A Rice Endogenous Small <scp>RNA</scp> ‐Binding Protein Improves Prime Editing for Precise Sequence Insertion and Replacement

To improve prime editing (PE) efficiency in plants, the prime editing guide RNA (pegRNA) has been unceasingly optimised in recent years. The optimisation efforts include, for example, stabilising the pegRNA by incorporating pseudoknots such as evopreQ1 at its 3′ end and processing pegRNA by RNA endoribonuclease Csy4 (Ni et al. 2023). Alternatively, manipulating the cellular environment of the prime editor, using the human small RNA-binding exonuclease protection factor La that functionally interacts with the 3′ end of pegRNA, has further improved the editing performance in mammalian cells (Yan et al. 2024). OsLa1 was reported as a rice RNA-binding protein containing an La domain and two RNA-binding motifs (Guo et al. 2022). We hypothesised that it may have similar functions with the human La protein (hLa) for improving PE (Yan et al. 2024). To test our hypothesis, the truncations containing La and RRM1 domains from hLa, OsLa1 and another La homologue in rice (here designated as OsLa-1 and OsLa-A, respectively) were fused to the C-terminus of M-MLV RT in a recently optimised ePE2 in plants (Li et al. 2023), generating hPE7, OsPE7 and OsPE7-A, respectively (Figure 1a). These editors coordinated with pegRNA or epegRNA (Figure 1b) were evaluated at eight targets in rice protoplasts. Our results demonstrated that OsLa-1 increases editing efficiencies at most of the targets (Figure S1), and thus was selected for introducing base substitutions at Ehd1, IPA1 and ACCase via stable transformation. Analysis for the transgenic lines showed that the editing efficiencies of hPE7 with pegRNA and epegRNA were 1.40-, 1.17- and 3.10-fold, and 1.50-, 1.71- and 4.38-fold over those of ePE2 at the Ehd1, IPA1 and ACCase targets, respectively. OsPE7 had comparable editing activities with hPE7 in either the pegRNA or epegRNA architecture while generating fewer byproducts than ePE2 (Figure 1c,d, Figures S2, S3), indicating that both hLa and OsLa-1 promote precise editing. Encouraged by these results, we subsequently tested OsPE7 for DNA fragment insertion and replacement. Commonly used protein tags of 6× HIS (18 bp), HA (27 bp) and 3× FLAG (66 bp) were designed to seamlessly insert into the coding sequences of UBQ10 and TubA1 (Figure 1e). Compared to the ePE2 control, the ratio of desirable edited mutants (including homozygotes, bialleles and heterozygotes) generated by OsPE7 was increased from 16.3% to 54.2%, 37.5% to 57.9% and 0 to 13.3% for the tagging of 6× HIS, HA and 3× FLAG, respectively (Figure 1f, Figure S4, Table S1). The TwinPE or GRAND editing strategy uses a pair of pegRNAs to facilitate longer sequence insertion and replacement in plants (Li et al. 2023; Liu et al. 2024). To test the compatibility of using OsLa-1 in this system, dual pegRNA or epegRNA was designed to install a 28-bp fragment into the coding sequence of a dysfunctional his1 allele, or to replace a 28-bp fragment within the TVHYNP domain of the SLR1 gene as we described previously (Liu et al. 2024). Similar strategies were also adopted to insert a 3× FLAG or 2× FLAG tag into TubA1 and IPA1 (Figures S5, S6). Genotyping results showed that the frequencies of successful edits generated by OsPE7 with epegRNA were 1.23-, 1.94-, 2.78- and 2.71-fold over those of ePE2 at his1, SLR1, TubA1 and IPA1, respectively (Figure 1g). The ratio of byproducts did not substantially increase (Figure S7, Table S2). As expected, the precise editing lines of HIS1 restored their resistance to the herbicide mesotrione, and the desired slr1 mutant displayed an obvious dwarf phenotype (Figure 1h,i). Successful 3× FLAG tagging events in TubA1 generated by OsPE7 were detected by PCR (Figure 1j), while homozygous and transgene-free IPA1-2× FLAG tagging line in the T1 generation could be confirmed by western blotting (Figure 1k). Collectively, these data demonstrated that the rice endogenous small RNA-binding protein, OsLa-1, is compatible with various pegRNA, epegRNA and dual pegRNA architectures in the PE system to boost precise sequence insertion and replacement. To evaluate the applicability of the OsPE7 system beyond rice, we tested its performance in maize protoplasts for base substitutions at two endogenous sites, ZmACC1 (W2284G) and ZmALS1 (S621L). Our results revealed that OsPE7 significantly enhanced editing efficiency at both target sites compared to the ePE2 system (Figure S8), demonstrating its functional advantage in diverse crop species. Degradation of the 3′ region of the pegRNA in the cellular environment impedes editing efficiency. Incorporating structured RNA motifs such as evopreQ1 to the 3′ terminus of pegRNAs directly enhances their stability (Nelson et al. 2022). A recent study demonstrated that the Csy4 ribonuclease may indirectly stabilise pegRNAs by binding and processing its recognition sites within the pegRNA repeats (Lu et al. 2025). In contrast, according to previous study (Yan et al. 2024) and the editing outcomes of our editors, the La protein may directly bind to the 3′ poly(U) tract of pegRNAs, thereby stabilising the transcript to prevent degradation and minimise self-hybridisation. In our study, the endogenous rice La protein yielded higher editing efficiency than its human orthologue (hLa). This is likely because endogenous proteins are intrinsically compatible with the translational and post-translational machinery of the host, which reduces the risk of triggering cellular immune responses and thereby minimises potential protein degradation and cytotoxicity. J.-K.Z. and M.W. conceived of and designed the research. Y.D., F.S., H.X., S.Z., M.J. and C.Z. conducted the experiments and analysed the data. J.-K.Z., M.W. and Y.D. prepared the manuscript. The project was supported by Nanfan special project, CAAS (YBXM2514) and National Natural Science Foundation of China (32188102, 32271524). The data that supports the findings of this study are available in the Supporting Information of this article. Data S1: Supporting Information. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.