Mass Spectrometric Characterization of Different Norandrosterone Derivatives by Low-cost Mass Spectrometric Detectors Using Electron Ionization and Chemical Ionization D. de Boer,* E. G. de Jong, and R. A. A. Maes Netherlands Institute for Drugs and Doping Research, State University of Utrecht, Vondellaan 14, 3521 GE Utrecht, The Netherlands
The abuse of nortestosterone in sport is an important problem in doping-control analysis. In order to detect the main urinary metabolite of nortestosterone, norandrosterone (NA), sensitive and specific methodology is necessary. In this context the use of a low-cost mass spectrometric detector such as the Finnigan MAT ion-trap detector (ITD) was studied. The electron ionization (EI) and positive-ion chemical ionization (PICI) mass spectra of the methoxime-trimethylsilyl, trimethylsilyl-enol trimethylsilyl ether and pentafluoropropionic ester derivatives of NA are described. The limits of detection of these derivatives are compared with those obtained by the Hewlett-Packard mass selective detector (MSD), another low-cost mass spectrometric detector and operating only in the EI mode. For the derivatives of the reference standard of NA the ITD has in the EI mode the same limit of detection, in the range of 0.5 to l n g injected on the column, as the MSD. However, under these conditions the ITD provides more spectrometric information, because it gives full scan data. Moreover, with the same or even improved limits of detection the ITD can operate in the PICI mode. On the other hand, for the analysis of NA isolated from urine samples, the performance of the MSD was better than that of the ITD. The ion trapping technique is probably limited when the chemical background is high.
Norandrosterone (NA) is the main urinary metabolite of nortestosterone (NT),’ which is an anabolic steroid extensively abused in sport. In 1987, the laboratories accredited to the International Olympic Committee (IOC) found 52% urine samples positive for anabolic steroids. NA was found in 50% of these cases. For 1988 these figures were, respectively, 79% and 38Y0.~This number of positive cases not only reflects its popularity among athletes, but also results from its pharmacodynamic properties. NT is administered intramuscularly as an ester and is released gradually into the bloodstream. As the metabolic clearance of NT is also relatively slow, the metabolites of NT can be found in the urine even several months after administration. The chance of being found positive is therefore relatively high albeit that the concentrations of the urinary metabolites usually will be relatively low. The abuse of NT in sport can therefore be considered as an important problem in doping-control analysis, and demands very sensitive and specific methodology. For doping control, the IOC requires identification by mass spectrometry, as other techniques such as immuno-assays3 are either not sensitive or specific enough. The availability of low-cost mass spectrometric detectors such as the ion trap detector (ITD, Finnigan-MAT, San Jose, CA, USA) and the mass selective detector (MSD, Hewlett-Packard, Palo Alto, CA, USA) makes it possible to fulfil this requirement. Each detector is based on a totally different massanalyzer principle. The MSD is a conventional quadrupole mass analyzer and can operate only in the electron ionization (EI) mode. The ITD is a mass analyzer based on ion-trapping techniques, and can operate in the E I mode as well as in the positive-ion chemical ionization (PICI) mode. Some EI mass spectrometric * Author to whom correspondence should be addressed.
characterizations of methoxime-trimethylsilyl (MOTMS), trimethylsilyl-enol trimethylsilyl (TMS-enol TMS) and trimethylsilyl (TMS) ether derivatives of N A have been r e p ~ r t e d . ~The , ’ TMS-enol TMS ether derivative in particular has proven its applicability in routine doping control. This study describes the EI and PICI mass spectra of TMS-enol TMS ether, MO-TMS ether and pentafluoropropionic (PFP) ester derivatives of N A using the ITD. The E I and PICI detection limits are compared with those obtained with the MSD in the EI mode. EXPERIMENTAL Materials and samples Sephadex LH-20 was obtained from Pharmacia (Woerden, The Netherlands). N,O-Bistrimethylsilyltrifluoroacetamide (BSTFA) , trimethylchlorosilane (TMCS) , and pentafluoropropionic anhydride (PFPA) were purchased from Pierce (Oud-Beyerland, The N-Methyl-N-trimethylsilyltrifluoroNetherlands). acetamide (MSTFA) was from Macherey-Nagel (Duren, FRG) and methoxylamine hydrochloride (grade I), dithioerythritol (DTE) and trimethyliodosilane (TMIS) were from Sigma (St. Louis, MO, USA). A stock solution of norandrosterone at a concentration of 100ppm in methanol was kept at 20°C. All other solvents and reagents were of analytical grade. Urine samples containing NT metabolites were obtained from routine doping-control analysis as performed in our laboratory. Sample preparation Working standard solutions of the reference standard NA were made from the stock solution and volumes of this stock solution were evaporated to dryness under
0951-4198/90/0181-0185$5.00
0
John Wiley & Sons Limited. 1990
RAPID COMMUNICATIONS IN MASS SPECTROMETRY, VOL. 4. NO. 6. 1990 181
NORANDROSTERONE ASSAY WITH LOW-COST DETECTORS
nitrogen at 55 "C. Urine samples were prepared either by XAD-2 extraction or by enzymatic hydrolysis and diethylether extraction.6 The residues containing the steroids were dried under reduced pressure over phosphorus pentoxide and potassium hydroxide for at least 1h. TMS-enol T M S derivatization. A derivatization solution was prepared of MSTFA, TMIS (1000:2, v/v) and 2 mg/mL DTE.' Of this reagent solution, 50 pL were added to the dry residue containing the steroids. The mixture was heated for 30 min at 60 "C. MO- T M S deriuatization. A derivatization solution was prepared of methoxylamine hydrochloride in dry pyridine (8% w/v) and 100yL was added to the residue containing the steroids.' The mixture was heated for 30min at 60°C. The solvent was removed under a stream of nitrogen and 50 pL of a second derivatization mixture consisting of BSTFA, TMCS and TSIM (1000:50:20, v/v) was added. This mixture was heated for 15 min at 60 "C. Excess of derivatization reagents was removed by filtration through a column (20X5mm) of Sephadex LH-20 slurry packed in a Pasteur pipette using chloroform + hexane (1:1) as eluant. The steroid derivatives were eluted in the first 2 m L of eluant and the solvent was removed under a gentle stream of nitrogen. The residue was dissolved in 50 pL isooctane for analysis by gas chromatography/ mass spectrometry (GUMS). PFP deriuatization. To the residue containing the steroids was added 100 pL of PFPA and the mixture was heated for 30 minutes at 60 "C. Excess of derivatization reagent was removed under nitrogen and the residue was dissolved in 50 yL isooctane for G U M S analysis. GCIITD-MS analysis. The analysis was carried out 6n a Finnigan MAT ITD 800 with CI option. The open-split interface was used as supplied by Finnigan MAT. The ITD was controlled by an AT-type personal computer (Bremen, FRG). The ITD was coupled to a Carlo Erba HRGC 5160 Mega series gas chromatograph (Carlo Erba, Milan, Italy) with an on-column injector. Automatic cooling of this injector was performed by the O C 516 controller. The G C was equipped with a DB-1 fused silica capillary column (J&W, Folsom, CA, USA), 3 0 m x 0 . 3 2 m m I D 0.25 pm film thickness. A deactivated fused silica column, 40 cm X 0.53 mm I D was used as a retention gap to protect the analytical column from contamination. Temperature program: the initial temperature was maintained at 80 "C for l m i n , ramped to 180°C at 40"C/min, and then to 280°C at 10"C/min and maintained there for 20min. The temperatures of the ion-trap manifold, interface and open-split were 230 "C, 265 "C and 260 "C, respectively. EI spectra were measured at 70eV and at an electron multiplier voltage 150 V above the value of the automatic tuning program. Normally the automatic gain control (AGC) was on and the B value was set at 6000. If specified, the AGC was off and the B value set at 25 000. PICI spectra were run with all parameters set according to the manufacturer's manual using isobutane as reagent gas. G C I M S D analysis. The analysis was carried out on a 5790 gas chromatograph (Hewlett-Packard) coupled to a 5970 MSD. The G U M S system was controlled by a 9816 workstation (Hewlett-Packard) with extra 20 Mb fixed disk drive (HP 9153B). The chromatograph was equipped with an HP-1 fused capillary column 18 m x
0.20 mm ID, 0.33 ym film thickness. Temperature program: the initial temperature was maintained at 180 "C for 1min, ramped to 220°C at 10 "C/min, then to 231 "C at 0.5 "Clmin, and finally to 280 "C at 20 "C/min, then maintained there for 10 min. The temperatures of the injection port and the transfer line were 250 "C and 270 "C, respectively. Helium was used as a carrier gas at a flow rate of 1mL/min. A sample volume of 1 pL was injected in the splitless mode and the split valve was opened 30 s after the injection. The MSD was operated in the EI mode at 70 eV. The instrument was autotuned according to the manufacturer's instructions. RESULTS AND DISCUSSION Ion-trap mass spectra In Table 1 the E I and PICI mass spectra are summarized for the GC/ITD analysis of the different NA derivatives. Similar to the fragmentation in nonderivatized steroids,q the losses of a methyl radical, a water molecule and a combination of the two are observed. As well as these losses, the characteristic losses of trimethylsilanol (TMSOH), pentafluoropropionic acid molecules and methoxy radicals are observed, respectively, for the different derivatives under both ionization conditions. The loss of 44 mass units in the fragmentation of the PFP ester in the EI mode could be explained by the loss of C 0 2 , involving a rearrangement in the PFP group. Isotope labelling studies are needed to prove this statement. With chemical ionization, abundant molecular ions and less fragmentation are expected, especially with the use of isobutane as a reagent gas, which has a relatively high proton affinity."' Using chemical ionization on the ITD for those N A derivatives some changes in the fragmentation patterns are observed, although not always as expected. For instance, in the PICI mass spectrum of the TMS-enol TMS ether derivative the [ M + H ] + ion is prominent, but the base peak is formed by the loss of TMSOH, a process that apparently is more favourable under CI conditions. The fragmentation mechanism of this loss involves a hydrogen abstraction," which could explain its high abundance under CI conditions, where [M H]+ ions are formed. The other prominent ions found under E I conditions are also observed under CI conditions, such as the ions at m/z 420 and 405. The PICI mass spectrum of the MO-TMS ether derivative does show a base peak at the molecular ion, but because of the presence of a TMS group, the loss of a TMSOH molecule is also prominent. The peak at mlz256, which forms the base peak under El conditions, is also observed in the PICI mass spectrum with a high abundance. The PICI mass spectrum of the PFP ester derivative shows a simple fragmentation pattern compared to the E I spectra. The loss of a water molecule, which originates from the nonderivatized keto group, is the most important fragmentation. This loss involves a hydrogen abstraction too, which might explain its high occurrence under the CI conditions.
+
The limit of detection in the EI mode The limit of detection is the quantity that can be detected with reasonable certainty. With a 99.6% confidence level the limit corresponds to the mean of the
182 RAPID COMMUNICATIONS IN MASS SPECTROMETRY, VOL. 4, NO 6, 1990
0
John Wiley 61 Sons Limited. 1990
NORANDROSTERONE ASSAY WITH LOW-COST DETECTORS
Table 1. Partial mass spectrometric data of derivatives of norandrosterone as obtained by the ion trap detector TMS-enol TMS ether
MO-TMS ether
Ionization mode
420
EI
420 (25%) 405 (100%) 315 (44%) 255 (30%)
M" [M- CH,]' [M-CH,-TMSOH]+ [M - CH, - 2TMSOH]+
CI
421 (50%) 420 (55%) 405 (40%) 331 (100%) 241 (83%)
[M+H]+ M'' [M-CH,]' [M+H-TMSOH]' [M + H - 2TMS0HIi
EJ
362 (3%) 346 (48%) 256 (100%)
[M-CH,]' [M- OCH,]' [M - OCH, - TMSOH]'
CI
378 (100%) 347 (5%) 288 (26%) 256 (79%)
[M HI+ [M+H-OCH,]" [M H - TMSOH]' [M - OCH, - TMSOH]'
EI
422 (33%) 404 (56%) 389 (11%) 378 (100%) 363 (25%) 240 (39%)
M'' [M - HzO]+' [M-HZO-CH,]' [M - 441' [M-CH3-441i [M - H20 - PFPOH]'
CI
423 (5%) 405 (100%) 241 (19%)
[M+H]' [M + H - HZO]' [M + H - HZO - PFPOH]'
377
422
PFP ester
a
Ions and possible
mlra
Wt.
Mol
Dcrivative
(abundance)
fragmcntation
+ +
Normalized on base peak.
blank measures plus three times the standard deviation of the blank measures.'* The limits with this certainty for the reference standard of NA as determined with the ITD and the MSD in the EI mode are given in Fig. 1. For the ITD the detection limits for the derivatives studied, using the full scan mode, range between 0.5 and 1ng injected on the column. The AGC option in the ITD software provides automatic control of the ionization times. It was introduced to limit the total number of ions formed in order to prevent saturation of the ion trap. The loss of mass resolution at high concentrations of the analyte13 is then kept to a minimum. A t low concentrations of the analyte, as in our case, it is not necessary to limit the ionization time as the ion trap will not be saturated with ions of the analyte. Therefore the detection limit can be improved by switching off the norandrosterone derivative TMS e n o l ~TMS ether ~
mass analyzer MSD ITD
MO TMS ether ~
MSD
ITD
scan, , conditions
AGC and fixing the ionization time at the maximum value of 25 msec. In this way the detection limits for all derivatives studied were improved by a factor of two. For the MSD the detection limits range from 5 to 20 ng injected on the column, depending on the kind of derivative. The TMS-enol TMS ether derivative gives the best results, whereas the limit for the PFP ester derivative is four times greater. By using the selected-ion monitoring (SIM) mode the limits of detection as determined by the MSD improve to the same level as for the ITD in the full scan mode. However, by using SIM, less mass spectrometric information is obtained. Using the SIM mode for the ITD does not improve the limits further, in contrast to the MSD. This is inherent in the ion-trapping principle on which the ITD is based. All ions are formed at the same time and trapped for a relatively long time. The detection range in ng injected on the column (logarithmic scale) 40 30 20
10
5
1.0
0.5
0.2
0.1
El (scan) El (SIM) El (AGC on) El (AGC off) CI
El (scan) El (SIM) El (AGC on) El (AGC off) CI
PFP ether
MSD ITD
El (-)
El (SIM) El (AGC on) El (AGC off) CI
Figure 1. Detection ranges for the derivatives of the reference standard of norandrosterone using the ion trap detector (ITD) or the mass selective detector (MSD) under different scan conditions.
0
John Wiley & Sons Limited. 1990
RAPID COMMUNICATIONS IN MASS SPECTROMETRY, VOL. 4. NO. 6, 1990 183
NORANDROSTERONE ASSAY WITH LOW-COST DETECTORS
number of ions formed and therefore the sensitivity, is mainly determined by the ionization time. The ions are detected by ejection out of the trap, and whether a full spectrum or a limited mass range (SIM) is measured does not influence the detection limit very much. So both detectors reach the same range of detection limits in the E I mode for the derivatives of NA studied here. More mass spectrometric information is obtained with the ITD because it can maintain its performance in the full scan mode while the MSD is limited to the SIM mode.
47% 421
A
I
The analysis of urine samples The performance of the ITD and of the MSD was also tested using real urine samples containing the NA-glucuronide. NA was isolated from the urine according to the method routinely used for doping control in sport. Again the ITD could provide full scan data at low levels where the conventional MSD had to use SIM (Fig. 2). The possibility of using CI to confirm positive cases is advantageous too. However, the detection limits for the ITD using real urine samples are considerably higher than in the standard references. For this reason, problems occur with urine samples of athletes with very low traces of NA as seen 3 to 4 months after administration. Concentrations of 1-5 ng NA on column, corresponding to 10-50 ng/mL urine, could not be detected using the ITD. Even with the PFP ester derivatives using the PICI mode no traces of NA could be found. These concentrations could still be measured on the MSD albeit using the SIM mode. The reason is probably that the ion-trapping technique is limited when the chemical background level is high. First, the AGC will decrease the ionization time, because of the presence of the background ions, and will therefore also decrease the formation of the analyte ions. Second, the collection of ions will be influenced by the background ions. In the ITD, the ions are trapped in a field formed by two end caps and a ring electrode. By ramping the voltage applied to the electrodes, stored ions will become unstable and will leave the trapping field in the direction of the end electrodes. The lower end-cap electrode is perforated in order to let these ions pass and be detected by an electron multiplier. Background ions will mask the field of the electrodes, and the ejection will be less efficient. Therefore, background ions will limit the number of analyte ions that can be formed and can be ejected. This will have a disadvantageous effect on the detection limit of the trap.
J
I
I
I
I
I
I
I
I
I
56%
-
100%
Internal standard
259-
(a)
The limit of detection in the PICI mode The limits of detection in the PIC1 mode are also given in Fig. 1. The MSD can not operate in the CI mode, which is a great disadvantage. The detection limit for the TMS-enol TMS ether derivative in the PICI mode is comparable to that in the EI mode. This is not surprising considering the relatively high degree of fragmentation observed. For the MO-TMS derivative the detection limit improves by a factor of two. The best results are given by the PFP ester derivative. The limit improves by a factor of twenty. This improvement in detection limit is obtained because of the reduced fragmentation of the PFP derivative under CI conditions.
Norandrosterone
-
760 12:41
780 13:OI
800 13:21
820 13:41
840 14.01
860 1421
mh
Figure 2. Gas chromatographic/mass spectrometric data of the trimethylsilyl-enol trimethylsilyl ether derivative of norandrosterone isolated from a urine sample and analyzed by the ion-trap detector in the positive-ion chemical ionization mode: (a) ion chromatograms at mlz 421 and 331; (b) full-scan mass spectrum.
These problems do not occur in a quadrupole mass filter like the MSD where ionization and mass separation are separated.
Conclusion Anabolic steroid doping control requires sensitive and specific methodology. The ITD is a low-cost mass spectrometric detector which in the El mode has the same detection limits as the conventional quadrupole MSD. However, the ITD will provide more mass spectrometric information because it gives full scan data. Also the use of CI can increase the specificity of detection. Considering the implications of a positive doping sample, the analytical chemist would rather confirm the presence of a steroid with full scan data containing abundant molecular ions. In that respect, the ITD can satisfy these needs better than a conventional quadrupole detector like the MSD. In extreme cases, however, with concentrations at the low ng range, the MSD can still detect NA using the SIM mode when ITD detection is no longer possible. In conclusion, one can say that ITD provides the most specific data needed for confirmatory analyses in doping control, but the MSD has the lowest absolute detection limit in real samples. The recently introduced ITS-40 ion trap (Finnigan MAT) uses an extra resonance potential on the end caps. As a result, the ionization time can be increased further without reducing the mass resolution. More ions are ejected, and the sensitivity is improved. The sensitivities specified are 10-100 times better than for the ITD 800. Future experiments will determine whether this ITS-40 can obtain full scan data even for today's limiting cases in steroid doping control.
184 RAPID COMMUNICATIONS IN MASS SPECTROMETRY, VOL. 4, NO. 6. I990
0
John Wiley & Sons Limited, 1990
NORANDROSTERONE ASSAY WITH LOW-COST DETECTORS
Acknowledgements The reference standard stock solution of NA in methanol (100 ppm) DSH, (FRG). was a generous gift of Prof. Dr. M. We also thank Mr. P. w . van Dorp Van Vliet for his assistance in preparing the illustrations.
REFERENCES 1 . R. Masst, C. Lalibertt, L. Tremblay and R. Dugal, Biomed. Mass Spectrom. 12, 115 (1985). 2. IOC communications LAB/68/88/SJG and FICILAB112/ 89/SJG. 3. A. T. Kicmac and R. V . Brooks, J . Pharm. Biomed. Anal. 6,413 (1988). 4. R . Masst, C. LalibertC and L. Tremblay, J . Chromatogr. 339, 11 (1985). 5. D. A. Cowan and G . Woffendin, Finnegan MAT, Spectra 11 4 (1988).
0
John Wtley & Sons Limtted. 1990
6. M. Donike, J. Zimmermann, K-R. Barwald, W. Schanzer, V. Christ, K. Klostermann and G. Opfermann, Deutsch. 2. Sportmedizin 35, 14 (1984). 7. M. Donike and J . Zimmermann, J . Chromatogr. 202,483 (1980). 8. E. Houghton, p. Taele, M, c, Dumasia and J. K, Biomed. Mass Spectrom. 9, 459 (1982). 9. L. Tokts, R. T. LaLonPe and C. Djerassi, J . Org. Chem. 32, 1012 (1967). 10. F. W. Karasek and R. E. Clement; Basic Gas ChrornatographyMass Spectrometry, p. 45. Elsevier, Amsterdam (1988). 11. J. Diekman and C. Djerassi, J . Org. Chem. 32, 1005 (1967). 12, Nomenclature, Symbols, units and Their Usage in Spectrochemical Analysis-I1 Data interpretation, International Union of Pure and Applied Chemistry, Anal. Chem. 48, 2294 (1976). 13. J . W. Eichelberger and W. L. Budde, Biomed. Enuiron. Mass Spectrom. 14, 357 (1987). Received 23 March 1990, accepted 17 April 1990
RAPID COMMUNICATIONS IN MASS SPECTROMETRY, VOL 4, NO. 6. 1990 185
The abuse of nortestosterone in sport is an important problem in doping-control analysis. In order to detect the main urinary metabolite of nortestosterone, norandrosterone (NA), sensitive and specific methodology is necessary. In this context the use of a low-cost mass spectrometric detector such as the Finnigan MAT ion-trap detector (ITD) was studied. The electron ionization (EI) and positive-ion chemical ionization (PICI) mass spectra of the methoxime-trimethylsilyl, trimethylsilyl-enol trimethylsilyl ether and pentafluoropropionic ester derivatives of NA are described. The limits of detection of these derivatives are compared with those obtained by the Hewlett-Packard mass selective detector (MSD), another low-cost mass spectrometric detector and operating only in the EI mode. For the derivatives of the reference standard of NA the ITD has in the EI mode the same limit of detection, in the range of 0.5 to l n g injected on the column, as the MSD. However, under these conditions the ITD provides more spectrometric information, because it gives full scan data. Moreover, with the same or even improved limits of detection the ITD can operate in the PICI mode. On the other hand, for the analysis of NA isolated from urine samples, the performance of the MSD was better than that of the ITD. The ion trapping technique is probably limited when the chemical background is high.
Norandrosterone (NA) is the main urinary metabolite of nortestosterone (NT),’ which is an anabolic steroid extensively abused in sport. In 1987, the laboratories accredited to the International Olympic Committee (IOC) found 52% urine samples positive for anabolic steroids. NA was found in 50% of these cases. For 1988 these figures were, respectively, 79% and 38Y0.~This number of positive cases not only reflects its popularity among athletes, but also results from its pharmacodynamic properties. NT is administered intramuscularly as an ester and is released gradually into the bloodstream. As the metabolic clearance of NT is also relatively slow, the metabolites of NT can be found in the urine even several months after administration. The chance of being found positive is therefore relatively high albeit that the concentrations of the urinary metabolites usually will be relatively low. The abuse of NT in sport can therefore be considered as an important problem in doping-control analysis, and demands very sensitive and specific methodology. For doping control, the IOC requires identification by mass spectrometry, as other techniques such as immuno-assays3 are either not sensitive or specific enough. The availability of low-cost mass spectrometric detectors such as the ion trap detector (ITD, Finnigan-MAT, San Jose, CA, USA) and the mass selective detector (MSD, Hewlett-Packard, Palo Alto, CA, USA) makes it possible to fulfil this requirement. Each detector is based on a totally different massanalyzer principle. The MSD is a conventional quadrupole mass analyzer and can operate only in the electron ionization (EI) mode. The ITD is a mass analyzer based on ion-trapping techniques, and can operate in the E I mode as well as in the positive-ion chemical ionization (PICI) mode. Some EI mass spectrometric * Author to whom correspondence should be addressed.
characterizations of methoxime-trimethylsilyl (MOTMS), trimethylsilyl-enol trimethylsilyl (TMS-enol TMS) and trimethylsilyl (TMS) ether derivatives of N A have been r e p ~ r t e d . ~The , ’ TMS-enol TMS ether derivative in particular has proven its applicability in routine doping control. This study describes the EI and PICI mass spectra of TMS-enol TMS ether, MO-TMS ether and pentafluoropropionic (PFP) ester derivatives of N A using the ITD. The E I and PICI detection limits are compared with those obtained with the MSD in the EI mode. EXPERIMENTAL Materials and samples Sephadex LH-20 was obtained from Pharmacia (Woerden, The Netherlands). N,O-Bistrimethylsilyltrifluoroacetamide (BSTFA) , trimethylchlorosilane (TMCS) , and pentafluoropropionic anhydride (PFPA) were purchased from Pierce (Oud-Beyerland, The N-Methyl-N-trimethylsilyltrifluoroNetherlands). acetamide (MSTFA) was from Macherey-Nagel (Duren, FRG) and methoxylamine hydrochloride (grade I), dithioerythritol (DTE) and trimethyliodosilane (TMIS) were from Sigma (St. Louis, MO, USA). A stock solution of norandrosterone at a concentration of 100ppm in methanol was kept at 20°C. All other solvents and reagents were of analytical grade. Urine samples containing NT metabolites were obtained from routine doping-control analysis as performed in our laboratory. Sample preparation Working standard solutions of the reference standard NA were made from the stock solution and volumes of this stock solution were evaporated to dryness under
0951-4198/90/0181-0185$5.00
0
John Wiley & Sons Limited. 1990
RAPID COMMUNICATIONS IN MASS SPECTROMETRY, VOL. 4. NO. 6. 1990 181
NORANDROSTERONE ASSAY WITH LOW-COST DETECTORS
nitrogen at 55 "C. Urine samples were prepared either by XAD-2 extraction or by enzymatic hydrolysis and diethylether extraction.6 The residues containing the steroids were dried under reduced pressure over phosphorus pentoxide and potassium hydroxide for at least 1h. TMS-enol T M S derivatization. A derivatization solution was prepared of MSTFA, TMIS (1000:2, v/v) and 2 mg/mL DTE.' Of this reagent solution, 50 pL were added to the dry residue containing the steroids. The mixture was heated for 30 min at 60 "C. MO- T M S deriuatization. A derivatization solution was prepared of methoxylamine hydrochloride in dry pyridine (8% w/v) and 100yL was added to the residue containing the steroids.' The mixture was heated for 30min at 60°C. The solvent was removed under a stream of nitrogen and 50 pL of a second derivatization mixture consisting of BSTFA, TMCS and TSIM (1000:50:20, v/v) was added. This mixture was heated for 15 min at 60 "C. Excess of derivatization reagents was removed by filtration through a column (20X5mm) of Sephadex LH-20 slurry packed in a Pasteur pipette using chloroform + hexane (1:1) as eluant. The steroid derivatives were eluted in the first 2 m L of eluant and the solvent was removed under a gentle stream of nitrogen. The residue was dissolved in 50 pL isooctane for analysis by gas chromatography/ mass spectrometry (GUMS). PFP deriuatization. To the residue containing the steroids was added 100 pL of PFPA and the mixture was heated for 30 minutes at 60 "C. Excess of derivatization reagent was removed under nitrogen and the residue was dissolved in 50 yL isooctane for G U M S analysis. GCIITD-MS analysis. The analysis was carried out 6n a Finnigan MAT ITD 800 with CI option. The open-split interface was used as supplied by Finnigan MAT. The ITD was controlled by an AT-type personal computer (Bremen, FRG). The ITD was coupled to a Carlo Erba HRGC 5160 Mega series gas chromatograph (Carlo Erba, Milan, Italy) with an on-column injector. Automatic cooling of this injector was performed by the O C 516 controller. The G C was equipped with a DB-1 fused silica capillary column (J&W, Folsom, CA, USA), 3 0 m x 0 . 3 2 m m I D 0.25 pm film thickness. A deactivated fused silica column, 40 cm X 0.53 mm I D was used as a retention gap to protect the analytical column from contamination. Temperature program: the initial temperature was maintained at 80 "C for l m i n , ramped to 180°C at 40"C/min, and then to 280°C at 10"C/min and maintained there for 20min. The temperatures of the ion-trap manifold, interface and open-split were 230 "C, 265 "C and 260 "C, respectively. EI spectra were measured at 70eV and at an electron multiplier voltage 150 V above the value of the automatic tuning program. Normally the automatic gain control (AGC) was on and the B value was set at 6000. If specified, the AGC was off and the B value set at 25 000. PICI spectra were run with all parameters set according to the manufacturer's manual using isobutane as reagent gas. G C I M S D analysis. The analysis was carried out on a 5790 gas chromatograph (Hewlett-Packard) coupled to a 5970 MSD. The G U M S system was controlled by a 9816 workstation (Hewlett-Packard) with extra 20 Mb fixed disk drive (HP 9153B). The chromatograph was equipped with an HP-1 fused capillary column 18 m x
0.20 mm ID, 0.33 ym film thickness. Temperature program: the initial temperature was maintained at 180 "C for 1min, ramped to 220°C at 10 "C/min, then to 231 "C at 0.5 "Clmin, and finally to 280 "C at 20 "C/min, then maintained there for 10 min. The temperatures of the injection port and the transfer line were 250 "C and 270 "C, respectively. Helium was used as a carrier gas at a flow rate of 1mL/min. A sample volume of 1 pL was injected in the splitless mode and the split valve was opened 30 s after the injection. The MSD was operated in the EI mode at 70 eV. The instrument was autotuned according to the manufacturer's instructions. RESULTS AND DISCUSSION Ion-trap mass spectra In Table 1 the E I and PICI mass spectra are summarized for the GC/ITD analysis of the different NA derivatives. Similar to the fragmentation in nonderivatized steroids,q the losses of a methyl radical, a water molecule and a combination of the two are observed. As well as these losses, the characteristic losses of trimethylsilanol (TMSOH), pentafluoropropionic acid molecules and methoxy radicals are observed, respectively, for the different derivatives under both ionization conditions. The loss of 44 mass units in the fragmentation of the PFP ester in the EI mode could be explained by the loss of C 0 2 , involving a rearrangement in the PFP group. Isotope labelling studies are needed to prove this statement. With chemical ionization, abundant molecular ions and less fragmentation are expected, especially with the use of isobutane as a reagent gas, which has a relatively high proton affinity."' Using chemical ionization on the ITD for those N A derivatives some changes in the fragmentation patterns are observed, although not always as expected. For instance, in the PICI mass spectrum of the TMS-enol TMS ether derivative the [ M + H ] + ion is prominent, but the base peak is formed by the loss of TMSOH, a process that apparently is more favourable under CI conditions. The fragmentation mechanism of this loss involves a hydrogen abstraction," which could explain its high abundance under CI conditions, where [M H]+ ions are formed. The other prominent ions found under E I conditions are also observed under CI conditions, such as the ions at m/z 420 and 405. The PICI mass spectrum of the MO-TMS ether derivative does show a base peak at the molecular ion, but because of the presence of a TMS group, the loss of a TMSOH molecule is also prominent. The peak at mlz256, which forms the base peak under El conditions, is also observed in the PICI mass spectrum with a high abundance. The PICI mass spectrum of the PFP ester derivative shows a simple fragmentation pattern compared to the E I spectra. The loss of a water molecule, which originates from the nonderivatized keto group, is the most important fragmentation. This loss involves a hydrogen abstraction too, which might explain its high occurrence under the CI conditions.
+
The limit of detection in the EI mode The limit of detection is the quantity that can be detected with reasonable certainty. With a 99.6% confidence level the limit corresponds to the mean of the
182 RAPID COMMUNICATIONS IN MASS SPECTROMETRY, VOL. 4, NO 6, 1990
0
John Wiley 61 Sons Limited. 1990
NORANDROSTERONE ASSAY WITH LOW-COST DETECTORS
Table 1. Partial mass spectrometric data of derivatives of norandrosterone as obtained by the ion trap detector TMS-enol TMS ether
MO-TMS ether
Ionization mode
420
EI
420 (25%) 405 (100%) 315 (44%) 255 (30%)
M" [M- CH,]' [M-CH,-TMSOH]+ [M - CH, - 2TMSOH]+
CI
421 (50%) 420 (55%) 405 (40%) 331 (100%) 241 (83%)
[M+H]+ M'' [M-CH,]' [M+H-TMSOH]' [M + H - 2TMS0HIi
EJ
362 (3%) 346 (48%) 256 (100%)
[M-CH,]' [M- OCH,]' [M - OCH, - TMSOH]'
CI
378 (100%) 347 (5%) 288 (26%) 256 (79%)
[M HI+ [M+H-OCH,]" [M H - TMSOH]' [M - OCH, - TMSOH]'
EI
422 (33%) 404 (56%) 389 (11%) 378 (100%) 363 (25%) 240 (39%)
M'' [M - HzO]+' [M-HZO-CH,]' [M - 441' [M-CH3-441i [M - H20 - PFPOH]'
CI
423 (5%) 405 (100%) 241 (19%)
[M+H]' [M + H - HZO]' [M + H - HZO - PFPOH]'
377
422
PFP ester
a
Ions and possible
mlra
Wt.
Mol
Dcrivative
(abundance)
fragmcntation
+ +
Normalized on base peak.
blank measures plus three times the standard deviation of the blank measures.'* The limits with this certainty for the reference standard of NA as determined with the ITD and the MSD in the EI mode are given in Fig. 1. For the ITD the detection limits for the derivatives studied, using the full scan mode, range between 0.5 and 1ng injected on the column. The AGC option in the ITD software provides automatic control of the ionization times. It was introduced to limit the total number of ions formed in order to prevent saturation of the ion trap. The loss of mass resolution at high concentrations of the analyte13 is then kept to a minimum. A t low concentrations of the analyte, as in our case, it is not necessary to limit the ionization time as the ion trap will not be saturated with ions of the analyte. Therefore the detection limit can be improved by switching off the norandrosterone derivative TMS e n o l ~TMS ether ~
mass analyzer MSD ITD
MO TMS ether ~
MSD
ITD
scan, , conditions
AGC and fixing the ionization time at the maximum value of 25 msec. In this way the detection limits for all derivatives studied were improved by a factor of two. For the MSD the detection limits range from 5 to 20 ng injected on the column, depending on the kind of derivative. The TMS-enol TMS ether derivative gives the best results, whereas the limit for the PFP ester derivative is four times greater. By using the selected-ion monitoring (SIM) mode the limits of detection as determined by the MSD improve to the same level as for the ITD in the full scan mode. However, by using SIM, less mass spectrometric information is obtained. Using the SIM mode for the ITD does not improve the limits further, in contrast to the MSD. This is inherent in the ion-trapping principle on which the ITD is based. All ions are formed at the same time and trapped for a relatively long time. The detection range in ng injected on the column (logarithmic scale) 40 30 20
10
5
1.0
0.5
0.2
0.1
El (scan) El (SIM) El (AGC on) El (AGC off) CI
El (scan) El (SIM) El (AGC on) El (AGC off) CI
PFP ether
MSD ITD
El (-)
El (SIM) El (AGC on) El (AGC off) CI
Figure 1. Detection ranges for the derivatives of the reference standard of norandrosterone using the ion trap detector (ITD) or the mass selective detector (MSD) under different scan conditions.
0
John Wiley & Sons Limited. 1990
RAPID COMMUNICATIONS IN MASS SPECTROMETRY, VOL. 4. NO. 6, 1990 183
NORANDROSTERONE ASSAY WITH LOW-COST DETECTORS
number of ions formed and therefore the sensitivity, is mainly determined by the ionization time. The ions are detected by ejection out of the trap, and whether a full spectrum or a limited mass range (SIM) is measured does not influence the detection limit very much. So both detectors reach the same range of detection limits in the E I mode for the derivatives of NA studied here. More mass spectrometric information is obtained with the ITD because it can maintain its performance in the full scan mode while the MSD is limited to the SIM mode.
47% 421
A
I
The analysis of urine samples The performance of the ITD and of the MSD was also tested using real urine samples containing the NA-glucuronide. NA was isolated from the urine according to the method routinely used for doping control in sport. Again the ITD could provide full scan data at low levels where the conventional MSD had to use SIM (Fig. 2). The possibility of using CI to confirm positive cases is advantageous too. However, the detection limits for the ITD using real urine samples are considerably higher than in the standard references. For this reason, problems occur with urine samples of athletes with very low traces of NA as seen 3 to 4 months after administration. Concentrations of 1-5 ng NA on column, corresponding to 10-50 ng/mL urine, could not be detected using the ITD. Even with the PFP ester derivatives using the PICI mode no traces of NA could be found. These concentrations could still be measured on the MSD albeit using the SIM mode. The reason is probably that the ion-trapping technique is limited when the chemical background level is high. First, the AGC will decrease the ionization time, because of the presence of the background ions, and will therefore also decrease the formation of the analyte ions. Second, the collection of ions will be influenced by the background ions. In the ITD, the ions are trapped in a field formed by two end caps and a ring electrode. By ramping the voltage applied to the electrodes, stored ions will become unstable and will leave the trapping field in the direction of the end electrodes. The lower end-cap electrode is perforated in order to let these ions pass and be detected by an electron multiplier. Background ions will mask the field of the electrodes, and the ejection will be less efficient. Therefore, background ions will limit the number of analyte ions that can be formed and can be ejected. This will have a disadvantageous effect on the detection limit of the trap.
J
I
I
I
I
I
I
I
I
I
56%
-
100%
Internal standard
259-
(a)
The limit of detection in the PICI mode The limits of detection in the PIC1 mode are also given in Fig. 1. The MSD can not operate in the CI mode, which is a great disadvantage. The detection limit for the TMS-enol TMS ether derivative in the PICI mode is comparable to that in the EI mode. This is not surprising considering the relatively high degree of fragmentation observed. For the MO-TMS derivative the detection limit improves by a factor of two. The best results are given by the PFP ester derivative. The limit improves by a factor of twenty. This improvement in detection limit is obtained because of the reduced fragmentation of the PFP derivative under CI conditions.
Norandrosterone
-
760 12:41
780 13:OI
800 13:21
820 13:41
840 14.01
860 1421
mh
Figure 2. Gas chromatographic/mass spectrometric data of the trimethylsilyl-enol trimethylsilyl ether derivative of norandrosterone isolated from a urine sample and analyzed by the ion-trap detector in the positive-ion chemical ionization mode: (a) ion chromatograms at mlz 421 and 331; (b) full-scan mass spectrum.
These problems do not occur in a quadrupole mass filter like the MSD where ionization and mass separation are separated.
Conclusion Anabolic steroid doping control requires sensitive and specific methodology. The ITD is a low-cost mass spectrometric detector which in the El mode has the same detection limits as the conventional quadrupole MSD. However, the ITD will provide more mass spectrometric information because it gives full scan data. Also the use of CI can increase the specificity of detection. Considering the implications of a positive doping sample, the analytical chemist would rather confirm the presence of a steroid with full scan data containing abundant molecular ions. In that respect, the ITD can satisfy these needs better than a conventional quadrupole detector like the MSD. In extreme cases, however, with concentrations at the low ng range, the MSD can still detect NA using the SIM mode when ITD detection is no longer possible. In conclusion, one can say that ITD provides the most specific data needed for confirmatory analyses in doping control, but the MSD has the lowest absolute detection limit in real samples. The recently introduced ITS-40 ion trap (Finnigan MAT) uses an extra resonance potential on the end caps. As a result, the ionization time can be increased further without reducing the mass resolution. More ions are ejected, and the sensitivity is improved. The sensitivities specified are 10-100 times better than for the ITD 800. Future experiments will determine whether this ITS-40 can obtain full scan data even for today's limiting cases in steroid doping control.
184 RAPID COMMUNICATIONS IN MASS SPECTROMETRY, VOL. 4, NO. 6. I990
0
John Wiley & Sons Limited, 1990
NORANDROSTERONE ASSAY WITH LOW-COST DETECTORS
Acknowledgements The reference standard stock solution of NA in methanol (100 ppm) DSH, (FRG). was a generous gift of Prof. Dr. M. We also thank Mr. P. w . van Dorp Van Vliet for his assistance in preparing the illustrations.
REFERENCES 1 . R. Masst, C. Lalibertt, L. Tremblay and R. Dugal, Biomed. Mass Spectrom. 12, 115 (1985). 2. IOC communications LAB/68/88/SJG and FICILAB112/ 89/SJG. 3. A. T. Kicmac and R. V . Brooks, J . Pharm. Biomed. Anal. 6,413 (1988). 4. R . Masst, C. LalibertC and L. Tremblay, J . Chromatogr. 339, 11 (1985). 5. D. A. Cowan and G . Woffendin, Finnegan MAT, Spectra 11 4 (1988).
0
John Wtley & Sons Limtted. 1990
6. M. Donike, J. Zimmermann, K-R. Barwald, W. Schanzer, V. Christ, K. Klostermann and G. Opfermann, Deutsch. 2. Sportmedizin 35, 14 (1984). 7. M. Donike and J . Zimmermann, J . Chromatogr. 202,483 (1980). 8. E. Houghton, p. Taele, M, c, Dumasia and J. K, Biomed. Mass Spectrom. 9, 459 (1982). 9. L. Tokts, R. T. LaLonPe and C. Djerassi, J . Org. Chem. 32, 1012 (1967). 10. F. W. Karasek and R. E. Clement; Basic Gas ChrornatographyMass Spectrometry, p. 45. Elsevier, Amsterdam (1988). 11. J. Diekman and C. Djerassi, J . Org. Chem. 32, 1005 (1967). 12, Nomenclature, Symbols, units and Their Usage in Spectrochemical Analysis-I1 Data interpretation, International Union of Pure and Applied Chemistry, Anal. Chem. 48, 2294 (1976). 13. J . W. Eichelberger and W. L. Budde, Biomed. Enuiron. Mass Spectrom. 14, 357 (1987). Received 23 March 1990, accepted 17 April 1990
RAPID COMMUNICATIONS IN MASS SPECTROMETRY, VOL 4, NO. 6. 1990 185
