Angela Sessitsch, AIT Austrian Institute of Technology GmbH, Bioresources Unit, A-2444 Seibersdorf, Austria. E-mail firstname.lastname@example.org
Aims: To characterize bacteria associated with Zn/Cd-accumulating Salix caprea regarding their potential to support heavy metal phytoextraction.
Methods and Results: Three different media allowed the isolation of 44 rhizosphere strains and 44 endophytes, resistant to Zn/Cd and mostly affiliated with Proteobacteria, Actinobacteria and Bacteroidetes/Chlorobi. 1-Aminocyclopropane-1-carboxylic acid deaminase (ACCD), indole acetic acid and siderophore production were detected in 41, 23 and 50% of the rhizosphere isolates and in 9, 55 and 2% of the endophytes, respectively. Fifteen rhizosphere bacteria and five endophytes were further tested for the production of metal-mobilizing metabolites by extracting contaminated soil with filtrates from liquid cultures. Four Actinobacteria mobilized Zn and/or Cd. The other strains immobilized Cd or both metals. An ACCD- and siderophore-producing, Zn/Cd-immobilizing rhizosphere isolate (Burkholderia sp.) and a Zn/Cd-mobilizing Actinobacterium endophyte were inoculated onto S. caprea. The rhizosphere isolate reduced metal uptake in roots, whereas the endophyte enhanced metal accumulation in leaves. Plant growth was not promoted.
Conclusions: Metal mobilization experiments predicted bacterial effects on S. caprea more reliably than standard tests for plant growth-promoting activities.
Significance and Impact of the Study: Bacteria, particularly Actinobacteria, associated with heavy metal-accumulating Salix have the potential to increase metal uptake, which can be predicted by mobilization experiments and may be applicable in phytoremediation.
Zn and Cd hyperaccumulators occurring in the northern temperate zone are typically small herbaceous Brassicaceae, such as Thlaspi cearulescens and Arabidopsis halleri (Baker and Brooks 1989). Despite high concentrations of metals accumulated in their leaves, the metal extraction efficiency of these plants is limited because of low biomass production (Chaney et al. 2000). Recently, Salix caprea (goat willow) trees, growing on heavy metal-contaminated sites in central Europe, have been found to accumulate up to 116 mg Cd kg−1 and 4680 mg Zn kg−1 in their leaves (Unterbrunner et al. 2007). Metal-accumulating Salix species are ideal extractor plants, as they produce as much as 10 dry t per ha per year of easily harvestable leaf biomass and develop a massive root system in the topsoil (Pulford and Watson 2003). Indeed, the Zn content and particularly the Cd content of a moderately contaminated soil (13·4 mg kg−1 Cd, 955 mg kg−1 Zn) could be successfully reduced by phytoextraction with Salix (Wieshammer et al. 2007).
ACCD is an enzyme that has no known function in bacteria but antagonizes ethylene synthesis in plants, by cleaving the ethylene precursor 1-aminocyclopropane-1-carboxylic acid (ACC). Accelerated ethylene synthesis (‘stress ethylene’) occurs in plants under environmental stress – including heavy metal stress – and causes damage to the plant organism (Glick 2003). ACCD-producing bacteria can inhibit stress-ethylene formation (Glick 2003) and have been reported to alleviate heavy metal toxicity (Burd et al. 1998). Bacterial siderophores are high-affinity Fe(III) chelators for the acquisition of iron under iron-limiting conditions. Under certain conditions, bacterial siderophores can be taken up by plant roots (Bar-Ness et al. 1992). Iron deficiency is a frequent symptom of plants under heavy metal stress and may be prevented by the import of bacterial siderophore–iron complexes (Glick 2003). Moreover, bacterial siderophores can complex a variety of heavy metal ions (Neilands 1981), and it has been speculated that they are involved in heavy metal mobilization (Abou-Shanab et al. 2003a; Kuffner et al. 2008). The auxin hormone IAA controls important processes in plants, such as growth and tissue differentiation. Auxin levels in plant tissues can be modulated by IAA-producing bacteria (Davies 1995), and bacterial supply of IAA may be important for successful growth in contaminated environments. Certain efflux-based systems of bacterial heavy metal resistance seem to involve postefflux sequestration of metals, i.e. the prevention of extruded metal ions from re-entering the cell by precipitation, chelation or by binding to exopolymers (Diels et al. 1995; Salt et al. 1999). Endophytes equipped with postefflux sequestration systems may contribute to heavy metal detoxification in plants (Lodewyckx et al. 2001). Bacteria producing ACCD, siderophores and IAA as well as metal-resistant phenotypes are commonly found in association with heavy metal-accumulating plants (Lodewyckx et al. 2001; Abou-Shanab et al. 2003b; Idris et al. 2004; Zaidi et al. 2006). Moreover, isolates with these characteristics have been shown to promote plant growth in the presence of heavy metals (Salt et al. 1999; Burd et al. 2000; Dell’Amico et al. 2008) and/or to enhance metal accumulation (Zaidi et al. 2006; Jiang et al. 2008; Braud et al. 2009). Presumably, the most direct way for bacteria to support heavy metal accumulation in plants is metal mobilization. Soil and rhizosphere bacteria can increase metal mobility and plant availability by various processes, potentially including redox transformations and the release of protons and organic acids (Gadd 2004). Although heavy metal mobilization has been assessed less frequently than IAA, ACCD, siderophores and metal resistance, individual bacteria have been demonstrated to promote heavy metal solubility and metal uptake in plants simultaneously (Whiting et al. 2001; Abou-Shanab et al. 2003a; Rajkumar and Freitas 2008).
The aim of this work was to characterize culturable bacteria associated with Zn/Cd-accumulating S. caprea trees regarding their potential to promote heavy metal phytoextraction. We compared rhizosphere and endophytic bacteria to gain insight into the characteristics of these subpopulations and into their role in increasing growth and heavy metal accumulation.
Materials and methods
For the isolation of bacteria, S. caprea trees growing on a former zinc/lead mining and processing site in Arnoldstein, Austria (Table 1) were sampled in June 2004. The site, which has been described in detail by Unterbrunner et al. (2007), is contaminated with Pb, Zn and Cd. Branches with leaves and fine roots with adherent rhizosphere soil from a depth of 0–25 cm were taken from four trees growing close to the contamination source, cooled to 4°C within 4 h and processed within 48 h.
Table 1. Selected parameters of the experimental soils
CAC, cation exchange capacity, measured at soil pH.
Sand/silt/clay (g kg−1)
CAC (mmol kg−1)
Organic carbon (g kg−1)
Total metal contents (in aqua regia)
Zn (mg kg−1)
Cd (mg kg−1)
Pb (mg kg−1)
Mobile fraction of metals (in 1 mol l−1 NH4NO3)
Zn (mg kg−1)
Cd (mg kg−1)
Pb (mg kg−1)
Isolation of rhizosphere bacteria and endophytes
For the isolation of rhizosphere bacteria, 5 g of fine roots and adherent soil was shaken in 50 ml of 1% tryptic soy broth (0·3 g l−1 TSB; Merck, Darmstadt, Germany) for 2 h at room temperature. Soil particles and roots were allowed to settle for 1 h, and tenfold dilutions were plated on three Zn-containing isolation media of different nutrient strength as different laboratory media select for different bacteria. Ten per cent tryptic soy agar (TSA) (3 g l−1 TSB, 15 g l−1 agar), 1% TSA and 0·08% diluted nutrient broth agar (0·08% DNBA) (Janssen et al. 2002) were used. All media were amended with cycloheximidine (100 μg ml−1) to inhibit fungal growth and with ZnSO4 (2 mmol l−1) to select for Zn-resistant bacteria.
For specific isolation of endophytes, 15 leaves (3·2–5·0 g) and five green branch segments of 0·5 cm diameter (2·0–3·5 g) were randomly picked from each tree, surface-sterilized in 5% (w/v) sodium hypochlorite for 5 min and rinsed with sterile water. Branches were further dipped into 70% ethanol, flamed and peeled. Surface-sterilized leaves and branches were cut into small pieces and ground with 50 ml of 0·9% NaCl in a stomacher (Stomacher Circulator; Seward, W. Sussex, UK) five times for 1 min. In the intervening intervals, the suspensions were cooled on ice. Tenfold dilutions were plated on 10% TSA, 1%TSA and 0·08% DNBA. Furthermore, xylem sap was extracted from lignified branches using a Scholander bomb and plated without diluting. Ten per cent TSA plates were incubated for 1 week, 1% TSA plates for 3 weeks and 0·08% DNBA plates for up to 12 weeks to allow colony formation of slowly growing bacteria. Colonies of all distinguishable morphology types were isolated on phosphate-poor morpholinepropanesulfonic acid medium (MOPS) (Neidhardt et al. 1974) containing 0·1% glucose and 1 mmol l−1 ZnSO4.
PCR-RFLP analysis of rhizosphere and endosphere isolates
To allow discrimination at the strain level, 16S–23S intergenic spacer (IGS) DNA was amplified from all isolates, using the primers p23SRO1 (5′-GGCTGCTTCTAAGCCAAC-3′) and pHr (5′-TGCGGCTGGATCACCTCCTT-3′) (Massol-Deya et al. 1995). DNA was released by boiling one loop of bacterial cells for 10 min in 150 μl sterile deionized H2O. One microlitre of lysis product was used as template in 50 μl PCR containing two units Taq DNA polymerase (Invitrogen, Carlsbad, CA), 0·2 mmol l−1 of each dNTP, 0·15 μmol l−1 of each primer and 1·5 mmol l−1 MgCl2. The thermal programme included an initial denaturation of 5 min at 95°C, 30 cycles of 1 min denaturation at 95°C, 1 min annealing at 55°C and 2 min elongation at 72°C and a final elongation of 10 min at 72°C. PCR products (15 μl) were digested for 4 h at 37°C with AluI (Invitrogen). Restriction fragments were electrophoretically separated in 3% agarose gels.
DNA extraction and partial 16S rDNA sequencing
For the amplification and analysis of the 16S rRNA gene, bacterial DNA was isolated by bead beating and phenol–chloroform extraction (Sessitsch et al. 2001). About 100 ng of DNA was used in 50 μl PCR with the primers 8f (5′-AGAGTTTGATCCTGGCTCAG-3′) (Weisburg et al. 1991) and 1520r (5′-AAGGAGGTGATCCAGCCGCA-3′) (Edwards et al. 1989). The composition of the reaction mix was identical with that described for the IGS PCR, the annealing temperature was 53°C. PCR products were purified through Sephadex G-50 (Amersham Biosciences, Buckinghamshire, UK) columns, and 2 μl was used as template in 10 μl sequencing reactions using 0·4 μmol l−1 of the primer 518r (5′-ATTACCGCGGCTGCTGG-3′) (Liu et al. 1997) and the BigDye terminator cycle sequencing kit (ABI Prism). After a second purification with sephadex G-50, the DNA fragments were sequenced with an ABI 373A automated DNA sequencer (Applied Biosystems Inc., Foster City, CA). Nucleotide Blast (Zhang et al. 2000) was used to search the sequence database of the National Center for Biotechnology Information (NCBI) for identified relatives of the isolates.
Nucleotide sequence accession numbers
The 16S rRNA gene sequences of rhizosphere and endophyte isolates described in this study have been entered to the database of the NCBI under the accession numbers GQ342532 to GQ342602. When identical sequences were obtained from several strains, one representative sequence was deposited. Sequences of the following strains are also representative for the strains indicated between parentheses: RD343 (RD358), RD344 (RD336), EI149 (ED123, EX145), ED162 (EX151, EX161), EX128 (ED306), RD74 (RX14, RD308), EI198 (EX107), EI187 (EI178, ED320, ED328), EI189 (EI174), RX196 (EX45), RI3 (RX2).
Heavy metal resistance, ACCD, IAA and siderophore production
Minimal inhibitory concentrations (MICs) of Zn and Cd, ACCD activity, IAA production and siderophore release were analysed in agar plate assays as described previously (Kuffner et al. 2008). Briefly, heavy metal MICs were determined as the lowest concentrations inhibiting growth on phosphate-poor minimal medium. ACCD production was determined as the ability to grow with ACC as sole nitrogen source. Siderophore production was determined as the ability to scavenge iron from chrome azurol-S agar. IAA was determined in bacteria grown in the dark (on 5 g l-1 glucose, 0.025 g l-1 yeast extract and 0.204 g l-1l-tryptophane, 15 g l-1 agar) by covering colonies with Salkowsky reagent (35% perchloric acid, 3 mmol-1 FeCl3) and dark incubation for 30 min.
Metal mobilization by bacterial metabolites
Bacteria were grown in 10% TSB medium (pH 7·2) at room temperature and 200 rev min−1. Late-log cultures were grown until the highest cell density was reached. Stationary phase cultures were allowed to grow twice as long. The optical density of the cultures was measured at 600 nm (OD600) with a DU® 640 Spectrophotometer (Beckman, Brea, CA). Cells were removed from the cultures by centrifugation (8000 g, 15 min, 4°C) and filtration (0·2 μm Ministart filters; Sartorius AG, Göttingen, Germany). The filtrates containing bacterial metabolites were stored at −20°C, and filtrate pH was measured prior to mobilization analysis. Air-dried and sieved (2 mm) soil from Arnoldstein (Kuffner et al. 2008) contaminated with Zn, Cd and Pb (Table 1) was used for the mobilization experiment. One gram of soil was shaken with 5 ml of culture filtrate (2 h, 20 inversions per min). The soil particles were removed by centrifugation (7000 rev min−1, 5 min) and filtration (0·45 μm filters, Roth), and Zn and Cd in the filtrates were quantified by Atomic Absorption Spectroscopy (AAS; Perkin-Elmer 2100). Fresh 10% TSB medium was used for control extractions. Bacterial cultures and control medium were prepared in triplicates. From each replicate, three aliquots were shaken with soil and analysed three times (n = 9). In a pre-experiment with one mobilizing and three immobilizing strains, the mobilization capacity of filtrates from late-log and stationary cultures was compared. In the stationary phase, immobilization effects were increased by 20–40%, and mobilization effects were more than doubled (data not shown). Therefore, only stationary culture filtrates were analysed from the remaining isolates.
Inoculation of Salix caprea plantlets
Cuttings of S. caprea, clone Boku 04 CZ-024 (derived from Kutna Hora, Czech Republic) (Unterbrunner et al. 2007), were pre-grown for 1 year in a sand–soil mixture under nonsterile conditions in a climate chamber (14/20°C day/night temperature; 80% air moisture, 16 h light per day). The experimental soil originated from Celje, Slovenia (Table 1). This moderately contaminated soil was used, because phytoextraction with trees will be principally applied to remediate intermediate heavy metal pollutions (Dickinson and Pulford 2005). The soil was air-dried, sieved (2 mm) and gamma ray irradiated with 25 kGray for 24 h by Mediscan GmbH (Seibersdorf, Austria). Bacteria were grown in 10% TSB until the late logarithmic phase, harvested by centrifugation (2420 g, 10 min, 4°C) and resuspended in 50 mmol l−1 potassium phosphate buffer (pH 7) to an OD600 of 0·4 (about 108 cells per ml). Willow plantlets were transplanted into pots containing 800 g of Celje soil, and 10 ml of bacterial suspension were applied with a pipet into the soil surface surrounding the plantlets. Thus, about 109 cells were applied to each plantlet. Pots were positioned in a greenhouse (16/22°C day/night temperature; 60% relative humidity, 16 h light per day) following a randomized design, and each treatment was replicated four times. After 12 weeks, roots, shoots and leaves were harvested and washed with tap water. To remove metals from the apparent free space of the root tissues, roots were sonicated in 0·05 mol l−1 CaCl2 for 10 min and rinsed with deionized water. Plant material was dried at 80°C for 24 h and weighed to assess the dry matter weight (dw). Subsamples of 0·5 g were digested in 4 ml HNO3 (Puriss. p.a.; Sigma–Aldrich Handels GmbH, Vienna, Austria) and 1 ml HClO4 (Puriss. p.a., Sigma–Aldrich Handels GmbH) at 225°C using an automated heating block (Digester DK 42/26; Velp Scientifica, Milano, Italy). Cd and Zn concentrations were determined by inductively coupled plasma mass spectrometry (ICP-MS, Elan 9000 DRCe; Perkin-Elmer, Waltham, MA).
Statistic analysis was carried out in statistica 6 (StatSoft, Tulsa, OK, USA). Analysis of variance (anova) followed by post hoc Dunnet test was carried out to identify significant effects of bacterial strains in mobilization assays and plant inoculation experiments. Correlations between the resistance to different metals as well as correlations between the mobilization of different metals were determined by product–moment correlation analysis.
Abundance and diversity of culturable Zn-resistant bacteria on different isolation media
For the isolation of plant-associated bacteria, xylem sap as well as extracts from leaves, branches and rhizosphere material were plated on three different media (10% TSA, 1% TSA, 0·08% DNBA) containing 2 mmol l−1 of Zn. The number of colony-forming units (CFUs) varied between samples from different trees. One gram of fresh leaves or branches yielded 8 × 102 ± 2 × 102 CFUs, 1 ml xylem sap 6 × 103 ± 2 × 103 CFUs, 1 g of rhizosphere soil 8 × 105 ± 2 × 105 CFUs. One hundred and eighty endophytes and 180 rhizosphere isolates were screened by RFLP analysis of 16S–23S IGS DNA. Forty-four different IGS-types were identified among the rhizophere bacteria, and 44 among the endophytes. Strains with identical IGS types were grouped into operational taxonomic units (OTUs). Forty-three per cent of the rhizosphere OTUs originated from 10% TSA, 11% from 1% TSA and 21% from 0·08% DNBA (Fig. 1a). Among endophytes, 10% TSA, 1% TSA and 0·08% DNBA, respectively, contributed 43, 18 and 23% to isolate diversity (Fig. 1b). Twenty-five per cent of the rhizosphere OTUs and 16% of the endophyte OTUs were found on more than one medium. Of the endophyte OTUs, 55% originated from xylem sap, 14% from leaves, 11% from twigs, and 20% were found in more than one plant organ (Fig. 1c).
Phylogenetic affiliation of rhizosphere and endosphere isolates
For phylogenetic identification, 16S rDNA was amplified from one representative strain of each OTU. Fragments of about 450 bp were sequenced and compared to NCBI sequence database entries (Table 2). Alphaproteobacteria, Betaproteobacteria, Actinobacteria and Bacteroidetes/Chlorobi accounted for 30, 16, 23 and 25% of rhizosphere isolate diversity, respectively. Furthermore, Gammaproteobacteria and Firmicutes were represented in the rhizosphere isolate collection. Culturable endophytes were phylogenetically distinct from rhizosphere bacteria. Although rhizosphere and endophyte isolate collections included both 44 different species, endophyte diversity was much lower at the genus level. More than 60% of the analysed endophytes clustered with alphaproteobacterial genera Sphingomonas and Methylobacterium. Twenty-three per cent of the endophytes were affiliated with the division Actinobacteria. Moreover, the divisions Betaproteobacteria, Firmicutes and Bacteroidetes/Chlorobi were represented by individual endophyte isolates. Eight rhizosphere bacteria and seven endophytes had <97% 16S rRNA gene sequence identity to described bacteria and may therefore represent novel bacterial lineages. Ten of the potentially novel species clustered with Bacteroidetes/Chlorobi group organisms.
Table 2. Phylogenetic affiliation, 1-aminocyclopropane-1-carboxylic acid deaminase (ACCD), indole-3-acetic acid (IAA) and siderophore (SID) production of (A) rhizosphere bacteria and (B) endophytes
Closest identified relative† [accession number]
*The strain nomenclature gives information about the isolation medium (X for 10% TSA, I for 1% TSA and D for diluted nutrient broth agar). Isolates selected for further analysis are highlighted in bold.
†Phylogenetic affiliations are based on sequence analysis of about 450 bp of the 16S rRNA gene.
‡ND For certain strains, the production of ACCD, IAA or siderophores could not be determined because of lack of suitable minimal media.
ACCD activity was determined as the ability to use ACC as sole nitrogen source. Of the 44 rhizosphere strains, 41% were able to metabolize ACC (Table 2). Among endophytes, 9% were ACCD positive, and 60% were negative. For the remaining endophyte isolates, ACCD production could not be determined because of lack of a suitable minimal medium. IAA formation was found in 23% of the rhizosphere bacteria and in 55% of the analysed endophytes (Table 2). Forty-one per cent of the selected rhizosphere bacteria produced siderophores. In contrast, only one siderophore-producing endophyte was identified (Table 2). The distribution of IAA production, ACCD activity and siderophore release was not correlated. Within the genera Bradyrhizobium, Variovorax, Methylobacterium and Frigoribacterium, isolates with very similar or identical 16S rDNA sequences differed in siderophore or IAA production. Fifteen rhizosphere isolates and five endophytes, representing dominant genera with and without IAA, ACCD and siderophore production were selected for further analysis. In general, the ability to produce IAA was frequently found among endophytes, whereas siderophore and ACCD production was more common among rhizosphere bacteria.
Zn and Cd resistance of endophyte and rhizosphere isolates
Table 3 shows Zn and Cd resistance of the 20 selected isolates. MICs of Zn ranged between 12 and 16 mmol l−1. The highest observed Cd MIC was 8 mmol l−1 for Bradyrhizobium RI12. Frigoribacterium EX166 was inhibited by 0·5 mmol l−1, the lowest Cd dose given. The Cd MICs of the remaining strains ranged between 1 and 6 mmol l−1. The levels of Cd and Zn resistance did not correlate with each other.
Table 3. Zn and Cd resistance of selected rhizosphere bacteria and endophytes
MIC (mmol l−1)
MIC, minimal inhibitory concentration.
Heavy metal mobilization from soil
The ability of bacterial metabolites to mobilize metals was tested by leaching contaminated soil with filtrates of stationary bacterial cultures and quantifying the extracted Zn and Cd (Fig. 2). Sterile TSB medium extracted 2·62 mg Zn (equivalent to c. 0·1% of the soil concentration) and 173 μg Cd per kg soil (equivalent to c. 0·5% of the soil concentration). Cd extractability was reduced by metabolites from 14 of 15 rhizosphere bacteria and by four of five endophytes. Eight of these rhizosphere bacteria and two of the endophytes also immobilized Zn. The Zn- and Cd-immobilizing rhizosphere bacteria included five of the eight analysed siderophore producers. None of the siderophore producers had a positive effect on Zn or Cd extractability. One rhizosphere bacterium (RD334) increased Cd mobility. Three endophyte strains (EX72, EX104 and EX166) doubled Zn mobility. EX72 also strongly enhanced Cd extraction. The effects of the bacterial metabolites on Zn and Cd mobility correlated positively with each other (r = 0·60; P < 0·05) but did not correlate with culture filtrate pH. The pH of fresh sterile 10% TSB was 7·0 and rose during the growth of all analysed strains to values between 7·5 and 8·7 in the stationary cultures (data not shown).
Inoculation of Salix caprea plantlets
The rhizosphere isolate RX232 and the endophyte EX72 were selected for an inoculation experiment with S. caprea. RX232 was a strong ACCD producer, siderophore producer and Zn/Cd immobilizer. In contrast, EX72 did not produce ACCD, IAA and siderophores but had the ability to mobilize Zn and Cd. About 109 cells were applied to 1-year-old nonsterile plantlets, which were grown in gamma-sterilized soil moderately contaminated with Zn and Cd (Table 1). Three months later, dry weight, Zn and Cd concentration of leaves and roots were determined. No visible symptoms of heavy metal toxicity (i.e. chlorosis and/or necrosis of leaves) were observed. Neither the rhizosphere strain nor the endophyte significantly influenced biomass production of roots or leaves. However, strain RX232 promoted root growth tendentiously (Fig. 3a). In all plants, Zn and Cd concentrations in leaves were significantly higher than in roots (Fig. 3b,c). With respect to noninoculated control plants, plants inoculated with RX232 showed significantly reduced concentrations of Zn and Cd in roots (P < 0·05), and plants inoculated with EX72 showed significantly increased Zn and Cd concentrations in leaves (P < 0·05).
The majority of environmental bacteria are unable to grow on laboratory media. However, culturable rhizosphere bacteria and endophytes may serve as model organisms for studying plant–microbe interactions. Moreover, culturable bacteria can be enriched and used for various applications such as for bioaugmentation (Pilon-Smits 2005). Highest diversity of heavy metal–tolerant rhizosphere bacteria and endophytes was found on 10% TSA, suggesting that many plant-associated taxa require higher nutrient concentrations for in vitro growth. In contrast to that, bulk soil bacteria tend to prefer diluted media (Janssen et al. 2002). Still, 1% TSA and 0·08% DNBA favoured slowly growing heavy metal-tolerating species, and each medium contributed 10–20% to endophyte and rhizosphere isolate diversity. In heavy metal accumulators, leaves are the site of metal detoxification and storage (Vazquez et al. 1994; Schat et al. 2000), and remarkably, several endophyte strains have been obtained from this environment. Most endophytes were isolated from the xylem, for which we have no information on the prevailing heavy metal concentration.
In accordance with observations on other plants (Berg et al. 2005), the rhizosphere and endosphere of S. caprea were colonized by different communities. Two of our observations on community composition corresponded well with data from culture-independent surveys. First, culturable endophytes were dominated by Sphingomonas and Methylobacterium, which also prevailed in endophytic 16S rDNA sequence libraries from Zn- and Cd-accumulating Thlaspi (Idris et al. 2004). These genera seem to have a competitive advantage in metal-enriched shoots independent of the plant species and may also have plant growth-promoting effects. Individual methylobacteria have been shown to promote the growth of heavy metal–accumulating plants in contaminated soil (Madhaiyan et al. 2007), and inoculation with Sphingomonas has improved heavy metal uptake (Abou-Shanab et al. 2003a). Second, rhizosphere and endosphere isolates included many different organisms of the division Actinobacteria. rRNA-based community analysis showed that Actinobacteria might be a dominant part in the metabolically active rhizosphere population of heavy metal accumulators (Gremion et al. 2003). Our data suggest that Actinobacteria were similarly important in rhizosphere and endosphere of S. caprea. In addition, culturable methylobacteria, sphingomonads and members of the division Actinobacteria have been isolated previously from Thlaspi and Alyssum species accumulating Zn/Cd and Ni, respectively (Lodewyckx et al. 2002; Idris et al. 2004; Barzanti et al. 2007).
Rhizosphere bacteria and endophytes were isolated on Zn-containing media and showed Zn and Cd resistances much higher than those reported for isolates from Zn-hyperaccumulating Thlaspi caerulescens plants (Lodewyckx et al. 2002). Moreover, the observed tolerance levels exceeded the concentration of mobile Zn in bulk soil of the sampling site (Wenzel and Jockwer 1999), suggesting Zn and Cd mobilization in the rhizosphere of S. caprea and high bioavailability in the endosphere. The resistant bacteria may contribute to metal detoxification via postefflux sequestration (Diels et al. 1995; Salt et al. 1999).
Plate assays for IAA, ACCD and siderophore production were carried out with all 44 rhizosphere bacteria and 44 endophytes to asses their potential for plant growth promotion. Important proportions of the rhizosphere isolates produced IAA (23%), ACCD (41%) and siderophores (41%). Among endophytes, IAA synthesis was more frequent (55%) than in the rhizosphere, whereas ACCD activity and siderophore release were only detected in few individuals. The number of ACCD-producing endophytes may have been underestimated in this study. For about 30% of the endophyte isolates, ACCD activity could not be determined as they failed to grow on the test medium. Idris et al. (2004) found a higher percentage of ACCD-producing bacteria in the endosphere than in the rhizosphere of Ni-accumulating Thlaspi goesingense and explained this with the more intimate relationship of endophytes to their host plant. These authors also observed siderophore production in all culturable endophytes and rhizosphere bacteria. In S. caprea, the competition for iron might be less pronounced. The highest Zn resistance was found in siderophore-producing bacteria. Siderophore release can be induced by the presence of heavy metals, and siderophores may be involved in bacterial Zn resistance, for instance in postefflux chelation (van der Lelie et al. 2000).
Apart from supporting growth of the accumulating biomass, rhizosphere bacteria may mobilize heavy metals for enhanced uptake by plant roots (Gadd 2004). Twenty isolates representing dominant rhizosphere and endophyte taxa with and without the capacity to produce IAA, ACCD and siderophores were tested for their Zn and Cd mobilization potential. Most of these strains reduced the extractability of Zn and Cd when contaminated soil was shaken with filtrates of stationary liquid cultures. The medium pH rose during bacterial growth in all cultures. However, the immobilization was not affected by pH, because pH adjustments of fresh 10% TSB (pH 7·2–8·9) did not alter Zn extraction (data not shown). Heavy metal immobilization has been observed previously, when rhizosphere bacteria were directly applied to contaminated soil (Abou-Shanab et al. 2003a; Madhaiyan et al. 2007). Various products of bacterial growth, ranging from organic acids and alcohols to exopolymers, can bind metal ions and may trap them to soil particles (Gadd 2004). Moreover, heavy metal–resistant bacteria have been observed to remove heavy metals from liquid medium by precipitation (Diels et al. 1995). Culture filtrates of four slowly growing strains affiliated with the division Actinobaceria strongly increased the extractability of Cd and/or Zn. This was not an effect of acidification, because the medium pH rose in metal-mobilizing and metal-immobilizing cultures at the same rate. Authors, who observed bacteria to raise pH and to mobilize metals at the same time, speculated that bacterial siderophores may extract heavy metals from soil along with iron (Whiting et al. 2001; Kalinowski et al. 2004). Additionally, in none of the studies on changes in rhizosphere pH, a significant pH decrease was found (Wenzel et al. 2004). This suggests that pH is generally not involved in metal mobilization by plants and their associated rhizosphere bacteria. None of the analysed siderophore producers mobilized Zn or Cd, and six of eight siderophore producers even immobilized both metals. However, the role of siderophores in the observed Zn/Cd mobilizations and immobilizations remains unclear because the experiment was carried out in an iron-containing medium, which may not have induced the release of bacterial siderophores. The four metal-mobilizing strains were unable to produce siderophores showing that other mechanisms or metabolites play an important role in mobilization. They may either synthesize-specific ligands for Zn and/or Cd or form nonspecific metal-binding compounds during growth in TSB, which were not formed in the siderophore test minimal medium. Chemical analysis is necessary to reveal the nature of the metal-mobilizing and metal-immobilizing compounds, and it remains to be confirmed whether they can be produced from the substrates available in the rhizosphere and endosphere. Interestingly, three of the identified metal-mobilizing Actinobacteria were endophytes. Many endophytes derive from the rhizosphere (Huang 1986) and may mobilize heavy metals during rhizosphere colonization. Inside the plant, they may be involved in metal translocation. Rhizosphere and endosphere Actinobacteria, particularly Microbacterium strains, have been observed previously to mobilize metals and to enhance their accumulation in several plants (Whiting et al. 2001; Abou-Shanab et al. 2003a; Kuffner et al. 2008; Sheng et al. 2008).
The strains RX232 and EX72 were selected for a bioaugmentation experiment with S. caprea, based on their contrasting characteristics. RX232 was a rhizosphere isolate, closely related to a powerful plant growth promoter, Burkholderia phytofirmans PsJN (Sessitsch et al. 2005). It showed particularly intense ACCD activity, siderophore production and Zn/Cd-immobilizing effects. EX72 was an endophyte affiliated with the family of Microbacteriaceae (Actinobacteria), unable of IAA-, ACCD- and siderophore production, but strongly mobilizing Zn and Cd in the mobilization assay. In a previous study, we demonstrated that growth of heavy metal-accumulating S. caprea can be promoted by bacterial inoculation (Kuffner et al. 2008). The ACCD- and siderophore-producing rhizosphere strain RX232 tendentiously promoted root growth. However, neither RX232 nor EX72 had a significant effect on root or shoot biomass production. Salix caprea tolerates higher concentrations of Zn and Cd than present in the experimental soil (Unterbrunner et al. 2007). Most likely, the metal stress given in our experimental system did not exceed plant’s internal detoxification capacities and did not limit root or leaf growth. The presence of the rhizosphere strain RX232 reduced Zn and Cd import into the roots of S. caprea but did not affect the amount of heavy metal translocated to shoots. RX232 may have produced the same metal-immobilizing compounds as in the in vitro assay and thus may have reduced Zn/Cd availability to S. caprea. The endophyte EX72-enhanced Zn/Cd accumulation in leaves but did not change the amount of metals retained in roots. EX72 may have colonized the rhizosphere and may have mobilized Zn/Cd by release of those metabolites that had been detected in the in vitro assay. The surplus of available Zn/Cd may have been taken up by S. caprea roots and may have stimulated root-to-leaf translocation activity. The endophyte EX72 may also have entered the plant interior and may have induced the expression of heavy metal transporters. In all treatments, the accumulation of Zn and Cd in S. caprea was affected in a similar manner, suggesting that both metals were influenced by the inoculated strains likewise and no competition for uptake occurred.
Screening for IAA, ACCD and siderophore production revealed a high plant growth-promoting potential among culturable rhizosphere bacteria and endophytes from S. caprea. But as these traits were frequent and their distribution was not correlated, plate tests failed to identify a group of ‘most promising growth promoters’ producing ACCD, IAA and siderophores. Metal mobilization capacity was more laborious to test than IAA, ACCD and siderophore production and identified only few mobilizers. However, results of mobilization assays predicted the effect of individual bacteria on heavy metal accumulation in S. caprea more reliably. In the greenhouse experiment, the Zn/Cd mobilizer increased, and the Zn/Cd immobilizer reduced the uptake of these metals. More isolates have to be tested to confirm this correlation. Furthermore, analysis of metal mobilization drew our attention to the division Actinobacteria. Compiling our observations and the results of other studies, Actinobacteria constitute diverse and active populations in heavy metal accumulators and may fulfil key functions in metal uptake and translocation. The Zn/Cd-mobilizing Microbaterium EX72 could be employed for improving phytoextraction efficiency of S. caprea. Future studies should address the specific requirements of Actinobacteria, and techniques to favour their growth and activity in phytoremediation environments should be developed.
This project was funded by the Wiener Wissenschafts-, Forschungs- und Technologiefonds (WWTF). Melanie Kuffner received a DOC fellowship from the Austrian Academy of Sciences (Doktorandenprogramm der Österreichischen Akademie der Wissenschaften). The authors acknowledge DI Martin Schwab, Austrian Agency for Food Safety, for preparing the Salix caprea plantlets.