Off-target Cleavage Research Articles

Learning to quantify uncertainty in off-target activity for CRISPR guide RNAs.

CRISPR-based genome editing technologies have revolutionised the field of molecular biology, offering unprecedented opportunities for precise genetic manipulation. However, off-target effects remain a significant challenge, potentially leading to unintended consequences and limiting the applicability of CRISPR-based genome editing technologies in clinical settings. Current literature predominantly focuses on point predictions for off-target activity, which may not fully capture the range of possible outcomes and associated risks. Here, we present crispAI, a neural network architecture-based approach for predicting uncertainty estimates for off-target cleavage activity, providing a more comprehensive risk assessment and facilitating improved decision-making in single guide RNA (sgRNA) design. Our approach makes use of the count noise model Zero Inflated Negative Binomial (ZINB) to model the uncertainty in the off-target cleavage activity data. In addition, we present the first-of-its-kind genome-wide sgRNA efficiency score, crispAI-aggregate, enabling prioritization among sgRNAs with similar point aggregate predictions by providing richer information compared to existing aggregate scores. We show that uncertainty estimates of our approach are calibrated and its predictive performance is superior to the state-of-the-art in silico off-target cleavage activity prediction methods. The tool and the trained models are available at https://github.com/furkanozdenn/crispr-offtarget-uncertainty.

Construction and Stability of All-in-One Adenovirus Vectors Simultaneously Expressing Four and Eight Multiplex Guide RNAs and Cas9 Nickase.

CRISPR/Cas9 technology is expected to offer novel genome editing-related therapies for various diseases. We previously showed that an adenovirus vector (AdV) possessing eight expression units of multiplex guide RNAs (gRNAs) was obtained with no deletion of these units. Here, we attempted to construct "all-in-one" AdVs possessing expression units of four and eight gRNAs with Cas9 nickase, although we expected obstacles to obtain complete all-in-one AdVs. The first expected obstacle was that extremely high copies of viral genomes during replication may cause severe off-target cleavages of host cells and induce homologous recombination. However, surprisingly, four units in the all-in-one AdV genome were maintained completely intact. Second, for the all-in-one AdV containing eight gRNA units, we enlarged the E3 deletion in the vector backbone and shortened the U6 promoter of the gRNA expression units to shorten the AdV genome within the adenovirus packaging limits. The final size of the all-in-one AdV genome containing eight gRNA units still slightly exceeded the reported upper limit. Nevertheless, approximately one-third of the eight units remained intact, even upon preparation for in vivo experiments. Third, the genome editing efficiency unexpectedly decreased upon enlarging the E3 deletion. Our results suggested that complete all-in-one AdVs containing four gRNA units could be obtained if the problem of the low genome editing efficiency is solved, and those containing even eight gRNA units could be obtained if the obstacle of the vector size is also removed.

Non-clinical evaluation of NT-175, an autologous T cell product engineered to express an HLA-A*02:01-restricted TCR targeting TP53 R175H and resistant to TGF-b inhibition.

2560 Background: Adoptive transfer of T-cell receptor (TCR)-engineered T-cells to target shared cancer neoantigens is a promising new immunotherapy approach for patients harboring mutations in tumor suppressor genes such as TP53. TP53 is the most commonly mutated gene across all cancer types, with the R175H mutation being the most prevalent across different indications. NT-175 is an autologous engineered T-cell product expressing an HLA-A*02:01-restricted TCR that specifically targets the TP53 R175H mutation. NT-175 is additionally engineered to lack TGF-β receptor II (TGFBR2) expression, rendering T-cells insensitive to TGF-b-mediated inhibition in the tumor microenvironment. Methods: For manufacturing of NT-175, CD4 and CD8 T-cells are enriched from leukapheresis material, activated in vitro and gene-edited to knock-out both endogenous TGFBR2 and TCR β constant 1/2 (TRBC1/2) and knock-in an HLA-A*02:01-restricted TP53 R175H neoantigen-specific TCR at the TCR α constant (TRAC) locus using CRISPR-Cas9 technology. NT-175 product functionality was evaluated in vivo models and in vitro. T-cell reactivity was assessed by measuring cytotoxicity, proliferation, and cytokine production. The benefit of TGFBR2 knock-out (KO) in the presence of TGF-b was evaluated by measuring phosphorylation of SMAD2/3 proteins, the impact on cell viability and serial killing of target cells. For determination of NT-175 product safety, cross-reactivity and HLA-specificity assessments were carried out. A comprehensive analysis of potential off-target editing by CRISPR/Cas9 was performed to assess potential risk of genotoxicity in clinical grade NT-175. Results: High reactivity of the NT-175 TCR against TP53 R175H and HLA-A*02:01 expressing target cells was observed. T-cell activation and functionality was highly specific, as demonstrated by a lack of reactivity against TP53 WT, minimal cross-reactivity against antigens with up to 4 mismatches to the minimal TP53 R175H encoding epitope recognized by the NT-175 TCR, and lack of T-cell activation in the absence of HLA-A*02:01. In the presence of TGF-b, TGFBR2 KO TCR-edited T-cells displayed inhibition of SMAD2/3 phosphorylation, increased cell viability and increased cytotoxicity and proliferation in serial stimulation assays. In vivo, NT-175 T-cells were able to induce tumor clearance in two independent models. Low frequency chromosomal translocation events (<0.1%) between on-target and off-target Cas9 cleavage sites were detectable in NT-175 T-cells. However importantly, these did not result in autonomous cytokine-independent growth. Conclusions: Non-clinical studies revealed a favorable safety and efficacy profile for NT-175 and supported further clinical development of a TGF-β-resistant TCR-edited T-cell product for mutant TP53-targeted cancer immunotherapy.

Non-clinical evaluation of NT-112, an autologous T cell product engineered to express an HLA-C*08:02-restricted TCR targeting KRAS G12D and resistant to TGF-b inhibition.

e14533 Background: Adoptive transfer of T-cell receptor (TCR)-engineered T-cells to target shared cancer neoantigens is a promising new immunotherapy approach for patients harboring mutations in oncogenes such as KRAS. Activating mutations in KRAS are among the most prevalent oncogenic driver mutations in human cancers, with KRAS G12D being the most frequent KRAS mutation in colorectal carcinoma and pancreatic cancer. NT-112 is an autologous engineered T-cell product expressing an HLA-C*08:02-restricted TCR that specifically targets the KRAS G12D mutation. NT-112 is additionally engineered to lack TGF-β receptor II (TGFBR2) expression, rendering T-cells insensitive to TGF-b-mediated inhibition in the tumor microenvironment. Methods: For manufacturing of NT-112, CD4 and CD8 T-cells are enriched from leukapheresis material, activated in vitro and gene-edited to knock-out both endogenous TGFBR2 and TCR β constant 1/2 (TRBC1/2) and knock-in an HLA-C*08:02-restricted KRS G12D neoantigen-specific TCR at the TCR α constant (TRAC) locus using CRISPR-Cas9 technology. NT-112 product functionality was evaluated in both in vivo models and in vitro co-culture experiments with a panel of cell lines including those endogenously expressing the KRAS G12D mutation. T-cell reactivity was assessed by measuring cytotoxicity, proliferation, and cytokine production. For determination of NT-112 product safety, cross-reactivity and allo-reactivity assessments were carried out. A comprehensive analysis of potential off-target editing by CRISPR/Cas9 was performed to assess potential risk of genotoxicity in clinical grade NT-112. Results: High reactivity of the NT-112 TCR against KRAS G12D and HLA-C*08:02-expressing target cells was observed. T-cell activation and functionality was highly specific, as demonstrated by a lack of reactivity against KRAS WT, minimal cross-reactivity in Alanine and Glycine scans, and lack of T-cell activation in the absence of HLA-C*08:02. In vivo, NT-112 T-cells were able to induce tumor clearance in two independent models. Low frequency chromosomal translocation events (<0.1%) between on-target and off-target Cas9 cleavage sites were detectable in NT-112 T-cells. However importantly, these did not result in autonomous cytokine-independent growth. Conclusions: Non-clinical studies revealed a favorable safety and efficacy profile for NT-112 and supported further clinical development of a TGF-β-resistant TCR-edited T-cell product for mutant KRAS-targeted cancer immunotherapy.

A bioinformatic analysis of gene editing off-target loci altered by common polymorphisms, using ‘PopOff’

ABSTRACT Gene editing therapies are designed to minimise off-target editing. However, it is not widespread practice for common polymorphisms to be considered when identifying potential off-target sites in silico. Nevertheless, genetic variants should be included as they have the potential to alter existing, or to generate new, off-target sites. To facilitate the consideration of common polymorphisms when designing targeted gene therapies we developed PopOff, a web-based tool that integrates minor allele frequencies from the gnomAD variant database into an off-target analysis. We used PopOff to analyse predicted off-target loci from guide RNAs used in four clinical trials and thirty-four research publications. From an analysis of sixty guides, we identified that approximately 20% of off-target loci overlap with a common polymorphism. Of these sites, 6.93% contained variants that reduce the level of mismatch between the off-target locus and guide, and therefore may increase off-target cleavage. In addition, we identified that 0.34% of common polymorphisms generated novel PAM sites, resulting in off-target loci that standard workflows would miss. Our findings demonstrate that common polymorphisms should be considered when designing guides to maximise the safety of CRISPR-based gene therapies. However, this may be problematic in populations where the breadth of genetic diversity remains uncharacterised.

Deciphering the Substrate Specificity Reveals that CRISPR-Cas12a Is a Bifunctional Enzyme with Both Endo- and Exonuclease Activities

The class 2 CRISPR-Cas9 and CRISPR-Cas12a systems, originally described as adaptive immune systems of bacteria and archaea, have emerged as versatile tools for genome-editing, with applications in biotechnology and medicine. However, significantly less is known about their substrate specificity, but such knowledge may provide instructive insights into their off-target cleavage and previously unrecognized mechanism of action. Here, we document that the Acidaminococcus sp. Cas12a (AsCas12a) binds preferentially, and independently of crRNA, to a suite of branched DNA structures, such as the Holliday junction (HJ), replication fork and D-loops, compared with single- or double-stranded DNA, and promotes their degradation. Further, our study revealed that AsCas12a binds to the HJ, specifically at the crossover region, protects it from DNase I cleavage and renders a pair of thymine residues in the HJ homologous core hypersensitive to KMnO4 oxidation, suggesting DNA melting and/or distortion. Notably, these structural changes enabled AsCas12a to resolve HJ into nonligatable intermediates, and subsequently their complete degradation. We further demonstrate that crRNA impedes HJ cleavage by AsCas12a, and that of Lachnospiraceae bacterium Cas12a, without affecting their DNA-binding ability. We identified a separation-of-function variant, which uncouples DNA-binding and DNA cleavage activities of AsCas12a. Importantly, we found robust evidence that AsCas12a endonuclease also has 3′-to-5′ and 5′-to-3′ exonuclease activity, and that these two activities synergistically promote degradation of DNA, yielding di- and mononucleotides. Collectively, this study significantly advances knowledge about the substrate specificity of AsCas12a and provides important insights into the degradation of different types of DNA substrates.

In Silico Modeling of a Mutant Cas9 Protein for High -Fidelity Genome Editing

Introduction: CRISPR/Cas system is a bacterial -acquired immune system against viruses. spCas9 protein -derived from Streptococcus pyogenes is the most common Cas protein used in genome editing. The Cas9 proteins suffer from some problems like unwanted off-target cleavage, which is the major limitation in the application of the CRISPR system, especially for therapeutic issues. The present study aimed to in silico design of a mutant Cas protein to enhance the fidelity of the enzyme. Method: In this study, a new Cas9 protein was designed in silico via the selection of some mutations. The interaction of the mutant and wild -type Cas9 proteins with DNA was determined by molecular docking. Finally, the stability of the mutant and wild -type proteins was evaluated using molecular dynamics. Results: The outcome of this study resulted in designing a mutant Cas9 protein bioinformatically containing 5 amino acid substitutions. Molecular docking scores for the wild -type and mutant proteins were -1073.86 and -349.41, respectively . Conclusion: Molecular docking indicated that the number of hydrogen bonds between protein and DNA reduced in the mutant state. Moreover, based on the molecular dynamic findings, the stability of both mutant and wild -type Cas9 proteins was similar.

Deep learning and CRISPR-Cas13d ortholog discovery for optimized RNA targeting

Effective and precise mammalian transcriptome engineering technologies are needed to accelerate biological discovery and RNA therapeutics. Despite the promise of programmable CRISPR-Cas13 ribonucleases, their utility has been hampered by an incomplete understanding of guide RNA design rules and cellular toxicity resulting from off-target or collateral RNA cleavage. Here, we quantified the performance of over 127,000 RfxCas13d (CasRx) guide RNAs and systematically evaluated seven machine learning models to build a guide efficiency prediction algorithm orthogonally validated across multiple human cell types. Deep learning model interpretation revealed preferred sequence motifs and secondary features for highly efficient guides. We next identified and screened 46 novel Cas13d orthologs, finding that DjCas13d achieves low cellular toxicity and high specificity-even when targeting abundant transcripts in sensitive cell types, including stem cells and neurons. Our Cas13d guide efficiency model was successfully generalized to DjCas13d, illustrating the power of combining machine learning with ortholog discovery to advance RNA targeting in human cells.

Development of a CRISPR-Cas12a system for efficient genome engineering in clostridia.

Clostridium species have gained attention in industrial and medical applications, and the development of genetic tools has enabled the advancement of the CRISPR-Cas systems for these bacteria. Compared to the primarily used Cas9 from Streptococcus pyogenes, the utilization of Cas12a (previously known as Cpf1) proteins remains incomplete in clostridia, although they exhibit potential advantages, including T-rich recognition for Clostridium genomes and lower toxicity. In this study, we expanded the CRISPR-Cas tools in clostridia by establishing a CRISPR-Cas12a system with two different cas12a genes (Ascas12a from Acidaminococcus and Fncas12a from Francisella novicida). The optimized tetracycline-inducible systems were initially determined by the glucuronidase reporter and were used to drive expression of the cas12a genes and crRNAs. Our results demonstrate that the CRISPR-FnCas12a system offers flexible target selection in clostridia, and a specific folding pattern of the precursor crRNA is important to enable high mutation generation efficiency. By using sacB (encoding levansucrase) as a negative marker for plasmid curing and determining the optimal size of the donor DNA template for gene integration in the CRISPR-FnCas12a system, we achieved highly efficient and rapid genome modification, exemplified by the successful engineering of clostridia (Clostridium butyricum and Clostridium sporogenes) to produce near-infrared fluorescence from biliverdin and hemin.IMPORTANCEContinued efforts in developing the CRISPR-Cas systems will further enhance our understanding and utilization of Clostridium species. This study demonstrates the development and application of a genome-engineering tool in two Clostridium strains, Clostridium butyricum and Clostridium sporogenes, which have promising potential as probiotics and oncolytic agents. Particular attention was given to the folding of precursor crRNA and the role of this process in off-target DNA cleavage by Cas12a. The results provide the guidelines necessary for efficient genome engineering using this system in clostridia. Our findings not only expand our fundamental understanding of genome-engineering tools in clostridia but also improve this technology to allow use of its full potential in a plethora of biotechnological applications.

Leveraging QM/MM and Molecular Dynamics Simulations to Decipher the Reaction Mechanism of the Cas9 HNH Domain to Investigate Off-Target Effects.

The clustered regularly interspaced short palindromic repeats (CRISPR) technology is an RNA-guided targeted genome-editing tool using Cas family proteins. Two magnesium-dependent nuclease domains of the Cas9 enzyme, termed HNH and RuvC, are responsible for cleaving the target DNA (t-DNA) and nontarget DNA strands, respectively. The HNH domain is believed to determine the DNA cleavage activity of both endonuclease domains and is sensitive to complementary RNA-DNA base pairing. However, the underlying molecular mechanisms of CRISPR-Cas9, by which it rebukes or accepts mismatches, are poorly understood. Thus, investigation of the structure and dynamics of the catalytic state of Cas9 with either matched or mismatched t-DNA can provide insights into improving its specificity by reducing off-target cleavages. Here, we focus on a recently discovered catalytic-active form of the Streptococcus pyogenes Cas9 (SpCas9) and employ classical molecular dynamics and coupled quantum mechanics/molecular mechanics simulations to study two possible mechanisms of t-DNA cleavage reaction catalyzed by the HNH domain. Moreover, by designing a mismatched t-DNA structure called MM5 (C to G at the fifth position from the protospacer adjacent motif region), the impact of single-guide RNA (sgRNA) and t-DNA complementarity on the catalysis process was investigated. Based on these simulations, our calculated binding affinities, minimum energy paths, and analysis of catalytically important residues provide atomic-level details of the differences between matched and mismatched cleavage reactions. In addition, several residues exhibit significant differences in their catalytic roles for the two studied systems, including K253, K263, R820, K896, and K913.

Predicting CRISPR-Cas12a guide efficiency for targeting using machine learning.

Genome editing through the development of CRISPR (Clustered Regularly Interspaced Short Palindromic Repeat)-Cas technology has revolutionized many fields in biology. Beyond Cas9 nucleases, Cas12a (formerly Cpf1) has emerged as a promising alternative to Cas9 for editing AT-rich genomes. Despite the promises, guide RNA efficiency prediction through computational tools search still lacks accuracy. Through a computational meta-analysis, here we report that Cas12a target and off-target cleavage behavior are a factor of nucleotide bias combined with nucleotide mismatches relative to the protospacer adjacent motif (PAM) site. These features helped to train a Random Forest machine learning model to improve the accuracy by at least 15% over existing algorithms to predict guide RNA efficiency for the Cas12a enzyme. Despite the progresses, our report underscores the need for more representative datasets and further benchmarking to reliably and accurately predict guide RNA efficiency and off-target effects for Cas12a enzymes.

Structural basis for the activation of a compact CRISPR-Cas13 nuclease

The CRISPR-Cas13 ribonucleases have been widely applied for RNA knockdown and transcriptional modulation owing to their high programmability and specificity. However, the large size of Cas13 effectors and their non-specific RNA cleavage upon target activation limit the adeno-associated virus based delivery of Cas13 systems for therapeutic applications. Herein, we report detailed biochemical and structural characterizations of a compact Cas13 (Cas13bt3) suitable for adeno-associated virus delivery. Distinct from many other Cas13 systems, Cas13bt3 cleaves the target and other nonspecific RNA at internal “UC” sites and is activated in a target length-dependent manner. The cryo-electron microscope structure of Cas13bt3 in a fully active state illustrates the structural basis of Cas13bt3 activation. Guided by the structure, we obtain engineered Cas13bt3 variants with minimal off-target cleavage yet maintained target cleavage activities. In conclusion, our biochemical and structural data illustrate a distinct mechanism for Cas13bt3 activation and guide the engineering of Cas13bt3 applications.

A cleavage rule for selection of increased-fidelity SpCas9 variants with high efficiency and no detectable off-targets

Streptococcus pyogenes Cas9 (SpCas9) has been employed as a genome engineering tool with a promising potential within therapeutics. However, its off-target effects present major safety concerns for applications requiring high specificity. Approaches developed to date to mitigate this effect, including any of the increased-fidelity (i.e., high-fidelity) SpCas9 variants, only provide efficient editing on a relatively small fraction of targets without detectable off-targets. Upon addressing this problem, we reveal a rather unexpected cleavability ranking of target sequences, and a cleavage rule that governs the on-target and off-target cleavage of increased-fidelity SpCas9 variants but not that of SpCas9-NG or xCas9. According to this rule, for each target, an optimal variant with matching fidelity must be identified for efficient cleavage without detectable off-target effects. Based on this insight, we develop here an extended set of variants, the CRISPRecise set, with increased fidelity spanning across a wide range, with differences in fidelity small enough to comprise an optimal variant for each target, regardless of its cleavability ranking. We demonstrate efficient editing with maximum specificity even on those targets that have not been possible in previous studies.

Engineered lentivirus-derived nanoparticles (LVNPs) for delivery of CRISPR/Cas ribonucleoprotein complexes supporting base editing, prime editing and in vivo gene modification.

Implementation of therapeutic in vivo gene editing using CRISPR/Cas relies on potent delivery of gene editing tools. Administration of ribonucleoprotein (RNP) complexes consisting of Cas protein and single guide RNA (sgRNA) offers short-lived editing activity and safety advantages over conventional viral and non-viral gene and RNA delivery approaches. By engineering lentivirus-derived nanoparticles (LVNPs) to facilitate RNP delivery, we demonstrate effective administration of SpCas9 as well as SpCas9-derived base and prime editors (BE/PE) leading to gene editing in recipient cells. Unique Gag/GagPol protein fusion strategies facilitate RNP packaging in LVNPs, and refinement of LVNP stoichiometry supports optimized LVNP yield and incorporation of therapeutic payload. We demonstrate near instantaneous target DNA cleavage and complete RNP turnover within 4 days. As a result, LVNPs provide high on-target DNA cleavage and lower levels of off-target cleavage activity compared to standard RNP nucleofection in cultured cells. LVNPs accommodate BE/sgRNA and PE/epegRNA RNPs leading to base editing with reduced bystander editing and prime editing without detectable indel formation. Notably, in the mouse eye, we provide the first proof-of-concept for LVNP-directed in vivo gene disruption. Our findings establish LVNPs as promising vehicles for delivery of RNPs facilitating donor-free base and prime editing without formation of double-stranded DNA breaks.

Applications of CRISPR Cas-9 in Ovarian Cancer Research.

Ovarian cancer is a highly prevalent malignancy among women and affects a significant population worldwide. Different forms of hormonal treatments or chemotherapies are used to treat ovarian cancer, but the possible side effects, including menopausal symptoms, can be severe, forcing some patients to prematurely stop the treatment. The emerging genome editing technology, known as clustered regularly interspaced short palindromic repeats (CRISPR)-caspase 9 (Cas9), has the potential to treat ovarian cancer via gene editing strategies. Studies have reported CRISPR knockouts of several oncogenes that are involved in the pathogenesis of ovarian cancer, such as BMI1, CXCR2, MTF1, miR-21, and BIRC5, and demonstrate the potential of the CRISPR-Cas9 genome editing technique to effectively treat ovarian cancer. However, there are limitations that restrict the biomedical applications of CRISPR-Cas9 and limit the implementation of Gene therapy for ovarian cancer. These include off-target DNA cleavage and the effects of CRISPR-Cas9 in non-target, normal cells. This article aims to review the current state of ovarian cancer research, highlight the significance of CRISPR-Cas9 in ovarian cancer treatment, and establish the groundwork for potential clinical research.

In vivo CRISPR gene editing in patients with herpetic stromal keratitis

Invivo CRISPR gene therapy holds large clinical potential, but the safety and efficacy remain largely unknown. Here, we injected a single dose of herpes simplex virus 1 (HSV-1)-targeting CRISPR formulation in the cornea of three patients with severe refractory herpetic stromal keratitis (HSK) during corneal transplantation. Our study is an investigator-initiated, open-label, single-arm, non-randomized interventional trial at a single center (NCT04560790). We found neither detectable CRISPR-induced off-target cleavages by GUIDE-seq nor systemic adverse events for 18months on average in all three patients. The HSV-1 remained undetectable during the study. Our preliminary clinical results suggest that invivo gene editing targeting the HSV-1 genome holds acceptable safety as a potential therapy for HSK.

Deep sampling of gRNA in the human genome and deep-learning-informed prediction of gRNA activities

Life science studies involving clustered regularly interspaced short palindromic repeat (CRISPR) editing generally apply the best-performing guide RNA (gRNA) for a gene of interest. Computational models are combined with massive experimental quantification on synthetic gRNA-target libraries to accurately predict gRNA activity and mutational patterns. However, the measurements are inconsistent between studies due to differences in the designs of the gRNA-target pair constructs, and there has not yet been an integrated investigation that concurrently focuses on multiple facets of gRNA capacity. In this study, we analyzed the DNA double-strand break (DSB)-induced repair outcomes and measured SpCas9/gRNA activities at both matched and mismatched locations using 926,476 gRNAs covering 19,111 protein-coding genes and 20,268 non-coding genes. We developed machine learning models to forecast the on-target cleavage efficiency (AIdit_ON), off-target cleavage specificity (AIdit_OFF), and mutational profiles (AIdit_DSB) of SpCas9/gRNA from a uniformly collected and processed dataset by deep sampling and massively quantifying gRNA capabilities in K562 cells. Each of these models exhibited superlative performance in predicting SpCas9/gRNA activities on independent datasets when benchmarked with previous models. A previous unknown parameter was also empirically determined regarding the “sweet spot” in the size of datasets used to establish an effective model to predict gRNA capabilities at a manageable experimental scale. In addition, we observed cell type-specific mutational profiles and were able to link nucleotidylexotransferase as the key factor driving these outcomes. These massive datasets and deep learning algorithms have been implemented into the user-friendly web service http://crispr-aidit.com to evaluate and rank gRNAs for life science studies.

PiCRISPR: Physically informed deep learning models for CRISPR/Cas9 off-target cleavage prediction

CRISPR/Cas programmable nuclease systems have become ubiquitous in the field of gene editing. With progressing development, applications in in vivo therapeutic gene editing are increasingly within reach, yet limited by possible adverse side effects from unwanted edits. Recent years have thus seen continuous development of off-target prediction algorithms trained on in vitro cleavage assay data gained from immortalised cell lines. It has been shown that in contrast to experimental epigenetic features, computed physically informed features are so far underutilised despite bearing considerably larger correlation with cleavage activity. Here, we implement state-of-the-art deep learning algorithms and feature encodings for off-target prediction with emphasis on physically informed features that capture the biological environment of the cleavage site, hence terming our approach piCRISPR. Features were gained from the large, diverse crisprSQL off-target cleavage dataset. We find that our best-performing models highlight the importance of sequence context and chromatin accessibility for cleavage prediction and compare favourably with literature standard prediction performance. We further show that our novel, environmentally sensitive features are crucial to accurate prediction on sequence-identical locus pairs, making them highly relevant for clinical guide design. The source code and trained models can be found ready to use at github.com/florianst/picrispr.

CasKAS: direct profiling of genome-wide dCas9 and Cas9 specificity using ssDNA mapping

Detecting and mitigating off-target activity is critical to the practical application of CRISPR-mediated genome and epigenome editing. While numerous methods have been developed to map Cas9 binding specificity genome-wide, they are generally time-consuming and/or expensive, and not applicable to catalytically dead CRISPR enzymes. We have developed CasKAS, a rapid, inexpensive, and facile assay for identifying off-target CRISPR enzyme binding and cleavage by chemically mapping the unwound single-stranded DNA structures formed upon binding of a sgRNA-loaded Cas9 protein. We demonstrate this method in both in vitro and in vivo contexts.

Cas9 off-target binding to the promoter of bacterial genes leads to silencing and toxicity.

Genetic tools derived from the Cas9 RNA-guided nuclease are providing essential capabilities to study and engineer bacteria. While the importance of off-target effects was noted early in Cas9's application to mammalian cells, off-target cleavage by Cas9 in bacterial genomes is easily avoided due to their smaller size. Despite this, several studies have reported experimental setups in which Cas9 expression was toxic, even when using the catalytic dead variant of Cas9 (dCas9). Specifically, dCas9 was shown to be toxic when in complex with guide RNAs sharing specific PAM (protospacer adjacent motif)-proximal sequence motifs. Here, we demonstrate that this toxicity is caused by off-target binding of Cas9 to the promoter of essential genes, with silencing of off-target genes occurring with as little as 4 nt of identity in the PAM-proximal sequence. Screens performed in various strains of Escherichiacoli and other enterobacteria show that the nature of toxic guide RNAs changes together with the evolution of sequences at off-target positions. These results highlight the potential for Cas9 to bind to hundreds of off-target positions in bacterial genomes, leading to undesired effects. This phenomenon must be considered in the design and interpretation of CRISPR-Cas experiments in bacteria.