Background

A study published in aBIOTECH optimized an existing rice genome-editing tool and tripled the proportion of plants with precise insertion of a protein-detection tag at one target. The key changes were introducing a Tf1 reverse transcriptase variant and optimizing codon usage. Unintended edits were also observed, so the findings do not establish a general improvement in editing precision.

The nuclease-mediated prime editing (NM-PE) system, previously developed by the researchers, combines the SpCas9 nuclease with prime editing based on microhomology-mediated end joining (MMEJ). It can insert sequences that label proteins into endogenous plant genes, but improving insertion efficiency for the relatively long 3 × FLAG tag remained a challenge. This study focused on improving performance by changing components of the existing system.

Key findings

The researchers replaced the Moloney murine leukemia virus (M-MLV)-derived reverse transcriptase with evoTf1M4, a Tf1-family variant used in the mammalian PE6 system. They then compared constructs encoding the same protein but using different codon choices. Codon optimization adjusts the nucleotide sequence encoding a protein while preserving its amino acid sequence. The experiments used the rice cultivar Kitaake and analyzed independent primary transformants (T0) into which editing constructs had been introduced using Agrobacterium.

The initial rPE20a produced no detectable precise tag insertion at any of the three tested loci. By contrast, rPE20aV3, optimized for maize codon usage, showed the highest insertion efficiency in the rice experiments. Replacing the reverse transcriptase with a Tf1 variant alone did not guarantee improved performance. The study did not establish the molecular cause of these differences.

In experiments inserting the 66-base-pair (bp) 3 × FLAG sequence, rPE20aV3 produced precise insertions in 9 of 48 T0 plants at the OsCPK6 target. This was 18.75%, three times the 6.25% (3 of 48) observed with the original rPE14b. These percentages represent the proportion of transformed plants with confirmed precise insertions, not the proportion of edited DNA molecules among all DNA molecules. At the other targets, OsCPK3 and OsCPK4, the results were 7/48 plants (14.6%) and 5/48 plants (10.4%), respectively. The researchers confirmed insertions by sequencing and detected signals consistent with the expected molecular weights of the tagged proteins by immunoblotting.

A separate construct, rPE20c, was used to introduce point mutations. At OsACCase, precise editing that was not detected with the comparison construct rPE14a was confirmed in 5 of 45 plants (11.1%) using rPE20c. At OsGS3, however, both rPE14a and rPE20c yielded the same result: 7/48 plants (14.6%). The improvement varied by target.

Meaning and outlook

The results can be interpreted as an improvement to a functional-genomics tool for tagging endogenous rice proteins for detection. A higher proportion of plants with precise insertions may help researchers obtain experimental lines for downstream analyses.

The limitations are clear. Unintended edits, including small insertions or deletions and incomplete tag insertions, were observed at all three loci tested for tag insertion. Detection of the tagged proteins does not demonstrate that the original proteins' functions were preserved. These results also do not establish reproducibility in other cultivars or crops, stability in subsequent generations, or improvements in yield or stress tolerance. Practical use will require target-specific optimization and individual validation of edited lines.

The results and experimental conditions are described in the original aBIOTECH paper.