Experimental design
The experiment was carried out in a greenhouse located in the Fontes do Ipiranga State Park (PEFI), located in the city of São Paulo (São Paulo State, SE Brazil). This site contains an important urban remnant of the Atlantic Forest, from which the soil was removed to carry out this experiment. The soil that predominates in the park is Oxisol (Fernandes et al. 2002). The history of anthropic impacts on PEFI was divided into three distinct phases, according to dating carried out on sediments from a lake located next to the forest remnant, in parallel with heavy metal and polycyclic aromatic hydrocarbon (PAHs) analyses (Costa-Böddeker et al. 2012). Phase I (~ 1894-1975) was characterized by low atmospheric contamination, phase II (~ 1975-1990) by an abrupt increase in urbanization and air pollution levels and phase III (~ 1990-2005) by the activities of a neighboring steel company.
The experiment was carried out with six tree species native to the Atlantic Forest in São Paulo: Croton floribundus Spreng., Inga sessilis (Vell.) Mart., Rapanea ferruginea (Ruiz & Pav.) Mez (pioneer species), Eugenia uniflora L., Esenbeckia leiocarpa Engl. and Ocotea odorifera (Vell.) Rohwer (non-pioneer species), belonging to Euphorbiaceae, Fabaceae, Primulaceae, Myrtaceae Rutaceae and Lauraceae, respectively. All species seedlings were donated by the Energy Company of São Paulo (CESP), located in the city of Paraibuna, São Paulo State. The pioneer plants were, on average, 38 cm in height with a 4 mm stem diameter. The non-pioneer plants were 21 cm in height with a 3 mm stem diameter, on average. All seedlings had 5 leaves on average. These plants were transplanted into pots with a 1.7 L capacity, containing topsoil (0-20 cm) collected in the PEFI forest remnant, excluding the litter layer.
The chemical and physical characteristics of the topsoil used in the experiment were determined in five air-dried and sieved (through a 2 mm sieve) samples, as described by Raij et al. (2001). On average, the soil had a medium pH (5.3) and high levels of Presin (14 mg/dm3), Ca (66 mmolc/dm3), Mg (11 mmolc/dm3), S (27 mg/dm3), Cu (3.4 mg/dm3), Mn (10.3 mg/dm3), Zn (14.2 mg/dm3), B (58 mg/dm3) and Fe (57 mg/dm3). The soil still had a high content of organic matter (40 g/dm3) and a high cation exchange capacity (109 g/dm3). The soil was also characterized by a clay texture (fine sand: 110 g / kg; coarse sand: 427 g / kg; clay: 380 g / kg and silt: 74 g / kg).
Before plant transplantation, the soil was air-dried, homogenized, ground, passed through a 10 mm mesh sieve and kept moistened (80% of the field capacity) for 15 days. After one week of planting, all plants received 100 ml of nutrient solution containing macronutrients and micronutrients, as synthesized for the control treatment (TC) in supplementary material, Table S2.
After a one month adaptation period, the seedlings of all species were subjected to the following soil treatments: 1) control with balanced fertilization (referred to as TC); 2) balanced fertilization plus a higher dose of Ni (TNi); 3) balanced fertilization plus a higher dose of Zn (TZn); 4) balanced fertilization plus higher doses of Ni and Zn (TZnNi). The doses of Ni and Zn in the solutions used in treatments 2 to 4 (supplementary material, Table S1) were three times higher than the concentrations proposed by the Environmental Company of São Paulo State (CETESB) to prevent risks associated with soil contamination in the State (30 mg. Kg-1 of soil for Ni and 86 mg.kg-1 of soil for Zn; CETESB, 2016). These doses were reached by means of three weekly applications. The experiment lasted 90 days, starting from the first weekly application of the established doses. It was carried out in a 6 x 4 factorial scheme (six species x four soil treatments), with five replicates per treatment and five plants per treatment replicate, totaling 100 plants for each species.
Measurements and sampling procedures
Height and number of leaves on the main stem of each plant per treatment replicate was determined every 30 days.
At the end of the experiment, a total of 120 composed soil samples (five composed samples per treatment and species) were submitted to sequential extraction of Ni and Zn.
At the end of the experiment, the roots, stems/branches and leaves of each plant were also separated, and dried in an oven (60oC) to obtain the dry mass. After weighing, the samples of leaves, roots and stem/branches of the plants from each treatment replicate were gathered together in order to obtain five composed samples of each organ, per treatment and species.
Analytical techniques
The sequential extraction of Ni and Zn from the soil samples was performed by using the EC Standards, Measurement and Test Program protocol (called the BCR Test) (Janoš et al. 2010; Rauret et al. 1999). The Ni and Zn concentrations were determined in four fractions extracted sequentially from 1 g of each composed soil sample: soluble, exchangeable and bound to carbonates (F1); linked to iron and manganese oxides / hydroxides (F2); linked to organic matter (F3) and residual metals such as silicates (F4). F1 was extracted with acetic acid (0.11M) for 16 h at room temperature; F2 was extracted with hydroxylamine hydrochloride (0.5 M) for 16 h at room temperature; F3 was digested at an elevated temperature with hydrogen peroxide (30%), acidified to pH 2–3 and extracted with ammonium acetate (1 M) and F4 was digested at an elevated temperature with aqua regia (HCL: HNO3, 3:1). The extraction of F1, F2 and F3 was separated by centrifugation. The supernatant was removed, filtered and nitric acid (2N) was added in order to adjust the pH to 2.0. In the residual fraction (F4) was diluted with deionized water.
The composed samples of leaves, stems and roots were ground in an agate mill and subjected to acid digestion in a closed microwave oven (CEM-Mars-Xpres), using 5 ml HNO3. After digestion, the extracts were diluted with 10 ml of deionized water and stored in polypropylene flasks.
The concentrations of Ni and Zn in soil and plant extracts were determined by inductively coupled plasma-optical emission spectroscopy (ICP-OES), using a piece of Varian equipment, model 720. Analytical curves were prepared with an external standard from the brand Perkin Elmer. The concentration range used to obtain the curves for both elements was between 100 ppb and 1000 ppb. The detection limits were 0.001mg L-1 for Zn and 0.002mg L-1 for Ni. The quality control of the analyzes was verified by analyzing certified reference material (SRM 1515 – Apple Leaves).
Ni and Zn concentrations in the four soil fractions are expressed as mg g-1 of dry soil. The Ni and Zn in leaves, stems/branches and roots per plant are presented as absolute concentrations (concentrations * dry biomass of each organ per plant).
Ratio/index calculations
Soil
The concentrations and bioavailable levels of Ni and Zn in the contaminated soil were accessed by estimating: a) the sum concentrations found in the soil fractions in the TC, TNi, TZn or TNiZn treatments, as well as concentration ratios of each fraction, between the values found in TNi, TZn or TNiZn and those found in TC ; b) The soil’s mobility factor (MF), using the formula [MF = (F1 + F2) / (F1 + F2 + F3 + F4)] x 100 proposed by Kabala and Singh (2001) with adaptations (F1 to F4 in the formula represent the concentrations of Ni or Zn in the four fractions extracted from the soil).
Plant material
The concentration ratios in the different organs were estimated between the values found in each fraction of the soil in the TNi, TZn or TNiZn treatments and the values found in the control (TC). A heatmap was then proposed using these absolute concentration ratios to visualize the levels of general proportional accumulation of Ni and Zn in the different organs and tree species. All values were transformed into log10 (data included in Table S2; supplementary material) and a heatmap was generated using the public software Morpheus (https://software.broadinstitute.org/morpheus). The blue color in the scale (0 to 1) indicates a comparatively low to medium accumulation level and the red color (> 1 to 2) indicates a comparatively high to very high accumulation level.
Absolute concentrations (Ca) obtained from the different plant organs were also used to calculate a translocation index (TI), following a formula proposed by Chanu and Gupta (2016):
TI = (Ca stems-branches + Ca leaves) / Ca roots
TI> 1 indicates that the element accumulated in a greater proportion in the shoots than in the roots.
Finally, relative growth rates (RGR) for height and leaf number, as well as for the dry biomass of each plant organ, were estimated using the measurements performed at the beginning and end of the experiment, based on the equation proposed by Benincasa (1988):
RGR=[(Ln2-Ln1/t2-t1)]
Where: Ln2 and Ln1 = natural logarithm of the final and initial value; t2 and t1 = final time and initial time in days of exposure. The initial values for the dry biomass of each organ were obtained at the beginning of the experiment from an additional lot of plants of each species.
Statistics
Differences in the sum of the concentrations of Ni and Zn in the four soil fractions and the mobility factors of these metals in the TC, TNi, TZn and TNiZn treatments were identified using parametric One-way ANOVA (F test) and Tukey test. All the results obtained from the soil used for cultivating the six species were included in these statistics, considering that the different species did not alter significantly the metal concentrations in the soil fractions themselves.
One-way ANOVA, followed by a multiple comparison test (Tukey test), was also applied to compare: the sum of the absolute concentrations (concentration * biomass) of the leaves, stem/branches and roots of each species, to the soil treatments; concentration ratios ([Ni]treatments/[Ni]control; [Zn]treatments/[Zn]control) for the whole plants (leaves+stems+roots) to the species; translocation indices from roots to shoots to the species and RGR in height, number of leaves and the total biomass of each species between the treatments.
In all cases, the data were transformed into log10, when necessary to meet the requirements of ANOVA (normal distribution and equality of variations).