Towards Living Technology: Building an Artificial Cell
The emergence of cellular life is one of the major transitions
in evolution. It involves the emergence of a well-defined boundary
allowing metabolism and genetic information to be part of a well-defined
compartment. Theoretical models and experimental data support
the idea that simple protocells should be obtainable from a simple
systems of coupled reactions dealing with the three previous components.
The building of an artificial cell would be a fundamental breakthrough in
our understanding of life, its origins and evolution, not to mention a
wide array of potential medical and technological applications.
The integrated project
PACE will focus
on the IT potential of truly artificial cells: addressing both the technical opportunities of
programmable artificial cells and an evolutionary roadmap to
producing them under the control of current computers. Such artificial
cells will be useful because of their distinctness from, rather than
similarity to current biology. In this context, the fundamental principles
of current biology might not emerge at the very beginning of life: this is the case
for example of
allometric scaling and stoichiometric inbalances
Understanding the origins of life requires the understanding of a few key events that define
the so called major transitions in evolution. One of them is the emergence of cellular
structures. Cellularization allowed the emergence of separated compartments (the protocells)
able to evolve and maintain a well-defined integrity of all the components. Within the 6th
European Framework initiative, we are part of an interdisciplinary group of researchers that
will work with the final goal of building a very simple,
nano-scale artificial protocell able to self-replicate and
evolve under controlled conditions.
Our Lab (the COMPLEX SYSTEMS LAB ) will participate at different levels, particularly
from the theoretical perspective, modeling protocell dynamics and their possible evolution,
the evolution of protocellular networks, with specific emphasis in
exploring the role of noise on their behavior. We will also develop several
modelling tools allowing to explore the behavior of in silico models of
protocell and replicator dynamics. These cells will be much smaller
and simpler than modern cells and eventually will allow to create the foundations for information
processing in living nano-materials. They can be thought as nano-robots, performing at the
molecular scale. They will integrate three basic, microscopic biochemical systems: a
container (the lipid membrane), a construction system (the metabolism) and a genetic system
involving information-bearing molecules (peptide-DNA) that encode the critical molecular
processes and simultaneously participates catalytically in the metabolism. As such, these
artificial cells will embody extremely simple versions of biochemical networks, weighting
many orders of magnitude less than the smallest modern cells. Given the small numbers
of copies intrinsic to the nanoscale involved, it is important to understand the limits imposed by
small copy numbers to reliable replication, stability and computation.
The integrated project PACE will explore the utilization of the simplest
technically feasible elementary living units (artificial cells much simpler
than current cells) to build evolvable complex information systems. We will
create, analyse and investigate the applications of such systems that process
information by self-organization starting at molecular scales. The project will also
determine whether life-like properties are necessary for computational systems
to be fully robust and adaptive and investigate the tension between evolvable
living autonomy and programmable utilization. We will explore the collective
properties of artificial cells and demonstrate that they are the right material
for building nanoscale robot ecologies. The particular molecular systems we will
consider will have genetically controlled catalytic reactions, self-assembly
of complex supramolecular structures, and energy transduction. Experimental teams within PACE will
investigate the stepwise evolution of such complex systems by machine complementation
and combinatorial search using a programmable microfluidic interface. This work
will provide theoretical and simulation frameworks for understanding emergent
computational properties of such systems, and experimental frameworks for
programming them by evolutionary exploration of chemical reactions.
See the recent
feature article on PACE in
NEW SCIENTIST
12 February 2005: "ALIVE!" by Bob Holmes .
European Center for Living Technology
The Center will conduct a research, outreach and training program
designed to equip a new generation of scientists and engineers to
take advantage of programmable artificial cell evolution. The new
European Center for Living Technology is to be established in Venice,
supported by the European Union, the city of Venice, and the
University of Venice Ca Foscari.
Researchers will be invited to the
Center to collaborate on PACE and related multidisciplinary projects
involving living technology. The Center will also proactively foster
informed public discussion of the novel social, safety and ethical
issues raised by living technology.
Applicants with appropriate educational background should have an ongoing
institutional affiliation and plan to make a definite contribution to the
PACE project in collaboration with one or more partners during their visit to the Center.
NANOBIOLOGY AND NANOTECHNOLOGY SITES:
Richard Feynmann's visionary talk on Nanotechnology
Nanotechnology site by Ralph Merkle
USA Nanotechnology Initiative
Nanotech at the Foresight Institute
Scientific American Special issue on Nanotechnology
PACE Research Projects at the Complex Systems Lab
Vesicle self-reproduction through Turing-like Instabilities
Using dynamical instabilities inspired in Turing reaction-diffusion models,
we have developed a new class of models defining a RD mechanism coupled
with a flexible membrane container. Using this approach, we have been able
to show that self-replicating protocells including membrane and metabolism
can be obtained under physically and chemically reasonable conditions.
Within this project, we have developed a software tool named
OCCAM . It is
a toolset developed within PACE project in order to study
in silico the effects of non-uniform
osmotic pressure coupled with membrane growth. Membrane growth under
non-uniform osmotic pressure could be the basis for devoloping new
active mechanisms, which control protocell division cycles. Here we
present different metabolic scenarios which are able to create non-uniform
osmotic pressures and the simulation tools to study the membrane shape
evolution in these contexts.
Nonlinear network replicator dynamics
A first theoretical step when looking at replicator dynamics involves considering
nonlinear interactions with no explicit or implicit definition of cell membranes.
Models of this type deal with generic features exhibited by replicator dynamics, which
have been widely explored over the last decades. We are exploring a new class of replicator
dynamics models involving as particular cases hipercyclic organization
(see Hypercycles: an introduction)
. Both spatial
degrees of freedom and host-parasite dynamics are being analysed. By using an explicit
definition of genome complexity in terms of small strings of traits, it is possible to
follow the evolutionary dynamics of simple models of replicators and to explore how
the presence of parasites can either destroy the organization of cycles or instead
trigger the emergence of complexity. Since several possible genomes can be simultaneously
present, these models actually explore a network of interacting replicators. The models
will be extended from the standard deterministic behavior to explicit stochastic implementation.
We will extend our results to models where compartents
are allowed to be formed, so that sets of replicators can be eventually isolated from each other
and new types of interactions, now at the protocellular level, can emerge.
Noise and self-replicating spatial reaction-diffusion models
The most studied examples of the two types of reaction-diffusion systems are the
Meinhardt system (Gierer &Meinhardt 2000) and the diffusive Gray-Scott system
(Pearson 1993), respectively. Interesting enough, the replication characteristic
is a particularity of the diffusive Gray-Scott model alone, which makes
it the ideal model for developmental research - see more details. In such cases,
cell-like localized structures grow, deform and make replica of themselves until
they occupy the entire space , a theoretical result that also was confirmed
experimentally (Lee, McCormick, Swinney & Pearson, 1994). This model was also a
target for investigations resulting from the considerable interest of the scientific
community in the implications of stochasticity in the evolution of biological and
chemical systems. In particular, Lesmes et al. (2003) have carried out the first study
of the noise-controlled pattern formation in the Pearson model, with emphasis on the
self-replicating patterns. They found that for a specific set of parameters' values,
the noise drives the system from the non-multiplicative, stripe-like pattern to the
spot-multiplication one . Our results suggest the existence of an interval
of optimum noise-intensity values leading to a maximum number of spots, rather
than a single optimum value, as suggested by Lesmes and co-workers. Other models from the
literature consider hypercycle networks modeling RNA-like polymers catalysing the
replication of each other in a cyclic way (Cronhjort & Bloomberg, 1997). They
obtain the formation and division of clusters or "spots" and will be also explored.
Nonlinearity in minimal protocellular networks
Abstract models of inimal cells, consisting of a
collectively autocatalytic network of reactions enclosed within a membrane, are a first step
in modelling simple cellular replicators. An example of them is the so called
Ganti's chemoton.
The chemoton differs from the minimal
autopoiesis model suggested by
Varela and co-workers in explicitly including a genetic subsystem. It is also rather
more detailed in its analysis of the required chemical dynamics, and
aims at supporting self-reproduction by growth and fission even in the minimal version.
The chemoton has been presented mainly in papers and books by his creator, Tibor Ganti.
Most of this work has been done in terms of formalizing the connection between the different
components and how they relate each other. A few attempts to explore the real
dynamical behavior have also been conducted, providing evidence for a primitive form
of replicating cellular dynamics which exhibits some interesting homeostatic capabilities.
We will explore the a whole family of simple chemoton-like models of replicating
protocells both involving deterministic and stochastic dynamics. These models will provide
useful intuition for further exploration of models with explicit membrane dynamics. Since they contain
three coupled components (metabolism, membrane and information molecules) they are actually describable
in terms of networks of reactions. We will study their stability, time evolution and adaptation
to different external conditions.
Evolving lattice cellular aggregates
A previous step before a realistic model of realistic protocell dynamics involves
considering toy models of replicating systems forming aggregates in space. One possible
approach involves considering dynamics of membrane components and water as happening on a
discrete, regular lattice where each site can be occupied by a single molecule or part of it.
Interactions include a simplified energy function which takes into account physical interactions
through an Ising-like Hamiltonian. Molecular movements in such lattice model are allowed
provided that energy is (tipically) minimized. The rules include a Boltzmann probabilistic
update that explicitly takes into account the noise. Beyond the modelling of micelle
formation and the presence of phase transitions as non-covalent forces are changed, we will
also explore evolutionary dynamics of lipid aggregates. By using a given pool of
available lipid molecules and other precursors, it is possible to model the emergence
of compositional genomes and to observe the selection of whole replicating networks of
interacting molecules. The model will be completed by considering in- and out-flow of particles
thus mimicking microfluidic environments.
Protocell replication dynamics
Towards the modeling of more realistic protocells, consideration of more accurate
physical interactions is required. Some teams within PACE will work towards realistic
chemical simulation using molecular dynamics. Our own approach will be a shortcut between
simple toy models and full physical implementations. Our main goal is to provide a
platform that includes the basic implementation of a physically reasonable picture
of micelle formation coupled with evolving dynamics. Beyond the specific properties
displayed by the PACE protocell design, there is a whole range of possible scenarios
of protocell assembly that could be explored using this framework. We will consider
a set of minimal reaction networks coupled to membrane instability that will help understand
the role of molecular fluctuations in the evolution of protocells. Using different
types of molecular sets we should be able to characterize, at this level, the impact of
external and internal noise on protocell evolution and relevant types of evolution
towards stable, robust replicating entities. The emergence of parasites, simbiotic relations
and other forms of cell-cell interaction will be also explored.
Reliable computation in protocellular networks
Together with the construction, characterization and analysis of artificial protocells,
the PACE project intends to create a new research field of "Living Technology". As part
of this initiative, we will contribute by exploring in depth the problem of how computation
emerges in both protocells and protocell assemblies. Single protocells should evolve computations
allowing them to guarantee reliable replication and eventually to become independent
from external fluctuations. Additionally, collective computations might be possible
among sets of cells displaying simple forms of communication. Such communication might
be possible either through chemical signals or through artificial (computer) interfaces, such as those
to be developped within PACE. We want to extend early studies by von Neumann, Cowan and Vinograd
on reliable computation towards a rather unknown domain where global computations are performed
by collectives or loosely coupled protocells on a given spatial domain. By exploring such simple
models we hope to provide insight into the origins of multicellular systems and new views of
looking at computation at the nanoscale.
PARALLEL SUPERCOMPUTER MODELING OF LIVING PROTOCELLS
Within the PACE framework, we have started a collaboration with the
Barcelona Supercomputing Center,
which hosts Marenostrum (right picture),
the fifth Largest World's Supercomputer. By using the parallel computation
capacities of this system, we expect to model realistic, whole-protocell
simulations of self-organizing aggregates containing the three key components
of life (membrane, metabolism and information).
By means of Marenostrum
we'll be able to study the spectrum of parameter combinations and molecular
reactions that allow the transition from non-living to living matter, and
provide useful input to ongoing experiments. The large-scale simulation of
living protocells also allows a deep exploration of their computational
and evolutionary potential, thus providing a powerful framework for
future development within the emerging field of living technology.