Introduction

Metal(loid)s are encountered in 50% of polluted soils (Khalid et al. 2016), which makes metal(loid) soil pollution an important issue worldwide, posing a great risk to both the environment and human health. Contaminated soils usually present a low diversity with a poor fertility and thus often lack of vegetation, increasing the risk of contamination spreading by leaching and erosion. Furthermore, when vegetation is present, plants often suffer from oxidative stress. Moreover, with the consumption of contaminated plants by animals, metal(loid)s can enter the food chain and cause great problems. In fact, many metal(loid)s, such as arsenic and lead, are classified as carcinogenic (Singh et al. 2015). Therefore, the remediation of such contaminated areas is the object of many researches.

One possibility is the use of plants to take up pollutants and store them in their different tissues. In addition, the implementation of a plant cover will reduce the risk of erosion and water leaching. Such technique is called phytoremediation (Thavamani et al. 2017). Salicaceae showed good potential due to their fast and important growth as well as their tolerance and accumulation capacity towards metal(loid)s (Bart et al. 2016; Lebrun et al. 2017). Metal(loid)s can be stabilized in the rhizosphere, i.e., accumulated in the roots, adsorbed on the root surface, and complexed with root exudates and rhizosphere microorganism compounds, with low translocation towards the plant upper parts. In this case, the phytoremediation is called phytostabilization. However, when metal(loid)s are translocated and (hyper)accumulated in the aerial parts, it is named phytoextraction (Awa and Hadibarata 2020).

However, due to the poor fertility of contaminated soils, amendments must be applied in order to diminish the toxic effects of the pollutants (Cristaldi et al. 2017). The application of diverse amendments such as biochar, compost, and iron grit to highly contaminated and poorly fertile soils improves the soil physico-chemical properties (reduction of acidity, supply of organic matter and nutrients) (Adriano et al. 2004; Lebrun et al. 2018a, b) and reduces metal(loid) stress (reduction of metal(loid) exchangeable, extractable fractions, and soil pore water concentrations) (Lebrun et al. 2018a, b). Such ameliorations allow a better plant growth.

As a biological process, phytoremediation success relies on plant development and activity. More particularly, root exudates can have an important role. Root exudates come from plant metabolism, especially photosynthesis. About half of the products obtained by the photosynthesis are transported to the roots, and among this, 12 to 40 % are released to the rhizosphere as exudates (Dong et al. 2007). The composition and amount of the different compounds in the root exudates are greatly influenced by the growing environment: presence or absence of particular minerals or toxic metal(loid)s, nutrient content, soil moisture, and soil texture (Badri and Vivanco 2009; Haichar et al. 2014). In return, root exudates can also affect phytoremediation by altering pollutant bioavailability as well as affecting the composition of plant associated microbiota (Gomez et al. 2019), due to the presence of compounds having either stimulatory or inhibitory effects on the microorganisms (Dennis et al. 2010). Among root exudates, organic acids are an important factor as they can increase plant tolerance to high metal(loid) concentrations (Goliński et al. 2015). They possess metal(loid) chelation properties that can either increase or decrease metal(loid) accumulation in plants (Bais et al. 2006; Seshadri et al. 2015). Thus, organic acid study is important to assess plant ability to remediate metal(loid)-contaminated soils.

In addition to the amendments, also microorganisms are important in the phytostabilization process. Indeed, even though metal(loid)s cannot be degraded, bacteria, for example, can control the transformation of metal(loid)s through various mechanisms, such as oxidation, reduction, (de)methylation, complex formation, and biosorption (Adriano et al. 2004), and influence the solubility and speciation of metal(loid)s (Wenzel 2008). Finally, the microbial diversity is hypothesized to play a crucial role in the stability of the ecosystem productivity, functions, and resilience towards stress (Wu et al. 2019)

Both microbial community composition and activities are important parameters for assessing the phytoremediation success. These can be evaluated through diverse techniques, for instance, by measuring soil enzymatic activities, community level physiological profiles (CLPP), and next-generation sequencing (NGS). Soil enzymatic activities are important indicators of soil fertility as well as the changes taking place within the soil environment (Mierzwa-Hersztek et al. 2016). Indeed, such activities reflect the status of the whole microbial community and represent one of the most reactive components of the soil (Epelde et al. 2008). Moreover, they are highly sensitive to excess concentrations of soluble metal(loid)s and have been recommended as biochemical indicators for the evolution of the soil quality during the remediation process (Touceda-González et al. 2017). It is generally recognized that high soil enzymatic activities imply a good quality of the soil (Mierzwa-Hersztek et al. 2016). However, the enzyme tests for soil allow to determine the activity of enzymes from both living and dead microbial cells (Al Marzooqi and Yousef 2017). In addition, soil microbial functionality can be assessed by using Biolog EcoplatesTM, which give the community level physiological profiles (CLPP) (Gomez et al. 2006). Finally, the composition of the microbial communities can also be revealed through sequencing the 16S rRNA gene. This high-throughput sequencing approach identifies the microbial communities based on OTUs (operational taxonomic units). Contrary to the two previous techniques, which give information on the activity of the microbial community, this technique informs on the composition in terms of domain, phylum, class, order, family, genus, and species (where possible).

Many studies analyzed the effects of amendments on soil, plants, and the microorganisms. However, root exudates have been mainly studied in hydroponic experiments or artificially contaminated soils (Agrawal et al. 2012; Hawrylak-Nowak et al. 2015; Javed et al. 2017), except in the study of Montiel-Rozas et al. (2016). In their study, root exudates were evaluated in artificially contaminated sand and in real contaminated soils, in association of amendments. However, these studies, evaluating amendment effect on root exudation profiles, remain at this day scarce. Furthermore, studies of the effects of combined amendment application on soil microbial communities are also scarce. The knowledge on the reaction of plants, at a biochemical levels, as well as microorganisms, to metal(loid) toxicity and amendment application needs to be deepened.

This study was a continuation of a previous work (Lebrun et al. 2019) evaluating, in mesocosm, the effects of amendment application to a mining soil on the soil physico-chemical properties, Salix viminalis growth, and metal(loid) accumulation pattern. Additionally, root exudates and soil samples were collected at the end of the experiment and analyzed. Accordingly, the aims of the present study were to evaluate the effect of biochar, compost, and/or iron grit amendment, as well as Salix viminalis growth, on (i) Salix viminalis organic acid exudation profiles, (ii) soil CHNS contents, (iii) soil enzymatic activities, (iv) soil CLPP, and (v) soil microbial community composition. The use of diverse techniques to assess the microbial community activity and structure allowed having the most complete picture of the soil communities and how they are shaped by amendments and/or plant growth.

Material and methods

Experimental design

The study focused on a former silver-lead mine extraction site, located in Pontgibaud (Auvergne-Rhône-Alpes, France). The mining activity led to high As and Pb accumulation in soil (Lebrun et al. 2017). The properties of the soil were determined in a previous study (Lebrun et al. 2019) and are as follows: pH 3.7, [As] 1.06 g·kg−1, [Pb] 23.39 g·kg−1, and organic matter content 2.6 %.

The experimental design and the first results of the study have been fully described in a previous paper (Lebrun et al. 2019), but a brief description of the experiment will be reminded here. The Pontgibaud technosol (P) was mixed or not with three different amendments: a hardwood biochar added at 5% (B), a commercial compost added at 5% (C), and an industrial iron grit added at 1.5% (I), giving seven treatments in total (P0%, PB, PC, PI, PBC, PBI, and PBCI). One non-rooted cutting of Salix viminalis was placed in 14 pots, and five were left un-vegetated. The experiment lasted for 69 days, under greenhouse conditions: temperature 22 ± 2 °C, light intensity 800 μmol.m−2·s−1, and photoperiod 16 h. Pots were irrigated every 2 days based on the water lost through evapotranspiration.

Root exudate collection and analysis

After 69 days of growth, plants were harvested and subjected to different treatments. Three plants were used for the analysis of the root exudates: the plants were removed from the pot, and the roots were carefully washed in distilled water until no soil particles were observed on the roots. Following, the roots were immerged in milliQ water during 4 h to let plant roots release organic acids. The volume of milliQ water was adjusted depending on the root system of each plant in order to cover all the roots (Aulakh et al. 2001). After 4 h, at room temperature, the plants were removed, and the solutions were filtered to remove any particle left and put at −20 °C. The frozen solutions were then lyophilized, and the dry powder was recovered, weighted, and solubilized in acetone, at different volumes (between 0.5 and 2 mL) depending on powder quantity (from 0.4 to 10 mg/ml). These extracts were kept at −20 °C until further analysis.

Based on previous studies performed on willow (Drzewiecka et al. 2014; Goliński et al. 2015) and especially on Salix viminalis (Mleczek et al. 2018), five organic acids were analyzed: citric acid, fumaric acid, malic acid, succinic acid, and tartaric acid. These organic acids are chelating substances and are known to have an affinity and form complexes with metal(loid)s and, thus, are important in the stress tolerance (Bais et al. 2006; UdDin et al. 2015).

The samples were analyzed by ultra-high-performance liquid chromatography (UHPLC) (Destandau et al. 2005) using an Ultimate 3000 RSLC system (Dionex, Germering, Germany) consisting of a binary pump, an online vacuum degasser, an autosampler, and a column compartment. Separation of compounds was achieved on a Phenomenex Luna Omega PS C18 column 1.6 μm, 100 mm × 2.1 mm, kept at 25°C. Mobile phase A was water containing 0.1% formic acid; mobile phase B was acetonitrile containing 0.1% formic acid. The flow was 0.5 mL/min, and the gradient profile was 0 at 50% B in 5 min, return in 0.1 min to initial conditions. The injection volume of both the standard solutions and the samples was 20 μL. Equilibration time between two injections was 2 min.

UHPLC was coupled with mass spectrometry detection. It was performed on a TSQ Endura triple quadrupole mass spectrometer (Thermo Fisher Scientific Inc., Waltham, MA) instrument equipped with an electrospray ionization ion source (H-ESI). Capillary voltage was −2.5 kV in negative mode; the vaporizer temperature was 500°C; and ion transfer tube temperature was 380°C. Gas flow was in arbitrary unit, and sheath gas was fixed at 27, Aux gas at 9, and sweep gas at 0. MS acquisitions were done in full scan mode from m/z 80 to m/z 250. Data processing was done using Xcalibur version 3.0 SP2.

Standard curves (seven points) were prepared for each organic acid to allow compound identification as well as concentration determination in the extracts. Organic acid concentration was calculated based on root fresh weight.

Soil sampling and analysis

At the beginning of the experiment, soil from each of the seven treatments was sampled for physico-chemical analysis. After 69 days, soil samples were collected from each pot, both vegetated (rhizosphere, n = 14) and un-vegetated (bulk, n = 5). For the vegetated condition, rhizosphere soil, i.e., the soil attached to the roots, was collected by shaking the roots inside a sterile plastic bag. For each treatment and condition (bulk/rhizosphere), a composite sample was put together by mixing the same volume of the samples of all the replicates of each treatment (Quoreshi et al. 2019; Stacey et al. 2019). The composite samples were used to perform the following analyses.

CHNS content analysis

On all the substrate samples, collected at T0 and T69 in non-vegetated and vegetated pots, the total C, H, N, and S contents were determined, using a Flash 2000 organic elemental analyzer (Thermo).

Soil enzymatic activities

Three enzymatic soil activities were measured: acid phosphatase, alkaline phosphatases, and fluorescein diacetate (FDA) hydrolysis. The methods described below have been optimized based on previous published protocols in order to have the best recovery of activity in our soil conditions.

Acid and alkaline phosphatases were determined using PNPP (p nitrophenyl phosphate disodium) substrate: 2 g of air dried soil was mixed with 2 mL of buffer (sodium acetate 0.1 M, pH 5 for acid phosphatase; Tris-HCl 0.1 M, pH 8 for alkaline phosphatase) and shaken overnight (150 rpm, ambient temperature). Suspensions were centrifuged and the supernatants collected. The reactions were then realized in a microplate: 100 μL of extract were mixed with 100 μL of PNPP 5 mM or buffer (for control), and the plate was incubated for 1 h at 25 °C. The reaction was stopped by adding 100 μL NaOH 0.1 M, and absorbance was read at 410 nm (against blank reactant). Calculations were done using ε (PNPP) = 19500 L·mol−1·cm−1.

For the FDA hydrolysis, 0.1 g of air dried soil was mixed with 5 mL of potassium phosphate buffer (60 mM, pH 7.6) and 50 μL FDA (50 mM), except in control tubes. The tubes were shaken for 3 h (150 rpm, 37 °C). Suspensions were centrifuged, and then 200 μL of the supernatant were put in the well microplate, and absorbance was measured at 490 nm. Calculations were done using ε (FDA) = 8000 L·mol−1·cm−1.

Biolog EcoplatesTM

The Biolog EcoplatesTM are based on the ability of microorganisms to oxidize different carbon substrates (Gomez et al. 2006). They consist of plates of 96 wells containing three replications of 31 different carbon sources plus a blank. When the carbon source is used by the cultivable portion of the microbial community, the tetrazolium violet present in each well is reduced, and a purple color develops (Cesarano et al. 2017; Epelde et al. 2008). This color development can be assessed spectrophotometrically. Ecoplates are a simple and rapid method for the observation of the biological response of the microbial community (Cesarano et al. 2017; Epelde et al. 2008).

To obtain the microbial suspensions, fresh soil (2.5 g) was mixed with 10 mL of sterile NaCl solution (0.9%) and vortexed for 3 min. Soil suspensions were let to settle, and 600 μL of supernatant were mixed with 17.4 mL of sterile NaCl solution (0.9%). Then, 150 μL of the microbial suspensions were put in each well of the microplate. One plate was prepared per treatment. The plates were incubated at 25 °C, and the absorbances were measured at 590 nm every day for 2 weeks. Absorbance data were used to calculate several indices:

  • Average Well Color Development: AWCD = average of absorbance values.

  • Shannon-Weaver index H’ = Σ pi * ln pi, with pi = proportion of the absorbance.

  • Eveness E = H’/ ln 31.

  • Richness = number of wells having an absorbance > 0.25.

Next-generation sequencing

Soil DNA was extracted using a PowerSoilTM DNA Kit (MO BIO Laboratories, Inc), following the manufacturer’s instructions. The NGS sequencing protocol was performed at the INRA Transfert (Narbonne, France): extracted DNAs were amplified by PCR using the primers 515F (5′-GTGYCAGCMGCCGCGGTA-3′) and 909R (5′-CCCGYCAATTCMTTTRAGT-3′) which target the variable regions V4–V5 of the 16S ribosomal RNA gene of the prokaryotes. The sequencing was performed by MiSeq Illumina. Following, sequences were analyzed using FROGS pipeline (Escudié et al. 2018): artefact sequences were removed using the tools pre-process, swarm clustering, chimera removal, and filters. The remaining sequences were grouped in operational taxonomic units (OTUs) that were affiliated to a taxa by using the SILVA 138 16S rRNA gene database. Finally, FROGSSTAT was used to calculate alpha diversity indices. In addition, Bray-Curtis dissimilarity coefficients were calculated and represented using multidimensional scaling (MDS) plot and clustering.

Statistical analyses

Data were analyzed using R software version 3.1.2 (R Development Core Team 2009). The global treatment (seven treatments) and plant (bulk/rhizosphere) effects were assessed using the following procedure: normality and homoscedasticity of the data were evaluated using Shapiro and Bartlett tests, respectively; and the treatment effect was then measured using an Anova or a Kruskal test, while plant effect was determined using a Student or a Wilcox test, for normal and non-normal data, respectively. For a more in-depth analysis, treatments were compared two at a time using the Student test for normal data and the Wilcox test for non-normal data. Bulk and rhizosphere conditions were compared within each treatment using the same protocol. Difference was considered significant at p < 0.05.

Results

Organic acids exuded by the roots

Among the five organic acids, citric and malic acids were the most abundant in all conditions except PI, which did not present those two acids but presented elevated contents of succinic acid and to a lesser extent tartaric acid (Fig. 1).

Fig. 1
figure 1

Organic acids (citric (A), fumaric (B), malic (C), succinic (D), and tartaric (E)) contents (μg·g−1 fresh weight) in Salix viminalis roots determined after 69 days of growth on Pontgibaud technosol (P) alone and amended with 5% biochar (B), 5% compost (C), or 1.5% iron grit (I), alone or combined. Letters indicate significant difference among each organ (p < 0.05) (n = 3 ± SE)

Full size image

After 69 days of growth, the content in citric acid (C6H8O7) in Salix viminalis root exudates was higher in the biochar and compost-amended substrates, compared to P0%, although this increase was significant only for their combined application (PBC) (Fig. 1A). Moreover, its content was below detection limit in PI and PBI.

Fumaric acid (C4H4O4) was not detected in P0%, PI, and PBI, while low concentrations were detected in the biochar and/or compost-amended treatments, although such higher fumaric acid exudation was not significantly different from P0% (Fig. 1B). Moreover, fumaric acid presented the lowest level of the five organic acids measured. Depending on the treatment, its levels were 9 to 150 times lower than the citric acid, 3 to 190 times lower than the malic acid, 2 to 4 times less than the succinic acid, and 2 to 55 times lower than the tartaric acid.

Although there was no significant difference between P0% and the amended conditions, malic acid (C4H6O5) content in Salix root exudates tended to be higher in PBC. It was below detection limit in PI (Fig. 1C).

No significant variation was observed in the succinic acid (C4H6O4) concentrations (Fig. 1D), and contents were low in all conditions. However, its content tended to be higher in PI.

Similar to succinic acid, no significant difference in tartaric acid (C4H6O6) amounts between the seven treatments was observed. However, concentrations were very low in P0%, as well as in PB, PC, PBC, and PBCI, while it was higher in PI and PBI (Fig. 1E).

Soil CHNS contents

Soil CHN contents were highly significantly affected by treatments, while plant development had no significant effects. Soil S content was significantly affected by treatment and plant (Table 1). Results showed that at all sampling time, in vegetated or non-vegetated conditions, carbon content increased in all amended treatments compared to P, except in PI. In general the best treatments were PBC and/or PBCI (Table 2). When comparing the sampling times, in PB and PI treatments, carbon content increased with time, while plant decreased C content; in PBC it increased between T0 and T69; plant growth on PC decreased C content, whereas C content decreased with time in PBCI, but plant growth increased C content. At the beginning of the experiment, hydrogen content was decreased by iron amendment, was not affected by the combination biochar-iron, and increased with all the other amendments. At the end of the experiment, in both rhizosphere and bulk, all amendments increased H content compared to P, and the best treatment was PBCI. Only few modifications in H content were observed between T0 and T69: in P and PBCI, H content decreased with time, while in PI, PBC, and PBI, it increased with time. Plant decreased H content in PBI (Table 2). The nitrogen content increased in the three treatments containing compost (PC, PBC, and PBCI) compared to P, at T0. In the bulk soil, at T69, N content increased in all treatments except PI and PBC. In the rhizosphere soil, at T69, only PBC, PBI, and PBCI increased N content (Table 2). When comparing values at T69 and T0, it can be seen that N content decreased with time in P, PB, and PBCI. Finally, sulfur content was only affected by amendments at the beginning of the experiment, in which it decreased in all treatments except PBC (Table 2). Moreover, with time, sulfur content decreased in P and PBC compared to T0.

Table 1 Treatment and plant effects, determined on the basis of measurements of three soil enzymatic activities (acid phosphatase, alkaline phosphatase, and hydrolytic (FDA) activities) as well as on the parameters measured with the Biolog Ecoplate test (average well color development (AWCD), Shannon-Weaver index H’, Eveness E, and richness)
Full size table
Table 2 Total soil carbon, hydrogen, nitrogen, and sulfur content (%) determined at the beginning of the experiment (T0) and at the end of the experiment in both non-vegetated (T69-Salix) and vegetated (T69+Salix) conditions on Pontgibaud technosol (P) alone and amended with 5% biochar (B), 5% compost (C), or 1.5% iron grit (I), alone or combined
Full size table

Soil enzymatic activities

Three soil enzymatic activities were measured at the end of the experiment for the different modalities, for both non-vegetated (bulk) and vegetated (rhizosphere) soils: acid phosphatase, alkaline phosphatase, and hydrolytic activity (against FDA).

Global treatment effect was highly significant (p < 0.001) for the three activities, whereas plant effect was significant (p < 0.01) only for alkaline phosphatase (Table 1).

Acid phosphatase activity increased following the addition of biochar, compost, and iron grit alone in the bulk soil, while in the rhizosphere, it decreased with biochar-compost treatment and increased in the three treatments containing iron grit (Fig. 2a). The presence of plant decreased acid phosphatase activity in PI and PBI treatments, while plant activity increased it when grown on PC and PBC. Alkaline phosphatase activity increased with all amendments except for biochar alone in bulk, while it increased with all amendments except for biochar and biochar-iron treatments in the rhizosphere (Fig. 2b). Salix viminalis plant growth increased alkaline phosphatase activity on P, PB, and PC treatments.

Fig. 2
figure 2

Soil enzymatic activities: acid phosphatase (a) (nmol PNP·g−1·min −1), alkaline phosphatase (b) (nmol PNP·g −1·min−1), and hydrolytic (c) (nmol FDA·g−1·min−1) measured on Pontgibaud (P) technosol amended or not with biochar (B), compost (C), or iron grit (I), alone or combined, after 69 days of Salix viminalis growth on bulk (gray bar) and rhizosphere (black bar) soils. Minuscule letters indicate significant difference (p < 0.05) between the treatments in the bulk condition, while capital letters indicate significant difference in the rhizosphere condition. Significant difference between bulk and rhizosphere condition is indicated by * (p < 0.05), ** (p < 0.01), and *** (p < 0.001)

Full size image

The overall hydrolytic activity was assessed by the FDA hydrolysis. Compared to the non-amended technosol, this activity greatly increased in the three compost treatments as well as with iron alone in the bulk soil. In the rhizosphere, it decreased following biochar application and increased in the three compost-amended substrates. It has to be noted that the level of response of the FDA hydrolytic activity was higher in the bulk compared to the rhizosphere compartment (Fig. 2c). Finally, plant growth decreased the hydrolytic activity of the soil in three cases: PB, PC, and PBC.

Altogether, this data showed that in general, soil amendment increased the soil enzymatic activities, and plant development had a reduced effect on soil enzyme activities.

Community level physiological profiles (Biolog EcoplatesTM)

In addition to the soil enzymatic activities, Biolog EcoplateTM tests were performed to assess the microbial functional diversity. The values of absorbance at 590 nm were used to calculate several indices: the average well color development (AWCD), the Shannon-Weaver diversity index (H’), the Eveness (E), and the richness.

Treatment had a highly significant (for AWCD and richness) and a very significant (for H’ and E) effects on the parameters measured, while plant had no significant global effect for any of the parameters (Table 1).

In the non-vegetated condition (bulk), AWCD values increased in PC, PI, and PBC treatments compared to P (Table 3), with the highest increase observed in PC (6.1-fold). In the vegetated condition (rhizosphere), AWCD only increased in the compost treatment (PC). Plant growth only affected AWCD values for P and PC substrates, in which it increased and decreased AWCD values, respectively.

Table 3 Biolog Ecoplate data (AWCD average well color development, H’ Shannon-Weaver index, E Eveness; Richness, number of positive well (OD590 > 0.25)) measured at the end of the experiment (T69) on the bulk and rhizosphere soil of Pontgibaud (P) amended or not with biochar (B), compost (C), or iron grit (I), alone or combined
Full size table

The Shannon-Weaver index (H’) and Eveness (E) developed similar trends. In the bulk compartment, they only increased in PC compared to P (Table 3), while in the rhizosphere zone, they decreased with the addition of biochar-iron grit (Table 3). Moreover, plant development led to a decrease in H’ and E for PC and PBI treatments.

The richness index represents the number of wells having an absorbance above 0.25; this reflects the number of utilized carbon substrates. Richness increased in all the amended treatments, except PB and PBCI compared to the non-amended Pontgibaud soil, in the bulk condition. The highest increase (5.7-fold) was observed in PC and the lowest (2-fold) in PBI (Table 3). For the rhizosphere compartment, it only increased following the single compost amendment (Table 3). Again, when regarding plant effect for each treatment, plant growth had little and contradictory effects on richness: it doubled richness value on P, whereas it decreased it on PC substrates.

Furthermore, for a deeper analysis, the different carbon substrates were grouped into six categories (carbohydrates, amino acids, carboxylic acids, polymers, amines, and phenolic compounds), and the percentages of utilization of the different categories were calculated based on the AWCD. The statistical analysis revealed a highly significant treatment effect for the carbohydrates, a very significant one for the carboxylic acids, and a significant one for the amino acids, whereas treatment had no effect on the three other substrates, and plant development did not show any significant effect for any of the six categories (Table 1). Moreover, the results showed that carbohydrates were the most used substrates (Fig. 3), while amines and phenolic compounds were the least. Those two last substrate categories, amines and phenolic compounds, and carboxylic acids did not show change in relative utilization following amendments in both bulk and rhizosphere compartments. Overall, few changes in the percentage of utilization were observed in the amended treatments compared to P. In the bulk condition, amino acid use increased in PC, while polymers increased in PB and decreased in PBI. In the rhizosphere, only carbohydrate utilization was decreased in PBC. Even though the statistical analysis showed few significant changes in utilization categories, it can be seen from Fig. 2 that, contrary to the other treatments, PBC presented a more diverse profile, with the six categories having more or less similar utilization percentages in both bulk and rhizosphere

Fig 3
figure 3

Substrate utilization (%) by categories (carbohydrates, amino acids, carboxylic acids, polymers, amines, phenolic compounds) determined by the Biolog Ecoplates measured at the end of the experiment (T69) on the bulk (b) and rhizosphere (r) soil of Pontgibaud (P) amended or not with biochar (B), compost (C), or iron grit (I), alone or combined. Minuscule letters indicate significant difference between the three treatments for the bulk condition, while capital letters indicate significant difference between the three treatments for the rhizosphere condition (p < 0.05) (n = 5). Significant difference between the bulk and rhizosphere conditions is indicated by * p < 0.05, ** p < 0.01, and *** p < 0.001

Full size image

Microbial community analysis

The characterization of the microbial communities by using NGS showed a dominance of the phyla Proteobacteria, Acidobacteria, Actinobacteria, Verrucomicrobia, Bacteroidetes, Chloroflexi, Planctomycetes, Gemmatimonadetes, Desulfobacteria, Halobacteria, Crenarchaeota, Cyanobacteria, and Myxococcota (Fig. 4). After the addition of biochar and iron grit, alone or in combination, an increase in relative abundance of Proteobacteria was observed compared to the unamended soils or with compost. Moreover, a decrease in Acidobacteria was shown in all the treatments. The relative abundance of Bacteroidetes increased significantly in all amended treatments except iron grit alone (in the bulk soil compartment), whereas Actinobacteria increased in biochar and iron grit-treated samples (alone and combined). Chloroflexi proportion increased in all the compost treatments compared to the non-amended Pontgibaud technosol, whereas the proportion of Crenarchaeota and Cyanobacteria decreased in the amended conditions.

Fig. 4
figure 4

Phylum-level microbial community composition in the bulk (b) and the rhizosphere (r) compartments of Pontgibaud (P) amended or not with biochar (B), compost (C), or iron grit (I), alone or combined. Phyla with relative abundance < 1% are grouped in the category dothers

Full size image

Regarding the alpha diversity indices (Table 4), in all the amendment treatments, a higher number of OTUs was observed compared to unamended Pontgibaud, except for iron grit alone and combined with biochar, which induced a decrease in the OTU number. Similarly, Shannon and InvSimpson indexes increased with the amendment treatments, except for PI and PBI treatments, which had lower index values than the non-amended Pontgibaud (Table 4). This shows that except when iron grit was added alone or combined to biochar, amending Pontgibaud technosol increased the diversity of the soil microbial communities.

Table 4 Alpha diversity index of the bulk and rhizosphere soils of Pontgibaud (P) amended or not with biochar (B), compost (C), or iron grit (I), alone or combined
Full size table

The MDS ordination based on the Bray-Curtis dissimilarity values (Fig. S1) showed a small effect of plant, as the bulk and the rhizosphere soil communities of the same treatments are close, except for the treatment PBCI. In addition, different clusters can be seen in this two-dimensional plot (represented by the circles in the figure): all the treatments containing compost are grouped together on the left, the treatments containing biochar are grouped at the bottom of the plot, while the treatments with iron grit are located on the right of the plot and the unamended treatments at the top right. This shows that treatments with the same amendment have microbial communities of similar structure.

Similarly, the clustering of the communities (Fig. S2) separates the treatments in two groups: the one containing the compost are grouped on the left, while the other treatments, without compost, are on the right. In the group “compost,” two sub-groups can be seen: biochar+compost vs. compost and biochar+compost+iron grit. The other group is also divided in two: a sub-group containing the treatments with biochar and the other sub-group formed by the treatments without biochar (P and PI).

Discussion

Organic acids exuded by roots

The results showed a general increase in organic acid exudation in most of the amended conditions presenting a reduced metal(loid) stress, in contradiction with previous studies. For instance, Javed et al. (2013) showed that metal stress increased organic anion concentrations in the rhizosphere. However, such improvement could be related to a better growth and leaf surface area observed in most of these amended conditions, as shown in our previous study (Table S1, Lebrun et al. 2019). Indeed, a higher leaf surface means a more important photosynthetic surface and thus more products synthesized.

Furthermore, malic and citric acids were the most important organic acids measured in all conditions, except PI and PBI. These observations are consistent with the study of Zeng et al. (2008), in which the authors observed that oxalic and malic acid contents were higher than the four others (lactic, acetic, citric, succinic acids) in the root exudates of rice plants exposed to chromium. Similarly, Haoliang et al. (2007) identified nine organic acids in the root exudates of mangrove plants exposed to cadmium and found that acetic, lactic, malic, and citric acids were the most abundant. Such higher contents of these two organic acids can be explained by their high ability to complex with metal(loid)s (Seshadri et al. 2015). Malic and citric acids have a function in microbial mineralization, phosphorus mobilization, aluminum detoxification, as well as in the mobilization of several elements (Fe, Cu, Mn, Ca, Mg, Zn, and Ni) (Martin et al. 2014). They tended to increase with biochar and compost-amended conditions, in which plant growth was better and metal(loid) accumulation was reduced. Such findings are in contradiction with previous studies showing higher citric and malic acid contents in root exudates after treatment with metal(loid)s. For instance, the treatment of Solanum nigrum L. by increasing levels of chromium raised root exudation of citric acid (UdDin et al. 2015). Similarly, Javed et al. (2017) showed that citric acid exudation increased with increasing cadmium concentration treatment. Our previous study showed that metal(loid) stress was reduced with biochar and/or compost amendments (Lebrun et al. 2019), but soil pore water (SPW) As concentrations were observed to increase in PB and PBC conditions, which could explain the increase in malic and citric acid exudation observed here. Moreover, such more important exudation of malic and citric acids in response to higher SPW As concentration could have led to the chelation of As to render it less available (Pinto et al. 2008), as root As concentrations were observed lower in those two conditions even with higher concentrations in SPW (Table S1, Lebrun et al. 2019). In addition, SPW Pb concentrations were decreased in the vegetated pots containing biochar and compost, compared to P0%. This could be attributed to the uptake of Pb by the plant but also to a chelation of Pb by malic and citric acids, since root and leaf Pb concentrations were also decreased with biochar and compost amendments (Lebrun et al. 2019). Indeed, Duarte et al. (2007) observed that with increasing concentrations of citric acid, Ni uptake was decreased due to the formation of a complex that prevented its uptake, and Cd translocation to the upper parts was decreased.

Citric and fumaric acids were not detected in PI and PBI, which could be explained by a low leaf surface area (Lebrun et al. 2019) but also by the high levels of Fe measured in SPWs. Both organic acids are potential complexing reagents for Fe in soil but can also induce the dissolution of non-available ferric oxyhydroxide to the soil solution (Jones 1998). Since here SPW Fe levels were already high, dissolution was not necessary.

Regarding the other organic acids, the low exudation of fumaric, succinic, and tartaric acids by the plants in all the conditions tested, except in PI for fumaric acid, could mean that these organic acids had no major role in As and Pb detoxification and tolerance, while they had a more important role in Fe tolerance.

However, more than metal(loid) mobility, it is their speciation that can greatly affect plant root activity and exudation profiles; in return, compounds exuded by the roots can affect metal(loid) speciation and thus toxicity. For instance, As(III) is known to be more toxic than As(V), and thus it would potentially affect more negatively plant exudation. However, such hypothesis could not be verified here, as only metal(loid) concentrations were measured and the speciation of As and Pb was not assessed. It would be thus interesting to look more deeply the effect of amendments on metal(loid) speciation and consequently plant root exudation profiles, as well as the effect of root exudates on metal(loid) forms.

Soil CHNS contents

The increases in terms of soil C, H, and N contents observed with the different amendments were in accordance with previous studies. In 2012, Jones et al. observed an increase in soil total carbon content from 2.27 to 2.78 % 3 years after the application to a sandy soil of 50 t·ha−1 of wood biochar (Fraxinus excelsior, Fagus sylvatica, and Quercus robur). Moreover, in 2015, Agegnehu et al. showed that the application of biochar and/or compost increased the total N content of the soil. Such improvements can be explained by the amount of C, H, and N found in the biochar and the compost used as amendment (Nigussie et al. 2012). On the contrary, the non- or negative effect of the amendments on soil S content can be explained by the low S content of the biochar and the compost. Moreover, the higher soil C, H, and N content improvements usually observed with biochar compared to compost can be attributed to the fact that the C, H, and N elements are more stable in biochar than in compost (Laird et al. 2010). Regarding the evolution of the soil CHNS contents during the experiment time course, although the increase in the non-vegetated condition compared to T0 was not expected and thus difficult to explain, the decrease observed in some treatments can be attributed to a leaching of CHNS. Moreover, plant growth can affect CHNS soil contents, resulting in either a decrease or an increase, through nutrient uptake or release of root exudates that contain these elements.

Soil enzymatic activities

Phosphatase enzymes are involved in the mineralization of organic phosphate, and they develop a rapid response to soil management (Al Marzooqi and Yousef 2017; Yang et al. 2016). Similar to what was observed here, compost has been shown to increase phosphatase activities. For instance, municipal solid waste compost applied on the tailings of a copper mine increased acid and alkaline phosphomonoesterase activities (Touceda-González et al. 2017). Moreover, Ros et al. (2006) applied different kinds of composts (urban organic waste, green waste, cattle manure, and sewage sludge) to the field, and all such composts increased phosphatase activity (evaluated at pH 6) compared to the control plot. Similarly, the application of biochar to soil induced an increase in alkaline phosphatase activity in several studies (Al Marzooqi and Yousef 2017; Oleszczuk et al. 2014). However, it was surprising to observe in this study an increase in acid phosphatase activity in the biochar-treated substrate, as an increase in SPW pH was observed in this case (Table S1, Lebrun et al. 2019). However, previous studies observed either an increase or a decrease in acid phosphatase activity after biochar application. Indeed, Liu et al. (2017) and Pukalchik et al. (2018) showed a decrease in acid phosphatase activity, whereas Oleszczuk et al. (2014) observed a rise in the same activity following biochar amendment. Moreover, Yang et al. (2016) evaluated the effect of two biochars and dose application and found that applying 5% bamboo biochar decreased acid phosphatase activity, while 1% fine rice straw biochar increased it.

The FDA activity is related to the overall microbial activity (Pukalchik et al. 2018). Indeed, FDA can be hydrolyzed by several different enzymes, which makes the FDA hydrolysis activity an index of the overall enzyme activity (Lee et al. 2008). Similar to the present study, Pardo et al. (2014) observed an increase in FDA hydrolysis after compost amendment, indicating a stimulation of the soil microbial community, while Yun et al. (2017) observed a rise in FDA hydrolysis when both biochar and compost were added and no effect when added alone.

Overall, amendment application increased soil enzyme activities, indicating an increase in the microbial activity and/or biomass (Khadem and Raiesi 2017; Karaca et al. 2010). Such improvement can be explained by an amelioration of the soil physico-chemical properties (Khadem and Raiesi 2017; Touceda-González et al. 2017) (increase in pH, EC, water holding capacity), a supply of organic matter by the amendments used (Nie et al. 2018), and a decrease in metal(loid) stress (Khan et al. 2007). Indeed, the measure of the soil physico-chemical properties (Lebrun et al. 2019) showed that amendment application increased soil pore water pH (Table S1) as well as soil water holding capacity and organic matter content, while decreasing metal(loid) availability (CaCl2- and NH4NO3-extractable fractions) and mobility (soil pore water concentrations). In turn, this could have reduced the negative pressure put on microorganisms and thus improved their growth and activity.

Finally, although plants are known to increase enzymatic activity through the gradual formation of the rhizosphere microbiota and the release of enzymes (Nannipieri et al. 2007, 2011), both an increase and a decrease were observed in this case. This could be due to different root exudations depending on the treatments, which could have a different effect on the microbial community composition and activity, as shown for the organic acid root exudation: citric, fumaric, and malic acids tended to increase with biochar and/or compost, while succinic and tartaric acids increased in PI and PBI. Moreover, in some cases, the plant cover can induce a loss of carbon due to mineralization, which leads to a decrease in enzyme activity (Pascual et al. 2000).

Microbial functional diversity (Biolog EcoplatesTM)

The AWCD is an important index of the microbial functional diversity, which represents the ability of the soil microorganisms to use different carbon sources (Zhu et al. 2017). The increases in AWCD values observed in this study have been previously observed, in other researches, following compost and biochar amendments (Liao et al. 2016; Pardo et al. 2014) and reflect the increase in microbial abundance and the stimulation of the soil microbial community (Chen et al. 2016; Pardo et al. 2014). Indeed, amendments can provide a habitat for the soil microorganisms and stimulate their activity (Liao et al. 2016). They can also provide nutrients (Perez-Piqueres et al. 2006) and reduce stress pressure (Garau et al. 2017), such as metal(loid) toxicity. Indeed, Kenarova et al. (2014) found that AWCD value was negatively correlated to As, Cu, Pb, and U contents, demonstrating the adverse effect that pollution can exercise on microbial activity. Similarly, the soil data (Lebrun et al. 2019) showed that amendment applications reduced CaCl2- and NH4NO3-extractable As and Pb as well as soil pore water Pb concentrations (Tables S1 and S2). Additionally, soil pore water toxicity decreased at the end of the experiment in the amended substrates, indicating a reduction of the stress, and may have led to an improvement in the microbial community growth and activity. Moreover, the higher increase observed in the PC treatment could be related to the elevated nutrient and organic carbon levels of the compost but also to the addition of microorganisms from the compost itself.

The functional diversity H’, or Shannon-Weaver index, can be defined as the number of distinct functions carried out by a microbial community (Boshoff et al. 2014). The increase induced by compost in the bulk condition can be explained by the incorporation of organic matter (Pardo et al. 2014) and the lowering of metal(loid) availability, rendering them less toxic to both plants and microorganisms (Hmid et al. 2015), as shown by the soil measurements (Lebrun et al. 2019). The non- or negative effect of biochar seemed surprising as these data showed that biochar application improved soil conditions. However, such results have also been found by Zhu et al. (2017) and could be attributed to a possible release of toxic compounds by biochar, which could affect the activity of some microorganisms. Moreover, in their study, authors also revealed the negative relation between Shannon-Weaver H’ index and soil pH. High pH could also contribute to the non- or negative effect of biochar, as PB showed a higher SPW pH than PC but lower than PBC. However, contrary to the biochar-compost combination, biochar alone or combined with iron grit could have not provided enough nutrients to sustain microbial diversity and activity (Lebrun et al. 2019).

The improvement in substrate utilization (richness) could be attributed to a supply of organic matter (Pardo et al. 2014), especially for compost, as well as a reduction in metal(loid) toxicity (as shown in Lebrun et al. 2019) that rendered the substrates more suitable for microbial growth and activity (Hmid et al. 2015).

The analysis by functional substrate categories showed that biochar and compost amendment shifted the microbial community structure, with a more diverse C source utilization. This can be attributed to the direct addition of microorganisms with compost and an increased survival when a carbon rich material (biochar) is also present in the soil.

Microbial community analysis

An understanding of the temporal and spatial structures, functions, interactions, and population dynamics of microbial communities is critical for many aspects of life, including scientific discovery, biotechnological development, sustainable agriculture, energy security, environmental protection, and human health (Bucci et al. 2017). Accordingly, several methods (cultivation-dependent and molecular approaches) have been employed to reveal microbial community composition and responses to environmental changes in several and various environments and in different contexts (Bucci et al. 2011, 2014, 2015a, b; Crescenzo et al. 2017; Di Luccia et al. 2018; Petrella et al. 2018; Pietrangelo et al. 2018).

The dominant phyla observed in this study are often found in heterogeneous and complex soil systems (Janssen 2006; Schulz et al. 2013; Wolińska et al. 2017). Cyanobacteria phylum comprises photosynthetic organisms, which can easily survive on a bare minimum requirement of light, carbon dioxide, and water (Woese 1987; Castenholz 2001). They are phototrophic, fulfill their own nitrogen requirement through nitrogen-fixation, produce some bioactive compounds which promote crop growth, protect them from pathogens, improve the soil nutrient status (Singh et al. 2016), and have the ability to degrade various toxic compounds and to detoxify metal(loid)s (Cohen 2006; Singh et al. 2016). Proteobacteria include heterotrophic microorganisms, generally found in higher abundance in nutrient and carbon-rich substrates (Li et al. 2019). Their abundance was shown to either increase or decrease with Cr and Cd, respectively (Liu et al. 2019), showing that microorganisms respond differently depending on the contaminants. This could explain the difference in abundance between treatments: Proteobacteria could increase following a decrease in As and Pb availabilities induced by biochar and iron, while being restrained by the increase in As availability in compost-amended soils (Tables S1 and S2, Lebrun et al. 2019). Similar to Proteobacteria, Actinobacteria are dominant in carbon-rich environment (Xu et al. 2017), which could explain their higher abundance in biochar-soils. They have a role in the carbon cycle and the decomposition of organic materials (Wu et al. 2019). Acidobacteria are involved in several biogeochemical cycles and are able to decompose and use natural polymers (Huang et al. 2018). Finally, Archaea play important roles in C and N cycling (de Araujo et al. 2018). Their different abundances in the treatment revealed that the addition of compost as an amendment, alone or in combination, determined shifts in the archaeal community composition promoting the growth of members of Euryarchaeota and excluding Crenarchaeota.

Amendments, especially biochar and compost, tended to increase the diversity of the microbial community present in the soil. Such effects were consistent with previous studies (Abujabhah et al. 2018; Chen et al. 2018; Li et al. 2019). This higher microbial diversity demonstrates the recovery of the soil community following amendment application, which lead to a re-functionalization of the soil, except with iron grit. The positive effect of the biochar and compost amendments can be related to their effects on the soil properties, such as pH and organic matter content increase and metal(loid) immobilization. The soil microbial richness and diversity were shown to be positively correlated with soil pH and organic matter content (Huang et al. 2018; Li et al. 2018; Nguyen et al. 2018). In addition, Chen et al. (2018) observed a lower microbial diversity in a metal-polluted soil compared to an unpolluted soil. A previous study showed that the addition of biochar and compost to this soil increased soil pH and organic matter content and decreased As and Pb availability, reducing the stress pressure on microorganisms and plants (Lebrun et al. 2019, Tables S1 and S2), whereas iron grit did not induce a liming effect and had a reduced effect on organic matter content and metal(loid) availability. Moreover, biochar surface and porous structure can serve as habitat for the microorganisms (Li et al. 2018), and compost could have added new microorganism species, increasing the community diversity.

The molecular analysis of the microbial community revealed that the application of amendments modified the soil properties, inducing shifts in the community by promoting some phyla and restraining others. The reduced effect of the plant over soil properties in terms of shaping the microbial community was also observed by Xu et al. (2014) after biochar addition. This demonstrated that the microbial communities were more sensitive to the addition of carbon and organic matter as well as to the modifications of soil properties induced by the amendments and less to the compounds released by plant roots (this could be related to the small effects of amendments on organic acid contents released by the roots).

Conclusion

An assisted phytostabilization experiment was set up using a former mine technosol (Pontgibaud) mainly contaminated by As and Pb and amended with biochar, compost, and/or iron grit. The present paper aimed at evaluating the effect of amendments on plant root exudates as well as the effects of amendment application and plant growth, in both bulk and rhizosphere soils, on microbial community composition and activity. The results showed low but significant effects of biochar, compost, and iron grit amendments on root exudate release, separating the organic acids in two groups, malic and citric acids, and to a lesser extent fumaric acid, that seemed to have an important role in As and Pb tolerance and maybe in the control of Fe deficiency, while succinic and tartaric acids seemed to have a role in Fe tolerance in Salix viminalis plants. The study also revealed that the amendments led to an increase in the soil enzymatic activities (acid and alkaline phosphatases and FDA hydrolysis). Their application increased the capacity of the microbial community to use diverse carbon sources. These two parameters were especially increased when compost was applied. Next-generation sequencing revealed the presence of phyla commonly retrieved in soil ecosystems such as Proteobacteria, Acidobacteria, Actinobacteria, Verrucomicrobia, Bacteroidetes, Chloroflexi, Planctomycetes, and Gemmatimonadetes. Nevertheless, in general the different amendment treatments, and compost application specifically, induced a shift in microbial communities and a re-functionalization of the soil. This effect could be the result of the addition of microorganisms present in the compost itself. On the contrary, iron grit amendment decreased the soil microbial diversity, which was consistent with its effect on plant growth.

In conclusion, amendment had small effects on organic acid root exudates, whereas compost, applied alone or combined with biochar and iron grit, was the amendment showing the better soil microbial activity increases, confirming previous results (Lebrun et al. 2019) and thus its potential for assisted phytostabilization.