View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Elsevier - Publisher Connector

Genomics 95 (2010) 299–305

Contents lists available at ScienceDirect

Genomics

journal homepage: www.elsevier.com/locate/ygeno

Polymorphisms in the human glutathione transferase Kappa (GSTK1) promoter alter expression

Alison J. Shield, Tracy P. Murray, Jean Y. Cappello, Marjorie Coggan, Philip G. Board ⁎

Molecular Genetics Group, Division of Molecular Biosciences, John Curtin School of Medical Research, Australian National University, Acton ACT 2601, Australia

article info abstract

Article history: The level of glutathione transferase Kappa (GSTK1-1) has been correlated with (Liu et.al. 2008 PNAS Received 16 November 2009 105: 18302-7) and a polymorphism in the hGSTK1 promoter has been associated with insulin secretion and Accepted 17 February 2010 fat deposition (Gao et al 2009 Endocr J 56: 487-94). We searched for additional polymorphisms that may Available online 26 February 2010 influence GSTK1-1 function or expression. Two SNPs were identified in the 5’ non-coding region. A SNP at -1308 that occurs in Chinese subjects is predicted to eliminate a FXR/RXR transcription factor-binding site Keywords: and causes a 55% increase in transcription rate in HepG2 cells and a 59% decrease in HEK293 cells. These data Glutathione transferase Kappa fi Single nucleotide polymorphisms suggest that the impact of this polymorphism is complex and tissue speci c. A SNP at -1032 alters a Promoter methylation site and represses transcription by 38%. These observations provide the first functional insight into genetic factors that regulate hGSTK1 expression and may directly influence insulin secretion and fat deposition. © 2010 Elsevier Inc. All rights reserved.

Introduction consistent with the structural similarity between GSTK1-1 and the disulfide bond forming protein (DsbA) in E.coli [9]. This study also The Kappa class of glutathione S-transferases (GSTK) was first reported that GSTK1-1 (termed DsbA-L) expression in adipose tissue identified in mitochondrial matrix from rat liver [1]. Further studies of was negatively correlated with obesity in humans and mice. Since human GSTK1-1 have confirmed not only its mitochondrial localisa- expression has been correlated with insulin resistance, tion but also its presence in [2–4].Thisdefined obesity and diabetes [10–13] it is possible that the regulation of subcellular localisation makes GSTK1-1 distinct from other soluble GSTK1-1 expression may be an important factor in these metabolic GSTs that are found primarily in the cytosol. disorders. Significantly, a polymorphism in the 5’ non-coding region of Recent structural studies of GSTK proteins have shown that they the human GSTK1 gene (hGSTK1) was recently found to be associated are more closely related to the bacterial proteins 2-hydroxychro- with insulin secretion and fat deposition in Chinese subjects [14]. mene-2-carboxylate (HCCA) isomerase and disulphide-bond-forming Little is known about the hGSTK1 gene beyond its basic oxidoreductase than to other cytosolic GSTs [5–7]. Despite structural organization and location on 7q34-35 [5]. In the light divergence, human GSTK1-1 does have significant glutathione (GSH) of the previous reports implicating GSTK1-1 in the metabolic conjugating activity towards halogenated aromatics such as 1-chloro- regulation of obesity and diabetes, we have investigated the 2,4- dinitrobenzene (CDNB) and some peroxidase activity with frequency and function of polymorphisms in the hGSTK1 gene. organic hydroperoxides [4,5]. Although a preferred endogenous Although we did not find any genetic variation in the coding sequence substrate has not been identified, it has been speculated that we identified two SNPs in the 5’ flanking region (5'FR) that appear to GSTK1-1 might be important in the maintenance of the redox balance alter hGSTK1 transcription. In particular we found that the SNP studied in subcellular organelles [4]. The knockdown of GSTK expression in by Gao et.al. [14] disrupts an FXR(NR1H4)/RXR(RXRA) binding site C elegans is reported to result in a significant decrease in respiration and has a strong but differing impact on basal level expression in rate and a change in fatty acid metabolism [8]. Both observations are HepG2 and HEK293 cells. consistent with a role in mitochondria and peroxisomes. A role for GSTK1 in the post translational disulfide mediated Results multimerization of adiponectin has been proposed and this is SNP discovery

⁎ Corresponding author. Molecular Genetics Group, Division of Molecular Biosciences, In silico analysis of several databases revealed seven putative SNPs John Curtin School of Medical Research, PO Box 334, Canberra, ACT 2601, Australia. ’ Fax: +61 2 61254712. within the protein coding region of GSTK1: three SNPs in the 5 E-mail address: [email protected] (P.G. Board). flanking region (5'FR), plus nine SNPs in intronic and 3’ untranslated

0888-7543/$ – see front matter © 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.ygeno.2010.02.007 300 A.J. Shield et al. / Genomics 95 (2010) 299–305

Table 1 Characterisation of GSTK1 polymorphism frequencies in three ethnic populations.

Gene Base Base Change Amino Mode of Discovery2 or Restriction Frequency of variant allele Position relative Acid dbSNP database code WT VAR European ancestory Chinese African to ATG1 Change

5'FR -1308 G T - sequencing rs1917760 ND* 0.20* ND* 5'FR -1032 G C - sequencing BsrBI 0.00 0.00 0.20 5'FR -765 G C - rs745890 ND ND ND 5'FR -255 T del - rs35125014 ND ND ND Intron 1 C G - rs17164171 0.05* 0.00* 0.03* 4 317 T C V106A web trace BstU1 ND ND ND Exon 4 347 A C K116T EST/blast BsmA1 ND ND ND Exon 4 355 C G R119G EST/blast Xma1 ND ND ND Exon 4 370 C T R124C EST/blast BstU1 ND ND ND Intron 5 del T rs11296238 - - - Intron 5 C T rs10275885 - - - Intron 5 C T rs2367807 - - - Intron 5 C T rs2367808 - - - Intron 5 C T rs13311357 - - - Exon 7 543 T A F181L EST/blast Bsu361 ND ND ND Exon 7 549 C G L183L EST/blast Fnu4H1 ND ND ND Intron 7 C G - rs7803893 0.00* 0.00* 0.08* Exon 8 673 A G R225G web trace rs11552016 Hpa2 ND ND ND Exon 8 707 C G rs1043436 - - - Exon 8 777 A G - rs7786667 0.00* 0.00* 0.00*

1For 5’ flanking and coding SNPs the position relative to the GSTK1 ATG is given along with amino acid change and diagnostic restriction digest where applicable. 2Polymorphisms found by searching the dbSNP database [23], Genome Annotation Initiative (GAI) [31] database, alignments of expressed sequence tags or direct sequencing are listed. The numbers given under mode of discovery refer to records in the NCBI dbSNP database; web trace indicates that a sequencing trace was found in the GAI database. *Starred frequencies indicate genotyping data from the HapMap project [32]. ND indicates not detected in our study.

regions (3'UTR; Table 1). Coding SNPs and 5’ flanking SNPs were detect additional novel polymorphisms. Two SNPs were identified by selected for further study because it is possible to characterise their this process, a transition of G to T was found at position -1308 effect on the GSTK1-1 activity or protein expression. Because (numbered from the ATG) in 2 Chinese subjects and a G to C transition predicted coding region SNPs were in 4,7 and 8 these exons at -1032 was detected in 2 African individuals. No evidence was found were amplified from 194 individuals coming from ethnically diverse for the proposed G-C polymorphism at -765 (rs7458960) or the T backgrounds, including 94 Australian Caucasian, 50 African and 50 deletion at -255 (rs35125014) described in dbSNP build 125; see Southern Chinese individuals. The PCR products were then subjected Table 1. to restriction enzyme digestion to test for the predicted SNPs (see The -1308 SNP has previously been described (rs1917760) with a T Table 2). We were unable to confirm any of these coding region SNPs allele frequency of approximately 0.20 in Asian populations. The -1032 in our samples. SNP has not been previously reported. Genotype frequencies for the Approximately 1500 bp of the 5'FR for the hGSTK1 gene was -1032 SNP were confirmed in Australian European, African and amplified from the DNA of 10 Australian Europeans, 10 Africans and Southern Chinese population samples (N=85, 69, and 64 respectively) 10 Southern Chinese individuals. These amplicons were sequenced to by BsrBI RFLP analysis. The C allele had a frequency of 0.20 in the African validate the SNPs identified by in silico analysis and to potentially population sample (see Table 1).Basedonpopulationspecificgenotype

Table 2 Oligonucleotides used for PCR amplification, sequencing, and electrophoretic mobility shift assays (EMSA).

Region Oligonucleotides1 Restriction RFLP temp

Forward PCR primers Reverse PCR primers

5’ flanking region -1556: ttctgccccagtttaggaga 49: ggacagcacgtcatagaaga -1197: gccagctaaacaggagacca -726: attgcatccccaaccaattagca BsrBI 37 °C Exon 4 caaggaagagctagtcctc ctcacccttgaccagacg BstU1 60 °C BsmA1 55 °C Xma1 25 °C Exon 7 gtaactgctttctccaggc taacttacttacccagcagg Bsu361 37 °C Fnu4H1 37 °C Exon 8 ctgagaacagtgtgcagag cgagtcccagtgccttatg Hpa2 37 °C

Forward sequencing primers Reverse sequencing primers

5’ flanking region -486: agaattcaggcccctctcat -847: ctgaggtgggtggatcactt -745: taattggttggggatgcaat -338: ccttggtcattcctgagcat

EMSA probes2

-1308 SNP actgcagagcaggacagtcaccct FXR/RXR aggtcaatgacct Non specific tattacccacagcgtagccaaattg

1Some of the oligonucleotides listed as being PCR primers were also used for sequencing. 2The EMSA probes sequences shown are for the wild type sense probe with the mutant position in bold and underlined. For the -1308 SNP the variant nucleotide is a T. FXR/RXR indicates a consensus sequence for a prototypic FXR/RXR binding site. A.J. Shield et al. / Genomics 95 (2010) 299–305 301 frequencies these SNPs do not appear to be linked with other polymorphisms in the immediate region spanning GSTK1.

Functional characterization of promoter variants

Luciferase reporter assays were used to determine if the 5'FR SNPs cause changes in GSTK1 expression. The human kidney (HEK293) and liver (HepG2) cell lines used in this study have been found to express endogenous GSTK1-1 (Fig. 1A) and hence should contain appropriate transcriptional elements needed to drive expression of the hGSTK1 promoter. As shown in Fig. 1B, the -1308T allele caused a 59% decrease (p=0.002) in the HEK293 cell line and a 55% increase (p=0.054) in HepG2 cells. The -1032 C allele caused at least a 38% decrease in activity for both of the cell lines (pb0.001).

Characterisation of the -1308T allele Analysis by MatInspector software [15] predicts that the -1308T allele may eliminate a binding site for the retinoid X receptor (RXR) and farnesoid X receptor (FXR) transcription factors. Western blot analysis confirmed that RXR and FXR are present in HEK293 and HepG2 cells (Fig. 2A). RXR was readily detected in HEK293 and HepG2 cells. In contrast endogenous FXR was not detectable in HEK293 cells; Fig. 2. The hGSTK1 promoter was repressed by the farnesoid X receptor (FXR) agonist expression of endogenous FXR in HepG2 was noted only when the chenodeoxycholic acid (CDCA). (A) Western blot showing endogenous expression of blot was exposed for an extended period (Fig. 2A). This confirms a FXR (top panel) and retinoid X receptor (RXR; bottom panel) in HepG2 and HEK293 cell lines. The lane labelled RRL is a positive control showing the in vitro synthesis of the FXR report by Chen et al. [16] that FXR is expressed in HepG2 but not or RXR proteins using a rabbit reticulocyte lysate. HEKov indicates HEK293 cells which HEK293 cells. are over-expressing recombinant FXR or RXR protein. HepG2* is an overexposure of the The data in Fig. 1 indicate that there are significant allelic differ- film to demonstrate the low levels of expression for endogenous FXR in the HepG2 ences for GSTK1 promoter function in HEK293 and HepG2 cells. These cell line. (B) HepG2 cells which had been transfected with the WT-pXPG or -1308T data suggest that FXR and RXR may repress the WT GSTK1 promoter in allele-pXPG constructs were treated with CDCA 24 hours after transfection. The fi resulting luciferase activity has been standardised to the solvent control (treated with HepG2 cells. When the FXR/RXR binding site is modi ed by the -1308 50% DMSO). Values are the mean±SEM of at least triplicate data; stars indicate pb0.05 SNP there is a significant increase in luciferase activity implying that compared with the solvent control by ANOVA. the promoter is no longer repressed by transcription factors binding to that site. To further examine this possibility we treated HepG2 cells with the FXR agonist CDCA to ensure that FXR and RXR were being dose responsive. This suggests that the -1308 SNP does not actively expressed. As we expected CDCA caused further repression of completely eliminate FXR/RXR binding. the WT GSTK1 promoter. Treatment with CDCA was dose responsive, In contrast to the results obtained with HepG2 cells, the data in Fig. 1 resulting in significantly decreased levels of activity for the WT GSTK1 indicate that the -1308T allele causes repression of luciferase activity in promoter (Fig. 2B). CDCA treatment also decreased activity of the HEK293 cells. This is not consistent with the results obtained with -1308 SNP construct by 20% (p=0.02; Fig. 2B) although this was not HepG2 cells however this result may reflect the absence of FXR expression in HEK293 cells. To further evaluate whether the -1308 SNP modulates GSTK1 expression via RXR and FXR binding we investigated the binding of GSTK1 WT and -1308 SNP DNA probes (see Table 2) to nuclear proteins using electrophoretic mobility shift assays (EMSA; Fig. 3). A FXR/RXR consensus sequence labelled with 32P was initially bound to HEK293 and HepG2 nuclear proteins to determine which nucleopro- tein complex had the same mobility as recombinant FXR and RXR expressed using rabbit reticulocyte lysate (RRL) or overexpressed in HEK293 cells (Fig. 3A; band marked F/R). Different mobility shift patterns were noted in HEK293 and HepG2 cells indicating that there are differences in the composition of the protein complexes that interact with the FXR/RXR binding site in these cell lines. When GSTK1 WT or -1308 SNP DNA probes were bound to HEK293 and HepG2 nuclear proteins up to four additional complexes were noted (Fig. 3B); complex 1 was not apparent in HEK293 cells. The formation of many of these complexes was prevented by addition of increasing amounts of the RXR antibody (Fig. 3C; starred); presence of antibody also caused a supershift for some of the upper bands. We were unable to obtain similar results using a FXR antibody (data not shown). Competition studies using up to 100 times excess of unlabelled WT Fig. 1. Variant hGSTK1 alleles influence promoter activity in the HepG2 and HEK293 cell probe or an unlabelled -1308 SNP probe confirmed the specificity of lines. (A) Western blot showing endogenous expression of GSTK1 in HepG2 and these complexes (Fig. 3D). Interestingly no difference in the pattern of HEK293 cells lines. (B) Relative luciferase activity for constructs containing either the nucleoprotein complexes was noted for the -1308 SNP DNA probe -1308T allele or -1032 C allele compared with the wild type (WT) construct. Error bars indicate the SEM for a minimum of 3 independent assays. Starred bars indicate pb0.05 (Fig. 3B) although there appears to be a difference in the relative compared with the WT by ANOVA. intensity for complex 1 and 2. 302 A.J. Shield et al. / Genomics 95 (2010) 299–305

Fig. 3. FXR and RXR bind to a FXR/RXR binding site in the hGSTK1 promoter. Electrophoretic mobility shift assays (EMSA) were used to confirm the binding of nuclear proteins to oligonucleotides. (A) A 32P end-labelled double-stranded probe containing the consensus FXR/RXR binding sequence (Table 2) was incubated with nuclear extracts from the HEK293 (HEK) and HepG2 (Hep) cell lines. Protein that had been synthesised in vitro using rabbit reticulocyte lysates (RRL) or nuclear extracts of the HEK293 cell line over- expressing FXR or RXR were used as positive controls. The nucleoprotein complex identified in this gel as binding to FXR/RXR is marked F/R in all panels. (B) HepG2 and HEK293 nuclear extracts were incubated with a 32P end-labelled double-stranded probe containing either the GSTK1 WT FXR/RXR binding sequence (left) or this sequence modified to contain the -1308 SNP (right). The resulting nucleoprotein complexes are labelled F/R or 1-4. Probe sequences are given in Table 2. (C) HepG2 nuclear extracts were preincubated with RXR antibody for 30 min prior to addition of radiolabelled WT probe. NE indicates nuclear extract in the absence of antibody with subsequent lanes containing increasing concentrations of antibody (Ab). * indicates the disappearance of complex in the presence of antibody while ↑ indicates supershift of high molecular weight complexes. (D) Nuclear extracts from the HepG2 cell line were incubated with radiolabelled GSTK1 WT probe. The specificity of binding was investigated by incubating the nuclear extract for 30 min (prior to the addition of radiolabelled probe) with unlabelled oligonucleotide probes representing GSTK1 WT, the -1308 SNP or an unrelated (NSP) probe. The respective probes are shown in increasing concentration from left to right (at a ratio of 1:100, 1:10 and 1:1). In all panels labelled probe in the absence of a protein source is shown as a negative control (bl).

Characterisation of the -1032 C allele the 5'UTR) plus nine in intronic and 3'UTR regions. To maximise Bioinformatic analysis of the region surrounding position -1032 success in verifying them we deliberately chose SNPs that were did not predict any transcription factor binding sites in either the wild present in two or more EST sequences or where there was clear type or variant sequences; however this SNP does appear to alter a experimental evidence, such as a peak present in a sequencing trace. CpG methylation site. To investigate whether this SNP does introduce We genotyped the coding region and 5'FR SNPs and were able to a methylation site we used M.SssI CpG methylase to methylate verify only one of the 5'FR region SNPs (rs1917760). The HapMap hGSTK1 WT and -1032 SNP DNA fragments. Differences in DNA project was similarly unable to verify a 3'UTR SNP (rs7786667) in a fragments after digestion with the methylation sensitive enzyme AciI population comprising of individuals from Yoruba African (60), indicated that the variant allele is susceptible to methylation (Fig. 4A). Japanese (44), Han Chinese (45) and Caucasian American (59) In other data not shown we found that global methylation of the backgrounds. These results highlight the number of data-mined 1500 bp GSTK1 promoter caused almost complete repression of both SNPs that are difficult to confirm. It has been estimated that the wild type and variant -1032 C alleles. This data suggests that approximately half of the SNPs found in databases, such as dbSNP, hGSTK1 promoter activity can be globally influenced by methylation have not been independently validated [17,18]. but does not explain the specific repression observed for the -1032 C It is interesting that so few SNPs were found in the GSTK1 gene. allele in Fig. 1. Many estimates have been made for SNP density in the with one SNP every 1000 base pairs being a commonly Discussion quoted value [19]. Large resequencing studies suggest that this figure is closer to 1 SNP every 500 bp however most of these SNPs have allele The GSTK1 gene is approximately 5.7 kb in size with 8 exons and 7 frequencies of less than 5% and 96% of them are found in the non- introns [5]. Databases predict seven polymorphisms in the coding coding regions of the gene [20]. Without entirely resequencing the region, three within the 5’ flanking region (1600 bp of 5'FR including GSTK1 gene in a large number of subjects to obtain a complete survey A.J. Shield et al. / Genomics 95 (2010) 299–305 303

Fig. 4. The SNP at -1032 in the GSTK1 promoter changes a methylation site. (A) Differences in methylation status were demonstrated by in vitro methylation of the GSTK1 WT and -1032 SNP promoters followed by overnight restriction digestion with Bgl2 and AciI. Methylation (m) blocked the AciI restriction site to give the same electrophoresis pattern for both the WT and SNP constructs. (B) Schematic showing the expected banding patterns for an AciI/Bgl2 restriction digest before and after methylation. Bands less than 200 bp are not visible on this gel. of polymorphisms and their frequency it is difficult to accurately by methylation sensitive restriction digest. However, when in vitro estimate SNP density for GSTK1. However with no SNPs validated in methylated plasmids were used to drive luciferase expression the WT the coding region and only 4 SNPs seen in non-coding regions in four and SNP constructs were equally repressed suggesting that the ethnically distinct populations (present results and HapMap data), it hGSTK1 promoter can be silenced by more global methylation. If this appears that GSTK1 has fewer than the average 4.6 SNPs per kilobase site plays a role in the regulation of hGSTK1 expression it may be the estimated in 75 by Halushka et al [21]. subject of site specific methylation. Our studies confirmed the presence of two SNPs in the 5'FR. The G This study investigated the polymorphisms in the hGSTK1 gene to T transition at -1308 that was previously studied by Gao et al [14] in and only found two SNPs in the 5'FR. We found functional differences relation to diabetes has only been observed in Chinese subjects so far. between these SNPs and the common wild type allele. In addition we The second 5'FR SNP at -1032 was discovered by direct sequence were able to demonstrate that the promoter region can be repressed analysis and was only found in African subjects. by methylation and chenodeoxycholic acid. These observations The T allele at -1308 resulted in an increase in expression for provide the first functional insight into the genetic factors that HepG2 cells and a significant reduction in expression for HEK293 cells. regulate GSTK1-1 expression and the potential mechanism by which This striking difference in response between cell lines suggests that the -1308T allele influences insulin secretion and fat deposition. this SNP may be an important genetic regulator of GSTK1-1 expression with effects on downstream processes that are modulated Materials and methods by GSTK1-1. This is consistent with the proposed role for GSTK1-1 in the post-translational processing of adiponectin, its association with Study population obesity [9] and the association of this variant GSTK1 allele with insulin secretion and fat deposition [14]. As a consequence of its position the The genomic DNA samples were studied with the approval of the -1308T allele may disrupt an FXR/RXR transcription factor-binding Human Ethics Committee of the Australian National University and site. We found differences in the expression of FXR and RXR between were obtained with informed consent from normal blood donors after HepG2 and HEK293 cells and these may contribute towards the ethics approval by local blood transfusion authorities in Canberra differences in hGSTK1 promoter activity that we observed. However Australia, Durban South Africa and Hong Kong China. repression of the -1308T allele by CDCA indicates that it is still a functional FXR/RXR binding site and EMSA studies found that the Polymorphism identification and analysis -1308T allele binds to the FXR/RXR nucleoprotein complex and can compete successfully with the wild type -1308 C allele. Additionally Putative SNPs in the human GSTK1 gene were identified by analysis the presence of multiple nucleoprotein complexes binding to this of web-based sequence and polymorphism databases as previously transcription factor binding site suggests that it may not be described [23–25]. For promoter region SNP discovery, 1500 bp of exclusively binding to FXR and RXR. Clearly this FXR/RXR binding hGSTK1 5'FR DNA was amplified and sequenced from 10 Australian site is important in the regulation of GSTK1 expression but its role is European, 10 African and 10 Southern Chinese individuals. Confirma- complex and may vary considerably in different tissues. tion of novel SNPs in the coding region or in the 5'FR was undertaken FXR plays an essential role in the control of bile acid synthesis and by PCR followed by restriction fragment length polymorphism (RFLP; in partnership with the transcription factors proliferator- see Table 2) analysis in up to 100 individuals from each ethnic group. activated receptor (PPARA), vitamin D receptor (VDR), pregnane X Primers for the amplification of the hGSTK1 exons or regions of the receptor (NR1I2), constitutive andosterone receptors (NR1I3) and promoter were purchased from Sigma-Proligo (Castle Hill, NSW) and liver X receptor (NR1H2) also helps to regulate lipid homeostasis [22]. are listed in Table 2. SNPs found in the 5'FR were screened for changes Analysis of the hGSTK1 promoter region has identified at least one in putative transcription factor binding sites using MatInspector [15]. binding site for each of these transcription factors in the 5'FR. Given Potential methylation sites were identified using Methylator [26]. the subcellular localisation of GSTK1-1 to mitochondria and the peroxisome [4], and the association of GSTK knock down and fatty Promoter function assays acid metabolism in C. elegans [8] it is conceivable that GSTK1-1 is also involved in this regulatory cascade. Approximately 1500 bp of the wild type (WT) 5'FR of hGSTK1 The C allele at -1032 decreased luciferase activity by approxi- including the untranslated region of exon 1 was amplified using Pfu mately 40% when compared to the WT construct. We were able to Turbo (Stratagene, La Jolla CA) and subcloned into the pXPG firefly demonstrate the presence of an additional methylation site in the luciferase reporter plasmid [27]. Variant alleles were either generated -1032 C allele versus the WT allele using in vitro methylation followed using site-directed mutagenesis as described by Ho et al. [28] or by 304 A.J. Shield et al. / Genomics 95 (2010) 299–305 subcloning the variant allele from an affected individual. The integrity of antibody or 1 μL of unlabeled probe was incubated with the nuclear of the constructs was verified by sequence analysis undertaken in the extract for 30 min prior to the addition of the radiolabelled probe. Non Biomolecular Resource Facility at the John Curtin School of Medical specific probes (NSP) were randomly generated and checked to Research. ensure that they contained no putative transcription factor binding These reporter constructs (200 ng per well) were used to transfect sites using MatInspector [15]. the HEK293 and HepG2 cell lines using Lipofectamine 2000 according to the manufacturers protocol (Invitrogen, Mount Waverley VIC). Methylation analysis Cells (2×104) were co-transfected in quadruplicate with 20 ng plasmid encoding Renilla luciferase (pTK-RL; Promega, Sydney The pXPG reporter constructs were transformed into E.coli strain fi NSW) to control for transfection ef ciency. After 48 hr cells were JM110 to remove the potential for bacterial methylation of the assayed for luciferase activity using the Dual-Glo Luciferase Assay constructs. The WT and -1032 SNP alleles were then excised from the System (Promega, Sydney NSW) as described by the manufacturer. vector and gel purified. Purified inserts or whole plasmid were in vitro Each experiment was repeated at least three times independently and methylated using the CpG Methylase, M.SssI (NEB) for 6 hrs at 37 °C. differences in luciferase activity were evaluated using either t-tests or Similar to the mammalian enzyme Dnmt1, M.SssI methylates cytosine ANOVA where applicable. in the CpG dinucleotide [30]. Methylated and non-methylated DNA was digested overnight at 37 °C using a methylation sensitive Protein synthesis and Western Blot analysis restriction enzyme (BsrBI or AciI) and the restriction enzyme Bgl2 that is not influenced by methylation status. Methylation status was The FXR and RXR proteins were synthesised using the TNT Coupled determined by analysis of digestion patterns on a 1.5% agarose gel. ’ Reticulocyte Lysate System according to the manufacturers instruc- Where treated DNA was used for transfection, plasmids were purified fl tions (Promega, Sydney NSW). Brie y, pCMX-FXR or pCMX-RXR using the Wizard SV Gel and PCR clean-up system (Promega, Sydney expression vectors (from Dr David Mangelsdorf at UT Southwestern NSW). Medical Centre) were combined with TNT rabbit reticulocyte lysate (RRL), TNT T7 RNA polymerase, 40 nM amino acid mix, and RNasin Acknowledgments ribonuclease inhibitor at 30 °C for 90 min. [35S]methionine was added to a proportion of the reaction as a positive control. These expression We thank Dr Anneke Blackburn for helpful discussions. This work constructs were also used to transfect the HEK293 cell line using was supported by grants from the Australian Research Council and the Lipofectamine 2000. Protein synthesis was checked by SDS-PAGE of National Health and Medical Research Council, including an NHMRC the 35S labelled reactions or Western Blot. Antibodies for Western Blot Australian Biomedical (Peter Doherty) Fellowship [AS]. included rabbit anti-RXR (ΔN 197; sc-774) and goat anti-FXR (C-20; sc-1204) (Santa Cruz Biotechnology, Santa Cruz CA). For GSTK1, polyclonal antiserum was prepared in rabbits by immunization with References recombinant human GSTK1-1 that had been expressed in E. coli and [1] J.M. Harris, D.J. Meyer, B. Coles, B. Ketterer, A novel glutathione transferase (13-13) purified by means of an N-terminal His tag [5]. The specificity of the isolated from the matrix of rat liver mitochondria having structural similarity to antiserum was confirmed on a Western blot against purified class theta enzymes, Biochem. J. 278 (1991) 137–141. [2] I.R. Jowsey, R.E. Thomson, T.C. Orton, C.R. Elcombe, J.D. Hayes, Biochemical and recombinant GSTK1-1 (data not shown). genetic characterization of a murine class Kappa glutathione S-transferase, Biochem. J. 373 (2003) 559–569. Electrophoretic Mobility Shift Assays (EMSA) [3] R.E. Thomson, A.L. Bigley, J.R. Foster, I.R. Jowsey, C.R. Elcombe, T.C. Orton, J.D. Hayes, Tissue-specific expression and subcellular distribution of murine glutathione S- transferase class kappa, J. Histochem. Cytochem. 52 (2004) 653–662. Nuclear extracts were prepared from HEK293 and HepG2 cell lines [4] F. Morel, C. Rauch, E. Petit, A. Piton, N. Theret, B. Coles, A. Guillouzo, Gene and using a modified salt extraction method [29]. Cells at 80% confluence protein characterization of the human glutathione S-transferase kappa and – were harvested, pelleted by centrifugation at 1000×g for 5 min and evidence for a peroxisomal localization, J. Biol. Chem. 279 (2004) 16246 16253. [5] A. Robinson, G.A. Huttley, H.S. Booth, P.G. Board, Modelling and bioinformatics then washed in phosphate-buffered saline. Cell pellets were then studies of the human Kappa class Glutathione Transferase predict a novel third resuspended in 10 mM HEPES buffer (pH 7.9) containing 10 mM KCl, Glutathione Transferase family with homology to prokaryotic 2-hydroxychromene- – 0.1 mM EDTA, 0.1 mM EGTA, 0.4% NP40, 0.5 mM AEBSF (ICN 2-carboxylate (HCCA) Isomerases, Biochem. J. 379 (2004) 541 552. [6]J.E.Ladner,J.F.Parsons,C.L.Rife,G.L.Gilliland,R.N.Armstrong,Parallel Biomedical, Seven Hills, NSW) and 1 mM DTT, and incubated on ice evolutionary pathways for glutathione transferases: structure and mechanism for 15 min. The nuclear pellet was spun down for 30 sec and washed of the mitochondrial class kappa enzyme rGSTK1-1, Biochemistry 43 (2004) in 1 ml of the HEPES buffer described above. The nuclear pellet was 352–361. [7] J. Li, Z. Xia, J. Ding, Thioredoxin-like domain of human kappa class glutathione resuspended in 20 mM HEPES buffer (pH 7.9) containing 400 mM transferase reveals and structure similarity to the theta class NaCl, 1 mM EDTA, 1 mM EGTA, 1 mM DTT and 0.5 mM AEBSF and enzyme, Protein Sci. 14 (2005) 2361–2369. mixed on a rotary wheel for 15 min at 4 °C. Cellular debris was [8] E. Petit, X. Michelet, C. Rauch, J. Bertrand-Michel, F. Terce, R. Legouis, F. Morel, Glutathione transferases kappa 1 and kappa 2 localize in peroxisomes and pelleted by centrifugation and the supernatant stored at -20 °C until mitochondria, respectively, and are involved in lipid metabolism and respiration use. in Caenorhabditis elegans, Febs J. 276 (2009) 5030–5040. Double stranded oligonucleotides used as probes (Table 2) were [9] M. Liu, L. Zhou, A. Xu, K.S. Lam, M.D. Wetzel, R. Xiang, J. Zhang, X. Xin, L.Q. Dong, F. Liu, A disulfide-bond A oxidoreductase-like protein (DsbA-L) regulates adipo- generated by 5'end labelling 100 ng single stranded oligonucleotide nectin multimerization, Proc. Natl. Acad. Sci. U. S. A. 105 (2008) 18302–18307. 32 with [γ- P]dATP (PerkinElmer, Rowville VIC) using T4 polynucleo- [10] Y. Arita, S. Kihara, N. Ouchi, M. Takahashi, K. Maeda, J. Miyagawa, K. Hotta, I. tide kinase (New England Biolabs, Ipswich MA). Labelled single Shimomura, T. Nakamura, K. Miyaoka, H. Kuriyama, M. Nishida, S. Yamashita, K. stranded oligonucleotides were purified using QIAQUICK Nucleotide Okubo, K. Matsubara, M. Muraguchi, Y. Ohmoto, T. Funahashi, Y. Matsuzawa, Paradoxical decrease of an adipose-specific protein, adiponectin, in obesity, removal kit (QIAGEN, Doncaster, VIC) and allowed to anneal with the Biochem. Biophys. Res. Commun. 257 (1999) 79–83. complementary oligonucleotide at room temperature for 10 min. [11] A.J. Hanley, D. Bowden, L.E. Wagenknecht, A. Balasubramanyam, C. Langfeld, M.F. Double stranded labelled probe was incubated with 5 μg of nuclear Saad, J.I. Rotter, X. Guo, Y.D. Chen, M. Bryer-Ash, J.M. Norris, S.M. Haffner, μ Associations of adiponectin with body fat distribution and insulin sensitivity in extract (or 10 L RRL) for 30 min at room temperature in 20 mM nondiabetic Hispanics and African-Americans, J. Clin. Endocrinol. Metab. 92 HEPES pH 7.9, 5% glycerol, 1 mM EDTA, 1% NP40, 5 mM DTT, 1 mg/mL (2007) 2665–2671. BSA, and 1 μg poly(dI-dc). Samples were resolved on a 5% non- [12] C. Weyer, T. Funahashi, S. Tanaka, K. Hotta, Y. Matsuzawa, R.E. Pratley, P.A. Tataranni, Hypoadiponectinemia in obesity and : close association denaturing polyacrylamide gel; these gels were dried and subjected to with insulin resistance and hyperinsulinemia, J. Clin. Endocrinol. Metab. 86 autoradiography. For antibody and probe competition studies, 1-2 μL (2001) 1930–1935. A.J. Shield et al. / Genomics 95 (2010) 299–305 305

[13] T. Yamauchi, J. Kamon, H. Waki, Y. Terauchi, N. Kubota, K. Hara, Y. Mori, T. Ide, K. [21] M.K. Halushka, J.B. Fan, K. Bentley, L. Hsie, N. Shen, A. Weder, R. Cooper, R. Murakami, N. Tsuboyama-Kasaoka, O. Ezaki, Y. Akanuma, O. Gavrilova, C. Vinson, Lipshutz, A. Chakravarti, Patterns of single-nucleotide polymorphisms in M.L. Reitman, H. Kagechika, K. Shudo, M. Yoda, Y. Nakano, K. Tobe, R. Nagai, S. candidate genes for blood-pressure homeostasis, Nat. Genet. 22 (1999) 239–247. Kimura, M. Tomita, P. Froguel, T. Kadowaki, The fat-derived hormone adiponectin [22] C. Handschin, U.A. Meyer, Regulatory network of lipid-sensing nuclear receptors: reverses insulin resistance associated with both lipoatrophy and obesity, Nat. roles for CAR, PXR, LXR, and FXR, Arch. Biochem. Biophys. 433 (2005) 387–396. Med. 7 (2001) 941–946. [23] S.T. Sherry, M. Ward, K. Sirotkin, dbSNP-database for single nucleotide [14] F. Gao, Q. Fang, R. Zhang, J. Lu, H. Lu, C. Wang, X. Ma, J. Xu, W. Jia, K. Xiang, polymorphisms and other classes of minor genetic variation, Genome Res. 9 Polymorphism of DsbA-L gene associates with insulin secretion and body fat (1999) 677–679. distribution in Chinese population, Endocr. J. 56 (2009) 487–494. [24] P. Board, N. Tetlow, A. Blackburn, G. Chelvanayagam, Database analysis and gene [15] K. Cartharius, K. Frech, K. Grote, B. Klocke, M. Haltmeier, A. Klingenhoff, M. Frisch, discovery in pharmacogenetics, Clin. Chem. Lab. Med. 38 (2000) 863–867. M. Bayerlein, T. Werner, MatInspector and beyond: promoter analysis based on [25] N. Tetlow, D. Liu, P. Board, Polymorphism of human Alpha class glutathione transcription factor binding sites, Bioinformatics 21 (2005) 2933–2942. transferases, Pharmacogenetics 11 (2001) 609–617. [16] W. Chen, E. Owsley, Y. Yang, D. Stroup, J.Y. Chiang, Nuclear receptor-mediated [26] M. Bhasin, H. Zhang, E.L. Reinherz, P.A. Reche, Prediction of methylated CpGs in repression of human cholesterol 7alpha-hydroxylase gene transcription by bile DNA sequences using a support vector machine, FEBS Lett. 579 (2005) acids, J. Lipid Res. 42 (2001) 1402–1412. 4302–4308. [17] H. Edvardsen, G. Irene Grenaker Alnaes, A. Tsalenko, T. Mulcahy, A. Yuryev, M. [27] A.G. Bert, J. Burrows, C.S. Osborne, P.N. Cockerill, Generation of an improved Lindersson, S. Lien, S. Omholt, A.C. Syvanen, A.L. Borresen-Dale, V.N. Kristensen, luciferase reporter gene plasmid that employs a novel mechanism for high-copy Experimental validation of data mined single nucleotide polymorphisms from replication, Plasmid 44 (2000) 173–182. several databases and consecutive dbSNP builds, Pharmacogenet. Genomics 16 [28] S.N. Ho, H.D. Hunt, R.M. Horton, J.K. Pullen, L.R. Pease, Site-directed mutagenesis (2006) 207–217. by overlap extension using the polymerase chain reaction, Gene 77 (1989) 51–59. [18] D.E. Reich, S.B. Gabriel, D. Altshuler, Quality and completeness of SNP databases, [29] S.W. Jiang, N.L. Eberhardt, A micro-scale method to isolate DNA-binding proteins Nat. Genet. 33 (2003) 457–458. suitable for quantitative comparison of expression levels from transfected cells, [19] J.C. Stephens, J.A. Schneider, D.A. Tanguay, J. Choi, T. Acharya, S.E. Stanley, R. Jiang, Nucleic Acids Res. 23 (1995) 3607–3608. C.J. Messer, A. Chew, J.H. Han, J. Duan, J.L. Carr, M.S. Lee, B. Koshy, A.M. Kumar, G. [30] I. Nur, M. Szyf, A. Razin, G. Glaser, S. Rottem, S. Razin, Procaryotic and eucaryotic Zhang, W.R. Newell, A. Windemuth, C. Xu, T.S. Kalbfleisch, S.L. Shaner, K. Arnold, V. traits of DNA methylation in spiroplasmas (mycoplasmas), J. Bacteriol. 164 (1985) Schulz, C.M. Drysdale, K. Nandabalan, R.S. Judson, G. Ruano, G.F. Vovis, Haplotype 19–24. variation and linkage disequilibrium in 313 human genes, Science 293 (2001) [31] R. Clifford, M. Edmonson, Y. Hu, C. Nguyen, T. Scherpbier, K.H. Buetow, Expression- 489–493. based genetic/physical maps of single-nucleotide polymorphisms identified by the [20] M. Przeworski, R.R. Hudson, A. Di Rienzo, Adjusting the focus on human variation, genome anatomy project, Genome Res. 10 (2000) 1259–1265. Trends Genet. 16 (2000) 296–302. [32] A haplotype map of the human genome, Nature 437 (2005) 1299–1320.