Chromosome conformation of circadian genes Jérôme Mermet1, Kyle Gustafson1, Céline Jouffe2, Cédric Gobet2, Frédéric Gachon2, Jacques Rougemont1, and Felix Naef1 1: Ecole Polytechnique Fédérale de Lausanne, Lausanne, Switzerland; 2: Nestle Institute of Health Sciences, Lausanne, Switzerland
Since topological organiza on of chroma n plays an ABSTRACT important role in transcrip on regula on and the QUESTION, METHOD and DESIGN circadian oscillator provides a unique model for dynamic gene expression, we explored the topological organiza on of chroma n surrounding promoters of core clock and rhythmic output genes in mouse ssues of WT and BMAL1 KO animal during the circadian cycle using 4C-sequencing (4C-seq). We found chroma n interac on pa erns that are highly reproducible between biological replicates, conserved Cyclic gene Distal site Promoter across ssues and genotypes, and gene specific. Nevertheless, we Modi ied from Ong and Gorces, Nature Reviews Genetics, 2011 iden fied me-varying chroma n reorganiza on and DNA loops that depend on a func onal molecular clock and correlate with transcrip on. X4 X4 As a highlight, we iden fied a robust, conserved across ssues and WT ZT08 WT ZT20 rhythmic interac on between the Cry1 promoter and its intronic 4C-seq work low, from Van de Werken et al, Nat. Methods 2012 enhancer that depends on a func onal molecular clock, sugges ng that To answer if promoter of circadian genes interact with other genomic element (genes, regula on of chroma n topology by the circadian clock is a regulatory enhancers…etc), we used 4C-seq assays that allows to reveal interac ons between a single restric on fragment called the “bait” and the en re genome. We performed 4C-seq in mouse layer for transcrip on control. Bmal1-/- ZT08 Bmal1-/- ZT20 liver and kidney of WT and BMAL1 deficient mice. Therefore, we could evaluate the dynamic, X3 X3 ssue specific and genotype specific genomic interac on profiles of our selected baits.
4C-seq REVEALS CLOCK DEPENDENT INTERACTIONS
CRY1
Cry1: time, genotype, tissue wt ZT08 WT08 GENOMIC INTERACTIONS ARE LOCALIZED NEAR BAITS
1500 wt ZT20 WT20 3 ko ZT08 ko ZT20 bait ZT08 wild−type proximal signal (2 Mbp) 1000 Dixon domains (cortex) 1 2.5 wt ZT08 intron enh. 1000 Distance of peaks of interac on, all baits, all condi ons kidney 0.8 500 right TAD 2 wt ZT20 0.6 bait TAD kidney left TAD 800 0 Rudan et al. domains (liver) 0.4 8.42 8.43 8.441.5 8.45 8.46 8.47 8.48 8.49 8.5 8.51 8.52 signal fraction
CRY1 600 7 0.2 WT08KO08x 10 1 0
15001500 KO20 3000 400 WT20 from (kb) bait distance 0.5 bait 2500 log10(4C signal model)
10001000 200 intron enh. 2000 0 intron enh. 500 1500
500 0 8.36 8.38 8.4 8.42 8.44 8.46 8.48 8.5 8.52 8.54 8.56 1000 chr10 7 x 10 500 Liver WT ZT08 Liver WT ZT20 Liver KO ZT08 Liver KO ZT20 Kidney WT ZT08 Kidney WT ZT20 00 Rudan (liver) TAD size (kb) 8.428.42 8.438.43 8.448.44 8.458.45 8.468.46 8.478.47 8.48 8.49 8.5 8.51 8.52 0 modeled pseudocounts CRY1 CRY1CRY1 CRY1 Ccne1 Cry1 DbpL DbpR Hoxd4 Nr1d1 Per1 Per2 Gys2 Hmgcr Lipg Mfsd2a Nampt Pfkfb3 Por Slc2a2 7
WT08 6.5 WT08WT08KO08WT08 kidneyx 10 WT08 Most of the 4C-seq signal is localized in the bait Topologically Associated Domain (TAD), and most of the interac on 1500 150015001500 1500 WT20 CRY1 intron WT20WT20KO20WT20 kidney WT20
6.0 peaks are within few kb to 500kb window around the bait bait baitbaitbait bait 1000 100010001000 1000
intron enh. 5.5 intronintronintron enh. enh.enh. intron enh. 500 500500500 500 5.0 4C-seq PATTERNS ARE CONSERVED ACROSS CIRCADIAN TIME POINTS, 0 000 0 Z-score Z-score WT 8.42 8.43 8.44 8.45 8.468.428.428.42 8.478.438.438.43 8.488.448.448.44 8.498.458.458.458.42 8.468.468.468.58.43 8.518.478.478.478.44 8.528.488.488.488.45 8.498.498.464.5 8.58.478.5 8.518.518.48 8.528.528.49 8.5 8.51 8.52 modeled pseudocounts chr10 GENOTYPES AND TISSUES. 7 77 7 KO08x 10 WT08KO08KO08 kidneyxx 1010 KO08x 10 4.0 1500 15001500 1500 KO20 WT20KO20KO20 kidney KO20 Nr1d1: time, genotype, tissue wt ZT08 bait 3.5 baitbaitbait bait wt ZT20 1000 10001000 1000 ko ZT08 intron enh. intronintronintron enh. enh.enh. intron enh. 3.0 3 ko ZT20 500 500500 500 0 200 400 600 800 1000 Dixon domains (cortex) wt ZT08 kidney 0 00 0 Distance from bait (kb) 8.42 8.43 8.44 8.45 8.468.42 8.478.438.43 8.488.448.44 8.498.458.458.42 8.468.468.438.5 8.478.518.478.44 8.488.528.488.45 8.498.498.46 8.478.58.5 8.518.518.48 8.528.528.49 8.5 8.51 8.52 wt ZT20 modeled pseudocounts modeled pseudocounts modeled pseudocounts modeled pseudocounts 8.42 8.43 8.44 8.45 8.46 8.47 8.48 8.49 8.5 8.51 8.52 modeled pseudocounts chr10 7 77 7 5.0 kidney WT08 kidneyx 10 WT08WT08 kidneykidneyxx 1010 2 WT08 kidneyx 10 1500 15001500 1500 WT20 kidney WT20WT20 kidneykidney Rudan et al.WT20 domains kidney (liver) bait baitbait bait
1000 10001000 1000 4.5 intron enh. intronintron enh.enh. 1 intron enh. 500 500500 500
0 00 0 4.0 log10(4C signal model)
8.42 8.43 8.44 8.45 8.468.428.42 8.478.438.43 8.488.448.44 8.498.458.458.42 8.468.468.438.5 8.518.478.478.44 8.528.488.488.45 8.498.498.46 Z-score KO 8.478.58.5 8.518.518.48 8.528.528.49 8.5 8.51 8.52 chr10 chr10chr10 7 chr10 77 7 x 10 xx 10 10 0 x 10 CRY1 intron is an enhancer: 9.76 9.78 9.8 9.82 9.84 9.86 9.88 9.9 9.92 9.94 9.96 • contains RRE rhythmically bound by REV-ERBα (-) (ZT10) and RORg (ZT22) 3.5 chr11 7 • provides phase delay of CRY1 expression transac va on effect on CRY1 x 10 reporter Ukai-Tadenuma et al, Cell 2011 GENOMIC INTERACTIONS ARE ENRICHED FOR ACTIVE CHROMATIN MARKS. Zang et al, Science 2015 0 50 100 150 200 Distance from bait (kb)
Nampt liverCry1 liver wt ZT08 3 wt ZT20 3 ko ZT08 Dixon domains (cortex) ko ZT20 2.5 2.5
2 2
1.5 Rudan et1.5 al. domains (liver) 1 1 0.5 0.5 log10(4C signal model) log10(4C signal model) 0 Cry1 promoter rhythmically interacts with its intronic 3.24 3.26 3.28 3.3 3.32 3.34 3.36 3.38 3.4 3.42 3.44 0 chr12 7 enhancer in mouse liver and kidney of WT animals. This 8.36 8.38 8.4 8.42 8.44 8.46 8.48 8.5 8.52 x8.54 10 8.56 interac on is maximum at ZT20 and minimum at ZT8, chr10 7 corresponding respec vely to the maximum and x 10 minimum of Cry1 transcrip on. This interac on is strong and constant in Bmal1-/- animals, while Cry1 is highly and constantly expressed. Thus, this par cular example highlights a rhythmic and clock dependent chroma n loop that correlates with gene transcrip on.
Our study of chroma n structure in the context CONCLUSIONS of the circadian clock for mul ple ssues in mouse reveals the general stability of contact profiles within 1 - 2 Mb of 4C baits. This observa on is consistent with the typical size of contact domains, and in most cases, precisely consistent with domain boundaries measured recently in mouse liver. Nevertheless, we find some excep ons to the general conserva on of contacts, notably at a known regulatory site (ROR/REVERB-response element) in the first intron of Cry1, where we observed a rhythmic and clock dependent chroma n 4C-seq profiles for Nr1d1 (top panel) and Nampt (bo om panel) genes show a robust conserva on of genomic loop. Finally, we observed the tendency for contacts to associate with oscilla ng interac ons pa ern between circadian me point, genotype (Nr1d1 & Nampt), and ssues (Nr1d1), independently of their transcrip on status. Moreover, we observed that most of the significant peaks are contained in the bait eRNAs and ac ve chroma n marks, indica ng that circadian gene regula on may TAD. Finally, we observed a strong enrichment of ac ve chroma n marks under 4C-seq peaks such as DNase1 occur due to the me-dependent recruitment of transcrip on factors on a stable Hypersensi ve sites, enhancer RNAs, enhancer associated histone modifica on (H3K4me1, H3K27ac). scaffold of interconnected chroma n.