Microplastics increase the toxicity of mercury, chlorpyrifos and fluoranthene to mussel and sea urchin embryos ☆

The objective of this study was to determine whether and to what extent microplastics (MPs) enhance the toxicity of pollutants as well as whether pollutant-loaded MPs act as relevant vectors of chemical pollutants. With this aim, the toxicity for mussel and sea urchin embryos of: 1) three dissolved pollutants (Pol): chlorpyrifos (CPF), fluoranthene (FLT) and mercury (Hg); 2) their mixture with Microplastics (MP + Pol); and 3) pollutant-loaded MPs (MP Pol ), was assessed. Analyses of CPF, FLT and Hg were also performed to evaluate the transfer among dissolved and particulate phases. In general, the ’ MP + Pol ’ treatments were more toxic as 48-h EC 50 ( μ g/L) than the ’ Pol ’ treatments for sea urchin or mussel. The 48-h and 120-h EC 50 s ( μ g/L) for sea urchin showed little variation for CPF and MP + CPF, and no clear pattern was found for FLT and MP + FLT. The performed chemical analysis in the MP Pol tests indicated that desorption was the main route to explain the observed toxicity of Hg and a relevant route for CPF and FLT. This study contributes to improve the knowledge about the interactions be-tween MPs and chemical pollutants, which is fundamental for a more realistic ecological risk assessment in aquatic ecosystems.


Introduction
Microplastics (MPs) are plastic particles ranging from 1 μm to 5 mm, which can be classified in relation to their origin and use as primary or secondary.Primary MPs are produced for direct (e.g.abrasives, cosmetics) or indirect use (pellets for the production of polymers).Secondary MPs are originated from the fragmentation of larger plastics due to abiotic (photolysis, thermal-oxidation, hydrolysis) and biotic factors (e.g.mastication, bioturbation) (Browne, 2015;Frias and Nash, 2019;Galgani et al., 2015).The fate of MPs in the marine environment (sea surface, water column or sediments) is mainly driven by the original density of the polymer as well as by density changes due to fouling and weathering (Erni-Cassola et al., 2019;Rummel et al., 2017).
The toxicity of plastics is related, among other factors, to their composition and the size of the particles.The review by Beiras and Schonemann (2020) indicates that macroplastics showed lower toxicity, MPs intermediate toxicity and nanoplastics (<1 μm) higher toxicity.
Virgin plastics are materials with very low reactivity due to their chemical structure and molecular weight (Rochman, 2015).However, plastic additives (e.g., antioxidants, biocides, defoamers, flame retardants, heat stabilizers, plasticizers or UV stabilizers), incorporated into the plastic to improve the properties of the polymers, have been found to cause adverse biological effects at relatively low concentrations (Beiras et al., 2021;Bolívar-Subirats et al., 2021;Rochman, 2015).Some concern has been raised regarding the ability of MPs to act as vectors of pollutants (Hartmann et al., 2017).The sorption of a pollutant to the plastic is determined by the type of monomer, which rules its hydrophobicity and physicochemical characteristics; the particle size, which sets the surface/volume ratio and a greater sorption capacity at a smaller size; and the degree of weathering, due to increased porosity, surface area and presence of a microbial film (Crawford and Quinn, 2016).The sorption of pollutants can be reversible or irreversible and desorption kinetics can differ markedly from that of sorption (Fernández et al., 2020;Liu et al., 2018).Longer time is usually required to reach sorption equilibrium for the most hydrophobic organic pollutants than for the polar ones (Bakir et al., 2012;Endo et al., 2013;Karapanagioti and Werner, 2019).Comparatively, less information is available on the time needed to reach the equilibrium between MPs and metals (Fernández et al., 2020).Regarding desorption, polar chemicals are expected to desorb faster than apolar ones (Fernández et al., 2020;Liu et al., 2018).However, the physicochemical conditions of digestive fluids differ considerably from those of seawater and desorption may be greater in the digestive tract (Bakir et al., 2014;Liu et al., 2020).
Standard ecotoxicological tests designed for dissolved compounds have clear limitations when assessing suspensions of MPs.Suspensions of MPs in water are not stable because these particles have different density than water; although the use of a surfactant (e.g.Tween) or a plankton wheel at low speed allows obtaining homogeneous suspensions by compensating the negative or positive buoyancies of the MPs (Bellas and Gil, 2020;ECETOC, 2018;Gerdes et al., 2019).Among the possible routes of exposure to MPs (leaching, ingestion, inhalation, contact and entanglement), only some of them are considered in standard ecotoxicological tests (Enyoh et al., 2020).In this sense, adaptations of methods have been proposed to evaluate specifically some of these routes, such as leaching (Beiras et al., 2019b) or ingestion of particles (Gerdes et al., 2019).Mussel and sea urchin embryo-larval tests are of short duration (48 h) and very sensitive to low concentrations of dissolved toxicants (Beiras and Bellas, 2008;Bellas, 2006).These types of tests are limited to the endotrophic period but it is possible to increase its duration to cover the active ingestion of particles in the mixo-or exotrophic stage.
The overall objective of this study was to evaluate the toxicity of three model pollutants (chlorpyrifos -CPF-, fluoranthene -FLT-and mercury -Hg-), of their mixture with MPs and of pollutant-loaded MPs, on mussel and sea urchin embryos.These toxicants were chosen as representatives of three of the most relevant groups of pollutants in the marine environment: organophosphate pesticides (CPF), polycyclic aromatic hydrocarbons (FLT) and heavy metals (Hg), and are included in international monitoring programs such as those implemented by the EU Water Framework Directive, EU Marine Strategy Framework Directive, or Regional Seas Conventions (e.g.OSPAR Convention, Barcelona Convention).One of the specific objectives of this work was to assess whether MPs enhance or reduce the toxicity of chemical pollutants as well as whether pollutant-loaded MPs act as relevant vectors of chemical pollutants and whether toxic effects are observed.Several works have been carried out in this regard, but contradictory results have been found, depending on the type of MPs and contaminants studied (Wang et al., 2022;Yoo et al., 2022) or depending on the methodologies used (e.g., use of Tween).In order to progress on this topic, the present study aimed to infer the change in the bioavailability of the pollutants sorbed to the MPs, an aspect of the interaction between MPs and chemical pollutants that remains relatively unstudied.An additional objective was to assess whether the response observed varies from the endotrophic to the exotrophic stage and thus to improve the description of the toxic response to polluted MPs.

Experimental solutions and suspensions
The MPs used in this study were high density polyethylene micronized particles, named as MPP 635-XF (plain polyethylene with mean particle size of 3.29 μm, size range of 1.4-42 μm and density of 0.97 g/ mL) (Garrido et al., 2019) and AquaTex 325 (oxidized polyethylene with mean diameter of 4.6 μm and particles between 1 and 18 μm representing the 90% according to Fernández et al. (2020); mean particle size of 10-15 μm and maximum particle size of 44 μm and density of 0.99 g/mL according to the manufacturer), purchased from Micro Powders, Inc. Aquatex 325 showed a much higher affinity for metals than MPP 635-XF (Pinto et al., 2023).The reason for using two types of MPs (MPP 635-XF and AquaTex 325) is that plain polyethylene has a greater affinity for organic compounds, while oxidized polyethylene has a greater affinity for metals (Fernández et al., 2020;Garrido et al., 2019).
Three groups of treatments were evaluated in each experiment: solutions of the substance in 1-μm filtered seawater, sterilized with UV light and ozone (FSW); FSW without MPs (Pol), suspensions of MPs and the dissolved chemical (MP + Pol), and MPs loaded with the chemical (MP Pol ) (Table S1; Figure S1).The following control treatments were performed: FSW; a MP control, corresponding to the maximum concentration of MP used in the experiment but without toxicant (MP treatment); and acetone as a solvent control at a concentration of 1 mL/L (no toxic at this concentration according to Bellas (2006)) in CPF and FLT experiments.
For the treatment of toxicant without MPs (Pol), the required concentrations were obtained by diluting a standard Hg solution (1 g/L) or a stock solution prepared in acetone (FLT and CPF) with FSW.For organic chemicals, dilutions of the stock were prepared in acetone, in the range 0-1000 mg/L for CPF and 0-250 mg/L for FLT.Each concentration of the compound in acetone was dissolved in FSW at a ratio of 1 mL/L to obtain the experimental concentrations shown in Table S1.
Suspensions of MPs were prepared by adding MPs to 2 L glass bottles with FSW (Table S1; Figure S1).The suspensions were shaken manually and in an ultrasonic bath for 20 min.The maximum amount of MPs used in the experiments was previously estimated on the basis of the type of plastic, the amount of pollutant sorbed to microplastics in the MP Pol experiments and the capacity for exerting toxicity (Fernández et al., 2020;Garrido et al., 2019).
CPF, FLT or Hg were added to the MP suspension to obtain the maximum toxicant concentration (MP + Pol treatment) (Table S1; Figure S1).The mixture of MPs and toxicant prepared in 2 L bottles was subsequently diluted with FSW to obtain the required experimental concentrations (Table S1; Figure S1).
Two different procedures were used to prepare the suspension of MPs loaded with the toxicant (MP Pol ).For the experiments with organic pollutants, a suspension of MPs (1 mg/L) exposed to the maximum concentration of the toxicant for 2 h was filtered through a 0.22 μm Millex-GS filter (Merck KGaA, Germany).An incubation time of 2 h was chosen since previous studies by our group showed that after 2 h more than 70% of the CPF was adsorbed to the surface of the MPs (Garrido et al., 2019).Subsequently, MPs were resuspended in clean FSW, the concentration (number of particles/mL) in the eluate was quantified by means of a Coulter Counter (Beckman Coulter Inc., USA) and the concentrations in water were calculated (Table S1; Figure S1).Hg-loaded MPs (MP Hg ) were obtained by exposure for 7 days, since previous studies by our group showed that Hg adsorption to MP reaches an almost stable level after 7 days of exposure (Fernández et al., 2020).Following this, MPs were filtered, rinsed with ultrapure (Milli-Q) water, dried, weighed and added to FSW.The MP suspension was subjected to an ultrasonic bath before dosing into the vials and diluting with FSW.

Fluoranthene and chlorpyrifos
The concentrations of FLT and CPF in FSW were determined by stir bar sorptive extraction (SBSE) coupled with gas chromatograph with mass spectrometry detection (GC-MS) following García-Pimentel et al. (2023).Briefly, SBSE is applied to 100 mL sample with 100 g/L NaCl (Merck, Darmstadt, Germany) using commercial polydimethylsiloxane stir bars (20 mm × 0.5 mm) (Gerstel, Mulheim a/d Ruhr, Germany) shaken at 750 rpm for 22 h.Subsequently analytes were desorbed from the stir bar at 280 • C for 12 min, cryofocused in a PTV injector at 40 and analyzed by GC/MS in full-scan mode.This method shows a good linearity between 5 and 300 ng/L for both compounds tested.All samples, except blanks were previously diluted to obtain samples at the validated concentration range.The limit of quantification of fluoranthene and chlorpyrifos in seawater (100 mL) was 0.5 ng/L and 0.4 ng/L, respectively and a signal to noise ratio of 10 was established (García-Pimentel et al., 2023).Other contaminants, including other PAHs and plastic additives, were simultaneously determined with the same validated method.Blanks samples using oceanic seawater (sampled at 10 miles from the coast) were also used to check the analytical procedure.FSW samples were previously diluted with oceanic seawater (total volume: 100 mL) to get concentrations at ng/L level for their analysis by SBSE/GC/MS.

Mercury
Hg solutions and the suspension of MPs with Hg (MP + Hg) were taken in borosilicate glass vials to analyze the total content of this element.The samples were analyzed by using the pyrolysis atomic absorption spectrometry with gold amalgamation (LECO analyzer AMA254 model), according to the methodology described by Costley et al. (2000).In order to ensure the quality of the results obtained, experimental controls, analytical blanks and certificate material (NIST 2976 and NMIJ CRM 8112-a) were employed.The precision of the method, calculated by analytical blanks coefficient of variation, was 9%.The recovery of the certificate material was 110% for the NIST 2976 and 98% for the NMIJ CRM 8112-a, and the analytical detection limit was 0.01 ng.

Adsorption and desorption of CPF, FLT and Hg by MPs
The behavior of the pollutants in solution, sorption to the particles and desorption in conditions analogous to those of the toxicity tests were studied.FLT and CPF dilutions were prepared at the maximum concentrations tested by adding a standard dissolved in acetone (250 and 1000 μg/L, respectively) and a sample was frozen for later analysis.A suspension of MPs (1 mg/L) was prepared and FLT (250 μg/L) or CPF (1000 μg/L) were added.After 2 h exposure, MPs were filtered (GF/F) and the filter and the eluate were frozen for ulterior analysis.Finally, a suspension of virgin MPs with each pollutant was prepared at the same conditions indicated above.After 2 h exposure, MPs were filtered (0.22 μm Millex-GS), resuspended in 2 L of FSW, rotatory shaken for 48 h in the dark (9 rpm), and filtered (GF/F) again.The filter and eluate were frozen until analysis.
The adsorption of Hg to MPs and desorption from MPs to water was estimated based on the concentration of Hg in MPs determined by Fernández et al. (2020) in the adsorption and desorption phases.

Embryo-larval bioassays
Embryo-larval bioassays were performed using mussel (Mytilus galloprovincialis) and sea-urchin (Paracentrotus lividus) embryos.Gametes were obtained from mature adults in the laboratory and in vitro fertilization was conducted in FSW (salinity ranged from 32 to 35 ppt).The mussel test was performed according to Beiras and Bellas (2008).Two type of sea urchin tests were performed: the standard 48-h test, according to Saco-Álvarez et al. ( 2010), and a slight modification of this assay with an incubation time of 120 h.Sea urchin larvae acquires exotrophy at 72 h after fertilization (Grosjean et al., 1998), therefore by increasing the test duration to 120 h the contribution of ingestion to the toxic response can be investigated.

Mussel embryo-larval bioassay
Mature mussels were acquired from local markets and spawning was induced by thermal shock (+5 • C).Mature ova (spheric) were transferred to a 50 mL graduated cylinder containing FSW (~10 ova/μL) and 1 mL of diluted sperm was added.The suspension was gently stirred to facilitate fertilization.Fertilized eggs were transferred to 25 mL glass vials filled with no head space (40 eggs/mL).Each experimental concentration was performed in quadruplicate.The vials were placed in a rack fixed to a rotary wheel set at 9 rpm (20 ± 1 • C, darkness).Two drops of 40% formalin were added to vials after 48 h of incubation and 100 embryos/larvae per vial were recorded.Larvae were classified as abnormal or normal, based on His et al. (1997).The inhibition response was calculated as: where N i is the percentage of normal larvae obtained for the treatments with the toxicant and N 0 the percentage of normal larvae in the control.

Sea urchin embryo-larval bioassay
Adult sea urchins (Paracentrotus lividus) were collected from the field, maintained in flow-through system and dissected to obtain gametes.Maturity of gametes measured as ovum sphericity and sperm mobility was checked using an inverted microscope.The ova were transferred to a 50 mL graduated cylinder containing FSW (~10 ova/μL).A few drops of sperm were added and the proportion of eggs with a fertilization membrane was determined per quadruplicate (n = 100, >97%).Fertilized eggs were transferred to 25 mL glass vials filled with the experimental solutions and no head space.Each treatment was performed in quadruplicate.Egg density differed depending on whether the incubation was for 48 h (40 eggs/mL) or 120 h (10 eggs/mL).
The vials were placed in a rack fixed to a rotary wheel set at 9 rpm (20 ± 1 • C, darkness).Two drops of 40% formalin were added to vials after 48 h or 120 h of incubation.The maximum length of 35 individuals was measured in each vial.Growth inhibition was quantified as: where ΔL 0 is the length increase from eggs to larvae in the control and ΔL i is the length increase for the treatments.

Statistical analyses
The model used for describing dose-response relationships was the cumulative function of the Weibull distribution: where R is the response, C is the concentration, K is the maximum value of response, m is the concentration corresponding to the semi-maximum response and a is a shape parameter.Equation ( 3) was re-parameterized to estimate the slope at the median abscissa (v m ): The concentration corresponding to a certain level of effect was obtained by re-parameterizing Equation (3): where EC 50 is the concentration required to obtain a 50% effect.Parametric estimations were performed by minimisation of the sum of quadratic differences between observed and model-predicted values using the 'Solver' add-in of Microsoft Excel spreadsheet (Microsoft Corp., USA).Confidence intervals of the parameters were determined with the freely available 'SolverAid' macro and the consistence of mathematical models was assessed by a Fisher's F test.Parameter values obtained from two curves were compared using a Student's t-test (Motulsky and Christopoulos, 2003).
The extra sum-of-squares F test was used for comparing the entire curves obtained by fitting two sets of data with the same equation (Motulsky and Christopoulos, 2003).F statistic summarize goodness of-fit as residual sum of squares (SS) against the degrees of freedom (df) with the aim of detecting differences among the two curves.

Chemical analyses
The analysis of Hg showed a high correlation between nominal and measured concentrations for both Hg (R 2 = 0.9829) and Hg and MPs (MP + Hg) treatments (R 2 = 0.9958) (Figure S2), confirming that no precipitation of dissolved Hg occurred even at the highest concentrations.The measured concentration of dissolved FLT at time 0 (194 μg/L) was slightly lower than the nominal value (250 μg/L).For CPF, the difference between nominal (1000 μg/L) and measured value (399 μg/L) at 0 h was even greater, probably due to the limited solubility of this contaminant in seawater, sorption on recipient walls or other dissipation processes.

Adsorption and desorption of CPF, FLT and Hg by MPs
The concentration of toxicants sorbed to MPs after 2 h of exposure was high for both CPF (198*10^6 μg/kg) and FLT (105*10^6 μg/kg).
However, the concentration in MPs dropped markedly after 48 h in FSW (1.1*10^6 μg/kg for CPF and 1.4*10^6 μg/kg for FLT).The desorption estimated from these values would indicate water concentrations of 197 μg/L and 103 μg/L for CPF and FLT, respectively; well above the measured values (37 and 34 μg/L for CPF and FLT, respectively).
Probably, simultaneous dissipation processes (degradation, sorption and/or volatilization) can reduce their concentrations along the experiment.In fact, the occurrence of 3,5,6-trichloro-2-pyridinol (TCP), a degradation intermediate of CPF, was confirmed in the seawater samples (mainly at 48 h) using GC/MS full scan mode (99% of similarity of mass spectra with NIST library), evidencing the degradation of this pollutant at the tested conditions.
Hg followed a sorption kinetics to MPs with a plateau at 168 h (1211 mg/kg) and a rapid and stable desorption of Hg to FSW (760 mg/kg after 2.5 h).

Chlorpyrifos
The toxicity of dissolved CPF for mussels (Fig. 1A left, Table S2) and sea urchins (Fig. 1B left, Table S2) was lower than that of MP + CPF (Fig. 1A and B center, Table S2).For mussels, the 48-h EC 50 of MP + CPF (150.9 μg/L) was lower than that of CPF dissolved in water (224.9μg/L) according to a t-test (p = 0.0107); although the 48-h EC 10 values were similar for both treatments (Table S2).The slope of the toxicity curve (v m ) of the MP + CPF treatment (0.008 L/μg) for mussels was also greater than that of the dissolved CPF (0.003 L/μg) according to a t-test (p = 0.0063) (Table S2).
The EC 50 s at 48 or 120 h of MP + CPF for sea urchin showed slightly lower values (253.3 and 270.4 μg/L, respectively at 48 and 120 h) than that of dissolved CPF (269.6 and 284.4 μg/L), although the differences were not statistically significant (Table S2).The v m of MP + CPF (0.002 and 0.003 L/μg at 48 and 120 h respectively) were higher than those of dissolved CPF (0.001 and 0.001 L/μg respectively at 48 and 120 h) according to a t-test (p = 0.0168 at 48 h and p = 0.0115 at 120 h).
The toxicity curves of the treatments of dissolved CPF (Fig. 1B left) and MP + CPF (Fig. 1B central) for sea urchin showed little variation from 48 to 120 h.According to the F test, a single curve would allow the whole dose-response data for each treatment (CPF or MP + CPF) at 48 and 120 h to be adequately described.In fact, the values of the parameters (K, EC 50 , EC 10 , v m and a) obtained for each treatment (CPF or MP + CPF) did not show significant differences at 48 or 120 h.
The CPF-loaded MPs showed significant toxicity for mussels (48-h EC 50 : 1.56 mg MP/L) and sea urchins (EC 50 at 48 and 120 h: 0.79 and 0.84 mg MP/L, respectively), but to a lower extent than that of dissolved CPF or MP + CPF (Fig. 1 and Table S2).In the sea urchin experiments the curves at 48 and 120 h were homogeneous (Fig. 1B, right) and the parameters of the curves (EC 50 , EC 10 , v m , a) did not differ statistically (Table S2).The estimated toxic contribution of CPF desorbed from MPs at 48 h was 11.6%.

Fluoranthene
The toxicity of FLT treatments to sea urchin larvae differed slightly from 48 to 120 h and there was no clear pattern since: a) a reduction in toxicity with increasing exposure time was observed for dissolved FLT (Fig. 2 left) and b) an increase in toxicity with time was obtained for MPs treatments (MP + FLT and MP FLT ) (Fig. 2 center and right).The maximum response value (K) in the treatments was conditioned by the solubility of FLT, so in some cases it was not possible to calculate the EC 50 (Table S2).
The toxicity curves of FLT at 48 and 120 h showed significant differences in the F test.In fact, the EC 10 of FLT was lower at 48 h (42.5 μg/ L) than at 120 h (80.0 μg/L) according to a t-test (p = 0.000) (Table S2).
The K value at 120 h (0.46) was lower than at 48 h (0.54), although the differences were not significant according to a t-test; and the value of v m was very similar for 48 h (0.004 L/μg) and 120 h (0.004 L/μg) (Table S2).
The toxicity of MP + FLT was higher at 120 than at 48 h according to a F test (p = 0.0473) (Fig. 2 center).Some changes in toxicity parameters were observed in treatments from 48 to 120 h: an increase in K from 0.5 to 0.6 and a slight reduction in EC 10 , although not statistically significant.In contrast, the slope of the curve was very similar at 48 and 120 h (0.007 L/μg in both cases) (Table S2).
The pattern of response of FLT and MP + FLT was different at 48 and 120 h (Table S2 and Fig. 2 left and center).The toxicity curves of FLT and MP + FLT at 120 h were different according to the F test (p = 0.000).The difference in the toxicity curves at 48 h could be explained based on slightly different values in terms of slope or shape of the curve (not significant differences); although EC 50 , EC 10 and K values were similar (Table S2).The MP + FLT treatment at 120 h showed higher toxicity than FLT, denoted by a higher value of K (0.6) and v m (0.007 L/μg) compared to that of FLT (K = 0.46 and v m = 0.004 L/μg), as well as a significant reduction in the EC 10 value according to a t-test (from 80.0 to 30.5 μg/L, p = 0.0000) (Table S2).
FLT-loaded MPs showed an appreciable toxicity for sea urchin larvae which was slightly higher at 120 than at 48 h (Fig. 2 and Table S2).The most evident effects of changes with time (from 48 to 120 h) in toxicity parameters could be found in v m (from 0.48 to 0.65 L/mg MP) and EC 50 (from 1.29 to 0.99 mg MP/L) (Table S2).The estimated toxic contribution of FLT desorbed from MPs at 48 h was 21.6%.
For both mussel and sea urchin, the curves for Hg and MP + Hg were significantly different according to a F-test (p = 0.0000 and 0.0090 respectively).For both species the EC 50 or EC 10 values of Hg (Fig. 3 left) were higher than for MP + Hg (Fig. 3 center left, Table S3) and the slope of Hg was lower than that of MP + Hg (Table S3).However, significant differences were only shown in EC 50 , EC 10 and v m for the sea urchin experiment (p = 0.0002, 0.0020 and 0.0086 respectively).
Toxicity expressed as EC 50 for the Hg-loaded MPs was moderate for both sea urchin (42.48 mg MP/L) and mussel (4.93 mg MP/L) (Fig. 3 and Table S3).The EC 50 and EC 10 of the Hg-loaded MPs for sea urchin were 8.6 and 102 times higher than that of mussel.
The inhibitory response against the concentrations of dissolved Hg, added to the MP suspension or estimated from desorption of plastics to water, is shown in Fig. 3 (right).The EC 50 values in terms of desorbed Hg are 2.2 and 19.1 μg/L for mussel and sea urchin, respectively.For mussel, the curves of the three treatments were very similar, which would indicate a high homogeneity in terms of effects caused by Hg added or estimated by desorption.The curves corresponding to the three treatments differ slightly for sea urchin (Fig. 3

Discussion
There is growing concern about the relevance of the interactions of MPs with pollutants in their surrounding environment, and about the potential for these interactions to disturb the balance of the ecosystem.In the current study, it is shown that MPs can modulate the toxicity of selected relevant pollutants, namely CPF, FLT and Hg, to sea urchin and mussel embryos.The toxicity values of dissolved CPF, FLT and Hg for sea urchin, expressed as 48-h EC 50 , were: 269.6, 159.5 and 9.8 μg/L, respectively.These values are within the range of those previously obtained by other authors for CPF (300 μg/L, Bellas et al. (2005)), FLT (>250 μg/L, Bellas et al. (2008)) and Hg (22.0 μg/L, Fernandez and Beiras ( 2001)) using the sea urchin embryo-larval test.For mussel, the 48-h EC 50 of the dissolved compounds were: 224.9 μg/L for CPF and 2.0 μg/L for Hg; comparable to the values obtained by Beiras and Bellas (2008) for CPF (153.5 μg/L) and by Beiras and Albentosa (2004) for Hg (5.1 μg/L).The Log K ow value was similar for CPF (4.96) and FLT (5.16); although the surface sum over all polar atoms of a molecule (i.e. the Topological Polar Surface Area, TPSA), of CPF (TPSA 72.7 Å 2 ) was greater than that of FLT (TPSA 0 Å 2 ) (Pubchem web).This would indicate a greater ability of FLT to cross biological membranes and exert toxicity than of CPF (Prasanna and Doerksen, 2009); which was in moderate agreement with the 48-h EC 50 and 48-h EC 10 values obtained in the treatments of dissolved FLT and CPF for the sea urchin embryolarval test (Table S2).
In general, the 'MPs plus Pollutant' treatments were more toxic than the 'dissolved pollutant' treatments, and the toxicity of the pollutantloaded MPs was lower than that of the other two treatments.This was validated with the 3 toxicants tested (CPF, FLT and Hg) in two different biological models (mussel and sea urchin).Previously, Bellas and Gil (2020) had already concluded with a similar experimental design that the treatment of MPs plus CPF had greater toxicity for Acartia tonsa than dissolved CPF and CPF-loaded MPs.Several exposure routes may serve to explain the increase in toxicity observed for the treatment of MP plus toxicant: 1) leaching, the additives contained in the MPs are partially extracted during sonication and subsequent agitation and would sum to the toxic effects of the pollutants; 2) ingestion, not plausible for 48-h D-veliger and echinopluteus larvae because those stages are yet endotrophic; 3) dermal uptake from contact with the surface of the polluted MP, which would imply absorption into the dermis and subsequent diffusion; 4) the ability of the MPs to destabilize lipid membranes, altering their integrity and facilitating the permeability and entry of pollutants into the cells (Fleury and Baulin, 2021); and 5) MPs act as a passive doser of the pollutant, with sorption kinetics at high concentrations of the pollutant, desorption kinetics when the concentration of the pollutant dissolved in water is low and the subsequent dermal uptake by embryos/larvae of the dissolved pollutant in water.These routes are particularly relevant for the smallest fractions of MPs and NPs regarding to the transfer of contaminants to organisms.
Leaching would involve the partition of the plastic additives to the liquid phase, and toxicity would be expected if the concentration of additives contained in the leachate exceed a certain threshold (Beiras et al., 2021).Concentrations of the tested MP were below the LOEC determined for sea urchin (LOEC MPP635-XF = 100 mg/L, LOEC Aquatex-325 = 100 mg/L) and mussel (LOEC MPP635-XF >100 mg/L) (Beiras et al., 2018).However, it is possible that the leachates have additive effects to the dissolved pollutant, which would help to explain the increase in toxicity observed for MP + Pol treatments (Bellas, 2008;Silva et al., 2002).In fact, although the products used here are considered 'virgin' MPs, the occurrence of low concentrations of different compounds such as plastic additives and PAHs (e.g.benzophenone, tributyl-2-acetyl citrate, triphenylphosphate, phenanthrene) was found in seawater and filter extract samples (data not shown).
The The MP Pol tests suggested that desorption and subsequent dermal uptake was the main route to explain the observed toxicity of Hg and a relevant route for .Larval nutrition at 48 h is endotrophic for mussel and sea urchin and the digestive system is not fully functional (Fenaux et al., 1985;Lucas et al., 1986).Therefore, it was not plausible that MPs filtered by 48-h larvae were exposed to digestive fluids and resulted in an increase of bioavailability.This agrees with the results of Beiras et al. (2019a) who exposed sea urchin embryos and larvae <48 h to a mixed treatment of MPs and nonylphenol or methylbenzylidene-camphor and did not find an increase in toxicity or bioaccumulation with respect to the dissolved pollutant.Fig. 3 (right) shows that the toxicity curve of Hg desorbed from MPs to water was very close to those of dissolved Hg and MPs plus Hg.It was not possible to estimate an analogous toxicity curve for FLT or CPF.
The ability of MPs to sorb, desorb, or transport pollutants depends on the monomer, the additives, and the structure of the plastic.Therefore, the toxicity of pollutants loaded into the MPs would be related to their polarity or hydrophobicity, the nature of the plastic, and the reversibility/irreversibility of the interaction (Wang et al., 2018).In the present work, two different polyethylene MPs (plain and oxidized) and three pollutants (CPF, FLT and Hg) were tested and the concentration of pollutants were measured after sorption and desorption.The concentration of FLT and CPF loaded into the MP after 2 h (198*10^6 μg/kg for CPF and 105*10^6 μg/kg for FLT) was much higher than the concentration after exposing those pollutant-loaded MPs to clean water for 48 h (1.1*10^6 μg/kg for CPF and 1.4*10^6 μg/kg for FLT).At the end of the loading stage (2 h) 69% of the CPF was dissolved and 31% particulate and after 48 h of exposure to water more than 99% was dissolved.For FLT, the percentage in solution at the end of the loading stage (2 h) was 33% (67% particulate) and 99% after 48 h of exposure to clean water (1% particulate).Although the performed measurements were punctual, and not kinetic, the results obtained for FLT and CPF would be indicative of a reversible process with rapid accumulation and desorption.In fact, the desorption of both pollutants from different plastic polymers to seawater was confirmed in previous studies (León et al., 2019;León et al., 2018).
Hysteresis is the tendency of a material to retain one of its properties, in the absence of the stimulus that generated it.The hysteresis index for a type of MP and pollutant is defined as a ratio between the sorbed concentration at equilibrium after sorption and desorption kinetics (Song et al., 2021).The higher the hysteresis index, the more tendency MP has for the pollutant to be sorbed; while a negative index indicates that the sorption is reversible and the absence of hysteresis in desorption.The irreversibility of the sorption determines the capacity that a specific MP has to transport a certain type of pollutant.Polymers are classified as glassy or rubbery depending on their glass transition temperature, which determines its state at room temperature.Rubbery polymers (e.g.polyethylene) are easily accessible for partitioning hydrophobic organic pollutants, have high values of sorption coefficients (K d ) and low hysteresis indices (Wang et al., 2020).Glassy polymers (polystyrene or polyvinyl chloride) are denser, and have lower K d and desorption hysteresis (Wang et al., 2020).Liu et al. (2018) found that: a) the hysteresis index for pyrene (log K ow = 4.88, TPSA 0 Å 2 ) and 4-nonylphenol (log K ow = 5.76, TPSA 20.2 Å 2 ) was very low for polyethylene MPs and b) the desorption of these compounds was very fast (2 h).This agrees with the results of Song et al. (2021) who found that the K d of petroleum hydrocarbons followed the order: polyamide > polyethylene > polyethylene terephthalate > polylactic acid > polyvinyl chloride and that the hysteresis index of these MPs was less than 0. Therefore, the results obtained here would agree with low values of hysteresis (and transport capacities) for the tested combinations of MPs and pollutants.
The 120-h sea urchin embryo-larval tests covered the entire endotrophic period (72 h) and the onset of the exotrophic period (72-120 h); so this test was useful to evaluate whether the toxic response was related to an increase in concentration through the digestive tract.For CPF, the curves at 48 and 120 h were identical, suggesting that the main mechanism that explained the observed toxicity was desorption from the MP and not uptake through the digestive tract.In contrast, there were indications that the digestive route was possible for FLT, but of less relevance.On one hand, a slight increase in toxicity was observed at 120 h in both MP + FLT and MP FLT compared to 48 h (Fig. 2 center and right).On the other hand, the concentration of FLT in the particulate fraction was 67% after loading for 2 h (30% for CPF) and 1% remaining in the MPs after 48 h in clean FSW (0.5% for CPF).It should be noted that digestion in sea urchin larvae is basic (pH 8.9-9.6)(Stumpp et al., 2013); therefore, the desorption under these conditions was possibly lower than that of acid digestion (Bakir et al., 2014;Liu et al., 2020).Additionally, it is necessary to consider that the simultaneous degradation process along the experiment can reduce the exposure concentration.Further experiments, such as bioaccumulation tests, need to be conducted to address the digestive route (Asmonaite et al., 2020;Bour et al., 2020).
This experimental approach served to suggest plausible toxicity pathways of polluted MPs or the mixture of MPs and pollutant.However, it is necessary to note that the concentrations of MP and pollutant used greatly exceed those observed under natural conditions.For instance, MP can reach concentrations up to 900-15600 MP/m 3 in polluted areas (Gorokhova, 2015;Kang et al., 2015;Zhao et al., 2015), whereas seawater concentrations of CPF, FLT and Hg of 0.20, 2.9 and 0.074 μg/L, respectively, have been reported in estuaries and impacted coastal areas (Beiras et al., 2002;Benson et al., 2014;Campillo et al., 2013).Therefore, the sorption and desorption of the dissolved pollutants to the MP in the natural environment occur under different conditions than those investigated here.It is estimated that most of the plastics in the ocean are in equilibrium with the concentrations of dissolved organic pollutants (Koelmans et al., 2021).Thus, the experimental conditions illustrated here would be only applicable in certain scenarios such as the spill of a pollutant (far from equilibrium conditions) and it could be applicable to study the ability of environmental plastics to sorb and release pollutants.The relevance of the interactions of pollutants with MPs has been confirmed in relation to their toxicity on invertebrate embryos, evidencing the role of desorption from plastics in this process.

Conclusions
The experimental design used here served to investigate the toxicity of dissolved pollutants with MPs as well as pollutant loaded-MPs and to discriminate the toxicity mechanisms.In general, the 'MP + Pol' were more toxic than 'Pol' treatments.It is plausible that the increase of toxicity in 'MP + Pol' treatments is due to the leaching of the additives contained in the MPs which increase the adverse effects provoked by the dissolved pollutants alone.The treatments with MP Pol showed toxicity in the μg MP/L to mg MP/L range.The analysis of polluted MPs after having put them in clean seawater showed that desorption is the main mechanism that explains Hg toxicity and a relevant mechanism for CPF and FLT.This study provides new knowledge about the interactions between MPs and relevant chemical pollutants, which is essential for better understanding the impact of MPs and associated pollutants in aquatic ecosystems.

D
.Rial et al.

Fig. 1 .
Fig. 1.Toxic response as enhancement of abnormalities in mussel embryos (A) or inhibition of larval growth in sea urchin (B) by nominal concentrations of chlorpyrifos (CPF, left), polyethylene microplastics MPP 635-XF and chlorpyrifos dosed jointly (MP + CPF, center) and microplastics exposed for 2 h to chlorpyrifos, filtered and dosed into the test vials with clean FSW (MP CPF , right).Response at 48 (ᅳ○ᅳ) and 120 (⋯•⋯) h.Error bars as standard deviation.

Fig. 3 .
Fig. 3. Toxic response as potentiation of abnormalities in mussel embryos (Top) or inhibition of sea urchin larval growth (Bottom) by nominal concentrations of Hg (left, ᅳ○ᅳ), Aquatex 325 polyethylene microplastics and Hg dosed in test vials (MP + Hg, center-left, --▢--), microplastics exposed for seven days to mercury, filtered and dosed into the test vials (MP Hg , center-right, ⋯△⋯) and dissolved or desorbed mercury from the microplastics (right).Response at 48 h and error bars as standard deviation.

D
.Rial et al.
concentrations of pollutant-loaded MPs (MP Pol ) that result in toxic effects were in the μg MP/L to mg MP/L range.For sea urchin, the