Environ Sci Pollut Res DOI 10.1007/s11356-013-2450-y
RESEARCH ARTICLE
The sequestration of trace elements by willow (Salix purpurea) —which soil properties favor uptake and accumulation? Benoît Cloutier-Hurteau & Marie-Claude Turmel & Catherine Mercier & François Courchesne
Received: 9 October 2013 / Accepted: 9 December 2013 # Springer-Verlag Berlin Heidelberg 2013
Abstract The effect of soil properties on trace element (TE) extraction by the Fish Creek willow cultivar was assessed in a 4-month greenhouse experiment with two contrasted soils and two mycorrhizal treatments (Rhizophagus irregularis and natives). Aboveground tissues represented more than 82 % of the willow biomass and were the major sink for TE. Cadmium and Zn were concentrated in leaves, while As, Cu, Ni, and Pb were mostly found in roots. Willow bioconcentration ratios were below 0.20 for As, Cu, Ni, and Pb and reached 10.0 for Cd and 1.97 for Zn. More significant differences in willow biomass, TE concentrations, and contents were recorded between soil types than between mycorrhizal treatments. A slight significant increase in Cu extraction by willow in symbiosis with Rhizophagus irregularis was observed and was linked to increased shoot biomass. Significant regression models between TE in willow and soil properties were found in leaves (As, Ni), shoots (As, Cd, Cu, Ni), and roots (As, Cu, Pb). Most of the explanation was shared between soil watersoluble TE and fertility variables, indicating that TE phytoextraction is related to soil properties. Managing interactions between TE and major nutrients in soil appeared as a key to improve TE phytoextraction by willows.
Keywords Phytoextraction . Fish Creek willow cultivar . Greenhouse experiment . Mixed contamination . Rhizophagus irregularis . Rhizosphere Responsible editor: Philippe Garrigues Electronic supplementary material The online version of this article (doi:10.1007/s11356-013-2450-y) contains supplementary material, which is available to authorized users. B. Cloutier-Hurteau (*) : M. K for cations and Cl > SO4 ≫ NO3 > PO4 (systematically below DLM) for anions (Table 1). The two-way ANOVA results indicated that the water-soluble major element concentrations significantly differed between bulk soils for K, Na, Ca, Mg, and NO3 and for K, Na, Ca, Mg, and SO4 in the rhizosphere. For all watersoluble major elements, there were no significant effect of the mycorrhizal treatment for both the bulk and the rhizosphere soils (data not shown). The water-soluble TE concentrations fell also within one order of magnitude and were all below 700 μg/kg (Table 1), with abundances following the sequence Cu > Ni > Zn > As ≫ Pb > Cd (Table 1). In both soil components, significant differences in As-H2O, Cu-H2O, Pb-H2O, and Zn-H2O concentrations were noted between the soils, whereas Ni-H2O differed only in the rhizosphere (Table 2). The effect of the mycorrhizal treatment was undetectable, except for Cu-H2O and Pb-H2O in the rhizosphere where lower concentrations were measured in the R− compared to the R+ pots (Table 2). The stronger contrast between soil types than between mycorrhizal treatments was confirmed by the LDA performed on water-soluble TE in the rhizosphere (Figure S1). The C− rhizosphere had higher Cd-H2O and Ni-H2O than the C+ rhizosphere, whereas the inverse trend was seen for As-H2O, Cu-H2O, Pb-H2O, and Zn-H2O (inset of Figure S1 and Table 1). The multivariate approach also showed that the mycorrhizal treatment was better differentiated in the C+ than the C− rhizosphere, mostly because of Cu-H2O. Aboveground plant materials represented more than 82 % of the total willow biomass mostly because of shoots (Table 3).
The individual total recoverable TE concentrations in willow tissues ranged from 0.04 to 3.82 mg/kg for As, from 1.54 to 17.9 mg/kg for Cd, from 4.44 to 32.7 mg/kg for Cu, from 0.20 to 10.4 mg/kg for Ni, from 0.15 to 12.2 mg/kg for Pb, and from 39.8 to 684 mg/kg for Zn. The TE were well contrasted between willow parts with Cd and Zn mostly concentrated in leaves, while As, Cu, Ni, and Pb reached generally maximum levels in roots (Table 3 and Fig. S2). When the total amount of TE sequestered was considered, the order follows: Zn ≫ Cu > Cd ∼ Ni > Pb > As. The sequestration of TE was concentrated in the aboveground biomass, and only As displayed a somewhat balanced sequestration pattern (Table 3). More significant differences in willow biomass, TE concentrations and TE contents were recorded between soil types than between mycorrhizal treatments (Table 2). The soil effect was characterized by significantly higher leaf biomass, shoot As and Cu concentrations, root As, Cu, and Pb levels, aboveground As and Cu contents, and belowground As, Cu, and Pb contents in C+ compared to C− soils. The inverse pattern was noted for Cd and Ni, with higher leaf Cd and Ni concentrations, shoot Ni and above- and belowground Ni contents in the C− than the C+ soil (Tables 2 and 3). The mycorrhizal treatment induced a slight effect on the aboveground and shoot biomass, the aboveground Cu and on root Zn. In all cases, the inoculated treatment had significantly higher values (Tables 2 and 3). Relationships between trace elements in willows and soil properties Simple linear regressions performed between TE in willows and water-soluble TE in either the bulk or the rhizosphere soils were computed and the significant relationships are displayed in Fig. 1. At least one significant relationship between willow tissues and soil concentrations was obtained for each TE, except for Zn. Arsenic and Cu were the TE showing the higher number of significant fit between plant concentrations or contents and water-soluble TE in soils, with most relationships observed for willow shoots and roots (Fig. 1). All the significant relationships, no matter the response variables, were positive with Cd in shoots being the only exception. The predictive models for leaves were weak and barely significant. Several shoot TE models were strongly predicted with adjusted R2 values of 0.72 for As and of 0.45 for Cu (Fig. 1). For roots, the As, Cu, and Pb models performed well with adjusted R2 of 0.66, 0.54, and 0.47, respectively. When the TE contents in either the above- or the belowground materials were considered, few significant models emerged except for As in both cases, Cu in aboveground parts and Pb in the belowground biomass (Fig. 1). However, only As was well described with predictive models explaining 43 and 57 % of the As variance in the above- and belowground biomass of willow, respectively. In general, the bulk and rhizosphere models performed similarly, although using the
Environ Sci Pollut Res Table 1 Mean (and standard deviation) values for soil properties and for water-soluble major and trace element concentrations in the bulk and the rhizosphere soil components at the end of the experiment Mycorrhizae
R−
Soil
C−
Soil component
Bulk
R+ C+ Rhizosphere
Bulk
C− Rhizosphere
Soil properties (n=5 for the R+/C− treatment and n=6 for the other treatments) SOC (g kg−1) 26.6 (2.89) 27.9 (5.18) 34.5 (2.80) 39.7 (4.35) CEC (cmolc kg−1) 30.1 (1.23) 30.2 (2.47) 31.0 (0.88) 31.6 (2.73) Water-soluble major elements (n=3 for all treatments) pH-H2O 7.16 (0.19) 7.27 (0.26) 7.16 (0.19) 7.31 (0.09) EC-H2O (mS cm−1) 0.11 (0.01) 0.11 (0.01) 0.14 (0.02) 0.15 (0.03) DOC-H2O (g kg−1) 0.33 (0.03) 0.29 (0.03) 0.50 (0.09) 0.38 (0.09) K-H2O (mg kg−1) 9.24 (1.81) 10.6 (0.52) 12.4 (1.13) 14.6 (1.35) Na-H2O (mg kg−1) 154 (21.4) 143 (19.8) 209 (12.3) 209 (20.2) Ca-H2O (mg kg−1) 67.9 (3.09) 64.7 (4.24) 76.3 (12.6) 76.8 (14.2) Mg-H2O (mg kg−1) 16.7 (0.54) 17.1 (1.08) 18.5 (3.86) 20.7 (4.17) Cl-H2O (mg kg−1) 103 (7.59) 97.8 (12.2) 96.8 (15.2) 103 (26.6) NO3-H2O (mg kg−1) 11.3 (3.64) 6.53 (9.44) 1.96 (1.87) 0.85 (1.47) SO4-H2O (mg kg−1) 63.0 (0.53) 72.7 (6.83) 94.3 (31.7) 131 (41.7) Water-soluble trace elements (n=5 for R+/C− treatment and n=6 for the other treatments) As-H2O (μg kg−1) 19.0 (1.75) 19.2 (2.68) 46.3 (6.16) 41.9 (7.72) Cd-H2O (μg kg−1) Cu-H2O (μg kg−1) Ni-H2O (μg kg−1) Pb-H2O (μg kg−1) Zn-H2O (μg kg−1)
1.81 (0.24) 428 (35.2) 107 (8.87) 5.67 (1.24) 47.8 (14.9)
1.90 (0.40) 427 (38.9) 116 (13.0) 7.79 (1.03) 57.1 (16.2)
2.14 (0.58) 562 (54.9) 109 (18.4) 14.1 (3.30) 64.3 (13.4)
1.65 (0.67) 586 (41.5) 103 (16.0) 12.5 (2.53) 67.9 (9.11)
C+
Bulk
Rhizosphere
Bulk
Rhizosphere
26.6 (3.03) 31.0 (0.99)
28.8 (1.76) 31.9 (1.99)
35.3 (1.79) 30.8 (0.73)
38.2 (2.09) 33.9 (2.91)
7.38 (0.02) 0.12 (0.02) 0.34 (0.05) 8.78 (0.81) 170 (30.3) 73.0 (7.10) 18.0 (2.49) 117 (25.9) 10.4 (4.30) 74.0 (20.7)
7.22 (0.12) 0.11 (0.01) 0.30 (0.03) 10.1 (1.34) 143 (7.92) 65.0 (0.93) 16.8 (0.31) 103 (21.2) 5.18 (3.57) 80.8 (10.8)
7.30 (0.06) 0.13 (0.01) 0.43 (0.06) 14.3 (2.50) 190 (16.4) 85.0 (2.26) 21.6 (2.16) 91.1 (21.0) 6.34 (1.86) 91.7 (39.5)
7.08 (0.19) 0.14 (0.01) 0.33 (0.05) 13.3 (1.53) 194 (5.89) 75.9 (9.66) 18.8 (2.08) 94.9 (4.19) 7.45 (4.46) 122 (20.8)
18.2 (2.83)
18.1 (1.72)
41.3 (6.94)
37.9 (5.17)
1.90 (0.39) 430 (53.4) 106 (14.0) 6.66 (1.22) 44.7 (10.5)
1.85 (0.27) 418 (26.5) 110 (13.6) 5.96 (1.83) 45.4 (12.5)
1.80 (0.39) 499 (29.2) 96.9 (13.9) 14.6 (4.31) 63.1 (18.0)
1.46 (0.27) 497 (79.7) 90.7 (9.63) 11.1 (1.67) 72.6 (18.2)
Variables followed by the suffix “-H2O” were measured in water extract (1 :10 soil-to-water ratio). The detection limit of the method of the trace elements measured in the water-soluble extract were set to 0.75, 0.34, 9.15, 11.0, 1.45, and 9.61 μg kg−1 , respectively, for As, Cd, Cu, Ni, Pb, and Zn R− without mycorrhizal inoculation, R+ inoculated with the mycorrhizal fungus Rhizophagus irregularis, C− non-contaminated soil, C+ contaminated soil, SOC soil organic carbon, CEC cation-exchange capacity, EC electrical conductivity, DOC dissolved organic carbon
rhizosphere soil improved prediction in the root compartment as well as for As and Cu in shoot compartment (Fig. 1). The potential effect of chemical properties associated to soil fertility on the growth of willow and on TE sequestration was quantified and partitioned from the significant predictive models of Fig. 1. No significant predictive model was found for Zn and only one model was obtained for Cd that is in the shoot compartment (Table 4). For almost all significant predictive models, the variance partitioning indicated that most of the explanation was due to the shared effect of water-soluble TE in soils and of soil variables related to fertility (Table 4). Yet, only the Ni models in leaves and shoots and the Cu model in the aboveground biomass were mostly based on soil fertility variables. Each significant regression model obtained is unique as there is not a single soil fertility variable that was systematically retained in all models. The models never involved soil pH, but mostly CEC, some exchangeable cations (K, Na, Mg, or Mn), or SOC. Table 4 further highlights the
fact that despite their significance, the models describing TE in willows have a limited scope with residuals generally equal to or in excess of 32 %.
Discussion The accumulation and storage of trace elements in willow The soils had distinct contamination profiles with the C+ soil being contaminated with PHC and PAH compounds while soil C− was not (Table S1). Moreover, the soils were not contaminated with TE according to the quality criteria used in Québec (MDDEFP 2013), and concentrations were within background levels except for Cu. It therefore suggested that Cu occurrence in these soils was related to the release of organic contaminants by anthropogenic activities, while the source for the other TE was mostly geogenic. Despite the non-
Environ Sci Pollut Res Table 2 Two-way ANOVA results to evaluate the effect of mycorrhizal inoculation treatment (MIT) and of soil types (S) on the water-soluble trace element concentrations in soil components (bulk and rhizosphere) as Response variables
Biomass
As
Cd
Cu
Ni
Pb
Zn
ANOVA effects
MIT S MIT × S MIT S MIT × S MIT S MIT × S MIT S MIT × S MIT S MIT × S MIT S MIT × S MIT S MIT × S
well as on the biomass and trace element concentrations (leaf, shoot, and root) and contents (aboveground and belowground) in willow (Salix purpurea, cultivar Fish Creek) tissues
Soil components
Willow tissues
Bulk
Leaf
Rhizosphere
Shoot
Root
R+ > R−*
Aboveground
Belowground
R+ > R−*
**
C+ > C−
C+ > C−***
C+ > C−***
C+ > C−***
C+ > C−***
C+ > C−***
C+ > C−***
C+ > C−***
* R+ > R−* C+ > C−***
C+ > C−**
C− > C+***
C− > C+**
C− > C+*
C+ > C−*** *
R− > R+** C+ > C−*** * C− > C+***
***
C+ > C−
C+ > C−***
C− > C+***
R− > R+** C+ > C−***
C− > C+***
C+ > C−***
C+ > C−**
R+ > R−* **
C+ > C−
***
C+ > C−
Empty cell: nonsignificant at α=0.10 Aboveground leaf + shoot biomass, belowground root biomass *significant at α=0.10; **significant at α=0.05; ***significant at α=0.01
contamination of soils by TE, willows accumulated substantial TE amounts, notably Cd and Zn (Table 3). For example, a total of 19 out of 23 leaf tissue samples and 3 out of 23 shoot samples had Cd concentrations in excess of the soil threshold for residential use (1.5 mg/kg). Three more leaf samples exceeded the residential threshold for soil Zn (110 mg/kg). Upon harvest, these residues will therefore need to be managed according to existing laws on toxic materials. Several studies used willow in hydroponic, pot, or field experiments to evaluate its potential for TE remediation. The concentrations in willow parts varied considerably between studies for a given TE and our TE data generally fell in the lower range of this spectrum (Bissonnette et al. 2010; De Maria et al. 2011; French et al. 2006; Jensen et al. 2009; Mleczek et al. 2010; Ruttens et al. 2011; Van Nevel et al. 2007; Van Slycken et al. 2013; Vandecasteele et al. 2005; Vangronsveld et al. 2009; Wieshammer et al. 2007; Zhivotovsky et al. 2011). This early performance of the Fish Creek willow cultivar should nonetheless be considered good, in particular, because the data were for the first growing
season and from young willows growing on soils not contaminated with TE. The partition of TE among willow parts, as expressed by concentrations data, was characterized by the preferential sequestration of Cd and Zn in leaves and that of As, Cu, Ni, and Pb in roots (Fig. S2 and Table 3). These observations agreed well with existing literature on TE in plants and in willow (Pulford and Watson 2003; Wieshammer et al. 2007). Computing a bioconcentration ratio (whole plant/soil TE concentration ratio) confirmed the contrast between Cd and Zn versus As, Cu, Ni, and Pb. The latter group of elements had bioconcentration ratios below 0.20, whereas ratios for Cd and Zn ranged from 1.98 to 10.0 and from 0.69 to 1.97, respectively. Maximum ratio values of 23.2 for Cd and 8.2 for Zn were reached in willow leaves. These Cd bioconcentration ratios agreed well with the willow literature reviewed by Dickinson and Pulford (2005), Pulford and Watson (2003), and Wieshammer et al. (2007). The magnitude of the translocation process within willows further discriminated Cd and Zn from the other TE studied. Mean aboveground/belowground
Environ Sci Pollut Res Table 3 Mean (and standard deviation) values for biomass per plant, trace element contents per plant, and trace element concentrations in willow (Salix purpurea, cultivar Fish Creek) tissues Mycorrhizae_soila
R−_C− R−_C+ R+_C− R+_C+
R−_C−
R−_C+
R+_C−
R+_C+
Willow tissues
Biomass
As
Aboveground Belowground Aboveground
g 40.7 (5.05) 5.77 (2.57) 46.1 (8.71)
Content per plant (μg)a 2.85 (0.73) 159 (42.1) 2.15 (0.46) 17.4 (8.65) 5.22 (1.20) 185 (105)
4.84 (2.44) 48.7 (5.66) 6.23 (0.95) 50.0 (9.74) 5.26 (2.13) g 6.33 (1.21) 34.3 (4.37) 5.77 (2.57) 8.40 (0.94) 37.7 (8.23) 4.84 (2.44) 7.36 (0.74) 41.4 (5.41) 6.23 (0.95) 8.56 (2.52) 41.4 (8.26) 5.26 (2.13)
Belowground Aboveground Belowground Aboveground Belowground Leaf Shoot Root Leaf Shoot Root Leaf Shoot Root Leaf Shoot Root
Detection limit of the method (DLM)
–
Cd
Cu
Ni
Pb
Zn
264 (43.0) 72.9 (27.1) 374 (54.3)
71.2 (21.4) 33.7 (9.53) 44.5 (11.5)
31.1 (20.1) 10.9 (3.61) 27.6 (22.2)
4 427 (1 676) 338 (128) 5 392 (1 939)
5.95 (3.14) 16.3 (13.1) 3.80 (0.61) 242 (89.6) 2.64 (0.38) 24.0 (4.68) 5.21 (2.24) 159 (34.7) 5.83 (2.11) 15.1 (5.98) Concentration (mg kg−1) 0.22 (0.06) 7.01 (1.11) 0.04 (0.01) 3.31 (0.73) 0.41 (0.13) 3.03 (0.55) 0.29 (0.08) 6.69 (2.93) 0.08 (0.02) 3.18 (1.32) 1.54 (1.23) 2.95 (1.13) 0.26 (0.06) 10.1 (5.00) 0.05 (0.01) 4.12 (1.55) 0.43 (0.05) 3.88 (0.69) 0.30 (0.12) 5.91 (1.98) 0.06 (0.01) 2.77 (0.94) 1.28 (0.66) 2.99 (0.97)
126 (60.6) 338 (65.3) 94.6 (19.4) 406 (110) 128 (39.8)
28.1 (13.8) 81.3 (15.6) 44.0 (8.07) 44.5 (16.1) 29.8 (8.80)
14.5 (7.37) 28.5 (12.2) 10.5 (1.81) 38.0 (37.1) 16.5 (5.70)
343 (224) 5 742 (1 060) 435 (62.3) 5 421 (1 669) 408 (191)
14.9 (1.86) 4.91 (0.50) 13.2 (2.46) 14.5 (2.77) 6.71 (0.59) 26.5 (3.89) 16.1 (8.52) 5.32 (0.51) 15.2 (1.61) 16.1 (5.25) 6.37 (0.63) 25.2 (3.07)
8.11 (1.30) 0.56 (0.21) 6.46 (2.42) 3.48 (0.72) 0.40 (0.16) 5.90 (1.30) 7.83 (1.59) 0.59 (0.19) 7.06 (0.71) 3.49 (0.78) 0.33 (0.09) 5.96 (1.24)
2.05 (1.81) 0.52 (0.33) 2.06 (0.68) 1.18 (0.31) 0.44 (0.37) 3.25 (1.27) 1.73 (0.86) 0.38 (0.23) 1.72 (0.41) 1.63 (1.79) 0.54 (0.34) 3.22 (0.57)
352 (170) 62.9 (9.92) 60.6 (7.55) 334 (97.4) 66.3 (13.0) 65.8 (15.3) 410 (132) 66.4 (4.88) 70.2 (5.76) 324 (64.4) 62.1 (3.56) 76.0 (12.9)
0.01
0.52
0.92
0.26
2.15
0.04
R− without mycorrhizal inoculation, R+ inoculated with the mycorrhizal fungus Rhizophagus irregularis, C− non-contaminated soil, C+ contaminated soil, aboveground leaf + shoot biomass, belowground root biomass a
The number of samples is five for the R+/C− treatment and six for the three other treatments
concentration ratios were higher than unity for Cd and Zn, while they were equal to or below 0.40 for As, Cu, Ni, and Pb, as documented elsewhere (De Maria et al. 2011; Dickinson et al. 2009; Pulford and Watson 2003). In contrast to concentration data, all TE contents were higher in the aboveground than in the belowground components, except for As (Table 3). This trend was due to the shoot compartment having a biomass four to five times higher than either leaves or roots (Table 3). The shoots act as a key TE sink because they grow every year after coppicing, they constitute an increasing proportion of total biomass over time, and they resist decomposition and the sequestration of TE in these metabolically inactive tissues protects the plant from TE toxicity (Pulford and Watson 2003). Controls on trace elements uptake and sequestration by willow The experimental approach was designed to test the effect of two factors, the mycorrhizal symbiosis and soil properties, on the uptake and sequestration of TE in willow. The influence of
the mycorrhizal symbiosis of willow roots was explored through the inoculation of the Rhizophagus irregularis AM species. Its effect is compared to that of native AM fungi since both R− and R+ treatments were equally mycorrhized for a given soil. Statistical analyses indicated that the effect of Rhizophagus irregularis on TE accumulation in willow was either slight or nonsignificant after one growing season (Table 2). Previous work of Bissonnette et al. (2010) performed on willow (Salix viminalis) and hybrid poplar concluded that inoculation with the Glomus intraradices did not contribute to increase the plant TE extraction capacity compared to soil native AM fungi. This was explained by the fact that the native microorganisms were well adapted to these soil conditions and that they were best suited for the situation. The better performance of the native microorganisms over the inoculated ones was often reported in phytoextraction studies (Kidd et al. 2009; Sessitsch et al. 2013). In the present study, a Rhizophagus irregularis signal did exist even after one growing season. It remained weak but was nonetheless promising for remediation purposes. Indeed, higher shoot biomass, root Zn concentration, and aboveground Cu content were recorded
0.05 0.8
0.010
R2 = 0.43***
−1.6
−1.4
−2.3 −2.5
Bulk R2 = 0.57***
0.002 −1.8
−1.6
−1.4
0.02
log As-H2O (mg kg-1)
0.03
0.04
0.05
As-H2O (mg kg-1)
0.6
Bulk
0.4
log Cd (mg kg-1)
log As (mg) −1.8
log As-H2O (mg kg-1)
log As-H2O (mg kg-1)
0.2
Cd
0.0
−1.4
−1.6
−1.8
Bulk
−0.4
−1.2
log As (mg kg-1) 0.04
As-H2O (mg kg-1)
−2.7
0.03
−1.4
0.4 0.3 0.02
R2 = 0.66***
0.006
Bulk
R2 = 0.72***
Belowground tissues content
As (mg)
Rhizosphere
Aboveground tissues content −2.1
Bulk R2 = 0.15**
0.4
Root tissue concentration log As (mg kg-1)
Shoot tissue concentration −1.0
Leaf tissue concentration
0.2
As
As (mg kg-1)
0.5
Environ Sci Pollut Res
R2 = 0.11* −2.85
−2.65
−2.75
−2.55
0.60 kg-1)
−0.3
0.50
Cu-H2O (mg
0.60 kg-1)
−0.5
log Cu (mg) 0.40
Bulk
R2 = 0.10*
−0.45
−0.35
−0.25
log Cu-H2O (mg kg-1)
6
0.6
R2 = 0.10*
R2 = 0.09*
0.2
0.4
Ni (mg kg-1)
8
0.8
Bulk
4
Ni
Ni ( mg kg-1)
10
Cu-H2O (mg
10
0.50
−0.7
30
R2 = 0.54*** 0.40
Rhizosphere
Rhizosphere
20
Cu (mg kg-1)
6.5
R2 = 0.45***
4.5
Cu
Rhizosphere
5.5
Cu (mg kg-1)
7.5
log Cd-H2O (mg kg-1)
0.12
0.09
0.11
0.13
Rhizosphere
4
R2 = 0.47***
3 2
Pb
Pb (mg kg-1)
5
Ni-H2O (mg kg-1)
Pb (mg)
0.10
Ni-H2O (mg kg-1)
0.004
0.008
0.012
Pb -H2O (mg kg-1)
0.010 0.015 0.020 0.025
0.08
Rhizosphere R2 = 0.15**
0.004
0.008
0.012
Pb-H2O (mg kg-1)
Fig. 1 Simple linear regression of As, Cd, Cu, Ni, and Pb concentrations and contents in willow (Salix purpurea, cultivar Fish Creek) tissues (Yaxis) against their water-soluble concentrations in either the bulk or the rhizosphere soil component (X-axis) after one growing season under controlled conditions. Only the significant (single asterisk, double
asterisks, and triple asterisks indicate significance at α 0.10, 0.05, and 0.01, respectively) relations are presented along with their adjusted regression coefficients (R2). Some variables were log-transformed in order to meet the postulate of the simple linear regression
in the R+ compared to the R− pots (Table 2). Rhizosphere results also indicated a decrease of available TE, mostly Cu, in the R+ willow compared to the non-inoculated in the C+ soil (Fig. S2). Despite lowering the TE availability in the rhizosphere soil, the Rhizophagus irregularis symbiosis slightly increased the TE extraction by willow (Table 2) through mostly an increase in shoot biomass production. The beneficial effect of AM symbiosis on plant growth has been extensively documented and constitutes the most important impact
of AM fungi on plant phytoextraction (Lopez de Andrade et al. 2008; Meharg 2003). The biogeochemical reactions occurring in the rhizosphere strongly depend on the ability of the plant to acquire sufficient nutrients for its growth and to protect itself from chemical or biological aggressions. The fact that the bulk and rhizosphere soil components were chemically similar at the end of our growth experiment (Table 1) might be explained by the combination of two mechanisms. First, the time needed to produce
Environ Sci Pollut Res Table 4 Variance partioning of multiple linear regressions for trace element concentrations and contents in willow (Salix purpurea, cultivar Fish Creek) tissues (Y) between water-soluble trace element Response variable (Y)
Total R2a
concentrations in soils (X1) and chemical properties related to soil fertility (X2). Only the significant relationships are displayed
Significant explanatory variables
%
Partition of total R2b
Residuals R2c
X1 %
X2
Shared fraction
Concentration in leaves As Ni
16* 82***
(+) As-H2O, Mn-BaCl2 (−) SOC, Na-, Mn-BaCl2 (+) Ni-H2O, CEC
15** 10*
16** 82***
15 10
84 18
Concentration in shoots As
75***
(−) CEC (+) As-H2O, SOC, Na-, Mg-BaCl2 (−) Cd-H2O (+) K-BaCl2 (+) Cu-H2O, Na-BaCl2 (+) Ni-H2O, Na-, Mg-BaCl2
72***
69***
67
25
10ns
12*
5
83
45*** 9*
62*** 31**
39 8
32 68
65***
64***
61
32
Cd Cu Ni Concentration in roots As Cu
17* 68*** 32** 68*** 76***
Pb 55*** Content in aboveground tissues As 44*** Cu 25** Content in belowground tissues As 58*** Pb 15ns
(−) K-BaCl2 (+) As-H2O, Mn-BaCl2 (−) CEC (+) Cu-H2O, SOC, Na-BaCl2 (+) Pb-H2O, SOC
54***
72***
50
24
47***
49***
41
45
(+) As-H2O, SOC, Mn-BaCl2 (+) Cu-H2O, SOC
43*** 10*
44*** 25***
43 10
56 75
(+) As-H2O, Na-, Mn-BaCl2 (+) Pb-H2O, Mn-BaCl2
57*** 15**
56*** 11*
55 11
42 85
The symbols + and – refer, respectively, to positive or negative relations between response and explanatory variables; The symbols -H2O and –BaCl2 refer, respectively, to water-soluble or BaCl2-exchangeable element concentrations; SOC=soil organic carbon; CEC=cation-exchange capacity. *
significant at α=0.10; ** significant at α=0.05; *** significant at α=0.01; ns nonsignificant (α>0.10)
a
The total R2 is expressed as the percentage of the total variance of the response variable (Y) explained by the combination of the two groups of explanatory variables (X1 and X2) b The R2 values, expressed as the percentage of the total variance, for each group of explanatory variables is composed of two fractions, a fraction that represents the exclusive group explanation (X1 or X2) and a fraction that represents the shared explanation of the two groups. The R2 of the shared explanation of the two groups is presented but cannot be tested for significance (Legendre and Legendre 2012) c
The residuals R2 indicates the fraction of the relationships not explained by the two groups of explanatory variables
a rhizosphere microenvironment that was significantly different from the bulk soil was not fully reached. During this unique growth season, it can be assumed that the willows acquired their nutrients preferentially from the most available and soluble soil pools. Under such conditions, roots and associated AM fungi will not strongly transform their soil surroundings to acquire nutrients. Since processes induced by roots and AM fungi in the rhizosphere build up over time and because soil TE are mostly in unavailable forms (Hinsinger and Courchesne 2008), it can further be hypothesized that the rhizosphere would be more strongly transformed during subsequent growth seasons, leading to increase soil TE
solubility and uptake by willow as well as to stronger contrasts between the bulk and rhizosphere soil components. Moreover, the capacity of the soil solid phase to supply TE to the soil solution at a rate similar to the TE changes induced in the rhizosphere by roots and microorganisms might also explained the similarity in water-soluble TE concentrations between bulk and rhizosphere soils (Wieshammer et al. 2007). The soils used in this experiment had a moderate buffering capacity (>30 % clay content; ∼30 g/kg of SOC;
RESEARCH ARTICLE
The sequestration of trace elements by willow (Salix purpurea) —which soil properties favor uptake and accumulation? Benoît Cloutier-Hurteau & Marie-Claude Turmel & Catherine Mercier & François Courchesne
Received: 9 October 2013 / Accepted: 9 December 2013 # Springer-Verlag Berlin Heidelberg 2013
Abstract The effect of soil properties on trace element (TE) extraction by the Fish Creek willow cultivar was assessed in a 4-month greenhouse experiment with two contrasted soils and two mycorrhizal treatments (Rhizophagus irregularis and natives). Aboveground tissues represented more than 82 % of the willow biomass and were the major sink for TE. Cadmium and Zn were concentrated in leaves, while As, Cu, Ni, and Pb were mostly found in roots. Willow bioconcentration ratios were below 0.20 for As, Cu, Ni, and Pb and reached 10.0 for Cd and 1.97 for Zn. More significant differences in willow biomass, TE concentrations, and contents were recorded between soil types than between mycorrhizal treatments. A slight significant increase in Cu extraction by willow in symbiosis with Rhizophagus irregularis was observed and was linked to increased shoot biomass. Significant regression models between TE in willow and soil properties were found in leaves (As, Ni), shoots (As, Cd, Cu, Ni), and roots (As, Cu, Pb). Most of the explanation was shared between soil watersoluble TE and fertility variables, indicating that TE phytoextraction is related to soil properties. Managing interactions between TE and major nutrients in soil appeared as a key to improve TE phytoextraction by willows.
Keywords Phytoextraction . Fish Creek willow cultivar . Greenhouse experiment . Mixed contamination . Rhizophagus irregularis . Rhizosphere Responsible editor: Philippe Garrigues Electronic supplementary material The online version of this article (doi:10.1007/s11356-013-2450-y) contains supplementary material, which is available to authorized users. B. Cloutier-Hurteau (*) : M. K for cations and Cl > SO4 ≫ NO3 > PO4 (systematically below DLM) for anions (Table 1). The two-way ANOVA results indicated that the water-soluble major element concentrations significantly differed between bulk soils for K, Na, Ca, Mg, and NO3 and for K, Na, Ca, Mg, and SO4 in the rhizosphere. For all watersoluble major elements, there were no significant effect of the mycorrhizal treatment for both the bulk and the rhizosphere soils (data not shown). The water-soluble TE concentrations fell also within one order of magnitude and were all below 700 μg/kg (Table 1), with abundances following the sequence Cu > Ni > Zn > As ≫ Pb > Cd (Table 1). In both soil components, significant differences in As-H2O, Cu-H2O, Pb-H2O, and Zn-H2O concentrations were noted between the soils, whereas Ni-H2O differed only in the rhizosphere (Table 2). The effect of the mycorrhizal treatment was undetectable, except for Cu-H2O and Pb-H2O in the rhizosphere where lower concentrations were measured in the R− compared to the R+ pots (Table 2). The stronger contrast between soil types than between mycorrhizal treatments was confirmed by the LDA performed on water-soluble TE in the rhizosphere (Figure S1). The C− rhizosphere had higher Cd-H2O and Ni-H2O than the C+ rhizosphere, whereas the inverse trend was seen for As-H2O, Cu-H2O, Pb-H2O, and Zn-H2O (inset of Figure S1 and Table 1). The multivariate approach also showed that the mycorrhizal treatment was better differentiated in the C+ than the C− rhizosphere, mostly because of Cu-H2O. Aboveground plant materials represented more than 82 % of the total willow biomass mostly because of shoots (Table 3).
The individual total recoverable TE concentrations in willow tissues ranged from 0.04 to 3.82 mg/kg for As, from 1.54 to 17.9 mg/kg for Cd, from 4.44 to 32.7 mg/kg for Cu, from 0.20 to 10.4 mg/kg for Ni, from 0.15 to 12.2 mg/kg for Pb, and from 39.8 to 684 mg/kg for Zn. The TE were well contrasted between willow parts with Cd and Zn mostly concentrated in leaves, while As, Cu, Ni, and Pb reached generally maximum levels in roots (Table 3 and Fig. S2). When the total amount of TE sequestered was considered, the order follows: Zn ≫ Cu > Cd ∼ Ni > Pb > As. The sequestration of TE was concentrated in the aboveground biomass, and only As displayed a somewhat balanced sequestration pattern (Table 3). More significant differences in willow biomass, TE concentrations and TE contents were recorded between soil types than between mycorrhizal treatments (Table 2). The soil effect was characterized by significantly higher leaf biomass, shoot As and Cu concentrations, root As, Cu, and Pb levels, aboveground As and Cu contents, and belowground As, Cu, and Pb contents in C+ compared to C− soils. The inverse pattern was noted for Cd and Ni, with higher leaf Cd and Ni concentrations, shoot Ni and above- and belowground Ni contents in the C− than the C+ soil (Tables 2 and 3). The mycorrhizal treatment induced a slight effect on the aboveground and shoot biomass, the aboveground Cu and on root Zn. In all cases, the inoculated treatment had significantly higher values (Tables 2 and 3). Relationships between trace elements in willows and soil properties Simple linear regressions performed between TE in willows and water-soluble TE in either the bulk or the rhizosphere soils were computed and the significant relationships are displayed in Fig. 1. At least one significant relationship between willow tissues and soil concentrations was obtained for each TE, except for Zn. Arsenic and Cu were the TE showing the higher number of significant fit between plant concentrations or contents and water-soluble TE in soils, with most relationships observed for willow shoots and roots (Fig. 1). All the significant relationships, no matter the response variables, were positive with Cd in shoots being the only exception. The predictive models for leaves were weak and barely significant. Several shoot TE models were strongly predicted with adjusted R2 values of 0.72 for As and of 0.45 for Cu (Fig. 1). For roots, the As, Cu, and Pb models performed well with adjusted R2 of 0.66, 0.54, and 0.47, respectively. When the TE contents in either the above- or the belowground materials were considered, few significant models emerged except for As in both cases, Cu in aboveground parts and Pb in the belowground biomass (Fig. 1). However, only As was well described with predictive models explaining 43 and 57 % of the As variance in the above- and belowground biomass of willow, respectively. In general, the bulk and rhizosphere models performed similarly, although using the
Environ Sci Pollut Res Table 1 Mean (and standard deviation) values for soil properties and for water-soluble major and trace element concentrations in the bulk and the rhizosphere soil components at the end of the experiment Mycorrhizae
R−
Soil
C−
Soil component
Bulk
R+ C+ Rhizosphere
Bulk
C− Rhizosphere
Soil properties (n=5 for the R+/C− treatment and n=6 for the other treatments) SOC (g kg−1) 26.6 (2.89) 27.9 (5.18) 34.5 (2.80) 39.7 (4.35) CEC (cmolc kg−1) 30.1 (1.23) 30.2 (2.47) 31.0 (0.88) 31.6 (2.73) Water-soluble major elements (n=3 for all treatments) pH-H2O 7.16 (0.19) 7.27 (0.26) 7.16 (0.19) 7.31 (0.09) EC-H2O (mS cm−1) 0.11 (0.01) 0.11 (0.01) 0.14 (0.02) 0.15 (0.03) DOC-H2O (g kg−1) 0.33 (0.03) 0.29 (0.03) 0.50 (0.09) 0.38 (0.09) K-H2O (mg kg−1) 9.24 (1.81) 10.6 (0.52) 12.4 (1.13) 14.6 (1.35) Na-H2O (mg kg−1) 154 (21.4) 143 (19.8) 209 (12.3) 209 (20.2) Ca-H2O (mg kg−1) 67.9 (3.09) 64.7 (4.24) 76.3 (12.6) 76.8 (14.2) Mg-H2O (mg kg−1) 16.7 (0.54) 17.1 (1.08) 18.5 (3.86) 20.7 (4.17) Cl-H2O (mg kg−1) 103 (7.59) 97.8 (12.2) 96.8 (15.2) 103 (26.6) NO3-H2O (mg kg−1) 11.3 (3.64) 6.53 (9.44) 1.96 (1.87) 0.85 (1.47) SO4-H2O (mg kg−1) 63.0 (0.53) 72.7 (6.83) 94.3 (31.7) 131 (41.7) Water-soluble trace elements (n=5 for R+/C− treatment and n=6 for the other treatments) As-H2O (μg kg−1) 19.0 (1.75) 19.2 (2.68) 46.3 (6.16) 41.9 (7.72) Cd-H2O (μg kg−1) Cu-H2O (μg kg−1) Ni-H2O (μg kg−1) Pb-H2O (μg kg−1) Zn-H2O (μg kg−1)
1.81 (0.24) 428 (35.2) 107 (8.87) 5.67 (1.24) 47.8 (14.9)
1.90 (0.40) 427 (38.9) 116 (13.0) 7.79 (1.03) 57.1 (16.2)
2.14 (0.58) 562 (54.9) 109 (18.4) 14.1 (3.30) 64.3 (13.4)
1.65 (0.67) 586 (41.5) 103 (16.0) 12.5 (2.53) 67.9 (9.11)
C+
Bulk
Rhizosphere
Bulk
Rhizosphere
26.6 (3.03) 31.0 (0.99)
28.8 (1.76) 31.9 (1.99)
35.3 (1.79) 30.8 (0.73)
38.2 (2.09) 33.9 (2.91)
7.38 (0.02) 0.12 (0.02) 0.34 (0.05) 8.78 (0.81) 170 (30.3) 73.0 (7.10) 18.0 (2.49) 117 (25.9) 10.4 (4.30) 74.0 (20.7)
7.22 (0.12) 0.11 (0.01) 0.30 (0.03) 10.1 (1.34) 143 (7.92) 65.0 (0.93) 16.8 (0.31) 103 (21.2) 5.18 (3.57) 80.8 (10.8)
7.30 (0.06) 0.13 (0.01) 0.43 (0.06) 14.3 (2.50) 190 (16.4) 85.0 (2.26) 21.6 (2.16) 91.1 (21.0) 6.34 (1.86) 91.7 (39.5)
7.08 (0.19) 0.14 (0.01) 0.33 (0.05) 13.3 (1.53) 194 (5.89) 75.9 (9.66) 18.8 (2.08) 94.9 (4.19) 7.45 (4.46) 122 (20.8)
18.2 (2.83)
18.1 (1.72)
41.3 (6.94)
37.9 (5.17)
1.90 (0.39) 430 (53.4) 106 (14.0) 6.66 (1.22) 44.7 (10.5)
1.85 (0.27) 418 (26.5) 110 (13.6) 5.96 (1.83) 45.4 (12.5)
1.80 (0.39) 499 (29.2) 96.9 (13.9) 14.6 (4.31) 63.1 (18.0)
1.46 (0.27) 497 (79.7) 90.7 (9.63) 11.1 (1.67) 72.6 (18.2)
Variables followed by the suffix “-H2O” were measured in water extract (1 :10 soil-to-water ratio). The detection limit of the method of the trace elements measured in the water-soluble extract were set to 0.75, 0.34, 9.15, 11.0, 1.45, and 9.61 μg kg−1 , respectively, for As, Cd, Cu, Ni, Pb, and Zn R− without mycorrhizal inoculation, R+ inoculated with the mycorrhizal fungus Rhizophagus irregularis, C− non-contaminated soil, C+ contaminated soil, SOC soil organic carbon, CEC cation-exchange capacity, EC electrical conductivity, DOC dissolved organic carbon
rhizosphere soil improved prediction in the root compartment as well as for As and Cu in shoot compartment (Fig. 1). The potential effect of chemical properties associated to soil fertility on the growth of willow and on TE sequestration was quantified and partitioned from the significant predictive models of Fig. 1. No significant predictive model was found for Zn and only one model was obtained for Cd that is in the shoot compartment (Table 4). For almost all significant predictive models, the variance partitioning indicated that most of the explanation was due to the shared effect of water-soluble TE in soils and of soil variables related to fertility (Table 4). Yet, only the Ni models in leaves and shoots and the Cu model in the aboveground biomass were mostly based on soil fertility variables. Each significant regression model obtained is unique as there is not a single soil fertility variable that was systematically retained in all models. The models never involved soil pH, but mostly CEC, some exchangeable cations (K, Na, Mg, or Mn), or SOC. Table 4 further highlights the
fact that despite their significance, the models describing TE in willows have a limited scope with residuals generally equal to or in excess of 32 %.
Discussion The accumulation and storage of trace elements in willow The soils had distinct contamination profiles with the C+ soil being contaminated with PHC and PAH compounds while soil C− was not (Table S1). Moreover, the soils were not contaminated with TE according to the quality criteria used in Québec (MDDEFP 2013), and concentrations were within background levels except for Cu. It therefore suggested that Cu occurrence in these soils was related to the release of organic contaminants by anthropogenic activities, while the source for the other TE was mostly geogenic. Despite the non-
Environ Sci Pollut Res Table 2 Two-way ANOVA results to evaluate the effect of mycorrhizal inoculation treatment (MIT) and of soil types (S) on the water-soluble trace element concentrations in soil components (bulk and rhizosphere) as Response variables
Biomass
As
Cd
Cu
Ni
Pb
Zn
ANOVA effects
MIT S MIT × S MIT S MIT × S MIT S MIT × S MIT S MIT × S MIT S MIT × S MIT S MIT × S MIT S MIT × S
well as on the biomass and trace element concentrations (leaf, shoot, and root) and contents (aboveground and belowground) in willow (Salix purpurea, cultivar Fish Creek) tissues
Soil components
Willow tissues
Bulk
Leaf
Rhizosphere
Shoot
Root
R+ > R−*
Aboveground
Belowground
R+ > R−*
**
C+ > C−
C+ > C−***
C+ > C−***
C+ > C−***
C+ > C−***
C+ > C−***
C+ > C−***
C+ > C−***
* R+ > R−* C+ > C−***
C+ > C−**
C− > C+***
C− > C+**
C− > C+*
C+ > C−*** *
R− > R+** C+ > C−*** * C− > C+***
***
C+ > C−
C+ > C−***
C− > C+***
R− > R+** C+ > C−***
C− > C+***
C+ > C−***
C+ > C−**
R+ > R−* **
C+ > C−
***
C+ > C−
Empty cell: nonsignificant at α=0.10 Aboveground leaf + shoot biomass, belowground root biomass *significant at α=0.10; **significant at α=0.05; ***significant at α=0.01
contamination of soils by TE, willows accumulated substantial TE amounts, notably Cd and Zn (Table 3). For example, a total of 19 out of 23 leaf tissue samples and 3 out of 23 shoot samples had Cd concentrations in excess of the soil threshold for residential use (1.5 mg/kg). Three more leaf samples exceeded the residential threshold for soil Zn (110 mg/kg). Upon harvest, these residues will therefore need to be managed according to existing laws on toxic materials. Several studies used willow in hydroponic, pot, or field experiments to evaluate its potential for TE remediation. The concentrations in willow parts varied considerably between studies for a given TE and our TE data generally fell in the lower range of this spectrum (Bissonnette et al. 2010; De Maria et al. 2011; French et al. 2006; Jensen et al. 2009; Mleczek et al. 2010; Ruttens et al. 2011; Van Nevel et al. 2007; Van Slycken et al. 2013; Vandecasteele et al. 2005; Vangronsveld et al. 2009; Wieshammer et al. 2007; Zhivotovsky et al. 2011). This early performance of the Fish Creek willow cultivar should nonetheless be considered good, in particular, because the data were for the first growing
season and from young willows growing on soils not contaminated with TE. The partition of TE among willow parts, as expressed by concentrations data, was characterized by the preferential sequestration of Cd and Zn in leaves and that of As, Cu, Ni, and Pb in roots (Fig. S2 and Table 3). These observations agreed well with existing literature on TE in plants and in willow (Pulford and Watson 2003; Wieshammer et al. 2007). Computing a bioconcentration ratio (whole plant/soil TE concentration ratio) confirmed the contrast between Cd and Zn versus As, Cu, Ni, and Pb. The latter group of elements had bioconcentration ratios below 0.20, whereas ratios for Cd and Zn ranged from 1.98 to 10.0 and from 0.69 to 1.97, respectively. Maximum ratio values of 23.2 for Cd and 8.2 for Zn were reached in willow leaves. These Cd bioconcentration ratios agreed well with the willow literature reviewed by Dickinson and Pulford (2005), Pulford and Watson (2003), and Wieshammer et al. (2007). The magnitude of the translocation process within willows further discriminated Cd and Zn from the other TE studied. Mean aboveground/belowground
Environ Sci Pollut Res Table 3 Mean (and standard deviation) values for biomass per plant, trace element contents per plant, and trace element concentrations in willow (Salix purpurea, cultivar Fish Creek) tissues Mycorrhizae_soila
R−_C− R−_C+ R+_C− R+_C+
R−_C−
R−_C+
R+_C−
R+_C+
Willow tissues
Biomass
As
Aboveground Belowground Aboveground
g 40.7 (5.05) 5.77 (2.57) 46.1 (8.71)
Content per plant (μg)a 2.85 (0.73) 159 (42.1) 2.15 (0.46) 17.4 (8.65) 5.22 (1.20) 185 (105)
4.84 (2.44) 48.7 (5.66) 6.23 (0.95) 50.0 (9.74) 5.26 (2.13) g 6.33 (1.21) 34.3 (4.37) 5.77 (2.57) 8.40 (0.94) 37.7 (8.23) 4.84 (2.44) 7.36 (0.74) 41.4 (5.41) 6.23 (0.95) 8.56 (2.52) 41.4 (8.26) 5.26 (2.13)
Belowground Aboveground Belowground Aboveground Belowground Leaf Shoot Root Leaf Shoot Root Leaf Shoot Root Leaf Shoot Root
Detection limit of the method (DLM)
–
Cd
Cu
Ni
Pb
Zn
264 (43.0) 72.9 (27.1) 374 (54.3)
71.2 (21.4) 33.7 (9.53) 44.5 (11.5)
31.1 (20.1) 10.9 (3.61) 27.6 (22.2)
4 427 (1 676) 338 (128) 5 392 (1 939)
5.95 (3.14) 16.3 (13.1) 3.80 (0.61) 242 (89.6) 2.64 (0.38) 24.0 (4.68) 5.21 (2.24) 159 (34.7) 5.83 (2.11) 15.1 (5.98) Concentration (mg kg−1) 0.22 (0.06) 7.01 (1.11) 0.04 (0.01) 3.31 (0.73) 0.41 (0.13) 3.03 (0.55) 0.29 (0.08) 6.69 (2.93) 0.08 (0.02) 3.18 (1.32) 1.54 (1.23) 2.95 (1.13) 0.26 (0.06) 10.1 (5.00) 0.05 (0.01) 4.12 (1.55) 0.43 (0.05) 3.88 (0.69) 0.30 (0.12) 5.91 (1.98) 0.06 (0.01) 2.77 (0.94) 1.28 (0.66) 2.99 (0.97)
126 (60.6) 338 (65.3) 94.6 (19.4) 406 (110) 128 (39.8)
28.1 (13.8) 81.3 (15.6) 44.0 (8.07) 44.5 (16.1) 29.8 (8.80)
14.5 (7.37) 28.5 (12.2) 10.5 (1.81) 38.0 (37.1) 16.5 (5.70)
343 (224) 5 742 (1 060) 435 (62.3) 5 421 (1 669) 408 (191)
14.9 (1.86) 4.91 (0.50) 13.2 (2.46) 14.5 (2.77) 6.71 (0.59) 26.5 (3.89) 16.1 (8.52) 5.32 (0.51) 15.2 (1.61) 16.1 (5.25) 6.37 (0.63) 25.2 (3.07)
8.11 (1.30) 0.56 (0.21) 6.46 (2.42) 3.48 (0.72) 0.40 (0.16) 5.90 (1.30) 7.83 (1.59) 0.59 (0.19) 7.06 (0.71) 3.49 (0.78) 0.33 (0.09) 5.96 (1.24)
2.05 (1.81) 0.52 (0.33) 2.06 (0.68) 1.18 (0.31) 0.44 (0.37) 3.25 (1.27) 1.73 (0.86) 0.38 (0.23) 1.72 (0.41) 1.63 (1.79) 0.54 (0.34) 3.22 (0.57)
352 (170) 62.9 (9.92) 60.6 (7.55) 334 (97.4) 66.3 (13.0) 65.8 (15.3) 410 (132) 66.4 (4.88) 70.2 (5.76) 324 (64.4) 62.1 (3.56) 76.0 (12.9)
0.01
0.52
0.92
0.26
2.15
0.04
R− without mycorrhizal inoculation, R+ inoculated with the mycorrhizal fungus Rhizophagus irregularis, C− non-contaminated soil, C+ contaminated soil, aboveground leaf + shoot biomass, belowground root biomass a
The number of samples is five for the R+/C− treatment and six for the three other treatments
concentration ratios were higher than unity for Cd and Zn, while they were equal to or below 0.40 for As, Cu, Ni, and Pb, as documented elsewhere (De Maria et al. 2011; Dickinson et al. 2009; Pulford and Watson 2003). In contrast to concentration data, all TE contents were higher in the aboveground than in the belowground components, except for As (Table 3). This trend was due to the shoot compartment having a biomass four to five times higher than either leaves or roots (Table 3). The shoots act as a key TE sink because they grow every year after coppicing, they constitute an increasing proportion of total biomass over time, and they resist decomposition and the sequestration of TE in these metabolically inactive tissues protects the plant from TE toxicity (Pulford and Watson 2003). Controls on trace elements uptake and sequestration by willow The experimental approach was designed to test the effect of two factors, the mycorrhizal symbiosis and soil properties, on the uptake and sequestration of TE in willow. The influence of
the mycorrhizal symbiosis of willow roots was explored through the inoculation of the Rhizophagus irregularis AM species. Its effect is compared to that of native AM fungi since both R− and R+ treatments were equally mycorrhized for a given soil. Statistical analyses indicated that the effect of Rhizophagus irregularis on TE accumulation in willow was either slight or nonsignificant after one growing season (Table 2). Previous work of Bissonnette et al. (2010) performed on willow (Salix viminalis) and hybrid poplar concluded that inoculation with the Glomus intraradices did not contribute to increase the plant TE extraction capacity compared to soil native AM fungi. This was explained by the fact that the native microorganisms were well adapted to these soil conditions and that they were best suited for the situation. The better performance of the native microorganisms over the inoculated ones was often reported in phytoextraction studies (Kidd et al. 2009; Sessitsch et al. 2013). In the present study, a Rhizophagus irregularis signal did exist even after one growing season. It remained weak but was nonetheless promising for remediation purposes. Indeed, higher shoot biomass, root Zn concentration, and aboveground Cu content were recorded
0.05 0.8
0.010
R2 = 0.43***
−1.6
−1.4
−2.3 −2.5
Bulk R2 = 0.57***
0.002 −1.8
−1.6
−1.4
0.02
log As-H2O (mg kg-1)
0.03
0.04
0.05
As-H2O (mg kg-1)
0.6
Bulk
0.4
log Cd (mg kg-1)
log As (mg) −1.8
log As-H2O (mg kg-1)
log As-H2O (mg kg-1)
0.2
Cd
0.0
−1.4
−1.6
−1.8
Bulk
−0.4
−1.2
log As (mg kg-1) 0.04
As-H2O (mg kg-1)
−2.7
0.03
−1.4
0.4 0.3 0.02
R2 = 0.66***
0.006
Bulk
R2 = 0.72***
Belowground tissues content
As (mg)
Rhizosphere
Aboveground tissues content −2.1
Bulk R2 = 0.15**
0.4
Root tissue concentration log As (mg kg-1)
Shoot tissue concentration −1.0
Leaf tissue concentration
0.2
As
As (mg kg-1)
0.5
Environ Sci Pollut Res
R2 = 0.11* −2.85
−2.65
−2.75
−2.55
0.60 kg-1)
−0.3
0.50
Cu-H2O (mg
0.60 kg-1)
−0.5
log Cu (mg) 0.40
Bulk
R2 = 0.10*
−0.45
−0.35
−0.25
log Cu-H2O (mg kg-1)
6
0.6
R2 = 0.10*
R2 = 0.09*
0.2
0.4
Ni (mg kg-1)
8
0.8
Bulk
4
Ni
Ni ( mg kg-1)
10
Cu-H2O (mg
10
0.50
−0.7
30
R2 = 0.54*** 0.40
Rhizosphere
Rhizosphere
20
Cu (mg kg-1)
6.5
R2 = 0.45***
4.5
Cu
Rhizosphere
5.5
Cu (mg kg-1)
7.5
log Cd-H2O (mg kg-1)
0.12
0.09
0.11
0.13
Rhizosphere
4
R2 = 0.47***
3 2
Pb
Pb (mg kg-1)
5
Ni-H2O (mg kg-1)
Pb (mg)
0.10
Ni-H2O (mg kg-1)
0.004
0.008
0.012
Pb -H2O (mg kg-1)
0.010 0.015 0.020 0.025
0.08
Rhizosphere R2 = 0.15**
0.004
0.008
0.012
Pb-H2O (mg kg-1)
Fig. 1 Simple linear regression of As, Cd, Cu, Ni, and Pb concentrations and contents in willow (Salix purpurea, cultivar Fish Creek) tissues (Yaxis) against their water-soluble concentrations in either the bulk or the rhizosphere soil component (X-axis) after one growing season under controlled conditions. Only the significant (single asterisk, double
asterisks, and triple asterisks indicate significance at α 0.10, 0.05, and 0.01, respectively) relations are presented along with their adjusted regression coefficients (R2). Some variables were log-transformed in order to meet the postulate of the simple linear regression
in the R+ compared to the R− pots (Table 2). Rhizosphere results also indicated a decrease of available TE, mostly Cu, in the R+ willow compared to the non-inoculated in the C+ soil (Fig. S2). Despite lowering the TE availability in the rhizosphere soil, the Rhizophagus irregularis symbiosis slightly increased the TE extraction by willow (Table 2) through mostly an increase in shoot biomass production. The beneficial effect of AM symbiosis on plant growth has been extensively documented and constitutes the most important impact
of AM fungi on plant phytoextraction (Lopez de Andrade et al. 2008; Meharg 2003). The biogeochemical reactions occurring in the rhizosphere strongly depend on the ability of the plant to acquire sufficient nutrients for its growth and to protect itself from chemical or biological aggressions. The fact that the bulk and rhizosphere soil components were chemically similar at the end of our growth experiment (Table 1) might be explained by the combination of two mechanisms. First, the time needed to produce
Environ Sci Pollut Res Table 4 Variance partioning of multiple linear regressions for trace element concentrations and contents in willow (Salix purpurea, cultivar Fish Creek) tissues (Y) between water-soluble trace element Response variable (Y)
Total R2a
concentrations in soils (X1) and chemical properties related to soil fertility (X2). Only the significant relationships are displayed
Significant explanatory variables
%
Partition of total R2b
Residuals R2c
X1 %
X2
Shared fraction
Concentration in leaves As Ni
16* 82***
(+) As-H2O, Mn-BaCl2 (−) SOC, Na-, Mn-BaCl2 (+) Ni-H2O, CEC
15** 10*
16** 82***
15 10
84 18
Concentration in shoots As
75***
(−) CEC (+) As-H2O, SOC, Na-, Mg-BaCl2 (−) Cd-H2O (+) K-BaCl2 (+) Cu-H2O, Na-BaCl2 (+) Ni-H2O, Na-, Mg-BaCl2
72***
69***
67
25
10ns
12*
5
83
45*** 9*
62*** 31**
39 8
32 68
65***
64***
61
32
Cd Cu Ni Concentration in roots As Cu
17* 68*** 32** 68*** 76***
Pb 55*** Content in aboveground tissues As 44*** Cu 25** Content in belowground tissues As 58*** Pb 15ns
(−) K-BaCl2 (+) As-H2O, Mn-BaCl2 (−) CEC (+) Cu-H2O, SOC, Na-BaCl2 (+) Pb-H2O, SOC
54***
72***
50
24
47***
49***
41
45
(+) As-H2O, SOC, Mn-BaCl2 (+) Cu-H2O, SOC
43*** 10*
44*** 25***
43 10
56 75
(+) As-H2O, Na-, Mn-BaCl2 (+) Pb-H2O, Mn-BaCl2
57*** 15**
56*** 11*
55 11
42 85
The symbols + and – refer, respectively, to positive or negative relations between response and explanatory variables; The symbols -H2O and –BaCl2 refer, respectively, to water-soluble or BaCl2-exchangeable element concentrations; SOC=soil organic carbon; CEC=cation-exchange capacity. *
significant at α=0.10; ** significant at α=0.05; *** significant at α=0.01; ns nonsignificant (α>0.10)
a
The total R2 is expressed as the percentage of the total variance of the response variable (Y) explained by the combination of the two groups of explanatory variables (X1 and X2) b The R2 values, expressed as the percentage of the total variance, for each group of explanatory variables is composed of two fractions, a fraction that represents the exclusive group explanation (X1 or X2) and a fraction that represents the shared explanation of the two groups. The R2 of the shared explanation of the two groups is presented but cannot be tested for significance (Legendre and Legendre 2012) c
The residuals R2 indicates the fraction of the relationships not explained by the two groups of explanatory variables
a rhizosphere microenvironment that was significantly different from the bulk soil was not fully reached. During this unique growth season, it can be assumed that the willows acquired their nutrients preferentially from the most available and soluble soil pools. Under such conditions, roots and associated AM fungi will not strongly transform their soil surroundings to acquire nutrients. Since processes induced by roots and AM fungi in the rhizosphere build up over time and because soil TE are mostly in unavailable forms (Hinsinger and Courchesne 2008), it can further be hypothesized that the rhizosphere would be more strongly transformed during subsequent growth seasons, leading to increase soil TE
solubility and uptake by willow as well as to stronger contrasts between the bulk and rhizosphere soil components. Moreover, the capacity of the soil solid phase to supply TE to the soil solution at a rate similar to the TE changes induced in the rhizosphere by roots and microorganisms might also explained the similarity in water-soluble TE concentrations between bulk and rhizosphere soils (Wieshammer et al. 2007). The soils used in this experiment had a moderate buffering capacity (>30 % clay content; ∼30 g/kg of SOC;
