Purification and semienzymic synthesis of flavin adenine dinucleotide-3′-phosphate

Purification and semienzymic synthesis of flavin adenine dinucleotide-3'-phosphate M. Fisher, S. Harbron, H. J. Eggelte and B. R. Rabin London Biotechnology Limited, Department of Biochemistry and Molecular Biology, University College London, Gower Street, London, UK

Nonradioactive immunoassays incorporating an element of amplification in their detection system require the use of components that are highly purified. Flavin adenine dinucleotide-3'-phosphate (FADP) is the primary substrate used in such an amplification assay'. For incorporation into a simple, single-pot assay system, the concentration of contaminating flavin adenine dinucleotide (a prosthetic group for the enzyme D-aminoacid oxidase used in the amplification cascade assay) in this primary substrate must be minimized to achieve maximum sensitivity. Production of the substrate to a high degree of purity has been achieved using apo-glucose oxidase to specifically remove contaminating flavin adenine dinucleotide from solution and hydrolysis of a cyclic intermediate as a final production protocol by ribonuclease Te to give the product in high yield. The use of continuous ultrafiltration reactors at each stage is described and compared to a final production step utilizing immobilized ribonuclease Te. These reactors allow large volumes of material to be handled and assist in the scale-up of these processes. The suitability of each protocol is assessed for the commercial production of FADP.

Keywords: Flavinadeninedinucleotide-3'-phosphate;apo-glucoseoxidase;ribonucleaseT2; amplificationassay

Introduction Enzyme amplification systems are increasingly being used in enzyme immunoassays (EIA). 1-4 In the system that we have developed, 3-6 a primary substrate, 3'-phosphoryl FAD (FADP), is hydrolyzed by the enzyme label, alkaline phosphatase (E.C. 3.1.3.1) to produce FAD which is the prosthetic group for a secondary enzyme, D-aminoacid oxidase (DAAO) (E.C. 1.4.3.3). The activity of this second enzyme is monitored via a coupled reaction utilizing horseradish peroxidase (E.C. 1,11.1.7). The components used in such an amplification reaction must be highly purified under conditions of GMP to reduce the background signal. We have already reported the production of apo-D-aminoacid oxidase containing less than 1 ppb of contaminating phosphatase, 7 and here we describe a method for the production of the substrate containing less than 0.0006% of FAD. Higher concentrations of contaminating FAD result in an increased background signal and

Address reprintrequeststo Dr. Fisherat LondonBiotechnologyLimited, Department of Biochemistryand MolecularBiology,UniversityCollege London,GowerStreet, LondonWCIE 6BT, UK Received 13 July 1993;revised30 September 1993

© 1994 Butterworth-Heinemann

consequent reduction in the sensitivity of the assay system. The contribution of FAD to the background signal is shown in Figure 1. The approach used is based on the removal of FAD using apo-glucose oxidase (E.C. 1.1.3.4). This enzyme has a high affinity for its prosthetic group 8 and can therefore be used to specifically bind FAD in a complex mixture. The commercial holoenzyme contains phosphatase contaminants which prevent it from being used to purify FADP directly. The immediate precursor, FADcP, is therefore purified and this is subsequently converted to FADP. A final synthesis stage involving acid hydrolysis of the cyclic phosphate moiety, to yield a mixture of the 2' and 3'phosphate isomers of FADP, results in an increase in the contaminating concentration of FAD to a level that renders the preparation unusable in the amplification assay (approximately 0.02%). Phosphatase-free ribonuclease T2 (E.C. 3.1.27.1) is therefore used in this last step. This enzyme has a specificity preferential for adenylic acid bonds that splits RNA chains into 3'-nucleotides with the intermediary formation of 2',3'-cyclic phosphates. The FADcP can therefore be utilized as a substrate to give the 3'-isomer of FADP in high yield. We have investigated the use of a variety of reactor configurations for these unit processes and report the observations made.

Enzyme Microb. Technol., 1994, vol. 16, April

2.81

Papers 1.2"

Estimation of flavin concentration Assay of FAD/FADcP/FADP. An extinction coefficient of 11,300 M - l c m - 1 at 450 nm was used to estimate the concentration of flavin.

1.0-

"~

~

Microtitre plate assay of FAD. Final concentrations of reagents in a total volume of 100 ~1 were: 0.1 M Tris-HCl pH 8.9, 10 mM sodium phosphate, 35 m s D-proline, 2.0 mM DHSA, 0.2 mM 4-AP, 1 ~g hrp, and 0.1 txM aDAAO. The rate of change of absorbance was monitored at 520 nm at 25°C. The concentration of FAD was determined by comparison with a standard curve of FAD (0-10 riM).

0.8'

0.6-

Assay of FADP. FADP was incubated with 5 × 10- 9 M alkaline

0.4

phosphatase (E.C. 3.1.3.1) in 0.1 M Tris-HC1, pH 8.0, containing 0.1 mM MgSO4 and 1.0 ~M ZnSO4, at room temperature for 60 rain. The concentration of FAD produced was determined as described above in the microtitre plate assay.

0.2 ¸

Assay of "FADP-ase" activities in ribonuclease T2 and apo-glucose oxidase

0.0 0.0

i

i

!

i

0.1

0.2

0.3

0.4

FAD

0.5

(nM)

Figure 1 Contribution of residual FAD in FADP to the background signal in the amplification assay of alkaline phosphatase. Concentration of FADP in the assay was 20 IxM

Materials and methods All experiments and assays containing FADcP, FADP, and FAD were performed under conditions of subdued lighting to limit the background production of hydrogen peroxide.

Reagents Apo-D-aminoacid oxidase (aDA.AO) was obtained from Calzyme Laboratories Inc., (St. Luis Obispo, CA, USA). Horseradish peroxidase (hrp) (grade 1) and calf intestinal alkaline phosphatase (EIA grade) were supplied by Boehringer Mannheim (Lewes, East Sussex, UK). Glucose oxidase (Grade XS), ribonuclease T2 (Grade V), FAD, D-proline, 4-aminoantipyrine (4-AP), 3,5dichloro-2-hydroxybenzene sulphonic acid (DHSA), NADPagarose, adenosine-2',3'-cyclic phosphate 5'-monophosphate, Tris-HC1, bis-Tris, and bis-Tris-propane were obtained from Sigma Chemical Co. (Poole, Dorset, UK). FADcP and FADP were supplied by London Biotechnology Limited. Other laboratory reagents were of Analar grade from Fisons (Loughborough, Leics, UK).

Equipment All microtitre plate-based assays were performed using a MR7000 plate reader fitted with a kinetics cartridge and thermostatically controlled plate holder (Dynatech, Billingshurst, West Sussex, UK). Speetrophotometric assays were performed using a Shimadzu UV-240 spectrophotometer (VA Howe, Banbury, Oxon, UK). Chromatographic equipment was supplied by Pharmacia (Milton Keynes, Bucks, UK). Centricon 10 microconcentrators and ultrafiltration apparatus (Model 8050 fitted with a YM 10 membrane) were supplied by Amicon (Gloucs, UK).

282

Enzyme Microb. Technol., 1994, vol. 16, A p r i l

The final concentrations of reagents in a total volume of I00 ~1 were: 0.1 M Tris-HCl pH 8.9, containing 0.1 mM MgSO4 and 1.0 /~M ZnSO4, 0.02 mM FADP, 35 mM D-proline, 2.0 mM DHSA, 0.4 mM 4-AP, 1 ~g hrp, and 0.1 ~M aDAAO. The sample containing "FADP-ase" activity (i0 p~l)was placed in a well of a microtitre plate and 90 o.l of the assay reagents was added. The reaction was monitored at 25°C by measuring the change in absorbance at 520 nm.

Assay of ribonuclease T2 The formation of adenosine-3'5'-bisphosphate (differential molar extinction coefficient at 270 n m = 224 cm -1) from 0.1 mM adenosine-2',3'-cyclicphosphate 5'-monophosphate in 20 mM bisTris-propane, pH 7.0, was monitored at 270 rim. One unit was the amount of enzyme hydrolyzing 1~mol of adenosine-2',3'-cyclic phosphate 5'-monophosphate in 1 minute.

Purification and immobilization of ribonuclease T2 This enzyme was purified by a modification of the method of Janski and Oleson.VRibonuclease T2 was eluted from the NADPagarose column with 0.2 M Tris pH 9.0 containing 2.5 M sodium chloride. Fractions containing enzyme activity were pooled and concentrated in a Centricon 10 microconcentrator to a final volume of approximately 200 ~1. The enzyme fraction was desalted by washing with a further 1.0 ml of 25 mM sodium acetate pH 4.5 and reconcentrating the enzyme in the Centricon 10 microconcentratot. The enzyme was aliquoted and stored at - 20°C. The enzyme was immobilized on concanavalin A-Sepharose by the method of Reddy and Shankar.10 Concanavalin A-Sepharose (1 ml) was washed with 4 volumes of 30 mM sodium acetate, pH 5.0, containing 1 mM MgC12, 1 mM MnCI2, and 1 mM CaC12. The suspended gel was incubated with 5 units of purified ribonuclease T2 in a total volume of 1 ml for 16 h at +4°C with constant agitation. The gel was separated from the bulk of the incubation buffer by centrifugation and washed 4 times with 1.0 ml of incubation buffer. No enzyme activity was detectable in the supernatant or the washing solution. The immobilized enzyme was stored in 30 m s sodium acetate, pH 5.0, containing 1 mM MgCI2, 1 mM MnC12, 1 mM CaCI2, and 30% glycerol at room temperature. To determine the efficiency of the binding of ribonuclease T2 to the matrix, the gel was packed into a column (1.0 cm i.d. × 1.0 cm) and equilibrated with 20 mM bis-Tris, pH 7.0. No ribonuclease T2 activity could be detected in the eluate. Adenosine-2',3'-cyclic phosphate 5'-monophosphate (0.2 m s ) was passed through the

Enzymic synthesis of FADP: M. Fisher column in the bis-Tris buffer at various flow rates (1-10 ml m i n - 1). Each fraction collected was incubated with 1 × 10-7M nuclease, P t (E.C.3.1.30.1) for 1 h at room temperature to convert the adenosine-3'5'-biphosphate to AMP and inorganic phosphate. The latter was measured by a variation of the method of Fiske and Subbarow.11

et al.

Continuous column reactor. FADcP (0.4 mM) was passed

through the column of immobilized ribonuclease T2 in 20 mM bis-Tris, pH 7.0, at a flow rate of 0.1 ml rain- 1 and 0.2 ml m i n - 1. Fractions were collected and the concentration of FADP in the elute was determined as previously described. Results

Determination of the apparent Km and Kp of soluble ribonuclease T2 The reaction of ribonuclease T2 and adenosine-2',3'-cyclic phosphate 5'-monophosphate (a substrate analog of FADcP) was monitored continuously, to completion, at 270 nm in 20 mM bisTris, pH 7.0 at 25°C. Substrate concentrations were varied in the range 150-200 p~M.The progress curves of each of the reactions were analyzed using the integrated Michaelis-Menten rate equation 12 to yield values for the apparent Km for the substrate and Kp for the product.

Removal of FAD from FADcP using apo-glucose oxidase

Purification and immobilization of ribonuclease T2 T h e ribonuclease T2 used contained approximately 1 p p m " F A D P - a s e " which was r e m o v e d by c h r o m a t o g r a p h y on a column o f N A D P - a g a r o s e , with a recovery of g r e a t e r than 90% of the enzyme activity. T h e kinetic p a r a m e t e r s determ i n e d for the p r o c e d u r e were an a p p a r e n t Km for the substrate of 0.216 mM and a Kp for the p r o d u c t equal to 0.103 raM. W h e n adenosine-2',3'-cyclic p h o s p h a t e 5 ' - m o n o p h o s p h a t e was utilized as the s u b s t r a t e for the i m m o b i l i z e d enzyme, the c o n c e n t r a t i o n o f active e n z y m e was m e a s u r e d as 1 unit ml - 1 o f gel, a s s u m i n g no a l t e r a t i o n in the kinetic c h a r a c t e r i s t i c s from t h o s e o f t h e soluble e n z y m e c a l c u l a t e d above.

Batch reactor. Apo-glucose oxidase was prepared by the method

of Morris and Buckler./3 FADcP, containing 0.001% FAD, was incubated at a fnal concentration of 0.5 mM with 1 ~zMapo-glucose oxidase in 20 mM bis-Tris, pH 6.0, in a volume of 1 ml. Aliquots (400 ~xl)were taken at intervals and the glucose oxidase was removed by ultrafiltration in a Centricon 10 microconcentrator. The concentration of FAD in the ultrafiltrate was determined as described above. Continuous reactor. For large-scale purification of FADcP, a continuous ultrafiltration reactor containing 25 nM apo-glucose oxidase was used. To characterize the reactor, 50 nM FAD in 20 mM Bis-Tris, pH 6.0, was passed through a 10-ml hold-up volume in the reactor. Flow rates of 0.32, 0.36, and 0.55 ml m i n - I were employed to alter the residence time of the FAD solution in the reactor. The ultrafiltrate was collected, and the concentration of FAD was determined as described above. Once characterized, the reactor was used to purify FADcP. FADcP (0.5 mM) containing 0.001% FAD was passed through a similar reactor at 0.53 and 0.65 ml min - 1. The FAD content of the ultrafiltrate was determined as previously described.

Production of FADP using ribonuclease T2

Removal of FAD employing apo-glucose oxidase Batch reactor. W h e n used in excess, apo-glucose oxidase yields F A D c P containing levels of contaminating F A D c o m p a r a b l e to those attained after multiple separations by reversed-phase chromatography, 3 but in much g r e a t e r yields. A t least 70% o f the F A D was b o u n d within the first 10 min of the incubation. TableI shows the concentration of F A D in a p r e p a r a t i o n o f F A D c P before a n d after incubation with 1 I~M apo-glucose oxidase, a n d the resulting concentration o f F A D after conversion to F A D P utilizing the purified ribonuclease T2.

Figure 2 shows the concentration of F A D in the ultrafiltrate when the r e a c t o r was o p e r a t e d at flow rates o f 0.32, 0.36, and 0.55 ml min - 1. M o r e than 90% of the F A D was removed from the solution during the initial stages o f the o p e r a t i o n in each instance. T h e r e d u c e d binding o f F A D after 8-10 ml was due to the apo-glucose oxidasc b e c o m i n g s a t u r a t e d with F A D . A t the m a x i m u m C o n t i n u o u s reactor.

Batch reactor. FADcP (8.6 mg) was incubated at room tempera-

ture, with 1.3 units of ribonuclease T2 in 4.0 ml of 20 mM bis-Tris pH 7.0, for 60 min. The concentration of enzyme required to achieve 99% hydrolysis of the substrate in 60 min was determined as previously described. The enzyme was removed by ultrafiltration in a Centricon 10 microconcentrator which had been prewashed with sterile water. After ultrafiltration of the flavin solution, the membrane was washed with an equal volume of sterile water. FADPwas stored at - 70°C or lyophilized at 0.05 bar for 44 h and stored at - 70°C. Continuous ultrafiitration reactor. A 10 ml ultrafiltration reactor

was also designed for the synthesis of FADP. FADcP (1.0 ~M) in 20 mM bis-Tris, pH 7.0, was passed through the ultrafilter containing 0.125 units m l - 1 of ribonuclease T2 at 1 ml rain-1. This residence time was calculated to achieve 99% conversion of the cyclic phosphate to F A D P in the ultrafiltrate. 12 The reactor was operated for 90 min. The ultrafiltrate was sampled intermittently and the percentage of substrate utilized was determined by measuring the concentration of FADP as described above.

Table 1 Concentration of FAD present at each stage of synthesis of FADP

Stage of synthesis HPLC (single pass) Batch treatment with apoglucose oxidase Continuous treatment with apoglucose oxidase Batch ribonuclease T2 reactor Continuous ribonuclease T 2 reactor Immobilized ribonuclease T2 reactor

FAD in FADcP (%) yield (%)

FAD in FADP (%) yield (%)

0.001 6O 0.0003 > 90 0.00003 > 90 0.0006 >90 0.0001 > 90 0.0001 50

E n z y m e M i c r o b . T e c h n o l . , 1994, v o l . 16, A p r i l

2.83

Papers 100

after ultrafiltration of the bulk of the material, but the increased process time would result in a greater probability of contamination with FAD or bacterial growth.

80'

C o n t i n u o u s ultrafiltration reactor. The residence time and

concentration of enzyme in the reactor were calculated to achieve 99% conversion of the substrate to FADP, based on the calculated Km and Kp with adenosine-2',3'-cyclic phosphate 5'-monophosphate. In practice, total recovery of F A D P was 90% _+ 7.4 over an operation time of 90 rain.

% F A D 60' bound

C o n t i n u o u s c o l u m n reactor. The Km and Kp for the immobilized enzyme could not be determined due to the limitations of diffusion of the substrate and/or product in the microenvironment of the matrix to which the ribonuclease T2 had been attached. Figure 3 shows that even when operated at low flow rates, only 50-60% of the substrate was hydrolyzed. The immobilized enzyme must therefore have an altered Km and/or a much reduced Kp under the conditions employed.

40'

20

0

I

0

. . . .

)

5

. . . .

i

. . . .

10

!

Discussion

. . . .

15

20

Ultrafiltrate(mL) Figure 2 Specific removal of binding of FAD in a continuous ultrafiltration reactor containing 50 nM apo-glucose oxidase. The reactor was operated at different flow rates: ( • ) 0.32 ml min T; (V1)0.355 ml r a i n - l ; ( I ) 0.55 ml rain -~. The concentration of FAD in the ultrafiltrate was determined by the extent of reconstitution of apo-oaminoacid oxidase

The enzyme amplification assay that we have developed requires substrate containing very low levels of FAD to reduce any background signal in the assay. The concentration of contaminating FAD must be less than 0.12 nM (0.0006%) of the concentration of FADP used in the assay. Previously, FADP has been synthesized via an intermediate cyclic phosphate, flavin adenine dinucleotide-2',Ycyclic phosphate (FADcP), which can be purified by consecutive runs of reversed-phase chromatography on a 70

flow rate of 0.55 ml min - 1, 90% of the FAD was still removed from solution when apo-glucose oxidase was present in excess. This is equivalent to a second-order rate constant of 9.9 x 106 M - i min - I at pH 6.0 and 20°C, and is in good agreement with that reported elsewhere. 14,15 At lower flow rates, 99% of the FAD is removed from solution. Assuming that a position of equilibrium has been attained in the reactor, i.e., the residence time is longer than is required, a dissociation constant can be determined for glucose oxidase and FAD. This is equivalent to 5 x 10- 12M. This is lower than that reported by Okuda et al. 14 However, a dissociation constant of less than 10-10M was expected from the studies of Swoboda.15 When FADcP was passed through a similar reactor at flow rates of 0.53 and 0.65 ml min -1, 99.1% + 0.4 and 98.5% +_ 0.5 of the contaminating FAD was removed from solution, respectively. A comparison of this procedure and the batch reactor is shown in Table 1.

60

50

8 40

30

20 0

i

i

i

10

20

30

40

Production o f F A D P from FADcP Batch processing employing soluble enzyme. Reproducible high-quality grade material has been synthesized using this process (Table 1). Typically, yields of material are greater than 90%. That portion of the flavin which was not recovered was retained in the ultrafiltration membrane. This should be recoverable with more extensive washing

284

EnzymeMicrob. Technol., 1994, vol. 16, April

Volume of eluate (mL) Figure 3 Production of FADP employing a column of immobilized ribonuclease T2. FADcP (0.4 raM) was passed through the column at ( 0 ) 0.1 ml min --1 and (I-q) 0.2 ml min -1. FADP was quantitated by measuring the concentration of FAD in solution after complete dephosphorylation of FADP

Enzymic synthesis of FADP: M. Fisher et al. Partisil O D S - 2 M a g n u m 20 H P L C column to eliminate byproducts of the synthesis, which inhibit the amplification assay, and reduce the contaminating concentration of F A D to less than 0.0004%, though yields are low (approximately 60% per run). New methodologies have been developed that result in the synthesis of the primary substrate, F A D P , conforming to the stringent requirements of the amplification assay. F A D P can be prepared in greater yield if the second pass on the H P L C is replaced with an incubation reactor with apoglucose oxidase fromAspergillus niger. W h e t h e r operated in a batch or a continuous mode, the amount of apo-glucose oxidase used to reduce the concentration of F A D in solution is identical. The apo-glucose oxidase must be replenished in the ultrafiltration reactor once it has been saturated with F A D . The advantages of a continuous operation over a batch operation are therefore minimal, since both operations involve the same down time. However, a continuous reactor in series with one containing ribonuclease T2 would be preferable to two separate unit operations. Ribonuclease T2 can also be used in a batch or continuous operation to produce F A D P . A continuous ultrafiltration reactor has been run over a period o f 90 rain, achieving an average conversion of 90% u n d e r the operating conditions. The actual Km ofribonuclease T2 is therefore possibly larger and/or the Kp is slightly smaller than that calculated for adenosine-2',T-cyclic 5'-phosphate monophosphate. W h e n not in use, the continuous column reactor was stored in 30 mM sodium acetate, p H 5.0, containing 1 mM MgC12, 1 mM MnCI2, 1 mM CaC12, and 30% glycerol at room temperature. After more than 20 h of operation, no decrease in the efficiency o f the column was detectable. Although the increased stability of the enzyme afforded by immobilization simplifies its recovery and reuse, the ribonuclease T2 immobilized by this procedure cannot be employed in the production of F A D P because of the steric hindrance of the matrix. Alternative methodologies for the immobilization o f the e n z y m e may prove to be m o r e applicable.

Abbreviations aDAAO apo-D-aminoacid oxidase AMP adenosine m o n o p h o s p h a t e bis-Tris bis(2-hydroxyethyl)iminotris(hydroxymethyl)methane bis-Tris1,3-bis[tris(hydroxymethyl)Propane methylamino]propane DAAO D-aminoacid oxidase (holoenzyme) DHSA 3,5-dichloro-2-hydroxybenzene sulphonic acid. 4-AP 4-aminoantipyrine EIA enzyme immunoassay. FAD flavin adenine dinucleotide

FADcP FADP GMP hrp Km

Kp

NADP ppb ppm Tris-HC1

flavin adenine dinucleotide-2',3'-cyclic phosphate flavin adenine dinucleotide-3'-phosphate good manufacturing practice. horseradish peroxidase. Michaelis rate constant. Dissociation rate constant of the enzyme-product complex. nicotinamide adenine dinucleotide phosphate (oxidized form) parts per billion parts per million tris(hydroxymethyl)aminomethane

References 1 Johannsson, A., Ellis, D. H., Bates, D. L., Plumb, A. M. and Stanley, C. J. Enzyme amplification immunoassays:Detection limit of one hundredth of an attomole. Z Immuno. Meth. 1986,87, 7-11 2 Bates,D. L Enzyme amplification in diagnostics. Tibtech. 1987, 5, 204-209 3 Harbron, S., Eggelte, H. J., Fisher, M. and Rabin, B. R, Amplified assay of alkaline phosphatase using FAD-phosphate as substrate. Anal. Biochem. 1992,206, 119-124 4 Obzansky,D. M., Rabin, B. R., Simons, D. M., Tseng, S. Y., Severino, D. M., Eggelte, H. J., Fisher, M., Harbron S., Stout, R. W. and Di Paolo, M. J. Sensitive,colorimetricenzymeamplificationcascade for determination of alkaline phosphatase and application of the method to an immunoassay of thyrotropin. Clin. Chem. 1991, 37, 1513-1518 5 Rabin,B. R., Harbron, S., Eggelte, H. J. and Hollaway, M. R. An amplification assay for hydrolase enzymes. 1989,UK patent GB-B2240845 6 Rabin,B. R., Hollaway,M. R. and Taylorson,C. J. Enzymicmethod of detecting analytes and novel substrates therefor. 1991,US patent US 5057412 7 Harbron, S., Fisher, M. and Rabin, B. R. Large scale preparation and purification of apo-D-aminoacidoxidase for use in novel amplification assays.Biotechn. Techniques 1991,6, 55-60 8 Massey,V. and Mendelsohn, L. D. Immobilizedglucoseoxidase and D-amino acid oxidase: A convenient method for the purification of fiavin adenine dinucleotide and its analogs.Anal. Biochem. 1979,95, 156-159 9 Janski,A. M. and Oleson, A. E. NADP-agarose: An affinityadsorbent for tobacco extracellular nuclease and other nucleases. Anal. Biochem. 1976,71, 471-480 10 Reddy,L. G. and Shankar, V. Preparation and properties of RNase T2 immobilizedon concanavalinA-Sepharose.Appl. Biochem. Biotechnol. 1989, 22, 237-246 11 Fiske,C. H. and Subbarow, Y. The colorimetric determination of phosphorus. J. Biol. Chem. 1925,66, 375-400 12 Cornish-Bowden, A. Fundamentals of Enzyme Kinetics. Butterworth, 1979,pp. 16-38 13 Morris,D. L and Buckler, R. T. Colorimetric immunoassayusing flavin adenine dinucleotide as label. Methods Enzymol. 1983, 92, 413-425 14 Okuda,J., Nagamine, J. and Yagi, K. Exchange of free and bound coenzymeof flavinenzymes studied with [14C]FAD. Biochim. Biophys. Acta. 1979,566, 245-252 15 Swoboda,B. E. P. The relationship between molecular conformation and the binding of flavin-adenine dinucleotide in glucose oxidase. Biochim. Biophys. Acta. 1969, 175, 365-379

Enzyme Microb. Technol., 1994, vol. 16, April

285