Environ. Sci. Technol. 2009, 43, 4818–4823
Arsenic Speciation of Terrestrial Invertebrates MAEVE M. MORIARTY,† IRIS KOCH,† ROBERT A. GORDON,‡ AND K E N N E T H J . R E I M E R * ,† Environmental Sciences Group, Royal Military College of Canada, P.O. Box 17000 Station Forces, Kingston, Ontario, K7K 7B4, Canada, Simon Fraser University, Burnaby, BC, Canada, and PNC/XOR, Advanced Photon Source, Argonne, Illinois
Received January 11, 2009. Revised manuscript received April 1, 2009. Accepted April 28, 2009.
The distribution and chemical form (speciation) of arsenic in terrestrial food chains determines both the amount of arsenic available to higher organisms, and the toxicity of this metalloid in affected ecosystems. Invertebrates are part of complex terrestrial food webs. This paper provides arsenic concentrations and arsenic speciation profiles for eight orders of terrestrial invertebrates collected at three historical gold mine sites and one background site in Nova Scotia, Canada. Total arsenic concentrations, determined by inductively coupled plasma mass spectrometry (ICP-MS), were dependent upon the classification of invertebrate. Arsenic species were determined by highperformance liquid chromatography (HPLC) ICP-MS and X-ray absorption spectroscopy (XAS). Invertebrates were found by HPLC ICP-MS to contain predominantly arsenite and arsenate in methanol/water extracts, while XAS revealed that most arsenic is bound to sulfur in vivo. Examination of the spatial distribution of arsenic within an ant tissue highlighted the differences between exogenous and endogenous arsenic, as well as the extent to which arsenic is transformed upon ingestion. Similar arsenic speciation patterns for invertebrate groups were observed across sites. Trace amounts of arsenobetaine and arsenocholine were identified in slugs, ants, and spiders.
Introduction Arsenic is a naturally occurring constituent in many minerals (1). Arsenic can become available to plants and animals for uptake through various natural processes and human activities. Natural sources of arsenic include volcanic eruptions and weathering of rocks and soil; anthropogenic sources include mining, smelting, and the use of arsenic-containing pesticides in agriculture. The chemical form, or species, that an organism encounters in the environment determines how much arsenic gets incorporated into its tissues and whether or how it will be metabolized (1, 2). Arsenic species are broadly categorized in this paper as inorganic (containing only oxygen or sulfur bound to arsenic) or organic (containing an arsenic-carbon bond). Two oxidation states of arsenic are commonly found in water and biological tissues (+3 and +5). The aqueous inorganic forms arsenite (As(OH)3, As(III)) and arsenate (H2AsO4-, As(V)) are toxic and interconvert with changes in redox conditions and * Corresponding author phone: (613) 541-6000, x6161; fax: (613) 541-6596; e-mail:
[email protected]. † Royal Military College of Canada. ‡ Simon Fraser University and Advanced Photon Source. 4818
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 13, 2009
pH (1). Terrestrial organism extracts contain mostly inorganic arsenic and simple methylated organoarsenicals (3-5), but trace amounts of the nontoxic organoarsenical arsenobetaine (AB) have been identified in a number of matrices, including earthworms, ants, and birds (3-5). The methylation of inorganic arsenic occurs to varying extents in most organisms through a well-known pathway first proposed by Challenger (1), whereas the origin of AB in terrestrial organisms remains elusive (6). The source of organoarsenicals, especially more complex ones like AB, whether through diet or de novo synthesis by an organism, is still largely unknown in the terrestrial environment. Terrestrial invertebrates are eaten by many higher organisms such as frogs, birds (5, 7), and shrews, and thus form an important part of the terrestrial food web. However, little is known about the degree to which invertebrates take up arsenic from contaminated sites and what forms of arsenic are found within invertebrates in both natural and contaminated field settings. From laboratory studies, the toxicity of different arsenic species is known to follow the same trend in invertebrates as in other organisms: As(III) > As(V) > pentavalent dimethylarsinic acid (DMA) (8, 9). Only two studies on invertebrate metabolism have been carried out and both suggest invertebrates do not biotransform arsenic in vivo: bark beetles ingesting an arsenic pesticide (monosodium methanearsonate) contained >90% of the compound unmodified (10), and drosophila did not methylate inorganic arsenic (11). Arsenic speciation is often characterized using highperformance liquid chromatography inductively coupled plasma mass spectrometry (HPLC ICP-MS) methods. AB is the major arsenical present in the marine environment and this compound, along with a number of other organoarsenicals, is easily and completely extracted with methanol/water (1). In terrestrial organisms, however, characterizing arsenic speciation is hampered by poor extraction efficiencies (EEs) with conventional aqueous methanol extraction methods (99%, Fluka) were used as chromatography standards. DORM-2 (dogfish muscle certified by the National Research Council of Canada) was used as the certified standard reference material (SRM) for invertebrate extractions. Orthophosphoric acid (85%, Fluka), pyridine (>99%, Aldrich), and ammonium hydroxide solution (Fluka) were used to make cation and anion chromatography mobile phases. Rhodium AAS standard solution (in 5 wt % HCl, Aldrich), indium and uranium (10 000 ppm) ICP-MS standard solutions (PlasmaCAL) and Tune A (Thermo Electron Corporation) were used to make internal standard solutions. Nitric acid (Fisher Scientific), hydrochloric acid (Fluka), and hydrogen peroxide (30%, Caledon) were used for total digestions. Methanol (HPLC grade, Aldrich) was used for speciation extractions. Study Sites. Montague (44°71′57′′ N, 63°52′26′′ W), Upper Seal Harbour (USH, 45°12′06′′ N, 61°38′12′′ W), and Lower Seal Harbour (LSH, 45°10′03′′ N, 61°35′53′′ W) are three historical gold mine tailings sites in Nova Scotia, Canada with high concentrations of arsenic (21). All three contaminated sites are largely unvegetated in the center of the pHcircumneutral (22) tailings ponds (exposed tailings at LSH and Montague and submerged tailings in the creek at USH), surrounded by mixed forest. The background site (East Brook, 45°09′44′′ N, 61° 34′03′′ W) is located near the two Seal Harbour sites and has vegetation similar to the three contaminated sites but has not been affected by mining (see Supporting Information for map). Samples and Collection Methods. Plant and soil samples were collected along the tailings ponds edges. Plants representative of the site were hand collected, washed with ddH2O, and blotted dry using a Kimtowel. Plants were divided into stem, leaf, and flower and stored frozen prior to analysis. Soils were collected at 0-15 cm (plant root depths) at each sampling location and stored frozen prior to analysis. All invertebrates were collected in late summer 2007. A malaise tent, pitfall traps, and a CO2 mosquito trap (SkeeterVac SV-35 Mosquito Exterminator) were placed at the edge of the vegetated areas surrounding the mine tailings, and a sweep net was used throughout the sampling sites to collect invertebrates. Multiple collection methods were used to maximize the variety of terrestrial invertebrates sampled. Slugs were depurated for 24 h at 4 °C. All invertebrates were frozen after collection, sorted at room temperature, and stored frozen at -20 °C. Invertebrates were not rinsed, with the exception of ants which were rinsed 2-3× with water to remove soil particles during sorting. Organisms were sorted as follows: spiders (order Araneae), Formica sp. ants (order Hymenoptera), slugs (order Pulmonata), dragonflies (order Odonata), grasshoppers and katydids (order Orthoptera), mosquitoes (order Diptera), caterpillars (order Lepidoptera, larval form), moths (order Lepidoptera, mature), bulk flies (order Diptera), and finally a bulk category for the remaining organisms (bulk samples contained a wide variety of smaller invertebrates, with the majority of the biomass comprising organisms from class Insecta, order Hemiptera, which
organism
Montague
Lower Seal Harbour
bulk grasshoppersd antse spiderse caterpillarsd mothsd dragonfliese slugsf fliese
57 ( 74 1.7 -c 8.5 11 22 22 53 -
15 ( 2 1.1 ( 0.8a 5 ( 2b 11 8.9 12 25 ( 3a -
a
a
Upper Seal Harbour
background
13 1.2 11 5.1 9.8 19 1.7 ( 0.2a 1.1
2.4 0.16 0.54 0.1 0.21 0.13 0.61
a ( Indicates standard deviation of replicate analysis results (n ) 2). b ( Indicates standard deviation of replicate analysis results (n ) 4). c “-” Indicates insufficient sample for analysis. d Herbivore. e Carnivore. f Omnivore/detrivore.
includes aphids, leafhoppers, and shield bugs). Insufficient sample was collected for all groups to be analyzed from all sites. Identification of Arsenic Compounds. Invertebrate samples (0.5-2.0 g wet) were homogenized, using either a mortar and pestle or a Tissue Tearor prior to extraction. Samples were extracted using 1:1 methanol/water in a 20:1 solvent-to-sample wet weight ratio. Samples were mixed with a vortex, shaken for 30 min at 275 rpm at 5 °C (Innova 4230, New Brunswick Scientific), sonicated for 20 min, and centrifuged for 10 min (3500 rpm, 3120g). The supernatant was filtered (Whatman P541) into a roundbottomed flask. The vortex, sonication, and centrifugation steps were repeated twice. The extract was rotovapped (R-124 Buchi Rotovaper) and transferred to a 15 mL centrifuge tube for a final extract volume of 1.5-2.5 mL. The sample was again centrifuged and filtered (0.45 µm Millipore Millex-HV Hydrophilic PVDF) prior to analysis. The samples were stored frozen and were analyzed within 2 weeks of extraction. Plant samples were extracted by Analytical Services Unit (ASU), Kingston, ON, using a procedure in which approximately 0.25 g of dried and ground plant was extracted first with 10 mL of 1:1 methanol/water and the residue was then extracted with 10 mL of 0.1 M hydrochloric acid resulting in two extracts per sample. The procedure is described in detail in Mir et al. (23). Determination of Total Arsenic. For invertebrate total arsenic analysis, 0.5-2.0 g wet weight of sample, or the residue from speciation extraction, was first digested in 10 mL of 70% nitric acid, then reboiled in 3 mL of 30% hydrogen peroxide as described in detail by Smith et al. (13). For extract total analysis, ∼1 mL of sample extracts was digested in 3 mL of 70% nitric acid, then diluted to 3-5 mL using ddH2O. Samples were analyzed using an X7 X-series, ICP-MS (Thermo Instruments) as described in Smith et al. (13). Total arsenic concentrations were derived from the addition of the extract and residue totals or from total sample digestion. All of the samples that were of sufficient size for analysis were analyzed for total arsenic, and results are listed in Table 1. Dried and ground plant stems, leaves, and flowers were analyzed for total arsenic by ASU as described in Mir et al. (23). Briefly, ashed and acid-digested samples were analyzed by ICP optical emission spectrometry or hydride generation atomic absorption spectroscopy. Soil arsenic analysis was carried out by neutron activation analysis using the SLOWPOKE-2 reactor at the Royal Military College of Canada as described by Smith et al. (13). Soil and plant total arsenic concentrations are reported in dry weight. Invertebrate arsenic concentrations are reported in wet weight. VOL. 43, NO. 13, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
4819
HPLC ICP-MS. Extracted arsenic compounds arsenite (As(III)), arsenate (As(V)), dimethylarsinic acid (DMA), pentavalent methylarsonic acid (MMA), AB, trimethylarsine oxide (TMAO), and arsenocholine (AC) were identified using anionand cation-exchange HPLC ICP-MS. Details of the setup are found elsewhere (13). The reporting limit (5 ppb) was based on the lowest quantifiable concentration of the arsenic standard used in the calibration curves. Thermo PLASMALAB v 2.5.5.290 was used to record data, while data analysis was performed using PLASMALAB and PEAKFIT 4.0 chromatography software. A flow rate of 1.5 mL · min-1 was used for both anion and cation exchange analysis. Quality assurance/ quality control procedures included blanks, spikes, and standard reference materials, and all results were within acceptable limits. A few samples with increased hetereogeneity or with values close to the detection limits had replicate analyses outside limits but results were nevertheless accepted (details are provided in the Supporting Information). X-ray Absorption Spectroscopy (XAS). X-ray analysis was performed at the Pacific Northwest Consortium X-ray Operations and Research (PNC/XOR) facility, both on the bending magnet (BM) and insertion device (ID) beamlines at sector 20 of the Advanced Photon Source in Argonne, Illinois. The experimental setup is described by Smith et al. (13, 24). All samples were placed between two layers of Kapton tape and subsequently affixed to a sample holder. Bulk analyses were performed on the bending magnet line under vacuum at -50 and -100 °C. Ten scans were collected and averaged before background-removal and normalization to edge jump. Examination of the X-ray absorption near-edge structure (XANES) spectra from first to last in each set showed no beam damage to the samples. Coarse imaging of fluorescence data was performed on the ID line and samples were cooled to -22 °C using an electric chiller. Fluorescence data were collected at 50 µm steps with a 0.3 s integration time for two-dimensional spectra. Spatially resolved XANES spectra are described in this paper as micro-XANES. XANES spectra of the arsenic K-edge (11868 eV) were fit within -20 to +30 eV from E0 using ATHENA software (25). The Si(111) double-crystal monochromators for ID and BM lines were calibrated using the first inflection point of the gold LIII absorption edge (11919.7 eV). A reference gold foil was measured simultaneously with samples. Data were compared to linear combinations of reference compounds previously shown to distinguish among seven different groups of arsenic compounds based on the position of the main peak feature (24). Frozen As(III) and As(V) and liquid As(glutathione)3 standards were used. Scorodite was used for As(V) in the ant mouth micro-XANES fitting. All fits were constrained to sum to 100% of measured arsenic. Arsenic distribution maps were produced using SURFER modeling software. Statistical Analysis. Statistical analyses were done using SYSTAT 10 software. Total arsenic concentration in plants, soils, and invertebrates were log10-transformed before conducting variance analysis (ANOVA) and pairwise comparisons using a Bonferonni correction.
Results and Discussion Total Arsenic Concentration. Soils, plants, and invertebrates were collected from three mine tailings ponds and one background site (see Figure 1). There are significant differences in soil arsenic concentrations among all of the three contaminated sites and the background site (ANOVA, F(3,79) ) 101.4, p < 0.001) (median soil concentrations LSH ) 310 ppm, USH ) 2900 ppm, Montague ) 10000 ppm, background ) 6.0 ppm) (26). Plant arsenic concentrations also differ significantly among sites (ANOVA, F(3,105) ) 77.473, p < 0.001) with the exception of LSH and Montague plants (p ) 1.00). However, the plant arsenic concentrations do not reflect 4820
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 13, 2009
FIGURE 1. Total arsenic concentrations in soil, plants, and invertebrates from three mine tailings sites and one background (Bkgd) site. Soil and plant data are reported in ppm dry weight and invertebrates are in ppm wet weight. Invertebrate samples include grasshoppers, spiders, moths, caterpillars, and bulk invertebrates. Error bars are based on standard error. the soil arsenic concentrations since soil arsenic is higher at USH than LSH, and the reverse is true for plant arsenic (26). Invertebrates collected at all four sites (grasshoppers, spiders, moths, caterpillars, and bulk samples of invertebrates) were compared, and significant differences were observed only between the background and contaminated sites (F(3,16) ) 11.503, p < 0.005) and not between contaminated sites (pairwise comparisons, p > 0.05, geometric mean [As]: background ) 0.25 ppm, LSH ) 7.2 ppm, USH ) 6.8 ppm, and Montague ) 11 ppm). These data show that in areas with elevated arsenic concentrations, soil arsenic concentrations alone may not predict arsenic concentrations in organisms living on the sites. Arsenic Speciation. Arsenic speciation in the soil and tailings at the contaminated sites was investigated previously by XAS and exists in mineral inorganic forms (predominantly scorodite, FeAs(V)O4) (21). XANES spectra of invertebrates at these soil locations indicate that whole organisms are storing up to 90% of arsenic as reduced arsenic coordinated to sulfur (As(III)-S) in vivo (see Table 2 here and Figure S3 in the Supporting Information). As(III)-S species are likely indicative of arsenic protein binding via cysteine residues, which has been observed previously in invertebrates (3, 18, 27) where >80% of the total arsenic in moths fed sodium arsenate (As(V)) was stored as As(III)-S (18). As(III)-S complexes have also been found in plants, and were also not organoarsenicals; that is, no As-C bond was identified (12, 17). The nature and function of most As(III)-S compounds in terrestrial organisms is unknown. The high proportions of As(III)-S compounds, which may provide strong binding of inorganic arsenic in vivo, may be the reason for low extraction efficiencies, using conventional methanol/water methods, in terrestrial organisms. In recent years formic acid extraction schemes have been developed to nearly quantitatively extract arsenic from plants, identified as phytochelatin As(III)-S compounds (12). Future testing of the applicability of these extraction schemes to other As(III)-S-containing terrestrial samples, such as invertebrates, would be an interesting avenue of research.
TABLE 2. X-ray Absorption Near-Edge Structure Fitting Results for Selected Organisms from Contaminated Sitesa organism LSH grasshopperf Montague slugh LSH spiderg Montague spiderg Montague mosquitog LSH antg (sonicated) LSH ant (original) LSH ant mouth (1)e LSH ant body (2)e
% As(III)-S % As(III) % As(V) 29 55 83 87 86 52 ( 11b 30 ( 3c 34 51 ( 8d
15 16 n/a n/a n/a n/a n/a n/a 28 ( 8d
56 29 17 13 14 48 ( 11b 70 ( 3c 66 21 ( 1d
reduced χg 0.007 0.001 0.002 0.002 0.002 0.006, 0.014 0.004-0.007 0.002 0.005-0.014
a Frozen standards were used for fitting except for the ant mouth, where solid scorodite As(V) was used. As(III)-S indicates an As(glutathione)3 standard. As(III) is arsenic trioxide. As(V) is sodium arsenate. n/a indicates standard was not used in the fit. Smaller reduced χ2 are better fits. Spectra and fits are included in S3. b SD are given when samples or location results were averaged (n ) 2). c SD are given when samples or location results were averaged (n ) 3). d SD are given when samples or location results were averaged (n ) 4). e These two samples were analyzed by micro-XANES where all other samples were analyzed by bulk XANES. f Herbivore. g Carnivore. h Omnivore/detrivore.
Varying amounts of As(III) and As(V) were also identified by XANES, with less than 30% As(V) in mosquitoes, spiders, and slugs but higher percentages of As(V) in grasshoppers and ants. The higher results in grasshoppers were likely attributable to the grinding step used to prepare the sample (other samples were not ground); previous work by Smith et al. indicates that reduced arsenic can be oxidized during grinding (14). The reason for higher As(V) in ants will be discussed later. According to previous research, methanol/water appears to be an effective extraction solvent for organoarsenicals (1, 23), and therefore this method, combined with HPLC ICP-MS speciation analysis, was used to obtain information about trace concentrations of organoarsenicals in invertebrates in the present study. The results, reported in Table 3, were pooled across the three contaminated sites because of similarities in the observed speciation patterns and the statistical similarity observed in total arsenic concentrations. Around 20% of total arsenic was identified in this way, and mostly as inorganic arsenic within invertebrates, with limited organoarsenicals. Plant tissues are a food source for invertebrates, and samples from USH and LSH were analyzed by HPLC ICP-MS. As(III), As(V) DMA, and MMA were identified in extracts and 80% of the arsenic species identified in invertebrate extracts. TMAO and AB was absent in plants, but TMAO was identified in all invertebrate subgroupings and AB was identified in four organism groupings. Two groups of carnivorous invertebrates (spiders and mosquitoes) contained 11-19% of the total extracted arsenic as organoarsenicals TMAO and AB. This accounts for
