Glucose oxidase (GOx) is an important oxidoreductase enzyme with many important roles in biological processes. It is considered an ideal enzyme and is often called an oxidase Ferrari because of its fast mechanism of action, high stability and specificity. Glucose oxidase catalyzes the oxidation of β-d-glucose to d-glucono-δ-lactone and hydrogen peroxide in the presence of molecular oxygen. d-glucono-δ-lactone is sequentially hydrolyzed by lactonase to d-gluconic acid, and the resulting hydrogen peroxide is hydrolyzed by catalase to oxygen and water. GOx is presently known to be produced only by fungi and insects. The current main industrial producers of glucose oxidase are Aspergillus and Penicillium. An important property of GOx is its antimicrobial effect against various pathogens and its use in many industrial and medical areas. The aim of this review is to summarize the structure, function, production strains and biophysical and biochemical properties of GOx in light of its various industrial, biotechnological and medical applications.
If you are looking for more details, kindly visit our website.
Keywords: glucose oxidase, catalytic mechanism, GOx producing organisms, FAD binding domain, biosensors, nanosensors, antimicrobial effect
Glucose oxidase is an enzyme that has widespread applications in industry and biotechnology. Due to this, a deep understanding of its structure and function are warranted. Glucose degradation is the most universal metabolic process. In addition to its breakdown in glycolysis, glucose can also be directly oxidized to glucono-δ-lactone by a number of enzymes.
These fall into two classes: (1) the dehydrogenases glucose dehydrogenase (β-d-glucose: NAD(P)+ 1-oxidoreductase, E.C. 1.1.1.47) [1] and quinoprotein glucose dehydrogenase (d-glucose:ubiquinone oxidoreductase, E.C. 1.1.5.2) [2] and (2) the oxidases glucose oxidase (GOx; β-d-glucose:oxygen 1-oxidoreductase, E.C. 1.1.3.4) [3] and pyranose oxidase (pyranose:oxygen 2-oxidoreductase, E.C. 1.1.3.10) [4].
The dehydrogenases oxidize glucose in one step using a co-factor, either nicotinamide adenine dinucleotide (phosphate) (NAD(P)+) or pyrroloquinoline quinone (PQQ), as the electron sink while the oxidases use a two-step mechanism in which a bound flavin adenine dinucleotide (FAD) co-factor is used to oxidize glucose to form glucono-δ-lactone and an enzyme-FADH2 intermediate followed by electron transfer to O2 to form H2O2 (Figure 1). The principal difference between GOx and pyranose oxidase is that the former is specific to β-d-glucose while the latter is also able to act on d-xylose, l-sorbose and d-galactose.
The general reaction of GOx.
The high specificity, high turnover and high stability of GOx make it an ideal enzyme for biosensor applications [5], some of which will be described below. Although its rate constant is still several orders of magnitude below the diffusion limit, GOx has a much higher kcat/KM (on the order of 106M1·s1) compared with most other oxidoreductases, prompting at least one researcher to call it the Ferrari of the oxidases [6].
GOx is a member of the glucose-methanol-choline oxidoreductase (GMC oxidoreductase) superfamily. The members of this family are all FAD-dependent oxidoreductases that share a common fold [7,8]. They consist of two functional domains, an N-terminal FAD-binding domain, which contains a strictly conserved βαβ mononucleotide-binding motif and a more variable substrate binding-domain. As the name suggests, the members of this family oxidize a variety of substrates containing hydroxyl functional groups, including mono and di-saccharides, alcohols, cholesterol and choline. GOx is perhaps the most thoroughly characterized of these, and its mechanism will be described more thoroughly below.
Glucose oxidase has widespread applications in industry and biotechnology. Due to this, a deep understanding of its structure and function are warranted. Perhaps surprisingly for such a well-studied enzyme, there are only seven crystal structures of glucose oxidase available, six from A. niger and one from P. amagasakiense [9,10,11,12]. The A. niger GOx structures range in resolution from 2.41.2 Å (Table 1); five of them are in hexagonal space groups (P or P), and one is in an orthorhombic one (C). Although the active form in solution is a dimer, the asymmetric unit of each structure contains only monomer with the second monomer occupying a symmetry-related site.
Tertiary structures of glucose oxidase determined by X-ray crystallography.
PDB ID Resolution (Å) Source Reference Remarks1gal
2.3 A. niger [9] The earliest and lowest resolution structure.1cf3
1.9 A. niger [10]1gpe
1.8 P. amagasakiense [10] The only structure with a complete dimer in the asymmetric unit.3qvp
1.2 A. niger [11] The highest resolution structure available.3qvr
1.3 A. niger [11]5nit
1.9 A. niger [12] A2 Mutant engineered for higher stability and turnover: T30V, I94V, A162T, R537K, M566V.5niw
1.8 A. niger [12] A9 mutant: T30V, R37K, I94V, V106I, A162T, M566V. Open in a new tabThese structures are all very similar, with a mean RMSD of 0.31 ± 0.06 Å over an RMSD range of 0.150.40 Å. The P. amagasakiense structure is unique in having a complete dimer in the asymmetric unit of space group P. This is very similar to the A. niger GOx structures, with an average RMSD of 0.68 ± 0.04 Å over a range of 0.640.77 Å. As it has been the most thoroughly studied, A. niger GOx will be described below; however, because of the high structural similarity, much of the description of the overall structure will also apply for P. amagasakiense GOx.
The A. niger GOx monomer forms a rough, rounded rectangular prism shape, measuring roughly 60×52×37 Å. The monomer is comprised of a single folding domain with a rather complicated topology. When viewed in a manner that emphasizes the secondary structure elements while minimizing the contribution of the unstructured loops (as in Figure 2a,b), two subdomains can be seen, each one centered around a five-stranded β-sheet. These two subdomains can be associated with FAD-binding on the one hand and substrate binding on the other.
The overall structure of A. niger glucose oxidase. (a) The overall structure of the GOx monomer. The two central β-sheets are colored orange (parallel β-sheet A) and light brown (anti-parallel β-sheet C) and labeled. The two loops that participate in the dimer interface are colored magenta. (b) The GOx dimer. The two monomers are colored green and cyan, the FAD co-factor is shown as van der Waals spheres, the carbohydrate chains attached to Asn-89 that participate in the dimer interface are shown as sticks, and the loops that participate in the dimer interface are colored magenta. The more important of the two loops is labeled. (c) A closer view of parallel β-sheet A, which lies at the topological center of the enzyme. The β-sheet, together with the α-helix that forms part of the βαβ motif, is colored by residue from the N-terminus to the C-terminus based on sequence position to illustrate that all regions of the sequence take part in its formation. The α-helices and anti-parallel β-sheet that flank β-sheet A are shown in light blue. (d) A closer view of the anti-parallel β-sheet C, which forms a Greek-key motif. The residues surrounding β-sheet C form the second sub-domain and are colored by position from the N-terminal end to the C-terminal. The cytosolic-facing α-helices following β-strands C2 and C4 lie behind the central β-sheet in this view. All panels show the PDB structure 1cf3 [10].
The first, FAD-binding domain is centered on a five-stranded parallel β-sheet (β-sheet A in Figure 2c), which is flanked by a smaller, three-stranded anti-parallel β-sheet on one side and three α-helices on the other. It contains the FAD-binding site and is similar to the FAD-binding motifs found in other FAD-binding, GMC oxidoreductase proteins, several of which are listed in Table 2. The most highly conserved part of this motif is the βαβ motif comprising strands A3, A4 and the α-helix between them; a ψ-BLAST search returns only local alignments to this region on the second iteration.
Proteins from the PDB similar in structure to A. niger glucose oxidase. Structures were identified by a BLAST search against the PDB using the A. niger sequence. Listed structures covered at least 75% of the query sequence and represent unique structures. All these proteins are members of the GMC oxidoreductase family.
PDB ID RMSD Description Identity (%) Reference6xut
1. Trametes cinnabarina Oligosaccharide dehydrogenase 35.89 [13]4ynt
1. Aspergillus flavus FAD glucose dehydrogenase 35.69 [14]6zh7
1. Chlorella variabilis Fatty acid Photodecarboxylase 27.99 [15]5oc1
1. Pleurotus Eryngii aryl-alcohol oxidase 29.67 [16]6ze2
1. Chaetomium thermophilum FAD-dependent oxidoreductase 31.24 [17]4h7u
1. Agaricus meleagris pyranose dehydrogenase 28.60 [18]5hsa
1. Pichia pastoris Alcohol Oxidase 23.56 [19]6h3g
1. Phanerodontia chrysosporium Alcohol oxidase 24.81 [20]3nne
1. Arthrobacter globiformis choline oxidase 26.31 [21]4ha6
1. Mesorhizobium loti pyridoxine 4-oxidase 26.85 [22]6f97
1. Methylovorus sp. 5-(hydroxymethyl)furfural oxidase 28.52 [23]3q9t
1. Aspergillus oryzae formate oxidase 25.55 [24]6o9n
2. Myceliophthora thermophila aryl-alcohol oxidase 30.75 [25] Open in a new tabThe parallel β-sheet at the center of this subdomain lies at the topological center of the molecule. Starting with strand A3, the chain traces out the βαβ motif (that is, A3α-helixA4) followed by a long loop forming part of the dimer interface and an extension forming about one-quarter of the second subdomain. The trace then returns to form strand A5. After A5, the chain forms a three-stranded anti-parallel β-sheet that forms part of the FAD-binding domain, and this is followed by strand A2. After A2 comes the remainder of the second subdomain followed by the last strand of the parallel β-sheet (A1). This is then followed by a long C-terminal α-helix.
In this context, it can be seen that the residues responsible for forming the FAD-binding pocket come from all parts of the polypeptide chain. The majority of the binding site is formed by residues from the N-terminal part of the sequence. Residues from the central region provide additional support, while most of the residues from the C-terminal region form the edges of the flavin binding pocket and also contribute two of the three active site residues.
The second subdomain is centered around a five-stranded anti-parallel β-sheet, which is supported by six α-helices (Figure 2d). The first of the five β-strands, C1, is much shorter than the others and arises from the extension following strand A4 of the FAD-binding domain. The remaining four strands are arranged in a Greek-key motif: strands C5 and C2 are connected through a loop that passes over strands C3 and C4, which are joined together by a hairpin loop.
Strands C2 and C3 are connected through a large excursion that forms three of the supporting α-helices. Following C4 is a long loop that forms the second part of the dimer interface followed by the last to helices of the subdomain. The six α-helices lie on the cytosolic side of the β-sheet, while the β-sheet itself forms one side of a deep pocket that has the active site and the flavin at the bottom.
Overall, this subdomain is similar to the binding sheet subdomain of cholesterol oxidase [26]. Several of the residues thought to be important for binding β-d-glucose are found here, including Trp-426, Phe-414 and Glu-412. The active site will be described more closely in Section 3 where the catalytic mechanism is described.
The A. niger GOx dimer measures approximately 60×52×77 Å. The FAD groups are located near the dimer interface but are more than 22 Å apart, making it unlikely that they communicate with one another through allostery (Figure 2b). The dimer interface is predominantly formed by residues from the long loop following strand A4 of the parallel β-sheet of the FAD-binding subdomain (residues 7598) and from the loop following strand C4 of the anti-parallel β-sheet of the substrate-binding subdomain (residues 432455). The loop containing residues 7598 covers part of the FAD-binding pocket.
It was previously thought that the apo form of GOx was a monomer [27] and that FAD binding was coupled to dimer formation [28,29], and the position of this loop appeared to support those conclusions. More recent work has shown, however, that after the dissociation of FAD, the enzyme is not in a monomeric state but appears to form aggregates [30] and that the dimer does not dissociate during thermal denaturation [31]. GOx does appear to dissociate into monomers at pH 5 and below in the presence of sodium n-dodecyl sulfate [32].
In addition to forming part of the FAD-binding site, residues 7598 also from part of the wall of the active site cavity of the second monomer in the dimer: the complete dimer is needed to properly assemble a complete active site cavity, which may account at least in part for the dimeric form of the holoenzyme [27,33].
In the complete dimer, the active site lies at the bottom of a deep, cone-shaped cavity whose apex is centered on the N5 atom of the middle flavin ring, the one that accepts the hydride during reduction. Only three amino-acid side-chains are near this center, His-516, His-559, and Glu-412 (Figure 3). His-559 is fixed by a strong hydrogen-bond to Glu-412 and Glu-412 is largely fixed by the side-chains of the surrounding residues (Ala-349, Phe-351, Phe-414, Trp-426 and Leu-428). The side-chain of His-516 is less constrained and is more conformationally flexible [12].
The active site of A. niger GOx. (a) The active site of the oxidized form of the enzyme in the absence of any substrate. The residues potentially important for catalytic activity are colored magenta; those which are thought to be involved in binding β-d-glucose are colored cyan. The water molecule near the center is thought to indicate the approximate location of the β-d-glucose O1 hydroxyl in the substrate-bound conformation. PDB structure 1cf3 [10] is shown. (b) The active site of an engineered GOx mutant showing an O2 molecule bound to the active site (PDB structure 5nit from Petrović et al. [12]).
As all these structures were produced from crystals grown at pH 5.16.9, it seems likely that both histidine residues are protonated on both N atoms. All but one of the GOx crystal structures also have a water molecule at the center of the active site, forming hydrogen-bonds to both His-516 and His-559 and lying 3.0 Å away from the N1 ring of FAD (Figure 3a). The location of this water molecule has been taken to represent the likely position of the O1 hydroxyl group of β-d-glucose [10]. Kinetic, structural and thermodynamic data have allowed likely roles to be assigned to these residues.
Despite numerous attempts, no structure of GOx with a bound substrate has been determined [10,12]. Wohlfahrt et al. [10] used manual docking and molecular dynamics simulation to characterize a likely enzymesubstrate complex using oxidized A. niger GOx and β-d-glucose. Their modeling suggested that the residues most important for the enzymesubstrate complex included the catalytic His-516 and His-559 along with Tyr-68, Thr-110, Phe-414, Trp-426, Arg-512 and Asn-514. His-516 and His-559 formed hydrogen-bonds directly with the O1 hydroxyl group, which is expected to lose a proton in the catalytic step, while Asn-514, Arg-512, Tyr-68 and O4 of the FAD all form hydrogen-bonds to the remaining glucose hydroxyl groups, thereby, anchoring the substrate in position.
The catalytic mechanism of GOx has been studied for many years, primarily using kinetic methods [34]. These, together with the enzyme structure, provide the identities of the active-site residues and allow a description of the mechanism of the reaction to be proposed. Many many different substrates and electron-acceptors have been employed (see Table 3); however, we will concentrate on the natural reaction using β-d-glucose as the substrate and O2 as the electron acceptor.
Some glucose oxidase substrates.
Substrate GOx Activity (%) 1 Referenceβ
-d
-glucose 100 [35,36,37,38] 2-deoxy-d
-glucose 2530 [36,37,38] 4-O-methyl-d
-glucose 15 [37] 6-deoxy-d
-glucose 10 [37] 4-deoxy-d
-glucose 2 [37] 2-deoxy-6-fluoro-d
-glucose 1.85 [38] 3,6-methyl-d
-glucose 1.85 [38] 4,6-dimethyl-d
-glucose 1.22 [38] 3-deoxy-d
-glucose 1 [37] 6-O-methyl-d
-glucose 1 [37]α
-d
-glucose 0.64 [37,38] mannose 0.2; 1 [36,37,38] altrose 0.16 [37,38] galactose 0.08 [36,37,38] xylose 0.03 [36,37,38] idose 0.02 [37,38] Open in a new tabGOx operates using a Ping-Pong Bi Bi mechanism in which β-d-glucose oxidation and O2 reduction occur in two different steps [35,39,40,41] (Figure 4). This allows the oxidation and reduction half reactions to be analyzed independently using steady state kinetics. pH profiling showed that the pKa of the amino-acid side-chains in the active site of the enzymesubstrate complex is between 6.97.8 [35,42].
A scheme illustrating the overall glucose oxidase reaction. A indicates glucose, P indicates glucono-δ-lactone, B represents O2, Q represents H2O2, E represents the oxidized form of GOx bound to FAD, and F stands for the reduced form of GOx bound to FADH2. The individual components of the reaction are illustrated below the scheme; from left to right they are reduced GOx (E), the most likely mechanism of the GOx reduction half-reaction (EA FP), the GOxO2 complex (FB) and the GOxH2O2 complex (EQ).
Several lines of evidence ascribed this pKa value to His-516 [40,41]. The most likely mechanism involves a base-catalyzed hydride transfer from the glucose C1 to the flavin N5. Glucose binding is thought to displace the water molecule observed in the active site of the GOx crystal structures, which abstracts a proton from His-516 as it leaves (this abstraction appears to be necessary, according to modeling studies, because His-516 cannot adopt a proper catalytic position when doubly-protonated) [10].
Hydride transfer occurs in a concerted step in which a proton is removed from the glucose O1 hydroxyl group by a basic group on the enzyme while a hydride is transferred from the glucose C1 to the flavin N5 [43]. The kinetic rate constants associated with this step show a large H/D kinetic isotope effect, suggesting that this transfer is the rate-limiting step of the reductive half-reaction [44]. The basic group on the enzyme is likely to be His-516: a His-516-Ala mutant had practically no catalytic activity [41].
Considerations of resonance suggest that hydride transfer should create a negative charge around the FAD N1 atom, which have been confirmed by NMR studies of GOx at pH 5.6 in the absence of oxygen [45]. Following the reduction half-reaction, the enzyme is left in a reduced form with a bound glucono-δ-lactone. The product is then displaced by water or possibly O2 and the oxidation half-reaction follows.
The kinetic and thermodynamic parameters for the oxygen-binding step suggest that O2 easily and rapidly diffuses into the enzyme [35,44,46]. Indeed, structural studies on mutant GOx enzymes engineered to be more stable and have greater catalytic efficiency found O2 bound in the active site [12] (Figure 3). The oxygen bound in a small pocket formed by shifting His-516 up towards the active-site opening and displacing the active-site water molecule so that it lay outside hydrogen-bonding range of FAD (its closest approach was 3.7 Å from the N5 atom of FAD).
In its ground state, molecular oxygen is a paramagnetic triplet, making its insertion into diamagnetic organic molecules a spin-forbidden process [47]. In the active site, one of the O2 atoms is 2.7 Å away from the N1 flavin ring while the second is 3.0 Å away from the His-516 ring. This position led Petrović et al. [12] to suggest that the oxidative half-reaction might rely on orbital coupling between the oxygen and the π electrons in the His-516 side-chain to overcome this limitation.
Generally, O2 oxidizes organic substrates by transferring electrons one at a time, forming free-radical intermediates, and most analyses of of the oxidative half-reaction assumed a step-wise electron transfer [34,48]. The rate-limiting step appears to be the transfer of the first electron from flavin to O2 to produce the flavin semiquinone radical and the superoxide anion [41]. Kinetic isotope effect experiments show that no proton transfer takes place during this step [35,44]. The second electron transfer step, from the flavin semiquinone radical to the superoxide anion, happens quickly, with a second-order rate constant of 109 M1· s1 [49].
Consistent with its role as an extracellular protein, GOx is glycosylated, with carbohydrates, mostly mannose-like sugars, comprising between 10 and 16% of its final molecular weight [27,36,50]. Both N-linked and O-linked sugar chains are present. The removal of 95% of the GOx carbohydrate content had a noticeable effect on the kinetics of glucose oxidation, its stability at low pH and the number of available isoelectric forms but did not affect its thermal stability or the optimal pH and temperature of the catalytic reaction [50,51].
More particularly, a study using H/D isotope substitution found that reduced glycosylation appeared to lower the enthalpy of activation and, therefore, increase the activity of the enzyme [52]. Deglycosylation typically removes all of the O-linked sugars at their linkage to the amino-acid side chain but leaves the first N-acetylglucosamine (Nag), potentially allowing those sites to be identified in the crystal structure [53].
Potential glycosylation sites can be identified by the consensus sequence N-X-T/S, where X is any amino-acid residue except proline [54]. Eight such sites are present in A. niger GOx (residues Asn-43, 89, 161, 168, 258, 355, 388 and 473); seven of these sites have been confirmed in at least one of the known A. niger GOx structures. Asn-43 does not appear to be glycosylated in any of these structures, and peptide sequencing work also suggested that no carbohydrate is attached here [54].
GOx is used in many branches of industry because of its ability to oxidize glucose and produce hydrogen peroxide. Its rapid turnover and high stability finds it many applications in the food, pharmaceutical, medical, textile and power industries. For many of these applications GOx is used in a biosensor or nanosensor [94,95,96,97,98,99], in nanoparticles [100,101,102] or in nanosheets [103].
In many modern applications, GOx is often used in combination with other enzymes, for example, tyrosinase in the analysis and discrimination of musts and wines [104], α-amylases and xylanases for improving the quality of dough and bread [105], peroxidase for accurately measuring the level of glucose in blood and saliva [106] and tears [94], the autophagy inhibitor chloroquine in cancer intervention therapy [101] and insulin for regulating blood glucose levels in diabetes [107,108]. Finally, it has been combined with the anti-cancer drug tirapazamine and human serum albumin to create a nanoreactor capable of increasing the levels of hypoxia and reactive oxygen species and inhibiting tumor growth [109].
The applicability of GOx primarily depends on its quantity, thermal stability and activity. Many studies focused on identifying which fungal strains are better for biosensor development, which are better for clinical studies and which are better for biochemical diagnostic tests [55]. Optimal GOx utilization also requires consideration of the type of matrix on which GOx is bound and type of media and conditions under which it is used. Consequently, in addition to identifying the best GOx producers and the ideal conditions for its stability and activity, the development of different binding materials, environmental conditions and detection systems are also very important for expanding the range of GOxs industrial applications.
Industrially, the most important producers of GOx are species belonging to the genera Aspergillus and Penicillium [55,110,111], especially A. niger and P. amagasakiense. These genera are important because of their metabolic versatility and because they are generally regarded as safe by regulatory agencies [112].
Overall, P. amagasakiense GOx is catalytically more effective than A. niger GOx, with a 6× lower Michaelis constant (Km) for β-d-glucose and 10× higher catalytic efficiency (kcat/Km) [113]; however, it is also less stable and has a lower antimicrobial activity [114]. This prompted at least one group of researchers to create a hybrid GOx with improved stability and catalytic efficiency by genetically combining elements from both A. niger and P. amagasakiense GOxes [115]. Other researchers have attempted to improve the antioxidant capabilities [116] and thermal stability [117] of A. niger GOx.
Fungal GOx is produced by solid state fermentation (SSF) and submerged fermentation (SmF) [118,119]. Submerged fermentation is the more useful method because environmental factors can be controlled more easily. A study on the variability of GOx from Aspergillus tubingensis CTM507 produced by SSF and SmF found that SmF produced GOx more efficiently but that the GOx produced by SSF had a higher activity (170 U mL1 for SSF versus 43.73 U mL1 for SmF) [120]. Both of these methods have limited capabilities, so more efficient ways of producing GOx are being researched using genetic recombination techniques, mutagenesis and immobilization methods [3,55].
Glucose and saccharose are the carbon sources most often used for the industrial production of GOx [3,121]. GOx production can be induced from A. niger by glucose [121], CaCO3 [122], Mn2+, Co2+, thioglycolic acid, gluconic acid [123], EDTA (ethylene diamine tetra-acetic acid) Zn2+ and Fe2+ [116]. Inhibitors of A. niger GOx include the Ag+, Hg2+, Cu2+ and Mg2+ ions, CaCl2 [124,125], hydrogen peroxide accumulation [116], arsenates, p-chloro-mercapto-benzoate, phenyl mercuric acetate [125], hydroxylamine, hydrazine, phenyl hydrazine, dimedone, NaHSO4 [126], guanidinium chloride, urea and SDS (sodium dodecyl sulfate) [116].
The low fermentation capacity, complicated purification process and effectiveness of GOx limit the applicability of GOx from natural sources. Attempts to improve GOx production through the mutagenesis and screening of production strains has enhanced GOx production in A. niger by about 77% [127]. Genetic recombination has also been used to express GOx heterologously. The most effective cloning techniques and highest overexpression levels were reported for Saccharomyces cerevisiae and Escherichia coli and other Aspergillus and Penicillium species have also been used as hosts [128,129,130].
Other candidates for the industrial production of GOx are H. polymorpha and P. pastoris [131,132]. Along with S. cerevisiae, these yeasts have the advantages of rapid growth and extracellular protein production; the highest level of A. niger GOx production from S. cerevisiae is currently 9 g L1. Unfortunately, GOx produced by S. cerevisiae and H. polymorpha tend to produce over-glycosylated forms of GOx with reduced activities [133]. However, P. pastoris has been successfully used for the production of a GOx from A. niger and P. variable P16, which was not over-glycosylated [134,135].
GOx has several important applications in the food industry [55], including the baking industry, the production of drinks, the production of gluconic acid and food preservation. In the baking industry, GOx is used as an oxidant to improve the quality of bakery products [136,137,138]. The hydrogen peroxide produced by GOx makes the dough more elastic and viscous [139] and lipase and GOx can increase the quality and durability of bread [140].
Generally, bread quality depends on wheat quality and dough rheology depends on the quantity of enzymes added [141,142]. GOx causes the formation of protein fibers in dough. Basal additives, such as GOx, ascorbic acid and α-amylases decrease bread fragility and enhance chewability, adhesion, elasticity and cohesion [143]. The addition of fungal xylanase has a positive effect on flours and crumb firmness [105,144].
GOx also finds use in the reduction of the alcohol content of wine. Warmer temperatures during the growth season are likely to increase the level of glucose in wine grapes. By decreasing the level of glucose, which would otherwise be transformed into alcohol through anaerobic fermentation, GOx could decrease the quantity of alcohol in the resulting wine [145]. The hydrogen peroxide produced by GOx has a bactericidal effect on the corrosive acidic and dairy microbes produced during the fermentation process. The hydrogen peroxide produced can be removed by the enzyme catalase, which transforms H2O2 into oxygen and water. GOx in combination with catalase decreases the quantity of alcohol better than GOx alone by converting glucose into gluconic acid [146].
For measuring the quantity of glucose in liquids, GOx has also been incorporated into analytical devices known as biosensors. These devices combine a biological component, frequently some kind of enzyme or antibody, with a physical or electrical transducer and an electrical component to provide a measure of the analyte of interest [147]. For example, Garcia-Hernandez et al. [104] constructed a bioelectronic tongue consisting of enzymes (tyrosinase and glucose oxidase) and polypyrrole or polypyrrole/Au nanoparticles to measure the alcoholic degree in musts and wine.
The information from this sensor can help to predict the characteristics of the finished wine at the beginning of the vinification process. Lopes et al. [148] created a biosensor that consisted of GOx and immobilized horseradish peroxidase to determine the quantity of glucose in beverages, such as orange juice and energy drinks. Finally, Mason, Longo and Scampicchio [149] created an electrochemical biosensor that consisted of GOx immobilized on a nylon nano-fiber membrane to determine the level of glucose in brewed beer.
Such devices also allow the level of glucose to be measured in sugar-free foods intended for diabetics. For example, Kalaivani et al. [98] constructed a reliable sensor for detecting nanomolar amounts of glucose in food products.
You will get efficient and thoughtful service from Yulin HB™.
Its production of gluconic acid also finds some uses for GOx in the food industry. Gluconic acid is used as an acid regulator, color stabilizer, antioxidant and chelating agent in food and drinks [150]. Gluconic acid is also applied in the dairy industry in cheese curd production, improving thermal stability of milk and cleaning of aluminum tins. The most popular application of gluconic acid in the food industry is its use as an acid regulator and antioxidant [151].
By removing oxygen and glucose from food, GOx can be used to lengthen shelf-life through two reasons. First, during food preservation, an unwanted reaction may occur between the nuceophilic functional groups on amino acids with the active carbonyls on sugars to cause non-enzymatic browning through the Maillard reaction. By removing unwanted sugars from the food to be preserved, this browning is prevented.
For example, using GOx to remove glucose residues from dried eggs improves their durability [152]. The hydrogen peroxide produced during this reaction also helps to destroy all pathogen microbes potentially present in raw eggs. Extra hydrogen peroxide can then be removed by the enzyme catalase [153]. The combination of GOx and catalase can also be used to control the non-enzymatic browning of fruit and tomato paste.
The second reason is that the extra oxygen can support bacterial growth, and removing the excess oxygen will inhibit the growth of aerobic bacteria. The removal of excessive oxygen is essential for canned foods [154] and Karimi et al. [155] studied the removal of dissolved oxygen in water through its reduction by glucose, catalyzed by glucose oxidase and catalase. GOx can also be used to remove of oxygen from canned drinks (such as beer and wine) to help them keep their color and taste [156,157]. The decomposition of mayonnaise is connected with lipid peroxidase [158] and GOx and catalase can slow down lipid peroxidation: oxygen removed during glucose reduction will not be available for lipid metabolism.
GOx is also used directly as an antimicrobial agent in the food industry. It was demonstrated to decrease the growth of many pathogenic bacteria, including Clostridium perfringens, Campylobacter jejuni, Salmonella infantis, Staphylococcus aureus and Listeria monocytogenes [159]. Vartiainen, Ratto and Paulussen [160] found that immobilized GOx inhibits the growth of Escherichia coli and Bacillus subtilis and Malherbe et al. [130] demonstrated that S. cerevisiae producing GOx from A. niger had antimicrobial activity against bacteria that produce lactic acid and acetic acid.
Polyamide and ionomer films with immobilized GOx inhibited the growth of E. coli CNCTC , Pseudomonas fluorescens CNCTC , Lactobacillus helveticus CH-1, Listeria ivanovii CCM and Listeria innocua CCM on agar medium [161]. Yuan et al. [162] developed a photodynamic antimicrobial system consisting of glucose, GOx and horseradish peroxidase to inactivate bacterial and fungal pathogens. Finally, Xu et al. [163] developed a biosensor comprising of antibodies, GOx, gold nanoparticles, shell magnetic beads, polydopamine and polymeric nanocomposites for the detection of pathogens in foods. This biosensor was able to detect E. coli O157:H7 with a detection limit of 102 cfu/mL.
The ability of GOx to consume intracellular glucose and oxygen to produce hydrogen peroxide and gluconic acid might allow it to be used in certain cancer therapy combinations. By consuming glucose, GOx could reduce the available metabolic energy sources of cancer cells, thereby, inhibiting their proliferation, and by consuming oxygen and producing gluconic acid, it could increase the hypoxia and acidity of the tumor microenvironment.
In several of these proposed combination treatments, GOx is embedded in nanocomposites. These come in a variety of forms, including hollow mesoporous silica nanoparticles, metalorganic frameworks, organic polymers, and magnetic nanoparticles are used for the construction of GOx-based nanocomposites for multi-modal synergistic cancer therapy [164].
In one proposed treatment [101], GOx chloroquine are attached to the surface and loaded into the cavity of rattle-structured polydopamine core-hollow mesoporous silica shell nanoparticles (PDA@hm). These nanoparticles are then used for a therapy combining energy metabolism regulation (by GOx), autophagy inhibition (by chloroquine) and low-temperature photothermal therapy (induced by the PDA nanocore). Photothermal therapy uses light-absorbing materials to convert photoenergy into local hyperthermia to destroy cancerous tissues [165,166].
In this treatment, GOx served to starve the tumor and directly suppressed the expression of the heat-shock proteins HSP70 and HSP90. In another proposed treatment [99], GOxMnO2 nanosheets were developed to destroy cancer cells using a combination of starvation and self-oxygenation by GOx and photothermal therapy by the MnO2 nanosheet. In these nanostructures, the GOx catalytic activity could be enhanced by the hyperthermia triggered by near-infrared laser radiation.
These GOxMnO2 structures exhibited pH and glucose responsive performance, activated by magnetic resonance and photoacoustic dual-modal imaging. Finally, GOx formed one component of a recently-designed nanoreactor consisting of a tirapazamine (TPZ)human serum albuminGOx mixture combined with a metalpolyphenol network consisting of Fe3+ ions and tannic acid [109].
This complex kills cancer cells by producing HO, a process termed chemodynamic therapy (CDT) and TPZ radicals. The HO radicals are produced from H2O2 by Fe2+/Fe3+ ions through the Fenton reaction [167], while TPZ develops into a toxic radical under conditions of tumor hypoxia. The role of GOx in this assembly is to consume oxygen to increase the tumors hypoxia level, to produce H2O2 for conversion to HO and to consume glucose for starvation therapy.
The earliest suggestion for a glucose sensor involving GOx was in [168]. Today, GOx is extensively used in the most common methods for measuring blood glucose levels [169,170]. Here, we limit ourselves to some relatively recent developments. GOx activity is highly oxygen dependent, which can lead to inaccuracies in amperometric β-d-glucose determinations.
Fokkert et al. [171] investigated the performance of fluorescence sensor-based and GOx-based glucose measurement during intensive exercise, when oxygen consumption is expected to be higher, and normal daily activities. They found that both methods were less accurate during exercise than during daily activities, and this finding persisted over the whole range of glucose concentrations examined.
Gutierrez et al. [172] attempted to overcome this problem by engineering GOx variants with lower oxygen dependency through random mutagenesis using error-prone PCR and sequence saturation. They discovered which positions seemed to be vital for oxygen sensitivity and for oxygen activity. One of their variants had a 37-fold reduced oxygen dependency but maintained the same β-d-glucose specificity and thermal resistance.
Other studies have attempted to use GOx to measure effective blood glucose levels non-invasively by monitoring other body fluids. Thus, Mohammadnejad et al. [106] developed a more precise, accurate and rapid method to measure blood glucose using GOx and peroxidase to oxidize a 4-[(hydroxy-3-methoxyphenyl)-azo]-benzenesulfonic acid (GASA) substrate. The stability of GASA and its oxidized products along with its direct and fast consumption by peroxidase, not only made it possible to determine blood glucose concentration with high reproducibility but also allowed it to be used for salivary samples.
A low-cost non-invasive paper-based biosensor for glucose measurements from tears has also recently been developed [94]. Finally, to eliminate some limitations and errors in some of the more commonly used monitoring systems, Jędrzak et al. [97] developed a biosensor consisting of magnetite, lignin and polydopamine bound to GOx together with ferrocene and a dedicated carbon paste electrode. The results showed that this biosensor has a potential for application in the determination of glucose in various commercial products.
One of the more interesting developments in the use of GOx in glucose monitoring is its potential to be coupled to an insulin delivery system, which could improve the health and quality of life for many diabetics. Yang et al. [173] developed a glucose and magnetic-responsive microvesicle delivery system, which can both regulate glucose levels and generate nitric oxide.
The injectable microvesicles are loaded with GOx, which can reduce hyperglycemia by consuming excess glucose. An applied magnetic field can then alter the permeability of the microvesicle shell, allowing the H2O2 produced by GOx to react with l-Arginine within the vesicle to produce nitric oxide, which has been shown to be important in the early stages of glucose-stimulated insulin secretion [174].
A glucose-responsive closed-loop insulin delivery system mimicking the function of pancreatic cells has potential to improve quality of life and health in diabetics. Yu et al. [107] created a glucose-responsive insulin delivery closed-loop system using a painless microneedle-array patch containing glucose responsive vesicles that are loaded with insulin and GOx. Under conditions of hyperglycemia, the GOx consumes glucose and oxygen, creating hypoxic conditions, which then trigger the release of the insulin.
This device effectively regulated the blood glucose in a mouse model of chemically induced type 1 diabetes. In a later refinement, the same group made an enhanced system where insulin release contingent on both hypoxia and H2O2 [108]. This system could effectively regulate blood glucose in mice with chemically-induced type 1 diabetes for 10 h.
Several glucose-responsive nanoparticles are based on phenylboronic acid, which has better stability compared with protein-comprised systems [175]. Chai et al. [100] created a glucose-responsive insulin delivery system comprised of poly(acrylamido phenylboronic acid)/sodium alginate nanoparticles loaded with GOx. The GOx-laded nanoparticles showed greater glucose sensitivity and faster glucose-responsive insulin release than nanoparticles loaded with insulin alone. These nanoparticles can also be easily prepared and have good biocompatibility.
As mentioned above, GOx has several important applications in the food and medical industries. GOxs ability to continuously produce hydrogen peroxide at low concentrations has attracted a great attention in the area of wound management. One wound treatment that makes use of the GOx-mediated release of hydrogen peroxide involves honey. Honey has been used for the treatment of a wide range of injuries, including acute and chronic wounds and burns [176].
A plethora studies have provided compelling evidence that hydrogen peroxide is the major antibacterial compound found in diluted honey. GOx, likely together with some polyphenols found in honey, is primarily responsible for the hydrogen peroxide generated during nectar processing and honey ripening. GOx, together with other enzymes catalyzing the metabolism of sugar, is produced by the hypopharyngeal glands of worker bees and secreted into honey [177].
Therefore, GOx is a regular but quantitatively variable compound in natural honey [178]. The final amount of hydrogen peroxide in dilute honey is a result of the enzymatic activity of both GOx and pollen-derived catalase enzymes. Interestingly, honeydew honeys, which are rich in polyphenols, are able to generate a higher level of hydrogen peroxide than many other honey types. It is believed that polyphenols, including flavonoids, might significantly contribute to higher levels of hydrogen peroxide [179].
The efficacy of the topical application of honey in the treatment of various wounds has been well-documented enough to allow it to be registered as a medical device in wound care management. GOx itself has become an interesting therapeutic platform in wound care [180,181,182].
There are several clinically tested GOx-based products available in the market for wound application containing a GOx enzyme immobilized in a suitable carrier (e.g., a hydrogel), which gradually releases a controlled amount of hydrogen peroxide [183,184]. However, a GOx-embedded hydrogel does not perfectly replicate antimicrobial honey-mimetic mechanisms because glucose is not added into the hydrogel.
Furthermore, the high concentration of H2O2 in the hydrogel could damage the normal tissues around the wound site. Most importantly, reactive oxidative species (ROS), such as HO are more effective in inhibiting and killing bacteria than H2O2. A variety of nanomaterials exhibiting the enzyme-mimicking activity (called nanozymes) are able to convert H2O2 into ROS species.
The development of nanozymes, such as metal-based nanoparticles and carbon-metal hybrid nanomaterials, which can simultaneously exhibit dual or multienzyme mimetic activity, are a promising antibacterial therapy in wound care.
Nanozymes with embedded GOx have very recently become the object of several studies investigating their wound-healing properties [185,186,187]. In a study by Du et al. [187], a nanozyme comprised of clinically approved iron oxide nanoparticles coated with GOx exhibited GOx, catalase and peroxidase-like activities.
Interestingly, in neutral and acidic microenvironments, it exhibited pH-switchable GOx/peroxidase and GOx/catalase cascade reactions, respectively. Thus, a fabricated Fe3O4GOx nanozyme can, on the one hand, eradicate bacterial biofilm and shorten the inflammatory phase of wound healing and, on the other hand, accelerate the epithelialization and remodeling phase of wound healing. In addition, in vivo testing reveals that the Fe3O4GOx nanozyme had superior healing activities against diabetic wounds caused by methicillin-resistant Staphylococcus aureus compared to unjoined Fe3O4 nanoparticles and GOx.
Although these nanozymes show effective antibacterial and wound healing properties, their direct use in clinical settings is considered questionable. Their unique characteristics, including their small size, chemical composition and solubility, may create a high risk and a hazard for human health. To address the limits of nanozymes in wound care, nanozymes can be incorporated into suitable wound dressings [188].
Zhang and co-workers [186] developed a novel wound dressing prepared from a nanozyme constructed by assembling GOx onto a hollow mesoporous carbon nanosphere doped with single-atom Fe and bacterial cellulose enveloped polypropylene composites. This nanozyme-based wound dressing exhibited outstanding breathability, biocompatibility and water uptake and antibacterial and antibiofilm ability.
Glucose oxidase catalyzes the oxidation of β-d-glucose to d-glucono-δ-lactone, which spontaneously hydrolyzes to d-gluconic acid and hydrogen peroxide in the presence of molecular oxygen. Due to its fast turnover and high stability and specificity, glucose oxidase has found many uses in industry and medicine.
Uses have been found for all aspects of its reaction: its consumption of glucose, production of H2O2, consumption of oxygen and production of gluconic acid have all found practical applications. Glucose oxidase is among the earliest of the enzymes to be used in this way, and new applications continue to be found.
Although the biochemical and biophysical characteristics of GOx have been thoroughly studied, there are still areas where additional study would be beneficial. For example, all but one of the structures of GOx presently known are variations of A. niger GOx. Although a number of industrial strains with enhanced stability or activity have been prepared, there is no structural analysis of what makes these forms more stable. Despite numerous efforts, there is still no structure of a GOxsubstrate complex.
Moreover, some uncertainty still remains as to whether GOx might also be found in some bacteria or whether animals other than insects might produce it. Industrial research is also expected to develop new materials for biosensors and nanosensors and new ways of attaching GOx to them. In addition, GOx represents a promising molecule, which, in the form of nanozymes incorporated into wound dressings, may be of great use in wound care.
The authors wish to thank Eva Kutejová for general support during the writing of this article.
The following abbreviations are used in this manuscript:
GOx Glucose Oxidase FAD Flavin Adenine Dinucleotide NAD(P) Nicotinamide Adenine Dinucleotide (Phosphate) PQQ PyrroloQuinoline Quinone GMC Glucose-Methanol-Choline RMSD Root-Mean-Squared Deviation NMR Nuclear Magnetic Resonance NCBI National Center for Biotechnology Information SSF Solid State Fermentation PDA PolyDopAmine HSP70/90 Heat-Shock Protein 70/90 TPZ Tirapazamine CDT ChemoDynamic Therapy GASA 4-[(hydroxy-3-methoxyphenyl)-azo]-benzenesulfonic acid ROS Reactive Oxygen Species Open in a new tabWritingoriginal draft preparation, J.A.B., M.Z., J.M. and V.B.-H.; writingreview and editing, J.A.B. and V.B.-H.; visualization, J.A.B. and V.B.-H.; project administration, V.B.-H. and J.M.; funding acquisition, V.B.-H. and J.M. All authors have read and agreed to the published version of the manuscript.
This research was funded by Vedecká Grantová Agentúra MŠVVaŠ SR and SAV grant numbers 2//20 and 2//22.
The authors declare no conflict of interest.
Publishers Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.
The evolution from first-generation through third-generation glucose sensors has witnessed the appearance of a number of very diverse oxidoreductases, which vary tremendously in terms of origin, structure, substrate specificity, cofactor used as primary electron acceptor, and acceptable final electron acceptor. This article summarizes our present knowledge of redox enzymes currently utilized in commercially available glucose monitoring systems to promote a fuller appreciation of enzymatic properties and principles employed in blood glucose monitoring to help avoid potential errors.
Keywords: biosensing, glucose dehydrogenase, glucose oxidase, point-of-care testing, self-monitoring of blood glucose
Ever since Clark and Lyons1 introduced the first enzyme sensor employing glucose oxidase (GOx) and an oxygen electrode for glucose monitoring, extensive studies have been carried out to develop improved enzyme-based systems for monitoring glycemic levels. First-generation blood glucose monitoring systems employed oxygen as the electron acceptor, determining glucose concentration by following either the consumption of oxygen or the liberation of hydrogen peroxide (Figure 1). In second-generation sensors, enzymes transfer electrons to artificial electron acceptors (also referred to as electron mediators or redox dyes) instead of oxygen to avoid interference from other redox species. The reacted (reduced) artificial electron acceptors are monitored colorimetrically or electrochemically. Third-generation sensors employ direct electron transfer to the electrode, thus eliminating toxic artificial electron mediators and avoiding errors due to variations in the concentration of oxygen in blood samples. In addition to the three generations, blood glucose monitoring systems are further categorized based on the type of enzyme employed, whose individual characteristics vary tremendously according to several diverse criteria: reaction toward substrates, availability of electron acceptors, the nonprotein cofactor component (primary electron acceptor), protein structure (from primary to quaternary), presence or absence of glycosylation, as well as the host microorganism for the natural or recombinant production of the enzyme.
Schematic representation of the principles of first-, second-, and third-generation glucose sensors. Electrons from the glucose oxidation reaction are first taken up by the enzyme's cofactor (primary electron acceptor) and transferred to either oxygen (first generation), an electron mediator (second generation), or directly to the electrode (third generation).
Oxidoreductases that act on the CHOH group of donor molecules are classified into the Enzyme Commission group EC 1.1. This group includes several oxidoreductases employed for glucose monitoring, all of which act at the reducing end of glucose, on the first hydroxyl group. These enzymes are divided into two groups, GOxs and glucose dehydrogenases (GDHs), depending on their ability to react with external electron acceptors. Glucose oxidases are defined as oxidoreductases that can utilize oxygen as the external electron acceptor, liberating hydrogen peroxide, and are categorized in the EC 1.1.3 group. Glucose dehydrogenases are defined as oxidoreductases that are unable to utilize oxygen as the electron acceptor and instead transfer electrons to various natural and artificial electron acceptors. Glucose dehydrogenases are further subdivided according to their redox cofactors, which are the essential nonprotein component that act as the primary electron acceptor (Figure 2). The GDHs utilizing the cofactor nicotine adenine dinucleotide (NAD) or nicotine adenine dinucleotide phosphate (NADP) are categorized in the EC 1.1.1 group, whereas those utilizing the pyrroloquinoline quinone (PQQ) cofactor belong to group EC 1.1.5. The remaining GDHs belong to the EC 1.1.99 group, corresponding to redox enzymes utilizing other electron acceptors and which includes the GDHs utilizing flavin adenine dinucleotide (FAD) as the cofactor. Interestingly, whereas GOxs are able to utilize oxygen as well as a variety of other electron acceptors, FAD-harboring GDHs are unable to utilize oxygen, despite their harboring the same redox cofactor and possessing significant structural similarities with GOx.
Cofactors used by GDHs. The structures of FAD, PQQ, and NAD are shown in their oxidized state. Regions that undergo reduction are in bold, with the reduced form of the region drawn nearby. The region that distinguishes NAD (OH) and NADP (PO4H2) is circled.
This short review summarizes our present knowledge of redox enzymes currently utilized in commercially available glucose monitoring systems, GOxs and GDHs. Our goal is to promote a fuller appreciation of enzymatic properties and principles employed in blood glucose monitoring to help avoid potential errors.
Glucose oxidase (β-D-glucose:oxygen 1-oxidoreductase, EC 1.1.3.4) is a flavoprotein that catalyzes oxidation of β-D-glucose at its first hydroxyl group, utilizing molecular oxygen as the electron acceptor, to produce D-glucono-delta-lactone and hydrogen peroxide. Glucose oxidase is a homodimeric enzyme, with an FAD molecule noncovalently but tightly bound at the active site of each 80 kDa subunit. Glucose oxidase is a representative enzyme of the glucose/methanol/choline (GMC) oxidoreductase family. This large and diverse family includes many industrially important enzymes, especially in the area of diagnostics, such as cholesterol oxidase, alcohol oxidase, amino acid oxidase, and pyranose oxidase. Members of the GMC oxidoreductase family share a homologous structural backbone, including an adenine-dinucleotide-phosphate-binding βαβ-fold close to their amino terminus, as found in other flavoenzymes, and five other segments of conserved sequence dispersed throughout their primary sequences.
Glucose oxidase was originally isolated and produced from the fungus Aspergillus niger, which remains the most common source of GOx utilized for glucose monitoring and other industrial applications. Glucose oxidases are commercially available as both reagent grade and diagnostic reagent grade from various enzyme companies. Glucose oxidases have since been isolated from several different fungi as well as a few insects. An important characteristic of A. niger-derived GOx is its very high substrate specificity for glucose. Some of the other GOxs have been reported to also oxidize other sugars, including maltose. The high degree of specificity of GOx for glucose and the enzyme's application for measuring glucose in the presence of other sugars was documented in the s.2
Natural production of GOx by fungi, usually A. niger or Penicillium, produces an extracellular enzyme that is over 10% glycosylated. Recombinant production of GOx in yeast, such as Saccharomyces cerevisiae and Hansenula polymorpha, often results in hyperglycosylation with a concomitant decrease in enzymatic activity.3 Methylotrophic yeast Pichia pastoris was found to be an effective host for the recombinant expression of GOx, producing large quantities of a more moderately overglycosylated enzyme with properties that differed slightly from the native enzyme. Recombinant production in Escherichia coli produces inactive GOx in inclusion bodies, which can be solubilized and renatured in vitro in low yield to an active unglycosylated enzyme with properties very similar to the native enzyme.4 Glucose oxidase is inactivated by H2O2, which accumulates during the oxidation reaction.5,6 Reduction in pH, which may accompany the breakdown of D-gluconolactone to gluconic acid, also inhibits the enzyme.7
Because native GOx is unable to transfer electrons to conventional electrode surfaces, the enzyme was inadequate for third-generation glucose sensors. Progress has focused on connecting the enzyme's redox center to the electrode. The main approach is by employing various molecular wires, such as osmium complex-linked polymers,8 artificial cofactor derivatives,9 thienoviologens,10 carbon nanotubes,11 or nanofibers.12 Alternatively, deglycosylating the enzyme increased the efficiency of direct electron transfer ability,13 presumably by decreasing the distance between the enzyme's redox center and the electrode, which are believed to be too far apart to transfer electrons in the native glycosylated enzyme.14
Pyrroloquinoline quinone was discovered in the late s as the third redox cofactor found in enzymes. Biosynthesis of PQQ has been confirmed only in prokaryotic microorganisms; however, many details of the biosynthetic pathway remain unknown. Unlike NAD or FAD, biogenesis of PQQ is ribosomal, produced through sequential posttranslational modifications of a peptide precursor. Dehydrogenases harboring PQQ are widely distributed, such as GDH, alcohol dehydrogenase, and sorbitol dehydrogenase, all of which have a beta-propeller structure, completely different from the structure of GMC oxidoreductases. The blades of the beta propeller consist of four-stranded antiparallel beta-sheet motifs arranged toroidally around a central axis.
Quinoprotein GDH (EC 1.1.5.2) does not represent one enzyme but instead refers to at least two distinct groups of GDHs harboring PQQ as the redox cofactor: membrane-bound PQQ GDH-A and water-soluble PQQ GDH-B. PQQ GDH-As have been reported from a variety of Gram-negative bacteria, and their substrate specificity, stability, and other features are strain-specific. In bacteria, the physiological role of PQQ GDH-A, which is coupled with the respiratory chain via ubiquinone, is the terminal oxidation of glucose.
PQQ GDH-B has only been reported from the Gram-negative bacterium Acinetobacter calcoaceticus, where it is located in the periplasmic space. A soluble PQQ-harboring aldohexose dehydrogenase, with a similar structure to PQQ GDH-B, was identified in E. coli by extensive genome database analyses; however, its extremely low enzymatic activity suggests that its true physiological function is not that of an aldohexose dehydrogenase.15 PQQ GDH-B is not coupled to the respiration chain of A. calcoaceticus, and its physiological role remains unclear. PQQ GDH-B has a broader substrate specificity profile than PQQ GDH-A and catalyzes oxidation of glucose, allose, 3-O-methyl-glucose as well as disaccharides lactose, cellobiose, and maltose.16 Diagnostic reagent-grade PQQ GDH-B is commercially available.
PQQ GDH-B is a homodimeric enzyme, with one PQQ molecule and three calcium ions bound to each subunit. Unlike the eight-bladed beta-propeller PQQ GDH-A, PQQ GDH-B has a six-bladed beta-propeller structure. PQQ resides in a deep, broad, positively charged cleft at the top of the propeller near the six-fold pseudo-symmetry axis.17 The six blades are labeled 1 to 6, and the four strands of each blade are labeled A to D. The cleft is constructed by the five loop regions 1D2A, 2D3A, 3BC, 4D5A, and 6BC (Figure 3A). The polar active site residues form hydrogen bond interactions with the glucose O1 and O2 hydroxyl groups, while hydrophobic interactions form between glucose and the nonpolar active site residues (Figure 3B). The amino acid residues reported to bind the substrate are located mainly on loops 1D2A, 2D3A, and 3BC, while residues of loop 6BC have no direct interaction with glucose. Although amino acid residues of loop 4D5A are within Van der Waals distance to the glucose O4 hydroxyl and C5 hydroxymethyl groups, the O4 hydroxyl group does not make hydrogen-bonding interaction with any amino acid.
The three-dimensional structure of PQQGDH from A. calcoaceticus. (A) The overall crystal structure (PDB 1CQ1) with the loop regions interacting with the substrate colored red, the PQQ cofactor colored yellow, and the glucose substrate colored green. (B) The active-site region shows the PQQ cofactor (green) and the amino acid residues that interact with the glucose substrate (red).
Among several PQQ-harboring enzymes, PQQ GDHs are the most industrially attractive. Although some PQQ GDH-As have high catalytic efficiency and narrow substrate specificity for glucose, they have never been employed for self-monitoring of blood glucose (SMBG) due to their membrane-binding character, which was unsuitable for recombinant production at the time when industries adopted GDH as a SMBG enzyme. In contrast, recombinant PQQ GDH-B can be conveniently produced in E. coli, where it is secreted into the periplasmic space. Due to its several attractive properties, soluble PQQ GDH-B was readily adopted by industries and became a major enzyme used in sensor systems for SMBG. PQQ GDH-B has high catalytic efficiency (over 510,000 units/mg protein), over 25 times higher activity than GOx, thus enabling rapid glucose sensing. Furthermore, because of the enzyme's inherent dehydrogenase character, its activity is not affected by variations of oxygen content in blood samples, thus providing greater accuracy. Because PQQ is tightly bound to GDH, additional cofactor does not need to be supplied to the reaction, as is the case with NAD(P) GDHs.
Native PQQ GDH-B does have some limitations for glucose sensing applications, especially when compared to GOx. The most significant limitation is in its broad substrate specificity, which can lead to potentially fatal errors in glucose sensing.18 Protein engineering studies have succeeded in improving its substrate specificity1921 and stability.2224 There have also been improvements in the recombinant preparation of PQQ GDH-B, such as the secretional production in yeast,25 and engineering the surface charge of the enzyme to facilitate its purification.26
Glucose-1-dehydrogenase [NAD(P) GDH, EC 1.1.1.47] represents a large number of enzymes from very diverse origins, from prokaryotes to higher eukaryotes and from mesophiles to extremophiles, such as halophilic and hyperthermophilic archaea. The enzyme catalyzes oxidation of the first hydroxyl group of glucose, utilizing NAD+ or NADP+ as the primary electron acceptor, and has the systematic name β-D-glucose:NAD(P)+1-oxidoreductase. Unlike GOx or other GDHs, the redox cofactor of this enzyme [NAD(P)] is not bound to the enzyme. Nicotine adenine dinucleotide (phosphate) GDHs are essentially water-soluble cytosolic proteins with homo-oligomeric structures (Figure 4).
The three-dimensional structure of the Sulfolobus solfataricus NAD(P) GDH complexed with NADP and glucose. The crystal structure of NAD(P) GDH (PDB 2CDB) is shown with the NADP cofactor colored orange, glucose colored yellow, and all four protein subunits colored differently to distinguish them.
Investigations of several elucidated three-dimensional structures revealed that NAD(P) GDHs share significant structural features, including their active site, which has made them a representative enzyme group for understanding protein structural stability. Most NAD(P) GDHs are glucose specific; however, some of the enzymes in this group lack recognition of the second hydroxyl group, thus making them reactive toward mannose, 2-amino-2-deoxy-D-glucose, and glucosamine.27 Some NAD(P) GDHs are also active toward galactose, xylose, and fructose, as well as disaccharides (e.g., maltose) and oligosaccharides.28
The NAD(P) GDH from Gram-positive bacterium Bacillus megaterium has been utilized for blood glucose monitoring. The enzyme has a homotetrameric structure (4 × 28.2 kDa), with one catalytic site per subunit. NAD(P) GDHs are commercially available as both reagent grade and diagnostic reagent grade from various enzyme companies. Under alkaline conditions, the enzyme reversibly dissociates into inactive monomers.29 Because the NAD(P) cofactor is not protein bound, it must be supplied exogenously together with specific artificial electron mediators for the electrochemical measurement. Since direct electrochemical oxidation of the reduced NAD(P) H cofactor is known to cause oligomeric inactivation, an applied potential that does not lead to oxidation of NAD(P)H is employed.
The FAD-dependent GDHs (FAD GDH, EC 1.1.99.10) have received much attention for their potential sensor applications. This group comprises oxidoreductases that catalyze the first hydroxyl group of glucose and other sugar molecules, utilizing FAD as the primary electron acceptor. FAD GDHs utilize a variety of external electron acceptors (but not oxygen), and their systematic name is D-glucose:acceptor 1-oxidoreductase. Although GDHs harboring FAD as the cofactor had been first reported in the s, this group of enzyme remains in the EC 1.1.99 group corresponding to redox enzymes utilizing other electron acceptors.
The reported FAD GDHs can be divided into three distinct groups according to their origin. FAD GDHs have been found expressed in the periplasm of Gram-negative bacteria, as extracellular enzymes in fungi, and as cytosolic enzymes in some insects. Both eukaryotic enzyme groups comprise glycosylated enzymes with quaternary structures ranging from monomeric to homo-oligomeric.
Bacterial FAD GDH is a thermostable hetero-oligomeric enzyme complex made up of a catalytic subunit harboring FAD in its redox center, a multiheme cytochrome-complex electron-transfer subunit, and a chaperone-like subunit required for proper folding and secretion of the catalytic subunit. Although the catalytic subunit alone shows high catalytic activity, the electron-transfer subunit facilitates the transfer of electrons between the active-site cofactor and the artificial electron mediator, giving this enzyme a catalytic activity similar to or even higher than that of PQQ GDH. Furthermore, the electron-transfer subunit confers the ability to transfer electrons directly to an electrode. Bacterial FAD GDH's unique direct electron transfer ability eliminates the need for toxic artificial electron mediators and makes this enzyme ideal for third-generation glucose sensors. Although a sensor employing bacterial FAD GDH is on the market, the enzyme is not commercially available.
Native bacterial FAD GDH has broad substrate specificity, catalyzing oxidation of glucose and disaccharides, such as maltose, to the corresponding lactones. Protein engineering studies of the catalytic subunit's active site have succeeded in specifically eliminating the enzyme's activity toward disaccharides, thus making it essentially specific for glucose.30
An emergent demand for a GDH that is specific for glucose has directed significant attention toward fungi-derived FAD GDHs, focusing on their narrow substrate specificity, especially their lack of activity toward maltose.31 The earliest report of a fungal GDH dates back to , describing the presence in Aspergillus subspecies of an enzyme capable of oxidizing glucose utilizing certain electron acceptors, but not oxygen.32 In , GDH from Aspergillus oryzae was characterized, including the identification of its FAD cofactor.33 FAD GDH was later studied from A. niger34 and Aspergillus terreus.31 Despite its long history, scientific information describing the fungi-derived FAD GDHs and their availability remains limited.
A screening of a fungal genome database using GOx active-site sequence motifs yielded putative proteins that could be divided into two structurally distinct phylogenetic clades, a GOx clade and a FAD GDH clade.35 Some members of the FAD GDH clade were successfully expressed in E. coli and, despite not having their native glycosylation, showed GDH activity and a high specificity for glucose. Furthermore, the predicted amino acid sequence of one of the putative FAD GDHs was identical to that of a patented FAD GDH. Diagnostic reagent-grade fungal FAD GDHs are commercially available.
The reasons behind the high substrate specificity for glucose exhibited by GOx can be understood by looking at the detailed structure of the enzyme complexed with glucose (Figure 5). The cavity forming the substrate-binding site, located in the center of the enzyme, is composed of residues that interact with each of the five hydroxyl groups that characterize glucose, thus preventing any epimer of glucose from binding at the active site. A comparison of GOx active site with a structural model of the active site of fungi-derived FAD GDHs revealed that their similar specificity for glucose was matched by their almost identical configurations around the substrate binding site. In particular, a conserved tyrosine (Tyr68 in A. niger GOx) interacting with the fourth hydroxyl group of glucose in both GOx and FAD GDH appears to be a hallmark residue behind the glucose specificity of these two fungal enzyme groups.
Structure of the GOx and fungal FAD GDH active sites complexed with glucose. The active site structures of the GOxs from A. niger (PDB 1CF3) and Penicillium amagasakiense (PDB 1GPE) were derived from their crystal structures. The three-dimensional structure models of the FAD GDHs from A. niger (ANG544) and Aspergillus flavus (AFL599) were constructed by Modeller 9v3 using the structure of A. niger GOx as template.35 The FAD cofactor is colored blue, and D-glucose is colored green. The conserved tyrosine residue interacting with the fourth hydroxyl group of glucose is colored pink.
The evolution from first-generation through third-generation glucose sensors has witnessed the appearance of a number of very diverse oxidoreductases, which vary tremendously in terms of origin, structure, substrate specificity, primary electron acceptor, and acceptable final electron acceptor (Tables 1 and 2). These enzymes belong to four main categories GOx, PQQ GDH, NAD GDH, and FAD GDH, each with their own special characteristics. However, the properties of any particular enzyme also vary depending on protein engineering or post-translational modifications. Some enzymes have been successfully engineered to eliminate certain limitations found in native enzymes. How the different oxidoreductases are produced may have an important impact; for example, the glycosylation level of the enzyme will depend on which host is used for the recombinant production, as well as any possible treatment of the purified enzyme (e.g., enzymatic deglycosylation in vitro). These are important points to consider when evaluating an enzyme for glucose monitoring applications.
Properties of Glucose Oxidases and Dehydrogenases
Enzyme Systematic name EC number Cofactor Microorganism Structure Glucose oxidase β-D-glucose: oxygen 1-oxidoreductase 1.1.3.4 FAD fungia homodimer, GMC oxidoreductases family Glucose dehydrogenase PQQGDH-A D-glucose:ubiquinone oxidoreductase EC 1.1.5.2 PQQ Gram negative bacteriab membrane-bound, monomer, 8-bladed β-propeller PQQGDH-B D-glucose:ubiquinone oxidoreductase EC 1.1.5.2 PQQ Acinetobacter calcoaceticus soluble, homodimer, 6-bladed β-propeller 1.1.1.118 (NAD+) NAD(P)-GDH D-glucose:NAD(P)+ 1-oxidoreductase NAD(P) (unbound) Gram positivec and negatived bacteria, eukaryotes homooligomer 1.1.1.119 (NADP+) Gram-negative bacteriae heterotrimer (catalytic, chaperone, cytochrome subunits) FADGDH D-glucose:acceptor 1-oxidoreductase 1.1.99.10 FAD fungif monomer-homooligomer insectsg homodimer Open in a new tabSummary of Major Features of Glucose Sensing Enzymes
Enzyme Advantages Disadvantages Sensor applications Possible error sources GOx Glucose specific Reacts with oxygen; low activity Electrochemical; colorimetric Variations of oxygen concentration PQQ GDH-A High activity; glucose specific Membrane bound Not applied in sensor PQQ GDH-B High activity Broad substrate specificitya Electrochemical; colorimetric Maltose in samplea NAD(P) GDH Glucose specific Cofactor must be supplied Electrochemical Bacterial FAD GDH High activity; direct electron transfer (to electrode) Broad substrate specificitya Electrochemical; colorimetric Maltose in samplea Fungal FAD GDH Glucose specific Low activity; activity toward xylose Electrochemical; colorimetric Xylose in sample Open in a new tabOf particular importance for glucose sensors is the enzyme's substrate specificity. Patients undergoing peritoneal dialysis or other treatments may have significant quantities of maltose in their blood.36 Glucose sensors that employ oxidoreductases with broad specificity may give falsely elevated glucose readings, leading to iatrogenic hypo-glycemia. A Food and Drug Administration public health notification18 warned against potential fatal errors from PQQ-GDH-based sensing technology. Glucose oxidase has long been envied for its high specificity for glucose. However, protein engineering studies on PQQ GDH and bacterial FAD GDH, as well as the increasing use of fungal FAD GDHs, are challenging GOx's historical supremacy.
It is important to understand the basic properties and limitations of the enzymes employed and the principles behind their catalytic activity to avoid potential dangers. When considering the possible suitability of an enzyme for a particular sensor application, one must not only consider the enzyme's main category or conventional name, but should also be aware of other features that may affect the enzymatic properties, such as host strain used for its preparation, engineered mutations, as well as chemical or enzymatic modifications. This short review of oxidoreductases employed in blood glucose sensors has attempted to provide the necessary background to better understand these features and hopefully avoid potential pitfalls.
flavin adenine dinucleotide
glucose dehydrogenase
glucose/methanol/choline
glucose oxidase
nicotine adenine dinucleotide
nicotine adenine dinucleotide phosphate
pyrroloquinoline quinone
self-monitoring of blood glucose