Whole-genome analysis reveals the tandem duplication of Cannabis sativa L. GATA and their stress-responsive expression during seed germination
1https://ror.org/00zxgrh39grid.452609.cDaqing Branch of Heilongjiang Academy of Agricultural Sciences, Daqing, 163319 China
2https://ror.org/0313jb750grid.410727.70000 0001 0526 1937Institute of Bast Fiber Crops, Chinese Academy of Agricultural Sciences, Changsha, 410205 China
3Yuelushan Laboratory, Changsha, 410128 China
4https://ror.org/00wawdr98grid.473294.fKenya Agricultural and Livestock Research Organization, Nairobi, 57811 Kenya
5https://ror.org/049pzty39grid.411585.c0000 0001 2288 989XDepartment of Agronomy, Bayero University Kano, PMB 3011, Kano, Nigeria
Abstract
Background
GATA transcription factors (TFs) are zinc finger proteins that regulate diverse developmental and stress-responsive processes in plants. Despite the growing economic importance of Cannabis sativa L. (hemp) as a source of fiber, seed oil, and bioactive compounds, the GATA TF gene family has not been systematically characterized in this species. This study aimed to perform a comprehensive genome-wide identification and characterization of GATA TF genes in C. sativa L. and to investigate their expression responses during early seed germination under abiotic stress conditions.
Results
A total of 17 GATA TF genes (CsGATAs) were identified in the C. sativa L. genome and phylogenetically classified into three clusters: Cluster I (12 members), Cluster II (one member), and Cluster III (four members). Evidence of localized gene family expansion was observed, with tandem duplication events identified in CsGATA2, CsGATA5, CsGATA6, CsGATA10, CsGATA11, and CsGATA14, distributed across Clusters I and III. Gene structure analysis revealed significant variation, with exon numbers ranging from 2 to 10. Promoter analysis of the 2000 base pairs (bp) upstream regions showed enrichment of stress and hormone-responsive cis-acting elements, with abscisic acid (ABA) responsive elements present in 12 of the 17 promoters. Transcriptomic profiling during early germination under cold stress (4 °C), salt stress (200 mM NaCl), and combined cold and salt stress revealed distinct expression patterns. CsGATA3 was consistently downregulated across all three stress conditions (log2FC of -1.14 under salt stress, -1.08 under cold stress, and -1.23 under combined stress). In contrast, CsGATA9 and CsGATA14 responded specifically to combined stress, with CsGATA9 downregulated (log2FC of -1.62) and CsGATA14 upregulated (log2FC of + 1.62). qRT-PCR analysis of five differentially expressed CsGATA genes confirmed the RNA-seq expression trends, showing strong concordance between the two platforms (Pearson r = 0.83, p < 0.001).
Conclusions
This study provides the first genome-wide characterization of the GATA TF family in C. sativa L. and reveals gene family expansion driven by tandem duplication. The identification of stress-responsive members, particularly CsGATA3 as a broadly downregulated gene across all stress conditions and CsGATA14 as specifically induced under combined stress, highlights candidate transcriptional modulators of abiotic stress adaptation during hemp seed germination. These findings lay a foundation for functional studies and molecular breeding strategies aimed at improving stress tolerance in hemp.
Supplementary Information
The online version contains supplementary material available at https://doi.org/10.1186/s12870-026-09308-w.
Introduction
Abiotic stresses, including drought, salinity, and extreme temperatures, are major constraints on crop growth and productivity, reducing yields and compromising the structural integrity of fiber tissues [1–5]. In fiber crops such as industrial hemp (Cannabis sativa L.), these stresses disrupt secondary cell wall biosynthesis, leading to reduced cellulose deposition, induced lignification, and altered fiber properties [6, 7]. As industrial hemp is increasingly cultivated globally for textiles, composites, and biobased materials [8], elucidating the molecular mechanisms involved in its growth, development, and adaptation to abiotic stresses is crucial for sustaining stable production and fiber quality.
Plants have evolved sophisticated mechanisms to perceive and respond to environmental changes, including reactive oxygen species (ROS) signaling, hormone-mediated pathways, and transcriptional reprogramming that modulate growth, metabolism, and defense [9, 10]. Among transcriptional regulators, the GATA family of zinc finger proteins is associated with responses to environmental and developmental stimuli [11]. These transcription factors (TFs) are evolutionarily conserved across eukaryotes, recognizing the consensus DNA motif (A/T)GATA(A/G) through a single, highly conserved type-IV zinc finger domain (C-X2-C-X17-20-C-X2-C) followed by a basic region. In angiosperms, the GATA family is typically large and diverse, with about 30 members in Arabidopsis thaliana and Oryza sativa L. [11, 12].
Functionally, plant GATA TFs participate in a wide range of developmental and physiological processes. For example, A. thaliana GATA21/GNC and GATA22/GNL regulate the transition from heterotrophic to autotrophic growth by coordinating chloroplast development, greening, and photosynthesis, while integrating hormonal and nitrogen-related signals that influence flowering time and senescence [12, 13]. More broadly, GATA TFs are involved in seed germination, root and leaf development, and responses to abiotic stresses such as cold, drought, and salinity. In O. sativa L., OsGATA16 enhances tolerance to cold stress at the seedling stage, and other members are critical for salt and drought stress responses [14]. Recent genome-wide characterizations in crops such as wheat and grapevine have further highlighted the crucial roles of GATA genes in mediating multi-stress resilience [15, 16].
Despite their recognized importance, a systematic characterization of the GATA TF family in C. sativa L. has not yet been reported. Key gaps remain regarding their phylogenetic relationships, genomic organization, and regulatory roles under abiotic stress. This is particularly during seed germination, a developmental stage highly sensitive to environmental conditions and critical for stand establishment and subsequent fiber yield [17]. In hemp production systems, early-season germination often coincides with low soil temperatures and residual salinity, making combined cold and salt stress a recurrent agronomic challenge [18, 19]. Moreover, the potential involvement of C. sativa L. GATA (CsGATA) genes in linking environmental stress responses with pathways influencing secondary cell wall formation, a primary determinant of fiber strength and quality, remains largely unexplored.
To address these critical knowledge gaps, this study conducted a comprehensive genome-wide identification and characterization of GATA TFs in C. sativa L. Using phylogenetic, conserved motif, synteny, promoter, and transcriptomic analyses under cold, salt, and combined cold-salt stress conditions, we identified candidate CsGATA genes potentially involved in abiotic stress-responsive regulatory pathways during early seed germination. In addition, alternative splicing analysis provided novel insights into the structural diversification and post-transcriptional regulation of the CsGATA family under stress. Compared with previous GATA family studies in model plant species, this study uniquely integrates combined abiotic stress transcriptomics with evolutionary and post-transcriptional alternative splicing analyses in an economically important industrial fiber crop. These findings establish a robust molecular framework for future functional studies and targeted molecular breeding strategies aimed at improving stress resilience, seed vigor, and agronomic performance in hemp.
Materials and methods
Plant material, germination assay, and stress treatments
Seeds of C. sativa L. var. ‘Qingdama 5’ were obtained from the Gene Bank of the Daqing branch of Heilongjiang Academy of Agriculture Science. ‘Qingdama 5’ is a high-yielding, fiber-oriented cultivar characterized by strong lodging resistance (< 5% under field conditions), an erect stem architecture, and significant tolerance to saline-alkaline soils (pH 7.2–8.5) supported by a well-developed root system. The cultivar has high fiber yield (9,015.3 kg ha−1), enhanced resistance to gray mold and stem rot compared with traditional varieties, and completes its growth cycle within 105–115 days. Seeds were stored at 4 °C in sealed foil packets with silica gel desiccant and used within six months of the harvest date. Viability was confirmed prior to the experiment by a preliminary germination test conducted according to ISTA protocols [20]; only seed lots with ≥ 85% germination under control conditions were used.
Seeds were surface sterilized in 2% (v/v) sodium hypochlorite for 10 min with gentle agitation and rinsed five times with sterile distilled water. A 7 day germination assay was conducted to evaluate the effects of cold stress, salt stress, and their combination on seed germination. Fifty seeds per dish were placed in 90 mm Petri dishes lined with two layers of sterile Whatman No. 1 filter paper moistened with 10 mL of sterile distilled water or NaCl solution as specified. To prevent evaporation while allowing gas exchange, dishes were sealed with Parafilm and perforated with three 1 mm pinholes. Moisture levels were checked daily; an additional 2 mL of the corresponding solution (distilled water or NaCl of the same concentration) was added when visual drying of the filter paper was observed. Each treatment consisted of three independent biological replicates, and all dishes were arranged in a completely randomised design within the incubator, with positions rotated 90° daily to minimise positional effects.
Salt concentrations were determined based on preliminary dose–response assays (0–500 mM NaCl) to identify levels that inhibited germination without causing complete lethality (Figure S1). In the preliminary assay, seeds (50 per dish, three replicates per concentration) were germinated in five NaCl concentrations (100, 200, 300, 400, and 500 mM) at 25 ± 1 °C under the standard conditions described above. The 200 mM NaCl concentration was selected for the main experiment as it produced consistent partial inhibition of germination without complete lethality.
Four treatments were applied: (i) control, seeds imbibed in distilled water (0 mM NaCl) at 25 ± 1 °C; (ii) cold stress, seeds exposed to 4 °C for 12 h in darkness during early imbibition and then transferred to 25 ± 1 °C; (iii) salt stress, seeds incubated in 200 mM NaCl at 25 ± 1 °C for 72 h followed by transfer to distilled water for recovery; and (iv) combined stress, seeds exposed to 200 mM NaCl at 4 °C for 12 h, then maintained in 200 mM NaCl at 25 ± 1 °C until 72 h and subsequently transferred to distilled water. Following any cold treatment, all dishes were maintained at 25 ± 1 °C under a 16/8 h light/dark photoperiod with a photosynthetic photon flux density of 120 μmol m−2 s−1. All 25 °C incubations were conducted in a programmable illuminated growth chamber. Cold treatments (4 °C) were performed in a separate temperature-controlled incubator; temperature accuracy was verified daily using a calibrated digital thermometer placed at dish level. NaCl solutions were prepared using analytical-grade NaCl dissolved in sterile distilled water.
Germination was recorded daily for 7 days, and seeds were considered germinated upon radicle emergence ≥ 2 mm. Counts were performed at the same time each day (09:00 h). Seeds showing visible fungal contamination were removed and recorded separately; replicates in which contamination exceeded 5% of the total seeds per dish were excluded from analysis. Germination percentage (GP) and mean germination time (MGT) were calculated according to Ellis and Roberts [21]:where n is the total number of germinated seeds, N is the total number of seeds tested, ni is the number of seeds germinated on day di, and di is the number of days from the start of imbibition.
The germination index (GI) [22] was calculated as:where Gi is the number of seeds germinating on day di.
Time to 50% germination (T₅₀) was derived from a three-parameter logistic model fitted to cumulative germination curves:where G(t) is cumulative germination percentage at time t, Gmax is the estimated maximum germination percentage, k is the rate constant, and T50 is the time at which 50% of Gmax is reached. Model fitting was performed using the drc package in R [23].
All germination parameters (GP, MGT, GI, and T50) were subjected to one-way analysis of variance (ANOVA) followed by Tukey’s Honest Significant Difference (HSD) post-hoc test for pairwise comparisons among treatments. Homogeneity of variance was verified using Levene’s test prior to ANOVA. Where assumptions of normality or homoscedasticity were violated, a non-parametric Kruskal–Wallis test was applied as a confirmatory analysis, with Dunn’s post-hoc test used for multiple comparisons. All statistical analyses were performed in R (v4.5) using the agricolae, stats, and FSA packages. Significance was defined at p < 0.05. All data are presented as mean ± standard error (SE) of three independent biological replicates.
For gene expression analyses, seedlings were grown under identical conditions and harvested at 48 h post-imbibition, corresponding to early germination prior to recovery in salt treatments. At harvest, the entire seedling (radicle and hypocotyl) was collected. For each of three biological replicates, 20 seedlings were pooled, flash-frozen in liquid nitrogen within 60 s of collection, and stored at −80 °C to capture early stress-responsive transcriptional events.
Identification and sequence analysis of CsGATA genes
The C. sativa L. reference genome (CS10, National Center for Biotechnology Information (NCBI BioProject PRJNA557030)) was queried using HMMER (v3.3.2) [24] with the ZnF_GATA zinc finger profile (Pfam: PF00320) [25] (E-value < 1 × 10–5). Candidate sequences were retained only if confirmed by at least one additional database (SMART [26] or NCBI CDD [27]). Sequences lacking the conserved zinc-finger motif (C-X2-C-X17–20-C-X2-C) upon manual inspection of the multiple sequence alignment were excluded. Predicted gene models were cross-validated against the CS10 genome annotation (annotation release 100) to confirm exon–intron boundaries and coding sequence (CDS) integrity, and genes with incomplete open reading frames (ORF) or annotation conflicts were excluded from downstream analyses.
Phylogenetic analysis
Full-length GATA proteins from C. sativa L., A. thaliana (TAIR10), and O. sativa L. (IRGSP-1.0) were aligned with MAFFT (v7.505) [28] using the L-INS-i algorithm. Poorly aligned regions were trimmed with trimAl (v1.4) [29]. Maximum likelihood (ML) phylogeny was inferred using IQ-TREE (v2.3) [30] with ModelFinder [31] selected substitution models, 1000 ultrafast bootstrap replicates [32], and 1000 SH-aLRT tests. Trees were visualized and annotated using iTOL [33].
Chromosomal mapping, synteny, and motif analyses
Chromosomal locations were obtained from CS10 genome annotations and visualized with TBtools-II (v2.329) [34, 35]. Synteny was assessed with MCScanX [36] using an E-value threshold of 1 × 10–5 and a minimum block size of five genes. Conserved motifs were predicted with Multiple Em for Motif Elicitation (MEME Suite—v5.5.5) [37] using a maximum of 10 motifs, motif width of 6–50, and E-value < 1 × 10–5. Promoter sequences (2000 base pairs (bp) upstream of each start codon) were analyzed with PlantCARE [38] for hormone and stress-responsive cis-elements.
Physicochemical property and subcellular localization analysis
The physicochemical properties of CsGATA proteins, including amino acid (aa) length, molecular weight (kDa), theoretical isoelectric point (pI), instability index, and grand average of hydropathicity (GRAVY), were calculated using the ExPASy ProtParam tool [39]. Subcellular localization of CsGATA proteins was predicted using WoLF PSORT [40] with default parameters, supplemented by plant-specific localization prediction tools.
Alternative splicing analysis
Isoform diversity was evaluated using StringTie assemblies [41]. Isoforms with transcripts per million (TPM) > 1 in at least one treatment were retained for downstream analysis. Exon and intron structures were visualized using TBtools-II [35], and plots summarizing alternative splicing patterns were generated using Python.
RNA extraction and sequencing
Total RNA was extracted from germinating seedlings using the SteadyPure Plant RNA Extraction Kit (Accurate Biotechnology, Changsha, China). RNA integrity was assessed using an Agilent 2100 Bioanalyzer, and only samples with an RNA Integrity Number (RIN) ≥ 7.0 were retained for library construction. RNA concentration and purity were determined using a NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific, USA), with A260/A280 ratios between 1.8 and 2.1 accepted.
For each sample, 1 μg of high-quality total RNA was used for strand-specific library construction using the NEBNext Ultra II Directional RNA Library Prep Kit for Illumina (New England Biolabs, USA). Poly(A)+ mRNA was enriched using oligo(dT) magnetic beads and fragmented into short fragments prior to first-strand cDNA synthesis using random hexamer primers. After second-strand synthesis, end repair, A-tailing, adaptor ligation, and PCR amplification were performed according to the manufacturer’s instructions. The resulting libraries were quantified by qPCR and quality-assessed on an Agilent 2100 Bioanalyzer prior to sequencing.
Libraries were sequenced on an Illumina NovaSeq 6000 platform (Illumina, USA) in paired-end 150 bp mode. An average of approximately 41 million clean reads were obtained per sample (range: 37.8–46.1 million), with Q30 scores exceeding 96% across all libraries (Supplementary Table S2).
Transcriptome assembly and differential expression analysis
Raw sequencing reads were quality-filtered using fastp (v0.23.1) [42] with the following parameters: adapter trimming enabled, bases with Phred quality scores below 20 removed by sliding window (window size 4 bp), reads with more than five ambiguous (N) bases discarded, and reads shorter than 100 nucleotide (nt) after trimming excluded together with their paired reads. Clean reads were aligned to the C. sativa L. reference genome (CS10; NCBI BioProject PRJNA557030, annotation release 100) using HISAT2 (v2.1.0) [43] in strand-specific mode, guided by the CS10 GFF3 annotation file. Alignment rates exceeded 91% for all libraries (Supplementary Table S3).
Transcript assembly and expression quantification were performed using StringTie (v1.3.3b) [41, 44] in reference-guided mode (–rf flag), retaining only annotated gene loci; novel intergenic transcripts were excluded. Gene expression levels were calculated as transcripts per million (TPM) and raw read counts. Genes with total read counts below 10 across all samples were excluded prior to statistical analysis. Differential expression analysis was performed using DESeq2 (v1.12.4) [45] in R. Genes with |log2FC|≥ 1 and Benjamini–Hochberg-adjusted p-value (q-value) < 0.05 were considered significantly differentially expressed. The following pairwise comparisons were performed: (i) salt stress alone (200 mM NaCl, 25 °C) vs. control (0 mM NaCl, 25 °C); (ii) cold stress alone vs. control; and (iii) combined stress vs. control. Gene functional annotation was based on the CS10 official annotation supplemented by BLASTp searches (E-value < 1 × 10⁻5) against the NCBI non-redundant protein database and cross-referenced with UniProt/Swiss-Prot. Heatmaps and expression visualizations were generated using the pheatmap and ggplot2 packages in R.
Co-expression networks were constructed using Pearson correlation coefficients calculated across condition means. An adjacency matrix was built using a correlation threshold of r ≥ 0.75. The resulting network was analyzed using the igraph [46] package in R to calculate degree centrality, betweenness centrality, and average correlation strength for hub gene identification.
qRT-PCR validation and statistical analyses
Total RNA was extracted from the same seedling samples used for RNA-seq using the SteadyPure Plant RNA Extraction Kit (Accurate Biology, Beijing, China). RNA concentration and purity were verified using a NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific, USA), accepting only samples with A260/A280 ratios between 1.8 and 2.1. RNA integrity was confirmed by 1% agarose gel electrophoresis.
First-strand cDNA was synthesized from 200 ng total RNA using the Evo M-MLV RT Mix Kit with gDNA Clean for qPCR Ver.2 (Accurate Biology, Beijing, China) according to the manufacturer’s instructions, which includes an on-column genomic DNA elimination step prior to reverse transcription. Gene-specific primers for five CsGATA genes selected for validation (CsGATA3, CsGATA8, CsGATA9, CsGATA10, and CsGATA14) were designed using Primer3 [47] and are provided in Supplementary Table S4. qRT-PCR was performed on a CFX96™ Real-Time PCR System (Bio-Rad, USA) using SYBR Green chemistry. Thermal cycling conditions were as follows: initial denaturation at 95 °C for 3 min; 40 cycles of 95 °C for 15 s, 60 °C for 30 s, and 72 °C for 30 s; followed by melt curve analysis from 65 °C to 95 °C in 0.5 °C increments to confirm amplicon specificity. CsActin was used as the internal reference gene [48]. Relative expression levels were calculated using the 2-∆∆ct method [49], with the 0 mM NaCl control treatment as the calibrator.
All statistical analyses were performed in R (v4.5.0). For germination parameters (GP, MGT, GI, and T₅₀) and qRT-PCR expression data, normality of residuals was assessed using the Shapiro–Wilk test and homogeneity of variance was verified using Levene’s test (R package car). Data meeting parametric assumptions were analyzed by one-way ANOVA followed by Tukey’s HSD post-hoc test using the agricolae package. Where assumptions were violated, the non-parametric Kruskal–Wallis test with Dunn’s post-hoc test (Benjamini–Hochberg correction; R package FSA) was applied as a confirmatory analysis. Agreement between RNA-seq and qRT-PCR expression profiles was assessed by Pearson correlation between |log2FC| values from both platforms using the cor.test() function, reported at both the overall (all five genes) and per-gene level. Differences were considered statistically significant at p < 0.05. All data are presented as mean ± SE of three independent biological replicates.
Results
Identification and chromosomal mapping of CsGATA genes in C. sativa L.
Seventeen GATA genes were identified in the C. sativa L. reference genome (CS10) and designated CsGATA1 to CsGATA17 following the sequential nomenclature conventions established in genome-wide GATA TF characterization studies in other plant species [50–52]. These genes encoded proteins ranging from 151 aa (CsGATA16) to 538 aa (CsGATA8), with predicted molecular weights between 16.2 and 59.4 kDa and theoretical pI values from 4.75 to 9.89 (Table 1). All GATA proteins exhibited negative GRAVY values (range: −1.025 to −0.542; mean −0.75), indicating predominantly hydrophilic character. Instability indices ranged from 34.85 to 65.41 (mean 51.02), with most values exceeding 40, suggesting variable in vitro stability among family members. Subcellular localization prediction indicated nuclear localization for all 17 CsGATA genes, consistent with their role as TFs.Gene Protein ID Length MW (kDa) pI GRAVY Instability index Localization Chromosome CsGATA1 XP_030480651.1 301 32.46 6.21 −0.654 38.01 Nucleus NC_044370.1 CsGATA2 XP_030480667.1 365 40.40 5.09 −0.679 46.96 Nucleus NC_044370.1 CsGATA3 XP_030487548.1 339 37.16 6.06 −0.721 44.15 Nucleus NC_044371.1 CsGATA4 XP_030487739.1 369 41.33 9.04 −0.973 46.38 Nucleus NC_044371.1 CsGATA5 XP_030488344.1 347 37.72 6.46 −0.566 58.72 Nucleus NC_044371.1 CsGATA6 XP_030498922.1 333 36.57 5.24 −0.777 62.72 Nucleus NC_044373.1 CsGATA7 XP_030499018.1 304 34.12 8.02 −1.025 55.6 Nucleus NC_044373.1 CsGATA8 XP_030501038.1 538 59.42 6.4 −0.648 55.68 Nucleus NC_044374.1 CsGATA9 XP_030502249.1 278 30.99 6.18 −0.867 54.5 Nucleus NC_044374.1 CsGATA10 XP_030503179.1 388 42.16 4.75 −0.542 49.51 Nucleus NC_044375.1 CsGATA11 XP_030504713.1 317 33.89 5.57 −0.656 34.85 Nucleus NC_044375.1 CsGATA12 XP_030505639.1 181 19.65 9.89 −0.75 40.56 Nucleus NC_044375.1 CsGATA13 XP_030506133.1 325 36.15 9.1 −0.765 49.72 Nucleus NC_044375.1 CsGATA14 XP_030508431.1 267 30.29 8.69 −0.777 54.21 Nucleus NC_044376.1 CsGATA15 XP_030508436.1 273 30.37 9.17 −0.697 60.44 Nucleus NC_044376.1 CsGATA16 XP_030508484.1 151 16.17 9.88 −0.948 65.41 Nucleus NC_044376.1 CsGATA17 XP_030511114.1 352 38.28 4.98 −0.714 49.91 Nucleus NC_044377.1
Chromosomal mapping showed that CsGATA genes were unevenly distributed across seven of the ten assembled chromosomes, with no members detected on NC_044372.1, NC_044378.1, or NC_044379.1 (Fig. 1). NC_044375.1 harbored the highest number of CsGATA genes (four), followed by NC_044371.1 and NC_044376.1 with three genes each. Two genes were located on NC_044370.1, NC_044373.1, and NC_044374.1, while NC_044377.1 contained a single member (CsGATA17). The physical clustering of CsGATA10/CsGATA11 and CsGATA14-CsGATA16 suggests that tandem duplication events may have contributed to the expansion of this gene family in C. sativa L.
Gene structure and conserved motifs of CsGATA TFs
The 17 C. sativa L. GATA genes showed significant diversity in exon and intron organization (Fig. 2A). Exon numbers ranged from 2 to 10, with corresponding intron counts from 1 to 9. CsGATA2 and CsGATA10 were the most structurally complex, each comprising 10 exons and 9 introns, whereas CsGATA3, CsGATA7, CsGATA9, CsGATA14, and CsGATA15 had only 2 exons and a single intron. Most other genes, such as CsGATA4-6, CsGATA12-13, and CsGATA16-17, had an intermediate structure of 3 exons and 2 introns.
Domain analysis confirmed the universal presence of the ZnF_GATA zinc finger across all CsGATA genes (Supplementary Table S5). Most members aligned with established domain models such as SMART ZnF_GATA (smart00401) or CDD ZnF_GATA (cd00202). However, CsGATA9 showed a weaker association with the GAT1 superfamily, suggesting possible divergence from canonical GATA functions. In addition, de novo motif discovery with MEME identified 10 statistically significant motifs across the family (E-value < 0.05; Fig. 2B). Motif 1, which corresponded precisely to the canonical GATA zinc finger, was detected in all 17 members with high conservation. Motif 2 was observed in a subset of longer proteins, positioned immediately C-terminal to the zinc finger. Motif 3, observed only in CsGATA1, CsGATA2, CsGATA10, and CsGATA11, was strongly enriched in basic residues, a pattern characteristic of nuclear localization signals. Several other accessory motifs were distributed unevenly across the family.
The overall distribution of motifs was consistent with phylogenetic clustering. Genes belonging to the CsGATA1/2/10/11 cluster shared a common suite of accessory motifs. In contrast, those with simplified exon and intron structures, such as CsGATA14 and CsGATA15, retained only the core zinc finger and a minimal set of additional elements. Intron-rich members, such as CsGATA2 and CsGATA10, carried the most diverse motif repertoires, indicating that structural complexity and motif diversification tend to coincide within the GATA family.
Phylogenetic analysis of GATA TFs in C. sativa L.
Phylogenetic analysis of the GATA TF family across C. sativa L., O. sativa L., and A. thaliana resolved the sequences into three well supported clusters (Fig. 3A). In C. sativa L., all 17 GATA genes were distributed among these clusters, with Cluster I having 12 members, including CsGATA3, CsGATA4, CsGATA5, CsGATA6, CsGATA7, CsGATA9, CsGATA12, CsGATA13, CsGATA14, CsGATA15, CsGATA16, and CsGATA17. This cluster represented the largest group. Cluster II had a single gene, CsGATA8. Cluster III had four members: CsGATA1, CsGATA2, CsGATA10, and CsGATA11. Notably, Cluster I members were dominated by conserved ZnF_GATA and GATA domains, suggesting functional conservation within this subgroup. Cluster III members, although fewer, included both O. sativa L. and C. sativa L. GATA genes, indicating potential evolutionary divergence between monocot and dicot lineages. Cluster II only had CsGATA8, suggesting a divergent lineage, consistent with its unique phylogenetic position.
Synteny relationships of GATA genes
A. thaliana vs C. sativa L.
Synteny analysis between C. sativa L. and A. thaliana identified conserved collinearity for seven out of the 17 CsGATA genes (Fig. 3B). In total, 11 syntenic links were detected, involving CsGATA5, CsGATA6, CsGATA7, CsGATA12, CsGATA13, CsGATA15, and CsGATA16. These genes were located on CsChr2, CsChr3, CsChr4, CsChr5, and CsChr6, and corresponded to Arabidopsis orthologs distributed across AtChr1, AtChr2, AtChr3, AtChr4, and AtChr5. Notably, CsGATA6 (CsChr4) and CsGATA16 (CsChr6) showed multiple syntenic connections with Arabidopsis genes, reflecting possible ancient duplication and subsequent retention events. In contrast, single copy syntenic relationships were observed for CsGATA5 (At3G54810, AtChr3), CsGATA7 (At2G18380, AtChr2), CsGATA12 (At5G26930, AtChr5), CsGATA13 (At1G08000, AtChr1), and CsGATA15 (At3G24050, AtChr3). The recurrent mapping of CsGATA16 to both At3G06740 (AtChr3) and At4G16141 (AtChr4) further suggests functional diversification following duplication. Interestingly, 10 CsGATA genes lacked detectable synteny with Arabidopsis.
O. sativa L. vs C. sativa L.
Comparative synteny analysis revealed that a subset of C. sativa L. GATA TFs established conserved genomic blocks with O. sativa L. Four syntenic relationships were identified, involving CsGATA3, CsGATA6, and CsGATA7 (Fig. 3C). CsGATA3 had two distinct syntenic links, one with Oryza chromosome 1 (Os01t0745700-01) and another with chromosome 5 (Os05t0520300-00), underscoring its potential evolutionary conservation across multiple genomic regions. CsGATA6 showed a conserved collinear block with Os04t0539500-01 on O. sativa L. chromosome 4, while CsGATA7 was syntenic with Os05t0578900-01 on chromosome 5. No other C. sativa L. GATA members displayed syntenic relationships with O. sativa L. in the present analysis.
Structural and regulatory diversity of CsGATA genes
Alternative splicing diversity of CsGATA genes
The 17 CsGATA genes showed significant complexity, generating one to five isoforms per gene (Fig. 4). CsGATA2 had the highest splicing diversity with five transcript variants, having 10–12 exons each. In contrast, CsGATA13 and CsGATA16 maintained a single isoform with 3–4 exons, suggesting conserved functional roles. CsGATA10 had four transcripts with exon variations concentrated at the 5’ and 3’ ends.
CsGATA1 isoforms demonstrated extended regions of exons in the central CDS, indicating potential modulation of protein-DNA interaction domains. At the isoform level, exon counts per transcript ranged from three (CsGATA16) to 12 (CsGATA2), with alternative splicing primarily involving exon skipping, observed in nine genes, and alternative donor/acceptor sites (six genes), reflecting gene-specific diversification mechanisms.
Cis-regulatory element (CRE) landscape of CsGATA promoters
The promoter regions, defined as 2000 bp upstream of the ATG start codon, of the CsGATA gene family were analyzed to elucidate the CRE landscape, revealing a highly heterogeneous distribution of motifs associated with phytohormone signaling and environmental response (Fig. 5). Elements responsive to key phytohormones were highly prevalent across the family, implicating the CsGATA genes in diverse developmental and stress-response pathways. ABA responsive elements (ABREs) were the most widespread, identified in the promoters of 12 out of 17 CsGATA genes, with a frequency ranging from 2–5 motifs per promoter. Genes with notable ABRE abundance included CsGATA1, CsGATA2, CsGATA5, and CsGATA10. Furthermore, auxin-responsive TGA-elements were concentrated in the promoters of CsGATA2, CsGATA8, and CsGATA11, suggesting regulatory links to cell elongation and tropisms. Gibberellin-responsive GARE-motifs, typically linked to germination and flowering, were detected in CsGATA2 and CsGATA11. The convergence of multiple hormone-related motifs in the CsGATA2 promoter indicates its potential function as a key transcriptional hub integrating complex upstream signals.
Regulation via light and stress stimuli was confirmed by the abundance of relevant CREs. Light responsive elements (G-box, Box 4, and AE-box) were found present in 15 of the 17 gene promoters. The highest motif density was observed in the CsGATA5 and CsGATA8 promoters, each housing up to six light-related elements, underscoring their potential role in controlling photomorphogenesis and photosynthesis in C. sativa L. Stress-specific CREs were also enriched in certain members, conferring enhanced environmental adaptability. Motifs associated with drought response (MBS) and low-temperature (LTR) were significantly clustered in CsGATA2, CsGATA10, and CsGATA14. This specific enrichment strongly suggests that these gene members may play a role as candidates for functional studies focusing on improving abiotic stress tolerance and resilience in C. sativa L. cultivars, which is a critical trait for industrial crop productivity.
Germination performance of C. sativa L. var. Qingdama 5 under cold, salt, and combined stress
Germination performance of C. sativa L. var. ‘Qingdama 5’ was significantly affected by stress treatment. One-way ANOVA across all 12 treatment groups (four stress regimes × five NaCl concentrations plus control and cold stress alone) revealed a highly significant treatment effect on final germination percentage at day 7 (p < 0.001). Levene’s test confirmed homogeneity of variance prior to ANOVA (p = 0.500). This result was consistent with Kruskal–Wallis test (p = 0.003), validating the parametric analysis.
For the four main treatment groups (control, cold stress, 200 mM salt stress, and combined cold-salt stress at 200 mM NaCl), GP differed significantly (p < 0.05; Table 2). The control achieved the highest GP (94.7 ± 2.9%), which was significantly greater than all stress treatments (Tukey’s HSD; p ≤ 0.01 for all comparisons with control). Cold stress alone (4 °C, 0 mM NaCl) reduced GP to 48.7 ± 11.2%, though this decline did not differ significantly from salt stress at 200 mM (41.3 ± 6.4%; p = 0.895) or combined stress (28.7 ± 6.8%; p = 0.298). Combined cold and salt stress produced the lowest GP, though it did not differ significantly from salt stress alone at the same concentration (p = 0.640).Treatment GP (%) MGT (days) GI T50 (days) Control 94.7 ± 2.9 a 3.23 ± 0.46 a 22.40 ± 3.17 a 2.73 ± 0.53 Cold 48.7 ± 11.2 b 2.94 ± 0.17 a 10.89 ± 2.90 b 2.35 ± 0.23 Salt 200 mM 41.3 ± 6.4 b 3.95 ± 0.12 a 6.05 ± 1.02 b 3.43 ± 0.03 Combined 28.7 ± 6.8 b 4.20 ± 0.25 a 3.69 ± 0.59 b 3.60 ± 0.13
MGT showed a significant overall treatment effect (p < 0.05), although no individual pairwise comparison reached significance after Tukey’s HSD correction (all p > 0.05). MGT was shortest in the cold stress treatment (2.94 ± 0.17 days) and longest under combined stress (4.20 ± 0.25 days; Table 2). The Kruskal–Wallis test returned a non-significant result for MGT (p = 0.092), consistent with the Tukey outcome.
The GI was significantly affected by treatment (p < 0.05). The control exhibited the highest GI (22.40 ± 3.17), significantly exceeding cold stress (10.89 ± 2.90; p < 0.05), salt stress at 200 mM (6.05 ± 1.02; p < 0.05), and combined stress (3.69 ± 0.59; p < 0.05). No significant differences in GI were detected among the three stress treatments (all p > 0.18).
The time to 50% germination (T50), derived from three-parameter logistic model fitting to cumulative germination curves (R2 > 0.97 for all fits), showed a trend toward later onset under stress conditions. T50 was shortest under cold stress (2.35 ± 0.23 days) and control (2.73 ± 0.53 days), and longest under combined stress (3.60 ± 0.13 days) and salt stress (3.43 ± 0.03 days; Table 2). However, neither ANOVA (p > 0.05) nor Kruskal–Wallis (p > 0.05) reached significance for T50 after Tukey correction, and no individual pairwise comparison was significant (all p > 0.05).
Across the full dose–response range (100–500 mM NaCl), a clear concentration-dependent reduction in GP was observed under both salt stress alone and combined cold-salt stress (Table 3). Under salt stress, GP declined from 56.0 ± 14.7% at 100 mM to 13.3 ± 2.9% at 500 mM NaCl, accompanied by a progressive increase in MGT from 3.74 ± 0.11 to 5.94 ± 0.24 days. Combined cold-salt stress consistently produced lower GP and higher MGT than the corresponding salt-alone concentration at 200–500 mM NaCl, with the 500 mM combined treatment yielding GP of only 14.0 ± 4.2% and MGT of 6.23 ± 0.34 days.NaCl (mM) GP—Salt (%) MGT—Salt (d) GP—Combined (%) MGT—Combined (d) 100 56.0 ± 14.7 3.74 ± 0.11 39.3 ± 8.1 3.44 ± 0.13 200 41.3 ± 6.4 3.95 ± 0.12 28.7 ± 6.8 4.20 ± 0.25 300 40.0 ± 6.1 5.21 ± 0.27 29.3 ± 2.4 4.53 ± 0.38 400 20.0 ± 2.0 5.15 ± 0.10 18.7 ± 5.5 5.23 ± 0.04 500 13.3 ± 2.9 5.94 ± 0.24 14.0 ± 4.2 6.23 ± 0.34
Transcriptional expression patterns of GATA genes in C. sativa L. under abiotic stress
To comprehensively characterize the transcriptional behavior of the CsGATA gene family under abiotic stress, expression profiles of all identified CsGATA genes were analyzed across salt, cold, and combined stress conditions during germination (Supplementary Table S6). Using stringent differential expression criteria (|log2FC|≥ 1 and q < 0.05), only a subset of genes exhibited significant transcriptional responses, whereas the majority showed relatively stable expression patterns with modest fluctuations. Differential expression analyses were mainly interpreted relative to the control treatment (0 mM NaCl, 25 °C) to identify stress-responsive CsGATA genes. Additional pairwise comparisons between combined stress and the corresponding single-stress treatments (combined vs. salt and combined vs. cold) were performed to identify transcriptional responses specifically associated with stress interactions.
Under salt stress (200 mM NaCl), CsGATA3 was the sole DEG, exhibiting consistent and statistically significant downregulation (log2FC = −1.14; q < 0.05). Several other genes, including CsGATA14, CsGATA15, and CsGATA16, displayed statistically significant but modest changes (|log2FC|< 1). Similarly, cold stress (low-temperature) elicited a limited response, with only CsGATA3 being significantly downregulated (log2FC = −1.08, q < 0.05). All other CsGATA members maintained stable expression levels (q > 0.05). Heatmap visualizations confirmed CsGATA3 as the primary cold-responsive gene (Fig. 6A).
In contrast, the combined cold and salt stress treatment (synergistic stress) elicited stronger and more complex transcriptional changes. Three genes were significantly differentially expressed: CsGATA3 (downregulated, log2FC = −1.23, q < 0.05), CsGATA9 (downregulated, log2FC = −1.62, q < 0.05), and CsGATA14 (upregulated, log2FC = + 1.62, q < 0.05). Several other genes, including CsGATA5, CsGATA12, and CsGATA15, showed moderate but statistically significant changes (|log2FC|< 1). The observed overlap of CsGATA3 regulation across cold, salt, and combined conditions established it as a potential broad abiotic stress responder (Fig. 6B and C).
Pairwise comparisons revealed additional stress-specific signatures. Between salt and combined stress, CsGATA9 was significantly more repressed in the combined treatment (log2FC = −1.37, q < 0.05). Between cold and combined stress, CsGATA9 was induced (log2FC = + 1.16, q < 0.05), whereas CsGATA14 was repressed (log2FC = −1.26, q < 0.05).
Validation of RNA-seq expression profiles by qRT-PCR
Five representative CsGATA genes, including CsGATA3, CsGATA8, CsGATA9, CsGATA10, and CsGATA14 were selected to validate the reliability of the transcriptome data analysis under salt, cold, and combined stress conditions using qRT-PCR. The qRT-PCR expression patterns were generally consistent with the RNA-seq results (Fig. 6D). CsGATA3 expression decreased to 0.45–0.52-fold under all stress treatments relative to control, consistent with transcriptome results. CsGATA9 decreased to 0.17-fold under combined stress but remained stable under single stresses. CsGATA14 was strongly induced under combined stress (2.71-fold), while CsGATA8 and CsGATA10 showed moderate responses (< 1.5-fold) that matched transcriptome trends. Pearson correlation analysis between RNA-seq |log2FC| values and qRT-PCR relative expression data demonstrated a strong positive correlation (r = 0.834, p < 0.001, n = 15), supporting the accuracy and reproducibility of the transcriptomic analysis (Fig. 7).
Co-expression of CsGATA TFs with cell wall biosynthesis genes
To explore potential co-expression relationships between GATA TFs and secondary cell wall biosynthesis genes under abiotic stress, we constructed condition-specific and overall hub co-expression networks using Pearson correlation analysis. For each condition, networks were built by integrating all 17 CsGATA genes from the respective transcriptome data with phenylpropanoid pathway genes identified in the corresponding co-expression datasets (Fig. 8).
The overall hub genes co-expression network across all conditions (Fig. 8D) revealed a highly interconnected module. Hub gene analysis based on degree centrality identified 4CL CCL10 as the top hub (Degree = 8), followed by CCR1, 4CL-like 1, CsGATA9, and CsGATA14 (all with Degree = 7) (Table 4). Importantly, the three stress-responsive CsGATA genes showed strong network connectivity: CsGATA9 and CsGATA14 (Degree = 7 each) and CsGATA3 (Degree = 5). These GATAs exhibited robust positive correlations with key phenylpropanoid pathway genes, including PAL1, COMT, CCR1, and 4CL family members.Gene Degree Betweenness Avg. Correlation 4CL CCL10 8 2.00 0.900 CCR1 7 4.17 0.793 4CL-like 1 7 2.00 0.826 CsGATA9 7 0.45 0.858 CsGATA14 7 0.45 0.800 PAL1 6 2.20 0.714 COMT 6 0.45 0.705 4CL CCL1 5 1.45 0.600 CsGATA3 5 0.83 0.613 4CL2 4 0.00 0.513
Condition-specific networks demonstrated dynamic rewiring depending on the stress type (Fig. 8A-C). The cold stress network (Fig. 8A) showed broad interconnectivity among CsGATA members but limited co-expression with phenylpropanoid pathway genes, with only COMT and PAL1 meeting the correlation threshold (r ≥ 0.75). The salt stress network (Fig. 8B) revealed expanded pathway gene associations, with CsGATA members co-expressed with PAL1, COMT, CCR1, 4CL-like 1, 4CL CCL1, and 4CL2. The combined stress network (Fig. 8C) displayed the most extensive associations, with particularly strong connectivity between CsGATA9, CsGATA14, and key lignin biosynthesis genes including 4CL CCL10, 4CL-like 1, CCR1, and PAL1, consistent with their identification as hub genes in the overall network (Fig. 8D). These patterns indicate stronger co-expression associations between GATA TFs and lignin biosynthesis genes under synergistic cold + salt stress, but they do not by themselves demonstrate direct regulatory relationships.
Discussion
Transcriptional regulation is crucial for plant adaptation to environmental stress, especially during seed germination, when early developmental processes are strongly influenced by environmental conditions [1]. Although GATA TFs have been extensively studied in model plants [11], their genome-wide organization and stress-responsive behavior remain largely unexplored in industrial fiber crops. This study provides the first comprehensive characterization of the GATA TF family in C. sativa L. and integrates evolutionary, structural, transcriptomic, and co-expression analyses to elucidate their potential roles in abiotic stress responses during early germination.
Seventeen CsGATA genes were identified, a family size comparable to that reported in A. thaliana and O. sativa L. [11]. Instead of large-scale whole-genome duplication, tandem duplication events appear to have driven localized expansion, particularly in Clusters I and III. Such localized duplications are common among TF families and frequently contribute to functional diversification [53]. Phylogenetic and synteny analyses revealed both conserved and lineage-specific evolutionary patterns. Notably, the stress-responsive members (CsGATA3, CsGATA9, and CsGATA14) clustered within the largest clade (Cluster I), which also showed conserved synteny with A. thaliana and O. sativa L. GATA genes. This suggests that stress-responsive regulation in hemp is primarily mediated by evolutionarily conserved GATA lineages.
Consistent with A. thaliana and O. sativa L. [50], CsGATA proteins showed significant diversity in gene structure, physicochemical properties, and alternative splicing patterns. The high isoform diversity observed in genes such as CsGATA2 and CsGATA10 indicates that post-transcriptional regulation may enhance functional plasticity under stress conditions. Alternative splicing is increasingly recognized as a mechanism that enables plants to rapidly adjust TF activity under fluctuating environmental conditions, especially during stress-sensitive developmental phases [54].
All CsGATA proteins contained the conserved ZnF_GATA domain and were predicted to localize predominantly to the nucleus, consistent with their role as transcriptional regulators. Negative GRAVY values across the family further support their hydrophilic nature and compatibility with dynamic protein-DNA and protein–protein interaction. Collectively, these structural features indicate that CsGATA proteins possess the molecular attributes required for responsive transcriptional regulation during early development. Promoter analysis revealed enrichment of hormone and stress-responsive cis-elements (ABRE, MBS, and LTR), supporting the observed transcriptional responsiveness of several CsGATA genes. The presence of multiple ABREs in several promoters aligns with the known role of ABA signaling in seed germination and stress responses [55]. This heterogeneity aligns with the concept of functional partitioning within TF families, where individual members respond to distinct combinations of environmental and hormonal signals. In the context of seed germination, such diversification may enable fine-tuned regulation of growth, energy allocation, and stress tolerance during a critical physiological transition [56].
Transcriptomic profiling under cold, salt, and combined stress revealed that CsGATA genes exhibit highly specific and condition-dependent expression patterns under abiotic stress. Among all family members, CsGATA3 was consistently downregulated across all conditions. This broad responsiveness suggests that CsGATA3 may function as a stress-responsive transcriptional modulator whose downregulation is associated with growth restraint or transcriptional reprogramming under unfavorable physiological conditions. In contrast, CsGATA9 and CsGATA14 showed specific responses to combined stress. These patterns are consistent with the concept of functional partitioning within TF families, allowing plants to fine-tune responses to complex environmental conditions encountered during germination. To further examine the downstream regulatory roles of these stress-responsive GATAs, we performed condition-specific co-expression network analyses focusing on their association with phenylpropanoid and cell wall biosynthesis genes. The three stress-responsive CsGATAs (CsGATA3, CsGATA9, and CsGATA14) exhibited high network connectivity across conditions, with particularly strong associations to key lignin pathway genes (PAL1, COMT, 4CL family members, and CCR1). These findings suggest that CsGATAs are associated with transcriptional responses linked to secondary cell wall remodeling under stress. However, the correlation network does not establish direct regulatory control over lignin biosynthesis genes. Similar regulatory links between transcription factors and cell wall defense genes have been reported in cotton under Verticillium wilt stress [57].
Seed germination represents a crucial physiological reference point in the plant life cycle, particularly for crops such as hemp that are often sown under suboptimal environmental conditions. Cold stress during imbibition can impair membrane fluidity and enzymatic activity [58], while salinity imposes osmotic and ionic constraints that disrupt cellular homeostasis [59]. The co-occurrence of these stresses can therefore exert compounded physiological pressure on germinating seeds. In this study, the focus on combined cold and salt stress addresses a biologically relevant scenario encountered during early-season hemp cultivation. The identification of CsGATA genes that respond specifically to combined stress conditions underscores the importance of regulatory mechanisms that integrate multiple environmental signals. Such integration is crucial for coordinating growth suppression, stress tolerance, and recovery during early establishment. Although this study does not establish direct causal roles for CsGATA genes, the integration of expression profiling with evolutionary and promoter analyses and co-expression analyses provides a strong foundation for functional inference. The consistent repression of CsGATA3 across stress conditions, coupled with the stress-specific behavior of CsGATA9 and CsGATA14, identifies these genes as promising candidates for future functional validation. Therefore, the potential involvement of CsGATA genes in coordinating stress responses with metabolic and developmental processes warrants further investigation.
Conclusion
This study provides the first comprehensive genomic and transcriptomic characterization of the GATA TF family in C. sativa L. Seventeen CsGATA genes showing significant structural diversity and regulatory complexity were identified. Phylogenetic and synteny analyses revealed both conserved and lineage-specific evolutionary patterns, suggesting functional stability alongside adaptive diversification within the hemp genome. Transcriptomic analyses under cold, salt, and combined stress conditions identified CsGATA3, CsGATA9, and CsGATA14 as key stress-responsive genes during early seed germination. Co-expression analysis across treatment conditions revealed that CsGATA14 and CsGATA9 emerged as highly connected hub genes within the overall correlation network linking GATA TFs with phenylpropanoid and secondary cell wall biosynthesis genes, including CsCCR-1, CsCOMT paralogs, and Cs4CL paralogs. Direct experimental validation of these regulatory relationships remains to be explored; these findings nonetheless identify CsGATA14 and CsGATA9 as promising candidates for future functional studies and molecular breeding strategies aimed at improving stress resilience and fiber quality in hemp.
Supplementary Information
Acknowledgements
The authors thank the Gene Bank of the Daqing branch of Heilongjiang Academy of Agriculture Science for providing seeds of Cannabis sativa L. var. Qingdama 5. The authors also thank Higentec (Beijing, China) for RNA sequencing services.
Funding
This study was supported by the Seed Industry Special Project of Yuelushan Laboratory (YLS-2025-ZY02059), China Agricultural Research System (CARS-16), the Agricultural Science and Technology Innovation Project of the Chinese Academy of Agricultural Sciences (CAAS-ASTIP-IBFC), and Hunan Province Science and Technology Innovation Program (2026RC9014). The funding bodies had no role in study design, data collection or analysis, interpretation of results, or preparation of the manuscript.
Data availability
The Cannabis sativa reference genome (CS10) used in this study is publicly available under NCBI BioProject accession PRJNA557030. The RNA-seq raw sequencing data generated in this study has been submitted to the NCBI Sequence Read Archive (SRA) and will be publicly available upon acceptance of this manuscript. All other data supporting the findings of this study are included in the article and its supplementary materials.
Declarations
Ethics approval and consent to participate
Not applicable. This study did not involve human participants, human data, human tissue, or vertebrate animals.
Consent for publication
Not applicable. This manuscript does not contain data from any individual person.
Competing interests
The authors declare no competing interests.