Split selectable marker systems utilizing inteins facilitate gene stacking in plants

Metabolic engineering in plants, aiming to create novel traits and produce valuable metabolites and proteins, often requires introducing multiple genes. This multigene engineering is challenging due to limitations in current plant transformation techniques. *Agrobacterium*-mediated transformation, while widely used, faces difficulties with large DNA fragments, leading to instability and truncation. This paper introduces a new method for gene stacking utilizing split inteins, which are protein domains that excise themselves post-translationally, facilitating the ligation of flanking protein sequences (exteins). Previous research demonstrated the use of split inteins in reducing the size of the CRISPR/Cas9 system for base editing. The current study aims to leverage this technology by developing split selectable markers (Kanamycin and Hygromycin resistance genes) enabling co-transformation with a single selectable marker, simplifying the process and enhancing efficiency. The chosen reporter genes, eYGFPuv and RUBY, allow for easy visual assessment of co-transformation success.

The literature highlights the challenges associated with multigene engineering in plants, particularly the limitations of *Agrobacterium*-mediated transformation with large DNA constructs. Previous work showed that large genomic DNA fragments can be unstable in *Agrobacterium*, leading to T-DNA truncation. The use of split inteins has been demonstrated in other applications, such as simplifying the CRISPR/Cas9 system for plant base editing. Prior research on split selectable markers has also been explored, although not extensively in plants. The existing methods often rely on the use of multiple selectable markers, increasing complexity and cost. This study aims to improve upon these limitations by employing a single split selectable marker for co-transformation, a novel approach for gene stacking in plants.

The study used *eYGFPuv* and RUBY as reporter genes for visual assessment of co-transformation success. The RUBY reporter was split into two parts (GTf1 and GTf2), each fused with a portion of the *NpuDnaE* split intein. The functionality of the split RUBY system was tested in *Nicotiana benthamiana* using *Agrobacterium*-mediated leaf infiltration, observing reconstitution of the functional RUBY reporter. Similarly, Kanamycin resistance (*nptII*) and Hygromycin resistance (*hpt*) genes were split into N-terminal (MarN, F1/F3) and C-terminal (Ker, F2/F4) fragments, each paired with a corresponding *NpuDnaE* intein fragment. These split marker constructs were then co-transformed into *Arabidopsis thaliana* via floral dip and poplar using a tissue culture method. Transgenic plants were selected on appropriate selective media, and expression of reporter genes was confirmed using fluorescence and PCR analysis. Western blot analysis confirmed protein trans-splicing of Hygromycin marker fragments. Various techniques like PCR genotyping, phenotypic observation under UV and visible light were employed for analysis across generations to assess heritability and stability of gene stacking. The study included detailed protocols for plant material preparation, vector construction, stable transformation in Arabidopsis and Poplar, Tobacco leaf infiltration, genotyping, phenotyping, protein extraction, western blot and statistical analysis.

The split RUBY reporter system demonstrated functionality in tobacco, indicating successful reconstitution of the functional reporter protein through intein-mediated splicing. In *Arabidopsis*, the split selectable marker system (both Kanamycin and Hygromycin resistance genes) successfully mediated co-transformation, leading to the simultaneous presence of both *eYGFPuv* and RUBY reporter genes. This co-transformation was confirmed by PCR genotyping, showing the presence of both reporter genes in the resistant plants. The traits of antibiotic resistance and reporter gene expression were stably inherited across generations in *Arabidopsis*. In poplar, the split Hygromycin resistance system effectively facilitated co-transformation and expression of *eYGFPuv*, but RUBY expression was less consistent. Western blot analysis confirmed the in vivo protein trans-splicing of Hygromycin marker components demonstrating the orthogonality of the split inteins.

This study successfully demonstrates, for the first time in plants, the effectiveness of split selectable marker systems (split-KanR and split-HygR) for both in planta and in vitro co-transformations in herbaceous and woody plants. The use of split inteins simplifies the process of multigene engineering by reducing the need for multiple selectable markers and the challenges of handling large DNA constructs. While T-DNA insertion sites are not controlled, the Mendelian segregation observed in the offspring is consistent with expectations for such a system. The results highlight the potential of this technology for pathway engineering and genetic improvement of polygenic traits, especially considering potential issues associated with repetitive sequences in plasmids.

The split selectable marker systems using inteins provide a robust and efficient method for gene stacking in plants. This approach simplifies the process of multigene transformation by reducing the reliance on multiple selectable markers and enabling the simultaneous introduction of multiple genes using *Agrobacterium*-mediated transformation. This technology has the potential to accelerate metabolic engineering and genetic improvement efforts in diverse plant species. Future research could explore optimization of the split-intein system, examining different intein sequences and split sites for improved efficiency and expanding the applicability across a wider range of plant species and transformation methods.

The study notes that the insertion sites of the two T-DNAs are not controlled, resulting in Mendelian segregation of the traits in sexually reproducing plants. While *eYGFPuv* expression was consistent across species, RUBY expression, particularly in poplar, was less consistent, suggesting the need for a more reliable reporter gene in future studies. The study mainly focused on two reporter genes and two selectable markers. Further research could investigate the efficacy of this method with additional genes and marker systems.