Virus-Induced Plant Genome Editing: Four Delivery Architectures and Their Trade-Offs

Plant genome editing has been limited by the difficulty of delivering CRISPR (clustered regularly interspaced short palindromic repeats) reagents to germline-competent cells without stable transgene integration. Viral vectors address this by exploiting endogenous replication, cell-to-cell movement and phloem transport, but cargo capacity and germline access have so far traded against each other. We classify virus-induced genome editing (VIGE) systems along two binary design choices: whether the vector itself encodes the nuclease and whether an engineered mobility mechanism carries the editing reagents to the germline. These two axes define four delivery architectures: (a) guide-only without a mobility element, in which guides are delivered into a nuclease-expressing host; (b) guide-only with a mobile RNA element that carries the guide into meristematic cells, yielding heritable edits without tissue culture; (c) nuclease-encoding without a route to the germline, achieved by shrinking the nuclease or enlarging the vector; and (d) nuclease-encoding with a mobility element, reaching the germline directly. We evaluate all four architectures against three persistent constraints. These are off-target risks from viral amplification of the editing reagents, the host range of the available vectors, and the gap between model-system demonstrations and breeding pipelines. Finally, we propose that combining engineered compact nucleases, mobile RNA signals, and morphogenic regulators could enable a single-vector pipeline for producing edited seeds. This would extend genome editing to refractory crop species and reduce reliance on tissue-culture-based transformation.