in urine of patients with cystinuria. J Biol Chem 236, PC51-PC53. (1961). ImmunoenzymometricAssay for Insulin InvolvingColumn Chromatography.
22. Kredich NM, Hershileld MS, Falletta JM, et al. Effect of 2’deoxycoformycin on homocysteine metabolism in acute lymphoblastic leukemia. Clin Res 29, 541A (1981). Abstract. 23. Isles TE, Jocetyn PC. The reaction of protein thiol groups with some disulphides. Biochem J 88, 84-88 (1963). 24. Lorber A, Chang CC, Masuoka D, Meacham J. Effect of thiols in biological systems on protein sulfydryl content. Biochem Pharmacol 19, 1551-1560 (1970). 25. Malloy MH, Rassin DK, Gaull GE. A method for measurement
of free and bound plasma cyst(e)ine. Anal Biochem
113, 407-415 (1981). 26. Stein WH, Moore S. The free amino acids of human blood plasma. J Biol Chem 211, 915-926 (1954). 27. Smolin LA, Benevenga NJ. The use of cyst(e)ine in the removal of protein-bound homocysteine. Am J Clin Nutr 39,730-737(1984). 28. Frimpter GW. The dulsulfide of L-cysteine and L-homocysteine in urine of patients with cystinuria. J Biol Chem 236, PC51-PC53 (1961).
CLIN. CHEM. 31/4, 628-630 (1985)
ImmunoenzymometricAssay for Insulin InvolvingColumn Chromatography and Insulin Immobilized on Sepharose Ryohel
Yamamoto, Shigeki Kimura, Shigeko Hattorl, Akira Matsuura, and Tetsuo Hayakawa1
This practical assay for measuring insulin involves use of a 4 x 8 mm chromatographic column. Serum samples are incubated at 30 #{176}C for 1 h with p-D-galactosidase-labeled
antibodyto insulin,then passed throughthe O.1-mLcolumn containing insulin immobilized on Sepharose 4B. After the column is washed to remove the bound label, the buffer in the column is replaced with a solution of o-nitrophenyl-p-ogalactoside. The column is then incubated at 30 #{176}C for 1 h, the enzyme reaction is stopped by washing the column with sodium carbonate solution, and the absorbance of the eluate is measured at 420 nm. Results obtained by this method were compared with those by a radioimmunoassay and a solid-phase enzyme immunoassay. Our recently developed enzyme immunoassay, which includes a covalent chromatographic-separation method (1), is useful for competitive immunoassays (1-3) and immunoassays for antibodies (1, 4) without interference from serum components (5). Now we have improved the column-separation method, making it more practicable and simple, and have applied this technique to assays for thyroxin (6), triiodothyronine (6), and secretory immunoglobulin A (7). We describe here an enzyme immunoassay for insulin, in which we use the improved column-separation method, antibody labeled with enzyme, and antigen immobilized on Sepharose 4B.
Materials and Methods Materials We use de-ionized water throughout. G. This was sodium phosphate buffer (10 mmolfL, pH 7) containing 0.3 mol of NaCL, 1 mmol of MgCl2, 1 g of NaN3, 1 g of bovine serum albumin (Cohn Fraction V; Armour Pharmaceutical Co., Chicago, IL), and 5 g of digested gelatin per liter. The digested gelatin was prepared by Buffer
Department of Research and Development, Amano Pharmaceutical Co., Kunotsubo, Nishiharu, Nishikasugai, Aichi 481, Japan. ‘Second Department of Internal Medicine, Nagoya University School of Medicine, Showa-Ku, Nagoya 466, Japan. Received November 29, 1984; accepted January 18, 1985.
628 CLINICALCHEMISTRY,Vol.31, No. 4, 1985
treating gelatin (Difco Laboratories, Detroit, MI) with a protease (Protease Ti; Amano Pharmaceutical Co., Nagoya, Japan) (8). Antibody. Guinea pig antiserum to porcine insulin was obtained from Medical and Biological Laboratories, Nagoya, Japan. The IgG fractions were isolated from antiserum by precipitation with (NH4)2S04, dialysis, and chromatography on diethylaminoethyl-cellulose (5). Antibody-/3-D-galactosidase conjugate. F(ab’)2 fragments of anti-insulin antibody, prepared by digesting (anti-insulin) IgG fractions with pepsin (EC 3.4.23.1), were reduced with 2-mercaptoethylamine and coupled to f-n-galactosidase (EC 3.2.1.23, from Escherichia coli; Boehringer Mannheim, Mannheim, F.R.G.) by use of N,N’-o-phenylenedimaleimide (9). The amounts of the conjugate were expressed in terms of units of enzyme activity, and 1 unit (U) of activity was defined as that which hydrolyzed 1 mol of o-nitrophenyl-i>galactoside per minute under the conditions described below. Insulin immobilized on Sepharose. We mixed 1 mg of crystalline insulin (Sigma Chemical Co., St. Louis, MO), dissolved in 80 mL of sodium carbonate buffer (0.1 mol/L, pH 9), with 20 mL of CNBr-activated Sepharose 4B (Pharmacia Fine Chemicals, Uppsala, Sweden), and stirred this at 4#{176}C overnight. The insulin immobilized on Sepharose was then washed with Tris HC1 buffer (0.1 mol/L, pH 8) and stored at 4#{176}C. itandard serum. A porcine insulin (“Actrapid”; Novo Industri AJS, Copenhagen, Denmark) was diluted with human serum that had been freed of insulin by passage through a column containing anti-insulin antibody immobilized on Sepharose (100 mL of serum was passed through a 25-mL column). The anti-insulin immobilized on Sepharose was prepared as described previously (6). Comparison method. Using our enzyme immunoassay method, we measured our standard serum and that in a radioimmunoassay kit (Insulin “Eiken” RIA Kit from Eiken Chemical Co., Tokyo, Japan), then calculated the value of our standard serum from the standard curve obtained with the kit’s standard serum. The value for our standard serum was 97.3% as great as that for the standard serum specified with the radioimmunoassay kit.
Methods Immunoassay procedure. We mixed 0.1 mL of standard serum or serum sample with 0.5 mL of the antibody--Dgalactosidase conjugate (26 U/L, in buffer G) and incubated the mixture at 30#{176}C for 1 h. Of the reaction mixture, 0.5 mL was applied to a 4 x 8 mm column with a funnel-shaped buffer reservoir on the top (7), that had been packed with 0.1 mL of insulin immobilized on Sepharose and equilibrated with buffer G at a flow rate of 3 niL’h. After washing the column twice with 0.5 mL of buffer G, we stood the column on the top of a 15 x 100mm test tube, applied 0.25 mL of an 8.3 mmol/L solution of o-mtrophenyl-j3-n-galactosidase in buffer G to the column, and allowed the enzyme reaction to proceed at 30#{176}C for 1 h. The reaction was stopped by washing the column with 1 mL of an 80 mmolJL solution of sodium carbonate. We measured the absorbance of the eluate in the test tube, which contained o-nitrophenol, at 420 nm with a Shimazu Spectrophotometer Model VU-240. Measurement of /3-n-galactosidase activity. One milliliter of the enzyme solution (pH 7) was preincubated at 37 #{176}C for 5 mm, and the reaction was then started by adding 0.25 mL of a solution of o-nitrophenyl-f3-n-galactoside. After incubation of the mixture at 37 #{176}C for 20 mm, the reaction was stopped by adding 0.25 mL of a 1 mol/L solution of sodium carbonate. The absorbance was measured spectrophotometrically at 420 nm. The number of units of enzyme activity was calculated by utilizing the molar absorptivity of onitrophenol (21 300) at 420 nm (10) and the appropriate dilution factor.
Results To determine the available portions of antibody-f3-ngalactosidase conjugate, we applied 0.5 mL of buffer G containing various amounts of the conjugate to the same column packed with 0.1 mL of insulin immobilized on Sepharose, like that used in the immunoassay procedure, followed by washing the column with 1 mL of buffer G. The same operations were also performed with the column packed with plain Sepharose 4B for the control experiments. Then the enzyme activity in the eluate was measured. The results are shown in Table 1. The data indicate that about 10% of the conjugate can bind to insulin immobilized on Sepharose, and the column used in our immunoenzymometic assay can bind available portions of the conjugate in more than 260 milli-int. units of that. A standard curve for the assay for insulin is shown in Figure 1. The relation between f3-n-galactosidase activity (as reflected in absorbance) and hormone concentration in the standard is shown for the range 3 to 100 micro-int. units of insulin per tube. We tested the stability of the standard curve by assaying each standard serum in duplicate in 10 consecutive assays. The mean value on subtraction of the absorbance of 3 milli-units per liter of serum from that of 0
Table 1. Binding of lnsulin-.$-D-Galactosidase Conjugate to Insulin Immobilized on Sepharose 4B p-o-GalactosldaaeactivIty, mull-mt.units Applied to column A
2.6
Boundto columna B
Conjugate bound (B/A, %)
0.244
9.38
26 2.42 9.31 260 27.8 10.7 tmme amount of enzyme activity thatpassedthrough the column containing insulin immobilized on Sepharose was subtracted from the amount that passed through the column containing plain Sepharose 48.
0.8
E
0.6
0 4.’
w U C
,#{176}
I-
0
0.4
0.2 0 0
1
10
Insulin
100
(tilU/tube)
Fig. 1. Typical standard curve forinsulin JJ#{149} a micro-internationalunit(s)
milli-int. units per liter of serum was 0.029 (SD 0.003), and values of 0.031 (SD 0.005) and 0.064 (SD 0.010) were obtained for 10 milli-int. units per liter of serum from one with 3 milli-int. units/L, and 30 milli-int. units per liter of serum from one with 10 milli-int. unitsfL respectively. We concluded that the minimum detectable amount of insulin was 0.3 micro-mt. unit per assay in our enzyme immunoassay, because a #{163}442 of 0.029 can be measured adequately with a conventional spectrophotometer. In experiments on analytical recovery of insulin, we used seven specimens of human sera containing 10 to 50 miiiint. units of endogenous insulin per liter. We assayed each sample in duplicate, with and without added insulin (50200 milli-int. unitsfL) and calculated recovery from the standard curve. The results are shown in Table 2. We tested the precision of the assay by assaying three sera 15 times in one assay (within-run) or in duplicate in 10 consecutive assays (between-run). The results are shown in Table 3. In interference studies, we assayed five specimens of human sera containing 30 to 150 milli-int. units of endogenous insulin per liter, in duplicate, with and without added human hemoglobin (5 g/L) or bilirubin (200 mg/L). The results are shown in Table 4. To evaluate the applicability of the present method, we compared values for 60 serum samples as measured by the present method with those obtained with the above-mentioned radioimmunoassay. As shown in Figure 2, there was good correlation between the present method and the radio-
Table 2. AnalytIcal Recovery of Insulin Added to Human Serum Insulin
concn,
mull-mt.
unlts/L
Added
Recovered
50 100 200
49.6 (6.5)a 94.8 (17.4) 187
(17.6)
Recovery, % 99.2 (13.0)a 94.8 (17.4) 93.5 (8.8)
No. of serum samples was seven throughout. a Mean (and SD).
CLINICAL CHEMISTRY, Vol. 31, No. 4, 1985 629
Table 3. PrecIsion of the Present insulin Assay S.mpl.
Mean (and SD), mIllI-Int. unlts/L.
Within-run (n
=
15 each)
1 2
CV, S
28.0 (2.1)
-
35.8 (3.1)
3 Between-run
156 (n
10 each)
(5.8)
28.3 (2.0)
7.1
1 2
9.0
35.4 (3.2)
155
3
52
(8.1)
Table 4. interference from HemoglobIn and Blllrubin in Insulin Assay InsulIn concn, mlIIHnt. units
-
Determined B
A
Hemoglobin,
5 g/L
1 2 3
58.1 90.5 29.9 48.1 117
48.4
88.5 31.3 44.0 121
4
5
assay (ELA Insulin Test-S from Medical and Biological Laboratories, Nagoya, Japan). Here again, there was a good correlation between values obtained by the two methods (y = 0.92x 2.9; r = 0.97; more complete data not shown). The assay method for insulin described here is practicable, simple, and suitable for routine use in clinical laboratories. It is sufficiently sensitive for determining the concentration of insulin in serum, and one can complete the assay within about 3 h. In our method, free labeled antibody is measured, which is different from most conventional immunoassays, such as the double-antibody radioimmunoassay or the sandwich-type immunosorbent assay, which require effective separation of free label from bound. The column-separation method is suitable for this purpose, because it is possible to immobilize a much larger amount of antigen on Sepharose matrix than on polystyrene beads, plastic tubing, and so on. From the results in preliminary experiments, 1 mull-mt. unit of /3-n-
120 102 95.5 109 96.7 a 105(10.1)
galactosidase activity corresponds to 2.2 x i0’ mol of the enzyme. In our method, the available conjugate per assay tube is about 1.3 milli-int. unit, from the results shown in Table 1, and this corresponds to 2.9 x 10_14 mol. The amount of insulin immobilized on Sepharose, 5 g or 8.3 x 10b0 mol of insulin per column, much exceeds the amount
112 96.0 97.4 103
of the conjugate available. The present method, performed by passing the solutions through the mimcolumn sequentially, requires no centrifugation or removal of a solid phase from the reaction mixture.
102 631 a
Therefore, method is considered mechanized this immunoassay.
BIlirubin, 200 mg/i.
1
48.4
2 3
88.5
54.0 85.0
31.3 44.0
30.5 45.3
4 5
121
122 to be suitable
for
Mean (and SD).
References
100
#{149}
.
S
.i
S
#{149}C 50 S
C
-.
0
50
100
Insulin (mIU/L, RIA) Fig.2.Assay of insulin in60 serum samples by the presentmethod (EIA)and a radioummunoassay (RIA) Theradloimmunoassay ofinsulin was performed withInsulin Elken”ALAkit from ElianChemicalCo.,Tokyo.Jwan. mlU’ ismillknternalional units immunoassay. The regression equation was y = 0.86x + 2.7, the correlation coefficient 0.98. We also compared the values as measured with those obtained by a solid-phase “sandwich”-type enzyme immune-
630 CLINICALCHEMISTRY, Vol. 31, No.4, 1985
1. Kate K. Use of activated thiol-Sepharose in a separation method for enzyme immunoassay. Methods Enzymol 92, 345-359 (1983). 2. Yamainote R, Hattori S. Inukai T, et al. Enzyme immunoassay for thyroxin and triiodothyronine in human serum, with use of a covalent chromatographic separation method. C/in Chem 27, 17211723 (1981). 3. Umeda Y, Suzuki F, Kosaka A, Kate K. Enzyme immunoassay for insulin with a novel separation method using activated thiolSepharose. C/in Chim Acta 107, 267-272 (1980). 4. Yamamoto R, Umeda Y, Koseka A, Kate K. Enzyme immunoassay for antibodies in serum using a covalent chromatographic method for separation of the bound label. J Biochem 89, 223-229 (1981). 5. Kate K, Umeda Y, Suzuki F, Kosaka A. Interference in a solidphase enzyme immunoassay system by serum factors. J Appl Biochem 1, 479-488 (1979). 6. Yamamoto R, Hattori 5, Matsuura A, et al. Enzyme immunoassay for thyroxine and triiodothyronine with an improved columnseparation method. J Appi Biochem 4, 168-174 (1982). 7. Yainamoto R, Kimura S, Hattori S, et al. Column enzyme immunoassay for secretory inununoglobulin A in serum. Cl/n Chem 29, 151-153 (1983). 8. Kate K, Umeda Y, Suzuki F, Kosaka A. Improved reaction buffers for solid-phase enzyme immunoassay without interference by serum factors. Clin Chim Acts 102, 261-265 (1980). 9. Kate K, Fukui H, Hamaguchi Y, Ishikawa E. Enzyme-linked immunoassay: Conjugation of the Fab’ fragment of rabbit IgG with -i)-galactosidase from E. coli and its use for immunoassay. J Immunol 116, 1554-1560 (1976). 10. Jones KM. Artificial substrates and biochemical reagents. In Data for Biochemical Research, Dawson RMC, et al., Eds., Oxford University Press, London, 1969, pp 436-465.
