Báo cáo khoa học: Mercury(II) binding to metallothioneins Variables governing the formation and structural features of the mammalian Hg-MT species pptx

Mercury(II) binding to metallothioneins Variables governing the formation and structural features of the mammalian Hg-MT species A ` ngels Leiva-Presa, Merce ` Capdevila and Pilar Gonza ` lez-Duarte Departament de Quı ´ mica, Facultat de Cie ` ncies, Universitat Auto ` noma de Barcelona, Spain With the a im of extending our knowledge on the reaction pathways of Zn-metallothionein (M T) and apo-MT species in the presence of Hg(II), we monitored the titration of Zn 7 -MT, Zn 4 -aMT and Zn 3 -bMT proteins, at pH 7 and 3, with either HgCl 2 or Hg(ClO 4 ) 2 by CD and UV-vis spectr- oscopy. Detailed analysis of the optical data revealed that standard variables, such as the pH of the solution, the binding ability of the counter-ion (chloride or perchlorate), and t he time elapsed b etween subsequent additions of Hg(II) to the protein, play a determinant role in the stoichiometry, stereochemistry and degree of folding o f the Hg-MT species. Despite the fact that the effect of these variables is unques- tionable, it is difficult to generalize. Overall, it can be c on- cluded that the reaction conditions [pH, time elapsed between subsequent additions of Hg(II) to the p rotein] affect the structural properties more substantially than the s toi- chiometry of the Hg-MT species, and that the role of the counter-ion becomes particularly apparent on the structure of overloaded Hg-MT. Keywords: mercury(II) binding; mercury-metallothionein; metallothionein; a-metallothionein; b-metallothionein. Mercury t hiolates provide representative examples of the structural diversity shown by the extensive family of metal thiolates [1–4]. The most striking features of mercury thiolates in the solid phase are the different structures obtained when Hg(II) is co-ordinated to very similar thiolate ligands [5,6] and the distinctive behavior of Hg(II) towards a particular thiolate compared with that of Zn(II) or Cd(II) [7], which has been referred to a s the zinc family paradox [3]. Moreover, correlations between solid-state and solution complexes cannot be easily established. Overall, the diverse co-ordination preferences of Hg(II) ions (mainly tetrahedral, trigonal-planar and digonal) and their coexist- ence in polynuclear complex species, the various ligation modes of the thiolate ligands (i.e. terminal, l 2 -bridging or l 3 -bridging) and the possibility of secondary Hg(II)–sulfur interactions [8] make it difficult to anticipate the structure of a particular mercury thiolate complex [1,3,9]. This results from the interplay of not only the above factors, but a lso the reaction conditions. Of these, the presence of additional co- ordinating species, such as halide ions, make the bonding situation for mercury even less straightforward than in the case of homoleptic mercury thiolates [10,11]. The biological chemistry of mercury is dominated by co-ordination t o cysteine thiolate groups in agreement with the preference of this metal ion for the soft sulfur ligands. The high binding constants for binding of Hg(II) to cysteine residues account for the irreversible replacement of essential metals (Zn, Cu) in cysteine-containing metalloproteins and thus for the high toxicity of mercury to living systems. Within the same context of the highly favored thermo- dynamically Hg-S bond, resistance to Hg(II) toxicity in several bacteria is based on an ensemble of proteins designated as Mer, most of which bind Hg(II) i ons through cysteine residues ([3] and references therein). In mammals, detoxification of mercury by metallothioneins (MTs) occurs via cysteine complexation a nd sequestration [12]. A major feature of this very large family of ubiquitous low molecular mass proteins is their extremely high content of cysteine residues, the binding of which to metal centers determines the 3D structure of the protein [13]. Consideration of the high flexibility and multidentate ligand nature of t he peptide chain in MTs together with the intrinsic complexity of mercury thiolate complexes suggests that elucidation of the stoichiometry and co-ordination geometries of mercury in solution Hg-MT species may be rather i ntricate. To date, optical spectroscopy (UV-vis a nd CD) has played a major role in the study of the mercury-binding properties of mammalian MTs, for which several Hg-MT stoichiometries have been reported [14]. Thus, a detailed analysis of the electronic spectra o f Hg(II)-reconstituted M T led Vas ˇ a ´ k et al. [15] to propose that Hg(II) in Hg 7 -MT is co-ordinated at sites w ith tetrahedrally related geometry. Subsequent studies by Johnson & A rmitage [16] of the UV spectral data obtained in the titration of C d(II) 7 -MT with Hg(II) showed that Hg(II) initially occupies tetrahedral sites but, above a Hg/MT stoichiometry of four, there is a shift to linear co-ordination. However, on the basis of X-ray Correspondence to M. Capdevila, Departament de Quı ´ mica, Facultat de Cie ` ncies, Universitat Auto ` noma de Barcelona, E-08193 Bellaterra, Barcelona, Spain. Fax: + 34 935813 101, Tel.: + 34 935813 323, E-mail: merce.capdevila@uab.es Abbreviations: MT, metallothionein; TDPAC, time differential per- turbed angular correlation of c-r ays; UV-vis, ultraviolet-visible elec- tronic absorption; t, stabilization time allowed for the co-ordination of Hg(II) to the protein; X, counter-ion of the Hg(II) salt added as titrating agent. (Received 19 July 2004, revised 21 October 2004, accepted 25 October 2004) Eur. J. Biochem. 271, 4872–4880 (2004) Ó FEBS 2004 doi:10.1111/j.1432-1033.2004.04456.x absorption studies conducted on some of the species observed in the titration of either apo-MT or Zn 7 -MT with Hg(II), monitored by optical spectroscopy, Lu & Stillman [17] proposed a d istorted tetrahedral co-ordination for Hg(II) in Hg 7 -MT with two short (2.33 A ˚ )andtwolong (3.4 A ˚ ) Hg-S distances [18]. Previous extended X-ray absorption fine structure (EXAFS) results for Hg 7 -MT were consistent with a Hg-S bond length of 2.42 A ˚ and suggested that Hg(II) was in a three-co-ordinate thiolate environment [19]. Although the protective role of MTs a gainst Hg(II) toxicity provides particular interest for the study of the Hg(II)-MT system, most existing results are difficult to reconcile. With t he aim of finding new strategies for this study, w e now report o n the effect of two variables, the reaction time and the presence of chloride ions, on the stoichiometry, stereochemistry and degree of folding of the Hg(II)-MT species formed by either the binding of Hg(II) to apo-MT or Zn/Hg replacement in Zn 7 -MT. Materials and methods Protein preparation and characterization Fermentator-scale cultures, purification of the glutathione- S-transferase-MT fusion p roteins, and recovery and ana- lysis of the recombinant mouse Zn 7 -MT1, Zn 4 -aMT1 and Zn 3 -bMT1 domains were performed a s p reviously described [20,21]. The Zn 7 -MT, Zn 4 -aMT and Zn 3 -bMT species were obtained in both Tris/HCl and Tris/HClO 4 buffer (50 m M , pH 7) [22]. The protein concentration was  0.1 m M in the six solutions, which were diluted to a final concentration of 10 l M (MT) or 20 l M (aMT and bMT fragments) with MilliQ-purified and Ar-degassed water before being titrated with Hg 2+ solutions at 25 °C. The apoproteins were prepared by acidification of the recombinant material with 10 m M HCl or HClO 4 , respect- ively, until pH 3. At pH values lower than 3.5 the Zn 7 -MT, Zn 4 -aMT and Zn 3 -bMT species are entirely devoid of metal, according to their respective CD spectra. In contrast, Hg(II) remains bound to SCys at this pH. Metal solutions Glassware and solutions used in metal ion-binding studies were prepared as described [20]. A Riedel-de Hae ¨ natomic absorption spectrometry Hg 2+ standard of 1000 p.p.m. was used as the HgCl 2 solution. The Hg(ClO 4 ) 2 solution was prepared from the corresponding salt in MilliQ-purified water, and the Hg(II) concentration was quantified by atomic absorption spectrometry using a Perkin–Elmer 2100 atomic absorption spectrometer. In both cases the Hg(II) concentration of the tit rating agents was in the 1–10 m M range. Metal ion-binding reactions Metal-binding experiments were carried out by sequentially adding molar-ratio a liquots of concentrated Hg(II) stock solutions to single solutions of either the holo proteins or apoproteins and followed spectropolarimetrically (CD) and spectrophotometrically (UV-vis). Two sets o f titrations, which differ in the time elapsed between subsequent additions of Hg(II) to the protein, were carried out. In one set, the standard titration procedure [ 22] was followed, whereas in the other consecutive additions of Hg(II) were made every 24 h. The electronic absorption and CD measurements were performed and corrected as already described [22]. All m anipulations involving the protein and metal ion solutions were performed in Ar atmosphere, and the titrations were carried out at least in duplicate t o ensure the reproducibility of each point. The pH (7 or 3) for all experiments remained constant throughout. A t pH 7, t he acidity of t he Hg(II) solutions required the addition of appropriate buffer solutions of Tris/HCl or Tris/HClO 4 (50 or 70 m M at pH 7), but no buffering was required for the titrations carried out at pH 3. Results and Discussion In view of the well-known complexity of Hg(II)–thiolate systems, the difficulties we encountered in analyzing the results obtained th rough preliminary titrations of the Zn-MT proteins with Hg(II) were not a surprise. They indicate that the nature of the counter-ion (X) and the time elapsed between subsequent additions of the Hg(II) solution (t) have a significant effect on the stoichiometry, stereo- chemistry and degree of folding of the species formed. Thus, to understand the reaction pathways followed by Zn-MT and apo-MT species in the presence of Hg(II), the e ffect of each of the previous variables was analyzed separately. T o this end, the titration of Zn 7 -MT, Zn 4 -aMT and Zn 3 -bMT proteins, at pH 7 and 3, with either HgCl 2 or Hg(ClO 4 ) 2 were spectroscopically monitored. The CD and UV-vis spectroscopic techniques used in this work are currently used to study metal-binding features of MT as they provide i nformation on the c o-ordinative features of the predominant metal-MT species present in solution at each titration point and on the number of s pecies formed during the titration. Furthermore, titration of the separate fragments provides information on the depend- ence/independence relationship between the t wo constitu- tive domains of the whole MT protein [21,23]. With regard to the two pH values, titrations at pH 7 a llow the subsequent substitution of Zn(II) and thus formation of heterometallic Zn,Hg-MT species, and titrations at acidic pH values provide information on the binding of Hg(II) to the corresponding apo-MT form [23]. In a ddition, compar- ison of the two sets of data gives an indication of the role of Zn(II) in the Hg(II)-containing species formed at physiolo- gical pH. The use of two different Hg(II) salts allowed analysis of the possible role of the physiologically relevant chloride anion, which has a strong tendency to co-ordinate and b ridge Hg(II) ions, i n the degree of folding and 3D structure of the Hg-MT species. The perchlorate anion is well known for its low co-ordinating ability towards metal centers. As regards the time variable, the spectroscopic changes observed in the titrations of Zn 7 -MT, Zn 4 -aMT and Zn 3 -bMT with Hg(II), after different times were allowed for the reaction between the MT protein and the added Hg(II) ions, were indicative of a strong dependence of the Hg-MT system on this variable (Fig. 1). Thus, titrations Ó FEBS 2004 Variables governing the binding features of Hg-MT (Eur. J. Biochem. 271) 4873 with HgCl 2 were carried o ut at two different times, t ¼ 0h and t ¼ 24 h, whereas those with Hg(ClO 4 ) 2 were only performed at t ¼ 24 h. The t ¼ 0 h label denotes that the titration was performed under k inetic control c onditions, which means that, for each addition, the protein sample was allowed to react with the metal ion u ntil subsequent CD spectra were essentially coincident [22]. However, for most samples, if the CD spectrum was recorded again after 24 h, it showed significant differences from that recorded at t ¼ 0 h. For this reason, titrations labeled t ¼ 24 h denote those carried out under thermodynamic control conditions, where each molar-ratio aliquot of Hg(II) was added every 24 h, as longer time intervals showed no further changes in the spectroscopic features. Overall, evaluation of all the variables in the Hg-MT system required the performance and analysis o f 18 titrations and the corresponding duplicates. The detailed and comparative analysis of the set of CD, UV-vis and difference electronic absorption spectra recorded for e ach titration (provided as Supplementary Material) provides information on the species formed by the Zn-MT peptides in the p resence of Hg(II) under t he d ifferent experimental conditions assayed and has allowed us to propose the reaction pathways (Schemes 1–3) for Zn/Hg replacement in Zn-MT species ( pH 7) and for the binding of Hg(II) to apo-MT (pH 3) that are discussed below. Mercury content in the Hg(II)-MT species at each titration point has traditionally been established b y assu- ming that, in solution, only one species is present, the metal c ontent of which coincides with the number of Hg(II) equivalents (eq) added . To validate the previous assumptions as well as to quantify the Zn content in t he Zn,Hg-MT species observed at pH 7 (Schemes 1A, 2A and 3A), we unsuccessfully devoted m uch e ffort to obtaining ESI-MS data. Thus, information on the Zn(II) content was retrieved from CD data and it is mainly of a qualitative nature. Reaction of recombinant mouse Zn 7 -MT with Hg(II) Analysis of the CD, UV-vis and UV-vis difference spectra obtained in the titration of Zn 7 -MT with Hg(II) at pH 7 (Fig. 2, S 1 and S2) and pH 3 (Figs S3–S5) for each set of X and t values led to the reaction pathways shown in Scheme 1. Comparative analysis of the three sets of data indicates that the stoichiometry of the species formed along the three titrations at pH 7 depends on neither the stabilization time, t, nor the nature o f th e counter-ion. The unique exceptions Fig. 1. Evolution w ith time of t he CD spectra corresponding to the addition of the tenth Hg(II) to Zn 7 -MT at pH 7. Scheme 1. Proposed reaction pathways for Hg(II) binding to recombinant Zn 7 -MT at pH 7 (A) and at pH 3 (B), under th ermodynamic (t ¼ 24 h) or kinetic (t ¼ 0 h) control conditions, using HgCl 2 or Hg(ClO 4 ) 2 as titrating agents. The  and „ symbols denote similarity and difference, respectively, between the structure of two species compared. 4874 A ` . Leiva-Presa et al.(Eur. J. Biochem. 271) Ó FEBS 2004 to this rule are : (a) Z n,Hg 2 -MT, observed as an i ntermediate species only at t ¼ 24 h; (b) the stoichiometries of the fully loaded species, Hg 15 -MT and Hg 16 -MT. Conversely, the chirality of the species is highly dependent on the previous variables, t ¼ 24 h and X ¼ Cl – affording t he most chiral species, as s hown by the intensity of t he CD bands of the Hg(II)-MT species formed under these conditions (Fig. 2). Similarly, t and X have a significant effect on the structure of the Hg-MT aggregates, with a Hg to MT ratio equal or higher than 7, as evidenced by the comparison of the CD spectra of isostoichiometric species obtained under different conditions. The contribution of the counter-ion to the 3D structure of the Hg-MT aggregates is demonstrated by the outstanding example of Hg 11 -MT, which becomes one of the most chiral species if formed in the presence of Cl – under both kinetic and thermodynamic control conditions (Fig. 3 ). Another relevant feature is the formation of hetero- metallic Zn,Hg 5 -MT and Zn,Hg 7 -MT, both present in the three titrations. T he former shows a very specific CD fingerprint. The significance of the latter lies in the Hg(II) stoichiometry, as previous studies proposed formation o f homometallic Hg 7 -MT species [17,24]. Under the experi- mental conditions used, the evolution of the CD spectra is fully consistent with the presence of heterometallic Zn,Hg 7 -MT as an intermediate species between Zn,Hg 5 - MT and Hg 9 -MT. Overall, the information obtained using the optical techniques allows Zn,Hg 5 -MT a nd Hg 11 -MT to be considere d the hallmark species formed in the Zn/Hg replacement in Zn 7 -MT. ABC Fig. 2. (A) CD, (B) absorp tion UV-vis, and (C) differe nce abso rption UV-vis s pectra obtained by s ubtracting the successive spectra of (B), corresponding to the titration of recombinant mouse Zn 7 -MT1 with HgCl 2 at pH 7 and t ¼ 24 h. The Hg(II) to MT m olar ratios are indicated within each fram e. Ó FEBS 2004 Variables governing the binding features of Hg-MT (Eur. J. Biochem. 271) 4875 Data obtained a t pH 3 show a strong influence of t and X on the stoichiometry and structure of the species formed, as shown i n Scheme 1B, and thus, the three reaction pathways followed at this pH are remarkably different. Notwithstanding this, there is a minor effect of t and X at the b eginning and end o f the titration. Thus, the addition of the first 4–6 of Hg(II) to apo-MT gives rise to Hg-MT species of comparable stoichiometry and structure, i.e. Hg 4 -MT and Hg 5)6 -MT,andalsothe presence of an excess of Hg(II) cation leads invariably to Hg 18 -MT. Furthermore, within the previous range [ from 4–6 to 18 Hg(II)], subsequent additions of Hg(II) led to low-chirality Hg-MT species under all conditions. T he only exception is Hg 13 -MT, formed at t ¼ 0handX¼ Cl – , which shows a well-defined CD fingerprint, also indicative of a highly chiral species. Concerning the role of the counter-ion, the differences observed in the CD spectra of overloaded Hg-MT species, such as Hg 10 -MT and Hg 18 -MT, formed at t ¼ 24 h, provide evidence for the interaction of the chloride anion with Hg(II), as already found at pH 7. Fig. 3. Role of the chloride anion in the d egree of folding of Hg-MT species observed by comparing the CD spectra of the Hg 11 -MT species obtained in the titration of Zn 7 -MT with either HgCl 2 (in black) or Hg(ClO 4 ) 2 (in grey), both at pH 7 and t ¼ 24 h. Scheme 2. Proposed reaction pathways f or Hg(II) binding to recombinant Zn 4 -aMT at pH 7 (A) and a t pH 3 (B), under thermodynamic (t ¼ 24 h ) or kinetic (t ¼ 0 h) control conditions, using HgCl 2 or Hg(ClO 4 ) 2 as titrating agents. The  and „ symbols denote similarity and difference, respectively, between the structure of two species compared. Fig. 4. CD spectra o f (A) the Z n 2 Hg 4 -aMT (in b lack) and Hg 5 -aMT (in grey), and Zn,Hg 4 -aMT (in black) and Zn,Hg 5 -aMT (in grey) species, respectively, obtained in the titrations of Zn 4 -aMT with HgCl 2 (solid lines) or Hg(ClO 4 ) 2 (dashed lin es), both at pH 7 and t ¼ 24 h and (B) the Hg 11 -aMT spe cies obtained in the titrations of Zn 4 -aMT with HgCl 2 (in black) or Hg(ClO 4 ) 2 (in grey), both a t pH 3 and t ¼ 24 h. 4876 A ` . Leiva-Presa et al.(Eur. J. Biochem. 271) Ó FEBS 2004 Reaction of recombinant mouse Zn 4 -aMT with Hg(II) Consideration of the optical spectroscopic data obtained in the titrations of Zn 4 -aMT with Hg(II) at pH 7 (Figs S6–S8) and pH 3 (Figs S9–S11) allows the proposal of the reaction pathways shown in Scheme 2. Analogously to Zn 7 -MT, the stoichiometry of the Hg-aMT species formed at pH 7 (Scheme 2A) along the three titrations does not depend on t and X. Notwith- standing this, the Hg 7 -aMT species is absent in the presence of Cl – at t ¼ 24 h, and t he species containing the highest H g(II) content, Hg 11 -aMT, is only obtained if t ¼ 0h and X¼Cl – . Conversely, t he structure a nd chirality of t he various Hg-aMT species are significantly influenced by t and X, as evidenced by their CD spectra. Thus, the species with a Hg to aMT molar ratio higher than 6–7 became mo re chiral if formed in the presence of Cl – , a mong which, tho se formed a t t ¼ 0 h show the highest degree of chirality. Exceptionally, only the Zn,Hg 4 -aMT species are comparable with respect to their chirality and structure under the three sets of experimental conditions. Interestingly, concerning the Zn,Hg 4 -aMT species, the 244(+) nm CD band recorded after t he addition of 4 H g(II) to Zn 4 -aMT under all sets of conditions not only gives a clear indication of the presence of Zn(II) in the aggregate, but its intensity also suggests that the highest Zn(II) content is found when X ¼ ClO 4 – (Fig. 4A). A similar analysis reveals the presence of Zn(II) in the Hg 5 -aMT species formed with X ¼ ClO 4 – but its absence for X ¼ Cl – . Chelex-100 treatment [23] of a n aliquot of the correspond- ing sample and subsequent analysis of the Zn and Hg content by inductively coupled plasma atomic emission spectroscopy and i nductively coupled plasma mass spectro- metry a llowed us to unequivocally establish the Zn 2 Hg 4 - aMT and Hg 5 -aMT stoichiometrie s for the species formed at t ¼ 24 h and X ¼ Cl – . Overall, all previous data indicate that the replacement o f Zn(II) by Hg(II) in Zn 4 -aMT proceeds more efficiently in the presence of Cl – than in the presence of ClO 4 – . At pH 3 (Scheme 2B) neither t nor X has a substantial effect on the stoichiome try of the s pecies formed during the titrations, except for the formation of two additional species, Hg 3 -aMT and Hg 7 -aMT, at t ¼ 24 h and X ¼ ClO 4 – . Conversely, the nature of the counter-ion strongly affects the chirality of the species. This effect is remarkable for those species with a H g(II) stoichiometry equal to o r higher than 6, X ¼ Cl – and t ¼ 24 h. In contrast, the Hg-aMT species formed in the presence of ClO 4 – show a very low degree of folding, indicating that Cl – ions strongly participate in the acquisition of the 3D structure of the Hg- aMT species (Fig. 4B). Reaction of recombinant mouse Zn 3 -bMT with Hg(II) The spectroscopic data obtained in the titrations of Zn 3 -bMT with Hg(II) at pH 7 (Figures S 12–S14) and pH 3 ( Figures S15–S17) are consistent with the reaction pathways shown in Scheme 3. Comparison of the three sets of data recorded at pH 7 (Scheme 3A) reveals that the Hg:bMT stoichiometry of the species does not depend on the nature of the counter-ion. Conversely, the stabilization time determines the Hg-bMTstoichiometryofmostofthe species formed and becomes particularly evident as the Scheme 3. Proposed reaction pathways for Hg(II) binding to recombinant Zn 3 -bMT at pH 7 (A) and at pH 3 (B), under thermodynamic ( t ¼ 24 h) or kinetic (t ¼ 0 h) control conditions, using HgCl 2 or Hg(ClO 4 ) 2 as titrating agents. The  and „ symbols denote similarity and difference, respectively, between the structure of two species compared. Ó FEBS 2004 Variables governing the binding features of Hg-MT (Eur. J. Biochem. 271) 4877 nuclearity of the species increases. Notwithstanding this, saturation occurs in all cases for 10 Hg(II). On the other hand, CD data indicate that the degree of chirality an d the structure of the species formed up to Zn,Hg 3)4 -bMT depend on t and X, the most chiral species being those obtained at t ¼ 24 h and X ¼ Cl – . As opposed to that observed for the aMT fragment, the CD spectra reveal that the presence of Cl – favors the Zn(II) ions remaining bound to the bMT protein in the first stages of the titration. Titrations carried out at pH 3 (Scheme 3B) reveal that the stoichiometries of the Hg-bMT species become dependent on t and X after t he forma tion of Hg 7 -bMT. Comparison of the three sets of CD data indicates that the degree of chirality of the Hg-bMT species is generally independent of t. However, the chirality o f the species obtained in the presence of Cl – is much higher than that achieved when X ¼ ClO 4 – ,exceptfortheHg 3 -bMT species, with a very low chirality in both cases, a nd the Hg 7 -bMT species, which show comparable chirality for X ¼ Cl – and ClO 4 – (Fig. 5). Comparison of the CD fingerprints of the H g-bMT species formed a long the three titrations sh ows t hat their 3D structure is s trongly dependent on t an d X, except for Hg 3 -bMT, which is poorly structured under all conditions. Co-ordination environments around Hg(II) in Hg-MT species The c omplexity of the Hg(II)-MT system, which is mainly the result of its Hg-thiolate nature, makes it difficult to obtain information on the co-ordination geometry around Hg(II) in the Hg(II)-MT aggregates from optical techniques (CD and/or UV-vis spectra) by simple treatment of the data. There are several reasons: (a) the presence of different chromophores in the same species including Zn and/or Hg as metal ions and SCys and/or Cl – as ligands; (b) the absence of w ell-established relationships between most of the previous chromophores and the corresponding absorp- tion wavelengths [3]; (c) the overlapping of the absorption bands corresponding to different chromophores, as shown by the spectral envelopes in the difference UV-vis spectra. Despite this, analysis of the difference UV-vis data, which discloses the effect of each Hg(II) addition, can give an insight into the evolution of the co-ordination geometry about Hg(II) in the M T species for med by either Zn/Hg replacement in Zn 7 -MT or the addition of Hg to apo-MT. By following this approach, comparison of the difference UV-vis spectra obtained in the titrations of Zn 7 -MT, Zn 4 -aMT and Zn 3 -bMT with HgCl 2 at pH 7 and t ¼ 24 h (Fig. 2, S6 and S12) indicates a parallel evolution of the co-ordination geometry about Hg(II) in the three peptides. These spectra evo lve according to the following pattern: (a) the addition of the first 7 Hg(II) eq to Zn 7 -MT, or the first 4 Hg(II) eq to any of t he aMT and bMT fragments, causes initially the appearance of an asymmetric broad band Fig. 5. CD spectra of the H g 7 -bMT species obtained in the titrations of Zn 3 -bMT at pH 3 with HgCl 2 at t = 24 h (solid black line) or t = 0 (solid grey line), or with Hg(ClO 4 ) 2 at t = 24 h (dashed g rey line). AB Scheme 4. An insight into t he evolution of the coordination geometries about Hg(II) in the Hg- MT species formed during the titrations of Zn 7 -MT, Zn 4 -aMT and Zn 3 -bMT with HgCl 2 at t ¼ 24 h and pH 7 (A) or pH 3 (B). The different coloured are as have been d educ ed from the d ifference UV- vis spectra. Preliminary TDPAC me asurements on the Hg-MT spec ies within a square enable co rrelation of e ac h area w ith an spec ific coordination geometry about Hg(II). 4878 A ` . Leiva-Presa et al.(Eur. J. Biochem. 271) Ó FEBS 2004 (230–340 nm), which eventually transforms into two new broad overlapping bands with absorption maxima at  230 and 320 nm; (b) the next Hg(II) eq added to the three peptides gives rise to a negative broad band with absorption minima at  260 and 310 nm, together with a positive absorption with a maximum intensity in the range 220– 230 nm; (c) further Hg(II) additions to Hg 11 -MT, Hg 6 -aMT and H g 5 -bMT cause the former envelope to turn into a positive broad band with an absorption maximum at  250 nm with a shoulder at  310 nm; (d) this profile collapses in the last steps of the titrations to give rise to very weak absorptions along the whole wavelength range. This common evolution of the three titrations gives force to different scenarios (denoted differently in Scheme 4A), which may be consistent with the presence of three different sets of co-ordination environments around Hg(II) in MT. Although the UV-vis difference spectra also suggest the existence of d ifferent scenarios in the binding of Hg(II) to Zn 7 -MT, Zn 4 -aMT and Zn 3 -bMT at pH 3 a nd t ¼ 24 h (Figures S3, S9 and S15), their evolution for the three peptides (Scheme 4B) does not show such good parallelism as that found at pH 7. Thus, at the beginning and end of the three titrations, the spectral e nvelopes compare well and suggest two different scenarios. The former includes all the species formed up to Hg 5 -MT, Hg 4 -aMT and Hg 4 -bMT, and consists of a positive very intense band with a maximum at  220 nm and a shoulder at  290 nm. The second scenario, which includes the species with the highest Hg(II) to MT ratios, is c haracterized by very low absorptions along the whole wavelength range. In addition, a broad band with amaximumat 250 nm and a shoulder at  310 nm denotes a t hird common feature apparent in different intermediate stages of t he three titrations. However, o nly MT and the aMT peptides give rise to a fourth common profile showing negative a bsorptions at  260 and 310 nm together with a positive absorption within the range 220– 230 nm. The evolution of the difference UV-vis spectra at pH 7 (Scheme 4A) and pH 3 (Scheme 4B) is consistent with preliminary time differential perturbed angular correlation of c-rays (TDPAC) measurements (A ` . Leiva-Presa, M. Capdevila, P. Gonza ` lez-Duarte & W. Tro ¨ ger, unpublished results) on several Hg-MT species. These results not only corroborate the proposals made from the d ifference UV-vis spectra but also suggest the specific co-ordination environ- ments a bout Hg(II) associated with each scenario. The correlation between optical and TDPAC data is summarized in Scheme 4, where the influence of the pH on the co-ordi- nation geometry about Hg(II) becomes apparent. One main difference is the predominance of tetrahedral geometry at pH 7 and digonal geometry at pH 3, both coexisting with other co-ordination geometries at increasing Hg to MT molar ratios. Interestingly, TDPAC measurements disclose two types of linear co-ordination environments about mercury: [Hg(SCys) 2 ] and [Hg(SCys)Cl]. Further TDPAC studies, now in progress, should provide definitive data on the co-ordinative features of the Hg-MT species. Concluding remarks The above results document the strong influence of standard variables (pH of the s olution, reaction time, a nd binding ability of the counter-ions) on the nature and structural features of the H g(II)-MT s pecies obtained by Zn/Hg replacement in recombinant Zn 7 -MT, Zn 4 -aMT and Zn 3 -bMT. Table 1 shows that this dependence is d iverse and thus difficult to generalize. However, it can be concluded that t he reaction conditions (pH, t) a ffect the structural properties more substantially than the stoichiom- etry of the Hg-MT species, and that the effect of the counter-ion (X) is particularly apparent on the structure of overloaded Hg-MT. Specific findings of this work are: (a) the high number of Hg-MT species observed (Schemes 1– 3); (b) the fo rmation of heterometallic Zn,Hg-MT aggre- gates, which include species such as Zn,Hg 7 -MT and Zn,Hg 4 -aMT, where the Hg(II) content equals that tradi- tionally expected for bivalent metal ions; (c) the nonadditive behavior of the a and b fragments with respect to the whole MT. Moreover, the stoichiometry found for the Zn 2 Hg 4 - aMT species indicates that the binding of one Hg(II) cation to MT does not require the displacement of one Zn(II) from the protein. N o such findings have previously been r eported. Earlier reports including CD and UV-vis data for the titration of native apo-MT2 and Zn 7 -MT2 with Hg(II) at pH 7 proposed formation of t he same set of species, Hg 7 - MT, Hg 11 -MT and Hg 20 -MT, along both titrations, the latter being replaced by Hg 18 -MT in the titration of apo- MT2 at pH 2. Similarly, the titration of both apo-MT2 and Zn 4 -aMT2 at pH 7 resulted in formation of Hg 4 -aMT and Hg 11 -aMT exclusively [14,17]. Possibly, the different source of the protein and the different experimental conditions used account for the discrepancy between these results and those reported i n t his work. Overall, the optical spectral data sets observed for Hg(II) binding to either Zn-MT or apo-MT confirm the requirement for accurate control o f the experimental conditions. Particularly relevant is the time variable, which has been scarcely considered in previous metal-M T binding studies. On the one hand, it has often been considered that metal displacement reactions in MT are kinetically facile and are generally complete within a few seconds [25]. Moreover, the kinetic lability and consequently continuous breaking and reforming of the metal-sulfur bonds are well documented for t he group 12 metal thiolates in solution [26]. On the other hand, the mechanism involved in the binding of Table 1. Influence of the reaction time (t) and binding ability of the counter-ions (X) on the nature and structural features of th e set of Hg(II)-MT species formed during the corresponding titration. Variables in bold deno te that the y have a stron g influence on most of the H g-MT species formed. Variables underlined affect only a minority of the species. Voids den ote that no general conc lusions can be drawn. The effect of t he p H can be deduced by co mparing the data of the same protein at the two pH values. Set of Hg-MT species Set of Hg-aMT species Set of Hg-bMT species pH 7 pH 3 pH 7 pH 3 pH 7 pH 3 Stoichiometry t, X t, X t, X t, X t, X Chirality t, X t, XX t, X t, X Structure t, X t, X t, X t, X t, X Ó FEBS 2004 Variables governing the binding features of Hg-MT (Eur. J. Biochem. 271) 4879 Hg(II) to MTs, which would determine its reaction rate, is unreported. Remarkably, our results show that not only do the reaction pathways at t ¼ 0handt ¼ 24 h differ considerably, but also that the CD features of a particular species formed along the titration at t ¼ 0 h do not evolve with time to those found for the isostoichiometric species at t ¼ 24 h. Acknowledgements This work was supported by a grant from the Spanish Ministerio de Ciencia y Tecnologı ´ a (BQU2001-1976 ). Dr Sı ´ lvia Atrian, who kindly provided us with the recombinant p roteins used in this work, acknowledges the Spanish Ministerio de Ciencia y Tecnologı ´ a for financial support (BIO2003-03892 ). We also acknowledge the Servei d’Ana ` lisi Quı ´ mica, Universitat Auto ` noma de Barcelona (CD, UV-vis) and the Serveis Cientı ´ fico-Te ` cnics, Universitat de Barcelona (inductively coupled plasma-atomic emission spectroscopy and inductively coupled plasma mass spectrometry) for allocating instrument time. References 1. Dance, I.G. (1986) The structural chemistry of metal thiolate complexes. Polyhedron 5, 1037–1104. 2. Dance, I.G., Fisher, K. & Lee, G. (1992) Metal-thiolate com- pounds. structure and dynamics of metal-thiolate and metal- sulfide-thiolate compounds. In Metallothioneins (Stillman, M.J., Shaw, C.F. I II & Suzuki, K.T., eds), pp. 284–345. VCH Publish- ers, New York. 3. Wright, J.G., Natan, M.J., MacDonell, F.M., Ralston, D.M. & O’Halloran, T.V. (1990) Mercury (II)-thiolate chemistry and the mechanism of the heavy metal biosensor MerR. Progr. Inorg. Chem. 38, 323–412. 4. Blower,P.J.&Dilworth,J.R.(1987)Thiolato-complexesofthe transition metals. Coord. Chem. Rev. 76, 121–185. 5. Kunchur, N.R. (1964) Polymer structure of mercury tert-butyl mercaptide. Nature (London) 204, 468. 6. Bradley, D.C. & Kunchur, N.R. (1965) Structures of mercury mercaptides. II. X-ray structural analysis of mercury ethylmer- captide. Can. J. Chem. 43, 2786–2792. 7. Casals, I., Gonza ´ lez-Duarte, P., Sola, J., Miravitlles, C. & Molins, E. (1988 ) Mercury-sulfur stretching frequencies, s ynthesis and x-ray crystal structure of caten a -l-(3-dimethylammonio-1- propanethiolato) dichloromercury (II). Polyhedron 7, 2509–2514. 8. Grdenic ´ , G. (1965) The structural chemistry of merc ury. Quart. Rev. Chem. Soc. 19, 303–328. 9. Govindaswamy, N., Moy, J., Millar, M. & Koch, S.A. (1992) A distorted mercury [Hg(SR) 4 ] 2– complex with alkanethiolate lig- ands: the fictile c oordination sph ere of monomeric [Hg(SR) x ] complexes. Inorg. Chem. 26, 5343–5344. 10. Biscarini, P., Foresti, E. & Pradella, G . (1984) Organothiometallic compounds. Crystal structure a nd spectroscopic prope rties of (isopropylthio)mercury (II) chloride. J. Chem. Soc. Dalton Trans. 953–957. 11. Alsina, T., Clegg, W., Fraser, K.A. & Sola, J. (1992) [Hg 7 (SC 6 H 11 ) 12 Br 2 ], a novel mercury cage molecule with bridging and t erminal thiolate ligands and with te rminal and l 6 bromide. J. Chem. Soc. Chem. Commun. 1010–1011. 12. Satoh, M., Nishimura, N., Kanayama, Y., Naganuma, A., Suzuki, T. & Tohyama, C. (1997) Enhanced renal toxicity by inorganic mercury in metallothionein-null m ice. J. Pharmacol. Exp. Ther. 283, 1529–1533. 13. Gonza ´ lez-Duarte, P. (2003) Met allothioneins. In Comprehensive Coordination Chemistry II (McCleverty, J.A. & Meyer, T.J., eds), Vol. 8, pp. 213–228. Elsevier-Pergamon, Amsterdam. 14. Stillman, M.J. (1995) Metallothioneins. Coord. Chem. Rev. 144, 461–511. 15. Vas ˇ a ´ k, M., Ka ¨ gi, J.H.R. & Hill, H.A.O. ( 1981) Zinc (II), cadmium (II), and mercury (II) thiolate transitions in metallothionein. Bio- chemistry 20, 2852–2856. 16. Jonhson, B.A. & Armitage, I.M. (1987) Equilibrium and kinetic analysis of the interaction of m ercury (II) w ith cadmium (II) metallothionein. Inorg. Chem. 26, 3139–3144. 17. Lu, W. & Stillman, M.J. (1993) Mercury thiolate cluster in metallothionein. Analysis of circular dichroism spectra of complexes formed between a-metallothionein, apometallothion- ein, zinc metallothionein and cadmium metallothionein and H g 2+ . J. Am. Chem. Soc. 115, 3291–3299. 18. Jiang, D.T., Heald, S.M., Sham, T.K. & Stillman, M.J. (1994) Structures of the cadmium, mercury and zinc th iolate cluster in metallothionein. XAFS study of Zn 7 -MT, Cd 7 -MT, Hg 7 -MT and Hg 18 -MT formed form rabbit liver metallothionein2. J. Am. Chem. Soc. 116, 11004–11013. 19. Charnock, J.M., Garner, C.D., Abrahams, I.L., Arber, J.M., Hasnain, S.S., H enehan, C. & Vas ˇ a ´ k, M. (1989) EXAFS studies of metallothionein. Physica B: Condensed Matter 158, 93–94. 20. Capdevila, M., Cols, N., Romero-Isart, N., Gonza ` lez-Dua rte , R. , Atrian, S. & Gonza ` lez-Duarte, P. (1997) Recombinant synthesis of mo use Zn 3 -b and Zn 4 -a metallothionein 1 domains and char- acterization of their cadmium (II) binding capacity. Cell. Mol. Li fe Sci. 53, 681–688. 21. Cols, N., Romero-Isart, N., Capdevila, M., Oliva, B., Gonza ` lez- Duarte, P., G onza ` lez-Duarte,R.&Atrian,S.(1997)Bindingof excess cadmium (II) to Cd 7 -metallothionein from recombinant mouse Zn 7 -metallothionein 1. UV-VIS absorption and circular dichroism studies and theoretical location approach by surface accessibility analysis. J. Inorg. Biochem. 68, 157–166. 22. Bofill, R., Palacios, O., Capdevila, M., Cols, N., Gonza ´ lez-Duarte, R., A trian, S. & Gonza ´ lez-Duarte, P. (1999) A new ins ight into the Ag + and Cu + binding sites in the metallothionein b domain. J. Inorg. Biochem. 73, 57–64. 23. Bofill,R.,Capdevila,M.,Cols,N.,Atrian,S.&Gonza ´ lez-Duarte, P. (2001) Zinc (II) is required for the in vivo and in vitro fold ing of mouse copper metallothionein in two domains. J. Biol. Inorg. Chem. 6, 405–417. 24. Bernhard, W., Go od, M., Vas ˇ a ´ k, M. & K a ¨ gi, J.H.R. (1983) Spectroscopic stud ies an d c haracteriza tion o f meta llothioneins containing mercury, lead and bismuth. Inorg. Chim. Acta 79 , 154 – 155. 25. Li, H . & Otvos, J.D. (1996) 111 Cd NMR studies of the domain specificity of Ag + and Cu + binding to m etallothionein. Bio- chemistry 35, 13929–13936. 26. Clark-Baldwin, K., Tierney, D.L., Govindaswamy, N., Gruff, E.S.,Kim,C.,Berg,J.,Koch,S.A.&Penner-Hahn,J.E.(1998) The limitations of X-ray abso rption spectroscopy f or determi ning the structure o f zinc sites in proteins. When is a t etrathiolate no t a tetrathiolate? J. Am. Chem. Soc. 120, 8401–8409. Supplementary material The following material is available from http://www. blackwellpublishing.com/products/journals/suppmat/EJB/ EJB4456/EJB4456sm.htm Figs. S1–S17. 4880 A ` . Leiva-Presa et al.(Eur. J. Biochem. 271) Ó FEBS 2004 . Mercury(II) binding to metallothioneins Variables governing the formation and structural features of the mammalian Hg-MT species A ` ngels. additions of Hg(II) to the p rotein] affect the structural properties more substantially than the s toi- chiometry of the Hg-MT species, and that the role of the counter-ion