Skip Navigation


Human Molecular Genetics Advance Access originally published online on May 24, 2006
Human Molecular Genetics 2006 15(13):2098-2105; doi:10.1093/hmg/ddl133
This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow Supplementary Material
Right arrow All Versions of this Article:
15/13/2098    most recent
ddl133v1
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 (6)
Right arrowRequest Permissions
Google Scholar
Right arrow Articles by Bernat, J. A.
Right arrow Articles by Kondrashov, A. S.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Bernat, J. A.
Right arrow Articles by Kondrashov, A. S.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

© The Author 2006. Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oxfordjournals.org

Distant conserved sequences flanking endothelial-specific promoters contain tissue-specific DNase-hypersensitive sites and over-represented motifs

John A. Bernat1, Gregory E. Crawford3,{dagger}, Aleksey Y. Ogurtsov4, Francis S. Collins3, David Ginsburg1,2,* and Alexey S. Kondrashov4

1 Department of Human Genetics and 2 Department of Internal Medicine, Howard Hughes Medical Institute, and Life Sciences Institute, University of Michigan, Ann Arbor, MI 48109, USA, 3 National Human Genome Research Institute and 4 National Center for Biotechnology Information, National Institutes of Health, Bethesda, MD 20892, USA

* To whom correspondence should be addressed. Tel: +1 7346474808; fax: +1 7349362888; Email: ginsburg{at}umich.edu

Received January 26, 2006; Revised April 4, 2006; Accepted May 17, 2006


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 RESULTS
 DISCUSSION
 MATERIALS AND METHODS
 SUPPLEMENTARY MATERIAL
 REFERENCES
 
The transcriptional regulation of genes is a complex process, particularly for genes exhibiting a tissue-specific pattern of expression. We studied 28 genes that are expressed primarily in endothelial cells, another 28 genes that are expressed highly, but not exclusively, in cultured endothelial cells, and three control sets, consisting of genes not expressed in endothelium, genes expressed in neural tissues and housekeeping genes. For each gene, we identified conserved non-coding sequences (CNSs) of lengths 50 to >1000 nucleotides, located within the upstream intergenic region (from 500 to as far as 200 000 nucleotides upstream from the transcription start) or within the first intron. As a functional test, we assayed the CNSs from the set of endothelial cell-specific genes (EC-CNSs) for DNase hypersensitivity. Among 262 distant EC-CNSs, 33% are hypersensitive (HS) in endothelial cells, whereas only 16% are HS in control fibroblasts. A search for short sequence patterns revealed a number of motifs which are over-represented in EC-CNSs relative to CNSs from the control gene sets. In particular, the motif SAGGAAR is strongly and consistently over-represented among EC-CNSs, and is more over-represented in HS CNSs than in non-HS CNSs. CNSs which contain this motif are no closer to the promoter than an average CNS. This motif contains the core element of binding sites from the Ets family of transcription factors. Thus, one or several factors from this family may play a key role in the regulation of endothelial gene expression.


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
RESULTS
 DISCUSSION
 MATERIALS AND METHODS
 SUPPLEMENTARY MATERIAL
 REFERENCES
 
The regulation of transcription in complex multicellular organisms, such as mammals, is a very elaborate process. In contrast to yeast, in which the upstream activating sequences that participate in the regulation of a particular promoter are located relatively close to it (1,2), mammalian promoters can be regulated by more distant sequences. For example, expression of von Willebrand factor (VWF), a highly tissue-specific gene, appears to depend substantially on such distant sequences. VWF is expressed at varying levels in nearly all vascular endothelial cells (3), and its expression is restricted to endothelial cells (4) and megakaryocytes (5). Transgenic mice carrying only proximal regulatory sequences (from –2.6 kb upstream of the transcription start site to the end of intron 1) express VWF in some, but not all, populations of endothelial cells (6). In vertebrate genomes, long intergenic regions in the order of 104 to 105 nucleotides (~76 kb for human VWF) (7) provide ample room for such distant regulatory sequences.

A substantial fraction of non-protein coding sequences, which comprise ~99% of a mammalian genome (8), appear to be functionally important. Many genomic segments with functional attributes are controlled by negative selection and, thus, evolve slowly. Alignments of orthologous regions from moderately distant genomes reveal that at least 5–10% of intergenic human sequences are under selective constraint (8–11); this estimate falls to ~1% when considering extremely conserved sequences (12). Conserved, and presumably functionally important, genome segments differ from the rest of the genome. In particular, a number of short (5–20 nucleotides) sequence motifs are over-represented in human conserved non-coding sequences (CNSs) (13). At least some CNSs regulate transcription from promoters of adjacent genes, often located at considerable distances. For example, key enhancer elements that control expression of the human ß-globin gene reside in a locus control region 50–70 kb upstream of the ß-globin transcription start site (7,14,15).

One way to characterize the function of a CNS is to assay for hypersensitivity to DNase I cleavage. The presence of a DNase hypersensitive (HS) site is thought to indicate an open chromatin state and to suggest the binding of a transcription regulatory complex (16,17). CNSs can be assayed for DNase hypersensitivity in a high-throughput way by performing quantitative PCR across or within the CNS (18,19). If a CNS is hypersensitive, it will require additional cycles of PCR to amplify an equivalent amount of product from DNase-digested template DNA than from undigested template DNA.

Here, we report data on the role of distant CNSs in the regulation of transcription of endothelial genes. We show that many such CNSs are DNase-hypersensitive, especially in endothelial cells, and that they contain characteristic sequence motifs.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 RESULTS
DISCUSSION
 MATERIALS AND METHODS
 SUPPLEMENTARY MATERIAL
 REFERENCES
 
Conserved non-coding sequences
Human–mouse sequence comparison reveals a substantial number of CNSs, adjacent to all the analyzed genes (Table 1). Most CNSs are located in intergenic regions, but the first introns also contain a substantial fraction of CNSs. In fact, for each gene set the density (per kb) of CNSs in the first introns is higher than in the intergenic regions. The GC content of the CNSs (~40%) was not substantially different from the average GC content of the analyzed intergenic regions.


View this table:
[in this window]
[in a new window]
  		Table 1. Overall characteristics of CNSs

 
DNase hypersensitivity of EC-CNSs
Using quantitative PCR, we assayed each of the 262 CNSs adjacent to the 28 endothelial cell-specific genes from set endothelial-1 (EC-CNSs) for DNase I hypersensitivity (Table 2 and Fig. 1). Thirty-three percent of the EC-CNSs from successful reactions (79/240) were hypersensitive in cultured human umbilical vein endothelial cells (HUVECs), and 16% (36/226) were hypersensitive in cultured primary human foreskin fibroblast cells (HFFs). Twenty percent of the EC-CNSs scorable in both cell types (45/222) were hypersensitive only in endothelial cells, suggesting that this group may contain novel endothelial-specific transcriptional elements.


Figure 1331
View larger version (33K):
[in this window]
[in a new window]
  		Figure 1. DNase hypersensitivity analysis of 262 EC-CNSs. Each of the 262 EC-CNSs is represented in the Venn diagram above. HS, hypersensitive; not HS, not hypersensitive; HUVECs, human umbilical vein endothelial cells; HFFs, human foreskin fibroblasts. For a particular EC-CNS, if the reaction to determine DNase hypersensitivity was successful in both cell types, it is counted in an intersection between two circles. If a reaction failed in one cell type, the EC-CNS is counted in a single circle. For failures in both cell types, the EC-CNS is counted outside of all circles (at the center of the diagram).

 


View this table:
[in this window]
[in a new window]
  		Table 2. DNase hypersensitivity analysis of 262 EC-CNSs

 
Figure 2 shows the locations, relative to the transcription start site, of various classes of HS EC-CNSs as well as the set of all 262 EC-CNSs. EC-CNSs HS in HUVECs or HFFs are generally distributed similarly when compared with all EC-CNSs, although the former include few very distal EC-CNSs. As these distal CNSs are few in number, this difference may not be significant.


Figure 1332
View larger version (42K):
[in this window]
[in a new window]
  		Figure 2. Locations of hypersensitive EC-CNSs and all EC-CNSs. The histogram shown represents the locations of various subsets or the set of all EC-CNSs, relative to the transcription start sites of the corresponding genes. Within the set or each subset, each EC-CNS was placed into one of 26 bins of width 10 kb, according to its location as measured from the transcription start site. The first bin contains EC-CNSs greater than 150 kb upstream, and the last bin contains EC-CNSs greater than 90 kb downstream. The number of EC-CNSs in each bin was divided by the number of genes having an intergenic region or first intron of at least that size, as the most distal bins were represented in few genes. All EC-CNSs denotes the set of 262 EC-CNSs identified from set endothelial-1. HUVEC HS EC-CNSs denotes the subset of 79 EC-CNSs hypersensitive in HUVECs. HUVEC-specific HS EC-CNSs denotes the subset of 45 EC-CNSs hypersensitive in HUVECs but not hypersensitive in HFFs. HFF HS EC-CNSs denotes the subset of 36 EC-CNSs hypersensitive in HFFs.

 
Motifs in CNSs
Table 3 lists the 10 most over-represented motifs of width 7 found by the discriminating matrix enumerator (DME) (20) in CNSs from sets endothelial-1 and endothelial-2, which were analyzed independently. The CNSs from the three control gene sets (non-endothelial, neural and housekeeping) were combined and used as a background set in both analyses.


View this table:
[in this window]
[in a new window]
  		Table 3. Over-represented motifs in CNSs near endothelial genes

 
A large fraction of over-represented motifs found in both endothelial sets of CNSs is, in fact, a variation on one common theme (Table 4). This abundant endothelial motif can be represented as SAGGAAR. Motif TGASTCA, which is highest scoring in set 2, is a palindrome, so that each of its occurrences is counted twice. This may inflate the significance of its over-representation. Other over-represented motifs are set-specific.


View this table:
[in this window]
[in a new window]
  		Table 4. A common motif, SAGGAAR, appears in multiple over-represented motifs

 
Figure 3 shows the occurrences of the SAGGAAR motif in CNSs from each endothelial gene set as a function of the distance from the transcription start site. The location of all CNSs in the set are also shown for comparison. There is no substantial difference between the distribution of the SAGGAAR motif in CNSs and the distribution of all CNSs in either endothelial gene set.


Figure 1333
View larger version (54K):
[in this window]
[in a new window]
  		Figure 3. Locations of the SAGGAAR motif in CNSs from each of the two endothelial gene sets. The histograms shown represent the locations of the set of 262 EC-CNSs found in set endothelial-1, the set of 356 CNSs found in set endothelial-2 and the locations of the SAGGAAR motifs detected by DME in each set, relative to the transcription start sites of the corresponding genes. Each CNS or motif was binned in the same manner as in the previous figure (Fig. 2).

 
The SAGGAAR motif is also over-represented when either endothelial set of CNSs is compared against a background set of randomly generated sequences. However, such comparison reveals several other motifs, which are not over-represented when CNSs from the three control gene sets are used as the background set (data not shown). It might be expected that all CNS sets are enriched for some generic regulatory motifs which, however, do not distinguish endothelial CNSs from others. In addition, using an independent algorithm (a standard single-link cluster analysis of over-represented motifs found by pairwise alignment of the same human–mouse alignments) (13), we again identified the SAGGAAR motif as the most common endothelial motif (data not shown).

Finally, we searched for over-represented motifs in the subset of EC-CNSs hypersensitive only in HUVECs, and the subset of EC-CNSs not hypersensitive in either HUVECs or HFFs, both against a background set of randomly generated sequences. The SAGGAAR motif was found in both, but over-represented to a greater degree in the HUVEC-specific hypersensitive set of EC-CNSs (Table 5).


View this table:
[in this window]
[in a new window]
  		Table 5. SAGGAAR is more over-represented in HUVEC-specific HS EC-CNSs than in non-HS EC-CNSs

 

   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 RESULTS
 DISCUSSION
MATERIALS AND METHODS
 SUPPLEMENTARY MATERIAL
 REFERENCES
 
Human–mouse comparisons of sequences flanking endothelial-specific and control genes revealed a number of CNSs. For the average number and the overall length of the CNSs which correspond to a protein-coding gene (Table 1), we obtain values which are consistent with previous estimates (9,13). When even weakly conserved sequences are taken into account, CNSs cover >10% of non-coding DNA (9). In contrast, if only strongly conserved sequences are considered, as in the current analysis, CNSs cover ~3% of non-coding DNA (8). CNSs frequently extend many hundreds of nucleotides in length and reach up to 90% human–mouse identity. Some genes had very many more elements than other genes, though roughly correlated to the length of the intergenic region. We excluded the proximal 500 nt of intergenic sequence as containing core promoter elements which would be expected to be conserved among genes of all types (not just tissue-specific). Thus, we considered only CNSs more distant from the start of transcription, located either upstream of these sites or in the first introns of the corresponding genes.

As the regulation of transcription is likely to be very similar in human and mouse, we would expect important transcriptional elements to be found among CNSs. The length and number of CNSs associated with a housekeeping gene is much smaller than in genes with tissue-specific expression (Table 1). However, the intergenic regions upstream of housekeeping genes are correspondingly shorter (a pattern previously reported (21,22)), and the CNSs identified account for a comparable proportion of intergenic sequence as in the other gene sets. Given that housekeeping genes presumably require a less complex gene expression program, it follows that they would also require fewer CNSs.

The presence of a DNase-hypersensitive site within a CNS suggests functional significance within the corresponding tissue. In testing the 262 EC-CNSs from endothelial set 1 for DNase hypersensitivity in fibroblasts, we found the background level of hypersensitivity to be ~16% (Table 2). Some of the 28 genes which we considered to be endothelial-specific may have some background level of transcription in fibroblasts, or illegitimate transcription in cell culture. This may contribute to the fraction of endothelial CNSs that were found to be hypersensitive in fibroblasts. Also, eight EC-CNSs were found to be hypersensitive in fibroblasts but not in endothelial cells. It is possible that these sequences are involved in the transcription of a nearby fibroblast-expressed gene but do not affect the neighboring endothelial gene.

The higher percentage of EC-CNSs that were hypersensitive in HUVECs (33%) than in HFFs (16%) suggests that many of these EC-CNSs may regulate endothelial-specific gene expression. In addition, not all of the endothelial-specific genes may have been expressed at in vivo levels in the cultured HUVECs. There is a great deal of diversity among endothelial cells, and although HUVECs are commonly used for in vitro experiments, it is known that expression levels of endothelial genes vary among endothelial subtypes (3). Also, non-HS EC-CNSs from genes that had few or no HS EC-CNSs may in fact be important for endogenous gene expression.

Transcription factors are generally envisioned to bind cooperatively to sequences that correspond to short and often degenerate motifs. Finding such functional motifs is a key challenge of the post-genome era. Our comparative analysis of CNSs from the two endothelial sets versus control CNSs revealed one primary motif, SAGGAAR, that is over-represented in endothelial CNSs (Table 3). In fact, several of its variants, differing from each other by nucleotide shifts, are often found by DME separately (Table 4). Although an individual motif might be over-represented by chance, it is unlikely that the same core motif will be over-represented several times in the same data set. Furthermore, the same motif is found in a second independent set of conserved sequences from genes with high endothelial expression. Analysis of our control sets also finds over-represented motifs, but no one motif is found in multiple high-scoring variations, as was the case for the endothelial sets. In addition, SAGGAAR is more over-represented in HS EC-CNSs than in non-HS EC-CNSs (Table 5). Thus, we believe that this motif has functional significance in the endothelial gene expression program.

SAGGAAR occurs in both intergenic regions and first introns with approximately the same distribution as the endothelial sets of CNSs (Fig. 3), suggesting that it may act in both proximal and long-range gene regulation. SAGGAAR contains the core binding sequence (GGAW) of the Ets family of transcription factors, which contains over 20 members. At least three of these members are known to be expressed in endothelial cells and thought to be involved in the regulation of endothelial gene expression (23–25). A specific Ets family member may be critical for determining the endothelial-specific gene expression program. For example, VWF (one of our 28 endothelial-specific genes) contains several known Ets sites in its promoter, one of which has been shown to activate expression in vitro (23). In addition, our analysis has discovered a pair of conserved Ets sites 57 kb upstream from the VWF transcription start, and a second pair, 1.7 kb downstream from the first pair. The functional importance of these sites is currently being examined.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 RESULTS
 DISCUSSION
 MATERIALS AND METHODS
SUPPLEMENTARY MATERIAL
 REFERENCES
 
Gene sets
We assembled five sets of genes as follows:

A set of 28 genes (endothelial-1) with well-characterized endothelial-specific expression was compiled from a microarray study (26) (19 genes) as well as a survey of the literature (nine additional genes).

A set of 28 genes (endothelial-2) with high endothelial expression was identified by searching SymAtlas (Genomics Institute of the Novartis Research Foundation) (27), Human GeneAtlas U95A, for genes with expression in HUVECs greater than 10 times the median expression (26 genes), and for genes that correlated with the VWF expression pattern (R>0.7, two additional genes).

A control set of 33 genes with very low or no endothelial expression was identified by searching SymAtlas (27), Human GeneAtlas U95A, for genes with expression in HUVECs 20 times less than the median expression and with an average difference value of less than 100. One human gene, CGB, lacked a clear murine ortholog and was excluded, resulting in a total of 32 genes for this set.

A control set of 20 genes with high neural expression was identified by searching SymAtlas (27), Human GeneAtlas U95A, for genes with brain expression greater than 30 times the median expression and HUVEC expression less than 0.5 times the median.

A control set of 18 housekeeping genes was identified from the literature, using three studies that examined endogenously expressed control genes (28–30).

Comparative sequence analysis
For each of the genes above, we obtained orthologous human and mouse sequences beginning at the proximal boundary (transcription start or end of last exon) of the adjacent 5' gene and extending to the end of intron 1 of the gene of interest. The resulting sequences encompassed the entire 5' intergenic region as well as exon 1 and intron 1. For genes with very large (>400 kb) intergenic regions, only the proximal 200 kb (for set endothelial-1) or 100 kb (for the remaining sets) of human sequence (and the orthologous mouse sequence) were used. For the one gene with a very large (>200 kb) intron 1, only the proximal 100 kb was used. The orthologous sequences were repeatmasked [RepeatMasker (31)] and then aligned using OWEN (32), using an initial maximal P-value of 10–6, requiring 10 matches to begin an alignment, and extending the alignment when at least six of eight positions matched. Alignments were manually curated, and those with higher identity and greater length were selected as CNSs. For genes with intergenic regions and first introns with high human–mouse similarity, we chose the best alignments (generally <30, with >80% identity). If relatively few (<10) alignments were found, the parameters were relaxed (identity >70%). Conserved sequences within 1 kb of each other were considered part of the same CNS and were grouped as such. Intergenic CNSs found within the proximal 500 nt (core promoter region) or the distal 1 kb (core elements for the upstream gene) as well as exon 1 CNSs were excluded from further analysis. Details for each gene are provided in a supplementary table; all genomic sequences and CNSs are available online at ftp://ftp.ncbi.nih.gov/pub/kondrashov/ENDOTHELIUM/.

Hypersensitivity analysis
Genomic DNA from HUVECs (Cambrex) and human foreskin fibroblasts (HFF-2, ATCC) were prepared as previously described (19). Briefly, cells were grown in culture to confluence, harvested by trypsinization, washed and nuclei collected by 0.1% NP40 treatment and subsequent centrifugation. Genomic DNA was digested with increasing amounts of DNaseI (Promega) and visualized on a 0.8% agarose gel. For quantitative PCR, digested and undigested DNA samples were diluted to 1.7 ng/µl and 5 µl were plated onto 384-well clear optical reaction plates (Applied Biosystems). Primers were designed to amplify a 200–300 bp region surrounding or contained by each of the 262 EC-CNSs described earlier. Quantitative PCR was performed using a SYBR-green QPCR master mix (Qiagen) and an ABI 7900 real-time PCR machine (UMCCC microarray core).

Quantitative PCR was utilized as a high-throughput method to test these sites. Amplification of a target across a hypersensitive site will require additional cycles to reach a particular amount of product than a non-HS site. The difference in the number of cycles to reach a threshold product level between a digested sample and an undigested control is called the {Delta}Ct value. Depending on the experiment, between 20 and 73 randomly picked regions of the genome were amplified to determine the background level of DNase digestion. Sites with a {Delta}Ct value greater than two standard deviations from the mean of the distribution of {Delta}Ct values from the random sites were considered to be hypersensitive.

Motif analysis
We used DME software, kindly provided by Andrew Smith, to search for over-represented motifs in our sets of conserved sequences (20). We generated sets of random sequence of similar length and nucleotide content to act as a background set.

Within DME, two primary user-defined parameters are motif width (w) and minimum information content (in bits per column, I). Through several trials, we determined that using w=7 and I=1.8 gave us a sufficiently large number of motif occurrences to discriminate them from background noise.


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


   ACKNOWLEDGEMENTS
 
We thank Andrew Smith (Cold Spring Harbor Laboratory) for providing the DME software, technical assistance and critical reading of this manuscript. This research is supported by grants from the National Institutes of Health, R37-HL39693 (DG), P01 HL057346 (DG) and 5 P30 CA46592 (University of Michigan Cancer Center). D.G. is a Howard Hughes Medical Institute Investigator.

Conflict of Interest statement. None declared.


   FOOTNOTES
 
{dagger} Current address: Institute for Genome Sciences & Policy and Department of Pediatrics, Duke University, Durham, NC 27708, USA. Back


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

  1. Petrascheck M., Escher D., Mahmoudi T., Verrijzer C.P., Schaffner W., Barberis A. (2005) DNA looping induced by a transcriptional enhancer in vivo. Nucleic Acids Res. 33:3743–3750.[Abstract/Free Full Text]

  2. Escher D., Bodmer-Glavas M., Barberis A., Schaffner W. (2000) Conservation of glutamine-rich transactivation function between yeast and humans. Mol. Cell. Biol. 20:2774–2782.[Abstract/Free Full Text]

  3. Yamamoto K., de Waard V., Fearns C., Loskutoff D.J. (1998) Tissue distribution and regulation of murine von Willebrand factor gene expression in vivo. Blood 92:2791–2801.[Abstract/Free Full Text]

  4. Jaffe E.A., Hoyer L.W., Nachman R.L. (1973) Synthesis of antihemophilic factor antigen by cultured human endothelial cells. J. Clin. Invest. 52:2757–2764.[Web of Science][Medline]

  5. Nachman R., Levine R., Jaffe E.A. (1977) Synthesis of factor VIII antigen by cultured guinea pig megakaryocytes. J. Clin. Invest. 60:914–921.[Web of Science][Medline]

  6. Guan J., Guillot P.V., Aird W.C. (1999) Characterization of the mouse von Willebrand factor promoter. Blood 94:3405–3412.[Abstract/Free Full Text]

  7. Kent W.J., Sugnet C.W., Furey T.S., Roskin K.M., Pringle T.H., Zahler A.M., Haussler D. (2002) The human genome browser at UCSC. Genome Res. 12:996–1006.[Abstract/Free Full Text]

  8. Waterston R.H., Lindblad-Toh K., Birney E., Rogers J., Abril J.F., Agarwal P., Agarwala R., Ainscough R., Alexandersson M., An P., et al. (2002) Initial sequencing and comparative analysis of the mouse genome. Nature 420:520–562.[CrossRef][Medline]

  9. Shabalina S.A., Ogurtsov A.Y., Kondrashov V.A., Kondrashov A.S. (2001) Selective constraint in intergenic regions of human and mouse genomes. Trends Genet. 17:373–376.[CrossRef][Web of Science][Medline]

  10. Keightley P.D., Kryukov G.V., Sunyaev S., Halligan D.L., Gaffney D.J. (2005) Evolutionary constraints in conserved nongenic sequences of mammals. Genome Res. 15:1373–1378.[Abstract/Free Full Text]

  11. Kryukov G.V., Schmidt S., Sunyaev S. (2005) Small fitness effect of mutations in highly conserved non-coding regions. Hum. Mol. Genet. 14:2221–2229.[Abstract/Free Full Text]

  12. Dermitzakis E.T., Reymond A., Scamuffa N., Ucla C., Kirkness E., Rossier C., Antonarakis S.E. (2003) Evolutionary discrimination of mammalian conserved non-genic sequences (CNGs). Science 302:1033–1035.[Abstract/Free Full Text]

  13. Kondrashov A.S. and Shabalina S.A. (2002) Classification of common conserved sequences in mammalian intergenic regions. Hum. Mol. Genet. 11:669–674.[Abstract/Free Full Text]

  14. Li Q., Harju S., Peterson K.R. (1999) Locus control regions: coming of age at a decade plus. Trends Genet. 15:403–408.[CrossRef][Web of Science][Medline]

  15. Li Q., Peterson K.R., Fang X., Stamatoyannopoulos G. (2002) Locus control regions. Blood 100:3077–3086.[Abstract/Free Full Text]

  16. Wu C., Wong Y.C., Elgin S.C. (1979) The chromatin structure of specific genes: II. Disruption of chromatin structure during gene activity. Cell 16:807–814.[CrossRef][Web of Science][Medline]

  17. Wu C. (1980) The 5' ends of Drosophila heat shock genes in chromatin are hypersensitive to DNase I. Nature 286:854–860.[CrossRef][Medline]

  18. McArthur M., Gerum S., Stamatoyannopoulos G. (2001) Quantification of DNaseI-sensitivity by real-time PCR: quantitative analysis of DNaseI-hypersensitivity of the mouse beta-globin LCR. J. Mol. Biol. 313:27–34.[CrossRef][Web of Science][Medline]

  19. Crawford G.E., Holt I.E., Mullikin J.C., Tai D., Blakesley R., Bouffard G., Young A., Masiello C., Green E.D., Wolfsberg T.G., et al. (2004) Identifying gene regulatory elements by genome-wide recovery of DNase hypersensitive sites. Proc. Natl Acad. Sci. USA 101:992–997.[Abstract/Free Full Text]

  20. Smith A.D., Sumazin P., Zhang M.Q. (2005) Identifying tissue-selective transcription factor binding sites in vertebrate promoters. Proc. Natl Acad. Sci. USA 102:1560–1565.[Abstract/Free Full Text]

  21. Nelson C.E., Hersh B.M., Carroll S.B. (2004) The regulatory content of intergenic DNA shapes genome architecture. Genome Biol. 5:R25.[CrossRef][Medline]

  22. Vinogradov A.E. (2004) Compactness of human housekeeping genes: selection for economy or genomic design? Trends Genet. 20:248–253.[CrossRef][Web of Science][Medline]

  23. Schwachtgen J.L., Janel N., Barek L., Duterque-Coquillaud M., Ghysdael J., Meyer D., Kerbiriou-Nabias D. (1997) Ets transcription factors bind and transactivate the core promoter of the von Willebrand factor gene. Oncogene 15:3091–3102.[CrossRef][Medline]

  24. Lelievre E., Lionneton F., Soncin F., Vandenbunder B. (2001) The Ets family contains transcriptional activators and repressors involved in angiogenesis. Int. J. Biochem. Cell Biol. 33:391–407.[CrossRef][Web of Science][Medline]

  25. De Val S., Anderson J.P., Heidt A.B., Khiem D., Xu S.M., Black B.L. (2004) Mef2c is activated directly by Ets transcription factors through an evolutionarily conserved endothelial cell-specific enhancer. Dev. Biol. 275:424–434.[CrossRef][Medline]

  26. Ho M., Yang E., Matcuk G., Deng D., Sampas N., Tsalenko A., Tabibiazar R., Zhang Y., Chen M., Talbi S., et al. (2003) Identification of endothelial cell genes by combined database mining and microarray analysis. Physiol. Genomics 13:249–262.[Abstract/Free Full Text]

  27. Su A.I., Wiltshire T., Batalov S., Lapp H., Ching K.A., Block D., Zhang J., Soden R., Hayakawa M., Kreiman G., et al. (2004) A gene atlas of the mouse and human protein-encoding transcriptomes. Proc. Natl Acad. Sci. USA 101:6062–6067.[Abstract/Free Full Text]

  28. Lee P.D., Sladek R., Greenwood C.M., Hudson T.J. (2002) Control genes and variability: absence of ubiquitous reference transcripts in diverse mammalian expression studies. Genome Res. 12:292–297.[Abstract/Free Full Text]

  29. Vandesompele J., De Preter K., Pattyn F., Poppe B., Van Roy N., De Paepe A., Speleman F. (2002) Accurate normalization of real-time quantitative RT-PCR data by geometric averaging of multiple internal control genes. Genome Biol. 3:RESEARCH0034.[Medline]

  30. de Kok J.B., Roelofs R.W., Giesendorf B.A., Pennings J.L., Waas E.T., Feuth T., Swinkels D.W., Span P.N. (2005) Normalization of gene expression measurements in tumor tissues: comparison of 13 endogenous control genes. Lab. Invest. 85:154–159.[Web of Science][Medline]

  31. Smit A.F.A., Hubley R., Green P. RepeatMasker. http://www.repeatmasker.org.

  32. Ogurtsov A.Y., Roytberg M.A., Shabalina S.A., Kondrashov A.S. (2002) OWEN: aligning long collinear regions of genomes. Bioinformatics 18:1703–1704.[Abstract/Free Full Text]


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
 Mol. Cell. Biol.Home page
A. M. Pilon, M. O. Arcasoy, H. K. Dressman, S. E. Vayda, Y. D. Maksimova, J. I. Sangerman, P. G. Gallagher, and D. M. Bodine
Failure of Terminal Erythroid Differentiation in EKLF-Deficient Mice Is Associated with Cell Cycle Perturbation and Reduced Expression of E2F2
Mol. Cell. Biol., December 15, 2008; 28(24): 7394 - 7401.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow Supplementary Material
Right arrow All Versions of this Article:
15/13/2098    most recent
ddl133v1
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 (6)
Right arrowRequest Permissions
Google Scholar
Right arrow Articles by Bernat, J. A.
Right arrow Articles by Kondrashov, A. S.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Bernat, J. A.
Right arrow Articles by Kondrashov, A. S.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?