We recently described the mitochondrial localization and import of the vitamin D receptor (VDR) in actively proliferating HaCaT cells for the first time, but its role in the organelle remains unknown. Many metabolic intermediates that support cell growth are provided by the mitochondria; consequently, the identification of proteins that regulate mitochondrial metabolic pathways is of great interest, and we sought to understand whether VDR may modulate these pathways. We genetically silenced VDR in HaCaT cells and studied the effects on cell growth, mitochondrial metabolism and biosynthetic pathways. VDR knockdown resulted in robust growth inhibition, with accumulation in the G0G1 phase of the cell cycle and decreased accumulation in the M phase. The effects of VDR silencing on proliferation were confirmed in several human cancer cell lines. Decreased VDR expression was consistently observed in two different models of cell differentiation. The impairment of silenced HaCaT cell growth was accompanied by sharp increases in the mitochondrial membrane potential, which sensitized the cells to oxidative stress. We found that transcription of the subunits II and IV of cytochrome c oxidase was significantly increased upon VDR silencing. Accordingly, treatment of HaCaT cells with vitamin D downregulated both subunits, suggesting that VDR may inhibit the respiratory chain and redirect TCA intermediates toward biosynthesis, thus contributing to the metabolic switch that is typical of cancer cells. In order to explore this hypothesis, we examined various acetyl-CoA-dependent biosynthetic pathways, such as the mevalonate pathway (measured as cholesterol biosynthesis and prenylation of small GTPases), and histone acetylation levels; all of these pathways were inhibited by VDR silencing. These data provide evidence of the role of VDR as a gatekeeper of mitochondrial respiratory chain activity and a facilitator of the diversion of acetyl-CoA from the energy-producing TCA cycle toward biosynthetic pathways that are essential for cellular proliferation.

For most biological processes, organisms must respond to extrinsic cues, while maintaining essential gene expression programmes. Although studied extensively in single cells, it is still unclear how variation is controlled in multicellular organisms. Here, we used a machine‐learning approach to identify genomic features that are predictive of genes with high versus low variation in their expression across individuals, using bulk data to remove stochastic cell‐to‐cell variation. Using embryonic gene expression across 75 Drosophila isogenic lines, we identify features predictive of expression variation (controlling for expression level), many of which are promoter‐related. Genes with low variation fall into two classes reflecting different mechanisms to maintain robust expression, while genes with high variation seem to lack both types of stabilizing mechanisms. Applying this framework to humans revealed similar predictive features, indicating that promoter architecture is an ancient mechanism to control expression variation. Remarkably, expression variation features could also partially predict differential expression after diverse perturbations in both Drosophila and humans. Differential gene expression signatures may therefore be partially explained by genetically encoded gene‐specific features, unrelated to the studied treatment.

SummaryLong non-coding RNAs (lncRNAs) can often function in the regulation of gene expression during development; however, their generality as essential regulators in developmental processes and organismal phenotypes remains unclear. Here, we performed a tailored investigation of lncRNA expression and function during Drosophila embryogenesis, interrogating multiple stages, tissue specificity, nuclear localization, and genetic backgrounds. Our results almost double the number of annotated lncRNAs expressed at these embryonic stages. lncRNA levels are generally positively correlated with those of their neighboring genes, with little evidence of transcriptional interference. Using fluorescent in situ hybridization, we report the spatiotemporal expression of 15 new lncRNAs, revealing very dynamic tissue-specific patterns. Despite this, deletion of selected lncRNA genes had no obvious developmental defects or effects on viability under standard and stressed conditions. However, two lncRNA deletions resulted in modest expression changes of a small number of genes, suggesting that they fine-tune expression of non-essential genes. Several lncRNAs have strain-specific expression, indicating that they are not fixed within the population. This intra-species variation across genetic backgrounds may thereby be a useful tool to distinguish rapidly evolving lncRNAs with as yet non-essential roles.

21For most biological processes, organisms must respond to extrinsic cues, while maintaining 22 essential gene expression programs. Although studied extensively in single cells, it is still 23 unclear how variation is controlled in multicellular organisms. Here, we used a machine-24 learning approach to identify genomic features that are predictive of genes with high versus 25 low variation in their expression across individuals, using bulk data to remove stochastic cell-26

The study of genetic variation has been revolutionized by the advent of high-throughput technologies able to determine the complete genomic sequence of thousands of individuals. Understanding the functional relevance of variants is, however, still a difficult task, especially when focusing on non-coding variants. Most of the variants associated with disease by Genome-Wide Association Studies (GWAS) are indeed non-coding, and presumably exert their effects by altering gene regulation. Expression Quantitative Trait Loci (eQTL) studies represent an important step in understanding the functional relevance of regulatory variants. We propose a new strategy to detect and characterize eQTLs, based on the effect of variants on the Total Binding Affinity (TBA) profiles of regulatory regions. Using a large dataset of coupled genome and expression data, we show that TBA-based inference allows the identification of eQTLs not revealed by traditional methods and helps in their interpretation in terms of altered transcription factor binding.