Efficient parthenogenesis via egg cell expression of maize BABY BOOM 1: a step toward synthetic apomixis

Dear Editor,

Apomixis or asexual reproduction through seed produces plants that are clones of the maternal parent. Clonal seed is important for hybrid crop propagation because hybrid seed production can be complicated and resource intensive. A critical step for introducing apomixis into a sexual plant is the activation of parthenogenesis. In rice (Oryza sativa), OsBABY BOOM (BBM) AP2/ERF transcription factor genes are initially expressed from the male gamete to trigger zygotic development in fertilized egg cells (Anderson et al. 2017, Khanday et al. 2019). Synthetic apomixis has been achieved by combining transgenic ectopic expression of OsBBM1 in rice egg cells with Mitosis instead of Meiosis (MiMe) mutants that abolish meiosis (Mieulet et al. 2016), allowing efficient clonal propagation of hybrid seed without loss of grain quality (Khanday et al. 2019, Vernet et al. 2022). BBM-like genes from the natural apomict Pennisetum squamulatum and foxtail millet (Setaria italica) induce low levels of parthenogenesis in rice and maize (Zea mays) but with pleiotropic defects likely from using a heterologous gene (Conner et al. 2017, Chahal et al. 2022). Determining appropriate BBM-like genes to use in target crops is essential for agricultural use of synthetic apomixis. We describe progress toward achieving apomixis in maize by showing that egg cell expression of a maize BBM transcription factor can efficiently induce parthenogenesis and form viable haploid plants.

At least 4 maize AP2/ERF transcription factors with high similarity to OsBBM1 are de novo expressed in early zygote (Supplemental Table S1 (Chen et al. 2017)). Of these, Zea mays BABY BOOM 1 (ZmBBM1) (Zm00001eb247080) and ZmBBM2 (Zm00001eb144510) are most highly expressed and include the bbm-1-specific region (E/GLSMIKT/NWLRN) (El Ouakfaoui et al. 2010). A recent attempt using trans-activated ZmBBM2 to induce parthenogenesis (Qi et al. 2023) led to low percentages of haploid seed (2%), showing the need for more efficient approaches.

We used ZmBBM1 to determine whether maize BBM genes can induce parthenogenesis and produce haploid plants. An expression cassette comprising a ZmBBM1 cDNA driven by the Arabidopsis EGG CELL 1.2 (EC1.2) promoter was used to create stable transgenic maize lines in the inbred B104 (Supplemental Methods S1 and Table S4). Independent T0 transgenic plants were backcrossed to B104 or selfed. Ploidy was determined for herbicide resistant (Blp^R) T1 progeny from 2 T0 plants by flow cytometry (Fig. 1A–C). T1 progeny from line B1.1A showed no haploids from 28 plants, while T1 progeny of line B5.1A (T0 self) showed 13 haploids from 30 Blp^R plants, including 1 pair of twin plants that emerged from a single seed (Fig. 1D, Supplemental Table S2). The B5.1A line is likely to carry the transgene at a single locus (since 73% of T1 progeny carried herbicide resistance), from which we can estimate that 2/3 of Blp^R T1 plants (20 of 30) had maternal transmission of EC1.2pro::ZmBBM1. Considering only the egg cells carrying the transgene (20), haploid-inducing efficiency was 65% (13/20). Haploid plants were shorter with narrower leaves but did not show aberrant phenotypes other than sterility (Fig. 1B). This shows that egg cell–directed expression of ZmBBM1 triggers parthenogenesis.

Figure 1.

Parthenogenesis triggered by ectopic expression of ZmBBM1 in egg cells. Wild-type B104 fertile tassel A) compared with sterile tassel B) of a haploid T1 plant from line B5.1A. Flow cytometry C) showing the dominant 1n peak from a sterile plant as compared with the 2n peak of the diploid parent; plant was confirmed as haploid by combining the known diploid with the putative haploid samples. The y axis shows the number of nuclei counted, and the x axis shows fluorescence intensity. D) Twin plants from a single seed. E) Parthenogenetic seed showing the presence of 2 embryos (arrows). Scale bars in A) and B) are 10 cm, scale bar in C) is 2 cm, and scale bar in D) is 0.5 cm. WT, wild-type B104; AtEC1.2pro, Arabidopsis EGG CELL 1.2 promoter; ZmBBM1, Zea mays BABY BOOM 1.

Hemizygous T1 plants were crossed to plants homozygous for the dominant, anthocyanin-accumulating R1-nj allele (B73 background) used to detect maize haploids (Nanda and Chase 1966). When an ear from a parthenogenesis-inducing line is crossed to R1-nj pollen, fertilized endosperm leads to purple aleurone on the kernel crown, and sexual fertilized embryos are purple, whereas a haploid embryo from parthenogenesis of an unfertilized egg cell appears white (Supplemental Fig. S1), though embryo color varies with environment and background (Lopez et al. 2023). A DNA marker assay based on a single-nucleotide polymorphism (SNP) difference between B104 and B73 was used to confirm whether an F1 plant resulted from sexual fertilization or parthenogenesis (Supplemental Fig. S1).

Purple embryos were always the result of fertilization of the transgenic line with R1-nj pollen (n = 59 across several lines). Germinated seeds with purple endosperm and presumed white embryos were genotyped with the following possible outcomes: (i) heterozygous B104/B73 plants (result of cross), (ii) B104 allele only plants (result of parthenogenesis), (iii) twin plants from a single white seed (Fig. 1E) (counted as a single parthenogenesis event but genotyped independently where possible), or (iv) no germination (removed from analysis). F1 progeny originating from 11 independent transformants were assessed, and 9 lines showed haploid induction at frequencies from 4% to 75% (Table 1). Variation in parthenogenesis frequency may result from positional effects of transgene insertion on expression levels. Native egg promoters or optimized transgenes may give consistent, higher levels of parthenogenesis.

Table 1.

Parthenogenesis efficiency of EC1.2pro::ZmBBM1 in independent T1 lines

ZmBBM line × R1-nja	Total F1 seed assessed	White seed	Predicted haploids by genotyping	Predicted parthenogenesis efficiencyb	Twin pairsc
B23.1A	35	16	13	74.82%	2
B23.3C	149	43	37	49.57%	4
B4.3A	66	24	16	48.39%	0
B5.1A	90	17	13	28.89%	4
B19.3A	138	17	15	21.74%	3
B20.4B	104	17	7	13.47%	4
B6.1B	49	8	2	8.09%	0
B38.2A	111	20	3	5.40%	0
B35.1A	55	19	1	3.62%	0
B6.2A	273	12	0	0.0%	0
B7.3A	96	8	0	0.0%	0

^aLines were hemizygous for the transgene. ^bPredicted parthenogenesis efficiency calculated as haploids observed/ZmBBM1 transgenic eggs in the T0 plant. ^cEach twin pair is counted as 1 parthenogenic event.

Flow cytometry was performed on several plants from 3 different families to confirm their haploid or diploid status (Supplemental Table S3). The DNA marker predicted plant ploidy in 35 of 38 tested plants. In 3 cases, the SNP marker indicated the absence of the B73 allele but the plants were diploid (Supplemental Fig. S2). Since F1 seed was verified as the result of outcrossing with R1-nj (B73) by purple endosperm, such plants likely arose from early, spontaneous chromosome doubling of a haploid zygote (Wu et al. 2014).

Twin plants were usually haploids, but in 3 of 13 pairs, 1 twin was haploid, while the other was a sexual diploid (from crossing to R1-nj). Since twins are either haploid–haploid or haploid–sexual diploid, they are unlikely to arise from neighboring 2n sporophytic cells. They may arise from a switch in cell fate of 1 synergid to an egg as a response to the loss of the original egg cell (Lawit et al. 2013) followed by fertilization or parthenogenesis triggered by ZmBBM1 or from early splitting of an embryo (monozygotic twins).

ZmBBM1 ectopically expressed using the AtEC1.2 egg cell–specific promoter can trigger embryonic development at efficiencies up to 74%. Based on our results, BBM sequences from the same species may perform better due to evolutionary constraints on interactions with regulatory DNA or other proteins. Our results exceed the 0.4−3.5% haploid frequency recently observed with transgenic transactivation of ZmBBM2 (Qi et al. 2023). Thus, ZmBBM1-induced parthenogenesis is promising for efficient synthetic apomixis in maize, if combined with mutants that circumvent meiosis (Underwood and Mercier 2022), for hybrid seed propagation.

Acknowledgments

The authors wish to thank Brian Hauge and Derek Drost for helpful suggestions and Lynne Hagelthorne and Adrian Garcia for technical assistance.

Author contributions

V.S. and D.J.S. designed the research strategy; M.-J.C. generated maize transformants; D.J.S., M.D.M., and K.Z. performed the experiments; D.J.S. drafted the manuscript; V.S. critically revised the manuscript. All authors have read and approved the final manuscript.

Supplemental data

The following materials are available in the online version of this article.

Supplemental Figure S1. Haploid seed selection and genotyping.

Supplemental Figure S2. Flow cytometry to determine ploidy.

Supplemental Methods S1.

Supplemental Table S1. Maize BBM-like genes.

Supplemental Table S2. Parthenogenesis efficiency of EC1.2pro::ZmBBM transgenic lines.

Supplemental Table S3. Flow cytometry of putative haploids.

Supplemental Table S4. Primers.

Funding

This research was funded by a STAIR grant to V.S. from the University of California, Davis (Innovation and Technology Commercialization and the Office of Research).

Data availability

The data supporting the findings of this study have been provided in the text and in the supplementary data files and are available upon request.

Dive Curated Terms

The following phenotypic, genotypic, and functional terms are of significance to the work described in this paper:

Author notes

Conflict of interest statement. None declared.