B RIEFINGS IN FUNC TIONAL GENOMICS . VOL 13. NO 2. 131^143
doi:10.1093/bfgp/elt044
Gene transcription in the zebrafish
embryo: regulators and networks
Marco Ferg, Olivier Armant, LixinYang, Thomas Dickmeis, Sepand Rastegar and Uwe Stra«hle
Advance Access publication date 22 October 2013
Abstract
The precise spatial and temporal control of gene expression is a key process in the development, maintenance and
regeneration of the vertebrate body. A substantial proportion of vertebrate genomes encode genes that control
the transcription of the genetic information into mRNA. The zebrafish is particularly well suited to investigate
gene regulatory networks underlying the control of gene expression during development due to the external development of its transparent embryos and the increasingly sophisticated tools for genetic manipulation available for
this model system. We review here recent data on the analysis of cis-regulatory modules, transcriptional regulators
and their integration into gene regulatory networks in the zebrafish, using the developing spinal cord as example.
Keywords: zebrafish; transcription regulator; gene regulatory network; cis-regulatory element; genomics
INTRODUCTION
The first step of gene expression is transcription.
Differential gene transcription plays a crucial role in
the development of the vertebrate body plan. Also,
many homoeostatic processes and responses in the
adult organism are regulated by transcription.
Thus, misexpression of genes due to deregulated
transcription can lead to a variety of diseases including cancer and type II diabetes [1–3]. The importance of differential transcriptional regulation as a
means to correctly express the information encoded
in the genome is further underscored by the fact that
a substantial proportion of genes in the genomes
code for proteins that are involved in the control
of transcription, such as transcription factors (TFs),
chromatin-remodelling factors and proteins of the
basic transcription machinery [4, 5]. Complex gene
regulatory networks (GRNs) form the basis of
cell differentiation during development of multicellular organisms and can control entire batteries
of genes coordinately. In GRNs, specific combinations of transcriptional regulators (TRs) interact
with cis-regulatory modules (CRMs) of responsive
genes [6–8].
CRMs can be functionally subdivided into two
groups: activating modules (promoter, enhancer,
locus control region [LCR]) and repressive modules
(silencer, insulator). Silencers inhibit gene expression
through the interaction with repressor proteins
that in turn initiate heterochromatin formation by
recruiting histone-modifying enzyme complexes.
Insulators confine the activity of an enhancer or
Corresponding author. Sepand Rastegar, Institute of Toxicology and Genetics, Karlsruhe Institute of Technology (KIT), Postfach
3640, 76021 Karlsruhe, Germany. Tel: þ49-721-608-22886; Fax: þ49-721-608-23354; E-mail: [email protected]
Marco Ferg is a post-doctoral fellow at the Institute of Toxicology and Genetics (ITG). His research interest lies in the transcriptional
regulation of genes by cis-regulatory elements.
Olivier Armant is a scientist at the ITG. The research he is interested in concerns early patterning mechanism in the developing
zebrafish brain.
Lixin Yang is a project leader at the ITG. His main research topic concerns the identification of gene regulatory networks in the
developing neural tube of the zebrafish.
Thomas Dickmeis is a group leader at the ITG. His group is interested in transcriptional regulatory mechanisms at the interface
between metabolism, endocrine signalling and the circadian clock.
Sepand Rastegar is a project leader at the ITG. His work is focused on the identification and functional analysis of cis-regulatory
elements active in the midline of developing zebrafish embryos. He is also interested in the molecular mechanisms controlling zebrafish
neurogenesis.
Uwe Stra«hle, professor at the University of Heidelberg, is interested in the molecular mechanisms controlling vertebrate development.
He is director of the ITG and established the European Zebrafish Resource Centre.
ß The Author 2013. Published by Oxford University Press. All rights reserved. For permissions, please email: [email protected]
132
Ferg et al.
silencer to its cognate gene promoter, and interaction
of CCCTC-binding factor (CTCF) with insulators
has been demonstrated to be crucial to shield distal
enhancers/silencers [9]. This classical view of CTCF
being a major marker for insulator elements has been
questioned by 4C experiments showing that many
of the CTCF-mediated long range interacting sites
correlate with enhancer–promoter interactions.
Further experiments showed that CTCF facilitates
enhancer–promoter interactions directly through
cohesin-mediated DNA looping [10–12].
Enhancers have been traditionally defined as a
class of cis-regulatory elements that are able to activate and maintain gene expression in a location- and
orientation-independent manner [13, 14]. They are
found in the up- and downstream regions as well as
in introns of the regulated gene. However, enhancers
can also overlap with coding exons [15] or the
50 -untranslated region (UTR) [16]. They may be
located as far as 1 megabase from the target gene,
as has been shown, for example, for a limb-specific
enhancer of the sonic hedgehog gene [17].
Enhancers are thought to represent islands of open
chromatin that act as docking stations for TFs [18].
By a combination of distinct TFs that bind to the
same enhancer, various signalling inputs can be integrated and interpreted, thereby leading to specific
gene expression outputs. In the prevailing model,
the physical interaction of enhancers with their
target promoters is facilitated by a decondensation
of the chromatin structure enabling the DNA to
loop out from the nuclear periphery to the centre
to form enhancer/promoter interactions [19–21].
A recent publication suggests CTCF/cohesinmediated DNA looping is a determinant for promoter choice [22]. However, not all loci require
looping-out for transcriptional activity [23].
Promoters can be functionally subdivided into the
proximal promoter and the core promoter. Proximal
promoters are located in the immediate vicinity of
the transcription start site (TSS) and contain, like
enhancers or silencers, recognition sites for DNAbinding factors that are involved in transcriptional
regulation. The core promoter is usually defined as
a 100–150 bp stretch of DNA encompassing the TSS
that is sufficient to direct initiation of transcription by
the RNA polymerase II machinery [24]. Core promoters can differ considerably in motif composition
(reviewed in [25–27]) and can be bound by a multitude of promoter recognition complexes [28, 29].
Not all enhancers are able to interact with every
type of core promoter [18, 30, 31], providing
another level of regulatory control.
LCRs are, like enhancers, activating CRMs. They
possess chromatin domain-opening activity and are
able to confer ‘copy number-dependent’ expression
on a linked gene, i.e. the number of integrated transgenes determines the expression level and not the
influence of the chromatin surrounding the transgenes (reviewed e.g. in [32, 33]).
Interestingly, enhancers and silencers share important properties, such as DNase I hypersensitivity,
active chromatin marks and interaction with transcription factors and RNA polymerase II. It has,
therefore, been proposed that it depends on the
gene target and the developmental stage whether a
given CRM activates or represses gene transcription
[34].
CRMs contain the information required for
precise transcriptional regulation. The implementation of this information is carried out by TRs
through sequence-specific interactions with CRMs
of their target genes. As described above, TFs activate
or repress gene transcription by binding to enhancers
or silencers, respectively. In addition, general TFs
are involved in the basic processes of transcription,
such as initiation, elongation and termination of
transcription. However, they may be less ‘general’
than previously thought, as recent evidence suggests
gene-specific employment of these factors [35, 36].
Finally, chromatin remodelling proteins alter the
state of the chromatin and thereby affect transcription by regulating access to the CRMs.
In summary, GRNs can be defined as the implementation of the regulatory information encoded
in the DNA of CRMs via their interactions with
TRs. Modelling a GRN related to any development
process requires knowledge of which TFs and signalling molecules are involved, when and where the
genes are expressed and how these factors interact
with each other and with CRMs [8]. The zebrafish
is an ideal model to study the components and composition of GRNs. Its fast and external development
and the high number of transparent embryos obtainable have proven very useful to address questions in
transcriptional regulation. Zebrafish and other fish
species have been utilized to analyse developmental
GRNs [37, 38] and promoter/enhancer interactions
[31], to validate and analyse enhancer function on
a large scale [39–44], to address the regulatory
potential of short sequences [45] and to identify
regulatory regions driving tissue-specific expression
Gene regulatory networks and control of gene expression
in enhancer trap assays [46, 47]. Furthermore, recent
progress in defining gene expression profiles on a
global scale [4] demonstrates that zebrafish is a
highly useful organism to comprehensively study
GRNs in a vertebrate organism.
In this review, we will illustrate how zebrafish can
be used to understand the regulatory logic of the
vertebrate genome. We will summarize the current
knowledge on specific aspects of CRMs regarding
their distribution and the genomic features that
have been associated with CRMs, examine the repertoire of TRs in the zebrafish genome and discuss
one example of a transcriptional regulatory network
that is well analysed in both zebrafish and mice.
SEQUENCE CONSERVATION AND
HISTONE MODIFICATIONS
IDENTIFY CRMS
CRMs are frequently characterized by conservation
of sequence across species. DNA sequences with
regulatory functions exhibit a lower mutation rate
than non-functional DNA and therefore, like protein coding sequences, show significant sequence
similarity across species. The Fugu rubripes genome
was the first vertebrate genome to be assembled
[48] after sequencing of the human genome [49,
50]. Comparative studies of the human and Fugu
genomes established the presence of a large number
of highly conserved non-coding elements (HCNEs)
[51–54]. Since then, the growing number of wholegenome sequences from various organisms and the
development of bioinformatic tools [55–59] have
made comparative genomics a powerful means to
identify putative CRMs [60–63].
The binding of TRs to CRMs provides one
mechanism of transcriptional regulation. Another
level is introduced by the structure of the chromatin,
which can control the accessibility of the CRMs.
The fact that active regulatory CRMs are in an
open chromatin state forms the basis of large-scale
mapping of CRMs by DNAse I hypersensitivity
assays [64]. DNAse I hypersensitivity provides a
comprehensive picture of the active regulatory landscape of the genome in a particular cell, because it is a
key feature of all classes of CRMs [65, 66]. Specific
post-translational modifications of histones such as
methylation and acetylation at particular positions
are further landmarks of active and inactive chromatin structure of CRMs. For example, active enhancers have been shown to be marked by
133
monomethylation of H3K4 (H3K4me1) [67–69].
However, this chromatin modification is not limited
only to active enhancers. Several studies suggested
the existence of both active and poised classes of
enhancer elements. Active enhancers in a cell are
engaged in the regulation of transcription, whereas
poised enhancers are not actively regulating transcription, but can be rapidly converted to active
enhancers, e.g. by external stimuli to the cell or in
the course of differentiation [70]. The histone marks
associated with poised enhancers contain both the
H3K4me1 mark and trimethylation at K27 of histone H3 (H3K27me3). Active enhancers are distinguished from poised enhancers by the presence of
acetylation of histone H3 at lysine 27 (H3K27ac)
and by the absence of H3K27me3 [70–72]. Besides
these post-translational modifications of histones,
binding of chromatin remodelling enzymes has
been shown to designate active enhancer modules.
Binding of the acetyltransferase and transcriptional
co-activator p300, for example, predicted with
high accuracy where active enhancers are located
in the genome in neural, limb and heart tissue [69,
73, 74]. Furthermore, the chromatin remodelling
enzymes CHD7 and BRG1 appear to be present at
active enhancers [71, 72].
CRMS IN THE ZEBRAFISH
GENOME
Most genome-wide studies of the regulatory landscape have focused on mammalian cell line, while
the dynamic changes of the histone code and TF
occupancy of CRMs on a global scale during ontogenesis of a vertebrate are still unresolved. The zebrafish with its experimentally easily accessible early
stages of development has a great potential to address
these questions. ChIP (chromatin immunoprecipitation) on chip experiments confirmed enrichment of
the histone mark H3K4me3 found in active promoters [75] also near the TSS in zebrafish [76].
Interestingly, many inactive developmental regulatory genes are marked by the repressive mark
H3K27me3, while at the same time containing the
activating chromatin modification H3K4me3, thus
showing the characteristic signature of poised genes
that is also found in developmental regulators in
mouse embryonic stem cells [77, 78]. This ‘bivalent’
signature was not detected before midblastula transition, when the activation of the zygotic genome
occurs [79]. It was suggested that such marks might
134
Ferg et al.
secure precise timing of the activation of developmental regulators [80].
Combining H3K4me1 and H3K4me3 profiles,
signatures of active enhancers (H3K4me1) and
active promoters (H3K4me3) in combination with
the readily available expression data in zebrafish
helped to computationally identify putative cisregulatory sequences [39]. The generation of genomic tracks of H3K4me3, H3K4me1 and H3K27ac
histone modifications at four developmental timepoints of zebrafish embryogenesis lead to the
identification of 50 000 potential cis-regulatory
elements operating during the first 48 h of zebrafish
development, a wealth of data that will help to
deduce regulatory networks on the DNA sequence
level [81].
Although many data are now available on
the location of putative CRMs involved in embryonic development, a systematic analysis of the function of CRMs has not yet been carried out on a large
scale in zebrafish. One crucial question is how conserved the functions of the CRMs encoded in
HCNEs are. Case studies illustrated several possibilities. For example, the lateral stripe enhancer of the
neurogenin1 gene retained its function for about 395
million years of evolution, as it drives expression in
the dorsal telencephalon of both mouse and zebrafish
[82]. In contrast, the function of CRMs in the shh
gene that are conserved between mouse and zebrafish seems to have changed: the zebrafish sonic hedgehog (shh) enhancer ar-C directs mainly notochord
expression in zebrafish embryos and also functions
in the midline of mouse embryos. However, the
mouse enhancer SFPE2, which exhibits sequence
similarity with zebrafish ar-C, is floor plate-specific
in the mouse and not functional in zebrafish.
Another important question is why CRMs of
developmental regulators show higher sequence
conservation than CRMs of other genes; why do
not all conserved sequences have an obvious
function? Why are not all functional CRMs highly
conserved? Moreover, it is more than questionable
whether up to 90% conservation is necessary to
maintain the binding of the same TFs, given the
degeneracy of the recognition sequences of many
TFs. In fact, functional dissection of five notochord
enhancers with evolutionarily conserved sequences
revealed two short DNA elements mediating notochord expression. However, these elements were
embedded within the conserved sequences of
the five enhancers in an almost random fashion,
suggesting that the conservation of the sequence is
only indirectly linked to binding of TFs [83]. The
findings that non-coding RNAs bi-directionally
transcribed on enhancers (eRNAs) [84] significantly
increase promoter–enhancer interactions [85] and
that activating ncRNAs (ncRNA-a) play an important role in establishing and maintaining chromatin
structure [86, 87] opened up a new chapter in cisregulation of gene expression and might help to
answer some of these questions.
Taken together, analysis of CRMs of the zebrafish
genome now provides a framework to understand
precisely how classes of CRMs are composed, how
they interact with epigenetic marks and why some
of them are conserved across large evolutionary
distances. More functional analyses will be required,
and one important element of these is knowledge
about which transcriptional regulators interact with
the CRMs.
CHARACTERIZATION OF THE
TRANSCRIPTIONAL
REGULATORY REPERTOIRE OF
THE ZEBRAFISH GENOME
The availability of the entire zebrafish genome
sequence [88] and the elucidation of its gene structures by computational means to identify or predict
TSSs, exon/intron structure, UTRs, protein
domains and regulatory elements, provides a rich
resource to mine genomes for genes involved in
transcriptional regulation. The InterPro database
provides a resource of functional annotation on proteins by classifying them into families. It compiles
protein signatures from a number of databases,
thereby integrating all data into a unique searchable
resource [89]. A thorough search of the InterPro
protein domains led to the selection of 483 domains
associated with transcriptional regulation activities
(Figure 1A). Mining the zebrafish genome for
genes encoding at least one protein domain linked
to transcriptional regulation revealed 3302 genes,
including TFs with specific DNA binding domains,
genes involved in chromatin modification and factors
of the general transcriptional machinery (Figure 1B).
Of these, 2488 TR genes can be detected reliably by
deep sequencing (349 million 76 bp long paired-end
reads) of mRNA isolated from 24 h post-fertilization
(hpf) embryos. This suggests that nearly 75% of the
TR gene repertoire of the zebrafish is expressed at a
significant level at 24 hpf [4], a stage of particular
Gene regulatory networks and control of gene expression
135
Figure 1: Zebrafish transcriptional regulators. (A) InterPro protein domains specific to transcriptional regulation.
The number of protein domains specific for each category of TR (transcription factor, chromatin remodelling,
basal transcriptional machinery) is indicated (InterPro release 19). (B) The number of genomic loci encoding genes
with at least one domain functionally linked to transcriptional control are indicated for each category (Zv9 assembly,
Ensembl version 60 annotation). (C) Expression patterns of chromatin remodellers of the BTB/POZ family in
24 hpf embryos. bach2 and zbtb46 are expressed in the telencephalon, the zbtb16 homologue is expressed in the
whole spinal cord and the posterior tuberculum in the diencephalon, zbtb43 is expressed in the retina and the
tectum and btbd6a is expressed in the epiphysis, the telencephalon, the hindbrain and the spinal cord, while its paralogue btbd6b is expressed in the somites, the telencephalon and other parts of the fore- and midbrain. hic1 and
bcl6b are expressed in the vascular system of the trunk and the head, whereas the bcl6b homologue is expressed
only in the vascular system of the head.
importance as it represents the evolutionarily conserved phylotypic stage of this model organism
[90]. The remaining TR genes may be expressed at
very low levels, below the detection limit, or at later
stages or only under certain physiological conditions.
Of the 3302 TR genes detected in the zebrafish
genome, 2677 genes can be assigned to the TF subclass of TRs. Precedent analysis of human and mouse
genomes suggested that 1500–2000 genomic loci
encode DNA-binding TFs [5], roughly 10 times
136
Ferg et al.
the size of the TF repertoire of the yeast Saccharomyces
cerevisiae (200–350 loci) [91]. However, the fraction
of the total number of genes that encode TFs does
not increase that much, as only a 2-fold increase is
observed from yeast (4% TF genes) to human (8%).
In zebrafish, 9% of all genes encode TFs, slightly
more compared with other vertebrates. This might
be due to the additional genome duplication during
the evolution of teleost fishes [92]. The increased
number of encoded TF genes alone cannot explain
the enormous complexity of the vertebrate body,
with many distinct cell types and complex body
functions, compared to unicellular organisms such
as yeast. The differences seem to lie mainly in the
combinatorial action of TRs as well as in specific
post-transcriptional and post-translational modifications that will give alternative outputs in different
cellular contexts [93–95].
The precise spatiotemporal expression of many
TR genes is essential for normal vertebrate development. In a systematic analysis of expression of 1711
TR genes in the 24 hpf zebrafish embryo, around
200 TR genes were found to be expressed in a single
tissue, whereas the vast majority of TRs (n ¼ 1504)
were detected either in multiple tissues or ubiquitously. The highly restricted expression pattern of
these 200 TR genes makes them prime candidates
for functional studies. This group of tissue-specific
genes is highly enriched for TFs (82%; n ¼ 207) [4].
The central nervous system and specifically the
spinal cord and telencephalon express a majority of
the TR genes (60% of the TRs analysed by in situ
hybridization). A significant correlation of genes coexpressed in different tissues was observed in a few
brain regions, including the ectodermally derived
cranial sensory ganglia (otic vesicle and olfactory
bulb) and several forebrain structures as well as the
retina and the tectum [4]. Interestingly, retina and
tectum are functionally coupled by topographical
projections of retinal axons into the tectum.
However, at a global level, there was no extensive
co-expression of specific TRs in multiple tissues at 24
hpf. This is in contrast to the components of cell
signalling pathways, which are frequently organized
into synexpression groups [96, 97]. Together, these
results show that there is no extensive overlap between the TR expression patterns. This suggests a
high flexibility in the combination of different factors
in order to generate alternative regulatory outputs.
Genes encoding general TFs and chromatin
remodelling proteins are usually ubiquitously
expressed at 24 hpf with the exception of the
BTB–POZ family (named after Broad complex,
Tramtrack and Bric à brac/poxviruses and zinc
finger). A significant proportion of these genes (38
out of the 118 mapped genes) have expression patterns restricted to the somites and the central nervous
system. The BTB/POZ domain is an evolutionary
conserved domain involved in protein–protein interactions leading to dimerization. Proteins containing
both the BTB/POZ domain and C2H2 zinc finger
DNA-binding domain have been shown to promote
transcriptional repression through the recruitment of
co-repressor proteins such as histone deacetylase
(HDAC), N-CoR and SMRT [98–100]. Six genes
of the BTB/POZ family are expressed in parts of the
nervous system of the 24 hpf zebrafish embryos
(Figure 1C). The tissue-restricted expression of these
TR genes suggests that some members of the BTB/
POZ—zinc finger family may control cellular differentiation during embryogenesis, possibly by the recruitment of transcriptional repressors such as HDACs.
Functions of ubiquitously expressed genes can still
vary dramatically from one tissue to another. For
example, brg1 (alias smarca4), a protein associated
with the Swi/Snf-like Brg1/Brm-associated factors
(BAF) chromatin remodelling complex, is expressed
in a broad range of tissues during development and
in the adult zebrafish. However, mutation of brg1
leads to specific defects in the zebrafish heart [101]
and in retinal neurogenesis [102, 103]. Furthermore,
disrupting the dosage balance between Brg1 and the
cardiac TF Tbx5 led to impaired heart development
in the mouse, indicating that normal formation of
this tissue relies on precise levels of functional BAF
complexes. These results highlight the importance of
ubiquitous chromatin remodelers in the acquisition
of specific cellular fates. The fact that most TRs are
expressed ubiquitously emphasizes the importance of
combinatorial action with tissue-restricted factors as
well as that of post-translational and post-transcriptional regulation.
Thus, data on the global repertoire of TRs and on
their expression during development are now complementing the CRM and epigenetic modification
data. In order to be able to model transcriptional
regulation during development, these data will
have to be assembled into a global GRN. This will
require large-scale functional analysis by perturbing
normal networks, e.g. by creating mutants or other
loss of function phenotypes for particular regulators
or even for CRMs. Such large-scale efforts might be
Gene regulatory networks and control of gene expression
built on smaller scale studies that were aimed at
understanding particular subprocesses of development. In the next paragraph, we will discuss an
example of such a study that illustrates the experimental advantages of the zebrafish model.
GENE REGULATORY NETWORKS
SPECIFYING ZEBRAFISH
VENTRAL SPINAL CORD
NEURONS
Many TFs act in specific cellular contexts, thereby
addressing distinct downstream genes or GRNs.
Frequently, TFs are organized in pathways, which
form a hierarchy of TF gene interactions. The challenge is to understand these cascades and to link
the action of TFs to specific cellular outcomes. The
vertebrate spinal cord offers well-studied examples
of how different cell types are specified by combinations of TFs.
The spinal cord is patterned along its dorsoventral
axis by antagonistic signalling interactions of the Shh,
the bone morphogenetic protein and the Wnt pathways [104]. In the ventral spinal cord, the morphogen Shh, which is secreted from the underlying
notochord and the floor plate of the spinal cord,
induces five distinct ventral neuronal subtypes via a
concentration gradient [105]. Distinct concentrations
of Shh induce or repress the spatial expression of TFs,
which belong predominantly to the homoeodomain
protein (HD) or basic helix–loop–helix (bHLH)
families (Figure 2; [105]). Class II TFs are induced,
while class I TFs are repressed by Shh (Figure 2C).
As a consequence of this Shh activity, the five ventral
neural progenitor domains express each a specific
combination of TF genes (Figure 2A and B).
Further cross-repressive interactions between these
TF proteins sharply delineate the boundaries
between the ventral neuronal domains [106]. These
progenitor domains will give rise to distinct populations of neuronal subtypes (Figure 2B). Close to the
ventral source of Shh, V3 interneurons form, followed at a more dorsal position by motoneurons
(MNs) and then by V2, V1 and V0 interneurons
further dorsally (Figure 2B).
The Gli family of zinc finger transcription factors
mediates Hedgehog (Hh) signalling in all vertebrates
by activating or repressing the expression of downstream target genes [108]. As is the case for mammals,
zebrafish Shh is expressed in the notochord and floor
plate and specifies ventral spinal cord fates [109].
137
Most of the TFs expressed in the spinal cord of the
mouse that mark its distinct domains are expressed in
the zebrafish spinal cord in a similar dorsoventral
pattern (Figure 2A). It was therefore proposed that
gene regulatory networks underlying spinal cord
patterning are highly conserved during vertebrate
evolution [109].
One of the advantages of the zebrafish model
organism is the existence of a huge collection of
mutants for all major signalling pathways. For
instance, more than 10 zebrafish mutations have
been shown to affect different components of the
Shh signalling cascade [110]. Analysis of zebrafish
mutants in Hh pathway components as well as
over-expression of Shh highlighted differences in
the requirement of Shh signalling in spinal cord
development in mouse and zebrafish embryos: in
contrast to mouse Shh signalling, zebrafish Hh
signalling appears to be partially dispensable for the
specification of some of the ventral neuronal subtypes [106, 111]. In the total absence of Hh signalling
in maternal and zygotic smu mutants, motoneuron
progenitors (pMN) are strongly reduced in numbers
and p3 progenitors are totally absent. In contrast, p2,
p1 and p0 progenitors as well as the post-mitotic V2,
V1 and V0v interneurons that develop from these
progenitor domains are present [111]. These results
suggest that the activity of this signalling pathway in
zebrafish is primarily required for the specification of
the p3 progenitor domains. As in mouse, the expression of nkx2.2 and nkx2.9 genes in the p3 domain is
dependent on Hh signals secreted from the floor
plate and the notochord [111–115]. The expression
of mouse Nkx2.2 and 2.9, and zebrafish nkx 2.2a and
2.9 is driven by conserved Shh responsive CRMs,
which are bound by Gli factors [115–118], suggesting that Hh directly regulates the expression of these
transcription factors. In line with this hypothesis, in
gli1(dtr) mutants the expression of the nkx2 genes is
greatly reduced, while misexpression of Gli1 results
in ectopic induction of nkx2.9. Furthermore, a construct containing the Gli consensus binding site of
the nkx2.9 enhancer drives green fluorescent protein
(GFP) expression in the ventral neural tube, and mutation of this Gli-binding site abolishes GFP expression [115]. In addition, morpholino-mediated
knockdown of nkx2a/b and nkx2.9 genes leads to an
expansion of olig2 expression, which is a marker of
pMN, into the p3 domain and abolishes the expression of the V3 marker sim1 (leucine zipper/PAS
domain) [38]. Thus, zebrafish nkx2a/b and nkx2.9
138
Ferg et al.
Figure 2: Schematic representation of the ventral spinal cord of mouse and zebrafish. (A) TF genes expressed
in the progenitor domains (p) of mouse (left) and zebrafish (right). The expression patterns of the TF genes in the
progenitors and post-mitotic interneurons of mouse and zebrafish are almost identical. (B) Ventral progenitor
domains and their corresponding post-mitotic neurons. (C) Distinct concentrations of Shh induce the spatial
expression of class II genes or repress class I gene expression (left). Progenitor domain borders are refined and
maintained by negative cross-regulatory interactions among proteins, which share common boundaries. Negative
cross-regulatory interaction can take place between class I and class II proteins or between proteins of the same
class (centre). Specific combinations of homoeodomain transcription factor proteins in each progenitor domain
determine the neuronal subtype (right). Figure modified from a previous study [107].
genes are required downstream of the Hh pathway
for the specification of the p3 domain and the production of V3 interneurons and they contribute to
the correct dorsoventral positioning of the pMN
domain by repressing olig2 expression in a negative
cross-regulatory interaction similar as in the mouse
[105].
Taken together, it appears that the transcription
factors and their role in the specification of different
progenitor domains and of post-mitotic interneurons
are largely conserved between mammals and zebrafish, but that the signals required for the correct
spatial and temporal expression of these factors are
only partially maintained. It has still to be determined
which other signals are required for the early expression of pMN, p2, p1 and p0 progenitor domain
genes in the zebrafish [111].
Thus, the patterns of expression of the TFs in the
mouse and zebrafish spinal cord are very similar and
the mechanisms underlying neuronal differentiation
appear to be conserved, a conclusion also supported
by other studies [112, 119, 120]. However, expression of TFs is not necessarily always an indicator
of maintained function. This is exemplified by the
GRNs controlling differentiation of two closely
related types of GABAergic interneurons, the
Kolmer–Agdhur neurons KA0 and KA00 [38].
Homologues of these cells have so far not been
described in the mouse spinal cord. In contrast to
other neurons in the spinal cord, the cell bodies of
both KA interneuron types are positioned close to
the central canal, into which they extend cilia [121].
KA0 cells are required for spontaneous free
swimming, while the function of KA00 neurons is
unknown [122]. The KA00 interneurons occupy a
ventral position and develop from the lateral floor
plate next to the progenitors of V3 interneurons.
The KA0 cells reside at a more dorsal position and
arise from the motoneuron progenitor domain [38,
123–125]. The specification and formation of KA
interneurons involve the Notch and Hh signalling
pathways [38, 123–125]. KA00 interneurons in the
lateral floor express the TFs nkx2.2a, nkx2.2b,
nkx2.9, tal2, gata2 and gata3 (Figure 3). The expression of tal2, gata2 and gata3 as well as of the GABAsynthesizing enzyme glutamic acid decarboxylase 67
(gad67) is shared between KA0 and KA00 cells
(Figure 3; [38, 119]). The functional relationships
Gene regulatory networks and control of gene expression
Figure 3: Scheme outlining the regulatory interactions in KA00 and KA0 interneurons. In KA0 cells
(right), tal2 and gata2 are expressed but not required
for gad67 expression. In contrast, these two genes are
necessary for gad67 expression in KA00 cells. In KA00
cells on the other hand, gata3 is expressed but not
functionally linked to gad67. The expression of TF genes
in different cells does not necessarily reflect conserved
function, even in a case as described here where the
same downstream gene (gad67) is expressed in both
cell types.
of these factors and their role in the specification of
the two distinct KA interneuron subtypes were
investigated in a systematic knockdown approach
using morpholinos designed against each one of
these TFs [38]. Differentiation of the correct
number of KA00 cells depends on the activity of
Nkx2.9 that acts in cooperation with Nkx2.2a and
b. All three of these nkx2 genes are necessary for the
expression of the zinc finger transcription factor gata2
and gata3 in KA00 cells. Gata2 but not Gata3 is necessary for expression of the bHLH transcription factor
tal2 that acts upstream of gad67 in KA00 cells
(Figure 3). The four genes, tal2, gata2 gata3, gad67 are
also expressed in KA0 cells, and they depend on the
bHLH transcription factor Olig2 rather than on
Nkx2 genes in these cells. Curiously, knock-down
of gata2 does not affect tal2 or gad67 expression in
KA0 cells (Figure 3) suggesting different functional
connections. Expression of tal2 and gad67 is rather
dependent on Gata3 in these cells. Moreover, tal2,
although like gata2 it is expressed in KA0 interneurons, is not required for gad67 expression in
these cells (Figure 3). Hence, the functional connections between the different regulatory genes differ in
139
the two KA cell types: although gata2 and tal2 are
also expressed in KA0 cells, they are dispensable for
gad67 expression in these cells. Instead, olig2 and
gata3 are required for the differentiation of gad67expressing KA0 cells (Figure 3; [38]).
These data suggest that the expression state of TFs
per se is not sufficient to determine the cell fate or
differentiation status of cells, and that one would
need to know the complete expression repertoire
of TRs to predict the fate of a cell. Also, the posttranscriptional and post-translational modifications of
the TRs might play a role as, for instance, the phosphorylation state of the Olig2 protein determines the
choice between the MN and the oligodendrocyte
cell fate [126, 127]. Even this information might
not be enough, as the epigenetic history of the cell
might also play a role by determining which CRMs
are accessible to the TR combination present in it.
CONCLUSIONS AND
PERSPECTIVES
Over the last 20 years, zebrafish has become a valuable model organism to address questions of gene
regulation and regulatory networks. This model has
been shown to be advantageous in reverse and
forward genetic screens, in reporter assays to study
enhancer/promoter function as well as in ChIP
studies targeting histone modifications and TRs,
thereby delivering the resources for an in-depth
analysis of GRNs. The above described analysis of
Kolmer–Agdhur interneuron differentiation demonstrated that although TFs can be expressed in two
specific cell types, they do not necessarily have the
same regulatory function in both cell types. It
simultaneously underlines the major advantage of
zebrafish: the potential to easily and rapidly address
complex questions where the questions were actually
phrased—in the organism. The examples of cell
specification in a single organ, which have been
presented here, also demonstrate what would be
required to investigate the mechanisms underlying
TR on a global scale throughout development.
A systematic knock out of TRs by the transcription
activator-like effector (TALE) [128] or CRISPR/
CAS9 [129] systems would help to understand
their integration in GRNs. TALEs could also be
utilized to identify and characterize CRMs.
Depending on the fused effector, nucleases can
excise CRMs, while repressors and activators could
temporarily change the activity of the targeted
CRM(s) [130]. Experiments like this could help to
140
Ferg et al.
resolve how multiple CRM modules work together
to define the expression domains of a gene and
would demonstrate if a CRM identified in reporter
assays is actually required for accurate expression in a
given context. The ENCODE project demonstrated
that functional enhancers can be identified by binding of TRs at distinct loci in multiple cell lines.
A similarly systematic ChIP-seq study in zebrafish
might shed light on the changes of CRM TR occupancy over time in development. Provided the
availability of good antibodies, TF-binding sites
could be discovered rapidly in fish. In summary,
the time is ready for a zebrafish ENCODE project
to systematically mine the embryo of this animal
system for regulatory interactions and to combine
these with expression states and systematic functional
studies of TRs.
4.
5.
6.
7.
8.
9.
10.
11.
Key Points
CRMs are DNA sequences that regulate gene expression.
TRs modulate gene expression on the protein level by
interaction with CRMs.
Gene regulatory networks are formed by a multi-layered
interaction of TRs with CRMs resulting in the precise regulation
of gene expression.
The midline composed of notochord and floor plate is an
organizing centre for the specification of neuronal fate in the
spinal cord controlling complex gene regulatory networks.
12.
13.
14.
15.
16.
FUNDING
The work in our laboratory was supported by
the EU IP ZF-Health FP7-Health-2009-242048;
NeuroXsys Health-F4-2009 No. 223262, the
Interreg network for synthetic biology in the
Upper Rhine valley (NSB-Upper Rhine); the
BMBF funded network Erasys Bio BMBF KZ:
0315716 and the BioInterfaces programme of the
Helmholtz association.
17.
18.
19.
20.
References
1.
2.
3.
Ghiasvand NM, Rudolph DD, Mashayekhi M, et al.
Deletion of a remote enhancer near ATOH7 disrupts retinal
neurogenesis, causing NCRNA disease. Nat Neurosci 2011;
14:578–86.
Lee TI, Young RA. Transcriptional regulation and its
misregulation in disease. Cell 2013;152:1237–51.
Lettice LA, Horikoshi T, Heaney SJ, et al. Disruption of
a long-range cis-acting regulator for Shh causes preaxial
polydactyly. Proc Natl Acad Sci USA 2002;99:7548–53.
21.
22.
23.
Armant O, Marz M, Schmidt R, et al. Genome-wide,
whole mount in situ analysis of transcriptional regulators
in zebrafish embryos. Dev Biol 2013;380(2):351–62.
Vaquerizas JM, Kummerfeld SK, Teichmann SA, et al.
A census of human transcription factors: function, expression and evolution. Nat Rev Genet 2009;10:252–63.
Davidson EH. Emerging properties of animal gene regulatory networks. Nature 2010;468:911–20.
Ettensohn CA. Encoding anatomy: Developmental gene
regulatory networks and morphogenesis. Genesis 2013;51:
383–409.
Peter IS, Davidson EH. Evolution of gene regulatory
networks controlling body plan development. Cell 2011;
144:970–85.
Bell AC, West AG, Felsenfeld G. The protein CTCF is
required for the enhancer blocking activity of vertebrate
insulators. Cell 1999;98:387–96.
Handoko L, Xu H, Li G, et al. CTCF-mediated functional
chromatin interactome in pluripotent cells. Nat Genet 2011;
43:630–8.
Parelho V, Hadjur S, Spivakov M, et al. Cohesins functionally associate with CTCF on mammalian chromosome
arms. Cell 2008;132:422–33.
Rubio ED, Reiss DJ, Welcsh PL, et al. CTCF physically
links cohesin to chromatin. Proc Natl Acad Sci USA 2008;
105:8309–14.
Banerji J, Olson L, Schaffner W. A lymphocyte-specific cellular enhancer is located downstream of the joining region in
immunoglobulin heavy chain genes. Cell 1983;33:729–40.
Banerji J, Rusconi S, Schaffner W. Expression of a betaglobin gene is enhanced by remote SV40 DNA sequences.
Cell 1981;27:299–308.
Tumpel S, Cambronero F, Sims C, et al. A regulatory
module embedded in the coding region of Hoxa2 controls
expression in rhombomere 2. Proc Natl Acad Sci USA 2008;
105:20077–82.
McLellan AS, Kealey T, Langlands K. An E box in the exon
1 promoter regulates insulin-like growth factor-I expression
in differentiating muscle cells. AmJ Physiol Cell Physiol 2006;
291:C300–7.
Lettice LA, Heaney SJ, Purdie LA, et al. A long-range Shh
enhancer regulates expression in the developing limb and
fin and is associated with preaxial polydactyly. Hum Mol
Genet 2003;12:1725–35.
Li Q, Barkess G, Qian H. Chromatin looping and the
probability of transcription. Trends Genet 2006;22:197–202.
Carter D, Chakalova L, Osborne CS, et al. Long-range
chromatin regulatory interactions in vivo. Nat Genet 2002;
32:623–6.
Chambeyron S, Bickmore WA. Chromatin decondensation
and nuclear reorganization of the HoxB locus upon induction of transcription. Genes Dev 2004;18:1119–30.
Su W, Porter S, Kustu S, et al. DNA-looping and enhancer
activity: association between DNA-bound NtrC activator
and RNA polymerase at the bacterial glnA promoter.
Proc Natl Acad Sci USA 1990;87:5504–8.
Guo Y, Monahan K, Wu H, et al. CTCF/cohesin-mediated
DNA looping is required for protocadherin alpha promoter
choice. Proc Natl Acad Sci USA 2012;109:21081–6.
Morey C, Da Silva NR, Perry P, et al. Nuclear reorganisation and chromatin decondensation are conserved, but
Gene regulatory networks and control of gene expression
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
distinct, mechanisms linked to Hox gene activation.
Development 2007;134:909–19.
Butler JE, Kadonaga JT. The RNA polymerase II core
promoter: a key component in the regulation of gene
expression. Genes Dev 2002;16:2583–92.
Kadonaga JT. Perspectives on the RNA polymerase II core
promoter. Wiley Interdiscip Rev Dev Biol 2012;1:40–51.
Juven-Gershon T, Kadonaga JT. Regulation of gene
expression via the core promoter and the basal transcriptional machinery. Dev Biol 2010;339:225–9.
Thomas MC, Chiang CM. The general transcription
machinery and general cofactors. Crit Rev Biochem Mol Biol
2006;41:105–78.
Muller F, Tora L. The multicoloured world of promoter
recognition complexes. EMBO J 2004;23:2–8.
Goodrich JA, Tjian R. Unexpected roles for core promoter
recognition factors in cell-type-specific transcription and
gene regulation. Nat Rev Genet 2010;11:549–58.
Espinosa JM, Verdun RE, Emerson BM. p53 functions
through stress- and promoter-specific recruitment of
transcription initiation components before and after DNA
damage. Mol Cell 2003;12:1015–27.
Gehrig J, Reischl M, Kalmar E, et al. Automated highthroughput mapping of promoter-enhancer interactions in
zebrafish embryos. Nat Methods 2009;6:911–6.
Maston GA, Evans SK, Green MR. Transcriptional regulatory elements in the human genome. Annu Rev Genomics
Hum Genet 2006;7:29–59.
Raab JR, Kamakaka RT. Insulators and promoters: closer
than we think. Nat Rev Genet 2010;11:439–46.
Kolovos P, Knoch TA, Grosveld FG, et al. Enhancers and
silencers: an integrated and simple model for their function.
Epigenetics Chromatin 2012;5:1.
Ferg M, Sanges R, Gehrig J, et al. The TATA-binding
protein regulates maternal mRNA degradation and differential zygotic transcription in zebrafish. EMBO J 2007;26:
3945–56.
Guo S, Yamaguchi Y, Schilbach S, et al. A regulator of
transcriptional elongation controls vertebrate neuronal
development. Nature 2000;408:366–9.
Morley RH, Lachani K, Keefe D, et al. A gene regulatory
network directed by zebrafish No tail accounts for its roles
in mesoderm formation. Proc Natl Acad Sci USA 2009;106:
3829–34.
Yang L, Rastegar S, Strahle U. Regulatory interactions
specifying Kolmer-Agduhr interneurons. Development
2010;137:2713–22.
Aday AW, Zhu LJ, Lakshmanan A, et al. Identification of cis
regulatory features in the embryonic zebrafish genome
through large-scale profiling of H3K4me1 and H3K4me3
binding sites. Dev Biol 2011;357:450–62.
Ariza-Cosano A, Visel A, Pennacchio LA, et al. Differences
in enhancer activity in mouse and zebrafish reporter assays
are often associated with changes in gene expression. BMC
Genomics 2012;13:713.
Chatterjee S, Bourque G, Lufkin T. Conserved and nonconserved enhancers direct tissue specific transcription in
ancient germ layer specific developmental control genes.
BMC Dev Biol 2011;11:63.
Yip KY, Cheng C, Bhardwaj N, et al. Classification of
human genomic regions based on experimentally
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
55.
56.
57.
58.
59.
60.
61.
141
determined binding sites of more than 100 transcriptionrelated factors. Genome Biol 2012;13:R48.
Tamplin OJ, Cox BJ, Rossant J. Integrated microarray and
ChIP analysis identifies multiple Foxa2 dependent target
genes in the notochord. Dev Biol 2011;360:415–25.
Ritter DI, Dong Z, Guo S, et al. Transcriptional enhancers
in protein-coding exons of vertebrate developmental genes.
PLoS One 2012;7:e35202.
Smith RP, Riesenfeld SJ, Holloway AK, et al. A compact,
in vivo screen of all 6-mers reveals drivers of tissue-specific
expression and guides synthetic regulatory element design.
Genome Biol 2013;14:R72.
Ellingsen S, Laplante MA, Konig M, et al. Large-scale
enhancer detection in the zebrafish genome. Development
2005;132:3799–811.
Levesque MP, Krauss J, Koehler C, et al. New tools for
the identification of developmentally regulated enhancer
regions in embryonic and adult zebrafish. Zebrafish 2013;
10:21–9.
Aparicio S, Chapman J, Stupka E, et al. Whole-genome
shotgun assembly and analysis of the genome of Fugu
rubripes. Science 2002;297:1301–10.
Lander ES, Linton LM, Birren B, et al. Initial sequencing and analysis of the human genome. Nature 2001;409:
860–921.
Venter JC, Adams MD, Myers EW, et al. The sequence of
the human genome. Science 2001;291:1304–51.
Pennacchio LA, Rubin EM. Genomic strategies to identify
mammalian regulatory sequences. Nat Rev Genet 2001;2:
100–9.
Plessy C, Dickmeis T, Chalmel F, et al. Enhancer
sequence conservation between vertebrates is favoured
in developmental regulator genes. Trends Genet 2005;21:
207–10.
Sandelin A, Bailey P, Bruce S, et al. Arrays of ultraconserved non-coding regions span the loci of key developmental genes in vertebrate genomes. BMC Genomics
2004;5:99.
Woolfe A, Goodson M, Goode DK, et al. Highly conserved
non-coding sequences are associated with vertebrate development. PLoS Biol 2005;3:e7.
Bray N, Dubchak I, Pachter L. AVID: A global alignment
program. Genome Res 2003;13:97–102.
Brudno M, Do CB, Cooper GM, et al. LAGAN and MultiLAGAN: efficient tools for large-scale multiple alignment of
genomic DNA. Genome Res 2003;13:721–31.
Dubchak I, Poliakov A, Kislyuk A, et al. Multiple wholegenome alignments without a reference organism. Genome
Res 2009;19:682–9.
Schwartz S, Zhang Z, Frazer KA, et al. PipMaker–a web
server for aligning two genomic DNA sequences. Genome
Res 2000;10:577–86.
Sanges R, Kalmar E, Claudiani P, et al. Shuffling of cisregulatory elements is a pervasive feature of the vertebrate
lineage. Genome Biol 2006;7:R56.
Kuntz SG, Schwarz EM, DeModena JA, et al. Multigenome
DNA sequence conservation identifies Hox cis-regulatory
elements. Genome Res 2008;18:1955–68.
Li Q, Ritter D, Yang N, et al. A systematic approach to
identify functional motifs within vertebrate developmental
enhancers. Dev Biol 2010;337:484–95.
142
Ferg et al.
62. Pennacchio LA, Ahituv N, Moses AM, et al. In vivo enhancer analysis of human conserved non-coding sequences.
Nature 2006;444:499–502.
63. Sanges R, Hadzhiev Y, Gueroult-Bellone M, et al. Highly
conserved elements discovered in vertebrates are present
in non-syntenic loci of tunicates, act as enhancers and
can be transcribed during development. Nucleic Acids Res
2013;41:3600–18.
64. Gross DS, Garrard WT. Nuclease hypersensitive sites in
chromatin. Annu Rev Biochem 1988;57:159–97.
65. Crawford GE, Davis S, Scacheri PC, et al. DNase-chip: a
high-resolution method to identify DNase I hypersensitive
sites using tiled microarrays. Nat Methods 2006;3:503–9.
66. Song L, Crawford GE. DNase-seq: a high-resolution
technique for mapping active gene regulatory elements
across the genome from mammalian cells. Cold Spring Harb
Protoc 2010;2010:pdb prot5384.
67. Barski A, Cuddapah S, Cui K, et al. High-resolution profiling of histone methylations in the human genome. Cell
2007;129:823–37.
68. Bernstein BE, Kamal M, Lindblad-Toh K, et al. Genomic
maps and comparative analysis of histone modifications in
human and mouse. Cell 2005;120:169–81.
69. Heintzman ND, Stuart RK, Hon G, et al. Distinct and
predictive chromatin signatures of transcriptional promoters
and enhancers in the human genome. Nat Genet 2007;39:
311–8.
70. Creyghton MP, Cheng AW, Welstead GG, et al. Histone
H3K27ac separates active from poised enhancers and
predicts developmental state. Proc Natl Acad Sci USA 2010;
107:21931–6.
71. Rada-Iglesias A, Bajpai R, Swigut T, et al. A unique
chromatin signature uncovers early developmental enhancers in humans. Nature 2011;470:279–83.
72. Zentner GE, Tesar PJ, Scacheri PC. Epigenetic signatures
distinguish multiple classes of enhancers with distinct
cellular functions. Genome Res 2011;21:1273–83.
73. Blow MJ, McCulley DJ, Li Z, et al. ChIP-Seq identification
of weakly conserved heart enhancers. Nat Genet 2010;42:
806–10.
74. Visel A, Blow MJ, Li Z, et al. ChIP-seq accurately predicts
tissue-specific activity of enhancers. Nature 2009;457:854–8.
75. Santos-Rosa H, Schneider R, Bannister AJ, et al. Active
genes are tri-methylated at K4 of histone H3. Nature
2002;419:407–11.
76. Wardle FC, Odom DT, Bell GW, et al. Zebrafish promoter
microarrays identify actively transcribed embryonic genes.
Genome Biol 2006;7:R71.
77. Bernstein BE, Mikkelsen TS, Xie X, et al. A bivalent
chromatin structure marks key developmental genes in
embryonic stem cells. Cell 2006;125:315–26.
78. Pan G, Tian S, Nie J, et al. Whole-genome analysis of
histone H3 lysine 4 and lysine 27 methylation in human
embryonic stem cells. Cell Stem Cell 2007;1:299–312.
79. Kane DA, Kimmel CB. The zebrafish midblastula transition.
Development 1993;119:447–56.
80. Vastenhouw NL, Zhang Y, Woods IG, et al. Chromatin
signature of embryonic pluripotency is established during
genome activation. Nature 2010;464:922–6.
81. Bogdanovic O, Fernandez-Minan A, Tena JJ, et al.
Dynamics of enhancer chromatin signatures mark the
82.
83.
84.
85.
86.
87.
88.
89.
90.
91.
92.
93.
94.
95.
96.
97.
98.
99.
100.
101.
transition from pluripotency to cell specification during
embryogenesis. Genome Res 2012;22:2043–53.
Blader P, Lam CS, Rastegar S, et al. Conserved and
acquired features of neurogenin1 regulation. Development
2004;131:5627–37.
Rastegar S, Hess I, Dickmeis T, et al. The words of the
regulatory code are arranged in a variable manner in highly
conserved enhancers. Dev Biol 2008;318:366–77.
Kim TK, Hemberg M, Gray JM, et al. Widespread
transcription at neuronal activity-regulated enhancers.
Nature 2010;465:182–7.
Li W, Notani D, Ma Q, et al. Functional roles of enhancer
RNAs for oestrogen-dependent transcriptional activation.
Nature 2013;498:516–20.
Lai F, Orom UA, Cesaroni M, et al. Activating RNAs
associate with Mediator to enhance chromatin architecture
and transcription. Nature 2013;494:497–501.
Wang KC, Yang YW, Liu B, etal. A long noncoding RNA
maintains active chromatin to coordinate homeotic gene
expression. Nature 2011;472:120–4.
Howe K, Clark MD, Torroja CF, et al. The zebrafish
reference genome sequence and its relationship to the
human genome. Nature 2013;496:498–503.
Hunter S, Apweiler R, Attwood TK, et al. InterPro: the
integrative protein signature database. Nucleic Acids Res
2009;37:D211–5.
Domazet-Loso T, Tautz D. A phylogenetically based
transcriptome age index mirrors ontogenetic divergence
patterns. Nature 2010;468:815–8.
Gordan R, Murphy KF, McCord RP, et al. Curated
collection of yeast transcription factor DNA binding
specificity data reveals novel structural and gene regulatory
insights. Genome Biol 2011;12:R125.
Taylor JS, Braasch I, Frickey T, et al. Genome duplication,
a trait shared by 22000 species of ray-finned fish. Genome
Res 2003;13:382–90.
Naef F, Huelsken J. Cell-type-specific transcriptomics in
chimeric models using transcriptome-based masks. Nucleic
Acids Res 2005;33:e111.
Ravasi T, Suzuki H, Cannistraci CV, et al. An atlas of
combinatorial transcriptional regulation in mouse and
man. Cell 2010;140:744–52.
Zhang W, Morris QD, Chang R, et al. The functional
landscape of mouse gene expression. J Biol 2004;3:21.
Niehrs C, Pollet N. Synexpression groups in eukaryotes.
Nature 1999;402:483–7.
Karaulanov E, Knochel W, Niehrs C. Transcriptional
regulation of BMP4 synexpression in transgenic
Xenopus. EMBO J 2004;23:844–56.
Albagli O, Dhordain P, Deweindt C, et al. The BTB/POZ
domain: a new protein-protein interaction motif common
to DNA- and actin-binding proteins. Cell Growth Differ
1995;6:1193–8.
Kelly KF, Daniel JM. POZ for effect–POZ-ZF transcription factors in cancer and development. Trends Cell Biol
2006;16:578–87.
Stogios PJ, Downs GS, Jauhal JJ, et al. Sequence and
structural analysis of BTB domain proteins. Genome Biol
2005;6:R82.
Takeuchi JK, Lou X, Alexander JM, et al. Chromatin
remodelling complex dosage modulates transcription
Gene regulatory networks and control of gene expression
102.
103.
104.
105.
106.
107.
108.
109.
110.
111.
112.
113.
114.
115.
factor function in heart development. Nat Commun 2011;2:
187.
Link BA, Fadool JM, Malicki J, et al. The zebrafish young
mutation acts non-cell-autonomously to uncouple differentiation from specification for all retinal cells. Development
2000;127:2177–88.
Gregg RG, Willer GB, Fadool JM, et al. Positional cloning
of the young mutation identifies an essential role for the
Brahma chromatin remodeling complex in mediating
retinal cell differentiation. Proc Natl Acad Sci USA 2003;
100:6535–40.
Le Dreau G, Marti E. Dorsal-ventral patterning of the
neural tube: a tale of three signals. Dev Neurobiol 2012;72:
1471–81.
Dessaud E, McMahon AP, Briscoe J. Pattern formation in
the vertebrate neural tube: a sonic hedgehog morphogenregulated transcriptional network. Development 2008;135:
2489–503.
Briscoe J, Ericson J. Specification of neuronal fates in the
ventral neural tube. Curr Opin Neurobiol 2001;11:43–9.
Jacob J, Briscoe J. Gli proteins and the control of spinalcord patterning. EMBO Rep 2003;4:761–5.
Hui CC, Angers S. Gli proteins in development and
disease. Annu Rev Cell Dev Biol 2011;27:513–37.
Lewis KE. How do genes regulate simple behaviours?
Understanding how different neurons in the vertebrate
spinal cord are genetically specified. PhilosTrans R Soc Lond
B Biol Sci 2006;361:45–66.
Koudijs MJ, den Broeder MJ, Groot E, et al. Genetic
analysis of the two zebrafish patched homologues identifies
novel roles for the hedgehog signaling pathway. BMC Dev
Biol 2008;8:15.
England S, Batista MF, Mich JK, et al. Roles of Hedgehog
pathway components and retinoic acid signalling in specifying zebrafish ventral spinal cord neurons. Development
2011;138:5121–34.
Barth KA, Wilson SW. Expression of zebrafish nk2.2 is
influenced by sonic hedgehog/vertebrate hedgehog-1
and demarcates a zone of neuronal differentiation in the
embryonic forebrain. Development 1995;121:1755–68.
Guner B, Karlstrom RO. Cloning of zebrafish nkx6.2 and
a comprehensive analysis of the conserved transcriptional
response to Hedgehog/Gli signaling in the zebrafish neural
tube. Gene Expr Patterns 2007;7:596–605.
Schafer M, Kinzel D, Neuner C, et al. Hedgehog and
retinoid signalling confines nkx2.2b expression to the
lateral floor plate of the zebrafish trunk. Mech Dev 2005;
122:43–56.
Xu J, Srinivas BP, Tay SY, et al. Genomewide expression
profiling in the zebrafish embryo identifies target genes
116.
117.
118.
119.
120.
121.
122.
123.
124.
125.
126.
127.
128.
129.
130.
143
regulated by Hedgehog signaling during vertebrate
development. Genetics 2006;174:735–52.
Oosterveen T, Kurdija S, Alekseenko Z, et al. Mechanistic
differences in the transcriptional interpretation of local
and long-range Shh morphogen signaling. Dev Cell 2012;
23:1006–19.
Peterson KA, Nishi Y, Ma W, et al. Neural-specific Sox2
input and differential Gli-binding affinity provide context
and positional information in Shh-directed neural patterning. Genes Dev 2012;26:2802–16.
Vokes SA, Ji H, McCuine S, et al. Genomic characterization of Gli-activator targets in sonic hedgehog-mediated
neural patterning. Development 2007;134:1977–89.
Batista MF, Jacobstein J, Lewis KE. Zebrafish V2 cells
develop into excitatory CiD and Notch signalling
dependent inhibitory VeLD interneurons. Dev Biol 2008;
322:263–75.
Kimura Y, Satou C, Higashijima S. V2a and V2b neurons
are generated by the final divisions of pair-producing
progenitors in the zebrafish spinal cord. Development
2008;135:3001–5.
Park HC, Shin J, Appel B. Spatial and temporal regulation
of ventral spinal cord precursor specification by Hedgehog
signaling. Development 2004;131:5959–69.
Wyart C, Del Bene F, Warp E, etal. Optogenetic dissection
of a behavioural module in the vertebrate spinal cord.
Nature 2009;461:407–10.
Schafer M, Kinzel D, Winkler C. Discontinuous organization and specification of the lateral floor plate in zebrafish.
Dev Biol 2007;301:117–29.
Shin J, Park HC, Topczewska JM, et al. Neural cell fate
analysis in zebrafish using olig2 BAC transgenics. Methods
Cell Sci 2003;25:7–14.
Huang P, Xiong F, Megason SG, et al. Attenuation
of Notch and Hedgehog signaling is required for fate specification in the spinal cord. PLoS Genet 2012;8:e1002762.
Li H, de Faria JP, Andrew P, et al. Phosphorylation
regulates OLIG2 cofactor choice and the motor neuronoligodendrocyte fate switch. Neuron 2011;69:918–29.
Sun Y, Meijer DH, Alberta JA, et al. Phosphorylation state
of Olig2 regulates proliferation of neural progenitors.
Neuron 2011;69:906–17.
Christian M, Cermak T, Doyle EL, et al. Targeting DNA
double-strand breaks with TAL effector nucleases. Genetics
2010;186:757–61.
Jinek M, Chylinski K, Fonfara I, et al. A programmable
dual-RNA-guided DNA endonuclease in adaptive bacterial immunity. Science 2012;337:816–21.
Crocker J, Stern DL. TALE-mediated modulation of transcriptional enhancers in vivo. Nat Methods 2013;10:762–7.