Revealing critical mechanisms involved in carbon nanosol-mediated tobacco growth using small RNA and mRNA sequencing in silico approach

Abstract

Nanomaterials have been shown to promote crop growth, yield and stress resistance. Carbon nanosol (CNS), a type of nanomaterial, is used to regulate tobacco shoot and root growth. However, information about the application of CNS to crop plants, especially tobacco, is still limited. Based on differential expression analysis and trend analysis, several miRNAs (miRN21-Novel-5p-mature, miR319b-Probable-5p-mature, miR160a-c-Known/Probable-5p-mature and miR156c-e-Known-5p-mature/star) and their target genes, including transcription factors (TFs), are likely responsible for the effect of CNS on promoting the growth of tobacco plants. In addition, we characterized nine TFs [Nitab4.5_00001789g0110 (NbbZIP), Nitab4.5_00001176g0010 (NbMYB), Nitab4.5_0001366g0010 (NbNAC), Nitab4.5_00000895g013 (NbMYB), Nitab4.5_0001225g0120 (NbNAC), Nitab4.5_0000202g0230 (NbDof), Nitab4.5_0002241g0010 (NbMYB-related), Nitab4.5_0000410g0060 (NbTCP), and Nitab4.5_0000159g0180 (NbC2H2)] associated with the response of tobacco to CNS according to the differential expression analysis, TF‒gene interaction network analysis and weighted correlation network analysis (WGCNA). Taken together, the findings of our study help understand CNS-mediated growth promotion in tobacco plants. The identification of candidate miRNAs and genes will provide potential support for the use of CNS in tobacco.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12870-024-05992-8.

Introduction

Nanomaterials are composed of 1–100 nm basic units in three dimensions, which is equivalent to a scale of approximately 10–1000 atoms [1]. Sustainable farming practices are able to take advantage of nanomaterials, a type of innovative agricultural fertilizer [2]. After adsorption into plants and subsequent transport to different tissues and organs, nanomaterials have been reported to positively influence plant growth and development [3]. Due to their unusual chemical characteristics, structures, and low biotoxicity, scientists are studying the effects of carbon nanosol (CNS), carbon nanotubes, graphene, carbon nanohorns, fullerenes, and carbon nanoparticles on plants [4, 5]. The physiological bases and mechanisms underlying the effects of carbon nanomaterials on promoting cucumber, tomato, rice, onion, maize, lettuce, tobacco, wheat and soybean growth and development have been preliminarily revealed [6–10]. This information will help to identify innovative approaches for agricultural sustainability.

Tobacco is an important cash crop worldwide, and its biomass is directly related to its economic value. Although the area planted with tobacco accounts for only approximately one-tenth of the total cultivated land in China, tobacco is one of the main cash crops and is essential for national revenue. The quality and yield of tobacco leaves are affected by many factors, including plant hormones [auxin (IAA), cytokinin (CK), abscisic acid, and gibberellin (GA)], plant nutrition and environmental stresses. In plants, an increase in the IAA content in roots promotes cell elongation and division, root hair formation, and main and lateral root growth [11]. CKs are crucial for the formation of the plant architecture, including plant height, axillary bud outgrowth and root hair growth [12, 13]. Exogenous GA application increases the height and photosynthesis of tobacco plants [14]. The potassium (K) content in tobacco leaves is usually regarded as one of the most important indices for evaluating tobacco quality [15]. Nitrogen (N) can increase the dry weight of tobacco roots, and high water–nitrogen irrigation is beneficial for promoting dry matter accumulation in whole tobacco plants [16]. Severe biotic and abiotic stresses inhibit tobacco growth, whereas appropriate stresses increase the accumulation of secondary metabolites, including nicotine, neophytadiene, ionone, and solanesol, in tobacco [17, 18].

In the context of tobacco, nanomaterials have been shown to enhance photosynthetic performance, cell growth and plant immunity [8, 9, 19]. In tobacco BY-2 cells, carbon nanomaterials induce the expression of K channel genes to increase K absorption [20]. Under cadmium toxicity, nanoparticles reprogram amino acids, alkaloids, and flavonoids to alleviate impaired tobacco growth [21]. A suitable CNS concentration facilitates the shoot and root growth of tobacco plants, significantly increasing the total leaf area, shoot dry weight and root length [22]. The ion (K, copper, zinc and boron) concentrations in tobacco seedlings obviously change after CNS treatment [22]. In primary roots, the net and dynamic flux rates of H⁺ and IAA clearly increase after CNS treatment, whereas those of GA decrease [22]. In addition, the transcriptional responses of genes involved in the above metabolic processes were identified after CNS treatment [22].

At the posttranscriptional level, microRNAs (miRNAs) are important regulators of plant growth, senescence, programmed cell death, and stress resistance [23–26]. For example, miR156 is involved in the temporal regulation of Arabidopsis shoot development [27]; miR395 targets WRKY to regulate apple resistance to leaf spot disease [28]; and in rice, miR159 is suppressed by brassinosteroids and targets the GAMYBL2 gene to regulate gibberellin biosynthesis and plant development [29]. However, the molecular regulatory mechanisms of miRNAs involved in CNS-mediated tobacco growth are not very clear. Transcription factors (TFs) are essential for initiating the transcription of mRNAs in eukaryotic plants, and they are usually located in the cell nucleus and bind to cis-acting elements in gene promoter regions [30, 31]. As transcriptional regulatory centres, TFs control many genes to regulate plant growth, development, secondary metabolism, and stress resistance [32]. For example, WRKY TFs are involved in mediating the crosstalk between plant growth and drought tolerance [33]. In Arabidopsis, the MADS-box TF family participates in flower development (ABCD model) [34]. In plants, bZIP TFs are involved in pathogen defence, light and stress signals, seed maturation and flower development [35]. However, systematic information on CNS-mediated tobacco growth and development is limited. The physiological and molecular mechanisms underlying CNS-mediated growth promotion in tobacco plants were preliminarily identified via integrated analyses of the phenome, ionome, transcriptome, and metabolome in our previous study [22].

Here, a comprehensive analysis of TFs, miRNAs and their targets was conducted through small RNA, degradome and transcriptome sequencing. In this study, our findings may help to elucidate the response of tobacco to CNS, and potential candidate miRNAs, genes and TFs involved in CNS-mediated tobacco yield were identified. Therefore, our findings may contribute to the application of CNS in agricultural production.

Methods

Plant materials

Tobacco (Nicotiana tobacum L., cv. K326) seedlings in uniform sizes were grown hydroponically in standard Hoagland and Arnon nutrient solution for 15 d. The physiological parameters of tobacco seedlings, including maximum root length, shoots and roots weight, root architecture, ion and phytohormone contents, were measured at 15 d after 0.3% CNS treatment. The shoots and roots of seedlings in control and CNS treatment groups were also sampled with three independent biological replicates for sRNA-seq, degradome and RNA-seq. The detailed description could be found in our previous study [22].

sRNA-seq, degradome and RNA-seq analyses

Total RNA was isolated with prechilled TRIzol (Takara Bio Inc., Kusatsu, Shiga, Japan). With an Agilent 2100 bioanalyzer (Santa Clara, CA, USA) and agarose gel electrophoresis, RNA integrity was assessed. Using T4 RNA ligase, 18–30 nt gel fragments were ligated to a pair of adapters. The small RNAs were transcribed into cDNAs, and polymerase chain reaction (PCR) was performed with the cDNA templates. With an Illumina HiSeq X Ten genome analyser (Illumina HiSeq 4000, San Diego, CA, USA), single-end small RNA libraries were sequenced by Novogene Corporation (Beijing, China). Using fastp software (v0.20.1) [36] with default parameters, the raw reads were trimmed and their quality was evaluated.

The indices of the tobacco reference genome, organelle, rRNA and gene coding sequences were built with the bowtie-build command in Bowtie software (v1.3.1) (https://github.com/BenLangmead/bowtie2/releases). For the small RNA transcriptome, high-quality reads were aligned with the tobacco reference genome (https://solgenomics.net/organism/Nicotiana_tabacum/genome), organelle and rRNA sequences using “One Step sRNAminer” in sRNAminer software [37]. The target genes of the miRNAs were predicted with “sRNA_target” in sRNAminer software [37]. The raw degradome data were downloaded from the Chinese Academy of Sciences (https://bigd.big.ac.cn/gsa.) with accession number CRA001801. Using CleaveLand software, miRNA cleavage sites were identified after the reads were mapped to the gene coding sequences [38]. A cut-off P value ≤ 0.05 was selected as the criterion for identifying significant cleavage sites. The identified cleavage sites in five different categories (0–4) were clustered according to the read abundance. Only one read supporting the cleavage site was in category 4 with minimum confidence; more than one read was grouped into categories 0–3, and category 0 shared maximum confidence.

HISAT2 software (v2.1.0) (https://daehwankimlab.github.io/hisat2/) was used to map high-quality reads to the tobacco reference genome. With SAMtools software (v1.6) [39], SAM (sequence alignment/map format) files were converted into BAM (binary alignment/map) files, after which the BAM files were sorted and indexed. Using a gene structure (gtf) file, StringTie (v1.3.3b) [40] software was employed to calculate the FPKM values of high-confidence genes. The DEGs [p < 0.05, false discovery rate (FDR) < 0.05 and |log2(fold-change)| ≥ 1] were identified with the R package “edgeR” [41].

Visualization, TF‒gene network construction, gene enrichment analysis and WGCNA

“HeatMap” in TBtools [42] was used to visualize the miRNA and gene expression profiles. The “Interactive Venn Graph” of TBtools was used to construct Venn diagrams. With the R package “Mfuzz” [43], miRNAs and RNAs were grouped into 12 clusters. The background files for GO and KEGG enrichment analyses were prepared with emapper.py (Python script, https://github.com/eggnogdb/eggnog-mapper) and “eggNOG-mapper Helper” in TBtools. Using the “GO Enrichment” and “KEGG Enrichment Analysis” functions of TBtools, GO and KEGG enrichment analyses were completed. “Enrichment Bar Plot” was used to visualize the results. In addition, gene function was also annotated through “Quick Protein Anno” in TBtools.

High-quality genes were analysed with WGCNA using the R wgcna package (v1.51) [44]. The correlations between gene modules and traits were determined by module–trait values (|r| ≥ 0.7 and P value ≤ 0.01). Among the significant modules, hub genes (with high weights and degrees) were identified. With a coefficient of variation, a minimal module size of 30 genes, a cut-off of 0.25, and a power β of 16, the Dynamic Tree Cut algorithm was used to reduce hierarchal clustering. Gene networks were constructed using Cytoscape software (v3.8.2) (https://cytoscape.org/) to identify hub genes. TF-gene networks were constructed using the “Plant TF Motifs Shift” and “Fimo: Binding Motif Scan” plugins of TBtools.

The gene promoters (2,000 bp) were submitted to PlantCARE (http://bioinformatics.psb.ugent.be/webtools/plantcare/html/) to obtain potential cis-elements. Specific cis-elements were also identified with the “Sequence Pattern Locate” function of TBtools. NbTF data were downloaded from the Plant Transcription Factor Database (https://planttfdb.gao-lab.org/). NbTFs were mapped to the tobacco protein sequences using the “Best ID Converter” function of TBtools.

Statistical analysis

The sRNA-seq and RNA-seq analyses were performed with three biological replicates, and the degradome analysis was performed with four replicates.

Results

Genome-wide transcriptional responses of miRNAs to CNS in tobacco plants

A total of 332 miRNAs with average TPM values greater than 5, including 52 novel and 280 known miRNAs, were identified in the control and CNS-treated samples (Fig. 1). In all the samples, the miRNAs showing the highest expression levels [> 10,000 TPM (transcripts per kilobase per million mapped reads)] were miR166a-m-Known-3p-mature, miR7122-Known-5p-mature, miR6149-Known-5p-mature and miR159b-Known-3p-mature. Nineteen miRNAs had higher expression levels (> 1,000 TPM), such as miR482a-Known-3p-mature, miRN5d-Novel-5p-mature, miRN5b-Novel-5p-mature, and miR168-Known-5p-mature. Seventy-nine miRNAs with high expression levels (> 100 TPM) were identified, and 144 miRNAs with moderate expression levels (> 10 TPM) were detected (Fig. 1). Many differentially expressed miRNAs (DEMs) were identified. Among them, the expression of several miRNAs, such as miR159b-Known-3p-mature and miR166a-k-Probable/Known-3/5p-mature, was induced by CNS in both the roots and shoots (Fig. 1a-d). Thus, more attention should be given to these DEMs. In addition, CNS upregulated or downregulated many miRNAs (i.e., miRN4-Novel-5p-mature, miRN5c-Novel-3p-star, miRN15-Novel-5p-mature, and miRN5d-Novel-5p-mature) either in roots or shoots (Fig. 1a-d).

Fig. 1.

The expression profiles of miRNAs in control and CNS-treated samples. a-d. Hierarchical clustering analysis of miRNAs whose expression was greater than 5 TPM

The putative targets of the above miRNAs were obtained through miRNA‒RNA sequence alignment and degradome sequencing (Tables S1–5). A Venn analysis revealed that 6 target genes were commonly identified using these two methods (Fig. 2a). The expression profiles of the miRNAs and their targets are shown in Fig. 2b. miR167-Known-3p-star and its target Nitab4.5_0002029g0040 presented opposite expression patterns. Based on the results of degradome sequencing, miR530b-Probable-3p-star, miRN24-Novel-3p-satr, miR167-Known-3p-star, miR7122-Known-3p-star, miR7757-Known-5p-star and miR5303as-Probable-3p-mature were predicted to bind to the sites in Nitab4.5_0000252g0110 (kinase interacting family protein NET1B), Nitab4.5_0004433g0060 (Probable LRR receptor-like serine/threonine-protein kinase IRK), Nitab4.5_0002029g0040 (PI-PLC X domain-containing protein), Nitab4.5_0001735g0080 (E3 ubiquitin-protein ligase UPL5), Nitab4.5_0003515g0010 (Homeobox-leucine zipper protein ATHB-15) and Nitab4.5_0003757g0090, respectively. Target plots were constructed to confirm the binding sites (Fig. 2c-h).

Fig. 2.

Analysis of miRNA targets. a Venn diagram showing the target genes of the miRNAs identified by sequence alignment and degradome sequencing. b Expression profiles of common target genes of miRNAs and corresponding miRNAs. c-h Target plots (t-plots) for the targets of miR530b-Probable-3p-star, miRN24-Novel-3p-satr, miR167-Known-3p-star, miR7122-Known-3p-star, miR7757-Known-5p-star and miR5303as-Probable-3p-mature confirmed by degradome sequencing

Trend analysis of the miRNA responses to CNS

The miRNAs were clustered into 12 profiles according to the trend analysis (Table S6 and Fig. 3a). Clusters 1 and 6/7 were significantly repressed and induced by CNS, respectively (Fig. 3a). According to the results of the GO enrichment analysis, the target genes of the miRNAs in Clusters 1, 6 and 7 were enriched in adenylyltransferase activity, DNA-binding transcription factor activity, nucleotidyltransferase activity, RNA biosynthetic process, regulation of cellular macromolecule biosynthetic process, and regulation of nitrogen compound metabolic process, among others (Fig. S1a). According to the KEGG enrichment analysis, the target genes were grouped into monobactam biosynthesis, selenocompound metabolism, environmental adaptation, and circadian rhythm-plant (Fig. S1b). miRNAs and their target genes usually show inverse expression patterns. Therefore, a clustering analysis was performed on the gene expression matrix (Table S7 and Fig. 3b).

Fig. 3.

Trend analyses of miRNAs and genes. a Clusters based on the expression levels of tobacco miRNAs identified via small RNA sequencing. b Clusters of tobacco genes identified via RNA sequencing based on the expression levels. The clusters that will be analysed in the future are shown in red

The genes in Cluster 1 were generally upregulated by CNS (Fig. 3b), and their expression profile contrasted with that of the miRNAs in Cluster 1 (Fig. 3a). The genes in Clusters 4/11 and the miRNAs in Clusters 6/7 presented opposite expression patterns (Fig. 3). A Venn diagram (Fig. 4a) shows the genes common to the targets of the miRNAs in Cluster 1 (Fig. 3a) and the genes in Cluster 1 (Fig. 3b). The genes common to the targets of miRNAs in Clusters 6/7 (Fig. 3a) and genes in Clusters 4/11 (Fig. 3b) are shown in Fig. 4b. The specific expression patterns of the miRNAs and their target genes are shown in Fig. 4c. Five miRNA‒target pairs, namely, miRN21-Novel-5p-mature/Nitab4.5_0000198g0140 (putative late blight resistance protein), miR530b-Probable-5p-mature/Nitab4.5_0000054g0100, miR5303bo/aq/ah/ba-Probable-3p-mature/Nitab4.5_0000647g0130 (pathogenesis-related gene), miR390b-d-Probable/Known-5p-mature/Nitab4.5_0000075g0030 (Oxygen-dependent coproporphyrinogen-III oxidase), and miR166h-Known-5p-star/Nitab4.5_0000989g0020 (glutamate receptor), which share obviously distant expression patterns, were identified. The expression of miRN21-Novel-5p-mature, miR530b-Probable-5p-mature, miR5303bo/aq/ah/ba-Probable-3p-mature and miR166h-Known-5p-star was induced by CNS, whereas the expression of miR390b-Known-5p-mature, miR390c-Known-5p-mature and miR390d-Probable-5p-mature was inhibited. However, their target genes presented opposite expression patterns.

Fig. 4.

Trend analyses of miRNAs and their targets. a Venn diagram showing the genes common to the targets of miRNAs in Cluster 1 and genes in Cluster 1. b Venn diagram showing the genes common to the targets of miRNAs in Clusters 6/7 and genes in Clusters 4/11. c Expression profiles of miRNAs and their target genes. The candidate miRNAs and their targets are labelled in red and green *

Identification of TFs among the targets of miRNAs

Transcription factors (TFs) interact specifically with cis-acting elements in gene promoter regions to regulate the initiation of gene transcription with specific strengths at specific times and locations [30, 31]. Therefore, TFs may be better candidate genes for understanding the process of plant growth and development.

A total of 2,777 TFs were identified in tobacco (Table S8). A total of 61 TFs were putative targets of the miRNAs identified in the control and CNS-treated samples (Fig. 5a). Among these TFs, approximately 24 TFs with at least 10 FPKM expression levels were found, whereas more than 10 TFs were not expressed (Fig. 5b). Seventy-three miRNAs were predicted to target these 61 TFs (Fig. 5c). Most miRNAs presented relatively high transcript levels (> 100 TPM), and fewer than 10 miRNAs presented low expression levels (< 10 TPM) (Fig. 5c). Multiple miRNA‒TF pairs that presented opposite expression patterns were identified. For example, miR319b-Probable-5p-mature was separately upregulated in CNS-treated roots and downregulated in CNS-treated shoots, whereas the expression of its target Nitab4.5_0000206g0010 (TCP2) was repressed by CNS in roots and induced by CNS in shoots; miR160a-Known-5p-mature, miR160b-Known-5p-mature and miR160c-Probable-5p-mature were expressed at low levels in CNS-treated samples, whereas the expression of their target Nitab4.5_0000071g0020 (ARF18) was induced by CNS; CNS inhibited miR156d-Known-5p-mature, miR156c-Known-5p-star and miR156e-Known-5p-star expression, while it stimulated the expression of their target Nitab4.5_0000861g0050 (SPL3).

Fig. 5.

Analysis of miRNAs and their target TFs. a Venn diagram showing the target TFs of miRNAs. b Expression profiles of the target TFs. c Expression profiles of the miRNAs targeting TFs

Genome-wide transcriptional responses of TFs to CNS in tobacco plants

A total of 294 and 315 DEGs responded to CNS in roots and shoots, respectively, and 29 DEGs were differentially expressed in both roots and shoots (Fig. 6a). We also identified the differentially expressed TFs (DETs) in response to CNS; 18 and 15 DETs were detected in the roots and shoots, respectively, and 2 common DETs were identified in these two organs (Fig. 6b). The expression profiles of DETs were investigated (Fig. 6c). The expression level of Nitab4.5_0000354g0040 was higher in CNS-treated roots (TRs) than in control roots (CRs), while its content was almost zero in shoots. Nitab4.5_0000048g0080, Nitab4.5_0000381g0130, Nitab4.5_0001372g0100, Nitab4.5_0000222g0080, Nitab4.5_0000249g0070, Nitab4.5_0000286g0080, Nitab4.5_0000171g0190, Nitab4.5_0001000g0160 and Nitab4.5_0000047g0140 were downregulated in the CNS-treated samples, and their expression levels were relatively high compared with those of the other DEGs. Two common DETs [Nitab4.5_0000508g0130 (NbWRKY40) and Nitab4.5_0002906g0080 (NbMYB62)] in roots and shoots were conformably repressed by CNS (Fig. 6c). Using all the DEGs, a TF–DEG network was constructed to identify key TFs regulating CNS-mediated tobacco growth (Fig. 6d). Four TFs, Nitab4.5_0000048g0080 (NbWRKY70), Nitab4.5_0000222g0080 (NbWRKY23), Nitab4.5_0000440g0010 (NbHHO5, a G2-like TF) and Nitab4.5_0000102g0300 (NbLUX, a G2-like TF), were likely candidate genes involved in CNS-mediated tobacco growth. CNS obviously increased the K concentration in tobacco shoots, and K likely participated in CNS-promoted tobacco growth [22]. In the network, a two-pore K channel gene (Nitab4.5_0000982g0120, NbTPK) was a potential target of Nitab4.5_0000048g0080 (NbWRKY70), and a W box element was identified in the promoter (-1131 bp) of NbTPK.

Fig. 6.

DEGs and DETs between control and CNS-treated tobacco seedlings. a Venn diagram showing the common DEGs between control and CNS-treated tobacco seedlings. b Venn diagram showing the common DETs between control and CNS-treated tobacco seedlings. c Expression profile of DETs in control and CNS-treated tobacco seedlings. d Regulatory network of the TF-mediated CNS response in tobacco

Identification of weighted gene coexpression network analysis (WGCNA) modules and hub genes associated with target traits

Using all genes with average expression levels greater than 0 fragments per kilobase per million (FPKM) values, WGCNA was performed to analyse the significant associations between gene sets (modules) and phenotypes. In our previous study, IAA, GA and K were likely involved in CNS-mediated tobacco growth [22]. Therefore, IAA, GA and K were set as target traits related to the gene modules.

A power value of 16 was selected to construct a gene cluster scheme as the basis of a WGCNA (Figure S2). Twenty-five modules were identified by WGCNA, with 26,524 nonredundant genes (Fig. 7a). An analysis of module‒trait relationships revealed that target traits (IAA, GA and K contents) were highly related to the “dark red”, “dark turquoise”, “light green” and “purple” modules (Fig. 7b). The “dark red” (r = -0.71, p = 0.01), “light green” (r = -0.9, p < 0.01) and “purple” (r = -0.71, p < 0.01) modules were negatively correlated with the GA content after CNS treatment; the “dark turquoise” module was negatively correlated with the K level (Fig. 7b). The genes in these modules were significantly enriched in signalling receptor activity, growth, developmental growth, cell growth, developmental growth involved in morphogenesis, and root system development (Figure S4a). In addition, messenger RNA biogenesis, cytoskeletal proteins, membrane transport, cellular processes, and the biosynthesis of unsaturated fatty acids were also overrepresented (Figure S4b).

Fig. 7.

WGCNA of genes identified in the CNS-treated and control tobacco samples. a Hierarchical clustering tree illustrating 25 modules of coexpressed genes. b Module–trait relationships presenting the significance of the module eigengene correlations with traits. The left panel provides a colour key to the 25 modules. The colour scale on the right shows the module-trait correlation ranging from − 1 (green) to 1 (red)

In the “purple” module, 11 tobacco TFs (one M-type NbMADS, NbC2H2, NbG2-like, NbbZIP and NbNAC; two NbHD-ZIPs, NbERF, and NbbHLH) were identified as important genes (Fig. 8a). The GA biosynthesis gene NbGA₂₀OX₃ (Nitab4.5_0000662g0210) was strongly associated with NbbZIP (Nitab4.5_0001789g0110) (Fig. 8a). In the “dark red” module, NbNAC and NbMYB were CNS-responsive genes, and the GA biosynthesis gene NbGA₂OX₁ was likely associated with the core genes (Fig. 8b). Eight hub genes [Nitab4.5_0000895g0130 (NbMYB), Nitab4.5_0001225g0120 (NbNAC), Nitab4.5_0000202g0230 (NbDof), Nitab4.5_0002241g0010 (NbMYB-related), Nitab4.5_0000410g0060(NbTCP), Nitab4.5_0000371g0030 (NbTrihelix), Nitab4.5_0003063g0040 (NbDof) and Nitab4.5_0000412g0010 (NbERF)] were screened from the “light green” coexpression gene network map, and Nitab4.5_0000662g0210 (NbGA₂₀OX₃) may be their potential target gene (Fig. 9a). NbMYB (Nitab4.5_0002293g0010) was identified as a critical gene in the “dark turquoise” module (Fig. 9b).

Fig. 8.

Hub genes identified in the “purple” and “dark red” modules. Cytoscape representation of the relationships of genes in the “purple” (a) and “dark red” (b) modules. Genes with large and small degrees are indicated with large purple and small green circles

Fig. 9.

Hub genes identified in the “light green” and “dark turquoise” modules. Cytoscape representation of the relationships of genes in the “light green” (a) and “dark turquoise” (b) modules. Genes with large and small degrees are indicated with large purple and small green circles

NbAP2 (Nitab4.5_0001003g0080) was another key TF identified in the “dark red” module (Fig. 10a). In the “purple” module, NbMYB (Nitab4.5_0003656g0040) and NbbHLH (Nitab4.5_0000475g0100) were identified as hub TFs (Fig. 10b). In the “dark turquoise” module, K transporter-related genes [NbHAK26 (Nitab4.5_0000485g0100) and NbKEA4 (Nitab4.5_0004087g0050)] were associated with the hub gene NbC2H2 (Nitab4.5_0000159g0180) (Fig. 10c). An NbDof (Nitab4.5_0000202g0230) TF in the “light green” module was identified as an important gene (Fig. 10d).

Fig. 10.

Cytoscape representation of the relationships of genes in the “purple” (a), “dark red” (b), “light green” (c) and “dark turquoise” (d) modules

Discussion

CNS is an important innovative nanoparticle that optimizes water and nutrient absorption in crops and thereby promotes crop growth and yield [44]. Many nanomaterials can absorb water and nutrients, which help break down organic matter in the soil and increase plant root vigour [46]. The crop biomass, fruit yield, root length and chlorophyll content all increase after the addition of nanometre-scale materials [47, 48]. Engineered nanomaterials have been reported to promote tobacco cell growth and disease resistance [9, 49, 50]. In general, plant growth was obviously promoted by nano-materials. The understanding of the molecular mechanism of tobacco in response to CNS has progressed. However, information on the miRNA‒target and TF–target regulatory modules involved in CNS-stimulated tobacco growth and development is limited. In the present study, the expression profiles of miRNAs and their target genes and of TFs and their target genes in response to CNS treatment were systematically characterized through gene expression analysis, differential expression analysis, miRNA‒gene network analysis, TF‒gene network analysis and WGCNA.

miRNAs and their targets play key roles in CNS-mediated tobacco growth

The expression patterns of miRNAs were determined to explore their functions. A number of known and novel miRNAs have been characterized in CNS-treated tobacco, some of which were differentially expressed after CNS treatment (Fig. 1). The putative targets of miRNAs, especially DEMs, were characterized (Fig. 2). Among the above miRNA‒target pairs, miR167-Known-3p-star and its target Nitab4.5_0002029g0040 may help elucidate the underlying mechanisms by which the CNS promotes tobacco growth and should receive increased attention. The other five miRNA‒target pairs, which presented opposite expression patterns, were identified via a trend analysis (Figs. 3 and 4). These five miRNA‒target pairs may also play essential roles in the CNS response, and the roles and relationships of novel miRNAs (miRN21-Novel-5p-mature) and Nitab4.5_0000198g0140 (putative late blight resistance protein homolog R1B-16) should be validated preferentially. Among tobacco TFs, the targets of miRNAs were also determined (Fig. 5a). The transcriptional profiles of many of the miRNA–TF pairs differed (Fig. 5b and c). Considering the important biological functions of TFs, miR319b-Probable-5p-mature/Nitab4.5_0000206g0010 (TCP2), miR160a-c-Known/Probable-5p-mature/Nitab4.5_0000071g0020 (ARF18), and miR156c-e-Known-5p-mature/star/Nitab4.5_0000861g0050 (SPL3) should be the most critical candidate genes for elucidating the diverse mechanisms involved in CNS-promoted tobacco plant growth. Nitab4.5_0000206g0010 (NbTCP24) is a homologous gene of AtTCP24 that is involved in the regulation of leaf differentiation and hypocotyl elongation [50, 51]. In addition, a relationship between miR319 and AtTCP24 has been identified [53]. Nitab4.5_0000071g0020 was annotated as NbARF16, and AtARF16 is involved in root cap cell differentiation under the control of miR160 [54, 55]. The miR156 target AtSPL3 (a homologue of Nitab4.5_0000861g0050) is involved in the regulation of Arabidopsis lateral root development [56]. Therefore, the similar functions of miRNA and their targets from model plant and tobacco may be predicted. These findings suggest that these four TFs and their corresponding miRNAs play essential roles in the response of tobacco to CNS treatment. In addition, the other miRNAs and their target genes (likely miR167/Nitab4.5_0002029g0040 and miRN21/Nitab4.5_0000198g0140) might interact with CNS.

TFs sharing key roles in CNS-mediated tobacco growth

In our present study, the DEGs associated with the response to exogenous CNS application were analysed, and they were enriched in Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) terms [22]. TFs take part in complex transcription initiation of genes, and they are located in cell nucleus to interact specifically with genes promoter regions at specific times and locations. In addition, plants growth, development, secondary.

metabolism, and stress resistance were controlled by TFs. However, the functions of TFs in CNS-mediated growth have not been studied systematically. In this study, TFs were investigated. The differential expression analysis revealed that 31 DETs, including 2 DETs (NbWRKY40 and NbMYB62) identified in both roots and shoots, may be involved in the CNS-promoted growth of tobacco plants (Fig. 6b). GA₃ was present at markedly higher concentrations in CNS-treated roots than in control roots [22]. In Arabidopsis, AtMYB62 has been shown to participate in gibberellic acid biosynthesis [57]. Therefore, the CNS-induced GA profile revealed a distinct mechanistic response to CNS in the roots, which was likely related to NbMYB62. In the TF–DEG network, two NbWRKY and NbG2-like genes were the hub genes and were predicted to interact with the K transport-related gene NbTPK (Fig. 6d). In tobacco shoots, CNS increased K concentrations [22]. The genes in above network were annotated in Figure S3. Therefore, the role of the NbWRKY–NbTPK pathway requires further investigation and exploration of potential regulatory mechanisms in future work.

WGCNA is widely used to classify genes with similar expression patterns, cluster the relationships between modules and phenotypes, and reveal the associations between target traits and genes [44]. IAA, GA and K are predicted to participate in tobacco shoot and root growth following CNS treatments [22]. Three modules (“dark red”, “light green” and “purple”) were involved in GA metabolism, and one module (“dark turquoise”) was strongly correlated with the K level (Fig. 7b). In the “purple” module, NbHD-ZIP (Nitab4.5_0001090g0010) may function in leaf development (Fig. 8a) based on the function of AtHD-ZIP [58]. Similar to OsPRE6 and AtPRE6, NbbHLH (NbPRE6, Nitab4.5_0000785g0050) may mediate brassinosteroid regulation of cell elongation and plant development [59]. The interactions between NbbZIP (Nitab4.5_0001789g0110) and NbGA₂₀OX₃ (a GA biosynthesis gene) were characterized (Fig. 8a), and another GA metabolism gene, NbGA₂OX₁ (Nitab4.5_0001533g0010), was closely correlated with NbNAC (Nitab4.5_0001366g0010) and NbMYB (Nitab4.5_0001176g0010) in the “dark red” module (Fig. 8b). In addition, eight hub TFs, including NbMYB, NbNAC, NbDof, NbMYB-related, NbTCP, NbTrihelix and NbERF, were likely to interact with NbGA₂₀OX₃ (Nitab4.5_000662g0210) in the “light green” module (Fig. 9a). In tobacco roots, the GA₃ level clearly changed after CNS treatment [22], and GA is well known to control plant growth and seedling vigour [60]. In the above modules, genes were enriched in growth and development processes (growth, developmental growth, cell growth, developmental growth involved in morphogenesis, cell morphogenesis, cellular developmental process and root system development) (Figure S4a). This result also demonstrated the importance of genes in the above four gene modules.

TF‒gene networks were also constructed to identify the hub genes regulating traits. In the “dark red” module, the hub gene AP2 was identified, and it was possibly involved in lateral root formation based on previous research (Fig. 10a) [61]. NbbHLH (Nitab4.5_0000475g0100), a key gene in the “purple” module, may participate in seedling growth, according to reported information (Fig. 10b) [62]. In the K-related module (“dark turquoise”), the STOP1 homologue NbC2H2 (Nitab4.5_0000159g0180) was predicted to target K-related transporter genes (NbHAK26 and NbKEA4) (Fig. 10c) [63]. The hub gene NbDof (a homologous gene of AtDof2.4) in the “light green” module is likely to regulate root development (Fig. 10d) [64].

Conclusions

In general, the candidate regulators of CNS-promoted growth of tobacco plants were revealed. By predicting miRNA targets and performing a trend analysis, differential expression analysis, TF–DEG network analysis and WGCNA, the responses of miRNAs and their targets (including TFs) to CNS were investigated. The potential miRNA‒target pairs [miRN21-Novel-5p-mature/Nitab4.5_0000198g0140 (NbR1B-16), miR319b-Probable-5p-mature/NbTCP24, miR160a-c-Known/Probable-5p-mature/NbARF16, and miR156c-e-Known-5p-mature/star/NbSPL3] and TFs [Nitab4.5_00001789g0110 (NbbZIP), Nitab4.5_00001176g0010 (NbMYB), Nitab4.5_0001366g0010 (NbNAC), Nitab4.5_00000895g013 (NbMYB), Nitab4.5_0001225g0120 (NbNAC), Nitab4.5_0000202g0230 (NbDof), Nitab4.5_0002241g0010 (NbMYB-related), Nitab4.5_0000410g0060 (NbTCP), and Nitab4.5_0000159g0180 (NbC2H2)] may be crucial for regulating the CNS-mediated growth of tobacco plants. The mechanisms mediated by CNS may be controlled by these miRNAs and genes through hormone metabolism, nitrogen metabolism, and stress adaption. The roles of these genes should be identified in the future, and the findings will provide possibilities for the use of CNS to improve tobacco growth and yield.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Abbreviations

CNS

Carbon nanosol

DEG

Differentially expressed gene

DETs

Differentially expressed TFs

FPKM

Fragments per kilobase per million

GA

Gibberellin

IAA

Auxin

K

Potassium

TFs

Transcription factors

TPM

Transcripts per kilobase per million mapped reads

WGCNA

Weighted correlation network analysis

Author contributions

CW, TZ, and QSC conceived the project and were responsible for acquiring funding; XAZ and XZQ performed the experiments; YLX, KLZ, and PJC analysed and interpreted the data. TZ and CW wrote the paper with significant input from all the other authors.

Funding

This study was financially supported by the Chief Scientist Innovation Project of the State Tobacco Monopoly Administration/China National Tobacco Corporation (902022CK0560) and the Science and Technology Project of CNTC (110202201004 (JY-04)). No conflicts of interest between the funding and the funding company exist.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethical approval and consent to participate

This article does not contain any studies with human participants or animals performed by any of the authors. No specific permits were needed. All methods were in compliance with relevant institutional, national, and international guidelines and legislation.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.