Effect of applying an arsenic-resistant and plant growth–promoting rhizobacterium to enhance soil arsenic phytoremediation by Populus deltoides LH05-17

Authors


Gejiao Wang, State Key Laboratory of Agricultural Microbiology, College of Life Science and Technology, Huazhong Agricultural University, Wuhan 430070, China. E-mail: gejiaow@yahoo.com.cn and Bingkun Tu, College of Horticulture and Forestry Science, Huazhong Agricultural University, Wuhan 430070, China. E-mail: bktu@mail.hzau.edu.cn

Abstract

Aims:  Bioremediation of highly arsenic (As)-contaminated soil is difficult because As is very toxic for plants and micro-organisms. The aim of this study was to investigate soil arsenic removal effects using poplar in combination with the inoculation of a plant growth–promoting rhizobacterium (PGPR).

Methods and Results:  A rhizobacterium D14 was isolated and identified within Agrobacterium radiobacter. This strain was highly resistant to arsenic and produced indole acetic acid and siderophore. Greenhouse pot bioremediation experiments were performed for 5 months using poplar (Populus deltoides LH05-17) grown on As-amended soils, inoculated with strain D14. The results showed that P. deltoides was an efficient arsenic accumulator; however, high As concentrations (150 and 300 mg kg−1) inhibited its growth. With the bacterial inoculation, in the 300 mg kg−1 As-amended soils, 54% As in the soil was removed, which was higher than the uninoculated treatments (43%), and As concentrations in roots, stems and leaves were significantly increased by 229, 113 and 291%, respectively. In addition, the As translocation ratio [(stems + leaves)/roots = 0·8] was significantly higher than the uninoculated treatments (0·5). About 45% As was translocated from roots to the above-ground tissues. The plant height and dry weight of roots, stems and leaves were all enhanced; the contents of chlorophyll and soluble sugar, and the activities of superoxide dismutase and catalase were all increased; and the content of a toxic compound malondialdehyde was decreased.

Conclusions:  The results indicated that the inoculation of strain D14 could contribute to the increase in the As tolerance of P. deltoides, promotion of the growth, increase in the uptake efficiency and enhancement of As translocation.

Significance and Impact of the Study:  The use of P. deltoides in combination with the inoculation of strain D14 provides a potential application for efficient soil arsenic bioremediation.

Introduction

China has become one of the countries that are severely affected by arsenic pollution. Large areas of arsenic-contaminated fields exist in southern China, which are mainly because of the mining of arsenic and other metal(loid)s (Sun 2004; Liao et al. 2005). Soil arsenic contamination resulted in desolation of agricultural land and damaging human health via the food chain (Huq et al. 2006). Arsenic contamination of soil is an important environmental problem when considering China as a leading agricultural country.

Phytoremediation is the use of green plants to remove or reduce toxicity of hazardous substances in the environment (Salt et al. 1998). Based on the removal efficiency, cost-effective and environmental-friendly characteristics, hyperaccumulators that can accumulate exceptionally high quantities of contaminants have received increasing attention (Wei et al. 2006). The ideal hyperaccumulators should possess the following characteristics: (i) high contaminant accumulation; (ii) high translocation efficiency; (iii) contaminant tolerance (Wei et al. 2005); and (iv) sufficient biomass enrichment (Wei et al. 2006). However, most metal-accumulating plants identified so far were less efficient for large field applications because of their small biomasses and slow growth in the presence of high concentrations of metal(loid)s (Glick 2003).

Trees and grasses have been actively evaluated for phytoremediation (Stoltz and Greger 2002). Fast-growing tree species, such as poplar, have been shown as suitable candidates to bioremediate heavy metal–polluted soils (Di Baccio et al. 2003). Moreover, Wagner (1993) reported that Populus species and more general Salicaceae could detect bioavailable metal(loid)s in soils with great sensitivity because of their widespread root system and great capability of accumulating trace metals. Besides these capabilities, poplar trees have been applied to bioremediate organic pollutant–contaminated soil and water, such as petroleum hydrocarbons (Jordahl et al. 1997) and chloroacetanilide herbicides (Gullner et al. 2001). They have also been studied to take up or immobilize different metal(loid)s in soils, such as As, Cd, Cu and Zn (Di Lonardo et al. 2011). In the case of grasses, a Chinese natural fern Pteris vittata (Ma et al. 2001) and a silverback fern, Pityrogramma calomelanos (Jankong et al. 2007), displayed highly efficient arsenic accumulation. Phytoremediation of arsenic-contaminated soil was also performed using Brassica napus (Nie et al. 2002) and Cicer arietinum (Gupta et al. 2008). However, to our knowledge, no arsenic phytoremediation using Populus deltoides has been reported so far.

Soil microbial activities could affect the mobility and bioavailability of metal(loid)s (Singh et al. 2010). Especially, some rhizobacteria with the characters of synthesizing phytohormone, producing siderophore, fixing nitrogen, or solubilizing phosphorus and other nutrients (Passardi et al. 2004), could promote the growth of plant (plant growth–promoting rhizobacterium, PGPR; Belimov et al. 2004). Inoculation of PGPR resulted in reduction of arsenic toxicity to Pi. calomelanos (Jankong et al. 2007) and Pt. vittata (Yang et al. 2011). Therefore, inoculation of arsenic-resistant PGPR with poplar may improve the phytoremediation efficiency in arsenic-contaminated soils.

Thus, the objectives of this study were to: (i) isolate and identify an arsenic-resistant rhizobacterium (strain D14) and determine its arsenic resistance and plant growth–promoting characteristics and (ii) investigate the arsenic accumulation of poplar (P. deltoides LH05-17) with or without the inoculation of strain D14. Our results provide evidences that P. deltoides is an efficient arsenic-accumulating plant. Addition of PGPR D14 enhanced P. deltoides soil arsenic removal efficiency.

Materials and methods

Isolation and identification of arsenic-resistant rhizobacteria

Rhizosphere soil was collected from roots of an arsenic hyperaccumulating plant Pt. vittata L., which had been planted in an arsenic phytoremediation experimental station located at Dengjiatang village, Chenzhou city, Hunan Province, Central South, China (25°48′N and 113°02′E, with an elevation of 185 m), where the soil was contaminated by arsenic because of the waterfall wastes from an adjacent arsenic smelting factory at higher elevation (Yang et al. 2011). Rhizosphere soil was washed out from plant roots using sterile water and incubated on chemically defined medium (CDM) plates (Weeger et al. 1999) with a final concentration of 800 μmol l−1 (63·5 mg l−1) NaAsO2 as described before (Fan et al. 2008). Bacterial arsenic resistance was defined as the lowest arsenic concentration that completely inhibited the growth of strains and was determined using liquid CDM medium amended with different concentrations of As(III) or As(V) and grown at 28°C for 3 day with 150 rev min−1 shaking.

For bacterial strain identification, the nearly full-length 16S rRNA gene sequence was amplified by polymerase chain reaction using the universal primers 27F (5′-AGAGTTTGATCCTGGCTCAG-3′) and 1492R (5′-GGTTACCTTGTTACGACTT-3′; Wilson et al. 1990) and analysed as described before (Fan et al. 2008). The sequence was compared with reference 16S rRNA gene sequences in the GenBank using ClustalX 1.83 software (Thompson et al. 1997). The 16S rRNA gene phylogenetic tree was constructed using the neighbour-joining distance method with the mega 3.1 software (Kumar et al. 2004), and the reliability of the inferred tree was tested by 1000 bootstrap. Whole-cell fatty acid analysis was analysed by GC (Hewlett-Packard 6890) according to the instructions of the Sherlock® Microbial Identification System (midi Sherlock version 4.5, midi database tsba40 4.10). Biochemical and physiological analyses were performed using API 20NE and API ID 32GN test strips according to the manufacturer’s instructions (BioMérieux, Lyon, France).

For determination of IAA products, bacterial strains were grown in 100 ml Luria Bertani (LB; peptone, 10 g l−1; yeast extract, 5 g l−1; NaCl, 10 g l−1) broth supplemented with 500 μg ml−1 tryptophan and incubated at 28°C for up to 3 days with 160 rpm shaking. The growth of the bacterium (OD600) and the concentrations of IAA (OD530) in the culture supernatants were monitored every 5 h using an ultraviolet spectrophotometer (DU800; Beckman Coulter Inc., Pasadena, CA) as described before (Bric et al. 1991). An uninoculated LB medium sample was used as a blank.

For determination of siderophore products, rhizobacterial strains were streaked out on chrome azurol S agar plates (CAS, 61 mg l−1; hexadecyl trimethyl ammonium bromide, 73 mg l−1; NaOH, 6·0 g l−1; 1, 4-piperazinediethanesulfonic acid, 30·2 g l−1; KH2PO4, 0·3 g l−1; NaCl, 0·5 g l−1; NH4Cl, 1·0 g l−1; MgCl2, 95 mg l−1; CaCl2, 11 mg l−1; FeCl3·6H2O, 2·6 mg l−1; glucose, 2·0 g l−1; casamino acid, 0·3%; HCl, 0·01 mmol l−1 and cultured at 28°C for 48 h. Orange halos surrounding the colonies on the CAS agar plate were used to determine the synthesis of siderophore by the bacterial strains (Schwyn and Neilands 1987).

Pot experiments in greenhouse

The soil was collected from the Lion Mountain of Huazhong Agricultural University (Wuhan, Hubei Province, China), with total As of 4·2 mg kg−1, total N of 4·7 g kg−1 (analysed according to Martínez-Sánchez 2006), available P of 37 g kg−1 (analysed as described by Feng et al. 2009), available K of 85 g kg−1 (Feng et al. 2009; dry matter basis) and a pH (in water) of 4·8. The soil was sterilized using 5% formalin for 7 days, air dried for another 7 day, ground and sieved through a 2-mm screen, and then incubated with 0, 150 and 300 mg kg−1 Na3AsO4·12H2O for 2 weeks to ensure a thorough mixing. The soil mixture was serially diluted with sterilized water and incubated in LB plates at 28°C for 7 days to confirm the completed sterilization.

Populus deltoides LH05-17 that is a fast-growing poplar clone in Hubei area was selected for this study. In March 2009, poplar cuttings with similar sizes (approx. 15 cm long, 6 mm dia) and bud numbers (3–4 buds) were selected for pot experiments in a greenhouse. Each cutting was planted into a 50-cm3 pot filled with 2 kg soil. Every treatment had four replications (four pots). Three soil As concentrations (0, 150 and 300 mg kg−1) were designed, which confer to control, low and high As-contaminated levels, respectively. For each of the 0, 150 and 300 mg kg−1 As treatment, strain D14 was inoculated; meanwhile, the pots with As treatment but without bacterial inoculation were used as controls. For bacterial treatments, the soil was mixed thoroughly with the bacterial cells with final bacterial concentration of 108 CFU kg−1 soil. The pots were placed in a controlled greenhouse at 25°C for 5 months and watered routinely as needed.

Analyses of soil characteristics and poplar parameters

Plant height was measured during the initial 5 months of growth; poplar plants were carefully removed, following this time, dried at 65°C for 4 day and weighed. Dry plant materials were acid digested as described by De Souza et al. (1999). The soil was thoroughly stirred and dried at 105°C until a constant weight was obtained. The As concentration was determined using atomic adsorption spectrometry (Varian, AA240FS-VGA77) in soils that were completely acid digested using nitric acid. The As translocation ratio from roots to above-ground tissues was calculated by As content(stems + leaves)/As contentroots. The translocated As percentage in the above-ground tissues was calculated by As content(stems + leaves)/As contentplants × 100%. In addition, the percentage of accumulated As from soil to poplar was calculated by As contentplants/As contentsoil × 100%.

For the poplar plant, the following parameters were determined after harvest. The soluble sugar acts as both energy source and signal that may regulate the plant growth (Koch et al. 2000). It was assayed by the anthrone colorimetric method (Buysse and Merckx 1993). The chlorophyll content and soluble protein can estimate plant growth status by different functions. They were measured by a SPAD-502 chlorophyll meter (Markwell et al. 1995) and coomassie brilliant blue G-250 staining (Reisner et al. 1975), respectively. Moreover, the superoxide dismutase (SOD) and catalase (CAT) are both stress defence proteins, and their activities can be used to evaluate the toxic effects of metal(loid)s to the plants (Gallego et al. 2001). Plant peroxidase (POD) is a bifunctional enzyme involved in both reduction of hydrogen peroxide (H2O2) and production of reactive oxygen species (ROS; Passardi et al. 2004). The activity of SOD was analysed by the nitro blue tetrazolium method (Beauchamps and Fridovich 1971). The activity of CAT was determined using a UV spectrophotometer with 4-amino-3-hydrazino-5-mercapto-1,2,4-triazole as a chromogen (Johanssona and Borg 1988). And the activity of POD was assayed by the guaiacol method (Mika and Lüthje 2003). Furthermore, malondialdehyde (MDA) is one of the compounds having a toxic effect on the plant cells, and its content was determined by the thiobarbituric acid method (Schmedes and Hølmer 1989).

anova analysis of the biological/physiological parameters of the plants and the arsenic concentration of soils, roots, stems and leaves was performed with R and VEGAN statistics packages (Dixon 2003).

Results

Isolation and identification of arsenic-resistant rhizobacteria

In this study, a total of 22 arsenic-resistant bacteria were isolated. One of these strains, named D14, which was highly resistant to As with a minimum inhibitory concentration of 14 mmol l−1 (1·05 g l−1) for As(III) and 150 mmol l−1 (11·24 g l−1) for As(V) in CDM medium, was further characterized. Both arsenic resistance levels were in the upper range of arsenic-resistant bacteria reported thus far (Fan et al. 2008; Cai et al. 2009).

The 16S rRNA gene sequence of strain D14 exhibited the highest nucleotide identity (99·4%) with a rhizobacterium Agrobacterium radiobacter strain ATCC 19358 (GenBank number AJ389904) and 99·3% identity with Ag. tumefaciens strain NCPPB 3554 (GenBank number D14500). The phylogenetic relationships among D14 and some GenBank 16S rRNA gene sequences are shown in Fig. 1a. The morphological, biochemical and physiological analyses, and the fatty acid data are shown in the (Appendix S1). In addition, strain D14 could be distinguished from Ag. tumefaciens as it did not contain a Ti or Ri plasmid to infect plants (data not shown, Sawada et al. 1993). Thus, strain D14 was identified as an Ag. radiobacter member. The 16S rRNA gene sequence of strain D14 has been deposited in the NCBI GenBank database as number GQ166481.

Figure 1.

 (a) The 16S rRNA gene phylogenetic tree of strain D14 and the standard strain of Agrobacterium sp. (b) The bacterial growth (bsl00001) and IAA production (bsl00066) by Agrobacterium sp. D14 and IAA production in Luria Bertani medium without bacterial inoculation which was used as a blank (•). Error bars represent standard deviations of triplicates.

Strain D14 possesses the ability to produce IAA. As shown in Fig. 1b, the cell growth and IAA production increased simultaneously, and a maximum IAA production (7·8 mg l−1) was observed after 28-h incubation. The strain also displayed siderophore production, as indicated by the orange halos surrounded the colonies on CAS agar plates, after 2-day incubation (data not shown). In consideration of the high arsenic resistance and the production of IAA and siderophore, strain D14 was selected for further arsenic bioremediation studies with P. deltoides.

Arsenic uptake by P. deltoides

Populus deltoides cuttings were planted into the soils and grown for 5 months in a greenhouse with or without the inoculation of strain D14. At the time of harvest, the poplar roots were expanded up to the full pots. Before the bacterial inoculations, the soil As concentrations were measured as 134·4 and 272·3 mg kg−1 for the amended As of 150 and 300 mg kg−1, respectively. After 5 months, without bacterial inoculation, the soil As concentrations were 76·1 and 155·4 mg kg−1, in the 150 and 300 mg kg−1 As treatments, respectively, which indicated that soil As was removed at almost the same efficiency (both 43%) in both the 150 and 300 mg kg−1 As treatments. With the bacterial inoculation, the soil As concentrations were 73·6 and 125·5 mg kg−1, which equalled to 45 and 54% soil As being removed, in the 150 and 300 mg kg−1 As-amended soils, respectively (Fig. 2a, < 0·001). The results indicated that P. deltoides itself showed an efficient As removal ability; however, inoculation with strain D14 enhanced the arsenic removal efficiency especially when there was high As concentration in the soils. Furthermore, the higher As content in the soil and the higher As removal efficiency were obtained by the addition of strain D14.

Figure 2.

 The enhanced As accumulating and growth effects of Populus deltoides with the inoculation of bacterium D14 after 5 months growth. (a–d) are the As concentrations in soil, roots, stems and leaves, respectively; (e) the plant height; (f) the root-collar diameter; (g, h and i) the dry weight of roots, stems and leaves, respectively. (inline image) the original As concentrations amended in the soils; (inline image), the non-inoculated treatments and (inline image) the bioinoculated treatments. Error bars represent standard deviations of quadruplicates. Percentages refer to variation compared to the non-As amended. Significantly different from the control which was non-bacteial inoculated or non-As amended at < 0·001.

After the 5 months, without the inoculation of strain D14, the As concentrations of the 150 mg kg−1 As-amended soils were 84·6, 1·5 and 8·2 mg kg−1 in roots, stems and leaves, respectively, while in the 300 mg kg−1 As-amended soils, they were 79·1, 2·8 and 17·7 mg kg−1 in roots, stems and leaves, respectively. The highest As concentration in roots was obtained in 150 mg kg−1 amended soils (Fig. 2b, < 0·001). The As concentrations in stems and leaves both experienced a small increase along with the increase in As concentration (Fig. 2c,d, < 0·001). With bacterial inoculation, the As concentrations in roots and stems were significantly increased except for the leaves in 150 mg kg−1 As-amended soils; however, in 300 mg kg−1 As-amended soils, the As concentrations in roots, stems and leaves were significantly increased by 229, 113 and 291%, respectively, compared to the uninoculated treatments (Fig. 2b–d, < 0·001). Moreover, the As translocation ratio from roots to above-ground tissues with the inoculation of strain D14 was calculated as 0·8, which was higher than the uninoculated treatments (0·5) in the 300 mg kg−1 As-amended soils (Fig. 2b–d). Furthermore, 45% As was translocated into the above-ground tissues {[(9·6 + 10·6) mg/(9·6 + 10·6 + 26·3) mg] × 100% = 45%}, and 8·5% soil As was accumulated into the poplar in the 300 mg kg−1 As-amended soils [(9·6 + 10·6 + 26·3) mg/(272·3 × 2) mg = 8·5%].

Bacterial inoculation enhanced the growth of P. deltoides

The P. deltoides grown in the greenhouse with different As concentration treatments and with or without bacterial inoculation showed variable results in plant growth and physiological characteristics. When As was added, without the bacterial inoculation, the plant height, root-collar diameter and dry weight of roots, stems and leaves were all decreased by different degrees, compared to the controls without amended As (Fig. 2e, < 0·001). After inoculation of strain D14, with 150 and 300 mg kg−1 As amended, these growth parameters were comparable to uninoculated plants (Fig. 2e–i, < 0·001). Except for the dry weights of stems with low As amended and leaves without As amended, the other parameters were all increased by different degrees (Fig. 2h,i, > 0·001). The results indicate that the growth of P. deltoides was negatively affected by increasing soil As concentrations; however, the addition of D14 significantly reduced As toxicity and promoted the growth of poplar.

Bacterial inoculation changed physiological parameters of P. deltoides

The poplars were exposed to soils with or without As amended. Soluble sugar, chlorophyll content, soluble protein, the activities of SOD, CAT and POD, and the MDA concentration were used to estimate As stress.

Without inoculation of strain D14, the soluble sugar contents in leaves increased in the low As-amended soil and decreased in the high As-amended soils, compared to As control treatment (Fig. 3a, < 0·001). With the addition of strain D14, the soluble sugar contents all increased, compared to the uninoculated controls (Fig. 3a, < 0·001).

Figure 3.

 The physiological parameters of Populus deltoides with and without the inoculation of Agrobacterium sp. D14 after 5 months growth. (a) the concentrations of soluble sugar in leaves; (b) the chlorophyll content of leaves; (c and d) the concentrations of soluble proteins of roots and leaves, respectively; (e and f) the activities of superoxide dismutase in roots and leaves, respectively; (g and h) the activities of catalase in roots and leaves; (i and j) the activities of plant peroxidase in roots and leaves, respectively; (k and l), the malondialdehyde concentrations in roots and leaves., inline image the un-inoculated treatments, and inline image, the D14 bio-inoculated treatments. Error bars represent standard deviations of quadruplicates. Percentages refer to variation compared to the non As amended. Significantly different from the control which is non-inoculated or non- amended at < 0·001.

Decreases in chlorophyll contents were found when the poplars were grown with increasing As concentrations without the inoculation of strain D14 (Fig. 3b, < 0·001). With the bacterial inoculation, the chlorophyll contents increased in both the low and high As-amended soils, compared to the uninoculated treatments (Fig. 3b, < 0·001). However, in the control soils, the chlorophyll contents did not change significantly (Fig. 3b, > 0·001).

The contents of soluble protein in roots and leaves both decreased in the high As-amended soils, compared to the nonamended controls (Fig. 3c,d, < 0·001). With the inoculation of strain D14, the contents of soluble protein in roots and leaves in the control and high As-amended soil along with the contents of soluble protein in leaves in the low As-amended soil were all significantly increased (< 0·001), compared to the uninoculated controls (Fig. 3c,d), except for the contents of soluble protein in roots in the low As-amended soil (> 0·001; Fig. 3b).

In our experiments, without inoculation of D14, the SOD activities of roots were increased in the low As-amended soils, whereas they were decreased in the high As-amended soils (Fig. 3e,f, < 0·001). With the inoculation of strain D14, the SOD activities of roots and leaves were all increased in the 0, 150 and 300 mg kg−1 As-amended soils, all compared to the uninoculated ones (Fig. 3e,f, < 0·001).

The activities of CAT decreased along with the increasing As concentrations without bacterial inoculation (Fig. 3g,h, < 0·001). When strain D14 was inoculated, the CAT activities of roots and leaves all increased in the low and high As-amended soils, compared to the uninoculated treatments (Fig. 3g,h, < 0·001). It is interesting to note that the activities of CAT increased more pronouncedly with the low As concentration (150 mg kg−1, < 0·001).

With the inoculation of strain D14, the POD activities of roots and leaves decreased in the low and high As-amended soils, compared to the nonbacterial treatments (Fig. 3i,j, < 0·001). MDA contents increased along with the increasing concentrations of As (Fig. 3k,l, < 0·001). When strain D14 was inoculated, the MDA concentrations in roots and leaves all decreased in the low and high As-amended soils (Fig. 3k,l, < 0·001).

Discussion

In this study, we have successfully isolated a strain D14 displaying high arsenic resistance and the ability of synthesizing IAA and siderophores. Using phylogenetic and phylotypic analyses, we confirmed this strain belonged to Ag. radiobacter. So far, the reported rhizobacterial species with PGPR abilities were mainly from Bradyrhizobium, Azotobacter, Rhizobium and Bacillus (Wu et al. 2009; Dary et al. 2010). To our knowledge, this is the first time to report an Ag. radiobacter species with both high arsenic resistance and plant growth–promoting abilities.

With the inoculation of strain D14, enhancements of As accumulation and translocation were significant in As-amended soil but both more pronounced in the higher As-amended soils (300 mg kg−1). It is reported that the IAA released by rhizobacteria could directly promote the growth of roots by stimulating elongation of the plant cells or increasing cell division (Minamisawa and Fukai 1991), which may enhance the root arsenic absorption. Moreover, the siderophore production by strain D14 may mobilize the As (V) in the soil in the process of taking up iron ions (Drewniak et al. 2008) which rendered As more soluble and bioavailable to plants. As a result, the bacterial inoculation enhanced arsenic accumulation, which is very similar to the arsenic hyperaccumulating fern Pt. vittata (Yang et al. 2011). In the research of Vamerali et al. (2009), 18% of soil As was accumulated into Populus alba after 2 years, but only about 0·4% As was translocated from roots to shoots. In our experiments, about 8·5% of soil As was accumulated into P. deltoides after 5 months, and 45% As was translocated into the above-ground tissues; the As translocation was significantly enhanced, which could be another key factor for the increased arsenic accumulation by P. deltoides. Moreover, we believed that if the poplar is planted for a longer time (i.e. 1 year rather than 5 months), more As will be accumulated into the plants.

The growth and physiological parameters of P. deltoides LH05-17 indicated the promotion of plant growth and the reduction of arsenic toxicity by strain D14. The production of IAA by strain D14 which is a plant growth hormone may contribute to the growth enhancement of P. deltoides. Jankong et al. (2007) reported that the inoculation of rhizobacteria increased the biomass of Pi. calomelanos and arsenic accumulation. In our study, all of the results revealed that the enhancement of arsenic resistance and growth of P. deltoides were probably caused by the inoculation of As-resistant strain D14. Therefore, the inoculation of strain D14 increased the As concentration in plant tissues and the total plant biomass, which could both promote the higher As accumulation by P. deltoides.

In summary, the results of this study indicated that P. deltoides LH05-17 was an efficient arsenic accumulator. Furthermore, strain D14 was a highly efficient PGPR that may protect P. deltoides against the toxic effects of As, promote the growth of the poplar, mobilize As in the soil, increase the uptake efficiency by the plants and enhance the ability of As translocation into above-ground plant tissues. Primary studies have successfully applied a Chinese fern Pt. vittata to phytoremediate soil arsenic (Ma et al. 2001). However, its application has environmental limitation because Pt. vittata preferably grows in warm and moist climates (Tripathi et al., 2007). Populus deltoides as a perennial woody poplar plant can grow over a wide range of temperatures in both humid and semi-arid environments and produce large inedible biomass that makes it an ideal plant for bioremediation.

PGPR plays a key role when plants need to adapt into a new environment, in both nature and managed ecosystems (Ma et al. 2011). The hormones and benefit compounds produced by these bacteria could help the plants to tolerate high concentrations of metals. Thus, the plants would be well adapted in a heavy metal–contaminated environment (Welbaum et al. 2004). Agrobacterium strains are widespread in soil. Most of the Agrobacterium species except for Ag. radiobacter were reported to induce cortical hypertrophy of the upper parts of roots (i.e. Ag. tumefaciens and Ag. rubi) or cause abnormal root growth (i.e. Ag. rhizogenes) on many kinds of plants (Sawada et al. 1993). Moreover, in our experiments, strain D14 did not produce the abnormal tumour in P. deltoides LH05-17 roots, which indicated that this strain should be safe and suitable for applying in the nature environment. For the features that Ag. radiobacter D14 is a no-pathogenic highly arsenic-resistant strain that produces IAA and siderophore, the use of P. deltoides in combination with the inoculation of Ag. radiobacter D14 provides a potential application for natural soil arsenic phytoremediation.

Acknowledgements

This work was supported by a Major International Joint Research Project, National Natural Science Foundation of China (31010103903), and by a National Natural Science Foundation of China (30970075). We would like to express our gratitude to Dr Christopher Rensing (The University of Arizona) for scientific comments.

Ancillary