Protein occupancy landscape of a bacterial genome

Abstract

Protein-DNA interactions are fundamental to core biological processes, including transcription, DNA replication, and chromosomal organization. We have developed in vivo protein occupancy display (IPOD), a technology that reveals protein occupancy across an entire bacterial chromosome at the resolution of individual binding sites. Application to Escherichia coli reveals thousands of protein occupancy peaks, highly enriched within and in close proximity to noncoding regulatory regions. In addition, we discovered extensive (>1 kilobase) protein occupancy domains (EPODs), some of which are localized to highly expressed genes, enriched in RNA-polymerase occupancy. However, the majority are localized to transcriptionally silent loci dominated by conserved hypothetical ORFs. These regions are highly enriched in both predicted and experimentally determined binding sites of nucleoid proteins and exhibit extreme biophysical characteristics such as high intrinsic curvature. Our observations implicate these transcriptionally silent EPODs as the elusive organizing centers, long proposed to topologically isolate chromosomal domains.

Figure 1. In vivo protein occupancy display (IPOD)

(A) Schematic for isolation and genome-wide display of protein-bound sites across a bacterial genome. Formaldehyde cross-linking preserves in vivo protein-DNA interactions. Following cell-lysis and sonication, protein footprints are minimized through DNase I treatment. Phenol extraction enriches for protein-DNA complexes at the interface between the aqueous and organic phases. Following interface isolation, cross-links are reversed, the resulting DNA fragments are end-labeled and hybridized to a tiling array. (B) Gel-fractionation shows that DNase I treatment leads to a drop in the mode of fragment-length distribution from ~1000 bp (no DNase I) to ~200 bp (½X DNase I), to below 100 bp for (1X DNase I). The samples were separated on the same gel and extraneous lanes were removed for clarity. (C) Cumulative probability distribution of occupancy (z-score: standard deviations from the mean) for both coding and non-coding regions determined during late exponential phase growth. The z-score values were smoothed by averaging within a moving window of 128 base pairs.

Figure 2. Protein occupancy profile of the E. coli genome during late exponential phase growth

(A) At low spatial resolution, high-amplitude occupancy peaks are largely confined to intergenic (non-coding) regions of the genome (red arrows). However, similar peaks can less frequently be seen within coding regions as well (blue arrows). Two independent biological replicates show highly reproducible occupancy profiles across this region. (B) At high spatial resolution, multiple occupancy peaks are discernable within a single intergenic region. Peaks are localized to the typical footprint of individual transcription factors and often overlap experimentally determined binding sites (PurR, RegulonDB).

Figure 3. Extended protein occupancy domains (EPODs)

Protein occupancy and RNA expression profiles are shown for early exponential phase growth, smoothed by averaging within a moving window of 512 base pairs. At the bottom, open reading frames (ORFs) are annotated on both strands. Automatically detected transcriptionally-silent EPODs are shown at the top. (A) A high protein occupancy and high RNA expression domain encompassing a region with genes encoding ribosomal protein subunits. (B) An EPOD automatically detected within a transcriptionally-silent region with genes encoding LPS biosynthesis products. A neighboring region encoding 50S ribosomal protein subunits (rpmG and rpmB) shows an equal level of protein occupancy but shows a high level of RNA expression. (C,D) Transcriptionally-silent EPODs detected within genes of unknown function encoding a predicted protein, yccE (C) and a conserved inner membrane protein, yliE (D).

Figure 4. Distinct occupancy composition and biophysical properties within extended protein occupancy domains

The cumulative distribution of various measures are shown for transcriptionally-silent EPODs (red), highly-expressed EPODs (blue), and a matched background control (black). The Wilcoxon rank sum test is used to determine statistical significance of observed deviations relative to background. (A) Experimentally-determined relative RNA polymerase occupancy. (B,C) Computationally-scored PWM binding preference for a nucleoid protein (H-NS) and a non-nucleoid transcriptional repressor (LacI) using a genome-wide relative measure (z-score). (D) Cumulative distributions of A:T frequencies within EPODs. (E) Cumulative distribution of predicted relative curvature values within EPODs. (F) Distribution of binding sites for various nucleoid proteins and CRP within tsEPODs and heEPODs.