- Email: [email protected]
Amplified colorimetric assay of alkaline phosphatase using riboflavin 4′-phosphate: A simple method for measuring riboflavin and riboflavin 5′-phosphate
ANALYTICAL
BIOCHEMISTRY
198,47-51
(1991)
Amplified Calorimetric Assay of Alkaline Phosphatase Using Riboflavin 4’-Phosphate: A Simple Method for Measuring Riboflavin and Riboflavin 5’-Phosphate Stuart Harbron,l
Henny J. Eggelte, and Brian R. Rabin
London Biotechnology Ltd., Department of Biochemistry and Molecular Biology, University College London, Gower Street, London WClE 6BT, England
Received
April
5, 1991
Alkaline phosphatase hydrolyzes riboflavin 4’-phosphate to produce riboflavin. This is converted to riboflavin S’phosphate, using riboflavin kinase, which reconstitutes apoglycolate oxidase to give hologlycolate oxidase. This enzyme catalyzes the oxidation of glycolate with simultaneous production of hydrogen peroxide which is detected via the formation of a colored product through the action of peroxidase. The system allows the detection of 4 amol after a 2-h incubation. o ISSI Academic
Press,
Inc.
Alkaline phosphatase (EC 3.1.3.1) is widely used as a label in immunoassays and in gene-probe-based assays. It can be assayed luminometrically (l-5), colorimetritally, or amperometrically (6). Other systems have relied on enzyme amplification (7). In the present study the principles of the prosthetogen-based enzyme cascades described by Rabin et al. (8) have been applied to the measurement of alkaline phosphatase. This amplification system relies on the production of a prosthetic group or precursor thereof by hydrolysis of a primary substrate. The prosthetic group combines with a suitable apoenzyme, leading to the formation of the corresponding holoenzyme which then catalyzes an appropriate reaction for determination. This amplification system is capable of great sensitivity since every tumover of the primary enzyme creates another enzyme. To work successfully it is important that the precursor itself does not reconstitute the holoenzyme. In the present study, 4’FMN2 is hydrolyzed to produce riboflavin,
’ To whom correspondence should be addressed. ’ Abbreviations used: 4’FMN, riboflavin 4’-phosphate; flavin 5’-phosphate. 0003-269’7/91$3.00 Copyright 0 1991 by Academic Press, All rights of reproduction in any form
5’F’MN,
ribo-
which is converted to 5’FMN by riboflavin kinase (EC 2.7.1.26). The 5’FMN so formed reconstitutes apoglycolate oxidase (EC 1.1.3.15) to produce the active holoenzyme which oxidizes glycolate. The hydrogen peroxide produced during this oxidation reaction can be measured using peroxidase (EC 1.11.1.7) (9), provided all components are free of contaminating catalase (see Scheme 1). MATERIALS 5’FMN-agarose, glycolate oxidase from sugar beet or spinach, 4-aminoantipyrine, and 3,5-dichloro-2-hydroxybenzenesulfonic acid were obtained from Sigma. Calf intestinal mucosa alkaline phosphatase, enzyme immunoassay grade, and horseradish peroxidase, grade I, were obtained from Boehringer. Chromatographic equipment and media were obtained from PharmaciaLKB and spectrophotometric measurements were made using a Shimadzu UV-240. METHODS
Preparation
of 4’FMN
This was prepared as described by Harbron et al. (10) and typically contained less than 0.01% 5’FMN.
Preparation
of Apoglycolute
Oxiuhe
Apoglycolate oxidase, from sugar beet or spinach, was prepared from the holoenzyme by acid precipitation (11,12) or by dialysis against potassium bromide in a fashion analogous to that described by Massey and Curti (13) for the preparation of apo-D-amino acid oxidase. Apoglycolate oxidase from sugar beet was also prepared by gel filtration at room temperature. Typically 2 mg of enzyme was dissolved in 0.2 ml of 200 mM 47
Inc. reserved.
48
HARBRON,
EGGELTE,
AND RABIN
Assay of 5’FMN
4’FMN-Riboflavin
+ Phosphate ~,bo~avin
ATP
kinass
ADP 1 SFMN Holqlycolate odase b Glycolate
n
l--
Apglycdde oxicbse
A standard curve for the estimation of 5’FMN was obtained as follows: 50 mM Tris-HCl buffer, pH 8.3,44 mU apoglycolate oxidase, and 0.02 to 200 PM S’FMN in a total volume of 0.05 ml was incubated for 1 h at room temperature. This was then added to 0.95 ml of 50 mM Tris-HCl buffer, pH 8.3, containing 3.47 mM phenylhydrazine and 5.26 mM glycolic acid, and the (linear) rate of change of absorbance at 324 nm was measured.
Glyoxylate
Assay of Alkaline
El
N-ka$yryl)-3chloro5wlphonate-p-benzoquinow monolmine
SCHEME
1.
Enzyme cascade for the assay of alkaline phospha-
tase.
Tris-HCl buffer, pH 8.0, containing 1.0 M potassium bromide and 40% (v/v) glycerol, and the solution applied to a column (1 cm diam X 30 cm) of Superose 12 equilibrated with the same buffer. Traces of hologlycolate oxidase remaining after these treatments were removed by affinity chromatography on 5’FMN-agarose at 4°C. Up to 12 mg of apoenzyme in 2 ml of 5 mM phosphate buffer, pH 8.0, was applied to the column (1 cm diam X 13 cm) equilibrated with the same buffer. The column was washed with 7 column vol of equilibration buffer followed by 5 column vol of 1 M potassium phosphate buffer, pH 8.0. Under these conditions the holoenzyme passed through the column to which apoenzyme remained bound. Apoglycolate oxidase was eluted with 0.1 M phosphate buffer, pH 8.0, containing 1.0 M potassium bromide. It was stored as a suspension in 3.2 M ammonium sulfate, pH 8.0, at 4°C. Assay of Glycolate Ox&se Glycolate oxidase was measured in a mixture containing 50 mM Tris-HCl, pH 8.0, 5.0 mM glycolic acid, 0.24 mM 5’FMN, and 3.3 mM phenylhydrazine in a total volume of 1.0 ml. The change in absorbance was monitored at 324 nm and the activity of the enzyme determined (molar extinction coefficient of the glyoxylatephenylhydrazone is 17,000). One unit of activity is the amount of enzyme which produces 1 pmol of glyoxylatephenylhydrazone in 1 min. Purijication of Riboflavin Kinuse This was purified from yeast as described by Harbron et al. (10).
Phosphatase
The primary incubation mixture comprised: 50 mM Tris-HCl buffer, pH 8.3, containing 1.0 mM MgSO,, 0.1 mM ZnSO,, 0.2 mM 4’FMN, and alkaline phosphatase in a total volume of 0.05 ml. After incubation for 1 h at room temperature the riboflavin produced was measured by adding to the primary mixture 0.05 ml of the following mixture: 100 mM potassium phosphate buffer, pH 8.3, 2.1 PM sugar beet apoglycolate oxidase, 10 mM glycolic acid, 8 mM 3,5-dichloro-2-hydroxybenzenesulfonic acid, 0.8 mM 4-aminoantipyrine, 0.006 mg peroxidase, 10 mM ATP, 1.0 mM MgSO,, 0.1 mM ZnSO,, and 1.5 mU riboflavin kinase. This secondary incubation mixture (0.1 ml) was further incubated for 1 h at room temperature and the absorbance at 520 nm measured. The concentration of 10 PM alkaline phosphatase, from which subsequent dilutions were made, was checked by assaying it in 10 mM Tris-HCl buffer, pH 8.0, containing 0.1 mM nitrophenylphosphate at 25°C. An absorbance change per minute of 0.00137 in a l-cm cuvette was obtained.
RESULTS
Preparation
of Apoglycolate
Oxidase
Attempts to prepare the apoenzyme as described by Zelitch and Ochoa (11) or Frigerio and Harbury (12) yielded variable results. Both depend on the release of 5’FMN from the apoenzyme as the latter begins to precipitate at low pH in a solution of ammonium sulfate. An assessment of the precise behavior of the enzyme under these conditions was made and is shown in Fig. 1. Below pH 4.5, the amount of enzyme recovered declined markedly, although the proportion of it found in the apo form increased. Above pH 4.5 more of the enzyme remained in solution. The maximum yield of apoenzyme was achieved at about pH 4.3, somewhat higher than that used in the previous studies described above. At this pH, over 80% of the enzyme precipitated was in the apo form. In view of the relatively low yield of apoenzyme (about 45%) under these conditions, the use of a high concentration of potassium bromide (1 M) at pH 8.0 to
AMPLIFIED
COLORIMETRIC
ASSAY
FOR
ALKALINE
0
FIG. 1. Preparation of apoglycolate oxidase from spinach by acid precipitation at 4°C. To 0.5 U of glycolate oxidase dissolved in 0.5 ml of water, 35% saturated with ammonium sulfate, was added an appropriate volume of 0.1 M sulfuric acid, 35% saturated with ammonium sulfate, to give the pH value indicated. The mixture was immediately centrifuged and the pellet resuspended in 0.5 ml of water. (0) Total enzyme recovered (apo plus holo forms) in the pellet and supernatant; (V) total enzyme recovered (apo plus holo forms) in the pellet; (A) enzyme recovered in the apo form in the pellet, all given as a percentage of the initial amount (0.5 U) of enzyme. (x) Proportion of enzyme recovered in the pellet in the apo form as a percentage of the total enzyme (apo plus holo forms) recovered in the pellet.
cause the dissociation of the YFMN was investigated. Dialysis under these conditions resulted in the complete loss of activity over the 2-day dialysis period. Chromatography on Superose 12 (Fig. 2) typically yielded more than 90% apoenzyme with a residual holoenzyme activity of less than 1%. Subsequent chromatography on 5’FMN-agarose (Fig. 3) gave at least 75% recovery and a residual holoenzyme activity of less than 0.01%. When stored at 4°C as a suspension in 3.2 M ammonium sulfate, only 5% of the activity was lost over a period of 10 weeks. Chromatography on Superose 12 and 5’FMN-agarose removed the small amount of catalase present in the commercial holoenzyme preparation and permitted the use of the peroxidase-linked assay of alkaline phosphatase described above.
49
PHOSPHATASE
50
xx) Elutim Vdun~ imll
FIG. 3. Chromatography of 24 U of spinach on 5’FMN-agarose as described potassium phosphate buffer started at buffer containing 1 M potassium bromide were collected and assayed.
The validity of the methodology was tested by assaying different-sized aliquots of a mixture of apoenzyme (0.88 U/ml) and 5’FMN (2 PM). Since the same specific activity and linear progress curves were always obtained, the reconstituted holoenzyme does not dissociate under the assay conditions. This reaction rate measured quantitatively the amount of holoenzyme present in the incubation mixture. Figure 4 shows the enzymic activity produced by the apoenzymes from sugar beet and spinach as a function of the concentration of added 5’FMN. The apoenzyme from sugar beet clearly has a higher affinity than the apoenzyme from spinach for 5’FMN. Assuming that 5’FMN is bound independently by each subunit of the apoenzyme, these data indicate that the Kd)s for 5’FMN of the enzymes from sugar beet and spinach are 1.2 f.&M and 41 FM respectively. These data can also be analyzed to give the concentration of reconstitutable subunits in the incubation mixture and the reaction rate when the apoenzyme has been saturated with 5’FMN. From these, subunit turnover numbers for the holoenzymes from sugar beet and
, 0.02
d 02
I 2 b’MN1
Vdume (ml)
FIG. 2. Chromatography of 5 U of glycolate oxidase from beet on Superose 12 as described in the text. 0.5 ml fractions collected and assayed.
sugar were
apoglycolate oxidase from in the text. Wash with 1 M A, elution of enzyme with started at B, 5 ml fractions
Assay of 5’FMN
I
Elutm
150
20
200
fl
FIG. 4. Activity of glycolate oxidase reconstituted by the addition of 5’FMN to apoenzyme from sugar beet (0) and spinach (A) to give the concentration indicated under conditions described in the text. Rates plotted after subtraction of background (0.00213 for sugar beet; 0.002833 for spinach).
50
HARBRON,
EGGELTE,
spinach of 8.3 and 4.6 s-l, respectively, were calculated; 1 PM of reconstitutable apoenzyme subunits is thus given by 497 or 276 mu/ml of sugar beet or spinach apoenzymes, respectively. Inclusion of 10 mM ATP, 1.0 mM MgSO,, 0.1 mM ZnSO,, and 0.75 mU of riboflavin kinase in the incubation mixture allowed riboflavin to be assayed. Curves similar to those obtained for 5’FMN (Fig. 4) were produced. Using the apoenzyme from sugar beet, 8 nM riboflavin could be detected, corresponding to 0.4 pmol in 0.05 ml. This calorimetric assay of riboflavin thus has a similar sensitivity to a recently described luminometric assay (10). Apoenzyme from both sources could not be activated by 4’FMN or by riboflavin, indicating that the presence and position of the phosphate group on the ribityl side chain is of great importance for the reconstitution process. The reconstitution by 5’FMN of both apoenzymes was unaffected by 4’FMN or riboflavin at concentrations up to 0.4 mM.
Kinetic Properties of Alkaline Phosphatase with 4’FMN
In an assay of alkaline phosphatase with varying amounts of 4’FMN, the apparent Km of alkaline phosphatase for 4’FMN was 0.29 mM and the value determined for &at was 35 s-l. The corresponding values for 5’FMN under similar conditions were 18 pM and 244 s-l, respectively (10). The hydrolysis of the phosphate ester of the secondary alcohol by alkaline phosphatase is thus much less efficient than the equivalent primary and the value of It,,lK,,, for 4’FMN is reduced by a factor of 100 compared with that of 5’FMN.
AND RABIN
0014
1
".',.I
. . . . . ..I 10
(...."I
100
amol Alkaline
loo0
Rwsphatase
FIG. 5. Assay of alkaline phosphatase using the method described in the text. Absorbance plotted after subtraction of the background reading of 0.347 (SE = 0.014, n = 3). The regression line gives a value of 0.967 for r2.
Although the kcat of alkaline phosphatase increases with pH, the K,,, is reduced as the pH is raised (14). Given the need to use a moderately low concentration of 4’FMN (to prevent its interference with the reconstitution of the apoenzyme, to reduce the amount of contaminating 5’FMN and riboflavin present in the assay, and to reduce any contribution to the absorbance at 520 nm by the flavin), together with the instability of 4’FMN and 5’FMN at higher pH’s, especially in the presence of light, a somewhat lower pH (8.3) than is customarily used was chosen for the assay. The other components of the assay (apoglycolate oxidase, 4’FMN, riboflavin kinase, and horseradish peroxidase) are not limiting to the assay under the conditions described.
DISCUSSION
Calorimetric Assay of Alkaline Phosphutase
The observation that 4’FMN and riboflavin do not interfere with the reconstitution of apoglycolate oxidase by 5’FMN enables alkaline phosphatase to be assayed according to the method given above. Phosphate buffer and a relatively high concentration of ATP are incorporated in the second, color-generating step to prevent the hydrolysis by alkaline phosphatase of 5’FMN produced from riboflavin through the action of riboflavin kinase. Under these conditions the performance of the assay is illustrated in Fig. 5. The detection limit is the amount of alkaline phosphatase needed to give an absorbance that is greater than the blank by three standard deviations of the blank. Since the data in Fig. 5 are given with the blank already subtracted, this corresponds to the amount of alkaline phosphatase required to give a net absorbance of 0.1. After a total of 2 h incubation, 4 amol of alkaline phosphatase can be thus detected. Most of the background is produced from the small amount (0.01%) of 5’FMN contaminating the 4’FMN substrate.
The high specificity of apoglycolate oxidase for its prosthetic group, 5’FMN, has enabled the construction of an enzyme cascade for the measurement of alkaline phosphatase. In the sequence of multiple reactions described in this cascade there are five distinct possible types of interference: (i) Presence of catalase in the apoenzyme, which would reduce the absorbance change produced by the peroxidase reaction. This contaminant has been removed by chromatography on Superose 12 and 5’FMNagarose. (ii) Reconstitution of active holoenzyme by 4’FMN, which would give a high background, or (iii) Prevention of the reconstitution of active holoenzyme by 5’FMN in the presence of 4’FMN, which would reduce the signal. The high specificity of apoglycolate oxidase for its prosthetic group means that neither of these occurs. (iv) Hydrolysis of 5’FMN produced in the second stage of the reaction by alkaline phosphatase carried over from the first stage, which would again lead to a
AMPLIFIED
COLORIMETRIC
ASSAY
reduction in the signal obtained. Alkaline phosphatase is inhibited by phosphate buffer and excess ATP. (v) Presence of 51;‘MN and riboflavin in the 4’FMN substrate, which would lead to a high background. 4’FMN contains less than 0.01% of these. Although the assay is not as sensitive as that described by Bronstein et al. (2,3), who were able to detect 0.001 amol of alkaline phosphatase in solution or 0.07 amol of Hepatitis B core antigen DNA after a 2-h camera exposure at room temperature, it compares well with a number of other recently described luminometric assays. For example, 20 amol of alkaline phosphatase was detected using an amplified luminometric assay (lo), and Geiger and Miska (1) could detect 11 amol of kallikrein in a sandwich immunoassay using a luminometer, although the incubation times of 30 min used in these assays were less than that used in the present investigation. The assay is, however, much more sensitive than the widely used calorimetric assay of alkaline phosphatase based on the hydrolysis of nitrophenylphosphate which can detect 250 amol (6). The assay is simple to use, involving only two incubation steps. It can detect 4 amol of alkaline phosphatase and the components show a high degree of stability. The formation of a colored product absorbing at wavelengths remote from the flavin absorbance maximum means that the assay is well-suited for use in microtitreplate-based immunoassays.
FOR
ALKALINE
51
PHOSPHATASE
REFERENCES 1. Geiger, R., and Miska, 25,31-38. 2. Bronstein,
W. (1987)
I., Edwards,
B., and
J. Clin. Chem. Clin. Biochem. Voyta,
J. C. (1988)
J.
Biolum.
Cfwmilum. 2,186. 3. Bronstein, 180,95-98.
I., Voyta,
4. Rabin, B. R., Hollaway, (1988) Riboflavin-Linked A-2197951,
J. C., and Edwards,
B. (1989)
M. R., Harbron, S., and Benson, Assay Procedures and Materials,
5. Rabin, B. R., Hollaway, M. R., Harbron, Eggelte, H. J. (1989) Riboflavin-Linked Materials, PCT GB89100546.
C. H. (1985)
S. M. GB-
S., Benson, S. M., and Assay Procedures and
6. Thompson, R. Q., Barone, G. C., Hal&l, W. R. (1991) Anal. Biochem. 192,90-95. I. Self,
And. Biochem.
B., and
Heineman,
J. Zmmunol. Methods 76,389-393.
8. Rabin, B. R., Hollaway, M. R., and Taylorson, mic Methods of Detecting Analytes and Novel for, GB-B-2156518. 9. Fossati, P., Prencipe, L., and Berti, G. (1980) 227-231.
C. J. (1988) EnzySubstrates ThereJ. Clin. C&cm.
26,
10. Harbron, S., Eggelte, H. J., Benson, S. M., and Rabin, B. R. (1991), Submitted for publication. 11. Zelitch, I., and Ochoa, S. (1953) J. Biol. Chem. 201, 707-718. 12. Frigerio, N. A., and Harbury, H. A. (1958) J. Bid. Chm. 231, 135-157. 13. Massey,
V., and Curti,
B. (1966)
J. Biol. Chem. 241,3417-3423.
14. Chappelet-Tordo, D., Fosset, M., Iwatsubo, M., Lazdunski, M. (1974) Biochemistry 13,1788-1795.
Gache,
C., and
BIOCHEMISTRY
198,47-51
(1991)
Amplified Calorimetric Assay of Alkaline Phosphatase Using Riboflavin 4’-Phosphate: A Simple Method for Measuring Riboflavin and Riboflavin 5’-Phosphate Stuart Harbron,l
Henny J. Eggelte, and Brian R. Rabin
London Biotechnology Ltd., Department of Biochemistry and Molecular Biology, University College London, Gower Street, London WClE 6BT, England
Received
April
5, 1991
Alkaline phosphatase hydrolyzes riboflavin 4’-phosphate to produce riboflavin. This is converted to riboflavin S’phosphate, using riboflavin kinase, which reconstitutes apoglycolate oxidase to give hologlycolate oxidase. This enzyme catalyzes the oxidation of glycolate with simultaneous production of hydrogen peroxide which is detected via the formation of a colored product through the action of peroxidase. The system allows the detection of 4 amol after a 2-h incubation. o ISSI Academic
Press,
Inc.
Alkaline phosphatase (EC 3.1.3.1) is widely used as a label in immunoassays and in gene-probe-based assays. It can be assayed luminometrically (l-5), colorimetritally, or amperometrically (6). Other systems have relied on enzyme amplification (7). In the present study the principles of the prosthetogen-based enzyme cascades described by Rabin et al. (8) have been applied to the measurement of alkaline phosphatase. This amplification system relies on the production of a prosthetic group or precursor thereof by hydrolysis of a primary substrate. The prosthetic group combines with a suitable apoenzyme, leading to the formation of the corresponding holoenzyme which then catalyzes an appropriate reaction for determination. This amplification system is capable of great sensitivity since every tumover of the primary enzyme creates another enzyme. To work successfully it is important that the precursor itself does not reconstitute the holoenzyme. In the present study, 4’FMN2 is hydrolyzed to produce riboflavin,
’ To whom correspondence should be addressed. ’ Abbreviations used: 4’FMN, riboflavin 4’-phosphate; flavin 5’-phosphate. 0003-269’7/91$3.00 Copyright 0 1991 by Academic Press, All rights of reproduction in any form
5’F’MN,
ribo-
which is converted to 5’FMN by riboflavin kinase (EC 2.7.1.26). The 5’FMN so formed reconstitutes apoglycolate oxidase (EC 1.1.3.15) to produce the active holoenzyme which oxidizes glycolate. The hydrogen peroxide produced during this oxidation reaction can be measured using peroxidase (EC 1.11.1.7) (9), provided all components are free of contaminating catalase (see Scheme 1). MATERIALS 5’FMN-agarose, glycolate oxidase from sugar beet or spinach, 4-aminoantipyrine, and 3,5-dichloro-2-hydroxybenzenesulfonic acid were obtained from Sigma. Calf intestinal mucosa alkaline phosphatase, enzyme immunoassay grade, and horseradish peroxidase, grade I, were obtained from Boehringer. Chromatographic equipment and media were obtained from PharmaciaLKB and spectrophotometric measurements were made using a Shimadzu UV-240. METHODS
Preparation
of 4’FMN
This was prepared as described by Harbron et al. (10) and typically contained less than 0.01% 5’FMN.
Preparation
of Apoglycolute
Oxiuhe
Apoglycolate oxidase, from sugar beet or spinach, was prepared from the holoenzyme by acid precipitation (11,12) or by dialysis against potassium bromide in a fashion analogous to that described by Massey and Curti (13) for the preparation of apo-D-amino acid oxidase. Apoglycolate oxidase from sugar beet was also prepared by gel filtration at room temperature. Typically 2 mg of enzyme was dissolved in 0.2 ml of 200 mM 47
Inc. reserved.
48
HARBRON,
EGGELTE,
AND RABIN
Assay of 5’FMN
4’FMN-Riboflavin
+ Phosphate ~,bo~avin
ATP
kinass
ADP 1 SFMN Holqlycolate odase b Glycolate
n
l--
Apglycdde oxicbse
A standard curve for the estimation of 5’FMN was obtained as follows: 50 mM Tris-HCl buffer, pH 8.3,44 mU apoglycolate oxidase, and 0.02 to 200 PM S’FMN in a total volume of 0.05 ml was incubated for 1 h at room temperature. This was then added to 0.95 ml of 50 mM Tris-HCl buffer, pH 8.3, containing 3.47 mM phenylhydrazine and 5.26 mM glycolic acid, and the (linear) rate of change of absorbance at 324 nm was measured.
Glyoxylate
Assay of Alkaline
El
N-ka$yryl)-3chloro5wlphonate-p-benzoquinow monolmine
SCHEME
1.
Enzyme cascade for the assay of alkaline phospha-
tase.
Tris-HCl buffer, pH 8.0, containing 1.0 M potassium bromide and 40% (v/v) glycerol, and the solution applied to a column (1 cm diam X 30 cm) of Superose 12 equilibrated with the same buffer. Traces of hologlycolate oxidase remaining after these treatments were removed by affinity chromatography on 5’FMN-agarose at 4°C. Up to 12 mg of apoenzyme in 2 ml of 5 mM phosphate buffer, pH 8.0, was applied to the column (1 cm diam X 13 cm) equilibrated with the same buffer. The column was washed with 7 column vol of equilibration buffer followed by 5 column vol of 1 M potassium phosphate buffer, pH 8.0. Under these conditions the holoenzyme passed through the column to which apoenzyme remained bound. Apoglycolate oxidase was eluted with 0.1 M phosphate buffer, pH 8.0, containing 1.0 M potassium bromide. It was stored as a suspension in 3.2 M ammonium sulfate, pH 8.0, at 4°C. Assay of Glycolate Ox&se Glycolate oxidase was measured in a mixture containing 50 mM Tris-HCl, pH 8.0, 5.0 mM glycolic acid, 0.24 mM 5’FMN, and 3.3 mM phenylhydrazine in a total volume of 1.0 ml. The change in absorbance was monitored at 324 nm and the activity of the enzyme determined (molar extinction coefficient of the glyoxylatephenylhydrazone is 17,000). One unit of activity is the amount of enzyme which produces 1 pmol of glyoxylatephenylhydrazone in 1 min. Purijication of Riboflavin Kinuse This was purified from yeast as described by Harbron et al. (10).
Phosphatase
The primary incubation mixture comprised: 50 mM Tris-HCl buffer, pH 8.3, containing 1.0 mM MgSO,, 0.1 mM ZnSO,, 0.2 mM 4’FMN, and alkaline phosphatase in a total volume of 0.05 ml. After incubation for 1 h at room temperature the riboflavin produced was measured by adding to the primary mixture 0.05 ml of the following mixture: 100 mM potassium phosphate buffer, pH 8.3, 2.1 PM sugar beet apoglycolate oxidase, 10 mM glycolic acid, 8 mM 3,5-dichloro-2-hydroxybenzenesulfonic acid, 0.8 mM 4-aminoantipyrine, 0.006 mg peroxidase, 10 mM ATP, 1.0 mM MgSO,, 0.1 mM ZnSO,, and 1.5 mU riboflavin kinase. This secondary incubation mixture (0.1 ml) was further incubated for 1 h at room temperature and the absorbance at 520 nm measured. The concentration of 10 PM alkaline phosphatase, from which subsequent dilutions were made, was checked by assaying it in 10 mM Tris-HCl buffer, pH 8.0, containing 0.1 mM nitrophenylphosphate at 25°C. An absorbance change per minute of 0.00137 in a l-cm cuvette was obtained.
RESULTS
Preparation
of Apoglycolate
Oxidase
Attempts to prepare the apoenzyme as described by Zelitch and Ochoa (11) or Frigerio and Harbury (12) yielded variable results. Both depend on the release of 5’FMN from the apoenzyme as the latter begins to precipitate at low pH in a solution of ammonium sulfate. An assessment of the precise behavior of the enzyme under these conditions was made and is shown in Fig. 1. Below pH 4.5, the amount of enzyme recovered declined markedly, although the proportion of it found in the apo form increased. Above pH 4.5 more of the enzyme remained in solution. The maximum yield of apoenzyme was achieved at about pH 4.3, somewhat higher than that used in the previous studies described above. At this pH, over 80% of the enzyme precipitated was in the apo form. In view of the relatively low yield of apoenzyme (about 45%) under these conditions, the use of a high concentration of potassium bromide (1 M) at pH 8.0 to
AMPLIFIED
COLORIMETRIC
ASSAY
FOR
ALKALINE
0
FIG. 1. Preparation of apoglycolate oxidase from spinach by acid precipitation at 4°C. To 0.5 U of glycolate oxidase dissolved in 0.5 ml of water, 35% saturated with ammonium sulfate, was added an appropriate volume of 0.1 M sulfuric acid, 35% saturated with ammonium sulfate, to give the pH value indicated. The mixture was immediately centrifuged and the pellet resuspended in 0.5 ml of water. (0) Total enzyme recovered (apo plus holo forms) in the pellet and supernatant; (V) total enzyme recovered (apo plus holo forms) in the pellet; (A) enzyme recovered in the apo form in the pellet, all given as a percentage of the initial amount (0.5 U) of enzyme. (x) Proportion of enzyme recovered in the pellet in the apo form as a percentage of the total enzyme (apo plus holo forms) recovered in the pellet.
cause the dissociation of the YFMN was investigated. Dialysis under these conditions resulted in the complete loss of activity over the 2-day dialysis period. Chromatography on Superose 12 (Fig. 2) typically yielded more than 90% apoenzyme with a residual holoenzyme activity of less than 1%. Subsequent chromatography on 5’FMN-agarose (Fig. 3) gave at least 75% recovery and a residual holoenzyme activity of less than 0.01%. When stored at 4°C as a suspension in 3.2 M ammonium sulfate, only 5% of the activity was lost over a period of 10 weeks. Chromatography on Superose 12 and 5’FMN-agarose removed the small amount of catalase present in the commercial holoenzyme preparation and permitted the use of the peroxidase-linked assay of alkaline phosphatase described above.
49
PHOSPHATASE
50
xx) Elutim Vdun~ imll
FIG. 3. Chromatography of 24 U of spinach on 5’FMN-agarose as described potassium phosphate buffer started at buffer containing 1 M potassium bromide were collected and assayed.
The validity of the methodology was tested by assaying different-sized aliquots of a mixture of apoenzyme (0.88 U/ml) and 5’FMN (2 PM). Since the same specific activity and linear progress curves were always obtained, the reconstituted holoenzyme does not dissociate under the assay conditions. This reaction rate measured quantitatively the amount of holoenzyme present in the incubation mixture. Figure 4 shows the enzymic activity produced by the apoenzymes from sugar beet and spinach as a function of the concentration of added 5’FMN. The apoenzyme from sugar beet clearly has a higher affinity than the apoenzyme from spinach for 5’FMN. Assuming that 5’FMN is bound independently by each subunit of the apoenzyme, these data indicate that the Kd)s for 5’FMN of the enzymes from sugar beet and spinach are 1.2 f.&M and 41 FM respectively. These data can also be analyzed to give the concentration of reconstitutable subunits in the incubation mixture and the reaction rate when the apoenzyme has been saturated with 5’FMN. From these, subunit turnover numbers for the holoenzymes from sugar beet and
, 0.02
d 02
I 2 b’MN1
Vdume (ml)
FIG. 2. Chromatography of 5 U of glycolate oxidase from beet on Superose 12 as described in the text. 0.5 ml fractions collected and assayed.
sugar were
apoglycolate oxidase from in the text. Wash with 1 M A, elution of enzyme with started at B, 5 ml fractions
Assay of 5’FMN
I
Elutm
150
20
200
fl
FIG. 4. Activity of glycolate oxidase reconstituted by the addition of 5’FMN to apoenzyme from sugar beet (0) and spinach (A) to give the concentration indicated under conditions described in the text. Rates plotted after subtraction of background (0.00213 for sugar beet; 0.002833 for spinach).
50
HARBRON,
EGGELTE,
spinach of 8.3 and 4.6 s-l, respectively, were calculated; 1 PM of reconstitutable apoenzyme subunits is thus given by 497 or 276 mu/ml of sugar beet or spinach apoenzymes, respectively. Inclusion of 10 mM ATP, 1.0 mM MgSO,, 0.1 mM ZnSO,, and 0.75 mU of riboflavin kinase in the incubation mixture allowed riboflavin to be assayed. Curves similar to those obtained for 5’FMN (Fig. 4) were produced. Using the apoenzyme from sugar beet, 8 nM riboflavin could be detected, corresponding to 0.4 pmol in 0.05 ml. This calorimetric assay of riboflavin thus has a similar sensitivity to a recently described luminometric assay (10). Apoenzyme from both sources could not be activated by 4’FMN or by riboflavin, indicating that the presence and position of the phosphate group on the ribityl side chain is of great importance for the reconstitution process. The reconstitution by 5’FMN of both apoenzymes was unaffected by 4’FMN or riboflavin at concentrations up to 0.4 mM.
Kinetic Properties of Alkaline Phosphatase with 4’FMN
In an assay of alkaline phosphatase with varying amounts of 4’FMN, the apparent Km of alkaline phosphatase for 4’FMN was 0.29 mM and the value determined for &at was 35 s-l. The corresponding values for 5’FMN under similar conditions were 18 pM and 244 s-l, respectively (10). The hydrolysis of the phosphate ester of the secondary alcohol by alkaline phosphatase is thus much less efficient than the equivalent primary and the value of It,,lK,,, for 4’FMN is reduced by a factor of 100 compared with that of 5’FMN.
AND RABIN
0014
1
".',.I
. . . . . ..I 10
(...."I
100
amol Alkaline
loo0
Rwsphatase
FIG. 5. Assay of alkaline phosphatase using the method described in the text. Absorbance plotted after subtraction of the background reading of 0.347 (SE = 0.014, n = 3). The regression line gives a value of 0.967 for r2.
Although the kcat of alkaline phosphatase increases with pH, the K,,, is reduced as the pH is raised (14). Given the need to use a moderately low concentration of 4’FMN (to prevent its interference with the reconstitution of the apoenzyme, to reduce the amount of contaminating 5’FMN and riboflavin present in the assay, and to reduce any contribution to the absorbance at 520 nm by the flavin), together with the instability of 4’FMN and 5’FMN at higher pH’s, especially in the presence of light, a somewhat lower pH (8.3) than is customarily used was chosen for the assay. The other components of the assay (apoglycolate oxidase, 4’FMN, riboflavin kinase, and horseradish peroxidase) are not limiting to the assay under the conditions described.
DISCUSSION
Calorimetric Assay of Alkaline Phosphutase
The observation that 4’FMN and riboflavin do not interfere with the reconstitution of apoglycolate oxidase by 5’FMN enables alkaline phosphatase to be assayed according to the method given above. Phosphate buffer and a relatively high concentration of ATP are incorporated in the second, color-generating step to prevent the hydrolysis by alkaline phosphatase of 5’FMN produced from riboflavin through the action of riboflavin kinase. Under these conditions the performance of the assay is illustrated in Fig. 5. The detection limit is the amount of alkaline phosphatase needed to give an absorbance that is greater than the blank by three standard deviations of the blank. Since the data in Fig. 5 are given with the blank already subtracted, this corresponds to the amount of alkaline phosphatase required to give a net absorbance of 0.1. After a total of 2 h incubation, 4 amol of alkaline phosphatase can be thus detected. Most of the background is produced from the small amount (0.01%) of 5’FMN contaminating the 4’FMN substrate.
The high specificity of apoglycolate oxidase for its prosthetic group, 5’FMN, has enabled the construction of an enzyme cascade for the measurement of alkaline phosphatase. In the sequence of multiple reactions described in this cascade there are five distinct possible types of interference: (i) Presence of catalase in the apoenzyme, which would reduce the absorbance change produced by the peroxidase reaction. This contaminant has been removed by chromatography on Superose 12 and 5’FMNagarose. (ii) Reconstitution of active holoenzyme by 4’FMN, which would give a high background, or (iii) Prevention of the reconstitution of active holoenzyme by 5’FMN in the presence of 4’FMN, which would reduce the signal. The high specificity of apoglycolate oxidase for its prosthetic group means that neither of these occurs. (iv) Hydrolysis of 5’FMN produced in the second stage of the reaction by alkaline phosphatase carried over from the first stage, which would again lead to a
AMPLIFIED
COLORIMETRIC
ASSAY
reduction in the signal obtained. Alkaline phosphatase is inhibited by phosphate buffer and excess ATP. (v) Presence of 51;‘MN and riboflavin in the 4’FMN substrate, which would lead to a high background. 4’FMN contains less than 0.01% of these. Although the assay is not as sensitive as that described by Bronstein et al. (2,3), who were able to detect 0.001 amol of alkaline phosphatase in solution or 0.07 amol of Hepatitis B core antigen DNA after a 2-h camera exposure at room temperature, it compares well with a number of other recently described luminometric assays. For example, 20 amol of alkaline phosphatase was detected using an amplified luminometric assay (lo), and Geiger and Miska (1) could detect 11 amol of kallikrein in a sandwich immunoassay using a luminometer, although the incubation times of 30 min used in these assays were less than that used in the present investigation. The assay is, however, much more sensitive than the widely used calorimetric assay of alkaline phosphatase based on the hydrolysis of nitrophenylphosphate which can detect 250 amol (6). The assay is simple to use, involving only two incubation steps. It can detect 4 amol of alkaline phosphatase and the components show a high degree of stability. The formation of a colored product absorbing at wavelengths remote from the flavin absorbance maximum means that the assay is well-suited for use in microtitreplate-based immunoassays.
FOR
ALKALINE
51
PHOSPHATASE
REFERENCES 1. Geiger, R., and Miska, 25,31-38. 2. Bronstein,
W. (1987)
I., Edwards,
B., and
J. Clin. Chem. Clin. Biochem. Voyta,
J. C. (1988)
J.
Biolum.
Cfwmilum. 2,186. 3. Bronstein, 180,95-98.
I., Voyta,
4. Rabin, B. R., Hollaway, (1988) Riboflavin-Linked A-2197951,
J. C., and Edwards,
B. (1989)
M. R., Harbron, S., and Benson, Assay Procedures and Materials,
5. Rabin, B. R., Hollaway, M. R., Harbron, Eggelte, H. J. (1989) Riboflavin-Linked Materials, PCT GB89100546.
C. H. (1985)
S. M. GB-
S., Benson, S. M., and Assay Procedures and
6. Thompson, R. Q., Barone, G. C., Hal&l, W. R. (1991) Anal. Biochem. 192,90-95. I. Self,
And. Biochem.
B., and
Heineman,
J. Zmmunol. Methods 76,389-393.
8. Rabin, B. R., Hollaway, M. R., and Taylorson, mic Methods of Detecting Analytes and Novel for, GB-B-2156518. 9. Fossati, P., Prencipe, L., and Berti, G. (1980) 227-231.
C. J. (1988) EnzySubstrates ThereJ. Clin. C&cm.
26,
10. Harbron, S., Eggelte, H. J., Benson, S. M., and Rabin, B. R. (1991), Submitted for publication. 11. Zelitch, I., and Ochoa, S. (1953) J. Biol. Chem. 201, 707-718. 12. Frigerio, N. A., and Harbury, H. A. (1958) J. Bid. Chm. 231, 135-157. 13. Massey,
V., and Curti,
B. (1966)
J. Biol. Chem. 241,3417-3423.
14. Chappelet-Tordo, D., Fosset, M., Iwatsubo, M., Lazdunski, M. (1974) Biochemistry 13,1788-1795.
Gache,
C., and