Synthesis, characterization and ... - Wiley Online Library

JOURNAL OF MOLECULAR RECOGNITION J. Mol. Recognit. 2003; 16: 148–156 Published online in Wiley InterScience (www.interscience.wiley.com). DOI:10.1002/jmr.621

Synthesis, characterization and immunochemical evaluation of cephalosporin antigenic determinants Francisco Sa´nchez-Sancho1†, Ezequiel Perez-Inestrosa1*, Rafael Suau1, Marı´a I. Montan˜ez1, Cristobalina Mayorga2, Maria J. Torres2, Antonino Romano3 and Miguel Blanca4 1

Department of Organic Chemistry. Faculty of Science. University of Malaga, Malaga, Spain Research Unit for Allergic Diseases, Carlos Haya Hospital, Malaga, Spain 3 Department of Internal Medicine and Geriatrics, UCSC-Allergy Unit, C.I. Columbus and IRCCS Oasi Maria S.S. Troina, Italy 4 Allergic Service, Hospital Universitario La Paz, Madrid, Spain 2

Lack of knowledge of the exact chemical structure of cephalosporin antigenic determinants has hindered clinical interpretation of adverse reactions to these drugs and delayed understanding of the mechanisms involved in the specific recognition and binding of IgE molecules to these antigenic determinants. We further resolve the relationship between structure and activity of proposed antigenic chemicals, including the rational design and synthesis of these haptenic structures. Comparative RAST inhibition studies of the synthesized molecules revealed that they were recognized by IgE antibodies induced by cephalosporin antibiotics. Thus, these data indicate that recognition is mainly directed to the acyl side chain and to the blactam fragment that remains linked to the carrier protein in the cephalosporin conjugation course. Copyright # 2003 John Wiley & Sons, Ltd. Keywords: immediate reactions; lactams; IgE; antibiotics; molecular recognition; cephalosporins; immunochemistry Received 31 October 2002; revised 3 March 2003; accepted 4 March 2003

INTRODUCTION b-Lactam antibiotics, mainly penicillins and cephalosporins (Fig. 1), are the drugs that most frequently cause adverse reactions by specific immunologic mechanisms (Dewdney, 1977; Weiss and Adkinson, 1988). The great use of these antibiotics, as well as the wide diversity of their chemical structures, has led to the appearance of immunologic responses that recognize each of these structures as different, with clinical relevance in the induction of allergic reactions (Blanca, 1995). Of all the b-lactams, penicillins are the best studied so far (Weiss and Adkinson, 1988; Blanca et al., 1994), probably due to their greater prescription and consumption (Batchelor et al., 1965; Levine and Ovary, 1961). Since penicillins have very similar chemical structures, differing only in the nature of the side-chain group (Fig. 1), benzylpenicillin has traditionally been considered as the reference model for the study of allergy to b-lactams. The ‘major’ antigenic determinant is benzylpenicilloyl (Fig. 1, R=PhCH2), formed by the nucleophilic opening of the b-lactam ring by the amino group of proteins found in either plasma or cell membranes (Dewdney, 1977; Batchelor et al., 1965; Moreno *Correspondence to: E. Perez-Inestrosa, Department of Organic Chemistry, Faculty of Science, University of Malaga, E-29071 Malaga, Spain. E-mail: [email protected] †Current address: Pharma Mar, S.A. Madrid, Spain. Contract/grant sponsor: FEDER; contract/grant number: 1FD97-0516. Abbreviations used: AMP, ampicillin; AX, amoxycillin; MDM, minor determinant mixture; PG, penicillin G; PLL, poly-L-lysine; PPL, penicilloylpolylysine; RAST, radioallergosorbent test.

Copyright # 2003 John Wiley & Sons, Ltd.

et al., 1995), resulting in a covalent binding of the antibiotic to the carrier protein to yield the hapten-carrier conjugate. Nevertheless, a number of studies have shown that betalactam structures with different side chains contribute to the specific response, and it has become clear that these new conjugates formed with these penicillins have also to be considered for immunologic tests (Moreno et al., 1995). Although cephalosporins have an increasing role in the induction of allergic reactions (Romano et al., 2000), they have barely been investigated and the determinants involved in the immunological responses are largely unknown. Despite the structural similarities between penicillins and cephalosporins, they differ considerably in the products obtained from the nucleophilic opening of the b-lactam ring. The equivalent ‘major’ antigenic determinant for cephalosporins would be the cephalosporoyl (Batchelor et al., 1966). However, this compound is unstable and suffers an extensive fragmentation process (Fig. 1). The immunologic behavior of these antibiotics is determined by their intrinsic chemical reactivity, which is related with the capacity of the b-lactam carbonyl group to act as the acylating agent to form the corresponding antigenic determinants with protein amino groups (Batchelor et al., 1966; Kaiser and Kukolja, 1972). The reactivity of this group is higher in penicillins, due to the strain in the chemical structure caused by the cis fusion between the two heterocycles of four and five atoms. In cephalosporins the heterocycles are of four and six atoms, which results in less strain in the b-lactam ring (Lowe, 1979). Thus, in penicillins the nucleophilic ring opening takes place more quickly and effectively. Furthermore, the conjugate obtained in this way, the ‘major’ antigenic determinant, is stable enough to be

CEPHALOSPORIN ANTIGENIC DETERMINANTS

149

Figure 1. Structures of some frequently used penicillins and cephalosporins and their nucleophilic ring opening. Whereas the penicilloyl derivative is a stable and chemically well-characterized structure, the cephalosporoyl analog evolves to further uncharacterized end products.

isolated and characterized by common methods. However, in cephalosporins, the lower reactivity of the b-lactam ring slows the reaction. Nevertheless, the existence of a good leaving group at position 3' of the dihydrothiazine ring, which occurs in most clinically important cephalosporins, enhances the reactivity of the b-lactam carbonyl via the associated elimination of R2. In fact, the rate of opening of the b-lactam ring and the antibacterial activity both increase as the ability of the group at 3' to act as a leaving group increases (Kaiser and Koklja, 1972; Holden, 1984). Experimental and theoretical studies have been interpreted in terms of b-lactam ring opening non-concerted with the departure of the leaving group R2 (Faraci and Pratt, 1984). Nevertheless, no intermediate metabolite corresponding to cephalosporoyl has yet been isolated and fully characterized. The conjugate obtained this way is unstable and Copyright # 2003 John Wiley & Sons, Ltd.

undergoes a process of multiple fragmentation in the dihydrothiazine portion, which leads to a great number of degradation products, more so if the great variety of chemical structures for cephalosporins is taken into account (Fig. 1). This complicates the isolation and characterization of the haptenic structures and the possible antigenic determinants formed. Several studies have attempted to elucidate the possible chemical structures obtained after the reaction of different cephalosporins with ammonia, amino acids and simple amino compounds, and under hydrolysis conditions on the basis of UV (Hamilton-Miller et al., 1970a) and NMR spectra (Grabowski et al., 1985; Vilanova et al., 1993; Llina´s et al., 1998), as well as isolation by HPLC analysis (Vilanova et al., 1993; Llina´s et al., 1998). Thus, structures such as polymers, piperazinediones, penaldates and penaJ. Mol. Recognit. 2003; 16: 148–156

150

´ NCHEZ-SANCHO ET AL. F. SA

maldates have been proposed (Dewdney, 1977; HamiltonMiller et al., 1970a; Grabowski et al., 1985; Vilanova et al., 1993; Llina´s et al., 1998; Manhas and Bose, 1971). However, the assignments are somewhat tentative and ammonolysis (Grabowski et al., 1985) in liquid ammonia at 50 °C has characterized the intermediate of cephamycin. Initially the cephalosporin undergoes b-lactam cleavage and has a sufficient lifetime for carbon-13 spectral characterization. Direct observation of cephalosporoate intermediates in aqueous solutions has been also reported (Pratt and Faraci, 1986). The combination of absorption and NMR spectral evidence shows that, in those cases investigated, 3' elimination from cephalosporins is not concerted. These cephalosporoates are chemical compounds whose stability and lability depend on the nature of the R2 substituents. The authors conclude that, with better leaving groups than those examined, a concerted elimination might occur. As a result, it is claimed that comparison of the hydroxide-catalyzed elimination rate constant suggests that the intermediate cephalosporoate would not be seen during alkaline hydrolysis of cephalosporins with good 3' leaving groups. The production of monoclonal antibodies has shown that cephalosporins can generate unique structures capable of inducing a specific immunologic response without crossreacting with the classical structures (Nagakura et al., 1990). Consequently, allergic reactions to cephalosporins may occur by sensitization to determinants similar to those of penicillins or to specific cephalosporin haptens (Blanca et al., 1994; Pham and Baldo, 1996; Baldo A, 1999). On the other hand, there is evidence that the C-7 substitution (R1 acyl chain) plays a dominant role in determining the specificity of immunologic reactions between individual cephalosporins and between penicillins and cephalosporins to a greater degree than that observed for penicillins (Manhas and Bose, 1971).

EXPERIMENTAL Chemical synthesis General. Melting points were determined on a Gallenkamp instrument and are given uncorrected. MS (EI) were recorded on an HP-MS 5988A spectrometer operating at 70 eV. HRMS were recorded on an AutoSpecE, CACTI, University of Vigo (Spain). NMR spectra were recorded on a Bruker WP-200 SY instrument at 200 MHz for 1H and 50.3 MHz for 13C. Chemical shifts are given relative to the residual signal of solvents, dH 7.24 ppm and dC 77.0 ppm for deuteriochloroform. NOE experiments were recorded on a Bruker (500 MHz), University of Santiago (Spain). Optical rotations were determined by using a Perkin-Elmer 241 digital polarimeter. Elemental analyses were carried out in the Microanalytical Laboratory, SCAI, University of Malaga. TLC analyses were performed on silica gel 60 F 256 plates and column chromatography was carried out on silica gel 60 (70–230 mesh). Organic solutions were dried with MgSO4 and concentrated in vacuo. Compounds 2 (Paruszewski et al., 1996) and 3 (Wagatsuma et al., 1973) were prepared according to the sequence Boc-Ala-OH → Boc-Ala-nBu(2) → Ala-nBu(3) by proceCopyright # 2003 John Wiley & Sons, Ltd.

dures common for the preparation of peptides and showed spectroscopic data according to that reported in the literature. The following compounds were prepared according to literature procedures: 4a (Bucourt et al., 1978); 4c (Boger et al., 1997) and 4f (Kuisle et al., 1999). 4b, 4d and all other reagents were purchased from Aldrich, Sigma or Fluka. Procedure for the synthesis of compounds 5a–f. Solutions of the corresponding acids 4a–f were sequentially treated with isobutylchloroformate and the amine, 3. After workup, the corresponding pure compounds 5a–f were obtained. (2S)-N1-butyl-2-({(methoxyimino)[2-(trytilamino)-1,3thiazol-4-yl]acetyl}amino)propanamide (5a): white solid (635 mg, 60% yield); m.p. = 96–100 °C; [a]D = 37.0 (c = 1 in CHCl3). Analysis calculated for C32H35N5O3S—C, 67.46; H, 6.19; N, 12.39; S, 5.63; found—C, 67.54; H, 6.08; N, 12.19; S, 5.34. (2S)-N1-butyl-2-[2-(thienylacetyl)amino]propanamide (5b): white solid (225 mg, 40% yield); recrystallized with CHCl3-hexane; m.p. = 154–155 °C; [a]D = 43.7 (c = 1, MeOH). Analysis calculated for C13H20N2O2S—C, 58.18; H, 7.51; N, 10.44; S, 11.95; found—C, 58.03; H, 7.49; N, 10.19; S, 11.84. HRMS (EI) calculated for C13H20N2O2S— 268.1245; found—268.1245. (2S)-2-{[(2R)-(tert-butoxycarbonylamino)(4-hydroxyphenyl)acetyl]amino}-N1-butylpropanamide (5c): white solid (243 mg, 55% yield); m.p. = 172–174 °C; [a]D = 111.8 (c = 1.7, CHCl3). Analysis calculated for C20H31N3O5—C, 61.05; H, 7.94; N, 10.68; found—C, 59.94; H, 8.06; N, 10.25. (2S)-2-{[(2R)-(tert-butoxycarbonylamino)(phenyl) acetyl]amino}-N1-butylpropanamide (5d): white solid (310 mg, 80% yield); recrystallized with CH2Cl2-Et2O; m.p. = 194–196 °C; [a]D = 105.9 (c = 1, CHCl3). Analysis calculated for C20H31N3O4—C, 63.64; H, 8.28; N, 11.13; found—C, 63.31; H, 8.48; N, 11.01. (2S)-N1-butyl-2-{[(2R)-phenyl(tetrahydro-2H-pyran-2-yloxy)acetyl]amino}propanamide (5e): thick oil (219 mg, 57% yield). Analysis calculated for C20H30N2O4—C, 66.27; H, 8.34; N, 7.73; found—C, 66.13; H, 8.43; N, 7.68. (2S)-N1-butyl-2-{[2-furyl(oxo)acetyl]amino}propanamide (5f): yellow solid (760 mg, 40% yield); m.p. = 84–86 °C; [a]D = 82.1 (c = 1, CHCl3). Analysis calculated for C13H18N2O4—C, 58.63; H, 6.81; N, 10.52; found—C, 58.63; H, 6.91; N, 10.51. Preparation of compounds 1a–f. (2S)-2-{[(2-amino-1,3thiazol-4-yl)(methoxyimino)acetyl]amino}-N1-butylpropanamide (1a): compound 5a was treated with a 50% aqueous formic acid solution for 30 min at 70 °C. After work-up, pure 1a (71%) was obtained: white solid; m.p. = 83–86 °C; [a]D = 46.9 (c = 1, MeOH). 1H NMR (200 MHz, CD3OD): = 7.05 (s, 1H), 4.51 (q, 1H, J = 7.3 Hz), 4.05 (s, 3H), 3.31 (m, 2H), 1.70–1.30 (m, 4H), 1.48 (d, 3H, J = 7.3 Hz), 1.01 (t, 3H, J = 7.3 Hz). 13C NMR (50 MHz, CD3OD): = 174.4, 171.5, 165.0, 150.3, 143.6, 111.9, 63.2, 50.7, 40.3, 32.6, 21.0, 17.8, 14.2. MS (EI) m/z (%): 327 (7) [M], 296 (3), 227 (92), 197 (52), 156 (80), 125 (100), 83 (23). Analysis calculated for C13H21N5O3S—C, 47.69; H, 6.47; N, 21.39; S, 9.79; found— C, 47.84; H, 6.12; N, 21.05; S, 9.62. HRMS (EI) calculated for C13H21N5O3S—327.1365; found— 327.1361. J. Mol. Recognit. 2003; 16: 148–156


Hydrolysis of tert-butoxycarbonyl group. General procedure for the synthesis of compounds 1c and 1d: an ethyl acetate solution of the corresponding 5c and 5d was treated with 3 M aqueous HCl solution, at room temperature. Quantitative yield, white solids. (2S)-2-{[(2R)-amino(4-hydroxyphenyl)acetyl]amino}N1-butylpropanamide, hydrochloride (1c): m.p. = 182–184 °C; [a]D= 85.0 (c = 0.2, CH3OH). 1H NMR (200 MHz, D2O): = 7.35 and 6.96 (AA'BB', 4H), 5.08 (s, 1H,), 4.24 (b q, 1H, J = 7.3 Hz), 3.20 (b t, 1H, J = 6.7 Hz), 1.54–1.13 (m, 4H), 1.25 (d, 3H, J = 7.3 Hz), 0.87 (t, 3H, J = 7.3 Hz). 13C NMR (50 MHz, D2O): = 175.4, 169.6, 158.2, 130.8, 124.6, 117.3, 57.0, 51.2, 40.1, 31.4, 20.3, 17.5, 13.9. Analysis calculated for C15H24ClN3O3—C, 54.62; H, 7.33; N, 12.74; found—C, 54.43; H, 7.01; N, 12.51. (2S)-2-{[(2R)-amino(phenyl)acetyl]amino}-N1-butylpropanamide, hydrochloride (1d): m.p. = 196–198 °C; [a]D = 50.0 (c = 0.5, CH3OH). 1H NMR (200 MHz, D2O): = 7.51 (s, 5H), 5.13 (s, 1H), 4.26 (q, 1H, J = 7.3 Hz), 3.21 (t, 2H, J = 7.3 Hz), 1.59–1.19 (m, 4H), 1.26 (d, 3H, J = 7.3 Hz), 0.89 (t, 3H, J = 7.3 Hz). 13C NMR (50 MHz, D2O): = 175.4, 169.4, 132.8, 131.4, 130.7, 128.9, 57.5, 51.2, 40.2, 31.4, 20.3, 17.5, 13.9. Analysis calculated for C15H24ClN3O2—C, 57.41, H, 7.71; N, 13.39; found— C, 57.03; H, 7.75; N, 13.26. (2S)-N1-Butyl-2-{[(2R)-hydroxy(phenyl)acetyl]amino} propanamide (1e): compound 5e was treated with p-TsOH in MeOH at room temperature (80% yield); m.p. = 138–140 °C; [a]D = 94.1 (c = 1, CHCl3). 1H NMR (200 MHz, D2O): = 7.32 (m, 5H), 7.05 (b d, 1H, J = 6.7 Hz), 6.18 (b t, 1H, J = 4.9 Hz), 5.02 (d, 1H, J = 3.5 Hz), 4.41 (dq, 1H, J = 6.7 Hz, J = 7.3 Hz), 3.68 (d, 1H, J = 3.5 Hz), 3.05 (distorted q, 1H, J = 6.7 Hz), 1.40–1.07 (m, 4H), 1.31 (d, 3H, J = 7.3 Hz), 0.81 (t, 3H, J = 7.3 Hz). 13C NMR (50 MHz, CDCl3): = 172.3, 171.8, 139.4, 128.7, 128.5, 126.7, 74.3, 48.7, 39.2, 31.3, 19.8, 18.1, 13.6. MS (EI) m/z (%): 278 (3 mm in diameter was present 20 min later. When prick tests were negative, 0.01 ml of the reagent solution was injected intradermally in volar forearm skin. Readings were made 20 min after injections. Results were considered positive when there was an increase in the wheal of >3 mm. Positive controls for prick and intradermal tests were made with histamine (at 10 and 1 mg/mL, respectively). Normal saline was used as a negative control. The concentration used for cephalosporins had proved to be nonirritant in a control group of 40 healthy subjects. Radioimmunoassay for IgE determination. These were made by RAST to PG, AX, AMP and to different cephalosporins (ceftazidime, ceftriaxone, cefuroxime, cefotaxime, cefadroxile, cephalexine) conjugated to poly-Llysine (PLL) (Sigma) as previously described (Blanca et al., 1992). Blood samples were obtained when patients were evaluated and sera were kept at 20 °C until assayed. Samples were considered positive if they were higher than 2.5% of label uptake, which was the mean 2 SD of the negative control group, with total IgE ranging from 8 to 1 300 kU/l (Blanca et al., 1992). Study of the immunochemical recognition. In order to determine the capacity of the different synthetic structures to be recognized by specific IgE antibodies, competitive inhibition immunoassays were carried out with sera from the four patients with immediate allergic reactions to cephalosporin (ceftriaxone and cefuroxime) as described (Moreno et al., 1995). This was undertaken using in the fluid J. Mol. Recognit. 2003; 16: 148–156


152

Figure 2. Proposed aminolysis pathway of cephalosporins and retrosynthetic analysis for proposed haptenic structure 1 derived after b-lactam opening by amine nucleophiles.

phase the sera from the patients incubated with 10-fold concentrations (from 150 to 15 mM) of the five compounds synthesized, 1a–f. Only two concentrations were used in order to save sera, because many determinations were required. Based on previous studies these concentrations, although high, were known not to induce non-specific binding (Moreno et al., 1995). After 18 h, disks sensitized with ceftriaxone for sera 1 and 2 and cefuroxime for sera 3 and 4, conjugated to PLL, were added. Results were expressed as percentage inhibition with respect to the noninhibited serum.

RESULTS AND DISCUSSION Synthesis Previous work by others on the conjugation of different cephalosporins with simple amino compounds (HamilstonMiller et al., 1970b) and our preliminary observations by NMR techniques employing n-butylamine as the nucleophile, point towards the formation of the cephalosporoyl conjugate at the first moment of the reaction. This compound suffers extensive fragmentation of the dihydrothiazine portion as soon as it is formed, leading to a complex mixture of compounds difficult to isolate and analyze. Nevertheless, as can be inferred by NMR data, the acyl side chain would appear to remain virtually unaffected, giving rise to a conjugate such as compound 1 (Fig. 2), with a penaldate-like structure. Obviously, the carbon atom from which the G group derives, in the final compound 1, is originally like a carbonyl group and its nature is a priori speculative, since in the physiological media it can undergo different side reactions and become oxidized to a diverse extent. This particular characteristic of 1 should certainly modulate the extent of recognition by IgE antibodies, but in order to Copyright # 2003 John Wiley & Sons, Ltd.

explore the inherent nature of this biological process, we decided initially to establish the basis of the structural requirements. Furthermore, the possibility of using different aminoacids would increase the versatility of this synthetic strategy to produce diverse molecules, which would help further understand the nature of the recognition process. Due to the difficulties encountered in the isolation of pure compounds from the complex mixtures obtained in the conjugation process, we decided to synthesize conjugates with compound 1 structure in which G is a methyl group. These would allow us to have access to essentially pure compounds with defined chemical structures and to evaluate their ability to recognize IgE antibodies to cephalosporin. A retrosynthetic analysis for the synthesis of these compounds is shown in Fig. 2. They were devised so as to be derived from the corresponding acyl side chain, plus an alpha aminoacid and butylamine, which are readily available starting materials. The synthetic route is depicted in Fig. 3. We chose Lalanine as the model aminoacid (G = CH3, Fig. 2) to evaluate the extent of the recognition process by the antibody. The synthetic scheme was essentially the same for all the compounds obtained, differing only in the nature of the protecting groups employed in each acyl side chain (4a– f) and therefore in the final deprotection step. Compound 2 was obtained in good yield by reaction of NBOC-L-alanine with n-butylamine via previous carboxyl group activation as a mixed anhydride. This compound was subsequently deprotected under acidic conditions to obtain compound 3, which is conveniently functionalized and ready for amide formation with acyl side chains 4a–f to obtain 5a–f. The last step of the synthetic scheme was carried out under the same conditions employed for the synthesis of compound 2 and took place in good to moderate yields for 1a–e. All the compounds obtained, as well as the intermediates in their syntheses, were conveniently analyzed by spectroscopic and analytical techniques, all the J. Mol. Recognit. 2003; 16: 148–156


153

Figure 3. Synthetic route for the products 1a±f. Compound 3 was obtained as the base structure, to which the libraries of acyl side chains were coupled by amide linkage.

Table 1. Results of skin tests to different cephalosporins and penicillin derivatives, including the culprit cephalosporin Penicillin derivatives Pat 1 2 3 4

Drug involved

PPL

MDM

Cephalosporin derivatives

PG

AX

AMP

Ceftriaxone Ceftriaxone Cefuroxime Cefuroxime

CL

CM

CZ

CT

CU

CX

Penicillin derivatives: PPL, benzylpenicilloyl determinant; MDM, minor determinant mixture; PG, penicillin G; AX, amoxicillin; AMP, ampicillin. Cephalosporin derivatives: CL, cephalothin; CM, cefamandole; CZ, ceftazidime; CT, ceftriaxone; CU, cefuroxime; CX, cefotaxime.

Table 2. Results of skin tests and speci®c IgE antibody RAST to different cephalosporins and penicillins, including the culprit cephalosporin Penicillin Pat 1 2 3 4

Cephalosporin

Drug involved

PG

AX

AMP

CDX

CLX

CZ

CT

CU

CX

Ceftriaxone Ceftriaxone Cefuroxime Cefuroxime

0.1 0.71 0.06 0.2

0 0.04 0.11 0

0 0.11 0.13 0.4

0.18 0.02 0.12 0.03

0.41 0.19 0.31 0.09

0.14 0.31 0.01 0.67

4.53 17.7 0.16 1.99

0.09 1.86 9.43 20.8

0.46 40.7 0.89 13.9

Penicillins: PG, penicillin G; AX, amoxicillin; AMP, ampicillin. Cephalosporins: CDX, cefadroxil; CLX, cephalexine; CZ, ceftazidime; CT, ceftriaxone; CU, cefuroxime; CX, cefotaxime. Copyright # 2003 John Wiley & Sons, Ltd.

J. Mol. Recognit. 2003; 16: 148–156


154

Figure 4. RAST inhibition studies with sera from four patients allergic to cephalosporins. (A) RAST inhibition with sera from the patients allergic to ceftriaxone at two concentrations (150 and 15 mM) of different compounds. In both sera (serum 1 and 2) the maximum inhibition was seen with 1a, the synthesized compound having the same acyl side chain as ceftriaxone. (B) RAST inhibition with sera from the patients allergic to cefuroxime at two concentrations (150 and 15 mM) of different compounds. For these patients, compounds with chemically similar acyl side chains show comparable inhibition levels, with the compound having the same side chain as cefuroxime producing the highest inhibition (1f).

data being in agreement with the proposed structures. The synthesis of 1f (syn configuration) required, in the final step, a dehydrating condensation between the keto-carbonyl group of the a-furanacetic moiety in 4f and methoxyamine. This yielded the two syn/anti isomers in the carbon nitrogen double bond and further exhaustive chromatographic separations were necessary to separate the two isomers, Copyright # 2003 John Wiley & Sons, Ltd.

the relative syn/anti configuration of which was determined by NOE experiments. Immunochemical studies The recognition of the different structures synthesized was J. Mol. Recognit. 2003; 16: 148–156


made by competitive inhibition studies of RAST immunoassay using human sera from the four patients with allergic reactions to cephalosporins. The patients were diagnosed by skin test (Table 1) and RAST studies (Table 2), which demonstrated IgE antibodies that specifically recognized the cephalosporin involved in the immediate reaction and, in some cases, cross-reactivity to other cephalosporins and/or penicillins. The RAST inhibition studies provided more precise information concerning the specific recognition by the IgE antibodies of the synthetic compounds, because their structure and concentration are well characterized. This method was undertaken with serum from each of the four patients included. The results with serum 1 [Fig. 4(A)], specific to ceftriaxone, showed that the optimal recognition was made with the synthetic structure 1a, which included the same side chain structure as ceftriaxone, and also to the other structures in decreasing order 1e, 1b, 1d, with no recognition of 1c. Serum 2 [Fig. 4(A)] showed a strong binding with structure 1a, which had the same structure as the side chain of several cephalosporins, i.e. ceftriaxone, which induced the reaction, cefotaxime and ceftazidime. These results were confirmed by skin test and, to a certain extent, RAST. The inhibition with serum 3 [Fig. 4(B)], specific to cefuroxime, showed that optimal recognition was obtained with the structure containing its corresponding side chain 1f, followed by 1a and 1b. The results obtained with serum 4 [Fig. 4(B)] showed a specific recognition of the structure 1f, containing the side chain of cefuroxime, which was the culprit cephalosporin in the reaction. There was also a specific binding with the structure 1a, which had a chemical similarity to 1f. The IgE antibodies present in this serum showed no specific binding with the other structures assayed. These results indicated that differences existed in the behavior of specific IgE antibodies of serum from the four patients studied. In sera 2, 3 and 4 the recognition was more specific to the acyl side chain structure, whereas in serum 1, besides the side chain, the butyl-alanyl moiety also seemed to contribute to the recognition process. These results reveal that structures 1a–f, representing a fragment of the proposed cephalosporoyl intermediate, are suitably recognized by IgE antibodies directed to cephalosporins, and that specificities are related to the acyl side chain as well as to the b-lactam fragment, which remains linked to the carrier protein in the conjugation process. In contrast to the great number of studies related to the chemistry of the synthesis and reactivity of penicillins, in which the chemical structures are unmistakably established, the chemistry of cephalosporins lacks this conclusive accuracy, mainly in reactivity, despite which the same chemical repercussions have been assumed for cephalosporins. This supposition has been reflected in the immunological and clinical postulates, the negative consequences of which may well be very important. Despite the lack of certainty about the chemical structure of the antigenic determinant responsible for allergic reactions to cephalosporins, several studies (Pham and Baldo, 1996; Baldo, 1999; Harle and Baldo, 1990) have reported immunological results, although the conclusions have to be taken with a certain degree of speculation. Because the haptenic determinant of cephalosporins produced by their degradation is largely unknown, more definite studies Copyright # 2003 John Wiley & Sons, Ltd.

155

require each free drug to be used as the skin test agent to detect antibodies reactive to these antibiotics (Baldo, 1999; Weiss and Adkinson, 1988). Besides the requirement for specific immunoassays to detect IgE antibodies to individual cephalosporins, and the need to define the allergenic determinants in these drugs, other potentially helpful data on a number of fundamental and applied aspects of cephalosporin immunochemistry are absent. These include allergic cross-reactivity with other b-lactams, the contribution of side-chain structures to allergenicity, and the extent of heterogeneity of cephalosporin allergenic determinants. Consequently, many questions on cross-reactivities between cephalosporins, and between cephalosporins and penicillins, cannot currently be answered with confidence. This raises difficulties in the selection of antibiotics for some penicillinand/or cephalosporin-allergic subjects. The antigen binding activity is found in the variable region of antibodies, but there are structural constraints conditioning its ability to bind specifically to one or more closely related small molecules. Thus, synthesis of 1a–f has provided, for the first time, structural information concerning the chemical implications in the recognition requirements, and has allowed us to evaluate the extent of the heterogeneity indicated by the evident and fine differences in recognition of different sera. We propose this synthetic methodology as the right course for the systematic study of the chemical implications in these recognition processes in order to define the allergenic determinants of cephalosporins.

CONCLUSION A series of six aN-acyl-L-alanylbutanamides was prepared in four steps from N-BOC-L-alanine and the appropriate acids. These chemically well-defined structures comprised the entire acyl side chain and the aminoacidic residue included in the b-lactam moiety of the cephalosporins studied, and were linked as amide functions to an aliphatic (n-butyric) chain. They were used to study the molecular basis of the recognition of different cephalosporins by IgE antibodies from subjects allergic to ceftriaxone and cefuroxime. Based on the results obtained in RAST inhibition studies, which were well correlated with both skin test and RAST results, fine structural recognition was detected between IgE antibodies in different sera and compounds 1a–f. Thus, these synthesized compounds incorporate the appropriate epitope recognizable by IgE antibodies and provide the basis to understand the pathway of cephalosporin conjugation to carrier proteins, to determine the structural requirements needed to be recognized by IgE antibodies and, finally, to obtain an efficient in vitro test to detect these antibodies. Chemical structure modifications are now under way in our laboratories to enhance the specificity and sensitivity of the recognition process.

Acknowledgements F.S.-S. is grateful for a fellowship linked to this project. The authors are also grateful to Professor D. Dominguez for recording the NOE experiments. The authors thank Ian Johnstone for help with the English language.

J. Mol. Recognit. 2003; 16: 148–156

156


REFERENCES Baldo BA. 1999. Penicillins and cephalosporins as allergensÐ structural aspects of recognition and cross-reaction. Clin. Exp. Allergy 29: 744±749. Batchelor FR, Dewdney JM, Gazzard D. 1965. Penicillin allergy: the formation of the penicilloyl determinant. Nature 206: 362± 364. Batchelor FR, Dewdney JM, Weston RD, Wheeler AW. 1966. The immunogenicity of cephalosporins derivatives and their cross-reaction with penicillin. Immunology 10: 21±33. Blanca M. 1995. Allergic reactions to penicillins. A changing world? Allergy 50: 777±782. Blanca M, Mayorga C, Perez E, Suau R, Juarez C, Vega JM, Carmona MJ, Perez-Estrada M, Garcia J. 1992. Determination of IgE antibodies to the benzyl penicilloyl determinant. A comparison between poly-L-lysine and human serum albumin as carriers. J. Immunol. Meth. 153: 99±105. Blanca M, Vega JM, Garcia J, Miranda A, Carmona MJ, JuaÂrez C. 1994. New aspects of allergic reactions to beta-lactamsÐ crossreactions and unique speci®cities. Clin. Exp. Allergy 24: 407±415. Boger DL, Borzilleri RM, Nukui S, Beresis RT. 1997. Synthesis of the vancomycin CD and DE ring systems. J. Org. Chem. 62: 4721±. Bucourt R, Heymes R, Lutz A, PeÂnasse L, Perronnet J. 1978. Cephalosporines a chaines amino-2 thiazolyl-4 acetyles. In¯uence de la presence et de la con®guration d'un grupe axyimino sur l'activite antibacterienne. Tetrahedron 34: 2233±2243. Dewdney JM. 1977. Antigens. In Immunology of the Antibiotics, Sela M (ed). Academic Press: New York; 73±245. Faraci WS, Pratt RF. 1984. Elimination of a good leaving group from the 3'-position of a cephalosporin need not be concerted with b-lactam ring opening: TEM-2 b-lactamase-catalyzed hydrolysis of pyridine-2-azo-4'-(N',N'-dimethylaniline) cephalosporin (PADAC) and of cephaloridine. J. Am. Chem. Soc. 106: 1489±1490. Grabowski EJJ, Douglas AW, Smith GB. 1985. Ammonolysis of cephalosporins: 13C NMR characterization of the intermediates from b-lactam ring cleavage prior to loss of the 3'-group. J. Am. Chem. Soc. 107: 267±268. Hamilton-Miller JMT, Newton GGF, Abraham EP. 1970a. Products of aminolysis and enzymic hydrolysis of the cephalosporins. Biochem. J. 116: 371±384. Hamilton-Miller JMT, Richards E, Abraham EP. 1970b. Changes in proton magnetic resonance spectra during aminolysis and enzymic hydrolysis of cephalosporins. Biochem. J. 116: 385. Harle DG, Baldo BA. 1990. Drugs as allergens: an immunoassay for detecting IgE antibodies to cephalosporins. Int. Arch. Allergy Appl. Immunol. 92: 439±444. Holden KG. 1984. Cephalosporins. In Comprehensive Heterocyclic Chemistry, Vol. 7, Katritzky AR (ed). Pergamon Press: Oxford; 285±339. Kaiser GV, Kukolja S. 1972. Modi®cation of the b-lactam ring. In Cephalosporins and Penicillins. Chemistry and Biology,

Copyright # 2003 John Wiley & Sons, Ltd.

Flynn EH (ed). Academic Press: New York; 74±133. Kuisle O, QuinÄoaÂ E, Riguera R. 1999. A general methodology for automated solid-phase synthesis of depsides and depsipeptidesÐpreparation of a valinomycin analog. J. Org. Chem. 64: 8063±8075. Levine BB, Ovary Z. 1961. Studies on the mechanism of the formation of the penicillin antigen III. The N-(D-a-benzylpenicilloyl) group as an antigenic determinant responsible for hypersensitivity to penicillin G. J. Exp. Med. 114: 875±904. LlinaÂs A, Vilanova B, Frau J, MunÄoz F, Donoso J, Page MI. 1998. Chemical-reactivity of penicillins and cephalosporinsÐintramolecular involvement of the acyl-amido side-chain. J. Org. Chem. 63: 9052±9060. Lowe G. 1979. b-Lactam antibiotics. In Comprehensive Organic Chemistry, Vol. 5, Barton D, Ollis D (eds). Pergamon Press: Oxford; 289±320. Manhas MS, Bose AK. 1971. b-Lactams: Natural and Synthetic. Part 1, Bose AK (ed). Wiley-Interscience: New York. Moreno F, Blanca M, Mayorga C, Terrados S, Moya MC, PeÂrez E, Suau R, Vega JM, GarcõÂa J, Miranda A, Carmona MJ. 1995. Studies of the speci®cities of IgE antibodies found in sera from subjects with allergic reactions to penicillins. Int. Arch. Allergy Immunol. 108: 74±81. Nagakura N, Shimizu T, Masuzawa T, Yanagihara Y. 1990. Anticephalexin monoclonal antibodies and their cross-reactivities to cephems and penams. Int. Arch. Allergy Appl. Immunol. 93: 126±132. Paruszewski R, Rostai®nska-Suchar G, Strupinska M, Jaworski P, Stables JP. 1996. Synthesis and anticonvulsant activity of some amino acid derivatives. Part 1: alanine derivatives. Pharmazie 51: 145±148. Pham NH, Baldo BA. 1996. b-Lactam drug allergens: ®ne structural recognition patterns of cephalosporin-reactive IgE antibodies. J. Mol. Recognit. 9: 287±296. Pratt RF, Faraci WS. 1986. Direct observation by 1H NMR of cephalosporoate intermediates in aqueous solution during the hydrazinolysis and b-lactamase-catalyzed hydrolysis of cephalosporins with 3' leaving groups: kinetics and equilibria of the 3' elimination reaction. J. Am. Chem. Soc. 108: 5328± 5333. Romano A, Mayorga C, Torres MJ, Artesani MC, Suau R, Sanchez F, Perez E, Venuti A, Blanca M. 2000. Immediate allergic reactions to cephalosporins: cross-reactivity and selective response. J. Allergy Clin. Immunol. 106: 1177±1183. Vilanova B, MunÄoz F, Donoso J, Blanco FG. 1993. HPLC and 1H NMR studies of alkaline-hydrolysis of some 7-(oxyiminoacyl)cephalosporins. Helv. Chim. Acta 76: 2789±2802. Wagatsuma M, Terashima S, Yamada SI. 1973. Amino acids and peptides. VI. Novel peptide bond formation catalyzed by metal ions. IV. Formation of optically active amino acid amides and peptide amides. Chem. Pharm. Bull. 21: 422±427. Weiss M, Adkinson NF. 1988. Immediate hypersensitivity reactions to penicillin and related antibiotics. Clin. Allergy 18: 515±540.

J. Mol. Recognit. 2003; 16: 148–156