1. Functional Non-Coding DNA Part II DNA Regulatory Elements BNFO 602/691 Biological Sequence Analysis Mark Reimers, VIPBG

2. Epigenetic Marks (as per current use) Any stable modification of DNA or chromatin DNA modifications – almost all methylation on C, mostly at ‘CpG’ – Some other methylation derivatives Histone modifications by adding groups – Majority on histone H3 – Mostly methyl, acetyl, phosphate groups – Sometimes ubiquitin!

3. What Epigenetics Originally Meant Heritable characteristics that did not involve DNA mutations Most mechanisms involved chromatin state Now means mitotically or meiotically heritable traits without change to DNA sequence, mediated by chromatin state Barbara McClintock’s ‘jumping genes’ were one of the early epigenetic phenomena

4. Nucleosomes and Histone Marks DNA wrapped around nucleosomes (8 histones) – ~142 bases wrapped twice Nucleosomes may be consistent (‘well positioned’) from cell to cell at key regulatory sites

5. Vertebrate Gene Promoters Regulatory region just before TSS Most coding genes have more than one promoter – median 3; some have > 20 Each promoter typically has binding sites for several general transcription factors and several gene-specific TFs – Many binding sites are repressive, e.g. NRSF

6. Three Common Promoter Types From Lenhard et al, Nat Rev Gen, 2012

7. Epigenetic Marks at Promoters H3K4me3 typically marks active promoters (and a few other places) H3K27ac typically marks active regulatory sites including promoters H3K27me3 marks repressed promoters – Typically developmental genes DNA methylation dips at silenced promoters One histone is kicked out at active promoters

8. Typical Promoter Marks: EHD3 in Cortex TSS

9. Six Highly Expressed Genes

10. Six Genes Expressed Little

11. CpG Islands and Promoters Although C & G constitute 42% of the human genome, less than 1% of pairs are CpG – Less than ¼ of expected frequency of 0.04 A ‘CpG island’ is a run of “CpG-rich” sequence – Typically taken as > 2% of pairs – This definition is not precise Many CpG islands occur within promoters

12. DEMO: Display BDNF in UCSC Browser

13. Enhancers and Silencers Many sites outside the promoter regulate gene expression One (or more commonly several) DNA-binding proteins attach at these loci Enhancers typically regulate a nearby gene but more commonly NOT the closest gene

14. Enhancers and Silencers It is thought that proteins bound to the enhancer make contact with the promoter Recent studies suggest these contacts are set up in embryonic cells From Michael Botchan labFrom Nature SciTable

15. More Evidence for Looping – 5C Data DNA is arranged in domains in eukaryotic cells Many co-transcribed genes are in ‘transcription factories’ Chromosome Conformation Capture Carbon Copy (5C) technology maps regions of the genome that are in close physical proximity in the cell

16. More Evidence for Looping – 5C Data Those enhancers that are highly correlated with a gene’s promoter often lie in fragments that are found joined in 5C data From Thurman et al, Nature, 2012

17. Perspective on Histone Marks Dozens of histone modifications are known, but usually their meanings are not yet clear Acetylation masks charge; methylation modifies binding of other proteins H3K4me3 marks active promoters H3K4me1 marks inactive enhancers H3K9ac marks transcribed promoters H3K27ac marks active enhancers H3K27me3 marks repressed promoters

18. Genomic Insulators Insulators are regulatory elements that establish independent domains of transcriptional activity within eukaryotic genomes. Insulators both block enhancer-promoter communication, and prevent the spread of repressive chromatin

19. Insulators Many insulators in vertebrates are bound by CTCF (CCCTC-binding factor) an 11-zinc-finger – But that is NOT the only role of CTCF CTCF was originally found to bind to three regularly spaced repeats of CCCTC – Now known to bind many similar sequences using different subsets of fingers CTCF seems to regulate 3D structure of DNA

20. CCCTC-Binding Factor (CTCF) Multi-functional DNA-binding protein binds together strands of DNA, forming chromatin loops, defining 3D structure of chromatin Anchors DNA to the nuclear lamina Insulates euchromatin from heterochromatin Highly conserved 11-zinc finger protein Binding sites heterogeneous – different fingers A common CTCF motif Springer Images

21. Open Chromatin at Regulatory Sites Almost all DNA-binding proteins bind to unwrapped DNA Either they must force open the DNA and displace histones or wait for another TF to do Open chromatin is not an epigenetic ‘mark’ but is a useful indicator of functional DNA Open chromatin can be assayed by DNase I sensitivity – where does DNase I cut?

22. Genes, regulatory DNA, and epigenetic features Graphic from NIH RoadMap Epigenomics Site

23. Genes, regulatory DNA, and epigenetic features - promoters - enhancers - silencers - insulators - etc.

24. Genes, regulatory DNA, and epigenetic features - promoters - enhancers - silencers - insulators - etc. DNaseI

25. ~100,000 – 250,000 DHSs per cell type (0.5-1.5% of genome) genome.ucsc.edu www.epigenomebrowser.org DNaseI Hypersensitive site (DHS) Promoters Enhancers DNaseI hypersensitive sites mark regulatory DNA Courtesy John Stamatoyannopoulos

26. Classifying the Genome Can combine histone marks, DNAse and CTCF to classify function of regions of genome Hidden Markov Model commonly used From Bernstein et al, Nature, 2012

27. DNA Methylation A further epigenetic modification directly on DNA is cytosine methylation Recently hydroxy-methylation shown to be important

28. DNA Methylation and Regulation Cytosine methylation blocks DNA- binding proteins’ access to regulatory sites and creates binding sites for repressive proteins Methylation often follows decrease in site use From Thurman et al Nature 2012

29. Methylation and Expression Some genes (e.g. E-cadherin in cancer cell lines) show strong correlation of average promoter methylation with expression

30. Methylation, Retroviruses and Repeats Bacteria use DNA methylation to limit invasive DNA from viruses A large fraction of the human genome consists of carcasses of retro-viruses and transposons Almost all DNA repeats are heavily methylated If they lose methylation they are more likely to be expressed

31. DNA Methylation and Development Almost all DNA de-methylated in embryo Increasing methylation at various times during fetal development restrict functionality – This is why cloning is difficult Wave of methylation in adolescence Gradual de-methylation in old age