Breaking the Transformation Bottleneck in CRISPR/Cas-Based Crop Improvement: From Genotype-Dependent Regeneration to DNA-Free and In Planta Delivery Systems

CRISPR/Cas genome editing offers new opportunities for crop improvement, but its wider application remains limited more by the difficulty of delivering editing tools into plant cells and regenerating whole plants than by the editing chemistry itself. Many elite cereal, legume, and vegetatively propagated crops remain difficult to transform and regenerate, restricting genome editing to a limited range of amenable genotypes and often requiring additional breeding to introduce desired traits into elite varieties. This review examines how these interconnected barriers influence the recovery of fertile, heritably edited plants and evaluates emerging strategies to overcome them. We discuss developmental regulators such as BBM, WUS2, and GRF-GIF fusions, approaches that induce new meristems, and alternative delivery systems involving nanoparticles, viral vectors, pollen tubes, and DNA-free Cas ribonucleoproteins or guide RNAs. We distinguish the major steps from delivery and editing to plant regeneration, recovery of fertile plants, and transmission of the desired edits to progeny, highlighting that improvements in one step do not necessarily overcome the entire process. We compare these approaches in terms of their applicability across species and genotypes, reported efficiencies, heritability, and practical limitations. Persistent challenges include mosaicism, unintended off-target changes, and limited applicability to non-model and polyploid crops. Finally, we discuss how improved regeneration systems, base and prime editing, machine-learning-guided design, and evolving precision-breeding frameworks in the European Union, United States, Japan, and India may contribute to broader applications of genome editing in crop improvement.