| Human Molecular Genetics | Pages |
Functional analysis of the Huntington's disease (HD) gene promoter
Introduction
Results
Interactions of transcription factors with putative binding sites
Deletion analysis of the HD promoter
Analysis of HD promoter polymorphisms
Transcription activation by the 20 bp direct repeat
Discussion
Materials And Methods
Electrophoretic mobility shift assays (EMSAs)
Promoter constructs for expression studies
DNA transfections, and luciferase and [beta]-galactosidase assays
Acknowledgements
References
Functional analysis of the Huntington's disease (HD) gene promoter
The basis for the highly specific neuronal vulnerability seen in Huntington's disease (HD) has not been determined. Recent studies have demonstrated that variation in HD protein expression occurs in the striatum, with affected regions showing increased HD immunoreactivity. Experiments in HD and SCA1 transgenic mice suggest a correlation between phenotypic severity and expression of the mutant transgene. To gain insights into control of HD gene expression, and to investigate the possibility of cell-cell differences in transcription, we have analysed the 5[prime] upstream region of the HD gene in a neuronal (SK-N-SH) and a non-neuronal (JEG3) cell line. Reporter gene assays demonstrated the presence of a key positive-acting region apparently arising from two Sp1 sites in a tandem repeat acting synergistically. This site is polymorphic, and a single Sp1 site is associated with reduced levels of transcription. These experiments also reveal differences in control of expression between neuronal and non-neuronal cell lines.
INTRODUCTION
The mutation underlying the neurodegenerative disorder, Huntington's disease (HD), is a CAG expansion in exon 1 of the IT15 gene (1). Normal chromosomes are polymorphic and contain <36 repeats, while chromosomes with 36 or more repeats are associated with the disease (2). The CAG repeat is translated into a polyglutamine tract. The exact function of the wild-type protein is unknown, but it has some essential role in embryonic development in mice (3-5). The mutant protein is thought to acquire a novel deleterious function (6), a process involving the formation of ubiquitinated neuronal intranuclear inclusions (NIIs), which are N-terminal huntingtin immunoreactive (7,8).
Eight neurodegenerative diseases have been identified which are associated with polyglutamine expansions, namely spinobulbar muscular atrophy (9), dentatorubralpallidoluysian atrophy (10), spinocerebellar ataxia (SCA) types 1 (11), 2 (12-14), 3 (15), 6 (16) and 7 (17), and HD. The size ranges of normal and disease-causing CAG tracts are similar in all, except for SCA6 which has disease expansions of <30 repeats and where pathology may be mediated via a different process. NIIs have also been detected in SCA1 (18) and SCA3 (19) and may be a general feature of these diseases. The expression of the pathogenic protein in each disease is widespread. However, despite the many molecular similarities, the patterns of cell death observed are highly specific to each disorder. In HD, the caudate nucleus, putamen and globus pallidus predominantly are affected, with neuronal loss in layers III, V and VI of the cerebral cortex occurring to a lesser extent (20). Within the striatum, medium spiny neurons are selectively lost whilst large aspiny neurons are spared (21). This highly selective neuronal loss could involve a cell-specific factor distinct from the huntingtin protein. Interaction of huntingtin with several other proteins has been demonstrated, such as HAP1 (22), HIP1 (23,24), ubiquitin-conjugating enzyme (25) and GAPDH (26), but none of these is expressed specifically in affected neurons. Alternatively, neuronal vulnerability may be related to subtle increases in expression of the HD protein. Variation in HD protein levels occurs within the human striatum (27), and the cell types displaying increased expression correspond to the sites affected (28). In addition, a transgene dosage-dependent phenotype is exhibited by HD and SCA1 transgenic mice (29,30). Variable expression of the HD protein may therefore be relevant to pathogenesis.
In order to study the control of gene expression, it is necessary to perform functional analyses. Although both the human (HD) and mouse (Hdh) upstream regions have been sequenced, current information regarding control of HD gene expression is limited. Sequencing analysis showed the promoter to be GC rich and lacking both TATA and CCAAT elements (31). Lin and colleagues mapped two transcription initiation sites at positions -145 and -135 (31; although the authors did not show their raw data); thus we have labelled all sites in this report relative to the translation start site (+1). A highly conserved region, showing 78.8% nucleotide identity between man and mouse, was identified between positions -198 and -49. Putative transcription factor-binding sites were identified on the basis of homology to consensus sequences, and included a conserved Sp1 site (GGGCGG) at -285 to -280, and a conserved AP2 site (CCGCAGGC) at -249 to-242. Another feature of the human sequence was a 20 bp direct repeat (-213 to -174) (31). An ~1 kb fragment upstream of the translation start site contains the regions necessary to drive widespread expression of a truncated human gene in transgenic mice (29). We subsequently showed the HD promoter to be polymorphic at four sites in both control and patient populations (32). The aim of the present study was to characterize the HD promoter further in both a neuronal and a non-neuronal cell line, in order to identify key regulatory elements, to investigate the possibility of cell-specific differences in expression control and to study the functional effects of the polymorphisms identified.
RESULTS
Interactions of transcription factors with putative binding sites
Our initial studies aimed to determine whether cis-acting sequences with homology to transcription factor recognition sites could bind these proteins. This would allow rational design of constructs for functional assays by successive deletion of transcription factor-binding sites. A potential Sp1-binding site (GGGCGG) (at position -285 to -280 relative to the +1 translation start site), a putative AP2 site (CCGCAGGC; -249 to -242), both of which are conserved between man and mouse, and another Sp1 motif in the 5[prime]-untranslated region (5[prime] UTR; position -15 to -10) were identifed on the basis of sequence homology (31). Our further analysis using a commercially available software package (BIMAS) showed possible Sp1 sites at positions -190 to -185 and -210 to -205, in both of the elements of the previously identified 20 bp direct repeat (31).
To determine whether these sequences bind transcription factors in vitro, electrophoretic mobility shift assays (EMSAs) were carried out using oligonucleotides spanning each site in turn. Incubation of an oligonucleotide spanning the conserved Sp1 site (-285 to -280) with a general (HeLa) nuclear extract produced a series of retarded bands, the uppermost of which was abolished when an unlabelled oligonucleotide containing a consensus Sp1 site (5[prime]-GGGGCGGGGC-3[prime]) was used as a competitor (Fig. 1A). In the presence of purified recombinant human Sp1 a single retarded band of similar mobility was produced, which was competed efficiently by a consensus Sp1 oligonucleotide, and weakly by an unlabelled oligonucleotide with the same sequence as the probe, but not by a TFIID-binding site (Fig. 1B). This provides evidence for the specificity of interaction between Sp1 and the conserved Sp1 site oligonucleotide.
Figure 1. The Sp1 site (-285 to -280) conserved between man and mouse binds Sp1 in vitro. (A) EMSA using HeLa nuclear extract and the radiolabelled conserved Sp1 site oligonucleotide. Reactions were incubated with a 10- or 20-fold molar excess of unlabelled consensus Sp1 site oligonucleotide, which competes for the uppermost band (arrow). The open arrow indicates free (unbound) probe. (B) EMSA using purified Sp1 and the labelled conserved Sp1 site oligonucleotide, with a 20- or 50-fold molar excess of the following oligonucleotide competitors: lanes 2 and 3, consensus Sp1 site; lanes 4 and 5, conserved Sp1 site; lanes 6 and 7, TFIID site. Purified recombinant human AP2 interacted with the conserved AP2 site (-249 to -242) to form a single retarded band which was competed efficiently by an oligonucleotide containing a defined consensus AP2 site (5[prime]-GCCCGCGGC-3[prime]) (Fig. 2A). Competition with an unlabelled conserved AP2 site oligonucleotide (same sequence as the probe) was effective at 100-fold molar excess of the competitor (Fig. 2B). However, competitors with unrelated binding sites (Sp1 and TFIID) had no effect (see Fig. 2A). This interaction therefore appears to be specific. Figure 2. Binding of AP2 to the conserved AP2 site (-249 to -242) is specific. EMSA was carried out using purified AP2 and the conserved AP2 site oligonucleotide as the probe. (A) Reactions were incubated with 20- or 50-fold molar excess of the following oligonucleotide competitors: lane 1, no competitor; lanes 2 and 3, defined consensus AP2 site; lanes 4 and 5, conserved AP2 site (same sequence as the probe); lanes 6 and 7, consensus Sp1 site; lanes 8 and 9, TFIID site. Only the defined AP2 site oligonucleotide competes efficiently for the band formed by AP2 binding (arrow). (B) Lane 1, no competitor; lane 2, a 10-fold molar excess of a defined AP2 site as competitor; lanes 3-6, increasing amounts of the conserved AP2 site oligonucleotide as competitor, as indicated. A probe spanning a single 20 bp repeat element (20 bp ×1) formed several complexes in the presence of a general (HeLa) nuclear extract (Fig. 3A). A consensus Sp1 site oligonucleotide used as a competitor virtually abolished the uppermost band, but a defined AP2 site had very little effect. Purified Sp1 bound and produced a single retarded band with the 20 bp site probe, and again this was competed by a defined Sp1 site oligonucleotide (Fig. 3B). In contrast, no evidence for binding by Sp1 to the 5[prime] UTR Sp1 site was seen (data not shown). Figure 3. Sp1 binds specifically to a site within a single copy of the 20 bp repeat element. (A) EMSA using a labelled single 20 bp repeat element oligonucleotide (lane 1: free probe), and HeLa nuclear extract (lanes 2-4). Lane 3 also contains a 30-fold molar excess of a consensus Sp1 site oligonucleotide which competes for the uppermost band (arrow). Lane 4 contains a 30-fold excess of a defined AP2 site, which does not compete for this band. Purified Sp1 also binds to the probe (lane 5). (B) EMSA using the same probe (lane 1: free probe) and purified Sp1 (lanes 2-4), together with a 10- and 50-fold molar excess of a defined Sp1 site. 
Deletion analysis of the HD promoter
In order to characterize the HD promoter functionally, a series of deletion constructs were made, and used in luciferase reporter gene assays in both neuronal (SK-N-SH; human neuroblastoma) and non-neuronal (JEG3; human choriocarcinoma) cell lines. Some of the constructs used in our initial experiments ended at +20 in order to encompass the entire 5[prime] UTR. These have two translation start sites: the HD start at +1 and that of luciferase, which are out-of-frame relative to each other. Any peptide from the HD site would be truncated at a stop codon shortly downstream relative to the luciferase start. In this situation, the translation initiation of luciferase will depend on leaky ribosomal scanning and/or translational re-initiation. In order to rule out the unlikely possibility that luciferase activity does not relate directly to mRNA concentrations in constructs with two start sites, we performed a dose-response experiment in both cell lines. Three different amounts of the [-324/+20] construct, cloned into the pGL3-basic promoterless vector, were transfected in triplicate into each cell line (see Table 1; details in table legend). These data demonstrate that luciferase levels for constructs ending at +20 do depend on the amount of promoter-containing vector transfected, and therefore on mRNA levels (R2 = 0.99 for each cell line).
Table 1.
An initial fragment, extending from position -1032 to +20, was chosen for deletion analysis, since an ~1 kb promoter region has been shown to be sufficient to drive widespread expression of a truncated human HD gene in transgenic mice (29). In order to examine the functional importance of the sites shown to bind Sp1 and AP2 in vitro, PCR primers were used to amplify specific regions of the HD promoter, allowing sequential removal of each site. The fragment from -324 to +20 was chosen as the reference construct (100% luciferase activity) since it spans all the aforementioned conserved sites, and this was transfected in parallel with other constructs in each experiment.
Figure 4. shows the relative levels of luciferase activity produced by each HD promoter construct ending at +20. No difference in activity was observed between the fragment extending from -1032 to +20 [-1032/+20] and the [-324/+20] fragment in JEG3 cells, which suggests that the critical controlling elements are within ~320 bp of the translation start site. However, in the neuronal SK-N-SH cells, a 20% reduction in activity occurs in the [-1032/+20] fragment compared with [-324/+20] (P = 0.002), which may suggest the presence of a weakly negative site between the -1032 and -324 positions in this cell line.
Figure 4. 5[prime] Deletion analysis of the HD promoter in neuronal (SK-N-SH) and non-neuronal (JEG3) cell lines. The 5[prime] and 3[prime] ends of fragments are numbered according to the translation start site (+1). Positions of putative transcription factor binding sites and transcription initiation sites are depicted diagrammatically (see key). Levels of luciferase activity are expressed as a percentage of the reference [-324/+20] construct, set at 100%, and results which differ significantly from those of the reference construct (P < 0.01) are marked with an asterisk (*). n, number of independent transfection experiments. Deletion of the conserved Sp1 site (from -324 to -271) had no significant effect on activity in either cell line (P = 0.20 for SK-N-SH and P = 0.77 for JEG3), nor did deletion of the conserved AP2 site (P = 0.21 for SK-N-SH and P = 0.29 for JEG3). These results are confirmed by the comparison of two fragments lacking the HD translation start site: the activity of the -1032/-15 fragment was not significantly different from that of the -242/-16 fragment in either cell line (P = 0.15 for SK-N-SH and P = 0.4 for JEG3). All three experiments on these fragments in SK-N-SH cells demonstrated a tendency towards lower levels of activity in the [-1032/-15] fragment compared with [-242/-16], and the lack of significance between mean values is compatible with the weak nature of any putative repressor in the -1032 to -324 region. Further 5[prime] deletion from -242 to position -145 totally abolished activity in both cell lines (P = 0.0001 for SK-N-SH and P < 0.0001 for JEG3, for [-324/+20] versus [-145/+20]), suggesting that important positive-acting factors bind in the region -242 to -145. In order to study sites closer to the translation start, the 5[prime] end was fixed at position -242, and 3[prime] deletions were made. For ease of reference and to allow assessment of relative luciferase activity across all experiments, these constructs were also expressed as a percentage of the [-324/+20] (100%) construct (see Fig. 5). The [-242/-2] fragment gave high levels of activity in both cell lines: 2020% in SK-N-SH cells and 518% in JEG3 cells. The increase in luciferase resulting from removal of the HD translation start site was expected: constructs including the HD start rely on leaky ribosomal scanning or translation re-initiation for luciferase translation. Further 3[prime] deletion to -16 gave no significant difference in activity in SK-N-SH cells (P = 0.23 for [-242/-2] versus [-242/-16]) and a slight increase in JEG3 cells (P = 0.026 for [-242/-2] versus [-242/-16]). It is possible that a weak repressor occurs between -16 and -2 in JEG3 cells; however, due to the large number of comparisons made in deletion experiments, it is probably prudent to ignore P-values >0.01. Figure 5. 3[prime] Deletion analysis of the HD promoter in neuronal (SK-N-SH) and non-neuronal (JEG3) cells. The 5[prime] and 3[prime] ends of fragments are numbered according to the translation start site (+1). Levels of luciferase activity are expressed as a percentage of the reference [-324/+20] construct, set at 100%. Statistically significant paired comparisons are indicated by superscript letters. n, number of independent transfection experiments. 3[prime] Deletion to position -126 gave similarly high levels of activity (1966%) in SK-N-SH cells as in the previous fragment (P = 0.84 for [-242/-126] versus [-242/-16]). In JEG3 cells, the mean activity level of [-242/-126] was less than half that of the [-242/-16] fragment (P = 0.003), which implies that a positive-acting element lying between positions -126 and -16 is used for transcription in JEG3 cells, but not in the neuronal cell line. Deletion from -126 to -141 resulted in a significant reduction in activity in SK-N-SH cells (P = 0.002) and in JEG3 cells (P = 0.002). It is likely that this is due to the deletion of the 3[prime] transcription start site which lies in this region. The fall in activity appears to be greater in the JEG3 cells than in SK-N-SH cells (12-fold reduction compared with 3-fold reduction, respectively). These observations are compatible with the hypothesis that the 3[prime] transcription initiation site is the site used predominantly in JEG3 cells. However, we cannot rule out the possibility that another positive-acting element is responsible for the effect seen. To define further the important positive elements in the -242 to -145 region, the 3[prime] end was fixed at position -141 and the 5[prime] end deleted to -171. This virtually abolished activity in SK-N-SH cells (9%, P = 0.0001) and did so entirely in JEG3 cells (0.25%, P < 0.0001). This loss of activity is unlikely to be due simply to the absence of the -141 to -126 sequence, since a fragment spanning from -171 to -128 had equally low activity. In summary, these results suggest the presence of strong positive elements between positions -242 and -171, and between -141 and +126, likely to be the 3[prime] transcription initiation site. The non-neuronal JEG3 cells have a positive-acting site between -126 and -16, and a possible weak repressor element occurs between -1032 and -324 in the neuroblastoma (SK-N-SH) cell line.
Analysis of HD promoter polymorphisms
Our previous studies have shown the HD promoter to be polymorphic in both control and patient populations (32). Seven alleles were identified, containing four polymorphic sites. Two sites involved single base pair substitutions (G->A at -148 and C->T at -103), and did not appear to create or delete any putative transcription factor-binding sites. However, a third polymorphic site is generated by the presence of one, two or three copies of a 20 bp repeat element, which is present in two copies on the commonest allele as the 20 bp direct repeat (-213 to -174). As described above (Fig. 3), Sp1 is able to bind to a single 20 bp stretch in vitro. This raises the possibility that duplication (or triplication) of this site may lead to a corresponding increase in Sp1 binding and increased transcription. Another polymorphic site involved the repetition of a 6 bp stretch (-292 to -287), which does not contain any consensus binding sites when present as a single copy (the commonest allele), but creates a putative Sp1-binding site when duplicated to produce a perfect tandem repeat. EMSAs demonstrated that while purified Sp1 extract cannot bind to a probe spanning a single 6 bp stretch, binding does occur when the site is duplicated (see Fig. 6A). Competition assays demonstrated the specificity of this interaction. The retarded band produced by Sp1 was competed efficiently by an unlabelled consensus Sp1-binding site; competition by a conserved Sp1 oligonucleotide also occurred, but an unrelated (AP2) site did not appear able to compete for binding (Fig. 6B). Differences in ability to bind Sp1 therefore occur at the 6 bp polymorphic site, and may also occur at the 20 bp site.
Figure 6. Sp1 binds to a duplicated 6 bp site, but not to a single copy of the 6 bp element. (A) EMSA using a labelled oligonucleotide containing a single 6 bp element (6 bp ×1) or an oligonucleotide containing two copies of the 6 bp site (6 bp ×2) as indicated. Lanes 1 and 4, free probe; lanes 2 and 5, each probe was incubated with Sp1; lanes 3 and 6, probes were incubated with AP2. The only binding detected was between Sp1 and the duplicated 6 bp site (arrow). (B) Competition EMSA. The labelled 6 bp ×2 fragment (lane 1: free probe) was incubated with Sp1 (lanes 2-6) and a 10- or 30-fold molar excess of the following unlabelled competitor oligonucleotides: consensus Sp1 site (lanes 3-4), defined AP2 site (lanes 5-6). To determine whether these two polymorphisms affect levels of transcription by the HD promoter, reporter gene assays were carried out. Again, the [-324/+20] fragment, amplified from the commonest allele containing one 6 bp stretch and a 20 bp direct repeat, was used as the reference construct. The same primers were used for cloning a 6 bp insertion allele, and an allele in which one 20 bp site was deleted (referred to as the 20 bp deletion allele), to maintain the same 5[prime] and 3[prime] termini. These constructs and their relative luciferase activites are shown in Figure 7. . The 6 bp insertion had no significant effect in either cell line (P = 0.41 for SK-N-SH; P = 0.96 for JEG3), despite the fact that Sp1 can bind to a duplicated 6 bp site in vitro. The 20 bp deletion allele, however, gave significantly lower levels of luciferase activity in JEG3 cells (77%, P = 0.02) and in SK-N-SH cells (57%, P = 0.003). This 20 bp polymorphism therefore does appear to affect transcription from the HD promoter in both cell lines. Figure 7. Functional analysis of the 6 bp and the 20 bp polymorphic sites. Each fragment was amplified using the same primers so that each contains the same 5[prime] and 3[prime] ends. The reference construct was amplified from the commonest allele, and contains one 6 bp element and two copies of the 20 bp element. The activity of an allele containing a 6 bp insertion, and an allele with a 20 bp deletion were measured and expressed as a percentage of the reference. Statistically significant results (P < 0.01) are marked with an asterisk (*). n, number of independent transfection experiments.
Transcription activation by the 20 bp direct repeat
The polymorphic 20 bp repeat element lies within the region shown to contain important positive elements (-242 to -171) in the HD promoter. In order to examine this site in more detail, the fragments depicted in Figure 8. were cloned into the pGL3-promoter vector (Promega) to assess their ability to enhance expression from a heterologous (SV40) promoter. The native vector was included in each transfection and used as the reference construct (100% luciferase activity). No significant increase in activity resulted from addition of a single 20 bp site in SK-N-SH cells (P = 0.63), although a small increase was seen in JEG3 cells (P = 0.03) (Fig. 8). Again, this P-value should be considered with some caution since we have performed multiple independent tests. However, insertion of a duplicated 20 bp stretch upstream of the vector's promoter resulted in enhanced luciferase activity in both cell lines: 287% in SK-N-SH and 206% in JEG3. In each cell line, this was significantly more than the single 20 bp fragment (P = 0.0005 for SK-N-SH and P = 0.0085 for JEG3.)
Figure 8. Enhanced transcription of a heterologous promoter by the 20 bp direct repeat. A single copy of the 20 bp repeat element, a duplicate copy, or a duplicate copy in which two base pairs in each Sp1 site were mutated, were inserted upstream of the SV40 promoter in the pGL3-promoter luciferase reporter gene vector. Relative levels of luciferase activity are expressed as a percentage of that of the native vector (100%), and those which are statistically significant (P < 0.01, see Results) are marked with an asterisk (*). n, number of independent transfection experiments. The enhanced transcription mediated by the 20 bp direct repeat may require either both Sp1 sites or a novel binding site created in the repeat adjoining region. To distinguish between these two possibilities, we generated a duplicated 20 bp fragment in which both Sp1 sites were mutated. These Sp1 sites are on the antisense strand, and each was mutated from 3[prime]-GGGCGG-5[prime] to GGGTTG. As shown in Figure 9. , Sp1 binds to the wild-type 20 bp direct repeat (20 bp ×2), but not to the mutated fragment. The band produced by Sp1 was competed efficiently by a 50-fold molar excess of a consensus Sp1 oligonucleotide (data not shown). When the mutated duplicated 20 bp repeat was cloned upstream of the SV40 promoter, no effect on transcription in SK-N-SH cells was seen (P = 0.81 when compared with the single 20 bp stretch; Fig. 8). In addition, this fragment gave almost identical levels of luciferase activity to the single (wild-type) 20 bp stretch in JEG3 cells (P = 0.84). These data suggest that both Sp1 sites within the duplicated 20 bp repeat are required for enhanced transcription. Figure 9. Sp1 and a factor in HeLa nuclear extract bind to a wild-type duplicated 20 bp repeat (20 bp ×2) but not when the two Sp1 sites within this fragment are mutated (mutated 20 bp ×2). Lanes 1-5: EMSA using the 20 bp ×2 oligonucleotide as a probe, with HeLa nuclear extract (lanes 2-4) or purified Sp1 (lane 5). Lanes 6-10: EMSA using the mutated 20 bp ×2 oligonucleotide as a probe, with HeLa extract (lanes 7-9) or Sp1 (lane 10). A 30-fold molar excess of consensus oligonucleotide competitors was included as indicated. The uppermost band produced by binding of HeLa extract to the wild-type 20 bp ×2 probe is competed specifically by a consensus Sp1 oligonucleotide (lane 3, arrow); no such band appears with the mutated 20 bp ×2 probe. Sp1 produces two bands with the 20 bp ×2 probe (arrowheads), but does not bind to the mutated 20 bp ×2 probe.
DISCUSSION
We have begun to characterize the HD promoter using a combination of sequence analysis, EMSA and luciferase reporter gene assays. Expression studies revealed that the main regulatory elements responsible for transcription lie within a fragment extending from position -324 to +20 relative to the translation start site. In both neuronal and non-neuronal cell lines, a key positive element was identified in the -242 to -171 region, and a further region from -141 to -126 is also required. Despite the ability of Sp1 and AP2 to bind specifically in vitro to their relevant conserved sites, these seem to have minimal effect in mediating transcription. However, the ability of AP2 to activate transcription has been shown to be increased in response to phorbol esters and agents which raise cAMP levels (33). Such agents do not affect the ability of AP2 to bind to DNA, so it is possible that stimulation of cells with similar agents is required for AP2 to enhance transcription from the HD promoter.
3[prime] Deletion from -126 to -141 creates a significant reduction in expression in both cell lines. The two transcription initiation sites (31) occur at positions -145 and -135. Although our results suggest that the 3[prime] site is responsible for the -141 to -126 deletion effect, we cannot exclude the possibility that another positive cis-acting sequence is present in this region.
In addition to the observation that JEG3 cells seem to be more dependent on the 3[prime] transcription initiation site than SK-N-SH cells, certain other cell-specific differences were observed. The most apparent difference is that the region between -126 and -16 has a positive effect only in JEG3 cells. This could be accounted for by a number of possibilities, for example the presence of a cell-specific transcription factor, or cell-specific activation of a binding factor. In addition, a site between -1032 and -324 weakly represses activity in SK-N-SH cells.
Studies of the -242 to -171 positive region demonstrate that a 20 bp direct repeat is sufficient to enhance the activity of a heterologous promoter. Enhancer activity was abolished when two Sp1 sites, one in each 20 bp repeat, were mutated, despite leaving 18 bp overlapping the central axis of the direct repeat intact. This suggests that, in contrast to a transcription factor binding to a site in the repeat adjoining region, stimulation is achieved by Sp1. Since a single Sp1 site had no effect in SK-N-SH cells, the high level expression by two sites in tandem suggests a synergistic, rather than an additive effect. This theory is supported by the differences in expression between alleles with one or two copies of the 20 bp site. Synergism between repeated Sp1 sites has been suggested previously by studies on the SV40 early promoter, since mutation of any one of the three Sp1 sites required here for transcription results in 80-90% reduction in SV40 promoter activity (34). Synergism between proximally and distally bound Sp1 has also been shown (35). In contrast to the SK-N-SH cells, the slightly enhanced transcription mediated by a single 20 bp stretch and by the mutated 20 bp ×2 fragment in JEG3 cells could be due to weak binding by Sp1, or some other factor.
It is possible that reduced expression mediated by a single copy of the 20 bp element on HD chromosomes may cause a delay in the age at onset, or reduced severity of symptoms in HD patients. However, a single copy was only found in 1.2% of HD patients, and since the phase of each polymorphic allele with regard to the CAG expansion in these patients was unknown, we were unable to determine any correlation with the HD phenotype (32). The effect of the 20 bp polymorphism may be slight if the amount of the mutant protein produced by the single 20 bp allele saturates the pathological process. It is interesting to note that this 20 bp element is present as a single copy in non-human primates (32), and can therefore mediate sufficient expression for normal functioning of the gene.
In summary, we have identified a key positive region involved in control of HD gene expression, and characterized it as a region with activity arising from two Sp1 sites acting synergistically. It is possible that other sites, such as distal enhancer elements, may also occur. Our study demonstrates the existence of functional differences between the HD promoter in a neuronal and a non-neuronal cell line: a positive-acting site between -126 and -16 confined to JEG3 cells and a weak negative element between -1032 and -324 in the SK-N-SH cell line.
Future analysis of HD transcript regulation in vivo will allow detailed characterization of cell type and temporal expression control by different cis-acting sequences. Intervention at the level of transcription is one possible means of disease therapy and, in order for expression levels to be reduced artificially, information regarding the control of HD gene expression would be beneficial.
MATERIALS AND METHODS
Electrophoretic mobility shift assays (EMSAs)
The following synthetic oligonucleotides were used as probes (in double-stranded form) for EMSA. [Fragments are numbered in relation to the +1 translation start site, according to the published sequence (GenBank accession no. Y07981) (32) which includes the extra G which we have previously noted, at position -172.] Sp1 conserved, -299 ACCTGCGGGGGCAGGGGCGGGCTGGTTCCC -270; AP2 conserved, -263 GCCATTGGCAGAGTCCGCAGGCTAGGGCTG -234; UTRSp1, -22 CCGTGCCGGGCGGGAGACCG -3; 6 bp ×1, -303 CAGAACCTGCGGGGGCAGGG -284; 6 bp ×2, CAGAACCTGCGGGGGCGGGGGCAGGG; 20 bp ×1, CGCGTGGCCCCGCCTCCGCCGGCGCAGC; 20 bp ×2, -216 CGTGGCCCCGCCTCCGCCGGCGCGGCCCCGCCTCCGCCGGCGCAGCG -170; mutated 20 bp ×2, -216 CGTGGCCCCAACTCCGCCGGCGCGGCCCCAACTCCGCCGGCGCAGCG -170.
The 20 µl binding reactions containing 2 µg of poly (dIdC) and 5 µg of HeLa nuclear extract or 0.25-1 U of purified recombinant Sp1 or AP2 (Promega), in 20 mM HEPES, 1 mM MgCl2, 4% Ficoll, 0.5 mM dithiothreitol (DTT), 50 mM KCl, were incubated with or without unlabelled competitor oligonucleotide (Promega), on ice for 10 min. Two µl of end-labelled probe was added and incubated for a further 30 min on ice. Reactions were analysed on a 4% non-denaturing polyacrylamide gel in 1× TBE, and exposed to X-ray film.
Promoter constructs for expression studies
Regions of the HD promoter were amplified by PCR using sense primers with an SacI-containing tag, and antisense primers with either an XhoI or an NheI tag. Sequences of forward primers were as follows: 5[prime]-GCGCGAGCTCTTCTCGCTGCACTAATCACA-3[prime] (-1262 to -1242), 5[prime]-GCGCGAGCTCAGCGGCTTGCTGTGTGAGG-3[prime] (-324 to -305), 5[prime]-GCGCGAGCTCCTGGCCAGCCATTGGCAGA-3[prime] (-271 to -252), 5[prime]-GCGCGAGCTCTAGGGCTGTCAATCATGCTGG-3[prime] (-242 to -221), 5[prime]-GCGCGAGCTCGCTGCCGGGACGGGTCCAA-3[prime] (-145 to -127); and reverse primers were: 5[prime]-GCGCCTCGAGCTTTTCCAGGGTCGCCAT-3[prime] (+20 to +1), 5[prime]-GCGCGCTAGCGGTCTCCCGCCCGGCA-3[prime] (-2 to -19), 5[prime]-GCGCGCTAGCGGCACGGCAGTCCCCGGAG 3[prime] (-15 to -34), 5[prime] GCGCCTCGAGGCACGGCAGTCCCCGGAG-3[prime] (-16 to -34), 5[prime]-GCGCGCTAGCTTGGACCCGTCCCGGCAG-3[prime] (-126 to -144), 5[prime]-GCGCCTCGAGCAGCCCCCACGGCGCCTT-3[prime] (-141 to -159). The two smallest fragments, -171/-141 and -171/-128, were generated by annealing synthetic single-stranded oligonucleotides designed to create a 5[prime] SacI site and a 3[prime] NheI site when in double-stranded form. The sense oligonucleotides of these fragments were as follows: -171/-141, 5[prime]-CGTCTGGGACGCAAGGCGCCGTGGGGGCTG-3[prime]; -171/-128, 5[prime]-CGTCTGGGACGCAAGGCGCCGTGGGGGCTGCCGGGACGGGTCCAG-3[prime]. PCR products were double-digested to create cohesive ends, and all fragments were purified before ligating into identical sites in the pGL3-basic promoterless luciferase reporter gene vector (Promega). To study enhancer activity of the 20 bp stretch, the 20 bp ×1, 20 bp ×2 and mutated 20 bp ×2 fragments were designed so that 5[prime] SacI and 3[prime] NheI cohesive ends were created when oligonucleotides annealed in double-stranded form. These fragments were ligated into the pGL3-promoter luciferase vector (Promega), which contains the SV40 promoter. Large-scale plasmid isolation and purification was carried out using the endotoxin-free Maxi Kit (Qiagen). Plasmids were diluted to 1 µg/µl, and the exact concentrations were determined again in triplicate using a spectrophotometer, immediately before each transfection.
DNA transfections, and luciferase and [beta]-galactosidase assays
The SK-N-SH neuroblastoma cell line (ECACC #86012802) and a choriocarcinoma cell line of epithelial morphology (JEG3) (ECACC #92120308) were used in transfections. The calcium phosphate co-precipitation method was used to transfect plasmids into JEG3 cells, and the SuperFect Transfection Reagent (Qiagen) was used for SK-N-SH cells. Five hundred ng of plasmid for JEG3 cells, or 1 µg for SK-N-SH cells, together with 100 ng of a plasmid containing the gene for [beta]-galactosidase linked to a constitutively active human elongation factor 1[alpha] promoter, were co-transfected into cells in a 24-well plate. Each promoter construct was transfected in at least three different experiments, in triplicate wells. Cells were harvested ~36 h post-transfection, and levels of luciferase and [beta]-galactosidase activity determined (36). In each experiment using pGL3-basic, the native plasmid and the [-324/+20] reference construct were transfected in parallel with the other HD promoter constructs. In experiments to determine enhancer activity, the native pGL3-promoter vector was transfected in parallel and used as the reference. The luciferase activity from each well was normalized to the [beta]-galactosidase activity to control for variations in transfection efficiency. In each experiment, the subsequent mean luciferase activity of the native plasmid was deducted from the mean of each HD promoter construct, and each result expressed as a percentage of that of the reference construct (100% activity). Statistical analyses were carried out using the Macintosh Statview[trade] software package. Two-tailed paired t-tests were used to compare each experimental fragment with its (100%) reference construct. Where two experimental fragments were compared with each other, only the results from experiments in which both fragments were transfected simultaneously were used. Two-tailed unpaired t-tests were used for the comparison between cell lines, and to compare the single 20 bp site with the 20 bp ×2 and the mutated 20 bp ×2 fragments.
ACKNOWLEDGEMENTS
We thank Krishna Chatterjee and Odelia Rajanayagam for advice and technical guidance. This research was supported by the Wellcome Trust (R.C.) and the Rehabilitation and Medical Research Trust (D.C.R.). D.C.R. is a Glaxo Wellcome Research Fellow.
REFERENCES
This page is run by Oxford University Press, Great Clarendon Street, Oxford OX2 6DP, as part of the OUP Journals
Comments and feedback: www-admin{at}oup.co.uk
Last modification: 4 Apr 1998
Copyright© Oxford University Press, 1998.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  H. Ryu, J. Lee, S. W. Hagerty, B. Y. Soh, S. E. McAlpin, K. A. Cormier, K. M. Smith, and R. J. Ferrante ESET/SETDB1 gene expression and histone H3 (K9) trimethylation in Huntington's disease PNAS, December 12, 2006; 103(50): 19176 - 19181. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. J. Ferrante, H. Ryu, J. K. Kubilus, S. D'Mello, K. L. Sugars, J. Lee, P. Lu, K. Smith, S. Browne, M. F. Beal, et al. Chemotherapy for the Brain: The Antitumor Antibiotic Mithramycin Prolongs Survival in a Mouse Model of Huntington's Disease J. Neurosci., November 17, 2004; 24(46): 10335 - 10342. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. Tanaka, J. Shouguchi-Miyata, N. Miyamoto, and J.-E Ikeda Novel Nuclear Shuttle Proteins, HDBP1 and HDBP2, Bind to Neuronal Cell-specific cis-Regulatory Element in the Promoter for the Human Huntington's Disease Gene J. Biol. Chem., February 20, 2004; 279(8): 7275 - 7286. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. L. Sugars, R. Brown, L. J. Cook, J. Swartz, and D. C. Rubinsztein Decreased cAMP Response Element-mediated Transcription: AN EARLY EVENT IN EXON 1 AND FULL-LENGTH CELL MODELS OF HUNTINGTON'S DISEASE THAT CONTRIBUTES TO POLYGLUTAMINE PATHOGENESIS J. Biol. Chem., February 6, 2004; 279(6): 4988 - 4999. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Lee, E. H. Park, G. Couture, I. Harvey, P. Garneau, and J. Pelletier An upstream open reading frame impedes translation of the huntingtin gene Nucleic Acids Res., December 1, 2002; 30(23): 5110 - 5119. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Wyttenbach, J. Swartz, H. Kita, T. Thykjaer, J. Carmichael, J. Bradley, R. Brown, M. Maxwell, A. Schapira, T. F. Orntoft, et al. Polyglutamine expansions cause decreased CRE-mediated transcription and early gene expression changes prior to cell death in an inducible cell model of Huntington's disease Hum. Mol. Genet., August 1, 2001; 10(17): 1829 - 1845. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. Narain, A. Wyttenbach, J. Rankin, R. A Furlong, and D. C Rubinsztein A molecular investigation of true dominance in Huntington's disease J. Med. Genet., October 1, 1999; 36(10): 739 - 746. [Abstract] [Full Text]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||