ZHOU Yu-qian, WANG Qin-yang, ZHAO Hai-liang, GONG Dian-ming, SUN Chuan-long, REN Xuemei, LlU Zhong-xiang, HE Hai-jun, QlU Fa-zhan

1 Crop Research Institute, Gansu Academy of Agricultural Sciences, Lanzhou 730070, P.R.China

2 National Key Laboratory of Crop Genetic Improvement/College of Plant Science and Technology, Huazhong Agricultural University, Wuhan 430070, P.R.China

1. lntroduction

Seed size play a significant role in determining maize yield production, which is strongly affected by various biotic,abiotic, and genetic factors. Epigenetic and phytohormones are also important factors affecting seed growth and development (Kesavan et al. 2013).

As a monocot, the endosperm constitutes the majority of maize kernel volume and dry matter (Lur and Setter 1993; Sabelli and Larkins 2009) and contributes to yield production. Both the embryo and the triploid endosperm are surrounded by and restricted in maternal tissue, and the nutrition level, growth, and senescence of the maternal individual will influence seed development. Interactions between the maternal integument and zygote components must be coordinated to support normal seed development(Garcia et al. 2005). Factors controlling the cell number can influence seed size (Gonzalez et al. 2012). Grain-filling conditions after cell proliferation and elongation also affect maize seed production, particularly seed mass (Wang et al.2008).

The regulation of cell proliferation influences seed size.Maize endosperm undergoes three different types of cell cycles before grain filling: The first is karyokinesis without cytokinin, the second is the normal mitosis, and the third is a process called endoreduplication. When the nuclear division stage last longer or when syncytium cellularization starts later, maternal individuals tend to produce larger seeds than normal. OsTGW6 encodes a protein with indole-3-acetic (IAA)-glucose hydrolase activity; the loss of function of OsTGW6 can delay the timing of the transition from the syncytial to the cellular phase and eventually lead to a higher grain weight (Ishimaru et al. 2013). Early cellularization in Arabidopsis can also cause smaller seeds at maturity,although the endosperm is only a temporal tissue. All of the three small-seed mutants of the Arabidopsis IKU pathway- iku1, iku2, and miniseed3- are accompanied by the early cellularization of the endosperm (Luo et al. 2005;Wang et al. 2010).

Changes in the grain-filling stage after mitosis in seed development can also influence yield production, especially grain weight. OsGIF1 is a maize CWIN2 homologue that is required for carbon partitioning during early grain filling(Wang et al. 2008). The gif1 mutant in rice showed a reduced grain weight although not so severely as maize mn1. Sekhon et al. (2014) examined the transcriptional and developmental changes during the seed development of two divergently selected large- and small-seed populations.They found that the two populations showed great differences in the grain-filling stage.

Plant assimilative organ or leaf senescence can also influence yield production. Weber et al. (1996) reported that large seeds always mature later, whereas small-seeds usually undergo early maturity (Weng et al. 2008). Leaf senescence can impact crop production by either changing the photosynthesis duration or modifying the nutrient remobilization efficiency and harvest index (Wu et al. 2012).OsFBK12 encodes an F-box protein containing a Kelch repeat motif that can target OsSAMS1 for degradation and trigger changes in the ethylene level for leaf senescence.When OsFBK12 is over-expressed, more assimilates can be sent to the pedicles, increasing not only the grain number per pedicles but also the rice seed size (Chen et al. 2013).The arf2/mnt mutant in Arabidopsis, characterized by a larger seed size, also showed later leaf senescence than that of the wild type (Schruff et al. 2006).

Hormones play an important role in seed development.Cytokinin (CK) and auxin have great influence on seed development. CK accumulation is correlated with mitosis activity in maize (Lur and Setter 1993). CK concentration peaks at 9 days after pollination (DAP) when the endosperm cell division rate is at its highest level (Lur and Setter 1993).IAA, the most predominant auxin in planta (Jensen and Bandurski 1994), increases when CK is decreasing and when the endosperm begins endoreduplication. The CK/IAA ratio plays a critical role in the mitosis-to-endoreduplication/storage accumulation transition (Lur and Setter 1993).Moreover, IAA promotes storage accumulation, especially beta- and gamma-zein. Recently, several genes related to the biosynthesis/perception of gibberellin (GA) have been reported to influence seed size. OsSGL encodes a kinesinlike protein that is involved in regulating GA synthesis and response; mutants of sgl have a decreased cell elongation and exhibit reduced grain length and plant height in rice(Wu et al. 2014).

In an earlier work (Liu et al. 2014), a population derived from the large-seed inbred line V671 and the small-seed line Mc was established, and QTL mapping for the controlling locus of seed size was also completed. To determine why the two inbred lines showed such differently sized seed, a phenotypic investigation and transcriptome analysis were conducted in the present study; we propose a hypothesis similar to that of Sekhon et al. (2014) that changing the time course of the mitosis-to-storage accumulation transition can increase the seed size in maize.

2. Materials and methods

2.1. Plant growth condition and phenotype evaluation

V671 and Mc were grown under natural conditions on the farm of Huazhong Agricultural University, Wuhan, China(30.35°N, 114.17°E, 30 m a.s.l.). Ears were bagged before silk emergence and self-pollinated on the same day. Seeds in the lower part of the pollination ears were harvested from at least six individuals at serial days after pollination, ground in liquid nitrogen immediately, and stored at –80°C for RNA extraction. Other age-matched seeds (6–42 DAP) were harvested, mixed together, and used for total zein extraction and 10-kernel width measurement.

2.2. RNA isolation and RNA-seq analysis

Kernels were collected from both Mc and V671 ears at 14 DAP respectively and frozen in liquid nitrogen. Total RNA was isolated using TRIzol Reagent (Invitrogen,Carlsbad, CA, USA) and then purified with RNase-free DNaseI (Invitrogen). The cDNA library construction and the sequencing of the transcriptome were performed by Tianyihuiyuan Bioscience-Technology Inc. (Beijing, China).The cDNA libraries were prepared using the Illumina Standard mRNA-seq Library Preparation Kit according to the manufacturer’s instructions and sequenced to generate 101-nucleotide paired-end reads on an Illumina HiSeq platform (Illumina, CA, USA). Clean reads were mapped to the reference genome (ftp://ftp.ensemblge-nomes.org/pub/release-30/plants/gtf/zea_mays/) using the TopHat2 software (Kim et al. 2013). The read count was converted to fragments per kilobase of transcript sequence per millions base pairs sequenced (FPKM) (Trapnell et al. 2012). The differential expression genes between the two parental lines were analyzed using the R-based DESeq Package(1.18.0). DESeq provides statistical routines for determining differential expression in digital gene expression data using a model based on a negative binomial distribution. The resulting P-values were adjusted using the Benjamini and Hochberg’s approach for controlling the false discovery rate(FDR) (Anders and Huber 2010). Genes with an adjusted P＜0.05 identified using DESeq were considered differentially expressed. A Gene Ontology (GO) analysis was conducted using the GOseq software based on the Wallenius noncentral hyper-geometric distribution (Young et al. 2010).

2.3. cDNA synthesis and quantitative RT-PCR

RNA samples at serial DAP were reverse transcribed using the SuperScriptRIII First-Strand Synthesis System(Invitrogen) according to the manufacturer’s protocol. The cDNA was diluted to the same concentration and stored at –20°C.

Quantitative RT-PCR was performed using iTaqTMUniversal SYBR Green Supermix (Bio-Rad, USA). An aliquot of 5 μL of cDNA was used as the template in the PCR reaction, which contained 12.5 μL of SYBR Green Supermix and 2 μL of 10 μmol L–1primer mixture. PCR reactions were performed in the CFX96TMReal-Time System C1000 Thermal Cycler (Bio-Rad, USA). The PCR conditions were as follows: 95°C for 3 min, followed by 42 cycles of 95°C for 10 s, and 56°C for 30 s, with a final incremental 0.5°C increase up to 95°C. The gene-specific primers are listed in Appendix A. The expression level was normalized to that of the internal control gene Actin using the 2–ΔΔCtmethod (Livak and Schmittgen 2001).

2.4. Total zein extraction and SDS-PAGE analysis

Total zein extraction and SDS-PAGE were performed according to Wu and Messing (2012). Endosperm with pericarp and embryo removed from the whole seed was weighed and added to 1 mL of extraction buffer (70%ethanol, 2% 2-mercaptoethanol, 3.75 mmol L–1sodium borate (pH 10.0), 0.3% SDS); the mixture was kept at 37°C for two hands, then shaken on a tissue grinder at 60 Hz for 60 s; the sample was centrifuged at 13 000 r min–1for 15 min; 100 μL of supernatant was transferred to a new tube,and 10% SDS was added; the mixture was dried by vacuum for 70 min; then, the zein was resolved in moderately diluted water according to its weight. A 3-μL sample was loaded for SDS-PAGE analysis.

3. Results

3.1. McV671 phenotype evaluation during seed development

V671 has a larger seed size than Mc at maturity (Liu et al.2014), and we investigated the kernel volume growth pattern at serial DAPs by measuring the kernel width for convenience. As shown in Fig. 1, both V671 and Mc showed an increasing-decreasing growth pattern during the investigated DAPs. The increasing rate of Mc peaked at 15 DAP, whereas V671 kept increasing until 18 DAP. Mc was significantly larger than V671 before 15 DAP; however,V671 exceeded Mc at 12 DAP in the rate of increase, and its kernel width became larger after 18 DAP (P=0.003). In addition, we compared the difference in Mc/V671 kernels during a neighbouring time period (30 DAP Mc vs. 33 DAP Mc for example) and found that the kernel width of Mc at 33 DAP was the same as that at 30 DAP (P=1); however, the V671 kernel width was significantly different between 33 and 30 DAP (P=0.0074), and the difference in the V671 kernel size disappeared between 36 and 33 DAP (P=1), indicating a longer increasing time period in V671. Moreover, the average increasing rate of V671 during grain filling was greater than that of Mc, with a rate of 0.167 mm d–1for V671 and 0.119 mm d–1for Mc, respectively (P=0.0021).

3.2. RNA-seq data and plant GO slim enrichment analysis

To determine why V671 and Mc differ in seed size and to determine the relationship between seed size and development process, a transcriptome analysis was conducted via RNA-seq using the Illumina Hi-Seq 2000 paired-end platform. The immature seeds at 14 DAP were selected for the analyses (each sample was analyzed with two biological replications) (Appendix B) because they play a critical transition role in seed development. The sequencing reads were selected according to the criteria of the quality of full-length reads＞Q30 and the forward 50 bp N＜1 and full sequence N＜3A, significantly high correlation (Fig. 2)was observed between the two biological replications,revealing steady expression profiling between the replicated samples. After filtering, 34 528 630 and 25 708 241 useful reads were generated in V671. A total of 31 583 352 and 22 661 161 reads were mapped to the B73 genome using TopHat, accounting for 91.47 and 88.15% of the useful reads respectively, among them, 30 231 111 and 21 715 702 were mapped to genes, 1 352 241 and 945 459 were mapped to intergenic regions. Mc had a similar percentage of reads that mapped to the B73 genome (Appendix B). In total,29 385 genes were expressed, of which 27 776 were expressed in V671 and 27 235 in Mc, accounting for 68.67–70.04% of the annotated genes. We counted the numbers and percentages of genes that were expressed in every single chromosome and compared them to the total number of genes that were expressed using a chi-square test(Appendix C); the results show that the number of genes that were expressed on chromosome 5 is significantly greater than average number in both Mc and V671. However, the number of genes that were expressed on chromosome 4 is less than the average number. There may be additional gene clusters or housekeeping genes that are expressed in the seed development stage on chromosome 5. According to the critical of 2-fold change and P＜0.05, 834 genes were differentially expressed (DEG, differentially expressed genes), with 492 genes more highly expressed in V671 and 342 genes more highly expressed in Mc.

Fig. 1 Growing patterns of seed size at serial days after pollination (DAPs) between Mc and V671. A, 10-kernel width and increasing rate (mm d–1). B, seed size after pollination from 6 to 42 DAP. Bar=1 cm.

Fig. 2 Principal component analysis (PCA) of RNA-seq datasets from Mc and V671 kernels at 14 days after pollination (DAP).

To clarify the developmental feature of 14 DAP kernels,GO classification was performed using the whole expressed genes in Mc and V671 (Fig. 3). In molecular function, the Mc-expressed genes were focused on binding, catalytic activity, nutrient reservoir activity, and related functions;in biological process, Mc-expressed genes were focused on response to stimulus, cellular process, single-organism process, metabolic process, and related functions; and in cellular component, most of the Mc-expressed genes were focused on cell, organelle, membrane, and other components. When V671 was analysed, similar GO classification components and scales were found, indicating a similar metabolic process at 14 DAP between Mc and V671. However, when we conducted a GO enrichment analysis using Mc-higher-expressed and V671-higherexpressed genes using AgriGO (Du et al. 2010), a different trend was found. Mc-higher-expressed genes were focused on nutrient reservoir activity (GO:0045735),endopeptidase inhibitor activity (GO:0004866, GO:0030414,GO:0004857, GO:0004867), enzyme regulator activity(GO:0030234), and hydrolase activity, acting on glycosyl bonds (GO:0016798), whereas V671-higher-expressed genes paid more attention to carbohydrate metabolic process (GO:0005975, GO:0044262, GO:0006073,GO:0044042), DNA conformation change (GO:0071103),protein-DNA complex assembly (GO:0065004), defense response (GO:0006952), nucleosome organization(GO:0034728), chromatin assembly (GO:0031497),nucleosome assembly (GO:0006334), DNA packaging(GO:0006323), and related functions (Appendix D). Genes that showed a higher expression in V671 were focused on DNA duplication-related processes, possibly indicating a higher mitosis activity. However, the genes that were expressed at higher levels in Mc shared a more intensive nutrient reservoir and enzyme regulator activity. Because nutrient accumulation begins after mitosis decreasing in maize endosperm development, we proposed that when the V671 endosperm undergoes mitosis, Mc has already begun to establish a nutrient reservoir activity. It is interesting to note that of the 22 expressing zein synthesis genes, 17 showed higher expression in Mc (Appendix E).

Fig. 3 Gene Ontology (GO) classification of 14 days after pollination (DAP) differentially expressed genes.

To determine biological processes regulated by Mc and V671 in 14 DAP, pathway analyses were performed based on differentially expressed genes (Fig. 4). These differentially expressed genes mainly involved in three biological processes: starch biosynthesis, storage proteins,and nutrient reservoir activity. Noteworthily, most of DEGs are concentrated in the starch synthesis pathway and some of DEGs are concentrated in the storage protein pathway and nutrient reservoir activity. Therefore, we can speculate that the synthesis of starch may be a key reason causing grain size difference at 14 DAP in maize. Furthermore,the genes over twice difference in these pathways were annotated (Table 1). According to the maize genome annotation (Zea_mays L. AGPv4.32), 19 annotated genes were performed, most genes are zinc-finger protein genes.

To investigate our hypothesis, genes that were related to storage protein accumulation and cell division were investigated at serial time periods of seed development.

3.3. Storage protein accumulation pattern

Maize (Z. mays L.) is the main source consumption of human and livestock, especially the storage protein in maize seed,which is more helpful to human and animal nutrition intake(Yue et al. 2014). Zein consists 70% of the storage proteins in maize endosperm (Wu et al. 2012), and its accumulation pattern was examined using SDS-PAGE; moreover, 15-kDa beta-zein and 16-kDa gamma-zein genes were selected to verify their expression by quantitative RT-PCR.

The accumulation of zein was examined at serial DAPs(Fig. 5). As shown in Fig. 3, no zein protein was observed at 10 DAP in either Mc or V671. At 12 DAP, Mc started accumulating gamma and alpha-zein, and these proteins continued to increase at 14 and 17 DAP. However, V671 started to accumulate zein later than did Mc, as only a trace amount of zein could be detected at 14 DAP, and the amount of zein that accumulated at 17 DAP in V671 was less than that in Mc at 14 DAP (Fig. 5).

According to the quantitative RT-PCR results (Fig. 6),zein synthesis genes experienced a huge jump from 9 to 11 DAP, increasing approximately 100 times than the levels at 8 DAP, and continued increasing to more than 350-fold.Consistent with the SDS-PAGE results, zein expression increased earlier in Mc. The 15-kDa beta-zein and 16-kDa gamma-zein began to increase at 8 DAP in Mc, whereas V671 maintained a rarely low zein synthesis activity before 9 DAP. The expression level of both gamma- and beta-zein in Mc was higher than that in V671 until 12 DAP. V671 began to have a higher level of gamma-zein expression from 13 DAP, whereas Mc maintained a higher beta-zein gene expression except at 18 DAP.

Although Mc and V671 share the same zein expression pattern during seed development, Mc begins zein accumulation earlier than does V671.

3.4. Quantitative RT-PCR verification for mitosis genes and gene dynamic expression

Fig. 4 Pathway analysis of differentially expressed gene transcripts. Log2Fold change heat maps of differentially expressed genes functioning in selected pathways in maize kernel.

Table 1 The description of differentially expressed genes in 14 days after pollination (DAP) between Mc and V671 kernels

Fig. 5 SD-PAGE of zein at serial days after pollination (DAP) between Mc and V671. M, marker.

The expression profiles of the two mitosis-related genes CycZme1 and CDC25 were examined between V671 and Mc by quantitative RT-PCR analysis (Fig. 6). As a B-type cyclin, CycZme1 works together with CDK1 to promote the G2-to-M transition in mitosis. The initiation of M phase indicates the beginning of cell proliferation. CDC25 acts as an activator of the CDK1/CycB complex.

In agreement with the reported decreased mitosis activity after 9 DAP, both CycZme1 and CDC25 showed a similarly decreased expression pattern in the first few days of the investigation (Fig. 6). CycZme1 expression was similar in V671 and Mc at 9 DAP, whereas V671 exhibited a significantly higher expression level on the remaining days. A similar situation occurred in CDC25 expression,except for a higher expression of Mc at 12 DAP and an equal expression at 13 DAP. Overall, V671 possessed a higher expression level at most time periods.

3.5. lAA synthesis genes expression pattern

IAA acts comprehensively during maize seed development(Lur and Setter 1993). IAA not only promotes storage accumulation but works together with CK to regulate the cell cycle; in other words, IAA can regulate the progression of seed development.

To investigate whether differences exist in IAA synthesis between Mc and V671, we examined the expression of the two zein synthesis genes ZmTar1 and ZmYuc1 at serial developmental stages using quantitative RT-PCR (Fig. 6).ZmTar1 and ZmYuc1 are rate-limiting genes that function in the IPA pathway of IAA synthesis, which is the main pathway for synthesis IAA in maize seed.

Both Mc and V671 showed a similar increasingdecreasing expression pattern in ZmTar1 and ZmYuc1. Mc displayed a higher ZmYuc1 expression at almost all of the investigated time points except for 11 DAP. Mc continually increased ZmTar1 expression from 8 to 11 DAP and had a higher expression than V671 before 12 DAP. V671 began increasing ZmTar1 expression later than did Mc (from 9 DAP) and continued increasing until peaking at 13 DAP.The earlier increasing pattern of ZmTar1 and ZmYuc1 in Mc is consistent with zein synthesis genes expression as well as the expression pattern of cell cycle-related genes (Fig. 6).

4. Discussion

4.1. The difference of phenotype and lAA synthesis between Mc and V671

Fig. 6 Expression patterns of indole-3-acetic (IAA) and zein synthesis genes as well as cell cycle related genes. DAP, day after pollination.

Mc and V671 are two inbred lines that show distinctively different seed sizes and kernel weights at maturity (Liu et al. 2014). Several QTLs that are related to seed size were mapped using a population derived from Mc and V671 in an earlier study. To identify the factors underlying seed size difference, phenotypic investigation and transcriptome analysis at key transitional stages of seed development were performed.

According to the seed size growth pattern investigation,we found that Mc develops faster than V671 but the rate of seed size increasing is slower in Mc. A similar result was found using RNA-seq at 14 DAP, which illustrated a higher storage accumulation activity in Mc and a higher DNA replication activity in V671. Regarding the hypothesis that Mc transitions earlier than does V671 during the mitosisto-storage accumulation transition stage, we designed and verified the accumulation of storage proteins in maize seed and the expression pattern of cell cycle-related genes at serial DAPs. The expression pattern of IAA synthesis genes demonstrates that it may be the more rapid increase of IAA in Mc that leads to the earlier transition from mitosis-tostorage accumulation in Mc and ultimately result in smaller seeds, as a higher expression level of IAA synthesis genes is consistent with a higher IAA level.

4.2. The mitosis-to-storage accumulation transition accounts for the difference in grain size between Mc and V671

Maize endosperm cells undergo a mixed developmental model from the center to the periphery at ～9 DAP (Sabelli and Larkins 2009), in which cells in the central starchy endosperm cease cell proliferation and transition to storage accumulation (Schweizer et al. 1995) while the peripheral cells are still undergoing mitosis. Over time, mitosis activity decreases and storage accumulation becomes the major process of seed development. The time period when seeds transition from mitosis to storage accumulation differs in different materials.

Prolamin and starch are the two most important storage materials in the maize endosperm at maturity. Storage accumulation can act as an indicator of the maize seed developmental progress. According to the RNA-seq results,zein synthesis genes expression in Mc was much higher than that in V671 at 14 DAP; however, no difference was found in the genes related to starch synthesis. In addition, the expression level of zein synthesis genes was much higher than that of starch accumulation genes, possibly indicating that zein synthesis occurs earlier than starch in maize endosperm; this phenomenon is consistent with results from the observation of developing maize and rice seeds.

When endosperm begins to transition from mitosis to storage accumulation, factors promoting the G2-to-M transition begin to be down-regulated; a higher G2-to-M transition activity means a higher mitosis rate. From the expression level of CycZme1 and CDC25, we inferred that V671 possesses a higher G2-to-M transition activity, and it can be inferred that the cell proliferation activity is more intensive in V671.

When seeking for factors controlling maize endosperm development, which constitutes the majority of the maize kernel volume and dry matter (Lur and Setter 1993; Sabelli and Larkins 2009), plant hormones are important candidates.Lur and Setter (1993) reported that the CK/IAA ratio in maize endosperm can regulate mitosis activity and other seed development processes. The locally synthesized IAA in maize seed is synthesized predominantly through the IPA pathway (Phillips et al. 2011). ZmVt2, ZmTar, and ZmYuc1 are rate-limiting enzymes in the IPA pathway. Based on the expression pattern of IAA-synthesized genes, we concluded that the IAA content increased earlier in Mc than in V671 in the first few days of the investigation. Furthermore, the increasing IAA content resulted in a decreased CK/IA1-A ratio. When the maize endosperm transitioned from mitosis to storage accumulation, the earlier decreasing CK/IAA in Mc also indicated an earlier decreasing mitosis and an increasing storage accumulation activity. In the early stages of seed development, Mc began IAA synthesis earlier than did V671. When V671 was still undergoing cell proliferation,the cell number in Mc had already peaked, and Mc endosperm cell began storage accumulation and increasing the cell volume. Mc seed width was larger than that of V671 before 15 DAP because of its precocity. However,when V671 completed mitosis and began cell elongation and storage accumulation, its seed size quickly exceeded that of Mc and it maintained this seed size difference until maturity. Because V671 possesses a larger sink strength and requires a longer time period to accumulate storage materials, a heavier kernel weight is necessary.

5. Conclusion

Developmental differences can have a function on seed size determination. Through this research, we postulate that the difference in IAA accumulation and some other regulators is responsible for distinct developmental progress of Mc and V671. The developmental progress changed mitosis and storage accumulation activity, as well as the time period of seed enlargement, and eventually leading to different seed size. In conclusion, this study showed that kernel development is a complex process involving an array of genetic and metabolic factors. Our comprehensive transcriptome analysis will provide valuable resource for further investigation into identify genes that are potential candidates for conferring small seed size.

Acknowledgements

This research was supported by the National Natural Science Foundation of China (91735306), the National Basic Research Program of China (973 Program, 2014CB138203),and the National Key Research and Development Program of China (2016YFD0100103-19). We also thank Prof.Wu Yongrui (Shanghai Institutes for Biological Sciences,Chinese Academy of Sciences) for his kindly hearted help on kernel zein extraction and SDS-PAGE analysis.

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

Anders S, Huber W. 2010. Differential expression analysis for sequence count data. Genome Biology, 11, R106.

Chen Y, Xu Y, Luo W, Li W, Chen N, Zhang D, Chong K.2013. The F-box protein OsFBK12 targets OsSAMS1 for degradation and affects pleiotropic phenotypes, including leaf senescence, in rice. Plant Physiology, 163, 1673–1685.

Du Z, Zhou X, Ling Y, Zhang Z, Su Z. 2010. AgriGO: A GO analysis toolkit for the agricultural community. Nucleic Acids Research, 38, W64–W70.

Garcia D, Fitz J G, Berger F. 2005. Maternal control of integument cell elongation and zygotic control of endosperm growth are coordinated to determine seed size in Arabidopsis. The Plant Cell, 17, 52–60.

Gonzalez N, Vanhaeren H, Inze D. 2012. Leaf size control:Complex coordination of cell division and expansion. Trends in Plant Science, 17, 332–340.

Ishimaru K, Hirotsu N, Madoka Y, Murakami N, Hara N, Onodera H, Kashiwagi T, Ujiie K, Shimizu B, Onishi A, Miyagawa H, Katoh E. 2013. Loss of function of the IAA-glucose hydrolase gene TGWss6 enhances rice grain weight and increases yield. Nature Genetics, 45, 707–711.

Jensen P J, Bandurski R S. 1994. Metabolism and synthesis of indole-3-acetic acid (IAA) in Zea mays (levels of IAA during kernel development and the use of in vitro endosperm systems for studying IAA biosynthesis). Plant Physiology,106, 343–351.

Kesavan M, Song J T, Seo H S. 2013. Seed size: A priority trait in cereal crops. Plant Physiology, 147, 113–120.

Kim D, Pertea G, Trapnell C, Pimentel H, Kelley R, Salzberg S L. 2013. TopHat2: Accurate alignment of transcriptomes in the presence of insertions, deletions and gene fusions.Genome Biology, 14, R36.

Liu Y, Wang L, Sun C, Zhang Z, Zheng Y, Qiu F. 2014. Genetic analysis and major QTL detection for maize kernel size and weight in multi-environments. Theoretical and Applied Genetics, 127, 1019–1037.

Livak K J, Schmittgen T D. 2001. Analysis of relative gene expression data using real-time quantitative PCR and the 2(–Delta Delta C(T)) method. Methods, 25, 402–408.

Luo M, Dennis E S, Berger F, Peacock W J, Chaudhury A.2005. MINISEED3 (MINI3), a WRKY family gene, and HAIKU2 (IKU2), a leucine-rich repeat (LRR) KINASE gene,are regulators of seed size in Arabidopsis. Proceedings of the National Academy of Sciences of the United States of America, 102, 17531–17536.

Lur H S, Setter T L. 1993. Role of auxin in maize endosperm development (timing of nuclear DNA endoreduplication, zein expression, and xytokinin). Plant Physiology, 103, 273–280.

Phillips K A, Skirpan A L, Liu X, Christensen A, Slewinski T L,Hudson C, Barazesh S, Cohen J D, Malcomber S, McSteen P. 2011. Vanishing tassel2 encodes a grass-specific tryptophan aminotransferase required for vegetative and reproductive development in maize. The Plant Cell, 23,550–566.

Sabelli P A, Larkins B A. 2009. The contribution of cell cycle regulation to endosperm development. Sexual Plant Reproduction, 22, 207–219.

Schruff M C, Spielman M, Tiwari S, Adams S, Fenby N, Scott R J. 2006. The AUXIN RESPONSE FACTOR 2 gene of Arabidopsis links auxin signalling, cell division, and the size of seeds and other organs. Development (Cambridge,England), 133, 251–261.

Schweizer L, Yerk-Davis G L, Phillips R L, Srienc F, Jones R J. 1995. Dynamics of maize endosperm development and DNA endoreduplication. Proceedings of the National Academy of Sciences of the United States of America, 92,7070–7074.

Sekhon R S, Hirsch C N, Childs K L, Breitzman M W, Kell P,Duvick S, Spalding E P, Buell C R, de Leon N, Kaeppler S M. 2014. Phenotypic and transcriptional analysis of divergently selected maize populations reveals the role of developmental timing in seed size determination. Plant Physiology, 165, 658–669.

Trapnell C, Roberts A, Goff L, Pertea G, Kim D, Kelley D R,Pimen- tel H, Salzberg S L, Rinn J L, Pachter L. 2012.Differential gene and transcript expression analysis of RNA-seq experiments with TopHat and Cufflinks. Nature Protocols, 7, 562–578.

Wang A, Garcia D, Zhang H, Feng K, Chaudhury A, Berger F, Peacock W J, Dennis E S, Luo M. 2010. The VQ motif protein IKU1 regulates endosperm growth and seed size in Arabidopsis. The Plant Journal, 63, 670–679.

Wang E, Wang J, Zhu X, Hao W, Wang L, Li Q, Zhang L, He W, Lu B, Lin H, Ma H, Zhang G, He Z. 2008. Control of rice grain-filling and yield by a gene with a potential signature of domestication. Nature Genetics, 40, 1370–1374.

Weber H, Borisjuk L, Wobus U. 1996. Controlling seed development and seed size in Vicia faba: A role for seed coat-associated invertases and carbohydrate state. The Plant Journal, 10, 823–834.

Weng J, Gu S, Wan X, Gao H, Guo T, Su N, Lei C, Zhang X,Cheng Z, Guo X, Wang J, Jiang L, Zhai H, Wan J. 2008.Isolation and initial characterization of GW5, a major QTL associated with rice grain width and weight. Cell Research,18, 1199–1209.

Wu T, Shen Y, Zheng M, Yang C, Chen Y, Feng Z, Liu X, Liu S,Chen Z, Lei C, Wang J, Jiang L, Wan J. 2014. Gene SGL,encoding a kinesin-like protein with transactivation activity,is involved in grain length and plant height in rice. Plant Cell Reports, 33, 235–244.

Wu X Y, Kuai B K, Jia J Z, Jing H C. 2012. Regulation of leaf senescence and crop genetic improvement. Journal of Integrative Plant Biology, 54, 936–952.

Wu Y, Messing J. 2012. RNA interference can rebalance the nitrogen sink of maize seeds without losing hard endosperm.PLoS ONE, 7, e32850.

Wu Y, Wang W, Messing J. 2012. Balancing of sulfur storage in maize seed. BMC Plant Biology, 12, 77.

Young M D, Wakefield M J, Smyth G K, Oshlack A. 2010. Gene ontology analysis for RNA-seq: Accounting for selection bias. Genome Biology, 11, R14.

Yue J, Li C, Zhao Q, Zhu D, Yu J. 2014. Seed-specific expression of a lysine-rich protein gene, GhLRP, from cotton significantly increases the lysine content in maize seeds.International Journal of Molecular Sciences, 15, 5350–5365.