ZmHsp18 screened from the ZmHsp20 gene family confers thermotolerance in maize

Abstract

Heat stress has become one of the abiotic stresses that pose an increasing threat to maize production due to global warming. The Hsp20 gene family confers tolerance to various abiotic stresses in plants. However, very few Hsp20s have been identified in relation to maize thermotolerance. In this study, we conducted a comprehensive study of Hsp20s involved in thermotolerance in maize. A total of 33 maize Hsp20 genes (ZmHsp20s) were identified through scanning for a conserved α-crystalline domain (ACD), and they were categorized into 14 subfamilies based on phylogenetic analysis. These genes are distributed across all maize chromosomes and nine of them are in regions previously identified as heat-tolerance quantitative trait loci (hrQTL). These hrQTL-associated ZmHsp20s show variation in tissue-specific expression profiles under normal conditions, and seven of them possess 1–5 heat stress elements in their promoters. The integration of RNA-seq data with real-time RT-PCR analysis indicated that ZmHsp23.4, ZmHsp22.8B and ZmHsp18 were dramatically induced under heat stress. Additionally, these genes exhibited co-expression patterns with key ZmHsfs, which are crucial in the heat tolerance pathway. When a null mutant carrying a frame-shifted ZmHsp18 gene was subjected to heat stress, its survival rate decreased significantly, indicating a critical role of ZmHsp18 in maize thermotolerance. Our study lays the groundwork for further research into the roles of ZmHsp20s in enhancing maize’s thermotolerance.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12870-024-05763-5.

Introduction

High temperatures, an increasingly threatening stress factor for crop production, have emerged as a focal area in agricultural research [1]. Heat stress, defined as sustained high temperatures exceeding the optimal growth range, adversely affects crop growth and reduces yield per unit area [2]. For maize, the optimal growth temperatures range from 25–33 ℃ during the day to 17–23 °C at night [3]. Heat stress can cause considerable damage to maize throughout its entire lifecycle, particularly during the critical flowering and grain filling stages [4–6]. For instance, under high temperature stress, maize undergoes a decrease in photosynthetic intensity [7]. To combat heat stress, a variety of heat shock proteins (Hsps) have been identified, and some of them have been proven to enhance heat stress tolerance in plants [8–12].

Hsps in eukaryotes are classified into several families, including Hsp100, Hsp90, Hsp70, Hsp60, and Hsp20. The Hsp20 family, also known as small Hsps, is conserved across different species and consists of small Hsps with a monomer molecular mass ranging from 12 to 42 kilodaltons (KDa) [13, 14]. All Hsp20s are characterized by a highly conserved alpha-crystalline domain (ACD) of about 90 amino acid residues, flanked by N- and C-terminal regions [15–17], but not all the ACD-containing proteins belong to the Hsp20 family [13]. In angiosperms, members of the Hsp20 family are further divided into several subfamilies based on their subcellular localizations. These include cytoplasmic/nuclear-localized C-I, C-II, C-III, C-IV, C-V, C-VI, and other cell compartments, such as ER (endoplasmic reticulum), CP (chloroplasts), M-I (mitochondria I), M-II (mitochondria II), Px (peroxisomal) [14], and P (plastidial) [18, 19].

Hsp20s function as ATP-independent molecular chaperones and accumulate in order to protect plants by preventing protein aggregation and facilitating refolding of damaged proteins under heat stress conditions [20]. They may participate in heat tolerance through pathways dependent on heat shock transcription factors (Hsf) [15, 21–24]. In Arabidopsis, HsfB1 and HsfB2b were reported to be required for the expression of ten Hsp20s under heat stress conditions, which is necessary for acquired thermotolerance in Arabidopsis [25]. Hsa32 is required not for the induction but for the maintenance of acquired thermotolerance [26], while Hsp21 regulates thermomemory in collaboration with FtsH6 [27]. Exogenous Hsp20s also work in transgenic Arabidopsis for heat stress tolerance, like Hsp26 from wheat [28], Hsp17.8 from Rosa chinensis [29], and Hsp18.2 from rice [30]. In rice, the silencing of class I small heat shock proteins decreased seedling thermotolerance, while the overexpression of sHsp17.7 increased it [31, 32]. Additionally, the up-regulation of five Hsps (Hsp26.7, Hsp23.2, Hsp17.9A, Hsp17.4, and Hsp16.9A) may also play important roles in rice resistance to heat stress [33]. In wheat, HsfA6f can directly regulate Hsp16.8, Hsp17, and Hsp17.3, contributing positively to thermotolerance [34]. TaHsfA1 can sense dynamically changing heat stress signals and regulate heat stress responses (HSRs) [35]. Likewise, HsfA6e was manifested as being involved in the regulation of Hsp17 in wheat [36]. In maize, some studies gave hints to the role of ZmHsp20s in heat tolerance. For instance, ZmHsp16.9 (corresponding to Zm00001d039935) showed heat tolerance when introduced into tobacco [37]. ZmHsp17.2 (corresponding to Zm00001d039936) showed transcriptional accumulation under high temperatures in maize [38, 39]. Additionally, the phosphorylation of sHsp17.4 (corresponding to Zm00001d028561) by ZmCDPK7 positively regulates heat stress tolerance in maize [40]. Our initial Hsp20 gene prediction, conducted by ACD domain scanning, revealed numerous ACD-domain-containing members in maize. However, whether these ACD-containing proteins belong to the Hsp20s family and are involved in heat stress tolerance remains unclear. In this study, we conducted a comprehensive analysis of ZmHsp20 genes and screened candidate thermotolerant ZmHsp20s through a stepwise analysis. Finally, one member of the ZmHsp20 genes, ZmHsp18, was confirmed to be responsible for thermotolerance in maize. This finding enables us to better understand the roles of ZmHsp20s in adapting to heat stress.

Materials and methods

Genome-wide identification of ZmHsp20 family genes in maize

Three approaches were adapted to identify ZmHsp20 family genes in maize step by step. Initially, the Hidden Markov Model (HMM) was used to obtain ACD-containing proteins in maize according to the Hsp20 family protein (PF00011) profile (P < 0.001) [18, 41–43]. Subsequently, protein sequences retrieved from known Hsp20 family members in Arabidopsis [44] and rice [19] were used as queries in a BLAST search against the maize genome database, filtering results with an e-value ≤ 1e⁻³. Additionally, a comprehensive search using “Hsp20” and “small heat shock protein” as keywords was performed in the maize genome database. After removing redundant sequences, unique sequences were verified for the presence of the conserved Hsp20 domain using CDD (https://www.ncbi.nlm.nih.gov/Structure/lexington/lexington.cgi) and Pfam (http://pfam.xfam.org/). Isoelectric point (pI) and molecular weight were calculated using the ExPASy tool (https://web.expasy.org/compute_pi/). These identified proteins were named according to their molecular weight [19, 45], and their chromosomal locations were determined by sequence alignment within the maize genome database (https://www.maizegdb.org/), corresponding to the Version 4 of B73 inbred line.

Multiple sequence alignment and phylogenetic analysis

In order to identify the ZmHsp20 family genes, the ACD-containing proteins from Zea mays, Oryza sativa [19], Arabidopsis thaliana [44] and Glycine max [46] were used to construct maximum likelihood phylogenetic tree. The amino acid sequences of all ZmHsp20s were aligned using the Muscle tool in MEGA7 software for phylogenetic tree construction. Neighbor joining analysis was performed with the pairwise deletion option and Poisson correction. Bootstrap analysis was conducted with 1000 replicates using the MEGA program. Amino acid sequences of maize, Arabidopsis [44], rice [19] and soybean [46] Hsp20 were sourced from the Ensemble Plants database (http://plants.ensembl.org/index.html). ZmHsp20s were classified into different subfamilies based on their proximity to these known Arabidopsis, rice and soybean Hsp20s in the phylogenetic tree. Gene duplication was verified using methods previously applied in potato Hsp20 and ZmHsf identification [41, 47]. For all candidate Hsp20s, the similarity threshold was set > 80%, and coverage > 70%, when their amino acid sequences were aligned with known Hsp20s.

Co-localization with heat stress QTL region

To date, considerable heat-resistant QTL loci related to various traits at different developmental stages have been identified and mapped in maize [48–50]. These QTLs are concentrated on several maize chromosomes. To find ZmHsp20s that localize within these heat-resistant QTL regions, we labeled the reported QTL regions on the same physical map (B73 V4) by calculating the positions of their flanking molecular markers.

Cis-acting regulatory element prediction in promoter regions of ZmHsp20

About 2.0 kilobases of promoter regions upstream of the initiation codon (ATG) of the 33 ZmHsp20s were downloaded from the MaizeGDB database. The prediction of conserved cis-acting regulatory elements within these promoters was conducted in the PlantCARE database [51].

Expression analysis

The expression patterns of ZmHsp20s were analyzed in fifteen different tissues of B73, including shoot, root, root hair, leaf, ovule, shoot apical meristem (SAM), 10-day ear, 10-day embryo, 10-day endosperm, seed, cob, tassel, anther, silk, and pollen (Supplemental Table 1) [52–58]. Based on the RNA-seq data sets retrieved from Li et al. [59], we also examined the expression patterns of these hrQTL-related ZmHsp20s under elevated temperature conditions in maize leaves. A threshold of a false discovery rate (FDR) corrected by a P-value < 0.05 and a fold change ≥ 2 was used to identify significant differential expression between the compared samples.

Co-expressed gene modules

Weighted gene co-expression network analysis (WGCNA) [60] in R software was applied to detect relative relationships among Hsp20 and Hsfs genes using RNA-seq datasets from various tissues (Supplemental Table 1) and RNA data sets under high-temperature treatment cited from Li et al. [59]). A weighted adjacency matrix was created using an automatic network construction and module detection method. The soft threshold power (β) was set at 9 to analyze the adjacency matrix. The modules were identified based on the parameters: minModuleSize = 3 and mergeCutHetght = 0.25. The topological overlap matrix similarity algorithm was then used to convert the adjacenty matrix into a topological overlap matrix at a threshold = 0.02. The visualization of the different modules was completed in Cytoscape, guided by the edge information [61].

Plant material and heat stress treatment

The maize (Zea mays L.) inbred line W22 (Wisconsin Agricultural Experiment Station) was grown in a chamber under normal conditions (32℃/22℃, day/night). At the V4 stage, half of the seedlings were kept in the normal condition, while the other half were transferred into a higher temperature environment (42℃, daytime) for heat stress treatment. From that point, leaves from each condition were collected every three hours and immediately snap‐frozen in liquid nitrogen, followed by storage at -80℃. These leaf samples were utilized for RNA extraction and to verify gene expression patterns under heat stress treatment.

For the thermotolerance analysis, a frame-shift EMS mutant of ZmHSP18 (EMS4-0be651) and its wild type B73 were purchased from the Maize EMS Database ( http://maizeems.qlnu.edu.cn/) [62]. The seeds of B73 and zmhsp18 were surface sterilized with 75% alcohol, followed by sterilization with a 10% NaClO solution for 30 min. The seeds were then soaked and germinated in a growth chamber at 28℃. Ten uniformly germinated seeds were planted in one pot, with a minimum of six pots prepared for each line. After growing for about 10 days, the seedlings at the V3 stage were exposed to heat stress treatment. Three pots were kept under normal temperature, while the others were simultaneously transferred to heat stress conditions (42℃/35℃, day/night) for 15 days, followed by a recovery period of 8 days at normal conditions. The survival rate in each pot was then calculated as the mean ± standard deviation. Each pot was adequately watered throughout the treatment period.

RNA extraction and RT-PCR analysis

Total RNA was extracted using the Trizol reagent (Vazyme, China). For RT-PCR analysis, 500 ng of total RNA were reverse-transcribed into cDNA by the PrimeScript™ Reverse Transcriptase Kit (TaKaRa, Japan). Quantitative RT-qPCR assays were then performed with a SYBR Green RT-PCR Kit 'Taq Universal SYBR Green Supermix' (Vazyme, China) according to the manufacturer’s instructions. Relative gene expression levels were calculated using the 2^−ΔΔCT method and normalized to the 0-h treatment under each condition. The maize GAPDH gene (Zm00001d049641) was employed as an internal control. Three biological replicates were included for gene expression analysis. The primers used in this study are detailed in Supplemental Table 2.

Results

Genome-wide scanning of ZmHsp20s

A total of 55 ACD-containing proteins were identified using BLAST, and these proteins were considered as candidate ZmHsp20 proteins. The sequences of these 55 proteins were retrieved from MaizeGDB (https://www.maizegdb.org/). In cases that genes had multiple transcripts, all protein sequences corresponding to these transcripts were downloaded and compared, with the most intact sequence being retained for each gene. Three proteins lacking the typical ACD domain were excluded after conserved domain analysis using the CDD and Pfam programs. Subsequently, a maximum likelihood phylogenetic tree was constructed using MEGA7, based on these 52 proteins and other ACD-containing proteins from rice and Arabidopsis. As a result, 33 non-redundant ZmHsp20s, including 7 previously reported ones [37], were identified by their proximity to known Hsp20s within the clusters of the phylogenetic tree (Supplemental Fig. 1). The predicted molecular weights of these 33 ZmHsp20s ranged from 14.0 kDa (ZmHsp14) to 27.4 kDa (ZmHsp27.4), with amino acid counts varying from 123 to 252. Their predicted isoelectric (PI) points ranged from 4.95 (ZmHsp21.9) to 8.66 (ZmHsp17.4B). Comprehensive information about gene name, sequence ID, protein size, molecular weight, and PI is listed in Table 1.

Table 1.

Basic information of HSP20 genes in maize

No	Gene ID	Chromosome	Location	Name	Introns	Size (aa)	PI	Mw (K Da)	ACD	Subfamily
1	Zm00001d028408	1	33,789,050–33790182	ZmHSP26.4	1	240.0	7.9	26.4	134–240	P
2	Zm00001d028555	1	38,997,819–38,998,295	ZmHSP17.8B	0	158.0	5.6	17.8	54–157	CI
3	Zm00001d028557	1	39,085,728–39086204	ZmHSP17.8A	0	158.0	5.6	17.8	54–157	CI
4	Zm00001d028561	1	39,214,417–39,214,896	ZmHSP17.9B	0	159.0	6.9	17.9	55–158	CI
5	Zm00001d031325	1	186,052,724–186053956	ZmHSP27.4	1	252.0	8.1	27.4	143–252	P
6	Zm00001d003554	2	47,866,673–47,867,299	ZmHSP22.8B	0	208.0	6.0	22.8	77–181	ER
7	Zm00001d004599	2	121,335,332–121,345,585	ZmHSP17.4B	2	156.0	8.7	17.4	57–140	CIII
8	Zm00001d007271	2	226,345,272–226,345,835	ZmHSP20.2	0	187.0	6.0	20.2	63–168	ER
9	Zm00001d039566	3	8,104,720–8105217	ZmHSP17.9A	0	165.0	6.0	17.9	60–163	CII
10	Zm00001d039933	3	19,846,081–19846545	ZmHSP17.1	0	154.0	6.8	17.1	50–153	CI
11	Zm00001d039935	3	19,868,724–19,869,483	ZmHSP14	1	123.0	8.0	14.0	51–122	CI
12	Zm00001d039936	3	19,966,205–19966663	ZmHSP17.2B	0	152.0	5.8	17.2	48–151	CI
13	Zm00001d039941	3	20,014,552–20015052	ZmHSP17.4A	0	166.0	5.0	17.4	46–148	CIX
14	Zm00001d039942	3	20,067,714–20068163	ZmHSP16.6	0	149.0	6.2	16.6	41–148	CVIII
15	Zm00001d052194	4	182,772,609–182773786	ZmHSP23.8	1	218.0	6.5	23.8	120–218	MI
16	Zm00001d053965	4	244,409,108–244409662	ZmHSP19.9	0	184.0	6.9	19.9	80–183	CX
17	Zm00001d015777	5	118,463,714–118,464,650	ZmHSP22.8A	1	211.0	6.9	22.8	112–211	MII
18	Zm00001d017813	5	207,525,809–207526282	ZmHSP16.7	0	157.0	7.8	16.7	26–145	CXI
19	Zm00001d018298	5	218,096,889–218097874	ZmHSP18.3	1	171.0	6.6	18.3	57–169	CIII
20	Zm00001d037633	6	132,470,913–132487481	ZmHSP15	1	138.0	5.7	15.0	57–136	CIII
21	Zm00001d038608	6	160,896,290–160898350	ZmHSP21.9	1	197.0	4.8	21.9	95–196	CVI
22	Zm00001d020390	7	110,937,105–110937825	ZmHSP17.5	1	161.0	6.2	17.5	61–144	CIII
23	Zm00001d021634	7	158,645,403–158646190	ZmHSP17.2A	1	155.0	6.6	17.2	44–139	CV
24	Zm00001d008577	8	13,397,318–13,397,785	ZmHSP17.3	0	155.0	5.8	17.3	51–154	CI
25	Zm00001d008841	8	22,307,488–22307982	ZmHSP17.8C	0	164.0	5.3	17.8	58–162	CII
26	Zm00001d010693	8	124,738,149–124,740,137	ZmHSP21.2	1	190.0	4.8	21.2	91–189	CVI
27	Zm00001d044728	9	593,529–593,969	ZmHSP15.8	0	146.0	8.1	15.8	49–153	Px
28	Zm00001d044874	9	6,096,248–6097444	ZmHSP22.3	1	208.0	5.5	22.3	110–208	MI
29	Zm00001d047542	9	134,707,072–134707554	ZmHSP17.4C	0	160.0	6.2	17.4	55–158	CII
30	Zm00001d047548	9	134,965,867–134,966,361	ZmHSP17A	0	154.0	7.8	17.0	33–143	CII
31	Zm00001d047553	9	135,126,430–135126924	ZmHSP17B	0	154.0	7.8	17.0	33–143	CII
32	Zm00001d047841	9	143,024,246–143024842	ZmHSP18	0	163.0	5.5	18.0	59–162	CI
33	Zm00001d025508	10	120,809,943–120810596	ZmHSP23.4	0	217.0	6.1	23.4	84–189	ER

Gene organization, location, and duplication

The structure analysis of the ZmHsp20 genes revealed that 60.61% (20/33) of these genes are intron-less. The other genes contain one intron each, except for ZmHsp17.4B, which contains two introns (Supplemental Fig. 2). Their schematic structures showed significant similarities with their counterparts in other species [19, 44].

The amino acid sequences of 33 ZmHsp20s, 23 OsHsp20s [19], 19 AtHsp20s [44], and 51 GmHsp20s [46] were used to reconstruct an unrooted phylogenetic tree (Fig. 1). The ZmHsp20s were divided into 14 different subfamilies based on the classifications of Hsp20s in rice, Arabidopsis and soybean. Twenty-four ZmHsp20s belong to the cytosol-localized subfamilies, with eight in CI, five in CII, four in CIII, two in CVI, and one each in CV, CVIII, CIX, CX, and CXI. The remaining nine ZmHsp20s were categorized into five different subfamilies: three in the ER, two in MI, two in P, one in MII, and one in Px.

Fig. 1.

Phylogenetic analysis of Hsp20 proteins. The unrooted neighbor-joining phylogenetic tree was conducted with Hsp20 proteins from Zea mays (Zm), Arabidopsis thaliana (At), Oryza sativa (Os) and Glycine max (Glyma). The different clades were marked with different background colors. These Hsp20s were grouped into six cytoplasmic/nuclear localized subfamilies (CI, C-II, C-III, C-IV, C-V, C-VI) and six organelle-localized subfamilies (ER: endoplasmic reticulum, CP: chloroplasts, M-I: mitochondria I, M-II: mitochondria II, Px: peroxisomal, and P: plastidial)

The 33 ZmHsp20s distribute across 10 chromosomes in the maize genome. The number of genes on each chromosome varied widely, with six ZmHsp20 genes on both chromosomes 3 and 9, and only one ZmHsp20 gene on chromosome 10 (Fig. 2). Consistent with the genome duplication event that occurred approximately 12 million years ago [63], seven pairs of duplicated ZmHsp20 genes were identified within chromosomes (Fig. 2 and Supplemental Table 3). One pair, ZmHsp17.2B and ZmHsp14, is on chromosome 3. Three homologs, ZmHsp17A, ZmHsp17B, and ZmHsp17.4C, are located on chromosome 9, with ZmHsp17A and ZmHsp17B sharing identical amino acid sequences. Likewise, on chromosome 1, ZmHsp17.8B and ZmHsp17.8A differ by only one amino acid. Apart from these homologs on the same chromosome, the other 16 pairs of duplicated genes are distributed across various chromosomes (Fig. 2). For instance, ZmHsp17.1 on chromosome 3 pairs with ZmHsp17.3 on chromosome 8, ZmHsp21.9 on chromosome 6 pairs with ZmHsp21.2 on chromosome 8, and ZmHsp18 on chromosome 9 is associated with ZmHsp17.9B, ZmHsp17.8B, and ZmHsp17.8A on chromosome 1.

Fig. 2.

Chromosomal location, gene duplication of ZmHsp20s and co-located heat-resistance QTL (hrQTL). The ring represents chromosome ideograms. The approximate distribution of each ZmHsp20 gene was marked with a short black line on the periphery of the circle. Possible duplicated gene pairs were linked together by marked colorful lines inside the circle. Genes in black were not involved in duplication. The regions marked with red color in the ring represent hrQTL intervals

Several quantitative trait loci (QTLs) related to heat stress tolerance have been identified in previous studies [48–50]. To explore the relationship between these QTLs and ZmHsp20 genes, the eight QTL intervals related to 13 different heat-related traits were mapped onto the maize genome too (Fig. 2 and Table 2). Compared with these heat-resistance QTLs, we found that nine ZmHsp20 members were co-localized within these intervals. This finding suggests a potential link between these ZmHsp20s and heat stress tolerance. Notably, among all the duplicated genes, at least seven members are associated with these QTL intervals. This implies that these ZmHsp20 genes may have been duplicated and subsequently developed new functions to help protect plants from heat stress.

Table 2.

ZmHSP20s co-located with QTL related to heat stress tolerance

Chromosome	Traits	Colocated ZmHsp20 genes	Reference
Chr2	50% of the plants of a plot showed male/female flowering	ZmHsp22.8B	Frey et al., 2016
	grain dry yield per hectare	ZmHsp17.4B	Frey et al., 2016
	adjusted dry yield	ZmHSP20.2	Inghelandt et al., 2019
Chr5	leaf greenness	ZmHsp22.8A	Inghelandt et al., 2019
	principal component analysis considering the calculated HSI of the traits	ZmHsp22.8A	Frey et al., 2016
Chr9	data for leaf scorching	ZmHsp17.4C, ZmHsp17A, ZmHsp17B, ZmHsp18	Frey et al., 2016
	leaf scorching of young leaves		Inghelandt et al., 2019
Chr10	The length of the fourth leaf	ZmHsp23.4	Inghelandt et al., 2019
	leaf firing, leaf blotching, leaf firing, leaf blotching, leaf firing	ZmHsp23.4	McNellie et al., 2018

Expression patterns of ZmHsp20 genes under normal growth conditions

To further investigate the potential functions of each ZmHsp20 gene, the expression levels of all 33 ZmHsp20 genes were analyzed, based on RNA-Seq data from various tissues. A heatmap was then constructed using normalized FPKM values by log₂^FPKM (Fig. 3 and Supplemental Fig. 3).

Fig. 3.

Analysis of the expression profile of ZmHsp20 genes under normal conditions. Cluster dendrograms and the color scale signal values were shown on the right. The genes located in the hrQTL regions were marked in red. The different tissues or organs were noted on the bottom of each lane

The ZmHsp20 genes showed significantly divergent expression patterns and were classified into four clusters (Fig. 3). The genes in cluster I, such as ZmHsp17.8A and ZmHsp17.8B, were found to be constitutively expressed in almost all tissues. In contrast, those in cluster IV demonstrated tissue-specific preferences, with ZmHsp19.9 and ZmHsp17.2B showing higher expression in the root and silk, respectively. Using FKPM > 100 as a criterion, some ZmHsp20 genes in clusters I and IV exhibited exceptionally high expression levels (FKPM > 100) in specific tissues, such as ZmHsp17.9A in the ear, ovule, and silk, and ZmHsp17.2B in silk. Conversely, ZmHsp20 genes in clusters II and III were relatively inactive in most tested tissues under normal conditions. Notably, most of the nine hrQTL-related ZmHsp20s belong to clusters II and III (Fig. 3), implying they are more likely to respond to heat stimuli than to be involved in basic developmental processes.

Analysis of cis-elements in the ZmHsp20 promoters

The analysis of stress-responsive cis-elements is crucial for deciphering the functions of each ZmHsp20 in stress response. We identified 12 types of stress-responsive cis-elements in the ZmHsp20 promoters, including ARE, DRE (DRE core), GC-motif, LTR, MBS, STRE, TC-rich repeats, W box, WRE3, WUN-motif, box S, and heat stress elements (HSE) (Fig. 4, Supplemental Fig. 4 and Supplemental Table 4).

Fig. 4.

Stress responsive cis-acting elements in 20 ZmHsp20 genes. A The different colors and numbers in the grids indicated the numbers of different elements in the proximal 2.0 kb promoter regions of 20 ZmHsp20 genes. The gene names were listed on the left and genes located in hrQTL region are highlighted in red. The different elements were noted on the bottom of each lane. ARE: involved in the anaerobic induction. DRE (DRE core): involved in dehydration responsive. GC-motif: involved in anoxic specific inducibility. LTR: involved in low-temperature responsiveness. MBS: involved in drought-inducibility. STRE: involved in stress responsiveness. TC-rich repeats: involved in defense and stress responsiveness. W box: involved in stress responsiveness. WRE3: involved in wound-responsive element. WUN-motif: involved in wound-responsive element. Box S: involved in stress response and signalling pathway. HSE: involved in heat stress-responsive element. B The number of cis-elements. Blue bars represent the number of total stress-responsive cis-elements. Orange rectangles represent for the number of heat stress cis-elements. The fraction on the right stands for the ratio of heat stress cis-elements to total stress responsive cis-elements

These elements were primarily associated with anaerobic stress, wound responsiveness, heat stress, and other abiotic stresses. Each ZmHsp20 promoter contains four to nine types of stress-responsive cis-elements, indicating a complex regulatory mechanism involved in the transcription of ZmHsp20 genes. Notably, only 20 ZmHsp20 genes have 1–5 HSEs in their promoters. Remarkably, seven out of the nine hrQTL-conjugated ZmHsp20s contain HSEs (Fig. 4), which constitutes a high proportion of 77.8%. Considering their co-localization with hrQTLs and the potential heat-responsive capabilities caused by HSEs, ZmHsp18, ZmHsp22.8A, ZmHsp17A, ZmHsp17B, ZmHsp17.4C, ZmHsp22.8B and ZmHsp23.4 were identified as promising candidates for heat-stress tolerance within the ZmHsp20 gene family.

Expression profiles of seven candidates in response to elevated temperatures in seedlings

The transcription patterns of seven candidate ZmHsp20 genes were compared using a set of RNA-seq data generated under continuously elevated high temperature treatments (31℃/21℃, 33℃/23℃, 35℃/25℃, and 37℃/27℃, day/night) in V4 and V5 leaves in W22 (Fig. 5A). In V4 leaves, ZmHsp23.4, ZmHsp18, and ZmHsp22.8B displayed up-regulated expressions as the temperature increased from 31℃ to 35℃, with a slight decrease when the temperature reached 37℃. In V5-stage seedlings, these three genes exhibited a similar yet enhanced expression. Conversely, ZmHsp17.4C and ZmHsp22.8A did not show notable induction in either V4 or V5 seedlings. As homologs, ZmHsp17A and ZmHsp17B showed consistent fluctuation patterns, with no significant correlation between temperature and gene expression. To validate the responsive expression pattern, we monitored their dynamic expression at 0 h, 3 h, 6 h, 9 h, and 12 h after heat stress treatment (42℃) using real-time RT-PCR analysis (Fig. 5B). Intriguingly, ZmHsp23.4/ZmHsp22.8B and ZmHsp18 showed a rapid increase in expression, peaking within the first 3 h of heat treatment, indicating a strong response to heat stress. In contrast, ZmHsp17A, ZmHsp17B, and ZmHsp17.4C exhibited about a ten-fold increase in expression after 3 h of treatment.

Fig. 5.

Expression profiles of seven ZmHsp20 genes in response to elevated temperature in W22. A Heatmap of expression of seven ZmHsp20s in V4 or V5 leaves in W22 with increasing temperature treatment (31℃ /21℃, 33℃ /23℃, 35℃ /25℃ and 37℃ /27℃, Day/night). The color scale on the right indicates the signal intensity. Relative expression levels of ZmHsp17.4C and ZmHsp22.8A were nearly not detected. B RT-qPCR analysis of six genes in W22 plants at room temperature (RT, 28℃) and heat stress (HT, 42℃). Relative levels of seedlings treated within different hours were presented. Due to high sequence identity between homologs, the expression of ZmHsp17A / ZmHsp17B / ZmHsp17.4C was detected using a same primer, as was ZmHsp23.4 / ZmHsp22.8B. Data were mean ± SD. Asterisks indicated statistically significant differences between the RT and HT treatments at each timepoint according to the Student’s t test. *P ≤ 0.05

Co-expression interaction networks between ZmHsp20s and ZmHsfs

The heat-responsive ZmHsp20s genes can be directly or indirectly triggered by upstream Hsfs, which are key factors in the heat resistance pathway [64, 65]. To date, 25 ZmHsf genes have been identified [22, 47]. The association between ZmHsp20s with ZmHsf genes is largely unknown. Therefore, we used WGCNA to explore their co-expression interaction network. The normalized FPKM expression values of ZmHsp20s and ZmHsf genes in different tissues were used as input to construct the co-expression network. A total of five modules containing different ZmHsp20s and ZmHsf genes were identified (Supplemental Fig. 5 and Fig. 6). There are 3 candidate genes in the blue module. Within this module, ZmHsp23.4, ZmHsp22.8B, and ZmHsp18 all exhibited strong connections with ZmHsf11, ZmHsf08, and ZmHsf20, suggesting that they may be regulated by ZmHsf genes and potentially play roles in heat tolerance in maize (Fig. 6).

Fig. 6.

Coexpression network analysis of the blue module. A Heatmaps of genes in this module based on their expression in different tissues. The log₂(RPKM) expression values of genes were applied to construct the heatmap. B Correlation networks of these genes involved in this module. Orange and blue color spheres represented ZmHsp20 and ZmHsf genes, respectively. The thickness of the line indicates the correlation degree between the two genes at either end of the line. The thicker the line, the higher correlation between them

ZmHsp18 is critical for maize tolerance to heat stress

ZmHSP18 belongs to the CI subfamily, while ZmHSP23.4 belongs to the ER subfamily. They are hypothesized to function as chaperons in the cytosol or respond to ER stress under heat stress conditions, respectively. EMS-induced mutation is an effective means to confirm their roles in heat stress resistance. Due to the unavailability of a double mutant of the homologs ZmHsp23.4 and ZmHsp22.8B in MEMD, we focused on the zmhsp18 mutant for heat stress treatment (Fig. 7). Initially, a single nucleotide mutation (G > A) in the exon leading to a premature stop codon (TGA) was confirmed by Sanger sequencing (Fig. 7A - 7B). Under normal growth conditions, no significant difference was observed between WT and zmhsp18 seedlings (Fig. 7C, Supplemental Fig. 6). However, both WT and zmhsp18 showed suppressed growth after a 15-day heat stress (42℃) treatment. As depicted in Fig. 7D, the height of the WT plants decreased by nearly 50% compared to the normal condition. Likewise, the height of zmhsp18 plants decreased even more severely. Moreover, zmhsp18 seedlings exhibited a significantly lower survival rate than WT after 8 days of recovery from heat stress (Fig. 7D – 7E). Three independent biological replicates were conducted with comparable results both before and after the 8-day recovery (Supplemental Fig. 6B - 6D). These findings demonstrated that the loss-of-function of ZmHsp18 undermined heat stress tolerance in maize, suggesting its vital role in maize thermotolerance.

Fig. 7.

The zmhsp18 seedlings exhibited lower survival rate after heat stress treatment. A Sequencing chromatograms of Zmhsp18 and zmhsp18 genomic fragment. Red arrow indicated the mutation sites. B The schematic structure of Zmhsp18 gene and the mutation site in the zmhsp18 mutant. C Ten-day seedlings treated by room temperature for 15 days. D Ten-day seedlings treated by high temperature (42℃, 16 h, day time/35℃, 8 h, night time) for 15 days and then recovered at normal conditions for 8 days. E Survival rate of the zmhsp18 and WT after heat stress treatment and recovery. Asterisks indicate significant differences according to the two-tailed student’s t test. **P ≤ 0.01. Scale bars, 2 cm

Discussion

As the average global temperature rises, maize is increasingly exposed to high-temperature environments. Numerous studies have revealed that Hsp20s, functioning as molecular chaperones, prevent the irreversible aggregation of denaturing proteins, thereby enhancing the stress tolerance of many plants [14, 15]. Forty-four Hsp20s have been identified recently [66], but not all ACD proteins can be classified as Hsp20 [13, 19, 44]. In the current study, we set more stringent parameters and identified 33 ZmHsp20s. These ZmHsp20s were grouped into 14 subfamilies, with 24 members predicted to be located in the cytosol, aligning with their potential chaperone functions in that location [67].

The phylogenetic analysis revealed homologs in Arabidopsis, rice and soybean within each ZmHsp20 subfamily, indicating a conserved evolution of Hsp20s in gramineous species. However, many ZmHsp20 subfamilies were enlarged in maize. For example, the CIII subfamily in maize has four ZmHsp20s versus only one in Arabidopsis and rice, and two in soybean. This pattern of expansion is also observed in other crops, as 13 Hsp20s in barley [68], 23 in rice [19], 94 in cotton [69], and 109 in wheat [42]. The diversification may reflect plant adaptations to environmental stresses [14, 15] and could be attributed to genome duplication or chromosomal polyploidy. Maize genomes have undergone duplication, and the ZmHsp20 gene family may have expanded and diversified concurrently, contributing to gene divergence [70]. In our study, 20 ZmHsp20 genes (61%) were found to be involved in gene duplication, indicating that duplications have significantly contributed to the amplification of the ZmHsp20 gene family.

Although the Hsp20 subfamilies vary among species, most Hsp20 genes tend to cluster into the CI, CII, and CIII subfamilies. The functions of CI and CII subfamily genes are not consistent across different species. For example, CI Hsp20s have been reported to provide thermoprotection [71, 72], while CII Hsp20s do not [73]. Consistently, our research discovered that ZmHsp18 in the CI subfamily is highly expressed under heat stress and confers thermotolerance in maize.

As previously reported, Hsfs can regulate the expression of a group of Hsp20s by recognizing heat stress elements (HSEs, 5’-GAANNTTC-3’) located in their promoters upstream of the TATA box [14, 74].For instance, one HSE was detected in ZmHsp17.2B promoter and it has been shown to confer heat tolerance in transgenic tobacco [37]. It is encouraging to find that HSEs were present in the promoters of 20 ZmHsp20 genes, indicating their potential heat-stress response under high temperature conditions. Both ZmHsp23.4 and ZmHsp22.8B contain HSEs in their promoters, showing dramatical upregulation under heat stress treatment and co-expressing with four Hsfs in a large regulation module. These findings give strong hints to their potential function in heat stress tolerance. However, ZmHsp23.4 and ZmHsp22.8B show high sequence identity and may function redundantly in maize thermotolerance. Double mutants or transgenic lines would provide more definitive evidence of their role in maize thermotolerance. The positive role of ZmHsp18 in maize thermotolerance was finally confirmed by EMS mutants. Given that ZmHsp18 also colocalizes with previously reported heat tolerance QTL on chromosome 9, it is plausible to consider that the QTL effect may be attributed to natural variation in ZmHsp18. ZmHSP18 may act as a chaperone in the cytosol, protecting plants against heat stress by preventing protein aggregation. Further experiments are required to elucidate the underlying mechanisms of ZmHsp18 in maize heat tolerance in future studies. Our discovery establishes the foundation for future investigations into how ZmHsp20s contribute to maize's thermotolerance.

Conclusion

In this study, 33 ZmHsp20s were identified and divided into 14 subfamilies. These genes were distributed across 10 chromosomes, and 23 gene pairs were derived because of genome duplication. There are nine hrQTL-related ZmHsp20 genes located within the quantitative trait loci of heat resistance. The ZmHsp20 genes showed significantly divergent expression patterns and were classified into four clusters under normal growth conditions. Seven hrQTL-related genes, ZmHsp18, ZmHsp22.8A, ZmHsp17A, ZmHsp17B, ZmHsp17.4C, ZmHsp22.8B and ZmHsp23.4, contain 1–5 HSE elements in their promoters. They were relatively silent in nearly all tested tissues under normal conditions, but most of them were induced under heat treatment. RNA-seq analysis and real-time RT-PCR verified that ZmHsp23.4/ZmHsp22.8B and ZmHsp18 were highly induced by heat stress and showed strong associations with ZmHsf11, ZmHsf20, and ZmHsf08. Finally, ZmHsp18 was proven to positively regulate thermotolerance in maize.

Authors’ contributions

M.X., Y.W.Y.and S.H.C. performed most of data analysis. Y.W.Y., L.Y.Z. and J.M.C. participated in the experiments. M.X., M.L.X. and S.H.C. wrote the main manuscript text. All authors reviewed the manuscript.

Funding

This work was supported by grants from the Jiangsu Province Government (JBGS [2021]002), the Hainan Yazhou Bay Seed Lab (B23YQ1510), the National Key Research and Development Program of China (2022YFD1201804) and the China Postdoctoral Science Foundation (2022M712701). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Data availability

All data supporting the findings of this study are available within the paper and its Supplementary Information.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.