Anal. Chem. 2008, 80, 775-782
Arsenic Speciation and Identification of Monomethylarsonous Acid and Monomethylthioarsonic Acid in a Complex Matrix Santha Ketavarapu V. Yathavakilla,†,‡ Michael Fricke,§ Patricia A. Creed,| Douglas T. Heitkemper,⊥ Nohora V. Shockey,⊥ Carol Schwegel,| Joseph A. Caruso,† and John T. Creed*,|
U.S. EPA, ORD, NERL, MCEARD, Cincinnati, Ohio 45268, Department of Chemistry, University of Cincinnati, Cincinnati, Ohio 45221, and U.S. FDA, Forensic Chemistry Center, Cincinnati, Ohio 45237
Anion-exchange chromatography was utilized for speciation of arsenite (As(III)), arsenate (As(V)), dimethylarsinic acid (DMA(V)), monomethylarsonic acid (MMA(V)), monomethylarsonous acid (MMA(III)), and the new As species monomethylthioarsonic acid (MMTA), using inductively coupled plasma mass spectrometric (ICPMS) detection. MMA(III) and MMTA were identified for the first time in freeze-dried carrot samples that were collected over 25 years ago as part of a joint U.S. EPA, U.S. FDA, and USDA study on trace elements in agricultural crops. The discovery of MMA(III) and MMTA in terrestrial foods necessitated the analytical characterization of synthetic standards of both species, which were used for standard addition in carrot extracts. The negative ion mode, high-resolution electrospray mass spectrometry (HR-ESI-MS) data produced molecular ions of m/z 122.9418 and 154.9152 for MMA(III) and MMTA, respectively. However, ESI-MS was not sensitive enough to directly identify MMA(III) and MMTA in the carrot extracts. Therefore, to further substantiate the identification of MMA(III) and MMTA, two additional separations using an Ion-120 column were developed using the more sensitive ICPMS detection. The first separation used 20 mM tetramethylammonium hydroxide at pH 12.2 with MMA(III) eluting in less than 7 min. In the second separation, MMTA eluted at 11.2 min by utilizing 40 mM ammonium carbonate at pH 9.0. Oxidation of MMA(III) and MMTA to MMA(V) with hydrogen peroxide was observed for standards and carrot extracts alike. Several samples of carrots collected from local markets in 2006 were also analyzed and found to contain low levels of inorganic arsenic species. Arsenic contamination in water and soil is due to the presence of natural hydrogeological backgrounds, volcanic activity, or * To whom correspondence should be addressed. E-mail:
[email protected]. † University of Cincinnati. ‡ Student Contractor at U.S.EPA, ORD, NERL, MCEARD, Cincinnati, Ohio 45268. § Oak Ridge Postdoctoral Research Fellow. | U.S. EPA, ORD, NERL, MCEARD, Cincinnati. ⊥ U.S. FDA, Forensic Chemistry Center. 10.1021/ac0714462 CCC: $40.75 Published on Web 01/09/2008
© 2008 American Chemical Society
anthropogenic sources.1 Arsenic undergoes biochemical transformations and organism-dependent metabolism, which influences its toxicity as it is assimilated into the food chain. Therefore, from a risk assessment perspective, it is important to analyze for not only total arsenic content but also for individual arsenic species. The Safe Drinking Water Act requires the U.S. EPA to estimate the aggregate exposure to a contaminant as a part of a Health Risk Reduction and Cost Analysis (HRRCA) with special emphasis on sensitive subpopulations. For arsenic, the aggregate exposure has two main pathways, including dietary intake and drinking water. Relative to drinking water, the assessment of dietary exposure is considerably more complex because of the diversity of arsenicals associated with a complex solid dietary matrix. More than 25 arsenic species have been identified in various plants and animals over the last 50 years and this list continues to grow.2 However, the precise mechanism of formation and/or functionality of some of the species identified in biological matrixes is not completely understood. It was not until recently that the commonly believed carcinogenic properties of inorganic arsenic have been attributed to its downstream metabolites, such as trivalent methylated arsenicals (dimethylarsinous acid, DMA(III), and monomethylarsonous acid, MMA(III)),3,4 which have also been identified in biological samples like animal tissues and urine.5 With this growing knowledge, some metabolic pathways predict arsenicals in biological matrixes that have yet to be discovered.6 At the same time, new arsenic compounds that have been discovered in biological matrixes do not have a place in the currently known metabolic pathways. A majority of arsenicals that have been discovered are associated with seafood7-12 such as fish, shellfish, and seaweeds, which (1) Zhang, W.; Cai, Y.; Tu, C.; Ma, L. Q. Sci. Total Environ. 2002, 300, 167177. (2) Sloth, J. J.; Larsen, E. H.; Julshamn, K. J. Anal. At. Spectrom. 2003, 18, 452-459. (3) Chen, G.-Q.; Zhou, L.; Styblo, M.; Walton, F.; Jing, Y.; Weinberg, R.; Chen, Z.; Waxman, S. Cancer Res. 2003, 63, 1853-1859. (4) Mass, M. J.; Tennant, A.; Roop, B. C.; Cullen, W. R.; Styblo, M.; Thomas, D. J.; Kligerman, A. D. Chem. Res. Toxicol. 2001, 14, 355-361. (5) Vasken Aposhian, H.; Zakharyan, R. A.; Avram, M. D.; Sampayo-Reyes, A.; Wollenberg, M. L. Toxicol. Appl. Pharmacol. 2004, 198, 327-335. (6) Suzuki, K. T. Anal. Chim. Acta 2005, 540, 71-76. (7) Gallagher, P. A.; Shoemaker, J. A.; Wei, X.; Brockhoff-Schwegel, C. A.; Creed, J. T. Fresenius J. Anal. Chem. 2001, 369, 71-80. (8) Van Hulle, M.; Zhang, C.; Zhang, X.; Cornelis, R. Analyst 2002, 127, 634640.
Analytical Chemistry, Vol. 80, No. 3, February 1, 2008 775
contain relatively high concentrations of arsenic. Terrestrial foods have not been studied as extensively because the concentration of arsenic found is usually in microgram per kilogram levels, which, until recently, were below the detection limits of most analytical instruments.13 Schoof et al.14 indicated that 25-100% of all terrestrial arsenic might be in its inorganic form. Some terrestrial foods for which arsenic speciation analysis has been performed include carrots,15-17 potatoes, Swiss chard, beets, tomatoes,18 mushrooms,19-22 spinach,23,24 lettuce,25,26 etc. However, rice has attracted a lot of attention because of the relatively high worldwide consumption rates27 and is a significant dietary staple in endemic areas. Recently, the elevated levels in U.S. rice samples were attributed to the extensive use of arsenic-containing pesticides.28 Depending on the geographical location, inorganic arsenic29 or DMA30 were observed as the major species in rice. Carrots are another terrestrial crop that have been identified as a target food for arsenic exposure, but have not been as extensively studied. Moreover, USDA reports that the per capita consumption of carrots in 2004 was 11.87 lbs,31 which was close to the 20.4 lbs per capita of rice consumed.32 Studies of carrot specimens, collected from an area where arsenic leached out of CCA treated wood, contained high levels of arsenic.33 A later report by Larsen et al.16 showed that carrots grown in arsenicsupplemented fields displayed stunted growth with higher accumulated total concentrations of arsenic (2.9 mg/kg). Only inorganic arsenic was found in these carrot extracts. Carrots (9) McSheehy, S.; Marcinek, M.; Chassaigne, H.; Szpunar, J. Anal. Chim. Acta 2000, 410, 71-84. (10) Fricke, M. W.; Creed, P. A.; Parks, A. N.; Shoemaker, J. A.; Schwegel, C. A.; Creed, J. T. J. Anal. At. Spectrom. 2004, 19, 1454-1459. (11) Schmeisser, E.; Raml, R.; Francesconi, K. A.; Kuehnelt, D.; Lindberg, A.-L.; Soeroes, C.; Goessler, W. Chem. Commun. 2004, 1824-1825. (12) McKiernan, J. W.; Creed, J. T.; Brockhoff, C. A.; Caruso, J. A.; Lorenzana, R. M. J. Anal. At. Spectrom. 1999, 14, 607-613. (13) Schoof, R. A.; Yost, L. J.; Eickhoff, J.; Crecelius, E. A.; Cragin, D. W.; Meacher, D. M.; Menzel, D. B. Food Chem. Toxicol. 1999, 37, 839-846. (14) Schoof, R. A.; Yost, L. J.; Crecelius, E.; Irgolic, K.; Goessler, W.; Guo, H. R.; Greene, H. Hum. Ecol. Risk Assess. 1998, 4, 117-135. (15) Zandstra, B. H.; De Kryger, T. A. Food Addit. Contam. 2007, 24, 34-42. (16) Helgesen, H.; Larsen, E. H. Analyst 1998, 123, 791-796. (17) Vela, N. P.; Heitkemper, D. T.; Stewart, K. R. Analyst 2001, 126, 10111017. (18) Pyles, R. A.; Woolson, E. A. J. Agric. Food Chem. 1982, 30, 866-870. (19) Slekovec, M.; Irgolic, K. J. Chem. Speciation Bioavailability 1996, 8, 6773. (20) Slejkovec, Z.; Byrne, A. R.; Goessler, W.; Kuehnelt, D.; Irgolic, K. J.; Pohleven, F. Acta Chim. Slov. 1996, 43, 269-283. (21) Byrne, A. R.; Slejkovec, Z.; Stijve, T.; Fay, L.; Goessler, W.; Gailer, J.; Irgolic, K. J. Appl. Organomet. Chem. 1995, 9, 305-313. (22) Stijve, T.; Bourqui, B. Dtsch. Lebensm.-Rundsch. 1991, 87, 307-310. (23) Saraswati, R.; Watters, R. L., Jr. Talanta 1994, 41, 1785-1790. (24) Zhang, Y.; Li, J.; Wang, G. Guangdong Weiliang Yuansu Kexue 2005, 12, 23-27. (25) Warren, G. P.; Alloway, B. J. J. Environ. Qual. 2003, 32, 767-772. (26) Campbell, J. A.; Stark, J. H.; Carlton-Smith, C. H. Heavy Met. Environ., Int. Conf., 5th 1985, 1, 478-480. (27) http://usinfo.state.gov/journals/ites/1005/ijee/buell.htm. (28) Williams, P. N.; Raab, A.; Feldmann, J.; Meharg, A. A. Environ. Sci. Technol. 2007, 41, 2178-2183. (29) Smith, N. M.; Lee, R.; Heitkemper, D. T.; DeNicola Cafferky, K.; Haque, A.; Henderson, A. K. Sci. Total Environ. 2006, 370, 294-301. (30) Williams, P. N.; Price, A. H.; Raab, A.; Hossain, S. A.; Feldmann, J.; Meharg, A. A. Environ. Sci. Technol. 2005, 39, 5531-5540. (31) http://www.ers.usda.gov/publications/vgs/2007/03Mar/VGS31901/ VGS31901.pdf. (32) http://www.ers.usda.gov/publications/agoutlook/aotables/2007/02Feb/ aotab39.xls. (33) Grant, C.; Dobbs, A. J. Environ. Pollut. 1977, 14, 213-226.
776
Analytical Chemistry, Vol. 80, No. 3, February 1, 2008
collected from Northern Chile gave a total arsenic concentration of 49 mg/kg, including inorganic arsenic and low levels of a few methylated forms such as arsenobetaine, AsB, monomethylarsonic acid, MMA(V), and dimethylarsinic acid, DMA(V).34 Vela et al.17 reported that carrots collected directly from agricultural fields in different parts of the United States during the early 1980s had high total arsenic concentrations up to 18.7 mg/kg. The presence of inorganic arsenic, MMA, DMA, and some unidentified species were reported for these samples. This article augments the Vela et al.17 study with a focus on the identification of the unknown arsenicals associated with these carrot samples. Ionexchange chromatography was used in conjunction with inductively coupled plasma mass spectrometry (IC-ICPMS) for this study. EXPERIMENTAL SECTION Reagents and Standards. Degassed, ultrapure 18 mΩ water (DDI; Millipore, Bedford, MA) was used for extraction and in the preparation of all chromatographic mobile phases in an attempt to preserve oxidation-specific information of MMA(III). ACS grade ammonium nitrate, ammonium dihydrogen phosphate, trace metal grade ammonium hydroxide (Fisher Scientific, Pittsburgh, PA), ACS grade ammonium carbonate (Sigma, St. Louis, MO), and 25% tetramethylammonium hydroxide (TMAH; Alfa Products, Danvers, MA) were utilized in chromatographic mobile phases. Ultrapure nitric acid was purchased from J.T. Baker (Phillipsburg, NJ). Stock solutions of 1000 mg L-1 inorganic arsenic (As(III) and As(V)) were obtained from Spex Industries (Metuchen, NJ). Dimethyl arsinic acid (98%) and disodium methylarsenate (99%) were purchased from Chem Service (West Chester, PA). A germanium (Ge) stock solution (1000 mg L-1 Plasma Chem Associates, Bradley Beach, NJ) was used as internal standard in total arsenic analysis. Professor William R. Cullen (Department of Chemistry, University of British Columbia, Vancouver, BC, Canada) kindly donated the crystals of tetramethyl-cyclo-tetraarsaoxane [cyclo-(CH3AsO)4] that were synthesized and characterized as described elsewhere.35 These crystals were stored at -21 °C and at the time of analysis they were hydrolyzed by degassed, deionized water to obtain a stock solution of MMA(III) [CH3As(OH)2 ]. This standard was further purified by 10 consecutive fraction collections using chromatography A conditions (Table 1), and the fractions were stored at 4 °C and used to fortify the carrot extract. Monomethylthioarsonic acid (MMTA) was synthesized as follows: a 4 mg/kg solution of MMA(V) in deionized water was bubbled with H2S gas (generated from the reaction of FeS and HCl)36 for ∼1 h at room temperature. The conversion of MMA(V) to MMTA was monitored via IC-ICPMS. The MMTA was purified by chromatographic fraction collection (chromatography A, Table 1) and characterized by high-resolution electrospray mass spectrometry (HR-ESI-MS). The parent ion produced by MMTA was fragmented in MS/MS mode via collisionally induced dissociation using normalized collision energy of 30% and detected (34) Pizarro, I.; Gomez, M. M.; Camara, C.; Palacios, A. Int. J. Environ. Anal. Chem. 2003, 83, 879-890. (35) Cullen, W. R.; McBride, B. C.; Manji, H.; Pickett, A. W.; Reglinski, J. Appl. Organomet. Chem. 1989, 3, 71-78. (36) Fricke, M. W.; Zeller, M.; Sun, H.; Lai, V. W. M.; Cullen, W. R.; Shoemaker, J. A.; Witkowski, M. R.; Creed, J. T. Chem. Res. Toxicol. 2005, 18, 18211829.
Table 1. Chromatographic Conditions Utilized for IC-ICPMS Analysis of Carrot Extracts Chromatography A Used for Separation of All Arsenic Species pump Agilent 1100 series column PRP-X 100 column dimension 4.6 × 250 mm, 10-µm particle size mobile phase 10 mM ammonium nitrate/10 mM ammonium dihydrogen phosphate, pH 4.5 flow rate 1 mL/min injection volume 100 µL Chromatography B Used for Secondary Confirmatory Analysis of MMA(III) pump Dionex column Ion 120 column dimension 4.6 × 120 mm, 9-µm particle size mobile phase A) 20 mM TMAH B) 20 mM TMAH/5 mM ammonium nitrate pH- 12.1 for both mobile phases flow rate 1 mL/min gradient (step 0-15 min, 100% A elution) 15-15.1 min, 0-100% B 15.1-40 min, 100% B 40-40.1 min, 0-100% A 40.1-70 min, 100% A Injection volume 100 µL Chromatography C Used for Secondary Confirmatory Analysis of MMTA pump Agilent 1100 series column Ion 120 column dimension 4.6 × 120 mm, 9-µm particle size mobile phase 40 mM ammonium carbonate, pH 9.0 flow rate 1 mL/min injection volume 100 µL
using a scan range of 90-400 Da. MMTA solutions were stored at 4 °C to minimize its conversion to MMA(V). Chromatography. Two anion-exchange columns were used in this study, PRP-X100 (PEEK, 4.6 × 250 mm + guard column, Hamilton Reno, NV) and an Ion-120 (4.6 × 120 mm + guard column, Transgenomic, San Jose, CA). Three types of chromatographic conditions, (chromatography A, B, and C; see Table 1) were utilized for the separation and identification of arsenic species. Long-term instrumental drift was corrected using a 20 µg/kg As(V) postcolumn flow injection standard, except in the case of chromatography B. All figures with overlaid chromatograms are offset by 30 s to enable clear viewing. Instrumentation. Inductively coupled plasma mass spectrometers, 7500a and 7500ce (Agilent Technologies, Wilmington, DE), were used for arsenic detection. 75As, 77Se, and 82Se isotopes were monitored during chromatographic runs with a 1-s dwell time for each mass. A Waters hybrid QTOF2 quadrupole time-of-flight mass analyzer from Micromass (Wythenshawe, Manchester, UK) was used for high-resolution and tandem mass spectrometric studies. The instrument was operated in negative ion mode with the electrospray voltage set at 3 kV, sheath gas 20 (unitless), and no auxiliary gas, while the capillary temperature was 300 °C. Chromatographies A and C were used with an Agilent 1100 HPLC (Agilent Technologies, Palo Alto, CA) equipped with a binary pump, a vacuum degasser, and an autosampler. A LabPro
Rheodyne valve (Rohnert Park, CA) was used for the flow injection of a standard (100-µL sample loop) at the beginning and end of each chromatographic run. The Rheodyne valve was controlled by the Agilent 1100. Chromatography B was performed with mobile phase at pH 12.1 using a Dionex GP50 HPLC (Dionex Corp., Sunnyvale, CA) equipped with a binary pump. The Dionex pump was used because the high pH of the mobile phase for this separation was not compatible with the stainless steel components of the Agilent system. Injections were made manually using the LabPro Rheodyne valve. Freeze-drying and homogenization of the fresh carrot samples were performed using a Freezone 4.5 (Labconc, Kansas City, MO) and centrifugal grinding mill (Retsch/Brinkmann, Westbury, NY), respectively. An Eppendorf model 5810R centrifuge (Brinkman Instruments, Westbury, NY) was used. A Mettler AG204 (0.0001 g) balance manufactured by Mettler Instrument Corp. (Hightstown, NJ) was used. A model MDS-2100 microwave digestion system from CEM (Matthews, NC) was used for acid digestion of carrot samples. Sample extraction by sonication was performed using a Cole Parmer sonicator model 08895-28 (Vernon Hill, IL) with temperature control. Sample Collection and Pretreatment. Two batches of carrot samples were analyzed identically in this study. The first batch consisted of carrot samples that were collected in the early 1980s from agricultural fields in several areas of the United States for a study on background levels of trace elements in raw crops.37,38 The samples were freeze-dried and stored in HDPE bottles at -10 °C since collection. The second batch of carrots was obtained from local markets in 2006. During preparation of fresh samples, carrots were scrubbed with a plastic brush under running water. Any rotten portion and ends of the carrot were discarded, and the remainder was rinsed with water and air-dried in a clean air hood. The dried carrots were chopped, freeze-dried, ground, and passed through a 40-µm mesh sieve. During the time of analysis, all samples were stored at 4 °C. Determination of Total Arsenic in Samples. A 0.5 g sample of freeze-dried carrots was weighed and soaked in 5 mL of ultrapure nitric acid overnight. The following day, the samples were digested using a microwave system at 50% power output. The programmable microwave was set at 100 psi pressure at 125 °C temperature limits with a total run time of 60 min. The digested samples were diluted to match the nitric acid concentration of the standards. Standard addition with Ge as an internal standard was used to determine the arsenic concentration on the 7500ce ICPMS with a He flow of 1.4 mL/min. Sample Extraction for Speciation Analysis. In a plastic centrifuge tube, ∼0.5 g of carrot sample and 6 mL of degassed DDI were combined at 60 °C. The sample was vortexed to ensure complete wetting and sonicated at 60 °C for 10 min, with a brief vortex at 5 min. The sonicated sample was centrifuged at 10414g for 10 min. The supernatant was collected, and the remaining residue was re-extracted in the same manner. Each carrot was subjected to at least three sequential extraction cycles. The clear supernatant obtained in each step was filtered individually through (37) Wolnik, K. A.; Fricke, F. L.; Gaston, C. M. Spectrochim. Acta, Part B 1984, 39, 649. (38) Wolnik, K. A.; Fricke, F. L.; Capar, S. G.; Braude, G. L.; Meyer, M. W.; Satzger, R. D.; Bonnin, E. J. Agric. Food Chem. 1983, 31, 1240.
Analytical Chemistry, Vol. 80, No. 3, February 1, 2008
777
0.45-µm Millipore HPF, glass fiber from Millipore Corp., and the filtrate was used for arsenic speciation and for the determination of extraction efficiency. To determine extraction efficiency, the filtered sample solution was treated with 10% w/w nitric acid and diluted with DDI, so that the resultant solution was 2% w/w HNO3. The total arsenic in this solution was determined by following the same protocol as total arsenic in samples. RESULTS AND DISCUSSION Analysis of Arsenic Species in Carrot Extracts. In the previous study, MMA(V), inorganic arsenic and an unidentified arsenic species were observed using anion-exchange chromatography.17 To verify that these unknown species were not those typically associated with marine samples (AsB, arsenocholine, arsenosugars), our laboratory utilized a PRP-X 100 column with a 20 mM ammonium carbonate eluent at pH 9.5. This verification was consistent with the initial investigation, which indicated that the intensity of the unknown in aqueous extracts decreases, as a function of time, with a proportionate increase in MMA(V) concentrations.17 Moreover, the alkaline mobile phase had degenerative effects on the unidentified peak, which coincided with an increase in peak area of MMA(V). Because these observations supported an oxidative conversion to MMA(V), this current study utilized degassed extraction fluids and mobile phases and initially avoided alkaline separation conditions. Among the various methods available for separation of arsenic species,39 anion-exchange chromatography was chosen because it has been widely utilized in arsenic speciation studies. The first separation that was evaluated was similar to that of a previous publication, which utilized a PRP-X100 column with 10 mM each of ammonium nitrate and ammonium dihydrogen phosphate at pH 6.2.40 The analysis using the initial separation indicated poor chromatographic resolution between one unknown peak and DMA. As a result, the pH of the mobile phase was changed to 4.5, which resulted in decreased retention time (tr) for DMA, while tr of the remaining arsenicals was unaffected. Consequently, baseline resolution of the unknown peak from DMA(V) was achieved and this mobile phase (Table 1, chromatography A) was chosen to investigate the presence of this unknown. Upon extending this chromatographic method to another carrot extract, in addition to the first unknown, a second unknown was discovered with a broad peak shape and tr of 16.6 min. On the basis of a tr match from an earlier report, the first unknown peak at 5.4 min was suspected to be MMA(III),40 which converts to MMA(V) upon oxidation. The broad peak shape associated with the second unidentified peak (tr ) 16.6 min) is characteristic of sulfur-containing analogues of arsenic oxide, which can be oxidized using hydrogen peroxide.41 To verify these assumptions, the carrot extract was treated with a small amount of 30% hydrogen peroxide. The chromatographic profile of the peroxide-treated sample showed degradation of both unknowns (Figure 1) and an increase in the integrated peak area of MMA(V). The results of the peroxide additions were consistent with the previous studies.17,40 The logical choice for assignment of the second unknown is the sulfur analogue of MMA(V), MMTA. To (39) Gong, Z.; Lu, X.; Ma, M.; Walt, C.; Le, X. C. Talanta 2002, 58, 77. (40) Pergantis, S. A.; Miguens-Rodriguez, M.; Vela, N. P.; Heitkemper, D. T. J. Anal. At. Spectrom. 2004, 19, 178-182. (41) Francesconi, K. A. Chem. Commun. 2004, 16, 1824-1825.
778 Analytical Chemistry, Vol. 80, No. 3, February 1, 2008
Figure 1. Arsenic speciation analysis of a carrot extract by ICICPMS, compared to extract subjected to oxidation. ? are unknown peaks. Chromatography A (Table 1) conditions were used. Chromatograms are offset for visual clarity.
Figure 2. High-resolution ESI-MS of standards of (a) MMA(III) and (b) MMTA. The inset (c) is an MS/MS of molecular ion peak of MMTA. Instrumental details in the Experimental Section.
verify these hypotheses, synthetic standards of MMA(III) and MMTA that were characterized by high-resolution ESI-MS (Figure 2) were used to fortify carrot extracts (Figure 3). Characterization of MMA(III) and MMTA Synthetic As Standards. The sample of MMA(III) [CH3As(OH)2] was characterized using HR-ESI-MS in the negative ion mode by direct infusion of 10 µL of a 10 mg/kg solution of the standard prepared in 20 mM TMAH at pH 12.1 with 5% MeOH (Figure 2a). The molecular ion peak at m/z 122.9418, with a mass difference (∆m) of 8.9 ppm, appeared only under highly basic conditions. However, the low signal intensity of molecular ion at pH 12.1 can be partially attributed to the oxidation of MMA(III). Because MMA(III) is readily oxidized, it is recommended that the standard be prepared
Figure 3. Arsenic speciation analysis of a carrot extracts by IC-ICPMS, compared to (a) fortification with MMA(III) and (b) fortification of MMTA. Chromatography A (Table 1). Chromatograms are offset for visual clarity.
immediately before introducing it to the ESI source. To the best of our knowledge, this is the first ESI-MS report of a neat MMA(III) standard. An earlier report by Le and co-workers published electrospray data of a complex of MMA(III) with sodium 2,3dimercapto-1-propanesulfonate.42 Synthesis of MMTA (CH3AsSO2H2) was described in the Experimental Section. The synthesized standard of MMTA was characterized by HR-ESI-MS (Figure 2b). A peak at m/z 154.9152 (∆m ) -2.8 ppm) was observed in negative ion mode and corresponds to the molecular ion of MMTA. High-resolution tandem mass spectrometry (Figure 2c) of the peak at m/z 155 showed ions with m/z values of 140 (loss of CH3), 137 (loss of H2O), 121 (due to CH3AsO2•), and 106 (due to AsO2•). The low ESI-MS sensitivity for MMA(III) and MMTA observed for the standards indicated that direct confirmatory analysis of the carrot extracts would not be possible. Hence, fortifying the carrot extracts with synthetic standards on two different chromatographic separations was pursued as a means of identification of the two unknowns. Fortification of Carrot Extracts with Synthetic Arsenic Standards. A 10 µg/kg MMA(III) standard had a retention time matching the first unidentified species (Figure 1). Fortifying the carrot extract with MMA(III) showed a similar peak profile and resulted in a 100% recovery of the standard (Figure 3a). This coelution provides evidence that the peak at 5.4 min could be MMA(III). The elution time and peak profile of the synthesized MMTA standard matched that of the second unknown (Figure 1). A 105% spike recovery was obtained from the MMTA fortified carrot extract with no significant change in peak shape (Figure 3b). Therefore, it was suspected that the unknown species eluting at 16.6 min was MMTA. The retention times of all the species studied by this chromatographic method varied by less than 3% over 4 months of analysis period. Alternate separation methods were developed to further corroborate the presence of MMA(III) and MMTA in the carrot extracts. Alternate Chromatography To Confirm the Presence of MMA(III) and MMTA. Previous reports observed that MMA-
(III) eluted in the void volume of a purely anion-exchange-based chromatography.40,43 An Ion-120 column that predominantly exhibits an anionic retention mechanism was chosen as the stationary phase with an alkaline mobile phase for the present study. Initial chromatographic conditions utilized 10 mM ammonium hydroxide at pH 9.0. However, it was observed that MMA(III) elutes in the void volume, indicating a neutral charge. Retention of the species was achieved by raising the pH of the mobile phase to 12.1 using TMAH (tr 5.6 min), while no other arsenical used in this study was retained at 5.6 min under these conditions. Due to the extreme pH conditions, an all-PEEK Dionex pump was employed for this separation. Since MMA(III) was retained on the column at this pH, the pKa value of MMA(III) could be ∼12. This approximate pKa value is consistent with MMA(III)’s increased sensitivity in ESI-MS analysis under negative ion mode at this pH. Under these conditions, a step elution using an additional 5 mM ammonium nitrate was required to elute the strongly retained MMA(V) (tr 23.8 min) (Table 1, chromatography B). Using the same conditions, a standard of MMA(III) matched the retention time of the unknown in a carrot extract at 5.6 min (Figure 4). The percent RSD of tr for MMA(III) was 2.1% (n ) 5); however, the percent recovery indicated some possible oxidation of MMA(III) to MMA(V). Speciation analysis on a carrot and its spike indicated coelution and a similar chromatographic performance with respect to partial oxidation. The coelution of the MMA(III) standard with the unknown arsenical using two independent chromatographic separations (A and B) was the basis for identification of MMA(III) in carrot samples. The partial oxidation of MMA(III) was analytically acceptable because the primary focus of this chromatography was to confirm the identification of MMA(III). The secondary chromatographic conditions chosen to substantiate the presence of MMTA utilized an Ion-120 column with a mobile phase of 40 mM ammonium carbonate as mobile phase (Table 1, chromatography C). All the synthetic standards used in this study were analyzed by this method. Under these conditions MMA(III) eluted in the void volume, a synthetic standard of MMTA was well-resolved, and other arsenicals demonstrated poor
(42) Gong, Z.; Jiang, G.; Cullen, W. R.; Aposhian, H. V.; Le, X. C. Chem. Res. Toxicol. 2002, 15, 1318-1323.
(43) Xie, R.; Johnson, W.; Spayd, S.; Hall, G. S.; Buckley, B. Anal. Chim. Acta 2006, 578, 186-194.
Analytical Chemistry, Vol. 80, No. 3, February 1, 2008
779
Figure 4. Confirmatory arsenic speciation analysis of MMA(III) in a carrot extract using IC-ICPMS: Chromatograms for unknown in extract of carrot 3 and carrot 3 fortified with MMA(III) standard; the separation used is chromatography B in Table 1. Chromatograms are offset for visual clarity.
Figure 5. IC-ICPMS for confirmation of MMTA: chromatograms for unknown in extract of carrot 4 and carrot 4 fortified with MMTA standard. The separation used is chromatography C in Table 1. Chromatograms are offset for visual clarity.
chromatographic resolution. A tr match of the unknown in carrot extract with a MMTA synthetic standard at 11.3 min (RSD of tr 1.5%, n ) 5) and their coelution in fortified samples was observed (Figure 5). This coelution of MMTA standard with the unknown arsenical in carrot extract by chromatographies A and C provides evidence that MMTA was indeed present in carrots. The literature contains one report which speculates that MMTA could be formed as a reduced species of MMA(V).6 The identification of MMTA (from this study) represents the first analytical data to substantiate its occurrence in a biological system or as a synthetic standard. Because both MMA(III) and MMTA have not been reported earlier in any terrestrial plant matrixes, there was some concern that these species could be produced during the extraction procedure followed in this study. To ensure that MMA(V) was not responsible for formation of MMA(III) or MMTA during extraction process, freeze-dried carrot 5 (Table 2), which had no 780
Analytical Chemistry, Vol. 80, No. 3, February 1, 2008
detectable amount of MMA(V), was fortified with ∼7000 µg/kg of MMA(V). This mixture was allowed to dry in a clean air hood for 18 h, and the sample was extracted and analyzed. The analysis indicated no detectable amount of MMA(III) or MMTA in the extract. This implies that the carrot matrix does not support conversion of MMA(V) to MMA(III) or MMTA during the extraction or chromatographic conditions, described in this work, in the given time frame of 18 h. In addition, as noted by Larsen et al.,16 the carrot matrix was not observed to support conversion of inorganic arsenic to methylated arsenicals. However, the influence of long-term storage of carrots on the formation of MMA(III) or MMTA could not be verified. Arsenic Speciation Analysis of Carrot Samples Using a Mass Balance Approach. The data in Table 2 are reported in a mass balance format, which allows the speciated arsenic (chemical form specific) to be determined relative to the total digested arsenic concentration. Carrots 1-5 were collected in the early 1980s, while carrots 6-8 are present-day samples obtained directly from farms. All of the analyses were conducted on freeze-dried samples, and the percent moisture in fresh carrots was found to be 68-90% (Table 2, column 3). The total arsenic (AsT) was determined after microwave digestion using the method of standard addition. The amount of AsT present in freeze-dried carrots varied over 3 orders of magnitude, ranging from 67 to 18 700 µg/kg of carrot (Table 2, column 2). The relatively high total arsenic concentration in carrots 1-4 may be due to arsenic contamination in the soil or misuse or overspray17 of MMA(V)-based pesticides,40 which might have been translocated to the carrot matrix. Among anthropogenic sources, lead arsenate,44 or methylated arsenicals used as pesticides, could be potential sources of arsenic contamination in soils.28 This may also explain the large variance in total arsenic concentration in the carrots in Table 2. Multiple extraction cycles were utilized in an attempt to liberate the maximum amount of arsenic prior to speciation analysis. To estimate the number of extraction cycles required, two carrots (1 and 5) were subjected to five extraction cycles and the resulting extraction fluids were analyzed for total arsenic. It was observed in both cases (see Supporting Information Figure 1) that the amount of arsenic extracted plateaus after the third extraction cycle, which is consistent with extraction procedure described for dogfish by Shibata.45 Therefore, three extraction cycles were used in this work for the speciation data in Table 2, and the extracts were combined prior to analysis. Speciation Studies by Chromatography A. Based on speciation studies, six arsenic compounds were observed in the present study, MMA(V), MMA(III), MMTA, As(III), As(V), and DMA(V), although not all species were found in all carrots (Table 2, columns 5-10). In carrots 1-4, the MMA(III) and MMA(V) concentration represents between 50 and 75% of the total arsenic. This distribution of arsenicals is considerably different relative to carrots 5-8, which do not contain measurable concentrations of these species. Carrots 1, 3, and 4 contain measurable amounts of DMA(V). It is important to note that DMA(III) could have been (44) Pendergrass, A.; Butcher, D. J. Microchem. J. 2006, 83, 14-16. (45) Shibata, Y.; Morita, M. Anal. Chem. 1989, 61, 2116-2118.
Table 2. Mass Balance Approach for Arsenic Speciation in Carrots Using IC-ICPMS total speciation studies by chromatography A (Table 1) concn, µg/kg mass balance total arsenic arsenic moisture extracted carrot (µg/kg) in carrot, (µg/kg) MMA MMA As As inorg extraction chromatographic overall no. AsT (III) (V) DMA MMTA (III) (V) Asc ΣChrom efficiencyd recoverye recoveryf % AsT.E
d
1a
230
2a
250
3a
9100
4a
18700
5a
130
6b 7b 8b
92 107 67
range of 88-6817 range of 88-6817 range of 88-6817 range of 88-6817 range of 88-6817 89 90 90
220 (35 220 (25 8600 (2000 16950 (4800 111 (12 73.5 88 59
ndg
2.5 (0.9 6.0 (5.0 620 (115 2400 (1780 nd
110 2.0 (24 (2.5 120(12 nd 6100 (1080 11300 (1660 nd
7.5 47 (2.0 (20 24 140.5 (10.4 (14 nd nd
nd nd nd
nd nd nd
nd nd nd
nd
nd nd nd
44 (8.5 56.5 (8.0 45(3.0
16 (7 14 (6 nd
60 (7.7 70.5 (7 45 (3 64.5(6.0 nd 64.5 (6 63.5(2.0 21.5 85 (5 (3 54 2 56 63 3 66 37 7 44
175 (50 200 (22 6800 (1200 14000 (3500 90(10 56 66 44
95 (14 75 (25 90 (25 95 (19 86 (10 80 82 88
80 (11 90 (27 82 (10 77 (27 77 (15 76 74 74
76 (18 69 (9 75 (13 74 (18 65 (5 61 61 65
a Carrots from 1980 field study, interday triplicate speciation ( 2σ. b Carrots from 2006, one time speciation analysis only. c As(III)+As(V). 100 × AsTE/AsT. e 100 × ∑Chrom/AsT.E. f 100 × ∑Chrom/AsT. g nd, below instrumental detection limit.
in the native carrot, but because of its tendency to be oxidized,46 it may not have survived the extraction process. Carrots 3 and 4 were found to contain MMTA. The only arsenical common to all carrot samples is inorganic arsenic (As(III) + As(V)). Columns 9 and 10 of Table 2 report As(III) and As(V) concentrations, but these species can readily interconvert, and for this reason, they may be best summed up as inorganic arsenic (Table 2, column 11). It should be noted that the inorganic arsenic concentrations in all eight carrot samples are within a factor of 2, while the AsT concentrations vary by a factor of 280. Based on the relatively constant amount of inorganic arsenic observed in the present study for this limited carrot sampling, carrots may provide an average of 0.2 µg of inorganic arsenic per day, based on a consumption rate of 15 g of carrot/ day.31 The large variation in the total arsenic concentration is mainly produced by the MMA(V) concentration in older carrots. The presence of MMA in older carrots coupled with the lack of MMA-related species in carrots from the present study provides some evidence that the MMA-related compounds found in older carrots are related in some way to the use of MMA(V)-based pesticides.47 Larsen et al.16 demonstrated that carrots grown in As(III)- and As(V)-enriched soils showed only inorganic arsenic in their extracts. These data suggest that inorganic arsenic in the soil is unlikely to be converted to MMA by the carrot. The speciation results are placed in a mass balance framework in the last three columns of Table 2. The sum of all the species in a given carrot that are quantified chromatographically is represented as ∑Chrom (Table 2, column 12). The extraction efficiency of this method (Table 2, column 13) represents the amount of the arsenic available for speciation analysis relative to the total arsenic ([AsT.E/AsT] × 100) and ranged from 80 to 96%. This indicates that potentially up to 20% of the arsenic remains in the carrot unextracted and, in turn, unavailable for speciation analysis. Chromatographic recovery ([∑Chrom/AsT.E] × 100) gives an estimate of how well the extracted species can be eluted off the (46) Gong, Z.; Lu, X.; Cullen, W. R.; Le, X. C. J. Anal. At. Spectrom. 2001, 16, 1409-1413. (47) Sierra-Alvarez, R.; Cortinas, I.; Yenal, U.; Field, J. A. Appl. Environ. Microbiol. 2004, 70, 5688-5691.
chromatographic column. This recovery was found to be in the range of 74-90% (Table 2, column 14). The lower chromatographic recovery could be attributable to two factors. First, some arsenic may not be eluting from the column. The second factor involves the natural variation associated with determining individual species at concentrations that challenge the instrumental capability. The final mass balance term, overall speciation recovery (Table 2, column 15), gives the percentage of arsenicals speciated relative to the total arsenic determined by microwave digestion. The overall recovery ranged from 61 to 76%, indicating that the chemical form specific information, which aids in predicting toxicity, is available for at least 61% of the arsenic in the carrots. This quantitative speciation analysis improves the risk assessment profiling by providing the percentage of the total arsenic for which the analytical method provides species-specific information. CONCLUSIONS Synthetic standards of MMA(III) and MMTA were characterized by HR-ESI-MS, and coelution of these species with the unknowns in carrots using two different chromatographic separations is the basis of their positive identification in this study. MMA(III) has been reported as a harmful metabolic intermediate in humans exposed to arsenic.48 It has also been identified as a urinary metabolite in exposed populations. However, toxicological effects of direct dietary consumption of MMA(III) are unknown as its biotransformation before cellular uptake has not yet been studied. The literature is void of citations regarding MMTA, and for this reason, it is too early to predict the risk assessment implications of this species. Since carrots were not found to metabolize inorganic arsenic to methylated arsenic species (from earlier studies), the presence of the later is possibly a result of their translocation from natural or anthropogenic contamination of carrot crops. Another plausible explanation is that either carrots or the growth environment of carrots was conducive for formation of MMA(III) or MMTA from the MMA(V) that is abundantly present in these carrots. In (48) Styblo, M.; Del Razo, L. M.; Vega, L.; Germolec, D. R.; LeCluyse, E. L.; Hamilton, G. A.; Reed, W.; Wang, C.; Cullen, W. R.; Thomas, D. J. Arch. Toxicol. 2000, 74, 289-299.
Analytical Chemistry, Vol. 80, No. 3, February 1, 2008
781
addition, the influence of long-term storage effects on formation of MMA(III) and MMTA has not been explored in this study. Thus, further studies must be performed to verify this speculation. Finally, it is important to note that these arsenicals were found only in the carrots collected from the 1980s. Inorganic arsenic concentrations were found to be within a factor of 2 for all eight carrots collected in the 1980s and 2006, although the total arsenic content varied by a factor of 280. Based on the carrots analyzed in this study, exposure to inorganic arsenic by consumption of carrots (0.2 µg/day) is below the U.S. EPA permissible limits in drinking water (20 µg/day).49 Given the limited number of samples evaluated, this is only an estimated average and extensive sample evaluation must be done to reach any consensus.
Agilent Technologies for their continuing research support. This project was supported in part through an Interagency Agreement (IAG# DW-75-92171501-4) between the U.S. FDA and U.S. EPA and by an appointment to the Research Fellowship Program at the U.S. FDA administered by Oak Ridge Associated Universities through a contract with the U.S. FDA. The U.S. EPA through its Office of Research and Development funded and managed the research described here under contract EP05D000677. It has been subjected to Agency review and approved for publication. Mention of trade names or commercial products does not constitute endorsement or recommendation for use. SUPPORTING INFORMATION AVAILABLE Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org.
ACKNOWLEDGMENT The National Institute of Environmental Health Sciences is recognized for partial support of this study. J.A.C. is grateful to
Received for review July 6, 2007. Accepted October 31, 2007.
(49) http://www.epa.gov/safewater/arsenic/pdfs/ars_final_app_n.pdf.
AC0714462
782
Analytical Chemistry, Vol. 80, No. 3, February 1, 2008
