Multienzymatic Cascades and Nanomaterial Scaffolding—A Potential Way Forward for the Efficient Biosynthesis of Novel Chemical Products

Synthetic biology is touted as the next industrial revolution as it promises access to greener biocatalytic syntheses to replace many industrial organic chemistries. Here, it is shown to what synthetic biology can offer in the form of multienzyme cascades for the synthesis of the most basic of new materials—chemicals, including especially designer chemical products and their analogs. Since achieving this is predicated on dramatically expanding the chemical space that enzymes access, such chemistry will probably be undertaken in cell‐free or minimalist formats to overcome the inherent toxicity of non‐natural substrates to living cells. Laying out relevant aspects that need to be considered in the design of multi‐enzymatic cascades for these purposes is begun. Representative multienzymatic cascades are critically reviewed, which have been specifically developed for the synthesis of compounds that have either been made only by traditional organic synthesis along with those cascades utilized for novel compound syntheses. Lastly, an overview of strategies that look toward exploiting bio/nanomaterials for accessing channeling and other nanoscale materials phenomena in vitro to direct novel enzymatic biosynthesis and improve catalytic efficiency is provided. Finally, a perspective on what is needed for this field to develop in the short and long term is presented.


Introduction 1.The Conceptual Promise of Synthetic Biology
In order to understand what specifically designer multienzymatic cascades assembled on nanomaterial scaffolding have to offer for DOI: 10.1002/adma.202309963 the synthesis of new non-natural products, it is important to understand what is sparking so much interest in synthetic biology (SynBio) and, more importantly, what its current limitations are specifically.Long-term global economic and infrastructure policies now focus on developing alternative sources for the everyday chemicals our technologically dependent society relies on.[3][4][5][6] Selected countries are already implementing policies to promote this agenda as exemplified by the recent United States President's Executive Order 14081 "Advancing Biotechnology and Biomanufacturing Innovation for a Sustainable, Safe, and Secure American Bioeconomy." [7]Similar efforts are underway in the European Union, China, and elsewhere. [8,9]he overarching goal of this foundational effort is to create a circular economy where renewable bulk feedstocks, and especially those derived from agriculture or waste, are converted into industrial chemical intermediaries and fine chemical products (e.g., pharmaceuticals) by SynBio-based processes. [10]These chemical processes will be undertaken by the enzymes present in cellular systems, commonly referred to as cellular chasses, along with cellular extracts and other enzyme solutions.In its most common current implementation, large bioreactors (≈10 3 -10 5 L) of recombinantly engineered and catalytically optimized microbial cell cultures would utilize this initial feedstock material as substrate and produce large amounts of the final desired product.When directly contrasted against a synthetic chemical plant, SynBio offers many conceptual benefits including: i) being more environmentally friendly, which can save money in terms of regulatory processes and allow for easier domestic production, ii) less energy intensive in terms of having to maintain high-temperature and pressure reactions, iii) having smaller, distributed production facilities with lower capital costs, iv) no organic waste or use of toxic chemicals, v) the use of Figure 1.Schematic representation of some synthetic biology approaches toward small molecule biosynthesis.Many synthetic methods for small molecule production exist, from using live cells to purified enzymes, some of which are depicted here.Each method has its own advantages and challenges; therefore, which method is best depends on a given biosynthetic pathway and its overall goals.
][13][14] The industrially transformative promise of SynBio is real and is certainly not in dispute here.

Cell-Based versus Cell Free Synthetic Biology
There are multiple different methodologies that fall within the SynBio umbrella for molecular production, running a gradient between using live cells (in vivo) to using purified enzymes (in vitro) each with their own (sometimes overlapping) terminology such as "biosynthesis," "bioproduction," "biocatalysis," "wholecell catalysis," "synthetic biochemistry," and "cell-free" among others.Figure 1 presents a brief-schematic overview of the major types of SynBio systems in current use or undergoing concerted development.Obviously, each has its own distinctive set of advantages (Pros) and challenges (Cons), some of which are listed therein as bullet points.Live cell or in vivo biosynthesis allows for automatic regeneration of biosynthetic enzymes and their requisite cofactors, the cells contain potentially beneficial endogenous metabolic pathways without requiring engineering (e.g., for glucose breakdown), and benefit from the efficiency of sub-nanoliter volume reactions within a cell.However, this method can suffer from limitations based on the necessity and energy intensive process required to keep cells alive, including: cellular regulatory mechanisms; off-target diversion of resources to cell growth; and susceptibility to substrate, intermediate, product, or heterologous enzyme toxicity.In vivo biosynthesis can also (although not always) involve complex downstream processing (DSP) and the need to move substrates/products in and out through the cell membrane.In vitro biosynthesis using purified enzymes allows for finer control and focus on just the desired reaction and the ability to produce cytotoxic small molecules, but as enzymes are not automatically regenerated, it is limited by the enzymes' stability and total turnover number (TTN) along with achievable reaction volumes.17] The main point to be appreciated in this context is that each approach has a benefit versus liability consideration and that some approaches are more preferable over others for a given target molecule depending upon the final intended application. [14]or example, engineered -proteobacterium cells are the primary active ingredient in the liquid fertilizer product PROVEN and are meant to fix nitrogen in situ with preliminary use suggesting a reduction in the need for chemical fertilizer of ≈25 lb acre −1 while simultaneously increasing corn yields by around 5.8 bushels. [14]This example also highlights how a cellbased SynBio product is ultimately used in situ and not even in a fermenter although that is where it would be initially produced in bulk under stringent conditions.In contrast, packed bed reactors hosting mammalian cells are being used to produce both monoclonal antibodies along with selectively glycosylating them. [18,19]Cell-free transcription and translation (TX-TL) systems are in essence optimized cellular extracts, which are more amenable to single-use sensors and point of need synthesis of single dosage drug compounds and so would be used for other applications in a completely different context. [20,21]Lastly, enzyme-only or minimalistic reaction systems are more geared toward focused multistep biosynthetic pathways along with incorporation of substrates/intermediaries/products that would potentially be cytotoxic or otherwise extremely challenging to implement. [15,16,22]cheme 1. Example of non-natural analog exploration of the antitumor compound rebeccamycin.In the host pathway, RebH halogenates tryptophan to 7-chloro-tryptophan, which then proceeds down the pathway to rebeccamycin.Sánchez et al. used combinatorial biosynthesis, along with examining a brominated analog, to create 30 indolocarbazole derivatives in vivo. [24]

Limited Chemical Space of In Vivo Synthetic Biology Systems
The main issue and, indeed challenge, to achieving the long-term and full potential of SynBio is to expand upon the chemical space available to Nature.While Nature is able to produce eloquent, complex natural products difficult to produce synthetically, it is confined in many other ways.The latter point is epitomized in this statement from Professor Chris Voigt "All of the chemistry performed by the natural world can be captured with ≈250 reactions, whereas there are 60000+ in the chemistry literature." [23]he chemical reaction space available to Nature is extremely limited and this, in turn, limits the number of viable reactions and ultimately the TYPE of product syntheses that are available.Moreover, existing as complex 3-dimensional protein structures, enzymes gain much of their substrate specificity by limiting what they can react with to that which can be recognized and bound by the enzyme along with functionally accessing the catalytic site.Further limitations are added to these issues, especially in vivo, based on substrate-intermediary solubility (as cells may have limited tolerance to solubility additives), membrane transport in/out capacity, kinetics, and reaction conditions along with other reaction considerations.Lastly, and perhaps most critically, Nature is nothing if not frugal and thus almost every biochemical pathway inside a cell is cross-linked or intertwined to another with almost no pathway redundancy.Thus, even though a given enzyme may tolerate a non-natural substrate to make non-natural intermediaries or products, the latter are more often than not invariably toxic to another pathway.As a representative example, these issues are highlighted immediately below examining the biosynthesis of rebeccamycin, a natural product with strong antitumor or chemotherapeutic properties, see Scheme 1 for an overview of its biosynthesis from tryptophan. [24][26][27] The next goal will likely be to explore the structure-activity relationship (SAR) of the compound, and/or to develop structural analogs that may improve its pharmacokinetic/pharmacodynamic properties, at which point in vivo biosynthesis may hit challenges.Can an analog substrate be taken up by the host and/or pass the cellular membrane of the host?Is the analog substrate, analog intermediates, or the analog product toxic to the host?Can the analog substrate be fed exclusively to the host, or will the analog substrate be competing with a natural substrate that is also required for growth and so cause difficulties for DSP to separate and process the two products (e.g., halogenated amino acids versus non-halogenated or "natural" amino acids)?Molecular substitution with halogen groups is a common strategy to increase a drug's pharmacological activity. [28]n Scheme 1 pathway, in vivo production would have to accommodate the cellular need for endogenously synthesized tryptophan, regardless of analogs added since it is a metabolically required amino acid.Therefore, if examining the brominated analog, for example, products would include multiple rebeccamycin analogs (R 1 = H, R 2 = H; R 1 = H, R 2 = Br; R 1 = Br, R 2 = H; R 1 = Br, R 2 = Br).While this may be beneficial for small-scale synthesis and testing, large-scale synthesis of, for example, the double-brominated analog, would likely have to accommodate for extensive DSP with extended purification needs.Further, the product rebeccamycin exhibits microbial toxicity; in this case, toxicity can be (partially) overcome with a rebeccamycin transporter.[27] An alternative could be organic synthesis, but depending on the product, this can face challenges as well, for example, specific methylation of a saccharide requiring protection/deprotection steps.Summing up from this example, almost all the chemical products and materials we utilize in our society are non-natural, meaning that they are produced or incorporate chemistries that living cells currently do not undertake as part of a concerted biosynthetic pathway.It is in precisely these circumstances where we believe in vitro biosynthesis can play an outsized role in overcoming some, but not all of the challenges that SynBio faces as it seeks to expand beyond the limited scope and chemical space of natural product synthesis (vide infra).Potential ways in vitro biosynthesis could address some of these challenges include: i) easier integration of non-natural substrates toxic to cells, either as the desired non-natural state of the product or with chemical handles for downstream coupled organic synthesis, [29] ii) in some cases, increased flexibility for using solubility additives, iii) ability to limit non-desired substrates, and iv) no need to transport substrate/product analogs in/out of a cell.

Multienzymatic Cascades
The full power of enzymatic biosynthesis is typically achieved via the use of multienzyme cascades, which is also the primary way that Nature utilizes these processes.The biochemical pathways contained in biochemistry textbooks and reference works epitomize the complexity and density of these pathways, see for example http://www.biochemical-pathways.com/#/map/1.The schematics in Figure 2 highlight the four principle or functional types of multienzyme cascades.As also highlighted in this figure, cofactors are a common element in many enzyme reactions and, in turn, many multienzyme cascades.They serve to facilitate the reactions energetically along with providing or accepting key subcomponents (e.g., electrons) that energetically drive the reactions.Expanding on the concept of eons of evolutionary pressure driving Nature to be both frugal and efficient with multiple overlapping cross-dependent pathways, the use of single and focused intracellular "one-pot" multienzymatic cascades designed to take one substrate to one product manifests multiple inherent (conceptual) benefits.These include: i) no processing steps (e.g., purification or sequestration) required as the reaction steps proceed, allowing the overall required processing time and waste to be minimized, ii) toxic/unstable/highly reactive intermediates do not accumulate in the cell and can be immediately utilized, minimizing the amount of competition from potential side reactions and other pathways that are present, and iii) kinetically unfavorable and potentially reversible reactions can be driven forward or minimized. [30,31][32][33] We specifically point out the excellent review from the Kroutil Group that provides a comprehensive overview of artificial biocatalytic linear cascades for preparation of organic molecules. [34]This publication further introduced a systematic set of parameters to help classify artificial cascades.They differentiate the cascades according to the number of enzymes found in the linear reaction sequence and differentiated between those cascades, which exclusively involved enzymes versus those that combined enzymes with non-natural catalysts or chemical steps (i.e., chemoenzymatic).
In the context of industrial development and exploitation, assembling multienzyme reactions in an in vitro "minimalist" or "one-pot" format can be particularly appealing as it can focus all the aforementioned benefits while minimizing and removing many of the liabilities such as competing reactions, the requirement for cross-membrane transport, and cellular toxicity.These formats typically consist of just the enzymes, substrates, cofactors, and a conducive reaction environment.Given that each enzyme present in a cascade will have a kinetic profile that differs from the others, and sometimes quite substantially so, focus is often put on balancing the ratio of enzymes to each other to better match kinetics and optimize the cascaded flux through the system.Within cells, this is accomplished by the careful control of relative enzyme expression levels in response to metabolic needs.Aside from initial enzyme expression and purification requirements, the primary issues with utilizing multienzyme cascades in such minimalist formats now become that of long-term enzyme stability and diffusional limitations and it is here where surface attachment or (nano)scaffolding along with access to enzyme channeling phenomena can become important facilitators.

Enzymatic Enhancement on Nanoscaffolds and Channeling
Displaying enzymes on planar surfaces or macroscale particles generally increase their long-term stability and/or reusability, but this is invariably achieved at the cost of a loss of catalytic efficiency. [35,36]This loss is partially attributable to the heterogeneous chemistry used to attach the enzymes to the materials along with laminar flow around the surface which gives rise to formation of stagnation layers that function as effective separation boundaries.In contrast, enzyme activity often appears to be enhanced or accelerated when an enzyme is attached to, or displayed on, some type of nanomaterial, including nanoparticles (NPs). [37,38]Although there is no fixed size-cutoff, the enhancement appears most often with smaller NPs of <30 nm diameter. [39]Studies with the enzyme phosphotriesterase (PTE) as displayed around the surface of semiconductor quantum dots (QDs, <10 nm diameter) revealed that enhancement (e.g., 4× increase in k cat ) occurred because the enzyme's rate-limiting step (enzyme-product release, EP → E + P) was alleviated by the unique nano-environment found around the NPs. [40,41]Moreover, this phenomena appears to be size-dependent as enhancement peaked with 10 nm diameter NPs and diminished as the NP size increased from there. [39]The environment around a NP still remains almost completely uncharacterized beyond seminal findings that colloidal NPs universally structure their surroundings leading to ionic, pH, and density gradients amongst other unknown physicochemical changes that can extend, in some cases, to approaching 2× the NP diameter. [42][49][50] However, here the underlying contributing factors are more complex and appear to be associated with the influence of DNA's strongly charged environment leading to localized sequestration of substrate or intermediaries along with perhaps making the reaction environment more favorable in certain cases. [51,52]There has been, and still continues to be, considerable scientific debate around the origins and complexity surrounding such phenomena. [53,54]revious studies of enzymatic enhancement on NP scaffolds in the context of a coupled enzymatic system revealed that enzyme stabilization could also contribute to the observed catalytic increases.Studies of pyruvate kinase (PykA) and lactate dehydrogenase (LDH) as a coupled catalytic system when assembled to QDs revealed multiple phenomena at play. [55][58] As obligate tetramers, PykA and LDH both display 4-distal (His) 6 sequences on their periphery and characterization of their QD assembly revealed that they cross-linked with the QDs into self-assembled The functionality or mechanism behind cascaded enzyme structures can be grouped into four primary categories: linear, orthogonal, parallel, and cyclic.In linear cascades, one substrate is ultimately converted to one product through multiple enzymes and intermediates.In orthogonal cascades, a cosubstrate or cofactor is generated by a second enzyme or a second enzyme removes the undesired byproduct.In parallel cascades, two reactions are coupled by cosubstrate or cofactor use; importantly, in this case both products are desired, though one could potentially consider the orthogonal cascade as just a combination of the linear and parallel cascades.In cyclic cascades, from a mixture of substrates (e.g., D and L enantiomers), only one type (e.g., the D enantiomer) is selectively converted to an intermediate and subsequently converted back to a mixture of substrates (e.g., D and L enantiomers); over multiple cycles, the nonreactive substrate (e.g., the L enantiomer) accumulates.The second step can be chemical instead of enzymatic, making the process a chemo-enzymatic cascade.Reproduced with permission. [31]Copyright 2019, American Chemical Society.

Figure 3.
Potential benefits suggested from assembling artificial multienzyme cascades on scaffolds.A) In this example of an artificial multienzyme scaffold, enzyme 1 (E 1 ), with its substrate (S) and product (the "intermediate," I), is in a cascade with enzyme 2 (E 2 ), which takes the intermediate as its substrate and makes the final product (P).The intermediate can either travel directly between the enzymes (f direct ) or diffuse from E 1 into bulk solution before encountering E 2 (f bulk ).The radius of the enzyme (r), the distance between enzymes (d e-e ), the type of scaffold, and the type of attachment (e.g., Tag/Dock) can all play a role in accessing the benefits from artificial multienzyme scaffolds.B) Overview of some potential mechanisms responsible for the benefits of artificial multienzyme scaffolds.Reproduced with permission. [31]Copyright 2019, American Chemical Society.
nanoclusters. [38]Studies of the PykA-LDH system revealed that QD display and cross-linking increased LDH activity by >50-fold and total turnover by at least 41-fold largely due to stabilization of its quaternary structure at low concentrations (≤1 nm) where the native tetrameric enzyme normally dissociates and falls apart (≈10 nm).This nanoclustering of both the enzymes with the QDs provided for a high-localized density of enzyme and soft confinement effects that also facilitated access to inter-enzyme channeling.
Channeling refers to any mechanism in a coupled or multienzyme cascade that limits out-diffusion of a reaction intermediate from an upstream enzyme while also increasing the probability that it will productively encounter the next enzyme.Channeling is a nanoscale phenomenon observable in the diffusionlimited regime whenever the "effective" multienzyme catalytic rate >> diffusion rate. [59,60]Accumulating evidence suggests that intracellular enzymes can similarly associate into complexes, referred to as metabolons, that can significantly increase the efficiency of a given metabolic pathway by engaging in substrate channeling between enzymes. [61,62]These metabolons are believed to increase the overall pathway's flux by decreasing the concentration of required enzymes, increasing concentration of intermediates in a localized manner, overcoming non-optimal kinetics, and minimizing the escape of intermediates amongst other benefits that arise from having a high local concentration of substrates and enzymes in a very small volume.The engineering of synthetic metabolons usually requires that the enzymes are first genetically fused to some type of assembly or scaffolding protein.When these are then expressed in a microorganism, the scaffolding forces the enzymes to assemble jointly into the functional metabolon architecture.Research has shown that such approaches can significantly increase intracellular production of target chemicals.In one prominent example, Dueber engineered a designer-modular protein scaffold to host three sequential biosynthetic enzymes associated with the mevalonate pathway at differing stoichiometry to better match catalytic rates and increased the mevalonate product titer 77-fold from acetyl-CoA substrate in E. coli. [63]This product was meant as a precursor for production of artemisinin, a key anti-malarial drug (see also Section 2.3 on amorpha-4-11-diene below).In another prominent example, Castellana assembled linked functional enzyme co-clusters intracellularly in agglomerates to gain control over a metabolic branch point between pyrimidine and arginine biosynthesis by directing synthetic flux through one arm of a metabolic branch point. [62]The benefits of channeling have also been demonstrated in completely in vitro minimalist systems and Figure 3 provides a schematic overview and a listing of some of the benefits of such focused approaches. [31]Overall, the salient points to be appreciated here are that nanoscale/nanomaterialbased approaches to stabilize enzymes, increase their kinetic properties, and significantly increase their coupled activity in an artificially assembled cascade by accessing channeling all represent ways to gain more product with less material in artificial reactions.Indeed, this may represent an excellent approach to increase the activity of the minimalist multienzymatic systems we now review.A brief overview of current nanoscaffolded systems for hosting multienzyme cascades is then provided to help look forward to how these systems can be concomitantly developed with SynBio to be most efficient in the future.

Representative Multienzymatic Cascades for the Synthesis of Novel Compounds and Those Previously Not Produced by Enzymes
In the following sections, we utilize representative examples from the literature to highlight the development of designer (primarily in vitro) multienzymatic cascades and discuss what specifically they can provide as compared to classical organic chemistry approaches in their pursuit of either new molecules or those that cannot be undertaken in cells or in cell extracts efficiently.Being representative, they certainly do not encompass most of the work in this space, but rather function as a cross-section.For these purposes, we focus on multienzymatic cascades that have been assembled for the synthesis of compounds not previously produced by enzymes while a novel compound in this biosynthetic context refers to one that has not been previously made by enzymes or even chemically before.The latter may include de novo biosynthesis where a complex molecule is assembled from simple starting materials.We begin with representative two-enzyme systems with a focus on control over chirality and derivation of high value compounds.This is followed by three-and four-enzyme systems, which are highlighted by examples of specific, focused multistep chemistries.Chirality and site-specific substitution also play a part in some of these latter synthetic approaches.Obviously, the number of available examples decrease with increasing number of enzymes being utilized in a given system as does the inherent complexity and number of variables that are required to initially set up and then optimize synthesis in the latter.Lastly, it is important to keep in mind that most of these biosyntheses are either extremely inefficient or not feasible at all in cell-based systems for a variety of reasons.

Synthesis of Enantiopure Chiral Amines
Chiral amines are desirable intermediates for the production of pharmaceuticals, fine chemicals, polymers, plasticizers, and agrochemicals, amongst other target molecules.Chiral amines are also involved in organic synthesis and organocatalysis. [64]For example, the Noyori catalyst is an asymmetric hydrogenation catalyst famous for the enantioselective reduction of ketones, which consists of an enantiopure diphenylethylenediamine within the inner coordination sphere of the ruthenium (Ru) catalyst. [65]Additionally, (R)-(+)--methylbenzylamine is used as a resolving agent and is required in the synthesis of phosphonomycin, an antibiotic used for the treatment of urinary tract infections. [66]n pharmaceutical synthesis, chirality is important to the final desired structure, function, and activity of the target compound since most isomers of chiral drugs manifest significant differences in biological activities including their pharmacological, toxicological, pharmacokinetic, and metabolic profiles. [67]For example, the -blocker propanolol is administered as a racemic mixture, but the S(−) isoform is 100× more active than the R(+) isoform. [68]In another prominent example, for the drug used to treat Parkinson's disease, dihydroxy-3,4-phenylalanine (DOPA), the L-isoform is therapeutically relevant while the D-isoform is highly toxic, resulting in agranulocytosis, and thus L-DOPA must be strictly purified from any of its D-isoform. [69]equisite chiral amines are scarce in Nature and their industrial synthesis from primary and secondary alcohols typically require expensive and unstable transition metal complexes, which are reacted under harsh conditions with only moderate conversion rates and enantioselectivity, the latter being the major requirement for optically active drugs. [70]Furthermore, multiple protecting/deprotecting steps are required during the synthesis of chiral amines, which increases the number of steps and complexity, subsequently generating copious amounts of hazardous solvent wastes.Seeking to address these and other related issues with a "greener" biological approach, Mutti et al. designed a bienzymatic cascade for the synthesis of enantiopure chiral (R)-amines from a structurally diverse set of aromatic and aliphatic primary and secondary alcohols (Scheme 2).This enzymatic cascade is based on exploiting an alcohol dehydrogenase (ADH, Enzyme Commission number or EC 1.1.1.1,Scheme 2, E1) for the interconversion of ketones and alcohols and an amine dehydrogenase (AmDH, EC 1.4.9.1, Scheme 2, E2) to facilitate the asymmetric reductive amination of ketones and aldehydes in the formation of the chiral amine product. [71]or this study, the authors wanted to test a variety of aromatic and aliphatic primary and secondary alcohols for conversion to their respective chiral amines to demonstrate the utility and versatility of the system.In the first step, an Aromatoleum aromaticum ADH (AA-ADH, MW ≈ 27 kDa, tetrameric) and an engineered ADH from Lactobacillus brevis (LBv-ADH, MW ≈ 7 kDa, tetrameric), and two NAD-dependent ADHs that display specificity for (S)-and (R)-alcohols, respectively, were used. [72,73]For the second step, an engineered phenylalanine dehydrogenase (Ph-AmDH) variant from Bacillus badius was initially used in the reaction which requires NADH as a cofactor. [74]However, this was replaced with a recently published chimeric AmDH (Ch1-AmDH, MW ≈41 kDa), as its substrate scope was determined to be much broader, including benzylic carbonyl substrates. [75]h1-AmDH was generated by domain shuffling residues 1-149 of B. badius phenylalanine dehydrogenase (PheDH, EC 1.4.1.20)and residues 140-366 from Bacillus sphaericus leucine dehydrogenase (LeuDH, EC 1.4.1.9).The reactions tested proceeded successfully to completion under physiological conditions requiring only ammonia as the nitrogen source and generated water as the sole by-product confirming their "green" profile.Moreover, the hydrogen-borrowing character of this coupled reaction meant it proceeded through a self-sustaining regeneration of the NADH cofactor.Notably, in later studies the Turner group engineered a single, non-selective (R/S)-ADH from Scheme 2. Two-step bienzymatic cascade for synthesis of enantiopure chiral amines from diverse aromatic, aliphatic and 1°and 2°alcohols. [71]Generalized scheme showing the coupled reactions of ADH (E1) and AmDH (E2) in the synthesis of enantiopure chiral amines.
Thermoanaerobacter ethanolicus (TeSADH W110A) which was capable of oxidation of racemic chiral secondary alcohols, alleviating the need to use two enantiocomplementary ADHs, and extending the utility of the cascade. [76]mpressive conversion rates of up to 96% and >99% enantiomeric excess (ee) could be achieved for the biosynthesis of (R)amines from structurally diverse sets of (S)-and (R)-secondary alcohols using either AA-ADH or LBv-ADH, respectively, along with Ch1-AmDH (Table 1).Starting with 1-phenyl-2-propanol (Table 1, entry 1), this substrate achieved excellent conversion rates and high ee.Various substitutions on the benzyl ring (Table 1, entries 2-5) resulted in reduced conversion rates, yet still maintained high stereoselectivity and ee values, with the exception of methoxy substitution at the meta-position of the benzyl ring (Table 1, entry 5), which had a significantly reduced ee (82%).Importantly, it seemed positioning of the benzyl ring relative to the alcohol affected ee values (Table 1, entries 6-8), where, and comparing to entries #1, 7, and 9, at the -position to the benzyl ring (Table 1, entry 7) resulted in significant loss of enantioselectivity (ee ≈ 82%), although this could be alleviated by having an ether in the -position to the benzyl ring (Table 1, entry 8).When using 1-phenylethanol and benzyl substituted derivatives (Table 1, entries 9-14), conversion rates were significantly decreased, while enantioselectivity remained exceptionally high (ee > 99%).Varieties of aliphatic secondary alcohols (Table 1, entries 15-19) were also assessed using this system.Again, exceptional enantioselectivity was achieved (ee > 99%) for all substrates tested, but conversion rates varied based on the alkyl chain length, with the longer alkyl chains performing better than those with shorter alkyl chains.Additionally, by combining both ADH enzymes in the reaction, racemic mixtures of the secondary alcohols could be used as substrates for the complete conversion to enantiopure chiral amines.
To further demonstrate the utility of this system, the authors also tested a thermostable ADH (hT-ADH, MW ≈ 37 kDa, tetramer) from Bacillus stearothermophilus, which has selectivity toward primary alcohols. [77]Combining hT-ADH with either Ph-AmDH or Ch1-AmDH, quantitative conversion of a variety of aliphatic primary alcohols (Table 1, entries 20-26) to the respective amines were achieved.While alcohols with chain lengths ranging from heptanol to butanol (Table 1, entries 21-25) demon-strated high conversion rates, longer alcohol chains such as octanol (Table 1, entry 20) and aromatic primary alcohols such as 2-phenylethanol (Table 1, entry 26) were poorly converted (8% and 10%, respectively).These data and other similar examples also provide another important contribution to this field and that is in the mapping of substrate specificity along with conversion efficiency and chiral yield as a function of a specifically sourced enzyme; this will contribute greatly to establishing retrobiosynthetic efforts for non-natural products (vide infra).
More recently Liu et al., improved upon a major limitation of the previous ADH/AmDH system. [78]While conversion rates and ee were exceptional depending on the enzyme/substrate pairs used, reaction rates were rather slow.This was determined to be inherent to the different catalytic efficiencies of ADH (k cat /K M ≈ 10 4 -10 5 m −1 s −1 ) and AmDH (k cat /K M ≈ 10 2 -10 3 m −1 s −1 ).This resulted in limited recycling of the critical NAD + /NADH cofactor by the much slower AmDH catalytic step.Additionally, the TTN was determined to be 76 for the in vitro system above.Liu overcame this limitation and enhanced cofactor recycling by using a combination of both whole-cell and cellfree systems to separately regenerate the NAD + /NADH cofactors (Scheme 3).In this design, both an (S)-selective (CpsADH) and (R)-selective (PfODH) ADH (Scheme 3, E1) were expressed and compartmentalized within E. coli T7 (DE3) cells where intracellular NADH oxidase (NOX, EC 1.6.3.4,MW ≈ 50 kDa, dimer) was also expressed to regenerate intracellular NAD + during the initial oxidation step (Scheme 3, E2). [78]Purified AmDH (Scheme 3, E3) and glucose dehydrogenase (GDH, EC 1.1.1.47,MW ≈ 28 kDa, tetramer) were then added to the extracellular supernatants to complete the biocascade reaction, where the presence of GDH allowed for the regeneration of NADH (Scheme 3, E4).In this study, the authors decided to use racemic mixtures of 4-phenyl-2-butanol or 1-phenyl-2-propanol as substrates for the bioconversion to their respective (R)-chiral amines.Both are important intermediates for the synthesis of the drugs (R,R)formoterol, tamsulosin, and dilevalol, which are used as antihypertensives and beta-blockers.Using the one-pot biocascade system, E. coli cells at 2 g cell dry weight (cdw L −1 ) expressing a trio of enzymes (CpsADH-PfODH-NOX) were mixed with purified AmDH (9 mg mL −1 ) and GDH (2 mg mL −1 ).This mixture was then supplemented with 0.5 m NH 4 Cl, 20 mg mL −1 Table 1.Substrate panel tested for the conversion of the stereospecific or racemic alcohol mixture to their respective (R)-chiral amines. Reaction: Two enzyme cascades consisting of ADH (E1) and AmDH (E2).All reactions were conducted at 30 °C for 48 h.ee %-percentage enantiomeric excess. [ 71]v.Mater.2024, 36, 2309963 © Published 2023.This article is a U.S. Government work and is in the public domain in the USA.Advanced Materials published by Wiley-VCH GmbH Scheme 3. Two-step coupled cellular-cell free bienzymatic cascade for synthesis of enantiopure chiral amines.Generalized scheme showing the coupled reactions where ADH (E1) and NOX (E2) functioned intracellular to convert racemic alcohols to their respective ketones, while AmDH (E3) and GDH (E4) functioned extracellular for the conversion of the ketones to their respective chiral (R)-amines.Using this cell free system, conversion rates of 79% with an ee of 99% could be obtained within 8 h and a TTN for NADH recycling of 350.Immobilization of AmDH and GDH to nanoparticles improved conversion rates up to 84% with TTN for NADH as high as 1410. [78]ucose, 1 mm NADH and 10 mm 4-phenyl-2-butanol to begin the cascade reaction to produce the chiral amine.Under these conditions, conversion rates of up to 79% were achieved within 8 h.By reducing the concentration of NADH to 0.01 mm, conversion rates were measured up to 70% with 99% ee within 24 h, with a TTN for NADH recycling of 350, one of the highest reported values at that time.However, the conversion rates were much lower (48%) when using 1-phenyl-2-propanol, which was attributed to the lower activity of AmDH toward the reductive amination of 3-phenylpropionaldehyde.The conversion rates of 4-phenyl-2-butanol and TTN for NADH recycling were increased via immobilizing the purified AmDH and GDH enzymes onto 138 ± 40 nm diameter Ni-NTA iron oxide derived magnetic NPs using the enzyme's terminal (His) 6 motifs in a manner similar to that discussed above.These immobilized enzymes were similarly used in the one-pot cascade reaction as previously conducted with conversion rates as high as 84%, ee of 99%, and a TTN for NADH recycling of 1410 within a 12 h reaction time. [78]This one-pot immobilized cascade showed substantial improvements over the initial bienzymatic cascade, resulting in higher productivity in a much shorter period of time with one of the highest TTN ever reported for NADH recycling.It will be interesting to see how this system works with more structurally diverse sets of substrates and other ADH and AmDH pairs with different specificities to expand the scope of available substrates and possibly improve productivity for the conversion of alcohols to amines.

Synthesis of Chiral 𝛽-Halohydrins
Another group of compounds with specific interest from the pharmaceutical, natural products, and agrochemicals industry are chiral -halohydrins. [79]For example, chiral -halohydrin is a key intermediate in the enzymatic synthesis of luliconazole, a common antifungal agent used to treat athletes foot. [80][83] Currently, several chemical and enzymatic methods are already available for the transformation of -halohydrins to a diverse range of chiral compounds displaying various functionalities such as aminoalcohols, azidoalcohols, hydroxynitriles, and 1,2-diols. [79]One of the challenges associated with the chemical synthesis of -halohydrins is the requirement for costly transition metal complexes and the production of toxic organic wastes.86][87] However, one of the more significant limitations of these biocatalytic routes is the absolute requirement for prior oxygenfunctionalization at the desired C─H bond site, limiting the available substrate pool to carbonyl or hydroxylated precursors.To address this latter issue, Cui et al. exploited the enzymatic promiscuity and powerful oxidation reactions of cytochrome P450s (CYP450, EC 1.14.14.86, monomer) to asymmetrically hydroxylate the C─H bond of halohydrocarbons to generate the requisite enantioenriched -halohydrins (Scheme 4, E1). [88]Screening a panel of CYP450s from Parvibaculum lavamentivorans and their ferredoxin/ferredoxin-reductase (Fdr-Fdx) partners [89,90] led to the isolation of the CYP450 PL2 -Fdx4-Fdr combination (CYP450 PL2 -4) as being the most active toward a model substrate (2-chlorethyl) benzene when heterologously expressed in E. coli with a preference for (R)-stereoselectivity.Optimized reaction conditions for the production of 2-chloro-1-phenylethanol were determined to be pH 8.5, 35 °C, and 30 g cdw L −1 of recombinant E. coli cells, which achieved a conversion rate of 80% and an ee of 82% within 12 h.The substrate tolerance chemical space was analyzed by testing various benzyl substitutions as well as the effect of the terminal halogen group.The results summarized in Table 2 demonstrate most of the chlorohydrins (Table 2, entries 1-6) tested display more favorable conversions than their bromohydrin counterparts (Table 2, entries 7-13), regardless of benzyl ring substitution.They also discovered Scheme 4. Two-step enzymatic cascade for synthesis of chiral -halohydrins.Generalized scheme showing the coupled reactions of CYP450 (E1) and HHDH (E2) to yield chiral -halohydrins with exceptionally high ee. [88]tho-substituted derivatives (Table 2, entries 2 and 8) as well as less electron-withdrawing group substitutions (Table 2, entries 5 and 6) resulted in decreased activities.However, in all cases the ee was relatively modest, ranging in most cases from 80-90%.
To improve the ee, a second halohydrin dehalogenase (HHDH, EC 4.5.1.,monomer) enzyme was incorporated to form a one-pot cascade (Scheme 4, E2).In particular, the HheA10 variant isolated from Tsukamurella sp.1534 (MW ≈ 27 kDa) displays strong stereoselectivity toward (S)-isomer halohydrins, [91] and when included in the reaction with CYP450 PL2 -4 allowed for the enantioselective consumption of the minor (S)-halohydrin to significantly improve the ee of the (R)-halohydrin (ee 98-99%).Importantly, to avoid excessive dehalogenation of the (R)-halohydrin in the second reaction step, CYP450 PL2 -4 oxidation was allowed to first proceed for 8 h, followed by addition of whole HheA10 cell extracts.Additionally, HheA10 was previously demonstrated to have higher activity against bromohalohydrins than the respective chlorohalohydrins.Consequently, reaction times proceeded faster in the second reaction for the bromohalohydrins versus the chlorohalohydrins.A potential limitation of this strategy was apparent in differential activities between the enzyme pairs, with CYP450 Pl2 -4 having higher activities toward chlorosubstrates versus HheA10 having higher activities toward bromosubstrates.Perhaps further screening of enzymes, enzyme evolution, or directed protein engineering can lead to the development of a dual cascade that shares similar preferences and activities for Table 2. Synthesis of chiral -halohydrins from a diverse set of substituted halohydrocarbons.
Overall reaction: Cascades used CYP450 PL2 -4 (E1) or a one-pot cascade of CYP450 PL2 -4 (E1) with HheA10 (E2) and their listed conversion and ee percentages.a) Reactions conducted in 5 mL PBS buffer pH 8.5 at 35 °C with 30 g cdw L −1 CYP450 PL2 -4 expressing E. coli cells for 12 h; b) Same as in condition a), but reactions proceeded for 8 h, followed by addition of Scheme 5. Synthesis of lignin derived high value aromatic compounds.
Based on an enzymatic biocascade within whole cells expressing a phenolic acid decarboxylase (PAD, E1) and an aromatic dioxygenase (ADO, E2).
Reaction conditions optimized at a pH of 9.5, 50 °C, for 2 h. [96]e substituted halogen group leading to improved conversion rates and ee.

Synthesis of Lignin Derived High Value Aromatic Compounds
Lignocellulosic biomass in its different forms represent one of the largest bioindustrial generated waste products, originating primarily from biorefineries along with the paper and wood pulp industries.In the United States, it has been estimated that there exists over 340 million tons of lignocellulosic biomass available for sustainable energy production. [92]Additionally, current waste disposal methods of lignocellulosic biomass involve landfill disposal and/or incineration due to the lengthy pretreatment and energy intensive procedures required to convert lignocellulosic biomass to a sustainable energy source.This biomass is comprised of primarily polysaccharide (cellulosic) and lignin polymers.The monomeric units of lignin have great potential for use as biorenewable aromatic materials for the production of bioplastics and biofuels. [93]Such monomeric units can be efficiently acquired by thermochemical depolymerization or enzymatic hydrolysis. [94,95]However, chemical methods of transforming lignin monomers to high value chemicals from lignocellulose require large energy inputs and bulk solvents, which generate excessive amounts of waste and add considerable cost to the process.Greener alternatives using biocatalysis are being explored but require initial physical breakdown of the mass itself and large inputs of ATP and coenzyme A (CoA) as cofactors, which are exceedingly costly, making large-scale expansion of such approaches economically non-viable at the current time.To potentially address this roadblock, Ni et al. recently assembled and tested a biocatalytic cascade that negates the requirement for the CoA cofactor by using a phenolic acid decarboxylase (PAD, EC 4.1.1.102,monomer) and an aromatic dioxygenase (ADO, EC 1.14.12., dimer) to transform lignin-derived aromatics into highvalue aromatic aldehydes (see Scheme 5). [96]n this chemistry, PAD catalyzes a non-oxidative decarboxylation and ADO catalyzes the oxidative cleavage of C═C double bonds in aromatic olefins.The predominant monomers present in lignin are ferulic acid and p-coumaric acid and Ni's enzymatic cascade was meant to convert them to vanillin and 4-vinylphenol, respectively.A previous report identified a carotenoid oxygenase (Cso2) that could convert the vinylphenol intermediate to the desired aromatic aldehyde; however, it was found that the catalytic efficiency was extremely low as the enzyme suffered from temperature sensitivity and further required an alkaline environment to function optimally. [97]Using a bioinformatics-driven approach, ADO-like enzymes with high similarities to Cso2 were discovered, and one in particular from the thermophilic fungus Thermothelomyces thermophila (ADO, MW ≈ 67 kDa) that demonstrated a remarkable catalytic efficiency of 1.2 mm −1 s −1 ; this is roughly 78 500-fold greater than Cso2 when converting the vinylphenol intermediate to the corresponding aromatic aldehyde.Furthermore, this enzyme demonstrated suitable thermostability over a temperature range of 40-50 °C, which was exploited for the deactivation of redundant alcohol dehydrogenases present when expressing the ADO enzyme in E. coli to accumulate the aromatic aldehydes.Identification of a suitable thermotolerant PAD enzyme was also identified via bioinformatic analysis.A putative PAD was discovered in the thermophilic bacterium Bacillus coagulans (BcPad, MW ≈ 22 kDa) and was subsequently shown to be highly thermostable (40-50 °C) and demonstrated exceptional catalytic efficiencies (58.4 mm −1 s −1 ) toward aromatic lignin compounds. [96]At face value, these significant enzyme efficiencies make these enzymes more desirable and appealing for direct transition to bulk industrial scale up of this chemistry.
In initial experiments, these two enzymes were co-expressed in E. coli cells, which were fed the lignin monomer substrate for conversion to the aromatic aldehyde.When supplemented with 30 mm ferulic acid, only 7.1 mm of vanillin was detected when reacted for 2 h at 30 °C and a pH of 7.5.However, the formation of vanillyl alcohol was detected in significant quantities (9.7 mm), presumably by the further reduction of vanillin by endogenous ADHs.Fortuitously, increasing the reaction temperatures to 50 °C and conducting the reaction at an alkaline pH of 9.5 resulted in the inactivation of the endogenous ADH enzymes while simultaneously enhancing catalysis by ADO (rate-limiting step).Under these conditions, 30 mm of ferulic acid was converted to 27.7 mm vanillin within 2 h, with only 0.2 mm of vanillyl alcohol generated.This corresponded to a product conversion of 92.3% and an overall productivity of 2.1 g L −1 h −1 , representing one of the highest reported values for a whole-cell biocatalyst.Furthermore, lignin waste products have been reported to release ferulic acid and p-coumaric acid at ratios of 1:1.8 after alkaline hydrolysis.Using a mixture of 30 mm ferulic acid and 54 mm p-coumaric acid (1:1.8),27.5 mm vanillin and 48.7 mm 4vinylphenol were generated, corresponding to conversion yields of 91.7% and 90.2%, respectively.This clearly confirms the feasibility for simultaneous production of value added aromatics from predominantly hydrolyzed lignin aromatics albeit at the initial bench or reaction scouting point.
To improve upon the biocascade and demonstrate scalability, a biphasic organic/aqueous system was developed for the biosynthesis of lignin-derived aromatics.The inclusion of water immiscible organic solvents allows for the partial extraction of the hydrophobic products vanillin and 4-vinylphenol to potentially alleviate issues with toxicity, prevent substrate inhibition, as well as drive reaction conditions toward product formation.Using this biphasic system, larger scale 1 L bioreactors were used to generate 13.3 g L −1 h −1 vanillin and 20.5 g L −1 h −1 4-vinylphenol within 18 h.To further show the versatility of this system, a variety of aromatic compounds were also incorporated as substrates and were demonstrated to produce a variety of high value aromatics with varying productivities.Not only does this study show very impressive product formation efficiency, it also demonstrates an important next step in the transition of these technologies, namely the 3 orders of magnitude scaling up from μL to mL and then L volume reactions.We note that scaling up these types of reactions toward industrial production is a nontrivial challenge and is currently viewed as more of an art form than science given the lack of concerted industrial experience in the field. [98,99]

Other Two-Enzyme Cascades
The studies outlined above are specific examples of two-enzyme cascades, which are not only used for the first time in this context but also demonstrate structurally diverse chemistries and strategies toward biosynthetic production of a desired target product.[102][103][104][105] For example, Wu et al. previously developed a two enzyme cascade for the production of chiral vicinal diols with both high ee and high yield across a variety of aryl olefins. [100]Further, Li et al. utilized methods of directed evolution to develop two-enzyme cascade reactions which could selectively produce three stereoisomeric cyclohexane-1,2diols. [105] Of all the different types of multi-enzyme cascades, twoenzyme cascades remain the most abundant in the literature to date likely due to overall system simplicity relative to more complex cascades.Still, it is important to highlight the contributions that such two-enzyme cascades provide as they represent a starting point, are the most tractable to set up, characterize, and optimize, and, as such, they lay the foundation for more complex systems.

Synthesis of Para-Vinylphenols
Para-vinylphenol derivatives are an interesting class of bulk chemical feedstock that comprise a range of useful scaffolds for further synthesis.p-vinylphenol has found use in pharmaceuticals, chemical and biological sensors, and fire retardants. [106]urther, homopolymers and co-polymers of p-vinylphenol have many uses, including in the manufacture of metal treatments and in photoresistors.A common chemical synthesis of p-vinylphenol consists of five steps, starting with phenol, which is twice acetylated to p-acetoxyacetophenone and then hydrogenated to p-acetoxyphenylmethyl carbinol, dehydrated to p-acetoxystyrene, and finally, saponified to p-vinylphenol using potassium hydroxide. [107]Methods for synthesizing parahydroxystyrene derivatives are limited and typically require a functional group in the para position; they usually focus on transforming p-halogenated phenols by relying on transition-metal catalysts in a variety of different reaction approaches to yield the corresponding aldehydes.The only alternative or "green" approach is limited to enzymatic decarboxylation of cinnamic acid derivatives prepared from the same type aldehydes. [108]Since no single enzyme could accomplish all this, Busto et al. constructed a three-enzyme cascade to produce a variety of p-vinylphenols using tyrosine phenol lyase (TPL, EC 4.1.99.2), tyrosine ammonia lyase (TAL, EC 4.3.1.23),and ferulic acid decarboxylase (FAD, EC 4.1.1.02)based on the findings of a previous biocatalytic retrosynthetic analysis (Scheme 6). [109]The authors chose to utilize TPL variant M379V from Citrobacter freundii (≈51 kDa, tetramer) over wild-type TPL as a previous study demonstrated its larger substrate acceptance and ability to produce non-natural tyrosine derivatives. [110]For initiating this cascaded reaction, TPL was used in the form of a cell-free extract, where the TPL was overexpressed in E. coli BL21(DE3), the cells were lysed, and the supernatant was used for the first step.For the subsequent two steps of the cascade, wild-type forms of both TAL from Rhodobacter sphaeroides (≈55 kDa, tetramer) and FAD from Enterobacter sp.(≈19 kDa, dimer) were used; both were overexpressed in E. coli BL21(DE3) and used as freeze-dried, whole-cell preparations.Each of these cell-free preparations were rehydrated in buffer, with pyruvate and NH 4 Cl available for initiating the cascade, and incubated with the input phenol substrate at 30 °C for 24 h.
In the one-pot reaction, the regioselective TPL carries out a C-C coupling at the para position between the phenol and pyruvate resulting in the formation of an L-tyrosine derivative, which occurred within the first hour under the described conditions.This tyrosine derivative is then deaminated by TAL, which produces a coumaric acid derivative and ammonia, the latter of which is recycled by TPL in the process of creating the tyrosine derivative.Finally, the coumaric acid derivative undergoes decarboxylation by FAD resulting in the desired p-vinylphenol.Evaluating the effectiveness of this assembled cascade, the authors found that each of the eight phenol substrates they tested were successfully converted to the desired product, each with a conversion rate of >99% (Table 3).Overall yields ranged between 65% and 83% and required reported reaction times of 48 h.Interestingly, both yields and conversion rates remained similar regardless of the substitution location or number (2-, 3-, or 2,3), suggesting a degree of flexibility in the overall cascade.The group noted that halving the amount of input pyruvate under their conditions from 46 to 23 mm significantly reduced the concomitant conversion rate.Importantly, since the ammonia was recycled, the only byproducts of this one-pot reaction were water and CO 2 , while byproducts in the traditional chemical synthesis are obviously more numerous and toxic. [111]

Synthesis of Cladribine Triphosphate
Biosynthesis of non-natural deoxyribonucleotides such as cladribine triphosphate is challenging but of great interest because this class of molecule is frequently identified as having anti-cancer properties or other important therapeutic attributes such as treating multiple sclerosis. [112]For treating lymphoma, the drug is used as an adenosine mimic that is adenosine deaminase (ADA) resistant.Following administration, it is taken up by cells, including lymphocytes, and phosphorylated into the pharmacologically active triphosphate form that is incorporated into chromosomes during cell division and DNA repair; the subsequent inhibition of replication is its pharmaceutical mechanism of action. [113]The synthesis of non-natural deoxyribonucleotides is not trivial and this is directly confirmed in this example where the estimated cost of synthetic product is ≈90 000 USD g −1 (at the time of writing).Scheme 6. Three-enzyme cascade for producing para-vinylphenols.Substituted phenols and pyruvate are converted to a tyrosine derivative by TPL (E1), and then converted to a coumaric acid derivative and ammonia by TAL (E2), and finally decarboxylation by FAD (E3) yields the desired vinylated phenol.The reactions described by Busto et al. [111] include functional groups at 2-, 3-, and 2,3-positions, displayed in the figure inset (bottom left).
Frisch et al. assembled a one-pot synthesis for producing cladribine triphosphate via the enzymes adenine phosphoribosyltransferase (APT, EC 2.4.2.7), polyphosphate kinase (PPK, EC 2.7.4.1), and ribonucleotide reductase (RNR, EC 1.17.4.1). [114]heir gene sequence for APT (dimer) was obtained from E. coli, PPK (tetramer) from Meiothermus ruber, a Gram negative non-motile rod-like bacteria, and RNR (monomer) from Thermus virus TV74-23, a bacteriophage.Each of the genes were encoded on a plasmid, cloned into E. coli for expression, and purified by encoded C-terminal His-tags.Molecular weights for each protein are ≈20, 35, and 74 kDa, for APT, PPK, and RNR, respectively.The four-step reaction shown in Scheme 7 first catalyzes the loading of the halogenated adenine (2Cl-adenine) onto the ribose of phosphoribosyl pyrophosphate (PRPP) to form 2Cl-adenosine via APT.Following this, PPK performs two consecutive reactions yielding 2Cl-adenosine triphosphate.Finally, RNR leads the reduction to the final product 2Cl-deoxyadenosine-triphosphate (called cladribine triphosphate) which is the drug in its active triphosphorylated form without need for further modification.However, minor levels of side products (other forms of 2Cladenosine and cladribine) were also noted to be present by the end of the reaction in their experiments confirming the need for DSP and careful purification.After a 150 min incubation at 40 °C, a final reaction yield of ≈80% was achieved for cladribine triphosphate with cladribine diphosphate being one major side product; in their conditions ≈800 μm cladribine triphosphate and ≈200 μm cladribine diphosphate, while other products were detected at negligible levels.
Several limitations in this initial proof-of-concept cascade were noted.Although APT does convert the non-natural substrate 2Cl-adenine, a 100× higher enzyme concentration was needed to convert 2Cl-adenine to 2Cl-adenosine monophosphate than Table 3. Conversion percentages for the three-enzyme cascade for producing p-vinylphenols.
Overall reaction: Substituted phenols and pyruvate are converted to a tyrosine derivative by TPL (E1), then converted to a coumaric acid derivative and ammonia by TAL (E2), and finally decarboxylation by FAD (E3) yields the desired vinylated phenol. [ 111]v.Mater.2024, 36 . [114] required to carry out its natural reaction (adenine to adenosine monophosphate).The equivalent calculation for PPK and RNR were 10× and 2.5×, respectively.Another area for improvement highlighted was the expensive co-substrate PRPP (≈3000 USD g −1 ), which contributes to the high synthesis cost for non-natural deoxyribonucleotides listed above, a cost that may be offset by expanding to a five-enzyme cascade with the addition of ribokinases and PRPP synthetases such that PRPP is produced in situ.The relative promiscuity of both PPK and RNR, suggests that the cascade can likely be used to produce compounds other than cladribine triphosphate, since both enzymes accept different nucleotides, including pyrimidines, [115] and the adenosine-specific APT can be swapped out for other phosphoribosyltransferases, suggesting the system can be modified to produce a range of other nucleotide-based drugs.Overall, the above is an excellent perhaps even emblematic example of exploiting enzymes to biosynthesize a fine chemical product-an active drug molecule-that is extremely hard to make with synthetic chemistry and also in cell-based SynBio systems.Although still expensive, the enzymatic product represents a significant reduction in cost to produce and ultimately provide to patients.Moreover, as mentioned coupling this cascade to others that produce the cofactors or perhaps even recycle them would reduce the cost even more.Given the number of competitive sugar and ribonucleotide pathways present in a cell, yields would likely be lower in vivo.Additionally, DSP and purifying this product from amongst the complex milieu of other cellular sugars and nucleotides would be extremely difficult.Lastly, being able to test and incorporate a variety of other substituted and modified substrate molecules into this cascade could lead to far more specific and potent drug molecule leads.The latter is not something that could be undertaken in cells due to toxicityonce again serving to highlight the unique possibilities that are driving the strong interest in this research area.

Synthesis of Chiral Substituted Piperidines and Pyrrolidines
Chiral substituted piperidines and pyrrolidines are important chemical intermediates used for the production of many different pharmaceuticals, such as a 1,3,3,4-tetrasubstituted pyrrolidinecontaining CCR5 receptor antagonist, which is used as a potent anti-HIV agent. [116]≈15% of the top 200 prescribed drugs in the United States contain intrinsic piperidines and pyrrolidines in their structures, therefore more efficient methods for the challenging targeted production of single piperidine or pyrrolidine stereoisomers are highly sought after. [117]In pursuit of a greener and more simplistic approach, France et al. exploited a linear three-enzyme artificial cascade to produce a wide variety of differentially substituted piperidines and pyrrolidines using carboxylic acid reductase (CAR, EC 1.2.1.30),-transaminase (-TA, EC 2.6.1.18),and an imine reductase (IRED, EC 1.5.1.48)(see Scheme 8). [118]They obtained the CAR (128 kDa, monomer) from Mycobacterium marinum, (R)-IRED (≈31 kDa, dimer) from Streptomyces sp.GF3587, and the (S)-IRED from Streptomyces sp.GF3546.The two IREDs were identified by Mitsukura et al. via screening for reduction of pyrroline with high conversion and enantioselectivity; they have 36% identity. [119]The -TA was a commercially available transaminase called ATA-113 as obtained from Codexis.It is important to appreciate how this initial cascade setup, along with several already mentioned above, highlight the strong role that bioinformatics and classical Scheme 8. Production of chiral substituted piperidines via a three-enzyme cascade.Keto acid substrate is bound by carboxylic acid reductase (CAR), reducing to keto aldehyde which is converted by -transaminase (TA) to an imine.The imine is then converted to the final 2,x-substitued piperidine (n = 1) or pyrrolidine (n = 0) by either (S)-imine reductase (IRED) or (R)-IRED.Chirality of the product is dependent on the (S) or (R) variant of the IRED used. [118]able 4. Results for chiral substituted piperidines via a three-enzyme cascade.
Keto acid substrate is bound by CAR (E1), reducing to keto aldehyde which is converted by TA (E2) to an imine.The imine is then converted to the final 2,x-substitued piperidine (n = 1) or pyrrolidine (n = 0) by either (S)/(R)-IRED (E3). [ 118]ochemistry in the form of enzyme activity screening assays play in these processes.
As shown in Scheme 8, CAR converts the keto acid into a keto aldehyde, which is then converted to an amine following transamination of the aldehyde group using -TA.After this step, the formed amino ketone spontaneously cyclizes into the imine, which subsequently undergoes reduction to finally yield either a chiral substituted piperidine or pyrrolidine.Using five different simple keto acid substrates, the group showed that a range of different useful products-2-substitued piperidines in their examples-all could be made using the same three-enzyme cascade.Both CAR and variants of IRED were utilized within whole cells while purified -TA was added exogenously.While initial attempts performed the reaction in a single stage, in one pot, their early results identified a problem where the activity of the transaminase was inhibited in the presence of IRED-containing whole-cells.As a solution, their cascade was broken up into two sequential stages.Stage one allows for complete consumption of the starting keto aldehyde by the -TA (24 h) and the second stage adds IRED whole-cell biocatalysts to the same pot for an additional 24 h, with both stages incubated at 30 °C with agitation.In experiments employing either the CAR--TA-IRED or -TA-IRED cascades, depending on the input substrate, their reported yields were between 75% and 95% corresponding to final product quantities of ≈50-75 mg, although their determined conversion percentages ranged between 20% and >90% depending on the substrate.Notably, because of their use of either (S)-IRED or (R)-IRED, they demonstrated stereoselective reduction of the generated cyclic imines, where, depending on which IRED-containing cell was used (S or R), they could produce opposite enantiomers of the piperidines.Beyond this mechanism of control, both IRED enzymes were shown to accept a variety of imines, making the process flexible.Importantly, in most of their reported experiments both enantiomeric excess and diastereomeric excess were >98%, suggesting selection of either (S)-IRED or (R)-IRED reliably produces the desired product (Table 4).Their approach of splitting the cascade into two sequential stages represents a func-tional and elegant way to address kinetic incompatibility while still achieving impressive stereoselectivity.Moreover, being able to implement this type of cascade in a scaled up manner in a minimalist format with just the enzymes needed present could allow for cost reductions and perhaps even industrial feasibility.

Synthesis of an O-Mannosyl Glycan-Based Tetrasaccharide Unique to 𝛼-Dystroglycan
Approximately half of the proteins in the human body are glycosylated and defects in either their glycosylation status or composition are associated with various diseases including muscular dystrophy. [120]Difficulty in obtaining or producing pure, homogeneous glycopeptides has thwarted both the growth of the glycoscience field and the systematic study of all glycan structures in an organism, and thus by extension this thwarts understanding of diseases associated within this biomolecule class. [121]The ability to efficiently synthesize a wide array of homogenous and pure glycopeptides with diverse sequences would greatly support research into a range of human diseases.Unlike peptide or oligonucleotide synthesis, [122,123] no universal process for chemical synthesis of structurally different oligosaccharides incorporating many different monomers is yet possible, and although one-pot chemical synthesis protocols exist for certain oligosaccharides, these remain challenging as they must be designed on a case-by-case basis. [124]o avoid chemical synthesis issues, Šardzík et al. demonstrated the synthesis of a tetrasaccharide via a three-enzyme cascade.By undertaking three consecutive enzymatic glycosylation steps, the group showed, utilizing both a one-pot reaction and on a solid phase (glycopeptide on a gold platform), that they could produce the O-mannosyl glycan NeuNAc2-3Gal1-4GlcNAc1-2Man (N-), a tetrasaccharide unique to -dystroglycan (-DG), a heavily glycosylated protein found in muscle and brain tissue. [125]This was the first described ex vivo synthesis of this particular tetrasaccharide regardless of method marking this as a novel molecule.Scheme 9. Three enzyme cascade for production of a tetrasaccharide.GlcNAc is added to the manno-threonine substrate with a 1-2 linkage by POMGnT1 (E1) producing the starting disaccharide peptide.This disaccharide is appended to with a galactose (1,4 linkage) by 1,4-GalT (E2), and finally this trisaccharide is extended to the final N- tetrasaccharide with the addition of sialic acid (2,3 linkage) by TcTS (E3). [125]r the enzyme cascade, they made use of protein O-mannose -1,2-N-acetyl-glucosaminyltransferase 1 (POMGnT1, EC 2.4.1.1),1,4-galactosyl-transferase (1,4-GalT, EC 2.4.1.38),and a transsialidase (TcTS, EC 3.2.1.18).Due to its inherent complexity, human POMGnT1 was expressed in the ascomycete (yeast) Pichia pastoris (≈75 kDa, monomer), 1,4-GalT was obtained from Bos taurus (≈45 kDa, monomer), while TcTS was obtained from Trypanozoma cruzi (≈110 kDa, monomer).
The starting point of the structure was made via a mannothreonine, with mannose linked to the side chain of threonine in an -DG fragment (amino acid residues 317−326).Using this scaffold, the natural, known enzyme for addition of the GlcNAc moiety, POMGnT1, was used for the first step in N- synthesis (Scheme 9).This addition was added with the desired -1,2 linkage, with an 85% yield.Interestingly, the authors noted that POMGnT1 activity is highly specific to peptide sequence, suggesting usage of alternative enzymes at this step may be difficult.Overcoming this may require considerable mutational optimization or enzyme evolution approaches, the point of which leads into the final part of this synthesis.For the final two steps of N- synthesis, the group used non-natural substrates as naturally derived precursor molecules from humans for this stage of the chemistry are unknown.For the second step, 1,4-GalT was used for addition of the galactose with the required 1,4 linkage, successfully added after 24 h incubation at 37 °C, with a yield of 87%.The final step was carried out by TcTS, to add the sialic acid with the desired -2,3 linkage.Although the yield for the final step was just 47%, the reaction was complete after 6 h at 37 °C.Overall, the reaction required 60 h for completion, but the group noted that this rate can be significantly increased on a smaller scale with increased enzyme concentrations.Critical for applications to other oligosaccharide syntheses, the final two enzymes are known to be flexible to other substrates, unlike POMGnT1.In addition, TcTS does not require an expensive CMP-NeuNAc cofactor (438 USD per 100 mg) and can catalyze trans-glycosylation with the less expensive cofactor fetuin, unlike other trans-sialidases.This clearly represents another example of the exquisite chemistry that enzymes are capable of along with being specifically utilized for overcoming considerable synthetic roadblocks in glycochemistry.It even suggests that careful thought and technoeconomic analysis has to be given with respect to the continued development of bulk and specialized oligosaccharide chemistries.Pursuing the development of alternative biosynthetic approaches may provide for access to a wider target chemical space and be more fruitful, dynamic, adaptable, and applicable in the long-run than trying to create the corresponding wet-chemical synthetic approaches for each target product based upon creating a library of chemically synthesized monomers.

Synthesis of Non-Canonical Amino Acids
Non-canonical amino acids (NcAAs) are essential building blocks for the synthesis of important compounds, including many drugs, or are drugs themselves.For example, NcAAs Dphenylglycine (D-PheGly) and D-para-hydroxyphenylglycine (D-pHPG) are used in the synthesis of a collection of different antibiotics, including cephalexin, amoxicillin, and ampicillin, each of which are included in the World Health Organization's (WHO) list of essential medicines.The earlier example of tryptophan analogs and their potential role as chemotherapeutics (Scheme 1) reflect another similar example. [24]Other, simpler NcAAs, including D-amino acid monomers like D-tyrosine have been suggested to trigger biofilm disruption in some bacteria, which, in turn, is known to promote antibiotic sensitivity, improving activity of existing antibiotics. [126]Beyond valuable antimicrobials, D-phenylalanine is used for production of nateglinide, a treatment for Type-2 diabetes, [127] while other NcAAs like L-5-hydroxytryptophan can directly be used for treatment of sleep terrors and other neurological issues. [128]lthough many of the products used in chemical synthesis of NcAAs, such as cyanide, are highly toxic, pursuing chemical synthesis approaches remains popular due to the cost effectiveness at the scales required for such important compounds even with low overall efficiencies.New strategies for NcAA biosynthesis have been generally highlighted as valuable, with Prof. Francis Arnold's group aiming to engineer or generate new proteins-NcAA synthases-to support effective and "green" production of NcAAs. [129]Therefore, the use of multienzyme cascades for NcAA synthesis appears to be a promising avenue in the search for less toxic and more efficient NcAA production as highlighted in the examples below.
Synthesis of D-Phenylglycine: D-PheGly is used in the synthesis of several important antibiotics, including the -lactam group antibiotics, which are also included in the WHO's list of essential medicines for treatment of a wide range of infections and especially those of the urinary tract and wounds. [130]Therefore, traditional chemical synthesis of enantiopure D-PheGly is essential for many drugs.However, the conventional synthetic process involves the Strecker reaction of benzaldehyde making use of cyanide, which is followed by crystallization with camphorsulfonic acid providing for an ≈68% yield.Other higher yielding processes have been developed but these included more complicated and expensive protection/deprotection steps. [131]To improve upon the synthetic process, Zhou et al. recently demonstrated a one-pot enantioselective synthesis of D-PheGly from several different substrates. [132]reviously, Wu et al. demonstrated the synthesis of chiral hydroxy acids, 1,2-amino alcohols, and -amino acids using a modular cascaded enzyme system, each encoded on separate plasmids for in vivo production in E. coli. [133]Zhou et al. continued pursuing this by building from a previously described enzymatic cascade using a range of substrates: racemic mandelic acids, styrenes, or L-phenylalanine derivatives for the enantioselective production of D-PheGly and derivatives.For the three enzymes, the group obtained D-PheGly aminotransferase (DpgTA, EC 2.6.1.72)from Pseudomonas stutzeri ST-201, and both (S)mandelate dehydrogenase (SMDH, EC 1.1.99.31) and mandelate racemase (MR, EC 5.1.2.2) were obtained from Pseudomonas putida ATCC 12633.Glutamate dehydrogenase (GluDT), used here for restoring or recycling glutamate in the reaction, was obtained from E. coli.Each of the enzymes was encoded onto two plasmids, SMDH (≈43 kDa, dimer) and MR (≈39 kDa, octomer) were placed on one plasmid and DpgTA (≈49 kDa, dimer) on another, and then the two plasmids were co-transformed into E. coli DL39(DE3).In their assembled system, with the two plasmids induced for expression, MR converts racemic mandelic acids to its (S)-configuration, where SMDH then oxidizes it to the second intermediate.After this, DpgTA converts this to the final product, where the glutamate cofactor is restored by GluDH, which itself requires NADPH as a cofactor.The stereoinverting L-to D-activity of DpgTA makes production of D-amino acids possible using this system in the final step.In this instance, DpgTA can be swapped out for the analogous PgaT to produce L-PheGly.Depending on the group added (F, Cl, Br, Me, or OMe) and its position, conversions ranging 58-94% were attained, with ee values ranging from 93% to 98% (Scheme 10).Although all obtained ee values were in a relatively narrow range, supporting specific chiral production, conversion rates for phenyl substitutions with o-Cl and p-Cl were particularly poor at 58% and 73%, respectively, while all other conversions were ≥80%, see Table 5.
Selective Stereoisomeric Synthesis of Trans-3-Hydroxy-L-Proline: In another three-enzyme cascade used for synthesis of a NcAA, Hara et al. constructed a sequential three-enzyme system for production of trans-3-hydroxy-L-proline from L-arginine. [134]Naturally occurring L-hydroxyproline in each of its four regioand stereoisomeric forms has been investigated as precursors for pharmaceutical agents.3-hydroxyprolines are important for the production of antibiotic compounds such as the DNA gyrase inhibitor cyclothialidine. [135]However, the synthesis of 3hydroxyproline in its trans-/L-form has not been achieved.Production of 3-hydroxyproline is generally arrived at via a combined chemoenzymatic synthesis, which is a combination of enzymatic and chemical synthetic steps used in the synthesis of a target molecule.For example, in one reported method, to reach cis-3-hydroxyproline, an intermediate in the synthesis of a Bristol-Myers-Squibb pharmaceutical candidate BMS-564929, L-proline was converted to 3-hydroxylase by a Streptomyces sp., which was then further converted by a series of chemical synthesis steps. [136]ara et al. reported a three-step procedure involving L-arginine 3-hydroxylase (VioC, EC 1.14.11.41, 39 kDa, monomer), arginase (EC 3.5.3.1, 33 kDa, hexamer), and ornithine cyclodeaminase (OCD, EC 4.3.1.12,39 kDa, dimer), which was performed using L-arginine as a starting substrate. [134]Arginase and OCD were obtained from Mesorhizobium loti and VioC was taken from Streptomyces sp.This process, in using VioC for generation of the intermediate (2S,3S)-3-hydroxyarginine, introduces a non-natural substrate to arginase (Scheme 11), which normally converts L-arginine.The kinetic parameters of the cascade, or at least for the final two enzymes used, arginase and OCD, suggest that while the cascade will tolerate the 3-hydroxylated analogs of L-arginine and L-ornithine, the catalytic efficiency is greatly reduced.Specifically, for arginase, the k cat /K M with 3hydroxyarginine was 0.046 compared to 6.58 s −1 mm −1 for the natural L-arginine substrate, as the K M was ≈8-fold higher and the k cat was ≈8-fold lower for 3-hydroxyarginine.Similarly, for OCD, the k cat /K M with 3-hydroxyornithine was 0.0001 compared to 3.54 s −1 mm −1 for the natural L-ornithine substrate; while the K M was approximately identical for the natural substrate and analog, the k cat was ≈20 000-fold lower for 3-hydroxyornithine.Despite the lower catalytic efficiency for the analogs, millimolar quantities of both intermediates and trans-3-hydroxy-L-proline were produced over the course of 1-3 h reactions.Importantly, neither regio-or stereoisomers were formed in the reaction, supporting its use for production of pure, specific trans-3-hydroxy-Lproline.OCD was identified as the rate-limiting step in this cascade as it very inefficiently converts 3-hydroxyornithine.Given the propensity of VioC to produce multiple hydroxyarginines, it seems likely that VioC and the relatively promiscuous arginase can find use in other synthetic applications beyond use in production of viomycin, an anti-tuberculosis agent, and other important NcAAs along with other compounds of pharmaceutical interest. [137]cheme 10.Production of D-phenylglycine via a three-enzyme cascade.Production of different D-PheGly-derivatives starting from -hydroxy-acids using mandelate racemase (MR, E1), (S)-mandelate dehydrogenase (SMDH, E2), and D-PheGly aminotransferase (DpgTA, E3).GluDH (glutamate dehydrogenase, E4) is responsible for restoring glutamate, which is needed to regenerate NADPH. [132]ble 5. Results for the synthesis of D-phenylglycine via a three-enzyme cascade.Production of different D-PheGly-derivatives starting from -hydroxyacids using MR (E1), SMDH (E2), and DpgTA (E3). [132]erall reaction:

Other Three-Enzyme Cascades
[140][141][142] Notably, Otte et al. developed a three enzyme cascade to produce the industrially relevant polymer building block azelaic acid. [142]Additionally, Engelmann et al. developed a three enzyme cascade for the production of cinnamyl cinnamate where they demonstrated potential scale-up methods for the production of the highvalue aromatic ester. [140]The three-enzyme cascades outlined in the above section demonstrate a higher level of complexity required to set up the initial approach relative to some of the other two-enzyme cascades prior.This complexity includes sourcing of the enzymes and then matching them kinetically where the number of variables increases with the number of steps/enzymes involved within a given cascade.

Four or More Enzyme Cascades
Few non-natural product deriving multi-enzymatic cascades containing four or more enzymes have been described in the literature to date.As alluded to above, this is likely compounded by the increasing system complexity and overall number of reaction steps, which accumulate with each enzymatic step added into a given cascade. [143]Each enzyme added into a cascade possesses its own properties in terms of specificity and promiscuity and these, in turn, must be conducive with each of the other enzymes when added into a multi-enzymatic system.These systems thus require kinetic and reaction optimization with the added steps also potentially reducing overall yield. [23]As BRENDA contains information for over 40 000 different enzymes, identifying the right enzyme that will perform a desired function on a specific intermediate within a multi-enzymatic cascade is often an overwhelming task. [144]When the necessary enzymes do not exist or are not specific or robust enough, then directed evolution approaches have to be added as shown below.This is another added complexity requiring very specific knowledge, protein engineering and screening capabilities, and, of course, previous successful experience.Despite all this, there do exist some model examples, which demonstrate the feasibility of designing such systems.47]

Deracemization of Primary Amines
Turner and co-workers previously took on the challenge of designing a five-enzyme cascade for the deracemization of primary amines, [145] a synthesis known for the competing imine hydrolysis which occurs in the presence of water. [148]The widespread prevalence of chiral amines in over 40% of all pharmaceutical products, including drugs involved in the treatment of hypertension, HIV, Parkinson's disease, and diabetes, has led to increased interest in finding greener/alternative enzymatic routes for the synthesis and production of optically pure amines. [149]Currently, methods for the synthesis of chiral amines not only lack the selectivity, which can be achieved with enzymatic systems, [149] but also require precious metal catalysts, organic solvents, and large amounts of hazardous chemicals such as sodium azide and sodium cyanoborohydride. [150,151]he cascade designed by Turner and co-workers incorporated a mutant (D9) of monoamine oxidase (EC 1.4.3.4,MW ≈215 kDa per tetramer) from Aspergillus niger, a catalase from bovine liver (EC 1.11.1.6,MW ≈60 kDa per homotetramer), an (R)ɷ-transaminase (EC 2.6.1.),and lactate dehydrogenase/glucose dehydrogenase for the one-pot synthesis of chiral amines with >99% ee for the (R)-amines (Scheme 12).In this cascade, the monoamine oxidase (E1) functions to selectively oxidize the methylbenzyl amine with the imine product undergoing rapid hydrolysis to form acetophenone.The (R)-ɷ-transaminase (Scheme 12, E3) functions to facilitate the reductive amination of the produced acetophenone to the (R)-amine, the final product in this cascade.The catalase (E2) functions to react with the hydrogen peroxide (H 2 O 2 ) by-product that is produced in the first step of the cascade, as the H 2 O 2 was speculated to decrease (R)ɷ-transaminase activity.Similarly, the incorporation of a pyruvate reduction step, with LDH (Scheme 12, E4), aided in shifting the equilibrium of the second step of the cascade in the forward direction. [152]The final enzyme in the cascade, glucose dehydrogenase (E5), functions to recycle the costly NADH cofactor.Turner also demonstrated the utility of their method across a series of primary chiral amine derivatives (Table 6) and even showed the utility of their method in conjunction with sterically hindered secondary amines.Specifically, the authors demonstrated that their five-enzyme cascade could also be applied for the >99% conversion of N-methyl-1-phenylmethanamine, Nbenzylethanamine, and N-benzylpropan-2-amine.While this system did require a large excess of alanine to drive intermediate conversion, the success of the limited substrate pool analyzed in this study indicates that additional amine derivatization could be possible with this system in order to obtain even more structurally diverse, and enantiomerically pure, chiral amines.Overall, this study demonstrates that with strategic enzymatic cascade design it is possible to turn what is considered a catabolic or degradation pathway, i.e., ketone formation via imine hydrolysis, into a viable synthetic route to achieve the production of enantiomerically pure amines.

Synthesis of Amorpha-4-11-Diene
The production of artemisinin and its derivatives is of continuing research interest for the treatment of malaria.While artemisinin is a natural compound produced in plants, non-natural derivatives of artemisinin are the core components of many of the drugs used to treat malaria.In 2016, nearly half a million deaths occurred due to malaria infections and it has been estimated that global demand for artemisinin along with its derivatives is approaching 119 metric tons per year. [153]The death toll and the threat of resistance to artemisinin-derived drugs for the treatment of malaria has resulted in increased research funding in this field, with the Gates Foundation committed to providing one billion dollars in funding by 2023. [154]One of the important precursors to artemisinin is the compound amorpha-4,11diene.While amorpha-4,11-diene is a natural compound, its Scheme 12. Multi-enzymatic cascade for the deracemization of primary amines.Reaction kinetics were not analyzed, however, in this study 1 mL experimental mixtures were incubated at 37 °C for 24-36 h prior to product analysis. [145]quirement for the potential synthesis of non-natural artemisinin derivative compounds to address drug resistance is the reason for its inclusion and discussion herein.
Recently, a six-enzyme cascade was developed by Chen et al. for the synthesis of amorpha-4,11-diene (Scheme 13). [146]n this cascade, mevalonate kinase (EC 2. one, to amorpha-4,11-diene (Scheme 13, E1-E6).While this study does not include a full substrate scope, it does serve as an elegant example as to how one can systematically optimize production in complex multi-enzyme cascades.In the study, the authors utilize the Taguchi method, which is a statistical method based on orthogonal arrays that allows one to optimize a system, its tolerance, and the parameters involved, [155,156] in order to identify optimal conditions for production in the multi-enzyme cascade.Using this, they were able to identify the optimal enzyme ratios, characterize the inhibitory effect associated with farnesyl pyrophosphate synthase, and increase the rate-limiting step of product release for amorpha-4,11-diene synthase.Additionally, the authors were able to optimize buffer conditions with respect to pH conditions and ion choice to further improve this multi-enzyme cascade to nearly 100% efficiency for the desired amorpha-4,11diene product.Table 6.Substrate scope and results from the multi-enzymatic cascade for the deracemization of primary amines. [145]erall reaction: See also Scheme 12 for information on the enzymes.
Adv. Mater.2024, 36, 2309963 © Published 2023.This article is a U.S. Government work and is in the public domain in the USA.Advanced Materials published by Wiley-VCH GmbH Scheme 13.Multi-enzymatic cascade for the production of amorpha-4,11-diene.Reactions with the purified enzymes were carried out in 100 mm Tris/HCl (pH 7.4), 10 mm MgCl 2 , 10 mm mevalonic acid, and 15 mm ATP at 30 °C. [146]

Synthesis of Islatravir
The viability of a nine-enzyme cascade, consisting of five primary and four auxiliary supporting enzymes, engineered for the synthesis of the non-natural product islatravir, an experimental compound potentially useful in the treatment of HIV, was demonstrated in a recent study by Huffman et al. (Scheme 14). [147][159] This allowed them to begin their experimental analysis starting from their desired product, islatravir.The more optimal enzymes the authors discovered in the bacterial nucleoside salvage pathway included purine nucleoside phosphorylase (EC 2. Scheme 14. Multi-enzymatic cascade for the production of islatravir.Reactions were performed in an automated lab reactor at 40 °C in pH 7.5 buffer to achieve an overall 51% yield. [147]reover, the authors specifically point out some of the same points that we allude to in the Introduction when discussing the benefits of minimalist biosynthetic systems, namely: i) the atom economy far exceeded that of previous syntheses implemented toward this target, ii) the number of steps required is less than half of previous, and more importantly, iii) the entire reaction sequence takes place under mild conditions in a single aqueous solution without the need to repeatedly isolate and purify intermediates before undertaking the next step.They also point out that bioinformatics followed by extensive screening and rescreening of individual enzyme mutants for initial low-level or promiscuous activity followed by directed evolution to evolve the necessary catalytic activity was repeatedly required in this approach.Moreover, thermodynamics and reaction modeling also had to be considered to set up the reactions to function in a productive direction.

Other Four or More Enzyme Cascades
][162][163] Importantly, Zhou showed the conversion of both styrenes (and derivatives) and L-phenylalanine to D-PheGly using different multi-enzyme cascades, where the number of enzymes involved numbered seven and nine, respectively, demonstrating the modular character of their system and expanding the range of the number of substrates that can be converted to D-PheGly.Like the three-enzyme system described above, both of these larger enzyme cascades were functioning in vivo within E. coli DL39(DE3). [132]Therefore, while the number of studies on multi-enzyme cascades that contain more than four enzymes is currently limited, the success shown in the two-and threeenzyme cascades outlined above remains promising for the development of more complex artificial multi-enzyme cascades.

Nanomaterial Scaffolds for Enhancing Product Formation in Multi-Enzymatic Cascades
With a representative overview of what multienzymatic cascades have to offer for the biosynthesis of products previously produced chemically, novel compounds, and especially those that cells cannot produce or would be extremely challenged to produce efficiently, we now turn to discuss some nanomaterials-based mechanisms to enhance final product formation and conversion efficiency.Since these are catalytic systems, the ability to increase their efficiency will allow for less costly enzyme materials overall to produce more product in an ideal system.Moreover, channeling, in particular, can allow not only the targeted multienzyme cascade to function more efficiently, it can also incorporate and improve the efficiency of the critical cofactor recycling components; this will reduce material needs and costs.Although channeling is a nanoscale phenomenon, it has macroscale implications since nanomolar concentrations of catalytic biomaterials can produce millimolar levels of product or more with far less enzyme than bulk reactions when they are actively constrained by diffusion-limited reactions. [22,59,60]The possibility also exists for functionally incorporating two or more channeled multienzyme cascades into the same reaction in a modular fashion; this would certainly help with making different non-natural products and also creation of molecular libraries of potential products (e.g., combi-chem libraries).
For in vitro minimalist approaches to multienzymatic cascades, the focus has been on trying to increase and expand reaction efficiency in one of three primary ways: i) stabilizing the enzyme's structure, ii) enhancing the enzyme's kinetic profile, and iii) accessing multienzyme channeling.Each of these have considerable challenges associated with them when implemented in the macroscale and especially when utilizing heterogeneous attachment chemistry with scaffolds.Barring the considerable engineering cost of redesigning an enzyme by rational mutation or directed evolution to enhance stability, [164][165][166][167] stabilizing the enzyme by surface attachment for long-term use and re-use depending upon reaction format is one of the more popular approaches.As mentioned, this often comes at the cost of a loss of catalytic efficiency when undertaken with macroscale materials. [35,36][166][167] Lastly, chemically crosslinking enzymes together or attaching them to (macroscale) scaffolding materials also does not guarantee the cascade will access channeling.[170][171][172][173] Additionally, cysteine-thiols are often required to form dithiol bonds to help drive protein folding and assume a functional structure.Thus, reducing these disulfides and then chemically modifying with crosslinking reagents many times kills the enzyme altering its required structure.Nor does simple chemical fusion or fused expression of two linked enzymes consistently deliver channeling. [54,174]][177] That these same approaches are now welltranslated into commercial and industrial application of enzyme technology also speaks to their relative maturity. [178]It is also worth noting that these applications may indeed be the source or original inspiration for using MOFs and other nanomaterials in a similar way. [179,180][183][184][185][186][187] Epitomizing this potential, we note that in the two-step coupled cellular-cell free bienzymatic cascade for synthesis of enantiopure chiral amines implemented by Liu and Li and described above, immobilization of AmDH and GDH to NPs improved the conversion rates up to 84% along with significantly increasing the TTN for NADH. [78]There are other examples where NPs have improved multistep enzymatic catalysis albeit in less defined systems.For example, Lupoi and Smith physisorbed cellulase onto 40 nm diameter silica NPs for simultaneous saccharification and fermentation (SSF) reactions. [188]In these reactions, the initial cellulose substrate is processed to glucose by the cellulase while yeast that is also present in the reaction mix then ferments the glucose to ethanol.Immobilizing the cellulase in this manner increased ethanol yields twofold versus free enzyme.In this example, the authors ascribe the improvement to stabilization of the cellulase structure on the solid support, which they also suggest helps promote its activity in the presence of increasing ethanol along with the non-optimal higher temperatures the reactions were run in.However, the focus in most of the examples above has been on achieving successful biosynthesis of product and not necessarily that of improving the mechanics of multienzyme cascades, hence the lack of other examples in this context is not surprising.
Table 7 presents a representative but certainly far from comprehensive list of some nanomaterial scaffolds being investigated for these purposes along with describing some of the enzyme systems that have been prototyped on them.Most of the systems in this table report some type of individual enzyme catalytic enhancement and increased stability on their scaffolds when these phenomena were specifically tested for.It is important to appreciate that each nanomaterial comes with its own benefits and liabilities (pros and cons) and that these may include how the enzymes are attached to the material, the amount of work that each approach requires, and whether they actually are able to provide the properties desired consistently.For example, DNA scaffolds are particularly intriguing in that the scaffold can be assembled into any arbitrary 1-, 2-, or 3-dimensional (3D) nanoscale structure as desired and the enzyme attachment points can be varied, which leads to strict control over enzyme order and density. [41,47,49,182,187,189,190]The latter properties are unique to working with DNA scaffolds and not achievable in any of the other scaffold systems to date. Figure 4 shows an example from the Morri group where xylose dehydrogenase/xylitol dehydrogenase (XR/XDH) derived from the D-xylose metabolic pathway were immobilized on a 3D DNA hexagonal prism and different open and closed forms of the structure tested while enzyme placement was varied. [189]Ingeniously, each of the enzymes was fused to a specific DNA binding domain, the latter of which were incorporated into the DNA structures in a modular fashion as desired to predefine enzyme placement.In this case, the authors were able to show a sixfold increase in coupled flux when the structure went from open to closed simply by addition of other oligonucleotides.Clearly DNA scaffolding is a fascinating material with much to offer, however, achieving site-specific attachment of enzymes by creating DNA binding fusions or utilizing chemoselective ligation to attach DNA complements to enzymes is not trivial. [52,53,182]Additionally, although most enzymes show some type of kinetic enhancement when attached to DNA, multienzyme channeling on DNA scaffolding has still not been unambiguously shown to date.Metal organic frameworks offer the benefit of being formed while encapsulating the enzymes at the same time and providing them with protection for long time periods.This does not provide for intimate control over enzyme ratios-but rather a more gross incorporation process and again, channeling processes have not been specifically shown with these systems.Protein-based nanosystems in the form of viral capsids also allow for enzyme encapsulation or display on the [191] particle itself but channeling itself has not been shown here either despite some kinetic improvements.As Minteer's group showed, direct fusion of enzymes together can lead to significant

Potential use in biofuels production
Expression from fused gene alcohol dehydrogenase→aldehyde dehydrogenase 500-fold increase in flux Designer channeling [191]   Gamma-prefoldin scaffold
improvements from catalytic channeling with a 500-fold increase in flux achieved.However, as they also showed, this required considerable modeling and engineering to achieve for just a single targeted fusion. [191]llowing enzymes to crosslink with NPs into dense clusters that engage in channeling appears to have much to offer in this regard and may also mechanistically mimic what happens in cells as these clusters resemble the transient intracellular metabolons that are believed to form intracellularly to increase metabolic flux in a localized manner. [61,62]Figure 5 shows a recent example where from 4 up to 9 enzymes associated with glycolysis were self-assembled with QDs into dense nanoclusters and where coupled flux was increased by channeling at least 100fold consistently. [17]Seven of the fourteen enzymes tested in this study also displayed kinetic enhancement when displayed on QDs.The presence of channeling in this system was unambiguously confirmed with several rigorous and classical assay formats.Moreover, the authors showed that flux could be enhanced several folds more by optimization of enzymatic stoichiometry with numerical simulations-this considered each enzyme's kinetic profile to better match the relative ratio of each as assembled in the QD clusters.Additionally, switching from spherical QDs to 2-D planar nanoplatelets and ordering of the enzyme assembly process also increased the multienzyme flux.Interestingly, the current understanding of multienzymatic channeling does not account for enzyme order having any effect on efficiency, rather it is achieving a critical localized enzyme density that is considered paramount. [59,60]However, systems with this many enzymes actively engaged in channeling have not been assembled and intimately characterized before.In the context of an extended cascade where unfavorable kinetics did not allow for contiguous channeling through the cascade, purification of an endproduct from the upstream sub-cascade, and then feeding it as a concentrated substrate to the downstream sub-cascade maintained overall channeling activity in both parts of this pathway.Lastly, the generalized applicability of this approach was verified by creating channeled assemblies with the same enzymes using other hard and soft NP materials such as AuNPs and dendrimers.These NP-scaffolded systems formed based solely on the binding interaction of each enzyme's distal (His) 6 motif with the QD ZnS surface.Since enzymes are typically expressed with these motifs for subsequent purification using metal chelate resin not much engineering is actually needed to form the requisite channeling clusters.All any other NP materials need to be used in a similar manner is the ability to coordinate to the (His) 6 motif on other enzymes and this can be provided by NPs displaying NTA groups; these are also widely available commercially. [172]The overall simplicity and engineering ease of this approach, which essentially QDs are mixed with stoichiometric ratios of enzymes that constitute a targeted cascade and selfassemble into nanoclusters.Addition of initial substrate such as linear starch is then processed to product by the multienzyme cascade in the cluster, which exploits localized intermediary channeling.B) Forming into NP-enzyme clusters and engaging in multistep channeling increases the overall catalytic flux by orders of magnitude over that of freely diffusing enzymes, which encounter significant diffusion limitations.The latter substantially reduces the overall transient time () for that reaction.C) Progress curve measuring NADH conversion over time for 520 QD clusters assembled with the 7 enzyme (7E) system (glucose→3-PG) at empirical enzyme ratios (red) versus same enzyme free in solution (blue).10 μm indicates the final amount of NADH converted in the free enzyme assay.Progress curves assembled using optimized enzyme ratios per QD determined after two consecutive rounds of numerical simulation (pink-Opt 2).Free enzyme controls for optimization had identical results as that of the empirical sample.520 QD concentration = 2.5 nm.D) Representative TEM micrographs of clusters formed with 520 QD materials (6.25 nm) using a 7 enzyme cascade at Opt 2 ratios processing glucose to 3 phosphoglycerate.Average cluster size is given above the micrograph along with the number of QDs counted in that determination.Inset, representative high-resolution micrograph of an individual cluster.Reproduced under the terms of the Open Access Creative Commons Attribution 4.0 International License. [17]Copyright 2023, Springer Nature.
only requires mixing enzymes with QDs or similar NPs in the right stoichiometry, suggests it as the most feasible for continued research implementation in this pursuit right now.Perhaps the most interesting future possibility would be if such scaffolding could help a given enzyme increase its substrate scope and tolerance, and there have been some interesting preliminary results that suggest this may be possible. [22]

Conclusions and Outlook
We begin this section by again stipulating that the representative example reactions highlighted above are certainly not comprehensive nor do they cover the breadth of progress being made within the in vitro and non-cellular portion of the nascent Syn-Bio field with respect to non-natural products.Rather they primarily reflect what is possible when enzymatic chemistry is coupled together and moved beyond the cell.SynBio is not meant to be a wholesale replacement for synthetic chemistry since there will always be a significant amount of chemical space that is beyond the capabilities of enzymes. [23]It is also important to re-alize that moving enzymatic reactions out of cells is not an all or nothing approach.There is no preclusion from undertaking mixed chemoenzymatic or even mixed cellular-chemoenzymaticchemical synthetic approaches, [136,188,198] and indeed some of the examples above confirm this approach.However, all of the benefits of the enzymatic approach will be imported into the reaction regardless of whether it is used in one or more steps.Ultimately, one reality for this scenario may be dictated by synthetic chemistry providing what it can only do and enzymes providing what they can only do toward creation of new molecules.
As already confirmed by some of the above examples, [10,113,114] high value pharmaceutical compounds probably represent the most desirable low-hanging fruit and will be amongst the first set of target compounds to be successfully developed using this approach while also providing a good return on investment.Driving much of this will be the unique regio-and stereo-specific chemistries that are only available to enzymes, and which are specifically needed for many drug compounds to be biologically active.There is already some precedent for such a production approach in the example of the drug paclitaxel or Taxol.This  [201] Copyright 2023, the American Chemical Society.
isoprenoid compound is found in the bark of the Pacific Yew tree and was identified in a screening of natural compounds as a potent cancer inhibitor.However, such minute amounts are naturally present in the bark that it could never be produced by farming these slow growing trees.This led to a concerted investment in its total synthesis, which although successfully accomplished is exceedingly difficult with some of the initial approaches requiring more than 40 steps and yielding only 1-2% of product. [199]t is currently produced by a mixed chemical-biosynthetic route since efficient total production in host cells is exceedingly challenging and it has already attained WHO essential medicine status. [200]As such, production of paclitaxel epitomizes much of what is driving the research discussed here-using enzymes outside of cells to do what cells either cannot undertake or undertake efficiently.
This drug represents one of the first success stories in this context.Another example in this vein specifically exploited substrate promiscuity in a mixed chemoenzymatic approach to produce a verruculogen, a class of rare fumitremorgin alkaloids containing a highly unusual eight-membered endoperoxide group that have shown toxicity toward several specific cancer cell lines along with reversing multidrug resistance in certain breast cancers. [201]tructurally, these natural fungal-produced alkaloids have a diketopiperazine core that is formed from proline and tryptophan with the prenyl side chains attached to the tryptophan indole embedded within the eight-membered endoperoxide ring.The Ting group utilized the verruculogen synthase FtmOx1 isolated from Aspergillus fumigatus (E.C. 1.14.11.38,MW ≈ 35.1 kDa per homodimer) to biomimetically complete the endoperoxidation step in their retrosynthesis of the target 13-oxoverruculogen product as they were not able to accomplish this synthetically under endoperoxidation conditions, see Scheme 15. [202] They were able to achieve a 62% conversion of the 13-epi-fumitremorgin B substrate to the 13-epi-verruculogen peroxide product when optimizing the enzymatic step by adding (NH 4 ) 2 Fe(SO 4 ) 2 as an iron source for the reaction.This represented the first successful use of this enzyme acting on a non-natural analog of its native substrate, fumitremorgin B. Reflecting the thesis of this review, the authors state "Our strategy highlights the use of native enzyme promiscuity in the chemoenzymatic synthesis of natural products, where substrate analogs are utilized rather than the native substrate." [201]roduction of high-value bioreagents and other fine chemicals will probably follow behind pharmaceuticals as also driven by return on investment.Enzymatic synthesis of complex glycans is a prime example of the latter area given the dearth of standardized and modular wet chemical synthetic approaches to these compounds. [203]Enantiomeric and stereoisomeric specificity are a big part of sugar/glycan chemistry because of their natural production by enzymes-so why try to compete with this?The examples above, including that of the O-mannosyl glycan NeuNAc2-3Gal1-4GlcNAc1-2Man (N-) and cladribine highlight some possibilities in this area, [114] however, there are other applications where enzymatic synthesis of glycans can contribute.The first is targeted toward that of complex glycan synthesis itself.Indeed, efforts are already underway to develop automated enzymatic glycan synthesis (AEGS), where the focus is on utilizing Leloir-glycosyltransferases to achieve the requisite quantitative regio-and stereoselective glycosylation in each step as part of a repeated sequential process. [204,205]Although exploiting instrumentation to control and optimize chemical conditions, the goal is to have the essential reactions carried out by enzymes using unmodified and unprotected monomers.The second area is that of sugar nucleotide recycling within reaction schemes.The costs of chemically synthesizing sugar nucleotides make recycling these key molecules in reactions an imperative for developing viable industrial applications.[17] Given the cost benefits and chemical space that can be accessed by using enzymes for these purposes, there will undoubtedly be much investment and progress toward these goals.
Commercial enzymatic synthesis of bulk (non-natural) industrial intermediaries will occur over a much longer timeframe.Here, cost and regulatory considerations will be the biggest determinants.Enzymatic biosynthesis of large amounts (>10 5 -10 6 tons or more) of a target molecule must become cost competitive with production by synthetic chemical methods and especially those that rely on initial petroleum feedstocks.Reduction of organic waste and concomitant waste disposal costs can become an important factor in this regard.Unlike centralized petroleum refineries and chemical plants, which are typically near seaports due to the need to transport in bulk, enzymatic biosynthesis can be more distributed allowing for a more localized production to match existing levels of local agricultural feedstock/substrate production and local need.This will also be another positive cost determinant.Indeed, having a given production facility focus on synthesizing only one or two specific products in this manner is not unrealistic because of the reduction in the required infrastructure.In an ideal scenario, the waste stream from another localized agricultural or (bio)synthetic production line would be used as substrate, again highlighting the concept of a circular bioeconomy.
The unique synthetic capabilities of the pathways discussed in this review arises not only from the complexity of the target molecule produced but also from the manner in which they are produced.To emphasize this, Table 8 compares five representative examples of the biosynthetic routes already highlighted above and picked at random versus examples of a traditional chemical synthetic route for the same molecule.Although certainly not meant to be predictive, in general, the conversion rates, percentage yield, and ee are somewhat comparable between the two approaches.The biosynthetic routes are all accomplished in fewer to sometimes just one-step, without requiring expensive catalysts, mostly without organic waste, and certainly without the harsh, sometimes dangerous conditions and multistep process of react/purify/characterize yield and then redo for the next chemical conversion that typically characterizes each sequential step of a chemical synthesis route.Notably, to our knowledge, no current purely chemical synthetic route exists for the production of cladribine triphosphate and the chemical synthetic route to islatravir involves a 12 step total synthesis to obtain less than half the overall yield. [206]However, one major and still mostly unaddressed issue with the development of these type of biosynthetic processes-be they cell based or cell free-is that of achieving industrial scale-up. [207,208]As mentioned this remains mostly an art form rather than an exercise in classical chemical engineering.Regardless of the way it will be implemented, there is still much to be learnt about these systems.
To fully realize SynBio and especially the minimalist enzymeonly version of it that we focus on and espouse here will absolutely require the establishment of a viable retrobiosynthetic framework.This will be directly predicated on populating databases not only with available enzymatic reactions, but more importantly, the available substrate tolerance of all related enzymes from all sources.This is not an easy or small undertaking, to say the least.Fortunately, artificial intelligence (AI) may have now matured enough to help with this endeavor. [212]Appropriately programmed AI systems are probably mature enough now that they can scour the literature to find, extract, interpret, and arrange all of the requisite data from the scientific literature.Although not discussed here, the bulk of this data will be drawn from the vast literature describing single enzyme substrate characterization and tolerance studies.Parsing of any quantitative data in these studies will be useful for comparative purposes between enzyme sources.Similar to chemical retrosynthesis, data can be collected all the way back to the first publications on enzyme characterization.Using AI to extract such complex data from the literature was not possible even 10 years ago when similar efforts where undertaken to collect and collate the accumulated literature data on NP toxicity studies. [213,214]A future, more advanced AI approach could then consider enzyme structure as predicted from raw genomic data by programs such as AlphaFold along with what is already in the current retrobiosynthetic and crystallographic databases to make predictions about the substrate specificity and substrate tolerance of newly sequenced enzymes. [215,216]It is probable that these same AIbased approaches could predict which changes need to be made to current enzymes to make them more tolerant to a given (nonnatural) substrate or reaction.This, in essence, represents a complementary, less labor intensive AI-based approach toward enzyme evolution.The impressive number of times that bioinformatics was utilized as part of the design approach in conjunction with all the different enzymes and enzyme source species already mentioned in just this brief review directly reflects the importance of such efforts in creating more complex enzymatic cascades.Attesting to how much untapped potential there is in this approach, a small fraction of one percent of the plant and bacterial species (and indeed all species) on the planet have had their genomes sequenced to date pointing out how much remains to be discovered.Not only are there many new enzyme reactions waiting to be discovered and characterized, but also an unimaginable amount of substrate tolerance space to be added for each enzyme source and variant.We do note the establishment of an initial retrobiosynthetic framework in the form of the RetroBioCat Database. [217,218]In line with this we also note the establishment of preliminary standards for reporting such biocatalytic data in the form of the Standards for Reporting Enzymology Data (STRENDA) database, which incorporates guidelines and recommendations for scientific journal publications. [219]evelopment of in vitro minimalistic enzymatic systems brings with it another unique benefit not available to individual prokaryotic cellular chasses.Expressing eukaryotic proteins in prokaryotes requires significant back engineering to remove introns, optimize codons, remove most post-translational modification sites, and allow for correct folding, amongst other requirements. [220,221]Non-cell extract-based ex vivo reactions should be completely agnostic to enzyme origin and should be able to functionally incorporate almost any protein.This will allow enzymes sourced from Monera (prokaryotes) and the 4 eukaryotic kingdoms: Animalia (Metazoa), Plantae, Fungi, and Protista to be expressed under optimized conditions in their preferred host with little to no protein engineering needed and then utilized together in designer reactions.The bulk of this review focused on enzyme catalysis and so inclusion of nanoscaffolding to improve multienzyme reactions may seem a bit of a non sequitur at first view.As mentioned above, approaches that can stabilize enzymes and allow them to be more efficient in the specific context of multienzymatic catalytic flux become key to utilizing these formats.Although displaying enzymes on NPs does seem to improve the kinetic profile of many enzymes along with their stability, channeling itself does not-it is just a way to improve flux through the coupled system while requiring less material.Obviously, increasing reactant and enzyme concentrations will produce the same outcome but at the penalty of significant extra cost.Moreover, a great deal of initial research focuses on reaction optimization prior to scale up and the formats used at this stage do not utilize large amounts of material.Thus improving their efficiency at this step can be very helpful.This is not just academic as companies are already investing in placing enzymes on DNA in pursuit of viable catalytic systems.For example, Touchlight Genetics is focused on immobilizing enzymes that use lactate on DNA scaffolds to create a lightweight nonhazardous biobattery technology. [222]Additionally, Zymtronix is developing multi-enzyme immobilization on ceramic scaffolds that are designed to incorporate tunable enzyme ratios of core pathways to produce target molecules in a cell free manner; they are focused on target molecules that are currently non accessible as they are "locked by traditional organism constraints." [204,223]iven the inherent benefits of the SynBio reaction formats outlined above, there are many other examples of startup companies pursuing similar types of technological approaches.Interestingly, although the idea is to exploit nanoscaffolds to access channeling ex vivo for improving non-natural product synthesis, the systems will still benefit from long term enzyme stabilization and potential reusability.Moreover, studying channeling phenomena may actually provide information about how enzymes function efficiently in cells as metabolons and allow us to emulate that and, in turn, improve channeling in both formats Lastly, as this field develops we can expect the reaction cascades to become far more complex, whether they are fully enzymatic, chemoenzymatic, or cascades that incorporate significant contributions from both.This will require highly interdisciplinary teams to undertake with contributions including: bioinformatics; AI; large data analysis; enzyme evolution; protein production; screening; enzymology; reaction modeling; chemical engineering; laboratory automation, IT, and analytical chemistry among many others. [224,225]This is clearly an exciting time to be working in this field and there certainly is a lot of truth to the idea that SynBio is the next industrial revolution. [226]Enzymes will be the key catalysts to realizing this revolution and they will be exploited in any manner that they are useful regardless of whether it is inside a living cell or in a minimalist-nanoscaffolded reaction.

Figure 2 .
Figure 2. Multienzymatic cascades.The functionality or mechanism behind cascaded enzyme structures can be grouped into four primary categories: linear, orthogonal, parallel, and cyclic.In linear cascades, one substrate is ultimately converted to one product through multiple enzymes and intermediates.In orthogonal cascades, a cosubstrate or cofactor is generated by a second enzyme or a second enzyme removes the undesired byproduct.In parallel cascades, two reactions are coupled by cosubstrate or cofactor use; importantly, in this case both products are desired, though one could potentially consider the orthogonal cascade as just a combination of the linear and parallel cascades.In cyclic cascades, from a mixture of substrates (e.g., D and L enantiomers), only one type (e.g., the D enantiomer) is selectively converted to an intermediate and subsequently converted back to a mixture of substrates (e.g., D and L enantiomers); over multiple cycles, the nonreactive substrate (e.g., the L enantiomer) accumulates.The second step can be chemical instead of enzymatic, making the process a chemo-enzymatic cascade.Reproduced with permission.[31]Copyright 2019, American Chemical Society.

, 2309963 ©Scheme 7 .
Scheme 7. Cladribine triphosphate production with three-enzyme cascade.2Cl-adenine and phosphoribosyl pyrophosphate (PRPP) are bound by APT.Two rounds of phosphorylation PPK yields 2Cl-adenosine triphosphate, which is reduced by RNR to form 2Cl-deoxyadenosine-triphosphate (cladribine triphosphate).For APT, conversion of adenine has a K M of 11.8 ± 5.2, and k cat of 119 ± 14; PPK has a K M of 41 ± 6.3, and k cat of 17 ± 1.3 for AMP, and a K M of 144 ± 14, and k cat of 0.43 ± 0.02 for ADP; RNR conversion of ATP has a K M of 199 ± 13, and k cat of 0.56 ± 0.02 (units for K M are μmol L −1 and s −1 for k cat ).[114]

Figure 4 .
Figure 4. Enzyme cascade reaction after the shape transformation of DNA scaffold.A) Shape transformation of the enzyme-loaded DNA scaffolds.HPO/XR-XDH was incubated in the presence or absence of CLK to obtain the closed state of DNA scaffold encapsulated with enzymes (HPC/XR-XDH).B) Time course for NADH regeneration of the cascade reaction of HPOa/XR-XDH or HPC/XR-XDH.C) Illustrations representing XR and XDH coassembled in the closed state of 3D DNA hexagonal prism (top, HPC L /XR-XDH) and the equimolar mixture of HPC L /XR and HPC L /XDH (bottom).D) An HPLC chromatogram shows the formation of products xylitol and xylulose in the reaction of xylose containing [1-3 H]xylose with HPC L /XR-XDH for 400 min.E) Time course of the xylulose production by the equimolar mixture of HPC L /XR and HPC L /XDH, and HPC L /XR-XDH.Reproduced under the terms of the Open Access Creative Commons Attribution 4.0 International License.[189]Copyright 2023, John Wiley & Sons, Inc.

Figure 5 .
Figure5.Channeling in multienzymatic glycolytic cascades self-assembled on nanoparticle scaffolds.A) Schematic depicting the self-assembled QD enzyme clusters forming multienzyme cascades.QDs are mixed with stoichiometric ratios of enzymes that constitute a targeted cascade and selfassemble into nanoclusters.Addition of initial substrate such as linear starch is then processed to product by the multienzyme cascade in the cluster, which exploits localized intermediary channeling.B) Forming into NP-enzyme clusters and engaging in multistep channeling increases the overall catalytic flux by orders of magnitude over that of freely diffusing enzymes, which encounter significant diffusion limitations.The latter substantially reduces the overall transient time () for that reaction.C) Progress curve measuring NADH conversion over time for 520 QD clusters assembled with the 7 enzyme (7E) system (glucose→3-PG) at empirical enzyme ratios (red) versus same enzyme free in solution (blue).10 μm indicates the final amount of NADH converted in the free enzyme assay.Progress curves assembled using optimized enzyme ratios per QD determined after two consecutive rounds of numerical simulation (pink-Opt 2).Free enzyme controls for optimization had identical results as that of the empirical sample.520 QD concentration = 2.5 nm.D) Representative TEM micrographs of clusters formed with 520 QD materials (6.25 nm) using a 7 enzyme cascade at Opt 2 ratios processing glucose to 3 phosphoglycerate.Average cluster size is given above the micrograph along with the number of QDs counted in that determination.Inset, representative high-resolution micrograph of an individual cluster.Reproduced under the terms of the Open Access Creative Commons Attribution 4.0 International License.[17]Copyright 2023, Springer Nature.

Table 7 .
Overview of representative nanoscaffold materials being investigated for enhancing multienzymatic cascades.

Table 8 .
Comparison of reaction conditions, conversion rates, and enantiomeric excess between select enzymatic and chemical syntheses.