See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/26768062
Metabolic systems maintain stable nonequilibrium via thermodynamic buffering
Article in BioEssays · October 2009
DOI: 10.1002/bies.200900057 · Source: PubMed
CITATIONS
READS
19
240
2 authors:
Abir U Igamberdiev
Leszek A Kleczkowski
219 PUBLICATIONS 3,407 CITATIONS
108 PUBLICATIONS 3,304 CITATIONS
Memorial University of Newfoundland
SEE PROFILE
Umeå University
SEE PROFILE
Some of the authors of this publication are also working on these related projects:
Several research projects on the topics of plant science, bioenergetics and theoretical biology View
project
All content following this page was uploaded by Leszek A Kleczkowski on 11 January 2017.
The user has requested enhancement of the downloaded file. All in-text references underlined in blue are added to the original document
and are linked to publications on ResearchGate, letting you access and read them immediately.
DOI 10.1002/bies.200900057
Problems and paradigms
Metabolic systems maintain stable
non-equilibrium via thermodynamic buffering
Abir U. Igamberdiev,1* and Leszek A. Kleczkowski2
1
2
Department of Biology, Memorial University of Newfoundland, St. John’s, NL, A1B 3X9, Canada
Department of Plant Physiology, Umeå Plant Science Centre, University of Umeå, 901 87 Umeå, Sweden
Here, we analyze how the set of nucleotides in the cell is
equilibrated and how this generates simple rules that
help the cell to organize itself via maintenance of a stable
non-equilibrium state. A major mechanism operating to
achieve this state is thermodynamic buffering via high
activities of equilibrating enzymes such as adenylate
kinase. Under stable non-equilibrium, the ratios of free
and Mg-bound adenylates, Mg2R and membrane potentials are interdependent and can be computed. The adenylate status is balanced with the levels of reduced and
oxidized pyridine nucleotides through regulated uncoupling of the pyridine nucleotide pool from ATP production in mitochondria, and through oxidation of
substrates non-coupled to NADR reduction in peroxisomes. The set of adenylates and pyridine nucleotides
constitutes a generalized cell energy status and determines rates of major metabolic fluxes. As the result,
fluxes of energy and information become organized spatially and temporally, providing conditions for self-maintenance of metabolism.
Keywords: adenylate kinase; metabolomics; regulated uncoupling; respiration; stable non-equilibrium; thermodynamic
buffering
Introduction
The field of metabolism, complex as it is, can be hardly
reduced to simple rules. However, the pathways of cellular
metabolism are almost identical across widely divergent
organisms, which may simplify finding the basic principles of
metabolome organization and functioning.(1) Actual rates of
metabolic fluxes depend on energy (stored and supplied
mainly via ATP) and redox potentials (estimated by reduction
levels of NAD for catabolic reactions and NADP for anabolic
pathways). Several concepts of cell energy status have
been developed, including adenylate energy charge
[(ATP þ ½ADP)/(ATP þADP þAMP)],(2) catabolic redox
charge [NADH/(NADH þ NADþ)] and anabolic redox charge
[NADPH/(NADPH þ NADPþ)].(3) These ratios have some
relation to the state of the cell and to the rate of metabolic
*Correspondence to: A. U. Igamberdiev, Department of Biology, Memorial
University of Newfoundland, St. John’s, NL, A1B 3X9, Canada.
E-mail: igamberdiev@mun.ca
BioEssays 31:1091–1099, ß 2009 Wiley Periodicals, Inc.
fluxes,(4,5) but they are too general and do not include
major factors governing these processes, e.g., free magnesium concentration in the case of the adenylate energy
charge.(6–8)
Although living systems operate far from equilibrium,(9)
non-equilibrium fluxes should be steady. This stable nonequilibrium state(10) can be reached via a ‘‘fitting function’’ of
special thermodynamic buffering enzymes that equilibrate
fluxes of load and fluxes of consumption of major metabolic
components, e.g., ATP and pyridine nucleotides.(11) Thermodynamic buffering generates steady metabolic fluxes via local
equilibrations and regulated uncoupling or ‘‘slippage’’.(12)
Understanding the rules of such equilibration will help in
developing a computational approach for calculating
major parameters of metabolism. With such approach, the
primary metabolomic data could be used for calculating
metabolic parameters, similar to the deciphering of huge
number of genes and proteins using the genomic and
proteomic data.
The equilibration of fluxes leading to stable non-equilibrium is established in seconds and results in the exchange of
metabolic material between catabolic and anabolic pathways.
In this process, the established near-equilibria manifest
themselves as stoichiometric linkage between different pools
of metabolites so that they cease to be independent
variables.(13) The dynamic equilibrium reached via thermodynamic buffering results in the appearance of sustained
oscillations in the systems.(14,15) The resulting ordered
complexity of biological systems out-performs the simplicity
of physical or chemical systems.(16)
Here, we analyze how the set of nucleotides is equilibrated
in the cell and how this generates simple rules helping the cell
to organize itself via maintenance of a stable non-equilibrium
state. This is an important condition for self-maintaining
metabolism, which operates in a homeodynamic rather than
homeostatic mode.(17,18) The stable non-equilibrium state is
supported through thermodynamic buffering, in which the
ratios of adenylates and pyridine nucleotides follow thermodynamically and kinetically defined computable principles.
This allows determination of the rates and regulation of
metabolic pathways from the basic nucleotide ratios governing energy metabolism.
1091
Problems and paradigms
Buffering of intermediates
The buffering of energy intermediates is the most widespread
source of free energy in biological systems.(19) Filling buffer
reservoirs corresponds to the accumulation of free energy.
The most striking example is the bicarbonate buffering via
carbonic anhydrase. This plays a significant role in many
physiological processes from carbon fixation in photosynthesis to respiration in animals.(20) This buffer operates
effectively at a pH close to neutral, as does phosphate buffer.
The role of phosphate in biological systems is of particular
importance, as the internal instability of the pyrophosphate
(PPi) bond was used in early evolution for energy purposes.
This type of bond became the basis of energy-rich
compounds such as ATP and other nucleoside triphosphates.
The breakage of terminal phosphoanhydride bond results in a
local acidification, and protons stored in phosphate buffer are
released for catalysis (enzymes) or mechanical movement
(actin).(21) This reaction can then be coupled with thermodynamically unfavorable reactions to give an overall negative
(spontaneous) DG for the reaction sequence. The actual
value of DG for ATP hydrolysis varies, primarily depending
on Mg2þ concentration, and under normal physiological
conditions is close to –50 kJ/mol.(22–25)
By using ATP (or even PPi) in enzymatic reactions or
cytoskeleton movement, the release of energy from phosphate buffer becomes vectorized.(19) Vectorization is
achieved by the use of ATP for local changes in pH during
enzymatic reactions through the link between ATP hydrolysis
(accompanied by the release of proton) and an endergonic
metabolic process. Therefore, the splitting of energy-rich
bonds leads to local (in time and space) pH changes
accompanying hydrolysis of these bonds. This results in
driving of the conformational changes in protein molecules
and in the maintenance of mechanical macromolecular
action.
While simple inorganic buffers store energy without
vectorization, the adenylate system itself needs buffering to
make vectorization most efficient. This buffering is not
possible with the PPi system, but the adenylate system
includes three acting species, ATP, ADP, and AMP. This allows
using the most dephosphorylated species (AMP) for fine
regulation of the ATP consumption/production and for the
maintenance of ADP concentration at the level sufficient to
support ATP synthesis by ATP synthase.(26) The major
process of ATP synthesis (oxidative phosphorylation) has a
fluctuating load conductance, so it needs a maximal buffering
of the ADP supply from the intermembrane space of
mitochondria. It was experimentally shown that the oxidative
phosphorylation obeys linear and symmetric relations
between flows and forces, reminiscent of the classical
Onsager relations.(27,28) It can operate at optimal efficiency
only if the conductance of the load, i.e., the ATP utilizing
1092
A. U. Igamberdiev and L. A. Kleczkowski
reactions in a living cell, is exactly matched by the output
conductance of oxidative phosphorylation.(11,29) Because the
load conductance in a living cell is rarely constant but
fluctuates dramatically, oxidative phosphorylation could
hardly ever operate at optimal efficiency due to a violation
of conductance matching.(30)
A way out of this dilemma is in the introduction of reversible
ATP-utilizing reactions, such as the ones catalyzed by
adenylate kinase (AK) or creatine kinase (CK), which have
been proposed to act as thermodynamic buffers.(31) Thermodynamic buffer enzymes represent a bioenergetic regulatory
principle for the maintenance of a regime that is far from
equilibrium. The intermembrane space of mitochondria
(containing a high activity of AK and, in animals, of CK) is
a major compartment for such buffering, acting as a filter
where the adenylate concentration is adjusted to a correct
value before being handed over by adenylate translocator. It
acts as a linear energy converter maintaining the linearity of
oxidative phosphorylation within a physiological range. Due to
thermodynamic buffer enzymes, linear non-equilibrium thermodynamics becomes applicable to biological systems, and
metabolic fluxes can be derived from the concentrations of
nucleoside phosphates and metal ions established under
such equilibrium.
Computability of the metabolic system
due to thermodynamic buffering
Thermodynamic buffering of adenylates
Within the framework of the theory of dissipative structures,(9)
thermodynamic buffering(31) represents the basic regulatory
principle for the maintenance of a stable regime, i.e., far from
equilibrium where the production of energy is minimal. Due to
ATP production in oxidative phosphorylation, the phosphate
potential is shifted far away from equilibrium, and, since ATP
drives cell metabolism, this shifts all other potentials into a farfrom-equilibrium regime.
The load of metabolism by ATP can fluctuate depending on
the state of the cell and its environment. To satisfy the
condition of stable non-equilibrium, it should be optimized by
the efficiency of ATP production fluctuating in concert with
fluctuations in ATP consumption. Deviations from optimal
efficiency can be diminished by the action of enzymes that
generate sensitive thermodynamic buffering.(31–34) Every
reversible ATP utilizing reaction can act as a thermodynamic
buffer, e.g., the AK reaction reversibly utilizes MgATP in the
reaction with free AMP and produces one MgADP and one
free ADP.(7) This type of buffering effectively equilibrates the
pool of adenylates.
Such an important function as thermodynamic buffering
should have its own compartments associated with it. The
BioEssays 31:1091–1099, ß 2009 Wiley Periodicals, Inc.
A. U. Igamberdiev and L. A. Kleczkowski
compartment(s) should be in close proximity with the
locations of ATP synthesis and consumption. This is satisfied
for intermembrane space of mitochondria and (in plants only)
for intermembrane spaces of chloroplasts. In fact, the
mitochondrial compartment exhibits extremely high activity
of AK, e.g., it is two to four times higher than the activity of ATP
synthase in plant mitochondria,(35) and the activities of
associated enzymes such as nucleoside diphosphate
(NDP) kinase and (for animals) a major part of CK activity
is also associated with the mitochondrial intermembrane
space. Indeed, one can consider thermodynamic buffering as
the major function for this compartment.
Problems and paradigms
ratios of magnesium-bound and unbound adenylates that
govern various aspects of cell energetics. For instance, the
MgATP/MgADP ratio (essential for the rates of energy
metabolism), ATPfree/ADPfree ratio (important for the translocation of adenylates) and ADPfree/AMPfree ratio (a parameter
in allosteric regulation) could all be derived from total
adenylate contents.(38) The AK equilibrium generates the
computational property of the adenylates-magnesium set. A
similar computational set can be established for guanylates,
cytidylates, and uridylates via corresponding buffering
equilibria.
Creatine kinase
Adenylate kinase
The AK activity generates simple linear relations between the
equilibrated adenylate species,(7,36,37) so the system itself
generates computation. This means that the adenylate
metabolome is organized computationally due to the ‘‘stable
non-equilibrium’’ principle maintained via thermodynamic
buffering of opposite (e.g., catabolic and anabolic) nonequilibrium fluxes. The AK is therefore a ‘‘measuring device’’
in the biological system that makes it computable.
Figure 1 shows how the mechanism of equilibration of
fluxes via AK works. ATP in the cell provides the energy for
anabolic reactions (load of cell biosynthesis) via two major
mechanisms, one including hydrolysis to ADP (load 1) with
corresponding phosphorylation of a substrate or Pi release,
and another to AMP (load 2) with pyrophosphorylation or
release of PPi. ATP is resynthesized via phosphorylation from
ADP. These fluxes are controlled by AK in such a way that,
upon reaching equilibrium, all concentrations of adenylates
(free and magnesium-bound) plus free magnesium become
computationally linked. This results in the establishment of
CK equilibrium is coupled to AK equilibrium in certain
biological systems, e.g., in muscle.(39) The large pool size
of creatine and creatine phosphate in muscle represents an
additional thermodynamic buffering system. Its high capacity
is beneficial for muscle where the fluctuations of the extrinsic
load conductance are bigger than, for instance, in liver. This
results in specific conditions for adenylate equilibration
essential for special needs, e.g., active muscle contraction,
where concentrations of adenylates, Mg2þ and Ca2þ are
strongly controlled. The link between the AK and CK equilibria
results in high MgATP/MgADP ratios and also provides
optimal Ca2þ and Mg2þ concentrations for muscle operation.(40) The resulting ATP/ADP ratios in the intermembrane
space of animal mitochondria are up to 20 and thus
mitochondria operate closer to state 4.(41)
Disruption of AK and CK activities(42) results in severe
reductions of the cellular capability to maintain total ATP
turnover under muscle functional load. This indicates that AK
and CK are major points in the coordinated network of
complementary enzymatic pathways that serve in the
maintenance of energetic homeostasis and physiological
efficiency.(43) CK and AK operate as interrelated cellular high
energy phosphoryl transfer systems from mitochondria to the
cytosol(44): each creatine phosphate molecule produced
undergoes about 50 unidirectional CK-catalyzed phosphotransfers in transit to an ATP consumption site in the intact
muscles. CK can be considered as a functional analog of
arginine kinase, the enzyme that fulfils thermodynamic
buffering role in lower animals.(45) In these species, the
arginine pool (but not creatine) is linked to adenylates.
Essential components of the adenylate system in
driving metabolism
Figure 1. Scheme of the operation of the adenylate system in
metabolism. Energy consuming metabolic reactions are fed by ATP
hydrolysis to ADP (load 1) or AMP (load 2). AK buffer equilibrates
three adenylate species (Mg2þ feedback is not shown). ATP is
synthesized by phosphorylation, while AMP and ADP can also be
produced by apyrase and other reactions.
BioEssays 31:1091–1099, ß 2009 Wiley Periodicals, Inc.
The adenylate system in driving metabolism consists of a
directing component (ATP and ATP synthase) and energysensing component (AMP and AMP kinase) so it has both
energetic and informational constituents. The energy-sensing
mechanism maintains the balance between ATP production
1093
Problems and paradigms
and consumption in all eukaryotic cells. Without considering
all factors (including free magnesium), the energy-sensing
ratio (ATP/AMP) changes as a square of metabolic driving
ratio (ATP/ADP).(26) In reality, the dependence is more
complicated and based on the established concentration of
free magnesium,(8) and the directing component is in fact the
MgATP/MgADP ratio while the sensing component is the
MgATP/AMPfree ratio.
In a more detailed view, the adenylate system has also an
activating or milieu component (free magnesium), i.e.,
essential to provide optimum condition for the activity of
major anabolic and catabolic enzymes. Other ion concentrations (potassium, manganese) can be established or
adjusted under AK equilibrium.(36) The system also includes
a translocation component (ATPfree/ADPfree ratio) determining the exchange of adenylates across membranes. Thus, in
approximation, the adenylate system can be represented as
a matrix of the four components: driving, sensing, activating,
and translocating. Adenylate species not included in these
ratios can be counted in some specific cases, e.g., the
MgATP-dependent phosphofructokinase is allosterically
regulated by all free adenylates, with MgATP acting as
an inhibitor of fructose 6-phosphate binding to the
enzyme.(46,47)
The knowledge of driving forces for metabolic fluxes is
essential in metabolomic research. It is easy to get the data of
total amounts of adenylates, i.e., ATP, ADP, and AMP. As they
are not distributed evenly in the cell, it is more valuable to
have the data on their subcellular amounts. From the total
amounts of adenylates one can get information on whether
they comply with AK equilibrium, i.e., whether the value of
[ATP][AMP]/([ADP])2 is within the range of apparent AK
constant.(7,8) Reliable data can be obtained for erythrocytes,(48) but also for chloroplast stroma and for intermembrane spaces of plant mitochondria and chloroplasts. The
mitochondrial AK also effectively equilibrates adenylates in
the cytosol of green leaves, but there is apparently no AK
equilibrium in the cytosol of some heterotrophic tissues, e.g.,
roots.(7)
Thus, provided that thermodynamic equilibria are established in the cell (which is detected via correspondence of the
values of the substrate ratios to established constants of
thermodynamic buffering enzymes), one can argue that major
metabolic parameters are computationally interrelated and
their balance is almost as firm as sequences of genes and
proteins. Under AK equilibrium, ratios of free and Mg-bound
adenylates and concentration of Mg2þ are established. This
equilibrium is also linked to the ratios of guanylates,
cytidylates, and uridylates and to concentrations of other
metal ions. Specific equations for the determination of these
parameters have to be established, but it is evident that, in the
case of metabolomic research, we have a solid computational
pattern that can be firmly analyzed.
1094
A. U. Igamberdiev and L. A. Kleczkowski
Adenylate system governs organellar membrane
potentials
The ratios of adenylates are linked to membrane potential
values by the set of formulae described in ref.(8) The measure
of subcellular adenylate concentrations (AK-equilibrated)
provides an effective tool for the estimation of in vivo
mitochondrial and chloroplastic membrane potentials, which
would be impossible or complicated by other methods. High
respiratory rates generate membrane potential across the
mitochondrial membrane. The membrane potential at
established pH values inside and outside of mitochondria
and chloroplasts is tightly linked to the ATP/ADP ratios inside
and outside the organelles. The values of membrane
potentials are tightly linked to total subcellular concentrations
of adenylates, as well as internal [Mg2þ] values and ratios of
free and/or Mg-bound adenylates.(8) The maintenance of
electric potential is one of the main functions of membranes
achieved by the balance of both non-equilibrium processes of
proton pumping, ATP synthesis and ion transport, and
equilibrium action of the AK enzyme.
Other nucleotide systems governing metabolism
Ratios of guanylates, uridylates, and cytidylates are responsible for driving metabolism of proteins, carbohydrates, and
lipids, respectively.(3) They are directly linked to ATP availability
since the essential phosphorylase/kinase reactions responsible for the production of the other triphosphate nucleotides in
the cell transfer one phosphate from ATP to another nucleotide
such as GDP, UDP, or CDP to create ADP and either GTP, UTP,
or CTP, respectively, by NDP kinases.(49) Because the energies
of the triphosphate bonds being created and destroyed in these
reactions are essentially identical, these reactions are driven
entirely by the concentration gradient between the reactants
and products. Driving the ATP/ADP ratio to very low levels, also
drives the ratios GTP/GDP, UTP/UDP, CTP/CDP, and nearly
every other triphosphate/diphosphate pair in the cell, to even
lower levels to maintain a favorable concentration gradient.(3)
GTP is important for protein synthesis, UTP for carbohydrate
synthesis and CTP for lipid synthesis, so these distinct groups
of metabolites are ‘‘marked’’ by corresponding nucleotides.
Buffering enzymes for the nonadenylic nucleotides are,
correspondingly, guanylate, uridylate, and cytidylate kinases,
all belonging to NMP kinase family.(50,51) Some of these
enzymes have mixed substrate specificity, and it is likely that
the nucleotide species are well equilibrated; AK itself is
frequently nonspecific for nucleotides (e.g., corn AK).(37) On
the other hand, the operation of guanylate, uridylate, and
cytidylate systems depends on the adenylate system,
because only ATP is synthesized by a direct mechanism
linked to membrane potential. The enzymes that link ATP to
other nucleotide systems are NDP kinases that catalyze
BioEssays 31:1091–1099, ß 2009 Wiley Periodicals, Inc.
A. U. Igamberdiev and L. A. Kleczkowski
Problems and paradigms
active phosphorylation of GDP, UDP, CDP, and IDP.(49,52)
These nucleoside triphosphates are not formed by the
oxidative phosphorylation.
Thermodynamic buffering of pyridine
nucleotide system and regulated
uncoupling
Pyridine nucleotide system
The pyridine nucleotide system is the most important,
together with the adenylate system, in governing metabolism.
It is composed of reduced and oxidized pyridine nucleotides:
one (NAD) is used mainly in catabolic processes and its
reduced form (NADH) supplies electrons for ATP synthesis,
while another (NADP) used is mainly in anabolic processes.
Its direct participation in ATP synthesis is limited to certain
groups of organisms, e.g., in plants, where NADPH can serve
as an electron donor for the mitochondrial electron transport
chain (ETC) in the case of a high reduction level.(53) Some
dehydrogenases use NADH or NADPH in a stereo a form,
while others use the b form. Although it is not possible to
distinguish the two forms as they are interconvertible, this
different usage may make sense in metabolic channeling
when nucleotide from the previous enzyme is supplied in
corresponding orientation. Also, different dehydrogenases
use NADH or NADPH species with different degrees of
protonation, e.g., external mitochondrial NADH and NADPH
dehydrogenases use protonated forms, which are related to
low pH optimum of their operation.
NAD (nicotinamide-ribose-ADP) and NADP (nicotinamideribose-phosphoADP) are both the derivatives of adenylates
carrying a residue that can be in a reduced or oxidized form.
We do not consider here covalently bound cofactors such as
FAD, FMN, or pyridoxal phosphate that cannot be equilibrated
freely by corresponding thermodynamic buffering systems.
Equilibration of pyridine nucleotides is achieved via transhydrogenase (Fig. 2). The proton-pumping transhydrogenase
equilibrates pyridine nucleotides together with the proton
gradient and hence membrane potential.(54) Plant mitochondria lack a proton-pumping transhydrogenase.(55) The nonenergy-linked transhydrogenase reactions are generally
important for NADH generation in conditions of high NADPH
concentration, like in photosynthetic cells, while the energylinked transhydrogenase is important for NADPH formation
using reducing power of NADH and proton motive force, which
is important for animal cells.(56)
Transhydrogenase equilibrium is an important regulatory
principle in metabolic systems. It regulates the flux through
the tricarboxylic acid cycle via regulation of NAD- and NADPdependent isocitrate dehydrogenases, one of which is
irreversible (NAD) and another reversible (NADP).(57) A
BioEssays 31:1091–1099, ß 2009 Wiley Periodicals, Inc.
Figure 2. Scheme of the operation of the pyridine nucleotide system in metabolism. Catabolism is fed by NADH while anabolism is
driven by NADPH. Oxidized pyridine nucleotides are reduced by
corresponding enzymes and anabolic and catabolic nucleotide pairs
are equilibrated by transhydrogenase, TH, which is linked to proton
pumping in bacteria and animals but not in plants.
similar mechanism, but in the absence of proton-translocating
transhydrogenase, may also function in plants, which
equilibrate pyridine nucleotides via a different mechanism
involving other transhydrogenase-like reactions.(55,58) When
the concentration of NADH rises, it transhydrogenates with
NADPþ forming NADPH. The consequence of this will be the
activation of additional oxidation flow via the internal NADPH
dehydrogenase of the ETC. Its capacity [up to 15% of the total
capacity for NAD(P)H oxidation] can provide an additional
power to increase flux through the ETC. The rise in NADPH
also contributes to the activation of the so-called (cyanideinsensitive) alternative oxidase (AOX),(59) so the total flux
through the ETC can increase even much more. It also
stimulates the reverse reaction of NADP-dependent isocitrate
dehydrogenase, leading to citrate efflux from mitochondria
and to the activation of AOX gene.(60)
Regulated uncoupling in the pyridine nucleotide
system in mitochondria
The irreversible exergonic reaction that is not coupled to ATP
synthesis can be tightly regulated by shifts in balance
between the reactions of load and consumption. It can be
called ‘‘regulated uncoupling’’ and, in general, occurs if one of
the two coupled reactions of a cyclic process proceeds without
its counterpart(12); e.g., an enzyme passes a proton without
ATP synthesis, e.g., the AOX in the mitochondrial ETC in
plants.(61) Although this slippage decreases the efficiency of
energy utilization, it enables controlling and regulating
metabolic demands.(12) Another example is the activation
of an uncoupling protein under certain conditions or regulated
uncoupling caused by long-chain fatty acids.(62) This type of
thermodynamic buffering equilibrates the ‘‘energy charge’’ of
the ATP pool with the redox charge of pyridine nucleotides. It
1095
Problems and paradigms
A. U. Igamberdiev and L. A. Kleczkowski
allows uncoupling of a high redox charge from the imminent
elevation of energy charge. As a result, metabolic processes
may be less affected by high values of redox charge.
Figure 3 shows the regulation of ATP synthesis (and hence
of the value of energy charge after its equilibration through
thermodynamic buffering) achieved via regulated uncoupling.
This uncoupling can be achieved via different mechanisms,
e.g., in plant mitochondria such a mechanism includes the
AOX, rotenone-insensitive NAD(P)H dehydrogenases, and
the uncoupling proteins. In animal mitochondria, although
non-coupled dehydrogenases, and oxidases are absent,
different types of uncoupling proteins, including those that
may be involved in Ca2þ release,(63) regulate the level of
coupling between NADH oxidation and ATP synthesis.
Regulated uncoupling represents a link between redox
charge and energy charge, and serves to establish the
energy charge at a certain level prior to buffering. The form of
metabolism is based on the fact that concentrations of the
metabolites NADH, NADþ, and ADP are all under strict
regulatory control. One purpose for maintaining the concentrations of these metabolites near the center of the
thermodynamically feasible range could be to maximize
flexibility.(3)
The thermodynamic equilibration of pyridine nucleotides
results in their mass action ratio (NADPH/NADPþ)/(NADH/
NADþ) of 10 or less in the absence of proton-pumping
transhydrogenase, as in some bacteria depending on oxygen
supply(56,64) and in plants.(58) For animal cells, this ratio has
much higher values (from 100 to 500 depending on
membrane potential).(65) Different species of pyridine nucleotides can participate in particular metabolic processes, e.g.,
external NADH and NADPH dehydrogenases in plant
mitochondria use protonated forms of reduced pyridine
nucleotides as substrates.(66) However, further details of
thermodynamic equilibration of pyridine nucleotides and a
possibility of local pH values as a feedback of this equilibrium
need to be clarified.(67)
The regulated uncoupling of NADH oxidation from ATP
synthesis results in a buffered maintenance of the ATP/ADP
and NADH/NADþ ratios under different physiological conditions. The ratio of ATP/ADP decreases with the decrease in
oxygen supply, while NADH/NADþ increases. The balance
between the catabolic and anabolic reduction charges is
established at different levels depending on the state of the
cell, e.g., in higher plants it depends on developmental stage
(floral induction, flowering) and light/dark conditions.(68)
Figure 3. Link between pyridine nucleotide and adenylate systems.
The balance between NADH/NAD and ATP/ADP ratios is supported
via switching to non-coupled pathways of NADH oxidation triggered at
high reduction level.
Figure 4. Generalized scheme of the peroxisomal metabolism,
representing the mechanism of dissipation of high NAD(P)H/NAD(P)þ
ratio, thus unlinking the transformation of the substance A to B from
redox control.
1096
Peroxisomes in regulated uncoupling
Whereas in mitochondria NADH oxidation can be unlinked
with ATP synthesis, peroxisome may provide uncoupling at
another level, i.e., when substrate oxidation is not coupled to
NAD reduction. Therefore, peroxisome can be considered as
a ‘‘regulated leak compartment’’.(69) Peroxisomes carry out
reactions that may lower reduction charge, e.g., fatty acid
oxidation in peroxisomes is not coupled to NADþ reduction as
in mitochondria, and in higher plants glycolate oxidation
occurs via an NAD-independent mechanism (contrary to that
existing in algae containing glycolate dehydrogenase).(69) In
this regard, peroxisomes can be considered as redox
buffering organelles representing metabolic steps regulating
NADH/NADþ ratio.
Usually, peroxisomes contain metabolic cycles in which one
(reductive) half is linked to the consumption of NADH and
another (oxidative) dissipates energy in the reaction of
hydrogen peroxide formation (Fig. 4). Since several oxidative
metabolic processes can be linked or non-linked to NADH
generation (such as b-oxidation of fatty acids in mitochondria
versus in peroxisomes), this may provide a fine metabolic
control of NAD reduction.(69) Plant photorespiration, a wasteful
process, attributed to peroxisomes in several key points, is
important for controlling the redox level in photosynthesizing
cells.(70) Some metabolites linked to reduction by NADH and
NADPH can be converted outside of the peroxisome,(71) while
those that undergo flavin-dependent oxidation are exclusively
located in the peroxisome.
BioEssays 31:1091–1099, ß 2009 Wiley Periodicals, Inc.
A. U. Igamberdiev and L. A. Kleczkowski
Pyrophosphate system
PPi can be used as a macroergic molecule; however, the
system of PPi-based energy transfer lacks the flexibility of the
adenylate system, since it is not capable of thermodynamic
buffering. This is because the PPi system cannot directly
regulate the load and consumption of PPi through a buffered
feedback compound such as the AMP in the adenylate
system. PPi energy transfer may be important under
conditions of ATP deficiency, e.g., during anoxia in plants
or under uncoupling conditions when the ATP production is
suppressed and low ATP level is not sufficient to fulfill energy
demands of the cell. Lowering of ATP in the cell results in the
increase in Mg2þ due to the shift in the AK equilibrium.(7) PPidependent enzymes such as PPi-dependent phosphofructokinase and pyruvate-Pi-dikinase thus become more active
upon elevation of Mg2þ concentration, and in hypoxia-tolerant
plants their expression is strongly increased.(72) The MgPPidependent phosphofructokinase needs higher [Mg2þ] than
the MgATP-dependent phosphofructokinase because of the
much lower affinity that Mg2þ has for complexation with PPi
than with ATP (KMgPPi ¼ 1.2 mM1 compared with
KMgATP ¼ 73 mM1).(73)
We do not have information on the regulatory role of free
PPi, so the system may lack a feedback in the regulation, such
as that demonstrated for the adenylate system. However, PPi
has a more stable pool at lower pH values characteristic for
low oxygen conditions because the enzyme splitting it
(inorganic pyrophosphatase) is less active at pH below 7
and is inhibited by Ca2þ accumulating under hypoxia.(74) This
may make PPi an important autonomous salvage energy
molecule when ATP production goes down. In plants, the
vacuolar proton pumping can be also PPi dependent, and its
role is increased upon elevation of Mg2þ concentration.(75)
While the ratios of nucleotides such as ATP/ADP and UTP/
UDP change in a similar manner, PPi concentration changes
independently, highlighting its importance as an autonomous
energy donor.(76)
Problems and paradigms
which is equilibrated by thermodynamic buffering enzymes.
There are several links between energy charge and RNA
processing, including the supply of nucleoside monophosphates released during RNA cleavage for buffering with NTP.
It is known, e.g., that NDPK cleaves DNA and RNA.(77) In
addition, AMP kinase having key functions in regulation of
energy metabolism also regulates transcription.(78) The level
of ATP directly affects adenylation level of messenger RNA
and hence its stability.(79)
The generalized view on metabolism should, ideally,
include the metabolic reproduction (via transcription and
translation) of catalysts that govern metabolism. In fact,
synthesizing and maintaining the catalysts is necessary for
sustained operation of metabolism and this is a part of the
metabolic network. This is based on the assumption that
proteins themselves are products of metabolism, and thus
metabolites.(80–82) The function of ATP, GTP, UTP, and CTP is
two-fold: these nucleoside triphosphates are used both in
general metabolism as free nucleotides and in nucleic acid
metabolism via covalent polymerization to RNA molecules.
Changes in their pools directly affect both energy metabolism
and RNA turnover. Nucleotides serve as metabolites (free
coenzymes) of enzymes and their association into nucleic
acids generates matrices for the reproduction of enzymes
themselves.(83) The main challenge here is the difficulty of
formal joint representation of energetic and information fluxes
in the frames of a whole system. It is important to mention that
the operation of the two systems (the genetic and the
metabolic) is based on simple computable rules that are
generated within these systems. For the metabolic system,
certain ratios between adenylates are established via
thermodynamic buffering of enzymes such as AK, NDP
kinases, and others. Deviations from such equilibrium will
result in an unstable state that tends to return to the initial
stable non-equilibrium state.
Acknowledgments: This work was supported by the Natural
Sciences and Engineering Research Council of Canada (to
A. U. I.) and by the Swedish Research Council (to L. A. K.)
Future challenges
We have argued in this paper, that the stable non-equilibrium
state is generated by steady fluxes established via thermodynamic buffering and regulated uncoupling. Under these
conditions, fluxes of energy and information in the system
operate in a constant steady-state way and are organized
spatially and temporally. The thermodynamic buffer theory of
Stucki(11,31) suggested a general explanation of steady
operation of oxidative phosphorylation. This theory can be
extended to include all major parameters related to cell
metabolism. Both energy charge and genetic information use
the components of the same pool (nucleoside triphosphates),
BioEssays 31:1091–1099, ß 2009 Wiley Periodicals, Inc.
References
1. Caspi, R., Foerster, H., Fulcher, C. A., Hopkinson, R., Ingraham, J.,
et al. MetaCyc: a multiorganism database of metabolic pathways and
enzymes. Nucleic Acids Res 2006. 34: D511–D516.
2. Atkinson, D. E., The energy charge of the adenylate pool as a regulatory
parameter. Interaction with feedback modifiers. Biochemistry 1968. 7:
4030–4034.
3. Henry, C. S., Broadbelt, L. J. and Hatzimanikatis, V., Thermodynamicsbased metabolic flux analysis. Biophys J 2007. 92: 1792–1805.
4. Andersen, K. B. and Meyenburg, K. V., Charges of nicotinamide adenine
nucleotides and adenylate energy charge as regulatory parameters of the
metabolism in Escherichia coli. J Biol Chem 1977. 252: 4151–4156.
1097
Problems and paradigms
5. Russell, J. B. and Cook, G. M., Energetics of bacterial growth – balance
of anabolic and catabolic reactions. Microbiol Rev 1995. 59: 48–62.
6. Fromm, H. J., Control by energy charge is an untenable theory. Trends
Biochem Sci 1977. 2: N198–N200.
7. Igamberdiev, A. U. and Kleczkowski, L. A., Implications of adenylate
kinase-governed equilibrium of adenylates on contents of free magnesium
in plant cells and compartments. Biochem J 2001. 360: 225–231.
8. Igamberdiev, A. U. and Kleczkowski, L. A., Membrane potential, adenylate levels and Mg2R are interconnected via adenylate kinase equilibrium in plant cells. Biochim Biophys Acta 2003. 1607: 111–119.
9. Prigogine, I., Thermodynamics of irreversible processes. New York,
Wiley, 1967.
10. Bauer, E. S., Theoretical biology. Moscow, VIEM, 1935.
11. Stucki, J. W., The thermodynamic-buffer enzymes. Eur J Biochem 1980.
109: 257–267.
12. Tomashek, J. J. and Brusilow, W. S., Stoichiometry of energy coupling
by proton-translocating ATPases: a history of variability. J Bioenerg Biomembr 2000. 32: 493–500.
13. Reich, J. G. and Sel’kov, E. E., Time hierarchy, equilibrium and nonequilibrium in metabolic systems. Biosystems 1975. 7: 39–50.
14. Chance, B., Williamson, G., Lee, I. Y., Mela, L., DeVault, D., et al.
Synchronization phenomena in oscillations of yeast cells and isolated
mitochondria. New York, Academic Press, 1973.
15. Aon, M. A., Roussel, M. R., Cortassa, S., O’Rourke, B., Murray, D. B.,
et al. The scale-free dynamics of eukaryotic cells. PLoS One 2008. 3:
e3624.
16. Lloyd, D., Oscillations, synchrony and deterministic chaos. Prog Bot
2009. 70: 69–91.
17. Lloyd, D., Aon, M. A. and Cortassa, S., Why homeodynamics, not
homeostasis? Scientific World J 2001. 1: 133–145.
18. Lloyd, D. and Murray, D. B., Redox rhythmicity: clocks at the core of
temporal coherence. BioEssays 2007. 29: 465–473.
19. Shnoll, S. E., Physico-chemical factors of the biological evolution. Moscow, Nauka, 1979.
20. Henry, R. P., Multiple roles of carbonic anhydrase in cellular transport and
metabolism. Annu Rev Physiol 1996. 58: 523–538.
21. Khan, L. A. and Amin, M., A kinetic scheme for the early phase of ATP
hydrolysis by actomyosin ATPase and its bioenergetic implications. Indian
J Biochem Biophys 1995. 32: 37–43.
22. Alberty, R. A., Standard Gibbs free energy, enthalpy, and entropy
changes as a function of pH and Mg for several reactions involving
adenosine phosphates. J Biol Chem 1969. 244: 3290–3302.
23. Alberty, R. A., Thermodynamics of the hydrolysis of adenosine triphosphate as a function of temperature, pH, Mg, and ionic strength. J Phys
Chem B 2003. 107: 12324–12330.
24. Shikama, K., Standard free energy maps for the hydrolysis of ATP as a
function of pH, Mg and Ca. Arch Biochem Biophys 1971. 147: 311–317.
25. Tulub, A. A. and Stefanov, V. E., New horizons of adenosinetriphosphate
energetics arising from interaction with magnesium cofactor. Eur
Biophys J 2008. 37: 1309–1316.
26. Hardie, D. G. and Hawley, S. A., AMP-activated protein kinase: the
energy charge hypothesis revisited. BioEssays 2001. 23: 1112–1119.
27. Rottenberg, H., Non-equilibrium thermodynamics of energy-conversion
in bioenergetics. Biochim Biophys Acta 1979. 549: 225–253.
28. Lemasters, J. J. and Billica, W. H., Non-equilibrium thermodynamics of
oxidative phosphorylation by inverted inner membrane vesicles of rat liver
mitochondria. J Biol Chem 1981. 256: 2949–2957.
29. Kedem, O. and Caplan, S. R., Degree of coupling and its relation to
efficiency of energy conversion. Trans Faraday Soc 1965. 61: 1897.
30. Kadenbach, B., Intrinsic and extrinsic uncoupling of oxidative phosphorylation. Biochim Biophys Acta 2003. 1604: 77–7794.
31. Stucki, J. W., The optimal efficiency and the economic degrees of
coupling of oxidative phosphorylation. Eur J Biochem 1980. 109: 269–283.
32. Stucki, J. W., Lehmann, L. H. and Mani, P., Transient kinetics of
thermodynamic buffering. Biophys Chem 1984. 19: 131–145.
33. Veuthey, A. L. and Stucki, J., The adenylate kinase reaction acts as a
frequency filter towards fluctuations of ATP utilization in the cell. Biophys
Chem 1987. 26: 19–28.
34. Demirel, Y. A. and Sandler, S. I., Nonequilibrium thermodynamics in
engineering and science. J Phys Chem B 2004. 108: 31–43.
1098
A. U. Igamberdiev and L. A. Kleczkowski
35. Roberts, J. K. M., Aubert, S., Gout, E., Bligny, R. and Douce, R.,
Cooperation and competition between adenylate kinase, nucleoside
diphosphokinase, electron transport, and ATP synthase in plant mitochondria studied by 31P-nuclear magnetic resonance. Plant Physiol 1997. 113:
191–199.
36. Blair, J. M., Magnesium, potassium, and the adenylate kinase equilibrium. Magnesium as a feedback signal from the adenine nucleotide pool.
Eur J Biochem 1970. 13: 384–390.
37. Kleczkowski, L. A. and Randall, D. D., Maize leaf adenylate kinase:
purification and partial characterization. Plant Physiol 1986. 81: 1110–
1114.
38. Igamberdiev, A. U. and Kleczkowski, L. A., Equilibration of adenylates in
the mitochondrial intermembrane space maintains respiration and regulates cytosolic metabolism. J Exp Bot 2006. 57: 2133–2141.
39. Vincent, A. and Blair, J. M., The coupling of the adenylate kinase and
creatine kinase equilibria. Calculation of substrate and feedback signal
levels in muscle. FEBS Lett 1970. 7: 239–244.
40. Kemp, G. J., Manners, D. N., Clark, J. F., Bastin, M. E. and Radda, G. K.,
Theoretical modelling of some spatial and temporal aspects of the mitochondrion/creatine kinase/myofibril system in muscle. Mol Cell Biochem
1998. 184: 249–289.
41. Gellerich, F. N., The role of adenylate kinase in dynamic compartmentation of adenine nucleotides in the mitochondrial intermembrane space.
FEBS Lett 1992. 297: 55–58.
42. Janssen, E., Terzic, A., Wieringa, B. and Dzeja, P. P., Impaired intracellular energetic communication in muscles from creatine kinase and
adenylate kinase (M-CK/AK1) double knock-out mice. J Biol Chem 2003.
278: 30441–30449.
43. Dzeja, P. P., Zeleznikar, R. J. and Goldberg, N. D., Adenylate kinase:
kinetic behavior in intact cells indicates it is integral to multiple cellular
processes. Mol Cell Biochem 1998. 184: 169–182.
44. Dzeja, P. P., Zeleznikar, R. J. and Goldberg, N. D., Suppression of
creatine kinase-catalyzed phosphotransfer results in increased phosphoryl transfer by adenylate kinase in intact skeletal muscle. J Biol Chem
1996. 271: 12847–128451.
45. Teague, W. E., Jr. and Dobson, G. P., Thermodynamics of the arginine
kinase reaction. J Biol Chem 1999. 274: 22459–22463.
46. Kemp, R. G. and Krebs, E. G., Binding of metabolites by phosphofructokinase. Biochemistry 1967. 6: 423–434.
47. Li, Y. L., Rivera, D., Ru, W., Gunasekera, D. and Kemp, R. G., Identification of allosteric sites in rabbit phosphofructo-1-kinase. Biochemistry
1999. 38: 16407–16412.
48. Rose, I. A., The state of magnesium as estimated from the adenylate
kinase equilibrium. Proc Natl Acad Sci U S A 1968. 61: 1079–1086.
49. Hasunuma, K., Yabe, N., Yoshida, Y., Ogura, Y. and Hamada, T.,
Putative functions of nucleoside diphosphate kinase in plants and fungi.
J Bioenerg Biomembr 2003. 35: 57–65.
50. Fukami-Kobayashi, K., Nosaka, M., Nakazawa, A. and Go, M., Ancient
divergence of long and short isoforms of adenylate kinase: molecular
evolution of the nucleoside monophosphate kinase family. FEBS Lett
1996. 385: 214–220.
51. Alexandre, J. A. C., Roy, B., Topalis, D., Pochet, S., Périgaud, C. and
Deville-Bonne, D., Enantioselectivity of human AMP, dTMP and UMPCMP kinases. Nucleic Acids Res 2007. 35: 4895–4904.
52. Ray, N. B. and Mathews, C. K., Nucleoside diphosphokinase – A functional link between intermediary metabolism and nucleic acid synthesis.
Curr Top Cell Regul 1992. 33: 343–357.
53. Rasmusson, A. G., Geisler, D. A. and Møller, I. M., The multiplicity of
dehydrogenases in the electron transport chain of plant mitochondria.
Mitochondrion 2008. 8: 47–60.
54. Hatefi, Y. and Yamaguchi, M., Nicotinamide nucleotide transhydrogenase: a model for utilization of substrate binding energy for proton translocation. FASEB J 1996. 10: 444–452.
55. Bykova, N. V., Rasmusson, A. G., Igamberdiev, A. U., Gardeström, P.
and Møller, I. M., Two separate transhydrogenase activities are present in
plant mitochondria. Biochem Biophys Res Commun 1999. 265: 106–
111.
56. Voordouw, G., van der Vies, S. M. and Themmen, A. P. N., Why are two
different types of pyridine nucleotide transhydrogenase found in living
organisms? Eur J Biochem 1983. 131: 527–533.
BioEssays 31:1091–1099, ß 2009 Wiley Periodicals, Inc.
A. U. Igamberdiev and L. A. Kleczkowski
57. Sazanov, L. A. and Jackson, J. B., Proton-translocating transhydrogenase and NAD-linked and NADP-linked isocitrate dehydrogenases operate
in a substrate cycle which contributes to fine regulation of the tricarboxylic
acid cycle activity in mitochondria. FEBS Lett 1994. 344: 109–116.
58. Igamberdiev, A. U. and Gardeström, P., Regulation of NAD- and NADPdependent isocitrate dehydrogenases by reduction levels of pyridine
nucleotides in mitochondria and cytosol of pea leaves. Biochim Biophys
Acta 2003. 1606: 117–125.
59. Gray, G. R., Villarimo, A. R., Whitehead, C. L. and McIntosh, L.,
Transgenic tobacco (Nicotiana tabacum L.) plants with increased expression levels of mitochondrial NADPR-dependent isocitrate dehydrogenase:
evidence implicating this enzyme in the redox activation of the alternative
oxidase. Plant Cell Physiol 2004. 45: 1413–1425.
60. Vanlerberghe, G. C. and McIntosh, L., Coordinate regulation of cytochrome and alternative pathway respiration in tobacco. Plant Physiol
1992. 100: 1846–1851.
61. Lambers, H., Cyanide-resistant respiration: a non-phosphorylating electron transport pathway acting as an energy overflow. Physiol Plant 1982.
55: 478–485.
62. Schönfeld, P. and Wojtczak, L., Fatty acids decrease mitochondrial
generation of reactive oxygen species at the reverse electron transport
but increase it at the forward transport. Biochim Biophys Acta 2007. 1767:
1032–1040.
63. Trenker, M., Malli, R., Fertschai, I., Levak-Frank, S. and Graier, W. F.,
Uncoupling-proteins 2 and 3 are elementary for mitochondrial Ca2R uniport. Nat Cell Biol 2007. 9: 445–452.
64. Matin, A. and Gottschal, J. C., Influence of dilution rate on NAD(P) and
NAD(P)H concentrations and ratios in a Pseudomonas sp. grown in
continuous culture. J Gen Microbiol 1976. 94: 333–341.
65. Hoek, J. B. and Rydström, J., Physiological roles of nicotinamide nucleotide transhydrogenase. Biochem J 1988. 254: 1–10.
66. Edman, K., Ericson, I. and Møller, I. M., The regulation of exogenous
NAD(P)H oxidation in spinach (Spinacia oleracea) leaf mitochondria
by pH and cations. Biochem J 1985. 232: 471–477.
67. Noctor, G., De Paepe, R. and Foyer, C. H., Mitochondrial redox biology
and homeostasis in plants. Trends Plant Sci 2007. 12: 125–134.
68. Bonzon, M., Simon, P., Greppin, H. and Wagner, E., Pyridine nucleotides and redox-charge evolution during the induction of flowering in
spinach leaves. Planta 1983. 159: 254–260.
69. Igamberdiev, A. U. and Lea, P. J., The role of peroxisomes in the
integration of metabolism and evolution of land plants. Phytochemistry
2002. 60: 651–674.
BioEssays 31:1091–1099, ß 2009 Wiley Periodicals, Inc.
View publication stats
Problems and paradigms
70. Igamberdiev, A. U., Bykova, N. V., Lea, P. J. and Gardeström, P., The
role of photorespiration in redox and energy balance of photosynthetic
plant cells: a study with a barley mutant deficient in glycine decarboxylase. Physiol Plant 2001. 111: 427–438.
71. Timm, S., Nunes-Nesi, A., Pärnik, T., Morgenthal, K., Wienkoop, S.,
et al. A cytosolic pathway for the conversion of hydroxypyruvate to
glycerate during photorespiration in Arabidopsis. Plant Cell 2008. 20:
2848–2859.
72. Huang, S., Colmer, T. D. and Millar, A. H., Does anoxia tolerance involve
altering the energy currency towards PPi? Trends Plant Sci 2008. 13: 221–
227.
73. O’Sullivan, W. J. and Smithers, G. W., Stability constants for biologically
important metal-ligand complexes. Meth Enzymol 1979. 63:
294–336.
74. Hemalatha, K. P. and Prasad, D. S., Purification, physicochemical
properties, and subcellular location of alkaline inorganic pyrophosphatase from sesame (Sesamum indicum L.) cotyledons. Biochem Cell Biol
2002. 80: 215–224.
75. Maeshima, M. and Yoshida, S., Purification and properties of vacuolar
membrane proton-translocating inorganic pyrophosphatase from mung
bean. J Biol Chem 1989. 264: 20068–20073.
76. Dancer, J., Veith, R., Feil, R., Komor, E. and Stitt, M., Independent
changes of inorganic pyrophosphate and the ATP/ADP or UTP/UDP ratios
in plant cell suspension cultures. Plant Sci 1990. 66: 59–63.
77. Hammargren, J., Salinas, T., Maréchal-Drouard, L. and Knorpp, C.,
The pea mitochondrial nucleoside diphosphate kinase cleaves DNA and
RNA. FEBS Lett 2007. 581: 3507–3511.
78. McGee, S. L. and Hargreaves, M., AMPK and transcriptional regulation.
Front Biosci 2008. 13: 3022–3033.
79. Eads, B. D. and Hand, S. C., Mitochondrial mRNA stability and polyadenylation during anoxia-induced quiescence in the brine shrimp Artemia franciscana. J Exp Biol 2003. 206: 3681–3692.
80. Montero, F., Nuno, J. C., Melendez-Hevia, E., Olasagasti, F., Vazquez,
S. and Moran, F., Stoichiometric analysis of self-maintaining metabolisms. J Theor Biol 2008. 252: 427–432.
81. Rosen, R., Life itself. New York, Columbia University Press, 1991.
82. Cornish-Bowden, A., Cárdenas, M. L., Letelier, J. C. and SotoAndrade, J., Beyond reductionism: metabolic circularity as a guiding
vision for a real biology of systems. Proteomics 2007. 7: 839–845.
83. Igamberdiev, A. U., Foundations of metabolic organization: coherence as
a basis of computational properties in metabolic networks. BioSystems
1999. 50: 1–16.
1099