Multiple Functional Sp1 Domains in the Minimal Promoter Region of the Neuronal Nicotinic Receptor α5 Subunit Gene*

The α5 subunit is a component of the neuronal nicotinic acetylcholine receptors, which are probably involved in the activation step of the catecholamine secretion process in bovine adrenomedullary chromaffin cells. The promoter of the gene coding for this subunit was isolated, and its proximal region was characterized, revealing several GC boxes located close to the site of transcription initiation (from −111 to −40). Deletion analysis and transient transfections showed that a 266-base pair region (−111 to +155) gave rise to ∼77 and 100% of the maximal transcriptional activity observed in chromaffin and SHSY-5Y neuroblastoma cells, respectively. Site-directed mutagenesis of five different GC motifs indicated that all of them contribute to the activity of the α5 gene, but in a different way, depending on the type of transfected cell. Thus, in SHSY-5Y cells, alteration of the most promoter-proximal of the GC boxes decreased α5 promoter activity by ∼50%, whereas single mutations of the other GC boxes had no effect. In chromaffin cells, by contrast, modification of any of the GC boxes produced a similar decrease in promoter activity (50–69%). In both cell types, however, activity was almost abolished when four GC boxes were suppressed simultaneously. Electrophoretic mobility shift assays using nuclear extracts from either chromaffin or SHSY-5Y cells showed the specific binding of Sp1 protein to fragment −111 to −27. Binding of Sp1 to the GC boxes was also demonstrated by DNase I footprint analysis. This study suggests that the general transcription factor Sp1 plays a dominant role in α5 subunit expression, as has also been demonstrated previously for α3 and β4 subunits. Since these three subunits have their genes tightly clustered and are expressed in chromaffin cells, probably as components of the same receptor subtype, we propose that Sp1 constitutes the key factor of a regulatory mechanism common to the three subunits.

The ␣5 subunit is a component of the neuronal nicotinic acetylcholine receptors, which are probably involved in the activation step of the catecholamine secretion process in bovine adrenomedullary chromaffin cells. The promoter of the gene coding for this subunit was isolated, and its proximal region was characterized, revealing several GC boxes located close to the site of transcription initiation (from ؊111 to ؊40). Deletion analysis and transient transfections showed that a 266base pair region (؊111 to ؉155) gave rise to ϳ77 and 100% of the maximal transcriptional activity observed in chromaffin and SHSY-5Y neuroblastoma cells, respectively. Site-directed mutagenesis of five different GC motifs indicated that all of them contribute to the activity of the ␣5 gene, but in a different way, depending on the type of transfected cell. Thus, in SHSY-5Y cells, alteration of the most promoter-proximal of the GC boxes decreased ␣5 promoter activity by ϳ50%, whereas single mutations of the other GC boxes had no effect. In chromaffin cells, by contrast, modification of any of the GC boxes produced a similar decrease in promoter activity (50 -69%). In both cell types, however, activity was almost abolished when four GC boxes were suppressed simultaneously. Electrophoretic mobility shift assays using nuclear extracts from either chromaffin or SHSY-5Y cells showed the specific binding of Sp1 protein to fragment ؊111 to ؊27. Binding of Sp1 to the GC boxes was also demonstrated by DNase I footprint analysis. This study suggests that the general transcription factor Sp1 plays a dominant role in ␣5 subunit expression, as has also been demonstrated previously for ␣3 and ␤4 subunits. Since these three subunits have their genes tightly clustered and are expressed in chromaffin cells, probably as components of the same receptor subtype, we propose that Sp1 constitutes the key factor of a regulatory mechanism common to the three subunits.
Nicotinic acetylcholine receptors (nAChRs) 1 are members of the gene superfamily of neurotransmitter-gated ion channels (1,2). These multimeric receptors are heteromers or, in some cases, homomers of subunits (␣2-␣9 and ␤2-␤4) that exhibit well defined and restricted expression patterns in vivo (1). The diversity of neuronal nAChRs arises, at least in part, from the different combinations of subunits able to form functional nAChRs (3). Thus, it is clear that their differential expression affects the electrophysiological and pharmacological properties of the resultant receptors (4). Moreover, potential changes in subunit expression in response to modulation of synaptic function might have important consequences on the signals transduced by nAChRs (5).
To understand how the regional and developmental expression of nAChR subunits is controlled, we have started to analyze the transcriptional mechanisms that regulate expression of the nAChR subunits expressed in chromaffin cells of the bovine adrenal gland. This cell type represents a relevant and accessible model in which to study a particular neuronal nAChR subtype with a well defined function. Previously, we cloned the bovine ␣3 (6), ␣5 and ␤4 (7), and ␣7 (8) subunits of neuronal nAChRs, which are expressed in chromaffin cells as components of the two nAChR subtypes typically present at the peripheral nervous system (9). We have also shown that nAChRs formed by ␣7 subunits are differentially expressed in adrenergic cells (10), probably as the result of transcriptional regulation, whereas ␣3, ␣5, and ␤4 subunits have a less restricted distribution in adrenergic and noradrenergic cells (7). Interestingly, the ␣3, ␣5, and ␤4 subunit genes have been found tightly clustered in the avian (11) and mammalian (12) genomes, with the ␣3 and ␣5 genes contiguous and having opposed transcription polarity. A number of studies have concentrated on the transcriptional regulation of the ␣3 and ␤4 subunits. Deneris and co-workers (13,14) have shown that the POU domain factor SCIP/Tst-1 is able to activate the ␣3 subunit promoter, probably as a consequence of protein-protein interactions at the level of the basal transcriptional machinery. Furthermore, an enhancer located in the 3Ј-untranslated exon of the ␤4 subunit (15,16), at the ␤4/␣3 intergenic region, activates transcription from the ␣3 and ␤4 subunit promoters in a cell type-specific manner, possibly via a novel ETS domain factor, Pet-1, whose expression is almost restricted to the adrenal medulla (17). Considerable effort has also been dedicated to the transcriptional regulation of the ␤4 subunit. Thus, Gardner and co-workers (18) have shown that Pur␣ interacts with a 19-bp element in the ␤4 promoter. In addition, Sp1 (19) and Sp3 (20) transactivate the ␤4 promoter in a synergistic way, an effect possibly mediated by heterogeneous nuclear ribonucleoprotein K, which affects the transactivation of ␤4 promoter activity by Sp1 and Sp3 differentially (21). As no information has been available until now for the promoter of the ␣5 subunit, we have chosen to focus on it, with the aim of finding a possible link in the regulation of the three subunits. This study reports that at least five positive regulatory elements exist in the ␣5 promoter proximal region. These elements, all of them GC boxes, were shown to interact with Sp1. Since the ␣3 promoter is also the target of Sp1 (22), we suggest the possible involvement of this transcription factor in a regulatory mechanism common to the ␣3, ␤4, and ␣5 subunits.

EXPERIMENTAL PROCEDURES
Isolation and Analysis of the 5Ј-Flanking Sequence of the ␣5 Subunit-A cDNA probe containing 152 bp of the 5Ј-end of exon 2 was used to screen a bovine genomic library in EMBL-3 SP6/T7 (CLONTECH, Heidelberg, Germany) as described previously (8). Several overlapping bacteriophage clones were purified and characterized.
RNase Protection-Poly(A) ϩ RNA was directly selected from a lysate of bovine adrenal medulla by oligo(dT)-Dynabeads (Dynal, Oslo, Norway) and used in the RNase protection experiments. Probes were generated with SP6 and T7 polymerases (Boehringer Mannheim, Barcelona, Spain), [␣-32 P]CTP (Amersham Pharmacia Biotech, Madrid, Spain), and the corresponding linearized templates (in the pSPT18 vector, Boehringer Mannheim). A 431-bp BglII-PvuII fragment of the ␣5 gene that included 163 bp 5Ј to exon 1 and 249 bp downstream in the same exon was subcloned into the BamHI and HincII sites of pSPT18. After linearization of the plasmid with EcoRI, a probe of 479 nucleotides was synthesized with SP6 polymerase. As control, a cRNA sense fragment of 368 nucleotides was synthesized by linearizing the same template with HincII and using T7 polymerase. This cRNA protected a fragment of 358 nucleotides upon RNase treatment. Parallel experiments were carried out with a smaller probe that overlapped the first one. For this purpose, a 341-bp fragment of the ␣5 gene that included 319 bp of the 5Ј-end of the first probe (downstream of the HincII site mentioned above) was also subcloned into pSPT18. After linearization of the plasmid with EcoRI, a probe of 403 nucleotides was synthesized with SP6 polymerase. The same control sense cRNA used above then produced a protected fragment of 328 nucleotides when used instead of adrenal medulla RNA (see Fig. 2 for further explanations). RNase protection experiments were performed using an RNase protection kit (Boehringer Mannheim) as indicated by the manufacturer. Protected fragments were separated on a 7 M urea and 6% acrylamide gel along with several other labeled RNAs of known size, which were also synthesized and used for calibration.
Plasmid Constructions-All ␣5 promoter-luciferase gene fusions were made in the pGL2-Basic vector (Promega, Madison, WI), introducing in its polylinker, upstream of the luciferase gene, the suitable ␣5 promoter fragments. These fragments were generated with restriction enzymes and directly cloned into pGL2-Basic or subcloned first in pBluescript and then transferred to pGL2-Basic. The vector pGL2-Control, which expresses the luciferase gene under the regulation of the SV40 promoter and enhancer sequences, was used to check luciferase activity. Deletion analysis of the most promoter-proximal region was performed by generating polymerase chain reaction fragments with suitable sense oligonucleotides and an antisense primer (5Ј-CTTTAT-GTTTTTGGCGTCTTCC-3Ј) that anneals to the pGL2-Basic vector downstream of the site of transcription initiation.
The basic strategy for site-directed mutagenesis of the different elements in region Ϫ111 to Ϫ40 of the ␣5 promoter (see Fig. 6) consisted of the following steps. (a) We performed polymerase chain reaction (25 cycles at 94°C for 10 s, 62°C for 30 s, and 68°C for 45 s) amplification of p111␣5LUC (or its single or double mutants when double or quadruple mutants were desired, respectively) with appropriate mutagenic primers in the sense orientation, which generated restriction sites useful for further mutant constructions and to confirm mutagenesis. We used the same oligonucleotide mentioned above as antisense primer. The introduced mutations are indicated in lowercase letters in Fig. 6A (sites 1-4) and Fig. 10A (site 5). The mutant sequences did not create any known binding site for transcription factors as predicted by the MatInspector data base search (23). (b) Polymerase chain reaction products were cloned into pBluescript, sequenced, and transferred to the appropriate construct into the pGL2-Basic vector.
Plasmids were banded in two gradients of CsCl. Both cell types were transfected by the calcium phosphate procedure (25). Chromaffin cells on 48-well plates (5 ϫ 10 5 cells/well) were incubated with 0.75 g of pGL2 vector or an equivalent amount (in molar terms) of the different constructs derived from this vector and with 0.75 g of ␤-galactosidase expression vector pCH110 (Amersham Pharmacia Biotech, Barcelona) as a control of transfection efficiency. SHSY-5Y cells (10 5 cells/well) or COS cells (5 ϫ 10 4 cells/well) on 24-well plates were incubated with 1.5 g of the different ␣7 constructs and 1.5 g of pCH110 per well. Cells were harvested after 48 h and lysed with reporter lysis buffer (Promega). ␤-Galactosidase and luciferase were then determined in the lysates with the corresponding assay systems (Promega). Luciferase activity was normalized to values obtained with the p163␣5LUC (see Fig. 3) or p111␣5LUC (see Fig. 6) plasmid in the same cell type. When comparing ␣5, ␣7, and ␤4 promoters (see Fig. 4), representative constructs for each of the subunits, giving the maximal promoter activity, were chosen. They are indicated in the corresponding figure legends.
Electrophoretic Mobility Shift Assay-Crude nuclear extracts were prepared from chromaffin and SHSY-5Y cells as described by Schreiber et al. (26). The DNA fragment corresponding to region Ϫ111 to Ϫ27 was obtained by digesting pBluescript subclones with EcoRI-KpnI and endlabeled by Klenow filling with [␣-32 P]dATP. The DNA-protein binding reaction volumes were 20 l containing 10 mM Tris, pH 7.5, 50 mM NaCl, 1 mM EDTA, 1 mM dithiothreitol, 10% glycerol, 5 g of bovine serum albumin, 2 g of poly(dA-dT)⅐(dA-dT) (Amersham Pharmacia Biotech), 2 g of nuclear extract protein, and 20,000 cpm of 32 P-labeled probe. Reactions were incubated for 10 min at room temperature; the labeled probe was added; and the incubation was continued for an additional 20-min period. For competition studies, the nuclear extract was incubated with the competing probe prior to the labeled probe during 20 min. Supershift assays were performed by preincubating nuclear extracts with 2 l of antibodies against different transcription factors (Santa Cruz Biotechnology, Santa Cruz, CA) or rabbit IgG (Sigma) for 3 h on ice before probe addition.
Western Blotting-Ten g of nuclear proteins/lane was separated by 10% SDS-polyacrylamide gel electrophoresis. Western blotting was carried out as described by Towbin et al. (27). After the transfer, nitrocellulose membranes were blocked overnight at 4°C with 5% dry milk in phosphate-buffered saline and incubated in the same way with the anti-Sp1 antibody (1:500) in phosphate-buffered saline and 5% dry milk. After incubation with the secondary antibody at room temperature for 2 h, the bands were visualized by a chromogenic reaction (Sigma Fast, nitro blue tetrazolium, Sigma).
DNase I Footprinting-The sense strand corresponding to region Ϫ111 to Ϫ27 of the ␣5 promoter was end-labeled by Klenow filling with [␣-32 P]dATP. Assays were performed with the Sure Track Footprinting kit (Pharmacia) according to the manufacturer's instructions. Recombinant Sp1 was incubated with the radiolabeled double-stranded fragment (ϳ25,000 cpm) using the binding reaction conditions described above in the EMSA experiments (except for the absence of EDTA and the presence of 2.5 mM MgCl 2 ). Immediately following the 30-min incubation at room temperature, 0.5 mM CaCl 2 and 1 mM MgCl 2 were added to the reactions. This was followed by the addition of 1 unit of DNase I. The reactions were incubated at room temperature for 1 min and stopped with the addition of stop solution (SDS/EDTA). The DNA was prepared for loading onto a 7 M urea and 8% polyacrylamide sequencing gel (ϳ15,000 cpm/lane), and the Maxam-Gilbert A/G chemical sequencing reaction was included as reference ladder.

Structure of the 5Ј-Flanking Region of the ␣5 Subunit
Gene-To examine the requirements for ␣5 subunit transcription, its promoter region was isolated and analyzed. A bovine genomic library was screened, and several overlapping clones were isolated. Clone ␣5-21 contained ϳ16 kilobases of bovine genomic sequence including exons 1 and 2 and ϳ1.6 kilobases of 5Ј-flanking region. This region was further subcloned and sequenced (Fig. 1A). Comparison of this sequence to a data base of binding sequences of known transcription factors revealed the main features of the promoter/regulatory region of the ␣5 gene: the lack of a TATA box and the presence of several GC boxes, all of them concentrated into ϳ110 bases located 5Ј to the transcription initiation site. Two perfect Sp1 consensus sites ((G/T)GGGCGGGGC) were present within this GC-rich region. Additional sites with one mismatch, regarding perfect consensus sequences, were also observed for Egr-1, Ap-1, Oct-1, and Ap-2. It is interesting to note the presence of a contiguous direct repeat of two 52-and 42-bp monomers (Fig. 1B), each of them containing one of the perfect Sp1 sites and two other elements to which this transcription factor could potentially bind.
The 5Ј-end of ␣5 mRNA was mapped by RNase protection analyses (Fig. 2). A 479-residue antisense riboprobe (Fig. 2, Probe 1) yielded two main protected fragments of ϳ250 and 249 bases. The major one mapped transcription initiation to an adenosine present in a group of three (position ϩ1, black arrow in Fig. 1) and located ϳ40 bp downstream of the GC-rich repeats. To improve precision in the determination of the transcription initiation site, a second overlapping probe was used (Fig. 2, Probe 2). In this case, two protected fragments were also observed, which were 157 and 156 bp long and mapped transcription initiation to the same place. Therefore, these were tentatively considered the main transcription initiation sites. Other initiation sites may exist, as this is a typical feature of promoters without TATA boxes, and in fact, minor protected fragments of smaller size and intensity were also observed.
Functional Analysis of the ␣5 Subunit Promoter-A series of constructs was generated to determine the regions of the ␣5 subunit promoter that contributed to its maximal activity (Fig.  3). These constructs were introduced into SHSY-5Y and chromaffin cells, a neuroblastoma cell line and a primary cell culture, respectively, that express the ␣5 subunit endogenously (7,28). In SHSY-5Y cells, the construct containing 111 bp of ␣5 promoter sequence plus 155 bp of 5Ј-noncoding region (p111␣5LUC) showed the maximal activity. This construct covered all Sp1 sites already mentioned. A shorter construct (p65␣5LUC) with only two Sp1 sites showed a 50% decrease in promoter activity, whereas further 5Ј-deletions that removed all the Sp1 sites (p39␣5LUC, pϩ7␣5LUC, and pϩ56␣5LUC) were inactive. In chromaffin cells, results were similar, suggesting that sequences in the minimal promoter, between 111 and 39 bp upstream of the start site of transcription, appear to be critical for basal transcription of the ␣5 subunit gene in transient transfection assays. Reporter constructs larger than p111␣5LUC did not show significant changes in relative luciferase activity when expressed in chromaffin cells. However, in SHSY-5Y cells, a small but constant decrease was observed from one construct to the next larger one, being maximal with p752␣5LUC, which showed a marked decrease in activity (36%). The largest construct tested (p1412␣5LUC), however, exhibited increased activity (73%). Therefore, in SHSY-5Y but not chromaffin cells, elements predominantly located between -600 and -750 with respect to the transcription initiation site have a negative effect on ␣5 promoter activity.
We next compared the ␣5 promoter with the SV40 viral promoter, present in the vector pGL2-Control, and the nAChR ␣7 and ␤4 subunit promoters (Fig. 4). These subunits are also expressed in chromaffin and SHSY-5Y cells (7,8,28), probably forming part of the same (␤4) or different (␣7) nAChR subtype in which ␣5 is present. In addition, the ␤4 subunit gene is located in the same gene cluster as the ␣5 subunit gene (11,12), as indicated previously. In chromaffin cells, the maximal nAChR subunit promoter activity corresponded to the ␣5 subunit, which accounted for about one-half of the activity shown by the SV40 promoter and was 10-and 2-fold higher that the activities of the ␤4 and ␣7 promoters, respectively. By contrast, in SHSY-5Y cells, the activities of the three subunit promoters were similar and about one-fifth the activity of the SV40 promoter. Therefore, it appears that some cell-specific differences exist among the three nAChR subunit promoters, which in any case are weaker than the SV40 viral promoter.
The ␣5 promoter constructs already tested in chromaffin and SHSY-5Y cells were also transfected into COS cells in an attempt to find elements that would confer cell specificity. Although this cell line does not express nAChRs, the pattern of promoter activity (Fig. 5) was similar to the one found in chromaffin cells. Thus, p111␣5LUC showed the maximal activity. A shorter construct (p65␣5LUC) with only two Sp1 sites showed a 44% decrease in promoter activity, whereas further 5Ј-deletions that removed all the Sp1 sites (p39␣5LUC, pϩ7␣5LUC, and pϩ56␣5LUC) were inactive. Constructs larger than p111␣5LUC did not show significant changes in luciferase activity relative to this construct. Interestingly, whereas in SHSY-5Y and chromaffin cells, the ␣5 promoter was clearly weaker than the SV40 promoter, present in the pGL2-Control vector, in COS cells, it was ϳ2-fold stronger than the viral promoter. Therefore, no cell-specific elements were found in the ␣5 promoter region analyzed.
Initially, four GC boxes present between -111 and Ϫ40, two of them representing perfect Sp1 consensus sites and another two with one mismatch, were chosen for mutagenic analysis. Single mutations of the four GC boxes indicated that all of them contribute to the activity of the ␣5 gene, but in a different way, depending on the type of transfected cell (Fig. 6). Thus, when the most promoter-proximal of the GC boxes was altered (site 1, Fig. 6A), ␣5 promoter activity in SHSY-5Y cells decreased to ϳ50% of that observed for the parent construct (p111␣5LUC), whereas single mutations of the other three had no effect (Fig. 6B). The double mutant of sites 1 and 2 produced an activity close to the obtained for site 1, whereas the simultaneous mutation of sites 3 and 4 did not have any effect, further suggesting that sites 2-4, alone or in pairs, are not required for optimal promoter activity. Surprisingly, activity was almost abolished when all four GC boxes were altered simultaneously, which may indicate that sites 2-4, together with site 1, do integrate a whole synergistic mechanism required for basal promoter activity. By contrast, in chromaffin cells, mutation of any of the GC boxes produced a similar decrease in promoter activity (50 -69%), and double mutants showed a further decline. Finally, as happened with SHSY-5Y cells, the simultaneous mutation of the four GC motifs produced the maximal decrease in activity.
Characterization of the Regulatory Elements Present at -111 to Ϫ27 of the ␣5 Promoter by EMSA-DNA fragments carrying wild-type promoter region Ϫ111 to Ϫ27 were labeled and incubated with nuclear extracts from chromaffin and SHSY-5Y cells (Fig. 7). Several retarded bands were observed in both cases (Fig. 7 A, lanes 2 and 4). Some of them were common to both extracts (q and E). Recombinant Sp1 produced a main retarded complex (lane 6), coincident in position with one of those observed with nuclear extracts (q). Larger complexes were also observed with recombinant Sp1, probably as the result of the simultaneous binding of two or more Sp1 mole- cules to the same DNA fragment. Antibody supershift analysis was employed in an attempt to identify the proteins producing the retarded bands. One of the main complexes (q) was retarded by an anti -Sp1 antibody (arrowhead, lanes 3, 5, and 7), whereas no supershift was observed with antibodies against several transcription factors able (Sp3, Egr-1, and Ap-2) or not (Myc, Max, USF1, and USF2) to bind to GC boxes (data not shown). In addition to the main band, several minor ones were also displaced by the anti-Sp1 antibody when using chromaffin cell nuclear extracts (compare lanes 2 and 3). They are probably degradation products of Sp1 protein, which keep the capacity of binding to the DNA probe and being recognized by the antibody. The band labeled with an open circle (E) was more prominent in SHSY-5Y extracts (lanes 4 and 5), but we were not able to identify it. Competition experiments confirmed the specificity and identity of the complex formed by Sp1 (Fig. 7B). Thus, an excess of unlabeled probe neutralized the formation of all complexes (lanes 10 and 13), whereas a synthetic oligonucleotide containing an Sp1 consensus sequence abolished the formation of the Sp1 complex (lanes 11 and 14).
Western blot analysis of SHSY-5Y and chromaffin nuclear proteins indicated that Sp1 protein was indeed expressed in these cells (Fig. 7C). Two proteins bands (ϳ95-100 kDa) were detected with anti-Sp1 antibodies in chromaffin (lane 15) and SHSY-5Y (lane 16) nuclear extracts, showing the same size as that previously described for Sp1 polypeptides (29). The amount of Sp1 protein detected in both extracts was approximately the same, and both bands had the same intensity. By contrast, recombinant Sp1 (lane 17) showed predominantly one species.
Interactions of Sp1 with the GC Boxes in the ␣5 Proximal Promoter-Recombinant Sp1 protein was used in DNase I foot- printing (Fig. 8) to document the preferences of this transcription factor for the four GC-rich elements previously characterized in transfection studies (Fig. 6). When using the same probe of the EMSA, Sp1 protected several domains (Fig. 8, lanes 3  and 4): the largest one corresponded to GC boxes 1 and 2 (large filled box on the left), whereas two additional domains were also observed, especially at the highest Sp1 concentration (lane 4). One of them corresponded to GC box 4, and the other overlapped the 3Ј-half of GC box 3 and a downstream GC-rich element. This element was not as close to the Sp1 consensus binding sequence as the ones previously analyzed and, for this reason, was not included in our previous functional characterization (Fig. 6), but, given the DNase protection results, it was further characterized.
To confirm that an additional Sp1 site exists in the proximal region of the ␣5 subunit promoter, EMSA experiments were performed with recombinant Sp1, and the DNA fragment carrying wild-type promoter region Ϫ111 to Ϫ27 and the corresponding fragment mutated at the four original Sp1 sites were compared (Fig. 9). At low Sp1 concentration, a retarded band was observed with the wild-type probe (lane 3). The same band was observed in the case of the mutant probe, but it was very faint (lane 7). A higher Sp1 concentration increased the intensity of the retarded band with the wild-type probe and even induced the formation of complexes of larger size (lane 4), probably as the result of Sp1 binding to more than one site within a single DNA molecule. In the case of the quadruple mutant, the faint band was more intense (lane 8), but no larger complexes were observed, suggesting that this DNA fragment still has one intact Sp1-binding site. That this Sp1 site is the one immediately downstream of site 3 was demonstrated by mutating it (Fig. 10A) and performing EMSA experiments (Fig.  10, B and C). The combination of the same wild-type probe, already used, with nuclear extracts from chromaffin cells yielded the typical pattern in which a prominent band (Fig.  10B, lane 2, q) was observed. The formation of this band, previously demonstrated to be antigenically related to Sp1 (Fig.   7B), was competed by the unlabeled probe (Fig. 10B, lanes 3  and 4), but not by the same probe mutated at the five Sp1 sites (lanes 5 and 6). Moreover, when the mutant fragment was used as the labeled probe, the prominent Sp1 band was not observed (lane 8). Finally, recombinant Sp1 was unable to produce retarded bands with the quintuple mutant (Fig. 10C, lanes 13  and 14), contrary to what occurred with the quadruple mutant (Fig. 9).
The functional significance of the fifth GC box was examined in chromaffin cells by transfecting constructs, made in the context of p111␣5LUC (see Fig. 6), in which this element was mutated individually or in combination with the other four. In principle, this element seems to be unable to activate transcription by itself alone since the mutation of the other four, leaving this site intact, decreased activity to very low levels (Fig. 6). Thus, the contribution of this element to transcriptional activation, if any, should be expected in a cooperative manner, as it happens with the other four previously characterized. In fact, the single mutation had an effect on promoter activity similar to the one observed for the other single mutants (50% of the activity of p111␣5LUC). The quintuple mutant had 14% of promoter activity (relative to p111␣5LUC), close to the value obtained for the quadruple mutant (14.7%). Therefore, it seems that this element mediates the transcriptional activation of the ␣5 promoter in a collective way, combined with the other GC boxes.

DISCUSSION
Neuronal nAChRs mediate chemical synaptic transmission, probably regulating transmitter release at many synapses (reviewed in Ref. 5). A relatively large number of genes that encode nAChR subunits have been identified (1), having distinct, although overlapping patterns of expression in the central and peripheral nervous systems. This diversity constitutes, in large part, the molecular basis on which the variety of nAChR properties and neural responses to acetylcholine is established. The nAChR ␣5 subunit is widely expressed in the FIG. 6. Sites 1-4 are functional elements required for ␣5 subunit gene expression. A, the proximal region of the ␣5 subunit promoter (nucleotides -111 to -40) is depicted with the putative regulatory elements underlined. This region contains putative binding sites for transcription factor Sp1 (boxed). Several nucleotides of each potential element were mutated as indicated below the sequence to yield constructs analyzed in transfection experiments (B). B, the name of each mutant construct indicates the element(s) that have been altered (also crossed out in the scheme). Plasmids were transfected into SHSY-5Y and chromaffin cells, and activities were measured. Luciferase activity was normalized to values obtained with the p111␣5LUC construct. Data are expressed as described in the legend to Fig. 3. peripheral nervous system (7,30,31) in combination with ␣3, ␤4, and, in some cases, ␤2 subunits (9,32). Moreover, it has been demonstrated recently that the ␣5 subunit combines with ␣4, ␤2, ␣3, and ␤4 subunits (33,34) in heterologous expression systems, modifying nAChR channel properties. Thus, some information has been obtained regarding ␣5 subunit function, but nothing was known until now about the cis-elements and trans-acting factors that regulate its expression.
The core promoter region of the ␣5 subunit does not contain TATA and CAAT boxes, but it does have several GC-rich domains, a feature found in the promoters of the ␣2 (35), ␣3 (13), ␤4 (20), and ␣7 (10, 36) subunits. Moreover, the structure of the ␣5 subunit promoter appears highly organized with two direct repeats (Fig. 1B), each containing crucial elements for promoter activity. From 5Ј-end deletion analysis (Fig. 3), we determined that the region located between nucleotides -111 and ϩ155 was necessary for the basal promoter activity detected in SHSY-5Y and chromaffin cells. A large loss in promoter activity was observed when 72 bp were deleted from the 5Ј-end (compare p39␣5LUC with respect to the larger construct p111␣7LUC) (Fig. 3). These constructs were also transfected into COS cells (Fig. 5), which do not endogenously produce ␣5 subunits, and were even more potent in luciferase activity  6 and 7). Lane 1 is probe run in the absence of protein extracts. Lanes 3, 5, and 7 represent complexes obtained upon adding anti-Sp1 antibodies (Ab). A prominent band (q) was supershifted by Sp1 antibodies (arrowhead). Other complexes were also observed (E and f), but were not displaced by any of the antibodies tested (see "Results"). B, the gel mobility assay was run using DNA fragment Ϫ111 to Ϫ27 as the labeled probe and nuclear extracts from chromaffin (lanes 9 -11) and SHSY-5Y (lanes 12-14) cells. Lanes 9 and 12 were without added competitor (Comp.; Ϫ). Competitor DNA fragment Ϫ111 to Ϫ27 (PR; lanes 10 and 13) and an oligonucleotide with a consensus site for Sp1 (Sp1; lanes 11 and 14) were added in 100-fold excess. Lane 8 is probe run in the absence of protein extracts. C, nuclear proteins from chromaffin (lane 15) and SHSY-5Y (lane 16) cells and recombinant Sp1 (lane 17) were separated by SDS-polyacrylamide gel electrophoresis on a 10% resolving gel. Following Western blotting with Sp1 antibodies, two molecular species of ϳ97 kDa were detected. N.P., nuclear protein. when using a viral promoter as a reference. Therefore, it is possible that promoter elements needed for cell-specific expression are not included within the promoter fragments used in this study. Nevertheless, the action of negative elements seems evident in SHSY-5Y cells, as demonstrated by the slight but gradual decrease in the activity of larger constructs. Elements located between -660 and -752 appear responsible for the largest decline in activity and could be involved in a silencing mechanism. It is evident that this mechanism might not be totally effective in SHSY-5Y cells since the largest construct (p1409␣5LUC) regains activity, and in fact, SHSY-5Y cells endogenously express ␣5 subunits. However, it could be of relevance in tissues where ␣5 subunit expression is repressed, although, at least in COS cells, such a mechanism does not seem to be operative.
The most remarkable feature in the deleted region, between -111 and -39, was the presence of at least four closely located Sp1 sites (labeled 1-4 in Fig. 6A). Therefore, these elements appeared to be suitable candidates for controlling promoter activity. Consequently, when these elements were simultaneously mutated in the context of p111␣7LUC, promoter activity was almost suppressed, thus suggesting that they play a crucial role in the transcriptional regulation of the ␣5 gene. Moreover, DNase I footprinting with purified Sp1 protein revealed regions of extended protection, which contained the four Sp1-binding sites and also an additional GC-rich element, close to site 3 (Fig. 8). This site, in contrast to the others, which have no or one mismatch with respect to a perfect Sp1 consensus binding sequence, has two mismatches and, for this reason, was not considered in the initial analysis of promoter activity. However, EMSA (Fig. 10) and gene reporter experiments confirmed the footprinting data indicating that Sp1 can also bind to this site with a functional significance. We postulate that Sp1 contacts the DNA through multiple interactions to form a higher order complex over the promoter region. Interestingly, the five sites are approximately located on the same side of the DNA, as each turn of B-DNA contains 10.5 bp (37), and sites 2, 5, 3, and 4 are 11, 31, 42, and 62 bp apart, respectively, from site 1. This special arrangement could generate an activation environment that would produce more effective interactions with the transcription apparatus, giving rise to the formation of a multiprotein complex that would accomplish transcription activation at a higher level (38). The individual contribution of each Sp1 element to the formation of this hypothetical higher order complex appears differently defined, depending on the cell context. Thus, in SHSY-5Y cells, the only critical site is the most promoter-proximal, perhaps because the possibility of FIG. 9. Binding of recombinant Sp1 to GC motifs within the promoter region of the ␣5 gene. Labeled wild-type DNA fragment Ϫ111 to Ϫ27 and the corresponding fragment mutated at the four Sp1 sites previously analyzed (Quadruple Mutant) were used as gel mobility shift probes in the presence of 0.02 (lanes 2 and 6), 0.1 (lanes 3 and 7), and 0.5 (lanes 4 and 8) footprint units of recombinant Sp1. Lanes 1 and 5 contain probe run in the absence of protein (Ϫ). q, Sp1 bound to a single site; qq, Sp1 bound to two sites.

FIG. 10. Gel mobility shift assays of the fifth Sp1-binding site.
A, region -80 and -71 contains an additional putative binding site for Sp1 (underlined and labeled as 5 in the upper sequence). Several nucleotides of this potential element were mutated, as indicated in the lower sequence, in combination with the previously mutated sites and used as probe (Quintuple Mutant). B, labeled wild-type DNA fragment Ϫ111 to Ϫ27 and the corresponding quintuple mutant were used as gel mobility shift probes in the presence of 2 g of crude chromaffin cell nuclear extracts. Competitor (Comp.) DNA fragments corresponding to the wild type (WT; lanes 3 and 4) and the quintuple mutant (QM; lanes 5 and 6) were added in 25-and 100-fold excesses. Lanes 1 and 7 were run in the absence of protein (Ϫ). q corresponds to the band that was supershifted by Sp1 antibodies (see Fig. 7), whereas the band indicated by f has not been identified. C, the same probes from B were used in the presence of 0.05 (lanes 10 and 13) and 0.2 (lanes 11 and 14) footprint units of recombinant Sp1. Lanes 9 and 12 contained probe run in the absence of protein (Ϫ). physically direct interactions between Sp1 at this site and the basal transcriptional machinery. The importance of the others only emerges if they are simultaneously altered, suggesting that Sp1 acts synergistically as a whole on the complete region. By contrast, in chromaffin cells, transcriptional activity seems to depend on each of these sites, as alteration of any of them produces a decrease in promoter function. If we assume that the formation of a higher level complex requires several Sp1 sites, then the possibility exists that in SHSY-5Y but not chromaffin cells, other elements can substitute for the Sp1 sites that were gradually altered by mutagenesis. This substitution would be not possible, however, when all of the Sp1 sites are modified. This assumption is supported by the fact that the double mutant of sites 3 and 4 shows only an insignificant decrease (11%) with respect to the activity of the parent construct p111␣5LUC (Fig. 6B), whereas the deletion construct p65␣5LUC, which lacks the region between -111 and -66, including sites 3-5, but also other sequences, shows a marked decrease (48%) in activity. Perhaps some additional element present in the deleted region becomes effective in SHSY-5Y cells when some but not all of the Sp1 sites are altered. Whatever the mechanism, the primary conclusion of this study is that the basal activity of the ␣5 subunit promoter is defined by an array of Sp1-binding elements to which this transcription factor can bind to activate transcription in a synergistic manner.
In addition to silencing mechanisms, the ␣5 subunit promoter, which appears strongly dependent on a ubiquitous transcription factor like Sp1, may be regulated in a tissue-specific manner by this factor in several ways. (a) Availability of Sp1, as levels of this protein have been shown to vary among different tissues (39). (b) Competition with other members of the Sp family that could bind to the same elements. For instance, Sp3 acts with Sp1 in a concerted way to transactivate the nAChR ␤4 subunit (21). However, we have not found binding of Sp3 to the ␣5 promoter in EMSA of chromaffin and SHSY-5Y nuclear extracts. In other cases, Sp3 has been shown to repress the action of Sp1 (40 -42). (c) Formation of heteromeric complexes between Sp1 and other proteins. Protein p107, a member of the retinoblastoma family of proteins, binds Sp1 and represses Sp1-dependent transcription (43). Sp1 also interacts with the RelA subunit of transcription factor NF-B (44) and the cellular protein YY1 (45). (d) Action of Sp1 as a physical link between proximal and distal promoter elements via DNA bending (46) or looping (47).
Finally, we would like to emphasize the role that Sp1 could play in a potential common regulatory mechanism of the ␣3, ␤4, and ␣5 subunits. As has been already mentioned, these subunits compose the predominant nAChR subtype in the peripheral nervous system (7,9,32) and have their genes clustered in the genome (11,12). Therefore, a concerted regulation of the expression of the three subunits is plausible. Given that Sp1 (and other members of the Sp family) seems to play a crucial role in the regulation of the ␣3 (22), ␤4 (20,21), and ␣5 (this study) subunits, we propose that this transcription factor, either directly or through other proteins that can regulate its activity, may play the coupling role in the coordinated regulation of the three subunits. Other factors, or perhaps a differential regulation of Sp1, specific to a determined neuronal cell subset, would be responsible for the independent regulation of each subunit, which also must take place, given the dissimilar pattern of expression of these subunits in certain places (48,49).