Strategies for delivery of CRISPR/Cas-mediated genome editing to obtain edited plants directly without transgene integration

Increased understanding of plant genetics and the development of powerful and easier-to-use gene editing tools over the past century have revolutionized humankind’s ability to deliver precise genotypes in crops. Plant transformation techniques are well developed for making transgenic varieties in certain crops and model organisms, yet reagent delivery and plant regeneration remain key bottlenecks to applying the technology of gene editing to most crops. Typical plant transformation protocols to produce transgenic, genetically modified (GM) varieties rely on transgenes, chemical selection, and tissue culture. Typical protocols to make gene edited (GE) varieties also use transgenes, even though these may be undesirable in the final crop product. In some crops, the transgenes are routinely segregated away during meiosis by performing crosses, and thus only a minor concern. In other crops, particularly those propagated vegetatively, complex hybrids, or crops with long generation times, such crosses are impractical or impossible. This review highlights diverse strategies to deliver CRISPR/Cas gene editing reagents to regenerable plant cells and to recover edited plants without unwanted integration of transgenes. Some examples include delivering DNA-free gene editing reagents such as ribonucleoproteins or mRNA, relying on reagent expression from non-integrated DNA, using novel delivery mechanisms such as viruses or nanoparticles, using unconventional selection methods to avoid integration of transgenes, and/or avoiding tissue culture altogether. These methods are advancing rapidly and already enabling crop scientists to make use of the precision of CRISPR gene editing tools.

change in a subset of cells within a plant, such as when one cell became transformed with a CRISPR cassette, that single cell grew into a population of transgenic cells, formed an embryo which formed a plant, but the gene editing occurred in only a subset of cells later. In practice it can be difficult to distinguish the origin (mosaic vs chimeric) in plants with a mixture of genotypes.
Most GE methods rely on the DNA repair mechanisms in plant cells to heal the double strand break or nicked DNA, since an unhealed double strand break would be lethal to the cell and useless to agriculture.

Supplementary Note 2: Calculation of editing efficiency
To optimize a process, it is necessary to be able to measure outcomes. The calculation of editing efficiency in regenerated tissue, plantlets, or pools of plantlets "has the potential to be confusing" (Poddar et al. 2023). The simplest case is when whole plantlets are categorized into WT vs Edited. In this case, one can calculate # edited plants / # total plants = fraction edited plants.
However, even within one plant, editing is not always clearly binary. Even in a diploid, the sample could be WT, heterozygous, biallelic, or homozygous. When multiple orthologs of a gene are targeted, and/or in a polyploid genome, the situation becomes more complex. One solution often employed is to set a threshold based on either the detection limit or on a case-by-case basis to continue using the fraction edited plants calculation to summarize the editing efficiency for a group of samples. Another calculation can be performed on individual samples to determine the fraction of DNA that is edited. For example, amplicon sequencing can result in some # edited reads / # total reads = fraction edited DNA in that sample. Further complicating the calculations: the tissue could be chimeric or mosaic, tissue samples may be pooled to save resources, or may be mixtures, such a protoplasts. These mixtures of cells need to be considered in the interpretation of the fraction edited DNA value. Finally, it is important to note that editing fractions may change over time for a few reasons. Editing reagents may continue to be present in the tissue, increasing the fraction edited. Edited and un-edited tissue may divide and grow at a different rate, changing the fraction edited. Also "only a subset of the originally edited tissue ultimately gives rise to an edited plant" (Poddar et al. 2023).

Supplementary Note 3: Additional info on delivery vehicles and tissue-culture-free methods
Several delivery methods that stopped being used for transgene integration are having a comeback in the era of gene editing. For example, use of the gene gun declined once consistent protocols were developed to generate single-copy transgenic events using Agrobacterium in a wide range of species and varieties. The vast majority of transgenic work now employs T-DNA delivery by Agrobacterium rather than the gene gun. The possibility of delivering gene editing reagents in a DNA-free way, such as with ribonucleoprotein revived interest in the gene gun. It is possible that other methods which were attempted and discarded for transgenics may be found to be effective for delivery of gene editing reagents.
Gene editing has brought a resurgence in plant transformation methods, as gene editing reagents have distinct requirements compared to "simple" transgene integration. The field is no longer dominated by Agro-transformation for T-DNA delivery and low-copy "quality" events. In gene editing delivery by mRNA and by ribonucleoprotein is replacing DNA-based delivery, mainly driven by regulatory considerations of transgenes and an expectation of lower off-target editing. Both in gene editing and in transgenics, there is growing demand for tissue-culture free delivery methods. For gene editing, tissue-culture free has recently been demonstrated via viral or grafting-based delivery of cell-to-cell mobile reagents, and we look forward to these techniques reaching more and more crops. In a trend mirroring the medical field, the availability of smaller gene editors is enabling some of these delivery methods including viral delivery.
Agrobacterium-based delivery may also be desirable due to its potential for tissue-culture-free delivery. For decades, Arabidopsis has been transformed routinely using floral dip methods (Clough and Bent 1998). More recently, Agro-transformation with Agrobacterium rhizogenes was shown to be possible without tissue culture using a root Cut-Dip-Budding method which relies on the innate root suckering ability of a few species including dandelion, sweet potato, woody medicinals, and some legumes (Cao et al. 2023). Unfortunately, to date, neither of these methods has been adapted to achieve gene editing that is both tissue-culture and also transgene-free.
Gene editing without transgenes and without tissue culture is highly desirable. A few reports have achieved both, while others have reported transgenic tissue-culture free or transgene-free with tissue culture. As described in the viral delivery section, tissue-culture free gene editing can be achieved if the reagents are delivered to meristems or if viral reagents spread throughout the entire plant (T. Li et al. 2021) and (Maher et al. 2020). Alternately, transgenes or gene editing reagents can be delivered to embryonic meristems (Hamada et al. 2017;Imai et al. 2020) or axillary meristems (Manickavasagam et al. 2004;Mayavan et al. 2015). In Arabidopsis floral dip Agro-transformation, tissue-culture-free methods are routine, but limited to this and closely related species due to flower anatomy and development (Bent 2023). Two additional tissue-culture free delivery methods rely on plant cells: pollen delivery in the form of HI-EDIT (Kelliher et al. 2019)and mobile RNA delivery though grafting (Yang et al. 2023).

Supplementary Note 4: Protoplasts and Microspores
Plant protoplasts, cells with the cell wall removed, resemble animal cells and are thus amenable to many of the same transformation (aka transfection) methods, such as lipofection Zhang, Iaffaldano, and Qi 2021), electroporation, and PEG/Ca 2+ -mediated transfection, which need to bypass only the cell membrane. These transformation methods can deliver gene editing reagents to a large proportion of the protoplast cells in a tubeoften over 50%, and in some optimized systems, very close to 100% (Lin et al. 2018;Woo et al. 2015;J.-F. Li et al. 2013;Yoo, Cho, and Sheen 2007). In addition, protoplasts are amenable to transfection with plasmid DNA, linear DNA, single-stranded DNA, RNA, protein, RNPs, and any combination of the above. A second advantage of protoplasts is that they are separate single cells, which can, in some species/varieties, be regenerated into entire plants. From this perspective, highly-efficient DNA-free genome editing of protoplasts is extremely attractive for producing transgene-free edited plants (Kim et al. 2017).
However, regeneration of plants from protoplasts is not widely adopted in the industry setting. Protoplast isolation (cell wall removal) and transfection are very stressful to delicate plant cells. These stresses significantly hinder the ability to regenerate a cell wall, trigger cell division, and eventually to regenerate a healthy plant. Protocols have been developed for some varieties of some species, but are not available for the majority of species. Even in crops where these protocols have been developed, the resulting plants are often mal-formed with epigenetic defects and are typically bred to remove these defects.
Even though protoplast regeneration is uncommon, the ease of reagent delivery makes protoplasts a very useful system for rapid screening of gRNAs, nucleases, vectors, etc. After delivering reagents and allowing editing to occur, DNA can be extracted for edit detection. Alternately, researchers have also used repair of "broken" GFP or conversion of GFP and BFP to rapidly quantify gene editing in cells without needing to rely on molecular methods (Glaser, McColl, and Vadolas 2016;Zong et al. 2017;Jiang et al. 2020). However, it must be noted that nonregenerating protoplasts are non-dividing and may not be the best model for DNA repair, which varies throughout the cell cycle (non-homologous end joining more in G1, homology-directed repair more in S and G2) (Huang and Puchta 2019).
In some crops, microspores can be isolated, edited, and regenerated using methods somewhat resembling the protoplast techniques described above (Bhowmik et al. 2018). Microspore gene editing offers the advantages of single cells with high transfection efficiencies amenable to transgenefree methods. Microspores are also haploid, which can be advantageous to creating homogeneous edits, but a disadvantage if loss of allele combinations in hybrids is undesirable. To date, microspore regeneration methods are limited to a few species, but improvements continue to be made (Berenguer et al. 2021).