Development and Application of Prime Editing in Plants

LIU Tingting, ZOU Jinpeng, YANG Xi, WANG Kejian, RAO Yuchun, WANG Chun

(1State Key Laboratory of Rice Biology and Breeding, China National Rice Research Institute, Hangzhou 310006, China; 2College of Life Science and Technology, Huazhong Agricultural University, Wuhan 430070, China; 3College of Life Sciences, Zhejiang Normal University, Jinhua 321004, China; #These authors contributed equally to this work)

Abstract: Clustered regularly interspaced palindromic repeats (CRISPR)/CRISPR-associated protein(Cas)-mediated genome editing has greatly accelerated progress in plant genetic research and agricultural breeding by enabling targeted genomic modifications. Moreover, the prime editing system,derived from the CRISPR/Cas system, has opened the door for even more precise genome editing.Prime editing has the capability to facilitate all 12 types of base-to-base conversions, as well as desired insertions or deletions of fragments, without inducing double-strand breaks and requiring donor DNA templet. In a short time, prime editing has been rapidly verified as functional in various plants, and can be used in plant genome functional analysis as well as precision breeding of crops. In this review, we summarize the emergence and development of prime editing, highlight recent advances in improving its efficiency in plants, introduce the current applications of prime editing in plants, and look forward to future prospects for utilizing prime editing in genetic improvement and precision molecular breeding.

Key words: prime editing; CRISPR/Cas; precision genome editing; crop breeding

Genetic variation is fundamental to crop breeding. The history of crop breeding mainly involves the discovery and screening of natural genetic resources. In addition,radiation and chemical mutagenesis are employed to accelerate the acquisition of genetic germplasm resources. However, introducing elite alleles into different varieties by crossing and screening is a time-consuming and labor-intensive process (Rukmini et al, 2021). Although transgenic technology has been developed to quickly transfer endogenous or exogenous genes into plants, which greatly speeds up the breeding process, the public still has serious doubts about the safety of transgenic crops.

Owing to the discovery of sequence-specific nucleases(SSNs), a new generation of genome engineering technology, gene editing, has emerged to make precise changes in crop genes. Compared with the previous breeding strategies, gene-editing technology greatly shortens the breeding cycle and independently changes a single gene or an agronomic trait without involving others (Gao, 2021). In addition, it does not introduce foreign genes, making it more secure. As a result, crop breeders can utilize gene-editing tools to improve plant genetic traits, including increasing yield and enhancing resistance to stress, herbicides and pests.

SSNs induce DNA double-strand breaks (DSBs) at target sites, causing mutations in target genes via error-prone non-homologous end joining (NHEJ) or donor-dependent homology-directed repair (HDR)(Zhu et al, 2020). The developed SSN technologies include the meganuclease/LAGLIDADG homing endonuclease (MN/LHE) system (Chevalier and Stoddard, 2001; Bochtler, 2016), the zinc-finger nuclease (ZFN) system (Bibikova et al, 2003), the transcription activator-like effector nuclease (TALEN)system (Boch et al, 2009; Moscou and Bogdanove,2009), and the clustered regularly interspaced short palindromic repeats (CRISPR)/CRISPR-associated protein (Cas) system (Jinek et al, 2012; Cong et al,2013; Mali et al, 2013). In contrast to other SSN technologies that depend on synthesizing specific binding proteins via DNA-protein interactions to recognize target sequences, the CRISPR/Cas system recognizes target sites via RNA, relying on DNARNA interactions for target recognition (Kosicki et al,2018). This characteristic of the CRISPR/Cas system simplifies genome editing and makes it highly effective, resulting in its wide use in genetic research and crop breeding (Távora et al, 2022). Furthermore,the CRISPR/Cas9 system, which consists of the nuclease Cas9 and artificially designed single-guide RNAs (sgRNAs), is the most commonly used.However, the majority of these gene-editing products insert or delete small nucleotides through the NHEJ pathway rather than achieving precise insertions through the HDR pathway, which is so inefficient as to be nearly impossible to apply for precise editing (Ira et al, 2004; Al-Zain and Symington, 2021).

The difficulties inherent in gene editing using CRISPR/Cas limit the capability to improve genetic traits, most of which require targeted gene modification rather than gene disruption, thus stimulating the development of other techniques to mediate precise gene editing (Cullot et al, 2019). The base editors(BEs), which combine the CRISPR/Cas system with different deaminases, have been developed to convert bases at target nucleotides without forming DSBs(Chen and Liu, 2023). The cytosine base editor (CBE)catalyzes cytosine to thymine (C-to-T) conversion using cytosine deaminases like APOBEC1 or AID(Beale et al, 2004; Conticello, 2008), and the adenine base editor (ABE) catalyzes adenine to guanine (A-to-G) conversion using an artificially modified adenosine enzyme (Gaudelli et al, 2017). Subsequently, the glycosylase base editor (CGBE or GBE) has emerged to efficiently induce targeted C-to-G and C-to-A base transversions, building upon the foundation of CBE(Kurt et al, 2021), and precise substitutions have been achieved in rice (Kurt et al, 2021; Tian et al, 2022).The systems of CBE, ABE, and GBE are advanced and have been applied in plants, partially addressing the need for site-directed mutations and directed evolution. However, BEs are limited in their ability to perform specific types of base conversions, have a narrow editing window, and are constrained in their genome editing range (Grünewald et al, 2019; Zuo et al,2019). These features make it challenging for BEs to perform accurate editing at all genomic positions.

To further expand the scope of precise editing and overcome the inefficiency of CRISPR/Cas9 systemmediated HDR and the inability of BE systems to achieve base reversal, prime editing (PE) was developed.PE is a ‘search-and-replace’ genome editing technology that precisely mediates all 12 types of precise base-tobase conversions and desired InDels (Anzalone et al,2019). As a result, PE can advance the research and correction of most genes by enabling precisely targeted InDels or combination edits without inducing DSBs or requiring donor DNA templates (Hao et al,2021). The PE system is rapidly being applied to plants, developing flexible and versatile precision gene-editing tools and providing a new path for plant genome editing. In this review, we summarized the emergence and development of PE, highlighted recent advances in improving PE efficiency in plants,introduced its current plant-related applications, and looked forward to the future possibilities for using PE in genetic improvement and precision molecular breeding of plants.

Emergence and development of PE

In 2019, Anzalone et al (2019) developed the PE system, which has the capacity to induce all 12 base-to-base conversions and produce accurate small DNA insertions or deletions in mammalian cells. The original prime editor 1 (PE1) system consists of two components: prime editor (nCas9-MMLV) and PE guide RNA (pegRNA). The nCas9-MMLV is a fusion protein formed by the Cas9 nickase (nCas9) (H840A)and the wild type reverse transcriptase from Moloney murine leukemia virus (MMLV). The pegRNA is an engineered sgRNA with an added primer binding site(PBS) sequence for directing nCas9-MMLV to the genomic target site and a reverse transcriptase template (RTT) sequence for mediating the desired edit at the 3′-terminus of the sgRNA (Fig. 1-A).

The PE1 system follows a comprehensive workflow to achieve precise genome editing. First, the nCas9-MMLV-pegRNA complex recognizes the target site and cleaves the desired strand to induce a nick at the third base upstream of the protospacer adjacent motif(PAM), releasing a 3′-DNA end (Fig. 1-B). Next, the PBS of the pegRNA extension can hybridize to the 3′-DNA end, and the MMLV reverse transcriptase participates in the reverse transcription of new singlestranded DNA (ssDNA) using the RTT of the pegRNA as a template (Fig. 1-B). The ssDNA generates a 3′-flap that contains the edited sequence to compete with the original 5′-flap, and excision of the displaced 5′-flap followed by ligation of the remaining nick results in a heteroduplex (Fig. 1-C). Finally, DNA repair or replication copies the edited sequence onto the complementary strand, achieving editing of both strands (Fig. 1-C) (Anzalone et al, 2019).

Fig. 1. Composition and mechanism of prime editing (PE) system.

However, the editing efficiency of the original PE1 is modest, and most targets have less than 5%precision editing efficiency (Anzalone et al, 2019). To optimize the PE1 system, prime editor 2 (PE2) and prime editor 3 (PE3 and PE3b) systems are generated(Fig. 1-C). In the PE2 system, three mutations(D200N/L603W/T330P) are introduced to the wild type MMLV to increase its thermostability, and other two mutations (T306K/W313F) are introduced to increase its reverse transcription process (Fig. 2-A).The editing efficiency achieved by the PE2 system is,on average, three times higher than that achieved by the PE1 system in mammalian cells. Building on the PE2 system, the PE3 system is developed by introducing an additional sgRNA to create another nick on the non-edited strand, 40-100 bp downstream of the pegRNA-induced nick (Fig. 2-A). As a result,the PE3 system increases editing efficiency 3-fold compared with PE2 in mammalian cells. When the additional sgRNA is complementary to the pegRNAinduced edited sequence, the system is referred to as PE3b, which prevents nicking of the non-edited strand prior to the completion of the editing process (Fig.2-A). The editing efficiency of the PE3b system is slightly higher than that of PE3, and the PE3b system significantly reduces InDel formation compared with the PE3 system in mammalian cells (Anzalone et al,2019; Chen and Liu, 2023).

Fig. 2. Emergence and development of prime editing (PE) systems in plants.

Similar to the PE system in mammalian cells, the plant prime editing (PPE) system consists of plant codon-optimized nCas9-MMLV and pegRNA. Lin et al(2020) established plant prime editor 2 (PPE2) and plant prime editor 3 (PPE3 and PPE3b) systems in rice and wheat protoplasts, verifying that the PPE system can introduce all types of base substitutions and small DNA insertions or deletions. Tang et al (2020) also proved the feasibility of PPE2 and PPE3 systems in rice protoplasts. Subsequently, different researchers tested the editing efficiency of the PPE system in transgenic T0plants. Lin et al (2020) used the PPE3 system to validate base substitutions and small DNA deletions at three sites in transgenic T0rice, with efficiencies ranging from 2.6% to 21.8%. Xu W et al(2020) also used the PPE3 system to validate editing efficiency at six different sites in transgenic T0rice and found that theOsACC-2site has no editing efficiency, while theOsALS-2site shows the highest editing efficiency of 26%. However, editing efficiencies at the other four sites averaged approximately 4%.Hua et al (2020) investigated the ability of the PE2 and PE3 systems to repair a non-functional green fluorescent protein (GFP) site through precise editing and found that both systems achieve an editing efficiency of over 15% for repairing the non-functional GFP. Nevertheless, when using the PE3 system to edit endogenous targets in transgenic T0rice, only theALSsite yields transgenic plants with base substitutions at an efficiency of 9.1%, and other sites do not yield precisely edited plants. Xu R F et al (2020) established an efficient PPE2 system in rice and applied it to six different types of editing on thePDSgene. The research findings showed that editing efficiencies vary from 0% to 31.3%, indicating that the structure of the pegRNA significantly influences PE editing efficiency in rice. Furthermore, they compared the editing efficiency difference between the optimized PPE3 and PPE3b systems with the PPE2 system in transgenic T0rice, and found that the editing efficiencies of the PPE3 and PPE3b systems are not higher than those of the PPE2 system. Butt et al (2020) also established the PPE2 and PPE3 systems to achieve precise editing ofALSgene and found no difference in editing efficiency between these two systems. To improve the efficiency of target editing, Li H Y et al (2020) developed a co-editing strategy using the PPE3 system and successfully edited the exogenoushptIImutant and endogenousOsEPSPSin transgenic T0rice with an editing efficiency of 9.38%. Jiang et al (2020)established efficient PPE3 and PPE3b systems in maize and successfully achieved homozygous double mutations (W542L and S621I) inZmALS1andZmALS2. Compared with the previous results in rice,they obtained higher editing efficiency in maize (up to about 72% to 75%), providing a possible method for further improving PE editing efficiency in rice.Additionally, Lu et al (2021) established the PPE3 system in dicotyledonous tomato plants by optimizing codons and promoters, achieving precise editing of theALS2andPDS1genes. However, the editing efficiencies were only 6.7% and 3.4%, respectively. Overall, the editing efficiency of the PPE system in plants is currently not at a level suitable for practical applications(Fig. 2-B). Therefore, further improvements are urgently needed to enhance the potential use of PPE in breeding.

Improvements of PPE

The feasibility of PPE has been demonstrated, but the editing efficiency of early PPE versions is too low,limiting their applications in crop breeding. Therefore,the most critical issue is improving the precise editing efficiency of existing PPE versions. Research on mammalian cells and plants has shown that various strategies are effective in improving PE editing efficiency.In the following sections, we will describe the latest research on key strategies to enhance PPE editing efficiency.

Optimization of prime editor (nCas9-MMLV)

Enhancing nCas9 activity

nCas9 has been employed as part of the prime editor due to its ability to cut ssDNA. Therefore, PE editing efficiency can be improved by enhancing Cas9 nuclease activity. Chen et al (2021) engineered an enhanced PE system, PEmax, by utilizing a variant of nCas9 (R221K/N394K/H840A) with stronger singlestrand cleavage capability and additional nuclear localization signal (C-terminal c-Myc NLS) tags to enhance nuclear localization in mammalian cells (Fig.3-A). With a similar strategy, Li J et al (2022) found that the plant PEmax (PPEmax) system can increase precision editing efficiency over the original PPE2 by up to 2.8-fold, on average, at four sites in rice. Jiang Y Y et al (2022) tested the editing efficiency of PPE at 20 sites in rice protoplasts and found that the PPEmax system has a higher editing efficiency compared with the original PPE2/3 systems. Meanwhile, the PPEmax system has also been applied to maize, resulting in the generation of herbicide-resistant mutant plants (Qiao et al, 2023).

Enhancing MMLV activity

MMLV is another essential component of the prime editor, achieving precise gene editing by synthesizing a new DNA sequence based on the unpaired pegRNA sequence and integrating it into the DNA. Therefore,enhancing the activity of MMLV is crucial for improving PE efficiency. The highest editing efficiency of the original PE1, using the wild type MMLV, is only 5%in mammalian cells (Anzalone et al, 2019). To enhance the activity of MMLV, Anzalone et al (2019)engineered the MMLV (D200N/L603W/T330P/T306K/W313F) in PE1 through mutagenesis to improve its thermostability and reverse transcription process,which leads to the development of the PE2 system(Fig. 3-A). On average, the PE2 system achieves a 3-fold improvement in editing efficiency compared with PE1. This engineered MMLV is widely used in mammalian cells and plants due to its high activity. In plants, Zong et al (2022) employed an ePPE strategy to further improve the activity of MMLV by removing the ribonuclease H domain (Rnase H) from engineered MMLV and inserting a viral nucleocapsid protein between the nCas9 and engineered MMLV. This ePPE system increases editing efficiency by an average of 5.8-fold in rice (Fig. 3-A). Moreover, in mammalian cells, Velimirovic et al (2022) fused amino acid peptides IGFpm1-NFATC2IPp1 (IN) to the N-terminus of nCas9, enhancing translation and forming the INPE2 system, which significantly increases prime editor activity and thus improves PE efficiency. Song et al (2021) fused hRad51-ssDBD, a single-stranded DNA-binding domain of human Rad51, between nCas9 and MMLV, forming the hyperactive PE2(phyPE2) system, which significantly improves PE editing efficiency. However, Li J et al (2022) found that both strategies have no effect on rice calli.

Optimization of pegRNA

Designing suitable pegRNA

The ‘search-and-replace’ function of PE is based on pegRNA, which includes well-known sgRNAs as well as PBS and RTT sequences located at its 3′-end(Anzalone et al, 2019) (Fig. 3-B). Therefore, designing an efficient pegRNA is crucial for ensuring the stability of PE efficiency. To address the difficulty of manually designing pegRNA, efficient software tools have been developed for both animals and plants. For mammalian cells, pegFinder, pegIT, PrimeDesign, and PnB Designer have been developed based on the unique structure of pegRNA using different algorithms(Chow et al, 2021; Hsu et al, 2021; Siegner et al, 2021;Standage-Beier et al, 2021). In plants, PlantPeg-Designer is developed using the melting temperature(Tm) of the PBS sequence as a basis. As a result, PPEs are most efficient when PBSTmapproaches 30 °C(Lin et al, 2021). Overall, the development of these tools has simplified the process of pegRNA design and improves its accuracy.

Improving pegRNA expression

Fig. 3. Overview of prime editor (PE) system optimization in plants.

The longer expression cassettes of pegRNA and the possibility of forming double-stranded structures between PBS and spacer sequences can inhibit pegRNA expression. Therefore, increasing the expression level of pegRNA is a crucial strategy for improving PE efficiency. Currently, strategies to enhance pegRNA expression have only shown significant effects in plants, with no relevant reports found in mammalian cells (Jiang et al, 2020). Jiang et al (2020) used two strategies to enhance pegRNA expression in maize through the PPE3 system: using aCaMV35SCmYLCV-U6composite promoter to drive pegRNA and doubling the number of pegRNA cassettes. The results showed that using the composite promoter significantly improves PE efficiency from 0% to 43.8%, while redoubling the number of pegRNA cassettes does not show a significant effect on PE efficiency. Furthermore, Qiao et al (2023) later found that redoubling the number of pegRNA cassettes contributes to PE efficiency through the PPEmax system.Meanwhile, Li J et al (2022) also found that replacing theOsU3promoter with theCaMV35S-CmYLCV-U6composite promoter significantly increases PE efficiency in rice (Fig. 3-B). The above results indicate that higher expression levels of pegRNA are highly useful for PE efficiency.

Maintaining structural stability of pegRNA

The structural design of pegRNA leads to the 3′-extension of pegRNA being likely to be exposed in cells and more susceptible to degradation by exonucleases,resulting in its instability (Nelson et al, 2022). To address this issue, Nelson et al (2022) incorporated two RNA structural motifs, a modified prequeosine1-1 riboswitch aptamer (evopreQ1) and a frameshifting pseudoknot from Moloney murine leukemia virus, to the 3′-terminus of pegRNA to prevent its degradation.The engineered pegRNA enhances PE efficiency by 3-4-fold compared with canonical pegRNA in mammalian cells. Currently, the RNA structural motif evopreQ1has been widely employed in plants to maintain the stability of pegRNA structure. In both rice and maize, evopreQ1significantly enhances PE efficiency (Zou et al, 2022; Qiao et al, 2023) (Fig.3-B). Meanwhile, to maintain the stability of the pegRNA secondary structure, Chai et al (2021)applied the MS2 RNA aptamer to the PPE system by adding MS2 RNA to the 3′-extension of pegRNA and fusing the MS2 coat protein with MMLV to express it together with nCas9, creating the MS2PE system (Fig.3-B). The results showed that, compared with the original PPE system, the editing efficiencies for five target sites using the MS2PE system increase by 1.2-10.1 times in rice. In addition, other structures have been developed to maintain pegRNA stability in mammalian cells, such as the Zika virus exoribonucleaseresistant RNA motif (xr-pegRNA) (Zhang et al, 2022),G-quadruplex-modified (G-PE) (Li X Y et al, 2022),and stem-loop RNA aptamer (sPE) (Liu et al, 2022),but the feasibility of these structures in plants has not been reported yet.

Other methods for improving editing efficiency

Inhibiting DNA mismatch repair

The endogenous DNA mismatch repair (MMR)pathway restrains the process of introducing targeted mutations into the genome using the PE system.Therefore, manipulating the key gene expression of the MMR pathway can affect PE efficiency. In 2021,using a CRISPR interference screen, Chen et al (2021)screened a key protein in the MMR pathway, MLH1,and found that its dominant negative variant(MLH1dn) can significantly inhibit the MMR pathway,leading to a significant increase in PE efficiency in mammalian cells. On the other hand, Li J et al (2022)applied MLH1dn to plants and found that overexpression of MLH1dn does not enhance PE efficiency in rice, possibly because the dominant negative protein is derived from mammalian cells rather than plants. To resolve this issue,OsMLH1dnandZmMLH1dn, which are endogenous to rice and maize,respectively, are employed to optimize the PPE system(Fig. 3-A). Interestingly, overexpressing plant-derived MLH1dn facilitates a remarkable improvement in PE efficiency (Qiao et al, 2023).

Suitable higher temperature treatment

The optimal operating temperatures for nCas9 and MMLV, the key components of the prime editor, are 37 °C and 42 °C, respectively. However, both of these temperatures are higher than the optimal plant culture temperature. Therefore, a strategy of appropriately increasing the temperature can enhance the activity of the prime editor, leading to an improvement in its efficiency. Lin et al (2020) compared PPE efficiency at 26 °C and 37 °C in rice protoplasts and found that the PPE activity at 37 °C (average 6.3% efficiency) is significantly higher than that at 26 °C (average 3.9%efficiency). However, Tang et al (2020) found that raising the temperature to 37 °C from 32 °C does not improve editing efficiency in rice protoplasts, possibly due to the difference in original temperature conditions.Because the optimum temperature for MMLV activity is approximately 42 °C (Anzalone et al, 2019), nearly 14 °C higher than the optimum culture temperature(approximately 28 °C to 32 °C) for rice transformation,Zou et al (2022) subjected the resistant calli to an extremely high-temperature treatment condition (42 °C for 2 h followed by culture at 34 °C for two weeks)and found that compared with its function at the default temperature, the PPE3 editing efficiency is significantly improved by about 4-fold.

Co-editing strategy

To enhance PE efficiency, a plant co-editing system including hygromycinY46*- andOsALSS627I-based resistance is developed for editing endogenous genes in rice (Li H Y et al, 2022). This system is able to enrich target editing through dual selection with hygromycin and bispyribac-sodium. Compared with the original PPE3, the co-editing system improves the editing efficiency of target genes by approximately 50-fold.

C-terminus of nCas9 and N-terminus of nCas9 strategies

The positional relationship between nCas9 and MMLV is also an important factor affecting PE efficiency.Anzalone et al (2019) observed higher editing efficiencies when MMLV is fused to the C-terminus of nCas9(cPE) rather than the N-terminus of nCas9 (nPE) in mammalian cells. Based on these findings, they designated the cPE strategy as the PE1 system. However,Xu et al (2022) made an interesting discovery that in transgenic rice plants and maize protoplasts, the editing efficiency of the nPE strategy is significantly higher than that of the cPE strategy. This finding is opposite to what has been observed in mammalian cells. The higher efficiency of the nPE strategy in plants may be attributed to its potential to impact protein expression and stability, leading to a favorable environment for the reverse transcription process.

Balancing flap equilibration process

Under the guidance of pegRNA, the 3′-flap containing the edited sequence is generated after MMLV reverse transcription. Through the flap equilibration process,the 5′-flap containing the unedited DNA sequence is cleaved, and the mutation corresponding to reverse transcriptase is introduced into the genome through the 3′-flap (Anzalone et al, 2019). Therefore, the rapid cleavage of the 5′-flap can accelerate the PE process and improve its efficiency. To validate the feasibility of this method, Liang et al (2023) established the PPE2-v2 system by fusing T5 exonuclease to the N-terminus of the nCas9-MMLV fusion protein in plants (Fig. 3-A). Compared with the original PPE2 system, the PPE2-v2 system can significantly improve PE editing efficiency. Interestingly, Chen et al (2023)found that the direct fusion of T5 exonuclease to PE2 negatively affects PE in mammalian cells, possibly because the DNA repair mechanisms at work differ in mammals and plants.

twinPE strategy

Although PE is a highly precise genome editing technology that can accurately mediate small DNA insertions, deletions, and all 12 base-to-base conversions,it still faces challenges inserting large fragments(Anzalone et al, 2019). To overcome this problem,Anzalone et al (2022) developed a new twinPE strategy in mammalian cells. The principle of the twinPE strategy is to use an nCas9-MMLV fusion protein and a pair of pegRNAs for the programmable replacement or excision of large fragment sequences at endogenous sites. Currently, multiple systems optimized based on the twinPE strategy have been developed in mammalian cells, including PRIME-Del, Bi-PE, GRAND, and PEDAR, which have achieved accurate deletions of 10 kb and precise insertions of 250 bp (Choi et al, 2022;Jiang T T et al, 2022; Tao et al, 2022; Wang et al,2022). Although the length of fragment insertion currently achieved by the twinPE strategy is limited to the optimization of different systems, longer target fragments can be accurately inserted. Likewise, in plants, new PE systems based on the twinPE strategy have also been employed. Lin et al (2021) developed the dual-pegRNA system, which aims to improve PE efficiency by simultaneously delivering two pegRNAs into cells for joint editing of the desired mutations.The testing results of 15 endogenous sites showed that the PE efficiency of dual-pegRNA is, on average,three times higher than that of pegRNA in rice protoplasts. Li et al (2023) developed a plant version of the mammalian GRAND system, effectively inserting protein tag sequences, such as histidine (His),hemagglutinin (HA), and FLAG, into the rice genome(Fig. 3-C). This plant GRAND system not only enables accurate labeling of proteins in living rice cells, but also provides a method for the precise insertion of different elements for crop breeding.

Inserting large DNA segments

The PE system’s editing efficiency still faces the challenge of efficiently integrating larger segments. To address this issue, new tools combining PE with serine integrases have been developed in mammalian cells,such as twinPE-mediated large sequence replacement and deletion, including programmable addition via site-specific targeting elements, which can insert about 40 bp of DNA sequence with an efficiency of up to nearly 25% (Anzalone et al, 2022; Yarnall et al,2023). In plants, Sun et al (2023) further optimized the ePPE system and combined it with the twinPE strategy to establish a dual-ePPE system that achieves precise and efficient insertion of short DNA fragments with an efficiency of over 50%. In order to achieve the insertion of longer DNA fragments, Sun et al (2023)combined the dual-ePPE system with the highly specific recombinase, Cre, to develop the prime editing-mediated recombination of opportune targets (PrimeRoot) system.The PrimeRoot system can accurately integrate large DNA fragments into rice and maize genomes with an efficiency of up to 6%, and the longest successfully inserted fragment length is 11.1 kb. Furthermore, they demonstrated that the PrimeRoot system enables rapid breeding for herbicide and disease resistance by precisely inserting the actin promoter (1.4 kb) into the 5′-untranslated region (UTR) of the herbicide geneHPPDand by integrating the rice blast resistance genepigmRinto the genome.

Expanding editing scope

Due to the ability to extend RTT length for editing sequences that are further from the target site, PE does not require a precisely positioned PAM at the target site. Traditional PE utilizes nCas9, a variant ofStreptococcus pyogenesCas9 (SpCas9), which recognizes genomic sites with an editing PAM sequence of NGG,where N is any nucleotide base (Yu et al, 2022).However, genomic sequences exhibit some preferences,making it challenging to edit sequences that are far from the target site. To further expand the editing scope of PE, different systems based on different variants ofSpCas9, including PE2-VQR, VRQR,PE2-VRER, PE2-SpG, and PE2-SpRY, have been successively developed in mammalian cells, achieving almost complete coverage of the entire genome sequence (Kweon et al, 2021). Currently, theseSpCas9 variants have been demonstrated to be applicable for gene editing in plants, but further investigation is needed to determine whether they are suitable for plant PE systems (Hu et al, 2017; Wang et al, 2019;Xu Y B et al, 2020; Ling et al, 2021). Zong et al (2022)replaced the nCas9 in the original PPE system with a plant codon-optimized nSpG variant to produce the PPE-SpG system (Fig. 3-A). The PPE-SpG system maintains a broad targeting range at NG PAM sequences, with efficiencies ranging from 0.4% to 7.5% in rice protoplasts.

Applications of PPE

The PE system has been widely applied in crop breeding, especially in the creation of herbicideresistant germplasm. With the continuous optimization of PE efficiency in plants, it will likely become the preferred tool for both site-directed mutagenesis at known sites and the creation of new mutations at unknown sites.

Creation of herbicide-resistant germplasm

Currently, the main targets of herbicides in plants include 5-enolpyruvylshikimate-3-phosphate synthase(EPSPS), acetolactate synthase (ALS), acetyl-CoA carboxylase (ACCase). Fine manipulation of target genes is crucial for improving plant resistance to herbicides (Fig. 4-A). TAP-IVS (T102I/A103V/P106S)mutations ofEPSPSwere discovered inAmaranthus hybriduspopulations from Argentina with significant resistance to glyphosate (Perotti et al, 2019). Jiang Y Y et al (2022) used the PPEmax system to precisely edit homologous sites of endogenousEPSPSin rice and generated a TAP-IVS (T173I/A174V/P177S)mutant plant. Compared with the control, the TAP-IVS mutant rice conferres 10 mmol/L glyphosate resistance.Qiao et al (2023) also used the same strategy to obtain TAP-IVS (T164I/A165V/P168S) mutant plants in maize. Subsequent herbicide screening showed that this TAP-IVS mutant maize also exhibits resistance to 10 mmol/L glyphosate.

Natural variation and ethyl methanesulfonate (EMS)mutagenesis screening have revealed that different allelic mutations in theALSgene can result in plants with broad-spectrum resistance against ALS-inhibiting herbicides (Endo et al, 2007; Kawai et al, 2007;Okuzaki et al, 2007; Powles and Yu, 2010; Han et al,2012). Zong et al (2022) used an optimized ePPE system to obtain the W548M allelic plant ofOsALSin rice, which shows dual resistance to 1.10 μL/L of imazapic and 0.09 mg/L of nicosulfuron. Qiao et al(2023) obtained a maize mutant that harbors W542L and S621I mutations inZmALS1andZmALS2via an optimized PPEmax system. This mutant maize exhibits tolerance to flucarbazone herbicide at a concentration of 0.1 g/L. Plants are also sensitive to ACCase-inhibiting herbicides, and resistant plants can be generated by mutatingACC1with excellent alleles.Xu R F et al (2020) successfully created a W2125C mutation in theOsACC1gene using the optimized PPE2 system in rice, which shows resistance to haloxyfop-R-methyl herbicide at a concentration of 5 mmol/L. In addition, Qiao et al (2023) obtained a W2284G mutant of the maizeACC1gene, which exhibits tolerance to 0.09 g/L gallant herbicide. In summary, the PPE system has been widely used in the breeding of herbicide-resistant crops.

Creation of effective alleles

The PPE system not only has the ability to design mutations at existing variant sites, but also to create new variant sites that are absent in nature. As a result,the PE system can be employed to generate novel crop germplasm resources. Based on riceACC1, Xu et al(2021) established a PE library-mediated saturation mutagenesis (PLSM) method by introducing all possible 64 NNN base combinations at the key amino acid sites ofACC1(Fig. 4-B). By conducting PLSM screening on six reported potential resistance sites, a total of 16 different amino acid substitution mutations closely related to herbicide resistance are identified,including the I1879S, P1927Y, and W2097G sites identified in plants for the first time (Jang et al, 2013;Li C et al, 2020). This study established a new PLSM strategy for screening amino acid saturation mutations at specific sites and successfully created sites that do not exist in nature.

Fig. 4. Applications of prime editing systems in plants.

Insertion of protein tags

Efficient insertion of protein tags, such as His, HA,and FLAG into plant genomes, has always been a challenge. Li et al (2023) achieved highly efficient homologous insertion of 6× His and HA protein tags using an optimized ePPE2 system, with insertion efficiencies of 15.63% and 5.21%, respectively.However, when they tested the insertion efficiency of a longer sequence, 3× FLAG (66 bp), it was found to have a low efficiency of around 2% (Fig. 4-C). To improve the efficiency of 3× FLAG insertion, they employed a plant GRAND pegRNA strategy from mammalian cells, achieving a high insertion efficiency of 25% (Li et al, 2023). The results provide a simple,user-friendly and highly efficient protein tagging strategy for research in plants.

Regulation of protein expression

As an mRNA element capable of precisely controlling protein translation, upstream open reading frames(uORFs) are widely present in eukaryotes. However,accurately manipulating uORFs to control plant target gene expression levels is challenging (Srivastava et al,2018; Zhang et al, 2018; Zhang et al, 2020). Xue et al(2023) used the ePPE system to generate two different strategies: extending the endogenous uORFs and generatingde novouORFs, finely manipulating uORFs,and accurately down-regulating the expression of downstream target proteins in rice (Fig. 4-D). Based on this strategy, a series of uORFs mutants with different inhibitory abilities are designed in the 5′-UTR region.As a result, the translation levels of target proteins are finely inhibited in a gradient manner, ranging from 2.5% to 84.9% of their original level, and gradient knockout of the gene is achieved. This universal and innovative method, which can predictably downregulate protein expression, provides a new technical method for future molecular design of crop breeding.

Creation of disease-resistant germplasm

Rice bacterial blight (BB) is a devastating disease caused byXanthomonas oryzaepathovaroryzae(Xoo),resulting in substantial yield losses.Xooemploys transcription activator-like effectors (TALEs) to activate the expression of target genes and disrupt the host immune system, and these virulent TALEs bind to the promoter sequences known as effector binding elements (EBEs) of susceptible genes in the host,ultimately leading to the manifestation of rice disease(Wang et al, 2014; Oliva et al, 2019). Based on the host counteracting mechanism against TALEs ofXoo,Gupta et al (2023) utilized an improved PPE system,PE5max, to engineer resistance against BB through two strategies. One strategy involves the knock-in of a 30-bp TALE EBE, derived from the susceptible geneSWEET14of BB into the promoter of a dysfunctional executor R genexa23, resulting in the generation of the dominant resistance alleleXa23. The other strategy focuses on creating the V39E allele (xa5) of the transcription factor TFIIA geneTFIIAγ5, thereby conferring resistance toXoo. These strategies demonstrate the immense potential of the PE system in creating disease-resistant germplasm and provide valuable references for disease-resistant breeding in other crops.

Conclusions and future perspectives

As a controllable and precise editing system, PPE offers an artificial approach to accelerate crop improvement.It can not only accelerate the production of excellent variants that have evolved naturally over time, but also create numerous variants that do not exist in nature and are designed to meet human-specific demands.Recently, extensive optimization has been carried out to enhance PPE efficiency, and its feasibility has been demonstrated in crop breeding, particularly in herbicide resistance.

At the current stage, there is still need for further optimization of PPE to overcome certain limitations.Specifically, due to the limitations of PAM sequences,some sites cannot be edited. To address this issue, the components of the PPE system can be further improved.In addition, exploring more applicable nCas9 variants can help expand the editing scope beyond current PAM limitations. Moreover, the PPE system has a lower efficiency in producing homozygous T0plants.This not only prolongs breeding time but also limits its application in hybrid rice breeding or polyploid breeding. Therefore, there is a need to further improve the efficiency of PPE in specific application scenarios.

In the future, as a precise and powerful genome editing technology, PPE can be utilized to improve crops in a customized manner without affecting their other desirable traits. Meanwhile, multiple genes can be precisely edited simultaneously through molecular design breeding using PPE, thus efficiently aggregating multiple desirable traits. Most importantly, PPE can target key genes for various important traits in crops,identify excellent allele genotypes, and enrich crop germplasm resources. This facilitates precise molecular crop breeding and advances plant synthetic biology.

ACKNOWLEDGEMENTS

This study was supported by the National Key Research and Development Program of China (Grant No. 2022YFC3400200),the Agricultural Science and Technology Innovation Program of the Chinese Academy of Agricultural Sciences (Grant No.CAAS-ZDRW202001) and the Earmarked Fund for China Agriculture Research System (Grant No. CARS-01-07).