Identification of microRNA158 from Anthurium andraeanum and Its Function in Cold Stress Tolerance

Anthurium andraeanum is a tropical flower with high ornamental and economic value. Cold stress is one of the major abiotic stresses affecting the quality and value of A. andraeanum; thus, improving the cold tolerance of this species is an important breeding objective. MicroRNAs (miRNAs) have a critical role in plant abiotic stress responses, but their specific molecular regulatory mechanisms are largely unknown, including those related to the cold stress response in A. andraeanum. Here, we identified and cloned the precursor of miR158 from A. andraeanum (Aa-miR158). Both Aa-miR158 and its target gene (c48247) had higher expression levels in strong leaves than in other tissues or organs. Further study revealed that the transcript level of Aa-miR158 was increased by cold stress. Heterologous overexpression of Aa-miR158 improved cold stress tolerance in Arabidopsis, which was associated with decreases in the malondialdehyde (MDA) concentration and relative electrical conductivity (REC) as well as increases in peroxidase (POD) and catalase (CAT) activity. Moreover, overexpressing Aa-miR158 significantly increased the expression of endogenous genes related to cold stress tolerance and reactive oxygen species (ROS) levels in transgenic Arabidopsis under cold stress. Overall, our results demonstrate that Aa-miR158 is significantly involved in the cold stress response and provide a new strategy for cold tolerance breeding of A. andraeanum.


Introduction
Anthurium andraeanum Lind., an important ornamental flower, originated in the tropical rainforest of South American and has been widely cultivated throughout the world. It needs to be cultivated in a high-temperature, high-humidity environment throughout the year. The optimal growth temperature of A. andraeanum is 20-28 • C, and cold damage symptoms occur when the temperature is below 15 • C [1]. During greenhouse cultivation of A. andraeanum, especially in winter, the heating costs are exorbitant; thus, improving the cold tolerance of A. andraeanum would help to further extend its planting range and increase its economic benefits.
MicroRNAs (miRNAs) are a class of endogenous non-coding small RNAs that negatively modulate target gene expression at the post-transcriptional level through mRNA cleavage or translational repression [2][3][4]. In plants, miRNAs regulate a broad group of transcription factors (TFs) and are involved in multiple biological processes, including development [5][6][7], biotic stresses [8] and abiotic stresses [9][10][11]. Specifically, miRNAs play a role in many horticultural traits (e.g., flowering, fruit quality, abiotic and biotic stress), and provide valuable information for future horticultural crop research and breeding [12].
Cold stress is among the major abiotic stresses and has a detrimental impact on almost every aspect of the physiology and biochemistry of plants [13]. Quite recently, many miRNAs such as miR319, miR394, miR396, miR397, miR402 and miR408 have been found to play various roles in mediating plant responses to cold stress [13,14]. The cold

Expression Analyses of Aa-miR158 and c48247
To explore the expression pattern of Aa-miR158, the transcript profile of mature Aa-miR158 was detected in the root, stem, spire, strong leaves, old leaves and spathe of A. andraeanum through qRT-PCR. Relative to the expression in the root, the expression of Aa-miR158 was highest in strong leaves (152.2-fold), followed by the stem (56.5-fold) and old leaves (3.4-fold) (Figure 2A).

Expression Analyses of Aa-miR158 and c48247
To explore the expression pattern of Aa-miR158, the transcript profile of mature Aa-miR158 was detected in the root, stem, spire, strong leaves, old leaves and spathe of A. andraeanum through qRT-PCR. Relative to the expression in the root, the expression of Aa-miR158 was highest in strong leaves (152.2-fold), followed by the stem (56.5-fold) and old leaves (3.4-fold) (Figure 2A). Meanwhile, the transcript profile of target gene c48247 was also detected in A. andraeanum using qRT-PCR. The expression level of c48247 was highest in the strong leaves followed by the spire and old leaves ( Figure 2B). The transcript level of c48247 in the spire was nearly 6-fold that in the root, whereas Aa-miR158 had lower expression in the spire than in the root ( Figure 2B). c48247 maintained a low transcript level in the stem, which Meanwhile, the transcript profile of target gene c48247 was also detected in A. andraeanum using qRT-PCR. The expression level of c48247 was highest in the strong leaves followed by the spire and old leaves ( Figure 2B). The transcript level of c48247 in the spire was nearly 6-fold that in the root, whereas Aa-miR158 had lower expression in the spire than in the root ( Figure 2B). c48247 maintained a low transcript level in the stem, which was similar to that in the root. Thus, Aa-miR158 and its target gene c48247 were both highly expressed in the strong leaves and old leaves of A. andraeanum, while their expression patterns were different in the stem and spire. These results suggested that Aa-miR158 may have diverse roles in various biological processes via upregulated or downregulated expression of c48247.

Aa-miR158 and c48247 Response to Cold Stress in A. andraeanum
To identify whether Aa-miR158 responds to cold stress, Aa-miR158 expression was measured in leaves of A. andraeanum seedlings exposed to cold stress treatment (4 • C). During 24 h of cold stress, the expression level of Aa-miR158 gradually increased and was about 5-fold higher at 24 h than at 0 h ( Figure 3A). This result indicated that Aa-miR158 was involved in the cold stress response at the transcriptional level in A. andraeanum. On the other hand, the expression of its target gene c48247 was downregulated at 12 and 24 h of 4 • C treatment ( Figure 3B).  In addition, the expression of two cold-related genes, ascorbate peroxidase 1 (AaAPX1) and catalase 3 (AaCAT3), was also detected in leaves of A. andraeanum seedlings exposed to cold stress treatment (4 °C). The results showed that these two genes' transcripts were significantly upregulated under cold stress treatment after 24 h, which is similar to the expression pattern of Aa-miR158 under the same cold stress conditions ( Figure 3C,D).

Aa-miR158 Represses the Expression of c48247 in A. andraeanum
To further investigate whether Aa-miR158 inhibited the expression of the c48247 gene in A. andraeanum, a pre-Aa-miR158 overexpression construct was infiltrated into A. andraeanum leaves for a transient expression assay. The transcript level of Aa-miR158 in p35S::pre-Aa-miR158 leaves was 3-fold than the level in control leaves at 28/20 °C ( Figure 4A). As expected, the expression of c48247 was decreased in p35S::pre-Aa-miR158 leaves compared with control leaves ( Figure 4B). Interestingly, Aa-miR158 in p35S::pre-Aa-miR158 leaves dis- In addition, the expression of two cold-related genes, ascorbate peroxidase 1 (AaAPX1) and catalase 3 (AaCAT3), was also detected in leaves of A. andraeanum seedlings exposed to cold stress treatment (4 • C). The results showed that these two genes' transcripts were significantly upregulated under cold stress treatment after 24 h, which is similar to the expression pattern of Aa-miR158 under the same cold stress conditions ( Figure 3C,D).

Aa-miR158 Represses the Expression of c48247 in A. andraeanum
To further investigate whether Aa-miR158 inhibited the expression of the c48247 gene in A. andraeanum, a pre-Aa-miR158 overexpression construct was infiltrated into A. andraeanum leaves for a transient expression assay. The transcript level of Aa-miR158 in p35S::pre-Aa-miR158 leaves was 3-fold than the level in control leaves at 28/20 • C ( Figure 4A). As expected, the expression of c48247 was decreased in p35S::pre-Aa-miR158 leaves compared with control leaves ( Figure 4B). Interestingly, Aa-miR158 in p35S::pre-Aa-miR158 leaves displayed a significant increase in expression of 7.6-fold at 4 • C, while c48247 showed a significant reduction in expression of nearly 90% ( Figure 4C,D). On the other hand, a co-expression assay showed that the fluorescence signal became weaker in co-expression with pre-Aa-miR158 and a GFP reporter bearing Aa-miR158 target site compared with that in a GFP reporter bearing Aa-miR158 target site alone (Supplementary Materials, Figure S1).

Heterologous Expression of Aa-miR158 in Arabidopsis Improves Tolerance to Cold Stress
To investigate the biological role of Aa-miR158 in detail under different abiotic stresses, the nucleotide sequence of pre-Aa-miR158 was amplified from genomic DNA of A. andraeanum. The p35S::pre-Aa-miR158 vector was constructed and transformed into Arabidopsis through Agrobacterium-mediated transformation. A total of 11 independent T1 transgenic lines of p35S::pre-Aa-miR158 were identified through screening on hygromycin and then confirmed via PCR. Among the 11 transgenic lines identified by hygromycin screening, the electrophoresis band corresponding to Aa-miR158 was detected in all 11 lines but not in WT plants (Supplementary Materials, Figure S2A). Aa-miR158 was found to be overexpressed at high levels in two transgenic lines (designated as p35S::pre-Aa-miR158-1 and -5), which were used for further analysis (Supplementary Materials, Figure  S2B).
Considering that miRNAs are involved in various stress responses [28,29], we investigated the responses of two p35S::pre-Aa-miR158 overexpressing transgenic Arabidopsis lines to salinity, drought and cold stress. Compared with WT, the phenotypes of p35S::pre-Aa-miR158 transgenic plants displayed no marked changes under salt stress (Supplementary Materials, Figure S3A). As expected, the MDA levels and REC exhibited no significant differences between WT and p35S::pre-Aa-miR158 transgenic lines under salt stress (Supplementary Materials, Figure S3B). We next tested the function of Aa-miR158 under drought stress in transgenic and WT plants. The soil moisture content decreased by nearly

Heterologous Expression of Aa-miR158 in Arabidopsis Improves Tolerance to Cold Stress
To investigate the biological role of Aa-miR158 in detail under different abiotic stresses, the nucleotide sequence of pre-Aa-miR158 was amplified from genomic DNA of A. andraeanum. The p35S::pre-Aa-miR158 vector was constructed and transformed into Arabidopsis through Agrobacterium-mediated transformation. A total of 11 independent T1 transgenic lines of p35S::pre-Aa-miR158 were identified through screening on hygromycin and then confirmed via PCR. Among the 11 transgenic lines identified by hygromycin screening, the electrophoresis band corresponding to Aa-miR158 was detected in all 11 lines but not in WT plants (Supplementary Materials, Figure S2A). Aa-miR158 was found to be overexpressed at high levels in two transgenic lines (designated as p35S::pre-Aa-miR158-1 and -5), which were used for further analysis (Supplementary Materials, Figure S2B).
Considering that miRNAs are involved in various stress responses [28,29], we investigated the responses of two p35S::pre-Aa-miR158 overexpressing transgenic Arabidopsis lines to salinity, drought and cold stress. Compared with WT, the phenotypes of p35S::pre-Aa-miR158 transgenic plants displayed no marked changes under salt stress (Supplementary Materials, Figure S3A). As expected, the MDA levels and REC exhibited no significant  Figure S3B). We next tested the function of Aa-miR158 under drought stress in transgenic and WT plants. The soil moisture content decreased by nearly 90% in the drought treatment compared with that in the control (Supplementary Materials, Figure S4B). There was also no clear effect of Aa-miR158 overexpression on the phenotype of p35S::pre-Aa-miR158 transgenic plants after withholding water (Supplementary Materials, Figure S4A). Under drought stress, MDA levels were significantly lower in the two p35S::pre-Aa-miR158 transgenic lines than in WT, but the soluble protein content showed no significant difference, and the POD and CAT activities were even lower in one or both of the p35S::pre-Aa-miR158 transgenic lines than in WT plants (Supplementary Materials, Figure S4C).
To assess the cold stress tolerance of transgenic Arabidopsis, we investigated the phenotypic characteristics of p35S::pre-Aa-miR158 and WT plants under cold stress (4 • C for 6 h followed by 0 • C for 1 h) and normal growth conditions (22/20 • C, 12/12 h). Under normal conditions, no marked phenotypic differences were found between transgenic lines and WT plants ( Figure 5A). Under cold stress, the leaves of WT Arabidopsis began to turn yellow and showed severe wilting, which eventually led to their death. However, p35S::pre-Aa-miR158 transgenic plants exhibited enhanced cold tolerance and a low degree of chilling injury compared with WT plants ( Figure 5A). Following the exposure to cold stress treatment, the survival rate of two p35S::pre-Aa-miR158 transgenic lines was >80%, which was significantly higher than that of WT (<60%) ( Figure 5B).  Figure S4C).
To assess the cold stress tolerance of transgenic Arabidopsis, we investigated the phenotypic characteristics of p35S::pre-Aa-miR158 and WT plants under cold stress (4 °C for 6 h followed by 0 °C for 1 h) and normal growth conditions (22/20 °C, 12/12 h). Under normal conditions, no marked phenotypic differences were found between transgenic lines and WT plants ( Figure 5A). Under cold stress, the leaves of WT Arabidopsis began to turn yellow and showed severe wilting, which eventually led to their death. However, p35S::pre-Aa-miR158 transgenic plants exhibited enhanced cold tolerance and a low degree of chilling injury compared with WT plants ( Figure 5A). Following the exposure to cold stress treatment, the survival rate of two p35S::pre-Aa-miR158 transgenic lines was >80%, which was significantly higher than that of WT (<60%) ( Figure 5B). The levels of MDA were significantly lower under cold stress in p35S::pre-Aa-miR158 plants than in WT plants ( Figure 6). REC was significantly lower under cold stress in p35S::pre-Aa-miR158 plants than in WT, whereas POD and CAT activities were significantly higher (Figure 6). These results suggested that p35S::pre-Aa-miR158 transgenic lines are more resistant to cold stress than WT plants. The levels of MDA were significantly lower under cold stress in p35S::pre-Aa-miR158 plants than in WT plants ( Figure 6). REC was significantly lower under cold stress in p35S::pre-Aa-miR158 plants than in WT, whereas POD and CAT activities were significantly higher ( Figure 6). These results suggested that p35S::pre-Aa-miR158 transgenic lines are more resistant to cold stress than WT plants.

Aa-miR158 Improves Cold Stress Tolerance by Affecting Expression of Genes Related to Cold Stress and ROS
In Arabidopsis, both cold-regulated 15a (COR15a) and C-REPEAT/DRE BINDING FAC-TOR 1 (CBF1) are involved in chilling tolerance [30,31]. In this study, the expression of COR15a and CBF1 was not significantly different between WT and p35S::pre-Aa-miR158 transgenic Arabidopsis under unstressed (control) conditions ( Figure 7A,B). Nevertheless, increases in transcript level in transgenic p35S::pre-Aa-miR158 compared to WT plants were observed for both COR15a and CBF1 under conditions of cold stress ( Figure 7A,B). In p35S::pre-Aa-miR158-1, the expression level of COR15a under cold stress was almost 5.0fold that in WT, and in p35S::pre-Aa-miR158-5, the expression levels of COR15a and CBF1 under cold stress were 7.5-fold and 4.9-fold those in WT, respectively ( Figure 7A,B). In addition, the expression levels of CAT3 and APX2 in two p35S::pre-Aa-miR158 transgenic lines were not significantly different from those in WT under unstressed (control) conditions, while under cold stress there was a significant increase in the transcript levels of both genes in p35S::pre-Aa-miR158-1 relative to WT ( Figure 7C,D). These results illustrated that Aa-miR158 may improve cold stress tolerance by activating genes related to cold stress and ROS at the transcriptional level, emphasizing the important function of Aa-miR158 in plant survival.

Aa-miR158 Improves Cold Stress Tolerance by Affecting Expression of Genes Related to Cold Stress and ROS
In Arabidopsis, both cold-regulated 15a (COR15a) and C-REPEAT/DRE BINDING FAC-TOR 1 (CBF1) are involved in chilling tolerance [30,31]. In this study, the expression of COR15a and CBF1 was not significantly different between WT and p35S::pre-Aa-miR158 transgenic Arabidopsis under unstressed (control) conditions ( Figure 7A,B). Nevertheless, increases in transcript level in transgenic p35S::pre-Aa-miR158 compared to WT plants were observed for both COR15a and CBF1 under conditions of cold stress ( Figure 7A,B). In p35S::pre-Aa-miR158-1, the expression level of COR15a under cold stress was almost 5.0-fold that in WT, and in p35S::pre-Aa-miR158-5, the expression levels of COR15a and CBF1 under cold stress were 7.5-fold and 4.9-fold those in WT, respectively ( Figure 7A,B). In addition, the expression levels of CAT3 and APX2 in two p35S::pre-Aa-miR158 transgenic lines were not significantly different from those in WT under unstressed (control) conditions, while under cold stress there was a significant increase in the transcript levels of both genes in p35S::pre-Aa-miR158-1 relative to WT ( Figure 7C,D). These results illustrated that Aa-miR158 may improve cold stress tolerance by activating genes related to cold stress and ROS at the transcriptional level, emphasizing the important function of Aa-miR158 in plant survival.

Discussion
miRNAs are involved in many kinds of abiotic stresses in various species; however, there are very few studies on the function of miRNAs in A. andraeanum. In this study we isolated Aa-miR158, a homolog of the miR158 family, which was previously reported to be a specific miRNA in Brassicaceae. The mature sequence of Aa-miR158 was poorly aligned (<70%) with miR158 sequences in other species and its highest expression was found in strong leaves, suggesting that the function of Aa-miR158 may have diverged during evolution (Figure 1 and 2). Actually, more than one member of the miR158 family exist in other plant species [32], while how many members are included in the miR158 family in A. andraeanum and their expression patterns in A. andraeanum remain unclear and need to be investigated in the future.
Low-temperature stress damages plant tissue and thus affects the growth and development of important ornamental flowers [33]. Our expression analysis showed that the transcript level of Aa-miR158 and cold-related genes was increased by low temperature (Figure 3), suggesting Aa-miR158 may be involved in the cold stress signaling pathway and a possible correlation between elevated levels of Aa-miR158 and enhanced cold tolerance in A. andraeanum.
Aa-miR158 was predicted to be an unconservative miRNA and a terpene synthase (TPS) gene, c48247, was predicted to be the target of Aa-miR158 [27]. The fluorescence assays showed that Aa-miR158 can silence the GFP reporter bearing Aa-miR158 target site (Supplementary Materials, Figure S1), while the 5′-RLM-RACE results showed that the target site was not cleaved by Aa-miR158, providing evidence that Aa-miR158 might regulate the function of c48247 by repressing its translation in A. andraeanum. However, each miRNA corresponds to more than one target genes in general; therefore, whether Aa-

Discussion
miRNAs are involved in many kinds of abiotic stresses in various species; however, there are very few studies on the function of miRNAs in A. andraeanum. In this study we isolated Aa-miR158, a homolog of the miR158 family, which was previously reported to be a specific miRNA in Brassicaceae. The mature sequence of Aa-miR158 was poorly aligned (<70%) with miR158 sequences in other species and its highest expression was found in strong leaves, suggesting that the function of Aa-miR158 may have diverged during evolution (Figures 1 and 2). Actually, more than one member of the miR158 family exist in other plant species [32], while how many members are included in the miR158 family in A. andraeanum and their expression patterns in A. andraeanum remain unclear and need to be investigated in the future.
Low-temperature stress damages plant tissue and thus affects the growth and development of important ornamental flowers [33]. Our expression analysis showed that the transcript level of Aa-miR158 and cold-related genes was increased by low temperature (Figure 3), suggesting Aa-miR158 may be involved in the cold stress signaling pathway and a possible correlation between elevated levels of Aa-miR158 and enhanced cold tolerance in A. andraeanum.
Aa-miR158 was predicted to be an unconservative miRNA and a terpene synthase (TPS) gene, c48247, was predicted to be the target of Aa-miR158 [27]. The fluorescence assays showed that Aa-miR158 can silence the GFP reporter bearing Aa-miR158 target site (Supplementary Materials, Figure S1), while the 5 -RLM-RACE results showed that the target site was not cleaved by Aa-miR158, providing evidence that Aa-miR158 might regulate the function of c48247 by repressing its translation in A. andraeanum. However, each miRNA corresponds to more than one target genes in general; therefore, whether Aa-miR158 has other potential target genes, especially related to cold stress tolerance, also needs to be further explored. The CBF-dependent pathway has been proved to play a leading role in plant tolerance to low-temperature stress [34]. Overexpression of pre-Aa-miR158 in Arabidopsis enhanced cold stress tolerance compared with that of WT plants ( Figure 5). In addition, the expression levels of endogenous cold stress-related genes (COR15a and CBF1) were significantly higher in p35S::pre-Aa-miR158 transgenic plants than in WT under cold stress, except for CBF1 in p35S::pre-Aa-miR158-1 (Figure 7). Our results suggested that Aa-miR158 enhanced lowtemperature tolerance in Arabidopsis by activating the expression of genes involved in the CBF-dependent pathway.
The production and removal of ROS maintains a balance in plants under normal physiological conditions [35,36]. However, cold stress leads to the excess accumulation of ROS, and this damages intracellular components and membrane lipids [37]. Our results showed that MDA levels and REC were significantly lower in p35S::pre-Aa-miR158 transgenic plants than in WT under cold stress ( Figure 6). They also showed that the POD and CAT activity were significantly higher in p35S::pre-Aa-miR158 transgenic plants than in WT under cold stress ( Figure 6). In addition, the transcript levels of CAT3 and APX2 were significantly higher in p35S::pre-Aa-miR158-1 than in WT under cold stress (Figure 7). These results suggested that Aa-miR158 improves low-temperature tolerance in plants, possibly due in part to the reduction of ROS damage.
Interestingly, although p35S::pre-Aa-miR158 transgenic plants did not appear to have significantly improved the tolerance to other abiotic stresses compared with WT, the MDA content decreased significantly compared with WT under drought stress (Supplementary Materials, Figures S3 and S4). The above results imply that the increased tolerance caused by the overexpression of pre-Aa-miR158 may be highly specific to cold stress in A. andraeanum, but its possible effects on other stress response processes need to be considered in future investigations.
Although Arabidopsis transgenics that overexpressed pre-Aa-miR158 showed more tolerance to cold treatment in this study, many of the details of the regulatory mechanisms of Aa-miR158 remain elusive. A recent review showed that miRNAs respond to lowtemperature stress by directly or indirectly responding to external stimuli by mediating target genes, or responding to multiple stresses, and thus starting the hydrolysis or redox processes [38]. In various species, the regulatory mechanisms of the miRNAs involved in low-temperature stress were different, for example, phytohormone signal transduction and the ROS scavenging pathway [20,39,40]. Therefore, the specific regulation mechanism of Aa-miR158 involved in cold tolerance in heterologous Arabidopsis requires further explication by experimental evidence such as degradome sequencing.
With the development of omics technology, the miRNAs sequence information can be obtained more conveniently by high-throughput sequencing. However, the stable genetic transformation of A. andraeanum remains difficult, is very time-consuming and has quite a low transformation efficiency. It is limited for investigating the potential function of miRNAs in A. andraeanum. In future research, the functional characterization of Aa-miR158 will be further investigated in A. andraeanum by establishing a rapid and efficient genetic transformation system, which will lay the foundation for utilizing miRNAs to improve stress resistance in other ornamental plants.

Plant Materials and Growth Conditions
In vitro plantlets of A. andraeanum cultivar 'Sonate' were grown in a tissue culture room under 12/12 h, 25/25 • C, day/night conditions at Nanjing Agricultural University. The potted plants were grown in the Baima Teaching and Research Base of Nanjing at 25/20 • C (day/night), 70-80% relative humidity, and natural lighting under standard greenhouse conditions. Leaves from plant tissue culture and different organs (root, stem, spire, strong leaves, old leaves and spathe) from potted plants were collected from at least three independent plants for further analysis. After collection, the samples were frozen immediately in liquid nitrogen and stored at −80 • C until use.
The A. thaliana Columbia (Col-0) ecotype was used for transgenic analysis. The seeds were surface-sterilized in 75% alcohol with 0.03% Triton X-100 for 2 min, rinsed three times in 70% alcohol and then washed three times in distilled water. Disinfected seeds were sown on Murashige and Skoog (MS) medium with 1% sucrose and placed in a 4 • C refrigerator for 3 days. After refrigeration, plates with seeds were transferred to a light incubator under a 12 h photoperiod with day/night conditions at 22/20 • C. The seedlings were transplanted into soil at the two-leaf stage.

Sequence Analysis
The precursor and mature sequences of Aa-miR158 were acquired from our inhouse small RNA sequencing library (Bioproject, PRJNA472856). The corresponding target gene was predicted using the plant small RNA target (psRNA Target) online sever (http://plantgrn.noble.org/psRNATarget/, accessed on 20 September 2016; now available at https://www.zhaolab.org/psRNATarget/, accessed on 25 May 2018) and verified by small RNA-mRNA co-analysis. Genomic DNA was isolated from leaves using a cetyltrimethylammonium ammonium bromide (CTAB) method for the amplification of the pre-Aa-miR158 sequence. The pre-Aa-miR158 fragment was amplified by PCR using KOD-Plus-Neo (TOYOBO, Osaka, Japan). The PCR reaction conditions were as follows: initial denaturation at 94 • C for 3 min; 36 cycles of denaturation at 98 • C for 10 s, annealing at 55 • C for 30 s, and extension at 72 • C for 10 s; and final extension at 72 • C for 7 min. Multiple sequence alignment was performed with DNAMAN software (Lynnon Biosoft, CA, USA). The primers used for pre-Aa-miR158 cloning are listed in the Supplementary Materials, Table S1.

Expression Analysis
Total RNA was extracted using an improved CTAB method [41] and then reversetranscribed with a Mir-X miRNA First-Strand Synthesis Kit (Takara, Kyoto, Japan) using oligo (dT) primers. Aa-miR158-specific forward primers and a universal reverse primer (mRQ3 primer) supplied by the manufacturer were used. Primers for target genes were designed by Primer Premier 5.0 and synthesized by Tsingke Biotechnology Company (Nanjing, China). SYBR ® Premix Ex Taq™ (Takara, Japan) was used for qRT-PCR. The 5S ribosomal RNA and a housekeeping gene, glyceraldehyde-3-phosphate dehydrogenase (Aa-GAPDH), were used as internal references for the quantification of miRNA and target gene expression, respectively [42,43]. qRT-PCR was performed using a QuantStudio™ 3 Real-Time PCR System (ABI, MA, USA) according to the manufacturer's instructions. The relative expression level was calculated using the 2 −∆∆Ct method [44] and the standard deviation was calculated using three biological replicates. The t-test (p < 0.05) was selected for statistical analysis. The primers used for qRT-PCR are listed in the Supplementary Materials, Table S1.
Leaves from wild-type (WT) and p35S::pre-Aa-miR158 transgenic Arabidopsis were frozen in liquid nitrogen and stored at −80 • C for qRT-PCR analysis. Total RNA extraction was performed as described above, and cDNA was synthesized using a PrimeScript RT reagent Kit with gDNA Eraser (Takara, Japan). Arabidopsis tubulin2 (AtTUB2) was used as the qRT-PCR internal control. The primers used for qRT-PCR are listed in the Supplementary Materials, Table S1.

Transient Overexpression of Pre-Aa-miR158 via Agrobacterium Injection
The pGreenII 62-SK vector (853 bp) was used for the transient overexpression of pre-Aa-miR158 in leaves of A. andraeanum. The recombinant overexpression vector p35S::pre-Aa-miR158 was transformed into Agrobacterium strain 'GV3101 (psoup)' and then infiltrated into the leaves. The Agrobacterium-infiltrated plants were incubated in a dark chamber at room temperature for 1 d and then in a natural greenhouse at 28/20 • C (day/night) or a temperature chamber at 4 • C, 12h/12h for 2 d. The samples were collected and frozen immediately in liquid nitrogen for RNA extraction and qRT-PCR analysis. Three biological replicates with three independent plants per replicate were included for the expression analysis of each gene . The primers used for transient overexpression are listed in the  Supplementary Materials, Table S1.

Fluorescence Assay
The green fluorescent protein (GFP)-fused the 3 UTR of target gene was inserted into the plasmid pGREEN through the XbaI and SmaI cleavage sites. The fused protein p35S::c48247-GFP plasmid was transformed into Agrobacterium tumefaciens 'GV3101 (psoup)' and then individually transfected and co-transfected with p35S::pre-Aa-miR158 plasmid into tobacco leaves (OD600 = 1.0), respectively. A fluorescence microscope (Leica DM6, Weztlar, Germany) was used to take images at 460-500 nm wavelengths for GFP. The primers for the fluorescence assay are listed in the Supplementary Materials, Table S1. 4.6. 5 -RLM-RACE Assay 5 -RLM-RACE was performed to determine the cleavage site of target using the FirstChoice RLM-RACE Kit (Ambion, TX, USA). The total RNA that removed genome DNA was ligated with a 5 -adaptor and reversed using random primers to obtain 5 -RACE cDNA. Nested PCR was performed using adaptor-derived and gene-specific primers to obtain the 3 -cleavage products. The primers used for the 5 -RLM-RACE PCR are listed in the Supplementary Materials, Table S1.

Vector Construction and Plant Transformation
The full-length of pre-Aa-miR158 was cloned into the pCAMBIA1302 vector behind the 35S promoter by using the BglII and SpeI cleavage sites. The construct was then introduced into Agrobacterium strain EHA105 and transformed into Arabidopsis WT (Col-0) plants through the floral-dip method [45]. The transgenic plants were screened on MS medium with 35 mg·L −1 hygromycin and confirmed by PCR amplification assay to detect the p35S::pre-Aa-miR158 transgene lines. The acquired hygromycin-resistant plants were used for further analysis. The primers used for producing the overexpression vector and DNA detection are listed in the Supplementary Materials, Table S1.

Abiotic Stress Treatments
In vitro plantlets of A. andraeanum cultivar 'Sonate' were used to examine the effects of cold stress on Aa-miR158 and c48247 gene expression. The plantlets were placed in a temperature chamber at 4 • C for 0, 12 or 24 h. After treatment, 0.1-0.2 g of leaves was harvested and immediately frozen in liquid nitrogen for expression analysis. The cold stress treatment assay was performed three times.
To examine the effects of pre-Aa-miR158 overexpression on stress tolerance, the surfacesterilized seeds of Arabidopsis WT plants and p35S::pre-Aa-miR158 transgenic plants were sown on MS medium (1% sucrose). For cold stress treatment, 3-week-old transgenic and WT plants grown in soil were placed in a light incubator at 4 • C for 6 h followed by 0 • C for 1 h. Phenotypic observation, physiological index determination and gene expression analysis were performed after the cold stress treatment. The survival rate was determined after resuming growth under a 12 h photoperiod and 22/20 • C light/dark temperatures for 24 h. For the salt stress treatment, 3-week-old transgenic and WT plants grown in Hoagland's nutrient solution with 150 mM NaCl were placed in a light incubator (22/20 • C, light/dark and 12 h photoperiod) for 3 d. For drought stress treatment, 3-week-old transgenic and WT plants grown in a light incubator (22/20 • C light/dark and 12-h photoperiod) were subjected to natural drought (withholding water) treatment for 21 d.

Determination of Physiological Indexes
The malondialdehyde (MDA) content was determined by the thiobarbituric acid (TBA) method [46] and the peroxidase (POD) activity and catalase (CAT) activity were determined by the guaiacol chromogenic method and ultraviolet absorption method,