Tài liệu Báo cáo khoa học: The two Caenorhabditis elegans metallothioneins (CeMT-1 and CeMT-2) discriminate between essential zinc and toxic cadmium docx

Zinc levels were significantly increased in all knockout strains, but were most pronounced in the CeMT-1 knockout, mtl-1 tm1770, while cadmium accumulation was highest in the CeMT-2 knock

and CeMT-2) discriminate between essential zinc and toxic cadmium

Sukaina Zeitoun-Ghandour1, John M Charnock2, Mark E Hodson3, Oksana I Leszczyszyn4,

Claudia A Blindauer4and Stephen R Stu¨rzenbaum1

1 School of Biomedical & Health Sciences, King’s College London, UK

2 School of Earth, Atmospheric and Environmental Sciences, University of Manchester, UK

3 Department of Soil Science, University of Reading, UK

4 Department of Chemistry, University of Warwick, Coventry, UK

Introduction

Metal pollution in the environment is a matter of

con-cern Many studies have focused on the use of

terres-trial biomonitors to determine how organisms, in

particular invertebrates, control and tolerate increased exposure to elevated levels of metals [1–7] Responses may include avoidance, excretion, chelation or

Keywords

affinity; C elegans; cadmium; metal

speciation; metallothionein; zinc

Correspondence

S Stu¨rzenbaum, School of Biomedical &

Health Sciences, Pharmaceutical Science

Division, King’s College London, 150

Stamford Street, London SE1 9NH, UK

Fax: +44 2078484500

Tel.: +44 2078484406

E-mail: stephen.sturzenbaum@kcl.ac.uk

(Received 22 January 2010, revised 23

March 2010, accepted 30 March 2010)

doi:10.1111/j.1742-4658.2010.07667.x

The nematode Caenorhabditis elegans expresses two metallothioneins (MTs), CeMT-1 and CeMT-2, that are believed to be key players in the protection against metal toxicity In this study, both isoforms were expressed in vitro in the presence of either Zn(II) or Cd(II) Metal binding stoichiometries and affinities were determined by ESI-MS and NMR, respectively Both isoforms had equal zinc binding ability, but differed in their cadmium binding behaviour, with higher affinity found for CeMT-2 In addition, wild-type

C elegans, single MT knockouts and a double MT knockout allele were exposed to zinc (340 lm) or cadmium (25 lm) to investigate effects in vivo Zinc levels were significantly increased in all knockout strains, but were most pronounced in the CeMT-1 knockout, mtl-1 (tm1770), while cadmium accumulation was highest in the CeMT-2 knockout, mtl-2 (gk125) and the double knockout mtl-1;mtl-2 (zs1) In addition, metal speciation was assessed by X-ray absorption fine-structure spectroscopy This showed that O-donating, probably phosphate-rich, ligands play a dominant role in maintaining the physiological concentration of zinc, independently of metallothionein status In contrast, cadmium was shown to coordinate with thiol groups, and the cadmium speciation of the wild-type and the CeMT-2 knockout strain was distinctly different to the CeMT-1 and double knock-outs Taken together, and supported by a simple model calculation, these findings show for the first time that the two MT isoforms have differential affinities towards Cd(II) and Zn(II) at a cellular level, and this is reflected at the protein level This suggests that the two MT isoforms have distinct in vivo roles

Abbreviations

EXAFS, extended X-ray absorption fine structure; ICP-OES, inductively coupled plasma optical emission spectrometry; XANES, X-ray absorption near-edge structure.

immobilization of metal ions, or activation of general

stress response mechanisms⁄ proteins [8,9] A

promi-nent response pathway involved in the chelation of

metal ions involves metallothioneins (MTs) These are

proteins of low molecular mass that are characterized

by a high cysteine content [15–30%], high heat stability

and lack of aromatic amino acids (including histidine)

[10,11] Although the discovery of MTs dates back to

1957 [12], their precise physiological functions are still

debated It has become evident that a single function

does not exist for this heterogeneous superfamily of

proteins, and that they are ‘multipurpose’ proteins

[13], with roles in protection against cadmium toxicity

[14], essential Cu(I) and Zn(II) homeostasis [15], and

response to oxidative stress [16]

There is growing evidence that the existence of

mul-tiple MT isoforms is associated with functional

differ-entiation, for example in snails [17], earthworms [18],

plants [19] and vertebrates [16] So far, studies have

focused on the discrimination between monovalent

Cu(I) and divalent Zn(II) and Cd(II) [20] As the

coor-dination geometries of mono- and divalent metal ions

are very distinct (digonal or trigonal planar versus

tet-rahedral), it is easily conceivable that the steric

require-ments imposed by binding of these metal ions will

differ, and this offers a straightforward mode of

dis-crimination

In contrast, discrimination between the essential

Zn(II) and toxic Cd(II), which have relatively similar

coordination chemistry, presents a major challenge for

organisms that are exposed to both metal ions The

soil nematode Caenorhabditis elegans is a case in point

[21,22], and offers a unique biological system for the

study of MT isoform specificity, because its fully

sequenced genome contains only two metallothioneins

CeMT-1 and CeMT-2 [23] The encoded proteins bear

the hallmarks of metallothioneins, i.e they are small

and cysteine-rich, and their expression is induced by

metals [24] More recently, RNA interference (RNAi)

and chromosomal deletion of the C elegans MT loci

have highlighted an increased sensitivity of mutant

strains to metal toxicity, reflected by reduced growth,

brood size and lifespan [23,25] In addition,

phytochel-atins, which are small, non-ribosomally synthesized,

Cd-binding peptides, play a prominent role in

protec-tive responses to Cd exposure [26–28]

Significantly, the two MT isoforms show differential

expression profiles [24] CeMT-2 is only induced in

intestinal cells in the presence of cadmium, but CeMT-1

is also constitutively active in three cells of the lower

pharyngeal bulb [24] These studies provided the first

evidence that CeMT-1 and CeMT-2 may have distinct

in vivo functions, but although additive sensitivity

towards cadmium was observed in C elegans metallo-thionein knockout alleles, isoform-specific in vivo effects have not been observed to date, even by detailed meta-bolomic profiling analysis [28]

At the protein sequence level, CeMT-1 and CeMT-2 display intriguing differences, and are more different from one another than vertebrate MT isoforms CeMT-1 contains a 15 amino acid insert with two additional histidines and one cysteine [23,24,29], with a further histidine at position 54 (see Fig S1A for sequence alignment) Recent in vitro characterization

of recombinantly expressed CeMT-1 and CeMT-2 by ESI-MS and CD spectroscopy has begun to determine the differences in metal binding properties of the two isoforms [30] A clear preference for divalent metal ions was discovered, but, most significantly, this study suggested that CeMT-1 and CeMT-2 show differential metal preferences, with CeMT-1 biased towards Zn(II) and CeMT-2 biased towards Cd(II)

In the present study, we explore whether these quali-tative findings are reflected by overall in vivo metal accumulation and speciation of metallothionein-mutated C elegans strains, as well as the in vitro metal ion affinities of the two isoforms under metal-replete and metal-excess conditions

Results Metal-binding properties of recombinant metallothioneins

For characterization and quantification of the metal-binding properties of CeMT-1 and CeMT-2, an expres-sion strategy was adapted that avoids the use of fuexpres-sion tags, as we have previously observed that tags can influ-ence the metal binding properties of recombinantly expressed metallothioneins [31] In contrast to most expression tag systems, our protocol also allows the expression of proteins with no additional residues at the termini Careful chemical precipitation followed by gel filtration chromatography yielded pure proteins (> 95%, as judged by ESI-MS analysis) with no addi-tional species (see Fig S1B,C for expression and purifi-cation, respectively)

Both metallothioneins were expressed in the presence

of either Zn(II) or Cd(II) in the culture medium The metal ion stoichiometry of the purified proteins was determined by inductively coupled plasma optical emission spectrometry (ICP-OES) and ESI-MS (Fig 1 and Table 1) Consistent with previous work [30,32], CeMT-2 expressed in the presence of Cd(II) had six Cd(II) ions bound We found the same stoichiometry for Zn(II), with no discernible peaks for metal-depleted

species in the mass spectrum, and corresponding

ICP-OES results Consistent with recent findings [30], our

analysis also confirmed that CeMT-1 binds seven metal

ions, with Zn7-CeMT-1 the only species observed in

mass spectra at neutral pH for Escherichia coli cells

grown in Zn(II)-supplemented medium (Fig 1)

To allow quantification of metal affinities and their

comparison, it was very important to obtain clearly

defined homo-metallic species, and the data compiled

for Zn6-CeMT-2, Cd6-CeMT-2 and Zn7-CeMT-1 in

Fig 1 and Table 1 show that this was achieved by

expression in the presence of the desired metal ion

However, the CeMT-1 form isolated from

Cd(II)-sup-plemented cultures was Cd6Zn-CeMT-1 (Fig S2), and

incorporation of seven cadmium ions was only possible

by reconstitution of metal-free CeMT-1 with rigorous

exclusion of Zn(II), using an established protocol [33]

Although the Cd7-CeMT-1 species was the major form

in this preparation, we observed a loss of definition in

metal binding stoichiometry despite extensive gel

filtra-tion and washing, as Cd8-CeMT-1 and Cd9-CeMT-1,

as well as a very small amount of Cd6Zn1-CeMT-1

me-talloforms, were present as minor species (Fig 1B)

The contribution of these over-metallated species

is also apparent in the stoichiometry determined by

ICP-OES given the larger than expected stoichiometry

for cadmium-bound CeMT-1

The overall in vitro affinities of CeMT-1 and CeMT-2

towards Zn(II) and Cd(II) were determined by

comp-etition experiments using the metal chelator 5F-BAPTA

[34] and 19F-NMR spectroscopy under conditions that allow direct comparison with literature values The stability constants obtained for the homo-metallic zinc and cadmium complexes of CeMT-1 and CeMT-2 are given in Table 2, and represent means over all six (CeMT-2) or seven (CeMT-1) binding sites As expected for predominantly thiol coordination, the sta-bility constant for cadmium binding in CeMT-2 was significantly larger than that for zinc binding, and was close to the value for human MT-2 measured under similar conditions [34] Remarkably, this was not the case for CeMT-1 Although both isoforms displayed identical affinities for Zn(II), cadmium binding in the

CeMT-1 CeMT-2

8000 8400 8800 9200

[–Met]

[–Met]

[–Met]

[–Met] [–Met]

C

D

0 50 100 50

100

A

B

[–Met]

[+Met]

Cd 6

Cd 6

Cd 6 Zn

Cd 8

Cd 7

Cd 9

[–Met]

Mass (Da)

Fig 1 Deconvoluted ESI mass spectra of

the various metalloforms of

Caenorhabd-itis elegans MTs Holo zinc (A) and cadmium

(B) species of CeMT-1; zinc (C) and

cad-mium (D) species of CeMT-2 obtained at

neutral pH (10 m M ammonium acetate, 10%

methanol) Samples (A), (C) and (D) result

from expression in the presence of the

respective metal ion; sample (B) was

obtained by expression in presence of Zn(II)

and reconstitution of the apoprotein with

Cd(II) )Met and +Met annotations refer to

the absence or presence of the N-terminal

methionine residue for each species The

peaks in (A), (C) and (D) to the right of the

main peaks correspond to Na+adducts.

Table 1 Metal to protein stoichiometries for CeMT-1 and CeMT-2 metalloforms determined by mass spectrometry and elemental analysis Theoretical and observed mass are given for the major species in each mass spectrum )MET, without Met; +Met, includ-ing Met.

Metalloform

Mass spectrometry

Stoichiometry (ICP-OES) Theoretical

mass (Da)

Observed

CeMT-1

Zn 7 8402.7 ( )Met) 8401.9 ± 0.7 6.6 ± 0.7 8.8 ± 0.8

Cd7 8731.9 ( )Met) 8731.5 ± 0.5 CeMT-2

Zn 6 6843.1 ( )Met) 6843.3 ± 0.7 5.6 ± 0.6 6.0 ± 0.6

Cd6 7256.5 (+Met) 7257.0 ± 0.6

Cd7-CeMT-1 complex was dramatically weaker than

that in Cd6-CeMT-2 and other MTs (Table 2), even

when the effect of the over-metallated species is taken

into account To date, the overall stability of cadmium

binding to CeMT-1 is the lowest value reported for

any MT However, it is important to note that, despite

this, the overall affinity of CeMT-1 towards Cd(II) is

still an order of magnitude larger than that for Zn(II),

so it would be inappropriate to claim that CeMT-1 is

a Zn(II)-specific MT, although one particular site does

indeed appear to have an absolute preference for

Zn(II)

In vivo metal speciation in wild-type C elegans:

excess zinc and cadmium are handled differently

To investigate the native organism-wide responses to

cadmium and zinc exposure, the wild-type (N2) C

ele-gans strain was grown on supplemented media, and

the collective ligand environment of intracellular

cad-mium and zinc was analysed by X-ray absorption

near-edge structure (XANES) and extended X-ray

absorption fine structure (EXAFS) spectroscopy

Because the low-energy photoelectrons have a long

mean free path, XANES spectroscopy is strongly

affected by multiple scattering, which means that it is

very sensitive to differences in geometry as well as

coordination number and oxidation state Although

this complexity complicates the analysis of XANES

data, it is valuable as a ‘fingerprint’ technique,

com-paring unknowns with model compound spectra

Indeed, XANES and EXAFS spectroscopy have

previ-ously been used successfully on rat liver samples to

distinguish different binding modes in Cd–S clusters

and metallothionein [35] The cadmium XANES spec-tra (Fig 2A) show that the edge shape and position are distinct from Cd–O-bonded complexes, and display features that are more similar to S-coordinated cad-mium models (Cd–S and rat Cd7-MT) The EXAFS results and associated Fourier transforms, together with the best possible fits, are shown in Fig S3, and indicate a single major transform peak at

R + D = 2.5 A˚ (other fits gave higher residuals, data not shown) When modelled, the cadmium EXAFS data produce a best fit with one shell of four sulfur scatterers at 2.49 A˚ (Table 3) However, due to the small size of the nematodes, even 300 000–500 000 syn-chronized nematodes generated only a dilute sample that, although sufficient for analysis, produced a short data range and had a poor signal-to-noise ratio, thus precluding fitting of further shells of scatterers Although the EXAFS data were admittedly noisy, they

Table 2 Zinc and cadmium binding affinities for MTs in

Caenor-habditis elegans and other species Log K binding constants for

zinc and cadmium metalloforms of C elegans MTs were

deter-mined and compared to those of other MTs The log K for

cadmium binding was determined by competition between protons

and metal ions for complexed thiolate ligands [33] In all cases, the

log K of zinc binding was measured by competition for metal ions

between the MTs and 5F-BAPTA (ionic strength 4 m M and pH 8.1)

[34,48,53].

a Recalculating this value to account for over-metallation yields a

log K value of 13.4.

Wildtype Cd-S Cd-MT Cd(OH)2 CdSO4

26 680 26 720 26 760 26 800 26 840

Energy (eV)

Wildtype Zn-S

Zn foil ZnSO4 H2O

Zn3(PO4)2

Energy (eV)

A

B

Fig 2 XANES profiles in wild-type nematodes and standards Cd XANES spectra (A) and Zn XANES spectra (B) For cadmium,

a minor monochromator drift during data collection made it necessary to correct the edge position using reference spectra, therefore the error in the absolute position of the edge was marginally larger than the station benchmark.

are of sufficient quality to justify the conclusion that

sulfur coordination gives the best single shell fit, which

is also consistent with XANES results

Previously recorded Zn(SO4)ÆxH2O, ZnS and

Zn3(PO4)2 spectra were used to model the zinc

XANES spectra of the wild-type (N2) nematode The

spectra show that, of all reference compounds, the

wild-type spectrum displayed features most similar to

those of the zinc phosphate standard (Fig 2B) This

was corroborated by EXAFS spectra (Table 3 and

Fig S4), which indicated the best fit to be four oxygen

atoms surrounding zinc in the first coordination shell,

with a mean Zn–O distance of 1.97 ± 0.03 A˚ This is

consistent with the tetrahedral coordination of zinc

phosphate [36]

Although, it may not be technically possible to

distin-guish between N⁄ O ⁄ F or between P ⁄ S ⁄ Cl as a scatterer,

the difference between O and S is substantial Therefore,

these data suggest that the mechanisms to deal with zinc

and cadmium employed by C elegans are separate and

distinct, as accumulated cadmium is predominantly

S-bound and zinc is predominantly O-bound

C elegans metallothioneins are not the only

players in metal detoxification and homeostasis

The effects on the ligand environments of cadmium

and zinc upon deletion of metallothioneins were

inves-tigated by comparative analysis of XANES spectra

(Fig 3) and EXAFS data (Table 3, Figs S3 and S4)

Cadmium XANES spectra (Fig 3A) for the MT

knockout strains do not show features significantly

dif-ferent from those observed for the wild-type (N2),

which suggests that the cadmium ions are still

predom-inantly coordinated by sulfur atoms However,

a broader edge and lower starting energy (1.5–2 eV)

were observed in spectra of the CeMT-1 KO and the double knockout, both were observed in spectra of the CeMT-1 knockout mtl-1 (tm1770) and the CeMT-1

Table 3 Cd⁄ Zn EXAFS parameters Best fit of the Cd ⁄ Zn K-edge data for Caenorhabditis elegans wild-type and metallothionein knockout strains, where r is the absorber–scatterer distance in A ˚ (± 0.02 A˚, inner shell; ± 0.05 A˚, outer shell), N is the number of scatterers around the central atom, 2d 2 is the the Debye–Waller factor in A˚2 , ± 25%, and the R factor is the least-squares residual, which indicates goodness

of fit.

Strain

Scatterer N r (A ˚ ) 2d 2 (A ˚ 2 ) R factor Scatterer N r (A ˚ ) 2d 2 (A ˚ 2 ) R factor

37.6

Energy (eV)

26 700 26 740 26 780 26 820 26 860

26 700 26 710 26 720

Wildtype CeMT-1 KO CeMT-2 KO Double KO

Wildtype CeMT-1 KO CeMT-2 KO Double KO

Energy (eV)

9657 9658 9659 9660

A

B

Fig 3 Metal speciation in Caenorhabditis elegans strains Compari-son of Cd XANES spectra (A) and Zn XANES spectra (B) obtained for

C elegans wild-type and metallothionein knockouts The inserts show the cadmium and zinc energy shifts between samples.

and CeMT-2 double knockout mtl-1;mtl-2 (zs1) Such

features are characteristic of a Cd–O phase This

observation was supported by EXAFS analysis, for

which data fitting was improved by addition of a shell

of oxygen scatterers and refining the distances, the

De-bye–Waller factors, and the ratio of S-bound to

O-bound cadmium (Table 3) The refined Cd–O distance

of 1.98 A˚ is arguably very short compared to

crystallo-graphic values for Cd–O in phosphates, carbonates,

etc However, this may be due to the Cd being

four-coordinate rather than six-four-coordinate, or may reflect a

larger than normal error in the EXAFS distance due

to weaker scattering of the oxygen than of the sulfur,

making the Cd–O contribution to the total EXAFS

spectra much smaller than the Cd–S contribution

Nevertheless, this confirms that the cadmium

coordina-tion environments of the CeMT-1 KO and double

KO differ from those of the wild-type (N2) and the

CeMT-2 KO, although it should be emphasized that

the majority of the cadmium remained bound to sulfur

in all strains (including the metallothionein deletion

stains) (Table 3)

The absence of either or both MT(s) had no

observa-ble effect on zinc speciation All XANES spectra

(Fig 3B) were similar, and EXAFS data analysis

(Table 3 and Fig S4) identified a common first shell

scatterer peak at 1.97 A˚, characteristic of

O-coordina-tion Adding a second shell of phosphorus scatterers

improved the fit for all four spectra, but this shell was

statistically significant only in the case of mtl-2 (gk125)

Although superbly fitted Zn and Cd XANES and

EXAFS data have previously illustrated that isolated

mammalian metallothioneins bind metals [37,38], the

data presented here reveal that the MT status of the

nematode does not significantly alter the overall

speci-ation of zinc and cadmium in cells, as the principal

ligand environment for both metals is similar to that

of the wild-type (N2) strain Nevertheless, the data

provide insights about the ultimate fate of each

metal ion As Cd–S bonds were maintained in the

double knockout strain, it is clear that the Cd–S

spe-cies observed do not correspond to

metallothionein-bound Cd Instead, it is likely that phytochelatins

dominate Cd speciation Excess zinc in C elegans is

clearly not MT- or phytochelatin-bound, but may be

sequestered through other means such as deposition in

phosphate-rich granules [39], possibly synonymous to

those found in earthworms [40,41]

However, these facts do not preclude a role for MTs

in metal handling, as binding of zinc and cadmium by

MTs may be transient, particularly as MTs are capable

of releasing metal ions relatively rapidly [42–44],

possi-bly to molecules downstream in the detoxification

pathway We therefore next address the question of whether MTs in C elegans influence overall zinc and cadmium levels at all

Metal levels in metal-exposed worms: CeMT-2 is important with regard to cadmium accumulation

In the wild-type (N2) strain, low levels of cadmium accumulation were observed when nematodes were grown (from L1 larval to pre-adult stage L4) on Cd-supplemented medium (Fig 4A and Table S1) An equivalent cadmium body burden was also observed in the CeMT-1 knockout This suggests that, upon cad-mium exposure, the CeMT-1 KO strain responds ‘as wild-type’, and the mechanism of this response is not hindered by lack of CeMT-1 in the cytosol In con-trast, the CeMT-2 KO strain shows an approximately twofold increase in cadmium levels compared to the wild-type (N2) strain, indicating that one of the mech-anisms by which C elegans normally responds to cad-mium exposure has been disrupted This is exacerbated

in the double knockout strain, in which the cadmium burden is significantly increased These data suggest that (a) if CeMT-2 is expressed, then CeMT-1 does not play a significant role in the cadmium response, (b) if CeMT-2 is absent, then CeMT-1 can fulfil the role carried out by CeMT-2, but not as effectively, and (c) if both metallothioneins are absent, the ‘normal’ and ‘back-up’ MT-mediated pathways of dealing with cadmium exposure are impaired, leading to hyperaccu-mulation of cadmium compared to the wild-type strain

Both CeMT-1 and CeMT-2 are important in maintaining physiological zinc levels Under control (non-metal-supplemented) conditions in the wild-type (N2), zinc was maintained at basal physiological levels (Fig 4B and Table S1) For the CeMT-2 and double mutant strains, no significant dif-ference from wild-type (N2) was observed; however, the CeMT-1 mutant accumulated slightly more Zn(II) Under Zn-supplemented conditions, all three knockout strains accumulated significantly more zinc compared

to the wild-type (N2) strain Of the single knockout strains, deletion of CeMT-1 resulted in accumulation

of the highest zinc concentration; however, deletion of CeMT-2 also led to a moderate increase in zinc levels The double knockout did not differ significantly from the CeMT-1 knockout This indicates that (a) CeMT-1 has a more significant role than CeMT-2 in the regula-tion of zinc levels, (b) both CeMT-1 and CeMT-2 are required to maintain physiological zinc levels, as lack

of CeMT-2 also disrupts the mechanism that prevents

zinc accumulation, and (c) CeMT-1 and CeMT-2

oper-ate in a synergistic manner in zinc trafficking

Figure 4B also includes data for Zn(II) levels after

Cd exposure, and these data offer further interesting

insights In the CeMT-1 knockout, which showed only

basal Cd levels, Zn levels were depressed, but were

ele-vated under Zn exposure This observation can only

be rationalized if we consider that the two isoforms are regulated differently, and that Cd(II) strongly induces CeMT-2 Hence, in the CeMT-1 knockout, induction of CeMT-2 may have led to enhanced excre-tion (or reduced uptake) of not only Cd(II), but also some of the basal Zn(II), possibly mediated by the same CeMT-2-dependent pathway No difference in

Zn levels was observed for the CeMT-2 knockout mutant, indicating that zinc homeostasis functioned normally even in the presence of Cd(II) Finally, in the double knockout, a significant increase in Zn(II) levels was observed, indicating significant disruption of Zn(II) homeostasis

CeMT-1 and CeMT-2 provide a system for discrimination between essential Zn(II) and toxic Cd(II)

The question of how cells select the correct metal ions

is of current interest [45] One emerging concept holds that it is not the absolute but the relative affinity of various metal-trafficking proteins towards various metal ions in a common cytosol that governs metal ion selection and distribution The in vitro and in vivo data presented here are consistent with this concept, and allow development of a framework that helps to understand the discrimination between Zn(II) and Cd(II) by the two metallothioneins in C elegans, as well as at a more general level

To illustrate this idea, we have used the in vitro (Table 2) and in vivo (Fig 4) data to approximate the proportion of metal ions bound to CeMT-1 and CeMT-2 if presented with Zn : Cd ratios as encountered

by C elegans Using a Cd : Zn ratio of 33 : 1 [21 nm Cd(II) and 0.7 lm Zn(II)] and 0.1 lm of CeMT-1 and CeMT-2 each, and the stability constants given in Table 2, it can be calculated that 98.6% of Cd(II) is bound to CeMT-2, and only 1.4% to CeMT-1 Zn(II) is more evenly distributed (45 : 55%) between CeMT-1 and CeMT-2 When equimolar amounts of Zn(II) and Cd(II) are used (0.65 lm each), 93% of Zn(II) is bound

to CeMT-1, and 85% of Cd(II) is bound to CeMT-2 With a 10-fold excess of MTs and the same metal con-centrations, 98.4% of Cd are bound to CeMT-2, and the Zn(II) distribution is 57 : 43% for CeMT-1 : CeMT-2 These numbers have been calculated based on two relatively crude simplifications: first that all binding sites

in CeMT-1 and CeMT-2 are equivalent, and second that

no other competing ligands are present It is conceivable that the overall reduction in Cd(II) affinity is to a con-siderable extent, but not exclusively, due to weaker binding to the histidine-rich site It is therefore likely that the difference in affinities for binding to the

Cadmium

Zn-exposed

a a

b

c

ND ND ND ND ND ND ND ND

KO

CeMT-2 KO

Double KO 0

500

1500

1000

2000

2500

3000

c

b

a a

a

b

c

b

a

3500

Zinc

*

**

**

**

**

Control

0

Cd-exposed

20

40

60

80

100

Double KO

CeMT-2 KO

CeMT-1 KO WT

A

B

Fig 4 Metal accumulation in nematodes Levels of cadmium (A)

and zinc (B) were quantified by ICP-OES in Caenorhabditis elegans

wild-type and metallothionein deletion strains cultured in the

pres-ence or abspres-ence of cadmium (25 l M ) or zinc (340 l M ) Values are

the means ± SEM of five replicates Different letters above bars

indicate statistical significance compared with each other.

*P < 0.05;**P < 0.01 ND, not detectable (below detection limits).

all-thiolate sites in the two proteins is < 1.9 orders of

magnitude However, even a difference of only 0.3 log

units would achieve a 66 : 34% distribution of Cd(II) in

CeMT-1 and CeMT-2, and, given the presence of

fur-ther mechanisms, this sorting level may be sufficient to

ensure tolerable management of both Zn(II) and Cd(II)

in C elegans Our simplistic model demonstrates that,

even though both MTs show an overall preference for

Cd(II) over Zn(II), as expected for predominantly

thio-late coordination, the decrease in affinity of Cd for

CeMT-1 may allow segregation of Cd(II) into

predo-minantly CeMT-2 in a common cytosol

Discussion

Like other soil-dwelling organisms, C elegans

nema-todes are constantly exposed to and ingest varying

lev-els of essential and toxic metal ions present in the

surrounding medium Consequently, such organisms

require mechanisms that capture and redistribute the

correct amounts of biologically essential metal ions

whilst preventing the accumulation of harmful levels of

toxic metal ions Within this framework, mechanisms

must exist that allow the cell to distinguish between

closely similar essential and toxic metal ions, such as

Zn(II) and Cd(II), respectively The predominating

Cd–S and Zn–O forms observed by X-ray absorption

analysis suggest that separate pathways exist for

traf-ficking of these two metal ions These pathways do not

appear to be MT-mediated, and the negligible effect on

in vivospeciation for either Cd(II) or Zn(II) in knockout

mutants has excluded the possibility that MTs function

as metal storage proteins in C elegans In contrast, the

reduced accumulation, or excretion, of cadmium and

zinc is MT-mediated, as there was a large effect on the

levels of accumulated zinc and cadmium when CeMT-1

and CeMT-2 were deleted We interpret this observation

as an indication that some processes, possibly excretion

of excess zinc and cadmium, do not function normally

in the double knockout strain Furthermore, and most

importantly, the extents to which these

MT-mediated processes are disrupted are isoform- and

metal-ion specific We have shown that CeMT-2 plays a

more significant role in preventing hyperaccumulation

of cadmium Conversely, both CeMT-1 and CeMT-2

are important in maintaining physiologically acceptable

zinc levels, and the lack of CeMT-1 had a more

deleteri-ous effect These metal-specific preferences at the

cellu-lar level are mirrored in the relative affinities of the

individual CeMT-1 and CeMT-2 proteins towards

Zn(II) and Cd(II) The thermodynamic data suggest

that, when presented with both MT isoforms, cadmium

ions preferentially bind to 2, thus leaving

CeMT-1 to deal with zinc The origin of this differential affinity

is most likely rooted in the structure of the two isoforms

It is conceivable that the differences in specificity are, at least to a considerable extent, associated with the four additional metal ligands in CeMT-1, particularly the his-tidine residues (see Fig S1A) Previous studies on both zinc fingers [46] and metallothioneins [47–50] have dem-onstrated that an increasing number of histidine residues

in a metal binding site shifts the preference towards Zn(II) Further studies, including determination of 3D structures for CeMT-1 and CeMT-2, are required to determine the precise cause of the observed metal speci-ficities

In conclusion, the nematode C elegans exhibits both MT-mediated and non-MT-mediated pathways to deal with cadmium and zinc We have shown for the first time that the responses to cadmium and zinc ions at the cellular level are isoform-specific, and that this specificity is reflected at the protein level

Experimental procedures Cloning of MT constructs

Total RNA was isolated from nematodes using TRI reagent (Sigma, St Louis, MO, USA) and reverse tran-scribed into cDNA from 1 lg RNA using oligo(dT) primers and MMLV reverse transcriptase (Stratagene, La Jolla,

CA, USA), all according to the supplier’s protocols

cDNA using isoform-specific primers containing SalI and NdeI restriction site extensions (mtl-1_fwd: 5¢-TATACAT ATGGCTTGCAAGTGTGACTGC-3¢; mtl-1_rev: 5¢-AGC TTGTCGACGTTAATGAGCCGCAGCAGTTCCC-3¢;

GC-3¢ and mtl-2_rev: 5¢-AGCTTGTCGACGTTAATGA GCAGCCTGAGCACAT-3¢), generating DNA fragments

of 247 and 211 bp for isoform 1 and isoform 2, respec-tively The purified PCR products, as well as the plasmid pET29a, were digested using SalI and NdeI (Promega,

trans-formed into DH5a-competent cells (Invitrogen, Carlsbad,

CA, USA) and positive clones were identified by PCR screening The identity of the insert was confirmed by sequencing both strands of the cloned inserts

In vitro protein expression and purification

Plasmids containing the respective metallothionein isoform were transformed into E coli Rosetta TM2 (DE3)pLysS (Merck, Nottingham, UK) using standard molecular clon-ing techniques Expression cultures (1 L) selective for

respectively) were induced using isopropyl

final concentration of 500 lm Protein expression was

centrifugation at 5000 g Cell pellets were resuspended in

ice-cold sonication buffer (50 mm Tris⁄ Cl, 0.1 m KCl,

by sonication This was followed by centrifugation at

45 000 g for 45 min to remove cell debris The resulting

lysate was subjected to a chemical fractionation similar

to that described by You et al [32] Briefly,

dropwise with continuous stirring The mixture was

centri-fuged for 5 min at 5000 g A further three volumes of the

resulting supernatant, and this mixture was stored

centrifu-gation at 5000 g, resuspended in 20 mm ammonium

HiLoad 75 Superdex prep grade, GE Healthcare, Little

Chalfont, UK) MT-containing fractions were pooled and

concentrated by ultrafiltration (Amicon Ultra; Millipore,

Billerica, MA, USA) The isolated proteins either retained

or did not retain the N-terminal methionine The cleavage

efficiency of the E coli Met aminopeptidase appeared to be

dependent on the metal ion supplied, such that MTs

expressed in the presence of Cd(II) mostly retained the

initi-ation methionine

Preparation of Cd7-CeMT-1

based on the method reported by Vasˇak [33] Briefly, an

was incubated at room temperature with dithiothreitol

(approximately 10 mm) for 1 h This mixture was acidified

to a pH of approximately 1 using 2 m HCl, and applied to

a gel filtration column (Sephadex G25, PD10, Amersham

Biosciences) The demetallated protein was eluted under

equiva-lents) was added to the eluate, and the pH was increased to

> 7.0 via addition of 2 m Tris base Extensive washing by

ultrafiltration ensured removal of unbound metal ions

Mass spectrometry

All isoforms (20 lm) were buffer exchanged into 10 mm

ammonium acetate (pH 7.2) by ultrafiltration Prior to the

analysis, methanol was added to a final concentration of

performed using either ESI-TOF (MicrOTOF; Bruker,

Bremen, Germany) or ESI-ion trap (HCT-UltraTM Dis-covery System; Bruker) mass spectrometers Data were

range 500–3000 Th Using data analysis software supplied

by Bruker Daltonics, smoothing and baseline subtraction were applied to averaged data, which were subsequently deconvoluted

19F-NMR spectroscopy

A sample of each C elegans metalloform, approximately

480 lm with respect to metal ion concentration, was

determinations of metal ion content were performed using

[1,2-bis(2-amino-5-fluoro-phenoxy)ethane-N,N,N¢,N¢-tetraacetic acid; 4 mm final con-centration] was added to the sample, and incubated

19 F-NMR spectroscopy was performed using a DRX400 spectrometer (Bruker) fitted with a quadruple nuclei probe

Chemical shifts are reported with respect to the signal

spectral width of 50 p.p.m., an acquisition time of 3.48 s and a relaxation delay of 1.0 s, with 12 288 scans Fre-quency Induction Decay (FID)s were apodized using

65 536 complex data points, and baseline-corrected Spectra were processed using topspin version 2.1 software (Bruker

ionic strength (4 mm) as described by Hasler et al [34] to give a log K value of 11.75 Calculations of apparent stabil-ity constants for metal–MT complexes were performed using a published procedure [34]

Sample preparation for in vivo studies

Wild-type (N2) and the CeMT-2 knockout strain mtl-2 (gk125) were obtained from the Caenorhabditis Genetics Center (CGC) at the University of Minnesota, Minneapolis,

MN, USA, and the CeMT-1 knockout strain mtl-1 (tm1770) was obtained from the Mitani Laboratory at the Tokyo Women’s Medical University School of Medicine, Japan The metallothionein double knockout mtl-1;mtl-2 (zs1) was generated previously [25] Each strain was syn-chronized (bleach prepped), and 300 000–500 000 L1

were cultured per plate (90 mm diameter), and grown at

quench-frozen by immersion in liquid nitrogen

Metal quantification

Nematodes were digested in 1 n concentrated nitric acid,

and metal concentrations were quantified by inductively

coupled plasma optical emission spectrometry (ICP-OES)

using standard methods [52]

X-ray absorption spectra collection and analysis

The samples were ground to a fine powder under liquid

nitrogen, and stored as fully hydrated deep-frozen samples

K-edge (approximately 26 710 eV) and the zinc K-edge

(approximately 9660 eV) were collected on station 16.5 of

the Synchrotron Radiation Source (now closed) at the

Science and Technology Facilities Council Daresbury

Labo-ratory, Warrington, UK The ring operated at 2 GeV with

a mean current of 140 mA: the station was equipped with a

vertically focusing mirror and a flat Si (220) double crystal

monochromator detuned to 70% transmission to minimize

harmonic contamination The monochromator was

cali-brated at each energy value using a 15 lm cadmium foil or

a 10 lm zinc foil Data were collected with the station

operating in fluorescence mode using an Ortec 30 element

solid-state Ge detector The samples were mounted onto

aluminium sample holders, and X-ray absorption

spectros-copy measurements were performed at cryogenic

tempera-ture (approximately 20 K) using an Oxford Instruments

helium closed-cycle cryostat The standard samples were

prepared by grinding in an agate pestle and mortar, diluted

with boron nitride to give an edge step of approximately 1,

and mounted in 1 mm thick aluminium sample holders with

Sellotape windows Single scans were collected for the

model compounds in the transmission mode, and 16–23

scans were collected and summed for each experimental

sample Background subtraction and analysis of EXAFS

spectra were performed as described previously [36]

Acknowledgements

This work was supported by the Biotechnology and

Biological Sciences Research Council (BBSRC grant

BB⁄ E025099), the Science and Technology Facilities

Council (STFC grant BB⁄ E05099), an Altajir Trust

PhD studentship (to S.Z.-G.), and the Royal Society

(Olga Kennard Fellowship to C.A.B.) The X-ray

absorption spectroscopy was performed at the

Dares-bury Synchrotron Radiation Source (station 16.5),

managed and kindly assisted by Mr Bob Bilsborrow

We wish to acknowledge Dr Suresh Swain (King’s

College London) and Dr Samantha Hughes (King’s

College London, now at Oxford University) for

valu-able advice and resources provided throughout the

project, and finally the Caenorhabditis Genetics Centre

(CGC), which is funded by the National Institutes of Health National Centre for Research Resources, for the supply of Caenorhabditis elegans wild-type (N2) and mtl-2 (gk125) and Escherichia coli OP50, and the Mitani Laboratory at the Tokyo Women’s Medical University School of Medicine, Japan, for the supply

of mtl-1 (tm1770)

References

1 Van Straalen NM & Roelofs D (2008) Genomics technology for assessing soil pollution J Biol 7, 19

2 Timmermans MJTN, de Boer ME, Nota B, de Boer

TE, Marie¨n J, Klein-Lankhorst RM, Van Straalen NM

& Roelofs D (2007) Collembase: a repository for springtail genomics and soil quality assessment BMC Genomics 8, 341

3 Bundy JG, Sidhu JK, Rana F, Spurgeon DJ, Svendsen

C, Wren JF, Stu¨rzenbaum SR, Morgan AJ & Kille P (2008) Systems toxicology approach identifies coordi-nated metabolomic responses to copper in a terrestrial non-model invertebrate, the earthworm Lumbricus rubellus BMC Biol 6, 25

4 Owen J, Hedley A, Svendsen C, Wren JF, Jonker M, Hankard KP, Lister L, Stu¨rzenbaum SR, Morgan AJ, Spurgeon DJ et al (2008) Transcriptome profiling of developmental and xenobiotic responses in a keystone soil animal, the oligochaete annelid Lumbricus rubellus BMC Genomics 9, 226

5 Menzel R, Swain SC, Hoess S, Claus E, Menzel S, Steinberg CE, Reifferscheid G & Stu¨rzenbaum SR (2009) Gene expression profiling to characterize sediment toxicity – a pilot study using Caenorhabditis elegans whole genome microarrays BMC Genomics 10, 160

6 Galay-Burgos M, Spurgeon DJ, Weeks JM, Stu¨rzenbaum SR, Morgan AJ & Kille P (2003) Developing a new method for soil pollution monitoring using molecular genetic biomarkers Biomarkers 8, 229–239

7 Spurgeon DJ, Stu¨rzenbaum SR, Svendsen C, Hankard

PK, Morgan AJ, Weeks JM & Kille P (2004) Toxico-logical, cellular and gene expression responses in earth-worms exposed to copper and cadmium Comp Biochem Physiol C Toxicol Pharmacol 138, 11–21

8 Newman MC & McIntosh AW (1991) Metal Ecotoxi-cology: Concepts and Applications Lewis Publishers, Chelsea, MI

9 Rainbow PS & Dallinger R (1986) Strategies of metal detoxification in terrestrial invertebrates In

PS eds), pp 245–290 Lewis Publishers, Boca Raton, FL

10 Hamer DH (1986) Metallothionein Annu Rev Biochem

55, 913–951