Distinct metal-binding configurations in metallothionein.

In a study of the binding stoichiometry of various metals to rat liver metallothionein, the protein appears to coordinate metals in 2 distinct configurations. Ions of at least 18 different metals were shown to associate with the protein suggesting that there is little specificity in binding. Most metals exhibited saturation binding at 7 mol eq forming M7-metallothionein. These included Bi(III), Cd(II), Co(II), Hg(II), In(III), Ni(II), Pb(II), Sb(III), and Zn(II). Others metals including Os(III), Pd(II), Pt(IV), Re(V), Rh(III), and Tl(III) give a positive indication of binding, but stoichiometries were unclear. Ag(I) and Cu(I) bound in clusters as M12-metallothionein. This binding stoichiometry was determined in 3 ways: (a) by determining the equivalence point in Cu- and Ag-titrated samples where resistance to proteolysis is maximal; (b) by determining the point where Zn ions are completely displaced from Zn7-metallothionein; and (c) by direct binding studies. Ag-reconstituted protein, recovered from gel filtration, had an average Ag content of 11.5 g atoms/mol of protein. A similar stoichiometry for the Cu-protein resulted from displacement of Zn from Zn7-metallothionein by Cu(I). The M12-protein was converted to the M7-protein by displacement of Ag(I) or Cu(I) with 7 mol eq of Hg(II). Whereas the distribution of metals in the 2 domains of M7-metallothionein is M4 alpha and M3 beta, the arrangement in the M12-molecule is probably M6 alpha and M6 beta. We propose that metallothionein ligates Ag(I) and Cu(I) in a trigonal geometry by bridging thiolates. This is in contradistinction to a tetrahedral binding geometry in the M7-protein. Distinct binding configurations may result in different tertiary structures for M7- and M12-proteins which may relate to metabolic specificity of Zn-metallothionein and Cu-metallothionein, respectively.

4 To whom correspondence should be addressed. bonding between side chain hydroxyl residues and amide carbonyls of the polypeptide backbone (7).
Although binding stoichiometries and coordination geometry have not been clearly established for Ag(I), Au(I), and Cu(1) (13)(14)(15), Cu-metallothionein is one form of the protein that deviates from the usual coordination of 7 tetrahedrally bound metals/polypeptide (15)(16)(17). We recently found that 11 or 12 Cu ions were bound to metallothionein and that the j3 domain can coordinate 6 Cu(1) ions, unlike the 3 Zn(I1) ions/ j3 domain in Zn-metallothionein (17). The higher apparent binding stoichiometry of Cu-metallothionein suggests that the Cu-protein adopts a conformation different from that of Cd,Zn-metallothionein. Cu(1) can farm tetrahedral complexes but usually does not with sulfur ligands (18). Several small Cu-S complexes with Cu4S6 and Cu5S7 cores have been characterized, and the usual coordination chemistry is trigonal Cu(1) ions with bridging sulfurs (19)(20)(21)(22).
It is important to understand the coordination properties and structure of both Zn-metallothionein and Cu-metallothionein, because the protein may function in cellular processes involving both Zn and Cu (23)(24)(25)(26)(27)(28). These two forms of metallothionein may play a central role in intestinal absorption, intracellular storage, transport, or excretion of Cu and Zn. Although Zn-metallothionein appears to be the predominant form of metallothionein in adult human tissues, Cu-metallothionein exists in significant concentrations in the liver of ruminants and fetal livers from various species (29)(30)(31)(32)(33). The copper protein is the predominant metallothionein species in crab and Neurospora crassa (34,35). The hypothesis to be tested is that metallothionein exists in two functional states. The different tertiary conformations of Cu-metallothionein and Zn-metallothionein permit the cell to discriminate between the two forms.
Since Cu(1)-metallothionein is susceptible to air oxidation, structural studies on using the Cu-protein abound with complexity. We have tested several metal ions for their ability to associate with the protein in attempt to find an experiment model for Cu-metallothionein. An oxygen-stable derivative of metallothionein that mimics the coordination geometry of the Cu-protein would aid in the estimation of the binding stoichiometry, In addition, we wanted to determine the selectivity of metal binding to the protein and to find the range of binding stoichiometries. In the present report, evidence is provided that ions from at least 18 different metals associate with the protein. Although the prevalent binding configura-

Metal Binding in Metallothionein 5343
tion is the M7-protein, Ag(1) coordinates to the protein similarly to Cu(1) with a equivalency of about 12.

RESULTS
Ultraviolet Absorption Assay-In order to study the specificity of metal binding to metallothionein, in vitro metal reconstitution of apometallothionein was used. The procedure involved anaerobic titration of the apoprotein with increasing mol eq of various metal ions. It is well established that renaturation of apometallothionein with Cd(I1) and Zn(I1) restores properties characteristic of native Cd,Zn-metallothionein (5,37). One characteristic feature of the metalloprotein is the prominent absorption spectrum from 240 to 300 nm which results from metal-ligand charge-transfer transitions. The shape and extinction coefficients of the spectrum are distinctive for different metals bound to the protein (8-13, 38). Increasing the metal equivalency in subsaturating amounts in metallothionein samples resulted in an increase in the absorbance (36,37).
Various metal ions were screened for their ability to enhance the ultraviolet absorption of apometallothionein. Only a few transition and group 111 to V cations (Cr(III), Mn(II), Ge(IV), Tl(I), and Zr(1V)) were without effect. Group I1 divalent ions (Mg, Ca, Sr, Ba) had no effect. Metal ions that yielded a significant increase in ultraviolet absorption as a function of added metal were tested in a second experiment. Renaturation studies were performed with both intact apometallothionein and proteolyzed apoprotein. Metals (Cd(II), Zn(II), and Cu(1)) ligated in chelated clusters showed greater absorption compared to ligation by cysteines from the proteolyzed protein. If binding of a particular metal occurs by only uninuclear thiolates, little difference may be observed in comparing intact to proteolyzed metallothionein. In Table I the ultraviolet absorption difference is shown for these 2 states of the protein mixed with various metals. Several cations (Fe(II), Fe(III), Ir(III), W(V1)) increased the absorbance of proteolyzed apometallothionein to about the same extent as the intact protein. Other metal ions augmented the ultraviolet absorbance of intact metallothionein more appreciably compared to their effect on the digested protein. One interpretation of the data is these latter metals may be associating with the protein in polynuclear clusters analogous to Cd, Zn, and Cu ions.
Proteolytic Protection Assay-Apometallothionein is digested to small peptides by proteases (subtilisin, proteinase K) whereas the native metalloprotein is completely resistant. We demonstrated previously that metallothionein renatured with 7 mol eq of Cd(I1) or about 12 mol eq of Cu(1) was resistant to incubation with these proteases (15,36). Various other metal ions were tested for their ability to protect the protein against proteolysis ( Table 11). The extent of digestion was monitored by reaction with fluorescamine, which quantifies primary amino groups. Metallothionein protected by a metal from proteolytic digestion has a fluorescamine fluorescence yield similar to that of the protein in the absence of a protease. The protein reconstituted in the presence of 10-12 mol eq of Ag(I), Bi(III), Cd(II), Cu(I), In(III), Ni(II), Os(III), Pb(II), Pd(II), Rh(III), Sb(III), Tl(III), and Zn(I1) was either completely or largely resistant to the proteases. Only partial protection was afforded by Fe(II), Fe(III), Ir(III), Pt(II), Pt(IV), and Sn(II), although higher metal ion concentrations of Fe(III), Mn(II), and Sn(I1) enhanced the resistance. Al(III), Ba(II), Ca(II), Cr(III), Ga(III), Ge(IV), K(I), Mg(II), Mn(II), Mo(VI), Sr(II), V(IV), W(VI), and Zr(1V) were without significant effect at 30 mol eq. The results refer only to apparent protection, since resistance would also be apparent if the metal inhibited the activity of the proteases. The metals were tested for protease inhibition in a proteolysis reaction using casein and ribonuclease as substrates for subtilisin and proteinase K. Most of the cations were without significant effect on the proteases. Addition of Au(III), Cu(I), Hg(II), Os(III), Pd(II), Re(V), Rh(III), and Ru(II1) in concentrations equivalent to those used in the studies with metallothionein resulted in partial inhibition of one or both of the proteases. If the metals are associated with metallothionein, greater concentrations of the metal were found necessary to affect the activity of the proteases. Decreasing the concentration of the protease 3-fold did not alter the results, so a partial inhibition of the enzyme should not affect the outcome.
The apparent metal-dependent protection against proteolysis was also checked by analysis of the peptides on thin layer chromatography. Numerous peptides were visualized by ninhydrin staining after digestion of apometallothionein with either sbutilisin or proteinase K. Only a few faint bands were observed other than the native protein which was at the origin when the protein reconstituted with Ag(I), Cd(II), Cu(I), In(III), Ni(II), Pb(II), Pd(II), Sb(III), Tl(III), and Zn(I1) was incubated with one of those proteases.
There are numerous factors which influence the proteolysis assay resulting in different optimal conditions for certain metals. Besides the obvious metal-anion solubility effect with certain buffers, the pH of the reaction is an important variable because of displacement of complexed metal ions by protons. With certain metals the protection decreased as the pH was lowered. Ions of Zn(II), Pb(II), In(III), and Cd(I1) became ineffective in protecting metallothionein from proteolysis by pepsin at pH values of 4,3,2.5, and 2.5, respectively. Whereas the protection rendered by Bi(III), Cd(II), Cu(I), Hg(II), and Pb(I1) was insensitive within the pH range of 5.5 to 7.8, the resistance of the molecule reconstituted with In(III), Sb(III), or Hg(I1) increased as the pH was lowered from 7.8 to 5.5. This effect appears to result from the pH-dependent instability of the aqueous metal ion due to hydrolysis (39). Support for this interpretation came from other experiments. In the case of Sb(III), the extent of protection at pH 7.8 became almost complete as the metal ion concentration was increased to 60 mol eq. Cation aqueous stability was also apparent when the order of addition of the metal ion and buffer was varied. When the reconstitution procedure was altered so the apoprotein was neutralized to pH 7.8 prior to addition of the metal ion, Sb(II1) and In(II1) became ineffective in exerting a protective effect on the protein. Conversely, only a slight effect was observed in preneutralizing the apomolecule to pH 7.8 and using Cd(I1) or Bi(II1) as the protecting metals. At pH 5.5 a concentration of In and Sb of 10 mol eq was adequate for complete proteolytic resistance. The pH results with In(II1) and Sb(II1) were confirmed by peptide analysis using thin layer chromatography.
The protease insensitivity of metallothionein renatured with various metals is also affected by the time of incubation. Resistance to proteinase K at pH 7 of the molecule containing bound Ag(I), Cu(I), Cd(II), Pb(II), Ru(III), or Rh(II1) was unchanged for times up to 40 h. To the contrary, the protein coordinating Bi(III), Hg(II), In(III), Ni(II), Sb(III), Sn(II), or Zn(I1) became increasingly sensitive to digestion after prolonged time under similar conditions. Although this lability may be related to the structure of the metalloprotein, aqueous instability of certain metal ions at pH 7 may also be a contributing factor.
The proteolytic protection assay was also used to assess binding stoichiometry with various metals. The procedure was good since the incubations could be performed anaerobically and large numbers of samples could be screened. Titrations of apometallothionein with increasing metal equivalents led to a steady attenuation in the fluorescence resulting from fluorescamine reaction with amino groups until a value was reached that was similar to that of the native protein.
The assumption is that maximum protection against proteolysis is achieved only upon saturation of all binding sites. In previous experiments with Cd(I1) and Zn(II), the equivalency at the break point in the fluorescence curve was 7 which is known to be the maximal number of metals bound (15).  (Bi(II1) and Pd(I1)) the metal ions in the buffered solutions absorbed ultraviolet light, thereby masking a transition point, In the titrations with Hg(II), the absorbance curve at 250 nm was complicated by a drop in the absorbance after 7 mol eq followed by a second rise above 11 mol eq (Fig. 2). The basis for these changes, although unclear, may represent a change from polynuclear cluster binding of Hg(I1) to ligation by nonbridging thiolates. A break point was not always obvious with Hg(I1) in the fluorescence curves, although maximal protection (about 80%) was usually attained by 7-8 mol eq. The number of Cu(1) equivalents necessary to render the protein resistant was previously reported to be approximately 11, but the transition point was not as obvious as with the Cd(I1) titrations (15). The lack of clarity is presumably related to valence instability in the Cu(1) protein. Results with Ag(1) were analogous to those with Cu(1) (Fig. 3). The minimum in the fluorescence curve could be more accurately established to occur near 12 mol eq. The curues shown in Figs. 1-3 are representative data of multiple titrations with each metal. Os(III), Pt(IV), Rh(III), and Ru(II1) were also screened in the protease assay at pH 6.5. In each case maximal protection was achieved by 15 mol eq, but transition points were not clear.
Displacement of Metalbthwnein-bound Metuls"Cd(I1) or Zn(1I) bound to metallothionein can be displaced by metals that have higher affinities for the protein. Native Cd,Znmetallothionein and metallothionein separately renatured with Ag(1) or Zn(I1) were incubated anaerobically with various metals to determine the ability of the added metals to displace the protein-bound cations. After incubation an aliquot of Chelex 100 was added to each tube to remove unbound cations. The samples were centrifuged, and metal analysis was performed on the supernatant (Table 111) study with Ag(I), Bi(III), In(III), and Pb(I1) ions the duration of the incubation period was varied from 0.5 to 24 h. No differences were observed in the extent of Cd and Zn displacement from the native protein. Similar results were obtained using the same metals in comparing Cd and Zn displacement from native and Cd,Zn-renatured proteins. The effect of pH on metal displacement was investigated. An increase in the extent of Zn displacement from the native protein was observed by decreasing the pH from 6.5 to 5.5 with 10 mol eq of In(III), Os(III), Pt(IV), Ru(III), Sb(III), andTl(II1). The other metals listed in Table I1 did not exhibit a pH effect. Zn was essentially completely displaced from the protein by 10 mol eq of In and Sb at pH 5.5, only 68 and 22% complete at pH 6.5, respectively, and less than 10% in each case at pH 7.8. This pH effect appears to be predominantly due to the pHdependent stability of aqueous metal ions. The incubation conditions at pH 5.5 had no apparent effect on the association of Zn with metallothionein. The molar Cd/Zn ratio in native Cd,Zn-metallothionein was unaffected in this Chelex procedure in a pH range from 5.5 to 7.8.
Titrations were carried out using this assay in order to obtain binding stoichiometries. Metals that were effective in the total displacement of Cd or Zn from the protein included Ag(I), Bi(III), Cd(II), Cu(I), Hg(II), In(III), Pd(II), Pt(IV), and Sb(II1). Addition of 7 mol eq of Bi(II1) or Hg(I1) yielded complete release of Cd and Zn from native metallothionein and Zn from the Zn,-renatured protein (Fig. 4). Likewise, Cd(I1) displaced the Zn ions in Zn7-metallothionein in a 1:1 relationship. With increasing Zn displacement, there was a corresponding increase in the Cd content of the protein (Fig  4). Whereas the protein was depleted of its Zn content by 7 mol eq of In(II1) and Sb(II1) at pH 5.5, somewhat higher equivalents of Pd(I1) and Pt(1V) appeared necessary for total Zn depletion (Fig. 5 ) . Binding studies with Pd(I1) and Pt(1V) were complicated by turbidity and occasionally precipitates of the metal-protein complexes. Ag(1) and Cu(1) titrations with metallothionein revealed that 12 mol eq led to complete displacement of protein-bound Zn and maximal concentrations of Ag and Cu in the supernatant (Fig. 6). Cupric ions were also effective in displacing Zn, but the known redox properties of the Cu(I1)-metallothionein interaction complicate the interpretation.
Inference of binding stoichiometry using the metal displacemelit assay was also carried out with CuI2-or Aglz-metallothionein as the starting material (Fig. 7). Seven mol eq of Hg(I1) totally displaced the protein-bound 12 Cu or Ag ions. There was a corresponding increase in the Hg content of the protein with decreasing Ag or Cu concentrations. The maximal Hg content occurred at 7 mol eq. Direct Binding Studies-Apometallothionein reconstituted with 10-14 mol eq Ag(1) was chromatographed by gel filtration. The elution volume containing the protein was quantified for metal and protein concentrations. In 11 separate experiments the protein was recovered with an average Ag content of 11.5 +-0.7 g atoms/mol protein. This metal content was unaffected if the sample was pretreated with 2 mM pmercaptoethanol prior to chromatography. The elution position of the Ag-protein was consistent with that of a monomeric protein. Increasing the Ag(1) concentration to 20 mol eq in the incubation mixture resulted in the recovery of molecules with approximately 20 Ag ions bound. Higher molecular weight Ag species were apparent by gel filtration.
Renaturation with a slight excess of Cu(1) led to the isohtion of metallothionein containing 10.9 bound Cu ions. However, the yield was quite low (10%) which may be attributed to complications arising from the oxidation lability of the molecule. If 12 mol eq of Cu(1) are allowed to displace Zn from Zn7-metallothionein prior to chromatography, the protein molecule is recovered in about 30% yield. The Cu content in two experiments was 10.9 and 12.4 g atoms/mol.
Relative in Vitro Metal-binding Affinity-The relative in vitro binding affinities of metals that associate with metallothionein can be ordered by comparing the displacement data in Table 111. The proposed ranking of relative affinities is presented in Table IV. The data are not definitive in distinguishing between metals in certain groups such as Hg, Pd, and Pt. The proposed order was substantiated by additional metal displacement studies. Separate incubations of 7 eq of Cd(I1) with In7-, Pb7-, and Sb7-metallothionein resulted in the apparent displacement of the protein-bound metals. This was inferred by detection of protein-bound Cd in the supernatant with recoveries of 90, 60, and 88%, respectively, of the Cd content of Cd7-metallothionein. In an analogous experiment, Pb7-metallothionein was incubated with 7 mol eq of Zn(II), and no significant quantities of Zn were found in the proteincontaining supernatant after the Chelex treatment.
A second approach to ranking in vitro metal-binding affinities was a pH titration of the protein with various metals ligated. As the pH is lowered, metals dissociate from the protein in an order related to the avidity of binding. The displacement of bound metals was monitored by the ultraviolet absorbance of the metalloproteins. Since the extinction in the near ultraviolet is dominated by sulfur -+ metal change transfer transitions, the pH-dependent loss of absorbance is a reflection of the dissociation of metal ions (37). From the dissociation curves, the pH where the absorbance is attenuated by 50% was determined (Table IV). Since the 2 domains differ in their avidities from metals, the midpoint pH will obviously be an average for the 2 halves. The following midpoint pH values were observed Cd (3.7), Co (5.8), Cu (2.7), In (4.4), Ni (5.7), Pb (3.8), and Zn (4.8).

In attempt to
determine the possible configurations of metal binding to metallothionein, we have screened numerous metal ions for binding to the apoprotein. Four assays have been employed to identify binding. Chelation was classified as positive if a metal ion: ( a ) generated a positive ultraviolet absorption difference spectrum comparing binding to the intact apoprotein versus the proteolyzed molecule; (6) protected apometallothionein against proteolysis by subtilisin or proteinase K; (c) displaced Zn or Cd from Cd,Zn-metallothionein; and ( d ) remained associated with the protein after gel filtration of reconstitution mixtures. All assays were not applicable to all the metal ions capable of binding to the protein. Certain metals (Co(II), Ni(I1)) did not displace protein-bound Zn(I1). Other metal ions, e.g., Hg(II), bound so tenaciously that any stoichiometry may be obtained in the direct binding assay depending on the equivalents added. Ions of at least 18 metals, when present in only a slight excess concentration over available binding sites, satisfied at least 2 of the tests for binding.  (9-14).
Experiments with certain metals presented difficulties. Aqueous instability due to disproportionation or hydrolysis, redox reactions of metal ions with the protein, and insolubility with certain anions were among the problems encountered. Therefore, experimental conditions had to be changed to circumvent these difficulties. Whereas In(I1I) and Sb(II1) were prone to hydrolysis near neutrality, buffers at pH 5.5 were acceptable. Although pH 5.5 is higher than pH values where hydrolysis of many of these metal ions normally occurs, complex formation with a ligand as metallothionein will prevent metal hydroxide precipitation (40).
A relative order of in vitro binding affinities determined for ;everal metals correlated well with published affinities of Hg Co to metallothionein (9,41). Contrary to this order, Waalkes et al. (42) recently reported the following order based on Zn displacement studies from Zn-metallothionein: Cd > Pb > Cu > Hg > Zn > Ag > Ni > Co. Boulanger et al. (7) proposed an order of affinity in the p domain to be Cu > Zn > Cd. Data presented in this study clearly makes these latter schemes questionable. The order proposed in Table IV agrees well with established association constants for metals and cysteine (Hg > Cu(1) > Ag > Pb > Cd > Ni > Zn > Co) and with sulfides (Hg > Ag > Pb > Cd > Zn) (40,43,44). The pH displacement results with metallothionein suggest that Cd and Pb bind with the similar affinities, whereas the Zn displacement data predicts greater avidity in the Cd binding. Contrary to binding studies with cysteine, Zn associated to metallothionein more tenaciously than did Ni.
Binding stoichiometries of these various metals were assessed by monitoring the equivalencies yielding complete protection against proteolysis and total displacement of Zn or Cd ions from Cd,Zn-metallothionein. The basis of the proteolytic protection assay was verified with Cd(II), Co(II), Ni(II), and Zn(I1) where maximal binding is known to be 7 ions/molecule (5,9). The Zn displacement assay is valid only for metals that possess a substantially greater affinity for the protein than does Zn(I1). Other metals that were found to exhibit saturation binding at 7 mol eq include Bi(III), Hg(II), In(III), Pb(II), Sb(III), and probably Pd(I1). Vasak et al. (8) and Bernhard et al. (10) reported tetrahedral binding of Bi(III), Hg(II), and Pb(I1) from spectral studies of the metal-reconstituted proteins, but titrations were not performed to verify binding stoichiometries. Others have observed metallothionein binding of Bi and Hg in metal-injected animals, but the bound Bi and Hg contents were low (45,46).
Metals such as Au(III), Re(V), Rh(III), Pt(IV), Ru(III), Os(III), and Tl(II1) give positive indication of binding, but data on stoichiometries were unclear. The ambiguity with these ions is probably related to aqueous instability of the ions or unstable complexes with the protein. Experiments with Au(II1) revealed a break point of about 12 in both the proteolytic protection assay and Zn displacement from Zn7metallothionein. Although the transition point seems clear, the nature of Au(II1) as a potent oxidant makes interpretation of these data unclear. Other redox-active metal ions would also confuse data interpretation. Direct binding studies were carried out with In(III), Pd(II), Pt(IV), Rh(III), and Sb(II1). Whereas In and Sb exhibited saturation binding between 5.5 to 7 metal ions/protein molecule with good recovery of protein, recovery of protein renatured with Rh, Pd, and Pt was quite low, thereby obscuring an interpretation.
Cu (1) and Ag(1) appear to bind in clusters as Mlz-metallothionein. In the proteolytic protection assay 12 mol eq of Ag(1) protects the protein against proteolysis. Approximately 12 mol eq of Ag(1) or Cu(1) are necessary for the complete displacement of Zn from Zn7-metallothionein. In direct binding studies, renaturation of the apoprotein with a slight excess of Ag(1) followed by gel filtration resulted in the recovery of Ag-metallothionein with an average Ag content of 11.5 g atoms of Ag/mol of protein. A comparable number is obtained in the molecule resulting from Cu(1) displacement of Zn from the Zn7-protein. Thus, a binding stoichiometry for Ag and Cu of 12 eq is apparent from both reconstitution and Zn-displacement studies. Addition of Ag(1) is excess of 20 mol eq results in collapse of the cluster structures to yield a metal cysteine ratio of about 1. At lower equivalents Ag(1) appears to bind in polynuclear clusters analogous to Cu(1). We have preliminary data that Ag(1) binds cooperatively and initially in the p domain as does Cu(1). Ag-metallothionein appears to represent an air-stable model for the MI2-protein complex.
Metallothionein binds metal ion in 2 distinct configurations. The distribution of metals in the M7 state is M4a and M3P, and the metals are coordinated tetrahedrally (3, 5). The 12 Cu(1) ions are bound in polynuclear clusters with a distribution of M6P and M6a. From the cysteine content in the domains, the p domain metabligand stoichiometry is Cu6:Cys9.
In order for the Cu ions to have quasi-equivalent bonds, the Cu(1) ions should be coordinated trigonally. One possible trigonal orientation is depicted in Fig. 9. In this complex, all thiolates have bridging sulfurs. In known structures of several small Cu-thiolate complexes, the usual coordination chemistry is trigonal Cu(1) ions and bridging sulfurs (19)(20)(21)(22). Ag(1) forms a M5S7 complex that is similar in structure and coordination stereochemistry to the Cu5S7 complex (22). Two distinct binding configurations have also been reported in another system. Beltramini et al. (47) recently reported that whereas the Neurospora %-residue metallothionein polypeptide normally binds 6 Cu(1) ions/molecule, the Cd(II), Hg(II), and Zn(I1) renatured molecules exhibit a metal to protein stoichiometry of 3. The two states in metallothionein are probably interconvertible. The Zn7-protein can be converted to the M12 state by metal displacement with 12 mol eq of Cu(1) or Ag (1). Conversely, the MI2-protein can be reverted to the M7 state by displacement with 7 mol eq of Hg(I1).
The higher binding content of Cu-metallothionein compared to the Zn-form may result in 2 different tertiary structures. If metallothionein functions in distinct processes for Zn and Cu, a conformational difference between the M7-and M12-proteins would permit the cell to discriminate between the 2 forms. In support of distinct structures, the Stokes radius of Zn-metallothionein is slightly greater than that of the Cu-protein. Resolution of Cu-and Zn-proteins has also been observed by Bremner and colleagues (48,49).
In many cells metallothionein appears to exist with varying ratios of bound Zn and Cu. It is unclear whether the protein exists in most cells as a mixed Cu,Zn-protein or whether pure Zn or Cu states of the mo!ecule occur. The 2 domains in the protein appear to be relatively independent of each other. However, as the interdomain-connecting peptide L y~~~L y s~~ FIG. 9. Proposed coordination complexes for the 6 domain of metallothionein. In the M? and MI* states of the protein, the p domain appears to coordinate 3 and 6 metals, respectively, via 9 cysteinyl residues. is short, the clusters may be spatially close. The stoichiometry of a given domain may be influenced by the tertiary conformation of the other domain. Prior formation of cluster A by Zn could impart constraints in the folding of the p domain, thereby precluding formation of a Zn-type tetrahedral center mixed with a Cu-type trigonal center. Answers to these questions will have an important bearing on the yet unresolved function of metallothionein. Further studies to elucidate the cluster structure of Cu-metallothionein and its coordination geometry are in progress.  .I I .03