How many micrornas are there

To investigate if this phenomenon is due to the difference in the cell-specific expression levels of target genes, we performed an analysis of all these targets. This was not the case as well. Experimentally identified microRNA-binding regions form a promising basis for further queries into the basics of the gene expression regulation and lead to uncovering novel disease-causing mechanisms.

At the next stage, the set of microRNA-binding regions was cleaned up to include only these satisfying following criteria: i every position in this microRNA-binding subsequence is supported by evidence from at least two different datasets or two different chimeric sequences and ii the length of at least 10 nt Figure 3A , Supplementary Table 3.

A Validation of the Exp-MiBR by their independent occurrence in two or more datasets, or in two or more chimeric sequences from one dataset. B Exp-MiBRs: distribution of the lengths.

Venn diagram depicting Exp-MiBRs detected in experiments employing three different types of identification techniques. The longest Exp-MiBR of nt was formed by the regions confirmed as microRNA-interacting in 54 different experiments in nine different cell lines. In addition, there were a few Exp-MiBRs located closely to each other. As an example, chromosome 2 contains a cluster of Exp-MiBRs covering an area of approximately 1. To characterize Exp-MiBRs further, we analyzed their tissue specificity.

Interestingly, some Exp-MiBRs were observed in a majority of studied cells, possibly reflecting a housekeeping function of these interactions. Experimental identification of microRNA-binding regions is an important prerequisite for querying into the basics of the gene expression regulation, and for uncovering novel disease-causing mechanisms.

In both studies, the primary goal was to develop and optimize the experimental protocol itself, while identifying miRNA—mRNA interactions in a particular cell line grown under different conditions.

Although these techniques provide a unique window into miRNA targeting, they are not free of limitations, which preclude thorough mapping of entire miRNA—mRNA interactome. Thus, our analysis supports observations of Ragan et al. Ragan et al. Grimson et al. In our study, we attempted painting a holistic picture of human miRNA—mRNA interactome by comparing the entries from experimentally collected datasets describing miRNA-binding activity to the gene expression data.

Interestingly, we found that more than half of mRNA transcripts do not bind to any miRNAs present in the same cellular environment. It was surprising to find that more than half of mRNA transcripts do not bind to any miRNAs present in the same cellular environment. These observations suggest that one can figure out whether some mRNAs may possess such property by analyzing the number of its interactions and the level of its expression: some genes are expressed at a high level but have much fewer interactions than other expressed at same tpm range.

This means that each mRNA differs in their miRNA-binding capacities, and some of them do it in more efficient manner than others. Remarkably, observed miRNA—mRNA sponge-like interactions were cell-line-specific, with very little overlap identified. For each of these mRNAs, amounts of detected interactions were comparable to that of a well-known circular RNA with sponge properties, Cdr1as 74 predicted sites Xu et al. In Huh7. It is peculiar that some Huh7.

About a hundred of such non-interacting miRNAs were present in both studied cell lines. There is a possibility that the natural targets for these microRNAs are either not expressed in studied cellular contexts, or that they have no targets at all. In total, only microRNAs had at least one interaction in each of studied cell lines. For individual miRNAs, levels of their expression have no bearing on amounts of interactions they display, possibly reflecting difference in their functions depending on the cellular context.

As an example, we revealed that, in Huh7. These observations complement previous findings of Mullokandov and colleagues Mullokandov et al. Further studies are required in order to investigate how RNA binding properties of individual miRNAs may change in response to context-dependent regulation by extrinsic or intrinsic factors. At least some Exp-MiBRs are tissue-specific, in agreement with Clark and colleagues, who revealed the differences in the microRNA targetomes across tissues Clark et al.

Previous studies showed four mitochondrial regions with high degree of homology to microRNAs, namely, hsa-miR chrM: 10,—10, , hsa-miR chrM: 13,—13, , hsa-miR chrM: 5,—5, , and hsa-miR chrM: 2,—2, Sripada et al. In both cases, previously identified cognate miRNAs hsa-miR and hsa-miR were among confirmed interactors.

Altogether, these findings support the notion that miRNA—mRNA interactions take place in a variety of cellular compartments, including mitochondria Ni and Leng, For individual miRNAs, levels of their expression have no bearing on amounts of interactions they display, possibly reflecting context depending difference in their functions. On the other hand, there is a set of microRNAs expressed at a very high level and interacting with only a few mRNAs, thus, indeed, regulating expression of their targets in a specific manner.

Notably, microRNAs are capable of switching between these two modes of action, depending on cellular context. The question of the biological significance of these two miRNA groups remains open. It is notable, however, that the presence of miRNA groups, one with a low expression level and a high number of interactions, and one with opposite characteristics, was independently detected in both cell lines profiled. We have also established a collection of reliable microRNA-binding regions that we systematically extracted in course of an analysis of 79 CLIP datasets.

The promise of microRNAs as potential diagnostic mean and therapeutic target got expanded with a number of pathogenic loss-of-function and, recently, gain-of-function mutations described Grigelioniene et al. Hence, our efforts in mapping the human miRNA—mRNA interactome may be aided in untangling molecular underpinnings of hereditary and acquired diseases. All data generated during this study are included in this published article and its supplementary information files.

MS and OP designed the study and carried out the research. AB contributed to the discussion of the results. OP and AB wrote the paper. All authors read and approved the final manuscript. This project has been funded in part by the Laboratory of functional genomics of the Research Centre for Medical Genetics and by the Laboratory of functional genome analysis of the Moscow Institute of Physics and Technology.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. Another concern is that some targets identified under miRNA overexpression in cell culture may not be physiologically relevant. Furthermore, miRNA overexpression analysis is also greatly limited by the lack of high-quality transcriptome-wide profiling data.

Specifically, most existing datasets are of small scale, focusing only on a few miRNAs in any single study, and thus are not ideal for training a general target prediction model. Although it is possible to combine data from multiple small-scale studies, significant heterogeneity among different experiments poses a major concern for accurate target modeling.

Despite the aforementioned challenges, microarray data from miRNA overexpression studies have been proven valuable for target analysis and have been used to train several widely used target prediction models [ 14 , 15 ].

As the first step, we performed a large-scale miRNA overexpression study that is specifically designed to identify transcripts downregulated by 25 individual miRNAs. This comprehensive dataset allowed us to quantitatively re-characterize the previously reported features in the context of target downregulation at the transcriptome level. In this way, our final target prediction model, MirTarget v4.

Comparative analysis indicates that MirTarget has improved performance over other state-of-the-art target prediction tools. It is well established that the binding of a miRNA to its target transcript does not necessarily result in gene expression downregulation.

The overall study design is summarized in Additional file 1 : Figure S1. These 25 miRNAs are listed in Table 1. To control for experimental variations, each miRNA was transfected into cells in duplicate on different days, and RNA-seq library construction and sequencing runs were also performed in duplicate on different days. In total, 1.

The profiling data are presented in Additional file 2 : Table S1. All sequencing data were combined to identify the genes downregulated by miRNA overexpression.

In contrast, transcripts that contain at least 1 seed site but had no expression change are designated as non-target controls.

Specifically, there were 90 targets identified for each miRNA on average, and the target numbers vary considerably among individual miRNAs ranging from 11 to , Table 1. Previous studies have identified several major types of canonical miRNA target sites, including those matching to the 6-mer, 7-mer, or 8-mer miRNA seed sequences Table 2. Sequence conservation analysis suggested that target sites pairing to longer miRNA seeds are more conserved across species and thus are more likely to be bona fide miRNA targets [ 27 ].

This hypothesis on the seed type strength has also been confirmed by analyzing heterogeneous microarray datasets in the context of target downregulation [ 15 , 28 ]. However, further analysis is needed to accurately quantify the contribution of each seed type.

Our newly generated large-scale RNA-seq dataset, encompassing 25 miRNAs assessed under uniform experimental conditions, provided a unique opportunity to quantitatively evaluate the strength of different miRNA seeds on target downregulation.

Specifically, we evaluated the enrichment of each seed type in downregulated target sites as compared to non-target sites. As shown in Table 2 and Fig.

On the other end, seed8A1 is the most selective type, with an enrichment ratio of 6. Among all 7-mer seeds, seed7b and seed7A1 have similar enrichment ratios, both of which are higher than the ratio for seed7a. The impact of miRNA seed types on target downregulation.

Six seed types were evaluated see Table 2 for seed definitions. All 25 miRNAs were included in the analysis. Another type of 8-mer seed, seed8, has the second highest enrichment ratio of 5.

These results suggest that terminal A-U perfect match has little impact on target recognition, as the presence of terminal A in target sites, regardless of its pairing status to the miRNA, is associated with target downregulation. Interestingly, we also observed a dramatically decreased enrichment ratio for seed8 from this miRNA subset. In fact, the seed8 ratio 3. Thus, a perfect terminal match other than A-U is detrimental rather than contributing to target recognition.

Based on the seed analysis, we decided to focus on 3 strongest seed types, including seed8A1, seed7b, and seed7A1, for target prediction modeling. One common concern with miRNA overexpression studies is that it is challenging to locate the exact miRNA binding site within the target transcript.

To alleviate this concern, we identified candidate target sites based on the presence of canonical 7-mer or 8-mer seed sites. In our analysis, we are interested in identifying common features that are characteristic of functional target regulation, including both miRNA binding and subsequent target downregulation.

In a recent target prediction analysis [ 18 ], we have compiled a miRNA target binding dataset derived from multiple public CLIP ligation studies [ 22 , 23 ]. In the present study, the CLIP binding dataset was further combined with new miRNA overexpression data to identify targeting features that are common to both miRNA binding and target suppression.

In this way, target sites and non-target sites, identified from both CLIP and miRNA overexpression studies, were combined and evaluated in subsequent feature analysis. Target and non-target sites in the combined dataset were compared to identify the features that are commonly associated with miRNA target regulation.

These features are listed in Additional file 3 : Table S2. It is well-established that miRNA target sites are evolutionarily conserved [ 7 , 28 ]. In our study, we evaluated target conservation using two complementary approaches. First, we calculated the difference in conservation scores between seed binding positions and flanking positions, as determined by phyloP scores from way multi-genome alignment [ 29 ].

Second, we also determined whether the entire seed site 7-mer or 8-mer is found across multiple species by word search. Both conservation analyses indicated that target sites were very significantly conserved as compared to non-target sites.

In fact, seed conservation was among the most significantly enriched features, whether miRNA overexpression and CLIP binding data were analyzed separately, or in combination. On the other end, non-conserved seed7A1 was the most depleted seed type 9. Besides seed conservation, there were many other features commonly found in both datasets.

One prominent example is related to the GC content of the target site. The depletion of C nucleotide was moderate in both datasets. One possible explanation could be related to RNase T1 used in CLIP studies, which preferentially cuts at G nucleotide, resulting in the depletion of internal G in sequencing reads. Nuclear localized miRISC was found to regulate both transcriptional rates and post-transcriptional levels of mRNA 59 — 61 and associate with euchromatin at gene loci with active transcription However, our understanding of when and how miRNAs exert their functions in the nucleus is still limited.

It has been reported that low molecular weight miRISC can interact with mRNAs within the nucleus and induce nuclear mRNA degradation, although the mechanism behind this is unclear 59 , 61 , Also, a subset of AGO-promoter bound genes was upregulated following senescence, and AGO2 was found to co-immunoprecipitate with euchromatin A more recent study by Miao et al.

The overall role of miRISC in the regulation of chromatin state and structure and transcriptional control remain to be determined, but these current data suggest a transcription factor-like role. It is also possible that miRISC may be involved in the establishment of de novo methylation, and by extension, the compactification of chromatin into nuclear compartments, and mediation of genomic remodeling. Studies have revealed that miRNA-mediated gene regulation is dynamic and helps to buffer gene expression to a steady state.

It is only recently that a more comprehensive understanding of miRNA dynamics has begun to shed light on the highly robust nature of miRNA-mediated gene regulation. Factors that may contribute to the robustness of miRNA-mediated gene regulation include the functionalized compartmentalization and shuttling of miRISC within the cells. The availability and abundancy of miRNAs and their target mRNAs are also contributing factors in determining which genes are regulated.

This in effect acts to spatially enrich mRNA and miRISC concentrations over time and promote efficient regulation of gene expression Figure 2. Proposed model of miRNA localization and function. In the nucleus, miRISC is enriched at sites of active transcription where it can interact with DNA to promote active or inactive chromatin states. It can also interact with nascent mRNA to promote more efficient splicing or alternate splicing profiles.

Cytoplasmic miRISC can diffuse throughout the cytosol or undergo shuttling, most likely via microtubules. Within the cytosol, miRISC can associate with polysomes, inhibit translation initiation, mediate mRNA decay, or promote translational activation. Lastly, vesicular or vesicle-free miRISC can be exocytosed from at least the late endosome into the extracellular milieu to mediate cell-cell communication. P-bodies were identified early on as possible sites involved with miRNA-mediated suppressive activity It is not clear why a microtubule stabilization would lead to an increase in P-body count.

Typical microtubule instability may in effect dilute microtubule-associated miRISC by preventing enrichment at sites of nucleation. However, the degree to which P-bodies are necessary for efficient miRNA-mediated suppression is uncertain as mRNA degradation machinery exists diffusely throughout the cytoplasm, and at other subcellular locations, and can confer RNAi activities even in the absence of P-bodies Polysomes, which are complexes of mRNA with multiple translating ribosomes, are generally found freely within the cytoplasm or bound to the cytoskeleton or membranous subcellular organelles, such as rER.

Together these data elude to a key aspect of miRNA functioning; there is a crucial equilibrium between MRE bound and unbound states and this equilibrium is important for the spatiotemporal dynamics of miRISC.

This allows for miRNA to respond quickly to changes in subcellular environments and dynamically regulate many target mRNAs. Moreover, approximately half of the expressed miRNAs are cell type enriched, one quarter are broadly expressed, and the remaining had low-level expression regardless of cell type.

These data help fortify the general roles that miRNA play within cells. In both cases, higher-order effects on gene regulatory networks can propagate 88 , such as regulation of transcription factor expression leading to changes in transcriptional profiles.

Intriguingly, miRNA do not solely, or maybe even predominantly, function as target-specific regulators but may play key roles in the post-transcriptional reduction of expression noise 89 , In this way miRNA promote stable gene expression by buffering out stochastic fluctuations in transcription. A common indicator of expression noise control is high mRNA to protein ratios, rendering reduced translational rates resistant to random fluctuations in mRNA concentration 91 , i.

Accordingly, the strongest predictor of protein level for any given gene is the rate of translation, followed by mRNA levels This has the effect of diluting cellular miRNAs amongst many potential targets such that only a small proportion of each target mRNA is bound to a cognate miRNA at any given time Synergistic effects of miRNAs have been shown to be important for biological processes, such as neurogenesis and human embryonic stem cell pluripotency and differentiation For example, cyclin-dependent kinase inhibitor 1A CDKN1A , a tumor suppressor that is downregulated in multiple cancers , is targeted by at least 28 miRNAs and many of which are upregulated together in cancers where CDKN1A has been implicated In a study by Wang et al.

Both cell stress induced by heat shock and translation inhibition following treatment with hippuristanol or cycloheximide induced translocation of miRISC from the nucleus and cytoplasm to transient, cytoplasmic SG 72 , SG are known to act as intermediate storage of messenger ribonucleoproteins mRNPs that have stalled during translation or following viral infection SG components can also return to the cytoplasm, exchange with P-bodies, or be digested by lysosomes 81 , Numerous studies have demonstrated that miRNAs can be released into extracellular fluids.

Extracellular miRNAs can be used as biomarkers for a variety of diseases. These studies have been extensively reviewed — and therefore will not be discussed here. In this regard, miRNAs have hormone-like activities. Contrary to cellular RNA species, extracellular miRNAs are highly stable, resisting degradation at room temperature for up to 4 days and in deleterious conditions such as boiling, multiple freeze-thaw cycles, and high or low pH , Two populations of extracellular miRNAs exist in biological fluids.

One can be found in vesicles such as exosomes, microvesicles, and apoptotic bodies , while the other is associated with proteins, especially AGO2 , There have been some discrepancies on the relative abundancies of these two populations. The presence of miRNAs in vesicles or with accompanying proteins is generally thought to protect extracellular miRNAs and increase their stability in the extracellular milieu Although some extracellular miRNAs are regarded as by-products of cellular activities, such as cell injury or death , increasing evidence suggests that the release of extracellular miRNAs is a regulated process.

It has been shown that the secretion of exosomal miRNAs is mediated by a ceramide-dependent pathway and the secreted miRNAs exert growth regulatory effects in target cells Recently, it was demonstrated that atheroprotective laminar shear stress induced the release of vesicle-free miRp and other miRNAs, as well as AGO2, from endothelial cells by activating vesicle-associated membrane protein 3 VAMP3 and synaptosomal-associated protein 23 SNAP23 This study also showed that miRNAs secreted from endothelium can regulate the activity of smooth muscle cells IL4-activated macrophages were found to secrete exosomes carrying oncogenic miRNAs to promote invasiveness of breast cancer cells On the other hand, DHA, which has anticancer and anti-angiogenic activities, induced the secretion of miRNA-containing exosomes that exert inhibitory effects on tumor angiogenesis Many studies have also demonstrated that extracellular miRNAs can exert biological functions in recipient cells to regulate their activity, thereby acting as intercellular signaling molecules.

For example, exosome mediated transfer of miR from metastatic breast cancer cells to endothelial cells directly targeted a tight junction protein, zonula occludens 1 ZO-1 , and this led to the destruction of the barrier function of endothelium and promoted metastasis Moreover, exosomes from umbilical cord blood were found to be enriched in miRp, which promoted the proliferation and migration of fibroblasts, and induced the angiogenic activities of endothelial cells, leading to accelerated wound healing Extracellular miRNAs have also been reported to bind to Toll-like receptors , activate downstream signaling events, and eventually lead to biological responses, such as tumor growth and metastasis , and neurodegeneration Thus, miRNAs may act as chemical messengers to regulate cell-cell communications.

The mechanisms of extracellular miRNA uptake are not well understood. It has been proposed that vesicle-associated extracellular miRNAs may enter cells by endocytosis, phagocytosis, or direct fusion with the plasma membranes, while vesicle-free secreted miRNAs may be taken up by specific receptors on the cell surface Indeed, several studies have shown that miRNAs enter recipient cells by endocytosis and micropinocytosis , This process has been reported to be dependent on clathrin, but not on caveolae or lipid rafts in PC12 cells However, another study conducted in AP cells showed that endocytosis of exosomal miRNAs is mediated by caveolae- and lipid raft-dependent pathways While these studies suggest that extracellular miRNAs can interact with recipient cells via multiple mechanisms, the factors that determine the specificity of such interactions need to be investigated.

Since the discovery of miRNAs in the earlier s, tremendous progress has been made on how miRNAs are produced within cells, how they exert regulatory effects on gene expression, and how they are involved in various physiological and pathological events.

It is now clear that miRNAs are powerful gene regulators, and that they not only help control mRNA stability and translation but are also involved in transcription. However, our understanding of when and how miRNAs can exert regulatory effects on transcription is limited.

Similarly, the conditions under which miRNAs elicit translational activation need to be further explored. In addition, careful analysis and consideration of experimental techniques and model systems should be employed when attempting to generalize miRNA capabilities.

The assaying of miRNA activity within a test tube may not be recapitulated within the cellular environment and thus should be viewed with caution. Many studies have been conducted in vitro by transfecting pre-miRNAs or mature mRNA mimics into immortalized and cancer cell lines.

The extent to which findings from such studies reflect the endogenous miRNA functions in vivo requires further study. Recent studies have shed light on the dynamic nature of miRNA actions and further revealed the complexity of miRNA-mediated gene regulation. Cytosolic miRISC components shuttle between different compartments. Recent advances in single molecule imaging will greatly impact the field, as has already begun. Investigation of large scale, global miRNA interactomes will also propel the field forward, allowing powerful mathematical models to be applied to highly complex regulatory networks.

Many studies have shown that extracellular miRNAs are functionally active in recipient cells. Some miRNAs can even interact with cell surface receptors, such as Toll-like receptors. Therefore, miRNAs have hormone-like activities. The remaining nucleotides corresponding to the template were added using Klenow DNA polymerase resulting in the desired dsDNA template. Whole NB images in this manuscript were further manually adjusted in terms of contrast and brightness.

For all three groups, we computed a high-throughput based exclusion rate based on the NGS data sets mentioned above denoted as h 1 , h 2 , h 3 , respectively as well as a low throughput validation rate l 1 , l 2 , l 3 , respectively.

We further assumed that high-throughput exclusion and low throughput validation are independent of each other. We collected 28 human small RNA sequencing samples containing Following stringent quality filtering, samples were excluded because of wrong annotations or low data quality. For mapping of those billions of small RNA sequencing reads, only 4. From our training set, miR—5p was found to have the highest number 16 of isomiRs detected.

MiRNA isomiR variants were not further validated by northern blotting due to the lack of specificity of RNA probes for the nucleotide changes at 5p or 3p ends.

To select miRNAs for experimental validation, we employed a recently developed algorithm that acknowledges key criteria characteristic for miRNAs. The complete listing of criteria to define high-confident miRNAs is given in Backes et al. After excluding precursors that did not meet the criteria of high-confident miRNAs, we employed an experimental validation step to further identify false-positive miRNAs. Altogether, mature miRNAs originating from precursor molecules were tested in this validation step.

Out of the three miRNA candidates that were not present in V21 yet, one hsa-miR has been added to the high-confidence set and was successfully validated by us, while the other two hsa-miR, hsa-miR—5p were added to the low-confidence set and did not pass our validation. In addition to miR—3p, miR—5p was also positively validated by us and should be included in miRBase. To search for endogenous miRNAs with expression levels high enough to be identifiable by northern blotting, we analyzed 12 human cell lines by microarray.

All other miRNAs including —3p, a-3p, —5p, 16—5p, 19b-3p, 20a-5p, 23a-3p and 23b-3p yielded signals in the expected size range for the mature form as shown in Figure 2. While the miRNAs 16—5p, 19b-3p, 20a-5p, 23a-3p and 23b-3p were detected in all 12 cell lines, miR—3p was found in 5 cell lines only, miRa-3p in 10 cell lines and miR—5p in 5 cell lines Figure 2B.

Out of the 11 miRNAs, the miRNAs —3p, a-3p, —5p, 23a-3p and 23b-3p were also identified by the exogenous expression analysis. Northern blots of endogenous miRNAs. The endogenous mature forms are shown for the miRNAs indicated on the left side of the figure.

B The number of mature and premature forms of the endogenous miRNA expressed in the 12 cells lines as indicated in Figure 2A.

Taken together, from our exogenous and endogenous experiments, we successfully validated 51 of 54 high-confidence miRNAs Details are provided in Supplementary Table S2.

All of these signals for mature miRNAs approximately matched the expected size range. Most of them were not discovered in the endogenous controls. We defined miRNAs as positive when signals for both a precursor and a mature form were detected and are stronger for the overexpression compared to control RNA lysates.

However, in the majority of cases, the size of the precursor did not correspond to the size indicated for the respective stem-loop forms by miRBase. A NB for hsa-miR—5p from high-confidence set A showing distinct bands for its precursor p and mature m form. B Hybridization against hsa-miRa low-confidence set B detects two small RNA fragments with similar signal intensities for the control.

C Probing for hsa-miR—5p low-confidence set B did not result in any specific bands. Ethidium bromide staining of RNA gels was used as a loading control.

Considering all cases, we found six recombinants and two endogenous miRNAs, as already described above, with signals neither in the size range of the precursor nor in the range of the mature forms. The only high-confident miRNAs that were not confirmed by our northern blot analysis were miRa-5p and miRa-3p, both of which showed signals for potential precursor forms, albeit in different sizes, but not for any of both mature forms and miRb-3p where only a premature form could be detected.

These results suggest that even in the high-confidence set a certain number of false-positive miRNAs exist. For the low-confident miRNAs, exogenous and endogenous expression analysis failed to confirm 22 and 3 miRNAs, respectively, that did not show signals for the mature and precursor form.

About 19 of these had at least some signals designated as potential precursor form or unclear signals details provided in Table 1. The remaining 10 miRNAs were confirmed by the identification of both the mature and the precursor forms. Although our analysis is largely consistent with the miRBase qualification of many low-confident miRNAs, our data also indicate a considerable number of false-negative miRNAs among the low-confident set in miRBase.

As for variances in signal intensities, 49 miRNAs showed stronger signals for the precursor than for the mature form indicating a reduced processing efficiency Table 1. As abovementioned, miRa-5p and 24 other miRNAs showed only signals for the precursor but not for the mature form.

Overall, the 5p-forms appear to be more efficiently processed into mature miRNAs than the 3p-forms. Out of the 14 miRNAs, which gave rise to one mature form only according to miRBase V22, we confirmed 4 miRNAs, 2 of which, including mirb and mir, showing stronger signals for their mature form than for the precursor Table 1.

In detail, all 33 miRNAs that were taken from miRBase versions 1 through 7 have been experimentally validated. Here, for mir, we only detected a very faint signal in precursor size. Only 3 out of 20 miRNAs taken form version 17 onwards have been verified. Consistent with the doubts raised in previous studies, these results further question the nature of miRNAs with high ID numbers recently deposited in miRBase that are mostly identified by NGS studies only.

At the time of the study setup miRBase V22 was not yet released. While the proportions changed substantially, our estimate remained stable and decreased only by 49 miRNAs 2. Here, we compared the targetomes of miRNAs verified by our above applied criteria versus the targetomes of miRNAs not confirmed in our analysis. To this end, we considered only miRNA targets that have been classified as verified targets by miRTarBase strong evidence. In the set of the not-validated miRNAs, each had a decreased number of 65 median target genes.

For example, both mature forms of hsa-mir that we could not validate in this study have been associated with complex target gene sets of for the 5p and targets for the 3p form. Altogether, 16 miRNAs not-validated by our approach have recently been associated with experimentally validated targets.

Although there are no studies that systematically analyzed endogenous miRNAs by NB, there are data sets listing endogenous miRNAs according to their signal intensities e. The overall low endogenous expression of miRNAs necessitates an exogenous expression system, which also allows monitoring the processing of the precursor into the mature form.

We chose HEK T cell culture as expression system that stems from a kidney of a healthy aborted fetus and that allows high transfection efficiency and high expression rates for the pSG5 vector 36 , The use of an exogenous expression system also allowed to systematically compare the processing of the 5p and the 3p form for each of the analyzed miRNAs. Depending on the tissue, the cell, and the applied condition, both mature forms have been reported as functional In case of miRNAs for which no mature but a precursor form can be detected by our NB procedure, it is conceivable that the amount of miRNA was too low to yield distinct signals.

Overall, our data indicated that the 5p-forms are more frequently processed into mature miRNAs than the 3p-forms. This observation is consistent with previous publications that analyzed miRNA strand selection with regards to thermodynamic stability 39 , Although our validation pipeline was not optimized for the detection of splicing-derived miRNAs, we obtained positive NB signals for mirtron precursor mir and its mature 3p form but not for mir and mir To the best of our knowledge, the only studies that also identified mirtrons by NB were by Schamberger et al.

Agotrons, another potential exception for our validation system, do not appear as miRNA like signals or even not as a premature like form on northern blots as they differ in size up to nt and irrespective of their association with Ago proteins.

For the specific validation of exogenous agotrons via northern blotting they should be coexpressed along with Argonaute proteins for stabilizing effects In contrast to the detection of single-nucleotide polymorphisms, isomiR detection is not readily possible by quantitative Real Time-PCR or northern blotting.

These methods do not allow to discriminate between miRNAs with length variants of up to 5 nt at their 5p or 3p ends. Almost all of the studies describing isomiR variants use enzyme-based methods making the bias-free validation of miRNA isoforms that were detected by NGS a major challenge 44— For example, we failed to validate mir exogenously and miR and miR endogenously, although others have already provided functional evidence for its derived miRNAs 47—53 and many others.

As addressed above, one has to differentiate between miRNAs that show a signal for the precursor miRNA only but no processing, a signal only for the processed form, and cases without any or with only faint signals.

While the latter cases likely do not represent true miRNAs, even the lack of identification of the two forms does not necessarily disprove that a tested sequence represents a true miRNA. Tissue specific factors that are required for processing in a given cell may not be present in the HEK cells used Finally, signal intensities also depend on the amount of miRNA expressed and processed and the number of radiolabeled nucleotides in probes.

To minimize the influence of biased labeling, we used radiolabeled GTP for all oligonucleotide probes that contained only two or less UTPs. A failure of processing a miRNA precursor in the according wild-type cells can likewise be linked to a specific cell type.

Complementary to experimental settings that manipulate miRNA expression of a specific cell type, endogenous miRNA expression can be analyzed.