• neural crest stem cells;
  • protein–protein interaction;
  • sympathetic neuron;
  • transcription factors;
  • tyrosine hydroxylase


  1. Top of page
  2. Abstract
  3. Experimental procedures
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

The zinc finger transcription factor GATA-3 is a master regulator of type 2 T-helper cell development. Interestingly, in GATA-3–/– mice, noradrenaline (NA) deficiency is a proximal cause of embryonic lethality. However, neither the role of GATA-3 nor its target gene(s) in the nervous system were known. Here, we report that forced expression of GATA-3 resulted in an increased number of tyrosine hydroxylase (TH) expressing neurons in primary neural crest stem cell (NCSC) culture. We also found that GATA-3 transactivates the promoter function of TH via specific upstream sequences, a domain of the TH promoter residing at −61 to −39 bp. Surprisingly, this domain does not contain GATA-3 binding sites but possesses a binding motif, a cAMP response element (CRE), for the transcription factor, CREB. In addition, we found that site-directed mutation of this CRE almost completely abolished transactivation of the TH promoter by GATA-3. Furthermore, protein–protein interaction assays showed that GATA-3 is able to physically interact with CREB in vitro as well as in vivo. Based on these results, we propose that GATA-3 may regulate TH gene transcription via a novel and distinct protein–protein interaction, and directly contributes to NA phenotype specification.

Abbreviations used

amino acid


basic region-leucine zipper


chick embryo extract


cAMP response element


cAMP response element binding protein


dopamine β-hydroxylase


Dulbecco's modified Eagle's medium


electrophoretic mobility shift assay


green fluorescent protein


multiplicity of infection




neural crest stem cells


peripheral nervous system


Rous Sarcoma Virus


sympathetic ganglia


sypathetic nervous system


tyrosine hydroxylase


type 2 T helper cells

Vertebrate nervous system development is regulated by a complex regulatory network of extracellular signals and nuclear transcription factors controlling the process of cell fate specification of vast neural subtypes (Marquardt and Pfaff 2001; Goridis and Rohrer 2002). Among the various phenotypes of a particular neuron, neurotransmitter identity is an important feature because it determines the nature of the chemical neurotransmission a given neuron will mediate, and influences its specific connectivity with target cells.

As indicated by the group nomenclature of noradrenaline (NA)-producing cells (i.e. group A1–A7) in the CNS, the NA cell type is one of the earliest neurotransmitter systems to be defined by neuroscientists (Cooper et al. 1996). In the CNS, NA is mainly produced in the locus coeruleus (A4 and A6 cell groups) by sympathetic ganglia (SG) in the peripheral nervous system (PNS). The regulatory cascade controlling the NA neurotransmitter phenotype has been extensively studied, leading to the identification and functional characterization of critical signaling molecules and transcription factors. For example, bone morphogenic proteins (BMPs) are essential for sympathetic nervous system (SNS) development and induce the expression of the proneural gene, Mash1 (or Cash1) (Brunet and Pattyn 2002; Goridis and Rohrer 2002; Howard 2005). Thus, Mash1 is the first transcription factor shown to be essential for NA neuron development (Guillemot et al. 1993). Downstream of Mash1 lies the homeodomain transcription factor, Phox2a, which is a critical regulator of NA cell lineage development (Morin et al. 1997; Guo et al. 1999; Lo et al. 1999; Stanke et al. 1999). The closely related transcription factor, Phox2b, is also induced by BMPs independently of Mash1 and is another essential regulator of NA neuron development (Pattyn et al. 1999).

A potentially new player in NA cell fate specification is the zinc finger transcription factor, GATA-3. GATA-3 was originally isolated as a T cell-specific DNA binding factor. Extensive in vivo and in vitro studies demonstrated that GATA-3 is essential for the development of T cell lineage and is a master regulator for the differentiation of type 2 T helper (Th2) cells (Murphy and Reiner 2002). GATA-3–/– embryos die between days 11 and 12 post coitum (d.p.c.) and display massive internal bleeding, marked growth retardation, severe deformities of the brain and spinal cord, and gross fetal liver haematopoietic aberrations (Pandolfi et al. 1995). Interestingly, a re-evaluation of the GATA-3–/– mouse showed that embryonic lethality was partially averted by feeding catechol intermediates, suggesting that NA deficiency is a proximal cause of death (Lim et al. 2000). Consistent with this, GATA-3 null mutation led to reduced accumulation of TH and dopamine β-hydroxylase (d.b.h.) mRNAs in the SNS, whereas several other SNS genes were unaffected (Lim et al. 2000). Furthermore, Tsarovina et al. (2004) recently re-investigated the GATA-3–/– mouse and confirmed that TH and d.b.h. expression is affected in this mouse, albeit to a somewhat different degree. This study demonstrated a more global effect of GATA-3 on sympathetic development, including the reduced size of the sympathetic chain (Tsarovina et al. 2004). Together, these in vivo findings suggest significant roles for GATA-3 in the development of the SNS. However, important questions regarding the function of GATA-3 remain unanswered, e.g. does GATA-3 directly regulate the induction of the NA neurotransmitter phenotype, are NA-synthesizing genes, such as TH and d.b.h., immediate target genes of GATA-3 in SNS development, and by what mechanism does GATA-3 control its target genes. To address these questions, we performed a series of experiments. Here, we show that overexpression of GATA-3 resulted in an increase of NA neurons among avian NCSCs in vitro, supporting the notion that GATA-3 has a direct role in NA neurotransmitter phenotype specification. Furthermore, we found that GATA-3 directly transactivates the promoter function of the TH gene, strongly suggesting that TH may be an immediate target of GATA-3. Surprisingly, our promoter analysis and protein interaction assays demonstrated that GATA-3 regulates the TH gene promoter via a distinct protein–protein interaction with the known transcription factor, CREB. Taken together, our results suggest the possibility that GATA-3 controls TH gene transcription via a novel and distinct protein–protein interaction.

Experimental procedures

  1. Top of page
  2. Abstract
  3. Experimental procedures
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

NCSC cultures

Primary culture of the trunk region of quail eggs was performed as described (Bilodeau et al. 2000) with some modifications. Fertilized Japanese quail (Conturnix japonica) eggs (CBT farms, Chestertown, MD, USA) were incubated for 48 h at 38°C in a humidified incubator (G. Q. F. Manufactory Co., Savannah, GA, USA). Neural tubes and associated structures at the thoracic level were dissected from the embryos and treated with dispase II (4 mg/mL; Roche, Indianapolis, IN, USA). The purified neural tubes were plated in dishes coated with fibronectin (40 µg/mL; Sigma, St. Louis, MO, USA) and cultured in Dulbecco's modified Eagle's medium (DMEM)/F12 medium containing 15% horse serum and 10% chick embryo extract (CEE) which was collected from embryonic day 9 (E9) to E11 white hens' egg embryos. After 24 h in culture, neural tubes were removed from the culture dish using an electrolytically sharpened tungsten needle. Brief treatment with 0.25% trypsin (Invitrogen, Carlsbad, CA, USA) detached migrating NCSC that were plated at a density of 1.8 × 104 cells in 24-well plates. After 5 h, the attached cells were infected with 9 × 104 Infectable Units (IU) of RCAS-cGATA-3 and control viruses. After 5 days in culture, cells were fixed (4% formaldehyde) and immunohistochemistry (IHC) was performed. For clonal assay (Duff et al. 1991), the E2 neural tube explants were infected with RCAS viruses overnight. Then, trypsinized cells were plated at low density (300 per 35 mm dish), cultured for 10 days, and the expression of TH analyzed as described below.

Construction of RCAS-cGATA-3 retrovirus and immunohistochemistry

Full-length chick GATA-3 cDNA was cloned to the pSlax13 shuttle vector (Logan and Tabin 1998) that is designed for efficient expression of transgenes. GATA-3 cDNA was cloned at the ClaI site of RCASBP(B) and the resultant plasmid was named RCAS-cGATA3. After transfection of DF1 chicken fibroblast cells (ATCC) with RCAS-cGATA3 or the control empty vector, the viruses were harvested and concentrated as described (Logan and Tabin 1998). Viral titers were determined using the anti-gag protein monoclonal antibody (mAb) AMV-3C2 (Developmental Studies Hybridoma Bank, University of Iowa, IA, USA). TH expression in fixed NCSC was analyzed by indirect immunofluorescence. Cells were treated with rabbit anti-TH antibody (Pel-Freeze, Rogers, AR, USA), followed by Texas Red conjugated donkey anti-rabbit IgG (Jackson ImmunoResearch Laboratory, West Grove, PA, USA). Ten random fields that contained cells stained for TH were selected, and the numbers of positive-signaled cells and total cells stained with Hoechst dye [5 µg/mL in phosphate-buffered saline (PBS)] were counted. More than three independent experiments were performed for the analysis.

Cell culture and transient transfection assays

Human neuroblastoma cells, SK-N-BE(2)C and SK-N-BE(2)M17, and HeLa cells were maintained in DMEM supplemented with 10% fetal calf serum (HyClone, Logan, UT, USA), 100 µg/mL streptomycin and 100 U/mL penicillin in a CO2 incubator (Kim et al. 1998). Transfection was performed by the calcium phosphate method, using effector plasmids expressing GATA-3 along with various TH reporter constructs. Each 60 mm dish was transfected with an equimolar amount (0.5 pmol) of each reporter construct, 0.5 µg of pRSV-LUC, 0.1 pmol of the effector plasmid and pUC19 plasmid to a total of 5 µg DNA. The RSV-luc construct was used as an internal control. To correct for differences in transfection efficiencies, chloramphenicol acetyl transferase (CAT) activity was normalized according to luciferase activity, as previously described (Hong et al. 2001).

Plasmid construction

The TH-chloramphenicol acetyltransferase (CAT) reporter constructs, containing upstream sequences of the rat TH gene fused to the CAT gene, were used as described (Kim et al. 1993). TH2400(CREm)CAT contains a point mutation abolishing the CRE function. The cDNA of the human GATA-3 was cloned into pcDNA3.1/Zeo from SK-N-BE(2)C mRNA, resulting in pcDNA/GATA-3. The 580 bp NdeI/HindIII fragment, which contains RSV (Rous Sarcoma Virus) promoter, of RSV-PKI (Day et al. 1989) was cloned at pGL3-basic (Promega, Madison, WI, USA) to make RSV-luciferase (LUC). The mutation of the GATA-3 binding site in the TH promoter was made using the transformer site-directed mutagenesis kit (Clontech, Chicago, IL, USA) with oligonucleotides 5′-CTCGGGGTACCATGGAAGAAAAAAACACTGAGTGTGCCC TTACA-3′ (underlined bases represent mutant sequences).

Glutathione S-transferase (GST) pull-down assays and co-immunoprecipitation

pGEX-2T (Amersham Biosciences, Little Chalfont, UK) was used to express GATA-3 and CREB proteins which were fused to GST. Deletion constructs were made by PCR and cloned in pGEX-2T. Expression of the GST-fusion proteins was induced in Escherichia coli BL21 (DE3) by addition of 0.5 mm Isopropyl-b-D-thiogalactopyranoside (IPTG). The proteins were purified according to the manufacturer's protocol (Pharmacia Biotech, Piscataway, NY, USA). The purified proteins were stored at −70°C in aliquots containing 10% glycerol. For in vitro GST pull-down assays, [35S]-methionine-labeled proteins were obtained using the TNT-coupled wheat germ extract system (Promega). Purified GST (5 µg) or GST fusion proteins were bound to glutathione Sepharose beads and incubated with the labeled proteins for 2 h at 4°C in binding buffer [50 mm potassium phosphate (pH 7.5), 150 mm KCl, 1 mm MgCl2, 10% glycerol, 1 mm Triton X-100, and protease inhibitor cocktail (Roche)]. Beads were washed and bound proteins were eluted by boiling in 20 µL sodium dodecyl sulfate (SDS) loading buffer [60 mm Tris-HCl (pH 7.0), 2% SDS, 6% glycerol, 0.1 m dithiothreitol (DTT), 0.01% bromophenol blue]. The products were subjected to SDS–polyacrylamide gel electrophoresis (SDS–PAGE), fluorographic reagent and autoradiography.

For in vivo co-immunoprecipitation (Co-IP), 293T cells were harvested 48 h after transfection with green fluorescent protein (GFP)-tagged GATA-3 and Flag-tagged CREB constructs, which are expressed by cytomegalovirus (CMV) promoter. Cells were lysed in a 1% NP-40 lysis buffer (25 mm HEPES, 1 mm EDTA, 1 mm EGTA, 150 mm NaCl, 1 mm DTT and 1% NP-40) supplemented with protease inhibitors [2 mm phenylmethylsulfonyl fluoride (PMSF), 10 µg/mL leupeptin, 10 µg/mL pepstatin and 1 µg/mL aprotinin]. Cell extracts were treated with either anti-GFP or M2 anti-Flag antibody (1 : 1000). After incubation overnight at 4°C, 30 µL Protein A-Sepharose were added, and the reaction mixture was incubated for 2 h. The beads were pulled down through centrifugation and washed three times with 1 mL NP-40 lysis buffer containing protease inhibitors. Bound proteins were analyzed by western blotting. Detections were performed with anti-GFP (1 : 5000) and M2 anti-FLAG (1 : 5000) antibodies using the enhanced chemiluminescence (ECL) western blotting system (Amersham).

DNase I footprinting analysis and electrophoretic mobility shift assay (EMSA)

The DNase I footprinting assay was performed using TH promoter fragments that were prepared by PCR as previously described (Kim et al. 2001). Approximately 30 000 c.p.m. labeled probe was incubated with 2 µg purified GST-GATA3 (242–364) protein, which contains two zinc finger motifs. The amount of DNase I was adjusted empirically for each reaction to produce an even pattern of partially cleaved DNA fragments. Location of the protected region was determined by the Sanger sequencing method. Sense and antisense oligonucleotides corresponding to the sequences of GATA-3 binding site in the TH promoter and its mutant form were synthesized with the following nucleotide sequences: 5′-CCATGGAAGAGATGATCACTGAGTGTG-3′ and 5′-GCACACTCAGTGATCATCTCTTCCATG-3′, 5′-CCATGGAAGAAAAAAACACTGAGTGTG-3′ and 5′-GCACACTCAGTGTTTTTTTTCTTCCATG-3′, respectively (underlined bases represent wild and mutant sequences, respectively). Oligonucleotides representing the Sp1 binding site were used as a negative control. The sense and antisense oligonucleotides were annealed, 32P-labeled, and used as a probe for EMSA. Purified full-length GST-GATA3 or GST-GATA3 (242–364) was used in EMSA. All binding reagents including the nuclear extracts were mixed together and incubated at 25°C for 5 min. Approximately 40 000–50 000 c.p.m. of labeled probe was added, and the reaction was allowed to incubate at 25°C for 25 min. Competition binding assays were performed by adding non-radioactive competitor oligonucleotides in a molar excess before adding radiolabeled oligonucleotides. The DNA–protein complexes were resolved on high ionic strength, non-denaturing 6% polyacrylamide gels and visualized.


  1. Top of page
  2. Abstract
  3. Experimental procedures
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Forced expression of GATA-3 increases the number of TH-expressing cells in primary NCSC culture

To determine whether GATA-3 can directly control fate determination of NCSC lineages, we performed a gain-of-function analysis of GATA-3 by infecting quail primary NCSC with a chick GATA-3-expressing avian virus construct, RCAS-cGATA3. To determine optimal MOI (multiplicity of infection), a titration of RCAS-GFP viruses expressing GFP was performed on NCSCs from the trunk region of E2 quail embryos. After 2 days of incubation with 5 MOI of RCAS-GFP viruses, most of the cells were GFP positive (data not shown). Therefore, NCSCs were infected with 5 MOI of RCAS-cGATA3 or RCASBP(B) viruses and cultured for 5 days. As shown in Fig. 1, immunohistochemistry (IHC) using TH-specific antibody indicated that forced expression of GATA-3 increased the number of TH-positive cells more than twofold when compared with cells infected with the empty virus. The total number of NCSCs was not detectably changed under these conditions; in 10 random fields at 50× magnification, 7414 ± 1023 and 7587 ± 1379 cells were counted in culture with RCAS-cGATA3 and RCASBP(B), respectively.


Figure 1.  GATA-3 overexpression increases the number of TH-expressing cells in neural crest stem cell culture. Primary neural crest stem cell cultures were infected with RCASBP(B) alone or RCAS-cGATA3 viruses. (a) Immunohistochemistry was performed to detect TH expression. (b) TH-positive cells are represented as a percentage. Data are presented as mean ± SEM from four independent experiments. Bar: 300 µm.

Download figure to PowerPoint

Although the total number of cultured cells was unchanged, the data do not exclude the possibility that there could be increased cell proliferation of committed NA progenitor cells after GATA-3 infection. To investigate these possibilities, we performed clonal analysis using isolated neural crest cells from E2 quail embryos after viral infection and overnight incubation. After 10 days of incubation, the clones were classified into three groups based on the number of TH-expressing cells (< 20, between 20 and 100, and > 100 TH-positive cells per clone). Ectopic expression of GATA-3 increased the number of clones of all three categories with comparable efficiencies. Overall, the numbers of all TH-expressing clones were increased from 8.3% to 17.0%(Table 1). This result strongly supports the hypothesis that GATA-3 directly controls the NA fate determination of NCSCs.

Table 1.   Clonal analysis of TH-expressing neural crest stem cells
Number of TH-positive cells per cloneNumber of clones (% of total clones)
  1. The colony counts were from three independent experiments. The clones were classified into three groups based on the number of TH-expressing cells.

< 2028 (3.0%)56 (5.7%)
20 < < 10021 (2.2%)40 (4.0%)
> 10029 (3.1%)72 (7.3%)
Total number of clones944988

GATA-3 activates the transcriptional activity of the TH gene promoter

Based on a previous loss-of-function study in knockout mice (Lim et al. 2000) and our gain-of-function data (Fig. 1), we hypothesized that the TH gene may be an immediate downstream target of GATA-3. To address this, we tested whether forced expression of GATA-3 directly transactivates the transcriptional activity of the TH promoter. These co-transfection assays showed that GATA-3 up-regulates reporter gene expression driven by TH2400CAT approximately fourfold in non-neuronal HeLa cells (Table 2). However, there was no up-regulation of reporter gene expression in two TH-expressing cells, SK-N-BE(2)C and SK-N-BE(2)M17 (Table 2).

Table 2.   GATA-3 activates the TH promoter in HeLa cells
  1. HeLa, SK-N-BE(2)C and SK-N-BE(2)M17 cells were co-transfected with TH2400CAT and empty or GATA-3-expressing vector. RSV promoter activities were set to 100 by transfecting cells with 0.1 µg RSVCAT.

TH2400CATEmpty12.3 ± 2.811.5 ± 1.116.0 ± 3.0
GATA-344.0 ± 5.09.5 ± 0.516.4 ± 1.2
RSV 100100100

A novel GATA-3 binding site in the TH promoter has no functional role in activation by GATA-3

GATA family transcription factors, which contain two C4-type zinc finger domains, bind to the DNA motif WGATAR and transactivate promoters (Patient and McGhee 2002). DNA footprinting analysis was performed to determine the GATA-3 binding site in the TH promoter. A protected region was found at −825 to −839 bp of the TH promoter (Fig. 2a) that contains GATGAT sequences instead of the GATA motif. Specific binding of GATA-3 to the sequences was addressed. Purified GST-GATA3 (242–368) protein, which contains two zinc finger motifs, formed a complex with the oligonucleotide (TH/GAT) containing the DNase I protected region (Fig. 2b, lane 2). This complex was completely abolished by incubation with a 1000 molar excess of cold oligonucleotide (Fig. 2b, lane 5). However, the GST-GATA3 protein–TH/GAT DNA complex was not affected by addition of the non-specific Sp1 sequences (Fig. 2b, lanes 6–8). When mutant oligonucleotide, which has AAAAAA instead of GATGAT, was used the complex was not formed (data not shown). These results demonstrate the specificity of the complex. Next, we mutated the GATA-3 binding site in the context of the TH reporter construct by site-directed mutagenesis. However, the induction of TH2.4(GATGATm)CAT by GATA-3 was unchanged compared with that of the wild type (Fig. 2c). Taken together, these results suggest that the novel GATA-3 binding site found by DNase I footprinting analysis is not responsible for transactivation of the TH promoter.


Figure 2.  GATA-3 binds to novel sequences in the TH promoter. (a) The upstream region of the rat TH gene was analyzed by DNase I footprinting using purified GST-GATA3 (242–364) protein. The coding strand was radiolabeled and used as a probe. Sequences of protected region are shown. (b) Oligonucleotide TH/GAT containing a novel GATA-3 binding site of the TH promoter was radiolabeled and used as a probe. Purified GST-GATA3 (242–364) protein was incubated with the radiolabeled probe (lane 2). For competition, a 50-fold (lanes 3 and 6), 200-fold (lanes 4 and 7) or 1000-fold (lanes 5 and 8) molar excess of unlabeled oligonucleotides TH/GAT (lanes 2–5) and Sp1 (lanes 6–8) were added to the reaction mixture before the addition of radiolabeled probe. The same amount of GST was incubated with the probe and did not generate any complex (lane 1). Specific protein–DNA complexes are indicated by an arrowhead. (c) Promoter activity of the TH gene is not affected by mutation of the GATA-3 binding site in the promoter. HeLa cells were co-transfected with TH2400CAT and pcDNA/GATA-3, or empty vector, at a molar ratio of 0.2. Fold induction by effector plasmid co-transfection is presented as mean ± SEM value from six to nine independent samples. The bent arrow represents the TH transcription start site. The bold thick line denotes the 5′ untranslated sequences and the thin line denotes the 5′ upstream sequences of the TH gene.

Download figure to PowerPoint

GATA-3 directly activates the transcriptional activity of the TH gene promoter via a proximal region encompassing the CRE site

We next sought to map the specific promoter region of the TH gene that mediates the transactivation function of GATA-3. TH-CAT reporter constructs containing different lengths of the TH promoter were co-transfected with GATA-3-expressing plasmids in HeLa cells. GATA-3 transactivated reporter gene expression when the reporter construct contained 61 bp (or more) of the upstream promoter sequences (Fig. 3a, and data not shown). When tested in TH-expressing SK-N-BE(2)C and SK-N-BE(2)M17 cells, reporter gene expression driven by the deletion constructs was not affected at all by GATA-3 (data not shown). Interestingly, deletions up to −39 bp were refractory to GATA-3 transactivation in HeLa cells, suggesting that the subdomain between −61 and −39 bp of the TH promoter mediates GATA-3 responsiveness. Surprisingly, the known consensus motif (GATA) was absent in this proximal region. Instead, within this subdomain resides a cis-acting element, the cAMP response element (CRE; Fig. 3b), whose function is critical for basal and cAMP-regulated TH gene transcription (Kim et al. 1993; Trocme et al. 1998). To determine whether CRE is required for GATA-3 transactivation, we tested a mutant reporter construct, TH2400(CREm)CAT, harboring a point mutation (Kim et al. 1993). Indeed, GATA-3 failed to transactivate reporter gene expression driven by TH2400(CREm)CAT (Fig. 3b), suggesting that GATA-3 transactivates TH gene transcription in concert with the CRE-binding protein (CREB). Another possible explanation is that disturbance of basal promoter function by CRE mutation causes the reduction of transactivation by GATA-3. To test this possibility, we co-transfected HeLa cells with the expression vector of the AP2α protein that can transactivate the TH promoter through the AP2 response elements in HeLa cells (Kim et al. 2001). We found that AP2α transactivated both the wild-type and CRE-mutated TH promoter approximately 10-fold (Fig. 3b). Taken together, these findings support the notion that specific GATA-3–CREB interaction may be important for transactivation of the TH promoter by GATA-3.


Figure 3.  GATA-3 transactivates rat TH promoter by interacting through the CRE site. (a) GATA-3 transactivation of serially-deleted TH promoters. (b) Mutation of the CRE site on TH2400CAT diminishes reporter gene activation by GATA-3 but not by AP2α. The bold characters represent the CRE site. HeLa cells were co-transfected with TH-CAT and pcDNA/GATA-3, pcDNA/AP2α or empty vector at a molar ratio of 0.2. The arrowhead represents the GATA-3 response region of the TH promoter. Co-transfections were performed in HeLa cells and the symbols are the same as those used in Fig. 2. The numbers on the left of the diagram represent the size of the TH promoter.

Download figure to PowerPoint

GATA-3 physically interacts with CREB in vitro and in vivo

To investigate whether GATA-3 interacts directly with CREB, we performed GST pull-down assays. Indeed, in vitro translated full-length GATA-3 was able to bind to CREB (Fig. 4a). GATA-3 contains two zinc finger domains, N-finger and C-finger, at amino acid (aa) residues 263–288 and 312–342, respectively (Yang et al. 1994). To test which GATA-3 protein domain is required for protein–protein interaction, we tested various fragments in our protein interaction assays. We found that the C-terminal half (aa 242–444) encompassing both the N- and C-fingers is required for protein–protein interaction with CREB (Fig. 4; data not shown). Next, we labeled the full-length and shorter fragments of CREB. Truncated proteins as well as correct-sized proteins were made during labeling (Fig. 4a, lower panel, lane 1). The labeled proteins were incubated with the minimal domain of GATA-3 (aa 242–444) bound to glutathione Sepharose beads. This analysis demonstrated that the C-terminal region of CREB (aa 270–327) encompassing the basic region and the leucine zipper is essential for interaction with GATA-3 (Fig. 4).


Figure 4.  GATA-3 physically interacts with CREB. (a) Determination of the minimal interaction domains of GATA-3 and CREB. (Upper panel) In vitro translated [35S]-methionine-labeled GATA-3 proteins were incubated with GST (lanes 4, 6 and 8) or with full-length GST-CREB (lanes 5, 7 and 9) bound to glutathione Sepharose beads. [35S]-methionine-labeled input proteins are shown (lane 1: aa 1–144; lane 2: aa 1–364; lane 3: aa 242–444). (Lower panel) In vitro translated [35S]-methionine-labeled CREB proteins were incubated with GST (lanes 5, 7, 9 and 11) or GST-GATA-3 (aa 242–444) corresponding to aa residues 242–444 (lanes 6, 8, 10 and 12) bound to glutathione Sepharose beads. [35S]-methionine-labeled input proteins are shown (lane 1: aa 1–327; lane 2: aa 212–327; lane 3: aa 270–327; lane 4: aa 1–218). The numbers at the top of the figure represent the amino acid residues of GATA-3 (upper panel) or CREB (lower panel). Protein size makers (kDa) are shown on the left. (b) Schematic representation of interaction results between GATA-3 and CREB. The two zinc finger motifs of GATA-3, N-finger (aa 263–288) and C-finger (aa 312–342), and the basic leucine zipper (bZIP) region of CREB are shown.

Download figure to PowerPoint

To test whether GATA-3 interacts with CREB in vivo, we used co-immunoprecipitation (Co-IP) assays. We co-transfected 293T cells with GFP-tagged GATA-3 and Flag-tagged CREB. Expression of tagged transcription factors was confirmed by western blot analysis. Cell lysates were precipitated with αGFP antibody and western blot analysis was performed with αFlag antibody to detect the presence of tagged proteins. In vivo interaction between GATA-3 and CREB (Fig. 5) was robustly detected.


Figure 5.  GATA-3 interacts with CREB in vivo. 293T cells were co-transfected with the empty vector, GATA-3 and CREB as indicated at the top of the figure. The cell lysates were precipitated with an anti-GFP antibody. Monoclonal anti-Flag was used to detect the Flag-tagged CREB protein. Upper panel, crude cell extracts; lower panel, immunoprecipitates.

Download figure to PowerPoint


  1. Top of page
  2. Abstract
  3. Experimental procedures
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Transcription factors belonging to the GATA family contain two C4-type zinc finger domains and bind to the DNA motif WGATAR (Patient and McGhee 2002). Among the six members of the GATA family, GATA-3 is best known as a master regulator of Th2 cell differentiation (Murphy and Reiner 2002). In the present study, we investigated the potential functional role of GATA-3 in the regulatory cascade of specification of the NA neurotransmitter identity during SNS development.

In the NA biosynthetic pathway, TH is the first and, presumably, rate-limiting step converting tyrosine to l-dopa. Thus, expression of TH is an essential feature of NA neurotransmitter identity. We hypothesized that GATA-3 may directly regulate NA phenotype specification, and that the TH gene is an immediate downstream target of GATA-3. Alternatively, GATA-3 can indirectly regulate TH gene expression by inducing proliferation of precursor cells and/or regulating other transcription factors, e.g. Phox2a and 2b, which are known to transactivate TH gene expression. Several lines of evidence described in this study support the first possibility. First, we found that GATA-3 was able to transactivate TH promoter prominently. Second, overexpression of GATA-3 increased the number of TH-expressing cells in NCSC culture. Third, in our clonal analysis, GATA-3 increased the number of TH-positive clones.

GATA family transcription factors bind to GATA-containing sequence motifs (Murphy and Reiner 2002). Interestingly, the TH subdomain responsive to GATA-3 does not contain any GATA-3 recognition motif. In the case of the TH proximal subdomain (at −61 to −39 bp), the CRE is the only known cis-regulatory motif, which is the binding site for the basic region-leucine zipper (bZip) transcription factor CREB. Mutation of this CRE motif almost completely abolished transactivation of the TH promoter by GATA-3, while it did not affect transactivation by another transcription factor, AP2α. These results suggest that GATA-3–CREB interaction may be involved in TH promoter regulation. Indeed, in vitro GST pull-down and in vivo Co-IP assays demonstrated that GATA-3 and CREB could physically interact with each other. Further protein interaction assays using a series of fragments delineated the regions of the proteins essential for their physical interaction. For GATA-3, the C-terminal half (aa 242–444) encompassing both fingers was sufficient while for CREB, the fragment (aa 270–327) encompassing bZip allowed the interaction. Taken together, we propose that GATA-3 is able to transactivate the TH gene via physical interaction with CREB. Despite prominent expression of CREB, forced expression of GATA-3 did not activate the TH promoter in TH-expressing cells. This observation can be explained by the assumption that GATA-3 is sufficiently expressed in the TH-expressing cells to maintain its activity. In line with this, Phox2a and Phox2b, NA-specific transcription factors, transactivated the d.b.h. promoter in d.b.h. non-expressing cell lines, but not in d.b.h.-expressing cell lines (Yang et al. 1998).

In developing embryos, neural crest cells originate from the roof plate of the neural tube and migrate between somites and the neural tube. At their final target sites, neural crest cells of the trunk region become different types of neurons or glia in response to environmental signals. To address further the in vivo function of GATA-3, we infected primary NCSC isolated from the trunk region of E2 quail embryos with replication competent RCAS-cGATA-3 or RCASBP(B) viruses. Whereas the total number of NCSC was not affected by GATA-3, the number of cells with the sympathoadrenergic phenotype was increased more than twofold as examined by TH expression. The increase in TH- positive cells could be explained by the increase in cell proliferation of NA progenitor cells or by fate changes of NCSCs by GATA-3. In our clonal analysis, the number of TH-expressing clones was increased from 8.3% to 17.0%. Thus, these results (i.e. increases of both TH-positive cells and clones) cannot be explained by simple up-regulation of TH gene expression by GATA-3 in TH-positive cells. Rather, our result strongly supports the hypothesis that GATA-3 directly controls the NA fate determination of NCSCs during sympathetic neuron development and that this process involves, at least in part, direct activation of TH gene transcription.

One interesting issue is the functional relationship between GATA-3 and GATA-2. Among the six GATA factors, GATA-3 and GATA-2 share the highest sequence homology, and only these two members are expressed in the nervous system in an overlapping manner. While a knockout study indicated a critical role of GATA-3 in mouse SNS development, a similar role was suggested for GATA-2 in chick embryo (Groves et al. 1995). In developing chick SNS, GATA-2 was expressed after Cash1 and Phox2 proteins, followed by SCG10 and TH (Groves et al. 1995). Moreover, Tsarovina et al. (2004) recently reported that GATA-2, but not GATA-3, is expressed in developing chick sympathetic precursors and controls SNS development, suggesting that GATA-2 in chick is the functional counterpart of GATA-3 in mouse in SNS development. Notably, GATA-3 is genetically downstream of Phox2b, because its expression is abolished in Phox2b knockout mice while Phox2b is induced in GATA-3 knockout mice (Lim et al. 2000). Though GATA-3 is critical for the development of SNS (Lim et al. 2000; Tsarovina et al. 2004), induction of NA phenotype by GATA-3 requires a certain cellular context. GATA-3 needs to work in concert with other factors, such as Phox2a/b and dHand, because the increase in TH-positive clones in NCSC culture was rather modest (2–2.5-fold).

In summary, our results show that GATA-3 can regulate the transcriptional activity of the TH gene via a novel and distinct protein–protein interaction. We propose that TH is an immediate target gene of GATA-3, and that GATA-3 directly participates in NA phenotype specification during SNS development. Although GATA-3 has been shown to interact with several proteins, such as FOG, ROG and Smad3 (Miaw et al. 2000; Blokzijl et al. 2002; Kurata et al. 2002), our study is the first to support the possibility that GATA-3 can regulate its target gene(s) via protein–protein interactions with a general transcription factor. In this context, it is noteworthy that another key transcription factor for dopamine neuron development, Nurr1, has been shown to enhance the transcriptional activity of the human dopamine transporter gene through a promoter region devoid of Nurr1 binding sites (Sacchetti et al. 2001). Thus, this potentially novel mechanism by which key fate-determining transcription factors (e.g. GATA-3 and Nurr1) control neurotransmitter identity warrants further investigation. In addition, as CREB is expressed in a vast number of tissues, it is of great interest to determine whether GATA-3 may regulate additional target genes via similar protein–protein interactions.


  1. Top of page
  2. Abstract
  3. Experimental procedures
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

The authors thank J. Engel for the chick GATA-3 clone, C. Tabin for the RCASBP vectors and pSlax13, O. Andrisani for the NCSC culture and M. Sieber-Blum for clonal assay. This work was supported by NIH grants (MH48866 and DC006501) and a NARSARD Independent Award.


  1. Top of page
  2. Abstract
  3. Experimental procedures
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  • Bilodeau M. L., Boulineau T., Hullinger R. L. and Andrisani O. M. (2000) Cyclic AMP signaling functions as a bimodal switch in sympathoadrenal cell development in cultured primary neural crest cells. Mol. Cell Biol. 20, 30043014.
  • Blokzijl A., Ten Dijke P. and Ibanez C. F. (2002) Physical and functional interaction between GATA-3 and Smad3 allows TGF-beta regulation of GATA target genes. Curr. Biol. 12, 3545.
  • Brunet J. F. and Pattyn A. (2002) Phox2 genes – from patterning to connectivity. Curr. Opin. Genet. Dev. 12, 435440.
  • Cooper J. R., Bloom F. E. and Roth R. H. (1996) The Biochemical Basis of Neurophamacology, 7th Edn. Oxford University Press, New York.
  • Day R. N., Walder J. A. and Maurer R. A. (1989) A protein kinase inhibitor gene reduces both basal and multihormone-stimulated prolactin gene transcription. J. Biol. Chem. 264, 431436.
  • Duff R. S., Langtimm C. J., Richardson M. K. and Sieber-Blum M. (1991) In vitro clonal analysis of progenitor cell patterns in dorsal root and sympathetic ganglia of the quail embryo. Dev. Biol. 147, 451459.
  • Goridis C. and Rohrer H. (2002) Specification of catecholaminergic and serotonergic neurons. Nat. Rev. Neurosci. 3, 531541.
  • Groves A. K., George K. M., Tissier-Seta J. P., Engel J. D., Brunet J. F. and Anderson D. J. (1995) Differential regulation of transcription factor gene expression and phenotypic markers in developing sympathetic neurons. Development 121, 887901.
  • Guillemot F., Lo L. C., Johnson J. E., Auerbach A., Anderson D. J. and Joyner A. L. (1993) Mammalian achaete-scute homolog 1 is required for the early development of olfactory and autonomic neurons. Cell 75, 463476.
  • Guo S., Brush J., Teraoka H., Goddard A., Wilson S. W., Mullins M. C. and Rosenthal A. (1999) Development of noradrenergic neurons in the zebrafish hindbrain requires BMP, FGF8, and the homeodomain protein soulless/Phox2a. Neuron 24, 555566.
  • Hong S. J., Kim C. H. and Kim K. S. (2001) Structural and functional characterization of the 5′ upstream promoter of the human Phox2a gene: possible direct transactivation by transcription factor Phox2b. J. Neurochem. 79, 12251236.
  • Howard M. J. (2005) Mechanisms and perspectives on differentiation of autonomic neurons. Dev. Biol. 277, 271286.
  • Kim H. S., Seo H., Yang C., Brunet J. F. and Kim K. S. (1998) Noradrenergic-specific transcription of the dopamine beta-hydroxylase gene requires synergy of multiple cis-acting elements including at least two Phox2a-binding sites. J. Neurosci. 18, 82478260.
  • Kim H. S., Hong S. J., LeDoux M. S. and Kim K. S. (2001) Regulation of the tyrosine hydroxylase and dopamine beta-hydroxylase genes by the transcription factor AP-2. J. Neurochem. 76, 280294.
  • Kim K. S., Lee M. K., Carroll J. and Joh T. H. (1993) Both the basal and inducible transcription of the tyrosine hydroxylase gene are dependent upon a cAMP response element. J. Biol. Chem. 268, 1568915695.
  • Kurata H., Lee H. J., McClanahan T., Coffman R. L., O'Garra A. and Arai N. (2002) Friend of GATA is expressed in naive Th cells and functions as a repressor of GATA-3-mediated Th2 cell development. J. Immunol. 168, 45384545.
  • Lim K. C., Lakshmanan G., Crawford S. E., Gu Y., Grosveld F. and Engel J. D. (2000) Gata3 loss leads to embryonic lethality due to noradrenaline deficiency of the sympathetic nervous system. Nat. Genet. 25, 209212.
  • Lo L., Morin X., Brunet J. F. and Anderson D. J. (1999) Specification of neurotransmitter identity by Phox2 proteins in neural crest stem cells. Neuron 22, 693705.
  • Logan M. and Tabin C. (1998) Targeted gene misexpression in chick limb buds using avian replication-competent retroviruses. Methods 14, 407420.
  • Marquardt T. and Pfaff S. L. (2001) Cracking the transcriptional code for cell specification in the neural tube. Cell 106, 651654.
  • Miaw S. C., Choi A., Yu E., Kishikawa H. and Ho I. C. (2000) ROG, repressor of GATA, regulates the expression of cytokine genes. Immunity 12, 323333.
  • Morin X., Cremer H., Hirsch M. R., Kapur R. P., Goridis C. and Brunet J. F. (1997) Defects in sensory and autonomic ganglia and absence of locus coeruleus in mice deficient for the homeobox gene Phox2a. Neuron 18, 411423.
  • Murphy K. M. and Reiner S. L. (2002) The lineage decisions of helper T cells. Nat. Rev. Immunol. 2, 933944.
  • Pandolfi P. P., Roth M. E., Karis A., Leonard M. W., Dzierzak E., Grosveld F. G., Engel J. D. and Lindenbaum M. H. (1995) Targeted disruption of the GATA3 gene causes severe abnormalities in the nervous system and in fetal liver haematopoiesis. Nat. Genet. 11, 4044.
  • Patient R. K. and McGhee J. D. (2002) The GATA family (vertebrates and invertebrates). Curr. Opin. Genet. Dev. 12, 416422.
  • Pattyn A., Morin X., Cremer H., Goridis C. and Brunet J. F. (1999) The homeobox gene Phox2b is essential for the development of autonomic neural crest derivatives. Nature 399, 366370.
  • Sacchetti P., Mitchell T. R., Granneman J. G. and Bannon M. J. (2001) Nurr1 enhances transcription of the human dopamine transporter gene through a novel mechanism. J. Neurochem. 76, 15651572.
  • Stanke M., Junghans D., Geissen M., Goridis C., Ernsberger U. and Rohrer H. (1999) The Phox2 homeodomain proteins are sufficient to promote the development of sympathetic neurons. Development 126, 40874094.
  • Trocme C., Sarkis C., Hermel J. M., Duchateau R., Harrison S., Simonneau M., Al-Shawi R. and Mallet J. (1998) CRE and TRE sequences of the rat tyrosine hydroxylase promoter are required for TH basal expression in adult mice but not in the embryo. Eur. J. Neurosci. 10, 508521.
  • Tsarovina K., Pattyn A., Stubbusch J., Muller F., Van Der Wees J., Schneider C., Brunet J. F. and Rohrer H. (2004) Essential role of Gata transcription factors in sympathetic neuron development. Development 131, 47754786.
  • Yang C., Kim H. S., Seo H. and Kim K. S. (1998) Identification and characterization of potential cis-regulatory elements governing transcriptional activation of the rat tyrosine hydroxylase gene. J. Neurochem. 71, 13581368.
  • Yang Z., Gu L., Romeo P. H., Bories D., Motohashi H., Yamamoto M. and Engel J. D. (1994) Human GATA-3 trans-activation, DNA-binding, and nuclear localization activities are organized into distinct structural domains. Mol. Cell Biol. 14, 22012212.