AtIDD protein sequences were used as queries to search homologous genes in tobacco genome (Edwards et al., 2017) with BLASTP (E value<100) (Camacho et al., 2009). InterPro (v93, https://www.ebi.ac.uk/interpro/) (Paysan-Lafosse et al., 2023) was used to predict the ID domain (IPR031140). Sequences were subjected to manual curation. Genes with incomplete ID domains were removed. The molecular weights and isoelectric points were calculated with EMBOSS (Rice et al., 2000).

IDD protein sequences in Arabidopsis thaliana, rice, and maize were collected from Phytozome (v10, https://phytozome-next.jgi.doe.gov/) (Goodstein et al., 2012). Multiple-sequence alignments were carried out by MAFFT (v7.520) (Katoh and Standley, 2013). A phylogenetic tree was constructed by MEGA (v7.0.21) (Kumar et al., 2016) using the maximum likelihood (ML) method with bootstrap value 1,000. The tree was edited on iTOL (https://itol.embl.de/) (Letunic and Bork, 2021).

Gene structure annotations were obtained from the GFF3 file of tobacco genome annotation. MEME (v4.9.1) (Bailey et al., 2015) was used to discover motifs in NtIDD protein sequences with parameters (minw = 8, maxw = 50, nmotifs = 10). The gene structure and motifs were plotted with the custom Python script.

The genomic locations of NtIDDs were retrieved from the tobacco genome GFF3 file (Edwards et al., 2017 version, http://solgenomics.net/ftp/genomes/Nicotiana_tabacum/edwards_et_al_2017). Synteny analysis was performed using MCScanX (Wang et al., 2012); minimum five genes were required to call a syntenic block. The types of duplication were identified using the duplicate_gene_classifier program resided in the MCScanX package. Genome distributions and syntenic blocks were visualized using Circos software (Krzywinski et al., 2009).

For cis-element prediction, 1.5-kb upstream sequences of NtIDD genes were extracted and submitted to PlantCARE (https://bioinformatics.psb.ugent.be/webtools/plantcare/html/) (Lescot et al., 2002). NtIDD1 was not performed, due to too many ambiguous bases (N) in its promoter region. After prediction, filtering was further carried out, and elements involved in three categories (hormone responsive, stress responsive, and development related) were considered for analysis.

Public RNA-Seq data generated by a previous study (GenBank accession code: SRP029183) (Sierro et al., 2014) were used for tissue expression analysis. The data were mapped to tobacco genome with HISAT2 (v2.1.0) (Kim et al., 2019). Gene expression levels (FPKM, Fragments Per Kilobase of transcript per Million mapped reads) were estimated using StringTie2 (v2.1.7) (Pertea et al., 2015).

Tobacco plants intended for hormone treatments were cultivated in a greenhouse maintained at temperatures of 28/24°C, with regulated light conditions of 16 h of light and 8 h of darkness. 3-week-old tobacco seedlings were soaked into liquid medium containing 50 μM methyl jasmonate (MeJA), 10 μM abscisic acid (ABA), 10 μM salicylic acid (SA), 10 μM gibberellin acid (GA), 10 μM 6-benzylaminopurine (6-BA, cytokinin, CK), and 5 μM GR24 (strigolactone, SL), where they were cultured for a duration of 5 h. Control seedlings were treated with a 1% (v/v) dimethyl sulfoxide (DMSO) solution.

Samples were collected with three biological replicates. Total RNA of different samples was extracted with RNA Kit (Imagene, Beijing, China) according to the instruction. The DNA was firstly removed using RNase-free DNase I (Takara, Beijing, China). Then, high-quality RNA was used for cDNA synthesis using Reverse Transcriptase M-MLV (Takara). The quantitative real-time PCR (qRT-PCR) was quantified on a LightCycler^® 96 Real-Time PCR System. The reaction program was set as the following: 95°C for 30 s, 40 cycles of 95°C for 10 s, 60°C for 30 s. The NtGAPDH gene was used as reference gene to standardize the expression level with the 2^−△△CT method. The primer sequences are listed in Supplementary Table 6 .

To construct RNA interference (RNAi) plants, the full-length coding sequence (CDS) of the NtIDD9 gene was firstly amplified. Subsequently, the derived PCR fragment was connected to the pBWA(V)HS vector through homologous recombination. The constructs were then transformed into tobacco with the help of Agrobacterium tumefaciens strain GV3101.

For cellular localization, RNA in situ hybridization was conducted according to the manufacturer’s protocol (Servicebio, Wuhan, China). The leaf petiole base tissue was fixed with in situ hybridization fixative (plant) and embedded in wax. Paraffin blocks were sliced 6 μm thick using a slicing machine. The slides were then dewaxed, dehydrated, digested, and hybridized to probes. After washing and dropping anti-Digoxin antibody, BCIP/NBT solution was used for chromogenic staining. Under microscopic observation, blue and blue-purple were interpreted as positive hybridization.

To verify the subcellular location of NtIDD9 protein, the full-length coding sequence (CDS) without stop codon was cloned into the pC1300 and C-terminal fused with enhanced green fluorescent protein (GFP). The product was reclaimed from the recombination ligation gel, and the recombinant plasmid obtained from the ligation was introduced into Agrobacterium tumefaciens strain GV3101. The vector harboring 35S::GFP-NtIDD9 and control vector were infiltrated into Nicotiana benthamiana leaves. After 24  h, fluorescence images were observed using a confocal microscope (FV1200 Olympus, Japan).

The samples were frozen in liquid nitrogen and crushed into fine powder. 50 mg of powdered samples was used to extract endogenous hormones with 1 mL of methyl−tert−butyl−ether (MTBE) solution containing MTBE, methanol, and water in a ratio of 15:4:1. After centrifugation, supernatant was collected to perform liquid chromatography with tandem mass spectrometry (LC-MS/MS) analysis. The data were captured on instrument system UPLC (ExionLC™ AD) coupled with MS/MS (QTRAP^® 6500+). By using each standard hormone calibration curve, the levels of seven endogenous hormones including auxins, cytokinins (CKs), gibberellin acids (GAs), methyl jasmonates (MeJAs), salicylic acids (SAs), abscisic acids (ABAs), and ethylenes (ETHs) were quantified. To get statistically meaningful results, three biological replicates were carried out. The differential analysis were determined using t-test with false discovery rate (FDR) < 0.05 and |log2(fold change)| ≥1.

Based on homology searching and ID domain prediction, 36 potential tobacco IDD genes were firstly screened ( Supplementary Table 1 ). After removing genes without complete ID domains, totally 26 IDDs were identified in the tobacco genome. The characteristics of NtIDD genes are listed in Table 1 and Supplementary Table 2 . In brief, the peptide lengths were ranged from 382 to 542. The molecular weights (Mw) were distributed from 44 kDa to 59.5 kDa. The isoelectric points (pI) were between 8.5 and 10.1. NtIDD protein sequences, together with IDDs from Arabidopsis thaliana, rice, and maize, were used to construct phylogenetic tree ( Figure 1 ). Based on the tree, NtIDDs could be classified into five groups: group I (NtIDD11/20/18), group II (NtIDD12/2/19/10), group III (NtIDD17/14/26/8), group IV (NtIDD24/7/15/5/4/23/1/16/13), and group V (NtIDD25/3/22/21/6/9).

Most of NtIDD genes comprised three to four exons ( Figure 2A ). NtIDD15 contains five exons, whereas NtIDD1, NtIDD6, and NtIDD21 possess only two exons. The intron length of NtIDD genes varied greatly. The shortest and longest introns were found in NtIDD25 and NtIDD15, respectively. Using MEME, 10 motifs were identified in NtIDD protein sequences ( Figure 2B ). Among them, motif 1/3 were found in all NtIDDs, motif 2/4/5/6/7 were distributed in most of NtIDDs, motif 8 was specifically found in NtIDD1/9/6/21, motif 9 was found in NtIDD5/15/7/24/1/23/4, and motif 10 was found in NtIDD5/15/7/24/1/23/6/21/22. Motifs 1/2/3/8 were associated with the ID domain, which consists of two C2H2 zinc fingers and two C2CH zinc fingers ( Figure 3 ).

By investigating distributions of NtIDDs in the genome, 17 genes were positioned on chromosomes, and 9 genes were located on scaffolds ( Figure 4 ). Using MCScanX, syntenic blocks were identified and the types of duplication were classified. There were 14 NtIDDs that were found involved in syntenic blocks ( Figure 4 ) and therefore were deemed as segmental and/or whole-genome duplications (WGD). There were six syntenic gene pairs detected among these NtIDDs (NtIDD2/12, NtIDD5/15, NtIDD6/9, NtIDD7/15, NtIDD10/19, NtIDD13/16). The other 12 NtIDDs distributed out of synteny, and they were categorized as dispersed duplications.

In order to gain more insights on regulatory mechanism, the 1.5-kb upstream sequences of NtIDDs were extracted, and their cis-elements were predicted. There were total 42 elements identified ( Figure 5 ; Supplementary Table 3 ). Among them, 10 elements were hormone responsive, 27 elements were stress responsive, and 5 elements were development related.

In the hormone-responsive elements, ABRE (abscisic acid responsive), CGTCA-motif/TGACG-motif (MeJA responsive), and TCA-element (salicylic acid responsive) were detected in most of NtIDDs. TGA-element/TGA-box (auxin responsive), TATC-box/GARE-motif/P-box (gibberellin responsive), and SARE (salicylic acid responsive) were discovered in few NtIDDs, such as NtIDD2 and NtIDD10.

In the stress-responsive elements, ARE (anaerobic induction), TC-rich repeats (defense and stress responsive), and G-box/GT1-motif/Box 4/TCT motif (light responsive) were identified in most of NtIDDs. GC-motif (anoxic induction), MBS (drought inducibility), LTR (low-temperature responsive), WUN-motif (wound responsive), and the remaining elements (light responsive) were distributed in several NtIDDs, like NtIDD2 and NtIDD4.

For the development-related elements, MSA-like (cell cycle regulation) was identified in NtIDD6/14/21, circadian (circadian control) was found in NtIDD3 and NtIDD17, GCN4_motif/AACA_motif (endosperm expression) were detected in NtIDD10/20/23, and CAT-box (meristem expression) was discovered in NtIDD9/12/13/17/18/26.

Utilizing public RNA-Seq data, their expressions in nine tissues (Dry Capsule, Root, Stem, Young Leaf, Mature Leaf, Senescent Leaf, Immature Flower, Mature Flower, Senescent Flower) were investigated. In general, NtIDDs expressed throughout all the nine tissues ( Figure 6 ; Supplementary Table 4 ). Most of NtIDDs expressed in multiple tissues. The expression patterns for NtIDDs were distinct. Figure 7 shows that members with similar expression profiles were clustered together. NtIDD16/13/11 preferentially expressed in the immature flower. NtIDD9/21/6/24/3/25/4/22 expressed higher in leaf. NtIDD10/19 dominantly expressed in capsule. NtIDD12/2 expressed in all tissues, but still higher in capsule. NtIDD17/14/8/26 expressed more specifically in root. NtIDD23 expressed in all tissues, but the expression level in root was the highest. NtIDD20/18/7 expressed higher in stem. NtIDD5/15/1 expressed higher in root, stem, and young leaf.

Under conditions with six exogenous hormones, qRT-PCR expression levels of NtIDDs were measured. Compared with the controls, nine NtIDDs showed significant differential expressions under hormone treatments ( Figure 8 ). NtIDD6/11/12/21 differentially expressed under ABA treatment. NtIDD5/6/11/20/21 were disturbed by GA. NtIDD5/21 and NtIDD19/21 responded in CK and MeJA conditions, respectively. NtIDD5/21/24 were fluctuated by SA application. NtIDD10/11/20/21 were found mediated under SL treatment.

IDDs were reported functioning in leaf angle regulation. Earlier studies in Arabidopsis characterized the roles of AtIDD14, AtIDD15, and AtIDD16 in leaf angle regulation (Cui et al., 2013). Later studies pinpointed AtIDD4 acting as a direct regulator on AtIDD14 expression (Völz et al., 2019). Thereby, AtIDD4 was also implicated in regulating leaf angle. In order to investigate roles of NtIDDs in leaf angle regulation, the ortholog of AtIDD4 was identified and analyzed.

Based on phylogenetic relationship and sequence similarity, NtIDD9 was determined as the ortholog of AtIDD4. Expression studies in Arabidopsis showed that AtIDD14 specifically expressed in leaves, AtIDD15 mainly presented in leaf petioles and stems, and AtIDD16 and AtIDD4 highly expressed in leaves and other tissues. In tobacco, as shown above, NtIDD9 also exhibited high and preferential expression in the leaves. Thus, NtIDD9 may carry on conserved functions.

To further test its functions in leaf angle, RNA silencing was then performed and NtIDD9-RNAi transgenic plants were generated. Quantitative RT-PCR showed that the transcripts of NtIDD9 were dramatically reduced ( Figure 9A ) in the transgenic RNAi lines, suggesting successful silencing. Further examining the phenotype, it is noticed that the orientation angles of leaves were obviously increased in the RNAi plants ( Figure 9C ). Degree measurement showed that the average leaf angle in the wild type (WT) was 33.8°, whereas it expanded to 76.6° in the RNAi plants ( Figure 9B ). Therefore, the experiment validated NtIDD9 engaging in leaf angle regulation in tobacco.

In order to investigate the expression characteristics of NtIDD9 at the cellular level, RNA in situ hybridization technology was used for localization analysis. Results showed that a dense hybridization signal corresponding to NtIDD9 mRNA concentrated in the endodermis at the base of leaf petiole, which is the site that perceives gravity (Tasaka et al., 1999) ( Figure 10A ).

The subcellular localization of NtIDD9 protein was predicted by using the online tool Cell-PLoc (Chou and Shen, 2010), and it was found located in the nucleus. To determine the actual subcellular localization, NtIDD9 protein fused with green fluorescent protein (GFP) was transiently expressed in tobacco leaves. As shown in Figure 10B , the fluorescent signal of 35S::GFP-NtIDD9 fusion protein was restricted to the nucleus, suggesting that the NtIDD9 protein is indeed located in the nucleus.

Hormone levels were further quantified in wild-type and NtIDD9-RNAi plants. Utilizing LC-MS/MS, total 37 compounds belonging to seven hormones were quantified ( Supplementary Table 5 ). Further comparison analysis revealed that 11 compounds were significantly different between wild-type and NtIDD9-RNAi plants ( Figure 11 ). The concentrations of ABA (abscisic acid), JA (jasmonic acid), JA-ILE (jasmonoyl-L-isoleucine), and OPDA (cis(+)-12-oxophytodienoic acid) decreased in RNAi plants. In contrast, the levels of TRP (tryptamine), Indole (indole), and cZROG (cis-zeatin-O-glucoside riboside) increased in RNAi plants. In addition, ABA-ald (abscisic aldehyde), IAA-Ala (N-(3-indolylacetyl)-L-alanine), GA12-ald (gibberellin A12 aldehyde), and MeSAG (2-methoxycarbonylphenyl beta-D-glucopyranoside) were specifically detected in RNAi plants.