Small RNA and degradome deep sequencing reveal respective roles of cold-related microRNAs across Chinese wild grapevine and cultivated grapevine

Background Chinese wild grapevine (Vitis amurensis) has remarkable cold stress tolerance, exceeding that of the common cultivated grapevine (Vitis vinifera L.). Result Here, we surveyed the expression dynamics of microRNAs (miRNAs) across Chinese wild grapevine (cv. Beibinghong) and cultivated grapevine (cv. Cabernet Sauvignon) under cold stress using high-throughput sequencing. We identified 186 known miRNAs in cultivated grape and 427 known miRNAs in Beibinghong. Of the identified miRNAs, 59 are conserved miRNAs orthologous in Cabernet Sauvignon and Beibinghong. In addition, 105 and 129 novel miRNAs were identified in Cabernet Sauvignon and Beibinghong, respectively. The expression of some miRNAs was related to cold stress both in Cabernet Sauvignon and Beibinghong. Many cold-related miRNAs in Cabernet Sauvignon and Beibinghong were predicted to target stress response-related genes such as MYB, WRKY, bHLH transcription factor genes, and heat shock protein genes. However, the expression tendency under cold treatment of many of these miRNAs was different between Cabernet Sauvignon and Beibinghong. Different modes of expression of cultivated and Chinese wild grape miRNAs were indicated in key pathways under cold stress by degradome, target prediction, GO, and KEGG analyses. Conclusion Our study indicated three likely reasons that led to the different cold stress tolerance levels of Cabernet Sauvignon and Beibinghong. Specifically, there may be (1) differential expression of orthologous miRNAs between cultivated grapevine and Chinese wild grape; (2) species-specific miRNAs or target genes; or (3) different regulatory models of miRNAs in cultivated and Chinese wild grape in some key pathways.


Background
MicroRNA (miRNA) is an 18-24 nucleotide-long noncoding small RNA (sRNA) that can regulate gene expression by blocking protein translation through binding with the 3′ untranslated region (3'UTR) of the target gene mRNA in animals [1]. In plants, miRNAs can negatively regulate target gene mRNA by two major mechanisms: transcript cleavage and translational inhibition [2,3]. MicroRNAs have important regulatory functions in the growth and development of plants. Moreover, miRNAs have been confirmed to be involved in various abiotic stress responses in plants, such as cold stress (chilling or freezing) [4][5][6][7].
Cold stress is known to impact the survival, productivity, and geographical distribution of plants [8]. Cold stress tolerance is a complex trait in plants [9,10]. Cold stress could activate the Inducer of CBF Expression (ICE) gene, which is a type of basic helix-loop-helix (bHLH) gene, via the ABAindependent pathway [1,[11][12][13][14], and ICE may activate the downstream transcription factor C-repeat (CRT)-binding factors (CBF) or dehydration-responsive element binding (DREB) that binds to C-repeat elements (CRT)/lowtemperature-response elements (LTRE) [1,11]. Inducing the expression of downstream genes, including cold-responsive (COR) [10,15], CBF, and DREB, could play critical roles in the cold acclimation process [10,16]. Micro-RNAs are also involved in the regulatory network of these proteins [1]. COR and other cold-regulated genes could produce two types of products. The first include heat shock proteins (HSP), antifreeze proteins, late embryogenesis abundant proteins (LEA), lipid transfer proteins, dehydrins, and compatible solutes such as raffinose, free sterols, sugars, glucosides, proline, and glycine betaine [11,[17][18][19]. The second products are various transcript factors (TFs). These TFs are involved in the regulation of signal transduction and the expression of downstream genes [20]. In addition, the abscisic acid (ABA) level in several coldtolerant plant varieties was found to be significantly higher than in cold-sensitive varieties during cold acclimation [1].
The same study showed that ABA hormone signaling, which is involved in the cold stress response of herbaceous tissues, is controlled by various miRNAs [1]. The expression of many miRNAs was dynamic under cold stress in plants. For example, miR397 and miR169 were found to be up-regulated in Arabidopsis, Brachypodium, and poplar under cold stress [21]. MiR168 was upregulated both in poplar and Arabidopsis [22], but miR168 was down-regulated in rice [23] during cold stress. Previous studies have shown that some miRNAs, such as miR394a, miR319, miR408, and miR396, were involved in the regulation of cold stress response in some plants [24][25][26]. For example, one previous study showed that 35S:At-miR394a OE Arabidopsis lines showed higher free proline levels and more total soluble sugars than wild type plants [24]. In addition to enhanced expression of other cold responsive genes (such as CBF1, CBF2, RD29A, and COR15a), an increased expression (up to 90-fold) of CBF3 was also observed in At-miR394a Arabidopsis OE lines [24]. Enhanced accumulation of free proline and increased expression of cold stress-related genes were detected in At-miR319 Arabidopsis OE lines [25]. The 35S:At-miR408 lines showed higher survival, higher optimal/maximal quantum yield of PSII (Fv/Fm) indicating the efficiency of photosystem II, low electrolyte leakage, and lower levels of MDA than At-miR408-knockout lines and wild type plants [26]. However, the orthologs of these miRNAs in grapevine were not shown to be involved in the regulation of cold stress response.
Cold stress is an important environmental factor that can negatively affect grapevine productivity and quality. Chinese wild grapevine (Vitis amurensis Rupr.) is more cold-tolerant than the common cultivated grapevine (Vitis vinifera L.) [27]. However, the molecular mechanisms that contribute to the enhanced cold stress tolerance of Chinese wild grapevine remain unknown. This study aimed to investigate the relationship between cold stress tolerance of Chinese wild grapevine and miRNA regulation. Here, sRNA and degradome deep sequencing were performed to detect and analyze cold-related miR-NAs and their target genes in Chinese wild grapevine. Cabernet Sauvignon, a cold-sensitive grapevine cultivar (cv), was also examined to detect and analyze the coldrelated miRNAs and their target genes.

Results
Data mining of small RNA sequencing To reveal grape miRNAs related to cold, 12 small RNA libraries were constructed and sequenced. These samples included three cold-treated Cabernet Sauvignon sample repetitions (cold cultivated grape), three untreated Cabernet Sauvignon sample repetitions (CK cultivated grape), three cold-treated Beibinghong sample repetitions (cold Chinese wild grape), and three untreated Beibinghong sample repetitions (CK Chinese wild grape). After sequencing, a total of 364,979,306 raw and 334,324,103 clean reads were obtained. There was an average of 30,414,942 raw and 27,860,342 (91.61%) clean reads in each library. The clean reads were used to analyze the small RNAs. sRNAs of 20 nt long were the most abundant sRNA in both Cabernet Sauvignon and Beibinghong (Additional file 22: Figure S1). On average, 81.61 and 81.18% of clean reads from untreated Cabernet Sauvignon and coldtreated Cabernet Sauvignon samples, respectively, could be mapped to the grape genome database v2.0 (http://genomes.cribi.unipd.it/grape/). The average mapping of clean reads from untreated Beibinghong samples and cold-treated Beibinghong on the grape genome database were of 81.66 and 80.05%, respectively. Using the BLASTn and blastall search options against the GenBank and Rfam databases [28], all small RNAs identified from the 12 small RNA libraries were annotated, including rRNA, snoRNA, tRNA, and snRNA-derived (Table 1).

Identification and expression patterns of known miRNAs
In Cabernet Sauvignon, we identified 186 known miR-NAs, which were also deposited in grape miRNA data from the miRBase database (v20) (Additional file 1: Table S1). In Beibinghong, we identified 427 known plant miRNAs (Additional file 2: Table S2), of which 59 known miRNAs are conserved miRNAs orthologous in Cabernet Sauvignon and Beibinghong (only exhibiting one to three nucleotide differences in the sequences; Additional file 3: Table S3). For example, orthologous pairs include Chinese wild grape miR172a with cultivated grape Vvi-172a; Chinese wild grape miR159a with cultivated grape Vvi-159a; and Chinese wild grape miR160b with cultivated grape Vvi-miR160b.
In Cabernet Sauvignon, three known miRNAs were expressed only in the untreated library, and eight known miRNAs were expressed only in the cold-treated library. In cold-treated Cabernet Sauvignon, 41 known miRNAs were significantly (log2 fold change > 1 and P < 0.05) upregulated and 33 known miRNAs were significantly (log2 fold change<− 1 and P < 0.05) down-regulated. In Cabernet Sauvignon, Vvi-miR395h was the most up-regulated by cold treatment, followed by Vvi-miR171c and Vvi-miR164a. Vvi-miR171a was the most down-regulated by cold treatment, followed by Vvi-miR395c and Vvi-miR395b. In Cabernet Sauvignon, most miR169, miR171, miR172, miR167, miR393, miR319, and miR396 family members were significantly (log2 fold change > 1 and P < 0.05) up-regulated by cold treatment. In Cabernet Sauvignon, most miR397 family members and miR408 were significantly (log2 fold change <− 1 and P < 0.05) downregulated by cold treatment (Additional file 4: Table S4). In Beibinghong, 78 known miRNAs were expressed only in the untreated library and 77 known miRNAs were expressed only in the cold-treated library. In cold-treated Beibinghong, 116 known miRNAs were significantly (log2 fold change > 1 and P < 0.05) up-regulated and 109 known miRNAs were significantly (log2 fold change <− 1 and P < 0.05) down-regulated. In Beibinghong, miR3932b-5p was the most up-regulated by cold treatment, followed by miR156b-3p and miR3633a-3p; miR2118f was the most down-regulated by cold treatment, followed by miR171 and miR8036-3p. In Beibinghong, the miR169 family member miR169z was significantly up-regulated by cold treatment. In Beibinghong, miR171 and the miR172 family member miR172a were significantly down-regulated by cold treatment (Additional file 5: Table S5). In the 59 known orthologous miRNAs, 40 miRNAs had a different expression tendency between the cultivated grape and Chinese wild grape in the cold treatment (Additional file 6: Table S6).

Identification of potentially novel miRNAs and analysis of their expression changes under cold stress
We identified 105 novel miRNAs and 129 novel miRNAs in cultivated grape and Chinese wild grape, respectively (Additional file 7: Table S7 and Additional file 8: Table   S8). All of the miRNA complement sequences of these novel miRNAs were found in our study, and their mature sequences and star sequences are listed in Additional file 7: Table S7 and Additional file 8: Table  S8. Some structures of novel miRNA precursors of cultivated grape or Chinese wild grape are listed in Fig. 1.
In Cabernet Sauvignon, no novel miRNAs were expressed only in the untreated library, and three novel miRNAs were expressed only in the cold-treated library. Thirty novel miRNAs were significantly (log2 fold change > 1 and P < 0.05) up-regulated and one novel miRNA was significantly (log2 fold change <− 1 and P < 0.05) down-regulated in cold-treated cultivated grape. In Cabernet Sauvignon, the novel miRNA Vvi-novel_mir69 was the most up-regulated by cold treatment, followed by Vvi-novel_mir47 and Vvi-novel_mir72. Vvi-novel_ mir6 was the most down-regulated by cold treatment (Additional file 3: Table S3). No novel miRNAs were expressed in only the untreated and cold-treated Chinese wild grape. However, six novel miRNAs were significantly (log2 fold change > 1 and P < 0.05) up-regulated and two novel miRNAs were significantly (log2 fold change <− 1 and P < 0.05) down-regulated in coldtreated Beibinghong. In Beibinghong, novel_mir120 was the most up-regulated by cold treatment, followed by novel_mir73 and novel_mir25. Novel_mir55 was the most down-regulated in the cold treatment, followed by novel_mir79 and novel_mir98 in cold-treated Beibinghong (Additional file 4: Table S4).

Target prediction and degradome analyses
In cultivated grape, a total of 1986 target genes (protein-coding genes) were predicted in silico (Additional file 9: Table  S9). These genes could be targeted by 230 miRNAs. Our results showed that 85 differentially expressed cultivated grape miRNAs were predicted to target 925 protein-coding genes in silico. These differentially expressed cultivated grape miR-NAs targeted AP2-like ethylene-responsive transcription factor genes, transcription factor HY5 genes, transcription factor TCP2 genes, squamosa promoter-binding-protein-like (SBP) genes, MYB transcription genes, transcription factor GAMYB genes, and heat shock protein 90.1 isoform 1 genes, which are stress-related genes. Differentially expressed cultivated grape miRNAs also targeted auxilin-like protein genes, auxin response factor 18 genes, and ETHYLENE INSENSI-TIVE 3-like genes, which are hormone signal transductionrelated genes (Additional file 10: Table S10). In Chinese wild grape, a total of 11,543 target genes (protein-coding genes) were predicted. In total, 499 Chinese wild grape miRNAs could target these genes (Additional file 11: Table S11). Our results showed that 212 differentially expressed Chinese wild grape miRNAs were predicted to target 4959 protein-coding genes (Additional file 12: Table  S12), which targeted WRKY transcription genes, MYB transcription genes, ethylene-responsive transcription factor genes, bHLH transcription genes, transcription factor TCP2 genes, NAC transcription factor genes, SBP genes, GATA transcription factor genes, TIR-NB-LRR type resistance protein genes, heat shock cognate 70 kDa protein genes, and universal stress protein genes, which are stress-related genes. Differentially expressed miRNAs of Beibinghong also targeted auxin-induced protein genes and abscisic acid receptor genes as well as photosynthesis, amino acid, and sugar transporter related genes such as chlorophyll a-b binding protein genes, bidirectional sugar transporter genes, and cationic amino acid transporter genes (Additional file 12: Table S12).
The 40 differentially expressed miRNA orthologs were predicted to target 427 genes including NAC domain-containing protein genes, ethylene-responsive transcription factor genes, AP2-like ethylene-responsive transcription factor TOE3 genes, microtubule-associated protein genes, auxin transport protein genes, and promoter-binding protein genes (Additional file 13: Table S13).
We employed degradome sequencing to identify miRNA targets. After sequencing, a total of 11,253,630, 12,596,309, 12,548,638, and 12,367,509 clean tags were obtained from untreated Cabernet Sauvignon, coldtreated Cabernet Sauvignon, untreated Beibinghong, and cold-treated Beibinghong, respectively. In cultivated grapevine, 141 target genes and 145 specific miRNA-mRNA pairs were identified using degradome sequencing (Additional file 14: Table S14). In Chinese wild grapevine, 359 target genes and 375 specific miRNA-mRNA pairs were identified using degradome sequencing (Additional file 15: Table S15). Target genes related to stimulus responses included WRKY transcription factor 47-like, heat shock cognate 70 kDa protein 2 gene cleaved by Chinese wild grape miR160b, and novel_ mir64 (Additional file 15: Table S15). Fifty-one predicted target genes in silico were validated using degradome sequencing (36% of target genes identified using degradome sequencing) in cultivated grape. In silico, 124 predicted target genes were validated using degradome sequencing (35% of target genes identified using degradome sequencing) in Beibinghong. The target genes identified included miRNA-mRNA pair novel_mir9-VIT_200s0179g00260.1 and vvi-miR3632-3p-VIT_ 200s0144g00270.1 (Fig. 2). This showed that the in silico prediction of target genes was relatively reliable.

qRT-PCR validation of miRNA expression
We performed qRT-PCR to validate the results of small RNA deep sequencing using six miRNAs from cultivated grape, Vvi-miR171c, Vvi-miR172d, Vvi-miR3624-5p, Vvi-miR166g, Vvi-miR397a, and Vvi-miR3634-5p. The primers used for the qRT-PCR are listed in Additional file 21: Table S21. qRT-PCR was performed using the Cabernet Sauvignon sample under control and cold conditions with three technical and three biological replicates. The expression results were similar to the deep sequencing data. Vvi-miR171c was significantly upregulated in the qRT-PCR analysis, showing a positive correlation with the deep sequencing results. Vvi-miR166g was significantly down-regulated in both qRT-PCR and small RNA deep sequencing results. Other miRNAs were not significantly changed in either the qRT-PCR or small RNA deep sequencing results (Fig. 6).

Discussion
As post-transcriptional gene regulators, miRNAs have been shown to be involved in the regulation of cold stress response in plants [24][25][26]29]. For example, overexpression of miR397a could increase the transcript levels of cold-induced CBF (CBF1, CBF2, and CBF3) and downstream cold responsive genes in Arabidopsis [30]. In Brachypodium, miR397, miR172, miR171, miR169, and miR408 could be induced by cold stress [31]. In Arabidopsis, miR393 and miR397 could be induced by cold stress [4]. In poplar, miR167, miR171, and miR1450 could be induced by cold stress [32]. In sugarcane, miR319 could be induced by cold stress [33]. In cultivated grapevine cv. Muscat Hamburg, most miR171 family members, miR172 family members, miR172d, and most miR395 family members could be down-regulated by cold stress [34]. Investigations into cold-related miR-NAs in cultivated grape and Chinese wild grape are limited; therefore, the aims of our research were to detect the cold-related miRNA in cultivated grape and Chinese wild grape to understand the roles of miRNAs in grape cold tolerance, and to compare their different functions in cold stress tolerance process between cultivated grape and Chinese wild grape.
Differential expression of orthologous miRNAs between cultivated grapevine and Chinese wild grape may Lead to different cold stress tolerance Here, we employed deep sequencing technology to systematically study the effect of cold stress on miRNA expression in cultivated grape and Chinese wild grape. We found that 59 known miRNAs are conserved miR-NAs orthologous in cultivated grape and Chinese wild grape. Forty of the identified miRNA orthologs showed different expression tendencies between the two taxa when exposed to the cold treatment. The cultivated grapevine Vvi-miR171a was significantly down-regulated in the cold treatment (Additional file 4: Table S4). This is consistent with the changes in the expression of the miRNA in cultivated grapevine cv. Muscat Hamburg [34]. In Chinese wild grape, miRNAs orthologous miR171a were up-regulated by cold treatment. The target genes of Vvi-miR171a were F-box/LRR-repeat protein, scarecrow-like protein 22 isoform X2, squamosa promoter-binding-like protein 13A, and 7-deoxyloganetic acid glucosyltransferase gene.
In cultivated grapevine, Vvi-miR172a was up-regulated under cold stress, and Vvi-miR164b and Vvi-miR535a were downregulated under cold stress; the expression tendency of the miRNA orthologs in Chinese wild grape was the opposite. Vvi-miR535a, vvi-miR172a, and Vvi-miR164b targeted the late embryogenesis abundant (LEA) protein, AP2-like ethylene-responsive transcription factor, and NAC transcription factor gene, respectively in cultivated grapevine. The targets of these miRNAs in Chinese wild grape were the same as in cultivated grapevine. Some AP2-like ethyleneresponsive transcription factors could interact with CRT/ DER, increasing the expression level of some COR genes and the cold tolerance of plants [35]. Some LEA proteins have been shown to be involved in the freezing tolerance of plants [36]. Previous research has shown that some NAC transcription factors negatively regulated cold tolerance via the CBF-dependent pathway [37]. These orthologous miRNAs were coldrelated and were differentially expressed. It is possible that differential expression of orthologous miRNAs between cultivated grapevine and Chinese wild grape may lead to the difference in their tolerance to cold stress.
Species-specific miRNAs or target genes may Lead to different cold stress tolerance of cultivated grapevine and Chinese wild grape The target of Chinese wild grape miR858a was MYB transcription factor, MYB3. Many MYB transcription factor genes have been shown to work together in the regulation of cold tolerance in plants [38][39][40]. The target of Chinese wild grape miR2109-5p was heat shock cognate 70 kDa protein gene. Heat shock proteins (HSPs) have also been shown to accumulate in response to low temperatures [41,42]. The target of Chinese wild grape miR169a-5p was WRKY transcription gene WRKY2. WRKY transcription factors have been shown to be involved in modulating gene expression in plants during cold stress [43,44]. The target of Chinese wild grape miR8179 was bHLH49 gene. Basic helix-loop-helix (bHLH) transcription factors positively modulate the CBF-pathway and confer tolerance to cold-stress in plants [45]. In Chinese wild grape, these miRNAs and target genes were identified to be cold-related. The above mentioned Chinese wild grape miRNAs were not found in cultivated grape. Some orthologous miRNAs between cultivated grapevine and Chinese wild grape have speciesspecific target genes. For example, Chinese wild grape miR171a could target endonuclease III homolog 1, chloroplastic-like isoform X1, polycomb group protein EMBRYONIC FLOWER 2-like isoform X3, and glucose-1-phosphate adenylyltransferase large subunit 1 gene. However, cultivated grapevine Vvi-miR171a could not target these genes. These data are consistent with a previous study on cultivated grapevine cv. Muscat Hamburg [34]. We suspect that speciesspecific miRNAs or target genes led to the different cold stress tolerance levels of cultivated grapevine and Chinese wild grape.

Different regulation models of miRNAs in key pathways may be essential to different cold stress tolerance
Results from the KEGG annotation showed that 38 targets of differentially expressed miRNAs of cultivated grape were involved in adaptation to the environment, and 460 targets of differentially expressed miRNAs of Chinese wild grape were involved in adaptation to the environment (Fig. 4). In view of the differences between the two cultivars of grapes indicated by the results of the KEGG annotation analysis, we analyzed the KEGG pathway location of targets of differentially expressed cultivated grape and Chinese wild grape miRNAs.
In the same pathways under cold stress, the miRNAs of the cultivated grape and Chinese wild grape may also play different roles. The ABA signal transduction pathway has been shown to be related to cold tolerance in plants [46]. In ABA signal transduction pathways (Fig. 5), the cultivated grape miRNA Vvi-miR166a, which targeted the ABAresponsive promoter element binding factor (ABF) gene VIT_208s0007g03420.1, was up-regulated under cold stress. In Chinese wild grape, miR166a was not found. Additionally, in Chinese wild grape, miR5234, miR7776-5p, miR166h-3p, and miR472b targeted ABF genes. However, these four miRNAs were down-regulated in cultivated grape. In Chinese wild grape ABA hormone signal transduction pathways, miR6020a-5p, which targeted the SNF1related protein kinase gene (SnRK2; upstream of ABF and could increase the expression of ABF) was down-regulated by cold treatment. MiR6020a-5p was not found in cultivated grape. The up-regulation of Vvi-miR166a may degrade the ABF gene VIT_208s0007g03420.1, affecting the ABA signal transduction pathway and ABA-dependent cold response network under cold stress. In Chinese wild grape, the down-regulation of miR6020a-5p may be conducive to the accumulation of SnRK2, allowing the normal function of the ABA signal transduction pathway and regulating downstream genes related to cold tolerance [46]. JA positively regulates the C-REPEAT BINDING FACTOR (CBF) transcriptional pathway to up-regulate downstream coldresponsive genes and ultimately improve cold tolerance [47]. In Chinese wild grape JA hormone signal transduction pathways (Fig. 6), miR7755-3p and miR9408-3p, which targeted MYC genes VIT_202s0025g03450.1-3 and VIT_215s0046g00320.1, were down-regulated. However, in cultivated grape, the expression of miR-NAs targeted to MYCs was unchanged. MYC has been shown to be related to cold tolerance in plants and to play a key role in ABA-dependent cold response network [46]. This showed the miRNAs may be regulated to promote the accumulation of MYC genes and the JA signal transduction in Chinese wild grape but not in cultivated grape under cold stress. In the ethylene signal transduction pathway, the over-expression of EIN3 has been shown to decrease cold tolerance in plants [48]. Interestingly, Chinese wild grape miRNAs novel_ mir61 and miR396f, which targeted the EIN3 gene VIT_200s0357g00120.1 and the EIN2 gene VIT_ 208s0040g01730.2, were down-regulated in the ethylene signal transduction pathway. These two miRNAs were not found in cultivated grape.
Our research showed that in the starch and sucrose metabolism pathway (Additional file 23: Figure S2), the cultivated grape miRNA Vvi-miR3629a-3p that targeted the Dglucose synthesis-related lysosomal beta glucosidase gene (VIT_213s0073g00550.1) was up-regulated under cold stress. In Chinese wild grape, miR3629a-3 was not upregulated and could not target lysosomal beta glucosidase gene. In contrast, miR5079a, miR6253, and miR9654a-3p, which targeted the lysosomal beta glucosidase genes VIT_ 207s0141g00250.1, VIT_216s0098g01760.4, and VIT_ 218s0089g00940.1, respectively, were down-regulated in Chinese wild grape. Glucose can improve the cold tolerance of plants [11,[17][18][19]. This showed that miRNAs may be regulated to promote the accumulation of the D-glucose synthesis-related lysosomal beta glucosidase gene in Chinese wild grape, but not in cultivated grape under cold stress.

Conclusions
In this study, we analyzed cold-related miRNAs and their target genes as well as the relationship between cold stress tolerance and miRNA regulation in two grapevine cultivars. We present three like explanations for the different cold stress tolerance levels of cultivated grapevine and Chinese wild grape, namely, (1) differential expression of orthologous miRNAs between cultivated grapevine and Chinese wild grape; (2) speciesspecific miRNAs or target genes; (3) different regulatory models of miRNAs in cultivated grapevine and Chinese wild grape in some key pathways.

Plant materials
The parents plants of Chinese wild grapevine cv. Beibinghong and cultivated grapevine cv. Cabernet Sauvignon used in this study were grown in the National Field Gene Bank for Amur Grapevine, Zuojia Town, Jilin Province, China and provided by Mr. Yiming Yang and Changyu Li. Chinese wild grapevine cv. Beibinghong was formally identified by Mr. Rungang Song and Wenpeng Lu (National Field Gene Bank for Amur Grapevine, Zuojia, China) and grapevine cv. Cabernet Sauvignon was formally identified by Baron Hector Brane (Chateau Mouton Rothschild, Pauillac, France). One-year-old self-rooted seedlings of grapevine cv. Cabernet Sauvignon and Chinese wild grapevine cv. Beibinghong were grown and maintained in the greenhouse under a 16-h light/8-h dark photoperiod at 26°C. For the cold stress treatment, plant materials under a 16-h light/8-h dark photoperiod were transferred to 4°C for 4 hours. For the control (CK), plants were kept under a 16-h light/8-h dark photoperiod at 26°C for 4 hours. The shoot apices with leaves were collected from each treatment. Three biological replicates were used for the CK and cold treatment samples. RNA of each biological replicate was isolated, used for the construction of small RNA and degradome libraries, and were also use for real-time RT-PCR analysis.

Construction of small RNA libraries and deep sequencing
RNAs were sent to BGI (Shenzhen, China) for small RNA library construction and high-throughput sequencing using BGISEQ-500 [49]. RNA of every biological replicate was isolated and was used in the construction of small RNA libraries for high-throughput sequencing.

Identification of known and novel miRNAs using deep sequencing
The raw reads were collapsed and used as input for the web-based tool miRMaster [50], allowing for integrated analysis of miRNA data. Bowtie [51] was used to map the clean reads to the grape genome v2.0 [52] and miRNA data from the miRBase database (v20) [53] incorporating data from grape and other species (one mismatch allowed). The miRDeep2 algorithm was used to quantify miRNAs [54]. The prediction of novel miRNAs was performed using an extended set of eight features built using novoMiRank [55]. For classification, an AdaBoost model using decision trees was applied. Novel miRNAs were cross-checked against other RNA resources, including the miRBase [56], NONCODE2016 [57], and Ensembl non-coding RNAs. The assessment of the quality of new miRNAs was carried out using the novoMiRank algorithm [49]. The miRNAs with zero reads in all three biological replicates were considered absent. Otherwise, the miRNAs were considered present.

Analysis of differentially expressed miRNAs
DEGseq [58] was used to identify the differentially expressed miRNAs based on an MA-plot [59]. Log2 fold change indicates the Log2 ratio (miRNA expression in cold/CK).

Target prediction, GO, and KEGG pathway analyses
For target prediction, we applied psRobot [60] and Tar-getFinder [61] to predicted target genes of all miRNAs. Target genes were annotated based on the Gene Ontology database (http://www.geneontology.org/). Pathway analyses of target genes were performed using the KEGG (The Kyoto Encyclopedia of Gene and Genome) database (http:// www.genome.jp/kegg/kegg1.html) [62]. GO enrichment analysis was conducted according a previous study [28].