Biology Reference
In-Depth Information
hysteretic switching, digital counting, signal amplification,
and noise-generated excitation.
regulator) by acting as a phosphatase that dephosphorylates
the response regulator, as in the case of the EnvZ
OmpR
system
[19]
. In these cases it is called a bifunctional sensor,
whereas when this second function is lacking, as in the case
of the CheA
e
Interface and Context
In this broad use of the term, function always occurs
within particular contexts that require specific interfaces.
Realizing a particular function in one context may require
a design that differs from that for realizing the same
function in a different context. For this reason, a class of
mechanisms that performs the same function will often
exhibit a common theme, but with important variations on
the theme that undoubtedly reflect differences in context.
This is clear from the examples of inducible catabolic and
repressible biosynthetic systems referred to earlier. For
example, the lactose and maltose systems are similar in
certain aspects of function: both represent inducible
catabolic systems and their principal effector function is
a disaccharidase activity. However, the interface of these
systems with their environment is materially different.
Moreover, a repressor controls the lactose system,
whereas an activator controls the maltose system. For
some time many molecular biologists considered these
different modes of control to be an inconsequential
accident of evolution because they performed essentially
the same 'function'. However, a more satisfying expla-
nation for the difference in control
CheY system
[19]
, it is called a monofunc-
tional sensor
[20]
. Careful analysis predicts that an
advantage of bifunctional sensors is the amplification of
cognate vs. non-cognate signals that reduces the impact
of noisy cross-talk. This is an irreducible consequence of
bifunctional over monofunctional mechanisms, indepen-
dent of the numerical values of the parameters
[20]
.
However, the magnitude of the effect will depend on the
specific parameter values. Recent experimental studies
have found evidence that supports the role of bifunctional
sensors in suppressing cross-talk
[21,22]
. Similarly, there
are numerous toxin
e
antitoxin systems that have been
characterized biochemically and genetically and yet their
cellular role is unclear
[23
e
25]
.
In other cases one starts from clear performance
requirements and attempts to understand which mecha-
nisms are best suited to meet those requirements. For
example, one might know from the ecology of an
organism that it must express a particular subsystem in
a specific context, and yet the detailed mechanism by
which the expression of the subsystem is controlled is still
unknown.
Any design principle that can be established for
a general class of such systems can be useful as a guide to
an experimental program aimed at discovering the envi-
ronmental context when the molecular mechanism is
established, or at discovering the molecular mechanism
when the environmental context is known
[15]
.
e
indeed a design
e
principle
was eventually found in relation to the
different environmental contexts in which these systems
operate. The performance of these systems differs in the
broader sense of system robustness and response time as
a consequence of their different modes of control.
Moreover, the selection of these alternative modes is
related to the duty cycle or demand for their expression in
different microenvironments
[13
e
Design
For a long time the issue of design principles in biology was
not a legitimate consideration, and in some quarters it
remains a taboo subject. This is a legacy of its misuse in
arguments against evolution and in the 'just so' stories used
to explain the evolution of biological systems. However, the
term design has a rich and well-established meaning, and
when used in the context of rigorous analysis and objective
performance criteria, provides a deep understanding of the
function and evolution of biological systems. This is now
more widely accepted, as is evident in topics being pub-
lished on biological design principles and conferences
being devoted to the subject.
Still, the question might be asked, are there design
principles or rules that govern the patterns observed among
biological systems? The answer depends upon whom one
asks. There are some biologists who would answer: 'Of
course there are rules, and it is the business of science to
discover them!' This view has a long tradition embedded in
positivist philosophy
18]
. For example, the
substrate of the lactose system is rarely present in the
natural environment of Escherichia coli, which selects for
a negative mode of control, whereas the substrate of the
maltose system is frequently present, which selects for
a positive mode of control
[13]
.
Many such biological design principles involve subtle
variations on a theme, and their identification requires
a comparative approach and quantitative criteria for
comparison
[2,12]
. In some cases one starts with a well-
characterized mechanism and attempts to identify which
performance criteria it best fulfills. Good examples here
include numerous two-component signal transduction
systems that have been characterized biochemically and
genetically and yet which are believed to be responding to
environmental signals that in many cases are unknown. The
design of prototype two-component systems exhibits
a common variation in which the unphosphorylated sensor
exhibits a second function (in addition to its primary
function
e
of
phosphorylating
the
cognate
response
the collection of empirical data,
e
