PFAS accumulation in several terrestrial plant and invertebrate species reveals species-specific differences

Abstract

Despite the known persistence and bioaccumulation potential of perfluoroalkyl substances (PFAS), much uncertainty exists regarding their bioavailability in the terrestrial environment. Therefore, this study investigated the influence of soil characteristics and PFAS concentrations on the adsorption of PFAS to soil and their influence on the PFAS bioavailability to terrestrial plants and invertebrates. PFAS concentrations and profile were compared among different invertebrate and plant species and differences between leaves and fruits/nuts of the plant species were assessed. Soil concentrations were primarily affected by organic carbon content. The PFAS accumulation in biota was, except for PFOA concentrations in nettles, unrelated to the soil concentrations, as well as to the soil characteristics. The PFAS profiles in soil and invertebrates were mainly dominated by PFOA and PFOS, whereas short-chained PFAS were more abundant in plant tissues. Our results show that different invertebrate taxa accumulate different PFAS, likely due to dietary differences. Both long-chained and, to lesser extent, short-chained PFAS were observed in herbivorous invertebrate taxa, whereas the carnivorous invertebrates only accumulated long-chained PFAS. Correlations were observed between PFOA concentrations in herbivorous invertebrates and in the leaves of some plant species, whereas such relationships were absent for the carnivorous spiders. It is essential to continuously monitor PFAS exposure in terrestrial organisms, taking into account differences in bioaccumulation, and subsequent potential toxicity, among taxa, in order to protect the terrestrial ecosystem.

Appendices

Appendix 1

Sample pretreatment and extraction

The soil samples were oven-dried for 3 days at 60 °C prior to the analyses. The invertebrate and plant samples (leaves and soft fruits) were homogenized by using a TissueLyser LT (Qiagen GmbH, Germany) with stainless steel beads (5 mm) after placing them in liquid nitrogen. The shells were removed from the acorns, and the shell and nut were analyzed separately. The nut and the shell were cut into small pieces (± 1 mm) using stainless steel scissors. The following analyses were performed on approximately 400 mg of dried soil, the whole body of invertebrates (only the soft tissue was used for the snails; shells were removed prior to homogenization), single leaves, and fruits/nuts. Three replicates per matrix per site were analyzed.

Each sample was spiked with 10 ng of a mass-labelled perfluoroalkyl carboxylic acid (PFCA) and perfluoroalkyl sulfonic acid (PFSA) mixture (MPFAC-MXA, Wellington Laboratories, Guelph, Canada), containing seven mass-labelled PFCAs (C₄, C₆, C₈, C₉, C₁₀, C₁₁, and C₁₂) and two mass-labelled PFSAs (C₆ and C₈). Hereafter, 10 mL of acetonitrile (ACN, Acros Organics BVBA, Belgium) was added to the samples. After vortex-mixing, the samples were sonicated (3 × 10 min, with vortex-mixing in between periods, Branson 2510) and left overnight on a shaking plate (135 rpm, room temperature). The samples were then centrifuged (4 °C, 1037 × g, 10 min, Eppendorf centrifuge 5804R) and the supernatant was transferred to a 15-mL PP tube.

The soil samples then followed a protocol described by Groffen et al. (2019b). Chromabond HR-XAW Solid Phase Extraction (SPE) cartridges (Macherey–Nagel, Germany) were conditioned and equilibrated with 5 mL of ACN and 5 mL of Milli-Q (MQ; 18.2 mΩ, TOC: 2.0 ppb, Merck Millipore, Belgium), respectively. After loading the samples, the cartridges were washed with 5 mL of a 25 mM ammonium acetate solution (dissolved in MQ; VWR International, Belgium) and 2 mL of ACN. Finally, the cartridges were eluted with 2 × 1 mL of a 2% ammonium hydroxide solution (dissolved in ACN; Filter Service N.V., Belgium). The eluent was dried completely using a rotational-vacuum-concentrator (Eppendorf concentrator 5301) and reconstituted with 200 μL of the 2% ammonium hydroxide solution. After vortex-mixing for at least 1 min, the samples were filtered through an Ion Chromatography Acrodisc 13 mm syringe filter with 0.2 μm Supor (polyethersulfone; PES) membrane (VWR International, Belgium) into a PP auto-injector vial.

The invertebrates and plant samples were further extracted following a modified procedure described by Powley et al. (2005). The supernatants were dried to approximately 0.5 mL in the rotational-vacuum-concentrator. To eliminate pigments, the concentrated extracts were then transferred to a PP Eppendorf containing 0.1 mL of graphitized carbon powder (Supelclean ENVI-Carb, Sigma-Aldrich, Belgium) and 50 μL of glacial acetic acid. In addition, 2 × 250 μL of ACN, used to rinse the 15 mL tubes, was added to the Eppendorf tubes. After vortex-mixing for at least 1 min, the samples were centrifuged (4 °C, 10 min, 9279.4 × g, Eppendorf centrifuge 5415R) and the supernatant was then treated equally as the eluent from the previously described method for soil.

UPLC-TQD analysis

Ultra-performance liquid chromatography coupled tandem ES(-) mass spectrometry (UPLC-MS/MS, ACQUITY, TQD, Waters, Milford, MA, USA) was used to analyze the different analytes. As target analytes, we selected eleven PFCAs (C₄–C₁₄), and four PFSAs (C₄, C₆, C₈, and C₁₀). The analytes were separated using an ACQUITY BEH C18 column (2.1 × 50 mm; 1.7 μm, Waters, USA) and an ACQUITY BEH C18 pre-column (2.1 × 30 mm; 1.7 μm, Waters USA) was inserted between the solvent mixer and the injector to retain any PFAS contamination originating from the system. The flow rate was 450 μL/min with an injection volume of 6 μL. As mobile phase solvents, we used 0.1% formic acid in water and 0.1% formic acid in ACN. The solvent gradient started at 65% of the 0.1% formic acid solution in water, decreased to 0% of this solution in 3.4 min and returned to 65% at 4.7 min. We used multiple reaction monitoring (MRM) of two diagnostic transitions per target analyte to identify and quantify the target analytes. The MRM transitions, cone voltages, and collision energy of each analyte, including the ISTDs, are displayed in Table 8, and have been validated previously (Groffen et al. 2019b).

Determination of soil properties

The soil clay content (particles with a size < 2 μm) was assessed by using a Malvern Mastersizer 2000 and Hydro 2000G. The samples (± 1 g ww) were pretreated with 25 mL of 33% hydrogen peroxide and 10 mL of 30% of hydrochloric acid to digest organic material and iron conglomerates in the soil. To speed up this digestion process, the samples were boiled and eventually sieved over a 2.0-mm test sieve, prior to the analyses.

The organic carbon content (TOC) was determined using the loss on ignition (LOI) method as described by Heiri et al. (2001). Approximately 10 g dw of soil was added to aluminum-foil bags that have been dried at 105 °C for at least 2 h. Hereafter, the samples were dried at 105 °C for at least 1 day. After cooling down in a desiccator, the dry-weight of the soil samples was determined, and the samples were incinerated in a muffle furnace at 550 °C for at least 5 h. Finally, after cooling down in a desiccator, the weight loss was determined, and the TOC was calculated using the following equations:

$${LOI}_{550}\left(\%\right)=100* \frac{\left({DW}_{105}-{DW}_{550}\right)}{{DW}_{105}}$$

$$TOC \left(\%\right)= \frac{{LOI}_{550}}{1.742}$$

with LOI as the loss on ignition after 550 °C, DW as the dry weight after drying at 105 °C or 550 °C, and 1.742 as the “Van Bemmelen” factor which assumes that 58% of the total organic matter is carbon (Nelson and Sommers, 1996).

The electrical conductivity of the soil was determined using a procedure described in ISO 11265:1994 (Soil quality—Determination of the specific electrical conductivity). Fresh soil was extracted for 30 min using demineralized water (1:5 m/V) at room temperature to dissolve the electrolytes. The conductivity was measured in the extracts using a conductivity meter (WTW Multi 3430 SET F, TetraCon 92 probe, Weilheim, Germany).

Finally, we determined the cation exchange capacity (CEC) in the soil samples. The soil moisture content was determined in oven-dried samples. To approximately 2 g of oven-dried soil, 25 mL of 1 M ammonium acetate (pH 7, adjusted with ammonium hydroxide) was added. Two procedural blanks (25 mL of 1 M ammonium acetate) were included as quality control. After three-dimensional shaking of the tubes for 1 h, the pH of the extract was measured, and the extracts were filtered using a 0.45-μm Mixed-Cellulose-Ester syringe filter. Two titration curves were made by adding steps of 0.1 mL of 0.1 N acetic acid to 25 mL of 1 M ammonium acetate (pH 7) up to a total volume of 10 mL acetic acid. After each addition of acetic acid, the pH was measured using a WTW Multi 3430 SET F multimeter (probe SenTix 940, Weilheim, Germany). Based on the titration curves, which provide information on the concentration of H⁺, and the moisture content of the samples, the exchangeable acidity was determined. The exchangeable bases and acidic cations were determined by analyzing the concentrations of aluminum, calcium, iron, potassium, magnesium, manganese, and sodium in the extracts using Inductive Coupled Plasma-Optical Emission Spectrometry (ICP-OES). The CEC was then calculated as the sum of the exchangeable acidity, bases, and acidic cations.

Appendix 2

Table 8 MRM transitions, internal standards (ISTDs), cone voltages (V), and collision energy (eV) for the target perfluoroalkyl substances and their internal standards. Table adopted from Groffen et al. (2019b)

Table 9 Individual limits of detection (LOD) and limits of quantification (LOQ) of the targeted analytes, determined in matrix, for soil (ng/g dw) and biota (ng/g ww)

Table 10 Response variables, explanatory variable(s), AIC values, and R² values of the final multiple regression models. The most suitable model was selected based on the AIC value. Regarding the plants, the concentrations were related to those in soil, taking into account the TOC in case of PFOA concentrations, and TOC and CEC in case of PFOS concentrations (as motivated in chapter 2.4). In the models of the invertebrates, besides the soil, also the various plant species (both leaves and fruits were included)