Population structure of the 285 diverse inbred lines in China
The main heterotic groups according to the pedigree information in China varied with different stages, that is Golden Queen, Huobai, TSPT, Ludahonggu and Lancaster in 1970s (Wu 1983); TSPT, Ludahonggu, Lancaster, and Reid in 1980s (Zeng 1990); Lancaster (Mo17 subgroup and Zi330 subgroup), Reid, TSPT, Ludahonggu and other germplasm in 1990s(Wang et al. 1998; Teng et al. 2004); Reid, P, Ludahonggu, TSPT, Lancaster in early 2000s (Teng et al. 2004; Sun et al. 2014). Recently, SS, NSS and X groups were generated with the hybrids of Monsanto and Pioneer companies introduced to China (Sun et al. 2014; Zhao et al. 2018a). The main heterotic groups of China were also determined by molecular evaluation (Xie et al. 2007, 2008; Wang et al. 2008; Wu et al. 2014a). A mini core set of 95 lines which represented the Chinese maize inbred lines were assigned to four groups, i.e., TSPT, Reid, Lancaster, and P group (Wang et al. 2008). Wu et al. (2014a) also subdivided 367 inbred lines historically and recently widely used in maize breeding of China into Reid, Lancaster, P, TSPT, and Tem-tropic I group. In the present study, 285 inbred lines were assigned to four major groups by the model-based method and PCA, i.e., TSPT, Reid, Lancaster, and P group, which were consistent with known pedigrees of the inbred lines and previous studies (Wang et al. 2008; Wu et al. 2014a).
The TSPT group played a dominant role in Chinese maize breeding and occupied an important position in Chinese maize heterotic patterns. For example, Zhengdan958 derived from the pattern of Reid × TSPT was the most important Chinese maize hybrid with the largest annual planting area in consecutive 16 years since 2004. Ludan981 derived from P × TSPT with annual planting area about 0.67 million hectares from 2004 to 2007. Some elite lines of Chinese maize breeding historically and currently, such as Chang7-2, Dan340 and Aijin525 were developed from Chinese landraces (Li and Wang 2010), most of which and their descendants were clustered into the TSPT group in the present study. These results were consisted with Xie et al. (2007, 2008), who integrated the Ludahonggu and TSPT group into the D group. The reciprocal genetic improvement of TSPT and other local germplasm e.g. Ludahonggu and Golden queen were also founded based on the pedigree of some excellent lines, i.e., 5237 derived from Huangzaosi × Dan340 (TSPT × Ludahonggu), Jing2416 derived from 5237 × Jing24, Xun92_8 derive from Chang7-2 × 5237, and Shuang741 derived from [(3Tuan-11 × Huafeng100) × Huangzaosi] × Aijin525. In additionally, Zi330 and its descendants were also always used to improve lines from TSPT by Chinese breeders. For example, Ji853, a famous parent in TSPT, was derived from Huangzaosi × Zi330. Furthermore, several reports classified Zi330 and Ludahonggu into one group (Teng et al. 2004; Wang et al. 2008). Therefore, lines from Ludahonggu, Zi330 and local landraces were important donors to improve the TSPT germplasm.
Complex genetic basis of Lancaster was confirmed by the present study. Some lines of Lancaster according to pedigree were assigned to TSPT and P + Reid by neighbor-jointing tree analysis, such as lines from the Zi330 subgroup and the Va35 subgroup. The PCA results also showed that the Lancaster germplasm generally close to other germplasm. One reason was that the Lancaster Surecrop germplasm was less than 50% in the lines of Lancaster in China (Zheng et al. 2002).
The Reid group still occupied an important position in Chinese maize heterotic patterns. The Reid germplasm, such as M14, W24, W59E and B73, were firstly introduced to China in the 1950s. Later, using a batch of hybrids introduced from the United States, such as XL80, 3147, U8 and 3382, Chinese breeders developed several excellent inbred lines, such as Ye107, U8112, Ye478, Tie7922, Shen5003 (Li and Wang 2010). Recently, the female parent of Denghai605, the fourth Chinese maize hybrids with planting area more than 0.85 million hectares in 2019, was also derived from Reid.
The P inbreds, such as 178, P138, 87001, Qi319, Dan598, Dan599, Ye107 and 18–599, have been developed from the Pioneer hybrid of “78599’’ or similar hybrids since the late 1980s (Xie et al. 2007, 2008; Li and Wang 2010). Due to the existence of tropical and subtropical germplasm in the P group, the resistance to biotic and abiotic stress is often strong. Therefore, the introgression of the P germplasm to Reid, SS or NSS germplasm attracted more attention of breeders. Recent years, the success of several hybrids with extensive planting in China, e.g. Denghai605, Yufeng303 and Zhongkeyu505, were partly attributed to the P germplasm.
Comparative analysis of the genetic basis for leaf angle
The heritability of leaf angle varied, depending on genetic populations and environments in which populations are evaluated. It was found that the heritability of leaf angle was 90.5% in different years (Wassom 2013), 54.56% -87.00% at different locations (Michelson et al. 2002; Chen et al. 2015; Ding et al. 2015), 68.00% − 81.15% in different years combined with locations (Ku et al. 2010; Li et al. 2015). In addition, the heritability of leaf angle was 84.00% (Wang et al. 2017) under different planting densities (52500 plants/ha, 67500 plants/ha and 82500 plants/ha). Except that the heritability of leaf angle in Xiema, Beibei and Hechuan in Chongqing was 54.56–58.75% (Chen et al. 2015), more than 68.00% was observed for the heritability of leaf angle in different environments in other reports. In this study involving 285 diverse inbred lines of maize, it was found that the heritability of leaf angle was also very high (93.94%), indicating that Leaf angle was mainly controlled by genetic factors and less influenced by environments.
Zhao et al. (2018b) integrated the results of 27 QTL mapping studies for leaf angle, leaf orientation, leaf length, leaf width, leaf area and leaf length/width, and 16 meta-QTLs (mQTLs) were found to be distributed on 8 chromosomes except Chr. 9 and Chr. 10, suggesting the complexity of the genetic basis of leaf angle in maize. In this study, a total of 96 SNPs significantly associated with leaf angle (P < 4.14E-5) were distributed in 27 bins on the nine chromosomes except Chr. 8. We further identified 43 QTLs for leaf angles, of which 7 stable major QTLs (explaining more than 8% of the phenotypic variation for leaf angle and could be stably detected in at least 2 years and also for BLUP values) were distributed in bin 2.01, bin 2.07, bin 5.06 and bin 10.04.
QTL11 located in bin 2.01 were also identified by Lu et al. (2007), Ding et al. (2015), Li et al. (2015), and Wang et al. (2017) using different linkage mapping populations, explaining 2.86%, 4.06%, 4.30%, 8.85% of phenotypic variation, respectively. Ku et al. (2012) and Zhao et al. (2018b) integrated the previous QTLs for leaf angle, and detected an mQTL in bin 2.01 by meta-analysis. This locus in bin 2.01 stably expressed in different environments and different genetic backgrounds might be an important target for further study on leaf angle.
Two stable major QTLs (QTL17 and QTL18) were detected in bin 2.07, and a total of 8 SNPs significantly associated with leaf angle within these two QTLs could be stably inherited in different years. Liu et al. (2012a) detected a major QTL explaining 10.8% of phenotypic variation for leaf angle in bin 2.02–2.07. Zhang and Huang (2021) also found a major QTL in bin 2.07, which accounted for 14.2% of phenotypic variation for leaf angle. Due to the limited number of parents, the interaction of QTL × genetic background, population size and marker density in linkages analysis, and population structure and rare alleles in association analysis, few studies identified QTLs for leaf angle in this bin.
QTL29 located in bin 5.06 was a hot spot locus for leaf angle founded by previous studies (Mickelson et al. 2002; Lu et al. 2007; Zhang et al. 2014a; Zhao et al. 2018b)., and could explain 4.17% − 9.40% of phenotypic variation. In addition, Ku et al. (2012) integrated the previous QTL mapping results of leaf angle, and found an mQTL in bin 5.06 by meta-analysis. The mQTL affected both leaf angle and leaf orientation, and four candidate genes, yabby9, SE, LIC and yabby15 were predicted. Therefore, bin 5.06 might play an important role in leaf angle.
In previous studies, QTLs for leaf angle detected were mainly located in bin 10.07 (Yu et al. 2006; Zhang et al. 2014a; Sun et al. 2018; Wang et al. 2017), while bin 10.01–10.02 (Zhao et al. 2018b), bin 10.06 (Li et al. 2015) and bin 10.08–10.09 (Yu et al. 2006) were also reported. However, the QTL in bin 10.04 was the most important locus found in the present study, which carried three stable major QTLs (QTL41, QTL42 and QTL43), and contained 24 SNPs significantly associated with leaf angle. Therefore, the present study found for the first time that bin 10.04 plays an important role in the regulation of leaf angle in maize, and there may be at least three major genes that control leaf angle within this region. It was necessary to further map-based clone the genes underline by constructing large-scale secondary segregation populations.
Analysis of candidate genes for leaf angle
In recent years, a few genes regulating leaf angle or tiller angle were cloned in rice, maize, etc. It has been clear that leaf ligular region is the key tissue to regulate leaf angle (Zhu et al. 2015; Kong et al 2017; Hu et al. 2019). Factors such as the imbalanced division and expansion of adaxial and abaxial cells, the strength of mechanical tissue and the gravity response in the ligular region will affect the leaf angle (Zhu et al. 2015; Hu et al. 2019). The molecular regulation mechanism of leaf angle mainly including (1) the regulatory pathway of plant hormone synthesis and signal transduction: brassinosteroid, auxin and gibberellin have been deeply studied on the regulation of leaf angle, while other hormones such as jasmonic acid, strigolactones and ethylene have also been found to be involved in the regulation of ligular region development. (2) other regulatory pathways: the gravity response of the ligular region (such as LPA1, which affects the division of adaxial cells in ligular region, and LAZY1, which regulates the polar auxin transport) or the strength of mechanical tissue in ligular region (such as ILA1, which controls the formation of vascular bundles and the content of cellulose and xylan in the cell wall; OsCesA7 and OsCD1 involved in cellulose synthesis; DL, the YABBY gene family, which controls the formation of leaf blade-midribs; OsIG1 involved in the division and differentiation of vesicular cell and the formation of ligular region). (Zhu et al. 2015; Luo et al. 2016; Hu et al. 2019)
The progress of identifying the regulatory genes for leaf angle in maize is relatively slow. Through the research on mutants, a few genes in regulating leaf angle was cloned, such as LG1, encoding squamosa promoter binds to protein (SBP); LG2, encoding a basic leucine zipper protein (bZIP); lg3 and lg4, encoding homeo domain protein. The mutants of lg1 and lg2 (Moreno et al. 1997; Walsh et al. 1998) and the dominant mutants of lg3-O and lg4-O showed that the abnormal development of ligular region resulted in a small leaf angle (Bauer et al. 2004). Through homologous cloning, ZmTAC1, a homologous gene of TAC1 that regulated leaf angle in rice, had been proved to be able to positively regulate the leaf angle (Ku et al. 2011). In recent years, ZmCLA4 underlined qLA4-1 in bin 4.03 (Zhang et al. 2014b), brd1 underlined UPA1 in bin 1.08 and two-base sequence polymorphism underlined UPA2 in bin 2.03 (Tian et al. 2019), and ZmILI1 underlined qLA2 in bin 2.02 (Ren et al. 2020) for leaf angle have successively map-based cloned. Seven stable major QTLs for leaf angle were identified in the present study, which will be useful for the improvement of leaf angle by molecular breeding and provide a basis for the cloning of the genes.
The candidate gene of QTL11 might be Zm00001d001961, which encoded SAUR-like auxin-responsive protein. The majority of SAUR genes can induce growth in leaves, stems and floral organs via cell elongation (Sun et al. 2016; Stortenbeker and Bemer 2019). In addition, the overexpression of SAUR could also influence the auxin levels, polar auxin transport and/or expression of auxin pathway genes (Stortenbeker and Bemer 2019). Auxin can also played role on the leaf angle via the regulation of the growth of the abaxial and abaxial cells of ligular region, auxin signal transduction (Song et al. 2009; Li et al. 2010). So Zm00001d001961 might regulate leaf angle by the division and expansion of adaxial and abaxial cells.
Zm00001d006348, the predicted candidate gene for QTL17, encoded a growth-regulating factor (GRF) with two conserved domains, i.e. WRC and QLQ (Choi et al. 2004). Through the interaction between QLQ domain in the N-terminal of GRFs and the SNH domain in GIF (GRF-interaction factor) protein, a functional complex which regulated the expression of genes downstream was formed (Kim and Kende 2004). In Arabidopsis thaliana, GRFs could promote or inhibit the growth of organs by regulating the cell size (Kim et al 2003) or the cell division (Kim and Lee 2006; Vercruyssen et al. 2015). Over-expression of BrGRF8 from Chinese cabbage (Wang et al. 2014) and BnGRF2 from rapeseed (Liu et al 2012b) in transgenic Arabidopsis plants increased the sizes of the leaves and other organs by regulation of cell proliferation. It was also found that genes for GRFs could regulate several traits, such as leaf size, plant height and flowering traits in maize (Wu et al 2014b; Zhang et al. 2008). Their transcriptional expression were affected by gibberellin (GA) (Wang et al. 2014) and miR396 (Jones-Rhoades et al. 2006). Because the protein encoded by the candidate gene of Zm00001d006348 was the homologous sequence of that encoded by AT2G45480 (GROWTH-REGULATING FACTOR 9) in Arabidopsis thaliana, the function of Zm00001d006348 can be predicted by AT2G45480. AT2G45480 negatively regulated leaf growth by activating the expression of a bZIP transcription factor gene, i.e. OBP3-RESPONSIVE GENE (ORG3), and restricted cell proliferation in the early stage of leaf development (Omidbakhshfard et al. 2018). Therefore, Zm00001d006348 might negatively regulate the proliferation of cells in the ligular region, and further affect the leaf angle.
Zm00001d006463, the candidate gene for QTL18, encoded C2C2-Dof-transcription factor with a highly conserved DNA-binding domain of Cys2/Cys2 (C2/C2) zinc finger (Yanagisawa, 2004). The DOF proteins were found involved in many biological processes, such as tissue differentiation, seed development and regulation of metabolism (Noguero et al. 2013). It is worth noting that theDOF transcription factor AtDOF5.1 could regulate leaf adaxial-abaxial polarity by controlling the expression of Revoluta (REV), a Class III homeodomain-leucine zipper (HD-ZIPIII) transcriptional regulatory proteins (Kim et al. 2010). Therefore, Zm00001d006463 might also regulate the leaf adaxial-abaxial polarity in maize.
Zm00001d017618, a possible candidate gene for QTL29, encoded a B3-domain transcription factor that might regulate the expression of genes in the BR synthesis pathway and the signal transduction pathway (Tian et al. 2019; Je et al. 2010). Recently, Tian et al. (2019) found that the B3-domain transcription factor gene Zm00001d002562 (ZmRAVL1) on Chrom. 2 activated the expression of brd1 downstream, increased endogenous brassinosteroid content, promoted proliferation of cells in ligular region, and enlarged leaf angle. Therefore, Zm00001d017618 was predicted to regulate the leaf angle via brassinosteroid content in the ligular region.
Zm00001d024919 encoding an adenylate kinase was the possible gene for QTL41. Two near isogenic lines with different leaf angle were analyzed by the method of comparative proteomic analysis, and adenylate kinase complicated in glycometabolism might be associated with leaf angle formation and the physical and mechanical properties of midribs (Wang et al. 2015).
The candidate gene for QTL42 might be Zm00001d025018 encoding alpha expansin. Expansins responsible for cell-wall relaxation are required for cell elongation (Kong et al. 2017). Furthermore, increased expansin expression at early stages of organ development induced cell proliferation and consequently increased growth (Wyrzykowska et al. 2002). Both the size of the cells determined by the cell elongation and the number of cells determined by the cell division influence the size of the auricle and ultimately leaf angle (Kong et al. 2017).
Zm00001d025033 encoding a TCP transcription factor was the possible candidate gene of QTL43. TCP transcription factors with TCP conserved domain formed a bHLH structure had been found to regulate many aspects of plant development, such as branching in maize (Doebley 1995; Bai et al. 2012)༎TB1 (Teosinte Branched 1) can inhibit the growth and development of lateral branches, and tb1 mutant gene leaded to increased number of lateral branches in maize (Doebley 1995). BAD1, which is another TCP transcription factor gene and could promote cell proliferation in a lateral organ, i.e. the ligular region, may influence inflorescence architecture by regulating the angle of lateral branch emergence ༈Bai et al. 2012). Zm00001d025036 encoding indole-3-acetate beta-glucosyltransferase also located in this QTL. The glycosylation of auxin might influence the phenotype of plant via the regulation of auxin levels (Jackson et al. 2002).