Fish and molluscan metallothioneins A structural and functional comparison Laura Vergani 1 , Myriam Grattarola 1 , Cristina Borghi 2 , Francesco Dondero 3 and Aldo Viarengo 3,4 1 Department of Biophysical Sciences and Technologies, M. & O. University of Genova, Italy 2 Department of Biology, University of Genova, Italy 3 Department of Environmental & Life Science, University of Piemonte Orientale, Alessandria, Italy 4 Center on Biology and Chemistry of Trace Metals, University of Genova, Italy Metallothioneins (MTs) are cytosolic polypeptides found in almost all organisms, including vertebrates, invertebrates, plants and bacteria [1]. They do not appear to be essential for life, even though they are involved in many pathways, such as sequestration of toxic (Cd, Hg) or essential (Zn, Cu) metals, scavenging of oxyradicals, inflammation, and infection [2]. MTs exhibit unusual primary sequence, lacking histi- dines and aromatic residues, and their 3D structure is unique [3,4]. Cysteines represent one-third of the total amino acids and are distributed in typical motifs con- sisting of CC, CXC or CXYC sequences [5]. The beha- viour of MTs is dominated by the nucleophilic thiol group reacting with electrophilic compounds, including many alkylating agents and radical species [6]. Verteb- rate MTs have a monomeric dumbbell shape, com- posed of two globular domains connected by a flexible linker consisting of a Lys-Lys segment. Each domain contains a ‘mineral core’ enclosed by two large heli- cal turns of the polypeptidic chain. The N-terminal Keywords absorbance spectroscopy; circular dichroism; metal release; structure ⁄ function relationship; thermal stability Correspondence L. Vergani, Department of Biophysical Sciences and Technologies, M. & O. University of Genova, Corso Europa 30, 16132 Genova, Italy Fax: +39 010 3538346 Tel: +39 010 3538404 E-mail: Laura.Vergani@unige.it (Received 6 July 2005, revised 14 September 2005, accepted 26 September 2005) doi:10.1111/j.1742-4658.2005.04993.x Metallothioneins (MTs) are noncatalytic peptides involved in storage of essential ions, detoxification of nonessential metals, and scavenging of oxyradicals. They exhibit an unusual primary sequence and unique 3D arrangement. Whereas vertebrate MTs are characterized by the well-known dumbbell shape, with a b domain that binds three bivalent metal ions and an a domain that binds four ions, molluscan MT structure is still poorly understood. For this reason we compared two MTs from aquatic organ- isms that differ markedly in primary structure: MT 10 from the inverteb- rate Mytilus galloprovincialis and MT A from Oncorhyncus mykiss. Both proteins were overexpressed in Escherichia coli as glutathione S-transferase fusion proteins, and the MT moiety was recovered after protease cleavage. The MTs were analyzed by gel electrophoresis and tested for their differen- tial reactivity with alkylating and reducing agents. Although they show an identical cadmium content and a similar metal-binding ability, spectro- polarimetric analysis disclosed significant differences in the Cd 7 -MT secon- dary conformation. These structural differences reflect the thermal stability and metal transport of the two proteins. When metal transfer from Cd 7 - MT to 4-(2-pyridylazo)resorcinol was measured, the mussel MT was more reactive than the fish protein. This confirms that the differences in the pri- mary sequence of MT 10 give rise to peculiar secondary conformation, which in turn reflects its reactivity and stability. The functional differences between the two MTs are due to specific structural properties and may be related to the different lifestyles of the two organisms. Abbreviations MT, metallothionein; GST, glutathione S-transferase; PAR, 4-(2-pyridylazo)resorcinol. 6014 FEBS Journal 272 (2005) 6014–6023 ª 2005 The Authors. Journal compilation ª 2005 FEBS right-handed b domain binds three bivalent metal ions. The C-terminal a domain is left-handed and binds four bivalent ions. Zinc is preferentially located in the b do- main and cadmium in the a domain. Therefore, the b domain would regulate zinc and copper homeostasis, whereas the a domain may play a central role in heavy metal detoxification [7]. The loosely structured b do- main is responsible for metal-bridge dimerization, whereas the a domain is involved in oxidative dimeri- zation. Metal bridge dimerization is reversed by dilu- tion or addition of chelating agents, whereas oxidative dimers are reduced by reducing compounds [8]. As reported previously [9], oxidative dimerization may also occur in vivo under conditions of stress, such as exposure to toxic metals and reactive oxygen species and in neurological disorders (e.g. Alzheimer’s disease). A different susceptibility to oxidation may be import- ant for the physiological role of the protein. Despite the high homology among vertebrate MTs, fish and mammalian MTs exhibit significant differences at the level of primary structure, i.e. displacement of one cysteine and fewer lysines [10]. Compared with vertebrates, invertebrate MTs show unusual features in their primary structure. The sequences of only a few MTs from aquatic inverte- brates (crab, mussel, sea urchin, snail and oyster) have so far been elucidated [11–17]. As in mammals and fish [18], echinoderm MTs contain two globular domains binding four and three bivalent ions [19]. On the other hand, in crab (Scylla serrata and Cancer pagurus) the two domains bind three bivalent metals each [20–22]. In comparison with mammalians, molluscan MTs usu- ally have higher glycine content ( 15% in mussels), randomly distributed throughout the sequence. Despite the differences, molluscan MTs appear to be more clo- sely related to vertebrate MTs than those from other invertebrate phyla [23,24]. In this study we focused on two MTs from different aquatic organisms which we had widely investigated in previous work [25,26]: MT A from Oncorhyncus mykiss and MT 10 from Mytilus galloprovincialis were selected as representative of vertebrates and invertebrates, respectively. Although MT 10 is longer than MT A, both have a similar number of cysteine residues and identical cadmium content. Both recombinant MTs were tested for reactivity to alkylating and reducing agents, to evaluate their susceptibility to oxidative and metal-bridge dimerization. Secondary conformation was analyzed in both the metal-free protein and Cd 7 - MTs. After metal binding, significant differences between the two forms were observed. The altered sec- ondary structure influenced the physicochemical prop- erties of the proteins, with MT 10 being more thermostable than MT A. When the redox-induced metal transfer from Zn 7 -MT or Cd 7 -MT to the specific acceptor 4-(2-pyridylazo)resorcinol (PAR) was meas- ured, MT 10 was much more reactive in terms of cad- mium release. This observation is interesting because the redox control of metal bioavailability seems to be an important physiological function of MTs [27]. Results Analysis of primary sequence When the primary sequence of mussel MT 10 was compared with that of fish MT A (Fig. 1) with Needle- man–Wunsch global alignments [28], a low identity was observed (39%). Because the first extra amino acid number is similar in the two recombinant MTs, we assumed that they affect the two proteins in a similar way. Accordingly, experiments using atomic absorp- tion spectroscopy estimated 7 mol cadmium bound per mol recombinant MT in both samples. The b domain of MT A has nine cysteines distri- buted in classic Cys motifs. In MT 10, this domain is two residues longer, but it has only eight cysteines with a similar arrangement of the CXC motifs. Major dif- ferences between the two proteins occur at the level of the a domain, which is longer in MT 10 than in MT A (42 vs. 29 residues) and has two additional cysteines. Moreover, the cysteines are organized differently in Fig. 1. Sequence alignment of fish MT A and molluscan MT 10. The sequences of the two recombinant MTs were aligned with the program Needleman-Wunsch global alignments. This program uses the Needleman–Wunsch global alignment algorithm [28] to find the optimum alignment (including gaps) of two sequences when considering their entire length. L. Vergani et al. Comparison between two MTs from aquatic organisms FEBS Journal 272 (2005) 6014–6023 ª 2005 The Authors. Journal compilation ª 2005 FEBS 6015 terms of both the CXC or CXYC sequence and Cys motif arrangement. In summary, MT A has six CXC, one CXYC and four CXYWC sequences, whereas mussel MT 10 has nine CXC, one CXYC and five CXYWC sequences. In MT 10, the last a domain Cys motif is CXXXCC, instead of CXCC, which is typical of other vertebrates. This feature has been reported in the MT from the Antarctic fish Notothenia coriiceps [18], but not in mussels. In conclusion, this comparat- ive analysis localizes the major differences between the fish and molluscan MTs at the level of the a domain. Because this cluster is mainly involved in cadmium binding and oxidative dimerization, these differences may reflect functional differences in the two MTs with regard to Cd release and oxidation. A further difference is the reduced number of lysines in MT 10 compared with MT A (5 vs. 7), correspond- ing to 6.8% of the amino-acid composition for the mussel protein and 11.5% for the fish one. In spite of the fewer lysines, MT 10 contains one more CK motif (5 vs. 4), whereas mammalian MTs have seven. The arrangement of the CK motifs also differs between mussel and fish proteins: the four CK motifs of MT A are equally distributed between the a and b domain, whereas in MT 10, four are at the C-terminus and only one at the N-terminus. As previously reported [29], the hydropathic index (a parameter that is inversely proportional to flexibility) is lower in fish MTs than in mammalian MTs. A higher flexibility should facilitate conforma- tional changes in organisms living at low temperatures. When the hydropathic index was calculated for the trout MT A and the mussel MT 10 using the protparam tool [30], MT A yielded a negative value ()0.110), similar to that recorded for N. coriiceps [29], whereas MT 10 gave a positive value (0.199) consis- tently higher than that for mammalian MTs (0.098). This points to mussel MT having a lower flexibility than either the fish or mammalian counterparts. Oxidative and metal bridge polymerization After chromatographic purification and enzymatic removal of the glutathione S-transferase (GST) tail, proteins were analyzed by SDS ⁄ PAGE (15% gel). As expected, MT A showed a lower molecular mass than MT 10, but also more marked smearing at high molecular mass than MT 10. This effect is due to the presence of polymeric forms typical of native MTs (Fig. 2). When both MTs were alkylated with N-ethyl- maleimide, a unique band at a lower molecular mass appeared, representing the monomeric form, and no differences in mass between the two MTs could be observed. Moreover, alkylation of the thiol group of MT A resulted in disappearance of the smearing at high molecular mass. A similar effect on the aggre- gates was observed when MT A was reduced with dithiothreitol, which caused the appearance of a single AB Fig. 2. Electrophoretic comparisons between fish MT A and mussel MT 10. MT A (A) and MT 10 (B) were electrophoresed on SDS ⁄ 15% polyacrylamide gel before (lane 1) and after the addition of an alkylating agent (N-ethylmaleimide) at two different concentrations: 40 and 80 m M (lanes 2 and 3) for 3 h. MTs were also treated for the same period with a reducing agent (dithiothreitol) at 40 and 80 mM (lanes 4 and 5). To reduce aggregation, MT samples were handled in anaerobic conditions under nitrogen atmosphere. Molecular markers (lane M) from the top: BSA, 66 kDa; chicken egg ovalbumin, 45 kDa; bovine chymotrypsinogen, 25 kDa; lysozyme, 14.3 kDa; ribonuclease A, 11.9 kDa; bovine lung aprotinin 6.5 kDa. Comparison between two MTs from aquatic organisms L. Vergani et al. 6016 FEBS Journal 272 (2005) 6014–6023 ª 2005 The Authors. Journal compilation ª 2005 FEBS band at  12 kDa, corresponding to the dimeric form of the protein. In contrast, no changes occurred in MT 10 when exposed to the reducing agent, indicating lower susceptibility of the mussel protein to oxidation. These results suggest that the smearing at high molecu- lar mass (12–18 kDa) is due to oxidative polymeriza- tion of the MT molecules, whereas dimer formation ( 12 kDa) is probably due to metal bridge effects. The marked differences between the a domain sequences of the two MTs described above is in line with the reduced sensitivity to oxidation exhibited by MT 10, as oxidative polymerization occurs mainly in the a domain. Characterization of structure and metal binding UV absorption spectra of both MTs were recorded on addition of increasing equivalents of cadmium ions to the metal-free apoprotein, at neutral pH. After Cd(II) titration, a shoulder peak appeared at 254 nm, reflect- ing the charge-transfer interaction of the cadmium– thiolate clusters. In both curves (Fig. 3) absorption at 254 nm increased steadily, until saturation was reached at seven metal equivalents. The slope of the curve was almost identical for the two MTs, indicating no signifi- cant differences with respect to cadmium-binding prop- erties. This agrees with the data acquired by atomic absorption spectroscopy, which estimated 7 mol cad- mium bound per mol MT. On analysis of the CD spectra of the MTs, both metal-free thioneins showed a strong negative band at  230 nm (Fig. 4), typical of proteins in random coil conformation [31]. This confirms that both apo-MTs were unfolded in the absence of metals, and only after binding of the correct number of cadmium equivalents did they assume a stable secondary structure. When complexed to the metal, both MTs showed a strong positive ellipticity band above 250 nm, but the peak was red-shifted in the MT 10 spectrum compared with that of MT A. The major differences were evident in the region below 250 nm. In fact, both the negative band at 245 nm and the positive one at 228 nm, char- acteristic of the fish MT A, were lost in the MT 10 spectrum. Considering the spectral peculiarities, we can infer that mussel MT 10 has an atypical secondary conformation, which is probably due to the differences in primary sequence. Fig. 3. Spectrophotometric titration following the binding of Cd(II) to the apo-MTs. The Cd-induced contribution to the absorption spectrum at 254 nm is plotted against the number of Cd equiva- lents added, from 0.3 to 8 ratio for both fish MT A (m) and mussel MT 10 (n). Each curve is representative of at least three independ- ent sets of measurements. Fig. 4. CD analysis. CD spectra were acquired in the near-UV region (from 190 to 290 nm) for fish MT A (A) and molluscan MT 10 (B) for both Cd 7 -MT forms and the apoproteins. The metal- free protein was obtained by acidification with HCl. The measure- ments were performed on three different MT preparations. L. Vergani et al. Comparison between two MTs from aquatic organisms FEBS Journal 272 (2005) 6014–6023 ª 2005 The Authors. Journal compilation ª 2005 FEBS 6017 Thermal stability We examined whether the differences in primary and secondary structure affected the thermal stability of the two MTs. UV absorption spectra were acquired for fish and mussel Cd 7 -MTs, after exposure to a ther- mal gradient. A 254 was plotted as a function of tem- perature. As expected, in both cases we observed a decrease in percentage absorbance with temperature increase (Fig. 5). The absorbance of fish MT A declined steadily starting at 30 °C, with a marked change in slope above 50 °C, similar to the description of D’Auria et al [18]. The thermal profile of MT 10 showed a similar trend, but the slope change occurred at a higher temperature (above 60 °C). Moreover, MT 10 maintained higher percentage absorbance at all temperatures than the fish protein (0.7 vs. 0.5 at 90 °C, respectively). These results suggest that the mussel MT is much more thermostable at high temperatures than the fish protein. This is in line with the greater rigidity suggested by the hydropathic index. Kinetics of metal release Cysteine residues can be oxidized in vitro by mild cellu- lar oxidants and release metals during the process [32–34]. It has been suggested that oxidoreductive mechanisms may also modulate in vivo the affinity of cysteines for metal ions and regulate the bioavailability of bivalent metals [35]. In the presence of the glutathione redox couple (GSH ⁄ GSSG), we observed zinc release from both recombinant MTs. The kinetics of this process were similar early on, but, at saturation, MT 10 seemed to release slightly more zinc than MT A (Fig. 6A). The difference in metal-releasing ability was much more evident when the Cd-complexed MTs were assayed in the presence of the H 2 O 2 redox partner. Because MTs have a higher affinity for cadmium than for zinc (typically, K d ¼ 5Æ10 )12 m for zinc and K d ¼ 5Æ10 )16 m for cadmium), a stronger oxidizing agent such as H 2 O 2 was needed to detach cadmium ions [32,36]. Cadmium release was much more marked for MT 10 than for MT A (Fig. 6B). These data point to a pronounced reactivity of the metal–thiolate clusters in the mussel MT 10, which Fig. 5. Thermal stability of fish and mussel Cd 7 -MTs. Absorption UV spectra were acquired for fish MT A (m) and mussel MT 10 (n) as a function of the temperature increase from 20 to 90 °C. The absorbance decrease at 254 nm was reported as a fraction of the standard absorbance (absorbance at room temperature) in order to compare the denaturation profile of the Cd–thiolate chromophore of the two MTs. Each curve is representative of four independent sets of measurements. Fig. 6. Kinetics of zinc and cadmium release from recombinant MTs. The metal release was followed by the formation of metal– (PAR) 2 complex at 500 nm. Each experimental point represents the difference between the absorbance measured in the presence and absence of the appropriate redox couple: GSH ⁄ GSSG for zinc and H 2 O ⁄ H 2 O 2 for cadmium. We measured (A) the kinetics of zinc release and (B) the kinetics of cadmium release for fish MT A (m) and mussel MT 10 (n). Each curve is representative of at least three independent sets of measurements. Comparison between two MTs from aquatic organisms L. Vergani et al. 6018 FEBS Journal 272 (2005) 6014–6023 ª 2005 The Authors. Journal compilation ª 2005 FEBS releases zinc and cadmium more quickly and effectively than fish MT A. The kinetic observations of a more pronounced release of cadmium than zinc fits well with our data indicating that the major structural differ- ences are at the level of the a domain, which is in fact responsible for cadmium binding. Discussion MTs are noncatalytic metalloproteins, the physiologi- cal function of which is not yet fully understood. The moderate variability of this class of proteins across phylogenetically distant organisms reflects the highly conserved function that they exert in living systems. In contrast, the specific environmental requirements explain the existence of numerous isoforms in the same organism. A comparative analysis of the functional and structural features of MTs from different organ- isms may help to clarify their physiological role. Usually, vertebrate MTs contain 61–62 amino-acid residues, whereas larger chains with 72–74 residues are found in molluscs and in nematodes, and shorter chains have been reported in insects and fungi [37]. In this report, we describe the production and char- acterization of two MTs from evolutionary distant aquatic organisms, the fish MT A and the mussel MT 10. Whereas much information is available for fish MTs [18,38], characterization of mussel MTs has been a problem until now [39], which we have overcome by using the recombinant protein. When MT 10 from M. galloprovincialis was com- pared with MT A from O. mykiss, the major finding was a difference in their primary sequence, mainly at the level of the a domain. These differences suggest that the mussel protein a domain is larger than the b domain. We therefore postulated an asymmetric dumbbell shape for MT 10 and different behaviour in terms of cadmium release and oxidative dimerization, which occur in this region. As expected, we found that oxidative dimerization was less marked in the mussel MT 10 than in the fish MT A. When the kinetics of metal release were investigated, MT 10 showed more pronounced reactivity than MT A. We wish to empha- size that this higher mobility was more marked for cadmium than for zinc. Accordingly, the major struc- tural differences are concentrated only in the domain binding this toxic metal. The different reactivity can be attributed to a different spatial arrangement of the mercaptide bonds, altering their accessibility to oxid- izing agents. Marked differences between the two proteins appeared also at the level of their secondary conforma- tion. The CD spectrum of MT 10 lacked both the 245 nm negative and 228 nm positive bands that are typical of vertebrate MTs. We hypothesized that these striking differences in the CD spectra are due mainly to the lysine residues, which are highly conserved in ver- tebrate MTs, but not in mussel. The lower number of lysine residues in MT 10 than in MT A (6.8% vs. 11.5%) may also explain the increased ability of the mussel protein to release metals. The increased mobility of cadmium and zinc of MT 10 may be due to a weaker metal–thiolate interaction because of the reduced num- ber of lysines. In fact, substitution of three lysines with glutamates in the CK motifs of the a domain modified the metal-binding ability of MT [40]. Finally, the mussel MT 10 showed greater thermal stability than the fish protein, probably because of its longer polypeptide chain. Moreover, MT 10 has a pos- itive hydropathic index (0.199), whereas fish MTs are usually characterized by a negative value ()0.110 for trout MT A). As a higher hydropathic index means lower flexibility, this feature may explain the higher thermal stability of MT 10. This is confirmed by 2D NMR spectroscopy data. A preliminary analysis of 2D homonuclear ( 1 H) NOESY spectra, acquired for both proteins, indicates a more rigid structure for MT 10 than for MT A, with both the number of NOE peaks and signal spread being greater in the former (Fig. 7). All the above data led us to conclude that the mus- sel MT is different, in terms of spatial conformation and functional properties, from vertebrate MTs, even if the cadmium content is identical. The higher metal mobility and rigidity exhibited by MT 10 is probably related to the environment inhabited by mussels, which are subjected to sudden changes in environmental vari- ables (temperature, anoxia, concentration of aquatic pollutants). The modified a domain, which plays a role in detoxification ⁄ sequestering of toxic metals (e.g. cad- mium), would allow adaptation to the requirements of these aquatic organisms. Resolution of the 3D structure of MT 10 at the atomic level will allow us to clarify the structural fea- tures supporting the observed different reactivity. For both MTs, besides 2D homonuclear ( 1 H) NOESY spec- tra, 2D heteronouclear ( 113 Cd) NMR spectra have also been acquired, and data processing is in progress. How- ever, from a comparison of the raw 2D homonuclear ( 1 H) NOESY spectra, the differences between the two proteins have already been confirmed (Fig. 7). When spectra are compared in the same chemical-shift win- dow, a greater number of NOE peaks and signal spread is palpable in the MT 10 sample, providing clear evi- dence of the difference in the level of structural organ- ization. The more the spectrum is ‘crowded’ and the wider the chemical-shift range over which the signal is L. Vergani et al. Comparison between two MTs from aquatic organisms FEBS Journal 272 (2005) 6014–6023 ª 2005 The Authors. Journal compilation ª 2005 FEBS 6019 spread, the more the structure can be assumed to be well defined, therefore these raw data confirm that MT 10 has a better defined and stable structure than MT A. Experimental procedures Materials Chemicals and molecular mass markers were supplied by Sigma Aldrich (Milan, Italy). Reagents for bacterial growth were purchased from Fluka (Milan, Italy). T4 DNA ligase and Taq polymerase were from Stratagene (La Jolla, CA, USA), and restriction enzymes and dNTPs from Promega Italia (Milan, Italy). Expression vector pGEX 6P-1, E. coli strains BL21 and JM109, precision protease and glutathi- one–Sepharose 4B matrix were purchased from Amersham Biosciences (Uppsala, Sweden). Primers for sequencing and mutagenesis were synthesized by TibMolBiol (Genoa, Italy). Cloning and amplification of MTs The coding sequence of the O. mykiss MT A gene [41] was a gift from Professor P. E. Olsson (Umea University, Umea, Sweden) and was cloned as previously described [42]. Recombinant molluscan MT 10 from M. galloprovin- cialis (NCBI GeneBank database accession number AY566248) was prepared starting from the 222-bp coding sequence, previously cloned by our group. By PCR we added a BamHI site upstream from the ATG codon, using the 5¢-end primer (5¢-CTACTACGAATTAGGATCCCCT GCACCTTG-3¢) and the 3¢-end primer (5¢-GTAATACGA CTCACTATAGGGCGAATTGGG-3¢). Amplification was performed as previously described [42]. The PCR fragment was eluted from gel using the NUCLEOSPIN-EXTRACT MN kit (Du ¨ ren, Germany) and subcloned into the expres- sion vector pGEX-6P-1. Both recombinant MTs were syn- thesized as fusion proteins, with a GST tail at the N-terminus. After enzymatic removal of the GST, MT 10 had four additional amino acids (Gly-Pro-Leu-Gly) with respect to the wild-type, with the initial Met substituted with a Ser (Fig. 1). The sequence of the recombinant vector and the correct orientation of the cDNA were checked by sequencing it in both directions using the appropriate pGEX primers (Amersham Biosciences). Bacterial expression and purification Large-scale expression was carried by inoculating 12.5 mL Luria–Bertani medium containing 100 lgÆmL )1 ampicillin and growing the cells at 37 °C overnight with vigorous sha- king. Then 1 L prewarmed 2XYT medium (16 gÆL )1 tryp- tone, 10 gÆ L )1 yeast extract, 5 gÆ L )1 NaCl, 100 lgÆmL )1 ampicillin) was inoculated with 10 mL of the overnight culture and grown until mid-exponential growth phase. To MTA MT 10 Fig. 7. Comparison of the whole 2D-NOESY spectra of fish MT A and mussel MT 10. The 2D nuclear Overhauser enhancement spectra (2D-NOESY) were acquired on a Bruker Advance 600 MHz spectrometer (Rheimstetten, Germany) using 2 m M solutions of the proteins in 95% H 2 O, 5% 2 H 2 Oor 2 H 2 OatpH 7.0 under a nitrogen atmosphere. Spectra are shown in the same chemical-shift window. Comparison between two MTs from aquatic organisms L. Vergani et al. 6020 FEBS Journal 272 (2005) 6014–6023 ª 2005 The Authors. Journal compilation ª 2005 FEBS overexpress the recombinant protein, we added isopropyl b-d-thiogalactopyranoside to a final concentration of 0.5 mm. The highest level of nondegraded MT was observed after 5 h of growth at 30 °C. For preparing Me 7 -MTs, 0.2 mm CdCl 2 was added to the culture medium, or alter- natively Zn 7 -MT the same concentration of ZnCl 2 . Recombinant MTs were purified by affinity chromatography using glutathione–Sepharose 4B matrix to selectively bind the GST tag of the fusion protein. The expression showed an average yield higher than 1 mgÆL )1 of culture. The bacterial pellet was resuspended in cold NaCl ⁄ P i (140 mm NaCl, 2.7 mm KCl, 10 mm Na 2 HPO 4 , 1.8 mm KH 2 PO 4 , pH 7.3) and lysed by mild sonication at 4 ° C. After addition of 1% Triton X-100, the suspension was mixed gently at 4 °C for 30 min, and the supernatant was mixed for 30 min with 2 mL 50% slurry resin previously equilibrated. Recombinant MT was recovered by enzymatic cleavage using ‘Prescission Protease’ (120 UÆmL )1 resin) to selectively remove the GST tail. Digestion was carried out at 4 °C for 16 h directly on the column equilibrated with digestion buffer (50 mm Tris ⁄ HCl, pH 7, 150 mm NaCl, 1 mm dithiothreitol) [42]. SDS/PAGE electrophoresis and metal quantification At each step of the purification procedure, the presence of the recombinant MT was checked by electrophoresis on 12.5% polyacrylamide gel, performed according to the clas- sical method of Laemmli [43]. Because of the small dimen- sions and the physicochemical features of MTs, the best resolution was obtained by 16% Tris ⁄ Tricine SDS ⁄ PAGE [44]. The cadmium content in the recombinant MTs was determined with a polarized Spectra AA 558 atomic absorption spectrophotometer (Varian, Torino, Italy). The number of molecules of cadmium bound per molecule of MT was determined using as a standard curve constructed using a standard solution of cadmium chloride. Both fish and mussel recombinant MTs contain 7 equivalents of cad- mium per mol protein. Protein quantification At each step of the purification, total proteins were quanti- fied by the Bradford assay [45], with BSA as standard. At the end of the purification, MT was quantified by measur- ing the absorbance of the metal-free protein at 220 nm in 0.1 m HCl using e 220 ¼ 47 300 m )1 Æcm )1 [46]. Although the absorption coefficient of molluscan apo-MT should be higher once the amino-acid content is higher (73 vs. 61 resi- dues), the protein concentration was calculated using the absorption coefficient of vertebrates [24]. Alternatively MT was quantified by estimating the -SH groups using Ellmans’ reagent in potassium phosphate buffer (2 m NaCl in 0.2 m potassium phosphate, pH 8), using the absorption coeffi- cient e 412 ¼ 13 600 m )1 Æcm )1 [47]. Absorption and CD spectroscopy Absorption spectra were acquired after resuspending each recombinant MT (0.025 mgÆmL )1 )in5mm Tris⁄ HCl (pH 7) ⁄ 100 mm NaCl. UV spectra were recorded in the wavelength range 200–300 nm, using a Jenway 6505 spec- trophotometer (Felsted Dunmow, Essex, UK), both in standard conditions and after exposure to a linear thermal gradient (25–90 °C). A broad absorption shoulder occurred near 250 nm when thionein binds cadmium. To analyse the formation of the metal–thiolate clusters, we subjected fish and molluscan MTs to titration with bivalent metals (zinc and cadmium): 0.025 mgÆmL )1 each protein was resus- pended in 5 mm Tris ⁄ HCl (pH 7.5) ⁄ 100 mm NaCl ⁄ 1mm dithiothreitol and the spectra were recorded in the range 220–300 nm at increasing metal ⁄ protein ratios [38]. CD spectra were recorded on a Jasco J-710 spectropola- rimeter (Jasco, Tokyo, Japan) calibrated with a standard solution of (+)-10-camphosulfonic acid. All spectra were recorded in a 0.05-cm path-length quartz cell, using the following parameters: time constant 4 s, scanning speed 20 nmÆmin )1 , band width 2 nm, sensitivity 10 millidegrees, step resolution 0.5 nm [48]. Photomultiplier high voltage did not exceed 600 V in the spectral region under analysis, and the absorbance never exceeded 1.0. Each spectrum was an average of five scans over 290–190 nm. Protein concen- tration was kept below 0.1 mgÆmL )1 in 5 mm Tris ⁄ HCl (pH 7) ⁄ 100 mm NaCl. All the acquired spectra were correc- ted for the baseline and normalized to the amino-acid con- centration, in order to calculate the mean residual molar ellipticity (degreesÆcm )2 Ædecimol )1 ). All experiments were performed in strictly anaerobic conditions, by purging high- grade nitrogen in the sample chamber. To characterize the secondary structure of the two proteins, the acquired CD spectra were analyzed by dedicated software [49–51]. Kinetics of metal release Zinc and cadmium release were estimated spectrophotomet- rically by following the formation of the metal–PAR com- plex at 500 nm. For zinc kinetics, the MT samples (1.3 lm protein) were resuspended in 0.2 m Tris ⁄ HCl, pH 7.4, and incubated with 100 lm PAR in the absence or presence of 1.5 mm GSH ⁄ 3mm GSSG [32]. Cadmium mobility was tested by measuring its transfer from Cd 7 -MT to PAR induced by the presence of the H 2 O 2 ⁄ H 2 O redox couple [33]. Samples of 4.6 lm MT were added to the reaction buf- fer (100 lm PAR, 50 mm Tris ⁄ HCl, pH 7.4) in the absence or presence of 1 mm H 2 O 2 . Acknowledgements We would like to extend our gratitude to Professor Gabriella Gallo for her scientific collaboration and L. Vergani et al. Comparison between two MTs from aquatic organisms FEBS Journal 272 (2005) 6014–6023 ª 2005 The Authors. Journal compilation ª 2005 FEBS 6021 Dr Mara Carloni for her experimental contributions. We thank Dr Giuseppe Digilio (Bioindustry Park del Canavese spa, Ivrea, Italy) and Professor Mauro Botta for NMR spectra. This research was supported by a grant from the National Research Council (within the program ‘Biomolecules for Human Health’) and from the University of Genova Project for the year 2002. This work was also supported by the 5th UE Framework Program project Biological Effects of Environmental Pollution in marine coastal ecosystems (BEEP) (contract No. EVK3-2000-00543). References 1 Hamer DH (1986) Metallothionein. Annu Rev Biochem 55, 913–951. 2 Coyle P, Philcox JC, Carey LC & Rofe AM (2002) Metallothionein: the multipurpose protein. Cell Mol Life Sci 59 , 627–647. 3 Braun W, Wagner G, Worgotter E, Vaa ´ kM,Ka ¨ gi JHR &Wu ¨ thrich K (1986) Polypeptide fold in the two metal clusters of metallothionein-2 by nuclear magnetic reso- nance in solution. J Mol Biol 187, 125–129. 4 Furey WF, Robbins AH, Clancy LL, Winge DR, Wang BC & Stout CD (1986) Crystal structure of Cd,Zn metallothionein. Science 231, 704–710. 5Ka ¨ gi JHR, Vaa ´ k M, Lerch K, Gilg DE, Hunziker P, Bernhard WR & Good M (1984) Structure of mammalian metallothionein. Environ Health Perspect 54, 93–103. 6 Romero-Isart N & Vaa ´ k M (2002) Advances in the structure and chemistry of metallothioneins. J Inorg Biochem 88, 388–396. 7 Zhou Y, Li L & Ru B (2000) Expression, purification and characterization of beta domain and beta domain dimer of metallothionein. Biochim Biophys Acta 1524, 87–93. 8 Zangger K & Armitage IM (2002) Dynamics of inter- domain and intermolecular interactions in mammalian metallothioneins. J Inorg Biochem 88, 135–143. 9 Zangger K, Shen G, Oz G, Otvos JD & Armitage IM (2001) Oxidative dimerization in metallothionein is a result of intermolecular disulphide bonds between cysteines in the alpha-domain. Biochem J 359, 353–360. 10 Scudiero R, Carginale V, Riggio M, Capasso C, Capasso A, Kille P, Di Prisco G & Parisi E (1997) Difference in hepatic metallothionein content in Antarctic red-blooded and haemoglobinless fish: undetectable metallothionein levels in haemoglobinless fish is accom- panied by accumulation of untranslated metallothionein mRNA. Biochem J 322, 207–211. 11 Lerch K, Ammer D & Olafson RW (1982) Crab Metal- lothionein primary structures of metallothioneins 1 and 2. J Biol Chem 257, 2420–2426. 12 Mackay EA, Overnell J, Dunbar B, Davidson I, Hunzi- ker PE, Ka ¨ gi JH & Fothergill JE (1993) Complete amino acid sequences of five dimeric and four mono- meric forms of metallothionein from the edible mussel Mytilus edulis. Eur J Biochem 218, 183–194. 13 Wang Y, Mackay EA, Kurasaki M & Ka ¨ gi JH (1994) Purification and characterisation of recombinant sea urchin metallothionein expressed in Escherichia coli. Eur J Biochem 225, 449–457. 14 Berger B, Dallinger R, Gehrig P & Hunziker PE (1997) Primary structure of a copper-binding metallothionein from mantle tissue of the terrestrial gastropod Helix pomatia L. Biochem J 328, 219–224. 15 Barsyte D, White KN & Lovejoy DA (1999) Cloning and characterization of metallothionein cDNAs in the mussel Mytilus edulis L. digestive gland. Comp Biochem Physiol C Pharmacol Toxicol Endocrinol 122, 287–296. 16 Tanguy A, Mura C & Moraga D (2001) Cloning of a metallothionein gene and characterization of two other cDNA sequences in the Pacific oyster Crassostrea gigas (CgMT1). Aquat Toxicol 55 , 35–47. 17 Narula SS, Brouwer M, Hua Y, Armitage IM (1995) Three-dimensional solution structure of Callinectes sapi- dus metallothionein-1 determined by homonuclear and heteronuclear magnetic resonance spectroscopy. Bio- chemistry 34, 620–631. 18 D’Auria S, Carginale V, Scudiero R, Crescenzi O, Di Maro D, Temussi PA, Parisi E & Capasso C (2001) Struc- tural characterization and thermal stability of Northenia coriiceps metallothionein. Biochem J 354, 291–299. 19 Riek R, Precheur B, Wang Y, Mackay EA, Wider G, Guntert P, Liu A, Ka ¨ gi JH & Wu ¨ thrich K (1999) NMR structure of the sea urchin (Strongylocentrotus purpura- tus) metallothionein MTA. J Mol Biol 291, 417–428. 20 Otvos S, Olafson RW & Armitage IM (1982) Structure of an invertebrate metallothionein from Scylla serrata. J Biol Chem 257, 2427–2431. 21 Hunt CT, Boulanger Y, Fesik SW & Armitage IM (1984) NMR analysis of the structure and metal seques- tering properties of metallothioneins. Environ Health Perspect 54, 135–145. 22 Overnell J, Good M & Vaa ´ k M (1988) Spectroscopic studies on cadmium (II) - and cobalt (II)-substituted metallothionein from the crab Cancer pagurus. Evidence for one additional low-affinity metal-binding site. Eur J Biochem 172, 171–177. 23 Dallinger R, Berger B, Hunziker PE, Birchler N, Hauer CN & Ka ¨ gi JH (1993) Purification and primary struc- ture of snail metallothionein. Similarity of the N-term- inal sequence with histones H4 and H2A. Eur J Biochem 216, 739–746. 24 Simens DC, Bebianno MJ & Moura JJ (2003) Isolation and characterisation of metallothionein from the clam Ruditapes decussates. Aquat Toxicol 63, 307–318. Comparison between two MTs from aquatic organisms L. Vergani et al. 6022 FEBS Journal 272 (2005) 6014–6023 ª 2005 The Authors. Journal compilation ª 2005 FEBS 25 Viarengo A, Ponzano E, Dondero F & Fabbri R (1997) A simple spectrophotometric method for metallothio- nein evaluation in marine organism: an application to mediterranean and antarctic molluscs. Mar Environ Res 44, 69–87. 26 Viarengo A, Burlando B, Ceratto N & Panfoli I (2000) Antioxidant role of metallothioneins: a comparative overview. Cell Mol Biol 46, 407–417. 27 Maret W (1994) Oxidative metal release from metallo- thionein via zinc-thiol ⁄ disulfide interchange. Proc Natl Acad Sci USA 91, 237–241. 28 Needleman SB & Wunsch CD (1970) A general method applicable to the search for similarities in the amino acid sequence of two proteins. J Mol Biol 48, 443–453. 29 Scudiero R, Temussi PA & Parisi E (2005) Fish and mammalian metallothioneins: a comparative study. Gene 345, 21–26. 30 Gasteiger E, Hoogland C, Gattiker A, Duvaud S, Wil- kins MR, Appel RD & Bairoch A (2005) Protein identi- fication and analysis tools on the EXPASY server. In The Proteomics Protocols Handbook (Walker JM, ed.), pp 571–607. Humana Press, Totowa, NJ. 31 Willner H, Vaa ´ kM&Ka ¨ gi JHR (1987) Cadmium-thio- late clusters in metallothionein: spectrophotometric and spectropolarimetric features. Biochemistry 26, 6287– 6292. 32 Maret W, Vallee B & L (1998) Thiolate ligands in metallothionein confer redox activity on zinc cluster. Proc Natl Acad Sci USA 95, 3478–3482. 33 Quesada AR, Byres RW, Krezoski SO & Petering DH (1996) Direct reaction of H 2 O 2 with sulfhydryl groups in HL-60 cells: zinc-metallothionein and other sites. Arch Biochem Biophys 334, 241–250. 34 Jacob C, Maret W & Vallee BL (1998) Control of zinc transfer between thionein, metallothionein, and zinc proteins. Proc Natl Acad Sci USA 95, 3489–3494. 35 Jiang LJ, Maret W & Vallee BL (1998) The glutathione redox couple modulates zinc transfer from metallothio- nein to zinc-depleted sorbitol dehydrogenase. Proc Natl Acad Sci USA 95, 3483–3488. 36 Ka ¨ gi JHR, Suzuki KT, Imura N & Kimura M (eds) (1993) Metallothionein III. Birkha ¨ user, Basel. 37 Ka ¨ gi JHR (1991) Overview of metallothionein. Methods Enzymol 205, 613–626. 38 Capasso C, Abugo O, Tanfani F, Scire A, Carginale V, Scudiero R, Parisi E & D’Auria S (2002) Stability and conformational dynamics of metallothioneins from the antarctic fish Notothenia coriiceps and mouse. Proteins 46, 259–267. 39 Dabrio M, Rodriguez AR, Bordin G, Bebianno MJ, De Ley M, Sestakova I, Vaa ´ k M & Nordberg M (2002) Recent developments in quantification methods for metallothionein. J Inorg Biochem 88, 123–134. 40 Pan PK, Hou FY, Cody CW & Huang PC (1994) Sub- stitution of glutamic acids for the conserved lysines in the alpha domain affects metal binding in both the alpha and beta domains of mammalian metallothionein. Biochem Biophys Res Commun 202, 621–628. 41 Olsson PE, Kling P, Erkell LJ & Kille P (1995) Struc- tural and functional analysis of the rainbow trout (Oncorhyncus mykiss) metallothionein-A gene. Eur J Biochem 230, 344–349. 42 Vergani L, Grattarola M, Dondero F & Viarengo A (2003) Expression, purification, and characterization of metallothionein-A from rainbow trout. Protein Expr Purif 27, 338–345. 43 Laemmli U.K. (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227, 680–685. 44 Scha ¨ gger H & von Jagow G (1987) Tricine-sodium dodecyl sulfate-polyacrylamide gel electrophoresis for the separation of proteins in the range from 1 to 100 kDa. Anal Biochem 166, 368–379. 45 Bradford MM (1976) A rapid and sensitive method for the quantification of microgram quantities of protein utilising the principle of protein-dye binding. Anal Bio- chem 72, 142–146. 46 Bu ¨ hler R & Ka ¨ gi JHR (1978) Spectroscopic properties of zinc-metallothionein. Experientia Supplement 34, 211–220. 47 Bu ¨ hler RHO & Ka ¨ gi JHR (1979) In Metallothionein (Kagi, JHR & Nordberg, M, eds), pp. 211–220. Birkha- user, Basel. 48 Bartolucci S, Gagliardi A, Pedone E, De Pascale D, Cannio R, Camardella L, Rossi M, Nicastro G, de Chi- ara C, Facci P, et al. (1997) Thioredoxin from Bacillus acidocaldarius: characterization, high-level expression in Escherichia coli and molecular modelling. Biochem J 328, 277–285. 49 Hennessey JP & Johnson WC (1981) Information con- tent in the circular dichroism of proteins. Biochemistry 20, 1085–1094. 50 Manavalan P, Taylor P & Johnson WC Jr (1985) Circu- lar dichroism studies of acetylcholinesterase conforma- tion. Comparison of the 11S and 5.6S species and the differences induced by inhibitory ligands. Biochim Biophys Acta 829, 365–370. 51 Provencher SW & Glo ¨ ckner J (1981) Estimation of globular protein secondary structure from circular dichroism. Biochemistry 20, 33–37. L. Vergani et al. Comparison between two MTs from aquatic organisms FEBS Journal 272 (2005) 6014–6023 ª 2005 The Authors. Journal compilation ª 2005 FEBS 6023 . Fish and molluscan metallothioneins A structural and functional comparison Laura Vergani 1 , Myriam Grattarola 1 , Cristina Borghi 2 , Francesco. primer (5¢-CTACTACGAATTAGGATCCCCT GCACCTTG-3¢) and the 3¢-end primer (5¢-GTAATACGA CTCACTATAGGGCGAATTGGG-3¢). Amplification was performed as previously