15q11-13 gaba a receptor genes are normally … · gen es is com mon to bo th di sor der s. ......

15q11-13 GABAA receptor genes are normallybiallelically expressed in brain yet aresubject to epigenetic dysregulation inautism-spectrum disordersAmber Hogart, Raman P. Nagarajan, Katherine A. Patzel, Dag H. Yasui and Janine M. LaSalle*

Medical Microbiology and Immunology, Rowe Program in Human Genetics, School of Medicine,University of California, One Shields Avenue, Davis, CA 95616, USA

Received November 17, 2006; Revised and Accepted January 5, 2007

Human chromosome 15q11-13 is a complex locus containing imprinted genes as well as a cluster of threeGABAA receptor subunit (GABR) genes—GABRB3, GABRA5 and GABRG3. Deletion or duplication of15q11-13 GABR genes occurs in multiple human neurodevelopmental disorders including Prader-Willisyndrome (PWS), Angelman syndrome (AS) and autism. GABRB3 protein expression is also reduced inRett syndrome (RTT), caused by mutations in MECP2 on Xq28. Although Gabrb3 is biallelically expressedin mouse brain, conflicting data exist regarding the imprinting status of the 15q11-13 GABR genes inhumans. Using coding single nucleotide polymorphisms we show that all three GABR genes are biallelicallyexpressed in 21 control brain samples, demonstrating that these genes are not imprinted in normal humancortex. Interestingly, four of eight autism and one of five RTT brain samples showed monoallelic or highlyskewed allelic expression of one or more GABR gene, suggesting that epigenetic dysregulation of thesegenes is common to both disorders. Quantitative real-time RT–PCR analysis of PWS and AS samples withpaternal and maternal 15q11-13 deletions revealed a paternal expression bias of GABRB3, while RTT brainsamples showed a significant reduction in GABRB3 and UBE3A. Chromatin immunoprecipitation and bisul-fite sequencing in SH-SY5Y neuroblastoma cells demonstrated that MeCP2 binds to methylated CpG siteswithin GABRB3. Our previous studies demonstrated that homologous 15q11-13 pairing in neurons wasdependent on MeCP2 and was disrupted in RTT and autism cortex. Combined, these results suggest thatMeCP2 acts as a chromatin organizer for optimal expression of both alleles of GABRB3 in neurons.

INTRODUCTION

Human chromosome 15q11-13 is a complex locus containingimprinted genes involved in two neurological disorders—Prader-Willi syndrome (PWS) and Angelman syndrome(AS). PWS, a disorder characterized by hyperphagia leadingto obesity with mild to moderate mental retardation, resultsfrom paternal deficiency of 15q11-13. AS, characterized bysevere mental retardation with a lack of speech and excessiveinappropriate laughter, is caused by maternal 15q11-13deficiency (1). The majority of PWS and AS cases (!70%)result from de novo deletions spanning the 4 Mb region of15q11-13. Uniparental disomy (UPD) of the maternal chromo-some is another common cause (!25% of cases) of PWS,

while paternal UPD is an infrequent cause of AS. Mutationswithin the maternally expressed imprinted gene UBE3A, aresufficient to cause AS (2,3), and occur in !10% of cases (1).

Autism is a complex neurodevelopmental disorder charac-terized by impairments in social interaction, restricted andstereotyped behaviors, and deficits in language and communi-cation (4). Although there is a strong genetic component toautism, few candidate genes have been identified (5). Chromo-some 15q11-13 is one of the genetic loci implicated in autism,as maternal duplications of this region remain one of the mostcommon cytogenetic abnormalities found in cases of idio-pathic autism (6). Additional evidence from multiple linkageand association studies suggests 15q11-13 genes are involvedin autism (7–10).

# The Author 2007. Published by Oxford University Press. All rights reserved.For Permissions, please email: [email protected]

*To whom correspondence should be addressed. Tel: !1 5307547598; Fax: !1 5307528692; Email: [email protected]

Human Molecular Genetics, 2007, Vol. 16, No. 6 691–703doi:10.1093/hmg/ddm014Advance Access published on March 5, 2007

Rett syndrome (RTT) is a severe neurodevelopmental dis-order that primarily affects females and is caused by mutationsin the X-linked gene MECP2, the methyl CpG binding protein2 (11). RTT is characterized by normal early post-natal devel-opment followed by an abrupt loss of developmental mile-stones around 6–18 months of age leading to absence ofspeech, severe mental retardation and loss of purposefulhand movements (12). Like autism, RTT is categorized as apervasive developmental disorder (PDD) and is the only oneof the five PDDs with a known genetic cause (4). In arecent molecular study of post-mortem human brain, bothRTT and autism samples exhibited expression defects in two15q11-13 genes, UBE3A and GABRB3, suggesting that over-lapping pathways are dysregulated in these disorders (13).

Chromosome 15q11-13 contains a cluster of three GABAA

receptor subunit genes, GABRB3, GABRA5 and GABRG3,encoding receptor subunits for the inhibitory neurotransmittergamma-aminobutyric acid (GABA). The mouse orthologues,contained within a syntenic block on chromosome 7qB5, arebiallelically expressed in brain (14–16), however, conflictingdata exist on the imprinting status of the human 15q11-13GABAA receptor subunit genes (17–21). An early study ofGABRB3 expression in human hydatidiform mole andnormal villi tissues suggested preferential maternal geneexpression (20). Gabriel et al. (21) later detected 15q11-13GABAA receptor expression in multiple human–mousesomatic-cell hybrid lines containing a maternal or paternalhuman chromosome 15, suggesting that these genes are non-imprinted. Meguro et al. (17) performed microcell-mediatedcell transfer to create mouse A9 hybrids harboring either anormal maternal or paternal human chromosome 15. Thehuman paternally expressed imprinted genes retained normalimprinting in the A9 hybrids, however, the GABAA receptorgenes were only expressed from the paternal chromosome.Bittel et al. (18,19) demonstrated evidence in support ofpaternal expression bias of human GABAA receptor geneswith microarray analyses. Both studies, comparing deletionand UPD lymphoblastoid cell lines from PWS and ASpatients, found consistent paternal bias of GABRB3 andGABRA5 gene expression.

Additional suggestive evidence for imprinting of human15q11-13 GABAA receptors comes from allele-specific repli-cation timing differences of genomic sequences near theGABAA receptor genes (22), as replication timing differencesbetween parental homologues is a characteristic of imprintedgenes (23). Both parental chromosomal contributions arerequired for proper replication timing of the 15q11-13GABAA receptor genes, suggesting trans interactionsbetween parental homologs (24). Furthermore, homologousassociation of 15q11-13 domains occurs in normal lympho-cytes, yet is deficient in PWS and AS samples, suggestingthat the trans interaction involves the association of oppositelyimprinted homologues (25). Homologous association of15q11-13 GABRB3 alleles occurs in mature neurons withinbrain, and is deficient in RTT and autism (26).

Since the 15q11-13 GABAA receptor genes are implicatedin multiple human disorders, including autism, definitivelydetermining the imprinting status of these genes in humanbrain is critical to understanding their potential role in thepathogenesis of neurodevelopmental disorders. In this study,

single nucleotide polymorphisms (SNPs) within the codingregion of GABRB3, GABRA5 and GABRG3 were used toinvestigate allelic expression of transcripts from cerebralcortex. Since several imprinted genes, including UBE3A,exhibit neuron-specific imprinting (16,27–29), this imprintingstudy was performed using post-mortem human brain tissue.The human 15q11-13 GABAA receptor subunit genes weredetermined to have equal biallelic expression in all controlsamples, demonstrating that these genes are not imprinted innormal frontal cortex. In contrast, allelic expression analysisrevealed that four idiopathic autism samples and one RTTsample exhibited monoallelic or highly skewed allelicexpression of one or more GABAA receptor gene. Further-more, quantitative analysis of GABRB3 transcript in PWSand AS brain samples showed evidence for a paternalexpression bias when a biparental 15q11-13 contributionwas lacking. These results explain prior observations ofpaternal bias of GABRB3 expression, and suggest that epige-netic abnormalities in 15q11-13 that serve to dysregulateGABRB3 expression may be relatively common in autisticindividuals without detectable cytogenetic alterations.

RESULTS

15q11-13 GABAA receptor genes are normallybiallelically expressed in cortex, but monoallelicallyexpressed in some RTT and autism samples

In order to conclusively determine if 15q11-13 GABAA recep-tor genes exhibit imprinted expression, allelic expression wasanalyzed in control brain samples from frontal cortex, Brod-mann area 9. RTT and autism brain samples were alsoincluded in the analysis to test for evidence of epigeneticdysregulation. DNA and RNA was isolated from frozen post-mortem brain samples with a post-mortem interval (PMI),30 h (details in Supplementary Material, Table S1). Copynumber of chromosome 15q11-13 was analyzed by DNA flu-orescence in situ hybridization (DNA FISH) using a GABRB3probe (26), hybridized to brain samples on a tissue microarray(30). Normal disomic 15q11-13 contributions were verified inall control, RTT and AUT brain samples (data not shown).

To identify polymorphisms to discriminate allelicexpression, SNPs within the coding region of GABRB3,GABRA5 and GABRG3 were screened for heterozygosity ingenomic DNA samples from human cerebral cortex. Geno-types were determined by allele-specific restriction enzymedigestion of PCR products or by direct sequencing. Allele fre-quencies within control and autism spectrum disorder samplesanalyzed were comparable to values reported in the SNP data-base (http://www.ncbi.nlm.nih.gov) for rs2912582, rs140679and rs140682 (Table 1). Allelic frequency data was not avail-able for SNP rs20318 within GABRB3, however heterozygos-ity for this SNP was common within the population studied(Table 1).

Allelic expression in heterozygous samples was determinedby RT–PCR, followed by allele-specific restriction digestionor cDNA sequencing (Fig. 1). All 21 heterozygous controlsanalyzed exhibited equal biallelic expression for each infor-mative GABAA receptor subunit gene SNP (Table 2 and repre-sentative data in Fig. 1). In addition, one control sample also

692 Human Molecular Genetics, 2007, Vol. 16, No. 6

exhibited equal biallelic expression of the GABRB3 30-UTRSNP rs2912582 (data not shown). Most of the RTT (four offive) and four of eight autism brain samples analyzed demon-strated equal biallelic expression of the GABAA receptorsubunit genes, however four of eight informative autism andone out of five informative RTT samples demonstrated mono-allelic or highly skewed expression, summarized in Table 3with representative data shown in Figure 1. Expression ofGABAA receptor transcripts was not detectable in severalRTT and autism samples (N.D., Table 3), likely due to verylow levels of expression, previously described for GABRB3(13). Transcript quality was verified by RT–PCR of the house-keeping gene, GAPDH. All samples, including RTT 1748,RTT 1815 and AUT 5173 exhibited normal levels ofGAPDH by RT–PCR (data not shown), suggesting thatGABAA receptor transcript levels are specifically undetectablein these samples. The total number of control samples ana-lyzed (21) was greater than the number of RTT (5) orautism (8) samples analyzed for each gene, suggesting thatmonoallelic expression of 15q11-13 GABAA receptor genesis not variable in the general population but specific to RTTand autism samples.

15q11-13 GABAA receptor loss-of-biallelic expressioncorrelates with GABRB3 protein defectsin autism brain samples

Previously, we found a significant reduction in GABRB3protein in a subset of autism brain samples (13). To determineif the loss-of-biallelic (LOB) expression of GABAA receptorsubunit genes in autism samples correlates with reducedGABRB3 protein levels, we compared previous normalizedGABRB3 protein values determined by semi-quantitativeimmunoblot (13) between groups of individuals based onallelic expression status. Autism samples with LOBexpression of one or more 15q11-13 GABAA receptorsubunit gene had highly significantly reduced GABRB3protein levels compared with controls and biallelically expres-sing autism samples (Fig. 2A). There was no difference inGABRB3 protein between control samples and autismsamples with biallelic GABAA receptor expression(Fig. 2A). These results suggest that allelic expression statusof 15q11-13 GABAA receptor subunit genes predicts proteinexpression defects of GABRB3 in autism.

Quantitative real-time RT–PCR was then performed withthe same autism brain cDNA samples analyzed in Tables 2and 3 to determine if allelic expression status reflects the

total transcript level. In contrast to the clear differences inGABRB3 protein, GABRB3 transcript levels were variable inLOB expression autism samples. Two individuals (AUT1174 and AUT 5173) showed significantly reduced GABRB3transcript and two individuals (AUT 797 and AUT 4925)showed transcript levels not significantly different than con-trols (Fig. 2B). Quantitative RT–PCR was also performedfor GABRG3 and GABRA5. The expression levels ofGABRG3 and GABRA5 were apparently 1000-fold lowerthan GABRB3 when normalized to GAPDH, but controls,autism and RTT samples exhibited similar low levels ofexpression for both GABRG3 and GABRA5 (SupplementaryMaterial, Fig. S1).

Quantitative RT–PCR reveals a paternal biasin GABRB3 expression in 15q11-13 deletion samplesand significantly reduced GABRB3 expression in RTT

To further investigate imprinting effects on GABRB3expression, quantitative real-time RT–PCR analysis was per-formed on PWS and AS brain cDNA with paternal or maternal15q11-13 deletions, respectively (Fig. 3A). Differences inRNA quality were controlled by normalization to the house-keeping gene GAPDH. GABRB3 transcript was significantlylower than control samples in both PWS and AS deletionsamples, but PWS deletion samples had significantly lowerexpression than AS deletion samples, suggesting a paternalbias in gene expression when a single 15q11-13 region ispresent. PWS maternal UPD samples, despite having twocopies of GABRB3, exhibited significantly reduced GABRB3expression when compared with controls, further suggestinga paternal expression bias in the absence of a biparentalchromosome 15 contribution.

UBE3A, a maternally expressed imprinted gene, andSNRPN, a paternally expressed imprinted gene were analyzedas control genes for the quantitative RT–PCR assay (Fig. 3Band C). As expected when compared with the controls, the ASdeletion samples exhibited significantly reduced UBE3Aexpression, the PWS deletion samples showed no change inUBE3A expression, and the PWS maternal UPD sampleshad significantly higher UBE3A levels (Fig. 3B). As expected,SNRPN was absent in PWS deletion and UPD samples(Fig. 3C). AS deletion samples, however, exhibited a surpris-ing subtle but significant defect in SNRPN (Fig. 3C).

Both RTT and Mecp2-deficient mouse brain samples werepreviously shown to have significant GABRB3 proteindeficiencies (13). Mecp2-deficient mouse brain also exhibited

Table 1. Summary of allele frequencies for SNPs used in allelic expression analysis

Gene SNP Position Type Number ofindividuals

Alleles Controlfrequency

ASD frequency Total frequency dbSNP frequency

GABRB3 rs20318 exon 1a, aa25 syn, Pro-Pro Controls: 31 C/T T ! 0.129 T ! 0.184 T ! 0.15 N.D.ASD: 19 C ! 0.871 C ! 0.816 C ! 0.85 N.D.

GABRB3 rs2912582 30-UTR Non-coding Controls: 34 C/T T ! 0.909 T ! 1 T ! 0.943 T ! 0.954ASD: 19 C ! 0.091 C ! 0 C ! 0.057 C ! 0.046

GABRA5 rs140682 exon 2, aa80 syn, Val-Val Controls: 30 C/T T ! 0.419 T ! 0.333 T ! 0.388 T ! 0.4ASD: 19 C ! 0.581 C ! 0.667 C ! 0.622 C ! 0.6

GABRG3 rs140679 exon 3 aa129 syn, Thr-Thr Control: 29 C/T T ! 0.724 T ! 0.528 T ! 0.649 T ! 0.524ASD: 18 C ! 0.276 C ! 0.472 C ! 0.351 C ! 0.476

Human Molecular Genetics, 2007, Vol. 16, No. 6 693

a transcriptional defect in Gabrb3, therefore RTT brainsamples were included in the quantitative GABRB3 transcriptanalysis. As expected from the mouse studies, a significantdefect in GABRB3 transcript was observed in RTT brainwhen compared with control (Fig. 3A). Although conflictingdata have been reported on the dysregulation of Ube3a inMecp2-deficient mouse (13,31,32), a significant decrease inUBE3A expression was found in RTT brain (Fig. 3B), consist-ent with the decreased protein observed previously (13). Eventhough MeCP2 binds to the promoter of SNRPN (13,26), nodifference in SNRPN transcript was observed (Fig. 3C), con-sistent with the maintenance of normal imprinting of SNRPNin RTT (33), and normal expression in Mecp2-deficientmouse brain (13,34).

MeCP2 binds to intronic DNA sequences downstreamof the 50-CPG islands in GABRB3 and GABRG3

The GABRB3 expression defects in RTT and Mecp2-deficientmouse brain (13) have implicated MeCP2 in epigenetic regu-lation of 15q11-13 GABAA receptors. In addition, LOB

expression of 15q11-13 GABAA receptors in one RTTsample and multiple autism samples with previously describedMeCP2 expression defects (30) suggests that MeCP2 plays arole in the regulation of these genes. To determine ifMeCP2 is directly involved in regulating the 15q11-13GABAA receptor subunit genes, chromatin immunoprecipita-tion (ChIP) was performed on chromatin isolated from PMA-differentiated human SH-SH5Y neuroblastoma cells. MeCP2has previously been shown to bind to methylated CpG siteswith adjacent A/T sequence motifs (35), therefore ChIPprimers were designed based on these criteria. The maternallymethylated promoter of SNRPN has been shown to be a targetof MeCP2 and was used as a positive control (13,26). Tworegions within the 50 end of GABRB3 and one region withinthe 50 end of GABRG3 were identified as positive forMeCP2 binding by ChIP (Fig. 4A). The ChIP-positivebinding sites are within intronic regions towards the 50 endof the genes. MeCP2 binding was not detectable in a compar-able site near the 50 end of GABRA5.

To confirm that the positive ChIP results were due to bindingof MeCP2 to methylated CpG sites, bisulfite sequencing was

Figure 1. Imprinting analysis of 15q11-13 GABAA receptor genes. (A) Sequencing chromatograms demonstrate equal biallelic expression of GABRB3 (SNPrs20318) and GABRG3 in control samples. Sequences in upper panels are genomic DNA and sequences in lower panels are from cDNA. Black arrows highlightthe SNP in each chromatogram. AUT 797 shows clear monoallelic expression of GABRB3 and GABRG3 and RTT 5214 has skewed expression of the T allele inGABRG3. Both forward and reverse directions were sequenced for each individual and the chromatograms were consistent. (B) Allele-specific restriction diges-tion analysis of GABRA5 demonstrates normal equal biallelic expression in controls and most autism and RTT samples. The upper panel contains AccI-digestedgenomic DNA from each individual and the lower panel contains AccI-digested cDNA products. AccI cuts the C allele but does not cut the T allele of theGABRA5 SNP. Controls and several autism and RTT samples demonstrate equal biallelic expression, however only the C allele of GABRA5 was detected inAUT 5173.


performed. Genomic DNA from SH-SH5Y cells was bisulfiteconverted and primers were designed within the ChIP-positiveregions when possible. Figure 4B and C illustrate the regionsof GABRB3 and GABRG3 that were analyzed. Although theanalyzed regions of the GABRB3 CpG island were almostcompletely unmethylated, the adjacent MeCP2-bound regionof intron 3 exhibited partial methylation (Fig. 4B). Similarly,bisulfite sequencing of the MeCP2-bound region of GABRG3demonstrated high levels of CpG methylation (Fig. 4C) furthersuggesting that both of these genomic sequences are targets ofMeCP2.

MeCP2-bound regions of GABRB3 and GABRG3are methylated in human brain

Human brain genomic DNA was analyzed by bisulfite sequen-cing to determine if the methylation observed in SH-SY5Ycells is also present in brain. At least 10 clones were analyzedfor each individual and the number of individuals included ineach category is listed below each graph (Fig. 5). The percen-tage of methylated sites reflects the number of methylatedCpG sites divided by all possible methylation sites analyzed.Control brain samples exhibited methylation patterns similarto the observed methylation in SH-SY5Y cells (Fig. 4B andC), shown in Figure 5A and B.

To determine if the observed methylation was parentalallele-specific, PWS deletion, AS deletion and PWS maternalUPD samples were analyzed. Interestingly, PWS samplesexhibited striking differences in GABRB3 methylation basedon gender, however similar gender differences were notobserved for controls or other disorder categories. BothPWS females and AS (male and female) samples exhibitedsignificant deficiencies in methylation for GABRB3 intron 3when compared with control samples (Fig. 5A), suggestingthat GABRB3 methylation is not parent-of-origin specificbut may depend on a normal biparental chromosomal

contribution. In contrast to GABRB3, GABRG3 clones didnot show significant differences in methylation betweencontrols and PWS/AS samples, although the overall levels ofmethylation were high (Fig. 5B). Autism brain samples wereanalyzed to investigate the possibility of methylation abnor-malities contributing to the LOB expression in some ofthese samples. No difference in methylation was observedbetween autism samples and controls for GABRB3 orGABRG3 (Fig. 5A and B).

To test the hypothesis that intronic methylation of GABRB3is positively associated with expression, we looked at themethylation patterns in additional tissue sources. Analysis ofGABRB3 methylation in blood and sperm revealed a signifi-cantly lower level of methylation compared with cortex(Fig. 5C), suggesting that increased methylation of GABRB3intron 3 is biallelically acquired in brain. In contrast to thetissue-specific GABRB3 methylation, GABRG3 methylationwas equally high in brain and blood (Fig. 5D).

DISCUSSION

Because human chromosome 15q11-13 is implicated in mul-tiple neurodevelopmental disorders, it is critical to understandhow genes within this locus are regulated and expressed. Inthis study, we have resolved the dispute over imprinting ofthe 15q11-13 GABAA receptor cluster by showing that thesegenes are biallelically expressed in human frontal cortex yetare aberrantly expressed in multiple neurodevelopmental dis-orders. Our data also provide evidence for homologous transeffects on expression levels of GABRB3, as samples with asingle parental contribution show sub-optimal expression.We have identified both trans and gender effects on brain-specific methylation of an intronic sequence of GABRB3 thatserves as a binding site for MeCP2. Finally, we have providedsupporting evidence that MeCP2 is a positive regulator of

Table 2. Genotype and allelic expression results for control brain samples

Disorder Sample ID Age Sex GABRB3(rs20318)

GABRB3cDNA

GABRA5(rs140682)

GABRA5cDNA

GABRG3(rs140679)

GABRG3cDNA

Control 390 125 days F C/T C/T T/T C/T C/TControl 510 2 years F C/T C/T T/T T/TControl 21 4 years M C/T C/T C/C T/TControl 1406 38 years F C/T C/T C/T T/TControl 1065 15 years M C/T C/T C/C C/TControl 1029 29 years M C/T C/T C/T T/TControl 1136 33 years F C/T C/T C/C C/T C/TControl 35 1 day M C/C C/T C/T T/TControl 125 76 days M C/C C/T C/T T/TControl 1055 96 days M C/C C/T C/T T/TControl 22 113 days M C/C C/T C/T T/TControl 738 8 years F C/C C/T C/T T/TControl 662 12 years F C/C C/T C/T C/CControl 1157 20 days F C/C C/T C/T C/T C/TControl 1206 57 years M C/C C/T C/T C/T C/TControl 1321 62 days F C/C C/C C/T C/TControl 1275 2 years F C/C C/C C/T C/TControl 812 18 years F C/C C/C C/T C/TControl 1027 22 years M C/C C/C C/T C/TControl 602 27 years M C/C T/T C/T C/TControl 4192 46 years M C/C T/T C/T C/T


GABRB3 expression and propose a novel mechanistic role forMeCP2 as an activator of gene expression.

While imprinting is well characterized for some 15q11-13genes, such as SNRPN (36) and UBE3A (27,28), the imprint-ing status of the GABAA receptor cluster has been disputed(18–21). We directly analyzed allelic expression of15q11-13 GABAA receptor subunit genes in human brainand report that GABRB3, GABRA5 and GABRG3 are normallybiallelically expressed in frontal cortex, supporting previousfindings in mouse (14,15). Interestingly, we detected apaternal expression bias of GABRB3 in brain samples withsingle 15q11-13 parental contributions, as previously observedin human PWS and AS lymphoblastoid cell lines (18,19) andmicrocell-mediated chromosome transfer (17). We hypoth-esize that GABRB3 is expressed more efficiently from thepaternal chromosome than the maternal chromosome in iso-lation because the paternal chromosome, containing manypaternally expressed imprinted genes, inherently has a moreactive chromatin configuration than the maternal chromosome.No evidence for a paternal bias was observed in seven differ-ent heterozygous control samples with equal biallelicGABRB3 expression (Table 2), suggesting that the maternalGABRB3 expression defect is rescued by complementationwith a paternal 15q11-13 allele.

Despite having biparental 15q11-13 contributions, one RTTand four autism brain samples exhibited loss of equal biallelicexpression of one or more of the 15q11-13 GABAA receptorsubunit genes (Table 3). Although LOB expression wasobserved for all three of the 15q11-13 GABAA receptorsubunit genes, GABRB3 was the most highly expressed andis the most relevant to autism-spectrum disorders. Autismsamples with LOB expression of any one of the 15q11-13GABAA receptor subunit genes had significantly reducedGABRB3 protein levels compared to samples with biallelicexpression. Furthermore, one sample, AUT 797, heterozygousfor all three GABAA receptor genes, showed LOB expressionfor both GABRB3 and GABRG3 (Fig. 1). These observationssuggest that the defects in autism samples exist throughout

the 15q11-13 GABAA receptor cluster and may be due to epi-genetic dysregulation of one parental allele. Parental DNAsamples were not available for the post-mortem brainsamples, therefore the parental origin of the expressed alleleis currently unknown.

We expected GABRB3 transcript to be significantly reducedin samples with LOB expression, however, transcript levelscorrelated to the reduced protein in only two of four autismsamples with LOB expression. This discrepancy may be par-tially explained by the complex genetic backgrounds ofautism samples. No mutations were detected within thecoding sequence of GABRB3 in autism samples, howevervariations in regulatory sequences may dramatically impacttranscription and stability of transcripts. Interestingly, aGABRB3 promoter polymorphism was recently reported toimpair transcriptional activity in a luciferase assay (37),however this polymorphism was found frequently within oursamples and could not explain differences in GABRB3expression (data not shown). Alternatively, individual differ-ences in post-transcriptional regulation of GABRB3 within theautism samples may explain discrepancies between transcriptand protein. In addition, GABAA receptor subunit genes havebeen shown to exhibit complex coordinate regulation (38),thus a compensation mechanism may account for the transcrip-tional differences observed in some autism brain samples.

Consistent with actively expressed genes, the promoter ofGABRB3 was hypomethylated, however a nearby intronicregion was more highly methylated in human brain than inblood (Fig. 5C). Our quantitative real-time RT–PCR datademonstrate that cerebral cortex has high levels of GABRB3expression (Fig. 3A), while lymphoblastoid cells have barelydetectable GABRB3 expression levels less than 1/1000 of thebrain transcript levels (data not shown). DNA methylationwithin intronic or intergenic regions has been previouslydescribed to confer a positive influence on gene expression(39,40), therefore we suggest that intronic methylation ofGABRB3 is positively associated with expression. WhilePWS and AS brain samples did not reveal parent-of-origin

Table 3. Genotype and allelic expression results for RTT and AUT brain samples

Disorder Sample ID Age Sex MECP2 mutation GABRB3(rs20318)

GABRB3cDNA

GABRA5(rs140682)

GABRA5cDNA

GABRG3(rs140679)

GABRG3cDNA

RTT 4687 8 years F R255X C/C C/C C/T C/TRTT 5214 10 years F R270X C/C C/T C/T C/T T . CRTT 1815 18 years F c. 378-2 A. G C/C C/T N.D. C/T N.D.RTT 5075 20 years F NM C/C C/T C/T T/TRTT 1420 21 years F NM C/T C/T C/C C/T C/TRTT 1748 22 years F NM C/C C/T N.D. C/CRTT 5020 24 years F R255X C/T C/T C/T C/T T/TAUT 3871 5 years M NM C/T C/T C/C T/TAUT 1174 7 years F NM C/C T/T C/T TAUT 4925 9 years M NM C/C C/C C/T TAUT 797 9 years M NM C/T C C/T C/T C/T TAUT 1182 9 years F NM C/C C/C C/CAUT 5342 11 years F g. 21398 T . C C/C C/C C/CAUT 3924 16 years F NM C/T C/T C/T C/T T/TAUT 5144 20 years M NM C/T C/T T/T C/T C/TAUT 5000 27 years M NM C/T C/T C/T C/TAUT 5173 30 years M NM C/C C/T C C/T N.D.AUT 4498 56 years M NM C/C C/C T/T


specific methylation patterns at GABRB3 and GABRG3, ASdeletion samples and PWS female samples had significantlyreduced levels of GABRB3 intron methylation. The hypo-methylation of both AS and PWS samples suggest that transeffects between parental alleles serve to regulate propermethylation levels in this locus in cortex. The gender differencein PWS samples for GABRB3 was unexpected, however,Liljelund et al. (41), also found gender and parent-of-origineffects on Gabrb3 protein expression in mice heterozygousfor a small Gabrb3 deletion that includes a syntenic regionfor the intron 3 methylation sites. Control and autism brainsamples did not exhibit gender effects on methylation andautism samples with LOB expression did not have methylationdefects within the assayed regions of GABRB3 or GABRG3(Fig. 5). Perhaps, the significantly reduced MeCP2 expression

observed in the autism samples with LOB expression (30,42)could be the major epigenetic defect causing dysregulatedGABRB3 expression. In addition, methylation defects maybe found in distal regions within or near GABRB3 using

Figure 2. GABRB3 protein and transcript levels in autism brain samples. (A)Previously published GABRB3 protein levels (13) expressed as mean+ SEMof all samples in each category were compared between controls (white bar),autism samples with biallelic expression (light grey bar) and autism sampleswith LOB 15q11-13 GABAA receptor expression for at least one GABRgene (LOB, dark gray bar). Number of individuals included in each categoryis shown in parentheses besides each x-axis label. LOB autism samples exhib-ited significantly lower GABRB3 protein expression than controls(P ! 0.0004) and biallelic autism samples (P ! 0.004), however there wasno significant difference (N.S.) between controls and biallelic autismsamples. (B) GABRB3 transcript levels were measured in autism and controlcerebral cortex samples by quantitative real-time RT–PCR. GABRB3 tran-script levels normalized to GAPDH are shown for controls (white bar, repre-senting mean+ SEM of seven samples) and individual autism samplesrepresenting the mean+ SEM of the three replicates for each autismsample. LOB autism samples are indicated with dark gray bars and lightgray bars represent autism samples with biallelic expression. Two of fourLOB autism samples, AUT 1174 and AUT 5173, exhibited significantlydecreased transcript levels compared with controls.

Figure 3. Quantitative RT-PCR of 15q11-13 genes in human cortex. Quanti-tative RT–PCR amplification of GABRB3, UBE3A and SNRPN was performedin triplicate experiments on post-mortem human brain cDNA from control,RTT, AS deletion, PWS deletion and PWS maternal UPD samples. Thenumber of individuals analyzed per category is shown in parentheses belowthe x-axis label. All genes were normalized against the housekeeping genecontrol GAPDH and y-axis values represent the mean+ SEM of all replicatesfor all individuals in each category. (A) GABRB3 transcript levels were signifi-cantly reduced in all disorder categories compared with controls (P-values areindicated with asterisks and are listed below the figure). AS deletion and PWSdeletion samples were compared and were also significantly different.(B) UBE3A expression was significantly lower in RTT and AS deletionsamples and significantly higher in PWS UPD samples. (C) SNRPN expressionwas absent in PWS deletion and UPD samples and significantly reduced in ASdeletion samples.


high-throughput techniques in autism samples with LOB15q11-13 GABAA receptor expression in future studies.

We have previously described a developmentally regulatedprocess of homologous pairing of human 15q11-13 GABRB3alleles in normal brain that is deficient in PWS, AS, RTT andautism samples (26). Deficiencies in neuronal trans inter-actions of 15q11-13 domains have been directly linked toMeCP2 by ablation of MeCP2 binding in SH-SY5Y neuro-blastoma cells that significantly reduces pairing of GABRB3alleles (26). Although MeCP2 mutation or deficiency is notsufficient to cause LOB expression of 15q11-13 GABAA

receptor genes in all RTT and autism samples, our resultssuggest that trans interactions between 15q11-13 parentalhomologues contribute to the proper normal biallelic geneexpression within the GABAA receptor domain. The unex-pected significant reduction of SNRPN expression in ASdeletion samples (Fig. 3C) further suggests that trans inter-actions are important for obtaining optimal gene expressionlevels within 15q11-13. Our data show that monoallelicexpression of GABRB3 does not occur in normal brain,which suggests that biallelic expression of GABRB3 isactively maintained. We hypothesize that the proximity ofthe 15q11-13 GABA receptor cluster to imprinted genesrequires an active process in which both GABRB3 alleles

physically associate in order to be expressed optimallyfrom both alleles.

RTT samples with biallelic expression also have significantlyreduced GABRB3 expression (Fig. 3A), implicating a positiverole for MeCP2 in the regulation of GABRB3 expressionby binding to methylated CpG sites in both cis and trans.The model shown in Figure 6 illustrates two possibilities toexplain reduced biallelic and monoallelic expression ofGABRB3 in RTT brain. In normal brain, MeCP2 binding tointron 3 methylated sites serves to positively regulateGABRB3 expression by bringing both parental alleles to anuclear location with optimal transcriptional machinery, suchas a transcription factory (43). In cases of reduced biallelicexpression, 15q11-13 homologues are not paired due todeficiencies in MeCP2, and neither parental chromosome isoptimally transcribed. Similarly, Mecp2-deficient mice alsoexhibit a significant two-fold reduction in Gabrb3 despitenormal biallelic expression (13). In autism and RTT sampleswith monoallelic expression, pairing is disrupted, but anadditional epigenetic or genetic defect prohibits transcriptionof one allele (Fig. 6). Although we speculate that the maternalallele is silent, due to evidence for a paternal bias in expression(Fig. 3A), the parental origin of alleles in the autism and RTTpost-mortem brain samples is currently unknown.

Figure 4. MeCP2 binds to methylated CpG sites within intronic sequences of GABRB3 and GABRG3. (A) Chromatin immunoprecipitation of 48 h differentiatedSH-SY5Y cells. Input DNA represents chromatin prior to immunoprecipitation, MeCP2 Ab represents chromatin immunoprecipitated with a C-terminal MeCP2antibody, and pre-immune (negative) control antibody. Genes assayed for binding are shown to the left of the panel. The SNRPN promoter was used as a positivecontrol (26). Two regions of GABRB3 and an intronic region of GABRG3 show detectable MeCP2 binding, however, MeCP2 binding to GABRA5 was not detect-able. (B) Schematic of 2.5 kb of the 50 end of GABRB3. Coding exons are indicated with black numbered boxes and gray boxes represent 50-UTR. Dashed lineabove the gene schematic represents the location of the 50 end CpG island. Black lines with arrows below the gene show the relative position of the ChIP PCRproducts (A) and the lines with numbered circles above the gene show the relative position of the bisulfite sequencing products. Bisulfite sequencing results from48 h differentiated SH-SY5Y cells are shown below the gene schematic. Each line represents an individual clone and circles indicate CpG sites. Filled circlesrepresent methylated CpG sites and unfilled circles represent unmethylated CpG sites. Stars above methylation sites indicated CpG sites with adjacent A/Tsequence motifs that are high affinity MeCP2 binding sites (35). (C) Schematic of 2.0 kb of the 50 end of GABRG3 showing the highly methylated intronicsequence, shown to be positive for MeCP2 binding in (A).


In this report we show by ChIP and bisulfite sequencing thatMeCP2 binds to intronic methylated CpG sequences withinGABRB3 in differentiated human SH-SH5Y cells (Fig. 4).Although traditionally viewed as a transcriptional repressor(44,45), MeCP2 has recently been shown indirectly to posi-tively regulate the target gene Bdnf (46) and genome-wideexpression profiling analyses have demonstrated many poten-tially down-regulated genes with MeCP2 deficiency (47–49).In our model (Fig. 6), we suggest that binding of MeCP2 tobrain-specific methylated CpG sites within GABRB3 posi-tively influences expression by promoting trans interactionsbetween homologues and by positively organizing chromatinin cis. Horike et al. (34) previously described a role forMeCP2 as an organizer of chromatin loops that silence geneexpression. We hypothesize that by binding to multiple sitesin cis within the GABAA receptor cluster, MeCP2 promotesformation of chromatin structures that are favorable foractive transcription, as illustrated by proximity to an activetranscription factory in the model in Figure 6.

GABA is the major inhibitory neurotransmitter in thebrain and plays an important role in brain function. Partialdisruption of GABAergic neurons in the cortex of mice causesseizures, anxiety and impaired social behavior, consistent withGABA signaling defects contributing to the etiology of autism-spectrum disorders (50). Unlike GABRB3, GABRA5 andGABRG3 proteins do not primarily localize in cerebral cortex(51–53), consistent with our expression data (SupplementaryMaterial, Fig. S1). Deletion of both Gabra5 and Gabrg3 doesnot cause an apparent phenotype in mice (54,55), however,Gabrb3-deficient mice exhibit phenotypes such as seizures,sleep abnormalities and stereotyped behaviors consistent withcharacteristics of AS, RTT and autism (55).

Genetic and phenotypic overlap in RTT, autism and ASsuggests that overlapping molecular pathways are dysregu-lated in these similar neurodevelopmental disorders. Althoughrare, MECP2 mutations have been reported in individualsclinically diagnosed with AS or autism (56,57), and MECP2mutations combined with rearrangements in 15q11-13 have

Figure 5. Bisulfite sequencing of human brain and blood DNA. Bisulfite sequencing was performed on human genomic DNA from control, PWS, AS and autismindividuals. At least 10 clones were analyzed for each individual included, and percent methylation reflects the percentage of methylated CpG sites out of thetotal number of CpG sites assayed for all clones of each sample. Results are graphed as the mean+ SEM of percent methylation for all samples included in eachcategory. The number of individuals analyzed in each category is shown below x-axis label. (A) GABRB3 methylated intronic sequence shown in Figure 4(B)was bisulfite-sequenced in brain genomic DNA from control, PWS male, PWS female, AS and autism samples. PWS male samples and autism samples did nothave significantly different methylation than controls. In contrast, the PWS female samples and AS samples exhibited significantly reduced GABRB3 methyl-ation. (B) The methylated sequence of GABRG3 intron 1 shown in Figure 4(C) was bisulfite-sequenced in control, PWS, AS and autism human brain samples. Allsamples analyzed exhibited similar high-level of methylation. (C) Control brain DNA methylation of GABRB3 intron 3 was compared with methylation in humancontrol blood samples and one sperm sample. Blood and sperm samples showed reduced methylation compared with brain (P, 0.02 in blood). (D) Control brainand blood methylation levels for GABRG3 intron 1 were compared and no significant difference was observed.


been reported in individuals with RTT (58). In this study weshow that GABRB3, a 15q11-13 gene, is positively regulatedby MeCP2 and significantly reduced in both RTT and AS.Clinical observations of EEG abnormalities typical of AS inRTT patients support the involvement of GABRB3 in theshared epileptic phenotypes in these disorders (59,60).Maternal duplications of 15q11-13, including GABRB3,occur in a small percentage of autistic individuals, howeverit is unclear how this duplication affects GABRB3 expressionin brain. Although increased copy number is predicted toincrease expression, our data demonstrate complex transregulation of GABRB3 in human brain suggesting that devi-ations from the normal biparental chromosomal contributionmay negatively impact GABRB3 expression. Our currentstudy also provides evidence for complex epigenetic dysregu-lation of the 15q11-13 GABAA receptor cluster in idiopathicautism samples with GABRB3 protein defects, furthersuggesting that shared molecular pathways are dysregulatedin related neurodevelopmental disorders.

MATERIALS AND METHODS

Human post-mortem brain samples

Frozen frontal cerebral cortex (BrodmannArea 9) samples wereobtained from the Autism Tissue Program from the Universityof Maryland Brain and Tissue Bank for NeurodevelopmentalDisorders, the Harvard Brain Tissue Resource Center and theUniversity of Miami Brain and Tissue Bank for Neurodevelop-mental Disorders. Brain samples were chosen from human

cadavers where the PMI was ,30 h. The cause of death, PMIand additional information about each tissue sample is providedin Supplementary Material, Table S1. All samples were storedat 2808C to preserve RNA integrity.

Human brain imprinting analysis

Human brain genomic DNA was isolated from frozen cerebralcortex with the Puregenew DNA Purification Kit (GentraSystems). SNPs were identified within the coding regions ofeach gene using the UC Santa Cruz Genome Browser(http://genome.ucsc.edu) and the NCBI SNP database (http://www.ncbi.nlm.nih.gov). Primers used to PCR amplifygenomic DNA samples for genotyping are found in Sup-plementary Material, Table S2. SNPs rs2912582 andrs140682 cause differential restriction enzyme digestioncleavage and genotypes were determined after 3 h of digestionwith an excess of ScrFI (rs2912582) or AccI (rs140682).Direct sequencing was used to genotype SNPs rs20318 andrs140679 (UC Davis Division of Biological Sciences DNASequencing Facility).

Allelic expression of each SNP in heterozygous individualswas analyzed by the same methods described above for geno-typing. RNA was extracted from frozen human cerebralcortex tissue using TRIzol reagent (Invitrogen). RNA concen-trations were measured with the Nanodropw D-1000 spectro-photometer and high quality RNA was used for cDNAsynthesis. cDNA was synthesized from one or two microgramsof DNase I (Invitrogen) treated RNA, with oligo(dT) primer andSS II Reverse Transcriptase (Invitrogen). Negative controls

Figure 6. Model for MeCP2 involvement in positive regulation of GABRB3 expression in brain. Maternal and paternal chromosomes are shown as pink(maternal) and blue (paternal) dotted lines, respectively. The size of arrows represents relative levels of gene expression. The large paternal-specific transcriptionunit from SNRPN through the antisense transcript of UBE3A is depicted as a cloud. In normal brain samples MeCP2 binds to methylated CpG sites within theintron of GABRB3 on both chromosomes and promotes GABRB3 expression and homologous pairing by localizing both alleles to an area of active transcriptionsuch as a transcription factory. In Rett syndrome patients with MeCP2 mutation or deficiency, homologous pairing of 15q11-13 is disrupted, as previouslydescribed in Thatcher et al. (26). In RTT individuals with reduced biallelic expression both maternal and paternal alleles are transcribed at low levelsbecause of suboptimal nuclear location. Individuals with LOB expression may have an additional epigenetic defect leading to selective reduction of oneallele (represented with an X). Because 15q11-13 deletion leads to loss of homologous pairing and a paternal bias in GABRB3 expression, we speculate thatthe paternal allele is preferentially expressed in these cases.


were created for each individual without the addition of the SSII Reverse Transcriptase enzyme to ensure genomic DNAcontamination or non-specific amplification did not occur.Approximately 50–100 ng of brain cDNA was PCR-amplifiedwith standard reaction conditions using the intron–exon bound-ary spanning primers found in Supplementary Material,Table S2. The same genomic DNA primers used to amplifythe GABRB3 30-UTR were used for cDNA amplification. Nega-tive control cDNA samples (-RT) were used for all RT–PCRreactions to ensure specific amplification of cDNA products.

Quantitative real-time RT–PCR of human brain cDNA

Quantitative real-time RT–PCR was performed using theRoche LightCycler system according to the manufacturer’sinstructions with 1" DNA master SYBR Green I reactionbuffer (Roche), 3 mM MgCl2, 0.5 mM of primers and 20–25 ng of cDNA. To confirm amplification specificity, thePCR products were subject to a melting curve analysis, inwhich only one peak was observed. Crossing point valueswere measured with the LightCycler analysis softwareversion 2.0 and final quantification was performed using thecomparative CT method (Applied Biosystems). Three repli-cate reactions were performed for each individual per geneanalyzed and values were normalized to the housekeepinggene GAPDH. Primers were designed to span at least oneintron–exon boundary and are listed in SupplementaryMaterial, Table S2.

Chromatin immunoprecipitation

Chromatin was prepared from human SH-SY5Y neuroblas-toma cells and purified by urea gradient centrifugation asdescribed previously (61,62). Immunoprecipitation, reverse-crosslinking and PCR amplification were performed asdescribed previously (26,63) with minor modifications.Briefly, chromatin was digested with a panel of restrictionenzymes (New England Biolabs) to 100–700 bp DNA frag-ments. Digested chromatin was precleared with PrecipHenagarose beads (Aves Labs) alone, then with pre-absorbedchicken IgY (Aves Labs) followed by agarose beads.Precleared chromatin was incubated overnight with 5 mg ofC-terminal anti-MeCP2 (26) or with equivalent amounts ofpre-immune IgY as a control for non-specific binding. PCRamplification was performed for 31 cycles with one-twentiethof the IP products in a reaction containing 1 M Betaine, 1"Invitrogen buffer, 1 U Taq (Invitrogen), 4 mM MgCl2,2.5 mM dNTPs and 0.4 mM of the following primers:GABRB3 50-F: CCAGAATCTCTTTCCAAACGA; R: CATCGACATGGTTTCCGA), GABRB3 intron 3 (F: CTTGGGGTCCCTGCATTTAA; R: AACAGCTTGGCGTGGTCAAA),GABRG3 (F: GGAGTGTGGGGACTCAGATTA; R: TGGAAGGTGTGTGTGTGTGTT), GABRA5 (F: CATCACGTTTTCAGTGGCTTT; R: TCTGGGGGTTTTCTGAAGTT),SNRPN (26).

Bisulfite sequencing

GenomicDNA fromSH-SY5Y cells and human tissues was iso-lated with the Puregenew DNA Purification Kit (Gentra

Systems) according to manufacturer’s protocol. DNA yieldwas measured with a Nanodropw D-1000 spectrophotometerand 1 mg of genomic DNA was used for bisulfite conversionwith the CpGenomeTM DNA Modification Kit following themanufacturer’s protocol (Chemicon International). BisulfitePCR primers were designed using MethPrimer (www.urogene.org/methprimer/index1.htm) (64) and are as follows:GABRB3 region 1 (F: GAGGGATATTTTGGAAGGAGTATTT; R: CAAAACTCCAACCTATCTTAATCAAA),GABRB3 region 2 (F: GGATTTAATATTTTGATTTGTGGGA; R: AAAAAAACCATCCTTCTCCTAATTC), GABRB3region 3 (F: GTTTTGAGGATTGGGGGTTAT; R: AAAACATATCCTTTATAAAAAAAAC), GABRB3 region 4(F: GTTTAGGAGGTAGTTTTGGAAAAAT; R: TAAACTCCCAACTCCTCCAAC), GABRG3 region 1 (F: GTAGGGGT-TATTTTTGGTTTTTTAT; R: ATTTAACCCTTTAAATCCCACTATC). PCR amplification was performed for 38 cycleswith standard reaction conditions and !25 ng of bisulfite con-verted genomic DNA. PCR products were gel-purified withthe Qiaquick Gel Purification kit (Qiagen) and ligated into thepGEM T-Easy cloning vector (Promega). JM109 competentcells (Promega) were transformed with ligated bisulfite pro-ducts and plated onto LB ampicillin agar plates with IPTGand X-gal. Single white colonies were selected, grown in LBwith ampicillin overnight and plasmids were purified usingthe Qiaprep kit (Qiagen). Digestion of plasmids with NotI ver-ified the presence of the correct insert and then plasmid insertswere sequenced at the UC Davis Division of BiologicalSciences DNA Sequencing Facility. At least 10 colonies weresequenced for each brain sample analyzed.

SUPPLEMENTARY MATERIAL

Supplementary Material is available at HMG Online.

ACKNOWLEDGEMENTS

The authors thank Dr Nobuko Hagiwara, Dr Susan Swanberg,Karen Thatcher, Sailaja Peddada and Roxanne Vallero forvaluable critiques of the manuscript. We thank Chris Iwahashiand Dr. Chris Bray for help obtaining blood and brainsamples. Human brain tissues were generously provided bythe Autism Tissue Program, The University of MarylandBrain and Tissue Bank for Developmental Disorders, The Uni-versity of Miami Brain and Tissue Bank for DevelopmentalDisorders, and the Harvard Brain Tissue Resource Center(supported in part by NIH R24MH-068855). This work wassupported by NIH 1R01HD048799 granted to J.L.

Conflict of Interest statement. The authors do not declare aconflict of interest.

REFERENCES

1. Nicholls, R.D. and Knepper, J.L. (2001) Genome organization, function,and imprinting in Prader-Willi and Angelman syndromes. Annu. Rev.Genomics Hum. Genet., 2, 153–175.

2. Kishino, T., Lalande, M. and Wagstaff, J. (1997) UBE3A/E6-APmutations cause Angelman syndrome. Nat. Genet., 15, 70–73.


3. Matsuura, T., Sutcliffe, J.S., Fang, P., Galjaard, R.J., Jiang, Y.H., Benton,C.S., Rommens, J.M. and Beaudet, A.L. (1997) De novo truncatingmutations in E6-AP ubiquitin-protein ligase gene (UBE3A) in Angelmansyndrome. Nat. Genet., 15, 74–77.

4. Volkmar, F.R. and Pauls, D. (2003) Autism. Lancet, 362, 1133–1141.5. Muhle, R., Trentacoste, S.V. and Rapin, I. (2004) The genetics of autism.

Pediatrics, 113, 472–486.6. Cook, E.H., Jr, Lindgren, V., Leventhal, B.L., Courchesne, R., Lincoln,

A., Shulman, C., Lord, C. and Courchesne, E. (1997) Autism or atypicalautism in maternally but not paternally derived proximal 15q duplication.Am. J. Hum. Genet., 60, 928–934.

7. McCauley, J.L., Olson, L.M., Delahanty, R., Amin, T., Nurmi, E.L.,Organ, E.L., Jacobs, M.M., Folstein, S.E., Haines, J.L. and Sutcliffe, J.S.(2004) A linkage disequilibrium map of the 1-Mb 15q12 GABA(A)receptor subunit cluster and association to autism. Am. J. Med. Genet,131, 51–59.

8. Nurmi, E.L., Amin, T., Olson, L.M., Jacobs, M.M., McCauley, J.L.,Lam, A.Y., Organ, E.L., Folstein, S.E., Haines, J.L. and Sutcliffe, J.S.(2003) Dense linkage disequilibrium mapping in the 15q11-q13maternal expression domain yields evidence for association in autism.Mol. Psychiatry, 8, 624–634.

9. Shao, Y., Cuccaro, M.L., Hauser, E.R., Raiford, K.L., Menold, M.M.,Wolpert, C.M., Ravan, S.A., Elston, L., Decena, K., Donnelly, S.L. et al.(2003) Fine mapping of autistic disorder to chromosome 15q11-q13 byuse of phenotypic subtypes. Am. J. Hum. Genet., 72, 539–548.

10. Menold, M.M., Shao, Y., Wolpert, C.M., Donnelly, S.L., Raiford, K.L.,Martin, E.R., Ravan, S.A., Abramson, R.K., Wright, H.H., Delong, G.R.et al. (2001) Association analysis of chromosome 15 gabaa receptorsubunit genes in autistic disorder. J. Neurogenet., 15, 245–259.

11. Amir, R.E., Van den Veyver, I.B., Wan, M., Tran, C.Q., Francke, U. andZoghbi, H.Y. (1999) Rett syndrome is caused by mutations in X-linkedMECP2, encoding methyl- CpG-binding protein 2. Nat. Genet., 23,185–188.

12. Zoghbi, H.Y. (2003) Postnatal neurodevelopmental disorders: meeting atthe synapse? Science, 302, 826–830.

13. Samaco, R.C., Hogart, A. and LaSalle, J.M. (2005) Epigenetic overlap inautism-spectrum neurodevelopmental disorders: MECP2 deficiencycauses reduced expression of UBE3A and GABRB3. Hum. Mol. Genet.,14, 483–492.

14. Buettner, V.L., Longmate, J.A., Barish, M.E., Mann, J.R. andSinger-Sam, J. (2004) Analysis of imprinting in mice with uniparentalduplication of proximal chromosomes 7 and 15 by use of a customoligonucleotide microarray. Mamm. Genome, 15, 199–209.

15. Nicholls, R.D., Gottlieb, W., Russell, L.B., Davda, M., Horsthemke, B.and Rinchik, E.M. (1993) Evaluation of potential models for imprintedand nonimprinted components of human chromosome 15q11-q13syndromes by fine-structure homology mapping in the mouse. Proc. NatlAcad. Sci. USA, 90, 2050–2054.

16. Yamasaki, K., Joh, K., Ohta, T., Masuzaki, H., Ishimaru, T., Mukai, T.,Niikawa, N., Ogawa, M., Wagstaff, J. and Kishino, T. (2003) Neurons butnot glial cells show reciprocal imprinting of sense and antisensetranscripts of Ube3a. Hum. Mol. Genet., 12, 837–847.

17. Meguro, M., Mitsuya, K., Sui, H., Shigenami, K., Kugoh, H.,Nakao, M. and Oshimura, M. (1997) Evidence for uniparental, paternalexpression of the human GABAA receptor subunit genes, usingmicrocell-mediated chromosome transfer. Hum. Mol. Genet., 6,2127–2133.

18. Bittel, D.C., Kibiryeva, N., Talebizadeh, Z. and Butler, M.G. (2003)Microarray analysis of gene/transcript expression in Prader-Willisyndrome: deletion versus UPD. J. Med. Genet., 40, 568–574.

19. Bittel, D.C., Kibiryeva, N., Talebizadeh, Z., Driscoll, D.J. andButler, M.G. (2005) Microarray analysis of gene/transcript expressionin Angelman syndrome: deletion versus UPD. Genomics, 85, 85–91.

20. Kubota, T., Niikawa, N., Jinno, Y. and Ishimaru, T. (1994) GABAAreceptor beta 3 subunit gene is possibly paternally imprinted in humans.Am. J. Med. Genet., 49, 452–453.

21. Gabriel, J.M., Higgins, M.J., Gebuhr, T.C., Shows, T.B., Saitoh, S. andNicholls, R.D. (1998) A model system to study genomic imprintingof human genes. Proc. Natl Acad. Sci. USA, 95, 14857–14862.

22. Knoll, J.H.M., Cheng, S.-D. and Lalande, M. (1994) Allele-speificity ofDNA replication timing in the Angelman/Prader-Willi syndromeimprinted chromosomal region. Nat. Genet., 6, 41–46.

23. Kitsberg, D., Selig, S., Brandeis, M., Simon, I., Keshet, I., Driscoll, D.J.,Nicholls, R.D. and Cedar, H. (1994) Allele-specific replication timing ofimprinted gene regions. Nature, 364, 459–463.

24. LaSalle, J.M. and Lalande, M. (1995) Domain organization ofallele-specific replication within the GABRB3 gene cluster requiresa biparental 15q11-13 contribution. Nat. Genet., 9, 386–394.

25. LaSalle, J. and Lalande, M. (1996) Homologous association of oppositelyimprinted chromosomal domains. Science, 272, 725–728.

26. Thatcher, K.N., Peddada, S., Yasui, D.H. and Lasalle, J.M. (2005)Homologous pairing of 15q11-13 imprinted domains in brain isdevelopmentally regulated but deficient in Rett and autism samples.Hum. Mol. Genet., 14, 785–797.

27. Rougeulle, C., Glatt, H. and Lalande, M. (1997) The Angelman syndromecandidate gene, UBE3A/E6-AP, is imprinted in brain. Nat. Genet., 17,14–15.

28. Vu, T. and Hoffman, A. (1997) Imprinting of the Angelman syndromegene, UBE3A, is restricted to brain. Nat. Genet., 17, 12–13.

29. Albrecht, U., Sutcliffe, J.S., Cattanach, B.M., Beechey, C.V.,Armstrong, D., Eichele, G. and Beaudet, A.L. (1997) Imprintedexpression of the murine Angelman syndrome gene, Ube3a, inhippocampal and Purkinje neurons. Nat. Genet., 17, 75–78.

30. Nagarajan, R.P., Hogart, A., Gwye, Y., Martin, M.R. and LaSalle, J.M.(2006) Reduced MeCP2 expression is frequent in autism frontal cortexand correlates with aberrant MECP2 promoter methylation. Epigenetics,1, 172–182.

31. Makedonski, K., Abuhatzira, L., Kaufman, Y., Razin, A. and Shemer, R.(2005) MeCP2 deficiency in Rett syndrome causes epigenetic aberrationsat the PWS/AS imprinting center that affects UBE3A expression.Hum. Mol. Genet., 14, 1049–1058.

32. Jordan, C. and Francke, U. (2006) Ube3a expression is not altered inMecp2 mutant mice. Hum. Mol. Genet., 15, 2210–2215.

33. Balmer, D., Arredondo, J., Samaco, R.C. and LaSalle, J.M. (2002)MECP2 mutations in Rett syndrome adversely affect lymphocyte growth,but do not affect imprinted gene expression in blood or brain. Hum.Genet., 110, 545–552.

34. Horike, S., Cai, S., Miyano, M., Cheng, J.F. and Kohwi-Shigematsu, T.(2005) Loss of silent-chromatin looping and impaired imprinting of DLX5in Rett syndrome. Nat. Genet., 37, 31–40.

35. Klose, R.J., Sarraf, S.A., Schmiedeberg, L., McDermott, S.M.,Stancheva, I. and Bird, A.P. (2005) DNA binding selectivity of MeCP2due to a requirement for A/T sequences adjacent to methyl-CpG. Mol.Cell, 19, 667–678.

36. Reed, M.L. and Leff, S.E. (1994) Maternal imprinting of human SNRPN,a gene deleted in Prader-Willi syndrome. Nat. Genet., 6, 163–167.

37. Urak, L., Feucht, M., Fathi, N., Hornik, K. and Fuchs, K. (2006) AGABRB3 promoter haplotype associated with childhood absence epilepsyimpairs transcriptional activity. Hum. Mol. Genet., 15, 2533–2541.

38. Ramadan, E., Fu, Z., Losi, G., Homanics, G.E., Neale, J.H. andVicini, S. (2003) GABA(A) receptor beta3 subunit deletion decreasesalpha2/3 subunits and IPSC duration. J. Neurophysiol., 89, 128–134.

39. Unoki, M. and Nakamura, Y. (2003) Methylation at CpG islands in intron1 of EGR2 confers enhancer-like activity. FEBS Lett., 554, 67–72.

40. Murrell, A., Heeson, S. and Reik, W. (2004) Interaction betweendifferentially methylated regions partitions the imprinted genes Igf2 andH19 into parent-specific chromatin loops. Nat. Genet., 36, 889–893.

41. Liljelund, P., Handforth, A., Homanics, G.E. and Olsen, R.W. (2005)GABAA receptor beta3 subunit gene-deficient heterozygous miceshow parent-of-origin and gender-related differences in beta3subunit levels, EEG, and behavior. Brain Res. Dev. Brain Res., 157,150–161.

42. Samaco, R.C., Nagarajan, R.P., Braunschweig, D. and LaSalle, J.M.(2004) Multiple pathways regulate MeCP2 expression in normal braindevelopment and exhibit defects in autism-spectrum disorders.Hum. Mol. Genet., 13, 629–639.

43. Osborne, C.S., Chakalova, L., Brown, K.E., Carter, D., Horton, A.,Debrand, E., Goyenechea, B., Mitchell, J.A., Lopes, S., Reik, W. et al.(2004) Active genes dynamically colocalize to shared sites of ongoingtranscription. Nat. Genet., 36, 1065–1071.

44. Nan, X., Campoy, F.J. and Bird, A. (1997) MeCP2 is a transcriptionalrepressor with abundant binding sites in genomic chromatin. Cell, 88,471–481.

45. Nan, X., Ng, H.H., Johnson, C.A., Laherty, C.D., Turner, B.M.,Eisenman, R.N. and Bird, A. (1998) Transcriptional repression by


the methyl-CpG-binding protein MeCP2 involves a histonedeacetylase complex. Nature, 393, 386–389.

46. Chang, Q., Khare, G., Dani, V., Nelson, S. and Jaenisch, R. (2006)The disease progression of Mecp2 mutant mice is affected by the levelof BDNF expression. Neuron, 49, 341–348.

47. Tudor, M., Akbarian, S., Chen, R.Z. and Jaenisch, R. (2002) Transcriptionalprofiling of a mouse model for Rett syndrome reveals subtle transcriptionalchanges in the brain. Proc. Natl Acad. Sci. USA, 99, 15536–15541.

48. Colantuoni, C., Jeon, O.H., Hyder, K., Chenchik, A., Khimani, A.H.,Narayanan, V., Hoffman, E.P., Kaufmann, W.E., Naidu, S. andPevsner, J. (2001) Gene expression profiling in postmortem Rettsyndrome brain: differential gene expression and patient classification.Neurobiol. Dis., 8, 847–865.

49. Peddada, S., Yasui, D.H. and Lasalle, J.M. (2006) Inhibitors ofdifferentiation (ID1, ID2, ID3 and ID4) genes are neuronal targetsof MeCP2 that are elevated in Rett syndrome. Hum. Mol. Genet., 15,2003–2014.

50. Levitt, P. (2005) Disruption of interneuron development. Epilepsia.46(Suppl. 7), 22–28.

51. Pirker, S., Schwarzer, C., Wieselthaler, A., Sieghart, W. and Sperk, G.(2000) GABA(A) receptors: immunocytochemical distribution of 13subunits in the adult rat brain. Neuroscience, 101, 815–850.

52. Lingford-Hughes, A., Hume, S.P., Feeney, A., Hirani, E., Osman, S.,Cunningham, V.J., Pike, V.W., Brooks, D.J. and Nutt, D.J. (2002)Imaging the GABA-benzodiazepine receptor subtype containing thealpha5-subunit in vivo with [11C]Ro15 4513 positron emissiontomography. J. Cereb. Blood Flow Metab., 22, 878–889.

53. Miralles, C.P., Li, M., Mehta, A.K., Khan, Z.U. and De Blas, A.L. (1999)Immunocytochemical localization of the beta(3) subunit of thegamma-aminobutyric acid(A) receptor in the rat brain. J. Comp. Neurol.,413, 535–548.

54. Culiat, C.T., Stubbs, L.J., Montgomery, C.S., Russell, L.B. andRinchik, E.M. (1994) Phenotypic consequences of deletion of thegamma 3, alpha 5, or beta 3 subunit of the type A gamma-aminobutyricacid receptor in mice. Proc. Natl Acad. Sci. USA, 91, 2815–2818.

55. DeLorey, T.M., Handforth, A., Anagnostaras, S.G., Homanics, G.E.,Minassian, B.A., Asatourian, A., Fanselow, M.S., Delgado-Escueta, A.,Ellison, G.D. and Olsen, R.W. (1998) Mice lacking the beta3 subunit ofthe GABAA receptor have the epilepsy phenotype and many of thebehavioral characteristics of Angelman syndrome. J. Neurosci., 18,8505–8514.

56. Carney, R.M., Wolpert, C.M., Ravan, S.A., Shahbazian, M.,Ashley-Koch, A., Cuccaro, M.L., Vance, J.M. and Pericak-Vance, M.A.(2003) Identification of MeCP2 mutations in a series of females withautistic disorder. Pediatr. Neurol., 28, 205–211.

57. Watson, P., Black, G., Ramsden, S., Barrow, M., Super, M., Kerr, B. andClayton-Smith, J. (2001) Angelman syndrome phenotype associated withmutations in MECP2, a gene encoding a methyl CpG binding protein.J. Med. Genet., 38, 224–228.

58. Longo, I., Russo, L., Meloni, I., Ricci, I., Ariani, F., Pescucci, C.,Giordano, C.T., Canitano, R., Hayek, G., Zappella, M. et al. (2004)Three Rett patients with both MECP2 mutation and 15q11-13rearrangements. Eur. J. Hum. Genet., 12, 682–685.

59. Laan, L.A. and Vein, A.A. (2002) A Rett patient with a typical AngelmanEEG. Epilepsia, 43, 1590–1592.

60. Valente, K.D. (2003) Another Rett patient with a typical Angelman EEG.Epilepsia, 44, 873–874.

61. de Belle, I., Cai, S. and Kohwi-Shigematsu, T. (1998) The genomicsequences bound to special AT-rich sequence-binding protein 1 (SATB1)in vivo in Jurkat T cells are tightly associated with the nuclear matrixat the bases of the chromatin loops. J. Cell Biol., 141, 335–348.

62. Orlando, V. and Paro, R. (1993) Mapping Polycomb-repressed domainsin the bithorax complex using in vivo formaldehyde cross-linkedchromatin. Cell, 75, 1187–1198.

63. Yasui, D., Miyano, M., Cai, S., Varga-Weisz, P. andKohwi-Shigematsu, T. (2002) SATB1 targets chromatin remodelling toregulate genes over long distances. Nature, 419, 641–645.

64. Li, L.C. and Dahiya, R. (2002) MethPrimer: designing primers formethylation PCRs. Bioinformatics, 18, 1427–1431.


15q11-13 gaba a receptor genes are normally … · gen es is com mon to bo th di sor der s. ......

Documents