- Email: [email protected]
Identification of a probable new adrenergic agonist by nuclear magnetic resonance and mass spectrometry
Analytica Chimica Acta 586 (2007) 223–227
Identification of a probable new adrenergic agonist by nuclear magnetic resonance and mass spectrometry Gianpiero Boatto a , Nicola Culeddu b , Cecilia Testa c , Bruno Neri d , Gianfranco Brambilla e,∗ , Jorge Barbosa f , Clara Cruz f a
Department of Toxicological Chemistry, University of Sassari, Sassari, Italy b CNR Biomolecular Chemistry Institute, Sassari, Italy c IZS della Sardegna, Sassari, Italy d IZS delle Regioni Lazio e Toscana, Rome, Italy e Istituto Superiore di Sanit` a, Environment Department, Toxicological Chemistry Unit, Rome, Italy f LNIV, Lisbon, Portugal Received 19 July 2006; accepted 20 September 2006 Available online 29 September 2006
Abstract In animal production, it is consolidated the synthesis and the illegal use of growth promoters of new generation, able to skip routine screening and confirmatory analysis. In this work it is reported the nuclear magnetic resonance (NMR) and the mass spectrometry identification of a probable new adrenergic drug found in a feed premix. The substance was selectively purified on alpha 1 acid glycoprotein affinity columns; then its structure was first achieved by recording the 13 C NMR spectrum that gave the total number of carbons of the molecule, successively sorted by DEPT experiments into quaternary, CH, CH2 , and CH3 groups. However, the complete assignments of all resonances were derived from the bi-dimensional analysis and the crucial indications from the 1 H–13 C reverse experiments. Further characterisation was performed by atmospheric pressure chemical ionisation both in positive and negative ion mode, matching the molecular ion and the fragmentation pattern with those of most recently described new adrenergic agonists. After the loss of a ter-butylic group, the structure shows an internal symmetry along with the presence of Chlorine clusters. The proposed formula of the compound, the 8,8 -diamino-9,9 -dichloro-1-terbutyl-1,1 ,4,4-tetrahydro-5H,5 H-2,2 -bi-1-benzazepine-5,5 -dione, partially resembles that of Zilpaterol for the presence of a heterocyclic ring; Further work is in progress to characterise the structure–activity relationship. © 2006 Elsevier B.V. All rights reserved. Keywords: Adrenergic agonist; Nuclear magnetic resonance; Mass spectrometry
1. Introduction Since 1988, when the use of beta adrenergic agonist drugs has been proposed in animal nutrition to improve quality and quantity of meat products [1], an almost constant turn-over in the chemical structures of the drugs administered to feedlots has been observed. Such pharmacological class has been regularly inserted among the analytical duties framed within the National monitoring plans on veterinary drugs residues analysis since 1990. At that time the attention was focused mainly on already known compounds, such as Clenbuterol, Mabuterol,
∗
Corresponding author. Tel.: +39 06 4990 2764; fax: +39 06 4938 7139. E-mail address: [email protected] (G. Brambilla).
0003-2670/$ – see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.aca.2006.09.045
Salbutamol, Terbutaline, already licensed as veterinary and/or human drugs. Along with the implementation of the monitoring strategies, new compounds of possible selective illegal use in animal production have been progressively discovered [2–4], probably with the intent to skip routine screening and confirmatory analysis. Recently, the registration of two adrenergic agonist zootechnical additives, such as Ractopamine and Zilpaterol has been finalised in the USA with the set up of the appropriate Maximum Residue Limits, while their use is still forbidden within the European Union [5]: at the same time molecules of new generation have been developed for human asthma therapy [6] (Fig. 1). Owing to the above, to improve the efficacy of monitoring plans, thus reducing possible health hazards for Consumers, it is mandatory to support the overall analytical strategy with the use of sophisticated techniques, able to identify possible new
224
G. Boatto et al. / Analytica Chimica Acta 586 (2007) 223–227
Fig. 1. Structures of some among most recent beta adrenergic agonist drugs of possible use in animal production.
analytical targets, such as nuclear magnetic resonance (NMR), already used for a full characterisation of anabolic preparations [7]. In this work, it is reported the NMR and LC-MS characterisation of a new adrenergic drug found in a feed premix, after its purification on alpha 1 acid protein affinity chromatography column [8], a chromatographic step selective for beta adrenergic drugs. 2. Experimental The feed premix was found during official monitoring plan within the European Union, in form of a white powder. After its solubilisation in Posphate Buffer Saline (PBS) pH 7.4 (10 mg mL−1 ), aliquots of 5 mL of the extract were serially
applicated on the top of a 50 mg (10 mL bedding) alpha 1 acid affinity column, prepared and conditioned as previously described [8]. Briefly, after a 10 mL washing with PBS, the fraction of interest was eluted by the mean of 10 mL MeOH (1% acetic acid) and brought to dryness under a gentle nitrogen stream. NMR experiments: the NMR spectra were recorded on Bruker Avance 600 and Varian VX 200 spectrometers, operating at 600.13 and 200.057 Hz for 1H and at 150.91 and 50.309 MHz for 13C, using TMS as external standard. About 20 mg of sample was dissolved in 0.6 mL of deuterated solvent. 1 H, 13 C{1 H}, 13 C DEPT-90 and 1 H TOCSY (MLEVPH pulse sequence spinlock time 80 ms), 1 H NOESY phase sensitive, 1 H–13 C HSQC, and 1 H–13 C HMBC were performed. All of the experiments were
G. Boatto et al. / Analytica Chimica Acta 586 (2007) 223–227
performed using standard Brucker and Varian programs, and most of the acquisition parameters were used as suggested by the programs. The temperature of the probes was 300 K. The 1 H NOESY phase sensitive spectrum was obtained with a mixing time of 800 ms. The structure was first achieved by recording the 13 C NMR spectrum that gives the total number of carbons of the molecule, successively sorted by DEPT experiments into quaternary, CH, CH2 , and CH3 groups. However, the complete assignments of all resonances were derived from the bidimensional experiments and the crucial indications from the 1 H–13 C HSQC, and 1 H–13 C HMBC. The LC-MS/MS system consisted on a PE Sciex (Foster City, CA) API 2000 tandem mass spectrometer equipped with a heated nebulizer (HN) interface. A HPLC Perkin-Elmer (Norwalk, CT, USA) micro pumps Series 200 and PE 200 autoinjector were used for all LC-MS analyses. The data were processed using MacQuan software (PE Sciex) on a MacIntosch Quadra 900 microcomputer. The HN probe was maintained at +500 ◦ C and gas phase chemical ionisation was effected by a corona discharge needle (+4 A) using positive ion atmospheric pressure chemical ionisation (APCI). The nebulizing gas (N2 ) pressure was set for the HN interface at 80 psi. The auxiliary flow was 2 L min−1 , the curtain gas flow (N2 ) was 0.9 L min−1 , and the sampling orifice was set at +50 V. The dwell time was 400 ms, and the temperature of the interface heater was set at +60 ◦ C. Infusion rate was 10 L min−1 of the purified compound redissolved in MeOH–NH4 HCO2 3 mM (50:50, v/v) solution. Acquisition was carried out in positive and negative ions mode, in the scan range 100–500 m/z. 3. Results and discussion The determination of the exact chemical structure of this compound was achieved by using a series of 1D, 2D NMR experiments. Signal assignments were extracted from 1D spectra and 2D H,H (Fig. 2) and C,H correlation spectra (Table 1). Signals were also assigned by long-range C,H correlation spectra (gradient-selected HSQC–TOCSY) (Fig. 3). In the HSQC–TOCSY spectrum, cross peaks of the NH1 –C4 , the NH1 –C3 and the C3 –C4 were identified. Moreover, there was cross peaks of the C4–C12 These results indicate the presence of two distinct framework. Moreover, the cross peaks from NH1 –C4 in HSQC–TOCSY (also shown in the complete TOCSY spectrum) also confirmed the assignments of signals within each amino-aromatic residues. The compound is a representative model of two canonical forms and they can be distinguish via NOESY experiments. In the NOESY spectrum, NOE correlation peaks of weak intensity were observed between of NH1 –C13 and H3–H3 are present, these results are consistent with assumption that the ter-butyul moiety is close to NH1 than CH3 . LC-MS positive ion spectrum is reported in Fig. 4; the m/z of [M + H]+ ion was found to be 471 according to the data of NMR that assigned a formula bruta C24 H24 N4 O2 Cl2 . The protonated molecular ion shows the two chlorine isotropic cluster. The m/z 415 ion can be addressed to the to the loss of the ter-butylic group (−56 amu) from [M + H]+ ion, as already described for
225
Fig. 2. TOCSY spectrum of the compound.
Clenbuterol [9]. The following fragments m/z 236 and m/z 208 can be reasonably addressed to the monomerisation of the compound after the loss of the ter-butyl group, and its subsequent re-arrangement, respectively. Both 208 m/z and 236 m/z ions share the typical chlorine isotopic cluster (Fig. 5). LC-MS negative ion spectrum (Fig. 6) indicates the presence of the [M − H]− Table 1 1 H and 13 C chemical shifts (ppm) of compound: the atoms numbering used are those commonly used for benzodiazepine (see Fig. 2) No.
d 1H
d 13 C
2 3 4 5 6 7 8 9 10 11 12 13 14 15 1 2 3 4 5 6 7 8 9 10 11
– 4.86 3.02 – – – – – 7.41 7.09 – 1.2 1.2 1.2 8.58 – 4.86 2.90 – – – – – 7.41 7.09
137.9 70.10 56.12 195.50 106.8 137.4 138.0 147.2 122.0 128.0 53.96 24.9 24.9 24.9 – 128.9 67.05 48.7 196.5 117.2 147.3 127.0 129.0 125.0 121.1
226
G. Boatto et al. / Analytica Chimica Acta 586 (2007) 223–227
Fig. 3. HSQC–TOCSY spectrum of the compound.
Fig. 4. Heated nebulizer LC-MS positive ion spectrum of the unknown -agonist.
G. Boatto et al. / Analytica Chimica Acta 586 (2007) 223–227
Fig. 5. Proposed chemical structure of the unknown compound, along with its fragmentation pattern: the atoms numbering is that commonly used for benzodiazepine.
227
the results of both C and H NMR experiments with the evidences of MS (APCI) analysis. The proposed formula, compared with those of classical ethanolamine-derived adrenergic drugs, highlights cyclisation and the presence of a heterocyclic ring in the structure, features shared with the more recent repartioning agents such as the USA licensed feed additive Zilpaterol, the illegal compound described by Nielen et al. [4] and some of the antiastmathic drugs of new generation [7]. This consideration constitutes a prompt to approach the matter in an interdisciplinary way, through a cross-checking of the biological, pharmacological and molecular evidences, tacking into account the molecular diversity of the beta adrenergic receptors population in the different animal target species. Acknowledgements Work supported by Italian Ministry of Health Grant no. IZSSA 005 2000 and accepted as poster presentation at the Fifth International Symposium on Hormone and Veterinary Drug Residue Analysis, Antwerp, Belgium, 16–19 May 2006. References
Fig. 6. Heated nebulizer LC-MS negative ion spectrum of the unknown agonist.
ion (m/z 469), that supports the previous molecular evidences sourcing from LC-MS APCI analysis and NMR experiments, respectively. Again, the deprotonated molecular ion shows the two chlorine isotropic cluster. 4. Conclusions In this work it has been possible to elucidate the structure of a new compound illegally used in pig production, by matching
[1] H. Timmerman, in: J.P. Hanrahan (Ed.), -Agonists and Their Effects on Animal Growth and Carcass Quality, Elsevier, New York, 1987, p. 13. [2] G. Boatto, C. Testa, G. Brambilla, B. Neri, M. Fiori, D. Cuccaro, G. Manca, in: L.A. van Ginkel, A. Ruiter (Eds.), Proceeding of Euroresidue IV Conference Veldhoven, NL, 2000, p. 237. [3] G. Mazzanti, C. Daniele, G. Boatto, G. Manca, G. Brambilla, A. Loizzo, Toxicology 187 (2003) 91. [4] M.W.F. Nielen, C.T. Elliott, S.A. Boyd, D. Courtheyn, M.L. Essers, H.H. Hooijerink, E.O. van Bennekom, R.E. Fuchs, Rapid Commun. Mass Spectrom. 17 (2003) 1633. [5] J. Blanca, P. Mu˜noz, M. Morgado, N. M´endez, A. Aranda, T. Reuvers, H. Hooghuis, Anal. Chim. Acta 529 (2005) 2199. [6] B. Waldeck, Eur. J. Pharm. 445 (2002) 1. [7] A. Lommen, R. Schilt, J. Weseman, A.H. Roos, J.W. van Velde, M.W.F. Nielen, J. Pharm. Biomed. Anal. 28 (2002) 87. [8] P. Gallo, G. Brambilla, M. Fiori, A. Scaramuzzo, L. Serpe, Chromatographia 53 (2001) 446. [9] F.W. Smyth, J. Chromatogr. B 824 (2005) 1.
Identification of a probable new adrenergic agonist by nuclear magnetic resonance and mass spectrometry Gianpiero Boatto a , Nicola Culeddu b , Cecilia Testa c , Bruno Neri d , Gianfranco Brambilla e,∗ , Jorge Barbosa f , Clara Cruz f a
Department of Toxicological Chemistry, University of Sassari, Sassari, Italy b CNR Biomolecular Chemistry Institute, Sassari, Italy c IZS della Sardegna, Sassari, Italy d IZS delle Regioni Lazio e Toscana, Rome, Italy e Istituto Superiore di Sanit` a, Environment Department, Toxicological Chemistry Unit, Rome, Italy f LNIV, Lisbon, Portugal Received 19 July 2006; accepted 20 September 2006 Available online 29 September 2006
Abstract In animal production, it is consolidated the synthesis and the illegal use of growth promoters of new generation, able to skip routine screening and confirmatory analysis. In this work it is reported the nuclear magnetic resonance (NMR) and the mass spectrometry identification of a probable new adrenergic drug found in a feed premix. The substance was selectively purified on alpha 1 acid glycoprotein affinity columns; then its structure was first achieved by recording the 13 C NMR spectrum that gave the total number of carbons of the molecule, successively sorted by DEPT experiments into quaternary, CH, CH2 , and CH3 groups. However, the complete assignments of all resonances were derived from the bi-dimensional analysis and the crucial indications from the 1 H–13 C reverse experiments. Further characterisation was performed by atmospheric pressure chemical ionisation both in positive and negative ion mode, matching the molecular ion and the fragmentation pattern with those of most recently described new adrenergic agonists. After the loss of a ter-butylic group, the structure shows an internal symmetry along with the presence of Chlorine clusters. The proposed formula of the compound, the 8,8 -diamino-9,9 -dichloro-1-terbutyl-1,1 ,4,4-tetrahydro-5H,5 H-2,2 -bi-1-benzazepine-5,5 -dione, partially resembles that of Zilpaterol for the presence of a heterocyclic ring; Further work is in progress to characterise the structure–activity relationship. © 2006 Elsevier B.V. All rights reserved. Keywords: Adrenergic agonist; Nuclear magnetic resonance; Mass spectrometry
1. Introduction Since 1988, when the use of beta adrenergic agonist drugs has been proposed in animal nutrition to improve quality and quantity of meat products [1], an almost constant turn-over in the chemical structures of the drugs administered to feedlots has been observed. Such pharmacological class has been regularly inserted among the analytical duties framed within the National monitoring plans on veterinary drugs residues analysis since 1990. At that time the attention was focused mainly on already known compounds, such as Clenbuterol, Mabuterol,
∗
Corresponding author. Tel.: +39 06 4990 2764; fax: +39 06 4938 7139. E-mail address: [email protected] (G. Brambilla).
0003-2670/$ – see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.aca.2006.09.045
Salbutamol, Terbutaline, already licensed as veterinary and/or human drugs. Along with the implementation of the monitoring strategies, new compounds of possible selective illegal use in animal production have been progressively discovered [2–4], probably with the intent to skip routine screening and confirmatory analysis. Recently, the registration of two adrenergic agonist zootechnical additives, such as Ractopamine and Zilpaterol has been finalised in the USA with the set up of the appropriate Maximum Residue Limits, while their use is still forbidden within the European Union [5]: at the same time molecules of new generation have been developed for human asthma therapy [6] (Fig. 1). Owing to the above, to improve the efficacy of monitoring plans, thus reducing possible health hazards for Consumers, it is mandatory to support the overall analytical strategy with the use of sophisticated techniques, able to identify possible new
224
G. Boatto et al. / Analytica Chimica Acta 586 (2007) 223–227
Fig. 1. Structures of some among most recent beta adrenergic agonist drugs of possible use in animal production.
analytical targets, such as nuclear magnetic resonance (NMR), already used for a full characterisation of anabolic preparations [7]. In this work, it is reported the NMR and LC-MS characterisation of a new adrenergic drug found in a feed premix, after its purification on alpha 1 acid protein affinity chromatography column [8], a chromatographic step selective for beta adrenergic drugs. 2. Experimental The feed premix was found during official monitoring plan within the European Union, in form of a white powder. After its solubilisation in Posphate Buffer Saline (PBS) pH 7.4 (10 mg mL−1 ), aliquots of 5 mL of the extract were serially
applicated on the top of a 50 mg (10 mL bedding) alpha 1 acid affinity column, prepared and conditioned as previously described [8]. Briefly, after a 10 mL washing with PBS, the fraction of interest was eluted by the mean of 10 mL MeOH (1% acetic acid) and brought to dryness under a gentle nitrogen stream. NMR experiments: the NMR spectra were recorded on Bruker Avance 600 and Varian VX 200 spectrometers, operating at 600.13 and 200.057 Hz for 1H and at 150.91 and 50.309 MHz for 13C, using TMS as external standard. About 20 mg of sample was dissolved in 0.6 mL of deuterated solvent. 1 H, 13 C{1 H}, 13 C DEPT-90 and 1 H TOCSY (MLEVPH pulse sequence spinlock time 80 ms), 1 H NOESY phase sensitive, 1 H–13 C HSQC, and 1 H–13 C HMBC were performed. All of the experiments were
G. Boatto et al. / Analytica Chimica Acta 586 (2007) 223–227
performed using standard Brucker and Varian programs, and most of the acquisition parameters were used as suggested by the programs. The temperature of the probes was 300 K. The 1 H NOESY phase sensitive spectrum was obtained with a mixing time of 800 ms. The structure was first achieved by recording the 13 C NMR spectrum that gives the total number of carbons of the molecule, successively sorted by DEPT experiments into quaternary, CH, CH2 , and CH3 groups. However, the complete assignments of all resonances were derived from the bidimensional experiments and the crucial indications from the 1 H–13 C HSQC, and 1 H–13 C HMBC. The LC-MS/MS system consisted on a PE Sciex (Foster City, CA) API 2000 tandem mass spectrometer equipped with a heated nebulizer (HN) interface. A HPLC Perkin-Elmer (Norwalk, CT, USA) micro pumps Series 200 and PE 200 autoinjector were used for all LC-MS analyses. The data were processed using MacQuan software (PE Sciex) on a MacIntosch Quadra 900 microcomputer. The HN probe was maintained at +500 ◦ C and gas phase chemical ionisation was effected by a corona discharge needle (+4 A) using positive ion atmospheric pressure chemical ionisation (APCI). The nebulizing gas (N2 ) pressure was set for the HN interface at 80 psi. The auxiliary flow was 2 L min−1 , the curtain gas flow (N2 ) was 0.9 L min−1 , and the sampling orifice was set at +50 V. The dwell time was 400 ms, and the temperature of the interface heater was set at +60 ◦ C. Infusion rate was 10 L min−1 of the purified compound redissolved in MeOH–NH4 HCO2 3 mM (50:50, v/v) solution. Acquisition was carried out in positive and negative ions mode, in the scan range 100–500 m/z. 3. Results and discussion The determination of the exact chemical structure of this compound was achieved by using a series of 1D, 2D NMR experiments. Signal assignments were extracted from 1D spectra and 2D H,H (Fig. 2) and C,H correlation spectra (Table 1). Signals were also assigned by long-range C,H correlation spectra (gradient-selected HSQC–TOCSY) (Fig. 3). In the HSQC–TOCSY spectrum, cross peaks of the NH1 –C4 , the NH1 –C3 and the C3 –C4 were identified. Moreover, there was cross peaks of the C4–C12 These results indicate the presence of two distinct framework. Moreover, the cross peaks from NH1 –C4 in HSQC–TOCSY (also shown in the complete TOCSY spectrum) also confirmed the assignments of signals within each amino-aromatic residues. The compound is a representative model of two canonical forms and they can be distinguish via NOESY experiments. In the NOESY spectrum, NOE correlation peaks of weak intensity were observed between of NH1 –C13 and H3–H3 are present, these results are consistent with assumption that the ter-butyul moiety is close to NH1 than CH3 . LC-MS positive ion spectrum is reported in Fig. 4; the m/z of [M + H]+ ion was found to be 471 according to the data of NMR that assigned a formula bruta C24 H24 N4 O2 Cl2 . The protonated molecular ion shows the two chlorine isotropic cluster. The m/z 415 ion can be addressed to the to the loss of the ter-butylic group (−56 amu) from [M + H]+ ion, as already described for
225
Fig. 2. TOCSY spectrum of the compound.
Clenbuterol [9]. The following fragments m/z 236 and m/z 208 can be reasonably addressed to the monomerisation of the compound after the loss of the ter-butyl group, and its subsequent re-arrangement, respectively. Both 208 m/z and 236 m/z ions share the typical chlorine isotopic cluster (Fig. 5). LC-MS negative ion spectrum (Fig. 6) indicates the presence of the [M − H]− Table 1 1 H and 13 C chemical shifts (ppm) of compound: the atoms numbering used are those commonly used for benzodiazepine (see Fig. 2) No.
d 1H
d 13 C
2 3 4 5 6 7 8 9 10 11 12 13 14 15 1 2 3 4 5 6 7 8 9 10 11
– 4.86 3.02 – – – – – 7.41 7.09 – 1.2 1.2 1.2 8.58 – 4.86 2.90 – – – – – 7.41 7.09
137.9 70.10 56.12 195.50 106.8 137.4 138.0 147.2 122.0 128.0 53.96 24.9 24.9 24.9 – 128.9 67.05 48.7 196.5 117.2 147.3 127.0 129.0 125.0 121.1
226
G. Boatto et al. / Analytica Chimica Acta 586 (2007) 223–227
Fig. 3. HSQC–TOCSY spectrum of the compound.
Fig. 4. Heated nebulizer LC-MS positive ion spectrum of the unknown -agonist.
G. Boatto et al. / Analytica Chimica Acta 586 (2007) 223–227
Fig. 5. Proposed chemical structure of the unknown compound, along with its fragmentation pattern: the atoms numbering is that commonly used for benzodiazepine.
227
the results of both C and H NMR experiments with the evidences of MS (APCI) analysis. The proposed formula, compared with those of classical ethanolamine-derived adrenergic drugs, highlights cyclisation and the presence of a heterocyclic ring in the structure, features shared with the more recent repartioning agents such as the USA licensed feed additive Zilpaterol, the illegal compound described by Nielen et al. [4] and some of the antiastmathic drugs of new generation [7]. This consideration constitutes a prompt to approach the matter in an interdisciplinary way, through a cross-checking of the biological, pharmacological and molecular evidences, tacking into account the molecular diversity of the beta adrenergic receptors population in the different animal target species. Acknowledgements Work supported by Italian Ministry of Health Grant no. IZSSA 005 2000 and accepted as poster presentation at the Fifth International Symposium on Hormone and Veterinary Drug Residue Analysis, Antwerp, Belgium, 16–19 May 2006. References
Fig. 6. Heated nebulizer LC-MS negative ion spectrum of the unknown agonist.
ion (m/z 469), that supports the previous molecular evidences sourcing from LC-MS APCI analysis and NMR experiments, respectively. Again, the deprotonated molecular ion shows the two chlorine isotropic cluster. 4. Conclusions In this work it has been possible to elucidate the structure of a new compound illegally used in pig production, by matching
[1] H. Timmerman, in: J.P. Hanrahan (Ed.), -Agonists and Their Effects on Animal Growth and Carcass Quality, Elsevier, New York, 1987, p. 13. [2] G. Boatto, C. Testa, G. Brambilla, B. Neri, M. Fiori, D. Cuccaro, G. Manca, in: L.A. van Ginkel, A. Ruiter (Eds.), Proceeding of Euroresidue IV Conference Veldhoven, NL, 2000, p. 237. [3] G. Mazzanti, C. Daniele, G. Boatto, G. Manca, G. Brambilla, A. Loizzo, Toxicology 187 (2003) 91. [4] M.W.F. Nielen, C.T. Elliott, S.A. Boyd, D. Courtheyn, M.L. Essers, H.H. Hooijerink, E.O. van Bennekom, R.E. Fuchs, Rapid Commun. Mass Spectrom. 17 (2003) 1633. [5] J. Blanca, P. Mu˜noz, M. Morgado, N. M´endez, A. Aranda, T. Reuvers, H. Hooghuis, Anal. Chim. Acta 529 (2005) 2199. [6] B. Waldeck, Eur. J. Pharm. 445 (2002) 1. [7] A. Lommen, R. Schilt, J. Weseman, A.H. Roos, J.W. van Velde, M.W.F. Nielen, J. Pharm. Biomed. Anal. 28 (2002) 87. [8] P. Gallo, G. Brambilla, M. Fiori, A. Scaramuzzo, L. Serpe, Chromatographia 53 (2001) 446. [9] F.W. Smyth, J. Chromatogr. B 824 (2005) 1.