Determination of Arsenobetaine in Fish Tissue by Species Specific

ARTICLE pubs.acs.org/ac

Determination of Arsenobetaine in Fish Tissue by Species Specific Isotope Dilution LC-LTQ-Orbitrap-MS and Standard Addition LC-ICPMS Lu Yang,*,† Jianfu Ding,‡ Paulette Maxwell,† Margaret McCooeye,† Anthony Windust,† Laurent Ouerdane,† Sezgin Bakirdere,†,§,^ Scott Willie,† and Zolt
an Mester† †

Institute for National Measurement Standard, National Research Council Canada, Ottawa, Ontario, Canada K1A 0R6 Institute for Chemical Process and Environmental Technology, National Research Council Canada, Ottawa, Ontario, Canada K1A 0R6 § Department of Chemistry, Middle East Technical University, 06531 Ankara, Turkey ^ Department of Chemistry, Zonguldak Karaelmas University, 67100 Zonguldak, Turkey ‡

ABSTRACT: An accurate and precise method for the determination of arsenobetaine (AsB, (CH3)3þAsCH2COO
) in fish samples using exact matching species specific isotope dilution (ID) liquid chromatography LTQ-Orbitrap mass spectrometry (LC-LTQ-Orbitrap-MS) and standard addition LC inductively coupled plasma mass spectrometry (LC-ICPMS) is described. Samples were extracted by sonication for 30 min with high purity deionized water. An in-house synthesized 13C enriched AsB spike was used for species specific ID analysis whereas natural abundance AsB, synthesized and characterized by quantitative 1H NMR (nuclear magnetic resonance spectroscopy), was used for reverse ID and standard addition LC-ICPMS. With the LTQ-Orbitrap-MS instrument in scan mode (m/z 170
190) and resolution set at 7500, the intensities of [M þ H]þ ions at m/z of 179.0053 and 180.0087 were used to calculate the 179.0053/180.0087 ion ratio for quantification of AsB in fish tissues. To circumvent potential difficulty in mass bias correction, an exact matching approach was applied. A quantitatively prepared mixture of the natural abundance AsB standard and the enriched spike to give a ratio near one was used for mass bias correction. Concentrations of 9.65 ( 0.24 and 11.39 ( 0.39 mg kg
1 (expanded uncertainty, k = 2) for AsB in two fish samples of fish1 and fish2, respectively, were obtained by ID LC-LTQ-Orbitrap-MS. These results are in good agreement with those obtained by standard addition LC-ICPMS, 9.56 ( 0.32 and 11.26 ( 0.44 mg kg
1 (expanded uncertainty, k = 2), respectively. Fish CRM DORM-2 was used for method validation and measured results of 37.9 ( 1.8 and 38.7 ( 0.66 mg kg
1 (expanded uncertainty, k = 2) for AsB obtained by standard addition LC-ICPMS and ID LC-LTQOrbitrap-MS, respectively, are in good agreement with the certified value of 39.0 ( 2.6 mg kg
1 (expanded uncertainty, k = 2). Detection limits of 0.011 and 0.033 mg kg
1 for AsB with LC-ICPMS and ID LC-LTQ-Orbitrap-MS, respectively, were obtained demonstrating that the technique is well suited to the determination AsB in fish samples. To the best of our knowledge, this is first application of species specific isotope dilution for the accurate and precise determination of AsB in biological tissues.

A

rsenic is a well-known toxic trace element in the environment. The toxicity of arsenic depends on its chemical form or speciation.1 In general, inorganic arsenic species, As (III) and As (V), are most toxic whereas organic arsenic species such as monomethylarsonic acid (MMA) and dimethylarsinic acid (DMA) are less toxic.1,2 On the other hand, arsenobetaine (AsB) and arsenocholine (AsC) are touted to be virtually nontoxic,3 but this belief is still for debate due to a recent study on AsB in humans.4 Marine organisms are known to accumulate and biotransform arsenic which often leads to a large number of arsenic species in tissue. This could provide a major route of exposure for humans to arsenic through consumption of marine food products.5,6 As a result of the above concerns, research in speciation of arsenic in biological organisms has increased in the past decade3,4,6
13 due to the need to estimate the risk associated with consumption of marine food products, to access environmental implication and to elucidate the metabolic processes. For arsenic speciation analysis, efficient extraction of arsenic species from a sample matrix without degrading the species is Published 2011 by the American Chemical Society

required. The most frequently used procedures for the extraction of arsenic species from biological samples are based on mixing/ shaking, sonication, accelerated solvent extraction, and microwave assisted extraction with water or a water
methanol mixture.3,4,6
19 Extraction procedures that are capable of quantitatively extracting arsenic species from every biological sample do not currently exist. It appears that the matrix composition is the determining factor for quantitative extraction.12,17,18 For example, for molluscs or particular organs of an organism, the extraction is rarely quantitative17,18 even with modern, aggressive extraction techniques. Nevertheless, for fish tissue, quantitative extraction is relatively successful using a variety of extraction procedures.6,12,18,19 Liquid chromatography (LC) in combination with atomic absorption spectrophotometry,20 atomic fluorescence spectrometry,21 Received: December 14, 2010 Accepted: April 1, 2011 Published: April 01, 2011 3371

dx.doi.org/10.1021/ac103258m | Anal. Chem. 2011, 83, 3371–3378

Analytical Chemistry and inductively coupled plasma mass spectrometry (ICPMS)2
4,6,7,10
12,14
19 for detection is currently the most commonly used technique for arsenic speciation. In particular, ICPMS has been used as a sensitive and selective detector for elemental speciation analysis in the past decade. In addition to its high sensitivity, large dynamic range, and multielement capability, quantitation by isotope dilution (ID) can also be implemented. The isotope dilution mass spectrometry (ID-MS) technique has been widely employed for trace element determinations in a variety of sample matrixes, as it serves to improve both the accuracy and precision of the final results.22 Although applications of species-specific ID have been limited due to a lack of commercially available species-specific spikes,23 this method has been successfully applied for some trace element species such as Sn,24 Hg,25 and Se.23 If ID can be performed, a number of advantages accrue, including the following: enhanced precision and accuracy of the results as the species specific spike serves as an ideal internal standard; matrix effects are accounted for since quantitation is based on ratio measurements; nonquantitative analyte recovery during subsequent sample manipulation does not impact the final results; species alteration during sample workup can be assessed;22 and an alternative and comparative quantitation strategy is provided. Unfortunately, arsenic is a monoisotopic element and the ICPMS does not detect C, O, and H well, thus preventing the use of ID for the determination of arsenic species. On the other hand, organic mass spectrometry with electrospray ionization (ES-MS), is a powerful technique for identifying organic molecules including arsenic species6 in biological tissues. The popular LC-ICPMS approach relies on retention time matching with standards for identification of arsenic species. Due to the limited number of commercially available arsenic species standards, identification of a particular species is impossible when unknown species coelute. Although, it is believed that an LC-ES-MS method for arsenic speciation is less sensitive than the more popular LC-ICPMS approach, the former allows the identification of arsenic species even when they coelute since the accurate mass of the target analyte can be extracted from the chromatogram when a high resolution mass spectrometric detector is employed. In addition, ID can be applied for the determination of arsenic species if 13C or 2H enriched spikes are available, resulting in improved accuracy and precision of the final results.22
25 Despite the advantages offered by LC-ES-MS, isotope dilution calibration with this technique has not yet been applied to arsenic speciation. Clearly, quantitative speciation of arsenic in biological samples requires efficient extraction and good separation and detection, as well as the use of high purity arsenic species standards. Compromise in any of these steps can contribute to the difficulty of the analysis and degrade the accuracy and precision of the results. Despite the increased interest in such determinations, the accuracy of the data cannot be directly verified due to a lack of matrix matched certified reference materials (CRMs) for arsenic species. Furthermore, the choice of CRMs for speciation analysis is currently limited. The National Research Council Canada, Institute for National Measurement Standards (NRCC-INMS) has been active in addressing the need for biological CRMs for the validation of measurements of arsenic species including AsB which is the major form of arsenic found in fish.6,12 As certification typically requires an agreement between at two or more independent analytical methods, development of independent methodologies possessing the highest possible accuracy and precision is required. The objective of this study was to evaluate the application of isotope dilution LC-LTQ-Orbitrap-MS for the determination of AsB using a 13C enriched species specific

ARTICLE

Table 1. LC-ICPMS Operating Conditions Agilent 7500 ICPMS Rf power

1400 W

Rf matching

1.75 V

sample depth

6 mm

Torch-H

0.2 mm

Torch-V


0.1 mm

Ar carrier gas flow rate

0.92 L min
1

Ar makeup gas

0.19 L min
1

He reaction cell gas sampler cone (nickel)

2.5 mL min
1 1.0 mm

skimmer cone (nickel)

0.4 mm

dead time

37.4 ns

extract 1

4.2 V

extract 2


145.2 V

Omega Bias-ce


24 V

Omega Lens-ce

1.2 V

cell entrance QP Focus


22 V 4V

cell exit


36 V

AMU gain

133

AMU offset

126

Axis gain

1.0018

Axis offset

0.05

QP bias


2.8 V

OctP RF OctP bias

140 V
7.4 V

data acquisition

integration time 1.0 s

OctP RF

140 V Agilent HPLC 1200 Series

column

Supelcosil LC-SCX (250  2.1 mm  5 μm)

mobile phase A

DIW

mobile phase B

20.0 mM of ammonium formate adjusted to

gradient elution

pH = 3.0 with formic acid 0
3.5 min, 65% A and 35% B; 3.5
5.5 min, 65
0% A and 35
100% B, 5.5
13.5 min, 100% B; 13.5
5 min, 0
65% % A and 100
35% B

isocratic elution

65% A and 35% B

injection volume

25 μL

spike synthesized in-house. For comparison purposes, standard addition calibration was applied to the determination of AsB in biological CRMs using the LC-ICPMS approach. To the best of our knowledge, this is the first report of the application of species specific isotope dilution calibration for the determination of AsB in biological samples using LC-LTQ-Orbitrap-MS with a 13C enriched spike.

’ EXPERIMENTAL SECTION Instrumentation. An Agilent 7500 ICPMS instrument (Agilent Technologies Canada Inc., Mississauga, Ontario, Canada) equipped with a double-pass spray chamber with Peltier cooler and a glass concentric nebulizer was used for all measurements with the LC-ICPMS approach. A quartz torch with a sapphire injector and a Pt guard electrode were used. Optimization of the Agilent 7500 ICPMS was performed as recommended 3372

dx.doi.org/10.1021/ac103258m |Anal. Chem. 2011, 83, 3371–3378

Analytical Chemistry

ARTICLE

Table 2. LC-LTQ-Orbitrap-MS Operating Conditions ESI source ion spray voltage

3.8 Kv

sheath gas flow rate

25 arb units

aux. gas flow rate

20 arb units

spray current

∼50 μA

capillary temperature

275 C

tube lens voltage

75 V

capillary voltage

24 V

ion spray voltage

3.8 Kv MS

scan mode

full scan mode with positive polarity

resolution

7500 Agilent HPLC 1200 Series

column

Supelcosil LC-SCX

mobile phase A

DIW

mobile phase B

20.0 mM of ammonium formate adjusted to pH = 3.0 with formic acid

mobile phase C

acetonitrile in H2O

isocratic elution

50% A, 35% B, and 15% C

injection volume

25 μL

(250  2.1 mm  5 μm)

(v/v = 35:65) solution

by the manufacturer; typical operating conditions are summarized in Table 1. Helium gas was used as reaction cell gas to minimize ArCl interferences on 75As. A Thermo high resolution LTQ-Orbitrap mass spectrometer (Thermo Fisher Scientific, San Jose, CA) was used for all measurements with the LC-LTQ-Orbitrap-MS approach. Optimization of the LTQ-Orbitrap was performed as recommended by the manufacturer using a 50 ppb AsB standard solution in LC buffer and tuning for m/z 179; typical operating conditions are summarized in Table 2. An Agilent HPLC 1200 Series (Agilent Technologies Canada Inc., Mississauga, Ontario, Canada) with a cation exchange column, Supelcosil LC-SCX (250  2.1 mm  5 μm) and a Supelquard SCX Guard Column from Supelco (Bellefonte, PA, USA) were used for separation of arsenic species. The coupling of LC to ICPMS or LTQ-Orbitrap-MS was accomplished by directing the eluent from the column to the nebulizer of the ICPMS or to the LTQ-Orbitrap-MS electrospray inlet through a 1 and 0.5 m length of PEEK tubing (0.13 mm id, 1.59 mm od), respectively. A Branson 3510 sonication system (Branson, Danbury, CT USA) was used for extraction of arsenic species from fish samples. A Thermo IEC Centra CL3 (Thermo Fisher Scientific, North Carolina, USA) was used for centrifuging samples. Reagents and Solutions. High purity deionized water (DIW) was obtained from a NanoPure mixed bed ion exchange system fed with reverse osmosis domestic feedwater (Barnstead/Thermolyne Corp, Iowa, USA). HPLC grade acetonitrile was purchased from EMD Chemicals INC (Darmstadt, Germany). OmniSolv methanol, ethanol (glass-distilled), and toluene were purchased from EM Science (Gibbstown, NJ, USA). High purity formic acid (88%) was obtained from GFS Chemicals Inc. (Powell, OH, USA). DIW was used as mobile phase A. A 20 mM solution of ammonium formate

(mobile phase B) was prepared by quantitative dissolution of solid ammonium formate (Certified, Thermo Fish Scientific, Ottawa, ON, Canada) in 1 L of DIW and adjusted to pH 3.0 with 3 mL of formic acid (88%). An acetonitrile in H2O (v/v = 35:65) solution was prepared and used as mobile phase C for cleaning the column at the end of the day. Trimethylarsine with 99% purity and bromoacetic acid, (both natural abundance and 13C enriched (BrCH2COOH and BrCH213COOH with 99% enrichment) were purchased from Strem Chemicals, INC (Newburyport, MA, USA). Natural abundance arsenobetaine hydrobromide was synthesized (see below) in-house. An arsenobetaine [(CH3)3AsþCH2COO
] stock solution of 1000 mg kg
1 was gravimetrically prepared in DIW and kept refrigerated until used. An AsB standard solution of 15.13 mg kg
1 was prepared by dilution of AsB stock in DIW. 13 C enriched arsenobetaine hydrobromide was also synthesized (see below) in-house and was used for the determination of AsB in fish tissues by species specific isotope dilution. A stock solution of 13C enriched arsenobetaine [(CH3)3AsþCH213COO
] at 1000 mg kg
1 was prepared by dissolution of this material in DIW. A working spike solution, approximately 13.83 mg kg
1, was prepared by gravimetric dilution of the enriched AsB stock. The final concentration of 13C enriched AsB spike was verified by reverse spike isotope dilution using the natural abundance AsB standard. Two fish samples of fish1 (tissue) and fish2 (liver) from National Research Council Canada (NRCC, Ottawa, ON, Canada) were used as test samples, and DORM-2 from NRCC was used for validation of the method. Synthesis of Natural Abundance and 13C Enriched Arsenobetaine Hydrobromide. Natural abundance arsenobetaine hydrobromide ([(CH3)3AsCH2COOH]Br) and 13C enriched arsenobetaine hydrobromide ([(CH3)3AsCH213COOH]Br) were synthesized in-house following a previously reported procedure.26 In brief, 4.0 g (33.3 mM) of trimethylarsine in 30 mL of dry toluene in a 250 mL flask was purged with argon under vacuum with the flask in an ice bath. 4.89 g (34.9 mM) BrCH2COOH (or BrCH213COOH for synthesis enriched spike) in 20 mL of dry toluene was also purged with argon and was slowly added into trimethylarsine toluene solution using a syringe. Powder formed in 5 min. The reaction mixture was stirred for 4.5 h and then cooled at
20 C for 2.0 h. The white powder was collected by filtration under an argon atmosphere to avoid any possible oxidation. However, filtration under an argon atmosphere is just a standard operation practice in organic lab for producing high quality chemicals. The white powder was rinsed with degassed absolute ethanol twice. Then, the powder was dissolved in 30 mL of absolute ethanol with heating under the protection of argon. The solution was kept at room temperature overnight. The white crystals formed were collected by filtration in an argon atmosphere and rinsed with degassed absolute ethanol twice. The obtained crystal material was recrystallized once again by repeating the above procedure. 5.1 g of white crystals were obtained after the product was dried under vacuum overnight (yield: 58.9%). The product was stored in a freezer prior to the preparation of stock solution. The purity of 0.983 ( 0.005 (expanded uncertainty, k = 2) for the natural abundance arsenobetaine hydrobromide was obtained by quantitative 1H NMR.27 NIST SRM 350b benzoic acid was used as an internal standard. The material was gravimetrically prepared (7 replicates) in CD3OD (Cambridge Isotope Laboratories, Andover, MA), transferred to precision 5 mm NMR tubes (Wilmad LabGlass, Buena, NJ), and flame-sealed. Proton NMR spectra were acquired at 399.94 MHz (Varian Innova, Santa Clara, CA) at 23 C, and free induction decays (FIDs) were acquired (32 scans, 44915 points), 3373

dx.doi.org/10.1021/ac103258m |Anal. Chem. 2011, 83, 3371–3378

Analytical Chemistry apodized with a decreasing exponential, Fourier transformed, and manually phased and integrated. A subsequent purity test on this material was performed with ICPOES, and a value of 0.998 ( 0.010 (expanded uncertainty, k = 2) was obtained on the basis of an acid digestion of AsB solution. Unfortunately, the ICPOES method can only provide total As determination; thus, purity obtained could be biased if the material contains other arsenic species. A further investigation was thus conducted. A 1 mg kg
1 AsB standard solution prepared from the starting material was subsequently injected to LC-ICPMS. In addition to a dominant AsB peak, a small inorganic peak was also found. On the basis of the above experimental results, the purity value obtained by quantitative 1H NMR was thus chosen for the calculation of final results. Safety Considerations: Arsenic compounds are toxic and organic solvents are flammable substances. Material Safety Data Sheets must be consulted and essential safety precautions employed for all manipulations. Sample Preparation and Procedure for the Determination of AsB in Biological CRMs by LC-ICPMS. For convenience, sample preparation was undertaken in a class-100 clean room. Twelve 0.25 g subsamples of each fish1 and fish2 were accurately weighed into individual precleaned 50 mL PE plastic tubes. A suitable mass of 15.13 mg kg
1 AsB standard was added to each of six spiked samples, resulting in a spiked concentration of 1.5fold that of the endogenous AsB concentration in the fish samples. The remaining 6 samples served as the unspiked samples. Similarly, for procedural blanks, spiked and unspiked solutions were prepared as above. Into each tube, 25.0 g of DIW were then added. The tubes were capped and sonicated for 30 min. After centrifugation at 2500 rpm for 10 min, an aliquot of the supernatant (10 mL) was transferred and filtered through a 0.45 μm filter into clean 15 mL PE plastic tubes prior to LCICPMS measurements. The Agilent 7500 ICPMS was optimized, and the HPLC, fitted with a a cation exchange column, Supelcosil LC-SCX (250  2.1 mm  5 μm), was installed and coupled to the ICPMS via a 1 m length of PEEK tubing (0.13 mm i.d., 1.59 mm o.d.). Following injection of the sample (25 μL) onto the HPLC column, data acquisition was automatically triggered by the Agilent 7500 control software. Samples were injected in a sequence of spiked sample
unspiked sample
unspiked sample
spiked sample. Response at 75As was monitored. At the end of the chromatographic run, the acquired data were processed via Chromatographic Data Analysis software on the ICPMS instrument to yield peak areas, which were used to calculate the analyte concentration in the fish CRMs. Sample Preparation and Procedure for the Determination of AsB in Fish by Isotope Dilution LC-LTQ-Orbitrap-MS. The extraction procedure used for the ID LC-LTQ-Orbitrap-MS approach was the same as described above. In brief, six 0.25 g subsamples of each fish sample were accurately weighed into individual precleaned 50 mL PE plastic tubes. A suitable amount of the 13C enriched AsB spike was then added to each vessel to achieve a near one ratio for reference ion and spike ion at m/z 179.0053 and 180.0087 [protonated ion of (CH3)3AsþCH2COOH] at 7500 resolution. Three process blanks (spiked with 50% of the amount of enriched AsB used for the samples) were run along with the samples. The remainder of the sample preparation was the same as described for LC-ICPMS measurements. Calibration of the 13.83 mg kg
1 13C enriched AsB spike was achieved by reverse spike isotope dilution. Four replicate reverse ID solutions were prepared to accompany each set of fish extracts

ARTICLE

Figure 1. Chromatogram of major As species in fish1 extract obtained under gradient elution conditions (see detailed info in Table 1). As: inorganic arsenic including As(III) and As(V); DMAs: dimethyarsenic; AsB: arsenobetaine; and TMAsO: trimethylarsine oxide.

by accurately weighing 0.15
0.58 g of 13.83 mg kg
1 13C enriched AsB spike solution into precleaned 50 mL PE plastic tubes. Aliquots of 0.15
0.58 g of 15.13 mg kg
1 natural abundance AsB standard solution were added. The contents of each tube were then diluted with 25 g DIW prior to LC-LTQOrbitrap-MS analysis. Similarly, a mass bias correction solution was prepared by gravimetrically mixing 0.1736 g of 15.13 mg kg
1 natural abundance AsB standard solution and 0.1852 g of 13.83 mg kg
1 13C enriched AsB spike solution, resulting in a ratio of 1 for m/z of 179.0053 and 180.0087. Each set of fish sample extracts and their corresponding four reverse ID calibration samples were analyzed by LC-ES-Orbitrap-MS on the same day. The mass bias correction solution was injected after every two samples. The integrated peak areas for m/z of 179.0053 and 180.0087 were used to obtain the 179.0053/180.0087 ratio. The mass bias corrected ratios were then used to calculate the analyte concentration in the samples.

’ RESULTS AND DISCUSSION Quantitation of AsB in Fish1 and Fish2 by LC-ICPMS. As noted earlier, the LC-ICPMS configuration is one of most popular techniques used for arsenic speciation due to the high sensitivity and selectivity of the ICPMS. In addition, helium gas was used as a reaction cell gas to minimize any possible polyatomic interferences such as 40Ar35Clþ and 38Ar37Clþ on 75 þ As . As show in Figure 1, chromatographic separation of all major As species6,12 in an extract of fish1 was obtained on the LCSCX cation exchange column with the use of gradient elution with water containing ammonium formate as the mobile phase. Since quantitative determination of only AsB is needed for the certification of fish CRMs fish1 and fish2 at this time, a shorter, 8 min isocratic eluting program using 35% mobile A and 65% mobile B was adopted subsequently to save time. At the end of the day, the column was cleaned with mobile phase A, 20 mM 3374

dx.doi.org/10.1021/ac103258m |Anal. Chem. 2011, 83, 3371–3378

Analytical Chemistry

ARTICLE

Table 3. Results for Extraction Study on Fish1 by LC-ICPMS # of extraction

sample þ tube weight (g)

DIW water added (g)

measured AsB peak area (counts)

extraction 1

12.3237

23.2552

3555000 ( 17700

extraction 2

13.2034

22.9884

130100 ( 1300

131026 ( 650

extraction 3

13.0519

23.2316

4005 ( 60

3990 ( 20

ammonium formate solution at pH 3.0, for 30 min and then mobile phase C, 35% acetonitrile solution, for 30 min to remove the small amount of trimethylarsine oxide (TMAsO) accumulated on the column. No adverse effect on the column was observed during the study period. As noted earlier, Wahlen et al.12 reported a quantitative extraction of AsB from fish tissue using DIW with a series of three extractions by sonication. Since standard addition calibration is used for the quantitation of AsB in fish samples, less manipulation in sample preparation is desired in order to minimize the uncertainty contributed to the final results. Thus, one step extraction is preferred. To investigate the efficiency of one step extraction, a subsample of 0.25 g of fish1 was extracted three times in series with 25 g of DIW for 30 min by sonication. Weights before and after each extraction were recorded in order to calculate expected signals contributed from previous extraction of remaining solution based on AsB peak area measured in the first or second extract. The three extracts of the sample were then analyzed by LC-ICPMS with isocratic elution as shown in Table 1. It is evident from Table 3 that the measured AsB signals in the second and third extracts agreed well with the expected values calculated from measured values in the first and second extracts, respectively. This observation confirms that one step extraction is sufficient and thus was used for all subsequent measurements. The absence of degradation of AsB during the sonication extraction was experimentally confirmed. Three replicate samples of 0.5 g each of 15.13 mg kg
1 AsB standard solution in 25 g DIW were sonicated for 30 min with an additional three replicate samples prepared in the same way without sonication. Concentrations of AsB in these extracts were quantified by one point external calibration. A concentration of 15.15 ( 0.12 mg kg
1 (one standard deviation, n = 3) for AsB was obtained in the sonicated samples, in agreement with the value of 15.13 ( 0.12 mg kg
1 (one standard deviation, n = 3) obtained in control samples. In addition, a subsequent spike recovery test was conducted and quantitative recovery was obtained for AsB in fish1. These results confirmed that there is no degradation of AsB incurred during the extraction process. Fish CRM DORM-2 was used to validate the accuracy of the developed method using one step sonication extraction with DIW and LC-ICPMS detection. One point standard addition calibration was used with the LC-ICPMS approach, and analyte concentration was obtained using the following eq 1: Cx ¼

Iun 3 Cstd 3 msp

!
Cblank mf ðspÞ 3 mxðunÞ Isp 3
Iun 3 mxðspÞ 3 w mf ðunÞ

ð1Þ

where Cx is the blank corrected concentration of AsB in the sample (mg kg
1); Cstd is the concentration of AsB in the natural abundance AsB standard solution (mg kg
1); msp is the mass of natural abundance AsB standard solution added to the spiked

expected AsB left over (counts)

Table 4. Results for AsB in Fish Samples name of standard addition LC-ICPMS samples fish1 fish2 DORM-2

(mg kg
1, U, k = 2)

isotope dilution LC-LTQOrbitrap-MS (mg kg
1, U, k = 2)

9.56 ( 0.32

9.65 ( 0.24

11.26 ( 0.44 37.9 ( 1.8

11.39 ( 0.39 38.7 ( 0.66

DORM-2 certified value

39.0 ( 2.6 (U, 95%)

sample (g); mf(sp) is the sum of the mass of msp and DIW (g); mx(sp) is the mass of sample used to prepare the spiked sample (g); mx(un) is the mass of sample used to prepare the unspiked sample (g); mf(un) is the mass of DIW used to prepare the unspiked sample (g); w is the dry mass correction factor, obtained from the ratio of sample masses after and before drying, respectively; Iun is the measured peak area of the analyte in the unspiked sample (counts); Isp is the measured peak area of the analyte in the spiked sample (counts); Cblank is the analyte concentration in the blank (mg kg
1). As shown in Table 4, a concentration of 37.9 ( 1.8 mg kg
1 (expanded uncertainty, k = 2) for AsB in DORM-2 was obtained by LC-ICPMS, in agreement with the certified value of 39.0 ( 2.6 mg kg
1 (95% confidence interval), confirming the accuracy of the method used. Uncertainty estimations for AsB concentrations in fish samples were made in accordance with JCGM 100:28 Evaluation of Measurement Data-Guide to the Expression of Uncertainty in Measurement,28 using the following law of propagation of uncertainty:  N  X Df 2 2 2 u ðxi Þ uci ðyÞ ¼ Dxi i¼1 !   NX
1 X N Df Df þ23 uðxi Þ 3 uðxj Þ 3 rði, jÞ Dxi 3 Dxj 3 i¼1 j¼i þ 1 ð2Þ where y = f(x1, x2, . . ., xN). The partial derivatives Mf/Mxi are often referred to as sensitivity coefficients, u(xi) and u(xj) is the standard uncertainty associated with the input xi and xj, and r(i,j) is the correlation coefficient,
1 e r(i,j) e 1. The combined uncertainty of the grand mean, uc, was obtained by combining the uncertainties of the individual estimates and the variations between these means as per recent guidelines from NIST.29 The following equation was used: 1X 2 u ci u2C ¼ SD2 þ 2 ð3Þ n where SD is the standard deviation of the n means and uci is the uncertainty of the individual measurand estimates, i = [1. . .n]. Six replicate samples of each of fish1 and fish2 were prepared for the final quantitation of AsB concentrations using one point standard addition LC-ICPMS as described for DORM-2 above. 3375

dx.doi.org/10.1021/ac103258m |Anal. Chem. 2011, 83, 3371–3378

Analytical Chemistry

ARTICLE

species arising from various combinations of isotopes of the reference and spike ions must be included in calculations to derive the true abundances of the reference and spike ions needed for the final quantitation. An isotopic distribution calculator in Analyst software 1.4.1 (MDS Sciex, Toronto, Canada) was used to obtain isotopic distributions in order to calculate the relative abundances of the protonated natural abundance and 13C enriched AsB molecular ions ((CH3)3AsCH2COOHþ and (CH3)3AsCH213COOHþ). The calculated relative abundances based on the IUPAC recommended isotopic abundance of C, H, and As and enriched C are presented in Table 5. Ions at m/z 179.0053 and 180.0087 were selected as reference and spike ions for ID analysis using a 13C enriched AsB to generate the final concentration of AsB in the fish CRMs. With information on calculated relative abundances, calculation of AsB concentrations in candidate CRMs fish1 and fish2 using ID LCLTQ-Orbitrap-MS was undertaken using the following equation:25

Figure 2. (a) Total ion chromatogram (m/z 170 to 190) of a spiked fish1 extract obtained by LC-LTQ-Orbitrap-MS in scan mode using isocratic elution. (b) Spectrum of extracted ion at m/z 179
179.10 for AsB peak. (c) Spectrum of extracted ion at m/z 180
180.10 for AsB peak.

Table 5. Isotopic Compositions of (CH3)3AsCH2COOHþ and (CH3)3AsCH213COOHþ Ions natural abundance

enriched

m/z

(CH3)3AsCH2COOHþ

(CH3)3AsCH213COOHþ

179.0053

0.94053

0.00941

180.0087

0.05419

0.93167

181.0102 182.0131

0.00504 0.00023

0.05370 0.00499

182.0153

0.00001

0.00023

Results are summarized in Table 4; concentrations of 9.56 ( 0.32 and 11.26 ( 0.44 mg kg
1 (expended uncertainty, k = 2) were obtained for AsB in fish1 and fish2, respectively. The detection limit for the standard addition LC-ICPMS technique was evaluated using three blank samples. A value of 0.011 mg kg
1 was estimated for AsB, on the basis of three times the standard deviation of measured concentrations normalized to a 0.25 g subsample. Quantitation of AsB in Fish1 and Fish2 by Isotope Dilution LC-LTQ-Orbitrap-MS. Since only quantitative determination of AsB is of interest at this time, the short isocratic eluting program used for LC-ICPMS was adapted here with a slight modification using 35% mobile A, 50% mobile B, and 15% mobile C for the AsB determination in fish CRMs. The small amount of acetonitrile in the mobile phase helps ionize the analyte more efficiently with LTQ-Orbitrap-MS detection. During the chromatographic runs, a resolution of 7500 was found to be sufficient to separate AsB ions from the background or matrix ions in the mass spectra while still maintaining relatively high sensitivity and, consequently, chromatograms free from interferences for all matrixes were obtained. A full scan approach (m/z 170
190) was used. As shown in Figure 2, good separation and peak profiles for AsB were obtained under the chosen experimental conditions and the entire chromatographic run was accomplished in 6 min. All

my mz Ay
By 3 Rn Bxz 3 R 0n
Axz Ci ¼ C z 3
fb 3 Cb ð4Þ w 3 mx 3 m0y 3 Bxz 3 Rn
Axz 3 Ay
By 3 R 0n where Ci is the blank corrected analyte concentration (mg kg
1); Cz is the concentration of natural abundance AsB standard (mg kg
1); my is the mass of enriched spike solution used to prepare the mixture of sample and spike (g); mx is the mass of sample used (g); w is the dry mass correction factor; mz is the mass of natural abundance AsB standard solution used (g); m0 y is the mass of enriched spike used to prepare the mixture of spike and natural abundance AsB standard solution (g); Ay is the abundance of the reference ion (m/z 179.0053) in the spike; By is the abundance of the spike ion (m/z 180.0087) in the spike; Axz is the abundance of the reference ion in the sample or natural abundance standard; Bxz is the abundance of the spike ion in the sample or natural abundance standard; Rn is the measured ratio of reference/spike ions (mass bias corrected) in the mixtures of sample and spike; R0 n is the measured ratio of reference/ spike ions (mass bias corrected) in the mixtures of spike and natural abundance standard; fb is the blank correction factor, when the blank only contributes to the ID process, fb = 1 and when the blank contributes to both the ID and the reverse ID process fb = (1
(my/ m0 y)  ((Ay
By 3 Rn)/(Bxz 3 Rn
Axz))  ((Bxz 3 R0 n
Axz)/ (Ay
By 3 R0 n))); and Cb is the analyte concentration in the blank (mg kg
1). As clearly expressed by this equation, only reference/ spike ion mass bias corrected ratios in the spiked samples and reverse ID sample solutions need to be known to derive the final analyte concentrations. In general, the mass bias correction factor can be simply calculated from the expected ratio of abundances of reference and spike ions to the measured ratio in a natural abundance analyte standard solution.23
25 Although the mass bias correction factor can vary slightly from day to day due to different instrument settings, it remains a relative constant for a particular analyte and is independent of analyte concentration within the detector range.23
25 The observed mass bias is generally in the range of a few percent in mass spectrometers.23
25 However, in a preliminary study, mass bias correction factors obtained from natural abundance AsB standard solutions varied significantly from 0.3741 ( 0.0038 to 0.5410 ( 0.0045 (mean, 1SD, n = 3) as concentrations of AsB in these standards increased from 100 to 500 ng g
1. To circumvent potential difficulty in mass bias correction, an exact matching approach30 was adopted in this study in order to obtain accurate results by isotope dilution. In addition, instead of the conventional 3376

dx.doi.org/10.1021/ac103258m |Anal. Chem. 2011, 83, 3371–3378

Analytical Chemistry way of the use of a natural abundance standard to obtain a mass bias correction factor, a mixture of natural abundance standard and enriched spike, which was quantitatively prepared to have a ratio near one, was used. Mass bias correction factors were found to remain unchanged in a ratio range of 0.95 to 1.05 and also were independent of analyte concentrations within the detector range. This finding confirms that accurate results can be obtained with use of the abovedescribed approach. If the accurate concentration of the spike is unknown initially, reverse ID can be applied to determine the concentration of the spike using an iterative calculation of the mass bias correction factor starting with a rough concentration of the spike. In addition, both spiked samples and reverse ID solutions were prepared to have ratios near one as well. Therefore, when direct ID and reverse ID are performed during the same measurement period and the intensities and ratios in the spiked samples and reverse ID calibration samples are matched (i.e., Rn ≈ R0 n ≈ 1), more accurate and precise results can be obtained because bias effects encountered during these two steps tend to cancel out,30 as indicated in eq 4. For the final quantitation of AsB concentrations using ID LCLTQ-Orbitrap-MS, six replicate samples of both, fish1 and fish2, and three replicate samples of fish CRM DORM-2 along with four replicates of reverse ID solutions were prepared. Extracts for each fish sample and its corresponding reverse ID solutions were analyzed by LC-ES-Orbitrap-MS on the same day. Results are summarized in Table 4. Concentrations of 9.65 ( 0.24 and 11.39 ( 0.39 mg kg
1 (expended uncertainty, k = 2) were obtained for AsB in fish1 and fish2, respectively. These values are in agreement with results obtained using the standard addition LC-ICPMS approach. Furthermore, results obtained for fish CRM DORM-2 by ID LC-LTQ-Orbitrap-MS are in excellent agreement with the certified value of AsB, confirming the accuracy of the proposed method. Precision improved up to 3-fold in results obtained by the ID LC-LTQ-Orbitrap-MS method as compared to the standard addition LC-ICPMS method. A detection limit of 0.033 mg kg
1 was estimated for AsB, on the basis of three times the standard deviation of the measured concentrations in three blanks normalized to a 0.25 g subsample, which is higher than that obtained by standard addition LC-ICPMS.

’ CONCLUSIONS An accurate and precise method is described for the determination of AsB in fish samples using ID LC-LTQ-Orbitrap-MS and standard addition LC-ICPMS. Concentrations of AsB measured in fish1 and fish2 using a 13C enriched AsB spike are in a good agreement with those obtained by standard addition LC-ICPMS. The exact matching approach with use of a quantitatively prepared mixture of the natural abundance AsB standard and the enriched spike to have a ratio near one for mass bias correction has been successfully applied to obtain accurate final results using ID LC-LTQ-Orbitrap-MS. The LC-ICPMS setup offered better limit of detection (LOD) for AsB; however, identification of As species was based on retention time matching only, and ID analysis was not possible due to the monoisotopic nature of arsenic and poor detection for C, O, and H by ICPMS. The LC-LTQ-Orbitrap-MS offered very high confidence in the identification of the analyte based on mass accuracy and the option of ID analysis using a 13C labeled spike. To the best of our knowledge, this is the first application of species specific isotope dilution for the accurate and precise determination of AsB in biological tissues.

ARTICLE

’ AUTHOR INFORMATION Corresponding Author

*E-mail: [email protected].

’ REFERENCES (1) Fodor, P. Arsenic Speciation in the Environment. In Trace Element Speciation for Environment, Food and Health; Ebdon, L.; Pitts, L.; Cornelis, R.; Crews, H.; Donard, O. F. X.; Quevauviller, P., Eds.; The Royal Society of Chemisty: Cambridge, U.K., 2001. (2) Verdon, C. P.; Caldwell, K. L.; Fresquez, M. R.; Jones, R. L. Anal. Bioanal. Chem. 2009, 393, 939–947. (3) Larsen, E. H.; Pritzl, G.; Hansen, S. H. J. Anal. Atom. Spectrom. 1993, 8, 1075–1084. (4) Newcombe, C.; Raab, A.; Williams, P. N.; Deacon, C.; Haris, P. I.; Meharg, A. A.; Feldmann, J. J. Environ. Monit. 2010, 12 (4), 832–837. (5) Francesconi, K. A.; Edmonds, J. S. Part 1: Cycling and Characterization in Arsenic in the Environment; Nriagu, J. O., Ed.; Wiley: New York, 1994. (6) McSheehy, S.; Mester, Z. J. Anal. Atom. Spectrom. 2004, 19, 373–380. (7) Ciardullo, S.; Aureli, F.; Raggi, A.; Cubadda, F. Talanta 2010, 81, 213–221. (8) Shiomi, K.; Sugiyama, Y.; Shimakura, K.; Nagashima, Y. Appl. Organomet. Chem. 1995, 9, 105–109. (9) Huang, Y. K.; Lin, K. H.; Chen, H. W.; Chang, C. C.; Liu, C. W.; Yang, M. H.; Hsueh, Y. M. Food Chem. Toxicol. 2003, 41, 1491–1500. (10) Slejkovec, Z.; Bajc, Z.; Doganoc, D. Z. Talanta 2004, 62, 931–936. (11) Zheng, J.; Hintelmann, H. J. Anal. Atom. Spectrom. 2004, 19, 191–195. (12) Wahlen, R.; McSheehy, S.; Scriver, C.; Mester, Z. J. Anal. Atom. Spectrom. 2004, 19, 876–882. (13) Francesconi, K. A.; Kuehnelt, D. Analyst 2004, 129, 373–395. (14) Dufailly, V.; No€el, L.; Fr
emy, J. M.; Beauchemin, D.; Gu
erin, T. J. Anal. At. Spectrom. 2007, 22, 1168–1173. (15) Mato-Fern
andez, M. J.; Otero-Rey, J. R.; Moreda-Pi~ neiro, J.; Alonso-Rodríguez, E.; L
opez-Mahía, P.; Muniategui-Lorenzo, S.; PradaRodríguez, D. Talanta 2007, 71, 515–520. (16) S
anchez-Rodas, D.; Luis G
omez-Ariza, J.; Oliveira, V. Anal. Bioanal. Chem. 2006, 385, 1172–1177. (17) Slejkovec, Z.; Van Elteren, J. T.; Byrne, A. R. Talanta 1999, 49, 619–627. (18) Kirby, J.; Maher, W. J Anal. At. Spectrom. 2002, 17, 838–843. (19) McKiernan, J. W.; Creed, J. T.; Brockhoff, C. A.; Caruso, J. A.; Lorenzana, R. M. J. Anal. At. Spectrom. 1999, 14, 607–613. (20) Soleo, L.; Lovreglio, P.; Iavicoli, S.; Antelmi, A.; Drago, I.; Basso, A.; Lorenzo, L. D.; Apostoli, P. Chemosphere 2008, 73, 291–299. (21) Simon, S.; Lobos, G.; Pannier, F.; De Gregori, I.; Pinochet, H.; Potin-Gautier, M. Anal. Chim. Acta 2004, 521, 99–108. (22) De Bi
evre, P. Isotope dilution mass spectrometry. In Trace Element Analysis in Biological Specimens; Herber, R. F. M.; Stoeppler, M., Eds.; Elsevier, Amsterdam, 1994. (23) Yang, L.; Mester, Z.; Sturgeon, R. E. Anal. Chem. 2004, 76, 5149–5156. (24) Bancon-Montigny, C.; Maxwell, P.; Yang, L.; Mester, Z.; Sturgeon, R. E. Anal. Chem. 2002, 74, 5606–5613. (25) Yang, L.; Sturgeon, R. E. J. Anal. At. Spectrom. 2005, 20, 724–729. (26) Minhas, R.; Forsyth, D. S.; Dawson, B. Appl. Organometal. Chem. 1998, 12, 635–641. (27) Windust, A. Anal. Bioanal. Chem. 2010, to be submitted for publication. (28) Evaluation of Measurement Data-Guide to the Expression of Uncertainty in Measurement; JCGM 100:28: http://www.bipm.org/ en/publications/guides/gum.html. 3377

dx.doi.org/10.1021/ac103258m |Anal. Chem. 2011, 83, 3371–3378

Analytical Chemistry

ARTICLE

(29) Levenson, M. S.; Banks, D. L.; Eberhardt, K. R.; Gill, L. M.; Guthrie, W. F.; Liu, H. K.; Vangel, M. G.; Yen, J. H.; Zhang, N. F. J. Res. Natl. Inst. Stand. Technol. 2000, 105, 571–579. (30) Yang, L.; Sturgeon, R. E. J. Anal. At. Spectrom. 2009, 24, 1327–1335.

3378

dx.doi.org/10.1021/ac103258m |Anal. Chem. 2011, 83, 3371–3378