224
GENERALMETHODOLOGY
[24]
[24] S e q u e n t i a l S o l i d - P h a s e A s s a y for B i o t i n B a s e d o n x2SI-Labeled A v i d i n
By DONALD M. MOCK Interest in accurate measurement of biotin concentrations in plasma and urine has been stimulated by wider application of avidin-biotin techniques in biomedical research, by recent advances in the understanding of biotin-responsive inborn errors of metabolism, and by several reports describing biotin deficiency associated with parenteral nutrition. This chapter describes a biotin assay utilizing radiolabeled avidin in a sequential solid-phase method. ~The assay has increased sensitivity compared to published methods based on avidin binding, correlates with expected trends in biotin concentrations in blood and urine in a rat model of biotin deficiency, and can utilize radiolabeled avidin available from a commercial source. The assay will detect several avidin-binding substances that have binding affinities equal to or not greatly decreased compared to that of biotin. Thus, the assay is not specific for biotin; when coupled with the HPLC separation described here, the method is a powerful tool for measuring biotin and biotin analogs in complex biological mixtures. Assay Method
Principle. A fixed amount of 125I-labeled avidin is incubated with varying concentrations of biotin in buffer (standard curve) and with several dilutions of the unknown samples (e.g., serum or urine). A portion of the first incubation is then transferred to microtiter plates previously coated with biotin covalently linked to bovine serum albumin (BSA). After a second incubation, the wells are washed with buffer and counted individually. Increasing amounts of biotin in the standards or unknowns occupy an increasing proportion of the biotin-binding sites on ~25I-labeled avidin in the first incubation, and progressively fewer counts are bound to the well in the second incubation. "Biotin" concentration (actually, total concentration o f avidin-binding substances) in the unknowns is determined by comparing counts bound to the standard curve. Safety Considerations. When the iodination procedure described here is properly performed, radioactive iodine gas should not be released. However, as a precaution against errors that result in production of ~z512 1 D . M . M o c k a n d D . B . D u B o i s , Anal. Biochem. 153, 272 (1986).
METHODS IN ENZYMOLOGY, VOL 184
Copyright © 1990 by Academic Press, Inc. All rights of reproduction in any, form reserved.
[24]
SOLID-PHASE RADIOASSAY FOR BIOTIN
225
(e.g., acidification), a properly operating fume hood should be used for all steps through the final molecular sieve separation of 125I- from ~25I-labeled avidin. Use of a hood also provides an appropriate area for containment of spills, which are more likely to occur than release of 12512. Materials and Reagents U-bottomed microtiter plates (Cat. No. 001-010-2201), Immulon II Removawell strips (011-010-6302), and Removawell holders (011010-6601), Dynatech Laboratories, Inc., Alexandria, VA d-Biotin and BSA, RIA grade (Sigma Chemical Co., St. Louis, MO) Avidin-D (Vector Laboratories, Burlingame, CA) N-2-Hydroxyethylpiperazine-N'-2-ethanesulfonic acid (HEPES, Calbiochem, La Jolla, CA) All other chemicals should be reagent grade or better. The water used is distilled and deionized in all buffers, washes, etc. Additional chemicals required are mono- and dibasic sodium phosphate, sodium chloride, sodium hydroxide, sodium carbonate and bicarbonate, and sodium azide. Additional materials will be needed if lZSI-labeled avidin is synthesized rather than purchased; these are mentioned in the description of the synthetic methods below. Equipment 7 counter for 1251quantitation Twelve (or 96-)-channel pipet (e.g., Titertek, Flow Laboratories, Inc., McLean, VA), optional Twelve (or 96-)-channel plate washer (e.g., Miniwasher, Skatron AS, Lier, Norway), optional Preparation of 1251-Labeled Avidin The assay works well with ~25I-labeled avidin synthesized in any of three ways, namely, radioiodination with chloramine-T, Iodo-Beads, or the Bolton-Hunter reagent. Preparation of HPP-Avidin. Avidin-D is derivatized with N-succinimidyl-3-(p-hydroxyphenyl)propionate (SHPP, Pierce Chemical Co., Rockford, IL) by the method of Finn et al. 2 The reagent is dissolved in 2-propanol: ethyl acetate (3:2) at a concentration of 5 mg/ml. A 100-/zl aliquot of the SHPP solution is added to 2 ml of avidin [9.5 mg/ml in 0.2 M borate buffer (pH 9.0) prepared from sodium borate and pH adjusted with 0.2 M boric acid]. The reaction mixture is incubated in an ice bath with occasional shaking for 3 hr and then dialyzed overnight against 4 liters of 2 F. M. Finn, G. Titus, J. A. Montibeller, and K. Hofmann, J. Biol. Chem. 225, 5742 (1980).
226
GENERALMETHODOLOGY
[24]
0.5 M phosphate buffer (pH 7.5). The dialyzed HPP-avidin product is aliquoted into vessels (e.g., I0 or 50/xl in 1.5-ml microcentrifuge tubes), capped, and stored at - 7 0 ° until used for iodination. Chloramine-T Radioiodination of HPP-Avidin. Labeling of HPPavidin is carried out by the method of Hunter and Greenwood. 3 A 40-/~1 volume of HPP-avidin (200 t~g) in 50 mM phosphate buffer (pH 7.5, prepared from Na2HPO4 and NaHzPO4 mixed in a ratio of 84 to 16% by weight) is added to a reaction vial containing 1 mCi of carrier-free 1251 (100 mCi/ml, Amersham Corp., Arlington Heights, IL). A 12.5-/zl aliquot of a chloramine-T solution (2 mg/ml in the same phosphate buffer) is added to the mixture, which is then allowed to incubate for 15 min at room temperature. The reaction is stopped by the addition of 50/zl of a solution of sodium metabisulfite (0.5 mg/ml in water). The entire reaction mixture is transferred to 1 ml of Tris-saline [20 mM Tris-HC1, 0.15 M saline (pH 7.4 at 25°)] containing 1 mg/ml BSA and 1 mg/ml sodium iodide. Separation of Free from Bound t25I. Free 1251- is separated from 1251labeled avidin by molecular sieve chromatography as follows. Apply the sample to the drained bed of a 0.9 × 25 cm Sephadex G-25 fine (Pharmacia, Uppsala, Sweden) column previously equilibrated with 50 mM phosphate buffer (pH 7.4) containing 0.1% BSA (w/v), which serves here as a carrier for the nSI-labeled avidin to prevent loss by adsorption to the column. Drain the sample into the bed of the column and begin collecting 1.0-ml fractions into tubes which contain 1.0-ml aliquots of 0.1% BSA in phosphate buffer. After the sample is drained into the bed, wash the column walls with several milliliters of buffer and continue collecting fractions. Additional buffer is added as necessary to collect 30 fractions. Quantitate the radioactivity in 50/xl of each fraction using a 3' counter; determine the position and separation between the two peaks. Pool the four or five fractions of the first peak that have the highest activity. The 125I-labeled avidin runs in the exclusion volume and thus elutes first. The specific activity of the ~zSI-labeled avidin should be approximately 5/.~Ci//zg. Specific activities up to 40/~Ci//xg can be obtained by using less avidin in the radioiodination step. The pooled 125I-labeled avidin is aliquoted (10/xl/tube); tubes are capped and frozen at -20 °. Radioiodination of HPP-Avidin by the Iodo-Beads Method. The Iodo-Beads method of iodination is less complex and less time-consuming than the chloramine-T method. The 125I-labeled avidin synthesized by this method is more uniform in stability and specific activity. Wash four Iodo-Beads (Pierce Chemical Co.) twice with 5 ml of 3 W. M. Hunter and F. C. Greenwood, Nature (London) 194, 495 (1%2).
[24]
SOLID-PHASE RADIOASSAY FOR BIOTIN
227
50 mM phosphate buffer (pH 7.4). Drain the beads on filter paper. Place 180/xl of the same phosphate buffer and the four Iodo-Beads in a vial containing 2 mCi of 1251; the original vial from the supplier (e.g., Amersham) is suitable. Mix by thumping the closed vial, and allow the mixture to stand at room temperature for 5 min. The Iodo-Beads are now preloaded with 1251. Pipet 4/zl of HPP-avidin into the vial containing the preloaded IodoBeads; close and mix for 20 min by thumping every 5 min at room temperature. Next, separate the aqueous reaction mixture from the Iodo-Beads by transferring the liquid from the reaction vial to a 1.5-ml microcentrifuge tube using a pipet with a disposable tip. Wash the Iodo-Beads with 200/zl of 50 mM phosphate buffer (pH 7.4) and transfer the wash to the same tube. Separation of 125I-labeled avidin from free 1251- is performed on a Sephadex G-25 column as described above. The final pool of 125I-labeled avidin should have a specific activity of 25-50 ~Ci//zg, and the concentration should be about 200,000 cpm/~l. The pooled 125I-labeled avidin is aliquoted ( - 2 5 /~l/tube), and the tubes are capped and stored at 4 ° or frozen at - 2 0 ° . 125I-Labeled Avidin by Bolton-Hunter Reagent. 125I-Labeled avidin synthesized by the Bolton-Hunter reagent method 4 can be obtained from New England Nuclear (Boston, MA). Comments. The 125I-labeled avidin synthesized by each of these methods will bind quantitatively (>95%) to an iminobiotin column 5 at pH 9 and will be released at pH 4. The 125I-labeled avidin declines in specific activity and affinity for binding to the biotin-BSA plates more rapidly than would be predicted from simple radioactive decay. We routinely discard 125I-labeled avidin that is more than 2 months old.
Biotinylation of BSA BSA is biotinylated according to the method of Heitzmann and Richards. 6 A 500-mg amount of BSA is dissolved in 50 ml of ice-cold 0.1 M NaHCO3 (pH 7.5). A 5-ml volume of biotin N-hydroxysuccinimide ester (BNHS, Pierce Chemical Co.) at a concentration of 12 mg/ml in N, Ndimethylformamide is added to the BSA solution, which is incubated with gentle stirring at 4 ° overnight. This mixture is then dialyzed exhaustively against distilled, deionized water to remove unreacted BNHS and finally buffered by dialyzing against HEPES buffer (see below). The biotin-BSA 4 A. E. Bolton and W. M. Hunter, Biochem. J. 133, 529 (1973). 5 K. Hofmann, S. W. Wood, C. C. Brinton, J. A. Montibeller, and F. M. Finn, Proc. Natl. Acad. Sci. U.S.A. 77, 4666 (1980). 6 H. Heitzmann and F. M. Richards, Proc. Natl. Acad. Sci. U.S.A. 71, 3537 (1974).
228
GENERALMETHODOLOGY
[24]
is aliquoted (200/zl/tube), and the microcentrifuge tubes are capped and frozen at - 2 0 °. The biotin-BSA preparation is stable for at least I year; each tube contains enough biotin-BSA to coat 10 plates (i.e., 96 wells: 8 strips of 12 wells each). The degree of biotinylation of BSA is assessed by measuring the amount of lzSI-labeled avidin that will bind to a well which is coated with biotin-BSA as follows. Polystyrene wells (96 wells/plate, Immulon II) are coated with either biotin-BSA or native BSA as described below. Serial dilutions of 125I-labeled avidin are incubated in the wells overnight at 4°; the wells are then washed, separated, and counted. Wells coated with biotin-BSA will reproducibly bind 50-100 times more ~zSI-labeled avidin than wells coated with native BSA.
Performing the Assay Buffers HEPES buffer [0.2 M HEPES, 2 M NaCI (pH 7.0)]: Dissolve 190.64 g HEPES, 467.52 g NaC1, and 0.8 g sodium azide in about 3 liters of water; adjust the pH to 7.0 with I0 M NaOH and bring to 4.0 liters Coating buffer [50 mM bicarbonate (pH 9.6)]: Dissolve 1.59 g sodium carbonate and 2.93 g sodium bicarbonate in 900 ml water and bring to 1.0 liter with water; this buffer can be stored at 4 ° for at least 2 weeks Washing buffer [20 m M HEPES, 0.2 M NaC1 (pH 7.0)]: Dilute HEPES buffer (above) 1/10 with water Blocking buffer [0.2 M HEPES, 2 M NaC1 (pH 7.0) with 0.01% BSA and 0.02% sodium azide]: Dissolve 0.1 g BSA in 1 liter of HEPES buffer; this buffer is stable at 4 ° for at least 2 weeks Avidin buffer (HEPES buffer with 0.1% BSA): Dissolve r g BSA in 1 liter of HEPES buffer; stable at 4 ° for at least 2 weeks Biotin Standards. A 100 nM biotin stock solution is prepared by dissolving 24.43 mg of d-biotin in 1 liter of distilled water. When aliquoted into 1 ml fractions and stored at - 7 0 °, this reference biotin stock is stable for at least 2 years. A series of biotin standards is prepared from the reference stock solution by dilution in HEPES buffer. The dilutions are performed in disposable graduated polyethylene specimen containers having a volume greater than 200 ml. Twenty milliliters of the reference stock is diluted 1/10 to yield 200 ml of a 1.00 x 10 7 fmol/ml solution. Three subsequent 1/10 dilutions are performed in the same manner to generate 1.00 x 10 6, 1.00 x 10~, and 1.00 x 104 fmol/ml standards. Fifty milliliters of the 1 x 104
[24]
229
SOLID-PHASE RADIOASSAY FOR BIOTIN
fmol/ml solution is diluted to 200 ml to yield the 2500 fmol/1 biotin standard. The final 16 dilutions of 1/1.5 in the series yield concentrations (in fmol/ml) of 1.67 × 10 3, 1.11 x 10 3, 741,494,330, 220, 146, 97.5, 65.0, 43.4, 28.9, 19.3, 12.8, 8.6, 5.7, and 3.8; the dilutions are prepared by diluting 120 ml of each stock to 180 ml. This scheme leaves 60 ml of the previous dilution to use as a standard. The dilutions are stored at - 7 0 ° after aliquoting 0.4-ml volumes. This dilution scheme yields 150 sets of tubes with 20 biotin concentrations ranging from 1 x 10 7 to 3.8 fmol/ml. Preparing Coated Plates. The biotin-BSA-coated wells (plates of 96 wells) must be prepared at least 1 day before the assay. The biotin-BSA stock (prepared and stored as described above) is diluted 1/1000 in coating buffer. A 96-well plate requires 20 ml of coating buffer. Prepare the coating solution by adding 20/zl of biotin-BSA stock (10 mg/ml) to 19.98 ml of coating buffer to produce a final concentration of biotin-BSA of 10/xg/ml. 7 Add 200/xl of this biotin-BSA coating solution to each well of the Immulon II strips (8 strips per plate). Cover the wells and incubate at 4 ° at least overnight but not more than 5 days; 4 days has given the best results. After incubation, aspirate the coating solution and wash each Immulon II strip twice with washing buffer as follows: fill the well to the brim with washing buffer and then aspirate the washing buffer. 8 Next add 400 ~1 of blocking buffer to each well of the washed plates. Incubate for at least 2 hr at room temperature before using. Plates can be stored in blocking buffer for I month at 4°. At the time of assay, aspirate the blocking buffer and wash the plate twice with washing buffer immediately before transferring the contents from the first incubation microtiter wells. Do not allow drying of the biotin-BSA-coated wells for more than 1 minute. Incubation 1. Add 100/zl of each biotin standard dilution to a well of the microtiter uncoated plate. Run triplicates of each standard. To determine maximum counts bound (i.e., zero concentration of biotin), use 100/zl of HEPES buffer; run in triplicate. These three wells are subsequently handled just as if they contained standards and yield a value for the maximum binding in the absence of biotin. To determine the number of counts transferred to Immulon II (and thus calculate a maximum percent binding which is equal to the maximum counts bound divided by the total counts transferred to Incubation 2), again add 100 ~1 HEPES buffer to three wells; after adding the ~25I-labeled avidin as described below, 7 Biotin-BSA coating solution should be prepared in polypropylene tubes, styrene tubes, because the biotin-BSA will adsorb to the polystyrene. 8 A multichannel well washer is a worthwhile timesaving device.
not
in poly-
230
GENERALMETHODOLOGY
[24]
100/~1 of the contents of each of these three wells is counted directly instead of being transferred to Incubation 2, thus yielding a mean value for the total counts in the final incubation. Also, add 100 /zl of each dilution of each unknown sample to a well in the microtiter plate. Triplicates are run for each dilution of each unknown. After all standards and unknowns have been added to the wells in the microtiter plate, add 50/zl of ~25I-labeled avidin solution to each well. The ~25I-labeled avidin solution is made by diluting the stock ~25I-labeled avidin with avidin buffer (see above) to a concentration of approximately 10,000 cpm/50/zl. As the labeled avidin degrades with age, more avidin will be required to maintain 10,000 cpm/50/~1. A 50-~1 aliquot of diluted avidin should be counted to measure total cpm//zl before each assay. Mix the contents of Incubation 1 by alternately drawing up and expelling the contents of each well with a pipet. Incubate for 1 hr at room temperature (the length of this incubation is not critical). Incubation 2. After the first incubation, transfer 100/zl from each well in the microtiter plate to a well of the Immulon II plate. Allow the samples to incubate in the coated wells for at least 4 hr at room temperature or up to 3 days at 4 °. Then aspirate the contents of the Immulon II wells and wash twice with washing buffer to remove all unbound ~zSI-labeled avidin. Quantitate the bound 125I-labeled avidin by counting. The Removawell strips of Immulon II can be broken apart for placement into a suitable carrier for the 3' counter. Data Analysis and Interpretation A standard curve (Fig. 1A) depicts uncorrected counts bound to the well versus the log of the biotin concentration. A log-logit transformation (Fig. 1B) does not consistently produce a linear dependence on the log of the biotin concentration. Competitive isotope dilution assays in liquid phase characteristically have linear log-logit plots; the fact that this assay is sequential rather than competitive and has a solid-phase binding step may be responsible for the more complex curve shape. Determination of biotin concentration graphically is acceptable, but tedious when assaying large numbers of unknowns. We use a polynomial curve-fitting program to fit the curve to r 2 > 0.99; typically a fourth or fifth degree polynomial is suitable. The assay sensitivity has typically been about 100 fmol/ml (10 fmol/ assay well) but has occasionally been 10 fmol/ml (1 fmol/assay well) with an avidin preparation of unusually high specific activity and unusually high affinity for the solid phase. The intraassay variability was determined on 10 replicates on rat urine (--1/400 dilution) in a single assay; the coeffi-
[24]
SOLID-PHASE RADIOASSAY FOR BIOTIN
231
2800
A 2400 z > 2000 in
a
1600
z
o >iI-U
1200
800
,¢ o 400
¢v I
100
I
300 BIOTIN ( f m o l / m l )
I
I
1000 log scale
2500
3.0
B c~ 2 . 0 z o
m
1.0
t,--
z
14.1 U
'~' 14.1 m
0
o -1.0; I--
9_2.o i
-3.0 ¸
I
100
lO O
2 ioo
BIOTIN f f m o l / m l , log scale FIG. 1. (A) Plot of uncorrected counts bound to the coated wells versus log of biotin concentration in the series of biotin standards. (B) Plot of logit of percent binding versus log of biotin concentration.
232
GENERALMETHODOLOGY
[24l
cient of variation was 10%. The long-term precision (interassay variability) was estimated from results of triplicate determinations from assays of a sample by three different technicians on 6 different days; the coefficient of variation (calculated from the standard deviation of the means of the triplicates) was 6%, and the intraclass correlation coefficient on that data was -0.77. We have investigated the use of an alternate coating protein (keyhole limpet hemocyanin), and the use of biotinylating agent with a longer spacer (NHS-LC-biotin, Pierce Chemical Co.). These modifications resuited in modestly higher maximum percent binding (85 versus 65%) but gave essentially the same value for biotin (i.e., total avidin-binding substances) in urine and ultrafiltrates of plasma. We have also investigated the detectability of biotin analogs in this assay. For example, a series of standard concentrations of biocytin, biotin sulfoxide, and bisnorbiotin were assayed. The biocytin standard curve was identical to the biotin standard curve, within experimental error. For the others, the point of half-maximal binding (the inflection point of the sigmoid curve) was shifted toward greater concentrations of the analog, and the true concentration of the analog would have been underestimated by a factor of 4 to 5 for both analogs. The lower binding affinities of some biotin analogs 9 may allow the solid-phase biotin (biotinylated BSA) to compete for binding to ~25I-labeled avidin in the second step and thus modify the standard curve. This observation emphasizes two limitations of the assay that may well apply to other avidin-binding assays: (1) in a complex biological mixture of biotin, biotin analogs (e.g., metabolites), and even protein-bound biotin, the measured total avidin-binding substances may not equal the stoichiometric sum of all avidin-binding substances, and (2) a more accurate measure could be made if the avidin-binding substances are separated chromatographically and each avidin-binding substance identified and quantitated versus the standard curve for that compound. A tacit assumption in previous measurements of biotin in human plasma has been that biotin is the only avidin-binding substance present (if using an avidin-binding assay) or that biotin is the only growth-promoting substance present (if using a bioassay). Using paper chromatography and a bioassay, Lee et al. l0 demonstrated the presence of substantial amounts of biotin metabolites in human urine. More recently this assumption concerning human plasma was examined by using reversed-phase HPLC to separate biotin analogs in ultrafiltrates of human plasma. Figure 2 depicts 9 N. M. Green, Ado. Protein Chem. 34, 1967 (1975). to H. Lee, L. D. Wright, and D. B. McCormick, J. Nutr. 11)2, 1453 (1972).
[24]
SOLID-PHASE RADIOASSAY FOR BIOTIN
233
100 C w 90 r,,3 z B0 Q~ 0 (/1 70 m < I-- 60 Ld -J 0 50
b
e
f
o
40 J
30 1,1
>_
F--
20
5 ¢Y
10 0
k______ 0
5
10
15
RETENTION TIME ( m i n u t e s ) FIG. 2. Chromatography of a mixture of biotin analogs. The biotin analogs are biotin l-sulfoxide (a), biotin d-sulfoxide (b), biocytin (c), biotin (d), dethiobiotin (e), and biotin methyl ester (f). Chromatography parameters were a linear gradient from 0% A to 50% A at 1.0 ml/min for 40 min. [Solution A, methanol; solution B, 10 m M potassium phosphate (pH 4.5)].
the HPLC separation of a mixture of biotin analogs; in this mixture, concentrations were sufficient (e.g., micromolar range) for detection by ultraviolet absorption. In human plasma, however, biotin and biotin analogs are present in nanomolar concentrations. Using the ]25I-labeled avidin assay to measure the fractions collected after chromatography, a substantial amount of avidin-binding substances that are not biotin were detected, u-~3 Recovery of total avidin-binding substances was 94 - 9%. Biotin accounted for only 60 - 6% (X - 1 SD) of total avidin-binding substances; four other peaks were detected that did not coelute with biocytin or biotin, the only recognized biotin vitamers which occur naturally.
n D. M. Mock, Fed. Proc., Fed. Am. Soc. Exp. Biol. 46, 1159 (1987). 12 D. M. Mock and G. L. Lankford, FASEB J. 3, A1058 (1989). t3 T. Suormala, E. R. Baumgartner, J. Bausch, W. Holick, and H. Wick, Clin. Chim. Acta 177, 253 (1988).
GENERALMETHODOLOGY
[24]
[24] S e q u e n t i a l S o l i d - P h a s e A s s a y for B i o t i n B a s e d o n x2SI-Labeled A v i d i n
By DONALD M. MOCK Interest in accurate measurement of biotin concentrations in plasma and urine has been stimulated by wider application of avidin-biotin techniques in biomedical research, by recent advances in the understanding of biotin-responsive inborn errors of metabolism, and by several reports describing biotin deficiency associated with parenteral nutrition. This chapter describes a biotin assay utilizing radiolabeled avidin in a sequential solid-phase method. ~The assay has increased sensitivity compared to published methods based on avidin binding, correlates with expected trends in biotin concentrations in blood and urine in a rat model of biotin deficiency, and can utilize radiolabeled avidin available from a commercial source. The assay will detect several avidin-binding substances that have binding affinities equal to or not greatly decreased compared to that of biotin. Thus, the assay is not specific for biotin; when coupled with the HPLC separation described here, the method is a powerful tool for measuring biotin and biotin analogs in complex biological mixtures. Assay Method
Principle. A fixed amount of 125I-labeled avidin is incubated with varying concentrations of biotin in buffer (standard curve) and with several dilutions of the unknown samples (e.g., serum or urine). A portion of the first incubation is then transferred to microtiter plates previously coated with biotin covalently linked to bovine serum albumin (BSA). After a second incubation, the wells are washed with buffer and counted individually. Increasing amounts of biotin in the standards or unknowns occupy an increasing proportion of the biotin-binding sites on ~25I-labeled avidin in the first incubation, and progressively fewer counts are bound to the well in the second incubation. "Biotin" concentration (actually, total concentration o f avidin-binding substances) in the unknowns is determined by comparing counts bound to the standard curve. Safety Considerations. When the iodination procedure described here is properly performed, radioactive iodine gas should not be released. However, as a precaution against errors that result in production of ~z512 1 D . M . M o c k a n d D . B . D u B o i s , Anal. Biochem. 153, 272 (1986).
METHODS IN ENZYMOLOGY, VOL 184
Copyright © 1990 by Academic Press, Inc. All rights of reproduction in any, form reserved.
[24]
SOLID-PHASE RADIOASSAY FOR BIOTIN
225
(e.g., acidification), a properly operating fume hood should be used for all steps through the final molecular sieve separation of 125I- from ~25I-labeled avidin. Use of a hood also provides an appropriate area for containment of spills, which are more likely to occur than release of 12512. Materials and Reagents U-bottomed microtiter plates (Cat. No. 001-010-2201), Immulon II Removawell strips (011-010-6302), and Removawell holders (011010-6601), Dynatech Laboratories, Inc., Alexandria, VA d-Biotin and BSA, RIA grade (Sigma Chemical Co., St. Louis, MO) Avidin-D (Vector Laboratories, Burlingame, CA) N-2-Hydroxyethylpiperazine-N'-2-ethanesulfonic acid (HEPES, Calbiochem, La Jolla, CA) All other chemicals should be reagent grade or better. The water used is distilled and deionized in all buffers, washes, etc. Additional chemicals required are mono- and dibasic sodium phosphate, sodium chloride, sodium hydroxide, sodium carbonate and bicarbonate, and sodium azide. Additional materials will be needed if lZSI-labeled avidin is synthesized rather than purchased; these are mentioned in the description of the synthetic methods below. Equipment 7 counter for 1251quantitation Twelve (or 96-)-channel pipet (e.g., Titertek, Flow Laboratories, Inc., McLean, VA), optional Twelve (or 96-)-channel plate washer (e.g., Miniwasher, Skatron AS, Lier, Norway), optional Preparation of 1251-Labeled Avidin The assay works well with ~25I-labeled avidin synthesized in any of three ways, namely, radioiodination with chloramine-T, Iodo-Beads, or the Bolton-Hunter reagent. Preparation of HPP-Avidin. Avidin-D is derivatized with N-succinimidyl-3-(p-hydroxyphenyl)propionate (SHPP, Pierce Chemical Co., Rockford, IL) by the method of Finn et al. 2 The reagent is dissolved in 2-propanol: ethyl acetate (3:2) at a concentration of 5 mg/ml. A 100-/zl aliquot of the SHPP solution is added to 2 ml of avidin [9.5 mg/ml in 0.2 M borate buffer (pH 9.0) prepared from sodium borate and pH adjusted with 0.2 M boric acid]. The reaction mixture is incubated in an ice bath with occasional shaking for 3 hr and then dialyzed overnight against 4 liters of 2 F. M. Finn, G. Titus, J. A. Montibeller, and K. Hofmann, J. Biol. Chem. 225, 5742 (1980).
226
GENERALMETHODOLOGY
[24]
0.5 M phosphate buffer (pH 7.5). The dialyzed HPP-avidin product is aliquoted into vessels (e.g., I0 or 50/xl in 1.5-ml microcentrifuge tubes), capped, and stored at - 7 0 ° until used for iodination. Chloramine-T Radioiodination of HPP-Avidin. Labeling of HPPavidin is carried out by the method of Hunter and Greenwood. 3 A 40-/~1 volume of HPP-avidin (200 t~g) in 50 mM phosphate buffer (pH 7.5, prepared from Na2HPO4 and NaHzPO4 mixed in a ratio of 84 to 16% by weight) is added to a reaction vial containing 1 mCi of carrier-free 1251 (100 mCi/ml, Amersham Corp., Arlington Heights, IL). A 12.5-/zl aliquot of a chloramine-T solution (2 mg/ml in the same phosphate buffer) is added to the mixture, which is then allowed to incubate for 15 min at room temperature. The reaction is stopped by the addition of 50/zl of a solution of sodium metabisulfite (0.5 mg/ml in water). The entire reaction mixture is transferred to 1 ml of Tris-saline [20 mM Tris-HC1, 0.15 M saline (pH 7.4 at 25°)] containing 1 mg/ml BSA and 1 mg/ml sodium iodide. Separation of Free from Bound t25I. Free 1251- is separated from 1251labeled avidin by molecular sieve chromatography as follows. Apply the sample to the drained bed of a 0.9 × 25 cm Sephadex G-25 fine (Pharmacia, Uppsala, Sweden) column previously equilibrated with 50 mM phosphate buffer (pH 7.4) containing 0.1% BSA (w/v), which serves here as a carrier for the nSI-labeled avidin to prevent loss by adsorption to the column. Drain the sample into the bed of the column and begin collecting 1.0-ml fractions into tubes which contain 1.0-ml aliquots of 0.1% BSA in phosphate buffer. After the sample is drained into the bed, wash the column walls with several milliliters of buffer and continue collecting fractions. Additional buffer is added as necessary to collect 30 fractions. Quantitate the radioactivity in 50/xl of each fraction using a 3' counter; determine the position and separation between the two peaks. Pool the four or five fractions of the first peak that have the highest activity. The 125I-labeled avidin runs in the exclusion volume and thus elutes first. The specific activity of the ~zSI-labeled avidin should be approximately 5/.~Ci//zg. Specific activities up to 40/~Ci//xg can be obtained by using less avidin in the radioiodination step. The pooled 125I-labeled avidin is aliquoted (10/xl/tube); tubes are capped and frozen at -20 °. Radioiodination of HPP-Avidin by the Iodo-Beads Method. The Iodo-Beads method of iodination is less complex and less time-consuming than the chloramine-T method. The 125I-labeled avidin synthesized by this method is more uniform in stability and specific activity. Wash four Iodo-Beads (Pierce Chemical Co.) twice with 5 ml of 3 W. M. Hunter and F. C. Greenwood, Nature (London) 194, 495 (1%2).
[24]
SOLID-PHASE RADIOASSAY FOR BIOTIN
227
50 mM phosphate buffer (pH 7.4). Drain the beads on filter paper. Place 180/xl of the same phosphate buffer and the four Iodo-Beads in a vial containing 2 mCi of 1251; the original vial from the supplier (e.g., Amersham) is suitable. Mix by thumping the closed vial, and allow the mixture to stand at room temperature for 5 min. The Iodo-Beads are now preloaded with 1251. Pipet 4/zl of HPP-avidin into the vial containing the preloaded IodoBeads; close and mix for 20 min by thumping every 5 min at room temperature. Next, separate the aqueous reaction mixture from the Iodo-Beads by transferring the liquid from the reaction vial to a 1.5-ml microcentrifuge tube using a pipet with a disposable tip. Wash the Iodo-Beads with 200/zl of 50 mM phosphate buffer (pH 7.4) and transfer the wash to the same tube. Separation of 125I-labeled avidin from free 1251- is performed on a Sephadex G-25 column as described above. The final pool of 125I-labeled avidin should have a specific activity of 25-50 ~Ci//zg, and the concentration should be about 200,000 cpm/~l. The pooled 125I-labeled avidin is aliquoted ( - 2 5 /~l/tube), and the tubes are capped and stored at 4 ° or frozen at - 2 0 ° . 125I-Labeled Avidin by Bolton-Hunter Reagent. 125I-Labeled avidin synthesized by the Bolton-Hunter reagent method 4 can be obtained from New England Nuclear (Boston, MA). Comments. The 125I-labeled avidin synthesized by each of these methods will bind quantitatively (>95%) to an iminobiotin column 5 at pH 9 and will be released at pH 4. The 125I-labeled avidin declines in specific activity and affinity for binding to the biotin-BSA plates more rapidly than would be predicted from simple radioactive decay. We routinely discard 125I-labeled avidin that is more than 2 months old.
Biotinylation of BSA BSA is biotinylated according to the method of Heitzmann and Richards. 6 A 500-mg amount of BSA is dissolved in 50 ml of ice-cold 0.1 M NaHCO3 (pH 7.5). A 5-ml volume of biotin N-hydroxysuccinimide ester (BNHS, Pierce Chemical Co.) at a concentration of 12 mg/ml in N, Ndimethylformamide is added to the BSA solution, which is incubated with gentle stirring at 4 ° overnight. This mixture is then dialyzed exhaustively against distilled, deionized water to remove unreacted BNHS and finally buffered by dialyzing against HEPES buffer (see below). The biotin-BSA 4 A. E. Bolton and W. M. Hunter, Biochem. J. 133, 529 (1973). 5 K. Hofmann, S. W. Wood, C. C. Brinton, J. A. Montibeller, and F. M. Finn, Proc. Natl. Acad. Sci. U.S.A. 77, 4666 (1980). 6 H. Heitzmann and F. M. Richards, Proc. Natl. Acad. Sci. U.S.A. 71, 3537 (1974).
228
GENERALMETHODOLOGY
[24]
is aliquoted (200/zl/tube), and the microcentrifuge tubes are capped and frozen at - 2 0 °. The biotin-BSA preparation is stable for at least I year; each tube contains enough biotin-BSA to coat 10 plates (i.e., 96 wells: 8 strips of 12 wells each). The degree of biotinylation of BSA is assessed by measuring the amount of lzSI-labeled avidin that will bind to a well which is coated with biotin-BSA as follows. Polystyrene wells (96 wells/plate, Immulon II) are coated with either biotin-BSA or native BSA as described below. Serial dilutions of 125I-labeled avidin are incubated in the wells overnight at 4°; the wells are then washed, separated, and counted. Wells coated with biotin-BSA will reproducibly bind 50-100 times more ~zSI-labeled avidin than wells coated with native BSA.
Performing the Assay Buffers HEPES buffer [0.2 M HEPES, 2 M NaCI (pH 7.0)]: Dissolve 190.64 g HEPES, 467.52 g NaC1, and 0.8 g sodium azide in about 3 liters of water; adjust the pH to 7.0 with I0 M NaOH and bring to 4.0 liters Coating buffer [50 mM bicarbonate (pH 9.6)]: Dissolve 1.59 g sodium carbonate and 2.93 g sodium bicarbonate in 900 ml water and bring to 1.0 liter with water; this buffer can be stored at 4 ° for at least 2 weeks Washing buffer [20 m M HEPES, 0.2 M NaC1 (pH 7.0)]: Dilute HEPES buffer (above) 1/10 with water Blocking buffer [0.2 M HEPES, 2 M NaC1 (pH 7.0) with 0.01% BSA and 0.02% sodium azide]: Dissolve 0.1 g BSA in 1 liter of HEPES buffer; this buffer is stable at 4 ° for at least 2 weeks Avidin buffer (HEPES buffer with 0.1% BSA): Dissolve r g BSA in 1 liter of HEPES buffer; stable at 4 ° for at least 2 weeks Biotin Standards. A 100 nM biotin stock solution is prepared by dissolving 24.43 mg of d-biotin in 1 liter of distilled water. When aliquoted into 1 ml fractions and stored at - 7 0 °, this reference biotin stock is stable for at least 2 years. A series of biotin standards is prepared from the reference stock solution by dilution in HEPES buffer. The dilutions are performed in disposable graduated polyethylene specimen containers having a volume greater than 200 ml. Twenty milliliters of the reference stock is diluted 1/10 to yield 200 ml of a 1.00 x 10 7 fmol/ml solution. Three subsequent 1/10 dilutions are performed in the same manner to generate 1.00 x 10 6, 1.00 x 10~, and 1.00 x 104 fmol/ml standards. Fifty milliliters of the 1 x 104
[24]
229
SOLID-PHASE RADIOASSAY FOR BIOTIN
fmol/ml solution is diluted to 200 ml to yield the 2500 fmol/1 biotin standard. The final 16 dilutions of 1/1.5 in the series yield concentrations (in fmol/ml) of 1.67 × 10 3, 1.11 x 10 3, 741,494,330, 220, 146, 97.5, 65.0, 43.4, 28.9, 19.3, 12.8, 8.6, 5.7, and 3.8; the dilutions are prepared by diluting 120 ml of each stock to 180 ml. This scheme leaves 60 ml of the previous dilution to use as a standard. The dilutions are stored at - 7 0 ° after aliquoting 0.4-ml volumes. This dilution scheme yields 150 sets of tubes with 20 biotin concentrations ranging from 1 x 10 7 to 3.8 fmol/ml. Preparing Coated Plates. The biotin-BSA-coated wells (plates of 96 wells) must be prepared at least 1 day before the assay. The biotin-BSA stock (prepared and stored as described above) is diluted 1/1000 in coating buffer. A 96-well plate requires 20 ml of coating buffer. Prepare the coating solution by adding 20/zl of biotin-BSA stock (10 mg/ml) to 19.98 ml of coating buffer to produce a final concentration of biotin-BSA of 10/xg/ml. 7 Add 200/xl of this biotin-BSA coating solution to each well of the Immulon II strips (8 strips per plate). Cover the wells and incubate at 4 ° at least overnight but not more than 5 days; 4 days has given the best results. After incubation, aspirate the coating solution and wash each Immulon II strip twice with washing buffer as follows: fill the well to the brim with washing buffer and then aspirate the washing buffer. 8 Next add 400 ~1 of blocking buffer to each well of the washed plates. Incubate for at least 2 hr at room temperature before using. Plates can be stored in blocking buffer for I month at 4°. At the time of assay, aspirate the blocking buffer and wash the plate twice with washing buffer immediately before transferring the contents from the first incubation microtiter wells. Do not allow drying of the biotin-BSA-coated wells for more than 1 minute. Incubation 1. Add 100/zl of each biotin standard dilution to a well of the microtiter uncoated plate. Run triplicates of each standard. To determine maximum counts bound (i.e., zero concentration of biotin), use 100/zl of HEPES buffer; run in triplicate. These three wells are subsequently handled just as if they contained standards and yield a value for the maximum binding in the absence of biotin. To determine the number of counts transferred to Immulon II (and thus calculate a maximum percent binding which is equal to the maximum counts bound divided by the total counts transferred to Incubation 2), again add 100 ~1 HEPES buffer to three wells; after adding the ~25I-labeled avidin as described below, 7 Biotin-BSA coating solution should be prepared in polypropylene tubes, styrene tubes, because the biotin-BSA will adsorb to the polystyrene. 8 A multichannel well washer is a worthwhile timesaving device.
not
in poly-
230
GENERALMETHODOLOGY
[24]
100/~1 of the contents of each of these three wells is counted directly instead of being transferred to Incubation 2, thus yielding a mean value for the total counts in the final incubation. Also, add 100 /zl of each dilution of each unknown sample to a well in the microtiter plate. Triplicates are run for each dilution of each unknown. After all standards and unknowns have been added to the wells in the microtiter plate, add 50/zl of ~25I-labeled avidin solution to each well. The ~25I-labeled avidin solution is made by diluting the stock ~25I-labeled avidin with avidin buffer (see above) to a concentration of approximately 10,000 cpm/50/zl. As the labeled avidin degrades with age, more avidin will be required to maintain 10,000 cpm/50/~1. A 50-~1 aliquot of diluted avidin should be counted to measure total cpm//zl before each assay. Mix the contents of Incubation 1 by alternately drawing up and expelling the contents of each well with a pipet. Incubate for 1 hr at room temperature (the length of this incubation is not critical). Incubation 2. After the first incubation, transfer 100/zl from each well in the microtiter plate to a well of the Immulon II plate. Allow the samples to incubate in the coated wells for at least 4 hr at room temperature or up to 3 days at 4 °. Then aspirate the contents of the Immulon II wells and wash twice with washing buffer to remove all unbound ~zSI-labeled avidin. Quantitate the bound 125I-labeled avidin by counting. The Removawell strips of Immulon II can be broken apart for placement into a suitable carrier for the 3' counter. Data Analysis and Interpretation A standard curve (Fig. 1A) depicts uncorrected counts bound to the well versus the log of the biotin concentration. A log-logit transformation (Fig. 1B) does not consistently produce a linear dependence on the log of the biotin concentration. Competitive isotope dilution assays in liquid phase characteristically have linear log-logit plots; the fact that this assay is sequential rather than competitive and has a solid-phase binding step may be responsible for the more complex curve shape. Determination of biotin concentration graphically is acceptable, but tedious when assaying large numbers of unknowns. We use a polynomial curve-fitting program to fit the curve to r 2 > 0.99; typically a fourth or fifth degree polynomial is suitable. The assay sensitivity has typically been about 100 fmol/ml (10 fmol/ assay well) but has occasionally been 10 fmol/ml (1 fmol/assay well) with an avidin preparation of unusually high specific activity and unusually high affinity for the solid phase. The intraassay variability was determined on 10 replicates on rat urine (--1/400 dilution) in a single assay; the coeffi-
[24]
SOLID-PHASE RADIOASSAY FOR BIOTIN
231
2800
A 2400 z > 2000 in
a
1600
z
o >iI-U
1200
800
,¢ o 400
¢v I
100
I
300 BIOTIN ( f m o l / m l )
I
I
1000 log scale
2500
3.0
B c~ 2 . 0 z o
m
1.0
t,--
z
14.1 U
'~' 14.1 m
0
o -1.0; I--
9_2.o i
-3.0 ¸
I
100
lO O
2 ioo
BIOTIN f f m o l / m l , log scale FIG. 1. (A) Plot of uncorrected counts bound to the coated wells versus log of biotin concentration in the series of biotin standards. (B) Plot of logit of percent binding versus log of biotin concentration.
232
GENERALMETHODOLOGY
[24l
cient of variation was 10%. The long-term precision (interassay variability) was estimated from results of triplicate determinations from assays of a sample by three different technicians on 6 different days; the coefficient of variation (calculated from the standard deviation of the means of the triplicates) was 6%, and the intraclass correlation coefficient on that data was -0.77. We have investigated the use of an alternate coating protein (keyhole limpet hemocyanin), and the use of biotinylating agent with a longer spacer (NHS-LC-biotin, Pierce Chemical Co.). These modifications resuited in modestly higher maximum percent binding (85 versus 65%) but gave essentially the same value for biotin (i.e., total avidin-binding substances) in urine and ultrafiltrates of plasma. We have also investigated the detectability of biotin analogs in this assay. For example, a series of standard concentrations of biocytin, biotin sulfoxide, and bisnorbiotin were assayed. The biocytin standard curve was identical to the biotin standard curve, within experimental error. For the others, the point of half-maximal binding (the inflection point of the sigmoid curve) was shifted toward greater concentrations of the analog, and the true concentration of the analog would have been underestimated by a factor of 4 to 5 for both analogs. The lower binding affinities of some biotin analogs 9 may allow the solid-phase biotin (biotinylated BSA) to compete for binding to ~25I-labeled avidin in the second step and thus modify the standard curve. This observation emphasizes two limitations of the assay that may well apply to other avidin-binding assays: (1) in a complex biological mixture of biotin, biotin analogs (e.g., metabolites), and even protein-bound biotin, the measured total avidin-binding substances may not equal the stoichiometric sum of all avidin-binding substances, and (2) a more accurate measure could be made if the avidin-binding substances are separated chromatographically and each avidin-binding substance identified and quantitated versus the standard curve for that compound. A tacit assumption in previous measurements of biotin in human plasma has been that biotin is the only avidin-binding substance present (if using an avidin-binding assay) or that biotin is the only growth-promoting substance present (if using a bioassay). Using paper chromatography and a bioassay, Lee et al. l0 demonstrated the presence of substantial amounts of biotin metabolites in human urine. More recently this assumption concerning human plasma was examined by using reversed-phase HPLC to separate biotin analogs in ultrafiltrates of human plasma. Figure 2 depicts 9 N. M. Green, Ado. Protein Chem. 34, 1967 (1975). to H. Lee, L. D. Wright, and D. B. McCormick, J. Nutr. 11)2, 1453 (1972).
[24]
SOLID-PHASE RADIOASSAY FOR BIOTIN
233
100 C w 90 r,,3 z B0 Q~ 0 (/1 70 m < I-- 60 Ld -J 0 50
b
e
f
o
40 J
30 1,1
>_
F--
20
5 ¢Y
10 0
k______ 0
5
10
15
RETENTION TIME ( m i n u t e s ) FIG. 2. Chromatography of a mixture of biotin analogs. The biotin analogs are biotin l-sulfoxide (a), biotin d-sulfoxide (b), biocytin (c), biotin (d), dethiobiotin (e), and biotin methyl ester (f). Chromatography parameters were a linear gradient from 0% A to 50% A at 1.0 ml/min for 40 min. [Solution A, methanol; solution B, 10 m M potassium phosphate (pH 4.5)].
the HPLC separation of a mixture of biotin analogs; in this mixture, concentrations were sufficient (e.g., micromolar range) for detection by ultraviolet absorption. In human plasma, however, biotin and biotin analogs are present in nanomolar concentrations. Using the ]25I-labeled avidin assay to measure the fractions collected after chromatography, a substantial amount of avidin-binding substances that are not biotin were detected, u-~3 Recovery of total avidin-binding substances was 94 - 9%. Biotin accounted for only 60 - 6% (X - 1 SD) of total avidin-binding substances; four other peaks were detected that did not coelute with biocytin or biotin, the only recognized biotin vitamers which occur naturally.
n D. M. Mock, Fed. Proc., Fed. Am. Soc. Exp. Biol. 46, 1159 (1987). 12 D. M. Mock and G. L. Lankford, FASEB J. 3, A1058 (1989). t3 T. Suormala, E. R. Baumgartner, J. Bausch, W. Holick, and H. Wick, Clin. Chim. Acta 177, 253 (1988).
