Genetics: Probing the phenomics of noncoding RNA
It has been known since the late 1970s that many DNA sequences are transcribed but not translated. Moreover, most protein-coding genes in mammals are fragmented, with only a small fraction of the primary RNA transcript being spliced together to form messenger RNA. For many years it was assumed that untranslated RNA molecules served no useful purpose but, starting in the mid-1990s, a small body of researchers, including the present author (Mattick, 1994), have been arguing that these RNAs transmit regulatory information, possibly associated with the emergence of multicellular organisms. This is supported by the observation that the proportion of noncoding genomic sequences broadly correlates with developmental complexity, reaching over 98% in mammals (Liu et al., 2013), although others have argued that the increase in genome size is due to the inefficiency of selection against non-functional elements as body size goes up and population size goes down (Lynch, 2007).
High-throughput sequencing analyses over the past decade have shown that the majority of mammalian genome is transcribed, often from both strands, and have revealed an extraordinarily complex landscape of overlapping and interlacing sense and antisense, alternatively spliced, protein-coding and non-protein-coding RNAs, the latter generally referred to as long noncoding RNAs (lncRNAs). Moreover, the repertoire of these lncRNAs is different in different cells (Carninci et al., 2005; Cheng et al., 2005; Birney et al., 2007; Mercer et al., 2012). While some transcripts may encode previously unrecognized small proteins, the function or otherwise of the vast majority of lncRNAs remains to be determined.
Because many lncRNAs appear to be expressed at low levels, and many have lower sequence conservation than messenger RNAs, one interpretation has been that these RNAs represent transcriptional noise from complex genomes cluttered with evolutionary debris. However, assessments of sequence conservation rely on assumptions about the non-functionality and representative distribution of reference sequences, which are not verified and cannot be directly tested (Pheasant and Mattick, 2007). Nonetheless, many lncRNAs show patches of relative sequence conservation (Derrien et al., 2012), and even more do so at the secondary structural level (Smith et al., 2013).
Expression analyses have shown that lncRNAs originate from all over the genome and are expressed at different times during differentiation and development (Dinger et al., 2008), often exhibiting highly cell-specific patterns (Mercer et al., 2008). The precision of lncRNA expression is consistent with evidence suggesting that many are associated with chromatin-modifying complexes, thereby acting as regulators of the epigenetic control of differentiation and development (Mercer and Mattick, 2013).
A number of lncRNAs have also been linked to complex diseases like cancer (Mattick, 2009) and other complex physiological processes (see, for example, Rapicavoli et al., 2013). However, these results seem at odds with the fact that few lncRNAs have been identified in traditional genetic screens. The reason for this is likely a combination of phenotypic, technical and expectational bias: mutations in protein-coding regions of the genome generally have phenotypes that are more severe, and are easier to identify, than those in non-coding regions. By contrast, in this context, it is worth noting that ∼95% of all variants associated with complex (as opposed to monogenic) diseases in humans map to non-coding, presumably regulatory, sequences (Freedman et al., 2011).
Still, the gold standard in this field is the targeted in vivo silencing or deletion of specific genes, and since few of these have been conducted to date, some researchers have remained sceptical about the biological significance of lncRNAs. Now, in eLife, John Rinn, Paolo Arlotta and co-workers at Harvard, MIT, the Broad Institute, Rutgers and Regeneron Pharmaceuticals—including Martin Sauvageau, Loyal Goff and Simona Lodata as joint first authors—report the results of the first large-scale attack on the question (Sauvageau et al., 2013). They selected 18 lncRNA genes in the mouse genome that had been stringently assessed for lack of protein-coding capacity and that did not overlap with known protein-coding genes or other known gene annotations—hence the name long intergenic noncoding RNAs (lincRNAs)—and generated knockout mouse mutants by replacing the lncRNA gene with a lacZ reporter cassette.
Sauvageau, Goff, Lodata et al. report discernable developmental problems in five of the 18 mutants, with three exhibiting embryonic or post-natal lethality, two of which exhibited growth defects in the survivors. The phenotypes of two of the mutants were analyzed in detail: one of the mutants that died showed defects in multiple organs (including the lung, heart and gastrointestinal tract), and one of the mutants that survived with growth defects also showed defects in the cerebral cortex. Other mutants that did not exhibit overt developmental defects showed brain-specific expression patterns and may be associated with cognitive defects that are not grossly apparent at the developmental level.
Another group (Grote et al., 2013) recently generated a different knockout allele for one of the 18 lincRNAs interrogated by Sauvageau et al., and also reported an embryonic lethal phenotype, albeit with some differences. Importantly, the approach used by Grote et al. also provided strong evidence that the mutant defects were not caused by an indirect effect on an overlapping genomic element, such as an enhancer for a nearby gene.
The work of Sauvageau, Goff, Lodata et al. is a mini tour-de-force that shows that there are lncRNAs with important developmental functions in vivo, and it joins a small number of studies from other pioneering groups that show the same thing (Lewejohann et al., 2004; Gutschner et al., 2013; Li et al., 2013), although not all of the targeted lncRNAs showed a phenotype. Similarly, other knockout experiments of widely expressed lncRNAs, as well as some of the most highly conserved elements in the mammalian genome, also did not yield discernable phenotypes (Ahituv et al., 2007; Nakagawa et al., 2011), which should sound a note of caution about the interpretation of negative results.
Indeed, since most lncRNAs are expressed in the brain (Mercer et al., 2008) and many are primate-specific (Derrien et al., 2012), it may be that much of the lncRNA-mediated genetic information in humans (and in mammals generally) is devoted to brain function, and therefore not easily detectable in developmental, as opposed to cognitive, screens. A good example is a noncoding RNA called BC1 that is widely expressed in the brain: knockout of BC1 causes no visible anatomical consequences, but it leads to a behavioural phenotype that would be lethal in the wild (Lewejohann et al., 2004).
Although evidence for the hypothesis that lncRNAs have a role in mammalian development, brain function and physiology is growing, there is also a clear need for more sophisticated and comprehensive phenotypic screens, especially with respect to cognitive function.
References
-
Principles for the post-GWAS functional characterization of cancer risk lociNature Genetics 43:513–518.https://doi.org/10.1038/ng.840
-
Role of a neuronal small non-messenger RNA: behavioural alterations in BC1 RNA-deleted miceBehavioural Brain Research 154:273–289.https://doi.org/10.1016/j.bbr.2004.02.015
-
Introns: evolution and functionCurrent Opinion in Genetics & Development 4:823–831.https://doi.org/10.1016/0959-437X(94)90066-3
-
The genetic signatures of noncoding RNAsPLOS Genetics 5:e1000459.https://doi.org/10.1371/journal.pgen.1000459
-
Specific expression of long noncoding RNAs in the mouse brainProceedings of the National Academy of Sciences of the United States of America 105:716–721.https://doi.org/10.1073/pnas.0706729105
-
Targeted RNA sequencing reveals the deep complexity of the human transcriptomeNature Biotechnology 30:99–104.https://doi.org/10.1038/nbt.2024
-
Structure and function of long noncoding RNAs in epigenetic regulationNature Structural and Molecular Biology 20:300–307.https://doi.org/10.1038/nsmb.2480
-
Paraspeckles are subpopulation-specific nuclear bodies that are not essential in miceJournal of Cell Biology 193:31–39.https://doi.org/10.1083/jcb.201011110
-
Raising the estimate of functional human sequencesGenome Research 17:1245–1253.https://doi.org/10.1101/gr.6406307
-
Widespread purifying selection on RNA structure in mammalsNucleic Acids Research 41:8220–8236.https://doi.org/10.1093/nar/gkt596
Article and author information
Author details
Publication history
- Version of Record published: December 31, 2013 (version 1)
Copyright
© 2013, Mattick
This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.
Metrics
-
- 1,126
- views
-
- 136
- downloads
-
- 13
- citations
Views, downloads and citations are aggregated across all versions of this paper published by eLife.
Download links
Downloads (link to download the article as PDF)
Open citations (links to open the citations from this article in various online reference manager services)
Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)
Further reading
-
- Developmental Biology
- Evolutionary Biology
Despite rapid evolution across eutherian mammals, the X-linked MIR-506 family miRNAs are located in a region flanked by two highly conserved protein-coding genes (SLITRK2 and FMR1) on the X chromosome. Intriguingly, these miRNAs are predominantly expressed in the testis, suggesting a potential role in spermatogenesis and male fertility. Here, we report that the X-linked MIR-506 family miRNAs were derived from the MER91C DNA transposons. Selective inactivation of individual miRNAs or clusters caused no discernible defects, but simultaneous ablation of five clusters containing 19 members of the MIR-506 family led to reduced male fertility in mice. Despite normal sperm counts, motility, and morphology, the KO sperm were less competitive than wild-type sperm when subjected to a polyandrous mating scheme. Transcriptomic and bioinformatic analyses revealed that these X-linked MIR-506 family miRNAs, in addition to targeting a set of conserved genes, have more targets that are critical for spermatogenesis and embryonic development during evolution. Our data suggest that the MIR-506 family miRNAs function to enhance sperm competitiveness and reproductive fitness of the male by finetuning gene expression during spermatogenesis.
-
- Developmental Biology
We previously showed that SerpinE2 and the serine protease HtrA1 modulate fibroblast growth factor (FGF) signaling in germ layer specification and head-to-tail development of Xenopus embryos. Here, we present an extracellular proteolytic mechanism involving this serpin-protease system in the developing neural crest (NC). Knockdown of SerpinE2 by injected antisense morpholino oligonucleotides did not affect the specification of NC progenitors but instead inhibited the migration of NC cells, causing defects in dorsal fin, melanocyte, and craniofacial cartilage formation. Similarly, overexpression of the HtrA1 protease impaired NC cell migration and the formation of NC-derived structures. The phenotype of SerpinE2 knockdown was overcome by concomitant downregulation of HtrA1, indicating that SerpinE2 stimulates NC migration by inhibiting endogenous HtrA1 activity. SerpinE2 binds to HtrA1, and the HtrA1 protease triggers degradation of the cell surface proteoglycan Syndecan-4 (Sdc4). Microinjection of Sdc4 mRNA partially rescued NC migration defects induced by both HtrA1 upregulation and SerpinE2 downregulation. These epistatic experiments suggest a proteolytic pathway by a double inhibition mechanism:
SerpinE2 ┤HtrA1 protease ┤Syndecan-4 → NC cell migration.