Biochemical activity is the default DNA state in eukaryotes

Abstract

Genomes encode for genes and the regulatory signals that enable those genes to be transcribed, and are continually shaped by evolution. Genomes, including those of human and yeast, encode for numerous regulatory elements and transcripts that have limited evidence of conservation or function. Here, we sought to create a genomic null hypothesis by quantifying the gene regulatory activity of evolutionarily naïve DNA, using RNA-seq of evolutionarily distant DNA expressed in yeast and computational predictions of random DNA activity in human cells and tissues. In yeast, we found that >99% of bases in naïve DNA expressed as part of one or more transcripts. Naïve transcripts are sometimes spliced, and are similar to evolved transcripts in length and expression distribution, indicating that stable expression and/or splicing are insufficient to indicate adaptation. However, naïve transcripts do not achieve the extreme high expression levels as achieved by evolved genes, and frequently overlap with antisense transcription, suggesting that selection has shaped the yeast transcriptome to achieve high expression and coherent gene structures. In humans, we found that, while random DNA is predicted to have minimal activity, dinucleotide content-matched randomized DNA is predicted to have much of the regulatory activity of evolved sequences, including active chromatin marks at between half (DNase I and H3K4me3) and 1/16th (H3K27ac and H3K4me1) the rate of evolved DNA, and the repression-associated H3K27me3 at about twice the rate of evolved DNA. Naïve human DNA is predicted to be more cell type-specific than evolved DNA and is predicted to generate co-occurring chromatin marks, indicating that these are not reliable indicators of selection. However, extreme high activity is rarely achieved by naïve DNA, consistent with these arising via selection. Our results indicate that evolving regulatory activity from naïve DNA is comparatively easy in both yeast and humans, and we expect to see many biochemically active and cell type-specific DNA sequences in the absence of selection. Such naïve biochemically active sequences have the potential to evolve a function or, if sufficiently detrimental, selection may act to repress them.

Exogenous chromosomes reveal how sequence composition drives chromatin assembly, activity, folding and compartmentalization

“…For instance, short stretches or small (13 kb) random DNA sequences can be used to generate yeast promoters and to decipher the regulatory logic of transcription activation 53–55 . Here we show that both normal GC% and AT-rich bacterial chromosomes are spontaneously transcribed when introduced in the yeast eukaryotic environment 6,13,56 . Active transcription follows, on average, the orientation of the bacteria genes ( Figure 3c ).…”

“…Here we show that both normal GC% and AT-rich bacterial chromosomes are spontaneously transcribed when introduced in the yeast eukaryotic environment 6,13,56 . Active transcription follows, on average, the orientation of the bacteria genes (Figure 3c).…”

Hold out the genome: A roadmap to solving the cis-regulatory code

“…While major efforts aim to make this more routine 148 , it is presently far too laborious and costly to produce the needed training data at a reasonable scale. Long random DNA sequences are predicted to be active in human cells 115 , suggesting that this strategy may work here as well. As before, specific elements could be spiked in to produce data more conducive to model training (e.g.…”

“…Thus, most regulatory sequences appear to have evolved from previously non-regulatory DNA [119][120][121][122] , demonstrating that the syntax by which TFs act must be relatively permissive such that functional elements can emerge by chance. Indeed, the proportion of random sequences with regulatory activity is similar to genomic sequences 25,31,115 . This rather counterintuitive finding makes sense from an evolutionary perspective since there is no advantage to increasing the specificity of a cis-regulatory element (let alone all elements) beyond what is required for its function, and the small marginal advantages to further specificity beyond this (if they exist at all) can be dominated by the effects of genetic drift 123,124 .…”

“…While learning cis-regulation from measurements of random DNA activity may seem counter-intuitive, the reason why this works is simple: there are many TFs (~1,639 in humans 2 ), and their binding sites occur frequently by chance 25,111,112 , enabling one to select for binding sites for individual TFs or pairs of TFs from random DNA 37,113,114 . While a random sequence having a binding site for a specific TF is unlikely, it is almost certain to have binding sites for some TF, leading to high expression diversity among random sequences 25 , and distinct expression programs across cell types 111,115 . This propensity for random sequences to frequently be active is also consistent with the soft regulatory syntax of the billboard model 23 .…”