Paraspeckles are nuclear bodies form around the long non-coding RNA, Neat1, and RNA-binding proteins. While their role is not fully understood, they are believed to control gene expression at a post-transcriptional level by means of the nuclear retention of mRNA containing in their 3’-UTR inverted repeats of Alu sequences (IRAlu). In this study, we found that, in pituitary cells, all components of paraspeckles including four major proteins and Neat1 displayed a circadian expression pattern. Furthermore the insertion of IRAlu at the 3’-UTR of the EGFP cDNA led to a rhythmic circadian nuclear retention of the egfp mRNA that was lost when paraspeckles were disrupted whereas insertion of a single antisense Alu had only a weak effect. Using real-time video-microscopy, these IRAlu were further shown to drive a circadian expression of EGFP protein. This study shows that paraspeckles, thanks to their circadian expression, control circadian gene expression at a post-transcriptional level.DOI: http://dx.doi.org/10.7554/eLife.14837.001

The daily temporal organization of rhythmic functions in mammals, which requires synchronization of the circadian clock to the 24-h light-dark cycle, is believed to involve adjustments of the mutual phasing of the cellular oscillators that comprise the time-keeper within the suprachiasmatic nucleus of the hypothalamus (SCN). Following from a previous study showing that the SCN undergoes day/night rearrangements of its neuronal-glial network that may be crucial for intercellular phasing, we investigated the contribution of glutamatergic synapses, known to play major roles in SCN functioning, to such rhythmic plastic events. Neither expression levels of the vesicular glutamate transporters nor numbers of glutamatergic terminals showed nycthemeral variations in the SCN. However, using quantitative imaging after combined immunolabelling, the density of synapses on neurons expressing vasoactive intestinal peptide, known as targets of the retinal input, increased during the day and both glutamatergic and non-glutamatergic synapses contributed to the increase (+36%). This was not the case for synapses made on vasopressin-containing neurons, the other major source of SCN efferents in the non-retinorecipient region. Together with electron microscope observations showing no differences in the morphometric features of glutamatergic terminals during the day and night, these data show that the light synchronization process in the SCN involves a selective remodelling of synapses at sites of photic integration. They provide a further illustration of how the adult brain may rapidly and reversibly adapt its synaptic architecture to functional needs.