High throughput sequencing of small RNA in grape berries of the stone-hardening stage
The sRNA library from the ‘Wink’ cultivar grape stone-hardening berries (SB) at 45 days after flowering (DAF) was constructed and sequenced in depth by Solexa technology to explore the role of sRNA in the grape stone-hardening process (Fig. 1a). Sequencing result analysis showed that total of 11,456,656 redundant and 3,591,857 unique clean reads were identified (Supplementary Table S1), of which about 18.40 and 8.79% were annotated, respectively. Out of them, 5.54 and 1.99% of these two reads, respectively, were further successfully mapped into the non-coding RNA of rRNAs, snRNAs, snoRNAs, and tRNAs. While, 5.71 and 0.04% of the redundant and unique reads, respectively, were determined to be putative known miRNAs, and the majority of the redundant (81.6%) and unique (91.21%) reads, were mapped to un-annotated regions in the grape genome (Fig. 1b).
The length distribution of sRNAs in the SB library were uneven, and most sRNA reads were 21 and 24 nucleotides (nt), which were the characteristic lengths of miRNAs and siRNAs, respectively, and consistent with the expected size range generated by Dicer . In contrast to two other sRNA libraries of 5 DAF berry (YB) and 90 DAF berry (MB) (data were not shown), the sRNA distribution at 21 and 24 nt at SB was more similar to that at MB than that at YB (Fig. 1c). The comparative expression level using normalized counts per million of miRNA (21 nt) and siRNA (24 nt) at grape YB, SB, and MB stages indicated a gradual increase in the known miRNA (21 nt) reads towards the MB stage. By contrast, the siRNA (24 nt) reads exhibited a reverse trend with the highest level observed at the early YB stage followed by a gradual decrease towards the MB stage (Fig. 1c). The above results suggested that known miRNA (21 nt) reads were highly abundant during the later stage (MB) of grape-berry ripening compared with siRNA (24 nt) reads.
Identification and characterization of known VvmiRNAs in the stone-hardening stage of grape berries
A total of 143 known VvmiRNAs and 18 corresponding precursors, namely, VvmiRNAs* belonging to 48 VvmiRNA families, were identified in the stone-hardening stage of grape berries (Supplementary Table S2). Although the number of known VvmiRNA members within each family varied from 1 to 24 during grape-berry development, they had the most variation at the SB stage compared with the other two stages (YB and MB) (Fig. 1d and Fig. 1e). Amongst 48 VvmiRNA families, VvmiR169 family was highly represented during all three stages with 22–24 members, followed by VvmiR395 family with 13 members, VvmiR156 family with 8–9 members, and VvmiR166 family with 8 members, whereas the remaining miRNA families consisted of 1–7 members (Fig. 1d). This result suggested a diversification in the functions of these VvmiRNA families during grape development and ripening. Interestingly, from the percent numbers of VvmiRNA members at different grape-berry developments, we found that VvmiR3628 family was sequenced only in mature berries (90 DAF), whereas VvmiR393 family was detected only in young berries (5 DAF) (Fig. 1e), indicating the stage-specificity of VvmiRNAs’ expression.
We further characterized most known VvmiRNA families with high expression abundances at the stone-hardening stage of grape berries. For instance, amongst the 143 known VvmiRNAs, about 68% exhibited high copy read number, including 42 known VvmiRNAs with read number > 1000, whereas 55 known VvmiRNAs had read numbers ranging between 100 and 1000 (Supplementary Table S2). Specifically, VvmiR166, VvmiR168, VvmiR479, VvmiR156, and VvmiR3636 families (not including their VvmiRNA* sequences) possessed more than 10,000 copy reads (Supplementary Table S2). Meanwhile, 83.3% of known VvmiRNA* family members exhibited lower copy number than their corresponding VvmiRNAs (Fig. 1f and g, and Supplementary Table S2), which may have been due to the fact VvmiRNAs* were easier to degrade than their corresponding VvmiRNAs. Thus, they usually had less copy numbers, similar to a previous report . Nevertheless, two VvmiRNA* families of VvmiR3623* and VvmiR2950* showed higher copy numbers (14,830 and 3029, respectively) than their corresponding mature sequences (4405 and 1773, respectively) (Fig. 1g and Supplementary Table S2). This finding implied that VvmiR3623* and VvmiR2950* might play potential roles during the development and ripening of grape berries, similar to VvmiRNAs [17, 24].
Screening of novel VvmiRNAs at grape stone-hardening stage and their validation using miR-RACE and qRT-PCR
According to the annotation criteria of novel miRNAs , all un-annotated sRNAs were used to explore the stem–loop structures of their precursors for the prediction of novel miRNAs. A total of 90,352 reads were identified as novel VvmiRNAs, including 72 novel VvmiRNAs and 12 novel VvmiRNAs*, in the grape stone-hardening stage (Table 1). These novel precursors were folded into stable hairpin structures, and their negative minimal folding free energy (MFE) ranged from − 108.2 kcal mol− 1 to − 20.88 kcal mol− 1 (Table 1), which was in line with the criteria of novel VvmiRNAs (MFE < − 20.0 kcal mol− 1) as previously reported . The novel VvmiRNAs and VvmiRNAs* were primarily 21 nt in length, accounting for 84.14% (69/82), and the first base with uracil (U) at the 5′-end of their mature sequences reached 56.0%, confirming that these were novel VvmiRNAs (Fig. 2a). From our datasets, the novel VvmiRNAs were unconserved, species specific, and low abundance; they usually exhibited lower accumulation level than conserved ones, and in agreement with previous results [23, 26, 27]. Interestingly, we also observed a few novel VvmiRNAs with high abundance and read number > 1000, such as VvmiR10, VvmiR13, VvmiR29, VvmiR30, VvmiR34, VvmiR37, VvmiR43, and VvmiR71 (Table 1). Amongst them, VvmiR37, VvmiR13, and VvmiR30 showed higher copy numbers (37,985, 13,748, and 10,199, respectively) than the others (Table 1). Furthermore, their corresponding VvmiR37* and VvmiR13* exhibited high copy numbers with 2588 and 1332, respectively, which implying that VvmiR37/VvmiR37* and VvmiR13/VvmiR13* may play significant roles during the stone-hardening stage of grape berries. The location of novel VvmiRNAs in their precursors showed that 40 novel VvmiRNAs were located in the 5′-arm of their precursors, whereas another 32 ones were located in the 3′-arm of their precursors (Table 1). Similarly, seven novel VvmiRNAs* were located in the 5′-arm of their precursors, whereas another 5 ones were located in the 3′-arm of their precursors (Table 1). These results indicated that the 5′-arm of miRNA precursors may be more efficient in generating miRNAs and miRNA* than the 3′-arm. However, further confirmation of this mechanism is necessary. The identified novel VvmiRNAs exhibited three types of VvmiRNA-3P or VvmiRNA-5P or the first sequences (3p and 5p) (Table 1). As shown in Table 1, the miRNA with 3p indicated this miRNA sequence originated only from the 3′ arm of its precursor, and the marked 5p denoting the corresponding miRNA originated just from the 5′ arm of its precursors. Conversely, the miRNA with 3p and 5p represented both arms of miRNA precursors that generated two sequences of miRNA and miRNA* (Table 1). The distribution of all these novel VvmiRNAs varied between the 19 grape Chrs and 1 unknown Chr (Fig. 2b). Amongst them, Chr19 possessed the highest number (10) of novel VvmiRNAs, followed by Chr8 with 9 novel VvmiRNAs and the unknown Chr with 8 ones (Fig. 2b), whereas Chr3, Chr4, and Chr11 did not harbor any novel miRNAs (Fig. 2b).
Subsequently, six novel VvmiRNAs closely related to berry development, such as VvmiR8, VvmiR16, VvmiR31, VvmiR38-5p, VvmiR44-3p, and VvmiR53-3p, were further validated by miR-RACE. Their precise sequences were detected by miR-RACE, and the sequences were consistent with those from the high-throughput sequencing dataset (Table 1 and Supplementary Table S2), which further verified the results of high throughput sequencing. Moreover, their qRT-PCR expression profiles during grape-berry development showed differential expression patterns similar to the high-throughput sequencing dataset (Fig. 2c). Therefore, our miR-RACE and qRT-PCR results confirmed the reliability and expression modes of VvmiRNA involved in the modulation of grape-berry development.
Identification of grape stone-hardening stage-specific VvmiRNAs
The identification of grape stone-hardening stage-specific VvmiRNAs is essential to gain insights into the regulatory roles of grape-berry development. Compared with our other two sRNA libraries from 5 and 90 DAF in our other work (Supplementary Tables S3 and S4 and Fig. 2d), 35 VvmiRNAs/VvmiRNAs* were identified only at the stone-hardening stage of grape berries, including 28 VvmiRNAs (1 known and 27 novel) and 7 VvmiRNAs* (2 known and 5 novel). Among them, there were a large number of stage-specific novel VvmiRNAs indicated that novel miRNAs may play significant roles in the stone-hardening stage of grape-berry development.
To identify the differential expression VvmiRNAs during grape-berry development, the fold changes log2 (YB/SB) or log2 (MB/SB) > 1 cut offs were selected, and the filtered VvmiRNA/VvmiRNA* possessed significant expression difference across diverse development stages of grape berries. Here, we discovered that 52 VvmiRNAs/VvmiRNAs* exhibited significant differences in their expression levels during grape-berry development and ripening, comprising 44VvmiRNAs (34 known and 10 novel ones) and 8 VvmiRNAs* (5 known and 3 novel ones) (Supplementary Tables S3 and S4 and Fig. 2d). This finding indicated that they may possess dynamic regulatory roles of grape-berry development and of them, more known VvmiRNAs may have dynamic variation in their regulatory roles than novel ones.
SNPs and their edit types of known VvmiRNAs/VvmiRNAs* from grape berries at stone-hardening stage
Numerous SNP variations of known VvmiRNAs/VvmiRNAs* and their Edit types were detected in our datasets (Fig. 3), which were consistent with our previous work in ‘Amur’ grape . Identifying the characteristics of VvmiRNA SNP helps to recognition of the evolution of VvmiRNA and its overall roles in the process ofthe stone-hardening stage of grape berries. Amongst 161 types of VvmiRNAs/VvmiRNAs*, 71 types of VvmiRNA SNPs and corresponding Edit types were identified, while the mature sequences of the remaining 90 types of VvmiRNA remained unchanged (Table 2 and Supplementary Table S5). Furthermore, several VvmiRNA families exhibited high SNPs amongst their members. For example, VvmiR166, VvmiR156, and VvmiR167 families exhibited the highest SNPs amongst their members. In contrast, VvmiR169 family possessed the lowest SNP amongst eight members (Table 2). This finding suggested that divergence of conservation in the sequences amongst various VvmiRNA families (Table 2 and Fig. 3), similar to the previous report .
Interestingly, we observed that diverse VvmiRNA families had diverse SNP variations in their Edit types and numbers (Fig. 3), supported by the Amur grape report . To depict this phenomenon clearly, all VvmiRNA families with SNPs were further classified into several groups (Fig. 3). In group I, most members of each VvmiRNA family had an SNP, and each member with an SNP possessed multiple Edit types of SNP. For example, the VvmiR166 family (VvmiR166s) had 8 members and 282 Edit types of SNV (8, 282), followed by VvmiR156s (9, 118), VvmiR167s (5, 40), VvmiR164s (4, 17), and VvmiR535s (3, 9), respectively. From these VvmiRNA families, diverse members with sequence variations obviously exhibited divergent Edit types of SNP. Although the diverse members of one miRNA family with various precursors possessed same mature sequences (e.g., VvmiR166b/c/d/e/f/g/h, VvmiR156b/c/d, VvmiR156f/g/i, VvmiR167b/c/d/e, VvmiR164a/c/d, and VvmiR535a/b/c), they had various Edit types. For instance, VvmiR166b and VvmiR166c/d/e/f/g/h had the same mature sequences, but they possessed 35 and 43 Edit types (35 and 43, respectively), resembling VvmiR156b, VvmiR156c and VvmiR156d (10, 10, and 11, respectively); VvmiR167b, VvmiR167c, VvmiR167d and VvmiR167e (12, 9, 5, and 12, respectively); and VvmiR164a, VvmiR164b, VvmiR164c, and VvmiR164d (5, 6, and 5, respectively) (Fig. 3). These findings suggested the diversification of the assorted VvmiRNA families in the evolution of the sequences. In group II, other VvmiRNA families only had one member, but it possessed multiple Edit types, such as VvmiR168 (34), VvmiR479 (31), VvmiR3636 (25), VvmiR3623*(19), VvmiR3624 (4), VvmiR3633a (4), VvmiR162 (3), VvmiR3623 (3), and VvmiR2950* (3). This finding implied that VvmiRNA families with single member exhibited drastic divergence and may thus be active factors during VvmiRNA sequence evolution. In group III, some VvmiRNA families such as VvmiR169b/c/g/h/i/l/r/u, VvmiR160a/b/c/d, VvmiR399a/b/c/d, and VvmiR3629 were also revealed only one Edit type even though they had multiple members with SNPs, indicating that they may possess relatively high conservation during VvmiRNA sequence evolution. In the final group, the remaining VvmiRNA families had fewer members and Edit types (Fig. 3). All these results confirmed the diversification of VvmiRNA families in the evolution of their mature sequences.
The total read number of VvmiRNAs with SNP was further revealed to reach 77,141,827, and diverse VvmiRNA families and their various members had conspicuous divergence in the number of sequences with SNP. Amongst them, VvmiR166 and VvmiR156 families had considerably more reads with SNP than the other families (Table 2). Generally, the number of SNPs in VvmiRNA families was less than that of the normal sequences. However, the SNP sequences of 21 VvmiRNAs had more read numbers than miRNAs themselves, including VvmiR156a, VvmiR156e, VvmiR156h, VvmiR160a, VvmiR160b, VvmiR169b, VvmiR169f, VvmiR169g, VvmiR169h, VvmiR169r, VvmiR169u, VvmiR171e, VvmiR3629a, VvmiR3629b, VvmiR3629c, VvmiR3631b*, VvmiR396b, VvmiR396c, VvmiR396d, VvmiR399b, and VvmiR399c, suggesting that these VvmiRNAs had stronger evolution than the others. Interestingly, compared with homologous VvmiRNAs from the grape cv. ‘Pinot Noir’ in miRBase 21.0 (http://www.mirbase.org/summary.shtml?org=vvi), some VvmiRNAs could not be identified in this work. However, their SNP sequences such as VvmiR164b, VvmiR169i, and VvmiR828b were identified (Supplementary Table S5; bold and italic words). Thus, SNPs may explain the generation of new members of VvmiRNA family in miRNA evolution.
Functional annotation of specific-VvmiRNA targets during grape stone-hardening stage
To further recognize the roles of VvmiRNA during grape stone-hardening stage, PsRNATarget software (http://plantgrn.noble.org/psRNATarget/result?sessionid=1503987414486479) was utilized to predict the potential miRNA target genes on the basis of our previous RNA-seq data (GEO Accession: GSE77218) by using mature miRNA sequences as queries. A total of 2124 targets for known VvmiRNAs and 885 targets for novel VvmiRNAs were predicted in this work, amongst which 1639 and 1635 target genes may be targeted by 24 stone-hardening-specific VvmiRNAs and VvmiRNAs with significant expressional difference, respectively. GO and KEGG analyses were performed with these predicted target gene sequences to improve our understanding of their functions in grape stone-hardening stage. Totally, 13 VvmiRNAs might be involved in the regulation of embryo development, 11 in lignin and cellulose biosynthesis, and 28 in the modulation of hormone signaling, sugar, and proline metabolism (Supplementary Table S6). From the Fig. 4a, the top 20 of GO enrichment of the stone-hardening-specific VvmiRNA targets included 18 GO enrichment of biological process and 2 of molecular function, whereas those of VvmiRNAs with significant difference had 6 GO enrichment of biological process, 13 of molecular function, and 1 of cell component (Fig. 4b). Further comparison of the biological processes between Fig. 4a and Fig.4b, the target genes for stone-hardening-specific VvmiRNAs primarily focused on the lipid, phosphatelipid biosynthesis/metabolic process (GO: 0008610; GO: 0008654; GO: 0006644), phosphatidylinositol biosynthesis/metabolic process (GO: 0006661; GO: 0046488), and glycerophospholipid/glycerolipid biosynthetic/metabolic process (GO: 0046474; GO: 0006650; GO: 0046486). These GO pathways were closely related to seed development, supporting the modulation of the stone-hardening-specific VvmiRNAs on grape seed development. Whereas, those of VvmiRNAs with significant difference largely participated in cellular metabolic process (GO: 004237), ligin metabolic process (GO: 0009808), secondary/aromatic metabolic process (GO: 0019748; GO: 0019439), cell-wall organisation or biogenesis (GO:0071555; GO: 0071554), and peptidyl-proline modification (GO: 0018208), thereby providing evidence of supporting their potential regulatory roles in grape-berry development and quality formation.
Meanwhile, an overview of the KO pathways of VvmiRNA-mediated target genes was further generated by KEGG pathway analysis (http://www.genome.jp/kegg/). A total of 73 KO pathways were identified by 228 targets for VvmiRNAs (Supplementary Table S6), among which the stone-hardening-specific VvmiRNAs-mediated target genes largely participated in the KO pathways as showed in Fig. 4c. Amongst them, metoblism pathway (ko01100), inostol phosphate biosynthesis (ko00562), phosphatidylinositol signaling system (ko04070), biosynthesis of secondary metabolites (ko01110), stilbenoid, diarylheptanoid and gingerol biosynthesis (ko00945), carbon metabolism (ko01200), plant hormone signal transduction (ko04075), and pentose and glucuronate interconversions (ko00040) were the main pathways. These results supported this view of these specific VvmiRNAs involved in the modulation of seed and berry development during grape stone-hardening stage. By contrast, VvmiRNAs had significant difference-mediated target genes in 17 KO pathways (Fig. 4d), primarily enriched in nucleotide excision repair (ko03420), metabolic pathways (ko01100), biosynthesis of secondary metabolites (ko01110), NF-kappa B signaling pathway (ko04064), apoptosis (ko04210), citrate cycle (TCA cycle) (ko00020), carbon metabolism (ko01200), and so on. These findings indicated the potential role of these VvmiRNAs in the regulation of grape-berry development.
Verification of target genes for novel VvmiRNAs related to berry development at the stone-hardening stage of grape berries
Based on the expression profiles of novel VvmiRNAs (Fig. 2c), together with their potential functional annotation, we found four novel VvmiRNAs closely related to the berry development of VvmiR8-5p, VvmiR31-3p, VvmiR38-5p, and VvmiR53-3p. Their corresponding target genes involved in embryo and seed stone development [VIT_204s0008g03060, DDB1- and CUL4-ASSOCIATED FACTOR HOMOLOG 1 (VvDCAF1)] , GA signaling [VIT_217s0000g10300, GIBERELLIN INSENSITIVE (VvGAI1)] , lignification [VIT_212s0028g03110, CAFFEOYL-CoA O-METHYLTRANSFERASE (VvCCoAOMT)] , and cell wall deacetylation [VIT_218s0041g02160 (ESTERASELLIPASE,VvGDSL)] , which were further selected to verify their roles at the stone-hardening stage of grape berries. The cleavage interactions of these four VvmiRNAs on their target genes above at the berries of grape stone-hardening stage were verified with our modified RNA ligase-mediated 5′-rapid amplifification of cDNA ends (RLM-RACE) and developed poly(A) polymerase-mediated 3′-rapid amplifification of cDNA ends (PPM-RACE) procedures [32, 33], which are high effeciency, low cost strategies for validating the true target genes for miRNAs by sequencing the 3′- / 5′- end cleavage products of their target genes.
First, using RLM-RACE, the sequencing of the amplified 3′-end products confirmed that the VvmiRNAs’ cleavage on their target genes occurred at the 9th–11th sites, amongst which the 10th was their main cleavage site with the highest cleavage frequency (Fig. 5). This result indicated the specificity of cleavage sites of miRNAs, consistent with previous findings [32, 33]. Although VvmiR8 and VvmiR31 had two cleavage sites on their corresponding targets (VvDCAF1 and VvCCoAOMT), the cleavage sites of the former were at the 9th and the 10th (the highest frequency 18/20), and those of the latter were at the 10th (the highest frequency 10/16) and the 11th. Similarly, VvmiR38-5p on VvGAI1 possessed three cleavage sites at the 9th–11th, amongst which the 10th had the highest frequency 18/22. Meanwhile, VvmiR53-3p on VvGDSL possessed only one cleavage site at the 10th with the same high frequency 18/22. Next, our developed PPM-RACE was used to further confirm the target genes of VvmiR8-5p, VvmiR31-3p, VvmiR38-5p, and VvmiR53-3p and their cleavage sites. The sequencing of the amplified 5′-end products identified the same cleavage sites as those of the 3′-end sequencing in the RLM-RACE experiment, but all their 5′-end cleavage product accumulation levels and their cleavage frequency detected in PPM-RACE were lower than their corresponding 3′-end cleavage frequency examined in RLM-RACE (Fig. S1; Fig. 5). This phenomenon may be due to the fact that the 5′-end cleavage products were more easily degraded than the 3′-end cleavage products, similar to previous reports [32, 33]. The consistent results of the RLM-RACE and PPM-RACE experiments demonstrated that VvDCAF1, VvCCoAOMT, VvGAI1, and VvGDSL were the true target genes of VvmiR8-5p, VvmiR31-3p, VvmiR38-5p, and VvmiR53-3p, thereby verifying their cleavage-interaction mode in the stone-hardening stage of grape berries.
Expression modes of VvmiRNAs and their target genes during GA-induced grape seedless-berry development
To gain insight into seed development during grape stone-hardening stage and GA-induced parthenocarpy process, we compared berries phenotype and seed morphology of GA-treated and untreated control (CK) ‘Wink’ grape cultivar at 5, 20, 45, and 90 DAF were carried out (Figs. 6a–d). Amongst them, the berries at 45 DAF in untreated control group had full seeds and hardened seed coats (Fig. 1a). GA-treated grape berries showed more distinct increased in vertical diameter of berries grains at 20 and 45 DAF as compared with untreated CK groups, but no variation were observed in horizontal diameter (Fig. 6b). Additionally, GA-treated grapes showed 99.6% seedless rate (Fig. 6d), compared with untreated CK groups, GA-treated grape seeds were drastically inhibited growth, and there were almost no growth from fruit setting to maturation. These results confirmed the profound effects of GA-induced ‘Wink’ grape parthenocarpy, suggesting that complicated regulatory networks might exist during GA-mediated grape berries and seed development.
To determine the long-term roles of VvmiRNAs and their target genes validated above during GA-induced grape seedless-berry development, the expression levels of VvMIR8-5p, VvMIR31-3p, VvMIR38-5p, and VvMIR53-3p and their corresponding target genes VvDCAF1, VvGAI1, VvCCoAOMT, and VvGDSL were examined in berries at 5, 20, 45, and 90 DAF, respectively. Results (Fig. 6e) showed that except for VvMIR53-3p and its target gene, the remaining three VvMIRNAs and their target genes exhibited significant expression differences in response to GA treatments relative to untreated control plants, whereas VvMIR53-3p and its target gene VvGDSL hardly differed in response to GA treatment than the control. Notably, the former three miRNAs VvmiR8-5p, VvmiR31-3p, and VvmiR38-5p expression were strongly up-regulated by GA treatment. Specifically, VvmiR8-5p and VvmiR31-3p displayed the highest expression level at the grape stone-hardening stage (45 DAF) in response to GA treatment relative to other stages (Fig. 6e), indicating that these two miRNAs may play significant roles by responding to GA during the stone-hardening stage of grape berries. By contrast, the expression of their target genes VvDCAF1, VvCCoAOMT, and VvGAI1 exhibited strong down-regulation in response to GA treatment (Fig. 6e). VvmiRNAs and their target genes above also displayed the opposite expression trends during GA-induced grape seedless-berry development, confirming that these VvmiRNAs negatively modulated their target genes’ expression during this process.
Interestingly, GA treatments significantly down-regulated the expression of VvGAI1, which is a key DELLA protein negative interaction factor in the GA signal pathway (Fig. 6e). Similarly, during grape stone-hardening stage GA treatment also significantly down-regulated the key genes VvCCoAOMT and VvDCFA1, which were involved in lignin biosynthesis and embryo development, (Fig. 6e). These results indicated that GA may repress grape stone hardening and embryo development by inducing the expression of VvmiR31-3p and VvmiR8-5p to negatively regulate the expression levels of VvCCoAOMT and VvDCAF1, as a key molecular mechanism involved in the modulation of GA-induced grape seedless-berry development. However, GA exhibited no effect on the expression levels of VvmiR53-3p and VvGDSL target genes compared with the untreated control, due to the fact that VvGDSL was a gene related to cell-wall development. Meanwhile, we revealed that VvGDSL exhibited the highest expression level at 20 DAF (berry-expansion stage), implying that it may participate in the modulation of the cell-wall expansion development of young berries.
Dynamic accumulation of cleavage products of the four VvmiRNAs’ target genes during GA-induced grape-berry development
Monitoring the accumulation patterns of these four VvmiRNAs’ cleavage products and target genes during GA-induced grape seedless-berry development could contribute to determining the variation of their cleavage roles. Here, the 3′-and 5′-end cleavage product accumulation levels were examined by RLM-RACE and PPM-RACE, respectively. Results showed similar dynamic accumulation modes of both ends-cleavage products during different stages of grape-berry development (Fig. 7), confirming the dynamic variation in cleavage roles of these VvmiRNAs on their target genes during this process. Moreover, the accumulation modes of cleavage products resembled the expression modes of the corresponding VvmiRNAs, indicating that miRNAs may be the main factors affecting their interactions. Interestingly, we found that GA evidently promoted the cleavage roles of VvmiR8-5p, VvmiR31-3p, and VvmiR38-5p on their corresponding targets by obviously up-regulating accumulation levels of the corresponding target cleavage products, while VvmiR53-3p/VvGDSL pair with nearly no change under GA treatment. In particular, the cleavage products of VvmiR31 and VvmiR8 on their corresponding targets VvCCoAOMT and VvDCFA1 were boosted at the highest level at 45 DAF (grape stone-hardening stage), which may derive from the correlation of these two targets’ potential functions with embryo and stone development [28, 30]. Unlike this, those of VvmiR38-5p on VvGAI1 were enhanced drastically by GA at all stages of GA-induced seedless berries used in this work. These results suggested that GA might involved in manipulating grape seedless-berry development primarily by promoting the cleavage role of VvmiR38-5p on VvGAI1 at all stages here. It is inferred that GA might repress grape stone hardening and embryo development through negatively regulating the expression levels of the lignin biosynthesis gene VvCCoAOMT and embryo development related gene VvDCAF1 by inducing the expression of VvmiR31-3p and VvmiR8-5p, as a potential key molecular mechanism involved in the modulation of GA-induced grape seedless-berry development.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.