Anal. Chem. 1998, 70, 1362-1368
The Potential of Restricted Access Media Columns As Applied in Coupled-Column LC/LC-TSP/MS/MS for the High-Speed Determination of Target Compounds in Serum. Application to the Direct Trace Analysis of Salbutamol and Clenbuterol Elbert A. Hogendoorn and Piet van Zoonen
National Institute of Public Health and Environmental Protection (RIVM), P.O. Box 1, 3720 BA, Bilthoven, The Netherlands Aldo Polettini,* Giorgio Marrubini Bouland, and Maria Montagna
Department of Legal Medicine and Public Health, Laboratory of Forensic Toxicology, University of Pavia, Via Forlanini 12, I-27100 Pavia, Italy
This study investigated the potential of restricted access media (RAM) columns used as a first column in coupledcolumn LC hyphenated to thermospray tandem mass spectrometry (LC/LC-TSP/MS/MS) for the fast, selective, and sensitive determination of target drugs in serum samples. Because of their wide range in polarity, salbutamol and clenbuterol were chosen as model compounds and representatives of the class of β2-agonists. Three types of RAM columns were tested: (i) Pinkerton ISRP (internal surface reversed phase, 5 µm), (ii) SPS (semipermeable surface, 5 µm C18), and (iii) RP-18 ADS (alkyl-diol silica, 25 µm). A 3-µm C18 column (50 mm × 4.6 mm i.d.) was chosen as the second column. Tandem mass spectrometric detection was carried out in the selected reaction monitoring (one parentfone daughter) mode. With regard to retention and, moreover, the peak elution volume of the analytes, the ISRP material was found to perform best: a 50-mm × 4.6-mm i.d. ISRP column in combination with a 100% aqueous buffer (pH of 7.0 ( 0.2) allowed the injection of large volumes (up to 200 µL) of sample without additional band broadening of the analytes and provided sufficient preseparation between analytes and large-molecule serum constituents. Under the selected conditions, both analytes could be determined in serum samples up to a limit of quantification (LOQ) of 0.5 ng/mL, with a sample throughput of 7 and 5 h-1 for salbutamol and clenbuterol, respectively. Method validation was carried out by analyzing, in the course of several days, calf and human serum samples spiked with the analytes. In the case of salbutamol, the overall recovery from serum samples spiked at levels between 0.5 and 50 ppb (n ) 33) was 103.4%, with a repeatability of 12.7% and reproducibility of 14.3%. The overall recovery for clenbuterol was 99.6% (n ) 15, spiked level 0.5-5 ppb), with a repeatability of 15.2% and reproducibility of 16.4%. The adopted LC/LC-TSP/MS/
MS analyzer appeared to be very robust under the selected conditions, and, after the period of analysis involving the processing of more than 100 mL of serum, neither loss of chromatographic performance nor pressure increase of columns or of the interface was observed.
1362 Analytical Chemistry, Vol. 70, No. 7, April 1, 1998
S0003-2700(97)01030-5 CCC: $15.00
β2-Agonists form a vast class of sympathomimetic drugs, with preferential affinity for β2-adrenergic receptors. Through the activation of pulmonary β2 receptors, these drugs produce relaxation of bronchial smooth muscle and decrease airway resistance, thus exerting a bronchodilatory action which is extensively exploited in the treatment of asthma. Together with this main therapeutic action, β2-agonists produce, at higher doses, side effects on protein synthesis and lipolysis which result in a growth-promoting action. Owing to both the stimulatory and the anabolic effects, β2-agonists possess a high potential for misuse in sports as well as in zootechnics.1 To monitor therapeutic use as well as to control illegal use of β2-agonists, many methods for the determination of these drugs in biological matrixes have been published and recently reviewed and discussed.1,2 As for many other classes of drugs, the most common strategy for the analysis of β2-agonists residues in biofluids (urine, serum, vitreous humor) is based on large-scale immunochemical screenings, followed by GC/MS confirmation of positives. However, the choice of GC/MS as the confirmatory technique is mainly based on practical considerations, such as the widespread availability of low-cost GC/MS equipment in analytical laboratories, more than on theoretical issues. In fact, the need for a derivatization procedure prior to analysis in order to reduce the polarity and to improve the mass spectral selectivity of these compounds makes GC/MS methods laborious and timeconsuming. Methods based on LC/MS/MS3 or on MS/MS4 * To whom correspondence should be addressed. Phone: ++39-382-527958. Fax: ++39-382-528025. E-mail:
[email protected]. (1) Polettini, A. J. Chromatogr. B 1996, 687, 27-42. (2) Boyd, D.; O’Keeffe, M.; Smyth, M. R. Analyst 1996, 121, 1R-10R. © 1998 American Chemical Society Published on Web 03/04/1998
analysis have been proposed as alternatives. Although having the undoubted advantage of not requiring derivatization, these methods cannot avoid the extraction of analytes to prevent matrix interference on detection. On the other hand, the possibility of direct analysis of the sample with no or minimal pretreatment using hyphenated liquid chromatographic and mass spectrometric techniques eliminates the typical drawbacks of sample preparation (loss of analyte through adsorption phenomena, hydrolytic or oxidative degradation, contamination) and, in addition, increases substantially the sample throughput, thus partly counterbalancing the high operative cost of the equipment required. In previous work,5,6 the feasibility of coupled-column liquid chromatography (LC/LC) hyphenated to thermospray tandem mass spectrometry (TSP/MS/MS) has been demonstrated for the direct determination of β2-agonists in bovine urine at the low ppb level. Providing on-line desalting of large-volume urine samples and a very stable flow during TSP-MS detection of the analytes, the use of LC/LC with two C18 separation columns instead of conventional LC was preferable.6 The aim of this study was to investigate the viability of the LC/LC-TSP/MS/MS approach for the determination of these analytes in the serum. However, for this matrix, the use of a fullsize C18 separation column as the first column is not possible, as the direct injection of large volumes of serum would rapidly deteriorate the column, owing to denaturization and precipitation of serum proteins. For this purpose, columns packed with restricted access media (RAM) materials have been successfully developed.7-16 These materials are nonadsorptive and/or exclusionary to large molecules such as serum proteins but allow hydrophobic interaction of small molecules, providing retention of analytes and almost no retention of proteins. Because of their difference in RAM concept and their commercial availability, three different RAM materials were selected in this study and tested to determine their performance in the LC/MS analysis of β2-agonists in serum. Because of their large difference in polarity, salbutamol (highly polar) and clenbuterol (moderately polar) were selected as model compounds to represent the class of β2-agonists and other drugs with similar properties. The structures of the two compounds are shown in Figure 1. EXPERIMENTAL SECTION Chemicals. Clenbuterol hydrochloride and salbutamol sulfate were obtained from Biomedica Foscama (Rome, Italy) and Glaxo (3) Doerge, D. R.; Bajic, S.; Blamkenship, L. R.; Preece, S. W.; Churchwell, M. I. J. Mass Spectrom. 1995, 30, 911-916. (4) van Rhijn, J. A.; O’Keeffe, M.; Heskamp, H. H.; Collins, S. J. Chromatogr. A, 1995, 712, 67-73. (5) Hogendoorn, E. A.; van Zoonen, P.; Polettini, A.; Montagna, M. J. Mass Spectrom. 1996, 31, 418-426. (6) Polettini, A.; Montagna, M.; Hogendoorn, E. A.; Dijkman, E.; van Zoonen, P.; Ginkel, L. A. J. Chromatogr. A 1995, 695, 19-31. (7) Anderson, D. J. Clin. Chem. 1993, 65, 434R-443R. (8) Haginaka, J. Trends Anal. Chem. 1991, 10, 17-22. (9) Unger, K. K. Chromatographia 1991, 31, 507-511. (10) Hagestam, I. H.; Pinkerton, T. C. Anal. Chem. 1985, 57, 1757-1763. (11) Cook, S. E.; Pinkerton, T. C. J. Chromatogr. 1986, 368, 233-248. (12) Pinkerton, T. C.; Koeplinger K. A. Anal. Chem. 1990, 62, 2114-2122. (13) Desilets, C. P.; Rounds, M. A.; Regnier, F. E. J. Chromatogr. 1991, 544, 25-39. (14) Wang, H.; Desilets, C.; Regnier, F. E. Anal. Chem. 1992, 64, 2821-2825. (15) Boos, K.-S.; Rudolphi, A.; Vielhauer, S.; Walfort, A.; Lubda, D.; Eisenbeiss, F. Fresenius J. Anal. Chem. 1995, 352, 684-690. (16) Vielhauer, S.; Rudolphi, A.; Boos, K.-S.; Seidel, D. J. Chromatogr. 1995, 666, 315-322.
Figure 1. Structures of selected target analytes.
Wellcome (London, UK), respectively. A 1.00 mg/mL stock solution of each analyte was prepared in methanol and stored at -20 °C. Stock solutions were diluted in HPLC-grade water and kept in a refrigerator at about 4 °C. Internal standards, clenbuterol-d6 hydrochloride and salbutamol-d6 sulfate, were obtained from the National Institute of Public Health and Environmental Protection (Bilthoven, The Netherlands) and from the State Institute of Quality Control of Agricultural Products (Wageningen, The Netherlands), respectively. The internal standards were individually dissolved in HPLC-grade water to obtain a concentration of 5 µg/mL. Methanol (UV-grade), water (HPLC-grade), and triethylamine p.a. (99.5%) were purchased from Carlo Erba (Milan, Italy). Ammonium acetate p.a. (99%) and formic acid p.a. (96%) were obtained from Aldrich (Milan, Italy). The mobile phase (M-1) applied on the first column (C-1) consisted of 0.1 M ammonium acetate and 0.01 M triethylamine in water, adjusted to pH 7.0 ( 0.2 with formic acid. On the second column (C-2), mobile phases (M-2) containing 0.1 M ammonium acetate, 0.01 M triethylamine, and 0.17 M formic acid consisting of methanol-water (30:70, v/v) and methanol-water (15:85, v/v) were used for clenbuterol and salbutamol analysis, respectively. Before use, a freshly prepared mobile phase was filtered over a type HA 0.45 µm filter from Millipore Waters (Milford, MA). A rinsing mobile phase (M-R) consisting of methanol-water (25:75, v/v) was used for the cleanup of C-1 after each analytical session. Columns. As C-1, the following column materials (A, B, or C) and dimensions were tested: (A) 5-µm Pinkerton ISRP GFFII-S5-80 (Regis, Morton Grove, IL) in columns of 10 mm × 3 mm i.d. and 50 mm × 4.6 mm i.d.; (B) 5-µm SPS-5PM-S5-100-ODS (Regis) in columns of 10 mm × 3 mm i.d. and 150 mm × 4.6 mm i.d.; and (C) 25-µm LiChrospher RP-18 ADS (Merck, Darmstadt, Germany) in a column of 25 mm × 4 mm i.d. Final analyses on C-1 were performed on a type A 50-mm × 4.6-mm-i.d. column. A 10-mm × 3-mm-i.d. precolumn packed with the same material was installed before the analytical column. As C-2, a 50-mm × 4.6mm-i.d. cartridge column packed with 3-µm MicroSpher C18 (Chrompack, Bergen op Zoom, The Netherlands) with a 10-mm × 3-mm-i.d. RP-guard column (Chrompack) was used. Apparatus. The LC/MS instrumentation (see also Figure 2) consisted of the following components: a Rheodyne (Cotati, CA, USA) 7125 manual injector, I; a Valco (VICI, Schenkon, Switzerland) EC4W security switching valve controlled by the MS system, situated between C-2 and the TSP/MS; two Rheodyne 7010 switching valves controlled by a MUST device from Spark Holland (Emmen, The Netherlands) (one valve, HV, displayed in Figure 2, is used for column switching, while the other can be used for Analytical Chemistry, Vol. 70, No. 7, April 1, 1998
1363
phase (M-R) on the ISRP column (C-1) was set at a flow of 0.2 mL/min, while on the second column (C-2) the flow was stopped. Biological Samples. Blank calf serum was obtained from the National Institute of Public Health and Environment, while blank lyophilized human serum (Seronorm) was purchased from Nycomed Pharma As (Oslo, Norway). It was prepared by dissolving the content of an ampule in 5 mL of distilled water. Blank bovine plasma was kindly provided by the Institute of Physiology and Biochemistry of the Faculty of Veterinary of the University of Milan. Sample Pretreatment. Serum and plasma samples (about 5 mL) were filtered over 0.22-µm Millex-GS filters (Millipore). To a volume of 1.0 mL of sample (filtered serum/plasma or aqueous solution of the analytes) was added 25 µL of the appropriate internal standard solution (see above) prior to injection. Figure 2. Scheme of the LC/LC-TSP/MS/MS analyzer. I, injector; P, LC pump; UV, UV detector; TSP/MS/MS, thermospray tandem mass spectrometer; C-1, RAM first separation column; C-2, C18 second separation column; M-1 and M-2, mobile phases used on C-1 and C-2, respectively; W, waste. Table 1. Experimental LC/LC-TSP/MS/MS Conditions for the Determination of Salbutamol and Clenbuterol part
parameter
LC/LC
injection volume (µL) cleanup volume (mL) transfer volume (mL) repeller (V) vaporizer temp (°C) source block temp (°C) collision energy (-eV) collision gas (argon) pressure (mTorr) multiplier (eV) reactions monitored (m/z) of analyte reactions monitored (m/z) of hexadeuterated analyte
TSP MS/MS
salbutamol
clenbuterol
100 1.3 0.35 45 110 200 25 2.5
200 3.4 1.0 60 105 200 25 1.7
1750 240f148
2000 277f203
246f148
283f203
rinsing the column); two Waters model 6000A isocratic LC pumps, P, for delivering M-1 and the rinsing mobile phase; a Waters model 600MS gradient LC pump equipped with a flow rate stabilization option and a helium degassing system, P, for the delivering of M-2; a Waters model 441 UV detector set at 280 nm, UV; a PerkinElmer (Norwalk, CT) model 561 recorder; a Finnigan MAT (San Jose, CA) TSQ-700 tandem mass spectrometer, TSP/MS/MS, equipped with a Finnigan MAT TSP-2 thermospray interface and a cold trap containing liquid nitrogen. Procedure. The flows of the mobile phases M-1 and M-2 were set at 1.0 and 1.3 mL/min, respectively. The TSP interface was operated in ion evaporation mode (filament and discharge off), and the MS/MS analysis was performed in the selected reaction monitoring, parent mode. The experimental LC/LC-TSP/MS/ MS conditions used for the determination of salbutamol and clenbuterol are given in Table 1. Quantification of clenbuterol or salbutamol in serum/plasma samples was performed by comparing the peak area ratio clenbuterol/clenbuterol-d6 or salbutamol/salbutamol-d6 of the sample with that obtained for the injection of standards. After each day of analysis, the thermospray interface was washed with methanol (5 mL) followed by water (5 mL); overnight, the rinsing mobile 1364 Analytical Chemistry, Vol. 70, No. 7, April 1, 1998
RESULTS AND DISCUSSION General LC Aspects. Three (A, B, and C) different RAM materials (see Experimental Section) were studied. Both ISRP and SPS consist of highly efficient, small (5 µm) particles; however, the concept of excluding large-molecule serum proteins from retention is different. The first developed RAM material, Pinkerton ISRP,10-12 uses chemically and enzymatically modified silica particles. They consist of a hydrophobic bonded phase (glycine-L-phenylalanineL-phenylalanine, GFF) inside the small pores (80 Å) and outside a hydrophilic glycine phase, excluding in this way hydrophobic interaction of large protein molecules. The SPS concept13,14 makes use of silica particles on which first a covalently bonded polyoxyethylene polymer network is formed, followed by bonding a conventional phase, here (ODS) C18, underneath the polymer. Lichrospher RP-18 ADS was recently developed by Boos et al.15,16 as a special reversed-phase sorbent for use as a precolumn packing. The chemically and enzymatically modified material consists of rather large, porous silica particles (25 µm) with small pores (60 Å), having an external surface of hydrophilic electroneutral diol groups and an internal surface of hydrophobic C18 alkyl chains. From ISRP and SPS materials, both precolumns (10 mm × 3 mm i.d.) and separation columns of 50 mm × 4.6 mm i.d. (ISRP) and 150 mm × 4.6 mm i.d. (SPS) were available (see Experimental Section). RP-18 ADS material is only supplied in a rather large precolumn (25 mm × 4 mm i.d.). The separation columns are able to perform an efficient one-column separation between proteins and analytes.11-17 However, to prevent elution of proteins into the interface of the TSP/MS, column switching is needed. As demonstrated in previous coupled-column RPLC studies5,6,21,22 involving the trace analysis of polar analytes in aqueous samples, the analyte’s retention (k value) and the separation power of the column are important, since they influence both the (17) Brewster, J. D.; Lightfield, A. R.; Barford, R. A., J. Chromatogr. 1992, 598, 23-31. (18) Chromatography Catalog/Guide; Regis Technologies, Inc.: Morton Grove, IL, 1993; p 78. (19) Yu, Z.; Westerlund, D. J. Chromatogr. A 1996, 725, 137-147. (20) van der Hoeven, R. A. M.; Hofte, A. J. O.; Frenay, M.; Irth, H.; Tjaden, U. R.; van der Greef, J.; Rudophi, A.; Boos, K.-S.; Marko Varg, G.; Edholm, L. E. J. Chromatogr. A 1997, 762, 193-200. (21) Hogendoorn, E. A.; Verschraagen, C.; van Zoonen, P.; Brinkman, U. A. Th. Anal. Chim. Acta 1992, 268, 205-215. (22) Hogendoorn, E. A.; van Zoonen, P. J. Chromatogr. A 1995, 703, 149-166.
Table 2. Elution Performance of Salbutamol,a Clenbuterol,a and Serumb on RAM Columnsa RAM packing 5 µm ISRP
column (l × i.d. (mm)) 10 × 3
5 µm ISRP
50 × 4.6 + 10 × 3
5 µm SPS
10 × 3
5 µm SPS 25 µm ADS
150 × 4.6 + 10 × 3 25 × 4
compound
tr (min)
k
peak vol. (µL)
salbutamol clenbuterol salbutamol clenbuterol salbutamol clenbuterol salbutamol clenbuterol salbutamol clenbuterol
0.30 0.68 1.55 3.6 0.20 0.60 1.63 8.7 3.0 >100
1.0 3.5 1.1 3.8 0.33 3.1 0.23 5.7 9
nmc 550 250 950 nmc 175 175 2100 4000
As
SEPb (min) 0.9
2.0 1.7 2.8
2.0 0.9
2.0 2.5 3.3 1.6
ntd 2.1
a Injection of 200 µL of an analyte standard (10 µg/mL) in water; mobile phase, buffer pH 6.9 at 1 mL/min. b SEP, serum elution performance expressed as the time needed for a 90% reduction of the maximal UV (280 nm) signal in the first part of the chromatogram for 200 µL of serum injected. c Not measurable because of partial coelution with solvent front. d Not tested.
maximum tolerable sample injection volume (sensitivity) and the potential for separation between analyte(s) and early-eluting interferences (selectivity). In fact, if hydrophobic retention is insufficient (k < 1 in 100% water), efficient RPLC analysis becomes highly unlikely. In comparison to a conventional C18 phase, RAM phases provide less hydrophobic retention.18 In addition, the fact that studies using RAM precolumns16,19-21 deal with rather nonpolar analytes (elution of analytes on C18 column with mobile phase containing 60% of modifier5) emphasizes that limitations of RAM materials to very polar analytes, e.g. salbutamol, can be expected. Testing of RAM Columns. To maximize retention of the analytes and to minimize protein precipitation, an aqueous mobile phase was selected as the RAM column eluent. Because of the basic and/or amphoteric properties of β2-agonists (see also Figure 1), the buffer constituents ammonium acetate, triethylamine, and formic acid from previous work5,6 were selected. To fulfill recommended ISRP and SPS conditions,10,12,18 the pH of the buffer was set at 7.0 ( 0.2 (see Experimental Section). Aiming at a sensitivity comparable to that obtained in previous work,5 the chromatographic performance of ADS precolumns and both ISRP and SPS precolumn and separation columns was tested by injecting 200 µL of the standard solutions of clenbuterol and salbutamol in water. The results obtained with UV detection at 280 nm, expressed in terms of retention time (tr), capacity factor (k), peak elution volume (peak vol.) and peak asymmetry factor (As), are given in Table 2. The latter two parameters were calculated at 10% of the peak height. The void volume of the columns was calculated by using the time at which the first disturbance of the baseline in the chromatogram caused by the solvent front occurred after injection. In case of analytes with little retention, the time required to elute the serum proteins from the column is important. Therefore, the serum elution performance (SEP) was studied by recording the UV response at 280 nm of 200 µL of serum injected. The SEP of the RAM column, expressed as the time needed for a 90% reduction of the maximal UV signal of serum constituents, is given in the last column of Table 2. From the results of Table 2, several conclusions about the RAM columns can be made. ISRP. Shows marginal retention for salbutamol (k ≈ 1) and a moderate retention for clenbuterol (k ≈ 3.5). The serum elution performance indicates that, even for the less polar clenbuterol, the use of a precolumn results in the transfer of a considerable
amount of serum constituents to the second column. The analytical column provides an appropriate separation between clenbuterol and serum constituents. Concerning salbutamol, coelution with serum constituents to the second column (C-2) would be unavoidable although limited owing to the small elution volume of the analyte. SPS. Despite the potentially higher hydrophobicity, salbutamol was less retained than by the Pinkerton ISRP material; hence, adequate separation between the polar analyte and the serum proteins cannot be expected. With regard to the analytical column, SPS provides more retention for clenbuterol than Pinkerton ISRP. From the point of view of efficiency, the 15-cm length of the analytical column (shortest length commercially available) is not favorable; beside the costs, the overkill in separation power enhances band broadening of the analyte and cleanup time. Therefore, this analytical column was not further tested for serum. ADS. The retention of clenbuterol makes this column suitable for trapping and allows large-volume sample injection, together with adequate exclusion of large molecular weight compounds. In comparison to those obtained with Pinkerton ISRP and SPS, the capacity factor of salbutamol was somewhat higher. However, band broadening of the analyte on the ADS column was distinctly higher. A 90% reduction of the UV signal (280 nm) required about 2 min, indicating that a considerable coelution of serum constituents would take place during the quantitative transfer of the analyte to C-2. Coupled-Column Conditions. Based on the results of Table 2 and on the considerations made above, both the analytical ISRP column and the ADS precolumn were further evaluated for their performance involving column switching. In particular, the ability to obtain peak compression of analytes was tested by analyzing standard solutions of clenbuterol and salbutamol with LC/LCUV (280 nm). Coupled-column LC can be performed in either the forward or back flush mode. When dealing with very polar analytes, e.g., salbutamol, which already migrate through the column during sample injection and cleanup, the forward flush mode in LC/LC is preferable21,22 because it avoids unnecessary band broadening of the analyte. However, the ADS precolumn is designed to be used in the back flush mode;15,16,19,21 therefore, both column-switching modes were tested on the ADS column. The results given in Table 3 clearly show that, with regard to the band broadening of the analytes on the ADS column, the back flush mode is distinctly more appropriate than the forward flush Analytical Chemistry, Vol. 70, No. 7, April 1, 1998
1365
Table 3. LC/LC-UV (280 nm) of Salbutamol and Clenbuterol Using ISRP and ADS Columns as C-1a C-1
compound
C-1 modeb
cleanup volume (mL)
transfer volume (mL)
trc (min)
peak volume (µL)
As
peak height (mm)d
25 µm ADS, 25 mm × 4 mm i.d.
salbutamol salbutamol clenbuterol clenbuterol salbutamol clenbuterol
FF BF FF BF FF FF
1.0 1.0 4.0 4.0 1.4 3.4
1.0 1.0 1.0 1.0 0.40 1.0
2.08 1.77 4.10 3.86 2.73 5.84
530 300 1540 800 170 365
1.9 2.0 2.6 2.2 1.4 1.4
69 108 15 35 225 63
5 µm ISRP, 50 × 4.6 mm i.d.
a LC conditions: injection, 200 µL of standard in water; C-2, 50 mm × 4.6 mm i.d., 3 µm C18; M-1, 100% buffer, pH 6.8 at 1 mL/min (ISRP column) or 2 mL/min (ADS column); M-2 for salbutamol, methanol-buffer pH 4 (15:85, v/v) at 1.4 mL/min; M-2 for clenbuterol, methanol-buffer pH 4 (32:68, v/v) at 1.4 mL/min. b FF, forward flush; BF, back flush. c Sum of retention on both columns. d Corresponding to 2 µg of analyte detected at 0.1 AUFS (10 mV).
mode. Furthermore, under the preferable conditions, viz. back flush on ADS and forward flush on Pinkerton ISRP, sensitivity expressed as the peak height was about 2 times higher when using the ISRP column. According to the results, separation of salbutamol from serum proteins would be incomplete on both RAM columns. However, the unavoidable transfer of serum proteins to the second column would be limited in the case of the ISRP column, owing to the lower transfer volume (see Table 3). These features contributed to our decision to select the ISRP column as C-1 for further experimental work. Determination of Salbutamol. To minimize transfer of proteins to C-2 and their possible further transport to the TSP/ MS, the amount of serum sample injected needed to be adjusted, also taking into account the desired sensitivity (1-5 ng/mL). Based on previous information3 and on the small elution volume of the analyte observed under the established LC/LC conditions (see Table 3), a sample injection volume of 100 µL was selected. Scouting experiments with LC/LC-UV (280 nm) indicated that, when analyzing calf serum, the retention of salbutamol increased by about 5% in comparison to retention of the analyte in the standard solution. However, even with an adapted transfer volume (0.26-0.33 mL), no deterioration (clogging/reduction in efficiency) of C-2 was observed during the repeated analysis of serum samples (n ) 12) and standards (10 µg/mL; n ) 25). Moreover, variation of retention and peak height was less than 0.5% and 2%, respectively. Also, when TSP/MS/MS analysis (see Table 1 for conditions) was applied after LC/LC, no interface artifacts caused by coeluted serum constituents were observed. The approach was validated by analyzing, on two different days, salbutamol in calf and human serum spiked at different levels. The overall results, summarized in Table 4, showed that both the repeatability and reproducibility were satisfactory. On the first day, the linearity was tested by randomly analyzing standards of salbutamol (placed between serum samples) with concentrations in the range of 5.0-100 ng/ mL. The calibration plot obtained (y ) 0.015x - 0.00322) showed an acceptable linearity (r2 ) 0.9956) over the tested range. However, calculation of the concentration in serum was done with a one-point calibration using a standard containing a concentration of salbutamol close to that in the sample. Similar to that adopted for urine samples,5 the analytical strategy consisted of the alternate analysis of a serum sample and a standard. This approach speeds up considerably the time of analysis by eliminating a washing and equilibration step of C-1 after the analysis of a serum sample. The performance of the method in terms of sensitivity and high speed 1366 Analytical Chemistry, Vol. 70, No. 7, April 1, 1998
Table 4. Overall Results of Salbutamol Recoveries from Serum (n ) 33) type of serum
day
spiked level (ng/mL)
no. of experiments
mean recovery (%)
RSD (%)
calf calf calf human human calf calf human
1 1 1 1 1 8 8 8
5.0 10 50 5.0 50 1.0 5.0 0.5-2.5
5 5 4 3 3 5 3 5
96 106 115 96 107 90 110 111
9.7 18.8 14.1 13.7 3.3 16.9 2.5 6.5
overall recovery (%) repeatability (%) reproducibility (%)
103.4 12.7 14.3
Figure 3. Ion chromatograms obtained for the LC/LC-TSP/MS/ MS analysis of human serum spiked at a level of 0.5 ng/mL with salbutamol. Peaks: 1, salbutamol; 2, salbutamol-d6 (12.5 ng/mL).
is illustrated in Figure 3, showing the LC/LC-TSP/MS/MS analysis of salbutamol in a human serum sample spiked at a level of 0.5 ng/mL. The instrumental time of analysis of less than 3 min makes a sample throughput of at least 7 h-1 possible. Determination of Clenbuterol. As the sensitivity for clenbuterol and salbutamol by TSP/MS/MS analysis appeared to be similar, more sample should be injected in the case of clenbuterol in order to compensate for its higher elution volume (see Table 3). With LC-UV (280 nm), the elution performance on the ISRP column of 200- and 500-µL serum volumes were studied. Similarly to salbutamol, an increase in retention of clenbuterol of 15 and 24 s was obtained for the injection of 200 and 500 µL of serum, respectively. The obtained separation between proteins and analyte allowed us to increase the transfer volume in order to take
Table 5. Overall Results of Clenbuterol Recoveries from Serum (n )15) type of serum
day
spiked level (ng/mL)
no. of experiments
mean recovery (%)
RSD (%)
calf calf human
1 21 21
0.5-5.0 2.5 2.5
6 4 5
108 88 99
18.3 18.6 3.5
overall recovery (%) repeatability (%) reproducibility (%)
99.6 15.2 16.4
into account the observed shift in retention. However, the 500µL serum injections provided a considerable increase (∼500 psi) in pressure, and about 30 min was required to reduce it. These drawbacks were not observed after the injection of 200-µL serum samples, and, consequently, this volume was selected for further work. The possibility to decrease the peak elution volume of clenbuterol by decreasing the k′ of the analyte on C-2 (5.5) was considered in order to improve the sensitivity of analysis. However, in a LC/MS/MS experiment with a 2.5% increase of organic modifier in M-2, the analytical response of clenbuterol became highly variable. Connection of C-2 to the UV detector (280 nm) revealed that, at the elution time of clenbuterol, the baseline was not sufficiently recovered from the column switching disturbance. This finding emphasized the necessity of stable elution conditions and clearly showed that, concerning this aspect, the conditions could not be further optimized. By applying the analytical strategy mentioned above for salbutamol, clenbuterol was determined in spiked calf and human serum samples on two different days. The overall results given in Table 5 demonstrate that the analytical performance of the method is satisfactory. An example of the LC/LC-TSP/MS/MS analysis of clenbuterol in a calf serum sample spiked at the level of 0.50 ng/mL is given in Figure 4. It clearly shows that this approach allows the determination of this analyte in serum samples at the sub-ppb level with a sample throughput of about 5 h-1. The described analytical method was also applied on a plasma sample with an incurred clenbuterol residue collected from a calf which had been treated with the drug. As shown in Figure 4, a clenbuterol concentration of 1.3 ng/mL was found with the LC/ LC-TSP/MS/MS method, which was in agreement with a previous analysis employing enzyme immunoassay. However, the pressure on C-1 increased when injecting plasma, owing likely to the presence of fibrin in the sample, and replacement of the precolumn was necessary after the analysis of a few samples. During a 1-month intensive use of the same ISRP column (C-1) involving the injection of more than 100 mL of serum, a slow decrease in the retention of clenbuterol (5 to 3 min) and in the column efficiency (about 3 times) and a small increase in pressure (about 200 psi) were observed. However, by adjusting daily the LC/LC conditions, no problems were encountered during TSP/MS analysis, thus indicating that, despite the slight decrease in performance, the applicability of the column remained satisfactory. (23) Van Vyncht, G.; Preece, S.; Gaspar, P.; Maghuin-Rogister, G.; DePauw, E. J. Chromatogr. A 1996, 750, 43-49.
Figure 4. Ion chromatograms obtained for the LC/LC-TSP/MS/ MS analysis of (A) calf serum spiked with 0.5 ng/mL clenbuterol and (B) positive bovine plasma sample containing a clenbuterol residue of 1.3 ng/mL. Peaks: 1, clenbuterol; 2, clenbuterol-d6 (25 ng/mL).
The repeatability data of the individual days (see Tables 4 and 5) indicate that the decrease in column performance does not affect the accuracy and precision of the method. For example, about 100 mL of serum had already been processed when the analysis of clenbuterol on day 21 was performed (see Table 5). CONCLUSIONS This study clearly showed that the application of restricted access media (RAM) columns as the first column in the coupledcolumn LC hyphenated to TSP/MS/MS analysis enables the sensitive, selective, and rapid determination of β2-agonists in serum. The analyzer (Figure 2) makes it possible to determine the target analyte in less than 10 min to a level of at least 0.5 ng/mL. As shown in a recent comparative study,23 improved sensitivity can be obtained when employing atmospheric pressure chemical ionization (APCI) instead of TSP. The relatively wide range in polarity of the tested amphoteric compounds salbutamol (very polar) and clenbuterol (moderately polar) indicates that the applied procedure is potentially suitable for many other compounds. Under the applied conditions, the Pinkerton ISRP column resisted the processing of more than 100 mL of serum. On the various tested RAM columns, salbutamol could not be fully separated from serum proteins. However, the small transfer volume obtained with the selected ISRP column allowed the processing of more than 50 mL of serum without clogging or efficiency decrease of the second column caused by the coelution of serum constituents. ACKNOWLEDGMENT This work has been carried out with financial support form the European Commission (DG XII, Science and Technology, Analytical Chemistry, Vol. 70, No. 7, April 1, 1998
1367
Division XII-H-1), in the framework of the Human Capital and Mobility Program, as part of the network project “Hyphenation Analytical Chemistry for Environmental and Public Health Research in the European Union (Grant No. ER-BCHRXCT 930274)”. The authors thank Prof. Camillo Secchi and Dr. Vitaliano Borromeo of the Institute of Physiology and Biochemistry, Faculty
1368 Analytical Chemistry, Vol. 70, No. 7, April 1, 1998
of Veterinary, University of Milan, for supplying a bovine plasma sample from a calf treated with clenbuterol. Received for review September 17, 1997. January 15, 1998. AC971030W
Accepted
