Deletion of SWI/SNF ATPase activity in hepatocytes of adult liver disrupts circadian gene expression and central liver functions [ChIPseq]

The circadian rhythm in the murine liver governs the activity of numerous enhancers which in turn coordinates diurnal gene expression. This process is controlled by oscillating activities of specific transcription factors (TFs) and recruitment of co-regulators, including histone modifying enzymes and chromatin remodeling complexes. Several circadian controlled TFs interact with the SWI/SNF family of chromatin remodeling complexes to control chromatin accessibility. To unravel the significance of SWI/SNF ATPase subunits in circadian chromatin remodeling, we mapped chromatin accessibility, SWI/SNF occupancy, and gene expression throughout a day in murine liver. We found remarkable remodeling during fasting-to-fed transitions, and a third of circadian enhancers exhibited circadian SWI/SNF occupancy and accessibility. Intriguingly, genetic disruption of either of the two mutually exclusive ATPases of SWI/SNF had minor effects on chromatin accessibility in circadian enhancers, indicating redundancy. However, simultaneous disruption of both ATPases caused a collapse of the chromatin landscape, liver damage and inflammation. This disruption abolishes rhythmic expression of metabolic genes without affecting oscillation of the core circadian clock. In summary, this suggests an indispensable role of SWI/SNF-mediated chromatin remodeling of enhancers for circadian transcriptomic rhythms and basic liver function. Overall design: To investigate the role of SWI/SNF chromatin remodeling complexes in circadian chromatin remodeling we mapped chromatin accessibility (ATAC-seq), SWI/SNF occupancy (Brg1 ChIP-seq), and gene expression (RNA-seq) throughout a day in murine liver. All mice were housed at ~23 °C and 12-hour light/dark cycle. Circadian experiments were performed in mice trained to night restricted feeding (NRF) for 4 days. Mice were sacrificed by cervical dislocation at different timepoints during the day (ZT 2, 6, 10, 14, 18, and 22). Wild type mice were C57BL/6N background and Smarca2-/- 44 and Smarca4fl/fl mice in a 129/Sv background were backcrossed with C57BL/6N mice and subsequently intercrossed. At age 8-10 weeks, the resulting Smarca2-/- mice, Smarca4fl/fl mice, and Smarca2-/-;smarca4fl/fl mice were tail vein injected AAV-TBG-GFP or AAV-TBG-CRE. Ten days after injection the mice were trained to NRF for another 4 days. Three mice per time point were used for RNA-seq and ATAC-seq of WT mice. Two mice per time point for Brg1 ChIP-seq in WT mice. Four mice were used for RNA-seq and ATAC-seq at ZT10 of mice with the different genotypes.