By continuing to browse this site you agree to us using cookies as described in About Cookies
Wiley Online Library is migrating to a new platform powered by Atypon, the leading provider of scholarly publishing platforms. The new Wiley Online Library will be migrated over the weekend of March 17 & 18. You should not experience any issues or loss of access during this time. For more information, please visit our migration page: http://www.wileyactual.com/WOLMigration/
The present address of Dr. P. J. McLean is Alzheimer's Disease Research Unit, Department of Neurology, Massachusetts General Hospital—East, 149 13th Street, Charlestown, MA 02129, U.S.A.
Abbreviations used: GABA, γ-aminobutyric acid; GABAA, type A γ-aminobutyric acid; Inr element, initiator element.
Address correspondence and reprint requests to Dr. D. H. Farb at Laboratory of Molecular Neurobiology, Department of Pharmacology, Boston University School of Medicine, 80 East Concord Street, Boston, MA 02118, U.S.A. E-mail: firstname.lastname@example.org
Abstract: The ability of nerve cells to regulate the expression of specific neurotransmitter receptors is of central importance to nervous system function, but little is known about the DNA elements that mediate neuron specific gene expression. The type A γ-aminobutyric acid (GABAA) receptor α6-subunit gene, which is expressed exclusively in cerebellar granule cells, presents a unique opportunity to study the cis elements involved in restricting gene expression to a distinct neuronal population. In an effort to identify the regulatory elements that govern cerebellar granule cell-specific gene expression, the proximal 5′ flanking regions for the human, rat, and mouse α6 genes were cloned and sequenced, and a major transcriptional initiation site was identified in the rodent genes. Functional analysis of rat α6 gene-reporter constructs in primary neuronal cultures reveals that a 155-bp TATA-less promoter region (-130 to +25 bp) constitutes a minimal promoter that can drive cerebellar granule cell-specific expression. Internal deletion and decoy competition studies demonstrate that the minimal promoter contains a 60-bp region (-130 to -70 bp) that is critical for enhanced promoter activity in cerebellar granule cells. Activity of the compromised promoter containing the deletion cannot be rescued by placing the 60-bp region downstream of the reporter gene, demonstrating that it is not a classical enhancer but rather a positionally dependent regulator. An additional cerebellar-specific activating sequence is located between -324 and -130 bp, and a downstream negative regulatory region (+158 to +294) has been shown to be active in fibroblasts but inactive in cerebellar granule cells. Taken together, the results suggest a possible mechanism for the control of cerebellar granule cell-specific expression of the GABAA receptor α6 subunit gene.
The mechanism(s) by which neuronal-specific gene expression is controlled represents a central problem in modern molecular neurobiology. The α6 subunit of the γ-aminobutyric acid type A (GABAA) receptor provides an excellent model for studying the events leading to neuron-specific gene expression as this subunit is expressed exclusively in the embryonically related granule cells of the cerebellum and cochlear nucleus (Lüddens et al., 1990; Varecka et al., 1994). The mechanisms involved in restricting expression of the α6 gene to cerebellar granule cells may have general implications for understanding the regulation of gene expression in the nervous system.
Previous studies have demonstrated that a region of the 5′ flanking sequence of the α6 subunit gene can confer tissue-specific expression in certain transgenic animals (Jones et al., 1996; Bahn et al., 1997). However, little is known about the function of the α6 promoter responsible for cerebellar granule cell-specific expression. In an effort to identify the minimal α6 promoter and important regulatory regions, we have compared the sequence of the human, rat, and mouse α6 subunit genes. Homology in consensus regulatory sequences suggests that they may share a similar mechanism of cell-specific expression. Reporter constructs containing regions of the rat α6 5′ flanking sequence upstream of the luciferase gene were tested for promoter activity by transient transfection into rat neuronal and nonneuronal primary cultures (McLean et al., 1997). Here, we report the identification of a minimal neuronal-specific promoter for the α6 subunit gene and regions contained within it that play a role in cerebellar granule cell-specific expression. This study represents the first analysis of the promoter region for the α6 gene and provides the basis for an investigation into the mechanisms of both neuronal- and cerebellar-specific gene expression.
Isolation and cloning of the 5′ end of the rat GABAA receptor α6 subunit gene using PCR
PCR was performed to amplify the 5′ end of the rat α6 gene from rat genomic DNA (Clontech) using the oligonucleotide primers rα6SP#2 and rα6AS#3 (Table 1), derived from the rat α6 cDNA sequence (Lüddens et al., 1990). PCR was performed in a 100-μl reaction using 500 ng of rat genomic DNA, 325 ng of each primer, 0.2 mM each deoxynucleotide triphosphate, 1.5 mM MgCl2, and 2.5 units of Taq DNA polymerase. The resulting PCR product was cloned into pBluescriptKSII+ (Stratagene). DNA sequencing confirmed that the product was specific to the published α6 cDNA sequence, except for a 198-bp insert found 65 bp from the 5′ end of the sequenced product that we designated intron-1. PCR was then performed with the oligonucleotide primers rα6SP#3 and rα6AS#6 (Table 1), which are directed to cDNA sequence downstream of intron-1. The presence of ∼800 bp of intervening sequence was detected within the PCR product and was designated intron-2.
Table 1. DNA sequence of primers used in PCR, inverse PCR, and primer extension analyses
Isolation of sequences upstream from the 5′ end of rat α6 cDNA using inverse PCR
Inverse PCR was performed as described (Triglia et al., 1988) with the PCR primer pairs rα6IVSP#1 and rα6AS#7 and rα6IVSP#3 and rα6IVAS#2. In total, 845 bp of sequence upstream of the 5′ end of the published rat α6 cDNA was isolated from two consecutive inverse PCR procedures.
Isolation of the 5′ end of the mouse α6 subunit gene using PCR
Oligonucleotide primers, mα6SP#1 and rα6AS#7, directed toward the 5′ end of the mouse α6 cDNA (Kato, 1990), were used to isolate a 250-bp region of the mouse gene from mouse genomic DNA (Clontech). Following verification of the PCR product as the mouse α6 subunit gene, the same primers were used to screen a mouse bacteriophage library (P1; Genome Systems). Subclones were sequenced with the oligonucleotide primers mα6AS#1-mα6AS#8 (Table 1). Results from sequencing indicated that the subclone contained ∼3.5 kb of mouse α6 gene sequence.
Isolation of the 5′ end of the human α6 subunit gene using PCR
Oligonucleotide primers, hα6SP#1 and hα6AS#1, directed toward the 5′ end of the human α6 cDNA (Hadingham et al., 1996), were used to isolate an ∼1-kb region of the human gene from human genomic DNA. The PCR product was cloned, and sequencing verified that the clone contained sequences specific to the first three exons and the first two introns of the human α6 subunit gene. The 5′ flanking sequence of the human α6 subunit gene was isolated from a P1-derived artificial chromosome library (PAC) by PCR screening using the primers hα6SP#2 and hα6AS#2 (Genome Systems).
Mapping the transcriptional start sites for the rat α6 subunit gene
Ribonuclease protection assay. A ribonuclease protection assay was performed as described (Ausubel et al., 1990). In brief, pBluescriptKSII+ plasmids containing the following regions of the 5′ end of the rat α6 subunit gene were used as templates for the generation of riboprobes using either T3 or T7 RNA polymerase (Pharmacia): nucleotides -354 to +70 (RP424) and -354 to -66 (RP288), where numbers express position relative to the 5′ end of published cDNA sequence. Each probe was hybridized to 10 μg of total rat brain RNA isolated from cerebellum, cortex, or hippocampus of adult rat or primary cultures of rat cerebellar granule cells. The reactions were then processed according to the established protocol. A DNA sequencing ladder was generated from the single strand of M13 phagemid by an M13 sequencing primer to estimate the size of the protected products, assuming a 5% difference between DNA and RNA.
Primer extension analysis. Antisense oligonucleotide primers rα6PE37 and rα6S1AS#1 were designed for hybridization to sequences located close to the 5′ end of the published rat α6 cDNA sequence. Primer extension analysis was performed using 20 μg of total rat brain RNA from cerebellum and cortex as described previously (Sambrook et al., 1989). A Sanger sequencing reaction was performed with the same primer to estimate the size of the extension products.
Mapping transcriptional start sites for the mouse α6 subunit gene
Antisense oligonucleotide primers that correspond to a region located downstream of the previously mapped 5′ end of the rat α6 gene (mα6AS#9 and rα6AS#7) were used in primer extension analysis with 20 μg of total mouse brain RNA from cerebellum and cortex as described previously.
Construction of chimeric rat α6-luciferase reporter constructs
PCR primers directed toward the 5′ flanking region of the rat α6 gene were used to amplify specific regions of the sequence for cloning into pGL2-Basic reporter vector (Promega). PCR procedures were performed as described previously. The combination of primers that were used to generate the PCR products are as follows: rα6SP#7/rα6AS#7 (-595/294); rα6SP#9/rα6AS#7 (-130/294); rα6SP#7/rα6S1AS#1 (-595/158); rα6IVSP#5/rα6S1AS#1 (-324/158); rα6SP#9/rα6S1AS#1 (-130/158); and rα6SP#7/rα6AS#12 (-595/-264). The PCR products were subcloned into the SmaI restriction site of the pGL2-Basic vector using the method of restriction selection cloning (Russek et al., 1993). PCR products were completely sequenced to verify the integrity of DNA sequence. The construct -595/294 Δ60:+60 was generated by cloning an oligonucleotide, 5′-TCAGCAAGATG CCACATTTCCAGATTTC-CTCACAGGTCCTTCTAATTCCATGCCAAAA-3′, down-stream of the luciferase gene. The 60-bp oligonucleotide was also cloned into the pBluescript KSII+ vector to generate KSII60.
Generation of internal deletions
The internal deletion construct -595/294Δ60 was constructed using a modification of the Kunkel oligonucleotide-directed mutagenesis method (Kunkel et al., 1987). Using this method, a 60-bp region from -130 to -70 was deleted from -595/294 to produce -595/294Δ60. A HindIII restriction enzyme site was engineered into -595/294 and -130/294 at position +27 to generate the constructs -595/25 and -130/25 following digestion with HindIII enzyme.
Rat cerebellar granule cell cultures. Primary cultures of cerebellar neurons were prepared from cerebella of 8-day-old rat pups as described previously (Novelli et al., 1988). Groups of 10 cerebella from 8-day-old Sprague-Dawley rats (Charles River Laboratories) were dissected and used for the preparation of cerebellar granule cell cultures. Dissociated cells were plated on individual plastic 60- or 35-mm-diameter dishes or six-well 35-mm plates (Nunc) coated with poly-L-lysine (0.1 mg/ml). Cultures were incubated at 37°C in 5% CO2. Cytosine arabinoside (final concentration, 10 μM) was added 24 h after plating to prevent replication of nonneuronal cells. Cultures were returned to the incubator for a total of 24 h before medium was changed.
Rat neocortical and hippocampal cultures. Primary cultures of neocortical and hippocampal neurons were prepared from 18-day-old rat embryos essentially as described previously (Brewer and Cotman, 1989) with modifications. Brains were removed from 12 to 18 embryos (stage E18), and neocortex and hippocampi were dissected, triturated, and then centrifuged at 200 g for 5 min. The pellet was resuspended in plating medium [Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum, 2.4 mg/ml bovine serum albumin, 26.5 mM sodium bicarbonate, 1 mM sodium pyruvate, 20 mM HEPES, 100 U/ml penicillin, 100 μg/ml streptomycin, and a modification of Brewer's B16 defined components (with 250 nM vitamin B12 and without catalase, glutathione, and superoxide dismutase)] and triturated before being added to a final volume of defined medium. Cultures were incubated at 37°C in 5% CO2. Cytosine arabinoside (final concentration, 1 μM) was added 24 (neocortex) or 48 h (hippocampus) after plating to prevent replication of nonneuronal cells. Cultures were returned to the incubator for a total of 48 h before the medium was changed to serum-free defined medium.
Rat fibroblast cultures. Primary cultures of rat fibroblasts were prepared from the hind limbs of E18 rat embryos. The hind limbs were removed, and the tissue was minced, centrifuged at 200 g for 5 min, and resuspended in trypsin-EDTA solution. Following incubation at 37°C for 7 min, plating medium was added, and the suspension was centrifuged at 200 g for 5 min. The pellet was resuspended in plating medium and triturated through a 5-ml pipette. Dissociated cells were plated in 100-mm-diameter tissue culture dishes and incubated at 37°C in 5% CO2.
Transfections into rat cerebellar granule cells, neocortical neurons, and rat fibroblasts were performed using a modification of the calcium phosphate coprecipitation method (Xia et al., 1996). Eight micrograms of plasmid DNA was transfected into each 60-mm-diameter dish, and the transfection was allowed to proceed for 2 h at room temperature for cerebellar granule cells or 30 min at 37°C in 5% CO2 for neocortical neurons. The cultures were incubated for 24 h posttransfection before being washed once with 1 × Hanks' balanced salt solution, lysed, and analyzed for luciferase reporter gene activity or fixed with 4% paraformaldehyde and stained for β-galactosidase activity (Ausubel et al., 1990). In addition, primary cultures of rat neocortical, neurons, hippocampal neurons, and fibroblasts were transfected by calcium phosphate coprecipitation (Chen and Okayama, 1987) using 35 μg of DNA per 100-mm-diameter dish, 14 μg of DNA per 60-mm-diameter dish, and 7 μg of DNA per 35-mm-diameter dish. Cells were incubated at 37°C in 5% CO2 for 18-20 h. Cells were then washed twice with 1 × Hanks' balanced salt solution, lysed, and analyzed for reporter gene activity. To control for difference in transfectional efficiency, the promoter activity of different α6 constructs was compared with that of the SV40 promoter, with its downstream enhancer, in sister cultures and with background activity as measured by the pGL2 promoter-less vector. Cotransfection was not used because it decreased α6 promoter activity, possibly reflecting competition for transcription factors.
Reporter gene assay
Luciferase reporter gene activity was measured according to the instructions of the manufacturer (Promega) and quantified in a Beckman model LS6000 scintillation counter with an attached single photon monitor. Protein content in the cellular extracts was determined by the method of Lowry et al. (1951). All reporter gene activity was normalized against protein concentration in each extract.
Using PCR with oligonucleotide primers derived from the published cDNA sequence (Lüddens et al., 1990), we have isolated the 5′ end of the rat GABAA receptor α6 subunit gene from rat genomic DNA. Analysis of the sequence and a comparison with the published cDNA sequence reveal the presence of two introns. In addition, we have isolated the sequences flanking the first two introns of the mouse and human α6 genes and determined that the position of the introns is conserved among these species (Fig. 1A). The organization of the mouse gene is consistent with the structure previously reported by Jones et al. (1996). The first intron in all genes is relatively small: 198 bp in the rat, 225 bp in the mouse, and 196 bp in the human. The conservation of nucleotide sequence in the first intron among the rat, human, and mouse is ∼55%, whereas the sequences constituting the second intron are more divergent (data not shown). Only the extreme 5′ and 3′ ends of the second intron have a high degree of sequence similarity, with the intervening sequences displaying little identity. The greatest conservation of sequence is observed in the protein-coding exons, where there is >90% identity among the rodent and human genes.
Inverse PCR was used to isolate sequence upstream of the published rat α6 cDNA sequence. Successive genomic walking with this technique yielded 845 bp of sequence upstream of the 5′ end of the published cDNA sequence. In addition, screening of a mouse P1 library resulted in the isolation of ∼3.5 kb of mouse α6 gene sequence, of which ∼2.6 kb is upstream of the 5′ end of the published mouse cDNA sequence (Kato, 1990). Sequencing of the 5′ end of the human α6 subunit gene contained within a PAC clone that was isolated by PCR screening reveals significant areas of sequence identity among the rat, mouse, and human genes in the regions upstream of the 5′ ends of the published cDNA sequences (Fig. 1B).
Transcriptional start site determination
Primer extension analysis and RNase protection were used to determine the site(s) of transcriptional initiation for the rat and mouse α6 subunit genes. Riboprobes RP288 and RP424 (see Experimental Procedures) were designed to detect any transcriptional initiation sites that may be located within 414 bp of the ATG initiation codon. Hybridization of RP424 and RP288 with total RNA from rat cerebellum reveals several protected species (Fig. 2A, lanes 3 and 7, respectively). Major protected bands correspond to 293 (lane 3) and 158 bp (lane 7) in size, which map to a position that is ∼227 nucleotides upstream of the 5′ end of the published rat α6 cDNA sequence. Additional protected species of intermediate intensity are detected at 315 (lane 3) and 183 bp (lane 7), and several minor bands are apparent. RNase protection of total RNA extracted from primary cultures of rat cerebellar granule cells also produces protected species of 293 and 158 bp (Fig. 2A, lanes 5 and 9). Consistent with the fact that α6 mRNA is solely expressed in cerebellar granule cells, total RNA from rat cortex (lanes 4 and 8) failed to produce protected bands. In agreement with the results of RNase protection assays, primer extension analysis reveals the presence of one major transcriptional initiation site located ∼227 bp upstream of the 5′ end of the cDNA sequence, as well as several minor start sites (data not shown).
Primer extension analysis using mouse cerebellar RNA produces an extension product of 173 bp, indicating that the transcriptional initiation site is 44 bp upstream of the 5′ end of the mouse α6 cDNA sequence (Fig. 2B, lane 1). In addition, the extension product is not observed in a reaction with total RNA isolated from mouse cortex (lane 2). Primer extension was also performed with a primer whose sequence is conserved between the rat and the mouse α6 genes (rα6AS#7). One extension product at 295 bp was detected in each reaction using total RNA isolated from mouse cerebellum and rat cerebellum (data not shown). These products are not observed in reactions with total RNA extracted from mouse cortex or rat cortex. Taken together, the extension products locate the major transcriptional initiation sites of the rat and mouse α6 genes to corresponding positions. Also, several putative transcriptional initiation sites were identified using either RNase protection or primer extension but were not consistently identified with both techniques. These potential initiation sites cannot be excluded at this point; however, for reasons of clarity, only those transcriptional initiation sites identified using both RNase protection and primer extension are indicated (Figs. 1B and 2C).
Identification of an initiator element
The approximate consensus Py2A+1NT/APy2 provides a starting point for predicting whether a given sequence will impart initiator element (Inr element) activity (Javahery et al., 1994). Analysis of the DNA sequences surrounding the major transcriptional initiation site in the rat and mouse α6 genes indicates the presence of a consensus sequence for an Inr element (Smale and Baltimore, 1989) (Fig. 2C), where the A nucleotide is designated +1.
The major goal of this study was to determine if sequences responsible for directing cerebellar granule cell-specific expression of the rat α6 gene are contained within the proximal 5′ flanking region. To this end, we established transfections into primary cultures of rat cerebellar granule cells using a modification of the calcium phosphate coprecipitation technique (Xia et al., 1996). Figure 3 demonstrates that rat cerebellar granule cells can be successfully transfected with a reporter construct expressing the gene for β-galactosidase. For the α6 gene, we used a functional assay to assess promoter activity where the expression of a luciferase reporter gene, driven by various regions of the α6 gene, is monitored. In primary cultures of rat cerebellum, which are ∼90% granule cells, striking cerebellar granule cell-specific expression of the reporter construct -595/294 is observed when compared with the activity obtained in primary cultures of rat neocortical and hippocampal neurons (Fig. 4). In rat cerebellar granule cells, the activity of -595/294 is 25-fold greater than background, as measured by the promoterless construct, pGL2-Basic. By contrast, in primary cultures of rat neocortical neurons, hippocampal neurons, and fibroblasts the activity of -595/294 is 4-, 5-, and 1.6-fold above pGL2-Basic, respectively (Fig. 4). This result indicates that an 889-bp region of the α6 gene contains sequences that are responsible for tissue- and cell-specific expression. We refer to this sequence as the full-length promoter region because it is the largest fragment under study.
Removal of 269 bp from the 3′ end of -595/294 (yielding -595/25) has no effect on α6 promoter activity in rat cerebellar granule cells (Fig. 5A and B). Additional truncation of 465 bp from the 5′ end of -595/25, which results in a 155-bp region (-130 to 25) of the α6 gene, also failed to affect promoter activity (Fig. 5A and B). The observation that activity remains specific for cerebellar granule cells (Fig. 5C) demonstrates that the sequence requirements for tissue and cell specificity are located between nucleotides -130 and +25, thus defining a minimal promoter for the α6 gene.
To define further the regulatory sequences located within this region, we produced a luciferase reporter construct that contained a 60-bp internal deletion, from -130 to -70, in the context of the full-length promoter (-595/294Δ60). As can be seen in Fig. 1B, there is a significant sequence identity between the rodent and human genes within this 60-bp region, and several consensus sequences for gene regulatory elements, including an Oct-1 motif (Staudt et al., 1986), NF-1 (Jones et al., 1987), AP-4 (Hu et al., 1990), and STAT factors (Darnell et al., 1994), have been identified. An internal deletion of the 60-bp region decreases promoter activity by 52% in cerebellar granule cells, increases promoter activity by 65% in neocortical neurons, but is without effect in fibroblasts (Fig. 5B and D), indicating a role for these sequences in tissue and cell type specificity. To determine whether the 60-bp region functions as a classical enhancer, we cloned this region downstream of the luciferase gene in 595/294Δ60 to generate 595/294Δ60: +60. As can be seen in Fig. 6A, addition of the 60-bp region to the construct does not restore the activity of the promoter, demonstrating that the positive regulatory region displays position-specific activity. The importance of this positive regulatory region to cerebellar specific expression was further investigated by performing decoy competition studies using cotransfection of -595/294 with a cloning vector that contains the 60-bp KSII60. Competition dramatically inhibits α6 promoter activity in cerebellar granule cells (Fig. 6B).
Truncation analysis has also identified several regions of the α6 gene that contain putative regulatory elements. Removal of 136 bp (between 158 and 294) from the 3′ end of both the full-length promoter construct (-595/294) and a 5′ truncated construct (-130/294) reveals the presence of downstream sequences that may enhance α6 promoter activity (Fig. 7A), although the observed decrease is not statistically significant (p = 0.10). Removal of an additional 133 bp from the 3′ end (25 to 158) increases activity, revealing the presence of a negative regulatory element (p < 0.05; Fig. 7). In addition, upstream sequence located between -324 and -130 may contain a positive regulatory region that acts in the absence of sequence between 158 and 294 (Fig. 7A); however, additional studies are needed to support these conclusions.
To determine whether sequences downstream of the initiation site in the α6 gene also contain important regulatory elements responsible for tissue specificity, we performed transient transfection of primary fibroblast cultures using reporter constructs containing different truncations of the α6 subunit gene. Interestingly, when 136 bp is deleted from the 3′ end of the full-length promoter (-595/294), activity increases by 236% in fibroblasts. This observation demonstrates that sequences in the 5′ untranslated region contribute to tissue specificity (Fig. 8A). The increase in activity seen on removal of the downstream sequence, however, is counteracted by the removal of sequence at the 5′ end (-595 to -324), revealing the presence of a positive regulatory element (Fig. 8A). Removal of additional downstream sequence (133 bp) from the 3′ end of -595/158 returns activity to the level of the full-length construct, suggesting the existence of a positive regulatory element between +25 and +158 (Fig. 8B). By contrast, removal of sequence from the 3′ end of the 5′ truncated constructs does not alter activity in fibroblast cultures (Fig. 8C), indicating that sequence at the 5′ end of the proximal promoter region (-595 to -324) acts in concert with sequence at the 3′ end (+158 to +294).
The exquisite cell type specificity of GABAA receptor α6 subunit gene expression makes it an attractive model for investigating the cis elements involved in restricting gene expression to a subset of neurons. As a consequence of this unique control over transcription, the granule cells of the cerebellum contain a pharmacologically distinct GABAA receptor that is relatively insensitive to the allosteric action of classical benzodiazepines, such as diazepam, whose action is characteristic of the majority of GABAA receptors in the brain (Korpi et al., 1993). The role played by this distinct subtype of GABAA receptors in brain function and in the etiology of disease states, such as alcoholism, is a major unresolved area of molecular neurobiology (Suzdak et al., 1986; Korpi and Uusi-Oukari, 1992; Makela et al., 1995).
Here, we describe the isolation and characterization of the 5′ ends of this important gene in the human, rat, and mouse. Sequence alignment reveals that the sequences of the first three exons are highly conserved among species. The position of the first two introns in the human, rat, and mouse α6 genes is identical, with the size of the first intron in the mouse being slightly larger (225 bp) than the equivalent intron in the human (196 bp) and the rat (198 bp) (Fig. 1A). Intronic sequences are less well conserved than exonic sequences between the three genes, with the sequences in the second intron being more divergent than those of intron 1.
Inverse PCR was used to isolate sequence upstream of the 5′ end of the published rat α6 cDNA sequence (Fig. 1B). The rat and mouse α6 genes contain a cluster of transcriptional start sites spanning 50-60 bp, in agreement with the results of Jones et al. (1996), who used both RNase protection and RACE (rapid amplification of cDNA ends) to identify a cluster of transcriptional initiation sites. However, the major transcriptional initiation site designated here is not in agreement with the major start site described therein (Fig. 2). One explanation for this discrepancy is that the RACE product identified by Jones et al. (1996) may represent a rare transcript rather than the major transcription product from the α6 promoter, as the authors stated in their article.
Several consensus sequences for cis elements are conserved among species, which suggests that they subserve an important functional role in the regulation of gene expression (Fig. 1B). The α6 promoter region lacks TATA motifs near the transcriptional start sites of either the rat or mouse genes. However, an Inr element (CACTTT) surrounds the major transcriptional initiation site (Fig. 2C). This Inr element is therefore part of the α6 core promoter. The presence of an initiator in the core promoter of the α6 gene is consistent with the observation that in other TATA-less promoters, initiators can localize a transcriptional start site and mediate the action of at least some upstream activators (Smale and Baltimore, 1989).
Numerous genes transcribed by RNA polymerase II have been identified that lack a TATA-box ∼30 bp upstream of their initiation site, including other GABAA receptor subunit genes isolated to date [α] (Kang et al., 1994; M. Leach, unpublished data), β1 (S. J. Russek, unpublished data), β3 (Kirkness and Fraser, 1993), and δ (Sommer et al., 1990)]. Whereas the absence of a TATA-box, together with the presence of multiple transcriptional start sites, is often taken as characteristic of promoters for housekeeping genes (Dynan, 1986), promoters for tissue-specific genes can also be TATA-less (Funk et al., 1989). The unique spatial and temporal patterns of expression exhibited by the α6 gene in the CNS demonstrate that a TATA-less gene can be regulated in a highly specific fashion to generate the unique phenotype of cerebellar granule cells. However, an 850-bp fragment of the α6 subunit gene when randomly integrated into the mouse genome did not show consistent expression specific to cerebellar granule cells, although expression was targeted to neurons (Jones et al., 1996). This could be due to the fact that TATA-less promoters when integrated into random regions of the mouse genome may come under the control of additional enhancers that are not present or functional in the endogenous chromatin environment.
We used primary cultures of rat neocortex, hippocampal formation, cerebellum, and hindlimb to determine whether the α6 5′ flanking sequence indeed contains the cis elements necessary to drive both tissue- and cell-specific expression. Transfections into primary cultures of rat cerebellar granule cells demonstrate that an active promoter is located within an 889-bp region of the rat α6 gene (-595/294) (Fig. 4). Moreover, the data suggest that a 155-bp region, spanning -130 to +25 (Fig. 5A-C), is able to drive tissue- and neuronal cell type-specific expression. Although the cis elements responsible for restricting α6 gene expression to cerebellar granule cells remain to be determined, a 60-bp sequence of the promoter (from -130 to -70) plays a major role in cell type-specific expression when assayed in primary brain cultures (Fig. 5D). The 60-bp region displays position-specific activity (Fig. 6), which suggests that it does not act as a classical enhancer but instead controls cell type-specific expression as part of the minimal promoter region. Moreover, the 60-bp region contains several consensus sequences for gene regulatory elements (see Results) that appear to be conserved among the rat, mouse, and human genes, and further analysis will be required to determine which of the elements are active and whether or not their activity is specific to the cerebellum.
Deletional analysis was also used to identify possible regulatory elements that may be active in nonneuronal cells, restricting expression of the α6 gene to neurons (Fig. 8). Removal of 136 bp from the 3′ end of construct -595/+294, a region that contains the 5′ untranslated sequence, increases promoter activity by about threefold in rat fibroblasts (Fig. 8A), whereas removal of this same 3′ sequence from construct -130/+294 has no effect on promoter activity (Fig. 8C). It seems unlikely that a change in RNA stability or secondary structure could account for the differential response of these constructs to deletion as both promoter constructs generate the same 5′ end of the luciferase transcript. That is, the region between +1 and +158 of the α6 gene should be incorporated into the 5′ end of the luciferase transcript in both experiments. Taken together, these observations suggest that the low level of α6 promoter activity in fibroblasts may be controlled by a combination of DNA-binding events, with the binding activities of the transcription factors that recognize sites upstream of the initiation site being interdependent with those that recognize downstream elements (Fig. 8D).
The α6 gene, together with the α1, β2, and γ2 genes, is clustered on human chromosome 5q (Buckle et al., 1989; Johnson et al., 1992; Wilcox et al., 1992; Hicks et al., 1994; Quirk et al., 1994; Russek and Farb, 1994; 1995; Russek, 1999), and its transcriptional unit is organized head to head with that of the β2 gene (Russek, 1999). It is unknown whether this organization is related to coordinate regulation of gene expression by long-range cis elements. The fact that the GABAA receptor subunit genes on chromosome 5 are all highly expressed in granule cells, in contrast to the GABAA receptor subunit genes on chromosomes 4 and 15, does suggest that their organization may contribute to their tissue-specific expression. However, the α6 promoter sequence has no obvious homology with the α1 gene promoter (Kang et al., 1994; Bateson et al., 1995; M. Leach, manuscript in preparation) or with promoters for other GABAA receptor subunit genes (β1, β3, and δ) (Sommer et al., 1990; Kirkness and Fraser, 1993; S. J. Russek, submitted for publication). Furthermore, the brain-specific factor-1 (BSF-1) binding sequence, identified as being involved in the regulation of the δ subunit gene and its cerebellar granule cell-specific expression (Motejlek et al., 1994), is also not present in the proximal region of the rat or mouse α6 genes.
Although we have not yet identified the transcription factors and their cognate regulatory elements that are responsible for the cerebellar granule cell-specific expression of the α6 gene, we have identified regions of DNA sequence that play an important role in granule cell-specific expression. In addition, our data are consistent with the view that negative regulation may be used as a general mechanism to restrict the expression of certain genes to subpopulations of neurons (Mandel and McKinnon, 1993). For instance, deletion of sequences from +158 to +294 results in a reduction of tissue or neuronal cell type specificity. This reduction on deletion indicates that activators required for α6 gene expression are not restricted to cerebellar granule cells, but rather that the binding properties of these factors are affected by the expression of inhibitory proteins specific to nonneuronal cells.
The present results also suggest that positive regulatory elements located in the core promoter bind proteins that are specific to the cerebellum. Deletion of the 60-bp region (-130 to -70) does not alter promoter activity in fibroblasts (Fig. 5D), indicating that regulatory elements within this region are not involved in inhibiting α6 gene expression in nonneuronal tissue; however, the deletion markedly reduces expression in cerebellar granule cells. Therefore, tissue-specific α6 gene expression may result from the lack of specific activator proteins in nonneuronal tissue as well as the presence of specific repressor proteins. Moreover, deletion of the 60-bp region also increases expression in neocortical neurons, indicating that there may be negative and positive regulation mediated by the same regulatory sequences to target precisely mRNA expression to the appropriate neuronal cell type. Studies in transfected primary cultures will lay the foundation for further investigation into the genomic mechanisms of neuron- and cell-specific gene regulation in the nervous system.