Poaceae-specific MS1 encodes a phospholipid-bindingprotein for male fertility in bread wheatZheng Wanga,1, Jian Lia,1, Shaoxia Chenb,1, Yanfang Hengc,1, Zhuo Chenc,1, Jing Yangc, Kuanji Zhoua, Jiawei Peic,Hang Heb,2, Xing Wang Dengb,2, and Ligeng Mac,2

aFrontier Laboratory of System Crop Design, Beijing 102206, China; bState Key Laboratory of Protein and Plant Gene Research, Peking–Tsinghua Center forLife Sciences, School of Advanced Agricultural Sciences and School of Life Sciences, Peking University, Beijing 100871, China; and cCollege of Life Sciences,Capital Normal University, and Beijing Key Laboratory of Plant Gene Resources and Biotechnology for Carbon Reduction and Environmental Improvement,Beijing 100048, China

Contributed by Xing Wang Deng, September 28, 2017 (sent for review September 6, 2017; reviewed by Daolin Fu and Christian S. Hardtke)

Male sterility is an essential trait in hybrid seed production formonoclinous crops, including rice and wheat. However, comparedwith the high percentage of hybrid rice planted in the world, littlecommercial hybrid wheat is planted globally as a result of the lackof a suitable system for male sterility. Therefore, understandingthe molecular nature of male fertility in wheat is critical forcommercially viable hybrid wheat. Here, we report the cloning andcharacterization of Male Sterility 1 (Ms1) in bread wheat by usinga combination of advanced genomic approaches. MS1 is a newlyevolved gene in the Poaceae that is specifically expressed in mi-crosporocytes, and is essential for microgametogenesis. Orthologsof Ms1 are expressed in diploid and allotetraploid ancestral spe-cies. Orthologs of Ms1 are epigenetically silenced in the A and Dsubgenomes of allohexaploid wheat; only Ms1 from the B subge-nome is expressed. The encoded protein, Ms1, is localized to plas-tid and mitochondrial membranes, where it exhibits phospholipid-binding activity. These findings provide a foundation for thedevelopment of commercially viable hybrid wheat.

Male Sterility 1 | epigenetic silence | phospholipid binding | wheat | hybridseed production

Wheat is an important staple food crop worldwide; it con-stitutes approximately 20% of the calories consumed by

humans and serves as the major food source for 30% of theworld’s population. Global wheat grain yields increased in the1960s and 1970s as new varieties with mutations in the “greenrevolution” gene, a GAI ortholog in wheat, were adopted (1, 2).However, to meet the demand of an increasing global populationfor a high-quality food supply, a substantial increase in wheatgrain yield is vital (3, 4). Thus, a new green revolution in wheatis necessary.Hybrid vigor is an important consideration in increasing crop

yields. The breeding and large-scale adoption of hybrid rice andcorn have contributed significantly to the global food supply,indicating that the use of hybrid crops is a feasible means ofincreasing crop yields (5, 6). However, hybrid wheat currentlyaccounts for less than 0.2% of the total planted wheat acreagearound the world despite several decades of development (7, 8).The lack of commercial progress in hybrid wheat is the result of alack of a practical male sterility trait, which is essential for hybridseed production by monoclinous crops (9, 10). Therefore, iden-tifying a nuclear recessive male sterility trait and its corre-sponding gene in wheat is a prerequisite for commercial hybridwheat breeding and hybrid seed production (11), which has beendemonstrated in maize (12, 13) and rice, another monoclinouscrop (5, 6). Among the five stable genic male sterility (GMS) loci(MS1–MS5) identified thus far in bread wheat (14–18), ms1 andms5 are recessive mutants (16, 19), whereas Ms2, Ms3, and Ms4are dominant mutants (20–22). Presently, only one dominantgene, Ms2, has been cloned (23, 24). Ms2 mutants have beenwidely used for wheat breeding and potentially for hybrid wheatbreeding (23). However, to date, recessive nuclear genes

affecting male fertility have not been cloned, even though mu-tants were identified almost 60 y ago.Here, we report the cloning and molecular, biochemical, and

cell-biological characterization of a nuclear recessive locus, MS1,in allohexaploid bread wheat. We developed a strategy forcloning wheat genes, MutMap-based cloning, by combiningMutMap (25) and traditional map-based cloning approaches.MS1 is a newly evolved gene that exists only in the Poaceae. It isspecifically expressed in microsporocytes, with ortholog sistergenes that are epigenetically silenced in the A and D sub-genomes of allohexaploid wheat. Our work details a nuclear-recessive gene that regulates male fertility in hexaploid wheatand provides a foundation for large-scale commercial hybridwheat breeding and hybrid seed production.

ResultsMs1 Is Required for Microgametogenesis in Wheat. Unlike WTplants, ms1e plants lack extruded anthers and the glumes remainopen at anthesis (Fig. 1A). There is no obvious difference in pistildevelopment between WT and ms1e plants (Fig. 1 B and C);however, ms1e anthers are slightly smaller and indehiscent. Fur-ther, ms1e anthers bear aborted pollen, leading to unfertilizedpistils and complete sterility (Fig. 1 C–E). To examine the role ofMs1 in pollen development, we analyzed microsporogenesis and

Significance

Heterosis provides an important strategy for increasing cropyield, and breeding and adoption of hybrid crops is a feasibleway to increase crop yields. Male sterility is an essential trait inhybrid seed production for monoclinous crops, including wheat.Heterosis in wheat was observed approximately 100 y ago.However, very little commercial hybrid wheat is planted in theworld because of the lack of a suitable male sterility trait.Therefore, understanding the molecular nature of male fertilityin wheat is critical for hybrid wheat development. Here, we re-port the cloning and molecular, biochemical, and cell-biologicalcharacterizations of Male Sterility 1 (Ms1) in bread wheat, andprovide a foundation for large-scale commercial hybrid wheatbreeding and hybrid seed production.

Author contributions: X.W.D. and L.M. designed research; Z.W., J.L., Y.H., Z.C., K.Z., andJ.P. performed research; Z.W., J.L., S.C., J.Y., H.H., X.W.D., and L.M. analyzed data; andZ.W., J.L., H.H., X.W.D., and L.M. wrote the paper.

Reviewers: D.F., University of Idaho; and C.S.H., University of Lausanne.

The authors declare no conflict of interest.

Published under the PNAS license.

Data deposition: The sequences reported in this paper have been deposited in the NCBISequence Read Archive (accession no. SRP113349).1Z.W., J.L., S.C., Y.H., and Z.C. contributed equally to this work.2To whom correspondence may be addressed. Email: hehang@pku.edu.cn, deng@pku.edu.cn, or ligeng.ma@cnu.edu.cn.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1715570114/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1715570114 PNAS Early Edition | 1 of 6

PLANTBIOLO

GY

Dow

nloa

ded

by g

uest

on

Oct

ober

24,

202

0

microgametogenesis in WT and ms1e plants by using DAPIstaining and histological sections. No difference in pollen devel-opment was observed up to the early uninuclear stage of micro-sporogenesis between WT and ms1e plants (SI Appendix, Fig.S1A); approximately 14% of ms1e pollen grains had becomeshrunken with a condensed nucleus at the late unicellular stage. No

normal binucleate or trinucleate pollen grains developed in ms1e(SI Appendix, Fig. S1A), suggesting a defect during microgameto-genesis. However, no defects were observed in the surroundingsomatic cell layers inms1e compared with WT, including defects inthe tapetal layer (SI Appendix, Fig. S1B). These data suggest thatMs1 is required for microgametogenesis in wheat.

MutMap-Based Cloning of Ms1 in Bread Wheat. Ms1 was formerlymapped to a region encompassing the distal 16% of the shortarm of chromosome 4B (4BS) (26). In our study, we used amodified version of the MutMap approach (25) to clone Ms1.First, we used transcriptome sequence (i.e., RNA-seq) data fromthree Ms1 genotypes—Ms1/Ms1, Ms1/ms1e, and ms1e/ms1e—segregated from the progeny of heterozygous ms1e plants (SIAppendix, Table S1). As ms1e was recovered from an EMS-mutagenized population of WT Chris (19, 27), it was expectedto carry a point mutation in Ms1. Through an alignment of thereads to predicted genes (28) and mapping to chromosome 4B(29), SNPs between the expressed genes from Ms1/Ms1 andms1e/ms1e were found to be concentrated in an ∼50-cM regionof chromosome 4BS (Fig. 1F), consistent with the notion that

Fig. 1. Phenotypic characterization of ms1 and the MutMap-based cloningof Ms1 from hexaploid wheat. (A) Spikes of Ms1 and ms1e. (Scale bar: 1 cm.)Anthers ofMs1 andms1e before (B) and after filament elongation (C). (Scalebar: 1 mm.) (D) Mature pollen grains of Ms1 and ms1e stained with I2-KI.(Scale bar: 200 μm.) (E) Seeds of Ms1 and ms1e at 20 d after pollination.(Scale bar: 1 mm.) Scatter diagrams of SNPs between Ms1 and ms1e plantsobtained from RNA-seq (F) and DNA-seq (G) by MutMap analysis. The y axisshows the indexMU/indexWT ratios [index = Nmutant/(Nreference+Nmutant), Nrepresents the number of accumulated reads with corresponding geno-types], and the x axis presents the relative physical (Upper) and genetic(Lower) positions of each SNP on chromosome 4B (Chr. 4B) (31). (H) Map-based cloning of Ms1 using SNP markers. (Top) Genetic map of the Ms1 re-gion on chromosome 4BS based on SNP markers obtained from RNA-seqdata. The SNP markers and their corresponding genetic positions are in-dicated. (Middle) Fine mapping of theMs1 region on chromosome 4BS basedon the SNP markers obtained from the DNA-seq data. The numbers beneaththe line in the second panel indicate the recombination frequency. (Bottom)Genetic structure of Ms1 and ms1 allele characterization. Closed boxesrepresent exons; connecting lines represent introns. ATG and TGA are in-dicated, and the mutation for each of 13 alleles is indicated. The mutationsite in each allele is based on the sequence in the variety from which theallele is generated.

Fig. 2. Genomic structure of Ms1 in the genomes of hexaploid, tetraploid,and diploid wheat and the origination of MS1 in Poaceae. (A) Southern blothybridization of Ms1 in hexaploid, tetraploid, and diploid wheat using Ms1-specific probes. Genomic DNA from AABBDD (Ms1/Ms1), ms1g, AABB(T. turgidum accession Langdon), AA (T. urartu accession G1812), and DD (Ae.tauschii accession AL8/78), respectively, was digested with HindIII. The sizemarkers are from a λ-DNA-HindIII digest. (B) Phylogenetic tree of Ms1 and itsorthologs in Poaceae plants. Proteins are named according to their species.The numbers at the nodes show bootstrap values obtained for 1,000 repli-cates. The accession number and species name for each gene are shown in SIAppendix, Table S6. (C) The origination of MS1 in the Poaceae family.

2 of 6 | www.pnas.org/cgi/doi/10.1073/pnas.1715570114 Wang et al.

Dow

nloa

ded

by g

uest

on

Oct

ober

24,

202

0

Ms1 is located in this region. However, the huge and highlysimilar three bread wheat subgenomes made it difficult toidentify the SNP responsible for the mutation in Ms1. Therefore,instead of identifying Ms1 from those SNP-containing genes,we used the SNPs as markers to further map Ms1 by using676 progeny of heterozygous ms1e plants. In this way, we nar-rowed down Ms1 to a 12.82-cM interval in chromosome 4BS(Fig. 1 F and H and SI Appendix, Table S2).As the SNP markers from the RNA-seq data were not suffi-

ciently dense to narrow the region containing Ms1, we rese-quenced the genomes of the Ms1/Ms1 and ms1/ms1 genotypessegregated from the progeny of the ms1e heterozygotes (SI Ap-pendix, Table S3) to generate additional SNP markers for finemapping (Fig. 1G and SI Appendix, Table S2). Eventually, wemapped Ms1 to a 198-kb interval by using the SNP markersgenerated from our DNA-seq data (Fig. 1 G and H and SI Ap-pendix, Fig. S2). Nine predicted genes were located within thisregion (SI Appendix, Table S4).To identify the gene responsible for the ms1 phenotypes from

these candidates, we took advantage of three reported ms1 al-leles as well as 11 newly recovered ms1 alleles from our EMS-mutagenized population of Chinese bread wheat (variety“Ningchun 4”; SI Appendix, Table S5). Sequencing revealedmutations or deletions for each of the 14ms1 alleles in candidategene 6 (C6; Fig. 1H and SI Appendix, Table S5), indicating thatC6 was Ms1. To confirm this, we transferred a genomic DNAfragment containing full-length C6 into immature embryos thatare homozygous for the ms1e allele. The genomic fragmentcompletely complemented the abnormal spike and pollen de-velopmental phenotypes of the mutant, and it restored malefertility (SI Appendix, Fig. S3). Thus, C6 was confirmed to beMs1from the hexaploid bread wheat genome.

MS1 Is a Newly Evolved Gene in the Poaceae Family. To determinethe genomic structure of Ms1, we performed Southern blottingby using probes specific for Ms1. One Ms1 was detected in thehexaploid bread wheat genome, from the B subgenome; no Ms1was detected in Triticum urartu (AA) or Aegilops tauschii (DD),and only one Ms1 was detected in allohexaploid wheat or Triti-cum turgidum (an allotetraploid; AABB; Fig. 2A). Moreover, noMs1 was detected in ms1g, indicating that the Ms1 region used asthe probe was deleted inms1g (Fig. 2A). A BLAST analysis usingMs1 revealed two Ms1-related genes from the A and D sub-genomes, respectively, of the hexaploid wheat genome, withDNA sequence identities of 82.9% and 83.9% between Ms1 andMs-A1 and between Ms1 and Ms-D1, respectively (SI Appendix,Fig. S4). Ms1 orthologs were identified only in the Poaceae,including all seven subfamilies for which the sequence of MS1was available (Fig. 2B and SI Appendix, Fig. S5 and Table S6).No full-length Ms1 homolog was identified from other families,but sequences corresponding to the 5′ and 3′ portions of Ms1,respectively, were identified across angiosperms (Fig. 2C and SIAppendix, Tables S7 and S8). These results suggest that MS1 is anewly evolved gene within the Poaceae lineage, but that it con-tains two conserved domains that arose independently in theancestral angiosperm lineages.

Ms1 Is Specifically Expressed in Microsporocytes and Its OrthologsAre Epigenetically Silenced in the A and D Subgenomes ofHexaploid Wheat. An RNA expression analysis indicated thatMs1 is not expressed in the roots, stems, or leaves of breadwheat; it is expressed only in the spikes. Further, it was highlyexpressed in developing anthers during microspore meiosis anddetectable during the uninuclear stage of microgametogenesis,but it was not expressed in the lemma, palea, or pistil (Fig. 3A).The same expression pattern was observed for Ms1 protein in

Fig. 3. Ms1 expression pattern in wheat. Expression analysis ofMs1 by RT-PCR (A, C, D, Upper; and E, Left) and quantitative RT-PCR (A, C, D, Lower; and E, Right)in hexaploid T. aestivum (A), diploid T. urartu (C), diploidAe. tauschii (D), and allotetraploid T. turgidum (E). BP, bicellular pollen stage; Le&pa, lemma and palea;MP, mature pollen stage; UM, unicellular microspore stage. ACTIN was used as an internal control. Error bars indicate the SD of three biological replicates. (B) Insitu hybridization analysis ofMs1 in WT andms1g anthers. MC, meiotic cell; MMC, microspore mother cell; SP, sporogenous cell; T, tapetum; Tds, tetrads. (F) DNAmethylation analysis of the Ms1 promoter in the genomes of hexaploid, allotetraploid, and diploid wheat species.

Wang et al. PNAS Early Edition | 3 of 6

PLANTBIOLO

GY

Dow

nloa

ded

by g

uest

on

Oct

ober

24,

202

0

wheat anthers (SI Appendix, Fig. S6A). In situ hybridizationrevealed that Ms1 was expressed in secondary sporogenous cellsand highly expressed in microsporocytes, but not expressed inpollen grains during microgametogenesis (Fig. 3B). As a control,neither Ms1 mRNA nor Ms1 protein was expressed in ms1g (Fig.3B and SI Appendix, Fig. S6B), consistent with the lack of Ms1 inthe ms1g genome (Fig. 2A); this supports the conclusion thatms1g carries a deletion in the Ms1 region. These expression datasuggest that Ms1 is expressed specifically in microsporocytes.Similar to most genes in the hexaploid wheat genome, Ms1 has

orthologs, Ms-A1 and Ms-D1, in the A and D subgenomes, re-spectively, of hexaploid wheat (SI Appendix, Fig. S4). NeitherMs-A1 expression nor Ms-D1 expression was detected in any ofthe tissues examined, including anthers, in which Ms1 is highlyexpressed (Fig. 3A). However, Ms1 orthologs from the genomes ofdiploid ancestral species of T. urartu (AA) and Ae. tauschii (DD)were expressed in anthers with a similar expression pattern to thatof Ms1 in hexaploid wheat (Fig. 3 C and D). In addition, Ms1orthologs from the A and B subgenomes were expressed in theanthers of allotetraploid T. turgidum (AABB), and their expressionpatterns were again similar to that of Ms1 in hexaploid wheat (Fig.3E). Furthermore, the transformation of Ms-A1 from T. aestivuminto immature embryos homozygous for thems1g allele completelyrescued the abnormal phenotypes of the mutant (SI Appendix, Fig.S7). These data suggest that Ms-A1 is functional, similar to Ms1,but that Ms-A1 and Ms-D1 are silenced in hexaploid wheat.To determine why Ms-A1 and Ms-D1 are silenced in hexaploid

wheat, we examined the DNA methylation level in the promoterregions ofMs1,Ms-A1, andMs-D1, and found that the promotersof Ms-A1 and Ms-D1 were hypermethylated at CG and CHGsites, whereas the Ms1 promoter was not (Fig. 3F). In addition,no obvious DNA methylation was detected in the promoters ofthe Ms1 orthologs in the genomes of T. urartu (AA), Ae. tauschii(DD), and allotetraploid T. turgidum (AABB; Fig. 3F). Together,our data indicate that Ms-A1 and Ms-D1 expression was epige-netically silenced during the generation and selection of allo-hexaploid wheat, and that only Ms1 from the B subgenome isexpressed in the microsporocytes of allohexaploid wheat. Ourdata also demonstrate that epigenetic regulation underlies theasymmetric expression and subgenome dominance ofMs1 and itsorthologs in hexaploid wheat.

Ms1 Is Localized to the Membranes of Plastids and Mitochondria.Ms1encodes a protein containing 220 amino acids, with a predictedsignal peptide between residues 1 and 23, and a predicted trans-membrane domain between residues 206 and 220 (SI Appendix,Fig. S8). However, Ms1 is not secreted from cells (SI Appendix,Fig. S9), indicating that it is a plasma or organelle membrane-localized protein. Ms1-GFP exhibited punctate and threadlikepatterns in cells; it was neither colocalized with the peroxisomemarker PST1 nor colocalized with the Golgi marker GmMan1,but it colocalized with the mitochondrial marker pFAγ. Ms1-GFPwas also localized in plastid membranes; it was detected in ringsaround plastids as a result of colocalization between Ms1 and theplastid marker WxTP (Fig. 4A). Further analysis substantiated thelocalization of Ms1 in the plastid membrane and outer mito-chondrial membrane (Fig. 4B). Deletion of the signal peptideor transmembrane domain of Ms1 disrupted its subcellular lo-calization (Fig. 5 A–C), whereas the signal peptide domain ofMs1 alone localized the tagged protein to the endoplasmic re-ticulum (ER; Fig. 5D). Additionally, truncated Ms1 lacking thetransmembrane domain accumulated in the ER (Fig. 5D), in-dicating that the signal peptide domain of Ms1 is required forlocalization of the protein to the ER, whereas the trans-membrane domain of Ms1 is required for integration into theplastid membrane and outer mitochondrial membrane from theER, at least under its overexpression condition.

Ms1 Is a Phospholipid-Binding Protein. Besides the N-terminal sig-nal peptide and C-terminal transmembrane domain, a lipidtransfer protein (LTP) domain was predicted in Ms1 followingthe N-terminal signal peptide (SI Appendix, Fig. S9). In plants,fatty acids are synthesized mainly in plastids and exported to theER, where they are incorporated into phosphatidic acid (PA),phosphatidylinositol-phosphates, and other membrane lipids(30). To determine whether Ms1 has lipid-binding activity,recombinant MBP-Ms1 was expressed in and purified frombacteria (SI Appendix, Fig. S6 C and D) and then used to mea-sure the binding activity of Ms1 to typical lipid constituents ofplasma or organelle membranes. Ms1 bound PA and severalphosphatidylinositols (PIs), including PI(3)P, PI(4)P, PI(5)P,PI(3,4)P, PI(3,5)P, PI(4,5)P, and PI(3,4,5)P (Fig. 6 A and B).However, no lipid-binding activity was detected for MBP (Fig. 6B and C), indicating that Ms1 is a phospholipid-binding protein,especially for PA and PIs. In addition, neither N-terminal norC-terminal portion of Ms1 bound lipids (Fig. 6C), suggesting thatthe whole Ms1 is required for its lipid binding activity. Therefore,Ms1 may transfer PA and PIs from the ER to plastids and mi-tochondria, and proper localization of PA and PIs is critical formicrogametogenesis in wheat.

DiscussionChanges to the architecture of a gene and its encoded protein maygive rise to new functions. We found that MS1 is a newly evolvedgene that exists specifically in Poaceae plants; it was not found inancestral or more recently evolved lineages other than Poaceae(Fig. 2B). Given that MS1 originated only in Poaceae, it may playan essential role in the development and survival of Poaceaeplants. Indeed, Ms1 is essential for male fertility and metagenesis

Fig. 4. Subcellular localization of Ms1 and its colocalization with organellemarkers. Onion epidermal cells were transiently cotransformed with con-structs encoding Ms1-GFP and mCherry-fused organelle markers driven by the35S promoter. GmMan1, Golgi marker; pFAγ, mitochondria marker; PST1,peroxisome marker; WxTP, plastid marker. (Scale bars: A, 5 μm; and B, 1 μm.)

4 of 6 | www.pnas.org/cgi/doi/10.1073/pnas.1715570114 Wang et al.

Dow

nloa

ded

by g

uest

on

Oct

ober

24,

202

0

in bread wheat (Fig. 1A). Thus, it would be interesting to knowwhether MS1 plays a similar role in other Poaceae plants, espe-cially rice and corn. Additionally, Ms1 contains two domains thatare each conserved across angiosperms (Fig. 2C); the N terminusof Ms1 is predicted to contain an LTP domain, which is requiredfor binding lipids (Fig. 6). However, compared with most plantLTPs (30), that of Ms1 is bigger (20 kDa compared with 7–10 kDafor most LTPs) as a result of the presence of the C terminus (SIAppendix, Fig. S8). Thus, as new architectures arose inMS1 and itsencoded protein during the evolution of Poaceae plants, newbiochemical and biological functions were acquired. A pointmutation in the C-terminal region of Ms1 (ms1j, S195F; Fig. 1H

and SI Appendix, Table S5) completely abolished the function ofMs1 in wheat, indicating that the C terminus of Ms1 is requiredfor proper protein function. Further study is needed to reveal therole of the C-terminal portion of Ms1 in male fertility.Heterosis in wheat was first reported approximately 100 y ago

(31); since then, effort has been made to develop a hybrid wheatseed production system. Compared with its parents, the yield fromhybrid wheat is as much as 30% greater (7, 8). Wheat is mono-clinous and autogamous; therefore, the effort required for me-chanical emasculation is prohibitive, and male sterility providesthe best and most practical method for blocking self-fertilizationduring hybrid seed production. Cytoplasmic male sterility (CMS)

Fig. 5. The signal peptide and transmembrane domain of Ms1 are required for proper subcellular localization in plastids and mitochondria. (A) Schematicrepresentations of the full-length and truncated Ms1 proteins used in our subcellular localization analysis. LTP_2, predicted LTP domain; SP, signal peptide;TM, transmembrane domain. (B–D) Colocalization of full-length or truncated Ms1 fused GFP with the mitochondrial marker pFAγ (B), plastid marker WxTP (C),or ER marker SP-mCherry-HDEL (D) in onion epidermal cells. (Scale bar: 5 μm.)

Fig. 6. Ms1 binds phospholipids in vitro. PurifiedMBP-Ms1-His or MBP-His (A and B), MBP-Ms1N-His, andMBP-Ms1C-His (C) were overlaid in an Echelon P-6001 lipidstrip (A and C) or Echelon P-6002 membrane strip (B). CL, cardiolipin; Cs, cholesterol; DAG, diacylglycerol; LPA, lysophosphatidic acid; LPC, lysophosphocholine; PC,phosphatidylcholine; PE, phosphatidylethanolamine; PG, phosphatidylglycerol; PI, phosphatidylinositol; PI(3)P, phosphatidylinositol-3-phosphate; PI(4)P, phos-phatidylinositol-4-phosphate; PI(5)P, phosphatidylinositol-5-phosphate; PI(3,4)P2, phosphatidylinositol-3,4-bisphosphate; PI(3,5)P2, phosphatidylinositol-3,5-bisphosphate; PI(4,5)P2, phosphatidylinositol-4,5-bisphosphate; PI(3,4,5)P3, phosphatidylinositol-3,4,5-triphosphate; PS, phosphatidylserine; SM, sphingomyelin;S1P, sphingosine-1-phosphate; TAG, triacylglycerol.

Wang et al. PNAS Early Edition | 5 of 6

PLANTBIOLO

GY

Dow

nloa

ded

by g

uest

on

Oct

ober

24,

202

0

was developed in 1951 by introducing Aegilops caudata cytoplasminto bread wheat (32); since then, cytoplasmically induced malesterile lines have been used as part of a three-line system for hy-brid wheat seed production (27, 33). To date, more than 70 dif-ferent male sterile cytoplasm systems have been reported in wheat(7). Chemically induced male sterility (using chemical hybridizingagents [CHAs]) has also been used to develop hybrid wheat sys-tems (34); currently, more than 40 chemicals are available aspotential CHAs (7, 8). The third approach used to develop hybridwheat is the environmentally sensitive GMS-dependent two-linesystem (e.g., photoperiod- and temperature-induced GMS) (35,36). However, the CMS system lacks effective fertility-restoringgenes, CHA suffers from problems with toxicity and selectivity,and the photoperiod/thermosensitive system is not stable (7, 9,37). These limitations have severely impeded the practicality ofthese systems for wheat hybrid seed production and may explainwhy only approximately 0.2% of all wheat acreage is currentlyhybrid wheat. However, the limitations inherent to these systemscan be overcome by the use of a nonconditional GMS system (9,12, 23), which was recently adapted for another monoclinous crop,rice (5). The basic requirements for a GMS system are corre-sponding mutations and nuclear genes controlling male sterility.

In this study, we cloned a wheat nuclear-recessive gene affectingmale sterility; therefore, our work has made it feasible to generatea stable and commercially viable hybrid seed production system inbread wheat.

Materials and MethodsBread wheat (Triticum aestivum L.), diploid ancestral species of T. urartu andA. tauschii, and allotetraploid T. turgidum were used in this study. Themutants ms1d.1 andms1e were obtained from the Wheat Genetics ResourceCenter at Kansas State University. We screened ms1d.2 and ms1h-p from anEMS-mutagenized population of bread wheat (variety Ningchun 4). Detailsof experimental procedures, such as cloning of Ms1, qPCR, in situ hybrid-ization, Southern blot, Western blot, DNA methylation analysis, protein-lipidoverlay assay, preparation of Ms1-specific antibodies, RNA-seq, resequenc-ing and bioinformatics processing of the sequence data, microscopic analy-ses, and subcellular localization of Ms1, are described in SI Appendix, SIMaterials and Methods.

ACKNOWLEDGMENTS. We thank Dr. Jessica Habashi and Prof. WenhuaZhang for critical reading of the manuscript. This work was supported byNational Transgenic Science and Technology Program of China Grant2010ZX08010-003 (to X.W.D.) and in part by Beijing Municipal GovernmentScience Foundation Grant CIT&TCD20150102 (to L.M.) and Peking-TsinghuaCenter for Life Sciences, Peking University, (X.W.D.).

1. Peng J, et al. (1999) ‘Green revolution’ genes encode mutant gibberellin responsemodulators. Nature 400:256–261.

2. Pingali PL (2012) Green revolution: Impacts, limits, and the path ahead. Proc Natl AcadSci USA 109:12302–12308.

3. Tilman D, Cassman KG, Matson PA, Naylor R, Polasky S (2002) Agricultural sustain-ability and intensive production practices. Nature 418:671–677.

4. Foley JA, et al. (2011) Solutions for a cultivated planet. Nature 478:337–342.5. Chang Z, et al. (2016) Construction of a male sterility system for hybrid rice breeding

and seed production using a nuclear male sterility gene. Proc Natl Acad Sci USA 113:14145–14150.

6. Deng XW, et al. (2013) Hybrid rice breeding welcomes a new era of molecular cropdesign. Scientia Sinica Vitae 43:864–868.

7. Singh SK, Chatrath R, Mishra B (2010) Perspective of hybrid wheat research: A review.Indian J Agric Sci 80:1013–1027.

8. Whitford R, et al. (2013) Hybrid breeding in wheat: Technologies to improve hybridwheat seed production. J Exp Bot 64:5411–5428.

9. Perez-Prat E, van Lookeren Campagne MM (2002) Hybrid seed production and thechallenge of propagating male-sterile plants. Trends Plant Sci 7:199–203.

10. Longin CF, et al. (2012) Hybrid breeding in autogamous cereals. Theor Appl Genet125:1087–1096.

11. Singh SP, Srivastava R, Kumar J (2015) Male sterility systems in wheat and opportu-nities for hybrid wheat development. Acta Physiol Plant 37:1713.

12. Albertsen MC, et al.; inventors (2006) Nucleotide sequences mediating plant malefertility and method of using same. US Patent US WO2007002267.

13. Wu Y, et al. (2016) Development of a novel recessive genetic male sterility system forhybrid seed production in maize and other cross-pollinating crops. Plant Biotechnol J14:1046–1054.

14. Pugsleay T, Oram RN (1959) Genic male sterility in wheat. Aust Plant Breed GenetNewsletter 14:10–11.

15. Fossati A, Ingold M (1970) A male sterile mutant in Triticum aestivum. Wheat Inf Serv(Kyoto) 30:8–10.

16. Sasakuma T, Maan SS, Williams ND (1978) EMS-induced male-sterile mutants in eu-plasmic and alloplasmic common wheat. Crop Sci 18:850–853.

17. Driscoll CJ (1977) Registration of Cornerstone male-sterile wheat germplasm. Crop Sci17:190.

18. Zhou K, Wang S, Feng Y, Ji W, Wang G (2008) A new male sterile mutant LZ in wheat(Triticum aestivum L.). Euphytica 159:403–410.

19. Klindworth DL, Williams ND, Maan SS (2002) Chromosomal location of genetic malesterility genes in four mutants of hexaploid wheat. Crop Sci 42:1447–1450.

20. Maan SS, Carlson KM, Williams ND, Yang T (1987) Chromosomal arm location andgene centromere distance of a dominant gene for male sterility in wheat. Crop Sci27:494–500.

21. Maan SS, Kianian SF (2001) Third dominant male sterility gene in common wheat.Wheat Inf Serv 93:27–31.

22. Qi LL, Gill BS (2001) High-density physical maps reveal that the dominant male-sterilegene Ms3 is located in a genomic region of low recombination in wheat and is notamenable to map-based cloning. Theor Appl Genet 103:998–1006.

23. Ni F, et al. (2017) WheatMs2 encodes for an orphan protein that confers male sterilityin grass species. Nat Commun 8:15121.

24. Xia C, et al. (2017) A TRIM insertion in the promoter of Ms2 causes male sterility inwheat. Nat Commun 8:15407.

25. Abe A, et al. (2012) Genome sequencing reveals agronomically important loci in riceusing MutMap. Nat Biotechnol 30:174–178.

26. Endo TR, Mukai Y, Yamamoto M, Gill BS (1991) Physical mapping of a male-fertilitygene of common wheat. Jpn J Genet 66:291–295.

27. Franckowiak JD, Maan SS, Williams ND (1976) A proposal for hybrid wheat utilizingAegilops squarrosa L. cytoplasm. Crop Sci 16:725–728.

28. International Wheat Genome Sequencing Consortium (IWGSC) (2014) A chromosome-based draft sequence of the hexaploid bread wheat (Triticum aestivum) genome.Science 345:1251788.

29. Chapman JA, et al. (2015) A whole-genome shotgun approach for assembling andanchoring the hexaploid bread wheat genome. Genome Biol 16:26.

30. Benning C (2009) Mechanisms of lipid transport involved in organelle biogenesis inplant cells. Annu Rev Cell Dev Biol 25:71–91.

31. Freeman GF (1919) Heredity of quantitative characters in wheat. Genetics 4:1–93.32. Kihara H (1951) Substitution of nucleus and its effects on genome manifestations.

Cytologia (Tokyo) 16:177–193.33. Wilson JA (1984) Hybrid wheat breeding and commercial seed development. Plant

Breed Rev 2:303–319.34. Moore RH (1950) Several effects of maleic hydrazide on plants. Science 112:52–53.35. Murai K, Tsunewaki K (1993) Photoperiod-sensitive cytoplasmic male sterility in

wheat with Aegilops crassa cytoplasm. Euphytica 67:41–48.36. Qian CM, Xu A, Uiang GH (1986) Effects of low temperatures and genotypes on

pollen development in wheat. Crop Sci 26:43–46.37. Chen G, Gong D, Guo X, Qiu W, He Q (2005) Problems of the hybrid with Chongqing

thermo-photo-sensitive male sterility wheat C49S in the plain of Jiang Han. MaileiZuowu Xuebao 25:147–148.

6 of 6 | www.pnas.org/cgi/doi/10.1073/pnas.1715570114 Wang et al.

Dow

nloa

ded

by g

uest

on

Oct

ober

24,

202

0