At the current level of scientific capabilities, reconstructing contemporary living cells from their parts is still out of reach. Although scientists have decoded the genomes of several simple organisms, this is still not enough to yield a working organism. In contrast, we believe that the design of an artificial cell, namely a highly simplified version of a biological cell, might be achievable in the near future. To lay the foundations for any discussion of this endeavor, it is essential to clearly define the properties by which the living character of any system is measured. The minimal requirements for a living system are often listed as follows: A living system must (1) have a specific identity and be able to preserve it (compartmentalization), (2) sustain itself by using energy from its environment to manufacture at least some of its components from resources in the environment (metabolism), the remainder being acquired from the environment without the need for further alteration. (3) The system must be capable of growth and self-replication (a prerequisite here being a genetic building plan). This point may represent the most challenging aspect of an artificial cell because it presupposes a true regulation of the simultaneous growth and self-replication of all the cell components by a simple genetic system. (4) Finally, it must be capable of adaptation (open-ended evolvability). We will take these requirements as given.

Historically, the design of chemical systems to mimic cells or create artificial ones has been based on amphiphilic containers (liposomes, vesicles, and emulsion compartments; see Fig. 1a) whose enclosed aqueous volumes harbored the enzymatic systems, even ones as complex as the transcription–translation machinery (Noireaux and Libchaber 2004; Yu et al. 2001). These systems, which typically had their substrates co-encapsulated, produced RNA (Chakrabarti et al. 1994; Monnard et al. 2007) and/or proteins. In some instances, precursors of container molecules were added to the external medium (the environment) and induced container growth and division (Oberholzer et al. 1995; Walde et al. 1994). However, both the container growth and the production of genetic material remained disconnected. Moreover, the self-replication of a liposome-based protocell with an internal content as complex as a minimal genome that can encode for the protein translation and the production of compartment molecules [i.e., 37 genes for a minimal transcription–translation machinery such as the Pure System™ (Shimizu et al. 2001) and several genes for the amphiphile production] seem incredibly difficult, especially considering that each protein or gene may require a different concentration in the protocell to ensure its proper functioning.

Fig. 1
figure 1

Two approaches to the protocell: a Enclosed model b Interfacial model. a The enzymatic machinery for transcription and translation is encapsulated in the aqueous volume surrounded by a membranous compartment—a liposome or vesicle. Substrates (S)—amino acids, nucleotides—and the precursors of compartment molecules (pL) are added in the external medium. They must either cross the membrane using, e.g., transient membrane packing defects (substrates, dotted arrow) to be incorporated in metabolic products (black arrow) or insert in the membrane (pL, dashed arrow) to be transformed into compartment molecules (double-line arrow). b The protocell system self-assembles due to the properties of all its components (photocatalytic metal catalysts, information molecule template, both with a hydrophobic moiety, and the amphiphiles). The compartmentalization occurs within the amphiphile structure itself, here the bilayer of a vesicle. The precursors for both the information oligomeric precursors and the compartment boundary molecules insert into the bilayers and upon irradiation are transformed in situ into full-length information and amphiphile molecules, respectively. In both systems, the production of new components should lead to growth and eventually division into “daughter” systems of equivalent complexity

Full size image

At Los Alamos National Laboratory, we are now attempting to design a chemical system (Rasmussen et al. 2003, 2004) in which the regulation of the metabolism (promoting growth and self-replication) is truly carried out by “genetic” information molecules (see Fig. 1b). Our system is based on a self-assembled chemical system composed of fatty acids (container molecules), amphiphilic peptide nucleic acids (PNA), and hydrophobic metabolic complexes (photosensitizers). The PNA acts not only as a “genetic” information molecule, but also as a sequence-dependent, metabolic mediator in conjuction with a ruthenium metal complex as the photosensitizer. The PNA–photosensitizer complex resides in a ternary colloid mixture consisting of a fatty acid and its oil-like precursor in water.

Due to the intrinsic properties of the PNA–photosensitizer complex, these components will accumulate at or near the surface of the amphiphilic fatty acid structures and promote the metabolic steps: the non-enzymatic polymerization of the information molecule and the photochemical fatty acid formation from its oil-like precursor. Thus, the metabolism will lead to replication of both the information material and the compartment, and in turn growth and division. Furthermore, we have proven theoretically that the replication rate of the information molecule and the container self-regulate one another (Munteanu et al. 2007; Rocheleau et al. 2007).

Such a system could clearly fulfill the three first prerequisites for any living system. What about evolvability? Evolution seems to be the product of three main processes: the pressure of the environment due to e.g. changes in physical conditions, the inevitable competition between different “species,” and the ability of the genetic information to change due to errors in its reproduction and a possible increase in its size. The chemical evolution of a simple system could be artificially triggered by changing the chemicals provided in its environment, but would it lead to gradually more complex systems that ultimately would be able to function like true cells or at least like reasonable, cellular mimics? The question remains open for now.