Advertisement

Analytical and Bioanalytical Chemistry

, Volume 410, Issue 22, pp 5689–5702 | Cite as

Speciation analysis of arsenic in seafood and seaweed: Part II—single laboratory validation of method

  • Mesay Mulugeta Wolle
  • Sean D. Conklin
Research Paper
Part of the following topical collections:
  1. Food Safety Analysis

Abstract

Single laboratory validation of a method for arsenic speciation analysis in seafood and seaweed is presented. The method is based on stepwise extraction of water-soluble and non-polar arsenic with hot water and a mixture of dichloromethane and methanol, respectively. While the water-soluble arsenicals were speciated by anion and cation exchange liquid chromatography inductively coupled plasma mass spectrometry (LC-ICP-MS), the non-polar arsenicals were collectively determined by ICP-MS after digestion in acid. The performance characteristics and broad application of the method were evaluated by analyzing eight commercial samples (cod, haddock, mackerel, crab, shrimp, geoduck clam, oyster, and kombu) and four reference materials (fish protein (DORM-4), lobster hepatopancreas (TORT-3), mussel tissue (SRM 2976), and hijiki seaweed (CRM 7405-a)) representing finfish, crustaceans, molluscs, and seaweed. Matrices spiked at three levels in duplicates were also analyzed. The stepwise extraction provided 76–106% extraction of the total arsenic from the test materials. The method demonstrated satisfactory repeatability for analysis of replicate extracts prepared over several days. The accuracy of the method was evaluated by analyzing reference materials certified for both total arsenic and a few arsenicals; the experimental results were 90–105% of the certified values. Comparison between the total water-soluble arsenic and the sum of the concentrations of the chromatographed species gave 80–92% mass balance. While spike recoveries of most arsenicals were in the acceptance range set by CODEX, a few species spiked into cod, haddock, and shrimp were poorly recovered due to transformation to other forms. After thorough investigations, strategies were devised to improve the recoveries of these species by averting their transformations. Limits of quantification (LOQ) for the extraction and quantification of 16 arsenicals using the current method were in the range 6–16 ng g−1 arsenic.

Keywords

Arsenic Seafood Seaweed Speciation Validation 

Introduction

Seafood and seaweed are regarded as important parts of healthy diets as they contain several nutrients associated with beneficial health effects [1]. On the other hand, these marine products contribute substantially to dietary arsenic (As) [2], which is one of the elements of concern in relation to food safety. Although it has the reputation of being highly toxic, the toxicity of arsenic is a function of its chemical forms. Seafood and seaweed contain arsenic in a wide range of chemical forms [2, 3]. Most seafood matrices accumulate arsenic mainly as arsenobetaine (AsB) [4], which is harmless and stable [5]. Some molluscs and seaweed are reported to contain inorganic arsenic (iAs), comprised of arsenite (As3+) and arsenate (As5+), [6] which is classified as a group 1 carcinogen by the International Agency for Research on Cancer [7]. Arsenoribosides (also known as arsenosugars) and lipid-soluble arsenicals, whose potential toxicological effects have yet to be fully elucidated [8, 9, 10, 11, 12], predominate in seaweed [4, 6] and oily products [4, 13], respectively.

Due to the variation in the toxicities of its chemical forms, meaningful risk assessment associated with arsenic must use speciation data. This requires the use of reliable and robust methods of analysis. Over the past decades, numerous methods have been developed for the determination of arsenic species in seafood and seaweed [14, 15]. Most of the methods, however, targeted only a few arsenicals [16, 17, 18, 19, 20, 21, 22, 23] in specific groups of samples [16, 19, 20, 21] and/or reference materials [18, 19, 20, 21, 22]. Moreover, apart from reports on validation of methods to determine a few common arsenic species in such products [22, 23], rigorous validation exercises that address the wide variation and complexity of seafood and seaweed in their matrix compositions and arsenical distributions are not available.

Recently, the US Food and Drug Administration (FDA) developed a method for the comprehensive determination of water-soluble and non-polar arsenic in seafood and seaweed [24]. The method was developed through extensive evaluation of several previously reported approaches. More than a dozen extraction systems including water, methanol-water mixtures, acidic, basic and enzymatic solutions integrated with different heating/agitation procedures were evaluated for their efficiency in extracting arsenic without affecting the chemical forms of its native species. A wide range of samples and reference materials representing finfish, crustaceans, molluscs and seaweed were used for the evaluation. It was found that water, as an extractant, effectively maintains the integrity of arsenic species and allows analysis of extracts by chromatography with no manipulation other than filtration and dilution. Sufficiently quantitative extraction yield of arsenic was achieved for most seafood and seaweed based on stepwise extraction of its water-soluble and non-polar fractions. While the water-soluble arsenicals were speciated by ion exchange chromatography, the non-polar species were collectively determined after digestion in acid. For mass balance, the sum of the water-soluble and non-polar arsenicals was compared with the total arsenic in the sample.

This paper presents the single laboratory validation (SLV) of the above-described procedure. The validation was conducted in accordance with FDA’s Office of Foods and Veterinary Medicine (OFVM) guidelines for validation of chemical methods. [25] Limits of detection (LOD) and quantification (LOQ), selectivity, accuracy, spike recovery, repeatability and measurement uncertainty were evaluated using blanks, standard solutions, and commercial samples and reference materials of finfish, crustaceans, molluscs and algae.

Materials and methods

Instrumentation

A Blixer 3 food processor (Robot Coupe, USA) with a stainless steel bowl and blade was used for sample homogenization. Aqueous extractions were performed in a DigiPREP MS 48-position hot block from SCP Science. A Multi-Purpose Rotator (Thermo Scientific) was used to shake extraction mixtures. Organic solvent was evaporated from non-polar extracts using a benchtop vacuum concentrator (CentriVap, Labconco). Samples and extracts were acid-digested in an UltraClave or UltraWave microwave system (Milestone).

Agilent 1260 LC system consisting of a temperature-controlled autosampler, a binary pump and a vacuum degasser was used. The ICP-MS was Agilent 7900 equipped with concentric nebulizer, Scott-type double pass spray chamber, octopole reaction system, quadrupole mass analyzer and an orthogonal detector. Ultra-high purity (99.999%) argon (Roberts Oxygen Company) and helium (Airgas) were used. The ICP-MS was tuned daily to ensure sufficiently low levels of oxide and doubly charged ions, and operated in helium collision mode to avoid polyatomic interference from 40Ar35Cl+ on 75As+. For ICP-MS analyses, analytical solutions were introduced using an ASX-500 Series autosampler (Agilent Technologies) kept in an enclosure (ENC500, CETAC). Analytical solutions were mixed 1:1, in a Teflon Tee-fitting, with a germanium internal standard solution (20 ng g−1) prepared in 5% (v/v) HNO3, 0.5% (v/v) HCl and 4% (v/v) isopropanol. For LC-ICP-MS analyses, the outlet of the column was connected to the nebulizer inlet with Teflon tubing. A 10 ng g−1 arsenic solution injected post-column was used as an internal standard to compensate for instrument drift over the course of the batch. Table 1 shows the ICP-MS and LC-ICP-MS operating conditions.
Table 1

ICP-MS and LC-ICP-MS operating conditions

ICP-MS

RF power

1550 W

RF matching

1.8 V

Sampling depth

8 mm

Plasma, carrier and makeup gas flow

15, 1.05 and 0.15 L min−1, respectively

Spray chamber temperature

2 °C

ORS3 gas (He) flow

4.5 mL min−1

Cones

Ni

Data acquisition

Spectruma, time resolved analysis (TRA)b

Integration time

0.5 sa, 1.0 sb

Peak pattern

3 points per massa

Replicates per analysis

7a, 1b

LC

Anion exchange

Cation exchange

Guard column

PRP-X100, Hamilton

Metrosep C4 Guard/4.0, Metrohm

Analytical columnc

PRP-X100 (10 μm, 4.1 × 250 mm), Hamilton

Metrosep C6 (5 μm, 4.0 × 250 mm), Metrohm

Mobile phases

(A) 5 mM NH4HCO3

(B) 50 mM (NH4)2CO3

(A) Deionized water

(B) 50 mM pyridine

(both at pH 2 with HNO3)

Gradient profile

0–15 min (3.5% B, 1.0 mL/min), 15.5–24 min (80% B, 1.0 mL/min), 24.5–36 min (99% B, 1.0 mL/min), 36.5–40 min (3.5% B, 1.5 mL/min)

0–22 min (0% B, 0.7 mL/min), 22.5–34 min (10% B, 1.0 mL/min) and 34.5–44 min (0% B, 1.2 mL/min)

Autosampler temperature

4 °C

4 °C

Injection volume

50 μL

50 μL

Column temperature

Ambient

Ambient

Sample diluent

Deionized water

Mobile phase A

aTotal arsenic by ICP-MS

bSpeciation analysis by LC-ICP-MS

Reagents and standards

Optima grade nitric acid (67–70%) and hydrogen peroxide (30–32%) were purchased from Fisher Scientific. Methanol (HPLC grade) and 2-propanol (electronic grade) were obtained from J. T. Baker Chemicals. Dichloromethane (Honeywell Burdick & Jackson), ammonium carbonate (99.999%, Alfa Aesar), ammonium bicarbonate (≥ 99.0%, Sigma-Aldrich), pyridine (HPLC grade, Acros Organics) and N-ethylmaleimide (99+%, Acros Organics) were used. Water deionized to > 18 MΩ cm (Milli-Q Element, Millipore) was used throughout.

A 10 μg mL−1 stock solution of arsenic (High Purity Standard), and 1000 μg mL−1 separate stock solutions of As3+ and As+5 (Inorganic Ventures) were used. A solution of AsB (1.03 g kg−1 as arsenic) was obtained from the Institute for Reference Materials and Measurements of the European Commission’s Joint Research Centre. Sodium salts of monomethylarsonic acid (MMA, 98.5% purity) and dimethylarsinic acid (DMA, 98.9% purity) were purchased from Chem Service. Standards of dimethylarsinoyl acetate (DMAA), dimethylarsinoyl ethanol (DMAE), dimethylarsinoyl propionate (DMAP) and trimethylarsoniopropionate (TMAP) were purchased from the Institute of Chemistry, University of Graz, Austria. Solutions of arsenocholine (AsC), tetramethylarsonium ion (TMA), trimethylarsine oxide (TMAO) and glycerol-, sulfonate-, sulfate- and phosphate-arsinoylribosides (arsenosugars 328, 392, 408 and 482, respectively) were kindly obtained from FDA’s Forensic Chemistry Center.

Spiking solutions were prepared with the following compositions: (I) 10,000 ng g−1 AsB, (II) 10,000 ng g−1 As5+, (III) 200 ng g−1 As3+, (IV) 2.0 ng g−1 arsenosugars 328, 392, 408, and 482, and (V) 200 ng g−1 of each of the remaining species, i.e. AsC, DMA, DMAA, DMAE, DMAP, MMA, TMA, TMAO and TMAP. All concentrations were as ng g−1 arsenic.

Samples and reference materials

Samples of Atlantic cod (Gadus morhua), haddock (Melanogrammus aeglefinus) and Pacific oyster (Crassostrea gigas) were purchased from local stores. Blue swimming crab (Portunus pelagicus), Cortes geoduck clam (Panopea globosa), Whiteleg shrimp (Litopenaeus vannamei) and mackerel (scientific name not available) were collected as part of other FDA research projects. All the samples were fresh except the oyster and mackerel, which were canned. The edible parts of the samples were ground in a food processor and stored at − 30 °C in polypropylene bottles (without freeze drying). Powdered kombu (scientific name not available) was purchased online. The sample was sieved through a 500-μm stainless steel mesh and kept in a polypropylene bottle at room temperature. The reference materials (all freeze-dried) were fish protein (DORM-4) and lobster hepatopancreas (TORT-3) from the National Research Council of Canada, mussel tissue (SRM 2976) from the US National Institute of Standards and Technology, and hijiki seaweed (CRM 7405-a) from the National Metrology Institute of Japan.

Sample digestion for total arsenic determination

The total arsenic concentration in the samples and reference materials was determined according to the method in FDA’s Elemental Analysis Manual Section 4.7 [26]. Briefly, the test material was weighed out (0.25 g dry or 0.5 g wet) into Teflon microwave vessels in triplicate plus a spiked portion fortified with 50 μL of a 10 μg mL−1 arsenic solution. After adding concentrated HNO3 (5 mL) and 30% hydrogen peroxide (1 mL), the mixtures were microwaved in an UltraClave at 250 °C for 15 min with a 30-min linear ramp. The digests were cooled to room temperature, quantitatively transferred to polypropylene tubes and gravimetrically diluted to 50 g with deionized water for analysis by ICP-MS.

Extraction batches and procedures

Extraction batches

The validation plan was to prepare replicate extracts (n = 5) over 3 days for each sample and reference material listed in the Materials and Methods section, along with duplicate spikes at three levels (low, mid and high) of each analyte (except arsenosugars due to limited supply). The four arsenosugars were spiked into one low-level spiked extract generated from each matrix. The three-level spike concentrations are given in the Electronic Supplementary Material (ESM, Table S1). On day one, two non-spiked and two low-level spiked extracts were generated. On the second day, medium-level spiked extracts were prepared along with two non-spiked extracts. The fifth non-spiked extract and the two high-level spiked extracts were prepared on day three. Each batch included three blanks and two fortified blanks. The procedures for stepwise extraction of water-soluble and non-polar arsenic are described in the following sections.

Extraction of water-soluble arsenic

The test material (0.25 g dry or 1.0 g wet) was weighed out into 50 mL polypropylene centrifuge tubes and the designated portions were spiked as described above. Deionized water (15 g) was added into each tube. The mixtures were vortexed and put into a hot block programmed to ramp to 90 °C over 45 min and hold for 30 min. Extraction mixtures were cooled to room temperature and centrifuged at 3000 rpm for 10 min. Supernatants were carefully filtered into polypropylene test tubes through a 0.45-μm pore size polyvinylidene difluoride syringe filter (Whatman, GE Healthcare Life Sciences) and kept at 4 °C. Extracts were analyzed by the anion and cation exchange chromatographic methods described in Table 1. A measured portion of each extract (2.5–5.0 g depending on the total arsenic in the sample matrix) was acid-digested following the procedure described above to determine the total water-soluble arsenic.

Extraction of non-polar arsenic

The solids leftover from aqueous extraction were rinsed with three portions of deionized water to remove residuals of dissolved arsenic and dried in an oven at moderate temperature (60–80 °C). The dry pellets were mixed with 5 mL of a 2:1 (v/v) dichloromethane–methanol mixture and gently shaken at room temperature for 60 min. The supernatants were filtered into disposable borosilicate glass microwave vessels (VWR) and heated to dryness in a centrifugal evaporator at 45 °C. The residues left in the glass microwave vessels were acid-digested in an UltraWave for total non-polar arsenic determination by ICP-MS.

Analyte quantification

The total arsenic, total water-soluble arsenic and total non-polar arsenic concentrations of the test materials were determined by analyzing the respective acid-digests by ICP-MS. Aqueous extracts were also analyzed by anion and cation exchange LC-ICP-MS (Table 1) to speciate the water-soluble arsenicals. In all cases, analytes were quantified based on external calibrations constructed using standards prepared on the day of the analysis. For total arsenic determination, calibration standards (0.5–200 ng g−1) were prepared by serially diluting the 10 μg mL−1 arsenic stock in a solution of 5% HNO3 and 0.5% HCl. For analysis of aqueous extracts by LC-ICP-MS, calibration standards (0.25–100 ng g−1) were prepared from intermediate standards containing all the target arsenicals (except the arsenosugars). The intermediate standards were prepared in deionized water and the calibration standards in the diluents shown in Table 1. Due to limited supply, the four arsenosugars were added into one of the calibration standards to confirm their retention times. All dilutions were made gravimetrically. Agilent MassHunter software was used for data acquisition, chromatographic peak integration and analyte quantification. Data were normalized to the Ge internal standard (ICP-MS) or post-column injected standard (LC-ICP-MS). Further calculations were performed using Microsoft Excel. Concentrations were calculated and reported as mass fractions.

Results and discussion

Chromatographic methods

Table 1 presents the anion and cation exchange chromatographic methods validated in this study for separation and quantitation of water-soluble arsenicals extracted from seafood and seaweed. While the anion exchange method separates As3+, As5+, DMA, DMAA/DMAP, MMA and arsenosugars 392, 408 and 482 on a PRP-X100 column, AsB, AsC, DMA, DMAA, DMAP, DMAE, TMA, TMAO, TMAP and arsenosugar 328 were separated on a Metrosep C6 cation exchange column, see Fig. 1. As can be seen from Fig. 1b, DMAP and DMAE were partially separated on a fresh Metrosep C6 column but the analytes start to co-elute after the column was used for about 2 weeks. Analytes were identified based on their retention times. Since retention times may be altered by sample matrix, fortified analytical portions were used for verification. DMA was quantified by both the anion and cation exchange methods to make sure that its counts are not included in the partially overlapping DMAA in the cation exchange separation. Quantification of DMAP was based on the difference in concentration between DMAA (cation exchange) and co-eluting DMAA/DMAP (anion exchange). DMAE was quantified by subtracting the concentration of DMAP from co-eluting DMAP/DMAE (cation exchange). Fresh Metrosep C6 column was conditioned overnight with 10% B at 0.4 mL min−1. The column was also flushed with 100% A (0.7 mL min−1) for at least an hour before use on the day of analysis to get reproducible separation between early eluting analytes; DMA and DMAA may elute together in the first couple of runs if the column is used without flushing.
Fig. 1

(a) Anion and (b) cation exchange chromatograms showing the separation of arsenic species from a standard solution containing 1.0 ng g−1 of each species. The front unlabeled peak represents the post-column injected standard

Method validation

An important aspect of this study was to evaluate the applicability of the developed extraction and chromatographic methods for analysis of a wide range of seafood and seaweed matrices. Due to practical limits to the number and types of samples to be included in validation exercises, care was taken to represent four broad categories of interest (finfish, crustaceans, molluscs and algae) both as samples and reference materials. Oily and leaner matrices were also considered by including samples like canned mackerel and cod in the list, respectively. The SLV was conducted according to FDA’s OFVM Guidelines for Validation of Chemical Methods [25]. The examined performance characteristics include LOD, LOQ, selectivity, accuracy, spike recovery, repeatability and measurement uncertainty.

Limits of detection and quantification

The LOD and LOQ of the optimized protocols were determined according to the procedure in FDA’s Elemental Analysis Manual Section 3.2 [27]. Deionized water fortified at 0.2 ng g−1 (as arsenic) of each of the 16 species listed in Table 2 was analyzed by anion and cation exchange LC-ICP-MS in replicates (n = 10). LOD and LOQ were calculated based on the standard deviations of the replicate data as LOD = (2) (DF) (t0.95) (σ) (1 + 1/n)1/2 and LOQ = (30) (DF) (σ); where DF is the matrix dilution factor, t0.95 is one-sided t-distribution at 95% confidence level, σ is the standard deviation and n is the number of injections. As can be seen from Table 2, quantification limits in the range of 6–16 ng g−1 were achieved for the arsenicals after 45 times dilution of 1.0 g of a wet sample.
Table 2

Limits of detection (LOD) and quantification (LOQ) of arsenic species (ng g−1)

Specie

As3+

As5+

AB

DMA

DMAA

DMAP

DMAE

MMA

TMAO

TMAP

AC

TMA

328

392

408

482

LOD

1.1

2.0

0.9

1.2

1.0

1.0

1.5

0.8

1.0

1.5

0.8

1.2

1.0

1.1

2.0

1.8

LOQ

8.9

15.7

7.4

9.1

8.0

8.2

11.5

5.9

7.9

11.6

5.9

9.1

7.9

8.6

16.0

13.7

Selectivity

Chloride interferes with the detection of As+ (m/z 75) through 40Ar35Cl+ formation in the ICP. Since the matrices considered in this study are of marine origin, the interference of chloride on the selectivity of the present methods was evaluated using a chloride solution of concentration matching ocean salt content, i.e. 3.5% (w/v). The solution was analyzed by anion and cation exchange LC-ICP-MS with and without the collision cell gas. Peaks representing 40Ar35Cl+ were detected in both analyses operated in ‘no gas’ mode at 21 and 3 min, respectively. The peaks, which did not overlap with any of the 16 arsenicals evaluated, were effectively eliminated by the collision cell with 4.5 mL min−1 He.

In the blank anion exchange chromatograms, a small peak was detected at the retention time of As5+. The As5+ concentration represented by this peak was mostly between the LOD and LOQ values. Previous studies utilizing similar elution gradients indicated that such a peak may be due to As5+ impurity in the mobile phase reagents [28]. The impurity is accumulated on the column during the last (regeneration) step of the gradient and eluted with the higher buffer concentration in the subsequent injection [28].

Accuracy

The accuracy of the method was evaluated by comparing the certified and experimentally found concentrations of analytes in the reference materials. All the reference materials were certified for total arsenic, DORM-4 and TORT-3 for AsB, and CRM 7405-a for As5+. While reference materials representing the four broad categories of interest (finfish, crustaceans, molluscs and algae) were used for validation, the insufficiency in terms of having certified values only for three of the 16 target arsenicals worth mentioning. The experimentally found total arsenic concentrations in all the reference materials were 90–105% of the certified values (see Table 3). The experimental results for AsB in DORM-4 and TORT-3, and As5+ in CRM 7405-a were 95, 93 and 91% of the certified values, respectively. All the experimental values were well within the CODEX acceptance criteria, i.e., 60–115% recovery for 10 μg kg−1 and 80–110% for 0.1–10 mg kg−1 [29].
Table 3

Concentrations of total arsenic and water-soluble arsenicals along with extraction efficiency and chromatographic recovery (n = 5, 95% CI)

Species (ng g−1)

Cod

Haddock

Mackerel

Crab

Shrimp

Geoduck

Oyster

Kombu

DORM-4

TORT-3

SRM 2976

CRM 7405-a

Inorganic arsenic

As3+

  

1.4±0.2

   

5.6±1.7

27±3.1

 

158±26

12±2.7

53±6.2

As5+

4.9±0.7

5.0±0.7

6.6±0.8

7.6±1.8

7.8±1.2

15±4.4

8.9±2.2

322±29

214±21

183±15

91±7.8

9143±108

(10,100±500)

iAs (As3++As5+)

          

103±10.5

110 [23]

 

Methylated arsenicals

AsB

1200±110

6000±1231

405±48

21,100±840

115±2.3

323±33

554±47

352±12

3740±326

(3950±360)

50,800±1254

(54,500±2500)

8700±226

10,300 [23] , 8870 [42]

 

AsC

3.1±0.7

7.8±2.0

6.2±1.3

5.2±1.2

 

9.5±0.9

5.2±1.0

12±1.1

24±5.1

103±21

32±3.0

 

DMA

17±1.8

14±1.7

39±2.9

7.8±1.4

 

26±2.2

58±2.1

427±30

541±55

1303±67

625±29

410 [23], 1270 [42]

468±23

DMAA

   

2.0±0.5

     

148±21

  

DMAE

 

14±2.9

 

15±3.8

 

19±3.3

   

343±59

65±16

 

DMAP

  

4.3±0.4

  

3.2±0.6

3.9±0.9

23±1.3

45±5.1

34±13

82±1.6

25±4.0

MMA

  

1.2±0.4

4.5±0.4

  

1.9±0.3

 

43±4.5

125±39

63±2.7

120 [23]

 

TMA

1.4±0.3

33±7

3.7±0.7

11±0.4

   

12±2.0

20±3.5

175±17

85±4.1

 

TMAO

  

7.2±1.5

3.3±0.6

   

18±2.3

88±15

194±20

41±2.6

36±5.1

TMAP

3.9±0.8

10.8±2.1

8.5±0.8

52±5.1

4.2±1.2

 

7.4±1.8

19±1.7

97±18

523±24

98±11

 

Arsenosugars

Sug 328

   

20±2.8

 

247±37

38±5.2

1876±50

26±1.7

2714±312

355±13

333±14

Sug 392

       

2370±82

 

260±45

 

546±41

Sug 408

   

6.5±1.0

 

5.9±1.0

 

7615±216

 

76±17

14±2.5

3155±123

Sug 482

   

5.6±1.2

 

529±83

328±70

1234±97

 

289±41

658±21

1113±29

Unknown anions

UA1

       

31±3.1

   

27±4.4

UA2

3.4±0.6

2.6±0.6

2.1±0.3

2.8±0.8

  

9.3±1.0

 

10±4.4

78±14

11±1.5

 

UA3

         

98±15

  

UA4

   

13±3.1

 

6.1±1.5

4.4±1.9

26±2.8

  

23±2.5

38±3.4

UA5

      

4.0±1.1

  

68±13

  

UA6

       

13±1.0

   

38±4.1

UA7

          

12±1.8

 

UA8

     

7.2±1.5

7.1±1.2

12±2.4

 

59±20

25±3.6

30±2.4

UA9

          

96±4.2

 

Unknown cations

UC1

       

137±15

    

UC2

     

14±2.8

 

35±2.9

   

40±7.4

UC3

   

9.5±1.7

     

242±56

  

UC4

     

59±8.8

6.2±0.4

87±6.4

 

181±33

33±4.7

534±21

UC5

3.1±0.4

4.8±0.6

          

UC6

         

35±9.3

  

UC7

  

20±1.9

   

25±2.9

42±9.1

27±5.7

63±45

  

UC8

10.4±3.7

24±8.7

 

13±1.2

   

57±7.8

    

UC9

   

80±4.4

    

25±4.7

   

UC10

         

42±4.3

  

Sum of species (μg g−1)

1.2±0.1

6.1±1.3

0.5±0.1

21±0.9

0.1±002

1.3±0.2

1.1±0.1

15±0.4

4.9±0.4

58±0.8

11±0.3

16±0.2

Total arsenic (μg g−1)

1.4±0.04

5.9±0.3

1.2±0.04

22.1±1.0

0.3±0.01

2.9±0.1

2.9±0.1

26.5±0.8

6.2±0.1

(6.87±0.44)

62.6±0.3

(59.5±3.8)

13.9±0.6

(13.3±1.8)

34.2±0.7

(35.8±0.9)

Water-soluble arsenic (%)a

91±7.3

105±22

48±5.2

97±1.6

66±4.1

67±7.5

56±3.3

65±1.2

76±7.1

106±7.0

90±2.8

66±1.4

Chromatographic recovery (%)b

98±2.9

100±2.6

98±0.9

100±1.2

88±3.4

89±3.7

99±3.5 97±3.1

74±5.4

80±13

64±1.8

90±1.0

67±4.6

78±3.2

85±2.2

88±2.4

94±3.0 91±2.2

92±6.4

92±3.3

93±3.8

66±1.4

67±1.4

Non-polar arsenic (%)c

3.7±0.9

1.2±0.6

32±5.3

1.0±0.3

22±3.8

13±2.4

24±4.2

11±2.1

8.1±0.5

2.3±1.0

3.9±0.4

11±3.2

Extraction efficiency (%)d

95±7.1

106±22

80±10

98±1.5

88±5.4

79±8.1

80±6.9

76±2.0

84±6.9

108±6.0

94±3.1

77±2.7

Bolded numbers in parentheses are certified values

Bolded numbers in italics are literature values

Non-bolded numbers in italics (x): LOD < x < LOQ

Underlined numbers are chromatographic recovery values accounting for species in the aqueous extract that are not retained on either the anion or cation exchange columns

a Water-soluble arsenic (%) = (total water-soluble arsenic/total arsenic) × 100%

bChromatographic recovery = (sum of species/total water-soluble arsenic) × 100%

cNon-polar arsenic (%) = (total non-polar arsenic/total arsenic) × 100%

dExtraction efficiency = water-soluble arsenic (%) + non-polar arsenic (%)

Spike recovery

Spike recovery was evaluated for 16 arsenic species. All the samples and two randomly selected reference materials (SRM 2976 and CRM 7405-a) were used for the recovery test. As described above, the arsenicals (except the four arsenosugars) were spiked at three levels in duplicates. The arsenosugars were single spiked into a low-level fortified extract of each test material. Table 4 summarizes average recoveries of the three-level spikes in duplicate (n = 6) for the 12 arsenicals along with the single-spike recoveries of the arsenosugars. Average recoveries in the range 78–124% were obtained for all the species spiked into mackerel, crab, SRM 2976 and CRM 7405a with most values being in the control limit (100 ± 20%) provided in FDA’s methods for arsenic speciation analyses in juice [30] and rice [31]. With the exception of As3+ and/or As5+, recoveries were in the control limits for all the arsenicals spiked into geoduck clam, oyster and kombu. The spike recoveries of iAs in these three matrices indicate that As3+ was partially oxidized to As5+, which was not unexpected.
Table 4

Recovery (%) of arsenic species spiked into the samples and reference materials. Averages of the three-level spike recoveries in duplicates are presented for all arsenicals (n = 6, 95% CI) except the four arsenosugars that were single spiked into one low-level fortified extract per matrix. Numbers in parentheses are %RSD (n = 6)

Species

Fortified blank

Cod

Haddock

Shrimp

Mackerel

Crab

Geoduck

Oyster

Kombu

SRM 2976

CRM 7405-a

As3+

89±3 (3.7)

9±1 (8.6)

100±2 (0.6)

23±3 (13)

97±4 (1.5)

44±5 (11)

100±4 (1.7)

87±2 (2.6)

81±2 (2.0)

67±3 (4.3)

78±8 (10)

51±1 (1.4)

78±8 (10)

84±3 (2.9)

As5+

111±7 (6.0)

112±3 (2.2)

104±2 (1.6)

120±11 (9)

106±4 (4.0)

116±4 (3.2)

120±4 (3.3)

140±8 (5.6)

119±5 (4.0)

124±11 (8)

102±3 (2.3)

iAs (As3++As5+)

98±5 (5.1)

46±2 (3.7)

53±7 (13)

71±6 (8.3)

94±2 (2.2)

93±2 (1.9)

86±2 (2.3)

92±5 (5.2)

88±3 (3.3)

94±5 (5.1)

101±2 (2.3)

AsB

103±5 (4.5)

107±11 (10)

108±11 (9.8)

102±6 (5.3)

102±9 (8.7)

102±10 (8.9)

102±9 (8.2)

110±6 (5.3)

96±8 (8.2)

99±5 (4.5)

102±6 (5.7)

AsC

100±5 (4.8)

101±3 (2.7)

107±2 (1.6)

101±6 (5.4)

103±6 (5.8)

101±3 (2.5)

105±4 (3.7)

102±5 (4.3)

82±3 (4.0)

101±4 (3.9)

84±7 (7.5)

DMA

100±3 (3.0)

98±9 (8.4)

103±5 (5.1)

96±3 (3.4)

101±3 (2.9)

104±2 (1.5)

98±6 (5.9)

98±4 (3.5)

102±7 (6.3)

98±6 (6.2)

101±5 (4.6)

DMAA

94±4 (4.3)

NC

107±9 (3.3)

NC

95±1 (1.6)

NC

109±6 (2.2)

101±5 (4.6)

93±4 (3.6)

97±11 (10)

100±2 (1.9)

90±5 (5.2)

105±6 (5.4)

99±6 (5.7)

DMAE

96±7 (6.7)

11±2 (13)

106±10 (3.7)

2±1 (54)

111±6 (2.2)

48±22 (43)

108±6 (2.1)

101±13 (13)

91±4 (4.2)

104±14 (13)

106±7 (5.9)

81±1 (3.4)

106±6 (5.4)

94±7 (6.7)

DMAP

104±4 (3.6)

ND

102±10 (3.8)

ND

99±10 (3.8)

ND

100±9 (3.7)

78±8 (9.6)

106±4 (3.7)

84±14 (16)

93±6 (6.1)

99±5 (4.4)

85±9 (9.8)

94±9 (9.4)

MMA

102±5 (4.4)

103±9 (8.6)

103±9 (7.9)

103±8 (7.6)

101±9 (8.3)

106±5 (4.9)

103±6 (5.2)

102±6 (5.6)

109±7 (6.1)

101±6 (6.1)

107±5 (4.3)

TMA

96±8 (7.9)

94±4 (4.4)

102±5 (4.4)

97±6 (5.4)

100±6 (6.2)

98±2 (2.4)

100±2 (1.5)

98±6 (5.4)

79±1 (1.5)

98±5 (4.7)

79±4 (4.6)

TMAO

93±5 (5.0)

3±2 (76)

90±4 (1.8)

4±3 (74)

97±2 (0.9)

26±10 (35)

100±12 (4.6)

83±8 (9.3)

100±5 (4.6)

86±4 (4.8)

97±4 (3.9)

87±6 (6.7)

97±5 (5.1)

90±3 (3.1)

TMAP

99±4 (3.7)

99±9 (8.4)

99±8 (7.9)

96±7 (6.9)

101±8 (7.7)

101±3 (2.5)

106±4 (3.2)

99±9 (8.7)

96±7 (6.7)

99±7 (7.0)

97±5 (5.0)

Sug 328

105

81

87

111

111

100

104

97

86

94

101

Sug 392

93

92

92

95

94

90

98

90

84

94

92

Sug 408

113

108

97

104

101

105

114

107

92

109

94

Sug 482

96

94

90

95

100

82

100

99

92

92

89

DMAP+DMAE

 

82±11 (12)

76±14 (14)

96±10 (10)

       

DMAA+TMAO

 

77±11 (13)

80±9 (11)

95±3 (3.3)

       

Spike recovery (%) = ((Analyte concentration in spiked sample – Analyte concentration in non-spiked sample) / Spiked concentration) × 100%

Bolded numbers are spike recoveries of analytes accounting for their transformation products. Due to co-elution of the transformation products (see Fig. 2), recoveries were calculated for the species in joint

Underlined numbers are recoveries of analytes separately spiked (mid-level, n = 3) into samples treated with NEM

NC: Spike recovery of DMAA was not calculated due to its overlap with the transformation product of TMAO

ND: Spiked analyte was not detected

Table 4 shows that As3+, DMAA, DMAE, DMAP and TMAO were poorly recovered from cod, haddock and shrimp. This observation was new as these samples were not included in the method development, [24] hence, it required investigation. Accordingly, the fate of these analytes was studied by individually spiking them into the three samples. For As3+, analysis of the extracts generated from the spiked samples by anion and cation exchange chromatography showed no transformation of the analyte except its slight oxidation to As5+ (4–11%). Acid-digestion of the solid leftovers from the aqueous extractions revealed that substantial fraction of the arsenite stayed in the matrices. A possible reason for this may be binding of the analyte to sulfhydryl groups of membrane proteins as trivalent arsenicals have high affinity for sulfur [32, 33, 34, 35]. To ascertain the effect of these thiol groups on the poor extraction of As3+, fresh samples of cod, haddock and shrimp were treated with a 1% (w/v) aqueous solution of a thiol-selective blocking agent, N-ethylmaleimide (NEM) [36]. After adding the NEM solution (5 mL), samples were vortexed for 30 s and spiked with As3+. The spiked arsenite was quantitatively extracted into hot water without undergoing any transformation, see Table 4.

Study on the fate of DMAA, DMAE, DMAP and TMAO showed that the methylated arsenicals were converted to other forms. The individual spikes of these analytes in cod, haddock and shrimp resulted in the appearance of new peaks in the cation exchange chromatograms clearly showing the conversion of the analytes to unknown forms. The chromatogram in Fig. 2 shows the transformation products of the four arsenicals in cod (all arsenicals were spiked together into the sample). Accounting for these newly formed species improved the spike recoveries of the methylated arsenicals (76–96%); see results in the last two rows of Table 4. Further experiments indicated that the transformation of these arsenicals may be caused by thiol/sulfide groups. None of the four arsenicals showed any transformation when spiked into samples of cod, haddock and shrimp treated with NEM. It was also found that the transformation products (with no use of NEM) can be quantitatively converted back to the original species by treating the extracts with a dilute solution (1% v/v) of H2O2.
Fig. 2

Cation exchange chromatograms showing the transformation of DMAA, DMAE, DMAP and TMAO spiked into cod. The front unlabeled peak represents post-column injected standard

Repeatability

Repeatability, expressed as relative standard deviation (RSD), was evaluated based on data from the chromatographic analyses of the replicate extracts (n = 5) generated from the samples and reference materials listed in the Materials and Methods section. The average concentrations of the arsenicals measured in the five analytical portions of the test materials are summarized in Table 3. The RSD values (not shown in Table 3) for each of these average arsenical concentrations (above LOQ) were in the ranges 7–28% (cod), 7–29% (haddock), 6–16% (mackerel), 3–20% (crab), 1.6–22 (shrimp), 7–16% (geoduck), 3–17% (oyster), 2–18% (kombu), 3–17% (DORM-4), 1–13% (TORT-3), 2–20% (NIST 2976) and 1–15% (CRM 7405-a). Most of the RSD’s were in the 15% control limit set in FDA’s methods for arsenic speciation analyses in juice [30] and rice [31]. The higher RSD’s were for analytes with concentrations close to their LOQ’s.

RSD’s were also evaluated for the three-level spike recovery values. The RSD values for the average spike recoveries (n = 6) in Table 4 were mostly below 10%. Higher values were obtained for the analytes poorly recovered from some of the matrices (see discussion in the Spike Recovery section).

Measurement uncertainty

Measurement uncertainty (U) was estimated using a simplified approach based on existing validation data as described elsewhere [37] as:

$$ U\left(\%\right)=\frac{1}{R}\left(\left(k\sqrt{(RSD)^2+{\left(\frac{RSD}{\sqrt{n}}\right)}^2}\right)+\left(\left|1-R\right|\right)\right) $$

where R is the average recovery of all spiked concentration levels, RSD is the relative standard deviation of the recovery values and n is the number of spiked samples. A coverage factor (k) of 2 was used to get expanded uncertainty at a confidence level of approximately 95%. The relative expanded uncertainty values for all the analytes were between 1.0 and 68.3%. The highest values were recorded for DMAE in cod (68.3%), and TMAO in cod (64.8%) and haddock (36.4%). All the other values were in the acceptable range, i.e., less than twice the respective RSD [37].

Distribution of arsenic species in the tested materials

Table 3 summarizes the concentrations of water-soluble arsenicals found in the samples and reference materials along with mass balance data for extraction and chromatographic recovery. The total arsenic concentration in the samples ranged between 0.3 and 27 μg g−1. As discussed above, the experimentally found and certified concentrations of total arsenic in the reference materials were in good agreement. Stepwise extraction of water-soluble and non-polar arsenic accounted for more than 80% of the total arsenic in most of the matrices. The non-quantitative extraction from a few of the materials may be due to the presence of strongly bound species that need more aggressive conditions for release. Such conditions, however, will affect the chemical forms of native arsenicals, as discussed in the first part of this study [24].

While more than 90% of the arsenic in cod, haddock, crab, TORT-3 and NIST 2976 was found to be soluble in water, non-polar arsenic comprised considerably high fractions in mackerel (32%), oyster (24%) and shrimp (22%). In geoduck clam, kombu and hijiki CRM 7405a, the water-soluble and non-polar arsenic were close to 65 and 12% of their total arsenic, respectively. The fractions of water-soluble arsenic previously reported in cod, [38] mackerel, [38] shrimp, [22, 39] and oyster [22, 39] were comparable to the present values. Regarding non-polar arsenic, Amayo et al. [40] sequentially extracted 65% of the arsenic in mackerel into hexane and a mixture of dichloromethane and methanol. Glabonjat et al. [41] extracted 17.5% of the arsenic in hijiki CRM 7405a into the same non-polar extractant utilized in the present study. The slightly higher non-polar arsenic found by the analysts compared to the present value (11%) may be because they conducted the organic extraction directly on the matrix without prior extraction of water-soluble arsenicals. Any arsenicals extractable into both aqueous and a dichloromethane–methanol mixture would be counted as water-soluble in the present study, but as non-polar in the previous work [41].

Table 3 shows that a total of 35 known and unknown water-soluble arsenicals were detected (above LOD) in the tested materials. Arsenobetaine, which was found in all of the materials except CRM 7405-a, accounted for the high total arsenic in most of the matrices. Both inorganic arsenicals (As3+ and As5+) were quantified in kombu, TORT-3, SRM 2976 and CRM 7405-a, and only As5+ in DORM-4. While DMA was detected in all the matrices except shrimp, samples other than TORT-3 and NIST 2976 seldom contain quantifiable levels of AsC, DMAA, DMAE, DMAP, MMA, TMA, TMAO and TMAP. Molluscs (geoduck clam, oyster and SRM 2976), seaweed (kombu and CRM 7405-a) and TORT-3 were found to be rich in arsenosugars. Among the finfish, arsenosugar was found only in DORM-4 (26 ng g−1 arsenosugar 328). Due to lack of standards, the analytes represented by some of the peaks remained “unknown.” The unknown arsenicals were quantified based on the calibration curve of the nearest eluting standard. Figure 3 shows chromatograms for an aqueous extract of TORT-3, which was found to contain the largest number of arsenicals among the matrices analyzed in this study.
Fig. 3

(a) Anion and (b) cation exchange LC-ICP-MS chromatograms for an aqueous extract generated from lobster hepatopancreas (TORT-3). The front unlabeled peak represents the post-column injected standard

Attempts were made to compare the concentrations of some of the arsenic species found in the reference materials with literature values. Only two studies were found reporting values for a few arsenicals in SRM 2976 based on dilute HNO3 [23] and methanol-water [42] extractions. As can be seen in Table 3, the reported iAs [23] and AsB [23, 42] concentrations were in good agreement (107–119%) with the values found in the present study. However, the DMA [23, 42] and MMA [23] concentrations were significantly different from the current values. One possible reason for the higher DMA concentration in the acidic extract [23] may be degradation of non-polar arsenicals in such extraction systems as demonstrated in the first part of this study [24] and the references therein.

Chromatographic recoveries were calculated as the ratio of the sum of the concentrations of the chromatographed species (including unknowns) and the total concentration of water-soluble arsenic determined after digesting the aqueous extract in acid. The values in Table 3 show that 64–99% of the water-soluble arsenic in the tested materials was speciated by the anion and cation exchange chromatographic methods. The relatively low chromatographic recoveries for geoduck clam (64%), hijiki CRM 7405-a (66%), oyster (67%) and shrimp (74%) may be due to the presence of arsenicals non-retained on either the anion or cation exchange columns and/or because of irreversible binding of analytes to the column(s). Better recoveries were obtained for geoduck clam (90%), oyster (78%) and shrimp (80%) when the species non-retained on both columns were taken into account.

Conclusions

Single laboratory validation of a method for the speciation analysis of arsenic in seafood and seaweed has been conducted according to FDA’s OFVM guidelines for validation of chemical methods. The method involved stepwise extraction of water-soluble and non-polar arsenic. Arsenicals in the aqueous extracts were speciated by ion exchange chromatography while the non-polar fractions were collectively determined after digestion. The stepwise extraction procedure helped to fractionate arsenicals based on their polarities and to achieve sufficiently quantitative extraction yield from most of the tested materials. The aqueous extraction was suitable to extract polar arsenicals while simultaneously keeping them in their native chemical forms. Chromatographic analysis of the aqueous extracts was straightforward as it involved no sample manipulation than filtration and dilution. More than 35 known and unknown water-soluble arsenicals were identified in all the tested materials, along with non-polar arsenic, using the validated procedure. Chromatographic recoveries were satisfactory for water-soluble arsenicals extracted from most of the matrices. The spike recovery values for most of the analytes were in good agreement with the acceptance criteria set by CODEX. A successful procedure was devised to improve the spike recoveries of arsenicals poorly recovered from some samples due to matrix-induced transformations. It should be noted that, apart from improving spike recoveries, this procedure was not used or required to extract native arsenicals form test materials. To the authors’ knowledge, the presented validation exercise is unique and more rigorous than most previous reports in the same field as it was exhaustive in including a wide range of samples and reference materials. The validation also looked at more arsenicals than the commonly targeted ones, and addressed the day-to-day variability of the method by preparing the extracts over several days.

Notes

Acknowledgements

The authors thank Oak Ridge Institute for Science and Education (ORISE) for financial support. Sarah Stadig (Center for Food Safety and Applied Nutrition, FDA) is acknowledged for barcoding the seafood samples, and Dr. Kevin Kubachka (Forensic Chemistry Center, FDA) for kindly providing standards of arsenosugars.

Compliance with ethical standards

Conflict of interest

The authors declare that they have no conflict of interest.

Supplementary material

216_2018_910_MOESM1_ESM.pdf (138 kb)
ESM 1 (PDF 138 kb)

References

  1. 1.
    Larsen R, Eilertsen K-E, Elvevoll EO. Health benefits of marine foods and ingredients. Biotechnol Adv. 2011;29:508–18.CrossRefPubMedGoogle Scholar
  2. 2.
    Molin M, Ulven SM, Meltzer HM, Alexander J. Arsenic in the human food chain, biotransformation and toxicology—review focusing on seafood arsenic. J Trace Elem Med Biol. 2015;31:249–59.CrossRefPubMedGoogle Scholar
  3. 3.
    Francesconi KA. Arsenic species in seafood: origin and human health implication. Pure Appl Chem. 2010;82:373–81.CrossRefGoogle Scholar
  4. 4.
    Taylor V, Goodale B, Raab A, Schwerdtle T, Reimer K, Conklin S, et al. Human exposure to organic arsenic species from seafood. Sci Total Environ. 2017;580:266–82.CrossRefPubMedGoogle Scholar
  5. 5.
    Kaise T, Watanabe S, Itoh K. The acute toxicity of arsenobetaine. Chemosphere. 1985;14:1327–32.CrossRefGoogle Scholar
  6. 6.
    Feldmann J, Krupp EM. Critical review or scientific opinion paper: arsenosugars—a class of benign arsenic species or justification for developing partly speciated arsenic fractionation in foodstuffs? Anal Bioanal Chem. 2011;399:1735–41.CrossRefPubMedGoogle Scholar
  7. 7.
    World Health Organization, International Agency for Research on Cancer (IARC Monographs 100), A review of human carcinogens. C. Metals, arsenic, dusts and fibres, 2012, Lyon, France.Google Scholar
  8. 8.
    Francesconi KA, Tanggaard R, McKenzie CJ, Goessler W. Arsenic metabolites in human urine after ingestion of an arsenosugar. Clin Chem. 2002;48:92–101.PubMedGoogle Scholar
  9. 9.
    Hulle M, Zhang C, Schotte B, Mees L, Vanhaecke F, Vanholder R, et al. Identification of some arsenic species in human urine and blood after ingestion of Chinese seaweed Laminaria. J Anal At Spectrom. 2004;19:58–64.CrossRefGoogle Scholar
  10. 10.
    Andrewes P, Demarini DM, Funasaka K, Wallace K, Lai VWM, Sun H, et al. Do arsenosugars pose a risk to human health? The comparative toxicities of a trivalent and pentavalent arsenosugar. Environ Sci Technol. 2004;38:4140–8.CrossRefPubMedGoogle Scholar
  11. 11.
    Schmeisser E, Goessler W, Francesconi KA. Human metabolism of arsenolipids present in cod liver. Anal Bioanal Chem. 2006;385:367–76.CrossRefPubMedGoogle Scholar
  12. 12.
    Meyer S, Matissek M, Muller SM, Taleshi MS, Ebert F, Francesconi KA, et al. In vitro toxicological characterization of three arsenic-containing hydrocarbons. Metallomics. 2014;6:1023–33.CrossRefPubMedGoogle Scholar
  13. 13.
    Sele V, Sloth JJ, Lundebye A-K, Larsen EH, Berntssen MHG, Amlund H. Arsenolipids in marine oils and fats: a review of occurrence, chemistry and future research needs. Food Chem. 2012;133:618–30.CrossRefGoogle Scholar
  14. 14.
    Francesconi KA, Kuehnelt D. Determination of arsenic species: a critical review of methods and applications, 2000–2003. Analyst. 2004;129:373–95.CrossRefPubMedGoogle Scholar
  15. 15.
    Maher WA, Ellwood MJ, Krikowa F, Raber G, Foster S. Measurement of arsenic species in environmental, biological fluids and food samples by HPLC-ICPMS and HPLC-HG-AFS. J Anal At Spectrom. 2015;30:2129–83.CrossRefGoogle Scholar
  16. 16.
    Larsen EH, Engman J, Sloth JJ, Hansen M, Jorhem L. Determination of inorganic arsenic in white fish using microwave-assisted alkaline alcoholic sample dissolution and HPLC-ICP-MS. Anal Bioanal Chem. 2005;381:339–46.CrossRefPubMedGoogle Scholar
  17. 17.
    Pétursdóttir ÁH, Gunnlaugsdóttir H, Krupp EM, Feldmann J. Inorganic arsenic in seafood: does the extraction method matter? Food Chem. 2014;150:353–9.CrossRefPubMedGoogle Scholar
  18. 18.
    Wahlen R, McSheehy S, Scriver C, Mester Z. Arsenic speciation in marine certified reference materials part 2. The quantification of water-soluble arsenic species by high-performance liquid chromatography-inductively coupled plasma mass spectrometry. J Anal At Spectrom. 2004;19:876–82.CrossRefGoogle Scholar
  19. 19.
    Ackley KL, B’Hymer C, Sutton KL, Caruso JA. Speciation of arsenic in fish tissue using microwave-assisted extraction followed by HPLC-ICP-MS. J Anal At Spectrom. 1999;14:845–50.CrossRefGoogle Scholar
  20. 20.
    Kuehnelt D, Irgolic KJ, Goessler W. Comparison of three methods for extraction of arsenic compounds from NRCC standard reference material DORM-2 and the brown alga Hijiki fuziforme. Appl Organomet Chem. 2001;15:445–56.CrossRefGoogle Scholar
  21. 21.
    Reyes LH, Mar JLG, Rahman GMM, Seybert B, Fahrenholz T, Kingston HM. Simultaneous determination of arsenic and selenium species in fish tissues using microwave-assisted enzymatic extraction and ion chromatography–inductively coupled plasma mass spectrometry. Talanta. 2009;78:983–90.CrossRefPubMedGoogle Scholar
  22. 22.
    Leufroy A, Noël L, Dufailly V, Beauchemin D, Guérin T. Determination of seven arsenic species in seafood by ion exchange chromatography coupled to inductively coupled plasma-mass spectrometry following microwave assisted extraction: method validation and occurrence data. Talanta. 2011;83:770–9.CrossRefPubMedGoogle Scholar
  23. 23.
    Zmozinski AV, Llorente-Mirandes T, López-Sánchez JF, da Silva MM. Establishment of a method for determination of arsenic species in seafood by LC-ICP-MS. Food Chem. 2015;173:1073–82.CrossRefPubMedGoogle Scholar
  24. 24.
    Wolle MM, Conklin SD. Speciation analysis of arsenic in seafood and seaweed: part I—evaluation and optimization of methods. Anal Bioanal Chem.  https://doi.org/10.1007/s00216-018-0906-0
  25. 25.
    US FDA Office of Foods and Veterinary Medicine, Guidelines for validation of chemical methods for the FDA Foods and Veterinary Medicine Program, 2nd ed., April 2015 (https://www.fda.gov/downloads/ScienceResearch/FieldScience/UCM273418.pdf).
  26. 26.
    US FDA Elemental Analysis Manual for Food and Related Products, Section 4.7. Inductively Coupled Plasma-Mass Spectrometric Determination of Arsenic, Cadmium, Chromium, Lead, Mercury, and Other Elements in Food Using Microwave Assisted Digestion, Version 1.1, March 2015 (https://www.fda.gov/Food/FoodScienceResearch/LaboratoryMethods/ucm2006954.htm).
  27. 27.
    US FDA Elemental Analysis Manual for Food and Related Products, Section 3.2. Terminology, September 2015 (https://www.fda.gov/Food/FoodScienceResearch/LaboratoryMethods/ucm2006954.htm).
  28. 28.
    Sloth JJ, Larsen EH, Julshamn K. Selective arsenic speciation analysis of human urine reference materials using gradient elution ion-exchange HPLC-ICP-MS. J Anal At Spectrom. 2004;19:973–8.CrossRefGoogle Scholar
  29. 29.
    Codex Alimentarius Commission, Procedural Manual, 24th ed., 2015, Rome, Italy.Google Scholar
  30. 30.
    US FDA Elemental Analysis Manual for Food and Related Products, Section 4.10. High Performance Liquid Chromatography-Inductively Coupled Plasma-Mass Spectrometric Determination of Four Arsenic Species in Fruit Juice, Version 1.0, July 2013 (https://www.fda.gov/Food/FoodScienceResearch/LaboratoryMethods/ucm2006954.htm).
  31. 31.
    US FDA Elemental Analysis Manual for Food and Related Products, Section 4.11. Arsenic Speciation in Rice and Rice Products Using High Performance Liquid Chromatography-Inductively Coupled Plasma-Mass Spectrometric Determination, Version Draft 1.1, November 2012 (https://www.fda.gov/Food/FoodScienceResearch/LaboratoryMethods/ucm2006954.htm).
  32. 32.
    Rey NA, Howarth OW, Pereira-Maia EC. Equilibrium characterization of the as(III)–cysteine and the as(III)–glutathione systems in aqueous solution. J Inorg Biochem. 2004;98:1151–9.CrossRefPubMedGoogle Scholar
  33. 33.
    Jiang G, Gong Z, Li X-F, Cullen WR, Le XC. Interaction of trivalent arsenicals with metallothioneins. Chem Res Toxicol. 2003;16:873–80.CrossRefPubMedGoogle Scholar
  34. 34.
    Lu M, Wang H, Li X-F, Lu X, Cullen WR, Arnold LL, et al. Evidence of hemoglobin binding to arsenic as a basis for the accumulation of arsenic in rat blood. Chem Res Toxicol. 2004;17:1733–42.CrossRefPubMedGoogle Scholar
  35. 35.
    Wang Z, Zhang H, Li X-F, Le XC. Study of interactions between arsenicals and thioredoxins (human and E. coli) using mass spectrometry. Rapid Commun Mass Spectrom. 2007;21:3658–66.CrossRefPubMedGoogle Scholar
  36. 36.
    Winther JR, Thorpe C. Quantification of thiols and disulfides. Biochim Biophys Acta. 1840;2014:838–46.Google Scholar
  37. 37.
    Llorente-Mirandes T, Calderón J, López-Sánchez JF, Centrich F, Rubio R. A fully validated method for the determination of arsenic species in rice and infant cereal products. Pure Appl Chem. 2012;84:225–38.CrossRefGoogle Scholar
  38. 38.
    Ruttens A, Blanpain AC, De Temmerman L, Waegeneers N. Arsenic speciation in food in Belgium, part 1: fish, molluscs and crustaceans. J Geochem Explor. 2012;121:55–61.CrossRefGoogle Scholar
  39. 39.
    Cao X, Hao C, Wang G, Yang H, Chen D, Wang X. Sequential extraction combined with HPLC–ICP-MS for As speciation in dry seafood products. Food Chem. 2009;113:720–6.CrossRefGoogle Scholar
  40. 40.
    Amayo KO, Raab A, Krupp EM, Marschall T, Horsfall M Jr, Feldmann J. Arsenolipids show different profiles in muscle tissues of four commercial fish species. J Trace Elem Med Biol. 2014;28:131–7.CrossRefPubMedGoogle Scholar
  41. 41.
    Glabonjat RA, Raber G, Jensen KB, Ehgartner J, Francesconi KA. Quantification of arsenolipids in the certified reference material NMIJ 7405-a (Hijiki) using HPLC/mass spectrometry after chemical derivatization. Anal Chem. 2014;86:10282–7.CrossRefPubMedPubMedCentralGoogle Scholar
  42. 42.
    Argese E, Bettiol C, Rigo C, Bertini S, Colomban S, Ghetti PF. Distribution of arsenic compounds in Mytilus galloprovincialis of the Venice lagoon (Italy). Sci Total Environ. 2005;348:267–77.CrossRefPubMedGoogle Scholar

Copyright information

© This is a U.S. Government work and not under copyright protection in the US; foreign copyright protection may apply 2018

Authors and Affiliations

  1. 1.Division of Bioanalytical Chemistry, Office of Regulatory Science, Center for Food Safety and Applied NutritionUS Food and Drug AdministrationCollege ParkUSA

Personalised recommendations