doi: 10.1186/gb-2014-15-4-r57.
Evidence for the biogenesis of more than 1,000 novel human microRNAs
- PMID:24708865
- PMCID: PMC4054668
- DOI: 10.1186/gb-2014-15-4-r57
Item in Clipboard
Evidence for the biogenesis of more than 1,000 novel human microRNAs
Marc R Friedländer et al. Genome Biol..
Display options
Format
Display options
Format
Abstract
Background: MicroRNAs (miRNAs) are established regulators of development, cell identity and disease. Although nearly two thousand human miRNA genes are known and new ones are continuously discovered, no attempt has been made to gauge the total miRNA content of the human genome.
Results: Employing an innovative computational method on massively pooled small RNA sequencing data, we report 2,469 novel human miRNA candidates of which 1,098 are validated by in-house and published experiments. Almost 300 candidates are robustly expressed in a neuronal cell system and are regulated during differentiation or when biogenesis factors Dicer, Drosha, DGCR8 or Ago2 are silenced. To improve expression profiling, we devised a quantitative miRNA capture system. In a kidney cell system, 400 candidates interact with DGCR8 at transcript positions that suggest miRNA hairpin recognition, and 1,000 of the new miRNA candidates interact with Ago1 or Ago2, indicating that they are directly bound by miRNA effector proteins. From kidney cell CLASH experiments, in which miRNA-target pairs are ligated and sequenced, we observe hundreds of interactions between novel miRNAs and mRNA targets. The novel miRNA candidates are specifically but lowly expressed, raising the possibility that not all may be functional. Interestingly, the majority are evolutionarily young and overrepresented in the human brain.
Conclusions: In summary, we present evidence that the complement of human miRNA genes is substantially larger than anticipated, and that more are likely to be discovered in the future as more tissues and experimental conditions are sequenced to greater depth.
Figures
Evidence for the biogenesis of more than one thousand novel human miRNAs. (a) Number of human miRNAs in the miRBase database over time. NGS, next-generation sequencing.(b) Advantage of dataset pooling. Each of seven sequencing experiments detects a single strand from a miRNA gene (in blue). When the datasets are analyzed separately (top), the single sequencing read does not constitute evidence of miRNA Dicer processing as opposed to random degradation. However, when the datasets are pooled (bottom), the numbers and positions of the reads constitute strong evidence of Dicer processing.(c) Synergistic miRNA prediction pipeline. The workflow is described in the Results section. The three plots below the flowchart indicate the length of the sequences retained after each step; notice the approximately 22-nucleotide peak characteristic of Dicer processing.(d-h) Four key factors in miRNA biogenesis were silenced with RNA interference. sRNA expression was profiled with high-throughput sequencing in silenced and mock-transfected cells, and the log2 fold-change shown. The curves indicate the cumulative fraction of RNAs with the indicated fold-change or lower. The numbers in the left margins indicate the fractions of miRNAs that are substantially down-regulated (>30% change). Control sequences comprise transfer RNAs (tRNAs), small nucleolar RNAs (snoRNAs) and miscellaneous RNAs (miscRNAs) (in grey). *P <0.01; **P <0.001.(i-k) RNAs interacting with DGCR8, Ago1 and Ago2 have previously been detected by crosslinking immunoprecipitation coupled with high-throughput sequencing (CLIP-seq) [22-24]. For each known or novel miRNA the overlaps with these RNAs were plotted as the difference between the 5′ end of the miRNA hairpin and the interacting RNA. The number of miRNAs that are supported by CLIP-seq evidence is shown above each subfigure. In the cases where more than one interacting RNA supported a given miRNA, one random overlap was chosen, such that each data point represents one miRNA. The blue bars indicate the consensus positions of the miRNA strands.
Experimentally identified miRNA-mRNA interactions. (a) A method was recently developed to sequence miRNAs ligated to their mRNA targets [31]. In these data, we identified 256 distinct interactions between 87 novel miRNAs and 245 mRNAs. The inferred binding energies of the novel miRNAs are similar to those of known miRNAs, and are significantly stronger than the binding of shuffled control sequences (P <0.001, sub-sampling).(b) The miRNA-mRNA interactions were grouped based on 5' seed pairing. Both novel and known miRNAs tend to bind by canonical seed pairing, in contrast with control sequences. nt, nucleotide.
Three novel miRNAs and experimentally identified targets. (a) miRNA candidate 2375 targetsDicer mRNA. The density plot and read alignments show the distribution of sequenced RNAs mapping to the miRNA precursor, summing over 15 sRNA-seq datasets. The two miRNA strands are indicated in light and dark green. Above, the RNA structure of the precursor. Below, the candidate 2375 miRNA (dark green) ligated to the interactingDicer mRNA (blue).(b) Candidate 153 is integrated in protein biosynthesis pathways. The miRNA is derived from an intron of the EIF2B3 translation factor, and interacts with mRNA ofPSGM1, an established chaperone, andFPBP9, involved in protein folding.(c) Candidate 1331 targets ribosomes. It interacts with three mRNAs, of which two are distinct parts of the same 60S ribosomal subunit (RPL8 andRPL13A).
Features of the identified novel human miRNAs. (a) Inferred evolutionary origin of known and novel human miRNAs. Representative species are shown for each clade: hominids are represented by chimp; old world monkey, baboon; primate, tarsier; simian, tree shrew; placental mammal, armadillo; mammal, platypus; tetrapod, clawed frog; and vertebrate, zebrafish. The miRNAs have been divided into those specific to hominids and humans (‘specific’ in light green or orange) and those conserved in old world monkeys or beyond (‘conserved’ in dark green or brown).(b,c) Genomic sources of known and novel miRNAs.(d) Specificity of expression. The horizontal axis shows the number of datasets (out of 94) in which the miRNA is detected. The vertical axis shows the cumulative fraction of miRNAs present in at least these many datasets. The horizontal blue line indicates the median number of datasets in which the miRNAs are present. Examples of miRNAs detected in all 94 datasets are noted.(e) Maximal expression. As in the previous figure, except the maximum expression of each miRNA in any of the 94 datasets is shown. Expression is normalized (transcripts per million reads) to adjust for varying sequencing depth between the samples. The miRNA with the highest normalized expression in any dataset islet-7 f.(f-i) The processing precision of known and novel miRNAs. The precision is defined as the fraction of mapping reads that correspond to the consensus end position of the sequence. The miRNAs are sorted on the x-axis such that the most precisely processed one is at percentile 1 and the least precisely processed one at percentile 100. Dark colors indicate conserved miRNAs, light colors non-conserved.
Design of the custom SureSelect miRNA capture system. (a) The workflow of the capture system. An sRNA-seq library contains miRNAs of interest and other small RNAs. The library is mixed with biotinylated cRNA baits in solution. The baits are complementary in sequence to the target miRNAs and specifically hybridize with these. Magnetic streptavidin-coated beads are added to the solution and bind to the biotin. Last, the baits and their bound targets are isolated by magnetic separation, and the target miRNA clones can be sequenced.(b) The short size of miRNAs requires specific design considerations, since the SureSelect baits are much longer, at 120 nucleotides. If the baits are designed to hybridize with the target miRNA only, the interaction might be weak and targets lost. If baits are complementary to the target miRNA and the entire length of the flanking ligation adapters, which are added during library preparation, then specificity might be lost because most binding is to the universal adapter sequences. We have designed the baits so that they are complementary to the target miRNA and part of the adapters, giving a strong and specific hybridization. For targets longer than typical miRNAs, the length of the hybridized region is kept constant by dynamically varying the part of the adapters which is baited. NGS, next-generation sequencing.
Profiling novel miRNAs during differentiation with capture. (a) Saturation curve of miRNA detection with or without SureSelect capture-based enrichment. For sequencing depths ranging from 10 thousand to 30 million reads, the number of known target miRNAs detected is shown. Differentiated neuroblastoma cells were profiled.(b) As before, but for novel miRNAs.(c-f) Profiling depth with or without capture. The histograms show for each target miRNA how many times it is detected.(g) miRNA expression changes measured with or without capture. Neuroblastoma cells were induced to differentiate, and fold-changes were estimated for 140 known miRNAs that could be reliably profiled both with and without using the capture system.(h) miRNA expression changes during differentiation. The capture data from the previous figure could be used to robustly profile 428 known and novel target miRNAs, as cells underwent differentiation to neuron-like state. For each miRNA the normalized expression in the two states is shown. Five miRNAs have previously been observed to be regulated during differentiation of SH-SY5Y cells [42]. Our results are in agreement with the up- or down-regulation of four of these miRNAs, while the fifth, miR-422, is a border case in our measurements.
Saturation of novel miRNA prediction. To assess the influence of data magnitude on the analyses, saturation curves of the 94 datasets were performed.(a) Saturation curve of sequencing depth, from 10% to 100% of reads retained. For each dataset this percentage of (randomly chosen) reads were retained and subsequently the miRNA prediction analysis was repeated. The total number of reported novel miRNAs (brown) or high-confidence novel miRNAs (orange) is shown. The number of known miRNAs that are detected by simple sequence matches is shown in green.(b) As before, except that entire datasets rather than individual reads were discarded or retained.
References
- Lodish HF. Molecular Cell Biology. 6. New York: W.H. Freeman; 2008. Molecular cell biology.
- Dunham I, Kundaje A, Aldred SF, Collins PJ, Davis CA, Doyle F, Epstein CB, Frietze S, Harrow J, Kaul R, Khatun J, Lajoie BR, Landt SG, Lee BK, Pauli F, Rosenbloom KR, Sabo P, Safi A, Sanyal A, Shoresh N, Simon JM, Song L, Trinklein ND, Altshuler RC, Birney E, Brown JB, Cheng C, Djebali S, Dong X, Dunham I. etal.An integrated encyclopedia of DNA elements in the human genome. Nature. 2012;489:57–74. doi: 10.1038/nature11247. - DOI - PMC - PubMed
- Gerstein MB, Lu ZJ, van Nostrand EL, Cheng C, Arshinoff BI, Liu T, Yip KY, Robilotto R, Rechtsteiner A, Ikegami K, Alves P, Chateigner A, Perry M, Morris M, Auerbach RK, Feng X, Leng J, Vielle A, Niu W, Rhrissorrakrai K, Agarwal A, Alexander RP, Barber G, Brdlik CM, Brennan J, Brouillet JJ, Carr A, Cheung MS, Clawson H, Contrino S. et al.Integrative analysis of the Caenorhabditis elegans genome by the modENCODE project. Science. 2010;330:1775–1787. doi: 10.1126/science.1196914. - DOI - PMC - PubMed
Publication types
MeSH terms
Substances
Grants and funding
LinkOut - more resources
Full Text Sources
Other Literature Sources