A mixture of molecules can be regarded as a network if all the molecular components participate in some kind of interaction with other molecules--either physical or functional interactions. Template-assisted ligation reactions that direct replication processes can serve as the functional elements that connect two members of a chemical network. In such a process, the template does not necessarily catalyze its own formation, but rather the formation of another molecule, which in turn can operate as a template for reactions within the network medium. It was postulated that even networks made up of small numbers of molecules possess a wealth of molecular information sufficient to perform rather complex behavior. To probe this assumption, we have constructed virtual arrays consisting of three replicating molecules, in which dimer templates are capable of catalyzing reactants to form additional templates. By using realistic parameters from peptides or DNA replication experiments, we simulate the construction of various functional motifs within the networks. Specifically, we have designed and implemented each of the three-element Boolean logic gates, and show how these networks are assembled from four basic "building blocks". We also show how the catalytic pathways can be wired together to perform more complex arithmetic units and network motifs, such as the half adder and half subtractor computational modules, and the coherent feed-forward loop network motifs under different sets of parameters. As in previous studies of chemical networks, some of the systems described display behavior that would be difficult to predict without the numerical simulations. Furthermore, the simulations reveal trends and characteristics that should be useful as "recipes" for future design of experimental functional motifs and for potential integration into modular circuits and molecular computation devices.

This Minireview gives an overview of recent progress in the design and analysis of chemical systems that utilize template-directed autocatalytic and cross-catalytic processes as a means of wiring dynamically interacting molecules. Synthetic networks comprising two to nine replicating species are discussed. It is shown that for larger systems, more catalytic pathways must be manipulated to control the entire network topology and specific functionality of the individual species or subnetworks. Cellular biochemistry is an example of a natural functional molecular network; synthetic self-organized networks can provide additional models of complex systems.

No abstract

An emerging new direction of research focuses on developing "self-synthesizing materials", those supramolecular structures that can promote their own formation by accelerating the synthesis of building blocks and/or an entire assembly. It was postulated recently that practical design of such systems can benefit from the ability to control the assembly of amphiphilic molecules into nanostructures. We describe here the self-assembly pathway of short amphiphilic peptides into various forms of soluble β-sheet structures--β-plates, fibrils, and hollow nanotubes--and their consequent activity as autocatalysts for the synthesis of monomeric peptides from simpler building blocks. A detailed kinetic analysis of both the self-assembly and self-replication processes allows us to suggest a full model and simulate the replication process, revealing that only specific structures, primarily fibrils that are stable within the solution for a time shorter than a few hours, can be active as catalysts. Interestingly, we have found that such a process also induces fibril reproduction, in a mechanism very similar to the propagation of prion proteins by transmission of misfolded states.

One of the grand challenges in contemporary systems chemistry research is to mimic life-like functions using simple synthetic molecular networks. This is particularly true for systems that are out of chemical equilibrium and show complex dynamic behaviour, such as multi-stability, oscillations and chaos. We report here on thiodepsipeptide-based non-enzymatic networks propelled by reversible replication processes out of equilibrium, displaying bistability. Accordingly, we present quantitative analyses of the bistable behaviour, featuring a phase transition from the simple equilibration processes taking place in reversible dynamic chemistry into the bistable region. This behaviour is observed only when the system is continuously fueled by a reducing agent that keeps it far from equilibrium, and only when operating within a specifically defined parameter space. We propose that the development of biomimetic bistable systems will pave the way towards the study of more elaborate functions, such as information transfer and signalling.

The kinetics of novel dynamic libraries that operate via reversible replication is described. In these systems, selective product formation is governed by peptides autocatalytic efficiency and by differences in their unfolding stability. We suggest ways to significantly alter the network behavior by chemical inputs (templates) or physical triggers (light).

Biopolymer syntheses in living cells are perfected by an elaborate error correction machinery, which was not applicable during polymerization on early Earth. Scientists are consequently striving to identify mechanisms by which functional polymers were selected and further amplified from complex prebiotic mixtures. Here we show the instrumental role of non-enzymatic replication in the enrichment of certain product(s). To this end, we analyzed a complex web of reactions in β-sheet peptide networks, focusing on the formation of specific intermediate compounds and template-assisted replication. Remarkably, we find that the formation of several products in a mixture is not critically harmful, since efficient and selective template-assisted reactions serve as a backbone correction mechanism, namely, for keeping the concentration of the peptide containing the native backbone equal to, or even higher than, the concentrations of the other products. We suggest that these findings may shed light on molecular evolution processes that led to current biology.

Bistable reaction networks provide living cells with chemically controlled mechanisms for long-term memory storage. Such networks are also often switchable and can be flipped from one state to the other. We target here a major challenge in systems chemistry research, namely developing synthetic, non-enzymatic, networks that mimic such a complex function. Therefore, we describe a dynamic network that depending on initial thiodepsipeptide concentrations leads to one of two distinct steady states. This bistable system is readily switched by applying the appropriate stimuli. The relationship between the reaction network topology and its capacity to invoke bistability is then analyzed by control experiments and theory. We suggest that demonstrating bistable behavior using synthetic networks further highlights their possible role in early evolution, and may shine light on potential utility for novel applications, such as chemical memories.