An Enhancer Element for Expression of the Ncx(Enx, Hox11L1) Gene in Neural Crest-derived Cells*

The murine Ncx (Enx,Hox11L1) gene is specifically expressed in a neuronal subset of neural crest-derived tissues. In attempts to elucidate the regulatory DNA element of the tissue-specific expression, we sequenced the 5′-flanking region of the Ncx gene. The transcriptional initiation site was determined at 297 nucleotides (−297) upstream from the ATG start codon (+1). A retinoic acid response element was located on the region between −1163 and −1150. Transient transfection assays with the 5′-flanking sequences fused to the luciferase gene showed that the region between −1387 and −1368 was crucial for the tissue-specific enhancer activity. Furthermore, nuclear proteins extracted from neural crest-derived cells such as murine and human neuroblastoma cells bind to the DNA region between −1387 and −1368. This DNA element was also conserved in the 5′-flanking region of the human NCX gene. Our observations strongly suggest that the DNA element (between −1387 and −1368) contributes to tissue-specific expression of the Ncx gene in murine and human species.

The murine Ncx (Enx, Hox11L1) gene is specifically expressed in a neuronal subset of neural crest-derived tissues. In attempts to elucidate the regulatory DNA element of the tissue-specific expression, we sequenced the 5-flanking region of the Ncx gene. The transcriptional initiation site was determined at 297 nucleotides (؊297) upstream from the ATG start codon (؉1). A retinoic acid response element was located on the region between ؊1163 and ؊1150. Transient transfection assays with the 5-flanking sequences fused to the luciferase gene showed that the region between ؊1387 and ؊1368 was crucial for the tissue-specific enhancer activity. Furthermore, nuclear proteins extracted from neural crest-derived cells such as murine and human neuroblastoma cells bind to the DNA region between ؊1387 and ؊1368. This DNA element was also conserved in the 5-flanking region of the human NCX gene. Our observations strongly suggest that the DNA element (between ؊1387 and ؊1368) contributes to tissue-specific expression of the Ncx gene in murine and human species.
The mammalian homeobox (Hox) genes encode proteins containing a conserved 60-amino acid sequence (homeodomain) originally identified in Drosophila (1). In humans and mice, many of the Hox genes are clustered on four different chromosomes (HoxA, -B, -C, and -D), and each cluster retains a linear arrangement of the gene that is similar to its Drosophila counterpart. Expression of each Hox gene in the cluster is strictly regulated temporally and spatially during embryogenesis, and the gene located on the more 3Ј side of the cluster is expressed in the more anterior part of mouse embryo. Those clustered Hox genes play an important role in specification of position along axes in the developing embryo (1). In addition to clustered Hox genes, some Hox genes do not locate in those four cluster regions. Expression of the orphan Hox genes is also tightly regulated temporally and spatially during embryogenesis.
The murine Ncx/Enx/Hox11L1 gene (2, 3), a member of the Hox11 gene family (4 -7) that locates outside Hox gene clusters (8), is expressed specifically in neural crest-derived tissues such as dorsal root ganglia, cranial ganglia (V, IX, and X), sympathetic ganglia, adrenal medulla, and enteric ganglia (9,10). Neural crest cells arise from the dorsal part of neural tube and migrate along a number of defined routes to various tissues from embryonic day 9.5. These neural crest cells mainly differentiate into neuronal cells in the peripheral nervous system, melanocytes, and cephalic mesenchymal cells such as smooth muscle and cartilage (11). Since the Ncx mRNA has not been detected in the cell lineage of melanocytes and mesenchymal cells (9), it is apparently specifically expressed in a neuronal subset of neural crest-derived tissues. The regulatory mechanisms of expression have remained to be identified.
Deregulated expression of the HOX genes can be oncogenic in humans (5,12,13). Mice with mutations in Hox genes also have a variety of congenital anomalies resembling human genetic diseases. For example, mice deficient for the Ncx gene developed megacolon with the hyperinnervation of enteric neurons in the narrow segment of megacolon (10,14). This phenotype resembles the human disease known as neuronal intestinal dysplasia. Therefore, a detailed analysis of regulatory mechanisms of Hox gene expression is required to better understand the genetic basis of such disease in humans.
To examine molecular mechanisms underlying the regulation of Ncx gene expression, we sequenced 5Ј-flanking regions of the murine and the human Ncx gene; both have a retinoic acid response element (RARE). 1 Furthermore, the enhancer element necessary for tissue-specific expression was identified in the 5Ј-flanking region by transient transfection of the murine Ncx promoter-luciferase chimeric genes into neuroblastoma cells. The element was also conserved in the human NCX gene. We also demonstrate that nuclear proteins from neural crest-derived cells such as neuroblastoma and melanoma cells (but not from fibroblast cells) bind to the enhancer element. The regulatory region of tissue-specific expression of the Ncx gene is discussed.
Cell Culture-C1300 (Neuro-2A) murine neuroblastoma cells, B16 murine melanoma cells, and five kinds of human neuroblastoma cells (SH-SY5Y (15), SMS-KCN (16), SK-N-AS (17), and NBTM and GANB 2 ) were cultured in RPMI 1640 medium (Life Technologies, Inc.) with 10% fetal calf serum (Bioserum, CSL Ltd.) at 37°C under 5% CO 2 . L mouse * This work was supported by a grant-in-aid for scientific research on priority areas from the Ministry of Education, Science, Sports and Culture, Japan. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM  fibroblast cells were maintained in Dulbecco's modified Eagle's medium (Life Technologies, Inc.) with 5% fetal calf serum. All-trans-retinoic acid (RA; Sigma) was dissolved in ethanol at a 10 mM concentration. RA was added to the medium in a volume corresponding to 0.1% (10 M) of the culture medium.
Northern Blot Analysis-Total RNAs were extracted from murine embryos and several cell lines using the Trizol reagent (Life Technologies, Inc.). Total RNAs (15 g) were electrophoresed through a 1.0% agarose gel containing formaldehyde, transferred to a nylon membrane (Roche Molecular Biochemicals). The filter was hybridized overnight at 50°C in 50% formamide hybridization buffer with 0.5% SDS, 1% blocking reagent, and the digoxigenin (DIG)-labeled murine Ncx homeodomain probe (193 base pairs (bp)) synthesized by polymerase chain reaction (PCR) on Ncx cDNA using primers (5Ј-ATCCCTACCAAAAC-CAAACC-3Ј and 5Ј-TTGGTGCGTCGGTTCTGGAA-3Ј) (9). Following hybridization, the filter was washed twice for 5 min with 2ϫ SSC and 0.1% SDS at room temperature and twice for 15 min with 0.1ϫ SSC and 0.1% SDS at 55°C. The DIG-labeled probe was detected with sheep anti-DIG antibody conjugated with alkaline phosphatase. The antibody detection reaction was performed using an enhanced chemiluminescent detection system (Roche Molecular Biochemicals).
Reverse Transcribed (RT)-PCR and Southern Blot Analysis-RT-PCR and Southern blot analysis were performed as described elsewhere (9). Briefly, cDNAs were made by reverse-transcribing total RNAs (100 ng) with oligo(dT) primers and amplified 30 times with 2.5 units of Taq DNA polymerase, 2 mM MgCl 2 , 5% formamide, and the primers (5Ј-TTTGCAAAGGACAGGCTCACG-3Ј and 5Ј-GGTGCAGCAGCAGGCG-ACCA-3Ј) complementary to the murine Ncx cDNA. The PCR products were separated on a 1% agarose gel and transferred onto a nylon membrane (Amersham Pharmacia Biotech). The filter was hybridized overnight with the DIG-labeled murine Ncx homeodomain probe at 42°C. The DIG-labeled probe was detected using the same method described above.
Isolation of the Ncx Genomic Clones-A murine Ncx genomic clone was isolated from the 129/Sv genomic library, using the murine Ncx homeodomain probe. A 10.0-kilobase pair (kb) XbaI fragment of the genomic clone was inserted into pBluescript KS(Ϫ) (Stratagene). The human NCX cDNA was isolated from the cDNA library of human neuroblastoma (SMS-KCN) cells in ZAPII (Stratagene), using the murine Ncx homeodomain probe. A human NCX genomic clone was isolated from a human placenta genomic library in EMBL3 (Stratagene) with the human NCX probe (a 241-bp BstXI/NotI fragment in exon 1). A 6.8-kb HindIII fragment derived from the human genomic clone was subcloned into pGEM-7Zf(ϩ) (Promega).
Rapid Amplification of 5Ј-cDNA Ends (5Ј-RACE) Analysis-The 5Ј-RACE analysis was performed using the 5ЈAmpliFINDER RACE kit (CLONTECH). The Ncx mRNAs in 2 g of poly(A) ϩ RNA isolated from the whole body of mouse embryos at embryonic day 12.5 using oligo(dT)cellulose columns (Life Technologies, Inc.) were reverse-transcribed using the primer 5Ј-AAAACGCCGCACTTTCTCC-3Ј, complementary to nucleotides between ϩ151 and ϩ133 of the Ncx cDNA. The cDNA was mixed with 2.5 units of Taq DNA polymerase, 1.5 mM MgCl 2 , 5% formamide, the primer (5Ј-GGATGACTGGAAAAGGTGTTTC-3Ј; complementary to nucleotides between Ϫ21 and Ϫ42 of the Ncx cDNA), and the anchor primer for PCR. PCR was run for 1 min at 94°C for denaturing, 1 min at 60°C for annealing, and 1 min at 72°C for extension. After 35 cycles of the PCR, the products were separated on a 1.5% agarose gel and stained with ethidium bromide. ϳ330-bp PCR products were ligated with pGEM-T (Promega) for sequencing.
Primer Extension-This analysis was made using by the method described elsewhere (18). An oligonucleotide primer (5Ј-AGTC-CGCTCTGGGACTCGCAT-3Ј; complementary to nucleotides between Ϫ187 and Ϫ207 of the Ncx genomic sequence) was end-labeled with [␥-32 P]ATP (Amersham Pharmacia Biotech). The labeled primer was mixed with 15 g of poly(A) ϩ RNA isolated from the whole body of murine embryos at embryonic day 12.5 in 30 l of hybridization buffer (40 mM PIPES, 1 mM EDTA, 0.4 M NaCl, 80% formamide) overnight at 30°C. 15 g of transfer RNA was used as a negative control. After ethanol precipitation, reverse transcription was performed at 37°C for 1 h, and products were loaded onto a 6% sequencing gel. Nucleotide sequencing of the murine genomic Ncx gene was carried out by the dideoxy chain termination method using the same primer.
Assay for the Promoter Activity with the Luciferase Reporter Gene-2.0 ϫ 10 6 cells (C1300, B16, L, and SH-SY5Y) were plated on a 100-mm dish 1 day before transfection. The luciferase reporter plasmid (5 g) and the sea pansy luciferase expression vector, pRL-SV40 (0.25 g) (TOYO INK, Tokyo, Japan), as a transfection efficiency control, were co-transfected into those cells using the DNA-calcium phosphate coprecipitation technique (19). SV40Luc, the pGL2 plasmid with the SV40 promoter and enhancer and the firefly luciferase reporter gene (Promega), was used for a positive control. The transfected cells were incubated for 8 h, washed three times with phosphate-buffered saline, and further cultured with corresponding culture medium for 24 h. RA stimulation was given 8 h before harvesting. Luciferase activities in cell lysates were determined using standard methods using the pikkagene dual (TOYO INK) with luminometer (Lumat LB9507; Berthold). The firefly luciferase activity from the reporter plasmids was normalized with the sea pansy luciferase activity. Preparation of Nuclear Extracts-Nuclear extracts from C1300, L, B16, and SH-SY5Y cells were prepared as described elsewhere (20). Briefly, cultured cells were resuspended in hypotonic buffer. After centrifugation, nuclei were resuspended in extraction buffer and incubated on ice for 60 min. The amounts of protein in the supernatant were estimated by micro-BCA protein assay (Pierce).

Ncx Expression and Up-regulation with RA Stimulation in
Neuronal Lineage Cells-The Ncx gene is specifically expressed in a neuronal subset of neural crest derived tissues in murine embryos (9,10). To further determine the cell type-specific expression, we examined the expression in several murine cell lines, using RT-PCR. The Ncx gene was expressed in C1300 neuroblastoma cells but not in B16 melanoma cells, L fibroblast cells, and M12 B lymphoma cells (Fig. 1A). We then investigated the NCX expression in several human neuroblastoma cell lines (NBTM, SH-SY5Y, SMS-KCN, SK-N-AS, and GANB), using Northern blot analysis; all expressed the NCX (Fig. 1B). PC12 rat pheochromocytoma cells also expressed the Ncx gene. (data not shown).
RA stimulation induces differentiation of some cell lines and embryonic tissues with expression of the Hox genes (22). Since RA induced morphological differentiation in C1300 and SH-SY5Y neuroblastoma cell lines (23,24), we examined the inducibility of Ncx expression in those differentiating cells after RA stimulation. As shown in Fig. 2A, the Ncx expression was up-regulated to 10-fold of the prestimulation level in C1300 cells at 2 h and then reverted to the prestimulation level within 9 h later. In the SH-SY5Y cells, the NCX expression was also up-regulated at 6 h and then down-regulated within 24 h later (Fig. 2B).

Regulatory Elements in the 5Ј-Flanking Region of the Murine and the Human NCX Gene-
To examine the regulatory elements for cell type-specific expression and the RA responsibility of the Ncx gene, we sequenced 1597 bp of the 5Ј-flanking region with the translation start codon (Fig. 3) and determined the transcription initiation site of the gene by primer extension analysis and the 5Ј-RACE with poly(A) ϩ RNA from the whole body of murine embryos. Fig. 4 shows that one initiation site was located 297 bp upstream from ATG (ϩ1). Comparison of the sequence of the 5Ј-RACE products to the genomic sequence confirmed that no intron existed on the 5Ј-flanking region from the ATG translation start codon (data not shown). When we compared the 5Ј-flanking sequence with established consensus sequences in a transcription factor data base, this regulatory region contained a DR2-RARE, which consists of direct repeats of two motifs (AGGTCA) separated by 2 nucleotides (25), located in the region between Ϫ1163 and Ϫ1150 and the putative binding site of Sp1 (GGGGCGGGGT; at positions Ϫ1463/ Ϫ1452) and AP-1 (TGAGTCA; at positions Ϫ461/Ϫ455) (Fig. 3). However, the region lacked the conventional TATA box and the CAAT box sequence.
Since the NCX gene was also expressed and up-regulated with RA stimulation in human neuroblastoma cell lines (Fig.  2B), we sequenced 2169 bp of the 5Ј-flanking region from the translation start codon of the human genomic NCX gene. When the sequence was compared with that from the mouse, the regions between Ϫ1397 and Ϫ1065 (Fig. 5) and between Ϫ682 and Ϫ522 (data not shown) in the murine Ncx gene were highly conserved in the human NCX gene (Ϫ1517/Ϫ1187 and Ϫ801/ Ϫ638, respectively). The region between Ϫ1517 and Ϫ1187 in the human NCX gene also contained a DR2-RARE sequence (Ϫ1283/Ϫ1270). These data suggest that these conserved regions are important for regulation of the cell type-specific expression of the Ncx gene.
Functional Analysis of the Murine Ncx Promoter-We next analyzed a cell type-specific promoter activity in the 5Ј-flanking region of the murine Ncx gene by transient transfection of the promoter-reporter fusion genes into various cell lines. Various lengths of the 5Ј-flanking region of the Ncx gene were fused to the firefly luciferase gene. Since the endogenous Ncx was expressed in SH-SY5Y and C1300 cells but not in B16 and L cells (Fig. 1), those chimeric constructs were transfected into those four cell lines to examine the cell type-specific promoter activity (Fig. 6A). A significant increase (3-8-fold) of the luciferase activity was observed in all four cell lines when the 1153NLuc gene was transfected, suggesting the presence of a minimal promoter element in the first 856 bp (Ϫ1153/Ϫ297) of this 5Ј-flanking region. When the 1595NLuc construct was transfected into SH-SY5Y and C1300 cells, the luciferase activity increased about 10-fold over the activity observed in the case of 1153NLuc transfection. However, this augmentation was not detected in B16 and L cells transfected with the 1595NLuc gene. Transfection of the 3300NLuc and the 7000NLuc construct demonstrated almost the same activity as that of the 1595NLuc, in all cell lines tested. These results suggested that the region between Ϫ1595 and Ϫ1153 is important for cell type-specific expression of the Ncx gene.
To further characterize the regulatory region of the murine Ncx gene, a series of deletion clones from the region between Ϫ1595 and Ϫ1153, prepared using the Erase-a-base system and fused to the luciferase reporter gene, were transfected into C1300 cells (Fig. 6B). The region between Ϫ1373 and Ϫ1153 did not enhance the promoter activity, irrespective of the presence of RARE. The promoter activity of the constructs with the region between Ϫ1414 and Ϫ1153 was more than 7-fold higher than that with the region between Ϫ1373 and Ϫ1153. These findings suggest that the region between Ϫ1414 and Ϫ1374 plays an important role in determining enhancer activity in neuroblastoma cells of the Ncx gene.
A DNA Element for the Cell Type-specific Expression in the 5Ј-Flanking Region of the Murine Ncx Gene-The reporter gene analysis using a series of deletion clones in the 5Ј-flanking region indicated that a DNA element regulating the cell typespecific transcription of the Ncx gene was located in the region between Ϫ1414 and Ϫ1374. Furthermore, the DNA sequence of the region between Ϫ1397 and Ϫ1065 in the murine Ncx gene was highly conserved in the human NCX gene. These observations strongly suggested that nuclear factors derived from neuroblastoma cells bound to the region between Ϫ1397 and Ϫ1374 to regulate cell type-specific expression of the Ncx gene. To examine this possibility, we synthesized three kinds of double-stranded oligonucleotides (Ϫ1397/Ϫ1358, Ϫ1387/ Ϫ1368, and Ϫ1377/Ϫ1338) that span this region and did EMSA experiments with nuclear proteins from C1300, L, B16, and SH-SY5Y cells. A specific band was detected in nuclear proteins from C1300, SH-SY5Y, and B16 but not in proteins of L cells when the Ϫ1387/Ϫ1368 oligonucleotides were used (Fig.  7). This binding was inhibited with unlabeled Ϫ1387/Ϫ1368 oligonucleotides but not with the unlabeled Ϫ1347/Ϫ1328 oligonucleotides in C1300 murine neuroblastoma cells (Fig. 8), suggesting that nuclear factors bind to the DNA region between Ϫ1387 and Ϫ1368. In order to delineate the binding site within the 20-bp element, we tested a series of mutations in EMSA competition experiments. Unlabeled oligonucleotides with a mutation at Ϫ1384 or Ϫ1373 could not inhibit the binding, although oligonucleotides with a mutation at other sites could do so. Similar results were obtained using nuclear proteins from SH-SY5Y human neuroblastoma cells (data not shown). Therefore, the DNA element between Ϫ1387 and Ϫ1368 controls the cell type-specific expression of the Ncx, and nuclear factors that bind to this element are present in cell lines derived from neural crest tissues in mouse and human.
Next we examined effects of this element on Ncx promoter activity by preparing deletion mutants of the reporter plasmids. Deletion from Ϫ1387 to Ϫ1368 was introduced to the 1492NLuc and the 1414NLuc reporter plasmid (1492MLuc and 1414MLuc, respectively), and the promoter activity of those plasmids in C1300 neuroblastoma cells was measured. When the 1492MLuc or the 1414MLuc reporter plasmid was transfected, luciferase activity dramatically decreased to 1 ⁄5-1 ⁄8-fold compared with the case of wild type reporter plasmids (Fig. 9). These results also confirm that the DNA element from Ϫ1387 to Ϫ1368 is required for display of the proper expression level of the Ncx gene in neural crest-derived cells.

DISCUSSION
Expression of the Ncx gene is tightly regulated in neuronal lineage of neural crest-derived tissues (9,10). We confirmed the tissue-specific expression of Ncx in various cell lines, using a sensitive RT-PCR assay. This gene was expressed in neuronal cells derived from the neural crest such as neuroblastoma cell lines but not in fibroblast and B lymphoma cells. Although melanocytes are derived from the neural crest (11), the expression was never detected in B16 murine melanoma cells. Therefore, we attempted to identify the regulatory region governing the tissue-specific expression in the 5Ј-flanking region of the Ncx gene, and for this we transfected promoter-reporter fusion genes into those cell lines. First, we sequenced 5Ј-flanking regions of the murine and the human Ncx gene. Neither the TATA box nor the CAAT box consensus sequence was identified in the 5Ј-flanking regions of murine and human Ncx genes. The promoter regions of Hoxa-4 (26), Hoxb-4 (27), Hoxc-4 (28), Hoxc-6 (29), and Hoxc-8 (30) also lack the TATA box. Transient transfection experiments (Fig. 6A) suggest that the minimal promoter element of the murine Ncx gene locates in the region between Ϫ1153 and Ϫ73. The regions between Ϫ1153 and Ϫ1065 and between Ϫ682 and Ϫ522 in the Ncx gene were highly conserved in humans (data not shown), suggesting a presence of the common basic regulatory element in those regions. However, none of the known consensus sequence was found in those regions.
Second, we found that the region between Ϫ1414 and Ϫ1373 of the murine Ncx gene plays an important role in determining its cell type-specific expression (Fig. 6B). Furthermore, we found that a putative nuclear factor specifically bound to the region between Ϫ1387 and Ϫ1368, as determined by EMSA (Fig. 7), and nucleotides at the position Ϫ1384 or Ϫ1373 are critical for the binding. The sequence of this region, highly conserved between murine and human, does not resemble any known consensus bindng sequence of transcription factors by data base search. This factor was detected in nuclear proteins from not only murine and human neuroblastoma cell lines but also the B16 melanoma cell line. However, it was not detected in nuclear proteins from L fibroblast cells. Therefore, the DNA region between Ϫ1387 and Ϫ1368 may be the element for determining the neural crest tissue-specific expression of Ncx. Since the promoter activity was only enhanced in neuroblastoma cells and not in the B16 melanoma cell line, additional nuclear factors specific to the neuronal lineage cells are required to control lineage-restricted expression of the Ncx gene of neural crest-derived cells. Further characterization of this factor may reveal not only the molecular mechanism of tissuespecific expression of the Ncx gene but also mechanisms related to lineage commitment of neural crest to neuronal cells.
Ncx expression was augmented in C1300 or SH-SY5Y cells with RA stimulation. The 5Ј-flanking regions of the human and the murine Ncx gene carried the DR2-RARE. When C1300 cells were transfected with the 1298NLuc construct containing RARE and stimulated with RA to examine the inducibility, the luciferase activity with RA stimulation was up-regulated 3-fold FIG. 7. EMSA of nuclear proteins interacting with the DNA region between ؊1387 and ؊1368 of the murine Ncx sequence. The labeled oligonucleotide (Ϫ1387/Ϫ1368) probe was incubated with 10 g of nuclear extracts from B16, L cell, C1300, or SH-SY5Y.
FIG. 8. EMSA with various mutated oligonucleotides as a competitor. A, the labeled oligonucleotide (Ϫ1387/ Ϫ1368) probe was incubated with 10 g of nuclear extracts from C1300 murine neuroblastoma cell lines with various competitors. For competition, 10-fold (ϩ) or 50fold (ϩϩ) molar excess of the unlabeled oligonucleotide Ϫ1387/Ϫ1368 or Ϫ1347/ Ϫ1328 or mutated oligonucleotides were used. Ϫ1387/Ϫ1368-specific DNA-protein complexes are indicated by arrows. B, a sequence of the 20-bp element in mice and humans. Critical nucleotides for factor binding are underlined. above that without RA stimulation (data not shown), suggesting that this RARE is functional. Several Hox genes also have RARE in their noncoding region (31). For example, the 3Јflanking sequence of the Hoxa-1 gene has the DR5-RARE, the direct repeats separated by 5 nucleotides, that regulate its expression (32). The Hoxb-1 gene has two DR2-RAREs, which locate in the 5Ј-and 3Ј-flanking regions. RARE in the 3Јflanking region functions as a tissue-specific enhancer. Point mutation of RARE in the 3Ј side prevented the neuronal expression of Hoxb-1 in transgenic mice (33). On the other hand, RARE in the 5Ј-flanking region is an essential component of the repressor element to refine the Hoxb-1 expression to hind brain (34). Thus, those two RAREs control the tissue-specific expression of the Hoxb-1 gene. Although RARE in the 5Ј-flanking region of the Ncx gene is functional in neuroblastoma cells (Fig.  2), RARE is not essential for the cell type-specific expression of the Ncx gene in vitro (Fig. 6B). Further in vivo studies using transgenic mice should reveal the precise function of this RARE element.
RA is known to induce differentiation in some neuroblastoma cell lines. Both C1300 and SH-SY5Y underwent morphological changes upon stimulation with RA, and expression of Ncx was up-regulated in these cells after the stimulation. To examine the relation between the expression of Ncx and neuronal differentiation, differentiation was induced in neuroblastoma cell lines by various stimuli other than RA. C1300 cells are known to differentiate after serum deprivation (35). However, Ncx expression in C1300 cells was not induced after serum deprivation; rather, it was down-regulated (data not shown). SH-SY5Y cells differentiate upon nerve growth factor stimulation (36). During the differentiation after nerve growth factor stimulation, the amount of the NCX message remained unchanged (data not shown). Moreover, the Ncx-deficient mice showed normal differentiation of neural crest-derived neurons such as dorsal root ganglia, trigeminus ganglia, and enteric ganglia (10,14). Therefore, Ncx is not directly related to neuronal differentiation and may not be essential for differentiation of neural crest lineage cells.
Several transcriptional factors important for differentiation of neural crest have been identified, and most act as a master gene for lineage determination of neural crest cells. The Mash1 (mammalian achaete-scute homologue 1) gene, encoding basic helix-loop-helix transcription factor, is expressed only in precursors of all autonomic neurons, not in those of sensory neurons (37). A targeted mutation in Mash1 blocks the development of sympathetic, parasympathetic, and a subset of enteric neurons (38), indicating that this gene product plays a critical role in autonomic neurogenesis. The Neurogenin 1 (39) and the neurogenin 2 (also known as Math4A) (40), expressed in sensory but not in autonomic ganglia, are thought to be important for lineage determination of sensory cells. The Ncx gene is expressed in all of the neuronal lineage derived from neural crest (9). Data presented here demonstrate that the DNA element (between Ϫ1387 and Ϫ1368) located in the 5Ј-flanking region of the Ncx gene is critical for enhancer activity in neuroblastoma cells. In Ncx-deficient mice, megacolon, which occurs with an increased number of enteric neurons, resembles neuronal intestinal dysplasia, a disease in humans (10,14). Further characterization of a regulatory mechanism of tissuespecific expression of Ncx should reveal a genetic basis of this disease.