StTGA gene identification in the potato genome was performed using BLAST and hidden Markov models (HMM) search methods (Wang et al., 2018). We downloaded the potato genome from the Phytozome database. Briefly, the HMM seed file of delay of germination 1 (DOG1) domain (accession number: PF14144), which belongs to TGA, was downloaded from the Pfam database (http://pfam.xfam.org/). A round of HMM scan was performed for all the obtained hits against the Pfam database. We used the HMMER program to search all putative TGA protein sequences and extracted the corresponding sequence IDs, with expectation value (E-value) set to 1.0. The sequences IDs were submitted to the Spud DB (http://solanaceae.plantbiology.msu.edu/). The results from the two methods were compared, and common sequence IDs were selected. In total, 42 members of the StTGA gene family in potato were predicted and named StTGA01–StTGA42. The information about the renaming StTGA genes in potato is provided in Supplementary Additional File S1. The physicochemical parameters, such as molecular weight and isoelectric points (pIs), and subcellular localization of each StTGA protein were predicted via ExPASy (http://web.expasy.org/protparam/) and PSORT (https://wolfpsort.hgc.jp/) online tools (Liang et al., 2017).

We constructed the phylogenetic tree by MEGA 7.0 via the neighbor-joining (NJ) method, using 1000 bootstrap iterations and default parameters (Gao et al., 2018). The potato gff3. annotation file was parsed to extract the genome locations of the identified TGA genes. According to this information, we visualized the in silico predicted chromosomal distribution of StTGA genes by TBtools software (Chen et al., 2020). The in silico exon–intron distribution of StTGA genes was predicted through the Gene Structure Display Server (GSDS) website (http://gsds.gao-lab.org/index.php). Conserved motifs of StTGA genes were predicted by the online service Multiple Em for Motif Elicitation (MEME) (Fan et al., 2021a). The information of in silico chromosome location of StTGA was retrieved from potato genome data downloaded from the phytozome database (Chen et al., 2020). The secondary and 3D structure prediction of StTGA proteins was performed using SOPMA and I-TASSER web servers, and the data were captured using Discovery Studio 4.5 (Geourjon and Deléage, 1995; Ambrish et al., 2010). The upstream 2000 bp region of the initiation codon ATG of each StTGA gene was used to search the promoter region in the downloaded sequence and predicted potential cis-acting elements of each binding site on PlantCARE website (Lescot et al., 2002; Zhu et al., 2020).

Pair-wise all-against-all BLAST was performed for potato, tomato, tobacco, Arabidopsis, and pepper protein sequences. The obtained results and the gff3. annotation file were then determined by the Multiple Collinearity Scan toolkit (MCScanX) for determination of the gene duplication type (Wang et al., 2012; Fan et al., 2021b). Microsynteny relationships between potato and the other four species were analyzed to show the gene homology relationship (Gutsche and Zachgo, 2016), and visualized by TBtools (Chen et al., 2020).

Based on the digital expression RNA-Seq data retrieved by previous methods (Cao et al., 2021), the expression level of StTGAs in different tissues and developmental stages under abiotic/biotic stress was indicated using Fragments Per Kilobase of transcript sequence per Millions base pairs (FPKM), which were retrieved through the website Spud DB (http://solanaceae.plantbiology.msu.edu/) (Cao et al., 2021). The heatmaps of StTGA expression and hierarchical clustering analysis were conducted using the TBtools software.

Furthermore, the co-expression pattern of StTGA genes was analyzed based on the Pearson correlation coefficient (PCC) and graphically presented using the Cytoscape package. The PCC and mutual rank (MR) were calculated according to previous reports (Da et al., 2019). Forty-two potato StTGA differently expressed genes and 24 related genes were analyzed by co-expression network analysis (Renamed in Supplementary Additional File S2). The interaction network of differentially expressed TFs was built by STRING: functional protein association networks (https://string-db.org/) and determined by Cytoscape (Kohl et al., 2011; He et al., 2022). When the PCC was greater than 0.8, it was assumed that they were co-expressed.

Potato plants (ZHONG 3,2 n = 48, tetraploid cultivation) used in this study were provided by the Chinese Academy of Agricultural Sciences (CAAS). The pots which were used to culture these plants were 10 cm in diameter and 15 cm in height, and they contained 450 ml peat and vermiculite (3:1, volume ratio). In the process of culture, S. tuberosum (ZHONG 3) wild-type plants were grown under long-day conditions (16 h, 26°C, day/8 h, 18°C, night) on the abovementioned mixed soil. The light intensity was about 3000 lx, and the relative humidity was 75% (Gao et al., 2010; Yu et al., 2021a).

R. solanacearum belongs to PO41 strain (seed type II, race 3 biotype 2) (He et al., 1983), which was cultured in a CPG medium, and then the solution was diluted to 10⁸ cfu/ml (Optical Density, OD600 = 0.2). Potatoes were inoculated by the root injury irrigation method when they grew to the stage of two-week-old potato plants with 7–8 leaves (Yu et al., 2021b). The control group was inoculated with the same amount of water. At 12, 24, 36, 48, 60, 72, 84, 96, 108, and 120 hpi after inoculation, we sampled the leaves of the experimental group and control group. All samples were frozen in liquid nitrogen immediately and then stored at −80°C (Kong et al., 2016). Three replicates were performed for both R. solanacearum and mock inoculation.

When potatoes grew to the stage of two weeks with 7–8 leaves, the leaves were sprayed with 100 μM abscisic acid (ABA), 350 μM, gibberellin 3 (GA₃), 50 μM SA, and 50 μM methyl jasmonate (MeJA) (Yu et al., 2021a). At 1, 2, 3, 4, and 5 dpi, we sampled the leaves of each hormone treatment. All samples were frozen in liquid nitrogen immediately and then stored at −80°C until these samples were used to extract RNA (Fan et al., 2021a). Each treatment was repeated three times.

The RNA-seq data of the StTGA39 gene were verified by qRT-PCR according to the methods of Schmittgen and Livak, 2008. The leaf samples for RNA extraction were previously stored at −80°C, and the total RNA was isolated using the TRIzol reagent (Invitrogen, Carlsbad, CA, United States) according to the manufacturer’s protocol. The mRNA was re-transcribed into cDNA using the Prime Script cDNA Synthesis Kit (TransGen, Beijing, China). Then, cDNA was used as a template for qRT-PCR, which was performed on the Quant Studio 3 qRT-PCR System (Thermo Fisher Scientific, Shanghai, China) (Kong et al., 2016). The following primers were used in the process of qRT-PCR with three replicates for each gene as described previously: StTGA-FP: TCC​AGC​ACA​TCC​AAC​ACC; StTGA-RP: TTC​ACC​AAG​ATT​TCC​CAC. ACTIN of potato served as the internal control (GenBank Accession: X55747), and the following primers were used in the process of qRT-PCR with three replicates for each gene as described previously: Actin-F: TAT​AAC​GAG​CTT​CGT​GTT​GCA​C; Actin-R: ACT​GGC​ATA​CAG​CGA​AAG​AAC​A.

After the potato seedlings were inoculated with R. solanacearum or water for 72 hpi, the leaves were collected for observing the infection of R. solanacearum. The leaves were left in trypan blue solution for 24 h (Yu et al., 2021b). Levels of H₂O₂ were detected using the DAB staining method by the DAB Substrate Kit (Solarbio, Beijing, China). Leaves were soaked in the solution to stain for 3–10 min under dark conditions (Christensen et al., 1997; Song et al., 2021). The leaves were sampled from inoculated plants by R. solanacearum at 12, 24, 36, 48, 60, 72, 84, 96, 108, and 120 hpi for quantifying levels of MDA. The content of MDA of these leaves was tested by using an MDA Assay Kit (Abbkine, Wuhan, China) according to the previous method (Islamoglu et al., 2018). OD at 532 and 600 nm was determined by a microplate reader.

In order to analyze whether there is a TGACG cis-element in the upstream promoter sequence of the gene homologous to S. tuberosum Brassinosteroid insensitive 1 (BRI1)-associated receptor kinase 1 (StBAK1), we analyzed the relationship between StTGA1 and StBAKl by the Y1H assay Kit (Clontech, State of California, United States) in this study. Briefly, the sequence of StBAK1 with 1512 bp was amplified from S. tuberosum cDNA by gene-specific primers and digested by EcoRI (CAATTC) and MluI (ACGCGT) so that the oligonucleotide sequences containing TGA cis-elements and other corresponding promoter sequences were cloned into the pHIS2 to generate the reporter vector. All of the constructs were co-transformed into yeast strain Y187 according to the manufacturer’s instructions (Liu et al., 2016). The full-length open-reading frame (ORF) of StTGA (855bp) was fused to the GAL4 activation domain in the vector pGADT7 digested with EcoRI (CAATTC) and BamHI (GGATCC) to get a fusion protein TGA-pGADT7 (effector vector) (Huang et al., 2010). The others were carried out following the manufacturer’s instructions.

Standard values and standard error of experimental data were calculated using Microsoft Excel 2016 and n = 3 for independent experiments according to a t-test. Analyses of the significance of differences were conducted via the data processing System. The PCC is calculated by the SPSS online services tool (https://spssau.com/indexs.html). PCR efficiency was estimated from the data obtained from the exponential phase of each individual amplification plot. The expression level of each gene of interest (GOI) is presented as 2^−∆∆Ct; where ∆∆Ct = ∆Ct_GOI-∆Ct_Contorl; ∆Ct_GOI = Ct_GOI-Ct_Actin; ∆Ct_Contorl = Ct_Control-Ct_Actin.

In total, 42 StTGA genes were identified in S. tuberosum, and they were named StTGA01–StTGA42. This study revealed that the largest number of amino acids is 488 in the StTGA gene family, and the relative molecular weight of these genes is between 17,290.85 and 54,371.83 Da. In addition, we predicted the pIs of StTGA protein. The results showed that the pI of StTGA17 is 4.96, which is the lowest in the StTGA protein, while the pI of StTGA01 is 9.1, which is the highest. Furthermore, we speculated the StTGA may be distributed in many parts of the cell and most of StTGA proteins are located in cytoplasm and nucleus (Supplementary Table S1).

Using TBtools, we drew the location map of StTGA and found they are distributed on four chromosomes of potato (Figure 1A). Chromosome St12 had 29 StTGAs, chromosome St08 contains six, St01 with four, and St09 contain three StTGAs. In order to study the StTGA family, we constructed the phylogenetic tree by MEGA 7 via the NJ method, which is shown in Figure 1B. The StTGA family was divided into six groups (designated classes I to VI) with 9, 7, 3, 12, 4, and 7 StTGAs. Then, the exon–intron structure of StTGAs was further analyzed. The number of exons of StTGA genes was between 1 and 11, of which six genes contained only one exon, and 42.86% of genes had five exons or less Figure 1D. Furthermore, 10 high conserved motifs were predicted in Figure 1C and Figure 1E. Compared to other classes, motifs 10 were exclusively found in classes IV, motifs 9 in classes I and II, and motifs 8 in classes IV and VI, separately Figure 1C.

In order to study the transcriptional regulation, the 2000 bp promoter region upstream of the StTGA genes was extensively analyzed by PlantCARE. The cis-elements responding to abiotic/biotic stresses and hormone treatments with the prediction score being greater than or equal to 5 were considered for further analysis (Figure 2). The results showed that the numbers and distribution patterns of the cis-elements also greatly varied among the promoters. This means that the StTGA genes might have different regulatory mechanisms in expression, implying their functional divergence in responses to abiotic/biotic stress and phytohormone treatments.

To study the evolutionary relationship of StTGA, we selected S. tuberosum as the core and identified collinearities between tomato, tobacco, Arabidopsis, pepper, and potato TGA genes. There are 46 pairs of one-to-one microsynteny relations (Figure 3). Some of the genes could not be assigned to any of the chromosomes. In addition, 21 orthologous gene pairs were detected between potato and tomato, 20 orthologous genes were found between potato and pepper, 10 orthologous gene pairs were identified between potato and Arabidopsis, and two orthologous gene pairs were detected between potato and tobacco. The collinear relationship among these five species is shown in Supplementary Additional File S3.

We have retrieved the expression data of StTGA via the RNA-Seq Expression Browser and the heat map of gene FPKM value constructed using the TBtools software. We found that 11 of the StTGAs (StTGA 01, StTGA 02, StTGA03, StTGA 04, StTGA 11, StTGA 16, StTGA 23, StTGA 28, StTGA 34, StTGA 41, and StTGA42) were not expressed in almost all organs. On the contrary, StTGA05, StTGA15, StTGA25, StTGA32, and StTGA36 played roles in different developmental stages and tissues (Figure 4). For verification of RNA-seq data, the potato seedlings were treated with the exogenous hormones ABA, GA₃, SA, and MeJA, and the results showed the inducible upregulated expression of the StTGA39 gene (Figure 5). By comparison, the effect of SA was relatively strong, while that of the MeJA, ABA, and GA₃ was light for inducing the upregulation.

To determine the relationship of StTGA proteins and their interactors further, we performed a gene co-expression network analysis based on the STRING database. Twenty-four other potato proteins, such as 60S ribosomal protein L18 (RPLs), DNA-binding proteins (DBPs), and Glutaredoxin (GRXs), were identified Figure 6, which exhibited a co-expression relationship with StTGA (Figure 6A). The co-expression network comprised 36 pairs, and each pair of nodes was directly or indirectly connected to each component (Figure 6A). StTGA and StTGA members, StTGA and other potato proteins presented co-expression relationships, albeit with different weight values (Figure 6B). In particular, StTGA01 (StTGA28), StTGA02 (StTGA40), StTGA05 (StTGA13), etc., presented co-expression relationships (Figure 6A).

The secondary structure of StTGA39 was predicted by SOPMA. It was found that the a-helix of StTGA protein consists of 188 amino acids, accounting for 67.14%; the random coil of StTGA protein consists of 78 amino acids, accounting for 27.86%; the extended strand of StTGA protein consists of nine amino acids, accounting for 3.21%; and the β-turn of StTGA protein consists of five amino acids, accounting for 1.79% (Figure 7A). A three-dimensional structure of StTGA39 was built with the I-TASSER web server. Five possible models of StTGA protein were predicted on the 3D structure, according to the website modeling with Discovery Studio 4.5 Client. The conservative domains of these models were basically the same (Figure 7B).

In order to confirm whether StTGA39 was induced by R. solanacearum, the induced expression of StTGA39 mRNA in potato seedlings inoculated with R. solanacearum was compared and analyzed by the qRT-PCR method (Figure 8A). We found that the expression level of StTGA39 was low when plants were inoculated with water. In comparison with the control group, the expression increased at 60 hpi and reached the highest level at 72 hpi after inoculating with the highly pathogenic strain PO41. That is to say, the expression of the StTGA39 gene was up-regulated when stressed by R. solanacearum and may play an important role in potato resistance to bacterial wilt. Similar results have been confirmed by other family members of StTGA (Supplementary Additional File S4).

In order to determine the defense response, various histochemical analyses were carried out on potato leaves (Figure 8). Under the stress of R. solanacearum solution of 10⁸ cfu/L, the effect of MDA content in potato seedling leaves is shown in Figure 8B. It could be seen from Figure 8B that with the prolongation of infection time of R. solanacearum, the change of MDA content in potato seedling leaves showed a trend, which was upregulated once at 24 h and then decreased at intervals of 24 h. Callose strengthens the plant cell walls and prevents cells from being invaded by pathogens. Aniline staining analysis and trypan blue analysis showed that the induction of callose deposition and the HR-like cell death was significantly increased in the treated potato leaves (Figure 8C and Figure 8D).

BAK1 is a well-established receptor for defense-related genes in plants. To identify if StTGA1 could interact directly with StBAK1, Y1H analysis was used (Figure 9). The yeast expression vector TGA-pGADT7 and BAK-pHis2 were constructed. The results of Y1H showed that all the Y1H gold strains co-transformed with pHIS2 vectors and clones grew on SD (-Leu, -Trp) medium, indicating that the co-transformation was successful. The growth was different at different concentrations of 3-AT-deficient SD (-His, -Leu, -Trp) medium. On SD (-His, -Leu, -Trp) medium, the growth inhibition of the pGADT7/BAK-pHis2 self-activated group deepened with the increase of 3-AT concentration. The growth condition of the TGA-pGADT7/BAK-pHis2 group was slightly better than that of the control group, although the experimental group could not grow on SD (-His, -Leu, -Trp; 90 mM) (Figure 9). The results showed that StTGA1 might bind to StBAK1.