Protein sequences of SAMSs from A. thaliana were obtained from the TAIR (https://www.Arabidopsis.org/). Genome and annotation files of Helianthus annuus XRQ) was downloaded from Ensembl plants (https://plants.ensembl.org/index.html). Helianthus annuus (HA89), Helianthus annuus (HA300), Helianthus annuus (IR), Helianthus annuus (LR1), Helianthus annuus (OQP8), Helianthus annuus (PI659440), Helianthus annuus (PSC8) and Helianthus annuus (RHA438) were downloaded from NCBI. The Hidden Markov Model (HMM) (PF02772, PF02773, PF00438) of S-adenosylmethionine synthase was downloaded from the Pfam database (https://pfam.xfam.org/), and were used to search protein databases by HMMER in TBtools-II (Chen et al., 2023) with an E-value<1e−5. Subsequently, all putative SAMS genes shared the three HMM domains were validated by batch-CD search (http://www.ncbi.nlm.nih.gov/Structure/cdd/wrpsb.cgi), Pfam, and HMMER (https://www.ebi.ac.uk/Tools/hmmer/) databases (Potter et al., 2018; Mistry et al., 2021; Wang et al., 2023). The SAMS genes in the XRQ cultivar are named using Latin abbreviations coupled with their chromosomal positions in the XRQ reference genome. For instance, the designation XRQ-HanSAMS1 indicates that XRQ represents the cultivar name, Han refers to Helianthus annuus, and the numeral in SAMS corresponds to the gene’s ordered position on the chromosome, listed from the smallest to the largest. Other cultivars’ genes keep their names but get a SAMS number based on where they group with XRQ-HanSAMS genes in the evolution tree. Furthermore, the biochemical parameters of HanSAMS were determined using the ProtParam tool (https://web.expasy.org/protparam/) (Gasteiger et al., 2005). Finally, the subcellular localizations of HanSAMS were predicted using the WoLF PSORT (https://wolfpsort.hgc.jp/). The NPS@: SOPMA secondary structure (https://npsa-prabi.ibcp.fr/cgi-bin/npsa_automat.pl?page=npsa_sopma.html) was used to predict the secondary structures of HanSAMS proteins. SWISS-MODEL (https://swissmodel.expasy.org/) was employed to 3D protein structure prediction and PyMOL software was used to draw 3D structures of SAMS proteins (PyMOL molecular graphics system; http://www.pymol.org) (DeLano, 2002).

Multiple sequences alignments of the full-length SAMS protein sequences was performed using the ClustalX (Larkin et al., 2007). The Neighbor-joining (NJ) tree was constructed by MEGA7.0 with the amino acid substitution Poisson model and 1000 bootstrap replicates test model (Kumar et al., 2016). The exon-intron structure of the SAMS genes was analysed using GSDS 2.0 (http://gsds.cbi.pku.edu.cn/) (Hu et al., 2015). Conserved domains of SAMS proteins were analysed by MEME (http://meme.sdsc.edu/meme/cgi-bin/meme.cgi). The upstream 2000 bp sequences relative to the start codon of each SAMS gene were obtained to analyze the promoter regions, and the cis-elements within these regions were predicted using the PlantCARE (http://bioinformatics.psb.ugent.be/webtools/plantcare/html/) (Lescot et al., 2002). We employed BLASTP to identify homologous genes, with key parameters set to an e-value threshold of 1e-3 and a maximum of 10 target sequences. To identify collinear genes using MCScanX (Wang et al., 2012), we applied the default parameters, include an E_VALUE of 1e-05 and a MAX_GAPS count of 25. The nonsynonymous substitution rate/synonymous substitution rate (Ka/Ks) values were calculated via the DnaSP 6.0 application released by Universitat de Barcelona.

A Python-compiled custom program was used to calculate the genomic composition of the SAMS gene family across nine cultivated sunflower varieties, determining the total GC content (GCall) as well as the GC content at the first (GC1), second (GC2), the average GC content at the first and second positions (GC12) and third (GC3) codon positions within the coding DNA sequences (CDS). Additionally, we utilized the software CodonW v1.4.4 (http://codonw.sourceforge.net) to assess the relative synonymous codon usage (RSCU), count the number of effective codons (ENc), and calculate the codon adaptation index (CAI), also determining the length of the amino acid sequences. Furthermore, we conducted a series of analyses to visualize the codon usage bias and neutrality: the Neutrality plot, the PR2 plot, the ENc-plot, and the RSCU plot were all generated using R software.

Hormonal response expression data (NCBI accession number SRP092742) were sourced from the SunExpress V1 database, which provides a comprehensive resource for exploring the expression patterns of genes under various conditions in sunflowers. The FPKM values for all XRQ-HanSAMS genes were extracted and subsequently processed using TBtools-II to create heatmaps.

The sunflower salt-tolerant inbred line 19S05 was used to explore the influence of salt and drought stress on sunflower seedlings. We sowed high-quality sunflower seeds in a perforated plastic container filled with nutrient-rich soil, ensuring they received regular watering every three days to support their healthy development. The plants were grown under a controlled photoperiod of 16 hours of light followed by 8 hours of darkness, all within a stable room temperature range of 21 to 25 degrees Celsius(Song et al., 2024b). Once the sunflower seedlings reached the four-true-leaf stage, seedings were treated with 150 mM NaCl solution and 15% PEG6000 solution, respectively. The leaves were then collected at 0 h, 1 h, 3 h, 6 h, 12 h, and 24 h, immediately frozen in liquid nitrogen, and stored at −80°C. The total RNA isolation and purification of samples were performed using an RNAprep Pure Plant Plus Kit (rich in polysaccharides and polyphenolics) (Tiangen, Beijing, China). The RNA isolation for gene expression was done in biological replicates for each sample analyzed. RNA integrity was visualized by 1% agarose gel electrophoresis. The concentration and purity of RNAs (OD260/OD280>1.95) were determined with a NanoDrop Onemicrovolume UVvis spectrophotometer (NanoDrop Technologies, DE, USA). Further, 1 ug of total RNA was reverse transcribed in a 20 ul reaction volume using a PrimeScript RT reagent kit with a gDNA eraser (Code No.6210A, Takara, Beijing, China) following the manufacturer’s instructions to remove traces of contaminant DNA and prepare cDNA. 1 µg of purified total RNA was reverse transcribed into the first strand cDNA that was used to qRT-PCR. Quantitative real-time polymerase chain reaction (qRT-PCR) analysis was used to analyze the expression level of the identified HanSAMSs. The standard qRT-PCR with SYBR Premix Ex Taq II (TaKaRa, Beijing, China) was repeated at least three times on a CFX96 real-time System (BioRad, Beijing). Subsequently, Cycling parameters were 95°C for 30 s, 40 cycles of 95°C for 5 s, and 60°C for 30 s. For melting curve analysis, a program including 95°C for 15 s, followed by a constant increase from 60°C to 95°C, was included following the PCR cycles. Primer Premier 6.0 software were used to designed the specific primers of HanSAMS genes according to their gene sequences, listed in Supplementary Table S1 . Results were analyzed by the 2^−△△Ct method using the HanActin as the endogenous reference gene (He et al., 2019).

A total of 58 SAMS genes are identified based on the nine sunflowers genomes, including 7 XRQ-HanSAMS, 6 HA89-HanSAMS, 7 HA300-HanSAMS, 6 IR-HanSAMS, 7 LR1-HanSAMS, 6 PI659440-HanSAMS, 6 PSC8-HanSAMS, 6 OQP8-HanSAMS, and 7 RHA438-HanSAMS ( Table 1 ). The physicochemical properties of the SAMS genes were presented in Table 1 . Their protein sequence length ranged from 390 to 391 aa, with a molecular weight (MW) varying from 42583.22 to 43012.92 Da. The isoelectric points (pI) of the protein ranged from 5.47 to 5.97. The grand average of hydropathicity (GRAVY) of the proteins ranged from -0.291 to -0.357, all were the hydrophobic proteins. Secondary structure prediction analysis revealed that the proteins encoded by all the genes were predominantly composed of α-helices, β-turns, random coils, and extended chains ( Supplementary Table S2 ). Among these, random coils were the most abundant structural element, accounting for 40.26% to 45.9% of the secondary structure. α-helices were the next most common, representing a proportion of 30.51% to 37.69%. Extended chains followed with a composition of 13.85% to 16.92%. The least prevalent structure was β-turns, which constituted only 6.92% to 8.97% of the total secondary structure content. Tertiary structure prediction showed that seven SAMS proteins from the reference genome XRQ were matching prediction s-adenosylmethionine synthase 2, which including 2 diphosphomethylphosphonic acid adenosyl ester and 2 potassium ion ( Supplementary Figure S1 ).

The phylogenetic analysis of nine cultivated sunflower SAMS proteins were performed to examine their relationships. Based on the constructed phylogenetic tree, the SAMS genes could be classified into two major clades with seven groups (SAMS1-SAMS7) ( Figure 1 ). Clade I, which is specific to sunflower SAMS genes, was found to branch into three main divisions, with each division containing a pair of distinct SAMS genes: SAMS1 with SAMS5, SAMS6 with SAMS2, and SAMS4 with SAMS7 ( Figure 1 ). It is proposed that the WGT-1 event around 38-50 million years ago (Badouin et al., 2017) was possibly responsible for generating three homologs within the sunflower SAMS gene family, establishing the three principal branches. Subsequently, the WGD-2 event, occurring approximately 29 million years ago (Badouin et al., 2017), is believed to have caused the duplication of each branch, resulting in two copies per branch and shaping the present structure of the clade I gene family. Notably, the clade II only consists of the SAMS3 group and is uniquely distributed on a separate branch and forms a striking cluster with three homologs from A. thaliana. This finding suggests that the SAMS3 group may share a common ancient ancestor with A. thaliana and appears to have not undergone the most recent whole-genome duplication event, due to lacking the partnered SAMS group that are found clustered together in other SAMS groups. To explore the expansion mechanism of the HanSAMS gene family, we analyzed gene duplication events in sunflowers using the reference genome XRQ. We found that the HanSAMS genes are distributed across seven chromosomes, with one gene per chromosome (see Figure 2 ), and no tandem clusters were identified. Subsequently, we investigated the gene collinearity within sunflowers and identified 12 pairs of duplicated genes ( Figure 2 ), suggesting that whole genome duplication (WGD) is the primary driver behind the expansion of the HanSAMS gene family. The Ka/Ks values were all lower than 1 for the duplicated genes ( Table 2 ), indicated that the SAMS gene family in sunflower has predominantly experienced purifying selection. The interspecies collinearity analysis of the HanSAMS gene families among XRQ and other eight sunflowers was further performed, and it was found that there were 134 pairs of collinearity, including 19 pairs of HA89, 19 pairs of OQP8, 18 pairs of HA300, 18 pairs of IR, 18 pairs of RHA438, 16 pairs of PSC8, 14 pairs of LR1 and 12 pairs of PI659440 ( Figure 3 ). The collinear relationship among HA89, OQP8 and XRQ genes is the strongest, followed by HA300, IR, RHA438, and the least in PI659440, which may reflect the divergency among nine different cultivated sunflowers.

Examination of gene structures revealed that all SAMS gene in nine sunflowers have only one exon and are devoid of introns ( Figure 4A ). The HA300-HanSAMS1, HA300-HanSAMS5, HA89-HanSAMS1, LR1-HanSAMS1.1, LR1-HanSAMS1.2, LR1-HanSAMS2, LR1-HanSAMS4, LR1-HanSAMS5, LR1-HanSAMS6, OQP8-HanSAMS1, OQP8-HanSAMS5, PI659440-HanSAMS1, and XRQ-HanSAMS1 were all lacked 5’ and 3’ untranslated regions (UTRs). The subcellular localization prediction suggested that the majority of SAMS proteins are predominantly found in the cytoskeleton, with the exception of SAMS3, which is localized in the cytoplasm, as shown in Figure 4B . The motifs of SAMS protein sequences were predicted using the MEME server, and all members of the SAMS contain motif1-motif10 ( Supplementary Figure S2A ), indicating highly conserved between different SAMS and different cultivars. The motif2 and motif5 were s-adenosylmethionine synthase domain (central domain), motif1 was s-adenosylmethionine synthase domain (N-terminal domain), motif3 and motif4 were s-adenosylmethionine synthase domain (C-terminal domain), about 50 amino acid residues long and is considered a key element ( Supplementary Figure S2B ).

The codon usage bias (CUB) of SAMS gene family in nine sunflowers species was investigated by analyzing the GC, GC1, GC2, and GC3 content ( Supplementary Table S3 ). The GC content of the HanSAMS genes among the nine sunflower genomes ranged from 48.95% to 52.59%, with all group of SAMS and group SAMS3 having the lowest value 50% ( Table 3 ). The GC1 content of all HanSAMS genes and the GC3 content of 91.38% of HanSAMS genes across nine sunflower species exceeded 50%, while the GC2 content remained below 50%. This suggests a notable variation in base composition at different positions and a pronounced bias towards G/C-rich start and stop codons. Although CUB across all HanSAMS genes was generally weak, as indicated by ENc values ranging from 41.76 to 53.62, there were variations among different SAMS groups. Specifically, HanSAMS2, HanSAMS4, and HanSAMS7 exhibited lower ENc values (<50%), suggesting a stronger preference for certain codons compared to HanSAMS groups 1, 3, 5, and 6 (see Supplementary Table S3 for details). The ENc-plots of HanSAMS genes exhibit deviations from the expected curve, suggesting that natural selection predominantly influences CUB ( Supplementary Figure S3 ).

The neutrality curve analysis of GC12 and GC3 values of nine sunflowers HanSAMS gene family revealed positive correlation between GC12 and GC3, with R values ranging from 0.11(IR-HanSAMS) to 0.72(HA89-HanSAMS) and the regression coefficients varying from 0.0576(IR-HanSAMS) to 0.18(HA89-HanSAMS), indicated that CUB of HansSAMS genes was mainly affected by natural selection ( Supplementary Figure S4 ). The PR2-plot analysis reveals the distribution of the third base at the codon. The results show an uneven distribution of scatters across the four regions ( Figure 5 ). Scatters in the top and bottom are predominantly in the lower half, indicating a preference for T at the third position. Those on the left and right are mostly in the left half, indicating a preference for C at the third position. Comparison among the quadrants shows the highest number of scatters in the quadrant three, suggesting a preference for C/T at the third position, implying that natural selection is the primary factor leading to CUB.

RSCU (relative synonymous codon usage) is a pivotal metric that quantifies CUB by comparing the observed frequency of each synonymous codon to its expected frequency under equal usage. The RSCU values of the HanSAMS genes were calculated and the results showed there were 22 codons shared by all nine sunflower cultivars with RSCU values greater than 1, of which 15 codons end with C/G ( Supplementary Table S4 ; Figure 6 ). Conversely, low-frequency codons, which end in A/U, were also prevalent (22 of 35), indicating a bias for these codons in the gene family. The top 3 codons with the largest average RSCU value were encode Arg (AGG with RSCU 2.36), Gly (GGU with RSCU 2.29) and Leu (CUU with RSCU 2.00). The RSCU value varied among cultivars but were generally similar, suggesting a consistent pattern of codon usage across the SAMS gene family.

Promoter cis-acting elements are crucial for regulating gene expression. We utilized PlantCARE to identify cis-acting elements in the promoter regions of 58 HanSAMS genes ( Supplementary Table S5 , Supplementary Figure S5 ). Our statistical analysis showed these elements involved in various plant processes, including growth, development, hormone response, light response, and stress response ( Figure 7A ). Notably, stress response elements predominated in the HanSAMS promoters ( Figure 7A ). In detail analysis, 302 MYB binding site, 191 anaerobic induction (ARE element),186 MYC binding site and 149 stress response element (STRE element) were predicted with high frequency in the promoter region of HanSAMS genes ( Figure 7B ; Supplementary Table S5 ). Additionally, 624 light response-related elements were identified, such as MRE (n=105), GT1-motif (n=103), G-box (n=94), Box 4 (n=87) ( Supplementary Table S5 ). Hormone response-related cis-regulatory elements were also observed, such as 108 salicylic acid responsiveness (TCA, as-1), 92 abscisic acid responsiveness (ABRE), 78 gibberellin-responsiveness, 74 MeJA-responsiveness (CGTCA-motif, TGACG-motif), and 41 ethylene response elements (ERE) ( Figure 7B ; Supplementary Table S5 ). We also noted that while most SAMS genes within the same group shared similar element distributions across different cultivars, some cultivars exhibited distinct differences (see marked with black boxes in Figure 7B ). For instance, the SAMS3 gene in the IR cultivar contained 12 salicylic acid response elements, which is significantly higher than other cultivars by at least three folds. These differences may be linked to the cultivars’ adaptability to environmental stresses and functional selection during evolution. Collectively, our findings suggest that SAMS genes are likely broadly involved in the regulation of hormones and stress responses.

HanSAMS genes may have different functions in the growth and development of sunflowers. To determine the spatial expression pattern of HanSAMS genes in sunflowers, we measured the expression levels of seven XQR-HanSAMS genes from three tissues (roots, stems, leaves) using qRT-PCR. As shown in Figure 8 , seven HanSAMS genes were expressed in all the tissues. Among them, HanSAMS1, HanSAMS3 and HanSAMS5 had similar expression patterns and were expressed highest in leaves. Meanwhile, the expression of HanSAMS4 and HanSAMS7 was higher in stems than in the other three tissues. HanSAMS2 and HanSAMS6, exhibited relatively high expression levels in the root. The results suggested that these genes showed a tissue-specific expression pattern and may play different roles in the growth and development of sunflowers.

Promoter analysis revealed there are many hormonal response elements ( Figure 7A ) the RNA-seq data available for cultivated sunflower (XRQ) were examined and to elucidate the expression patterns of SAMS genes in response to hormones. A responsive pattern was observed across all HanSAMS genes upon IAA treatment in both leaves and roots ( Figure 9 ), suggesting a pronounced sensitivity to auxin signaling (Song et al., 2024a). In leaf tissues, the HanSAMS gene family—excluding HanSAMS1—demonstrated a significant upregulation in expression following BRA (brassinosteroids) treatment, as depicted in Figure 9A . This underscores their crucial role in the regulatory pathways activated by BRA. However, in root tissues, all HanSAMS genes under BRA treatment conditions did not exhibit a significant increase compared to the control samples ( Figure 9B ). For other different hormone treatments, we also observed the different expression patterns in different tissues. For instance, HanSAMS3 demonstrates its highest expression levels in leaves following MeJA treatment ( Figure 9A ), whereas in roots, the ABA treatment elicits its peak expression ( Figure 9B ). These results indicate that the HanSAMS gene expression is modulated in a tissue-specific manner in response to hormonal signals.

Considering that the cis-elements responding to various stress existed in the promoter sequences of HanSAMS genes ( Figure 7A ), we conducted a quantitative analysis of the HanSAMS gene using qRT-PCR to examine their expression profiles under drought and salt stresses ( Figure 10 ). Our findings revealed that the HanSAMS genes exhibited distinct expression patterns at various time intervals (0 h, 1 h, 3 h, 6 h, 12 h, 24 h) following exposure to drought and salt stresses. All of the HanSAMS genes showed increased expression levels at different times under stresses, and some differences were extremely significant when compared with the untreated group (CK, 0h). In the case of drought treatment, four of the seven HanSAM genes, including HanSAMS3, HanSAMS4, HanSAMS5 and HanSAMS6, showed the highest upregulation at 12th hour, while HanSAMS1 showed the highest upregulation at the 6th hour. Notably, we observed that the expression level of the HanSAMS5 gene under drought stress is the highest among all SAMS genes (exceeding 55-fold at the 12th hour). In the case of salt treatment, five of the seven HanSAM genes, including HanSAMS2, HanSAMS4, HanSAMS5, HanSAMS6 and HanSAMS7, showed the highest upregulation at the 3th hours and gradually downregulated thereafter, while HanSAMS3 was up-regulated to highest point at 6 th of treatment. Notably, we also observed that the expression level of the HanSAMS5 gene under salt stress is the highest among all SAMS genes (exceeding 10-fold at the 3th hour). In summary, most of HanSAMS genes exhibit responsiveness to both salt and drought stress treatments, with a more rapid response observed for salt stress (peak at 3th hour) compared to drought stress (peak at 12th hour).