UV-Traceable Reagent for Controlled Biotinylation of Antibodies

UV-Traceable Reagent for Controlled Biotinylation of Antibodies

Abstract

Biotin incorporation remains the most popular method of tagging antibodies for affinity capture and detection. In this report, we describe a UV-traceable biotinylation reagent that permits non-destructive quantification of biotin incorporation by means of a simple UV scan of the labeled antibody. The reagent was used to label a model antibody substrate (bovine IgG) in an attempt to better understand the interaction between antibody concentration and reagent equivalents required to achieve a controlled biotin molar substitution ratio (MSR). In related experiments, an IL-2 specific detector antibody was biotinylated at different molar substitution ratios to assess how the degree of incorporation affects immunoassay sensitivity within a sandwich ELISA format. Bovine IgG labeling results indicate that within a given antibody concentration range the biotin molar substitution ratio doubles as reagent equivalents double. In related experiments, the data reveal that at a fixed antibody concentration biotin labeling efficiency increases in a predictable manner expressible as a percentage of the number of reagent equivalents used in the reaction. Finally, IL-2 sandwich ELISA results demonstrate a 2.7-fold increase in detection sensitivity as a biotinylated detector antibody’s molar substitution ratio increases from 0.70 through 8.4.

Introduction

Enzyme-linked immunosorbent assays (ELISAs) are a commonly used method of detecting and quantifying antigens (1, 2). In one type of ELISA format, matched pairs consisting of a capture and biotinylated detector antibody are used to quantify antigen (3, 4). In this now widely used format, the degree of biotin incorporation on a detector antibody has been shown to play a key role in immunoassay specificity (5, 6). Although the level of biotin incorporation on a detector antibody is known to be important for assay performance, it is often ignored or overlooked by researchers because rapid and direct methods of measuring biotin incorporation have not been available until recently.

Over the years, two major indirect assays have prevailed as the standard for measuring biotin incorporation on antibodies and other proteins. The first of these assays, known as the HABA assay (2- (4’-hydrozyazobenzen benzoic acid) was developed more than 40 years ago (7). In the HABA assay, incorporation is measured when biotin or a biotin-labeled molecule displaces the weaker binding avidin- bound HABA dye into the surrounding bulk solution causing a reduction in the dye’s absorbance at 500 nm. This reduction in signal is then quantified with an externally generated avidin/biotin calibration curve. In more recent years, a similar yet more sensitive indirect fluorescent assay has been developed by Johnson et. al (8) called the FluoroReporter™ assay. In this assay, biotin binding displaces an avidin- bound quencher dye back into the bulk solution, relieving the quenching of the covalently bound fluorophore. Neither of these indirect assays actually measure the number of biotin molecules on a protein’s surface, rather they measure the number of biotin molecules capable of binding to avidin. As a consequence, both assays tend to underestimate the absolute number of incorporated biotin molecules because any two molecules spaced closer than avidin’s (or streptavidin’s) binding footprint of 34 nm² (9) will bind to only one of them. For this reason, improvements to these lengthy and complicated assayshave included either an acid hydrolysis or protease digestion step to increase access of surface bound biotin molecules (10, 11, 12, and 13). Although both of these indirect assays are still in use today they are inconvenient, destructive in nature and time consuming continuing to pose significant barriers to their routine implementation.

In this short report, we discuss a traceable biotin labeling reagent that addresses many of these shortcomings. This reagent can be used to measure biotin incorporation directly from a simple UV scan of the labeled antibody. The linker’s intrinsic high molar extinction coefficient permits sensitive and direct measurement of biotin incorporation. Using this reagent, we labeled a model antibody substrate in order to explore the interplay between antibody concentration and reagent equivalents required to achieve a controlled biotin molar substitution ratio (MSR). In related experiments, a rat anti-mouse IL-2 detector antibody was labeled at defined biotin molar substitution ratios in order to better understand the relationship between the degree of biotin substitution and relative assay sensitivity within the confines of a sandwich ELISA format.

Materials and Methods

Biotinylation of Antibodies

All antibodies in this report were biotinylated using Sulfo-ChromaLink Biotin reagent (SoluLinK, San Diego, CA). Briefly, solid antibodies were re-suspended in phosphate labeling buffer (100 mM sodium phosphate, 150 mM NaCl, pH 7.4) followed by buffer exchange into the same buffer using Zeba™ desalt spin columns (Pierce/ThermoScientific, Rockford, IL). Purified antibody concentrations were quantified by their intrinsic A280 absorbance as measured on a NanoDrop™ spectrophotometer (ThermoFisher Scientific, Wilmington, DE). Bovine IgG stock solutions (0.5, 1.0, and 2.5 mg/ml) were prepared from a more concentrated stock (13.9 mg/ml) in phosphate labeling buffer. Sulfo-ChromaLink Biotin labeling reagent, an amorphous solid (>95% pure as confirmed by ¹HNMR, reverse phase HPLC 95%, and mass spectrometry (M.W. 912.96)) was prepared by dissolving known quantities into 100 mM sodium phosphate, 150 mM NaCl at pH 7.4. Model bovine IgG substrate solutions (100 µl) were biotinylated in triplicate by addition of equal volumes (5 µl) of Sulfo-ChromaLink Biotin reagent corresponding to 5-, 10- or 20-fold mole equivalents for 2 hours at room temperature. Similarly, rat anti-mouse IL-2 specific detector antibody (100 µl at 1 mg/ml) (BioLegend, San Diego, CA) was biotinylated by incubation of the antibody with 0, 2, 4, 10, and 28-fold excess of Sulfo-ChromaLink Biotin. After labeling, excess reagent was desalted using Zeba™ spin columns equilibrated in PBS. Aliquots from the purified labeling reactions (2 µl) were scanned on a NanoDrop™ spectrophotometer (220-420 nm) and the resulting A280 and A354 values used to calculate biotin molar substitution ratios.

IL-2 Sandwich ELISA

Initially, IL-2 specific capture and detection monoclonal antibodies were titrated against each other in a preliminary checkerboard optimization process. Sandwich ELISAs were performed by coating a 96-well plate (100 µl per well) with rat anti-mouse IL-2 capture antibody (1.25 µg/ml) for 30 minutes at 37°C. Unbound capture antibody was washed away with 10 mM sodium phosphate, 250 mM NaCl, 0.05% Tween-20, pH 7.2 (3X) followed by careful blotting of plates between washes using paper towels. Capture antibody coated wells were then blocked for 2 h at room temperature using 200 µl blocking buffer (10 mM sodium phosphate, 250 mM NaCl, 1% BSA, 0.05% Tween-20, pH 7.2). All ELISA capture plates were sealed and stored at 4°C until used. Lyophilized carrier-free recombinant IL-2 antigen (BioLegend, San Diego, CA) was re-suspended in PBS at 100 ng/ml and two-fold serially diluted across the plate (100 ng/ml to 19.5 ng/ml) using replicates. After incubating for 1 hour at room temperature unbound antigen was washed away. Biotinylated IL-2 detector antibody (100 µl at 0.5 µg/ml) was then incubated for 1 hour at room temperature followed by additional washes. All wells were then incubated for 1 h with horseradish peroxidase-streptavidin at 1.5 µg/ml (KPL, Gaithersburg, MD). Immune complexes were detected by addition of 100 µl tetramethybenzidine substrate (Pierce/ThermoScientific, Rockford, IL). Substrate development was stopped after 15 minutes using 1M H2SO4. ELISA signals were read at 450 nm on a SpectraMax™ Plus™ 384 plate reader (Molecular Devices, Sunnyvale, CA) and the data plotted using the instrument’s four-parameter curve fitting software routine.

UV Spectra and Biotin Molar Substitution Ratios

All UV spectra were obtained from a calibrated NanoDrop™ ND-1000 spectrophotometer employing a 1 mm path length. Data were collected by scanning 2 µl aliquots using the instrument’s Protein A280 menu (340 nm normalization feature turned-off). The A280 and A354 from each sample’s absorption spectrum were then used to calculate the biotin molar substitution ratio as follows:

1. Ac₂₈₀ = A₂₈₀ - ( A₃₅₄ × 0.23)
2. moles of antibody = V (ml ) × [( Ac ₂₈₀) / (E1%)] × [(10 mg/ml ) / (1000 mg/ml ) / M .W . IgG]
3. moles of biotin = ( A₃₅₄ / ε ₃₅₄) × [V (ml ) / 1000 ( ml/l )]
4. biotin molar substitution ratio = moles of biotin / moles of antibody

Equation 1 is used to measure the corrected A280 of the labeled antibody sample. It employs a correction factor that subtracts the A280 contribution originating from the labeling reagent itself. This correction factor was originally determined by scanning an ethanolamine quenched Sulfo-ChromaLink Biotin solution (35 µM, 1 mm path length) in phosphate buffer and expressing the A280 correction factor as a fraction of the quenched reagent’s absorbance at 354 nm. Equation 2 is used to quantify the moles of antibody from the labeled antibody’s known sample volume (V) in milliliters, the corrected A280 as well as the antibody’s mass extinction coefficient (E1%) and its molecular weight in Daltons. Equation 3 is used to quantify the moles of antibody incorporated biotin using the reagent’s quenched molar extinction coefficient (ε354 =29,000), its sample’s volume (ml), and its A354. Equation 4 then combines the results of equation 2 and 3 to obtain the biotin molar substitution ratio of the labeled antibody.

Results and Discussion

The chemical structure of the biotin labeling reagent (Sulfo-ChromaLink™ Biotin) used in this report is illustrated in Figure 1, Panel A. As seen in the figure, this compound possesses a water-soluble N-hydroxysulfosuccinimidyl ester that acylates antibody lysine residues under mild aqueous buffer conditions (100 mM sodium phosphate, 150 mM NaCl, pH 7.4). The UV traceable portion of the linker consists of an embedded bis-arylhydrazone chromophore attached through a PEG3 spacer to the biotin moiety. The chromophore’s high molar extinction coefficient (ε354 = 29,000) originates from the extended conjugation of the bis-arylhydrazone. Figure 1, Panel B illustrates the unquenched solution phase UV absorption spectrum of the labeling reagent whereas Figure 1, Panel B is the spectrum of a biotin labeled antibody.

Using this traceable biotin reagent we labeled a model antibody substrate (bovine IgG) at three different protein concentrations to better understand the interplay between initial antibody concentration and the number of reagent mole-equivalents required to achieve a given biotin molar substitution ratio. Aliquots of bovine IgG at 3 different concentrations (0.5, 1.0, and 2.5 mg/ml) were labeled using 5, 10, and 20-fold mole excess reagent in triplicates. After desalting all 27 reactions using Zeba™ spin columns to remove excess reagent, the UV spectrum (220-420 nm) of each sample was acquired. The resulting A280 and A354 absorbance values along with measured sample volumes was then used to calculate biotin molar substitution ratios. A bar graph summarizing these biotin incorporation data is illustrated in Figure 2 below. As seen from the graphical data, at lower fixed antibody concentrations (0.5 and 1 mg/ml) the biotin molar substitution approximately doubles as reagent equivalents double. For example at 0.5 mg/ml IgG, we observe a 2.3-fold and 1.9-fold increase in biotin MSR as we double the number of equivalents from 5 to 10-fold and again from 10 to 20-fold, respectively. At 1 mg/ml IgG, the corresponding increases are 2.1 and 1.90-fold. While at 2.5 mg/ml the increase is 2.0 and 1.56-fold, respectively. The observed drop in 2-fold proportionality at 2.5 mg/ml and 20-equivalents implies label saturation of the antibody. This saturation phenomenon has also been observed by Smith (11) when attempting to label other types of antibodies using biotin NHS esters. We surmise as have others (6), that label saturation occurs because of reduced antibody solubility (aggregation/precipitation) at the highest molar substitution ratios.

The experimental results in Figure 2 also reveal another interesting relationship between antibody concentration and molar substitution ratio. Namely, that biotin labeling efficiency (at a fixed excess of reagent equivalents) is predictable across different antibody concentrations. That is, labeling efficiency is expressible as a simple percentage of the number of input equivalents used to modify the antibody. For example, at a fixed 5-fold mole excess of Sulfo-ChromaLink Biotin the labeling efficiency averages 29 ± 2.3% the number of input equivalents at all three concentrations tested. At a fixed 10-fold mole excess the labeling efficiency increases to 42 ± 1.7% at all three IgG concentrations, while at 20 equivalents the labeling efficiency rises to 66 ± 0.4% at 0.5 and 1.0 mg/ml but declines to 52% at 2.5 mg/ml as label saturation is reached. The observed decline at 2.5 mg/ml is consistent with label saturation causing aggregation and precipitation of the antibody at highest substitution ratios.

Using this predictable relationship, we biotinylated a rat anti-mouse IL-2 antibody (1 mg/ml) at various mole-equivalents in order to generate a defined set of biotin molar substitution ratios with the same antibody. Overlaid absorption spectra of this IL-2 detector antibody are illustrated in Figure 3 (A). This detector antibody labeled at three substitution levels (MSR 0.7, 3.7 and 8.38) was then evaluated for relative assay sensitivity in an IL-specific sandwich ELISA. The resulting dose-response curves are illustrated in Figure 3 (B). The data reveal a gradual increase in IL-2 detection sensitivity as biotin molar substitution increases through the series. The leftward shift in the inflection point of the response curves reveals a 2.7-fold increase in IL-2 detection sensitivity as the biotin MSR progress from 0.7 to 8.4.

In summary, we demonstrate the utility of using a UV traceable biotin labeling reagent to achieve controlled and quantifiable biotin incorporation with antibodies. The reagent’s traceable nature allows rapid and non-destructive measurement of biotin incorporation; greatly simplifying immunoassay standardization. These experiments reveal that antibody biotin labeling efficiency is predictable and highly reproducible under well defined reaction conditions. Lastly, IL-2 sandwich ELISA results confirm that biotin molar substitution ratios do affect relative assay sensitivity as demonstrated in the specific ELISA format chosen for this study.

ACKNOWLEDGMENTS

We wish to thank Jamie McDonald for her insightful comments and editorial assistance with the manuscript.

REFERENCES

1. Voller, A. 1978. The enzyme-linked immunosorbent assay (ELISA) (theory, technique and applications). Ric Clin Lab 8(4): 289-98.
2. Lequin, R. M. 2005. Enzyme immunoassay (EIA)/enzyme-linked immunosorbent assay (ELISA). Clin Chem 51(12): 2415-8.
3. Yolken, R.H., Leister, F.J., Whitcomb, L.S, and Santosham, M. 1983. Enzyme immunoassays for the detection of bacterial antigens utilizing biotin-labeled antibody and peroxidase biotin-avidin complex. J Immunol Methods, Feb 11; 56(3):319-27
4. Yolken, R.H. 1986. Enzyme immunoassays for the detection of microbial antigens and prospects for improved assays. Yale J Biol Med. 1986 Jan–Feb; 59(1): 25–31.
5. Høyer-Hansen, G., Hamers, M.J.A.G., Pedersen, A.N., Nielsen, H.J., Brünner, N., Danø, K. & R.W. Stephens. 2000. Loss of ELISA specificity due to biotinylation of monoclonal antibodies. J. Immunol. Meth., 235:91-99.
6. Wadsley, J.J., and Watt, R.M. 1987. The effect of pH on the aggregation of biotinylated antibodies and on the signal-to-noise observed in immunoassays utilizing biotinylated antibodies. J. Immunol. Methods. 1987,103:1-7.
7. Green, N.M. 1965. A spectrophotometric assay for avidin and biotin based on binding of dyes by avidin. Biochem.J. 94:23C-24C.
8. Batchelor, R. H., Sarkez, A., Cox, W.G., and I. Johnson. 2007. Fluorometric assay for quantitation of biotin covalently attached to proteins and nucleic acids. Biotechniques 13:543-546.
9. Weber, P.C., Ohlendorf, D.H., Wendoloski, J.J., and F.R. Salemme. 1989. Structural origins of high- affinity biotin binding to streptavidin. Science 6;243(4887):85-8
10. Hermanson, G.T. 2008. Bioconjugate Techniques (2nd edition): 921-923
11. Smith, G.P. 2006. Kinetics of amine modification of proteins. Bioconjugate Chem. 17, 501- 506.
12. Bayer, E. A., and M. Wilcheck. 1990. Protein biotinylation. Methods Enzymol.184, 138-160
13. Rao, S.V., Anderson, K.W., and L.G. Bachas. 1997. Determination of the extent of protein biotinylation by fluorescence binding assay. Bioconjugate Chem. 8, 94-9