High-throughput haploid induction in species with bisexual flowers

Dear Editor,

Doubled haploid (DH) technology can significantly accelerate the development of homozygous lines. DH breeding has achieved great success in maize because of the discovery of the first haploid inducer, Stock6, and the development of a series of high-efficiency haploid inducers (Hu et al., 2016). Pioneering studies on the genetic basis of haploid induction (HI) revealed that loss-of-function mutation of the phospholipase gene ZmPLA1/MATL/NLD triggers HI and that the HI rate (HIR) can be dramatically enhanced by a single nucleotide substitution from T to C in ZmDMP (Jacquier et al., 2020). Remarkably, knockout of ZmPLA1/MATL/NLD homologs in rice, wheat, and foxtail millet results in HIRs of 2%–6%, 5%–15%, and 2%–3%, respectively (Jacquier et al., 2020; Cheng et al., 2021). In addition, loss of function of ZmDMP-like genes enables HI in species including Arabidopsis, tomato, rapeseed, tobacco, etc., with an average HIR of around 2% (Zhong et al., 2020, 2022a, 2022b). These successes have laid solid foundations for the construction of a universal DH breeding system in different crop species. More importantly, HI-Edit/IMGE systems that enable gene editing in elite germplasms have been established on the basis of HI, making HI even more important (Kelliher et al., 2019; Wang et al., 2019).

One key advantage that makes DH breeding popular in maize is its separate male and female flowers that facilitate convenient cross pollination. Nevertheless, the majority of crop species, including wheat, rice, tomato, rapeseed, etc., have typical bisexual flowers whose emasculation and hybridization are time consuming, thus hindering large-scale HI. Theoretically, the use of male sterile (MS) lines makes crosses easy. However, the majority of MS genes are recessive and cannot be used for HI because haploids will inherit the MS phenotype. Zhang et al. (2014) reported the haploid induction from wheat plants harboring MS2 dominant male sterility gene by interspecies hybridization with corn inducer; however, this method is difficult to be widely adopted by other bisexual-flower plants. Design of a high-throughput HI strategy for plants with bisexual flowers could be crucial for DH breeding in numerous crop species. To this end, we propose a scheme for introducing a dominant MS allele into a breeding population. We provide a proof of concept in wheat and Arabidopsis, demonstrating that use of this method can avoid complex hybridization procedures and easily produce large numbers of haploids in both monocot and dicot species.

In this scheme, a single copy of the dominant MS allele (Msms) leads to the sporophytic MS phenotype (Figure 1A). After crossing with a normal male parent (msms), the resulting progenies exhibit a 1:1 ratio of fertile and sterile plants. In DH breeding, these Msms plants are chosen as female parents to cross with a haploid inducer. Haploids with either the Ms or ms allele will be obtained, and haploids carrying the ms allele have the potential to produce fertile male and female gametes and, eventually, DH lines. Such a method could significantly reduce the amount of work required for large-scale HI.

Figure 1.

Schematic design of large-scale HI and a proof of concept in wheat and Arabidopsis.

(A) Schematic design of large-scale DH breeding in species with bisexual flowers, with dominant MS plants used as female parents.

(B) Phenotypes of dwarf plants of hexaploid (He) and haploid (H) wheat. m, M, r, and R represent ms2, Ms2, rht-D1c, and Rht-D1c, respectively.

(C) Plant heights of hexaploids (left) and haploids (right). Bars represent mean ± SEM. ∗∗∗P < 0.001 (two-tailed Student’s t-test).

(D) Representative figures of ploidy verification for hexaploids (He) and haploids (H) by flow cytometry.

(E) Numbers of haploid and hexaploid plants obtained.

(F) Genotyping of the Ms2 allele in normal-height and dwarf haploids; bands near 2000 bp indicate the presence of the Ms2 allele. M, 2K DNA ladder.

(G) Maps of plasmids used to create MS lines in Arabidopsis by expressing ZmMs7 (top) and Barnase (bottom). The red dotted line represents the fragment deleted after Dex treatment.

(H–J) Representative images of anthers (i), flowers (ii), and self-pollinated siliques (iii) of p5126:ZmMs7(H), pTA29:Barnase(I), and wild-type (J) plants. The anthers were visualized by Alexander staining.

(K) Representative diploid and haploid plants.

(L) Statistics of diploid and haploid seeds obtained.

(M) Detection of MS alleles in haploids. Two haploids (II and III) and 5 haploids (VII, X, XII, XIII, and XIV) from p5126:ZmMs7 × dmp8 dmp9 and pTA29:Barnase × dmp8 dmp9, respectively, lacked MS alleles. M, 5K DNA ladder; WT, wild type.

(N) Screening of haploid seeds lacking MS genes indicated by the absence of a GFP signal. Seeds are shown under white light (top) and fluorescent light (bottom).

(O and P) Dex-induced deletion of MS genes resulted in fertile pollen (O) and elongated self-pollinated siliques (P) in diploid plants of p5126:ZmMs7 (left) and pTA29:Barnase (right). The anthers were examined by Alexander staining.

(Q) Verification of the induced deletion of target fragments in Dex-treated haploid (H) and diploid (D) plants. Bands near 300 bp indicate successful deletion of the target fragment. CK1 and CK2 showed no PCR products from the untreated haploid and diploid, respectively. Scale bars represent 10 cm in wheat plants (B), 100 μm in anthers (H–J and O), 0.5 mm in flowers and seeds (H–J and N), 1 cm in inflorescence phenotypes (H–J and P), and 5 cm in Arabidopsis plants (K).

To test the feasibility of this method, we chose the natural dwarf dominant MS wheat AiBai as the female parent. The MS phenotype of AiBai is caused by expression of Ms2, which is tightly linked to Rht-D1c, which causes the dwarf phenotype (Ni et al., 2017; Xia et al., 2017) (Figures 1B and 1C). After natural pollination, 139 progenies were obtained, including 117 hexaploids and 22 haploids evaluated by flow cytometry (Figures 1D and 1E). Among the 22 haploids, 13 showed normal plant height and 9 showed the typical dwarf phenotype (Figures 1B, 1C, and 1E). Because of the linkage between Ms2 and Rht-D1c, these dwarf haploid plants were expected to carry Ms2, whereas normal-height haploid plants carried the ms2 allele. This speculation was validated by genotyping of the Ms2 allele (Figure 1F). For practical purposes, normal-height haploids could be used for chromosome doubling.

In Arabidopsis, two types of dominant MS systems have been generated by anther expression of ZmMs7 (An et al., 2020) and Barnase (Jagannath et al., 2001), respectively (Figure 1G). A GFP marker driven by the embryo- and endosperm-expressed OLEO1 promoter (Zhong et al., 2020) was introduced to both systems to enable the identification of MS seeds. Fertility analysis revealed that both p5126:ZmMs7 and pTA29:Barnase transgenic plants were completely MS with shriveled anthers but normal female fertility (Figures 1H–1J). These two lines were then crossed with the haploid inducer dmp8 dmp9 (Zhong et al., 2020). Five out of 510 progenies from p5126:ZmMs7 × dmp8 dmp9 and 9 out of 503 progenies from pTA29:Barnase × dmp8 dmp9 were identified as haploids (Figures 1K and 1L). Further analysis showed that two haploids derived from p5126:ZmMs7 and five haploids derived from pTA29:Barnase did not show a GFP signal (Figure 1N) and thus appeared to lack the MS allele. These plants were further verified by PCR (Figure 1M) and would be fertile upon successful chromosome doubling.

To take advantage of the haploids carrying MS alleles, we introduced a dexamethasone (Dex)-inducible Cre/loxP-mediated gene deletion system to remove ZmMs7 or Barnase (Figure 1G). Both ZmMs7 and Barnase plants produced well-developed pollen in the anthers and elongated siliques upon selfing after Dex treatment (Figures 1O and 1P), suggesting that the MS genes had been deleted. PCR assays and sequencing of MS alleles verified the presence of a 340-bp recombined fragment in treated diploid and haploid plants (Figures 1G and 1Q). These data indicated that the target fragment was deleted successfully after Dex treatment. Thus, this method can make full use of haploids obtained for DH breeding.

In conclusion, the introduction of dominant MS systems has significantly simplified cross pollination and made large-scale HI feasible in species with bisexual flowers. Although the breeding populations with MS and marker genes were genetically modified organisms, the DH lines produced from the population were not. In practice, the MS population also facilitates the introduction of elite germplasms, thus enabling continuous improvement. Overall, the strategy presented here can be important not only for large-scale DH breeding but also for germplasm innovation and potential use of HI-Edit/IMGE in many crop species with bisexual flowers (Kelliher et al., 2019; Wang et al., 2019).

Funding

This research was supported by the Hainan Yazhou Bay Seed Laboratory (project of wheat haploid induction B21HJ0501), the National Natural Science Foundation of China (32001554), the China Agricultural Research System (CARS-02), the Chinese Universities Scientific Fund (no. 2022TC141), and the China Postdoctoral Science Foundation (2022TQ0368).

Author contributions

C.L., S.C., Y.Z., and Z.N. conceived and designed the study. X.Q., J.L., Z.L., and S.G. performed the experiments. X.Q. and J.L. performed the data analyses. C.L., Y.Z., X.Q., J.L., C.C., B.C., and Z.N. wrote and revised the manuscript.

Published: September 28, 2022