The life context: cells,
nutrients and signals
G.P. Pescarmona
Department
of Genetics, Biology and Biochemistry
University
of Torino
Running title:
Cells, nutrients, signals
ABSTRACT
The current trend in biology
is to collect more and more data as fast as possible; their number is so
overwhelming that it is almost impossible to fit them into a coherent picture
starting from data analysis. The only possibility to make experimental results
meaningful is to try to fit them into a general model of a living organism,
starting from the assumption that life in itself is very simple and that actual
complexity arises from the interaction of organisms with numberless and chaotic
environments. The aim of this paper is to define the basic factors ruling all
forms of life and their interactions with the environment.
1. INTRODUCTION
One of
the most intriguing problems of the current biological research is the fact
that the more information is collected about genes and metabolic pathways in
different organisms, the more difficult it is to fit it into a coherent
picture. When glucose catabolism was simply described in mammals as glycolysis
in anaerobiosis and Krebs cycle associated to mitochondrial electron transport
in aerobiosis, it was easy to distinguish between tissues like muscles with
high aerobic metabolism and erythrocytes with an almost complete dependence
from glycolysis for energy supply. But when we started to include all the
branches and the connections of these pathways to aminoacid metabolism not only
in mammals, which are reasonably similar, but also in very different species,
we realized that glycolysis was no longer a kind of highway for glucose
breakdown but a part of an intricate network of fluxes, whose entity and
direction was largely dependent on the type of organism and on the surrounding
environment. The use of a reductionistic approach to the description of
innumerable living organisms has led to a huge collection of data, barely
comparable.
To make a
long story short, at the present day the risk is that the more we know, the
less we understand. One of the possible approaches to overcome this difficulty
is the use of models to simulate the behavior, at different levels of
complexity, of living systems. The advantage of models is that they are much
faster to test and cheaper to run than classical in vivo or in vitro
experiments; apparently, they are not as sound as "real" experiments,
but all scientists know how many constraints are imposed on experiments to make
them reproducible and how often, more or less consciously, the results of this
kind of experiments are biased by the selection, out of the available results,
of those judged more sound on the basis of many factors. These factors include
the original working hypothesis of the authors, the prevailing opinions in the
specific scientific sector, and the opinions of the referee to whom the paper
is sent for review. That means that it is very difficult to get new information
to support original hypotheses; experiments are expensive and time-consuming,
and the judgement about the feasibility of the projects, their funding and the
acceptance of the results depends on the opinions of well-established
scientists, who would find hardly acceptable a drastic change of the current
opinion.
On the
contrary, models may be tested even with a relatively low budget, due to the
low cost of computing time. Modeling in biology is still a young science and
therefore the most common approach in building models is to create models able
to reproduce experiments. This approach is based on the hypothesis that
experiments are neutral, but it is now quite clear that experimental results
are biased by many factors, such as the type of organism used, its age, the
number of organisms tested (the response of one cell to an agonist is
completely different from the average response of 50.000 cells to the same
agonist), the local conditions, and so on. Moreover, if the final result of the
experiment is known in advance, it is quite easy to manipulate the value
assigned to the initial parameters in the model to fit almost perfectly the
experimental results. But, as we said before, there are now so many
experimental data that people are more and more confused about their meaning
and the idea of creating models with the aim of reproducing experimental
results simply makes no sense. Modeling makes sense if it allows a strong
reduction in the number of the parameters involved in the models and focuses on
the dynamical interactions between them. The core of the model is the
relationship between the parameters; the search has to be focused on the
robustness of the set of relationships, that is to say that a sound model
should be able, with minor adjustments, to predict biological events independently
of the scale (molecular or cellular or population scale). The model
construction, according to this approach, is not influenced by the final
results of any specific experiments, but it is based on a set of rules, based
on known or argued relationships between parameters. This approach has been
widely used in Artificial Intelligence (A. I.) and it is defined as a
"deep model". Models developed in this way can be checked for either
their intrinsic behavior (sensitivity to the initial parameters, sensitivity to
the environment, chaotic behavior, etc) or for their ability to predict real
world events. In the case of some lack of correspondence between the calculated
and the expected results, if the model is sound, criticism may be advanced
against both the model and the experiment, trying to find out which one is the most biased.
2. A
"DEEP MODEL" OF LIFE
There are
so many forms of life in our present environment that it seems very hard to
choose the properties that they share and can be used to establish a common
model. In the past naturalists classified organisms according to their shape,
nutrition, reproduction mode and so on. This approach was based on the analysis
of the end products of evolution. Looking into the genome we now know that
these features depend on an intricate exchange of genetic information between
both individuals and species and that genomic data correlate only poorly with
phenotype similarities, making this approach too complex and unrealistic. We
have therefore followed a different approach [1,2] defining some very simple
rules ("deep model" [3]) that can be applied to the behavior of any
living system, from the very early forms of life to the most complex societies.
Once defined, rules can be checked for their applicability and falsifiability
in any known environment, from the most primitive environments with few
molecules (CO, CH4, NH3, SH2) to the recent
environmental conditions with thousands of new molecules polluting water and
atmosphere.
All life
forms in our environment share some properties and can be described as:
Dissipative:
life takes place only in environments with an excess of energy: in our case, in
the sunlight.
Evolutionary:
formation of evolutionary trees for any molecule and organism mediated by
irreversible bifurcation followed by selection. The driving forces of
bifurcation and selection depend on the environment and can be considered
“local”.
Cyclic:
at any level from the molecules to the species, precise feedback mechanisms can
be identified that regulate the number of the objects involved in the
equilibrium.
Oscillating:
in any self-regulatory cyclic system the number of any item is changing in time
with a periodicity that depends on the size of the system – from seconds for
chemical reactions to years for prey/predator relationships – and the time the
feedback signals need to diffuse across the system.
Competitive:
as biological systems tend to expand exponentially in a finite environment,
they become – sooner or later – limited in their growth due to a shortage in
some essential factor (“nutrient”). The competition for the limiting nutrient
will locally drive the selection.
The
relationships between different properties vary and can be described as:
·
Hierarchical,
when a property is not possible unless the other(s) is (are) present. This is
the case of the dissipative behavior that is a prerequisite for any other
property of living systems.
·
Complementary,
when properties describe, from different points of view, the same set of
events. Cyclicity of a metabolic pathway is a structural feature and does not
depend on time, whereas the concentration of metabolites involved oscillates
with time. When the concentration of a metabolite reaches its lowest value, it
usually becomes limiting
and
therefore competition arises between users of the same molecule.

Fig.1:
Schematic representation of the main properties of living systems
3. DETAILED
DESCRIPTION AND EXAMPLES OF LIVING SYSTEMS PROPERTIES
1) Living
systems are dissipative [4].
Dissipative
systems do not conform to the second law of classical thermodynamics because
they are not closed systems, but they continuously receive energy from the
outside. On the earth crust excess energy comes from sunlight. Photosynthetic
organisms (plants, bacteria, algae) use radiant energy to move electrons from
more to less electronegative elements leading to the present atmosphere, which
is completely different from the original one surrounding earth before life.
The most
striking difference is the continuous increase of gaseous N2 and O2
in the atmosphere over the past 4 billion years [5]. The large availability of
a strong oxidant like O2 allows survival of all the forms of life
depending on respiration for energy production.
The whole
picture of the different forms of life possible in an environment fuelled by
sunlight is summarized in Fig.2.

Fig. 2:
The life context on the earth crust (redrawn from Morowitz H.J.[6])
In the
case of sunlight and life on earth surface, the concept of dissipative system
is quite intuitive, but how can we describe in both qualitative and
quantitative ways dissipative systems, from molecular to social level?
The
answer is that dissipative systems, as they are continuously getting energy
from the environment and using it to do something (it doesn’t matter what), are
always asymmetric. Indeed they cannot create new atoms (the energy required for
nuclear fusion is too high); they can just build up and destroy molecules, move
them around, move themselves and modify their own shape.
As a
matter of fact, in order to describe properly any dissipative system we have to
identify and quantify first all its intrinsic asymmetries and then to evaluate
which are the environmental asymmetries that can supply enough energy to drive
the intrinsic ones.
Ionic
asymmetry is shared by all living organisms; it involves both cations like Na+,
K+, Ca++ and H+ and anions like Cl-,
Pi (HPO4--) and protein carboxyl groups [7].
As
biological membranes are not fully impermeable to ions, the ion gradients between
the inner and outer compartments would spontaneously disappear. Their
persistence (with a very small shift from the average value) means that
gradients are sustained by an active process driven by different types of
ion-sensitive- ATPases [8].
ATP synthesis
requires a supply of nutrients, both reducing (carbon skeletons:
monosac-charides, fatty acids (FA), aminoacids (AA) or the light-dependent
reduced cofactors in photosynthetic organisms) and oxidizing (O2,
nitrate , sulfate). As electrons spontaneously flow from less electronegative
(C, H) to more electro-negative atoms like oxygen, life is possible wherever we
have asymmetry in the electron distribution: more electrons than expected from
electronegativity on C or H (e.g. CH4 instead of CO2 and
H2O), less on oxygen (e.g. O2 instead of H2O
and CO2). For any given organism it is always possible to identify
the local conditions (reducing and oxidizing nutrients availability) necessary
for its life. In in vitro cell growth (a good model of a closed system until
the growth medium is changed), initial conditions are characterized by plenty
of nutrients, but after a few days the nutrient will be broken down and waste
products will accumulate leading to cell death. This happens in closed systems,
while in natural systems organisms have developed motility and they move in the
direction of new sources when nutrients are locally scarce [9]. After billions
of years organisms have adapted themselves to almost any environment, and have
developed every type of machinery (enzymes, organelles) to take advantage of
the locally available nutrients [10]; indeed no form of life is possible in the
absence of nutrients in whatsoever form. Identification of nutrient type,
concentration and diffusion rate in the environment is therefore the first step
in the definition of any form of life.
Actually,
the lack of one or more nutrients (e.g. ischemia or hypoxia) leads to
disruption of intracellular ion gradients with a significant increase of
cytoplasmic H+ and/or Ca++ that activate degradation of
cellular components and induce cell death by necrosis or apoptosis [11];
through this process atoms of the dying cell are made available to the
healthier cells.
To cope
with their continuous energy requirement, in the presence of a space- or time-
dependent local lack of nutrients, cells have evolved two specific functions:
·
Motility
to move in the direction of nutrient source. In this function we can include
motility of suspended cells (bacteria), adhering cells (eukaryotic) and hyphae
formation by fungi [12].
·
Nutrients
storage (glycogen, fatty acids) to cope with the problem of unpredictability of
feeding and work requirements.
1) Living
systems are evolutionary.
Many
evidences show that living organisms are unstable and can change with time. The
time course of modifications depends on many factors linked to both the
environment and to the type of organism. The highest rates of change are
displayed by viruses; months or a few years are sufficient for the creation and
selection of new variants [13]. A hostile environment is a good reason for
change: a group of Spirochetas (Pillotina Calotermitidis) living in the termite
intestines have not changed in the past 1 billion years (the termites
supposedly did not change their diet during this time) [14]. On the other hand,
in the past few years the introduction of chemotherapy rapidly induced the rise
of new variants of parasites resistant to the therapy (Plasmodium falciparum
and chloroquine, most bacteria and antibiotics, HIV and ziduvidine) [15].
Why only
dissipative systems can be evolutionary? Because evolution is based on the
selection of the fittest in a changing environment. To cope with these external
changes organisms have to change themselves, and they can do that in two ways:
·
To change
randomly, an energetically very expensive process based on the random synthesis
of new molecules followed by selection of the fittest
·
To change
according to a blueprint, synthesizing
only the molecules fitting the new environment
As a
matter of fact a blueprint requires forecasts on the future environment; to get
this kind of information organisms should develop sensors and that would
require further blueprints, which would require additional information and so
on. Evolution without forecasts is the only possible strategy to adapt to a
changing environment. This approach requires excess nutrients for both energy
production and the synthesis of new structures.
Energy
production: glucose and fatty acids (FA) can be stored and used for the synthesis
of ATP necessary for movement; the bigger the stores, the larger the space the
organisms can wander into looking for new nutrients. Since without O2
one molecule of glucose produces 2ATP and FA are not metabolized at all,
whereas with O2 38 ATP are produced and FA are fully burnt, the
terrestrial life, which requires long trips in search of food, evolved only
when atmospheric pO2 approached its present value.
New
structures synthesis: aminoacids are essential as building blocks for synthesis
of both proteins and purine/pyrimidine bases of nucleic acids. These processes
require also a lot of ATP. Nucleotides and aminoacids are stored in a more
dynamic state than glucose and FA: they are stored as mRNA and nascent proteins
respectively. The mRNA and protein synthesis is a process continuously going
on, but the percentage of proteins effectively reaching the functional state is
very low and hence all the intermediate molecules (nascent mRNA and proteins)
can be quantified as intracellularly available stores [16].
In
synthesis, plenty of ATP, glucose, FA and scarcity of aminoacids will stimulate
cell motility and hence migration, while protein evolution will be slow due to
the relative lack of aminoacids. In the presence of excess AA, new and new
proteins can be synthesized, with a run up of evolutionary process in situ. In
any case the choice between movement and protein evolution is driven by the
local type and concentration of nutrients.
In the
case of protein evolution the birth of cell clones with different isoforms of a
specific protein usually will be followed by irreversible selection of the
clone with the highest fitness to the environment. Selection depends usually on
the fact that nutrients are always limited in any environment: the population with
the highest fitness and the highest reproduction rate will eliminate, sooner or
later, all the other less fit populations. All the evolutionary trees, either
in temporal (cytochrome C in different species, millions of years [17]) or in
spatial terms (virus variety in different geographical areas [18]) can be
considered features representative of a dissipative form of life in an
environment with limited resources. The liveliest description of dissipative
systems is the Pinball (Fig. 3).

Fig. 3: The Pinball, a
comprehensive metaphor of life.
It is
always possible to play again, provided you have enough coins (sun, nutrients,
oxygen - they depend on the type of
life). But once the ball is at the top it has only one aim: to go down. The
pathway doesn’t matter. Life behaves in the same way: it has to go on, it
doesn’t matter where and how. The winners survive and eventually reproduce, the
losers disappear. In the bifurcation of every evolutionary tree only the result
of winning players can be read. The winning choice is valid as long as the
local conditions are the same. When they change, a set of new bifurcations will
take place as well as a new selection. Life is an endless pinball game.
The
driving forces of bifurcation and selection depend on the environment and can
be considered “local”.
2) Living
systems are cyclic.
If we
look at the current forms of life we get the impression that evolution, in the
time scale of man life span, is actively at work as far as some proteins of
bacteria and viruses are concerned, but most of the basic cellular metabolisms
are stable. How do dissipative and evolutionary systems stabilize instead of
going on with an endless change? The main reason for this stability is the
existence, at any level from the molecules to the species, of feedback
mechanisms downregulating specific steps of the system. Usually, some of the
end products inhibit the first step of the corresponding metabolic pathway
(e.g. heme inhibits ALA-synthetase, estrogens inhibit cholesterol synthesis
etc.). This mechanism of downregulation of the first step of a series of events
obeys the same general rules independently of the scale and of the molecules
involved. Suitable examples of cyclic behavior are: a) calcium influx at
cellular level [19], b) TSH release from hypophysis [20], and c) a predator/prey system [21].
A further
constraint to an endless evolution is the lack of atoms: on the earth crust the
number of atoms is fixed (atom number can change in environments with
temperatures of millions of degrees) and competition for them strongly limits
the evolution of new molecules and metabolic cycles.
3) Living
systems are oscillating.
In any
self-regulatory cyclic system the number of any item is changing with time with
a periodicity that depends on the size of the system: seconds for chemical
reactions at the cellular level (NADH, ATP/ADP, Ca++) and heartbeat; a few minutes for oxygen consumption; from minutes to
hours for blood hormones; months for the menstrual cycle; years for
prey/predator relationships. The period length depends not only on the
geometrical size of the system but also on the diffusion rate of the feedback
signals across the different compartments of the system [22].
4) Living
systems are competitive.
As
biological systems tend to expand exponentially in a finite environment they –
sooner or later – become limited in their growth by the scarcity of some
essential factor (“nutrient”) and the competition for the limiting nutrient
will locally drive the selection. Since life started to evolve billions years
ago, all possible nutrient sources have been exploited and life is now possible
only at the expense of other forms of life.
Complex cyclic and
oscillating systems also evolve with time exactly as molecules do, in order to
become as fit as possible to the environment. This fitness can be evaluated
measuring the behavior of the system with time: the most studied example is the
heartbeat, but calcium oscillations, breathing or metabolic rates can be
studied as well. Heartbeat series show
a chaotic behavior that can be quantified with a fractal number [23]; the
higher the number the lower the predictability of the characteristics of the
single heartbeat. This poor reproducibility is typical of an unstable system
(at the chaos edge) which is very sensitive to even the smallest change of
local conditions (rest, fatigue, digestion, emotions) and therefore is
indicative of a good degree of fitness (remember that at molecular level
proteins usually work at a substrate concentration near Km in order
to get maximal dependence of reaction rate from the environment). This
instability is lost and the heartbeat or breathing becomes oscillating almost
independently from the environment in the presence of diseases (e.g.
Cheyne-Stokes breathing) [24].
A second form of
fitness for oscillating systems is the strong reduction of the oscillations of
the parameter under control. A good example is the serum Ca++; the
range of oscillation of Ca++ in serum is very narrow (from 2.05 to
2.8 mM) but this stability is kept in a dynamic way. When Ca++ is
high, it is stored in bones, and when it lowers immediately parathormone (PTH)
is released activating Ca++ release from bone [25].
As a consequence,
serum Ca++ oscillations with time are barely detectable, but PTH
oscillations are huge and display the same chaotic behavior of heartbeat.
Stability of Ca++ availability improves the reliability of all
signaling pathways using Ca++ as a second messenger. The costs of a
huge store, like bone, and of a continuous turnover of PTH are compensated by
the evolutionary advantage of a more efficient signaling.
Stability in
biological systems is therefore constitutively different from stability of
rocks and diamonds: rocks and diamonds are characterized by stable multiple
bonds with electrons perfectly shared by adjacent atoms and therefore not
breakable unless very high temperatures are used. In living organisms stability
is attained through very expensive processes planned to keep the internal
milieu constant and to make enzyme activity reproducible. Enzyme activity
depends on substrate concentration, temperature, metal ions, pH. Multiple
systems have been worked out to allow the control of these key factors: the H+ exchanger, Ca++ channels,
Na+ channels, Ca++ ATPases, glucose carriers, and so on
[26]. Moreover, many of these proteins can exist in phosphorylated and not
phosphorylated forms. The need for multiple forms of proteins exerting very
similar effects can be explained assuming that in order to survive cells must
be sensitive to the environment, that is to say, to substrates, ions or H+
concentration changes. Protein interaction with ligand is expressed through Kb
or Km.

Fig.4: Protein
activity (carriers, enzymes, receptors) depends on the number of ligand
molecules bound to the protein. Kb (carriers) and Km
(enzymes) define the affinity of the protein for the ligand. V is the enzymatic
reaction rate.
Kb (or Km)
is, by definition, the ligand concentration at which 50% of the protein is in
the bound form and at which very small changes in ligand concentration strongly
affect the binding and consequently the protein activity (see Fig. 4). Any
system gets its highest responsivity to environmental ligand changes near its Kb
or Km and in a world where the concentration of molecules changes,
the proteins must also change their Kb to fit local conditions.
Biological systems have evolved different solutions to put up with this
variability:
a) Phosphorylation. Ca++ ATPase e.g.
exists in a nonphosphorylated form (Km for Ca++ = 10-7M)
and in a phosphorylated form (Km for Ca++ = 10-6M).
Quiescent cells have an average Ca++ concentration of 10-7
and the ATPase is fitted to manage small random changes in this range. A
stimulated cell phosphorylates the enzyme via protein kinase A or C, converting
it to a form sensitive to small changes of Ca++ in the range of 10-6,
typical of activated cells [27].
b) Multiple forms with different Km or
pH dependent activity. Good examples are proteases, DNAases etc, involved in
apoptosis or digestion. Protein kinases
and protein phosphatases are the enzymes that exist in the highest number of
isoforms; so far they are only partially characterized. A fruitful approach to
the characterization of their exact function should start from the definition
of the conditions of maximal sensitivity (pH, Mg++, ATP, Ca++,
Na+, K+ etc) to environmental factors. The advantage of
protein phosphorylation is the partial lack of specificity of these enzymes
that allows them to control multiple metabolic pathways simultaneously [28].
The Kb
curve is one of the few parameters useful to describe the properties of a
system based on chemical reactions, as it is the case with all living
organisms. Chemical reaction rates depend on temperature, reactant
concentrations and protein/ligand affinity. Temperature can be easily measured
and evaluated; the strong dependence of chemical reactions on temperature
explains why keeping constant the body temperature has been the choice of so
many organisms: the gain in predictability of chemical reactions is worth to
have a costly central heating.
Concentration is
usually symmetric in dilute aqueous media like sea or synthetic culture media,
but in most environments, where more cells or species live and the diffusivity
of reactants is not free, gradients of almost every molecule can be described.
Every time we see a shape arising in a biological system we have to wonder
which molecular gradient is underlying it. In some cases the answer is easy:
sunlight for the tree branches, water, nitrogen and iron for the roots, but
what about mammal-ian lungs, so exquisitely fitted to exchange oxygen, but
shaped during the fetal life in the absence of air? [29] Gradient evaluation
requires the knowledge of many factors including, for each molecule, synthesis
and degradation rates, diffusion from different environments, and spatial
localization of metabolic steps.
Affinity is the
ability of molecules to bind to each other. In order to react, molecules have
to meet in a proper way and this goal may be attained either in a statistical
way, by increasing the average speed of molecules (high temperature), or in a
structural way, by stretching the shape of the molecules to increase their
reciprocal fitness.
As living systems
exist only at relatively low temperatures (mostly between 20 °C and 37 °C)
their high metabolic efficiency can be achieved only by increasing the
molecular fitness [30].
Tricks available to
do that rely primarily on the different electronegativities of atoms that
regulate both polarity and acidity of molecules; small localized differences in
charge can change the tertiary structure of proteins and hence their Kb
to specific ligands including binding between subunits of a functional unit
(homo- and heterodimers e.g. of growth factor receptors).
All biochemical
systems from the simplest to the most complex can be described in terms of
binding kinetics identified by a Kb, correspond-ing to the substrate
concentration at which the system displays the maximal instability and
sensitivity, and by a slope identifying the concentration range in which the
system can be regulated.
Some examples:
·
The behavior of acids
(strong and weak) depends on the affinity of the conjugated base (A-)
for H+. The Kb in this case is called Ka. pKa
of a group also depends on the micro-environment and proteins can supply many
different microenviron-ments to the same group modulating its pKa.
Changes of protein shape can affect pKa,
but also protonation of a weak acid
group can change protein structure [31].
·
The hemoglobin
dissociation curve and its dependence on pO2, pH, 2,3-DPG and type
of globin [32].
·
The allosteric
behavior of enzymes [33]
·
The shift of Kb
or Km of many proteins (enzymes, carrier etc.) upon phosphorylation
[34]
4. CELLS
Organisms (from
bacteria to nations) are delimited by dynamic boundaries (cell membrane, skin,
borders) where essential nutrients or goods are exchanged with the environment
(including the whole of all other organisms) [35]. The working hypothesis is
that rules valid at the microscopic level (cell) will be applicable to upper
levels too, since complex organisms are made of cells and societies are made of
organisms. We can define the cell as the simplest living unit, able to
duplicate itself and to exchange nutrients with the environment (competition for
nutrients). Chemical composition (characterized from the structural and
molecular points of view), energy requirements and boundary properties are the
parameters that must be taken into account for a correct modeling of cell
behavior. Chemical composition identifies all atoms and molecules necessary for
the synthesis of structural molecules (essential "nutrients"
including some aminoacids, vitamins, ions), energy requirements identifies the
daily uptake of oxidant (e.g. O2) and reducing (e.g. glucose) agents
necessary for energy metabolism. Cells have an additional property: each of
them differs from the other objects of the same class and from itself at a
different time. The limits of the changes allowed to the single parameter
(remember that cells are identified by dynamic boundaries) without leaving a
specific class have to be included in the cell class definition. Living objects
are self-made and their structure strongly depends on both the genetic
background (that affects the general properties of the system) and the local
supply of energy and molecules required as building blocks for the new
organism. The possibility for the cell of changing with time depends strongly
on the energy supply and on the conservation of all its asymmetries (ions, H+,
membrane phospholipids, etc) even when stressed by environmental stimuli. Since
at the microscopic level stimuli from the environment are randomly distributed,
any single cell is always different from its previous self and from the
neighboring cell with a different history. We have previously described how
evolutionary living systems, when adapted to their environment, are apparently
stable, but this stability has a high energy cost.
Cell morphology
follows the same rule; for 200 years from its foundation by the French
researcher Bichat, histology has described hundreds of different types of
cells. It was a morphological classification based on reproducibility of cell
shapes in the same organ of different individuals and shape differences between
cells of different organs in the same individual. Now we know that even if the
vast majority of the cells of an organ is differentiated and displays the
expected features (according to the textbooks), it is always possible to find
out a small number of undifferentiated, totipotent cells in any tissue and that
this pool of cells is a reservoir for the replacement of dying cells [36]. If
we combine this information with the fact that genetic information is exactly
the same in all the cells of an individual, it becomes clear that all the
morphological patterns we know depend on the local conditions where the cells
live. Stability and reproducibility of cell shapes is not due to some hidden
information of the genome or some blueprint but to the stability and the
reproducibility of the milieu surrounding the cells.
Let's try to describe
why liver is liver as a function of its local conditions. It is the heaviest
organ in the body because it collects 100% of the nutrients we introduce with
the diet. It gets 50% venous blood very rich in any kind of nutrients (portal
vein) and 50% fully oxygenated blood (hepatic artery); a lot of nutrients but
50% less oxygen than the other tissues. Moreover, hepatic cells are not
directly stimulated by nerves and do not require high rate of ATP synthesis to
perform heavy-duty mechanical work like muscle. ATP is therefore utilized for
synthetic processes: glycogen is produced and stored, cholesterol,
triglycerides and proteins are produced and poured into the blood. The entire
liver enzymatic machine is not evenly distributed in the liver, but at the
microscopic level there is a clear zonation of enzymes and cytochromes
according to the distance from the cell to the artery. The oxygen gradient is
responsible for this behavior. Up to now the liver structure and function can
be explained on the basis of local conditions, determined by nutrient
availability and gradients. To fill the picture we can evaluate the molecular
mechanisms (nuclear factors, Ca++, cAMP e.g.) of the induction of
the different proteins but we do not have to invoke new mechanisms in addition
to nutrient and oxygen gradients [37].
Liver is also
sensitive to hormones like insulin, glucagon, and growth hormone (GH), which
modify its function. Hormones themselves are not nutrients (usually present in
10-3 M concentration) as they exert their effect at very low
concentrations (10-9M) and cannot modify substantially the cell
composition. Insulin release from pancreas tells the body cells that in the
next half an hour there will be plenty of glucose. Insulin cannot be used for
energy metabolism, but can increase the number of glucose carriers on the cell
plasma membrane, improving the efficiency of glucose uptake. Insulin has no
effect unless it activates glucose uptake. Glucose without insulin and insulin
without glucose are both threatening to cell life.
The half-life of
small molecules, like cofactors or metabolites, is in the range from minutes to
hours, the half life of structural molecules, like proteins from hours to
months [38]; the metabolic state of a cell (ATP/ADP ratio, NAD/NADH ratio,
glycogen stores) is representative of the past of the cell life in terms of
minutes or hours, the proteins pattern in terms of hours or months, according
to the protein. As in many cases, the life of cells is orders of magnitude
longer than that of their components. Surviving in a changing environment has a
high energy cost: the more rapidly the environment is changing the shorter the
half-life of cell components to fit the changes.
This continuous turnover
of molecules requires an incessant uptake of new molecules from the outside,
making feeding the most prominent activity of the cell. Feeding includes also
respiration (O2 uptake) and in a way reproduction; feeding indeed
requires a set of complex strategies like acid digestion, receptor- mediated
uptake, and movement in the direction of the nutrient source. The splitting of
one cell into more cells allows exploratory movements in search of food in
different directions, substantially increasing the probability of survival of
the species.
5. NUTRIENTS
Local availability of
nutrients plays a fundamental role in cell life.
Factors affecting
nutrient diffusion (medium viscosity, temperature, blood flow, angiogenesis,
etc) will affect the growth and shape of the organisms as well as the energy
supply. Cells can actively move in search of nutrients: in bacteria a rotating
flagellum drives the cell in the right direction, while in eukaryotic cells
nutrient uptake is mediated in most cases by endocytosis coupled to vesicle
acidification, a process that is also responsible for the movement in the
direction of the nutrient source [39]. Movement involves a) secretion of
proteases able to digest surrounding proteins and to decrease medium viscosity
(metallo-proteases) [40] b) adhesion to extracellular matrix c) cytoskeleton
rearrangement and d) directional acid vesicle flow [41].
Details of the
mechanisms involved in movement show that they follow the same rules: to use
the local tools and the available nutrients to get the best result with the
minimal effort. Strategies differ from case to case. Bacteria living as
separate organisms in aqueous solutions usually move using their flagellum as a
propeller, but under certain conditions the rely for movement on the Brownian
motion of water and the use of their
flagellum as a rudder. In both cases
their movements will be driven by nutrient gradients[42].
As a whole, the
winning strategies of living organisms (remember that only winners survive and
we have no information on the strategies of the losers) share a common feature:
they are strictly dependent on and adapted to a perennially changing
environment. As this adaptability cannot depend on a blueprint based on
forecasts, it has to be based on a very efficient network of information about
both the environment and the other populations living there.
6. SIGNALS
Think global, act
local
In a complex
environment like ours even an outstanding improvement in factors involved in local
competition for nutrients cannot give a strong evolutionary advantage to a
cell. To increase their fitness cells have had to learn to get information
about their present and future environment. We call signals all (physical or
chemical) events able to transmit information to a cell. Signaling requires at
least two components:
a) a physical (e.g. sound) or chemical (e.g.
hormone) event that it is released into the environment in specific conditions
and that we can call a sign (S)
b) a cellular structure able to react with
(S) modifying the cell behavior
according to the concentration and direction of (S). These structures are
called Receptors (R).
Modeling of complex systems requires a correct
definition of all parameters involved in signaling. They include:
a) all
local conditions required for the release of (S) n
b) diffusion rate of (S) n in the
specific environment
c) local distribution and characterization
of different (R) n
d) signal transduction pathways at cellular level, whose degree of signal amplification
depends on local conditions
e) combination of all previous factors to get all
or most of the information spread along
the whole signaling network influenced by
(S) n
A proper combination
of the aforementioned factors leads to thousands of different patterns of
information about local and distant availability of essential nutrients
delivered to cells or groups of cells with different metabolic requirements.
The diffusion rate of (S) n
in this context plays a major role as a delay in the signal transmission
(common and unpredictable in complex systems) can induce signal oscillations
out of phase with nutrient oscillations resulting in misleading information.
(S) n are classified as sound, light, hormones,
cytokines, chemokines, growth factors, and so on, according to their major
effect, but they share a common property: the ability to give information about
the existence of specific environments or about the availability of one or more
nutrients. The minimal information content carried by any sign is that all the
conditions required for its synthesis are present.
The exact definition
of the role of any molecule involved in signalling (cytokines, hormones etc.)
should always include the search of all the factors necessary for its synthesis
(Fig. 5).

Fig.
5: The search for all factors required for the synthesis of a sign should be
always thoroughly performed.
For example, thyroid
hormones depend on oxygen supply, estrogens on heme and oxygen, insulin on
intestinal glucose absorption, LPS
(lipopolysaccharide) on bacterial infections, nitric oxide (NO) requires
arginine, oxygen, heme, NADPH, GSH [43]. This set of information is what we can
define as the "meaning" of the signal (S) n. As far as the target cells are concerned, only
skeletal muscles and adipocytes will get from insulin the information that
glucose will increase in the next ten minutes. All other cells just will sense
the local glucose increase.
Modeling of the
living state has therefore to take into account modeling of both nutrients and
signal molecules independently (uptake, diffusion, binding constants), with
their adequate time scales. What is
important is to avoid excessive emphasis on the role of signals in models. Cell
life requires energy and essential nutrients for work performances or
duplication; signals can freely modulate cell activity in the presence of
unlimited nutrients, but can never overcome the lack of some essential
nutrient. Models based on cells + nutrients are possible and meaningful and can
validly mimic local effects; modeling of complex organisms or environments
requires also the incorporation of signals into the model. Models solely based
on cell/signal interactions are meaningless.
Modeling a complex
environment including both nutrients and signals requires previous assignment
of molecules to one of the two classes of molecules.
Nutrients (glucose,
aminoacids, fatty acids) are in the range of 10-3 M, while signs
(chemotactic factors, hormones, cytokines, odors, growth factors) are comprised
between 10-6 and 10-9 M.
Every sign (S) should
be investigated to identify all the conditions necessary for its synthesis
(precursors, pO2, pH) and that have to be included as information
carried by the (S) molecule.
As transfer of
information takes place only when cells have proper receptors (R) also
receptors have to be treated as signs. Signs carry information from the near or
far environment, while receptors carry infor-mation about the inside of the
cell: nutrients and energy stores, filling of calcium stores, ATP and cyclic
nucleotides level, and so on.
Signaling is possible
only when all the requirements for (S) and
(R) synthesis are fulfilled.
Additional factors
that have to be checked are:
·
Binding constants (Kb)
of receptors for their ligand; they are usually in the same range of prevailing
concentration of the ligand/sign in body fluids (between 10-6 and 10-9M).
·
Degree of linearity
of the signal as a function of ligand concentration. Role of homodimers and
heterodimers, dimers or tetramers, allosteric effectors.
·
Signal amplification;
it depends on the asymmetries of ions (Na+, K+, Ca++)
in the different cell compartments (calciosomes, mitochondria, nuclear
membrane). The properties of both passive and active ionic channels spanning
the different membranes have to be taken into account.
As in the
post-genomic era the structure of all proteins is known, it is also possible to
try to answer some questions about the evolution of signaling systems.
What is born before, the
sign or the receptor?
From the logical
point of view the sign should always come earlier [44], but if we want also
some additional biological evidence we can just check the AA composition of
signs and their corresponding receptors. The AA pattern is different in
environments with more or less oxygen. Usually the aminoacid percentage in
proteins corresponds to the prevailing AA pattern when the protein was created
in the evolution. As pO2 has been continuously growing during
evolution, AA composition can give some reasonable hint about the age of
proteins and can help us to identify which one came earlier.
One striking example
of applicability of this approach is that of proteins working as protease
inhibitors [45]. They usually have a structure rich in aminoacids frequent when
pO2 is low and usually their synthesis is induced when tissues are
hypoxic: in this situation they act as antiapoptotic agents and seem to play an
intelligent role. But when billions of years ago pO2 was low and
what we call protease inhibitors were the prevailing proteins, no evolution at
all was possible. The increase in the number of signs is a prerequisite for the
development of information exchange between molecules and between cells. The
combination of a protein with two proteases of different specificity can
produce tens of signals (Fig. 6).
But in the presence
of protease inhibitors the number of signs will be drastically reduced. In a
world based on signaling protease inhibitors have to be quiescent; they are
strongly anti-
evolutionary and
become useful only when cells are dying.

Fig. 6: Role of
proteases in the evolution of signaling. Splitting of one protein into many
peptides diffusing with different rates and interacting with different
receptors dramatically increases the signalling power of the system.
A similar approach
can be applied to many organisms and cell types, allowing a better
comprehension of the living systems "intelligence" based on a
powerful hierarchical system of information, whose features can be better
understood if interpreted according to the approach described in this paper.
7. APPENDIX
Life in a pond
Life of a bacterial colony in an isolated pond
can be a good field application to check the correctness of the approach to the
description of a living system detailed in the previous sections.
Let us
consider:
·
An
isolated pond with a cyclic nutrient supply (representative e.g. of seasonal
variations in nutrient availability, Fig. 7), allowing description of cell life
under both high and low nutrient conditions.
·
a single
cell population whose behavior, as a function of nutrients, can be described by
a logistic curve, with bacterial growing rates dependent on

Fig. 7: Nutrient
behavior in an isolated pond with cyclic feeding and
growing organisms

Fig. 8: Logistic
curve of bacterial growth in a closed environment
nutrient availability (Fig. 8).
·
an
oversimplified life style (for a bacterium) with lack of resistance forms (like
spores e.g.) to overcome starvation. Cells can reproduce until nutrients are
available, afford starvation for a while, and then die. This assumption makes
the model more general and applicable also to eukaryotic cells. The cell count
in this case will display an oscillating behavior very similar to that of
nutrients but slightly shifted to the right (dotted line), a typical
predator/prey behavior (Fig.9).
But we know that long-lived systems, even if they
are structurally oscillating, display very reduced oscillations. How did they
get that?

Fig. 9: Time course
of cell number (dotted line) as a function of nutrient availability
Let's try to build up
a possible scenario leading to oscillation smoothening, given an organism fully
described by its general properties (dissipative, evolutionary etc.).
Our dissipative
bacteria feed and grow while there are enough nutrients and then die. But we
know that protein synthesis is not a straightforward process, much more mRNA
and many more proteins start to be synthesized than actually get to a mature
form. Both nascent mRNA and proteins can undergo splicing and rearrangement.
New and new proteins in search of their function are produced by well-fed
cells. The fate of these orphan proteins without a well-defined role for the
cell producing them may be very different from case to case. Albumin produced
by liver cells is discarded by an overfed liver (it takes up 100% of what we
eat) and becomes a nutrient for all other body cells, even if it is a waste for
the liver. Another possibility is that the orphan protein becomes a membrane
protein; it will not be discarded but it will enter the basic structure of the
membrane.
Different cells can
develop different types of soluble or membrane proteins; the absolute number of
molecules produced in any case will depend on the nutritional status of the
cells. It will be higher when the nutrients are more abundant. Let us assume
that one of these proteins dimerizes (we know plenty of homodimeric membrane
proteins) and that dimers behave as active protein kinases or as calcium
channels leading to a local calcium influx. There are hundreds of protein
kinases activated either directly or indirectly by calcium. In case one of them
phosphorylates a carrier for a specific nutrient (glucose, aminoacids,
phosphate) and the phosphorylation lowers the Kb of the carrier for
the nutrient (dotted line: native form, open circles: phosphorylated form) we
would have a well-defined evolutionary advantage for the clone derived from
this cell (Fig. 10).

Fig. 10: Nutrient
uptake rate as a function of extracellular concentration for two carriers with
different Kb
Indeed, at lower
substrate concentrations, these cells are able to take up nutrients at a faster
rate (open circles) and on the short or long run, depending on the difference
between the two Kb's, they will be able to eliminate all
other cells that are
less effective (continuous line) in feeding at low nutrient concentrations
Fig. 11: Cells with
lower Kb (open circles) take up more nutrients than cells with
higher Kb (continuous line)
These cells
nevertheless are still subject to death by starvation; they are able to sense
the nutrient (the dimer is a "sign" of nutrient abundance) and to
transform this information into a behavioral change (the "sign" is
used as a "signal" via protein phosphorylation and Kb
change), but they are still unable to sense the future lack of nutrients. Which
molecule can tell a cell that in the near future nutrients will decrease? Any
molecule produced by a living cell; live cells (don’t forget they are
dissipative organisms) include in their definition a continuous nutrient
consumption and therefore any molecule produced and released by these cells
contains the information of a future constant nutrient breakdown. Plenty of
these molecules have been described in the control of bacterial growth and are
grouped under the name of "quorum sensing", a "sign" of the
number of living cells [46].
Different mechanisms
can be devised to transform this "sign" into a "signal" and
some of them have already been characterized [47].

Fig. 12: Biphasic
effect of a ligand on a dimeric receptor. Activation at low concentration, inhibition
at high concentration
One of the simplest
models relies on the behavior of intrinsic membrane proteins working e.g. as
calcium channels in the homodimeric form and inactive in monomeric form. In the
simplest form the model requires
as a first step the
already described evolution of a dimeric protein kinase or calcium channel
giving an evolutionary advantage to the cells
bearing it. Further
evolution (it doesn’t matter how much time it requires) can produce monomers
able to bind with high affinity (lower Kb) a protein secreted (the
ligand) by the same cell. Low concentrations of the secreted protein will
increase the percentage of dimers giving to these cells an additional advantage
in feeding rate over the cells with membrane
proteins not binding
the ligand (dimer formation leads to nutrient uptake activation).
The final result will
be a further shift of the Kb curve to the left. The effect of this
shift in an environment with a limited and oscillating supply of nutrients
could be dangerous for the survival of the cell population due to the
lengthening of the period of food deprivation (Fig. 13: dotted line versus
continuous line, the time on abscissa is representative of the interval between
two successive nutrient supplies).

Fig. 13: Cell
population with higher uptake rate wins the competition for nutrients, but dies
sooner and has no evolutionary advantage
But as every monomer
has a binding site for the ligand, high ligand concentrations will lead to a
competition of ligand molecules for monomers, decreasing the dimer formation
and the corresponding nutrient uptake activation. The final result will be a
net reduction of the cell replication rate (Fig. 14: dotted line after time 5).

Fig. 14: Cell
population with biphasic uptake rate wins the competition for nutrients and
survives longer evolving toward an almost stationary state
The overall effect
will be to postpone cell death due to lack of nutrients and to reduce the
amplitude of the cell number oscillations. Provided constant oscillations in
nutrient supply (whose period corresponds to the length of the oscillation
shown in the previous figures) and enough time for an evolutionary process to
take place, this mechanism will lead to the development of a ligand/receptor
coupling with a Kb that permits a very fine regulation of the
population size with barely detectable oscillations.
This very simple kind
of dimeric receptor displays a biphasic behavior as a function of ligand
concentration; activation at low concentrations and inhibition at high
concentrations. This feature, mimicking an “intelligent” behavior, clearly
offers a strong evolutionary advantage in adaptation to environments with a
limited supply of nutrients, thus explaining why this kind of dimeric structure
is currently used by the vast majority of membrane receptors.
The
model can easily be adapted to mixed cell populations, assuming that the receptor
evolves in one population and the ligand in a different one [48]: the result
will be the behavior typical of a complex organism, where the cell growth in
different organs is finely tuned by reciprocal interactions.
REFERENCES
1.
Scalerandi, M., Romano, A., Pescarmona,
G.P., Delsanto, P.P. and Condat, C.A. 1999, Phys. Rev. E, 59, 2206.
2. Scalerandi, M., Pescarmona, G.P., Delsanto, P.P.and Capogrosso Sansone,
B. 2001, Phys. Rev. E, 63,11901.
3.
Yuan, Z.and Wen, C.
1992, Lect. Notes Comput. Sc., 585, 307.
4. Toussaint, O.O. and Schneider E.D., 1998, Comparative
Biochemistry and Physiology - Part A: Molecular & Integrative Physiology,
120, 3.
5. Williams, R.J.P. and Frausto da Silva, J.J.R.
1996, The Natural Selection of the Chemical Elements, Oxford University Press.
6. Morowitz ,H.J. 1968, Energy flow in Biology, Academic Press, New York
7. Devlin, T.M. 1997, Biochemistry, Wiley-Liss,
New York, 7.
8. Sperelakis, N. 1998, Cell physiology, Academic
Press, 171.
9. Bray, D. 1992, Cell movements, Garland
Publishing, inc., New York & London, 31.
10. Harold, F.M. 1986, A study of bioenergetics,
W.H. Freeman and Company, New York, 428.
11. Yu, S.P., Canzoniero, L.M.and Choi D.W. 2001,
Current Opinion in Cell Biology, 13, 405.
12. Rayner, A.D.M. 1997, Degrees of Freedom, Imperial College Press,
181.
13. Fitch,
W.M. 1996, Mol. Phylogenet. Evol., 5, 247.
14. Margulis, L. 1993, Symbiosis in cell evolution, W.H. Freeman and
Company, New York, 282.
15. Roos,
T.C., Alam, M., Roos, S., Merk H.F. and Bickers, D.R. 2001, Pharmacotherapy of
ectoparasitic infections., Drugs, 61, 10.
16. Aon,
M.A.and Cortassa, S. 1997, Dynamic Biological Organization, Chapman & Hall,
402.
17. Meyer,
T.E. 1991, Biochim. Biophys. Acta., 1058, 31.
18. Rodriguez, L.L., Fitch, W.M.and Nichol S.T. 1996, Proc. Natl.
Acad. Sci. U S A , 93, 13030.
19. Goldbeter, A. 1996, Biochemical oscillations and cellular rhythms,
Cambridge University Press, 351.
20. Frayn,
K.N. 1996, Metabolic Regulation, Portland Press, 92.
21. DeAngelis, D.L. 1992, Dynamics of nutrient cycling and Food webs,
Chapman & Hall, 186.
22. Goldbeter, A.., 1996, Biochemical oscillations and cellular
rhythms, Cambridge University Press, 491.
23. Havlin ,S., Buldyrev, S.V., Bunde, A.,
Goldberger, A.L., Ivanov, P.C., Peng , C.K. and Stanley, H.E. 1999, Physica A., 273, 46.
24. Francis, D.P., Willson, K., Davies, L.C.,
Coats, A.J. and Piepoli, M. 2000, Circulation, 102, 2214.
25. Angeletti, R.H., D'Amico, T.and Russell, J. 2000, Regulation of
parathyroid secretion. Adv. Exp. Med. Biol., 482, 217.
26. Nilius, B.and Droogmans,
G. 2001, Physiol. Rev., 81, 1415.
27. Sperelakis, N. 1998, Cell Physiology, Academic Press, 132.
28. Shi,
L., Potts, M. and Kennelly, P.J. 1998, FEMS Microbiology Reviews, 22, 229.
29. Weibel, E.R. 1991, Am J Physiol, 261, L361.
30. Sheridan,
P.P., Panasik, N., Coombs, J.M. and Brenchley, J.E. 2000, Biochimica et
Biophysica Acta (BBA)/Protein Structure and Molecular Enzymology, 1543, 417.
31. Jaenicke, R. 1999, Progress in Biophysics and
Molecular Biology, 71, 155.
32. Garrett, R.G. and Grisham C.M. 1995, Biochemistry, Saunders
College Publishing, 417.
33. Newsholme, E.A. and Start, C. 1977, Regulation in Metabolism, John
Wiley and Sons, 34.
34. Dean,
W.L. and Sullivan, D.M. 1982, J. Biol. Chem., 257, 14390.
35. Rayner, A.D.M. 1997, Degrees of Freedom, Imperial College Press, 1
36. Clarke, D. and Frisén, J. 2001,
Current Opinion in Genetics and Development, 11, 575.
37. Samson, R.L., Scoazec, J.Y., D'Errico, A., Fiorentino, M.,
Christa, L., Moreau, A., Roda, C., Grigioni, W.F.and Feldman, G. 1996, Hepatology, 24, 104.
38. Hochstrasser, M. 1995, Current Opinion in Cell Biology, 7, 215.
39. Pescarmona, G.P., Morra, E., Aldieri, E.,
Ghigo, D. and Bosia, A. 1998, MRS Proceedings, Materials Science of the Cell,
vol. 489, 205.
40. Chang, C., Werb, Z. 2001, Trends in Cell
Biology, 11, S37.
41. Petit,
V., Thiery J.P. 2000, Biology of the
Cell, 92, 477.
42.
Armitage,
J.P. 1999, Adv. Microb. Physiol., 41, 229.
43.
Ghigo, D., Geromin, D.,
Franchino, C., Todde, R., Priotto, C., Costamagna, C., Arese, M., Garbarino,
G., Pescarmona, G.P. and Bosia, A. 1996, Amino Acids, 10, 277.
44. Herder,
G.J. 1791-94, Samtliche werke, Weidmannsche Buchhandlung, Berlino 1887
45. Fager, N. and Jaworski, D.M. 2000, Mechanisms
of Development, 98, 105.
46. Lazazzera, B.A. 2000, Current Opinion in
Microbiology, 3, 177.
47. Aroeti, B., Okhrimenko, H., Reich, V. and
Orzech, E. 1998, Biochimica et Biophysica Acta (BBA)/Reviews on Biomembranes,
1376, 57.
48. Crespi, B.J. 2001, Trends in Ecology &
Evolution, 16, 178.