Environmental Science and Pollution Research

6 downloads 0 Views 1MB Size Report
Jun 24, 2014 - Carson et al. (2000). ...... Sophie Raous for providing the purified smectite and André ... Carson KC, Meyer JM, Dilworth MJ (2000) Hydroxamate ...
Environmental Science and Pollution Research Effect of pyoverdine supply on cadmium and nickel complexation and phytoavailability in hydroponics --Manuscript Draft-Manuscript Number:

ESPR-D-14-01147R1

Full Title:

Effect of pyoverdine supply on cadmium and nickel complexation and phytoavailability in hydroponics

Article Type:

Research Article

Corresponding Author:

Jean-Yves Cornu FRANCE

Corresponding Author Secondary Information: Corresponding Author's Institution: Corresponding Author's Secondary Institution: First Author:

Claire Ferret

First Author Secondary Information: Order of Authors:

Claire Ferret Jean-Yves Cornu Mourad Elhabiri Thibault Sterckeman Armelle Braud Karine Jezequel Marc Lollier Thierry Lebeau Isabelle Schalk Valérie Geoffroy

Order of Authors Secondary Information: Abstract:

Siderophores are chelators with a high selectivity for Fe(III) and a good affinity for divalent metals, including Cd(II) and Ni(II). Inoculation with siderophore-producing bacteria (SPB) has thus been proposed as an alternative to chelator supply in phytoremediation. Accurate assessments of the potential of this association require a dissection of the interaction of siderophores with metals at the soil-root interface. This study focuses on pyoverdine (Pvd), the main siderophore produced by Pseudomonas aeruginosa. We first assessed the ability of Pvd to coordinate Ni(II). The stability constant of Pvd-Ni(II) (log KL'Ni = 10.9) was found to be higher than that of Pvd-Cd(II) (log KL'Cd = 8.2). We then investigated the effect of a direct supply of Pvd on the mobilization, speciation and phytoavailability of Cd and Ni in hydroponics. When supplied at a concentration of 50 µM, Pvd selectively promoted Ni mobilization from smectite. It decreased plant Ni and Cd contents and the free ionic fractions of these two metals, consistent with the free ion activity model. Pvd had a more pronounced effect for Ni than for Cd, as predicted from its coordination properties. Inoculation with P. aeruginosa had a similar effect on Ni phytoavailability to the direct supply of Pvd.

Response to Reviewers:

See attachment

Powered by Editorial Manager® and ProduXion Manager® from Aries Systems Corporation

Authors' Response to Reviewers' Comments Click here to download Authors' Response to Reviewers' Comments: Cover letter.docx

Dr J.Y. CORNU INRA -UMR 1391 ISPA Biogeochemistry of trace elements group 71 avenue Edouard Bourlaux 33882 Villenave d’Ornon cedex FRANCE Tel: +33 5 57 12 25 22 [email protected] Bordeaux, 24th June 2014

Dear Editor-in-Chief,

Thanks for having taken into consideration our manuscript. Please find attached the revised version. The suggestions and helpful comments of the two reviewers helped us improve our paper. Specific responses to the 1st reviewer 1) 2) 3) 4)

The English language has been revised The formatting requirements have been followed cautiously Pages in the manuscript have been numbered in the middle The term MES has been explained (Line 200), the units have been harmonized (e.g. Lines 156, 172 and 199) and the term “speciation” has been defined (Line 272) 5) Throughout the manuscript, Ni(II) has been changed to Ni and Cd(II) to Cd, except when their divalent nature needed to be recalled (e.g. Lines 64, 138, 165) Specific responses to the 2nd reviewer 1) As suggested, log K has been used (instead of K) for equilibrium constants 2) "Pyoverdine concentration was assessed in solution by measuring absorbance at 400 nm". Was it only this wavelength that was used or a range of wavelengths? Please supply a reference for this method. In a previous paper by Cornu et al. (2014) the wavelength 380nm was used, why was this wavelength not used? Was this only used for the measurement of the unprotonated ligand or protonated or complexed? Could different forms be distinguished? This needs to be clarified. If so at this point it would be good to refer to F2”

The UV-visible absorption spectrum of pyoverdine in the absence of metal (apo pyoverdine) is pH-dependent because of the protonation of the catechol moiety. At pH 5 (pH of the pyoverdine solution in Cornu et al. 2014), the maximum of absorbance is centered at 380 and 360 nm. At neutral pH (pH of the pyoverdine solution in this study), the maximum of absorbance of apo Pvd is bathochromically shifted to 400 nm. This was shown previously in Albrecht-Gary et al. (1994). This reference has been added to the manuscript (Line 118). The pyoverdine-Fe complex has a maximum of absorbance at 400 nm and 450 nm, whatever the pH (Albrecht-Gary et al. 1994). The pyoverdine-Ni complex has a different spectrum with a maximum of absorbance only at 450 nm (Online Resource 2). The wavelength 400 nm was thus used to measure the concentration of apo Pvd and Pvd-Fe which are the two main forms of Pvd in our experiments. 3) Please give a reference for this statement The major metal binding sites of pyoverdine are the catecholate and the two hydroxamates groups which are by far more efficient chelators of a wide range of metal cations. It has been also described in the literature that the nature of the acyl chain does not affect iron transport (and by consequence its metal chelating capacity), and these variant molecules rather represent isoforms of an otherwise identical pyoverdine. As suggested, a reference (Meyer et al. 2002) has been added to the statement “metal coordination properties are supposed to be similar between the three isoforms of Pvd since the succinic, succinamide and -cetoglutaric moieties are not involved in the metal coordination” (Line 124). Meyer JM, Geoffroy VA, Baida N, Gardan L, Izard D, Lemanceau P, Achouak W, Palleroni NJ (2002) Appl Environ Microbiol 68:2745-2753 4) “I find this section not easy to understand and leaves me with some questions despite reading the cited paper. Presumably Pvd was measured by spectrophotometric adsorption. So was only LH4 measured at 400nm or where other species measured? Was an indicator used as in Elhabiri 2004? It is not clear to me exactly which parameters were measured for the purpose of calculating the stability constants. This section needs to be made a lot clearer” We apologize if this section was not clear enough. Pyoverdine is a valuable chromophore (primarily the catecholate unit) whose absorption properties are affected by its protonation state and its metal loading. As stated previously, at neutral pH pyoverdine is characterized by a main absorption band centered at 360 and 380 nm. The decrease of the pH induces a shift of the -* transitions (catecholate unit) to 400 nm. Consequently, this absorption band that is pH sensitive is a valuable probe of the protonation state of pyoverdine. This has been thoroughly discussed in Albrecht-Gary et al. (1994). Albrecht-Gary AM, Blanc S, Rochel N, Ocacktan AZ, Abdallah MA (1994) Bacterial iron transport: coordination properties of pyoverdin PaA, a peptidic siderophore of Pseudomonas aeruginosa. Inorg Chem 33:6391-6402 Similarly, pyoverdine displays spectral properties which are sensitive to the metal uptake since the ionizable binding sites, namely the catechol and the two hydroxamic acids, have to

be deprotonated in the final complexes. Therefore, the absorption characteristics of pyoverdine as a function of pH (not only at 400 nm but on the whole spectral window from 220 nm - 600 nm; see Online Resource 2 and the above mentioned reference) were used to determine both the stability constants of the pyoverdine metal complexes and the spectral properties of each of the protonated metallic species. Absorption spectra were recorded as a function of pH (in a large pH range spanning from ~ 2 to ~ 11) and the corresponding spectral and pHmetric datasets were processed statistically with the Specfit program. The Specfit program adjusts the stability constants and the corresponding molar extinction coefficients (M-1 cm-1) of the species formed at equilibrium. Specfit software uses factor analysis to reduce the absorbance matrix and to extract the eigenvalues prior to the multiwavelength fit of the reduced dataset, according to the Marquardt algorithm. For the sake of clarity, we have accordingly modified the corresponding section (Lines 131-146). 5) The sentence has been reworded as suggested (Lines 162-165). 6) You don't mention Pvd in this section although you should be measuring the amount of free Cd when in solution with Pvd. Please clarify. The following sentence has been added to the manuscript (Lines 177-179) to clarify the protocol: “This determination was carried out with the background solution for Exp. 2 (725 µM Ca(NO3)2, 2 mM MES, 1 µM Cd(NO3)2), to which 50 µM Pvd was added, at pH values of 6.0, 6.4 and 7.0” 7) In the text and in figure 1 and figure F4 for the species that will form for Ni and Cd is suggested to be LMH2 at higher pH. However it seems to me that the stability constants for these complexes were not measured. Therefore how can the model predict they were formed? Please clarify. We again apologize if this was not clear enough. LMH2 actually corresponds to the Ni or Cd complexes with Pvd for which the arginine and succinate moieties have been considered (i.e. protonation and charge states relevant for our discussion), while in Table 1, L’ designates the pyoverdine for which the ionizable sites of the arginine and succinate moieties have been omitted. It is noteworthy that since the two groups do not participate to the metal binding, their pKa values (12.2 and 4.8) are not affected by metal complexation, as demonstrated in the case of Fe(III) complexes by Albrecht-Gary et al. (1994). For the sake of clarity, we have therefore modified the captions of Figure 1 and Figure F4 (now called Online Resource 4). 8) As suggested, the concentration of Ni and Cd used for speciation calculations (i.e. 1 µM) was specified in the title of Figure F4 (now called Online Resource 4). Hoping it will now be suitable for publication in Environmental Science and Pollution Research. Sincerely yours,

Dr Jean-Yves CORNU

Manuscript Click here to download Manuscript: Manuscript.docx Click here to view linked References

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

1

Effect of pyoverdine supply on cadmium and nickel complexation and phytoavailability

2

in hydroponics

3

C. Ferret1*, J.Y. Cornu2,3,4*, M. Elhabiri5, T. Sterckeman6, A. Braud7, K. Jezequel4, M.

4

Lollier4, T. Lebeau4,7, I. J. Schalk1 and V. A. Geoffroy1

5 6

*

These authors contributed equally to this work

1

UMR 7242, Université de Strasbourg-CNRS, ESBS, 300 Boulevard Sébastien Brant, F-

7 8 9 10 11 12

67412, Illkirch cedex, Strasbourg, France 2

INRA, UMR 1391 ISPA, F-33140 Villenave d’Ornon, France

3

Bordeaux Sciences Agro, UMR 1391 ISPA, F-33170 Gradignan, France

4

Université de Haute Alsace, EA 3991 LVBE (Laboratoire Vigne Biotechnologies

13

Environnement), Equipe Dépollution Biologique des Sols, BP 50568, 68008 Colmar cedex,

14

France.

15 16 17 18 19 20

5

CNRS-Université de Strasbourg, UMR 7509 Laboratoire de Chimie Moléculaire, Equipe

Chimie Bioorganique et Médicinale, 25 rue Becquerel, 67200 Strasbourg, France 6

Laboratoire Sols et Environnement, Université de Lorraine - INRA, 2 avenue de la Forêt de

Haye, TSA 40602, 54518 Vandœuvre-lès-Nancy cedex, France. 7

LUNAM, LPGN UMR 6112 CNRS, Université de Nantes, 2 rue de la Houssinière, BP

92208, 44322 Nantes cedex 3, France.

21 22

Corresponding author: Jean-Yves Cornu

23

Tel: 33 5 57 12 25 22

24

Fax: 33 5 57 12 25 15

25

E-mail: [email protected]

1

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

26

Abstract

27

Siderophores are chelators with a high selectivity for Fe(III) and a good affinity for divalent

28

metals, including Cd(II) and Ni(II). Inoculation with siderophore-producing bacteria (SPB)

29

has thus been proposed as an alternative to chelator supply in phytoremediation. Accurate

30

assessments of the potential of this association require a dissection of the interaction of

31

siderophores with metals at the soil-root interface. This study focuses on pyoverdine (Pvd),

32

the main siderophore produced by Pseudomonas aeruginosa. We first assessed the ability of

33

Pvd to coordinate Ni(II). The stability constant of Pvd-Ni(II) (log KL’Ni = 10.9) was found to

34

be higher than that of Pvd-Cd(II) (log KL’Cd = 8.2). We then investigated the effect of a direct

35

supply of Pvd on the mobilization, speciation and phytoavailability of Cd and Ni in

36

hydroponics. When supplied at a concentration of 50 µM, Pvd selectively promoted Ni

37

mobilization from smectite. It decreased plant Ni and Cd contents and the free ionic fractions

38

of these two metals, consistent with the free ion activity model. Pvd had a more pronounced

39

effect for Ni than for Cd, as predicted from its coordination properties. Inoculation with P.

40

aeruginosa had a similar effect on Ni phytoavailability to the direct supply of Pvd.

41 42

Keywords

43

Bacterial siderophore, divalent metals, phytoremediation, pyoverdine, smectite, speciation

44 45

Introduction

46

Methods using living organisms (microorganisms, earthworms, plants) to rehabilitate

47

contaminated soils are gaining increasing attention, as a sustainable approach to this problem.

48

One such method, phytoextraction, is currently the only way to clean up metal-contaminated

49

soils in situ. However, improvements in its efficacy and reliability are required, to meet the

50

requirements of planners and users. The main drawback of this approach is the long time

2

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

51

required for clean-up (Baker et al. 2000), principally due to the low phytoavailability of

52

metals in the soil. Synthetic chelators (e.g. EDTA, EDDS) have been shown to enhance the

53

phytoavailability of metals (Huang et al. 1997; Luo et al. 2005). However, so-called “chelant-

54

assisted” phytoextraction is also subject to several limitations. At the doses required to

55

enhance metal phytoavailabity (mmol kg-1 soil), chelators have been shown to have toxic

56

effects on plants and microorganisms (Evangelou et al. 2007) and to be associated with a high

57

risk of metal leaching into groundwater (Nowack et al., 2006). Inoculation with siderophore-

58

producing bacteria (SPB) has recently been proposed as an alternative strategy to the direct

59

supply of chelators (Braud et al. 2009; Rajkumar et al. 2010). The continuous localized

60

production of siderophores in the close vicinity of plant roots, where most of the SPB

61

establish themselves (Deweger et al. 1995), should help to promote the phytoextraction of

62

metals whilst minimizing the risk of metal leaching. Siderophores are low-molecular weight

63

organic chelators (150-2000 Da) with a high selectivity for Fe(III) (e.g. log K= 42 for

64

enterobactin produced by E. coli, Hider and Kong 2011) and a good affinity for divalent

65

metals, including Cd(II), Cu(II) and Ni(II) (Schalk et al. 2011). Siderophore synthesis is

66

regulated by the intracellular concentration of Fe and is enhanced by Fe starvation (Visca et

67

al. 2007). The presence of divalent metals, including Cu and Ni, has also been shown to

68

enhance the production of siderophores, as reported for the production of yersiniabactin by E.

69

coli (Chaturvedi et al. 2012) and pyoverdine (Pvd) by P. aeruginosa (Braud et al. 2010).

70

Siderophores are thought to protect bacteria against the toxic effects of metals, by

71

sequestering them outside the bacterial cell, in their siderophore-chelated form (Braud et al.

72

2010; Schalk et al. 2011). Thus, in addition to their known role in supplying bacteria with

73

Fe(III), siderophores are thought to play a significant role in the biogeochemical cycling of

74

metals. The combination of phytoextraction with SPB inoculation has been tested several

75

times in soil, but conflicting results were obtained (Rajkumar et al. 2010). For example, SPB

3

76 1 2 77 3 4 5 78 6 7 79 8 9 10 80 11 12 81 13 14 82 15 16 17 83 18 19 84 20 21 22 85 23 24 86 25 26 27 87 28 29 88 30 31 89 32 33 34 90 35 36 91 37 38 39 92 40 41 93 42 43 44 94 45 46 95 47 48 49 96 50 51 97 52 53 98 54 55 56 99 57 58 100 59 60 61 62 63 64 65

has been shown to increase the uptake of Cr and Pb by maize (Braud et al. 2009) but to decrease the uptake of Cd by brown mustard (Sinha and Mukherjee 2008). As suggested by Ma et al. (2011), the efficiency of SPB-assisted phytoextraction probably depends on the combination of plant, bacterium and metal. Soil properties, such as soil pH in particular, should also be considered, because protons may compete with metals for complexation by siderophores (Tansupo et al. 2008). Unfortunately, there has been little exploration of the mechanisms underlying SPB-assisted phytoextraction. One reason for this is the difficulty of monitoring siderophores and investigating the complexation of metals in soil solution. Simplified systems are thus required to provide guidance concerning the selection of the optimal plant-bacterium association to maximize the phytoextraction of a given metal from a given soil. The addition of siderophore directly to the contaminated matrix is a relevant alternative, assuming that siderophore production is the principal mechanism by which SPB alter the phytoavailability of metals in soils. We recently initiated a study on pyoverdine (Cornu et al. 2014), the major siderophore produced by P. aeruginosa. This work aimed to dissect the effect of Pvd supply on the phytoextraction of Cd and Cu in a calcareous soil. We found that the addition of Pvd to the soil increased the phytoextraction of Cu, consistent with the much higher stability constant of Pvd-Cu than of Pvd-Cd. These findings highlighted the important role of complexation in Pvd-assisted phytoextraction and led on to this study, in which we aimed to dissect the relationships between metal speciation in solution and metal phytoavailability in the presence of Pvd. We focused on Cd and Ni, two metals intensively studied in phytoremediation and phytomining. Experiments were performed in hydroponic conditions, in the presence and absence of a solid mineral phase (smectite). The main objectives were: (i) to study the ability of Pvd to coordinate Ni(II), its coordination properties with Cd(II) having already been characterized (Cornu et al. 2014), (ii) to assess the changes in Cd and Ni speciation and phytoavailability of Cd and Ni due to the supply of Pvd and (iii) to

4

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

101

determine whether inoculation with P. aeruginosa caused similar changes in the fate of Ni to

102

the direct supply of Pvd.

103 104

Materials and methods

105

Pyoverdine (Pvd)

106

Production, purification and quantification

107

A wild-type laboratory strain of P. aeruginosa (ATCC 15692, also called PAO1) was used to

108

produce Pvd. Bacteria were grown at 30 °C in iron-deficient casamino-acid (CAA) medium

109

for 24 h, this time point corresponding to peak Pvd production at the start of the stationary

110

growth phase. The culture medium was then removed by centrifugation and the supernatant

111

containing Pvd was filtered and its pH adjusted to 6.0 before purification. Pvd was purified by

112

ion-exchange chromatography on an Amberlite-XAD-4 column (Sigma), as described by

113

Carson et al. (2000). Pvd was eluted with a 1:1 (v/v) mixture of ethanol and H2O, freeze-dried

114

and stored at -20 °C. The level of contamination of Pvd with Fe, Cu, Cd and Ni was checked

115

at this stage in the purification procedure and was considered suitable for the experiments

116

presented here (Online Resource 5). For the determination of Pvd affinity constants for Ni(II),

117

Pvd was further purified by HPLC, as described by Albrecht-Gary et al. (1994). Pyoverdine

118

concentration was assessed in solution by measuring absorbance at 400 nm ( = 19000 M-1

119

cm-1 for Pvd at pH 7, Albrecht-Gary et al. 1994).

120 121

Physicochemical characterization

122

The physicochemical investigation was restricted to the succinic isoform of Pvd (Online

123

Resource 1). The metal coordination properties of the three isoforms of Pvd are thought to be

124

similar, because the succinic, succinamide and -ketoglutaric moieties are not involved in

125

metal coordination (Meyer et al., 2002). The protonation and Cd(II) coordination properties

5

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

126

of this isoform of Pvd were described in a previous study (Cornu et al. 2014). We investigated

127

the Ni(II) coordination properties of this isoform here, by plotting

128

absorption against pH, as described by Elhabiri et al. (2004) and Albrecht-Gary et al. (1994).

129

Pyoverdine was found to be a valuable spectrophotometric probe, sensitive to both pH and

130

metal uptake (i.e. the catecholate unit). It was therefore possible to use differences in

131

absorption in the presence and absence of Ni(II) as a function of pH to evaluate the

132

protonation and stability constants of Ni(II) complexes. Briefly, we introduced an aliquot of

133

0.04 liters containing equimolar concentrations of Pvd and Ni(II) ([Pvd]tot = 1.93 x 10-5 M and

134

[Pvd]tot/[Ni]tot = 1.04) dissolved in Millipore-Q water solvent (I = 0.1 M NaClO4) into a

135

jacketed cell maintained at 25.0 ± 0.2 °C. The initial pH of the solution was adjusted to ~2

136

with HClO4 and the UV-visible absorption spectrophotometric titration of Ni(II) complexes

137

(2.3< pH < 11.3) was then performed by adding known volumes of NaOH with a

138

microburette. After each addition of base, special care was taken to ensure that complete

139

equilibration was achieved, by measuring absorption spectra over time (Ni(II) complexes with

140

Pvd were found to form within one minute). A final absorption spectrum was then obtained

141

and considered for the final datasets (absorption and potentiometric measurements).

142

Absorption spectra versus pH were then recorded with a Varian Cary 50 spectrophotometer

143

fitted with Hellma optical fibers and a Suprasil quartz immersion probe. Specfit software (see

144

Elhabiri et al. 2004 for further details) was used to adjust the stability constants and the

145

corresponding molar extinction coefficients (M-1 cm-1) of the species formed at equilibrium.

146

This software uses factor analysis to reduce the absorbance matrix and to extract the

147

eigenvalues before multiwavelength fitting to the reduced dataset, according to the Marquardt

148

algorithm.

149 150

Smectite

6

spectrophotometric

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

151

Smectite was extracted from bentonite obtained from the Ozurgeti mine (Georgia, Eurasia). A

152

suspension of 20 g of bentonite in 0.7 liters of distilled water was shaken for 16 h with

153

Amberlite IR 1200 Na resin (40 ml), to facilitate dispersion. The mixture was filtered, poured

154

into 30 cm-high cylinders and allowed to decant for 16 h. The clay fraction (diameter < 2

155

µm), corresponding to the upper 20 cm according to Stokes’ law, was then recovered and

156

transferred to small bottles for centrifugation (1 h at 4000 x g). The pellets were then dried at

157

75 °C for 24 h. We contaminated the smectite with Cd and/or Ni, by suspending the smectite

158

in a solution containing 1 µM Cd(NO3)2 and/or 1 µM NiCl2, at a concentration of 0.2 g l-1 and

159

incubating this mixture for one day. The suspension was then filtered and the smectite

160

enriched in Cd (Sm-Cd), Ni (Sm-Ni) or Cd and Ni (Sm-M) was dried and sterilized by

161

tyndallization to preserve the layers.

162 163

Free ionic fractions of Cd and Ni

164

The free hydrated ionic fractions of Cd and Ni (i.e. the M2+ form) in solutions to which Pvd

165

had been added, was calculated from the stability constants shown in Table 1 and for H+,

166

Cd(II), Ni(II), Fe(III) and Cu(II) (traces of Fe and Cu were detected in Pvd), with Hyss 2009

167

software (Alderighi et al. 1999). Competition with Ca(II) was not taken into account because

168

hydroxamate moieties have a much lower affinity for Ca(II), as previously demonstrated by

169

Farkas et al. (1999). We investigated the influence of both Pvd concentration (0 < Pvd < 100

170

µM) and pH value (6 < pH < 7) on the free ionic fractions of Cd and Ni. The input parameters

171

used for calculations are specified (Online Resource 3). For Cd, the free ionic fraction was

172

also determined experimentally by the ion exchange method described by Schneider (2006).

173

This method is based on the Cd/Ca exchange properties of a cation exchange resin (Amberlite

174

IR-120) converted to the Ca form. Briefly, we added 0.006 g of Ca-resin to 0.003 liters of

175

solution in a 0.005-liter polyethylene tube. A known amount of radioactive

7

109

Cd was

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

176

immediately added to the suspension and the tube was shaken for 24 h at 20 ± 1 °C on a

177

roller. The free ionic Cd fraction was determined from the radioactivity remaining after

178

contact, the total Ca concentration in the solution before and after contact and the resin

179

Vanselow selectivity coefficient for Cd/Ca exchange (vKCd/Ca= 0.661), as described by

180

Schneider (2006). This determination was carried out with the background solution for Exp. 2

181

(725 µM Ca(NO3)2, 2 mM MES, 1 µM Cd(NO3)2), to which 50 µM Pvd was added, at pH

182

values of 6.0, 6.4 and 7.0. The same determination was not carried out for Ni, because the

183

resin Vanselow selectivity coefficient for Ni/Ca exchange has not yet been characterized.

184 185

Plant material

186

The cropping device used for the bioassays was adapted from that designed by Chaignon and

187

Hinsinger (2003). Ten seeds of tomato (Lycopersicon esculentum cv. Saint Pierre) were

188

surface-sterilized with 6% H2O2 and sown in a container, on the surface of the grid. The seeds

189

were allowed to germinate and the plants were grown in hydroponic conditions for 13 days.

190

Ten containers were placed on top of a 6-liter bucket, containing an aerated nutrient solution.

191

During the first five days, plants were supplied with 600 µM CaCl2 and 2 µM H3BO3. After

192

germination, the plants were supplied with a complete nutrient solution for eight days, of the

193

following composition: 2 mM KNO3, 2 mM Ca(NO3)2, 1 mM MgSO4, 0.5 mM KH2PO4, 50

194

µM FeNaEDTA, 10 µM H3BO3, 2 µM MnCl2, 1 µM ZnSO4, 0.2 µM CuCl2 and 0.05 µM

195

Na2MoO4. This initial period of culture was conducted in a growth chamber with the

196

following day–night conditions: 25 °C with 300 µmol photons m2 s-1 during the 16 h day and

197

20 °C during the 8 h night. The relative humidity of the air was set at 70%.

198 199

Experiments

200

Effect of pyoverdine on the mobilization of Ni and Cd from smectite (Exp. 1)

8

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

201

We suspended 0.01 g of smectite enriched in Cd and Ni (Sm-M) (see section 2.2) in 0.05

202

liters

203

morpholino)ethanesulfonic acid (MES, pH 6.4), in the presence (Pvd+, concentration fixed at

204

50 µM) or absence (Pvd-) of pyoverdine. The samples were incubated for 24 h with shaking

205

(200 rpm), and were then centrifuged (20 min, 8000 x g). The supernatants were filtered (0.22

206

µm), acidified and their total Cd, Ni, Fe and Al concentrations were determined by

207

inductively coupled plasma-optical emission spectroscopy (ICP-OES). Each treatment

208

(Pvd+/-) was replicated five times.

of

a

background

solution

of

725

µM

Ca(NO3)2

and

2

mM

2-(N-

209 210

Effect of pyoverdine on the uptake of Cd (Exp. 2) and Ni (Exp. 3) by the plants

211

Plant containers were placed on the top of individual 0.15-liter polyethylene pots containing a

212

background solution of 725 µM Ca(NO3)2 and 2 mM MES (pH 6.4). In Exp. 2, plant roots

213

were exposed to Cd in the presence (Pvd+, concentration fixed at 50 µM) or absence (Pvd-) of

214

pyoverdine. We tested two sources of Cd: Cd was introduced either as soluble Cd(NO3)2

215

(noted Cd) or sorbed onto smectite (Sm-Cd). The concentration of Cd(NO3)2 was fixed at 1

216

µM and that of Sm-Cd was fixed at 0.2 g l-1. We also set up an experiment without Cd, as a

217

control. Plants were exposed to the metal for 48 h in a growth chamber with the same day-

218

night conditions as the initial plant culture. The solution was stirred at 130 rpm and harvested

219

after 48 h for analyses of pH and determinations of the concentrations of Cd, Ni and Pvd.

220

Each treatment (Cd source, Pvd +/-) was replicated six times. The same protocol was carried

221

out in Exp. 3, but with the replacement of Cd with Ni, and with 1 µM NiCl2 and 0.2 g l-1 Sm-

222

Ni as the sources of nickel.

223 224

Effect of P. aeruginosa (PAO1) on the uptake of Ni by plants (Exp.4)

9

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

225

The background solution was modified to sustain bacterial growth, based on the composition

226

of the CAA medium: 5 g l-1 casamino acids, 6.77 mM K2HPO4, 1 mM MgSO4, 500 µM

227

Ca(NO3)2 and 1.83 mM MES, with the initial pH adjusted to 6.4. P. aeruginosa PAO1 was

228

added to this solution, at a density of 1011 CFU l-1, just before plant exposure. At harvest

229

(after 48 h), bacterial density and Pvd concentration were estimated by monitoring absorbance

230

at 600 and 400 nm, respectively. All other experimental conditions were similar to those used

231

in experiments 2 and 3.

232 233

Plant analyses

234

After exposure, plant shoots and roots were collected separately and rinsed in deionized

235

water. Roots and shoots were dried for 48 h at 65 °C in a ventilated oven and were then

236

weighed to determine their dry matter content. Dried plant roots and shoots were milled and

237

digested in a 4:1 mixture of 65% HNO3 and 30% H2O2 in a microwave mineralizer (Ethos

238

One, Milestone). The concentrations of Cd and Ni in the plant digest were assayed by

239

inductively coupled plasma optical emission spectrometry (ICP-OES; Liberty II, Varian). The

240

validity and accuracy of these two procedures were checked with a certified sample of peach

241

leaves (NIST-SRM 1547).

242 243

Statistical analyses

244

Analyses of variance were performed, with Tukey’s test, to identify significant differences

245

between treatments. The figures show mean values, with 95% confidence intervals calculated

246

from the six replicates per treatment. We used SYSTAT 10 Edition 2000 software for

247

statistical analysis (SPSS Inc., 233 S. Wacker Drive, Chicago, USA).

248 249

Results and Discussion

10

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

250

Coordination of Cd and Ni by Pvd

251

The protonation and stability constants of free pyoverdine and of its complexes with Cd(II),

252

Cu(II) and Fe(III) were determined in previous studies (Albrecht-Gary et al. 1994; Cornu et

253

al. 2014) and are shown in Table 1. For Ni(II), Online Resource 2a displays the spectral

254

variations measured during the absorption titration of complexes of nickel with Pvd as a

255

function of pH. The large bathochromic shift ( > 32 nm for LNi vs. LH4, L being the fully

256

deprotonated form of Pvd) recorded upon Ni(II) complexation made it possible to probe the

257

ability of Pvd to coordinate Ni(II). Statistical processing of these spectrophotometric and

258

potentiometric data led to the characterization of a single Ni(II) monochelate in two different

259

protonation states (LNi and LHNi) and to the determination of the corresponding stability

260

constants (Table 1). The electron absorption spectra of LNi and LHNi are shown in Online

261

Resource 2b. Throughout the statistical analysis, the protonation constant of free Pvd and the

262

hydrolysis constant of Ni(II) were fixed. As the succinate and arginine residues of Pvd are not

263

involved in Ni(II) coordination, their protonation constants were expected to be similar to

264

those of free Pvd, as already reported for ferric Pvd complexes by Albrecht-Gary et al.

265

(1994). The stability constant of Pvd-Ni (log KL'Ni = 10.9, L' designates the Pvd system in

266

which the ionizable sites of the arginine and succinate residues are omitted) was found to be

267

higher than that of Pvd-Cd (log KL'Cd = 8.2, Cornu et al. 2014), but much lower than those

268

determined for Cu(II) (log KL'Cu = 20.1, Cornu et al. 2014) or Fe(III) (log KL'Fe = 30.8,

269

Albrecht-Gary et al. 1994). The ranking of Pvd-M stability constants (KL'M, Table 1) is thus

270

consistent with the Irving-Williams series, according to which, the classical order for

271

transition-metal complexes is: Mn(II) < Fe(II) < Co(II) < Ni(II) < Cu(II) > Zn(II).

272

Interestingly, the sequence of the Pvd-M monochelate protonation constants (KL'HM, Table 1):

273

Cd (9.7) > Ni (7.7) > Cu (5.5) > Fe (