Cell-free synthetic biochemistry aims to engineer chemical biology by exploiting biosynthetic dexterity outside of the constraints of a living cell. One particular use is for making natural products, where cell-free systems have initially demonstrated feasibility in the biosynthesis of a range of complex natural products classes. This has shown key advantages over total synthesis, such as increased yield, enhanced regioselectivity, use of reduced temperatures and less reaction steps. Uniquely, cell-free synthetic biochemistry represents a new area that seeks to advance upon these efforts and is particularly useful for defining novel synthetic pathways to replace natural routes and optimising the production of complex natural product targets from low-cost precursors. Key challenges and opportunities will include finding solutions to scaled-up cell-free biosynthesis, as well as the targeting of high value and toxic natural products that remain challenging to make either through whole-cell biotransformation platforms or total synthesis routes. Although underexplored, cell-free synthetic biochemistry could also be used to develop ‘non-natural’ natural products or so-called xenobiotics for novel antibiotics and drugs, which can be difficult to engineer directly within a living cell.
Cell-free synthetic biochemistry seeks to explore the boundaries of what we define as ‘natural’ and ‘synthetic’ in biosynthesis, with recent progress including building synthetic CO2 fixation , reshaping a synthetic glycolysis pathway  and non-standard amino acid incorporation . In essence, this can be portrayed as a form of enzyme alchemy. For readers such as graduate students and postdoctoral researchers who are unfamiliar with the ‘cell-free’ terminology, essentially this is about engineering chemical biology within a test tube, using purified enzymes or cell extracts containing biosynthetic machinery. Moreover, cell extracts are also capable of catalysing coupled transcription and translation to make proteins from DNA. This is referred to as cell-free protein synthesis (CFPS), which is summarised in these reviews [4,5]. Distinctly, this perspective specifically offers an opinion on the potential of cell-free synthetic biochemistry for natural products. Here, it is important to reflect upon early achievements of demonstrating feasibility of cell-free approaches to natural product biosynthesis, before discussing the new opportunities that cell-free synthetic biochemistry presents. This includes examining the potential of cell-free processes for scaled-up natural product synthesis, combinatorial biosynthesis and the development of new ‘non-natural’ natural products as xenobiotics for drug discovery.
Natural product synthesis: caveats and opportunities
How are natural products obtained and why does cell-free biosynthesis provide a distinct advantage over traditional routes? So far, there is a myriad of over 210 000 natural product compounds isolated from microbial, plant and marine sources . Some of these are pigmented due to extensive conjugation and many are exquisitely crafted into complex cyclic aromatic structures and contain multiple chiral centres. Indeed, all natural products have some nature-intended biological purpose, which can be serendipitously useful to humans as medicines, herbicides, antibiotics and immunosuppressants . Ideally, these are harvested from natural sources; however, many are produced in low quantities, and thus for industrial production, it is often favourable to use organic chemistry — also referred to as total synthesis. Or if a basic scaffold can be isolated from the natural source, semi-synthesis is used to provide the final polishing modifications through organic chemistry to yield the desired product. One example includes the anti-malaria agent artemisinin . However, organic synthesis may not be feasible due to substantially more complex aromatic scaffolds, multiple chiral centres and reactive species that require protection groups . Alternatively, in natural product biosynthesis, enzymes typically employ the chemical logic of scaffold assembly, followed by decoration with a range of chemical functional groups. In comparison with total synthesis, enzymes provide many distinct advantages, such as enhanced regioselectivity, reduced temperatures and the ability to build C–C bonds from simple precursors under physiological conditions. While the use of large quantities of heavy metal catalysts are avoided, it should be considered that >25% of all proteins require a metal for function, albeit at lower concentrations . To orchestrate natural product biosynthesis, typically whole-cell microbial platforms are used to recombinantly produce the enzymes for a natural product pathway. However, this can be compromised by a need to divert carbon flux away from central metabolism, substrate import, product export and metabolic toxicity . As an intriguing compromise, cell-free approaches offer an entirely distinct direction.
Cell-free-based synthesis of natural products provides an opportunity to manually exploit biosynthetic dexterity without the competing constraints of a living whole-cell biocatalyst such as metabolic cross-talk, toxicity and transport. However, enzymes when handled outside of a living cell present some important limitations. Fundamentally, due to the stability and cost of purification, most enzymes are unsuitable for industrial-scale synthesis. More typically, enzymes also have narrow substrate tolerances and can use specialised cofactors or substrates, which may be difficult to acquire or are costly . While these are important caveats to consider, this also creates opportunities to exploit. For example, in terms of scale, a recent study showed the potential of a 20 000 l scale cell-free biosynthesis of inositol from maltodextrin, by semi-purifying thermostable enzymes via heating . In addition, to increase thermostability and broaden substrate specificity, enzymes can be engineered with directed evolution  or provide a template for synthetic protein design of bespoke catalytic reactions . Notwithstanding, to regenerate reducing equivalents and transferable groups for expensive cofactors (e.g. ATP, NADH, NADPH, malonyl-CoA), alternative lost-cost substrates and recycling enzymes are available for cell-free applications . For example, stable and inexpensive nicotinamide biomimetics can be used for NAD(P)H-dependent reductase enzymes [16,17].
The ‘one-pot’ biosynthesis approach
The concept of using cell-free approaches to assemble natural products is often referred to, within the literature, as ‘one-pot’ biosynthesis or total enzyme synthesis. The latter terminology is used based on the analogy to total synthesis for a specific target, but replacing organic solvents and catalysts, with enzymes. In distinction, cell-free synthetic biochemistry applies the same concept but seeks to advance into new areas, which is discussed later. To begin, seminal work by Professor Ian Scott, an organic chemist at Texas A&M, originally demonstrated a partial ‘one-pot’ biosynthesis of vitamin B12 [18–20]. While the full pathway (∼30 enzymes) was never completed using cell-free approaches, these pioneering efforts were more important about demonstrating the feasibility of cell-free biosynthesis at this time. To put this into context, although not fully reported, the landmark total synthesis of vitamin B12 by the groups of Robert Woodward and Albert Eschenmoser required ∼80–100 chemical steps . Initially, Scott and co-workers delivered a 12-enzyme ‘one-pot’ biosynthesis of hydrogenobyrinic acid at ∼20% yield, a late-stage vitamin B12 precursor (Figure 1) using an oxygen-dependent pathway . This feat was later recreated with the alternative oxygen-independent pathway using a multi-staged 15-enzyme synthesis of cobyrinic acid at ∼30% yield . Moving to the more exclusive members of secondary metabolism, there has been a few examples for ‘one-pot’ biosynthesis of natural products, predominantly from the polyketide family that are biosynthesised from building blocks such as acetyl-CoA, malonyl-CoA and derivatives (Figure 1). One noteworthy example is the elucidation of tetracenomycin biosynthesis and the type II polyketide synthase (PKS) assembly by Hutchinson and co-workers . Other polyketide ‘one-pot’ biosynthesis examples include type II polyketides norsolorinic acid , rabelomycin , defucogilvocarcin M , enterocin [28,29] and type III polyketide merochlorin A . Where reported, it is important to highlight that a number of these ‘one-pot’ reactions typically require fewer steps and provide higher yields than total synthesis, with reported yields between 9 and 80% from the starting material. Notably, these reactions begin from complex and expensive primary metabolic building blocks such as malonyl-CoA (∼£10 000 a gram), a dilemma that cell-free synthetic biochemistry seeks to address. Moreover, for type I PKSs and non-ribosomal peptide synthetases (NRPS), these are large enzymes composed of a modular domain architecture. These enzymes alone can be difficult to produce and purify to provide sufficient quantities for scaled-up synthesis. Consequently, only a few complete examples of ‘one-pot’ biosynthesis have been demonstrated with the type I PKS pathways including dihydromonacolin L , 6-deoxyerythronolide B  and tetronate RK-682 , as well as the hybrid PKS/NRPS-derived ikarugamycin . However, even at the preparative scale, the need for high quantities of enzyme material, as well as the requirement for post-translational modification of PKS/NRPS enzymes with a phosphopantetheine group, limits their role at present for scalable enzyme synthesis. A specific review on cell-free approaches to PKS and NRPS natural products is covered by Li et al. .
‘One-pot’ biosynthesis of natural products.
Cell-free synthetic biochemistry for natural products?
As an emerging field, there are only a handful of cell-free synthetic biochemistry publications so far that advance upon earlier efforts in ‘one-pot’ biosynthesis. In contrast, cell-free synthetic biochemistry provides greater focus on reshaping natural biocatalytic processes into synthetic pathways, using an entire plethora of enzymes from different genetic sources, including plants, bacteria, fungi and higher eukaryotes. Moreover, particular attention is drawn to refining cofactor regeneration, the efficient provision of building blocks, demonstrating scalability, longevity and the use of synthetic biology for rapid optimisation . In a noteworthy example, the cannabinoid pathway was recently built in cell-free by Professor James Bowie and co-workers, using a synthetic module of enzymes to convert d-glucose into the terpenoid precursor acetyl-CoA, based on their previous efforts . This is remarkable, since the previous reported yields of cannabinoids in whole-cell biotransformation were minimal, while this cell-free synthetic biochemistry system used 25 enzymes to provide titres exceeding 1 g/l . One of the advantages of cell-free synthetic biochemistry demonstrated by this research is the use of synthetic enzyme modules to efficiently build complex and high-value natural products directly from basic inexpensive precursors . This provides a new development where different areas of synthetic metabolism can be initially tested within a cell-free environment, before potential forward engineering into living cells. However, to expand synthetic biochemistry further, there is still an essential need to characterise a greater variety of enzymes and pathways. In particular, for natural products, bioinformatic studies based on genomic sequencing have highlighted that approximately <1% of all natural product biosynthetic pathways have been explored at either the genetic or biochemical level .
An underexploited area is the development of ‘non-natural’ natural products or so-called xenobiotics for tailored design of new drugs or antibiotics. For this area, cell-free is particularly useful for directing biosynthesis in a controlled, low-scale environment and is advantageous for combinatorial biosynthesis where different sets of enzymes and substrates can be explored, either as purified enzymes or cell extracts, as recently demonstrated . To consider how xenobiotics can potentially be synthesised using cell-free synthetic biochemistry, this certainly will also rely upon extensive enzyme engineering and directed evolution efforts [11,13], as well as exploring natural enzyme promiscuity. In Kalaitzis et al. , a range of benzoic acid analogues were provided as the starting substrate for a bespoke type II PKS machinery in the enterocin biosynthetic pathway. By substituting the precursor benzene ring with –OH, –Cl, –F groups or altering the ring with pyridine and thiophene motifs, 24 new unnatural analogues of enterocin and the derivative wailupemycin were generated using ‘one-pot’ cell-free synthesis . Here, the introduction of halogens is particularly important for medicinal chemistry, where the pharmacokinetics of the drugs can be altered favourably by manipulating the halogen motifs. Although halogens are not found in primary metabolism, there are several reported halogenated natural products and corresponding halogenation enzymes . While it can be difficult to directly modify a biosynthetic pathway from the starting substrate, alternatively, late-stage tailoring steps with modifiable transfer groups could also provide an important tool. One example is the potential promiscuity offered by the S-adenosyl-l-methionine (SAM) methyltransferase (MTase) family, where SAM provides a universal donor for a range of C-, O-, S- and N-type MTases. Here, promiscuous SAM synthetases can use ATP and a range of methionine derivatives to generate up to 29 distinct SAM analogues, including allyl, propyne and nitrile transfer groups . For the development of xenobiotics, 8 out of 18 SAM analogues were incorporated into the rebeccamycin indolocarbazole scaffold in a ‘one-pot’ biosynthesis approach. Also, in the context of ribosomal and non-ribosomal peptides, cell-free offers an enticing opportunity for expanding natural product chemistry through incorporation of non-canonical amino acids, including l- and d-amino acid analogues. Although still an actively developing area, the use of CFPS systems can potentially leverage the power of automation and combinatorial design for screening different DNA sequences. Initially, CFPS has orchestrated the biosynthesis of non-ribosomal peptides  and ribosomally synthesised and post-translationally modified peptides directly from DNA . Here, a range of different bacterial CFPS systems are available for this purpose [43–46]. In summary, although still a developmental area, cell-free synthesis of ‘non-natural’ natural products provides a controlled environment for directing biosynthesis with potential for combinatorial opportunities. However, further developments are required to bridge the gap of simply demonstrating feasible synthesis, towards tailor designed xenobiotics with a desired biological activity.
Cell-free synthetic biochemistry is beginning to reignite the potential of ‘one-pot’ biosynthesis by diversifying biosynthetic capabilities . This includes demonstrating low-scale feasibility challenges, such as replacing natural biosynthetic pathways with synthetic routes [1,2,37,47], which may be useful for solely in vitro purposes, or for direct engineering of synthetic metabolism in living cells. Moving forward, key questions and challenges remain to be addressed to fully exploit the potential of cell-free synthetic biochemistry. In terms of scale-up, while single-step cell-free enzyme processes are well established in industrial processes (e.g. textile and paper manufacturing), technical challenges surrounding complex natural product pathways have limited progress beyond the preparative scale. For this to be cost-effective, natural products that are high value, toxic to cell growth or essential biopharmaceuticals are the most suitable focus points . To permit scale-up, since enzymes can have restrictive conditions for activity, either continuous-flow systems or multi-staged cell extract-based reactions [23,39] may present important solutions, while energy and reducing equivalents can also be regenerated within the cell extract . On the screening scale, cell-free will also be advantageous for combinatorial approaches. This can be performed with either cell extracts containing pathway enzymes, purified enzymes or potentially using CFPS, which provides extra control with the template DNA sequence. In addition, it will also be essential to expand the toolset of enzymes and the broadening of substrate scope, to accelerate the development and capabilities of cell-free synthetic biochemistry. Finally, an underexplored but important challenge for cell-free synthetic biochemistry is the potential to develop ‘non-natural’ natural products or so-called xenobiotics for drug discovery purposes, where this perspective only begins to highlight the essence of this potential. In summary, cell-free synthetic biochemistry provides a new field for synthetic biologists, chemists and enzymologists alike to collaborate and push the frontier of this exciting but uncharted territory.
Cell-free synthetic biochemistry offers an attractive route to manipulate biosynthetic pathways for natural products.
Cell-free systems provide potential advantages for natural product biosynthesis with increased yield, less steps, absence of heavy metal catalysts and enhanced regioselectivity over total synthesis.
Potential opportunities include the design of enhanced synthetic pathways, combinatorial screening, scale-up or generation of ‘non-natural’ natural product analogues.
The author acknowledges support from the Royal Society [RGS\R1\191186].
The author would like to thank Mr Terrence Lai, a PhD student at Imperial College London, for critical discussion.
The Author declares that there are no competing interests associated with this manuscript.