Project description:Both upregulation and downregulation by cis-regulatory elements help establish precise gene expression. Our understanding of how elements repress transcriptional activity is far more limited than activating elements. To address this gap, we characterized RE1, a group of transcriptional silencers bound by REST, on a genome-wide scale using an optimized massively parallel reporter assay (MPRAduo). MPRAduo empirically defined a minimal binding strength of REST required by silencer (REST m-value), above which multiple cofactors colocalize and act to directly silence transcription. We identified 1,500 human variants that alter RE1 silencing and found their effect sizes are predictable when they overlap with REST binding sites above the m-value. In addition, we demonstrate that non-canonical REST binding motifs exhibit silencer function only if they precisely align two half sites with specific spacer length. Our results show mechanistic insights into RE1 silencer which allows us to predict its activity and effect of variants on RE1, providing a paradigm for performing genome-wide functional characterization transcription factors binding sites.
Project description:Both upregulation and downregulation by cis-regulatory elements help establish precise gene expression. Our understanding of how elements repress transcriptional activity is far more limited than activating elements. To address this gap, we characterized RE1, a group of transcriptional silencers bound by REST, on a genome-wide scale using an optimized massively parallel reporter assay (MPRAduo). MPRAduo empirically defined a minimal binding strength of REST required by silencer (REST m-value), above which multiple cofactors colocalize and act to directly silence transcription. We identified 1,500 human variants that alter RE1 silencing and found their effect sizes are predictable when they overlap with REST binding sites above the m-value. In addition, we demonstrate that non-canonical REST binding motifs exhibit silencer function only if they precisely align two half sites with specific spacer length. Our results show mechanistic insights into RE1 silencer which allows us to predict its activity and effect of variants on RE1, providing a paradigm for performing genome-wide functional characterization transcription factors binding sites.
Project description:Gas-fermenting acetogens, such as Clostridium autoethanogenum, have emerged as promising biocatalysts capable of converting CO and CO2-containing gases into fuels and chemicals relevant for a circular economy. However, functionalities of the majority of genes in acetogens remain uncharacterised, hindering the development of acetogen cell factories through targeted genetic engineering. We previously identified gene targets through adaptive laboratory evolution (ALE) that potentially realise enhanced autotrophic phenotypes in C. autoethanogenum. In this study, we deleted one of the targets – CLAU_0471 (proposed amino acid permease) – with high mutation occurrence in ALE isolates and extensively characterised autotrophic growth of strain RE3 in batch bottle and bioreactor continuous cultures. In addition, we characterized two previously reverse-engineered strains RE1 (deletion of CLAU_3129; putative sporulation transcriptional activator Spo0A) and RE2 (SNP in CLAU_1957; proposed two component transcriptional regulator winged helix family). Strikingly, the strains recovered the superior phenotypes of ALE isolates, including faster autotrophic growth, no need for yeast extract, and robustness in bioreactor operation (e.g. low sensitivity to gas ramping, high biomass, and dilution rates). Notably, RE3 exhibited elevated 2,3-butanediol production while RE1 performed similar to the best-performing previously characterised ALE isolate LAbrini. The three reverse-engineered strains showed similarities in proteome expression and bioinformatic analyses suggest that the targeted genes may be involved in overlapping regulatory networks. Our work provides insights into genotype-phenotype relationships for a better understanding of the metabolism of an industrially-relevant acetogen.
Project description:Genome annotation of the chelicerate Tetranychus urticae revealed the absence of many canonical immunity genes. T. urticae either does not mount an immune response or it induces uncharacterized immune pathways. To disentangle these two hypotheses, we performed transcriptomic analysis of mites injected with bacteria vs mites injected with LB-buffer. Two types of bacteria were injected: E. coli and B. megaterium and transcriptomes were sampled 3, 6 and 12 hrs after injection. We found no consistent differential expression after bacterial infection, supporting the hypothesis that spider mites do not mount an immune response. We hypothesize that the apparent absence of inducable immunity pathways in T. urticae is a result of relaxed selective pressure due to ecological factors.
Project description:The delineation of genes in bacteria has remained an important challenge because prokaryotic genomes are often tightly packed frequently resulting in overlapping genes. We hereby present a de novo approach called REPARATION (RibosomeE Profiling Assisted (Re-)AnnotaTION) to delineate translated open reading frames (ORFs) in bacteria independent of (available) genome annotation. By deep sequencing of ribosome protected mRNA fragments (RPF) to map translating ribosomes across the entire genome, REPARATION takes advantage of the recently developed ribosome profiling (Ribo-seq) technique. REPARATION starts by traversing the entire genome to generate all possible ORFs and then collects their corresponding RPF signal information. Based on a growth curve model to estimate minimum ORF read density and Ribo-seq RPF coverage, thresholds indicative of translation is estimated. Finally, our algorithm applies a random forest model to build a classifier to classify putative protein coding ORFs. We evaluated the performance of REPARATION on 3 annotated bacterial species using in-house generated Ribo-seq data and matching N-terminal and shotgun proteomics data next to publically available Ribo-seq data. In all cases, about 80% of the ORFs predicted by REPARATION were previously annotated as protein coding. While 13-20% were variants of previously annotated ORFs and about 3-4% point to novel translated ORFs within intergenic or other regions previously annotated as non-coding. Without stringent length restrictions REPARATION was able to identify several small ORFs (sORFs). Multiple supportive evidence from matching MS data and sequence conservation analysis was obtained to validate predicted ORFs.
