Bioconjugate Chem. 2004, 15, 1125−1136
1125
Novel Monoclonal Antibodies with Specificity for Chelated Uranium(VI): Isolation and Binding Properties Robert C. Blake II,‡ Andrey R. Pavlov,§,| Mehraban Khosraviani,§,⊥ Harry E. Ensley,X Garry E. Kiefer,# Haini Yu,§ Xia Li,§,3 and Diane A. Blake§,* Division of Basic Pharmaceutical Sciences, College of Pharmacy, Xavier University of Louisiana, New Orleans, Louisiana, Department of Biochemistry, Tulane University Health Sciences Center, New Orleans, Louisiana, Department of Chemistry, Tulane University, New Orleans, Louisiana, Dow Chemical Company, Freeport, Texas, and Beijing Vegetable Research Institute, Beijing, China. Received May 7, 2004; Revised Manuscript Received August 2, 2004
A derivative of 1,10-phenanthroline that binds to UO22+ with nanomolar affinity was found to be a very effective immunogen for the generation of antibodies directed toward chelated complexes of hexavalent uranium. This study describes the synthesis of 5-isothiocyanato-1,10-phenanthroline-2,9dicarboxylic acid and its use in the generation and functional characterization of a group of monoclonal antibodies that recognize the most soluble and toxic form of uranium, the hexavalent uranyl ion (UO22+). Three different monoclonal antibodies (8A11, 10A3, and 12F6) that recognize the 1:1 complex between UO22+ and 2,9-dicarboxy-1,10-phenanthroline (DCP) were produced by the injection of BALB/c mice with DCP-UO22+ covalently coupled to a carrier protein. Equilibrium dissociation constants for the binding of DCP-UO22+ to antibodies 8A11, 10A3, and 12F6 were 5.5, 2.4, and 0.9 nM, respectively. All three antibodies bound the metal-free DCP with roughly 1000-fold lower affinity. The secondorder rate constants for the bimolecular association of each antibody with soluble DCP-UO22+ were in the range of 1 to 2 × 107 M-1 s-1. Binding studies conducted with structurally related chelators and 21 metal ions demonstrated that each of these three antibodies was highly specific for the soluble DCP-UO22+ complex. Detailed equilibrium binding studies conducted with three other derivatives of DCP, either complexed with UO22+ or metal-free, suggested that the antigen binding sites on the three antibodies have significant functional and structural similarities. Biomolecules that bind specifically to uranium will be at the heart of any new biotechnology developed to monitor and control uranium contamination. The three antibodies described herein possess sufficient affinity and specificity to support the development of immunoassays for hexavalent uranium in environmental and clinical samples.
INTRODUCTION
The mining and processing of uranium ores have contaminated significant quantities of soil and groundwater in the US, Europe, and Russia (1-3). In the US alone, more than 200 million metric tons of mine tailings and other waste have been identified for cleanup and disposal (4). The health effects associated with oral or dermal exposure to natural and depleted uranium appear to be solely chemical in nature and not related to uranium’s radiological properties (5). The toxicity of natural uranium varies according to its chemical form, and the more soluble, hexavalent form has been shown to be the most potent systemic toxicant (6). Uranium is absorbed from the intestine or the lungs. Although most * To whom correspondence should be addressed at 1430 Tulane Ave., SL-43, New Orleans, LA 70112. E-mail: blake@ tulane.edu; Tel: 504-988-2478; Fax: 504-988-2684. ‡ Xavier University of Louisiana. § Department of Biochemistry, Tulane University Health Sciences Center. X Department of Chemistry, Tulane University. # Dow Chemical Company. 3 Beijing Vegetable Research Institute. | Current address: Fidelity Systems Inc., Gaithersburg, MD 20879. ⊥ Current address: Amgen Incorporated, Thousand Oaks, CA 91320.
uranium is rapidly excreted in the urine, a small percentage is deposited in the tissues, predominantly kidney, muscle, and bone (7, 8). Chronic exposure to uranium in drinking water is weakly associated with altered proximal tubulus function without a clear threshold, which suggests that even low uranium concentrations in drinking water can cause nephrotoxic effects (9). Current research suggests that the safe concentration of uranium in drinking water may be close to the value proposed by the US Environmental Protection Agency, 30 µg/L or ∼130 nM (10). As part of an ongoing effort in our laboratories to develop immunoassays for heavy metals in environmental and clinical samples, we have produced and characterized a growing number of monoclonal antibodies directed toward epitopes of metal-chelate complexes (11-17). Although other laboratories have described the generation and properties of monoclonal and recombinant antibodies directed toward chelated metal ions (18-23), a detailed characterization of antibodies directed toward an epitope that includes some form of ionic uranium has not been reported. Previous attempts by our laboratories to generate antibodies to chelated hexavalent uranium using bifunctional acyclic polyaminocarboxylate chelators available from similar studies were not successful. These failed attempts led us to hypothesize that bifunctional chelators derived from EDTA and DTPA did not coordi-
10.1021/bc049889p CCC: $27.50 © 2004 American Chemical Society Published on Web 08/20/2004
1126 Bioconjugate Chem., Vol. 15, No. 5, 2004
nate hexavalent uranium with an affinity high enough to preserve the metal-chelate complex during immune processing in the injected animal. This paper describes the generation and binding properties of three different monoclonal antibodies directed toward an epitope comprised of the hexavalent UO22+ and a novel bifunctional derivative of 2,9-dicarboxy-1,10-phenanthroline (DCP)1. DCP bound to UO22+ with an affinity approximately 5 orders of magnitude greater than that determined for EDTA or DTPA. Each of the three monoclonal antibodies produced using this new amine-reactive chelator bound to the soluble DCPUO22+ complex with an equilibrium dissociation constant in the low nM range. Equilibrium and kinetic binding studies conducted with different chelators and metal ions indicated that each of these three antibodies possessed sufficient specificity and affinity to create an immunoassay for soluble natural uranium in the concentration range currently proposed by the US Environmental Protection Agency for drinking water (10). EXPERIMENTAL PROCEDURES
Materials. 2,9-Dicarboxy-1,10-phenanthroline (DCP) was synthesized (24) or purchased from Alfa Aesar (Wood Hill, MA). 1,10-Phenanthroline, 2,9-dimethyl-1,10-phenanthroline, and 4-(2-pyridylazo)resorcinol (PAR) were products of Aldrich, Inc. (Milwaukee, WI). The 2,9-diphosphonatomethyl-, 2,9-dicarboxymethyl ester-, 2,9-dihydroxymethyl-, and 2,9-ditrichloromethyl-substituted 1,10phenanthrolines were generously provided by Dow Chemical Co. (Freeport, TX). Uranyl diacetate (ACS-grade) was a product of Mallinckrodt Chemical Works (St. Louis, MO). Other metals (atomic absorption grade) were obtained from Perkin-Elmer (Norwalk, CT). Prepacked Sephadex G-25 columns (PD-10) were products of Pharmacia, Inc. (Piscataway, NJ). Centricon filters were purchased from Amicon (Beverly, MA). Keyhole limpet hemocyanin (KLH) was purchased from CalBiochem (LaJolla, CA). BALB/c inbred mice were obtained from Charles River Laboratories (Wilmington, MA). SP 2/0Ag14 myeloma cells were from the American Type Culture Collection (Rockville, MD). RIBI and TiterMax adjuvants were purchased from RIBI Immunochemicals (Hamilton, MT) and CyxRx (Norcross, GA). Culture media and other reagents for the preparation of hybridoma cells were obtained from Stem Cell Technologies (Vancouver, BC, Canada). An IsoStrip mouse monoclonal antibody isotyping kit was purchased from BoerhingerMannheim (Indianapolis, IN) and used according to the manufacturer’s instructions. Fetal bovine serum (low IgG) was the product of Hyclone Laboratories (Logan, UT). 3,3′,5,5′-Tetramethylbenzidine peroxidase substrate (TMB Microwell Substrate) was from Kirkegaard-Perry Laboratories (Gaithersburg, MD). Microwell plates for ELISA (high-binding, flat bottom) and standard tissue culture plasticware were from Corning/Costar (Cambridge, MA). CellLine culture flasks were distributed by Fisher Chemical Co. (Houston, TX). Polystyrene beads (98 micron diameter) for binding studies were purchased from Sapidyne Inc (Boise, ID). HRP- and Cy5-labeled goat 1 The following abbreviations are used in the text: BSA, bovine serum albumin; CHXDTPA, trans-N-(2-aminoethyl)-1,2diaminocyclohexane-N,N′,N′′-pentaacetic acid; DCP, 2,9-dicarboxy-1,10-phenanthroline; DHM, 2,9-dihydroxymethyl-1,10phenanthroline; DME, 2,9-dicarboxymethyl ester-1,10-phenenthroline; DPP, 2,9-diphosphonatomethyl-1,10-phenanthroline; DTPA, diethylenetriaminepentaacetic acid; EDTA, ethylenediaminetetraacetic acid; PAR, 4-(2-pyridylazo)resorcinol.
Blake et al.
anti-mouse IgG were products of Jackson Immunoresearch Laboratories (West Grove, PA). Ethylenediaminetetraacetic acid (EDTA), diethylenetriaminepentaacetic acid (DTPA), pristane, HEPES, RPMI medium, Lglutamine, antibiotics, bovine serum albumin (fraction V, metal-free), and antibodies specific for mouse IgG1, IgG2a IgG2b, IgG3, IgM, and IgA were purchased from Sigma Chemical Co. (St. Louis, MO). The DTPA analogue, N-(2-aminoethyl)-trans-1,2-diaminocyclohexaneN,N′,N′′-pentaacetic acid (CHXDTPA), was available from a previous study (16). Elemental analyses were performed by Atlantic Microlab, Inc., Norcross, GA. Determination of Dissociation Constants for the Interaction of Chelators with UO22+. The binding of uranyl ion to colorimetric and noncolorimetric chelators was determined by spectrophotometric titrations performed on an OLIS-14 renovated Cary Spectrophotometer (On-Line Instruments, Inc. Bogart, GA). Visible absorbance spectra of 26 µM PAR in HEPES-buffered saline (HBS, 137 mM NaCl, 3 mM KCl, 10 mM HEPES, pH 7.4) were performed in the absence and presence of varying concentrations of uranyl ions (5-50 µM), and values of ∆ absorbance were determined as differences in the spectrum of the metal-free chelator versus the spectrum of the chelator in the presence of uranyl ions. For determination of the dissociation constants of noncolorimetric chelators with uranyl ion, the concentrations of PAR and uranyl ion were fixed at 26 and 30 µM, respectively, and spectra were obtained after adding varying concentrations of the noncolorimetric chelator to the preformed complex. Data were fit to the equations described in Results to obtain equilibrium dissociation constants. Synthesis of Bifunctional Chelator for UO22+. Synthesis of 5-isothiocyanato-1,10-phenanthroline-2,9dicarboxylic acid was performed in four steps, as shown in Figure 1. 5-Nitro-2,9-dimethyl-1,10-phenanthroline (2 in Figure 1). To a mixture of 5 mL of nitric acid and 10 mL of sulfuric acid was added 0.5 g (2.4 mmol) of 2,9-dimethyl1,10-phenanthroline (1 in Figure 1), and the solution was heated at 115 °C for 1 h. The solution was poured onto 100 g of ice, and the mixture was adjusted to pH 8 with a concentrated solution of NaOH. The precipitate was filtered and dried at 110 °C to give 248 mg of 2 (0.98 mmol, 41% yield); mp 201-203 °C, 1H NMR δ 2.97 (3H, s), 2.99 (3H, s), 7.66 2H, app t, J ) 8.5 Hz), 8.26 (1H, d, J ) 8.3 Hz), 8.60 (1H, s) 8.91 (1H, d, J ) 8.7 Hz). 5-Nitro-1,10-phenanthroline-2,9-carboxylic Acid (4 in Figure 1). A mixture of compound 2 (1.0 g, 3.6 mmol) and selenium dioxide (1.0 g, 9.0 mmol) was dissolved in 5 mL of dioxane containing 4% water and heated at reflux for 3 h, then filtered through a pad of Celite while hot. The dialdehyde 3 was isolated from the filtrate as a yellowred precipitate in 64% yield. Anal. (C14H7N3O4) C, H, N. The dialdehyde was subsequently oxidized to the dicarboxylic acid using 10 mL of 80% nitric acid under reflux for 3 h. The solution was then cooled and poured into ice. The resulting precipitate was collected and recrystallized from aqueous THF to give 580 mg of the bis-acid 4 (1.9 mmol, 46% from 2); mp 217-220° C; IR 3100 and 1733 cm-1, 1H NMR δ 8.47 (1H, d, J ) 8.0 Hz), 8.95 (1H, d, J ) 8.0 Hz), 9.04 (1H, d, J ) 8.8 hz), 9.17 (1H, s). Anal. (C14H7N3O6-1/2 H2O) C, H, N. 5-Isothiocyanato-1,10-phenanthroline-2,9-dicarboxylic Acid (5 in Figure 1). Compound 4 (110 mg, 0.3 mmol) was hydrogenated in 5 mL of ethanol at 60 psi in the presence of 15 mg of 10% Pd/C. On exposure to air, the initially yellow solution became bright red. Therefore, the
Monoclonal Antibodies to Chelated Uranium
Bioconjugate Chem., Vol. 15, No. 5, 2004 1127
Figure 1. Synthesis of 5-isothiocyanato-1,10-phenanthroline-2,9-dicarboxylic acid. Scheme for the synthesis of 5-isothiocyanato1,10-phenanthroline-2,9-dicarboxylic acid in from 2,9-dimethyl-1,10-phenanthroline.
amine was directly converted to the isothiocyanate. 3 N HCl (10 mL) was added, and the solution was filtered to remove the catalyst. The ethanol was evaporated at reduced pressure, and the solution was cooled to 0 °C. Thiophosgene (5 mL of a 10% solution in chloroform) was then added with rapid stirring, which was continued overnight at room temperature in a chemical hood. Evaporation of the two-phase mixture under reduced pressure gave a solid that was washed with 10 mL of cold water and dried under vacuum to give 98 mg (0.30 mmol, 86% from 4) of 5. The isothiocyanate derivative could be used immediately or stored for up to 3 months at -20 °C. Synthesis of Protein-Chelate Conjugates. 5-Isothiocyanato-1,10-phenanthroline-2,9-dicarboxylic acid (7.1 mg or 22 µmol, ∼equal molar with respect to total lysine residues in these proteins) was suspended in 1.0 mL of 0.1 M HEPES buffer, pH 9.0, containing 20 mg of BSA or KLH. The pH of the reaction mixture was readjusted to 9.0 if necessary with KOH, and the reactions were stirred overnight at room temperature. The supernatant from this reaction mixture was desalted by filtration through a prepacked Sephadex G-25 column equilibrated with HEPES-buffered saline (HBS, 137 mM NaCl, 3 mM KCl, 10 mM HEPES, pH 7.4). The protein-DCP conjugates eluted in the void volume of the column and could be further concentrated (if necessary) by ultrafiltration in Centricon filters (30 000 mol. wt. cutoff). The concentration of the conjugates was determined using the BCA protein assay kit (Pierce Chemical Co., Rockford, IL) with supplied BSA standards. The degree of lysine substitution in the conjugates was estimated by a modification of a literature method (25). The original procedure required additional background measurements because the conjugates had significant absorbance at 345 nm. Protein-chelate conjugates prepared by this method were estimated to have between 20% and 80% lysine substitution (depending upon the preparation) and were light yellow-brown in color when present at a concentration > 2.5 mg/mL. Mouse Immunization, Hybridoma Production, and Purification of Monoclonal Antibodies. The KLH-DCP conjugate (80% substitution, 0.7 mg in 0.28 mL of HBS) was mixed 0.07 mL of 10× HBS and 0.07 mL of uranyl diacetate (62.5 mM solution in dilute acetic acid, 3.5 µmol), followed by immediate addition of 0.05 mL of KOH (1.8 M aqueous solution) to bring the pH of the mixture to 7.0. The mixture was diluted by the addition of 0.15 mL of water, the pH was adjusted to 7.0, and the final volume of the solution was brought to 0.7 mL. This mixture was emulsified with an equal volume
of RIBI or TiterMax adjuvant. The components of this preparation were not fully soluble, but produced homogeneous emulsions with both adjuvants. The final immunogen contained 0.5 mg/mL of the KLH-DCP conjugate and 2.5 mM uranyl diacetate. Three 6-week old female BALB/c mice were each injected intraperitoneally with 0.1 mL of the prepared immunogen (2 with RIBI and 1 with TiterMax). The injections were repeated after 3 weeks, and the polyclonal antibody responses were measured by indirect ELISA as described below. The mouse with the best polyclonal antibody response to uranium-loaded BSA-DCP was given a final intraperitoneal injection of 50 µg of KLHDCP with 2.5 mM uranyl diacetate in HBS, and after 72 h the animal was sacrificed and its spleen was removed. Hybridomas were prepared by fusion of splenocytes and SP 2/0-Ag14 myeloma cells using the reagents provided by Stem Cell Technologies and plated into a semisolid growth medium also provided by this manufacturer. After incubation at 37 °C in 95% air/5% CO2, hybridoma cell colonies began to grow out in 7-10 days. These colonies, visible to the naked eye, were individually transferred to the wells of 96-well plates that contained 0.15 mL of culture medium and incubated for several days. The culture supernatants were assayed for antibody production by indirect and competitive ELISA (described below) and representative cells were cloned twice by limiting dilution before storage in liquid nitrogen. Ascites fluid was produced in pristane-primed BALB/c mice. For some experiments, large quantities of antibodies were prepared by hybridoma culture in 1-L CellLine culture flasks, used according to the manufacturer’s instructions. Monoclonal antibodies were purified by affinity chromatography on protein G as previously described (11). The homogeneity of the purified proteins was checked by gel electrophoresis in the presence of sodium dodecyl sulfate and β-mercaptoethanol (26). The concentration of the antibodies was determined using the BCA protein assay, and the immunoglobulin subclass was determined and quantified by sandwich ELISA using antibodies specific for IgG1, IgG2a, IgG2b, IgG3, IgM, and IgA. Light chain isotype was identified by using the IsoStrip antibody isotyping kit according to the manufacturer’s instructions. Indirect and Competitive ELISA. Addition of uranyl diacetate to concentrated solutions of the DCPprotein conjugates (>100 µg/mL) caused precipitation. This precipitation could be avoided either by dilution of the conjugate to concentrations less that 10 µg/mL before addition of uranyl ion, or by immobilizing metal-free DCP-protein conjugate onto the ELISA plate and sub-
1128 Bioconjugate Chem., Vol. 15, No. 5, 2004
Blake et al.
sequently incubating the plate with a solution of uranyl diacetate. Both procedures gave equivalent results in control experiments (data not shown). ELISA plates were incubated with 2-5 µg/mL of DCP-BSA in HBS amended with 5 µM uranyl diacetate for at least 1 h at 37 °C or overnight at 4 °C. The plates could be stored for several weeks in this coating buffer with no loss of activity. Immediately before use, the plates were washed three times with PBS-Tween (137 mM NaCl, 3 mM KCl, 10 mM phosphate, pH 7.4, and 0.05% Tween-20), blocked for 15 min with 3% BSA in HBS, and washed again three times with PBS-Tween. For indirect ELISA, diluted serum or hybridoma culture supernatant was incubated for 1 h at room temperature in the coated wells. After three washes in PBS-Tween, the amount of primary antibody bound on the plate was quantified using an HPR-conjugated goat anti-mouse antibody and TMB microwell substrate (see refs 11 and 12 for a more detailed description). For competitive ELISA, diluted hybridoma culture supernatants were incubated in microwell plates coated with BSA-DCP-UO22+ in the presence of one of the following: 12.5 µM DCP in HBS or 12.5 µM DCP + 5 µM uranyl diacetate in HBS. The binding of the antibodies in HBS alone served as a control in the competitive ELISA’s. The amount of primary antibody bound to the plate was quantified as described for indirect ELISA. Equilibrium Binding Studies and Kinetic Analyses. Equilibrium binding studies and measurements of association rate constants were performed using a KinExA 3000 (Sapidyne Instruments Inc, Boise, ID) flow fluorimeter. The general procedures have been described in detail elsewhere (11, 27, 28). For these experiments, polystyrene beads coated with UO22+-DCP-BSA were used as the capture reagent. Polystyrene beads (98 µm, 200 mg) were incubated with 1 mL of 100 µg/mL DCPBSA for 1 h at 37 °C. The beads were washed two times with 1 mL of HBS, then any remaining binding sites were blocked by incubation for 1 at 37 °C with BSA (10 mg/ mL in HBS containing 0.03% NaN3.) Beads could be stored in this blocking solution for up to 2 months at 4 °C. Immediately before use, the beads were diluted into 30 mL of HBS amended with 30 µM uranyl diacetate to saturate the immobilized DCP with UO22+. The excess uranyl ions were washed away when the coated beads were packed into the capillary observation cell. RESULTS
Binding of Uranyl Ions to Chelators. Our laboratories have successfully isolated and characterized monoclonal antibodies to chelated forms of ionic cadmium and lead using bifunctional derivatives of EDTA (11), CHXDTPA (16), and DTPA (14) to create metal-binding epitopes on appropriate carrier proteins. In all cases to date, the resulting covalent complexes of chelated metal ions appeared to be weak immunogens, since only a few clones out of hundreds assayed actually produced antibodies that exhibited acceptable specificity for the chelated metal ion in question. In the present case, initial efforts to generate monoclonal antibodies directed against chelated forms of UO22+ utilized bifunctional acyclic polyaminocarboxylate chelators (DTPA and CHXDTPA). Not a single clone (out of approximately 2000 examined) produced an antibody with acceptable affinity and specificity for chelated ionic uranium (29).
fraction of PAR occupied with UO22+ )
One hypothesis to account for these observations is that the affinity of DTPA and CHXDTPA for uranyl ions was insufficient to maintain the intact chelated uranium complex for the period of time necessary to elicit an immune response to the chelated metal. To test this hypothesis, the binding affinity of various chelators for uranyl ions was determined by competitive uranyl binding experiments conducted in the presence of 2-(4pyridylazo)resorcinol (PAR), a spectrophotometric compleximetric indicator for cationic metal ions (30, 31), including UO22+. Figure 2A shows representative absorbance spectra of PAR obtained in the absence and presence of different concentrations of UO22+. As the concentration of UO22+ increased, the absorbance peak of the metal-free PAR at 410 nm decreased concomitantly as a new absorbance peak at 520 nm increased. These UO22+-dependent changes in the absorbance spectrum of PAR were exploited to determine the equilibrium dissociation constant for the binding of UO22+ to the colorimetric chelator. Differences in the spectra were quantified as follows: for the interval from 340 to 453 nm, the area under the spectrum obtained in the presence of UO22+ was subtracted from the area under the spectrum obtained in the absence of UO22+; for the interval from 455 to 640 nm, the area under the spectrum obtained in the absence of UO22+ was subtracted from that obtained in the presence of UO22+; the values of the resulting two quantities were then combined to create the values represented on the ordinate of Figure 2B. Figure 2B shows the dependence of the spectral changes extracted from the primary data in Figure 2A on the concentration of total uranyl ion. The dashed lines represent the theoretical curve anticipated for a 1:1 stoichiometric binding event with infinite affinity. The actual data points clearly conform to a model where PAR binds a single uranyl ion with relatively high affinity. The value of the equilibrium dissociation constant for the binding of UO22+ to PAR, KPAR-U, was obtained from a nonlinear regression fit of quadratic equation 1 where [P] and [U] are the concentrations of total PAR and UO22+, respectively. Equation 1 corresponds to a 1:1 homogeneous binding model where the binding event causes a significant depletion in the concentrations of both of the free, uncomplexed reagents. The value of KPAR-U that provided the best fit of eq 1 to the data in Figure 2B was 4.0 ( 0.3 × 10-7 M. The next experiment was to study the binding of UO22+ to a noncolorimetric chelator by quantifying the effects that the colorless chelator had on the interaction of UO22+ with the colorimetric chelator, PAR. When different concentrations of DTPA (a colorless chelator) were equilibrated with a preformed complex of PAR-UO22+, uranyl ions transferred from the PAR to the DTPA, thus regenerating the metal-free form of the colored chelator (primary data not shown). Figure 2C shows the dependence of the spectral changes taken from these 3-component competition experiments on the concentration of total DTPA. The value of the equilibrium dissociation constant for the binding of UO22+ to DTPA, KDTPA-U, was obtained from a nonlinear regression fit of quadratic equation 2 where [D] is the concentration of total DTPA. Equation 2 differs from eq 1 only in that the concentration of total uranyl ion is divided by the factor (1 + [D]/KDTPA-U), which represents the influence of UO22+ complexation with DTPA on the concentration of
([P] + [U] + KPAR-U) - {([P] + [U] + KPAR-U)2 - 4[P][U]}1/2 2
(1)
Monoclonal Antibodies to Chelated Uranium
Bioconjugate Chem., Vol. 15, No. 5, 2004 1129
Figure 2. Determination of binding constants for uranyl ion to colorimetric and noncolorimetric chelators. A, visible absorbance spectra of 26 µM 4-(2-pyridazo)resorcinol (PAR) in the absence (a) and presence of 5.0 (b), 12.5 (c), 20 (d), and 50 (e) µM uranyl ion. B, dependence of the changes in the absorbance of 26 µM PAR as a function of the concentration of uranyl ion. The values of ∆ absorbance were determined as the differences in the spectrum of the metal-free chelator versus the spectrum of the chelator observed in the presence of uranyl ions. C, dependence of the changes in the absorbance of the PAR-uranyl ion complex as a function of the concentration of DTPA. The concentrations of PAR and uranyl ion were 26 and 30 µM, respectively. The values of ∆ absorbance were determined as the differences in the spectrum of the PAR-uranyl complex in the absence of DTPA versus the spectrum of the complex observed in the presence of DTPA. D, dependence of the changes in the absorbance of the PAR-uranyl ion complex as a function of the concentration of DCP. The conditions were the same as those described in C, except that DCP was substituted for DTPA. The parameters for the curves drawn through the data points in B, C, and D were determined by nonlinear regression analyses using eqs 1, 2, and 3, respectively, in the text. Table 1. Equilibrium Dissociation Constants for Uranyl Ion with Selected Complexants chelatora
equilibrium dissociation constant, M
PAR DTPA EDTA CHXDTPA DCP
4.0 ( 0.3 × 10-7 3.1 ( 0.2 × 10-4 2.4 ( 0.2 × 10-4 3.3 ( 0.3 × 10-4 3.5 ( 0.4 × 10-9
a Abbreviations are as follows: PAR, 4-(2-pyridylazo)resorcinol; DTPA, diethylenetriaminepentaacetic acid; EDTA, ethylenediaminetetraacetic acid; CHXDTPA, trans-N-(2-aminoethyl)-1,2diaminocyclohexane-N,N′,N′′-pentaacetic acid; DCP, 2,9-dicarboxy1,10-phenanthroline.
free UO22+ that is available to bind to the PAR. The value of KDTPA-U that provided the best fit of eq 2 to the data in Figure 2C was 3.1 ( 0.2 × 10-4 M. Analogous competition experiments were conducted with two other colorless chelators, EDTA and CHXDTPA (primary data not shown). The values of the corresponding equilibrium dissociation constants for the binding of UO22+ to each
of these chelators are summarized in Table 1. These competition experiments provided uranium complexation data consistent with the hypothesis that the low affinity of DTPA and CHXDTPA for UO22+ was the most probable origin of the failure to identify antibodies directed against the corresponding chelator-uranyl ion complexes. Figure 2D shows the dependence of spectral changes taken from analogous competition experiments on the concentration of 2,9-dicarboxy-1,10-phenanthroline (DCP), an alternative colorless chelator examined as part of an effort to identify chelators that bound to UO22+ with higher affinities. Unlike DTPA, where millimolar concentrations of the colorless chelator were required to effectively compete with 26 µM PAR, concentrations of DCP stoichiometric with that of PAR were sufficient to remove UO22+ from its complex with the colored chelator. The value of the equilibrium dissociation constant for the binding of UO22+ to DCP, KDCP-U, was obtained from a nonlinear regression fit of quadratic equation 3 where [DCP] is the concentration of total DCP. Equation 3 corresponds to a competition binding model where the
fraction of PAR occupied with UO22+ ) ([P] + [U]/(1+[D]/KDTPA-U) + KPAR-U) - {([P] + [U] /(1+[D]/KDTPA-U) + KPAR-U)2 - 4[P][U] /(1+[D]/KDTPA-U) }1/2 2 (2)
fraction of PAR occupied with UO22+ ) (2KDCP-U[P] + KPAR-U[DCP]-KPAR-U[P]) - {(2KDCP-U[P] + KPAR-U[DCP] - KPAR-U[P])2 - 4(KDCP-U - KPAR-U)KDCP-U[P]2}1/2 2 × (KDCP-U - KPAR-U)
(3)
1130 Bioconjugate Chem., Vol. 15, No. 5, 2004
transfer of UO22+ from PAR to DCP causes significant changes in the concentrations of the metal-free forms of both chelators. Since the concentrations of both chelators were significantly higher than their respective dissociation constants for their binding with UO22+, the concentration of free, uncomplexed UO22+ was assumed to be negligibly small. In addition, all of the PAR initially present in the absence of DCP was assumed to be coordinated with UO22+. The value of KDCP-U that provided the best fit of eq 3 to the data in Figure 2D was 3.5 ( 0.4 × 10-9 M. A summary of the equilibrium dissociation constants determined for these complexants at the pH and ionic strength dictated by the immunization protocol is provided in Table 1. Hybridoma Production and Screening. Clearly, the DCP chelator was superior to EDTA, DTPA, or derivatives thereof in its ability to bind strongly to the UO22+ ion at the pH and ionic strength required by the immunization protocol. The UO22+-DCP complex should therefore be better at eliciting a metal-specific immune response than the forms of chelated UO22+ previously employed as immunogens. However, this complex was not of sufficient size to generate an immune response, and the chelator needed to be covalently coupled to a carrier protein to be an effective immunogen. Bifunctional derivatives of DCP were not commercially available, and the synthesis of 5-isothiocyanato-1,10-phenanthroline-2,9-dicarboxylic acid from 2,9-dimethyl-1,10phenanthroline was performed in four steps, as shown in Figure 1. The isothiocyanato derivative of DCP was subsequently coupled to a carrier protein and injected into mice. After a 4-week injection series, all the animals showed high serum titers of antibodies with the ability to bind immobilized UO22+-DCP-BSA (data not shown). The animal with the highest titer was given an intraperitoneal boost of antigen, and its spleen was used 72 h later for the preparation of hybridoma cells. A novel culture procedure that utilized semisolid media (see Experimental Procedures) resulted in the isolation of 1146 individual hybridoma colonies from a single fusion. The culture supernatants from these hybridomas were initially screened for their ability to bind to UO22+DCP-BSA immobilized in 96-well plates, and 486 clones (42.4%) secreted antibodies with reactivity toward uranium-loaded DCP-BSA (defined as an absorbance450 nm > 0.3 in indirect ELISA). When these 486 supernatants were rescreened by competitive ELISA (shown in Figure 3), 89.9% of the reactive supernatants (437/486) or 36.8% (437/1186) of the total hybridomas generated in these experiments were inhibited by soluble UO22+-DCP. Of the remaining 49 clones, 2.3% (11/486) were more strongly inhibited by metal-free DCP, and 7.8% (38/486) could not be inhibited by either metal-free DCP or UO22+-DCP and were assumed to be directed toward the thioureido group that linked the chelator to the carrier protein. Three of these clones, 8A11, 10A3, and 12F6, were chosen for further experiments because they had desirable growth characteristics, showed the least reactivity to metal-free DCP, and/or produced supernatants with high specific titers at limiting dilutions (data not shown). These hybridomas were subcloned by limiting dilution to confirm homogeneity. All three hybridomas secreted antibodies of the IgG1 subtype with a κ light chain. Production of antibody for subsequent characterization was performed either by culture as an ascites in pristineprimed BALB/c mice or by in vitro culture in a specialized tissue culture apparatus, followed by purification on protein G. Monoclonal antibodies prepared by these two
Blake et al.
Figure 3. Specificity of antibodies synthesized by hybridomas. The 486 reactive hybridoma (abs450nm > 0.3 when tested against immobilized BSA-DCP-UO22+) were rescreened by competitive ELISA as described in Methods. Of the 486 supernatants examined in theses experiments, 437 (89.9%) of the hybridomas synthesized antibodies with a primary specificity for DCPUO22+, and 11 (2.2%) showed specificity for metal-free DCP. A small number (38, 7.8%) of the antibodies in hybridoma supernatants bound to immobilized BSA-DCP-UO22+ but not to immobilized BSA and could not be competitively inhibited by either DCP-UO22+ or DCP; these antibodies were assumed to be directed toward the linker region between the DCP and the carrier protein.
culture methods showed identical binding properties for metal-free DCP and UO22+-DCP (data not shown). Binding Specificity. The binding specificity of each of the three anti-uranium monoclonal antibodies was investigated by conducting equilibrium binding measurements on a KinExA 3000 immunoassay instrument. The KinExA is a flow fluorimeter designed to achieve the rapid separation and quantification of free, unbound antibody present in reaction mixtures of antibody, antigen, and antibody-antigen complexes; its operation has been described in detail elsewhere (11, 27, 28). The equilibrium binding data presented in Figure 4 demonstrate that each of the three anti-uranium antibodies bound to the DCP-UO22+ complex with much higher affinity than it bound to the metal-free DCP. The low affinity binding curves in each panel of Figure 4 were obtained with metal-free DCP, while the corresponding high affinity binding curves were obtained with different concentrations of DCP in the presence of a constant total concentration of UO22+ of 1.0 µM. Given the affinity of DCP for UO22+ as determined above, these solution conditions dictated that greater than 99.6% of the DCP was present as the 1:1 complex with UO22+; control experiments indicated that 1.0 µM UO22+ alone did not inhibit the binding of any of the antibodies to the immobilized antigen. In all cases in Figure 4, the value of the equilibrium dissociation constant, Kd, for the binding of the antibody to the soluble ligand was obtained as previously described (11, 27, 28). Values for the dissociation constants thus obtained are given in Table 2. Antibody 12F6 had the highest affinities for both the metal-free DCP and the DCP-UO22+ complex; antibody 8A11 had the lowest affinities for the same ligands. Antibody 10A3 showed the highest level of discrimination between the metal-free DCP and the DCP-UO22+ complex; the difference in Gibbs free energy between the binding of DCP-UO22+ and metal-free DCP to 10A3 was -4.2 kcal/mol. However, the least discriminatory antibody, 8A11, still exhibited a corresponding difference in Gibbs free energy of -3.9 kcal/mol between its affinity for DCP-UO22+ versus that for the metal-free chelator. Each of the three anti-uranium antibodies exhibited appreciable affinity for other chelators that bore a close
Monoclonal Antibodies to Chelated Uranium
Bioconjugate Chem., Vol. 15, No. 5, 2004 1131
Table 2. Equilibrium Dissociation Constants for the Binding of Selected Phenanthroline Derivatives in the Presence and Absence of Uranyl Ion to Monoclonal Antibodies 12F6, 10A3, and 8A11
structural resemblance to DCP. Table 2 contains values of the equilibrium dissociation constants for the binding of each antibody to the dimethyl ester of DCP in the presence and absence of excess UO22+. In general, esterification of the two carboxyl groups on DCP had only a small (
