glucose oxidase

11040-8347/95/$.50 1995 by CRC Press, Inc.

Critical Reviews in Analytical Chemistry,25(1):142 (1995)

Lees forecast, quoted above, has beenproved true. The real impact of enzymes asanalytical reagents, however, was not feltuntil the mid-1960s and early 1970s, de-spite the fact that in 1951 Stetter1 describedapplications of impure enzymes to a varietyof analytical problems. Even as early as1845, Osann2 determined hydrogen perox-ide using peroxidase. The use of enzymepreparations as analytical reagents, bothsoluble or immobilized on inert carriers,has grown exponentially since the 1970s.3

The enzyme glucose oxidase (EC 1.1.3.4)is, without doubt, the one employed mostwidely. Its use is so frequently mentioned

in the analytical literature that glucose oxi-dase is sometimes suspected to be abusedrather than used. Such popularity is, how-ever, the result of a combination of factors.Glucose oxidase is a useful reagent for theselective determination of glucose, ananalyte of clinical as well as of industrialinterest. Moreover, it is used as an analyti-cal reagent as a marker of antigens andantibodies in enzyme immunoassays. It isused in the food industry to remove smallamounts of oxygen from food products orglucose from diabetic drinks.4 In addition,glucose oxidase has been proposed as ananticancer drug5a,5b because it may damage

Glucose Oxidase as an Analytical Reagent

Julio Raba and Horacio A. Mottola*Department of Chemistry, Oklahoma State University, Stillwater, OK 740780447

There is no doubt that methods based on catalysis by enzymes will increase in number andpracticability as advances are made in the study of these substances and as more enzymes ofknown activity and purity become commercially available

T.S. Lee, in Organic Analysis, Vol. 2; J. Mitchell, Jr., I.M. Kolthoff, E.S. Proskauer, A. Weissberger,Eds.; Interscience: New York, 1954, p. 242.

* Author to whom correspondence should be addressed.

Julio Rabas permanent address is: Departamento de Qumica Analtica, Facultad de Qumica, Bioqumicay Farmacia, Universidad Nacional de San Luis, 5700-San Luis, Argentina.

ABSTRACT: Glucose oxidase (EC 1.1.3.4) is the most widely employed enzyme as analyticalreagent. This is the result of (1) its utility in the determination of glucose, an analyte of wideanalytical interest, and (2) its relatively low cost and good stability that make the glucose/glucoseoxidase system a very convenient model for method development (particularly in the area ofbiosensors). This review discusses in detail enzyme structure, biocatalysis, enzymes as analyticalreagents, properties of glucose oxidase (including a historical account), and the use of glucoseoxidase as an analytical reagent in homogeneous systems as well as an immobilized reagent.

KEY WORDS: enzymes as analytical reagents, glucose oxidase

2cancerous tissue and cells as a result ofhydrogen peroxide formation.

Glucose oxidase is commercially avail-able with high purity and at relatively lowcost. Also, its immobilization is rather eas-ily achieved and results in preparations ofadequate stability. Even in solution, its sta-bility is competitive in view of its cost.6 Inany event, the vast number of applicationshas not found an integrating forum. Conse-quently, this review is an attempt to criti-cally describe the analytical uses of solubleand immobilized glucose oxidase. No at-tempt is made to cite all published workconcerning its use as an analytical reagent,but only those reports in the authors per-ception that have resulted in advances ofthe topic area. In many instances publica-tions offer minor contributions to the ad-vancement of the use of glucose oxidase(e.g., use of a different chromophore in anindicator reaction without substantial im-pact on the value of the method, minorvariations in the chemistry of immobiliza-tion, or slightly different strategic ap-proaches to the utilization of redox media-tors in electrochemical sensing).

I. ENZYME STRUCTURE,BIOCATALYSIS, AND ENZYMES ASANALYTICAL REAGENTS

Catalysis involves catalytic species thatrange from the simple proton to very com-plex molecular species in their architecturalatomic arrangement such as enzymes. Be-cause enzymes are biocatalysts, their use inanalytical chemistry as reagents fits withinthe scope of kinetic-based determinations.Consideration to the large number of enzy-matic determinations performed at clinicallaboratories around the world has justifiedthe statement that the number (of determi-nations) carried out by kinetic-based meth-ods probably exceeds that carried out bythermodynamic (equilibrium) methods anddirect instrumental measurement combined.7

A catalyst is a chemical species that isboth a reactant and product of a reaction.8 Itsconcentration enters into the kinetic equa-tion but not into the equilibrium constant ofthe reaction. The presence of a catalytic cycle,as illustrated below for a generalized sys-tem, singles out catalysis from other rate-modifying effects (e.g., promotion, in whichthe rate modification is transitory and themodifier does not appear as a product of thereaction):

[AE]

[AE] P + E

A + E

in which E is the enzyme and A a reactant,commonly known as substrate. [AE] is anintermediate species involving associationbetween the enzyme and the substrate. Theoverwhelming majority of analytical meth-ods using enzymes as analytical reagents aredesigned for the determination of speciesacting as A in the example above, that is, thesubstrate is the analyte under determination.There are cases, however, in which the ana-lytical interest is the determination of thecatalytic activity of the enzyme. In such casesthe enzyme plays the role of analyte. In anyevent, a distinctive analytical advantage ofusing enzymes as analytical reagents is thehigh selectivity (even eventual specificity)of the interaction that leads to [AE] forma-tion. Jenks,9 in a very interesting review,argues that specificity (and by extension highselectivity) and rate modification (the twospecial features of enzymatic catalysis) bothderive from the utilization of the free energymade available from binding interactions withhighly selective or specific substrates. Jenksqualifies further his argumentation by point-ing out that ...The principal difference be-tween enzymatic and ordinary chemical ca-talysis is that enzymes can utilize noncovalent

binding interactions with substrates to causecatalysis, in addition to the chemical mecha-nisms utilized by ordinary catalysts.

Enzymes are proteins, and the majorconstituent of proteins is an unbranchedpolypeptide chain consisting of L-a-aminoacids linked together by amide bonds be-tween the a-carboxyl group of one residueand the a-amino group of the next. The pri-mary structure is the sequence in which theamino acids form the polymer. Although theprimary structure of almost all intracellularproteins are linear polypeptide chains, manyextracellular proteins contain covalent -S-S-crossbridges that result from two cysteineresidues linked by their thiol groups.10 Thesecondary structure of proteins is recognizedas polypeptide chains organized into hydro-gen-bonded structures. The three-dimen-sional structure of enzymes, starting fromthe polypeptide chains, unfolds in the ter-tiary structure of covalently linked polypep-tide chains. There are proteins composed ofsubunits that are not covalently linked. Theoverall organization of these subunits is whatis known as the quaternary structure. Achange in quaternary structure indicates thatthe subunits move relative to each other.These three-dimensional proteinic structuresare present in enzymes that can be relativelysmall molecules with molecular masses ofthe order of 10,000 Da, while others are verylarge molecules, with molecular masses from150,000 to over 1 million Da. No matterwhat the size, chemical composition, or struc-ture of the enzyme the catalytic action isconfined to a specific region of the enzymeknown as the active site, because the AEcomplex does not form in a topologicallyrandom manner. The substrate binds to thisregion on the enzyme in every catalytic cycle,and catalysis takes place only at active sites.The structure of enzymes and the uniquefunction of the active site have resulted in anillustration of enzyme catalysis by what isknown as the lock-key model.11a,11b Thissimplistic model is no longer satisfactory toevaluate molecular interactions because the

physical rigidity that it conveys does notagree with a conformationally mobile en-zyme-substrate interaction. It, however, doesmetaphorically illustrate the unique selectiv-ity (specificity) of enzymes. Because thelock-key model does not explain all cases ofenzyme catalysis, these anomalies have re-sulted in the postulation of revisions to themodel. Koshland12 suggested a modifiedmodel called for brevity the induced fitmodel (or theory). In essence, the induced fitmodel rests on: (1) a precise orientation ofcatalytic groups needed for enzymatic ac-tion, (2) changes in the three-dimensionalrelationship of the amino acids at the activesite caused by the substrate, and (3) changesin the protein structure caused by the sub-strate to bring the catalytic groups into theproper orientation for reaction; a nonsubstratewill not do this. Figure 1 illustrates both thelock-key and the induced fit models.

The active site (or active center) is rela-tively small and is formed by three or fouramino acids in the twisted polypeptide chain,separated considerably from each other inthe amino acid sequence, but very near toeach other spatially.13 Certain side chains ofthese amino acids hold or anchor the sub-strate at the active site, whereas others modifythe bonding forces at the reacting group ofthe substrate. Presumably, the remainder ofthe enzyme proteinic structure is needed tomaintain the unique conformation to formthe active center. Because the catalytic ac-tivity of the enzyme depends on the presenceof a given conformational structure in thefolded polypeptide chain, even minor alter-ations in the tertiary structure result in lossof activity.13 Because the catalytic activityof an enzyme is proportional to the numberof operating active sites at the time of themeasurement, the catalytic activity for dif-ferent preparations of the same enzyme maynot be the same. Consequently, the analyti-cal concentration of a given enzyme prepa-ration cannot accurately describe the enzymeactivity. This lack of correlation betweenconcentration and activity plus the inherent

4FIGURE 1. Models describing the relationship between the active site of an enzyme and thesubstrate on which that enzyme performs catalysis.

potential instability of enzyme preparationsare properties that analytical chemists had toadapt to in order to use them as analyticalreagents. This was not an immediate adapta-tion; we need to realize that gravimetry,

titrimetry, and other equilibrium-based de-terminations all rely on stable reagents andchemical species.

Many enzymes need the presence of otherspecies, known as cofactors, to exert cataly-

5sis. The enzyme-cofactor complex is calleda holoenzyme; the proteinaceous portion isknown as an apoenzyme (holoenzyme =apoenzyme + cofactor). Apoenzymes mayform complexes with different types of co-factors such as: (1) a coenzyme (a non-proteinaceous organic species loosely at-tached to the apoenzyme), (2) a prostheticgroup (an organic species firmly attached tothe apoenzyme), and (3) a metallic ion.

II. GLUCOSE OXIDASE

The discovery of glucose oxidase is at-tributed to Mller,14 who found this enzymein Aspergillus niger and Penicillium glaucum.Mller established in 1928 that the enzymecatalyzes the oxidation of glucose to glu-conic acid in the presence of dissolved oxy-gen. In the same year, Bernhauer15 concludedthat the conversion of glucose into gluconicacid by A. niger was due to an enzyme thathe designated glucoxidase. Nord andEngel16 found a similar enzyme in Fusariumlini. Franke et al.17a,17b purified the enzymefrom A. niger to a moderate extent, and foundthat the activity of different preparations wasqualitatively correlated with the flavine con-tent of the preparation. This was the firstindication that the enzyme is a flavoprotein.Two years later, Mller18 studied the oxida-tion of various sugars in the presence andabsence of oxygen, and ventured that twodifferent enzymes were present. Somewhatearlier, Ogura19 claimed to have isolated fromA. oryzae a glucose-oxidizing enzyme thatdid not utilize O2 as the hydrogen acceptor.These dualities were put to rest by Franke,20

who in 1944 concluded that the glucose oxi-dases from P. glaucum, F. lini, A. niger, andA. oryzae are virtually identical, consistingof a single enzyme utilizing O2 as the hydro-gen acceptor. In 1948, Keilin and Hartree21

cleared another confusion by demonstratingthat glucose oxidase, notatin (penicillin A),and penicillin B are the same enzyme, whichcatalyzes the aerobic oxidation of glucose to

gluconic acid. These authors can also becredited as probably the first to make ananalytical application of glucose oxidase inthe manometric measurement of oxygen todetermine glucose in some biologicalmaterials.22a,22b

The purification, crystallization, andsome properties of crystalline glucose oxi-dase from P. amagasakiense were reportedby Kusai et al.23 The interesting demonstra-tion that the enzyme is a glycoprotein, be-sides being a flavoprotein, is due to Pazur etal.24

III. PROPERTIES OF GLUCOSEOXIDASE [EC 1.1.3.4]

Glucose oxidase from A. niger is a dimerof 186 103 mol wt with two molecules offlavin adenine dinucleotide (FAD) tightlybound per dimer. The FAD in glucose oxi-dase stabilizes the three-dimensional struc-ture of the enzyme and cannot be removedby dialysis at neutral pH. Evidently, the pro-cess by which FAD becomes bonded to theapoenzyme is a step in the synthesis of theholoenzyme.

Glucose oxidase is composed of twopolypeptide chains of approximately equalmolecular weight held together by disulfidebonds and with carbohydrate content amount-ing to 16%.25 Recently, Chi et al.26 providedpictures illustrating the first direct observa-tion of native and unfolded glucose oxidasestructures obtained by scanning tunnellingmicroscopy. The images show an openingbutterfly-shaped pattern in accord with thegeneral topology assigned to the structure ofthe enzyme. As a first step in attempting toelucidate the residues necessary for catalysisand improve the properties of glucose oxi-dase by protein engineering, Frederick etal.27 have described the cloning and expres-sion in yeast of A. niger glucose oxidase.

Some differences are observed in glu-cose oxidase obtained from different sources;glucose oxidase from P. amagasakiense, for

6instance, has a mol wt of 154 103 and con-tains 2 mol of FAD per mole of proteinicenzyme.23 Table 1, adapted from Swobodaand Massey,25 illustrates the properties ofglucose oxidase obtained from three differ-ent sources.

The redox mechanism of a flavoproteinoxidase such as glucose oxidase involves the

catalysis of a redox reaction. During the re-action either one or two electrons from theelectron donor are transferred to the isoal-loxazine nucleus of the flavin coenzyme(FAD, flavin mononucleotide [FMN], orderivative) and then to the electron acceptor.

The prosthetic group of a flavoproteinusually contributes only 0.2 to 2% of the

TABLE 1Some Properties of Three Glucose Oxidase Preparations Obtainedfrom Fungus (Adapted from Reference 25)

Source

Penicillium Aspergillus PenicilliumProperty nonatum niger amagasakiense

Wavelength ofmaximun 270280,absorbance, (nm) 278, 383, 452 278, 380, 460 377, 455

Inhibition by Virtually More than 90%mercury none by 105 M HgCl2

or p-chloromercuri- benzoate

Michaelis-Menten 0.033 M 0.015 M 0.096 Mconstant (in air, at 25C at 30C at 20Ccatalase present,pH 5.6)

Molarabsorptivity ofenzyme flavin, 1.41 104 1.10 104 1.09 104

450 nm, M1 cm1

Formation offlavin semiquinoneby dithionite Yes No

Sedimentationcoefficient (s)a 8.00 7.93 8.27

Diffusioncoefficientcm2 s1a 4.12 107 5.02 107 5.13 107

Molecular weight 186 103 154 103 152 103

Standardizedspecific activity 80 112c 64at 25Cb 77c

a In aqueous solution and at 20C.b In units of micromoles per milligram (dry weight), infinite glucose concentration,

0.25 mM oxygen, excess catalase, and phosphate or acetate buffer of pH 5.60.c The values of 112 and 77 are from References 23 and 24, respectively.

Data from References 22a, 22b, 24, 29.

7total molecular weight. Because of the vari-ety of chemically active residues in thesegroups, multiple stabilizing bonds are formedbetween the prosthetic group and the rest ofthe proteinic framework.28 The flavin nucleusexists in three oxidation states, each of whichcan adopt different states of ionization. Fig-ure 2 illustrates the redox and acid-base be-havior of these flavin species. As Figure 2shows, the three states of ionization involvethe proton on N-3 for HFo and the proton onN-1 in both H2F and H3Fr.29 The R group atN-10 is either FAD or FMN.

Glucose oxidase should be defined as anoxidoreductase that catalyzes the reactionby which all electrons taken from the sub-strate (glucose) are transferred to O2 to formH2O2. Catalysis by this flavoprotein dependson the reduction-oxidation of its flavin group.

FIGURE 2. Flavin species as they occur at different pH and oxidation states.

The mechanism of the overall process ispostulated to include an oscillation betweenthe oxidized and the reduced forms of theflavin:30

E-FADH2-P

E-FADH2 + O2 E-FAD-H2O2

E-FAD + G E-FADH2 + P

E-FAD + H2O2

k1 k2

k3 k4 (1)

Glucose, G, reduces the FAD to FADH2 with-out the formation of the free radicalsemiquinone as intermediate and producesgluconic acid, P, as product. Subsequently,oxygen, the acceptor, oxidizes the FADH2back to FAD and H2O2 is released as product.

Keilin and Hartree,22b in their pioneeringstudies of glucose oxidase, report that theenzyme does not fluoresce under ultravioletlight within the pH limits in which it hascatalytic activity (pH 2.0 and 8.0). Outsidethis range, however, glucose oxidase loses

8activity and becomes fluorescent. Accord-ing to Gibson et al.,31 glucose oxidase showscatalytic activity over the pH range from 3.0to 10.0 (maximum seems to center at ca. pH5.5) as the result of the rather unusual stabil-ity of this enzyme.

Using a flow apparatus and spectropho-tometric detection, Nakamura and Ogura32

studied the kinetics of catalysis by the glu-cose oxidase from P. amagasakiense, andidentified as rate limiting the step corre-sponding to the conversion of the glucose-enzyme complex to reduced enzyme andproducts. Gibson et al.31 convincingly re-ported on the mechanism of catalysis byglucose oxidase from A. niger. They em-ployed stopped-flow mixing and manomet-ric measurements at a pH of 5.6 and in the0 to 38C temperature range. They foundno evidence of a kinetically significant fla-vin-semiquinone intermediate, and all theirkinetic data can be analyzed in terms of thedistribution of the enzyme in two forms:the fully oxidized and the fully reduced(Figure 2). The results from all the sugarsubstrates tested fit the general scheme ofEquation 1 comprising a reductive half-re-action and an oxidative one. Several otherkinetic studies33a33f tend to confirm the find-ings of these earlier reports.

The enzyme is highly selective towardb -D-glucose21 (Table 2), and any chemicalalteration departing from the basic molecu-lar structure of this species results in sub-strates with considerably reduced reactivityin the presence of glucose oxidase as thecatalyst. The b -anomer reacts about 150 timesas fast (at 20C) as the a -anomer of D-glu-cose. Substitution on C-2 (except of -H for-OH) or on C-3 erases all catalytic activity,but substitution on C-4 and C-6 consider-ably decreases but does not totally eliminatethe catalytic activity of the enzyme towardthe corresponding substrates. This high se-lectivity, obviously, singles out glucose oxi-dase as a unique analytical reagent for thedetermination of glucose in general and b -D-glucose in particular.

IV. ANALYTICAL USES OFGLUCOSE OXIDASE

Considering its price per unit activity,glucose oxidase is one of the less expensiveenzymes for analytical use as a reagent, par-ticularly when its inherent competitive sta-bility is also taken into consideration. His-torically, its use in solution preceded itsanalytical use in immobilized form. The mainargument for immobilization is the recoveryof the enzyme for repetitive use as a reagent.In actuality, however, if an enzyme is rela-tively inexpensive and relatively stable (sothat it retains its activity toward the substrateof interest at reasonably constant level atroom temperature with time), it can be useddirectly and competitively in solution forrepetitive determinations.6 Competitive useis afforded by implementation of unseg-mented closed-loop continuous-flow sys-tems.34 The virtue of using glucose oxidasein solution has also been argued and demon-strated in connection with open-loop auto-mated unsegmented continuous-flow con-

TABLE 2Relative Oxidation Rates, in thePresence of Glucose Oxidase asCatalyst, for Several Sugarsa

Relative rate ofSugar oxidation

b -D-glucose 1002-Deoxy-D-glucose 256-Deoxy-6-fluoro-D-glucose 36-Methyl-D-glucose 1.854,6-Dimethyl-D-glucose 1.22D-mannose 0.98D-xylose 0.98a -D-glucose 0.64Tetralose 0.28Maltose 0.19Galactose 0.14Melobiose 0.11

a Rates are normalized to the rate of b -D-glucose oxidation and for the general reac-tion: sugar + O2 corresponding acid +H2O2.

9figurations.35 Why then, is glucose oxidaseused so extensively in immobilized form?The fashionable use of packed-column reac-tors in unsegmented-continuous-flow sys-tems (e.g., in flow injection analyses) andthe ease of immobilization of glucose oxi-dase undoubtedly play some role. Of greaterimpact is perhaps the convenience of immo-bilized glucose oxidase and the oxidation ofglucose to gluconic acid as a model systemfor the development and characterization ofnew analytical approaches (e.g. new reactorconfigurations or strategies in so-calledbiosensing). As such, in the remainder ofthis review, we will address the followingthemes: (1) the use of glucose oxidase as ananalytical reagent in solution, (2) the use ofimmobilized glucose oxidase for determina-tive purposes (including the use of glucoseoxidase in model systems for method orapproach development), and (3) the determi-nation of glucose oxidase activity.

V. USE OF GLUCOSE OXIDASE ASAN ANALYTICAL REAGENT INHOMOGENEOUS SYSTEMS

In 1956 Keston36 proposed the first prac-tical analytical application of glucose oxi-dase as a homogeneous as well as a hetero-geneous biocatalyst. In 1959 Marks37

suggested that older, nonselective, methodsbe replaced by enzymatic ones. The bulk ofthese enzymatic methods, for obvious rea-sons, are kinetic (rate-based) methods. Thechemistry of the enzyme-catalyzed reactionprimarily dictates the approach (includinginstrumentation) to be used in monitoringthe reaction rate. In the case of glucose de-termination, each chemical species involved(except of course the analyte itself and thebiocatalyst) has been used to provide themeans for determination. Table 3 gives asummarized overview of the typical detec-tion approaches used in the enzymatic deter-

TABLE 3Typical Detection Approaches Used in the EnzymaticDetermination of Glucose Using Glucose Oxidase

Chemical speciesresponsible for Coupled process Detection

detection to aid detection approach

H2O2 None AmperometricOxidation of anindicator species(a) organic dye Spectrophotometric(b) Fe(CN)63 Amperometric or

spectrophotometricOxidation of a Fluorometricchemical speciesleading to fluorescence

Oxidation of a chemical Chemiluminometricspecies leading tochemiluminescence

Oxidation of a chemical Potentiometricspecies resulting in achange in potential of thesystem under observation

O2 None AmperometricGluconic acid None pH-stat

(potentiometric)

10

mination of glucose using glucose oxidase.It should be noted that the information givenin this table, in general, applies equally wellto soluble or immobilized glucose oxidase.

Kestons proposal36 for using glucoseoxidase in solution to determine glucose inurine samples involved spectrophotometricmeasurement (480 nm) of the colored prod-uct of o-dianisidine (3,3 -dimethoxy-benzidine) oxidation in the presence of glu-cose oxidase and peroxidase. This couplingof the main enzyme-catalyzed reaction witha second (indicator reaction) has become avery common practice in enzymatic meth-ods. Because reactions involving oxidasesyield H2O2 as one of their products, the oxi-dizing power of H2O2 is exploited in suchcoupled schemes. The oxidation of the re-duced form of the dye is aided by the actionof peroxidase:

GLUCOSE OXIDASEglucose + H2O + O2 gluconic acid + H2O2H2O2 + dye (reduced form) dye (oxidized form) + H2O

PEROXIDASE

Keston36 also proposed the determination ofglucose in urine by impregnating paper stripswith glucose oxidase, peroxidase, and o-to-lidine (3,3 -dimethylbenzidine) as a chro-mogenic oxygen acceptor. Strips of filterpaper impregnated with the appropriate re-agents are dried in air and stored protectedfrom light. The reagent papers are dippedinto the test solution, removed, and com-pared (after a minute or two) with colorsproduced by standard glucose solutions. Inthese tests the glucose oxidase enzyme canbe considered immobilized (by physical ad-sorption) on the filter paper support. Kestonsproposal has been widely adapted for thedetermination of glucose in clinical, food,and agricultural analysis and the literaturerecords numerous applications involvingminor variations to the original conditions.Different dyes, for instance, have been pro-posed in place of o-dianisidine; some typi-cal examples include 2,6-dichloropheno-lindophenol,38 o-toluidine,39 guaiacum(guaiac or guaiacum resin obtained from

the wood of Guaiacum officinale and Guai-acum sanctum; Chemical Abstract ServiceRegistry Number 900-29-7; Merck IndexNo. [11th ed., 1989] 4455),40 2,2 -azino-di-[3-ethylbenzthiazoline-6-sulfonic acid](ABTS),41a,41b 4-aminophenazone or ad-renaline,42 in situ coupling of 3-methyl-2-benzothiazolinone hydrazone with N,N-dimethylaniline,41a oxidative coupling ofN,N-diethylaniline with 4-aminophenazoneor phenol,43 and leuco Patent Blue Violet.44

The quest for alternative dyes was ingreat part motivated by the fact that the useof o-dianisidine may expose those using it tolong term-health problems. ABTS, with anabsorption maximum at 420 nm, is a veryconvenient chromogen because it is safe,very soluble in water, practically insensitiveto light, and does not undergo autooxidation.Moreover, it is more sensitive for H2O2 de-tection than o-dianisidine. Regarding thechemical performance of the most commonlyused dyes in enzymatic determinations withglucose oxidase, however, the results arecomparable and the enzymatic-based meth-ods are competitive with previously wellestablished methods.4547 Readers interestedin optimal conditions for the use of glucoseoxidase in solution to directly determine glu-cose in plasma and urine, and without pre-liminary preparation of protein-free filtrates,are referred to the work of Kingsley andGetchell.48

The availability of improved electronicdevices led to the development of automaticmethodologies in the 1960s. Malmstadt andHicks,49 for instance, assembled an instru-ment from commercially available units andinterfaced it with a control system and en-zyme injector. The commercially availableunit was the Spectro section of the Spectro-Electro titrator (E.H. Sargent & Co., Skokie,IL). The practical end was to shorten to about1 min the clinical determination of glucosebased on glucose oxidase catalysis by spec-trophotometric monitoring of H2O2 releasedby oxidation of a dye. One year later,

11

Malmstadt and Pardue50a,50b introduced apotentiometric reaction rate method in whichthe H2O2 formed during the enzyme-cata-lyzed reaction reacts with an excess of io-dide, in the presence of molybdate as thecatalyst, and produces an equivalent amountof iodine. In this paper the applicability ofthe variable-time procedure51 to nonlinearresponse curves is demonstrated. A spectro-photometric variant of the same method(measurement of the absorbance of I3 at360 nm) was proposed by Malmstadt andHadjiioannou 1 year later.52 The same chem-istry was utilized by Pardue53 to introduceamperometric monitoring for continuousmeasurement of reaction rates. The rate ofincrease in iodine concentration was mea-sured using a polarized rotating platinumelectrode (polarizing source: a 1.5-V batteryin series with a 100-W potentiometer; rota-tion velocity: 2000 rpm). The approach wasapplied to the determination of glucose inserum, plasma, and whole blood.

An ingenious piece of work was de-scribed by Blaedel and Hicks54 in what con-stitutes the first implementation of unseg-mented continuous-flow sample/reagent(s)processing for the measurement of the rateof an enzyme-catalyzed reaction in a con-tinuous fashion. This paper describes a flowsystem that permits determinations using thefixed-time approach. As an extension of thiscontinuous-flow approach, the same glucosedetermination was used by Blaedel andOlson,55 to introduce a setup that allows thecontinuous measurement of reaction rates.The measurement involved a differentialamperometric procedure in which the H2O2produced oxidizes hexacyanoferrate(II) tohexacyanoferrate(III), the concentration ofthe latter measured at a tubular platinumelectrode. However, when enzymes are usedas analytical reagents for the determinationof substrates, their catalytic nature indicatesthat repetitive use should be possible. Im-mobilization is an avenue that moves in thatdirection. Enzyme regeneration in homoge-

neous systems is, however, possible. Reagentrecirculation in closed flow-through sys-tems34,56 permits the reutilization of solubleenzyme preparations. An example of imple-menting recycling of the enzyme solution,sample injection into a closed flow-throughsystem, and amperometric detection of thechange in dissolved oxygen level as a resultof the oxidation of glucose to gluconic acidin the presence of glucose oxidase has beenoffered for glucose determination.34 The H2O2released in this reaction interferes with themonitoring of O2. This is circumvented, how-ever, by using glucose oxidase that containscatalase as an impurity, which is inexpen-sive and ensures practically instantaneousdestruction of the H2O2 formed. The oxygenproduced by this reaction is one half theoxygen consumed in the main catalyzed re-action, and monitoring based on oxygenconsumption is still possible.

Although the bulk of the proposed meth-ods for glucose using glucose oxidase uti-lizes spectrophotometric monitoring, the re-action can be followed electrochemically (bymonitoring amperometrically the H2O2formed or the O2 consumed34,5759) orpotentiometrically. Kadish and Hall57 com-pared results obtained with a polarographicoxygen sensor with standard AutoAnalyzerdata. The compared population data con-tained 1,056 pairs of points obtained over a5-month period. They measured the decreasein dissolved oxygen and used iodide, etha-nol, and ammonium molybdate to minimizeproblems claimed to be associated with theH2O2 produced.

A special consideration of continuous-flow systems and the so-called AutoAnalyzertechnology seems necessary at this point ofthe review. Continuous-flow procedures haveprovided advantageous alternatives for wetchemical methods. The practice of wetchemical analysis has undergone drasticchanges, particularly since the mid-1950swhen Skeggs60,61 proposed a novel mannerof sample-reagent mixing and transport to

12

detection. The elements of Skeggss modu-lar, continuous-flow concept resulted in theworkhorse of practically every clinical labo-ratory and many industrial analytical facili-ties, the so-called AutoAnalyzer developedand marketed by Technicon (Technicon In-struments Corp., Tarrytown, NY). Methodsfor glucose using the AutoAnalyzer weredeveloped in the late 1960s and early to mid1970s using hexacyanoferrate(III), hexoki-nase, neocuprione, o-toluidine, and glucoseoxidase.62

Monitoring of the reaction has involveddetection based not only on the O2 consumedor the H2O2 formed, but in the change in pHresulting from the conversion of glucose togluconic acid, if low ionic strength buffersare used. Malmstadt and Piepmeier,63 as partof the electronic impact of the mid 1960s,described a pH-stat with digital readout thatthey used to develop a method for glucosedetermination in the 50 to 250 ppm range. Inthis procedure the pH was held constant at6.5 by adding small (equal) increments of0.0020 M NaOH; the number of aliquots ofreagents delivered during a preset time wascounted. The count proved to be directlyproportional to the glucose concentration.The approach is very reproducible but lacksthe sensitivity of methods based, for instance,on the coupling of reactions based on H2O2production. Guilbault et al.,64 on the otherhand, proposed a potentiometric rate-basedmethod monitoring the main reaction by re-cording the change in the difference in po-tential between two platinum thimble elec-trodes polarized with a constant current of40 A. The method was developed for thedetermination of either glucose or glucoseoxidase itself. Diphenylaminesulfonic acidwith a reported transition potential of+0.8 V vs. the NHE was used for establish-ing the initial potential. The rate of potentialchange after addition of the sample providedthe bases for calibration plots. Another twistin coupling the main enzyme-catalyzed re-action with a second chemical-indicating

system can be seen in the proposal ofAlexander and Seegopaul65 for using a po-tentiometric SO2 probe to monitor the H2O2production by the progress of the followingreactions:

Because of glucose relevance in patientsdemographic information via normal/abnor-mal test results, it is not surprising that clini-cal applications of glucose oxidase as ananalytical reagent outweigh applications inother areas. Glucose, as already mentionedin this review, plays an important role infoods and the food industry, and applica-tions to the analysis of food products abound.White,66 for example, determined glucose inhoney by a photometric procedure in whichthe glucose oxidase used was contaminatedwith a -glucosidase (invertase) but the inter-ference was circumvented by inhibitionworking in a Tris [tris-(hydroxymethyl)-aminomethane] buffer. Results (accuracy andreproducibility) were comparable to those ofrecognized standard methods.

As indicated in Table 3, the H2O2 can beused to effect different changes on coupledchemical species. Some of these changes arephysicochemical transformations that resultin the formation of an excited chemical spe-cies that, after relaxing to its ground stateemits radiant energy (fluorescence or chemi-/bioluminescence). Guilbault et al.,67 forexample, proposed the oxidation of homo-vanillic acid (4-hydroxy-3-methoxyphenyl-acetic acid) by the H2O2 to produce a highlyfluorescent product. In a subsequent paperthey evaluated 25 indicator species for thesame fluorometric determination.68 Theyconcluded that p-hydroxyphenylacetic acidconstitutes the best target species because itis less expensive and provides a higher fluo-rescent coefficient (fluorescence intensity/

2HSO3-

H2O2 + 2HSO3-

S2O62- + 2 H2OS2O5

2- + 2 H+ 2SO2 + H2O

S2O52- + H2O

13

molar concentration). The procedures are alsoillustrated for the determination of enzymeactivity. Leucodiacetyldichlorofluorescein,originally synthesized by Brandt andKeston69a and proposed as a fluorophor forH2O2 determination by the same authors,69b

was used by Kelly and Christian.70a,70b TheH2O2 (in the presence of horseradish per-oxidase) produces fluorescent dichloro-fluorescein (excitation at 448 nm, emissionat 525 nm). Keston et al.,71 however, re-ported photons of 500 nm as the best excita-tion wavelength. Excitation was accom-plished with an argon ion laser and theinstrumental setup involved a capillary sheathcell and continuous-flow processing. Air-segmented continuous-flow processing forcollecting data and stopping the flow wasused by Hsieh and Crouch72 in a kineticmethod for determining glucose with theTrinder reaction73 in the presence of glucoseoxidase. Using this approach glucose wasdetermined in wines and serum samples. Anunsegmented continuous-flow/stopped-flow/continuous-flow system has also been re-ported.74 The sample and enzyme solutionare simultaneously injected and a gas-per-meable silicone-rubber reaction coil (to en-hance mixing) as well as pulsed ampero-metric detection of H2O2 were employed. Incontinuous-flow systems, mixing can alsobe enhanced by the use of what has beentermed a single bead string reactor.75

Chemi- and bioluminescence providekinetic-based methods involving measure-ments under dynamic conditions of transientlight emission (Chapter 8 in reference 51).Their amenability to the determination ofspecies of biomolecular interest and a rela-tively simple implementation have resultedin a considerable increase in popularity inthe past 10 years or so,76 popularity thatcontinues as sustained application.77 Enzy-matic chemiluminescence determination ofglucose, of course, did not escape this popu-larity. A review is available in which theadvantages and disadvangates of chemi- and

bioluminescence determinations in clinicalchemistry have been considered.78 Luminol(5-amino-2,3-dihydrophthalazine-1,4-dione)has found several analytical applications andis perhaps the most commonly usedluminophore in chemiluminescence deter-minations. Because H2O2 is commonly usedas an oxidizing agent in luminol chemilumi-nescence, it is not surprising that in situ gen-eration of this reactant via glucose oxidationaided by glucose oxidase is also commonlyused. Auses et al.79 proposed the determina-tion of glucose by coupling the H2O2 pro-duced with the oxidation of luminol in thepresence of Fe(CN)63 as a rate promoter.The high pH conditions required in theluminol/H2O2 reaction (pH 10 to 11) are notcompatible with the relatively mild pH (ca.pH 7.00) needed for efficient functioning ofmost enzymes as biocatalysts. An alterna-tive approach that successfully overcomesthe pH mismatch has been proposed.80 Thepresence of an inverted (reversed) micellarmedium of hexadecyltrimethylammoniumchloride permits the enzymatic and chemilu-minescence reactions to take place simulta-neously at a mild pH (7.8) and in the absenceof rate modifiers. The reversed micellar bulksolvent was a 6:5 (v/v) mixture of chloro-form and cyclohexane.

An interesting strategy to the use of glu-cose oxidase in solution for the determina-tion of glucose, avoiding the pH mismatch,was advanced in 1982 by Pilosof andNieman.81a In an unsegmented continuous-flow system, the enzyme solution is sepa-rated from the flow of other reagents andsample by a microporous membrane. Theglucose solution is allowed to flow, underpressure, through the membrane, acting as acontainment barrier, and a pH gradient iscreated into the flow cell. A 0.10 M phtha-late buffer carries the enzyme through themembrane and assures a pH of 5.0 adjacentto the membrane wall, which is close to theoptimum for the enzyme-catalyzed reaction.The sample is carried by a 0.50 M KOH

14

solution containing the luminol and tris[1,10-phenanthroline]copper(II) as a rate promoter.The pH in the bulk of the solution close tothe optical window for chemiluminescencemeasurement is about 11, optimal for thechemiluminescence reaction between luminoland the H2O2 released in the enzyme-cata-lyzed step. If immobilization is defined in ageneral form as the localization or confine-ment of a reagent,82 this contribution couldbe grouped with other applications utilizingother forms of immobilized glucose oxidase.We have opted in this review, however, forthe more restrictive definition of immobili-zation in which the immobilized reagent isinsoluble in the medium, and the inclusionof this strategy in this section of the review.The same applies to a contribution by Changet al.83 in which an outer silicone membrane(for oxygen supply) and an inner polyamidemembrane (for substrate permeation) wereused as an immobilized enzyme reactorfor glucose determination. The glucose oxi-dase is used, however, in solution and con-tained between a silicone tube and a spongelayer.

Species other than luminol and whichalso lead to chemiluminescence for determi-nation of glucose have been proposed. Wil-liams et al.,84 for example, utilized (2,4,6-trichlorophenyl)oxalate in a mixed ethylacetate-methanol-aqueous buffer system.

The use of glucose as an analytical re-agent is nearly monopolized by glucose asthe analyte. Few examples in which glucoseoxidase acts as an analytical reagent for spe-cies other than glucose can, however, becited. Toren and Burger85 studied the inhibi-tory effect of metal ions such as silver, mer-cury, and lead on the activity of the enzymeand developed indirect methods for determi-nation of these species. The indicator reac-tion was the classical o-dianisidine (in pres-ence of peroxidase) and photometricmonitoring at 440 nm was used. Silver(I)was determined in the 0.005 to 0.2 g/mlrange, mercury (II) in the 0.1 to 0.4 g/ml

range, and concentrations of lead(II) greaterthan 260 g/ml are needed for inhibition.Considerably lower concentrations of lead(II), however, seem to be amenable to de-termination using inhibition of peroxidaseand fluorescence monitoring.67 The determi-nation of molecular oxygen dissolved inaqueous and nonaqueous solvents is anotherexample of the use of glucose oxidase forthe determination of a chemical species otherthan glucose. Ghosh et al.86 proposed such adetermination by using an excess of glucoseand determining the glucose consumed by aglucose oxidase procedure87 involving per-oxidase and o-dianisidine. After mixing theexcess glucose with the glucose oxidase ando-dianisidine (peroxidase was also present),bubbling with N2 (g) was effected for 5 to7 min, the sample was incubated at 37C for30 min, the enzyme remaining active inacti-vated by heating at 100C for 4 min, andfinally the glucose consumed determined.The method is elaborate and more straight-forward means of determining dissolvedoxygen are available; its mention here isonly to point to other uses of glucose oxi-dase than for glucose determination. Theglucose oxidase-catalyzed oxidation of glu-cose by dissolved oxygen has also been usedas a coupled indicator reaction for the deter-mination of other chemical species produc-ing D-glucose in their main enzyme-cata-lyzed reaction. Starch, maltose, and lactose,for instance, have been determined in a va-riety of food products by using thisapproach.88a,88b

An interesting twist to the use of glucoseoxidase (and any other enzyme) in solutionhas been provided by Thompson et al.89 Theydeveloped a strategy for the determinationof glucose in whole blood employing, inimmobilized form, the dye of the indicatorreaction. The covalently immobilized dyeused was 3-hydroxyphenylacetic acid, andthe inert matrix for immobilization was pro-vided by micron-size porous glass beads.Glucose oxidase and horseradish peroxidase

15

in solution catalyzed reactions that resultedin converting the immobilized species into asurface-bound red quinoneimine dye, anddetection involved diffuse reflectance spec-trophotometry.

VI. ANALYTICAL APPLICATIONS OFIMMOBILIZED (INSOLUBLE)GLUCOSE OXIDASE

Immobilization refers here to the local-ization or confinement of a given chemicalspecies in such a manner that it remainsphysically separated from the substrate solu-tion and the products of the enzyme-cata-lyzed reactions. The biocatalyst is in insolubleform in the medium, and heterogeneousbiocatalysis occurs in a restricted space inwhich diffusional considerations play a morecritical role than in homogeneous catalysis.The four most common methods of enzymeimmobilization involve:90 (1) containmentby a membrane, (2) entrapment in a poly-meric gel matrix, (3) surface immobilizationby physical adsorption, and (4) surface im-mobilization by covalent binding. Contain-ment by a membrane and gel entrapmentfind use in the construction of the so-calledenzyme electrodes and biosensors, and ap-plications of covalent binding outweigh thoseusing physical adsorption, although physicaladsorption or containment by membranes isprevalent in early developments. Basic andapplied aspects of enzyme immobilization canbe consulted in pertinent monographs.9092c

The use of insoluble glucose oxidaseprovides some advantages that overcomelimitations encountered when using thesoluble enzyme in solution. For example, (1)there is some increase in the retention ofenzyme activity with time, (2) easy separa-tion and recovery are accomplished withminimum (if any) contamination of the en-zyme preparation by reactants and products,and (3) the insoluble enzyme preparationsare ubiquitous in the design of reactor/sens-

ing units and are adaptable to continuous-flow sample and reagent processing. In viewof this ubiquity, this part of the review fo-cuses on the use of insoluble glucose oxi-dase in (1) reactors located at some distancefrom the detection system and (2) reactorsintegrated with the detection unit. The over-whelming use of external reactors is realizedin the form of packed reactors in unsegmentedcontinuos-flow processing (e.g. flow injec-tion analyses). Integrated reactor/detectionunits appear predominantly in what is de-nominated as biosensors. The same type ofdetection (electrochemical, optically aided[e.g., photometric], or enthalpimetric), how-ever, can be used with either of the twostrategies; therefore, the focus here is on thelocation of the enzyme reactor and its char-acteristics.

VII. GLUCOSE OXIDASEREACTORS LOCATED AT ADISTANCE FROM THE DETECTOR

The immense majority of applicationsutilizing reactors located at some distancefrom the detector comprise flow systems,particularly the so-called flow injectionanalyses. Monographs on the topic providethe background for interested readers.93a,93b

Short reviews focused on reactors of thistype are also available.82,94 In essence, mostof them involve short (1 to 3 cm in length)reactors of small diameter (0.5 to 2 mm i.d.)packed with the immobilized enzyme prepa-ration. In a few cases the column contains asingle-bead string packing with the enzymeimmobilized on the beads, or is located inthe sample loop of the injection valve.82 Alsoin a few examples, the enzyme was directlyimmobilized on the walls of the reactor. Thepacked column located after injection in thecontinuous-flow manifold is, however, themost commonly used variation, although thissimple approach does not fully utilize thepotential of the immobilized enzyme

16

preparation.95a,95b Table 4 is a selective com-pilation of the reactors containing immobi-lized glucose oxidase located at a certaindistance from detection and involving packedcolumn or open column reactors. Also in-cluded at the end of Table 4 is a sample ofmiscellaneous reactors in which the bio-reactor part is located close to the point ofdetection.

VIII. GLUCOSE OXIDASEREACTORS INTEGRATED WITHTHE DETECTOR

A. Electrochemical Sensing

Since the basic concept of an enzymeelectrode was first described by Clark andLyons in 1962,133 there has been an impres-sive proliferation of biosensing design usingglucose oxidase. The first glucose oxidaseenzyme electrode was described by Updikeand Hicks134 and was designed for the con-tinuous measurement of blood glucose. Ac-tually, it was Updike and Hicks who coinedthe term enzyme electrode, a questionabledesignation because these electrodes are notused for the determination of enzyme activ-ity and the electrode per se is not made ofenzyme.135 In any event, the reactor/elec-trode designed by these authors was basedon the amperometric measurement of O2 (g)depletion in an immobilized glucose oxidasegel of uniform particle size. The enzymereactor part is a glucose oxidase enzymeentrapped within an acrylamide gel mem-brane prepared by mixing, at pH 7.4, thesoluble enzyme preparation with solutionsof acrylamide and N,N-methylenebisacryl-amide in 0.10 M phosphate buffer.

An interesting variation afforded by elec-trochemical measurement was described byWilliams et al.136 They proposed to usebenzoquinone instead of oxygen in the mainenzyme-catalyzed reaction and based theelectrochemical measurement on oxidation

of the hydroquinone formed. Thus, the elec-trochemical strategy is based on the follow-ing reactions:

GODGlucose + Benzoquinone + H2O Gluconic acid + Hydroquinone

Hydroquinone Benzoquinone + 2H+ + 2e- (E = 0.40 V vs. SCE)

Because the quinone is regenerated inthe electrochemical half-reaction, the con-centration of the oxidant is for all practicalpurposes constant, a property of kinetic sig-nificance. The authors proposed a reactor/detector system composed of the electro-chemical sensor (platinum electrode), en-zyme reaction layer (enzyme trapped in po-rous or jelled layer), and diffusion anddialysis layer (e.g., dialysis membrane). A0.006-in., 100-mesh nylon screen was usedas the porous structure to entrap the enzymesolution, and the diffusion barrier was pro-vided by a cellophane film.

Amperometric reactor/detector units havebeen classified as follows:137 (1) first-gen-eration units, based on the detection of H2O2or oxygen consumption, (2) second-genera-tion units, which employ an electron media-tor to facilitate electron communication be-tween the active site of the enzyme and theelectrode surface, and (3) third-generationunits, which make use of special electrodematerials allowing direct electron trans-fer between enzyme and electrode.This classification was used as the basisfor evaluating glucose electrodes for thedetermination of glucose in whole bloodby Gunasingham et al.138 Guilbault andLubrano139a,139b attached a glucose oxidasereactor layer to platinum used as an anodefor H2O2 detection. The reactor layer wasglucose oxidase physically entrapped in apolyacrylamide gel. Glucose oxidase can alsobe physically immobilized by direct admix-ing in a carbon paste formulation foramperometric measurements in continuous-flow systems.140

Shu and Wilson141 have attached glucoseoxidase to carbonaceous rotating electrodesurfaces. The bioreacting surface was pre-

17

TABLE 4Selected Examples of Reactors Containing Immobilized Glucose Oxidase

I. Reactors located after sample introduction but before detection in continuous-flowmanifolds

Spectrophotometric detection

Reactor type Comments Ref.

Packed A 1-ml disposable syringe packed with glucose 96oxidase immobilized on polyacrylamide gel used asreactor; unsegmented continuous-flow system;glucose determination used to characterize the system

Open Glucose oxidase covalently attached to the inner 97wall of a coiled diazotized polyaminostyrene tube(375 cm long, 0.2 cm i.d.); air-segmentedcontinuous-flow system

Packed and open Glucose oxidase covalently attached to type 6 98nylon powder, membrane, and tubing (1-mmbore); glutaraldehyde attachment; air-segmentedcontinuous-flow system for glucose determination

Open Tubular glucose oxidase wall reactors; air-segmented99a, 99b, 99ccontinuous-flow system

Open Glucose oxidase physically entrapped between 100two dialysis membranes and the resulting reactorlocated in the dialyzer unit of a Technicon air-segmented analyzer

Packed Enzyme covalently bound to controlled-pore 101glass; air-segmented continuous-flow system

Open Enzyme immobilized on the walls of a hydrolyzed 102nylon tubing; the reactor is intercalated in one ofthe arms of a stopped-flow mixing unit, and betweenthe push liquid syringe and the flow cell; the secondarm of the unit is used to deliver the indicator reagents(iodide and Mo[VI])

Packed Two packed column reactors in tandem, one containing103glucose oxidase and the other horseradish peroxidase;both enzymes immobilized on controlled-pore glass

Packed The system was designed for the determination of 104dissolved oxygen; the effluent from the reactor (theglucose-glucose oxidase system acts as indicator) ismerged with a chromogen and fed into a secondpacked reactor with immobilized peroxidase

Packed Glucose oxidase immobilized on a single-bead-string 105reactor for the study of the kinetics of D-glucosemutarotation

Packed Enzyme immobilized on glass beads; on-line 106monitoring of glucose for control of fermen-tation processes

Packed Glucose oxidase and peroxidase immobilized on 107controlled-pore glass and packed in the same reactor;different flow programming evaluated

18

TABLE 4 (continued)Selected Examples of Reactors Containing Immobilized Glucose Oxidase


Electrochemical detection


Packed Enzyme-gel capillary column; glucose oxidase 108entrapped in the gel matrix by photocatalyticpolymerization of acrylamide and N,N-methylbis-acrylamide; Clark-type oxygen electrode used fordetection

Packed Closed-loop recirculating system; Enzyme immobi- 109lized on a water-insoluble carrier; oxygen detection

Packed Enzyme immobilized on controlled-pore glass; 110a, 110bmonitoring of dissolved O2 using Clark-type electrode

Open Glucose oxidase immobilized on a nylon tube 111activated by alkylation with dimethyl sulfate usingamine and glutaraldehyde spacers; monitoring ofdissolved oxygen

Packed Glucose oxidase covalently immobilized on porous 112alumina; amperometric monitoring of H2O2

Packed Glucose determination in plasma; glucose oxidase 113immobilized on controlled-pore glass; hydrogenperoxide amperometrically detected in a flow-throughcell with two Pt electrodes with a 0.60-V potentialdifference

Packed Glucose oxidase immobilized on controlled-pore glass114by using an immunological reaction; unsegmentedcontinuous-flow system; hydrogen peroxide monitoring;glucose determination

Packed Determination of sucrose by converting to glucose 115in a reactor containing immobilized glucose isomeraseand glucose oxidase; in series reactors containing theenzymes separated also described; amperometricmonitoring of dissolved oxygen

Packed Continuous monitoring of glucose; a microdialysis 116fiber used to withdraw the sample; after passingthrough the fiber, the perfusion liquid entered amicroreactor with glucose oxidase immobilized oncontrolled-pore glass; electrochemical monitoringof dissolved oxygen levels

Packed Glucose determination; glucose oxidase immobilized 117on aminopropyl-derivatized controlled-pore glass andpacked in a polycarbonate tube; continuous-flowsystem and amperometric detection of H2O2 with awall-jet electrode

19



Electrochemical detection


Membrane Glucose oxidase immobilized on poly(2-hydroxyethyl 118methacrylate) membranes by entrapment and polyme-rization of a coat of hydroxyethyl methacrylate;this resulted in a sandwich with glucose oxidase inthe center with high loading and enzyme activitycontinuous-flow system and monitoring of dissolvedoxygen; potential application for glucose determination

Packed Enzyme immobilized with near 100% efficiency onto 11910-m tresyl-activated silica beads (1000 to 500 poresize); bead slurry packed into 2- 20-mm columns);columns stable for more than 40 d; limit of detectionfor glucose in the picomole range; amperometric detection

Luminescence detection


Packed Glucose oxidase immobilized on Sepharose and packed120a, 120b,in a Pyrex tube (4 mm i.d., 2 cm length); luminol 120cchemiluminescence with Fe(CN)63 as rate modifier;determination of glucose in urine samples

Packed Glucose oxidase and peroxidase coimmobilized on 121diazotized polyaminostyrene beads packed in glasscoils; continuous-flow system; fluorescence of theoxidation product of homovanillic acid (excitation:315 nm, emission: 425 nm)

Packed Glucose oxidase immobilized on controlled-pore 122glass and packed in a column located away from thepoint of detection; peroxidase immobilized on a glassdisk located in the flow cell; determination of glucosein serum by luminol chemiluminescence; peroxidaseacts as rate modifier

Packed Miniaturized (microconduit) flow injection system; 123glucose oxidase immobilized on controlled-poreglass; luminol chemiluminescence (glucose determi-nation) with Fe(CN)63 as rate modifier

Open Glucose oxidase immobilized on the wall of a nylon 124a, 124b,tubing; monitoring of glucose concentration during 124cfermentation using a continuous-flow system; luminolchemiluminescence

20


I. Reactors located after sample introduction but before detection in continuous-flow mani-folds

Luminescence detection


Packed Glucose oxidase immobilized with glutaraldehyde 125attachment to aminopropyl controlled-pore glass;luminol chemiluminescence with peroxidase in thecarrier solution as rate modifier; determination ofglucose in animal cell cultures

Image analysis


Packed Image analyzer system for multicomponent analysis;126glucose oxidase and glucose-6-phosphate immobilizedon a highly crosslinked polydextran support; enzymepreparations packed into a capillary tube; a section ofcapillary tube captured for image analysis; stopped flowused in conjunction with a charged coupled device and avideo cassette recorder for spatial resolution of closelyspaced responses

Post-column reactors for high-performance liquid chromatography


Packed Sequential immobilized enzyme reactors to hydrolyze127b -D-glucosides to b -D-glucose (using b -glucosidase)and then produce H2O2 with glucose oxidase; luminolchemiluminescence with horseradish peroxidase as ratemodifier

II. Miscellaneous reactor configurations

Comments Ref.

Glucose oxidase immobilized on the inner wall of a nylon tube inserted into the128observation cell of a commercial stopped-flow spectrophotometer

Cylindrical magnetic stirrer designed to hold glucose oxidase and peroxidase129(immobilized on cellulose); a nylon cloth tube acts as barrier permitting the flow/stopped-diffusion of substrate and products of the enzyme-catalyzed reactions

Integrated rotating bioreactor-amperometric detection unit; bioreactor contains130a, 130b,glucose oxidase immobilized on top of a rotating disk; rotation minimizes 130c, 130ddiffusional constraints and allows high rates of the catalyzed reaction with verysmall amounts of enzyme; amperometric detection afforded by using astationary Pt-ring electrode concentric to the rotating bioreactor; continuous-flow/stopped-flow/continuous-flow processing of sample and reagents

21


II. Miscellaneous reactor configurations

Comments Ref.

Spectrophotometric cell comprising parallel bioreactors facing each other; the131upper reactor is fixed and contains a film of immobilized glucose oxidase; thelower reactor rotates and contains a film of immobilized horseradish peroxidase;operating characteristics of the cell illustrated with the determination of glucosein serum samples utilizing Trinders reaction system

Three different surface strategies based on sol-gel-derived glasses evaluated132as sensing platforms to detect glucose by doping with glucose oxidase; thethree strategies tested were based on: (1) physisorption, (2) microencapsulation,and (3) a sol-gel/glucose oxidase/sol-gel sandwich; the sandwich configurationwas found to provide fast response and high enzyme loading; amperometricand photometric detection employed to determine glucose

pared by forming a carbon paste with graph-ite, Nujol as pasting liquid, and n-octa-decylamine. The amine group was then re-acted with glutaraldehyde and bovine serumalbumin, and finally the enzyme (in a glut-araldehyde solution) was incorporated in thegelatinous membrane formed. This is an-other example in which the use of immobi-lized glucose oxidase is incidental but justi-fied to demonstrate the potentials of anenzyme-reactor layer in the disk of a rotatingring-disk electrode configuration. Later on,Kamin and Wilson142 extended the studies tographitic oxide (prepared by wet chemicaloxidation or dry oxygen plasma ashing) andplatinum surfaces. An interesting observa-tion emanating from these studies is theirconclusion that at rotation speeds equal orlarger than 1600 rpm, the overall processoperates under catalytic control. Contempo-rarily to these studies, Bourdillon et al.143

covalently attached glucose oxidase to gra-phitic surfaces that were first chemicallyoxidized (nitric acid + dichromate treatment)with the help of electrochemical cycling andfinal enzyme attachment using carbodiimideactivation.90 Glucose has also been covalentlyimmobilized on graphitic electrodes by alinking procedure involving a carbodiimidereagent.144 The carbonaceous surface is elec-

trochemically pretreated in a 0.10 M KNO3solution (oxidation at +2.3 V vs. SCE) be-fore attaching the enzyme.

Strategies used to design the glucoseoxidase-reactor unit are numerous, and thisappears as a point that has stimulated theimagination of those working with ampero-metric methods, and chemically modifiedelectrodes in particular. Chi and Dong145

modified a glassy carbon surface by electro-chemically codepositing palladium and glu-cose oxidase, and covering the surface witha thin film of Nafion (perfluorinated ion-exchange membrane). The authors report thatthere is no obvious interference from sub-stances such as ascorbate and saccharides.Glucose oxidase has also been adsorbed ona modified electrode of palladium/gold sput-tered on graphite.146

Although recently the focus has been onamperometric monitoring (mainly of H2O2production), some potentiometric reactor/detection units can be found in contributionsof some years back. Liu et al.,147 for in-stance, coentrapped glucose oxidase andcatalase in a polyacrylamide gel layer arounda platinum screen to provide a potentiomet-ric unit in which the potential difference wasfound to be logarithmically related to glu-cose concentration.

22

A microhole-array electrode (bearing1000 microholes of 7 m diameter and 50 to100 m depth) was fabricated by immobiliz-ing glucose oxidase on the surface of theplatinized microhole array.148 Immobiliza-tion was accomplished by immersion into asolution containing the enzyme and 1-cyclohexyl-3-(2-morpholinoethyl)carbo-diimide metho-p-toluenesulfonate. A non-aqueous photopolymer (UV irradiation ofpartially hydrolyzed poly[methyl methacry-late] as binder, bisphenol A-bis[2-hydroxy-propyl methacrylate] as monomer, and aketone-benzophenone system as initiator)containing dispersed glucose oxidase wasused to construct photolithographically pat-terned membranes for amperometric glucosedetermination.149 These glucose electrodescan have an adjustable linear response rangeprovided by deposition of membranes con-taining a surfactant (dodecyltrimethyl-ammonium chloride) that is leached out byconditioning.

Blanger et al.150 incorporated platinummicroparticles into a polypyrrole/glucoseoxidase film grown potentiostatically(+0.65 V vs. SCE) from aqueous solutionscontaining pyrrole and the enzyme. The dis-persed platinum particles were incorporatedby immersion of the film in a hexachloro-platinate(IV) solution. The platinized elec-trode gave amperometric responses at +0.7 Vvs. SCE that were 40% higher than thoseobtained in the absence of platinummicroparticles.

A simple strip-type electrode preparedfrom platinized Vulcan XC-72 carbon par-ticles by using silk screen printing techniqueswas described by Cardosi and Birch.151 Glu-cose oxidase was covalently attached to thesurface of the carbon particles by acarbodiimide-based bonding. Thick-film andlaser micromachining procedures have beenused to electrochemically localize glucoseoxidase within the pores of 15-m disks viacodepositing with rhodium or platinum.152

The approach is considered potentially use-

ful in the production of glucose oxidase-based diagnostic strips.

Gough et al.153a,153b described a reactor/sensor unit in which oxygen diffuses fromtwo directions (a hydrophobic membranesurrounding a cylindrical oxygen sensor andthe enzyme-containing gel), whereas theanalyte (glucose) diffuses only through theenzyme-containing membrane located at thetip of the unit. The sensing element mea-sures the excess oxygen that is not consumedas a result of the glucose oxidase-catalyzedreaction. The glucose oxidase is immobi-lized within a gel made with denatured bo-vine Achilles tendon collagen crosslinkedwith glutaraldehyde. The configuration is animprovement over previous designs basedon the same sensing strategy. Some theoreti-cal considerations about the performance ofcylindrical amperometric reactor/detectorunits employing soluble redox mediators canbe found in the literature.154

Glucose-sensitive field-effect transistorsensors have been prepared by crosslinkingglucose oxidase with bovine serum albuminin an atmosphere saturated with glutaralde-hyde vapor.155 The crosslinked enzyme waslocated on top of the gate area and coveredwith a Nafion membrane deposited by a spin-coating procedure. Mizutani et al.156 alsocovered an amperometric sensing surfacewith a Nafion membrane utilizing a layer oflipid-modified glucose oxidase. A glassycarbon surface was first dipped into a ben-zene solution of the modified enzyme anddried. This was followed by dipping into aNafion solution and drying again. The elec-trode was used for at least 6 weeks, exhib-ited rather rapid response, and was appliedto the determination of glucose in fruit juices(orange and apple).

A composite barrier made of a diamond-like carbon-coated microporous polycarbon-ate membrane, impregnated with a cross-linking glucose oxidase/bovine serumalbumin/glutaraldehyde mixture, was usedto cover a commercial oxygen electrode

23

assembly.157 The membrane acts as an en-zyme reactor, protects the electrode, andpermits the determination of glucose inwhole blood.

The so-called biosensors based on im-mobilized glucose oxidase are obviouslynumerous. Besides the examples alreadymentioned in this review and those includedin Table 4, glucose oxidase has been, forinstance, incorporated into polypyrrole filmselectrically produced on glassy carbon orplatinum electrodes,158 poly(amphilic pyr-role) films,159 and poly(N-methylpyrrole)electrochemically deposited on gold sur-faces;160 immobilized in poly(o-phenyl-enediamine) films by potentiometric electro-polymerization on platinum surfaces,161a,161b

entrapped into a lipid matrix,162 adsorbed onplatinized carbon paper163 and platinum,164

electrodeposited on platinum black togetherwith bovine serum albumin, and finallycrosslinked with glutaraldehyde.165 Centonzeet al.166 discussed the influence of ascorbicacid on the response of a glucose sensor withthe enzyme immobilized on electropoly-merized poly(o-phenylenediamine) oroveroxidized poly(pyrrole) films on aplatinum electrode. They concluded that thedecrease in sensor response cannot be attrib-uted to depletion of H2O2 via the homo-geneous reaction with ascorbate. The culpritseems rather to be the electrode surface foul-ing by electrooxidation products of ascorbicacid. The overoxidized poly(pyrrole) filmcompletely eliminates the interference ofascorbate.

A great deal of interest in the develop-ment of electron shuttling to facilitate theelectron transfer between the active centerof the enzyme and an electrode can be ob-served in the last two decades. The majorityof these efforts involve the use of chemicalspecies used as redox mediators. Figure 3schematically shows how redox mediatorscan facilitate the electrical communicationbetween the enzyme active site and the elec-trode surface.

Table 5 provides a tabulation of redoxmediators that have been proposed and usedin conjunction with glucose oxidase.

Direct communication can be achieved,with the enzyme practically becoming anintrinsic part of the electrode. The admixingwith carbon paste already cited96 is an ex-ample, and the composite formulation pro-posed by Cspedes et al.192a,192b represents aneffort in the same direction. The compositecontained graphite, palladium-gold in anepoxy resin, and glucose oxidase. The ma-trix was used to detect the electrocatalyticoxidation of H2O2 at the gold-palladium con-ductor, and simple polishing using a 3-malumina paper wetted with distilled watergenerated fresh surfaces for detection pur-poses. These efforts were preceded by abioreactor/detector design in which the elec-trode was a graphite-epoxy composite andthe enzyme reactor part was a nylon 6,6membrane with glucose oxidase immobilizedafter activation with dimethyl sulfate.192b

More recently, Sakslund et al.193 describedthe preparation of a glucose biosensor byelectrochemically codepositing palladiumand glucose oxidase on a glassy carbon elec-trode.

Nolte et al.194a,194b reported on a biosensorapproach involving glucose oxidase, whichthey claim results in direct communicationbetween the enzyme and polypyrrole as aconducting polymer synthesized inside thepores (tubules) of a filtration membrane.More recently, however, Kuwabata andMartin195 studied the mechanism of theamperometric response to glucose of thistype of sensor and concluded that the currentsignal results from direct electrochemicaloxidation of glucose at the platinum filmcoated onto one face of the membrane. Thisfinding opens an interesting question withrespect to the role of the many redox media-tors proposed in the literature. Moore et al.196

turned the table in the tactical approach ofusing mediators and presented the determi-nation of some pharmaceuticals (e.g., ac-

24

FIGURE 3. Possible ways of organizing electron transport from the enzyme active site to theelectrode. (A) Mediator and enzyme in solution. (B) Enzyme immobilized and mediator in solution.(C) Mediator immobilized and enzyme in solution. (D) Mediator and enzyme both immobilized onthe surface of the electrode.

etaminophen, norepinephrine, and chlor-promazine) in which the analyte plays therole of redox mediator. The amplificationscheme allows limits of detection in the 0.1to 0.3 M range.

Of special interest in the development ofelectrodes with a direct electrical communi-cation between the redox enzyme and theelectrode itself are efforts centered on chemi-cal modification of the enzyme. Degani andHeller,197a197c for instance, have attachedinside the enzyme different chemical spe-cies acting as electron relays. Direct com-munication entails electron transfer from theenzymes redox centers to relays that arecloser to the periphery of the enzyme; theseelectrons are then transferred at practical rates

to the electrode surface. Therefore, directelectrical communication is established be-tween the FAD/FADH2 centers of the en-zyme and gold, platinum, or carbon-con-ducting surfaces. Examples of relaysinclude: (1) amides formed between en-zyme amines and ferrocenylacetic acid,ferrocenecarboxylic acid, or rutheniumpentaamine complexes of isonicotinic acid;(2) azo compounds made by reacting thediazonium salt of 4-aminopyridine withthe enzyme, followed by complexing theenzyme-bound pyridine with rutheniumpentaamine, and (3) ruthenium pentaaminechemically bound to the enzyme. 3-Carboxymethylpyrrole has been attachedcovalently to lysyl residues of glucose oxi-

25

TABLE 5Typical Redox Mediators Used in Conjunction with Glucose Oxidase inBioreactor/Detection Systems

Mediator Comments Ref.

Hexacyanoferrate(III), Slightly recessed Pt electrode coated with 167p-benzoquinone, 2,6- a thin layer of enzyme solution; enzymedichlorophenolindo- solution separated from test solutionphenol, pyocyanine, containing the mediator by a cellophanethionine, or methylene membraneblue

Organic metal complexes Organic metal complexes pressed into conducting168of N-methylphenazinium (metallic) disks glued to a glass tube;or N-methylacridinium and solution containing enzymes (glucose oxidase,the anion radical tetracyano-cytochrome b, and peroxidase) entrapped onp-quinodimethane. the surface of the electrode by using a dialysis

membrane7,7,8,8-Tetracyano-p-quino- Glucose oxidase in solution entrapped by a 169dimethane, quinone, pyocy- dialysis membrane as a layer adjacent to aanine dichlorophenolindo- glass-carbon electrode coated with mediatorphenol, dextran-dopamine

Ferrocene and derivatives Mediator deposited onto the surface of the 170electrode and air dried; enzyme covalentlyattached (carbodiimide) to the surface of theelectrode and covered with a polycarbonatemembrane

Carbon paste electrodes modified by direct 171a, 171b,admixing of glucose oxidase and modifier 171c, 171d

Benzoquinones Glucose oxidase and mediator entrapped 172(glutaraldehyde crosslinking) on the surfaceof a graphite foil disk

Glucose oxidase immobilized (adsorption) on 173a carbon electrode by covering its surface witha solution of the enzyme and allowing solventto evaporate; mediator in solution

Crosslinked and noncrosslinked poly(ether- 174aminequinone) used as electron-transfer relaysin carbon paste electrodes

p-Benzoquinone and ferrocenemonocarboxylic 175acid; polyaniline-coated platinum electrode

Ruthenium compounds, Mediator and enzyme used in solution; catalytic 176octacyanotungstate(VI) currents observed at carbon electrodes. 177and -molybdate(VI)

Viologens Several water-soluble viologens and glucose 178oxidase incorporated into carbon pasteformulations

Cobalt phthalocyanine Enzyme and mediator immobilized as a colloidal 179graphite dispersion matrix applied to a glassycarbon disc electrode; very good storage stability

N-Methylphenazin-5-ium Mediator adsorbed on graphite; enzyme 180a, 180bcovalently attached (via activation with cyanuricchloride) to oxidized graphite surfaces afterselective reduction by LiAlH4

26

TABLE 5 (continued)Typical Redox Mediators used in Conjunction with Glucose Oxidase inBioreactor/Detection Systems

Mediator Comments Ref.

Tetrathiatetracene (TTT), TTT and TTF in solution, DMTSF adsorbed 181tetrathiafulvalelene (TTF), together with glucose oxidase on a graphiteand 1, 2-dimethyltetra- electrodeselenafulvane (DMTSF)

Tetrathiafulvalene Ink-jet printing on poly(vinylchloride) for 182deposit of mediator and glucose oxidase

Mediator in solution; microelectrochemical 183enzyme transistor prepared by connectingtwo carbon band electrodes; glucose oxidaseimmobilized on an anodically grown filmof poly(aniline)

Osmium-bipyridyl Mediator incorporated as a crosslinkable 184a, 184b,complexes poly(vinylpyridine) complex; the mediator 184c, 184d

was used in a reactor/sensor system in whichthe measured response corresponds toreaction of all substrate (equilibrium-basedmeasurement)

Osmium-containing redox polymer with 185bipyridyl, poly-(4-vinylpyridine) and chloride;enzyme and mediator codeposited on thesurface of a Pt electrode with glutaraldehyde;the resulting surface covered with anelectropolymerized layer of pyrrole alsocontaining glucose oxidase

Enzyme and redox polymer codeposited by 186embedding vitreous carbon rods in a Teflonshroud using a low-viscosity epoxy resin

Glucose oxidase immobilized in a hydrogel 187formed by crosslinking poly(1-vinylimidazole)complexed with Os(4,4 -dimethylbipyridyl)2Clwith poly(ethylene glycol) diglycidyl ether

7-Dimethylamino-1,2- Mediator and glucose oxidase admixed in a 188benzophenoxazinium carbon paste electrodesalt (Meldola Blue)tetrathiafulvane- Mediator and glucose oxidase incorporated 1897,7,8,8-tetracyano- together in carbon paste electrodesquinodimethane

4,4 -Dihydroxybiphenyl Mediator tested in solution, adsorbed on glassy 190carbon, and also electropolymerized on glassycarbon

Hydroquinone Glucose oxidase and hydroquinone coimmo- 191bilized in a carbon paste electrode; surfaceactivation shown to improve electrochemicalreversibility and analytical performance;activation by a linearly varying potentialbetween 0.6 and 2 V at 50 mV/s for a giventime in unstirred solution of 0.50 M NaOHor NaHCO3

27

dase and copolymerized with unmodifiedpyrrole to produce films of high conduc-tivity and enzyme activity.198 Of singularinterest is a recently introduced sensor inwhich the glucose-sensing layer was madeby crosslinking a genetically engineeredglucose oxidase (Chiron Corp., Emeryville,CA) with a polymer based on poly(vinyl-imidazole) and made by complexing partof the imidazole moieties to [Os(bipyri-dine)2Cl]+/2+.199 The enzyme wired layeris part of a subcutaneously implanted glu-cose sensor operating as a one-point invivo calibration unit.

It seems appropriate to close this sec-tion of the review with the mention thatamperometric detection has been imple-mented into worldwide commercially avail-able pen-size digital glucose meters. Thesedevices follow the original design pio-neered by Matthews et al.200 and comprisea disposable strip in which the glucoseoxidase reactor and electrode system arelocated; the potentiostat/amplifier systemis contained in the pen body in which aliquid crystal display and an on/off switchare also located. The disposable strip isinserted into one of the ends of the penbody (Figure 4).

B. Optically Based Detection inReactor/Detector Units

The type of units covered here are alsocommonly known as optoelectrodes or opti-cal biosensors. A distinctive feature of thesedevices is the use of fiber optics that shuttlephotons from a source of radiant energy tothe enzyme reactor and/or from the enzymereactor to the detection devices (commonlya photomultiplier tube, photocell, or photo-diode). Consequently, these devices maymeasure absorbance, fluorescence, or chemi-luminescence. Besides the fact that photonstravel faster than electrons, these opticallybased devices avoid or minimize problemssuch as electrical interferences, electricalconnections, junction potentials, and refer-ence/auxiliary electrode pairs. The interestedreader would benefit by consulting availablemonographs201a201d and reviews.202a202e Inconsideration of the chemistry of the mainenzyme-catalyzed reaction for the determi-nation of glucose (glucose + O2 + H2O =gluconic acid + H2O2), optical sensing, incontrast to electrochemical sensing, has paidlittle attention to the H2O2, and has ratherfocused on the O2 consumption or pH de-crease. Table 6 presents examples of these

FIGURE 4. Pen-size glucose meter. (From Reference 200. With permission.)

28

TABLE 6Typical Examples of Reactor/Detection Units of the Optrode Type Based onO2-Quenching of Luminescence or pH Change

Ref

Oxygen-quenching-based devices

Quenching the fluorescence of bis(2-ethylhexyl) phthalate; first glucose oxidase-204a, 204bbased optrode; Complex signal response controlled by bidirectional masstransfer across the membrane holding the enzyme layer in place

Quenching the fluorescence of the dye decacyclene dissolved in a very thin 205silicone-based membrane located beneath the enzyme layer; dynamic range:1 to 20 mM; response stable for at least 5 months

Glucose oxidase immobilized on the surface of the oxygen optrode by 206adsorption onto carbon black and crosslinking with glutaraldehyde; carbonblack used as an optical insulator to protect the oxygen probe from ambientlight and sample fluorescence interference

Quenching the fluorescence of tris(1,10-phenanthroline)ruthenium(II) cation; 207complex adsorbed onto silica gel incorporated in a silicone matrix possessinghigh oxygen permeability, and placed at the tip of the optical fiber of the oxygensensor that also contained the enzyme

Hydrogen ion (pH)-sensing devices

Bromocresol Green indicator; monitoring of absorbance change; optoelectronic208light-emitting diodes used

Fluorescence indicator (1-hydroxypyrene-3,6,8-trisulfonate); glucose detected 209in the 0.10 to 2 mM range; Slow response (8 to 12 min for 90% signal) that dependson the buffer capacity of the medium

approaches. An exception seems to be thework of Trettnak and Wolfbeis203 in whichthe intrinsic fluorescence of glucose oxidaseis utilized as the basis for measurement ofglucose concentration. Excitation is at 450nm and measured emission at 550 nm. Asolution of glucose oxidase or a gel-entrapped(bovine serum albumin and glutaraldehyde)enzyme was compartmentalized by a dialy-sis membrane at the end of the fiber opticconduit. Both approaches provided compa-rable responses to glucose concentration.

C. Thermal Detection

During the 1970s thermal detection ofreaction enthalpy in glucose oxidase-cata-lyzed oxidation of glucose received consid-

erable attention and commercial units en-tered the market. A unique property ofenthalpimetric monitoring is that there arevery few chemical processes that do not in-volve at least 5 (kcal.mol,1)87a and this is ofparticular analytical interest. Experimentalaspects and calculations related to glucoseoxidase thermal probes can be found in theliterature210 as well as in reviews pertinent tothe topic.211a,211b

The detection unit (mostly thermistors)is located in very close proximity to the siteof reaction (reactor), because the heat changeis most pronounced in the vicinity of theenzyme active site environment. Accordingto Mosbach and Danielson212 there are twoways of accomplishing this: (1) direct im-mobilization of the enzyme onto the ther-mistor and (2) encirclement of the thermistor

29

by a coil of tubing packed with matrix-boundenzyme.

One of the first applications of thermalmonitoring for glucose determination wasintroduced in 1973 by Johansson.213 Subse-quent examples of thermometric monitoringfor glucose determination using immobilizedglucose oxidase are summarized in Table 7.Muehlbauer et al.214 have mathematicallymodeled the response of a calorimetric glu-cose sensor with the purpose of describingthe energy and mass balance. The validity ofthe model was ascertained with experimen-tal data obtained with a prototype sensorconsisting of a thermopile to which a mem-brane containing immobilized glucose oxi-dase and catalase was attached.

IX. THE USE OF GLUCOSEOXIDASE IN ENZYMEIMMUNOASSAYS

Immunoassays designate a variety ofdeterminative procedures that are widely usedin biomedical research and in clinical labo-ratories.222 Immunology, protein chemistry,and enzymology have greatly benefited from

applications involving enzyme immunoas-says. Early developments centered on whathas become to be known as radioimmunoas-says; since then, the literature has recordedan exponential growth in the development ofisotopic as well as nonisotopic immunoas-says. Basically, immunoassays rely on a re-versible antigen-antibody reaction that canbe represented as follows

Ag + Ab AgAb

where Ag is free antigen, Ab free antibodysites, and AgAb the antigen-antibody com-plex. Antibodies are bifunctional moleculesthat bind antigens at specific sites and serveas linkers of the specific antigen to immunesystem cells. Enzymes are used as labelingagents to aid in the detection; examples ofenzyme immunoassays are presented inTable 8, which adopts the classification ofWisdom.223

Although glucose oxidase fulfills theconditions of compatibility for use as a labelin immunoassays,225a,b including the desir-able property of almost nil background inmammalian tissue,225b it has been used lessfrequently than other enzymes such as per-

TABLE 7Examples of Reactor/Detector Units Employing Immobilized Glucose Oxidaseand Thermometric Detection

Ref.

Thermistor (sensor) in contact with the enzyme that is gel immobilized and 215fills a microcolumn in contact with the detection unit

Capillaries used for heat transfer from the enzyme reactor part to the thermistor216(no direct contact); time needed per determination: 20 min

Use of a tubular reactor; time per determination: 5 min. 217Small glass-encapsulated thermistor containing immobilized catalase and located218at the end of a packed column containing immobilized glucose oxidase

Divided-flow enzyme thermistor composed of two similar microcolumns;219a, 219b,one of the columns contains the immobilized enzyme and the other is filled 219cwith glass beads

Thermistor coated with a membrane containing the glucose oxidase and prepared220by glutaraldehyde crosslinking of serum albumin

Array of p-type semiconducting silicon-aluminum strips integrated onto a thin 221silicone membrane; glucose oxidase and catalase directly immobilized on theback side of the thermopile

30

TABLE 8Operational Classification of Enzyme Immunoassays, EIA

DiagrammaticIllustration Comments

Homogeneous EIAa

The covalently enzyme-labeled antigen competes withthe unlabeled antigen in the sample for a limited concen-tration of antibody; it has been suggested that the inhi-bition of enzyme activity is caused by the complexationof antibody molecules to haptens located sufficientlynear the active site of the enzyme;224 the substrate isthen sterically excluded

Hetereneous EIAb

Competitive enzyme immunoassay for antigen. Adheresto the general principles of saturation analysis; The la-beled antigen competes with unlabeled antigen from thesample for a limited amount of solid-phase-immobilizedantibody; after brief incubation, the antibody-bound Ag-E is separated from the unbound Ag-E; the enzymeactivity associated with the solid phase is inversely re-lated to concentration of the analyte

Competitive enzyme immunoassay for antibody. Labeledand unlabeled antibody compete for immobilized anti-gen; the amount of antibody is determined by measuringthe enzyme activity in the free or bound fractions, aftercentrifugation

Immunoenzymometric assay. Excess of labeled antibodyreacts with antigen; then an excess of immobilized anti-gen is added; after centrifugation, the activity associatedwith the soluble antigen fraction is measured

Sandwich enzyme immunoassay. Requires antigens withmultiple binding sites (epitodes); the antigen being deter-mined is held between two different antibodies; excess ofimmobilized antibody is added to the sample, and afterincubation, followed by washing, the enzyme-labeled an-tibody is added; the enzyme activity associated with theimmobilized antibody provides a direct measurement ofthe amount of antigen present in the sample

31

TABLE 8 (continued)Operational Classification of Enzyme Immunoassays, EIA

DiagrammaticIllustration Comments

Hetergeneous EIAb

Enzyme immunoassay for antibody. Very similar to thesandwich immunoassay; the antigen is immobilized, how-ever; provides direct measurement of the amount ofspecific antibody present

a This approach does not require separation of free and bound antigen.b More sensitive and less prone to interferences. A separation step is required.

oxidase, alkaline phosphatase, and b -galac-tosidase.226 Examples of immunoassays inwhich glucose oxidase is used as label aretabulated in Table 9.

X. THE DETERMINATION OFGLUCOSE OXIDASE ACTIVITY

A review on the use of glucose oxidaseas an analytical reagent cannot be closedwithout mentioning how to ascertain theactivity of the reagent. The catalytic activityof an enzyme, as mentioned earlier in thisreview, is proportional to the number ofoperating active sites at the time of its mea-surement. Although related to the concen-tration of the protein content in the medium,the activity may be different for differentpreparations of the same enzyme, and theanalytical concentration becomes useless indescribing enzyme activity. Moreover, theactivity of a given enzyme preparation maychange with time. As such, the standard-ization of enzyme preparations must bebased on activity rather than concentration.A very common procedure for assessment

of glucose oxidase activity is to determinethe reactio

glucose oxidase

Documents