Current experimental strategies for intracellular target identification of microRNA
Intracellular target identification of microRNA (miRNA), which is essential for understanding miRNA-involved cellular processes, is currently the most challenging task in miRNA-related studies. Although bioinformatic methods have been developed as the most efficient strategy for miRNA target identification, high-throughput experimental strategies are still highly demanded. In this review paper, we summarize and compare current experimental strategies for miRNA target identification, including gene expression profiling, immunoprecipitation and pull-down methods. Gene expression profiling methods mainly rely on the measurement of target gene expression through overexpression or inhibition of specific miRNA, which are indirect strategies to unveil miRNA targets. Immunoprecipitation methods use specific antibody to isolate RISC and bound mRNAs, followed by analysis with high-throughput techniques and bioinformatics to reveal miRNA-mRNA interactions. Pull-down methods use tagged miRNA mimics as probes to isolate associated target genes through affinity purification, which directly indicate miRNA-mRNA interactions after analysis of isolated target genes. Each method has its own advantages and limitations, which will be summarized and discussed in details. Overall, this review paper aims to provide a brief outline of recent achievements at experimental strategies for miRNA target identification. With the further development or improvement, we envision these experimental strategies will ultimately contribute a lot to the research on miRNA and miRNA-targeted biomedicine.
KeywordsmicroRNA Target identification Experimental strategies
Crosslinking, immunoprecipitation and sequencing of hybrids
Crosslinking and immunoprecipitation
miRNA crosslinking and immunoprecipitation
miRNA target RNA affinity purification
RNA-induced silencing complex
Stable isotope labeling by amino acids in cell culture
MicroRNAs (miRNAs) are endogenous small-noncoding RNAs with the length of ~ 22 nucleotides, which can regulate gene expression at the post-transcriptional level . Since the first discovery of lin-4 [2, 3] and let-7 [4, 5] in C.elegans, more than 2500 miRNAs have been found and identified in human beings . Meanwhile, a single miRNA could target multiple genes and over one third of human genes were predicted as conserved miRNA targets , suggesting miRNAs participate in almost all cellular processes through regulating their target genes. Recent evidences also revealed miRNAs were involved in not only normal physiological processes but also pathologies [8, 9]. The abnormal expression or function of miRNAs were closely related with diverse human diseases, such as cancers. MiRNAs are thus emerging as novel endogenous bio-targets for diagnostics and therapeutic treatments [10, 11]. Understanding miRNA-involved cellular processes, including a clear picture of regulatory networks of intracellular miRNAs, is therefore essential and critical for miRNA-targeted biomedicine [12, 13], which still represents a big challenge in miRNA-related investigations. It is worth noting that phase I clinical trials of miR-34 in cancer treatment were recently terminated due to severe immune-reactions with unknown reasons , which is mainly due to the lack of information about miR-34 regulatory networks and further highlights the importance of miRNA target identification before proceeding to therapeutics.
Current experimental strategies for isolation and identification of miRNA targets mainly rely on three methods (Fig. 1) [23, 24]. (1) Gene expression profiling methods, which indirectly indicate miRNA targets through measuring gene expression changes after overexpression or inhibition of specific miRNA. (2) Immunoprecipitation methods, which isolate RISC using specific antibody to capture miRNA targets in RISC for further analysis. (3) Pull-down methods, which use chemical tags-labeled miRNA mimics as probes to enrich miRNA-associated target genes through affinity purification for further analysis. In this review paper, we introduce the general principles and applications of current experimental strategies for miRNA target identification. Comparison and discussion on the advantages and limitations of these strategies will also be presented.
Experimental strategies for miRNA target identification
Summarization and comparison of current experimental strategies for intracellular target identification of miRNAs
Experimental strategies and references
Gene expression profiling through overexpression or inhibition of specific miRNA
•A straightforward method to identify direct targets for miRNAs
•Easy to adopt
•Lack of 3’-UTR libraries
•Unable to identify non-canonical targets
•Simultaneous identification of a subset of genes
•Difficult to distinguish direct and indirect miRNA targets
•No information about miRNA-mRNA interaction
•High false-positive and false-negative results
•Easy quantification of protein production through metabolic labeling
Ribosome profiling 
•Measuring mRNA translation
Immunoprecipitation of RISC with specific antibody
•Avoid false-positive targets outside RISC
•Limited by the specificity of antibody
•Non-specific to miRNA
•Increase in capture efficiency due to photo-crosslinking
•Clear information about miRNA-mRNA interaction due to ligation of miRNA-mRNA in RISC
•Limited by the specificity of antibody and crosslinking efficiency
Pull-down with labeled miRNA mimics
•Specific to miRNA
•Side effect of 3’-biotinylation on miRNA function
MiRNA crosslinking and immunoprecipitation (miR-CLIP) 
•Avoid side effect of biotinylation on miRNA function
•Specific to miRNA
•Limited by the specificity of antibody and crosslinking efficiency
•Not universal for other miRNAs
Photo-clickable miRNA 
•A universal strategy for different miRNAs
•Specific to miRNA
•Possible dissociation between photo-clickable miRNA and target mRNAs during pull-down
Gene expression profiling
Since miRNAs regulate gene expression through interaction with the 3’-UTR of target mRNAs, screening possible miRNA targets using cellular reporter systems bearing 3’-UTR of different mRNAs is also a straightforward way to identify miRNA targets. The reporter systems were constructed by transfection of luciferase reporter genes containing 3’-UTR of human genes into cells, followed by introducing miRNA of interest into these cells to modulate luciferase expression (Fig. 2). MiRNA targets could then be indirectly indicated through measuring luciferase signals. Using this strategy, Mangone et al. engineered 275 luciferase reporter genes with human 3’-UTRs and two cancer relevant miRNAs, let-7c and miR-10b, were chosen to screen possible targets . A large number of novel genes were then identified for these miRNAs, among which only 32% of them were consistent with bioinformatic predictions. Similarly, 139 luciferase reporter genes with predicted human 3’-UTRs were also constructed by Penalva et al. for screening possible targets for liver-specific miR-122, showing the prediction accuracy was ~ 37% . This method is sensitive and can identify direct targets for miRNAs, but it is limited by the high costs, shortage of 3’-UTR libraries and low throughput.
To realize high-throughput identification, indirect strategies based on quantification of global gene expression changes following miRNA overexpression or inhibition were developed (Fig. 2). After collection of possible targets through detecting gene expression changes, miRNA-mRNA interactions could be further indicated by bioinformatics. For instance, Johnson et al. overexpressed brain-specific miR-124 or muscle-specific miR-1in HeLa cells and analyzed the gene expression profiles through microarray, showing down-regulation of genes with special expression patterns in brain or muscle and the 3’-UTRs of these mRNAs tended to pair to the 5’-end of miRNAs . Similarly, expression of mRNAs was profiled by microarray analysis after overexpression or inhibition of cartilage-specific miR-140 in murine C3H10T1/2 fibroblast cells, resulted in 49 genes were simultaneously detected in mRNA samples from cells overexpressed or repressed with miR-140 . With technique stable isotope labeling by amino acids in cell culture (SILAC), protein expression changes post modulation in miRNA expression could be read out. SILAC was then used to indicate targets for several miRNAs through overexpressing them in different cells, showing hundreds of proteins were modulated by these miRNAs [29, 30]. Additionally, through measuring mRNA translation rates with ribosome profiling, Bartel et al. compared intracellular protein levels and mRNA levels after overexpression of miRNA in HeLa cells and showed mammalian miRNAs regulate gene expression mainly through mRNA degradation . These methods are quantitative and high throughput, while it is unable to distinguish the direct or indirect targets of miRNAs, since primary and secondary targets are both yielded. Meanwhile, these methods cannot provide detailed information about miRNA-mRNA interactions. Additional bioinformatic methods are thereby always needed for further analysis.
Provided that some nucleic acids and amino acids are photo-sensitive and could be crosslinked upon 254 nm irradiation, capture efficiency could thus be improved through photo-crosslinking of AGO with bound RNAs. Crosslinking and immunoprecipitation (CLIP) method that uses ultraviolet (UV) light to covalently conjugate protein-RNA was then developed (Fig. 3). After immunoprecipitation with a specific AGO antibody, miRNAs, their targets and AGO protein are precipitated together for further sequencing analysis. For example, Darnell et al. used CLIP to map interaction networks for miR-124, which simultaneously generated AGO-miRNA and AGO-mRNA data sets through high-throughput sequencing . To further increase the capture efficiency, Tuschl et al. developed photoactivatable ribonucleoside-enhanced CLIP (PAR-CLIP) method, which incorporated photo-reactive 4-thiouridine into RNAs to more efficiently crosslink to nearby biomolecules upon UV irradiation . While, due to the indirect isolation and identification, additional bioinformatic analysis are still needed to reveal miRNA-mRNA interactions from the CLIP data [38, 39]. To address this issue, crosslinking, immunoprecipitation and sequencing of hybrids (CLASH) method, which is similar to CLIP but ligates miRNA and target mRNA in RISC together for further sequencing analysis, was developed (Fig. 3). Using this method, Tollervey et al. obtained data sets of many miRNA-mRNA conjugates and revealed frequent non-canonical bindings for human miRNAs [19, 40]. Even though CLASH could reveal direct interaction between miRNA and target mRNAs, the efficiency of this method is relatively low. Moreover, immunoprecipitation strategies are not miRNA specific. Further improvements of these immunoprecipitation methods are still highly demanded before they can be used to map global miRNA-mRNA networks.
In addition to strategies based on using biotinyated miRNAs as probes, Tsai et al. developed an alternative strategy, which used digoxigenin (DIG)-labeled pre-miRNA as probe and was termed as labeled miRNA pull-down (LAMP) assay system . The DIG-labeled pre-miRNA probe was incubated with cell extracts, leading to generation of DIG-labeled mature miRNA probe upon cleavage by Dicer and further binding of the probe with target genes. Through immunoprecipitation by anti-DIG antiserum, DIG-labeled miRNA and bound mRNA complex were obtained for further analysis. With this strategy, they found a novel target gene hand2 for zebrafish miR-1. While, the effect of DIG on miRNA function and the possibility of introduction of DIG-labeled miRNA probes into living cells for miRNA target identification remain unknown.
Conclusion and perspective
In this review paper, we summarize and compare current experimental strategies for intracellular target identification of miRNAs. Each strategy has its inherent advantages and limitations, which require further refinements of these methods before proceeding to globally mapping miRNA regulatory networks. Accuracy and efficiency are the two major factors that need to be considered during development and improvement of experimental strategies. In comparison with gene expression profiling methods, accuracy of immunoprecipitation methods is greatly improved, since false-positive target genes outside RISC are excluded. However, relying on specific antibody to isolate target genes further decreases the efficiency of target isolation and identification. Meanwhile, due to the indirect isolation and identification, bioinformatics are always needed to reveal miRNA-mRNA interactions. Currently the most promising strategy is the pull-down method, since it uses tagged miRNAs as probes to directly isolate miRNA-associated targets. The biocompatibility of chemical tags toward miRNA modification is then critical for miRNA target identification. Recent results revealed 3’-biotinylation greatly hampered the association of miRNA with their targets in RISC [22, 45], indicating direct biotinylation is not suitable for miRNA target identification. To address this issue, we recently developed photo-clickable miRNA that pre-tagged miRNA with tetrazole groups on 3’-miRNAs without affecting their function, which allowed further attachment of affinity tags onto miRNA-mRNA complexes post their binding . Moreover, combination of other bio-orthogonal reactions, such as click reaction and tetrazine reaction , should further improve the accuracy and efficiency of miRNA target identification through miRNA probes bearing bio-orthogonal groups and should allow simultaneous target identification for different miRNAs in same biological environment. With the development and improvement of experimental strategies for miRNA target identification, a clear picture of miRNA regulatory networks inside cells will be drawn in the future, which will ultimately lead to huge progresses in therapeutic treatments with miRNAs as targets.
National Natural Science Foundation of China (21877058, 21672103, 21572102, 21372115).
Availability of data and materials
JL and YZ wrote and revised the manuscript. Both authors read and approved the final manuscript.
Ethics approval and consent to participate
Consent for publication
The authors declare that they have no competing financial interest.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
- 6.miRBase: http://www.mirbase.org /.
- 15.Singh NK. miRNAs target databases: developmental methods and target identification techniques with functional annotations. Cell Mol Life Sci. 2017;74(12):2239–61.Google Scholar
- 39.Uhl M, Houwaart T, Corrado G, Wright PR, Backofen R. Computational analysis of CLIP-seq data. Methods. 2017;118-119:60–72.Google Scholar
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. 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.