The cAMP Pathway Regulates Both Transcription and Activity of the Paired Homeobox Transcription Factor Phox2a Required for Development of Neural Crest-derived and Central Nervous System-derived Catecholaminergic Neurons*

Pluripotent neural crest (NC) cells differentiate to diverse lineages, including the neuronal, sympathoadrenal lineage. In primary NC cultures, bone morphogenetic protein 2 (BMP2) requires moderate activation of cAMP signaling for induction of the sympathoadrenal lineage. However, the mechanism by which cAMP signaling synergizes with BMP2 to induce the sympathodrenal lineage is unknown. Herein, we demonstrate that moderate activation of cAMP signaling induces both transcription and activity of proneural transcription factor Phox2a. In NC cultures inhibition of cAMP-response element-binding protein (CREB)-mediated transcription by expression of dominant-negative CREB suppresses Phox2a transcription and sympathoadrenal lineage development. Interestingly, the constitutively active CREBDIEDML, despite inducing Phox2a transcription, is insufficient for sympathoadrenal lineage development, requiring activation of the cAMP pathway. Because CREBDIEDML-mediates cAMP-dependent transcription without requiring activation by the cAMP-dependent protein kinase A (PKA), these results identify PKA activation as necessary in sympathoadrenal lineage development. Treatment of NC cultures with the PKA inhibitor H89 or 1-10 nm okadaic acid (OA), a serine/threonine PP2A-like phosphatase inhibitor, suppresses sympathoadrenal lineage development. Likewise, OA treatment of the CNS-derived catecholaminergic CAD cell line inhibits cAMP-mediated neuronal differentiation. Specifically, OA inhibits cAMP-mediated Phox2a dephosphorylation, cAMP-dependent Phox2a DNA binding in vitro, and cAMP- and Phox2a-dependent dopamine-β-hydroxylase-luciferase reporter expression. Together, these results support cAMP-dependent Phox2a dephosphorylation is required for its activation. We conclude that moderate activation of cAMP signaling has dual inputs in catecholaminergic, sympathoadrenal lineage development; that is, regulation of both Phox2a transcription and activity. These results provide the first mechanistic understanding of how moderate activation of the cAMP pathway in synergy with BMP2 promotes sympathoadrenal lineage development.

The neural crest (NC) 3 is a pluripotent cell population derived from the lateral ridges of the neuroepithelium during closure of the neural tube (1). NC cells migrate along defined paths in the developing embryo, generating diverse cell types (2). The sympathoadrenal (SA) lineage originating from the trunk region of the neural tube (1) in response to instructive, microenvironmental signals develops to sympathetic neurons and chromaffin cells of the adrenal medulla (3). Cells committed to the SA lineage are characterized by expression of tyrosine hydroxylase (TH), the rate-limiting enzyme in catecholamine biosynthesis, and dopamine-␤-hydroxylase (DBH), catalyzing the conversion of dopamine to norepinephrine (3).
Bone morphogenetic proteins (BMPs) produced by the developing aorta (4 -9) and unidentified signals originating from the notochord and ventral neural tube (10 -14) are required for SA cell development. However, in murine (15) and avian (16,17) NC cultures, SA cell development requires not only BMP2 but also moderate activation of cAMP signaling. Importantly, our earlier studies (16) demonstrated moderate activation of cAMP signaling synergizes with BMP2 in promoting development of the SA lineage. However, the molecular mechanism by which cAMP signaling in synergy with BMP2 promotes SA cell development is unknown.
The cAMP pathway (18), via the cAMP-dependent protein kinase A (PKA), activates the transcription factor CREB by Ser-133 phosphorylation (19,20). Phospho-CREB interacts with the coactivator CBP/p300, mediating cAMP-dependent transcription (21). CREB null mice (22) demonstrate CREB-dependent, neurotrophin-mediated survival and growth of peripheral neurons as well as a 65% CREB-dependent reduction in superior cervical ganglia sympathetic neuron development, occurring before neurotrophin dependence. These results implicate CREB in sympathetic neuron development by an unknown mechanism. Herein, we investigate how cAMP signaling promotes SA cell development by examining the role of CREB and PKA activation in development of NC cells to the SA lineage.
Phox2a Ϫ/Ϫ mice lack the locus ceruleus, a major catecholaminergic center in the central nervous system (CNS), but Phox2b expression and sympathetic neuron development is largely normal (30). In Phox2b null embryos, NC cells migrate to the dorsal aorta and express MASH1, but catecholaminergic neurons fail to develop (31). In vivo Phox2a or Phox2b overexpression generates additional noradrenergic and cholinergic neurons (33). By contrast, in vitro in primary NC cultures Phox2a overexpression induces TH expression and SA cell development only with activation of cAMP signaling (15). Thus, these studies (15) underscore the importance of the cAMP pathway in Phox2a action and SA cell development. cAMP signaling is also required for Phox2a-mediated transcription from the TH-and DBH-luciferase reporters studied in NC-derived PC12 cells and CNS-derived catecholaminergic Cath.a cells (34 -36). Thus, cAMP signaling is important in Phox2a activation, studied by employing Phox2a-responsive reporter constructs in established cell lines. However, the mechanism by which cAMP signaling induces development of primary NC cells to the SA lineage and, specifically, the role of cAMP signaling in Phox2a regulation in NC cells is unknown.
Herein, we examine the role of the cAMP pathway in catecholaminergic cell development employing avian primary NC cultures, in conjunction with the CNS-derived catecholaminergic Cath.a cell line (37). This cell line provides another cellular model for comparative studies of NC-derived and CNS-derived catecholaminergic neuron development. The Cath.a cell line originated from brain tumors of transgenic mice expressing the SV40 T-antigen under control of the TH promoter (37). Cath.a cells, like catecholaminergic SA cells, are characterized by expression of TH and DBH, synthesis of catecholamines, and development of neurites. The CAD cell line used herein, a variant of the Cath.a cell line, undergoes neuronal differentiation characterized by expression of TH and development of neurites (17,38), in response to serum withdrawal (38) or signals activating the cAMP pathway (17).
Herein, we demonstrate the molecular mechanism and the essential role of the cAMP pathway in catecholaminergic, sympathoadrenal lineage development; the cAMP pathway has dual inputs in sympathoadrenal cell development, namely, transcriptional induction of proneural Phox2a by CREB and regulation of Phox2a activity by PKA.

MATERIALS AND METHODS
Neural Crest Cell Cultures-Primary cultures of NC cells were prepared from J. quail embryos, stage 14 -15 (39), as described (16). Briefly, 48-h embryo explants from trunk region of neural tube were cultured for 42 h. After removal of neural tube (Fig. 1A, day 0), the remaining cells are NC cells, assessed by HNK1-positive immunostaining. Mass cultures of NC cells (day 0) were harvested by trypsin treatment and replated at a density of 320 cells/mm 2 (16). Clonal cultures of NC cells (day 0) were replated at density of 300 cells/35-mm dish in dishes coated with Vitrogen 100 (Collagen Corp.) and bovine fibronectin (20 g/ml).
Transient Transfections-DBH-luciferase reporter (50 ng) co-transfected with Phox2a (36) expression vector (500 ng) in NC or CAD cells by the FuGENE 6 method (Roche Applied Science). Cells transfected for 24 h were pretreated (30min.) with 10 -100 nM OA followed by the addition (2 h) of BMP2 and IBMX (36). After the 2-h treatment, cells were assayed for luciferase activity. Luciferase activity was normalized per g of protein extract.
Electrophoretic mobility shift assays were performed as described (42) employing 20 g of nuclear extract (42) isolated from CAD cells treated with BMP2 and IBMX (24 h) followed by treatment (6 h) with 10 nM OA. 32 P-Radiolabeled oligonucleotide probes containing the Phox2a cis-acting elements of DBH promoter were described in Adachi and Lewis (36).
Real-time PCR-Total RNA was isolated by the Trizol method (Invitrogen) from NC cultures (grown in 35-mm dishes) or CAD cells. cDNA (20 l) was synthesized from 2 g of RNA. cDNA (2 l) was used in quantitative real-time PCR as described in detail in Lee et al. (51).

RESULTS
cAMP Signaling in Synergy with BMP2 Regulates Phox2a mRNA Transcription-Primary murine (26) and avian (16) NC cultures require moderate activation of the cAMP pathway in synergy with BMP2 for SA lineage development. However, the mechanism by which cAMP signaling promotes SA cell development is unknown. To determine whether the cAMP pathway participates in SA cell development by regulating cAMP-responsive, CREB-mediated transcription, we expressed in primary NC cultures a dominant negative (44) or a constitutively active (52) variant of CREB, employing avian RCAS retroviral vectors (48). RCAS vectors were also used encoding the wild type (WT) Ad2 12 S E1A, which interferes with CBP function required for CREB-mediated transcription. The inactive N-terminal deletion ⌬ 2-32 E1A was used as negative control (45,46).
To establish the infection kinetics of the RCAS viruses and the expression of the foreign proteins, a time course was performed employing NC cultures infected at day 0 (Fig. 1A). Western blot analyses of cellular extracts isolated from virus-infected NC cultures demonstrate the expression of the foreign proteins (Fig. 1B). Immunofluorescence microscopy at 24, 48 (Fig. 1C), and 72 h (data not shown) postinfection demonstrates that nearly 100% of the cells express the foreign proteins.
Moderate activation of cAMP signaling by treatment with the cAMPelevating agents IBMX (100 M) or forskolin (0.1-1.0 M) synergizes with BMP2, inducing expression of Phox2a (16). Herein, employing quantitative real-time PCR, we quantified Phox2a expression after treatment with IBMX in combination with BMP2. In comparison to IBMX, co-treatment with BMP2 and IBMX mediated a synergistic more than additive induction in Phox2a mRNA expression ( Fig. 2A), suggesting the cAMP pathway together with BMP2 is involved in Phox2a transcription. Importantly, NC cultures treated with either IBMX or a low concentration of forskolin (0.1-1.0 M) display CREB phosphorylation required for cAMP-mediated transcription ( Fig. 2A).
To demonstrate the transcriptional involvement of the cAMP pathway in Phox2a expression, we examined the effects of dominant negative ACREB (44) and constitutively active CREB DIEDML (52) via infection of NC cultures with the respective RCAS viruses. Infection of NC cultures at day 0 ( Fig. 1A) with the control RCAS virus in the presence of both BMP2 and IBMX results in a progressive increase of Phox2a mRNA at day 2, reaching a maximal level by day 3 (Fig. 2B). By comparison, infection with RCAS-ACREB virus in the presence of BMP2 and IBMX interferes with Phox2a mRNA expression, reducing its expression by nearly 10-fold (Fig. 2B). Infection with RCAS-CREB DIEDML virus in the presence of BMP2 mediates Phox2a mRNA induction (Fig. 2C), comparable with that with BMP2 and IBMX (Fig. 2B). Because CREB DIEDML is transcriptionally active (52) without activation of the cAMP pathway, these results ( Fig. 2C) directly demonstrate the transcriptional involvement of the cAMP pathway in Phox2a mRNA expression via active CREB.
The co-activator CBP is required for CREB-mediated transcription (53). WT E1A represses CREB-dependent CBP function via direct interaction with CBP, whereas ⌬ 2-32 E1A does not interact with CBP and does not interfere with CBP function (45,46). Accordingly, we examined the effect of WT E1A and its inactive ⌬ 2-32 mutant on Phox2a mRNA expression in NC cultures infected with the respective viruses. WT E1A reduces Phox2a mRNA by day 3, reaching a 50% reduction by day 4. By contrast, the ⌬ 2-32 E1A mutant has no noticeable effect (Fig.  2D). Although the mechanism of this E1A-mediated inhibition of

The cAMP Pathway in Catecholaminergic Neuron Development
Phox2a mRNA expression has not been further investigated, the results support CBP involvement in Phox2a expression.
ACREB and WT E1A Repress the SA Phenotype-SA cells express TH, the rate-limiting enzyme in catecholamine biosynthesis, and DBH, the last enzyme required for the synthesis of catecholamines (CA). Thus, detection of the synthesized and stored catecholamines by histofluorescence (43) constitutes a definitive marker of SA cell development.
NC cultures infected with RCAS-ACREB, RCAS-WT E1A, and RCAS-⌬ 2-32 E1A viruses were examined for the appearance of the SA phenotype at day 6 of NC culture (Fig. 1A). SA cell markers assayed include both TH expression monitored by immunofluorescence microscopy and catecholamine biosynthesis by histofluorescence. ACREB or WT E1A interfered with the appearance of both TH-immunoreactive and catecholamine-positive cells in response to BMP2 and IBMX (Fig. 3A). By contrast, the control virus or expression of the inactive ⌬ 2-32 E1A had no effect (Fig. 3A). Similar to Fig. 1C, parallel immunostaining experiments employing the M2 FLAG antibody, or the E1A antibody confirmed expression of these foreign proteins (data not shown). Because ACREB and WT E1A interfere with CREB and CBP, respectively, their inhibitory effect on the appearance of TH-immunoreactive and catecholamine-positive cells supports the direct transcriptional involvement of the cAMP pathway in SA cell development. CREB DIEDML Is Insufficient for SA Lineage Development-NC cultures infected with the RCAS-CREB DIEDML virus in the presence of BMP2 alone did not increase the appearance of TH-and catecholamine-positive cells relative to cultures infected with control RCAS virus (Fig. 3B). This result was surprising because the expression of CREB DIEDML with BMP2 did increase Phox2a mRNA expression (Fig.  2C). Interestingly, in CREB DIEDML -expressing cultures, IBMX and BMP2 enhanced the appearance of both TH-and catecholamine-positive cells (Fig. 3B). Conversely, inhibition of PKA by the addition of the PKA inhibitor H89 (10 M) reversed the induction of the SA phenotype (Fig. 3B). Because increased Phox2a expression in the absence of PKA activation is insufficient for SA lineage development (Figs. 2C and 3B), we conclude the 1) CREB is necessary but insufficient in SA cell development and 2) cAMP signaling via PKA activation regulates additional events necessary for SA cell development.
To confirm these results we quantified TH protein expression by Western blot assays (16,54). Consistent with our earlier observations (16), BMP2 and IBMX co-treatment of NC cultures infected with control RCAS virus resulted in a 2.6-fold increase in TH protein (Fig. 4 CREB and PKA Activation Are Necessary and Instructive in SA Lineage Development-The effect of microenvironmental signals on NC cell differentiation is either selective, by influencing cell proliferation and survival of a committed progenitor, or instructive, restricting the pluripotent nature of NC cells. Our earlier studies in NC cells (16) support that the influence of cAMP signaling on SA cell development is instruc-   DECEMBER 9, 2005 • VOLUME 280 • NUMBER 49 tive (5). To gain further understanding of the mechanism by which CREB and E1A affect development of SA cells, we studied at clonal density NC cell survival and development (TABLE ONE). The total number of colonies generated in the assay is similar in all treatment groups, supporting that viral infection and expression of ACREB, CREB DIEDML , E1A, and ⌬ 2-32 E1A do not affect survival of the colony founder cells. However, dominant negative ACREB and WT E1A blocked formation of catecholamine-positive colonies, suppressing the SA phenotype. These results demonstrate CREB-mediated transcription is instructive, acting by restricting pluripotent NC cells to the SA lineage.

The cAMP Pathway in Catecholaminergic Neuron Development
On the other hand, NC cultures expressing CREB DIEDML in the presence of BMP2 without activation of the cAMP pathway did not generate catecholamine-positive colonies (TABLE ONE). By contrast, in CREB-DIEDML -expressing clonal NC cultures IBMX mediates a 28-fold increase in catecholamine-positive colonies, in comparison to those grown without IBMX (TABLE ONE). Thus, the affect of ACREB and CREB DIEDML in clonal assays confirms that CREB activation is necessary but insufficient for SA cell development. Furthermore, because CREB-DIEDML with BMP2 alone increased Phox2a mRNA expression ( Fig. 2C) but required IBMX for TH expression and catecholamine biosynthesis (Figs. 3B and 4), the results of TABLE ONE are interpreted to mean that cAMP-dependent PKA activation is a necessary instructive signal for SA lineage development.
A Ser/Thr Phosphatase Is Required for SA Cell Development-To further understand the mechanism by which activated PKA regulates SA lineage development, we based our investigations on the studies by Adachi and Lewis (36). In PC12 cells Phox2a is constitutively phosphorylated in vitro and in vivo; the cAMP pathway promotes Phox2a dephosphorylation, increased DNA binding, and DBH-luciferase reporter expression (36). In CAD cells cAMP signaling regulates Phox2a-dependent DBH-luciferase reporter expression, which is inhibited by OA, an inhibitor of Ser/Thr phosphatases (36). Because in NC cells Phox2a is required for SA lineage development, we investigated whether the cAMP pathway regulates Phox2a activation via the same mechanism. Accordingly, NC cultures were induced to the SA lineage by BMP2 and IBMX as a function of OA.
In intact cells 1-10 nM OA specifically inhibits the Ser/Thr phosphatase PP2A (55) and PP2A-like phosphatases (56). Continuous 6-day treatment of NC cultures with 1 nM OA significantly decreased TH protein expression (Fig. 5A, compare lane 4 to lane 5). To exclude an OA effect on NC cell survival, we investigated cell survival and development of NC cultures plated at clonal density (TABLE TWO). Continuous treatment with 1 nM OA or 2 M H89 did not alter the total number of colonies formed, whereas each inhibitor decreased by 50% the appearance of TH-immunoreactive (data not shown) and catecholamine-positive colonies (TABLE TWO). This observation suggests that PP2A or a PP2A-like phosphatase, which can be inhibited by 1 nM OA, is activated by cAMP signaling, in turn activating a key regulator of SA cell development.
Continuous 6-day treatment with 1 nM OA could affect many cellular targets. Therefore, we examined the OA effect on SA lineage development after a short term treatment, i.e. at an interval when the activity of proneural transcription factors is required to initiate SA lineage development. The rationale for this approach is based on our recent studies (57) demonstrating that inhibition of the SA lineage-determining factor ASH1 within 24 h of secondary NC culture (Fig. 1A) is sufficient to abrogate development of the SA lineage. These observations (57) illustrate the principle that in the dynamic developmental system of primary cultures of pluripotent NC cells the interval of expression and activation of lineage-regulating transcription factors is crucial in the determination of a cell lineage.
Accordingly, NC cultures were treated for 3 h with 10 nM OA (Fig.  5B). OA was added to the secondary NC cultures on day 0, 4 h after replating, or on day 1 (Fig. 1A). To exclude the OA effect(s) on endogenous CREB activity, NC cultures were treated with BMP2 and IBMX and infected with RCAS-CREB DIEDML -expressing virus to ensure Phox2a expression (Fig. 2C). OA added for 3 h on day 0 of secondary NC culture decreased TH protein expression by 50% (Fig. 5B, compare lane 2 to lane 3). The addition of 10 nM OA for 3 h on day 1 of secondary NC culture mediates a smaller but reproducible decrease in TH expression (Fig. 5B, compare lane 5 to lane 6). This effect of OA on day 1 becomes more pronounced with increased duration (24 h) of treatment (Fig. 5C,  compare lane 1 to lane 4). Furthermore, 24 h of OA treatment on day 1 reduced TH mRNA by ϳ30% without a significant effect on Phox2a mRNA (data not shown).
Because the inhibitory OA effect on TH expression and SA cell development occurs between 4 and 48 h of secondary NC culture (Fig. 5, B and C), we interpret this result to mean the OA effect is specific, targeting the activity of early acting SA lineage-promoting transcription factor(s). Based on these results (Fig. 5) and those of Adachi and Lewis (36), we reasoned Phox2a is a likely target of the cAMP-mediated, OA-sensitive dephosphorylation required for SA cell development.
A Ser/Thr Phosphatase Is Required for Neuronal Differentiation of the CNS-derived CAD Cell Line-To confirm the role of the OA-sensitive Ser/Thr phosphatase in SA cell development via regulation of Phox2a activity, we employed the CAD cell line (37). CAD cells undergo neuronal differentiation characterized by expression of TH and development

The cAMP Pathway in Catecholaminergic Neuron Development
of neurites in response to serum withdrawal (38) or to cAMP-elevating agents (17). Treatment of CAD cells with BMP2 and IBMX increases CAD cell neuronal differentiation, measured by TH and peripherin immunoreactivity (Fig. 6A). Interestingly, 1 nM OA or 5 nM (data not shown) OA repressed both TH and peripherin immunoreactivity without an affect on Phox2a immunoreactivity (Fig. 6A). To confirm these results, we quantified by real-time PCR Phox2a and TH mRNA expression in CAD cells induced to differentiate by the addition of BMP2 and IBMX as a function of OA (1 nM). BMP2 and IBMX induce transcription of both TH and Phox2a mRNAs. By contrast, 1 nM OA inhibits TH mRNA expression (Fig. 6B) without a significant effect on Phox2a expression (Fig. 6C). These results are also supported by the immunofluorescence data of CAD cells (Fig. 6A), demonstrating inhibition of TH and peripherin immunostaining in the presence of OA without an affect on Phox2a immunostaining. These observations suggest a regulatory role for the OA-sensitive PP2A-like phosphatase, regulating the activity of Phox2a, which in turn mediates TH expression and neurite outgrowth. Amino acid sequence alignment of Phox2a proteins identified putative, evolutionarily conserved protein kinase phosphorylation sites.
These include phosphorylation sites for glycogen synthase kinase 3, extracellular signal-regulated kinase 1/2, PKA, calmodulin kinase II, and protein kinase C (Fig. 7A). Although more detailed studies are required to demonstrate the functional significance of these putative phosphorylation sites in Phox2a, the evolutionary conservation of functionally important regulatory sequences is a well accepted biological phenomenon. Thus, we hypothesized that these conserved Phox2a phosphorylation sites are of regulatory significance and are potential targets of dephosphorylation by cAMP-mediated activation of a PP2A-like phosphatase.
To directly demonstrate that Phox2a phosphorylation is regulated by cAMP signaling and the OA-sensitive PP2A-like phosphatase, we examined Phox2a immunoprecipitates for the presence of phosphorylations at Ser/Thr residues that are likely phosphorylation sites of proline-directed kinases, such as mitogen-activated kinases. Phox2a immunoprecipitates derived from CAD cells grown under control conditions or after treatment with IBMX, IBMX and OA (10 nM), or PD 98059 were immunoblotted with a phospho-Thr/Ser-Pro-specific antibody (Fig.  7B). The Phox2a phosphorylation status demonstrates a statistically significant reduction (p Ͻ 0.05) after treatment with IBMX, and impor-  Effect of H89 and okadaic acid on NC cell growth NC cells were plated at clonal density, as described under "Materials and Methods," and grown for 13 days with BMP2 and IBMX as indicated. CA-positive colonies were identified by histofluorescence (43).  DECEMBER 9, 2005 • VOLUME 280 • NUMBER 49 tantly, this reduction is reversed by OA. Likewise, treatment with PD 98059, a specific inhibitor of MEK1, also reduces Phox2a phosphorylation, confirming the specificity of the phospho-Thr/Ser-Pro-specific antibody and suggesting these phosphorylations include putative S-P phosphorylation sites (Fig. 7A). These results (Fig. 7B) demonstrate that cAMP signaling modulates via an OA-sensitive phosphatase the phosphorylation of Phox2a. Phox2a Dephosphorylation Induces Phox2a DNA Binding in Vitro and Phox2a Transactivation-To determine the functional importance of the cAMP-dependent and OA-sensitive dephosphorylation of Phox2a, we initially determined the interval of OA sensitivity, abrogating CAD cell neuronal differentiation. As shown in Fig. 8A, the addition of 5 nM OA for 12-18 h after 24 h of treatment with BMP2 and IBMX inhibits neuronal differentiation of CAD cells, assessed by immunofluorescence microscopy for peripherin expression. Our current studies 4 have defined that 24 h of BMP2 and IBMX treatment of CAD cells is necessary for cell cycle exit and differentiation. Accordingly, CAD cells were treated with BMP2 and IBMX for 24 h followed by 6 h of treatment with OA. CAD cell nuclear extracts were isolated at 6 h Ϯ OA treatment and utilized in electrophoretic mobility shift assays employing the Phox2a cis-acting elements of the DBH promoter (36). A sequencespecific Phox2a⅐DNA complex is detected in extracts isolated from CAD cells treated with BMP2 and IBMX, based on competition assays with a 30-fold excess of unlabeled WT versus mutant oligonucleotides (Fig. 8B). This sequence-specific complex is supershifted with the Phox2a antibody but not with IgG, thus confirming the identity of the Phox2a⅐DNA complex (Fig. 8B). Importantly, this Phox2a⅐DNA complex is absent or reduced in extracts isolated from control, untreated CAD cells or extracts isolated from CAD cells grown in the presence of OA (Fig. 8B). These observations support that cAMP-dependent Phox2a dephosphorylation is required for Phox2a DNA binding, in agreement with earlier studies (36).

The cAMP Pathway in Catecholaminergic Neuron Development
To directly link the functional importance of the OA effect on Phox2a activity, a Phox2a expression vector (36) was transiently co-transfected in CAD and NC cells with the DBH-luciferase reporter as a function of BMP2 and IBMX treatment with and without OA addition (Fig. 8C). In CAD cells (data not shown) and NC cells (Fig. 8C), OA suppresses BMP2-and IBMX-mediated expression from the Phox2a-dependent 4 M. Paris and O. Andrisani, manuscript in preparation. FIGURE 6. Okadaic acid effect on CAD cell neuronal differentiation. A, phase contrast microscopy and immunofluorescence microscopy of TH, peripherin, and Phox2a of CAD cells grown for 2 days in serum, serum-free medium with or without BMP2 and IBMX, or with BMP2 and IBMX in the presence of continuous treatment with 1 nM OA, as indicated. B and C, real-time PCR analyses of TH (B) and Phox2a (C) mRNAs using RNA from CAD cells grown as indicated for 2 days. Results represent three independent RNA preparations analyzed in PCR reactions using identical reaction triplicates. The asterisk indicates the absence of statistical significance. Data are expressed relative to 18 S rRNA. DAPI, 4,6-diamidino-2-phenylindole.

DISCUSSION
In this study we investigated the molecular mechanism by which the cAMP pathway induces SA lineage development, employing primary cultures of NC cells. We demonstrate CREB activation is necessary in synergy with BMP2 to induce Phox2a transcription. Furthermore, PKA activation via a PP2A-like phosphatase is required for activation of the transcriptional activity of Phox2a and development of the SA lineage. Because transcription mediated by Phox2a is necessary for SA cell development (29 -31), this PKA-dependent transcriptional activation of Phox2a leading to SA lineage development establishes for the first time that moderate activation of cAMP signaling is an essential, instructive signal in SA lineage specification.
Although in vivo gain of function (24,26,33) and loss of function studies (23,30,31) established that BMPs induce MASH1 and Phox2a expression required for SA cell development, overexpression of Phox2a in murine NC cultures in vitro (15) is insufficient for SA lineage development. Activation of the cAMP pathway is necessary to promote TH expression, suggesting involvement of additional, regulatory signaling networks in NC cell differentiation. Herein, our studies provide the first mechanistic evidence of how moderate activation of the cAMP pathway, in synergy with BMP2, promotes SA lineage specification. Specifically, the cAMP pathway has a dual role in NC cells, regulating both transcription and activation of proneural Phox2a. This observation conclusively establishes that moderate activation of the cAMP pathway is an essential, instructive signal in SA lineage development in vivo. ACREB Inhibits Phox2a Transcription and SA Cell Development-In NC cultures, ACREB suppresses Phox2a expression and the appearance of TH-and catecholamine-positive cells without affecting NC cell survival. Accordingly, we conclude cAMP-dependent transcription mediated by CREB is necessary in instructing NC cells to the SA lineage. This conclusion agrees with the phenotype of CREBϪ/Ϫ mice (22), which displays a 65% reduction in superior cervical ganglia sympathetic neuron development. This 65% reduction, as opposed to 100%, which  DECEMBER 9, 2005 • VOLUME 280 • NUMBER 49 occurs earlier than the CREB-dependent, neurotrophin-mediated survival and growth (22), is probably due to the redundancy of CREB/ATF family of proteins. Importantly, our studies extend these observations by identifying CREB as necessary in Phox2a transcription.

The cAMP Pathway in Catecholaminergic Neuron Development
The direct involvement of CREB in transcription from the Phox2a promoter in synergy with BMP2 is supported by our recent studies (58). A cluster of functional CRE half-sites have been identified in proximity to E-box DNA binding sites that are potential binding sites for ASH1. These CRE and E-box sites are evolutionarily conserved in mouse and human Phox2a promoters. Importantly, the Phox2a gene has been identified by the serial analysis of chromatin occupancy (SACO) method (59) as a CREB-regulated gene, in agreement with our findings (58).
CREB DIEDML Induces Phox2a mRNA but Not SA Cell Development-In NC cultures, CREB DIEDML in the presence of BMP2 increased Phox2a mRNA expression, thus, directly demonstrating that activated CREB is necessary for Phox2a mRNA expression. Interestingly, CREB DIEDMLexpressing NC cells, despite the increased Phox2a expression, do not develop to the SA lineage unless the cAMP pathway is activated, in agreement with the Phox2a overexpression studies of Lo et al. (15). Because Phox2a transcription induced by CREB DIEDML is insufficient for SA lineage development unless PKA is activated, our studies establish the PKA dependence in the development of the SA lineage.

Involvement of a Ser/Thr Phosphatase in SA Lineage Development-
In PC12 and Cath.a cells, Phox2a activation, studied via the DBH-driven reporter (36), requires activation of the cAMP pathway and is inhibited by OA. In this study employing pluripotent NC cells, Phox2a activation studied in the context of SA lineage development requires the cAMP pathway mediating activation of an OA-sensitive Ser/Thr phosphatase. This OA-sensitive event does not affect NC cell survival and occurs early in NC cell development, when proneural Phox2a initiates development of the SA lineage. Because low concentrations of OA inhibit the phosphatase 2A family of proteins (55, 56), we propose PP2A or a PP2Alike phosphatase (56) regulates SA lineage specification by dephosphorylating Phox2a.
Interestingly, during retina development in Xenopus, the activity of NeuroD is positively regulated by dephosphorylation at its glycogen synthase kinase 3 phosphorylation site (60). Furthermore, in Drosophila, PP2A in association with the BЈ subunit is essential in positively regulating the activity of the homeodomain protein SCR (61). In addition, PP2A in complex with the B56␦ subunit positively regulates the activity of the basic helix-loop-helix transcription factors HAND1 and HAND2 involved in heart, vascular, neuronal, and extraembryonic development (62). Thus, the precedent exists for a role of PP2A in pos- FIGURE 8. Okadaic acid effect on CAD cell differentiation, Phox2a DNA binding, and transactivation. A, CAD cells were grown with BMP2 and IBMX for 24 h followed by the addition of 5 nM OA for 12 or 18 h as indicated. Differentiation was assessed by immunofluorescence microscopy employing perpherin immunostaining. B, electrophoretic mobility shift assays performed with nuclear extracts (20 g) isolated from CAD cells grown without BMP2 and IBMX (control), with BMP2 and IBMX for 24 h, or with BMP2 and IBMX for 24 h and OA (10 nM) addition for 6 h. 32 P-Radiolabeled probes are Phox2a cis-acting elements of DBH promoter, indicated as HD3 and DB1, according to Adachi and Lewis (36). Shown are wild type (wt) and mutant (mt) oligonucleotides for HD3 and DB1 sites (36) used at 30-fold molar excess as unlabeled competitors. The sequence-specific Phox2a⅐DNA complex and the supershifted complex are indicated by brackets and dots, respectively. C, DBH-luciferase reporter co-transfected with the Phox2a expression vector (36) in NC cells at day 0 of culture. Transfected cells at 24 h were treated without BMP2 and IBMX (control), with BMP2 and IBMX for 2 h, or with BMP2 and IBMX and OA for 2 h, including a 30-min pretreatment with 10 or 100 nM OA as indicated. Results are from three independent assays.
itively regulating the activity of transcription factors involved in differentiation (61) and specifically in NC differentiation (62).
A PP2A-like Phosphatase Regulates Neuronal Differentiation in the CAD Cell Line-The unavailability of avian Phox2a antibody to further study Phox2a modifications during NC cell development prompted us to explore another catecholaminergic cellular model, the murine CAD cells. In CAD cells, BMP2 and IBMX induce neuronal development, monitored by Phox2a mRNA expression, TH expression, and neurite development. OA inhibits TH transcription and neurites development but not Phox2a transcription. Furthermore, IBMX treatment of CAD cells decreases the phosphorylations of Phox2a, whereas OA reverses this effect, indicating that cAMP signaling promotes the dephosphorylation of Phox2a. Likewise, nuclear extracts isolated from CAD cells treated with BMP2 and IBMX demonstrate in electrophoretic mobility shift assays a sequence-specific Phox2a-immunoreactive DNA/protein complex that is absent in control extracts or in extracts isolated from CAD cells treated with BMP2 and IBMX and OA. We conclude the DNA binding potential of Phox2a is regulated via a cAMP-dependent dephosphorylation event involving an OA-sensitive PP2A-like phosphatase, in agreement with earlier observations (36). Furthermore, transient transfections of a Phox2a expression vector (36) with the Phox2aresponsive DBH-luciferase reporter in CAD cells or NC cells demonstrate that whereas cAMP signaling induces DBH-luciferase expression, co-treatment with OA inhibits DBH-luciferase expression. We conclude the transcriptional activity of Phox2a is directly regulated via a cAMP-dependent, OA-sensitive dephosphorylation event. Together, these results confirm, using another catecholaminergic neuronal cellular model, the essential role of the cAMP-regulated, OAsensitive, PP2A-like phosphatase in regulating the transcriptional activity of Phox2a.
In summary, we demonstrate the essential, instructive role of the cAMP pathway in SA cell development (Fig. 9). The transcriptional components of the cAMP pathway, CREB and CBP, induce Phox2a transcription acting instructively in SA cell specification. In addition, the cAMP-dependent PKA regulates the activity of a Ser/Thr PP2A-like phosphatase involved in Phox2a activation. Thus, the cAMP pathway mediates a multilevel control in SA cell development, regulating both Phox2a transcription and activity. We conclude the cAMP pathway is necessary and instructive in SA cell specification.