Drosophila miR-87  promotes dendrite regeneration by targeting the transcriptional repressor Tramtrack69

There is no author summary for this article yet. Authors can add summaries to their articles on ScienceOpen to make them more accessible to a non-specialist audience.

Abstract

To remodel functional neuronal connectivity, neurons often alter dendrite arbors through elimination and subsequent regeneration of dendritic branches. However, the intrinsic mechanisms underlying this developmentally programmed dendrite regeneration and whether it shares common machinery with injury-induced regeneration remain largely unknown. Drosophila class IV dendrite arborization (C4da) sensory neurons regenerate adult-specific dendrites after eliminating larval dendrites during metamorphosis. Here we show that the microRNA miR-87 is a critical regulator of dendrite regeneration in Drosophila. miR-87 knockout impairs dendrite regeneration after developmentally-programmed pruning, whereas miR-87 overexpression in C4da neurons leads to precocious initiation of dendrite regeneration. Genetic analyses indicate that the transcriptional repressor Tramtrack69 (Ttk69) is a functional target for miR-87-mediated repression as ttk69 expression is increased in miR-87 knockout neurons and reducing ttk69 expression restores dendrite regeneration to mutants lacking miR-87 function. We further show that miR-87 is required for dendrite regeneration after acute injury in the larval stage, providing a mechanistic link between developmentally programmed and injury-induced dendrite regeneration. These findings thus indicate that miR-87 promotes dendrite regrowth during regeneration at least in part through suppressing Ttk69 in Drosophila sensory neurons and suggest that developmental and injury-induced dendrite regeneration share a common intrinsic mechanism to reactivate dendrite growth.

Author summary

Dendrites are the primary sites for synaptic and sensory inputs. To remodel or repair neuronal connectivity, dendrites often exhibit large-scale structural changes that can be triggered by developmental signals, alterations in sensory inputs, or injury. Despite the importance of dendritic remodeling to nervous system function, the molecular basis for this remodeling is largely unknown. Here we used an unbiased genetic screen and in vivo imaging in Drosophila sensory neurons to demonstrate that the microRNA miR-87 is a critical factor required in neurons to reactivate dendritic growth both in developmental remodeling and following injury. Our work supports the model that miR-87 promotes dendrite regeneration by blocking expression of the transcriptional repressor Tramtrack69 in neurons. This study thus establishes a role for miRNAs in temporal control of dendrite regeneration.

Related collections

Most cited references 47

Record: found
Abstract: found
Article: not found

miRNA-mediated gene silencing by translational repression followed by mRNA deadenylation and decay.

S. Djuranovic, A. Nahvi, R. Green (2012)

microRNAs (miRNAs) regulate gene expression through translational repression and/or messenger RNA (mRNA) deadenylation and decay. Because translation, deadenylation, and decay are closely linked processes, it is important to establish their ordering and thus to define the molecular mechanism of silencing. We have investigated the kinetics of these events in miRNA-mediated gene silencing by using a Drosophila S2 cell-based controllable expression system and show that mRNAs with both natural and engineered 3' untranslated regions with miRNA target sites are first subject to translational inhibition, followed by effects on deadenylation and decay. We next used a natural translational elongation stall to show that miRNA-mediated silencing inhibits translation at an early step, potentially translation initiation.

0 comments Cited 303 times – based on 0 reviews      Review now

Bookmark

Record: found
Abstract: found
Article: not found

Identification of Drosophila MicroRNA Targets

Alexander Stark, Julius Brennecke, Robert B. Russell … (2003)

Introduction MicroRNAs (miRNAs) are small RNAs, typically of approximately 21–23 nt, that direct posttranscriptional regulation of gene expression by binding to messenger RNAs (mRNAs). Many endogenously encoded miRNAs have been cloned from plants and animals (Lagos-Quintana et al. 2001, 2002; Lau et al. 2001; Lee and Ambros 2001; Mourelatos et al. 2002; Reinhart et al. 2002; Ambros et al. 2003; Aravin et al. 2003; Lim et al. 2003). Combining these data with computational cross-genome comparison predicts 100–120 miRNA-encoding genes in Caenorhabditis and Drosophila and approximately 250 in mouse and human (Ambros et al. 2003; Grad et al. 2003; Lai et al. 2003; Lim et al. 2003a, 2003b). However, functions have been assigned to only four animal miRNAs (Reinhart et al. 2000; Brennecke et al. 2003; Lee et al. 1993; Wightman et al. 1993; Xu et al. 2003), in part owing to the difficulty in identifying mutations in such small genes. A method to identify the target genes that are regulated by miRNAs would greatly help the study of miRNA function in animals (Ambros 2001). Two modes of miRNA-directed target inhibition have been demonstrated. The same small RNA can cause degradation of its target mRNA or block its translation depending on the degree of miRNA–target sequence complementary (Hutvagner and Zamore 2002; Doench et al. 2003). Target RNAs containing sequences with perfect complements of the miRNA (or exogenously supplied short interfering RNA [siRNA]) are cleaved by the RNA-induced silencing complex (RISC) ribonuclease (Hutvagner and Zamore 2002; Martinez et al. 2002; Zeng et al. 2002). Endogenous plant miRNAs have been shown to regulate target RNAs by RNA interference (RNAi) involving perfect or near-perfect target site complementarity (Llave et al. 2002b; Kasschau et al. 2003; Palatnik et al. 2003; Tang et al. 2003; Xie et al. 2003). Targets for plant miRNAs have been identified on a genome-wide scale by searches that require a high degree of sequence complementarity (Rhoades et al. 2002). However, this approach did not find targets for known animal miRNAs. The animal miRNAs tested until now pair imperfectly with their targets and act to control translation. Indeed, systematic analysis of the complete C. elegans miRNA complement has confirmed the absence of targets with perfect or near-perfect sequence complementarity (Ambros et al. 2003). To date targets have been experimentally validated for just three animal miRNAs: the lin-4 targets lin-14 and lin-28 (Wightman et al. 1993; Ha et al. 1996; Moss et al. 1997; Olsen and Ambros 1999; Seggerson et al. 2002), the let-7 targets lin-41 and lin-57/hbl-1 (Reinhart et al. 2000; Slack et al. 2000; Abrahante et al. 2003; Lin et al. 2003), and the bantam target hid (Brennecke et al. 2003). These miRNA–target duplexes contain mismatches, gaps, and G:U basepairs at different positions. Even allowing for G:U basepairs, the longest contiguous alignments in these examples range from 8 to 10 nt. Such limited information content makes it difficult to identify targets within whole-genome or transcriptome databases, since standard alignment methods produce many false positives with such short variable sequences. Furthermore, the small number of validated examples makes the development and benchmarking of a generally applicable computational method problematic at present. Here we present a screen for miRNA targets in Drosophila that identifies all of the previously known miRNA targets and we demonstrate that it successfully predicts new targets that we validate experimentally. Results Database Design For each of the validated miRNA–target pairs, functional target sites are located in the 3′ untranslated region (UTR) of the mRNA and are conserved in the 3′ UTRs of the homologous genes from related species (Wightman et al. 1993; Moss et al. 1997; Pasquinelli et al. 2000; Brennecke et al. 2003). We used pairwise comparison of the 3′ UTRs of orthologous genes in related genomes to identify conserved 3′ UTR sequences. Figure 1A shows the resulting pattern of 3′ UTR conservation for the known targets in worms and flies. The vast majority of miRNA target sites (red bars in Figure 1A) are located in blocks of conserved sequence (white blocks in Figure 1A). Figure 1B shows cross-genome conservation of these miRNA target sites. A striking pattern of uninterrupted conservation emerges at the end of the target sequences that pair with the 5′ end of the miRNAs. To permit genome-wide searches for targets of Drosophila miRNAs, a conserved 3′ UTR database was prepared by comparison of Drosophila melanogaster and Drosophila pseudoobscura 3′ UTRs. As very few 3′ UTRs are defined by cDNA sequence data in D. pseudoobscura, we used genomic sequence following the last exon of the D. pseudoobscura gene as the orthologous UTR (see Materials and Methods). Last exons were reliably detected in D. pseudoobscura for approximately two-thirds of D. melanogaster genes. On average, 22% of the D. melanogaster 3′ UTR sequence is conserved in the predicted D. pseudoobscura 3′ UTR. Much of this reflects isolated blocks of very high conservation interspersed among less-conserved sequence. Use of conserved 3′ UTRs reduces the expected number of sequence matches that would occur at random by 4- to 5-fold in relation to full-length 3′ UTRs, and severalfold further compared to the full transcriptome. We considered using the Anopheles gambiae genome to extend the cross-species' comparison. Although genome annotation identifies orthologs for two-thirds of D. melanogaster genes (Zdobnov et al. 2002), we were unable to identify the last D. melanogaster exon for approximately half of these. We therefore chose not to require conservation in Anopheles, but use it as an additional level of validation for predicted Drosophila targets, where possible. Screening Strategy We have adopted a two-step approach to target identification that combines a sensitive sequence database search with an RNA folding algorithm to evaluate the quality of the RNA duplex formed between the miRNA and its predicted targets. We examined the known target sites for lin-4, let-7, and bantam for common features. All of these sites showed better complementarity to the 5′ end of the miRNA, with no obvious common features elsewhere (Figure 2A and 2B). There were few sequence mismatches or G:U basepairs in the alignment of the first eight residues at the 5′ end of the miRNA. We used the alignment tool HMMer (Eddy 1996) to search for sequences complementary to the first eight residues of the miRNA, allowing for G:U mismatches. Where possible, the corresponding sites were also identified in the D. pseudoobscura 3′ UTR, and the sites from both genomes were considered, since the regions outside of the sequence match can vary between the two organisms, leading to differences in subsequent steps (see below). The identified sequences were extended to the length of the miRNA plus five residues to allow for bulges and were evaluated for their ability to form energetically favorable RNA–RNA duplexes with the miRNA using Mfold, which combines knowledge of known RNA structures with thermodynamic parameters, such as those involved in basepairing to evaluate the free energy of folding (ΔG) (Mathews et al. 1999; Zuker et al. 1999). Mfold requires a single linear sequence as input, so each predicted target was linked to the miRNA using a standard hairpin-forming linker sequence (GCGGGGACGC). An example of the Mfold output is shown in Figure 2C for the top-scoring bantam miRNA target site that we had previously identified in the 3′ UTR of hid (Brennecke et al. 2003). The Mfold free energy of folding (ΔG) was determined for each predicted target, which allows predicted sites to be ranked according to ΔG. However, ΔG depends on miRNA length and GC content, so it is not possible to distinguish systematically real targets from random matches using the raw ΔG score or to compare different miRNAs. Instead, we calculated Z-values as a measure of nonrandomness, with an average random site scoring Z = 0 (Z = standard deviations [SD] above the mean of background matches). Figure 2D shows the distribution of folding energies for predicted targets of the bantam miRNA compared to 10,000 randomly selected target sequences. Most of the previously validated targets have more than one predicted miRNA-binding site in their 3′ UTRs. Use of the Z-value allows us to add the scores of several sites within one UTR by selecting only those scores that are different from background matches. This is not possible with ΔG alone because even average random matches have favorable energy values (Figure 2D) and the sum of several average random matches in a UTR could score better than a single true site. We have selected Z ≥ 3 as a cutoff value, as folding energies of more than 3 SD above the mean (Z ≥ 3) are expected to occur for only 0.3% of random matches. Use of a higher Z-value increases the likelihood that predictions are correct, but also increases the risk of missing out contributions from real sites of lower folding energy. For example, only three of the five conserved bantam sites previously identified in the hid 3′ UTR score Z ≥ 3 (with the best site at Z = 7.4). We evaluated our predictions by the best single site in the 3′ UTR (Z max) and by the sum of sites with Z ≥ 3 (Z UTR). Tests with Previously Validated Targets Table 1 summarizes the performance of the method in predicting the known targets of the C. elegans miRNAs lin-4 and let-7 and the Drosophila miRNA bantam. The Drosophila hid gene ranked first of all predicted bantam targets sorted by the single best site (Z max) or by the sum of sites (Z UTR). All of the known targets of lin-4 and let-7 were found when their 3′ UTRs were added to the Drosophila 3′ UTR database. Like hid, the let-7 target lin-57 ranked near the top of the list by both measures. With several sites predicted of Z ≥ 3, lin-57 ranked first by Z UTR. Its best single site ranked in position 2 (Z = 6.8). C. elegans lin-14 was predicted to contain multiple lin-4 sites (Wightman et al. 1993). Three of these scored Z ≥ 3. lin-14 ranked first when the list of predicted lin-4 targets was sorted for Z UTR,although the best single site in lin-14 placed it in position 20 (Z = 4.3). The lin-4 target lin-28 and the let-7 target lin-41 ranked highly when the lists were sorted for the best single site, but ranked lower when multiple sites were summed because they had few high-scoring sites. The Drosophila homolog of lin-41, dpld, also ranked high among let-7 targets (Z = 5.6; see below). We compared our results with previous target predictions from the literature that have not been experimentally validated (Table 1). Our screen supports some of them (e.g., let-7 regulating lin-14), but we consider others unlikely because they rank very low on their lists or have no sites of Z ≥ 3 (e.g., let-7 and lin-28 or miR-4 and m4). The predicted miR-14 target Drice (Xu et al. 2003) is unlikely to be valid because the site is not conserved in the predicted Drice 3′ UTR from D. pseudoobscura. This analysis shows that all known targets were detected and ranked among the top-scoring predictions in genome-wide searches. This suggests that other valid targets should rank among the small number of best predictions that can be tested experimentally. Of particular interest were three miRNAs for which we predicted clusters of functionally related targets: miR-7, the miR-2 family, and miR-277 (Table 2; Table 3). Clustering of top-scoring sites in a group of related genes is likely to be significant when it arises from an unbiased genome-wide analysis. miR-7 and miR-2 were selected for target validation. miR-7 Regulates Notch Targets Among the top 10 predictions for miR-7, we found Enhancer of split (E(spl)) and Bearded (Brd) complex genes (Figure 3A). HLHm3 encodes a basic helix–loop–helix (bHLH) transcriptional repressor; Tom and m4 encode Brd family proteins. The bHLH repressor hairy was also among the top 10. These sites were conserved in the orthologous genes from Anopheles, when those could be identified. This prompted us to examine all the genes in E(spl) and Brd complexes for miR-7 sites. We found possible target sites in many of them. Alignment of these sites showed a pattern of 5′ end conservation quite similar to that for validated targets, with no mismatches and few G:U basepairs for about half of these genes (Figure 3B). To assess the validity of some of the predicted targets, transgenic flies expressing the miR-7 miRNA and several sensor transgenes were prepared. A genomic fragment containing the miR-7 hairpin was cloned into the 3′ UTR of a UAS–DSRed2 plasmid to allow GAL4-dependent expression of miR-7. The 3′ UTRs of HLHm3, m4, and hairy were cloned into a tubulin promoter–EGFP (enhanced green fluorescent protein) reporter plasmid. As a control, a specific miR-7 sensor transgene was produced by cloning two copies of a perfect complement of the miR-7 miRNA sequence into the 3′ UTR of the tubulin promoter–EGFP reporter. The miR-7 sensor was expressed uniformly in the wing imaginal disc. GAL4-dependent expression of miR-7 miRNA reduced expression of miR-7 green fluorescent protein (GFP) sensor transgene (Figure 3C). As the target sites in the sensor construct are perfect complements of the miR-7 miRNA, this is expected to be due to RNAi. GAL4-dependent expression of miR-7 also reduced expression of the m4 3′ UTR sensor transgene (Figure 3D). The miR-7 site in m4 is identical in D. pseudoobscura and conserved in Anopheles. Expression of miR-7 also caused a clear downregulation of the hairy 3′ UTR sensor transgene, although its overall level of expression was lower in the wing disc (Figure 3E). The hairy gene has been cloned and cDNAs sequenced from two additional insect genomes: the flour beetle Tribolium castanaeum and Drosophila simulans. The predicted miR-7-binding site is conserved in these genomes, as well as in Anopheles, and shows striking conservation of alignment at the 5′ and 3′ ends of the predicted miRNA-binding site (Figure 3F). The level of expression of the HLHm3 3′ UTR sensor was too low to be reliably studied, but also showed regulation by miR-7. Again, the miR-7 site in HLHm3 is identical in D. pseudoobscura. These observations validate the utility of the screen in predicting new miRNA targets. To assess miR-7 function in vivo, we examined wings in which miR-7 was overexpressed under ptc–Gal4 control. Notching of the wing margin was observed (Figure 3G), which is characteristic of reduced Notch signaling (de Celis and Garcia Bellido 1994; Diaz-Benjumea and Cohen 1995; Rulifson and Blair 1995; Micchelli et al. 1997). The Notch target cut was expressed at reduced levels in the miR-7-expressing cells at the wing margin (Figure 3E); wingless expression was also reduced (data not shown). Although bHLH transcription factors and Brd-like genes of the E(spl) complex are not strictly required for all aspects of Notch activity at the wing margin, they are required for cut expression (Ligoxygakis et al. 1999). miR-7 expression might provide a means to simultaneously downregulate these and other proteins, which might otherwise function redundantly to mediate Notch activity in the wing margin. We also noted reduced spacing of veins 3 and 4 in these wings, which may also reflect reduced Notch activity in controlling tissue growth (Baonza and Garcia-Bellido 1999). E(spl) genes are also expressed in proneural clusters, where they are required for sense organ determination and bristle patterning. Clones of cells lacking multiple genes of the E(spl) complex form extra bristles (de Celis et al. 1996). Consistent with this, we observed that expression of miR-7 under ptc–Gal4 control causes ectopic bristles and bristle duplication in the scutellum (data not shown). Taken together, these findings support the prediction that miR-7 miRNA regulates expression of bHLH and Brd-like proteins encoded by hairy and the E(spl) and Brd complex genes and implicates miR-7 as a possible regulator of Notch target gene expression. A more detailed analysis of miR-7 function will require isolation of lack of function mutations in the miR-7 gene. Lai (2002) has reported complementarity between some miRNAs and sequence elements known as K boxes, Brd boxes, and GY boxes in the 3′ UTRs of E(spl) and Brd complex genes. K boxes and Brd boxes have been implicated in posttranscriptional regulation, though no function was assigned to the miR-7 complementary GY boxes (Lai and Posakony 1997; Lai et al. 1998). The presence of GY boxes in several E(spl) and Brd complex genes, as well as in hairy and extramachrochatae has been reported (Lai and Posokony 1998). Based on the presence of these boxes, Lai (2002) predicted miR-7 target sites in HLHm3 and in Tom. We extend these predictions to a much larger gene family and provide experimental validation for some of them, indicating that GY boxes participate in the regulation of Notch target genes. miR-2 Regulates Proapoptotic Genes The proapoptotic genes reaper and grim were among the top predictions for miR-2a and miR-2b (see Table 2). reaper, grim, and the third proapoptotic gene sickle are clustered in the genome and show blocks of high conservation in their 3′ UTRs, which include the miR-2 family target sites (underlined in Figure 4A). Alignment of these sites shows a very similar pattern of predicted miRNA binding (Figure 4B). The corresponding sites are highly similar in D. pseudoobscura, but the orthologous genes cannot be identified in Anopheles. To evaluate these predictions, we made 3′ UTR sensor transgenes for reaper, grim, and sickle. The expression level of the reaper 3′ UTR sensor transgene was too low to be reliably studied in transgenic flies, so we used an in vitro assay to assess its function. Drosophila Schneider S2 cells express the miR-2 family of miRNAs (Lagos-Quintana et al. 2001). S2 cells were transfected with the reaper 3′ UTR construct or with a version of the construct in which the miR-2-binding site was mutated (the residues shown in Figure 4B were replaced by a NotI site). A low level of GFP expression was detected in immunoblots of cells transfected with the reaper 3′ UTR construct (Figure 4C, lane 2). The level of GFP expression was much higher in cells transfected with the mutated UTR construct, suggesting that the endogenous miR-2 family miRNAs in S2 cells can repress expression of a reporter construct via the reaper 3′ UTR. The grim and sickle 3′ UTR sensor transgenes were expressed at detectable levels in transgenic flies and were both downregulated by expression of miR-2b in the wing disc (Figure 4D and 4E). The miR-13 family is similar in sequence to the miR-2 family. Experimental validation will be needed to determine whether reaper, grim, and sickle are also regulated by miR-13. Identification of reaper, grim, and sickle as targets suggests that miR-2 family miRNAs may be involved in control of apoptosis. Statistical Evaluation of Target Predictions Although a number of the top-ranking sites identified in our screen have been experimentally validated, we wanted to assess the likelihood that sites with equivalent scores can be found by chance. To do so, we calculated E-values for the bantam miRNA based on the tail of the cumulative distribution of ΔG values for 10,000 random matches. An E-value predicts the number of background matches with a similar or better score (E-values scale with database size and are applicable to any distribution profile). Values greater than 1 are not significant, while those close to 0 are very significant. The results are presented on a logarithmic scale for UTRs containing one, two, or four sites of a given ΔG value (Figure 5). The best single bantam site in the hid UTR had an E-value of 1.3. This means that background matches reach RNA–duplex energies similar to the best sites, even in the smaller conserved 3′ UTR database. Indeed, target sites predicted using shuffled bantam miRNA sequences give folding energy distributions very similar to the native sequence (data not shown). Although single sites are not statistically significant, the presence of multiple sites within a single UTR can greatly increase the significance of the prediction. Combining the three bantam sites (Z > 3) predicted in the hid 3′ UTR gives an E-value of 1.8 × 10−5. Some single sites are sufficient to mediate regulation by a miRNA; however, we emphasize that the lack of statistical significance for even the best single site means that they require experimental validation. Additional Validation by Cross-Genome Comparison One means to improve the significance of the predictions would be to require conservation in a third genome. The two Drosophila species are separated by an estimated 30 million years. The mosquito A. gambiae is separated from Drosophila by 250 million years. Orthologous mosquito genes have been defined for approximately two-thirds of Drosophila genes; however, systematic comparison showed great differences in length between orthologous gene pairs (Zdobnov et al. 2002). Indeed, we were able to identify orthologous last exons with confidence for only half of these pairs, or one-third of D. melanogaster genes. We have therefore chosen to use conservation in Anopheles to provide more stringent evaluation of target site conservation, instead of requiring it generally. The presence of a conserved site with a high Z score across all three genomes increases the confidence that the site is functional. To illustrate the utility and limitations of this, we examined the top 100 predictions for miR-7 and miR-2. The Anopheles orthologs were identified for 52 of the top 100 predicted miR-7 targets. Of these, 11 had conserved target sites (Z ≥ 3), including four of the top ten predictions: hairy, Tom, m4, and CG14989 (see Table 2). For miR-2a, forty of the top hundred predictions had a detectable ortholog in Anopheles. Of these sites, five were conserved in Anopheles (Z ≥ 3), and none of these were among the top ten predictions. Conservation in Anopheles can be used to enrich for sites with a higher probability of being valid, but increases the risk of missing real targets. It is only useful in cases where the orthologous UTR region can be identified, which, for example, is not the case for the validated miR-2a targets grim, reaper, and sickle. miR-277: A Metabolic Switch? Table 3 shows predicted miR-277 targets that are conserved (Z ≥ 3) in Anopheles. Of the top 11, seven are enzymes involved in branched chain amino acid degradation (Figure 6). At more relaxed stringency (Z ≥ 2), two additional enzymes were identified (Figure 6) along with a number of unrelated loci. This striking clustering of functionally related enzymes suggests that miR-277 regulates the pathway for valine, leucine, and isoleucine degradation by downregulating many of its enzymes and thus acts as a metabolic switch. The degradation of these essential amino acids is presumably regulated under conditions of starvation or excess dietary intake. miR-277 expression has so far only been detected in adult flies (Aravin et al. 2003; Lai et al. 2003), suggesting a role in regulating metabolic responses to environmental conditions. Interestingly, the human homolog of CG8199 is mutated in maple syrup urine disease. It remains to be determined whether these enzymes are regulated by miRNAs in vertebrates. Features Shared by Validated Targets Comparison of the five previously validated targets and the six new targets validated here revealed three features shared by all sites. First, cross-genome comparison showed perfect sequence identity in the target site residues that basepair with residues 2–8 of the miRNA (see Figure 1B). This was also true for the newly validated target sites (data not shown). Second, the pattern of basepairing between the miRNAs and their targets, shown in Figure 2A, suggested that a continuous helix of at least six of the first eight basepairs might be required (allowing G:U basepairs). This was also true for the newly validated target sites (see Figures 3B and 4B). Third, many transcripts in the D. melanogaster genome overlap other transcripts on the same strand or on the opposite strand of the DNA. There are many examples of alternate splicing that produce alternate 3′ UTRs so that one UTR variant may include coding exons from another variant. In such cases, the basis for the sequence conservation between genomes is unclear. None of the validated sites from Drosophila overlaps coding sequence (CDS) on either strand (this feature was not examined for the C. elegans sites). Target sites that do not share these features are indicated in Table 2. These features can be used to increase the stringency of the screen, by discounting sites that differ from validated targets. For miR-7 this would eliminate two of the top ten predictions so that the validated targets would constitute three of the remaining top eight predictions. For miR-2a this would eliminate four of the top ten predictions, so that the validated targets reaper and grim would rank in positions 2 and 6. We have chosen not to implement the flags as filters to exclude predictions because they are based on a relatively small set of 11 validated targets, although we note that all nine predicted miR-277 targets would pass such a filter. When more targets are validated, we will learn whether these features have a general predictive value. Discussion One of the major limitations in studying animal miRNA function is the difficulty in identifying miRNA targets. Our screening strategy has proven to be useful for predicting new miRNA targets. Three new targets have been experimentally validated for miR-7 and for miR-2, bringing the total number of validated targets of animal miRNAs to 11. In addition, we predict a number of miRNA–target pairs or target families that seem likely to be valid, but require experimental validation. Our study depended on the high-quality annotation of the D. melanogaster genome and the availability of the D. pseudoobscura genome sequence. Where possible, we have extended the analysis to include evaluation of predicted sites in the A. gambiae genome. More complete annotations of the fly and mosquito genomes, aided by cDNA sequencing projects, will increase the number of genes for which orthologous UTR sequences can be defined. This will permit more sensitive and more extensive cross-genome comparison. We also expect improvements to come from further knowledge of the structural requirements of miRNA-target pairing. Evaluation of Target Predictions In designing the screening strategy, we considered the balance between sensitivity and specificity. We chose a search strategy that was based on the known examples, but generalized to allow detection of similar targets. By doing so, we risk missing fewer valid targets at the expense of including more false positives, as indicated by our statistical analysis (see Figure 5). To help distinguish false positives from potentially valid targets, we identify features shared by valid targets and, where possible, test predictions for conservation in a third, more distantly related, genome. Both positive and negative results in tests of new predictions will provide a better understanding of how miRNAs bind their targets, perhaps highlighting positions that are particularly critical. This may permit us to achieve both high sensitivity and high specificity in target prediction. Complete tables of target site predictions are available as Dataset S1 and at http://www.miRNA.embl.de. These tables report Z scores and sequences for the D. melanogaster and D. pseudoobscura target sites and, where possible, for the Anopheles target site. The tables contain flags to identify sites that share the features described above. We recommend using these flags to filter the predictions, but note that this may exclude valid targets. (Filtered tables containing the top 100 predictions are available as Dataset S2 and at http://www.miRNA.embl.de) We recommend making use of the Anopheles data to discount predictions where the othologous gene is identified and the site is absent or has a low Z score (Z 50 nt were selected. Duplicate UTRs from different splice variants of the same transcript were removed. For each of the resulting 10,196 nonredundant 3′ UTRs, we mapped the last 50 amino acids of the corresponding open reading frame to the D. pseudoobscura genome sequence with TBLASTN (E ≤ 10−5; hgsc.bcm.tmc.edu/projects/Drosophila). We selected UTR matches that included the last 10 residues and had a sequence identity ≥80% or E ≤ 10−10 and compared these UTRs to the 3,000 nt downstream of the putative D. pseudoobscura ortholog with BLASTN (word size = 7; E ≤ 10,000, assuming a database the size of the whole D. pseudoobscura genome). Nonconserved nucleotides or those outside the matched regions were replaced by Ns in the D. melanogaster 3′ UTR database to produce the conserved 3′ UTR database. The D. pseudoobscura genome has not been fully assembled. This means that some D. pseudoobscura genes are located close enough to the end of a contig that the UTR sequences may be missed. 386 D. melanogaster genes mapped to the D. pseudoobscura genome less than 1 kb from a contig end; 189 mapped less than 500 nt from a contig end. UTR conservation may be underestimated for these genes. For 3,564 genes, we did not detect a suitable ortholog using this protocol. Of these, 571 are known genes; the others are predicted genes about which little is known. For the 4,662 D. melanogaster genes lacking annotated UTRs, we assumed 3′ UTRs of 2 kb after of the stop codon and built a separate database of predicted UTRs. The search for Anopheles orthologs was done using TBLASTN for the last 50 amino acids of each D. melanogaster ORF. Owing to the more extensive sequence divergence, a lower cutoff threshold was allowed (E 3: total number of sites of Z ≥ 3 found in that UTR. UTR: whether the site is an experimentally defined or predicted UTR. 5′ conservation, 5′ helix, CDS+, CDS−, and mispairing flags are described in text and in Table 2. Z(ps): Z score for the corresponding site from D. pseudoobscura. It is not always possible to unambiguously identify the corresponding site from D. pseudoobscura, even though the sequence is in the database (because the genome is not assembled and some regions appear two or more times). We chose to omit ambiguous cases. This problem will disappear when the D. pseudoobscura genome sequence is assembled so that unambiguous gene assignment is possible. dG(ps): folding energy for the site in the D. pseudoobscura UTR. The folding energy can be used to determine a Z score on the same scale as the site from D. melanogaster. This allows a direct comparison of site quality. Alignment (ps): as above, for the D. pseudoobscura site. Mfold (ps): as above, for the D. pseudoobscura site. Ano-tblastn: TBLASTN score for the Anopheles ortholog (explained in the text). Z(ano): as above, for Anopheles. dG(ano): as above, for Anopheles. Aln(ano): as above, for Anopheles. Mfold(ano): as above, for Anopheles. Click here for additional data file. Dataset S2 Top 100 List This dataset presents our recommended filtering of the lists. Only sites in experimentally validated UTRs were included. The 5′ conservation, 5′ helix, CDS+, and CDS− flags were used to discard sites that we consider less likely to be valid. This allows use of a lower Z score cutoff (Z ≥ 2). Sites of Z < 2 were removed and Z UTR includes all sites Z ≥ 2. Columns are as above, except the flags have been removed. Click here for additional data file. Accession Numbers The accession numbers for the miRNAs discussed in this paper are bantam (AJ550546, Rfam MI0000387), let-7 (NR_000938), lin-4 (NR_000799), miR-2a (RF00047, AJ421757), miR-4 (AJ421762), miR-7 (AJ421767), miR-9 (AJ421769), miR-11 (AJ421771), miR-13a (AJ421773), miR-14 (AJ42177), and miR-277 (Rfam MI0000360). The accession numbers for the target genes discussed in this paper are CG1140 (NM_167928), CG1673 (NM_132656), CG5599 (NM_132772), CG8199 (NM_141648), CG15093 (NM_166306), CG17896 (NM_130489), dpld (NM_080033), Drice (NM_079827), grim (NM_079413), hairy (NM_079253), hairy (D. simulans) (AY055843), hairy (T. castaneum) (AJ457831), hid (NM_079412), HLHm3 (NM_079785), lin-14 (NM_077516), lin-28 (NM_059880), lin-41 (NM_060087), lin-57 (NM_076575), Lyra (NM_080079), m4 (NM_079786), reaper (NM_079414), scu (NM_078672), sickle (AF460844), and Tom (NM_079349).

0 comments Cited 199 times – based on 0 reviews      Review now

Bookmark

Record: found
Abstract: found
Article: not found

Regulation of microRNA function in somatic stem cell proliferation and differentiation.

Archana Shenoy, Robert Blelloch (2014)

microRNAs (miRNAs) are important modulators of development. Owing to their ability to simultaneously silence hundreds of target genes, they have key roles in large-scale transcriptomic changes that occur during cell fate transitions. In somatic stem and progenitor cells--such as those involved in myogenesis, haematopoiesis, skin and neural development--miRNA function is carefully regulated to promote and stabilize cell fate choice. miRNAs are integrated within networks that form both positive and negative feedback loops. Their function is regulated at multiple levels, including transcription, biogenesis, stability, availability and/or number of target sites, as well as their cooperation with other miRNAs and RNA-binding proteins. Together, these regulatory mechanisms result in a refined molecular response that enables proper cellular differentiation and function.

0 comments Cited 153 times – based on 0 reviews      Review now

Bookmark

All references

Author and article information

Contributors

Yasuko Kitatani: Role: ConceptualizationRole: Formal analysisRole: InvestigationRole: MethodologyRole: ValidationRole: VisualizationRole: Writing – original draft

Akane Tezuka:

ORCID: http://orcid.org/0000-0002-0831-3496

Role: ConceptualizationRole: Data curationRole: InvestigationRole: MethodologyRole: VisualizationRole: Writing – original draft

Eri Hasegawa: Role: Data curationRole: Formal analysisRole: InvestigationRole: MethodologyRole: ValidationRole: VisualizationRole: Writing – original draft

Satoyoshi Yanagi:

ORCID: http://orcid.org/0000-0002-6860-4690

Role: Investigation

Kazuya Togashi: Role: Investigation

Masato Tsuji:

ORCID: http://orcid.org/0000-0001-8181-7968

Role: Data curationRole: Formal analysisRole: Validation

Shu Kondo:

ORCID: http://orcid.org/0000-0002-4625-8379

Role: Resources

Jay Z. Parrish: Role: ConceptualizationRole: Data curationRole: Formal analysisRole: Funding acquisitionRole: InvestigationRole: MethodologyRole: SupervisionRole: ValidationRole: Writing – review & editing

Kazuo Emoto:

ORCID: http://orcid.org/0000-0003-4194-801X

Role: ConceptualizationRole: Funding acquisitionRole: SupervisionRole: Writing – review & editing

Bing Ye: Role: Editor

Journal

Journal ID (nlm-ta): PLoS Genet

Journal ID (iso-abbrev): PLoS Genet

Journal ID (publisher-id): plos

Journal ID (pmc): plosgen

Title: PLoS Genetics

Publisher: Public Library of Science (San Francisco, CA USA )

ISSN (Print): 1553-7390

ISSN (Electronic): 1553-7404

Publication date (Electronic): 7 August 2020

Publication date Collection: August 2020

Volume: 16

Issue: 8

Electronic Location Identifier: e1008942

Affiliations

[1 ] Department of Biological Sciences, Graduate School of Science, University of Tokyo, Hongo, Bunkyo-ku, Tokyo, Japan

[2 ] Genetic Strains Research Center, National Institute of Genetics, Yata, Mishima, Shizuoka, Japan

[3 ] Department of Biology, University of Washington, Seattle, Washington, United States of America

[4 ] International Research Center for Neurointelligence (WPI-IRCN), University of Tokyo, Hongo, Bunkyo-ku, Tokyo, Japan

University of Michigan, UNITED STATES

Author notes

The authors declare that no conflicts of interest exist.

* E-mail: jzp2@ 123456uw.edu (JZP); emoto@ 123456bs.s.u-tokyo.ac.jp (KE)

Author information

Akane Tezuka http://orcid.org/0000-0002-0831-3496

Satoyoshi Yanagi http://orcid.org/0000-0002-6860-4690

Masato Tsuji http://orcid.org/0000-0001-8181-7968

Shu Kondo http://orcid.org/0000-0002-4625-8379

Kazuo Emoto http://orcid.org/0000-0003-4194-801X

Article

Publisher ID: PGENETICS-D-20-00327

DOI: 10.1371/journal.pgen.1008942

PMC ID: 7439810

PubMed ID: 32764744

SO-VID: 3596d125-10b4-4777-8262-e54591e22c48

License:

This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

History

Date received : 3 March 2020

Date accepted : 17 June 2020

Page count

Figures: 7, Tables: 0, Pages: 22

Funding

Funded by: MEXT Grants-in-Aid for Scientific Research on Innovative Areas

Award ID: KAKENHI 16H06456

Award Recipient :

ORCID: http://orcid.org/0000-0003-4194-801X

Kazuo Emoto

Funded by: JSPS

Award ID: KAKENHI 16H02504

Award Recipient :

ORCID: http://orcid.org/0000-0003-4194-801X

Kazuo Emoto

Funded by: AMED-CREST

Award ID: JP18gm0610014

Award Recipient :

ORCID: http://orcid.org/0000-0003-4194-801X

Kazuo Emoto

Funded by: funder-id http://dx.doi.org/10.13039/100000065, National Institute of Neurological Disorders and Stroke;

Award ID: NINDS R01 NS076614

Award Recipient : Jay Z. Parrish

This work was supported by MEXT Grants-in-Aid for Scientific Research on Innovative Areas ‘‘Dynamic regulation of brain function by Scrap & Build system’’ (KAKENHI 16H06456), JSPS (KAKENHI 16H02504), WPI-IRCN, AMED-CREST (JP18gm0610014), JST-CREST, the Strategic Research Program for Brain Sciences, Toray Foundation, Naito Foundation, Takeda Science Foundation, and Uehara Memorial Foundation to KE; and by grants from the National Institutes of Health (NINDS R01 NS076614) and a JSPS invitational fellowship to JZP. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Custom metadata

PLOS Publication Stage vor-update-to-uncorrected-proof

Publication Update 2020-08-19

Data Availability All relevant data are within the manuscript and its supporting information files.

Drosophila miR-87 promotes dendrite regeneration by targeting the transcriptional repressor Tramtrack69

Read this article at

Abstract

Author summary

Related collections

Microbiology Independent Research Journal (MIR Journal)

Most cited references 47

miRNA-mediated gene silencing by translational repression followed by mRNA deadenylation and decay.

Identification of Drosophila MicroRNA Targets

Regulation of microRNA function in somatic stem cell proliferation and differentiation.

Author and article information

Contributors

Journal

Affiliations

Author notes

Author information

Article

History

Page count

Funding

Categories

Custom metadata

Comments

Comment on this article

Similar content 70

Cited by 10

Most referenced authors 411