5720–5738 Nucleic Acids Research, 2017, Vol. 45, No. 10 Published online 8 March 2017 doi: 10.1093/nar/gkx156 A class of circadian long non-coding RNAs mark enhancers modulating long-range circadian gene regulation Zenghua Fan1,2, Meng Zhao3, Parth D. Joshi4,PingLi5, Yan Zhang5, Weimin Guo3, Yichi Xu1,2, Haifang Wang3, Zhihu Zhao5 and Jun Yan3,* 1 CAS-MPG Partner Institute for Computational Biology, Shanghai Institutes for Biological Sciences, Chinese Downloaded from https://academic.oup.com/nar/article-abstract/45/10/5720/3063381 by guest on 06 March 2019 Academy of Sciences, 320 Yue Yang Road, Shanghai 200031, China, 2University of Chinese Academy of Sciences, Shanghai 200031, China, 3Institute of Neuroscience, State Key Laboratory of Neuroscience, CAS Center for Excellence in Brain Science and Intelligence Technology, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences, Shanghai 200031, China, 4Department of Genes and Behavior, Max Planck Institute for Biophysical Chemistry, Am Fassberg 11, 37077 Gottingen,¨ Germany and 5Beijing Institute of Biotechnology, 20 Dongdajie Street, Fengtai District, Beijing 100071, China Received February 04, 2017; Editorial Decision February 23, 2017; Accepted February 24, 2017 ABSTRACT regulation and shed new lights on the evolutionary origin of lncRNAs. Circadian rhythm exerts its influence on animal physiology and behavior by regulating gene expres- sion at various levels. Here we systematically ex- INTRODUCTION plored circadian long non-coding RNAs (lncRNAs) Circadian rhythm is an intrinsic 24 h oscillation of var- in mouse liver and examined their circadian reg- ious physiological processes and behaviors synchronized ulation. We found that a significant proportion of with daily light/dark cycle in a wide-range of species. In circadian lncRNAs are expressed at enhancer re- mammals, suprachiasmatic nucleus (SCN) in the brain is gions, mostly bound by two key circadian transcrip- known as the control center of circadian rhythm that or- tion factors, BMAL1 and REV-ERB␣. These circadian chestrates rhythms of peripheral tissues. Circadian con- lncRNAs showed similar circadian phases with their trol of many aspects of physiology and behavior is man- nearby genes. The extent of their nuclear localiza- ifested by the widespread circadian regulation of gene tion is higher than protein coding genes but less expression. It is now known that the central molecular than enhancer RNAs. The association between en- clock is formed by two main feedback loops consisting hancer and circadian lncRNAs is also observed in of core clock genes (1–4). In the first loop, BMAL1 and tissues other than liver. Comparative analysis be- CLOCK form heterodimers to initiate the expression of Pe- riod (Per1/2/3)andCryptochrome (Cry1/2) genes that in tween mouse and rat circadian liver transcriptomes return repress BMAL1/CLOCK activities. In the second showed that circadian transcription at lncRNA loci loop, BMAL1 and CLOCK activate the expression of Rev- tends to be conserved despite of low sequence con- erbα/β (Nr1d1/2) that in return repress the expression of servation of lncRNAs. One such circadian lncRNA Bmal1 and Clock. The studies of genome-wide circadian termed lnc-Crot led us to identify a super-enhancer transcription have mainly focused on the protein coding region interacting with a cluster of genes involved genes (5–8) and small non-coding RNAs such as microR- in circadian regulation of metabolism through long- NAs (9,10). range interactions. Further experiments showed that In recent years, a new class of non-coding RNAs referred lnc-Crot locus has enhancer function independent to as long non-coding RNAs (lncRNAs) has gained much of lnc-Crot’s transcription. Our results suggest that attention. It was shown that lncRNAs are pervasively tran- the enhancer-associated circadian lncRNAs mark the scribed from mammalian genomes (11) and tens of thou- sands of lncRNAs have been detected in a variety of tis- genomic loci modulating long-range circadian gene sues. In contrast to the protein coding mRNAs, lncRNAs are considered to have lower expression level, stronger tis- *To whom correspondence should be addressed. Tel: +86 21 54920474; Fax: +86 21 54921735; Email: [email protected] Disclaimer: The funders had no role in study design, data collection and analysis, decision to publish or preparation of the manuscript. C The Author(s) 2017. Published by Oxford University Press on behalf of Nucleic Acids Research. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by-nc/4.0/), which permits non-commercial re-use, distribution, and reproduction in any medium, provided the original work is properly cited. For commercial re-use, please contact [email protected] Nucleic Acids Research, 2017, Vol. 45, No. 10 5721 sue specificity and lower sequence conservation (12–16)and circadian TFs’ regulation on nearby genes through long- have been found to play important roles in many biological range interaction. processes (17,18). Recently, lncRNAs showing circadian expression have MATERIALS AND METHODS been reported in several species. In Neurospora, qrf,a lncRNA residing on the anti-sense strand of a core clock Animal preparation and sample collection gene, frequency (frq), oscillates in anti-phase to frq and in- All mice were male, C57BL/6J strain and aged around 8 hibits the expression of frq through histone modification weeks. All animals were provided with food and water avail- and premature termination of transcription (19). In mouse, able ad libitum under 12/12 h light-dark (LD) conditions. a lncRNA anti-sense to the core circadian gene, Per2,has After the entrainment under light-dark condition for at also been found oscillating anti-phase to Per2 but its ex- least 7 days, animals were then transferred to constant dark- act function is still unknown (20). A recent study of mouse ness (DD) condition and sacrificed after 24 h of DD. Livers Downloaded from https://academic.oup.com/nar/article-abstract/45/10/5720/3063381 by guest on 06 March 2019 circadian transcriptome (21) suggested that at least 1000 from three mice at each time point every 4 h for 48 h were known lncRNAs showed circadian expression in multiple collected, quickly frozen in liquid nitrogen and then stored mouse tissues. However, the roles of these lncRNAs in circa- at −80◦C. Liver samples of two pairs of REV-ERB␣ knock- dian rhythm remain unclear. Also another class of short and out (KO) and wild-type (WT) mice were collected at CT0 non-spliced non-coding RNAs transcribed from enhancers, and CT12 respectively.All animal experiments performed in called enhancer RNAs (eRNAs), were found with circadian this study were approved by the Institutional Animal Care expression in mouse liver (22). However how these eRNAs and Use Committee of Shanghai Institutes for Biological function in circadian gene regulatory network was unclear Sciences and conformed to institutional guidelines of verte- and their relationship with lncRNAs is yet to be discovered brates study. (23). In this study, we compiled a comprehensive list of cir- cadian lncRNAs in mouse liver. Among them, we iden- RNA sequencing tified a class of circadian lncRNAs that are transcribed Total RNA was extracted from tissue samples by Tri- from enhancer loci in mouse liver and regulated by cir- zol (Invitrogen). Liver RNA samples of three liver-specific cadian transcription factors (TFs) including BMAL1 and BMAL1 KO mice and three WT control mice at CT0 and REV-ERB␣. These lncRNAs were found to show close CT12 respectively were provided by Prof. Yi Liu (Institute circadian phases and similar responses upon BMAL1 or for Nutritional Sciences, Shanghai Institutes for Biologi- REV-ERB␣ knockout in mouse liver with their nearby cir- cal Sciences). Each of the RNA samples was then used cadian protein coding genes. We hypothesized that these for strand-specific library construction and sequencing with lncRNA loci may harbor enhancers to form chromatin single-ended 100 bp reads and the WT samples were used loops to facilitate BMAL1 or REV-ERB␣ regulation on for lncRNA de novo assembly. The RNAs of two pairs of nearby genes through long-range interaction. Comparing REV-ERB␣ KO and control mouse livers were used for nascent-seq with RNA-seq data, circadian lncRNAs were un-stranded RNA sequencing. RNA sequencing was per- found more nucleus-enriched than circadian protein cod- formed using the Illumina HiSeq 2000. All sequencing raw ing mRNAs but less than previously defined circadian eR- data were submitted to Gene Expression Omnibus (GEO) NAs. Furthermore, strong association between enhancers with accession number GSE87299. and circadian lncRNAs were also discovered in non-hepatic mouse tissues. Our analysis of circadian lncRNAs from RNA-seq data in rat liver showed that circadian transcrip- qPCR analysis tion of lncRNA locus is more conserved than lncRNAs For quantitative polymerase chain reaction (qPCR) anal- themselves. This suggests that circadian lncRNAs may re- ysis, 500 ng of total RNA was reverse transcribed using sult from clock-controlled transcription at the enhancers, random hexamer and the Superscript II reverse transcrip- some of which are shared across tissues and even conserved tase (Invitrogen). A total of 4 ␮l of reverse transcribed between species. We found that a circadian lncRNA, termed (RT) product (1:10 diluted) and SYBR Green I Master Mix lnc-Crot, is expressed at a super-enhancer upstream of the (Roche) were used in qPCR on a LightCycler 480 (Roche). circadian oscillating gene Crot and is regulated by BMAL1 Expression of all genes and lncRNAs were normalized by and REV-ERB␣. With circular chromosome conformation control gene, Actb. Primers used were listed in Supplemen- capture sequencing (4C-seq), we showed that lnc-Crot lo- tary Table S7. cus interacts with the circadian oscillating genes that tend to oscillate with similar phases through long-range interac- Assembly of lncRNAs in mouse tion while these interactions appeared invariant through- out the day.