Skip Navigation


Human Molecular Genetics Advance Access originally published online on January 12, 2007
Human Molecular Genetics 2007 16(5):537-546; doi:10.1093/hmg/ddl488
This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow Supplementary Data
Right arrow All Versions of this Article:
16/5/537    most recent
ddl488v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (3)
Right arrowRequest Permissions
Google Scholar
Right arrow Articles by Wilkins, J. M.
Right arrow Articles by Loughlin, J.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Wilkins, J. M.
Right arrow Articles by Loughlin, J.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

© The Author 2007. Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oxfordjournals.org

Extreme context specificity in differential allelic expression

James M. Wilkins*, Lorraine Southam, Andrew J. Price, Zehra Mustafa, Andrew Carr and John Loughlin

University of Oxford, Institute of Musculoskeletal Sciences, Botnar Research Centre, Nuffield Orthopaedic Centre, Oxford OX3 7LD, UK

* To whom correspondence should be addressed. Tel: +44 1865227963; Fax: +44 1865227966; Email: james.wilkins{at}ndos.ox.ac.uk

Received December 1, 2006; Accepted January 3, 2007


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 RESULTS
 DISCUSSION
 MATERIALS AND METHODS
 SUPPLEMENTARY MATERIAL
 REFERENCES
 
Variability in cis-regulation of gene expression has been implicated in the phenotypic manifestation of complex traits including common, multifactorial diseases. The differential expression of alleles due to polymorphism in cis-regulatory elements is common in the human genome, but there is a paucity of information about the context specificity of these control elements. In this study, we examined the differential allelic expression (DAE) of BMP5 in human mesenchymal tissues obtained from 16 donors undergoing joint replacement for treatment of osteoarthritis. We observed significant differences in BMP5 allelic output, with allelic ratios greater than 4:1 (P < 10–20) in the tissues of some donors. We also discovered a significant variability in allelic expression within the different tissues of donors. For 12 of our donors, we examined the allelic expression of BMP5 in two different regions of cartilage: cartilage adjacent to the site of the osteoarthritic lesion and cartilage distal from the lesion. Five of these 12 donors demonstrated highly significant differences (P ≤ 10–8) in allelic expression between the different regions of their cartilage. Using DAE as a phenotype, we attempted to map tissue-specific cis-regulatory polymorphisms, and we identified a single nucleotide polymorphism located downstream of BMP5, which was significantly associated with DAE in some but not all of the examined tissues. These findings suggest that allelic expression can be highly context specific and that when interrogating the cis-regulatory control of a particular gene, one cannot necessarily assume that allelic expression is conserved across different tissues or even across different regions of the same tissue.


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
RESULTS
 DISCUSSION
 MATERIALS AND METHODS
 SUPPLEMENTARY MATERIAL
 REFERENCES
 
The genetic basis for the phenotypic diversity observed in natural populations is increasingly being linked to polymorphisms residing in regulatory elements. Whether these polymorphisms are cis-acting or trans-acting, their functional consequence is variability in gene expression. Regulatory polymorphisms affecting transcriptional regulation have been postulated to be the driving force behind phenotypic evolution (1) as well as a significant component of one's susceptibility to complex, multifactorial diseases (2). Although the origins of modern humans are relatively recent, there is considerable variability in cis-regulatory elements within the human population, with an average individual estimated to be heterozygous at functional cis-regulatory sites in over 40% of all genes (3,4).

Polymorphisms in the coding regions of human genes have generally been well characterized (5,6), but functional variants in regulatory regions of a gene remain much harder to discern from nucleotide sequences alone for a number of reasons. For instance, regulatory regions can reside large distances from the transcriptional unit, and in some cases, these control regions are found hundreds of thousands of bases away from the transcriptional start site (7,8). Furthermore, the action of cis-regulatory elements may be context specific in that one allele can show enhanced expression in one cell type and repression or no effect in other cellular environments (9,10). Consequently, direct sequence-based analyses of a particular gene locus have generally proved inadequate in elucidating the true complexity of regulatory polymorphisms governing gene expression.

Experimentally, one can measure the amount of transcript from a gene in a number of individuals as a surrogate estimate of the functional contribution of variants within regulatory elements at that locus, but these studies are complicated by their inability to differentiate between cis-acting and trans-acting factors as well as environmental influences (11). To circumvent these complications, one can instead measure the allelic output from a particular gene to directly assess the extent of cis-regulatory variation at that locus. The central assumption of this type of investigation is that in the absence of cis-regulatory polymorphisms, both copies of an autosomal gene will contribute equal amounts of message. Thus, if differential allelic expression (DAE) is detected for an individual, the implication is that that individual is heterozygous at a cis-regulatory site responsible for transcription, mRNA stability or processing and/or translational efficiency (12). This approach focuses on allelic variation within an individual, so each allele serves as an internal control for confounding factors that contribute to the total amount of message in a particular sample, such as polymorphisms in trans-acting factors, environmental factors and tissue quality and preparation (11,12).

A number of recent studies have examined the extent of DAE in subsets of human genes using a variety of quantitative methods for allelic discrimination (12–15), and the results of these studies reveal that variation in cis-regulation is quite common in the human genome, with 20–50% of genes demonstrating DAE depending on the methodology used. Additionally, several studies have reported that for a percentage of the genes demonstrating DAE, the allelic imbalances were tissue specific, with differences in allelic expression reported for kidney and liver (15) and for spleen, liver and brain (16). Because these tissues are physiologically, developmentally and anatomically distinct from each other, however, it is not necessarily surprising that context specificity in cis-regulation was reported for a number of genes in these tissues.

We have previously observed that DAE of the BMP5 gene (chromosome 6p12.1, MIM 112265 [OMIM] ) is quite common in articular cartilage samples (unpublished data). BMP5 encodes bone morphogenetic protein 5, which is a member of the TGF-ß superfamily of secreted proteins. Members of this family are critically involved in synovial joint development, tissue homeostasis and a variety of cellular processes such as differentiation, proliferation and apoptosis (17,18). In the study reported here, we measured the allelic output of BMP5 in mesenchymal tissues obtained from the synovial joints of donors undergoing either a total knee replacement (TKR) or a total hip replacement (THR) for treatment of osteoarthritis (OA), a degenerative joint disease (MIM 165720 [OMIM] ). Our aim was to elucidate patterns of allelic expression in a panel of closely related tissues and in different anatomical regions of the same tissue to determine whether allelic expression was similar in developmentally similar environments.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 RESULTS
DISCUSSION
 MATERIALS AND METHODS
 SUPPLEMENTARY MATERIAL
 REFERENCES
 
DAE is common in mesenchymal tissues
From the 38 individual donors in our cohort who were genotyped for single nucleotide polymorphisms (SNPs) residing in the transcribed region of the BMP5 gene, 16 donors (nine TKRs and seven THRs) heterozygous for at least one BMP-transcribed SNP were identified. Total RNA was extracted from the various joint tissues of these 16 donors, and the mRNA was carried forth for the DAE analysis using one of the transcribed SNPs as a marker to distinguish the output from the two alleles. The corrected allelic ratios for the TKR and THR donors are summarized in Table 1. The 16 donors analyzed all demonstrated DAE at BMP5 in at least one of their tissues, and the greatest difference in allelic expression was for the proximal articular cartilage (PAC) sample of donor 16, with a corrected allelic ratio of 4.77:1 (P = 2 x 10–21).


View this table:
[in this window]
[in a new window]

  		Table 1. DAE analyses of BMP5 for the TKR donors 1–9 (A) and THR donors 10–16 (B)

 
To validate the reproducibility of our DAE assay, cDNA was synthesized from all of the tissues (n = 6) from donor 1 in two separate rounds of reverse transcription (RT) reactions, and there was excellent correlation between the corrected allelic ratios from the first RT reactions and those from the second RT reactions (Pearson correlation = 0.99, P = 3 x 10–5). Additionally, we validated the consistency of DAE measurements from different transcribed marker SNPs by assaying a donor with two independent marker SNPs. Donor 7 was identified as a compound heterozygote at the transcribed BMP5 marker SNPs rs3734444 and 846G > A. The corrected allelic ratios generated from rs3734444 for each of the tissues were significantly correlated with the corrected allelic ratios generated from 846G > A for donor 7 (Pearson correlation = 0.93, P = 0.003). These data imply that there is high correlation between the measurements generated from the various marker SNPs and that the data from one marker SNP can substitute for the data from a second marker SNP.

Variability in the tissues demonstrating DAE within the individual donors
We found that donors who demonstrated BMP5 DAE in one tissue did not necessarily demonstrate BMP5 DAE in all of the other tissues. This is clearly the case for donor 3, who demonstrated DAE in the distal articular cartilage (DAC), PAC, infrapatellar fat pad (IFP), meniscus (ME) and anterior cruciate ligament (ACL), but not in patellar tendon (PT) or synovium (SY) (Fig. 1A). In fact, of the 16 donors assayed, 11 demonstrated DAE in one or more of their synovial joint tissues but not in all tissues examined (donors 1–3, 5, 6, 8, 10–13 and 15). There was no pattern as to which tissues harbored differences in allelic expression, however, as some donors showed DAE only in their cartilage tissues (donors 1, 10 and 13), some showed equal allelic expression in cartilage but DAE in some other tissues (donors 5 and 15) and the majority of donors was found to have a range of allelic ratios in their various tissues.


Figure 1
View larger version (15K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Figure 1. DAE of the BMP5 gene in a variety of mesenchymal-derived tissues. The corrected allelic ratios between the two BMP5 alleles and the error bar for 1 SD are shown for the tissue specimens analyzed for donor 3 (A), donor 7 (B) and donor 14 (C). The marker SNP used to generate these data was rs3734444, and the allelic ratio shown is T/C. The gray bar demarcates a relative difference in expression of 20% between the T allele and the C allele. GD is genomic DNA.

 
It was also clear from our analysis that donors who demonstrated BMP5 DAE in a number of their tissues often had different allelic ratios between those tissues. For example, donor 7 was found to have altered allelic expression in all of the joint tissues, but with variable corrected allelic ratios as high as 2.19:1 (P = 3 x 10–28) in the DAC and 1.96:1 (P = 3 x 10–20) in the ME and as low as 1.49:1 (P = 1 x 10–15) in the PAC and 1.53:1 (P = 4 x 10–21) in the PT (Fig. 1B). Additionally, one of the donors (donor 14) was found to have DAE in all of the tissues assayed; however, the allele that was over-represented in the DAC and the PAC was the under-represented allele in the ligamentum teres femoris (LTF) and the SY (Fig. 1C).

Intra-tissue variation in DAE
For the 12 donors for whom both DAC and PAC samples were analyzed, highly significant differences (P ≤ 10–8) in corrected allelic ratios were found between these different cartilage samples for five of the donors (donors 1, 4, 6, 7 and 11), with donors 1 and 4 showing a greater allelic imbalance in PAC relative to DAC and donors 6 and 7 demonstrating the converse (Fig. 2). For donor 11, the DAC sample did not demonstrate DAE (corrected allelic ratio = 1.00 ± 0.07), whereas the PAC sample did (corrected allelic ratio = 1.48 ± 0.28), generating a highly significant difference between the two cartilage samples (P = 3 x 10–10) for this donor.


Figure 2
View larger version (23K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Figure 2. Comparison of corrected allelic ratios between the distal and proximal cartilage samples within individual donors. The corrected allelic ratios between the two BMP5 alleles and the error bar for 1 SD are shown for the distal and proximal cartilage samples for 12 donors. The marker SNPs used for the DAE analysis for each donor are given in Table 1. Comparison between the distal and proximal cartilage samples was performed using the two-tailed Mann–Whitney exact test. NS, not significant.

 
Mapping tissue-specific cis-regulatory elements
Our data clearly demonstrated highly significant differences both in the occurrence and in the degree of BMP5 DAE between the different mesenchymal tissues of many of our donors, which implies the existence of polymorphic, tissue-specific, cis-acting control elements. We next set out to identify these elements by correlating haplotype-tagging SNP genotypes with the DAE status. Since the tissues of each individual donor were assayed for the presence of DAE with the same marker SNP, we were able to make intra-individual comparisons of the allelic ratios across the various tissues. Additionally, because we found that the different marker SNPs yielded comparable corrected allelic ratios across the various tissues, we were able to look for tissue-specific trends across our entire donor cohort.

We combined the allelic ratios from our donor cohort into discrete tissue-specific cohorts. For these cohorts, the DAC, PAC and SY allelic ratios generated from the TKR donors were combined with the DAC, PAC and SY allelic ratios generated from the THR donors, respectively. Additionally, the ACL and LTF allelic ratios were combined to create a generic ligament cohort delineated as LI. We then selected haplotype-tagging SNPs across a 416 kb interval including the entire BMP5 gene, 50 kb upstream of the gene and 376 kb downstream of the gene. We limited our analysis to 50 kb upstream because of the known occurrence of an evolutionary breakpoint between humans and mice at this point (19), making it unlikely that conserved regulatory elements exist beyond it. In setting the downstream limit, we were guided by a study that reported an extensive 3' regulatory control region for the mouse Bmp5 gene, with regulatory sequences mapping up to 270 kb from the transcription start site (20). A homology search revealed this interval to be equivalent to 376 kb downstream of the transcription start site in humans.

We identified 44 haplotype-tagging SNPs with minor allele frequencies ≥ 20% in this 416 kb interval, and these SNPs were genotyped in each of our donors. The corrected allelic ratios for the tissue samples were transformed by taking the absolute value of the log10 to account for the bidirectional allelic imbalances evident in our cohort. The donors were then classified as homozygous or heterozygous for each SNP, and the transformed allelic ratios of the homozygotes were compared with those of the heterozygotes to assess whether there was a correlation between the heterozygous genotype and DAE. Our assumption here was that if a tagging SNP was a functional polymorphism, or in strong linkage disequilibrium (LD) with a functional polymorphism, then a heterozygote at that SNP would possess both a high-expressing and a low-expressing allele and would therefore demonstrate DAE. Conversely, a homozygote for the SNP would possess two copies of a low-expressing allele or two-copies of a high-expressing allele and would therefore not demonstrate DAE. This analysis is complicated by the possible existence of multiple functional polymorphisms on different haplotypic backgrounds, but a correlation should arise as one approaches the functional SNP.

One of the 44 SNPs genotyped showed significant correlation (nominal P < 0.05) between the heterozygote genotype and DAE status: rs9475394 (Table 2), which is located downstream of BMP5. This SNP was correlated with DAE status in four tissues: IFP, PT, LI and SY (Fig. 3). LD mapping of CEPH data using Haploview (21) and the Human HapMap release 21 revealed that this SNP lies in an LD block (D' = 1.0) of ~80 kb, and that within this block there are at least 13 other SNPs with a pair-wise r2 ≥ 0.8 with rs9475394 (data not shown). The exact location and P-values for the correlation between the genotype and DAE status for all of the 44 haplotype-tagging SNPs are listed in Supplementary Material, Table S1.


Figure 3
View larger version (16K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Figure 3. Plots of the absolute values of log10 transformed allelic ratios for AA homozygotes and CA heterozygotes at rs9475394 for the various tissue cohorts. Comparisons between the transformed allelic ratios between homozygotes and heterozygotes were made with the two-tailed Mann–Whitney exact test. LI, ligament (combination of ACL from TKRs and LTF from THRs).

 


View this table:
[in this window]
[in a new window]

  		Table 2. Genotypes at rs9475394 for each donor and absolute values of the log10-transformed allelic ratios for the various tissues used for the correlation analysis between the heterozygote genotype and DAE

 

   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 RESULTS
 DISCUSSION
MATERIALS AND METHODS
 SUPPLEMENTARY MATERIAL
 REFERENCES
 
In this study, we sought to elucidate allelic expression patterns in a group of anatomically and developmentally related tissues. As our model, we chose to investigate the allelic expression of BMP5 in a variety of mesenchymal-derived tissues that were made accessible during elective joint-replacement surgery. Our results indicate that DAE of BMP5 is very common in the tissues of synovial joints, but there is considerable inter-individual and intra-individual variation both to the existence and to the extent of an allelic imbalance in a particular tissue in a particular individual. Additionally, these results suggest that allelic expression is highly context dependent such that different cellular environments can produce significantly different allelic ratios, regardless of any developmental, anatomical and/or structural similarities among those environments.

We have also observed that there can be significant differences in allelic expression in different anatomical regions of the same tissue through our analysis of cartilage samples obtained from OA donors. OA is a disease characterized by the focal degeneration of the smooth articular cartilage surface in synovial joints, and the cartilage at the osteoarthritic lesion may undergo a number of changes such as fibrillation, matrix destabilization and de-differentiation of the chondrocytes to fibroblast-like cells (22). Thus, the cartilage defined as PAC (cartilage proximal to the OA lesion) may be environmentally and contextually distinct from the cartilage defined as DAC (cartilage distal to the OA lesion) within an individual donor, which could explain some of the observed differences in allelic expression in these two regions. These data suggest that alterations in cellular phenotype can be accompanied by changes in allelic expression patterns.

We genotyped 44 haplotype-tagging SNPs in our donors and correlated genotype at each SNP with BMP5 DAE status in a first attempt to map functional variants responsible for the tissue-specific expression patterns that we observed for this gene. Since our donor sample size was small (16 donors heterozygous for at least one BMP5-transcribed SNP), we chose tagging SNPs with relatively high minor allele frequencies. This would have limited our capacity to identify rare functional variants. Nevertheless, we did obtain nominally significant data for one SNP (rs9475394) in multiple tissues, but these data must be interpreted with caution because of our small sample size and the multiple tests performed. Although the significance was marginal and will need confirming in much larger sample sizes, these preliminary data are encouraging as they show that DAE is a mapable phenotype.

SNP rs9475394 was identified as correlating with DAE status in four different tissues: IFP, PT, LI and SY, which suggests that this variant, or a variant/s in LD with it, can influence BMP5 allelic expression in some tissues, but not in others. Such context specificity in the functional effects of cis-regulatory polymorphisms has been previously linked to allele-specific binding of nuclear factors in different cell types. For example, Koch et al. (9) reported that one allele of a cis-regulatory polymorphic site in the IFNGR1 gene showed a 4-fold reduction in gene expression in endothelial cells, a 30% increase in gene expression in B-lymphocytes, and no significant effect on expression in T-lymphocytes. Thus, such cell-specific control of expression may partially explain the tissue-specific DAE that we have observed with the BMP5 expression, as some tissues may utilize a cis-regulatory polymorphism for induction or suppression of expression depending on cell-specific signals.

Because we have observed no consistent pattern in allelic expression between the tissues within the individual donors in our cohort, however, we suspect that there may be multiple cis-acting regulatory polymorphisms interacting to govern the expression of BMP5. For example, it has recently been reported that certain haplotypic combinations of cis-regulatory polymorphisms in the KRT1 gene can explain a significant portion of the allele-specific expression differences of this gene, with each individual variant of the haplotype contributing only fractionally to the overall expression (23). Although rs9475394 demonstrated marginal association with DAE, additional LD mapping and experimental tests such as luciferase reporter assays and electrophoretic mobility shift assays will need to be performed to distinguish between the functional variant and those in LD with it and to determine whether there are in fact multiple cis-acting regulatory polymorphisms working independently or interacting in functional haplotypes to cause DAE.

A number of recent studies have highlighted the previously unappreciated prevalence of structural genetic variation including insertions, deletions and duplications, which have been collectively defined as copy-number variations (CNVs) (24,25). Copy-number variable regions (CNVRs), which can encompass genes, functional regulatory elements and conserved non-coding sequences, have been reported to impact gene expression and complex phenotypes. Such variants can lead to an imbalance in the level of expressed RNA at the allelic level and may therefore generate a DAE phenotype at a particular gene. In our study, we observed significant differences in BMP5 DAE status between closely related tissues within individuals. If there is a theoretical CNVR encompassing BMP5 contributing to the DAE phenotype, our data would suggest then that there would have to be a high level of somatic mutation to explain the intra-individual differences, which seems unlikely.

BMP5 is a secreted signaling molecule involved in the development of a number of diverse anatomical structures in vertebrates, such as the limb (26), ovary (27), heart (28) and skeleton (29). Through germline mutagenesis experiments in mice, an extensive 3' regulatory region containing a large number of discrete regulatory elements was elucidated for Bmp5 (20,30). These 3' regulatory elements displayed context specificity in their control of Bmp5 expression even in developmentally and anatomically related structures. For example, a particular 3' cis-regulatory element was found to profoundly affect the Bmp5 expression in the external ear but not in the middle ear, inner ear or temporal bone elements. Additionally, DiLeone et al. (30) reported that even similar patterns of Bmp5 expression in different mesenchymal tissues could be controlled by distinct cis-acting regulatory sequences. The results of our study evidence that the complex, context-specific control of the Bmp5 expression discovered in mice is also present in the regulation of human BMP5 expression. Moreover, as context specificity in the cis-regulation of Bmp5 has provided a molecular mechanism by which the morphology and development of particular skeletal structures is controlled in mice, the context-specific regulation that we report for BMP5 provides a plausible mechanism by which the development of complex structures such as synovial joints can occur in humans.

Thus, to gain the most accurate understanding of the cis-regulation of a particular gene, our findings suggest that one cannot assume that one tissue can authentically predict the allelic expression of another tissue. We have examined closely related, mesenchymal-derived tissues that were taken at the same time point for each donor, and we still observed highly significant differences in allelic expression within each individual donor. Furthermore, our analysis of cartilage sampled adjacent to and away from the osteoarthritic lesion has revealed that even for a particular tissue, one cannot assume that allelic expression patterns are consistent across that tissue. There has been a trend in analyses of complex traits to assign functional significance to putative cis-regulatory elements on the basis of in vitro experiments such as reporter gene assays and nuclear-factor-binding assays while largely ignoring the complexity and dynamics of gene expression in native tissues (2). Our analysis of the BMP5 gene, however, suggests that the most significant aspect of gene expression is the natural context in which it takes place.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 RESULTS
 DISCUSSION
 MATERIALS AND METHODS
SUPPLEMENTARY MATERIAL
 REFERENCES
 
Samples
Synovial joint tissue was obtained from 38 individuals (24 TKRs and 14 THRs) undergoing elective joint replacement for OA. For TKR and THR patients, two different cartilage specimens were obtained: DAC and PAC, with the osteoarthritic lesion used as a reference point for the collection of these samples. Thus, proximal cartilage was taken from an area adjacent to the lesion, whereas distal cartilage was sampled from an area away from the lesion. Other tissues collected from TKR patients were IFP, ME, PT, ACL and SY. For THR patients, the other tissues collected were SY and LTF. Ethical approval for the collection of these tissue specimens was granted by local Ethics Committees, and informed consent was obtained from each donor.

Nucleic acid extraction and quantitation
Within approximately 1 h of joint-replacement surgery, the tissue specimens were collected from the surgical theatres and snap-frozen in liquid nitrogen. For each individual tissue sample, 0.5–1.0 g of frozen tissue was subsequently ground to a powder using a Retsch mixermill 200 (Retsch Limited, Leeds, UK) under liquid nitrogen, which causes the sample to become fracturable and prevents the RNA from degrading. Genomic DNA and RNA were both extracted from the powderized tissue samples using a protocol established for articular cartilage, but effective for the other tissue specimens analyzed (31). Briefly, each 1 g of powderized tissue was mixed with 5 ml of lysis buffer RLT (Qiagen, Crawley, UK) supplemented with 10 µl/ml ß-mercaptoethanol (Sigma-Aldrich, Gillingham, UK) and centrifuged at 10 000g for 1 h. The supernatant was then mixed with one volume of 70% ethanol and spun through an RNeasy Midi column (Qiagen). The column was washed with buffer RW1 (Qiagen), and from this buffer RW1 wash, the genomic DNA was isolated. Column-bound RNA was then treated with 81 Kunitz units of DNase I (Qiagen) at room temperature for 25 min, washed once with buffer RW1, twice with buffer RPE (Qiagen) and eluted in diethylpyrocarbonate-treated water (Invitrogen, Paisley, UK). The extracted RNA was quantitated using the Ribogreen Quantitation Reagent (Molecular Probes, Inc., Eugene, USA) and/or a NanoDrop ND-1000 Spectrophotometer (NanoDrop Technologies, Wilmington, USA).

Genotyping
Genomic DNA isolated during the nucleic extraction from either the DAC or the PAC sample from each donor was used for genotyping at four transcribed BMP5 SNPs: a T/C SNP in exon 1 (rs3734444), an A/G SNP in exon 4 (846G > A) and an A/G SNP (rs3756990) and a C/A SNP (rs3756991), both in the 3' untranslated region. Heterozygosity was determined by PCR-restriction fragment-length polymorphism analysis. PCR was performed in a total reaction volume of 30 µl, containing 50 ng of genomic DNA, 1 x PCR buffer, MgCl2 at 1–2 mM, dNTPs at 0.2 mM, primers at 0.33 µM and 0.16 U of AmpliTaq® DNA polymerase (Applied Biosystems, Warrington, UK). Primer sequences, PCR conditions and restriction enzymes used for this genotyping are listed in Supplementary Material, Table S2.

DAE Analysis
To distinguish the mRNA output from the two copies of BMP5, a DAE analysis was conducted on individuals who were heterozygous for at least one of the genotyped SNPs located in the transcribed region of the gene. These transcribed SNPs were used as markers to delineate the mRNA derived from the two BMP5 alleles. cDNA was synthesized from the tissues of donors who were heterozygous at one of the transcribed SNPs. Briefly, 2 µg of total RNA was incubated with 50 ng of random hexamers and dNTPs at 1.25 mM for 5 min at 65ºC. To each reaction, MgCl2 at 5.26 mM, 1 x first-strand synthesis buffer (Invitrogen), DTT at 10.53 mM and 40 U of RNaseOUTTM inhibitor (Invitrogen) were added and incubated at 25ºC for 1 min. Then, 12.5 U of SuperScriptTM RT enzyme (Invitrogen) were added to each reaction and incubated at 25ºC for 10 min, 42ºC for 50 min and 70ºC for 15 min. Finally, 3.33 U of RNaseH (NEB, Hitchin, UK) were added to the reaction and incubated at 37ºC for 20 min.

This cDNA was then PCR amplified using primers specific for the transcribed SNP at which the donor was heterozygous. Blood genomic DNA from an individual heterozygous for the SNP of interest was also PCR amplified. The primer sequences for these PCR amplifications are listed in Supplementary Material, Table S3. The primer-binding sites were sequenced in 48 individuals (96 chromosomes) to ensure that there were no polymorphisms that could influence allele-specific binding of the primers. PCR was performed in a total reaction volume of 15 µl, containing either 1 µl of cDNA or 50 ng of genomic DNA, 1 x PCR buffer, MgCl2 at 1–2 mM, dNTPs at 0.2 mM, primers at 0.33 µM and 0.08 U of AmpliTaq® DNA polymerase (Applied Biosystems). Thermocycling conditions were an initial denaturation at 96ºC for 10 min, followed by 35 cycles of 96ºC for 1 min, annealing for 1 min and 72ºC for 1 min, with a final extension at 72ºC for 10 min. Annealing temperatures are listed in Supplementary Material, Table S3.

From each RT reaction, 20 individual cDNA PCR amplifications were carried out. Minus RT controls did not yield detectable products. Additionally, 20 individual PCR amplifications of the blood genomic DNA were carried out on six different occasions to give a total of 120 genomic DNA PCR amplifications. These PCR products were incubated with 1.5 U of shrimp alkaline phosphatase (SAP) (Amersham, Little Chalfont, UK) and 3 U of exonuclease I (EXO) (Amersham) at 37ºC for 1 h, followed by 15 min at 75ºC. These products then served as the template used to test for DAE by a single-base extension (SBE) reaction using the ABI Prism SNaPshot Multiplex kit (Applied Biosystems). SBE involved the extension by a single base of a primer located immediately upstream of the SNP of interest. The SBE primer sequences are listed in Supplementary Material, Table S4. The primer was extended into the polymorphic site using fluorescently labeled dideoxy nucleotides (ddNTPs). Each ddNTP had a unique fluorescently labeled tag that allowed for discrimination of the extension products from the two BMP5 alleles. SNaPshot primer extension reactions were performed in a total reaction volume of 3 µl, containing 1.2 µl of the SAP/EXO-treated product, 1.5 µl of SNaPshot reaction mix and extension primer at 1 µM. Thermocycling conditions were 25 cycles of 96ºC for 10 s, 50ºC for 5 s and 60ºC for 30 s. Following this reaction, extension products were treated with 1.0 U of SAP for 1 h at 37ºC, followed by 15 min at 75ºC.

One microliter of the fluorescently labeled extension products was combined with 9 µl of Hi-Di formamide, and these products were separated by electrophoresis through a 36 cm capillary array with POP-4TM polymer (Applied Biosystems) at 60ºC using an Applied Biosystems 3100 Genetic Analyzer. These data were analyzed using GeneMapper 3.5 software (Applied Biosystems) that produced electropherograms that listed peak heights for the two alleles, which are proportional to the amount of each allele present. The blood genomic DNA products were used to ascertain the peak pattern for an assumed 1:1 ratio between alleles. For each transcribed SNP, 20 individual SBE reactions were performed on six different occasions to give a total of 120 measurements for genomic DNA. Six blocks were performed to account for the systematic variation encountered in the steps of the DAE analysis. For the cDNA from the tissue samples, one block of 20 replicates was subjected to SBE.

We have assumed that genomic DNA samples contained an allelic ratio of 1:1, but because of differences in fluorescent yield and terminator dye incorporation, the measured peak height ratios for genomic DNA samples deviated from the 1:1 ratio. Thus, the average genomic DNA peak height ratio of the 120 individual measurements for each SNP was used to correct each DNA and cDNA measurement by the following equation: corrected allelic ratio = cDNA peak height ratio (or genomic DNA peak height ratio)/average genomic DNA peak height ratio.

Statistical analysis of DAE data
A particular tissue sample was considered to demonstrate DAE if all the following criteria were met: (i) after correction by the average genomic DNA peak height ratio, the fold difference between the alleles was ≥ 20%; (ii) the distribution of the 20 corrected cDNA replicates was significantly different than the distribution of the 120 corrected DNA replicates at P < 0.01 (two-tailed Mann–Whitney exact test); (iii) the 95% confidence interval for the mean of the 20 corrected cDNA replicates did not overlap the 95% confidence interval for the mean of the 120 corrected genomic DNA replicates. A 20% fold difference was set as the threshold, as it has been demonstrated previously that the quantitative allelic analysis performed here can detect differences in allelic output at or greater than this threshold (12,13). Furthermore, we verified that this threshold can be accurately measured with a standard curve experiment for marker SNP rs3756990 using mixtures of homozygous genomic DNA to create PCR reactions containing 20, 30, 40, 50, 60, 70 and 80% of one allele relative to the other. Five replicates were tested for each mixture at two different dilutions: 1:0 and 1:10. We obtained an R2 of 0.99 against the standards for both dilutions. Statistical analysis was performed using SPSS 13.0 software for Windows.

Mapping tissue-specific cis-regulatory elements
Haploview (21) was used to analyze CEPH genotyping data from the Human HapMap release 21 (http://www.hapmap.org) to identify haplotype-tagging SNPs within the 416 kb interval encompassing and flanking BMP5. Using pair-wise tagging, we identified 44 haplotype-tagging SNPs on the basis of inclusion criteria of r2 ≥ 0.8 and a minor allele frequency ≥ 20%. We set the cutoff at the relatively high level of 20% because our tissue-specific cohorts were small in size, making it extremely unlikely that we would be able to detect effects mediated by rare alleles. The tagging SNPs were genotyped in our donors by restriction fragment length polymorphism analysis or by direct DNA sequencing (primer sequences, PCR conditions and restriction enzymes are available on request). Corrected allelic ratios were transformed by taking the absolute value of log10. This was necessary because our samples demonstrated bidirectional allelic imbalances, which can result from allelic heterogeneity at one or more cis-regulatory sites or from incomplete LD between the marker SNP and a cis-regulatory polymorphism, assuming that there is only one causative polymorphism (14). Comparisons were made between the transformed DAE folds of homozygotes and heterozygotes for each SNP using the two-tailed Mann–Whitney exact test.


   SUPPLEMENTARY MATERIAL
TOP
 ABSTRACT
 INTRODUCTION
 RESULTS
 DISCUSSION
 MATERIALS AND METHODS
 SUPPLEMENTARY MATERIAL
REFERENCES
 
Supplementary Material is available at HMG Online.


   ACKNOWLEDGEMENTS
 
We would like to thank Professor Nicholas Athanasou, Mr Stephen McDonnell, Mr Richard Benson and Mr Timothy Matthews for the collection of tissue samples used in this study. The Marshall Aid Commemoration Commission (Marshall Scholarship to J.M.W.), the Arthritis Research Campaign and Research into Ageing supported this work.

Conflict of Interest statement. None declared.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 RESULTS
 DISCUSSION
 MATERIALS AND METHODS
 SUPPLEMENTARY MATERIAL
 REFERENCES
 

  1. Wray G.A., Hahn M.W., Abouheif E., Balhoff J.P., Pizer M., Rockman M.V., Romano L.A. (2003) The evolution of transcriptional regulation in eukaryotes. Mol. Biol. Evol. 20:1377–1419.[Abstract/Free Full Text]

  2. Knight J.C. (2005) Regulatory polymorphisms underlying complex disease traits. J. Mol. Med. 83:97–109.[CrossRef][ISI][Medline]

  3. Ayala F.J., Escalante A., O'Huigin C., Klein J. (1994) Molecular genetics of speciation and human origins. Proc. Natl. Acad. Sci. USA 91:6787–6794.[Abstract/Free Full Text]

  4. Rockman M.V. and Wray G.A. (2002) Abundant raw material for cis-regulatory evolution in humans. Mol. Biol. Evol. 19:1991–2004.[Abstract/Free Full Text]

  5. Cargill M., Altshuler D., Ireland J., Sklar P., Ardlie K., Patil N., Shaw N., Lane C.R., Lim E.P., Kalyanaraman N., et al. (1999) Characterization of single-nucleotide polymorphisms in coding regions of human genes. Nat. Genet. 22:231–238.[CrossRef][ISI][Medline]

  6. Halushka M.K., Fan J.B., Bentley K., Hsie L., Shen N., Weder A., Cooper R., Lipshutz R., Chakravarti A. (1999) Patterns of single-nucleotide polymorphisms in candidate genes for blood-pressure homeostasis. Nat. Genet. 22:239–247.[CrossRef][ISI][Medline]

  7. Liu J. and Francke U. (2006) Identification of cis-regulatory elements for MECP2 expression. Hum. Mol. Genet. 15:1769–1782.[Abstract/Free Full Text]

  8. Kleinjan D.A. and van Heyningen V. (2005) Long-range control of gene expression: emerging mechanisms and disruption in disease. Am. J. Hum. Genet. 76:8–32.[CrossRef][ISI][Medline]

  9. Koch O., Kwiatkowski D.P., Udalova I.A. (2006) Context-specific functional effects of IFNGR1 promoter polymorphism. Hum. Mol. Genet. 15:1475–1481.[Abstract/Free Full Text]

  10. Qiao L., Maclean P.S., Schaack J., Orlicky D.J., Darimont C., Pagliassotti M., Friedman J.E., Shao J. (2005) C/EBPalpha regulates human adiponectin gene transcription through an intronic enhancer. Diabetes 54:1744–1754.[Abstract/Free Full Text]

  11. Stamatoyannopoulos J.A. (2004) The genomics of gene expression. Genomics 84:449–457.[CrossRef][ISI][Medline]

  12. Bray N.J., Buckland P.R., Owen M.J., O'Donovan M.C. (2003) Cis-acting variation in the expression of a high proportion of genes in human brain. Hum. Genet. 113:149–153.[ISI][Medline]

  13. Yan H., Yuan W., Velculescu V.E., Vogelstein B., Kinzler K.W. (2002) Allelic variation in human gene expression. Science 297:1143.[Free Full Text]

  14. Pastinen T., Sladek R., Gurd S., Sammak A., Ge B., Lepage P., Lavergne K., Villeneuve A., Gaudin T., Brandstrom H., et al. (2004) A survey of genetic and epigenetic variation affecting human gene expression. Physiol. Genomics 16:184–193.[Abstract/Free Full Text]

  15. Lo H.S., Wang Z., Hu Y., Yang H.H., Gere S., Buetow K.H., Lee M.P. (2003) Allelic variation in gene expression is common in the human genome. Genome Res. 13:1855–1862.[Abstract/Free Full Text]

  16. Cowles C.R., Hirschhorn J.N., Altshuler D., Lander E.S. (2002) Detection of regulatory variation in mouse genes. Nat. Genet. 32:432–437.[CrossRef][ISI][Medline]

  17. Chen Y.G. and Meng A.M. (2004) Negative regulation of TGF-ß signaling in development. Cell Res. 14:441–449.[CrossRef][ISI][Medline]

  18. Francis-West P.H., Parish J., Lee K., Archer C.W. (1999) BMP/GDF-signaling interactions during synovial joint development. Cell Tissue Res. 296:111–119.[CrossRef][ISI][Medline]

  19. Fitzgerald J. and Bateman J.F. (2004) Why mice have lost genes for COL21A1, STK17A, GPR145 and AHRI: evidence for gene deletion at evolutionary breakpoints in the rodent lineage. Trends Genet. 20:408–412.[CrossRef][ISI][Medline]

  20. DiLeone R.J., Russell L.B., Kingsley D.M. (1998) An extensive 3' regulatory region controls expression of Bmp5 in specific anatomical structures of the mouse embryo. Genetics 148:401–408.[Abstract/Free Full Text]

  21. Barrett J.C., Fry B., Maller J., Daly M.J. (2005) Haploview: analysis and visualization of LD and haplotype maps. Bioinformatics 15:263–265.

  22. Aigner T. and McKenna L. (2002) Molecular pathology and pathobiology of osteoarthritic cartilage. Cell Mol. Life Sci. 59:5–18.[CrossRef][ISI][Medline]

  23. Tao H., Cox D.R., Frazer K.A. (2006) Allele-specific KRT1 expression is a complex trait. PLoS Genet. 2:e93.[CrossRef][Medline]

  24. Redon R., Ishikawa S., Fitch K.R., Feuk L., Perry G.H., Andrews T.D., Fiegler H., Shapero M.H., Carson A.R., Chen W., et al. (2006) Global variation in copy number in the human genome. Nature 444:444–454.[CrossRef][Medline]

  25. Feuk L., Marshall C.R., Wintle R.F., Scherer S.W. (2006) Structural variants: changing the landscape of chromosomes and design of disease studies. Hum. Mol. Genet. 15:R57–R66.[Abstract/Free Full Text]

  26. Zuzarte-Luis V., Montero J.A., Rodriguez-Leon J., Merino R., Rodriguez-Rey J.C., Hurle J.M. (2004) A new role for BMP5 during limb development acting through the synergic activation of Smad and MAPK pathways. Dev. Biol. 272:39–52.[CrossRef][ISI][Medline]

  27. Shimizu T., Yokoo M., Miyake Y., Sasada H., Sato E. (2004) Differential expression of bone morphogenetic protein 4–6 (BMP-4, -5 and -6) and growth differentiation factor-9 (GDF-9) during ovarian development in neonatal pigs. Domest. Anim. Endocrinol. 27:397–405.[CrossRef][ISI][Medline]

  28. Yamagishi T., Nakajima Y., Nishimatsu S., Nohno T., Ando K., Nakamura H. (2001) Expression of bone morphogenetic protein-5 gene during chick heart development: possible roles in valvuloseptal endocardial cushion formation. Anat. Rec. 264:313–316.[CrossRef][Medline]

  29. King J.A., Marker P.C., Seung K.J., Kingsley D.M. (1994) BMP5 and the molecular, skeletal, and soft-tissue alterations in short ear mice. Dev. Biol. 166:112–122.[CrossRef][ISI][Medline]

  30. DiLeone R.J., Marcus G.A., Johnson M.D., Kingsley D.M. (2000) Efficient studies of long-distance Bmp5 gene regulation using bacterial artificial chromosomes. Proc. Natl Acad. Sci. USA 97:1612–1617.[Abstract/Free Full Text]

  31. McKenna L.A., Gehrsitz A., Soder S., Eger W., Kirchner T., Aigner T. (2000) Effective isolation of high-quality total RNA from human adult articular cartilage. Anal. Biochem. 286:80–85.[CrossRef][ISI][Medline]


Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us    What's this?


This article has been cited by other articles:


Home page
 Schizophr BullHome page
N. J. Bray
Gene Expression in the Etiology of Schizophrenia
Schizophr Bull, May 1, 2008; 34(3): 412 - 418.
[Abstract] [Full Text] [PDF]

Home page
 Hum Mol GenetHome page
L. Southam, J. Rodriguez-Lopez, J. M. Wilkins, M. Pombo-Suarez, S. Snelling, J. J. Gomez-Reino, K. Chapman, A. Gonzalez, and J. Loughlin
An SNP in the 5'-UTR of GDF5 is associated with osteoarthritis susceptibility in Europeans and with in vivo differences in allelic expression in articular cartilage
Hum. Mol. Genet., September 15, 2007; 16(18): 2226 - 2232.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow Supplementary Data
Right arrow All Versions of this Article:
16/5/537    most recent
ddl488v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (3)
Right arrowRequest Permissions
Google Scholar
Right arrow Articles by Wilkins, J. M.
Right arrow Articles by Loughlin, J.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Wilkins, J. M.
Right arrow Articles by Loughlin, J.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?