Deciphering the Molecular Architecture and Cross-species Plasticity of in Vivo Haploid Induction: A Critical Narrative Review

Abstract

In vivo haploid induction has moved from a maize-specific breeding phenomenon to a general experimental strategy for perturbing fertilisation, parental genome transmission and embryogenic initiation. This critical narrative review evaluates the molecular architecture that underlies seed-based haploid production and asks why some induction modules transfer readily across species whereas others remain lineage-restricted. Literature was selected through live searches of PubMed, PubMed Central, AGRIS, Google Scholar, DOAJ and Semantic Scholar, supplemented by citation chaining and direct verification of bibliographic records and digital object identifiers. Evidence converges on several partially overlapping mechanistic modules. In maize, the strongest genetic foundation centres on the pollen-associated phospholipase MATRILINEAL/NOT LIKE DAD/ZmPLA1, with DOMAIN OF UNKNOWN FUNCTION 679 MEMBRANE PROTEIN, PHOSPHOLIPASE D3, ZmPOD65 and additional loci modifying penetrance. These factors implicate membrane organisation, lipid signalling, redox balance, sperm competence and post-fertilisation paternal genome stability rather than a single linear pathway. DMP-family loss of function is especially transferable across dicots, while MTL orthologues have shown practical transfer within cereals. Centromere engineering through CENH3 and related kinetochore components represents a mechanistically distinct route based on competitive chromosome inheritance. Newer factors such as GEX3 strengthen the view that perturbing conserved gamete-adhesion and fusion machinery can provide broader cross-lineage portability. Nevertheless, induction efficiency, fertility penalties, polyploid gene redundancy, maternal-genotype effects and uncertain routes from fertilisation failure to embryo rescue constrain universality. The most defensible model is therefore modular: conserved reproductive checkpoints can be engineered to create haploids, but their quantitative outcome is conditioned by species-specific reproductive architecture and genomic buffering. Future progress will depend on mechanistic dissection at single-cell resolution, combinatorial allele design and rigorous cross-genotype validation.