The experimental materials—I. pes-caprae, ‘ZiShu8’, Nicotiana tabacum, and Arabidopsis thaliana—were cultivated at the Xuzhou SweetPotato Research Center, China. Nicotiana and Arabidopsis were grown at 20–28 °C under a 16 h/8 h dark/light cycle with 50–60% humidity and 500 µmol·m⁻²·s⁻¹ light intensity. Ipomoea pes-caprae was cultivated in Hoagland nutrient solution. After 3–5 weeks of growth, 7–9 functional leaves and 8–12 cm roots were present, and I. pes-caprae plants were transferred to soil mixed with fine sand and a nutritional substance (3:1). Then, after 4–5 weeks, we selected plants with similar growth characteristics for experiments.

The amount of expressed genes was obtained from transcriptome data (Liu et al., 2020). Combined with the level of I. pes-caprae NAC differentially expressed genes under salt stress, 12 TFs were preliminarily screened as research targets and named IpNAC1–12. The IpNAC nucleotide sequence was converted into an amino acid sequence by NCBI ¹ . The physicochemical properties of the IpNAC1–12 proteins, including the theoretical molecular weight (MW), isoelectric point (pI), and hydrophobicity, were analyzed using the Expasy ProtParam tool ² . The full-length amino acid sequences of the IpNAC and ANAC proteins were used to generate a phylogenetic tree by deleting gaps and blanks. The unrooted neighbor-joining tree was constructed after ClustalW alignment using MEGA (Tamura et al., 2011). Bootstrap analysis was performed with 100 replicates to assess the level of statistical support for each tree node (Hu et al., 2015). EvoView was used for subsequent beautification.

The MEME ³ online tool was used to identify the protein domains of the IpNAC, with the following parameters: site distribution = zero or one occurrence per sequence, largest number of motifs = 10, and optimum motif width of 6–50. Default settings were used for the other parameters. The results were presented by GSDS 2.0 ⁴ . The gene sequences of IpNACs were compared with the full-length sequences by SnapGene 4.36 to determine the exon–intron structure and were visualized using TBtools software. Collinearity between I. pes-caprae (IpNACs) and A. thaliana (TAIR) ⁵ (ANACs) was determined using MCScanX (Wang et al., 2013), and the figures were generated using Circos (Darzentas, 2010).

Most gene regulatory regions are located 2 kb upstream of the translation start site. The IpNAC promoter sequences were obtained from the I. pes-caprae genome database (unpublished). Then, the PlantCARE ⁶ online database was used to identify the cis-acting elements of these promoters.

We selected plants with similar growth rates for plant hormone and abiotic stress treatments. Untreated plants were used as controls. The leaves or adventitious roots were collected after 0, 3, 6, 12, 24, or 48 h of treatment. Heat treatment was performed by transferring I. pes-caprae to a 45 °C climate box, and the roots were submerged in 20% PEG 6000 for 48 h for dehydration treatment. The leaves and adventitious roots were harvested from the plants. The roots of I. pes-caprae were submerged in a salt solution at three concentrations of 300, 600, and 900 mM NaCl, and adventitious roots were harvested. For hormone treatment, adventitious roots were submerged in 0.1 mM abscisic acid (ABA), 2 mM methyI jasmonate (MeJA), or 2 mM salicylic acid (SA) for 0, 3, 6, 12, 24, or 48 h, and the adventitious roots were collected. The roots were collected from three separate plants at each time point of each treatment and combined to form one sample. All experiments were performed independently and in triplicate. The collected samples were immediately frozen in liquid nitrogen and stored at −80 °C.

Following the manufacturer’s instructions, the FastPure Plant Total RNA Isolation Kit (Vazyme, China) was used to isolate the total RNA in the leaves and roots. The quality of the RNA was determined using a NanoDrop2000c spectrophotometer (Thermo Fisher Scientific, US), and total RNA integrity was identified by electrophoresis on 1.0% agarose gel. The HiScript II 1st Strand cDNA Synthesis kit (Vazyme, China) was used to obtain cDNA from total RNA. Real-time quantitative PCR was carried out using SYBR green (TOYOBO, Japan) on an ABI QuantStudio 6 Flex (Thermo Fisher Scientific, US) with a final volume of 10 µL per reaction. Each reaction mixture contained 5 µL SYBR Green Realtime PCR Master Mix (TOYOBO, Japan), 0.4 µL forward primer (10 μM), 0.4 µL reverse primer (10 μM), 3.2 µL ddH₂O, and 1 µL cDNA template (diluted at 1:10). Specific primers for each gene are listed in Supplementary Table S3 . The cycling parameters were 95 °C for 1 min and 40 cycles of 95 °C for 15 s, 60 °C for 15 s, and 72 °C for 45 s. The Actin gene of I. pes-caprae was used as an internal control. Each measurement was performed with three biological replicates. Data were analyzed using the 2^-ΔΔCt method.

The coding DNA sequences of IpNAC1–12 were obtained from the transcriptome datasets of I. pes-caprae. We designed the specific primers ( Supplementary Table S2 ). The Phanta Max Super-Fidelity DNA Polymerase (Vazyme, China) (18 μL ddH₂O; 25 μL 2 × Phanta Max Buffer; dNTP Mix (10 mM each); forward primer (10 μM); reverse primer (10 μM); 1 μL Phanta Max Super-Fidelity DNA Polymerase; 1 μL cDNA) were used to amplify the IpNAC1–12 genes from the cDNA of I. pes-caprae. The PCR parameters were set as follows: initial denaturation at 95 °C for 3 min, 30 cycles of denaturation at 95 °C for 15 s, annealing at 58 °C for 30 s, extension at 72 °C for 1 min 30 s, and final extension at 72 °C for 10 min. The PCR products were subcloned into the pHB-GFP vector with a Hind III restriction site to form pHB-IpNAC1-12-GFP. pHB-IpNAC1-12-GFP and pHB (set as a control) were transformed into Arabidopsis. In addition, using pCAMBIA0390-DsRed as the backbone vector, the coding regions of IpNAC1–12 were inserted into the pCAMBIA0390-DsRed expression vector constructing pUBI.U4:: IpNAC1-12-CaMV35S::DsRed for Agrobacterium rhizogenes-mediated transformation. pCAMBIA0390-DsRed was used as a control. Each experiment was conducted using more than three biological replicates.

The 35S: GFP-IpNAC5/IpNAC8/IpNAC10/IpNAC12 plasmid were transformed into Agrobacterium tumefaciens strain GV3101 for plant transformation. Transient expression in tobacco leaves was determined according to a published method (Zhou et al., 2018). After transformation, tobacco was kept in the light room for 36 h before examination by fluorescence microscopy. Then, the tobacco leaf epidermis was peeled to make a temporary squash. After incubation with phosphate-buffered saline containing 40, 60-diamidino-2-phenylin dole (DAPI), the temporary squashes were observed under a Leica Fluorescence Microscope at 10 × 20 (Leica Microsystems, Wetzlar, GmbH). The image data were processed using Adobe Photoshop (Mountain View, CA, USA). All transient expression assays were repeated at least three times.

The recombinant vectors pHB-IpNAC5/IpNAC8/IpNAC10/IpNAC12-GFP were transferred into Arabidopsis through the Agrobacterium tumefaciens strain GV3101 using the floral dipping method. After transformation, the T3 (third generation) seeds were germinated on 1/2 MS medium with 50 mg L⁻¹ hygromycin for screening. Then, the WT and overexpressing transgenic seeds were cultivated on 1/2 MS medium with 180 mM NaCl at 22 °C under 16 h of daylight. After 10 d, the primary root length was measured and the growth state was observed.

The pUBI.U4:: IpNAC5/IpNAC8/IpNAC10/IpNAC12-CaMV35S::DsRed constructs were introduced into A. rhizogenes strain K599. ZiShu 8 with the same growth state was selected for Agrobacterium infection according to the published method (Yu et al., 2020). Subsequently, the transformed plants were transferred into the soil under a high-humidity (28 °C/65%) environment. After three weeks, the positive plants were selected through the portable fluorescent lamp instrument, immersed in hydroponic solution for 3 d, and then collected.

The WT and IpNAC5/8/10/12 transgenic sweetpotato were cultured in 150 mM NaCl solution for 24 h. First, 3–4 cm of intact root tips were collected and fixed in a buffer solution for 30 min. Then, Na ⁺ flux in the root elongation zone was recorded for 5 min. The root mature zone was measured for another 5 min. Each group had five replicates.

The data are presented as the mean ± standard deviation of the mean (SD), and statistically significant differences were assessed using Dunnett’s test at p < 0.05 as the point of minimal statistical significance in all analyses.

We identified 12 NAC genes in I. pes-caprae, named IpNAC1–12. We found that IpNAC1–12 had high homology with Arabidopsis, sweetpotato, and their wild species. IpNAC2 is one of the reported genes, and NAC1 was screened from the cDNA library of I. pes-caprae, but its function was not clear. We verified that IpNACs contained a complete ORF. Analysis of the physicochemical properties revealed amino acid lengths of IpNACs ranging from 242 to 344. The molecular weight varied from 27.58 to 39.21 KDa, and the pI ranged from 5.24 to 8.52. Except for IpNAC2, IpNAC3, and IpNAC6, IpNAC proteins were weakly acidic and were predicted to be located in the nucleus ( Supplementary Table S1 ). IpNAC proteins contain NAC family characteristic subdomains (A–E) ( Figure 1 ; Supplementary Figure S1 ).

To investigate the taxonomic and phylogenetic relationships of IpNACs, we used the model plant Arabidopsis to analyze the potential function of the IpNAC protein. Using MEGA-X software, we performed phylogenetic analysis of the IpNAC and ANAC proteins (without duplicated and erroneous sequences) to establish an unrooted phylogenetic tree ( Figure 2A ). The NAC proteins were classified into nine subfamilies. IpNACs were divided into ONAC22 (including IpNAC1, IpNAC3, IpNAC5, and IpNAC7), ATAF (including IpNAC2, IpNAC6, IpNAC8, IpNAC10, and IpNAC12), OsNAC7 (including IpNAC11), and NAM (including IpNAC4 and IpNAC9). To obtain more information about IpNAC genes, we identified syntenic relationships between IpNACs and ANACs and found six pairs of syntenic NAC genes between I. pes-caprae and Arabidopsis ( Figure 2B ). Interestingly, one IpNAC gene (IpNAC6) and one ANAC gene (ANAC072) were associated with three or two syntenic blocks, respectively. At the same time, other IpNAC genes have a close relationship with ANAC genes, indicating that their functions may be similar.

Motifs are key functional units. The IpNAC motif components were analyzed using MEME online software. All IpNAC proteins contained the NAC family characteristic subdomains (A–E), named motif1/2/3/4/5, located at positions 0–210. Except for IpNAC1/4/7/9, other protein sequences had more than five motifs, with a maximum of eight. The 12 IpNACs had three exons, separated by two introns ( Figure 1 ), forming a highly conserved gene structure. The untranslated region (UTR) lengths differed significantly, indicating a difference in the efficiency of IpNAC gene expression. TFs regulate gene function by binding cis-acting elements to the promoter. The PLACE Online website was used to analyze the 2 kb upstream region of the transcription start site of IpNACs, and a series of stress response elements were found ( Figure 3 ). A cis-acting element involved in drought-inducibility (MBIS) was found in three IpNAC genes, an enhancer-like element involved in anoxic specific inducibility cis-acting element (GC-motif) was detected in four IpNAC genes, and a defense and stress response cis-acting element (TC-rich) was detected in five IpNAC genes. Anaerobic induction cis-acting regulatory elements (AREs) were identified in 11 IpNAC genes. Interestingly, there were five IpNAC genes with low-temperature responsive cis-acting elements (LTRs). Many hormone-responsive elements were found in the IpNAC promoter sequence. GA responsive elements were found in three IpNAC genes. SA responsive elements were found in four IpNAC genes. IAA responsive elements were found in 11 IpNAC genes, and ABA responsive elements were found in 12 IpNAC genes. This indicates that plant hormones play a central role in the regulation of the salt stress response.

Ipomoea pes-caprae is a highly salt-tolerant crop, and the expression pattern of IpNACs was determined after treatment with different concentrations of NaCl solution. Among the 12 genes, there were significant differences ( Figure 4 ). Compared with 600 and 900 mM NaCl solutions, IpNAC genes treated with 300 mM NaCl solution had the highest response to salt induction. Interestingly, IpNAC10 showed the strongest significance after different salt solution treatments and maintained a high mRNA level (>2000-fold) at 6 and 24 h treatment periods. IpNAC11 was upregulated significantly, reaching the maximum change (47-fold) at 3 h, and returned to baseline over time. In contrast, IpNAC2 reached maximum expression at 48 h (6-fold), and the trend was downregulated and only slightly upregulated throughout the treatment period. IpNAC5, IpNAC8, and IpNAC12 expression were induced 19–300-fold at 6 h. The expressions of IpNAC3, IpNAC6, and IpNAC9 were increased 15–300-fold at 12 h, and those of IpNAC4 and IpNAC7 expression reached 25–130-fold at 24 h. Under 600 mM NaCl treatment, the 12 genes were not strongly induced at 3 and 6 h. IpNAC1 and IpNAC2 were upregulated after 12 h (16–19-fold) and remained stable or were downregulated in the early stage. The induction level of IpNAC5, IpNAC8, IpNAC11, and IpNAC12 increased 10- to 60-fold (24 h) and were upregulated steadily throughout the period. The expressions of IpNAC3, IpNAC4, IpNAC6, IpNAC7, and IpNAC9 reached 15- to 280-fold (48 h). Interestingly, IpNAC6, IpNAC7, and IpNAC9 were upregulated, but the trend was high–low–high. The IpNAC genes were generally downregulated after 900 mM NaCl solution treatment, and IpNAC2, IpNAC3, IpNAC4, IpNAC6, IpNAC7, IpNAC8, IpNAC9, IpNAC11, and IpNAC12 reached the maximum significance level at 48 h, among which IpNAC2, IpNAC3, and IpNAC4 were downregulated (3–24 h), and IpNAC5 and IpNAC10 maintained a high induction at 6 h. Interestingly, IpNAC1 did not respond to 900 mM NaCl solution treatment.

Ipomoea pes-caprae is a high temperature- and drought-tolerant crop. The expression level of most IpNAC genes was increased under heat stress (45 °C) and drought stress (20% PEG 6000), but the change in the expression level under drought stress was more obvious than that under high temperature stress ( Figure 5 ). Under a high temperature (45 °C), the IpNAC5 induction level was the highest (451-fold), followed by that of IpNAC1, IpNAC2, IpNAC3, and IpNAC9, which reached the highest significant expression (10–90-fold) at 24 h. However, the maximum induction of IpNAC4, IpNAC6, IpNAC7, IpNAC8, IpNAC9, IpNAC10, IpNAC11, and IpNAC12 was less than 10-fold. Interestingly, all IpNAC genes were strongly induced at 24 h under heat stress. Most IpNAC genes were near or below the baseline at 3–24 h and near the baseline at 48 h. Under drought stress, IpNAC12 reached a maximum induction at 48 h (442-fold), followed by IpNAC1, IpNAC2, IpNAC3, IpNAC5, IpNAC8, IpNAC9, and IpNAC11, reaching a maximum at 12 h (1–80-fold). IpNAC4, IpNAC6, and IpNAC10 increased 10- to 300-fold at 3 and 48 h. However, the expression trends of IpNAC2 and IpNAC3 after high temperature and drought stress were consistent ( Figure 5 ). In addition, compared with high temperature stress, IpNAC2 and IpNAC3 were less responsive to drought stress. Under the same stress, IpNAC6 and IpNAC10 showed similar trends but were less responsive to high temperature stress.

In response to internal and external stress, plant hormones (abscisic acid [ABA], salicylic acid [SA], and methyl jasmonate [MeJA]) play an important role in the stress signal network. We studied the relationship between the transcriptional levels of the 12 IpNAC genes under different plant hormones ( Figure 6 ). Gene expression of IpNAC4, IpNAC6, IpNAC7, IpNAC9, IpNAC10, IpNAC11, and IpNAC12 after ABA treatment increased throughout the treatment period, and most genes were rapidly upregulated 5- to 500-fold at 24 and 48 h, while IpNAC10 quickly reached 477-fold at 6 h. Similarly, the induction levels of IpNAC1, IpNAC2, IpNAC3, IpNAC5, and IpNAC8 reached maximum significance (4–60-fold) at 12–24 h, and the early response was weak or downregulated. Under SA stress, IpNAC1, IpNAC8, and IpNAC12 were significantly upregulated throughout the treatment period. The induced expression of IpNAC12 reached 2064-fold at 48 h. IpNAC2, IpNAC3, IpNAC4, IpNAC5, IpNAC6, IpNAC7, IpNAC9, IpNAC10, and IpNAC11 were mainly upregulated (2–500-fold) at 6–24 h, and except for the highest induction time, the induction times were close to baseline or downregulated. Based on the results of the MeJA treatment, the transcript level of IpNAC9 reached a peak of 241-fold at 48 h. IpNAC4, IpNAC6, IpNAC7, IpNAC8, IpNAC11, and IpNAC12 also maintained a high level (5–150-fold) throughout the period. Most trends continuously increased from 3 to 24 h and steadily decreased from 24 to 48 h. In contrast, IpNAC1, IpNAC2, IpNAC3, IpNAC5, and IpNAC10 were downregulated (2–60-fold) in different periods (3–12 h). At the same time, IpNAC5 did not respond to MeJA induction and was not sensitive throughout the period ( Figure 6 ).

DNA- and TF-specific binding is involved in transcriptional regulation activities, and most TFs with substantial functions are located in the cell nucleus. We screened four candidate salt-tolerance genes and performed subcellular localization for IpNAC5/8/10/12 in vivo ( Figure 7 ). The positions of IpNAC5/8/10/12 in plant cells were detected by GFP markers under the control of the CaMV 35S promoter. In the transient expression of GFP-IpNAC5/IpNAC8/IpNAC10/IpNAC12 in tobacco mesophyll protoplasts, the GFP signal was located in the nucleus and mixed with nuclear marker dye DAPI ( Figure 7 ). However, GFP was located in the nucleus of tobacco mesophyll cytoplasm when tobacco mesophyll protoplasts carrying GFP alone were used as controls. These results suggest that GFP-IpNAC5, GFP-IpNAC8, GFP-IpNAC10, and GFP-IpNAC12 are located in the nucleus ( Figure 7 ), which is consistent with the results presented in Table 1 .

Heterologous overexpression in Arabidopsis is a useful technique for verifying gene function. The physiological phenotypes of the WT and transgenic Arabidopsis on 1/2 MS were similar ( Figure 8A ). At 180 mM NaCl, the root length of transgenic plants was longer than that of WT plants ( Figures 8B, C ). Arabidopsis overexpressing IpNAC5/8/10/12 was highly salt tolerant at the post-germination stages. Among these, plants overexpressing IpNAC5 and IpNAC10 showed excellent characteristics.

To determine whether IpNAC5/8/10/12 affect root Na⁺ homeostasis, we carried out root transgenics and obtained many complete transgenic roots ( Figure 9 ). The target genes showed higher expression in the positive transgenic roots (TR) than in the non-transgenic roots (AR) ( Figures 9B, D, F, H ). The salt-induced Na ⁺ fluxes in the elongation and maturation zones of AR and TR were measured using the noninvasive microtest technique (NMT) after 24 h of NaCl stress (150 mM). Na ⁺ efflux in the elongation and mature regions of AR and TR differed ( Figure 10 , Figure 11 ). Among the four genes, the Na ⁺ efflux of the IpNAC10-overexpressing plant TR was more obvious than that of AR. The average Na ⁺ efflux rates in the two root zones of TR were 3.2- and 6.7-fold that of AR, respectively ( Figures 11A, B ). Interestingly, the Na ⁺ efflux of TR in the IpNAC5-overexpressing plant was 3.3-fold that of AR in the elongation zone, but there was no significant difference in the mature zone ( Figures 10A, B ). The average Na ⁺ efflux rate of TR in IpNAC8 was 1.8-fold that of AR in the mature zone ( Figures 10C, D ). The average Na ⁺ efflux rate of TR in IpNAC12 was 2.2-fold that of AR in the elongation zone ( Figures 11C, D ). These results indicate that IpNAC5/8/10/12 overexpression probably activated the ion channel and increased the root Na⁺ ion efflux efficiency.