Glucose determination in samples taken by microdialysis by peroxidase-catalyzed luminol chemiluminescence

ANALYTICAL

BIOCHEMISTRY

192,237-242

(1991)

Glucose Determination in Samples Taken by Microdialysis by Peroxidase-Catalyzed Luminol Chemiluminescence Birgitta

Ntislund,

Peter

Arner,

Jan Bolinder,

Lars

Clinical Research Centre and Department of Medicine, Karolinska Huddinge University Hospital F61, S-141 86 Huddinge, Sweden

Received

August

and Arne

Lundin

Institute,

20, 1990

An automatic, luminometric assay of glucose in samples of the extracellular water space obtained by microdialysis is described. The assay involves oxidation by glucose oxidase (EC 1.1.3.4) and mutarotation of glucose by aldose mutarotase (EC 5.1.3.3). The H,Oz formed is subsequently determined in a reaction catalyzed by horseradish peroxidase (EC 1.11.1.7) using luminol as electron donor. The assay is linear between 0.01 and 1 nmol in the cuvette. The detection limit, defined as 3 standard deviations of the reagent blank, was 0.008 pmobliter in the cuvette. A complete oxidation of glucose is obtained within 4 min and 25 samples are automatically assayed within 75 min. Addition of microdialysate sample obtained from human adipose tissue in vivo did not interfere with the standard curves. Glucose added to microdialysate resulted in a complete recovery compared to a H,O, standard. Analytical interference from different factors was investigated. No interference was observed up to the following concentrations: 5 rmollliter epinephrine, 1 pmol/liter norepinephrine, 100 pmollliter insulin, 500 rmollliter pyruvate, 50 mmollliter lactate, and 1 pmol/liter ascorbate. The glucose values with the present method correlated strongly (r = 0.984) with values obtained using a routine method involving glucose oxidase and peroxidase. 0 1991 Academic Press, Inc.

Several methods for analysis of glucose in human blood and in urine have been developed (for review see Ref. (1)). The most specific methods rely on the use of glucose oxidase (GOD)’ to measure the H,O, formed by peroxidase-catalyzed reactions, resulting in colored ’ Abbreviations peroxidase; DPTA,

Hallander,

used: GOD, glucose oxidase; diethylenetriaminepentaacetic

0003-2697/91 $3.00 Copyright 0 1991 by Academic Press, Inc. All rights of reproduction in any form reserved.

HRP, acid.

horseradish

products (for review seeRef. (2)). However, these methods are insensitive, with a lower limit in the range 5-30 nmol in the sample (see Table 1). Biological samples often contain substances which interfere with the assay. Thus dilution of the sample is required, resulting in an increased demand for sensitive methods. This demand is particularly strong when the level of glucose is low or when it is difficult to obtain large amounts of human biological samples, for example, those obtained by microdialysis. This technique can be used to sample the extracellular water space of different tissues and organs such as adipose tissue (13) and brain (14). In microdialysis the recovery of the metabolite to be investigated is often incomplete (5-75%, depending on the type of probe and perfusion speed), and the sample volume is in the microliter range. Earlier reports show that measurement of low concentrations of glucose can be achieved using luminometric methods to determine H,O,, i.e., the product from the GOD reaction. However, the existing methods are complicated and tedious, since the GOD and the luminol reaction are performed in separate media or phases. In the present investigation we have developed a sensitive and simple assay in which three enzymatic reactions, involving GOD, mutarotase, and horseradish peroxidase (HRP), are performed in the same buffer system. The method is specially adapted to the assay of samples obtained by microdialysis of the extracellular water space of adipose tissue. MATERIALS

AND

METHODS

Chemicals

The following chemicals were obtained from Sigma Chemical Co. (St. Louis, MO): glucose oxidase (EC 1.1.3.4) from Aspergillus niger (Type V; 325-344 U/mg) 237

238

N&LUND

40 0

2

4

5

8

10

12

14

16

18

ET

I 20

AL.

water acidified with phosphoric acid (0.5 mmol/liter), and the concentration of hydrogen peroxide was determined by titration against potassium permanganate in the presence of sulfuric acid according to Kolthoff and Sandell (16). All other chemicals were of analytical grade and were used without further purification. Cuvettes were obtained from Boehringer Mannheim Scandinavia AB (Bromma, Sweden). The glucose standard and reagents for the routine methods used at the Department of Clinical Chemistry were obtained from Apoteksbolaget AB, Produktionsenheten, Sodersjukhuset (Stockholm, Sweden).

Time (ml)

FIG. 1. The effect of different concentrations of mutarotase and pH on the assay measured as chemiluminescence obtained from glucose and compared in percentage of the H,O, standard. The incubation system (200 ~1) was as described under Standard Procedure with the following final concentrations; 5 pmol/liter glucose and 40 U/ml GOD. One experiment was performed at pH 7.75 with 60 U/ml mutarotase (W), and another experiment at pH 8.0 without mutarotase (+), with 30 U/ml mutarotase (A), and with 60 U/ml (0) mutarotase.

in 0.1 mol/liter sodium acetate buffer (pH approx 4), Trizma base, DTPA (diethylenetriaminepentaacetic acid), dimethyl sulfoxide, epinephrine-bitartrate, norepinephrine-bitartrate, and horseradish peroxidase (EC 1.11.1.7, Type VI; 250-330 U/mg). The latter (10 mg of protein/mliter) was stored in double-distilled deionized water and propylene glycol (50% v/v) at -20°C. Insulin (bovine) was obtained from Novo BioLabs (Bagsvaerd, Denmark). Ringer’s solution (concentrations in mmol/liter; 148 Na+, 4 K+, 2.4 Ca’+, 156 Cl-) was purchased from Pharmacia (Uppsala, Sweden). Hydrogen peroxide and L(+)-ascorbic acid were obtained from Merck (Darmstadt, West Germany). Sodium pyruvate and sodium lactate of biopure grade were obtained from Apoteksbolaget AB (Gothenburg, Sweden). D-Glucose (AnalaR) was purchased from BDH Limited (Poole, England). Glucose standard solutions (0.1 mol/liter) were prepared in double-distilled deionized water and stored in portions of 200 ~1 at -20°C. The concentrations of the glucose standards were determined by the Department of Clinical Chemistry at the hospital using the hexokinase glucose-6-phosphate dehydrogenase method according to Schmidt (15). Luminol(5 - amino - 2,3 - dihydro - 1,4 - phthalazinedione) used in this study was obtained from BioThema AB (Dalaro, Sweden) and was stored as a 100 mmol/liter solution in dimethyl sulfoxide at +4”C. For the selection of luminol, hydrogen peroxide curves were analyzed using this luminol and luminols obtained from LKB Wallac (Turku, Finland) and Aldrich-Chemie (Steinheim, Germany). Mutarotase (EC 5.1.3.3), from hog kidney was centrifuged and dissolved (1 mg/ml) in double-distilled deionized water. The hydrogen peroxide (1 mol/ liter) was stored at +4”C in double-distilled deionized

Principle

of the Assay

In Step 1 cu-D-glucose is transferred to /?-D-glucose by mutarotase (reaction [la]). Concominantly, P-D-glucose is oxidized by GOD in the presence of oxygen to D-gluconic acid and H,O, (reaction [lb]). Subsequently, in Step 2, the H,O, produced participates in a luminol reaction catalyzed by horseradish peroxidase with the emission of light (reaction [2]). The two steps are automatically and sequentially performed in the luminometer, and the total amount of light integrated during 106 s is directly proportional to the initial amount of glucose in the assay. a-D-glucose

4E-s

/3-D-glucose

(36%) ,&D-glucose

+ 0, + H,O GOD_ D-gluconic

2H,O,

[la1

(64%)

+ luminol

[lb]

5

3-aminophthalic IQ324

acid + H,O,

acid + N, + 2H,O

+ light

[2]

y.i.Lm+z3.2 (Gkrmseti.) y= 0.98x+22.7 (Glucose sld. + mhdiilysate) I

10 4 10-Q

10-a

10-7

10-e

I 10-S

Conmntraliw (mo!lL)

FIG. 2. Double concentrations in solution (m) or 20 cuvettes containing crodialysate have standard.

logarithmic plots of glucose standard curves (final cuvette) obtained after addition of 20 rl of Ringer’s ~1 microdialysate sample (diluted 1:5 A) added to glucose standard. The values in mV * s of the mibeen subtracted from each individual value of the

LUMINOMETRIC

Standard

ASSAY

OF

GLUCOSE

IN

239

MICRODIALYSIS

Procedure

The luminometer (Luminometer 1251, Bio-Orbit, Turku, Finland) automatically assayed 25 samples within 75 min, including a preincubation time of 5 min to allow for equilibration to 25°C. The sensitivity of the luminometer was reduced by a factor of 16 using the special Service program (instructions can be obtained from the manufacturer or from the authors). The reaction procedure was incubation for 10 min with GOD and mutarotase followed by the peroxidase-luminol reaction for 2 min. In the luminometer the GOD reaction was started in a new cuvette every second minute. The peroxidase-luminol reaction was measured for 106 s starting 10 min after the addition of GOD. Step 1. Endpoint incubation was performed in cuvettes containing 170 ~1 of 50 mmol/liter Tris-phosphate buffer, pH 7.75, and mutarotase (added from the water solution immediately before starting the assays). Twenty microliters of sample (diluted 1:20 in Ringer’s solution), glucose or H,O, standard (0.5-50 pmol/liter diluted in Ringer’s solution), or Ringer’s solution was added (blanks). After cuvettes were transferred to the luminometer, 10 ~1 of GOD reagent (800 U/ml, diluted in double-distilled deionized water) was automatically added, resulting in the following concentrations (in the 200-~1 incubation mixture): 0.05-5 pmol/liter glucose, 40 U/ml GOD, 60 U/ml mutarotase in 50 mmol/liter Tris-phosphate buffer, pH 7.75. Step 2. After 10 min 790 ~1 of the luminol reagent (200 pmol/liter luminol, 50 mmol/liter DTPA in 50 mmol/liter Tris-phosphate buffer, pH 7.75), and then, to start the reaction, 10 ~1 of the HRP reagent (0.9 mg/ ml HRP in 50 mmol/liter Tris-phosphate buffer, pH 7.75) were automatically added with the final concentrations of the components in 1 ml: 160 pmol/liter luminol, 40 mmol/liter DTPA, 9 pg/ml HRP, in 50 mmol/ liter Tris-phosphate buffer, pH 7.75, and the corresponding concentration of 0.01 to 1 pmol/liter glucose. Data reduction. The light emission was integrated for 106 s and the values (in mV * s) were first corrected for individual instrument blanks (signal before addition of HRP reagent to each individual cuvette), and then for the reagent blank measured in a separate cuvette with Ringer’s solution as sample blank. The luminometer was connected to a Z-88 computer for data collection. Data were transferred to and calculated on a Macintosh computer using MacTerminal and Microsoft Excel programs. Comparison of the Present with a Routine Method The concentration dialysate of human

Method

of glucose in samples from microsubcutaneous adipose tissue during

0 10-7

10-e

10-S Concentratmn

10-d

10-a

10-z

I 10-l

(mol/L)

FIG. 3. The effect of different factors on the assay. Cuvettes containing 250 pmol glucose were supplied with increasing concentrations of factor. The factors studied were 20 ~1 of 0.6 pmol/liter to 5 mmol/liter epinephrine (a, Cl), 0.3 pmollliter to 5 mmol/liter norepinephrine (a, n ), 0.2 to 100 rmol/liter insulin (a, A), 3 rmollliter to 50 mmol/liter sodium pyruvate (b, q ), 3 rmollliter to 50 mmol/liter sodium lactate (b, n ), and 0.14 rmol/liter to 0.57 mmol/liter (24 ng-0.1 mg/ml) sodium ascorbate (b, A). The stock solutions of the factors were prepared in Ringer’s solution and adjusted to pH 7.75 with sodium hydroxide. Dilutions were prepared in Ringer’s solution.

oral glucose loading was determined by assaying 20 ~1 of sample (diluted 1:20 in Ringer’s solution). These samples (60 ~1) were also analyzed by a routine method performed at the Department of Clinical Chemistry, using a calorimetric method as described by Barham and Trinder (17) including glucose oxidase, 4-aminophenazone, parahydroxybenzoate, and peroxidase. Microdialysis

Experiments

Detailed descriptions of the microdialysis device (14) and the experimental protocol (18) have been published previously. In brief, a microdialysis probe (10 X 0.5 mm; cutoff point M, 20,000) was inserted in the abdominal subcutaneous adipose tissue of nonobese healthy subjects and dialyzed with a CMA/lOO microinjection pump using Ringer’s solution. The dialysis probe and the micropump were obtained from Carnegie Medicin AB, Stockholm, Sweden. The perfusion speed was 5 pl/min. At this speed the in vitro recovery of glucose was 20%, which remained constant up to at least 20 mmol/liter of glucose (18). The subjects were investigatedin the morning after an overnight fast. After a 30- to 45-min equili-

240

N&LUND

ET

AL.

Glucose Standard

Time

(mm)

FIG. 4. Determination of glucose in microdialysate obtained from probes at 15 min intervals up to 165 min during oral glucose loading. Samples of 20 (diluted 1:20 in Ringer’s solution) and 60 al were analyzed by the present method and the routine method, respectively. (a) Results obtained by the present method. (b) The correlation between the present method and the routine method.

bration period 15-min dialysate fractions were collected. After two baseline samples were collected, a 75-g oral glucose load was given and the glucose level in the dialysate was followed for 165 min. Each individual was given a detailed description of the study and his or her consent was obtained. The study was approved by the Ethics Committee of the Karolinska Institute.

Statistical

Methods

Linear regression analysis and calculation of the coefficients of variation were performed according to Snedecor and Cochran (19). The linear regression analysis was performed after calculation of logarithms of light emission and concentration values. Thus a slope of 1 is required for a linear relation between nonlogarithmic values.

RESULTS

The GOD and Mutarotase

A linear relation between analyte and signal requires that the slope be unity in double logarithmic plots after subtraction of blanks. For this study a purified luminol from BioThema AB was selected. With this luminol a higher light intensity was obtained, and the hydrogen peroxide curves were more linear than those of the other luminol preparations investigated. The slope for five glucose standard curves was 0.996 _+ 0.0366 (mean f SD). The possible interference from microdialysate medium on the standard curves was studied. Twenty microliters of Ringer’s solution or 20 ~1 of microdialysate (originally containing 73 pmol/liter glucose) diluted 1:5 in Ringer’s solution was added to each concentration of standard. In these experiments the integrated millivolt second values for the sample of the microdialysate were subtracted before the slopes were calculated. The two curves are almost identical and no interference can be seen from the microdialysate (Fig. 2). Recovery In recovery experiments parallel cuvettes containing 20 ~1 of diluted (1:20) microdialysate samples with and without 20 ~1 of separate glucose standards (50 pmoll liter) were assayed in duplicate. H,O, standards treated in the same way were run in parallel. The recovery of separate glucose standards corresponding to 1 mmol/ liter in the original sample added to microdialysate was 99.9 f 2.03% (4 experiments) compared to that of H,O, standards. Addition of 20 ~1 of Ringer’s solution to the incubation system affected neither the glucose standard curve nor the recovery of glucose. The recovery of light from the H,O, standard was not reduced after 10 min of incubation with GOD and mutarotase, indicating no significant catalase activity in this assay system. It is recommended that this experiment be repeated for every new batch of enzymes. Precision The precision of the luminometric step was determined by assaying 10 replicates of 20 ~1 of microdialysate sample (originally containing 250 and 350 gmol/ liter glucose diluted 1:20). Resulting CV values were 2.03% for 12.5 pmollliter and 0.74% for 17.5 pmol/liter. Effects

Reaction

Incubation of 5 pmol/liter glucose at pH 7.75 with 40 U/ml GOD and 60 U/ml mutarotase resulted in a complete oxidation of glucose within 4 min as compared to a H,O, standard. At pH 8.0 or in the absence of mutarotase, longer incubation times were required (Fig. 1).

Curves

of Different

Factors

on the Assay

The influence of different factors on the overall assay was investigated using a glucose concentration of 0.25 pmol/liter in cuvette. Figure 3 shows the dose-response curves for epinephrine, norepinephrine, insulin, pyruvate, lactate, and ascorbate. The concentration values correspond to the concentration in the undiluted sample. A decrease in light emission of at least 50% was

LUMINOMETRIC

ASSAY

OF

GLUCOSE

TABLE

Comparison Method Spectrophotometric Spectrophotometric Calorimetric

Range

(nmol)”

33-504, endpoint up to 28, kinetic 22-666

Spectrophotometric Amperometric/FIA

26-396, endpoint up to 139, kinetic 16.6-150 4.8-240

Luminometric

(lo-8-1o-4

Luminometric Luminometric

1.11-111 0.2-20

Luminometric

25-5400

Luminometric

0.01-l

M)

a The linear range for the glucose the sample volume is unknown, the * The time for final measurement ’ Ti-PAPS; a mixture of titanium d Cu(phen)F; a complex obtained

of the Present

241

MICRODIALYSIS

1

of Methods for Glucose Determination Enzymes

and chemicals

involved

Hexokinase, glucose-6-phosphate dehydrogenase, ATP, NADP Glucose dehydrogenase, NAD, mutarotase GOD, peroxidase, 4-aminophenazone, phenol GOD, Ti-PAPS reagent’ GOD, mutarotase, catalase (immobilized enzymes) GOD (immobilized), luminol, We(CN1, GOD, luminol, K,Fe(CN), GOD, luminol, Cu(phen)F d (a microporous membrane flow cell) GOD, luminol (an inverted micellar system) GOD, mutarotase, luminol, peroxidase

Incubation

Assay with a Routine Method

The concentrations of glucose in samples obtained from microdialysis of human subcutaneous adipose tissue were determined using 1~1 (20 ~1 of sample diluted 20 fold) of microdialysate (Fig. 4a). The results were compared to values obtained by the routine assay used at the Department of Clinical Chemistry. The correlation coefficient between the two methods was 0.984 (Fig. 4b). DISCUSSION

In this investigation we have developed a sensitive, simple, specific, and automatic method for glucose determination in microdialysis samples. The method is based on the same principles as those recently worked out for free fatty acid determination (20); i.e., H,O,

time *

Reference

5-30 min, endpoint 67 s, kinetic 15 min

(3)

35-60 min, endpoint 67 s, kinetic 15 and 10 min 12 samples/60 min

(5,6)

4 samples/l0

(4)

(7)

(8)

min

(9)

5 min 20-30 s

(10) (11)

15 min

(12)

25 samples/75

concentration in the assay reported or recalculated using the molecular glucose concentration range in the assay is given in parentheses. in manual methods is not included. (IV) and 2-((5-bromopyridyl)azo)5-(N-propyl-N-sulfopropylamino)phenol. by mixing copper nitrate and l,lO-phenanthroline.

obtained at 200 pmol/liter epinephrine, 10 pmol/liter norepinephrine, 10 mmol/liter pyruvate, and 100 pmol/ liter (0.018 mg/ml) ascorbate. However, for epinephrine and ascorbate, a preceding increase of emitted light of 10 and 20%, respectively, at a concentration of 10 pmol/ liter was observed. Concentrations up to 5 pmol/liter epinephrine, 1 pmol/liter norepinephrine, 500 hmol/ liter pyruvate, and 1 pmol/liter ascorbate did not interfere with the assay. Lactate and insulin at concentrations up to 50 mmol/liter and 100 pmol/liter, respectively, did not interfere with the assay. Comparison

IN

weight

min

of 180.16

Present

for glucose.

method

When

formed in an oxidase reaction is assayed by the HRPcatalyzed luminol reaction. The recovery and the linearity of glucose through all the steps in the assay were related to the final product, i.e., H,O,. A 100% recovery and a linear range covering two orders of magnitude were obtained. The assay involves only three coupled enzymatic steps, and the reagent cost is approximately 15 cents per assay. The method allows 25 samples to be automatically assayed within 75 min, and it is possible to measure 100-150 samples per day. Several methods involving the specific enzyme GOD (Table 1) have recently been published. In spectrophotometric methods allowing an error of 5%, the measurement range will generally be limited to 0.05-1.5 absorbancy units (21). This corresponds to a useful range with a factor of 30 between upper and lower limit. The spectrophotometric methods are less sensitive (Table 1). Previous luminometric methods are sensitive but they all share the major disadvantages that the GOD and the luminol reactions are performed at different pH (9-11) or even in different phases (12). The present method based on HRP-catalyzed luminol reaction utilizes the fact that all three enzymatic reactions involved can be performed at the same pH. This makes the assay easier to perform. In order to avoid possible differences in the nonenzymatic mutarotation rate between the sample and the standard, which is both ion (22,23) and pH dependent (24), mutarotase was included in the assay mixture.

242

N&LUND

This resulted in a complete oxidation of (Y- and p-glucase and a 100% recovery in the assay. After a survey of different commercial luminol preparations, a purified luminol preparation was chosen. With this luminol the light emission was doubled. Compared to the pH in the previous study (20) the pH was reduced from pH 8.0 to pH 7.75. At the lower pH, nonspecific side reactions in the luminol system are reduced. Furthermore this pH is closer to the pH optimum of GOD from A. niger (25) and for hog kidney mutarotase (24). Since addition of microdialysate did not affect the standard curve, it was possible to analyze samples using an ordinary glucose standard curve. The effects of some common metabolites and hormones on the assay were investigated. In general, interference with the assay was observed only at concentrations above those that may appear in microdialysate samples. Samples containing high concentrations of ascorbate should, however, be diluted or treated with ascorbate oxidase to a concentration below 1 pmol/liter in the 203.41sample added in the first step of the assay. Inhibition of the peroxidase reaction by ascorbic acid and epinephrine has been demonstrated previously (26-28). The correlation between the present method and the routine method for determination of glucose in microdialysate was excellent. However, for the routine method 60 ~1 of sample was used compared to 1~1 in the present method. Using microdialysate samples obviates the need for protein precipitation or enzyme inactivation. High-molecular-weight components do not pass through the membrane, which has a cutoff point at a molecular weight of 20,000. Since only l-2 ~1 of samples is needed, possible interference in the assay by components in the sample is minimized. The sensitivity of the luminescent assay makes it possible to follow rapid biological fluctuations by the microdialysis technique. A unique advantage of using microdialysis is that the extracellular space of the investigated tissue can be locally exposed to pharmacologically active substances, which are delivered through the microdialysis probe. ACKNOWLEDGMENTS Thanks are due to Mrs. Eva Sjijlin for excellent technical assistance, to the Department of Clinical Chemistry at the Huddinge University Hospital for performing the glucose analyses, and to Boehringer Mannheim GmbH for providing the mutarotase for this study. This investigation was supported by grants from the Swedish Medical Research Council (MRF 19X-01034), the Swedish Diabetes Association, the Nordic Insulin Foundation, the Karolinska Institute, the Osterman and Stohne Foundations, and the Swedish Sports Research Council.

ET

AL.

REFERENCES Pileggi, V. J., and Szustkiewicz, C. P. (1974) in Clinical Chemistry Principles and Technics (Henry, R. J., Cannon, D. C., and Winkelman, J. W., Eds.), pp. 1265-1325, Harper & Row, Maryland. 2. Free,

A. H. (1963)

Aduan.

Clin.

Chem.

6,67-96.

3. Kunst, A., Draeger, B., and Zeigenhorn, J. (1984) Enzymatic Analysis (Bergmeyer, U., Bergmeyer, M., Eds.), pp. 163-172, Verlag Chemie, Weinheim. 4. Vormbrock, R. (1984) meyer, U., Bergmeyer, lag Chemie, Weinheim.

in Methods of Enzymatic J., and Grassl, M., Eds.),

Analysis (Bergpp. 172-178, Ver-

5. Kunst, A., Draeger, B., and Zeigenhorn, J. (1984) Enzymatic Analysis (Bergmeyer, U., Bergmeyer, M., Eds.), pp. 178-185, Verlag Chemie, Weinheim. 6. Trinder,

P. (1969)

7. Matsubara, (1985) Anal.

Ann.

Clin.

Biochem.

in Methods of J., and Grassl,

in Methods of J., and Grassl,

6, 24-27.

C., Kudo, K., Kawashita, Chem. 57,1107-1109.

T.,

8. Matsumoto, K., Kamikado, H., Matsubara, (1988) Anal. Chem. 60, 147-151. 9. Bostick, D. T., and Hercules, D. M. (1975) 452.

and

Takamura,

K.

H., and Osajima, Anal.

Chem.

Y.

47,447-

10. Auses, J. P., Cook, S. L., and Maloy, J. T. (1975) Anal. Chem. 47, 244-24s. 11. Pilosof, D., and Nieman, T. A. (1982) Anal. Chem. 54,1698-1701. 12. Igarashi,

S., and Hinze,

W. L. (1988)

Anal.

Chem.

13. Arner, P., Bolinder, J., Eliasson, U. (1988) Amer. J. Physiol. 255 E742.

A., Lundin, (Endocrinol.

14. Tossman, 128.9-14. 15. Schmidt,

U. (1986)

U.,

Ungerstedt,

F. H. (1961)

16. Kolthoff, I. M., tative Inorganic 17. Barham,

and

Klin.

D., and Trinder,

P. (1972)

19. Snedecor, G. W., and Cochran, ods, 7th ed., Iowa State Univ. 20. Niislund, B. M. A., Bernstrom, (1989) J. Biolumin. Chemilumin.

24. Li, L.-K,

Analyst

1244-1247.

97,

142-145.

U., and Arner,

P. (1989)

W. G. (1980) in Statistical Press, Ames, IA. K., Lundin, 3, 115-124.

T., and Tomimura,

C. L. M., Wade, 59,2245-2247. (1965)

Stand.

A., and

MethArner,

P.

and Sandell, E. B. (1963) in Textbook of QuantiAnalysis, pp. 613-641, Macmillan, New York.

J., Taguchi, 22. Okuda, Bull. 35,4332-4337. 23. Stults, Chem.

39,

Physiol.

and Sandell, E. B. (1963) in Textbook of QuantiAnalysis, pp. 564-578, Macmillan, New York.

18. Bolinder, J., Hagstrom, E., Ungerstedt, Stand. J. Clin. Lab. Invest. 49, 465-474.

21. Kolthoff, I. M., tative Inorganic

Actu

Wochenschr.

60,446-450.

A., and Ungerstedt, Metabol. S), E737-

Arch.

A. P., and

B&hem.

A. (1987) Crouch,

Biophys.

Chem.

Pharm.

S. R. (1987)

Anal.

110. 156-162.

25. Underkofler, L. A. (1958) in Proceedings of the International Symposium on Enzyme Chemistry, Tokyo-Kyoto, 1957 (Ichihara, K., Ed.), pp. 486-490, Maruzen, Tokyo. 26. Sharp, P. (1972) Clin. Chim. Acta 40, 115-120. G. A., and Price, C. P. (1983) Clin. Chem. 29, 181027. Maguire, 1812. 28. Naganna, B., Rajamma, Chim. A& 17,219-221.

M.,

and Vasudeva

Rao,

K. (1967)

Clin.