Project description:Mutation effects prediction is a fundamental challenge in biotechnology and biomedicine. State-of-the-art computational methods have demonstrated the benefits of including semantically rich representations learned from protein sequences, but leave structural constraints out of reach. Here we developed Protein Mutational Effect Predictor (ProMEP), a general and multimodal deep representation learning method that simultaneously learns sequence context and structural constraints from proteins at the scale of evolution. ProMEP markedly outperforms current leading methods and enables accurate zero-shot mutational effects prediction across a variety of deep mutational scanning experiments. The application of ProMEP in the transposon-associated TnpB enzyme engineering task further demonstrates its ability for high-throughput protein space exploration. Without prior knowledge of TnpB, ProMEP accurately identifies multiple mutations that significantly improve the editing efficiency from millions of variants.
Project description:MicF is a textbook example of a small regulatory RNA (sRNA) that acts on a trans-encoded target mRNA through imperfect base paring. The discovery of MicF as a post-transcriptional repressor of the major Escherichia coli porin OmpF established the paradigm for a meanwhile common mechanism of translational inhibition, through antisense sequestration of a ribosome binding site. However, whether MicF regulates additional genes has remained unknown for almost three decades. Here, we have harnessed the novel superfolder variant of GFP for reporter-gene fusions to validate newly predicted targets of MicF in Salmonella. We show that the conserved 5’ end of MicF acts by seed pairing to repress the mRNAs of global transcriptional regulator Lrp, periplasmic protein YahO, and lipid A-modifying enzyme LpxR. Whilst MicF binds lrp and yahO in the 5’ UTR, it targets lpxR at both the ribosome binding site and deep within the coding sequence. Repression in the coding sequence of lpxR may be achieved by decreasing mRNA stability through exacerbating the use of a native RNase E site proximal to the short MicF-lpxR duplex. Altogether, this study assigns the classic MicF sRNA to the growing class of Hfqassociated regulators that use diverse mechanisms to impact multiple loci.
Project description:MicF is a textbook example of a small regulatory RNA (sRNA) that acts on a trans-encoded target mRNA through imperfect base paring. The discovery of MicF as a post-transcriptional repressor of the major Escherichia coli porin OmpF established the paradigm for a meanwhile common mechanism of translational inhibition, through antisense sequestration of a ribosome binding site. However, whether MicF regulates additional genes has remained unknown for almost three decades. Here, we have harnessed the novel superfolder variant of GFP for reporter-gene fusions to validate newly predicted targets of MicF in Salmonella. We show that the conserved 5’ end of MicF acts by seed pairing to repress the mRNAs of global transcriptional regulator Lrp, periplasmic protein YahO, and lipid A-modifying enzyme LpxR. Whilst MicF binds lrp and yahO in the 5’ UTR, it targets lpxR at both the ribosome binding site and deep within the coding sequence. Repression in the coding sequence of lpxR may be achieved by decreasing mRNA stability through exacerbating the use of a native RNase E site proximal to the short MicF-lpxR duplex. Altogether, this study assigns the classic MicF sRNA to the growing class of Hfqassociated regulators that use diverse mechanisms to impact multiple loci. To determine the targets of the small regulatory RNA MicF in S. Typhimurium, we looked at the effect of a short pulse of MicF over-expression on the Salmonella transcriptome. To achieve over-expression, the micF gene was cloned in the pBAD plasmid and induced with 0.2% L-arabinose for 10 min. We then extracted the total RNA for transcriptional profiling. A strain carrying the pBAD plasmid w/o insert was used as negative control (also induced by L-arabinose). 3 biological replicates were performed. This sRNA target identification strategy has been described in Papenfort et al; Molecular Microbiology (2006) 62(6), 1674?1688.
Project description:We compared several E. coli AlkB variants for their applications in facilitating sequencing of tRNA species in total RNA. These proteins are wild type AlkB, mutant D135S and mutant D135T.
Project description:It is known that current the-art-of-the-state TadA8 and TadA8e which evolved from E. coli TadA. They inherited the 'YA' context from tRNA deaminase. We started with wildtype E. coli TadA and designed an evolution campaign to force TadA variants to deaminate “GA” with fast kinetics. Three rounds of de novo directed evolution followed by DNA shuffling led to TadA8r, a TadA variant of superior “RA” deamination activity. TadA8r acts on a broadened editing window when fused to Streptococcus pyogenes Cas9 (SpCas9) and delivers robust editing at PAM distal positions. While highly active at on-target sites, ABE8r shows off-target DNA and RNA editing much lower than ABE8e. The off-target effects of ABE8r can be further mitigated by introducing a V106W substitution23, a R153 deletion22, or by mRNA delivery. Lastly, we demonstrate ABE8r-mediated correction of G1961E in ABCA4, the most prevalent mutation driving Stargardt disease (STGD1), in a “GA” context. ABE8r, with its superior activity and broadened context compatibility, complements and expands the current ABE family.
Project description:The sex chromosome-encoded RNA helicases DDX3X and DDX3Y play important roles in RNA metabolism. Heterozygous mutations of DDX3X frequently occur in cancers and neurodevelopmental disorders which have strong sex biases. However, how different DDX3X variants impair cellular function in sex specific genetic background is not understood. Herein, we found that DDX3X variants with significantly impaired ATPase activities demixed into the shells of unique hollow condensates, the dynamics of which were further differentiated by the RNA binding affinities of the different DDX3X variants. Proteomic and imaging studies revealed that DDX3X variant condensates sequestered wild-type DDX3X, DDX3Y, and other proteins important for various signaling pathways. Intriguingly, wild-type DDX3X improved the dynamics of heterogenous variant/wild-type hollow condensates more than DDX3Y. These results suggest that DDX3X variants with distinct enzymatic and condensation propensities may interact uniquely with wild-type DDX3X or DDX3Y to cause sex-specific cellular impacts.
Project description:Sequencing studies of autism spectrum disorder (ASD) cases have revealed a causal role for mutations to chromatin remodeling genes. Chromodomain helicase DNA binding protein 8 (CHD8) encodes a chromatin remodeler with one of the highest de novo mutation rates in sporadic ASD. However, the relationship between CHD8 genomic function and autism-relevant biology remains poorly elucidated. CHD8 binding studies have relied on Chromatin Immunoprecipitation followed by sequencing (ChIP-seq), but these datasets exhibit significant variation. ChIP-seq has technical limitations in the context of weak or indirect protein-DNA interactions or when high-performance antibodies are unavailable. Thus, complementary approaches are needed to establish CHD8 genomic targets. Here we used Targeted DamID in utero to characterize CHD8 binding activity in the developing embryonic mouse cortex. CHD8 Targeted DamID followed by sequencing (CHD8 TaDa-seq) revealed binding at previously identified genomic targets as well as at genes sensitive to Chd8 haploinsufficiency. CHD8 TaDa-seq showed greater sensitivity for CHD8 binding near a subset of genes specific to brain development and neuron function. These studies establish TaDa-seq as a useful alternative for mapping protein-DNA interactions in vivo and provide insights into the relationship between chromatin remodeling by CHD8 and autism-relevant pathophysiology associated with CHD8 mutations.