Abstract
Long non-coding RNAs (lncRNAs) are non-protein-coding transcripts longer than 200 nt that are distributed widely in organisms and play many physiological roles. The BoNR8 lncRNA is a 272 nt long transcript yielded by RNA polymerase III in cabbage that was identified as the closest homolog of the AtR8 lncRNA in Arabidopsis. The BoNR8 lncRNA was expressed extensively in the epidermal tissue in the root elongation zone of germinated seeds, and its accumulation was induced by abiotic stresses, auxins and ABA. To investigate the correlation between the BoNR8 lncRNA and germination, BoNR8-overexpressing Arabidopsis plants (BoNR8-AtOX) were prepared. Three independent BoNR8-AtOX lines showed less primary root elongation, incomplete silique development and decreased germination rates. The germination efficiencies were affected strongly by ABA and slightly by salt stress, and ABA-related gene expression was changed in the BoNR8-AtOX lines.
Introduction
Non-coding RNAs (ncRNAs) are RNA species that do not encode proteins and are widely distributed among organisms, accounting for approximately 98% of the mammalian, approximately 71% of the Arabidopsis and approximately 29% of the budding yeast transcriptomes (Hüttenhofer et al. 2002, Wilusz et al. 2009, Zhang et al. 2013). Generally, it is thought that ncRNAs play many important physiological roles (Hüttenhofer et al. 2002, Wilusz et al. 2009, Zhang et al. 2013). According to transcript length, ncRNAs can be divided into two categories (Zhang et al. 2013). Small ncRNAs are shorter than 200 nt, and include microRNAs (miRNAs), small interfering RNAs (siRNAs) and piwi-ineracting RNAs (piRNAs). The primary transcripts (pri-RNAs) of miRNAs are mainly transcribed by RNA polymerase II and cleaved by Dicer to yield miRNAs of approximately 20–30 nt. miRNAs are complementary to their target mRNAs and work as guide regulators in post-transcriptional gene silencing (Costa 2007). ncRNAs over 200 nt are classified as long ncRNAs (lncRNAs). lncRNAs are more abundant than miRNAs in cells, e.g. lncRNAs account for about 80% of the total RNAs (Zhang et al. 2013). Hon et al. (2017) identified 19,175 potentially functional lncRNAs in the human genome using a FANTOM5 cap analysis of gene expression (CAGE) data set, and approximtely 2,000 lncRNAs were associated with diseases. lncRNAs are mainly transcribed from intron regions, intergenic sequences and antisense sequences encoding protein genes, and are 5’ capped, 3’ polyadenylated and expressed in a tissue- and developmental stage-specific manner (Mercer et al. 2009, Nie et al. 2012). Some lncRNAs are known to participate in diverse physiological processes through endogenous antisense mechanisms, epigenetic regulation and activation or repression of specific transcription factor functions, affecting the expression of flanking genes and altering the properties of DNA-binding proteins (Wilusz et al. 2009). The discovery of biological functions for lncRNAs has necessitated an expansion of the legacy genome concept. Thus, the elucidation of lncRNA functions and regulation mechanisms has attracted much interest.
In contrast to animal research, plant ncRNAs have been less studied; in particular, there have been few functional analyses of lncRNAs. Recently, the functions and working mechanisms of some lncRNAs have been studied in detail, including AtIPS1 (target mimicry) (Franco-Zorrilla et al. 2007), COOLAIR (antisense transcription in a condition-/stage-dependent manner) (Swiezewski et al. 2009), COLDAIR (polycomb-dependent model) (Heo and Sung 2011), LDMAR (DNA methylation) (Ding et al. 2012), ASCO-RNA (alternative splicing competitor) (Bardou et al. 2014), HID1 (regulation of PIF3) (Wang et al. 2014), APOLO (regulation of PID by dual transcription of Pol II and V) (Ariel et al. 2014), DRID (modulating expression of genes) (Qin et al. 2017), asHSFB2a (‘Yin–Yang’ regulation) (Wunderlich et al. 2014), FLORE (antiphasic regulatory) (Henriques et al. 2017), Enod40 (RNA–protein interaction) (Campalans et al. 2004) and ELENA1 (transcriptional regulation of plant innate immunity) (Seo et al. 2017).
The mustard family member Brassica oleracea L. is a diploid species with a CC-type genome containing nine chromosomes (2 n = 18) (Lukasik et al. 2013). It includes important vegetable subspecies, such as cauliflower, turnip, cabbage, broccoli, Brussels sprouts, kohlrabi and kale. Cabbage (Brassica oleracea var. capitata) is an agriculturally significant variety of B. oleracea that is rich in vitamin C, dietary fiber, minerals, carotenoids, lupeol and glucosinolates, and has anti-inflammatory and anti-carcinogenic effects. At present, limited numbers of ncRNAs have been identified in B. oleracea.Wang et al. (2012) computationally identified 193 miRNA candidates in broccoli that might regulate lipid synthesis, energy, carbohydrate and nitrogen metabolite pathways. Tian et al. (2014) found 81 broccoli miRNAs responsive to high-salt conditions by high-throughput Solexa sequencing and biological analysis. Thirteen miRNAs from mustard were included in the latest plant non-coding RNA database (PNRD; http://structuralbiology.cau.edu.cn/PNRD) (Yi et al. 2015).(Liu et al. (2014) predicted 3,756 ncRNAs [including miRNA, tRNA, rRNA and small nuclear RNA (snRNA)] in the cabbage genome using next-generation sequencing (NGS), but no lncRNAs were identified. Lukasik et al. (2013) found 261 conserved miRNAs in cabbage leaves using NGS that may be involved in glycolysis, glycerolipid metabolism, flavonoid biosynthesis and oxidative phosphorylation pathways.
Because the seed is a plant-specific reproductive and sink organ, the germination efficiency affects plant proliferation. The physiological processes of seed germination are strongly affected by environmental factors and precisely regulated by phytohormone signaling pathways such as ABA and gibberellic acid (Weitbrecht et al. 2011). At present, the detailed molecular mechanism of seed germination is not clear. Recently, however, several miRNAs involved in Arabidopsis seed germination have been identified. miR160 changes germination ratios through the interaction between ARF10- (AUXIN RESPONSE FACTOR10) dependent auxin signaling and ABA signaling (Liu et al. 2007). miR159 affects seed germination by reducing MYB101 and MYB33 (two positive regulators of ABA responses) transcription levels (Reyes and Chua 2007). miR156 delays germination by down-regulating target SPL (SQUAMOSA PROMOTER-BINDING PROTEIN LIKE) genes (Martin et al. 2010). miR167 also affects germination by regulating ARF expression and auxin signal transduction (Wu et al. 2006). Notably, several seed-related lncRNAs have been found and analyzed in the last 2 years. asDOG1 is a non-coding antisense RNA from DOG1 (Delay of Germination 1) in Arabidopsis, that suppresses DOG1 expression during seed maturation in cis and promotes germination (Fedak et al. 2016). A total of 753 lncRNA candidates were identified in maize seed; Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analyses demonstrated that seven novel lncRNAs can have major impacts on general developmental and metabolic processes of maize seed (Zhu et al. 2017). Multiple lncRNAs linked to DNA methylation were found from early developing castor bean seeds; they play potential roles in regulating the development of castor bean endosperm and embryo (Xu et al. 2018). From soybean seed transcriptomes, 1,575 lncRNAs were identified; some lncRNAs potentially regulate soybean seed development (Yu et al. 2017). A total of 22,430 lncRNAs were identified from developing tree peony seeds, and 39 lncRNAs were predicted to target lipid-related genes, suggesting that these lncRNAs are possibly involved in fatty acid synthesis and lipid metabolism in tree peony seeds (Yin et al. 2018). Kiegle et al. (2018) demonstrated that lncRNAs underwent alternative splicing during the transition from milk seed of rice to mature embryo and endosperm, and exon retention of the lncRNAs was enhanced in embryos.
In a previous work, we predicted a lncRNA designated the AtR8 lncRNA from the Arabidopsis genome by focusing on the conserved gene structure required for transcription by Pol III) The expression of this lncRNA was experimentally confirmed to respond to hypoxic stress (Wu et al. 2012). In parallel, the closest homolog of AtR8 (BoNR8) was found in the cabbage genome which showed 78% nucleotide similarity in the transcribed region (203 of 259 nt were identical) and had a highly conserved USE (upstream sequence element), TATA-like sequence and terminator. These data suggest that the BoNR8 locus of cabbage can be transcribed by Pol III. In this study, we characterized the BoNR8 lncRNA, which was 272 nt long and transcribed by Pol III similarly to Arabidopsis AtR8. The BoNR8 gene is single copy in the cabbage genome, expressed extensively in the epidermis tissue of the root elongation zone in germinating seeds, and responds to abiotic stress conditions. BoNR8 overexpression lines in Arabidopsis showed decreased seed germination rates, and root and silique growth. These phenomena were affected by ABA and salt stress treatments.
Results
Identification of BoNR8 lncRNA in cabbage
To identify the AtR8 lncRNA closest homolog sequence in cabbage, specific primers (primer1 and primer2) were designed from the Brassica draft genome sequence, and a 1,306 bp sequence was amplified from cabbage total DNA. Highly conserved USE, TATA-like and Pol III terminator (a T-stretch >4 bp) sequences were found, indicating that this sequence might have Pol III-dependent transcription features (Supplementary Fig. S1). Moreover, a possible transcribed region starting 25 bp downstream of the TATA motif was highly similar to the AtR8 lncRNA gene, suggesting that this sequence was the AtR8 lncRNA closest homolog in cabbage (Supplementary Fig. S2A).
Next, the 5’ end of the ncRNA was determined by primer extension of an in vitro transcript from a tobacco nuclear extract and from total RNA of cabbage seedlings. The 5’ end was an adenine at 25 bp upstream of the TATA-like sequence (Fig. 1A,C). However, we do not have any information about its cap structure.
Determination of the BoNR8 lncRNA 5’ and 3’ ends. (A) 5′ end mapping of the RNA by primer extension assay. Total RNA from cabbage seedlings (in vivo) and in vitro transcribed BoNR8 lncRNA were subjected to primer extension analysis with the BoNR8-PE primer, and then the extended products and sequencing ladders, prepared using the same primer, were separated by PAGE. An asterisk indicates the 5′ end of the BoNR8 lncRNA. EP, extended products. (B) 3′ end mapping of the RNA by 3′ RACE assay. Cabbage total RNA was polyadenylated by poly(A) polymerase and subjected to reverse transcription using oligo(dT) primer. The obtained cDNAs were amplified by PCR, cloned and sequenced. An arrowhead indicates the 3′ end of the RNA. (C) Nucleotide sequence of the BoNR8 lncRNA gene. USE, TATA-like, possible terminator and ORF sequences are boxed. The conserved salt stress motif located in the BoNR8 lncRNA transcribed region is underlined. The BoNR8 forward and BoNR8 reverse primers are underlined. The putative transcribed region (∼272 nt) is in bold upper case letters. Numbering is with respect to the 5′ and 3′ ends of the BoNR8 lncRNA. (D) The possible secondary structures of the BoNR8 lncRNA and the UCC salt stress motif were predicted by the RNALogo program. Bold letters indicate the conserved salt stress sequences.
The 3’ end was determined by 3’ rapid amplification of cDNA ends (RACE), and a T residue on the Pol III terminator was identified (Fig. 1B,C). Therefore, the length of the AtR8 lncRNA closest homolog in cabbage was 272 nt, which was 12 nt longer than the AtR8 lncRNA (Fig. 1C), and its sequence was 77% identical to that of AtR8 lncRNA (Supplementary Fig. S2A). The RNA was named BoNR8 lncRNA (Brassica oleracealong non-coding RNA). Because the whole genome of B. oleracea has been sequenced, we investigated the position of the BoNR8 lncRNA on the genome at http://plants.ensembl.org/Brassica_oleracea/Info/Index, and found that the BoNR8 lncRNA gene was mapped to chromosome C4: 45,322,023–45,322,291 (+). No longer open reading frames (ORFs) were found in the transcribed region. The longest ORF was seven amino acids in the BoNR8 lncRNA (Fig. 1C), and 15 amino acids in the AtR8 lncRNA (Wu et al. 2012). The RNA secondary structure was predicted by the RNALogo program (rnalogo.mbc.nctu.edu.tw), and the BoNR8 lncRNA formed a very stable stem–loop structure (dG = − 82.71; Fig. 1D). Based on the AtR8 lncRNA and BoNR8 lncRNA alignment, the RNALogo program showed that the conserved nucleotides were mainly positioned in the stem regions of the RNA secondary structure, whereas nucleotides in the loop regions were less conserved (Supplementary Fig. S2B).
We confirmed that transcription of the BoNR8 lncRNA gene was Pol III dependent using an in vitro transcription system prepared from tobacco cultured cells. A low concentration of α-amanitin (0.5 μg ml–1) had no effect on BoNR8 (Fig. 2, lanes 3 and 4) or AtR8 (Fig. 2, lanes 5 and 6) lncRNA transcription, but Pol II-dependent U2 snRNA transcription was completely inhibited (Fig. 2, lanes 1 and 2), revealing that the BoNR8 lncRNA was transcribed by Pol III.
Pol III-dependent transcription of BoNR8 lncRNA. An in vitro transcription assay of the BoNR8 gene from a tobacco nuclear extract was performed with a low concentration of α-amanitin (0.5 μg ml–1). The Arabidopsis U2 snRNA and AtR8 lncRNA were used as controls for Pol II and Pol III activity, respectively. Transcription reactions were performed with (+) or without (−) α-amanitin, and then the de novo BoNR8 lncRNAs (272 nt) were subjected to primer extension reaction, and detected as 101 nt cDNA extended from labeled BoNR8-PE primer (cf. Fig. 1C).
To examine the copy number of the BoNR8 lncRNA gene in the cabbage genome, genomic DNA from cabbage was digested with EcoRI, HindIII and SacII separately (the BoNR8 lncRNA transcriptional region does not contain these cleavage sites), and hybridized with a digoxigenin (DIG)–DNA probe containing the USE, TATA and complete BoNR8 lncRNA transcribed region. Southern blot analysis showed single bands in the digested genomic DNA (Fig. 3. right panel), suggesting that the BoNR8 lncRNA gene had a single locus in the cabbage genome.
Determination of the BoNR8 gene copy number in the cabbage genome. Cabbage and Arabidopsis genomic DNAs were digested with EcoRI, HindIII and SacI separately, and then Southern blot analysis was performed using a DIG-labeled DNA probe for BoNR8 (∼422 bp, from base pair −145 to +277 containing the USE, TATA-like and transcribed regions of the BoNR8 lncRNA). Genomic Southern blotting of Arabidopsis was used as a single-copy gene control (AtR8).
Characterization of the BoNR8 lncRNA during cabbage development
Many reports have shown that ncRNAs play important roles in growth and development processes in flowering plants. The localization of the BoNR8 lncRNA was determined in order to help elucidate its role in the growth and development of cabbage. Low and high molecular weight RNAs were extracted from dry seeds, imbibed seeds, germinating seeds and seedlings [7 days after germination (DAG)] of cabbage as described by Martin et al. (2005). The RNAs were then subjected to Northern blotting using a DIG-labeled AtR8 riboprobe. As shown in Fig. 4A and Supplementary Fig. S3, the BoNR8 lncRNA was highly accumulated in germinating seeds (1–2 DAG), and a trace amount was detected in seedlings. However, no BoNR8 lncRNA signal was detected in dry or imbibed seeds. To confirm further the BoNR8 lncRNA localization, whole-mount in situ hybridization (WISH) using germinating seeds (Fig. 4B) and in situ hybridization (ISH) using cross-sections of the root elongation zone of germinating seeds (Fig. 4C) were performed. We used BM-Purple as an alkaline phosphatase substrate, so that hybridization signals were observed as a purple color. The BoNR8 lncRNA was partially localized in the central cylinder, but was mainly localized in the epidermal tissues of the root elongation zone.
Characterization of the BoNR8 lncRNA during cabbage seed germination. (A) Temporal expression of the BoNR8 lncRNA shown by Northern blot analysis. Low molecular weight RNAs were extracted from dry seed (ds), seeds 24–72 h after imbibition, germinating seeds 24–48 h after germination and seedlings 7 days after germination (DAG). (B) Spatial expression of the BoNR8 lncRNA indicated by whole-mount in situ hybridization (WISH) analysis. Cabbage seedlings (2 DAG) were hybridized with or without AtR8 lncRNA DIG-labeled riboprobes. A red arrowhead indicates BoNR8 lncRNA signals. (C) Localization of the BoNR8 lncRNA shown by in situ hybridization (ISH) of horizontal sections at the root elongation zone. Purple indicates BoNR8 lncRNA signals.
ABA is a key phytohormone directing seed germination, seed maturation and stress responses. Many ABA-responsive genes carry the ABA-responsive element (ABRE) including the ACGT core motif in their promoter regions (Nakabayashi et al. 2005). According to the PLACE (plant cis-acting regulatory DNA elements) database, four ABREs and five W-boxes (TGAC) (Huang et al. 2016) were predicted within the 1,000 bp sequence upstream of the BoNR8 lncRNA transcription start site (Supplementary Fig. S1). The conserved salt stress-responsive structural motif (UCC motif, UCUUCUUCUUUA) (Di et al. 2014), which is usually found in plant mRNAs, was found in the BoNR8 lncRNA sequence with 67% similarity (Fig. 1C). This motif forms a distinctive ‘dumbbell’-like shape in the RNA (Fig. 1D). These data suggested that BoNR8 lncRNA expression might be responsive to ABA and salt stress during cabbage seed germination.
To investigate these possibilities, seedlings at 1 DAG were treated with 150 mM NaCl and 250 mM mannitol for 24 h and Northern blot analyses were carried out (Fig. 5A). The BoNR8 lncRNA was induced slightly by NaCl and substantially by mannitol. Additionally, three types of auxins [10 μM IAA, 10 μM 1-naphthaleneacetic acid (NAA) and 10 μM 2,4-D] and ABA (50 μM) clearly induced BoNR8 lncRNA accumulation after 24 h treatment of seedlings at 1 DAG (Fig. 5B). Furthermore, an early induction peak was observed under 50 μM ABA at 6 h (Fig. 5D). Under the above abiotic stress conditions, the BoNR8 lncRNA was not processed into smaller ncRNAs (Supplementary Fig. S4). In imbibed seeds [2 days after imbibition (DAI)], the BoNR8 lncRNA was weakly induced by 150 mM NaCl and 250 mM mannitol after 24 h of treatment (Supplementary Fig. S5A), but no change was observed under hormone treatments (Supplementary Fig. S5B).
Abiotic stress responses of the BoNR8 lncRNA. Cabbage seedlings (1 DAG) were exposed to 150 mM NaCl, 250 mM mannitol (A), 10 μM IAA, 10 μM NAA, 10 μM 2,4-D and 50 μM ABA (B) for 24 h, and subjected to Northern blot assay. Simultaneously, a time course experiment was performed under 150 mM NaCl (C) and 50 μM ABA (D) for 0, 6, 12, 18, 24 and 30 h, respectively. For each sample, approximately 5 µg of low molecular weight RNAs were hybridized with the AtR8 riboprobe. Signals were detected with a LAS 4000 instrument and quantified with Multi Gauge v3.2 software (Fujifilm). Bar graphs indicate mean values (±SE) with two independent experiments. Ethidium bromide-stained tRNA bands are shown as loading controls.
BoNR8 lncRNA functions during normal development of Arabidopsis
To study further the function of the BoNR8 lncRNA, we tried to overexpress it in cabbage. However, it was technically difficult to make transgenic cabbage plants, so BoNR8 lncRNA overexpression lines were made in Arabidopsis (BoNR8-AtOX) (Fig. 6A; Supplementary Fig. S6). Three independent T3 homozygotic transgenic Arabidopsis plants (BoNR8-AtOX1#, BoNR8-AtOX2# and BoNR8-AtOX3#) were screened out (Fig. 6B). The seed germination rates of the three BoNR8-AtOX lines were lower than that of the wild type (WT; 70–80% in BoNR8-AtOX and 90% in the WT at 7 DAG) (Fig. 6C). Furthermore, less primary root growth in seedlings and incomplete silique development were observed in all three BoNR8-AtOX lines (Fig. 6D,E; Supplementary Fig. S8).
BoNR8 lncRNA functions in Arabidopsis. (A) Construction of the BoNR8 overexpression (BoNR8-AtOX) vector. (B) Relative expression levels of Arabidopsis BoNR8-AtOX lines (1#, 2# and 3#). Total RNA from BoNR8-AtOX seedlings (14 DAG) was subjected to Northern blotting. Ethidium bromide-stained 5.8 S rRNAs are indicated as loading controls. (C) Germination ratios of BoNR8-AtOX seeds. Imbibed seeds kept on 1/2 MS phytoagar (0.8%) at 4°C for 72 h were transferred to 22°C conditions under a 16 h light/8 h dark cycle, and then seed germination rates were counted at the indicated time. Each data point represents the mean ± SE of three replicates each using 30 seeds. (D) Comparison of primary root length. Seedlings at 3 DAG were transferred to 1/2 MS phytoagar (0.8%) and vertically cultured for an additional 7 d, and then the primary root length was measured. Each data point represents the mean ± SE of three replicates each using six seedlings. Asterisks indicate statistically significant differences compared with the wild type: **P < 0.01 as determined by t-test. (E) BoNR8-overexpression in Arabidopsis inhibited silique development. Red arrowheads indicate incomplete siliques.
BoNR8 lncRNA reduced resistance of Arabidopsis to high salt during germination
To evaluate the function of the BoNR8 lncRNA during seed germination in relation to stress responses, the WT and BoNR8-AtOXs were compared. The germination rates on half-strength Murashige and Skoog (1/2 MS) phytoagar (0.8%) containing various concentrations of NaCl (0–200 mM) at 3 DAI were measured (Fig. 7A). The WT and BoNR8-AtOX seed germination rates were similar at 0–50 mM, but at concentrations > 100 mM, the germination rates of the BoNR8-AtOX seeds showed greater decreases than those of the WT. These results indicated that overexpression of the BoNR8 lncRNA in Arabidopsis reduced resistance to high salt concentrations during germination. Generally, salt stress increases the ABA level and suppresses seed germination (Li et al. 2013). Thus, we next tried to assay the effect of sodium tungstate (Na6WO4), which is an endogenous ABA synthesis inhibitor, on germination (Li et al. 2013). As shown in Fig. 7B, the addition of 0.1 mM Na6WO4 did not recover the low germination rates of BoNR8-AtOX seeds under 150 and 200 mM NaCl. Instead, the germination rates of both the WT and BoNR8-AtOX were decreased by Na6WO4 under 200 mM NaCl. These results implied that the increased ABA level induced by high-salt conditions was not the reason for the decrease in BoNR8-AtOX seed germination.
Germination efficiencies of BoNR8-AtOX under high-salt stress. (A) Three BoNR8-AtOX lines and WT seeds were imbibed on 1/2 MS phytoagar (0.8%) containing 0–200 mM NaCl in the dark at 4°C for 72 h, and then transferred to growth conditions with the same NaCl concentration at 22°C under a 16 h light/8 h dark cycle, and germination rates were counted at the indicated times. (B) Sodium tungstate (Na6WO4; an inhibitor of ABA biosynthesis) supplementation did not noticeably rescue the germination of BoNR8-AtOX under high salt. Seeds on 1/2 MS phytoagar (0.8%) were supplemented with NaCl (150 or 200 mM) and Na6WO4 (0.1 mM) at 4°C for 72 h, and then transferred to growth conditions at 22°C with a 16 h light/8 h dark cycle, and seed germination rates were counted at the indicated times. Each data point represents the mean ± SE of three independent experiments using 30 seeds.
BoNR8 lncRNA caused insensitivity of Arabidopsis to high ABA during germination
As mentioned above, BoNR8 lncRNA expression responded to ABA during seed germination in cabbage (Fig. 5B). Therefore, we also investigated BoNR8-AtOX seed germination under exogenous ABA application. Under low concentrations of ABA (0.5–2.0 μM), the germination rates of BoNR8-AtOX seeds were not affected (Supplementary Fig. S7). However, under 15–50 μM ABA, the germination rates were time dependently increased in all plants, and under high concentrations of ABA (75–100 μM), the germination rates of BoNR8-AtOX seeds were higher than those of the WT (Fig. 8). These results indicated that overexpression of the BoNR8 lncRNA caused insensitivity to high concentrations of ABA during germination.
BoNR8-AtOX germination efficiencies were less affected by higher ABA concentrations. Seed dormancy was broken on 1/2 MS phytoagar (0.8%) containing 0–100 μM ABA, and then the seeds were transferred to growth conditions as described in Fig. 7. Germination rates were counted at the indicated times. Data are shown as means ± SE of three independent experiments using 30 seeds.
BoNR8 lncRNA alters the expression of ABA-responsive genes in Arabidopsis
Tremendous research progress has been made on the genes involved in ABA signaling during seed germination and early seedling development (Huang et al. 2016). The BoNR8 lncRNA gene responded to exogenous ABA at the seed germination stage in cabbage (Fig. 5B), and contained four ABREs in its promoter region (Supplementary Fig. S1).
To investigate further the effects of BoNR8 overexpression, nine pivotal genes related to ABA signaling, namely the RD29A (Responsive to Desiccation 29 A, also called COR78), SnRK2.2 and SnRK2.3 (SNF1-related protein kinase 2), WRKY6 (WRKY family transcription factor 6), AtRAV1 (Related to ABI3/VP1, transcription factor), ABI3 (ABA-insensitive 3, B3 transcription factor) and ABI5 (ABA-insensitive 5, bZIP transcription factor), EM1 and EM6 [Late Embryogenesis Abundant (LEA) proteins] genes, were investigated (Huang et al. 2016). The WT and three BoNR8-AtOX lines (1 DAG) were treated with 75 μM ABA for 18 h and then subjected to quantitative real-tiime reverse transcription–PCR (RT–qPCR). SnRK2.3 and AtRAV1 mRNA levels were increased, while EM1 and EM6 mRNA levels were decreased in the BoNR8-AtOXs under normal conditions (Fig. 9C,E,H, I). However, ABA-inducible RD29A (Fig. 9A), WRKY6 (Fig. 9D), ABI3 (Fig. 9F) and ABI5 (Fig. 9G) mRNA levels in BoNR8-AtOX germinating seeds were higher than in the WT, and those of SnRK2.3 (Fig. 9C) were lower than in the WT under ABA-treated conditions. Therefore, we speculated that the BoNR8 lncRNA could up-regulate the ABI3 and ABI5 genes ABA dependently during seed germination of Arabidopsis.
Comparison of ABA-responsive gene expression in BoNR8-AtOX. Seedlings of the wild type and the three BoNR8-AtOX lines at 1 DAG were transferred to 1/2 MS phytoagar (0.8%) with 75 μM ABA or 150 mM NaCl for 18 h, and then total RNAs were extracted. The mRNA levels of ABA-responsive genes, RD29A, SnRK2.2, SnRK2.3, WRKY6, AtRAV1, ABI3, ABI5, EM1 and EM6, were measured by RT–qPCR using Arabidopsis act2 as an internal standard. RNAs from germinated seeds and seedlings at 7 DAG on 1/2 MS phytoagar (0.8%) were used as controls. Data are shown as means ± SE of three independent experiments. Asterisks indicate statistically significant differences compared with the wild type: *P < 0.05 and **P < 0.01 as determined by Student’s t-test.
As mentioned above, salt stress leads to ABA accumulation in the seed (Li et al. 2013). Therefore, the expression levels of the same nine genes were also investigated under 150 mM NaCl for 18 h. The RT–qPCR results showed that RD29A (Fig. 9A) and SnRK2.3 (Fig. 9C) mRNA levels in NaCl-treated BoNR8-AtOX germinating seeds were lower than in the WT. The other genes showed no significant differences (Fig. 9).
Similarly, we also examined the expression of these genes in BoNR8-AtOX and WT Arabidopsis seedlings at 7 DAG under normal conditions (Fig. 9). SnRK2.2 (Fig. 9B) and SnRK2.3 (Fig. 9C) levels were significantly higher than in germinating seeds. Conversely, WRKY6 (Fig. 9D), ABI3 (Fig. 9F), ABI5 (Fig. 9G), EM1 (Fig. 9H) and EM6 (Fig. 9I) levels were lower. The other genes showed no significant changes (Fig. 9).
Discussion
Discovery of the BoNR8 lncRNA
Through various new technologies applied to ncRNA research, a large number of lncRNAs transcribed by Pol II have been found. Some of them have important regulatory roles in biological processes, e.g. the cell cycle, programmed cell death, the establishment of cell identity and environmental stress responses (Wilusz et al. 2009). In contrast, Pol III transcribes many stable housekeeping ncRNAs in organisms, such as tRNA, 5 S rRNA and snRNA, which generally play important roles in RNA processing and translation. Recently, several Pol III-dependent ncRNAs associated with environmental responses and human diseases have been found. Functional ncRNAs such as 21 A (Pagano et al. 2007), 17 A (Massone et al. 2011), NDM29 (Gavazzo et al. 2011), 45 A (Penna et al. 2013) and 51 A (Ciarlo et al. 2013) that are transcribed by Pol III in the human and mouse genomes from intronic and intergenic regions have been reported. We previously computationally predicted 20 ncRNA candidates from the Arabidopsis genome using conserved class III gene promoters as bait sequences (Wu et al. 2012). The AtR8 lncRNA was an abundant RNA and responded to hypoxia stress in seedlings. Here, we found the AtR8 closest homolog (BoNR8 lncRNA) in the cabbage genome, which was confirmed as Pol III dependent using the tobacco in vitro transcription system. The BoNR8 lncRNA was greatly expressed in the epidermal tissue of the root elongation zone during the germination of cabbage seeds. Cauliflower mosaic virus (CaMV) 35 S promoter-driven BoNR8 gene overexpression in Arabidopsis (BoNR8-AtOX) indicated that it acted as a negative regulator of germination and seedling root growth. Recently, we also found that the AtR8 lncRNAs were expressed abundantly during germination (Li et al. 2016). This indicates that PoI III is highly active during germination, and raises the possibility that many other germination-related ncRNAs are transcribed by Pol III. Clearly, germination is a critical stage directly affecting crop yields and quality. This phase is characterized by changes in transcription, translation and phytohormone levels, and follows the resumption of energy metabolism and cellular repair (Weitbrecht et al. 2011). Genes encoding enzymes involved in starch degradation, DNA replication, transcription, translation, recombination, DNA repair, and ribosome biogenesis are activated in germination (Wang et al. 2011). These physiological processes are regulated by the ambient temperature, light conditions and phytohormones (Weitbrecht et al. 2011). To date, several seed-related lncRNAs have been reported, in maize (Zhu et al. 2017), rice (Kiegle et al. 2018), tree peony (Yin et al. 2018), castor bean (Xu et al. 2018), soybean (Yu et al. 2017) and Arabidopsis (Fedak et al. 2016). The AtR8 and BoNR8 lncRNAs can be added to the list of germination-associated lncRNA gene members. The discovery of these lncRNAs provides us with a new avenue to explore a previously unknown aspect of germination physiology.
Characterization of BoNR8 lncRNA structure
The BoNR8 lncRNA is 272 nt long, and its gene is a single locus in the cabbage genome, with five W-boxes and four ABREs in the 5’ upstream promoter region. This lncRNA is expressed extensively in cabbage epidermal tissues in the elongation zone of germinating roots in response to abiotic stresses. Recently, Di et al. (2014) found an evolutionarily conserved RNA structure called the UCC motif among Arabidopsis polyadenylated lncRNAs that was responsive to salt stress. This motif might be recognized by some regulatory factors (e.g. RNA-binding proteins) that can change the RNA secondary structure depending on environmental salt conditions. Additionally, Ding et al. (2014) used the structure-seq method to determine the relationship between RNA secondary structure and stresses including low temperature, heavy metals and salt, and found a distinctive RNA structure (the maximal loop) responsive to abiotic stress. A similar RNA motif was found in the BoNR8 lncRNA gene with 67% nucleotide sequence similarity (Fig. 1C; Supplementary Fig. S2A). A stress-responsive ‘dumbbell’-like dual loop feature was found in the BoNR8 lncRNA as a putative UCC motif (Fig. 1D). These features suggest that some plastic RNAs can change their RNA conformations depending on cellular conditions, and initiate responses against stress. We speculate that the BoNR8 lncRNA secondary structure can also be modified at the UCC-like motif by salt stress in co-operation with unknown regulators. To prove this, further studies are needed to investigate whether conformational changes of the BoNR8 lncRNA can regulate developmental and environmental stress responses. In addition, we need to identify binding factors for the BoNR8 lncRNA. Such studies will lead to better comprehension of the functions of the BoNR8 lncRNA.
BoNR8 lncRNA functions related to AtRAV1 in Arabidopsis
AtRAV1 (Related to ABA-insensitive3/Vivparous1) is a transcription factor in Arabidopsis that contains both plant-specific AP2 and B3 DNA-binding domains (Feng et al. 2014). The major B3 domains are found in ABA-responsive factors, and a large number of AP2-containing proteins have been identified as regulatory factors involved in various developmental processes and stress responses (Hu et al. 2004). The AtRAV1 gene is mainly expressed during seed germination and early seedling development stages (Feng et al. 2014), and is induced by wounding and low temperatures (Fowler et al. 2005, Kagaya and Hattori 2009) but repressed by brassinosteroids (Hu et al. 2004), ABA (Feng et al. 2014), drought and high-salt conditions (Fu et al. 2014). Under high-salt conditions, AtRAV1 overexpression in Arabidopsis decreased seed germination efficiencies, while atrav1 mutants showed increased efficiencies; both observations were ABA independent (Fu et al. 2014). In this study, BoNR8 expression in cabbage was slightly induced by 150 mM NaCl (Fig. 5A). Additionally, the BoNR8-AtOX lines showed high accumulation of the AtRAV1 mRNA in both germinating seeds and seedlings, and an inhibited seed germination phenotype (Figs. 6C, 9E). Under high-salt conditions (100–200 mM NaCl), the germination efficiencies of both the BoNR8-AtOX and WT lines were decreased, but were ABA independent (Fig. 7). Interestingly, the germination rates of the WT were more affected by higher ABA (75–100 µM) than those of BoNR8-AtOX (Fig. 8); thus, the BoNR8 lncRNA might have some antagonistic effect against ABA in germination processes.
At the beginning of this research, we wanted to use Pol III-driven overexpression. However, since we could not find such a technique at that time, we alternatively used the Pol II constitutive promoter (35 S) to overexpress the BoNR8 RNA in Arabidopsis. Indeed, we were worried that the 35 S-dirven Pol II could not produce the proper lncRNA because it lacks precise termination at the Pol III terminator, but Northern blot showed substantial RNA with the same size as expected for BoNR8 RNA. We could not perfectly conclude that Pol II-transcribed BoNR8 RNA was identical to the native BoNR8 RNA. However, we believed that we had many things to do at this time using three independent T3 homozygotic transgenic Arabidopsis. Now we are developing a technique to overexpress using Pol III; this will provide a more accurate answer to this study.
In addition, AtRAV1 overexpression retarded lateral root and rosette leaf growth (Woo et al. 2010), while the atrav1 mutant showed an earlier flowering phenotype, suggesting that AtRAV1 regulates plant development (Hu et al. 2004). In good agreement with these observations, the BoNR8-AtOX lines showed phenotypes of retarded primary root and silique development. Under drought stress, AtRAV1-overexpressing Arabidopsis plants exhibited higher transpirational water loss (Fu et al. 2014). Thus, previous studies suggest that AtRAV1 must be involved in growth and responses to abiotic stresses such as drought and high salinity. Although the contribution of the BoNR8 lncRNA to drought stress responses is not clear at present, it could have certain functions in stress responses and developmental processes related to the AtRAV1 transcription factor (Fig. 10).
Possible regulation model of the BoNR8-dependent seed germination pathway. The BoNR8 lncRNA acts as a negative regulator in AtRAV1-mediated normal development and high-salt stress responses in Arabidopsis. In the presence of ABA, the BoNR8 lncRNA affected the sensitivity of seed germination to ABA by regulating RD29A, SnRK2.3, WRKY6, ABI3, ABI5, EM1 and EM6, which are important ABA-responsive genes in ABA signaling. Solid arrows indicate ABA-dependent steps and dotted arrows indicate ABA-independent steps.
The BoNR8 lncRNA and WRKY transcription factors are related to ABA signaling and seed germination
At present, it is not completely clear how ABA regulates seed germination and early seedling growth. In recent years, the ABA receptors PYR/PYL/RCAR (Park et al. 2009) and many downstream genes have been found to affect seed germination in Arabidopsis. Two SNF1-related protein kinase 2 genes (SnRK2.2 and SnRK2.3) are expressed mainly in seeds, are induced by higher ABA concentrations and can activate ABI5 (ABA-insensitive 5), ABFs (ABRE-binding factors) and EEL (Enhanced late Embryogenesis abundant Level), and repress AtRAV1 (Feng et al. 2014). AtEM1 and AtEM6 (Early Methionine-labeled 1 and 6) are two important LEA protein genes. ABA strongly induces ABI5 expression, and then ABI5 activates ABI3, AtEM1 and AtEM6 by binding to their promoters, ultimately resulting in arrested germination and early growth (Gaubier et al. 1993).
In addition to the above-mentioned well-studied basic components, new factors involved in ABA signaling/seed germination have also been identified. WRKY is the largest transcription factor family in plants, and its members contain the conserved WRKY domain and zinc finger motif, and bind to the W-box (T)(T)TGAC(C/T) motif in target gene promoters (Huang et al. 2016). Recently, it has been found that several WRKY transcription factors participate in ABA signaling for seed germination. AtWRKY6 is located upstream of AtRAV1 in the ABA signaling pathway and can directly down-regulate AtRAV1 and directly up-regulate SnRK2.3, SnRK2.6, EM1 and EM6 (Huang et al. 2016). Although the ABI3 and ABI5 promoters contain the W-box, AtWRKY6 cannot directly regulate their expression (Huang et al. 2016); conversely, AtRAV1 can negatively regulate ABI3 and ABI5 expression by directly combining with their promoters (Feng et al. 2014). AtWRKY41 can control seed dormancy via directly regulating ABI3 expression (Ding et al. 2014). AtWRKY18, AtWRKY40 and AtWRKY60 are negative regulators of germination in ABA signaling, and knockout mutants of these three genes show ABA-hypersensitive phenotypes and arrested germination and post-germination growth (Shang et al. 2010). The knockout mutants of AtWRKY63 (Ren et al. 2010) and AtWRKY2 (Jiang and Yu 2009) are hypersensitive to exogenous ABA during seed germination and the vegetative growth stage. Recently, Shang et al. (2010) found a unique high concentration ABA signaling pathway from early signaling events. AtWRKY40 was recruited from the nucleus to the cytosol and promoted ABAR–AtWRKY40 interaction. ABAR is a chloroplast/plastid protein that functions as a receptor of ABA in Arabidopsis. ABAR relieved the inhibition of the ABI5 gene by repressing AtWRKY40 expression and regulated seed germination (Shang et al. 2010). These results show that WRKY transcription factors play an important role in ABA signaling for seed germination and early seedling growth.
As mentioned above, five W-boxes and four ABREs were found in the 5’ upstream promoter region of the BoNR8 gene (Supplementary Fig. S1), and the BoNR8 lncRNA was induced by a high ABA concentration (50 µM) during seed germination in cabbage (Fig. 5B). BoNR8-AtOX seeds showed an ABA-insensitive phenotype that promoted germination at high ABA concentrations (75–100 µM) (Fig. 8), and BoNR8 overexpression in Arabidopsis clearly promoted AtWRKY6, ABI3 and ABI5 expression under a higher ABA concentration (75 µM) (Fig. 9). One copy of the WRKY-binding motif (W-box) was found in the EM1 promoter, two in the SnRK2.3 promoter and three in the EM6 promoter. AtWRKY6 may directly increase SnRK2.3, EM1 and EM6 expression, but does not regulate SnRK2.2 expression (Huang et al. 2016). Consistent with this, the AtWRKY6 mRNA was greatly accumulated in BoNR8-AtOX seeds treated with higher ABA concentrations (Fig. 9D) and did not change SnRK2.2 (Fig. 9B), but inhibited SnRK2.3 (Fig. 9C). Our data support the possibility that the BoNR8 lncRNA can also regulate the expression of other WRKY genes under high ABA concentrations, so we need to investigate this further in detail.
BoNR8 lncRNA is possibly related to ABA-responsive genes in Arabidopsis seedlings
Some studies have shown that deletion of WRKY6 induces AtRAV1 expression in seedlings under normal conditions, but has no effect on root elongation (Huang et al. 2016). AtRAV1 overexpression inhibited root elongation (Hu et al. 2004) and ABI3, ABI5, EM1 and EM6 expression, but did not affect SnRK2.2 or SnRK2.3 expression (Huang et al. 2016). In this study, the root elongation of BoNR8-AtOX seedlings was weaker than that of the WT (Fig. 6), SnRK2.2, SnRK2.3 and AtRAV1 mRNA levels were higher, and WRKY6, ABI3, ABI5, EM1 and EM6 mRNA levels were lower than in the WT (Fig. 9). Therefore, excess BoNR8 lncRNA may inhibit the root elongation of Arabidopsis seedlings by down-regulating WRKY6 and up-regulating SnRK2.2 and SnRK2.3, which affect AtRAV1 expression through its transcription and phosphoregulation status (Figs. 9,10). Additionally, we also noted that both EM1 (Fig. 9H) and EM6 (Fig. 9I) in seedlings and seeds were inhibited by BoNR8 lncRNA overexpression. However, the mechanism of EM1 and EM6 regulation is still unclear. In this study, we could not rule out the possibility that the BoNR8 lncRNA participates in other regulatory processes involved in root growth.
Materials and Methods
Plant materials
Cabbage (Brassica oleracea var. capitata, cv. Jingfeng No. 1) and Arabidopsis thaliana L. accession Columbia (Col-0) were used. The seeds were sterilized once with 75% (v/v) ethanol and with 5% (v/v) sodium hypochlorite. To break dormancy, the seeds were imbibed on 1/2 MS phytoagar (0.8%) containing 3% sucrose at 4°C for 72 h under dark conditions. Then, they were transferred to a growth chamber at 22°C with a 16 h light/8 h dark cycle.
The BoNR8 cDNA was amplified by RT–PCR from cabbage total RNA using specific primers with XbaI and SacI sites (BoNR8 forward, 5′-GCTCT AGA AAC GGG GTG GGC CCC AGG AG-3’; and BoNR8 reverse, 5′-CCGAG CTC AAA TTT GGG GGT GGG AGG GA-3’; XbaI and SacI sites are underlined) (Fig. 1C; Supplementary Fig. S6). The amplified fragment was cloned into the pBI121 binary vector to generate pBI121-BoNR8 (Fig. 6A). The plasmid was sequenced and transformed into Agrobacterium tumefaciens strain EHA105. Plant transformation was done by the floral dip method (Clough and Bent 1998). Three independent transgenic plants were screened using 50 µg ml–1 kanamycin and Northern blot analyses. Homozygotic T3 generation plants were used for further analyses.
Stress treatment
Cabbage seeds at 2 DAI on 1/2 MS phytoagar (0.8%) containing 3% sucrose at 4°C in the dark were transferred to various stress conditions (150 mM NaCl, 250 mM mannitol, 10 μM IAA, 10 μM NAA, 10 μM 2,4-D and 50 μM ABA) for 24 h at 4°C in the dark. Cabbage seedlings at 1 DAG on 1/2 MS phytoagar (0.8%) were also transferred to 150 mM NaCl, 250 mM mannitol, 10 μM IAA, 10 μM NAA, 10 μM 2,4-D and 50 μM ABA for 24 h.
Measurement of germination rate and root growth
The same batch of seeds was used for germination rate and root length measurements. Thirty seeds of each genotype were subjected to various concentrations of NaCl or ABA at 4°C for 72 h in the dark and incubated at 22°C with a 16 h light/8 h dark cycle for 7 d, and then the germination rate was counted. For root growth comparison, six seedlings (3 DAG) were transferred from phytoagar plates to fresh 1/2 MS plates. These seedlings were grown vertically for 7 d at 22°C with a 16 h light/8 h dark cycle, and then their root lengths were measured.
In vitro transcription of the BoNR8 lncRNA gene
For in vitro transcription, template plasmids including the BoNR8 lncRNA (pYY1103), AtR8 lncRNA and Arabidopsis U2 snRNA genes and promoters were constructed using the pBluescript II KS+ vector, and amplified in Escherichia coli strain XL1-Blue. In vitro transcription of the genes was performed as previously described (Yukawa et al. 1997, Wu et al. 2012). De novo transcripts were produced from 0.4 pmol circular plasmids in a 30 μl reaction with 60 μg of nuclear soluble protein prepared from cultured tobacco BY-2 cells. After incubation at 28°C for 90 min, the resultant transcripts were purified and subjected to a primer extension assay using 5′ fluorescein-labeled primers (BoNR8-PE, 5’-CGG CCA AAG ACC CCC CAA CAG CTT AAA ATC-3’ for BoNR8 RNA; AtR8-PE, 5’-ACG CCG GCC TAA GAC CCC CAG TAA TTA ATT-3’ for AtR8 RNA; and T7DKS-PE, 5’-GGG CGA ATT GGA GCT CCA CCG CGG TGG CGG CCG CTC TAG A-3’ for U2 snRNA). The extended products were fractionated by PAGE and detected and quantified using a Typhoon 9400 analyzer (GE Healthcare).
5’ and 3’ end mapping of the BoNR8 lncRNA
To identify the 5′ end of the BoNR8 lncRNA, primer extension analysis was performed with the BoNR8-PE primer as previously described by Yukawa et al. (1997). Sequencing ladders from BoNR8 were prepared using the BoNR8-PE primer, the pYY1103 and Thermo Sequenase Fluorescent Labelled Primer Cycle Sequencing Kit (GE Healthcare).
The 3′ ends of the RNA were determined using a 3′ RACE method. Cabbage total RNA was 3′ polyadenylated by poly(A) polymerase (TAKARA BIO INC.), and cDNA was synthesized using a d(T) adaptor primer (5′-GGC CAC GCG TCG ACT AGT ACT TTT TTT TTT TTT TTT T-3′). The obtained cDNA was amplified by PCR with an Abridged Universal Amplification Primer (5′-GGC CAC GCG TCG ACT AGT AC-3′) and a GSP3 primer (5′-AAT TGA GTT GGG CGG GTT GGT TTG-3′). After cloning of the cDNAs, the 3′ end sequences were determined.
Southern blot analysis
A DIG end-labeled DNA probe was amplified from cabbage genomic DNA (422 bp, the region from −145 to +277 containing the USE, TATA-like sequence and transcribed regions of the BoNR8 lncRNA gene) using the forward primer 5′-TCA CTA GTG ACG AAA CAG TG-3′ and reverse primer 5′-GAA CAA AAT TTG GGG GTG G-3′.
Next, 15 μg of cabbage and Arabidopsis genomic DNA was digested with EcoRI, HindIII and SalI, separated by 0.8% agarose and blotted onto a Hybond N+ membrane (GE Healthcare). The membrane was incubated with hybridization solution [7% SDS, 50% formamide, 5× SSC, 2% blocking reagent (Roche Lifescience), 50 mM Na-phosphate (pH 7.0), 0.1% sarcosyl] with the DNA probe overnight at 50°C. The hybridization signals were detected using CDP-Star (Roche Lifescience) and a LAS-4000 (GE Healthcare) instrument.
RNA expression assays
Total RNAs of seedlings were extracted with TRIzol Reagent (Invitrogen). The low molecular weight RNAs and high molecular weight RNAs from cabbage seeds were extracted as described by Martin et al. (2005). Then, 5 μg of low molecular weight RNAs and DIG riboprobe containing the AtR8 lncRNA transcribed regions were subjected to Northern blot analysis as previously described (Wu et al. 2012). Signals were detected with a LAS 4000 instrument and quantified as integrated luminescent values with Multi Gauge v3.2 software (Fujifilm) from the chemiluminescence image. Each true signal value was calculated by subtraction of the background area intensity from the signal area intensity.
cDNA was prepared from 2 μg of Arabidopsis total RNA after various stress treatments using the PrimeScript RT Reagent Kit with gDNA Eraser (TAKARA BIO INC.). Next, 50 ng of cDNA and 40 nM primers (see Supplementary Table S1) were used for each RT–qPCR with 2× Brilliant III SYBR Green QPCR Master Mix (Agilent) and a Mx3000P Real-Time Thermal Cycling System (Agilent). The expression levels were normalized with the AtACT2 (At3g18780) gene.
In situ hybridization
Cabbage seedlings (2 DAG) were used for WISH, and transverse sections (10 μm) of the root elongation zone from cabbage seedlings (2 DAG) were used for ISH. The DIG riboprobe containing the AtR8 lncRNA transcribed regions was used and the procedures were as described previously (Yamauchi et al. 2004, Wu et al. 2012).
Statistical analysis
The data were analyzed using one-way analysis of variance by SPSS, and statistically significant differences were calculated based on Student’s t-test, with P < 0.05 (*) and P < 0.01 (**) as thresholds for significance.
Funding
This work was supported by the Fundamental Research Funds for the Central Universities [DL12BA38 and DL13EA04-02]; Scientific Research Foundation for Returned Scholars of Ministry of Education of China [47th Batc]; and the Key Laboratory of Saline-alkali Vegetation Ecology Restoration, Ministry of Education (Northeast Forestry University) [SAVER1702].
Disclosures
The authors have no conflicts of interest to declare.
Footnotes
Subject areas: growth and development; environmental and stress responses
