From OpenWetWare
Revision as of 12:47, 8 January 2013 by Michael E Hughes (talk | contribs)
Jump to: navigation, search


Home        Lab Members        Research        JTK_Cycle        MuscleDB        Publications        Protocols        Academic Analytics        Open Positions        Contact       

Discovery of circadian harmonics


We profiled global gene expression over two full days using Affymetrix microarrays. We identified rhythmic transcripts in the mouse liver and pituitary, as well as fibroblasts (NIH3T3) and osteosarcoma cells (U2OS). These data have been made freely available on CircaDB as a resource to the field. To our surprise, we found several hundred genes cycling with period lengths much shorter than 24 hours (Fig. 1). These ultradian rhythms had period lengths of ~8 and ~12 hours -- i.e., the second and third harmonics of 24 hour oscillations.

Subsequently, we have shown that these rhythms are found in tissues throughout the body. Moreover, they are found in fruit flies as well, suggesting that circadian harmonics are a common feature of animal transcriptional rhythms. At a mechanistic level, 12 hour rhythms require both a central and peripheral circadian oscillator, indicating that these rhythms are ultimately downstream of the conventional circadian clock. Typically, they are involved in cellular responses to stress, suggesting that ultradian transcriptional rhythms respond to twice daily stresses.

System-driven circadian oscillations

Fig. 2: Hughes et al. (2012b)
~100 genes oscillate in the mouse liver, despite ablation of the local, peripheral clock.

In collaboration with the Takahashi laboratory, we profiled transcriptional rhythms in mice with and without a defective circadian clock in the liver. Although most rhythmic genes are lost due to the ablation of the local circadian oscillator, nearly 100 genes continue to cycle with appropriate period lengths and phases (Fig. 2). Consequently, these persistent rhythms may form the molecular basis by which the central clock drives peripheral oscillations. Strikingly, many core clock genes continue to oscillate in the absence of a local clock, implying that the promoters of these genes have evolved to be directly responsive to circulating, systemic cues from the central nervous system.

Rhythms of snoRNA host genes

Fig. 3: Hughes et al. (2012a)
Non-coding, snoRNA host genes oscillate in the fly brain

We used Illumina sequencing to systematically profile RNA expression in fly heads over two consecutive days. We identified hundreds of cycling genes, including many non-coding RNAs (ncRNAs) that had not been identified in previous microarray studies. Most interestingly, an entire family of ncRNAs, the Uhg genes, oscillated over the course of 24 hours. These genes are hosts for snoRNA expression -- i.e., their exons are spliced together, and their introns are further processed to generate mature snoRNAs.

Since snoRNAs are involved in ribosomal maturation, this observation raises the possibility that these daily oscillations directly influence the assembly and stability of the ribosome, one of the most fundamental processes of cell biology. Alternatively, these snoRNAs may exert their influence over mRNAs instead, thereby regulating the stability or translation of circadian target genes. Consistent with this possibility, we found that snoRNAs can target several known clock genes in the fly, such as Takeout (to).

UMSL logo.jpg

Hughes Lab
Department of Biology
University of Missouri, St. Louis
(Starting in August, 2013)