Figures
Abstract
Human dopamine transporter gene (DAT1 or SLC6A3) has been associated with various brain-related diseases and behavioral traits and, as such, has been investigated intensely in experimental- and clinical-settings. However, the abundance of research data has not clarified the biological mechanism of DAT regulation; similarly, studies of DAT genotype-phenotype associations yielded inconsistent results. Hence, our understanding of the control of the DAT protein product is incomplete; having this knowledge is critical, since DAT plays the major role in the brain's dopaminergic circuitry. Accordingly, we reevaluated the genomic attributes of the SLC6A3 gene that might confer sensitivity to regulation, hypothesizing that its unique genomic characteristics might facilitate highly dynamic, region-specific DAT expression, so enabling multiple regulatory modes. Our comprehensive bioinformatic analyzes revealed very distinctive genomic characteristics of the SLC6A3, including high inter-individual variability of its sequence (897 SNPs, about 90 repeats and several CNVs spell out all abbreviations in abstract) and pronounced sensitivity to regulation by epigenetic mechanisms, as evident from the GC-bias composition (0.55) of the SLC6A3, and numerous intragenic CpG islands (27 CGIs). We propose that this unique combination of the genomic features and the regulatory attributes enables the differential expression of the DAT1 gene and fulfills seemingly contradictory demands to its regulation; that is, robustness of region-specific expression and functional dynamics.
Citation: Shumay E, Fowler JS, Volkow ND (2010) Genomic Features of the Human Dopamine Transporter Gene and Its Potential Epigenetic States: Implications for Phenotypic Diversity. PLoS ONE 5(6): e11067. https://doi.org/10.1371/journal.pone.0011067
Editor: Axel Imhof, Ludwig-Maximilians-Universität München, Germany
Received: January 25, 2010; Accepted: May 18, 2010; Published: June 10, 2010
Copyright: © 2010 Shumay et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This work was performed at Brookhaven National Laboratory with infrastructure support from the Department of Energy, Office of Biological and Environmental Research and funded by the National Institute on Drug Abuse, grants KO1 DA025280-01A1 (ES), K05 DA20001 (JSF) and NIH Intramural Research Program, NIDA (NDV). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Dopamine (DA) neurotransmission underlies core brain functions, including locomotion, behavior, cognition and motivation; consequently, disruption in dopamine signalling gives rise to various neuropsychiatric disorders and conditions i.e., Parkinson's disease, schizophrenia, attention deficit/hyperactivity disorder (ADHD) and addiction [1]. A key player in regulation of DA signalling is the dopamine transporter (DAT), it modulates the dynamics and the levels of DA in the synaptic cleft by recycling extracellular DA back into the presynaptic terminal. Alterations in the DAT availability in the brain directly affects the concentration of synaptic DA and the kinetics of its reuptake [2].
DAT expressed in the brain in a region-specific manner, most abundantly in the striatum (the mesostriatal DA pathway), where its concentration is more than ten fold higher than in the frontal regions (Figure 1A and [3]). Even in the same brain region, DAT expression significantly differs among individuals, reflecting high phenotypic heterogeneity (Figure 1B shows actual measures of the amount of the DAT1 mRNAs in the midbrain; see also [4]). DAT expression is highly dynamic: the brain's DAT level adjusts to accommodate DA signaling, decreasing when DA release is low and increasing when DA release is high [5]. They change markedly in response to drugs [6], salient stimuli [6], environmental factors [7], and pathogens [8]. DAT expression alters throughout the life span; first detectable in developing brain relatively late, starting from 32 gestation weeks, the DAT level reaches maximum during adolescence, and then gradually decreases at an estimated rate of 5–6% per decade [9].
A - DAT distribution in the brain. Averaged DVR [11C]cocaine image at the level of the striatum and ROI measures of DAT availability in caudate and putamen. Values represent means±standard deviations across 14 healthy men. B - DAT availability in the striatum of healthy controls of different ages. Measures were obtained with PET and [11C]cocaine and quantified as Bmax/Kd (using Logan plots for irreversible systems). Note the range of variation in DAT availability across individuals.
The possibility that the individual variations of basal DAT levels in the brain are due to genetic differences focused much attention on the DAT1 gene (SLC6A3). It has been rigorously investigated in numerous clinical studies that aimed at revealing an association between the DAT1 genotype and the risks for various brain-related diseases (reviewed in [7], [10]). By and large, these attempts had narrow scope: most focused on a single DAT1 polymorphism, the functional 3′-UTR VNTR (variable number of tandem repeats) [11]. This polymorphism was associated with numerous neurological- and psychiatric disorders, including ADHD [7], cocaine-induced paranoia [12], tobacco smoking [13], and alcohol dependence [14], however, the findings have not been consistently replicated and most of the studies reported no effect of this polymorphism on neurophysiologic and psychiatric measures. Contribution of the 3′-UTR VNTR to differences in density and affinity of DAT was also investigated by several imaging studies; several studies provided some evidence of support [15], though the others did not [16].
There is increasing recognition of the role of environment in the etiology of neuropsychiatric disorders; in most instances such disorders arise through the interactions of genetic and environmental factors (reviewed in [17]). The environment is a potent genetic modifier, influencing gene expression via epigenetic mechanisms. Ubiquitous epigenetic mechanisms represent an essential element of normal development and maturation. Aberrant epigenetic processes can cause maladaptive changes, (gene dys-regulations and dys- functions) and consequently lead to disease. In the brain, in contrast to somatic tissues, epigenetic processes remain active throughout the lifespan: they ultimately are involved in the maintaining brain functions, enabling adaptive plasticity and the ability to accommodate varying environmental challenges [18]. The fidelity of epigenetic processes is critical for the human brain since its development creates enormously complex biological patterning. Consequently, it is most susceptible to aberrant activity of epigenetic modifiers: epigenetic dysregulation is implicated in pathogenesis of a variety of brain-related diseases, including mental retardation and complex psychiatric disorders [19].
The intricate pattern of DAT regulation points to its control by epigenetic mechanisms. Indeed, experimental and clinical studies support the concept of DAT gene-environment interaction. Data from animal models demonstrated that environmental insults during embryonic stage trigger changes in DAT level and delay the development of plasticity in the mesolimbic dopamine system [20] while in humans, early developmental exposure (in utero exposure to tobacco smoke) contributes to the risks and the pathogenesis of ADHD [7].
Currently many research groups are intent on unraveling the mechanisms regulating DAT availability in the brain and on identification of the genetic determinants that are involved in the SLC6A3 control. Disparities in the results of traditional genetic studies of the DAT1 genotype-phenotype relationship and evidence for gene-environment interaction call for a novel approach to studying the SLC6A3 gene variations accounting for epigenetic mechanisms. We concentrated on developing a novel regulatory paradigm for the SLC6A3, wherein genetic- and epigenetic features are interconnected. Considering the facts about the human DAT1 expression (summarized above), we hypothesize that it has unique genomic features that provide a basis for variable epigenetic sensitivity. Under this assumption, the individual variations in the DAT1 gene sequence can directly affect its epigenetic potential, so that we might construe large inter-individual differences in brain's DAT level as phenotypic manifestations of the combinatorial interplay of the DAT1 haplotype and epigenetic marks that jointly regulate the DAT. Empirical evidence supports our hypothesis; thus, the study of the effect of developmental exposure to tobacco smoke clearly demonstrated the risks of developing ADHD and the severity of the disease depends on the DAT1 genotype [7]. Likewise, the DAT1 genotype moderates the correlation between the risk for ADHD and early childhood adversity (institutional deprivation) [21].
To elaborate our hypothesis and to gain supporting theoretical evidence, we undertook a bioinformatic analysis of the SLC6A3 locus using various genomic databases and computational analytical tools available via public domains (Table 1). Knowing that different software used to analyze same functional category are based on different algorithms and used different set of assumptions, and that random selection of one computational approach over another could bias results and introduce errors of interpretation, whenever possible, we applied several non-redundant programs in the same analysis. We selected the programs best suited to perform tasks and then ran them and compared their outputs to compile exhaustive results. If, however, the outputs obtained by different application were similar results, we choose a representative one to report in the manuscript. Our approach was to compare the serotonin transporter gene (SERT or SLC6A4). The SLC6A4 gene belongs to the same family of monoamine transporters and exhibits a high degree of sequence similarity with the SLC6A3 (BLAST E-value = 2e-166); nevertheless, these genes have strikingly different expression patterns suggesting differences in their regulation. In addition, we considered that our comparative analysis would clarify whether the genomic features of the SLC6A3 gene are truly unique or shared with phylogenetically related genes.
We first scrutinized in detail the gene sequences to identify various genomic characteristics may factor into expressional diversity. Next, we looked at different arms of gene regulatory network individually to identify which are involved in the SLC6A3 regulation (Figure 2). Subsequently, we assessed a potential impact of the individual variations in the gene sequence on the gene's sensitivity to a particular regulatory mode and on its downstream phenotypic manifestations.
Gene expression represents a multistep pathway that involves highly regulated and interconnected events; conceptually, it could be divided on transcription, mRNA processing and RNA turnover. Transcription rate is settled on genetic and epigenetic features of gene sequence; differential mRNA processing includes alternative splicing of transcripts, and RNA turnover rate which is determined by RNA stability depends on the RNA interference and nonsense mediated decay (NMD).
Our findings reported here deepen current understanding of the functional significance of the variations in the DAT1 sequence. We provide evidence that the individual variations on a sequence level ultimately affect DAT's epigenetic sensitivity and the interplay of genetics and epigenetic mechanisms gives rise to wide diversity of dopaminergic phenotype. We identified new genetic targets; should experimental testing validate their biological significance, these targets potentially will have practical implication in advancing the development of treatment strategies and improving reliability of genetic screening of devastating diseases arising from impairments of the dopaminergic circuits.
Results
Conserved elements in the genetic loci
The protein composition of dopamine transporter is highly evolutionally conserved reflecting the fact that dopaminergic neurotransmission is an ancient signaling mechanism. However, conservation on the protein level is not reciprocal with conservation on the genome level. As is evident from the Figure 3A, sequence conservation of the SLC6A3 locus across phylogenies at most is very limited: Sparse conserved elements are confined to coding sequences whilst the regulatory regions, including promoter, introns and 3′-UTR do not exhibit conservation (Vertebrate Multiz Alignment and Conservation, 44 species). Far more pronounced conservation is apparent in the SLC6A4 locus with major conservation spikes corresponding to the coding regions, 5′ regulatory region and 3′-UTR (Figure 3B), suggesting that the other genes of monoamine transporter family are much more conserved. Phylogenetic analysis using other programs produced similar results, thus, our inspection of the entire SLC6A3 and the SLC6A4 loci (ORKA toolkit, PhastCons multi-species analysis) revealed only 3 conserved regions in the SLC6A3 locus (including 5 kb flanks) versus 13 regions in the SLC6A4. Since evolutionary conservation commonly serves as a metric for identifying putative regulatory regions, the results of our analyses indicate that the regulatory mechanisms controlling the human SLC6A3 evolved recently. Strikingly, even core promoter of the human SLC6A3 (Figure 4 A) displays similarity with primate but not with other mammals where low complexity repeat (green) occur only in human genome. The SLC6A4 core promoter is far more phylogenetically conserved (Figure 4 B, blue dotted box). Extended promoter region of the human SLC6A3 is unique (Figure 4 A, blue arrow), while the sequences of extended promoter of the SLC6A4 are similar in human and chimp (Figure 4 B).
Conservation tracks of the cognate genes loci shown in UCSC Genome Browser interface. A new Conservation track includes a 44-way vertebrate alignment; the subtracks display phyloP scores (in blue and red). Red arrow points to the promoter regions. The 3′-end of the genes framed in the green box. Note that the conservation of the regulatory regions across vertebrate species is evident only in the SLC6A4 locus. In contrast, in the SLC6A3 locus the only visibly prominent is the “negative” phyloP track, which score indicates accelerated evolution of the region.
The ECR Browser was used to identify conserved regions within the cis-regulatory sequences of the SLC6A3 and the SLC6A4 genes. In the SLC6A3 (panel A) sequence similarity of core promoter is limited to the Pan troglodyte; a low complexity repeat (indicated by green arrow) that resides in the core promoter of the human gene, lacks in the primate gene. The sequence of extended promoter presumably containing promoter enhancer elements (shown in red) is uniquely human as well (blue arrow). Core promoter of the SLC6A4 is phylogenetically conserved across multiple species (panel B, blue box); its extended promoter is shared between humans and primates.
The poor conservation of the DAT1 locus might signify that the abundant changes in the sequence all are “effectively neutral”, that is, they do not affect fitness [22]. However, this is not the case because we noticed that the surrounding chromosomal regions do not exhibit such conservation dips. Accordingly, the changes in the SLC6A3 sequence are rather locus-specific and can not be explained by isochore structure of the chromosomal region thus suggesting that the other evolutional forces drive this process, a bias gene conversion [23] for example. On the evolutional timescale, it appears that the major changes on the DAT sequence occur after a chromosomal translocation event during which the SLC6A3 locus changed its genomic position from chromosome13 in rodents to chromosome5 in primates. Expectedly, the loss of the synteny of the genomic locus, which entails a very dissimilar chromosomal environment, adds to the phylogenetic divergence of the regulatory mechanisms of the SLC6A3.
Reconstruction of hominid evolutionary history revealed that the ancestral branch leading to primates and humans experienced a fourfold acceleration of segmental duplication accumulation whereas the rate of other types of genomic rearrangements, such as single-base-pair mutations, was far slower [24]. The SLC6A3 locus is characterized by abundance of repeats and occurrence of chromosomal rearrangements (see below). Moreover, much of repeat polymorphisms in the DAT1 sequence, including 3′UTR VNTR, is specific for human lineage and not shared even with primates, suggesting that these variations are positively selected. Interestingly, it is postulated that the sequences that are not shared between human and chimps may be important for human-specific traits [25].
Sequence Variations
Much recent progress defining the molecular basis of phenotypic variability resulted from studying genomic DNA. Since this topic is extremely broad, in this analysis we focus on three major types of sequence variations, that is, single nucleotide polymorphisms, repeats and large-scale copy number variations. The surprising ubiquity of these variations in population greatly increases the spectrum of potential genetic variability, both enhancing the complexity of ‘normal’ genome and highlighting these forms of genetic variation as potential disease mechanisms.
Simple Nucleotide Polymorphisms (SNPs).
Point mutations in the DNA sequence (SNPs) are considered as major source of genome variations and are the best understood mechanism for phenotypic change [26]. Traditionally, genetic studies addressing such polymorphisms focused only on non-synonymous coding SNPs assuming that only these are functionally significant. Recently, it became apparent that so-called synonymous mutations are not neutral [27] but affect gene expression in several different ways. For example, SNPs occurring in 3′-UTR SNPs modulate the translation process by impacting mRNA transport [28], interference with the compartmentalized expression of tissue-specific genes in particular cellular loci [29] and altering mRNA stability [30]. Non-coding SNPs may trigger nonsense-mediated decay [31], inactivate the native splicing site [32], hinder posttranslational mRNA modifications [33]. Synonymous SNPs, while not affecting amino-acid composition of the cognate protein, may lessen accuracy and reduce translation rate [34], thereby leading to aberrant protein folding [35] because all these processes rest upon codon specificity.
Multiple SNPs are mapped to the SLC6A3 locus: The NCBI SNP database (frozen April 7, 2009) lists 897 of them. This number significantly exceeds the number of SNPs in the other genes of the monoamine transporters family; thus the SLC6A4 gene has 441 SNPs and the SLC6A2 (norepinephrine transporter gene) has 514 SNPs (Figure 5A). Not only the numbers, but also the patterns of the SNPs distribution in the SLC6A3 and in the SLC6A4 loci are markedly distinclt: SNPs are found in virtually all SLC6A3 domains (flanking regions, coding sequences, exons and introns), whereas in the SLC6A4, they largely are confined to the UTR region (Figure 5B).
A – Chromosomal maps were produced by the Genome Variation Server in a “summary SNPs display” view (the SLC6A3 map - top panel, and the SLC6A4 map – the lower panel). Single peaks are indicating genomic positions of SNPs where peaks height corresponds to the SNP frequency (filtering – default options, merge option - common samples with combined variations). B - Regional frequency of the SNPs in the SLC6A3 (left) and the SLC6A4 (right).
The thirty-three SNPs mapped to the SLC6A3 3′-UTR potentially can hinder localized translation of the DAT1. We noted (GVS) that the 3′UTR SNPs have low conservation scores (mostly 0.000), indicating an evolutionally recent origin, so they likely are human-specific. Considering that five 3′UTR SNPs are common in human populations, with minor allele frequencies of 16% and higher (Table S1 lists the SNPs), they might add a factor to the variation in DAT1 expressional. Interestingly, two SNPs in the SLC6A4 3′UTR have high conservation score (0.99 and 0.98), suggesting that they occur in different species.
The 5′ gene region is pivotal for regulation, it encompasses binding sites for transcription factors and components of basic transcriptional machinery; consequently, it is unsurprising that 5′UTR often is highly conserved and rarely carries variations. We found that the 5′ regions of the SLC6A3 and the SLC6A4 genes have 31 and 12 SNPs, respectively, suggesting that they might confer expressional variability.
Below, we elaborate on possible functional consequences of SNPs in regards to their potential impact on particular regulatory mechanisms.
DNA repeats.
Repeated DNA sequences are common feature of the human genome (209,214 loci, according to [36]). Organized in tandem DNA repeats (tandem repeats) are inherently unstable, and known to be the “hotspots” for mutations [37]. However, their abundance in the human genome posits that they may not be strictly deleterious but might confer some evolutionary advantage, such as affording transcriptional divergence, for example [38].
The 3′ UTR VNTR of the SLC6A3 is the most studied polymorphism in this gene and has been linked to variety of clinical conditions [39]. The VNTR in the intron8 of the DAT1 was also explored in clinical- and preclinical settings and was linked to cocaine abuse [40] and smoking cessation [41]. The high SNPs frequency in the SLC6A3 gene led us search for possible repeat variations in its sequence. We analyzed the current version of the SLC6A3 sequence (Human 2009 Assembly (hg19) running several ab initio prediction programs. We repeated the same analysis with different software and found that even thought the outputs were different in actual numbers of the repeats, the results were consistent. The Tandem Repeats Finder (TRF 4.04) [42] that uses conservative approach, detected 71 repeats, the e-tandem identified 30 regions and the TandemSWAN [43], designed to identify imperfect or “fuzzy” repeats generated more than 600 hits.
Intrigued by the high number of repeats, we wanted to define some principal parameters associated with tandem repeat polymorphism in order to predict their variability (or hypervariability). G + C content of the sequence is a strong predictor of minisatellite polymorphisms [44], and, since the SLC6A3 sequence composition is CG bias (0.55, geecee of Mobyle Portal) it is likely that at least some of newly predicted repeats are polymorphic. The SERV software [45] assessed the polymorphic potential of the predicted repeats by assigning the score to the repeat region (VARscore). The score value between 1 and 3 was proposed as an indicator of potentially good genotyping target, and it was found that the VARscore and experimentally determined polymorphisms are in good correlation [45]. Seven of the SLC6A3 repeats had the optimal score thus representing potential new genotyping targets (Table S2). We noted that the VARscore 2.31 was assigned to the 3′UTR VNTR region; this fact strengthens the validity of our computation. It is noteworthy that the sequences of the longer tandem repeats in the SLC6A3 locus often itself contain simple repeats. Simple repeats, especially trinucleotide repeats such as CAG, CTG, CGG, are the most unstable and often are associated with human diseases.
We used the same methodology to analyze the SLC6A4 gene: in this sequence the SERV detected 40 repeats, the Mobile portal found 11 and the TRF found 30 repeats. In contrast to the SLC6A3, the repeats in the SLC6A4 have shorter consensuses overall and only three repeats have a period longer than 30 bp (Table S3). For the potentially novel genotyping targets, we might consider only one region (VARscore 1.36). Of notice, analyses were limited to the gene sequences; therefore the region of the 5-HTTLPR polymorphism was outside the scope.
Copy number variations (CNVs).
Copy number variations refer to the incidence of additional or missing chromosomal segments in some individuals, adding another dimension of variation to the genome [46]. This genomic feature is highly prevalent: about 13% of human genes overlap with CNVs [47]. CNVs are significantly overrepresented in telomere-proximal regions and in simple tandem repeat sequences; they often detected in genes that have significantly elevated nucleotide substitution rates. The non-uniform distribution of CNVs in the human genome suggests that they have functional consequences; according to Fraser and others, [48], CNV's contribution to genetic diversity might exceed the combined contributions of all other genetic variations. As thought, the CNVs might help to explain some cases of diseases with evident genetic causation but lack of a genetic clarification [49].
Multiple segmental duplication in human genome are recent and a large fraction of these regions is not duplicated in the chimpanzee (Pan troglodytes) genome [50]. Of direct relevance to the subject of our analysis is the fact that the localization of these human-specific duplications is bias for chromosomes 5 and 15 [51].
Our inspection of the SLC6A3 locus (DGV) revealed three structural variants mapped within the gene body (Figure S 1. Note that the largest CNV (36355) involves two exons – exon 5 and exon 6). However, that the actual number of the CNVs in the SLC6A3 locus might be much larger; indeed, the annotations to the latest release of Human Genome Assembly (hg19, 2009, UCSC) include references to the multiple deletions of different sizes, including a deletion of 489 nucleotides (dbSNP build 130 rs71309291). Considering that CNVs data are obtained by analyzing genomic DNA from healthy individuals [46], the CNVs detected in the DAT1 sequence have no apparent phenotypic consequences or have only subtle effects. Alternatively, these findings might be limited to a particular subpopulation of the blood cells (mosaicism) that arise due to the inherent genomic instability of the SLC6A3 locus. The high rate of the CNV occurrence in the DAT1 locus (three out of eight individual genomes tested revealed CNVs in the DAT1 gene), indicates that the copy number variations encompassing the DAT1 locus are frequent in the normal population and, therefore, may be advantageous. Lack of CNVs in the SLC6A4 and the SLC6A2 loci suggests that, in contrast to the SLC6A3, they are controlled by purifying selection.
Transcriptional regulation
Inspection of the Promoter region.
Our next step was to explore the promoter region of the SLC6A3. Experimental and clinical evidence indicates that sequence variations upstream of the transcription start site (TSS) affect DAT1 transcriptional regulation [52]; therefore, we evaluated the promoter region in greater details. A promoter sequence represents a template recognized by the transcriptional machinery; therefore promoter analysis is essential for understanding transcriptional regulation of the cognate gene. It is difficult to accurate locate the position of the TSS and to exactly specify the promoter region because the sequences of the human promoters have exceedingly diverse characteristics.
Since the SLC6A3 core promoter lacks “TATA” and “CAT” boxes (these DNA sequences provide docking sites for basal transcriptional complex) [53], we must identify other regulatory sequence elements that could substituting for them. Transcriptional initiation of neuronal genes often is facilitated via binding of core transcriptional machinery to the CCAAT element [54]; not unexpectedly, we detected several CCAAT consensuses upstream and downstream of the SLC6A3 TSS suggesting that the SLC6A3 might be one of these. Human genes that have CCAAT-promoters display several common characteristics; in general, they are less precise in terms of TSS then the genes with TATA-promoters and they mostly overlap with CpG islands. Transcription initiation from the CCAAT box involves NF-Y, an element with histone-like features, and a particular subset of transcription factors [55]. Both features are relevant to the epigenetic sensitivity of the SLC6A3 as we elaborate later.
Comparing the regulatory regions of the SLC6A3 and the SLC6A4 revealed their markedly different regulatory potential. While regulatory region of the SLC6A4 displays characteristics of a strong promoter (high significance of predicted TSS), the 5′ region of the SLC6A3 does not display the typical sequence characteristics of a strong promoter. Here, de novo predicted TSSs are only marginal (Promoter 2.0 Prediction server) or off-target (Softberry). The software Eponine (Sanger) predicts multiple low-score (ambiquious) TSS sites are predicted by the Eponine, the CorePromoter [56] denotes ten putative TSSs with comparable low scores and the AceView [57] predicts three putative alternative promoters for the SLC6A3 gene.
The complex structure of human gene promoters with a range of alternative transcription start sites (TSSs) [58] supports differential temporal- and spatial- patterns of gene expression and provide an additional level of gene regulation by modulating translational efficiency [59]. About one fifth of human genes have alternative promoters; this phenomena is most frequent in brain-related genes [60]. Based on our analysis we might categorize the SLC6A3 promoter as a broad one; in addition, absence of a strong promoter indicates the potential for off-target transcription initiation.
The heterogeneous dopaminergic neurons within the mesocorticolimbic dopamine system have distinct anatomical and physiological properties [61], therefore, the biological identity of a specific subpopulation of dopaminergic neurons might be enabled by preferential transcription of particular mRNA. Our computational prediction that the SLC6A3 transcription might lead to production of different mRNA isoforms is strengthened by in vivo-derived evidence: the SLC6A3 transcripts of different length were detected in different brain regions (long transcript M95167 in the brain stem [62], and short transcript, L24178, in Substantia Nigra [63]), suggesting that alternative transcription initiation might enable region-specific expression of the SLC6A3. Furthermore, differential transcription initiation might be a means for a rapid switch in response to the acute environmental cues: Prefolded RNA structures are less efficient in adopting environmental changes into expressional changes [64], [65], which might be explained as reflecting the high thermodynamic barriers to interconversion between the secondary structure isoforms.
Whether the SLC6A4 regulation involves similar regulatory mode remains unclear since almost all RefSeq annotated human SLC6A4 transcripts are of somatic origin (placenta, lung) and of similar length. However, it is likely that it, indeed, utilizes AS, given that the rat Sert gene has two alternative promoters and its transcripts are alternatively spliced in a tissue-specific manner [66].
Transcription factor binding sites (TFBSs).
The interaction of the cis-regulatory elements of a gene with transcription factors (TF) largely determined transcription event, therefore, an assessment of the putative TF binding sites in the regulatory region of the gene under analysis yields important information on this gene's regulation. In the nervous system, TFs define the basic framework [67]; their availability varies across the brain regions and cell types thus contributing to the phenotypic diversity [67];. As we mentioned, the nature of the SLC6A3 promoter (CCAAT promoter) implies its sensitivity to selective TFs, because NF-Y synergistically interacts with a subset of TFs [55]. Our inspection of the SLC6A3 5′-flanking sequence (−2 kb, JASPAR database) revealed that it contains binding sites for Sp1, GATA-1, CREB, and c-Myc cis-acting regulatory elements –all those TFs interact with NF-Y [55]. The Sp1-sensitivity of the DAT transcription was previously demonstrated in vitro [68], thus validating the biological significance of the sites we detected. Sp1 is a key element of non-classical genomic nuclear receptor pathways mediating hormone-dependent gene activation and repression suggesting the possibility of the involvement of a hormonal element in DAT regulation.
The nucleotide sequences recognized by individual TFs are too small and degenerate to enable detection of functional enhancer and silencer elements within the genomic context. Therefore, it was postulated that dense clustering of the motifs may improve the accuracy of their recognition and enable better control of the gene expression level [69]. Using the Cluster-Buster web server [69] to query promoter sequences of the SLC6A3 and SLC6A4 genes, we found a high-scoring (9.15) cluster in the SLC6A4 but not in the SLC6A3. Next, we compared the TFs profiles of both genes (CisRED [70]) noting their distinctions (Figure S2). While the lengths of the cisRED regions are comparable, they differ in their architecture: The atomic TF motifs are densely clustered in the vicinity of the SLC6A4 TSS, but and rather scattered in the SLC6A3 regulatory region. Collectively, these results suggest that TFs-mediated transcription more important for the SLC6A4 regulation than for that of the SLC6A3.
Quadruplex forming G-rich sequences in pre-mRNAs and mRNAs (QGRS).
The unusually frequent occurrence of CG nucleotides in the SLC6A3 sequence (CG content 0.55 compare to 0.47 in the SLC6A4 sequence) likely influences the physical properties of the DNA molecule, since guanine-rich DNA tends to adopt a four-stranded structural form known as a G-quadruplex, or G4 DNA [71]. QGRS are highly conservative and often found in key chromosomal regions, suggesting that they have functional significance. When located in the promoter region, G-quadruplex structures can act as cis-regulatory elements and either enhance transcription of, or silence the respective gene. Apparently, natural selection progressively favored the canonical G4 DNA motifs within regulatory regions and, consequently, G4-formation has been proposed as a mechanism for regulation of gene expression in eukaryotes [71]. CG-rich mRNAs also form highly stable G-quadruplex structures [72], which play major role in mRNA turnover [73] and translation [74].
To gain an insight on the involvement of G4 DNA formations in the regulation of the SLC6A3, we analyzed the landscape of G4 DNA (PG4Ms) motifs in the human SLC6A3 gene using the GRS database [73] with the default settings and filters of the server. Unexpectedly, we found that the multiple G4 motifs that are not restricted to the TSS-proximal regions. The number of non-overlapping QGRS in the SLC6A3 sequence (474) almost doubles the number of G4s in the SLC6A4 and the SLC6A2 sequences (203 and 281, respectively). This high score of detected QGRSs increases the likelihood that G-quadruplex structures form in vivo, thus, adding another factor to the SLC6A3 regulation. Importantly, several QGRS were detected in the vicinity of the intron-exon junctions of the SLC6A3: Because G-rich quadruplexes are actively involved in alternative and tissue-specific regulated splicing events [75], their positional effect on gene expression and the implication for DAT function could be most profound.
Post-transcriptional regulation
Post-transcriptional level of gene regulation comprises a series of post-transcriptional and -translational mechanisms that increases the complement of neuronal transcripts and protein isoforms [76]. Neuronal mRNAs are diversified through several processes including alternative splicing, compartmentalized translation, and RNA editing. In most tissues, transcription-generated RNA base sequences experience downstream alterations; this ubiquitous process is particularly active in the nervous system [77].
Short interspersed repeats.
One of the mechanisms of gene regulation involves controlled transport of the mRNA to the subcellular compartment where the gene's protein products are used. The half-life of a particular mRNA depends in part on the sequence of the 3′ UTR of the transcript such as clusters of short interspersed repeats (4–7 bp) that can regulate alternative splicing and mRNA stability [78]. We identified more that fifty short repeated units in the 3′- UTR of the SLC6A3 (about 2 kb) (REPFIND [79]) with a P-value less than 1.11e-16, signifying that they are not random but likely to be of functional significance (Figure S 3A). A similar analysis of 3′-UTR of the SLC6A4 revealed only ten short repeated units, with weaker significance where the smallest of P- values, assigned to one of the repeats, was 7.78e-06. We noted that short interspersed repeats are scattered across the 3′ region of the SLC6A4, whereas in the SLC6A3 gene, they are largely confined to the middle of the 3′-UTR. This is an important fact, since this positional overlap suggests that the length of the polymorphism directly affects the number of the short interspersed repeats.
In neurons, effective mRNA translation occurs near dendritic terminals [80], [81]. The binding of specific proteins to cognate mRNA sequences and formation of translocationally-competent nucleoprotein complexes enables the intracellular trafficking of mRNA molecules. The proteins of translocation machinery recognize CAG or CAG-containing repeated motifs in the 3′-UTRs of the mRNA molecule; such short motifs are commonly detected in genes that are translated in dendrites [79], [82]. Assuming that the SLC6A3 is one of such genes, and its translation likely occurs near synaptic terminals, we screened its 3′ UTR for CAG motifs; as expected, we found multiple such tri-nucleotides repeats and repeats clusters (Figure S3B). Surprisingly, we did not detect CAG motifs in the 3′ UTR of the SLC6A4. The basis for this difference is unclear, since in applying the same reasoning, SLC6A4 translation in neurons is likely to take place near synaptic terminals. Possibly, the lack of a translocation signal is due to the abundant expression of the SLC6A4 in non-neuronal cells. Alternatively, the intracellular translocation machinery might recognize the sites within the coding sequence of the SLC6A4 mRNAs, as was established for the dendritic trafficking of the BDNF transcripts [83]. Notably, the short interspersed repeats and the 3′-UTR VNTR polymorphism of the SLC6A3 reside in the same region, suggesting the complexity of possible biological effects of sequence variations occurring therein. We will discuss this issue in more detail later.
Splicing and Alternative splicing.
Because pre-messenger RNA of contain multiple introns, the splicing is a critical step in its expression. Alternative splicing (AS) generates multiple splicing isoforms with different combination of exons thus dramatically increasing the molecular complexity of the expressed proteins [84]. The AS rate progressively increased during evolution: in yeast only a small portion of genes is alternatively spliced, while in the human genome, the majority of protein-coding genes are known to undergo alternative splicing [85], [86], [87]. The production of differential isoform varies across tissues and cells: the rate of AS events and the level of alternatively spliced mRNA in the brain are much higher than in somatic tissues [88], [89].
We searched for evidence supporting the involvement of AS in the SLC6A3 regulation. Both, the SLC6A3 and the SLC6A4 genes have 15 exons, viz., more than the average of 9 exons for a human gene; and genes with higher number of introns and exons more often utilize AS. We also found a) there are at least three different mRNAs transcribed from this locus (AceView); b) the SLC6A3-mRNA L241781 encodes 11 exons (exon skipping, SliceInfo); and, c) two plausible intron/exon structures for the SLC6A3 mRNA containing 15 and 16 exons (SPLICY database) as evident from microarray experiment data (probeset annotations). In addition, an ambiguity in position mapping for the first exon of the SLC6A3 (First EF) and the number of ab initio predicted exons (26, as by the ExonScan) corroborate the possibility of AS of the SLC6A3 transcripts.
The splicing pattern might be dynamically regulated by external stimuli; in fact, most cellular signaling pathways that change gene transcription also alter groups of alternative exons [90]. Neuronal activity and drugs can induce alternative pre-mRNA splicing [77], [91], [92]. For example, cocaine and dopamine trigger the synthesis of the ania-6 (cyclins protein family) in the brain, whereby, depending on the nature of the stimulus, AS generates two distinct ania-6 mRNAs that then are translated to functionally different proteins [93].
Fox-1 and Fox-2 proteins were identified as regulators of the neuron-specific splicing pattern [94] [95]. They contain a single highly conserved RNA-binding domain acting as a splicing enhancer in the intron downstream from many exons, and might induce exon skipping [95]. Fox proteins mediate splicing of many genes critical for neuronal development and brain function [96]. We found multiple high-scoring binding sites for Fox2 family proteins on the SLC6A3 and SLC6A4 transcripts suggesting they may represent an AS regulatory element.
Splicing machinery recognizes flanking sequences around the exon-intron junctions (∼20 bp); they are pivotal for the fidelity of the splicing process. Mutations within the splice sites of their auxiliary elements can either affect the splice site's efficiency abolishing splicing or alter the balance of the alternatively spliced forms, so modulating the phenotypes of the cognate gene's [97], [98]. The functional consequences of SNPs in the exon-intron junctions are as profound as those for SNPs in the coding regions [99], [100]. Splicing errors are implicated in pathogenesis of various diseases, including neurological diseases and malfunctions [101]. Strict selection control of exonic splicing elements explains the rarity of SNPs near intron-exon boundaries [102]. Contrary to our expectation, we found twenty- five DAT''s SNPs and eleven SERT's SNPs in the exon-intron junctions (Galaxy Browser). Apparently, these SNPs increase a probability of splicing errors, thereby engendering aberrant transcripts.
MicroRNA (miRNA).
A miRNA-mediated mechanism that destabilizes transcripts regulates gene expression at the level of translation (reviewed in [103]). MiRNAs play a primary role in controlling gene expression during development but also in post-mitotic cells, such as neurons, and apparently contribute to age-dependent variation in gene expression [104], [105]. While repressing target genes in actively proliferating cells, MiRNAs can activate gene expression in non-dividing cells such as neurons [106]. The sites of miRNA recognition often are found in the 3′UTR region [107]. To retrieve information on miRNA binding sites for the SLC6A3 and the SLC6A4 sequences we used the MirBase Target Database software (Welcome Trust), which recognizes 14 and 57 targets, while EBI ASTD, recognizes 107 and 46 targets, respectively. The identification of multiple targets supports the possible involvement of the miRNA machinery in regulating both genes.
Recently, our understanding of the mechanisms of miRNA-mediated gene regulation was broadened due to the discovery that human-specific miRNAs interact with sites in the 5′-UTR motifs [108]. An important sophisticated facet of this work was the finding that the miRNAs of this novel class interact simultaneously with both 3′ and 5′ UTRs targets; this “double” targeting greatly enhances miRNA's ability to down-regulate the gene expression. We found that the SLC6A3 and the SLC6A4 -genes are amongst the targets for this novel class of miRNAs (miBridge, [108]). In particular, the SLC6A3 UTRs contain consensus sequences recognizable by eight miBridge' miRNAs, while the SLC6A4 UTRs contain consensuses for 4 miRNA of this class. The target prediction algorithm implemented in the miBridge is more stringent than those in other software, and consequently, the number of genes predicted as miRNA targets by miBridge is low; however, the multiple identification of the SLC6A3 gene as a target underscores the importance of miRNA-mediated control in its regulation.
Differences in conformational structure of SLC6A3 transcripts and their stability offer a putative mechanism explaining biological significance of the 3′VNTR polymorphism.
The half-life of a particular mRNA depends in part on the sequence of the 3′ UTR that can regulate alternative splicing, the intrinsic stability of a cognate mRNA and its sensitivity to regulation by miRNA [78]. Transcript length, however, also determines the RNA's secondary structure; hence, various SLC6A3 transcripts might fold differently. The conformation of the RNA molecule is essential for its biological function; it determines the translational rate and the biological catalysis of a transcript [109], [110]. We determined that the length of the SLC6A3 mRNAs correlates with the stability of their compacted structures (results in Table 2), suggesting that there is yet another regulatory mechanism for the DAT gene that might operate via preferential transcription of one mRNA over another. More specifically, cells might employ a rapid switch in transcriptional mode to achieve fast increase in DAT level, hence, rapid synthesis of DAT protein; whereas the production of longer, more stable transcripts prolongs translation time determining basal DAT's level.
The 3′UTR VNTR polymorphism of the SLC6A3 apparently changes the transcript length; therefore we sought that these differences might be a factor contributing to the phenotypic differences ascribed to the particular alleles. Following the original discovery of this polymorphism [111], there have been many experimentally and clinical investigations. The finding that the number of the tandem repeats affects the transcriptional efficacy of the DAT was confirmed by different experimental means but failed to provide satisfactory explanation to tshe phenotypic differences observed in vivo. By definition, 3′UTR variation does not affect the protein product; hence, it ought to exert its influence on the level of transcripts. We hypothesized that the actual number of the repeated consensuses in the cognate allele would determine the secondary structure of the corresponding SLC6A3 transcript. We tested our hypothesis by modeling the secondary structures of putative mRNAs containing 9 versus 10 repeats, which correspond to the most common alleles in the population (GeneBee server [112]. We found a notable difference in the free energy of the predicted secondary structures of the respective mRNAs, viz., −135 kcal per mol (10 repeats), versus −95 kcal per mol (9 repeats) (Figure S4A). Further, we noted that the folding structures of mRNAs i.e., the tendency towards hairpin formation also are subordinate to the repeats number: Duplexes formed by the 9-repeat sequences had longer stems and higher energy than those formed by the 10-repeat sequences (−14.4 kcals versus 7.4 kcals) (SECentral Clone Manager Suite 7 program) (Figure S 4B). In sum, our analyses revealed that the number of the 3′ UTR repeats might affect the SLC6A3 expression in several different ways; specifically, by transforming the secondary structure of the cognate transcript and modulating its propensity to form hairpin structures. The conformational differences of the mRNAs (determined by the transcripts lengths) might tangentially make an impact on other regulatory mechanisms, including the number of putative miRNA target sites facing the surface of the molecule could vary, changing the transcript's degradation rate. In a similar fashion, the colocalization of the 3′UTR VNTR region with the region of short interspersed repeats points to their co-regulation.
Epigenetic regulation of the SLC6A3 and the SLC6A4 genes
Several studies have analyzed known polymorphisms in the DAT and the SERT genes as a function of specific environmental factors [113] proving that adverse environments entail disordered behavioral manifestations more often in the carriers of particular polymorphisms. There is a common agreement that the environmental factors play a role in the pathogenesis of brain disorders, but the mechanisms of gene-environment interactions on molecular level remain poorly understood. Our knowledge on the epigenetic status and dynamics of epigenetic processes regulating the SLC6A3 and the SLC6A4 is very limited due to significant technological challenges; therefore, systematic identification of possible targets amenable to epigenetic modifications by computational means intrinsically is worthwhile. Further, since a computational analysis is sequence-based, it is not influenced by the innate variability of experimental conditions.
DNA methylation.
DNA methylation in the mammalian genome predominantly occurs on cytosine in context of the 5′-CpG-3′ dinucleotides; this is the only type of epigenetic modifications to directly change the DNA molecule. Stretches of GC-rich sequences in the genome called CpG islands (CGIs) that are associated with open transcriptionally competent chromatin structure were discovered in gene promoters [114]. The relevance of GC content and CpG dinucleotide concentration to the regulation of gene activity points to its physiological significance. The methylation status of the promoter-overlapping CpG island plays a central role in the epigenetic control of gene expression thus, aberrant methylation patterns (hypo- and hyper-methylation) can engender gene dys-regulation. Conceptually, an alteration in the pattern of DNA methylation might serve as an efficient mechanism for genes to adapt functionality in response to a variable environment and also might compensate for the effects of a mal-functional polymorphism. The DNA methylation pattern established during the development and differentiation is preserved with high fidelity during cell division by DNA methyltransferases (DNMTs). DNMTs are highly expressed in developing tissues but their activity declines during differentiation in all tissues except the brain wherein they are expressed throughout the lifetime in the brain [115], [116]. Dynamic DNMT activity in the brain is essential for synaptic plasticity and memory formation [48], [117].
Because CpGs act as a mutation hotspot (deamination of methylated CpGs to TpGs) and the estimated mutation rates at CpG sites are about 10–50 times higher than other transitional mutations, CpG dinucleotides are under-represented in mammalian genomes where they occur at one-fifth of their expected frequency [118]. The human SLC6A3 gene is remarkably GC dense and has multiple CpG islands: in contrast to most vertebrate genes that have only a promoter-overlapping CpG island, all gene regions of the SLC6A3 display characteristics attributed to this genomic feature. We mapped twenty-seven bona fide CpG islands [119] to the human SLC6A3 locus (Figure 6). Notably, even the “conventional” promoter-overlapping CpG island of the SLC6A3 is unusual: on average, in active promoters of human genes cytosine and guanine account for 57% of the nucleotides [120] GCs represent 79% of the SLC6A3 promoter sequence. The frequency of the intronic CGIs is approximately four-times lower than the expected value [121], yet we found 21 such CGIs in the SLC6A3. The SLC6A4 locus contains six CGIs, which is also above average but lower than that for SLC6A3 (Figure 6, bottom panel).
Visual displays of the bona fide CGIs (CpG islands) are obtained using Custom Tracks option (UCSC Genome Browser). Here, blue boxes represent predicted CGIs based on combined epigenetic score. Note, that in the SLC6A3 locus (top panel), multiple GCIs are distributed within all gene regions whereas in the SLC6A4 locus (lower panel), CGIs are evidently less numerous.
CGIs associated with genes underwent gains and losses during vertebrate evolution [122]. Human genes generally contain more promoter-associated CGIs and have more pronounced CGI characteristics [123] than do mouse genes. Seemingly, the number of CpG islands in the DAT gene dramatically increased during evolution: The mouse Slc6a3 has no CGIs, the rat's gene has one promoter-associated CGI, while the rhesus monkey's DAT gene has two CGIs (one in the promoter and one intragenic). It is tempting to speculate that the increase in the number of CGIs in the DAT gene could be evolution-driven and reflects the growing demands dictated by higher brain function in humans.
Chromatin organization, Nucleosomes.
In vivo, the DNA molecule forms a complex with proteins that allows its packaging into chromatin. Nucleosomes are the structural units of chromatin represented by histone octamers around which the DNA coils. The close interaction of the DNA molecule with a nucleosome core results in a condensed chromatin that is inaccessible to the transcription machinery; hence, the transcriptional activation of a gene requires the local transition of compact chromatin domains into decondensed loops. Nucleosome remodeling and covalent modifications of histones provide the basis for epigenetic gene regulation that occur via the modulation of the accessibility of the genomic loci to transcriptional machinery [124]. Dynamic chromatin- remodeling events largely determine the frequency of promoter activation, thus, the expression of many genes could be mapped to chromatin modifiers, rather than to TFs [125].
CG-rich motifs in DNA sequences inherently disfavor nucleosomes, and are referred to as “nucleosome exclusion sequences” (NX) [126]. We found that both, the SLC6A3 and the SLC6A4 genes have high NX-scoring sequences near the TSS (Figure 7). The predicted nucleosome positioning in the SLC6A3 and the SLC6A4 loci notably differ: Entire SLC6A3 locus comprises of numerous nucleosome-dysfavouring sequences, while in the SLC6A4, NX-peaks are sparse (Figure 7, top and bottom panels). It was suggested that intragenic regions with high NXScores might function as transcriptional enhancers (Sawsan Khuri, personal communication); upon this basis, we interpret the low nucleosome occupancy of intragenic regions of the SL6A3 gene as indicating that they might bear regulatory potential. Indeed, several intragenic NXScore peaks within the SLC6A3, including one in the 3′ UTR, have NXScore value about 600, suggesting that the open chromatin structure in these regions renders them susceptible to targeting by chromatin modifiers and possibly by other DNA interacting proteins.
Nucleosome-exclusion regions (NXRegions) are shown for the SLC6A3 (top panel) and the SLC6A4 (lower panel) genes. The figure was prepared by uploading NXScore results as a Custom Track on the UCSC Genome Browser interface where the snapshots showing the Known Genes track and Custom Track option were taken for the cognate gene's loci. NXScore peaks in the SLC6A4 are largely confined to the TSS vicinity (as it is seen in most of the human genes) and the height of the TSS-overlapping peaks is dominant. Distinctly different pattern of the nucleosomes is predicted for the chromosomal region corresponding to the SLC6A3: here multiple NXScore peaks are densely mapped to the different regions of the gene body and to the flanks. Notice that the height of the TSS-proximal NXScore peak is comparable with the other intragenic peaks.
The same computation assigns the NXScore to the promoter region that can serve as a proxy to the expression rate of a cognate gene. The NXScore value for the SLC6A4 promoter (623) agrees well with the average score value for the promoters of the RefSeq human genes (about 600). Accordingly, the expression of the SERT is expected to be rather broad. The score predicted for the SLC6A3 promoter (820) is significantly higher than the average, thus indicating that expression of the SLC6A3 is not constitutive but contingent upon stimulation.
Histone modifications.
DNA In cell nucleus is found in the form of a nucleoprotein, wherein histones represent the protein moiety. The N-terminal tail of the core histone proteins (H2A, H2B, H3 and H4) is a subject to various types of covalent modification, including methylation, acetylation, phosphorylation, sumoylation, and ubiquitination. Combinatorial arrangement of histone modifications determine the local packaging of the DNA, thereby influencing gene expression [127]. Acetylated histones mark actively transcribed genes. Enzymes with opposite functions, histone acetyltransferases (HATs) and histone deacetylases (HDACs), maintain the highly dynamics of histone acetylation/deacetylation status. Histone modifications are highly dynamic and reversible [128]; they are co-regulated with DNA methylation. For example, DNA methylation depends on the pre-established histone 3 methylation at lysine 9 [129] and aberrant DNA methylation occurs when the local pattern of histone modifications is distorted [130].
In the adult brain, changes in histone acetylation underlay synaptic plasticity [117]. Some drugs, including drugs of abuse, influence histone acetylation status; for example, administratered cocaine causes histone hyperacetylation [131] and, accordingly, clinical studies suggest that cocaine craving diminishes after treatment with valproate [132], [133], [134]. This evidence sheds some light on the pathogenesis of addiction: Since valproate has HDAC inhibiting activity, it is plausible that cocaine-mediated disregulation of gene that causes pathology is mitigated by the treatment. Both direct- and indirect- empirical evidence supports the involvement of epigenetic mechanisms in the SLC6A3 regulation: Thus, it was established that valproic acid, acting via the Sp1 (see above), up-regulates the SLC6A3 expression in a dose- and time-dependent manner [135]. In addition, researches using the ADHD animal model [136] reported that valproate attenuated hyperactive behaviors caused by the incomplete knockout of the Slc6a3 gene. The fact that the SLC6A3 is TATA-less (discussed above) must also be considered since TATA-less genes, generally, are biased towards regulation by HAT activity [125].
The collaborative efforts of many research groups have made experimental data on histone modification available via public domains. The high-throughput technologies, used to facilitate such analyses are very powerful yet suffer from limitations including the potential for inaccurately interpreting the biological significance of the data. Using cell lines as model systems implies that the culturing conditions affect their proliferation rate, transcriptional events and differentiation state. Apparently, the histone modification patterns that were detected in actively dividing cells could be partially related to malignant transformation hence, they can not be generalized to predict chromatin status in vivo, especially when considering terminally differentiated post-mitotic cells such as neurons. However, this information may offer some basic guidelines. We retrieved a compendium of the available information via the UCSC Browser (Figure S 5). Histone modifications mapped to the SLC6A4 locus corroborate well with the anticipated pattern (Figure S 5B): The TSS proximal region (+/− 2 kb) displays a characteristic combination of positive- and negative- histone marks. The SLC6A3 promoter is strikingly different (Figure S 5A); here, the positive histone marks merely are present. It is possible that open chromatin conformation in 5′ to the SLC6A4 TSS reflects its broader expression pattern whilst the lack of acetylated and tri-methylated histones in vicinity of the SLC6A3 TSS suggests that the experimental cell lines do not express DAT.
Role of epigenetic mechanisms in transcription initiation.
Histone modifications of the promoter region reflect the transcriptional state of genomic region, therefore the human core-promoter prediction software CoreBoost_HM integrates data on histone modification profiles with important sequence features. CoreBoost_HM is based on two classifiers, one for CpG related promoters and the other for non-CpG related promoters. We applied both classifiers to analyze the SLC6A3 and SLC6A4 so, to ascertain whether the epigenetic state of the regulatory region is critical for their transcription (Figure S 6A, B). As evident from the Figure, DNA methylation does not contributes equally to regulation of these genes. The SLC6A3 promoter was detectable only when analyzed as CpG-dependent, while a set up for CpG island-independent promoter produced no results. In contrast, for the SLC6A4, both analyses detected a putative promoter in the region corresponding to the actual TSS.
Discussion
The strikingly polymorphic DAT phenotype in humans and the biological mechanisms that connect individual genetic backgrounds with the phenotypes still are understood poorly. The functional demands on the DAT1 regulation are seemingly contradictory since the vital necessity of assuring the robust functioning of the dopaminergic system coincides with the need for plasticity and rapid fluctuations in the transcription rate. To better understand the distinct sources and mechanisms that might facilitate a phenotypic diversity of this magnitude, we comprehensively interrogated the sequence of the SLC6A3 gene applying computational methods.
Our analysis revealed several unique features of the human SLC6A3 gene sequence, including (1) a very high frequency of SNPs (897, as in NCBI SNP) compared with the SLC6A4 (441) and other brain-related genes, such as BDNF and DRD4 (454 and 154 SNPs, respectively), that, depending on their positional effect, engender various biological consequences; (2) the abundance of VNTRs (more than 90 in the SLC6A3 gene body alone), which is indicative of a tendency to open chromatin structure in the locus, and increased accessibility to chromatin modifiers; and (3) presence of intragenic CNVs. Most notable characteristic of the human SLC6A3 is its high sensitivity of to epigenetic regulation: in contrast to the relative enrichment in GC nucleotides in the promoter-proximal region as occurs in most human genes, the entire SLC6A3 locus has GC-bias sequence composition (0.55) and comprises multiple CpG sites comprising 27 bona fide CGIs (CpG islands).
Importantly, our analysis revealed that the distinctive sequence features of the SLC6A3 gene recognized by an array of regulatory mechanisms are evolutionally recent. In fact, they are either uniquely human or shared between human and primates. We noted that the epigenetic sensitivity of the DAT1 gene increased during evolution, and this process likely involves GC-bias gene conversion (GCBC) that is viewed as a major force driving genome evolution [137]. Indeed, whilst the mouse Slc6A3 gene has GC content 0.46 and its promoter has no CpG island, promoter-overlapping CpG island is present in the rat's Slc6a3 gene; furthering this tendency, the rhesus monkey's SLC6A3 also has two intragenic CpG islands. During evolution, the increase in demands on the DAT1 expression might well have driven changes in the gene sequence, expanding its sensitivity to the different regulatory mechanisms. In this view, the GC-density and the number of CGIs in the DAT gene might co-evolve with the growing role of epigenetic mechanisms in its control. In other words, the evolutional drift of the SLC6A3 sequence towards accumulating GC-nucleotides might reflect its enhancing epigenetic potential, needed to accommodate regulatory demands, as dictated by the increasingly complex functions of the human brain.
The genomic regions experiencing GCBC are seen as recombination hotspots [137] and, as such, the inherent instability of the SLC6A3 locus could explain its unique sequence features and regulatory attributes. These changes might be linked in time to the genome rearrangement event that resulted in the translocation of a syntenic block encompassing the DAT gene from the chromosome 13 (rodents) to the telomere-proximal region of the chromosome 5 (human and primates). In general, the propensity of genomic regions to instability is deleterious characteristic, but, theoretically, hypervariability arising from SNPs and tandem repeated sequences could be beneficial by allowing rapid adaptation [120], [138], [139]. Accordingly, the high variability of the DAT locus might support a broad basis for phenotypic diversity, and thus, be advantageous. Our detection of co-occurrence of the SNPs and the repeated sequences with the CGIs (116 SNPs and 2 tandem repeats) infers that the DAT1 might still be evolving.
Our observation that there are several CNVs in the SLC6A3 locus is not unique; they are detected in several human genes [140] that seemingly can asymptomatically tolerate this type of genetic variations. Here the human CNV genes stand apart; evidence from animal studies demonstrated that CNVs are depleted from laboratory mouse strains during selective breeding [141]. Consequently, it was postulated that that natural selection acts discriminately among human CNV genes and thus, at least for a subset of human genes, CNVs might confer adaptive benefits [140].
If propensity of the DAT1 locus to instability is high and can give rise to frequent mutations, there must be other means to compensate for potentially sub-optimal gene function resulting from deleterious changes in its sequence. Considering the main concept of systems biology, that emphasizes the biological networks and hubs rather than a single specific process, we posit that the multi-modality of the DAT expressional control with multiple feed-back control loops ensures its functional robustness. Our analysis suggests that the several arms of gene regulatory network modulate the expression of the SLC6A3, whereby differential engagements and/or combinatorial interplay of different mechanisms of this network, DAT expression can be robust and reliable yet plastic and adjustable to accommodate environmental demands.
The novel regulatory paradigm for the SLC6A3 gene that built upon our results might have direct practical implication. Acknowledging that DAT regulation is far more complex than that of average human gene implies that a comprehensive investigation is needed to reliably predict functional changes associated with the DAT polymorphism(s) and to infer possible phenotypic manifestations of sequence variation(s),. That is, we should consider broadening the scope of the future studies by interrogating the multiple features of the SLC6A3 gene rather than focusing on any single polymorphism in its sequence. Also, the functional- and anatomical- diversity of the brain's dopamine system [61] must be accounted for. We elaborate this point taking the 3′-UTR VNTR polymorphism as an example. This polymorphism was investigated in relation to a variety of behavioral traits and diseases [142], [143], [144], and the findings of those research are conflicting. That is, the 10 repeats-allele (10/10 genotype) in some studies, i.e., [145] was associated with a lower translation rate and increased extracellular dopamine; while in the others, i.e., [146] with higher efficacy of the DAT translation in vitro and in vivo [147]. We suggest that considering a genomic position of the respective polymorphism might resolve this controversy. Indeed, the repeats number (the length of the polymorphic region can affect exclusively the longer mRNAs since the VNTR resides in the middle of 3′-UTR. Evidently, the shorter transcript isoforms (akin to S44625) are insensitive to this variation since here the transcription is terminated upstream of the VNTR. Considering the transcription bias towards preferential production of shorter isoforms in specific brain regions, we expect that a subset of brain functions executed through the specific brain regions is inherently independent of the polymorphism. Conversely, for the behavioral traits that rely on the transcription of long mRNA isoforms, innate differences in transcripts stability and conformational folding, subordinate to the genotype, will be manifest phenotypically.
We want to reiterate that the analyses and observations presented here are theoretical by nature and require experimental validation. However, we believe, that they help broaden our current understanding on the complex regulatory mechanisms imbedded in the human DAT gene sequence.
Methods
The computational tools used to run the analyses and the databases used to retrieve the sequences and to visualize genomic features are listed in Table 1.
Supporting Information
Figure S1.
Copy number variations in the SLC6A3 locus. Information on validated structural variation maps of eight human genomes with resolution at the sequence level for selected regions is integrated in the UCSC Genome Browser. The CNVs in the SLC6A3 genetic locus are visualized in the Browser. The orange bar, indicating the longest deletion, encompasses two exons (indicated by arrows).
https://doi.org/10.1371/journal.pone.0011067.s001
(0.10 MB EPS)
Figure S2.
The SLC6A3 and the SLC6A4 regulatory regions. Red bar in this visual display represents cisRED annotation-based regulatory modules and predicted TFBSs are shown as brown blocks. In the DAT1 (top) red bar is long and predicted TFBSs are scattered over a long sequence stretch, whereas in the SERT (bottom), red bar is composed of discrete units and TFBSs are located within the TSS-proximal bar.
https://doi.org/10.1371/journal.pone.0011067.s002
(0.09 MB EPS)
Figure S3.
Clusters of short interspersed repeats in the DAT1 3′UTR. A - REPFIND program was used to identify clustered, exact repeats within the 5 kb region located on the 3′ - end of the SLC6A3 that includes 3′UTR. The program calculates a P-value indicating the probability of finding such a concentration of that particular repeat by chance for each repeat cluster found, and then selects the cluster with the most significant P-value. Top ten out of generated multiple repeated units with significant P-values multiple repeats found by the program are shown at the panel A. P-value of 8.18e-14 suggests that the closeness of these repeats is not random, and together they form a unit likely to have functional significance. B -CAG or CAG-containing motifs are the most abundant motif in 3′-UTRs of neuronal genes, therefore our identification of multiple CAG motif clusters in the 3′-UTR of the SLC6A3 posits that translation of its mRNAs is likely occur near the synaptic terminals. For this test, we used human 3′-UTR statistical background and P-value cutoff: 0.000001.
https://doi.org/10.1371/journal.pone.0011067.s003
(0.19 MB EPS)
Figure S4.
Secondary structure of the SLC6A3 transcript isoforms. A - The variation in the 3′UTR region of the SLC6A3 results in production of different length mRNAs. We used the GeneBee server to predict the secondary structure of RNAs transcribed from the sequences containing 10 versus 9 repeats which correspond to the most common allelic variations. As evident from the figure, there the presence of the 10th repeated increment profoundly changes the structure of the transcript. B - According to analysis of dyad symmetries (hairpins) predicted for the respective mRNA sequences (the SECentral Clone Manager Suite 7 program), the folding structure and hairpin formation generated by the 3′-UTR sequences containing 9 and 10 repeats are strikingly different. We found that the duplexes in 9 repeat's sequence have longer stems and higher energy than those from the 10 repeats (−14.4 kcals versus 7.4 kcals), suggesting that the 3′-UTR VNTR might influence the SLC6A3 gene expression by modulating the rate of mRNA degradation, which depends on conservation of the hairpin structures formed on mRNA 3′-UTR 's ends.
https://doi.org/10.1371/journal.pone.0011067.s004
(0.10 MB EPS)
Figure S5.
Experimental histone methylation data for the promoter region of the SLC6A3 and SLC6A4 genes. Compendium of these data is integrated in the UCSC genome browser, so, we retrieved the data on “active” histone modification mapped to the promoter regions of two genes. Histone profiling of the SLC6A3 (A) demonstrates lack of the acetylated histones and di-and tri-methylated H3K4 (green arrows), whereas the profile of the SLC6A4 regulatory region unequivocally reflects active status of the chromatin (red arrows) (B).
https://doi.org/10.1371/journal.pone.0011067.s005
(0.17 MB EPS)
Figure S6.
Prediction of the promoter region using CoreBoost_HM. We used two algorithms of the program to assess differences in the promoter prediction that depend on the epigenetic sensitivity. A - For the SLC6A3, density scores calculated for CpG-related promoters (upper panel) display the highest peak overlapping with the putative TSS (red arrow). Calculation under the assumption of CpG-independent promoter fails to produce a peak near the canonical TSS; and the position of the dominant peak is shifted downstream (green arrow). B - similar analyses carried out for the SLC6A4 gene showed that the position of the predicted promoter remains the same when CpG-association is considered (upper panel, red arrow) or discounted (lower panel, red arrow).
https://doi.org/10.1371/journal.pone.0011067.s006
(0.13 MB EPS)
Table S1.
Known SNPs in the SLC6A3 locus.
https://doi.org/10.1371/journal.pone.0011067.s007
(0.20 MB PDF)
Table S2.
Tandem repeats in the SLC6A3 locus with a period longer that 30 nt.
https://doi.org/10.1371/journal.pone.0011067.s008
(0.03 MB PDF)
Table S3.
Tandem repeats in the SLC6A4 locus.
https://doi.org/10.1371/journal.pone.0011067.s009
(0.04 MB PDF)
Acknowledgments
We thank Dr. Jim Swanson and Dr. Ruben Baler for helpful discussions and suggestions on the organization of the manuscript, Dr. Fritz Henn for long-standing support, and Dr. Avril Woodhead for editing.
Author Contributions
Conceived and designed the experiments: ES NDV. Performed the experiments: ES. Analyzed the data: ES. Wrote the paper: ES JSF NDV.
References
- 1. Mehler-Wex C, Riederer P, Gerlach M (2006) Dopaminergic dysbalance in distinct basal ganglia neurocircuits: implications for the pathophysiology of Parkinson's disease, schizophrenia and attention deficit hyperactivity disorder. Neurotox Res 10: 167–179.
- 2. Greenwood TA, Schork NJ, Eskin E, Kelsoe JR (2006) Identification of additional variants within the human dopamine transporter gene provides further evidence for an association with bipolar disorder in two independent samples. Mol Psychiatry 11: 125–133, 115.
- 3. Ciliax BJ, Drash GW, Staley JK, Haber S, Mobley CJ, et al. (1999) Immunocytochemical localization of the dopamine transporter in human brain. J Comp Neurol 409: 38–56.
- 4. Bannon MJ, Whitty CJ (1997) Age-related and regional differences in dopamine transporter mRNA expression in human midbrain. Neurology 48: 969–977.
- 5. Zahniser NR, Sorkin A (2004) Rapid regulation of the dopamine transporter: role in stimulant addiction? Neuropharmacology 47: Suppl 180–91.
- 6. Volkow ND, Wang G-J, Fowler JS, Ding Y-S (2005) Imaging the Effects of Methylphenidate on Brain Dopamine: New Model on Its Therapeutic Actions for Attention-Deficit/Hyperactivity Disorder. Biological Psychiatry 57: 1410–1415.
- 7. Swanson JM, Kinsbourne M, Nigg J, Lanphear B, Stefanatos GA, et al. (2007) Etiologic subtypes of attention-deficit/hyperactivity disorder: brain imaging, molecular genetic and environmental factors and the dopamine hypothesis. Neuropsychol Rev 17: 39–59.
- 8. Wang GJ, Chang L, Volkow ND, Telang F, Logan J, et al. (2004) Decreased brain dopaminergic transporters in HIV-associated dementia patients. Brain 127: 2452–2458.
- 9. Cruz-Muros I, Afonso-Oramas D, Abreu P, Perez-Delgado MM, Rodriguez M, et al. (2009) Aging effects on the dopamine transporter expression and compensatory mechanisms. Neurobiol Aging 30: 973–986.
- 10. Haddley K, Vasiliou AS, Ali FR, Paredes UM, Bubb VJ, et al. (2008) Molecular genetics of monoamine transporters: relevance to brain disorders. Neurochem Res 33: 652–667.
- 11. Vandenbergh DJ, Persico AM, Hawkins AL, Griffin CA, Li X, et al. (1992) Human dopamine transporter gene (DAT1) maps to chromosome 5p15.3 and displays a VNTR. Genomics 14: 1104–1106.
- 12. Greenwood TA, Alexander M, Keck PE, McElroy S, Sadovnick AD, et al. (2002) Segmental linkage disequilibrium within the dopamine transporter gene. Mol Psychiatry 7: 165–173.
- 13. Yang YK, Yao WJ, Yeh TL, Lee IH, Chen PS, et al. (2008) Decreased dopamine transporter availability in male smokers – a dual isotope SPECT study. Prog Neuropsychopharmacol Biol Psychiatry 32: 274–279.
- 14. Samochowiec J, Kucharska-Mazur J, Grzywacz A, Jablonski M, Rommelspacher H, et al. (2006) Family-based and case-control study of DRD2, DAT, 5HTT, COMT genes polymorphisms in alcohol dependence. Neurosci Lett 410: 1–5.
- 15. Heinz A, Goldman D, Jones DW, Palmour R, Hommer D, et al. (2000) Genotype influences in vivo dopamine transporter availability in human striatum. Neuropsychopharmacology 22: 133–139.
- 16. Krause J, Dresel SH, Krause KH, La Fougere C, Zill P, et al. (2006) Striatal dopamine transporter availability and DAT-1 gene in adults with ADHD: no higher DAT availability in patients with homozygosity for the 10-repeat allele. World J Biol Psychiatry 7: 152–157.
- 17. Jaffee SR, Price TS (2007) Gene-environment correlations: a review of the evidence and implications for prevention of mental illness. Mol Psychiatry 12: 432–442.
- 18. McGowan PO, Meaney MJ, Szyf M (2008) Diet and the epigenetic (re)programming of phenotypic differences in behavior. Brain Res 1237: 12–24.
- 19. Cirulli F, Francia N, Berry A, Aloe L, Alleva E, et al. (2009) Early life stress as a risk factor for mental health: role of neurotrophins from rodents to non-human primates. Neurosci Biobehav Rev 33: 573–585.
- 20. Frohna PA, Neal-Beliveau BS, Joyce JN (1997) Delayed plasticity of the mesolimbic dopamine system following neonatal 6-OHDA lesions. Synapse 25: 293–305.
- 21. Stevens SE, Kumsta R, Kreppner JM, Brookes KJ, Rutter M, et al. (2009) Dopamine transporter gene polymorphism moderates the effects of severe deprivation on ADHD symptoms: developmental continuities in gene-environment interplay. Am J Med Genet B Neuropsychiatr Genet 150B: 753–761.
- 22. Duret L, Galtier N (2009) Biased Gene Conversion and the Evolution of Mammalian Genomic Landscapes. Annu Rev Genomics Hum Genet.
- 23. Hurst LD (2009) Evolutionary genomics: A positive becomes a negative. Nature 457: 543–544.
- 24. Marques-Bonet T, Kidd JM, Ventura M, Graves TA, Cheng Z, et al. (2009) A burst of segmental duplications in the genome of the African great ape ancestor. Nature 457: 877–881.
- 25. Kehrer-Sawatzki H, Cooper DN (2007) Understanding the recent evolution of the human genome: insights from human-chimpanzee genome comparisons. Hum Mutat 28: 99–130.
- 26. Stranger BE, Forrest MS, Dunning M, Ingle CE, Beazley C, et al. (2007) Relative impact of nucleotide and copy number variation on gene expression phenotypes. Science 315: 848–853.
- 27. Vandenbergh DJ, Thompson MD, Cook EH, Bendahhou E, Nguyen T, et al. (2000) Human dopamine transporter gene: coding region conservation among normal, Tourette's disorder, alcohol dependence and attention-deficit hyperactivity disorder populations. Mol Psychiatry 5: 283–292.
- 28. Ninio J (2006) Multiple stages in codon-anticodon recognition:double-trigger mechanisms and geometric constraints. Biochimie 88: 963–992.
- 29. Okazaki N, Imai K, Kikuno RF, Misawa K, Kawai M, et al. (2005) Influence of the 3′-UTR-length of mKIAA cDNAs and their Sequence Features to the mRNA Expression Level in the Brain. DNA Res 12: 181–189.
- 30. Duan J, Wainwright MS, Comeron JM, Saitou N, Sanders AR, et al. (2003) Synonymous mutations in the human dopamine receptor D2 (DRD2) affect mRNA stability and synthesis of the receptor. Hum Mol Genet 12: 205–216.
- 31. Sethupathy P, Giang H, Plotkin JB, Hannenhalli S (2008) Genome-wide analysis of natural selection on human cis-elements. PLoS ONE 3: e3137.
- 32. Kimchi-Sarfaty C, Oh JM, Kim IW, Sauna ZE, Calcagno AM, et al. (2007) A “silent” polymorphism in the MDR1 gene changes substrate specificity. Science 315: 525–528.
- 33. Murphy FVt, Ramakrishnan V, Malkiewicz A, Agris PF (2004) The role of modifications in codon discrimination by tRNA(Lys)UUU. Nat Struct Mol Biol 11: 1186–1191.
- 34. Semon M, Duret L (2006) Evolutionary origin and maintenance of coexpressed gene clusters in mammals. Mol Biol Evol 23: 1715–1723.
- 35. Sauna ZE, Kimchi-Sarfaty C, Ambudkar SV, Gottesman MM (2007) Silent Polymorphisms Speak: How They Affect Pharmacogenomics and the Treatment of Cancer. Cancer Res 67: 9609–9612.
- 36. O'Dushlaine C, Shields D (2008) Marked variation in predicted and observed variability of tandem repeat loci across the human genome. BMC Genomics 9: 175.
- 37. Orr HT, Zoghbi HY (2007) Trinucleotide repeat disorders. Annu Rev Neurosci 30: 575–621.
- 38. Vinces MD, Legendre M, Caldara M, Hagihara M, Verstrepen KJ (2009) Unstable tandem repeats in promoters confer transcriptional evolvability. Science 324: 1213–1216.
- 39. Lim MH, Kim HW, Paik KC, Cho SC, Yoon DY, et al. (2006) Association of the DAT1 polymorphism with attention deficit hyperactivity disorder (ADHD): a family-based approach. Am J Med Genet B Neuropsychiatr Genet 141B: 309–311.
- 40. Guindalini C, Howard M, Haddley K, Laranjeira R, Collier D, et al. (2006) A dopamine transporter gene functional variant associated with cocaine abuse in a Brazilian sample. Proc Natl Acad Sci U S A 103: 4552–4557.
- 41. O'Gara C, Stapleton J, Sutherland G, Guindalini C, Neale B, et al. (2007) Dopamine transporter polymorphisms are associated with short-term response to smoking cessation treatment. Pharmacogenet Genomics 17: 61–67.
- 42. Benson G (1999) Tandem repeats finder: a program to analyze DNA sequences. Nucleic Acids Res 27: 573–580.
- 43. Boeva V, Regnier M, Papatsenko D, Makeev V (2006) Short fuzzy tandem repeats in genomic sequences, identification, and possible role in regulation of gene expression. Bioinformatics 22: 676–684.
- 44. Denoeud F, Vergnaud G, Benson G (2003) Predicting human minisatellite polymorphism. Genome Res 13: 856–867.
- 45. Legendre M, Pochet N, Pak T, Verstrepen KJ (2007) Sequence-based estimation of minisatellite and microsatellite repeat variability. Genome Res 17: 1787–1796.
- 46. Cooper GM, Zerr T, Kidd JM, Eichler EE, Nickerson DA (2008) Systematic assessment of copy number variant detection via genome-wide SNP genotyping. Nat Genet 40: 1199–1203.
- 47. Conrad DF, Pinto D, Redon R, Feuk L, Gokcumen O, et al. (2009) Origins and functional impact of copy number variation in the human genome. Nature advance online publication.
- 48. Frazer KA, Ballinger DG, Cox DR, Hinds DA, Stuve LL, et al. (2007) A second generation human haplotype map of over 3.1 million SNPs. Nature 449: 851–861.
- 49. Dear PH (2009) Copy-number variation: the end of the human genome? Trends Biotechnol 27: 448–454.
- 50. Bailey JA, Eichler EE (2006) Primate segmental duplications: crucibles of evolution, diversity and disease. Nat Rev Genet 7: 552–564.
- 51. Cheng Z, Ventura M, She X, Khaitovich P, Graves T, et al. (2005) A genome-wide comparison of recent chimpanzee and human segmental duplications. Nature 437: 88–93.
- 52. Drgon T, Lin Z, Wang GJ, Fowler J, Pablo J, et al. (2006) Common human 5′ dopamine transporter (SLC6A3) haplotypes yield varying expression levels in vivo. Cell Mol Neurobiol 26: 875–889.
- 53. Martinat C, Bacci J-J, Leete T, Kim J, Vanti WB, et al. (2006) Cooperative transcription activation by Nurr1 and Pitx3 induces embryonic stem cell maturation to the midbrain dopamine neuron phenotype. Proceedings of the National Academy of Sciences of the United States of America 103: 2874–2879.
- 54. Alberini CM (2009) Transcription factors in long-term memory and synaptic plasticity. Physiol Rev 89: 121–145.
- 55. Dolfini D, Zambelli F, Pavesi G, Mantovani R (2009) A perspective of promoter architecture from the CCAAT box. Cell Cycle 8: 4127–4137.
- 56. Zhang MQ (2005) Using CorePromoter to find human core promoters. Curr Protoc Bioinformatics Chapter 2: Unit 2 9.
- 57. Thierry-Mieg D, Thierry-Mieg J (2006) AceView: a comprehensive cDNA-supported gene and transcripts annotation. Genome Biol 7: Suppl 1S12 11–14.
- 58. Finotti A, Treves S, Zorzato F, Gambari R, Feriotto G (2008) Upstream stimulatory factors are involved in the P1 promoter directed transcription of the A beta H-J-J locus. BMC Mol Biol 9: 110.
- 59. Carninci P, Sandelin A, Lenhard B, Katayama S, Shimokawa K, et al. (2006) Genome-wide analysis of mammalian promoter architecture and evolution. Nat Genet 38: 626–635.
- 60. Schulz R, McCole RB, Woodfine K, Wood AJ, Chahal M, et al. (2009) Transcript- and tissue-specific imprinting of a tumour suppressor gene. Hum Mol Genet 18: 118–127.
- 61. Lammel S, Hetzel A, Hackel O, Jones I, Liss B, et al. (2008) Unique properties of mesoprefrontal neurons within a dual mesocorticolimbic dopamine system. Neuron 57: 760–773.
- 62. Vandenbergh DJ, Persico AM, Uhl GR (1992) A human dopamine transporter cDNA predicts reduced glycosylation, displays a novel repetitive element and provides racially-dimorphic TaqI RFLPs. Brain Res Mol Brain Res 15: 161–166.
- 63. Pristupa ZB, Wilson JM, Hoffman BJ, Kish SJ, Niznik HB (1994) Pharmacological heterogeneity of the cloned and native human dopamine transporter: disassociation of [3H]WIN 35,428 and [3H]GBR 12,935 binding. Mol Pharmacol 45: 125–135.
- 64. Lin JC, Thirumalai D (2008) Relative stability of helices determines the folding landscape of adenine riboswitch aptamers. J Am Chem Soc 130: 14080–14081.
- 65. Ringner M, Krogh M (2005) Folding free energies of 5′-UTRs impact post-transcriptional regulation on a genomic scale in yeast. PLoS Comput Biol 1: e72.
- 66. Ozsarac N, Santha E, Hoffman BJ (2002) Alternative non-coding exons support serotonin transporter mRNA expression in the brain and gut. Journal of Neurochemistry 82: 336–344.
- 67. Ma Q (2006) Transcriptional regulation of neuronal phenotype in mammals. J Physiol 575: 379–387.
- 68. Wang J, Bannon MJ (2005) Sp1 and Sp3 activate transcription of the human dopamine transporter gene. J Neurochem 93: 474–482.
- 69. Frith MC, Li MC, Weng Z (2003) Cluster-Buster: Finding dense clusters of motifs in DNA sequences. Nucleic Acids Res 31: 3666–3668.
- 70. Robertson G, Bilenky M, Lin K, He A, Yuen W, et al. (2006) cisRED: a database system for genome-scale computational discovery of regulatory elements. Nucleic Acids Res 34: D68–73.
- 71. Verma A, Yadav VK, Basundra R, Kumar A, Chowdhury S (2009) Evidence of genome-wide G4 DNA-mediated gene expression in human cancer cells. Nucleic Acids Res.
- 72. Lane AN, Chaires JB, Gray RD, Trent JO (2008) Stability and kinetics of G-quadruplex structures. Nucleic Acids Res 36: 5482–5515.
- 73. Kikin O, Zappala Z, D'Antonio L, Bagga PS (2008) GRSDB2 and GRS_UTRdb: databases of quadruplex forming G-rich sequences in pre-mRNAs and mRNAs. Nucl Acids Res 36: D141–148.
- 74. Kumari S, Bugaut A, Huppert JL, Balasubramanian S (2007) An RNA G-quadruplex in the 5′ UTR of the NRAS proto-oncogene modulates translation. Nat Chem Biol 3: 218–221.
- 75. Didiot MC, Tian Z, Schaeffer C, Subramanian M, Mandel JL, et al. (2008) The G-quartet containing FMRP binding site in FMR1 mRNA is a potent exonic splicing enhancer. Nucleic Acids Res 36: 4902–4912.
- 76. Barlati S, Barbon A (2005) RNA editing: a molecular mechanism for the fine modulation of neuronal transmission. Acta Neurochir Suppl 93: 53–57.
- 77. Price RD, Sanders-Bush E (2000) RNA editing of the human serotonin 5-HT(2C) receptor delays agonist-stimulated calcium release. Mol Pharmacol 58: 859–862.
- 78. Kuyumcu-Martinez NM, Cooper TA (2006) Misregulation of alternative splicing causes pathogenesis in myotonic dystrophy. Prog Mol Subcell Biol 44: 133–159.
- 79. Andken BB, Lim I, Benson G, Vincent JJ, Ferenc MT, et al. (2007) 3′-UTR SIRF: a database for identifying clusters of whort interspersed repeats in 3′ untranslated regions. BMC Bioinformatics 8: 274.
- 80. Aranda-Abreu GE, Hernandez ME, Soto A, Manzo J (2005) Possible Cis-acting signal that could be involved in the localization of different mRNAs in neuronal axons. Theor Biol Med Model 2: 33.
- 81. Kremerskothen J, Kindler S, Finger I, Veltel S, Barnekow A (2006) Postsynaptic recruitment of Dendrin depends on both dendritic mRNA transport and synaptic anchoring. J Neurochem 96: 1659–1666.
- 82. Corral-Debrinski M (2007) mRNA specific subcellular localization represents a crucial step for fine-tuning of gene expression in mammalian cells. Biochim Biophys Acta 1773: 473–475.
- 83. Chiaruttini C, Vicario A, Li Z, Baj G, Braiuca P, et al. (2009) Dendritic trafficking of BDNF mRNA is mediated by translin and blocked by the G196A (Val66Met) mutation. Proc Natl Acad Sci U S A 106: 16481–16486.
- 84. Abdel-Ghany M, el-Gendy K, Zhang S, Raden D, Racker E (1989) Brain protein kinase C phosphorylating poly(arginine,serine) or lamin B is stimulated by anions and by an activator purified from bovine serum albumin preparations. Proc Natl Acad Sci U S A 86: 1761–1765.
- 85. Crayton ME 3rd, Powell BC, Vision TJ, Giddings MC (2006) Tracking the evolution of alternatively spliced exons within the Dscam family. BMC Evol Biol 6: 16.
- 86. Cusack BP, Wolfe KH (2005) Changes in alternative splicing of human and mouse genes are accompanied by faster evolution of constitutive exons. Mol Biol Evol 22: 2198–2208.
- 87. Chen M, Manley JL (2009) Mechanisms of alternative splicing regulation: insights from molecular and genomics approaches. Nat Rev Mol Cell Biol 10: 741–754.
- 88. Chinta SJ, Pai HV, Ravindranath V (2005) Presence of splice variant forms of cytochrome P4502D1 in rat brain but not in liver. Brain Res Mol Brain Res 135: 81–92.
- 89. Tzvetkov MV, Meineke C, Oetjen E, Hirsch-Ernst K, Brockmoller J (2007) Tissue-specific alternative promoters of the serotonin receptor gene HTR3B in human brain and intestine. Gene 386: 52–62.
- 90. Stamm S (2008) Regulation of alternative splicing by reversible protein phosphorylation. J Biol Chem 283: 1223–1227.
- 91. Loria PM, Duke A, Rand JB, Hobert O (2003) Two neuronal, nuclear-localized RNA binding proteins involved in synaptic transmission. Curr Biol 13: 1317–1323.
- 92. Rahman L, Bliskovski V, Reinhold W, Zajac-Kaye M (2002) Alternative splicing of brain-specific PTB defines a tissue-specific isoform pattern that predicts distinct functional roles. Genomics 80: 245–249.
- 93. Nairn AC, Greengard P (2001) A novel cyclin provides a link between dopamine and RNA processing. Neuron 32: 174–176.
- 94. Zhou HL, Baraniak AP, Lou H (2007) Role for Fox-1/Fox-2 in mediating the neuronal pathway of calcitonin/calcitonin gene-related peptide alternative RNA processing. Mol Cell Biol 27: 830–841.
- 95. Yeo GW, Coufal NG, Liang TY, Peng GE, Fu X-D, et al. (2009) An RNA code for the FOX2 splicing regulator revealed by mapping RNA-protein interactions in stem cells. Nat Struct Mol Biol 16: 130–137.
- 96. Kuroyanagi H (2009) Fox-1 family of RNA-binding proteins. Cell Mol Life Sci 66: 3895–3907.
- 97. Nissim-Rafinia M, Kerem B (2006) Splicing modulation as a modifier of the CFTR function. Prog Mol Subcell Biol 44: 233–254.
- 98. Kim SR, Sai K, Tanaka-Kagawa T, Jinno H, Ozawa S, et al. (2007) Haplotypes and a novel defective allele of CES2 found in a Japanese population. Drug Metab Dispos 35: 1865–1872.
- 99. Kawase T, Akatsuka Y, Torikai H, Morishima S, Oka A, et al. (2007) Alternative splicing due to an intronic SNP in HMSD generates a novel minor histocompatibility antigen. Blood 110: 1055–1063.
- 100. Zhang XH, Chasin LA (2006) Comparison of multiple vertebrate genomes reveals the birth and evolution of human exons. Proc Natl Acad Sci U S A 103: 13427–13432.
- 101. Licatalosi DD, Darnell RB (2006) Splicing regulation in neurologic disease. Neuron 52: 93–101.
- 102. Carlini DB, Genut JE (2006) Synonymous SNPs provide evidence for selective constraint on human exonic splicing enhancers. J Mol Evol 62: 89–98.
- 103. Hebert SS, De Strooper B (2007) Molecular biology. miRNAs in neurodegeneration. Science 317: 1179–1180.
- 104. Wang E (2007) MicroRNA, the putative molecular control for mid-life decline. Ageing Res Rev 6: 1–11.
- 105. Ibanez-Ventoso C, Yang M, Guo S, Robins H, Padgett RW, et al. (2006) Modulated microRNA expression during adult lifespan in Caenorhabditis elegans. Aging Cell 5: 235–246.
- 106. Vasudevan S, Tong Y, Steitz JA (2007) Switching from repression to activation: microRNAs can up-regulate translation. Science 318: 1931–1934.
- 107. Kawahara Y, Zinshteyn B, Sethupathy P, Iizasa H, Hatzigeorgiou AG, et al. (2007) Redirection of silencing targets by adenosine-to-inosine editing of miRNAs. Science 315: 1137–1140.
- 108. Lee I, Ajay SS, Yook JI, Kim HS, Hong SH, et al. (2009) New class of microRNA targets containing simultaneous 5′-UTR and 3′-UTR interaction sites. Genome Res 19: 1175–1183.
- 109. Chen JM, Ferec C, Cooper DN (2006) A systematic analysis of disease-associated variants in the 3′ regulatory regions of human protein-coding genes II: the importance of mRNA secondary structure in assessing the functionality of 3′ UTR variants. Hum Genet 120: 301–333.
- 110. Kozak M (2005) Regulation of translation via mRNA structure in prokaryotes and eukaryotes. Gene 361: 13–37.
- 111. Mill J, Asherson P, Browes C, D'Souza U, Craig I (2002) Expression of the dopamine transporter gene is regulated by the 3′ UTR VNTR: Evidence from brain and lymphocytes using quantitative RT-PCR. Am J Med Genet 114: 975–979.
- 112. Brodskii LI, Ivanov VV, Kalaidzidis Ia L, Leontovich AM, Nikolaev VK, et al. (1995) [GeneBee-NET: An Internet based server for biopolymer structure analysis]. Biokhimiia 60: 1221–1230.
- 113. Singh M, Khan AJ, Shah PP, Shukla R, Khanna VK, et al. (2008) Polymorphism in environment responsive genes and association with Parkinson disease. Mol Cell Biochem 312: 131–138.
- 114. Wang Y, Leung FC (2004) An evaluation of new criteria for CpG islands in the human genome as gene markers. Bioinformatics 20: 1170–1177.
- 115. Levenson JM, Roth TL, Lubin FD, Miller CA, Huang IC, et al. (2006) Evidence that DNA (cytosine-5) methyltransferase regulates synaptic plasticity in the hippocampus. J Biol Chem 281: 15763–15773.
- 116. Nelson ED, Kavalali ET, Monteggia LM (2008) Activity-dependent suppression of miniature neurotransmission through the regulation of DNA methylation. J Neurosci 28: 395–406.
- 117. Miller CA, Campbell SL, Sweatt JD (2008) DNA methylation and histone acetylation work in concert to regulate memory formation and synaptic plasticity. Neurobiol Learn Mem 89: 599–603.
- 118. Simmen MW (2008) Genome-scale relationships between cytosine methylation and dinucleotide abundances in animals. Genomics 92: 33–40.
- 119. Bock C, Lengauer T (2008) Computational epigenetics. Bioinformatics 24: 1–10.
- 120. Cooper SJ, Trinklein ND, Anton ED, Nguyen L, Myers RM (2006) Comprehensive analysis of transcriptional promoter structure and function in 1% of the human genome. Genome Res 16: 1–10.
- 121. Illingworth R, Kerr A, Desousa D, Jorgensen H, Ellis P, et al. (2008) A novel CpG island set identifies tissue-specific methylation at developmental gene loci. PLoS Biol 6: e22.
- 122. Varriale A, Bernardi G (2006) DNA methylation and body temperature in fishes. Gene 385: 111–121.
- 123. Jiang C, Han L, Su B, Li WH, Zhao Z (2007) Features and trend of loss of promoter-associated CpG islands in the human and mouse genomes. Mol Biol Evol 24: 1991–2000.
- 124. Zelko IN, Mueller MR, Folz RJ (2008) Transcription factors sp1 and sp3 regulate expression of human extracellular superoxide dismutase in lung fibroblasts. Am J Respir Cell Mol Biol 39: 243–251.
- 125. Choi JK, Kim YJ (2008) Epigenetic regulation and the variability of gene expression. Nat Genet 40: 141–147.
- 126. Radwan A, Younis A, Luykx P, Khuri S (2008) Prediction and analysis of nucleosome exclusion regions in the human genome. BMC Genomics 9: 186.
- 127. Kouzarides T (2007) Chromatin modifications and their function. Cell 128: 693–705.
- 128. Oki M, Aihara H, Ito T (2007) Role of histone phosphorylation in chromatin dynamics and its implications in diseases. Subcell Biochem 41: 319–336.
- 129. Kass SU, Wolffe AP (1998) DNA methylation, nucleosomes and the inheritance of chromatin structure and function. Novartis Found Symp 214: 22–35; discussion 36-50.
- 130. Santos-Reboucas CB, Pimentel MM (2007) Implication of abnormal epigenetic patterns for human diseases. Eur J Hum Genet 15: 10–17.
- 131. Colvis CM, Pollock JD, Goodman RH, Impey S, Dunn J, et al. (2005) Epigenetic mechanisms and gene networks in the nervous system. J Neurosci 25: 10379–10389.
- 132. Michopoulos I, Douzenis A, Christodoulou C, Lykouras L (2006) Topiramate use in alprazolam addiction. World J Biol Psychiatry 7: 265–267.
- 133. Myrick H, Brady KT (2003) The use of divalproex in the treatment of addictive disorders. Psychopharmacol Bull 37: Suppl 289–97.
- 134. Reid MS, Casadonte P, Baker S, Sanfilipo M, Braunstein D, et al. (2005) A placebo-controlled screening trial of olanzapine, valproate, and coenzyme Q10/L-carnitine for the treatment of cocaine dependence. Addiction 100: Suppl 143–57.
- 135. Wang J, Michelhaugh SK, Bannon MJ (2007) Valproate robustly increases Sp transcription factor-mediated expression of the dopamine transporter gene within dopamine cells. Eur J Neurosci 25: 1982–1986.
- 136. Ralph-Williams RJ, Paulus MP, Zhuang X, Hen R, Geyer MA (2003) Valproate attenuates hyperactive and perseverative behaviors in mutant mice with a dysregulated dopamine system. Biol Psychiatry 53: 352–359.
- 137. Duret L, Galtier N (2009) Biased gene conversion and the evolution of mammalian genomic landscapes. Annu Rev Genomics Hum Genet 10: 285–311.
- 138. Levdansky E, Romano J, Shadkchan Y, Sharon H, Verstrepen KJ, et al. (2007) Coding tandem repeats generate diversity in Aspergillus fumigatus genes. Eukaryot Cell 6: 1380–1391.
- 139. Verstrepen KJ, Jansen A, Lewitter F, Fink GR (2005) Intragenic tandem repeats generate functional variability. Nat Genet 37: 986–990.
- 140. Nguyen DQ, Webber C, Hehir-Kwa J, Pfundt R, Veltman J, et al. (2008) Reduced purifying selection prevails over positive selection in human copy number variant evolution. Genome Res 18: 1711–1723.
- 141. Chiang CW, Derti A, Schwartz D, Chou MF, Hirschhorn JN, et al. (2008) Ultraconserved elements: analyses of dosage sensitivity, motifs and boundaries. Genetics 180: 2277–2293.
- 142. Burt SA, Mikolajewski AJ (2008) Preliminary evidence that specific candidate genes are associated with adolescent-onset antisocial behavior. Aggress Behav 34: 437–445.
- 143. Gizer IR, Ficks C, Waldman ID (2009) Candidate gene studies of ADHD: a meta-analytic review. Hum Genet 126: 51–90.
- 144. van der Zwaluw CS, Engels RC, Buitelaar J, Verkes RJ, Franke B, et al. (2009) Polymorphisms in the dopamine transporter gene (SLC6A3/DAT1) and alcohol dependence in humans: a systematic review. Pharmacogenomics 10: 853–866.
- 145. Wonodi I, Hong LE, Stine OC, Mitchell BD, Elliott A, et al. (2009) Dopamine transporter polymorphism modulates oculomotor function and DAT1 mRNA expression in schizophrenia. Am J Med Genet B Neuropsychiatr Genet 150B: 282–289.
- 146. VanNess SH, Owens MJ, Kilts CD (2005) The variable number of tandem repeats element in DAT1 regulates in vitro dopamine transporter density. BMC Genet 6: 55.
- 147. Forbes EE, Brown SM, Kimak M, Ferrell RE, Manuck SB, et al. (2009) Genetic variation in components of dopamine neurotransmission impacts ventral striatal reactivity associated with impulsivity. Mol Psychiatry 14: 60–70.