Identification of Low-Nitrogen-Related miRNAs and Their Target Genes in Sugarcane and the Role of miR156 in Nitrogen Assimilation

Chemical nitrogen (N) fertilizer is widely used in sugarcane production, especially in China and India. Understanding the molecular mechanisms and mining miRNAs and their target genes associated with nitrogen use efficiency (NUE) in sugarcane can aid in developing the N-efficient varieties, and thus is beneficial to reduce N fertilizer application. In this study, the root miRNA database of N-efficient sugarcane variety ROC22 under low N stress (0.3 mM NH4NO3) for 3 h was constructed, along with their transcriptome-rearranged data. KEGG analysis indicated that those candidate target genes, corresponding to differentially expressed miRNAs, were mainly enriched in N metabolism, amino acid metabolism, carbohydrate metabolism, photosynthesis, and hormone signal transduction pathways. It was found that under low N stress for 0–24 h, there was a negative correlation between miR168 and SPX, along with miR396 and acnA. Furthermore, the expression of miR156 in the roots of ROC22 was significantly up-regulated under low N treatment. Compared with the wild-type, the Arabidopsis plants overexpressing sugarcane miR156 exhibited significantly improved length and surface area of roots, while the expression of one NO3− transporter gene NRT1.1, three N assimilation key genes (NR1, NIR1, and GS), and the activity of two N assimilation key enzymes (NR and GS) were up-regulated under low N treatment. It can be reasonably deduced that sugarcane miR156 can enhance the nitrogen assimilation ability of the overexpressed Arabidopsis plants under low N application, and thus has a potential ability for improving sugarcane NUE. The present study should be helpful for understanding the molecular regulatory network in the N-efficient sugarcane genotype responding to low N stress and could provide the candidate miRNAs with a potential function in improving sugarcane NUE.


Introduction
N is an essential macro nutrient element for crops, and it is an important component of nucleic acid, phospholipid, and protein [1]. Sugarcane is the most important sugar crop [2,3]. N is one of the limiting factors for cane tonnage. Due to the limited arable land in China, sugarcane has been continuously cultivated for 20 to 30 years without rotation in most planting areas. To improve the yield, the N fertilizer application amount per sugarcane season was 400-800 kg/ha in China, which was much higher than that in other countries [4]. Excessive application of chemical N fertilizer has led to a cost increase, Int. J. Mol. Sci. 2022, 23, 13187 2 of 16 acidic soil, eutrophic water, and non-point source pollution, as well as the decrease in sucrose content in sugarcane [4,5]. It is gratifying that there are significant differences in NUE among different genotypes of sugarcane. The cultivation of N-efficient sugarcane varieties is an effective way to improve NUE and thus leads to a reduction in N application. The molecular mechanism of N uptake and utilization in plants is the basis of breeding N-efficient varieties.
In many biological metabolic processes, especially in stress response, miRNAs play an important regulatory role by inducing or inhibiting gene expression [3,6,7]. As reported, miRNAs play an important role in the regulation of plant nutrition [8][9][10]. Under low N stress, plants perceive the signal of N deficiency, amplify and transmit this signal to the responding molecules. Several miRNAs can cause mRNA degradation or translation inhibition of target genes, and regulate the expression of genes related to low N stress at the transcriptional and post transcriptional levels, thus open up the resistance mechanism in plants under low N stress [11].
Understanding the molecular response mechanism of crops under low N stress and analyzing the biological function of miRNA can lay a scientific foundation for improving crop NUE [8,12]. The miRNAs profiles and their regulatory mechanisms under low N stress were investigated in several kinds of plants, such as Arabidopsis thaliana [13], Oryza sativa [14], Zea may [9,15], Coffea arabica [16] and Dendranthema morifolium [12], etc.
MiR156 plays an important role in plant growth and development [17,18] and in the adversity [8,12,19,20]. The expression of miR156 could be induced by salt and drought in rice [21] and maize [22], and low temperature in rice [21]. The up-regulated miR156 might have contributed to the growth of peanut roots under K deficiency [23]. The miR156 could even regulate root regeneration and N fixation activity in alfalfa [24]. However, there has been no report of miRNA involvement in the regulatory response mechanism under low N stress in sugarcane until now. Therefore, the differentially expressed miRNAs and their candidate target genes associated with N utilization in sugarcane roots were analyzed based on the root miRNA in this study and the mRNA transcriptome databases [25] constructed from the N-efficient sugarcane variety ROC22 under low N treatment. In addition, the physiological and molecular regulation mechanism of the transgenic Arabidopsis plants overexpressing sugarcane miR156 in low N treatment were investigated. This study aims to provide the candidate miRNAs with a potential function in improving sugarcane NUE, which should also help to understand the molecular regulatory network in sugarcane response to low N stress.

sRNA Sequencing Data
The quality of sRNA sequencing samples met the requirements of library construction (class A), and the error rate of raw data from the six libraries constructed in this experiment was less than 0.02%, Q20 > 98.00% and Q30 > 97.50% (Table S1). The GC content was 52.00%. Clean reads were obtained by removing low-quality reads and reads with connectors from raw data (Table S2). The sequence of sRNA without annotation was the largest, followed by rRNA. The length distribution of sRNA sequences was shown in Figure S1. The clean sequence length of sugarcane sRNAs was generally 18-30 nt, and the peak value was concentrated at 24 nt which is similar to the peaks of 21 nt and 24 nt in plants [26]. These indicated that the quality of sequencing data was high.

Weighted Gene Co-Expression Network Construction and Module Identification
A total of 78 miRNAs were identified in the six libraries. There were 15 differentially expressed miRNAs in the roots of ROC22 under low N stress (RRT vs. RRCK), including three known miRNAs: miR156 (significantly up-regulated), miR396 (significantly down-regulated) and miR393 (significantly down-regulated). Among the 12 differentially expressed new miRNAs, seven miRNAs (novel-22, novel-23, novel-47, novel-57, novel-122, novel-158, and novel-169) were up-regulated and the other five miRNAs (novel-15, novel-31, novel-173, novel-175 and novel-215) were down-regulated. Figure 1 shows the hierarchical clustering results of the expression patterns of differentially expressed miRNAs.  Figure 1 shows the hierarchical clustering results of the expression patterns of differentially expressed miRNAs.

Prediction and Function of Target Genes of miRNA
In plants, miRNAs participate in various life processes by regulating their target genes [9,11]. Based on the root transcriptome database of ROC22 under low N stress [25] and the sugarcane EST database (about 284,818 ESTs) in NCBI, 10,454 target genes were predicted. GO enrichment and KEGG analysis were then carried out by GOseq and KO-BAS (2.0) software [27]. According to GO analysis ( Figure S2), the 2,056 predicted target genes of differentially expressed miRNAs were involved in the biological process, cellular component, and molecular function. Among them, the predicted target genes were mainly involved in the molecular functions of starch synthase, mannosyltransferase, asparagine hydrolase, and other biological processes, such as abiotic stimulation and hormone response. KEGG analysis showed that the predicted target genes were mainly concentrated in RNA transport, N metabolism, carbohydrate metabolism (starch and sucrose metabolism), amino acid metabolism (cysteine and methionine metabolism; glycine, serine and threonine metabolism; valine, leucine and isoleucine degradation) and hormone signal transduction pathways.

Crosstalk with miRNA and mRNA Transcriptome Sequencing Data
Those target genes in the roots were mainly concentrated in the N metabolism, amino acid metabolism, carbohydrate metabolism, photosynthesis, and hormone signal transduction pathways under low N stress ( Figure 2). In the N metabolism pathway, novel-158 and novel-22 might target GS, and miR393 might target CA, all participated in the N Figure 1. Hierarchical clustering map of the differentially expressed miRNAs. Notes: Clustering with log 10 (TPM + 1) value, red indicated up-regulated miRNAs, and blue indicated down-regulated miRNAs. RRCK, the roots of ROC22 cultured in the low N solution (N source: 0.3 mM NH 4 NO 3 ) for 0 h; RRT, the roots of ROC22 cultured in the low N solution for 3 h.

Prediction and Function of Target Genes of miRNA
In plants, miRNAs participate in various life processes by regulating their target genes [9,11]. Based on the root transcriptome database of ROC22 under low N stress [25] and the sugarcane EST database (about 284,818 ESTs) in NCBI, 10,454 target genes were predicted. GO enrichment and KEGG analysis were then carried out by GOseq and KOBAS (2.0) software [27]. According to GO analysis ( Figure S2), the 2056 predicted target genes of differentially expressed miRNAs were involved in the biological process, cellular component, and molecular function. Among them, the predicted target genes were mainly involved in the molecular functions of starch synthase, mannosyltransferase, asparagine hydrolase, and other biological processes, such as abiotic stimulation and hormone response. KEGG analysis showed that the predicted target genes were mainly concentrated in RNA transport, N metabolism, carbohydrate metabolism (starch and sucrose metabolism), amino acid metabolism (cysteine and methionine metabolism; glycine, serine and threonine metabolism; valine, leucine and isoleucine degradation) and hormone signal transduction pathways.

Crosstalk with miRNA and mRNA Transcriptome Sequencing Data
Those target genes in the roots were mainly concentrated in the N metabolism, amino acid metabolism, carbohydrate metabolism, photosynthesis, and hormone signal transduction pathways under low N stress ( Figure 2). In the N metabolism pathway, novel-158 and novel-22 might target GS, and miR393 might target CA, all participated in the N assimilation process. In the plant hormone signal transduction pathway, miR167 targeted auxin response factor (ARF), and miR393 might target transport inhibitor response 1 (TIR1), both participate in the IAA pathway, novel-22 might target SnRK 2 to participate in the ABA signaling pathway; novel-23 might participate in ETH pathway through targeting EIN 3 , and novel-173 might participate in JA signaling pathway through targeting MYC2. assimilation process. In the plant hormone signal transduction pathway, miR167 targeted auxin response factor (ARF), and miR393 might target transport inhibitor response 1 (TIR1), both participate in the IAA pathway, novel-22 might target SnRK2 to participate in the ABA signaling pathway; novel-23 might participate in ETH pathway through targeting EIN3, and novel-173 might participate in JA signaling pathway through targeting MYC2. The predicted target genes of miRNAs in the roots were also concentrated in alanine, aspartate and glutamic acid metabolism, phenylalanine metabolism, phenylalanine, tyrosine and tryptophan biosynthesis, glycine, serine and threonine metabolism, cysteine and methionine metabolism, lysine biosynthesis ( Figure 3) and the tricarboxylic acid cycle, galactose metabolism, starch and sucrose metabolism, pentose phosphate pathway in carbohydrate metabolism ( Figure 4).  The predicted target genes of miRNAs in the roots were also concentrated in alanine, aspartate and glutamic acid metabolism, phenylalanine metabolism, phenylalanine, tyrosine and tryptophan biosynthesis, glycine, serine and threonine metabolism, cysteine and methionine metabolism, lysine biosynthesis ( Figure 3) and the tricarboxylic acid cycle, galactose metabolism, starch and sucrose metabolism, pentose phosphate pathway in carbohydrate metabolism ( Figure 4).  The predicted target genes of miRNAs in the roots were also concentrated in alanine, aspartate and glutamic acid metabolism, phenylalanine metabolism, phenylalanine, tyrosine and tryptophan biosynthesis, glycine, serine and threonine metabolism, cysteine and methionine metabolism, lysine biosynthesis ( Figure 3) and the tricarboxylic acid cycle, galactose metabolism, starch and sucrose metabolism, pentose phosphate pathway in carbohydrate metabolism ( Figure 4).  In addition, some predicted target genes in the roots were also related to the photosynthetic pathway. Among them, novel-173 might participate in the photosynthetic pathway by targeting PsbP and ATPase (ATPF1D and ATPF0A), and novel-22 might participate in the electron transport in photosynthesis by targeting PetH, while novel-31 might participate in the photosystem I pathway by targeting Psbk, and miR408 might participate in the pathway of the photosystem by targeting plastocyanin (PC) and NiFU ( Figure 2). Among the identified miRNAs, the squamosa promoter binding protein-like gene (SPL, sample|Cluster-9046.200116), which was assumed to be the predicted target gene of miR156 in the roots of ROC22, was down-regulated by 0.61-fold at 3 h under low N stress. miR156 might also participate in the metabolism of alanine, aspartate and glutamate by targeting 1-pyrroline-5-carboxylate dehydrogenase gene (P5CS, sample|Cluster-9046.62773, log 2 FC = −1.4) (Figure 3). In the roots of ROC22, GRF (growth regulating factor, sample|cluster-9046.236180), the predicted target gene of miR396, was up-regulated by 2.48-fold. miR396 might also participate in the tricarboxylic acid cycle by targeting aconitate hydratase (acnA, ncbi|Cluster-1234.10611, log 2 FC= 0.46) (Figure 4). The predicted target gene of miR408, plastocyanin (PC, sample|cluster-9046.239128) was up-regulated by 5.39-fold ( Figure 2). miR393 might participate in IAA pathway by targeting TIR1 (sample|cluster-82409.0, log 2 FC = −0.50) ( Figure 2).
In addition, some predicted target genes in the roots were also related to the photosynthetic pathway. Among them, novel-173 might participate in the photosynthetic pathway by targeting PsbP and ATPase (ATPF1D and ATPF0A), and novel-22 might participate in the electron transport in photosynthesis by targeting PetH, while novel-31 might participate in the photosystem I pathway by targeting Psbk, and miR408 might participate in the pathway of the photosystem by targeting plastocyanin (PC) and NiFU (Figure 2). Among the identified miRNAs, the squamosa promoter binding protein-like gene (SPL, sam-ple|Cluster-9046.200116), which was assumed to be the predicted target gene of miR156 in the roots of ROC22, was down-regulated by 0.61-fold at 3 h under low N stress. miR156 might also participate in the metabolism of alanine, aspartate and glutamate by targeting 1-pyrroline-5-carboxylate dehydrogenase gene (P5CS, sample|Cluster-9046.62773, log2 FC= -1.4) (Figure 3). In the roots of ROC22, GRF (growth regulating factor, sample|cluster-9046.236180), the predicted target gene of miR396, was up-regulated by 2.48-fold. miR396 might also participate in the tricarboxylic acid cycle by targeting aconitate hydratase

Verification of the Transcriptome Sequencing Data of miRNAs and Candidate Target Genes by qRT-PCR
Ten miRNAs and 13 candidate target genes were randomly selected for qRT-PCR expression validation. As shown in Figure 5A,B, the qRT-PCR results of miRNA and candidate target genes were basically consistent with the miRNA sequencing data (miRNA-Seq) and transcriptome sequencing data (RNA-Seq), but the degree of up-regulation or down-regulation was different from the transcriptome, which might be due to the reason that sequencing is more sensitive than qRT-PCR detection [28].

Verification of the Transcriptome Sequencing Data of miRNAs and Candidate Target Genes by qRT-PCR
Ten miRNAs and 13 candidate target genes were randomly selected for qRT-PCR expression validation. As shown in Figure 5A,B, the qRT-PCR results of miRNA and candidate target genes were basically consistent with the miRNA sequencing data (miRNA-Seq) and transcriptome sequencing data (RNA-Seq), but the degree of up-regulation or down-regulation was different from the transcriptome, which might be due to the reason that sequencing is more sensitive than qRT-PCR detection [28].

Expression Trend of miRNA and Candidate Target Genes in Sugarcane under Low N Stress
According to the negative correlation between the expression of the miRNAs and their predicted corresponding target genes in the sequencing data, eight miRNAs and 12 candidate target genes were selected and detected by qRT-PCR in the roots of ROC22 at 0-16 d under low N stress ( Figure 6). Under low N stress for 0 h-1 d, novel-57 was negatively correlated with AUX/IAA, and similar for novel-158 and GS, miR167 and ARF, respectively. Under low N stress for 0-16 d, there was a negative correlation between novel-215 and ACO1. Under low N stress for 3 h, 8 d and 16 d, miR156 was negatively correlated with SPL and P5CS. miR168 was negatively correlated with SPX and AGO at 0-1 d and 1-16 d under low N stress, respectively. The negative correlation between miR396 and acnA was obvious under stress for 0 h-1 d, and so was between NiFU and miR408a/miR408e for 0h-6h.
Compared with the control, the expression level of miR156 was significantly up-regulated by 2.3-, 2.1-and 1.6-fold in the roots of ROC22 under low N stress for 3 h, 6 h and 1d, respectively ( Figure 6). miR156 may be one of the key genes in sugarcane under low N stress. Considering that the genetic transformation efficiency of sugarcane was extremely low and unstable, the physiological and molecular evaluation results of the gene in different plants could also help to understand the biological function of the gene. Therefore, we transformed sugarcane miR156 into Arabidopsis, which laid the foundation for the subsequent identification of the gene function in sugarcane.

Expression Trend of miRNA and Candidate Target Genes in Sugarcane under Low N Stress
According to the negative correlation between the expression of the miRNAs and their predicted corresponding target genes in the sequencing data, eight miRNAs and 12 candidate target genes were selected and detected by qRT-PCR in the roots of ROC22 at 0-16 d under low N stress ( Figure 6). Under low N stress for 0 h-1 d, novel-57 was negatively correlated with AUX/IAA, and similar for novel-158 and GS, miR167 and ARF, respectively. Under low N stress for 0-16 d, there was a negative correlation between novel-215 and ACO1. Under low N stress for 3 h, 8 d and 16 d, miR156 was negatively correlated with SPL and P5CS. miR168 was negatively correlated with SPX and AGO at 0-1 d and 1-16 d under low N stress, respectively. The negative correlation between miR396 and acnA was obvious under stress for 0 h-1 d, and so was between NiFU and miR408a/miR408e for 0 h-6 h.
Compared with the control, the expression level of miR156 was significantly upregulated by 2.3-, 2.1-and 1.6-fold in the roots of ROC22 under low N stress for 3 h, 6 h and 1 d, respectively ( Figure 6). miR156 may be one of the key genes in sugarcane under low N stress. Considering that the genetic transformation efficiency of sugarcane was extremely low and unstable, the physiological and molecular evaluation results of the gene in different plants could also help to understand the biological function of the gene. Therefore, we transformed sugarcane miR156 into Arabidopsis, which laid the foundation for the subsequent identification of the gene function in sugarcane.

Characterization of the miR156-Overexpressed Arabidopsis Plants under Low N Treatment
Root phenotypic characteristics of the ov-miR156 transgenic and wild type Arabidopsis under low N treatment were compared in Figure 7A. The length and surface area of roots in the miR156-overexpressed Arabidopsis plants were 1.39-and 1.31-fold longer than those of wild type under low N treatment ( Figure 7B,C). Compared with the wild type, the expression level of miR156 was significantly up-regulated in the transgenic plants under low N treatment ( Figure 7D). Moreover, three target genes of miR156, i.e., SPL, P5CS1, and P5CS2, all showed a significant downward trend simultaneously ( Figure 7E).

Characterization of the miR156-Overexpressed Arabidopsis Plants under Low N Treatment
Root phenotypic characteristics of the ov-miR156 transgenic and wild type Arabidopsis under low N treatment were compared in Figure 7A. The length and surface area of roots in the miR156-overexpressed Arabidopsis plants were 1.39-and 1.31-fold longer than those of wild type under low N treatment ( Figure 7B,C). Compared with the wild type, the expression level of miR156 was significantly up-regulated in the transgenic plants under low N treatment ( Figure 7D). Moreover, three target genes of miR156, i.e., SPL, P5CS1, and P5CS2, all showed a significant downward trend simultaneously ( Figure 7E).

Changes in the N Assimilation Key Genes Expressions and Enzymes Activities in the miR156-Overexpressed Arabidopsis under Low N Treatment
At the molecular level, the expression levels of ammonium transporter gene AMT1.1, nitrate transporter gene NRT1.1, N assimilation key genes NR1, NIR1, GS and NADH-GOGAT in the miR156-overexpressed Arabidopsis were significantly higher than those in the wild-type plants under low N stress ( Figure 8A).
Under low N treatment, its activity of the NR enzyme in the miR156-overexpressed plants, which limited N primary assimilation rate, was significantly increased and was 3.56-fold higher than that of the wild type. Additionally, the enzyme activities of NIR and GS were 12.30% and 38.03% higher than those of the wild type, respectively ( Figure 8B).

Changes in the N Assimilation Key Genes Expressions and Enzymes Activities in the miR156-Overexpressed Arabidopsis under Low N Treatment
At the molecular level, the expression levels of ammonium transporter gene AMT1.1, nitrate transporter gene NRT1.1, N assimilation key genes NR1, NIR1, GS and NADH-GOGAT in the miR156-overexpressed Arabidopsis were significantly higher than those in the wild-type plants under low N stress ( Figure 8A).
Under low N treatment, its activity of the NR enzyme in the miR156-overexpressed plants, which limited N primary assimilation rate, was significantly increased and was 3.56-fold higher than that of the wild type. Additionally, the enzyme activities of NIR and GS were 12.30% and 38.03% higher than those of the wild type, respectively ( Figure 8B).

The miRNA Data of Sugarcane Root under Low N Stress were Reliable
miRNA plays an important role in the molecular regulatory network of crop response to low N stress [29,30]. Vidal et al. [13] found 40 nitrate-responsive genes and novel miRNAs actively respond to nitrate changes by RNA-seq and sRNA-seq in the roots of Arabidopsis under different N treatments, such as miR5640 and its predicted target gene AtPPC3. After analyzing the expression trends of IncRNAs, miRNAs and predicted target genes in rice under N deficiency, Shin et al. [14] found the targeting relationship between root-specific miR444a4-3p and MADS25. Santos et al. [16] identified 86 miRNAs and 253 predicted target genes in coffee under low N conditions. In this study, 78 miRNAs were identified in sugarcane under low N stress. The peak of sequence length (24 nt) of sRNA in this study is comparable to them in plants (generally 21 nt and 24 nt) [26].

The Predicted Target Genes of Differentially Expressed miRNA were Mainly Concentrated in N Metabolism, Carbohydrate Metabolism, and Amino Acid Metabolism.
miRNA regulates the target genes' expression by binding with 3'UTR of the target gene, inhibiting or degrading mRNA translation [31]. The most direct way to explore the regulation mechanism of miRNA is to study the biological functions of its targeted genes [32], and bioinformatics analysis can better predict the target genes of miRNA [33]. In  [29,30]. Vidal et al. [13] found 40 nitrate-responsive genes and novel miRNAs actively respond to nitrate changes by RNA-seq and sRNA-seq in the roots of Arabidopsis under different N treatments, such as miR5640 and its predicted target gene AtPPC3. After analyzing the expression trends of IncRNAs, miRNAs and predicted target genes in rice under N deficiency, Shin et al. [14] found the targeting relationship between root-specific miR444a4-3p and MADS25. Santos et al. [16] identified 86 miRNAs and 253 predicted target genes in coffee under low N conditions. In this study, 78 miRNAs were identified in sugarcane under low N stress. The peak of sequence length (24 nt) of sRNA in this study is comparable to them in plants (generally 21 nt and 24 nt) [26].

The Predicted Target
Genes of Differentially Expressed miRNA Were Mainly Concentrated in N Metabolism, Carbohydrate Metabolism, and Amino Acid Metabolism miRNA regulates the target genes' expression by binding with 3'UTR of the target gene, inhibiting or degrading mRNA translation [31]. The most direct way to explore the regulation mechanism of miRNA is to study the biological functions of its targeted genes [32], and bioinformatics analysis can better predict the target genes of miRNA [33]. In order to explore the regulatory function of differentially expressed miRNAs, the potential target genes of miRNAs were predicted by using the mRNA transcriptome database [25] of ROC22 root under low N stress.
In the KEGG analysis of predicted target genes of differentially expressed miRNA in Soybean under low N stress, it was found that these genes were involved in protein degradation, N metabolism, carbon metabolism, amino acid metabolism, hormone signaling pathway, cell transport and other metabolic pathways and regulatory pathways [29]. The predicted target genes of differentially expressed miRNAs in the root of Arabidopsis under different nitrate treatments were mainly concentrated in carbon (C) and N metabolism and amino acid metabolism [13], which was similar to what we found (Figure 1). Here, under low N stress, the expression levels of key genes of N metabolism and C metabolism in sugarcane were significantly changed. In the process of the absorption of nitrate and ammonium, nitrate assimilation and biological synthesis of N-containing macromolecules required a lot of energy provided by C metabolism [29]. In addition, under low N stress, the predicted target genes of differentially expressed miRNAs were more concentrated in amino acid metabolism.

miRNA Was Differentially Expressed in Sugarcane under Low N Stress
Wang et al. [8] found that miR156 and miR396 responded positively to environmental stress in the new leaves of poplar cuttings under N deficiency. Under N starvation stress, the expression of miR156 was up-regulated in Arabidopsis seedlings [34]. The expression of miR156 was up-regulated in the roots of Chrysanthemum under low N stress [12]. Under low N stress (0.02 mM N), the expression of miR396 was up-regulated in poplar seedlings [20]. In this research, sugarcane miR156 was up-regulated and miR396 was downregulated under low N stress (0.06 mM N) for 3 h. The expression trend of the sugarcane miR396 was not consistent with that reported in poplar [20], which might be due to various action patterns of miRNA in different crops.
The expression of miR408 was down-regulated in Arabidopsis seedlings under N starvation stress [34]. Under short-term (0.5 h, 2 h, 6 h and 12 h) low N stress (0.75 mM N), the expression of miR408 was down-regulated in the roots of soybean varieties with low N adaptability and up-regulated under long-term low N stress. However, miR167 and miR168 did not respond to either short-term or long-term stress [29]. The expression of miR159, miR167 and miR168 was down-regulated in Populus seedlings under low N stress (0.02 mM N) [20]. In this research, the different expressions of miR159, miR167, miR168 and miR408 compared with the control were identified in the roots of ROC22 under low N stress, while without significance.

Diverse Expression Trends of miRNA and Predicted Target Genes in Sugarcane under Low N Stress
miRNAs regulate crop growth and development by acting on target genes and play an important role in the molecular regulatory network under low N stress [9]. For example, the expression levels of key miRNAs change significantly, which can inhibit or induce the expression of target genes, thus affecting the growth and development of crops [9,11]. In our sequencing database, the identified predicted target genes of miRNA had different biological functions, which could be divided into the following three categories.
The first was transcription factors. According to the literature reports, the target genes of miRl56 and miR396 were SPL [35,36] and GRF [8], respectively. Overexpressing miR156 significantly enhanced root regeneration and N fixation by regulating SPL in Medicago sativa [24]. Here, the target gene SPL of differentially expressed miR156 was down-regulated by 0.61-fold in the roots of ROC22 under low N stress for 3 h. Wang et al. [8] found that miR396 responded positively to N deficiency by targeting GRF in poplar. In this study, miR396 was significantly down-regulated (log 2 FC = −1.35, pval = 0.02) and its target gene GRF was up-regulated by 2.48-fold in the roots of ROC22 under low N stress for 3 h. The target gene of miR159 is MYB [37,38]. MYB was the most differentially expressed transcription factor in the roots of ROC22 under low N stress [25]. In the regulatory networks, transcription factors belong to upstream genes, which can combine with cis acting elements of downstream genes to regulate their expression [39]. miRNA targeted transcription factors might be involved in the regulation of some downstream genes in response to low N stress in sugarcane.
The second type was genes related to the hormone signal transduction pathway. miRNAs play an important regulatory role in the plant hormone signaling pathway and are induced by plant endogenous hormones. The target gene of miRl67 was ARF, a response factor of auxin in the plant hormone signal transduction pathway [40]. It played an important regulatory role in the growth and development of the leaves and roots and flower organ development of plants [41]. ARF also responded positively to NO 3 − signal, adapted to a low N environment, and regulated the structure and growth of lateral roots [42]. In this study, miR167 was down-regulated by 0.54-fold and ARF was up-regulated by 1.29-fold under low N stress.
The third was enzyme protein genes. The target genes of miR408 were PC [38] and laccase [43]. The PC gene was up-regulated by 5.39-fold in the roots of ROC22 under low N stress for 3 h, which positively responded to low N stress. The target gene of miR168 was AGO [38]. AGO was up-regulated by 2.1-fold in the roots of ROC22 under low N stress for 3 h. These proteins belonged to enzyme coding genes and might play an important role in response to low N stress in sugarcane.
In addition, some of the target genes predicted by novel miRNA also had clear biological functions, while some had unclear biological functions. We speculated that the up-regulated expression of miRNA might lead to the degradation of target genes under low N stress and the down-regulated expression of miRNA might promote the overexpression of target genes, thus changing some metabolic pathways and signal transduction pathways and affecting the growth and development of sugarcane. The metabolic and signaling pathways regulated by miRNA and its predicted target genes would provide a new research direction for the molecular regulation mechanism of sugarcane under low N stress.

The miR156 from Sugarcane Roots Could Enhance the N Assimilation Ability of Arabidopsis under Low N Application
Exploring the targeting relationship between miRNA and its target genes could further understand the molecular regulatory mechanism of miRNA [8,20]. miR156 had been extensively studied in response to abiotic stress [8,12,[19][20][21][22], but the function of miR156 in N assimilation has not been studied, including its role in N assimilation in Arabidopsis.
The relationship between miR156 and its target gene SPL was studied in Arabidopsis [44], rice [21], and wheat [45]. Previous studies had found that OsmiR156 might change the morphogenesis and grain size of rice by down-regulating the target gene OsSPL, thereby increasing the rice yield [46,47]. In Arabidopsis seedlings, miR156 participated in the leaf development [44] and flowering process by regulating the target gene SPL [48]. In switchgrass, miR156 regulated the apical dominance and flowering status by inhibiting the SPL gene [49]. In addition, miR156 responded positively to salt stress by down regulating SPL family genes and key genes responding to salt stress [50]. These results indicated that miR156 regulated plant growth and development by targeting SPL. In this research, compared with wild-type Arabidopsis, the SPL gene of transgenic Arabidopsis overexpressing miR156 was significantly down-regulated under low N treatment, which verified the target relationship between miR156 and SPL. Under low N treatment, the expression of P5CS1 and P5CS2, a candidate target gene of miR156, was down-regulated, showing a targeted relationship with miR156. We speculated that sugarcane miR156 might respond positively to low N treatment by negatively regulating SPL, P5CS1, and P5CS2. In future research, the targeting relationship between sugarcane miR156 and candidate target genes P5CS1 and P5CS2 needed to be further verified by other experimental means.
Transgenic plants overexpressing miR156 were often used as experimental materials to investigate the role of miR156 in N stress [18,51]. Gou et al. [51] found that miR156 promoted anthocyanin accumulation by regulating SPL9 in transgenic Arabidopsis overexpressing miR156 under N deficiency stress. In order to adapt to the low N stress environment, transgenic rice overexpressing miR156 under low N stress down-regulated the key genes CYP724 and CYP90 in the brassinosteroid biosynthesis pathway [18]. In this research, the root length and root surface area of transgenic Arabidopsis overexpressing miR156 were significantly higher than those of wild type under low N treatment. Nitrate which was absorbed by the plant roots, which was converted into ammonium salt by NR and NIR enzyme, and then converted into glutamine (Gln) and glutamate (Glu) mainly through two reaction pathways, one of which was GDH, and the other two were GS and GOGAT as the catalytic enzymes [52,53]. Gln and Glu, which served as precursors of N compound biosynthesis, were further converted into organic N molecules [54]. Therefore, NR, NIR, GS, GDH, and GOGAT were the key enzymes in the process of primary N assimilation. Furthermore, the expression of NRT1.1, NR1, NIR1, and GS of transgenic Arabidopsis overexpressing miR156 were significantly higher than those of wild type under low N treatment. According to the root phenotype, physiological indexes, and the expression of key N assimilation genes, sugarcane miR156 could enhance the N assimilation ability of Arabidopsis under low N application.

Cultivation of Plant Materials
Tissue cultured plantlets of ROC22 (low N tolerant variety) [55]

cDNA Library Construction and Small RNA (sRNA) Sequencing
The roots of four-leaf stage ROC22 plantlets of uniform size at 0 h and 3 h under low N treatment were collected. We set up 0.6 mM N concentration (N source: 0.3 mM NH 4 NO 3 ) for 3 h as low N treatment and for 0 h as control (N source: 3.0 mM NH 4 NO 3 ) [25]. The treatment and control were designated as RRCK and RRT. Three randomly selected plantlets were pooled into biological duplicates, and each experiment was performed using three biological replicates. The total RNA of samples was extracted with TRIzol reagent (Invitrogen, Shanghai, China), and RNA quality was examined using a NanoDrop (Thermo Fisher Scientific, Inc., Waltham, MA, USA) and an Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA). The RNA was divided into two parts, one was sent to Novogene Corporation (Tianjin, China) for miRNAsequencing, and the other for qRT-PCR validation analysis.
The analysis method of miRNAsequencing data was in reference to Yang et al. [27]. DESeq2 was used to analyze the difference of miRNA expression level. According to the p value of miRNA (p < 0.05), the significant level of miRNA was screened. The software of psRNATarget was used to predict the target genes of miRNA. Then, GOseq [27] was used to analyze the candidate target genes of differentially expressed miRNAs, and the functional differences of the genes between different samples were obtained. Finally, a hypergeometric test was used to analyze the KEGG enrichment of candidate target genes with differential expression of miRNAs, and to explore the biochemical metabolism and signal transduction pathway of candidate target genes.

Detection of miRNA and Target Gene Expression under Low N Stress
Ten miRNAs (novel-57, miR156, novel-158, miR168a, miR396, miR408a, novel-215, miR393, miR408e, miR167) and 13 candidate target genes (AUX/IAA, P5CS, GS, SPX, acnA, NiFU, ACO1, TIR1, SPL1, GRF, AGO, PC, ARF) were randomly selected, and their gene expression levels were validated by qRT-PCR. The primer sequences were listed in Tables S3 and S4. The genes of 18S rRNA [56] and glyceraldehyde-3-phosphate dehydrogenase, GAPDH [57] were used as the internal reference genes. ROC22 plantlets at a four-leaf stage were used for low N treatment, and the samples were collected for the follow-on experiments at 0 h, 3 h, 6 h, 1 d, 4 d, 8 d and 16 d after treatment. We set up 0.6 mM N concentration (N source: 0.3 mM NH 4 NO 3 ) as low N treatment, and the sample collected at 0 h as control.

Acquisition of the miR156-Overexpressed Arabidopsis
The precursor of sugarcane miR156 was constructed in the overexpression vector pCAMBIA1301 and transformed into Agrobacterium tumefacines strain EHA105 [58]. Transformation of pCAMBIA miR156 into Arabidopsis was performed according to a previously described method [27,59]. T3 transgenic Arabidopsis homozygotes were obtained and used for the subsequent experiments.

Low N Treatment of the miR156-Overexpressed Arabidopsis
MS medium containing 1/2 N (low N treatment, LN) was used, pH was adjusted to 5.8, and agar powder was added to 4.5 g/L. After disinfection, wild-type Arabidopsis seeds and homozygous miR156-overexpressed Arabidopsis seeds were sowed on the prepared plate. Each plate was a biological repeat, and three biological repeats were set for each transgenic line. The plates were wrapped with tin foil, treated for 3 days in the dark at 4 • C, and transferred to the incubator. After 20 days of cultivation (22 • C, 75% relative humidity, 16 h photoperiod), the whole plant samples were collected. Some fresh plant samples were used to determine the activities of key enzymes in N metabolism, and some plant samples were frozen in liquid N and stored in a −80 • C refrigerator for qRT-PCR detection of miRNAs and genes. Plant roots under low N treatment were analyzed by root analyzer (WinRHIZO system, Regent Instruments, Québec, Canada).

Physiological Changes and qRT-PCR Detection of the miR156-Overexpressed Arabidopsis under Low N Treatment
After 20 days of low N treatment, the activities of key N metabolism enzymes NR (nitrate reductase), NIR (nitrite reductase), GS (glutamine synthase), GOGAT (glutamate synthase) and GDH (glutamate dehydrogenase) in the miR156-overexpressed Arabidopsis were determined. At the molecular level, the gene expression levels of the miR156 and key genes of N metabolism in the miR156-overexpressed Arabidopsis under low N treatment for 20 days were detected by qRT-PCR. The stem ring primer (5 -3 ) of the miR156 was CT-CAACTGGTGTCGTGGAGTCGGCAATTCAGTTGAGGTGCTCAC. The qRT-PCR primers of the miR156 and N metabolism key genes (NRT1.1, AMT1.1, NR2, NR1, GS1, GDH1, FD-GOGAT, NADH-GOGAT, NIR1) were shown in Table S5. The internal reference gene of the miR156 and N metabolism key genes were 18S rRNA [56] and UBQ10 [60], respectively.

Conclusions
In the present study, the root miRNA database of the N-efficient sugarcane variety ROC22 under low N stress was constructed. According to GO and KEGG analysis, the predicted target genes of the differentially expressed miRNAs were mainly involved in the molecular functions of amylase synthase, mannosyltransferase, asparagine hydrolase, and in the biological processes of abiotic stimulation and hormone response. These target genes were mainly enriched in N metabolism, amino acid metabolism, carbohydrate metabolism, photosynthesis and hormone signal transduction pathways. Interestingly, there was a negative correlation between miR168 and SPX, along with miR396 and acnA under low N stress for 0-24 h. Moreover, the expression of miR156 in the roots of ROC22 was significantly up-regulated under low N treatment and was negatively correlated with SPL and P5CS at stress for 3 h, 8 d, and 16 d. Furthermore, the length and surface area of roots in the miR156-overexpressed Arabidopsis were significantly higher than those of wild type under low N treatment, and the expression of key genes NRT1.1, NR1, NIR1, GS and the enzyme activity of NR and GS of N assimilation was all significantly higher than those of the wild type. It can thus be concluded that sugarcane miR156 can enhance the nitrogen assimilation ability of Arabidopsis plants under a low N environment. In the future, miR156 can be overexpressed or silenced in sugarcane, and the mechanism of miR156 regulating NUE in sugarcane can be investigated to provide a theoretical basis for the development of sugarcane varieties with high NUE.