MicroRNA-9 directs late organizer activity of the midbrain-hindbrain boundary

Abstract

The midbrain-hindbrain boundary (MHB) is a long-lasting organizing center in the vertebrate neural tube that is both necessary and sufficient for the ordered development of midbrain and anterior hindbrain (midbrain-hindbrain domain, MH). The MHB also coincides with a pool of progenitor cells that contributes neurons to the entire MH. Here we show that the organizing activity and progenitor state of the MHB are co-regulated by a single microRNA, miR-9, during late embryonic development in zebrafish. Endogenous miR-9 expression, initiated at late stages, selectively spares the MHB. Gain- and loss-of-function studies, in silico predictions and sensor assays in vivo demonstrate that miR-9 targets several components of the Fgf signaling pathway, thereby delimiting the organizing activity of the MHB. In addition, miR-9 promotes progression of neurogenesis in the MH, defining the MHB progenitor pool. Together, these findings highlight a previously unknown mechanism by which a single microRNA fine-tunes late MHB coherence via its co-regulation of patterning activities and neurogenesis.

Full text

Translate

Turn on search term navigation

ARTICLES

http://www.nature.com/natureneuroscience

Web End = MicroRNA-9 directs late organizer activity of the midbrain-hindbrain boundary

http://www.nature.com/natureneuroscience

Nature Publishing Group

Christoph Leucht1,5, Christian Stigloher1,5, Andrea Wizenmann2,4, Ruth Klafke3, Anja Folchert1 & Laure Bally-Cuif1

The midbrain-hindbrain boundary (MHB) is a long-lasting organizing center in the vertebrate neural tube that is both necessary and sufcient for the ordered development of midbrain and anterior hindbrain (midbrain-hindbrain domain, MH). The MHB also coincides with a pool of progenitor cells that contributes neurons to the entire MH. Here we show that the organizing activity and progenitor state of the MHB are co-regulated by a single microRNA, miR-9, during late embryonic development in zebrash. Endogenous miR-9 expression, initiated at late stages, selectively spares the MHB. Gain- and loss-of-function studies, in silico predictions and sensor assays in vivo demonstrate that miR-9 targets several components of the Fgf signaling pathway, thereby delimiting the organizing activity of the MHB. In addition, miR-9 promotes progression of neurogenesis in the MH, dening the MHB progenitor pool. Together, these ndings highlight a previously unknown mechanism by which a single microRNA ne-tunes late MHB coherence via its co-regulation of patterning activities and neurogenesis.

The architecture of the vertebrate CNS, rst subdivided into prosencephalon, mesencephalon, hindbrain and spinal cord, is progressively rened by local organizing centers, of which the best characterized is the MHB1. Organizers are not only involved in providing graded patterning cues to neighboring areas, but also act as long-lasting coordinators of many cellular events, such as cell fate, survival, proliferation, differentiation and migration. At early developmental stages, strong signaling from organizers, reinforced by positive regulatory loops and facilitated by the relatively short distances in the neural tube, ensures the spatial coherence of these cellular events. For instance, a crucial event in MHB maintenance and activity is signaling by the diffusible protein Fgf8 (ref. 1), which mainly exerts its function via its receptor Fgfr1 (ref. 2). However, a major question remains as to how this coordination can be maintained at later stages; organizer activity tends to decrease over time, and the distances that signals have to travel in the embryo and the diversity of cellular states around the organizer increase. The mechanisms active at these late stages remain unknown.Notably, organizers are often found in spatial overlap with areas of delayed cellular differentiation (reviewed in ref. 3). As an example, MHB activity in all vertebrates also coincides with a zone of long-lasting progenitors that separate midbrain from anterior hindbrain neuronal clusters46. In zebrash, the bHLH Hairy/E(spl) transcription factors Her3, 5, 9 and 11 inhibit neurogenesis in this location79 (C.S.,

unpublished data), and experimentally induced neurogenesis across the MHB ultimately causes late MHB loss6. In the mouse, a lack of Hes1 and Hes3 also leads to premature differentiation at the MHB and to the failure to maintain (but not to initiate) MHB activity10. Therefore, it

appears to be crucial for proper MHB maintenance to ensure the spatial coincidence of the Fgf and neurogenesis inhibition pathways. Looking for a mechanism involved in this process, we searched for microRNAs that could simultaneously target the Fgf signaling and neurogenesis inhibition pathways.

We found that miR-9 is expressed in the late embryonic zebrash CNS in a prole that selectively avoids the MHB. Using loss- and gain-of-function experiments, as well as sensor and target protection assays, we identied her5, her9 and several components of the Fgf signaling pathway (fgf8, fgfr1 and canopy1) as in vivo targets of miR-9. We demonstrate that some of these targets mediate the simultaneous interference of miR-9 with both Fgf signaling and the maintenance of the neural progenitor state in vivo, and that these activities converge to negatively delimit the MHB, where miR-9 is not expressed. These results provide a mechanism for maintaining a coherent MHB where organizer activity and neurogenesis inhibition are in spatial register and suggest a new role for microRNAs as major components of the cascades ne-tuning late organizers in the neural tube.

RESULTS Overexpression of miR-9 RNA causes MHB loss We found that the 3 UTRs of zebrash MHB genes share putative miR-9 binding sites. This is, in particular, true for important effectors of MHB activity, such as fgf8 (two sites), fgfr1 (two sites) and canopy1 (one site)2,11,12, and genes encoding MHB neurogenesis inhibitors, such as

her5 (one site) and her9 (one site)6 (C.S., unpublished data) (Supplementary Fig. 1 online). To test whether these sites might be functionally

200 8

1Department of Zebrash Neurogenetics, Institute of Developmental Genetics, 2Institute of Stem Cell Research, 3Institute of Developmental Genetics, Helmholtz-Zentrum Mnchen, German Research Center for Environmental Health, Ingolstadter Landstrasse 1, D-85764 Neuherberg, Germany. 4Present address: Institute of Anatomy, University of Tbingen,sterbergstrasse 3, D-72074 Tbingen, Germany. 5These authors contributed equally to the work. Correspondence should be addressed toC.L. (mailto:[email protected]

Web End [email protected]) or L.B.-C. (mailto:[email protected]

Web End [email protected]).

Received 27 February; accepted 24 March; published online 4 May 2008; http://www.nature.com/doifinder/10.1038/nn.2115

Web End =doi:10.1038/nn.2115

NATURE NEUROSCIENCE VOLUME 11 [ NUMBER 6 [ JUNE 2008 641

ARTICLES

d e

f g

h i

j k

premiR-9

Figure 1 Gain of miR-9 function causes MHB loss. (ak) Embryos were injected at the one-cell stage with premiR-9-1 (c,e,g,i,k)or thesame concentration of a pre-miR control (b,d,f,h,j), with no predicted binding sites on her5, her9, canopy1, fgf8 and fgfr1; a shows an uninjected embryo (wild type, WT). Arrows indicate the location of the MHB; asterisks indicate a missing MHB. Morphology of injected embryos at 30 hpf (lateral views, anterior left; ac). The MHB is missing on overexpression of premiR-9 (c). Expression of neural tube regionalization markers in premiR-9overexpressing embryos (sagittal views, anterior left; embryos in dg are deyolked and at-mounted). MHB markers were downregulated (dg), but other regional markers were not (hk)(see also Supplementary Fig. 3). Som, somites.

30 hpf WT

10 M pre-miR control

20 som wnt1 pre-miR control

12 som eng2a pre-miR control

12 som

http://www.nature.com/natureneuroscience

premiR-9

egr2b pre-miR control

premiR-9

10 M premiR-9

12 somnkx6.1 pre-miR control premiR-9

Nature Publishing Group

relevant in vivo, we monitored the effect of miR-9 RNA overexpression at an early stage in zebrash. Given the known function of the abovementioned genes in early MHB development, we expected a detectable MHB phenotype if one or more of these genes were ectopically targeted by miR-9 in vivo in these experimental conditions.

We injected 10 mM miR-9, either as a duplex or as a precursor hairpin molecule (premiR-91, sequences in Supplementary Fig. 1), into fertilized oocytes and analyzed the resulting embryos by morphology at 30 h postfertilization (hpf). All embryos (n 4 100) had a prominent and marked phenotype that was characterized by a strong reduction of the MHB and the cerebellum (Fig. 1ac). Additionally, somitic boundaries appeared to be blurred and the otic vesicles were reduced, an effect that was previously reported for fgf8 mutants and fgfr1 knockdown embryos2,13 (Supplementary Fig. 2 online). This phenotype was not observed on injection of several other CNS-expressed miRNAs (Supplementary Fig. 2 and data not shown) and was not accompanied by typical nonspecic effects (reviewed in ref. 14) such as enlarged hearts or truncated tails (Supplementary Fig. 2), which only occurred at exaggerated doses (20 mM or more) (Supplementary Fig. 2). The selectivity of the miR-9induced MHB phenotype was further substantiated by the expression analysis of MHB markers, which were lost on miR-9 injection (Fig. 1dg and Supplementary Fig. 3 online), whereas other neural tube patterning markers were unaffected (Fig. 1hk and Supplementary Fig. 3).

In vivo conrmation of predicted miR-9 binding sites To test whether overexpressed miR-9 was capable of directly interacting with its predicted binding sites (Supplementary Fig. 1) and silencing the expression of its candidate genes in vivo, we conducted sensor assays for fgf8, fgfr1, cnpy1 (canopy1), her5 and her9 (Fig. 2). We engineered fusions of the d2egfp cDNA to the full 3 UTR or to the putative miR-9 binding sites of each gene (Fig. 2a) and injected capped mRNA generated from these constructs, together with miR-9 duplex or a control microRNA with a shufed miR-9 sequence, into one-celled embryos. The amount of d2EGFP protein expressed after 7 h was

quantied via western blot (Fig. 2b,c). We found a strong interaction of miR-9 with the 3 UTR elements of her5, her9, canopy1, fgf8 and fgfr1, as shown by a downregulation of d2EGFP. For the latter two genes, the putative binding sites fgf8-1 and fgfr-1, but not fgf8-2 and fgfr1-2, mediated this downregulation (Fig. 2b, see Supplementary Fig. 1 for sequences of the binding sites).

We next focused on one example gene for each regulatory pathway considered (fgfr1 for Fgf signaling and her5 for neurogenesis inhibition) and further conrmed these ndings using 3 UTR reporter constructs carrying engineered point mutations in their predicted miR-9 binding sites (site fgfr1-1 for fgfr1), which abolished downregulation by miR-9 (constructs her5mt and fgfr1mt; Fig. 2b,c). We conclude that her5, her9, canopy1, fgf8 and fgfr1 are probably in vivo targets of miR-9 and, in the case of the latter two genes, probably interact with miR-9 via the predicted binding sites fgf8-1 and fgfr1-1.

miR-9 overexpression downregulates Fgf signaling The phenotype triggered by miR-9 overexpression (Fig. 1 and Supplementary Figs. 2 and 3) is very reminiscent of that of the zebrash fgf8 mutant ace, which also lacks the cerebellum, MHB and expression of MHB marker genes13, and of fgfr1 morphants that show many aspects of the ace phenotype2. We therefore addressed whether miR-9 over-expression blocks Fgf signaling. Consistent with this hypothesis, we found that the expression of Fgf target genes such as dusp6 (ref. 15) and pea3 (ref. 16) was strongly downregulated throughout the embryo after miR-9 injections (Fig. 3ad), although the 3 UTR of dusp6 does not interact with miR-9 (ref. 17) and we could not predict a miR-9binding site in the 3 UTR of pea3 (data not shown). Another hallmark of Fgf loss of function at the MHB, unraveled in ace mutants, is that MH cells express abnormal identities18; posterior MH cells, normally free of expression of the midbrain marker otx2 (Fig. 3e), aberrantly expressed otx2 in ace. Therefore, we also assessed the effects of miR-9 injections on MH cell fate using the transgenic line Tg(her5PAC:egfp)ne1939 (later

referred to as her5PAC:egfp)18. In this line, GFP traces all MH cells until approximately 24 hpf. her5PAC:egfp embryos injected with miR-9 were stained for otx2 expression, and we quantied the territory coexpres-sing GFP and otx2 (indicative of midbrain identity) relative to the entire MH (GFP-positive cells) (Fig. 3f). This ratio was signicantly higher in ace and miR-9injected embryos than in wild type, indicating that both backgrounds similarly promote the transformation of posterior MH to midbrain identity (wild type and ace, Po0.01; wild type and miR-9, Po0.05; ace and miR-9, Po0.05). Together, these observations strongly suggest an interaction of miR-9 with components of the Fgf signaling pathway.

To formally prove that the targeting of Fgf signaling by miR-9 was instrumental in causing the MHB deletion of miR-9overexpressing embryos, we protected the miR-9 binding sites of endogenous fgfr1 transcripts concomitantly to miR-9 overexpression. To achieve this, we injected the miR-9 duplex into one-celled embryos together with a target protector19 morpholino antisense oligonucleotide specic to

200 8

642 VOLUME 11 [ NUMBER 6 [ JUNE 2008 NATURE NEUROSCIENCE

ARTICLES

Figure 2 Sensor assay to reveal direct interaction of miR-9 with predicted binding sites. (a) Scheme of the sensor mRNA that was co-injected with the miR-9 duplex or a control miR duplex (a shufed miR-9 sequence, miR-9sh). (b) We subjected embryos to western blot analysis 7 h after injection. b-tubulin was used as an internal loading control. Asterisks indicate cases where d2EGFP expression was downregulated in a statistically signicant manner by miR-9 compared with miR-9sh. 3 BS, three copies of the predicted binding site were fused to d2egfp;3 UTR, the full 3 UTR was fused to d2egfp; mt, full 3 UTR carrying engineered point mutations in predicted miR-9 binding sites (see Supplementary Fig. 1). (c) Densitometric analysis showed a downregulation of d2EGFP by miR-9 via the fgfr1, canopy1 and her5 3 UTRs, and the fgf8-1, fgfr1-1 and her9 binding sites. We found no downregulation of d2EGFP via the predicted miR-9 binding sites fgfr1-2 and fgf8-2. Point mutations engineered in the 3 UTR of her5 and fgfr1 (site fgfr1-1) abolished the downregulation by miR-9. Data are means s.d., P o 0.05 (t test, unpaired). All lanes were normalized to the b-tubulin signal.

Stop

d2egfp SV40pA

3 UTR sequences

100

Percentage of control

fgf8-1 (3BS)

fgf8-2 (3BS)

fgfr1-1 (3BS)

fgfr1-2 (3BS)

her9 (3BS)

canopy1 (3 UTR)

miR-9shmiR-9 miR-9shmiR-9 miR-9shmiR-9 miR-9shmiR-9 miR-9sh miR-9 miR-9sh miR-9

miR-9 miR-9 miR-9

-tubulin

d2EGFP *

* * *

http://www.nature.com/natureneuroscience

her5mt (3 UTR)

fgfr1 (3 UTR)

fgfr1mt (3 UTR)

her5 (3 UTR)

miR-9sh miR-9sh miR-9sh miR-9sh miR-9

her5

0 UTR fgfr1 3

UTR fgfr1 mt 3

UTR canopy1

UTR

UTR fgfr1-1

BS fgf8-1 3

BS her9 3

BS her

5 mt 3

Nature Publishing Group

the predicted miR-9 binding site fgfr1-1. We observed that protecting fgfr1-1 was sufcient to rescue the MHB and expression of MHB markers in a substantial proportion of miR-9overexpressing embryos (Fig. 3gj and Supplementary Table 1 online). We conclude that overexpression of miR-9 impairs Fgf signaling at the MHB, at least in part by targeting fgfr1 transcripts, and that this effect is instrumental in causing MHB loss. The interaction of microRNAs with their target genes does not necessarily result in reduced mRNA levels, but rather inhibits protein translation (for a review, see ref. 20). Together with the fact that we could not detect miR-9 binding sites on, for example, wnt1 (data not shown), the effect of miR-9 on Fgf signaling suggests that the loss of MHB marker genes reported above (Fig. 1, Supplementary Fig. 3) is not a direct effect, but, as in ace mutants, is instead a result of interference in the MHB regulatory loop.

miR-9 promotes neurogenesis in vivoIn addition, two sets of observations suggest that an interaction of miR-9 with pathways other than Fgf contributes to the MHB pheno-type of miR-9injected embryos. First, downregulation of Fgf signaling, as shown by pea3 and dusp6 expression (Fig. 3ad), was initiated at a later stage on miR-9 injection (1012 somites) than in ace mutants (5 somites), and chemical blockade of Fgf signaling in wild-type embryos at this later stage was not sufcient to consistently produce MHB loss (data not shown). Second, and paradoxically, the loss of expression of MHB markers in miR-9injected embryos (Fig. 1dg and Supplementary Fig. 3) was more complete and occurred earlier than in ace mutants; before 18 somites, ace embryos only show ventrally reduced expression of, for example, eng2a and wnt1 at the MHB13,

whereas these were already almost completely downregulated at 12 somites on miR-9 injection (Fig. 1dg). We conclude that miR-9 targets more than the Fgf pathway when affecting MHB maintenance on overexpression.

Abrogation of Hes1 and Hes3 function in the mouse induces premature neurogenesis at the MHB and thereby triggers MHB loss10. Thus, one process contributing to MHB failure on miR-9 injection might be the promotion of neurogenesis in the MHB

200 8

miR-9

miR-9 + fgfr1BSMO

a b

c d

g h

i j

Control miR

miR-9

WT 12 som Overlay

** *

12 som dusp6

12 som pea3 pea3

dusp6

0.8

Her5-GFP+;otx2+/ Her5-GFP+

24 hpf pax2a pax2a

wnt1 wnt1

0.6

0.4

0.2

otx2

Her5-GFP

ace miR-9 WT

Figure 3 miR-9 overexpression downregulates Fgf signaling. (ad) Expression of Fgf target genes (dusp6 and pea3, as indicated) at the 12-somite stage in embryos injected with 10 mM miR-9 duplex (b,d) or the same amount of the control miR-9 shufed duplex (a,c) (lateral views of the head of at-mounted embryos, anterior left; insets show the posterior trunk and tail of each embryo; arrows to the MHB and arrowheads to other expression domains in the telencephalon, hindbrain and tail). miR-9 downregulated expression of dusp6 and pea3 at all sites. (e,f) Assessment of posterior MH fate on miR-9 overexpression. Comparing the expression of Her5-GFP (GFP immunocytochemistry, green) and otx2 (ISH, blue) allowed the identication of the midbrain (Her5-GFP+;otx2+) and the posterior MH (rhombomere 1, Her5-GFP+;otx2)in her5:egfp transgenic wild-type embryos (posterior limits of otx2 and Her5-GFP expression indicated by color-coded arrowheads) (confocal views of sagittal cryosections, focus on the head, anterior left; e). Quantication of the relative extent of midbrain identity in the MH of wild-type embryos, ace mutants and miR-9overexpressing embryos (f). There was a signicant increase in midbrain extent in the two latter backgrounds, indicative of a fate transformation of posterior MH into midbrain (n 3 embryos, means s.d., *P o 0.05, **P o 0.01; t-test, unpaired). (gj) Targeting of fgfr1 mRNA by overexpressed miR-9 was instrumental in causing MHB loss. Expression of the MHB markers pax2a and wnt1 (as indicated) in embryos injected with 10 mM miR-9 duplex alone (g,i) or in combination with a morpholino (fgfr1BSMO) protecting the miR-9 binding site fgfr1-1 on endogenous fgfr1 transcripts (h,j) (sagittal views of at-mounted embryos, anterior left). Arrows point to rescued expression domains at the MHB (see Supplementary Table 1 for values and statistics).

NATURE NEUROSCIENCE VOLUME 11 [ NUMBER 6 [ JUNE 2008 643

ARTICLES

Control miR miR-9 WT her5/9gripNAs

vcc

vcc vcc vcc

r2mn r2mn

3 som neurog1

Control miR

miR-9

MHB

Her5-GFP pH3 HuC Overlay

f g

** **

* **

HuC/Her5-GFP cells (%)

0+Control miR ace +miR-9

Phosphohistone cells (%)

0+Control miR ace +miR-9

http://www.nature.com/natureneuroscience

ace

Figure 4 miR-9 overexpression causes premature neurogenesis across theMHB. (ad) Compared effects of overexpressing miR-9 (b) and blockingHer5/Her9 function (d), as revealed using neurog1 expression at threesomites (all views are at-mounted embryos, anterior up, focus on the MHarea). Both manipulations (b,d) led to an identical premature induction ofneurog1 expression across the MHB (asterisks). In control embryos, the MHB at this stage is the neurog1-free zone (brackets in a and c) separating the proneural clusters vcc (ventro-caudal cluster) and r2mn (presumptive motorneurons of rhombomere 2). (eg) Quantication of the miR-9induced neurogenesis phenotype at 18 somites, using immunohistochemistry for the early pan-neuronal marker HuC (blue) and the proliferative marker pH3 (red) in the her5PAC:gfp background (GFP, green; MH) (all panels in e are cross sections of embryos at the MH level). (f,g) Statistical analysis of the percentage of HuC-positive cells (d) and pH3-positive cells (e) in the MH. Embryos overexpressing miR-9, but not ace mutants, showed an increased number of postmitotic neurons at MH levels (f). In contrast, miR-9injected and ace embryos showed a comparable decrease in proliferation (g). Means s.d.; *P o 0.05, **P o 0.01 (t test, unpaired). (hk) Expression of the MHB markers pax2a and wnt1 in embryos co-injected with miR-9 duplex and a her5 target protector (her5BSMO) (i,k) compared with embryos overexpressing miR-9 duplex alone (h,j). Blockade of miR-9 binding to her5 by the target protector rescued the loss of MHB markers (see Supplementary Table 1 for values and statistics).

h i j k

miR-9 miR-9 her5BSMO

miR-9 her5BSMO miR-9

* *

pax2a wnt1 24 hpf 24 hpf

Nature Publishing Group

progenitor pool, possibly by targeting the second set of predicted miR-9 targets, antineurogenic genes such as her5 or her9 (Fig. 2). To assess this issue, we rst tested whether miR9 overexpression induces a neurogenic phenotype and whether this phenotype is comparable to that of abrogating Her5 and Her9 function. Simultaneous blockade of her5 and her9 by injection of antisense oligonucleotides triggered the activation of neurogenin1 (neurog1) expression across the medial aspect of the MHB at three somites (Fig. 4 and ref. 6), a zone that was normally free of neurogenesis (Fig. 4c,d). miR-9 overexpression resulted in an identical phenotype (Fig. 4a,b), strengthening the hypothesis that her5/her9 are direct targets of miR-9, as suggested by the predictions (Supplementary Fig. 1) and sensors assays (Fig. 2). To further test for an inuence of miR-9 on neurogenesis and to make sure that this effect was independent of the regulation of Fgf signaling by miR-9, we analyzed cell proliferation and differentiation in the MH of miR-9injected embryos in comparison with ace and control embryos at 18 somites (Fig. 4eg). We again made use of the her5PAC:egfp background, where GFP identies the MH at this stage18, combined with immunostaining for the M phase marker phosphorylated histone H3 (pH3) or the neuronal differentiation marker HuC (Fig. 4e). In the MH, we observed signicantly less mitotic cells in miR-9injected embryos compared with control embryos (P o 0.01; Fig. 4g). Because this is also the case in ace mutants (Fig. 4g), we attribute this effect to the impairment of Fgf signaling as a result of miR-9 overexpression. Notably, however, we found a marked increase in the number of HuC-positive cells in the MH of miR-9injected embryos compared with both ace and control embryos (Fig. 4f), demonstrating that miR-9 strongly promotes MH neurogenesis in vivo, independently of its action on Fgf signaling.

Finally, to prove that targeting of the Her5/9 pathway was contributing to the MHB-loss phenotype that follows miR-9 overexpression, we

injected embryos with both miR-9 and a target protector19 against the predicted miR-9binding site in the her5 3 UTR. This manipulation should reactivate Her5 function and leave other miR-9 targets unaffected. We observed that reactivation of Her5 rescued MHB markers (Fig. 4hk and Supplementary Table 1). Together with the above results on Fgf signaling, these observations show that blockade of both the Her and Fgf pathways by miR-9, at least via binding her5 and fgfr1, contributes to the complete MHB-loss phenotype induced by miR-9 overexpression.

miR-9 expression spares the MHB at late embryonic stages Expression of miR-9 started at 2024 hpf (approximately 30 somites) in the telencephalon (Fig. 5a,b) and later spread (starting at approximately 30 hpf) throughout the CNS (Fig. 5c,d). At 30 hpf and later stages, in situ hybridization (ISH) in the her5PAC:egfp background (GFP selectively labeling the MHB at that stage18) revealed that miR-9 expression consistently spared the MHB (Fig. 5e,f, see also Fig. 5c,d), but was found in immediately adjacent territories, such as the cerebellar plate and the ventricular zone of the tectum (Fig. 5cf). The Her5-GFPpositive MHB domain avoided by miR-9 expression also precisely coincided with the territory expressing fgf8 and fgfr1 (Fig. 5gj). This prole, together with the gain-of-function effects reported above, is compatible with miR-9 having a role in the anteroposterior restriction of Her and Fgf activities, hence of the MHB and its properties, at these stages.

In addition, along the mediolateral axis, miR-9 expression mostly highlighted the ventricular zone, directly adjacent to the HuC-positive domain, with few cells expressing both factors (Fig. 5f,k,l). This suggests that miR-9 expression stops when cells differentiate and is consistent with miR-9 driving the commitment of progenitors toward neurogenesis.

200 8

644 VOLUME 11 [ NUMBER 6 [ JUNE 2008 NATURE NEUROSCIENCE

ARTICLES

Figure 5 Endogenous miR-9 expression in theMH area avoids the MHB and postmitoticdomains. (ad) ISH for miR-9 (whole-mountembryos with focus on the brain area, anteriorleft, dorsal views in a and c, sagittal views in band d). Expression was rst detectable at 24 hpfin the telencephalon (a and b, arrows) and thenextended to the rest of the neural tube. At 35 hpfin the MH, miR9 stained the midbrain andhindbrain (c and d, arrows) but avoided the MHB(white arrowheads). (ej) ISH for miR-9, fgf8 orfgfr1, as indicated (black staining), in theher5PAC:egfp transgenic background at 35 hpf(GFP, green or red, highlights the MHB) (all viewsare sagittal sections observed under confocalmicroscopy with focus on the head, anterior left;e and f, g and h,and i and j are pairs of identicalsections viewed under bright eld alone or withsuperimposed uorescence, respectively; f is atriple staining to detect the postmitotic neuronal marker HuC, blue). miR-9 was expressed in the mid- and hindbrain (e, arrows; CB, cerebellar plate; Te, tectum), but avoided the Her5-GFPpositive area (MHB, arrowheads). fgf8 and fgfr1 expression overlapped with Her5-GFP (gj, arrowheads). (k,l) Expression of miR-9 (black) compared with HuC (blue on l, same section as k visualized under bright eld and uorescence) at hindbrain levels. miR-9 expression was largely conned to the ventricular zone (see also f).

a b c d

e f g h

i j k l

http://www.nature.com/natureneuroscience

Nature Publishing Group

miR-9 activity delimits Fgf signaling and the MHB We used a morpholino oligonucleotide specic to miR-9 (miR-9MO) to inactivate miR-9 function in vivo. Injection of miR-9MO resulted in a complete blockade of miR-9, as shown by miR-9 ISH after injection (Fig. 6a,b). We rst analyzed the effects of miR-9 blockade on the expression of Fgf targets at the MHB. Obvious changes in the spatial extent of expression of most MHB markers were difcult to assess by ISH, but we did nd signicant differences in the relative transcripts levels of these genes at 35 hpf using qPCR; the expression of fgf8, pea3, fgfr1 and canopy1 were signicantly upregulated on miR-9MO injection (P o 0.05; Fig. 6c). Again, because microRNAs generally do not modify expression of their target mRNAs20 and because we did not

predict miR-9 binding to pea3, this general upregulation probably reects increased Fgf signaling amplifying the expression of MHB markers via the MHB regulatory loop. In addition, we observed an obvious enlargement of the most sensitive Fgf read-out, dusp6 expression, at 35 hpf at the MHB using ISH. dusp6 expression expanded into the cerebellum (Fig. 6dg), a region where miR-9 is normally expressed (Fig. 5cf), as well as along the mediolateral axis at levels immediately posterior to the MHB (Fig. 6h,i).

To further support the hypothesis that the upregulation of dusp6 in the absence of miR-9 activity resulted from increased Fgf signaling, we treated miR-9MOinjected embryos with the Fgf signaling inhibitor SU5402. In all such embryos, dusp6 expression was abolished, including

b d e

f g

h i

Figure 6 Morpholino knockdown of miR-9 affects Fgf signaling in a manner opposite to miR-9 gain of function. (a,b) miR-9MOinjected embryos and embryos injected with the control miR-9MO were analyzed for miR-9 expression using a digoxigenin-labeled probe against miR-9 at 35 hpf (sagittal views of whole-mount embryos, anterior left). Injection of miR-9MO at 2 mM completely blocked detectable miR-9 expression in all cases (b,compare with a). (c)A whisker-box plot showing the results of the quantitative PCR on several components of the Fgf signaling pathway in miR-9MOinjected versus uninjected samples.Expression of fgf8, pea3, fgfr1 and canopy1 were increased. P o 0.05 for all genes. (dk)Effect of blocking miR-9 activity on dusp6 expression revealed by ISH (blue) (d,e,j,k, sagittal views of whole-mount embryos, anterior left; f,g,sagittal sections, anterior left; h,i, cross sections at the level of the anterior hindbrain, dorsal up; double arrow to MH expansion in j). dusp6 expression expanded along the anteroposterior (e,double arrow; g, arrow) and mediolateral (i, arrow) axes in embryos injected with the miR-9MO (e,g,i), but not with the control MO (d,f,h)(d is an uninjected embryo). dusp6 expression and expansion was blocked when miR-9MOinjected embryos were treated with the Fgf signaling inhibitor SU5402(k; compared with mock-treated embryos, j; double arrow to MH expansion in j).

NATURE NEUROSCIENCE VOLUME 11 [ NUMBER 6 [ JUNE 2008 645

200 8

Expression ratio

j k

bactin2 fgf8 pea3 fgfr1 canopy1 gapdh

ARTICLES

a b

Figure 7 Morpholino knockdown of miR-9 affects neurogenesis in a manner opposite to miR-9 gain of function. (ac) Control embryos (a), miR-9MO injected embryos (b), or embryos injected with both miR-9MO and her9gripNA (c) were analyzed for the expression of the proliferation marker pH3 (red) and the neuronal differentiation marker HuC (blue) (cross-sections at anterior hindbrain levels). (d) The relative extent of the HuC-positive area was quantied. miR-9MOinjected embryos showed reduced neuronal differentiation, and this defect was rescued by the co-inhibition of Her9 activity. Data are means s.d.; two-sample t-test. * P 6.37 106, ** P o 0.02.

area /

neural tube area, %)

70 60 50 40 30 20 10

* **

Area ratio (HuC +

http://www.nature.com/natureneuroscience

+miR-9MO

+her9 gripNA +miR-9MO

Control

on the cerebellar plate (Fig. 6j,k). Consistent with our gain-of-function experiments, these results demonstrate that the absence of miR-9 function increases activity of the Fgf signaling pathway at the MHB, in turn amplifying expression of MHB markers and MHB extent.

miR-9 activity controls MH neurogenesis progression We next analyzed the requirement for miR-9 in neurogenesis progression, focusing on regions that show high levels of miR-9 expression, such as the anterior hindbrain. Immunostaining for HuC on cross sections at hindbrain levels revealed a statistically signicant reduction of the relative HuC-positive area at 35 hpf on miR-9 blockade (P 0.01161; Fig. 7a,b), demonstrating an impairment of neuronal differentiation in the absence of miR-9. The proportion of mitotic cells (pH3 positive) in the ventricular zone was not substantially affected (data not shown). Hence, endogenous miR-9 is required for neurogenesis progression in this neural tube domain. This effect was not accompanied by increased expression of the predicted miR-9 targets involved in neurogenesis inhibition, that is, her5 and her9, as revealed by qPCR (data not shown). Again, however, miR-9 activity probably does not directly modify expression of its targets, and we used epistasy experiments to address the implication of her genes downstream of miR-9 in neurogenesis control in vivo.

We focused on her9 because its expression extends into the anterior hindbrain21. Blocking Her9 activity by the injection of an antisense her9gripNA oligonucleotide into one-celled embryos led to ectopic neurog1 expression in inter-proneural domains of the presumptive spinal cord at three somites, as reported using a her9 morpholino oligonucleotide (her9MO)21 (data not shown). Notably, we observed that abrogating Her9 activity antagonized the phenotype of decreased neurogenesis induced by miR-9 blockade, bringing neurogenesis back to higher levels in embryos co-injected with her9gripNA and miR-9MO (Fig. 7c,d). Again, this effect did not involve substantial alterations in cell proliferation, which supports a direct rescue of the miR-9induced phenotype. These results demonstrate that miR-9 is necessary to control neuronal differentiation in the MH in vivo and strongly suggest that this occurs, at least in part, via its regulation of her genes.

DISCUSSION Our data demonstrate that miR-9 has at least two roles in the vertebrate CNS: it globally regulates Fgf signaling by inhibition of fgf8, fgfr1 and canopy1, and it exerts a proneurogenesis effect, notably by inhibiting expression of antineurogenic bHLH transcription factorencoding genes such as her5 and her9. It is probable that the targeting of Fgf

pathway genes is even more extensive than we report here, as we could also predict miR-9binding sites in the 3 UTRs of fgf3, fgf17a and fgf18l (C.L., unpublished data). miR-9 thus appears to be a major regulator of the Fgf pathway during CNS development, an activity that is probably conserved across species, as we could also predict miR-9 binding sites on the 3 UTRs of human FGFR1 and multiple Fgf genes in human and mouse (C.L., unpublished data).

The expression of miR-9 in neural progenitors through most of the CNS of zebrash, chicken and mouse2225 and the effects of its gain-and loss-of-function on neuronal differentiation that we observed here suggest a role in vertebrate neurogenesis control that is not limited to the MH. Specically, the effect of miR-9 blockade via Her factors is compatible with a role for miR-9 in pushing commitment toward the differentiation state without having direct control over the proliferation of early progenitors. The promotion of neurogenesis progression by miR-9 might a priori seem surprising given its expression in the ventricular zone of the neural tube, where progenitors reside. At the stages analyzed, however, this is also the domain of expression of proneural genes like neurog1 (ref. 26), hence the zone where commitment toward neurogenesis takes place.

An involvement of miR-9 in neurogenesis was previously suggested on the basis of in silico predictions and expression analysis of miR-9a in Drosophila27. However, experimental assignment of miR-9a function in Drosophila led to results that diverge from ours28,29; a previous study

illustrated an antineuronal role of miR-9a in the Drosophila PNS29,

where it maintains cells in the non-neuronal (epidermal) state by repressing the proneural gene senseless. The expression of Drosophila miR-9a highlights epidermal, as opposed to neuronal, precursors, a situation that parallels the expression of vertebrate miR-9 in neural progenitors, as opposed to differentiated neurons. miR-9 activity therefore results in two markedly different outcomes between species. This functional divergence might be a result of different types of genes being targeted in vivo by miR-9 depending on the species, for example, proneural genes in Drosophila versus neurogenesis inhibitors in zebra-sh. miR-9 was also predicted to target human, mouse and chicken Hes1 (data not shown), which maintains progenitor pools30. The effect

of miR-9 on neurogenesis might also be aided, possibly in a variable manner between species, by interaction with the transcriptional repressor REST31,32, although in vivo evidence for this interaction remains to be provided.

The endogenous expression of miR-9 in the MH started after 30 hpf and was consistently found immediately adjacent to, but nonoverlap-ping with, the MHB. The targets of miR-9 identied in this study (fgf8, fgfr1, canopy1, her5 and her9) are all expressed at the MHB at late developmental stages, similar to the predicted targets fgf3, fgf17a and fgf18l. These ndings are consistent with the target/antitarget theory, which argues that microRNAs and their targets are expressed in a largely nonoverlapping manner27. In the present case, this expression pattern, together with the dual effects of miR-9 on Fgf signaling and her genes, imposes a special status on the MHB as a domain where miR-9 effects are concurrently released. Our results therefore support a model

Nature Publishing Group

200 8

646 VOLUME 11 [ NUMBER 6 [ JUNE 2008 NATURE NEUROSCIENCE

ARTICLES

whereby miR-9 expression and functions converge to contribute to the late coherence of the MHB in vivo; through its absence at the MHB, miR-9 regulates MHB correct positioning, spatial restriction and the coincidence of MHB patterning and neurogenesis inhibition activities (Supplementary Fig. 4 online). One implication of our model is that microRNAs not only control isolated events during development or cell specication, but probably provide a metabolically cheap method for the organism to regulate complex processes, such as the maintenance of progenitor pools or organizing centers, by targeting several converging components of these processes. In addition, our results identify, to the best of our knowledge, the rst known mechanism involved in assisting the spatial coherence of late neural tube organizers and suggest that such coordinating activity of the multiple functional inputs and outputs of organizing centers might constitute a previously unknown function of microRNAs.

METHODSFish strains. Embryos were obtained from natural spawning of AB wild-type or transgenic sh, aceti282a or her5PAC:egfp [Tg(her5PAC:EGFP)ne1939]18,33. Embryos were staged according to a previous study34. All experiments were performed in accordance with the regulations of the Regierung von Oberbayern.

Computational analysis of miR-9binding sites. Binding sites were analyzed using the programs MIRANDA35, MicroInspector36, RNA22 (ref. 37) and RNAHybrid38. Sequences of predicted sites are provided in Supplementary Figure 1.

miR-9 duplexes, and premiR-9 and miR-9 morpholino injections. Doses injected were always two- to vefold lower than those reported to trigger general nonspecic effects (Supplementary Fig. 2 and see ref. 14 for a review). miR-9, miR-124 and miR-138 RNA duplexes and a duplex containing a shufed miR-9 sequence were obtained as siRNAs (miR-9: sense 5-UCU UUG GUU AUC UAG CAG AAU GARNA, antisense 5-AUA CAG CUA GAU AAC CAA AGARNA TTDNA; miR-124: sense 5-UCA CAG UGA ACC GGU CUC UUU URNA, antisense 5-AAG AGA CCG GUU CAC UGU GARNA TTDNA; miR-138: sense 5-AGC UGG UGU UGU GAA UCA GGC CRNA, antisense 5-CCU GAU UCA CAA CAC CAG CURNA TTDNA; shufed miR-9: sense 5-UAU CAC UUC UAU AUG GUU UGG UGRNA, antisense 5-CCA AAC CAU AUA GAA GUG AUARNA TTDNA) and injected into one-celled fertilized embryos at a concentration of 10 mM. PremiR-9-1 RNA (sequence in Supplementary Fig. 1) and a pre-miR negative control (Ambion pre-miR control #1) were obtained from Ambion and used at 10 mM. All morpholinos were obtained from Gene Tools and used as follows: miR-9MO (5-TCA TAC AGC TAG ATA ACC AAA GA-3) was injected at 2 mM, the control morpholino oligonucleotide (a shufed sequence of miR9MO: 5-CAC CAA ACC ATA TAG AAG TGA TA-3) was injected at 2 mM, her5 target protector (her5BSMO, 5-ATC TTT GGC ATC TAC TGT ACA AAA T-3) was injected at0.1 mM or 0.5 mM, and fgfr1-1 target protector (fgfr1BSMO, 5-CTT TGG CGG TTT TGT GTG CAG CTG T-3) was injected at 0.1 mM or 0.5 mM. her9gripNA was obtained from Active Motifs (Carlsbad) (5-TGA TTT TTA CCT TTC TAT-3) and was based on the published her9MO21. It was favored over her9MO for its absence of toxicity (data not shown).

Quantitative real-time PCR. Total RNA was extracted from sample and control group embryos (a pool of 3035-hpf-old embryos) using the RNeasy kit (Qiagen). We used 1 mg of total RNA to generate cDNA using the Transcriptor First Strand cDNA Synthesis Kit (Roche). Quantitative real-time PCR was carried out on a LightCycler 1.2 system (Roche) using probes from the Universal Probe Library (Roche) and the TaqMan Master Mix (Roche) (for sequences of primers and the respective Universal Probe Library probes, see Supplementary Table 2 online). For each transcript, eight replications were performed and the results were normalized using bactin2 and gapdh. The relative expression software tool REST39 was used to analyze and normalize the data. It uses a hypothesis test to determine signicant differences between control and sample groups that performs 50,000 random reallocations of samples and controls

between the groups and counts the number of times the relative expression of the randomly assigned group is greater than that of the sample data.

SU5402 treatments. Dechorionated embryos were incubated from the 24-hpf stage into 10 mM SU5402 in Hanks embryo medium (0.137 M NaCl, 5.4 mM KCl, 0.25 mM Na2HPO4, 0.44 mM KH2PO4, 1.3 mM CaCl2, 1.0 mM Mg SO,4.2 mM NaHCO3), 0.1% DMSO. Mock-treated embryos were incubated for the same duration in Hanks medium containing only 0.1% DMSO. Treated and control embryos were xed for ISH at 35 hpf.

ISH and immunohistochemistry. Probe synthesis, ISH and immunohisto-chemistry were carried out as described8. The following in situ antisense RNA probes were used: her5 (ref. 40), fgf8 (ref. 13), fgfr1 (ref. 41), pax2a42, wnt1

(ref. 43), eng2a44, pea3 (ref. 16), dusp6 (ref. 15), shh45, nkx6.1 (ref. 46), foxa2 (ref. 47) and egr2b48. For ISH of miR-9, a digoxigenin tail was added to the miRCURY detection probe (LNA) hsamiR-9 (Exiqon) using the DIG tailing Kit (Roche). The miR-9 LNA probe was hybridized at 45 1C, and all other steps were carried out as described8. For sagittal or cross sections, embryos were cryostat-sectioned. We used chicken antibody to GFP (1:500, Aves Labs), rabbit antibody to GFP (TP401, 1:500, AMS), mouse antibody to human neural protein HuC/HuD (A-21271, 1:300, MoBiTec) and human antibody to Hu (kindly provided by B. Zalc (INSERM U711, Hpital de la Salptrire), 1:600) as primary antibodies for immunohistochemistry, and goat antibody to rabbit Cy2, goat antibody to rabbit Cy3, goat antibody to human Cy5, goat antibody to mouse Cy5, goat antibody to mouse Cy3 (all Jackson Laboratories) goat antibody to chicken Alexa488, goat antibody to rabbit Alexa555 and goat antibody to mouse Alexa647 (Invitrogen) as secondary antibodies. Embryos and at mounts were photographed under a Zeiss Axioplan photomicroscope, and sections were photographed and analyzed under a Zeiss confocal microscope (LSM 510 Meta). For the statistical analysis in Figure 4f,g, we used three embryos per background, and counted 1,505 cells for control, 2,040 for miR-9 injected and 1,853 for ace embryos to calculate the percentage of HuC-positive cells in the MH and the percentage of pH3-positive cells. In Figure 7d, the size of the HuC-positive area was calculated as a percentage of the size of the neural tube (n 3 embryos per assay; 3 cross-sections per embryo were analyzed). Sensor assay. d2egfp was cloned into pCS2+ using EcoRI and XhoI. Three copies of putative miR-9binding sites of fgf8 (fgf8-1 and fgf8-2), fgfr1 (fgfr1-1 and fgfr1-2), canopy1, her5 and her9 were fused to the 3 end of d2egfp using XhoI and XbaI, as described previously49. The oligonucleotides in Supplementary Table 3 online were phosphorylated, annealed and ligated in pCS2+d2egfp (at XhoI/Xba1) for each binding site. The 3 UTRs of fgfr1, her5 and canopy1 were amplied using the primers in Supplementary Table 3 and fused to the 3 end of d2gfp. For fgfr1, we foundthe 3 UTR to be at least 226 bp longer than the published sequence (NM_152962), as we could amplify the 3 UTR from embryonic cDNA using the primers in Supplementary Table 3 that extend further in the 3 direction than in the published sequence. Full 3-UTR sequences carrying point mutations in the predicted miR-9binding sites of her5 and fgfr1 (binding site fgfr1-1) were engineered by Sloning, PCR amplied and fused to d2egfp, as described above. For all constructs, capped mRNA was generated using the mMESSAGE mMACHINE Kit (Ambion). The mRNA was injected with either miR-9 duplex or the control miR duplex containing a shufed miR-9 sequence. Embryos were pooled and prepared for western blot analysis at 7 hpf.

Western blot. Embryos were pooled (ten embryos each) and SDS sample buffer was added, followed by a denaturation at 95 1C for 5 min, vortexing and a second incubation at 95 1C. The protein solution was centrifuged for 2 min in a table-top centrifuge. The samples were directly loaded onto a NuPage 10% BisTris gel (Invitrogen) with MOPS running buffer. The gels were blotted onto a PVDF Hybond-P membrane (GE Healthcare) and blocked with 4% nonfat dry milk (wt/vol) in phosphate-buffered saline (with 1% Tween-20, wt/vol). Immunodetection was carried out using rabbit antibody to GFP (TP401, 1:10,000 to 1:2,000, Torrey Pines Biolabs) and b-tubulin (ab6046, 1:15,000, Abcam) as primary antibodies, and antibody to rabbit HRP (1:10,000, Jackson Laboratories) as the secondary antibody. Detection was carried out using the Western Lightning chemiluminescence reagent (Perkin Elmer). From pools of 10 embryos, the equivalent of one embryo was loaded per lane; for statistical

http://www.nature.com/natureneuroscience

Nature Publishing Group

200 8

NATURE NEUROSCIENCE VOLUME 11 [ NUMBER 6 [ JUNE 2008 647

ARTICLES

analysis, three injected batches were each analyzed by densitometry (n 30 embryos for each binding site). The blots were scanned and the bands were quantied using ImageJ (US National Institutes of Health). The intensities of all bands were normalized with respect to the b-tubulin signal to circumvent loading inaccuracies.

Note: Supplementary information is available on the http://www.nature.com/natureneuroscience

Web End =Nature Neuroscience website. ACKNOWLEDGMENTSWe are grateful to members of the L.B.-C. laboratory for discussions and toM. Gtz, W. Norton and M. Wassef for their insightful ideas and critical reading of the manuscript. This work was funded by a junior group grant from the Volkswagen Association, the EU 6th framework integrated project ZF-Models (contract No. LSHC-CT-2003-503466), the Life Science Association (No. GSF 2005/01), a special research grant from the Institut du Cerveau et de la Moelle pinire, the Excellence Center for Protein Science, Munich, and the Helmholtz Impuls und Vernetzungsfond.

AUTHORS CONTRIBUTIONSC.L. and C.S. jointly conducted the experiments. A.W. and R.K. conducted parallel analyses in chicken to support the ndings described here. A.F. provided technical assistance and L.B.-C. supervised the project. C.L. and L.B.-C. wrote the manuscript.

Published online at http://www.nature.com/natureneuroscience

Web End =http://www.nature.com/natureneuroscience Reprints and permissions information is available online at http://npg.nature.com/reprintsandpermissions

Web End =http://npg.nature.com/ http://npg.nature.com/reprintsandpermissions

Web End =reprintsandpermissions

1. Wurst, W. & Bally-Cuif, L. Neural plate patterning: upstream and downstream of the isthmic organizer. Nat. Rev. Neurosci. 2, 99108 (2001).

2. Scholpp, S., Groth, C., Lohs, C., Lardelli, M. & Brand, M. Zebrash fgfr1 is a member of the fgf8 synexpression group and is required for fgf8 signalling at the midbrain-hindbrain boundary. Dev. Genes Evol. 214, 285295 (2004).

3. Stigloher, C., Chapouton, P., Adolf, B. & Bally-Cuif, L. Identication of neural progenitor pools by E(Spl) factors in the embryonic and adult brain. Brain Res. Bull. 75, 266273 (2008).

4. Bally-Cuif, L., Goridis, C. & Santoni, M.J. The mouse NCAM gene displays a biphasic expression pattern during neural tube development. Development 117, 543552 (1993).

5. Vaage, S. The segmentation of the primitive neural tube in chick embryos (Gallus domesticus). A morphological, histochemical and autoradiographical investigation. Ergeb. Anat. Entwicklungsgesch. 41,387 (1969).

6. Geling, A. et al. bHLH transcription factor Her5 links patterning to regional inhibition of neurogenesis at the midbrain-hindbrain boundary. Development 130, 15911604 (2003).

7. Geling, A., Plessy, C., Rastegar, S., Strahle, U. & Bally-Cuif, L. Her5 acts as a prepattern factor that blocks neurogenin1 and coe2 expression upstream of Notch to inhibit neurogenesis at the midbrain-hindbrain boundary. Development 131, 19932006 (2004).

8. Ninkovic, J. et al. Inhibition of neurogenesis at the zebrash midbrain-hindbrain boundary by the combined and dose-dependent activity of a new hairy/E(spl) gene pair. Development 132, 7588 (2005).

9. Hans, S. et al. her3, a zebrash member of the hairy-E(spl) family, is repressed by Notch signalling. Development 131, 29572969 (2004).

10. Hirata, H., Tomita, K., Bessho, Y. & Kageyama, R. Hes1 and Hes3 regulate maintenance of the isthmic organizer and development of the mid/hindbrain. EMBO J. 20, 44544466 (2001).

11. Hirate, Y. & Okamoto, H. Canopy1, a novel regulator of FGF signaling around the midbrain-hindbrain boundary in zebrash. Curr. Biol. 16, 421427 (2006).

12. Lun, K. & Brand, M. A series of no isthmus (noi) alleles of the zebrash pax2.1 gene reveals multiple signaling events in development of the midbrain-hindbrain boundary. Development 125, 30493062 (1998).

13. Reifers, F. et al. Fgf8 is mutated in zebrash acerebellar (ace) mutants and is required for maintenance of midbrain-hindbrain boundary development and somitogenesis. Development 125, 23812395 (1998).

14. Fjose, A. & Drivenes, O. RNAi and microRNAs: from animal models to disease therapy. Birth Defects Res. C. Embryo Today 78, 150171 (2006).

15. Kawakami, Y. et al. MKP3 mediates the cellular response to FGF8 signaling in the vertebrate limb. Nat. Cell Biol. 5, 513519 (2003).

16. Munchberg, S.R., Ober, E.A. & Steinbeisser, H. Expression of the Ets transcription factors erm and pea3 in early zebrash development. Mech. Dev. 88, 233236 (1999).

17. Yan, X. et al. Improving the prediction of human microRNA target genes by using ensemble algorithm. FEBS Lett. 581, 15871593 (2007).

18. Tallafuss, A. & Bally-Cuif, L. Tracing of her5 progeny in zebrash transgenics reveals the dynamics of midbrain-hindbrain neurogenesis and maintenance. Development 130, 43074323 (2003).

19. Choi, W.Y., Giraldez, A.J. & Schier, A.F. Target protectors reveal dampening and balancing of Nodal agonist and antagonist by miR-430. Science 318, 271274 (2007).

20. Valencia-Sanchez, M.A., Liu, J., Hannon, G.J. & Parker, R. Control of translation and mRNA degradation by miRNAs and siRNAs. Genes Dev. 20, 515524 (2006).

21. Bae, Y.K., Shimizu, T. & Hibi, M. Patterning of proneuronal and inter-proneuronal domains by hairy- and enhancer of splitrelated genes in zebrash neuroectoderm. Development 132, 13751385 (2005).

22. Wienholds, E. et al. MicroRNA expression in zebrash embryonic development. Science

309, 310311 (2005).

23. Darnell, D.K. et al. MicroRNA expression during chick embryo development. Dev. Dyn. 235, 31563165 (2006).

24. Kapsimali, M. et al. MicroRNAs show a wide diversity of expression proles in the developing and mature central nervous system. Genome Biol. 8,R173 (2007).

25. Kloosterman, W.P., Wienholds, E., de Bruijn, E., Kauppinen, S. & Plasterk, R.H. In situ detection of miRNAs in animal embryos using LNA-modied oligonucleotide probes. Nat. Methods 3, 2729 (2006).

26. Mueller, T. & Wullimann, M.F. Anatomy of neurogenesis in the early zebrash brain. Brain Res. Dev. Brain Res. 140, 137155 (2003).

27. Stark, A., Brennecke, J., Bushati, N., Russell, R.B. & Cohen, S.M. Animal MicroRNAs confer robustness to gene expression and have a signicant impact on 3 UTR evolution. Cell 123, 11331146 (2005).

28. Leaman, D. et al. Antisense-mediated depletion reveals essential and specic functions of microRNAs in Drosophila development. Cell 121, 10971108 (2005).

29. Li, Y., Wang, F., Lee, J.A. & Gao, F.B. MicroRNA-9a ensures the precise specication of sensory organ precursors in Drosophila. Genes Dev. 20, 27932805 (2006).

30. Nakamura, Y. et al. The bHLH gene hes1 as a repressor of the neuronal commitment of CNS stem cells. J. Neurosci. 20, 283293 (2000).

31. Wu, J. & Xie, X. Comparative sequence analysis reveals an intricate network among REST, CREB and miRNA in mediating neuronal gene expression. Genome Biol. 7,R85 (2006).

32. Conaco, C., Otto, S., Han, J.J. & Mandel, G. Reciprocal actions of REST and a microRNA promote neuronal identity. Proc. Natl. Acad. Sci. USA 103, 24222427 (2006).

33. Brand, M. et al. Mutations in zebrash genes affecting the formation of the boundary between midbrain and hindbrain. Development 123, 179190 (1996).

34. Kimmel, C.B., Ballard, W.W., Kimmel, S.R., Ullmann, B. & Schilling, T.F. Stages of embryonic development of the zebrash. Dev. Dyn. 203, 253310 (1995).

35. Enright, A.J. et al. MicroRNA targets in Drosophila. Genome Biol. 5,R1 (2003).36. Rusinov, V., Baev, V., Minkov, I.N. & Tabler, M. MicroInspector: a web tool for detection of miRNA binding sites in an RNA sequence. Nucleic Acids Res. 33, W696700 (2005).

37. Miranda, K.C. et al. A pattern-based method for the identication of MicroRNA binding sites and their corresponding heteroduplexes. Cell 126, 12031217 (2006).

38. Rehmsmeier, M., Steffen, P., Hochsmann, M. & Giegerich, R. Fast and effective prediction of microRNA/target duplexes. RNA 10, 15071517 (2004).

39. Pfaf, M.W., Horgan, G.W. & Dempe, L. Relative expression software tool (REST) for group-wise comparison and statistical analysis of relative expression results in real-time PCR. Nucleic Acids Res. 30, e36 (2002).

40. Mller, M., von Weizsacker, E. & Campos-Ortega, J.A. Transcription of a zebrash gene of the hairy-Enhancer of split family delineates the midbrain anlage in the neural plate. Dev. Genes Evol. 206, 153160 (1996).

41. Tonou-Fujimori, N. et al. Expression of the FGF receptor 2 gene (fgfr2) during embryogenesis in the zebrash Danio rerio. Mech. Dev. 119 Suppl 1: S173S178 (2002).

42. Krauss, S. et al. Zebrash pax(zf-a): a paired box-containing gene expressed in the neural tube. EMBO J. 10, 36093619 (1991).

43. Molven, A., Njolstad, P.R. & Fjose, A. Genomic structure and restricted neural expression of the zebrash wnt-1 (int-1)gene. EMBO J. 10, 799807 (1991).

44. Ekker, M., Wegner, J., Akimenko, M.A. & Westereld, M. Coordinate embryonic expression of three zebrash engrailed genes. Development 116, 10011010 (1992).

45. Hammerschmidt, M., Bitgood, M.J. & McMahon, A.P. Protein kinase A is a common negative regulator of Hedgehog signaling in the vertebrate embryo. Genes Dev. 10, 647658 (1996).

46. Cheesman, S.E., Layden, M.J., Von Ohlen, T., Doe, C.Q. & Eisen, J.S. Zebrash and y Nkx6 proteins have similar CNS expression patterns and regulate motoneuron formation. Development 131, 52215232 (2004).

47. Strahle, U., Blader, P., Henrique, D. & Ingham, P.W. Axial, a zebrash gene expressed along the developing body axis, shows altered expression in cyclops mutant embryos. Genes Dev. 7, 14361446 (1993).

48. Oxtoby, E. & Jowett, T. Cloning of the zebrash krox-20 gene (krx-20) and its expression during hindbrain development. Nucleic Acids Res. 21, 10871095 (1993).

49. Giraldez, A.J. et al. MicroRNAs regulate brain morphogenesis in zebrash. Science 308, 833838 (2005).

http://www.nature.com/natureneuroscience

Nature Publishing Group

200 8

648 VOLUME 11 [ NUMBER 6 [ JUNE 2008 NATURE NEUROSCIENCE

Word count: 8478

Show less

MicroRNA-9 directs late organizer activity of the midbrain-hindbrain boundary

Content area

Abstract

Full text