Review
doi: 10.1146/annurev-genet-120213-092023.
Epub 2015 Oct 14.
A Uniform System for the Annotation of Vertebrate microRNA Genes and the Evolution of the Human microRNAome
Affiliations
- PMID: 26473382
- PMCID: PMC4743252
- DOI: 10.1146/annurev-genet-120213-092023
Item in Clipboard
Review
A Uniform System for the Annotation of Vertebrate microRNA Genes and the Evolution of the Human microRNAome
Annu Rev Genet.
2015.
Abstract
Although microRNAs (miRNAs) are among the most intensively studied molecules of the past 20 years, determining what is and what is not a miRNA has not been straightforward. Here, we present a uniform system for the annotation and nomenclature of miRNA genes. We show that less than a third of the 1,881 human miRBase entries, and only approximately 16% of the 7,095 metazoan miRBase entries, are robustly supported as miRNA genes. Furthermore, we show that the human repertoire of miRNAs has been shaped by periods of intense miRNA innovation and that mature gene products show a very different tempo and mode of sequence evolution than star products. We establish a new open access database--MirGeneDB ( http://mirgenedb.org )--to catalog this set of miRNAs, which complements the efforts of miRBase but differs from it by annotating the mature versus star products and by imposing an evolutionary hierarchy upon this curated and consistently named repertoire.
Keywords: MirGeneDB; genome duplication; miRBase; miRNA; molecular evolution; vertebrate.
Figures
Examples of canonical, equivocal, non-canonical, and likely mis-annotated, human miRNA sequences. A. hsa-mir-224, a miRNA shared among placental mammals. Note that all of the criteria established by the community for bona fide miRNAs are met including length of each of the arms and the loop, the complementarity, 2 nucleotide overhangs, and 5′ end homogeneity. The mature arm (5p) is shown in red and the star arm (3p) in blue – note the greater than 2X differential between the expression of the two arms (bottom). B. Another example of a canonical miRNA gene, hsa-mir-212, shared amongst vertebrates. This miRNA shows the same features as mir-224, except in this case the two arms are expressed in a nearly equal ratio, and thus is an example of a miRNA that has two mature (or co-mature) arms (red). C. An example of an equivocal miRNA gene, hsa-mir-498, where expression of only one arm has been detected, abrogating the ability to ascertain key criteria for miRNA annotation including the 2-nucleotide offset. Nonetheless, if the 3p is expressed with the correct offset, then this sequence will fall within the parameters (numbers in parenthesis) established herein (see Table 1) for miRNA annotation. D. An example of a non-canonical miRNA, hsa-mir-451, a deeply conserved sequence that bypasses Dicer processing and thus only expresses one of the two arms. E, F. Two examples of sequences that are unlikely to be derived from a miRNA gene as they show none of the criteria established by the community for miRNA annotation including lack of phylogenetic conservation.
The miRNA gene mir-8 goes by six different names depending on the taxon of origin and the orientation of transcription. The last common ancestor of flies and rats had a single mir-8 gene, and this sequence is called mir-8 in both the annelid Capitella teleata (cte) and in the fruit fly Drosophila melanogaster (dme). This same gene though goes by the name of mir-236 in nematodes like C. elegans (cel), and hence is an example of a redundant orthologue (orange, see Fig. 3). In deuterostomes, because of gene duplication events (open circles), this same gene goes by three different names, mir-141, mir-200 and mir-429, all paralogues (blue, Fig. 3B) of the mir-8 gene. Finally, an antisense read of the rat (rno) mir-200a sequence exists and it is called mir-3548.
Curation of the miRBase (v. 21) entries for human miRNA sequences (A) and for the entire collection of numbered animal-specific gene families (B). Only about 27% of the human miRNA sequence entries, and 16% of the animal family-level entries, are supported as bona fide miRNAs (white) using a consistent set of criteria (see Fig. 1A, B) whereas about 58% of the sequences (left) and 48% of the families (right) can be rejected (red), including mir-1202 (Fig. 1E) and mir-8484 (Fig. 1F). Redundant entries (orange) are those miRNAs where the same sequence has been given two different names in two different species (e.g., mir-8 and mir-236, Fig. 2). Paralogous entries (light blue) are those miRNAs where two or more copies of the gene are given two or more different names (e.g., mir-141, -200 and -429, Fig. 2). Equivocal entries (purple) are those entries that do not show all the necessary data to robustly either support or reject the entry, usually due to the fact that only one arm was reported (e.g., Hsa-mir-498, Fig. 1C). Non-canonical entries are those “miRNAs” that fail at least one of the criteria, but are deeply conserved (e.g., mir-451, Fig. 1D). Antisense entries (light green) are entries that are simply the antisense read of another accepted entry (e.g., mir-3548, Fig. 2). False negatives (grey) are genes that are likely to be present in the human genome, but are not yet deposited in miRBase (see Supp. File 1).
The relationship between the evolutionary history of a miRNA family (MIR-15) and the nomenclature system proposed herein. Early in vertebrate evolutionary history a single Mir-15 gene was duplicated in tandem generating two copies of this gene (red and blue), in contrast to the single gene still found in the genome of the ascidian urochordate Ciona (purple). The red gene is labeled Mir-15-P1 and the blue gene is Mir-15-P2. Then, vertebrates underwent two rounds of whole genome duplication, generating four clusters of two genes, and these clusters are labeled a–d. Thus, there are four copies of the P1 gene (P1a, b, c and d) and four copies of the P2 genes (P2a, b, c and d). These four clusters are then passed on to the zebrafish, chicken and human lineages through a series of speciation events (black circles), each with their own examples of gene loss, and in the human lineage, gene gain. In the human lineage the “d” cluster was lost while the syntenic genes were retained, whereas in the chicken, the “c” cluster and the anchoring gene (Alox-12) were lost. On the lineage leading to zebrafish a third round of genome duplication occurred primitively generating 8 clusters of genes, five of which were retained in zebrafish with the “b2”, “c2” and “d2” clusters lost, as well as the P1d gene. Importantly, both the phylogeny (bottom) and the synteny (left) are concordant, allowing for an internally consistent scenario and for a robust nomenclature system. Note that Mir-15-P3 (= mir-424) is a eutherian-specific paralogue of the MIR-15 family and is not shown here.
The evolutionary history of microRNA genes across the animal kingdom. For each named node (i.e., branching point) the number of both families and genes (in parentheses) gained (top) and lost (bottom) are indicated. See Supplemental File 4 for the entire list of every gene gained and lost for every node shown, as well as the taxonomic names for each numbered node or branch, and the genus and species identities for all figured animals. Divergence times are taken from Erwin et al. (36), Near et al. (94), and dos Reis et al. (34). Geological time is shown on the bottom (in millions of years) and geological abbreviations are as follows: C – Cambrian; O – Ordovician; S – Silurian; D – Devonian; C – Carboniferous; P – Permian; T – Triassic; J – Jurassic; K – Cretaceous; Pe – Paleogene; N – Neogene + Quaternary. On the right are shown the rate of acquisition (red) and loss (blue) for both miRNA families (dotted lines) and genes (solid lines) for four key taxa, human, chicken, fish and fruit fly, with outlier periods of miRNA acquisition indicated by the numbered arrows. The increases in species-specific rates of miRNA genes for fly, fish and chicken are given with an “S” with the slope likely dictated by the depth of sequencing for each of these taxa (14; 89). See the text for further details. Times of genome duplication are shown in orange.
Mutational and nucleotide profiles of mature and star sequences of 234 genes present in the last common ancestor of tetrapods. A. The rate of nucleotide substitution per position of the sequences, from position 1 to 22, is shown for both the star (blue) and mature (red) sequences (n = the total number of tallied mutations). From a functional perspective the mutational profile for both the mature and star are grossly similar, except that there is no difference between nucleotides 1 and 2–8 in the star sequence. In addition, many more mutations occur in positions 2-8 of the star sequence (426 substitutions) relative to the mature sequence (10 substitutions). B. Rate of nucleotide substitution per position of the mature, star, and loop regions. The distribution of rates per region is summarized by the boxplots. The bold horizontal line through the box represents the median rate. The lower and upper edges of the box represent the 1st and 3rd quartiles, respectively. The vertical bars represent the range of values that are not outliers. The unfilled circles represent outliers. C. From a structural perspective the pattern of star mutation mirrors the mature, with regions of high conservation of the mature paired with regions of high conservation of the star, and vice versa, consistent with the notion that the mutational profile of the star is constrained by the conservation of the sequence of the mature miRNA. Because of bulges, many miRNAs are asymmetrical; the trends elucidated herein might even be more apparent if this was taken into account. We assumed for the construction of these logos that both arms are symmetrical, similar to what is shown for Hsa-Mir-126 (top). D. Nucleotide base frequencies for both the mature (bottom) and the star (top) for each of the 199 genes present in human that were inherited from the last common ancestor of tetrapods. On the right are shown the sequence logos aligned so that mature position 1 is opposite that of star position 20, and vice versa, in line with panel C. The asterisks indicate significance of the skew (if any) based on a chi-square test at three different levels of significance. All significant positions of the star correspond to significant positions of the mature, consistent with the hypothesis that base-pairing with the mature largely governs star evolution. Importantly, positions 1 and 9 of the mature are not matched by a corresponding bias in the star with both positions highly skewed towards “U.” When the polarity of change is established for mature position 1 at the family level (right), and each change recorded over the nearly four billion year evolutionary history of the 21 considered taxa (50 for mature, 63 for stars), the bias in possessing a U at position 1 is not retained, only the continued underrepresentation of G, and therefore once established miRNAs are relatively free to evolve to either A or C (see also panel A). In addition, although U’s are dramatically underrepresented at star position 1, again once the miRNA gene is established this position can take on any identity, including G. Together these data show the influences that the three different macromolecular partners have on miRNA mature strand evolution: the role target interaction has on conservation of mature seed and, to a lesser degree, 3′ complementarity region (3′CR); the role the opposite (= star) strand has on base composition at positions 2, 5, and 13-20, and AGO 2 on positions 1 and, presumably, 9.
References
-
- Alvarez-Garcia I, Miska EA. MicroRNA functions in animal development and human disease. Development. 2005;132:4653–62. - PubMed
Publication types
MeSH terms
Substances
Grants and funding
LinkOut - more resources
Full Text Sources
Other Literature Sources
Molecular Biology Databases
Miscellaneous
