A Deep Learning Approach to Antibiotic Discovery.

Due to the rapid emergence of antibiotic-resistant bacteria, there is a growing need to discover new antibiotics. To address this challenge, we trained a deep neural network capable of predicting molecules with antibacterial activity. We performed predictions on multiple chemical libraries and discovered a molecule from the Drug Repurposing Hub-halicin-that is structurally divergent from conventional antibiotics and displays bactericidal activity against a wide phylogenetic spectrum of pathogens including Mycobacterium tuberculosis and carbapenem-resistant Enterobacteriaceae. Halicin also effectively treated Clostridioides difficile and pan-resistant Acinetobacter baumannii infections in murine models. Additionally, from a discrete set of 23 empirically tested predictions from >107 million molecules curated from the ZINC15 database, our model identified eight antibacterial compounds that are structurally distant from known antibiotics. This work highlights the utility of deep learning approaches to expand our antibiotic arsenal through the discovery of structurally distinct antibacterial molecules.

CRISPR-Based Technologies for the Manipulation of Eukaryotic Genomes.

The CRISPR-Cas9 RNA-guided DNA endonuclease has contributed to an explosion of advances in the life sciences that have grown from the ability to edit genomes within living cells. In this Review, we summarize CRISPR-based technologies that enable mammalian genome editing and their various applications. We describe recent developments that extend the generality, DNA specificity, product selectivity, and fundamental capabilities of natural CRISPR systems, and we highlight some of the remarkable advancements in basic research, biotechnology, and therapeutics science that these developments have facilitated.

Genetic Code Expansion History and Modern Innovations.

The genetic code is the foundation for all life. With few exceptions, the translation of nucleic acid messages into proteins follows conserved rules, which are defined by codons that specify each of the 20 proteinogenic amino acids. For decades, leading research groups have developed a catalogue of innovative approaches to extend nature's amino acid repertoire to include one or more noncanonical building blocks in a single protein. In this review, we summarize advances in the history of in vitro and in vivo genetic code expansion, and highlight recent innovations that increase the scope of biochemically accessible monomers and codons. We further summarize state-of-the-art knowledge in engineered cellular translation, as well as alterations to regulatory mechanisms that improve overall genetic code expansion. Finally, we distill existing limitations of these technologies into must-have improvements for the next generation of technologies, and speculate on future strategies that may be capable of overcoming current gaps in knowledge.

Directed evolution of hyperactive integrases for site specific insertion of transgenes.

The ability to deliver large transgenes to a single genomic sequence with high efficiency would accelerate biomedical interventions. Current methods suffer from low insertion efficiency and most rely on undesired double-strand DNA breaks. Serine integrases catalyze the insertion of large DNA cargos at attachment (att) sites. By targeting att sites to the genome using technologies such as prime editing, integrases can target safe loci while avoiding double-strand breaks. We developed a method of phage-assisted continuous evolution we call IntePACE, that we used to rapidly perform hundreds of rounds of mutagenesis to systematically improve activity of PhiC31 and Bxb1 serine integrases. Novel hyperactive mutants were generated by combining synergistic mutations resulting in integration of a multi-gene cargo at rates as high as 80% of target chromosomes. Hyperactive integrases inserted a 15.7 kb therapeutic DNA cargo containing von Willebrand Factor. This technology could accelerate gene delivery therapeutics and our directed evolution strategy can easily be adapted to improve novel integrases from nature.

Publisher Correction: Programmable base editing of A•T to G•C in genomic DNA without DNA cleavage.

In this Article, owing to an error during the production process, in Fig. 1a, the dark blue and light blue wedges were incorrectly labelled as 'G•C → T•A' and 'G•C → A•T', instead of 'C•G → T•A' and 'C•G → A•T', respectively. Fig. 1 has been corrected online.

Bacterial translation machinery for deliberate mistranslation of the genetic code.

Inaccurate expression of the genetic code, also known as mistranslation, is an emerging paradigm in microbial studies. Growing evidence suggests that many microbial pathogens can deliberately mistranslate their genetic code to help invade a host or evade host immune responses. However, discovering different capacities for deliberate mistranslation remains a challenge because each group of pathogens typically employs a unique mistranslation mechanism. In this study, we address this problem by studying duplicated genes of aminoacyl-transfer RNA (tRNA) synthetases. Using bacterial prolyl-tRNA synthetase (ProRS) genes as an example, we identify an anomalous ProRS isoform, ProRSx, and a corresponding tRNA, tRNAProA, that are predominately found in plant pathogens from Streptomyces species. We then show that tRNAProA has an unusual hybrid structure that allows this tRNA to mistranslate alanine codons as proline. Finally, we provide biochemical, genetic, and mass spectrometric evidence that cells which express ProRSx and tRNAProA can translate GCU alanine codons as both alanine and proline. This dual use of alanine codons creates a hidden proteome diversity due to stochastic Ala→Pro mutations in protein sequences. Thus, we show that important plant pathogens are equipped with a tool to alter the identity of their sense codons. This finding reveals the initial example of a natural tRNA synthetase/tRNA pair for dedicated mistranslation of sense codons.

Programmable base editing of A•T to G•C in genomic DNA without DNA cleavage.

The spontaneous deamination of cytosine is a major source of transitions from C•G to T•A base pairs, which account for half of known pathogenic point mutations in humans. The ability to efficiently convert targeted A•T base pairs to G•C could therefore advance the study and treatment of genetic diseases. The deamination of adenine yields inosine, which is treated as guanine by polymerases, but no enzymes are known to deaminate adenine in DNA. Here we describe adenine base editors (ABEs) that mediate the conversion of A•T to G•C in genomic DNA. We evolved a transfer RNA adenosine deaminase to operate on DNA when fused to a catalytically impaired CRISPR-Cas9 mutant. Extensive directed evolution and protein engineering resulted in seventh-generation ABEs that convert targeted A•T base pairs efficiently to G•C (approximately 50% efficiency in human cells) with high product purity (typically at least 99.9%) and low rates of indels (typically no more than 0.1%). ABEs introduce point mutations more efficiently and cleanly, and with less off-target genome modification, than a current Cas9 nuclease-based method, and can install disease-correcting or disease-suppressing mutations in human cells. Together with previous base editors, ABEs enable the direct, programmable introduction of all four transition mutations without double-stranded DNA cleavage.

Continuous evolution of Bacillus thuringiensis toxins overcomes insect resistance.

The Bacillus thuringiensis δ-endotoxins (Bt toxins) are widely used insecticidal proteins in engineered crops that provide agricultural, economic, and environmental benefits. The development of insect resistance to Bt toxins endangers their long-term effectiveness. Here we have developed a phage-assisted continuous evolution selection that rapidly evolves high-affinity protein-protein interactions, and applied this system to evolve variants of the Bt toxin Cry1Ac that bind a cadherin-like receptor from the insect pest Trichoplusia ni (TnCAD) that is not natively bound by wild-type Cry1Ac. The resulting evolved Cry1Ac variants bind TnCAD with high affinity (dissociation constant Kd = 11-41 nM), kill TnCAD-expressing insect cells that are not susceptible to wild-type Cry1Ac, and kill Cry1Ac-resistant T. ni insects up to 335-fold more potently than wild-type Cry1Ac. Our findings establish that the evolution of Bt toxins with novel insect cell receptor affinity can overcome insect Bt toxin resistance and confer lethality approaching that of the wild-type Bt toxin against non-resistant insects.

Continuous directed evolution of proteins with improved soluble expression.

We report the development of soluble expression phage-assisted continuous evolution (SE-PACE), a system for rapidly evolving proteins with increased soluble expression. Through use of a PACE-compatible AND gate that uses a split-intein pIII, SE-PACE enables two simultaneous positive selections to evolve proteins with improved expression while maintaining their desired activities. In as little as three days, SE-PACE evolved several antibody fragments with >5-fold improvement in expression yield while retaining binding activity. We also developed an activity-independent form of SE-PACE to correct folding-defective variants of maltose-binding protein (MBP) and to evolve variants of the eukaryotic cytidine deaminase APOBEC1 with improved expression properties. These evolved APOBEC1 variants were found to improve the expression and apparent activity of Cas9-derived base editors when used in place of the wild-type cytidine deaminase. Together, these results suggest that SE-PACE can be applied to a wide variety of proteins to rapidly improve their soluble expression.

Continuous directed evolution of DNA-binding proteins to improve TALEN specificity.

Nucleases containing programmable DNA-binding domains can alter the genomes of model organisms and have the potential to become human therapeutics. Here we present DNA-binding phage-assisted continuous evolution (DB-PACE) as a general approach for the laboratory evolution of DNA-binding activity and specificity. We used this system to generate transcription activator-like effectors nucleases (TALENs) with broadly improved DNA cleavage specificity, establishing DB-PACE as a versatile approach for improving the accuracy of genome-editing agents.

Negative selection and stringency modulation in phage-assisted continuous evolution.

Phage-assisted continuous evolution (PACE) uses a modified filamentous bacteriophage life cycle to substantially accelerate laboratory evolution experiments. In this work, we expand the scope and capabilities of the PACE method with two key advances that enable the evolution of biomolecules with radically altered or highly specific new activities. First, we implemented small molecule-controlled modulation of selection stringency that enables otherwise inaccessible activities to be evolved directly from inactive starting libraries through a period of evolutionary drift. Second, we developed a general negative selection that enables continuous counterselection against undesired activities. We integrated these developments to continuously evolve mutant T7 RNA polymerase enzymes with ∼10,000-fold altered, rather than merely broadened, substrate specificities during a single three-day PACE experiment. The evolved enzymes exhibit specificity for their target substrate that exceeds that of wild-type RNA polymerases for their cognate substrates while maintaining wild type-like levels of activity.

Efficient genetic code expansion without host genome modifications.

Supplementing translation with noncanonical amino acids (ncAAs) can yield protein sequences with new-to-nature functions but existing ncAA incorporation strategies suffer from low efficiency and context dependence. We uncover codon usage as a previously unrecognized contributor to efficient genetic code expansion using non-native codons. Relying only on conventional Escherichia coli strains with native ribosomes, we develop a plasmid-based codon compression strategy that minimizes context dependence and improves ncAA incorporation at quadruplet codons. We confirm that this strategy is compatible with all known genetic code expansion resources, which allowed us to identify 12 mutually orthogonal transfer RNA (tRNA)-synthetase pairs. Enabled by these findings, we evolved and optimized five tRNA-synthetase pairs to incorporate a broad repertoire of ncAAs at orthogonal quadruplet codons. Lastly, we extend these resources to an in vivo biosynthesis platform that can readily create >100 new-to-nature peptide macrocycles bearing up to three unique ncAAs. Our approach will accelerate innovations in multiplexed genetic code expansion and the discovery of chemically diverse biomolecules.

Directed evolution of hyperactive integrases for site specific insertion of transgenes.

The ability to deliver large transgenes to a single genomic sequence with high efficiency would accelerate biomedical interventions. Current methods suffer from low insertion efficiency and most rely on undesired double-strand DNA breaks. Serine integrases catalyze the insertion of large DNA cargos at attachment (att) sites. By targeting att sites to the genome using technologies such as prime editing, integrases can target safe loci while avoiding double-strand breaks. We developed a method of phage-assisted continuous evolution we call IntePACE, that we used to rapidly perform hundreds of rounds of mutagenesis to systematically improve activity of PhiC31 and Bxb1 serine integrases. Novel hyperactive mutants were generated by combining synergistic mutations resulting in integration of a multi-gene cargo at rates as high as 80% of target chromosomes. Hyperactive integrases inserted a 15.7 kb therapeutic DNA cargo containing Von Willebrand Factor. This technology could accelerate gene delivery therapeutics and our directed evolution strategy can easily be adapted to improve novel integrases from nature.

Natural and engineered precision antibiotics in the context of resistance.

Antibiotics are essential weapons in our fight against infectious disease, yet the consequences of broad-spectrum antibiotic use on microbiome stability and pathogen resistance are prompting investigations into more selective alternatives. Echoing the advent of precision medicine in oncology, precision antibiotics with focused activities are emerging as a means of addressing infections without damaging microbiomes or incentivizing resistance. Historically, antibiotic design principles have been gleaned from Nature, and reinvestigation of overlooked antibacterials is now providing scaffolds and targets for the design of pathogen-specific drugs. In this perspective, we summarize the biosynthetic and antibacterial mechanisms used to access these activities, and discuss how such strategies may be co-opted through engineering approaches to afford precision antibiotics.

Turn-on DNA damage sensors for the direct detection of 8-oxoguanine and photoproducts in native DNA.

The integrity of the genetic information in all living organisms is constantly threatened by a variety of endogenous and environmental insults. To counter this risk, the DNA-damage response is employed for repairing lesions and maintaining genomic integrity. However, an aberrant DNA-damage response can potentially lead to genetic instability and mutagenesis, carcinogenesis, or cell death. To directly monitor DNA damage events in the context of native DNA, we have designed two new sensors utilizing genetically fragmented firefly luciferase (split luciferase). The sensors are comprised of a methyl-CpG binding domain (MBD) attached to one fragment of split luciferase for localizing the sensor to DNA (50-80% of the CpG dinucleotide sites in the genome are symmetrically methylated at cytosines), while a damage-recognition domain is attached to the complementary fragment of luciferase to probe adjacent nucleotides for lesions. Specifically, we utilized oxoguanine glycosylase 1 (OGG1) to detect 8-oxoguanine caused by exposure to reactive oxygen species and employed the damaged-DNA binding protein 2 (DDB2) for detection of pyrimidine dimer photoproducts induced by UVC light. These two sensors were optimized and validated using oligonucleotides, plasmids, and mammalian genomic DNA, as well as HeLa cells that were systematically exposed to a variety of environmental insults, demonstrating that this methodology utilizing MBD-directed DNA localization provides a simple, sensitive, and potentially general approach for the rapid profiling of specific chemical modifications associated with DNA damage and repair.

Evaluating the global CpG methylation status of native DNA utilizing a bipartite split-luciferase sensor.

Epigenetic modifications play an essential role in the regulation of gene expression and ultimately cell fate. Methylation of cytosine at CpG dinucleotides (mCpG) is an important epigenetic mark that has been correlated with cancer when present at promoter sites of tumor suppressor genes. To develop a rapid methodology for the direct assessment of global levels of DNA methylation, we first interrogated the methyl-CpG binding domains (MBDs), the Kaiso family of Cys(2)-His(2) zinc fingers, and an SET- and RING-associated domain using a split-luciferase reassembly methodology. We identified MBD1 as the most selective domain for the discrimination between mCpG and CpG sites with over 90-fold selectivity. Utilizing a bipartite strategy, we constructed a purely methylation-dependent bipartite sensor for the direct detection of global levels of DNA methylation by attaching MBD1 domains to each of the split-luciferase halves. This new sensor was validated for the direct determination of genomic DNA methylation levels in in vitro studies without any intervening chemical or enzymatic processing of DNA. Finally, we demonstrated that this bipartite sensor can be utilized for monitoring dose-dependent changes in global levels of methylation in DNA from HeLa cells challenged with 5-aza-2'-deoxycytidine, a DNA methyltransferase inhibitor.

Synthetic Biological Circuits within an Orthogonal Central Dogma.

Synthetic biology strives to reliably control cellular behavior, typically in the form of user-designed interactions of biological components to produce a predetermined output. Engineered circuit components are frequently derived from natural sources and are therefore often hampered by inadvertent interactions with host machinery, most notably within the host central dogma. Reliable and predictable gene circuits require the targeted reduction or elimination of these undesirable interactions to mitigate negative consequences on host fitness and develop context-independent bioactivities. Here, we review recent advances in biological orthogonalization, namely the insulation of researcher-dictated bioactivities from host processes, with a focus on systematic developments that may culminate in the creation of an orthogonal central dogma and novel cellular functions.

Directed evolution of rRNA improves translation kinetics and recombinant protein yield.

In bacteria, ribosome kinetics are considered rate-limiting for protein synthesis and cell growth. Enhanced ribosome kinetics may augment bacterial growth and biomanufacturing through improvements to overall protein yield, but whether this can be achieved by ribosome-specific modifications remains unknown. Here, we evolve 16S ribosomal RNAs (rRNAs) from Escherichia coli, Pseudomonas aeruginosa, and Vibrio cholerae towards enhanced protein synthesis rates. We find that rRNA sequence origin significantly impacted evolutionary trajectory and generated rRNA mutants with augmented protein synthesis rates in both natural and engineered contexts, including the incorporation of noncanonical amino acids. Moreover, discovered consensus mutations can be ported onto phylogenetically divergent rRNAs, imparting improved translational activities. Finally, we show that increased translation rates in vivo coincide with only moderately reduced translational fidelity, but do not enhance bacterial population growth. Together, these findings provide a versatile platform for development of unnatural ribosomal functions in vivo.