Transcriptomic analysis reveals the regulation of early ear-length development in maize

Ear length is an important component of maize grain yield. However, the ear length is a complex quantitative trait, and the underlying molecular mechanisms remain poorly understood. Here, the chromosome segment substitution line (CSSL) 1283 displayed a longer ear length compared with the recipient parent Xu178. An RNA sequencing analysis of Xu178 and CSSL1283 ears during three undifferentiated ear developmental stages identified 1,991 differentially expressed genes (DEGs). A gene ontology analysis of the DEGs showed that genes related to transcription factors and response to abiotic stimulus were significantly enriched. Furthermore, the expression of DEGs associated with AP2/EREBP and WRKY transcription factors and heat shock proteins was upregulated in CSSL1283. In addition, several genes encoding protein kinase were differentially expressed between Xu178 and CSSL1283. Our study provided a genetic resource for the dissection of the molecular mechanisms of ear-length development and for uncovering candidate genes to increase maize ear length.


Introduction
Maize (Zea mays L.) is an important resource for food, feed and biofuel materials (Godfray et al. 2010). As the population rapidly increases, so will the demand for maize grain. Maize grain yield is determined by planting density, kernel weight and kernel number per ear (Huo et al. 2016). Ear length is closely associated with kernel number per ear, which ultimately affects the grain yield, therefore, it is a key Communicated by Ben Zhang. Xiaoyang Chen cxy759020@henau.edu.cn Jihua Tang tangjihua1@163.com 1 by mediating transport in the vascular tissue of the young ears (Pei et al. 2022). Although significant progress has been made in understanding the regulation of ear length, knowledge of genes and biological pathways involved in ear-length development remains limited.
Maize ears start from the axillary meristems (AMs) in the axils of the leaves. In early ear development, switching from vegetative to reproductive development turns AMs into ear IMs. The IMs then elongate and produce spikeletpair meristems (SPMs). Each SPM develops into two spikelet meristems (SMs), which become floral meristems that form kernels after fertilization (Eveland et al. 2013;Vollbrecht and Schmidt 2009). The initiation and maintained activity of these meristems determine the grain yield-related traits in maize. For instance, the kernel row number and grain yield are enhanced in the elite alleles of FASCIATED EAR2 (FEA2), FASCIATED EAR 3 (FEA3), UNBRANED3 (UB3) and KRN4, which are involved in the development of IMs and the number of SPMs (Bommert et al. 2013b;Chuck et al. 2014;Je et al. 2016;Liu et al. 2015).
Transcriptome profiling studies can explore spatial and temporal gene expression patterns. Thus, this method has been used to identify important genes and biological pathways responsible for tissue and organ development (Zhao et al. 2020). In maize, Yi et al. performed RNA sequencing (RNA-seq) and detected 1,093 genes, including 110 transcription factors (TFs) specifically expressed in seed (Yi et al. 2019). Notably, with a transcriptome dataset and gene overexpression lines, an AP2 TF BABY BOOM2 was shown to promote maize callus formation and transformation (Du et al. 2019). In addition, a transcriptomic comparison of maize non-stressed and heat-stressed pollen indicated that starch, lipid and energy biosynthetic processes are crucial for pollen development at the tetrad stage after transient heat stress (Begcy et al. 2019). The transcriptomic analysis revealed that some TFs correlated to drought stress and abiotic stress responses were enriched in the overexpression of AtGA2ox1 in maize .
Here, CSSL1283 displayed significantly longer ears and a greater kernel number per row compared with the recipient parent, Xu178. To reveal the underlying molecular mechanisms, we report the transcriptome profiling of CSSL1283 and its recipient parent, Xu178, at three early ear developmental stages. We identified the enriched processes and important genes involved in early ear-length development. This study promotes the understanding of the molecular mechanisms underlying ear-length development and will aid in the genetic improvement of maize ear length.

Phenotypical analysis of CSSL1283
In total, 150 CSSLs were constructed using the inbred line Zong3 as the donor parent and Xu178 as the recipient parent (Mao et al. 2013). One of the lines, CSSL1283, displayed significantly longer mature ears and a higher kernel number per row compared with the recipient parent, Xu178 (Fig. 1A, J, K).
To gain a better insight into ear development, we tracked the growth of young Xu178 and CSSL1283 ears and measured the ear lengths at four different stages (stg 1-stg 4). Stg1 and stg2 ears initiated IMs and SPMs, respectively, A The mature ears of Xu178 and CSSL1283. Bar = 1 cm. B-I The ear primordia of Xu178 (B, D, F, H) and CSSL1283 (C, E, G, I) from stg1 (B and C), stg 2 (D and E), stg3 (F and G) and stg4 (H and I). Bars = 100 μm in B to E, 500 μm in F to I. J and K The comparison of ear length (J) and kernel number per row (K) at maturity ear. L The length of ear inflorescence meristem. Datas are presented as means ± SD, p-value is estimated by the two-tailed t-test, *P < 0.05, **P < 0.01 with SMs formed at stg3 and floral meristems formed by stg4. The CSSL1283 ears were significantly longer than those of Xu178 at all four stages ( Fig. 1B-I, L), indicating that the increased ear length of CSSL1283 occurs in an early stage of ear-length development.

Global transcript profiles of undifferentiated ear inflorescences
The phenotypic difference of CSSL1283 revealed that the expression of genes in early ear-length development contributed to the increased ear length. To explore regulatory networks of ear length in CSSL1283, we performed RNAseq on Xu178 and CSSL1283 ears from stg1 to stg3. In total, 410.93 million high-quality reads were generated ( Table 1) and then mapped to the maize B73 reference genome (Ref-Gen_V4). On average, 87.89% of the total clean reads were uniquely mapped (Table 1). Gene expression levels were calculated as FPKM (fragments per kb exon model per million mapped fragments) values. Determinations of the Pearson correlation coefficients indicated that the expression levels between biological replicates were highly correlated ( Table  S1). We next compared the number of expressed genes for different samples. Genes with FPKM ≥ 1 were considered expressed genes. A total of 21,518 genes were expressed in at least one sample. Among these genes, 20,842 and 20,997 genes were detected in Xu178 and CSSL1283, respectively. In total, 94.44% (20,321/21,518) of the genes were detected in both Xu178 and CSSL1283 ( Fig. 2A).
Nine DEGs were selected for verification by qRT-PCR. The correlation coefficient of the qRT-PCR results and RNA-seq data was 0.85, indicating the RNA-seq data were reliable (Fig. 3).

An overview of the differentially expressed genes (DEGs)
To identify the factors involved in the increased ear length of CSSL1283, we analyzed the DEGs between Xu178 and CSSL1283. Compared with CSSL1283, a total of 1,991 DEGs were identified in Xu178 using the cutoff criteria (FDR ≤ 0.05 and fold change ≥ 2), including 1,178 upregulated genes and 904 downregulated genes (Fig. 2B). In total, 908, 644 and 1,129 genes were differentially expressed in stg1, stg2 and stg3, respectively, suggesting more complicated regulatory networks in the latter stage than in the former two stages (Fig. 2C).
A gene ontology (GO) term enrichment analysis was used to elucidate the biological pathways of these DEGs. In stg1, enriched processes included TF activity, sequencespecific DNA binding, nucleic acid binding TF activity, oxylipin biosynthetic and metabolic process and oxidoreductase activity, and involved 65, 65, 7, 7 and 82 genes, respectively (Fig. 2D). However, in stg2, only two categories of TF activity, sequence-specific DNA binding and nucleic acid binding TF activity were significantly enriched, involving 38 and 38 genes, respectively (Fig. 2E). In stg3, the enriched terms were mainly associated with photosynthesis and response to abiotic stimulus (Fig. 2F). Thus, the mechanisms in stg1 and stg2 that underlie CSSL1283 earlength development were similar but different from those in stg3. and WRKY TFs appears to be related to the increased ear length of CSSL1283.

Protein kinase and heat shock protein are required for ear-length development in CSSL1283
A serine/threonine-protein kinase encoding gene, KNR6, was reported to determine pistillate floret number and ear length, suggesting the significance of protein kinases during maize ear-length development (Jia et al. 2020). Thus, DEGs related to protein kinases in stg1 were analyzed in detail. In total, 47 genes belonging to the protein kinase family were differentially expressed between Xu178 and CSSL1283 (Table S3). This included 14 genes encoding receptor-like protein kinases (RLKs) and 7 genes involved in the mitogen-activated protein kinase (MAPK) signaling pathway (Fig. 5A, B). In addition, opposite expression trends for the two kinds of protein kinase genes were observed. In CSSL1283, 12 of 14 DEGs encoding RLKs were downregulated, whereas all the DEGs involved in the MAPK signaling pathway were upregulated (Fig. 5A, B), implying that RLKs and members of the MAPK signaling pathway play different roles in CSSL1283 ear-length development.
The GO analysis revealed that the term "response to abiotic stimulus" (GO: 0009628) was the most significantly enriched in stg3 (Fig. 2F). This term included 68 DEGs, with 27 being assigned to the GO term "response to heat"

Activated expression of AP2/EREBP and WRKY family genes in CSSL1283
In plants, TFs are important regulators of growth and development. The GO analysis showed that the term of TF activity and sequence-specific DNA binding (GO: 0003700) was significantly enriched in both stg1 and stg2, suggesting that TFs play key roles in CSSL1283 ear-length development at these two stages (Fig. 2D, E). There were 65 and 38 differentially expressed TFs (DE TFs) in stg1 and stg2, respectively (Table S2).

Potential roles of protein kinases in ear-length development
Protein phosphorylation by protein kinases can regulate plant growth and environmental responses. In Arabidopsis, most of the protein kinases are RLKs, accounting for approximately 60% of the total protein kinases (Shiu and Bleecker, 2001). In maize, a leucine-rich repeat receptorlike kinase, thick tassel dwarf1, is involved in male and female inflorescence development (Bommert et al. 2005). In addition, the two kernel row number genes, FEA2 and FEA3, encode leucine-rich repeat receptor-like proteins (Bommert et al. 2013b;Je et al. 2016; Taguchi-Shiobara et (GO: 0009408) (Fig. 2F). The other DEGs in the term "response to heat", except for three genes, were upregulated in the CSSL1283 (Table S4). Furthermore, 15 of 27 DEGs encoded heat shock proteins (HSPs), and all of them are upregulated in CSSL1283 (Fig. 5C), suggesting that there is a positive connection between ear-length development and response to heat and that the HSPs are involved in these two biological processes.

The TFs play key roles in ear-length development
Several TFs have been reported in maize ear inflorescence development, such as UNBRANED 2, UB3 and FASCIATED EAR4 (Chuck et al. 2014;Pautler et al. 2015). In line with these results, our transcriptomic data revealed that TFs are important for early ear-length development.
AP2/EREBP family genes are involved in the development of IMs in maize. For instance, the SMs fate is controlled by two AP2/EREBP family genes, Indeterminate Spikelet1 and Branched Silkless1 (Chuck et al. 1998;Fig. 3 Relative expression levels of nine genes as assessed by qRT-PCR. The histogram represents qRT-PCR data and the broken line represents RNA-seq data. Gray and black histograms indicate the Xu178 and CSSL1283 ears, respectively. Broken lines with hollow points and solid points indicate the Xu178 and CSSL1283 ears, respectively. Datas are presented as means ± SD, p-value is estimated by the twotailed t-test, *P < 0.05, **P < 0.01 the predicted α-subunit of a heterotrimeric GTP-binding protein (Bommert et al. 2013a). These findings indicate that signal transduction is essential for ear inflorescence development. In this study, 14 RLKs were differentially expressed in stg1 ear inflorescences between Xu178 and CSSL1283 (Fig. 5A). Thus, we speculate that the signal transduction mediated by RLKs is an important mechanism underling the increased ear length of CSSL1283.
Auxin is a central regulator in ear inflorescence development (Gallavotti et al. 2008;Galli et al. 2015;McSteen et al. 2007;Phillips et al. 2011). Auxin transport can be affected by protein kinases, such as BARREN INFLORES-CENCE2 in maize (Skirpan, 2009), and PINOID and MAP KINASE KINASE7 in Arabidopsis (Dai et al. 2006;Friml et al. 2004). During stg1, 7 genes implicated in the MAPK signaling pathway and 8 auxin-related genes were identified as being differentially expressed between Xu178 and CSSL1283 (Fig. 5B, Table S5). Considering the potential function of protein kinase KNR6 in auxin-dependent earlength development (Jia et al. 2020), it is likely that the cascade of "MAPK signaling pathway-auxin" is a contributor to the increased ear length of CSSL1283.

The ear-length development involves complicated biological processes
Ear length is a typical quantitative trait and is controlled by many genetic loci. Consistently, the GO analysis showed that TFs are the major regulators of ear-length development in stg1 and stg2 (Fig. 2D, E), whereas processes related to photosynthesis and responses to abiotic stimulus played important roles in stg3 (Fig. 2F). This suggested that various biological processes are required for ear-length development in maize.
Adverse environmental conditions, such as high temperature, influence ear development and grain yield. Emerging evidence indicates that HSPs are involved in thermotolerance (Maestri et al. 2002). In addition, HSPs regulate the formation of chalky grains under heat-stress conditions (Tsutomu et al. 2018). HSP90 and a gene encoding a protein kinase WAK may bind to Ca2 + signaling related genes to regulate lipid and nitrogen metabolism, thereby improving salt tolerance in barley (Zhu et al. 2020). WRKY TFs modulate many developmental processes and stress resistance (Rushton et al. 2010;Wang et al. 2018). Thus, HSPs and WRKY TFs have dual functions in plant growth. In our data, the expression levels of DEGs related to HSPs and WRKY TFs were induced in CSSL1283 (Fig. 4C, 5C), suggesting that they can coordinate ear-length development and heat tolerance in maize. Therefore, the identified genes provide key targets for the improvement of maize grain yield. al. 2001). Because it lacks a signaling domain, the function of FEA2 depends on COMPACT PLANT2, which encodes  expression values. The differential expression analysis was carried out using Cuffdiff v2.2.1 3 . Genes with an expression fold change ≥ 2 and FDR ≤ 0.05 were defined as differentially expressed. AgriGO v2.0 (http://systemsbiology.cau. edu.cn/agriGOv2/index.php) was used for the GO enrichment analysis (Du et al. 2010).

qRT-PCR
In total, 1 µg RNA was reverse transcribed following the procedures of the PrimeScript™ , ® Reagent Kit with gDNA Eraser (TaKaRa). qRT-PCR was performed on a CFX96 Real-Time PCR Detection System (Bio-Rad) using TB Green™ Premix Ex Taq™ II (TaKaRa). Three biological replications were included, and each biological replicate contained three technical replications. The maize Actin gene (Zm00001d010159) was used as the internal normalization control. The gene expression level was evaluated using the 2 −ΔΔCt method (Livak and Schmittgen, 2001). All the primers were designed using NCBI (https://www.ncbi.nlm.nih. gov/tools/primer-blast) and are listed in Table S6.

Declarations
Ethics approval and consent to participate All experimental studies on plants have complied with relevant institutional, national, and international guidelines and legislation.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended In addition, EAD1 expression was downregulated in Xu178 in stg1. Overexpression of EAD1 resulted in longer ear length and an increase in kernel number per row (Pei et al. 2022), which is consistent with our results. These results further suggest that the transport and distribution of malate regulates ear development in maize.

Plant materials and phenotypic analysis
CSSLs were constructed using two maize elite inbred lines, Zong3 and Xu178, which are the donor and recipient parents, respectively (Mao et al. 2013). These lines were planted in the summer of 2019 in the experimental field of Henan Agricultural University in Zhengzhou, Henan Province. At maturity, randomly selected ears were used to measure the ear length and kernel number per row. The observation of young ears was conducted using a Zeiss dissecting microscope. The lengths of young ears were calculated using ZEN 2011 software. The stg1 to stg3 young ears were collected for RNA-seq. This included two biological replicates in stg1 and three biological replicates in stg2 and stg3. The young ears were stored at -80 ℃ for further use.

Total RNA extraction and transcriptome sequencing
The total RNA was extracted from ear inflorescence using an RNAprep Pure Plant Plus Kit (Tiangen). mRNAs were enriched with oligo (dT) magnetic beads and then fragmented into short fragments by adding a fragmentation buffer. The first cDNA strand was synthesized using random hexamers and the associated buffer. Then, dNTPs, RNase H and DNA polymerase I were added to synthesize the second cDNA strand. Double-stranded cDNA was terminal repaired and added to the 3′ ends. The QiaQuick PCR kit was used for purification, and samples were eluted in EB buffer. The ends were repaired and the sequencing adapter was connected. Agarose gel electrophoresis was used to select the appropriate fragment size, and finally, PCR amplification was performed to enrich the cDNA. Library quality was determined using the Agilent 2100 system. The library was then sequenced using Illumina Novaseq to generate 150-nucleotide paired-end reads.