Background

Traditional CRISPR gene editing has primarily relied on cutting both strands of target DNA and relying on the cell's repair response. While effective for knocking out gene functions, precise insertion of desired sequences requires donor DNA and homology-directed repair. In plant cells, the activity of this repair pathway is low, limiting the efficiency of precise substitutions or insertions, and leaving the risk of unintended insertions or deletions at the cleavage site.

Base editing can change specific bases without DNA double-strand breaks, but the types of possible substitutions and the editing range are narrow. Prime editing (PE) was developed to complement this. It combines Cas9 nickase with reverse transcriptase and uses prime editing guide RNA (pegRNA) that contains both the target location and the new sequence to be written. Theoretically, it can implement all 12 types of base substitutions, as well as short insertions and deletions, without donor DNA.

The problem is that PE, which originated in mammalian cells, does not function consistently in plants. Efficiency varies greatly depending on the target locus, and even when editing occurs in somatic cells, it often fails to be transmitted to germline cells. Somatic chimerism, where different genotypes exist within a single individual, also acts as an obstacle to variety fixation.

Key Findings

This paper is not an experimental study testing a new editor, but a review that provides a multi-layered synthesis of accumulated PE technologies in monocots and eudicots. Rather than simply ranking the performance of different PE variants, the researchers reconstructed them into design axes: protein, pegRNA, expression regulation, DNA repair, and delivery vectors. They concluded that editing efficiency depends on the combination of these elements and the target sequence rather than any single component.

The basic mechanism of PE proceeds as follows: Cas9 nickase creates a nick in one DNA strand, exposing the end to which the primer binding site of the pegRNA binds; subsequently, reverse transcriptase copies the sequence from the reverse transcription template into the DNA. Subsequently, cellular flap cleavage and mismatch repair determine whether the new sequence is established. During this process, if the length of the pegRNA primer binding site, the length and secondary structure of the reverse transcription template, additional nicking positions, or enzyme expression levels are mismatched, the editing rate can drop sharply.

The paper identifies protein engineering of Cas9 and reverse transcriptase, stabilization of the 3' end of pegRNA, selection of promoters tailored to plant species and tissues, and regulation of mismatch repair pathways as major improvement strategies. However, it points out that comparing results has become difficult as various laboratories test different structures following PE1 across different crops, loci, and culture conditions. This means that optimal conditions obtained in monocots like rice cannot be directly applied to eudicots like tomato or Arabidopsis.

Meaning and Prospects

The next competitive advantage of PE is expected to lie in 'result predictability' rather than 'editability'. Since the performance of even the same editor varies depending on the target flanking sequences, chromatin accessibility, and the cell's repair status, standard test methods for each crop and common evaluation metrics are required. The mere fact that the desired sequence was detected in some tissues of T0 plants is insufficient. To determine whether it can be considered a real breeding resource, one must also verify the chimera ratio, byproducts, off-target editing, germline transmission rate, and stable inheritance in subsequent generations.

The researchers suggest AI-based protein engineering and data-driven pegRNA design as promising solutions. The idea is to use accumulated sequence and efficiency data to select the optimal reverse transcriptase and guide structure for each target, and to design editors that increase intracellular stability and expression. Delivery technologies that reduce dependence on tissue culture must also be pursued in parallel. In major crops where regeneration is difficult, the utility of the technology decreases significantly if a complete plant cannot be obtained despite high editing efficiency.

Currently, PE is less a universal breeding tool ready for immediate use in all crops and more of a platform requiring optimization for specific conditions. Nevertheless, the advantage of being able to design precise sequence changes at the level of natural variation without double-strand breaks or external donor DNA is clear. Once technical standardization and delivery issues are resolved, the focus of breeding is likely to shift from 'deleting' agricultural traits in editing to accurately 'writing' useful alleles.