Anal. Chem. 2006, 78, 4501-4508
Isolation and Characterization of a Thermally Stable Recombinant Anti-Caffeine Heavy-Chain Antibody Fragment Ruth C. Ladenson, Dan L. Crimmins, Yvonne Landt, and Jack H. Ladenson*
Department of Pathology and Immunology, Division of Laboratory Medicine, Washington University School of Medicine, 660 South Euclid Avenue, Campus Box 8118, St. Louis, Missouri 63110
We have isolated and characterized a caffeine-specific, heavy-chain-only antibody fragment (VHH) from llama that is capable of being utilized to analyze caffeine in hot and cold beverages. Camelid species (llama and camel) were selected for immunization because of their potential to make heat-stable, heavy-chain-only antibodies. Llamas and camels were immunized with caffeine covalently linked to keyhole limpet hemocyanin, and recombinant antibody techniques were used to create phage displayed libraries of variable region fragments of the heavy-chain antibodies. Caffeine-specific VHH fragments were selected by their ability to bind to caffeine/bovine serum albumin (BSA) and confirmed by a positive reaction in a caffeine enzyme-linked immunosorbent assay (caffeine ELISA). One of these VHH fragments (VSA2) was expressed as a soluble protein and shown to recover its reactivity after exposure to temperatures up to 90 °C. In addition, VSA2 was able to bind caffeine at 70 °C. A competition caffeine ELISA was developed for the measurement of caffeine in beverages, and concentrations of caffeine obtained for coffee, Coca-Cola Classic, and Diet Coke agreed well with high performance liquid chromatography (HPLC) determination and literature values. VSA2 showed minimal cross reactivity with structurally related methylxanthines. There is increasing interest within the general population to monitor nutrition and limit unwanted components in food, whether natural or added. Specifically, caffeine, a potent central nervous system stimulator, is a compound frequently avoided, as reflected in the increasing array of decaffeinated beverages available in restaurants and grocery stores. Currently, caffeine is measured by a variety of methods,1 including ultraviolet spectroscopy, thin-layer chromatography, gas chromatography, high-performance liquid chromatography, and capillary electrophoresis, none of which is applicable to home or restaurant use. Immunoassays2,3 have been described for use in * Corresponding author. Phone: 314-454-8436. Fax: 314-454-5208. E-mail:
[email protected]. (1) Hurst, W. J.; Martin, R. A., Jr.; Tarka, S. M., Jr. In Caffeine; Spiller, G. A., Ed.; CRC Press: New York, 1998; pp 13-33. (2) Zysset, T.; Wahlla¨nder, A.; Preisig, R. Ther. Drug Monit. 1984, 6, 348354. (3) Pearson, S.; Smith, J. M.; Marks V. Ann. Clin. Biochem. 1984, 21, 208212. 10.1021/ac058044j CCC: $33.50 Published on Web 04/26/2006
© 2006 American Chemical Society
biological fluids but are also not applicable to home or restaurant use. We envisioned that a simple method to measure caffeine, even in hot beverages, such as coffee, would be of value to individuals and institutions wanting to verify the absence of caffeine. One potentially useful approach to this problem would be a qualitative immunoassay in a “dipstick” format, as is used in home pregnancy kits. Caffeine-specific antibodies both polyclonal and monoclonal are commercially available; however, such antibodies have been shown to irreversibly denature at high temperatures4 and would have difficulties in measuring caffeine in hot coffee. Members of the Camelidae family have been shown to produce a form of antibody that is devoid of light chains.5 The variable domains of heavy-chain-only antibodies (VHH) have been cloned from peripheral blood lymphocytes of camels (Camelus dromedarius)6 and llamas (lama glama).7 These single-domain antibody fragments can refold8 and maintain functionality after thermal denaturation.4,9 In some cases, VHH fragments can bind specifically at temperatures up to 90 °C.4 Moreover, hapten-specific VHH fragments have been reported,10,11 so we decided to develop a VHH that is reactive to caffeine. We have successfully developed a thermally stable, caffeine-specific VHH from llama and shown that it can be used to accurately determine the caffeine concentration in beverages. (4) van der Linden, R. H. J.; Frenken, L. G. J.; de Geus, B.; Harmsen, M. M.; Ruuls, R. C.; Stok, W.; de Ron, L.; Wilson, S.; Davis, P.; Verrips, C. T. Biochim. Biophys. Acta 1999, 1431, 37-46. (5) Hamers-Casterman, C.; Atarhouch, T.; Muyldermans, S.; Robinson, G.; Hamers, C.; Bajyana Songa, E.; Bendahman, N.; Hamers, R. Nature 1993, 363, 446-448. (6) Arbabi Ghahroudi, M.; Desmyter, A.; Wyns, L.; Hamers, R.; Muyldermans, S. FEBS Letters 1997, 414, 521-526. (7) van der Linden, R.; de Geus, B.; Stok, W.; Bos, W.; van Wassenaar, D.; Verrips, T.; Frenken, L. J. Immunol. Methods 2000, 240, 185-195. (8) Pe´rez, J. M. J.; Renisio, J. G.; Prompers, J. J.; van Platerink, C. J.; Cambillau, C.; Darbon, H.; Frenken, L. G. J. Biochemistry 2001, 40, 74-83. (9) Ewert, S.; Cambillau, C.; Conrath, K.; Plu ¨ ckthun, A. Biochemistry 2002, 41, 3628-3636. (10) Frenken, L. G. J.; van der Linden, R. H. J.; Hermans, P. W. J. J.; Bos, J. W.; Ruuls, R. C.; de Geus, B.; Verrips, C. T. J. Biotechnol. 2000, 78, 1121. (11) Spinelli, S.; Frenken, L. G. J.; Hermans, P.; Verrips, T.; Brown, K.; Tegoni, M.; Cambillau, C. Biochemistry 2000, 39, 1217-1222. (12) Cook, C. E.; Tallent, C. R.; Amerson, E. W.; Myers, M. W.; Kepler, J. A.; Taylor, G. F.; Christensen, H. D. J. Pharmacol. Exp. Ther. 1976, 199 (3), 679-686.
Analytical Chemistry, Vol. 78, No. 13, July 1, 2006 4501
Figure 1. Caffeine carboxylate derivative and its conjugation to BSA. (A) Chemical structure of caffeine carboxylate: 7-(5-carboxypentyl)1,3-dimethylxanthine. (B) MALDI-TOFMS spectra of unconjugated control BSA, top panel, and purified conjugated BSA, bottom panel. The +1 m/z region is shown for both and suggests the reaction product as a species of ∼10 mol of caffeine carboxylate/mol of BSA, bottom panel.
EXPERIMENTAL SECTION Caffeine Derivative. We had a caffeine carboxylate derivative, 7-(5-carboxypentyl)-1,3-dimethylxanthine (Figure 1A), synthesized as described by Cook et al.,12 at Daniels Fine Chemicals (Edmonton, Alberta). This derivative facilitates conjugation of caffeine to proteins or peptides via their amino groups by standard EDC chemistry. The caffeine derivative has a formula weight of 294.31 g/mol, melting point of 129 °C, and the expected 1H NMR spectra. Immunogen and Screening Reagent Preparation and Characterization. The caffeine derivative was covalently linked to Mariculture keyhole limpet hemocyanin (KLH) following the Imject Immunogen EDC (1-ethyl-3-{3-dimethylaminopropyl}carbodiimide hydrochloride) Conjugation Kit protocol from Pierce Chemical (Rockford, IL). The protein conjugate was desalted and buffer-exchanged into 0.083 M sodium phosphate, pH 7.2, containing 0.9 M NaCl prior to use as an immunogen. We also prepared a BSA conjugate using the EDC procedure described above to be used as a screening reagent during antibody assessment following immunization. Production of a Soluble Caffeine/Biotinylated Nonapeptide. To be able to connect biotin to caffeine, we designed the following nonapeptide, long chain biotin-Gly-Gly-Ser-Gly-Gly-SerGly-Gly-Lys (amide), and had it synthesized at Biomolecules Midwest, Inc. (Waterloo, IL). The crude peptide was purified following resin cleavage by preparative RP-HPLC, and peak fractions were pooled and lyophilized. This nonapeptide was designed to allow coupling of biotin to the caffeine derivative so as to be quite soluble in aqueous solution 4502
Analytical Chemistry, Vol. 78, No. 13, July 1, 2006
with no near-UV absorbance, so the adduct could readily be followed. It was also flexible and allowed for a reproducible sitespecific reaction. Furthermore, the C terminus was amidated to avoid intramolecular linkage mediated by EDC. A soluble caffeine/biotinylated peptide was formed by the conjugation of the carboxyl moiety of the caffeine derivative to the biotinylated nonapeptide at the -NH2 group of the C-terminal lysine using EDC chemistry. Immunization Protocol. Two types of camelids were immunized. Three llamas (Fanfare, Virtual, and Very Senorita) were housed at Triple J Farms/Kent Laboratories, Bellingham, WA, and two camels (Assab and Massawa) resided at the Veternary Research Station, in Hagaz, Eritrea. Each of the llamas received an initial intramuscular (IM) injection of 250 µg of caffeine/KLH in complete Freund’s adjuvant (Sigma-Aldrich, St. Louis, MO), followed by three boosts of 250 µg caffeine/KLH in incomplete Freund’s adjuvant (Sigma-Aldrich) at monthly intervals. The same schedule was used for the camels with MPL (monophosphoryl lipid A) + TDM (trehalose dicorynomycolate) + CWS (cell wall skeleton) (Sigma-Aldrich, M6661) used as adjuvant for all injections. Test bleeds were taken at day 0 (preimmune) and 14 days after each of the first three immunizations (2, 6, and 10 weeks). One unit of blood (∼500 mL) collected with sodium citrate as anticoagulant, including phosphate and dextrose, (CPD) was taken one week after the final boost. Construction of Phage Displayed VHH Libraries. The units of blood from llamas and camels were processed within 24 and
Table 1. Summary of the Three ELISA Formats Useda step
phage caffeine ELISA
standard caffeine ELISA
coating blocking primary antibody
caffeine/BSA TBS/BSA M13-phage-displayed llama VHH
caffeine/BSA TBS/BSA llama VHH
secondary antibody reporter antibody
none mouse anti-M13/horseradish peroxidase conjugate ABTS
mouse anti-E-tag goat anti-mouse IgG/alkaline phosphatase conjugate p-npp
substrate
competition caffeine ELISA caffeine/BSA TBS/BSA llama VHH ( caffeine standards or samples mouse anti-E-tag goat anti-mouse IgG/alkaline phosphatase conjugate p-npp
a Abbreviations: TBS, Tris-buffered saline; BSA, bovine serum albumin; ABTS, [2,2′-azino-di-[3-ethylbenzthiazoline-6-sulfonate]; p-npp, para-nitrophenylphosphate
48 h, respectively. Peripheral blood lymphocytes (PBLs, 107-108 cells) from each animal were isolated on density gradients using Histopaque 1077 and 1088 (Sigma-Aldrich).13 Messenger RNA (mRNA) was isolated from the purified PBLs using the Fast Track 2.0 Kit (Invitrogen Corporation, Grand Island, NY) and converted to first-strand DNA using the oligo dT primer of the cDNA Cycle Kit (Invitrogen Corporation). The variable regions of heavy-chain immunoglobulins (VHH) were amplified by nested PCR, as described by Ghahroudi et al.,6 using the initial primer pair CH2FORTA4 (5′-CGCCATCAAGGTACCAGTTGA-3′) and VHBACKA6 (5′-GATGTCCAGCTGCAGGCGTCTGG (A\G) GGAGG-3′). The products were analyzed by agarose gel electrophoresis using 1.5% (w/v) NuSieveGTG. Gel plugs from the bands near 600/680 base pairs (bp) were used as templates for the secondary PCR reaction. This was performed with primers6 VHBACKA4 (5′-CATGCCATGACTCGCGGCCCAGCCGGCCATGGCCGA (G\T) GT (G\C) CAGCT-3′) and VHFOR36 (5′ GGACTAGTGCGGCCGCGTGAGGAGACGGTGACCTG-3′), which incorporate the underlined SfiI and NotI restriction sites on the 5′ and 3′ ends, respectively. The products were run on 1.25% (w/v) NuSieve 3:1 preparative agarose gels, and the band at 450 bp was extracted using an Ultrafree-DNA filter system (Millipore, Bedford, MA). Each band is a pool of total amplified VHH DNA. We next incorporated the amplified VHH DNA from each animal into an M13 phagemid vector, pCANTAB 5E (Amersham Biosciences Corporation, Pisataway, NJ). This vector allows the covalent linkage of VHH protein to one of the phage coat proteins. After growth and assembly of the phage, the VHH protein will be displayed on the outside of the phage particle. In addition, a sequence of 16 amino acids designated as the E-tag is added to the C terminus of the VHH molecule by the vector. Before incorporation into the vector, the VHH DNA pools were digested by SfiI and NotI restriction enzymes and purified on StrataClean resin (Stratagen, La Jolla, CA). A 50-ng aliquot of each of the five restriction enzyme digested VHH DNA pools was ligated into 125 ng of SfiI/NotI-cut pCANTAB 5E using T-4 DNA ligase (Invitrogen). The ligated material was desalted and transformed into electrocompetent Escherichia coli XL1-Blue MRF′ cells (Stratagene). Six electroporations were done for each of the five ligations and pooled to form five separate bacterial libraries, one from each of the immunized animals. Library size was determined by plating an aliquot of the library and counting the resulting (13) Zweygarth, E. Zbl. Vet. Med. B 1984, 31, 786-789.
colonies. An extrapolation to the total volume of the library gives the total library size in colony forming units (cfu). Selection of Caffeine Specific VHH Fragments. Phagedisplayed caffeine-specific VHH fragments were isolated by processes called phage rescue and panning following the protocols of Harrison et al.14 The transformed bacteria are grown to produce phage particles with displayed VHH fragments, which are then incubated with immobilized caffeine/BSA. A washing step removes unbound phage, leaving those phages displaying caffeinespecific VHH fragments bound to the caffeine/BSA. The bound phages are eluted with 100 mM triethylamine and after neuralization with 1 M Tris-HCl, pH 7.4, are used to infect bacteria. The bacteria are then plated at various dilutions. This represents one round of panning. This process can be repeated with each round of panning, allowing for further enrichment for specific binders. Single colonies from the plates from each round of panning were selected and grown for phage production as described.14 The success of the panning procedure to isolate specific binders was assessed by measuring binding of phage displayed VHH fragments to caffeine/BSA immobilized in the wells of microtiter plates (phage caffeine ELISA) as described below and in Table 1. Phage Caffeine ELISA. Microtiter plates were coated overnight at 4 °C with caffeine/BSA at a concentration of 2 µg/mL in 50 mM sodium phosphate, pH 7.2, containing 150 mM NaCl (PBS) (coating step). The wells were then blocked with blocking buffer (PBS containing 2% nonfat dry milk (NFDM)) for 2 h at 37 °C (blocking step). Phages from individual colonies diluted 1:2 in blocking buffer were added and incubated for 90-120 min at 37 °C (primary antibody step). Bound phages were detected by the addition of mouse anti-M13-antibody/horseradish peroxidase conjugate (Amersham Biosciences Corporation) diluted 1/1000 in 20 mM Tris, pH 7.2, containing 150 mM NaCl and 2% BSA (TBS/BSA) (reporter antibody step) and ABTS substrate (Kirkagaard and Perry, Gaithersburg, MD) (substrate step). The wells were washed with Tween/saline (0.05% Tween/150 mM NaCl) between steps, and the reaction was monitored at A405 (see also Table 1.) Standard Caffeine ELISA. After caffeine-specific clones had been identified, soluble VHH proteins not linked to phage particles were expressed in E. coli.14 Binding activity of the soluble VHH proteins was assessed in a standard caffeine ELISA (Table 1). (14) Harrison, J. L.; Williams, S. C.; Winter, G.; Nissim, A. In Methods in Enzymology; Abelson, J. N., Ed.; Academic Press: San Diego, 1996; Vol 267, pp 83-109.
Analytical Chemistry, Vol. 78, No. 13, July 1, 2006
4503
Microtiter plates were coated and blocked as above, except that TBS/BSA was used in the blocking step. The VHH proteins serially diluted in 20 mM Tris, pH 7.2, containing 150 mM NaCl and 0.1% Tween (TBS/BSA/Tween) were added and incubated for 90 min at 37 °C (primary antibody step). Bound VHH fragments were detected by the addition of a 1/1000 dilution in TBS/BSA/Tween of mouse anti-E-tag antibody (Amersham Bioscience Corporation) for 90 min at 37 °C (secondary antibody step), followed by the addition of a 1/1000 dilution in TBS/BSA/Tween of goat antimouse IgG/alkaline phosphatase conjugate for 90 min at 37 °C (reporter antibody step) before the addition of para-nitrophenyl phosphate substrate (p-npp) (Sigma-Aldrich) (substrate step). The wells were washed and monitored as above. Competition Caffeine ELISA. Binding of VHH proteins to unconjugated caffeine was also assessed by a competition assay (competition caffeine ELISA, Table 1) in which 0-800 µg/mL caffeine (Sigma-Aldrich) was added as a competitor along with the VHH during the primary antibody step of the standard caffeine ELISA. All other steps are the same as for the standard caffeine ELISA. A405 readings with and without competitor were compared. Selection of VSA2. Four representative clones (two llama and two camel) showed the expected immunoreactive competition with standard solutions of caffeine. Larger amounts of soluble antibodies from these clones were purified from the periplasmic fraction of 1-L cell cultures14 and further purified using an anti-E-tag (Amersham Biosciences Corporation)/Sepharose affinity column following the manufacturer’s protocol. The antibodies were assessed for thermal stability and reactivity. Clone VSA2, which was isolated from the llama Very Senorita library showed the best protein expression and the greatest degree of heat stability and was, thus, selected for further expansion and characterization. The heat stability was assessed by randomly picking two to three clones for each library and testing their ability to bind caffeine after the partially purified antibody was incubated at 70 °C for 20 min (data not shown). Production and Purification of VSA2. VSA2 was cloned into pET 28a vector (Novagen, Madison, WI) at the EcoR1 site and transformed into chemically competent E. coli BL21 Star (DE3) pLysS cells (Invitrogen Corporation). This vector incorporates a His tag sequence and 19 additional amino acids encompassing the T7 promoter tag at the N terminus. The expressed VSA2 was purified from lysed cells utilizing NiNTA agarose (Qiagen, Valencia, CA). Typically, 1 L of culture yielded 4 mg of VHH protein. After thrombin cleavage to remove the His tag, the VSA2 was dialyzed against PBS and stored in aliquots at -20 °C. DNA and Amino Acid Sequence of VSA2. VHH cDNA from the VSA2 clone in pCANTAB 5E was amplified using an ABI PRISM Big Dye Terminator Cycle Sequencing Kit v3.1 (Applied Biosystems, Foster City, CA) and pCANTAB 5E (sequencing primers S1 and S6 (Amersham Biosciences Corporation)and sequenced by an ABI automatic DNA sequencer. Data analysis was performed using Vector NTI software (Informax, North Bethesda, MD). VHH cDNA from the VSA2 clone in the pET 28 vector was amplified and sequenced as above, but using the T7 promoter primer (Promega, Madison, WI). The amino acid sequence was deduced from the DNA sequence. Characterization of VSA2. Purified VSA2 (0.1-0.5 µg) was run in three lanes on 10-20% SDS-PAGE. One lane was stained 4504
Analytical Chemistry, Vol. 78, No. 13, July 1, 2006
with Coomassie blue for total protein visualization. The two other lanes were transferred to a PVDF membrane15 and probed with goat anti-llama IgG (Bethyl Laboratories, Inc., Montgomery, TX) or mouse anti-E-tag, followed by rabbit anti-goat IgG or goat antimouse IgG alkaline phosphatase conjugate (Sigma-Aldrich), respectively. All antibodies were diluted 1/1000 in TBS/BSA. The signal was developed with BCIP/NBT substrate (Kirkagaard and Perry, Gaithersburg, MD). Molecular weight markers were run in tandem with each of the VSA2 lanes. In addition, the molecular weight of the purified VSA2 was determined by MALDI-TOFMS with an accuracy of ∼1 part/500015 and compared to the calculated value from the deduced sequence. Cross Reactivity of Common Caffeine Competitors. Two additional methylxanthines with structures similar to caffeine, theobromine, and theophylline were tested for cross reactivity with VSA2. The competition caffeine ELISA described above and in Table 1 was run using theobromine and theophylline (SigmaAldrich) as competitors at concentrations up to 250 µg/mL and compared to competition by caffeine. Thermal Stability and Reactivity. The thermal stability of VSA2 was determined and compared to that of two commercial mouse anti-caffeine monoclonal antibodies designated M1 and M2 (Catalog Nos. C0110-07 and C0110-06, US Biological, Swampscott, MA). Antibody solutions of 2 µg/mL VSA2, 5 µg/mL M1, and 2 µg/mL M2 were prepared in TBS containing 0.1% Tween. The solutions were incubated for 20 min at temperatures ranging from room temperature (RT) to 90 °C and then reequilibrated to RT before measuring residual caffeine-binding activity. For VSA2, the standard caffeine ELISA protocol was followed. For the commercial mouse monoclonal antibodies, the secondary antibody step was omitted. The thermal reactivity of VSA2 was determined. VSA2 and the soluble caffeine/biotinylated nonapeptide described above were mixed and incubated in a 70 °C water bath for 20 min. Streptavidincoated magnetic beads (Dynabeads, MyOne, Streptavidin, Dynal Biotech Inc., Lake Success, NY) were added, and the mixture was incubated at 70 °C to allow the caffeine/biotinylated nonapeptide and any VSA2 bound to it to attach to the beads. The beads were pelleted using a magnetic device according to the manufacturer’s recommendation to separate the bound from the unbound VSA2. The amount of unbound VSA2 remaining in the supernatant was determined by comparison to a standard curve of known VSA2 concentrations in a standard caffeine ELISA (Table 1). Determination of Caffeine Concentrations in Beverages by Competition Caffeine ELISA. Regular and decaffeinated coffee (Seattle’s Best) were obtained from Aramark (St. Louis, MO). Regular and caffeine-free Coca-Cola Classic and Diet Coke were purchased from vending machines at Barnes-Jewish Hospital (St. Louis, MO). Stock solutions of caffeine (Sigma-Aldrich, reference standard C1778) at 5 mg/mL were prepared in either TBS, decaffeinated coffee, caffeine-free Coca-Cola Classic or caffeine-free Diet Coke. Caffeine standards from 0 to 800 µg/mL were prepared from the stock caffeine solutions using the above as diluents, except that TBS was used as the buffer diluent. Caffeinated beverages were assayed neat and serially diluted 2-fold. Decaffeinated coffee, (15) Dieckgraefe, B. K.; Crimmins, D.L; Landt, V.; Houchen, C.; Anant, S.; Porsche-Sorbet, R.; Ladenson, J. H. J. Invest. Med. 2002, 50, 421-434.
caffeine-free Coca-Cola Classic, or caffeine-free Diet Coke were used as diluent for coffee, Coca-Cola Classic or Diet Coke, respectively. Decaffeinated coffee was assayed neat and compared to standards diluted in TBS. The standards and beverages were incubated for 30 min at RT with equal volumes of VSA2 (2 µg/ mL) in 40 mM Tris, pH 7.2, containing 300 mM NaCl, 4% BSA, and 0.1% Tween-20. The preincubated mixtures were added to the standard caffeine ELISA (Table 1) at the primary antibody step. The assay was very matrix-dependent, so samples were compared only to standards prepared in the same diluent. A logit/log transformation was used to linearize the data. Determination of Caffeine Concentrations in Beverages by C18 RP-HPLC. Theophylline and caffeine standards having nominal concentrations of 1 mg/mL in methanol were purchased from Sigma-Aldrich. A Clipeus C18, 5 µm, 150 × 4.6 mm column (P. J. Colbert Associates, Inc., St. Louis, MO) was operated isocratically at 1 mL/min in 25% v/v methanol at 30 °C. Runs were for 12 min with a theophylline retention time of ∼5 min and caffeine at ∼7 min at 270 nm. It was readily possible to detect an injected amount of 0.05 µg for these methylxanthines. RESULTS AND DISCUSSION Immunoassays have been developed for the detection of a large number of substances both quantitatively and qualitatively. Antibodies can have very specific and sensitive binding to their antigens and are used in many immunoassay formats. Conventional IgG antibodies are composed of two identical heavy and light chains, each chain having a variable and a constant region. The interaction of the heavy and light variable regions forms the three-dimensional site of antigen binding. This can be irreversibly disrupted by denaturating conditions, such as chaotropic reagents or high temperatures. In 1993, Hamers-Casterman et al.5 published the first description of a unique form of antibody that consists of only two heavy chains. These heavy-chain-only antibodies are naturally occurring in camelids along with conventional antibodies. Because interaction of the heavy and light polypeptide chains is not required for binding of antigens by heavy-chain-only antibodies, they are inherently more stable and tend to refold to their functional form after denaturation far more readily than conventional antibodies.4,9 Indeed, this form of antibody seemed well-suited to an immunoassay format involving elevated temperatures such as would occur with hot beverages. Therefore, we decided to immunize camels and llamas as a potential source of heavy-chain antibodies for use in a caffeine immunoassay. The caffeine derivative we utilized was successfully conjugated to KLH based on its λmax of 275 nm, as compared to 280 nm for KLH alone. The massive size of KLH (up to 1.3 × 107 Da) precluded further characterization. Successful conjugation of caffeine to BSA was readily apparent, as exemplified by both a retention time increase and peak broadening of the BSA adduct in C18 RP-HPLC and slower and broader electrophoretic migration in native gel electrophoresis (data not shown). Last, the matrix assisted laser desorption ionization time-of-flight mass spectroscopy (MALDI-TOFMS) spectra (Figure 1B) for the unconjugated (top) and conjugated (bottom) BSA demonstrate covalent coupling of the caffeine carboxylate derivative to BSA. We estimate ∼10 mol of covalently linked caffeine to each mole of BSA.
Figure 2. C18 RP-HPLC purification of the caffeine carboxylate/ nonapeptide conjugate. Top panel, A275, and bottom panel, A214, show elution of nonapeptide, caffeine derivative, and caffeine/nonapeptide conjugate. Note the unreacted peptide absorbs essentially only at 214 nm, whereas the caffeine derivative and the conjugate absorb at both 275 and 214 nm. Fractions with elution times from 42 to 45 min were pooled.
The caffeine derivative was successfully coupled to the biotinylated nonapeptide. The reaction mixture of the caffeine derivative and the biotinylated nonapeptide was purified by C18 RP-HPLC with fractions collected from 30 to 50 min at 0.5-min intervals and monitored at 275 (Figure 2, top panel) and 214 nm (Figure 2, bottom panel). The UV properties of the caffeine carboxylate derivative showed a red shift from ∼270 to 275 nm after conjugation (data not shown). The yield was estimated at 2530% with a MALDI-TOFMS determined molecular weight of 1276.5 g/mole, as compared to the expected value of 1276.6 g/mol. All five immunized camelids showed good immune response in their serum (data not shown). Although polyclonal antisera can be used in many assay formats, we wanted to take advantage of phage display techniques for the reproducible and unlimited production of variable region fragments with selected specificity.16,17 We utilized a commercially available vector, pCANTAB 5 E, to produce M13 phage displayed libraries of the variable regions of the heavy-chain antibodies (VHH) from immunized animals. This vector is a phagemid vector which allows linking of phenotype with genotype (see ref 18 for a detailed description of phage display technology). Expressed VHH is covalently linked to a coat protein of the phage as a fusion partner. The pCANTAB 5E vector incorporates a C-terminal E-tag extension that can be used to aid purification and detection. We produced five separate libraries, each containing the cloned PCR fragments encoding VHH regions amplified from the cDNA from one immunized animal. A process of selection and enrichment, called panning (described in the Experimental Section), was utilized to isolate and enrich clones specific to caffeine. Specific clones were selected by two rounds of panning. Table 2 shows details of the library sizes and panning results. One llama library (Fanfare) produced no caffeine-specific (16) Sastry, L.; Alting-Mees, M.; Huse, W. D.; Short, J. M.; Sorge, J. A.; Hay, B. N.; Janda, K. D.; Benkovic, S. J.; Lerner, R. A. Proc. Natl. Acad. Sci. 1989, 86, 5728-5732. (17) Ward, E. S.; Gu ¨ ssow, D.; Griffiths, A. D.; Jones, P. T.; Winter, G. Nature 1989, 341, 544-546. (18) Barbas, C. F., III; Burton, D. R.; Scott, J. K.; Silverman, G. J. Phage Display: A Laboratory Manual; Cold Spring Harbor Laboratory Press: Cold Spring Harbor, New York, 2001.
Analytical Chemistry, Vol. 78, No. 13, July 1, 2006
4505
Table 2. VHH Libraries Prepared from Immunized Camels1 and Llamas2 positive clones (by ELISA) immunized animal Assab1 Massawa1 Very Senorita2 FanFare2 Virtual2
library size (cfu)
prepan
round 1
round 2
1× 4 × 106 2 × 105 4 × 105 1 × 106
0/96 0/96 0/96 76/96 0/96
0/48 1/48 17/48 0/48 18/48
2/48 31/48 37/48 0/48 18/48
106
Figure 4. Composite of SDS-PAGE and Western blot analysis of purified anti-caffeine VHH (VSA2). Panel A: SDS-PAGE of 0.5 µg VSA2 run on a 10-20% gradient gel and stained with Comassie Blue. Panel B: Western blot of 0.5 µg VSA2 transferred to PVDF membrane and blotted with goat anti-llama IgG alkaline phosphatase conjugate and developed with BCIP/NBT substrate. Panel C: Western blot of 0.1 µg VSA2 transferred to PVDF membrane and blotted with mouse anti-E-tag, followed by goat anti-mouse IgG/alkaline phosphatase conjugate and developed with BCIP/NBT substrate, Standard (Std): molecular weight (kDa) markers run in tandem with each lane of VSA2.
Figure 3. Deduced amino acid sequence of anti-caffeine VHH (VSA2) shown in standard type with the complementarity determining regions boxed. The pET 28 vector N-terminus amino acids are in boldface with the T7 tag residues underlined. The pCANTAB 5E vector E-tag residues at the C terminus are in boldface type with the E-tag residues underlined.
clones, even though prepanning showed positive clones. These early positive clones were possibly unstable or of low affinity and, thus, not robust enough to survive a round of panning and selection. These same characteristics could explain the little to no enrichment from round 1 to round 2. Preliminary specificity and thermal stability studies allowed us to select one clone (VSA2) from another llama library (Very Senorita) to expand and further characterize. This clone showed the best thermal stability and also expressed the most antibody in cell culture. Although pCANTAB 5E is a useful vector for selection of specific clones, it is not well-suited for recombinant protein production. Recloning into expression vectors can facilitate the production of larger quantities of antibody. We utilized the pET 28 vector, which adds a N-terminal extension consisting of a His tag and a T7 tag. After expression and purification on a nickel agarose column, thrombin cleavage removes the His tag portion of the construct. Figure 3 shows the protein sequence for the purified thrombincleaved VSA2 construct as deduced from the DNA sequencing data. We compared the sequence of VSA2 with published llama sequences.19 On the basis of key landmark consensus residues and using the Immunogenetics (IMGT) numbering system18 in the framework and complementarity determining regions (CDRs), we have assigned our VHH to the VHH1 subfamily. The VHH1 subfamily contains two cysteine residues N-terminal to CDRs 1 and 3, respectively, that are assumed to form a conserved disulfide bridge of the Ig fold.19 Figure 3 shows the CDR regions (enclosed in boxes), which when properly folded combine to form the antigen binding site. The boldface N-terminal and C-terminal extensions were incorporated by the pET 28 and pCANTAB 5E vectors, respectively. The E-tag and T7 tag amino acids are underlined. (19) Harmsen, M. M.; Ruuls, R. C.; Nijman, I. J.; Niewold, T. A.; Frenken, L. G. J.; de Geus, B. Mol. Immunol. 2000, 37, 579-590.
4506 Analytical Chemistry, Vol. 78, No. 13, July 1, 2006
Figure 5. Thermal stability of anti-caffeine VHH (VSA2) (-9-), as compared to two commercially available mouse monoclonal anticaffeine monoclonal antibodies, MAb1 (‚‚‚b‚‚‚) and MAb2 (- -2- -). Antibodies were incubated at increasing temperatures for 20 min, reequilibrated to RT, and then assayed in the standard caffeine ELISA. Shown is the average of four measurements ( 1 SD.
SDS-PAGE and Western blots of VSA2 are shown in Figure 4. Coomassie blue staining of VSA2 shows a single band at ∼15 kD (panel A). This band corresponds to the Western blot reactive bands using two different antibodies, anti-llama IgG (panel B) and anti-E-tag (panel C). In addition, an immunostained band with identical mobility was observed following anti-T7-tag Western blot (data not shown). These results validate the protein as a llamaderived antibody fragment and show full-length expression from the N-terminal T7 promotor tag to the C-terminal E-tag. Molecular weight determination of VSA2 by MALDI-TOFMS gave a value of 16 603.2 (data not shown) which agrees well with the MW of 16 600.1 calculated from the deduced sequence, the difference of 3.1 being within the accuracy of the method. Along with the gel data above, this confirms the VHH identity and purity. We investigated the thermal stability and reactivity of our caffeine-specific VHH. Figure 5 compares the binding in the standard caffeine ELISA (Table 1) of VSA2 with that of two commercial mouse anti-caffeine monoclonal antibodies after incubation at temperatures from RT to 90 °C for 20 min. At temperatures of 60 °C and below, there is no effect on either the llama VHH or the mouse monoclonal antibodies. In contrast, at higher temperatures, the VHH shows far greater stability, as compared to the mouse monoclonal antibodies. This VHH (VSA2)
Figure 6. Reactivity of anti-caffeine VHH (VSA2) at 70 °C as determined in a soluble competition caffeine assay. VSA2 (2 µg/mL) was incubated with equal volumes of biotinylated caffeine/nonapeptide (0, 20, and 200 µg/mL). The bound and free VSA2 were separated by means of streptavidin-coated magnetic beads, which were pelleted using a magnet. The bound antibody was removed with the pelleted beads, and the amount of unbound antibody remaining in the supernatant was determined using the standard caffeine ELISA. The bars represent the average of three determinations ( 1 SD.
can recover >90% of its activity after incubation at temperatures up to 90 °C, whereas virtually all of the binding activity of the mouse antibodies is lost after incubation at 70 °C and higher. It is possible that the low measurable binding of the mouse monoclonal antibodies after heat treatment is due to a denatured Fc portion and subsequent nonrecognition by the reporter antibody. We were unable to investigate the reactivity of our caffeinespecific VHH at high temperature in the standard caffeine ELISA (Table 1). At a temperature of 70 °C or above, the immobilized caffeine/BSA and the BSA carrier protein used in the diluent buffer became cloudy, causing high backgrounds and poor precision. We therefore designed a caffeine/peptide conjugate that was soluble and thermally stable. We designed a nonapeptide with a N-terminal biotin for avidin/strepavidin capture that had a flexible and hydrophilic core devoid of aromatic amino acids and internal amines and carboxylates and a C-terminal residue that is capable of single-site reaction with any carboxylate reagent via EDC chemistry. The C terminus was strategically amidated to avoid any intramolecular bond formation between the -NH2 of lysine and a normal C-terminal carboxylate, and aromatic residues were omitted so as not to obfuscate the conjugate 250-300-nm UV region, so conjugation to caffeine could be readily followed. The amino acid sequence of the nonapeptide is described in the Experimental Section. Note that it should be entirely possible to design a corresponding peptide with a C-terminal acid amide capable of reacting with amine groups using the same EDC chemistry. This biotinylated nonapeptide was conjugated to caffeine, and the product was used to demonstrate the thermal reactivity of our VHH. The amount of VSA2 remaining after incubation at 70 °C with increasing concentrations of the caffeine/biotinylated nonapeptide conjugate and removal of the VSA2 bound to the conjugate by streptavidin-coated magnetic beads was measured. At added concentrations of 20 and 200 µg/mL of caffeine/ biotinylated nonapeptide, 44.6 and 9.0%, respectively, of the VHH
Figure 7. Cross reactivity of anti-caffeine VHH (VSA2) with three methylxanthines. A competition caffeine ELISA was run with increasing concentrations of caffeine (-9-), theophylline (‚‚‚b‚‚‚), or theobromine (- -2- -), and response (B) was compared to that of no competitor (Bo). The average of three determinations ( 1 SD is shown.
Figure 8. Standard curves of competition caffeine ELISAs showing effects of different diluents. Caffeine was diluted with decaf coffee (-9-), Classic Coke (‚‚‚b‚‚‚), Diet Coke (- -2- -), or TBS (-O-) to produce standard concentrations of 0-800 µg/mL). Competition was determined by comparing the dose response at each concentration (B) to that of no competitor (Bo). Each point represents the average of three determinations ( 1 SD.
remains in solution. This indicates that VSA2 still binds to caffeine at elevated temperatures. Theophylline and theobromine are structurally related to caffeine and can compete for antibody binding. A competition caffeine ELISA (Table 1) was used to determine the degree to which these compounds will bind to VSA2 and prevent its binding to immobilized caffeine/BSA. The results are shown in Figure 7. Theophylline at 250 µg/mL partially competes with VSA2, but ∼14 times as much theophylline as caffeine is needed for equivalent competition. This corresponds to a cross reactivity of 7.4%, whereas theobromine at concentrations up to 250 µg/mL shows essentially no competition with VSA2. The two commercially available mouse monoclonal antibodies used in this study are reported to have cross reactivities of 3.0 and 11.4% with theophylline and 2.7 and 2.1% with theobromine (Biodesign International, Seco, ME, Specification Sheets). Thus, VSA2 showed cross reactivity comparable to the mouse monoclonal antibodies. Theophylline and theobromine are present in coffee at ∼1 µg/mL (compared to caffeine at 350-1200 µg/mL), whereas standard cola drinks contain
