WANG Yu-peng, TANG shuang-qin, WU Zhi-feng, SHI Qing-hua, WU Zi-ming

Key Laboratory of Crop Physiology, Ecology and Genetic Breeding, Ministry of Education/Collaborative Innovation Center for the Modernization Production of Double Cropping Rice/College of Agronomy, Jiangxi Agricultural University, Nanchang 330045,P.R.China

1. Introduction

Rice is an important food crop that feeds over half of the world’s population, and it is also one of the most important monocot model systems because of characteristics such as a small sequenced genome and a matureAgrobacteriummediated genetic transformation system. Rice production is correlated with multiple factors, such as tiller number, grain number per panicle, grain size, and plant height. Plant height,which is established during the vegetative growth periods,plays a crucial role in plant architecture, photosynthetic efficiency, nutrient absorption, and lodging resistance (Krag and Nielsen 1989; Wanget al. 2016). Generally, taller plants are advantageous for maintaining a higher photosynthetic efficiency, but have reduced lodging tolerance; dwarf plants have increased lodging tolerance, but severe dwarfism affects the development of other organs, which is not conducive to the improvement of rice yield (Dinget al. 2015; Zhaoet al. 2015;Zhanget al. 2016). The appropriate plant height, therefore,is a critical factor in improving rice yield.

During the last 20 years, considerable progress has been made in elucidating the regulatory mechanisms controlling plant height in rice. Previous studies have shown that plant hormones such as gibberellin (GAs) (Daviere and Achard 2013), brassinolide (BRs) (Yeet al. 2011), and strigolactone(SLs) (Screpantiet al. 2016), and genes associated with cell wall synthesis, cellular differentiation, and cell elongation(Yanget al. 2011), are the main factors controlling plant height. Mutations in these genes inhibit the development of stem internodes, leading to a dwarf phenotype. So far, the genes known to be related to plant dwarfism are mainly involved in GA, BR, and SLs biosynthesis and signal transduction. Additionally, some genes that are not associated with hormone metabolic pathways, such asOsKinesin-13A(Denget al. 2014),OSH15(Satoet al.1999),OsGLP1(Banerjee and Maiti 2010), andDTH8(Weiet al. 2010), have been found to in fluence plant height.

Although determined during the vegetative growth stage,plant height has a significant effect on reproductive traits,such as heading date, flower development, and rice quality(Songet al. 2008; Duanet al. 2012a, b). Many dwarf mutants are associated with the abnormal flower development. For example, a mutation inDDF1, which encodes an F-box protein, leads to much shorter plant height compared with wild type and significant structural abnormalities in florets (Duanet al. 2012b). Theddf2mutant has a severe dwarf phenotype and defects in spikelet and floral organs (Zhanget al. 2015).DEFORMED FLORAL ORGAN1(DFO1) was identified as a rice epigenetic repressor. Thedfo1mutant has a small stature and defects in palea identity (Zhenget al. 2015). Although a few genes have been reported to control both plant height and floral organ development, flower-deformation/dwarf genes remain largely undiscovered.

To gain further insight into the relationship between plant height and flower development and the molecular basis for these phenotypes, we identified and characterized thedwarf and deformed flower 3(ddf3) mutant, which in addition to a severe dwarf phenotype, exhibits pollen abortion and pistils with multiple stigmas. The mutant gene was mapped between insertion-deletion (InDel) markers M15 and M16 on the long arm of chromosome 7, which corresponds to an approximate physical distance of 45.21 kb. Sequencing revealed a 13.98-kb deletion in theddf3mutant genome,which caused the deletion of three genes. Our work has laid the foundation for further gene cloning and functional analysis ofDDF3.

2. Materials and methods

2.1. Plant materials and growth conditions

A dwarf mutant, which was nameddwarf and deformed flower3(ddf3), was identified in the progeny of anOryza sativacultivar Dongjin plant derived from tissue culture transformation. We con firmed that the mutant phenotypes were stably inherited after five years of cultivation in natural condition. To map the mutant locus, we generated an F2mapping population derived from a cross between theddf3mutant and the cultivar Yangdao 6 (9311). For phenotypic characterization, microscopic observation, analysis of floral organ development, mapping, and qRT-PCR assays,ddf3,wild-type (WT) Dongjin and the F2mapping population were cultivated in an experimental field at Nanchang (28°41′N,115°55′E) during the normal growing season.

2.2. Phenotypic characterization

At the heading stage, plant height, leaf length, leaf width,and internode length were measured in the mutant and WT.Yield-related traits, including panicle length, no. of grains per panicle, seed-setting rate, grain size, and 1 000-grain weight, were measured at harvest. All trait measurements are presented as the average from 10 plants.

2.3. Histological analysis of internodes

At the heading stage, the middle sections of the second internodes were collected from theddf3mutant and WT and then fixed in FAA (formalin:acetic acid:50% ethanol, 2:1:17(v/v)) overnight. After a series of dehydration and in filtration steps, the tissues were embedded in paraffin (Paraplast Plus; Sigma-Aldrich). The embedded tissues were cut into 8 μm sections with a microtome (Leica RM2265, Germany).Then the paraffin was removed from the sections with xylene. This was followed by a dehydration through an ethanol gradient, and toluidine blue staining. The stained sections were observed and photographed with a Nikon 50i microscope (Nikon, Japan).

2.4. Analysis of pollen activity and stigma development

The I2-KI staining method was used to detect pollen activity as described previously (Jianget al. 2007). Briefly, the anthers of the mutant and WT were crushed on a slide in a drop water and then stained with I2-KI solution. In this study, pollen type includes normal and abortion, and abortion type is divided into typical abortion, spherical abortion, and stained abortion type (Li 1980; Xianget al. 2011). Typical aborted pollen grains are shrunken and unstained. Spherical aborted pollen grains appear circular but cannot be stained.Stained aborted pollen grains appear circular and weakly stained. A Nikon 50i microscope (Nikon) was used to observe the stained pollen grains and count the number of different pollen types. To analyze stigma development, the pistils of the mutant and WT were observed under a stereo microscope (Nikon SMZ800, Japan).

2.5. Map-based cloning of the mutant gene

At harvest, the segregation of mutant traits was determined by counting the rate of plants with WT and mutant phonotypes in the F2mapping population. Rice leaf genomic DNA was extracted using the CTAB method(Murray and Thompson 1980). Among 512 simple sequence repeat (SSR) markers, 242 were found to be polymorphic between the mutant and 9311. Bulked segregant analysis (BSA) was used for rough mapping of the mutant locus (Zhanget al. 1994). New InDel markers were developed using primer 5.0 based on the gaps in sequence between thejaponicaNipponbare variety and theindicavariety 9311, which were found on the gramene website (http://www.gramene.org/). Gene sequences and gene models were obtained from RGAP7 (http://rice.plantbiology.msu.edu/cgi-bin/gbrowse/rice/). The candidate genes were identified by PCR ampli fication and sequence analysis.

2.6. qRT-PCR

The relative expression levels of the candidate genes in 20-day-old mutant and WT seedlings were compared.The relative expression levels of the candidate genes in different tissues were determined in WT plants at the heading stage. Total RNA was extracted using the MiniBEST Plant RNA Extraction Kit (TaKaRa, Japan).Then, cDNA was synthesized from 1 μg RNA with PrimeScript? RT Reagent Kit with gDNA Eraser (TaKaRa)according to manufacturer’s instructions. qRT-PCR was carried out in a Bio-Rad CFX96 Touch with a 25-μL reaction volume, which included 2 μL cDNA, 0.2 μmol L–1gene-specific primers, ddH2O, and SYBR?PremixEx Taq? (TaKaRa) reagent. The rice ubiquitin gene was used as an internal control. The 2–ΔΔCTmethod was used to calculate the relative expression (Yuet al. 2007).

3. Results

3.1. Phenotypic characterization of the ddf3 mutant

Compared with WT, theddf3mutant displayed reduced height and increased tillering; the plant height ofddf3was 69.23% shorter, and tiller number was 56.58% higher(Fig. 1-B and C). In addition, panicle length, leaf length,and grain shape, were shorter inddf3(Fig. 1-D, E, and H). Interestingly, the lengths of the flag leaf, the 2nd leaf,the 3rd leaf, and the grain inddf3were significantly shorter compared with those in WT, but there was no significant difference in leaf width, grain width, and grain thickness(Fig. 1-F, G, and I). These results indicate thatDDF3may not influence the transverse development of organs. We further investigated yield-related traits. We found that panicle length, number of grains per panicle, seed-setting rate, and 1 000-grain weight were reduced by 40.96, 53.87,61.78, and 27.49%, respectively, inddf3compared with WT(Appendix A).

3.2. The ddf3 mutant exhibits severe defects in cell elongation

To investigate the reason whyddf3plants show a severely dwarfed phenotype, we measured the lengths of each internode in WT andddf3. We found that all internodes were significantly shorter inddf3compared with WT (Fig. 2-A and B). Analysis of longitudinal sections revealed that the lengths of the cells inddf3internodes were significantly shorter than those in WT, whereas there was no significant difference in internode cell width (Fig. 2-C–F). Taken together, these results suggest that theDDF3mutation inhibits cell elongation, but might not affect the transverse development of internode cells.

3.3. Mutation of DDF3 affects pollen activity and stigma development

In order to determine why the seed-setting rates ofddf3were lower than those in WT, we compared pollen activity inddf3and WT. We found that pollen activity was significantly lower in theddf3mutant than that in WT and that there was an abundance of abnormal pollen inddf3(Fig. 3-A and B).Analysis of pollen types revealed thatddf3had a great deal of abortive pollen. The aborted pollen grains were mainly spherical and stained abortion types, although some were typical abortion types (Fig. 3-C). The pistil is an important reproductive organ, and the normal development of pistils is critical for obtaining good seed-setting rates. Therefore, we wanted to know whether pistil development was normal in theddf3mutant. Interestingly, we found that many pistils inddf3showed a multi-stigma phenotype (Fig. 3-D and E), and in some pistils, it appeared that the stigma didn’t elongate(data not shown). These results suggest that mutation ofDDF3affects pollen activity and stigma development.

Fig. 1 Comparison of phenotype characteristics between ddf3 and wild-type (WT) plants. A, WT and ddf3 plants during the heading stage. B and C, plant heights (B) and tiller numbers (C) statistical analysis. D, panicle lengths of WT and ddf3. E, leaf characteristics of WT and ddf3. F and G, comparison of leaf lengths (F) and leaf widths (G) between ddf3 and WT. H, phenotypes of grain shapes in WT and ddf3. I, grain lengths, grain widths, and grain thicknesses of WT and ddf3. Values are mean±SD (B,C, F, and G, n=10 plants; I, n=30 replicates). The Student’s t-test analysis indicates a significant difference (compared with WT,＊＊, P＜0.01). Scale bars, 10 cm (A, D, and E) and 1 cm (H).

3.4. There is a multi-gene deletion in the ddf3 mutant

To identify theDDF3locus, we evaluated the phenotypes of F1plants and an F2mapping population derived from a cross betweenddf3and 9311. All F1plants displayed thick and tall stems, broad leaves, and low seed-setting rates.Furthermore, the segregation ratio of WT to mutant plants in the F2population was 3:1 (normal:mutants=529:150;χc

2=3.06,P＞0.05). These results indicated that the mutant traits ofddf3are controlled by a single or tightly linked nuclear gene. TheDDF3locus was roughly mapped to a 3.3-Mb region between two SSR markers, RM6403 and RM234, on the long arm of chromosome 7 (Fig. 4-A and B). Using 663 mutant plants in the F2population, theDDF3locus was fine mapped to a 45.21-kb region between the InDel markers M15 and M16 (Fig. 4-C). According to the RGAP7 gene predictions, there are six genes in this region (Fig. 4-D).

To identify the gene responsible for theddf3mutant phenotype, we amplified all six genes located between markers M15 and M16 using gene-specific primers.Interestingly, only three genes could be successfully amplified from both WT andddf3. Sequence analysis revealed that the sequences of these three genes were the same in WT andddf3. However, the other three genes,which encode a zinc finger, DHHC-type domain containing protein (ZF), an expressed protein (EP), and an actinbinding FH2 domain containing protein (FH2), respectively,could be amplified from WT but could not be completely amplified fromddf3(Fig. 5-A and B). We next evaluated the expression of these genes at seedling stage inddf3and WT using qRT-PCR. This assay suggested that compared with WT, the expression ofZFinddf3was increased about 15-fold, while the expression ofFH2could not be detected inddf3(Fig. 5-C). Considering the direction and states of these three genes, we inferred that there was a large deletion in this region. Then we amplified genomic DNA fragment deletion by designing primers (P0) at both ends of the region that couldn’t be amplified inddf3. Consistent with our inference, this region contained a gene large deletion inddf3, compared with WT (Fig. 5-D). Sequence analysis revealed a 13.98-kb deletion in theddf3genomic DNA.These results suggest that theddf3mutant is caused by a multi-gene deletion, and the three genes in the deleted region areDDF3candidate genes.

Fig. 2 Morphological characteristics and histological characterization of internodes of wild type (WT) and ddf3. A, internodes of WT and ddf3. B, comparison of internode lengths between WT and ddf3 plants. C and D, longitudinal section analysis of WT (C)and ddf3 (D) in second internode (internode II). E and F, the statistical analysis of parenchyma (PC) cell length (E) and width (F)in internode II. Scale bars, 20 cm (A) and 100 μm (C and D). Values are mean±SD (B, n=10 plants; E and F, n=30 replicates).The Student’s t-test analysis indicates a significant difference (compared with WT, ＊＊, P＜0.01). NS, not significant.

Fig. 3 The pollen activity and stigma development state of wild type (WT) and ddf3. A and B, the pollen activity of WT (A) and ddf3 (B). C, the abortive pollen types of WT and ddf3. The Student’s t-test analysis indicates a significant difference (compared with WT, ＊＊, P＜0.01). Values are mean±SD (n=5 plants). D and E, stigma development state of WT and ddf3. Scale bar, 2 mm.

Fig. 4 The molecular mapping of the DDF3 gene. A and B, rough mapping of DDF3. C, fine mapping of ddf3. D, gene predication in fine mapping region of DDF3.

Fig. 5 Identification of the mutant site of DDF3. A and B, the amplified fragments of ZF, EP, and FH2 genes in wild type (WT,A) and ddf3 (B). ZF-1 and ZF-2, the 1st and 2nd fragments of ZF; FH2-1 and FH2-2, the 1st and 2nd fragments of FH2. C, the relative expression of ZF, EP, and FH2 between WT and ddf3. D, the amplified fragment around the deletion region in WT and ddf3. Values are mean±SD (B, n=3 plants). The Student’s t-test analysis indicates a significant difference (compared with WT,＊＊, P＜0.01). Red arrows indicate the DNA fragments which can be successfully amplified in WT genome but can’t be amplified in ddf3 genome.

3.5. ZF and FH2 could be involved in the regulation of rice morphogenesis or flower organ development

To further understand the effects of mutating eachDDF3candidate gene, we first analyzed which parts of these genes have been deleted. According to the RGAP7 gene prediction models, WTZFhas nine exons and eight introns, but the promoter, two exons and some parts of introns have been deleted inddf3. In WT,EPonly has two exons and one intron, and this gene was completely deleted inddf3. WTFH2has 15 exons and 14 introns, but inddf310 exons, 9 introns, and the 3′ untranslated region have been completely deleted (Fig. 6-A). These results indicate these three genes are completely or partly nonfunctional inddf3.

The functions ofZFandEPare not known, butFH2is an allele ofBENT UPPERMOST INTERNODE1(BUI1)/RICE MORPHOLOGY DETERMINANT(RMD) (Yanget al. 2011;Zhanget al. 2011), which has been reported to be involved in the regulation of cell development. To gain insight into the roles of these genes in rice plant development, we investigated their expression patterns in different organs.We found that the expression ofZFwas the highest in panicles, butZFexpression was very low in stems, leaves,and roots (Fig. 6-B).EPwas not expressed in any of these organs, whereas the expression ofFH2was relatively high in all four organs (Fig. 6-C). These results indicated thatZFandFH2could potentially be involved in the regulation of rice morphogenesis or flower organ development, whileEPmight not function at the heading stage.

4. Discussion

4.1. ddf3 is a novel flower-deformation dwarf mutant

In this present study, we found a novel flower-deformation dwarf mutant,ddf3. The gene(s) responsible for theddf3mutant phenotype is(are) located in an interval of 45.21 kb on the long arm of chromosome 7. There was a 13.98-kb mutation in theddf3mutant genome in this region, in which three genes are deleted. This indicates that theddf3mutant phenotype is due to the loss of function of one or more of these deleted genes. Further transgenic complementation of the mutant and gene knockout experiments in WT will allow the responsible gene(s) to be identified.

4.2. DDF3 gene(s) affects the seed-setting rate by regulating the development of pollen and stigma

Pollen abortion is divided into three types, typical abortion,spherical abortion, and stained abortion, based on the stage when abortion occurs. Typical abortion occurs during the early stages of pollen development and pollen generally only develop to the single nucleus pollen stage; spherical abortion usually occurs during the dual core pollen stage;stained abortion type mainly happens during the trinucleate pollen stage (Li 1980; Xianget al. 2011). The pollen abortion types observed in theddf3mutant were mainly the spherical abortion and stained abortion types (Fig. 3-C), indicating that the mutated gene(s) mainly affect the dual core and trinucleate stages of pollen development. In addition to pollen defects, many pistils in the mutant showed a multistigma phenotype (Fig. 3-D and E), and in some pistils the stigma didn’t elongate. Therefore, it is very likely that theDDF3gene(s) affects the seed-setting rate by regulating the development both of pollen and stigma.

Fig. 6 Diagram of deletion segment and analysis of mutant genes relative expression. A, diagram of deletion segment in genome.B and C, the analysis of ZF (B) and FH2 (C) relative expression in different organs of wild type (WT). Values are mean±SD (n=3).–, means the site of the primer;or mean the direction of the primer.

4.3. FH2 (BUI1/RMD) may be an important factor regulating morphological and flower development

BUI1/RMD, which encodes the Class II Formin Homology5(FH5), controls the rearrangements of the actin cytoskeleton and the spatial organization of actin filaments and plays a crucial role in proper cell expansion and rice morphogenesis(Yanget al. 2011; Liet al. 2014). Thebui1/rmdmutant, as well theddf3mutant reported in this study, display dwarfism,short leaves, a decrescent grain shape, and blocked cell elongation. Additionally,ArabidopsisFH5 was reported to facilitate pollen tube growth, and RNAi suppression ofFH5expression inNicotiana tabacumled to diminished subapical actin structure and compromised tip-focused growth of pollen tubes (Cheunget al. 2010). In this study,there was a complete deletion of 10 exons, 9 introns, and the 3′ untranslated region ofFH2(BUI1/RMD) in theddf3mutant (Fig. 6-A), resulting in the deletion of most of the conserved FH2 functional domain.FH2(BUI1/RMD) was relatively highly expressed in the panicle, stem, leaf, and root(Fig. 6-C). Therefore,FH2(BUI1/RMD) may be an important factor regulating morphological and flower development,and loss ofFH2may explain the morphological and floral defects in theddf3mutant.

4.4. ZF may be responsible for the morphogenic defects and deformed floral organs observed in ddf3

ZFencodes an S-acyltransferases (PATs) with a conserved Asp-His-His-Cys (DHHC) motif. The promoter, two exons and some parts of introns are deleted in theddf3mutant(Fig. 6-A). Evidence from previous studies suggests that S-acyltransferases are involved in cell expansion, pollen tube growth, ovule fertilization, and release of pollen inArabidopsis(Hemsleyet al. 2005; Chaeet al. 2009; Qiet al.2013). SeveralOsDHHCshave also been reported to play important roles in plant growth and development, and stress responses in rice (Liet al. 2016; Zhouet al. 2017). We found thatZFwas expressed in the stem, leaf, and root and was especially highly expressed in the panicle (Fig. 6-B).Therefore, loss ofZFfunction may also be responsible for the morphogenic defects and deformed floral organs observed in the mutant.

4.5. EP may not be involved in the regulation of plant development during the heading stage

EPencodes an expressed protein, and its function is unclear.It was not expressed in any tissue, which indicates thatEPmay not be involved in the regulation of plant development during the heading stage or there is the possibility of gene annotation wrong. However, future studies should also focus on investigating the role ofEPat other stages of plant development.

5. Conclusion

The mutant gene,DDF3,is involved in the regulation of cell elongation and pollen and stigma development in rice.A 13.98-kb deletion in theddf3resulted in the complete or partial deletion of three genes. Disruption of the function of one or more of these genes is the ultimate cause of theddf3mutant phenotype.

Acknowledgements

The research was supported by the National Natural Science Foundation of China (31560350 and 31760350)and the Science and Technology Program of Jiangxi, China(20171ACF60018).

Appendixassociated with this paper can be available on http://www.ChinaAgriSci.com/V2/En/appendix.htm

Banerjee J, Maiti M K. 2010. Functional role of rice germin-like protein1 in regulation of plant height and disease resistance.Biochemical and Biophysical Research Communications,394, 178–183.

Chae K, Kieslich C A, Morikis D, Kim S C, Lord E M. 2009. A gain-of-function mutation ofArabidopsislipid transfer protein 5 disturbs pollen tube tip growth and fertilization.ThePlant Cell, 21, 3902–3914.

Cheung A Y, Niroomand S, Zou Y J, Wu H M. 2010. A transmembrane formin nucleates subapical actin assembly and controls tip-focused growth in pollen tubes.Proceedings of the National Academy of Sciences of the United States of America, 107, 16390–16395.

Daviere J M, Achard P. 2013. Gibberellin signaling in plants.Development, 140, 1147–1151.

Deng Z Y, Liu L T, Li T, Yan S, Kuang B J, Huang S J, Yan C J, Wang T. 2014.OsKinesin-13Ais an active microtubule depolymerase involved in glume length regulationviaaffecting cell elongation.Scienti fic Reports, 5, 1702–1710.

Ding Z, Lin Z, Li Q, Wu H, Xiang C, Wang J. 2015.DNL1,encodes cellulose synthase-like D4, is a major QTL for plant height and leaf width in rice (Oryza sativaL.).Biochemicaland Biophysical Research Communications, 457, 133–140.

Duan Y L, Guan H Z, Ming Z, Chen Z W, Li W T, Pan R S,Mao D M, Zhou Y C, Wu W R. 2012a. Genetic analysis and mapping of an enclosed panicle mutant locusesp1in rice (Oryza sativaL.).Journal of Integrative Agriculture,11, 1933–1939.

Duan Y L, Li S P, Chen Z W, Zheng L L, Diao Z J, Zhou Y C,Lan T, Guan H Z, Pan R S, Xue Y B, Wu W R. 2012b. Dwarf and deformed flower 1, encoding an F-box protein, is critical for vegetative and floral development in rice (Oryza sativaL.).ThePlant Journal, 72, 829–842.

Hemsley P A, Kemp A C, Grierson C S. 2005. The TIP GROWTH DEFECTIVE1 S-acyl transferase regulates plant cell growth inArabidopsis.ThePlant Cell, 17, 2554–2563.

Jiang S Y, Cai M, Ramachandran S. 2007.ORYZA SATIVA MYOSIN XI Bcontrols pollen development by photoperiodsensitive protein localizations.Developmental Biology, 304,579–592.

Krag E, Nielsen P B. 1989. Analysis on difference of dry matter production between rice cultivars with different plant height in relation to gas diffusion inside stands.Japanese Journal of Crop Science, 58, 374–382.

Li G, Liang W, Zhang X, Ren H, Hu J, Bennett M J, Zhang D. 2014. Rice actin-binding protein RMD is a key link in the auxin-actin regulatory loop that controls cell growth.Proceedings of the National Academy of Sciences of the United States of America, 111, 10377–10382.

Li Y, Lin J, Li L, Peng Y, Wang W, Zhou Y, Tang D, Zhao X, Yu F,Liu X. 2016. DHHC-cysteine-rich domain S-acyltransferase protein family in rice: Organization, phylogenetic relationship and expression pattern during development and stress.Plant Systematics and Evolution, 302, 1405–1417.

Li Z B. 1980. A preliminary discussion about the classi fication of male sterile lines of rice in China.Acta Agronomica Sinica,6, 27–34. (in Chinese)

Murray M G, Thompson W F. 1980. Rapid isolation of high molecular weight plant DNA.Nucleic Acids Research, 8,4321–4325.

Qi B, Doughty J, Hooley R. 2013. A golgi and tonoplast localized S-acyl transferase is involved in cell expansion,cell division, vascular patterning and fertility inArabidopsis.New Phytologist, 200, 444–456.

Sato Y, Sentoku N, Miura Y, Hirochika H, Kitano H, Matsuoka M. 1999. Loss-of-function mutations in the rice homeobox geneOSH15affect the architecture of internodes resulting in dwarf plants.The EMBO Journal, 18, 992–1002.

Screpanti C, Fonne-Pfister R, Lumbroso A, Rendine S, Lachia M, De Mesmaeker A. 2016. Strigolactone derivatives for potential crop enhancement applications.Bioorganic &Medicinal Chemistry Letters, 26, 2392–2400.

Song L K, Lee S, Kim H J, Hong G N, An G. 2008.OsMADS51is a short-day flowering promoter that functions upstream ofEhd1,OsMADS14, andHd3a.Plant Physiology, 145,1484–1494.

Wang J X, Jian S, Li C X, Liu H L, Wang J G, Zhao H W, Zou D T. 2016. Genetic dissection of the developmental behavior of plant height in rice under different water supply conditions.Journal of Integrative Agriculture, 15, 2688–2702.

Wei X, Xu J, Guo H, Jiang L, Chen S, Yu C, Zhou Z, Hu P,Zhai H, Wan J. 2010.DTH8suppresses flowering in rice,influencing plant height and yield potential simultaneously.Plant Physiology, 153, 1747–1758.

Xiang X C, Zhang X H, Xu Y F, Long X L, Zhang Y L. 2011.Identification of a new sterile cytoplasmic fertility line H236A fromindicaMiyang 46.Journal of Plant Genetic Resources,12, 1014–1018.

Yang W, Ren S, Zhang X, Gao M, Ye S, Qi Y, Zheng Y, Wang J, Zeng L, Li Q. 2011.BENT UPPERMOST INTERNODE1encodes the class II formin FH5 crucial for actin organization and rice development.The Plant Cell, 23, 661–680.

Ye H, Li L, Yin Y. 2011. Recent advances in the regulation of brassinosteroid signaling and biosynthesis pathways.Journal of Integrative Plant Biology, 53, 455–468.

Yu S, Liu H, Luo L. 2007. Analysis of relative gene expression using different real-time quantitative PCR.Acta Agronomica Sinica, 33, 1214–1218. (in Chinese)

Zhang L, Guo S, Wang L, Zhang T Q, Zhuang H. 2015. Gene mapping and candidate gene analysis of adwarf and deformed flower 2(ddf2) mutant in rice (Oryza sativa).Scientia Agricultura Sinica, 48, 1873–1881. (in Chinese)

Zhang Q, Shen B Z, Dai X K, Mei M H, Maroof M A S, Li Z B.1994. Using bulked extremes and recessive class to map genes for photoperiod-sensitive genic male sterility in rice.Proceedings of the National Academy of Sciences of the United States of America, 91, 8675–8679.

Zhang Y H, Bian X F, Zhang S B, Ling J, Wang Y J, Wei X Y,Fang X W. 2016. Identiifcation of a novel gain-of-function mutant allele,slr1-d5, of rice DELLA protein.Journal of Integrative Agriculture, 15, 1441–1448.

Zhang Z, Zhang Y, Tan H, Wang Y, Li G, Liang W, Yuan Z, Hu J, Ren H, Zhang D. 2011.RICE MORPHOLOGY DETERMINANTencodes the type II formin FH5 and regulates rice morphogenesis.ThePlant Cell, 23, 681–700.

Zhao L, Tan L, Zhu Z, Xiao L, Xie D, Sun C. 2015. PAY1 improves plant architecture and enhances grain yield in rice.ThePlant Journal, 83, 528–536.

Zheng M, Wang Y, Wang Y, Wang C, Ren Y, Lv J, Peng C,Wu T, Liu K, Zhao S, Liu X, Guo X, Jiang L, Terzaghi W,Wan J. 2015.DEFORMED FLORAL ORGAN1(DFO1)regulates floral organ identity by epigenetically repressing the expression ofOsMADS58in rice (Oryza sativa).New Phytologist, 206, 1476–1490.

Zhou B, Lin J Z, Peng D, Yang Y Z, Guo M, Tang D Y, Tan X, Liu X M. 2017. Plant architecture and grain yield are regulated by the novel DHHC-type zinc finger protein genes in rice(Oryza sativaL.).Plant Science, 254, 12–21.