Verbena bonariensis seeds were obtained from Benary Seed Company (Hann. Münden, Germany). Seedling maintenance and multiplication were carried out in the Germplasm Resource Nursery at the Key Laboratory of Mountain Plant Resources Protection and Germplasm Innovation (Ministry of Education), Guizhou University. Wild-type tobacco seeds were provided by our laboratory’s nursery, and all plant materials were identified and authenticated by Dr. Yan Li and Dr. Sumeera Asghar.

The V. bonariensis genome annotation (PRJCA027875) was downloaded from the National Genomics Data Cen ter (NGDC, https://ngdc.cncb.ac.cn) [20], downloaded the TIFY conserved domain (PF06200) from Pfam database (http://pfam.xfam.org/). Candidate members containing the conserved structural domain were obtained by aligning the conserved structural domain to the genomic data of V. bonariensis using HMMER 3.1 (http://hmmer.org/download.html). In addition, 18 Arabidopsis TIFY members were compared to the Verbena willow genome using BLASTP (E-value < 1 × 10⁻⁵) to obtain candidate members. All the candidate members obtained were submitted to: SMART (http://smart.embl-heidelberg.de/) and NCBI-CDD (http://www.ncbi.nlm.nih.gov/Structure/cdd/wrpsb.cgi) to determine the conserved structural domains of TIFY, and finally obtain the TIFY family members.

The physicochemical parameters of V. bonariensis TIFY family proteins, including protein length, molecular weight, theoretical isoelectric point (pI), instability index, and grand average of hydropathicity (GRAVY), were analyzed using the ProtParam (https://web.expasy.org/compute_pi/) [21]. The secondary structure composition was analyzed using the SOPMA (http://npsa-pbil.ibcp.fr/cgi-bin/npsa_automat.pl?page=npsa_sopma.html). Transmembrane helices were predicted via the TMHMM (https://services.healthtech.dtu.dk/services/TMHMM-2.0/), subcellular localization was determined by WoLF PSORT (https://wolfpsort.hgc.jp/) [22]. And the precise positions of conserved structural domains were mapped using NCBI’s Conserved Domain Database (https://www.ncbi.nlm.nih.gov/Structure/cdd/wrpsb.cgi).

The intron-exon organization profiles of all identified VbTIFY genes were retrieved from the official genome annotation dataset of V. bonariensis. Conserved domain information was acquired by searching against the NCBI CDD database (https://www.ncbi.nlm.nih.gov/Structure/cdd/wrpsb.cgi). A total of 10 evolutionarily conserved motifs were identified using the MEME (http://meme-suite.org/tools/meme) [23] with default parameter configurations. All the above structural features were ultimately visualized using TBtools software (v2.042) [24].

Using the MUSCLE program within MEGA software, multiple sequence alignment was performed between the identified TIFY family genes from V. bonariensis and from A. thaliana. A phylogenetic tree was generated using MEGA version7.0 Bootstrap with parameter 1000 [25]; Subsequently the phylogenetic tree of VbTIFY genes was then reconstructed via the neighbour-joining method. Finally, the constructed phylogenetic tree was annotated and visualized using the Evolview v2 (https://evolgenius.info/evolview-v2/) [26]. Based on the genomic annotation data of V. bonariensis, the chromosomal localization of members of the VbTIFY gene family was visualized using TBtools v2.042 [24].

The 2000 bp genomic sequences upstream of the translation initiation codon (ATG) of each identified VbTIFY gene were extracted from the V. bonariensis genome file using TBtools v2.042 [24]. Putative cis-acting regulatory elements embedded in these promoter sequences were annotated via the PlantCARE online database (http://bioinformatics.psb.ugent.be/webtools/plantcare/html/) [27]. The distribution patterns and functional classification of all identified cis-elements were subsequently visualized and plotted using the same TBtools software [24].

Genome-wide datasets and annotation files were acquired on the basis of V. bonariensis genome annotation. Collinear relationships among the gene family members were analyzed using the MCScanX algorithm [28], and TBtools v2.042 was subsequently applied for figure visualization [24].

Using the transcriptome dataset generated from V. bonariensis, we retrieved the expression level (FPKM) values of all identified VbTIFY genes across three distinct tissues (flowers, stems, and leaves) and in samples collected pre- and post-12-hour 4℃ cold stress. To eliminate the influence of extreme values, we normalized the raw expression data via log₂(FPKM + 1) conversion, and subsequently generated heatmaps to visualize the expression profiles using TBtools 2.042 [24].

Two-month-old healthy Verbena bonariensis plants were selected as experimental materials. Different organs (leaf, flower, and stem) were sequentially collected, placed into sterile, nuclease-free cryotubes, and immediately frozen in liquid nitrogen. These samples were then stored at −80°C for subsequent RNA extraction. Total RNA was isolated using the OMEGA Extraction Kit (OMEGA, Norwalk, CT, USA), and first-strand cDNA was synthesized via reverse transcription using the StarScript III RT Kit. Quantitative real-time PCR (qRT-PCR) reactions were performed on a qTOWER³G Real-Time PCR System (Analytik Jena, Beijing, China) using 2× RealStar Fast SYBR qPCR Mix (GENStar, Beijing, China), with the VbACTIN gene serving as the internal reference. The 10 μL reaction system consisted of 1 μL of cDNA template, 0.5 μL of reverse primer, 0.5 μL of forward primer, 5 μL of 2× RealStar Fast SYBR qPCR Mix, and DEPC-treated water to a final volume of 10 μL. Three biological replicates and three technical replicates were set up for each sample. Additionally, primers were designed using NCBI Primers—blast and Primer Premier 5 software. The relative expression level was calculated using the 2^−ΔΔCT method [29], and results are expressed as mean ± standard deviation.

The pCAMBIA1300-VbTIFY16-EGFP fusion expression vector was constructed using the seamless cloning method, and transformed into Nicotiana benthamiana leaves via Agrobacterium tumefaciens strain GV1301-mediated transient transformation. After 2 d of dark incubation, the subcellular localization of VbTIFY16 protein in tobacco leaf epidermal cells was observed using a laser scanning confocal microscope.

The complete coding sequence of VbTIFY16 was amplified via PCR and cloned into the pCAMBIA1301-35S vector using double restriction enzyme digestion, generating the overexpression construct pCAMBIA1301-35S-VbTIFY16. The constructed vector was transferred into Agrobacterium tumefaciens strain EHA105, followed by Agrobacterium-mediated transformation of Nicotiana benthamiana referring to the protocol described by Ning et al. [30]. As reported by Jefferson et al. [31], histochemical staining of β-glucuronidase (GUS) was employed to validate the identity of transgenic plants.

Both transgenic overexpression (OE) plants and wild-type (WT) control plants were maintained in a greenhouse culture system for comprehensive phenotypic characterization. At 65 days following transplantation, multiple growth-related phenotypic indices (plant height, leaf length, leaf width and leaf area) were collected and statistically analyzed; the floral traits (flower diameter and petal area) were assayed at the full blooming (FB) developmental stage. To characterize the cellular features of petals, petal tissue samples were harvested from OE and WT for sectioning, which was followed by morphological observation and size measurement of petal cells. The morphology of petal epidermal cells was observed and photographed using a microscope (Olympus, BX43, Tokyo, Japan) [32]. Cell length and area were determined using ImageJ software (NIH, Bethesda, MD, USA), and the total number of epidermal cells in petals was calculated by multiplying the average number of cells per 1 mm² by the total petal area.

After comprehensive screening and identification, 16 VbTIFY genes were identified as from VbTIFY1 to VbTIFY16 (Table 1). The encoded proteins had amino acid lengths of 136–372 molecular weights of 15,075.11–40,696.34 Da, and theoretical pI values from 4.74 to 9.75. Thirteen out of 16 VbTIFYs had values higher than 7.0, indicating that most of the V. bonariensis TIFYs were alkaline proteins. Protein instability coefficients of VbTIFYs proteins ranged from 41.55 (VbTIFY15) to 78.85 (VbTIFY3), and the instability index was noticed higher than 40, which was a variable protein. The aliphatic index varied from 53.7 (VbTIFY6) to 82.76 (VbTIFY13). The overall mean of hydrophobicity ranged from −0.842 (VbTIFY2) to −0.007 (VbTIFY9), The average value of grand average of hydropathicity. Additionally, subcellular localization analysis suggested that all VbTIFY proteins are localized to the nucleus (Table 1). Among these, the nuclear localization of VbTIFY16 was further validated experimentally (see Section 3.8, Fig. 1).

Transmembrane domain analysis confirmed that among the 16 members of the VbTIFY family only VbTIFY9 and VbTIFY12 possessed a single transmembrane domain, while the remaining 15 members exhibited no transmembrane domain characteristics. Secondary structure predictions showed that all VbTIFY members had no β bridges, among which random coil was the most common secondary structure in the amino acid sequence of V. bonariensis TIFY, accounting for 70.81% (VbTIFY13) to 85.95% (VbTIFY9), followed by α helix 8.75% to 21.08%, and extended chain 4.21% to 10.29% (Table 2).

The results revealed the presence of four putative conserved domains within the VbTIFY family proteins: the TIFY domain, JAS motif domain, CCT domain, and GATA domain. All 16 VbTIFY proteins contain the TIFY domain, confirming their classification as members of the TIFY family. The CCT and GATA domains were exclusively identified in VbTIFY10/11/15, indicating that these three proteins are classified into the ZML subfamily. VbTIFY06 contains only the TIFY domain, suggesting possible structural loss during evolution. The remaining 12 members all possess two conserved domains: TIFY and JAS (Fig. 2). A total of 10 conserved motifs were identified among the TIFY family members in V. bonariensis (Fig. 3). The number of conserved motifs varied among different VbTIFY proteins, ranging from 1 to 6. All VbTIFY members shared Motif 1, which corresponds to the TIFY domain (Fig. 2). The findings confirmed that the VbTIFY gene family has retained high evolutionary conservation. Further phylogenetic analysis indicated that proteins within a given clade displayed analogous domain organizations and conserved motif patterns, which suggests that genes with comparable motif compositions are likely to have related biological functions. Furthermore, Exon-intron structures and 5′/3′ UTR sequences of the 16 identified VbTIFY genes in V. bonariensis were systematically analyzed, which provides essential support for clarifying the structural heterogeneity of the VbTIFY gene family. The findings indicated that the number of exons varied significantly among different members, ranging from 2 to 10. Except for VbTIFY06, which contained 3 non-coding regions, the number of non-coding regions in other VbTIFY members ranged from 0 to 2, indicating that the VbTIFY family genes are relatively conserved during evolution.

To elucidate the evolutionary divergence of TIFY proteins in V. bonariensis, multiple sequence alignment was performed on 34 TIFY protein sequences, (16 from V. bonariensis and 18 from A. thaliana), followed by the construction of a phylogenetic tree using the neighbor-joining (NJ) method. (Fig. 4). Phylogenetic analysis results showed that the 34 TIFYs were divided into six independent groups, which were sequentially named Group 1 to Group 6. Members of the VbTIFYs family were mainly concentrated in Groups 1, 3, 5 and 6, while only one VbTIFY family member was contained in each of Group 2 and Group 4. Based on the structural characteristics of the TIFY domain, this gene family can be divided into four major subfamilies: JAZ, ZML, TIFY and PPD. Group 1 belongs to the ZML subfamily, containing three members including VbTIFY10/11/15, as well as three members derived from A. thaliana. Group 2 falls into the TIFY sub-family and contains one member, VbTIFY6. Group 4 is assigned to the PPD sub-family and includes one member, VbTIFY4. The remaining three groups Group 3/5/6 belong to the JAZ subfamily and contain a total of 11 members. Furthermore, the phylogenetic tree analysis demonstrated tight clustering of TIFY proteins within the same group, a pattern consistent with the tree constructed using VbTIFY proteins in V. bonariensis. (Fig. 2 and Fig. 4).

The tree includes 16 VbTIFY proteins (red ★) and 18 AtTIFY proteins (black ●). The proteins are divided into six subgroups (Groups 1–6), corresponding to four subfamilies: ZML (Group 1), TIFY (Group 2), JAZ (Groups 1). Actually (Groups 3,5,6), and PPD (Group 4). The NJ tree was generated with 1000 bootstrap replicates.

16 VbTIFY genes were mapped to the chromosomes of V. bonariensis (Fig. 5). It is important to note that V. bonariensis is an autotetraploid species (4n), and the genome assembly used in this study represents a single subgenome (n). Therefore, the 16 VbTIFY genes identified here are expected to represent only a subset of the total TIFY complement, with additional copies potentially present in the three remaining subgenomes. All members were successfully located, with the distribution across chromosomes as follows: Chr.02 and Chr.06 contained the highest number of VbTIFY genes, each harboring four members, followed by Chr.05 with three members. Chr.03 and Chr.07 each carried two VbTIFY genes, while Chr.04 contained only one.

To investigate the evolutionary conservation of the TIFY gene family, collinearity analysis detected four homologous gene pairs (Red line in Fig. 6). As tabulated in Table 3, Ka/Ks ratios for these 4 pairs were all <1, demonstrating that the homologous genes evolved under negative selection (purifying selection pressure).

To elucidate the putative regulatory roles of the TIFY gene family in V. bonariensis, we screened the 2000 bp promoter regions upstream of the start codon for cis-acting regulatory elements (Fig. 7). The results revealed that VbTIFY genes harbor a diverse array of cis-elements, with light-responsive elements being the most prevalent (93), abscisic acid responsiveness (64), MeJA-responsiveness (28), anaerobic induction (35), low-temperature-responsive elements (10), auxin responsiveness (7), gibberellin responsiveness (8), among others. Specifically, 14 members contained abscisic acid-responsive cis-elements, apart from VbTIFY6 and VbTIFY13. MeJA-responsive elements were identified in eight VbTIFY genes (VbTIFY2/3/4/8/10/11/15/16). Regarding other phytohormone-related cis-elements, for instance, VbTIFY8 contained five types of auxin, gibberellin, salicylic acid, abscisic acid, and MeJA responsiveness. The presence of these cis-regulatory elements indicates that the TIFY gene may play a role in the cross-regulation of multiple hormones.

Regarding environmental response elements, all members carried multiple light-responsive cis-elements. Thirteen genes contained cis-elements essential for anaerobic induction, excluding VbTIFY6, VbTIFY11, and VbTIFY13. Additionally, six genes (VbTIFY3/4/5/6/12/13) contained defense and stress responsiveness elements; seven (VbTIFY2/3/5/9/10/13/15) contained low-temperature responsiveness elements; and nine (VbTIFY2/3/4/6/7/11/13/14/15) contained drought-inducibility elements. Furthermore, several genes carried tissue-specific or functionally specific cis-elements. For example, VbTIFY4/6/10/15 contained elements associated with endosperm expression; VbTIFY4/6/7/10/11 possessed elements related to meristem expression; and VbTIFY8 contained a cis-acting element involved in seed-specific regulation (Fig. 7).

To further investigate the regulatory roles of TIFY gene family members in the growth and development of V. bonariensis, we calculated and visualized the transcriptomic FPKM values of 16 TIFY family members across three different tissues of V. bonariensis (Fig. 8 and Fig. 9). The results demonstrated that VbTIFY genes exhibited distinct expression profiles across different tissues of V. bonariensis. Specifically, 13 VbTIFY genes were expressed in all tissues, while three genes showed no or very low expression in most tissues. VbTIFY07/01/08/03 were highly expressed in leaves; VbTIFY05/09/11 exhibited higher expression in stems. Notably, VbTIFY16 exhibited a marked upregulation in floral tissues relative to other plant organs, implying a potential role in flower development and function. Analysis of the expression patterns of TIFY gene family members in V. bonariensis under cold stress indicated that the expression levels of 12 members were significantly elevated after low-temperature treatment. Among them, two members (VbTIFY14/15) reached peak expression at 3 h; six members (VbTIFY01/02/05/08/12/13) peaked at 8 h; two members (VbTIFY03/09) peaked at 12 h; and two members (VbTIFY07/16) peaked at 24 h. In contrast, four members showed decreased expression levels. A series of abiotic stress-responsive cis-acting elements were detected in the promoter regions of 12 upregulated VbTIFY genes, which suggests that these members may play potential regulatory roles in plant responses to cold stress and the development of cold tolerance. Collectively, our results lay a solid theoretical foundation for subsequent research on deciphering the cold stress response mechanism in V. bonariensis.

To validate the RNA-seq results from different tissues of V. bonariensis, qRT-PCR analysis on six (VbTIFY01, VbTIFY03, VbTIFY05, VbTIFY09, VbTIFY12, and VbTIFY16) randomly selected genes from the RNA-seq dataset (Fig. 10) results demonstrated expression profiles consistent with RNA-seq analysis (see Fig. 8). Notably, VbTIFY16 showed significant up-regulation in flowers, where its expression level was the highest among all the genes analyzed. Our findings indicated that VbTIFY16 may play a critical regulatory part in the growth and development processes of V. bonariensis flowers.

To investigate the subcellular localization of VbTIFY16, the empty vector pCAMBIA1300-EGFP and recombinant vector pCAMBIA1300-VbTIFY16-EGFP were transformed separately. After 2-3 days of low-light incubation, the subcellular distribution of the Protein-GFP fusion protein in the agroinfiltrated leaves was observed under a laser scanning confocal microscope (LSCM) with excitation at 488 nm. The results showed that strong green fluorescent signals of VbTIFY16-EGFP were detected in the nucleus (Fig. 1), indicating that VbTIFY16 is localized to the nucleus.

To further elucidate the function of VbTIFY16, the 35S: VbTIFY16-GUS construct was introduced into tobacco, leading to the generation of three independent transgenic lines. Observation results showed that at the full petal expansion stage, the VbTIFY16 OE plants exhibited a significantly reduced inflorescence diameter and shorter petals, with their average petal length being 20.3% lower than that of the wild-type (WT) (Fig. 11C,F,G).

To explore the reasons for shorter petals, sections of petals from both transgenic plants and WT plants were prepared following the method described by Ren, and cytological observations were carried out [32] (Fig. 11D). Morphological observation and quantitative measurement of petal epidermal cells revealed that the cell diameter and cell area of VbTIFY16-OE petals were reduced by 20.54% and 40.19% compared with those of WT petals, respectively (Fig. 11I,J). Meanwhile, the total number of epidermal cells in petals showed no significant difference (Fig. 11H). Therefore, we infer that the inhibition of petal growth in VbTIFY16-OE plants is mainly attributed to the impaired cell expansion. In addition, the overexpression of VbTIFY16 transcription factor significantly affected plant growth potential at the vegetative growth stage. Compared with WT plants, VbTIFY16-OE plants exhibited a shorter height phenotype with smaller leaves (Fig. 11). Further statistical analysis confirmed that the plant height, leaf length and leaf width of the overexpression plants were all significantly decreased (Fig. 11 and Fig. 12).