24-epibrassinolide confers tolerance against deep-seeding stress in Zea mays L. coleoptile development by phytohormones signaling transduction and their interaction network

ABSTRACT

Coleoptile/mesocotyl elongation influence seedling emergence and establishment, is major causes of maize deep-seeding tolerance (DST). Detailed analyses on molecular basis underlying their elongation mediated by brassinosteroid under deep-seeding stress (DSS) could provide meaningful information for key factors controlling their elongation. Here we monitored transcriptome and phytohormones changes specifically in elongating coleoptile/mesocotyl in response to DSS and 24-epibrassinolide (EBR)-signaling. Phenotypically, contrasting maize evolved variant organs to positively respond to DST, longer coleoptile/mesocoty of K12/W64A was a desirable organ for seedling under DSS. Applied-EBR improved maize DST, and their coleoptiles/mesocotyls were further elongated. 15,607/20,491 differentially expressed genes (DEGs) were identified in W64A/K12 coleoptile, KEGG analysis showed plant hormone signal transduction, starch and sucrose metabolism, valine, leucine, and isoleucine degradation were critical processes of coleoptile elongation under DSS and EBR signaling, further highly interconnected network maps including 79/142 DEGs for phytohormones were generated. Consistent with these DEGs expression, interactions, and transport, IAA, GA₃, ABA, and Cis-ZT were significantly reduced while EBR, Trans-ZT, JA, and SA were clearly increased in coleoptile under DSS and EBR-signaling. These results enrich our knowledge about the genes and phytohormones regulating coleoptile elongation in maize, and help improve future studies on corresponding genes and develop varieties with DST.

Introduction

Drought is the primary abiotic constraint to global crop production, and moisture deficit is expected to become worse along with the general temperature increase in next few decades. Maize (Zea mays L.) is the third most consumed grain in the world. The soil water content of 0 ~ 10 cm soil layer is accounting for approximately 15% when seeds is planted in arid and semiarid areas, consequently, maize is constantly confronted with drought stress from seed germination to seedling emergence period, and causing great yield loss in these regions.^1,2 Fortunately, the deep-seeding of maize seeds is an effective measure to ensure seeds absorbing water from deep soil layer and germinate normally in these regions.^3,4 Consequently, untying the molecular basis of maize deep-seeding tolerance, to improve early seedling establishment in such arid and semi-arid regions, will be a pertinent in maize breeding programs.

Previous studies showed when seeds are located deep from the soil surface, some organs, e.g. the first internode and coleoptile in Triticum aestivum L. and Hordeum vulgare L.,^5,6 and the mesocotyl in Oryza sativa L.,⁷ could markedly elongate to reach the soil surface. In contrast, the elongated mesocotyl could pushes the coleoptile (a conical structure) across the soil surface in maize during germination,^2,8 whereas the mesocotyl was inhibited by light as soon as the coleoptile sprouts from the soil surface.⁹ During seedling emergence, the leaf elongation keeps pace that of the coleoptile, following the coleoptile arrests its and opens at its apex, wherefrom leaves emerge.⁸ In this regard, the collaborative elongation of both mesocotyl and coleoptile in maize contributes to seedling emergence from the deep soil layer.

What are the mechanisms that regulate the elongation of coleoptile and mesocotyl when maize seeds are subjected to deep-sowing pressure? Increasing evidences suggested that the phytohormones as growth regulators were involved in cell proliferation or elongation to control coleoptile or mesocotyl elongation in a wide range of crop species.^1,2,7,10 The indole-3-acetic acid (IAA) applied to promote cell elongation and further induce rapid mesocotyl elongation in maize under deep-seeding stress and even this process was regulated by an auxin-binding protein (ABP1).¹¹ Maize mesocotyl elongation was promoted by exogenous gibberellic acid (GA) stimulation under deep-seeding stress,¹² and ZmMYB59 played a negative regulatory role in maize seed germination in deep soil and the regulation was involved in GA signaling pathway.¹³ The abscisic acid (ABA) in maize mesocotyl was accumulated at the 4th day of seed germination,² further QTL-analysis showed that the phytochromeB2 (phyB2) was detected in Bin 9.03 that responsible for mesocotyl length and ABA accumulation.¹⁴ Moreover, the growth-limiting proteins in maize coleoptiles model and the auxin-brassinosteroid (AUX-BR) hypothesis of mesocotyl elongation showed maize mesocotyl elongation commonly depended on AUX and BR,¹⁵ as well as the expression of MSTRG.30577 (brassinosteroid insensitive 2, BIN2) changed in parallel with the maize mesocotyl elongation under deep-seeding conditions.² However, the coleoptile elongation regulated by phytohormones signaling under deep-seeding stress has been somewhat less studied, which may be equally important in the elongation of maize mesocotyl under deep-seeding stress. Herein, the transcriptomic features in both contrasting maize inbred lines were investigated following culture for 10 days after receiving 4.16 × 10⁻³ M exogenous 24-epibrassinolide (EBR) under 3 or 20 cm seeding depths using RNA-sequencing (RNA-Seq) analysis, to elucidate the dynamic molecular mechanisms underpinning the deep-seeding stress responses in maize coleoptile, and in addition to providing a basis for further targeted cloning and transgenic studies.

Materials and methods

Plant materials and sampling

One deep-seeding tolerant inbred lines W64A and one intolerant inbred lines K12 were used in this study.^2,16 The uniform and plump seeds of the two inbred lines were surface sterilized with 0.5% sodium hypochlorite for 10 min and then rinsed with double distilled water five times. The sterilized seeds were soaked in 0 and 4.16 × 10⁻³ M EBR (CAS: 78821–43-9, Sigma-Aldrich, USA) solution for 24 h under dark environment, respectively. The different concentrations of the EBR solution was initially dissolved in ethanol (98%, v/v) and made up with double distilled water containing 0.1% (v/v) Tween 20 (CAS: 9005–64-5, Sigma-Aldrich, USA) as an adhesive agent. The two EBR solution and dry vermiculite were mixed in proportion (1 mL: 5 g) as the two culture substrate, respectively, which was put into the PVC tubes (50 cm height, 17 cm diameter) and reached a certain PVC tubes height, then 30 soaked seeds were evenly planted onto the culture substrate surface and then covered with the 3 and 20 cm depth culture substrate. Later on, that were incubated for 10 days in a greenhouse (22 ± 0.5°C with 12/12 h light/dark cycle, and 60% relative humidity), and 40 mL exogenous EBR solution of 0 or 4.16 × 10⁻³ M concentrations was added to corresponding culture substrate at 2 day intervals. The research included three treatments, i.e. 3 cm seeding depth (abbr., 3 cm), 20 cm seeding depth (abbr., 20 cm), and 4.16 × 10⁻³ M EBR induction at 20 cm seeding depth (abbr., EBR+20 cm), and each treatment was conducted in three replicates. Then the prepared coleoptiles were collected, frozen in liquid nitrogen, and stored at −80°C for subsequent RNA extraction, as well as the growth parameters and phytohormones of their mesocotyl and coleoptile were evaluated.

The growth parameters and phytohormones level assay

The growth parameters, including coleoptile length, mesocotyl length, and length ratio of coleoptile to mesocotyl of 10-day-old seedlings were investigated under three treatments via ruler. Then 0.5 g coleoptile or mesocotyl was ground in liquid nitrogen and digested in 5 mL methanol-formic acid solution (99:1, v:v) for 12 h at 4°C and then centrifuged at 12,000 rpm at 4°C for 20 min, and the supernatant was collected. The residue composition was further digested in 5 mL methanol-formic acid solution (99:1, v:v) and recentrifuged as described above. Then the supernatant were pooled. After pigment removed by Cleanert ODS C18 plastic column, which was dried by nitrogen flow at 25°C, and was further dissolved using 1 mL methanol. Finally the solution was filtered with a 0.22 μM membrane filter, and 5 μL was injected for analysis. The level of IAA, GA₃, ABA, trans-/cis-zeatin (Trans-/Cis-ZT), jasmonic acid (JA), and salicylic acid (SA) were analyzed by high performance liquid chromatography (HPLC), and EBR was analyzed via ultrahigh performance liquid chromatography-tandem mass spectrometry (UHPLC-MS/MS). Standards of IAA (CAS: 87–51-4, Sigma-Aldrich, USA), GA₃ (CAS:77–06-5, Sigma-Aldrich, USA), Trans-ZT (CAS: 1637–39-4, Sigma-Aldrich, USA), Cis-ZT (CAS: 32771–64-5, Sigma-Aldrich, USA), ABA (CAS:21293–29-8, Sigma-Aldrich, USA), JA (CAS: 77026–92-7, Sigma-Aldrich, USA), SA (CAS: 69–72-7, Sigma-Aldrich, USA), and EBR (CAS: 78821–43-9, Sigma-Aldrich, USA) were used to optimize the mass spectrometric parameters and fragment spectra.

RNA isolation and Illumina sequencing

Total RNA was extracted from each sample using commercial kits (TRIZOL reagent, Invitrogen, Carlsbad, CA, USA) according to the manufacturer’s protocol. Total RNA was extracted from each coleoptile sample and DNA was digested using DNase. Eukaryotic mRNA was enriched using magnetic beads with Oligo (dT), and mRNA was fragmented using an interrupting reagent. The mRNA was used as a template for single-strand cDNA synthesis using random hexamer primers. Next, second strand synthesis was performed and the resultant double-stranded cDNA was purified using a kit (Invitrogen, Carlsbad, USA). The purified double-stranded cDNA was subjected to end repair and a tail was added. Sequencing adaptors were ligated, fragment size was selected, and PCR amplification was performed. The constructed library was quantified using an Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA) and sequenced using an Illumina NovaSeq PE150 sequencer by the Nanjing Genepioneer Biotechnologies Company (Nanjing, Jiangsu, China) to generate 150 bp paired-end reads.

Assembly, data analysis, and functional annotation

Some low-quality reads and reads containing adapters or poly-N were removed from the raw data to generate cleaned reads. The cleaned reads were aligned to the maize reference genome (Zea_mays B73_V4; ftp://ftp.ensemblgenomes.org/pub/plants/release-46/fasta/zea_mays/dna/) using Tophat2. The aligned reads of each sample were assembled using Cufflinks. To produce nonredundant transcripts, assembled transcripts from three biological replicates were merged using Cufflinks. The expression levels of merged transcripts were counted and fragments per kilobase of transcript per million fragments sequenced (FPKM) values were calculated as by Trapnell et al.¹⁷ Differentially expressed genes (DEGs) were analyzed using the DESeq R Package in Bioconductor (http://www.bioconductor.org/), and significant DEGs among different samples were identified as having a threshold level P value < 0.05, FDR < 0.001, and log2 fold-change > 1. Further analysis of DEGs, including gene ontology (GO) enrichment analysis (http://bioinfo.cau.edu.cn/agriGO/), cluster of orthologous groups of proteins (COG) analysis (https://www.ncbi.nlm.nih.gov/COG/), nonredundant (Nr) annotation (http://www.ncbi.nlm.nih.gov/pubmed), and Kyoto Encyclopedia of Genes and Genomes (KEGG) (http://www.genome.jp/kegg/) were performed, and phytohormones interactome network mapping was analyzed by Cytoscape 3.8.2 software (https://cytoscape.org/).

Quantitative real-time PCR (qRT-PCR) analysis

Fifteen DEGs were selected for qRT-PCR analysis. Purified RNA was reverse-transcribed into cDNA using a first strand cDNA synthesis Kit according to the manufacturer’s instructions (TaKaRa, Japan). qRT-PCR analysis was conducted using TransStart Tip Green qPCR SuperMix (TRAN, Beijing, China), and primers were designed using Primer premir 5.0 and listed in Table 1. The cycling parameters of PCR were optimized in Supplementary Table 1. The relative expression level was estimated using the (2^−ΔΔCt) quantification method with maize Actin1 (Zm.010159) serving as the endogenous control.¹⁸

Table 1.

The quantitative real-time PCR (qRT-PCR) primer sequence

Gene name	Gene ID	Forward primer (5ʹ to 3ʹ)	Reverse primer (5ʹ to 3ʹ)
IAA7	Zm.016277	GGATTATGAAGGCGATAGGGTG	GGCCAGGAGGTGCTGTTAGA
Histidine-containing phosphotransfer protein 4	Zm.037694	GGCGAACATAGAGCAAGCAC	GAGGGAAATAGGAAGAGTCGTCAG
CKX12	Zm.002989	GACGACGGCACCAACAAGA	CCGACACCAGGTCCACAACA
Putative cytochrome P450 superfamily protein	Zm.020418	TTATCGACGAGTGCAAGACG	AGCTTGGAGAGGTCGTCGTA
ciszog1	Zm.000237	TGGGAGAAACACGCTGAAAT	CGAGATCCTTACGCGAGTTC
PP2C7	Zm.009747	GTGCTCCCAACCTTCTTCCA	CCAGTACCTCGTGCATTCTGTC
SnRK2.3	Zm.029975	AGGGCGATTCAGTGAAGATGA	TTTGGAATGCAGCAAAGCAG
Hypothetical protein	Zm.048093	TTTCCTTGCCAGGGATTCC	TCTATTTGCTCCTCCTCGACC
DELLA protein SLN1-like	Zm.044065	GATGCTGGTTGGGCTCTTCT	TCCTGCTCCTGCTGTCACTCT
TIFY25	Zm.020614	CCGACCACCATGAACTTGC	AACCTGTCGAAGACGACCACT
BIN2	MS.30577	ACGCCCATTTCCTCCACTAT	AGCAGCCCAACATTAAAACG
cyclin11	Zm.005293	TGTCACGCCCTTCTCCTACC	ACACTCGGCGGTCTTGTCCT
protein ETHYLENE-INSENSITIVE 3-like 2	Zm.022530	GCGGCTTGGACTTGGATT	TGCGCTCTGCTGTATCACATT
MKK4 – putative MAPKK	Zm.018326	CGCATCAACACCGACCTCA	CGGTTCACCGGGTTCTTCT
Bzip70	Zm.046751	GCTCCAAGGCAGCAATCAA	TTCTTGCCATCTGCATCTCC
Actin-1	Zm.010159	CGATTGAGCATGGCATTGTCA	CCCACTAGCGTACAACGAA

Results

Response of maize coleoptile to deep-seeding stress and exogenous EBR stimulation

Here’s a very interesting phenomenon, compared to traditional sowing depth of 3 cm, the seeds of both inbred lines were cultured 10 days under 20 cm deep-seeding treatment, their coleoptile and mesocotyl were significantly elongated (P < .05; i.e. the coleoptile length and mesocotyl length of W64A/K12 were 2.13/2.17 cm and 5.67/2.20 cm at 3 cm sowing depth, respectively, and which even grow to 3.17/6.07 cm and 11.77/3.57 cm at 20 cm deep-seeding stress, respectively) (Figure 1a–d), coleoptile length, however, in intolerant K12 was clearly larger than that of its mesocotyl length under 20 cm deep-seeding stress, instead, which in deep-seeding tolerant W64A was obviously shorter than that of its mesocotyl length, then resulting in significant difference on length ratio of coleoptile to mesocotyl in W64A and K12, that were 0.27 and 1.69, respectively (Figure 1a–e). Suggesting that the coleoptile and mesocotyl elongated positively for the emergence of maize seeds under deep-seeding stress; however, why is the elongation degree of coleoptile and mesocotyl clearly different both W64A and K12 under deep-seeding stress? which needs further study. Like growth with coleoptile and mesocotyl of the W64A and K12 at 20 cm sowing pressure, their seeds were further incubated for 10 days with 4.16 × 10⁻³ M exogenous EBR under 20 cm planting-seeding stress, the length of coleoptile (i.e. 3.57/6.40 cm) and mesocotyl (i.e. 13.63/5.17 cm) in W64A/K12 were further elongated 1.13/1.05 and 1.16/1.45 times, respectively (Figure 1a–d). Indicating that the exogenous EBR stimulation could promote longitudinal elongation of maize coleoptile and mesocotyl to achieve normally seedling emergence under deep soil layer.

Figure 1.

Coleoptile elongation characteristics, phytohormones level, and their relationships between W64A and K12 under three treatments [3 cm seeding depth (3 cm), 20 cm seeding depth (20 cm), and 4.16 × 10⁻³ M EBR stimulation under 20 cm stress (20 cm+EBR)]. Performance of coleoptile in W64A and K12 under three treatments (a). Performance of mesocotyl in W64A and K12 under three treatments (b). Statistics of coleoptile length (c), mesocotyl length (d), and length ratio of coleoptile to mesocotyl (e) under three treatments. Different lowercase letters with a single inbred line among three treatments indicated a significant difference with P < 0.05. Polar coordinate graph for phytohormones [indole-3-acetic acid (IAA), gibberellic acid (GA₃), 24-epibrassinolide (EBR), cis-zeatin (Cis-ZT), trans-zeatin (Trans-ZT), jasmonic acid (JA), and salicylic acid (SA)] ratio of coleoptile to mesocotyl in both inbred lines under three treatments (f). Correlation coefficient diagram between coleoptile length and eight phytohormones level in both inbred lines under three treatments (g). Correlation coefficient diagram between length ratio of coleoptile to mesocotyl and eight phytohormones ratio of coleoptile to mesocotyl in both inbred lines under three treatments (h)

Transcriptome analysis with Zea mays reference genome

RNA-Seq was performed on 18 libraries of coleoptiles in W64A and K12 under three treatments, i.e., 3 cm sowing depth, 20 cm seeding depth, and 4.16 × 10⁻³ M exogenous EBR induction under 20 cm plant depth (three biological replicates for each sample), respectively, and compared to DEGs that respond to deep-seeding stress and exogenous EBR stimulation. An average of 50,842,839 reads were obtained of individual sample, with a Q20/Q30 quality score ≥ 98.09%/94.40%, while the average GC content of each sample was approximately 54.31% and 84.73 ~ 99.25% of clean reads mapped to the Zea_mays B73_V4 reference genome (ftp://ftp.ensemblgenomes.org/pub/plants/relese-46/fasta/zea_mays/dna/.) (Supplementary Table 2). Multiple mapped clean reads in each library were excluded from further analysis. Moreover, principal component analysis (PCA) of the RNA-Seq data for all samples showed that the experiment was reliable and the sample selection was reasonable (Supplementary Figure 1a). The FPKM density profile was a nonstandard normal distribution with a regional area size of 1, representing a sum of approximately 1 for probability (Supplementary Figure 1b).

Maize DEGs responses to deep-seeding stress and exogenous EBR treatment

In the study, FPKM values ≥ 1, log2 fold-change ≥ 1, and criteria of adjusted P value < 0.05 were used to determine genes expressed. Totally, 15,607 and 20,491 DEGs were identified on both W64A and K12 from the three treatments, respectively (Figure 2a). Moreover, the venn diagram (Figure 2b–c) showed the number of up- and down-DEGs exclusively expressed in different treatments, overlapping DEGs among all treatments. Whether 4.16 × 10⁻³ M exogenous EBR application under 20 cm seeding depth, 213 (1.45%) common up-DEGs of coleoptile in deep-seeding tolerant W64A and intolerant K12 were detected, and 121 (0.07%) common down-DEGs of the two inbred lines' coleoptile were expressed. The results implied the response of between deep-seeding tolerant and intolerant maize coleoptile to deep-seeding stress and exogenous EBR stimulation may vary at the transcriptional level, and differential expression level of these DEGs leads to the significant differences of coleoptile.

Figure 2.

Differentially expressed genes (DEGs) identified in both maize inbred lines (W64A and K12) coleoptile under under three treatments [3 cm seeding depth (3 cm), 20 cm seeding depth (20 cm), and 4.16 × 10⁻³ M EBR stimulation under 20 cm stress (20 cm+EBR)]. All DEGs of coleoptile between W64A and K12 under three treatments (a). The bar charts indicated DEGs under multiple or single treatments (in the x-axis, black dots represented a single treatment, black lines connected by dots represented multiple treatments, and the y-axis represented the number of genes corresponding to them). Veen of up-regulated DEGs identified between W64A and K12 under 20 cm and 20 cm+EBR treatments (b). Veen of downregulated DEGs identified between W64A and K12 under 20 cm and 20 cm+EBR treatments (c)

Because the intolerant line K12 had a longer coleoptile under 20 cm deep-seeding stress and 4.16 × 10⁻³ M exogenous EBR induction, so the GO annotation and KEGG pathway analyses were analyzed with K12 in all treatments. The GO annotation analysis showed “metabolic process,” “cellular process,” “biological regulation,” and “response to stimulus” were the most significantly enriched GO terms in the biological process category (Figure 3a–b). Additionally, the top 20 pathways identified by KEGG pathway analysis displayed “plant hormone signal transduction (map04075),” “starch and sucrose metabolism (map00500),” and “valine, leucine and isoleucine degradation (map00280)” were the most significantly enriched KEGG pathway (Figure 3c–d). Therefore, the findings implied multiple phytohormones signaling transduction pathways may play critical roles in maize coleoptile elongation under deep-seeding stress and exogenous EBR supplement.

Figure 3.

Visualization annotation of DEGs in K12 coleoptile under three treatments [3 cm seeding depth (3 cm), 20 cm seeding depth (20 cm), and 4.16 × 10⁻³ M EBR stimulation under 20 cm stress (20 cm+EBR)]. GO enrichment analysis of DEGs in K12 coleoptile under 20 cm (a) and 20 cm+EBR treatments (b). KEGG pathways enriched DEGs in K12 coleoptile under 20 cm (c) and 20 cm+EBR (d) treatments

Multiple phytohormones signaling transduction pathways involving in maize coleoptile development

The phytohormones, including AUX, BR, cytokinin (CTK), ethylene (ET), ZT, GA, ABA, JA, and SA, can regulate various processes of plant growth, development, and environmental adaptation independently and synergistically. As far as AUX signaling transduction (Figure 4a), two differentially expressed auxin influx carrier (AUX1) family DEGs (2.18 ~ 4.02-fold) were up-regulated in between deep-seeding tolerant (W64A) and intolerant (K12) inbred lines under 20 cm deep-seeding stress, and only one AUX1 DEG (−1.19-fold) were down-regulated in K12 with the 4.16 × 10⁻³ M EBR application under 20 cm pressure, that led to 27 DEGs of auxin-responsive protein IAA (AUX/IAA; −5.55 ~ 5.29-fold) were up/downregulated in both inbred lines under 20 cm stress or 4.16 × 10⁻³ M EBR mediation, along with five/four up/downregulated auxin response factors (ARFs; 1.41 ~ 9.99/-3.11~-1.35-fold) were also altered in K12 under 20 sowing depth or 4.16 × 10⁻³ M EBR stimulation, which directly activate or repress the AUX response DEGs, i.e. 27 DEGs responsible for AUX/IAA (−5.55 ~ 5.29-fold), six DEGs of auxin responsive promoter (GH3; −2.85 ~ 5.11-fold), and 23 DEGs controlling SAUR family protein (SAUR; −6.74 ~ 5.07-fold) were observed in both inbred lines under 20 cm planting condition or 4.16 × 10⁻³ M EBR treatment. These DEGs involved in AUX transport, and signaling pathway modified IAA metabolism may provide a critical evidence of cell enlargement in maize coleoptile subjected to 20 cm sowing depth or exogenous EBR feeding.

Figure 4.

DEGs involved in multiple phytohormones signaling transduction in both maize inbred lines (W64A and K12) coleoptile under three treatments [3 cm seeding depth (3 cm), 20 cm seeding depth (20 cm), and 4.16 × 10⁻³ M EBR stimulation under 20 cm stress (20 cm+EBR)]. The absolute values of the log2 fold-change [K12 (20 cm+EBR_vs_3 cm) (VS1), K12 (20 cm_vs_3 cm) (VS2), W64A (20cm+EBR_vs_3 cm) (VS3), and K12 (20cm_vs_3 cm) (VS4)] > 1 and FDR < 0.001 were used as the criteria for DEGs. The red/green box represented up-/down-regulated DEGs in VS1, VS2, VS3, or VS4, respectively, and the value in the box was the log2 fold-change (VS1, VS2, VS3, or VS4) of the DEGs. DEGs controlling the auxin (AUX; a), cytokinin (CTK; b), zeatin (ZT; c), gibberellin (GA; d), abscisic acid (AB; e), jasmonic acid (JA; f), brassinosteroid (BR; g), ethylene (ET; h), and salicylic acid (SA; i) signaling transduction

As shown in Figure 4b, complex CTK response metabolism existed the two contrasting inbred lines that upon to 20 cm sowing depth or 4.16 × 10⁻³ M EBR treatment. Three cytokinin receptor DEGs of Arabidopsis histidine kinase 2/3/4 (CRE1) were differentially expressed between K12, that were exposed to 20 cm deep-seeding depth and 4.16 × 10⁻³ M EBR supplement, and W64A that was cultured under 20 cm growing depth, respectively. Then the CTK signaling was further transmitted based on five histidine-containing phosphotransfer protein (AHP; −10.40 ~ 3.60-fold), five two-component response regulator ARR-B family (B-ARR; −1.58 ~ 5.59-fold), and four two-component response regulator ARR-A family (A-ARR; −2.19 ~ 5.00-fold) in both inbred lines under planting pressure and exogenous EBR application of 4.16 × 10⁻³ M concentration. In addition, Cis-/Trans-ZT of the CTK are the main driving forces in plants. So one DEG encoding cytokinin trans-hydroxylase (CYP735A; 3.99 ~ 5.90-fold), two DEGs encoding dimethylallyltransferase (IPT; 4.78 ~ 6.48-fold), three DEGs encoding O-glucosyltransferase (CISZOG; −3.06 ~ 4.68-fold), eight DEGs encoding cytokinin dehydrogenase (CKX; −3.01 ~ 6.31-fold), and five DEGs encoding UDP-glucosyltransferase 73 C (UGT73C; 1.03 ~ 3.97-fold) were detected in both inbred lines under multiple treatments (Figure 4c), that were further involved in the Cis-/Trans-ZT biosynthesis, and resulting in regulating cell division of maize coleoptile under deep-seeding stress or exogenous EBR stimulation.

In GA signaling (Figure 4d), one DEG of gibberellin receptor GID 1 (GID1; 1.82-fold) was up-regulated in K12 with 4.16 × 10⁻³ M EBR application under 20 cm stress, one DEG of transcriptional regulator DELLA protein (DELLA; 1.29 ~ 3.58 fold) was positively regulated, as well as two DEGs of phytochrome-interacting factor 3 (PIF; −1.49~-1.87-fold) were down-regulated between K12 and W64A under under 20 cm depth or 4.16 × 10⁻³ M EBR induction, then the differential expression level of these DEGs may induce seed germination and coleoptile growth in maize under 20 cm stress and exogenous EBR treatment.

Interesting, for ABA signaling response (Figure 4e), six DEGs controlling ABA receptor PYR/PYL family (PYR/PYL; −2.17 ~ 1.77-fold), six DEGs controlling protein phosphatase 2 C (PP2C; −3.44 ~ 3.09-fold), and eight DEGs of ABA response element binding factor (ABF; −7.90 ~ 5.51-fold) were expressed at different levels in both inbred lines under multiple treatments.

Intriguingly, in JA signaling transduction (Figure 4f), the two maize inbred lines under different treatments, JA in coleoptiles was transformed into active jasmonoyl-isoleucine (JA-Ile) under one DEG of jasmonic acid-amino synthetase (JAR1), one down-regulated DEG of coronatine-insensitive protein 1 (COI1; −1.63 fold) sensed JA-Ile and then triggered 20 DEGs of jasmonate ZIM domain-containing protein (JAZ; 1.04 ~ 10.99-fold) positive expression to active five DEGs of transcription factor MYC2 (1.12 ~ 4.54-fold) regulated JA response genes.

BR signaling for coleoptile between W64A and K12 under different conditions, further detecting one DEG responsible for BRI1 kinase inhibitor 1 (BKI1; 4.48 fold) was upregulated, which then interacted with brassinosteroid insensitive 1 (BRI1) associated kinase receptor 1 (BAK1) to negatively regulate BR signaling transmission, and BR-signaling kinase (BSK; −1.89 ~ 13.19-fold). Except for BAK1, four DEGS responsible for BR-signaling kinase (BSK; −1.89 ~ 13.19-fold) also interacted with BRI1 to positively regulate one DEG of protein brassinosteroid insensitive 2 (BIN2; 10.81-fold), and two upregulated DEGs of brassinosteroid resistant 1/2 (BZR1/2) further bound to above BIN2 and directly regulated the downstream DEGs upregulated expression, i.e. three DEGs of xyloglucosyl transferase TCH4 (TCH4; 1.15 ~ 4.54-fold) and three DEGs of cyclin D3, plant (CYCD3; 2.57 ~ 5.20-fold) involved in cell elongation and division of maize coleoptile under deep-seeding stress and exogenous EBR stimulation, respectively (Figure 4g).

ET also is an important phytohormone to cope with multiple stress pressure and adjust tissue growth. When the seeds of two inbred lines were incubated under different sowing depths and exogenous EBR treatments, two downregulated DEGs of ethylene receptor (ETR; −4.69~-1.30-fold) and one DEG of serine/threonine-protein kinase (CTR1; −1.95-fold) were observed, then further activating two DEGs of mitogen-activated protein kinase kinase 4/5 (MKK4_5; 1.07 ~ 2.04-fold) were up-regulated. Consequently, the two down-regulated DEGs of ethylene-insensitive protein 2 (EIN2; −2.57~-1.02-fold) were positively regulated ET reaction. Moreover, four upregulated DEGs of ethylene-insensitive protein 3 (EIN3; 1.39 ~ 7.14-fold) located in the downstream of EIN2, then even bound to two DEGs of EIN3-binding F-box protein (EBF1/2) and further induced EBF1/2 (1.17 ~ 3.33-fold) expression, as well as the excess EIN3 were degraded to keep it dynamic equilibrium (Figure 4h).

For SA signaling transduction (Figure 4i), 12 DEGs of regulatory protein NPR1 (−6.24 ~ 3.12-fold) displayed differential expression level in both inbred lines under 20 cm stress and 4.16 × 10⁻³ M exogenous EBR feeding, which then was able to translocate to the nucleus and regulate 12 downstream transcription factor TGA (−3.70 ~ 4.22-fold) positively expression, further these DEGs of TGA were able to modulate five DEGs of pathogenesis-related protein 1 (PR-1; −8.06 ~ 9.41-fold) high expression levels.

Discussion

Clear divergence exist between inbred lines W64A and K12 in their deep-seeding stress response

The deep-seeding tolerant maize varieties, with the rapid germination of seeds and robust emergence of seedlings at deep-seeding depth, were planted deep soil layer, their seeds could take advantage of soil moistened by the underground water, to promote seedling emergence normally, that as an advantage strategy has been widely applied to maize production in (semi-) arid regions.^2,3,8 As early as the 1990s, in the arid southwestern region of the United States and parts of western Mexico, native Americans had planted their local maize varieties “P1213733 (Komona)” at 30 cm seeding depth or more in an attempt to reach soil moisture.¹⁹ In fact, to emergence under drought stress with moisture available in the deeper soil layer, maize seed firstly had to elongate its mesocotyl, further growth of the seedling then depend on the elongation of plumule enveloped by coleoptile.⁴ Liu et al.⁴ reported the contribution to deep-seeding germination ability in maize at 12.5 cm seeding depth was ranked as the following order: mesocotyl > plumule > coleoptile via the correlation coefficients, which was consistent with our study, i.e. the seeds of both deep-seeding tolerant W64A and intolerant K12 were sowed the 20 cm deep-seeding depth, their coleoptile length and mesocotyl length elongated significantly compared to 3 cm condition (Figure 1a–d). Pointing elongation of coleoptile and mesocotyl act as the crucial response to make germplasm tolerant to deep-seeding in maize. And what’s more interesting about elongation characteristic of both coleoptile and mesocotyl in both contrasting inbred lines, we found coleoptile played a more important role in intolerant K12, and which were increased by 1.48 and 2.19 times in W64A and K12, respectively. Instead, mesocotyl displayed a more positive role in deep-seeding tolerant W64A, and which showed 107.58 and 62.27% increase in the two inbred lines, respectively, and further due to significant difference on length ratio of coleoptile to mesocotyl in W64A (0.27) and K12 (1.69) (Figure 1a–d). Suggesting like Triticum aestivum L. and Oryza sativa L.,^10,20 the contrasting maize with tolerance against deep-seeding stress may evolve variant organs (coleoptile and mesocotyl) to respond more positively to deep-seeding tolerance, because the recent evidences of Triticum aestivum L. and Oryza sativa L. observed that longer coleoptile and mesocotyl was a desirable organ for seedling under deep-seeding stress environments, respectively.^10,20

The phytohormones profiling in the elongating mesocotyl of Oryza sativa L. in response to light exposure over time had revealed that light could inhibit mesocotyl elongation by increasing JA level, and decreasing IAA, GA3, as well as Trans-ZT accumulation.²¹ The elongation of both coleoptile and mesocotyl in maize were essentially their cell elongation and cell division, and the changes of these cells were largely regulated by multiple phytohormones.^1,2,22 We further tried to reveal the broad responses of multiple phytohormones including coleoptile and mesocotyl elongation at the deeper soil of 20 cm in present study, and the results showed eight phytohormones contents of IAA, GA₃, ABA, EBR, Cis-ZT, Trans-ZT, JA, and SA of coleoptile and mesocotyl between W64A and K12 had remarkably changes (P < .05), and were always in a subtly dynamic balance during the maize coleoptile and mesocotyl elongation under 20 cm deep-seeding stress, and even existed polar transport or interactions of phytohormones between coleoptile and mesocotyl (Figure 1f; Supplementary Figure 1a–h). Moreover, the Pearson correlation analysis also showed there were close correlation among coleoptile elongation, multiple phytohormones level, length ratio of coleoptile to mesocotyl, phytohormone ratio of coleoptile to mesocotyl (Figure 1g–h). In parallel, Sasse also reported the ET generator ethephon overcame BR-induced elongation in an antagonistic interaction in Pisum sativum L. seedling, and the characteristics of AUX-induced elongation were not displayed in BR-induced growth of the upper stem segment.²³ However, there have been few studies reported in molecular mechanism of coleoptile/mesocotyl elongation under deep-sowing stress, and were utilized to clarify the interaction of phytohormones.

EBR stimulation promote coleoptile and mesocotyl elongation to improve maize deep-seeding tolerance

BR is an essential regulator of many aspects of plant development including seed germination,²⁴ organ elongation,²⁵ and stress response.²⁶ Moreover, increasing evidences indicated that BR and its stereoisomers also involved in elongation of hypocotyl structures in a wide range of plant species, such as the the BR biosynthesis was essential for hypocotyl etiolation in the dark, and application of bioactive BRs could promote Arabidopsis hypocotyl elongation;²⁷ BR caused up to fourfold increase in Glycine max L. epicotyl length, and further enhanced SAUR and AUX/IAA family protein of GH1 transcripts after 18 h;²⁸ BR also stimulated elongation of etiolated Cucurbita maxima Duch. hypocotyl segments with outer tissues removed, as well as that of unpeeled segments, and BR changed the mechanical properties of cell walls of the inner tissue, hence the inner tissue was probably the target tissue in BR induced elongation.²⁹ In these regards, we further explored the effects of 4.16 × 10⁻³ M concentration of applied exogenous EBR on maize coleoptile and mesocotyl elongation under 20 cm deep-seeding stress, and then showing that the average length of coleoptile and mesocotyl in both inbred lines were further elongated 1.09 and 1.31 times, respectively (Figure 1a–d). Intimating exogenous EBR may confer induced cell longitudinal growth to promote coleoptile and mesocotyl elongation of maize under deep soil layer, and the positive effects of EBR on mesocotyl was slightly stronger than that on coleoptile. Further analysis of the study also indicated there had markedly difference among IAA, GA₃, ABA, EBR, Cis-ZT, Trans-ZT, JA, and SA level of coleoptile and mesocotyl between W64A and K12 when they exposed to 4.16 × 10⁻³ M EBR induce under 20 cm stress, and these phytohormones ratio of coleoptile to mesocotyl in W64A were significantly higher than that in K12 (Figure 1e). This observation was consistent with previous study,³⁰ pointing BR treatment promoted both shoot and root growth of Malus hupehensis, increased their AUX level, while declined GA₃ and ABA level, and the expression level of MdUUCCA6, MdYUCCA10, MdPIN1, MddPIN2, and MddPIN3 for AUX synthesis and transport were significantly upregulation in BR-treatment. Therefore, the changes in corresponding gene expression and phytohormones level may due to an elongated coleoptile/mesocotyl in EBR-treated maize at 20 cm soil layer.

Molecular mechanism of maize deep-seeding tolerance by phytohormones signaling transduction

Further, the RNA-Seq analysis of coleoptile between W64A and K12 subjected to three treatments of 3 cm seeding depth, 20 cm deep-seeding depth and 4.16 × 10⁻³ M EBR stimulation under 20 cm stress were compared. Taken collectively, our results have shown that divergent response to deep-seeding stress exist between W64A and K12, and that there was coherence between the physiological characterization and transcriptome profiling data of the two inbred lines, i.e. totally 15,607 and 20,491 DEGs were detected in two maize coleoptiles among all treatments (Figure 2a). Even we verified the reliability of our RNA-Seq data by qRT-PCR analysis of 15 selected DEGs relative expression level (Supplementary Figure 3a–o). Moreover, the DEGs in K12 under different treatments displayed “plant hormone signal transduction (map04075),” “starch and sucrose metabolism (map00500),” and “valine, leucine and isoleucine degradation (map00280)” were the most significantly enriched KEGG pathway (Figure 3c–d). Therefore, we further deeply dissected the molecular mechanism of multiple phytohormones signaling transduction in contrasting maize coleoptile elongation under deep-seeding stress and applied exogenous EBR treatments.

AUX regulates different plant growth and developmental processes. The three AUX-responsive SAUR genes and two GH3 genes were downregulated of Oryza sativa L. mesocotyl in response to light at three time points by transcriptome,²¹ whose homolog gene AtSAUR24 function in promoting cell expansion and hypocotyl growth in Arabidopsis,³¹ and the OsGH3.1 of Oryza sativa L. encoding IAA-amido synthetases, its mutant had low level of free IAA and was insensitive to IAA.³² IAA33 (AUX/IAA protein) maintained root distal stem cell identity and negatively regulated AUX signaling by interacting with ARF10 and ARF16.³³ We also identified seven AUX/IAA, two GH3, and two SAUR DEGs were significantly down-regulated in deep-seeding signaling, as well as ten AUX/IAA, four ARF, one GH3, and eight SAUR DEGs were significantly down-regulated with the EBR stimulation under deep-planting stress (Figure 4a). So these findings suggested the decrease of IAA level of coleoptile by deep-sowing stress and EBR induction may be attributed to the decreased expression of AUX/IAA, ARF, GH3, and SAUR genes.

Almost all organisms make CTK, in plants have cytokinin been definitively shown to act in cellular signaling, and the most prevalent CTK is ZT included trans and cis configurations. The CTK biosynthesis was initiated by catalyzing with IPT.³⁴ Trans-ZT was produced via hydroxylation of the isoprenoid side chain in a process that is carried out by CYP735A1 and CYP735A2 of CTK trans-hydroxylase, and produced Trans-ZT played an important role in promoting shoot growth in Arabidopsis.³⁵ Moreover, the Trans-ZT could be cleaved by CKX, and overexpression of CKX led to reduce active CTK level and result in developmental defects.³⁶ Similarly, we identified two IPT, one CYP735A, six CKX, four UGT73C, and three CISZOG EDGs of two maize coleoptiles were significantly up-regulated in EBR treatment, in contrast, only three CKX, two UGT73C DEGs had significantly positive expression (Figure 4c), which were also consistent with the levels of Cis-/Trans-ZT under corresponding conditions (Supplementary Figure 2e–f), therefore, the findings may explain why deep-seeding signaling and EBR stimulation could lead to increase of Cis-/Trans-ZT content. In addition, for CTK signaling transduction, AHP2, AHP3, AHP5 of AHP in Arabidopsis were partially redundant positive regulators of CTK signaling, but they also act downstream of other plant histidine kinase (HK), e.g. CKI1.³⁷ There also had two types ARRs including B-ARR and A-ARR controlled CTK signaling, that were activated by phosphorylation of the Asp residue in their receiver domain by AHP, and were essential for the initial transcriptional response to CTK in Arabidopsis.³⁸ Interesting, multiple DEGs of CRE1, AHP, A/B-ARR that exhibit rapid expression changes were observed in contrasting maize coleoptile among three treatments in our study (Figure 4b).

GA regulates growth and development throughout the plant life cycle, and multiple endogenous and environmental signaling integrated to influence GA accumulation and transport. In the presence of GA in the cytoplasm and nucleus, GID1 bound GA and this complex promoted the degradation of DELLA proteins that inhibited the expression of GA mediated genes.³⁹ Different from the previous studies, the seeds of intolerant K12 were exposed to EBR application under deep-seeding stress, one up-regulated DEG of GID1 interacted with one upregulated DEG of DELLA to induce one DEG of PIF was negative expressed (Figure 4d), which may provide a new evidence for better interpretation of GA signaling in the future.

ABA perception occurred when it bound to the PYR/PYL/RCAR (regulatory component of ABA receptors.⁴⁰ Six DEGs of PYR/PYL were differentially expressed in coleoptile of both W64A and K12 in response to deep-seeding stress and EBR treatment, which were consistent with the expression analysis of members of the Oryza sativa L. PYR/PYL/RCAR family in distinct tissues in ABA stimulation,⁴¹ suggesting there were the specificity in ABA signaling transduction and diverse biological function. Then binding of ABA to PYR/PYL/RCAR receptors led to inhibition of PP2C activity, which in turn activated SnRK2, and SnRK2 further phosphorylated relevant substrates such as ABF of the ABA response element binding factor, ultimately resulting in ABA-related physiological responses under environment stress.⁴² By contrast, we found three downregulated DEGs of PYR/PYL induced four DEGs of PP2C positively expression, following four DEGs of SnRK2 negatively expression to generate three DEGs of ABF downregulation in coleoptile of deep-seeding tolerant W64A with EBR treatment and deep-seeding stress (Figure 4e).

JA inhibited coleoptile and mesocotyl elongation in etiolated Oryza sativa L. seedling,⁴³ and repressed hypocotyls elongation and stimulated cotyledon unfolding in etiolated Arabidopsis seedling.⁴⁴ And that six upregulated JAZ genes were found that involved in light-dependent regulation of Oryza sativa L. mesocotyl.²¹ The substantial differences were also evident in this study. We observed the exogenous EBR-activation of multiple DEGs involved in the JA biosynthesis and signaling transduction in intolerant K12 under deep-seeding stress, such as two JAR1, 15 JAZ, and three MYC2 were differentially positive expression level (Figure 4f), and these changes were also consistent across correlational observational of GA₃ accumulation (Supplementary Figure 2g). All these results suggested unlike light stimulation, JA may positively regulate coleoptile elongation at deep-seeding pressure and exogenous EBR treatment.

BR, relied on autologous signaling transduction, involved in multiple aspects of physiological responses essential to growth and development in whole life course. In the absence of BR, BKI1 as a negative regulator of BR signaling, which could bind to BRI1, and presumably inactivating its function by preventing association with BRI1-associated receptor kinase 1 (BAK1).⁴⁵ Interesting, our result supported the above study of Clouse again,⁴⁵ pointing only one DEG of BKI1 was up-regulated when W64A seeds was no EBR treatment under deep-seeding stress (Figure 4g). In contrast, BSK was direct substrate of BRI1 and positive regulator of BR signaling, and Tang et al.⁴⁶ had authenticated BSK1 and BSK3 were show to interact directly with BRI1 in vivo in the absence of ligand, the BSK1 then was phosphorylated by BRI1, causing its activation and release from the receptor complex, and promoting its interaction with the negative regulator BIN2.⁴⁷ Further BIN2 also phosphorylated its substrate BZR1, i.e. a 12–amino acid BIN2 docking motif adjacent to the C terminus of BZR1 allowed interaction with BIN2 and subsequent phosphorylation on specific BZR1 residues. Deletion of the docking motif prevents BIN2-BZR1 interaction and in vivo phosphorylation of BZR1 and leads to the nuclear accumulation of BZR1-GFP in dark-grown hypocotyls.⁴⁵ Similarly, the transcriptome also identified three DEGs of BSK, one DEG of BIN2, two DEGs of BZR1/2, three DEGs of TCH4, and two DEGs of CYCD3 were significantly upregulated in two inbred lines’ coleoptile under different treatment (Figure 4g).

One of the earliest reported response to ET was the triple response, that was common in eudicot seedling grown in the dark and was characterized by reduced growth of the hypocotyl and root.⁴⁸ To response ET signaling, two DEGs of ETR, one DEG of CTR1, and two DEGs of EIN2 were down-regulated, but two DEGs of MKK4_5, and two DEGs of EBF1/2 were upregulated in both inbred lines under different treatment in this study (Figure 4h), then to regulate development of these maize coleoptile.

In addition, we also found a pervasive and unexpectedly strong connection between EBR-treated/deep-seeding signaling and genes to SA, the differential expression level of multiple NRP1, TGA, and PR-1 DEGs may medicate the dynamic change of SA level in both maize materials under deep-seeding stress and EBR treatment (Figure 4i; Supplementary Figure 2h). Suggesting SA signaling to be also an important coleoptile growth regulator in maize.

Interaction network and regulation mechanism of DEGs in phytohormones signaling transductions

It appears that corresponding DEGs interactions have an important role in phytohormones signaling integration. Here, we experimentally generated two systems-level maps of phytohormones signaling network in maize coleoptile elongation under deep-seeding stress and EBR-mediated treatment, respectively, i.e. totally 79 DEG-DEG interactions in deep-seeding signaling, which involved in signaling transduction of AUX (two AUX1, two SAUR, ten AUX/IAA, and six GH3), CTK (one AHP, three A-ARR, and four B-ARR), ZT (two CKX), GA (one DELLA and one PIF), ABA (three PYR/PYL, four PP2C, seven SnRK2, and one ABF), BR (one BSK and two CYCD3), ET (one ETR, one EIN2, one EIN3, and one EBF1/2), JA (one COI1, nine JAZ, and one MYC2), and SA (three NPR1, six TGA, and five PR-1). Instead, totally 142 DEG-DEG interactions in EBR-mediated signaling, which regulated AUX (three AUX1, 12 SAUR, 14 AUX/IAA, five GH3, and five ARF), CTK (one CRE1, two AHP, three A-ARR, and two B-ARR), ZT (two CISZOG, five CKX, three IPT, and one CYP735A), GA (one GID1, one DELLA, and two PIF), ABA (six PYR/PYL, six PP2C, six SnRK2, and three ABF), BR (three BSK, one BIN2, one BZR1/2, and two CYCD3), ET (one ETR, one CTR1, one MKK4_5, two EIN2, four EIN3, and one EBF1/2), JA (two JAR1, 17 JAZ, and four MYC2), and SA (five NPR1, seven TGA, and four PR-1), etc. signaling transduction (Figure 5a–b). Suggesting there were highly interconnected network among phytohormones during maize coleoptile elongation under deep-seeding stress and EBR stimulation, further illustrated pathway communities and multiple previously unknown pathway contacts among phytohormones that represent potential points of crosstalk.

Figure 5.

DEGs of multiple phytohormones interactome network mapping between W64A and K12 under 20 cm deep-seeding depth (a), and with 4.16 × 10⁻³ M EBR stimulation under 20 cm deep-seeding depth (b), as well as the schematic molecular model of maize coleoptile deep-seeding tolerance in deep-seeding signaling and exogenous EBR stimulation (c)

Based on our main findings of the key seed-seeding responsive DEGs and their associated phytohormones signaling transduction pathways, interaction networks, and the relevant published citations contained in present study, we have developed a molecular model for deep-seeding tolerance in maize coleoptile elongation in deep-seeding signaling and exogenous EBR stimulation as shown in Figure 5c, respectively. Overall, the transcriptome data generated here help guide further research to develop novel strategies for promoting deep-seeding tolerance in maize.

Conclusion

Synergistic elongation of coleoptile and mesocotyl in maize under deep-seeding stress, demonstrated in this study, pointing to intolerant K12 and deep-seeding tolerant W64A may evolve coleoptile and mesocotyl to more positively respond to deep-seeding stress, respectively. We also concluded 4.16 × 10⁻³ M EBR treatment could significantly improve maize deep-seeding tolerance. Transcriptome further revealed plant hormone signal transduction (map04075), starch and sucrose metabolism (map00500), and valine, leucine and isoleucine degradation (map00280) were the critical pathway processes to regulate maize coleoptile elongation under 20 cm deep-seeding depth and EBR signaling treatments, multiple DEGs and TFs for phytohormones (AUX, CTK, ZT, GA, JA, ABA, BR, SA, and ET) signaling transduction were identified, and which formed a highly interconnected network to involve in multiple phytohormones metabolism, and resulting in the level of IAA, GA₃, ABA, and Cis-ZT were significantly reduced, the content of EBR, Trans-ZT, JA, and SA, however, were clearly increased in maize coleoptile under deep-seeding stress and EBR treatment.

Funding Statement

The study was partially supported by the National Natural Science Foundation of China [32060486], the Research Program Sponsored by Gansu Provincial Key Laboratory of Aridland Crop Science, Gansu Agricultural University, China [GSCS-2019-8; GSCS-2020-5], the Scientific Research Start-up Funds for Openly-recruited Doctors, Science and Technology Innovation Funds of Gansu Agricultural University, China [GAU-KYQD-2018-19], the Developmental Funds of Innovation Capacity in Higher Education of Gansu, China [2019A-052].

Disclosure statement

No potential conflicts of interest were disclosed.

Supplementary material

Supplemental data for this article can be accessed on the publisher’s website.