Mercury distribution in plants and soils from the former mining area of Abbadia San Salvatore (Tuscany, central Italy)

The distribution of heavy metals in plants growing in soils from active and abandoned mining areas is of scientific significance as it allows one to recognize their ability to survive in a hostile environment and to provide useful indications for phytoremediation operations. In this work, soils developed in the former Hg-mining area of Abbadia San Salvatore (Tuscany, Central Italy) were analyzed for total, leached Hg, % of organic- and inorganic-related Hg. The dehydrogenase enzyme activity (DHA) was also measured with the aim to evaluate the status of the soil, being characterized by high Hg content. Eventually, the concentration of Hg in the different parts of the plants growing on these soils was analyzed. The soils showed Hg content up to 1068 mg kg− 1 and in most of them is dominated by inorganic Hg (up to 92%). The DHA concentrations were < 151 μg TPF g− 1day− 1, suggesting that the presence of Hg is not significantly affecting the enzymatic soil activity. This is also supported by the bioaccumulation factor (BF) that is < 1 in most of the studied plants. Generally speaking, the plant leaves appear to be one of the main pathways of Hg uptake, as also observed in other mining areas, e.g. Almaden (Spain), suggesting that particulate-Hg and Hg0 are the main forms entering the plant system, the latter derived by the GEM emitted by both the edifices hosting the roasting furnaces and the soils themselves.


Introduction
Since 2011 the world-class Hg mining district, centered in the Municipality of Abbadia San Salvatore (Tuscany, Italy) and located in the eastern part of the Mt. Amiata silicic volcanic complex (Conticelli et al., 2004Ferrari et al., 1996;Laurenzi et al., 2015), has been undergoing remediation operations. Since the 19th century, Abbadia San Salvatore (ASS) was the site of one of the most important areas for the exploitation of cinnabar and production of liquid Hg. The ore deposit was indeed excavated, dried and roasted and, through a condensation system, liquid Hg was produced. Through the years, different furnaces were used: from those fed by wood to Spirek-Cermak, from the Gould Paci c to Nesa. It has been estimated that about 70% of the total production of Hg from the whole Mt. Amiata mines was from ASS (Cipriani & Tanelli, 1983). The mining structures of ASS produced more than 100,000 tons of liquid Hg and about 10,000 tons were dispersed into the environment (Bacci et al., 1998;Vaselli et al., 2015).
Past and recent geochemical investigations (e.g. Vaselli et al., 2019 and references therein) carried out in soils, waters and air inside the mining area and surroundings have highlighted that most environmental matrices have Hg concentrations much higher than those regulated by the European legislative levels.
According to Iverfeldt (1991) and Munthe et al. (1995), the atmospheric Hg assimilation from vegetation and its transfer to soil and water via throughfall and litterfall are the main sources of Hg in the terrestrial ecosystem. According to Barquero et  speaking, at least one order of magnitude lower than those related to inorganic Hg, although the accumulation of Hg in food webs is still not clear (Bailey et al. 2002. The primary warning system of alterations in terms of soil health and quality are soil enzymes, which can be considered potential bio-indicators to assess the soil health status (Datta et al., 2021 and reference therein). Among all soil enzymes, DHA (Dehydrogenase), alkaline phosphate (ALP) and urease (UR) are sensitive to both potentially toxic elements (PTEs) and minimal environmental changes (Datta et al., 2021;Elmayel et al., 2020;Gallego et al., 2021). In particular, DHA resides within all living microbial cells and, for this reason, it is considered a critical indicator of the enzymatic (and therefore, microbiological) activities of soils (Yuan and Yue, 2012). According to Mahbub et al. (2016) and Tazisong et al. (2012), scarce information is presently available about the biological functionality of Hg in plants. Nevertheless, this element has the ability to decrease enzyme activity by binding to both the protein's -SH residues and the active sites of enzyme and protein substrate complexes, or substituting metal cofactors such as Ca or In this paper, the distribution of Hg between soils and the most common plants growing in the former mining area of ASS and selected parts of them (e.g. roots, trunk, leaves) is presented and discussed in order to understand at which level they are either stressed by or recalcitrant to high Hg soil content.
Moreover, for the rst time in this area, the possible interaction between the enzyme Dehydrogenase and Hg in the pedological cover is evaluated.

The Study Area: ASS mine
The ASS mining area is located in the NW part of the local urban center. The main deposits consisted of cinnabar (HgS), with smaller amounts of pyrite, marcasite, orpiment and realgar (Rimondi et al., 2015). All the exploitation was underground and the galleries reached down to -400 m below the ground level (e.g. Lazzaroni et al., 2022). Mercury production began in 1899, following the abandonment of the area for about 2,000 years since Etruscans and Romans used cinnabar as pigment (Botticelli, 2019;Fantoni et al., 2022). The old mining area includes a large wood deposit for the old furnaces, some ancient dryers, and a few tanks that were used to cool gaseous Hg as it passed through the condensers. New dryers, belt conveyor systems, and horizontal (Gould) and vertical (Nesa) furnaces were installed in the following years, along with more e cient condensation systems. In 1976, the production activity at ASS dramatically slowed down since the exploitation of Hg was not economically sustainable, since its use had become conspicuously harmful and toxic. In 1982, the whole mining plant was de nitively shut down. In the 1990s, ENI (Ente Nazionale Idrocarburi) -AGIP (Azienda Generale Italiana Petroli) Unitproposed a reclamation project to permanently close the former mining and industrial activity. In 2008, an agreement between the Municipality of ASS and the former owner of the mining concession (ENI-AGIP) was signed. The mining concession and the reclamation project was thus transferred to the public administration. The ENI project was fully revised by the local municipality and the reclamation operations were then addressed to the environmental restoration of the mining areas and buildings for museum and public utility purposes (Vaselli et al., 2019). Consequently, the whole mining concession was divided into seven different units (Fig. 1a), including the reclamation of about 65 ha (black contour in Fig. 1a

Materials And Methods
Twenty-four plants (eight different species: Castanea sativa, Sambucus nigra, Verbascum thapsus, Popolus spp., Salix spp., Acer pseudoplatanus, Robinia pseudoacacia, Cytisus scoparius) and related soils were collected. Soils were sampled at a depth between 15 and 20 cm in order to analyze the total Hg distribution between roots and soils (Johnson et al., 2005). Two years before the sampling date, a complete vegetation clearcutting was carried out. Thus, the age of the sampled plants was known. It was decided to divide the plant samples into bark, internal and external roots bark and internal trunks, medulla part (when the trunk or root presented it), and foliage. Table S1 reported the geographic coordinates in WGS84-UTM 32N, the soil IDs, the sampling location, the Latin name of the sampled plants and the parts into which each plant was divided and analyzed. For each plant sample, the soil particles were brushed off manually. Subsequently, the plant samples were washed in an ultrasonic bath until the MilliQ water was clean. Then, they were heated in an oven for at least three days at 35°C. Eventually, each plant portion was ground into small pieces using a coffee grinder.
The soil samples were stored in an oven at a temperature of 35°C (to prevent any Hg 0 loss) until they were dried, and then sieved at 2 mm. A representative aliquot of the < 2 mm samples (about 100 g) was pulverized in a planetary mill (Pulverisette 5) equipped with agate mortars and balls. The pH values for each soil were determined following the UNI_EN 15933:2012 norm using a multi-probe Hanna HI98194.
Mercury concentrations in soils and plants were determined by a Lumex RA 915+ (Atomic Absorption Spectrometry with Zeeman effect) instrument (Sholupov et al., 2004), equipped with a Pyro-915 + device. Additionally, the percentage of Hg related to the organic and inorganic fractions was also estimated (Rumayor et al., 2016). In fact, it must be taken into account that biological samples burn at 275-290°C. At this temperature interval, all organic Hg compounds decompose and release Hg. and eventually, cooled in a freezer for 10 min to stop the reaction. Subsequently, 5 mL of methanol were added to extract the formed TTC. The colored extract was measured at 485 nm by molecular spectrophotometry. All these analyses were carried out at the IGeA Laboratories (Instituto de Geología Aplicada, University of Castilla La Mancha) in Almadén (Ciudad Real, Spain).
The leachable Hg was determined by ICP-MS using the USEPA 1312 method (USEPA, 1994) at the CSA Laboratories in Rimini (Italy), in order to simulate the fraction of soluble Hg in acid mine drainage conditions. This extraction entails mixing 5 g of soil sample with 100 mL of the EPA solution. A 60/40 combination of H 2 SO 4 and HNO 3 was added to 2 L of distilled water to create the EPA solution (pH of 4.5 ± 0.05). After mixing the soil with the EPA solution, the samples were heated in a stirrer thermostatic bath for 18 hours at 30 revolutions per minute. Water was added to maintain a constant temperature of 25°C.
After 18 hours, the samples were ltered using glass ber lters with a 0.45 µm pore size.

Hg Bioaccumulation Factor
According to Campos et al. (2018), the Bioaccumulation Factor re ects the bioavailability of Hg in plants. Bioaccumulation Factor (BF) is a parameter used to measure the transfer capacity of PTEs from soil to plant (Wang et al., 2016). In this study, BF was calculated as the ratio between the Hg concentration in the different parts of the plants and the soil leachable Hg content (Eq. 1):

Hg in plants
The minimum, maximum of BF values and minimum, maximum and median of Hg concentrations in the eight types of plants, without distinguishing the different plant portions, are summarized in Table 2. All data are listed in Table S3 (Supplementary Material).

Discussion
As far as the total Hg distribution in soils is concerned, the GC zone results to be the area with the highest concentration of Hg, followed by TMB > GO > FN (Fig. 1b). No signi cant correlation between total and leached Hg is observed, indicating that Hg is heterogeneously distributed in the investigated soils, being likely related to different sources. In contrast to Campos et al. (2018), who reported a correlation between total Hg and soil leached Hg of 0.79, and a correlation between soil leached Hg and humic acid Hg of 0.65, in this work soil leached Hg did not present any correlation between the Hg species analyzed. The top-soils are indeed affected by the presence of anthropogenic materials. In the past, in some portions of the former mining area, post-roasting and anthropic man-made (e.g. bricks, tiles, fragments of concrete) materials were used to ll a small paleo-valley positioned in front of the edi ce hosting the Gould and Nesa furnaces (Fig. 1b) (Vaselli et al., 2015).
The thermal speciation data evidenced that most Hg is inorganic although eight top-soils (i.e., ASS3, ASS4, ASS10, ASS14, ASS17b, ASS18, ASS19, and ASS20a) have a percentage of organic-related Hg that prevails over that related to inorganic Hg, as evidenced in the bar plot chart of Fig. 2.  (Fig. 3a,b and Fig. 4). In Fig. 3a, two distinct trends can be observed. The rst one corresponds to a positive correlation between total Hg and organic Hg whereas the second trend is mainly delineated by four samples (ASS8a, ASS8b, ASS17a and ASS21), which are characterized by an increasing concentration of computed inorganic Hg (up to > 80%) whilst that of organic Hg maintains almost unchanged. We can hypothesize that these soil samples are possibly indicating the presence of higher contents of residual mining materials. When these four samples are not considered, the correlation between the two parameters signi cantly increases as a Pearson coe cient of 0.92 (Fig. 3b) was computed. A similar positive correlation (r = 0.92) is also obtained when total Hg is plotted vs. inorganic-related Hg (Fig. 4). This interdependence is likely indicating that the fractionation of Hg compounds in the soils of the ASS mine is a distinct process unaffected by the relative position of the samples, the amount of organic matter present, or the activity due to enzymatic processes, as also reported by Campos et al. (2018). It is be noticed that the soil samples located in GO (Fig. 1b and Table S1) are more enriched in organic-related Hg with respect to those collected close to the mining facilities, suggesting that the proximity to the machineries to produce liquid Hg affected the soil matrix.
In addition, a second, weaker correlation (r = 0.5) between leachable Hg (in µg L − 1 ) and organic-related Hg (in mg kg − 1 ) is reported in Fig. 5. According to Campos et al. (2018), the most labile species of Hg are those containing organic Hg. However, it is to be pointed out that the high concentration of leached Hg (up to 20 µg L − 1 ) can also be released by solid phases and not necessarily only related to organic Hg. The concentration of DHA in the soils from the mining and production area appears to be even lower than those measured by Hinojosa et al. (2004) in the reclaimed area of the Aznalcollar mine (SW Spain). According to Pan and Yu (2011), the presence of PTEs in soils can have negative effects on the enzymatic activity, affecting either the enzyme-substrate complexation or the structure of the amino acids. In this case, the total Hg vs. DHA enzyme diagram (Fig. 6) shows a poor correlation (Pearson correlation r = 0.52, p < 0.05) with scatter distribution between the two parameters, suggesting that the presence of Hg, independently by its speciation, is not able to affect the microbial activity in the ASS soils, similarly to what observed by Campos et al. (2018) for the Hg-rich soils from Almadenejos. Figure 6 Scatterplot of DHA (mg TPF g − 1 d − 1 ) vs. Hg (mg kg − 1 ) in the soils from the ASS mining area. Blue circles: samples from CG, red circles: samples from FN, cyan circle: samples from GO and dark yellow circle: samples from TMB Considering the metals and metalloids usually found in the AAS ore deposits and the surrounding Hgmining areas (Rimondi et al., 2014a), the analytical spectrum should be enlarged to evidence whether, the enzymatic activity may be jeopardized by other PTEs (e.g. As and Sb).

Hg and BF in plants
The bar graphs in Fig. 7 depict the Hg distribution in each portion of the sampled plants, except for Acer pseudoplatanus and Salix spp., for which only one sample (bark trunk and root, respectively) was collected. Robinia pseudoacacia, Sambucus nigra, Castanea sativa and Popolus spp. are characterized by the highest Hg concentrations in the roots, as well as Salix spp. while the bark trunk of Acer is enriched in Hg. Different is the behavior of Cytisus scoparius and Verbascum thapsus as Hg is found in high contents in the foliage. Notably, is the fact that the highest Hg concentrations are related to the leaves of Cytisus scoparius, located in the TMB zone, where the gaseous elemental Hg in the atmosphere were found almost constantly up to 50,000 ng m − 3 or even higher (Vaselli et al., 2013). This is likely related to the fact that the leaf system is one of the main pathways of Hg uptake due to both dry deposition, as also suggested by Chiarantini et al. (2016) and Campos et al. (2018), as well as gaseous Hg from Hg-rich environments (such as that recorded in the air nearby the mining machineries and furnaces) and diffuse Hg from soil, although no Hg ux measurements are presently available.
The high Hg concentrations detected in the leaves of Verbascum thapsus (related to ASS 17 soil), collected from GO (Fig. 1b), can be explained by the main winds at ASS that blow from NNE/NE (https://www.meteoblue.com/it/tempo/historyclimate/climatemodelled/abbadia-san-salvatore_italia_3183581), thus, favoring the deposition of the Hg-rich atmospheric particulate and atmospheric Hg from TMB to GO (Fig. 1b). Therefore, according to the investigation on the different parts of plants analyzed in this study (Fig. 7), foliage is likely the main mechanism of Hg-uptake that can be invoked for the ASS plants, thus con rming previous investigations, e.g. Naharro et al., (2019).
Occasionally, roots seem to play a role in the Hg-uptake. However, further analyses on the leaf apparatus for those plants where the foliage was not collected are necessary.
The distribution of Hg between external and internal roots is reported in Fig. 8. All the studied root samples indicate that Hg concentration increases, as expected, in the external roots, with the exception of Sambucus nigra, showing an external root/internal ratio of about 1.7, i.e. more than one order of magnitude lower that those recorded for those samples characterized by Hg content > 20 mg kg − 1 . To the best of our knowledge, few are the studies related to the partitioning of Hg between internal and external roots and this calls for more detailed investigations.
According to Hussain et al. (2022), BF is a partitioning coe cient that mimics the ability of plants to absorb PTEs and, in this study, it was applied to the concentration of Hg in each part of the analyzed plants and that related to the soil leachable Hg. The BF values are highly variable when the different plant sectors are considered (Table S3). All BF values are < 1 (Table S3), but BF > 0.6 values correspond to the bark trunk and bark root of Sambucus nigra, and the leaves of Verbascum thapsus (0.63, 0.93, and 0.65 respectively). The BF values in the leaves of Verbasum thapsus seem to con rm that the leaves are likely the main path of Hg-uptake by plants. On the other hand, the relatively high BF values measured in the bark trunk and bark roots of Sambucus nigra are possibly due to the di culty in e ciently and completely removing all the soil-related particles during cleaning. This means that the concentration of mercury in the outermost part of roots and trunk is likely affected by the presence of soil material.

Conclusions
In this study, the distribution of Hg in soils and plants growing in the former mining area of Abbadia San Salvatore was investigated and, to best of our knowledge, DHA concentrations and the BF values were determined for the very rst time in one of the most important Hg sites worldwide. The content of Hg in the soils located in the highly Hg contaminated Sector 6 is heterogeneously distributed in the four, high Hg-contaminated, areas where the plants were collected (Fig. 1b). The highest concentrations were measured close to the edi ces hosting the Nesa and Gould furnaces, since here, when the mining works were active, mine tailings were used to ll a small paleo-valley. Thermal speciation allowed to recognize that inorganic Hg, presumably associated with cinnabar, with the exception of eight soils, was the prevailing species over the organic component. The low DHA concentrations, in agreement with Hinojosa et al. (2004), indicates that the area is likely contaminated by PTEs, although investigations are required to evidence their presence. However, the poor, though positive, correlation between total Hg and DHA shows that the Hg compounds do not affect the enzymatic action of DHA and do not inhibit but, conversely, enhance microbial activity. Mercury concentrations in different portions of the analyzed plants show that the main pathway of mercury uptake is the leaf system. For those samples where the foliage was not analyzed, roots are likely playing an important role in the uptake of Hg. However, more detailed investigations are needed to fully understand i) the role played by roots when developed in a Hgrich pedological environment and ii) the partitioning of Hg between external and internal roots. The Hg-BF is < 1 in all samples, indicating a weak translocation of mercury from the soil to the plant. In order to avoid possible errors in computing BF calculation, more careful and repeated washing of the different parts of the plants (especially those parts that are most in contact with the soil) is recommended. According to this study, phytoremediation projects should take into account the ability of Sambucus nigra to uptake Hg. A pilot site consisting of a Sambucus nigra plantation, positioned in a secure disposal location within the remediation area, could be established to verify whether the Hg removal is effective. Nevertheless, it can be recommended that more indigenous plants should be thoroughly analyzed to verify whether other plant species may have a stronger Hg adsorption capacity than that of Sambucus nigra.  Table S1