www.elsevier.com/locate/ygeno
Genomics 85 (20
In vivo characterization of a vertebrate ultraconserved enhanceri
Francis Poulin1, Marcelo A. Nobrega, Ingrid Plajzer-Frick, Amy Holt, Veena Afzal,
Edward M. Rubin, Len A. Pennacchio*
DOE Joint Genome Institute, Walnut Creek, CA 94598, USA
Genomics Division, Lawrence Berkeley National Laboratory, Berkeley, CA 94720, USA
Received 19 January 2005; accepted 7 March 2005
Abstract
Genomic sequence comparisons among human, mouse, and pufferfish (Takifugu rubripes (Fugu)) have revealed a set of extremely
conserved noncoding sequences. While this high degree of sequence conservation suggests severe evolutionary constraint and predicts a lack
of tolerance to change to retain in vivo functionality, such elements have been minimally explored experimentally. In this study, we describe
the in-depth characterization of an ancient conserved enhancer, Dc2, located near the dachshund gene, which displays a human-Fugu identity
of 84% over 424 basepairs (bp). In addition to this large overall conservation, we find that Dc2 is characterized by the presence of a large
block of sequence (144 bp) that is completely identical among human, mouse, chicken, zebrafish, and Fugu. Through the testing of reporter
vector constructs in transgenic mice, we observed that the 424-bp Dc2-conserved element is necessary and sufficient for brain tissue enhancer
activity. In vivo analyses also revealed that the 144-bp 100% conserved sequence is necessary, but not sufficient, to replicate Dc2 enhancer
function. However, the introduction of two separate 16-bp insertions into the highly conserved enhancer core did not cause any detectable
modification of its in vivo activity. Our observations indicate that the 144-bp 100% conserved element is tolerant of change at least at the
resolution of this transgenic mouse assay and suggest that purifying selection on the Dc2 sequence might not be as strong as we predicted or
that some unknown property also constrains this highly conserved enhancer sequence.
D 2005 Elsevier Inc. All rights reserved.
Keywords: Comparative genomics; Fugu; Pufferfish; Gene regulation; Enhancer; Transgenic mice
Introduction
Vertebrate comparative genomics has proven to be an
effective approach for uncovering functional elements in the
human genome. This has been accomplished through
comparison of the human genome with the genomes of a
wide range of species from primates to fish [1–4]. The
0888-7543/$ - see front matter D 2005 Elsevier Inc. All rights reserved.
doi:10.1016/j.ygeno.2005.03.003
i Sequence data from this article have been deposited with the EMBL/
GenBank Data Libraries under Accession No.
* Corresponding author. Genomics Division, One Cyclotron Road, MS
84-171, Lawrence Berkeley National Laboratory, Berkeley, CA 94720,
USA. Fax: +1 510 486 4229.
E-mail address: [emailprotected] (L.A. Pennacchio).1 Present address: Department of Integrative Biology, University of
California, Berkeley, Berkeley, CA 94720-3140.
underlying success of this strategy is based on comparing
sufficiently divergent genomes to distinguish neutral versus
functionally constrained sequence elements [5,6]. At one
extreme, comparison of the human genome to the genome of a
teleost fish, Takifugu rubripes (Fugu), has identified a set of
anciently conserved coding and noncoding sequences [7,8].
Humans and Fugu last shared a common ancestor
approximately 400 million years ago [9] and while the Fugu
genome is one of the smallest known in vertebrates (365Mb),
its gene repertoire is similar to that of humans [10]. These
characteristics led to the original proposal to sequence the
Fugu genome to assist with the annotation of human genes
based on comparative genomics [10]. Furthermore, the
availability of the Fugu genomic sequence also revealed that
human-Fugu-conserved noncoding sequences can be used to
delineate gene regulatory sequences [1,11–15].
05) 774 – 781
F. Poulin et al. / Genomics 85 (2005) 774–781 775
Many of the ancient elements conserved between human
and Fugu overlap with recently described ultraconserved
human/rodent DNA elements [16]. Through the use of
human/mouse/rat whole-genome comparisons, ultracon-
served elements were defined as sequences that are at least
200 bp and are absolutely conserved in these three species
[16]. Intriguingly, the nonexonic ultra- and human-Fugu-
conserved elements do not show a random distribution in
the human genome, but tend to cluster near genes encoding
transcription factors that are involved in developmental
processes [16,17].
One example of clustered human-Fugu-conserved non-
coding elements are those surrounding DACH1, a hom*olog
of the Drosophila dachshund gene [18]. In both vertebrates
and invertebrates dachshund displays a complex temporal
and spatial pattern of expression, and the gene product is
critical in the development of the central nervous system,
sensory organs, and limbs [19–23]. In vivo analysis of nine
DNA elements conserved from human to Fugu demon-
strated that seven displayed enhancer activities recapitulat-
ing some aspects of DACH1 expression [15]. The most
surprising characteristic of these enhancers is their
extremely high degree of conservation between humans
and fish. For instance Dc2, a conserved regulatory element
of human DACH1 that is known to drive expression in
hindbrain, forebrain, and retina, is 84% identical over 424
bp to its Fugu ortholog. As impressive is the high degree of
conservation within subregions of this element, with
portions displaying greater than 99% identity over 270 bp
among human, mouse, and rat and greater than 98% identity
over 219 bp between human and Fugu. This is especially
striking considering the known degeneracy of the sequences
recognized by transcription factors [24], and this observa-
tion of extreme conservation suggests that other limitations
may prevent these enhancers from changing over evolu-
tionary time. To better characterize the nature of the
sequence constraints in highly conserved vertebrate
enhancers, we manipulated a single known element, the
human Dc2 enhancer, in vivo through a series of reporter
constructs tested in transgenic mouse assays.
Results
In this study, we characterized a single human-Fugu
enhancer of the DACH1 locus (Dc2) through (1) compara-
tive genomic analyses to define ‘‘phylogenetic footprints,’’
(2) deletion constructs in transgenic mice to experimentally
define the minimal sequence necessary and sufficient for in
vivo activity, and (3) targeted mutagenesis to assess whether
the enhancer is tolerant of insertional disruption events.
Comparative genomics delineates conserved modules
The Dc2 enhancer was previously identified by compar-
ing the sequences of the orthologous human and Fugu
DACH1 genes [15]. Briefly, a human DNA fragment of
2086 bp encompassing the sequence conserved in Fugu was
tested in transgenic mice and demonstrated to enhance the
activity of a minimal heat shock promoter (HSP68) fused to
b-galactosidase (LacZ) [25]. The Dc2 enhancer was found
to drive the expression of LacZ in the hindbrain, forebrain,
and retina of the developing mouse embryo [15]. To better
define the sequence elements required for the enhancer
function of Dc2, we performed comparative analysis of the
functional element in multiple vertebrate species (Fig. 1A).
Comparison of the human, mouse, and rat sequences
revealed several discrete elements of conservation through-
out the enhancer when using an 80 bp and 70% identity
cutoff, slightly less stringent than the classically defined 100
bp and 70% identity previously used in the identification of
functional mammalian regulatory elements [3] (Fig. 1A; 4
human-mouse elements). Comparison of the human and
chicken sequences shows a decrease in the number of
conserved elements, but extensive conservation is still
readily observed (Fig. 1A; two human-chicken elements).
This is in contrast to the alignment of the human sequence
with that of the more distant vertebrates, which highlights a
single region of 424 bp that is more than 84% conserved in
all the species analyzed (Fig. 1A, human Dc2 to frog,
zebrafish or Fugu). This region is 96% identical to the
mouse Dc2, and displays 84% identity between the human
and the Fugu Dc2 (Fig. 1A).
Closer inspection of the alignment of this 424-bp region
reveals that the largest uninterrupted block of perfect
identity between human and mouse is 195 bp and that
between human and Fugu is 144 bp in length (Fig. 2). This
block of perfect identity can be expanded to a remarkable
270 bp with a single mismatch between human and mouse
Dc2 (Fig. 2, human nucleotides 75 to 345). Similarly, the
human-Fugu conservation is 98% for a block of 219 bp
from human nucleotides 90 to 308 (Fig. 2). This large block
is flanked by shorter regions where the similarity between
species is lower (Fig. 2), suggesting that the enhancer is
composed of a core sequence that is highly constrained, and
satellite sequences that can evolve more rapidly. Analysis of
the 424-bp conserved region with rVista 2.0 [26] predicts
111 and 114 putative transcription factor-binding sites in the
human and Fugu sequence, respectively. Of these sites, 72
are conserved in both species (Fig. 3), with 44 sites landing
in the 144 bp/100% conserved core (61%; Fig. 3, green
box).
Comparison of Dc2 to the human genome also uncov-
ered a single additional region of sequence similarity (which
we will refer to as Dc2V) on chromosome X. Upon detailed
further examination of the flanking sequence, we found that
Dc2V lies within DACH2, a known paralog of DACH1 that
arose from an ancient genomic duplication event predating
the divergence of the mammalian and fish lineages (Fig. 4).
Similar to Dc2, Dc2V is located in the first intron of DACH2,with both displaying similar enhancer properties in trans-
genic mice, further supporting their common origin (data
Fig. 1. Sequence comparison of Dc2 enhancer in multiple species. mVista alignment (http://gsd.lbl.gov/vista/) between the human DACH1 Dc2 enhancer and
(A) orthologous Dc2 sequences from the DACH1 gene, or (B) paralogous Dc2V sequences from the DACH2 gene of the indicated species. Alignments were
performed with an 80-bp window size and a 70% identity threshold.
F. Poulin et al. / Genomics 85 (2005) 774–781776
not shown). Alignment of the human Dc2 to the paralogous
Dc2V sequence from multiple vertebrates shows a more
limited region of similarity between the two elements (Fig.
1B). The human Dc2V contains the most conserved portion
of the DACH1 Dc2 enhancer (Fig. 2; 144 bp, human-Fugu
Dc2 100% identity), but overall has a lower degree of
sequence conservation (Fig. 4; 144 bp, human Dc2-Dc2V86% identity).
Refinement of the minimal necessary Dc2 enhancer
Using the results from our comparative analysis as a
guide, we designed several Dc2 constructs to test for
enhancer function in transgenic mouse embryos (Fig. 5).
As previously demonstrated [15], we found that the wild-
type construct displayed enhancer activity in all transgenic
lines tested (Fig. 5B, WT), with specific staining in the
forebrain, hindbrain, and retina (Fig. 5B, WT). To define the
minimum fragment necessary for this activity, we tested a
construct consisting of only the core 424 bp of sequence that
are conserved in the DACH1 Dc2 orthologues from human
to Fugu (Fig. 5B, 424 bp/84%). We found this 424-bp
fragment consistently drove expression in transgenic
embryos in a pattern indistinguishable from the parent
2086-bp construct. Conversely, we found that deletion of
this 424-bp sequence from the parent (WT) construct
completely abolished its activity in our transgenic assay
(Fig. 5B, D424 bp). This finding suggests the small 424-bp
element is able to carry out all the observed enhancer
activities found in the full construct.
To further explore the minimal sequence necessary for
enhancer activity, we reduced the size of the conserved 424-
bp fragment to the 144-bp Dc2 region 100% conserved
between human and Fugu (Fig. 2, dashed line), a segment
that also shows similarity to the paralogous human Dc2Venhancer (Fig. 4). We found that while this 144-bp sequence
does not enhance transcription in transgenic mice (Fig. 5B,
144 bp/100%), its removal from the WT parental Dc2
construct completely abolished Dc2 enhancer activity (Fig.
5B, D144 bp). Taken together, these results demonstrate
that, although human-mouse comparison shows several
regions of sequence conservation (Fig. 1A), the necessary
and sufficient portion of the element is contained within the
region shared between human and Fugu. Moreover, the
region of the element that is shared between Dc2 and Dc2V isnecessary, but not sufficient, for the function of the
enhancer.
The highly conserved minimal enhancer core is tolerant of
insertion mutations
Having defined the 144-bp core element as essential for
enhancer function and 100% conserved between human and
Fugu, we sought to determine its tolerance to sequence
change. The large size of the minimal Dc2 enhancer and its
extreme conservation raised several questions regarding the
types of constraints acting on its sequence. To address
whether the internal organization of the enhancer could limit
its variation, we introduced DNA linkers into the phyloge-
netically conserved core of the enhancer (Fig. 2, arrowheads).
Fig. 2. Sequence alignment of the evolutionarily conserved region from the DACH1 Dc2 enhancer. Dashed lines above the sequence indicate the 144-bp region
that is 100% conserved between human and Fugu Dc2. Arrowheads indicate the position of insertions in the mutated version of the enhancer. Alignment was
preformed using ClustalW.
F. Poulin et al. / Genomics 85 (2005) 774–781 777
These linkers were 16 bp long, representing an appro-
ximately one and a half DNA helix turn, and were designed
to not introduce new predicted transcription factor binding
sites into the enhancer (see Materials and methods). We
chose the location of the insertions to disrupt the most
constrained region of the enhancer (Fig. 2, arrowheads), with
one linker (insert 1) located in the segment that is absolutely
conserved and has hom*ology to DACH2 Dc2V (Fig. 2,
dashed line). Surprisingly, in studying the various transgenic
lines containing these reporter constructs, we found that the
two different linker insertions did not affect the WT reporter
construct activity in the mouse embryos (Fig. 5B, Insert 1
Fig. 3. Conserved transcription factor-binding sites in the DACH1 Dc2 enhancer. The 424-bp human and Fugu Dc2 enhancers were analyzed for the presence
of conserved vertebrates transcription factor-binding sites using rVista 2.0 (http://rvista.dcode.org), with a matrix similarity cutoff of 0.9. Alignment between
human and Fugu DACH1 Dc2 (424 bp) is depicted as blocks ranging from 50 to 100% conservation. Putative transcription factor-binding-site positions
(identified on the left) are indicated by colored boxes above the alignment. The position of the 144-bp/100% conserved Dc2 core is highlighted in green.
Arrowheads indicate the position of insertions in the mutated version of the enhancer.
F. Poulin et al. / Genomics 85 (2005) 774–781778
and Insert 2), and the pattern of expression of these mutant
constructs was not distinguishable from that of the WT
construct.
Discussion
Comparative genomics is continuing to help localize
conserved noncoding sequences with gene regulatory
activity, and distant evolutionary comparisons between
mammals and teleosts have proven especially efficient
[1,13,15,27]. In this study we characterized a gene
enhancer (Dc2) identified through human-Fugu compara-
tive genomics, and refined the sequences necessary and
sufficient for its function, as well as assessed the impact of
insertional mutations on its gene regulatory activity.
Fig. 4. Local alignment of the human DACH1 Dc2 and DACH2 Dc2V enhancers. Twith the paralogous sequence from DACH2 Dc2V. Numbering of DACH1 Dc2 nuc
The Dc2 enhancer is characterized by the presence of a
large block of sequence (144 bp) that is entirely identical
between human and multiple distantly related species
(mouse, rat, chicken, frog, zebrafish, and Fugu). This
observation is similar to recent reports for mammalian
ultraconserved sequences [16], despite the fact that the
Dc2 enhancer falls just short of the requirements of the
ultraconserved set (>200 bp, 100% identity). Nonetheless,
a similarity in human-Fugu and ultraconserved noncoding
sequences is their enormous degree of sequence conserva-
tion, extending hundreds of basepairs with minimal
substitutions over hundreds of million years of evolution.
Regardless of the precise definition of extreme sequence
conservation, the question is raised of why such striking
constraint. One possible explanation is their enrichment
near developmental genes, suggesting that they are
he most conserved portion of the human DACH1 Dc2 enhancer was aligned
leotides is the same as in Fig. 2. Alignment was performed using ClustalW.
Fig. 5. Functional analysis of the human DACH1 Dc2 enhancer. (A) mVista alignment between the Dc2 enhancer from the human and the mouse (H/M), or the
Fugu (H/F) DACH1 gene. Nucleotide positions are indicated for the human wild-type Dc2 sequence. (B) The indicated DNA fragments from human Dc2 were
assayed for in vivo enhancer activity on the minimal HSP68 promoter driving LacZ expression. The wild-type (WT) enhancer corresponds to the DNA
fragment tested by Nobrega et al. [15]. A 424-bp fragment (nt 318 to 741 of WT) that is 84% identical between human and Fugu was tested by itself (424 bp/
84%), or the corresponding sequence was deleted from the WT fragment (D424 bp). A 144-bp fragment (nt 404 to 547 of WT) that is 100% identical in human
and Fugu was tested by itself (144bp/100%), or the corresponding sequence was deleted from the WT fragment (D144 bp). The WT Dc2 enhancer was
modified to insert a 16-bp linker at two different locations in the human-Fugu-conserved region (Insertion 1 and Insertion 2). The number of mouse embryos
displaying the Dc2 expression pattern is reported over the total number of transgenic embryos (Pattern/Tg). ND: not detected.
F. Poulin et al. / Genomics 85 (2005) 774–781 779
involved in the tight regulatory control of genes directing
the basic vertebrate body plan. But why would gene
regulatory sequences require hundreds of basepairs of
sequence perfectly conserved over such long time periods
when most transcription factors are capable of recognizing
short (6–12 bp) degenerate sites? It may be due to a large
number of overlapping transcription factor-binding sites
with highly rigid binding requirements as well as severe
constraint on the spacing of putative modules within the
enhancer.
Another powerful aspect of vertebrate comparative
genomics and gene regulatory sequence characterization
is the occurrence of gene paralogs resulting from genomic
duplications over the course of evolution. This is readily
apparent with the human DACH1 gene and its Dc2
enhancer sharing similarity with the DACH2 paralog and
the adjacent Dc2V. This duplication event is very ancient,
predating the last common ancestor of human-Fugu, and
both loci have strongly resisted sequence change in these
functional sequences. Comparison between these paralogs
revealed a small and highly conserved core (144 bp) that is
necessary to preserve the enhancer function, but is not
sufficient to recapitulate it. This suggests that while it is
critical to conserve the sequence of certain modules for the
enhancer to function, some flanking sequences are more
flexible to change. It is known that stabilizing selection can
maintain an enhancer function in different species even
when sequence conservation is limited because of turnover
in transcription factor-binding sites [28]. However, the fact
that the flanking sequences are flexible in the Dc2 paralogs
is contrasted by their extreme conservation in each of the
individual orthologs. It therefore appears that the sequence
F. Poulin et al. / Genomics 85 (2005) 774–781780
of each of the paralogous elements changed rapidly after the
duplication event, and that each eventually became fixed
independently [16,29].
While enormous sequence conservation of human-Fugu
and ultraconserved elements alone predicts that any change
in such an element would destroy its activity, our analyses
indicate that the DACH1 Dc2 enhancer is much more
flexible than anticipated. Two separate insertions in the
highly conserved core of this enhancer did not cause any
detectable modification in its activity in vivo. This could be
due to the limitations of our assay, which occur at a single
time point and lack sensitivity to detect small quantitative
changes, or to the presence of another unidentified function
within this conserved module. However, the observation
that the complexity of the Dc2 expression pattern is not
affected by insertions indicates that many of its general
activities are still retained. While negative selection on the
Dc2 sequence could not be assessed in these in vivo
studies, its enormous conservation more likely suggests
that additional unknown biological properties also con-
strain Dc2.
Materials and methods
Cloning
Dc2 enhancer constructs were PCR-amplified from
human genomic DNA (BD Biosciences) and directionally
cloned into the pENTR/D-TOPO vector (Invitrogen). The
wild-type (WT), 424, and 144 Dc2 constructs were PCR-
amplified with the corresponding forward and reverse
primers described in Table 1. Constructs containing site-
specific mutations were generated by overlap extension
PCR [30], in which the mutagenic primers were used in
combination with WT primers of opposite orientation
(Table 1). All constructs were sequenced and transferred
to the Gateway-HSP68-LacZ vector using the LR recom-
bination reaction (Invitrogen). The elements were cloned in
the same orientation relative to the HSP68 promoter as
they are to the endogenous DACH1 promoter. A 16-bp
linker (5V- GCTGCCCGCGCAGTAC) was inserted at two
locations (nucleotides 128 and 236 of human DACH1 Dc2,
Fig. 2) in the wild-type human Dc2 enhancer to test for
Table 1
Primers used to generate Dc2 constructs
Construct Forward primer
WT 5V-GCAATTTTGAAAAAGAAAACAATGG424 5V-AATTCTTTGCCTGATTTTC144 5V-TCAGGGTGCCTTTGAGD424 5V-CTTATTATTAAAATATAGGCTGTCTTCCAGTCTTTGAATACD144 5V-GCCTAAAAAAATCTACTACACATTTCCCTTGGAGCTGCInsert 1 5V-GCTGAACGATGTGCATATTCATTAAGGCTCACATAInsert 2 5V-GCTGAACGATGTGCATCCCTTGGAGCTGCCTGC
disruption of Dc2 function. The linker was scanned with
the MatInspector software (Genomatix, Matrix Family
Library Version 4.2) to ensure the absence of putative
transcription-factor binding sites. Vertebrates matrices were
used with the following parameters: core similarity, 0.75;
matrix silmilarity, optimized.
Generation of transgenic mice
Plasmid DNA was purified using the EndoFree plasmid
maxi kit (Qiagen). One hundred micrograms of plasmid
was linearized with XhoI, followed by purification on
Micropure EZ columns and Montage PCR filter units
(Millipore). The purified DNA was dialyzed for 24 h
against injection buffer (10 mM Tris, pH 7.5; 0.1 mM
EDTA), and its concentration determined fluorometrically
and by agarose gel electrophoresis. The DNA was diluted
to a concentration of 1.5 to 2 ng/ll and used for
pronuclear injections of FVB embryos [31], in accordance
with protocols approved by the Lawrence Berkeley
National Laboratory.
Embryo staining
Embryos were harvested at 12.5 dpc and dissected in
cold 100 mM phosphate buffer pH 7.3, followed by 30 min
of incubation with 4% paraformaldehyde at 4-C. The
embryos’ heads were punctured with a 27 G needle to
facilitate the penetration of the staining solution and washed
three times for 30 min with wash buffer (2 mM MgCl2;
0.01% deoxycholate; 0.02% NP-40; 100 mM phosphate
buffer, pH 7.3). Embryos were stained for 24 h at room
temperature with freshly made staining solution (0.8 mg/ml
X-gal; 4 mM potassium ferrocyanide; 4 mM potassium
ferricyanide; 20 mM Tris, pH 7.5, in wash buffer). Stained
embryos were rinsed 3 times in 100 mM phosphate buffer,
pH 7.3, and postfixed in 4% paraformaldehyde. Yolk sacs
were carefully dissected from embryos and DNA was
prepared by boiling the tissue for 20 min in 75 Al of
solution 1 (25 mM NaOH; 0.2 mM EDTA), followed by
neutralization with 75 Al of solution 2 (40 mM Tris-HCl).
Yolk sac DNA was screened by PCR with LacZ primers
(LacZ forward, 5V-TTTCCATGTTGCCACTCGC; LacZ
reverse, 5V-AACGGCTTGCCGTTCAGCA).
Reverse primer
5V-TAGACAGCTCATGCTGAGAAAACTG5V-TTTTGGTGATGAAGACAG5V-GCTCCTTTCATACTTG5V-GTATTCAAAGACTGGAAGACAGCCTATATTTTAATAATAAG5V-GCAGCTCCAAGGGAAATGTGTAGTAGATTTTTTTAGGC5V-ATGCACATCGTTCAGCGCAGGATTAGATTTTAATTA5V-ATGCACATCGTTCAGCAAATGTGCTCCTTTCATAC
F. Poulin et al. / Genomics 85 (2005) 774–781 781
Acknowledgments
We thank N. Ahituv for the Gateway-HSP68-LacZ vector.
Research was conducted at the E.O. Lawrence Berkeley
National Laboratory, supported by Grant HL88728, Berke-
ley-PGA, under the Programs for Genomic Application,
funded by National Heart, Lung, and Blood Institute, USA,
and performed under Department of Energy Contract DE-
AC0378SF00098, University of California. F.P. was sup-
ported by a Fellowship from the Canadian Institutes of
Health Research.
References
[1] N. Ghanem, et al., Regulatory roles of conserved intergenic
domains in vertebrate Dlx bigene clusters, Genome Res. 13
(2003) 533–543.
[2] D. Boffelli, et al., Phylogenetic shadowing of primate sequences to
find functional regions of the human genome, Science 299 (2003)
1391–1394.
[3] L.A. Pennacchio, E.M. Rubin, Genomic strategies to identify
mammalian regulatory sequences, Nature Rev. Genet. 2 (2001)
100–109.
[4] J.W. Thomas, et al., Comparative analyses of multi-species sequences
from targeted genomic regions, Nature 424 (2003) 788–793.
[5] E.H. Margulies, M. Blanchette, D. Haussler, E.D. Green, Identification
and characterization of multi-species conserved sequences, Genome
Res. 13 (2003) 2507–2518.
[6] L. Elnitski, et al., Distinguishing regulatory DNA from neutral sites,
Genome Res. 13 (2003) 64–72.
[7] S. Aparicio, et al., Whole-genome shotgun assembly and analysis of
the genome of Fugu rubripes, Science 297 (2002) 1301–1310.
[8] S. Aparicio, et al., Detecting conserved regulatory elements with the
model genome of the Japanese puffer fish, Fugu rubripes, Proc. Natl.
Acad. Sci. USA 92 (1995) 1684–1688.
[9] B. Venkatesh, P. Gilligan, S. Brenner, Fugu: a compact vertebrate
reference genome, FEBS Lett. 476 (2000) 3–7.
[10] S. Brenner, et al., Characterization of the pufferfish (Fugu)
genome as a compact model vertebrate genome, Nature 366
(1993) 265–268.
[11] D.H. Rowitch, et al., Identification of an evolutionarily conserved
110 base-pair cis-acting regulatory sequence that governs Wnt-1
expression in the murine neural plate, Development 125 (1998)
2735–2746.
[12] H. Marshall, et al., A conserved retinoic acid response element
required for early expression of the homeobox gene Hoxb-1, Nature
370 (1994) 567–571.
[13] C. Kimura-Yoshida, et al., Characterization of the pufferfish Otx2 cis-
regulators reveals evolutionarily conserved genetic mechanisms for
vertebrate head specification, Development 131 (2004) 57–71.
[14] H. Popperl, et al., Segmental expression of Hoxb-1 is controlled by a
highly conserved autoregulatory loop dependent upon exd/pbx, Cell
81 (1995) 1031–1042.
[15] M.A. Nobrega, I. Ovcharenko, V. Afzal, E.M. Rubin, Scanning human
gene deserts for long-range enhancers, Science 302 (2003) 413.
[16] G. Bejerano, et al., Ultraconserved elements in the human genome,
Science 304 (2004) 1321–1325.
[17] D. Boffelli, M.A. Nobrega, E.M. Rubin, Comparative genomics at the
vertebrate extremes, Nature Rev. Genet. 5 (2004) 456–465.
[18] X. Li, et al., Eya protein phosphatase activity regulates Six1-Dach-Eya
transcriptional effects in mammalian organogenesis, Nature 426
(2003) 247–254.
[19] S.R. Martini, G. Roman, S. Meuser, G. Mardon, R.L. Davis, The
retinal determination gene, dachshund, is required for mushroom body
cell differentiation, Development 127 (2000) 2663–2672.
[20] G. Mardon, N.M. Solomon, G.M. Rubin, dachshund encodes a nuclear
protein required for normal eye and leg development in Drosophila,
Development 120 (1994) 3473–3486.
[21] M. Kurusu, et al., Genetic control of development of the mushroom
bodies, the associative learning centers in the Drosophila brain, by the
eyeless, twin of eyeless, and Dachshund genes, Proc. Natl. Acad. Sci.
USA 97 (2000) 2140–2144.
[22] R.J. Davis, W. Shen, T.A. Heanue, G. Mardon, Mouse Dach, a
hom*ologue of Drosophila dachshund, is expressed in the developing
retina, brain and limbs, Dev. Genes Evol. 209 (1999) 526–536.
[23] R.J. Davis, et al., Dach1 mutant mice bear no gross abnormalities in
eye, limb, and brain development and exhibit postnatal lethality, Mol.
Cell. Biol. 21 (2001) 1484–1490.
[24] G.D. Stormo, DNA binding sites: representation and discovery,
Bioinformatics 16 (2000) 16–23.
[25] R. Kothary, et al., Inducible expression of an hsp68-lacZ hybrid gene
in transgenic mice, Development 105 (1989) 707–714.
[26] G.G. Loots, I. Ovcharenko, rVISTA 2.0: evolutionary analysis of
transcription factor binding sites, Nucleic Acids Res. 32 (2004)
W217–W221.
[27] F. Spitz, F. Gonzalez, D. Duboule, A global control region defines a
chromosomal regulatory landscape containing the HoxD cluster, Cell
113 (2003) 405–417.
[28] M.Z. Ludwig, C. Bergman, N.H. Patel, M. Kreitman, Evidence for
stabilizing selection in a eukaryotic enhancer element, Nature 403
(2000) 564–567.
[29] S. Ohno, Evolution by Gene Duplication, Springer-Verlag, Berlin,
1970.
[30] J. Sambrook, D.W. Russell, Molecular Cloning: A Laboratory Manual,
Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, 2001.
[31] A. Nagy, M. Gertsenstein, K. Vintersten, R. Behringer, Manipulating
the Mouse Embryo: A Laboratory Manual, Cold Spring Harbor
Laboratory Press, Cold Spring Harbor, NY, 2002.
