Nuclear scaffold attachment sites within ENCODE regions associate with actively transcribed genes

Abstract

The human genome must be packaged and organized in a functional manner for the regulation of DNA replication and transcription. The nuclear scaffold/matrix, consisting of structural and functional nuclear proteins, remains after extraction of nuclei and anchors loops of DNA. In the search for cis-elements functioning as chromatin domain boundaries, we identified 453 nuclear scaffold attachment sites purified by lithium-3,5-iodosalicylate extraction of HeLa nuclei across 30 Mb of the human genome studied by the ENCODE pilot project. The scaffold attachment sites mapped predominately near expressed genes and localized near transcription start sites and the ends of genes but not to boundary elements. In addition, these regions were enriched for RNA polymerase II and transcription factor binding sites and were located in early replicating regions of the genome. We believe these sites correspond to genome-interactions mediated by transcription factors and transcriptional machinery immobilized on a nuclear substructure.

Figure 1. Isolation and identification of HeLa scaffold attachment regions.

A.) Overview of nuclear scaffold isolation procedure. B.) Recovered DNA from Loop (Lane 1) and Scaffold (Lane 2) fractions after electrophoresis on a 1.5% agarose gel. MW represents molecular weight markers: Lambda DNA digested with StyI and 100 bp DNA ladder (NEB). C.) Quantitative PCR results for the ApoB 3′MAR and ApoB negative region. Mean ± S.D of three measurements. D.) Results of microarray analysis of scaffold DNA for ENCODE region ENm001. The MAT score profile and regions identified as SARs are shown in relationship to UCSC annotated genes.

Figure 2. Validation of microarray results and SAR characteristics.

A.) Summary of qPCR validation for 48 identified SARs and 16 negative regions chosen from the microarray data. The average enrichment in the scaffold fraction relative to total genomic DNA was normalized to the ApoB negative control by calculating Z scores (+ indicates ApoB 3′MAR positive control, − indicates ApoB negative control). 96% of the sites tested validated with a Z score corresponding to ≥8 standard deviations away from the negative control. B.) Histogram of SAR AT content. C.) Histogram of SAR size plotted in 2 kb bins. D.) Histogram of inter-SAR distance plotted in 25 kb bins.

Figure 3. SARs preferentially associate with expressed genes.

A.) Pie charts showing the distribution of bp that correspond to genic (introns, exons) and intergenic (5 kb upstream, intergenic) regions of the genome across all ENCODE regions and identified SARs. B.) Percentage of SARs that directly overlap or lie within 5 kb of a gene or an expressed gene and the corresponding enrichment as compared to a random model. C.) For each SAR, the distance to the nearest transcriptional start site (TSS) or transcriptional end site (TES) within the ENCODE regions was determined and plotted as a histogram. D.) Percent enrichment of SARs, as compared to a randomized data set, that directly overlap a TSS or TES of expressed genes or a RNA Pol II binding site.

Figure 4. SARs are enriched in euchromatic regions and overlap functional elements.

Percent enrichment (or depletion) of SARs in (A) early, mid, or late replicating regions of the genome, and (B) acetylated histones H3 and H4, trimethylation of histone H3K27, or Lamin B1 associated domains in comparison to a randomized data set. Enrichment of SARs that overlap (C) DNAse I hypersensitive sites (DHS) and formaldehyde-assisted isolation of regulator elements (FAIRE) as well as (D) CTCF insulator binding sites, regulatory factor binding sites (RFBR), conserved sequence element (CE), and origins of replication (ORIs), as compared to a randomized data set.

Figure 5. SARs may represent a variety of transcriptionally-mediated genomic interactions.

UCSC Genome Browser images of representative loci containing SARs that may correspond to A.) a transcription factory, B.) gene looping, C.) an active chromatin hub, and D.) transcription factor-mediated cis-element interaction.