Biology Reference
In-Depth Information
expression. Third, reporter DNA constructs injected into
the egg are subsequent stably incorporated with high effi-
ciency, allowing accurate expression of these reporter
constructs. This allows testing of the tissue-specific activity
of putative CRMs. Thus, the connections between
sequence-specific TFs and their target sites in the CRMs of
their target genes have been dissected extensively in
different tissues throughout the developmental course of
S. purpuratus
[54]
. Comparisons of reporter activity
between wild-type and mutant versions of CRMs, in which
predicted TF-binding sites have been mutated, reveal
functional TF-binding sites; analogously, performing such
experiments in wild-type and mutant animals, in which the
amount of a particular TF is altered genetically, indicates
the contribution of the perturbed TF to reporter activity.
Recently, a medium-throughput approach was developed to
simultaneously inject and analyze multiple bar-coded
reporter constructs
[175]
. There is an extensive under-
standing of GRNs in the sea urchin primarily from an
abundance of data on the activities of CRMs and the effects
of perturbations of TFs during development (notably,
without genome-wide ChIP data on TF binding). As
a result, sea urchin is perhaps best characterized in terms of
the regulatory logic of the impact of TF input on target gene
output, as well as how such individual connections are
assembled into a GRN
[54,176]
.
Progress in understanding GRNs in mouse and human
has been far slower than in non-mammalian model organ-
isms, with genome-scale progress in only a handful of
systems, such as embryonic stem cells
[177]
and hemato-
poiesis
[178,179]
. Human and mouse have much larger
genomes than many other model organisms. For instance,
whereas the C. elegans genome is 100 Mb
[165]
, the human
is 30 times larger at 3 Gb
[172]
, even though it harbors
roughly the same number of protein-coding genes. One
significant challenge is that CREs can be located far
(hundreds of kilobases or even further away) from the
genes they regulate in large genomes. In addition, many
genes, including those encoding TFs, have undergone
substantial gene duplication events, resulting in large
families of at least partially functionally redundant genes,
making it more difficult to use genetic approaches to
unravel their function(s). There are also numerous practical
limitations for experimental work in mouse, in particular
the longer generation time than with simpler model
organisms (see
Box 4.3
). Progress in the understanding of
human GRNs is limited to experiments that can be per-
formed using samples obtained from donors or on cell lines
grown in cell culture, which are outside the organismal
context and may not accurately reflect true physiology.
Identification of CREs in these genomes has come largely
from gene expression profiling studies combined with
computational sequence motif analysis
[64]
, cross-species
sequence
including many as part of the human ENCODE project
[77]
, and enhancer assays using reporter constructs tested
in mice
[70]
or in human cell culture
[96]
. Using Y1H
resources that were recently developed for the human
genome, it will be possible to identify the TFs that can
physically interact with identified CRMs
[26]
.
The thale cress Arabidopsis thaliana has served as the
primary model organism for plants (see Chapter 20). Aside
from their importance in agriculture, plants have many of
the same levels of complexity in gene regulation as
animals, such as expansion of TF families and tissue-
specific gene expression. Elegant studies on different
tissues dissected from A. thaliana roots examined by gene
expression profiling have uncovered gene expression
patterns important in plant root development
[180]
. Such
gene expression data, combined with other types of
genomic, proteomic, and metabolite data, will be important
for deciphering the GRNs in A. thaliana and other plants.
Thus far, however, relatively few CREs and TF-binding
sites have been mapped throughout the A. thaliana genome
[181]
. The mapping of GRNs has recently been facilitated
by the development of a set of Y1H resources and their
application to genes involved in the root
[24,182]
.
A number of other model organisms has proven to be
powerful in various ways for other types of studies such as
those of patterning in vertebrate embryonic development,
the generation of cellular extracts for in vitro investigations
of cell biological and biochemical processes, or forward
genetic screens
[183
e
185]
. These include: chicken (Gallus
gallus), representing birds; African clawed frog (Xenopus
laevis), representing amphibians; and zebrafish (Danio
reiro), representing fish. However, we do not list these
organisms in
Box 4.3
because they have not yet been used
extensively for large-scale studies of GRNs. We anticipate
that the continued development of genomic resources in
these and other model organisms
[183,184,186]
will
continue to advance their use in such studies, and our
understanding of how their gene regulatory and network
properties may differ from those in the other major model
organisms that thus far are better characterized.
FUTURE CHALLENGES
Although much effort has gone into using ChIP technolo-
gies to identify genomic sites occupied by TFs in vivo,
a major question that persists is: which of the experimen-
tally measured TF-binding events are direct and correspond
to functional CREs that regulate gene expression? Simi-
larly, with a few exceptions, the regulatory function of the
vast majority of interactions detected by Y1H assays
remains to be determined. While the functions of most
ChIP-bound sites for any particular sequence-specific TF
are as of yet unknown, most are not expected to correspond
to active CREs. Other types of experiment will need to be
conservation
analysis
[38]
, ChIP studies
