Abstract
Adrenomedullary chromaffin cells express at least two subtypes of acetylcholine nicotinic receptors, which differ in their sensitivity to the snake toxin α-bungarotoxin. One subtype is involved in the activation step of the catecholamine secretion process and is not blocked by the toxin. The other is α-bungarotoxin-sensitive, and its functional role has not yet been defined. The α7 subunit is a component of this subtype. Autoradiography of bovine adrenal gland slices with α-bungarotoxin indicates that these receptors are restricted to medullary areas adjacent to the adrenal cortex and colocalize with the enzyme phenylethanolamine N-methyl transferase (PNMT), which confers the adrenergic phenotype to chromaffin cells. Transcripts corresponding to the α7 subunit also are localized exclusively to adrenergic cells. To identify possible transcriptional regulatory elements of the α7 subunit gene involved in the restricted expression of nicotinic receptors, we isolated and characterized its 5′ flanking region, revealing putative binding sites for the immediate early gene transcription factor Egr-1, which is known to activate PNMT expression. In reporter gene transfection experiments, Egr-1 increased α7 promoter activity by up to sevenfold. Activation was abolished when the most promoter-proximal of the Egr-1 sites was mutated, whereas modification of a close upstream site produced a partial decrease of the Egr-1 response. Because Egr-1 was found to be expressed exclusively in adrenergic cells, we suggest that this transcription factor may be part of a common mechanism involved in the induction of the adrenergic phenotype and the differential expression of α-bungarotoxin-sensitive nicotinic receptors in the adrenal gland.
Neuronal nicotinic acetylcholine receptors (nAChR) sensitive to α-bungarotoxin (α-Bgt) are widely expressed in the central and peripheral nervous systems (for review, see Sargent, 1993; McGehee and Role, 1995). The molecular cloning of nAChR subunits that assemble into channels activated by cholinergic agonists and are blocked by α-Bgt (termed α7, α8, and α9;Couturier et al., 1990; Schoepfer et al., 1990; Elgoyhen et al., 1994) has allowed structural and functional studies of these nAChR subtypes. Although α-Bgt-sensitive nAChRs on neurons function as ligand-gated ion channels (Alkondon and Albuquerque, 1993; Zhang et al., 1994), their physiological role in situ have not yet been elucidated fully. It has been suggested that they could enhance fast excitatory transmission via a presynaptic mechanism (McGehee et al., 1995). In addition, they could have a growth or trophic role, because there is evidence that these receptors are involved in the guidance of nerve fibers (Pugh and Berg, 1994), inhibition of neurite extension in cultured cells (Chan and Quik, 1993), regulation of neurotrophic growth factor expression in rat hippocampus (Freedman et al., 1993), motoneuronal death (Hory-Lee and Frank, 1995), and modulation of proliferative responses in lung tumor cells (Codignola et al., 1994;Quik et al., 1994). Recently, postsynaptic responses for α-Bgt-sensitive nAChRs on neurons have been detected (Zhang et al., 1996).
We have shown previously that the α-Bgt-sensitive nAChRs from the chromaffin cells of the adrenal medulla contain an α7 subunit, homologous to those previously cloned from other species (Garcı́a-Guzmán et al., 1995). In the adrenal gland, acetylcholine released on stimulation of the splanchnic nerve activates nAChRs in chromaffin cells and triggers catecholamine secretion, but, interestingly, this functional response is not blocked by α-Bgt. This suggests the involvement of other nAChRs in the activation of exocytosis and would imply a segregation of functions for different nAChRs subtypes in the same cell. Alternatively, different nAChRs with distinct functions might be expressed in specific chromaffin cell subsets. Given any of these cases, complex molecular mechanisms may underlie nAChR subunit expression, which in turn may influence the physiological characteristics of a cellular group or a neuronal circuit. To investigate these mechanisms, we studied the expression of α-Bgt-sensitive nAChRs in the adrenal gland of several mammals, consistently finding that they were present only in adrenergic cells. This heterogenous distribution is attributable, at the very least, to restricted transcription of the α7 subunit gene in this chromaffin cell subset. Furthermore, we have identified transcription factor Egr-1 (Cao et al., 1990), which also is restricted to adrenergic cells, as the candidate potentially involved in this differential expression.
MATERIALS AND METHODS
α-Bgt autoradiography. The procedure followed for [125I]α-Bgt binding has been described previously (Clarke et al., 1985). Cryostat sections (15–20 μm) were incubated with 5 nm [125I]α-Bgt for 2 hr at room temperature. After nonbound toxin was washed, sections were dried and stored in cassettes for 48 hr at 4°C with Kodak X-OMAT x-ray film.
Immunohistochemistry. Rat tissue was perfused and processed as previously described (Domı́nguez del Toro et al., 1994). Cat and bovine adrenal glands were fixed by immersion in 4% paraformaldehyde in 75 mm sodium phosphate buffer, pH 7.35, and then processed in the same manner as the rat tissue. Cryostat sections (25–30 μm) were free-floated in continuously agitated solutions at 4°C. Phenylethanolamine N-methyl transferase (PNMT; EC2.1.1.28) was immunolabeled with a rabbit antiserum against bovine PNMT (Eugene Tech, Ridgefield Park, NJ), diluted 1:1000. Egr-1 was localized with an affinity-purified polyclonal antibody (588) raised in rabbit against a peptide that corresponds to the 14 C-terminal residues of mouse Egr-1 p82 (Santa Cruz Biotechnology, Santa Cruz, CA). The antibody was used at a 1:1000 dilution.
In situ hybridization. Sense and antisense digoxigenin-labeled riboprobes were synthesized with SP6 polymerase from the corresponding linearized template (pSPT18) with a Boehringer Mannheim (Barcelona, Spain) RNA labeling kit. Probes were directed to a region of the M3–M4 cytoplasmic loop unique to the α7 subunit [amino acids 315–382; see Garcı́a-Guzmán et al. (1995)for the bovine α7 sequence and Séguéla et al. (1993) for its rat counterpart]. Frozen adrenal gland tissue was cut in 10 μm sections, mounted on silanated slides, fixed in 4% formaldehyde (5 min at room temperature), deproteinated (10 min in 0.2 m HCl at room temperature), and dehydrated in ethanol. Hybridization was for 16 hr at 45°C in a buffer containing 50% formamide, 5× SSC, 1× Denhardt’s solution, 10% dextran sulfate, 0.2% SDS, 10 mm dithiothreitol, 250 μg/ml herring sperm DNA, and 250 μg/ml yeast tRNA. Digoxigenin-labeled probes were diluted (in the range from 1:25 to 1:400) with hybridization buffer to obtain a suitable and equivalent concentration of the different probes, as verified by dot blot assays. After hybridization, sections were washed for 2 × 30 min in 0.1× SSC at 60°C. Visualization of the probe was performed with anti-digoxigenin antiserum, as indicated by the manufacturer (Boehringer Mannheim).
Reverse transcription-PCR (RT-PCR) analysis. Total RNA was extracted by guanidinium isothiocyanate and acid phenol extraction (Promega, Barcelona, Spain) from adrenomedullary tissue dissected from areas near to or far from the adrenal cortex. α7 subunit transcripts were detected by a combined RT-PCR assay (Singer-Sam et al., 1990). Briefly, samples of total RNA (1.2 μg) in a final volume of 30 μl were reverse-transcribed with 1.25 U avian myeloblastosis virus (AMV) reverse transcriptase (10 min at 50°C); then PCR was performed for 30 cycles (1 min at 94°C, 2 min at 58°C, and 2.5 min at 72°C). The following primers were chosen: sense, 5′-CCAGGGCTGGTTTCCGAGA-3′; antisense, 5′TGTCCAAGTCATTTGTAGCCA-3′. The PCR fragment had 281 base pairs (bp) from the 5′ noncoding region to amino acid 80; the amplified DNA fragment expanded over three introns to rule out the possibility of amplifying a potential contamination of genomic DNA. In fact, RT-PCR performed in the absence of AMV reverse transcriptase (see Fig. 2B, inset) did not yield any fragment.
Localization of α7 transcripts in the adrenal gland. In situ hybridization of α7 transcripts in the bovine (A–C) and rat (D, E) adrenal glands. A, Hybridization with bovine α7 antisense probe. B, Hybridization with bovine α7 sense probe. C, Detail of A. B, Inset, Reverse transcription and further PCR amplification of α7 subunit mRNA isolated from sites close (Me) or far (Mi) from the adrenal cortex (Co); + and − indicate reaction performed in the presence or absence of reverse transcriptase, respectively. D, Hybridization with rat α7 antisense probe; Co, adrenal cortex;M, adrenal medulla. E, Detail ofD. Scale bars: 300 μm in A,B; 100 μm in D; 25 μm inC, E.
Isolation and analysis of the 5′ flanking sequence of the α7 subunit. A cDNA probe containing 87 bp of the 3′ end of exon 1 and 105 bp of the 5′ end of exon 2 was used to screen a bovine genomic library in EMBL-3 SP6/T7 (Clontech, Heidelberg, Germany), as previously described (Garcı́a-Guzmánet al., 1995). Several overlapping bacteriophage clones were purified and characterized. Clone λα7–11, which contained ∼17 kb of bovine genomic sequence, including exons 1 and 2, and ∼7 kb of 5′ flanking sequence, was characterized further.
5′ RACE analysis of 5′ mRNA ends. The rapid amplification of cDNA ends (RACE) was performed as described (Garcı́a-Guzmán et al., 1995). The first-strand cDNA, after reverse transcription of bovine adrenal medulla mRNA, was polyadenylated and amplified by two PCR rounds. An oligo dT was the sense primer, and the two antisense primers were 5′-TGTCCAAGTCATTTGTAGCCA-3′ (first round) and 5′-ATGTTGGTGGTCAACAC CTG-3′ (second round), coding for amino acids 74–80 and 67–73, respectively, of the α7 sequence. The resultant DNA was cloned into pBluescript (Stratagene, Heidelberg, Germany), and the sequence of the inserts was determined and compared with genomic DNA to establish the 5′ end of α7 mRNA.
RNase protection. A 305 bpSacII–BamHI fragment of the α7 gene that included 126 bp 5′ and 46 bp 3′ to exon 1 was subcloned into pBluescript. After linearization of the plasmid with EcoRI, a probe of 397 nucleotides was synthesized with T7 polymerase and [α-32P]CTP. For calibration, several other RNAs of known size were synthesized also. RNase protection experiments were performed with the RNase Protection Kit (Boehringer Mannheim). Protected fragments were separated on a 6% acrylamide/urea gel.
Plasmid constructions. All α7promoter–LUC gene fusions were made in the pGL2-Basic vector (Promega), introducing in its polylinker upstream of the luciferase gene the suitable α7 promoter fragments. These fragments were generated with restriction enzymes (as indicated in Fig. 4) and cloned directly into pGL2-Basic or subcloned first in pBluescript and then transferred to pGL2-Basic. The vector pGL2-Control, which express the luciferase gene under the regulation of the SV40 promoter and enhancer sequences, was used to check luciferase activity.
Determination of the α7 subunit gene transcription initiation site. A, Sequence of the 5′ end of the RACE product obtained by reverse transcription and PCR amplification of α7 subunit mRNA. B, Mapping 5′ ends of α7 subunit mRNA by RNase protection by using an α7 probe that included exon 1 and its 5′ and 3′ flanking regions. T, Undigested probe, 397 bases; N, probe digested in the presence of 5 μg of yeast tRNA; P, protected fragments using 5 μg of bovine adrenal medulla poly(A+) mRNA. The size of several RNA fragments used for calibration of the gel is indicated at the left.
Deletion analysis of the most promoter-proximal region was performed by partially digesting an ApaI–HindIII fragment (from −339 to +41 in the α7 sequence) with BstUI, which leaves blunt ends at GCGC sequences, and subcloning the partial digests into pBluescript vector cut with HincII–HindIII. Sequence analysis allowed selection of the appropriate fragments, which were cloned into pGL2-Basic and transfected further.
For site-directed mutagenesis of the putative Egr-1 sites at the −38 to −15 region of the α7 promoter (see Fig. 6), the following mutagenic primers were synthesized: mut1, 5′-CaaaaaCGGTCGGGGCGTGGGCG-3′; mut2, 5′-CGGGGGCGGTaaaaaCGTGGGCGCGCGCTGGG-3′; mut3, 5′-CGGGGGCGGTCGGGGaaaaaGCGCGCGCTGGGCTTTTTA-3′ (the introduced mutations are indicated in lower case letters). PCR (25 cycles of 94°C for 10 sec, 62°C for 30 sec, and 68°C for 45 sec, using the Expand kit from Boehringer Mannheim) was performed with these oligonucleotides and an antisense primer (5′-CTTTATGTTTTTGGCGTCTTCC-3′) that anneals to the pGL2-Basic vector downstream of the site of transcription initiation. PCR products were cloned into pBluescript, sequenced, and transferred further to pGL2-Basic.
Egr-1 induction of luciferase expression from the α7 subunit promoter. Neuro-2a and bovine chromaffin cells were cotransfected with pα7LUC plasmids and Egr-1 expression vectors encoding a nonfunctional (+Mut. Egr-1) or a functional (+Egr-1) Egr-1 protein. Numbers above the columns indicate the fold induction by Egr-1. Data are expressed as in Figure 5.
Cell culture and reporter assays. Neuro-2a mouse neuroblastoma cells were cultured in 90% Eagle’s minimal essential medium (EMEM), 10% fetal calf serum (FCS), and 2 mmglutamine. SH-SY5Y human neuroblastoma cells were grown in a 1:1 mixture of Ham’s F12 medium and EMEM containing 1% nonessential amino acids and 10% FCS. Chromaffin cells were isolated from bovine adrenal glands as described (Gandı́a et al., 1991) and cultured in 90% DMEM, 10% FCS, and 10 μm cytosine arabinoside to prevent fibroblast proliferation.
Plasmids were purified by Wizard maxipreps (Promega) and then banded in CsCl. All cell types were transfected by the calcium phosphate procedure (Graham and van der Eb, 1973). Cells (1–3 × 105) on 3.5 cm plates were incubated with 3 μg of pGL2 vector or an equivalent amount (in molar terms) of the different constructs derived from this vector and with 3 μg of β-galactosidase expression vector pCH110 (Pharmacia, Uppsala, Sweden) as a control of transfection efficiency. Two micrograms of Egr-1 expression plasmid (pCMVEgr-1) (Gupta et al., 1991) or of its mutated inactive form (pCMVEgr-1Δ331–374) also were present in the transfection mixture when the effect of this transcription factor was studied. Cells were harvested after 48 hr and lysed with reporter lysis buffer (Promega). Then β-galactosidase and luciferase were determined in the lysates with the corresponding assay systems (Promega). Luciferase activity was normalized to values obtained with the pGL2-Control vector in the same cell type. Relative light units obtained with pGL2-Control in the different cell types are as follows (mean of three individual experiments and corrected for transfection efficiency): chromaffin, 273; neuro-2a, 371; SHSY-5Y, 801.
Electrophoretic mobility shift assay. Crude nuclear extracts were prepared from chromaffin cells, as described by Schreiber et al. (1989). DNA fragments corresponding to the region −38 to +41, mutated or not at the putative Egr-1 sites, were obtained by digesting pBluescript subclones with EcoRI–HindIII and end-labeled by Klenow filling with [α-32P]dATP. The DNA–protein binding reaction volumes were 20 μl containing (in mm): 10 Tris, pH 7.5, 50 NaCl, 1 EDTA, and 1 dithiotreitol with 10% glycerol, 5 μg of bovine serum albumin, 2 μg of poly(dA-dT)·(dA-dT) (Pharmacia), 5 μg of nuclear extract protein, and 20,000 cpm of 32P-labeled probe. Reactions were incubated for 10 min at room temperature, labeled probe was added, and the incubation was continued for an additional 20 min. For competition studies the nuclear extract was incubated with the competing probe before the labeled probe during 20 min. Competition was performed either with the same purified fragments or with double-stranded oligonucleotides containing consensus sites for Egr-1 (Santa Cruz Biotechnology), Sp1, or cAMP response element binding proteins (CREB; Promega). Supershift assays were performed by incubating with 2 μl of anti-Egr-1 antibody (588X from Santa Cruz Biotechnology) or rabbit IgG (Sigma, Madrid, Spain) after probe addition, and the incubation continued for an additional 30 min at room temperature.
In some experiments purified Egr-1 protein fused to glutathioneS-transferase (GST) was used. The expression vector for the Egr-1/GST fusion protein was kindly provided by Dr. X. Cao (National University of Singapore), and the protein was expressed in bacteria and purified as described by Jain et al. (1996). The mutated inactive form of Egr-1 (Egr-1Δ331–374) fused to GST also was cloned in the same expression vector (pGEX-KG; Guan and Dixon, 1991) and used as a negative control.
RESULTS
α-Bgt binding sites and α7 subunit mRNA colocalize in chromaffin adrenergic cells
Autoradiography of adrenal gland slices labeled with [125I]α-Bgt was performed to localize α-Bgt-sensitive nAChRs in chromaffin cells (Fig.1). Three mammalian species were tested: cow, cat, and rat. The bovine adrenal medulla showed a characteristic labeling: the highest α-Bgt levels were found in the areas close to the adrenal cortex, at the periphery, and around the portal vein at the center (Fig. 1A). Labeling of the rat adrenal medulla was almost complete (Fig. 1C), whereas the cat showed a patchy distribution of binding sites, which was difficult to categorize (Fig. 1E). In all cases labeling was abolished by preincubation with nicotine (Fig. 1B,D,F). The fact that the adrenergic cell-rich rat adrenal medulla was labeled almost totally by α-Bgt, whereas its bovine counterpart, in which only ∼50% of cells are adrenergic, was partially labeled, suggested that α-Bgt-sensitive nAChRs were differentially expressed in adrenergic cells. To verify this, we performed immunolocalization of PNMT, the enzyme converting norepinephrine to epinephrine and, therefore, conferring the adrenergic phenotype. PNMT was present only in the bovine chromaffin cells close to the cortex or the portal vein (Fig. 1G), whereas most rat chromaffin cells were labeled (Fig. 1H). PNMT immunostaining in the cat gland had a patchy distribution (Fig. 1I). Hence, PNMT localization for all three mammals was shown to be very similar to the one observed for α-Bgt binding sites.
Localization of α-Bgt binding sites and PNMT in the adrenal gland. A–F, Autoradiography of [125I]α-Bgt binding to adrenal gland slices.A, B, Bovine; C,D, rat; E, F, cat. Total binding is observed in A, C, andE, whereas nonspecific binding, determined by incubating in the presence of 0.1 mm nicotine, is observed inB, D, and F.G–I, Immunolocalization of PNMT in the bovine (G), rat (H), and cat (I) adrenal glands. Scale bars: 1000 μm inA–G; 500 μm inH–I. Note that sizes are not comparable between species, because sections were not obtained from analogous locations in each gland.
Because the α7 subunit present in bovine chromaffin cells is probably a major component of α-Bgt-sensitive nAChRs (Garcı́a-Guzmán et al., 1995), it seemed possible that the restricted expression of α-Bgt-sensitive nAChRs was regulated by the transcriptional activity of the α7 subunit gene. This hypothesis was supported by a qualitative assay, the RT-PCR analysis of α7 mRNA levels in bovine adrenomedullary tissue (Fig.2B, inset) dissected from places near to (Me) or far (Mi) from the adrenal cortex, i.e., where α-Bgt-sensitive nAChRs were in high or low abundance, respectively. This experiment demonstrates that α7 mRNA was more abundant in the region of the medulla close to the cortex. This evidence was confirmed by in situ hybridization of bovine α7 transcripts: as shown by an antisense probe, they were present in the area adjacent to the adrenal cortex (Fig. 2A,Me; 2C), where adrenergic cells had been localized previously. The same probe in the sense orientation gave no labeling at all (Fig. 2B). On the other hand, most of the rat adrenal medulla was labeled with an antisense rat α7 probe (Fig. 2D,E), as in the case of PNMT immunolabeling. Thus transcriptional regulation of the α7 subunit gene seems to be involved in the differential expression of α-Bgt-sensitive nAChRs in adrenergic cells.
Functional analysis of the α7 subunit promoter
To examine the requirements for α7 subunit transcription, we isolated and analyzed its promoter region. A bovine genomic library was screened, and several overlapping clones were isolated. Their mapping showed the presence of the α7 5′ untranslated region (first exon) and further upstream sequences, as well as the second and third exons; the latter were separated by an intron of ∼30 kb. Subcloning and sequencing of an isolated EcoRI–PstI fragment from one of the clones (Fig. 3) and further comparison of this sequence to a database of binding sequences of known transcription factors revealed the main features of the promoter/regulatory region of the α7 gene: lack of a TATA box and the presence of perfect matches to three potential Sp1 sites, one Egr-1 site, one Myc-Max site, and one E-box (Fig. 3), all of them concentrated into ∼360 bases located 5′ to transcription initiation site. Additional sites with one mismatch, regarding perfect consensus sequences, also were observed.
The 5′ region of the bovine α7 subunit gene. A 2 kb EcoRI–PstI fragment of genomic clone λα7–11 carrying exons 1 and 2 (with the protein sequence indicatedunderneath in italics), the intron in between (within arrowheads, ▾), and 1500 bp of 5′ flanking sequence, was analyzed. The translation start codon isunderlined, and the major transcription initiation site (+1) is denoted by the arrow. Perfect matches to transcription factors Sp1 (solid lines rectangle), Egr-1 (long dashed lines rectangle), Myc-Max (short dashed lines rectangle), and E-box (hatched lines rectangle) are indicated.
The 5′ end of α7 mRNA was mapped by 5′ RACE and RNase protection analyses (Fig. 4), because primer extension experiments produced inconsistent results. In 5′ RACE only one DNA fragment was detected, the 5′ end sequence of which is shown in Figure 4A. This would map transcription initiation to a stretch of 5Ts and 3As (see black arrow in Fig. 3). RNase protection analyses with a 397 residue antisense riboprobe complementary to +180 to −125 and poly(A+) RNA from bovine adrenal medulla yielded two main protected fragments of ∼139 and 137 bases (Fig. 4B). The major one mapped to the first adenosine detected in the 5′ RACE fragment and was located downstream of a GC-rich region. Therefore, this tentatively was considered the main transcription initiation site (position +1,black arrow in Fig. 3), although other initiation sites may exist, because this is a typical feature of promoters without TATA boxes.
The transcriptional activity of the α7 promoter was examined with a series of constructs made by fusing the α7 5′ flanking region to a luciferase reporter gene. The nested set of α7 expression plasmids, containing 5900–128 bp of α7 promoter sequence plus 41 bp of 5′ noncoding region was transfected into a variety of cell types. Two of them, bovine chromaffin cells and the human neuroblastoma SH-SY5Y (which is of sympathetic adrenergic ganglion origin; see Biedler et al., 1973), endogenously express α-Bgt-sensitive nAChRs (Lukas et al., 1993), whereas the other, mouse neuroblastoma neuro-2a, does not (our unpublished results). All reporter constructs were found to be active in chromaffin and SH-SY5Y cells, showing much reduced activity in neuro-2a cells (Fig. 5). The smallest α7 promoter fragment tested, 128 bp, was sufficient to drive transcription, although larger fragments increased activity, especially in chromaffin cells. No luciferase expression was detected in any of the cell types tested when no promoter segments were present upstream of the luciferase gene (pGL2-Basic vector; data not shown).
Analysis of α7 gene promoter activity. The indicated cell types were transfected with each of the plasmids (namedpα7LUC, with the number of promoter base pairs included in the construct) containing the luciferase reporter under the control of the different fragments of the α7 subunit promoter and pCH110/β-galactosidase as a transfection efficiency control. All experimental points were run in triplicate. The mean ± SE (error bars) are given for three individual experiments. The restriction enzymes used to clone the different fragments also are shown in the scheme.
The similar pattern of expression in the bovine adrenal gland of the enzyme PNMT and the α7 subunit suggested that they share common transcriptional regulatory elements. This idea was reinforced when the promoter regions were compared: both were markedly G/C rich and contained several consensus sequences for transcription factors Sp1 and Egr-1. Moreover, it is known that these factors regulate PNMT expression in PC12 cells (Ebert et al., 1994; Ebert and Wong, 1995). To determine the possible regulation of α7 expression by Egr-1, we cotransfected different α7-luciferase constructs into neuro-2a and chromaffin cells with either a construct allowing strong expression of Egr-1 (pCMVEgr-1) or a control yielding a truncated, nonfunctional Egr-1 protein (pCMVEgr-1Δ331–374) (Fig.6). In neuro-2a cells the functional Egr-1 increased the luciferase activity up to fourfold with respect to that observed with the truncated Egr-1. In chromaffin cells activation was between three- and fivefold, depending on the construct used.
In both cell types activation already was observed with the smallest construct, which contains 128 bp of α7 promoter region. Several potential binding sites for Egr-1 are found in this region, although only one, located 31 bp 5′ to the transcription initiation site (+1) of the α7 gene, comprises a perfect match to the Egr-1 consensus binding element (GCGGGGGCG) (Gashler and Sukhatme, 1995). To demonstrate a possible direct interaction of this transcription factor with the α7 gene, we performed several deletions, studying in each case the activation produced by Egr-1 expression on transfection into chromaffin cells (Fig. 7A). Thus, removal of nucleotides (nt) −128 to −78 did not affect Egr-1 activation (p77α7LUC, Fig. 7A), which was even larger than with previous constructs. The promoter basal activity was similar to that supported by p128α7LUC. Further removal of nt −77 to −39 (p38α7LUC, Fig. 7A) resulted in ∼30% decrease in promoter activity (relative to p77α7LUC). However, Egr-1 was able to activate this shortened construct to a similar extent. Additional removal of nt −38 to −16 (p15α7LUC, Fig. 7A) drastically reduced promoter activity (∼20% of p77α7LUC), which was not modified by the presence of Egr-1. A construct lacking the site of transcription initiation (p+7α7LUC, Fig. 7) was used to demonstrate further that downstream sequences as well as vector sequences were not involved in the observed effects. Therefore, these experiments indicate that the region between nt −38 to −16, although probably lacking important elements for basal activity, can support Egr-1 activation. This region contains several potential sites for Egr-1 (boxed and termed 1, 2, and 3 in Fig. 7B), which were mutated in an attempt to identify the area of interaction between this factor and the α7 gene. When site 1, starting at the 5′ end, was abolished (p38α7LUCmut1, Fig. 7A), the activation produced by Egr-1 decreased with respect to that observed with the wild construct but was still significant. This site also was modified in p128α7LUC by mutating the area at −37 to −39 from GCG to CAT, yielding a similar degree of Egr-1 activation to the one observed for p38α7LUCmut1 (data not shown). Mutation of site 2 (p38α7LUCmut2, Fig. 7A) did not produce any decline on Egr-1 activation. Finally, mutation of site 3 (p38α7LUCmut3, Fig. 7A) totally abolished the effect of Egr-1. Therefore, two areas (at −38 to −31 and −24 to −16) seem to be responsible for the effect of Egr-1, emerging as more important the most proximal to the initiation start site (site 3, Fig. 7B). In addition, it is important to note that the three mutant constructs showed reduced basal promoter activity, indicating that the whole region is involved in providing essential elements for transcriptional activity.
Two sites in the proximal promoter region of the α7 gene are involved in Egr-1 activation of transcription initiation.A, Deletion constructs (from p77α7LUCto p+7α7LUC) were obtained, using BstUI restriction enzyme, as described in Materials and Methods. The α7 promoter sequences contained between the constructs p38α7LUC and p15α7LUC were mutated by PCR (mut1–mut3). Each construct was cotransfected with Egr-1 expression vectors into chromaffin cells, and its activity was measured. Numbers above the columns indicate the fold induction by Egr-1. Luciferase activity was normalized to values obtained with the p77α7LUC construct. Data are expressed as in Figure 5.B, The region between −38 and −15 contains several putative binding sites for Egr-1 (boxed inp38). Five nucleotides of each potential element were mutated as indicated in mut1, mut2, andmut3 to yield constructs analyzed in transfection (p38α7LUCmut1–3 in A) and bandshift (Figs. 8, 9) experiments.
To examine further the binding sites of Egr-1 detected in the −38 to −16 region of the α7 promoter region, we performed gel mobility shift assays. DNA fragments carrying the wild-type −38/+41 promoter region and the corresponding mutants (mut1, mut2, andmut3, Fig. 7B) were labeled and used either with nuclear extracts from chromaffin cells or with purified Egr-1 protein. A prominent retarded band was observed when Egr-1 was used either with the wild-type DNA fragment or with mut1 or mut2 (Fig.8A, lanes 2, 5, and8, respectively, filled dot). The same band also was obtained with mut3 but with much less intensity (Fig.8A, lane 11). Because a mutant, inactive Egr-1 protein was unable to bind any of the DNA fragments (Fig.8A, lanes 3, 6, 9, and 12), we deduced that Egr-1 was binding specifically to the DNA probes. Moreover, the fact that mut3 showed very reduced capacity to bind Egr-1 confirms the significance, with respect to Egr-1 activation, of the Egr-1 binding site most proximal to the transcription initiation site (site3). When nuclear extracts were used, several retarded bands were observed (Fig. 8B). One of them (indicated by adot) was retarded further by an anti-Egr-1 antibody (arrowhead Fig. 8B, lane 3), but not by an unrelated IgG. The same band also was observed with mut1, mut2, and mut3, but, again, the latter fragment showed reduced binding capacity. Interestingly, mut3 revealed the presence of a new band, only weakly observed with the other constructs, that suggests the existence of another protein, the binding of which is possible only when Egr-1 is not bound. Alternatively, the introduced mutation would create a new recognition site not present in the other fragments.
Gel mobility shift assays. A, Labeled wild-type (WT-38/+41) and mutated (Mut1, Mut2, Mut3; see Fig. 7) DNA fragments were used as gel mobility shift probes in the presence of 0.3 μg of purified Egr-1/GST fusion protein (E) or its mutated, inactive counterpart (Δ). B, The same DNA fragments were used with 5 μg of crude chromaffin cell nuclear extracts (N). The results of a supershift assay with Egr-1 specific antibody (E) or control IgG (I) are shown in lanes 3 and 4, respectively, of this panel.Dot, Egr-1 bound probe; arrowhead, supershifted Egr-1 bound probe; −, probe run in the absence of protein extracts. C, D, The gel mobility assay was run, using the WT-38/+41 DNA fragment as the labeled probe and nuclear extracts from chromaffin cells. The first lane in each series (−) is always without added competitor. C, Competitor DNA fragments WT-38/+41, Mut1, and Mut3were added in 15-, 45-, 150-, or 300-fold excess. D, Competitor oligonucleotides with consensus sites for Egr-1 (E), Sp1 (Sp1), or CREB (CREB) were added in 400-fold molar excess.Dot, Egr-1 bound probe.
To evaluate further the specificity of Egr-1 binding to the proximal promoter region of the α7 gene, we performed competition experiments with oligonucleotides containing Egr-1, Sp1, or CREB consensus sequences or the same DNA fragments used in the bandshift assays (Fig.8C,D). Preincubation of the nuclear extracts with increasing amounts (ranging from a 15- to a 300-fold excess) of the WT-38/41 fragment gradually decreased the intensity of the different complexes (Fig. 8C, lanes 2–5). Competition with mut1 produced a similar decrease (Fig. 8C, lanes 6–9), although the band corresponding to Egr-1 appeared slightly more intense (compare, for instance, the corresponding lanes 2 and 6 in Fig.8C), suggesting that the absence of the Egr-1 site 1 in this fragment caused a decrease in its competition ability. The mut3 fragment was the less effective competitor, although it also induced a decrease in band intensity (Fig. 8C, lanes 10–13), suggesting that (1) the proximal Egr-1 site 3, absent in this fragment, has a higher affinity for the factor, and (2) the distant Egr-1 site, although weaker, can bind Egr-1 and induce competition. Finally, when competition was performed with a 400-fold molar excess of synthetic oligonucleotides (Fig. 8D, lanes 1–4), only the Egr-1 consensus fragment abolished the formation of the Egr-1 complex. Therefore, Egr-1 specifically interacts with the most promoter-proximal region of the α7 gene.
The above data also suggest that Egr-1 categorizes two nonequivalent sites in this area. Experiments with purified Egr-1 confirmed this assertion (Fig. 9). Thus, by raising the amount of protein used to shift the nonmutated DNA fragment WT −38/+41, it was possible to observe the formation of a second, larger complex (Fig. 9, lanes 2 and 3, double dot). This complex was not observed when any of the two Egr-1 sites (sites1 and 3) were mutated (Fig. 9, lanes 5–6 and lanes 8–9), suggesting that it arises from the binding of Egr-1 polypeptides to both sites. The different intensity of the bands obtained with mut1 and mut3 DNA fragments (compare lanes 5 and 6 vs lanes 8 and9) indicated, again, that the Egr-1 affinity of site 1, which is kept intact in mut3, is much lower than the one of site 3. If this is the case, one could assume that Egr-1 binds first to site 3, and, subsequently, a second Egr-1 binds to site 1. Moreover, the binding of the second Egr-1 molecule might be facilitated when the first one is already bound, because the band intensity of the larger complex, which would result on binding of Egr-1 to site 1 when site 3 is already occupied (Fig. 9, lane 3, double dot), is clearly stronger than the one produced on binding of Egr-1 to site 1 when site 3 is not available for binding (Fig. 8, lane 9). Accordingly, the enhancing activity of Egr-1 may take place via a cooperative interaction with two close elements at the α7 gene promoter region.
Purified Egr-1 binds two nonequivalent sites within the most promoter-proximal region of the α7 gene. Labeled wild-type (WT-38/+41) and mutated (Mut1,Mut3) DNA fragments were used as gel mobility shift probes in the presence of 0.7 and 4 μg of purified Egr-1/GST fusion protein. Dot, Egr-1 bound to a single site;double dot, Egr-1 bound to two sites; −, probe run in the absence of protein.
Egr-1 transcription factor is expressed only in adrenergic cells
If Egr-1 activates α7 and PNMT transcription, it should be present in the cells that express these two proteins. Immunolocalization of Egr-1 in the bovine adrenal gland (Fig.10A) with a polyclonal antibody against this transcription factor shows its expression in only the areas previously demonstrated to express PNMT, α-Bgt-sensitive nAChRs, and α7 transcripts, i.e., the domains close to the cortex and the portal vein. In these cells Egr-1 immunostaining was concentrated in the cell nuclei (Fig. 10B), corresponding to a transcriptional factor. In contrast to this restricted localization, Egr-1 labeling of the rat adrenal medulla was almost complete in accordance with PNMT and α-Bgt binding localization. In this case, Egr-1 also was found in the cell nuclei (Fig. 10C).
Egr-1 immunolocalization in bovine and rat adrenal gland sections. A, A general view of the bovine gland shows that immunoreaction is present in the areas previously demonstrated to express α-Bgt-sensitive nAChRs and PNMT.B, Detail of the bovine gland showing nuclear labeling.C, High magnification of immunostaining in the rat adrenal medulla also shows nuclear labeling of Egr-1. Scale bars: 1000 μm in A; 50 μm in B; 25 μm inC.
DISCUSSION
At least two classes of nAChRs have been identified on chromaffin cells. One class, probably formed by α3 (Criado et al., 1992), α5, and β4 subunits, is present in all chromaffin cells of the adrenal medulla (Campos-Caro et al., 1997) and seems to be responsible for mediating catecholamine secretion. The other binds α-Bgt and contains α7 subunits (Garcı́a-Guzmán et al., 1995). It was demonstrated only recently that these receptor subtype functions as a nAChR (Alkondon and Albuquerque, 1993; Zhang et al., 1994, 1996). The first conclusion that can be drawn from the present study is that, in contrast with the other nAChR subtype, α-Bgt-sensitive nAChRs are localized heterogeneously in the adrenal medulla at sites where the enzyme PNMT also is present. Moreover, α7 subunit transcripts also are detected in the same place. Cell type-specific gene expression frequently depends on cis-acting DNA sequences that are primarily located within the 5′ flanking region of the gene. Activatory or inhibitory trans-acting proteins can bind to these sequence elements and regulate transcription. Thus, the 5′ flanking region of the α7 subunit was isolated and characterized. The 128 bp sequence immediately upstream from the transcription initiation site was able to drive transcription of reporter genes in cells that express endogenous levels of the α7 gene, such as chromaffin cells (Garcı́a-Guzmán et al., 1995) and SH-SY5Y neuroblastoma cells (Peng et al., 1994). Thus it was deduced that this region seems to contain the essential sequence elements required to establish a core promoter, although other elements further upstream also may be important, as indicated by the fact that the larger α7 promoter constructs yielded increased activity. The core promoter region does not contain TATA and CAAT boxes, but it does coincide with a G+C-rich domain, characteristic of many promoters, including the chick α7 promoter (Matter-Sadzinski et al., 1992). When this domain was inspected for motifs that might be relevant for the regulation of transcription, several putative binding sites for transcription factors Sp1 and Egr-1 were detected. Because Sp1 generally is expressed constitutively, our interest was directed to Egr-1 (also termed NGF1-A, krox-24, and zif268), an immediate early gene transcription factor known to activate PNMT gene expression in PC12 cells (Ebert et al., 1994; Ebert and Wong, 1995). Egr-1 activated the different α7 gene constructs tested in chromaffin and neuro-2a cells, indicating the requirement of this transcription factor for α7 gene expression.
The minimal promoter region of the α7 gene, which shows activity enhanced by Egr-1 (−38 to −16), contains several sequences that are similar to Egr-1 binding sites. Egr-1 binds to this region at two locations. Interestingly, the site located −24 to −16, which contains a 1 bp mismatch regarding the Egr-1 consensus sequence, seems to have the highest affinity in bandshift experiments; consequently, its modification abolishes Egr-1 activation in transfection experiments. By contrast, the more distal site, which comprises a perfect Egr-1 consensus sequence, seems to play a secondary role, because Egr-1 activation is not eliminated totally when this site is modified. Accordingly, bandshift experiments showed that the site is recognized by Egr-1, although with lower affinity. These facts suggest that other elements, perhaps additional factors and/or DNA sequences outside the proximal Egr-1 binding site, may favor the binding of Egr-1 to this site. On the other hand, it may worth mentioning that the sequence of the proximal site (GCGTGGGCG), although containing a 1 bp mismatch, has been shown previously to be able to bind with high affinity an Egr-1 zinc finger peptide with which it was cocrystalized (Pavletich and Pabo, 1991). In any case, we propose that Egr-1 exerts its enhancing action via a mechanism that involves direct binding to two very close nonequivalent sites. However, this fact does not preclude the involvement of additional factors on α7 gene transcription. Consider, for instance, that the maximal activity obtained with Egr-1 in neuro-2a cells is always lower than the one exhibited by chromaffin cells, suggesting that, in the latter case, additional factors, probably not available in neuro-2a cells, are needed for enhanced transcription. This fact also is suggested by previous studies on the expression of the chicken α7 subunit, which described the ubiquitous activity of its gene promoter in undifferentiated neural cells and mesoderm stem cells (Matter-Sadzinski et al., 1992), as well as the presence of α7 transcript and protein in embryonic muscle (Corriveau et al., 1995). Thus α7 regulatory elements would acquire their cell specificity in the course of development (Matter-Sadzinski et al., 1992), depending on the availability of determined factors.
The Egr-1 polypeptide was localized exclusively to the same chromaffin cell subset as PNMT and α-Bgt-sensitive nAChRs; thus it seems likely that Egr-1 regulation is a common enhancing mechanism of transcription for these proteins. The transcription factors Egr-1–Egr-4 constitute a family of DNA binding proteins termed “early growth response genes,” which are induced rapidly and transiently in a variety of cells by mitogenic stimulation and by signals involved in cell growth and differentiation (for review, see Gashler and Sukhatme, 1995). In the brain Egr-1 displays manifest basal expression (Mack et al., 1990), which is highly responsive to neuronal stimulation. The level of Egr-1 transcript expression increases when high-frequency electric stimulation is associated with the induction in vivo of long-term potentiation. This increase is inhibited by NMDA receptor antagonists (Cole et al., 1989; Wisden et al., 1990) and suggests the potential involvement of Egr-1 in synaptic plasticity. In the bovine adrenal gland a specific stimulation of Egr-1 expression seems very plausible, because localization of adrenergic cells expressing Egr-1 (and also PNMT and α7 nAChRs) is so peculiar: they are only in the region of the adrenal medulla close to the adrenal cortex and to the portal vein. This might indicate that factors arising from the cortex diffuse into the nearby area of the adrenal medulla and induce, directly or indirectly, the expression of Egr-1, which, in turn, regulates the expression of PNMT and nAChRs. If this scenario is correct, glucocorticoids are strong candidates for substances that are released from the cortex and then act directly on nearby chromaffin cells, especially because the glucocorticoid receptor has been localized only to the cytoplasm of adrenergic cells (Ceccatelli et al., 1989). Interestingly, a functional glucocorticoid response element in the 5′ flanking region of the enzyme PNMT has been described (Ross et al., 1990). Moreover, by using a PC12 cell line that was stably expressing a glucocorticoid receptor, Ebert et al. (1994) subsequently demonstrated a concerted activation of Egr-1 and glucocorticoids. Preliminary experiments indicate that α7 promoter activity increases substantially when α7 constructs and a glucocorticoid receptor vector are cotransfected into chromaffin cell cultures (M. Criado, C. Carrasco-Serrano, S. Viniegra, unpublished observations), reinforcing our hypothesis that a direct or indirect interaction via the glucocorticoid receptor regulates α7 expression.
A final important question highlighted by this study is the functional significance of α-Bgt-sensitive nAChRs in adrenergic cells. A potential role is indicated by the high Ca2+permeability of this receptor subtype (Séguéla et al., 1993), allowing the capacity for substantial control of intracellular Ca2+ levels, which may modulate exocytosis. Alternatively, the expression of α7 nAChRs in adrenergic cells may provide them with an additional mechanism, perhaps faster and specifically directed to trigger epinephrine release. This hypothetical mechanism might have been obscured until now if we consider that (1) α-Bgt-sensitive receptors have very rapid rates of desensitization, and the method of agonist application when testing catecholamine release would have not been fast enough to detect responses caused by α-Bgt-nAChR (see Zhang et al., 1994), and (2) the existence of other nAChR subtypes that desensitize more slowly than does α-Bgt-nAChR (Campos-Caro et al., 1997) would have darkened responses from the latter. Whatever the case, many studies point out significant differences in the secretory process of adrenergic and noradrenergic cells (Choi et al., 1993; Langley and Grant, 1995; Lomax et al., 1997), and the restricted expression of α-Bgt-sensitive nAChRs may contribute to this heterogeneity.
Footnotes
This work was supported by Grants from the Ministries of Education (Dirección General de Investigación Cientifica y Ténica: PB92-0346, PB95-0690, and PB93-0931) and Health (Fis 95/1672) of Spain, the Commission of the European Economic Community (SC1*CT91-0666), Generalitat Valenciana (GV-D-VS-20-158-96), and the Wellcome Trust (039284). E.D. del T. was the recipient of a predoctoral fellowship from the Ministry of Education of Spain. F.I.S. was a postdoctoral fellow of the Wellcome Trust. C.C.-S. was the recipient of a fellowship from Generalitat Valenciana. We thank Y. Wang, V. P. Sukhatme, X. Cao, and R. Bravo for Egr-1 plasmids; J. Strotmann for communicating his in situ hybridization protocol to us; and A. Garcı́a for the suggestion of using PNMT antisera to label adrenergic cells. The excellent technical assistance of Eva Martı́nez is also appreciated.
Correspondence should be addressed to Dr. M. Criado, Department of Neurochemistry and Instituto de Neurociencias, Universidad Miguel Hernandez, Campus de San Juan, 03550 San Juan, Alicante, Spain.
Dr. Smillie’s present address: Wythenshawe Hospital Research Centre, Manchester M23 9LT, England.
