Hierarchical Action of Mulberry miR156 in the Vegetative Phase Transition

The vegetative phase transition is a prerequisite for flowering in angiosperm plants. Mulberry miR156 has been confirmed to be a crucial factor in the vegetative phase transition in Arabidopsis thaliana. The over-expression of miR156 in transgenic Populus × canadensis dramatically prolongs the juvenile phase. Here, we find that the expression of mno-miR156 decreases with age in all tissues in mulberry, which led us to study the hierarchical action of miR156 in mulberry. Utilizing degradome sequencing and dual-luciferase reporter assays, nine MnSPLs were shown to be directly regulated by miR156. The results of yeast one-hybrid and dual-luciferase reporter assays also revealed that six MnSPLs could recognize the promoter sequences of mno-miR172 and activate its expression. Our results demonstrate that mno-miR156 performs its role by repressing MnSPL/mno-miR172 pathway expression in mulberry. This work uncovered a miR156/SPLs/miR172 regulation pathway in the development of mulberry and fills a gap in our knowledge about the molecular mechanism of vegetative phase transition in perennial woody plants.


Introduction
During their post-embryonic development, plants need to go through three vital developmental processes from seed to mature plant, including seed germination, vegetative phase transition (juvenile-to-adult transition), and flower transition [1]. Vegetative phase transition determines the time of plant vegetative growth, and plants must go through sufficient vegetative growth from juvenile to maturity and eventual reproductive growth [2].
SPL genes have been shown to be repressed by miR156, and the miR156/SPLs/miR172 regulation pathway in vegetative phase transition was systematically revealed in Arabidopsis thaliana [3]. From seedlings to mature plants, the expression level of miR156 gradually decreased, while the expression levels of SPLs and miR172 gradually increased in model annual forbs, such as A. thaliana [3], Nicotiana benthamiana [4,5], and rice [6,7]. In A. thaliana, the roots were not involved in vegetative phase transition regulation; however, sugar in the leaves could trigger the vegetative phase transition by the suppression of miR156 expression [4,8,9]. Perennial woody plants had significantly longer vegetative growth times than annual plants, thus it is of more pragmatic value to study the vegetative phase transition mechanism of perennial woody plants [10]. Compared with annual plants, miR156 and miR172 showed similar expression trends in vegetative phase growth in some perennial woody plants such as Acacia confusa, Acacia colei, Eucalyptus globulus, Hedera helix, Quercus acutissima, P. × canadensis and some tropical/subtropical trees, and the over-expression of miR156 in P. × canadensis delayed its vegetative phase change [11,12]. Researchers also found that the expression of SPLs decreased under heavy fruit load in Citrus clementina, and Citrus SPL was able to promote flowering independent of photoperiod in Arabidopsis [13]. EjSPL4, EjSPL5, and EjSPL9 could significantly activate the promoters of EjSOC1-1, EjLFY-1, and EjAP1-1 in loquat and the over-expression of EjSPL3, EjSPL4, EjSPL5, and EjSPL9 in A. thaliana could promote obvious flowering [14]. However, the hierarchical action of miR156 in vegetative phase growth in perennial woody plants is still unknown.
Mulberry is a typical perennial woody plant with very important economic and medicinal values [15,16]. Mulberry trees have been utilized by humans for thousands of years, the leaves are used for raising silkworms, and the fruit is consumed fresh [17]. The medicinal value of mulberry has been explored, and abundant flavonoids and polysaccharides have been detected, especially in the leaves and fruits [18,19]. Mulberry plants also show strong resistance to abiotic stresses, and this crop can significantly improve the eco-environment [20]. Like other perennial woody plant breeding [21], a long juvenile period limits the mulberry resource development and utilization. In recent years, the mulberry genome sequencing has been gradually completed [16,22] and high-throughput sequencing of small RNA also has been finished in some varieties of mulberry [23,24]. All of these have laid a foundation for basic research on mulberry.
The molecular mechanism of the juvenile-to-adult transition in mulberry, however, is completely unknown. In this study, seven miR156s (mno-miR156a to mno-miR156g) and six miR172s (mno-miR172a to mno-miR172f) were identified using small RNAs highthroughput sequencing. Fourteen MnSPL genes were cloned. Our data revealed that miR156c directly repressed the transcription levels of nine MnSPL genes, including Mn-SPL2, MnSPL15, and MnSPL16A. Six MnSPL genes (MnSPL2, MnSPL8, MnSPL10A, Mn-SPL10B, MnSPL15, and MnSPL16A) could recognize the GTAC cis-element in the promoter region of mno-miR172a and activate the expression of mno-miR172a based on yeast onehybrid experiments. An existing miR156/SPLs/miR172 pathway involved in the mulberry vegetative phase transition has been proposed.

Identification of the miR156 Target Genes in Mulberry
Our results from in silico prediction showed that 115 genes from the mulberry genes database (https://morus.swu.edu.cn/, accessed on 3 March 2019) were selected as the presumptive miR156 target genes (Table S1). GO analysis showed that over half of the presumptive miR156 target genes were clustered into the secondary level "binding" group in the first level molecular function group ( Figure 2).
For further analysis of these presumptive target genes, we counted those genes whose expectation values were less than 2.5, and found seven SPL genes (Morus010792, Morus010123, Morus015493, Morus018032, Morus014488, Morus017456, and Morus026457), three WD40 repeat-like genes (Morus026689, Morus002371, and Morus011698), one fructose 1,6-diphosphatase gene (Morus017876) and two genes (Morus021290 and Morus006566) with no annotation ( Table 2). The results from the degradome sequencing verified seven SPL genes, including MnSPL2, MnSPL4, MnSPL6, MnSPL13, MnSPL15, MnSPL16A, and Mn-SPL16B, with the p values less than 0.05. The other genes with p values more than 0.05 were not considered as miR156 target genes ( Figure 3 and Table 3).     Except for MnSPLs, no other genes were identified by degradome sequencing, which demonstrated that the SPLs genes were the main or exclusive target genes of miR156 in mulberry. Moreover, our results from the RT-qPCR showed that all of MnSPLs except for MnSPL2 had a higher expression level in elder leaves, suggesting the opposite expression profile to mno-miR156 (Figure 4a,b). Next, all of MnSPL genes predicted as miR156 target genes by psRNATarget were tested using dual-luciferase assays. Compared to the CK group, the pGreenII 0800 LUC vector with the supposed miR156 splice site of six MnSPLs (MnSPL2, MnSPL7, MnSPL14, MnSPL15, MnSPL16A, and MnSPL16B) showed significantly lower luciferase activity when miR156 was over-expressed (Figure 4c,d). We integrated the results of degradome sequencing and dual-luciferase assays, and identified nine MnSPLs (MnSPL2, MnSPL4, MnSPL6, MnSPL7, MnSPL13, MnSPL14, MnSPL15, MnSPL16A, and MnSPL16B) as miR156 target genes in mulberry (Figure 4e).

MnSPLs Bound to the Promoter of mno-miR172a and Promoted Its Expression in Mulberry
The result of the BLASTN and HMMER search showed there were 15 MnSPL genes in mulberry ( Figure S1a and Table S3). All of the 15 MnSPL genes had the SBP conserved domains ( Figure S1a). The neighbor-joining (NJ) phylogenetic tree for the SPL genes from M. notabilis, P. trichocarpa, and A. thaliana showed that the MnSPL genes were separately distributed to five groups ( Figure S1b). Tissue expression profiles showed that MnSPL genes were differentially expressed in five tissues (winter bud, male flower, root, branch bark, and leaf). MnSPL6, MnSPL8, MnSPL10, MnSPL15, and MnSPL16A had significantly high expression in winter buds, while MnSPL2, MnSPL7, and MnSPL16B had much higher expression in male flowers. MnSPL1, MnSPL13, and MnSPL14 were mainly expressed in the root, and MnSPL4 and MnSPL12 showed much higher expression in the leaf and branch bark, respectively ( Figure S2). Six highly conserved mno-miR172s were identified in mulberry (Figure 5a,b), and the expression of mno-miR172a was significantly higher than the mno-miR172s in mulberry roots, barks, and leaves (Table 1). The miR172 had a higher expression level in 9-month-old leaves than in younger leaves, which indicated that miR172 might be involved in the regulation of vegetative development in mulberry (Figure 5c). GUS activity was detected in CK, OE-156, and STTM156cd samples, indicating that genes were over-expressed by transient transgenic technology (Figure 6a,b). The over-expression of miR156 in mulberry leaves significantly decreased the expression of miR172 and MnSPL genes, except for MnSPL1 and MnSPL12, while the down regulation of miR156 by STTM in mulberry seedlings significantly increased the expression of miR172 and most MnSPL genes except for MnSPL1, MnSPL8, MnSPL12, and MnSPL14 ( Figure 6). We wondered if MnSPLs acted as upstream regulators of mno-miR172 and how MnSPLs functioned in this process. Thus, yeast one-hybrid assays were carried out to identify which MnSPL genes could recognize the predetermined GTAC elements in the mno-miR172a promoter. Except for three genes (MnSPL6, MnSPL8, and MnSPL12), which were toxic in yeast, the remaining 11 MnSPL genes were tested by the yeast one-hybrid assay, and six of those genes (MnSPL2, MnSPL7, MnSPL10A, MnSPL10B, MnSPL15, and MnSPL16A) could grow normally on the -Leu/A 150 screening medium (Figure 7a,b). Then, we used dual-luciferase assays to further verify these predicted miR156 target MnSPL genes (Table S2) and found that six MnSPL genes (MnSPL2, MnSPL8, MnSPL10A, MnSPL10B, MnSPL15, and MnSPL16A) could increase luciferase activity. Taking the results of the yeast one-hybrid and dual-luciferase assays into consideration, we confirmed that six MnSPL genes (MnSPL2, MnSPL8, MnSPL10A, MnSPL10B, MnSPL15, and MnSPL16A) directly bound to the promoter of mno-miR172a and promoted its expression in mulberry.

Discussion
Woody plants were favorable systems for studying vegetative phase change because the stability and prolonged duration of the various stages of shoot development make these phases easy to observe and characterize [25]. However, the stability and prolonged duration of the vegetative stages limit breeding and genetic research in perennial woody plants. The miR156 over-expression can suppress vegetative phase transition and promote vegetative growth in poplar [11]. To date, however, researchers are still unclear about the molecular mechanism of vegetative phase development in perennial woody plants due to the limitations of the genetic research. Therefore, understanding the molecular mechanism of this process in perennial woody plants has very important research and practical value.
The sequence of miR156 was shown to be highly conserved among angiosperms [26]. Small RNA sequencing in mulberry showed that the expression level of mno-miR156c in three tissues (leaves, bark and male flowers) was significantly higher than other miR156 family members, and mno-miR156c also was shown to have the same sequences as miR156 found in A. thaliana, G. max, and M. domestica [23]. Similar to the discovery in A. thaliana [3] and P. trichocarpa [11], miR156 was expressed at higher levels in juvenile mulberry tissues than in mature mulberry tissues, which suggested that this conserved miRNA was probably involved in the vegetative phase transition in mulberry. In A. thaliana, AtSPLs were downstream factors of miR156-regulated vegetative stage development [3,27,28]. The SPL family gene function as promoters of vegetative phase transition and are distributed across angiosperms [29][30][31][32]. In total, there are 16 AtSPL genes in A. thaliana and 10 members (AtSPL2, AtSPL3, AtSPL4, AtSPL5, AtSPL6, AtSPL9, AtSPL10, AtSPL11, AtSPL13, and AtSPL15) were the miR156 target genes [3]. A total of 15 MnSPL genes were predicted by BLASTN and HUMMER modeling in MorusDB, and 14 gene sequences with an SBP-box domain were cloned successfully (MnSPL10 has two alternative spliceosomes named MnSPL10A and MnSPL10B, but MnSPL3 and MnSPL5 were not cloned). A total of nine MnSPLs (MnSPL2, MnSPL4, MnSPL6, MnSPL7, MnSPL13, MnSPL14, MnSPL15, MnSPL16A, and MnSPL16B) were identified as miR156-targeted genes in mulberry in the present work. This finding implied that the SPL genes were the exclusive miR156 target genes across herbs and woody plants. The miR156/SPLs modules can also control the development of lateral roots and miR156-targeted SPL10-controlled root meristem activity in A. thaliana [29,33]. We also found three MnSPL genes (MnSPL1, MnSPL13, and MnSPL14) prominently expressed in the root, which implied that these genes might be involved in root development regulation.
As a conserved miRNA, miR172 participated in the regulation of flowering and was regulated by SBP-box transcription factors in A. thaliana [34,35]. Additionally, the miR172 expression level increased with age in A. thaliana and P. × canadensis [11,36]. In A. thaliana, SPL9/10 activates the vegetative phase transition by promoting miR172 expression [3], while the over-expression of miR172 in A. thaliana can increase the expression of SPL3/4/5 [37]. In mulberry, there were six mature miR172 transcripts, named mno-miR172a-f, and six MnSPL genes (MnSPL2, MnSPL8, MnSPL10A, MnSPL10B, MnSPL15, and MnSPL16A) were verified as the direct upstream regulators of mno-miR172a. Additional miR172-upstream MnSPL genes were identified in mulberry, which suggested there were more complex regulatory relationships between SPLs and miR172 in mulberry.
In summary, we found that the expression of miR156 subsequently reduced as mulberry aged, while SPLs and miR172 showed the opposite profile. A total of nine MnSPL genes as exclusive miR156 target genes were verified and six MnSPL genes could promote the expression of mno-miR172a in mulberry. The over-expression of miR156 reduced the expression levels of miR172 and SPL genes, except for two miR156 non-target genes (MnSPL1 and MnSPL12) and decreasing the expression level of miR156 by STTM increased the miR172 and SPL genes expression levels, except for three non-target genes of miR156 (MnSPL1, MnSPL12, and MnSPL14). All of these experimental results indicated that the miR156/SPLs/miR172 regulation network also existed in mulberry, which implied this regulation pathway was conserved across angiosperms ( Figure 8). Our work should facilitate the integration of information about vegetative phase transition across perennial woody plants and give significant insight into how to reduce the breeding cycle of perennial woody plants.

Plant Materials and Growth Conditions
Tobacco seeds (N. tabacum L) were planted in sterilized soil and left at 4 • C for 2 d before transfer to a climate chamber at 25 • C under a long-day condition (16 h light/8 h dark), as were mulberry seedlings (Morus atropurpurea cv. Guiyou 62, abbreviated as GY62, which can bloom after 13 months). The older mulberry trees were grown on the open balcony of our laboratory. Plant age was measured from the date that seeds were transferred to the growth chamber.
Various mulberry tissues were used to analyze the expression of miRNAs and genes in this work. Young leaves, barks, and roots from 3-month-old (3 M), 11-month-old (11 M) and 16-month-old (16 M) mulberry were used to detect the expression of miR156 in mulberry by RT-qPCR. Young leaves from juvenile phase mulberry (J-YL, 3 months old) and young leaves of mature phase mulberry (M-YL, 16 months old) were used to analyze miR156 expression level by small RNA Northern blot assay. Young leaves of 2-month-old (2 M), 4-month-old (4 M), and 9-month-old (9 M) mulberry were used to detect the expression profile of miR172 in mulberry by RT-qPCR. Mature leaves from juvenile phase mulberry (J-ML) and mature leaves from mature phase mulberry (M-ML) were used to investigate the expression level of MnSPLs.

Epression Analysis of miRNAs and Genes
Small RNAs (<200 bp) and total RNA were both extracted using the miRcute Plant miRNA Isolation Kit purchased from TIANGEN (DP504, Beijing, China, OD 260/280 of all RNA samples were determined using a NanoDrop2000 (Thermo Scientific, Waltham, MA, USA), and agarose gel electrophoresis (AGE) was used to verify the integrity of RNA samples (data not shown). RNA probes were labeled with DIG by Riboprobe ® Systems (P1440, Promoga, Madison, WI, USA) and purified using an EasyPure ® RNA Purification Kit (ER701-1, TransGen Biotech, Beijing, China). Northern blots were performed as follows: twenty micrograms of small RNA were isolated on a 15% Ura-PAGE gel, transferred to a membrane using a Semi-Dry Transfer apparatus, cross-linked using ultraviolet crosslinking, and visualized using the protocol of a DIG DNA Labeling and Detection Kit (11093657910, Roche, Switzerland). Small RNA reverse transcriptions were performed using the miRNA First Strand cDNA Synthesis Kit (tailing reaction) (B532451, Sangon Biotech, Shanghai, China). The levels of mature miR156 and miR172 were quantified using the MicroRNAs qPCR Kit (SYBR Green method) (B532461, Sangon Biotech, Shanghai, China). mno-miR166b and MnU6 were selected as reference genes together in our small RNA RT-qPCRs. Heatmaps were made using TBbtools [39], and the mimax normalization method was utilized to normalize this data. The PrimeScript™ RT reagent Kit with gDNA Eraser (perfect real time) (RR047Q, Takara, Japan) was used to reverse transcribe total RNA to cDNA. Quantitative reverse transcription polymerase chain reaction (RT-qPCR) was performed to analyze the transcript abundance of various SPL genes using the TB Green ® Premix Ex Taq™ II (Tli RNaseH Plus) (RR820Q, Takara, Japan). All primers and probes used in this work are listed in Table S4.

Transient Trnsgene Expression in Mulberry
An 800-bp-predicted pre-miR156 sequence was cloned from the genome of GY62 into the pCAMBIA1301 vector to over-express mature miR156. Short tandem target mimicking (STTM) was performed to repress the expression of mature miR156. The STTM vectors were constructed as described previously [40,41]. For STTMd vector, three extra bases (cta) were inserted into the reverse complement sequences of mno-miR156c and mno-miR156d between the 11th and 12th bases. These two modified sequences were inserted into the end of 88-bp spacer sequence which could form the hairpin structure. pCAMBIA1301 was used to over-express this construct. The protocol of transient transgene expression of pCAMBIA-156 and pCAMBIA-STTM156cd in mulberry was as described in [42]. For GUS staining aay, samples were incubated with staining buffer (G3061, Solarbio, Beijing, China) at 37 • C for 12 h. The stained samples were washed by 75% (v/v) ethanol and photographed using a Leica microscope. All the primers and probes used in this work are listed in Table S4.

Dual-Luciferase Repoter Assay
A 200-bp sequence containing the predicted cleavage of target SPL genes was cloned into the pGreenII 0800-LUC vector, which had the multiple cloning site after the 3 end of the luciferase gene. The resultant vector was verified by sequencing. pCAMBIA-156 was used in dual-luciferase assays to over-express mature mno-miR156c. Vector structures are showed in (Figure 4c). Agrobacterium-mediated co-transformation of the pGreenII 0800-LUC and pCAMBIA-156 vectors into tobacco leaves was performed as described previously [43]. After infiltration fothree days, the ratio of LUC to REN activity was measured using the Dual-Luciferase Reporter Gene Assay Kit (11402ES80, Yeasen, Shanghai, China) on a configurable multi-mode microplate reader (Synergy™ H1, BioTek, Beijing, China). MnSPL geneloned into theGreenII 62-SK vector. A 1500-bp promoter sequence from mno-miR172a predicted by Promoter 2.0 Prediction Server (http://www. cbs.dtu.dk/services/Promoter/, accessed on 31 May 2020) was amplified and cloned into the pGreenII 0800-LUC vector. Vector structures are shown in (Figure 7c). Agrobacteriummediated co-transformation and LUC/REN activity detection were processed as described above. All the primers and probes used in this work are listed in Table S4.

Yeast One-Hybrid Assay
The fulllength sequences of SPL genes were amplified and then cloned into the pGADT7 vector. Because the full length of three genes were over 3000 bp, only partial sequences (1000 bp) from Morus024784, Morus013868 and Morus025152 were amplified. A 50-bp sequence containing a predicted cis-acting element was repeated three times and cloned into the pAbAi vector. A mutant cis-acting element (modified from GTACTGTAC to GATCTGATC) was used as a negative control. A verified concentration (150 ng/mL) of Aureobasidin A (AbA) was used to screen positive yeast strains in a yeast one-hybrid system. The assays were performed using the Matchmaker ® Gold Yeast One-Hybrid System (630491, Takara, Japan) according to the manufacturer's instructions. All the primers and probes used in this work are listed in Table S4.

Degradome Sequencing
We selectedroots, leaves, and bark from juvenile mulberry and winter buds from mature mulberry to constructed a degradome library. The purified cDNA library was sequenced on an Illumina HiSeq2000 Sequencing System (LC-BIO Sciences, Beijing, China).