Bioconjugate Chem. 1997, 8, 281−288
281
Synthesis of Conjugates for a Barbiturate Screening Assay Maciej Adamczyk,* Jonathan Grote, Jeanine Douglas, Robert Dubler, and Charles Harrington Department of Chemistry (D9NM), Abbott Diagnostics Division, Abbott Laboratories, 100 Abbott Park Road, Abbott Park, Illinois 60064-3500. Received November 29, 1996X
Novel derivatives of barbiturates functionalized with free carboxylic acids were designed and synthesized. Coupling of 5-cyclopentyl-5-carboxycrotylbarbituric acid via its active ester to an aminofluorescein derivative produced a fluorescent tracer. Conjugation of the 5-cyclopentenyl-5carboxyethylbarbituric acid via its mixed anhydride to thyroglobulin allowed for subsequent development of a polyclonal antibody which was evaluated for binding in a fluorescence polarization immunoassay format with various barbiturates. The binding studies showed good cross-reactivity of a variety of barbiturates containing both aromatic and aliphatic 5-substituents with the tested antisera. The relationship between the immunogen architecture, the chemical structure of the binding analytes, and the characteristics of the antisera is also presented.
INTRODUCTION
Barbiturates are central nervous system depressants that are frequently administered on a therapeutic basis as sedatives, hypnotics, and anticonvulsants. These drugs have been known for several decades, having been introduced into therapeutic use by Fisher and von Mering in the early 1900s (1). A large number of different barbiturates have since been synthesized, several of which are still in use (2). Disubstitution at the 5-position is required for pharmacological activity, as unsubstituted or 5-monosubstituted barbituric acids demonstrate no CNS activity. Small structural variations in these 5-substituents cause substantial changes in the drug’s physiological effects and duration (2). Because of the rapid onset of their CNS activity, the most frequently abused barbiturates (including secobarbital, pentobarbital, and amobarbital) have alone or in combination with other drugs been frequently used to commit suicide (3, 4). Therefore, rapid determination of the presence or absence of barbiturates in a comatose patient prior to emergency medical treatment can be the difference between life and death (4). Structurally, barbiturates can be divided into two classes: those containing aliphatic 5-substituents (such as secobarbital, Figure 1) and those containing an aromatic 5-substituent (such as phenobarbital, Figure 1). The ideal screening assay for barbiturates would allow for rapid determination of the presence or absence of a variety of structures with subtle structural differences. Competitive binding immunoassay techniques are advantageous over chromatographic techniques such as high-performance liquid chromatography (HPLC) and gas chromatography (GC) for this purpose, since chromatographic techniques need initial sample extraction and have lengthy assay time requirements (5). However, immunoassay techniques such as enzyme immunoassay (EIA) and substrate-linked fluorescence immunoassay (SLFIA) both contain enzymes that are sensitive to enzyme inhibitors or other reactive enzymes contained in the sample (6), while radioimmunoassay (RIA) has other severe shortcomings, including short shelf life * Author to whom correspondence should be addressed [telephone (847) 937-0225; fax (847) 938-8927; e-mail adamczykm@ apmac.abbott.com]. X Abstract published in Advance ACS Abstracts, April 15, 1997.
S1043-1802(97)00034-7 CCC: $14.00
Figure 1. Structures of secobarbital and phenobarbital. Scheme 1. Synthesis of the Barbiturate Immunogen
reagents and the hazards of working with, storage of, and disposal of radioactive materials (7). Alternatively, fluorescence polarization immunoassay (FPIA) is ideally suited for automation and employs fluorescent tracers and antibodies, thus avoiding the disadvantages associated with enzyme sensitivity and the hazards of radiolabeled tracers (8). The quality of a screening immunoassay greatly depends on the availability of antibodies that are specific for the target class of analytes. In this work, we present the design and synthesis of a hapten that shows the structural properties of both the aliphatic and aromatic classes of barbiturates (Scheme 1). This hapten was conjugated to prepare an immunogen useful for eliciting antibodies in sheep. A complementary fluorescent tracer was also synthesized (Scheme 2). The binding characteristics of the obtained antibodies were assessed with various barbiturates in competition with the fluorescent tracer. © 1997 American Chemical Society
282 Bioconjugate Chem., Vol. 8, No. 3, 1997 Scheme 2. Synthesis of the Barbiturate Tracer
MATERIALS AND METHODS
All reagents were purchased from Aldrich Chemical Co., Inc., Milwaukee, WI, and were used without further purification, except where noted. Solvents (HPLC grade) and silica gel were obtained from E. Merck Science, Darmstadt, Germany, and were used without further purification. 1H NMR and 13C NMR were recorded at 300 and 75 MHz, respectively, on a Varian Gemini 300 spectrometer with tetramethylsilane (TMS) as internal standard. Electrospray mass spectrometry (ESMS) was performed on a Perkin-Elmer Siex API III. HPLC was performed on a Waters RCM C18 (8 × 100 mm) reversed phase µBondapak column at 2 mL/min (detection at 220 nm) with the solvent indicated. Preparation of the Barbiturate Immunogen (5). Dicyclopentadiene (50 mL) was depolymerized (9), providing 31.22 g of cyclopentadiene. To the cyclopentadiene cooled to -78 °C was added gaseous HCl (9) in portions until the weight had increased to 46.4 g. The resulting crude 3-chlorocyclopentene was added in portions to a preformed solution of sodium diethylmalonate, prepared by adding diethyl malonate (73 mL, 0.48 mol) to a solution of sodium (11.0 g, 0.48 mol) in ethanol (200 mL). Each addition of the 3-chlorocyclopentene produced a noticeably exothermic reaction. After the addition was complete, the resulting yellow suspension was refluxed for 14 h. Most of the ethanol was removed by rotary evaporation from the cooled solution, and the residue was diluted with 300 mL of water. The solution was extracted with ether (3 × 100 mL), and the combined ether extracts were washed with water (2 × 50 mL), dried over MgSO4, and concentrated to a red oil. Distillation provided 45.194 g (42%) of diethyl cyclopentenylmalonate (1) as a colorless oil: bp 99-103 °C (0.8 mm); 1H NMR (CDCl3) δ 5.81 (ddd, olef H, 1H, J ) 5.7, 4.5, and 2.3 Hz), 5.66 (ddd, olef H, 1H, J ) 5.7, 4.3, and 2.1 Hz), 4.18 (q, ethyl CH2, 4H, J ) 7.1 Hz), 3.34 (m, allyl CH, 1H), 3.22 (d, 1H, malonate CH, J ) 9.6 Hz), 2.34 (m, allyl CH2, 2H), 2.13 (m, aliph CH, 1H), 1.60 (m, aliph CH, 1H), 1.25 (t, ethyl CH3, 6H, J ) 7.2 Hz); 13C NMR (CDCl3) δ 168.9, 168.8, 132.9, 131.6, 61.1, 57.0, 45.2, 31.6, 27.7, 14.0; ESMS, m/z 227 (M + H)+, 244 (M + NH4)+.
Adamczyk et al.
To a solution of sodium (3.51 g, 153 mmol) in ethanol (50 mL) was added urea (3.05 g, 51 mmol) and diethyl cyclopentenylmalonate (1, 10.00 g, 44.2 mmol). After 14 h of refluxing, the white suspension was dissolved in water and acidified to pH ∼2. The off-white solid collected by filtration was crystallized from ∼200 mL of 30% ethanol/water, to provide 5.656 g (66%) of cyclopentenylbarbituric acid (2) as tan plates, mp 192-193 °C. An additional 1.314 g of light brown powdery crystals (17%) with mp 186-188 °C was obtained by concentration to 80 mL and crystallization over 14 h. Both materials showed the following: 1H NMR (DMSO-d6) δ 11.17 (s, barb. NH, 1Η), 11.13 (s, barb. NH, 1H), 5.80 (m, olef CH, 1H), 5.59 (m, olef CH, 1H), 3.30 (br s, allyl CH + barb. CH, 2H), 2.22-2.08 (m, allyl CH2, 2H), 2.011.78 (m, aliph. CH2, 2H); 13C NMR (DMSO-d6) δ 170.4, 170.2, 151.3, 133.2, 130.4, 52.2, 48.5, 31.7, 26.2; ESMS, m/z 195 (M - H)-. Sodium hydride (95%, 135 mg, 5.36 mmol) was suspended in 2 mL of tetrahydrofuran (THF), and a solution of 5-cyclopentenylbarbituric acid (2, 1.00 g, 5.10 mmol) in 18 mL of dimethylformamide (DMF) was added. After 4 h of stirring, ethyl 3-bromopropionate (0.72 mL, 5.61 mmol) and potassium iodide (169 mg, 1.02 mmol) were added, and the reaction was refluxed for 24 h. The reaction mixture was poured into 100 mL of water and extracted with ethyl acetate (3 × 50 mL). The combined extracts were washed with water (2 × 50 mL), dried over MgSO4, and concentrated. Chromatography on a 2 × 18 cm column of silica gel, eluting with 100 mL each of 30, 40, and 50% EtOAc/hexanes, provided 1.159 g (77%) of the disubstituted barbituric acid ester (3) as a white solid: mp 93-94 °C; 1H NMR (CDCl3) δ 8.77 (s, barb. NH, 1Η), 8.67 (s, barb. NH, 1Η), 5.93 (m, olef CH, 1H), 5.64 (m, olef CH, 1H), 4.08 (q, ethyl CH2, 2H, J ) 7.5 Hz), 3.29 (m, allyl CH, 1H), 2.42-2.25 (m, allyl CH2 + propionate CH2CH2, 6H), 2.05-1.97 (m, aliph cyclopentenyl CH2, 2H), 1.21 (t, ethyl CH3, 3H, J ) 7.2 Hz); 13C NMR (CDCl3) δ 172.6, 171.3, 170.9, 149.2, 136.0, 60.8, 57.9, 56.6, 31.9, 29.7, 27.5, 24.5, 14.0; ESMS, m/z 293 (M - H)-. The barbiturate ester (3, 200 mg, 0.68 mmol) was suspended in a solution of 1.0 mL of 10% HCl in 1.0 mL of dioxane and stirred for 24 days at ambient temperature. HPLC showed that >90% of a single product had been formed. The resulting solution was purified by preparative reversed phase HPLC on a C18 µBondapak 40 mm × 100 mm PrepPak column, eluting at 40 mL/ min with 35% acetonitrile/65% 0.05% CF3COOH in water, to produce 167 mg (92%) of the acid 4 as a white solid: 1H NMR (DMSO-d6) δ 12.10 (br s, COOH, 1H), 11.43 (s, barb. NH, 1H), 11.37 (s, barb. NH, 1H), 5.88 (m, olef CH, 1H), 5.54 (m, olef CH, 1H), 3.11 (s, allyl CH, 1H), 2.19-2.12 (allyl CH2, 2H), 2.05-1.80 (m, aliph CH2s, 6H); 13C NMR (CDCl3) δ 173.7, 172.0, 171.8, 150.3, 134.6, 56.7, 55.5, 31.6, 29.5, 27.2, 24.1; ESMS, m/z 265 (M H)-. To a solution of the acid (4, 20.7 mg, 77 µmol) in p-dioxane (1.6 mL) were added isobutyl chloroformate (17 µL, 134 µmol) and triethylamine (17 µL, 125 µmol). After 2 h of stirring at ambient temperature, the dioxane mixture solution was added to a rapidly stirred solution of 318 mg of bovine thyroglobulin in 17 mL of 0.05 M, pH 9.5, borate buffer. After 2 h of stirring at ambient temperature, the solution was dialyzed against 0.05 M, pH 7.5, phosphate (five changes, at least 8 h between changes), to produce the immunogen 5. TNBS titration (10) indicated that 70% of the available amino groups had been modified.
Synthesis of Conjugates for a Barbiturate Screening Assay
Preparation of the Barbiturate Tracer (10). Sodium (9.2 g, 0.40 mol) was carefully dissolved in ethanol (150 mL), and diethyl malonate (60 mL, 0.40 mol) was added. After a brief period of stirring, cyclopentyl bromide (44 mL, 0.40 mol) was added, and the reaction mixture was refluxed for 14 h. The suspension was poured into 100 mL of water and extracted with ethyl acetate (3 × 100 mL). The combined extracts were dried over MgSO4 and concentrated. Distillation afforded 66.535 g (73%) of diethyl cyclopentylmalonate (6) as a clear colorless oil: bp 104-106 °C (3 mm Hg); 1H NMR (CDCl3) δ 4.16 (q, ethyl CH2, 4H, J ) 7.1 Hz), 3.15 (d, malonate CH, 1H, J ) 10.3 Hz), 2.45 (m, cyclopent CH, 1H), 1.85-1.51 (m, cyclopent CH2, 8H), 1.24 (t, ethyl CH3, 6H, J ) 7.2 Hz); 13C NMR (CDCl3) δ 169.2, 61.0, 57.3, 39.4, 30.6, 24.8, 13.9; ESMS, m/z 229 (M + H)+, 246 (M + NH4)+. Sodium (11.3 g, 490 mmol) was carefully dissolved in 150 mL of ethanol, and urea (9.80 g, 163 mmol) was added, followed by diethyl cyclopentylmalonate (32.409 g, 142 mmol). After 14 h of refluxing, the suspension was poured into water, and the mixture was acidified to pH ∼2. The white crystals collected by filtration were recrystallized from ∼200 mL of 30% ethanol/water, to provided 21.243 g (80%) of cyclopentylbarbituric acid (7) as white plates: mp 222-223 °C; 1H NMR (DMSO-d6) δ 9.56 (br s, barb. NH, 2H) 3.37 (d, 1H, barb. CH, J ) 10.8 Hz), 2.42 (m, cyclopent CH, 1H), 1.71-1.44 (m, cyclopent CH2, 8H); 13C NMR (DMSO-d6) δ 170.8, 151.4, 51.7, 42.4, 29.0, 24.3; ESMS, m/z 195 (M - H)-. Sodium hydride (95%, 271 mg, 10.7 mmol) was suspended in DMF, and a solution of 5-cyclopentylbarbituric acid (2.00 g, 10.2 mmol) in 40 mL of 85% DMF/THF was added. After 3 h of stirring, ethyl 4-bromocrotonate (1.54 mL, 11.2 mmol) and potassium iodide (423 mg, 2.55 mmol) were added, and the reaction was refluxed for 14 h. The reaction was poured into 100 mL of water and extracted with ethyl acetate (2 × 100 mL). The combined extracts were washed with water (3 × 50 mL), dried over MgSO4, and concentrated. Chromatography on a 3 × 30 cm column of silica gel, eluting with 100 mL each of 20, 30, 40, 50, and 70% EtOAc/hexanes, provided 1.917 g of a clear pale yellow oil. Crystallization of this material from EtOAc/hexanes provided 1.602 g (51%) of disubstituted barbiturate ester (8) as a white solid: mp 141142 °C; 1H NMR (CDCl3) δ 7.98 (br s, barb. NH, 2H), 6.65 (dt, 1H, β-olef CH, J ) 14.3 and 7.5 Hz), 5.90 (d, R-olef CH, 1H, J ) 14.3 Hz), 4.15 (q, ethyl CH2, 2H, J ) 7.2 Hz), 2.92 (dd, allyl CH2, 2H, J ) 7.7 and 1.1 Hz), 2.43 (m, cyclopent CH, 1H), 1.78-1.49 (m, cyclopent CH2, 8H), 1.26 (t, ethyl CH3, 3H, J ) 7.1 Hz); 13C NMR (CDCl3) δ 171.0, 166.2, 149.3, 126.3, 60.6, 57.7, 50.0, 36.2, 27.1, 24.1, 21.1, 14.0; ESMS, m/z 307 (M - H)-. The barbiturate ester (8, 200 mg, 0.65 mmol) was suspended in 12 mL of concentrated HCl and refluxed for 45 min. Concentration furnished a white solid, which was crystallized from water to yield 173 mg (95%) of the disubstituted barbiturate acid (9) as white needles: mp 217-218 °C; 1H NMR (DMSO-d6) δ 11.55 (br s, NH and COOH, 3H), 6.48 (dt, β-olef CH, 1H, J ) 15.5 and 7.7 Hz), 5.72 (d, R-olef CH, 1H, J ) 15.5 Hz), 2.76 (d, allyl CH2, 1H, J ) 7.0), 2.26 (m, cyclopent CH, 1H), 1.721.34 (m, cyclopent CH2, 8H); 13C NMR (CDCl3) δ 181.8, 176.7, 160.0, 152.1, 135.7, 66.3, 58.8, 45.5, 36.8, 34.0; ESMS, m/z 279 (M - H)-. To a solution of the acid 9 (20 mg, 71 µmol) in DMF (100 µL) were added N-hydroxysuccinimide (11 mg, 93 µmol) and dicyclohexylcarbodiimide (16 mg, 78 µmol). After 8 h of stirring, 4′-aminomethylfluorescein hydrochloride (28 mg, 71 µmol) and triethylamine (10 µl, 71
Bioconjugate Chem., Vol. 8, No. 3, 1997 283
µmol) were added, and the suspension was stirred for 96 h. Concentration furnished an orange solid, which was dissolved in methanol and purified by preparative reversed phase HPLC on a C18 µBondapak 40 mm × 100 mm PrepPak column, eluting at 40 mL/min with 38% CH3CN/62% 0.05% CF3COOH in water (220 nm), to produce 26 mg (59%) of the fluorescent tracer 10 as a bright yellow solid: RT ) 5.68 min (45% CH3CN/55% 0.05% CF3COOH in water); ESMS, m/z 622 (M - H)-. Immunization of Animals and Preparation of Antisera. Eight sheep were initially immunized with 1 mg of immunogen in 1.0 mL of complete Freund’s adjuvant. The sheep were subsequently bled 7-10 days after immunization and then boosted monthly with 0.5 mg of the immunogen in 0.5 mL of incomplete Freund’s adjuvant until the response was mature, after which time the animals were boosted with 0.5 mg of the immunogen monthly. From the blood obtained from each animal every 2 weeks, the serum was separated and stored at -20 °C. Antisera titers rose slowly and demonstrated consistent titer after about 6 months. Production bleeds (∼200 mL) collected for 16 weeks were combined to give the antisera pool, which is characterized below. Evaluation of the Pooled Antisera. The antisera pool was evaluated by fluorescence polarization immunoassay on an Abbott TDx analyzer equipped with Revision 15 software and operated as described in the TDx system operation manual (8) and Popelka et al. (11) using TDx dilution buffer (0.1 M phosphate buffer, pH 7.4, with 0.1% bovine γ-globulin and 0.1% sodium azide). Except where noted, antisera and tracer were added to the internal 3-POT serum vial (S) and tracer vial (T) containers, respectively. Analyte was added to the sample well of the disposable sample cartridge. Polarization measurements were recorded in millipolarization units (mP). To measure antisera activity, serial dilutions of antibody were added directly to the cuvette, and polarization measurements were made with and without analyte. In these experiments, the serum vial was filled with TDx dilution buffer. Displacement (∆mP) was calculated by taking the difference between the polarizations of tracer binding without analyte minus tracer binding with analyte. The working titer was selected from the dilution having the maximum displacement. The antisera were further evaluated by comparing the zero calibrator and the A-F span. Once consistent pool-to-pool titer and performance were obtained, the antisera pools were evaluated on the AxSYM analyzer. Standards were prepared by the addition of analytically pure secobarbital to normal human urine diluent (confirmed to be negative for barbiturates) at concentrations of 0, 200, 400, 700, 1200, and 2000 ng/mL. Using a four-parameter weighted curve fit (duplicate determinations), a calibration curve was constructed (Figure 2). Mean responses to these standards were established using multiple instruments and stored in barcode form for each antisera pool. The two-point Master Calibration Curve thus derived is specific for each pool and refits the standard six-point calibration curve based on the response of an individual instrument to the two-point curve. Sensitivity of the assay was determined by identifying the lowest measurable concentration of drug in human urine that could be distinguished from a sample known to contain 0 ng/mL of the drug with a 95% confidence limit (n ) 20). Replicates of a human urine sample known to be negative for barbiturates and the 0 ng/mL standard were both analyzed. The standard deviation of the replicates was then calculated, and twice this
284 Bioconjugate Chem., Vol. 8, No. 3, 1997
Adamczyk et al. Table 1. Cross-Reactivities of Barbiturates in Barbiturate Screening Assay
compd
Figure 2. Standard curve with secobarbital as calibrator.
number was subtracted from the average of the replicates in each case. The antisera was tested for binding with several different barbiturates. Percent cross-reactivity was calculated as 100× the measured concentration of drug divided by the concentration of analyte added. RESULTS
Immunogen and Tracer Synthesis. The first step toward development of an immunoassay is to prepare a carefully designed immunogen. This was accomplished by selective conjugation of the unsaturated acid (4, Scheme 1) linked at its 5-position to bovine thyroglobulin. This acid was derived from ester (3) via hydrolysis at ambient temperature. Attempted hydrolysis at higher temperatures resulted in lower yields of difficultly purified products. Formation of cyclopentenylbarbituric acid (2) was derived ultimately from dicyclopentadiene, via cracking to cyclopentadiene (9), low-temperature addition of HCl gas (9), and addition of the resulting allylic chloride to a preformed solution of sodium diethylmalonate. Formation of the monosubstituted barbituric acid was performed prior to propionate alkylation instead of attempting to form the barbituric acid ring with a bisalkylated malonate, since the presence of the third propionate ester could result in side reactions and lower yields in the formation of the barbiturate ring. The isobutyl mixed anhydride was formed by reaction with isobutyl chloroformate and triethylamine in pdioxane, and reaction with bovine thyroglobulin produced the desired immunogen (5). Analysis by TNBS (10) showed that 70% of the available amino groups of the thyroglobulin had been modified. The multistep synthesis of the tracer proceeded along a similar path (Scheme 2). Cyclopentylmalonate (6) was synthesized by alkylation with commercially available cyclopentyl bromide. The second alkylation was again performed after barbiturate formation for the same reasons noted above. Thus, the crotonate side chain was introduced by alkylation of barbituric acid 7 with sodium hydride in the presence of potassium iodide. Carboxylic acid 9 produced via acid hydrolysis of ester 8 was activated as described above and reacted with 4′-aminomethylfluorescein to produce the fluorescent tracer 10. Antisera Characterization. All animals responded by producing consistently titered antiserum after about 6 months. The maximum displacement identified for the mature pool of antisera was obtained at 1/76000 dilution, which gave a dynamic range of 107 mP (Figure 2) when tested against secobarbital. Using a two-point calibration curve, negative urine yielded a sensitivity of 34 ng/mL, and the 0 ng/mL standard a sensitivity of 37 ng/mL. With a standard six-point calibration curve, values of 24 and
concn added (ng/mL)
concn obsd (ng/mL)
% crossreactivity
cyclopentobarbital butabarbital talbutal alphenal butalbital secobarbital brallobarbital aprobarbital pentobarbital phenobarbital
200 200 200 200 200 200 200 200 200 200
898.9 420.2 361.7 312.9 226.9 200.0 133.3 128.0 107.5 105.6
449.45 210.10 180.85 156.45 113.45 100.00 66.65 64.00 53.75 52.83
butobarbital allobarbital amobarbital
400 400 700
199.2 90.9 183.6
49.80 22.73 26.23
2000 2000 2000
140.3 94.9 70.6
7.02 4.75 3.53
100000 1000000 1000000
291.6 438.3 344.3
0.292 0.044 0.034
thiopental veronal 5-ethyl-5-(4-hydroxyphenyl)barbituric acid hexobarbital metharbital methohexital
Table 2. Cross-Reactivities of Structurally Similar Compounds
compd glutethimide primidone aminoglutethimide 5-(p-hydroxyphenyl)-5-phenylhydantoin (HPPH)
concn added (ng/mL)
concn obsd (ng/mL)
% crossreactivity
10000 100000 100000 500000
451 248.7 203.9 243.6
4.510 0.250 0.204 0.049
27 ng/mL were obtained. These values are statistically indistinguishable. The assay was thus able to detect concentrations of secobarbital at 22% (Table 1, upper two groups). Other barbiturates showed lower cross-reactivity: veronal, thiopental, and 5-ethyl-5-(4-hydroxyphenyl)barbituric acid (4-hydroxyphenobarbital, a phenobarbital metabolite) showed limited (3-7%) cross-reactivity at 2000 ng/mL, while hexobarbital, metharbital, and methohexital showed (
