In Vivo Maternal Haploid Induction by Disrupting KOKOPELLI in Medicago truncatula

In vivo haploid induction (HI) has attracted interest in crop breeding and genetic research. HI facilitates creation of pure lines for hybrid seed production, development of quantitative trait locus (QTL) mapping populations, fixing hybrid vigour through synthetic apomixis, and direct gene editing in commercial varieties via HI editing (HI-Edit), also known as haploid-inducer mediated genome editing (IMGE) (Qu et al. 2024; Li et al. 2025; Ravi et al. 2014). Haploids can be generated in diverse plants through loss-of-function mutations or overexpression of paternal and maternal HI genes (Quiroz et al. 2024; Wang et al. 2025). However, slow progress in developing seed-based in vivo HI systems for legumes has impeded the application of doubled haploid technology in this agriculturally vital family. Disrupting AtKPL (KOKOPELLI, AT5G63720) in Arabidopsis thaliana impaired gamete interactions and induced haploid embryo formation, triggering maternal HI (Jacquier et al. 2023). AtKPL is highly expressed in pollen and encodes a protein without recognisable domains. Atkpl mutants exhibit seed abortion and arrested ovule development resulting from male gametophyte defects and frequent single fertilizations. AtKPL-like proteins are ubiquitous in monocots and dicots, suggesting that KPL-mediated in vivo HI systems could be promising for leguminous crops (Jacquier et al. 2023). In the model legume Medicago truncatula, a single KPL-like protein, MtKPL (Medtr6g078430), shares 44% identity with AtKPL and 51%–56% identity with soybean (Glycine max) GmKPLs (Figure S1). MtKPL is expressed specifically in anthers at different developmental stages and in mature pollen (Figure 1a). We identified two mutant lines from the Tnt1 retrotransposon-tagged population of M. truncatula ecotype R108: Mtkpl-1 (NF19728) and Mtkpl-2 (NF5171), with Tnt1 insertions in exons 1 and 3, respectively (Figure 1b). Both mutations were null alleles (Figure 1c). Under greenhouse conditions, Mtkpl mutants exhibited no phenotypic differences from wild type (WT) during the vegetative stage. However, the mutants displayed reduced seed-setting rates with fewer whorls per pod (from 5 ± 0.28 in WT to 4.81 ± 0.39 in Mtkpl-1 and 4.64 ± 0.32 in Mtkpl-2) and seeds per pod (from 6.98 ± 1.14 in WT to 5.44 ± 1.23 in Mtkpl-1 and 5.17 ± 1.09 in Mtkpl-2) (Figure 1d,e), although there was no difference in pollen viability between Mtkpl and WT (Figure 1f). The reduced seed set caused by MtKPL mutations is consistent with findings in Arabidopsis, supporting an evolutionarily conserved role for this gene in reproductive development. To determine whether Mtkpl mutants exhibit HI, we examined the progeny of self-pollinated WT, Mtkpl-1, and Mtkpl-2. We initially screened for typical haploid characteristics of reduced stature and sterility. Using flow cytometry (Figure 1g), we identified one haploid among the 342 progeny of Mtkpl-1 (HI rate [HIR] of 0.29%), two haploids from 452 Mtkpl-2 progeny (HIR of 0.44%) (Figure 1n) and none from WT. Similar to haploids derived from self-pollinated Mtdmp mutants, these haploid plants exhibited sterility and morphological changes including reduced leaf size, smaller flowers, underdeveloped ovules, non-viable pollen (Figure 1h–k), and lower epidermal and guard cell areas compared to their diploid counterparts (Figure 1l,m). We hand-pollinated M. truncatula ecotype Jemalong A17 with Mtkpl-1, Mtkpl-2, or WT R108 pollen. This resulted in HIRs of 0.61% (3/494), 0.64% (3/471), and 0.0% (0/600), respectively (Figure 1n). The haploid plants were morphologically similar to the female parent A17, but with an overall reduced size (Figure 1o). We subjected three haploids to whole-genome sequencing, and none carried paternally derived SNPs (Figure S2). We conclude that disrupting MtKPL induces in vivo maternal HI in M. truncatula. In Arabidopsis sperm, KPL and its inversely transcribed neighbour ARIADNE14 (ARI14) generate a natural antisense transcript-derived small interfering RNA (nat-siRNA) pair (Ron et al. 2010), leading to the degradation of ARI14 transcripts. In the absence of KPL, high ARI14 transcript levels impair fertilisation. However, M. truncatula has neither an ARI14 homologue nor genes oriented opposite the KPL locus (Figure S3), suggesting that M. truncatula lacks sperm-specific cis-nat-siRNA during double fertilisation. In summary, the inactivation of MtKPL triggers HI in M. truncatula, providing a foundation for haploid gene editing in legumes and offering opportunities to study HI mechanisms in this plant family. N.W., H.L., and L.N. designed research. N.W., Y.F., J.C., Y.G., and L.N. performed experiments and analysed data. Y.P., J.W., and K.S.M. contributed analytical tools. N.W., H.L., and L.N. wrote the manuscript with contributions from all authors. We thank Prof. Kejian Wang, Prof. Chun Wang, and doctoral student Like Chen for help with genotype calling. This work was supported by grants from the National Key Research and Development Program of China (2022YFF1003204), National Natural Science Foundation of China (32325035), Hohhot Key R&D Project (2023-JBGS-S-1), and Central Laboratory of Biotechnology Research Institute, Chinese Academy of Agricultural Sciences. The authors declare no conflicts of interest. The data that supports the findings of this study are available in the Supporting Information—S1 of this article. Figure S1: pbi70628-sup-0001-Supinfo.docx. M. truncatula KPL and sequence identity with GmKPLs. Figure S2: pbi70628-sup-0001-Supinfo.docx. Mtkpl-1 induces maternal haploid formation. Figure S3: ARI protein phylogeny and MtKPL genomic organisation. Table S1: Primers used in this study. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.