Genome-Wide Identification of the Heat Shock Transcription Factor Gene Family in Rosemary (Salvia rosmarinus)

Abstract

Rosemary (Salvia rosmarinus) is a world-famous plant frequently subjected to various environmental stresses. Heat Shock Transcription Factor (HSF) has been shown to be essential for plant growth and for resistance to environmental stresses. This study utilized bioinformatics techniques to identify the SrHSF gene family in the rosemary genome. A total of 49 SrHSFs were detected, unevenly distributed across 12 chromosomes. The SrHSF genes were classifiable into 3 subfamilies and contained in 14 subgroups. They were relatively conserved during the evolutionary process based on gene structure and conserved motif analysis. There were 22 kinds of cis-acting elements in the promoter regions of SrHSF genes, mostly related to hormones, stress, growth, and development. The interactions among 16 highly conserved SrHSF proteins were also identified. Gene collinearity analysis showed that 51 segmental duplication events were undergone among 41 SrHSF genes. Ka/Ks ratios were all less than 1, suggesting a purifying selection of SrHSF homologous genes. The expression pattern of SrHSF genes revealed that the majority of them are highly expressed in the secondary stems. After 0.1% MeJA treatment, SrHSF36 and SrHSF11 showed a significant upregulation in leaves. This research provides valuable insights into the functions and regulatory mechanisms of the SrHSF gene family.

1. Introduction

Rosemary (Salvia rosmarinus) is a small aromatic shrub in the Labiatae family. Its leaves are rich in potent antioxidant secondary metabolites, such as rosmarinic acid, carnosic acid, and carnosol, which have antibacterial, anti-inflammatory, antioxidant, antitumor, and immunomodulatory properties and are used in the fields of pharmaceuticals, food, and cosmetics [1,2,3]. With the intensification of the greenhouse effect and the frequent occurrence of extreme weather, rosemary is also frequently affected by various adverse environments during its growth, particularly heat, cold, drought, and sunburn stresses. Currently, the research on rosemary is mainly focused on the extraction and utilization of bioactive substances, and there are few reports on the molecular mechanisms of rosemary resistance to various stresses. Revealing its response mechanisms to stresses is significant for its cultivation management and the resistant breeding in rosemary.

Plants perpetually adapt to diverse environmental factors throughout their lives [4]. In response to the complex changeable environmental conditions, they develop a series of signaling systems to sense environmental variations, regulating the expression of transcription factors and functional proteins. Among these, the HSF–heat shock protein (HSP) pathway is the earliest discovered primary mechanism in plant response to heat stress. HSF is a central component in the regulation of plant resistance to external stresses. Under normal conditions, HSF proteins usually exist in a monomeric form, and when cells are subjected to heat stress, they are activated to form a trimeric structure [5], which can specifically recognize and bind to the specific cis-element heat shock elements (HSE), thereby regulating the expression of HSP and other stress-response genes. HSE is the conserved motif in the promoter regions of HSP genes, and HSP is a molecular chaperone that helps to maintain protein stability and mitigate the damage to plant cells under stresses [6,7]. Moreover, under heat stress, the binding of HSF and HSP is broken, and then HSP repairs misfolded proteins, thereby maintaining the structure and function of cells. Rosemary is often affected by various stresses in its native region. Revealing the mechanisms of the HSF-HSP regulatory pathway in rosemary could provide new insights into its environmental adaptability and molecular breeding for stress resistance.

Typical HSF proteins contain five conserved oligomerization domains [8,9]: the N-terminal DNA binding domain (DBD), the oligomerization domain (OD), the nuclear localization signal (NLS) [10], the nuclear export signal (NES), and the activator peptide motif (AHA) [11,12]. The DBD is highly conserved and binds to the HSE of stress-response genes [13]. The OD has a heptapeptide structure of two lipophilic amino acid residues (HR-A/B region), which can be linked to the DBD by a flexible linker, ranging from 15 to 80 amino acids. The formation of helical structures by HR-A/B promotes HSF trimerization [14]. HSF genes of plants were categorized into three subfamilies (A, B, and C) according to the sequence structure of amino acids from DBD to HR-A/B and the sequence structure between HR-A and HR-B [15]. The AHA motifs belong to Subfamily A, which can recruit transcriptional co-activators like p300 and CREB binding protein (CBP) [16,17]. Conversely, subfamilies B and C, without AHA motifs for transcriptional activation, are regarded as potentially serving as transcriptional co-activators or repressors to regulate heat shock response genes [18].

The results of previous studies show the involvement of HSF in plants’ response to various abiotic stresses, including heat, drought, salt, and oxidative stresses [19,20]. In the Mediterranean plant Ammopiptanthus mongolicus, the AmHSF gene shows significantly differential expression at different time points under heat stress [21]. In Apium graveolens, AgHSFa6-1 can enhance plant heat tolerance by upregulating heat-sensitive genes [22], and GmHSFB2b in soybeans promotes the accumulation of flavonoids to improve plant salt tolerance [23]. OsHSFA2b boosts rice resistance to drought and salinity [24], and PsHSFA1d in peas enhances antioxidant enzyme activity in leaves to reduce H₂O₂ levels under heat stress [25]. These results suggest that HSF regulates multiple pathways in plants’ adversity adaptation. Additionally, plant hormones are crucial in plant abiotic stress response by modulating transcription factor expression [26]. AtHSFA6b of Arabidopsis thaliana (A. thaliana) also participates in the ABA signaling pathway [27]. HSF genes were significantly upregulated by exogenous hormones ABA and GA3 while down-regulated by IAA [28]. The exogenous ABA, GA, BAP, and SA regulated most of the SlHSF genes in tomatoes, with ABA and MeJA specifically affecting the expression of SlHSFB3a in roots [29]. These studies indicate that HSF is also regulated by exogenous hormones. However, whether there is a crosstalk between the hormonal regulation and stress response of HSF needs to be further demonstrated.

Recently, the HSF gene families of numerous plants have been detected and functionally identified, including numerous aromatic and medicinal plants [30,31,32,33]. However, the identification of rosemary has not yet been performed. The molecular properties of SrHSF members are still unclear, and their expression patterns and cis-regulating elements are unknown. Based on the assembly of the rosemary genome [34,35], this study characterized and analyzed the SrHSF gene in rosemary at the genomic level, focusing on its features and expression patterns. The results could provide valuable insights into the functional identification of the HSF gene family in response to stresses and provide candidate genes for stress-resistant molecular breeding in rosemary.

2. Materials and Methods

2.1. Data Resources

The whole-genome sequence and annotation files of rosemary, completed by Yang et al. [35], were used to create a local database. A. thaliana genome and protein sequences were obtained from TAIR (https://www.arabidopsis.org/download, accessed on 20 February 2024). Additionally, protein sequences of 25 AtHSFs were retrieved from PlantTFDB (https://planttfdb.gao-lab.org/quick_search_result.php, accessed on 20 February 2024).

2.2. Identification of SrHSF Gene Family Members

Based on the local database of rosemary genome, a BLAST analysis was conducted to identify homologous proteins of 25 AtHSF members from A. thaliana (E-value < 0.05). Simultaneously, we downloaded the Hidden Markov Model (HMM) for the HSF domain (PF00447) from InterPro (https://www.ebi.ac.uk/interpro/download/, accessed on 21 February 2024). HMMER 3.0 software was used to identify proteins containing at least one HSF domain by searching across the rosemary proteome database (E-value < 0.05). Subsequently, we employed the NCBI-CDD tool and SMART (https://smart.embl.de, accessed on 21 February 2024) to verify the integrity of the conserved domains for the obtained proteins. Redundant sequences and those lacking the HSF domain were removed, obtaining the final members of the SrHSF gene family.

2.3. Chromosome Localization and Numbering of SrHSF

Based on the rosemary genome annotation file and identified SrHSF gene family members, the chromosome localization of SrHSF genes was visualized using TBtools v.2.069 and numbered according to their sequential order at chromosomes Chr1–Chr12.

2.4. Physicochemical Properties of SrHSF Proteins

The physicochemical characteristics of the SrHSF proteins were evaluated using the ExPASy tool (https://web.expasy.org/compute_pi/, accessed on 21 February 2024). Additionally, the online tool WoLF PSORT (https://wolfpsort.hgc.jp/, accessed on 22 February 2024) was utilized to predict subcellular localization.

2.5. Phylogenetic Tree Construction of SrHSF Protein Sequences

The Neighbor-Joining (NJ) method was employed to construct a phylogenetic tree for SrHSF and AtHSF proteins. Multiple sequence alignment was initially performed on 74 protein sequences of rosemary and A. thaliana using Clustal W, with parameters of gap opening at 10, extension of 0.2, delay divergent sequences at 30, and DNA transversion weight at 0.5. Subsequently, the aligned sequences were imported into MEGA11, and a phylogenetic tree was constructed using the 1000 bootstrap replicates.

2.6. Analysis of SrHSF Gene Structure and Conserved Motifs

Based on the annotation data from rosemary genome, an analysis of the SrHSF gene structure was conducted, with visualization by TBtools version 2.069. A conserved motif analysis of the SrHSF protein sequences was performed using the online software MEME (https://meme-suite.org/meme/tools/meme, accessed on 23 February 2024), setting it to display ten motifs, and the visualization of these findings was then carried out with the aid of TBtools v.2.069.

2.7. Prediction of SrHSF Promoter Functions

The promoter regions of the SrHSF members were defined as the 2000 base pair sequences upstream of the genes, isolated from the rosemary genome. Subsequently, these sequences were predicted and analyzed with the aid of the PlantCARE database (https://bioinformatics.psb.ugent.be/webtools/plantcare/html/, accessed on 24 February 2024). The visualization of the results was then carried out by using TBtools v.2.069.

2.8. Structure and Interaction of SrHSF Proteins

The prediction of the secondary structures for the SrHSF proteins was conducted by using the website SOPMA (http://npsa-prabi.ibcp.fr/cgi-bin/npsa_automat.pl?page=npsa_sopma.html, accessed on 25 February 2024). Moreover, the tertiary structures of SrHSF protein sequences were analyzed by using the CPHmodels database (https://services.healthtech.dtu.dk/services/CPHmodels-3.2/, accessed on 25 February 2024) by homology modeling. Subsequently, taking A. thaliana as the reference organism, the SrHSF protein sequences and the main action patterns of homologous proteins in Arabidopsis thaliana were searched for by using the online tool STRING (https://cn.string-db.org/, accessed on 26 February 2024). Then, TBtools v.2.069 software was utilized to visually analyze the results.

2.9. SrHSF Gene Duplication and Collinearity

Genomic information of Salvia miltiorrhiza (S. miltiorrhiza) was downloaded from the NCBI genome database (https://www.ncbi.nlm.nih.gov/, accessed on 26 February 2024). MCScanX v.2.0 was employed to identify gene pairs (GPs) within rosemary genome and between rosemary and S. miltiorrhiza genomes based on protein sequence comparisons. Collinearity relationships among A. thaliana, rosemary, and S. miltiorrhiza were visualized using TBtools v.2.069 (E value < −10). The ratios of non-synonymous to synonymous (Ka/Ks) were calculated by TBtools v.2.069 to evaluate the evolution rate of the gene-encoded protein and assess the selection pressure of genes during the evolutionary process.

2.10. Expression Pattern of SrHSF Genes

Based on the publicly available rosemary transcriptomic database (https://www.ncbi.nlm.nih.gov/bioproject/?term=PRJNA80069, accessed on 27 February 2024), we retrieved the RPKM values of the SrHSF genes in five different tissues (mature flowers, young leaves, mature leaves, secondary stems, and whole roots) and in leaves treated by 0.1% methyl jasmonate (MeJA) treatment for 48 h and 72 h (the treatment groups, T), without MeJA treatment (0 h) as the control (CK). Data with RPKM values less than 20 were excluded. The fold change in expression between the MeJA treatment group and the CK (control) group was calculated. Based on the log₁₀(RPKM) and log₂(T/CK) values, heatmaps were generated by using TBtools v.2.069 for expression pattern analysis. The parameters were set as follows: Fixed Mid Color Value = 0; Auto Polar and Cluster Rows were selected.

3. Results

3.1. Chromosomal Localization and Fundamental Information of SrHSF

Based on the genomic information of rosemary, we identified 49 SrHSF genes in total (Table S1), and their uneven distribution on the 12 chromosomes was revealed, with 8 SrHSF genes on chromosome 12 and the fewest SrHSF genes (2) on chromosomes 3, 5, and 8. One pair of tandem duplicates (spacing less than 200 kb) was on Chr12.

The sequence lengths of the 49 SrHSF proteins ranged from 82 amino acids (aa) (SrHSF40 and SrHSF48) to 750 aa (SrHSF4). The range of molecular weights was between 9.49 kDa (SrHSF40) and 82.78 kDa (SrHSF4), and the isoelectric points (pI) were from 9.46 (SrHSF21) to 4.67 (SrHSF49). Among 49 SrHSF proteins, 42 members were acidic, and 7 members were basic. Except SrHSF44, the instability coefficients of all SrHSF proteins were above 40, and the average hydrophobicity index of SrHSF proteins was below 0, implying that they were all hydrophilic proteins. Subcellular bioinformatic analysis of SrHSF proteins revealed that 45 proteins were functional in the nucleus, 3 proteins were present in chloroplasts, and 1 protein was localized in the cytoplasmic matrix. In conclusion, SrHSF proteins are mainly acidic, hydrophilic sulfur-containing proteins, which function primarily in the nucleus (Figure 1).

3.2. Phylogenetic Characteristic of SrHSF Protein Sequences

To further understand the phylogenetic characteristics between SrHSF and AtHSF, a phylogenetic tree of A. thaliana and rosemary was established (Figure 2), which was grouped into three subfamilies: (A, B, and C). There were nine subgroups (from A1 to A9) within Subfamily A, containing 27 SrHSF genes and 19 AtHSF genes. Four subgroups (from B1 to B4) are formed within Subfamily B, with 15 SrHSF genes and 3 AtHSF genes. Subfamily C contained only one subgroup (C1), containing seven SrHSF genes and one AtHSF gene. This is consistent with previous studies that there is an absence of subgroup C2 in dicots [36]. The small branches A7 and A9 of the A subfamily contained only AtHSF genes, and other branches included the SrHSF and AtHSF genes. The number of SrHSF genes in the C1 subgroup was the highest, whereas the lowest was in subgroup A5, with only one member. The results showed that the SrHSF and AtHSF gene families had similar evolutionary structural relationships, suggesting the high homology and conservative evolution of HSF genes.

3.3. SrHSF Gene Structure and Conserved Motifs

The analysis of gene structure and conserved motifs is conducive to gain more insights into gene function and evolution. As shown in Figure 3A, there were nine exons in SrHSF4, while others ranged from one to four exons. Approximately 81.6% of SrHSF genes had two exons, and 65.3% had one intron, with three genes lacking introns. Among the 49 SrHSF genes, 51% of them had one to two untranslated regions (UTRs), and no UTRs were found in 16 genes, suggesting that the functions of the SrHSF family members have changed through the evolutionary process. The gene structure between SrHSF39 and SrHSF46 (subgroup C1) was similar, correspondingly between SrHSF20 and SrHSF24 (subgroup A4). SrHSF1 and SrHSF7 (subgroup B2) were also similar in gene structure. The sequence length of the SrHSF genes varied among subfamilies, with subgroups A8 and B4 having shorter lengths than others. Except for SrHSF4, most SrHSF genes in the B2 subgroup had more introns and exons. These results indicate there were similar structures for genes in the same subfamily.

Based on the analysis of the protein sequences for 49 SrHSF, ten conserved motifs containing 11 to 41 amino acids were identified (Figure 3B). Except for SrHSF49, Motif 1 was broadly detected in the SrHSF proteins. Combined with the CDD analysis, Motif 1, Motif 2, and Motif 3 were characterized as the DBD domain. In combination with the results of the evolutionary tree, the composition of motifs within different subfamilies differed greatly. Motif 5 is only found in Subfamily B, and Motif 7 is unique in subgroups A4, A6, and A7. Motif 6 is only observed in subgroup A8.

3.4. Promoter Function of SrHSF

To further understand the regulatory mode and function of SrHSF genes, their promoter region was analyzed. In total, 22 kinds of functional elements were detected in 49 SrHSF genes, mostly involved in phytohormone response, abiotic stress response, and growth and development response (Figure 4 and S1). A total of 40 cis-elements were detected both in SrHSF31 and SrHSF32, whereas the least (12) in SrHSF36. Light-response elements and MeJA-response elements were the most widely distributed among phytohormone responsive elements. The findings suggested that the SrHSF genes were instrumental in hormonal signaling, growth and development, and the response to environmental stress in rosemary.

3.5. Structure and Protein Interaction of SrHSF Proteins

As shown by the secondary structure of SrHSF proteins in Table S2, 49 SrHSF proteins contained four secondary structures: α-helix, extended strand, β-turn, and random coil. The α-helices and random coils emerged as the predominant secondary structure, accounting for 22.57–57.89% and 27.49–57.88%, respectively. The proportion of the extended strand accounted for 3.35%-25.61%, while the proportion of β-turn was less than 15%. In addition, there were generally structures of long-chain α-helices and α-helix-β-turn α-helices in the three-dimensional structural simulation of 49 SrHSF proteins (Figure S2), which relatively agreed with the prediction results of protein secondary structures.

To explore the functional connections between proteins, we utilized homology mapping to predict the interaction network of the SrHSF proteins (Table S3, Figure 5). The interactions were found among 16 SrHSF proteins, such as SrHSF6 interacting with SrHSF30, SrHSF4, and SrHSF49. There were also interactions between SrHSF6 and AtHSF proteins like AtHSBP, AtFKBP62, AtDREB2A, AtCLPBI, AtAPX2, AtHSP90-1, AtMBF1C, AtHSP70-5, AtHSP70-4, and AtMPK3. Among the 16 interactive proteins, SrHSF6, SrHSF27, and SrHSF40 had high connectivity.

3.6. Duplication Events and Collinearity of SrHSF Genes

Gene duplication is crucial for plant genome evolution. To determine the duplication events of SrHSF genes, GPs were counted, and a homology analysis was performed within species (Figure 6A). The results revealed that 51 segmental duplications (SDs) occurred among 41 SrHSF genes, with no tandem duplication occurring. Notably, SrHSF10 and SrHSF16 participated in five SDs. The results revealed that SDs played a key role in the expansion of SrHSF genes. In addition, the Ka/Ks of all SrHSF GPs were below 1 (Table S4), suggesting that the SrHSF genes had undergone purifying selective pressure.

To more clearly elucidate the evolutionary process of SrHSF genes, collinear maps of rosemary, along with the model plant A. thaliana and the Lamiaceae plant S. miltiorrhiza, were constructed (Figure 6B). Between 26 SrHSF genes and 17 AtHSF genes, a total of 33 pairs of collinear GPs were detected, and there were 78 collinear GPs between 44 SrHSF genes and 27 SmHSF genes, implying a stronger collinear relationship between the SrHSF family and SmHSF family.

3.7. Expression Pattern of SrHSF Genes in Different Tissues and Leaves Under MeJA Treatment

The expression pattern of SrHSF genes in five different tissues (mature flowers, young leaves, mature leaves, secondary stems, and entire roots) and in leaves under 0.1% MeJA treatment at different time points (0 h, 48 h, and 72 h) are shown in Figure 7. There were 42 SrHSF genes expressed in the five rosemary tissues, displaying a tissue-specific expression. Most SrHSF genes were highly expressed in secondary stems. SrHSF27 was specifically expressed in mature flowers, and SrHSF30 and SrHSF28 were specifically expressed in entire roots. After 0.1% MeJA treatment, gene expression levels of 35 SrHSF genes were altered. Among them, SrHSF44 and SrHSF5 were significantly down-regulated at 48 h and 72 h, and SrHSF44 was down-regulated at 48 h and upregulated at 72 h. The other SrHSF genes mainly displayed an increasing trend, with SrHSF36 being markedly upregulated, followed by SrHSF11. This result implies that SrHSF36 and SrHSF11 may be regulated by JA signal.

4. Discussion

HSF serves as a key regulator in plant development and stress responses. Based on rosemary genomic data, 49 SrHSF gene family members were identified in this study. This number of SrHSF genes was more than that of S. miltiorrhiza HSF genes [37], and more SrHSF genes may be beneficial for rosemary to adapt to the hot and arid climate in its native region. The SrHSF gene family in the phylogenetic tree was grouped into three subfamilies, A, B, and C, containing 14 subgroups, while no HSF members of subgroups A9 and A7 were found in rosemary. Similarly, A9 subgroup genes were not observed in sesame [38], kiwifruit [39], and Brassica oleracea [40]. The A7 subgroup was not found in passion fruit [41], and both the A9 and A7 subgroups were not found in Phoebe bournei [42] and soybean [43]. It has been shown that after whole-genome duplication, the majority of genes exist in two copies as paralogs, and one copy of the paralog genes tends to undergo a pseudogenization process and eventually disappear. The absence of subgroups A9 and A7 in rosemary is presumably due to gene loss events during its evolution [44]. It has been suggested that genes with more exons may be the original genes in the subgroup [45]. In the SrHSF gene family, SrHSF4 contained the most exons (9), implying that it may be the original gene.

Jasmonic acid (JA) is an injury-related phytohormone and a potent signaling molecule. When plants are under external aggression or in an adverse environment, the synthesis of jasmonic acid is rapidly induced and its content increases, thus initiating a series of complex signal transduction pathways [46]. In rosemary, most of the SrHSF cis-elements were light response elements and MeJA response elements. Meanwhile, 35 SrHSF genes were regulated by 0.1% MeJA (Figure 7B). JA may regulate these genes to be involved in defense response. There are also numerous abiotic stress-related elements, including low-temperature and defense response elements, as well as MYB binding sites, suggesting that the majority of SrHSF genes participated in stress response. HSFs are transcriptional activators of HSPs, which have been revealed in eggplant [47] and Populus [48], showing that HSPs are HSF’s downstream genes. Protein interactions showed that most of the SrHSF proteins interacted with AtHSPs. However, the regulatory or interactive relationships between HSFs and HSPs and their exact function in rosemary still need further investigation. In response to stress signals, whether HSF co-activates the expression of stress-related genes with other factors also needs to be further studied.

In A. thaliana, AtHSFB4 is mainly expressed in roots with low expression in leaves, and Subfamily B may have a crucial contribution to root development [49]. Consistently in rosemary, the high homology gene of AtHSFB4 was SrHSF17, which showed a high expression level in the entire root and secondary stem while a low level in mature and young leaves. Similarly, SrHSF30 and SrHSF28 of Subfamily B also exhibited high expression levels in the root. It has been shown that the AtHSFA6b gene in A. thaliana is regulated by ABA signaling [27], which may be triggered by drought and salt stress, leading to crosstalk with hormones such as MeJA [50,51]. Exogenous MeJA can stimulate the expression of plant defense genes [52]. In peach, exogenous MeJA can upregulate the expression of PpCOI1, PpJAZ, and PpMYC2 to promote the JA signaling pathway and enhance its cold tolerance of plants [53]. Exogenous MeJA increases the accumulation of small HSP and HSP70 in tomatoes [54] and mediates HSF-HSP network expression under heat stress in perennial ryegrass [55]. In this study, the homologous genes of AtHsfA6b (SrHSF11 and SrHSF36) contained MeJA response elements and were upregulated under MeJA treatment. Whether SrHSF11 and SrHSF36 are regulated by ABA signaling or by the interaction of ABA with other signals needs to be further verified.

HSF is an important transcription factor that positively regulates plant stress resistance. In this study, based on bioinformatics analysis, we found that the expression levels of SrHSF genes differed in various tissues and external environments, indicating that their function in plant resistance to stress is not exactly the same. In the future, the knockout and overexpression of SrHSF genes need to be performed to identify further their molecular function, which will provide new insights into the molecular mechanism of rosemary resistance. Some members of the SrHSF gene family may be valuable candidate genes for resistant molecular breeding in rosemary.

5. Conclusions

This research employed bioinformatics methods to identify 49 SrHSF genes in the rosemary genome. These genes were unevenly distributed at twelve chromosomes with 51 SDs. SrHSF genes underwent purifying selection and were regulated by 22 kinds of cis-acting elements. The expression patterns of SrHSF family members differed in different tissues and in leaves when treated with MeJA. It is beneficial for further research on the function and regulatory mechanism of SrHSF genes and provides candidate genes for resistant breeding in rosemary.