Identification and Functional Analysis of Variant Haplotypes in the 5′-Flanking Region of Protein Phosphatase 2A-Bδ Gene

Serine-threonine protein phosphatase 2A (PP2A) is a trimeric holoenzyme that plays an integral role in the regulation of cell growth, differentiation, and apoptosis. The substrate specificity and (sub)cellular localization of the PP2A holoenzymes are highly regulated by interaction with a family of regulatory B subunits (PP2A-Bs). The regulatory subunit PP2A-B/PR55δ (PP2A-Bδ) is involving in the dephosphorylation of PP2A substrates and is crucial for controlling entry into and exit from mitosis. The molecular mechanisms involved in the regulation of expression of PP2A-Bδ gene (PPP2R2D) remain largely unknown. To explore genetic variations in the 5′-flanking region of PPP2R2D gene as well as their frequent haplotypes in the Han Chinese population and determine whether such variations have an impact on transcriptional activity, DNA samples were collected from 70 healthy Chinese donors and sequenced for identifying genetic variants in the 5′-flanking region of PPP2R2D. Four genetic variants were identified in the 1836 bp 5′-flanking region of PPP2R2D. Linkage disequilibrium (LD) patterns and haplotype profiles were constructed for the genetic variants. Using serially truncated human PPP2R2D promoter luciferase constructs, we found that a 601 bp (−540 nt to +61 nt) fragment constitutes the core promoter region. The subcloning of individual 5′-flanking fragment revealed the existence of three haplotypes in the distal promoter of PPP2R2D. The luciferase reporter assay showed that different haplotypes exhibited distinct promoter activities. The EMSA revealed that the −462 G>A variant influences DNA-protein interactions involving the nuclear factor 1 (NF1). In vitro reporter gene assay indicated that cotransfection of NF1/B expression plasmid could positively regulate the activity of PPP2R2D proximal promoter. Introduction of exogenous NF1/B expression plasmid further confirmed that the NF1 involves in the regulation of PPP2R2D gene expression. Our findings suggest that functional genetic variants and their haplotypes in the 5′-flanking region of PPP2R2D are critical for transcriptional regulation of PP2A-Bδ.


Introduction
Protein phosphatase 2A (PP2A) is a highly conserved serine/ threonine phosphatase which may account for up to 1% of cellular protein and the majority of serine/threonine phosphatase activity in the cell. The physiological functions of PP2A have been implicated in all aspects of cellular processes [1,2]. The PP2A heterotrimeric holoenzyme is composed of a catalytic subunit (PP2A-C), a scaffold subunit (PP2A-A) and one member of four families of regulatory subunits (PP2A-Bs). The structural (A) and catalytic (C) subunits form the constitutive core enzyme, which associates with one of a large number of regulatory B subunits. The PP2A-Bs subunits are derived from four diverse gene families (B, B9, B0, and B09). There is little sequence similarity between families, but members of a family maintain high similarity. Given the large number of PP2A subunits, it is thought that each cell expresses a dozen or more distinct holoenzyme complexes, which act on a diverse array of substrates [1,3].
The B family (B55, PR55) of regulatory subunits is encoded by four genes (PPP2R2s): Ba, Bb [4], Bc [5,6], and Bd [7]. B family members (PR55s) exhibit temporal and spatial expression patterns; B55a (PR55a) and B55d (PR55d) are expressed almost ubiquitously, while B55b (PR55b) and B55c (PR55c) are highly enriched in the brain [4,5,7]. The primary structures and functions of the B subunits are conserved from yeast to mammals [8]. The structures of the PR55 subunits are predicted to form b-propellers with divergent N-terminal tails that act as signals for subcellular targeting [9,10,11]. A structural feature of the PR55 subunits is the presence of five degenerate WD-40 repeats. WD-40 repeats are minimally conserved sequences of approximately 40 amino acids that typically end in tryptophan-aspartate (WD) and are thought to mediate protein-protein interactions [12].
In vitro dephosphorylation assays support a role for Bd in the targeting of the PP2A heterotrimer to the dephosphorylation and inactivation of ERKs [13]. Mochida et al. reported that a particular form of PP2A containing a B55d subunit (PP2A-B55d) has antimitotic activity and is the major protein phosphatase to act on model cyclin-dependent kinase (CDK) substrates in Xenopus egg extracts [14]. An interplay between the Drosophila serine/threonine kinase Greatwall (gwl) and PP2A/Bd during mitotic entry was recently observed in two separate studies [15,16]. PP2A/Bd activity prevents mitotic entry by maintaining Cdc25 in a dephosphorylated and inactive state [17]. Similar cell cycle regulatory activity has been observed with the mammalian ortholog of gwl, microtubule-associated serine/threonine kinaselike (MASTL); however, the human MASTL-PP2A interaction has not been characterized [17]. Together, these data indicate that PP2A/Bd plays essential roles in the regulation of substrate dephosphorylation and in cell cycle signaling pathways.
Currently, no functional variants of the PP2A-Bd subunit gene (PPP2R2D) have been found, and little is known about the regulation of PP2A-Bd expression. Moreover, the allele distribution in specific population is yet to be defined and a functional characterization of genetic variants or single nucleotide polymorphisms (SNPs) in the 59-flanking region of the PPP2R2D gene has not yet been undertaken. In this study, we identify four variants and multiple various haplotypes in the 59-flanking region of PPP2R2D in the southern Han Chinese population and demonstrate that certain functional variants in diverse haplotypes and nuclear factor 1 (NF1) may regulate the transcription of PPP2R2D gene and the expression of PP2A-Bd.

Study subjects
Seventy healthy Han Chinese donors were recruited for this study. The individuals were randomly selected from residents of Guangzhou city (Guangdong province, China) in southern China. At recruitment, each participant was personally interviewed to obtain detailed information on their demographic characteristic such as gender, age and related family background. All subjects were unrelated ethnic Han Chinese confirmed by ID cards and ''Household Register books'', which serve as identification in China as described in our previous study [18]. After physical examination, 5 ml of blood was collected, labeled, and delivered to the laboratory for immediate DNA isolation. This study was approved by the ethics committee of the School of Public Health, Sun Yat-sen University, and written informed consent was obtained from each subject.

Variants screening
Genomic DNA samples derived from peripheral blood were used to search for genetic variants within the 59-flanking region (21775 nt to +61 nt) of the PPP2R2D gene (the first nucleotide of the RNA transcript is defined as +1 nt). The 1836 bp 59-flanking fragment of PPP2R2D gene from each subject was amplified by PCR using a pair of primers (Table 1). The PCR cycling conditions were 94uC for 5 minutes, followed by 35 cycles at 94uC for 30 seconds, 55uC for 30 seconds, and 72uC for 2 minutes, with a final extension step at 72uC for 10 minutes. Genetic variants were detected by direct sequencing of the purified PCR products with the ABI PRISM DYE Terminator Sequencing Kit (Applied Biosystems, USA) using three sequencing primers (Table 1) to sequence the complete fragment. Finally, the Sequencer program was used to identify genetic variants that were then confirmed by 2 independent observers. We further confirmed these variant positions and individual genotypes by reamplifying and resequencing the variant sites from the opposite strand.

Linkage disequilibrium analysis and haplotype construction
The Hardy-Weinberg equilibrium (HWE) and pairwise linkage disequilibrium (LD) indices were calculated with the Haploview version 3.2 software (Broad Institute). We estimated the degree of pairwise LD between variants as quantified by the disequilibrium coefficients D9 or r 2 [19]. They define pairs to be in 'strong LD' if the one-sided upper 95% confidence bound of D9 is .0.98 (that is, consistent with no historical recombination) and the lower bound is above 0.7. A P value of ,0.05 was used as the criterion of statistical significance. All relevant haplotypes in the current population were also generated with Haploview.

Plasmids construction
To evaluate the promoter activity of different parts of the 59flanking region of the PPP2R2D gene with various haplotypes, we made serial truncations of the PPP2R2D promoter fragment and analyzed the activity of the reporter constructs. Three different upstream primers and one downstream (+61HR3) primer (Table 1), flanked by restriction sites at the 59-end, were designed for PCR reactions that employed human genomic DNA (wildtype) as template. The truncated fragments, which included F1 (21775 nt to +61 nt, 1836 bp as distal promoter), F2 (21245 nt to +61 nt, 1306 bp as middle promoter), and F3 (2540 nt to +61 nt, 601 bp as proximal promoter), were cloned into the pGL3-Basic Luciferase Reporter Vector (Promega), respectively. The recombinant plasmids, named pGL3b-2R2Dp-F1, -F2, and -F3, were confirmed by sequencing.
To examine the function of different haplotypes on PPP2R2D promoter activity, a series of reporter plasmids with different haplotypes encompassing the F1 fragments of human PPP2R2D were constructed by PCR from three genomic DNA samples with native haplotypes. The resulting constructs were termed pGL3b-2R2Dp-F1H1 (for distal 59-flanking region F1, haplotype 1, and the rest follow in this manner), -F1H2, and -F1H3, each of which harbors 59-flanking region with different haplotypes.
To study the effect of the 2462 G.A variant on the transcriptional activity of the proximal promoter, reporter constructs containing either G or A variant were generated by subcloning from the PCR products of the genomic DNA samples that are heterozygous for 2462 G.A variant (2462 GA heterozygote). The resulting two luciferase constructs with different allelotypes were designated as pGL3b-2R2Dp-462G and -462A, respectively. All fragments were digested with Kpn I and Hind III restriction enzymes and cloned in 59-39-orientation into a pGL3-Basic luciferase vector.
To construct the NF1/B expression plasmid (pEGFP-NF1/B), NF1/B coding sequence was amplified from total human cDNA by PCR using primers flanked with restriction sites ( Table 1). The fragment was then cloned into the EcoR I and Xho I sites of pEGFP-N1 vector that contains an expression cassette of enhanced green fluorescent protein (EGFP). The pEGFP-N1 vector was purchased from Clontech (TaKaRa Bio Company). All reporter constructs and expression plasmids were further verified by direct nucleotides sequencing.
For luciferase reporter assays, L02 cells (1610 4 ) were plated in 96-multiwell plates and grown to 80-90% confluence. Transfections were performed using Lipofectamine TM 2000 Reagent (Invitrogen) according to the manufacturer's instructions and a previous study from our group [18]. Cells were cotransfected with 50 ng of pRL-TK vector DNA (Promega) and 200 ng of either empty pGL3-Basic plasmid (a promoterless control; Promega) or one of several pGL3b-2R2Dp constructs harboring different lengths and allelotypes or haplotypes of the PPP2R2D promoter. The pRL-TK vector, which provided the constitutive expression of Renilla Luciferase, was cotransfected as an internal control to correct for differences in both transfection and harvesting efficiency. For cotransfection assays, various doses (250 ng, 500 ng, 1000 ng, respectively) of transcription factor NF1/B expression plasmid (pEGFP-NF1/B) or corresponding empty vector (pEGFP-N1), and reporter gene plasmids (200 ng of pGL3b-2R2Dp-462G or -462A, and 50 ng pRL-TK vector) were transfected into L02 cells. After 24 hours of incubation, cells were collected and analyzed for luciferase activity with the Dual-Luciferase Reporter Assay System (Promega). Promoter activity was reported as relative light units (RLUs) and normalized against the activity of the empty pGL3-Basic vector. All transfections were carried out in triplicate and repeated at least thrice in independent experiments.
In Silico analysis of the allelic variants in 59-flanking region of PPP2R2D The genetic variants identified in this study were compared with data from the National Center for Biotechnology Information (NCBI) database (http://www.ncbi.nlm.nih.gov/dbSNP) and phase 3 data from the HapMap Public Release database (http://hapmap.ncbi.nlm.nih.gov/). All identified variants were then subjected to computer-aided analyses with the online program AliBaba 2.1 (http://www.alibaba2.com) to predict physical transcription factor binding sites (TFBS) in the given DNA sequence based on the TRANSFAC database of eukaryotic transcription factors (TFs; TRANSFAC10). The Transcription Element Search System (TESS) was also used to integrate potential TFBSs in the 59-flanking region of PPP2R2D with various polymorphic alleles.

Electrophoretic mobility shift assay (EMSA)
To investigate whether sequence variants alter DNA-protein interactions, EMSAs were performed with the PPP2R2D promoter fragment using the LightShift Chemiluminescent EMSA kit (Thermo Fisher), in accordance with the manufacturer's instructions. Nuclear proteins were extracted from confluent cultures of human L02 cells with a Nuclear Extraction Kit (Active Motif). Double-stranded oligonucleotides to be used as DNA probes were labeled with biotin at the 59-end. Probe sequences corresponding to the 2462 G.A variant in the PPP2R2D promoter region were 59-biotin-GAATATTACATTGTACAGAAGCCAACAAAAGT-TATTGCATAG-39 for 2462 G, and 59-biotin-GAATATTA-CATTGTACAGAAACCAACAAAAGTTATTGCATAG-39 for 2462 A. The consensus sequence of the 59-biotin-labeled doublestranded probe designed to recognize transcription factor NF1 was 59-GGGCTAGATGGCTGCCAGCCAAGGCTTCAT-39 [21]. All labeled and unlabeled oligonucleotides were ordered from Sangon Biotech Co., Ltd (Shanghai, China). Complementary oligonucleotides were annealed and used either as probe or competitor DNA. In each EMSA reaction, an aliquot of 59biotinylated double-stranded oligonucleotide (10 fmole) was incubated with 10 mg nuclear protein extract for 20 minutes at room temperature in 20 ml of binding buffer (100 mM Tris, 500 mM KCl, 10 mM DTT, pH 7.5) with 1 ml poly(dINdC) (1 mg/ ml). For competition experiments, a 200-fold molar excess of unlabeled cold probe was added to the mixture as a specific competitor before the incubation period. For the supershift experiment, 10 mg of nuclear extract was preincubated with anti-NF1 antibody (Santa Cruz Biotechnology Inc, CA) for 30 minutes at room temperature before the addition of labeled probe. After the addition of the labeled probe, the reactions were incubated for another 45 minutes on ice, then submitted to 6% native polyacrylamide gel electrophoresis, and transferred (0.56 TBE, 100 V, 60 minutes) onto a positively charged nylon transfer membrane (Magmaprobe, Osmonics Inc.). Bands were visualized by stabilized streptavidin-horseradish peroxidase conjugate.

RNA extraction and real-time PCR
To analyze PPP2R2D mRNA levels, total RNA was extracted from L02 cell lines transfected with NF1/B encoding plasmids (250 ng, 500 ng, 1000 ng of pEGFP-NF1/B, respectively) or its corresponding empty vector (pEGFP-N1) using TRIzolH reagent (Invitrogen). Total RNA (500 ng) was reverse transcribed into signal-strand complementary DNA using oligo dT primer and PrimeScript TM RT Enzyme Mix I (SYBRH PrimeScript TM RT-PCR kit, TaKaRa). Relative gene expression quantitation for PPP2R2D, with ACTB as an internal reference gene, was performed with ABI 7500 system (Applied Biosystems) based on the SYBR-Green method. Primers used were: PPP2R2D, Forward, 59-CGTGAACAAGAGAATAAAAGCCG-39, Reverse, 59-CTTCAATATTGGGACCCGTAG-39; and ACTB, Forward, 59-CACCAGGGCGTGATGGT-39, Reverse, 59-CTCAAACATGATCTGGGTCAT-39. The PCR reaction mixture consisted of 0.1 mmol/L of each primer, 16 SYBRH Premix EX Taq TM (Perfect Real Time) premix reagent (TaKaRa), and 50 ng complementary DNA to a final volume of 20 ml. Cycling conditions were 95uC for 30 seconds, followed by 40 cycles at 95uC for 5 seconds and 60uC for 34 seconds. The averaged cycle threshold (C T ) values of each reaction derived from the target gene, determined with ABI 7500 system software, were normalized to ACTB levels. The DC T was calculated by subtracting the C T of ACTB from the C T of the target gene of interest. The DDC T was calculated by subtracting the DC T of the reference sample from the DC T of each sample. Fold change was generated using the equation 2 2DDCT .

Western blot analysis
Total proteins were extracted from L02 cells that cotransfected with pEGFP-NF1/B plasmid with lysis buffer (50 mM HEPES pH 7.6, 250 mM NaCl, 0.1% NP-40, 5 mM EDTA, 0.5 mM PMSF and the proteinase inhibitor cocktail). 70 mg of total proteins were used for each lane of loaded sample. The protein blots were blocked with 5% fat-free milk in TBS buffer for 1 hour and then incubated for overnight at 4uC with antibody against PP2A-Bd (Santa Cruz Biotechnology Inc, CA), and b-actin antibody (Cell Signaling Technology, MA) at a dilution of 1:500 and 1:3000, respectively. The secondary antibody was anti-goat IgG or anti-mouse IgG (Santa Cruz Biotechnology Inc, CA) at a dilution of 1:1000 and 1:3000 respectively. The blots were detected using the enhanced chemiluminescence method (Pierce).

Statistical analysis
LD analysis was performed using Haploview. The genotype frequencies in the study subjects were compared with the frequencies expected by Hardy-Weinberg equilibrium (HWE) by goodness-of-fit x 2 , and the x 2 values for statistical significance and their corresponding P values were calculated. Pairwise LD between SNPs was measured using the disequilibrium parameters D9 and r 2 . D is a measure of the difference between the frequency of a haplotype and the frequency that is expected under Hardy-Weinberg Equilibrium. D9 is a standardized version of D that ranges from 0 to 1. The higher the D9-value, the greater the association between the two loci being studied. Heterozygosity (p) was calculated as the average of 2jk/n(n-1) for all base pairs screened, with j and k referring to the number of copies of the minor and major alleles, respectively (n = j+k). Student's t test for two groups or ANOVA with an S-N-K test for multiple groups was used to examine the differences in luciferase reporter gene expression. P,0.05 was used as the criterion for statistical significance. All statistical tests were two-sided.

Results
Identification of genetic variants within the 59-flanking region of PPP2R2D in the southern Han Chinese population We sequenced a 1836 bp segment of the 59-flanking region of PPP2R2D in seventy healthy subjects and identified four genetic variants ( Table 2). The four variants found in this population are registered in the current NCBI SNP database (dbSNP): 21684 A.G (rs11146169) (Fig. 1A), 21610 A.T (rs77523588) (Fig. 1B), 21196 C.T (rs9419202) (Fig. 1C), and 2462 G.A (rs7074421) (Fig. 1D). Although 21684 A.G (rs11146169) is present in the dbSNP database, the minor allele frequency (MAF) for this site is less than 1% in our study population. Among the seventy healthy Han Chinese subjects, the MAF for the four genetic variants in all the variant allelotypes was 21684 G (0.71%), 21610 T (5.71%), 21196 T (2.14%), 2462 A (2.14%). All the allele frequencies were in HWE (P.0.05). The MAFs for three of the variants (21684 A.G, 21196 C.T, and 2462 G.A) were less than 5%. Not  surprisingly, none of the minor alleles of the three variants were observed in the homozygous state in any of the samples in the current population. Information regarding the SNP coordinates, MAF, and HWE of the variants is presented in Table 2.
To characterize the distribution of these variants in the 59flanking region of PPP2R2D, the MAFs in this southern Han Chinese population were compared to those in HapMap populations. The four genetic variants have been observed previously in various populations and their allele frequencies (from HapMap phase 3 data) are reported in Table 3. The MAF of 21684 A.G in our sample (0.71%, n = 140) was lower than that reported for the Asian populations in HapMap-HCB (2.22%, n = 90) and HapMap-JPT (1.11%, n = 90), but the frequencies were not statistically different from each other (P.0.05). The MAF of 21610 A.T in our study (5.71%) was similar to that of Asian populations in HapMap-HCB and -JPT (5.0%, n = 120). The MAF of 21196 C.T in our study (2.14%, n = 140) was identical to that in HapMap-HCB (1.2%, n = 86) and HapMap-JPT (2.3%, n = 172) (P.0.05). However, in the HapMap-CEU (14.6%, n = 226) and HapMap-YRI (21.7%, n = 226) populations, the MAFs of 21196 C.T were significantly higher than the MAF of this variant in the Han population of this study (P,0.05). The MAF of 2462 G.A in our study (2.14%, n = 140) was distributed similarly to the 21196 C.T. Taken together, these results indicate that there are significant differences in the genotypic distributions of the 21196 C.T and 2462 G.A variants between our southern Han Chinese group and European/sub-Saharan African groups in HapMap.
Next we assessed the degree of heterozygosity (p) in the southern Han population with HapMap phase 3 data. The heterozygosity of 21684 A.G (p = 0.014, n = 140) in the Han Chinese population was found to be slightly lower than that in the HapMap-HCB (p = 0.044, n = 90) and HapMap-JPT (p = 0.022, n = 90) populations, but showed no significant difference. The p value for 21610 A.T (p = 0.109, n = 140) was similar to that in HapMap-HCB+JPT (p = 0.096, n = 120) (P.0.05). The p value for 21196 C.T (p = 0.042, n = 140) was also found to be similar to that in HapMap-JPT (p = 0.046, n = 172) and slight higher than that in HapMap-HCB (p = 0.023, n = 86) but displayed no significant difference. However, the p values of 21196 C.T (p = 0.042, n = 140) was significantly lower than that in HapMap-CEU (p = 0.251, n = 226) and HapMap-YRI (p = 0.341, n = 226) populations (P,0.01). The p value for 2462 G.A (p = 0.042, n = 140) was distributed similarly to the 21196 C.T. These results indicate that the differences in heterozygosity might be due to racial differences. Variations with high heterozygosity values usually have functional significance and merit further study.

Linkage disequilibrium (LD) and haplotype analysis
To examine the degree of linkage among the four variants identified in the 59-flanking region of PPP2R2D, We quantified the degree of pairwise linkage disequilibrium (LD) using the disequilibrium coefficient D9 or r 2 and the EM-based algorithm of Haploview 3.2. Because there is no single best measure of LD that is universally applicable, the complementary measures D9 and r 2 were used in this study [22]. The 21196 C.T and 2462 G.A variants were found to be in high LD in this Han Chinese population, with r 2 = 1.0 and the lower and upper bounds of the 95% confidence interval (95% CI) for D9 at 0.59 and 1.00, respectively (Fig. 2). However, other variants were found to be in low LD, with an upper 95% CI bound for D9 of 0.97 and a lower bound of 0.04 (r 2 ,0.01) in this population sample. The LD results are consistent with the distribution of genotypes and the heterozygosity of these genetic variants in the southern Han Chinese population.
Haplotype analysis was performed on variants in the 59flanking region (21775 nt to +61 nt; F1 fragment) of PPP2R2D to identify haplotypes among the southern Han Chinese population and to estimate their frequencies. Three major novel haplotypes (F1H1, F1H2, and F1H3) were identified and their frequencies were calculated in our sample population; the other haplotype identified was rare and accounted for 0.72% (Table 4). Consistent with the LD analysis described above, alleles 21196 C and 2462 G were found to be transmitted together, as were the 21196 T and 2462 A alleles. However, no separate haplotype blocks were observed might help to suggest that the different 59-flanking regions are independent functional domains in PPP2R2D gene. Tag SNPs were found for three of the variants (21684 A.G, 21610 A.T, and 2462 G.A) in the distal and proximal 59-flanking region of the PPP2R2D in current study sample.
Promoter activity of the 59-flanking region of PPP2R2D We next attempted to determine whether this fragment (21775 nt to +61 nt) constitutes an active promoter. We amplified a 1836 bp fragment containing the 59-flanking region of PPP2R2D using wild-type human genomic DNA as the template (Fig. 3A). Subsequently, we generated truncated constructs (F1, F2, and F3) by progressive deletion of nucleotides from the 59-end and cloned these fragments into the pGL3-Basic luciferase vector (Fig. 3B). These constructs were transiently transfected into human L02 cells and tested for luciferase activity to define the minimal sequence for the transcription of PPP2R2D. As shown in Fig. 3C, the promoter activity of all the constructs was higher than the basal activity of the pGL3-Basic vector used as control (Ctrl) (P,0.05). Sequential deletions made to the 59-end of the 1836 bp F1 fragment to generate the construct from F1 to the 1306 bp F2 fragment resulted in a decrease in luciferase activity of approximately 42% (relative to the longest construct, F1). This result indicates that transcriptional repression exists within this distal promoter region. Removal of 705 bp from the 59-end of the F2 fragment generated construct F3 (2540 nt to +61 nt). F3 generated 125% of the luciferase activity of the F1 construct and 213% of the activity of the F2 construct. This indicates that critical cis-acting elements required for PPP2R2D transcription are located in the F3 region harboring the proximal promoter and suggests that the region between 2540 nt and +61 nt is responsible for most of the promoter activity of PPP2R2D. Therefore, the F3 fragment of the proximal 59-flanking region of PPP2R2D is recognized as the active core promoter. The F3 fragment contains one of the four variants identified earlier (2462 G.A). To investigate the effect of this potential functional SNP on the expression of PP2A-Bd, the 59-flanking region of PPP2R2D containing the 2462 G or 2462 A allele was subjected to further functional analysis for the promoter activity.

Differential proximal promoter activity of PPP2R2D with 2462 G.A variants
We wished to assess the impact of the 2462 G.A variant on the transcriptional activity of the PPP2R2D proximal promoter. We generated two luciferase reporter gene constructs, pGL3b-2R2Dp 2462 G and 2462 A, that contained different alleles at position 2462 nt in the proximal promoter region of PPP2R2D. The recombinant plasmids were transfected into L02 cells. The empty vector (pGL3b) and a Renilla construct (pRL-TK) were used as controls (Ctrl). As shown in Fig. 4, the two PPP2R2D proximal promoter constructs generated higher levels of luciferase expression than the empty vector control (P,0.05). The two constructs also differed from each other in luciferase expression. The F3 fragment containing variant 2462 G displayed the higher transcriptional activity than the fragment containing 2462 A (P,0.05). These data indicate that different allelotypes have a significant effect on the transcriptional activity of the PPP2R2D proximal promoter.

Differential transcriptional activity of different PPP2R2D promoter haplotypes
To assess the impact of the potentially functional variant identified above (Fig. 5A) on the transcriptional activity of the different PPP2R2D promoter haplotypes, we generated three luciferase reporter gene constructs (pGL3b-2R2Dp-F1H1 --F1H3), each of which contains a different haplotype in the promoter region of PPP2R2D (Fig. 5B). The recombinant plasmids were transfected into L02 cells along with the empty vector (pGL3b) and a Renilla construct (pRL-TK), which were used as controls. As shown in Fig. 5C, all the constructs provided higher  The 2462 G.A variant affects NF1 binding to the proximal promoter of PPP2R2D In Silico analysis of the 59-flanking region of PPP2R2D using a computer algorithm revealed several putative binding sites for transcriptional factors (TFs). Some of TFs binding patterns changed when the observed genetic variation in 59-flanking region of PPP2R2D was taken into account (Table 2). Interestingly, the 2462 G.A variant located in the proximal promoter of PPP2R2D altered the binding site of the transcription factor NF1 ( Table 2). To further confirm this finding, we performed an electrophoretic mobility shift assay (EMSA). Each fragment with wild-type alleles was cloned into the pGL3-Basic vector. C. Relative luciferase activity of series truncated constructs of 59-flanking region of PPP2R2D in human L02 cells. Each construct or empty vector was transfected into L02 cells with pRL-TK plasmid as an internal intrinsic control of transfection efficiency. After 24 hours, the luciferase activities from each testing construct and also from the internal control plasmid were measured using Dual-Luciferase Reporter Assay System. RLU indicates relative light units. The relative luciferase activity was normalized to Renilla activity and was relative to pGL3-Basic control (Ctrl), which was set as 1.0 RLU. Compared with the basal activity of Ctrl, the promoter activity of all constructs was increased at statistically significant levels (*P,0.05). Compared with the activity of the F3 construct, all constructs showed significant lower activity ( # P,0.05). The results represent the mean 6 SD of three independent experiments. doi:10.1371/journal.pone.0035524.g003 To identify whether these cis-acting elements physically interact with trans-factors, we incubated nuclear extracts from human L02 cells with two 59-biotinylated double-stranded oligonucleotides containing the 2462 G or the 2462 A variant. In mobility shift assays where only one band was observed on the gel, a DNAprotein complex was detected with the 2462 G probe, but not with the 2462 A probe (indicated by the arrow in Fig. 6, A and B). The difference in the binding affinity of these two probes was shown by competition experiments using unlabeled 2462 G or 2462 A cold probes. The DNA-protein complex containing the 2462 G probe (Fig. 6A, lane 1) was completely outcompeted by the addition of a 200-fold molar excess of unlabeled 2462 G probe (Fig. 6A, lane 2). Besides, the DNA-protein complexes in lanes 1 (Fig. 6A) are in the same position as the labeled NF1 consensus probe (Fig. 6A, lane 7), suggesting that NF1 in the nuclear extract bound to the probe.
Consistent with the competition assay, the cross-competition assay showed that this DNA-protein complex was completely outcompeted by the same amount of unlabeled NF1 consensus probe (Fig. 6A, lane 3; Fig. 6B, lane 5), but not outcompeted by the same amount of unlabeled 2462 A probe (Fig. 6B, lane 4), suggesting that the protein bound to the 2462 G probe may be NF1 and the 2462 G probe has specific affinity for nuclear proteins. Supershift assay was performed at the same time to determine whether NF1 could bind this sequence containing 2462 G variant. As shown in Fig. 6B, the presence of anti-NF1 antibody in the sample containing nuclear protein and labeled 2462 G probe resulted in a clear supershift complex of the band (Fig. 6B, lane 7). These results indicate that the presence of 2462 G or 2462 A in the proximal promoter region of PPP2R2D may affect the binding of NF1 to its target region in a sequence-specific manner.

NF1 plays an important regulatory role in PP2A-Bd gene expression
Next, we designed cotransfection experiments in L02 cell to provide direct evidence of identified transcription factor NF1 regulating PPP2R2D gene promoter activity and PP2A-Bd expression. As shown in Fig. 7A, overexpression of the NF1/B induced a dose-dependent increase in the luciferase activity of 2462 G luciferase reporter in L02 cells. However, this dosedependent response was abolished in 2462 A luciferase reporter. Furthermore, we also detected the PPP2R2D mRNA and protein expression levels of L02 cells (2462 GG genotype) transfected with NF1/B encoding plasmids by using real-time RT-PCR and western blot analysis. As shown in Fig. 7B and Fig. 7C, the mRNA and protein expression levels of endogenous PPP2R2D were enhanced in a dose-dependent manner after overexpression of 250 ng to 1000 ng exogenous NF1/B expression plasmid. Together, these results demonstrate that NF1/B displays a positive regulation on the transcription of PPP2R2D gene and upregulates the expression of PP2A-Bd.

Discussion
The results of this study demonstrate, for the first time, that genetic variants and their haplotypes in the 59-flanking region of PPP2R2D gene can contribute to variations in the transcriptional regulation of PP2A-Bd. The identified four known genetic variants, three haplotypes in the promoter region and their distribution in the population provide fundamental information for genotyping PPP2R2D in the southern Han Chinese population. Experimental data indicate that the functional variant 2462 G.A involves in the transcriptional regulation of PPP2R2D and the transcription factor NF1 interacts with the 2462 G.A variant present within the proximal promoter region of PPP2R2D. Moreover, promoter fragments harboring different haplotypes exhibit different promoter activities, which imply that genetic variants are likely to regulate PP2A-Bd expression and, consequently influence its dephosphorylation functions.
Protein phosphatase 2A (PP2A) is one of the most abundantly expressed serine/threonine protein phosphatases. A large body of evidence suggests that PP2A is a tumor suppressor [23,24,25], and that it plays critical roles in the regulation of many cellular processes [26,27,28]. The regulation of regulatory B subunits of PP2A (PP2A-Bs) ensures that dephosphorylation of specific substrates occurs in distinct cellular compartments and within specific signal transduction pathways [8,29,30]. The functions of B family subunit (PR55, B55) include the regulation of cytoskeletal dynamics and mitogen-activated protein kinase (MAPK) signaling, and apoptosis [10,31,32,33]. Silencing of the B-family regulatory subunit, Bd, leads to hyperactivation of the extracellular signalregulated kinase (ERK) caused by constitutively active MEK1. To date, there have been no publications characterizing the regulatory regions of the human PPP2R2D gene. More importantly, little work has been done to investigate potentially functional genetic variants in the 59-flanking region of the regulatory Bd subunit gene and characterize their distribution in different populations. In order to gather more information about regulatory polymorphisms (rSNPs) in the 59-flanking region of the Firstly by analyzing the variant characteristics in target population, we found that these four genetic variants, in the 59flanking region (21775 nt to +61 nt) of the PPP2R2D, differ greatly in frequency from those in other ethnic populations in the HapMap phase 3 database. In this southern Han population, the allele frequencies of the 21684 A.G, 21610 A.T, 21196 C.T, and 2462 G.A variants were determined to be 0.993:0.007, 0.943:0.057, 0.979:0.021, and 0.979:0.021, respectively; the minor G, T, T, and A alleles (for 21684 A.G, 21610 A.T, 21196 C.T, and 2462 G.A, respectively) are rare and only heterozygous genotypes were found in this population. Interestingly, our study shows that the MAFs of the variants differ between our study population and the Han Chinese in Beijing, China. This is consistent with a previous study showing that the greatest genetic differences among the Han Chinese are between the northern Han Chinese (N-Han) and southern Han Chinese (S-Han) [34]. Moreover, studies based on analyses of archeological, anatomical, linguistic, and genetic data have consistently suggested the existence of a significant boundary between the northern and southern populations in China [35]. These results suggest that genetic variants differ in their occurrence across ethnic groups. Our data not only update the HapMap database, but they also reveal significant differences between different ethnic subgroups.
SNPs and their characteristic distribution in a target population are valuable resources for investigating the genetic basis of individual susceptibility to damage or even disease. These variants can now serve as markers for fine-scale genetic mapping experiments and genome-wide association studies (GWAS) [36]. Both tumorigenicity and functional haploinsufficiency have been attributed to mutations in the A subunit of PP2A (PP2A-A). These mutations promote degradation of A subunit [37,38]. With regard to the B subunits of PP2A (PP2A-Bs), a recent study demonstrated that a SNP in PPP2R5E (the gene coding for B56e, a B56 family regulatory subunit of PP2A) is associated with human soft tissue sarcoma [39]. Furthermore, a case-control study using haplotype analysis of tagged SNPs in the PPP2R2A (coding for PP2A-Ba) and PPP2R1A (coding for PP2A-Aa) genes demonstrated that certain genotypes were protective for breast cancer, while others modified the risk for women with proliferative breast lesions [40]. A candidate gene association study identified a SNP (rs319217, A.G) in the PPP2R2B gene (coding for PP2A-Bb) that manifests allelic differences in the cellular responses to treatment with chemotherapeutic agents [41]. In addition, a study suggested that a CAG repeat polymorphism in the PPP2R2B gene may be functional and may, in part, play a role in conferring susceptibility to Alzheimer's disease (AD) and essential tremor (ET) in Taiwan [42]. A similar study from Kimura et al. suggested that CAG repeat lengths in the PPP2R2B gene may be potential genetic markers for AD susceptibility in the Japanese population [43]. Lin and colleagues have reported that the CAG repeat length, in addition to the CREB, SP1, and TFAP4 proteins, plays a critical role in regulating PPP2R2B promoter activity [44]. More recently, a study indicated that the A variant of a SNP (rs319217 A.G) in the PPP2R2B gene is a marker for improved prognosis of breast cancer [45]. However, the genetic variants in the PPP2R2D gene remain largely unknown. In the light of previous reports implicating PP2A-Bd in the regulation of the cell cycle and the dephosphorylation of specific substrates, we designed experiments aimed at functionally analyzing the identified variants in the 59-flanking region of PPP2R2D.
We designated a region of 601 bp (2540 nt to +61 nt) as the core proximal promoter of PPP2R2D because it displayed the highest transcriptional activity in luciferase assays. We constructed proximal promoter fragments mirroring the PPP2R2D allelotypes, including either one or the other allele of the 2462 G.A variant, and tested them for transcriptional activity. Promoter activity assay revealed that the most common allelotype (F3H1), containing the G 2462 polymorphism, exhibited a higher transcriptional activity. The other construct bearing A 2462 displayed a lower level of promoter activity. Our results suggest that different allelotypes composed of the 2462 G.A variant influence the transcriptional activity of PPP2R2D.
To determine whether the transcription of PPP2R2D was regulated by different haplotypes, we constructed the distal fragments (F1) in the 59-flanking region bearing different haplotypes (F1Hs) found in the Han Chinese population. Promoter activity analysis revealed that the most common haplotype, F1H1 (namely A 21684 /A 21610 /C 21196 /G 2462 ), with an overall frequency of 92.1%, showed the highest transcriptional activity. The other two variant haplotypes, F1H2 (A 21684 / T 21610 /C 21196 /G 2462 ) generated less transcriptional activity than F1H1, while F1H3 (A 21684 /A 21610 /T 21196 /A 2462 ) exhibited the least activity of the three constructs. This result suggested that transcription may be more tightly regulated in F1H2 than in F1H3. Consistent with the result of core proximal promoter (601 bp, 2540 nt to +61 nt) activity assay, the luciferase reporter gene assays also showed that haplotypes bearing the 2462 G allele had higher promoter activity than those that carried the 2462 A allele. Our results implied that different haplotypes composed of different genetic variants can influence the transcriptional activity of the PPP2R2D.
Bioinformatic analysis revealed that the 2462 G.A variant is located in a recognition sequence for NF1, which implies that this portion of the 59-flanking sequence may be important for NF1mediated regulation of gene expression. The NF1 family is encoded by four genes, NF1/A, /B, /C, and /X, which express a large number of splice variants that can act either as transcriptional activators or repressors depending on cellular and promoter context [46,47,48,49]. The proteins of NF1 family play a wide role in the replication of several vital DNA molecules and in the transcription of many cellular genes. A significant role for NF1 in regulating cellular growth, hormonal induction and repression, and oncogenic processes has been also described [46]. In this study, an EMSA revealed that the NF1 forms a complex with a DNA probe containing 2462 G, but not with an identical probe in which G at position 2462 nt was substituted for a 2462 A allele. Supershift and cross-competition assay demonstrated that human NF1 binds to the proximal 59-flanking region of PPP2R2D that contains 2462 G variant. Introduction of exogenous NF1/B expression plasmid into L02 cells induced dose-dependent increase of the PPP2R2D promoter activity and also upregulated the expression of the endogenous PPP2R2D. Because NF1 can act as an activator or repressor of the same gene depending on the cell type or promoter context, it can be suggested that the variant 2462 G.A with different affinity for NF1 might help to alter transcriptional activity and, thus, increase mRNA and protein expression of PPP2R2D. In accordance with this hypothesis, our results further confirmed that the NF1 may implicate in the positive control of PP2A-Bd expression.
This study reveals, for the first time, the occurrence of variations and distinct haplotypes in the 59-flanking region of PPP2R2D in the southern Han Chinese population. We provide data to indicate that the 2462 G.A variant can functionally regulate the transcriptional activity of PPP2R2D. The data provide support for further investigation into the molecular and cellular mechanisms behind the significant allelic differences between haplotypes and the effect of different haplotypes on the regulation of PPP2R2D. In the future, our haplotype information should allow for more efficient association studies, permit comparisons with studies of other ethnicities and help to identify people who are susceptible to damage or disease linked to these genes.