A Curiosity of Accurate Mass Analysis of Sulfadimethoxine in Pond

Jan 27, 2004 - Department of Analytical Chemistry, University of Almeria, Carreterra ... Liquid chromatography with time-of-flight mass spectrom- ... ...
0 downloads 0 Views 157KB Size
Download PDF
Anal. Chem. 2004, 76, 1228-1235

Intramolecular Isobaric Fragmentation: A Curiosity of Accurate Mass Analysis of Sulfadimethoxine in Pond Water E. Michael Thurman,*,† Imma Ferrer,† Mark Benotti,‡ and Curt E. Heine§

Department of Analytical Chemistry, University of Almeria, Carreterra Sacramento, Can˜ada San Urbano, 04120 Almeria, Spain, Marine Sciences Research Center, Stony Brook University, Stony Brook, New York 11794-5000, and Waters Corporation, 100 Cummings Center, Beverly, Massachusetts 01915-6101

Liquid chromatography with time-of-flight mass spectrometry (TOF-MS) and quadrupole-time-of-flight (Q-TOF) mass spectrometry/mass spectrometry (MS/MS) were used for the accurate mass analysis of sulfadimethoxine in pond water of a fish hatchery. Sulfadimethoxine is the most important sulfa antimicrobial used in aquaculture to treat bacterial disease in a wide variety of fish. Because correct identification is essential to environmental monitoring of antimicrobial pharmaceuticals, accurate mass analyses (TOF and Q-TOF-MS/MS) were compared to nominal mass measurement (quadrupole ion trap). It was known that all six members of the sulfa antimicrobial family gave a common 6-sulfanilamido ion at a nominal mass of m/z 156; thus, this ion was the focus of TOF confirmation (exact mass 156.0119 u) along with the protonated molecule (exact mass 311.0814 u). In the process of accurate mass confirmation of the 156 m/z fragment ion, a second isobaric ion (exact mass m/z 156.0773), was discovered at the same nominal mass, which was not differentiated by quadrupole ion trap. The structure was assigned as 2-4-dimethoxypyridine and is exactly the other protonated half of the sulfadimethoxine molecule. This discovery led to the subsequent use of Q-TOF-MS/MS and high-resolution identification of five other important ion fragments for the identification of sulfadimethoxine in pond water at environmental concentrations. The caveats of using low-resolution mass spectrometry without MS/MS for environmental monitoring are discussed in the light of high profile monitoring of sulfa antimicrobial pharmaceuticals in the aquatic environment. Antibiotics and antimicrobials are an important environmental concern at this time because of the potential for microbial resistance caused by widespread agricultural use of antibiotics * To whom correspondence should be addressed. Department of Hydrogeology and Analytical Chemistry, University of Almeria, 04120 Almeria, Spain. E-mail: [email protected]. † University of Almeria. ‡ Stony Brook University. § Waters Corporation.

1228 Analytical Chemistry, Vol. 76, No. 5, March 1, 2004

and antimicrobials in raising both hogs and poultry.1-5 Furthermore, correct identification of human and veterinary pharmaceuticals in natural water is an important analytical environmental issue that has been pointed out in a series of recent Letters to the Editor of Environmental Science and Technology6-8 concerning the national reconnaissance (United States) of pharmaceuticals in water, published in 2002.1 This paper was challenged analytically for the identification of human pharmaceuticals (in particular, hormones) in some water samples. The work was completed as part of a larger survey of hormones, over-the-counter medications, and antibiotics in the environment. The analytical methods were a combination of gas chromatography/mass spectrometry and liquid chromatography/mass spectrometry. As a coauthor on that paper, who also has published in Analytical Chemistry recently3 on the use of single quadrupole LC/MS analysis of antibiotics in the aquatic environment, it was deemed important to also analyze some aquatic environmental samples by accurate mass time-offlight mass spectrometry and quadrupole-time-of-flight mass spectrometry and to compare these results with another method (LC/MS/MS, quadrupole ion trap) and previously published results using single quadrupole LC/MS analysis. The purpose was for correct identification of not only the protonated molecule but also the structure of key fragment ions used for unequivocal identification of these pharmaceuticals in water samples. Thus, this paper examines the use of quadrupole ion trap, TOF, and Q-TOF methods for antimicrobial analysis and shows the increased knowledge of ion fragmentation as well as the increased confidence of correct identification when using accurate mass (1) Kolpin, D. W.; Buxton, H.; Furlong, E.; Meyer, M.; Thurman, E. M.; Zaugg, S.; Barber, L., Jr. Environ. Sci. Technol. 2002, 36, 1202-1211. (2) Hirsch, R.; Ternes, T.; Haberer, K.; Kratz, K. Sci. Total Environ. 1999, 225, 109-118. (3) Lindsey, M. E.; Meyer, M.; Thurman, E. M. Anal. Chem. 2001, 73, 46404646. (4) Thurman, E. M.; Dietze, J. E.; Scribner, E. A. Occurrence of antibiotics in water from fish hatcheries; U.S. Geological Survey Fact Sheet 120-02; U.S. Government Printing Office: Washington, DC, 2003. (5) U.S. Food and Drug Administration, Center for Veterinary Medicine, FDAapproved animal drug products-on line database system, 2003. Information available on the World Wide Web at URL http://dil.vetmed.vt.edu/. (6) Kolpin, D. W.; Furlong, E. T.; Meyer, M. T.; Thurman, E. M.; Zaugg, S. D.; Barber, L. B.; Buxton, H. T. Environ. Sci. Technol. 2002, 36, 4004. (7) Kolpin, D. W.; Furlong, E. T.; Meyer, M. T.; Thurman, E. M.; Zaugg, S. D.; Barber, L. B.; Buxton, H. T. Environ. Sci. Technol. 2002, 36, 4007-8. (8) Kolpin, D. W.; Furlong, E. T.; Meyer, M. T.; Thurman, E. M.; Zaugg, S. D.; Barber, L. B.; Buxton, H. T. Environ. Sci. Technol. 2003, 37, 1054. 10.1021/ac035094k CCC: $27.50

© 2004 American Chemical Society Published on Web 01/27/2004

analysis. Furthermore, this paper outlines the caveats of using single quadrupole analysis and low resolution mass spectrometry for monitoring sulfa antimicrobials (pharmaceuticals) in the aquatic environment. The compound chosen for this study is the sulfa antimicrobial, sulfadimethoxine, which is the most important sulfa drug used to treat fish with bacterial infections at fish hatcheries.5 It and sulfamerazine are the only two sulfa antimicrobials approved by the Federal Drug Administration for aquaculture,5 and only sulfadimethoxine is routinely used.4-5 Sulfamerazine has limited use for salmon.5 Sulfadimethoxine is supplied as a food product for fishsan example is Romet-30swhich is given to fish as an edible mixture of sulfadimethoxine and ormethoprim. Ormethoprim is an antibiotic activity agent that is also FDA-approved that enhances the effectiveness of sulfadimethoxine.5 Because sulfadimethoxine is a synthetic compound rather than natural, it is classified as an antimicrobial rather than antibiotic. Antibiotics are naturally occurring and are a subclass of antimicrobials.3 The first national reconnaissance of pharmaceuticals in surface water of the United States1 reported only 1.2% detections of sulfadimethoxine, and a more recent survey of selected U.S. fish hatcheries4 found that sulfadimethoxine and oxytetracycline (a third antimicrobial approved for aquaculture by FDA) were found with detections of 15% for sulfadimethoxine and 13% for oxytetracycline in fish hatchery ponds following correct use. In an earlier study, Hirsh et al.2,9 reported other sulfa antimicrobials in groundwater (sulfamethoxazole and sulfamethazine) at concentrations of ∼0.2 to 0.5 µg/L using liquid chromatography/mass spectrometry/mass spectrometry with multiple reaction monitoring. Similarly, Bratton et al.10 reported sulfamethoxazole in river water and sewage from Kansas (U.S.A.) as part of a larger study on the determination of pharmaceuticals in water samples. They also used LC/MS/MS triple quadrupole analysis and found concentrations in the range of 0.001-0.19 µg/L. Thus, these studies show that sulfa antimicrobials do occur in surface and groundwater and that sulfadimethoxine, specifically, is an antimicrobial that can occur in natural water samples associated with fish hatcheries. Previous methods for sulfadimethoxine and the family of sulfa antimicrobials in water include our work in 2001,3,4 which used single quadrupole selected ion monitoring with two selected ions for six different sulfa antibiotics, and the work of Hirsch et al.,9 which used liquid chromatography/mass spectrometry/mass spectrometry (LC/MS/MS), for the analysis of sulfadimethoxine and sulfamethazine using multiple reaction monitoring. Likewise, the work of Bratton et al.10 used LC/MS/MS with triple quadrupole using multiple reaction monitoring for the sulfa antimicrobial, sulfamethoxazole. Finally, Benotti et al.11 compared the analysis of three types of LC/MS instrumentation: single quadrupole, triple (9) Hirsch, R.; Ternes, T. A.; Haberer, K.; Mehlich, A.; Ballwanz, F.; Kratz, K. L. J. Chromatogr., A 1998, 815, 213-223. (10) Bratton, K. D.; Lillquist, A. S.; Williams, T. D.; Lunte, C. E. In Liquid Chromatography/Mass Spectrometry, MS/MS and Time-of-Flight MS: Analysis of Emerging Contaminants; Ferrer, I., Thurman, E. M., Eds.; American Chemical Society, Oxford University Press: New York, 2003; Chapter 12, pp 188-206. (11) Benotti, M. J.; Ferguson, P. L.; Rieger, R. A.; Iden, C. R.; Heine, C. E.; Brownawell, B. J. In Liquid Chromatography/Mass Spectrometry, MS/MS and Time-of-Flight MS: Analysis of Emerging Contaminants; Ferrer, I.; Thurman, E. M., Eds.; American Chemical Society, Oxford University Press: New York, 2003; Chapter 7, pp 109-127.

Figure 1. Milford Hatchery, site of water sample collection.4

quadrupole, and time-of-flight mass spectrometry for the analysis of a suite of pharmaceuticals in water samples. They analyzed one sulfa antimicrobial, sulfamethoxazole, by the three methods and found that a sewage effluent contained sulfamethoxazole but that interferences from nontarget compounds created an apparently artificially high background. The accurate mass analysis gave the lowest concentration by a factor of 5 when compared to single quadrupole analysis and a factor of 3 when compared to triple quadrupole analysis using MRM. These results were discussed, but no definitive conclusion was reached for the differences in concentrations by the three methods, although MRM is generally considered the “gold standard” for environmental analytical analysis.11 No example of the use of LC/Q-TOF-MS/MS for the analysis of sulfa antimicrobials in water could be found in a literature search for the analysis of sulfa antimicrobials or sulfadimethoxine or in a recent literature review of environmental time-of-flight mass spectrometry.12 Thus, our study on the use of MS/MS with accurate mass analysis, LC/MS Q-TOF, gives new insights into the fragmentation behavior of sulfa antimicrobials (i.e., sulfadimethoxine), which is applicable to other compounds in the family.1-4,10-11 Furthermore, this paper will show the discovery of two intramolecular isobaric fragment ions of sulfadimethoxine by accurate-mass determination. Five other accurate mass structural fragments of sulfadimethoxine will be shown for both the standard and a fish hatchery sample for unequivocal identification. Finally, the caveats of low resolution analysis of antimicrobials in the environment are discussed and an identification protocol for unknown pharmaceuticals in the environment is given. Hopefully, these results will prove useful for future studies of the analysis of pharmaceuticals in the aquatic environment. EXPERIMENTAL METHODS Sample Collection. Figure 1 shows the location of the sample collection sites for a monitoring study of sulfa antimicrobials used in aquaculture.4 The site at Milford, KS, was chosen for confirmation by time-of-flight mass spectrometry, both LC/TOF-MS and LC/Q-TOF-MS after a monitoring study of antimicrobials in fish (12) Ferrer, I.; Thurman, E. M. Trends Anal. Chem. 2003, 22, 750-756.

Analytical Chemistry, Vol. 76, No. 5, March 1, 2004

1229

hatcheries had detected sulfadimethoxine.4 The site is an intensive fish hatchery, which means that concrete raceways are used to raise the fish, and when fish are ill, such as suffering from any bacterial infections, they are treated in accordance with federal regulations5 with a medicated fish feed, called Romet-30. This feed contains a mixture of sulfadimethoxine and ormethoprim at concentrations of ∼50 ppm in the feed.4-5 Intensive raceways are operated under a controlled environment. Hand feeding or mechanical fish feeders meet the nutritional needs of the fish. Oxygen, ammonia, and nitrate levels are maintained by daily monitoring and the use of filters or rapid exchanges of water in the system to maintain the environmental needs of the fish. Both warm-water and cold-water fish are raised in intensive fish hatcheries; however, only channel catfish (Ictalurus) were raised at the Milford sampling ponds. Grab water samples were re-collected from two ponds, RW-21 and RW-24, in the summer of 2002, ∼3 months after the last feeding of Romet-30. Samples were filtered through glass fiber filters at 0.7 µm and stored in glass bottles in the refrigerator at 5 °C until analysis by LC/Q-TOF and LC/ion trap-MS. Samples of 100 mL were passed through an Oasis solid-phase extraction cartridge by gravity at 5 mL/min. The cartridge was eluted with 2 mL of methanol, evaporated to ∼25 µL by nitrogen evaporation, and reconstituted to a total volume of 200 µL in LC/MS buffer, which consisted of 10 mM ammonium formate adjusted to pH 3.7 with formic acid. LC/MS/MS Ion Trap. Liquid chromatography/electrospray ion trap mass spectrometry (LC/ESI-MS) in positive ionization, full-scan operation, was used to separate and identify sulfadimethoxine. The analyte was separated from matrix interferences using an HPLC (series 1100, Agilent Technologies, Palo Alto, CA) equipped with a reversed-phase C18 analytical column (Phenomenex, Torrance, CA) of 250 × 3 mm and 5-µm particle diameter. Column temperature was maintained at 25 °C. Mobile phase A was a 10 mM ammonium formate buffer, and mobile phase B consisted of acetonitrile. A linear gradient progressed from 15% B (initial conditions) to 100% B in 40 min, after which the mobile phase composition was maintained at 100% B for 5 min. The flow rate was 0.60 mL/min, and 50 µL of the SPE extract was injected. This HPLC system was connected to an ion trap mass spectrometer (Esquire LC, Bruker Daltonics, Billerica, MA). LC/TOF-MS. A Waters series 2795 HPLC was used for the LC/TOF-MS analyses. The standard and extracted water samples were eluted from a MetaChem MetaSil AQ C18 column with dimensions of 2 × 150 mm, at a flow rate of 0.5 mL/min. Mobile phases A and B were water with 0.1% formic acid and acetonitrile with 0.1% formic acid, respectively. The chromatographic method held the initial mobile phase composition (10% B) constant for 8 min, followed by a linear gradient to 100% B at 25 min. Fifteenmicorliter portions of the extracts were injected. An orthogonal acceleration TOF mass spectrometer, the LCT (Waters Corp., Manchester, U.K.), was used in positive ion mode with electrospray ionization. This instrument was equipped with a 3.6-GHz time-to-digital converter and a reflectron, which generated a resolving power of 7500 at m/z 556 (FWHM definition). Leucine enkephalin ([M + H]+ ) 556.2771 u) was added postcolumn as a lock mass to compensate for drift of the external calibration. The instrument was operated with a cone voltage of 22 V. Accurate 1230

Analytical Chemistry, Vol. 76, No. 5, March 1, 2004

mass data were processed (centroiding of continuum data with lock mass and digital deadtime correction) using Masslynx 3.5 software. Exact masses corresponding to particular elemental compositions were also calculated by Masslynx 3.5. LC/Quadrupole/Time-of-Flight Mass Spectrometry. The chromatography was carried out on a Waters HPLC model 2795 using a Metachem MetaSil AQ-C18 column of 2 × 150 mm with a flow rate of 500 µL/min. The flow was split in half to 275 µL/ min prior to the ion source. Solvent A was water, and solvent B was acetonitrile; both contained 0.1% formic acid. The gradient was 10% B for 8 min, then ramped to 100% B at 25 min. Injection volume was 3 µL for the standard and 50 µL for the fish hatchery sample. A quadrupole orthogonal acceleration time-of-flight mass spectrometer, the Q-TOF micro (Waters Corp., Manchester, U.K.), was used in positive ion mode with electrospray ionization. This instrument was equipped with a 3.6-GHz time-to-digital converter, and a reflectron, which generated a resolving power of 5900 at m/z 556 (FWHM definition). Leucine enkephalin ([M + H]+ ) 556.2771 u) was infused through the reference sprayer of a lockspray source as a lock mass to compensate for drift of the external m/z calibration in LC/MS analysis. A sample cone voltage of 33 V was used for LC/MS. An optimized sample cone voltage for sulfadimethoxine of 38 V was used in LC/MS/MS analyses. MS/MS experiments relied on residual precursor (protonated sulfadimethoxine, 311.0814 u) signal for lock mass correction of fragment ions formed through collision-induced dissociation (CID). The collision cell of the Q-TOF micro was filled with argon through a capillary tube (head pressure set to 17 psi), and a collision energy of 22 eV initiated the CID. Accurate mass data were processed (centroiding of continuum data with lock mass and digital deadtime correction) using Masslynx 3.5 software. Exact masses corresponding to particular elemental compositions were also calculated by Masslynx 3.5. The lock mass for the MS/ MS experiment was the protonated molecule of sulfadimethoxine, 311.0814 u. RESULTS AND DISCUSSION Low-Resolution Fragment Ions of Sulfadimethoxine. Figure 2 shows the low-resolution ion trap MS/MS spectrum of a sulfadimethoxine standard after liquid chromatographic separation from an ormethoprim standard (peak 1 in Figure 2). In positiveion mode, the protonated molecule of sulfadimethoxine is m/z 311 and fragments in the ion trap to three abundant ions at nominal masses of m/z 156, 218, and 245. Less abundant ions are also observed at m/z 108 and 230. Previous methods measuring sulfadimethoxine in the environment have used the protonated molecule at m/z 311 and the m/z 156 ion as the confirming ion,1,3-4 and no other fragment ions were monitored or reported.1,3-4 Furthermore, the work of Lindsey et al.3 pointed out that all of the six sulfa antimicrobials tested (sulfachloropyridazine, sulfadimethoxine, sulfamerazine, sulfamethazine, sulfamethoxazole, and sulfathiazole) contained the m/z 156 fragment ion and that this ion was a diagnostic ion of the sulfa class of antimicrobials. The structure of the m/z 156 ion, the sulfanilamido ion (resulting from a charge-transfer fragmentation) as proposed by Lindsey et al., is shown also in Figure 2. Because product ion spectra from all of the sulfa antimicrobials contain this m/z 156 ion, this structure shown in Figure 2 was not challenged. The

Figure 2. Ion trap mass spectrum of sulfadimethoxine standard, peak number 2 after chromatographic separation from ormethoprim (peak 1).

Figure 3. Ion trap MS/MS spectrum of the 311 peak eluting at the same retention time as standard of Figure 2 in the Milford pond sample, RW-24.

m/z 156 diagnostic ion was shown to be a useful tool for monitoring other sulfa antibiotics in water samples.1,3-4 Furthermore, the work of Bratton et al.,10 Benotti et al.,11 and Stolker et al.13 also used the m/z 156 fragment ion to monitor other members of the sulfa antimicrobials in water samples, although they do not discuss the structure of the fragment ion. Figure 3 shows the ion trap LC/MS/MS analysis of the Milford pond sample, RW-24. The MS/MS analysis of the m/z 311 gave the spectrum shown in Figure 3, which matches the standard closely, albeit not perfectly (note the ion at 293 m/z). The retention match was correct for sulfadimethoxine. The other pond sample, RW-21, did not contain the m/z 311 ion. The m/z 156 fragment ion was the base peak ion in both the standard and water samples (13) Stolker, A. A. M.; Dijkman, E.; Niesing, W.; Hoendoorn, E. A. In Liquid Chromatography/Mass Spectrometry, MS/MS and Time-of-Flight MS: Analysis of Emerging Contaminants; Ferrer, I., Thurman, E. M., Eds.; American Chemical Society, Oxford University Press: New York, 2003; Chapter 3, pp 32-49.

and the two mass spectra matched well with ions at m/z 245, 218, and 108. The concentration was 0.57 µg/L based on external standard calibration. With this mass spectra identification, we examined this sample again by accurate mass analysis with timeof-flight mass spectrometry, as shown in the following section. Accurate Mass Analysis of Sulfadimethoxine by TOF and Q-TOF. Figure 4 shows the accurate mass analysis of the sulfadimethoxine standard after chromatography using the timeof-flight MS (LCT). The protonated molecule has a mass of m/z 311.0816, which is 0.2 mmu different from the exact mass of 311.0814 u. To also see the important fragment ion of m/z 156, the cone voltage was increased from 20 to 30 V to promote fragmentation, and the standard was reanalyzed. The resulting mass spectrum and doublet peak shown in Figure 4 was obtained. The m/z 156 ion was assumed to be the sulfamido ion shown in Figure 2, which was reported in our earlier paper.3 With a resolution of approximately 7500, the two m/z 156 ions shown in Figure 4 are nearly baseline-resolved. The ions had exact masses of m/z 156.0118 and 156.0778. The first accurate mass of m/z 156.0118 was within -0.1 mmu of the correct mass for the sulfamido ion (C6H6NSO2) of 156.0119 u. Our first assumption was that the second ion of m/z 156.0778 was an impurity of some kind in the sulfadimethoxine standard. Because liquid chromatography was carried out on this standard, the idea of an impurity was unlikely, and upon closer examination of the structure of sulfadimethoxine, it was noted that the other half of the molecule also has an isobaric mass of m/z 156 when it is protonated (Figure 4). Furthermore, the exact mass of this fragment gives m/z 156.0773, which is within 0.5 mmu of the measured mass of the second ion with a mass of m/z 156.0778. The best match to the empirical formula gave the 2,4-dimethoxypyridine structure. The conclusion was reached that sulfadimethoxine underwent an intramolecular isobaric fragmentation. A literature search using these key words was not able to find such a previously reported phenomenon. There was no chemical or a priori reason to think that this is an incorrect fragmentation pathway, although it was fortuitous that the peak intensities were nearly equal (Figure 4). However, to be sure that both m/z 156 ions were originating from the precursor m/z 311 ion, the analysis was repeated using liquid chromatography with quadrupole-time-of-flight mass spectrometry. The protonated molecule was selected as the isolated ion of nominal mass m/z 311. The accurate mass was found to be m/z 311.0818, which is within 0.4 mmu of the correct accurate mass. This result further suggested that the standard was pure and that no other interfering ion was present. The fragmentation in the collision cell of the Q-TOF-MS/MS for the precursor m/z 311 ion gave the same doublet result with masses of m/z 156.0122 and 156.0762 (Figure 5 upper). The best-fit empirical formulas matched the same two structures shown in Figure 4 and concurred with the earlier conclusion that intramolecular fragmentation yields two isobaric ions of nominal mass m/z 156. Further evidence was found by looking at the +1 and +2 isotope patterns of the 156 isobars, which gave the carbon isotope signatures at +1 that are consistent with the second peak containing six carbons versus the four carbons of the first peak (Figure 5 lower). In addition, the sulfur +2 isotope was seen in the first peak and was not present in the second peak Analytical Chemistry, Vol. 76, No. 5, March 1, 2004

1231

Figure 4. Time-of-flight accurate mass analysis of sulfadimethoxine by LCT.

(Figure 5 lower). Furthermore, because these two isobaric ions result from the same molecule, they do not cause error in quantitation for lower resolution instruments when monitoring the m/z 156 ion. Figure 6 (upper) shows the mass spectrum for the remaining fragment ions found in the spectrum of the sulfadimethoxine standard. The accurate mass and assignments for the ion fragments are given in Table 1. Beginning with the highest mass ions, there is the m/z 245.1038 ion, which results from the loss of SO2 and H2, and was also seen earlier in the ion trap mass spectrum (Figure 2). The error on this assignment was -0.1 mmu. Next is the m/z 230.0807, which is the loss of methyl from the m/z 245 (error of 0.3 mmu), the m/z 218.0237 ion, which is the loss of anilino (error of 0.1 mmu), and the m/z 108.0467 ion, which is the anilino with a rearranged oxygen atom (error of 1.8 mmu). Finally, there is the m/z 92.0514 ion, which is the anilino ion (error of 1.4 mmu). The ion trap saw all the same ions, with the exception of the m/z 92 ion, which probably fell below the m/z stability region of the ion trap. The quality of product ion spectra obtained by MS/MS on the Q-TOF exceeds that generated by in-source fragmentation on the single-stage time-of-flight instrument (LCT). This is due to the specificity of the fragmentation that is linked to a narrow precursor m/z range passed through the quadrupole. Figure 6 shows the agreement between the Q-TOF product ion spectrum measured from the sulfadimethoxine standard and that measured at the same retention time for the RW-24 pond sample. Comparison of the accurate mass measurements and fragmentation patterns gives the unequivocal identification of sulfadimethoxine in the Milford pond. The catfish (Ictalurus) in this pond had received treatment with the fish food Romet-30 several months earlier (oral communication with the operator of the Milford Fish Hatchery). Romet-30 contains the antimicrobial sulfadimethoxine. Our field-analysis result is consistent with a laboratory study of Romet-30 degradation,14 which showed that sulfadimethoxine 1232

Analytical Chemistry, Vol. 76, No. 5, March 1, 2004

may have a long half-life in water, with degradation half-lives greater than one year.14 These field results from Milford show that even after more than 3 months after application of Romet-30, sulfadimethoxine was detected in one of two pond samples and corroborate the finding that sulfadimethoxine may be a long-lived antimicrobial in water.2,14 Caveats of Low-Resolution Mass Spectrometry. There are several pitfalls in the analysis of pharmaceuticals in water samples using low-resolution mass spectrometry, especially the use of selected ion monitoring and to a lesser extent, the use of multiple reaction monitoring. First is the lack of a full-scan spectrum when analyzing water samples in either selected ion monitoring (single quadrupole) or multiple reaction monitoring (triple quadrupole). The lack of full-scan spectra prevents the reanalysis of the data file for new compounds or unknowns that may be present. Because samples and their extracts may be unstable, this is an important consideration. Ion traps do acquire full spectral data, but at lower sensitivity than time-of-flight mass spectrometers. Both TOF and Q-TOF are the more sensitive instruments currently available in full-scan mode,15-16 based on a recent review of MS/MS analysis of emerging contaminants. Another useful feature of time-of-flight full-scan accurate mass spectra is the ability to check for matrix interferences at the same nominal mass. For example, Benotti et al.11 pointed out in a comparison of single quadrupole, triple quadrupole, and time-offlight mass spectrometry that the carbon-13 isotope of a large coeluting peak one mass unit less than the selected peak may interfere with a SIM or MRM analysis. His example was the C-13 isotope peak from a large m/z 194 ion that interfered with the (14) Bakal, R. S.; Stoskopf, M. K. Aquaculture 2001, 195, 95-102. (15) Ferrer, I.; Thurman, E. M. Liquid Chromatography/Mass Spectrometry, MS/ MS and Time-of-Flight MS: Analysis of Emerging Contaminants; American Chemical Society, Oxford Press: New York, 2003. (16) Thurman, E. M.; Ferrer, I. In Liquid Chromatography/Mass Spectrometry, MS/MS and Time-of-Flight MS: Analysis of Emerging Contaminants; Ferrer, Imma, Thurman, E. M., Eds.; American Chemical Society, Oxford University Press: New York, 2003; Chapter 2, pp 14-31.

Figure 5. Q-TOF-MS/MS mass spectra of sulfadimethoxine at 156 m/z fragment ion.

m/z 195 of caffeine. With full-scan accurate mass spectra, this interference is easily spotted, especially with time-of-flight, because of the sensitivity in full-scan mode. With either SIM or MRM data, this interfering ion is not found, and a possible interference peak is not known to be present. This is especially problematic with single quadrupole instruments when working at low concentrations and is probably one of the main reasons that pharmaceuticals and other small molecules are overestimated or incorrectly identified in water samples.11 This problem is coupled to the fact that natural organic substances also are ubiquitous in the solidphase-extraction eluents of natural waters and will give strong

spectra across the entire mass range typically studied, from m/z 100 to 50017 in both positive and negative ion electrospray. Thus, there is the possibility of these interferences giving the fragment ion that is being monitored for identification, especially when analyzing low levels of pharmaceuticals in water (0.1 µg/L) samples with poor fragmentation characteristics. As explained by Stolker et al.,13 there are European Guidelines for the LC/MS/MS identification of trace amounts of veterinary drugs in food samples. Such guidelines may be appropriate for (17) Leenheer, J. A.; Rostad, C. E.; Gates, P. M.; Furlong, E. T.; Ferrer, I. Anal. Chem. 2001, 73, 1461-1471.

Analytical Chemistry, Vol. 76, No. 5, March 1, 2004

1233

Figure 6. Q-TOF-MS/MS analysis and comparison of the sulfadimethoxine standard and sample pond RW-24 from Milford Hatchery. Table 1. Measured Mass, Elemental Composition, Exact Mass, Error, and Assignment of Fragment Ions for Sulfadimethoxine Standard measured mass

elemental composition

exact mass

error, mDa

error, ppm

comments

92.0514 108.0467 156.0122 156.0762 218.0237 230.0807 245.1038

C6H6N C6H6N C6H6NO2S C6H10N3O2 C6H8N3O4S C11H10N4O2 C12H13N4O2

92.0500 108.0449 156.0119 156.0773 218.0236 230.0804 245.1039

1.4 1.8 0.3 -1.1 0.1 0.3 -0.1

14.9 16.3 1.8 -6.8 0.5 1.4 -0.2

anilino anilino + rearranged O sulfanilamido 2,4-dimethoxypyridine minus anilino m/z 245 ion minus CH3 minus H2,SO2 rearrangement

environmental samples as advocated in this paper. In their protocol,13 points are given for each identification method by liquid chromatography/mass spectrometry, with a minimum of three identification points required for identification of legal use pharmaceuticals, such as the example here with sulfadimethoxine. The single quadrupole analysis of sulfadimethoxine in pond water would receive the lowest value of 1 point for the method that uses the intensity ratio of the protonated molecule and the m/z 156 fragment ion. The time-of-flight analysis of the protonated molecule and the accurate mass of the m/z 156 ion would receive 2.0 points. The triple quadrupole analysis of the protonated molecule and the 156 m/z fragment ion would receive 2.5 identification points, as would the ion trap analysis of the same two ions. The Q-TOF accurate mass MS/MS analysis would receive 4.5 points for the same transition from protonated molecule to the 156 m/z ion. More points could be accumulated by use of multiple transitions with the single quadrupole, triple quadrupole, ion trap, or Q-TOF instruments. A total of three points must be accumulated in this system to credit a correct identification of a pharmaceutical compound. For single quadrupole analysis, the ratio of the [M + H]+ peak to 1234 Analytical Chemistry, Vol. 76, No. 5, March 1, 2004

three other peaks must be monitored correctly; two peak ratios are required for accurate mass analysis, triple quadrupole MS/ MS, or ion trap MS/MS; and only one transition is necessary for MS/MS accurate mass analysis. For the routine analysis of pharmaceuticals in water, this proposed protocol will fail the identification points, in many cases with the commonly used LC/ MS methods, due to the lack of fragmentation and sensitivity required for environmental monitoring. Only Q-TOF-MS/MS meets the needs of sensitivity and accurate mass monitoring with a single transition. This may be especially important in negative ion mode, where fragmentation is sometimes lacking. Finally, several published studies have found that the combination of ion trap followed by time-of-flight MS is a powerful combination for identification of environmental unknowns and gives a set of complementary information for identification of unknown pharmaceuticals in the environment that is nearly as powerful as Q-TOF-MS/MS.18-24 Our conclusion is that either Q-TOF-MS/MS or the combination of MS/MS and TOF-MS is needed for unequivocal identification of new pharmaceuticals in (18) Ferrer, I.; Thurman, E. M.; Heine, C. Anal. Chem., in press.

the aquatic environment, especially in light of recent discussions in the current environmental literature.7,11,12,18 ACKNOWLEDGMENT We acknowledge the help of Julie Dietz and Betty Scribner of the U.S. Geological Survey for sample collection and the (19) Gilbert, J. R.; Lewer, P.; Duebelbeis, D. O.; Carr, A. W.; Snipes, C. E.; Williamson, R. T. In Liquid Chromatography/Mass Spectrometry, MS/MS and Time-of-Flight MS: Analysis of Emerging Contaminants; Ferrer, I., Thurman, E. M., Eds.; American Chemical Society Oxford University Press: New York, 2003; Chapter 4, pp 52-65. (20) Thurman, E. M.; Ferrer, Imma in Liquid Chromatography/Mass Spectrometry, MS/MS and Time-of-Flight MS: Analysis of Emerging Contaminants; Ferrer, I., Thurman, E. M., Eds.; American Chemical Society, Oxford University Press: New York, 2003; Chapter 8, pp 128-144. (21) Malato, S.; Albanis, T.; Piedra, A.; Aguera, D. H.; Fernandez-Alba, A.; in Liquid Chromatography/Mass Spectrometry, MS/MS and Time-of-Flight MS: Analysis of Emerging Contaminants; Ferrer, I., Thurman, E. M., Eds.; American Chemical Society, Oxford University Press: New York, 2003; Chapter 5, pp 66-95.

operators of the Milford Fish Hatchery. We also give special thanks to Waters Corp. for environmental collaboration studies. Finally, we acknowledge the partial support of this work (Benotti) by EPA Science to Achieve Results (STAR) Program Grant R-82900701-0 (manuscript not approved by the U.S. EPA). Received for review September 18, 2003. Accepted December 2, 2003. AC035094K (22) Bobeldijk, I.; Vissers, J. P. C.; Kearney, G.; Major, H.; Van Leerdam, J. A. J. Chromatogr., A 2001, 929, 63-74. (23) Marchese, S.; Gentili, A.; Perret, D.; D’Ascenzo, G.; Pastori, F. Rapid Comm. Mass Spectrometrom. 2003, 17, 879-886. (24) Ferrer, I.; Thurman, E. M. In Liquid Chromatography/Mass Spectrometry, MS/MS and Time-of-Flight MS: Analysis of Emerging Contaminants; Ferrer, I., Thurman, E. M., Eds.; American Chemical Society, Oxford University Press: New York, 2003; Chapter 22, pp 376-396.

Analytical Chemistry, Vol. 76, No. 5, March 1, 2004

1235