Biology Reference
In-Depth Information
To achieve a spontaneous increase in membrane binding
of MinD in the presence of its antagonist, MinD binding
must be self-referencing. This means that the presence of
bound MinD favors further MinD binding, which can be
caused by, for example, dimerization of MinD in the bound
form. This introduces a first non-linearity in the reaction
network. However, this needs to be balanced by a similar
non-linearity on the antagonistic side, otherwise only high
levels of membrane-bound MinD will be a stable solution.
As postulated in the above paper, rapid rebinding of free
MinE to MinD at high MinE concentrations is likely. This
e
without increasing the number of observables at the same
time runs not only the risk of over-fitting, but creates a false
sense of security if the model can account for all previously
observed experimental aspects. A more detailed model will
create testable new features that need to be checked against
experiments. In this way experiments and modeling must
improve upon each other, and need to be interwoven more
completely. The search for interaction partners of proteins
(via screening or FRET approaches), their localization and
transient translocation, the physicochemical properties that
govern interaction and translocation timescales (reaction
and association rates, diffusion coefficients) and their
activity needs to continue. Equally essential are modeling
approaches which can build on the existing experimental
data to validate hypotheses and at the same time predict the
missing experiments to complement current paradigms.
partly by experimental knowledge, partly by exclusion
via trial and error in simulations
reconstituted system is
capable of oscillatory behavior in '0D' of a well-mixed
reaction vessel and standing or travelling waves in 1D
(
Figure 17.3
). When extended to 2D, these waves will self-
propagate comparably to the reconstituted system
described earlier
[36]
. Furthermore, a decrease in the MinE
feedback strength results in localized Turing patterns.
When restricting the 1D system to finite lengths, we realize
that below a critical length wave propagation is impossible,
whereas above a certain length multiple waves can form.
Only at a certain length does the time-averaged maximum
of MinE coincide with a minimum of MinD at the center of
the cell, which could provide the spatial cue on where to
divide, coincidentally with a cue on when the cell has
grown large enough. This network operating with nm-sized
molecules and reactions at microsecond timescales
decides the fate of micrometer-sized bacteria after hours
of growing. The above simulation hinges on the collision
of scales in two instances: diffusion is about an order of
magnitude slower in the membrane than in the cytosol, and
recruitment of MinE to MinD is an order of magnitude
slower than recruitment of MinD to the PM. The first lets
a local reaction deplete a cytosolic pool fast enough to stay
local, and the second is responsible for the delay of the
MinE-peak in the wave with respect to the progressing
MinD peak. These two assumptions need to be experi-
mentally verified to validate the simulation. However, when
considering the profile of the 1D travelling wave, a clear
discrepancy with the experiment is obvious. The simplified
model generates an almost symmetrical MinE profile of the
1D wave, whereas the experiment showed a linearly rising
frontal edge with an abruptly decreasing tail edge. Even
though the MinD profile is more similar, the MinE profile's
only reproduced feature is its delay in respect to the MinD
peak.
This points to the necessity for further experiments
coupled with more detailed simulations to narrow down the
reason for this discrepancy. For example, one could intro-
duce further states to detail the order of MinD
e
CONCLUSIONS
Biological systems manage to increase their own organi-
zation and maintain it at the expense of the order taken
from the environment. Energy is consumed and disorder
(entropy) is exported. It has been suggested that for
a sufficiently complex network of chemical reactions, self-
organization into autocatalytic loops will necessarily occur
[37]
. Such loops, as a form of closure, are a necessary
topology for any system that needs to maintain itself, and
hence are a prerequisite for life. Evolution shapes the path
by favoring those structures adapted to the environment. In
such self-organized systems, global coherence spontane-
ously emerges out of local interactions. This effect cannot
be understood without the effect of simultaneous upward
and downward causation of the involved scales. Closure is
again present wrapping up the global and local scale.
Systems biology tries to describe living organisms
beyond reductionism to create a coherent view of their
operational principles and mechanisms. Our description
starts at the scale which is most appropriate for the
phenomenon that we are studying, being able to reach out
to other levels when understanding increases. In the words
of Sydney Brenner, we describe the system neither bottom-
up nor top-down, but middle-out. Closure of biological
systems makes the starting point logically irrelevant:
'There is no privileged level in biology that dictates the
rest'
[38]
. And understanding how closure is achieved aids
our understanding of living matter.
In cellular systems biology we work towards under-
standing how nanometer-sized molecules generate cell-
spanning patterns from which biological function emerges.
As we have described in this chapter, this requires bridging
conceptually, theoretically and experimentally scales in
space and time. A causal description of the molecular
ecology that runs the cellular operation can aid in bridging
the scale gap. Deriving such a causality map requires a tight
MinE
complex dissociation with regard to the state of PM binding
and the ADP
e
ATP switch of MinD recently detached
from the PM. Increasing the complexity of the model
/
