Biology Reference
In-Depth Information
Regulatory states are generated in a specific sequence as
determined in the genome. Transitions of regulatory state
are controlled between as well as within developmental
domains, by the use of signaling interactions encoded in
GRNs. These interactions ensure the coordinated activity of
cell fate specification GRNs across the organism. Tempo-
rally, regulatory state transistion occurs continuously.
However, and this is their main significance, regulatory
states are finite in space, and their function is to discrimi-
nate cellular domains according to the cell fates each
domain will give rise to. In space, regulatory states have
clear borders. These borders are immediately apparent in
any in situ hybridization experiment where gene expression
is spatially non-uniform. The regulatory state in cells
expressing given genes must be different from the regula-
tory state present in cells not expressing them.
Even though we can refer to the regulatory state as
a single entity, it is actually a composite product of diverse
regulatory instructions; some of these instructions depend
on signaling interactions with adjacent domains, others on
prior developmental inputs, others on temporal factors,
others on cell lineage-specific regulatory factors, etc. All of
these different upstream inputs together are required for the
regulation of target regulatory genes. Here the combina-
torial function of transcriptional control is the dominant
feature. A subtle but essential consequence follows. This is
that each cis-regulatory node controlling expression of the
genes that constitute the regulatory state will be encoded
differently, so that it will respond to different inputs. This
can be clearly seen if we compare the control landscape
underlying regulatory state to that underlying the differ-
entiation gene battery, where all the genes respond to the
same small set of regulators. The non-uniform use of
regulatory information in the control of regulatory genes
has its particular significance. An artist uses a palette of
many colors which are kept separate to produce an infinite
variety of different paintings, an impossibility were they all
mixed together before application to the canvas.
The formation of regulatory states in embryonic devel-
opment is directed and irreversible, and therefore also
hierarchical. One intrinsic reason for the directedness of
GRNs is the biochemical quality of regulatory interactions.
The linkage between two nodes of a GRN represents the
occupancy of a cis-regulatory module by a transcription
factor, and results in a transcriptional control function which
is unidirectional. The irreversibility is in addition ensured
by specific GRN architectures. For example, signaling
interactions are usually transient. Within a given time
window signaling is required for the expression of regula-
tory genes, and perturbation of the signal leads to a failure in
regulatory state propagation and cell fate acquisition. Later,
the presence of the signaling ligand is no longer required for
expression of these genes, and they are instead controlled by
cross-regulation within the GRN; the new regulatory state is
now irreversibly established. In general, GRN circuit design
ensures the steady progression of regulatory states. The first
zygotic developmental GRN(s) are activated by localized
maternal inputs of regulatory significance. Thereafter, as
embryonic domains are progressively subdivided, regula-
tory states propagate hierarchically. The end result is to
ensure that the genes at the downstream periphery of GRNs,
which encode the molecules that actually perform the
functional and structural cellular processes, are expressed in
the correct spatial location of the body plan.
SYSTEM-WIDE AND DEEP INFORMATION
FLOW IN DEVELOPMENT
Novel developmentally functional assemblages, each con-
sisting of many causally interrelated components, emerge at
different levels. At the level of the terminal differentiation
process, for instance, the terminal regulatory state causes
more or less coordinate expression of multiple genes
encoding functionally related proteins, and these interact in
their own cellular domains to fulfill the requirement for
specialized functionalities: think of genes encoding enzymes
constituting pigment synthesis pathways
[16]
, or the battery
of genes producing eye lens crystallins
[17,18]
, or the many
other examples of specialized, dedicated, cell type-specific
sets of effector gene batteries for which we now have some
knowledge of transcriptional control
[19]
. In the strange-
looking diagram in
Figure 11.3
we see a cartoon that stresses
the different levels of informational transaction in the hier-
archical process of development. At the bottom is the
genomic DNA, the unchanging source code, from which
GRNs generate regulatory outputs in various parts of the
embryonic organism. The GRNs are indicated in the gray
platform representing the system of spatial specification of
regional regulatory states. Progressively, these spatial
domains are subdivided in each body part, as indicated by the
red lines, and each new subdomain is defined by a new or
partially new regulatory state. Black time arrows indicate
developmental directionality. In the figure, vertical cylinders
arising from the various spatial domains indicate that the
local regulatory states (red vertical arrows) cause particular
diverse morphogenetic gene cassettes (orange) and differ-
entiation gene batteries (blue) to be activated. The impact of
Figure 11.3
is that even though cell types can be compared at
the effector gene level (horizontal planes), the causal
explanation of their difference and specificity is rooted in the
GRN control system (gray platform) and in the vertical flow
of regional regulatory information (red arrows). Of course,
few organisms operate in sequential lockstep as in this
abstraction; that is, some differentiation genes are expressed
in some parts before any morphogenetic genes are in other
parts, andmorphogenetic and differentiation gene expression
may often temporally overlap. But
Figure 11.3
is not
