Predicting and Improving Insertions by Prime Editing

Abstract

GEN BiotechnologyVol. 2, No. 3 Views & NewsFree AccessPredicting and Improving Insertions by Prime EditingMandana ArbabMandana Arbab*Address correspondence to: Mandana Arbab, Rosamund Stone Zander Translational Neuroscience Center, F. M. Kirby Neurobiology Center, Boston Children's Hospital, Department of Neurology, Harvard Medical School, Boston, MA 02115, USA, E-mail Address: Mandana.Arbab@childrens.harvard.eduRosamund Stone Zander Translational Neuroscience Center, F. M. Kirby Neurobiology Center, Boston Children's Hospital, Department of Neurology, Harvard Medical School, Boston, Massachusetts, USA.Search for more papers by this authorPublished Online:19 Jun 2023https://doi.org/10.1089/genbio.2023.29096.marAboutSectionsPDF/EPUB Permissions & CitationsPermissionsDownload CitationsTrack CitationsAdd to favorites Back To Publication ShareShare onFacebookTwitterLinked InRedditEmail A meticulous new study conducted by researchers at the Wellcome Sanger Institute reveals important regulators of prime editing insertion efficiency including key enzymes and secondary structures.Prime editing is a versatile genome editing technology that can precisely introduce a wide array of modifications into the genome, including every base-to-base conversion and small insertions and deletions, with minimal undesired byproducts and without requiring double-strand DNA breaks.1 Inserting short DNA sequences into the genome has vast utility in biotechnology and holds great potential for the development of gene-based therapeutics to treat genetic diseases, the majority of which are caused by small indels and nucleotide changes.2,3 However, the efficiency of prime edits is subject to many parameters that require optimization in ways that are largely unintuitive, even for expert users.An interesting recent article by Koeppel et al. at the Wellcome Sanger Institute and colleagues in Estonia, led by Leopold Parts, investigates the factors that govern prime editing-mediated insertions of small sequences and provides practical guidelines for prime editing single-guide (sg) RNA (pegRNA) design to improve the efficiency of these edits.4CRISPR genome editing has revolutionized our understanding and manipulation of the genome, helping to shed light on the expression of genes, the function of gene products, and enabling modifications of protein activity. First generation CRISPR technologies have had a major impact in the life sciences, mainly by their ability to induce deletions that facilitate studies of protein loss-of-function. Base editors, which use a modified Cas protein fused to a nucleotide deaminase enzyme, improved the precision of genome editing tools by enabling the conversion of single nucleotides in the genome without requiring the generation of double-strand DNA breaks, and have the potential to correct mutations that underlie the majority of genetic diseases. These CRISPR-based technologies enable the loss or change of genetic sequences but cannot facilitate the precise insertion of sequences in the genome without the use of a donor template, which is limiting for many research and therapeutic applications. The insertion of short DNA sequences including reporters and purification tags has helped to improve our understanding of gene expression and cellular protein interactions, and can be used for a wide range of applications such as protein engineering to introduce exogenous functional domains that alter protein activity, introducing neoantigens into cells, modifying gene regulatory sequences, or the correcting pathogenic sequences such as the F508 deletion in the CF transmembrane conductance regulator (CFTR) gene, that underlies the majority of cystic fibrosis cases.5,6Prime editing enables the programmable insertion of short sequences without requiring a separate donor template or the generation of double-strand DNA breaks.1 This enables more efficient and precise targeting in a broader range of cell types than was possible with the prior state-of-the-art technology that relied on homology-directed repair mediated by programmable nucleases to drive insertions.7Prime editors are composed of a reverse transcriptase that is fused to a Cas9-nickase to induce a nick at the targeted DNA locus and release a 3′-DNA flap.1 Prime editors use a modified sgRNA, the pegRNA, with a 3′-extension that encodes a primer binding site (PBS) that is complementary to a portion of the nicked strand. The pegRNA furthermore includes a reverse transcription (RT) template that encodes the intended edit and a homology arm (HA) to enable the permanent integration of the edited strand into the genome.The pegRNA:DNA heteroduplex formed between the PBS and nicked DNA strand serves as a primer template for RT by the prime editor, which polymerizes a 3′ DNA flap off the RT template that contains the edit and HA. This flap is subsequently incorporated at the target site endogenous DNA repair mechanisms, which vary depending on the nature of the edit.Some aspects of the prime editing system can be enhanced to increase desired product formation. For instance, codelivery of an sgRNA to nick the DNA strand that opposes the 3′-flap stimulates synthesis of the edited sequence into the nonedited strand, and thereby increases prime editing efficiency. Increasing the abundance of pegRNAs and prime editing components in cells by sequence or peptide modifications that enhance their production or stability also improves prime editing at many sites.8,9 In addition, downregulating the mismatch repair (MMR) DNA damage response, which recognizes and excises small mismatches in the genome, can improve the incorporation of short sequences introduced by prime editing.10,11Nevertheless, the efficiency of prime edits is mainly determined by the target locus and sequence features of the pegRNA and target site. Optimizing pegRNA design still requires the cumbersome and costly experimental validation of PBS and HA length permutations across many sgRNA spacers at the target locus. A deeper understanding of how pegRNA features affect prime edits could improve the workflow and outcomes of prime editing experiments by enabling a priori identification of local optimum pegRNA designs.Large-scale libraries of paired sgRNA–target sequences can provide great insight into the determinants of CRISPR-mediated genome editing outcomes, and have informed computational models such as inDelphi and BE-Hive that predict genome editing outcomes for CRISPR nucleases and base editors.12,13 These models facilitate the in silico design of genome editing strategies to reduce the burden of empirical evaluation.14–16 Similar large-scale libraries of paired pegRNAs and their target sites have identified general features of pegRNAs that are important determinants of prime editing efficiency, including sgRNA-dependent SpCas9 activity, GC content, and standard PBS and RT template lengths. These data have been used to train algorithms such as DeepPE, PRIDICT, and Easy-Prime that assist in pegRNA design.17–19And yet, due to the complexity of prime editing and the wide range of edits that it enables, a single library that is designed to investigate all classes of prime edits (insertions, deletions, and base changes) simultaneously is likely to underrepresent the breadth of sequences necessary to fully characterize the pegRNA features that determine their efficiency. Moreover, different classes of edits are handled differently by endogenous DNA repair machinery, thus some pegRNA features might affect one class of edit differently than another and such findings may be underpowered in broad-scale pegRNA libraries. Measuring prime editing activity across a wider range of sequences may uncover pegRNA features that are of particular importance to some classes of edits and increase the accuracy of predictive models to improve prime editing strategy design.Prime FactorsIn this new study, Koeppel et al. specifically focus on identifying factors that affect prime editing-mediated insertion of small sequences.4 The authors designed a library of >3,600 pegRNAs encoding various types of insertions—including reporters and purification tags and all 1–4 nt changes—with a fixed PBS and HA length, and identified pegRNA sequence features that specifically affect insertion efficiency. Prior studies have identified insertion size as a strong negative correlate of editing efficiency and found that shorter RT templates (10–12 nt) are typically optimal.1,17,19 Together, these findings suggest that balancing the size of a prime-editing insertion and HA sequence can improve editing outcomes. However, these studies were mainly performed in cells that are deficient in MMR and, therefore, overrepresent the insertion frequency of small edits compared with MMR proficient cell types.This study shows that in MMR proficient cells, the insertion frequency of short sequences (1–4 nt) is not substantially greater than that of longer sequences, and that insertion size, rather than overall flap length, impacts insertion frequency. The authors then used their data to train a machine learning model, MinsePIE, to accurately predict prime editing insertion efficiencies and facilitate pegRNA design to increase these outcomes by 1.4-fold on average.The findings of this study enable more than the computational identification of local optimum PBS and HA lengths to facilitate pegRNA design. They uncover the Three Prime Repair Exonuclease 1 and 2 genes (TREX1/2) as regulators of prime editing insertion efficiency that particularly affect the genetic incorporation of long insertions. Moreover, the authors demonstrate that greater secondary structure of the RT template increases prime editing efficiency, without affecting pegRNA stability in the cell.Prior studies have shown that a structured 3′-end of the pegRNA avoids exonuclease activity that reduces pegRNA abundance, and that encoding synonymous sequence changes into the RT template can increase prime editing outcomes by avoiding MMR repair.8,10,20 In this study, the authors demonstrate that a structured RT template encodes a more structured 3′-DNA flap at the target locus that suppresses TREX1/2-mediated exonuclease activity on the prime edited DNA (Fig. 1), and their computational tool facilitates the design of more structured RT templates by recoding with synonymous codons to improve prime editing outcomes. These RT template optimizations can be combined with 3′ engineered pegRNAs (epegRNAs)8 to further increase the frequency of prime editing insertions.FIG. 1. Three Prime Repair Exonuclease 1 and 2 (TREX1/2) are regulators of prime editing insertion efficiency.Long prime edited 3’-DNA flaps undergo TREX1/2-mediated degradation (above). Increased secondary structure of the prime edited insertion protects the 3’-DNA flap from degradation and improves the efficiency of prime editing insertions (below). (Adapted from Koeppel et al.4).This study illustrates how paired pegRNA–target libraries dedicated to a single class of edit can improve our understanding and application of prime editing. The identification of TREX1/2 as key regulators of long-insertion efficiency underscores how DNA repair pathways can differentially modulate some prime editing outcomes. Combining gene activation and inhibition functional screening with pegRNA–target libraries focused on specific classes of prime edits could allow for the unbiased detection of other actionable DNA repair pathways to improve prime editing outcomes.10Future studies may also focus on technical improvements to the assay design; the current library included different types of inserts up to 69 nt but used few spacers, and unequal representation of library members due to biases in molecular cloning and lentiviral vector recombination reduces the power to calculate potentially subtle but important effects of some sequences. Many recent advances in prime editing leverage small insertions for a multitude of applications, including the insertion of recombinase sites to enable the incorporation of large DNA sequences into the genome, which is the next major frontier in precision genome editing and therapeutics development.21,22 The present study contributes to the ongoing improvement of these technologies that are poised overcome the outstanding challenges of the genome editing field.