Synthetic enzyme supercomplexes: co-immobilization of enzyme cascades

A sustainable alternative to traditional chemical synthesis is the use of enzymes as biocatalysts. Using enzymes, di ﬀ erent advantages such as mild reaction conditions and high turnover rates are combined. However, the approach of using soluble enzymes su ﬀ ers from the fact that enzymes have to be separated from the product post-synthesis and can be inactivated by this process. Therefore, enzymes are often immobilized to solid carriers to enable easy separation from the product as well as stabilization of the enzyme structure. In order to mimic the metabolic pathways of living cells and thus to create more complex bioproducts in a cell-free manner, a series of consecutive reactions can be realized by applying whole enzyme cascades. As enzymes from di ﬀ erent host organisms can be combined, this o ﬀ ers enormous opportunities for creating advanced metabolic pathways that do not occur in nature. When immobilizing this enzyme cascades in a co-localized pattern a further advantage emerges: as the product of the previous enzyme is directly transferred to its co-immobilized subsequent catalyst, the overall performance of the cascade can be enhanced. Furthermore when enzymes are in close proximity to each other, the generation of by-products is reduced and obstructive e ﬀ ects like product inhibition and unfavorable kinetics can be disabled. This review gives an overview of the current state of the art in the application of enzyme cascades in immobilized forms. Furthermore it focuses on di ﬀ erent immobilization techniques for structured immobilizates and the use of enzyme cascade in specially designed (micro ﬂ uidic) reactor devices.


Introduction
The use of enzymes for the fermentation of food and beverages as well as in medicinal applications is almost as old as mankind.Even in the bible (Isaiah, 2 nd Book of Kings) it is reported, that a wound was healed by applying a patch of g.Nowadays we know that considerable amounts of the enzyme cain were responsible for the healing effect. 1 Since the beginning of the 20 th century, single enzymes are specically isolated from crude materials and used in industrial elds such as food, pharmacology and textile industry, for production of ne chemicals, for electricity generation in biofuel cells and in diagnosis.Compared to classical chemical synthesis, the use of enzymes offers crucial advantages: 2 while organic synthesis is oen conducted in pollutive organic solvents, most enzyme reactions take place under mild pH and temperature conditions in aqueous milieu as given in the natural environmentinside a cell.Most enzymes show high specicity both to their substrates and products, which reduces the formation of unwanted byproducts that subsequently have to be separated from the product.Furthermore by protein engineering, specicity, stability and enzyme characteristics can be modied according to specic industrial applications. 3,4However, if enzymes are applied in soluble form, they have to be separated from the product post-synthesis.This process is oen expensive, timeconsuming and the catalysts are mostly inactivated.The immobilization of enzymes can solve this problem: by converting the enzymes to an insoluble form, they can be easily separated from the reaction solution. 5Enzymes can be crosslinked to each other, 6 entrapped into matrices 7,8 or attached to solid supports such as microparticles, bers or other functionalized surfaces.Thus, the activity and stability of the enzymes can be enhanced and their selectivity can even be tuned, depending on the immobilization strategy. 9However, in some cases the enzyme activity can also be reduced by immobilization, for example if the exibility of the enzyme is disturbed or the binding occurs in or near the active center of the enzyme required for substrate conversion.Therefore an optimal immobilization protocol has to be developed for each enzyme.For the production of more complex products, one reaction step may not be enough.Thus, a series of consecutive enzyme catalyzed reactions may be required.In this case whole enzyme cascades are implemented.In 2011 Ricca et al. excellently reviewed the advantages of using enzyme cascades for the one-pot production of chiral chemicals, such as alcohols, amines and amino acids. 10The co-localized immobilization of enzyme cascades consisting of two or more types of biocatalysts offers further advantages: a specic arrangement of the enzymes that enables close proximity to each other leads to an effect called substrate channeling: the product of an enzyme is directly transferred to the co-localized following enzyme, where it can act in turn as substrate. 11,12Those short diffusional distances accelerate the speed of the reaction and lead to an enhanced performance of the cascade compared to their soluble form. 13Another advantage compared to whole-cellcatalysis is the opportunity of building up articial enzyme complexes whose components can be derived from an immense variety of different host organisms with distinct characteristics and advantages.Thus, articial metabolic pathways can be engineered that do not occur in nature. 14here are different approaches for the realization of enzyme cascades.Enzymes can be co-immobilized by different techniques in a more or less specic pattern, which leads to close proximity to each other and facilitates substrate channeling.Biocatalysts can also be separated into different reaction compartments that the product stream passes subsequently.As in separated reaction compartments the process parameters can be adapted, this approach is favored if the selected enzymes differ in their requirements concerning optimal process conditions.

Examples of an important enzymatic systems
The most prominent example of an enzyme cascade used technically is the bi-enzymatic system for sugar detection consisting of Glucose Oxidase (GOx) and a peroxidase, mostly derived from horseradish (HRP, Armoracia rusticana).Glucose is oxidized by the GOx using ambient oxygen and gluconolactone while hydrogen peroxide (H 2 O 2 ) is generated.The hydrogen peroxide is used by the peroxidase to oxidize a dye that is added to the substrate solution from its colourless to its coloured form.This leads to a quantiable absorbance signal which is in case of constant reaction term proportional to the glucose concentration in the medium.Carr et al. published already in 1946 the use of this enzyme system for blood sugar detection in the form of a rapid bedside test for diabetic patients. 15The whole history of monitoring blood glucose using enzyme based biosensors, as well as the main aspects concerning technical improvements, standardized analytics and performance levels were reviewed by Yoo and Lee in 2010. 16If further enzymes are added to the cascade, the detection spectrum can be extended to other types of sugar molecules.For instance, van Dongen et al. published in 2000 the extension of the system by the Lipase B from Candida antarctica (CalB), which enables the system to convert an acetate-protected glucose to glucose and its subsequent detection. 17 A second important eld, in which enzyme cascades are used, is the production of electricity using enzyme based biofuel cells.Enzymes used for this type of application normally belong to the family of oxidoreductased.The topic was very well reviewed in 2007 by Minteer et al. highlighting the advantages and disadvantages of enzyme fuel cells compared to microbial fuel cells.Advantages are the higher power densities that can be achieved by the immobilization of the biocatalysts and their higher specicity.However, they suffer from short life times (7-10 days) and only partly oxidized fuel substrates. 20The rst enzyme cascade for electricity generation was already applied in 1998 by Palmore et al.The authors used alcohol dehydrogenase, aldehyde dehydrogenase and formate dehydrogenase for the oxidation of methanol to carbon dioxide. 21Another work was published by Akers et al. in 2004, where ethanol was oxidized in a two-step reaction to acetate by combining an alcohol dehydrogenase and an aldehyde dehydrogenase.

Related work
There have been many important contributions reviewing the generation and application of enzyme cascades whereby the most important ones will be briey outlined in this section.In 2010, Betancor and Luckari reviewed the application of enzyme cascades in biosensing and production. 22The excellent paper by Schoffelen and van Hest published in 2013 summarizes the chemical strategies in covalent assembling of multienzyme complexes. 23The cell-free production of ethanol by enzyme cascades was reviewed in 2014 by Khattak et al. and gives example of the technical use of enzyme cascade in technical productions. 24

Scope of this article
This review summarizes the different techniques for immobilization of enzymatic cascades and focuses on classifying the different co-immobilization techniques while giving examples of sophisticated approaches.Another excellent review, highlighting the current state of the art, the principles of enzymatic fuel cells, unsolved problems and possible strategies for addressing them was published in 2011 by Osman et al. 25

Principles of random co-immobilization of enzymatic cascades
Random co-immobilization is mainly achieved by crosslinking the members of the cascade to solid supports or the entrapment of the biocatalysts into polymer lms.It leads to a statistical distribution of the enzyme, depending on the ratio of applied enzyme masses, and the density of functional groups on the support material.This method is oen used for biosensing application.A simple and fast opportunity for attaching enzymes to solid surfaces is the use of chemical linking agents as for instance classical aldehyde-amino-crosslinking by glutaraldehyde 26 or 1-ethyl-3-(4-diaminopropyl) carbodiimide (EDC) to connect functional amino-sidechains from the protein surface to an activated support 27 by the activation of carboxylic acids for a nucleophilic attack by an amine.The most commonly employed activation is the conversion of the carboxylic acid to a so-called active ester e.g., the N-hydroxysuccinimide or pentauorophenol ester.These approaches are fast and simple and enzymes do not need to be extensively modied before immobilization.However, as the distribution of functional groups is random, a specic control of localization and orientation is not possible.

Random co-immobilization on surfaces
Two or more enzymes can be co-immobilized to functionalized surfaces in a statistically distributed manner.For some control of the immobilization pattern, either the ratio of the applied enzymes can be tuned or the distribution of different functional groups for enzyme attachment at the surface can be engineered.Deng et al. published an elegant approach for the co-immobilization of proteins on a patterned surface that was generated by chemical vapour deposition (CVD). 28By copolymerization of a controlled ratio of different monomer types, a statistical distribution of alkyne and pentauorophenyl groups was generated on a surface.Thus, two different proteins were immobilized via azide-alkyne-cycloaddition and activated esteramine reaction, providing two orthogonal reaction types.Although in this case proteins for cell adhesion were immobilized instead of an enzyme cascade, this approach can be a versatile tool for the realization of a cascade of biocatalysts.However, this approach suffers from one disadvantage: the enzyme groups are statistically distributed, depending on the ratio of monomer.An exact patterning and thereby an ensured maximum distance between the individual enzymes is not possible using this approach.

Random co-immobilization by encapsulation
Zhu et al. immobilized GOx and HRP in an electropolymerized pyrrole lm that was deposited on an electrode coated with singe-wall carbon nanotubes (SWCT) in order to generate a glucose biosensor. 29By determination of the amperiometric response of the bioelectrode, a signal was recorded that seemed to be proportional to an applied glucose concentration of 3 Â 10 À5 to 2.43 Â 10 À3 M. Furthermore, the results indicated a 6.8-fold greater sensitivity, when the enzymes were co-immobilized, compared to a sensor with separately immobilized biocatalysts in different polymer layers.Comparable systems for glucose detection by co-encapsulation of HRP and GOx were also investigated by other groups.In 2002 Wei et al. published the successful incorporation of both enzymes into mesoporous sol-gel materials. 30Ji et al. achieved a co-encapsulation of the cascade by diluting the enzyme into a polyurethane based solution and subsequent co-axial electrospinning, creating a hollow nanobre membrane that is able to serve as glucose detection strip. 31Eguilaz et al. immobilized a different enzyme mixture, cholesterol oxidase (ChOx) and HRP to composites consisting of coated multiwall carbon nanotubes and magnetic nanoparticles, thus creating a biosensor for the detection of cholesterin. 32ndom co-immobilization by supportless crosslinking Another prominent example of random co-immobilization of enzymatic cascades is the immobilization by interconnecting enzymes into so-called combined crosslinked enzyme aggregates (combi-CLEAs).This approach combines two advantages: the co-localization of enzymes and the absence of carriers that lead to a dilution of enzyme activity. 33For instance Mateo et al. developed a combi-CLEA, that consisted of a S-selective oxynitrilase derived from Manihot esculenta and a nonselective nitrilase derived from Pseudomonas uorescens for the enantioselective conversion of benzaldehyde to (S)-mandelic acid. 34,35y doing so, another advantage of co-immobilized enzyme cascades was exploited: the in situ conversion of the enantioselective product produced by the oxynitrilase is directly converted by the nitrilase, whereby the equilibrium of the rst reaction step is driven towards the product. 33Thus, even unfavorable kinetics can be disabled by the co-immobilization of enzyme cascades.

Principles of site-specic co-immobilization of enzymatic cascades
This section will describe approaches by which enzymes are not only brought in statistically controlled close proximity by coencapsulation or co-crosslinking, but also in dened patterns or shapes and with dened spacing between them.For the immobilization of an enzyme cascade in an organized pattern, additional modication steps are necessary.However, due to an enhanced performance of the cascade it is oen worth the effort.There are different ways of generating enzyme patterns.For a specic attachment of a dened enzyme to a specic binding site, orthogonal binding mechanisms are required.That means, that binding occurs exclusively between target enzyme and target binding site without any unspecic attachment.Therefore the pattern has to be dened by the distribution of different functional groups on the respective surface.When more than one enzyme type has to be immobilized, several different orthogonal binding mechanisms are required.Here a promising yet somewhat limited approach are methods based on so-called "click chemistry".These "bio-orthogonal" reactions occur only between the functionalized material surface and specically introduced residues on the protein surface, enabling the generation of protein patterns. 36One prominent example is the copper-catalysed 1,3-dipolar cycloaddition of azide and alkyne groups or Huisgen-reaction.It occurs at mild reaction conditions without the formation of unwanted byproducts.One group is introduced at the protein surface, while the other group is attached to the desired surface before the coupling step.However, as these groups are oen introduced into a protein unspecically by crosslinking chemistry, the immobilization still takes place in a random orientation.An elegant, but sophisticated way to circumvent this limitation and to generate site-specically labelled enzymes is the incorporation of unnatural amino acids containing the respective groups, for instance p-azido-phenylalanine, by means of an expanded genetic code. 37Another way of perfectly controlling the orientation of immobilized proteins is the use of genetically encoded tags that are attached to the desired enzyme by genetic engineering leading to the expression of a fusion protein.Many reviews deal with the description of commercially available tagging systems, as for example the paper published by Terpe in 2003 that summarizes molecular basics and the development of such systems. 38An overview of the most prominent and promising examples used for site-directed immobilization is given in Table 1, demonstrating the binding partners and revealing selected examples of sophisticated immobilization approaches.Fig. 1 shows the immobilization of an enzyme by a Histidine-Tag using dip-pen-nanolithography.

Site-specic co-immobilization to protein scaffolds
For the generation of distinctive patterns, specically structured scaffolds are necessary.One potential approach to this is the application of protein scaffolds with specic domains to which modied proteins can bind orthogonally.In 2012 You et al. published a general approach for a self-assembling multienzyme-complex basing on a protein scaffold. 13The orthogonal binding mechanisms are mediated by the specic interactions between cohesin and dockerin domains (see Fig. 2).They are derived from a scaffold protein of the cellulosome, which constitutes the cellulose complex by binding different enzymes, carrying dockerin domains.Three enzymes from the glycolysis/ gluconeogenesis pathway, triosephosphate isomerase (TIM), aldolase (ALD) and fructose 1,6-bisphosphatase (FBP), were genetically modied with specic dockerin domains and bound by self-assembly to a synthetic trifunctional scaffoldin carrying the appropriate cohesin domains.The authors found that the overall performance of the cascade in co-immobilized form was enhanced up to 21.1-fold compared to soluble enzymes, especially at low substrate concentrations.

Site-specic co-immobilization to DNA scaffolds
DNA-macromolecules can also be used for creating specic patterns of immobilized enzymes, because of their capability for self-assembly, their physical and chemical stability and their backbone stiffness.Already in 1994, Niemeyer et al. coupled proteins to oligonucleotide strands and immobilized them to a complementary single strand of DNA, leading to macroscopic protein arrays. 50A more sophisticated method is the use of specically designed DNA-macromolecules that self-assemble by base hybridization into dened 2-and 3-dimensional shapes.This technology, called DNA-origami was originally invented by Rothemund in 2006. 51Underlying molecular principles and general considerations in the process of generating suitable scaffold structures were excellently reviewed by Feldkamp et al.   in 2006. 52Many approaches that couple enzymes to DNA microstructures use the biotin-streptavidin binding system for immobilization.However, due to the tetrameric structure of streptavidin and avidin the stoichiometry of the DNA-proteinconjugates is difficult to control.Thus, if the stoichiometry is important for the respective approach other binding mechanisms can be used. 53Most prominent example for the use of such DNA-protein-conjugates is the protein-microarray, where DNA-labeled proteins are site-specically immobilized to a matrix of complementary DNA-strands attached to a surface.For the generation of soluble biocatalytically active nanostructures, DNA-labeled proteins can be attached to a complementary single strand of DNA, leading to so-called linear protein-DNA-assemblies.One early example was published by Niemeyer et al. in 2002.Here, a bienzymatic assembly of NAD(P) H-FMN oxidoreductase and luciferase were site-specically immobilized to an complementary single strand DNA via the biotin-streptavidin binding system.The results clearly show, that the spatial orientation of the enzymes is of importance for the performance of the enzyme cascade. 54The same effect could be shown for the enzyme cascade described above, consisting of GOx and HRP. 55If two-or even three-dimensional scaffolds for protein attachment are required, the above described DNAorigami structures engineered by rational strand design can be applied.So far, only simpler approaches with model proteins are reported.For instance, in 2007 Duckworth et al. decorated a DNA tetrahydron 56 site-specically with GFP-molecules, using a click chemistry approach. 57In 2012 Fu et al. were able to immobilize HRP and GOx site-specically on DNA-origami tiles.Different distances between the enzymes, ranging from 10 to 65 nm were created (see Fig. 3) and further enzymes were immobilized, acting as bridges between the cascade enzymes.Enhanced activity could be observed for enzyme pairs who were in close proximity.However, activity decreased when enzymes were closer than 20 nm suggesting Brownian diffusion of intermediates are responsible for the variation in enzyme activity.The use of further noncatalytic proteins, connecting the hydration shells of the cascade enzymes also led to an enhanced cascade activity. 58As DNA proved to be a suitable scaffold for the site-specic immobilization of enzymes, it is likely to become established as a versatile tool for the immobilization of enzyme cascades exploiting the substrate channeling effect.

Site-specic co-immobilization in nanocontainers
An elegant way of the structured immobilization of a threeenzyme-cascade was published by van Dongen et al. in 2009 (see Fig. 4).The approach is based on porous polymersomes composed of isocyanopeptides and styrene block copolymers. 17n order to obtain a structured co-immobilization, mimicking the compartmentalization in living cells, three enzymes were immobilized to different locations of the polymersome: CalB was embedded in the bilayer membrane, GOx was encapsulated in the lumen and HRP was attached to the polymersome surface by means of click chemistry.A specic labelling with metal-ions and subsequent mass spectroscopic analysis revealed the desired distribution of enzymes.Those nanocontainers were shown to be able to convert glucose-acetate and generate a detectable signal of the dye 2,2-azinobis(3-ethylbenzothiazoline-6-sulfuric acid) (ABTS) upon its oxidation from colourless Fig. 3 Site-specific immobilization of HRP and GOx in defined distances ranging from 10 to 65 nm.Close proximity of the enzymes leads to an enhanced performance of the cascade due to substrate channeling.Reprinted with permission from. 58Copyright (2015) American Chemical Society.to coloured form.Furthermore it could be demonstrated that the containers can be easily separated from the reaction solution by ltration (Fig. 5).

Site-specic co-immobilization in microuidic devices
Microuidic devices exploit compartmentalization and efficient control of product and reactant streams, thereby aiming on mimicking the microcompartments of living cells.Microuidic scaffolds are oen combined with immobilized enzymes e.g. for sensor applications, analytics and the small-scale production of several agents.This topic was reviewed by Asanomi et al. in 2011 summarizing recent advantages in the development of micro-uidic reactors using immobilized enzymes.The authors focussed on the materials and production of such devices and the advantages of microuidics in general.Moreover commonly used immobilization techniques were highlighted. 59In this section, only a few of the most relevant examples will be discussed.The described enzymatic system for glucose detection was used by Boehm et al. in 2013 who designed a ow microreactor for synthetic enzyme reactions in vitro (see Fig. 3).A model enzyme cascade, consisting of b-Galactosidase (bGal), GOx and HRP for the conversion of galactose and a uorescent dye was implemented.The reactor along with its structures was produced by moulding polydimethylsiloxane (PDMS) on a fabricated master and closing it with a glass slide.In order to investigate two different compartmentalization mechanisms, enzymes were immobilized to microbeads and packed subsequently into a microuidic channel or attached directly to the inner surface of the microchannels.By streaming the channels with substrate solution and readout of the product formation, different kinetic parameters of the reaction were determined and the packed bed reactor with enzymes immobilized to microbeads was shown to be 1.5-fold more efficient than the reactor device with enzymatically active microchannels. 18The same enzyme system was implemented by Fornera et al. in 2012, who introduced a ow-through microuidic device containing the enzymes immobilized to a dendronized polymer in a predetermined pattern that was obtained by a valve system. 19The system can be applied for the determination of lactose in different lactose-containing solutions by measuring the resulting concentration of uorescent resorun, generated by the enzyme cascade.Another sophisticated microuidic system for the realization of complex enzymatic cascades was currently published by Huebner et al. in 2015.In the introduced system, reaction environments are realized by aqueous plugs separated by immiscible uidic plugs that are pumped through the reaction cascade of enzymes.The applied biocatalysts are immobilized to magnetic microparticles, that allow the fast and easy separation from the product stream and can be resuspended in the reaction solution by applying alternating electromagnetic elds.(DOI: 10.1002/elsc.201400171).

Summarization of co-immobilization techniques
A concluding overview over all discussed techniques, including their advantages and disadvantages is given in Table 2.

Summary and outlook
As enzymes provide some important advantages over traditional chemical syntheses, they have been established as green and cost-saving alternative in many elds.The use of enzyme cascades broadens the potential applications due to complex reaction opportunities, while obtaining the high reaction specicity of enzyme.Immobilization of enzyme cascades allows additional advantages.The catalytic complexes can be easy separated from the product, unfavourable kinetics can be circumvented and by co-localization the performance of the cascade can be enhanced by substrate channelling.
In this review an overview over different immobilization techniques has been given.The focus was on the random or site-specic immobilization of enzyme cascades leading to highly active multi-enzyme complexes with enhanced stability and activity.A great variety of techniques and different supports with sophisticated features exists nowadays in order to provide an optimal solution for the realization of enzyme cascades in many elds of application.

Notes and references
Fornera et al. and Böhm et al. extended  the system by b-galactosidase which enables the detection of lactose.18,19

Fig. 4
Fig. 4 Site-specific immobilization of an enzyme cascade in polymersome nanocontainers: Candida antarctica Lipase B (CalB) is embedded in the polymersome membrane, glucose oxidase (GOx) is entrapped in the inner lumen of the container and horseradish peroxidase (HRP) is attached to the outer polymersome surface by a Click chemistry approach.Reprinted from ref. 17 with permission of John Wiley and Sons.

Fig. 5
Fig.5Co-immobilization of b-galactosidase (bGal, green), glucose oxidase (GOx, here GOD, yellow) and horseradish peroxidase (HRP, red) in a microfluidic channel by two approaches: (A) immobilization to microbeads that are subsequently packed in the channel.(B) direct attachment to the inner surface of the microchannels.The packed bed reactor (A) proved to be the more efficient approach.Reprinted from ref.18 with permission of The Royal Society of Chemistry.

Table 1
Summarization of the most prominent and promising tools for enzyme immobilization via genetically encoded tags.Listed are the respective tags that are fused to the desired enzyme, their binding partner which facilitates binding to a support and selected publications, in which the immobilization of proteins by genetically encoded tags was used

Table 2
Classification and summarization of techniques for the immobilization of enzymatic cascades.The immobilization chemistry, the respective supports, advantages and disadvantages together with selected literature examples are listed Open Access Article.Published on 22 April 2015.Downloaded on 3/26/2024 6:06:20 AM.This article is licensed under a Creative Commons Attribution 3.0 Unported Licence.