Introduction

The use of free-living organisms as bioindicators of contamination has been proposed in numerous studies [1–4] since they can reflect the effect of pollutants on cellular metabolism and global homeostasis. Generally, environmental stress situations are marked by different metabolites (biomarkers), such as cytochromes P450, glutathione-S-transferases, metallothioneins, superoxide dismutases (SODs), glutathione peroxidases, peroxiredoxins, and many others [5, 6]. However, the use of biomarkers requires a deep knowledge of their toxicity mechanisms and only provides a biased assessment about the contamination process, since the biological response is defined by a limited and pre-established number of metabolites excluding others whose relationship with pollution is unknown. Therefore, the use of massive information approaches, the -omics, is the promising alternative to biomarkers in environmental contamination assessment [7, 8]. Metallomics is one of the most recent -omic [9] which uses metals or metalloids, present in one third of biomolecules in cells, as heteroatomic markers or tags to track these molecules in complex biological matrices [10]. In these approaches, the use of high-sensitivity atomic detectors, mainly inductively coupled plasma mass spectrometry (ICP-MS) [11] generally coupled to a chromatographic module (in single or multidimensional arrangements), and mass spectrometry for parallel biomolecules identification in an integrated workflow [9, 11, 12] is fundamental. Metallomics provides a good alternative to deep insight into the fate of elements in exposed organisms and gives information about metal trafficking, interactions, and homeostasis. For example, Se [13] and Zn [14] have been cited by the protective action against toxic metals such as Hg or Cd, and Mn in the cellular defense against reactive oxygen species [15] and cancer protection [16]. In addition, metallomics allows the characterization of a variety of trace element species in cells, body tissues, and fluid extracts using the same experimental setup [17], and this multielemental detection provides in-parallel information about elements with interconnected metabolic functions, such as Cu, Zn, Cd, and Pb in human brain cytosol [18]; Fe, Mn, Cu, and Zn in fresh porcine liver extracts [19]; and Cu, Zn, Mn, Fe, and Cd in a bovine liver reference material [20]. Other examples are the identification and determination of metallothionein metal-complexes induced in liver of rat exposed to Cd by intravenous injection [21]; the study of the role of different metallothionein isoforms to sequester Cd, Cu, and Zn in eel liver after exposure to Cd in aquarium tanks [22]; and the speciation of Cd, Cu, and Zn with metallothionein-like proteins in snails induced by Cd exposure [23].

The mouse Mus spretus is an aboriginal species frequently used as bioindicator in the southwest Spain, especially in Doñana National Park (DNP), which is an important European biological reserve for migrating birds. However, the Park is affected by agricultural, mining, and industrial activities [2, 24] that make necessary the performance of regular monitoring of the environment quality based on the analysis of pollutants and biological responses to them. The application of proteomics for this purpose using the mouse M. spretus as bioindicator has represented a good option [25], especially by the demonstrated genetic homology of this mouse with the classical inbred laboratory mouse Mus musculus that has been already sequenced [26]. This fact allows the use of database from M. musculus in protein and metalloprotein identification of M. spretus [25] as well as in transcriptomics studies [26], avoiding the cumbersome work associated with de novo sequencing.

The coupling size-exclusion chromatography with inductively coupled plasma mass spectrometry (SEC-ICP-MS) has been used for size characterization of metal species in liver and brain extracts of M. spretus from contaminated area in comparison with the laboratory mouse M. musculus used as control [27]. The present work explores the use of this metallomic approach to compare the biological response of M. spretus living in five different areas of Doñana National Park which are affected by differential contamination due to the presence of organic matter, dissolved inorganic salts, and substances rich in nitrogen and phosphorous [28], as well as pharmaceutical active compounds [29], nitrate from agricultural sources [30], and metals from the Iberian Pyrite Belt [31]. Previous proteomic studies performed in this area using the crab Procambarus clarkii [2] revealed upregulation of proteins in sites with contamination and inhibition of enzyme esterase, a biomarker that denotes the presence of pesticides. In addition, the presence of non-essential elements (such as Cd and Hg) in tissues of shrews associated to mine activities [32] and DNA damage caused by pollution, as denoted by Comet assay in blood of M. spretus [33], have been detected. In the present study, high-metabolic-activity organs such as liver and kidneys have been used to establish the differences in the intensities of metal-binding biomolecules in connection with environmental contamination issues, and interactive relationships among metabolically active metals, such as Mn, Cu, Zn, and Cd have been considered to explain the mice response in contaminated and non-contaminated sceneries.

Experimental

Standard solutions and reagents

All reagents used for sample preparation were of the highest available purity. Phenylmethanesulfonyl fluoride (PMSF) and tris(2-carboxyethyl)phosphine hydrochloride (TCEP) (BioUltra grade, >98 %) were obtained from Sigma Aldrich (Steinheim, Germany). Helium and hydrogen, used as collision and reaction gas, respectively, in a liquid chromatography–inductively coupled plasma-octopole reaction cell-mass spectrometry (LC‐ICP‐OCR‐MS) system, was of high-purity grade (>99.999 %).

Standards used for mass calibration of analytical size-exclusion chromatography (SEC) columns (mass range, 70–3 kDa) were: ferritin (440 kDa) (purity 95 %), bovine serum albumin (67 kDa) (purity 96 %), superoxide dismutase containing Cu and Zn (32 kDa) (purity > 70 %), myoglobin (14 kDa) (purity > 98 %), metallothionein I containing Cd, Cu, and Zn (7 kDa) (purity > 95 %), and arsenobetain (179 Da) (purity > 98 %). All these reagents were purchased from Sigma-Aldrich (Steinheim, Germany). The mobile phase solution used in SEC was 20 mM of ammonium acetate (Suprapure grade) purchased from Merck (Darmstadt, Germany), which was prepared daily with ultrapure water (18 MΩcm) from a Milli-Q system (Millipore, Watford, UK) and the pH adjusted to 7.4 with ammonia solution, this later prepared by dilution of 20 % (w/v) ammonia solution (Suprapur, Merck) with ultrapure water. The void volume was determined using ferritin (440 kDa).

In the case of SEC columns for low-molecular-mass species (10–0.1 kDa), mass calibrants were the following: bovine serum albumin (67 kDa) (purity 96 %), metallothionein I containing Cd, Cu, and Zn (7 kDa) (purity > 95 %), vitamin B12 (1.35 kDa) (purity > 96 %), reduced glutathione (307 Da) (purity 98–100 %), and arsenobetaine (179 Da). All these reagents were purchased from Sigma-Aldrich (Steinheim, Germany). Standard stock solutions at a concentration of 10 mg mL−1 were prepared by dissolving the respective compound in ultrapure water. Arsenic (III) and arsenic (V) standards were prepared from sodium arsenite and sodium arsenate in water, respectively, and arsenobetaine (AsB) were purchased from Sigma-Aldrich (Steinheim, Germany). Methylarsonic acid and dimethylarsinic acid were purchased from Supelco (Bellefonte, USA). Tetramethylarsonium ion and trimethylarsine oxide was supplied by Dr. Vélez (Instituto de Agroquímica y Tecnología de Alimentos, Valencia, Spain). Individual working solutions were prepared at 1 mg mL−1 of each of the compound for high-performance liquid chromatography (HPLC)-ICP-MS analyses. Working solutions were daily prepared by the appropriate dilution of the stock solutions with water. The stock solutions were stored at 4 °C in the dark.

The mobile phase solution used in SEC was 20 mM of ammonium acetate as described previously, and the void volume was determined in this case using bovine serum albumin (BSA) (67 kDa). For anion exchange chromatography, 30 mM of sodium dihydrogen phosphate was used as mobile phase, which was purchased from Sigma-Aldrich (Steinheim, Germany) and the pH adjusted at 7.4 by addition of 20 % NH3 (w/v) aqueous solution. However, for cation-exchange chromatography, a 20 mM of pyridine purchased from Sigma-Aldrich (Steinheim, Germany) in Milli-Q water, adjusting the pH at 2.5 by adding formic acid (>98 %, Fluka puriss, p.a.), was used for this purpose.

Amicon Ultra centrifugal filters (containing a 3,000-Da membrane of regenerated celulosa) were from Millipore (Billerica, MA). The trypsine (EC 3.4.21.4) TPCK was obtained from Sigma-Aldrich (Steinheim, Germany) and the urea, iodoacetamide, and dithiotreithol from Bio-Rad (Madrid, Spain).

Apparatus

A cryogenic homogenizer SPEX SamplePrep (Freezer/Mills 6770) was used to prepare the homogenates. Homogenized tissues were subsequently disrupted with a glass/Teflon homogenizer. The extraction was followed by ultracentrifugation with an ultracentrifuge Beckman model L9-90 K (rotor 70 Ti). Polycarbonate bottles of 10 ml with cap assembly (Beckman Coulter) were used for this purpose. A microwave oven (CEM Matthews, NC, USA, model MARS) was used for the mineralization of extracts.

Trace metals and metal-linked biomolecules were analyzed with an inductively coupled plasma mass spectrometer Agilent 7500ce (Agilent Technologies, Tokyo, Japan) equipped with an octopole collision/reaction cell. Chromatographic separations were performed using a Model 1100 HPLC pump with detector UV (Agilent, Wilmington, DE, USA) as the delivery system. ICP-MS measurement conditions (Table 1) for mode He and H2 were optimized by using a HNO3 2 % (v/v) aqueous solution of 59Co, 7Li, 89Y, and 205Tl (1 μg L−1). The flows of collision/reaction gases were fixed at 3.7 mL min−1 for He and at 3.5 mL min−1 for H2 in order to avoid or reduce the interferences.

Table 1 Operating conditions for ICP-MS, SEC, and IEC
Full size table

Mass spectrometry measurements were performed on a nano-electrospray ionization tandem mass spectrometer (API Qstar XL Hybrid system; Applied Biosystems, Foster City, CA, USA).

Sampling area and animals

Free-living mice (M. spretus) were collected during September of 2011 in five sampling areas from DNP, southwest Spain, in which contamination differs between different areas (Fig. 1). That: (1) “Lucio del Palacio” (LDP, green spot) located in the center of the park that was initially considered as a control area with low contamination, although the results show the presence of contaminants; (2) “La Rocina Stream” (ROC, red spot) with strawberry, citrus fruit, and grape fields in the surroundings is additionally affected by diffused pollution from petrochemical and chemical activities caused by the industrial belt of Huelva and acid waters and metals from north-west mining metallurgical activities in the Huelva province (Riotinto mine); (3) “El Partido” (PAR, orange spot) and (4) “El Ajolí” (AJO, brown spot) sites that correspond to up- and downstream from El Partido stream, respectively, which are under the influence of citrus fruit and grape fields; and (5) “El Matochal” (MAT, blue spot) is next to Guadiamar river that is affected by rice growing fields and suffered the input of metals transported by the Guadiamar river during the rupture of Aznalcollar mine tailing pond in 1998 [34]. The sampling sites are shown in Table S1, which includes the codes for the samples, UTM coordinates, type of sample (reference/problem), and number of mice of each sex sampled (see Electronic supplementary material, Table S1). M. spretus mice were caught with Sherman live-traps baited with hazelnut cream over bread, which were mounted during the evenings and checked the next morning. Adult animals were taken alive to a laboratory at Doñana Biological Reserve, and site/date of capture, sex, weight, and external measurements were recorded.

Fig. 1
figure 1

Sampling area in Doñana National Park and surrounding (SW Spain). a Area location; b localization of sampling points: Center of Doñana National Park, “Lucio del Palacio” (LDP, green spot); surrounding sites: low contamination: “El Ajolí” (AJO, brown spot); high contamination: “El Matochal” (MAT, blue spot) > “La Rocina”(ROC, red spot) > “El Partido” (PAR, orange spot)

Full size image

Mice were individually sacrificed by cervical dislocation and dissected using a ceramic scalpel and transferred rapidly to dry ice. Individual organs were excised, weighed in Eppendorf vials, cleaned with 0.9 % NaCl solution, frozen in liquid nitrogen, and then stored at −80 °C until they were used for extract preparation. Mice were handled according to the norms stipulated by the European Community. The investigation was performed after approval by the Ethical Committee of the University of Huelva (Spain).

Sampling of soil and sediments from the areas of study

Samples were collected in three different areas for each sampling site for mice. Sediments and soils samples were collected using a polypropylene spatula and subsequently transferred to acid-rinsed polypropylene bottles. The samples were air-dried, sieved to select the fraction <2 mm, and ground in an agate mortar for later sieving to select the fraction <0.63 microns.

Analytical procedures

Sample preparation

Pools of livers and kidneys from five different male mice (Table S2, see Electronic supplementary material) were treated following a procedure described elsewhere [27]. Briefly, the organs were disrupted by cryogenic homogenization in a 6770 freezer/mill apparatus (2 min at rate 15) from Spex SamplePrep (Metuchen, NJ, USA). After that, the metal–biomolecules were extracted with a solution (3 ml g−1) containing 20 mM ammonium acetate buffer solution at pH 7.4, 1 mM TCEP, and 1 mM PMSF using a glass/Teflon homogenizer in a cold chamber at a constant temperature of 4 °C. Then the extracts were centrifuged at 120,000×g for 1 h at 4 °C. Extracts were stored under nitrogen atmosphere to avoid oxidation by air and at −80 °C until analysis. Total protein contents (Bradford method) in the extracts of liver, kidney, and brain from M. spretus are given in Table 2.

Table 2 Protein content in the extracts of livers, brains, and kidneys from M. spretus
Full size table

Determination of total metal concentration in extracts from different organs

For this purpose, the extracts were exactly weighed (0.250 g) in 5-ml microwave vessels, and 800 μL of a mixture containing nitric acid and hydrogen peroxide (4:1 v/v) was added. After 10 min, the PTFE vessels were closed and introduced into the microwave oven. The mineralization was carried out at 400 W from room temperature ramped to 160 °C for 15 min and held for 40 min at this temperature. Then the solutions were made up to 2 ml with ultrapure water and the metals analyzed by ICP-MS. The element Rh was added as internal standard (1 μg L−1). All the analyses were performed using three replicates.

Determination of total metals concentration in soil and sediment

The fraction <63 μm was digested in a microwave-accelerated reaction system. For this purpose, the samples were weighed exactly (0.250 g) in PTFE microwave vessels, and 5 mL of nitric acid, 2 mL of hydrofluoric acid, and 1 mL of hydrochloric acid were added; the vessels were closed and introduced into the microwave oven. The mineralization was carried out at 800 W from room temperature ramped to 180 °C for 10 min and held for 50 min at this temperature. After digestion 6 mL of saturated boric acid to neutralize the excess of hydrofluoric acid was added. Then the solutions were made up to 25 ml, filtered through Iso-Disc poly-(vinylidene difluoride) filters (25 mm diameter, 0.45 μm pore size) and stored at 4 °C until analysis by ICP-MS. In all analyses, three replicates were performed, and Rh was added as internal standard (1 μg L−1). Recovery experiments were performed by spiking the different samples with 1, 5, 50, or 100 μg L−1 of metals depending on the relative concentration of each metal.

Analysis of tissue extracts using analytical and preparative SEC-ICP-MS

The extracts were filtered through Iso-Disc poly(vinylidene difluoride) filters (25 mm diameter, 0.2 μm pore size) to avoid column overloading or clogging. The SEC-ICP-MS online coupling was performed by connecting the outlet of the chromatographic column to the Micromist nebulizer inlet (GlasExpansion, Switzerland) of the ICP-MS by means of a 30 cm PEEK tubing (0.17 i.d. mm). The quality control of the SEC-ICP-MS system to overcome problems related to contamination, loss, and stability of species has been described elsewhere [27].

Separation of the analytes was performed with a glass analytical size exclusion column SuperdexTM-75 (10 × 300 mm; 13 μm) (GE Healthcare, Uppsala, Sweden) with an exclusion limit of 100 kDa (effective separation range from 3 to 70 kDa). Ammonium acetate 20 mM (pH 7.4) was used as mobile phase, at flow rate 0.7 ml min−1 and injection volume of 20 μL.

Other size exclusion chromatographic separations were performed using an analytical SEC column (10 × 300 mm, 13 μm) SuperdexTM-Peptide (GE Healthcare, Uppsala, Sweden) with an exclusion limit of 10 kDa (with an effective separation range from 10 to 0.1 kDa), using mobile phase of ammonium acetate 20 mM (pH 7.4), at flow rate of 0.7 ml min−1 and injection volume of 50 μL. Preparative size exclusion chromatography was carried out by using a Hiload 26/60 Superdex 30 Prep for separation range <10 kDa (Amersham Biosciences, Uppsala, Sweden). The insolation of the fraction located at about 30 kDa from liver extract was carried out by using a Hiload 26/60 Superdex75 Prep for separation range 3–70 kDa (Amersham Biosciences, Uppsala, Sweden). An AKTA-Prime system (pump and UV detector at 254 nm) (Amersham) was used as eluent delivery system, equipped with a 500-μl sample loop. The separation was performed with the mobile phase described previously, at flow rate 2 ml min−1 and injection volume of 500 μL.

After SEC preparative separation of Cu- and Zn-containing fractions, they were collected and loaded onto a HiTrap desalting column to remove salts. The mobile phase used for desalting was ammonia solution (pH 8) at flow rate 4 ml min−1 [35]. The fraction excluded from the desalting column was collected and lyophilized.

The retention times corresponding to the peaks of standards used for SuperdexTM-75 column calibration were the following—ferritin, 11.5 min; BSA, 13.7 min; superoxide dismutase containing Cu and Zn (Cu,Zn-SOD), 16.3 min; myoglobin (an iron containing protein), 18.9 min; metallothionein I containing Cd, Cu, and Zn (Cd,Cu,Zn-MT1), 21.0 min; and AsB, 26.1 min. The retention times corresponding to the peaks of standards used for SuperdexTM-Peptide column calibration were the following—BSA, 10.5 min; metallothionein I containing Cd, Cu, and Zn (Cd,Cu,Zn-MT1), 14.5 min; vitamin B12 (1.35 kDa) (a cobalt-containing protein), 18.5 min; reduced glutathione (307 Da), 24.0 min; and arsenobetaine (179 Da), 25.9 min. Finally, the retention times corresponding to standards used for Hiload 26/60 Superdex 30 Prep column were—BSA, 45.0 min; metallothionein I containing Cd, Cu, and Zn (Cd,Cu,Zn-MT1), 64.8 min; vitamin B12 (1.35 kDa) (which contains cobalt), 85.2 min; reduced glutathione (307 Da), 101.7 min; and arsenobetaine (179 Da), 109.0 min. On the other hand, the retention times corresponding to standards used for Hiload 26/60 Superdex 75 Prep column were—ferritin (440 kDa), 45 min; BSA, 53 min; superoxide dismutase (Cu,Zn-SOD), 70 min; myoglobin (an iron containing protein), 85 min; metallothionein I containing Cd, Cu, and Zn (Cd,Cu,Zn-MT1), 90 min; and vitamin B12 (1.35 kDa), 110 min.

Analysis of tissue extracts using analytical and preparative IEC-ICP-MS

Anion-exchange HPLC separations of arsenic species were carried out using a Hamilton PRP X-100 (250 × 4.6 mm × 5 μm) column, mobile phase sodium dihydrogen phosphate 30 mM (pH 6), at flow rate 1.0 ml min−1 and injection volume of 20 μl. Cation-exchange chromatography of these species was performed with a Supelcosil LC-SCX (250 × 4.6 mm × 5 μm) column, using pyridine 20 mM (pH 2.5) as mobile phase, at flow rate of 1.2 ml min−1 and injection volume of 20 μl.

Preparative anion-exchange HPLC separations for the purification of fraction containing Cu-Zn at about 30 kDa was carried out with preparative anion-exchange chromatography (AEC) column (Protein-Pak DEAE 5PW 21.5 × 150mm Semi-prep column) using ammonium acetate as eluent, with a gradient from 2 to 200 mM of this buffer at flow rate of 5 ml min−1 and injection volume of 500 μl. The AEC-ICP-MS online coupling was performed by connecting the outlet of the chromatographic column to the Miramist nebulizer inlet (Burgener, Ontario, Canada) of the ICP-MS by means of a PEEK tube, using post-column flow splitting (5:1) before the ICP-octopole reaction system (ORS)-MS nebulizer.

Identification of Cu,Zn-traced peak from liver extract

It was performed as described elsewhere [36]. Briefly, liver cytosolic extract was submitted to SEC separation (Hiload 26/60 Superdex 75 Prep), and Cu/Zn containing fraction was collected and loaded onto a HiTrap desalting column to remove the salts. The mobile phase used for desalting was ultrapure water (pH 8) at 4 mL min−1 [35]. Finally, the fraction excluded from the desalting column was collected and lyophilized. The lyophilizate was dissolved in ultrapure water and purified by preparative AEC (Protein-Pak DEAE 5PW 21.5 × 150 mm Semi-prep column) using a gradient of ammonium acetate buffer from 2 to 200 mM at flow rate of 5 ml min−1 and injection volume of 500 μl. After analysis, 2 mM of ammonium acetate buffer was switched on for 15 min in order to re-equilibrate the column before the next injection. A total volume of 2 mL of the fraction isolated by AEC (retention time between 22 and 24 min) was lyophilized to approximately 4 mL and loaded onto a HiTrap desalting column to remove salts. The mobile phase used for desalting was ammonia solution (pH 8) at flow rate 4 ml min−1 [35]. The fraction excluded from the desalting column was collected and lyophilized.

The lyophilized fraction was re-dissolved with 100 μL of 6 M urea and 50 mM ammonium bicarbonate (pH 8.3). An aliquot of 50 μL of this solution was separated, and 5 μL of 180 mM DTT was added to reduce disulphide bonds in the proteins. After 30 min at 37 °C, 5 μL of 400 mM iodoacetamide was added to the reaction mixture and left in dark at room temperature for 30 min. Finally, the sample was dissolved in 290 μL water to reduce the urea concentration to 2 M, to retain the activity of trypsin. Then, 50 μL of trypsin (0.1 μg μL−1) was added and the mixture incubated at 37 °C overnight. The reaction was then stopped by addition of 10 μL acetic acid (glacial) [37]. After tryptic digestion, the peptides were desalted, preconcentrated, and purified using ZipTips C18 (Millipore, Massachusetts, USA). The time of flight (TOF) mass analyzer was calibrated immediately prior to sample analysis using renin as standard. Electrospray ionization (ESI)-TOF-MS data acquisition was performed in positive ion mode, and the MS spectra of peptides were acquired in the range 400–1600 m/z. The values for ion spray voltage, electron multiplier voltage, curtain gas, and declustering potential were set to—1,100 V; 2,200 V; 20 psi, and 90 V. Data analyses were performed using the Analyst QS software (Applied Biosystems). After recording the MS spectra, doubly charged peptide ions were selected, and tandem mass (MS/MS) spectra obtained with collision energy of 45 V. Peptide sequences were then searched by database (NCBI) using MASCOT searching engine.

Data analysis

Statistical data analyses (principal component analysis (PCA)) were performed by means of statistical software packages Statistica 6.1 (StatSoft Inc., Tulsa, OK, 1984–2004).

Results and discussion

The present study considers and compares the biological response of mice Mus spretus from contaminated and non-contaminated sites of Doñana National Park and surrounding areas in the southwest Spain. Organs with high metabolic activity such as liver and kidneys were used for this purpose. The changes in the levels of metal–biomolecules caused by pollutants were traced by SEC-ICP-MS. The most interesting results were observed in relation to Mn, the group Cu, Zn, and Cd, and the toxic element As.

Total concentration of metals in the organs

The presence of metals in the extracts of different organs (liver, brain, and kidney) of M. spretus from the five sampling sites previously described (Fig. 1) was evaluated, and the results are shown in Table 3. Recovery experiments were performed by spiking the extracts with 1, 5, 10, or 50 μg L−1 of metals depending on the relative concentration of each metal in the extracts; the results are also shown in Table 3 and confirm quantitative recoveries in all the cases. Instrumental detection limits are also given in this table.

Table 3 Concentration of elements (micrograms per gram) in the extracts of the different organs of mice M. spretus
Full size table

The higher presence of copper, zinc, and manganese in liver extract is remarkable. Toxic elements, such as cadmium and especially arsenic, show very low concentrations in both liver and kidney; in brain extracts, the presence of these elements even decreases. These toxic elements have been mainly found in liver and kidney of free-living mice captured in MAT, although their concentrations in analogous organs from ROC and LDP are also remarkable. The relative high concentration of As in these sites but especially of Cd in mice from LDP is very surprising, which raises serious suspicion about the absence of contamination in the center of the park. Further analysis of metals in soils and sediments from the zone (Table 4) revealed that concentrations of Cd and As in LDP are higher than that in AJO, PAR, and ROC and the presence of Mn, Cu, and Zn unexpectedly high. Therefore, in spite of the location of LDP at the center of the Doñana Park, metal diffuse contamination reaches this area, which can affect metal–biomolecule expression in mice.

Table 4 Concentration of elements (micrograms per gram) in soils and sediments from Doñana National Park and surroundings
Full size table

For a better understanding of significance of metal concentrations determined in the different areas under study, principal component analysis was applied to data from Tables 3 and 4, and results are shown in Fig. 2a (soils and sediments) and b (mice organs). In soils, (Fig. 2a) two first principal components that account for 87 % of total system variability (PC1 for 53 % and PC2 34 %) can be seen. The PCA distinguish three groups that correspond to: (1) ROC, AJO, and PAR; (2) MAT; and (3) LDP. Considering the loadings of features, Mn and As are the dominant variables in the first principal component (both negative correlated with PC1), while Cd and Cu dominate in the second principal component (negative and positive, respectively, that correlate with PC2).

Fig. 2
figure 2

PCA scatterplot. a Samples of soils and sediments; b samples or organs (livers, kidneys, and brains)

Full size image

Figure 2b shows a natural separation of the three types of organs studied (kidneys, brains, and livers). Inside each group, it is possible to distinguish between the different areas, especially when livers and kidneys are considered. However, for brains, the discrimination is not so evident between samples from different locations since the concentrations of metals are very low in all the cases. In this case, the first two principal components account for 87 % of total system variability; the dominant variable of first component that accounts for 73 % of total variability (Fig. 2b) is related to Mn, Zn, and Cu which are negatively correlated. Samples of liver from mice captured in MAT, LDP, and AJO exhibit higher content of these elements, which is in good agreement with the previous discussion. On the other hand, the dominant variable of the second principal component is As followed by Cd, both negatively correlated. In this case, the higher content of dominant elements is found in the kidneys of mice from ROC and LDP. Therefore, Mn is a key element to classify the studied areas considering organs, soils, and sediments.

Differences in Mn-species in liver and kidney from M. spretus in the sites under study

The highest concentration of manganese was found in liver extracts (Table 3), with a mean value of 0.620 μg g−1 that contrasts with the higher concentrations of this element in pig liver, up to 3 μg g−1 [38]. It is necessary to consider that liver is the major Mn-reservoir organ, and the element is transported to other organs by Mn transporters such as Mn-transferrin and/or albumin [15, 39].

The relative abundance of Mn-species in liver and kidney from mice captured at the sites under study was evaluated in the cytosolic fraction of these organs by SEC-ICP-MS using a SuperdexTM-75 column with effective range of separation between 3 and 70 kDa. Figure 3 shows the results obtained, which reveal that intensity of signals traced by manganese is higher in liver than in kidney. In the former (Fig. 3a), three peaks can be observed: (1) at void volume that can be related to Mn-containing enzymes present in mouse and rat liver, such as catalase (Mr = 345 kDa) and arginase (Mr = 75 kDa) [40, 41], (2) at 13 min that matches well with the standard of BSA (66.7 kDa), and (3) a peak with retention time matching with the standard of Cu/Zn-SOD (32 kDa). Several unsuccessful attempts were made to resolve the peaks using previously described procedures based on a second dimension reverse-phase chromatographic separation [42]. However, the findings from Michalke et al. [39], using SEC followed by capillary zone electrophoresis-ICP-MS in the study of Mn species in porcine liver, could help to explain the present results. The peak located at the void volume could integrate a number of Mn-containing enzymes and proteins, such as concanavalin A, arginase, isocitrate dehydrogenase, pyruvate carboxylase, oxalate oxidase, and Mn-transferrin [39]. The peak at about 67 kDa could include Mn-albumin and galactosyltransferase prolidase, while the most intense peak in the mass interval 32 to 55 kDa could be related with Mn-superoxide dismutase. However, Mn-SOD is present only in mitochondria [43], and these organelles are not supposed to be disrupted by the extraction procedure used, designed for soluble cytosolic species. An alternative hypothesis to explain the presence of this peak is the replacement of active-sites of Zn-carbonic anhydrase by Mn, yielding manganese-substituted carbonic anhydrase (Mn-CA), a peroxidase activity enzyme with molecular mass in the range 29–54 kDa depending on isoform considered [44]. This assumption could explain that intensity of this peak increases with the concentration of Mn in liver extract (Table 3) that follows the order: LDP > ROC > MAT > AJO > PAR. This trend can also be observed in kidney (Fig. 3b), in which the concentration of Mn is lower (Table 3) and correlatively the intensity of peak decrease. In kidney, only this Mn-traced peak can be observed. However, further studies should be performed to confirm these results.

Fig. 3
figure 3

Mn-binding biomolecules in liver (a) and kidney (b) extracts of M. spretus from differential contaminated areas characterized by SEC-ICP-MS. Chromatographic conditions—column, SuperdexTM-75 (10 × 300 × 13 μm); mobile phase, ammonium acetate 20 mM(pH 7.4); flow rate, 0.7 ml min−1; injection volume, 20 μL

Full size image

The lower intensity of Mn-peaks in liver from MAT (Fig. 3a), which is a contaminated area in comparison to other sites apparently less exposed to pollutants such as LDP, is remarkable. However, the lower total concentration of manganese in liver extracts from MAT with respect to LDP (Table 3) can explain these differences.

Size characterization of Cu–, Zn–, and Cd–biomolecules in liver and kidney extracts from contaminated and non-contaminated areas

The distribution, accumulation, and transference of zinc, copper, and cadmium in living organisms have been considered in detail in the literature. Several facts such as the modulation of zinc concentration by homeostatic mechanisms [45], and the importance of transport mechanisms on copper distribution [46] are relevant issues to explain the relative presence of metal-binding molecules in the different organs of exposed organisms. In relation to this, the induction of Cd and Zn-metallothioneins in mice (M. musculus) exposed to industrial dust rich in metals has been reported [47], and these experimental data confirm the antagonistic interactions among Cd, Zn, Cu, and Pb, as well as the differential rate of excretion of these elements from kidney/liver under increasing exposure [47].

To check the presence and potential interactions of metal–biomolecules in organs of M. spretus from the areas under study, the coupling SEC-ICP-MS was used to obtain the Cu, Zn, and Cd-traced peaks in the cytosolic fractions of liver and kidney of M. spretus (Fig. 4). These chromatograms show great similarities in metalloproteins expression, being remarkable the high intensity of the peaks traced by Cu, Zn, and Cd at 7 kDa in liver extract, which match well with metallothionein I standard (Fig. 4a, c, and e). The intensity of Cu-peak from the contaminated area MAT (Fig. 4a) is clearly higher than that from other sites also affected by contamination, such as ROC and AJO (that represent a 40 % reduction of intensity respect to MAT) and LDP and PAR (60–70 % reduction of intensity respect to MAT). A similar peak located at 7 kDa was observed for Zn (Fig. 4c), although, in this case, LDP exhibits the highest intensity followed by MAT, which in fact can be related with the correlative concentration of this element in the cytosolic extract (Table 3). The equivalent peak traced by Cd (Fig. 4e) shows lower sensitivity, with remarkable intensity of peaks corresponding to MAT and LDP. Another peak traced by Cu at 32 kDa shows higher intensity for samples from LDP, MAT, ROC, and PAR, but the response is lower in the area with low pollution (AJO) (Fig. 4a). In Fig. 4c, two signals traced by Zn that matches with standards at 67 kDa and 32 kDa, respectively, can be seen. In this latter case, the highest response corresponds again to LDP (Fig. 4c) with lower response from MAT and ROC despite being contaminated areas. To confirm that this Cu,Zn-peak of molecular mass of 32 kDa is superoxide dismutase, it was collected from liver extract in which it was more concentrated (LDP) using a preparative SEC coupled to ICP-MS (Fig. 5a). The fraction was desalted, lyophilized, and analyzed by AEC (Fig. 5b). In this later chromatogram, it is possible to observe a predominant peak traced by Cu and Zn which can be collected with a preparative AEC column, lyophilized, desalted, and then submitted to tryptic digestion for identification by nano-ESI-triple quadrupole-TOF. The mass spectrum of peptides resulted after tryptic digestion (Fig. 5c) was used for protein identification, using doubly charged peptide ions of m/z 584.31, 684.39, and 756.84 for MS/MS analysis and protein identification in MASCOT database. This confirms the presence of Cu,Zn-superoxide dismutase associated to the peak of 32 kDa in Fig. 4c.

Fig. 4
figure 4

Up/downregulation of metal–biomolecule complexes in liver (a, c, e) and kidney (b, d, f) of M. spretus from different environmental areas assessed by molecular mass distribution with SEC-ICP-MS. Chromatographic conditions—column, SuperdexTM-75 (10 × 300 × 13 μm); mobile phase, ammonium acetate 20 mM (pH 7.4); flow rate, 0.7 ml min−1; injection volume, 20 μL

Full size image
Fig. 5
figure 5

a Liver extract of M. spretus from LDP area characterized by SEC-ICP-MS. Chromatographic conditions—column, Hiload 26/60 SuperdexTM-75 (10 × 300 × 13 μm); mobile phase, ammonium acetate 20 mM (pH 7.4); flow rate, 2 ml min−1; injection volume, 500 μL. b AEC-ICP-ORS-MS chromatogram of the Cu- and Zn-containing fractions of 32 kDa collected by SEC. Chromatographic conditions—column, Protein-Pak DEAE 5PW Semi-prep (21.5 × 150 mm); mobile phase, a gradient from 2 to 200 mM of ammonium acetate buffer pH 7.4 at flow rate of 5 ml min−1; injection volume, 500 μl. c Mass spectrum obtained by nano-ESI-MS of the fraction collected by AEC after tryptic digestion

Full size image

Finally, it is interesting to observe the presence of Cu-, Zn-, and Cd-peaks with differential intensity in kidney extracts at about 7 kDa (Fig. 4b, d, and f). In this case, the response evaluated as intensity of Cu and Zn peaks is higher for MAT, PAR, and AJO, which correlate with the concentration of these elements in the cytosolic extract (Table 3). However, the intensity of Cd-peak (Fig. 4f) is higher in LDP, again related to the highest presence of this element in the corresponding extract.

Presence of small metabolites of arsenic in kidneys

It is well known that organoarsenic species such as arsenobetaine, arsenocholine, arsenosugars, and arsenolipids are present in aquatic organisms and marine animals [48, 49]; however, mammals metabolize toxic inorganic arsenic in the liver [50], after absorption from the gastrointestinal tract, producing methylated species that are excreted in the urine as pentavalent methylated arsenic forms [51–54]. In addition, it has been recently demonstrated that binding of arsenic to certain proteins may be a detoxification process [55], considering that trivalent arsenic is chemically more reactive than pentavalent species and binds many proteins, such as tubulin and actin [56], hemoglobin [57, 58], metallothionein [59], and others.

The presence of low-molecular-mass As species in kidney cytosolic extracts analyzed by SEC-ICP-MS (column SuperdexTM-75) can be seen in Fig. 6; the higher intensity of signal was obtained from ROC, in which As concentration in kidney extract is the highest (Table 3). Generally, the intensity of signals traced by arsenic is higher in kidney than in liver; for this reason, further studies were performed on kidney extract from ROC. In order to obtain better-resolved chromatograms for these low-molecular-mass As-molecules, the SEC separation was performed on a SuperdexTM-peptide column with effective range from 7 to 0.1 kDa (Fig. 6a). Two peaks related with low-molecular-mass metabolites can be observed at 26 and 27.5 min, respectively. To identify these As species, the extract was fractionated using a preparative column Hiload 26/60 Superdex 30 Prep and the peak associated with the molecular mass range 100–300 Da collected (Fig. 7b).

Fig. 6
figure 6

Arsenic metabolites in kidney from M. spretus from ROC traced by SEC-ICP-MS. Chromatographic conditions—column, SuperdexTM-75 (10 × 300 × 13 μm); mobile phase, ammonium acetate, 20 mM (pH 7.4); flow rate, 0.7 ml min−1; injection volume, 20 μL

Full size image
Fig. 7
figure 7

a Arsenic metabolites in kidney from M. spretus captured in ROC traced by SEC-ICP-ORS-MS. Chromatographic conditions—column, SuperdexTM-Peptide (10 × 300 × 13 μm); mobile phase, ammonium acetate 20 mM (pH 7.4); flow rate, 0.7 ml min−1; injection volume, 50 μL. b Arsenic metabolites in kidney from M. spretus captured in ROC traced by SEC-ICP-ORS-MS. Chromatographic conditions—column, Hiload 26/60 Superdex 30 Prep; mobile phase, ammonium acetate 20 mmol L−1 (pH 7.4); flow rate, 2 ml min−1; injection volume, 500 μL. c As-metabolites in kidney extracts from M. spretus mice. Chromatographic conditions—column, Hamilton PRP X-100 (250 × 4.6 mm × 5 μm), mobile phase, sodium dihydrogen phosphate, 30 mM (pH 6); flow rate, 1.0 ml min−1; injection volume, 20 μL. d As-metabolites in kidney extracts from M. spretus mice. Chromatographic conditions—column, Supelcosil SCX (250 × 4.6 mm × 5 μm), mobile phase, pyridine 20 mM (pH 2.5); flow rate, 1.2 ml min−1; injection volume, 20 μl

Full size image

However, the confirmation of specific arsenic-metabolites is not possible with only a SEC separation [60], and a complementary IEC method was applied for further purification of the arsenic SEC fraction (Fig. 7c, d). The chromatogram obtained by anion exchange chromatography show that methylated arsenicals (MMA and DMA) are the predominant species in the kidney (Fig. 7d). When cation exchange chromatography is used (Fig. 7c), a peak at 4 min can be observed, which could be associated to the presence of arsenobetaine because its retention time matches well with the corresponding standard of this arsenic specie frequently found in aquatics organisms and marine mammals [48, 49], although this compound is not present in terrestrial organisms. Nevertheless, other arsenic species can also be related to this peak due to their similar chromatographic behavior, such as dimethylarsinoylethanol, which is present in human urine and in urine from experimental animals [61], and thioarsenic species such as MMMTA(V), DMMTA(V), and DMDTA(V) [62]. In addition, an unknown peak at about 7 min can be related with the presence of arsenocholine according to the retention time reported by the same column and chromatographic conditions by Chatterjee [63].

Conclusions

The use of SEC-ICP-MS coupling combining analytical and preparative columns is a good choice to assess changes in the regulation of metal-binding biomolecules in environmental bioindicators (e.g., M. spretus) caused by contamination episodes. Intensities of SEC-ICP-MS signals from Mn-containing molecules are higher in liver than in kidneys, showing a complex Mn-profile that cannot be well resolved, and for this reason, multidimensional chromatographic approaches are used to overcome this problem. The Mn-peak located at 32 kDa is the only one present in both liver and kidney and matches well with SOD standard; however, this protein can only be present in mitochondria that contrast with the use of cytosolic extracts in this experiment. The relationship between the concentration of Mn in the extracts from the different areas under study and the intensity of this peak has been proposed as hypothesis to explain these facts, considering that replacement of active-sites of Zn-carbonic anhydrase by Mn could be favored by the presence of this element. The higher concentration of Mn in liver respect to kidneys and the correlative increases of the intensity of this peak support this hypothesis. However, further studies are mandatory to confirm it.

Higher-intensity Cu-, Zn-, and Cd-peaks with molecular mass about 7 kDa were observed in samples from contaminated areas such as MAT, although other peaks related with molecules of 32 and 67 kDa traced by Cu and Zn exhibit high intensity in samples from areas with low contamination such as LDP, in which the 32 kDa-peak from liver was identified as Cu,Zn-SOD by organic mass spectrometry.

Two peaks traced by As with small molecular mass were observed from kidney extracts, with higher intensity in samples from ROC. The combination of preparative SEC and anion exchange chromatography allowed the identification of methylated arsenicals (MMA and DMA).

These results reveal the possibilities of the approach proposed in metal-related metabolism and homeostasis. One important use of this methodology is the study of metal trafficking, especially metal interactions with active sites in the biomolecules. Future works are necessary to identify by mass spectrometry the metal–biomolecules traced by SEC-ICP-MS as well as the identification and quantification of the arsenic-containing peaks by HPLC-MS and the exhaustive study of the trafficking of metals between the organs of mice exposed to environmental pollution.