The RNA Pol II elongation factor Ell3 marks enhancers in ES cells and primes future gene activation.

Enhancers play a central role in precisely regulating the expression of developmentally regulated genes. However, the machineries required for enhancer-promoter communication have remained largely unknown. We have found that Ell3, a member of the Ell (eleven-nineteen lysine-rich leukemia gene) family of RNA Pol II elongation factors, occupies enhancers in embryonic stem cells. Ell3's association with enhancers is required for setting up proper Pol II occupancy at the promoter-proximal regions of developmentally regulated genes and for the recruitment of the super elongation complex (SEC) to these loci following differentiation signals. Furthermore, Ell3 binding to inactive or poised enhancers is essential for stem cell specification. We have also detected the presence of Pol II and Ell3 in germ cell nuclei. These findings raise the possibility that transcription factors could prime gene expression by marking enhancers in ES cells or even as early as in the germ cell state.

Deep representation learning improves prediction of LacI-mediated transcriptional repression.

Recent progress in DNA synthesis and sequencing technology has enabled systematic studies of protein function at a massive scale. We explore a deep mutational scanning study that measured the transcriptional repression function of 43,669 variants of the Escherichia coli LacI protein. We analyze structural and evolutionary aspects that relate to how the function of this protein is maintained, including an in-depth look at the C-terminal domain. We develop a deep neural network to predict transcriptional repression mediated by the lac repressor of Escherichia coli using experimental measurements of variant function. When measured across 10 separate training and validation splits using 5,009 single mutations of the lac repressor, our best-performing model achieved a median Pearson correlation of 0.79, exceeding any previous model. We demonstrate that deep representation learning approaches, first trained in an unsupervised manner across millions of diverse proteins, can be fine-tuned in a supervised fashion using lac repressor experimental datasets to more effectively predict a variant's effect on repression. These findings suggest a deep representation learning model may improve the prediction of other important properties of proteins.

The little elongation complex functions at initiation and elongation phases of snRNA gene transcription.

The small nuclear RNA (snRNA) genes have been widely used as a model system for understanding transcriptional regulation due to the unique aspects of their promoter structure, selectivity for either RNA polymerase (Pol) II or III, and because of their unique mechanism of termination that is tightly linked with the promoter. Recently, we identified the little elongation complex (LEC) in Drosophila that is required for the expression of Pol II-transcribed snRNA genes. Here, using Drosophila and mammalian systems, we provide genetic and molecular evidence that LEC functions in at least two phases of snRNA transcription: an initiation step requiring the ICE1 subunit, and an elongation step requiring ELL.

Enhancer-associated H3K4 monomethylation by Trithorax-related, the Drosophila homolog of mammalian Mll3/Mll4.

Monomethylation of histone H3 on Lys 4 (H3K4me1) and acetylation of histone H3 on Lys 27 (H3K27ac) are histone modifications that are highly enriched over the body of actively transcribed genes and on enhancers. Although in yeast all H3K4 methylation patterns, including H3K4me1, are implemented by Set1/COMPASS (complex of proteins associated with Set1), there are three classes of COMPASS-like complexes in Drosophila that could carry out H3K4me1 on enhancers: dSet1, Trithorax, and Trithorax-related (Trr). Here, we report that Trr, the Drosophila homolog of the mammalian Mll3/4 COMPASS-like complexes, can function as a major H3K4 monomethyltransferase on enhancers in vivo. Loss of Trr results in a global decrease of H3K4me1 and H3K27ac levels in various tissues. Assays with the cut wing margin enhancer implied a functional role for Trr in enhancer-mediated processes. A genome-wide analysis demonstrated that Trr is required to maintain the H3K4me1 and H3K27ac chromatin signature that resembles the histone modification patterns described for enhancers. Furthermore, studies in the mammalian system suggested a role for the Trr homolog Mll3 in similar processes. Since Trr and mammalian Mll3/4 complexes are distinguished by bearing a unique subunit, the H3K27 demethylase UTX, we propose a model in which the H3K4 monomethyltransferases Trr/Mll3/Mll4 and the H3K27 demethylase UTX cooperate to regulate the transition from inactive/poised to active enhancers.

Engineering an allosteric transcription factor to respond to new ligands.

Genetic regulatory proteins inducible by small molecules are useful synthetic biology tools as sensors and switches. Bacterial allosteric transcription factors (aTFs) are a major class of regulatory proteins, but few aTFs have been redesigned to respond to new effectors beyond natural aTF-inducer pairs. Altering inducer specificity in these proteins is difficult because substitutions that affect inducer binding may also disrupt allostery. We engineered an aTF, the Escherichia coli lac repressor, LacI, to respond to one of four new inducer molecules: fucose, gentiobiose, lactitol and sucralose. Using computational protein design, single-residue saturation mutagenesis or random mutagenesis, along with multiplex assembly, we identified new variants comparable in specificity and induction to wild-type LacI with its inducer, isopropyl ß-D-1-thiogalactopyranoside (IPTG). The ability to create designer aTFs will enable applications including dynamic control of cell metabolism, cell biology and synthetic gene circuits.

A deep learning approach to programmable RNA switches.

Engineered RNA elements are programmable tools capable of detecting small molecules, proteins, and nucleic acids. Predicting the behavior of these synthetic biology components remains a challenge, a situation that could be addressed through enhanced pattern recognition from deep learning. Here, we investigate Deep Neural Networks (DNN) to predict toehold switch function as a canonical riboswitch model in synthetic biology. To facilitate DNN training, we synthesize and characterize in vivo a dataset of 91,534 toehold switches spanning 23 viral genomes and 906 human transcription factors. DNNs trained on nucleotide sequences outperform (R2 = 0.43-0.70) previous state-of-the-art thermodynamic and kinetic models (R2 = 0.04-0.15) and allow for human-understandable attention-visualizations (VIS4Map) to identify success and failure modes. This work shows that deep learning approaches can be used for functionality predictions and insight generation in RNA synthetic biology.

An unbiased index to quantify participant's phenotypic contribution to an open-access cohort.

The Personal Genome Project (PGP) is an effort to enroll many participants to create an open-access repository of genome, health and trait data for research. However, PGP participants are not enrolled for studying any specific traits and participants choose the phenotypes to disclose. To measure the extent and willingness and to encourage and guide participants to contribute phenotypes, we developed an algorithm to score and rank the phenotypes and participants of the PGP. The scoring algorithm calculates the participation index (P-index) for every participant, where 0 indicates no reported phenotypes and 100 indicate complete phenotype reporting. We calculated the P-index for all 5,015 participants in the PGP and they ranged from 0 to 96.7. We found that participants mainly have either high scores (P-index > 90, 29.5%) or low scores (P-index < 10, 57.8%). While, there are significantly more males than female participants (1,793 versus 1,271), females tend to have on average higher P-indexes (P = 0.015). We also reported the P-indexes of participants based on demographics and states like Missouri and Massachusetts have better P-indexes than states like Utah and Minnesota. The P-index can therefore be used as an unbiased way to measure and rank participant's phenotypic contribution towards the PGP.

Dataset of transcriptional landscape of B cell early activation.

Signaling via B cell receptors (BCR) and Toll-like receptors (TLRs) result in activation of B cells with distinct physiological outcomes, but transcriptional regulatory mechanisms that drive activation and distinguish these pathways remain unknown. At early time points after BCR and TLR ligand exposure, 0.5 and 2 h, RNA-seq was performed allowing observations on rapid transcriptional changes. At 2 h, ChIP-seq was performed to allow observations on important regulatory mechanisms potentially driving transcriptional change. The dataset includes RNA-seq, ChIP-seq of control (Input), RNA Pol II, H3K4me3, H3K27me3, and a separate RNA-seq for miRNA expression, which can be found at Gene Expression Omnibus Dataset GSE61608. Here, we provide details on the experimental and analysis methods used to obtain and analyze this dataset and to examine the transcriptional landscape of B cell early activation.

Divergence of transcriptional landscape occurs early in B cell activation.

BACKGROUND: Signaling via B cell receptor (BCR) and Toll-like receptors (TLRs) results in activation of B cells with distinct physiological outcomes, but transcriptional regulatory mechanisms that drive activation and distinguish these pathways remain unknown. RESULTS: Two hours after ligand exposure RNA-seq, ChIP-seq and computational methods reveal that BCR- or TLR-mediated activation of primary resting B cells proceeds via a large set of shared and a smaller subset of distinct signal-selective transcriptional responses. BCR stimulation resulted in increased global recruitment of RNA Pol II to promoters that appear to transit slowly to downstream regions. Conversely, lipopolysaccharide (LPS) stimulation involved an enhanced RNA Pol II transition from initiating to elongating mode accompanied by greater H3K4me3 activation markings compared to BCR stimulation. These rapidly diverging transcriptomic landscapes also show distinct repressing (H3K27me3) histone signatures, mutually exclusive transcription factor binding in promoters, and unique miRNA profiles. CONCLUSIONS: Upon examination of genome-wide transcription and regulatory elements, we conclude that the B cell commitment to different activation states occurs much earlier than previously thought and involves a multi-faceted receptor-specific transcriptional landscape.

The Mll2 branch of the COMPASS family regulates bivalent promoters in mouse embryonic stem cells.

Promoters of many developmentally regulated genes, in the embryonic stem cell state, have a bivalent mark of H3K27me3 and H3K4me3, proposed to confer precise temporal activation upon differentiation. Although Polycomb repressive complex 2 is known to implement H3K27 trimethylation, the COMPASS family member responsible for H3K4me3 at bivalently marked promoters was previously unknown. Here, we identify Mll2 (KMT2b) as the enzyme catalyzing H3K4 trimethylation at bivalentlymarked promoters in embryonic stem cells. Although H3K4me3 at bivalent genes is proposed to prime future activation, we detected no substantial defect in rapid transcriptional induction after retinoic acid treatment in Mll2-depleted cells. Our identification of the Mll2 complex as the COMPASS family member responsible for H3K4me3 marking bivalent promoters provides an opportunity to reevaluate and experimentally test models for the function of bivalency in the embryonic stem cell state and in differentiation.

Sequencing of the sea lamprey (Petromyzon marinus) genome provides insights into vertebrate evolution.

Lampreys are representatives of an ancient vertebrate lineage that diverged from our own ∼500 million years ago. By virtue of this deeply shared ancestry, the sea lamprey (P. marinus) genome is uniquely poised to provide insight into the ancestry of vertebrate genomes and the underlying principles of vertebrate biology. Here, we present the first lamprey whole-genome sequence and assembly. We note challenges faced owing to its high content of repetitive elements and GC bases, as well as the absence of broad-scale sequence information from closely related species. Analyses of the assembly indicate that two whole-genome duplications likely occurred before the divergence of ancestral lamprey and gnathostome lineages. Moreover, the results help define key evolutionary events within vertebrate lineages, including the origin of myelin-associated proteins and the development of appendages. The lamprey genome provides an important resource for reconstructing vertebrate origins and the evolutionary events that have shaped the genomes of extant organisms.

Poised RNA polymerase II changes over developmental time and prepares genes for future expression.

Poised RNA polymerase II (Pol II) is predominantly found at developmental control genes and is thought to allow their rapid and synchronous induction in response to extracellular signals. How the recruitment of poised RNA Pol II is regulated during development is not known. By isolating muscle tissue from Drosophila embryos at five stages of differentiation, we show that the recruitment of poised Pol II occurs at many genes de novo and this makes them permissive for future gene expression. A comparison with other tissues shows that these changes are stage specific and not tissue specific. In contrast, Polycomb group repression is tissue specific, and in combination with Pol II (the balanced state) marks genes with highly dynamic expression. This suggests that poised Pol II is temporally regulated and is held in check in a tissue-specific fashion. We compare our data with findings in mammalian embryonic stem cells and discuss a framework for predicting developmental programs on the basis of the chromatin state.