Deep sequencing of
Platynereis small RNAs identified 34 microRNA families common to protostomes and deuterostomes (
Supplementary Fig. 1, Supplementary Table 1), in accordance with recent studies
2. To investigate the temporal and spatial localization profile of these conserved bilaterian microRNAs, whole mount
in situ hybridization (WMISH) was conducted using locked nucleic acids as probes. In all cases, localization of mature
Platynereis microRNAs was spatially restricted and almost exclusively occurred in actively differentiating tissues (
Supplementary Fig. 2).We also analysed tissue-specific expression of predicted targets (
Supplementary Fig. 3 a, b and Supplementary Table 2) and identified a subset of tissues in which microRNAs were less frequently co-expressed with their predicted targets than expected to occur by chance, whereas other tissues showed the opposite trend (
Supplementary Table 3 a, b and Supplementary Data), indicating different modes of involvement of the conserved bilaterian microRNAs in regulating the establishment and/or maintenance of tissue identity
5.
We found the oldest conserved microRNA, miR-100, shared by cnidarians and bilaterians
1,2, highly specifically localized in two small groups of cells of the larval foregut (). These cells also expressed let-7 () and miR-125 (), which play a conserved role in developmental timing
6–8, and miR-375 (), demarcating foregutrelated, neurosecretory/endocrine cell populations of vertebrate pituitary and pancreas
9. Marker gene analysis revealed that these cells are differentiated secretory neurons (
Supplementary Fig. 4c). As in fly
7,8,
PlatynereismiR-100 and let-7 are processed froma single polycistronic transcript (
Supplementary Table 4) that in many bilaterians also includes miR-125 (refs
7,8). Since miR-100, miR-125 and let-7 are expressed much more broadly in fly
8 and vertebrate
10 (
Supplementary Table 5), we challenged evolutionary conservation of the highly specific
Platynereis pattern by investigating the expression of miR-100 in a cnidarian, the sea anemone
Nematostella. miR-100 was exclusively detected in single cells located around the pharynx anlage ( and
Supplementary Fig. 5a, b), demarcated by
brachyury11 and
foxa expression (blue, red and yellow in ). Intriguingly, miR-100,
brachyury12 and
foxa2 also colocalize in the
Platynereis foregut (; blue, red and yellow in ). These findings indicate that early in animal evolution miR-100 was active in a small population of cells located around a digestive opening. This expression was inherited by let-7 and by miR-125 once they evolved in the bilaterian stem line, to expand into other tissues only later in evolution. Corroborating this, expression of miR-100, miR-125 and let-7 was similarly restricted to foregut tissue in
Capitella (, data not shown) and in the sea urchin
Strongylocentrotus ( and
Supplementary Fig. 5c, e).
Locomotor ciliary bands are characteristic for the swimming larvae of marine protostomes and deuterostomes
13. In
Platynereis, miR-29, miR-34 and miR-92 demarcated the ciliary bands from early larval stages onwards () and were also detected in motile ciliated cells of the apical organ and later in the medial head region (). While expression was similar for all ‘ciliary’ microRNAs at early larval stages, partly complementary patterns were observed in the developing young worm (arrowheads in ).
Capitella showed a similar localization of these microRNAs to ciliary bands ( and
Supplementary Fig. 6a, b). We also investigated the sea urchin pluteus larva and indeed found miR-92 expressed in the ciliary bands () and miR-92 and miR-34 in motile ciliated cells lining the foregut and stomach (
Supplementary Fig. 5g, h). The similar specific affiliation of these microRNAs to motile ciliated cells in both protostome and deuterostome larvae is best explained by evolutionary conservation.
Two sets of conserved bilaterian microRNAs showed localized expression in distinct parts of the brain. For miR-7 (ref.
4), miR-137 and miR-153 we observed very similar expression in the dorso-medial neurosecretory tissue of the
Platynereis brain () comprising differentiated vasotocinergic and FMRFamidergic
4 as well as serotonergic neurons (). The same microRNAs showed brain-restricted expression in
Capitella (
Supplementary Fig. 6c–e) and have also been reported to show spatially localized expression in the zebrafish brain, including neurosecretory brain parts of the hypothalamus
14 (
Supplementary Table 5). Our comparative data thus indicate that these three microRNAs co-evolved in neurosecretory brain tissue. A conserved pair of complementary microRNAs, miR-9 and miR-9*/miR-131 (ref.
15) also showed highly restricted expression in the annelid brain, in two ventro-lateral sets of differentiated neurons ( and
Supplementary Fig. 6f). In mouse, miR-9 and miR-9*/miR-131 are detected broadly in neuronal precursors but among all differentiated neurons are expressed only in the telencephalon
15, comprising olfactory brain centres. In
Platynereis, the most apical cells expressing mature miR-9 and miR-9* were located at the base of the antennae, a pair of head appendages considered to be chemosensory sense organs (). This indicates that miR-9 and miR-9* may ancestrally locate to neurons involved in olfactory/chemosensory information processing. As in vertebrates, both miR-9 and miR-9* are detected in relatively high abundance in
Platynereis and thus represent a pair of complementary microRNAs conserved in bilaterians
2.
Other conserved bilaterian microRNAs were expressed more broadly in one of four major organ systems, representing the central nervous system, peripheral sensory tissue, musculature or gut. For each of these groups, expression of individual microRNAs was largely overlapping, but at the same time complementary to those of other groups. miR-124 (), known to maintain neuronal identity in the vertebrates
14,16 and expressed in developing central nervous system (CNS) in fly
17 and planarian
18, was found in differentiating neurons in the brain and ventral nerve cord in
Platynereis. miR-71, lost in vertebrates and insects
2 and restricted to CNS and parenchyma in planarians
18, is also nervous system-specific in
Platynereis ( and
Supplementary Fig. 6h). miR-8 proved to be an excellent marker for differentiating sensory organs, including eyes, antennae, palpae and sensory organs of the parapodia
19, covering both neural and non-neural tissue ( and
Supplementary Fig. 6j). A sensory tissue affiliation is likely to be evolutionarily ancient for this microRNA since the related miR-200a, miR-200b and miR-141 likewise show restricted expression in sensory organs such as nose and lateral line in the vertebrates
10 (
Supplementary Table 5). miR-183 and miR-263, clustered in the
Lottia genome and processed from the same transcript in
Platynereis (
Supplementary Table 4), also showed a conserved affiliation with sensory organ differentiation ( and
Supplementary Fig. 6i), as previously reported for other bilaterians
14,17. Notably, the expression of the sensory tissue-specific miR-183 and of the CNS-specific miR-124 was mutually exclusive (compare ). In fly and vertebrate, miR-1 is active during muscular differentiation
10,17. In
Platynereis, miR-1 and miR-133, clustered in vertebrates and in molluscs (
Supplementary Table 4), showed almost identical expression in the differentiating musculature (). In vertebrates, these microRNAs are robustly induced upon myotube differentiation concomitant with reduced expression of their target messengers
16,20,21. Similar musculature-specific expression was observed for miR-22 () that together with miR-1 and miR-133 has been reported to have myoD and myogenin upstream binding sites in the vertebrates
20. Finally, miR-12, miR-216 and miR-283 showed identical expression in the differentiating midgut at 5 days post fertilization (d.p.f.) (). miR-12 clusters with miR-216 in
Platynereis and in
Lottia22 and with miR-283 in
Drosophila17 (
Supplementary Table 4), indicating that these three microRNAs evolved from the same precursor gene. In the vertebrates, expression of miR-216 is characteristic of pancreatic tissue
23 and targets of
miR-216 are expressed at lower levels in pancreatic than in other tissue
21, indicating that the ancient site of activity of the miR-12/-216/-283 precursor had indeed been the gut.
Our results indicate that the conserved bilaterian microRNAs evolved in a strictly tissue-specific context. Corroborating this, brain-or musculature-specific microRNAs that we found in
Platynereis show similar specificity in fish, human and mouse expression profiling, with high confidence (
Supplementary Table 5). Although we cannot exclude that other expression sites initially existed for some of these microRNAs, these would have been lost in subsequent evolution in multiple lineages. At the present state of analysis, we can only speculate about the tissue-specific ancient roles of these microRNAs, by extrapolating from observations in other animal models. For example, miR-100 and the related miR-125 and let-
7 may have acquired an ancient role in developmental timing: The very late onset of
let-7 expression at 5 d.p.f. in
Platynereis (before settlement) is consistent with earlier observations in nematode
6, fly
7,8, mollusc
24, zebrafish
24 and in another annelid
24, indicative of a role in the control of late developmental transitions
6–8,24. With time, such role may then have spread to other tissues, such as target tissue differentiating in the course of metamorphosis, as observed in fly
7,8 and nematode
6,24. Similarly, the affinity of miR-29, miR-34 and miR-92 to motile ciliated cells sheds new light on the evolution of ventricular neuron types expressing these microRNAs in the vertebrate CNS
14 (
Supplementary Table 5), some of which are known to bear motile cilia
25. A comparison of targets in slow-evolving species should reveal the functional evolution of the conserved bilaterian microRNAs and also allow understanding why in many cases these microRNAs were co-opted by new tissues that did not express them before.
The identification of ancient expression sites for conserved bilaterian microRNAs implies that these microRNA-defined tissues existed in the protostome-deuterostome ancestor (
Supplementary Fig. 1). Our data indicate that it possessed a miR-124+ central nervous tissue as opposed to miR-8/-183/-263+ peripheral sensory nervous tissue
19, consistent with the notion that nervous system centralization predated the protostome/deuterostome ancestor
19. The ancient brain comprised miR-7+, miR-137+ and miR-153+ neurosecretory parts, as recently suggested
4 and miR-9/9*+ parts that may relate to sensory information processing. Our data also indicate the existence of miR-1/-22/-133+ body musculature, in line with the proposed conservation of somatic muscle cell types
19, and of a miR-12/-216/-283+ gut. Finally, the protostome/deuterostome ancestors possessed miR-100/-125/let-7/+ neurosecretory cells along the mouth and miR-29, miR-34, miR-92+ motile ciliated cells, possibly forming part of
otx+ larval ciliary bands
12,26. We have thus established microRNAs as an important new tool for reconstructing ancient animal body plans at important evolutionary nodes, focusing here on the protostome-deuterostome divergence. More complete inventories of microRNAs—yielding a refined picture of gains (and losses) of microRNAs in the diverging lineages—and concomitant expression analysis will allow expanding this approach to other key events of animal evolution.