Genome-wide analysis of the AP2 transcription factor family in cultivated sugarcane reveals ShAP2-1 as an enhancer of drought tolerance

Abstract

Background

Sugarcane (Saccharum spp. hybrid) is a primary global source of sugar and a vital feedstock for bioethanol, yet its productivity is severely limited by abiotic stresses, particularly drought. The AP2 family, belong to the APETALA2/ ethylene-responsive factor (AP2/ERF) superfamily, plays crucial roles in plant development and stress adaptation. However, a comprehensive analysis of the AP2 family in the complex polyploid genome of cultivated sugarcane remains lacking, limiting its potential for genetic improvement.

Results

A genome-wide survey of the AP2 family in the sugarcane cultivar XTT22 identified 211 members, representing a significant expansion compared to wild relatives. Phylogenetic analysis classified them into euAP2, euANT, and basalANT clades, with a notable dominance of the euANT subfamily. Their expansion was primarily driven by whole-genome/segmental duplication, with only 4.59% of duplicated pairs under positive selection. Promoter analysis revealed a predominance of hormone- and stress-responsive cis-elements. Expression profiling demonstrated distinct tissue-specific patterns (e.g., leaf-preferential ShAP2-1 and stem-specific ShAP2-56) and rapid, transient induction by abscisic acid, with stronger responses in young leaves. Functional validation showed that ShAP2-1 overexpression enhanced drought tolerance in transgenic rice by both markedly promoting primary root growth (2.4-fold) to improve water uptake, and mitigating oxidative stress by elevating superoxide dismutase activity while reducing malondialdehyde accumulation.

Conclusion

This study provides the first comprehensive characterization of the AP2 transcription factor family in cultivated sugarcane, elucidating its expansion history, structural diversity, and spatiotemporal regulation. ShAP2-1 enhanced drought tolerance by promoting root growth for water acquisition and activating antioxidant defenses for cellular protection.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12870-026-08633-4.

Introduction

Drought and salinity are among the most pervasive abiotic stresses that limit crop productivity worldwide, driving an urgent need for more resilient varieties [1]. To survive these challenges, plants undergo extensive transcriptional reprogramming, a process orchestrated by transcription factors (TFs) that act as central regulators, translating environmental cues into adaptive cellular programs [2]. The APETALA2/Ethylene-Responsive Factor (AP2/ERF) superfamily represents one of the largest groups of plant-specific TFs. Its members are defined by a conserved AP2 domain and are renowned for their ability to integrate growth and stress signaling pathways [3, 4]. This superfamily is divided into five main lineages—AP2, ERF, DREB, RAV, and Soloist—based on their domain architecture. The AP2 lineage is unique in possessing two tandem AP2 domains, a feature that sets it apart structurally from the single-domain ERF and DREB families [5, 6]. Within angiosperms, the AP2 family further diversifies into three phylogenetically conserved clades: euAP2, basalANT, and euANT [7, 8]. Members of this family typically bind to AT-rich cis-elements, thereby directly regulating vast networks of target genes that govern developmental processes [9–11]. This functional focus on development contrasts with the roles of other major lineages: ERF proteins primarily mediate responses to ethylene, pathogens, and wounding through GCC-box elements [12, 13], while DREB family members are key activators of drought and salinity tolerance via DRE/CRT elements [14, 15].

The story of AP2 transcription factors began with foundational studies in Arabidopsis developmental genetics. Here, the AP2 protein itself emerged as a key regulator of floral organ identity. It directly guided sepal and petal development by repressing the C-class gene AGAMOUS within specific floral whorls [9, 16]. Another family member, AINTEGUMENTA (ANT), controlled organ size by promoting cell proliferation during primordia development. When expressed ectopically, ANT caused a dramatic enlargement of leaves, flowers, and seeds, clearly demonstrating its growth-promoting role [17]. This core function in growth regulation is not only conserved in cereal crops but also expanded to govern critical agronomic traits. In rice, AP2 family genes are master regulators of inflorescence architecture and grain yield. For instance, FRIZZY PANICLE (FZP) strictly governed the transition from spikelet to floral meristem. Loss-of-function mutants showed uncontrolled branch elongation and failed to produce florets, while its overexpression transforms sterile lemmas into lemma-like organs. This highlighted its essential role in maintaining meristem determinacy and organ identity [18, 19]. Working in a related pathway, SUPERNUMERARY BRACT (SNB) and Indeterminate spikelet 1 (IDS1) acted in a partially redundant manner to regulate spikelet meristem fate and floral organ patterning, which ultimately affected grain shattering and size [20, 21]. Notably, specific mutations in SNB could increase grain length and weight, providing a direct genetic link between AP2 function and a key yield component [22]. Another critical player, RICE STARCH REGULATOR1 (RSR1), negatively regulated starch biosynthesis genes, thereby shaping grain quality by modulating amylose content and gelatinization temperature [23]. Collectively, these examples show how AP2 transcription factors serve as central hubs within genetic networks that orchestrate fundamental agronomic traits.

Moving beyond developmental patterning, AP2 transcription factors are also deeply embedded in hormonal signaling networks. This integration allows plants to precisely coordinate their growth with environmental conditions [24]. Essentially, these factors act as central mediators of crosstalk between major phytohormone pathways. A prime example is OsAP2-39, which modulated the balance between abscisic acid (ABA) and gibberellin (GA) to regulate processes spanning from seed germination to plant height and panicle development [25]. In the context of root development, the auxin-induced gene CROWN ROOTLESS5 (Crl5) promoted crown root formation by repressing cytokinin signaling. This illustrated a recurring theme: AP2 factors effectively translate hormonal gradients into specific morphological outcomes [26]. By serving as such integral nodes in hormonal networks, AP2 members become crucial arbitrators of the growth-defense balance, enabling plants to adaptively tailor development in response to internal and external signals. While the DREB family is more specialized for stress responses, many AP2 members also contribute to abiotic stress tolerance in significant and nuanced ways, often leveraging the hormone pathways they regulate. For instance, the cold-induced gene HcTOE3 from Halostachys caspica boosted freezing tolerance in Arabidopsis. It did so by upregulating two sets of targets: classic CBF-dependent cold-response genes and key ABA signaling components like ABI1, ABI5, and RD29A. This pointed to a coordinated, dual-pathway strategy for enhancing resilience [27]. The connection between stress and development is often direct. In rice, OsIDS1 bound to GCC-boxes in the promoters of salt-stress genes such as LEA1 and SOS1. To repress their expression, OsIDS1 recruited the transcriptional co-repressor TOPLESS-RELATED PROTEIN1 (TPR1) and the histone deacetylase HDA1, forming a dedicated repressive complex that actively reduced salt tolerance [28].Therefore, stress adaptation by AP2 factors involves specific transcriptional and developmental reprogramming.

Sugarcane (Saccharum spp. hybrid) stands as a cornerstone of global agriculture, producing nearly 80% of the world’s sugar and serving as a key feedstock for bioethanol, thereby underpinning major bioeconomies [29]. Its productivity, however, is critically constrained by environmental stresses. Drought represents the primary limitation, frequently responsible for yield losses of up to 30% [30]. This vulnerability is especially pressing in major producing regions such as China, where rain-fed cultivation dominates and large-scale irrigation is often unavailable. Developing cultivars with enhanced intrinsic drought tolerance is therefore a paramount breeding goal for ensuring sustainable and resilient sugarcane production. Despite the clear agronomic imperative, the molecular basis of stress tolerance in cultivated sugarcane remains poorly understood relative to model plants. A major reason for this knowledge gap lies in the organism’s extraordinary genomic complexity. Modern commercial cultivars are highly polyploid and aneuploid, stemming from the interspecific hybridization of S. officinarum and S. spontaneum [31, 32]. While foundational studies have characterized AP2/ERF families in the wild relative S. spontaneum and the related genus Erianthus fulvus [33, 34], their genomic architectures and evolutionary histories differ significantly from those of commercial hybrid cultivars. Consequently, the specific gene complement, expression profiles, and functional roles of the AP2 family within the complex, hybrid genome of elite sugarcane are largely unexplored. This study aims to address this gap through a comprehensive genome-wide analysis of the AP2 transcription factor family in a commercial sugarcane cultivar. We pursue three specific objectives: (1) to systematically identify all AP2 members and characterize their phylogenetic relationships, gene structures, conserved motifs, and promoter cis-regulatory elements; (2) to profile their spatiotemporal expression patterns across tissues and in response to abscisic acid (ABA)—a central hormone mediating plant adaptation to drought, salinity, and temperature stress; and (3) to functionally validate the contribution of a candidate gene, ShAP2-1, to drought tolerance. By integrating comparative genomics, transcriptomics, and transgenic approaches, this work will provide new insights into the evolution and functional specialization of the AP2 family in a polyploid crop. We anticipate that our findings will uncover valuable genetic resources for molecular breeding, paving the way for developing sugarcane varieties with superior drought resilience—a trait essential for stabilizing production in an era of climate change.

Results

Identification and phylogenetic analysis of the AP2 family genes

Our genome-wide analysis of the Saccharum spp. hybrid XTT22 genome identified 232 putative AP2 family genes, each predicted to encode a protein containing two conserved AP2 domains. Following the removal of redundant sequences, 211 non-redundant AP2 transcription factors were retained for in-depth characterization. Analysis of their physicochemical properties revealed substantial diversity. The encoded proteins varied in length from 295 to 1647 amino acids, with molecular weights ranging from 31.68 to 178.84 kDa. Their aliphatic indices spanned from 53.18 to 93.07, while the grand average of hydropathicity (GRAVY) values fell between − 0.804 and − 0.038 (Table S1). Most of these proteins (205 out of 211) were predicted to be unstable (instability index > 40), and 93 were basic in nature (pI > 7). To resolve the evolutionary relationships within this family, we constructed a maximum likelihood (ML) unrooted phylogenetic tree using the 211 ShAP2 proteins along with their homologs from Arabidopsis and Oryza sativa (Fig. 1A). Based on well-established phylogenetic criteria, the ShAP2 members were classified into three subfamilies: euAP2, euANT, and basalANT. The euANT subfamily was the largest, containing 89 members, while the basalANT group was the smallest with 55 members. Integrating chromosomal localization data with the phylogenetic structure revealed a hierarchical organization of the gene family. The 211 genes were consolidated into 70 distinct gene models (designated ShAP1–ShAP70) using a sequence identity threshold of ≥ 95%. Each model was further subdivided into subtypes (e.g., ShAP2-2.1, ShAP2-2.2) that reflect specific genomic contexts and evolutionary lineages. Given that the XTT22 reference genome comprises 97 chromosomes organized into 10 homologous groups (HGs) [32], we performed a focused phylogenetic analysis on a single, representative chromosome set to reduce complexity. This analysis showed that the 27 ShAP2 members present on this chromosome set were distributed across all three subfamilies (Fig. 1B; Table S1).

Fig. 1.

Phylogenetic tree of ShAP2 family genes. A Unrooted phylogenetic tree showing relationships among AP2 family genes of Saccharum spp. hybrid, Arabidopsis thaliana, Oryza sativa. B Phylogenetic tree constructed by maximum likelihood (ML) estimation of the AP2 TFs in the single chromosome set of Saccharum spp. hybrid XTT22. Different colors were used to classify subfamilies

Gene structure and conserved motif analysis

We next examined the structural diversity of ShAP2 genes by analyzing their exon-intron organization within the representative chromosome set. Distinct splicing patterns emerged across the three subfamilies (Fig. 2). Genes in the basalANT subfamily were the most compact, typically containing 6 exons (range 6–8). In contrast, euANT members predominantly had 8 exons (range 6–9). The euAP2 subfamily displayed the widest variation in exon number, ranging from 2 to 17 exons, with 9 exons being the most common. Notably, ShAP2-34.1 represented the minimal structure with only 2 exons, while ShAP2-29 was the most complex with 17. Analysis of intron lengths further highlighted subfamily differences: the euAP2 clade exhibited the highest degree of intron length polymorphism, whereas basalANT genes possessed the shortest average intron length. Analysis of conserved protein domains confirmed that all members contained the defining two AP2 domains, with one exception. The ShAP2-57 protein harbored one complete AP2 domain and a second, N-terminally truncated version. We also identified several accessory domains. Most strikingly, ShAP2-29 encoded a full SSM4 domain, suggesting a potential role as an E3 ubiquitin-protein ligase (DOA10). Several other proteins featured partial domains, including PRK11633, Atrophin-1, and Pollen_Ole_e_1. The pattern of conserved motifs largely reinforced the phylogenetic grouping. Aside from AP2-34.1, which contained only motif1, the majority of proteins possessed more than seven conserved motifs. These motifs were arranged in a highly conserved, ordered sequence (motif2-4-7-5-3-1). Furthermore, motif8 was exclusive to the euAP2 clade, and motif9 was uniquely found in the euANT clade, providing distinct molecular signatures for these lineages.

Fig. 2.

Phylogenetic relationships, gene structure, and conserved protein motifs of sugarcane AP2 genes. In the schematic, yellow boxes represent the 5’- and 3’- UTR regions, green boxes denote exons, and gray lines indicate introns. Distinct protein domains are highlighted in different colored boxes. Conserved motifs (numbered 1–10) are also shown in individual colored boxes, with corresponding sequence details provided in Table S2. Protein and gene lengths can be estimated using the scale bar at the bottom of the figure

Chromosome distribution, duplication and evolutionary analysis of AP2 genes

To trace the evolutionary origins of the ShAP2 genes, we leveraged the partitioned nature of the XTT22 allopolyploid genome, which comprises subgenomes derived from S. officinarum (So) and S. spontaneum (Ss). A clear bias was observed in their contributions: 156 (73.9%) of the ShAP2 genes originated from the So subgenome, which is known to be dominant in modern sugarcane hybrids, while 49 (23.2%) were assigned to the Ss subgenome; the origin of two genes remained ambiguous. These 205 ShAP2 genes (excluding the six located on unplaced contigs) were mapped across the chromosomal landscape. They were distributed unevenly across nine of the ten HGs. Homologous group Chr8 was an exception, containing no ShAP2 genes on its six constituent chromosomes. Among the remaining groups, the average number of ShAP2 genes per chromosome was highest in HG Chr1 (4.3) and lowest in HG Chr4 (1.1), indicating significant variation in local gene density. We next investigated the mechanisms behind this gene family’s expansion by analyzing duplication events. The vast majority of ShAP2 genes (193) were products of whole-genome duplication (WGD) or segmental duplication. In contrast, dispersed, proximal, and tandem duplications accounted for only 12, 4, and 2 genes, respectively. Two clear examples of tandem duplication were identified: the ShAP2-2.1/ShAP2-2.2 pair on chromosome 1 A and the ShAP2-13/ShAP2-52 pair on chromosome 7 H (Fig. 3A). Notably, the ShAP2-13/ShAP2-52 pair showed significant sequence divergence following duplication, resulting in markedly reduced protein sequence identity. To assess the evolutionary forces shaping these genes, we calculated the nonsynonymous-to-synonymous substitution rate ratio (Ka/Ks) for syntenic ShAP2 gene pairs (Table S3). The Ka/Ks ratios for segmentally duplicated pairs ranged from 0.10 to 1.92. While most ratios were below 1 (indicating purifying selection), 24 gene pairs (4.59% of analyzable pairs) had ratios greater than 1, providing evidence of positive selection. Furthermore, we estimated the divergence times for these syntenic pairs based on their Ks values. The duplications occurred over a wide temporal range, from approximately 0.51 to 62.6 million years ago (Mya), with the lower end of this spectrum pointing to very recent duplication events in the evolutionary history of sugarcane.

Fig. 3.

Chromosomal distribution and duplication analysis of AP2 family genes. A Schematic of the chromosomal locations of sugarcane AP2 genes. Lines connecting genes denote duplication events: red, tandem duplication; purple, proximal duplication; green, non‑homologous segmental duplication. B, C Synteny analysis of AP2 genes between sugarcane and O. sativa (B) or E. fulvus (C). Gray lines in the background depict collinear blocks across the genomes; red lines highlight syntenic AP2 gene pairs

To trace the evolutionary trajectory of the AP2 family, we performed a comparative synteny analysis with two key reference species: rice (a monocot model) and Erianthus fulvus (a close wild relative of sugarcane). This analysis identified 30 syntenic gene pairs between sugarcane and rice, and 28 pairs between sugarcane and E. fulvus (Fig. 3B, C). The high degree of conservation was striking: 26 sugarcane ShAP2 genes had syntenic counterparts in both species. Only one gene pair was unique to the sugarcane-rice comparison, and one to the sugarcane-E. fulvus comparison, underscoring the deep evolutionary conservation of this gene family across the grasses. A notable pattern emerged from the synteny map: several sugarcane ShAP2 genes were associated with a single ortholog in rice or E. fulvus. This “one-to-many” relationship suggests that these ancestral loci were pivotal hotspots for lineage-specific expansion within the sugarcane AP2 family. For example, ShAP2-36 (including Sh_Ss06K0265903.t1, Sh_So06K0263924.t1, Sh_So06K0265503.t1) shows synteny with Erufi.06G0024620.t1 and LOC_Os04g55970; ShAP2-3.1 (Sh_Ss01A0031173.t1, Sh_Ss02F0115175.t1), ShAP2-4.1 (Sh_Ss01A0029784.t1), and ShAP2-20.1 (Sh_So02F0115836.t1) are syntenic with Erufi.02G0010470.t2; The aforementioned ShAP2-3.1, ShAP2-4.1, and ShAP2-20.1 also exhibit synteny with rice LOC_Os07g13170.

Analysis of putative cis-regulatory elements (CREs) in AP2 promoters

Analysis of the promoter regions revealed a complex regulatory landscape for the sugarcane ShAP2 genes, dominated by CREs associated with plant hormone response, transcription factor binding, and environmental signaling (Fig. 4A, Table S4). Quantitatively, hormone-responsive elements constituted the largest category (30%), followed by transcription factor binding sites (23%), light-responsive elements (20%), stress-responsive elements (20%), and development-responsive elements (7%). The hormone-responsive category was notably diverse, implicating pathways for ABA, MeJA, GA, ethylene, salicylic acid, and auxin. Elements related to ABA (e.g., ABRE) and MeJA (e.g., CGTCA-motif) were the most abundant, highlighting these hormones as likely key regulators of ShAP2 expression (Fig. 4B). Among the 27 genes analyzed, the number of ABA-responsive elements varied widely, ranging from 1 to 25 per promoter, with a mean of 8.3 (Fig. 4C). ShAP2-12.1 stood out with 25 such elements, whereas ShAP2-36.1 and ShAP2-6.1 each contained only one. Promoters were also enriched for binding sites of major transcription factor families, including MYB, MYC, DRE, and WRKY. MYB and MYC sites were particularly prevalent. The count of MYB sites per gene ranged from 2 to 15 (mean 7.7), with ShAP2-13.1 possessing the maximum. All but one gene (ShAP2-28.1) contained MYC sites, averaging 4.4 per gene, and ShAP2-34.1 had the highest number (10). Stress-responsive elements pointed to adaptation to diverse environmental challenges, including wounding (STRE, WRE3), anaerobic conditions (ARE), low temperature (GC-motif, LTR), and general defense (TC-rich repeats). STRE and ARE elements were the most common within this group. The average promoter contained 3.9 STRE elements (with ShAP2-2.1 and ShAP2-27.1 each having 10) and 2.7 ARE elements (ShAP2-27.1 had a maximum of 11). In contrast, development-related elements were less widespread but highly specific. These included motifs for meristem-specific activation (CCGTCC-box, CAT-box) and seed-specific regulation (RY-element). Their presence was concentrated in a subset of genes, suggesting specialized roles. For example, ShAP2-8.1 and ShAP2-28.1 harbored 6 and 4 CCGTCC-boxes, respectively, while ShAP2-36.1 and ShAP2-26.1 contained 5 and 4 RY-elements, implicating them in the control of meristem and seed development.

Fig. 4.

Analysis of cis-regulatory elements in AP2 gene promoters. A Pie chart depicting the classification of predicted cis-regulatory elements. B Stacked bar chart illustrating the relative frequency of each element type. C Schematic representation of the 2.5 kb promoter region upstream of each gene; distinct colored boxes correspond to different CRE

Functional enrichment analysis of the AP2 gene family

The Gene Ontology (GO) enrichment analysis of 211 AP2 family genes revealed that these genes were predominantly involved in transcriptional regulation, biosynthetic processes, hormone signaling pathways, organ morphogenesis, and stress responses (Fig. 5). The most significantly enriched functional categories were “regulation of transcription, DNA-templated” and “regulation of biosynthetic process”, with all 211 genes annotated to both terms, underscoring their central role in controlling gene expression and metabolic pathways. Hormone signaling pathways were also highly represented, including ethylene-activated signaling pathway, hormone-mediated signaling pathway, auxin-mediated signaling pathway involved in phyllotactic patterning, cellular response to hormone stimulus, negative regulation of cytokinin-activated signaling pathway, and response to abscisic acid. These results indicate that the AP2 family genes are extensively involved in mediating responses to multiple phytohormones, which are critical for growth, development, and adaptation to environmental cues. Furthermore, the analysis highlighted significant enrichment in developmental processes, particularly those related to root and shoot architecture. Terms such as root development, lateral root formation, adventitious root development, post-embryonic plant organ morphogenesis, radial pattern formation, leaf phyllotactic patterning, and specification of floral organ identity were all significantly enriched. This suggests that these genes play key roles in shaping plant architecture and organogenesis. Additionally, the presence of terms like regulation of cutin biosynthetic process, positive regulation of fatty acid biosynthetic process, and seed germination further supports their involvement in lipid metabolism, barrier formation, and reproductive development, which are often linked to stress tolerance. In summary, the GO enrichment analysis demonstrates that the AP2 gene family is functionally multifaceted, with a primary role in transcriptional and biosynthetic regulation, extensive participation in hormone signaling networks, and crucial contributions to organ development and stress responses.

Fig. 5.

Gene Ontology (GO) enrichment analysis of biological processes for AP2 gene family. The color gradient from red to yellow to blue represents the −log₁₀(q-value), where red indicates higher enrichment significance (smaller q-values) and blue indicates lower significance. For each term, the number on the right denotes the count of enriched genes, and the value in parentheses is the corresponding −log₁₀(q-value)

Dynamic expression profiles of ShAP2 genes across developmental tissues and under ABA stress

Expression profiling revealed diverse and specific patterns for ShAP2 gene family members across sugarcane tissues and developmental stages. Several genes, including ShAP2-60.1, ShAP2-1, ShAP2-13.1, and ShAP2-13.2, were predominantly expressed in seedling leaves but showed minimal expression in stem tissues, linking them to leaf development (Fig. 6A). Conversely, members such as ShAP2-56.1/2/3/4, ShAP2-55, and ShAP2-27.1/2/3 exhibited high expression in stems, with peaks in mature stem samples, suggesting specialized roles in stem development and maturation. Subtypes of the same gene model (e.g., ShAP2-56.1/2/3/4) typically shared similar expression trends, indicative of functional redundancy orsynergy. Notable exceptions, such as the divergent expression of ShAP2-27.4 compared to its paralogs (ShAP2-27.1/2/3), hinted at potential subfunctionalization. To assess whether promoter divergence underlies this expression diversification, we compared cis-elements in the five ShAP2-27 paralogs using PlantCARE (Table S5). Clear differences in cis-regulatory element composition were evident. ShAP2-27.1 and ShAP2-27.3 were enriched in hormone-responsive (ABRE, CGTCA/TGACG, TATC-box) and stress-related (LTR, ARE) elements. ShAP2-27.2 and ShAP2-27.4 contained multiple MeJA-responsive motifs (CGTCA/TGACG). The GCN4 motif (endosperm expression) was present only in ShAP2-27.1, ShAP2-27.2, and ShAP2-27.3. These differences provide a likely explanation for the expression divergence among ShAP2-27 paralogs and support subfunctionalization after gene duplication. ABA treatment elicited strikingly tissue-specific responses (Fig. 6B). In leaves, expression of the ShAP2-56 model (particularly ShAP2-56.1 and ShAP2-56.3) and ShAP2-1 peaked sharply at 48 h. ShAP2-56.1 expression increased approximately two-fold, while ShAP2-1 was strongly induced from a very low basal level, demonstrating classic positive regulation by ABA. In contrast, ShAP2-60 displayed a unique biphasic response in leaves, being transiently suppressed at 24 h before rebounding to higher-than-control levels by 48 h. Genes with strong induction in leaves often showed weak or suppressed responses in stems. For instance, ShAP2-56.1 expression declined in stems following 24 h of ABA treatment. Some genes, like ShAP2-60, exhibited complex spatiotemporal patterns—upregulated in leaves at 48 h and in stems at 96 h, but downregulated at other timepoints—revealing intricate roles within ABA signaling networks. Cross-analysis of tissue-specific expression and ABA response profiles revealed a systematic relationship between organ identity and stress signaling: leaf-preferential genes exhibited stronger ABA induction, whereas stem-preferential genes showed attenuated responses, indicating that leaf-preferential AP2 members are preferentially integrated into ABA signaling networks. Collectively, these patterns indicate the ShAP2-1 gene model and ShAP2-56 as putative regulators of the leaf stress response. The widespread divergence in expression between stems and leaves underscores organ-specific adaptation strategies, wherein the AP2 family contributes to tailored physiological responses to abiotic stress.

Fig. 6.

Comprehensive expression analysis of ShAP2 gene across sugarcane tissues and in response to ABA treatment. A Tissue-specific expression patterns of ShAP2 genes. B Dynamic expression of ShAP2 genes in cultivar XTT22 following ABA treatment. C Temporal expression profiles of eight selected ShAP2 gene models in cultivar YZ141313 after ABA treatment. For each gene model, the plotted expression value represents the aggregate signal from all its paralogous members, quantified using a common primer pair designed against conserved regions. Values are mean ± SD (n = 3 biological replicates). Different lowercase letters indicated significant differences among mean values (one-way ANOVA with Ducan’s multiple range test; P < 0.05)

To further resolve the spatiotemporal dynamics of ABA responsiveness, we performed qPCR analysis on two distinct leaf positions—the younger leaf3 and the more mature leaf7—comparing ABA-treated samples with controls (CK) (Fig. 6C). Eight candidate ShAP2 genes were selected based on a complementary set of criteria derived from the public RNA-seq analysis.These genes were selected based on constitutive expression, with FPKM > 1 in at least 80% of the sampled tissues or conditions. From this set, we further selected representatives exhibiting distinct expression profiles—including leaf-preferential and stem-specific patterns—as well as varied responsiveness to ABA treatment. It should be noted that the qPCR primers were designed to target the gene models (e.g., ShAP2‑56), each of which consolidates multiple paralogous copies with high sequence identity; therefore, the detected signals reflect the combined expression of the underlying paralogous subtypes (e.g., ShAP2‑56.1, ShAP2‑56.2, etc.). The results revealed clear differences in both the timing and tissue specificity of the response. During the early response phase (1 day post-treatment), ABA induction was significantly more pronounced in the younger leaf3. In this tissue, ShAP2-1, ShAP2-55, ShAP2-56, and ShAP2-59 each showed approximately 3-fold upregulation compared to CK. The response of ShAP2-60 was particularly striking, with an 8-fold induction. In contrast, leaf7 exhibited attenuated responses: ShAP2-55 and ShAP2-56 were upregulated about 2-fold, ShAP2-1 only 1.5-fold, while ShAP2-59 and ShAP2-60 showed no significant change. An important baseline difference was also evident: constitutive expression levels in the untreated control were substantially higher in the mature leaf7 than in leaf3, suggesting that these ShAP2 gene models undergo developmental upregulation during leaf maturation. By 7 days post-treatment, the ABA-responsive expression differences between treated and control plants had diminished for most genes and were no longer statistically significant. This developmental modulation of ABA responsiveness—stronger inducible responses in younger leaves versus higher constitutive expression in mature leaves—suggests a strategic shift from an inducible stress response strategy in developing tissues to a primed state in mature, photosynthetically critical leaves. In summary, the ABA response of these ShAP2 gene models is both transient, being strongest in the early phase, and spatially modulated, with a preferential induction in younger leaf tissues. Concurrently, their baseline expression increases during leaf development, pointing to a complex interplay between developmental programming and inducible stress signaling.

Effects of ShAP2-1 overexpression on root growth and drought tolerance in rice

In this study, the selection of ShAP2-1 for functional characterization was based on its strong ABA inducibility in leaves, as root transcriptomic data were not available during the initial screening phase. Although roots are the primary sensors of soil water deficit, leaf-expressed ABA-responsive genes are well positioned to mediate systemic stress signals and coordinate whole-plant acclimation. Notably, our previous work indicated that ShAP2-1, a nuclear-localized transcription factor as confirmed by GFP fusion assays, was also highly transcribed in root tissues, supporting its potential direct role in drought stress responses [33]. To validate the function of sugarcane ShAP2-1 in a heterologous system, we overexpressed the gene (cloned from cultivar YZ141313) in transgenic rice. Analysis of root architecture immediately revealed a clear phenotype: two independent overexpression lines (ShAP2-1-OE5 and ShAP2-1-OE7) developed significantly longer primary roots than wild-type (WT) plants (Fig. 7A). Quantification showed that primary root length in these transgenic lines increased by approximately 2.4-fold compared to WT (P < 0.0001; Fig. 7B). This establishes ShAP2-1 as a positive regulator of primary root growth in rice. Under osmotic stress conditions (200 mM Mannitol), ShAP2-1 overexpression lines (ShAP2-1-OE5 and ShAP2-1-OE7) developed a more extensive root system than wild-type plants, with increased lateral root number and longer primary roots (Fig. 7C). Quantitative measurements confirmed that primary root length in both transgenic lines remained significantly greater than that of wild-type plants under drought stress (Fig. 7D), demonstrating that ShAP2-1 overexpression promotes primary root elongation even under water-limited conditions. These root architectural advantages directly enhance the plant’s capacity for water acquisition during drought stress, providing a morphological basis for the enhanced drought tolerance observed in transgenic plants. Under well-watered conditions, the shoot of the transgenic and WT plants were morphologically indistinguishable (Fig. 7E). Drought stress, however, unmasked a profound difference in resilience. After five days of water withholding, WT plants showed severe wilting and chlorosis, whereas the overexpression lines remained turgid and green. Following a six-day drought and a five-day recovery period, nearly 80% of transgenic plants survived, compared to fewer than 20% of WT plants (Fig. 7F). These data confirm that ShAP2-1 overexpression confers strong drought tolerance. To understand the physiological basis of drought tolerance, we measured key stress indicators. After five days of drought, transgenic plants maintained a significantly higher leaf relative water content (RWC) than wild-type plants (Fig. 7G), demonstrating their superior capacity to retain water. This was accompanied by a clear reduction in oxidative damage, as levels of the lipid peroxidation marker malondialdehyde (MDA) were approximately 2.1-fold lower in the overexpression lines (Fig. 7H). Analysis of the antioxidant system revealed a nuanced response. While superoxide dismutase (SOD) activity was enhanced by over 40% in transgenic plants, peroxidase (POD) activity showed a concurrent 2.6-fold decrease relative to the wild type under drought stress (Fig. 7I, J). This distinct activity profile suggests that ShAP2-1 overexpression alters the antioxidant response, leading to an overall reduction in oxidative stress, as evidenced by the lower MDA levels.

Fig. 7.

Functional characterization of sugarcane ShAP2-1 in promoting root growth and drought tolerance in transgenic rice. A Representative phenotype of primary root growth in WT and ShAP2-1-overexpressing lines (ShAP2-1-OE5, ShAP2-1-OE7). Scale bar = 5 mm. B Quantitative analysis of primary root length. Data represent mean ± SD (n = 15). ****P < 0.0001 (one-way ANOVA with Duncan’s multiple range test). C Root phenotyping at 5 days under osmotic stress. Scale bar = 1 cm. D Quantitative analysis of primary root length under drought stress. Data represent mean ± SD (n = 15). ****P < 0.0001 (one-way ANOVA with Duncan’s multiple range test). E Drought tolerance performance evaluated at three stages: pre-drought conditions, after 5 days of drought stress, and following 6 days of drought with 5 days of rewatering. Scale bar = 2 cm. F Survival rates after drought recovery across genotypes. Data represent mean ± SD (n = 3 biological replicates, 10–15 plants per replicate). ****P < 0.0001. (G-J) Physiological parameters under drought stress: (G) leaf relative water content (RWC), (H) malondialdehyde (MDA) concentration, (I) superoxide dismutase (SOD) activity, and (J) peroxidase (POD) activity. Values are mean ± SD (n = 5 biological replicates). Statistical significance was determined by one-way ANOVA with Tukey’s multiple range test (ns: not significant, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001) between WT and transgenic lines at each time point

Discussion

Comprehensive analysis of the sugarcane AP2 family reveals lineage-specific expansion and evolution shaped by polyploidy

Our genome-wide analysis identified 211 non-redundant AP2 transcription factors in the commercial sugarcane cultivar XTT22. This repertoire represents a substantial expansion over those of its wild relatives, far exceeding the 43 members reported in wild S. spontaneum AP85-441 [34] and the 14 found in the diploid relative E. fulvus [35]. When normalized per chromosome, the AP2 gene density in sugarcane is approximately 2.1, which is comparable to rice (~ 2.0) [36] but markedly higher than the densities in S. spontaneum (1.3) and E. fulvus (1.4). This expansion pattern suggests that the AP2 family’s growth in modern sugarcane is not merely a byproduct of polyploidy but reflects distinct, lineage-specific evolutionary paths, likely shaped by the crop’s complex domestication and hybridization history.

Tracing the subgenome origins of these genes revealed a pronounced asymmetry: 156 AP2 genes (73.9%) were derived from the S. officinarum subgenome, compared to only 49 (23.2%) from the S. spontaneum subgenome. This bias aligns with the recognized transcriptional dominance of the S. officinarum subgenome in sugarcane hybrids [32], indicating that the cultivated AP2 gene pool is an enriched assemblage drawn from both progenitors. The close match between the number of S. spontaneum-derived AP2 genes (49) and the total in wild S. spontaneum (43) implies limited gene loss or innovation from this progenitor. In contrast, the S. officinarum lineage appears to have been the primary driver of the family’s expansion. Chromosomally, AP2 genes were distributed unevenly, with notable clustering on specific chromosomes (e.g., homologous group Chr1) and a complete absence on Chr8. Such a distribution is conducive to coordinated gene regulation and has been observed in other plant species [37, 38].

Gene duplication serves as a major engine for functional innovation [39]. Comparative studies of the AP2/ERF superfamily show that different species employ distinct expansion strategies; Arabidopsis relies more on segmental duplications, whereas rice expansion is fueled by tandem duplication [5, 40]. Notably, in rice and E. fulvus, tandem duplication is largely confined to the DREB and ERF subfamilies and is uncommon within the AP2 clade itself [5, 35]. Mirroring this pattern, the expansion of the AP2 family in cultivated sugarcane was overwhelmingly driven by whole-genome and segmental duplication events, which account for 193 genes (91.5%). This mode of expansion is a likely direct consequence of sugarcane’s polyploid origin and the extensive chromosomal reshuffling that characterizes its genome [41, 42]. Tandem duplication played a minimal role, contributing only two genes. Expansion via large-scale genomic duplication facilitates functional differentiation, as dispersed gene copies integrate into diverse genomic and regulatory landscapes. This process can promote the neofunctionalization or subfunctionalization of AP2 genes, enabling their recruitment into the complex regulatory networks that govern sugarcane’s development and stress responses [43, 44].

Analysis of evolutionary pressures indicated that approximately 4.59% of duplicated AP2 gene pairs experienced positive selection (Ka/Ks > 1), suggesting they may have undergone neofunctionalization events that contributed to adaptive evolution. The estimated divergence times for these pairs span a wide range (0.51 to 62.6 million years ago), with the most recent duplications (as recent as 0.51 Mya) potentially coinciding with sugarcane speciation and domestication. This timeline partially overlaps with but extends beyond estimates for E. fulvus [35]. Synteny analysis further highlighted the deep evolutionary conservation of this family, identifying 28 and 30 syntenic AP2 gene pairs between sugarcane and E. fulvus or rice, respectively. A core set of 26 sugarcane genes showed synteny with orthologs in both reference species. The frequent observation of multiple sugarcane AP2 genes corresponding to a single ortholog in related species is a classic signature of polyploidization. This genomic context provides fertile ground for subfunctionalization or neofunctionalization, thereby enhancing genetic diversity and adaptive potential [45–47].

Structural diversity suggests a broad functional repertoire for sugarcane AP2 transcription factors

Phylogenetic analysis of the sugarcane AP2 proteins resolved them into three conserved subfamilies: euAP2, euANT, and basalANT. The euANT subfamily was the most populous, a distribution that mirrors the subfamily composition in model plants like Arabidopsis and rice. This confirms that the fundamental tripartite architecture of the AP2 family is widely conserved across flowering plants [5, 48].

Substantial divergence in gene structure distinguished the subfamilies. The euAP2 clade showed remarkable exon number plasticity, ranging from 2 to 17 exons, while the basalANT subfamily was more structurally conserved (6–8 exons). Notably, this pattern inverts that observed in pecan, where euANT genes are uniform and basalANT genes are variable [49]. The high structural plasticity of sugarcane euAP2 genes likely serves as a substrate for functional innovation through alternative splicing, a key mechanism for generating regulatory complexity [50, 51]. The extreme case of ShAP2-29 (17 exons) underscores this potential. Compared to the often intron-poor DREB and ERF families [35], the multi-exonic structure of AP2 genes may reflect distinct evolutionary constraints; genes with numerous introns can evolve novel regulatory variants, while those with fewer are primed for rapid transcriptional responses [52]. Analysis of conserved motifs and domains provided further functional clues. A core sequential arrangement of motifs (motif2-4-7-5-3-1) was ubiquitous, pointing to fundamental structural requirements. Subfamily-specific motifs, such as motif8 in euAP2 and motif9 in euANT, signify clade-specific functional adaptations, consistent with reports in other species [53]. We also identified accessory domains beyond the core AP2 repeats. For instance, ShAP2-29 encodes a complete SSM4 domain, suggesting evolutionary innovation through domain shuffling that may integrate AP2 function with processes like ubiquitination.

Cis-regulatory element analysis of promoter regions revealed a complex regulatory landscape, predisposing these genes to multifunctional integration. Hormone-responsive elements were the most abundant category (30%), with a strong emphasis on ABA and MeJA pathways. This aligns with the well-established role of AP2 factors as hubs in hormonal crosstalk [6, 54] and is supported by complementary lines of evidence. The number of ABA-responsive elements (ABRE) varied substantially across ShAP2 promoters, ranging from 1 to 25 per gene with a mean of 8.3 (Fig. 4). This regulatory potential was functionally validated by qPCR assays demonstrating rapid induction of multiple ShAP2 members upon ABA treatment, including ShAP2-1/55/56/60 (Fig. 6C). Collectively, these findings position sugarcane ShAP2 genes within well-established ABA-dependent regulatory networks, where AP2/ERF transcription factors integrate hormonal signals to orchestrate stress adaptation programs [6, 54]. A significant enrichment of stress-responsive elements (20%) further corroborates their central role in abiotic stress adaptation, consistent with findings in S. spontaneum and other crops [33, 55]. Promoters were also rich in transcription factor binding sites (23%), including those for MYB and MYC, hinting at intricate higher-order regulatory networks. The presence of development-related elements (7%), including those associated with meristem and seed development as well as circadian rhythms, implicates these transcription factors in fundamental developmental processes, expanding their functional repertoire well beyond stress responses.

Sugarcane AP2 genes function by linking development to stress responses

Expression patterns provide essential clues to gene function. AP2 transcription factors are established regulators of diverse processes, including plant growth, hormone signaling, and adaptation to environmental stresses such as drought, salinity, and temperature extremes [56, 57]. Our transcriptome analysis in sugarcane revealed a complex spatiotemporal expression landscape for the AP2 family. Members displayed pronounced tissue specificity; for instance, ShAP2-1 and ShAP2-60 were strongly leaf-preferential, whereas members of the ShAP2-56 model (e.g., ShAP2-56.1/2/3/4) were specifically expressed in stems. This points to deep functional specialization in organ development, mirroring patterns observed in the desert legume Eremosparton songoricum [58] and aligning with organ-preferential expression in E. fulvus and Coptis chinensis [34, 59]. Conserved expression among duplicate genes suggests functional redundancy, while divergent expression between paralogs (e.g., ShAP2-27.4 vs. ShAP2-27.1/2/3) hints at subfunctionalization, a common evolutionary outcome in polyploids [46, 60].

Although the ERF and DREB subfamilies are recognized as primary mediators of abiotic stress responses, the roles of AP2 members under conditions like ABA or drought are less clearly defined. Nonetheless, growing evidence underscores their importance in ABA-mediated signaling. In Arabidopsis, for example, ANT was shown to negatively regulate salt tolerance by repressing the calcium sensor SOS3-LIKE CALCIUM BINDING PROTEIN 8 (SCABP8/CBL10), which is vital for ion homeostasis under salinity [61]. Complementary findings in other species support this functional connection. In S. spontaneum, three AP2 genes (SsAP2-9, SsAP2-21, and SsAP2-25) were induced by NaCl, PEG6000, and ABA treatments [34], while in E. fulvus, EfAP2-13 was significantly downregulated after ABA exposure [35]. Similarly, in soybean, overexpression of the AP2 gene Glyma01g39520 suppressed the drought-tolerant phenotype of an Arabidopsis mutant by reducing ABA sensitivity and downregulating key signaling components ABI3 and ABI5 [62]. Our findings align with and extend this emerging paradigm. We confirmed that several sugarcane AP2 genes, most notably ShAP2-1, are rapidly and transiently induced by ABA. Heterologous overexpression of ShAP2-1 (from cultivar YZ141313) in transgenic rice demonstrated its capacity to confer drought resilience through a dual-mechanism strategy. First, ShAP2-1 enhances drought tolerance by promoting primary root growth, leading to a ~ 2.4-fold increase in primary root length and thereby improving water foraging capacity. This root phenotype is consistent with previously reported roles of AP2/ERF transcription factors in root development. For instance, CRL5 in rice has been shown to promote crown root formation through repression of cytokinin signaling [26], demonstrating that AP2 family members can influence root architecture. These observations provide a plausible context for ShAP2-1 function and suggest that it may intersect with regulatory pathways involved in root meristem activity or cell elongation—a hypothesis that warrants further molecular investigation. Second, it bolsters physiological tolerance by maintaining higher leaf water content, reducing membrane lipid peroxidation (indicated by lower MDA levels), and elevating the activity of antioxidant enzymes such as SOD. This successful cross-species validation confirms the functional conservation of ShAP2-1 and highlights its potential as a candidate genetic module for further investigation in drought improvement. The favorable allele identified from the elite drought-tolerant cultivar YZ141313 could potentially be further explored as a candidate for marker-assisted selection, while its coding sequence or native regulatory elements can be harnessed through transgenic approaches or CRISPR-mediated genome editing to optimize expression patterns. These strategies offer a pathway to enhance drought resilience in elite sugarcane cultivars. To further evaluate the potential of ShAP2-1 as a target for molecular breeding, several key questions remain to be addressed. While our heterologous overexpression in rice provides valuable proof-of-concept, validation at the reproductive stage, under native regulatory control, and through loss-of-function approaches in sugarcane itself is essential. Given the technical challenges posed by sugarcane’s polyploid genome, future work should prioritize CRISPR/Cas9-based gene editing to generate knockout mutants or allelic series for ShAP2-1 and other strong candidates such as ShAP2-60. Parallel efforts to identify the direct downstream targets of these transcription factors will further elucidate the regulatory networks underlying drought adaptation. Together, these approaches will provide the foundational knowledge needed to deploy ShAP2 genes in breeding programs aimed at developing drought-resilient sugarcane varieties.

Materials and methods

Identification of AP2 family members in sugarcane

The sugarcane cultivar XTT22 is a high-sucrose, high-yielding, and drought-tolerant commercial variety widely cultivated in China, and was chosen for this study due to the availability of its high-quality, chromosome-level reference genome [32]. Genomic data for the XTT22 were sourced from the Sugarcane Genome Database (https://sugarcane.g.xu.edu.cn/scdb/download). To guide our identification, we retrieved the full set of AP2 family protein sequences from two model species, Arabidopsis and O. sativa (listed in Table S6), from Phytozome V13. These sequences served as queries for a BLASTP search against the XTT22 proteome, applying a stringent E-value cutoff of < 1 × 10⁻¹⁰. In parallel, we employed a Hidden Markov Model (HMM)-based approach. The HMM profile for the AP2 domain (PF00847.21) was obtained from the Pfam database and used to scan the XTT22 genome via the “Simple HMM Search” tool in TBtools, with an E-value threshold of < 0.01. The candidate gene sets from these complementary homology-based and HMM-based searches were merged. We then used TBtools [63] to remove redundant sequences, generating a preliminary non-redundant list. To ensure all candidates contained the defining AP2 domain, each sequence was validated using the NCBI Conserved Domain Database. We subsequently filtered this list to retain only those proteins encoding two AP2 domains. To finalize the gene repertoire, we first addressed allelic redundancy within the polyploid genome using the “Fasta Get Representative” function in TBtools, which selects the longest protein sequence for each unique genetic locus. We then integrated chromosomal localization data to define the final membership. Putative paralogs sharing ≥ 95% amino acid sequence identity were consolidated into distinct gene models. Within each model, specific subtypes (e.g., ShAP2-2.1, ShAP2-2.2) were designated based on their unique chromosomal coordinates and supported by phylogenetic relationships.

Physicochemical properties and phylogenetic analysis of AP2 family members

The physicochemical properties of the identified AP2 proteins—including amino acid length, molecular weight, and theoretical isoelectric point (pI)—were calculated using the “Protein Parameter Calc” module in TBtools. For phylogenetic analysis, protein sequences of AP2 members from sugarcane, Arabidopsis, and rice (Table S6) were aligned to construct an unrooted phylogenetic tree using MEGA‑X software. In addition, a maximum‑likelihood (ML) phylogenetic tree was generated specifically for the AP2 transcription factors identified in the single‑haplotype genome of XTT22. The ML analysis was performed under the following parameters: the p‑distance substitution model, pairwise deletion, Poisson correction, and 1,000 bootstrap replicates, providing detailed evolutionary resolution within the sugarcane lineage.

Analysis of gene structure, conserved protein motifs, and domains

The exon–intron architecture of each ShAP2 gene was determined and visualized using TBtools. To confirm the presence and completeness of the conserved AP2 domain (PF00847) within these proteins, all protein sequences were scanned against the NCBI Conserved Domain Database (CDD). For a finer-scale examination of conserved regions beyond the core DNA-binding domain, putative protein motifs were predicted with the online MEME Suite (https://meme-suite.org/meme/), setting the search parameters to identify up to 10 distinct motifs, each 6–50 amino acids in length, with zero or one occurrence allowed per sequence. The results from these structural, domain, and motif analyses were subsequently integrated and compiled into a unified graphical representation using TBtools.

Chromosomal distribution, gene duplication, and synteny analysis of ShAP2 genes

The chromosomal positions of the identified ShAP2 genes were extracted from the XTT22 genome annotation file (GFF3 format) and visualized using the “Gene Location Visualize” function in TBtools. To infer gene duplication events and elucidate evolutionary relationships, synteny analyses were conducted. Intraspecific (within sugarcane) and interspecific (between sugarcane, rice, and Erianthus fulvus) collinear gene pairs within the AP2 family were identified using MCScanX [64]. The resulting syntenic relationships were plotted with the “Dual Synteny Plot” module in TBtools. For duplicated ShAP2 gene pairs, the nonsynonymous (Ka) and synonymous (Ks) substitution rates were calculated using simple KaKs_Calculator in TBtools. Divergence times (T) were subsequently estimated with the formula: T = Ks/(2 × 6.1 × 10^− 9)×10^− 6 million years ago (Mya) [35].

Analysis of putative cis-regulatory elements (CREs)

Genomic sequences spanning 2,500 bp upstream of the transcription start site for each ShAP2 gene were extracted from the XTT22 genome. These promoter sequences were submitted to the PlantCARE database (http://bioinformatics.psb.ugent.be/webtools/plantcare/html) for in silico prediction of CRE elements. Based on the prediction results, the identified CREs were systematically classified, and their distribution was visualized using TBtools.

Analysis of expression patterns using transcriptomic data

To characterize the expression patterns of ShAP2 genes, publicly available RNA-seq data (FPKM values) for the XTT22 cultivar were obtained from the Sugarcane Genome Database. The dataset encompassed samples from multiple developmental tissues (including seedling stem and leaf, pre-mature stem internodes 3/6/10 and corresponding rolled/expanded leaves, mature stem internodes 3/9/15 and corresponding rolled/expanded leaves) and from a time-course ABA stress experiment. For the ABA experiment, samples were collected at 0, 24, 48, and 96 h post-treatment, with both mock-treated control (C) and ABA-treated replicates at each time point. To focus the analysis on actively expressed genes, we filtered the initial expression matrix by removing genes with an average FPKM below 0.9 across all samples. A heatmap visualizing the expression patterns of the remaining ShAP2 genes across tissues and ABA time points was then generated using TBtools.

Plant materials, ABA treatment, and qPCR analysis

For expression validation, we used ‘YZ141313’, a drought-tolerant commercial sugarcane cultivar. Plants were grown from single-bud stalk cuttings (2–3 buds per pot) in a greenhouse at the Sugarcane Research Institute (Kaiyuan City, Yunnan, China) under controlled conditions: daytime temperature 24–35 °C, nighttime 14–20 °C, relative humidity 35–65%, and light intensity 500–940 µmol m^− 2 s^− 1. At the 8–10 leaf stage, plants were foliar-sprayed with either 1 mM abscisic acid (ABA) or water (mock control) [34]. The third (Leaf3) and seventh (Leaf7) fully expanded leaves were harvested at 1 and 7 days post-treatment, immediately frozen in liquid nitrogen, and stored at − 80 °C. Three biological replicates were collected per treatment and time point.

Total RNA was extracted from these samples and reverse-transcribed into cDNA. We selected eight candidate ShAP2 genes based on prior transcriptomic data for expression validation by qPCR. Gene-specific primers were designed using Primer Premier 5 (Table S7). The qPCR reactions were performed using GAPDH as an internal reference. Each 20 µL reaction contained 3.0 µL cDNA, 0.4 µL each of forward and reverse primer (10 µM), and 10 µL of 2× ChamQ Universal SYBR qPCR Master Mix. The thermal profile comprised an initial denaturation at 95 °C for 2 min, followed by 40 cycles of 95 °C for 10 s and 60 °C for 30 s, concluding with a melting curve analysis (60–95 °C, ramp rate 0.15 °C s⁻¹). Gene expression was quantified using the 2^−ΔΔCt method [65]. All assays included three biological replicates, each with three technical replicates.

Heterologous expression of ShAP2-1 in rice and phenotypic assessment

The full-length coding sequence of ShAP2-1 was amplified from cultivar ‘YZ141313’ via 5′ RACE PCR, verified by sequencing, and cloned into the binary vector pBWA(V)HS-ccdB. The resulting overexpression construct was introduced into embryogenic calli of rice (Oryza sativa cv. Zhonghua 11) via Agrobacterium tumefaciens (strain EHA105)-mediated transformation. Primary transformants were selected on hygromycin medium and validated by PCR. Homozygous T₂ plants, identified through segregation analysis, were used for subsequent experiments. In addition, we confirmed transgene expression in the T₂ homozygous lines by qPCR prior to phenotypic analyses. For root phenotyping, surface-sterilized seeds of wild-type and transgenic lines were germinated on 1/2 MS medium. Primary root length was measured at 4 days post-germination. For root phenotyping under osmotic stress, seeds were germinated on 1/2 MS medium supplemented with 200 mM mannitol. Primary root length and lateral root number were recorded at 5 days after germination. For the drought tolerance assay, seedlings at the three-leaf stage were transplanted into soil and subjected to water withholding. Phenotypic responses were monitored daily. Leaf samples for physiological assays were collected at 1 and 5 days after drought initiation. At 5 days, the whole-plant fresh weight (FW) was recorded. Samples were then immersed in water at room temperature for 8 h to obtain the turgid weight (TW), followed by oven-drying (108 °C for 1 h, then 80 °C for 48 h) to determine the dry weight (DW). The Relative Water Content (RWC) was calculated using the formula: RWC (%) = [(FW − DW) / (TW − DW)] × 100. The analysis included five biological replicates, with 10–15 plants pooled per replicate.

The remaining plants underwent continued drought until day 6, followed by a 5-day re-watering period to assess recovery and survival rate. Physiological indices associated with drought stress were measured using commercial assay kits (Suzhou Mengxi Biomedical Technology Co., Ltd). MDA content and the activities of SOD and POD were determined according to the manufacturer’s protocols. Briefly, 0.1 g of leaf tissue collected at 1 and 5 days of drought stress was homogenized in 1 mL of extraction buffer on ice, centrifuged at 12,000 g for 10 min at 4 °C, and the supernatant was used for measurements. Six technical replicates were performed for each sample. The remaining plants were subjected to continued drought until day 6, after which they were re-watered. Phenotypic recovery was assessed 5 days post re-watering. The survival rate was calculated as follows: Survival Rate (%) = (Number of surviving plants / Total number of plants) × 100.

Authors’ contributions

Conceptualization: X.L.L., and F.S.L.; Data curation: Z.Y.L., X.H., and X.J.L.; Investigation: Z.Y. L., X.H., L.P.Z., and J.Y.L.; Methodology: X.H., and Z.Y. L.; Resources: L.P.Z., and J.Y.L.; Supervision: X.L.L. and F.S.L.; Validation: X.H., and X.J.L.; Writing original draft: Z.Y.L., and X.H.; Writing review & editing: Z.Y.L., X.L.L., and F.S.L.; Funding acquisition: X.L.L, and X.H. All authors have read and agreed to the published version of the manuscript.

Funding

This research was supported by the Project of State Key Laboratory of Tropical Crop Breeding (No. SKLTCB-YAAS-2025-08), the Yunnan Seed Industry Joint Laboratory (202205AR070001-13), the Yunnan Preliminary Fundamental Research Project of Yunnan Academy of Agricultural Sciences (2024KYZX-03), and Key Laboratory for Crop Production and Smart Agriculture of Yunnan Province (No. 2024ZHNY08).

Data availability

All data generated or analyzed during this study are included in this published article.

Declarations

Ethics approval and consent to participate

Not applicable. This manuscript does not involve researching about humans or animals.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.