Seeds of maize inbred line B73 were cultivated in a growth chamber under conditions of 16h light, 25 °C, and 8h darkness, 22 °C. The seedlings at the three-leaf stage were subjected to the corresponding stress treatments (Guan et al., 2025). For drought treatment, the seedlings were watered with 35% PEG-6000 (w/v) for 0, 1, 3, 6, 12, and 24 h, respectively. For high salt treatment, the seedlings were watered with 200 mM NaCl solution. Samples were collected at 0, 1, 3, 6, 12 and 24 h after the treatment. Seedlings were subjected to heat stress treatment in a plant growth chamber maintained at 42 °C for 0, 4, and 8 h, respectively. For hormone treatment, seedlings were sprayed with 100 μM abscisic acid (ABA), and collected at various time intervals (0, 1, 3, 6, 12, and 24 h). Seedlings with no treatment (0 h) served as control. Leaf samples were harvested at the designated time points and snap-frozen in liquid nitrogen, then stored at -80 °C for subsequent RNA extraction.

The genomes fasta file, gff3 file, and protein fasta file of four species, including Zea mays (Zm-B73-REFERENCE-NAM-5.0.55), Arabidopsis thaliana (TAIR10), Sorghum bicolor (NCBIv3), and Oryza sativa subsp. japonica (IRGSP), were downloaded from Phytozome database (https://phytozome-next.jgi.doe.gov) (Goodstein et al., 2012) (accessed on 10 March 2025). The structural domain file Pfam-A.hmm was downloaded from the Pfam database (http://pfam.xfam.org/) (Mistry et al., 2021) (accessed on 11 March 2025). The Simple HMM Search function in the TBtools V2.110 (Chen et al., 2023) software was used to identify the maize LSD family, by using all the protein sequences of maize and the structural domain login number of the LSD family (PF06943.16, LSD1 zinc finger), and the E-value was set to 10^-5. The Domain of the protein obtained was analyzed by using the online software NCBI Batch CD-Search (https://www.ncbi.nlm.nih.gov/Structure/bwrpsb/bwrpsb.cgi) (Wang et al., 2023) (accessed on 12 March 2025). The common sequence of the zinc finger structural domain of the LSD1 gene family was checked by aligning the sequences with MEGA V11.0 (Mega Limited, Auckland, New Zealand) (Tamura et al., 2021).

Physicochemical characteristics such as the number of amino acids, molecular weight, theoretical isoelectric point, instability index, etc. of ZmLSD family members were analyzed by using the online tool Expasy ProtParam (http://web.expasy.org/protparam/) (Wilkins et al., 1999). The subcellular localization of ZmLSD family members was predicted using the online tool WoLF PSORT (https://wolfpsort.hgc.jp/) (Horton et al., 2007).

The protein sequences of the LSD family members of Arabidopsis, sorghum, and japonica rice were downloaded from PlantTFDB (https://planttfdb.gao-lab.org/index.php?sp=Zma) (Jin et al., 2017). Using the software MEGA V11.0, the phylogenetic tree of the LSD families in these four species was constructed using the neighbor-joining (NJ) method, with bootstrap replicates set to 1000 and other parameters as system defaults (Tamura et al., 2021). The phylogenetic tree was display using the iTOL V6 (https://itol.embl.de/) (Letunic and Bork, 2024).

Based on the genome fasta and gff3 annotation file, the CDS sequences were obtained. According to the genome sequence and CDS sequence, the gene structure of ZmLSD family members can be obtained by using TBtools V2.110 software. The protein conserved motifs of the maize LSD gene family members were analyzed using the online tool MEME (https://meme-suite.org/meme/) (Bailey et al., 2015), and the maximum number of Motifs was set to 3. The Domain of LSD genes was analyzed by using SMART (http://smart.embl-heidelberg.de/) (Letunic et al., 2021), and visual mapping was performed using TBtools V2.110.

From the maize whole genome annotation file (gff3), the position information of ZmLSD family genes on chromosomes was extracted. The chromosome location maps were drawn based on the location of the genes on the chromosomes. The covariance relationship within the maize genome was analyzed using the MCScanX toolkit in the TBtools V2.110 (Wang et al., 2012).

Based on the gff3 annotation file and genome sequences, the CDS and promoter (2000 base pairs (bp) upstream of the start codon ATG) sequences of ZmLSDs were extracted using TBtools V2.110. Cis-acting elements on each ZmLSD promoter was predicted using PlantCARE (http://bioinformatics.psb.ugent.be/webtools/plantcare/html/) (Lescot et al., 2002).

Transcriptome data of ZmLSD gene expression patterns were obtained from the NCBI database accession numbers PRJNA171684 and SRP010680 (Stelpflug et al., 2016), and the heat map was drawn using TBtools software V2.110.

Total RNA was extracted using RNAiso Plus (TaKaRa, Beijing, Japan). After testing for purity and quality using Nanodrop 2000 spectrophotometer (Thermo Scientific), 1 μg of total RNA was reverse-transcribed into first-strand cDNA using Evo M-MLV RT Kit with gDNA Clean for qPCR (Accurate Biology, Hunan) according to the manufacturer’s instructions. The cDNA template was diluted for 30-fold. The qRT-PCR reaction system consisted of 7.5 μL of SYBR Green Pro Taq HS Premix (Accurate Biology, Hunan), 0.3 μL upstream and downstream specific primers, 1.9 μL of ddH₂O and 5 μL template. The reaction program was pre-denaturation at 95 °C for 30 sec; denaturation at 95 °C for 5 sec, annealing and extension at 60 °C for 30 sec, and the samples underwent 45 amplification cycles. After completion, a melting curve was recorded by setting the temperature to start at 65 °C and gradually increased at 0.5 °C/s until 95 °C. Three biological replicates and three technical replicates per target gene were performed for qPCR. Ct values from technical replicates were averaged to reduce noise. ZmActin 1 was regarded as reference gene and the relative expression levels of ZmLSDs were calculated using the 2^-△△Ct method (Schmittgen and Livak, 2008). The primer sequences used were listed in Supplementary Table 1.

All the experimental measurements were repeated for three times. Data processing was performed with Microsoft Excel 2021, while statistical plotting, analysis of variance, and comparisons of differences (Student’s t-test) were completed with GraphPad Prism 6 (GraphPad Software Inc.; San Diego, CA, USA). Significant difference was defined as p < 0.05 (*) and p < 0.01 (**).

For analyzing the LSD gene family in maize, the Simple HMM Search toolkit in the TBtools wasemployed to search LSD genes against local maize genome databases. After NCBI’s Conserved Domain Database (CDD) verification, a total of 23 proteins containing conserved structural domains of the LSD were identified at the genome-wide level, encoded by nine genes location (Supplementary Table 2). The genes were renamed ZmLSD1 to ZmLSD9 based on their location on the chromosome, and 23 protein isoforms were renamed ZmLSD1.1 to ZmLSD9.3 (Table 1). The number of amino acids of the ZmLSD family members varied from 113 to 898, the relative molecular weights ranged from 12.133 KD to 93.568 KD, and the theoretical isoelectric points (pI) ranged from 4.46 to 9.63. Among them, ZmLSD7.2 protein sequence has the shortest length and the smallest molecular weight, which is only 12133.08 Da. There were 14 basic proteins (pI values greater than pH 7.0) and 9 acidic proteins (pI values less than pH 7.0). The instability index of ZmLSD ranged from 33.77 to 77.39, with 9 proteins with instability index less than 40 and 14 proteins with index greater than 40, indicating most ZmLSD protein structures were unstable. The values of grand average of hydropathicity (GRAVY) indicated most proteins were hydrophobic proteins. Subcellular localization prediction results showed that four members were localized in the nucleus, one (ZmLSD3.3) was located in the cytoplasm, and eighteen were localized in the chloroplasts (Table 1). Secondary structure analysis revealed that all ZmLSD proteins contain alpha helix,extended strand, beta turn, and random coil. ZmLSD1 and ZmLSD9 exhibit relatively high proportions of alpha helix and extended strand, whereas ZmLSD5 contains over 90% random coil (Supplementary Table 2). These results demonstrate significant divergence in the basic properties of ZmLSDs, implying functional diversity among family members.

The MEGA software was used to construct a phylogenetic tree containing 55 LSD proteins, including 12 proteins from Arabidopsis, 8 proteins from sorghum, 12 proteins from japonica rice and 23 proteins from maize. The phylogenetic tree was annotated with the online software iTOL. According to the evolutionary relationships, the 55 LSD members can be divided into five subfamilies: LSD1, Group 1, Group 2, LOL1, and LOL2 (Figure 1). In LSD1, there are only Arabidopsis members belonging to dicotyledonous plants, while maize, sorghum and japonica members belonging to monocotyledonous plants are absent. Group 1 contains three ZmLSDs and group 2 has eight ZmLSDs, while LOL1 contains only one ZmLSD member, namely ZmLSD6.1. The largest number of ZmLSD family members belong to LOL2, with eleven members. The results indicate that LSD1, Group 1, and Group 2 have remained highly conserved throughout plant evolution, whilst LOL1 and the LOL2 family underwent expansion during the evolution of monocotyledons and dicotyledons.

We constructed a phylogenetic tree using the protein sequences of 23 maize LSD family proteins. The results showed that ZmLSD protein classification was consistent with Figure 1 (Figure 2A). In order to investigate the protein conservation of ZmLSD family members, MEME online software was used to predict the conserved Motifs of proteins with a maximum of 3 Motifs. Meanwhile, the online software Conserved Domain Database in NCBI was used to predict the domain of ZmLSD family members. The results showed that 23 ZmLSDs all contained Motif1, and there was a correspondence between Motif1 and zf-LSD1 or zf-LSD1 superfamily in position (Figure 2B). Motif1 contained the C2C2 zinc finger structure, and the conserved domain of its zinc finger structure consisted of 22 amino acids with the sequence CxxCxxxLxxxxGAxxxxCxxC (Supplementary Figure 1).

Analysis of the gene structure of ZmLSD family members revealed that ZmLSD family members consist of 4–7 exons. The largest number of members contained five exons, including five genes (nine transcripts), namely ZmLSD4.1, ZmLSD9.2, ZmLSD9.3, ZmLSD9.1, ZmLSD2.1, ZmLSD2.2, ZmLSD1.2, ZmLSD5.4, ZmLSD5.3; followed by genes containing four exons, consisting of seven transcripts (ZmLSD7.1, ZmLSD7.2, ZmLSD8.1, ZmLSD6.1, ZmLSD1.1, ZmLSD5.1, ZmLSD5.2). There were six transcripts containing six exons: ZmLSD4.4, ZmLSD4.3, ZmLSD4.2, ZmLSD3.1, ZmLSD3.3, ZmLSD3.2; One gene containing seven exons (ZmLSD4.5). The ZmLSD7.2 did not have UTR, while the remaining 22 ZmLSD members all contained 5’ and 3’UTRs. The 5’ UTR of ZmLSD6.1 is spaced by an intron, and the 3’ UTR of ZmLSD9.1, ZmLSD9.2, and ZmLSD2.1, are spaced by an intron. The corresponding positions of domains were made in the gene structure, and it was found that Zf-LSD1 and Zf-LSD1 superfamily were mostly distributed on different exon fragments (Figure 2C).

To clearly illustrate the abundance and chromosomal distribution patterns of ZmLSD gene family members, we employed TBtools software for visualization. As demonstrated in Figure 3, nine genes exhibit a non-random distribution across five maize chromosomes, with notable clustering observed on specific chromosomal regions. They are mainly distributed on Chr1 and Chr3, with three genes (corresponding to seven transcripts) and two genes (corresponding to nine transcripts), respectively. Chr4 has two genes (three transcripts), Chr6 has one gene (one transcript), and Chr9 has one gene (three transcripts).

The collinear relationships of ZmLSD family members within the maize genome and among other species, including Arabidopsis, sorghum, and rice were analyzed respectively using the MCScanX function in the TBtools software. There were three intra-species covariance gene pairs for the ZmLSD family members, namely, ZmLSD2.2 and ZmLSD9.1, ZmLSD3.1 and ZmLSD4.3, and ZmLSD7.1 and ZmLSD8.1 (Figure 3). The ZmLSD1.1 and Zm00001eb392460_T002 are also a covariant gene pair, butZm00001eb392460_T002 does not belong to the ZmLSD family (Supplementary Table 3).

Comparative genomics analyses reveals that maize LSD genes exhibited a significant divergence in syntenic relationships between monocot and dicot species. No syntenic gene pairs were detected between the ZmLSD family and dicotyledonous plants, including Arabidopsis thaliana, soybean (Glycine max), or cabbage (Brassica oleracea), indicating lineage-specific genomic reorganization after species divergence. Strikingly, synteny was observed with monocot species, particularly sorghum (Sorghum bicolor) and rice (Oryza sativa subsp. japonica). There are seven ZmLSD transcripts, namely ZmLSD2.2, ZmLSD3.1, ZmLSD4.3, ZmLSD5.4, ZmLSD6.1, ZmLSD8.1, ZmLSD9.1, exhibited syntenic homologs with both sorghum and japonica rice LSD genes (Figure 4). However, ZmLSD7.1 exhibits a unique colinear relationship only with the sorghum LSD gene, not with the rice LSD gene (Figure 4, Supplementary Table 3). These findings indicated that the LSD gene family has followed distinct evolutionary trajectories between monocot and dicot, with stronger functional conservation retained among monocot species.

The promoter sequence of ZmLSD genes was extracted by TBtools, and thecis-elements were analyzed by PlantCARE. Using Excel, three classes of cis-acting elements were screened, which is associated with phytohormone response, abiotic and biotic stresses tolerance, and plant growth and development (Supplementary Table 4). Among them, all ZmLSD family genes contained ABRE and MYC hormone-responsive elements, and all ZmLSDs except for ZmLSD8, contained as-1, CGTCA-motif, and TGACG-motif elements (Figure 5). Among the biotic and abiotic stress-responsive elements, all ZmLSD genes contained MYB elements, and more than 70% of ZmLSD family members contain ARE, MYB-like sequence, and STRE cis-elements (Figure 5). The promoter of ZmLSDs also contain cis-elements associated with plant growth and development responses. For instance, ZmLSD7, ZmLSD6, ZmLSD4, and ZmLSD9 contain the CAT-box, expressed in meristematic tissues. The promoter of ZmLSD7, ZmLSD8, ZmLSD6, and ZmLSD9 contain the GCN4_motif, expressed in the endosperm. The ZmLSD5 promoter has a plant_AP-2-like element expressed in the seed. The ZmLSD7 promoter contains an RY-element. These results suggested that LSD genes may participate in regulating plant growth and development, and stress response.

To investigate the expression patterns of the ZmLSDs in different tissues duringmaize development, we analyzed genes’ fragments per kb exon model per million mapped fragments (FPKM) values using previously reported RNA-seq data (Supplementary Table 5). The results showed that among all tissues, only embryos and endosperm exhibit lower expression levels of ZmLSDs (Figure 6). The expression of ZmLSD genes in other parts remained relatively stable throughout all stages of growth and development. The expression levels of ZmLSD3, ZmLSD4, and ZmLSD9 were relatively high in all detected tissues, while ZmLSD1, ZmLSD7 and ZmLSD8 were relatively lower. The ZmLSD6 expression level was high in some tissues, such as internode, nonpollinated leaf. The above results indicate that the expression of ZmLSDs exhibit significant differences and specificity in tissue, suggesting that this family may possess diverse functions during maize growth and development.

Previous studies have demonstrated that the LSD family members can participate in responses tobiotic and abiotic stress (Guan et al., 2016; Jiang et al., 2019; Bernacki et al., 2021; Chao et al., 2024). To investigate the expression patterns of ZmLSD genes under drought, salt, heat, and hormone ABA stresses, we employed qRT-PCR to detect genes’ expression levels after PEG-6000, NaCl, high temperature, and ABA treatments, respectively (Supplementary Table 6). The results showed that after PEG-6000 simulated drought treatment, the expression of genes ZmLSD2, ZmLSD3, ZmLSD4, ZmLSD5, ZmLSD6, and ZmLSD7 were significantly down regulated at all time points except for ZmLSD8 and ZmLSD1 (Figure 7). The expression levels of ZmLSD1 were significantly downregulated at 1h, 6h, and 9h, after drought stress treatment (Figure 7). These findings indicated that ZmLSDs participate in maize’s response to drought stress pathway. After salt stress treatment, the expression levels of ZmLSDs exhibited an overall pattern of initial increase followed by decline (Figure 8). The expression levels of ZmLSD3 markedly decreased at 12h and 24h after NaCl treatment (Figure 8). The expression levels of ZmLSD5 and ZmLSD6 showed significant upregulation at 3h post-NaCl treatment, while ZmLSD7 exhibited marked upregulation at 1h post-treatment; ZmLSD6 and ZmLSD7 then exhibited significant downregulation at 6h, 12h, and 24h post-NaCl treatment (Figure 8). These results indicated that prolonged salt stress suppresses the expression of ZmLSD family genes. After 4h and 8h of heat treatments, expression levels of all ZmLSDs showed a downregulation trend, with the exception of ZmLSD2, whose expression at 8h remained largely unchanged compared to the control. Notably, ZmLSD3, ZmLSD4, ZmLSD6, ZmLSD7, and ZmLSD8 exhibited significantly reduced expression levels (Figure 9). These findings indicated that ZmLSD family genes participate in maize’s response to heat stress. After ABA hormone treatment at different time points, the relative expression levels of most ZmLSDs exhibited a pattern of initial increase, followed by decrease, then rise, and subsequent decline (Figure 10). The expression levels of ZmLSD6 exhibited most robust ABA responsiveness. Compared to the control, the expression levels of ZmLSD6 showed significant changes at 1h, 3h, 12h, and 24h post-treatment. Notably, its expression level was more than threefold increased after the 1h treatment. In addition, ZmLSD2, ZmLSD3, and ZmLSD5 exhibited significant upregulation compared to the control at 12 hours, while ZmLSD4 showed marked downregulation at 3h after ABA treatment (Figure 10). These findings indicated that ZmLSDs participate in stress responses to drought, high temperature, salt stress, and the hormone ABA. When subjected to stress, some ZmLSDs exhibit consistent expression profile, suggesting that their functions may be redundant.

Determining the subcellular localization of proteins is crucial for studying gene’s function. The PSORT website predicts that most ZmLSDs are localized in the nucleus, cytoplasm, and chloroplast (Table 1). To further confirm the localization of this protein family, we selected ZmLSD3 and ZmLSD4, the expression of which showed significantly changed under drought, high temperature, salt stress, and hormonal stress. After designed specific primers (Supplementary Table 1), they were cloned from B73 maize using molecular biology techniques. We then constructed 35S::ZmLSD3-eGFP and 35S::ZmLSD4-eGFP expression vectors (Figure 11A). Following transient transformation of maize protoplasts, subcellular localization of the fusion proteins was examined using laser confocal microscopy. The results revealed that ZmLSD3-eGFP predominantly localized to the cytoplasm, while ZmLSD4-eGFP localized to both the nucleus and cytoplasm (Figure 11B). These findings indicated that ZmLSD family members may exert significant functions not only in the nucleus and chloroplasts, but also within the cytoplasm.