The peanut genome data and the Arabidopsis genome data were downloaded from the Peanut Base (https://www.peanutbase.org/) and TAIR (https://www.arabidopsis.org/), respectively. Arabidopsis Hsf full-length protein sequences were used as queries to perform BLASTP program against the peanut genome database with an E-value of 0.0001. The resulting sequences were then subjected to Pfam (El-Gebali et al., 2019) and SMART (Letunic and Bork, 2018) analyses to detect the presence of the Hsf-type DBD and OD domain. The ProtParam (Wilkins et al., 1999) was used to calculate Mw (molecular weight) and theoretical pI (isoelectric point) of the putative Hsf proteins from peanut.

MAFFT (Katoh et al., 2019) under the default parameters was applied in protein sequences alignment analyses of newly identified peanut Hsf members and previously reported Arabidopsis Hsf members. A neighbor-joining (NJ) tree was constructed by MEGA 11 with the following parameters: Poisson correction, 1000 bootstrap values, and pairwise deletion (Tamura et al., 2021).

The structure of peanut Hsf genes was analyzed using the gene structure display server (Hu et al., 2015) by comparing the coding sequence (CDS) and genomic sequence obtained from the peanut genome database. The 2000-bp sequences upstream of peanut Hsf genes were obtained from peanut genome database as the promoter (Wang et al., 2021). And then the cis-regulatory elements of these promoters were identified by PlantCARE (Lescot et al., 2002).

The typical functional structure domains of peanut Hsf protein were analyzed by Pfam and SMART. The nuclear localization signal (NLS) prediction was done by using the online tool (Nguyen Ba et al., 2009). MEME tools (Bailey et al., 2009) was applied to identify conserved motifs of the AhHsfs full-length proteins.

The chromosomal location image of peanut Hsf genes was generated by TBtools (Chen et al., 2020), according to the data obtained from the peanut genome database. The tandem and segmental duplications were identified by using MCScanX program (Wang et al., 2012) and the results were visualized by Circos (Krzywinski et al., 2009). The syntenic analysis of the orthologous genes obtained from peanut and other three plant species were investigated by TBtools (Chen et al., 2020). Subsequently, the synonymous substitution (Ks) and non-synonymous substitution (Ka) rates were calculated using DnaSP 5.0 software (Librado and Rozas, 2009).

The peanut cultivar HY9312 plants were used to analyze the expression of AhHsf genes in this study. The peanut seeds were germinated on MS medium in a light incubator at 25 °C for two weeks. The seedlings were then transferred to 20% PEG 6000 and 200 mM NaCl for drought and salt stresses, respectively. After treatments of 0, 6, 24, and 48 h, the leaves were harvested for RNA extraction and qRT-PCR analysis. The samples were frozen in liquid nitrogen immediately and then transferred to -80 °C for RNA extraction. Three biological replicates were used for each sample.

Total RNA was extracted by Trizol Reagent according to the manufacturer’s instructions. RNA quality and concentration were tested by using a Nanodrop 2000 spectrophotometer (Thermo Fisher Scientific, Wilmington, DE, United States). The synthesis of fist-strand cDNA by using the PrimeScript ™ RT reagent Kit (TaKaRa, Shiga, Japan) with Oligo (dT) primers. The qRT-PCR was performed on an ABI7500 Real-Time PCR System (Applied Biosystems, Foster City, CA, United States) with 2 µL template cDNA. The peanut Actin11 gene was adopted as the internal control (Li et al., 2021). The relative primer sequences were listed in Supplementary Table 1 . All reactions were performed with three biological replications and the resulting data were analyzed by using the 2 ^−ΔΔCT method (Livak and Schmittgen, 2001).

The coding regions (without stop codon) of AhHsf20 were amplified from the cDNA of peanut leaves and inserted into the pCHF3-cGFP vector by Infusion (Invitrogen), generating the AhHsf20-GFP fusion fragment under control of the CaMV-35S promoter. The 35S::AhHsf20-GFP plasmid was introduced into Agrobacterium competent cell GV3101 for transient expression in the leaf of Nicotiana benthamiana. After growth of three days in light chamber, these leaves were subjected to 4,6-diamidino-2-phenylindole (DAPI) staining to confirm the position of the nucleus. The fluorescence signals were monitored by using the confocal microscope (TCS-SP8 Leica, Wetzlar, Germany), as previously reported (Li et al., 2019).

The coding sequence of AhHsf20 was inserted into the SacI-digested pCHF3 vector by Infusion (Clontech) to produce 35S::AhHsf20, which was then transformed into Agrobacterium tumefaciens GV3101 and used to transform Arabidopsis Col-0 plants by the floral dip method (Zhang et al., 2006). T3 homozygous seedlings were used for further analyses. The sterilized wild-type and transgenic Arabidopsis seeds were evenly sown on 1/2 MS media. The seeds were stratified at 4 °C in darkness for two days, and then transferred to culture room (23 °C, continuous light). Wild-type and transgenic Arabidopsis seedlings growing normally for seven days were transferred to 1/2 MS media containing 100 and 150 mM NaCl. The primary root lengths were measured after seven days of upright growth in 1/2 MS media. Three biological replicates.

The GraphPad Prism 8 (t test) was used to analyze significant differences (P < 0.05, P < 0.01, P < 0.001). All data were obtained from three replicates.

To identify Hsf genes in peanut, the BLASTP search was performed using the Pfam Hsf-type DBD domain (PF00447) as query. After deleting redundant genes, we identified a total of 46 Hsf family proteins in peanut, designating the newly identified peanut Hsf genes AhHsf1 to AhHsf46 according to their position on chromosome. The Chr5, Chr13, and Chr15 contained the largest number of AhHsf genes, with six genes. There was followed by Chr3, which had five AhHsf genes. The Chr6 and Chr16 contained four AhHsf genes, respectively. The Chr8, Chr9, Chr17, and Chr19 contained two AhHsf genes, respectively. Additionally, Chr1, Chr2, Chr7, Chr10, Chr11, Chr12, and Chr20 contained only one AhHsf gene ( Supplementary Table 2 ). The lengths of the predicted AhHsf proteins ranged from 207 amino acids (AhHsf29) to 599 amino acids (AhHsf44). The molecular weights (Mw) of Hsf proteins ranged from 23.6 (AhHsf29) to 66.7 (AhHsf44) kDa, and the theoretical isoelectric points (pI) ranged from 4.82 (AhHsf8 and AhHsf32) to 8.81 (AhHsf4) ( Supplementary Table 2 ).

In order to further study the evolutionary relationship among the peanut Hsf proteins, we constructed a neighbor-joining phylogenetic tree from alignments of 46 peanut Hsf proteins and 21 Arabidopsis Hsf proteins ( Figure 1 ). The results showed that all the Hsf proteins from peanut and Arabidopsis were divided into three groups: A, B, and C. Furthermore, 28 AhHsfs in group A were divided into nine subgroups comprising A1-A9; 16 AhHsfs in group B were divided into four subgroups comprising B1, B2, B3, and B4; and 2 AhHsfs in group C were divided into C1 subgroup. The A1 and B4 subgroups contained the largest number of AhHsf members with six members. There were followed by subgroup A2, A4, A5, B2, and B3 subgroups, all of which had four members of the AhHsf family. Then subgroup A3, A6, A7, A8, A9, B1, and C1 subgroups all contained two members of the AhHsf family ( Figure 1 ).

We identified five typical conserved domains in peanut Hsf proteins, including DBD, OD, NLS, AHA, and NES domains ( Supplementary Table 3 ). All of the AhHsf proteins possessed a highly conserved section, DBD domain, which contained three α-helices and four β-sheets (α1-β1-β2-α2-α3-β3-β4) in N-terminal region ( Supplementary Figure 1A ). Adjacent to the DBD domain is the HR-A/B domain, which is characterized by a coiled-coil structure in C-terminal region of the AhHsf proteins. We found that group A and group C Hsfs have different amounts of amino acid residues are respectively inserted between the HR-A and HR-B, whereas it is not found in the group B Hsfs ( Supplementary Figure 1B ). We also found that most of AhHsf proteins have NLS and NES domain, which are essential for maintaining the changes dynamically of Hsf proteins between nucleus and cytoplasm (Scharf et al., 1998; Heerklotz et al., 2001). 32 AhHsf proteins have NLS domain and NES was detected in 16 AhHsf members. Notably, the AHA motifs exist only in class A Hsfs. Among of them, seven AhHsf members (AhHsf8, AhHsf9, AhHsf23, AhHsf25, AhHsf32, AhHsf33, and AhHsf46) contained two AHA motif ( Supplementary Table 3 ).

We analyzed the gene structure of all the AhHsf genes by GSDS, including introns and exons. The results showed that all AhHsf genes contained introns, but the number and length of introns were different ( Figure 2B ). In group A, there were four introns in AhHsf3 gene, two introns in four AhHsf genes (AhHsf6, AhHsf16, AhHsf40, and AhHsf44), only one intron in others. In group B and group C, all AhHsf genes contained one intron. AhHsf genes had similar gene structures in the same subgroup.

We used MEME to analyze the motif composition of AhHsf proteins ( Figure 2C ). A total of 10 motifs (named motif 1-10), ranging from 6 to 50 amino acids, were predicted based on the feature of AhHsf protein sequences, and all sequences were listed in Supplementary Figure 2 . Motif 1, motif 2, and motif 4 correspond to the DBD. Motif 3 and motif 5 correspond to the OD. Motif 10 correspond to the NLS. Motif 8 correspond to the NES.

To further understand about the phylogeny of the peanut Hsf family members, syntenic analysis was performed between the Hsf genes of peanut and the Hsf genes of the other four plant species, including Arabidopsis, soybean (Glycine max), tomato (Solanum lycopersicum), and rice (Oryza sativa) ( Figure 3 ). The results revealed that 39, 38, 27, and eight AhHsf genes were synchronized with Hsf genes in soybean, tomato, Arabidopsis, and rice, respectively. The number of collinear pairs between peanut and other four plant species (soybean, tomato, Arabidopsis, and rice), were 121, 52, 36, and 14, respectively, suggesting that the genetic relationship between peanut Hsf genes and soybean Hsf genes was close. Meanwhile, we found that the AhHsf45 of peanut was associated with six, three, two, and two Hsf genes in soybean, Arabidopsis, tomato, and rice, respectively, suggesting that AhHsf45 may play an important role during the evolution in peanut. Notably, a total of six Hsf genes formed collinear pairs with Hsf genes from all of other four plants, suggesting that these Hsf genes may have existed before the divergence of these plant species. The detailed information of syntenic gene pairs is provided in Supplementary Table 4 .

In this study, 46 AhHsf genes were mapped on 17 peanut chromosomes, and each of which contains different number of the AhHsf genes ( Figure 4A ). The most peanut Hsf genes (six) were found on Ah5, Ah13, and Ah15, while Ah1, Ah2, Ah7, Ah10, Ah11, Ah12, and Ah20 only have one peanut Hsf gene. It has been reported that when a chromosome region within 200 kb possesses two or more genes of the same family, these genes are defined to be a gene cluster and genes sharing an identity of more than 70% in a cluster are considered to be tandem duplication genes (Holub, 2001). Homology analysis showed that there were two tandem duplication pairs (AhHsf8/AhHsf9 and AhHsf32/AhHsf33) in peanut Hsf gene family ( Figure 4A ).

We used the MCScanX to analyze segmental duplication or whole genome duplication of the AhHsf genes. In a total, 35 segmental duplication pairs were identified in 42 AhHsf genes ( Figure 4B ). 23 pairs of segmental duplication occurred in group A Hsf genes. 11 pairs of segmental duplication occurred in group B Hsf genes. However, only one pair of segmental duplication (AhHsf12/AhHsf36) occurred in group C Hsf genes. These results suggested that the generation of some Hsf genes may be due to gene duplication events. All of the tandem and segmental duplication genes were listed in Supplementary Table 5 .

Ka/Ks refers to the ratio between non-synonymous and synonymous substitutions, which can estimate whether the selective pressure acts on the protein-coding gene. There are three selection roles in evolutionary analysis, including the positive selection, neutral selection, and purifying selection. All Ka/Ks ratios of the 35 segmental and two tandem duplication pairs were less than one, suggesting that these Hsf genes may have undergone purifying selective pressure in process of evolution ( Supplementary Table 5 ).

In order to study the potential function of AhHsf genes in stress responses, plantCARE was used to predict the cis-elements in promoter regions ( Figure 5 ). Four hormone-responsive elements were identified in most AhHsf gene promoters, including ABRE, GARE, TCA-element, and CGTCA-motif, which regulate the plant responses to ABA, gibberellin, salicylic acid (SA), and methyl jasmonate (MeJA), respectively. The result suggested that most AhHsf genes may be involved in ABA-mediated and MeJA-mediated stress responses. Notably, the AuxRR-core element of regulating the plant responses to auxin was only detected in promoter regions of certain AhHsf genes, including AhHsf2, AhHsf3, AhHsf5, AhHsf12, and AhHsf25. Six stress-responsive elements including anaerobic induction element (ARE), wound-responsive element (WUN-motif), stress-responsive element (TC-rich repeats), WRKY binding site (W-box), MYB binding site (MBS), and low-temperature-responsive element (LTR) were also detected in the promoter regions. These AhHsf genes may be involved in responses to different stress conditions. Furthermore, a cis-element related to flavonoid biosynthetic (MBSI) was found in promoter regions of five AhHsf genes.

To determine the expression patterns of the AhHsf genes in different tissues, the corresponding transcriptome data with 22 tissues were accessed in the Peanut Base. It was found that the expressions of the AhHsf genes were significant differences in different tissues. As shown in Figure 6 , AhHsf5 from group A was highly expressed in the all-tested tissues, whereas the expression of its homologue AhHsf30 was barely detectable. AhHsf1 and AhHsf24 were highly expressed in seed Pat. The expression levels of AhHsf23 and AhHsf46 in leaves were higher than other tissues. In group B, AhHsf14 and AhHsf39 were highly expressed in root and nodule. However, the expression of these two genes were lower in other tissues. Additionally, in group C, the expression levels of AhHsf12 and AhHsf36 were similar in the all-tested tissues. The various expression patterns suggested that AhHsf genes have different functions in peanut growth and development.

To reveal peanut Hsf genes expression patterns under abiotic stresses, we utilized transcriptome data after drought and salt stresses from NCBI (Zhao et al., 2018; Zhang et al., 2020). As shown in Figure 7 , most peanut Hsf genes exhibited several different expression patterns. Some genes, such as AhHsf11, AhHsf24, and AhHsf35, were upregulated under drought and salt stresses. Several peanut Hsf genes showed conflicting expression patterns under the two stress conditions. For instance, AhHsf1 was insensitive to salt stress, whereas it was significantly induced by drought stress. Additionally, some genes did not show significant expression changes after drought and salt stress treatments, such as AhHsf46. The various expression patterns reflected the different roles of peanut Hsf genes in abiotic stress-response pathways.

qRT-PCR was used to verify the expression pattern of peanut Hsf genes in response to abiotic stresses, including drought and salt. According to the results, we found that several AhHsf genes were induced by drought and salt stresses ( Figure 8 ). For example, the expression levels of AhHsf1, AhHsf20, and AhHsf24 were continually up-regulated under drought stress, AhHsf4 and AhHsf24 were continually up-regulated under salt stress. Moreover, expression levels of AhHsf1/4/20/24/29/35/43 were significantly increased (>10-fold) in drought stress condition. AhHsf4, AhHsf20, and AhHsf29 were significantly increased (>7-fold) in salt stress condition. Notably, seven AhHsf genes, including AhHsf1, AhHsf4, AhHsf20, AhHsf24, AhHsf29, AhHsf35, and AhHsf43, were induced under both drought and salt stresses. These results suggested that peanut Hsf genes might be involved in response to drought and salt stresses.

In order to further explore the potential function of the AhHsf genes. The stress-induced AhHsf20 gene was selected for subcellular localization analysis ( Figure 9 ). The coding region of AhHsf20 without the stop codon was fused to the GFP reporter gene under control of the CaMV35S promoter. The Agrobacterium cultures with the recombinant construct were used to inject the tobacco leaf epidermal cells, using 35S::GFP vector as the control. Subsequently, the signal of GFP protein was observed by confocal microscopy. It can be seen from Figure 9 that AhHsf20 fusion protein was specifically located in the nucleus, whereas the control group was uniformly distributed throughout the whole cell. These results showed that AhHsf20 protein was located in nucleus.

To explore the underlying mechanisms of action of AhHsf members, interaction networks of AhHsfs in peanut were predict using STRING online tool (https://cn.string-db.org/). We found that protein-protein interaction occurred between 18 peanut Hsf members ( Supplementary Figure 3 ). There were four hub nodes in the AhHsf protein interaction network, including AhHsf13, AhHsf10, AhHsf2, and AhHsf18. Among them, AhHsf13 was found to interact with ten other Hsf members, followed by AhHsf2 (4), AhHsf10 (3), AhHsf18 (3). The results showed that these interacting proteins might co-exert specific biological functions in peanut.

To further investigate the function of AhHsf20 in response to abiotic stresses, we generated transgenic lines overexpressing AhHsf20 under the control of the 35S promoter in Arabidopsis Col-0 plants (AhHsf20-OE). Two overexpression lines AhHsf20-OE1 and AhHsf20-OE3 were selected for further studies ( Figure 10B ). Compared with wild-type, AhHsf20-OE1 and AhHsf20-OE3 plants displayed a longer root phenotype under 100 and 150 mM NaCl ( Figures 10A, C ). The results revealed that the overexpression of AhHsf20 enhanced the salt tolerance in transgenic Arabidopsis.