Yusong Lyu, Xingjin Wei, Min Zhong, Shipeng Niu, Shkeel Ahmd,Goneng Sho, Guii Jio, Zhonghu Sheng, Lihong Xie, Shiki Hu, Ywen Wu,Shoqing Tng, Peisong Hu

aState Key Laboratory of Rice Biology,China National Center for Rice Improvement,China National Rice Research Institute,Hangzhou 310006,Zhejiang,China

bNational Key Laboratory of Crop Genetic Improvement,Huazhong Agricultural University,Wuhan 430070,Hubei,China

Keywords:Oryza sativa L.Mesocotyl Transcriptome MicroRNAome Degradome

A B S T R A C T The mesocotyl, a structure located between the basal part of the seminal root and the coleoptile node of seedlings,contributes to pushing the shoot tip through the soil surface,a function that is essential for the uniform emergence of direct-seeded rice.Its elongation is inhibited by light and induced in darkness. This investigation of an indica rice (P25) with vigorous mesocotyl elongation was aimed at identifying the “omics” basis of its lightinduced growth inhibition. A transcriptomic comparison between mesocotyl tissues that had developed in the dark and then been exposed to light identified many differentially expressed genes (DEGs) and differentially abundant microRNAs (miRNAs). Degradome sequencing analysis revealed 27 negative miRNA-target pairs. A co-expression regulatory network was constructed based on the miRNAs,their corresponding targets,and DEGs with a common Gene Ontology term.It suggested that auxin and light,probably antagonistically,affect mesocotyl elongation by regulating polyamine oxidase activity.

1. Introduction

Rice (Oryza sativa L.) crops are increasingly direct-seeded in Asia, a change driven by the challenges of increasing labor and resource costs [1,2]. Direct seeding into soil protects young seedlings from water shortage, which is common in sown crops in north China, and it also reduces the risk of lodging [3,4]. Direct seeding relies on the ability of the germinating seedling to emerge, a trait that is strongly influenced by the capacity of the mesocotyl, the structure separating the cotyledon from the coleoptile,to elongate[5,6].Prolonged selection for productive paddy rice varieties has led, however, to a loss in the rice mesocotyl's ability to elongate. For this reason, modern breeders prefer to select varieties with long mesocotyls[7,8].

As the key environmental factor affecting plant growth and development, light is a signal controlling mesocotyl elongation [9]. Mesocotyl elongation is strongly stimulated in the dark but inhibited in light, suggesting that the mesocotyl of etiolated rice elongates to adapt to the dark soil environment and ceases to grow after crossing the soil surface [10].However, the molecular basis for this switch in growth remains unclear. Studies have suggested that mesocotyl elongation responds to levels of specific phytohormones. In maize(Zea mays L.),the transport of auxin from the coleoptile to the mesocotyl inhibited mesocotyl elongation [11]. The content of the hydrogen peroxide-producing enzyme polyamine oxidase (PAO) was promoted by light and suppressed by auxin in the maize mesocotyl.Hydrogen peroxide(H2O2)is the main substance responsible for wall stiffening, causing cell wall mechanical fortification and resulting in growth inhibition [12,13]. PAOs play a key role in H2O2production in the cell wall [14]. In Arabidopsis thaliana, light inhibits the elongation of the hypocotyl (similar to the rice meoscotyl) by regulating the stability of brassinosteroid(BR)signaling factor BES1/BZR1 [15]. Variants of the rice GSK2 protein, which coordinates strigolactone (SL) and BR signaling, have been associated with variation in the mesocotyl domestication of rice [16]. Rice seedling emergence depends on ethylene inhibition of the production of jasmonic acid (JA) via an OsEIL2-GY1 signaling pathway [17]. Seedling emergence in A.thaliana is regulated mainly by ethylene and a photomorphogenesis factor network [18–20]. A recent transcriptomic analysis of rice mesocotyl [21] also suggested that phytohormone signal transduction is implicated in the inhibition of mesocotyl elongation (ME) imposed by light. Although the effect of light on the elongation of the A. thaliana hypocotyl has been much studied, the regulatory network of rice gene expression during the dark–light transition remains unknown.

miRNA, a major type of endogenous non-coding RNA in plants, functions mainly in post-transcriptional gene silencing via mRNA cleavage or translational repression [22,23]. A growing body of evidence [24,25] suggests that miRNAs exert substantial influence on the growth and development of plants, as well as on their response to stress, by targeting specific genes. As yet, however, their involvement in light inhibition of mesocotyl growth, especially in rice, has not been well investigated.The objective of the present study was to take a genome-wide omics approach to identifying lightresponsive mRNAs and miRNAs in the rice mesocotyl.

2. Materials and methods

2.1. Plant materials, growing conditions, and assessment of phenotype

A panel of rice accessions (Table S1) was screened for variation in mesocotyl growth. The indica cultivar P25 (developed in Cuba) was identified as the accession with most rapid growth and was chosen for the study of its mesocotyl transcriptome, miRNAome, and degradome. The Chinese indica cultivar Peiai 64, which develops an intermediatelength mesocotyl, was used for the quantification of mesocotyl response to N,N′-dimethylthiourea (DMTU), 2-hydroxyethylhydrazine (2-HEH), 1-naphthylene acetic acid(NAA), and N1-naphthylphthalamic acid (NPA) treatments.To measure mesocotyl length,seeds were cultured at 30°C in the dark for five days on Murashige and Skoog medium and then exposed to light for 30 min,1 h,or 24 h.The length of the mesocotyl was measured with a ruler. At least 10 seedlings per treatment were measured.

2.2. Transcriptome library construction, sequencing, and analysis

Total RNA was extracted from mesocotyl tissues using TRIzol reagent (Invitrogen, Waltham, MA, USA) and RNase-free DNase treatment (TaKaRa, Otsu, Japan), following the supplier's protocol. Approximately 10 μg of total RNA from three biological replicates were combined into one library for sequencing. Paired-end sequencing was performed on an Illumina HiSeq 2500 instrument(LC-BIO,Hangzhou,Zhejiang,China), following a standard protocol. Prior to the assembly step, low-quality reads (adaptor or primer sequences and reads with a q score < 20) were removed, and the remaining sequences were mapped onto the rice indica reference genome(http://rice.hzau.edu.cn/rice/) [26,27]. To estimate relative transcript abundances (based on the fragments per kilobase million parameter, FPKM), the aligned read files were processed using Cufflinks software(www.cbcb.umd.edu/software/cufflinks). The criterion applied for calling differential transcription was a q-value<0.05.

2.3.Small RNA library construction,sequencing,and analysis

RNA samples from mesocotyl tissues at each point from darkness to light exposure were gathered and used to generate two sRNA libraries. Filtered reads 18–25 nt in length were purified after separation by 15% polyacrylamide gel electrophoresis and ligated with 3′ and 5′ adaptors. The remaining clean and unique sequences were aligned against miRBase 21 (http://www.mirbase.org/) by BLAST search to identify known and novel miRNAs. Four groups of miRNAs were revealed by bioinformatics analysis of sRNA sequencing based on the classification method. The differential abundances of miRNAs were assessed by Student's t-test(P<0.05).All the sequencing experiments (transcriptome, small RNA,and degradome sequencing) are available in the Sequence Read Archive of National Center for Biotechnology Information (NCBI, https://www.ncbi.nlm.nih.gov/) under accession GSE136320.

2.4.Degradome sequencing,target identification and analysis

Total RNA was used to prepare a degradome library. First,poly(A)+RNA was used as input RNA and annealed with random primers. Then, RNA fragments were captured on streptavidin-coated beads using biotinylated random primers. A 5′ adaptor was ligated to only those RNAs containing 5′ monophosphates, and reverse transcription and PCR were performed.Libraries were sequenced using the 5′ adapter, leading to the acquisition of the first 36 nucleotides of the inserts, representing the 5′ ends of the original RNAs. Finally, the set of putative target genes was subjected to a BlastX search and their functional categorization was based on GO (Gene Ontology, www.geneontology.org/) and KEGG (Kyoto Encyclopedia of Genes and Genomes, www.kegg.jp/). Networks involving miRNAs and their targets were visualized using Cytoscape 3.0 software (www.cytoscape.org). t-plot figures of degradome reads were generated with CleaveL 3.0 (www.CleaveL.org).

2.5. Validation of differential abundance by quantitative realtime PCR

To validate estimates of transcript abundance, a selection of differentially expressed genes (DEGs), negatively regulated targets, and miRNAs were subjected to quantitative realtime PCR (qRT-PCR) on a Light Cycler 480 II Real-Time PCR System device (Roche, Basel, Switzerland). The RNA was reverse-transcribed to obtain single-stranded cDNA (ss cDNA) using either a First-Strand cDNA Synthesis Kit(Toyobo, Tokyo, Japan) or a miRcut Plus miRNA First-Strand cDNA Synthesis Kit (Tiangen, Beijing, China). The reverse sequences of the miRNAs were used as forward primers for the respective miRNAs. The rice Ubiquitin (AF184280) and 5S rRNA genes were chosen as reference sequences for measuring the abundance of, respectively DEGs, and miRNAs[28]. All primer sequences are given in Table S2. The qRTPCRs were based on either a SYBR qPCR mix (Toyobo, Osaka,Japan) or a miRcute Plus miRNA qPCR Detection Kit(Tiangen). Each reaction was represented by three biological replicates.

2.6. Validation of miRNA/target interactions

Transient co-expression experiments were performed in rice protoplasts following Zhang et al. [29]. The flanking sequence of each miRNA (including its fold-back structure)was amplified from a template of genomic DNA extracted from mesocotyl tissue, and the target gene sequences were amplified from a preparation of mesocotyl cDNA (primer sequences in Table S2). The amplified fragments were inserted into pCAMBIA1305 (Addgene, https://www.addgene.org), producing a transgene driven by the CaMV 35S promoter,which was then transformed into protoplasts. The transfected protoplasts were held at 28 °C for 24 h, after which total RNA was isolated using the TRIzol reagent (Invitrogen, Carlsbad,CA, USA).

2.7. Measurement of tissue indole-3-acetic acid (IAA) content and PAO activity

The IAA content of mesocotyl tissues was assessed in 200-mg samples following Fu et al. [30]. Briefly, PAO activity was measured spectrophotometrically following the oxidation/condensation of 4-aminoantipyrine and 3,5-dichloro-2-hydroxybenzene sulfonic acid (purchased from Sigma-Aldrich, St. Louis, MO, USA) as a result of its catalysis by peroxidase, following Cona et al. [31].

3. Results

3.1. Phenotypic variation in mesocotyl length

The mesocotyl length screen of the germplasm entries revealed that the indica variety P25 had the potential to develop the longest mesocotyl and that Nipponbare (NIP)developed one of the shortest (Fig. 1A, B). Seedlings of P25 emerged at a higher rate than did those of NIP (Fig. 1C). P25 seedlings germinated in the dark for five days were divided into two groups, with one group left to grow in the dark (group A) and the other exposed to light (group B). Mesocotyl elongation ceased in the group B seedlings within 1 h of their exposure to light, whereas it continued in the seedlings of group A, such that their mean mesocotyl length after 48 h was 1.6-fold greater than that of the group B seedlings (Fig. 1D,E). Inspection of the cells in the central portion of the mesocotyl showed that cell size was much larger in the group A than in group B seedlings (Fig. 1F, G). Thus, lightinduced inhibition of mesocotyl elongation operated by preventing cell expansion rather than cell division.

3.2. The mesocotyl transcriptome from darkness to light exposure

Six RNA pools were used to characterize the transcriptomic response of the mesocotyl to light. They were sampled from one- to five-day-old seedlings grown in the dark (D_1d, D_3d,and D_5d respectively) and then from five-day-old darkgrown seedlings exposed to light for 0.5, 1, or 24 h (L_0.5 h,L_1h, L_24h; Fig. S1A). The contrast D_1d vs D_3d vs D_5d vs L_0.5 h vs L_1h. L_24h identified 7505, 5391 and 4851 DEGs during the whole process from darkness to light, in the dark(D_1d vs D_3d vs D_5d) and light exposure (L_0.5 h vs L_1h vs L_24h), respectively (Fig. 2A). The DEGs during the dark-tolight transition were classified into a low-to-high or a high-tolow expression pattern, indicating that some key factors controlling ME were included in these DEGs. The GO analysis of these dark–light shared 2703 DEGs indicated enrichment for genes involved in transcription, phytohormone response,protein phosphorylation, photosynthesis, chloroplast development, and cell wall (Fig. 2B). A putative dormancy/auxinassociated family gene (ARP1), a B-type response regulator gene (ORR6), an ethylene signaling key factor gene (OsEIL1)and an auxin response factor gene (ARF22) were all strongly transcribed in the dark, but were suppressed after light exposure. In contrast, a polyamine oxidase gene (PAO5), a gibberellin metabolism gene (GA2ox-10), GY1 (the product of which is involved in jasmonic acid synthesis) and the lightresponsive gene HY5 all behaved in the opposite manner,indicating that they were suppressed in the dark and rapidly activated upon exposure to light (Figs. 2C, S1B).

3.3. The set of miRNAs responding to light exposure

Given the rapidity of both the mesocotyl's transcriptomic and growth response to light exposure (Figs. 1D, 2A), the focus of small RNA sequencing and degradome analysis was directed at the two libraries D_5d and L_1h. The data were based on three biological replicates for each environmental condition (Table S3). The size range of the resulting miRNAs was 18–25 nt,with those of length 24 and 21 nt displaying the highest degree of sequence redundancy (Figs. 3, S2). A total of 967 miRNAs were identified in rice seedling mesocotyls in response to light exposure and were categorized into four groups based on their abundance insequencing reads and the miRNA database (Fig. 3, Table S4). Of these,778 were classified as known(gp1,gp2,and gp3),and the other 189 as novel (gp4); altogether, the miRNAs were grouped into 70 miRNA families (Fig. S3). A total of 83 miRNAs were differentially abundant in the two libraries: 45 were more abundant in L_1h than in D_5d, and vice versa for the other 38(Fig. 4A). Of the 83 miRNAs, 16 were novel (Table S5). The application of qRT-PCR to a selection of the differentially abundant miRNAs validated the small RNA sequencing data for seven of the nine sequences tested(Fig.4B).

Fig.2–The mesocotyl transcriptome as affected by exposure to light. (A)Differentially expressed genes(DEGs)in mesocotyls from darkness to light exposure.Dark,D_1d VS D_3d VS D_5d;Light,L_0.5 h VS L_1h VS L_24h;Dark-light,D_1d VS D_3d VS D_5d VS L_0.5 h VS L_1h VS L_24h.(B)GO enrichment terms for the shared 2703 DEGs.(C)qRT-PCR validation of the differential transcription of a selection of the DEGs identified via RNA sequencing.Details of the genes are given in Table S2.All values are shown as mean±SD(n=3); those marked with a different letter differ at P< 0.05.

3.4. Target prediction of miRNAs derived via degradome sequencing

Based on the signature abundance at each occupied transcript position, the potential cleaved transcripts were classified into five categories (Fig. S4). In all, 823 genes,generating 3262 transcripts, were revealed to be putatively targeted by 305 light-responsive miRNAs (Table S6). Twentyseven negative pairs of miRNA/target genes were identified,including five novel miRNA/target pairs (Fig. 5A, Table S6). A subset of cleavage sites of targets, such as miR160a cleavage sites in ARF10/22 mRNAs, was also detected with high prediction confidence (Fig. S5, Table S6). An analysis based on GO enrichment (Fig. S6A) showed that the most common GO terms were associated with the nucleus (GO:0005634),transcription factor activity (GO:0003700), DNA binding(GO:0003677), and auxin-activated signaling pathway(GO:0009734). The parallel KEGG pathway enrichment analysis(Fig.S6B)suggested that the products of most of the target genes participate in plant phytohormone signal transduction(ko04075), RNA degradation (ko03018), oxidative phosphorylation(ko00190), and photosynthesis (ko00195).

3.5. A co-expression regulatory network based on miRNAs,targets,and DEGs

A combined analysis of miRNAs and DEGs associated with the same GO terms of targets in the auxin and cell wall pathways was conducted. Four miRNAs: miR160, miR396, miR5826, and miR812,were selected on the basis that their targets shared a GO classification associated with one or more DEGs (Fig. 5B).miR396 was associated with four targets and two DEGs associated with auxin, along with seven targets and two DEGs associated with cell wall. miR5826 and miR812 were both more strongly associated with cell wall, while miR160 was more strongly associated with auxin (Fig. 5B). The negative interactions between miR160a and ARF22 and between miR7694 (sharing the same target GO terms of cell wall with osa-miR5826)and PAO8 were confirmed experimentally using qPCR in rice protoplasts and mesocotyl tissues,respectively(Figs.5C,S7).

3.6. Antagonistic regulation of mesocotyl elongation by auxin and light

Fig.3– Size distribution of miRNAs.gp1,miRNA sequences present in miRbase and mapped to a genomic location,plus premiRNAs mapped to a genomic location genome or expressed sequence tag;gp2,miRNA sequences present in miRbase and mapped to a genomic location,but pre-miRNAs not mapped to a genomic location;gp3,miRNA sequences present in miRbase but not mapped to a genomic location,and pre-miRNAs not mapped to a genomic location;gp4,novel miRNAs;Counts MIRb,the counts of miRNAs from miRBase;Hairpin length,pre-miRNA length.Expression level:low indicates<10;middle indicates>10 but less than the mean;high indicates greater than the mean.

Fig.4–Differential abundances of miRNAs during the dark–light transition.(A)Differentially expressed miRNAs in response to light exposure by hierarchical clustering.Red indicates higher and green lower levels of miRNAs.Names of samples are shown at left.The original expression values of the miRNAs were normalized using Z-score normalization.The absolute signal intensity ranged from ?2.04 to+2.04,with corresponding color changes from green to red.(B)qRT-PCR validates expressions of 9 randomly selected miRNAs.The amount of expression was normalized to the level of 5.8s rRNA.The normalized miRNA levels in the control were arbitrarily set to 1.

PAO genes were strongly transcribed in the mesocotyl immediately upon exposure to light (Figs. 2C, 6A), whereas an increase in PAO enzyme activity required a half hour of exposure (Fig. 6B). The IAA content of the mesocotyl was significantly reduced by light exposure (Fig. 6C). After the seedlings were further cultivated in darkness for five days(D-5d), etiolated seedlings were sprayed with exogenous NAA(auxin)and NPA(auxin transport inhibitor),respectively.After 30 min of darkness, PAO activity was measured. The results showed that NAA can rapidly inhibit the PAO activity of mesocotyl tissues of etiolated seedlings. In contrast, NPA increases PAO activity in darkness, a trend that reflects the high IAA contents and low PAO activity of etiolated seedlings in darkness(Fig.6D,E).Thus,NAA and NPA showed different effects of PAO activity in the dark. Furthermore, the PAO activity was highly increased after light exposure under both conditions, possibly because of the decline of IAA contents after light treatment (Fig. 6D, E). PAO catalyzes the oxidative deamination of polyamines(PAs),along with H2O2production in maize [14]. H2O2accumulation in the mesocotyl was high after light exposure and was distributed mainly in the cell wall (Fig. S8A). The scavenger of H2O2DMTU and inhibitor of PAO 2-HEH promoted mesocotyl elongation under darkness(Fig. S8B). NAA and NPA further increased and inhibited mesocotyl elongation, respectively (Fig. S8C). These results suggested that light exposure mainly reduce the auxin content in the rice seedling mesocotyl and increases the PAO activity and H2O2production, ultimately retarding the elongation of the mesocotyl.

4. Discussion

4.1. Dynamic transcriptome profiling sheds light on the complex regulation network of mesocotyl elongation

Fig.5– Network analysis of light-responsive miRNAs and their target genes.(A)miRNAs and targets are marked in red and green,according to whether their abundance was increased (red)or decreased(green)by exposure to light.(B) Combined analysis of miR160,miR396,miR5826,miR812,and DEGs associated with GO designation related to auxin or cell wall pathway.DEGs(targets of miRNAs)were based on three repetitive degradome sequencing data.The miRNAs,their targets,and the associated DEGs are marked in red and green,according to whether their abundance was increased or decreased by exposure to light.(C)Validation of selected miRNA/target interactions using a transient expression system in rice protoplasts.

Fig.6– IAA and light antagonistically control mesocotyl elongation by regulating PAO activity.(A)Transcriptional changes of PAO genes.(B,C)Response of mesocotyl PAO activity and IAA content to light exposure. (D,E)Response of PAO activity to treatment with NAA or NPA.Values are shown in the form mean± SD(n= 3);those marked with different letters differ significantly at P< 0.05 between mesocotyl tissues exposed for various periods of darkness and light.

The set of DEGs responding to light exposure was involved in the regulation of transcription, phytohormone response, and cell wall pathway(Fig.2B).Although many of these DEGs have been previously identified[21],in addition to an investigation of differential transcription in response to light exposure, an attempt was made to identify factors controlling mesocotyl elongation in the dark (Fig. 2C). A key finding was the involvement of the putative auxin/dormancy-associated family protein encoded by ARP1, a homolog of AtDRM1/2[32]. ARP1 was transcribed in mesocotyl tissues deprived of light, and was suppressed when they were exposed to light,indicating a linkage between auxin and light signaling in the regulation of mesocotyl elongation of etiolated rice. During the elongation of maize mesocotyl in the dark, there is an exponential increase in the presence of a growth response factor [33]. In agreement with this report, ORR6, encoding a cytokinin-responsive B-type response regulator, was strongly transcribed in the dark and inhibited by light in our study.Another important gene is likely OsEIL1, a component of ethylene signaling[34].Plants engineered to overexpress EIL1 produced longer mesocotyls than that of the wild type [35].GY1 is another putative inhibitor of mesocotyl elongation,given that dark-grown seedlings of the loss-of-function gy1 mutant produced a longer mesocotyl than the wild type [17].Another gene induced by light was OsHY5, a homolog of AtHY5 that encodes a repressor of skotomorphogenesis.Thus,the process of mesocotyl elongation is controlled by a complex transcriptional regulation network involving both phytohormones and light sensing.

4.2. Post-transcriptional regulation network of miRNAs that responds to light exposure in the rice seedling mesocotyl

Large numbers of miRNAs were identified by small RNA sequencing(Table S3).Many members of the same families of miRNAs showed similar expression patterns.The abundances of four members of the miR395 family and three of the miR399 family were reduced by light exposure (Fig. 4). Several of the miRNAs are known to be critical for the development of the rice plant. Examples are miR156 and miR172, which coordinate vegetative and reproductive branching [36]. miR156 increased in abundance in response to light exposure,whereas miR172 was reduced (Fig. 4). Many of the differentially abundant miRNAs identified have not been reported to date in rice,providing possible new leads towards identifying the genes involved in the light response of mesocotyl growth.

The degradome analysis identified numerous target transcripts, involving both known and novel miRNAs (Fig. 5A,Table S4). Most miRNA targets, including genes encoding transcription factors, signal transduction proteins and the light response, for example, miR164/NAC, miR396/GRF and miR156/SPL, were conserved. NAC29/31 interacts with genes encoding cellulose synthase, supporting the notion that NAC proteins participate in the regulation of cell wall pathway in rice [37]. In maize, the increased abundance of miR396 induced by UV-B irradiation inhibited the transcription of GRFs, suggesting that it acts as a repressor of gibberellinmediated pathways[38,39].An equivalent regulatory network involving miR396 and GRFs was revealed in the regulation of rice mesocotyl elongation (Fig. 5A). In A. thaliana, the interaction between miR156 and SPL is a major module of the shade-avoidance syndrome [40]. In our study too, a negative regulatory relationship between miR156 and SPL was established, suggesting that this interaction operates during the dark–light transition.

4.3. PAO activity of mesocotyl is a key indicator during light exposure

The co-expression regulatory network analysis raised the possibility that auxin-mediated pathways and the cell wall act together to regulate mesocotyl elongation during the dark–light transition (Fig. 5B, C). The cell elongation required for mesocotyl elongation depends on cell wall loosening,which is mediated mainly by H2O2[41,42]. In the maize mesocotyl,genes encoding PAO are induced by light and suppressed by auxin [31]. In the present study, PAO activity with H2O2production was rapidly promoted by exposure to light (Figs.6B, S8). Exogenous auxin depressed PAO activity in the mesocotyl and promoted its elongation(Figs.6D,S8),suggesting that auxin promotes mesocotyl elongation in the dark, at least in part by its inhibition of PAO activity. The OsARF22 gene was down-regulated by light, whereas OsPAO8 was induced (Figs. 1C, 6A). Several potential auxin and lightresponsive elements,such as AuxRE and G box,are present in the promoter region of PAO8 (results not shown), suggesting that auxin and light competitively regulate PAO activity via transcriptional regulation by auxin response factors and lightresponsive trans-activators. Whether transcription factors directly or indirectly regulate expression of PAOs awaits further study.

5.Conclusions

In agreement with the conclusion of Feng et al. [21], the present experiments indicated that in the mesocotyl, the synthesis of various phytohormones, specifically auxin,gibberellin, and cytokinin, is promoted in the absence of light, allowing the cells to elongate, while at the same time both PAO activity and H2O2production are suppressed (Fig.S9). When the coleoptile reaches the soil surface, auxin,gibberellin and cytokinin in the mesocotyl are reduced in response to light, while jasmonic acid and strigolactone are increased, resulting in a decline in the expression of genes involved in cell expansion, a cessation in mesocotyl elongation,and a rise in both PAO activity and H2O2content.During this dynamic process, many negative pairs of miRNA–mRNA targets such as miR396-GRFs, miR160-ARFs, and miR7694-PAOs are enriched (Fig. S9).

Supplementary data for this article can be found online at https://doi.org/10.1016/j.cj.2020.05.002.

Declaration of competing interest

The authors declare no conflict of interest.

Acknowledgments

This research was financially supported by the National S&T Major Project of China (2016ZX08001006), the National Key Research and Development Program of China(2016YFD0101801 and 2017YFD0100300) and the Agricultural Science and Technology Innovation Program of CAAS.

Author contributions

Yusong Lyu, Xiangjin Wei and Shipeng Niu performed the experiments. Yusong Lyu, Min Zhong, Gaoneng Shao, Shikai Hu, and Shakeel Ahmad analyzed the data. Peisong Hu and Xiangjin Wei designed the project. Zhonghua Sheng, Guiai Jiao,Lihong Xie,and Yawen Wu helped draft the manuscript.Peisong Hu, Xiangjin Wei, and Shaoqing Tang critically revised the article. All authors have read and approved the final manuscript.