In this study, three sweet potato cultivars; ‘Guangshu 79’ (GS79), ‘Guangshu 146’ (GS149) and ‘Shangshu 19’ (SS19) were grown in the National Sweet Potato Germplasm Nursery (Guangzhou), China. ‘GS 79’ and ‘GS 146’ were developed by Crops Research Institute, Guangdong Academy of Agricultural Sciences, whereas SS 19 was bred by the Shangqiu Academy of Agricultural and Forestry Sciences. Among these cultivars, GS 146 (Maternal Parent) and GS 79 (Pollen parent) constituted a cross-compatible combination, while GS 146 (Maternal Parent) and SS 19 (Pollen parent) represented a cross-incompatible combination.

The coding sequences (CDS), protein sequences, whole genome sequences and GFF3 annotation files of Ipomoea batatas were obtained from the ‘Taizhong 6’ genome database (https://sweetpotao.com/download_genome.html) (Yang et al., 2017). The hidden Markov model (HMM) profile of EXO70 (PF03081) was downloaded from the Pfam database (http://pfam-legacy.xfam.org/) and used to identify candidate EXO70 genes in the sweet potato genome using HMMER v3.3.3 software (Mistry et al., 2021). Redundant sequences were removed, and candidate genes were further validated using the CDD-research (https://www.ncbi.nlm.nih.gov/Structure/bwrpsb/bwrpsb.cgi) to confirm the presence of the conserved EXO70 domain, resulting in the final identification of IbEXO70 family members. The CDS, protein sequences, whole genome sequences and GFF3 annotation files of the diploid Ipomoea trifida and Ipomoea triloba were downloaded from Sweet Potato Genomics Resource (http://sweetpotato.uga.edu/gt4sp_download.shtml) (Wu et al., 2018; Li et al., 2022; Du et al., 2023). EXO70 gene family members in diploid genome were identified using the same pipeline. The EXO70 protein sequences of Arabidopsis thaliana were obtained from TAIR (https://www.Arabidopsis.org/) while those of Oryza sativa and Brassica oleracea were retrieved from Phytozome13 (https://phytozome-next.jgi.doe.gov/).

Multiple sequence alignment of EXO70 proteins from I. batatas, I. trifida, I. triloba, A.thaliana and O. sativa was performed using the Clustalw algorithm implemented in MEGA v6.0 (Thompson et al., 2002; Larkin et al., 2007; Tamura et al., 2013). Phylogenetic trees were constructed using the neighbor-joining (NJ) method with the Poisson correction model and branch support was evaluated with 1000 bootstrap replicates (Yang et al., 2017). The phylogenetic trees were visualized and optimized using ChiPlot (https://www.chiplot.online/).

The physicochemical properties of the EXO70 proteins from I. batatas, I. trifida, and I. triloba, including molecular weight and isoelectric point, were analyzed using ExPASy ProtParam (https://web.expasy.org/protparam/) (Wilkins et al., 1999). The subcellular localization of EXO70 proteins was predicted using WoLF PSORT (https://wolfpsort.hgc.jp/). Chromosomal distribution, gene structure (exon-intron organization), and visualization were performed using TBtools (Chen et al., 2020). Conserved protein motifs were identified using the MEME online suite and subsequently visualized with TBtools. The 2000 bp upstream promoter sequences of IbEXO70 genes were extracted using TBtools, and cis-acting regulatory elements were predicted using PlantCARE (https://bioinformatics.psb.ugent.be/webtools/plantcare/html/) (Lescot et al., 2002; Chen et al., 2020). The distribution of cis-acting elements was visualized in TBtools.

To investigate the gene duplication events, MCScanX was used to identify tandem and segmental duplication events among EXO70 genes in I. batatas, I. trifida, I. triloba, A. thaliana, and O. sativa (Wang et al., 2012). Syntenic relationships and collinearity of EXO70 genes across these species were visualized using TBtools.

To investigate the expression profiles of IbEXO70 genes in different tissues, a total of 16 tissues types were collected from the grafted GS 146 plants, including petal, budding flower period, buds period, receptacle, stigma, pollen, the compatible pollinated stigma (GS 146 × GS 79), incompatible pollinated stigma (GS 146 × SS 19), as well as the functional leaves, newly expanded leaves, petioles and stems from both flowering and non-flowering plants.

Total RNA was extracted using the RNAprep Pure Polysaccharide Polyphenol Plant Total RNA Kit (TIANGEN, Beijing, China), and cDNA was synthesized with the HiScriptIII 1st Strand cDNA Synthesis Kit (TIANGEN, Beijing, China). Gene-specific primers were designed (Supplementary Table S1), and quantitative real-time PCR (qRT-PCR) was performed on a Bio-Rad fluorescence quantitative PCR (Bio-Rad CFX96,Bio-Rad Laboratories, Inc) system. Each sample was analyzed with three biological replicates, and the relative expression levels were calculated using the 2^-ΔΔCt method (Livak and Schmittgen, 2001).

The full length coding sequence of IbEXO70–26 was amplified and cloned into pCAMBIA1300-GFP vector. The recombinant construct was then introduced into Agrobacterium tumefaciens strain GV3101 and subsequently infiltrated into fully expanded leaves of healthy Nicotiana benthamiana plants. After infiltration, plants were incubated under low light conditions for 2 days. Infiltrated leaf sections (approximately 1 cm²) were then treated with 0.8 M mannitol for 1–2 h before microscopic observation. GFP fluorescence signals were examined using a Zeiss LSM880 laser scanning confocal microscope (Jena, Germany). Leaves infiltrated with the empty pCAMBIA1300-GFP vector were used as controls. Three independent biological replicates were performed.

All date were expressed as mean values calculated using WPS Office Excel. Statistical analyses were conducted using SPSS v22.0 software. One-way Analysis of Variance (ANOVA) was performed, and mean comparisons were conducted using the least significant difference (LSD) test at the P < 0.05 level. Differences among treatments were indicated by different lowercase letters, with letters from a to z representing descending significance level.

A total of 94 EXO70 genes were identified across three Ipomoea species, including 35 genes in the hexaploid Ipomoea batatas, and 30 and 29 genes in the diploid relatives Ipomoea trifida and Ipomoea triloba, respectively. The comparable numbers of EXO70 genes among cultivated and wild sweet potato species indicate a high degree of conservation of the EXO70 gene family in Ipomoea, with no extensive gene expansion following polyploidization. Chromosomal distributions indicated that EXO70 genes were unevenly distributed across chromosomes (Figure 1). Interestingly, the distribution patterns of ItfEXO70 and ItbEXO70 genes were highly similar, with identical gene numbers and largely conserved chromosomal positions. Minor differences were observed, two additional ItfEXO70 genes on chromosome 7, and the absence of one ItfEXO70 gene on chromosome 4 (Figures 1B, C). In both diploid species, EXO70 genes were mainly concentrated on chromosomes 2, 4, 5, and 7, with each chromosome harboring 3 to 6 genes.

In contrast, the IbEXO70 genes in hexaploid sweet potato were distributed across all 15 chromosomes except chromosome 3 and were predominantly located on chromosomes 1, 2, 12, 13, and 14. These chromosomes contained 4, 7, 6, 3 and 3 IbEXO70 genes, accounting for 11.42%, 20%, 17.14%, 8.57%, and 8.57% respectively (Figure 1A). The distinct chromosomal distribution patterns observed between hexaploid, and diploid Ipomoea species suggest potential divergence in gene regulation or functional specialization. Based on their chromosomal positions, EXO70 genes were subsequently named IbEXO70–1 to 35, ItfEXO70–1 to 30 and ItbEXO70–1 to 29.

Analysis of physicochemical properties of IbEXO70 proteins showed substantial variation in protein length and molecular characteristics (Table 1).IbEXO70–1 was the shortest protein (406 amino acids), whereas IbEXO70-17 (1163 amino acids) was the longest protein. The remaining members ranged between 438 and 933 amino acids. The predicted molecular weights ranged from 45.32 to 128.65 kDa, and the theoretical isoelectric points (pI) varied from 4.74 to 9.16. Subcellular localization prediction revealed that most IbEXO70 proteins were localized to the chloroplasts and the cytoplasm, except for IbEXO70–14 and IbEXO70-32, which were localized to the nucleus.

For diploid species, the amino acid lengths of ItfEXO70 proteins ranged from 449 to 808 (aa), with molecular weights of 50.11-92.42 kDa, and pI values of 4.61-9.89. The amino acid length of ItbEXO70 proteins ranged from 478 to 786 (aa), with molecular weights between 53.37 to 88.63 kDa, and the pI values ranging from 4.55 to 8.97. The results of subcellular localization prediction showed that both ItfEXO70 and ItbEXO70 proteins were predominantly localized to chloroplasts and the nucleus (Supplementary Tables S2, S3).

To elucidate the evolutionary relationships of EXO70 genes, a phylogenetic tree was constructed using EXO70 protein sequences from I. batatas (35), I. trifida (30), I. triloba (29), A. thaliana (23) and O. sativa (41) (Figure 2A). The phylogenetic analysis classified all EXO70 genes into three major clades, designated EXO70.1, EXO70.2, and EXO70.3, which were further subdivided into nine subgroups (EXO70A-EXO70I) (Figure 2B).

The EXO70.1 clade consisted solely of EXO70A subgroup (including 29 genes, accounting for 18.35% of the total EXO70 genes analyzed). The EXO70.2 clade contained six subgroups (EXO70B-EXO70F and EXO70H) containing 97 genes (61.40%), whereas the EXO70.3 clade included the remaining two subgroups, EXO70G and EXO70I, with a total of 32 genes (20.25%) (Figure 2C). Consistent with previous reports, the EXO70H subgroup showed evidence of rapid expansion in dicotyledonous plants (Sekereš et al., 2017) and contained the largest number of EXO70 genes, followed by EXO70A, while EXO70F contained the fewest members.

The distribution of EXO70 genes among subgroups was highly conserved across dicotyledonous species, particularly within Ipomoea. In the EXO70B-F subgroups gene numbers were nearly Identical among Ipomoea species (Figure 2D). Specifically, the EXO70B and EXO70F each contained one gene, EXO70D and EXO70E contained three genes, and EXO70C contained five genes. This conserved pattern indicates strong evolutionary stability of the EXO70 gene family in Ipomoea. In contrast the monocotyledonous species O. sativa harbored a higher number of EXO70 genes in the EXO70I subgroup. In previous studies, the EXO70I subgroup was commonly found in monocotyledonous plants (Chong et al., 2010). In this study, the absence of the EXO70I subfamily in dicotyledons indicates that EXO70 genes indeed exhibit differences between dicotyledons and monocotyledons.

Based on the evolutionary relationships of EXO70 genes in Ipomoea species, the conserved protein motifs and gene structures of EXO70 family members were systematically analyzed (Figure 3). Motif analysis identified a total of ten conserved motifs among EXO70 proteins. These motifs are widely distributed across most EXO70 members of the hexaploid species I. batatas, and diploid species I. trifida and I. triloba. Members of the EXO70 gene family in sweet potato species all contain the conserved EXO70 domain. Analysis of exon/intron organization revealed substantial structural diversity among EXO70 genes. Genes with relatively high numbers of exons and introns were predominantly clustered within the EXO70A subgroup. For instance, IbEXO70–15 contained 13 exons, while ItfEXO70–14 and ItbEXO70–15 each harbored 12 exons. In contrast, EXO70 genes belonging to the other subgroups generally contained possessed only 1 to 3 exons. Notable structural variations were also observed among genes within the same subgroup. For example, within the EXO70H subgroup, IbEXO70–17 contained 8 exons, whereas ItfEXO70–10 and ItbEXO70–11 contained only 1 exon. Similarly, in the EXO70E subgroup, IbEXO70–13 contained 5 exons, while all other members of this subgroup possessed only one exon. These results indicate that EXO70 genes within the same subgroup exhibit considerable divergence in gene structures, which may reflect functional diversification or linear specific evolutionary adaptation.

To investigate the potential functions of the IbEXO70 gene and its transcriptional regulatory mechanisms, a 2kb promoter region upstream of each IbEXO70 gene was extracted. The PlantCARE online tool was then employed to analyze potential cis-regulatory elements within this region (Figure 4). A total of 30 cis-regulatory elements were identified in the promoter regions of 35 IbEXO70 genes and classified into five major categories: hormone responsive, light responsive elements, plant growth and development related elements, abiotic and biotic responsive elements, and transcription factor binding sites (Figure 4B). Hormone responsive elements were the most abundant category, comprising nine types: ABRE, AuxRR-core, TGA-element, CGTCA-motif, TGACG-motif, GARE-motif, P-box, TATC-box, and TCA-element, which were associated with responses to abscisic acid, auxin, methyl Jasmonate (MeJA), gibberellin, salicylic acid. Except for the IbEXO70-6, all IbEXO70 genes contained at least one plant hormone responsive element in their promoter regions. Among these, ABRE elements involved in abscisic acid signaling were the most prevalent, suggesting a potential role of IbEXO70 genes in abscisic acid mediated regulatory pathways.

Additionally, multiple light responsive elements were detected, including the 3-AF1 binding site, ACE, G-box, GT1-motif, MRE, and Sp1, suggesting possible involvement of IbEXO70 genes in light regulated biological processes. Seven cis-elements related to plant growth and development were also identified, including CAT-box, circadian, GCN4_motif, motif I, MSA-like, O2-site, and RY-element, which are implicated in meristem expression, circadian rhythm, endosperm expression, root specific regulation, cell cycle control, zein metabolism, and seed specific regulation.

Promoter regions further contained elements associated with stress responses, such as ARE, GC-motif, LTR, and TC-rich repeats, which are involved in anaerobic induction, hypoxia, low temperature stress response, and defense related mechanisms. The presence of these elements suggests that IbEXO70 genes may participate in abiotic and biotic stress responses. Furthermore, transcription factor binding sites including MBS, MBSI, MRE, and CCAAT-box were identified, which are known to be involved in drought responsiveness and light signaling. Collectively, these results indicate that IbEXO70 genes are likely regulated by diverse environmental and developmental cues and may participate in a broad range of biological processes.

A collinearity analysis of the EXO70 gene family in Ipomoea batatas was performed to investigate their evolutionary relationships and expansion patterns. Our analysis identified both tandem and segmental duplication events. Specifically, two tandem duplication pairs were detected, IbEXO70-19/IbEXO70–20 and IbEXO70-32/IbEXO70-33. Tandem duplications typically result in the adjacent replication of genes on the same chromosome, which can lead to increased gene dosage or functional divergence. While these tandem duplications contribute to the expansion of the EXO70 gene family, their role is relatively minor compared to segmental duplications. A total of eight segmental duplication pairs were identified, including IbEXO70-31/IbEXO70-18, IbEXO70-13/IbEXO70-16, IbEXO70-1/IbEXO70-15, IbEXO70-33/IbEXO70-15, IbEXO70-34/IbEXO70-15, IbEXO70-14/IbEXO70-16, IbEXO70-3/IbEXO70-12, and IbEXO70-35/IbEXO70-12 (Figure 5). Segmental duplications typically arise from large scale chromosomal rearrangements and play a more dominant role in the expansion and diversification of gene families. These results underscore that segmental duplications have been the primary mechanism driving the evolution of the EXO70 gene family in I. batatas, while tandem duplications have contributed to a lesser extent. The collinearity analysis was further visualized in a circular plot, which highlighted the duplications and gene locations across the chromosomes, providing a clear view of the evolutionary dynamics of gene family. These findings contribute valuable insights into the expansion mechanisms of the EXO70 gene family and its functional diversification in I. batatas.

To further elucidate the evolutionary history and gene family expansion of EXO70 genes in sweet potato, a comparative synteny analysis was performed across multiple species, including I. batatas, I. trifida, I. triloba, A. thaliana, and O. sativa (Figure 6). Overall, the analysis revealed extensive genomic conservation within the Ipomoea genus. Specifically, we identified 46 orthologous EXO70 gene pairs were identified between I. batatas and I. trifida, and 40 orthologous pairs between I. batatas and I. triloba. This high degree of synteny reflects the shared evolutionary ancestry and conserved genomic architecture among these closely related species. Notably, several of these orthologous pairs corresponded to duplicated IbEXO70 gene pairs (Figure 6) suggesting that segmental duplications have contributed to the expansion of this gene family in sweet potato. In contrast, comparisons with more distantly related species showed a reduced number of homologous EXO70 pairs. Between I. batatas and the dicotyledonous species A. thaliana we identified 22 homologous gene pairs, while only eight homologous gene pairs were identified between I. batatas and the monocotyledonous species O. sativa. These results suggest that the EXO70 gene family sweet potato shares higher evolutionary conservation with dicotyledonous species than with monocotyledonous plants. Our analyses suggest that the expansion and diversification of EXO70 genes in sweet potato likely occurred after its divergence from monocots but before its speciation within the Ipomoea genus. The retained syntenic blocks and duplicated gene pairs highlight the role of whole genome or segmental duplication events in shaping the EXO70 gene repertoire in sweet potato, potentially contributing to its adaptive evolution and functional specialization.

To investigate the potential biological roles of IbEXO70 genes, the expression patterns of all 35 IbEXO70 family members were examined in various tissues of I. batatas using quantitative real-time PCR (qRT-PCR) (Figure 7A). Overall, most IbEXO70 genes exhibited relatively low expression levels budding flower period, buds period, petioles of flowering plants, stems of non-flowering plants, and petioles of non-flowering plants. In contrast, elevated expression of at least one IbEXO70 gene was detected in each of the remaining tissues examined. Notably, peak expression was observed for 7 IbEXO70 genes in pollen, 8 genes in stigmas subjected to compatible pollination (GS 146 × GS 79), and 9 genes in stigmas subjected to incompatible pollination (GS 146 × SS 19). These tissue and pollination specific expression patterns suggest that IbEXO70 genes are actively involved in reproductive processes, particularly in pollen function and stigma responses during pollination, and may play important roles in regulating pollen-stigma interactions in sweet potato.

To further investigate the involvement of IbEXO70 genes in the regulation of sweet potato ICI regulation, we focused on genes that were significantly upregulated in compatibly pollinated stigmas compared with incompatible or unpollinated stigmas. Based on qRT-PCR results, eight IbEXO70 genes: IbEXO70-7, IbEXO70-13, IbEXO70-24, IbEXO70-26, IbEXO70-27, IbEXO70-28, IbEXO70-30, and IbEXO70–34 were identified as candidate genes potentially involved in ICI regulation (Figure 7B). Among these candidate genes, IbEXO70-13, IbEXO70-30, and IbEXO70–34 exhibited relatively low expression levels and were ubiquitously expressed across all 16 tissues examined, suggesting more generalized cellular functions. In contrast, IbEXO70–26 and IbEXO70–27 showed the highest expression levels and displayed pronounced tissue specificity, with predominant expression in post pollination stigmas, non-pollination stigmas, pollen, and receptacle tissues, but minimal expression in other examined tissues. This distinct expression pattern indicates that IbEXO70–26 and IbEXO70–27 are likely to play specialized roles in reproductive development and pollen-stigma interactions, making them strong candidates for involvement in sweet potato ICI.

In a previous transcriptomic analysis of sweet potato intraspecific cross-incompatibility (ICI), six IbEXO70 genes were identified using four types of stigma samples: CK (unpollinated stigma), MT (stigma treated by a cross-incompatibility overcoming reagent), FT (incompatible pollinated stigma) and MFT (compatible pollinated stigma) (Yang et al., 2022). These genes included IbEXO70-4, IbEXO70-12, IbEXO70-15, IbEXO70-16, IbEXO70-17, IbEXO70-26 (Figure 8). Among these genes, all exhibited differential expression in MFT compared with CK, except IbEXO70-16. However, IbEXO70–26 was the only gene that showed significant differential expression in MFT compared with FT, whereas no significant expression differences were detected in FT vs CK and MT vs CK. This unique expression pattern suggests that IbEXO70–26 is specifically associated with compatible pollination rather than general stigma activation or stress responses. Therefore, these transcriptomic data further support IbEXO70–26 as a key candidate gene involved in the regulation of sweet potato intraspecific cross-incompatibility.

To investigate the potential function of IbEXO70-26, its full length coding sequence was cloned and fused to green fluorescent protein (GFP) and subjected to subcellular localization analysis. The bEXO70-26–GFP fusion construct was transiently expressed in Nicotiana benthamiana leaves, with free GFP used as a control. As expected, the free GFP control showed fluorescence signals in the cytoplasm, cell membrane, and nucleus. In contrast, the fluorescence signals from the IbEXO70-26-GFP fusion protein were exclusively detected in the nucleus (Figure 9), indicating that IbEXO70–26 is predominantly localized to the nucleus.