Compartmentalization is an organization principle widely used in nature. On one hand, eukaryotic cells contain various vesicle-like sub-compartments (such as nucleus, endosomes, mitochondria, etc.). On the other hand, tissues can be understood as compositions of membrane-bound processing units (cells).
The reasons for this compartmentalization are manifold and partially caused by the fact that natural systems are the result of evolutionary processes and not of rational planning. Some aspects, however, are rooted in chemistry and physics and are also of importance in artificial systems: The implementation of a complex reaction network, for example, is difficult if one is only allowed to use one reaction container (e.g. one single vesicle). This for the following reasons:
- Individual steps of the reaction may interact with each other in an undesired manner.
- Many process steps only run efficiently under relatively specific conditions, say with respect to pH. These conditions may vary quite considerably for the different steps. Combining all steps in one reaction container implies to choose "one fits for all" conditions, which may lead to a significantly sub-optimal result.
Compartmentalization facilitates the de-coupling and optimization of different process steps. But there is a price to pay: The products of the individual steps have to be transported between the vesicles. Modern process design relies on the usage of multiple reaction vessels, which are connected such that they form a finely tuned process chain. It is exactly this process chain that we want to realize by the spatial structuring of the vesicles. Exchange of material between vesicles happens either via pores between adjacent (hemi-fused) vesicles or via diffusion. This material exchange is maximal between nearby vesicles. Indeed, already vesicles which are randomly placed on a substrate may yield considerable "compartmentalization gains" (Fig. 1a,c,d); by employing the selective coupling of vesicles to the substrate or to other vesicles (Fig. 1b) such a compartmentalized system can additionally be spatially organized and its efficiency may be further enhanced (Fig. 1e). Crucial for us is that spatial structuring is beneficial to the first instance because the material exchange between reaction containers can be controlled and optimized. But additionally, we claim that spatial structuring enables to realize different reaction networks by one and the same set of chemical modules.
An example: Many reactions are pH-dependent. Assume that we have modules Mi, which produce, as a function of pH, either product Ai or Bi. Assume further, that we have a sensory module S, which, triggered by an external chemical or optical signal, changes the pH-value in its immediate neighborhood. Depending on the relative placement of the Mi and S, a different reaction cascade is deployed. We emphasize the efficiency gain in the development process: the modules Mi can be optimized independently of each other and are applicable in different contexts. Vesicle-based multi-step reactors are interesting for several reasons:
1 The construction of a process chain with conventional means is expensive. This holds especially if one works with very small amounts of substance. Exactly this will be important in prospective applications of biochemistry, such as personalized medicine.
2 The walls of the vesicles, the membranes, are chemically active. The crucial fact is here that the membrane contains substances which are soluble only in lipid environments, and therefore enables the direct contact of these molecules with the aqueous solution (containing water soluble substances) of the interior of the vesicles. We aim at a specific exploitation of these properties with the goal of implementing a cell-like process management.
3 Constructing two-dimensional vesicle-based reactors allows the usage of technologically well-known procedures such as lithography or self-structuring by selective coupling.
4 The small size of the vesicles leads to short distances between different reaction vessels. In macroscopic processors, substances, which are only stable for seconds can be employed only with considerable effort. On the scale of vesicles, however, a few seconds are sufficient to transport a molecule form one vessel to the other. This implies that novel reaction pathways can be exploited.
5 Processes can be controlled by optical means, e.g. laser. Regions of a two-dimensional array can be selectively excited and thereby, a novel type of spatially resolved process control can be implemented.
We are guided by the idea that micro-structured reaction environments which are built from bio-inspired materials form a suitable platform for the implementation of cell-like process management. Of course, we don't expect to exploit the biochemical potential demonstrated by living cells. But we believe that we will provide the fundamentals for a form of process management which couples molecular, supra-molecular and membrane-physical processes in a technologically novel way.
Assisted Self-assembly
We emphasize that the notion of self-assembly has to be used with care. Individual vesicles may emerge autonomously in according mixtures of lipids and water. This doesn't hold anymore for vesicle assemblies, at least not for vesicles in the micro-meter range. This, because the thermal or Brownian motion is too small to achieve the necessary collision frequency between the components. During PACE several methods for "assisted" self-assembly were developed. The "self" in self-assembly refers to the autonomous and local control of the assembly process, but not anymore to the autonomous placement of components.
The Step to Programmability
Vesicle-based reactors, which, as a function of the employed modules and spatial structuring, show specific behavior, are scientifically interesting. But they are only suitable as a base for a novel production- and process-technology if their behaviors are not found just by accident and depend in an erratic manner on the choice of the used components and their coupling. Therefore, in future projects, it will be our aim to create "programmable matter". This means: We identify a number of chemical modules and depending on the way these components are arranged in assemblies, different emergent functionalities will result, as sketched in an example above. The structure of these assemblies is given by the choice of the adhesion elements used for the decoration of the vesicles. A main task in the realization of applications consists in the development of a scheme that at least stochastically predicts the functionality that emerges from a given set of linkers and vesicle contents and thereby forms the basis for establishing "programmable" reactors. We thereby envisage employing a methodology which is a mix of bottom-up simulation and top-down statistical means, which have been specifically developed by Irene Poli and co-workers at the University Venice.
- Individual steps of the reaction may interact with each other in an undesired manner.
- Many process steps only run efficiently under relatively specific conditions, say with respect to pH. These conditions may vary quite considerably for the different steps. Combining all steps in one reaction container implies to choose "one fits for all" conditions, which may lead to a significantly sub-optimal result.
Compartmentalization facilitates the de-coupling and optimization of different process steps. But there is a price to pay: The products of the individual steps have to be transported between the vesicles. Modern process design relies on the usage of multiple reaction vessels, which are connected such that they form a finely tuned process chain. It is exactly this process chain that we want to realize by the spatial structuring of the vesicles. Exchange of material between vesicles happens either via pores between adjacent (hemi-fused) vesicles or via diffusion. This material exchange is maximal between nearby vesicles. Indeed, already vesicles which are randomly placed on a substrate may yield considerable "compartmentalization gains" (Fig. 1a,c,d); by employing the selective coupling of vesicles to the substrate or to other vesicles (Fig. 1b) such a compartmentalized system can additionally be spatially organized and its efficiency may be further enhanced (Fig. 1e). Crucial for us is that spatial structuring is beneficial to the first instance because the material exchange between reaction containers can be controlled and optimized. But additionally, we claim that spatial structuring enables to realize different reaction networks by one and the same set of chemical modules.
An example: Many reactions are pH-dependent. Assume that we have modules Mi, which produce, as a function of pH, either product Ai or Bi. Assume further, that we have a sensory module S, which, triggered by an external chemical or optical signal, changes the pH-value in its immediate neighborhood. Depending on the relative placement of the Mi and S, a different reaction cascade is deployed. We emphasize the efficiency gain in the development process: the modules Mi can be optimized independently of each other and are applicable in different contexts. Vesicle-based multi-step reactors are interesting for several reasons:
1 The construction of a process chain with conventional means is expensive. This holds especially if one works with very small amounts of substance. Exactly this will be important in prospective applications of biochemistry, such as personalized medicine.
2 The walls of the vesicles, the membranes, are chemically active. The crucial fact is here that the membrane contains substances which are soluble only in lipid environments, and therefore enables the direct contact of these molecules with the aqueous solution (containing water soluble substances) of the interior of the vesicles. We aim at a specific exploitation of these properties with the goal of implementing a cell-like process management.
3 Constructing two-dimensional vesicle-based reactors allows the usage of technologically well-known procedures such as lithography or self-structuring by selective coupling.
4 The small size of the vesicles leads to short distances between different reaction vessels. In macroscopic processors, substances, which are only stable for seconds can be employed only with considerable effort. On the scale of vesicles, however, a few seconds are sufficient to transport a molecule form one vessel to the other. This implies that novel reaction pathways can be exploited.
5 Processes can be controlled by optical means, e.g. laser. Regions of a two-dimensional array can be selectively excited and thereby, a novel type of spatially resolved process control can be implemented.
We are guided by the idea that micro-structured reaction environments which are built from bio-inspired materials form a suitable platform for the implementation of cell-like process management. Of course, we don't expect to exploit the biochemical potential demonstrated by living cells. But we believe that we will provide the fundamentals for a form of process management which couples molecular, supra-molecular and membrane-physical processes in a technologically novel way.
Assisted Self-assembly
We emphasize that the notion of self-assembly has to be used with care. Individual vesicles may emerge autonomously in according mixtures of lipids and water. This doesn't hold anymore for vesicle assemblies, at least not for vesicles in the micro-meter range. This, because the thermal or Brownian motion is too small to achieve the necessary collision frequency between the components. During PACE several methods for "assisted" self-assembly were developed. The "self" in self-assembly refers to the autonomous and local control of the assembly process, but not anymore to the autonomous placement of components.
The Step to Programmability
Vesicle-based reactors, which, as a function of the employed modules and spatial structuring, show specific behavior, are scientifically interesting. But they are only suitable as a base for a novel production- and process-technology if their behaviors are not found just by accident and depend in an erratic manner on the choice of the used components and their coupling. Therefore, in future projects, it will be our aim to create "programmable matter". This means: We identify a number of chemical modules and depending on the way these components are arranged in assemblies, different emergent functionalities will result, as sketched in an example above. The structure of these assemblies is given by the choice of the adhesion elements used for the decoration of the vesicles. A main task in the realization of applications consists in the development of a scheme that at least stochastically predicts the functionality that emerges from a given set of linkers and vesicle contents and thereby forms the basis for establishing "programmable" reactors. We thereby envisage employing a methodology which is a mix of bottom-up simulation and top-down statistical means, which have been specifically developed by Irene Poli and co-workers at the University Venice.
a) Distribution of a multi-step reaction on several compartments: This enables to choose optimal conditions for the individual processes and prohibits unwanted interactions between different processes. This at the cost of having the necessity to organize the material transport between the vesicles. b) Vesicles can be coupled selectively either to each other (shown here) or to a substrate. This allows the construction of spatially structured vesicle assemblies. c) Based on the fact that vesicles enable a direct interaction between membrane-soluble and water-soluble substances, already homogeneous multi-vesicle reactors are of technological interest. But we want to go further: d) shows structures that are composed of several vesicle types and finally e) in which the vesicles are arranged according to specific correlations (Here pictorially: Red and blue vesicles are not in direct contact and don't (directly) exchange material (see reaction scheme in 1a). In order to obtain such patterns, we need the selective coupling illustrated in 1b. Fig. 1b: sequence specific coupling, realized at the AILab by DNA-DNA-linkers, in future DNA-PNA-constructs. Realization by Prof. Peter Eggenberger and Maik Hadorn, figure 1b by Maik Hadorn.