robustly just before the disappearance of the oct-3/4 transcript, suggesting that COUP-TFI plays an important role during the 12-48 h of the late-early stages of neural differentiation. The induction of RAR␣ occurred in a biphasic manner, and the first peak appeared within 1 h of RA treatment. On the other hand, the expression of RAR␤ began after 6 h and reached a peak at 48 h, coinciding with the second peak of RAR␣ expression. Because the sustained up-regulation of oct-3/4 expression is required for neuroectoderm formation in the immediate-early stages of neural differentiation of ES cells (47), it appears that CypA could act as a transcriptional activator for the oct-3/4 gene perhaps through the conformational change in the RAR␣/RXR␣ heterodimer, through which the neuroectoderm formation identified by the transient expression of otx2 within 12 h of the immediate-early stages of RA-primed neural differentiation of P19 cells occurred.
PRP19␤ Suppresses CypA-dependent Activation of RAREoct Promoter Activity-CypA possesses a peptidylprolyl isomerase activity and is a target of the immunosuppressive drug cyclosporin A (44). Therefore, we investigated the effect of CypA on the RAREoct promoter activity using the CypA expression vector and the pLuc-RAREoct reporter plasmid (Fig. 13A). In RA-primed P19 cells, CypA enhanced RAREoct-mediated gene expression in a dose-dependent manner (Fig. 13B, bars  1-3). The activation of gene expression by CypA was completely inhibited by the addition of cyclosporin A (bars 10 and 11), confirming that CypA can activate RAREoct-mediated gene expression. We also examined the effect of PRP19␤ on the CypA-dependent RAREoct promoter activity. RAREoct-mediated gene expression by CypA was decreased by cotransfection with the PRP19␤ expression vector (bars 4 -7), although no effect on the basal level of RAREoct promoter activity by PRP19␤ was observed. To further define the role of RAR␣ and CypA in the activation of the oct-3/4 promoter, we analyzed soluble chromatin from P19 cells treated with or without RA for 6 h by the chromatin immunoprecipitation assay (Fig. 13C). We found that RAR␣ and CypA were present on the oct-3/4 promoter in the absence of RA. After 6 h of RA treatment, robust acetylation of histone H3 was observed, suggesting that the recruitment of the transcriptional coactivator is caused by the CypAdependent conformational change in RAR␣, possibly the RAR␣/RXR␣ heterodimer, in the presence of RA. Thus, these results imply that PRP19␤ inhibits CypA-dependent oct-3/4 gene expression via direct binding with CypA.

DISCUSSION
The data present here provide the first insight into the molecular mechanism underlying the involvement of PRP19␣ (a component of the spliceosome) and its splice variant PRP19␤ in neuronal and glial cell fate decisions in different manners (schematically illustrated in Fig. 14). The induction of Prp19␣ mRNA expression in RA-primed P19 cells was observed within 3 h, as observed for otx2 mRNA expression (a neuroectodermal marker), whereas Prp19␤ mRNA expression was temporary enhanced after 24 -48 h of RA treatment (Fig. 4). The overexpression of Prp19␣ RNA was sufficient to convert P19 cells into ␤-tubulin III-positive neurons (but not GFAP-positive astrocytes) when the cells were cultured in aggregated form without RA, although the differentiation of both neuronal and astroglial cells was stimulated in the presence of RA (Fig. 5). Conversely, the down-regulation of Prp19␣ by the forced expression of antisense Prp19␣ RNA suppressed RA-primed neural differentiation. Therefore, it appears that PRP19␣ is involved in the immediate-early stages of RA-primed P19 cell neural differentiation, includ-ing neuroectodermal differentiation and neuronal fate decision. In contrast, in the presence of RA, the overexpression of Prp19␤ RNA promoted GFAP-positive astroglial cell differentiation, whereas neurogenesis was impaired. The results were further supported by the experiment using PRP19␤-specific antisense ODNs, by which astrogliogenesis was dose-dependently blocked concomitant with augmentation of neurogenesis in RA-primed P19 cells (Fig. 7). Thus, these results suggest that PRP19␣ plays a crucial role in the immediate-early stages of neural differentiation, whereas PRP19␤ could function as a neuron-glial switch.
Critical levels of oct-3/4 and nanog gene expression are required for the maintenance of pluripotency of ES cells (48). A Ͻ2-fold increase in expression causes differentiation into the endoderm and mesoderm, whereas reduction to Ͻ50% of the normal expression level triggers differentiation into the trophectoderm (49). During stroma cell-derived inducing activity-mediated neurogenesis of ES cells, the up-regulation of oct-3/4 mRNA expression is sustained (47). Suppression of oct-3/4 mRNA expression abolishes neuronal differentiation even after stimulation by stroma cell-derived inducing activity. Likewise, nanog-deficient ES cells lose pluripotency and differentiate into the extraembryonic endoderm lineage (22); in P19 cells, Oct-3/4 protein is also present, but is absent in all differentiated somatic cell types in vivo and in vitro (21). In this study, we found that the constitutive expression of Prp19␣ RNA in P19 cells caused the up-regulation of oct-3/4 and nanog mRNA expression within 12 h of RA treatment and that the expression of these genes was thereafter rapidly down-regulated (Fig. 8, A and B). Furthermore, the induction of the neurogenic bHLH transcription factor Mash-1 in ␣-S1 cells was more rapid and intense than that in V1 cells. Yeast two-hybrid and co-immunoprecipitation analyses showed that PRP19␣ could directly bind to the spliceosome components PSF and CDC5L (Fig. 9). These results support the idea that, downstream of the signal evoked by a simple aggregation culture, PRP19␣ participates in the down-regulation of genes that maintain the multipotency of P19 cells.
The neuronal differentiation of P19 cells is dependent on cell aggregation during the induction period (36). Conditions that reduce adhesion in the aggregates inhibit the formation of neurons (50). Although aggregation is clearly essential, little is known about the mechanism by which it promotes differentiation. Recently, Teramoto et al. (51) reported that Wnt-1, the vertebrate counterpart of the Drosophila wingless gene encoding a secreted protein, is up-regulated by aggregation alone, whereas Id-1 (inhibitor of DNA binding-1) is down-regulated. Wnt-1 signals through a receptor complex composed of members of the Frizzled and low density lipoprotein-related protein families and activates the ␤-catenin/transcriptional T-cell factor signaling pathway (known as the canonical Wnt pathway) (52). The overexpression of Wnt-1 directs P19 cells to differentiate into neurons, but not astrocytes (25), perhaps via the enhancement of Ngn-1 gene expression (52). On the other hand, the Id-1 gene encodes an HLH factor to form a heterodimer with bHLH factors and inhibits the DNA-binding function of bHLH factors (53,54). Consequently, the reduced expression of Id-1 may activate the neurogenic bHLH factors (including Mash-1) to aid in the neuronal differentiation of P19 cells. Thus, aggregation is important for the regulation of genes involved in the neural fate determination of P19 cells and may provide a cellular environment necessary for efficient PRP19 activity. Presumably, the forced expression of Prp19␣ RNA can only partially substitute for the function of RA, and cell aggregation is still needed to initiate the neuronal differentiation of P19 cells.
As in neurogenesis in vivo, neurons appear earlier than astroglial cells during RA-primed P19 cell neural differentiation (36). Neurogenic bHLH transcription factors promote neuronal determination (55,56). Comparable bHLH factors that promote the formation of astrocytes have not been isolated. Nonetheless, recent studies have shown that changes in the expression levels of neurogenic bHLH genes might be an important event in the switch from neurogenesis to gliogenesis. In Mash-1/Math-3 or Mash-2/Ngn-2 double knockout mice, neurogenesis is blocked at the precursor stage, and astrocytes are excessively produced (20,57). Ngn-1 promotes neurogenesis via the Ngn-1⅐cAMPresponsive element-binding protein-binding protein⅐p300 complex and inhibits astrocyte differentiation by sequestering the cAMP-responsive element-binding protein-binding protein⅐Smad transcription complex away from astrocyte differentiation genes and by inhibiting the activation of STAT transcription factors (58). In Drosophila, neuronal and longitudinal glial fates are determined by the gcm gene. In gcm-deficient flies, glial cells are converted to neurons, whereas ectopic expression of gcm in neurons causes neuron-glial transformation (1,59). Two mammalian homologs (Gcm1 and Gcm2) have been identified (2)(3)(4)(5)(6). However, Gcm1 exhibits only the potential to induce gliogenesis, but may function in the generation of a minor subpopulation of glial cells (10). Thus, in mammals, the genes that might control a GCM-like switching mechanism have not yet identified.
In this situation, we found that PRP19␤ could function as a neuronglial switch. How does PRP19␤ contribute to a glial cell fate decision? Yeast two-hybrid analysis showed that CypA is one of the PRP19␤specific binding partners. The overexpression of CypA RNA in P19 cells was no longer required for RA treatment to initiate neurogenesis, but not gliogenesis. RAREoct-mediated reporter assay indicated that CypA could function as a transcriptional activator through the conformational change in the liganded RAR/RXR heterodimer and that its activity was repressed by PRP19␤ via binding with CypA on the oct-3/4 promoter (Figs. 13 and 14). Collectively, it is possible that, within 12 h of the immediate-early stages of neuronal differentiation defined by the transient expression of the neuroectodermal marker otx2, the sustained enhancement of oct-3/4 expression by CypA was essentially required for neurogenesis and that the expression of oct-3/4 was subsequently reduced by the following transient induction of PRP19␤ expression (Figs. 12 and 14). It has also been reported that, in P19 cells, knockdown of CypA by RNA interference causes loss of RA-primed neuronal differentiation, but not dimethyl sulfoxide-induced mesodermal differentiation, and a marked reduction of RA-induced RARE-mediated reporter activity, showing that CypA promotes RARE-mediated neurogenic gene expression (45). Thus, the enhancement of astrocytogenesis by the forced expression of Prp19␤ RNA might be due at least in part to the inhibition of neurogenesis.
LIF-induced Jak (Janus kinase)/STAT signaling is a critical part of the astrogliogenic machinery (15,60). Mouse knockout studies have demonstrated a genetic deficiency in major components of this pathway, including LIF, its receptors LIF receptor-␤ and gp130, and the signaling molecules STAT1 and STAT3 (61)(62)(63). In addition, factors that promote astrogliogenesis, including BMPs, basic fibroblast growth factor, and Notch signaling, all require pre-activation of the Jak/STAT pathway (64 -67). Furthermore, the forced expression of the EGF receptor in early progenitors induces precocious astrocyte differentiation in response to LIF by increasing STAT3 protein levels (68), showing that, during the neurogenic period, the Jak/STAT pathway is only weakly activated. Recently, He et al. (41) reported a positive autoregulatory loop of the Jak/STAT pathway, by which STAT1/3 directly activates the expression of many components of this pathway. This feed-forward loop provides a positive mechanism to direct the progenitor stem cells toward astrogliogenesis, showing that sufficient activation of the Jak/ In the immediate-early stages of neural differentiation of RA-primed P19 cells, Prp19␣ gene expression is up-regulated to induce oct-3/4 gene expression, by which P19 cells are committed to neuroectodermal lineage cells. In the late-early stages, PRP19␣ differentiates neuroectodermal cells to NS cells through the COUP-TFI-dependent down-regulation of oct-3/4 mRNA expression. PRP19␤ also participates in this step. Thereafter, PRP19␣ promotes neuronal differentiation in cooperation (at least in part) with Wnt-1, whose expression could be up-regulated by cell aggregation alone. In contrast, PRP19␤ inhibits neurogenesis and inversely stimulates astrogliogenesis during the induction of neuronal and glial/astroglial progenitor cells from the NS cells, respectively. Presumably, PRP19␤ promotes astrogliogenesis via stimulation of the Jak/STAT pathway in cooperation with the signals evoked by autocrine/paracrine factors, including BMPs. In the immediate-early stage, CypA enhances the RAREoct promoter activity via the activation of the liganded RAR/RXR heterodimer and leads to the differentiation of ES cells to neuroectodermal cells. In the late-early stage, PRP19␤ inhibits oct-3/4 gene expression through direct binding with CypA. Thereafter, the RAR⅐RXR complex existing on the RAREoct promoter is replaced with the COUP-TFI homodimer, which conducts the complete suppression of oct-3/4 gene expression. FIGURE 13. Inhibitory effect of PRP19␤ on CypA-dependent oct-3/4 promoter activity. A, schematic illustration of the luciferase (Luc) reporter plasmid linked to the RAREoct promoter. The promoter region of oct-3/4 (Ϫ412 to ϩ39) was introduced into the pTRE2-Luc vector, yielding pLuc-RAREoct. The recognition sequence ofRARintheoct-3/4promoter(Ϫ48toϪ28)isalsoindicated (46). B, suppression of CypA-dependent RAREoct promoteractivitybyPRP19␤.P19cellsweretransfected with 1 g of pLuc-RAREoct and 0.5 g of pcDNA3.1/Myc-His/LacZtogetherwithatotalof1gof pcDNA3-EF1-␣-sense CypA, pcDNA3-EF1-␣-sense Prp19␤, and empty pcDNA3-EF1-␣ expression vectors in various combinations by lipofection. After 24 h of transfection, the cells were treated with RA for 24 h, and luciferase activities in the cell lysates were determined according to the recommendations of Promega Corp. and normalized based on ␤-galactosidase activity as an internal control. The effect of cyclosporin A (CsA) was also determined. Each value is the mean Ϯ S.E. of triplicate culture dishes. *, p Ͻ 0.002 compared with 0.25 g ofCypAexpressionvector;**,pϽ0.002comparedwith 0.5 g of CypA expression vector. C, occupancy of the oct-3/4 promoter by RAR␣ and CypA. Soluble chromatins were prepared from P19 cells treated with or without RA for 6 h and immunoprecipitated with anti-RAR␣ or anti-CypA antibody. The DNAs extracted from the immunoprecipitates were amplified using primer pairs covering the oct-3/4 promoter (Ϫ412 to ϩ39). The acetylation of histone H3 (AcH3) was simultaneously analyzed.
STAT pathway function is a switch for transition from neurogenesis to gliogenesis. During the neural differentiation of aggregated P19 cells in the presence of RA, EGF receptor and BMP4 mRNA expression is markedly up-regulated (25,69). We observed that both aggregation and RA treatments were required for the astroglial cell differentiation of ␤-S1 cells, in which sense Prp19␤ RNA was overexpressed (Fig. 5). In a preliminary experiment, we also investigated the phosphorylation level of STAT3 during the RA-primed neural differentiation of ␤-S1 cells using anti-Tyr(P) 705 STAT3 antibody. The phosphorylation of STAT3 in ␤-S1 cells could be detected after 3 days of RA treatment, when the phosphorylation in the other transfectants, including V1, ␣-A1, and ␣-S1 cells, could not (data not shown). Thus, these data strongly imply that only PRP19␤ overexpression is insufficient for the functional activation of the Jak/STAT pathway. Probably, PRP19␤ participates in the sufficient activation of the Jak/STAT pathway in cooperation (at least in part) with the signals evoked by LIF and BMPs to direct P19 cells toward astrogliogenesis.
P19 cells possess many properties similar to those of ES cells isolated from mice and humans (70). Therefore, it may be possible that the expression system of PRP19␣ can be utilized for the production of large amounts of neurons from human ES cells in combination with the down-regulation of PRP19␤.