1 Introduction

Soil environments contain complex microbial communities, with diverse microorganisms capable to perform a huge array of biogeochemical processes. However, pollution may impact on the overall soil state and dynamics, affecting microbial diversity and functions, such as nutrient transformations and organic matter mineralisation (Roane and Pepper 1999; Gans et al. 2005; He et al. 2005; Joynt et al. 2006; Khan et al. 2010). Soil pollution is often a result of anthropogenic activities that include inputs of trace elements through use of fertilisers, organic manures and industrial and municipal wastes, irrigation and wet or dry deposits (He et al. 2005), clearly contributing to the presence of unusually high concentrations of heavy metals in the environment. The effect of pollutants on soil microbiota is related to the inherent toxicity of chemicals, as well as to their availability, which strongly depends on the physicochemical properties of the soil (Boivin et al. 2006; Lee et al. 2009; Mäkelä et al. 2012). Therefore, studies of microorganisms in relation to pollutants are usually included in soil ecotoxicity assessments (Eisentraeger et al. 2000; Hollender et al. 2003; Winding et al. 2005; Niklinska et al. 2006; Khan et al. 2010; Wang et al. 2010).

Microbial communities have been exposed to most toxic heavy metals and metalloids periodically since the origin of cellular life. Toxic metal resistance systems probably arose shortly after life started, in an early metal-polluted world (Silver 1998). The tolerance mechanisms are often plasmid borne, which facilitates dispersion from cell to cell (Ryan et al. 2005). Unlike organic contaminants, which can be degraded to harmless chemical species, heavy metals cannot be destroyed. Remediating the pollution they cause can therefore only be envisioned as their immobilisation in a non-bioavailable form or their re-speciation into less toxic forms (Valls and de Lorenzo 2002). Metal bioremediation based on microbial activity mainly focuses on biosorption processes, precipitation and enzymatic transformation of metals and metalloids.

Among potentially toxic elements, Cr, Cd and Pb have been considered within the first 20 pollutants at the USEPA priority list (http://www.atsdr.cdc.gov/cercla/07list.html). The widespread use of Cr in the metal industries and subsequent contamination problems has led to a strong interest on this metal. Although trace quantities are required for some metabolic activities, e.g. glucose and lipid metabolism, Cr is considered to be toxic (Lloyd 2003). Indeed, Cr(III) is considered 1,000 times less mutagenic than Cr(VI) (Lloyd 2003, and references within). Interestingly, both biotic and abiotic reduction of Cr(VI) to Cr(III) are possible (Kozuh et al. 2000). Cd and Pb have no known beneficial effects and are toxic at low concentrations for living cells (Trajanovska et al. 1997; Rivera-Utrilla et al. 2003). Their adsorption to microorganisms, as well as to organic or inorganic particles, has also been reported (Valls and de Lorenzo 2002; Rivera-Utrilla et al. 2003).

The effect of pollutants in soil microorganisms is an important issue in order to understand their toxic effects in the environment. In this sense, knowing the involvement of the microbiota of previously unpolluted soils in the presence of pollutants is useful to predict their consequences in potentially contamination events, as well as for developing adequate bioremediation strategies. The main objective of this research was to assess the impact of Cr(VI), Cd(II) and Pb(II) on the indigenous microbial community of an artificially polluted acidic Mediterranean soil. Soil respirometric assays allowed assessing the effect of pollutants on the soil microbial activity, showing the global effect on the carbon cycle. Molecular techniques were used for analysing the impact on the soil microbial community and for the detection of resistant microbiota. Isolation of metal-resistant bacteria and fungi was performed from these experimental incubation procedures, and finally several strains were tested to consider their potential usefulness in soil bioremediation processes.

This study is part of a wider research including the assessment of chlorophenolic pollutants in relation to this soil and also the effects of both heavy metals and chlorophenols to an agricultural calcareous soil (Cáliz et al. 2011a, b; Marti et al. 2011). This global focus also allows analysing these results in the frame of the influence of a wider array of pollutants on different kinds of soil environments, as well as considering the potential of the microbiota to face up to co-contamination events.

2 Materials and methods

2.1 Soil samplings and microcosm incubations

Experiments were performed with a pine woodland soil without any prior history of exposure to the heavy metals assayed. The soil was collected in November 2006 from the superficial layer (A horizon) of a site located in the Mediterranean area of Vilassar de Dalt, Barcelona (UTM, 31T 444376E 4596459N). After sampling, the soil was sieved (<2 mm) and stored by air-drying conservation method until the forthcoming experiments. This soil was classified as haplic arenosol (FAO-UNESCO 1998). It was an acidic (pH 5.8, measured in water soil extract, proportion 1:2.5) and coarse-textured (87 % sand) soil of granitic origin, with low concentrations of organic carbon (0.71 %) and carbonates (0.1 %). This soil type has adequate C/N ratio (10.14) to perform respirometric assays, which corresponds to the properties of standard soil proposed by the OECD for ecotoxicity testing in terrestrial environments.

After a pre-incubation throughout 5 days at the field soil humidity (5.1 % dry weight), 50 g of soil sample was treated separately with 0.1, 1, 10, 100, 1,000 and 5,000 mg kg−1 Cr(VI), Cd(II) or Pb(II), and without any pollutant (control). The contamination of soil microcosms was performed with aqueous solutions of K2CrO4, CdSO4 or PbCl2. In order to perform respirometric analysis, soil samples were placed in closed reactors (500 mL) with an equivalent water content to the 60 % of the soil water holding capacity and incubated in the dark at 25 °C for 28 days. Soil pH was measured after the addition of the metals and monitored during incubations to confirm that it was not subjected to a wide range of variation. All experiments were performed in triplicate. Control and polluted soil microcosms destined to molecular analysis (polymerase chain reaction denaturing gradient gel electrophoresis, PCR–DGGE) were incubated likewise in triplicate, but harvested and sampled destructively after 7, 14 and 28 days. Non-incubated soil samples (without addition of any pollutant) were used as initial controls. After each harvest, soil samples were stored at −80 °C until extraction of nucleic acids.

2.2 Analyses of metals in the soil solution

Soluble fractions of Cr, Cd and Pb were determined from aqueous extracts (DIN 1984) of triplicate soil microcosms after the addition of metals and at different periods of incubation. The concentrations were measured by inductively coupled plasma using optical detection (ICP-OES) for Cr and mass spectrophotometry (ICP-MS) for Cd and Pb.

2.3 Respirometric analysis of microbial activity

The aerobic heterotrophic microbial activity was analysed by respirometry. Microcosms processed with this purpose were incubated in manometric respirometers as described above, to allow the determination of oxygen consumption (Oxitop®, WTW). The cumulative oxygen consumption (cumulative respiration, CR) was periodically registered throughout the incubation period. To determine the statistical significance of the differences between treated samples and controls, an ANOVA followed by Duncan’s post hoc test (p < 0.05) (SPSS 16) was performed.

Data derived from respirometric assays were considered to select treatments for a thorough analysis of the changes in the microbial community composition by PCR–DGGE. All the treatments with concentrations that resulted in a significant inhibitory effect on respirometry tests and those with the highest concentration of pollutants that showed low or no inhibitory effect were analysed by molecular methods.

2.4 DNA extraction

DNA was extracted from 0.5 g of soil samples using the FastDNA® SPIN Kit for Soil (Bio101, Carlsbad, CA, USA), as described in Cáliz et al. (2011b). Previously, the optimal amount of soil sample was experimentally determined from the relationship between the weight of processed soil samples and the concentration of extracted DNA (data not shown) to accomplish a high nucleic acid yield while avoiding saturation of the extraction procedure. The integrity of DNA was checked by agarose gel electrophoresis and the amount and purity of nucleic acid were determined using a NanoDrop ND-1000 UV–vis spectrophotometer (NanoDrop, Technologies, Inc., Wilmington, DE, USA). DNA extracts were stored at −80 °C until processed.

2.5 PCR amplification

All chemicals and Taq polymerase used for the PCR amplification of DNA extracts were provided by Applied Biosystems (Foster City, CA, USA). Amplifications were done in a 9700 GeneAmp thermal cycler (Applied Biosystems, Perkin-Elmer, CA, USA). Partial 16S rRNA gene fragments covering the variable V3 to V5 regions were obtained by PCR amplification using the Bacteria universal primers 357F (5′-CCTACGGGAGGCAGCAG-3′) and 907R (5′-CCGTCAATTCCTTTGAGTTT-3′) (Muyzer et al. 1993, 1995). Partial sequences of the internal transcriber gene spacer ITS from fungi were obtained by PCR amplification using primers ITS1-F (5′-CTTGGTCATTTAGAGGAAGTAA-3′) (Gardes and Bruns 1993) and ITS2 (5′-GCTGCGTTCTTCATCGATGC-3′) (White et al. 1990). A 44-bp GC-rich clamp sequence was added to the 5′ ends of primers 357F and ITS1-F to allow further separation of PCR products by DGGE (Muyzer and Smalla 1998). Reaction mixtures and PCR amplification conditions, specific for both primer pairs, were applied as previously described (Cáliz et al. 2011a). PCR amplification products were checked by agarose gel electrophoresis, and those with the correct size and similar yields (~100 ng μL−1) were used for DGGE analysis.

2.6 DGGE analysis

Denaturing gradient gel electrophoresis was performed using an INGENY PhorU system (Ingeny International BV, Goes, The Netherlands). PCR products were separated in 6 % (w/v) polyacrylamide gels prepared with a 35–75 % and 20–80 % urea–formamide vertical gradient for bacteria and fungi, respectively, according to the instructions of the manufacturer [100 % denaturant agent contains 7 M urea and 40 % deionised formamide (McCaig et al. 2001)]. Electrophoresis was performed for 12 h with 1× TAE buffer at 60 °C, at a constant voltage of 160 V. Gels were stained for 30 min with 1× SybrGold (Invitrogen Molecular Probes, Eugene, OR, USA) and visualised under UV excitation. Bands of interest were chosen after detailed analysis of fingerprint images and excised from gels for further reamplification and sequencing.

Digital images were analysed by using the GELCompar II v.6.1 software package (Applied Maths BVBA, Sint-Martens-Latem, Belgium). Comparison between samples loaded on different DGGE gels was completed using normalised values derived from control samples common in all of them. Calculation of the pairwise similarities of densitometric profiles was based on Pearson's correlation coefficients with an optimisation of 2 %. Cluster analysis based on this similarity matrix was done by unweighted pair-group method with arithmetic averages (UPGMA).

2.7 Isolation of metal-resistant bacteria

Soil samples contaminated with 100 or 1,000 mg kg−1 Cr, or 5,000 mg kg−1 Cd or Pb, respectively, were chosen for the isolation of resistant bacteria after 28 days of incubation under the same conditions described above for respirometric and molecular analysis. Samples consisting of 1 g of fresh soil from each microcosm were suspended in 100 mL of sterile Ringer’s solution (Scharlab, Barcelona, Spain) in 250-mL conical flasks and dispersed by stirring for 20 min at 200 rpm. The resultant suspensions were serially diluted and plated in triplicate on solid 10-fold diluted Luria–Bertani medium (Sigma-Aldrich GmbH, Steinheim, Germany) supplemented with K2CrO4, CdSO4 or PbCl2 to final concentrations of 50, 100, 500 and 1,000 mg L−1. Plates were incubated up to a week at 25 °C in the dark. Several colonies were selected from each soil microcosm according to their different morphology. Most of them belonged to bacteria, but a few fungi could also be grown in culture plates as well. Pure cultures were obtained after repetitive inoculation in fresh medium and saved for further identification and characterisation.

2.8 Sequencing DNA from DGGE bands and isolates

DNA from excised DGGE bands was rehydrated in 50 μL of sterile ddH2O, eluted after incubation at 65 °C for 30 min and reamplified using 2 μL of the eluate with the PCR conditions and the corresponding primers (without GC clamp) described above for bacteria and fungi. PCR products were stored at −80 °C until sequencing.

Nucleic acids were also extracted from colonies picked up directly from agar plates using Wizard™ Genomic DNA Isolation Kit (Promega, Madison, WI, USA). The isolation of DNA from bacteria and fungi was performed following the manufacturer’s indications. The amount of DNA was quantified using a NanoDrop ND-1000 UV–vis spectrophotometer (NanoDrop Technologies) and stored at −80 °C until processed. Partial 16S rRNA gene fragments were amplified by PCR using the Bacteria universal primers 27F (5′-AGAGTTTGATCCTGGCTCAG-3′) and 1492R (5′-ACGGTTACCTTGTTACGACTT-3′) (Lane 1991). Reaction mixtures and PCR amplification conditions were applied as previously described (Cáliz et al. 2011a). Partial sequences of the internal transcriber gene spacer ITS were amplified as described above for fungi. PCR amplification products were checked by agarose gel electrophoresis. Products of the correct size were purified with the QIAquick PCR purification kit (Qiagen, Hilden, Germany) and stored at −80 °C until sequencing.

Finally, the PCR-amplified DNA products, obtained from the DGGE bands and the isolates, were sequenced by Macrogen Inc. (Seoul, South Korea). Primers 357F and 907R were used in order to sequence the PCR products from the DGGE bands of bacteria. Primers 27F and 1492R were used to sequence the 16S rRNA gene of the isolates, in combination with the universal primer Eb787F (Baker et al. 2003) as reverse and forward, respectively. Both fungi-derived DGGE bands and isolates were sequenced using primers ITS1-F and ITS2.

2.9 Phylogenetic identification of bacteria and fungi

All bacterial- and fungal-retrieved sequences were compared for the closest relatives in NCBI database (http://www.ncbi.nlm.nih.gov/blast/) using the BLASTN algorithm tool (Altschul et al. 1990). The presence of chimera was checked using the Bellerophon tool (Huber et al. 2004). Bacterial sequences were properly aligned using the online automated aligner SINA (SILVA Incremental Aligner) available at SILVA website (http://www.arb-silva.de/; Pruesse et al. 2007). Alignments were imported into the ARB software package (http://www.arb-home.de/; Ludwig et al. 2004) and loaded with the SILVA 16SrRNA–ARB-compatible database (SSURef-102, February 2010). The phylogenetic tree was constructed by maximum likelihood (RAxML) analyses using reference sequences and sequences of the isolates longer than 1,200 bp. Subsequently, the shorter sequences obtained from DGGE bands were added by applying the parsimony tool implemented in ARB, thereby maintaining the overall tree topology. Closest relative bacteria were identified based on the phylogenetic tree affiliations, and sequence similarities were calculated using the ARB distance matrix tool.

Sequences of the bacterial 16S rRNA and fungal ITS regions derived either from DGGE bands or from isolates were deposited in GenBank under accession numbers HE577948 to HE578011.

2.10 Characterisation of resistant bacteria

Isolates were incubated at 25 °C during 1 week in the medium described by Francisco et al. (2010) with slight modifications (Cáliz et al. 2011b) in order to test for their resistance to different heavy metals and a metalloid: Cr(VI), Cd(II), Pb(II) and As(V). The pH was adjusted to 7.5 using NaOH. To prepare solid plates, agar was added to a final concentration of 15 g L−1. After autoclaving, each pollutant was added to the medium by using filtered stock solutions of K2CrO4, CdSO4, PbCl2 or KH2AsO4 to obtain plates with different concentrations: 0.5, 1, 2 or 4 mM Cr(VI); 0.5 or 2 mM Cd(II); 0.2 or 1 mM Pb(II); and 3 or 5 mM As(V). Resistance of the isolates was checked by qualitative observation of the colony development in polluted plates in comparison to the controls without pollutants.

The most resistant isolates were checked for their capacity to reduce the concentration from 0.5 mM Cr(VI) in 100-mL liquid cultures with the medium described above. Experiments were conducted in Erlenmeyer flasks (500 mL) under aerobic conditions on an orbital shaker (150 rpm) at 25 °C for 48 h. Controls without inocula were set up likewise. Growth was estimated at the end of the incubations by measuring optical density (OD) at 600 nm. Cr(VI) concentration was analysed from culture supernatant after the centrifugation of cells using diphenylcarbazide method (American Public Health Association 1998).

3 Results

3.1 Respirometric estimation of microbial activity inhibition

The overall state of the soil microbiota was assessed by cumulative respiration analyses. No inhibitory effects were detected at the microcosms incubated with up to 1 mg Cr kg−1 (Fig. 1), but oxygen consumption decreased at treatments with higher amounts of added Cr(VI). The highest amounts of Cr added as Cr(VI) to soil microcosms (1,000 and 5,000 mg kg−1) resulted in strong inhibitory effects of the respirometric activity values, more than 80 % inhibition in respect to the control soil values. Considering treatments with Cd and Pb, inhibitory effects were missing even at higher concentrations than Cr, producing a significant reduction of the oxygen consumption only above 1,000 mg kg−1 of both pollutants. At the same added amount of each pollutant, a slightly lower inhibition of the respirometric activity was obtained with Pb than with Cd.

Fig. 1
figure 1

Cumulative oxygen consumption (expressed as a percentage in relation to the respective controls) obtained from the CR assays of the polluted microcosms after 28 days of incubation. Variation coefficient is indicated around the mean values. Soluble metal concentrations determined for each treatment after the addition of pollutants are also indicated in the x-axis; standard deviations show no relevant differences among these values. *Significant CR inhibition compared to controls (p < 0.05)

Full size image

Analyses of soluble metal concentrations in soil microcosms during the incubations indicated that they were considerably lower than the initially spiked amounts for all tested metals (Fig. 1; see also additional data of soluble metal concentrations at different periods of incubation in Electronic Supplementary Material, Table 1). In most of the treatments, bioavailable Cr decreased readily after its addition to soil microcosms up to values below 20 % in comparison to the amounts actually amended. These values decreased up to 78 % and 63 % in the treatments of 1,000 and 5,000 mg kg−1 Cr, respectively. Moreover, almost all the added Cd and Pb was unavailable to the microbiota, recovering up to 32 % and 4.2 %, respectively, in the treatments from 1,000 mg Cd kg−1 and 5,000 mg Pb kg−1. The soluble metal concentrations tended to decrease throughout the incubation in all treatments, especially at the highest concentrations of Cr. According to these data, the lowest soluble concentrations of metals producing significant decreases of the respirometric values corresponded to the treatments of 10 mg Cr kg−1 and 1,000 mg Pb kg−1, while no inhibitory effects were detected at similar soluble concentrations of Cd. Moreover, considering the treatments of 1,000 mg Cr kg−1 and 5,000 mg Cd kg−1, a higher decrease in respirometric activity was detected at soluble amounts of Cr even lower than Cd.

3.2 DGGE fingerprints of soil samples

Changes in the microbial community composition of soil microcosms were analysed from PCR–DGGE fingerprints of 16S rRNA (Bacteria) and ITS (Fungi) gene fragments to assess the effects of the pollutants and to determine the concentrations that produced different degrees of alteration.

Taking into account bacterial fingerprints, soil samples clustered into several groups according to similarities of the densitometric DGGE profiles (Fig. 2). The fingerprints of the non-incubated and incubated controls were clearly separated in different groups (group I and group V, respectively) with more than 70 % similarity within each one, thus indicating that several changes appeared on soil bacterial community as an effect of the incubation itself. The fingerprints of treatments with low concentrations of pollutants (10 mg kg−1 Cr and 100 mg kg−1 Cd and Pb) also clustered together with the incubated control samples in group V. Soil samples amended with the highest concentrations of Cr(VI), which revealed a strong inhibition in respirometric activity, were included within group I, closely related to the non-incubated control samples.

Fig. 2
figure 2

DGGE fingerprints of partial 16S rRNA gene fragments obtained with Bacteria universal primers, as they have been grouped in a dendogram by GelCompar clustering analysis. Arrowheads indicate the bands that were excised and sequenced. Bands codes indicate whether the bands were recovered from Cr, Cd or Pb treatments. The dendrogram is based on calculated pair-wise similarities of densitometric profiles (Pearson's correlation coefficients with an optimization of 2 %). Grouping has been made by using a UPGMA method. Codes for soil samples indicate consecutively the pollutant used in the treatment, its concentration (milligrams per kilogram) and the period of incubation (in days)

Full size image

Three additional groups (II, III and IV) could be distinguished, with changes involving the disappearance of certain bands present in the control samples, probably related to pollutant-sensitive bacteria, and the appearance of a few new bands that could be related to the development, as dominant populations, of several presumably resistant bacteria. Most of these changes were detected after only 7 days of incubation and persisted throughout all the incubation period. Groups II and III, represented by samples amended with 5,000 mg kg−1 Pb and Cd, respectively, showed the most different DGGE fingerprints.

The analysis of the molecular fingerprints of fungi led to similar results: community composition changed along with the increased concentrations of the pollutants (Fig. 3). Soil samples of non-incubated and incubated controls also clustered separately, showing the effects of the incubation. Group III included the treatments with low concentrations of all pollutants, as well as the incubated controls, with more than 80 % similarity, indicating that no effects were produced at these concentrations. On the opposite, highly polluted samples showed significant differences with the controls, according to the clear appearance and disappearance of several bands. Group II, which included the Pb-treated samples, showed the highest similarities with the group containing the incubated control, while samples amended with the highest concentrations of Cr and Cd were included in group I and exhibited the most different fingerprints. No PCR products were obtained at 5,000 mg Cr kg−1.

Fig. 3
figure 3

DGGE fingerprints of the internal transcriber gene spacer fragments obtained with ITS fungal primers, as they have been grouped in a dendrogram by GelCompar clustering analysis. Arrowheads indicate the bands that were excised and sequenced. The dendrogram is based on calculated pair-wise similarities of densitometric profiles (Pearson's correlation coefficients with an optimization of 2 %). Grouping has been made by using a UPGMA method. Codes for soil samples indicate consecutively the pollutant used in the treatment, its concentration (milligrams per kilogram) and the period of incubation (in days)

Full size image

3.3 Phylogenetic identification of sensitive and resistant populations of microorganisms

Up to 82 relevant bands were excised from bacteria- and fungi-derived DGGE gels and treated for further PCR reamplification and sequencing (see Figs. 2 and 3, respectively). From them, 71 bands produced useful sequences, without ambiguous positions, and were used for identification purposes (Table 1 and Fig. 4). Most sequences yielded very high similarity values with previously published sequences. Some of the DGGE bands resulted in identical sequences despite appearing in different positions of the gel (i.e. bands Pb08/Pb15, Pb11/Pb12/Pb13/Pb14/Pb16 and Cd01/Cd09_1/Cd14_1, in Fig. 2) probably due to variable melting behaviours or the presence of multiple ribosomal gene copies in a single organism (Prat et al. 2009). On the other hand, some bands appearing in different samples at the same gel position resulted in different sequences (e.g. band Cd01, Cd15.1 and Cr05, in Fig. 2).

Table 1 Closest matches of 16S rRNA and ITS gene sequences obtained from bacterial and fungal DGGE analyses, respectively
Full size table
Fig. 4
figure 4

Maximum likelihood phylogenetic tree calculated for 16S rRNA gene sequences of Actinobacteria obtained in this study from both DGGE bands and isolates (in boldface) and those of their closest relatives. Bootstrap values >60 % are indicated at branch nodes. Representative sequences retrieved from DGGE bands are indicated as “Band”, and those obtained from the isolated resistant bacteria are indicated as “Isolate”. Type strain sequences included in SILVA database are marked with an asterisk (*). The scale bar indicates 5 % estimated sequence divergence

Full size image

Several bands, present in the fingerprints of incubated controls, were absent in metal-treated samples that showed inhibition of the respirometric activity, suggesting they probably corresponded to pollutant sensitive bacteria. Sequences retrieved from these bands belonged to Sphingomonas jaspsi (Cd01, Cd09_1 and Cd14_1), Thermomonas sp. (Cd18_1), Ramlibacter sp. (Cd03, Cd17_1 and Pb03) and Tumebacillus permanentifrigori (Pb05). These last two phylotypes could only be considered Cd- and Pb-sensitive bacteria because their related bands were still detected at 100 mg Cr kg−1 (bands Cr15 and Cr25_1, respectively). In all cases, homologies with previously published sequences were higher than 97.2 %.

In contrast, other bands were exclusively detected in the fingerprints of polluted soil samples, probably belonging to resistant bacteria. In Cr(VI)-treated soils, retrieved sequences from these bands were related to Geodermatophilus obscurus (Cr30_1), Phenylobacterium mobile (Cr23_1), Rhizobium leguminosarum (Cr05), Niastella koreensis (Cr02_1 and Cr03_1), Cupriavidus campinensis (Cr02), Massilia aerilata (Cr04) and Paenibacillus contaminans (Cr08). Most of these phylotypes were only detected at 100 mg kg−1, and homologies with previously published sequences were higher than 98.5 %. Retrieved sequences from bands that were exclusively detected at the highest concentration of Cd were closely related (>99.4 %) to Leifsonia kribbensis (Cd24_1), Actinoallomurus coprocola (Cd27_1) and Methylobacterium radiotolerans (Cd23_1). Other bands, detected at a lower concentration of Cd, showed high homologies (most of them above 97.5 %) to several members of the Bacteroidetes (Cd07_1, Cd04_1, Cd05_1, Cd06_1, Cd08_1 and Cd10_1), as well as to Burkholderia graminis (Cd04), Cupriavidus metallidurans (Cd11_1) and Frateuria aurantia (Cd02 and Cd05). Concerning Pb, up to 11 identified bands (from Pb08 to Pb18), most of them exclusively detected at 5000 mg Pb kg−1, showed high homologies (>98.7 %) with species of Streptomyces that are very closely phylogenetically related (see Fig. 4). Other bands, common in the fingerprints of different metal-treated samples, could be related to presumable resistant bacteria to several of the assayed pollutants. Retrieved sequences from these bands belonged to Burkholderia caledonica (Cd15_1 and Pb07), Bacillus acidiceler (Cr03 and Pb01) and an unidentified member of the Bacillales, strain Gsoil 1105 (Cr20, Cd06 and Pb06).

Considering the fungal sensitive populations, sequences affiliated to Cryptococcus terreus (F08) and an unclassified fungus (F01) were retrieved from bands disappearing at high concentrations of Cr and Cd. In contrast, these fungi were probably resistant to Pb since related bands (F09 and F02, respectively) were present at high concentrations of this pollutant. Other bands related to a sensitive Zygomycete sp. (F03 and F04) were absent at high concentrations of Cr and Pb, but this species was probably stimulated by the addition of Cd, as indicated by the intense band detected at 1,000 mg Cd kg−1. In all cases, homologies with previously published sequences were higher than 99.1 %. Several members of the Ascomycota were exclusively detected at the highest concentrations of the pollutants, probably related to resistant fungi. Sequences with high homology (>99.1 %) with different strains of Trichoderma atroviride, DAOM 233966 (F18 and F20) and NG 13 (F19), were found at these treatments with high amounts of Cr, and the former was also detected in Cd- and Pb-treated samples. Two additional fungi, related to Engyodontium album (F21) and Penicillium corylophilum (F22), were exclusively detected at the highest concentration of Cr.

3.4 Isolation and identification of resistant bacteria and fungi

Several isolates were obtained from soil microcosms supplied with 100 and 1,000 mg kg−1 of Cr, and 5,000 mg kg−1 of Cd or Pb, which DGGE fingerprints provided evidences of presumable resistant microbial populations. They were selected according to the observable differences in colony morphology. 16S rRNA or ITS gene sequence analysis indicated that most of the isolates showed more than 98.1 % similarity to sequences of cultured bacteria and fungi (Table 2). Bacterial isolates distributed mainly in two divisions, Actinobacteria and Firmicutes, but several strains obtained only from the Cd-polluted microcosm were also related to M. radiotolerans. Three species of fungi, all of them clustered into the group of Ascomycota, were isolated mainly from Pb treatments.

Table 2 Phylogenetic affiliation of bacterial and fungal isolates based on 16S rRNA and ITS gene analysis and characterisation of their resistance to the tested metals and metalloid, as well as their capacity to reduce the concentration of Cr(VI)
Full size table

3.5 Characterisation of metal resistance and reduction of Cr (VI) concentration by the isolates

Bacterial isolates were tested in minimal media supplied with different amounts of Cr, Cd, Pb and As to assess their resistance to these toxic elements (see Table 2). A single representative for each group of isolates with identical 16S rRNA gene sequences was chosen. Most of the isolates grew in the presence of 1–2 mM Cr(VI), but only Streptomyces strains (IMCr02 and IMCr03) were able to withstand 4 mM Cr(VI). Resistance to Pb was usually achieved by most of the isolates, and a few strains could also grow in the presence of Cd or As. The strain IMCd01 (M. radiotolerans) showed the highest resistance to Cd, although it was unable to tolerate any of the assayed concentrations for Cr and As. However, a few strains, all of them related to species of Bacillus (IMCr01, IMCr07 and IMCd03), could withstand most of the tested metals and metalloid.

The most resistant strains to Cr, belonging to species of Streptomyces (IMCr02, IMCr03 and IMCr09) and Bacillus (IMCr01, IMCr07 and IMCd03), were also tested for their capacity to reduce the concentration of added Cr(VI) in liquid cultures with minimal media and an organic carbon source. However, the strain IMCr07 (Bacillus weihenstephanensis) could not be properly assessed since its culture in liquid medium could not be achieved. All strains were able to reduce the concentration of Cr(VI), in a 48-h period, between 54 % and 70 % from an initial concentration of 0.5 mM Cr(VI) (see Table 2). At the end of the incubations, growth led to similar OD values for Bacillus strains (around 0.55). OD values for Streptomyces strains were not determined since their cells were grouped in aggregates, but their growth was clearly observed in the liquid cultures.

4 Discussion

4.1 Evaluation of the effects of the metals added to the soil microcosms by both respirometric analyses and molecular methods

Since soil microorganisms constitute the main part of the biomass and regulate all nutrient cycles, they are good ecological receptors for the assessment of metal toxicity (Lazzaro et al. 2006). Microorganisms assessed by various physiological, biochemical or molecular techniques have been recommended as biological indicators of metal contamination (He et al. 2005). In the forest soil surveyed in this study, an overview of the influence of the different heavy metals (Cr, Cd and Pb) on the soil system has been revealed by assessing activity and composition of the microbial community. Fungal populations have been considered, besides bacteria, since the soil assessed is clearly acidic and its climatic origin includes a long dry period, factors favouring the development of these microorganisms (Reith et al. 2002; Hobbie and Gough 2004; Kimura and Asakawa 2006). Despite that, DGGE fingerprints of the control samples evidenced a lower diversity of the soil fungal community composition in comparison to bacteria, as previously reported by Fierer et al. (2005) and Wang et al. (2010).

Metal-treated soil microcosms showed inhibition in the respirometric activity and variations in the community composition of both bacterial and fungal groups. According to the clustering analysis, significant changes in the bacterial and fungal populations were detected above the same added amounts for each metal (from 100 mg kg−1 Cr or 1,000 mg kg−1 Cd and Pb), thus indicating that sensitive microorganisms of both groups had similar pollutant tolerances. The higher the inhibition of the activity, the lower the similarities in the fingerprints between polluted soil samples and incubated controls. Thus, detrimental effects on the microbial community increased with the pollutant concentration, linking the level of toxicity with the added amount of each metal (Speir et al. 1995; Khan et al. 2010).

The most noticeable and drastic effects on the microbial community were detected at the highest treatments of Cr, which led to a strong inhibition of the respirometric activity (more than 80 % in comparison to unpolluted samples) and even of the development of microbial populations. DGGE fingerprints of these samples clustered with non-incubated controls in group I (see Fig. 2), indicating the lack of significant changes in bacterial community composition throughout the incubation with these high concentrations of Cr. Similar results were obtained in other experiments with the same soil spiked with other kinds of pollutants (Cáliz et al. 2011a). A strong impairment of the fungal populations in soil samples amended with 5,000 mg Cr kg−1 was also suggested by the negative PCR amplification of fungal ITS gene fragments, which probably indicated even a damage of the DNA of the dead fungi.

The comparison between both methods shows that a significant decrease in the respirometric activity values corresponded to changes in the DGGE fingerprints, in respect to the incubated controls, except for the treatment of 10 mg Cr kg−1. Clear variations in the microbial community were detected, however, only at treatments with the highest metal concentrations, in agreement to the most drastic decreases in the activity as well. In contrast, other experiments performed with the same soil (Cáliz et al. 2011a) showed that changes in the microbial composition of several chlorophenol-treated samples were detected even in treatments producing no significant alterations in the cumulative respiration. Respirometric assays could be more suitable and specific to detect ecotoxicological effects of heavy metals than organic pollutants due to the overestimation of the activity status caused by chlorophenol-degrading microorganisms, as previously reported in similar experiments with a calcareous soil (Cáliz et al. 2011b).

4.2 The influence of bioavailability on the toxic effects of heavy metals

The mobility and availability of trace elements in soils are controlled by many chemical and biochemical processes, such as precipitation–dissolution, adsorption–desorption, complexation–dissociation and oxidation–reduction (He et al. 2005), which are affected by pH and biological activity. In general, free cationic metal species are more bioavailable at acidic pH values (Hughes and Poole 1991; Smith 1994; Antoniadis et al. 2008) since they may not be sorbed by binding sites of the soil matrix, clay materials or organic matter, and therefore the interactions between metals and potential metal-complexing ligands are limited. Moreover, under basic conditions, cations tend to form unavailable hydroxy- or carbonate–metal complexes (Bataillard et al. 2003). According to these statements, Cd and Pb were sparingly available in our acidic soil microcosms when low amounts of these cations were spiked and high amounts were to be added to significantly increase their presence in the soluble fraction. Despite that, the bioavailability of Pb was largely the lowest in comparison to the other tested pollutants, confirming the difficulty of this metal to remain in the soluble fraction (Echeverría et al. 1998; Covelo et al. 2007). In contrast, the bioavailability of Cr is faintly influenced by the interactions with the soil matrix since it is an anionic species on its Cr(VI) ionic form, but its mobility is related to the oxidation state. Cr was added to soil microcosms in its oxidised form Cr(VI), which is the most toxic and water soluble (James and Bartlett 1983). However, at acidic pH values, the chemical reduction reactions favoured the predominance of Cr(III), whose features limit its bioavailability and mobility (Richard and Bourg 1991), thus explaining the lower values of soluble Cr detected in comparison to the total amounts spiked to soil microcosms. These results agree with the capacity of the soil physicochemical properties to reduce the potentially higher negative impact of Cr on the microbiota (Shi et al. 2002; Gzik et al. 2003).

In turn, environmental factors such as soil pH or nutrient availability can be altered by the addition of metals. pH may have significant effects on soil microbial populations, sometimes even higher than those related to heavy metal toxicity, as pointed out by several studies (Diaz-Ravina and Baath 1996; Fernández-Calviño et al. 2010, 2011). However, it was experimentally confirmed (data not shown) that the addition of metals in this soil had few influence in pH (at the worst cases, corresponding to the incubations with the highest amounts of Cd and Cr, produced a decrease or increase, respectively, lower than one pH unit), and thus in the nutrient availability.

Respirometric and molecular analyses showed that Cr is more toxic than Cd and Pb for the microorganisms of this acidic soil, according to the total added amounts of these metals. However, since large metal fractions can be present in biologically unavailable forms, total concentrations of heavy metals are poor indicators of their presumable toxicity effects on soil environment (Lazzaro et al. 2006). Therefore, the soluble fraction of the added metals should be considered as an indicator of the potentially toxic forms to target organisms (Blaser et al. 2000; Turpeinen et al. 2004; Lazzaro et al. 2006), enabling to better understand the microbial exposure to metals. According to this statement, the results suggest that Cr and Pb have theoretically a similar potential toxicity, higher than Cd, although the comparison among the different treatments indicated that, once in the soil environment, Pb showed a lower negative effect at the same added amounts. Therefore, the impact of these metals on the microbiota of this acidic soil can greatly be explained by their differential bioavailability.

4.3 Resistance of bacteria and fungi to high concentrations of heavy metals

Soil microcosms exposed to Cr, Cd or Pb showed a fast response of the microbial community to the new adverse situations by selection and development of supposedly resistant populations. In addition, most of the isolates obtained from these artificially polluted soil microcosms were also confirmed as resistant to several of the pollutants assayed. Thus, these results prove the survival of several metal-resistant microorganisms in the surveyed soil, even without any prior exposure to these toxic compounds, in agreement with similar studies performed with unpolluted soils (Diaz-Ravina and Baath 1996; Viti et al. 2006; Lazzaro et al. 2008).

While most of the fungi (a few species, compared to bacteria) detected by the molecular approach at highly polluted soil microcosms were restricted to Ascomycota lineage, as already expected (Vadkertiová and Sláviková 2006; Zafar et al. 2007), presumable resistant bacteria were found to be widespread among several phylogenetic groups. These differences may be explained by the low complexity of the natural fungal composition of the soil itself. The greater diversity of presumable resistant bacterial community inhabiting this soil may confer a better adaptability and flexibility (Boles and Singh 2008; Bodelier 2011) in comparison to the fungal assemblage, and thus, a higher potential to survive in more diverse events of contamination or perturbation. In the case of Cr-treated soil microcosms, however, DGGE showed that some fungi developed in highly polluted treatments, while no bacteria were detected, suggesting a higher efficiency of members of the fungal group to cope with high concentrations of Cr (Khan and Scullion 2000; Khan et al. 2010; Wang et al. 2010). In contrast, several isolated bacteria from these polluted microcosms showed high resistance to Cr in the tests performed. Therefore, resistant bacteria were probably present at too low abundances to be detected by the molecular methods, but longer incubations could evidence their later emergence. Similar studies also suggested a slowdown in the development of resistant microorganisms, especially at high concentrations of pollutants, which only became detectable after long incubation periods (Cáliz et al. 2011a, b).

Isolation procedures allowed recovering mainly gram-positive bacteria from metal-treated soil microcosms, in agreement with several previous studies (Viti and Giovannetti 2001; Branco et al. 2005; Viti and Giovannetti 2005). Some of the isolates showed identical affiliation, or were closely related, to most representative phylotypes identified from DGGE fingerprints. Most of the resistant fungi detected by the molecular analysis were obtained in pure cultures, thus showing a good correlation between culture-dependent and culture-independent approaches. Some bacteria and fungi could be detected either in soil samples amended with any of the pollutants assayed. Moreover, most of the isolates tested to assess their resistance to the heavy metals (Cr, Cd and Pb) and a metalloid (As) showed the capacity to grow in the presence of several of these toxic elements, in agreement with the coexistence in plasmids of the genes usually related to multiple resistances (Bruins et al. 2000).

The most resistant isolates to Cr, which were related to several species of Streptomyces and Bacillus, were found to reduce more than 50 % from an initial concentration of 0.5 mM Cr(VI) in a 48-h period, which indicated their potential for the soil restoration. These species could be responsible for the decrease of the soluble fraction of Cr throughout the incubation of soil microcosms with the highest amendments of this metal. Different species of Bacillus are well known as Cr resistant (Camargo et al. 2005), able to transform Cr(VI) to Cr(III) through enzymatic reduction (Campos et al. 1995; Camargo et al. 2003; Abou-Shanab et al. 2007; Verma et al. 2009), but this ability is less known in the genus Streptomyces (Poopal and Laxman 2009; Polti et al. 2010). A wide diversity of phylotypes related to Streptomyces was also detected by the molecular analysis at highly polluted soil microcosms with Pb, pointing out that members of this group successfully coped with different kinds of toxic metals as well.

4.4 Is resistant microbiota specific to soil environment and to the type of pollutants?

This study, as part of a wider research including the assessment of several pollutants (heavy metals and chlorophenols) in relation to two different Mediterranean soils, has shown that resistant microbiota is mostly specific for each environment since no common species have been detected within both soils. Resistant fungi were exclusively detected in this forest acidic soil (Cáliz et al. 2011a), as well as several bacteria such as some species of the genera Burkholderia or Streptomyces, which were the most resistant to high concentrations of pollutants in this soil, in agreement with other results obtained with similar soils (Lazzaro et al. 2008). In contrast, some other well-known resistant bacteria belonging to Gammaproteobacteria, such as Pseudomonas spp. or Stenotrophomonas spp., could only be detected in the calcareous soil (Cáliz et al. 2007, 2008, 2011b). Therefore, emerging resistant microbiota may be considered as the best suited to each environment, probably in agreement with the soil characteristics and the growing requirements that favoured the development of such microorganisms. These findings indicate that the microbial features and the soil characteristics must be considered before designing bioremediation strategies that use bioaugmentation in order to achieve a good performance of the process.

Several bacterial strains previously isolated from this acidic soil (Cáliz et al. 2011a), which were found to be resistant to chlorophenols, show identical phylogenetic affiliation with metal-resistant isolates retrieved in the present work. These strains, which seem to be able to cope with different kinds of pollutants, were related to Bacillus thuringiensis, B. weihenstephanensis and B. acidiceler. Moreover, two species of Burkholderia (B. graminis and B. caledonica) detected in the presence of Cd were also found in most of the chlorophenol-treated soil microcosms. The same fungi (Engyodontium album, Penicillium corylophilum and the isolated strains related to Trichoderma atroviride) were also found after contamination with heavy metals and chlorophenols. Therefore, these results pointed out the presence of multiresistant microorganisms in this Mediterranean soil, without previous adaptation or contact with all these pollutants. This finding suggests the capacity of this soil to cope with pollution of both heavy metals and chlorophenols, and thus presumes the potential of the soil and some of the isolated strains to face up to co-contamination events.

5 Conclusions

The combination of respirometric assays with molecular methods has been useful to assess the impact of metals on the soil microbial community. Cr is showed as the most damaging metal in soil microcosms. However, if soluble fractions of the metals are considered instead of their total added amounts, a similar potential toxicity of Cr and Pb could be achieved. Therefore, these results indicate that the impact of metals can greatly be explained by their differential bioavailability. Cultivation-dependent and -independent approaches have proved the presence and development of resistant microorganisms, which could play an important role protecting the soil functions in potential contamination events. In relation to a calcareous Mediterranean soil that was also assessed with metals and chlorophenols (in the frame of a wider research project including both soils and types of pollutants), resistant microbial populations have found to be mostly specific for each soil environment. Moreover, the detection of resistant microorganisms to both inorganic and organic pollutants points out the potential of previously unpolluted soils to face up even to co-contamination events.