R2 Retrotransposons Enable Efficient, DSB-Free Gene Insertion into Plant Genomes
核心洞察
Two independent studies demonstrate that R2 retrotransposon (搜索) systems can mediate precise, site-specific DNA insertion into plant ribosomal DNA loci without inducing double-strand breaks.
The optimized R2Tg editor achieved up to ~24% mCherry-positive cells in N. benthamiana and approximately one integrated copy per genome, outperforming Cas9 HDR by ~30-fold at the same rDNA locus.
R2-mediated integration successfully delivered large payloads including a 5-kb metabolic pathway (RUBY) in tobacco and kanamycin/ALS resistance cassettes in rice calli, though heritable transmission remains unconfirmed.
Two landmark studies published in Nature Biotechnology establish R2 retrotransposons as a powerful new platform for targeted gene insertion in plants, achieving precise, double-strand break (DSB)-free integration of multikilobase DNA payloads into ribosomal DNA (rDNA) safe-harbor loci across multiple plant species.
The work, conducted by independent research teams, demonstrates that a single retrotransposon-encoded protein—paired with an engineered RNA donor template—can mediate site-specific integration of full-length gene cassettes via target-primed reverse transcription (TPRT), a mechanism that bypasses the host DNA repair pathways required by CRISPR-Cas9 (搜索) homology-directed repair (HDR).
Optimization yields high-efficiency integration in tobacco
The first team, led by researchers at Caltech, systematically screened R2 proteins from four species—Bombyx mori (R2Bm), Zonotrichia albicollis (R2Za), Taeniopygia guttata (R2Tg), and a rationally improved variant (R2Tg-OPT)—paired with different untranslated region (UTR) combinations. The wild-type R2Tg protein paired with its native 5′ and 3′ UTRs yielded the highest initial integration efficiency at 0.5% of leaf cells showing mCherry fluorescence.
Through iterative optimization—including splitting the system into two plasmids to reduce transcriptional interference, introducing Arabidopsis thaliana introns into the R2Tg coding sequence, and incorporating geminiviral (GV) replicons to boost protein expression—the team achieved a 3.3-fold improvement. Heat-shock treatment (37°C during dark periods) further increased efficiency, ultimately reaching over 2% targeted integration.
A breakthrough came with the design of a compact RNA template featuring minimal homology arms (33 bp upstream, 4 bp downstream) and a hepatitis delta virus (HDV) ribozyme for precise 3′ RNA processing. This R33-R4-HDV payload, combined with intronized R2Tg in a two-plasmid format, achieved 23.6% mCherry-positive cells. Digital droplet PCR (ddPCR) confirmed approximately one integrated copy per genome on average, with roughly 90% of those copies representing full-length integration events.
R2 outperforms Cas9 HDR by 30-fold
In a direct head-to-head comparison targeting the identical 25S rDNA (搜索) locus in Nicotiana benthamiana (搜索) leaves, the R2Tg editor achieved approximately one copy per genome, while Cas9 HDR yielded only 0.035 copies per genome—a roughly 30-fold advantage for the retrotransposon system. Cas9 HDR at single-copy genomic loci (NbPDS1 and NbPDS2) performed even more poorly at approximately 0.01 copies per genome.
Large payload delivery and crop applications
The optimized R2Tg system successfully integrated a 5-kb RUBY reporter cassette encoding a three-enzyme betalain biosynthetic pathway. Visible red pigmentation confirmed functional expression, with ddPCR showing an average of one copy per genome at the 5′ junction and 1.5 copies at the 3′ junction using the one-plasmid format.
The second research team independently validated R2-mediated integration in plants, demonstrating successful insertion of GFP expression cassettes and orthogonal recombinase landing pads (attP, lox66, and FRT) into N. benthamiana rDNA. Their work extended the platform to Oryza sativa (搜索) (rice), where R2-TG and R2-TGopt effectors achieved targeted integration of kanamycin-resistance and acetolactate synthase (ALS) herbicide-resistance cassettes in rice calli, with TG constructs yielding more than 17% kanamycin-positive calli.
High target specificity and junction architecture
Tagmentation-based tag integration site sequencing (TTISS) revealed that 96% of genomic insertions mapped to 25S rDNA (搜索) sites in N. benthamiana, demonstrating high target-site specificity comparable to values reported in mammalian cells. Nanopore long-read sequencing showed that insertions mapped predominantly to chromosomes 6 and 19, which contain the two nucleolar organizer regions (NORs) harboring the majority of the 1,277 R2-compatible 25S sites.
Analysis of 5′ junction architecture indicated that most full-length insertions were seamless anneal-type junctions, with rare join-type and snap-back junctions also detected. The second study reported that 75% of reads corresponded to full-length insertions, while 25% exhibited 5′ truncations, consistent with incomplete reverse transcription during TPRT.
RNA virus-mediated delivery and temporal control
A key innovation from the second study was the development of an RNA virus-mediated donor delivery strategy. By supplying donor RNA templates via recombinant tobacco rattle virus (TRV) or tomato spotted wilt virus (TSWV), the team decoupled donor availability from sustained effector expression. This approach enabled R2-mediated integration in N. benthamiana plants constitutively expressing R2-ZA, with ddPCR estimating 1.97–4.06 donor copies per genome.
Limitations and challenges
Despite these advances, heritable transmission of R2-mediated insertions remains unconfirmed. In rice, although herbicide-resistant shoots were observed under tissue-culture selection, regenerated seedlings failed to establish following transfer to soil. Western blot analysis confirmed R2 effector protein in callus tissue but not in T0 regenerated plants, suggesting effector silencing or selective regeneration from cells lacking R2 activity.
Both studies note that sustained R2 expression and repeated rDNA targeting may impose viability constraints. The highly repetitive nature of plant rDNA arrays—ranging from approximately 800 copies in A. thaliana to over 6,600 in wheat—presents challenges for long-term retention of foreign sequences due to homology-driven recombination and concerted evolution during meiosis.
Future directions
The authors recommend future efforts focus on direct delivery of R2 ribonucleoprotein complexes using nanoparticles or biolistics to eliminate unintended T-DNA insertions, development of inducible or transient expression systems, and extension of targeting beyond rDNA through R2 protein-RNA engineering or fusion to programmable nucleases. The platform is considered particularly well-suited for applications requiring strong expression of large genetic constructs, such as biomanufacturing, metabolic pathway reconstruction, and heterologous protein production in plants.