In this study, the data used included 20 adult-plant stage agronomic traits of 513 maize inbred lines (Yang et al., 2014), and genotype data corresponding to the population; 20 aboveground and root traits of maize seedlings of the same population under low phosphorus and normal phosphorus conditions were investigated (Zhang et al., 2014). The genotype data included 556,809 high-quality single nucleotide polymorphism (SNPs) markers (Minor Allele Frequency, MAF ≥0.05) (Fu et al., 2013; Li et al., 2013), downloaded from http://www.maizego.org/Resources.html.

To identify the C2H2-ZFPs, we searched the Zea mays proteome (ftp://ftp.ensemblgenomes.org/pub/plants/release-29/fasta/zea_mays/pep/) using HMMER package 3.1b2 (Eddy, 1998) and the Pfam domain ZF-C2H2 (PF00096) (Bateman et al., 2004). The minimal cut-off score for the search was 0. The choice of this rather low threshold permits the detection of all ZFs/ZFPs and several false positives. The C2H2-ZFs and other conserved domains in the proteins coded by these genes were determined using SMART (http://smart.embl-heidelberg.de/) and InterProScan (http://www.ebi.ac.uk/interpro/search/sequence-search/), along with manual inspection.

We extracted all the C2H2-ZF domains of maize ZFPs according to the SMART (http://smart.embl-heidelberg.de/)-identified C2H2-ZF motif sequence, a neighbor-joining (NJ) phylogenetic tree was also constructed (El-Sappah et al., 2021a). The Bootstrap test was repeated 1,000 times.

Microarray-based data analyses of both development and stress responses were performed to gain insight regarding the family-wide expression profile of C2H2-ZFPs and obtain the genes with highly significant differential expression. To analyze the spatial and temporal expression patterns of the ZmC2H2 genes during development, transcriptome data of the genome-wide gene expression atlas of the maize inbred line B73, generated using NimbleGen microarray technology, were downloaded from Plexdb (ZM37) (http://www.plexdb.org/modules/PD_browse/experiment_browser.php?experiment=ZM37&genechip=nimbMaizeV2). Gene expression data from Agilent-025271 Zea mays Mais_array_v1 (GPL14913) under conditions of low T, N, and P were downloaded from Gene Expression Omnibus with accession numbers GSE46704 (Low P, T, N). In the preparatory stages of this study, our research team selected a drought-tolerant maize inbred line Ac7643 (P1) and a drought-sensitive inbred line Ac7729/TZSRW (P2) to simulate drought treatment and extract RNA. Two inbred maize lines were subjected to whole transcriptome sequencing (RNA-seq) using the Illumina HiSeq 2000 technology platform. We extracted the expression data of the C2H2-ZF gene from this data (RNA-seq), and used it to analyze the differential expression pattern of C2H2-ZF genes between different maize inbred lines under drought conditions. All microarray data were imported into R and Bioconductor (http://www.bioconductor.org/) for expression analyses. The ‘gplots’ package was then used to generate heatmaps.

An association panel of 513 inbred maize lines was used to identify single nucleotide polymorphisms (SNPs) in the C2H2-ZF family genes significantly associated with 20 adult-plant stage agronomic traits and 19 shoot and root traits in the seeding stage under normal afforded and non-afforded phosphorus. This panel contained 556,809 high-quality SNPs (Fu et al., 2013; Li et al., 2013), downloaded fromhttp://www.maizego.org/Resources.html. The SNPs of the 247 C2H2-ZF genes were obtained from the panel using perl script. In order to eliminate the influence of environmental factors, best linear unbiased predictors were generated for each trait in different environments using SAS v8 package (SAS Institute Inc., Cary, NC, USA). The general linear model (GLM) and GLM combined with Q matrix (population structure) in TASSEL v3.0 (Bradbury et al., 2007) were used to conduct association mapping between the SNPs of candidate genes with the 20 traits in the 513 maize inbred lines, with p < 0.0001 being considered the significant threshold.

NCBI PubMed literature database (https://pubmed.ncbi.nlm.nih.gov/32061337/) has been used to find the protein sequences of C2H2-ZF genes with known function for constructing NJ phylogenetic tree with the maize C2H2-ZF genes. Furthermore, the iTOL online server (https://itol.embl.de/) has been used for further beautify.

The modified CTAB method (Porebski et al., 1997) extracted genomic DNA of 109 inbred maize lines from different countries and regions (Supplementary Table S1). Using the Primer-BLAST program of NCBI’s BLAST tool (https://blast.ncbi.nlm.nih.gov/Blast.cgi), specific primers were designed using the DNA sequence of the first transcript of ZmZFP126 gene (4,516 bp) as a template (Table 1). Using specific primers and based on TaKaRa’s high-fidelity PCR amplification enzyme, ZmZFP126 gene DNAs of 109 inbred maize lines were amplified and sequenced. The obtained sequences were subject to multiple sequence alignment analysis with MAFFT software. Principal component analysis (PCA) was performed on the genotypes of 109 inbred maize lines using R software. SNPs and Indels were extracted from the ZmZFP126 gene sequence using TASSEL v3.0 software, and the general linear model (GLM) combined with PCA was selected to perform correlation analysis on SNPs and Indels of ZmZFP126 gene and 20 agronomic traits of 109 inbred maize lines. At the same time, the linkage disequilibrium (LD) level of the ZmZFP126 gene in maize was analyzed using TASSEL v3.0 software.

Three hundred twenty-six sequences from maize proteome version 2 were searched using HMMER and manual inspection, out of which 203 were unique, whereas the remaining 123 belonged to different transcripts of 44 genes (Supplementary Table S2). The SMART analysis of these 326 sequences revealed that all contained ZF domain(s). Some proteins were also found to contain KOW, DnaJ, SET, CactinC_cactus, Kin17_mid, C3H-ZF, U1-ZF, RING finger, Act-Frag_cataly, Amino_oxidase, transmembrane region, CMAS, RPT, SCOP, ANK, SEC14, EGF, G-patch, CHROMO, DSPc, F420_oxidored, SANT, IPPT, Di19_C, DUF3546, LYAR-ZF, AN1-ZF, Jmj, EXO3, DFD1, DUF1644, ARS2, Telomere_Sde2_2, PUG, SET, SF3a60_bindingd, and UBA domains as determined using the SMART analysis.

Four hundred and seventy-two individual ZFs were found in the 247 C2H2-ZFPs. Two extremely different kinds of motif, bearing no sequence similarity except for the presence of two C and two H residues, were observed (Table 2). The first type contained a conserved sequence “QALGGH” in the DNA-recognition motif and was designated as Q-type. This sequence has been shown to be specific to plants (Takatsuji et al., 1994), and 147 such ZFs have been found in maize. The other distinct kind of ZF was designated as C-type, as these did not contain any conserved motif in the ZF region. Certain modifications observed in the Q-type ZFs were classified as M-type and were numbered from M1-M11, indicating the type of modification as shown in Table 2. The M-type fingers have also been found in petunia (Kubo et al., 1998), Arabidopsis (Englbrecht et al., 2004), and rice (Agarwal et al., 2007). Among the C-type ZFs, some contained highly conserved motifs in the finger and flanking regions and were designated as Z-type; the 12 variants of Z-type were designated Z1-Z12 (Table 2).

An unrooted NJ-plotted phylogenetic tree was constructed using the 472ZFs domains sequences (Figure 1). This revealed that the following. 1) Maize C2H2-ZFs can be classified into two primary types, namely C-type and Q-type, of which the C-type was divided into two discontinuous groups (Figure 1, outer circle). 2) The position and order of ZFs appeared critical for genes containing multiple C2H2-ZFs, as numerous C2H2-ZF genes remained unchanged throughout evolution. Therefore, the ZFs exhibiting the same position and order were classified together (Figure 1, middle and inner circles). For example, the Q-type C2H2-ZF mainly comprises two kinds of ZF genes, those having a single (A) or two (B) ZFs, with the first and second ZFs of B genes being classified together. In addition, the first, second, and third ZFs of group D and E genes, containing three and four ZFs, respectively, were classified together.

Two hundred and twenty-eight transcripts of the C2H2-ZF domain-containing proteins were detected, belonging to 168 corresponding C2H2-ZF genes., Hierarchical clustering was performed to visualize the global transcription profile of the C2H2-ZF genes across the 11 maize organs and to understand the temporal and spatial transcription patterns in different developmental stages. As illustrated in Supplementary Figure S1, the resultant heatmap could be divided into three clusters. Clusters 1, 2, and 3 comprise 92, 39, and 38 members, respectively. Notably, the ZmZFP236 gene has six transcripts, of which transcripts 1, 2, 3, and 4 belong to cluster 2, whereas transcripts 5 and 6 belong to cluster 3. Genes in cluster 1 exhibit relatively low expression, with the mean of the log-signal values of each gene ranging from 6 to 8. In contrast, cluster 3 contains genes with relatively high expression, with the mean of the log-signal values of each gene within the range of 12–14. In cluster 2, genes are expressed over a large range (from 6 to 14). Furthermore, 33.3 and 39.5% of genes in clusters 2 and 3, respectively, produce at least two transcripts, whereas only 5.5% of cluster 1 genes generate multiple transcripts.

To identify putative differentially expressed genes in specific organs or stages, we calculated the coefficient of variation (CV value; CV = SD/mean; where, SD represents the standard deviation and mean represents the mean expression level of each gene across all the tissues in the three clusters) to compare the degree of variation of each gene among distinct organs (Wei K.-F. et al., 2012; Wei K. et al., 2012) (Supplementary Table S3). The results showed that the CV of the expression of these transcripts in 11 maize tissues or organs exhibited a large range, from 1.4 to 39.9%. The gene with the lowest CV was ZmZFP195, belonging to cluster 3, highly expressed in the 11 maize tissues or organs (mean = 13.847), indicating that this gene may be a housekeeping gene in maize. To the best of our knowledge, there are no other reports of this gene in maize.

The CVs of cluster 3 were the smallest (1.4–13.2%), with the CV of 72% genes 46) being <5%. The results indicated that most of the genes in this group exhibited consistent expression patterns relative to those other C2H2-ZF genes. The gene with the highest expression variation in this group was ZmZFP033 (ZmZFP033.1, ZmZFP033.2, and ZmZFP033.3) with a CV of 13.2%. This gene showed low expression in filaments (R1_Silks) but was highly expressed in most other tissues or organs, especially at 12–24 days after-pollination (DAP) in the whole seed and endosperm. In cluster 2, the CVs of 30% of the genes (20) were >17%, with the CVs of 15 genes exceeding 20%, including the ZmZFP056 gene with the highest CV. The results indicated that most of the genes exhibit tissue-specific expression in this group. In particular, the expression CVs were 20 and 22% for the ZmZFP144 and ZmZFP178 (ZmZFP178.1 and ZmZFP178.2) genes, respectively, which are highly expressed at 16–24 DAP in embryos and immature cobs. In turn, the ZmZFP098 gene was highly expressed in the leaves of maize during the whole growth period, having an expression CV of 19%, suggesting that the gene might be involved in the development of leaf or response to light. The ZmZFP152 gene, highly expressed at 2–24 DAP in the whole seed, innermost husk, and pericarp, had an expression CV of 18%, indicating that the gene is involved in the development of grain. The ZmZFP142 and ZmZFP166 genes had the same CV (20%) and were highly expressed in the leaves and anthers (R1_Anthers), and the gene structure and protein domains of these two genes are very similar; they belong to group III. The results suggested that these two genes perform the same function and may regulate reproductive growth.

The CVs of gene expression in cluster 1 were relatively stable compared with those in cluster 2, with expression CVs of 60% of the genes (59) being <10%, 19% of the genes 19) > 15%, and 7% genes 7) > 20%. The CVs of the ZmZFP035, ZmZFP128, and ZmZFP209 genes were 21, 23, and 18%, respectively. These three genes were highly expressed in the anther (R1_Anthers). In turn, the CVs of genes ZmZFP247 and ZmZFP074 (ZmZFP074.1, ZmZFP074.2, and ZmZFP074.3) were 20 and 19.7%, respectively, both of which were highly expressed in meiotic tassels (V18_Meiotic tassel) and exhibited low expression in other tissues or organs. The results indicated that these two genes are closely related to tassel development.

As a ubiquitous transcription factor, the C2H2-ZF genes can regulate the expression of genes related to broad-spectrum stress. One hundred and eighteen probes were found to match 118 transcripts of 112 C2H2-ZF genes in the GPL14913 gene chip (Supplementary Table S4). The log₂ (treated/control) ratio was illustrated using a heatmap (Figure 2), showing the fold change of each maize C2H2-ZF gene under low temperature (T), nitrogen (N), and phosphorus (P) conditions compared with the control (Figure 2; Supplementary Table S5). Under low T, N, and P conditions, the expression of 15, 10, and 8 genes changed by more than 2-fold (| Log₂ (treated/control) |> 1) compared with that of the control (Supplementary Table S5). The maize C2H2-ZF genes exhibited different response patterns to different stress conditions, such as downregulated expression of the ZmZFP106 gene under low N conditions, whereas no expression changes under low T and low P conditions. The response only to low N stress suggested that the gene may play an important negative regulatory role under low N conditions. In turn, the ZmZFP072 and ZmZFP181 genes were downregulated only under low T conditions but not under low N and low P conditions, which indicated that the two genes were mainly responsive to low T stress and had negative regulatory effects. The expression of the ZmZFP101 and ZmZFP150 genes was significantly upregulated under low T conditions; however, the ZmZFP150 gene expression was not changed under low N and low P conditions, whereas ZmZFP101 expression was down-regulated under low N condition, but not under low P condition. The results indicated that the ZmZFP101 and ZmZFP150 genes exhibit co-expression characteristics under low T conditions and stable low expression in various tissues or organs throughout the growth period of maize. The expression patterns of ZmZFP160 and ZmZFP053 showed similar responses under low T conditions, whereas the expression of both was significantly higher under low P conditions than that under low N conditions. These genes exhibit relatively stable and low expression in maize tissues or organs (Supplementary Figure S1). Notably, 25 maize C2H2-ZF genes exhibited > 2-fold changes in expression under low T, low N, and low P conditions, and 17 of which belonged to cluster 1 and showed stable expression in all tissues or organs at maize developmental stages (Supplementary Figure S1: Cluster 1). This indicated that these genes likely only play a role under stress conditions. In addition, four genes (ZmZFP023, ZmZFP056, ZmZFP144, and ZmZFP178) belong to cluster 2, of which ZmZFP056 was highly expressed at 12–24 DAP the whole seeds and endosperm but also exhibited negative regulation under low N stress. ZmZFP060, belonging to cluster 3, showed stable and high expression in various tissues or organs at maize development stages, exhibited considerably decreased expression under low N stress and appeared to play a positive regulatory role under low T conditions.

A drought-sensitive maize inbred line P2 (Ac7729 × TZSRW) and drought-tolerant maize inbred line P1 (Ac7643) were used to investigate the effect of maize C2H2-ZF genes on drought stress resistance. The RNA of these two inbred lines under simulated drought conditions was extracted and sequenced. One hundred and ninety maize C2H2-ZF genes were detected from the RNA sequencing data (Supplementary Table S6). The other 50 C2H2-ZF genes were not detected in either inbred strain upon treatment or control, although the expression of 32 of these C2H2-ZF genes had been detected in the whole growth stage of maize (Supplementary Figure S1; Supplementary Table S3). We speculated that these 32 genes might be expressed under specific conditions and specific tissues. In P1 and P2 inbred lines under drought conditions, we identified 63 and 54 C2H2-ZF genes with 2-fold downregulated expression and 18 and 17 genes with >2-fold downregulated expression (Figure 3; Supplementary Table S6). The results showed that the C2H2-ZF genes were mainly subjected to negative regulation under drought stress, although some showed positive regulation effects. The number of C2H2-ZF genes expressed was significantly increased under drought stress compared with that under low T, N, and P stress (Supplementary Table S5; Supplementary Table S6), indicating that drought regulation is a more complex system that requires the involvement of more genes, and also revealing that the C2H2-ZF genes play an important role in regulating drought responses in maize.

A total of 2998 SNPs of 159 C2H2-ZF genes were extracted from an association panel of 513 inbred maize lines, which contained 556,809 high-quality SNPs. Of these, 664 SNPs of 107 C2H2-ZF genes were significantly associated with 20 adult-plant stage agronomic traits (Supplementary Table S7). Specifically, 56% C2H2-ZF genes significantly associated with seed traits (e.g., 100-grain weight, kernel width, kernel length, and kernel thickness), 47.7% significantly associated with ear traits (e.g., ear diameter, ear length, ear row number, and kernel number per-row), 32.7 and 17.8% significantly associated with plant height and ear height, respectively, and 16.8% significantly associated with flowering traits (e.g., heading date, pollen shed, and silking time) (Supplementary Table S7). For example, all SNPs of the ZmZFP077, ZmZFP091, ZmZFP110, ZmZFP138, ZmZFP150, ZmZFP144, ZmZFP169, ZmZFP183, and ZmZFP191 genes were significantly associated with seed traits. In most cases, the seed and ear traits were genetically linked. Most SNPs of the ZmZFP024, ZmZFP064, ZmZFP095, ZmZFP227, ZmZFP124, ZmZFP238 and ZmZFP242 genes were significantly associated with flowering traits. In addition, 241 SNPs of 60 C2H2-ZF genes were significantly associated with 19 shoot and root traits in the seeding stage under normal afforded (AP) and non-afforded P conditions (NAP) (Supplementary Table S8), of which 16 genes (e.g., ZmZFP006, ZmZFP056, ZmZFP193, and ZmZFP202) were only significantly associated with the shoot and root traits under NAP, whereas 27 were associated under both AP and NAP (Supplementary Table S8). Moreover, every SNP (except chr6: 39893953 site) of the ZmZFP126 gene was significantly associated with at least four traits, and even the chr6: 39894469 sites was significantly associated with 24 traits under normal afforded and NAP.

Sixty-seven C2H2-ZF genes with known functions (Supplementary Table S9) have been reported to date, of which 66 have homologous genes in maize. Figure 4 shows the relationships between the 11 C2H2-ZF genes with known functions and 40 maize C2H2-ZF genes. The AtYY1 gene in Arabidopsis (Li et al., 2016) has three homologous genes, namely, ZmZFP126, ZmZFP093, and ZmZFP117. The AtYY1 gene is a novel negative regulator of the Arabidopsis ABA response network. Association analysis results showed that the maize ZmZFP126 gene has four SNP loci (chr6.S_39893949, chr6.S_39893950, chr6.S_39893951, and chr6.S_39893952) that are associated with 100-grain weight, two SNP loci (chr6.S_39893951 and chr6.S_39894217) that are significantly associated with kernel length, and one SNP locus (chr6.S_39894034) that are significantly associated with tassel main axis length. In addition, this gene has 13 SNP sites that are simultaneously significantly associated with multiple root traits under NAP, such as the root forks under NAP (RF-NAP), root surface area 1 (the average root surface area in diameter between 0.0–0.5 mm of five plants) under NAP (RSA1-NAP). The maize ZmZFP117 gene has two SNP loci (chr5.S_172019275 and chr5.S_172019305) that are significantly associated with leaf number above the ear. In addition, one SNP (SYN19115) is significantly associated with the root forks under AP (RF-AP), root length 1 (the average root length in diameter between 0.0–0.5 mm of five plants) under AP (RL1-AP), and total root length under AP (TRL-AP) traits. The functions of the Arabidopsis AtSTOP1 gene (Iuchi et al., 2007; Sawaki et al., 2009), the Gossypium hirsutum GhSTOP1 gene (Kundu et al., 2018), the rice OsART1/STAR3 gene (Yamaji et al., 2009; Tsutsui et al., 2011), and the sorghum SbSTOP1 genes (SbSTOP1a, SbSTOP1b, SbSTOP1c, and SbSTOP1d) (Huang et al., 2018; Gao et al., 2019), are all associated with aluminum stress, and these genes are homologous to the maize ZmZFP031, ZmZFP118.1, ZmZFP194.1, ZmZFP077, and ZmZFP171 genes. The maize ZmZFP118 gene has four SNP loci (chr5.S_175685185, chr5.S_175689281, chr5.S_175689296, and chr5.S_175689575) that are significantly associated with kernel thickness and kernel number per-row and two SNP loci (chr5.S_175687470 and chr5.S_175688702) that are significantly associated with kernel length. In addition, two SNP loci of this gene are significantly associated with leaf number under NAP (LN-NAP), and four SNP loci are significantly associated with root volume 2 (the average root volume in diameter between 0.5–1.0 mm of five plants) under NAP (RV2-NAP). The maize ZmZFP194 gene has three SNP loci (chr10.S_10140760, chr10.S_10140901, and chr10.S_10140902) that are significantly associated with ear leaf width. The chr10.S_10140760 locus is also significantly associated with the ear diameter. Two SNP loci (chr10.S_10140916 and chr10.S_10143604) are significantly associated with leaf number above the ear, and the chr10.S_10142324 SNP locus is significantly associated with kernel length. In addition, the chr10.S_10136241 SNP locus of this gene is significantly associated with multiple seedling traits in maize under NAP (e.g., RDW-NAP, RL2-NAP, RSA2-NAP, RV2-NAP, TDW-NAP, TFW-NAP, and TRV-NAP). The Arabidopsis AtWIP2/NTT gene is associated with fruit dehiscence and replum development (Chung et al., 2013; Marsch-Martínez et al., 2014), and this gene is homologous to the maize ZmZFP132, ZmZFP163, ZmZFP049, ZmZFP139, ZmZFP151, ZmZFP020, ZmZFP100, ZmZFP237, ZmZFP131, and ZmZFP177 genes. However, no SNP locus significantly associated with maize traits has been detected in these genes. The Arabidopsis AtSGR5 gene (Morita et al., 2006; Kim et al., 2015) is involved in early events of gravitropism in Arabidopsis inflorescence stems. This gene is homologous to the maize ZmZFP005, ZmZFP189, ZmZFP021, ZmZFP055, and ZmZFP150 genes. The SNP locus chr9.S_144035764 in the maize ZmZFP189 gene is significantly associated with ear leaf length. The ZmZFP150 gene has two SNP loci (chr7.S_124146886 and SYN38697) that are also significantly associated with kernel length. The Arabidopsis AtIDD1/ENHYDROUS gene promotes the transition to germination by regulating light and hormonal signaling during seed maturation (Feurtado et al., 2011; Fukazawa et al., 2014). This gene is homologous to 17 maize C2H2-ZF genes (Figure 4). The ZmZFP060 gene has four SNP loci (chr2.S_207380320, chr2.S_207381507, chr2.S_207381508, and chr2.S_207381537) that are significantly associated with kernel length. Three among them (chr2.S_207381507, chr2.S_207381508, and chr2.S_207381537) are simultaneously significantly associated with ear length. The chr2.S_207380848 and chr2.S_207380851 loci of this gene are significantly associated with 100-grain weight, while the chr2.S_207380910 and chr2.S_207380911 loci of the same gene are also significantly associated with cob diameter and ear diameter. Four SNPs (chr2.S_207378625, chr2.S_207378663, chr2.S_207380989, and chr2.S_207381192) are significantly associated with tassel branch number. The chr2.S_207379647 and chr2.S_207379650 loci are significantly associated with plant height. Four SNP loci of the ZmZFP155 gene (chr7.S_160601597, chr7.S_160601615, chr7.S_160601621, and chr7.S_160601626) are significantly associated with three flowering traits: heading date, pollen shed, and silking time. The chr7.S_160601466, chr7.S_160601469, chr7.S_160601395, and chr7.S_160600519 loci are significantly associated with kernel number per-row. This gene has seven SNP loci (chr7.S_160598451, chr7.S_160599436, chr7.S_160599437, chr7.S_160599580, chr7.S_160600407, chr7.S_160600877, and chr7.S_160601398) that are significantly associated with kernel length. The chr7.S_160600894 SNP locus of the gene is significantly associated with LN-NAP, and the chr7.S_160600586 and chr7.S_160600590 loci are significantly associated with the longest root length under AP (LRL-AP). In the ZmZFP024 gene, three SNP loci (SYN38850, SYN38852, and PZB02144.4) are significantly associated with silking time. In the ZmZFP119 gene, 27 SNPs are significantly associated with ear leaf length, ear length, kernel length, plant height, and tassel main axis length, and 13 of these are significantly associated with kernel length. In addition, the gene has 14 SNP loci that are significantly associated with seedling traits under AP, of which three SNP loci (chr5.S_194170562, chr5.S_194170563, and chr5.S_194170564) are significantly associated with root/shoot ratio under AP (R/S-AP). In the ZmZFP033 gene, seven SNP loci are significantly associated with kernel width. The chr2.S_17095383 locus is also significantly associated with 100-grain weight and ear diameter. In addition, the gene has four SNP loci that are significantly associated with seedling traits under NAP. In the ZmZFP208 gene, 22 SNP loci are more associated with 100-grain weight, ear row number, kernel thickness, kernel width, kernel length, and other grain-related traits. This gene has four SNP loci more associated with root dry weight under AP and NAP (RDW-AP and RDW-NAP). Three SNP loci are significantly associated with R/S-AP. In the ZmZFP023 gene, three SNPs (chr1.S_206123397, chr1.S_206123974, and SYN14326) are significantly associated with cob weight, ear length, and ear diameter in maize, respectively. Four SNP loci in the ZmZFP153 gene are significantly associated with grain traits (kernel thickness, ear row number, kernel width, and kernel thickness). In addition, the gene has two SNP loci that are significantly associated with three NAP traits: shoot dry weight under NAP (SDW-NAP), total fresh weight NAP (TFW-NAP), and plant height under NAP (PH-NAP). The nine SNP loci in the ZmZFP056 gene are all significantly associated with tassel branch number. Among them, the chr2.S_197936812 locus is also significantly associated with kernel width. In addition, this gene has 18 SNP loci that are significantly associated with multiple seedling traits under NAP, and each SNP locus is significantly correlated with LN-NAP. The ZmZFP003 gene has five SNPs significantly associated with grain traits (e.g., kernel thickness and kernel length) and ear traits (ear diameter and ear length). One SNP locus in the ZmZFP075 gene is significantly associated with kernel thickness. The chr3.S_5884603 locus of the ZmZFP069 gene is associated primarily with flowering traits (heading date, pollen shed, and silking time), and the chr3.S_5886671 and chr3.S_5886865 loci are significantly associated with ear diameter. The chr3.S_5884749 locus of this gene is significantly associated with multiple seedling traits under AP. Eight SNPs in the ZmZFP168 gene are significantly associated with kernel length. The chr8.S_139345593 locus is significantly associated with both ear length and kernel number per-row. The three SNP loci PZE-108082637, chr8.S_139345593, and chr8.S_139346331 are also significantly associated with root traits of maize seedlings under AP.

The Zoysia japonica ZjZFN1 gene is involved in salt stress in Arabidopsis (Teng et al., 2018). This gene is homologous to the maize genes ZmZFP026 and ZmZFP105. The association analysis results showed that the ZmZFP026 gene has six SNP loci associated with female ear traits (e.g., cob diameter, ear diameter, cob weight, and ear length). In addition, the chr1.S_279912850 locus is significantly associated with the longest root length under NAP (LRL-NAP), and the chr1.S_279912890 locus is significantly associated with plant height under AP (PH-AP). The ZmZFP105 gene has two SNP loci (chr5.S_6315680 and chr5.S_6316039) that are significantly associated with 100-grain weight, four SNP loci (chr5.S_6316055, chr5.S_6316450, chr5.S_6316451, and chr5.S_6316452) that are significantly associated with kernel length, and the chr5.S_6316039 and chr5.S_6316452 loci that are significantly associated with ear length and ear diameter, respectively. In addition, the chr5.S_6315637 locus is significantly associated with leaf number under AP (LN-AP). The Arabidopsis AtCZS/SUVR5 gene encodes a histone methyltransferase that functions together with its binding partner LDL1/SWP1 as one of the regulators of flowering timing in Arabidopsis (Krichevsky et al., 2007; Grewal et al., 2012); it is homologous to three maize genes: ZmZFP096, ZmZFP085, and ZmZFP120. Five SNP loci in the ZmZFP096 gene are significantly associated with plant height; two SNP loci with cob diameter; and one locus (chr4.S_160935992), with kernel length and kernel number per row. The chr5.S_201600785 locus of ZmZFP120 is significantly associated with ear height and root diameter under NAP (RD-NAP). The rice OsEMF2B and OsEMF2A (Li et al., 2009; Luo et al., 2009), Arabidopsis AtEMF2 (Yoshida et al., 2001), and cabbage BoEMF2.1 and BoEMF2.2 genes (Liu et al., 2012) regulate diverse genetic pathways and are required for vegetative development and plant survival. These genes are homologous to the maize ZmZFP241 gene, and the chr7.S_84537039 locus of ZmZFP241 is significantly associated with cob weight. The Arabidopsis AtSE gene controls leaf development, meristem activity, inflorescence architecture, and developmental phase transition (Prigge and Wagner, 2001; Yang et al., 2006). This gene has four homologous genes in maize, namely, ZmZFP214, ZmZFP227, ZmZFP238, and ZmZFP234. The ZmZFP214 gene has four SNP loci that are significantly associated with tassel branch number and three SNP loci that are significantly associated with ear leaf width. In addition, this gene has two SNP loci that are significantly associated with root traits under NAP, and the other 12 SNP loci are significantly associated with root traits under AP. Seven SNPs in the ZmZFP227 gene are simultaneously significantly associated with ear leaf length, heading date, plant height, pollen shed, and silking time. The chr4.S_42145910 and chr4.S_42145955 loci are also significantly associated with the flowering traits of pollen shed and silking time. Eight SNP loci are significantly associated with kernel length, and the chr4.S_42148067 and chr4.S_42148068 loci are significantly associated with kernel thickness. The ZmZFP238 gene has six SNP loci that are significantly associated with silking time and two SNP loci that are significantly associated with kernel length. The Arabidopsis AtJAG (Dinneny et al., 2004; Ohno et al., 2004; Dinneny et al., 2006) and AtJGL (Dinneny et al., 2006) and rice OsSL1(Xiao et al., 2009) genes are homologous to the maize ZmZFP159 gene. AtJAG and AtJGL define stamen and carpel shape in Arabidopsis, and OsSL1 regulates floral organ identity in rice. The association analysis results showed that the PZE-108011044 locus of the ZmZFP159 gene is significantly associated with kernel number per row, and the PZE-108011046 locus is significantly associated with plant height and tassel main axis length.

ZmMRPI-1 (ZmZFP174) and ZmMRPI-2 (ZmZFP080) (Royo et al., 2009) interact with ZmMRP-1 and modulate its activity on transfer cell-specific promoters. The results of association analysis showed that the chr8.S_171810459 and chr8.S_171805538 loci of the ZmZFP174 gene are significantly associated with the kernel width and RT2-AP traits, respectively. The chr3.S_190064606 locus of the ZmZFP080 gene is significantly associated with kernel thickness, and the chr3.S_190065192 locus is significantly associated with multiple root traits under NAP. The Arabidopsis AtSUF4 (Resentini et al., 2017) gene supports gamete fusion via regulating Arabidopsis EC1 gene expression. This gene is homologous to the ZmZFP058 gene, and the chr2.S_198580873 locus of the ZmZFP058 gene is significantly associated with multiple traits, including heading date, pollen shed, silking time, kernel length, leaf number above the ear, ear leaf length, ear length, and plant height. The chr2.S_198582212 and chr2.S_198582325 loci are significantly associated with LN-NAP. The Arabidopsis SUPERMAN/FLO10/SUP gene is required for maintaining the boundaries between stamens and carpels and for regulating the development of the outer ovule integument (Sakai et al., 1995; Yun et al., 2002; Isernia et al., 2003; Ito et al., 2003; Prunet et al., 2017; Xu et al., 2018). The cucumber SUPERMAN gene has a conserved function in stamen and fruit development and a distinct role in floral patterning (Wu et al., 2014). The petunia PhSUP1 gene plays a distinct role in floral organ morphogenesis (Nakagawa et al., 2004), and the SlSUP gene has a positive role in the female flower developmental pathways of S. latifolia (Kazama et al., 2009). These genes are homologous to the maize ZmZFP087 gene, and the results of association analysis showed that the chr4.S_610398 locus of the ZmZFP087 gene is significantly associated with cob weight. The Arabidopsis AtZFP8 gene plays partially redundant and essential roles in inflorescence trichome initiation and its regulation by gibberellic acid and cytokinins (Gan et al., 2007). This gene is homologous to the maize ZmZFP122 gene, and the chr5.S_211549650 and chr5.S_211549874 loci of the ZmZFP122 gene are significantly associated with pollen shed and ear row number, respectively. In addition, the chr5.S_211549697 and chr5.S_211549790 loci of this gene are significantly associated with multiple maize seedling traits under AP.

Through cloning and sequencing, ZmZFP126 gene sequences of 109 inbred maize lines (Supplementary Table S1) from different countries and regions were obtained, and the gene sequences were analyzed. The SNPs and Indels loci in the sequence were extracted and analyzed for correlation with 20 agronomic traits of maize adult-plant stage (Supplementary Table S10; Supplementary Figure S2) and 19 shoot and root traits in the maize seedling stage under AP and NAP (Figure 5; Table 3). The results showed that 31 SNPs and 11 Indel loci of the ZmZFP126 gene were significantly correlated with six agronomic traits (Supplementary Table S10). A total of 29 SNPs and 10 Indel loci of the gene were significantly correlated with an ear row number (ERN) (Supplementary Table S10; Supplementary Figure S2). In addition, from 24 SNPs and 6 Indel loci which were simultaneously significantly correlated with SDW-NAP, 8 SNPs and only one Indel were significantly correlated with a total dry weight under NAP (TDW-NAP) (Figure 5; Table 3). This result is consistent with the conclusion drawn from the correlation analysis, indicating that the ZmZFP126 gene plays a key role in the development of the ear, the female flower, in corn, and low phosphorus stress. Linkage disequilibrium (LD) analysis shows that the LD degree of the ZmZFP126 gene is very high from the 5′-end to the 3′-end. Almost the entire gene region forms an LD block, and the LD attenuation rate of the gene remains almost unchanged. It is at a high level of r ² = 0.6 (Figure 5D). LD analysis results show that this gene is very conservative during evolution, and large fragment recombination rarely occurs.