High-throughput sequencing of short RNAs in tomato
High-throughput sequencing has been used to study miRNAs at the whole genome level in several model plant species, including Arabidopsis , rice , and wheat . The successful application of high-throughput sequencing technology to systemically identify plant miRNAs in recent years has greatly advanced our knowledge on the functions of miRNAs in plants.
The composition of small RNAs often reflects roles of different kinds of small RNAs in a specific tissue or species. Several plant species, such as Pinus cordata, Populus balsamifera and grapevine were shown to contain substantially more 21-nt than 24-nt sRNAs [37, 39, 65]. However, we observed an unusually high level of 24-nt sRNAs compared with the 21-nt class (Figure 1; Figure 2), as in several other plant miRNAs studies [23, 36, 37, 66]. The number of 24-nt small RNAs was almost three times that of the 21-nt class, such a high percentage of 24-nt small RNAs may reflect the complexity of the tomato genome because 24-nt siRNAs are known to be involved in heterochromatin modification, especially for genomes with high content of repetitive sequences [67, 68].
Plant MIR genes are possible to arise from gene duplication events which then evolved by random mutations into short, imperfectly paired hairpins [46, 69]. Those non-conserved miRNAs are believed to be evolutionarily recent and generally represented by single copy MIR genes . Recently, many non-conserved miRNAs were reported in several species [36, 37, 64, 66]. Some of these non-conserved miRNAs are not found in other species/families and several non-conserved miRNAs are found in different phylogenetic families. We also found nine non-conserved miRNAs in our database, although most of them were present at a very low level. However, the low number of sequence reads does not rule out that they could be expressed at a high level in specific cells. Few non-conserved miRNAs were expressed at a relatively high level, suggesting that they may represent an intermediate status between the deeply conserved and the less conserved miRNAs.
New species-specific miRNAs are considered to be young miRNAs that have evolved recently, and are often expressed at a lower level than conserved miRNAs, as was reported for Arabidopsis and wheat [23, 36, 38]. This observation is true for many new tomato miRNAs identified here. However, few new miRNAs were expressed at a high level in a tissue-specific manner (miRZ6). In some cases we observed inconsistency between the level of miRNAs identified by Solexa sequencing and experimental verification. It is possible that during library generation or sequencing some bias could occur for certain sequences in some samples.
In contrast to the 454 pyrosequencing sequencing in tomato previously , more reads were obtained in each of our three libraries, and miRNAs at different fruit ripening stages were additionally analyzed in our study. In our database, most conserved miRNAs have been detected, including 33 conserved and non-conserved miRNA families, which cover all the published miRNAs in the two databases for tomato miRNAs (http://www.mirbase.org/, http://ted.bti.cornell.edu/cgi-bin/TFGD/sRNA/miRNA.cgi).
Dalmay and colleagues studied the correlation between the short RNAs and fleshy fruit development using deep sequencing technology . However, mostly sRNAs were analyzed, which is different from our miRNAs study. But they reported several miRNAs, and validated one novel miRNA related to the glutamate accumulation which contributes to the tomato fruit taste.
Targets of known and novel miRNAs in tomato fruit
Target prediction is necessary for assessing miRNAs' putative functions. Currently, the most efficient tool available for this is the bioinformatics approach facilitated by the high degree of homology between miRNA and its target sequences in plants . Our analysis reveals that most of the predicted targets in tomato have a conserved function among a variety of plant species . Consistent with previous reports, the largest of these targets in tomato fruit were plant-specific transcription factors, such as AP2, NAC, SBP and the ARF family . The second largest targets encoded a range of different proteins implicated in a variety of metabolic processes, such as ATP sulfurylase, Pectate Lyase, endo-1, 4-beta-glucanase, Laccase. In addition, functions of several targets are largely unknown.
Although many conserved miRNA targets were predicted, only a limited number of miRNA targets were identified experimentally [31, 39]. Several miRNAs slice their targets and can be validated by RACE. Alternatively, other miRNAs inhibit target gene expression through translational arrest, such as miR156 and miR172, which have been shown to regulate their target genes (SBPs and AP2) predominantly by inhibiting their translation [73–76]. Indeed, genetic and biochemical evidence is accumulated that plant miRNA-guided silencing has a widespread translational inhibitory component [2, 77].
The novel tomato miRNAs target different genes with a wide variety of predicted functions. It may be worthy to note that those miRZ6 targets six members of disease resistance protein (CC-NBS-LRR class), which are likely to be involved in host-pathogen interactions. Surprisingly, one target of miRZ7 is beta-galactosidase which is a crucial enzyme for fruit softening, which is beneficial for clarify the cooperating roles in freshly fruit ripening and senescence. The target of another member in miRZ7 family is starch synthase which participates in regulating starch biosynthesis (Additional file 1) .
Fruit ripening and softening related miRNAs
Functional analysis has been carried out only for role of a few tomato miRNAs morphological development [28–30]. Conserved and non-conserved miRNAs regulate genes involved in fleshy fruit development were reported previously . A target gene of miR156 belong to squamosa-promoter binding protein (SBP) family called CNR was validated, which plays pivotal roles in fruit ripening . Over-expression of miR156 can cause the fruit red colour slightly lighter than the wild type, which was thought that the expression of miR156 and CNR overlaps only partially and the function of the miRNA is to suppress CNR expression in specific cell types. CNR mRNA is up-regulated in fruit at the breaker stage , coincidently, our sequence results showed a negative correlation as expected. Two important targets of miR396 are endo-1, 4-beta-glucanase and MADS-box protein, both of which are important regulators in fruit ripening and softening [78, 79]. Our sequencing results showed that the abundance of miR396 peaked at the breaker stage and decreased sharply in red ripening stage suggesting its possible function in fruit ripening. A target of miR482 is pectate lyase which is an important enzyme in fruit softening [80, 81]. Meanwhile, a novel miRNA, whose target is beta-galactosidase, which is also a key softening related enzyme [82, 83]. The expression levels of the two miRNAs were also coinciding with the assumption including the softening ripening stage (Figure 8).
Intriguingly, an in-depth study of the tomato transcriptome and proteome unraveled the regulation mechanisms during fruit ripening stages. The transcriptional regulation is mostly responsible for global reductions in plastid gene expression prior to or at the onset of fruit formation whereas post-transcriptional events become predominant at the breaker stage [84, 85]. These post-transcriptional events did not lead to significant reductions in mRNA levels, although mRNA stability was reported to be modestly affected for some genes . miRNAs and their targets have been shown to regulate in translational inhibition way in a fairly large fraction .
Ethylene biosynthesis and signal transduction related miRNAs in tomato
Recent researchers have witnessed the great progress in the area of identification of interaction between miRNAs pathways and phytohormone responses, which improves our understanding of mechanism of plant development controlled by miRNAs and hormone action to a large extent. One of the most important phytohormone is ethylene which plays vital roles during all stages of the plant life cycle, functioning through seed germination to ripening and various abiotic stress conditions . MiR159 and miR394 were reported to be associated with ethylene in rice . Most targets of miR159 are MYB, and a novel target not related to MYB in tomato recently . Another important target of miR159 is 1-aminocyclopropane-1-carboxylate synthase which plays vital roles in ethylene biosynthesis . As a typical climacteric plant, abundance ethylene was synthesized at the breaker stage approximately, and the expression level of miR159 was suppressed which match our results well. MiR394 and miR414 also target F-box family proteins which are important participants in the signal transduction pathways of different plant hormones [88, 89]. The stability of the ethylene signaling regulators EIN2 and EIN3 are modulated by the F-box proteins ETP1/2and F-box proteins EBF1/2 respectively. Ethylene Insensitive 2 (EIN2), a target of miR828, is the key positive regulator of ethylene signal transduction and other hormones such as abscisic acid, auxin, cytokinin and jasmonate and thus may represent a point of crosstalk between multiple hormone signaling pathways [90–92]. CTR1 is a negative regulator of ethylene response that likely interacts directly with receptor molecules to form a signaling complex [93, 94]. As with the ethylene receptors, all tissues evaluated express CTR1 genes and their mRNAs are differentially accumulated depending on tissue. LeCTR1 induction is associated with tissues at stages of development associated with increased ethylene, including fruit ripening and is a validated target of miR1917 . Exogenous ethylene can influence the miRNAs expression patterns, most of which shown negative correlation to their targets well except for miR828, was not affected by exogenous ethylene treatment (Figure 9). The most likely reason was that miR828 affected this target slightly and we cannot catch the discrepancy.