Anal. Chem. 2006, 78, 4624-4629
LC-MS/MS Screen for Penitrem A and Roquefortine C in Serum and Urine Samples Elizabeth R. Tor,* Birgit Puschner, Michael S. Filigenzi, Asheesh K. Tiwary, and Robert H. Poppenga
California Animal Health and Food Safety Laboratory System, Toxicology Laboratory, University of California, Davis, California 95616
A rapid LC-MS/MS method, using a triple quadrupole/ linear ion trap mass spectrometer, was developed for determination of penitrem A and roquefortine C in serum and urine samples. Penitrem A and roquefortine C were extracted from samples with methylene chloride. The extracts were injected onto a liquid chromatograph coupled with a hybrid triple quadrupole/linear ion trap mass spectrometer. Seven replicate fortifications of serum at 0.001 µg/g (1 ppb) each of penitrem A and roquefortine C gave average recoveries of 90% with 10% CV (relative standard deviation) and 97% with 3% CV, respectively. Seven replicate fortifications of urine at 0.001 µg/g (1 ppb) each of penitrem A and roquefortine C gave average recoveries of 98% with 12% CV and 100% with 6% CV, respectively. This is the first report of a positive mass spectrometric identification and quantitation of both compounds in urine and serum samples from dog intoxication cases. Penitrem A [12627-35-9], C37H44ClNO6, and roquefortine C [58735-64-1], C22H23N5O2, Figure 1, are tremorgenic mycotoxins produced most commonly by the fungi from the genera Penicillium, Aspergillus, and Claviceps.1,2 Intoxication with these mycotoxins has been documented in humans3 and animals including dogs4-10 and cattle.11 Clinical signs of tremorgenic mycotoxin poisoning are well documented and include severe generalized muscle tremors and incoordination,3-5 seizures3-6, muscle fasciculations,6 and generalized convulsions,9,10 followed by death in many cases. In severe intoxications increased muscle activity can ultimately lead to exhaustion, metabolic changes, rhabdomyolysis, * Corresponding author. Phone: 530-752-6322. Fax: 530-752-3361. E-mail:
[email protected]. (1) Puschner, B. Vet. Clin. Small Anim. 2002, 32, 409-419. (2) Bu ¨ gner, J.; Westphal, G.; Mo¨nnich, A.; Hinnendahl, B.; Hallier, E.; Mu ¨ ller, M. Toxicology 2005, 202, 199-211. (3) Lewis, P. R.; Donoghue, M. B.; Hocking, A. D.; Cook, L.; Granger, L. V. Med. J. Aust. 2005, 182, 582-584. (4) Boysen, S, R.; Rozanski, E. A.; Chan, D. L.; Grobe, T. L.; Fallon, M. J.; Rush, J. E. J. Vet. Med. Assoc. 2002, 221, 1441-1444. (5) Walter, S. L. Can. Vet. J. 2002, 43, 372-374. (6) Young, K. L.; Villar, D.; Carson, T. L.; Imerman, P. M.; Moore, R. A.; Bottoff, M. R. J. Vet. Med. Assoc. 2003, 222, 52-53. (7) Lowes, N. R.; Smith, R. A.; Beck, B. E. Can. Vet. J. 1992, 33, 535-537. (8) Hocking, A. D.; Holds, K.; Tobin, N. F. Aust. Vet. J. 1988, 65, 82-85. (9) Puls, R.; Ladyman, E. Can. Vet. J. 1988, 29, 569. (10) Richard, J. L.; Bacchetti, R. L.; Arp, L. H. Mycopathologia 1981, 76, 55-58. (11) Cysewski, S. J.; Baetz, A. L.; Pier, A. C. Am. J. Vet. Res. 1975, 36, 53-58.
4624 Analytical Chemistry, Vol. 78, No. 13, July 1, 2006
Figure 1. Penitrem A [12627-35-9] C37H44ClNO6 (top) and roquefortine C [58735-64-1] C22H23N5O2 (bottom)stremorgenic mycotoxins produced most commonly by the fungus Penicillium crustosum.
and death. In most cases in dogs, vomiting occurs prior to the onset of neurological signs. Analytical methods for the diagnosis of tremorgenic mycotoxin poisoning are currently limited to the analysis of penitrem A and roquefortine C in source material (mold cultures, food) or GI contents. These include TLC,8,12 GC/MS,12,13 HPLC,14 and LC/ MS techniques including triple quadrupole,15 ion trap,16-18 and time-of-flight mass spectrometry.19 In clinical cases where penitrem A or roquefortine C poisoning is suspected a rapid, specific, and accurate method is needed to detect them at toxicologically significant, low ppb concentrations in urine and serum from live (12) Braselton, W. E.; Johnson, M. J. Vet. Diagn. Invest. 2003, 1, 42-45. (13) Braselton, W. E.; Rumler, P. C. J. Vet. Diagn. Invest. 1996, 8, 515-518. (14) Noroozian, E.; Lagerwerf, F.; Lingeman, H.; Brinkman, U. A.; Kerkhoff, M. A. J. Pharm. Biomed. Anal. 1999, 20, 611-619. (15) Sumarah, M. W.; Miller, J. D.; Blackwell, B. A. Mycopathologia 2005, 159, 571-577. (16) Rundberget, T.; Skaar, I.; O’Brien, O.; Flaoyen, A. Mycopathologia 2004, 157, 349-357. (17) Naude, T. W.; O’Brien, O. M.; Rundberget, T.; McGregor, A. D.; Roux, C.; Flaoyen, A. J. S. Afr. Vet. Assoc. 2002, 73, 211-215. (18) Rundberget, T.; Wilkins, A. L. J. Chromatogr., A 2002, 964, 189-197. (19) Nielsen, K. F.; Smedsgaard, J. J. Chromatogr., A 2003, 1002, 111-136. 10.1021/ac0601617 CCC: $33.50
© 2006 American Chemical Society Published on Web 05/09/2006
Figure 2. Positive ion ESI-MS/MS product ion spectra of penitrem A (A) and roquefortine C (B) in pure analytical standard prepared in methanol. The MS/MS spectra were obtained by fragmenting the protonated molecular ion of m/z 634 (penitrem A) and m/z 390 (roquefortine C). Roquefortine C and penitrem A eluted from the HPLC column at 2.9 and 6.0 min, respectively.
animals. To the best of our knowledge there are no reports of such analytical methodologies. This investigation was undertaken to develop a rapid, sensitive, and highly specific screen for penitrem A and roquefortine C in serum and urine samples using an LC-tandem triple quadrupole/ linear ion trap mass spectrometer. This procedure is especially suited to the veterinary diagnostic or public health laboratory situations for which rapid antemortem diagnosis of exposure to tremorgenic mycotoxins is desirable. MATERIALS AND METHODS Reagents. Water, methanol, and formic acid were of HPLC grade (Fisher Scientific) and methylene chloride was of
Optima grade (Fisher Scientific). All HPLC running solvents were filtered through 0.45 µm, nylon filters (Gelman Sciences, Ann Arbor, MI). Preparation of Standard Solutions. Penitrem A and roquefortine C standards (95% and 90% purity, respectively) were purchased from Sigma Chemical Co. (St. Louis, MO). A stock solution of 10 µg/mL each of penitrem A and roquefortine C was made in methanol. Subsequent dilutions of the standard were made daily in methanol from the stock solution. Five-point calibration curves of roquefortine C/penitrem A mix in serum or urine extracts were prepared at 0.0005/0.005, 0.002/0.02, 0.01/ 0.1, 0.05/0.5, and 0.1/1 µg/mL by adding aliquots of roquefortine Analytical Chemistry, Vol. 78, No. 13, July 1, 2006
4625
Figure 3. Typical LC-MS/MS chromatograms (m/z 558 for penitrem A and m/z 322 for roquefortine C) of a negative control canine serum.
C and penitrem A standards to dry, negative control canine serum or urine extracts, prepared as described below. Sample Preparation. One gram of sample was weighed into a 15 mL, screw-cap, disposable tube, the pH was adjusted to above 10 using ammonium hydroxide, 10 mL of methylene chloride was added, and the sample was rotated at medium speed for 30 min. The mixture was centrifuged at 2000 rpm (260g) for 5 min using an IEC Centra-7 centrifuge (International Equipment Co., Needham, MA). An aliquot (9 mL) of the methylene chloride extract was evaporated to dryness, using an N-Evap nitrogen evaporator (Analytical Evaporator, Organomation Assoc. Inc., Berlin, MA) set at 60 °C. The extract was redissolved in 180 µL of methanol/ water (1:1, v/v) and filtered through a 0.45 µm HPLC filter (Millipore Corp., Milford, MA) into a small volume autosampler vial. All control and fortified samples were prepared in the same manner. LC-MS/MS Analysis. An Agilent model 1100 (binary) highperformance liquid chromatograph coupled with a hybrid triple quadrupole/linear ion trap mass spectrometer, model 4000 Q TRAP (Applied Biosystems/MDS SCIEX, Concord, Canada) was 4626 Analytical Chemistry, Vol. 78, No. 13, July 1, 2006
used in all analyses. The analytical column was a 100 mm × 4.6 mm Chromolith Performance RP-18e with 10 mm × 4.6 mm RP18e guard column (Merck KGaA, Darmstadt, Germany). The injection volume was 10 µL. The mobile phase consisted of (A) 0.1% formic acid in water (v/v); (B) 0.1% formic acid in methanol (v/v) at a flow rate of 1200 µL/min under a linear gradient of 60% B to 90% B over 10 min. The retention times for roquefortine C and penitrem A were 2.9 and 6.0 min., respectively. MS data for this screen were acquired in the positive ion electrospray ionization (ESI) mode, using the following TurboIonSpray source conditions: temperature ) 500 °C, curtain gas ) 40 (arbitrary units), GS1 ) 70, GS2 ) 60, CAD gas pressure high, ion spray voltage ) 5500. The MS/MS acquisition program consisted of a single period with two experiments using full scan MS/MS in enhanced product ion (EPI) mode. In this mode of operation, the third quadrupole stage functions as a linear ion trap, providing a high level of sensitivity with full scan data. In EPI experiment 1, roquefortine C was analyzed using the following parameters: precursor ion ) m/z 390, collision energy (CE) ) 35, declustering potential (DP) ) 63, collision energy spread (CES) ) 5.0, dynamic
Figure 4. (A) LC-MS/MS chromatogram (m/z 558) of control canine serum fortified with penitrem A at 0.001 µg/g (1 ppb), S/N ) 61. (B) LC-MS/MS chromatogram (m/z 322) of control canine serum fortified with roquefortine C at 0.001 µg/g (1 ppb), S/N ) 388.
fill time on, Q3 entry barrier ) 8 V, product ion scan range ) m/z 100-400. In EPI experiment 2, penitrem A was analyzed using the following parameters: precursor ion ) m/z 634, CE, DP, and CES parameters as per those of roquefortine C, product ion scan range ) m/z 400-650. These parameters were optimized while separately infusing 10 µg/mL roquefortine C and 10 µg/mL penitrem A standards at 10 µL/min into the mobile phase as above. Ten microliters of standards in matching matrix or sample extracts was injected into the system described above. Each set of samples contained a reagent blank, control, and fortified samples. Quantification was by comparison with a five-point calibration curve using external standards in matching matrix and nonweighted linear regression using the Analyst version 1.4 software.
Method Validation. The method was validated by analyzing control canine serum or urine samples (n ) 7 each) fortified with roquefortine C and penitrem A at the 0.001 µg/g (1ppb). The fortifications were prepared by adding 10 µL of 0.1 µg/mL roquefortine C/penitrem A standard mix in methanol to 1 g of negative, control canine urine or serum samples and analyzing them using the method described above. The method was also tested by routine analyses of samples from veterinary diagnostic dog intoxication cases. RESULTS AND DISCUSSION Protonation was the main ionization process observed in the positive ESI mass spectrum of penitrem A and roquefortine C, with [M + H]+ (m/z 634 and 390, respectively) as base peaks. Analytical Chemistry, Vol. 78, No. 13, July 1, 2006
4627
Figure 5. (A) LC-MS/MS chromatogram (m/z 558) of penitrem A; (B) LC-MS/MS chromatogram (m/z 322) of roquefortine C in a serum sample from a dog with a history of muscle tremors and seizures. The concentration of penitrem A and roquefortine C in the serum was 0.0042 µg/g (4.2 ppb) and 0.0133 µg/g (13.3 ppb), respectively. The retention times of roquefortine C and penitrem A were 2.9 and 6.0 min, respectively.
Figure 2 shows typical LC-MS/MS product ion spectra of (A) penitrem A and (B) roquefortine C standards in methanol. The MS/MS spectra were obtained by fragmenting the protonated molecular ions of m/z 634 (penitrem A) and m/z 390 (roquefortine C). The m/z 558 and 322 ions were used for quantitation of penitrem A and roquefortine C, respectively. The spectra observed under the EPI scan conditions described and ions used for quantitation were similar to those previously published.17,18 The sensitivity of the instrument was approximately 10 times greater for roquefortine C than for penitrem A. Fifty picograms of penitrem A standard in methanol injected on column gave S/N of 115. Five picograms of roquefortine C standard in methanol on column gave S/N of 139. 4628 Analytical Chemistry, Vol. 78, No. 13, July 1, 2006
When ESI is used for quantitative analysis, the effect of ionization suppression by matrix components must be considered. Quantitation of both compounds in urine and serum using electrospray ionization was affected by ion suppression. Therefore, it was essential to perform quantitation using standards in matrixes that matched those of the samples. The standard curves in matrix followed linear regression with r2 values typically in the 0.999 range. Seven replicate fortifications of serum at 0.001 µg/g (1 ppb) each of penitrem A and roquefortine C gave average recoveries of 90% with 10% CV (relative standard deviation) and 97% with 3% CV, respectively. Seven replicate fortifications of urine at 0.001 µg/g (1 ppb) each of penitrem A and roquefortine C gave average recoveries of 98% with 12% CV and 100% with 6% CV, respectively.
The use of tandem triple quadrupole/linear ion trap mass spectrometry produced clean chromatograms of penitrem A and roquefortine C with minimal background contribution from urine or serum matrixes, Figure 3. LC-MS/MS ion chromatograms of penitrem A and roquefortine C in control negative canine serum fortified with penitrem A and roquefortine C at the 0.001 µg/g (1 ppb) are shown in Figure 4 to demonstrate instrument sensitivity and selectivity. The described method was applied to the routine analysis of serum and urine samples from veterinary diagnostic samples. Figure 5, parts A and B, shows chromatograms of a canine serum extract that tested positive for both penitrem A at 0.0042 µg/g (4.2 ppb) and roquefortine C at 0.0133 µg/g (13.3 ppb). The serum and urine samples were collected from a dog that had a history of ingesting moldy walnuts and was presented to an emergency clinic with tremors and seizures. Both penitrem A and roquefortine C were clearly identified in the serum sample of the exposed dog with spectra matching those of the analytical standards. There is no data on the correlation of serum penitrem A and/ or roquefortine C concentrations and the development of clinical signs. Serum concentrations vary depending on the time interval between exposure and sample collection. However, based on
preliminary data, a detection limit of 1 ppb for penitrem A and roquefortine provides an adequate diagnostic tool for case investigations of suspect exposures to these mycotoxins. It is important to analyze serum for both mycotoxins, penitrem A and roquefortine C. Although penitrem A is expected to cause the tremorgenic effects, it is more likely to detect roquefortine C, which is usually produced in higher concentrations and more slowly metabolized than penitrem A. This method offers a significant improvement in diagnosis of intoxication with tremorgenic mycotoxins in that it provides a rapid and unequivocal determination of penitrem A and roquefortine C in serum and urine and it is validated at very low, diagnostically relevant concentrations. ACKNOWLEDGMENT The authors thank Dr. Jennifer Koon for providing diagnostic material.
Received for review January 23, 2006. Accepted April 6, 2006. AC0601617
Analytical Chemistry, Vol. 78, No. 13, July 1, 2006
4629
