The long QT syndrome genes human ether-a-go-go-related gene (HERG1) and voltage-gated K+ channel, KQT-like subfamily, member 1, gene (KCNQ1), encoding K+ channels critical to the repolarization rate and repolarization reserve in cardiac cells, and thereby the likelihood of arrhythmias, are both composed of two isoforms: HERG1a and HERG1b and KCNQ1a and KCNQ1b, respectively. Expression of these genes is dynamic, depending on the differentiation status and disease states. We identified their core promoter regions and transcription start sites. Our data suggest that HERG1a and HERG1b, and KCNQ1a and KCNQ1b, represent independent transcripts instead of being alternatively spliced variants of the same gene, for they each have their own transcription start sites and their own promoter regions. We obtained data pointing to the potential role of stimulating protein 1 (Sp1) in the transactivation of these genes. We compared expression profiling of these genes across a variety of human tissues. Consistent with the general lack of cis elements for cardiac-specific transcription factors and the presence of multiple sites for ubiquitous Sp1 sites in the core promoter regions of HERG1a/HERG1b and KCNQ1a/KCNQ1b genes, the transcripts demonstrated widespread distribution across a variety of human tissues. We further revealed that the mRNA levels of all HERG1 and KCNQ1 isoforms were asymmetrically distributed within the heart, being more abundant in the right atria and ventricles relative to the left atria and ventricles. These findings open up an opportunity for studying interventricular gradients of slow and rapid delayed rectifier K+ current and of cardiac repolarization as well. Our study might help us understand the molecular mechanisms for arrhythmias since heterogeneity of ion channel activities is an important substrate for arrhythmogenesis.
- potassium channels
- human ether-a-go-go-related gene
- voltage-gated potassium channel
- KQT-like subfamily
- member 1
- stimulating protein 1
to date, no less than eight long QT syndrome (LQTS) genes have been isolated, which, when genetically defected (loss of function mutations), can lead to different forms of LQTS. Among these LQTS genes, KvLQT1 [voltage-gated K+ channel, KQT-like subfamily, member 1 (KCNQ1) or Kv7.1] and human ether-a-go-go-related gene (HERG1 or KCNH2) are responsible for the majority (∼85–90%) of cases of inherited LQTS. Moreover, HERG1 protein is also a pharmacological target for the majority of acquired LQTS as a result of drug blockade. HERG1 encodes the pore-forming α-subunit of a K+ channel with biophysical and pharmacological properties similar to those of rapid delayed rectifier K+ current (IKr), a critical repolarizing current in cardiac cells (26). In the mammalian heart, two HERG1 transcripts (long isoform HERG1a and short isoform HERG1b) encode proteins differing in their NH2-terminal sequence and gating properties (8, 11, 14, 16, 23). A recent study (11) proposed that cardiac IKr channels are minimally composed of HERG1a and HERG1b α-subunits that coassemble in the membrane. On the other hand, KCNQ1 encodes the pore-forming α-subunit of the K+ channel underlying slow delayed rectifier K+ current (IKs), another critical repolarizing current in cardiomyocytes (25). Intriguingly, KCNQ1 is also composed of two isoforms differing in their NH2-termini: isoform 1 and isoform 2; isoform 2 encodes a translational start in the middle of membrane segment 1 (9, 10, 20). When expressed in a heterologous context, the isoform 2 protein functions as a dominant negative isoform (9). For convenience, we designate here the two isoforms KCNQ1a (the long isoform or isoform 1) and KCNQ1b (the short isoform or isoform 2).
Strikingly, the expression of both HERG1 and KCNQ1 genes is dynamic, depending on the differentiation status and cell cycle of the cells, contributing importantly to the developmental evolution of myocardial action potential morphology. For example, IKr is the sole component of delayed rectifier K+ current in fetal day 18 mouse ventricles, yet both IKr and IKs have been observed on postnatal day 1. By day 3, IKs is the dominant component. With further development into adulthood, neither IKr nor IKs has been observed (31, 32). With respect to the canine heart, it was found that IKs is absent or minimal before an age of 4 wk, at which time IKr is the major repolarizing current (32). With progression toward adulthood, IKs appears and increases in density, whereas IKr density diminishes. Moreover, when adult cardiac cells become dedifferentiated or cancerous, such as in AT-1 and HL-1 cells (murine atrial tumor cell lines), IKr regains its predominance among expressed K+ channels (6, 37). While in tissues other than the heart and brain there is restricted HERG1 expression under normal situations, tumor cells of various histological origins overexpress HERG1a and HERG1b genes and their protein products (1, 2, 4, 7, 30, 33). It appears that expressions of HERG1 and KCNQ1 genes are tightly controlled by certain factors according to a defined genetic program related to morphogenesis during development. Elucidating the transcriptional control of HERG1 and KCNQ1 genes is necessary for our better understanding of the molecular mechanisms for certain pathological conditions such as arrhythmogenesis and tumorigenesis associated with abnormalities of expression of the channels encoded by these genes.
These considerations promoted us to perform the present study with the following objectives: 1) to elucidate the genomic structure and promoter regions of HERG1 and KCNQ1 genes and 2) to map the distributions of HERG1 and KCNQ1 transcripts in multiple human tissues and various regions of the heart. Our study defined the transcription start sites (TSSs) and 5′a-flanking genomic sequences of each isoform and delineated critical genomic elements necessary for the transcriptional control of these genes. Our data demonstrated tissue-dependent distributions of the transcripts and revealed regional differences of expressions of these isoforms within cardiac tissues. These findings shed light on the molecular mechanisms for transcriptional control of these genes and enable future work to determine factors responsible for tissue-specific expression.
Rapid amplification of cDNA ends.
TSSs of HERG1b, KCNQ1a, and KCNQ1b genes were determined with Ambion's RNA-Ready cDNA Human Heart RNA Ligase-Mediated Rapid Amplification of cDNA Ends (5′-RACE) kit, as previously described (15). The human ventricular RNA sample was purchased from Clontech. Gene-specific primers (GSPs) were designed based on the human HERG1b sequence (Genbank Accession No. NM_172057); GSP1 was 5′-CTTCCGTCTCCTTCAGCAGG-3′, and GSP2 (nested) was 5′-ATGCAGGATGGTCCAGCGGT-3′, corresponding to 597–616 and 491–510 bp, respectively. For KCNQ1a (GenBank Accession No. NM_000218), GSPs used for 5′-RACE were GSP1: 5′-AGGAAGACGGCGAAGTGGTA-3′ and GSP2: 5′-CGCGTGCTGTAGATGGAGAC-3′. For KCNQ1b (GenBank Accession No. NM_181798.1), GSP1 was 5′-ACGTACTCCGTCCCGAAGAA-3′ and GSP2 was 5′-CACGATGAGGTCGATGATGG-3′.
RNase protection assay.
To determine the exact TSSs of HERG1 genes, we hybridized 30 μg of total RNA from human hearts (Clontech) with human genomic antisense riboprobes in a solution hybridization/RNase protection assay (RPA) (15, 38). For HERG1b, a human genomic antisense riboprobe (1,000 bp) corresponding to the downstream 900 bp of intron 5 and upstream 100 bp of exon 6 of the HERG1 genomic sequence (NM_172057) was used. For KCNQ1a, a human genomic antisense riboprobe (900 bp) corresponding to 580 bp upstream of the ATG of KCNQ1a and the first 320-bp region of exon 4 of the KCNQ1 genomic sequence (NT_009237.17) was used. For KCNQ1b, a human genomic antisense riboprobe (800 bp) corresponding to a 576-bp sequence [which contains a 400-bp flanking region plus the TSS (ATG) and the immediate next 2 nucleotides] in intron 1 and the following 224-bp region within exon 2 of the KCNQ1 genomic sequence (NT_009237.17) was used. The fragments were PCR synthesized and cloned to the pGEMT easy vector for the preparation of RNA probes. RNA probes were synthesized using in vitro transcription kits (Ambion). RPAs were performed according to HybSpeed RPA (Ambion) protocols. Yeast tRNA (20 μg) was used as a negative control to test for the presence of probe self-protection bands.
PCR amplification of putative promoter regions and construction of promoter-luciferase fusion plasmids.
A series of fragments of varying lengths including the TSSs of HERG1 and KCNQ1 promoters were amplified with human genomic DNA and Homo Sapiens bacterial artificial chromosome clone RP11–166D23 (GenBank Accession No. AC011234) for HERG1 and RP5–915F1 (GenBank Accession No. AC124057) for KCNQ1 genes as templates using specific primers and PCR Advantage and Advantage-GC genomic polymerase mixes (Clontech K-1905-Y). PCR products were subcloned into the luciferase-containing PGL3-Basic (Promega) vector. The integrity and orientation of all constructs were confirmed by restriction endonuclease analysis and DNA sequencing.
The cell lines used in this study were all purchased from the American Type Culture Collection (Manassas, VA). H9c2 cells (rat ventricular cell line) and HEK-293 cells (human embryonic kidney cell line) were cultured in DMEM. SKBr3 cells (human breast cancer cell line) were maintained in McCoy's 5A modified medium, and LNCaP cells (human prostate cancer cell line) were maintained in RPMI 1640 medium. Cultures were all supplemented with 10% FBS and 100 μg/ml penicillin-streptomycin. Noncardiac cell lines were used to serve as a comparison to H9c2 cardiac cells. SKBr3 cells have been shown to express endogenous HERG1 proteins (30), and HEK-293 and LNCaP cells do not express functional HERG1 proteins (33).
Transfection and luciferase assay.
Cells (1 × 105 cells/well) were transfected with 1 μg PGL3-target DNA (firefly luciferase vector) and 0.1 μg PRL-TK (TK-driven Renilla luciferase expression vector) with Lipofectamine 2000 (Invitrogen). After transfection (48 h), luciferase activities were measured with a dual-luciferase reporter assay kit (Promega) on a luminometer (Lumat LB9507) (15, 18).
mRNA samples from various human tissues were purchased from Ambion and used for testing the expression of HERG1 and KCNQ1 isoforms. For each tissue, mRNA samples were pooled from three samples extracted from different individuals (all Caucasians, both genders, age: 59–72 yr). For the comparison of expression levels of isoforms among the four chambers of the heart, mRNAs were isolated from the hearts of the same individuals. TaqMan quantitative assay of transcripts was performed with real-time two-step RT-PCR (GeneAmp 5700, PE Biosystems), involving an initial RT with random primers and subsequent PCR amplification of the targets (15, 18). Master RT reaction mix (50 μl) [containing 10× RT buffer (10 μl), 25× dNTPs (4 μl), 10× random primers (10 μl), MultiScribe Reverse Transcriptase (50 U/μl) (5 μl), and nuclease-free H2O (21 μl)] and 50 μl of RNA sample (10 μg) were mixed into a PCR tube. The reaction was carried out at 25°C for 10 min and 37°C for 120 min to obtain first-strand cDNAs. Master PCR mix (23 μl) [including 2× TaqMan universal PCR master mix (12.5 μl), 20× assay mix (0.63 μl), and H2O (9.9 μl)] and 2 μl of the first-strand cDNA sample were pipetted into 96-well plates and centrifuged at 2,000 rpm for 2 min at 4°C. The amplification reaction was set as follow: 50°C for 2 min and 95°C for 10 min, followed by 15 s at 95°C and 1 min at 60°C for 40 cycles. Human GAPDH control reagents (Applied Biosystems) were used as internal controls. We used the comparative cycling method (ΔC) to quantify the level of target transcripts (GeneAmp 5700 sequence detection system). ΔC values were calculated as follows: ΔC = CTarget − CGAPDH and ΔΔC = ΔCDrug − ΔCNC, where C is the number of cycles for the amplification of target transcripts (CTarget) or GAPDH (CGAPDH) and ΔCDrug represents the ΔC obtained from RNA samples treated with drugs the ΔC from RNA samples without drug treatment. Finally, the amount of target transcripts, normalized to the endogenous reference GAPDH and relative to control, was calculated by 2−ΔΔC.
Group data are expressed as means ± SE. Statistical comparisons (performed using ANOVA followed by Dunnett's method) were carried out using Microsoft Excel. A two-tailed P < 0.05 value was taken to indicate a statistically significant difference. Nonlinear least-square curve fitting was performed with CLAMPFIT in pCLAMP 8.0 or GraphPad Prism. Cis elements for transcription factor binding sites were analyzed with MatInspector version 2.2.
Identification of TSSs of HERG1 and KCNQ1 genes.
HERG1 genes refer to two different variants that diverge at the NH2-terminus: HERG1a (HERG) and HERG1b. As documented in our previous study (15), HERG1a contains a single TSS (designated as +1) located 79 bp upstream of the translation start codon (ATG) of the HERG1a 5′a-untranslated region (5′a-UTR; Fig. 1A). For HERG1b, 5′a-RACE identified one major fragment of 403 bp upstream of ATG, along with two minor fragments with sizes of 150 and 296 bp, respectively, in the 5′a-UTR of HERG1b (Fig. 1, B and C). Sequencing of these fragments indicated that the 403- and 296-bp bands represented two separate TSSs of HERG1b. The potential TSSs of HERG1b described above were verified by RPAs using RNA samples from the human heart (Fig. 1C). The major fragment of 400 bp length, corresponding to the larger fragment obtained by 5′a-RACE, was identified, along with a weak band of ∼300 bp. These results defined two principal TSSs in HERG1b: the larger fragment containing the most upstream TSS revealed by 5′a-RACE is likely the major form of HERG1b transcript in the human heart, and the smaller fragment represents a relatively rare form of HERG1b transcript.
Similar to HERG1 genes, KCNQ1 genes are also composed of two isoforms: the long isoform KCNQ1a and the short isoform KCNQ1b, which also differ in their NH2-termini. 5′a-RACE was used to obtain the 5′a-ends of KCNQ1a and KCNQ1b transcripts (Fig. 2, A and B). For KCNQ1a, the nested primer (the second reverse primer) was designed from the very end of the NH2-terminus of KCNQ1a because this region is unique to KCNQ1a and is absent in KCNQ1b. The experiment using RNA samples from the human heart yielded two discrete fragments. Sequence analysis suggested that the two fragments represent separate TSSs: TSS1 (designated as +1) and TSS2 (position +12), which are located 80 and 68 bp, respectively, upstream of the translation start codon (ATG) of the KCNQ1a gene (Fig. 2C). For KCNQ1b, 5′a-RACE identified a single discrete fragment of ∼400 bp. Sequencing analysis indicated that the fragment spanned a portion of the first intron of KCNQ1 genomic DNA and a portion of the KCNQ1b NH2-terminus (Fig. 2). The TSS of KCNQ1b was defined to 274 bp upstream of the translation start codon (ATG) of KCNQ1b. The potential TSSs of KCNQ1a and KCNQ1b described above were verified by RPAs using RNA samples from the human heart, which yielded the expected fragments of 500 and 400 bp, respectively, with our probes (Fig. 2C). Our RPAs were not able to distinguish the 12-bp difference between the two fragments of KCNQ1a, as identified by 5′a-RACE.
Genomic arrangement of HERG1 and KCNQ1 genes and their promoter regions.
Figure 1D shows a schematic view of the genomic arrangement of HERG genes with HERG1a and HERG1b aligned for comparison. Based on the results described above, it was clear that the HERG1a promoter is localized to the 5′a-flanking region of the HERG1a gene and that the HERG1b promoter falls within intron 5 of HERG genomic DNA. Genomic exon 6 corresponds to the NH2-terminus of HERG1b, whereas HERG1a uses genomic exon 6 as part of its intron that spans genomic intron 5, exon 6, and intron 6. As such, HERG1b uses the downstream 403-bp region of intron 5 as its 5′a-UTR and the 250-bp region upstream of the TSSs as its core promoter sequence.
Genomic analysis indicated that the promoter sequence of KCNQ1a falls into the 5′a-flanking region of the KCNQ1 gene. KCNQ1b uses a portion of intron 1 of KCNQ1 genomic DNA as its promoter region, and its TSS and translation start codon (ATG) plus the next two nucleotides both fall within the intron 1 region (Fig. 2D).
Structural analysis of 5′-flanking regions of HERG1 and KCNQ1 genes.
Computer analysis of 3,000-bp 5′-flanking regions of HERG1 and KCNQ1 using the MatInspector program revealed that none of them contain the canonical mammalian TATA box within 1 kb upstream of their TSSs. They also lack other known common promoter elements that are required for the transcription initiation complex, including the initiator element, the downstream promoter element, and the transcription factor IIB recognition element. Less common promoter elements such as the downstream core element and multiple start site element downstream 1 are also missing. There is one putative CCAAT site in the HERG1a 5′a-flanking region, which is absent in the HERG1b promoter sequence. KCNQ1b has five CCAAT boxes, but KCNQ1a has none. However, among these five CCAAT consensus sites, only one is located within the core promoter region (−33), and others are at least 2.3 kb away from the TSS.
Cis elements for heart-specific GATA4 as the potent transactivator in cardiac cells were absent in the 5′a-flanking regions of HERG1 and KCNQ1 genes. Consensus sites for other members of the GATA family were identified. The distal 5′a-flanking region of HERG1b contains two putative GATA1 binding sites (−628 and −1431) and one GATA3 binding site (−1446). The KCNQ1a 5′a-flanking region contains three GATA1 consensus sites located distal to TSS1 (1.4, 1.6, and 2.0 kb upstream). KCNQ1b contains two GATA1 sites with one site at 50 bp and the other site at 1.5 kb upstream of its TSSs. Cardiac-specific homeobox Nkx2.5 was found in none of the genes. Cis elements for other transcription factors important in cardiac development and function, including serum response factor, Hand2, and MyoD, were found in the 5′a-flanking regions of HERG1 and KCNQ1 genes (Table 1).
Characterization of promoter regions of HERG1 and KCNQ1 genes.
To assess the functional role of the 5′a-flanking regions of HERG1 and KCNQ1 in the transcriptional regulation of these genes, various lengths of putative promoter fragment-luciferase constructs were generated and tested for their ability to drive the expression of the reporter gene in a rat ventricular cell line (H9c2) and a human embryonic kidney cell line (HEK-293).
The core promoter of the HERG1a gene was defined to 487 bp, which showed a maximum luciferase activity greater than eightfold times the activity of the promoterless vector (PGL3-basic; Fig. 3A). For HERG1b, the +85/+191 fragment demonstrated small but statistically significant luciferase activity (Fig. 3B). This fragment is contained within the 5′a-UTR of the TSS1 variant but is in the proximal promoter region of the TSS2 isoform. Significant luciferase activity was obtained with a longer fragment (−60/+191), indicating that the basal promoter activity is contained in the very proximal 5′a-flanking region (i.e., the first 60 bp upstream of TSS1). Maximum activity was reported by −250/+191, −390/+191, and −595/+191 constructs, suggesting that the −250/+191 fragment contains the core promoter sequence of HERG1b (Fig. 3B).
KCNQ1a promoter constructs had ∼10–15 times the activity of the promoterless vector (pGL3 basic) and 50% of the promoter activity of the pGL3 construct, which contained both the SV40 promoter and enhancer (pGL3 control; data not shown). Overall transcription activities of different KCNQ1a promoter constructs were similar in all three cell lines, and, consistently, the −329/+60 fragment elicited the maximum promoter activity (Fig. 4A). The core promoter of the KCNQ1a gene was thus defined to −329/+60. For KCNQ1b, although a statistically significant luciferase activity was seen with the −117/+71 fragment relative to the promoter-free PGL3 vector, more robust activities were consistently observed with longer fragments (Fig. 4B). For example, in both H9c2 and HEK-293 cells, luciferase activities increased by ∼15-fold with the vector containing the −1336/+71 fragment of the KCNQ1b promoter.
Multiple stimulating protein 1 cis elements and CpG islands in HERG1 and KCNQ1a core promoter regions.
Of particular note is the high GC content of the core promoter regions of HERG1a, HERG1b, and KCNQ1a. The high GC content confers three important features to these genes. First, HERG1a, HERG1b, and KCNQ1a all contain multiple stimulating protein 1 (Sp1) consensus sequences within their core promoter regions: −224, −366, −380, and −428 loci for HERG1a (Table 2); +118, +54, +26, −17, and −116 loci for HERG1b (relative to TSS1 in Figs. 1 and 2); and −61, −114, −134, −190, −259, −277, −295, −347, −407, and −460 for KCNQ1a. By comparison, although there are seven putative Sp1 binding sites within 3 kb length of the KCNQ1b 5′-flanking region, only one site is proximal to its TSS (at position −15) and the other six sites are ∼1 kb distal to its TSS. Strikingly, the positions of Sp1 clusters corresponded well to the fragments that demonstrated significant promoter activities, as shown in Figs. 3 and 4. For example, the −487/+4 fragment, which contains a cluster of four Sp1 sites in HERG1a, showed the maximum promoter activity, and similar situations were seen with HERG1b and KCNQ1a. For KCNQ1b, the minimal promoter activity corresponded roughly to the presence of a single Sp1 site proximal to the TSS, and the greatest promoter activity was coincident with the −1336/+71 fragment, which contains a cluster of three Sp1 elements.
The second feature of a GC-rich promoter is often the presence of so-called CpG islands (28, 36), which are clusters of GC dinucleotides near the TSS. Using the CpG Island Searcher (http://www.cpgislands.com) (27), we made a prediction of CpG islands in the promoter regions (500 bp upstream of TSSs) of HERG1a, HERG1b, and KCNQ1a, and we identified two CpG islands in HERG1a and KCNQ1a and three CPG islands in HERG1b (Fig. 5).
Another important feature associated with Sp1-rich promoters is the presence of GAGA boxes around Sp1 elements (3, 34). We identified a total of seven GAGA boxes within 3-kb of the 5′-flanking region of KCNQ1b, among which four are in a cluster within the −1336/+71 fragment (Table 2). By comparison, there is no GAGA box in the KCNQ1a 5′-flanking region, and there is only one GAGA box in each HERG1a and HERG1b promoter region.
Expression and tissue distribution of HERG1 and KCNQ1 genes.
The expression of HERG1 and KCNQ1 genes was assessed from two aspects. First, the presence of HERG1 and KCNQ1 transcripts was determined by TaqMan real-time RT-PCR with RNA samples extracted from various human tissues. We found that the transcripts of HERG1 and KCNQ1 isoforms existed in various tissues, although the relative expression levels varied (Fig. 6, A and B). In particular, HERG1 and KCNQ1 isoforms were all abundantly expressed in the heart, pancreas, and colon, but poorly in skeletal muscle and breast tissues.
Second, we compared the relative abundance of transcripts across the four chambers of the human heart: left ventricle (LV), right ventricle (RV), left atrium (LA), and right atrium (RA). The results shown in Fig. 6, C and D, revealed that the four genes have a similar pattern of expression with significant interventricular and interatrial gradients of mRNA concentrations: RV > LV and RA > LA. The regional differences of the four mRNAs were within the same range. There were no atrial-ventricular differences of expression.
K+ channels are represented by an extremely large and varied superfamily of genes including pore-forming α-subunits and regulatory auxiliary β-subunits. Many of these channels undergo remodeling processes, mostly with changed gene expressions, during pathogenesis and disease progression. A number of K+ channels have been investigated in regard to their genomic structures for transcriptional regulation, with their promoter regions identified and characterized. However, a majority of these studies were conducted with rat and mouse genes, and the findings may or may not be applicable to human genes considering the large interspecies variations in the 5′-flanking regions of the genes. Research on promoter elements of human K+ channel pore-forming α-subunits has been sparse despite that a recent report (17) described the transcriptional control of several human KCNE genes (KCNE1–5) encoding a family of single-transmembrane domain K+ channel β-subunits that modulate the properties of several K+ channel α-subunits. HERG1 and KCNQ1 genes encode K+ channel α-subunits that critically determine the repolarization rate and repolarization reserve in cardiac cells and thereby the likelihood of arrhythmias. Expression levels of these genes define the functional capacity of their gene products, IKr and IKs, which manifest in the course of the early development of the heart and during pathogenesis with dramatically altered gene expression. Elucidation of the promoter elements and transcriptional control of these genes is therefore a necessary step toward understanding the functional adaptation and impairment during developmental stages and pathological processes associated with electrical remodeling.
Transcriptional control of HERG1 and KCNQ1 genes.
HERG1a and HERG1b have been believed to be alternatively spliced variants of the same gene, as do KCNQ1a and KCNQ1b (8, 14). The present study does not support this notion; instead, our data suggest that they all represent independent transcripts because they each have their own TSSs and their own promoter regions.
Regulation of transcription is a complex set of events controlled by DNA sequences positioned in proximity to the genes (promoters) and by elements acting at a distance (enhancers). Promoters and enhancers that activate polymerase II transcribed mRNA genes are formed by a combinatorial puzzle of short sequences recognized by sequence-specific regulators.
One common feature of the promoter regions of HERG1 and KCNQ1 genes is that they are GC rich; there exist multiple Sp1 consensus binding sites in the proximal promoter regions (>5 sites) of HERG1a, HERG1b, and KCNQ1a genes and in the 5′a-flanking region of KCNQ1b. Sp1 is a widely distributed member of a multigene family of zinc finger transcription factors that bind DNA in mammalian cells primarily via interactions with GC box elements (5, 13, 21, 24). The binding of Sp1 to GC boxes is often critical in achieving significant levels of transcription from TATA-less promoters and is intimately involved in the determination of the TSS(s). Our previous study (15) confirmed the role of Sp1 as a transactivator of HERG1a.
For KCNQ1b, there is one putative Sp1 site and one CCAAT box within its core promoter region, and there six more Sp1 sites and four more CCAAT boxes distal to the TSS within the 3-kb frame of the 5′a-flanking region. Protein binding to CCAAT boxes can activate gene transcription (12, 19, 35). It is possible that the Sp1 and CCAAT within the promoter region cooperate to transactivate KCNQ1b and the distal Sp1 and CCAAT sites act as enhancers of KCNQ1b transcription. Our data therefore suggest that HERG1 and KCNQ1 genes are controlled by Sp1 for their transcription activation. Alternatively, the cluster of GAGA boxes, which exist only in KCNQ1b but not in KCNQ1a and HERG1a/HERG1b, may play a role in a positively cooperative way to enhance the activation, as it has been documented that the GAGA factor acts as a transcriptional coactivator/potentiator of Sp1 (3).
Another property of HERG1a, HERG1b, and KCNQ1a promoters is the presence of CpG islands (clusters of CpG dinucleotides) near their TSSs. CpG islands often reside in the promoter region of genes, and the methylation of CpGs in these regions is thought to affect the expression of their downstream genes (28, 36). Whether the CpG islands play a role in regulating the activities of HERG1a, HERG1b, and KCNQ1a promoters is worthy of detailed studies considering that the expression levels of these genes can have a great impact on cardiac repolarization and thereby the likelihood of arrhythmias.
Expression profiles of HERG1 and KCNQ1 genes.
Consistent with the general lack of cis elements for cardiac-specific transcription factors in the core promoter regions of HERG1a/HERG1b and KCNQ1a/KCNQ1b genes, the transcripts demonstrated widespread distribution across a variety of human tissues. This could be explained by the presence of multiple sites for ubiquitous Sp1 or of CCAAT boxes. Prominently, the ion channel gene promoters that have been identified to date, regardless of the animal species from which they are from, share some general structural features common to housekeeping-type promoters. That is, the promoters lack consensus TATA boxes and are GC rich with multiple putative Sp1 consensus elements. Our data, however, cannot be used for comparisons of expression levels of HERG1a/HERG1b and KCNQ1a/KCNQ1b transcripts among the various tissues because the mRNA samples were from different individuals of varying ages and both genders.
Our study revealed regional differences or interchamber gradients of expression of HERG1 and KCNQ1 genes. Strikingly, all four isoforms demonstrated the same interventricular and interatrial gradients: RV > LV and RA > LA. These findings provide an explanation for the known interventricular gradients of IKs and IKr and cardiac repolarization as well (22, 29). It has been well documented that the current densities of both IKs and IKr are larger in the RV than in the LV, and, accordingly, action potential duration is generally shorter in the RV than in the LV. This difference may be of physiological importance against arrhythmogenesis under normal conditions. However, when exacerbated under abnormal or pathological conditions, the repolarization heterogeneity becomes a substrate or prerequisite for arrhythmias to occur and sustain (such as torsade de pointes). Our previous study (18) revealed that the Sp1 level is higher in the RV than in the LV. This may account for the interventricular gradients of HERG1 and KCNQ1 subunits. At this stage, the functional consequences of the observed differential transcript expression patterns are not clear. Future studies are required to clarify whether the interventricular inhomogeneity of expressions of these genes contributes to arrhythmogenesis.
It should be noted that the roles of HERG1b and KCNQ1b in forming IKr and IKs and their exact relationships to HERG1a and KCNQ1a, respectively, are to be better defined. It has been claimed that cardiac IKr channels minimally comprise HERG1a and HERG1b subunits, implying that HERG1b coassembles with HERG1a to form complete HERG channels (11, 16). Based on the available data, KCNQ1b is a negative dominant isoform that coassembles with KCNQ1a to dampen the function of the latter (9, 10, 20). Based on these available data, the present study would suggest that the same mechanisms of transactivation between HERG1a and HERG1b should produce a positively cooperative effect on IKr function, whereas the same mechanisms of transactivation between KCNQ1a and KCNQ1b should produce a negative feedback effect on IKs function, at the gene transcription level. However, it should be noted that the relationships between these transcripts, the expression levels of the encoded proteins, and the generation of cardiac IKr/IKs channels remain to be determined.
In summary, we identified the core promoter regions and TSSs of HERG1a, HERG1b, KCNQ1a, and KCNQ1b genes. We compared the relative expression patterns of these genes across a variety of human tissues and demonstrated, unexpectedly, widespread tissue distributions of all the transcripts in addition to the heart, where the genes were originally cloned. Our data further revealed that the mRNA levels of all HERG1 and KCNQ1 isoforms are asymmetrically distributed within the heart, being predominant in the right chambers relative to the left ones. This finding may help us understand the molecular mechanisms for arrhythmias since heterogeneity of ion channel activities is an important substrate for arrhythmogenesis. Our study therefore provides highly valuable knowledge of core elements related to transcriptional regulation and identifies targets for studies of genetic variants in diseases associated with HERG1 and KCNQ1 genes.
This work was supported in part by the Canadian Institutes of Health Research and Fonds de la Recherche de l'Institut de Cardiologie de Montreal (to Z. Wang). Z. Wang is a senior research scholar of the Fonds de Recherche en Sante de Quebec and a Longjiang Scholar Professor of Heilongjiang province, China.
The authors thank XiaoFan Yang for excellent technical support.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2008 by the American Physiological Society