Biology Reference
In-Depth Information
2.1. Circuits are recurrent patterns in large-scale networks
Large-scale networks can be deconstructed into circuits composed of smal-
ler groups of nodes. Analysis of prokaryotic TF-target gene networks led to
the discovery that certain types of circuits occur inside networks at frequen-
cies much higher than in randomized control networks (
Shen-Orr
et al
.,
2002
). Such overrepresented circuits are called network motifs, and they
constitute “building blocks” of larger networks that preserve their functions
independent of the network environment in which they are embedded
(
Alon, 2007
;
Milo
et al
., 2002
).
One class of network motif is a circuit called a feedforward loop (FFL).
FFLs have two paths or arms of regulation, one direct (short arm) and one
indirect (long arm). The upstream node in the loop regulates the down-
stream node directly and indirectly through an intermediate node (
Fig. 9.5
).
Eight different FFL configurations exist, and four variables determine which
combination of dynamic behaviors emerges in a particular FFL (
Box 9.1
).
Computational modeling and experimental studies of prokaryotic FFLs
show that such loops have specific information processing properties that
differ from direct circuits (
Goentoro
et al
., 2009
;
Kaplan
et al
., 2008
;
Mangan and Alon, 2003
;
Mangan
et al
., 2003
;
Shen-Orr
et al
., 2002
).
Such properties include acceleration or delay of a response, generation of
signal pulses, and the ability to buffer the downstream node against fluctua-
tions in the upstream node such that only persistent changes in the upstream
node are transduced through the loop. FFLs provide robustness against
stochastic fluctuations in the upstream node of the circuit.
Another type of circuit is the feedback loop (FBL). Positive and negative
FBLs are known to be of central importance in biological processes (
Fig. 9.5
).
Positive FBLs can amplify signals, create ultrasensitivity, and enable irrevers-
ible states of gene expression to occur (
Brandman and Meyer, 2008
;
Chang
et al
., 2010
;
Ferrell, 2002
). Positive FBLs can give rise to bistable switches,
that is, two alternative stable states without stable intermediates in between
them (
Ferrell, 2002
). Double-negative FBLs can also stabilize gene expres-
sion in one state, though simulations have shown that double-negative FBLs
are not sufficient to create bistable switches. Other features such as nonlinear
positive feedback or balanced link strength are needed for a double-negative
FBL to generate bistable behavior (
Ferrell, 2002
;
Graham
et al
., 2010
). Single-
negative FBLs are associated with homeostasis and desensitization (
Ferrell,
2002
). Thus, in different ways, positive and negative FBLs provide robustness
of a circuit against fluctuation or perturbation.
We review the role of circuits that contain or are regulated by miRNAs,
focusing on their possible roles in providing robustness. We refer the reader
to four reviews on the topic that relate these circuits to developmental
canalization (
Hornstein and Shomron, 2006
;
Wu
et al
., 2009
), noise
(
Herranz and Cohen, 2010
), and signal transduction (
Inui
et al
., 2010
).
