Abstract miRNAs play essential roles in the mechanics of gene regulation, however, on an organismal-scale, the processes in which they are deployed are not well understood. Here, we adopt an evolutionary developmental approach to study miRNA function by examining their expression throughout embryogenesis in both Caenorhabditis elegans and Drosophila melanogaster. We find that, in both species, miRNA transcriptomic shifts in a punctuated fashion during the mid-developmental transition, specifying two dominant modes of early and late expression profiles. Strikingly, late-expressed miRNAs are enriched for phylogenetic conservation and function by fine-tuning the expression of their targets, implicating a role in the canalization of cell types during differentiation. In contrast, early expressed miRNAs are inversely expressed with their targets suggesting strong target-inhibition. Taken together, our work exposes a bimodal role for miRNA function during animal development, involving late-expressed physiological roles and early expressed repressive roles. miRNAs, evolutionary and developmental biology, embryogenesis, gene regulation Introduction Consistent with an ancestral role for RNA in the origin of the cell, the modern RNA world encompasses myriad RNAs working in concert to regulate gene expression in most biological processes (Mattick 2003; Chen and Rajewsky 2007; Shomron et al. 2009; Dahan et al. 2011). miRNAs are a class of small RNAs that exert posttranscriptional regulation by controlling the abundance of target molecules by either repression or mRNA degradation (Nakahara and Carthew 2004; Fabian et al. 2010; Catalanotto et al. 2016). miRNAs tend to number ∼1–5% of the number of protein-coding genes in animals, and typically target hundreds of mRNAs each by sequence pairing of 3′-UTRs (Bartel 2009). Evidence for their functional importance comes from the observation of lethal phenotypes following their perturbation (Bushati and Cohen 2007; Friedman et al. 2009; Alvarez-Saavedra and Horvitz 2010). However, while the molecular mechanisms of miRNA processing and targeting have been characterized, as well as miRNA regulatory functions throughout the life of the organism—such as maternal deposits (Giraldez et al. 2006; Schier and Giraldez 2006), signaling (Inui et al. 2010), specification (Johnston and Hobert 2003; Cochella and Hobert 2012), commitment (Ivey et al. 2008), and differentiation (Melton et al. 2010; Pauli et al. 2011)—it remains an open question how miRNA expression and target interaction are coordinated through embryonic development. miRNAs have been proposed to have two general modes of regulating their targets (Hornstein and Shomron 2006; Peterson et al. 2009; Wu et al. 2009): 1) setting the mean expression level of their targets, and 2) fine-tuning the target expression levels to decrease stochastic noise. The interaction between the miRNA let-7 and its target the mRNA lin-41 provides an example of the first mode, where let-7 strongly reduces lin-41 expression levels to ensure proper larval development in Caenorhabditis elegans (Lee et al. 1993; Reinhart et al. 2000; Rougvie 2001). As such, their developmental gene expression profiles appear anticorrelated, since the former represses the latter. The highly conserved mir-1 is an example of the second mode. mir-1 is expressed in muscle and its perturbation results in lethality (Brennecke et al. 2005; Sokol and Ambros 2005). While muscle tissue appears to develop normally in this perturbation, evidence for its weakness is revealed in some conditions, suggesting that mir-1 buffers the expression of its targets and stabilizes embryonic development by reducing stochastic noise (Brennecke et al. 2005; Hornstein and Shomron 2006). To understand these two distinct modes of gene regulation by miRNAs, one possible approach is to study these processes in developmental systems. Evolution and development are intertwined processes: development constrains which evolutionary changes may appear, while evolution constrains which developmental processes are possible (Gould 1977). By studying mRNA expression over time, the dynamic nature of gene expression has been described (Arbeitman et al. 2002; Baugh et al. 2003; Yanai et al. 2011). In particular, one stage during embryogenesis of the nematode C. elegans was identified as punctuating development in terms of gene expression changes (Levin et al. 2012). This stage corresponds to ventral enclosure, a period known to involve integration among the germ layers. Moreover, evidence was revealed that developmental constraints act to maintain the conservation of this stage (Zalts and Yanai 2017). Broadening out to ten species, each from a different phylum, a recent gene expression analysis identified a mid-developmental transition (Levin et al. 2016) in each species, including the ventral enclosure stage in C. elegans. The period of expression before and after the mid-developmental transition interestingly distinguished programs for proliferation and differentiation. In addition, a recent analysis comparing miRNA expression across two Drosophila species, identified an hourglass-shaped pattern of conservation, in which the middevelopment transition is conserved (Ninova et al. 2014). In this work, we invoked an evolutionary developmental approach to study miRNA expression throughout embryonic development of C. elegans and D. melanogaster. This approach provides a conceptual framework for studying miRNAs and also carries two direct benefits for our approach. We expect any program important for development to be conserved across animals and thus manifested in both species. This allows for the comparison of C. elegans with D. melanogaster to reveal conserved principles of miRNA regulation. Second, a comparison of miRNAs across species allows us to study them as a function of their evolutionary ages. In contrast to mRNAs, miRNAs are thought to be fast evolving and few of them are deeply conserved along animal phylogeny (Niwa and Slack 2007; França et al. 2016). Therefore, by assessing their evolutionary conservation, we can infer the age of a miRNA and distinguish between miRNAs that are conserved and those that are specific to the proximal phylogenetic branch of each species (França et al. 2016). We report here two main programs of miRNA function throughout development. These correspond to two opposing periods of expression throughout embryogenesis, namely early and late, relative to the mid-developmental transition (Levin et al. 2016). Interestingly, conserved miRNAs are expressed late in development and appear to fine-tune the expression of their targets. In contrast, early expressed miRNAs are young, that is, restricted in their presences to their respective genus, and regulate late expressing targets, suggesting that they act as repressors. Collectively, our results reveal a bimodal deployment of miRNA regulation throughout animal development. Results To study miRNA and their targets in both C. elegans and D. melanogaster, we isolated embryos at distinct developmental stages during embryogenesis (fig. 1A). The developmental stages for analysis were selected with the mid-developmental transition occupying the midpoint among our samples (Levin et al. 2016). From each sample, we extracted RNA and processed a fraction of it using CEL-Seq (Hashimshony et al. 2012) for mRNA RNA-Seq and the rest using a protocol designed for small RNA (fig. 1A, see Materials and Methods section). We detected expression of 50% and 48%, of the 244 and 260 annotated miRNA precursors in the worm and fly, respectively, with an average of 373, 570, and 38, 262 detected transcripts per sample of worm and fly (supplementary fig. S1A and B, Supplementary Material online). Of these, 75% and 71%, respectively, had a greater than 10-fold change in expression throughout the examined stages of development, indicating that they are dynamically expressed. We assayed the quality of the miRNA transcriptomes by examining correlation coefficients among replicates and recapitulation of known profiles for previously studied genes. The correlation coefficients among replicates was high for both species (R > 0.85, C. elegans and R = 0.88, D. melanogaster, supplementary fig. S1C, Supplementary Material online). We also compared our expression profiles for 3 and 46 C. elegans genes with those of Wu et al. (2010) and Stoeckius et al. (2009), respectively, and detected good correspondence (supplementary table S1, Supplementary Material online). For 19 D. melanogaster miRNAs analyzed by Ninova et al. (2014), we detected consistent expression for all but 4 miRNAs (supplementary table S1, Supplementary Material online). Three of these genes showed similar expression suggesting that the difference in expression stemmed from higher temporal resolution of our data. While miR-2a-1 was detected as early expressed in our data, Ninova et al. reported miR-2a-1 as expressed during morphogenesis, and the difference may be attributed to this miRNA low and noisy expression. Fig. 1. View largeDownload slide Comparing the developmental miRNA and mRNA transcriptomes of Caenorhabditis elegans and Drosophila melanogaster. (A) Embryos at the indicated stages of C. elegans and D. melanogaster embryogenesis, represented by red dots, were collected and processed for small-RNA and mRNA transcriptome analysis. The developmental stage and time of collection are indicated for each sample. (B) Heatmaps represent developmental gene expression profiles for miRNAs and mRNAs. Each row indicates the expression profile for a particular gene across the stages (columns). Expression was standardized (see Materials and Methods section) and sorted by time of expression (Levin et al. 2016; Zalts and Yanai 2017). Fig. 1. View largeDownload slide Comparing the developmental miRNA and mRNA transcriptomes of Caenorhabditis elegans and Drosophila melanogaster. (A) Embryos at the indicated stages of C. elegans and D. melanogaster embryogenesis, represented by red dots, were collected and processed for small-RNA and mRNA transcriptome analysis. The developmental stage and time of collection are indicated for each sample. (B) Heatmaps represent developmental gene expression profiles for miRNAs and mRNAs. Each row indicates the expression profile for a particular gene across the stages (columns). Expression was standardized (see Materials and Methods section) and sorted by time of expression (Levin et al. 2016; Zalts and Yanai 2017). To study the dynamics of miRNA and mRNA gene expression programs, we first sorted the profiles according to time of expression (Levin et al. 2016; Zalts and Yanai 2017). We found that miRNA profiles are qualitatively similar to the patterns of mRNAs, in that temporally restricted patterns of expression were detected in both (fig. 1B). Overall, one-third of both the miRNAs and mRNAs are dynamic in both C. elegans and D. melanogaster (fig. 1B). We analyzed the global pattern of miRNA expression using principal components analysis on the miRNA profiles (fig. 2A). For both species, we found that PC1 captures the temporal order of the developmental stages (fig. 2A). Strikingly, PC2 distinguished the middle of development (Stage 4) from the earlier and later stages, also in both species. Interestingly, this stage maps to the mid-developmental transition (Levin et al. 2016) (fig. 2B), which corresponds to the ventral enclosure stage and following the extended germ band stage in C. elegans and D. melanogaster, respectively (Levin et al. 2016). Fig. 2. View largeDownload slide miRNA turnover occurs at the mid-developmental transition. (A) Principal components analysis (PCA) on the developmental stages of Caenorhabditis elegans (left) and Drosophila melanogaster (right). Each circle represents the miRNA transcriptome at a developmental stage. (B) Principal component 2 score for the stages is shown below the stages. (C) The miRNA rate of expression change between stages is plotted, where between each pair of subsequent stages the transcriptomic distance between them is divided by the number of elapsed minutes. Error bar indicate SE. Fig. 2. View largeDownload slide miRNA turnover occurs at the mid-developmental transition. (A) Principal components analysis (PCA) on the developmental stages of Caenorhabditis elegans (left) and Drosophila melanogaster (right). Each circle represents the miRNA transcriptome at a developmental stage. (B) Principal component 2 score for the stages is shown below the stages. (C) The miRNA rate of expression change between stages is plotted, where between each pair of subsequent stages the transcriptomic distance between them is divided by the number of elapsed minutes. Error bar indicate SE. The transcriptional changes at the mid-developmental transition suggest that the dynamics are different at this stage. This was further evident when we studied the rate of miRNA expression changes by computing the distance between subsequent transcriptomes and normalizing by the number of elapsed minutes. Consistent with the PC analysis, we detected a different expression rate across time, with the transition from stage 3 to 4 amounting to the highest rate among all stages (fig. 2C). This suggests that at stage 4 many miRNAs are both up- and down-regulated and collectively amounts to the largest change of rates throughout embryogenesis. We next sought to examine the miRNA expression profiles that lead to the dynamic patterns we observed in figure 2. While many possible clusters of expression profiles are consistent with the patterns of correlations, we surprisingly detected only two major clusters accounting for most of the miRNAs (fig. 3A). We observed that the miRNA expression is clearly organized into two modules, corresponding to early and late expression, in both analyzed species (fig. 3B). Approximately 30% of the detected miRNAs shows an early profile, that generally includes expression in the first four stages. The late profile contains roughly 50% of miRNA in C. elegans and in D. melanogaster. Since many of the miRNAs are grouped into seed families and tend to cluster together in the genome, we examined the expression profiles of miRNAs that correspond to the same miRNA seed family (Kozomara and Griffiths-Jones 2014) and are clustered in the same genomic location (França et al. 2016) (supplementary fig. S3A and B, Supplementary Material online). We found that most miRNA seed families have similar expression profile (supplementary fig. S3A, Supplementary Material online), however miRNA from the same genomic cluster do not always present the similar expression profile, and thus should not be accounted as a cotranscribed group (supplementary fig. S3B, Supplementary Material online). Thus, to a first approximation, miRNA expression in embryogenesis can be classified into pre- or post- mid-developmental transition expression. Fig. 3. View largeDownload slide miRNA expression profiles show two modules of expression. (A) Heatmaps showing the correlations among the miRNA expression profiles of Caenorhabditis elegans (up) and Drosophila melanogaster (bottom). miRNA were ordered using hierarchical clustering of the pairwise distance that was calculated based on the Pearson correlation between miRNA log expression values. (B) miRNA expression profiles sorted as in a. miRNA from the same family are indicated by the markers along with the name of the C. elegans orthologue. The bar on the side indicates the assigning of miRNA to early (red) or late (green) expression module. (C) Expression profiles of three selected miRNA families with C. elegans (blue) and D. melanogaster (red) orthologues (see Materials and Methods section). The miRNA name is indicated above each profile. Fig. 3. View largeDownload slide miRNA expression profiles show two modules of expression. (A) Heatmaps showing the correlations among the miRNA expression profiles of Caenorhabditis elegans (up) and Drosophila melanogaster (bottom). miRNA were ordered using hierarchical clustering of the pairwise distance that was calculated based on the Pearson correlation between miRNA log expression values. (B) miRNA expression profiles sorted as in a. miRNA from the same family are indicated by the markers along with the name of the C. elegans orthologue. The bar on the side indicates the assigning of miRNA to early (red) or late (green) expression module. (C) Expression profiles of three selected miRNA families with C. elegans (blue) and D. melanogaster (red) orthologues (see Materials and Methods section). The miRNA name is indicated above each profile. Since miRNAs are not well conserved (Niwa and Slack 2007), only six miRNAs had orthologous groups according to miRBase annotations between C. elegans and D. melanogaster for a direct comparison of their expression (fig. 3C and supplementary fig. S3C, Supplementary Material online). mir-1, the muscle-specific miRNA, is expressed at the stage 4 transition in both species. In addition, mir-87 exhibits late embryogenesis expression in both species. The remaining C. elegans and D. melanogaster conserved miRNAs are expressed in the late module (supplementary fig. S3C, Supplementary Material online). Moreover, although most of the miRNAs are poorly conserved between these two species, we note that the pattern of early and late miRNA expression is a general property across these two animal phyla. As it was previously demonstrated that mRNAs are also expressed in a similar fashion during development (Levin et al. 2016), our results suggest that the crosstalk between miRNAs and their targets is coupled to the bimodal gene expression patterns during animal embryogenesis. Given that miRNAs are expressed in a bimodal manner during development, we hypothesized that the early and late modules would have distinct functional and evolutionary properties. To explore whether there is a difference in the phylogenetic conservation of early and late miRNAs, we computed for each miRNA the phylogenetic distribution across miRNA families (see Materials and Methods section and supplementary table S2, Supplementary Material online). We define miRNAs as conserved if their phylogenetic distribution extends beyond presence in the respective genus of the species. Studying the expression profiles of the conserved and nonconserved miRNAs, we found that conserved miRNAs are generally expressed late, whereas nonconserved miRNAs are expressed early during development (P <0.0062, Pearson’s correlation) in both C. elegans and D. melanogaster (fig. 4A). Thus, over evolutionary time-scales, late-expressed miRNAs appear to be more universal to animal development than early expressed miRNAs. To further study this result, we asked whether late miRNAs are enriched in increasingly conserved classes. We calculated the fraction of late and early miRNAs that were classified as “species specific,” “genus-specific,” or “phylum-specific” (see Materials and Methods section). We found that compared with the early expressed miRNAs, the late miRNAs are enriched in the “phylum” classification as compared with the early (P < 0.05 Hypergeometric cumulative distribution function) (supplementary fig. S4A, Supplementary Material online), which further supports the notion that late miRNAs have a deeper phylogenetic distribution than early miRNAs. Fig. 4. View large Download slide Evolutionary and functional properties of the miRNA modules. (A) Boxplots indicate the expression levels of nonconserved (gray), and conserved (black) miRNA for each developmental stage (see Materials and Methods section) for Caenorhabditis elegans (left) and Drosophila melanogaster (right). The plotted line indicates the median of expression levels of the nonconserved (gray), and conserved (black) miRNA. (B) The frequencies of correlation between miRNA and their mean targets are plotted for all miRNAs (blue), early miRNAs (red), and late miRNAs (green) for C. elegans (left) and D. melanogaster (right). The asterisk indicates P < 0.005. (C) Gene Ontology (GO) terms of early and late targets (see Materials and Methods section) for C. elegans (left) and D. melanogaster (right). The colored squares on the left indicate functional classifications of GO terms (supplementary fig. S4, Supplementary Material online): Transcription (orange), Signaling (turquoise), Morphology (purple), Cell division (pink), other (gray). Fig. 4. View large Download slide Evolutionary and functional properties of the miRNA modules. (A) Boxplots indicate the expression levels of nonconserved (gray), and conserved (black) miRNA for each developmental stage (see Materials and Methods section) for Caenorhabditis elegans (left) and Drosophila melanogaster (right). The plotted line indicates the median of expression levels of the nonconserved (gray), and conserved (black) miRNA. (B) The frequencies of correlation between miRNA and their mean targets are plotted for all miRNAs (blue), early miRNAs (red), and late miRNAs (green) for C. elegans (left) and D. melanogaster (right). The asterisk indicates P < 0.005. (C) Gene Ontology (GO) terms of early and late targets (see Materials and Methods section) for C. elegans (left) and D. melanogaster (right). The colored squares on the left indicate functional classifications of GO terms (supplementary fig. S4, Supplementary Material online): Transcription (orange), Signaling (turquoise), Morphology (purple), Cell division (pink), other (gray). We hypothesized that the difference in the temporal mode of expression between nonconserved and conserved miRNAs reflects their functional attributes. As was previously proposed (Hornstein and Shomron 2006; Wu et al. 2009), miRNAs can act to repress or buffer the expression of their targets. In order to distinguish between these two functional modes, we tested the correlation among the expression profiles of miRNAs and their predicted targets (see Materials and Methods section). While overall no correlation was observed when all miRNAs and their targets were considered together (fig. 4B, blue), when we distinguished miRNAs according to their expression, the early miRNAs showed a negative correlation with their targets whereas the late did not show any particular bias (P < 0.005, Wilcoxon rank-sum test, measure of effect size, >0.7, Hedges’ g; Hentschke and Stüttgen 2011) (fig. 4B). This pattern was not biased by the number of targets of each miRNA which shows no statistical difference between early and late miRNAs (supplementary fig. S4B, Supplementary Material online). We observed this same for both C. elegans and D. melanogaster, suggesting that early miRNAs mostly exert a strong repressive function on their targets compared with late miRNAs. To test the hypothesis that early and late miRNAs employ different repression strategies on their targets, we examined the expression profiles of miRNA and their known targets. We found that the miR-35 family has an early expression profile and the egl-1 target gene has a late expression profile (Sherrard et al. 2017), whereas the miR-51 family and lsy-6 miRNA both exhibit a late expression profile as do their targets, the cdh-3 and cog-1 genes, respectively (Johnston and Hobert 2003; Shaw et al. 2010) (supplementary fig. S4C, Supplementary Material online). Finally, we examined several miRNAs with characterized organ specific expression patterns: miR-124 in neurons (Clark et al. 2010), miR-236 in the intestine (Martinez et al. 2008) and miR-255 in the pharynx (Martinez et al. 2008). For each of these, we detected a late expression profile, supporting the notion that these miRNAs fine-tune their target during late stages of development (supplementary fig. S4D, Supplementary Material online). To further explore the potential functional properties of early and late expressed miRNAs, we performed Gene Ontology (GO) enrichment analysis on miRNA target genes (see Materials and Methods section). The targets of early miRNAs were not found to be highly enriched in different GO terms, as compared with the targets of late miRNAs. However, the targets of late expressed miRNAs in both species were strongly enriched for functions such as transcription, cell division, and morphology (fig. 4C and supplementary fig. S4E, Supplementary Material online). The morphology GO term in particular suggests that the late-expressed miRNAs fine-tune processes occurring during differentiation. Together with the conservation and target-profile analyses (fig. 4A and B), we conclude that the early and late miRNA modules are characterized by distinct functional and evolutionary properties. Discussion In this work, we studied the expression of miRNAs in C. elegans and D. melanogaster throughout embryogenesis, together with the expression of mRNAs. We report that miRNA expression is punctuated by the mid-developmental transition analogously to previously patterns observed for mRNA (Levin et al. 2016). In both species, we detected two canonical miRNA expression profiles, an early and a late, relative to the mid-developmental transition. For each mode of expression, we report different evolutionary and functional properties in terms of their evolutionary age, mode of repression and regulatory function (fig. 4). Our work has implications for miRNA function, evolutionary conservation, and also provides a general model for miRNA programs throughout development. Two regulatory modes—inhibitory and buffering—have been ascribed to miRNAs (Hornstein and Shomron 2006; Wu et al. 2009). We show evidence that these functions are deployed differentially throughout development. By studying the expression profiles of the targets of the early and late miRNAs, we found that early miRNAs have profiles that are inversely correlated to those of their mRNA targets. In contrast, late-expressed miRNAs appear to fine-tune the expression of their targets as they generally have correlated expression profiles (fig. 4B). Another explanation is that the late miRNAs repress the translation of their targets (Baek et al. 2008; Wilczynska and Bushell 2015) which would also appear as correlated expression profiles in our data. A second alternative explanation is that among the targets of a miRNA, there may be a subset of targets that are repressed (Garcia et al. 2011). The bimodal miRNA expression profiles we detect may be attribute to distinct cellular pathways. For example, in C. elegans there are two Argonautes genes that were previously shown to have spatiotemporal differential expression and different miRNA association during embryogenesis (Vasquez-Rifo et al. 2012). Thus, one underlying mechanism for the early and late expression profiles may be that each is processed by its own dedicated Argonaute. Together, these observations are consistent with the notion that miRNA function to control mRNAs that are active during the end of development: early expressed miRNAs prevent their targets expression in the first half of embryogenesis while late-expressed miRNAs fine-tune their expression. Although most miRNAs are not thought to have a deep evolutionary history (Niwa and Slack 2007), we found that the two programs show disparate levels of conservation: nonconserved miRNA are early expressed, while conserved miRNAs have a late expression profile. Further support for the conservation of the late-expressed miRNAs come from the functions of their targets which we found to be enriched for “core” functions such as transcription and cell division (fig. 4C). To account for the role of the early expressed miRNAs, we propose that they function to repress evolutionary novelties. Under this scenario, the early expressed miRNAs arose to repress newly evolved developmental programs, generally expressed at the end of development. Interestingly our characterization of miRNAs into temporal modes with distinct conservation and functional properties holds for both C. elegans and D. melanogaster, suggesting that the identified programs may be general to animal development. Based upon our results, we propose a model by which miRNAs do not regulate early stages in development but rather are required later for physiological function to maintain the integrity of cell types. The specific function of miRNAs may be to “lock-in” a particular physiological state characteristic of a cell type to prevent disruption and transition to another cell type. One prediction that follows from this model is higher levels of late miRNAs with increasing cell-type complexity across organisms. This prediction is consistent with our results that both early and late miRNA act to regulate late mRNA expression, and that there are more late miRNAs than early miRNAs (fig. 3). Collectively, our work suggests that the main developmental role of miRNAs is to canalize (Waddington 1957) the differentiating cell types occurring throughout embryogenesis. Materials and Methods Embryo Collection and Sample Preparation Caenorhabditis elegans embryos were isolated from cut N2 adult worms using mouth pipette (Baugh et al. 2003) and staged at eight developmental time-points (60, 120, 180, 225, 270, 330, 410, and 500 min after the 4 cell stage). About 50 embryos were collected at each time-point, in triplicates. From each sample, total RNA was extracted using TRIzol and eluted in 1.5 μl of RNase-free water of which 1 μl was used for Small-RNA library preparation. The remaining 0.5 μl were used for RNA-Seq using the CEL-Seq protocol (Hashimshony et al. 2012), processed and analyzed as previously described (Hashimshony et al. 2014). For D. melanogaster, we used the RNA from a previous study (Levin et al. 2016) with the following samples IDs (DM0011, DM0014, DM0023, DM0022, DM0039, DM0036, DM0052, DM0038, DM0062, DM0066, DM0078, DM0081, DM0083, DM0089) (Levin et al. 2016). We selected seven developmental time-points (160, 300, 530, 645, 930, 1,160, 1,300 min), and considered the two closest samples as duplicates, adjusted the RNA from each sample to 1 ng/μl for the Small-RNA library preparation. Small-RNA Library Preparation Illumina TruSeq Small-RNA kit (15004197) was used with the following modifications: 3′ and 5′ ligations were performed at a quarter of a reaction volume, with 1 µl of RNA and RA3 diluted 1:5. Reverse transcription was performed at half of a reaction volume using the entire ligation product. PCR reaction was performed at half of a reaction volume with a total of 15 cycles of PCR. Libraries were sequenced on the Illumina HiSeq2500 according to the standard 50-bp protocol. miRNA Analysis The 50-bp reads were trimmed using Scythe to remove the 3′ adapter sequence. For each sample, reads were mapped to the reference genome (WS230 for C. elegans and dm6 for D. melanogaster) using bowtie with default parameters. Using HTSeq-count, annotated miRNA (based on miRBase annotation; Kozomara and Griffiths-Jones 2014) were counted and normalized by dividing by the total number of counted reads and multiply by million (tpm). Expression Analysis We selected expressed miRNA by the highly expressed mature miRNA (5p or 3p), with at least two time-point with expression level of >10tpm. Mean expression of the replicates was calculates based on the normalized TPM expression. For the mRNA expression, we selected all genes with sum expression >10tpm. Using the ZAVIT method (Levin et al. 2012; Zalts and Yanai 2017) we standardize the gene expression of each gene and temporally arranged their expression. Stages Comparison Pearson correlation between the developmental stages was calculated based on the log values the mean expressed miRNAs. Principle component analysis (PCA) was performed on standardized mean expressed miRNAs. PC2 values were linearly interpolated taking into account the difference between the time-points. Rate of expression change was calculated based on the pairwise distance between each pair of subsequent stages and divided by the time between the two stages. miRNA Conservation Each species in the miFam miRBase database got a score indicating the phylogenetic distance from the species (C. elegans or D. melanogaster). 1) Species specific (C. elegans or D. melanogaster). 2) Genus (Caenorhabditis or Drosophila). 3) Phylum (Nematode or Arthropoda) and higher (supplementary table S2, Supplementary Material online). The conservation score was assigned for each miRNA family, as the highest phylogenetic distant score. For each species, we divided the miRNA to “nonconserved” or “conserved” so that miRNA that have family members within the genus or are species specific are nonconserved, and the rest are conserved. The expression of the10 highly expressed miRNA from each group was normalized and averaged to one expression profile. For six miRNA families C. elegans and D. melanogaster had miRNA orthologous. For their comparison, we collapsed miRNA paralogs and calculated a mean expression profile. miRNA Targets Prediction, Classification, and Correlation To select the sets of target genes of early and late expressed miRNAs for C. elegans and D. melanogaster, we first used the TargetScan 6.0 algorithm to predict miRNA binding sites on the 3′-UTR sequences downloaded from www.targetscan.org and miRNA annotations were obtained from miRBase v.21. Only site types of 7mer-m8 and 8mer with individual context scores < −0.1 were included. To refine the target selection, for each gene, we calculated the proportion of binding sites corresponding to early and late miRNAs. Only those target genes with higher proportion of binding sites for either early or late miRNAs were considered. For each miRNA, we generated a mean expression profile based on the normalized expression of its targets. We calculated the Pearson correlation coefficient between each miRNA and the mean target profile. Gene Ontology For targets of late miRNA, we computed the enrichment for gene ontology (GO) terms (hypergeometric distribution) using annotations from Ensembl ([CSL STYLE ERROR: reference with no printed form.]). We examined only GO terms with more than 5 or 30 genes for C. elegans and D. melanogaster, respectively. For the top 20 GO terms that we found for the targets of late miRNA, we calculated the P value for the targets of early miRNAs. Supplementary Material Supplementary data are available at Molecular Biology and Evolution online. Acknowledgments This work was supported by the Israel Science Foundation 1457/14. We also acknowledge Dr Tamar Hashimshony and Dr Michal Levin for their help in embryo collection. References Alvarez-Saavedra E, Horvitz HR. 2010. Many families of C. elegans MicroRNAs are not essential for development or viability. Curr Biol . 20( 4): 367– 373. http://dx.doi.org/10.1016/j.cub.2009.12.051 Google Scholar CrossRef Search ADS PubMed Arbeitman MN, Furlong EEM, Imam F, Johnson E, Null BH, Baker BS, Krasnow MA, Scott MP, Davis RW, White KP. 2002. Gene expression during the life cycle of Drosophila melanogaster. Science 297( 5590): 2270– 2275. Google Scholar CrossRef Search ADS PubMed Baek D, Villén J, Shin C, Camargo FD, Gygi SP, Bartel DP. 2008. The impact of microRNAs on protein output. Nature 455( 7209): 64– 71. Google Scholar CrossRef Search ADS PubMed Bartel DP. 2009. MicroRNAs: target recognition and regulatory functions. Cell 136( 2): 215– 233. http://dx.doi.org/10.1016/j.cell.2009.01.002 Google Scholar CrossRef Search ADS PubMed Baugh LR, Hill AA, Slonim DK, Brown EL, Hunter CP. 2003. Composition and dynamics of the Caenorhabditis elegans early embryonic transcriptome. Development 130( 5): 889– 900. Google Scholar CrossRef Search ADS PubMed Brennecke J, Stark A, Cohen SM. 2005. Not miR-ly muscular: microRNAs and muscle development. Genes Dev . 19( 19): 2261– 2264. http://dx.doi.org/10.1101/gad.1363905 Google Scholar CrossRef Search ADS PubMed Bushati N, Cohen SM. 2007. microRNA functions. Annu Rev Cell Dev Biol. 23: 175– 205. http://dx.doi.org/10.1146/annurev.cellbio.23.090506.123406 Google Scholar CrossRef Search ADS PubMed Catalanotto C, Cogoni C, Zardo G. 2016. MicroRNA in control of gene expression: an overview of nuclear functions. Int J Mol Sci . 17: 1712. Google Scholar CrossRef Search ADS Chen K, Rajewsky N. 2007. The evolution of gene regulation by transcription factors and microRNAs. Nat Rev Genet . 8( 2): 93– 103. http://dx.doi.org/10.1038/nrg1990 Google Scholar CrossRef Search ADS PubMed Clark AM, Goldstein LD, Tevlin M, Tavaré S, Shaham S, Miska EA. 2010. The microRNA miR-124 controls gene expression in the sensory nervous system of Caenorhabditis elegans. Nucleic Acids Res . 38( 11): 3780– 3793. Google Scholar CrossRef Search ADS PubMed Cochella L, Hobert O. 2012. Embryonic PRIMING Priming of a miRNA locus predetermines postmitotic neuronal left/right asymmetry in C. elegans. Cell 151( 6): 1229– 1242. http://dx.doi.org/10.1016/j.cell.2012.10.049 Google Scholar CrossRef Search ADS PubMed Dahan O, Gingold H, Pilpel Y. 2011. Regulatory mechanisms and networks couple the different phases of gene expression. Trends Genet . 27( 8): 316– 322. http://dx.doi.org/10.1016/j.tig.2011.05.008 Google Scholar CrossRef Search ADS PubMed Fabian MR, Sonenberg N, Filipowicz W. 2010. Regulation of mRNA translation and stability by microRNAs. Annu Rev Biochem . 79: 351– 379. http://dx.doi.org/10.1146/annurev-biochem-060308-103103 Google Scholar CrossRef Search ADS PubMed França GS, Vibranovski MD, Galante PAF. 2016. Host gene constraints and genomic context impact the expression and evolution of human microRNAs. Nat Commun . 7: 11438. Google Scholar CrossRef Search ADS PubMed Friedman LM, Dror AA, Mor E, Tenne T, Toren G, Satoh T, Biesemeier DJ, Shomron N, Fekete DM, Hornstein E, et al. 2009. MicroRNAs are essential for development and function of inner ear hair cells in vertebrates. Proc Natl Acad Sci U S A . 106( 19): 7915– 7920. Google Scholar CrossRef Search ADS PubMed Garcia DM, Baek D, Shin C, Bell GW, Grimson A, Bartel DP. 2011. Weak seed-pairing stability and high target-site abundance decrease the proficiency of lsy-6 and other microRNAs. Nat Struct Mol Biol . 18( 10): 1139– 1146. Google Scholar CrossRef Search ADS PubMed Giraldez AJ, Mishima Y, Rihel J, Grocock RJ, Van Dongen S, Inoue K, Enright AJ, Schier AF. 2006. Zebrafish MiR-430 promotes deadenylation and clearance of maternal mRNAs. Science 312( 5770): 75. http://dx.doi.org/10.1126/science.1122689 Google Scholar CrossRef Search ADS PubMed Gould SJ. 1977. Ontogeny and phylogeny. Cambridge (MA): Belknap Press of Harvard University Press. Hashimshony T, Feder M, Levin M, Hall BK, Yanai I. 2014. Spatiotemporal transcriptomics reveals the evolutionary history of the endoderm germ layer. Nature 519( 7542): 219– 222. Google Scholar CrossRef Search ADS PubMed Hashimshony T, Wagner F, Sher N, Yanai I. 2012. CEL-Seq: single-cell RNA-Seq by multiplexed linear amplification. Cell Rep . 2( 3): 666– 673. http://dx.doi.org/10.1016/j.celrep.2012.08.003 Google Scholar CrossRef Search ADS PubMed Hentschke H, Stüttgen MC. 2011. Computation of measures of effect size for neuroscience data sets. Eur J Neurosci . 34( 12): 1887– 1894. Google Scholar CrossRef Search ADS PubMed Hornstein E, Shomron N. 2006. Canalization of development by microRNAs. TL – 38 Suppl. Nat Genet . 38( 6s): 4. Google Scholar CrossRef Search ADS Inui M, Martello G, Piccolo S. 2010. MicroRNA control of signal transduction. Nat Rev Mol Cell Biol . 11( 3): 264– 275. http://dx.doi.org/10.1038/nrm2868 Google Scholar PubMed Ivey KN, Muth A, Arnold J, King FW, Yeh R-F, Fish JE, Hsiao EC, Schwartz RJ, Conklin BR, Bernstein HS, et al. 2008. MicroRNA regulation of cell lineages in mouse and human embryonic stem cells. Cell Stem Cell 2( 3): 219– 229. Google Scholar CrossRef Search ADS PubMed Johnston RJ, Hobert O. 2003. A microRNA controlling left/right neuronal asymmetry in Caenorhabditis elegans. Nature 426( 6968): 845– 849. http://dx.doi.org/10.1038/nature02255 Google Scholar CrossRef Search ADS PubMed Kozomara A, Griffiths-Jones S. 2014. miRBase: annotating high confidence microRNAs using deep sequencing data. Nucleic Acids Res. 42(Database issue): D68– D73. Google Scholar CrossRef Search ADS Lee RC, Feinbaum RL, Ambros V. 1993. The C. elegans heterochronic gene lin-4 encodes small RNAs with antisense complementarity to & II-14. Cell 75( 5): 843– 854. http://dx.doi.org/10.1016/0092-8674(93)90529-Y Google Scholar CrossRef Search ADS PubMed Levin M, Anavy L, Cole AG, Winter E, Mostov N, Khair S, Senderovich N, Kovalev E, Silver DH, Feder M, et al. 2016. The mid-developmental transition and the evolution of animal body plans. Nature 531( 7596): 637– 641. Google Scholar CrossRef Search ADS PubMed Levin M, Hashimshony T, Wagner F, Yanai I. 2012. Developmental milestones punctuate gene expression in the Caenorhabditis embryo. Dev. Cell 22( 5): 1101– 1108. Google Scholar CrossRef Search ADS PubMed Martinez NJ, Ow MC, Reece-Hoyes JS, Barrasa MI, Ambros VR, Walhout AJM. 2008. Genome-scale spatiotemporal analysis of Caenorhabditis elegans microRNA promoter activity. Genome Res . 18( 12): 2005– 2015. Mattick JS. 2003. Challenging the dogma: the hidden layer of non-protein-coding RNAs in complex organisms. BioEssays 25( 10): 930– 939. http://dx.doi.org/10.1002/bies.10332 Google Scholar CrossRef Search ADS PubMed Melton C, Judson RL, Blelloch R. 2010. Opposing microRNA families regulate self-renewal in mouse embryonic stem cells. Nature 463( 7281): 621– 626. http://dx.doi.org/10.1038/nature08725 Google Scholar CrossRef Search ADS PubMed Nakahara K, Carthew RW. 2004. Expanding roles for miRNAs and siRNAs in cell regulation. Curr Opin Cell Biol . 16( 2): 127– 133. http://dx.doi.org/10.1016/j.ceb.2004.02.006 Google Scholar CrossRef Search ADS PubMed Ninova M, Ronshaugen M, Griffiths-Jones S. 2014. Conserved temporal patterns of microRNA expression in Drosophila support a developmental hourglass model. Genome Biol Evol . 6( 9): 2459– 2467. http://dx.doi.org/10.1093/gbe/evu183 Google Scholar CrossRef Search ADS PubMed Niwa R, Slack FJ. 2007. The evolution of animal microRNA function. Curr Opin Genet Dev . 17( 2): 145– 150. http://dx.doi.org/10.1016/j.gde.2007.02.004 Google Scholar CrossRef Search ADS PubMed Pauli A, Rinn JL, Schier AF. 2011. Non-coding RNAs as regulators of embryogenesis. Nat Rev Genet . 12( 2): 136– 149. http://dx.doi.org/10.1038/nrg2904 Google Scholar CrossRef Search ADS PubMed Peterson KJ, Dietrich MR, McPeek MA. 2009. MicroRNAs and metazoan macroevolution: insights into canalization, complexity, and the Cambrian explosion. BioEssays 31( 7): 736– 747. http://dx.doi.org/10.1002/bies.200900033 Google Scholar CrossRef Search ADS PubMed Reinhart BJ, Slack FJ, Basson M, Pasquinelli a. E, Bettinger JC, Rougvie a. E, Horvitz HR, Ruvkun G. 2000. The 21-nucleotide let-7 RNA regulates developmental timing in Caenorhabditis elegans. Nature 403( 6772): 901– 906. Google Scholar CrossRef Search ADS PubMed Rougvie A. 2001. Control of developmental timing in animals. Nat Rev Genet . 2( 9): 690– 701. http://dx.doi.org/10.1038/35088566 Google Scholar CrossRef Search ADS PubMed Schier AF, Giraldez AJ. 2006. MicroRNA function and mechanism: insights from zebra fish. Cold Spring Harb Symp Quant Biol . 71: 195– 203. http://dx.doi.org/10.1101/sqb.2006.71.055 Google Scholar CrossRef Search ADS PubMed Shaw WR, Armisen J, Lehrbach NJ, Miska EA. 2010. The conserved miR-51 microRNA family is redundantly required for embryonic development and pharynx attachment in Caenorhabditis elegans. Genetics 185( 3): 897– 905. http://dx.doi.org/10.1534/genetics.110.117515 Google Scholar CrossRef Search ADS PubMed Sherrard R, Luehr S, Holzkamp H, McJunkin K, Memar N, Conradt B. 2017. miRNAs cooperate in apoptosis regulation during C. elegans development. Genes Dev . 31( 2): 209– 222. Google Scholar CrossRef Search ADS PubMed Shomron N, Golan D, Hornstein E. 2009. An evolutionary perspective of animal MicroRNAs and their targets. J Biomed Biotechnol . 14: 1– 9. Sokol NS, Ambros V. 2005. Mesodermally expressed Drosophila microRNA-1 is regulated by Twist and is required in muscles during larval growth. Genes Dev . 19( 19): 2343– 2354. http://dx.doi.org/10.1101/gad.1356105 Google Scholar CrossRef Search ADS PubMed Stoeckius M, Maaskola J, Colombo T, Rahn H-P, Friedländer MR, Li N, Chen W, Piano F, Rajewsky N. 2009. Large-scale sorting of C. elegans embryos reveals the dynamics of small RNA expression. Nat Methods 6( 10): 745– 751. Google Scholar CrossRef Search ADS PubMed Vasquez-Rifo A, Jannot G, Armisen J, Labouesse M, Bukhari SIA, Rondeau EL, Miska EA, Simard MJ, Hart AC. 2012. Developmental characterization of the MicroRNA-specific C. elegans Argonautes alg-1 and alg-2. PLoS One 7( 3): e33750. Google Scholar CrossRef Search ADS PubMed Waddington CH. 1957. The strategy of the genes. London: Allen & Unwin. Wilczynska A, Bushell M. 2015. The complexity of miRNA-mediated repression. Cell Death Differ . 22( 1): 22– 33. http://dx.doi.org/10.1038/cdd.2014.112 Google Scholar CrossRef Search ADS PubMed Wu CI, Shen Y, Tang T. 2009. Evolution under canalization and the dual roles of microRNAs – a hypothesis. Genome Res . 19( 5): 734– 743. http://dx.doi.org/10.1101/gr.084640.108 Google Scholar CrossRef Search ADS PubMed Wu E, Thivierge C, Flamand M, Mathonnet G, Vashisht AA, Wohlschlegel J, Fabian MR, Sonenberg N, Duchaine TF. 2010. Pervasive and cooperative deadenylation of 3’UTRs by embryonic microRNA families. Mol Cell 40( 4): 558– 570. Google Scholar CrossRef Search ADS PubMed Yanai I, Peshkin L, Jorgensen P, Kirschner MW. 2011. Mapping gene expression in two xenopus species: evolutionary constraints and developmental flexibility. Dev Cell 20( 4): 483– 496. http://dx.doi.org/10.1016/j.devcel.2011.03.015 Google Scholar CrossRef Search ADS PubMed Zalts H, Yanai I. 2017. Developmental constraints shape the evolution of the nematode mid-developmental transition. 1( 5): 113. Ensembl genome browser 89. © The Author 2017. Published by Oxford University Press on behalf of the Society for Molecular Biology and Evolution.All rights reserved. For Permissions, please e-mail: email@example.com
Molecular Biology and Evolution – Oxford University Press
Published: Mar 1, 2018
It’s your single place to instantly
discover and read the research
that matters to you.
Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.
All for just $49/month
Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly
Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.
Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.
Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.
All the latest content is available, no embargo periods.
“Hi guys, I cannot tell you how much I love this resource. Incredible. I really believe you've hit the nail on the head with this site in regards to solving the research-purchase issue.”Daniel C.
“Whoa! It’s like Spotify but for academic articles.”@Phil_Robichaud
“I must say, @deepdyve is a fabulous solution to the independent researcher's problem of #access to #information.”@deepthiw
“My last article couldn't be possible without the platform @deepdyve that makes journal papers cheaper.”@JoseServera