Cell-free transcription–translation is an expanding field in synthetic biology as a rapid prototyping platform for blueprinting the design of synthetic biological devices. Exemplar efforts include translation of prototype designs into medical test kits for on-site identification of viruses (Zika and Ebola), while gene circuit cascades can be tested, debugged and re-designed within rapid turnover times. Coupled with mathematical modelling, this discipline lends itself towards the precision engineering of new synthetic life. The next stages of cell-free look set to unlock new microbial hosts that remain slow to engineer and unsuited to rapid iterative design cycles. It is hoped that the development of such systems will provide new tools to aid the transition from cell-free prototype designs to functioning synthetic genetic circuits and engineered natural product pathways in living cells.
Cell-free systems represent a historically important component of the founding of the field of biochemistry. Ever since the pioneering efforts of the Nobel laureate Eduard Buchner (Nobel Prize in Chemistry in 1907) and his discovery of fermentation in yeast cell extracts , cell-free systems have been repurposed towards the further understanding of biological processes. Indeed, arguably one of the most notable biological discoveries of the 20th century was the unravelling of the genetic code by Nirenberg and colleagues [2–4], which was underpinned by the use of Escherichia coli cell extracts to study coupled transcription–translation (TX–TL). Together with Har Khorana and Robert Holley, this resulted in a shared Novel Prize in Physiology or Medicine in 1968. On this theme, the efforts of Alfred Goldberg led to the unveiling of an ATP-dependent mechanism for protein degradation by ubiquitin in a mammalian cell-free system .
Cell-free synthetic biology
Today, with the rise of synthetic biology and the design and construction of synthetic life , cell-free systems have yet again found a niche towards the understanding of biological networks and biosynthetic pathways [7,8]. Indeed, by isolating the cellular components of core metabolism and the TX–TL network within a test tube, this allows the synthetic biologist to study systems without the regulatory constraints and limitations of a dividing, evolving or adapting living cell. This mini-review summarises the efforts of recent cell-free synthetic biology research and the opportunities it provides for the future.
Cell-free coupled TX–TL utilises the core machinery of RNA polymerase holoenzyme, the translation apparatus (ribosomes, tRNA synthetases and translation factors) and energy regeneration enzymes to amplify a set of DNA instructions into the target protein(s) of choice (Figure 1A). Therefore, the study of cell-free presents an enticing opportunity to the synthetic biologist to design and engineer living systems from the bottom-up as prototype designs. Exemplar demonstrations of cell-free synthetic biology include their use as biomolecular ‘breadboards’ [9,10], healthcare biosensors [11,12] and enzyme cascades [13–16]. Coupled with the aid of computational design approaches [10,17–19], these early developments in cell-free synthetic biology will endeavour to aid the engineering of more complex systems. We shall now summarise the cell-free platforms available, with a specific focus to its use in prototyping genetic circuits.
Summary of cell-free TX–TL.
E. coli cell-free — purified recombinant elements or crude cell extract?
The choice of a well-characterised cell-free system almost entirely resides with E. coli platforms, which are based on either a crude cell extract [20–23] or a system of purified recombinant elements (PURE) [24,25]. A vital area of importance to cell-free systems is the process of energy regeneration, which represents the major cost factor and limitation for both the PURE and cell extract-based routes. First, transcription requires nucleotide triphosphates (NTPs — ATP, UTP, GTP and CTP), with each mRNA transcript utilised multiple times for protein synthesis . Protein translation is the major energy cost factor and requires two ATP equivalents for tRNA aminoacylation and two GTP equivalents per peptide bond formed . In addition, a single GTP equivalent is required for each of the initiation and termination steps. Therefore, a small sized 25 kDa protein costs ∼35–44 mM of ATP to synthesise 1 mg/ml under batch synthesis .
First, in respect to the PURE system , this includes the purified components (108 in total) of the entire E. coli translation machinery including ribosomes, 22 tRNA synthetases, initiation factors, elongation, release and termination factors, which when combined with T7 RNA polymerase, tRNA, energy regeneration enzymes, substrates (amino acids and creatine phosphate) and synthetic DNA instructions, this reconstitutes the entire TX–TL network within a test tube. This rather remarkable engineering feat is commercially available as the PURExpress® kit (New England Biolabs). While the high cost of the system prohibits scaled-up applications, a variety of cell-free researchers utilise the PURExpress® system to study the dynamics and kinetics of TX–TL [24,28–32]. The major advantage of the PURE system is it's high efficiency due to an absence of competing side reactions such as non-specific phosphatases , which rapidly degrade the energy source.
In contrast, a crude cell extract provides an inexpensive route to protein synthesis. In addition, unlike the PURE system, reactions are scalable into high-volume fermentation conditions [33,34]. However, with the presence of other primary and secondary pathway enzymes (phosphatases and amino acid biosynthesis), this leads to undesirable side reactions during catabolism of the starting energy source. Importantly, based on improvements in energy regeneration schemes by the groups of Swartz [27,35–37], Jewett [38,39] and Noireaux [40–42], powerful cell extract-based batch systems can now reach recombinant protein yields of up to 2.34 mg/ml [33,40], while extended steady-state synthesis can be achieved through the use of a semi-permeable dialysis membrane device, thus elevating protein yields up to 6 mg/ml . In addition, inexpensive energy sources such as glucose , glutamate , maltose/maltodextrin  and succinate  can be utilised to reduce the cost of energy regeneration in cell-free systems (Figure 1B). To this end, various cell extract protocols have been developed and are based on the harvesting of cells at exponential phase, when typically intracellular translation is at its peak. Standardised protocols involve washing the cells, mechanical lysis  and activation of the extract through a run-off reaction, a process believed to degrade endogenous mRNA transcripts and genomic DNA that can reduce cell-free translation efficiency . Additional dialysis can also remove inhibitory small molecules, but the requirement of this varies between E. coli strains and user preference .
Cell-free TX–TL provides the ability to study gene expression in isolation with the timescale from DNA to experimental results taking a few hours [10,44,45], whereas depending on the host chassis, an in vivo based approach can take several days to weeks. Thus, cell-free provides a prototyping approach (Figure 2) for rapid cycling between circuit experimental design and debugging . To enable cell-free prototyping, fluorescence tags that monitor both mRNA and protein synthesis can be studied in real time [29,40,46,47], thus providing dual microscale quantitative data of the TX–TL cascade that can be difficult to achieve within a living cell. In addition, the starting concentration of the substrates and relative enzyme stoichiometry can be determined , thus aiding system identification and mathematical modelling of the chemical reaction dynamics [9,17,18]. These models can be used to inform future circuit designs as part of an iterative design process. In vivo, the cellular components are constantly being diluted by cell growth and division as well as being synthesised. In contrast, batch cell-free reactions are closed systems starting with a limited set of initial resources. These differences make direct comparisons between in vivo and cell-free reaction dynamics of complex multi-promoter circuits difficult. One method to combine the rapid prototyping benefits of cell-free while emulating the conditions found in living cells is to use microfluidic devices to allow the continuous dilution and replenishment of the reaction substrates. This method was used to design three- and five-node ring oscillators in cell-free, based on the utilisation of PCR templates to test initial prototypes, before a model-inspired design-build-test cycle led to circuit designs that were also found to function in cells . On this theme, cell-free provides a dynamic biochemical system that can be accurately described by mathematical modelling .
Prototyping cell-free TX–TL systems.
In another context, cell-free prototyping can also be useful towards the design of synthetic cells. At the systems level, central to this effort is the further understanding of cellular compartmentalisation. Owing to its difficulty, especially at the structural level, a perhaps understudied area of biology is the dynamics of protein folding in the lipid membrane bilayer. Cell-free uniquely provides an opportunity to study the folding of membrane proteins , while in synthetic microfluidic-based liposomes, enzymes and substrates can be transported from one cell to another, demonstrating a simple recreation of membrane trafficking . Towards complexity, cell-free systems have also begun to be implemented for the assembly of large protein complexes. The bacteriophage represents a simple life form with a package of genes within a protein shell, which is released upon invasion of a cell to hijack the hosts TX–TL apparatus. A classically studied system is the T7 bacteriophage  that infects E. coli. Through cell-free, it has now been shown possible that the 40 kb dsDNA genome of the T7 bacteriophage, which constitutes 57 genes, can be reconstituted in vitro to demonstrate the assembly of a natural protein compartment . This is also expandable to other bacteriophage systems . Moving aside beyond biological compartmentalisation, cell-free is also portable to the interface of nanotechnology for studying gene expression and synthesising protein nanotube on biochips . Together these examples of compartmentalisation demonstrate an extra level of complexity in cell-free systems for prototype designs, which may aid in the design of new synthetic cells in the future.
Non-model cell-free platforms
Viewed from a different perspective, synthetic biology has begun to examine the prospects of engineering non-model microbial hosts  that can provide unique advantages for biotechnological application, such as rapid growth with inexpensive substrates, growth in extreme conditions or unique enzyme machinery, which in some cases can only accessed within non-standard microbial hosts. However, the greatest disadvantage of such cultivatable microbes is a combination of one or more of the following traits, such as a general lack of characterised gene expression tools, poor genetic tractability or insufficient knowledge towards the microbe metabolism. While cell-free cannot directly address genetic competence, it could provide a starting point to understanding the host's inherent TX–TL kinetics, genetic tools and enzymology, without the time limitations associated with direct engineering of the host. Noticeably, the methodology for cell-free extract preparation [42,55] has shown universal application to a variety of microbial cell-free platforms such as Saccharomyces cerevisiae [56,57], Streptomyces spp. [58–61] and Bacillus spp. [43,62]. Such interest in the use and application of alternative cell-free systems as a prototyping device is likely to grow; however, the cooperative development of synthetic biology tools with translational application into live cells may provide the greatest opportunity to access the design space of traditionally difficult to engineer microbes. In particular reference to the Streptomyces family, the high G + C (%) soil bacteria, it has long been appreciated that these hosts provide unique and well-characterised platform for the assembly of a rich repertoire of natural products . For Streptomyces cell-free, the recently developed high-yielding Streptomyces lividans and Streptomyces venezuelae host platforms [58,61] can potentially provide an opportunity to access high G + C (%) enzymes from secondary metabolism directly within a test tube for combinatorial biosynthesis. With further advances in efficiency and yield, Streptomyces cell-free could be utilised for incorporating non-natural or potentially toxic substrates into natural products, towards expanding the chemical space of biosynthesis. A proof of concept of how cell-free can be utilised to incorporate non-natural amino acids into protein backbones was demonstrated for GFP synthesis in E. coli cell-free . While this technology is in its infancy, it is also possible to engineer this in living cells in high yield , which has been made available through the multiplex automated genome engineering (MAGE) technology . However, this methodology is currently only accessible to specially engineered strains of E. coli. Thus, with further developments, cell-free potentially provides a potentially fast route to prototype and engineer the application of novel chemistry to natural product biosynthesis .
The rise of cell-free systems from its historical links in foundational biochemistry has provided a platform to this expanding field in synthetic biology. Perhaps, the greatest challenge of cell-free studies is to establish and define the boundaries and limitations of mimicking cellular biology within cell-free systems. One understudied area is the impact of molecular crowding on enzyme velocities [68,69] and spatial organisation , which can only be artificially controlled in cell-free reactions. In fact, cell-free systems in essence are reminiscent of primordial biology , whereby enzymes (or ribozymes) and chemicals once freely tumbled without the restrictions of biological compartmentalisation and the regulatory control of the genome. With the growing interest in the design of a minimal synthetic cell [72–74], cell-free systems can provide a base towards the design of synthetic life from individual components. We anticipate that the prototyping and modelling of gene expression and enzyme machinery from understudied arcane microbes will provide important new tools for the cell-free synthetic biologist's disposal.
The Authors declare that there are no competing interests associated with the manuscript.