Reaction of the Zinc Sensor FluoZin-3 with Zn₇-metallothionein: Inquiry into the Existence of a Proposed Weak Binding Site

Abstract

It has been reported that Zn₇-metallothionein (MT), contains one weak binding site for Zn²⁺. To test this conclusion, rabbit liver MT isolated at pH 7 was reacted with chelating agents of modest affinity for Zn²⁺. Contrary to the previous study, no evidence was found for Zn²⁺ stoichiometically bound to the protein with an apparent stability constant of about 10⁸. Indeed, stability constant measurements based upon competition between Zn₇-MT and ligands of known stability with Zn²⁺ showed that all of the protein bound Zn²⁺ displayed the same stability constant at pH 7.4 and 25° C of (1.7±0.6) × 10¹¹. Brief reaction of Zn₇-MT with strong acid converted it into MT* and upon reneutralization into Zn₇-MT*, which demonstrated reactivity of about 1 Zn²⁺/mol MT with competing ligands. Acid titration of Zn₇-MT to pH 2 or below rapidly resulted in the formation of Zn₇-MT* that displayed biphasic titration with base, revealing the rebinding of lower affinity Zn²⁺ between pH 5 and 7. Since MT is commonly acidified during preparation, care must be taken to document which form of the protein is present in subsequent experiments at pH 7.

Introduction

Metallothionein (MT) is a small, thiol rich metal binding protein that reacts with divalent metal ions such as Zn²⁺ and Cd²⁺ to form a two-domain conformation [1]. Each domain is comprised of a metal (M)-thiolate (S) cluster that organizes part of the polypeptide chain around it [1]. The N-terminal domain contains a M₃S₉ cluster, the C-terminal domain, a M₄S₁₁ cluster. Each is characterized by a set of metal ions linked through bridging sulfhydryl ligands and to the peptide backbone through bridging and terminal thiolate ligands.

It is well established that MT contributes to protecting cells against the toxicity of Cd²⁺ and other metal ions through direct chelation [2]. Its multiplicity of sulfhydryl groups also inactivates a variety of oxidants and electrophiles [2]. Normally present in many cells as a zinc-protein and induced by Zn²⁺, it has been hypothesized that metallothionein also is involved in intracellular Zn²⁺ trafficking [3–6]. One view is that MT simply acts as a transient Zn²⁺ storage site, buffering cells from changes in “free” or “loosely bound” Zn²⁺ [5].

$$7{\text{Zn}}^{2+}+\text{apo}-\text{MT}\rightleftharpoons {\text{Zn}}_7-\text{MT}$$ (1)

An extension has been to suggest that Zn₇-MT acts as a source of Zn²⁺ conversion of apo-proteins to Zn-proteins [7,8].

$$1/7{\text{Zn}}_7-\text{MT}+\text{apo}-\text{protein}\rightleftharpoons \text{Zn}-\text{protein}+1/7\text{apo}-\text{MT}$$ (2)

In this proposed reaction, the favorable stability constant of Zn-protein is sufficient to drive the equilibrium toward products.

Earlier reports of stability constant measurements of Zn₇-MT have only quoted values of greater than 10¹⁰ per Zn²⁺ at pH 7.4 [9,10]. That magnitude would allow for the observed reaction of Zn₇-MT with apo-carbonic anhydrase (K = 10¹²) [7,11]. However, a stability constant of this size would preclude its reaction with apo-Zn-proteins of lower affinity for Zn²⁺.

Recently, Maret and Krezel reported distinctly different results and concluded that apo-MT displays step-wise association of Zn²⁺ to form the holo-protein with the associated stability constants at spanning 4 orders of magnitude [12]:

$$\text{Apo}-\text{MT}+4{\text{Zn}}^{2+}\rightleftharpoons {\text{Zn}}_4-\text{MT}{\text{K}}_{1-4}={10}^{11.8}\text{per}{\text{Zn}}^{2+}$$ (3)

$${\text{Zn}}_4-\text{MT}+{\text{Zn}}^{2+}\rightleftharpoons {\text{Zn}}_5-\text{MT}{\text{K}}_5={10}^{10.9}$$ (4)

$${\text{Zn}}_5-\text{MT}+{\text{Zn}}^{2+}\rightleftharpoons {\text{Zn}}_6-\text{MT}{\text{K}}_6={10}^{10.7}$$ (5)

$${\text{Zn}}_6-\text{MT}+{\text{Zn}}^{2+}\rightleftharpoons {\text{Zn}}_7-\text{MT}{\text{K}}_7={10}^{7.7}$$ (6)

Reaction 3 cooperatively constitutes the 4-metal cluster. Then, Zn²⁺ adds sequentially to the 3 metal cluster. With its range of stability constants, particularly K₇, Zn₇-MT could potentially donate Zn²⁺ to apo-proteins that weakly bind Zn²⁺.

The experimental determination of these stability constants began with the observation that the fluorescent Zn²⁺ sensor, FluoZin-3, reacts with Zn₇-MT to form a fluorescent Zn²⁺ product [12]. The FluoZin-3 structure includes a fluorophore linked to a set of amine and carboxylate metal binding ligands with an aggregate modest stability constant for Zn²⁺ at pH 7 of 10^7.8 (Figure 1) [13]. Data in support of the weak stability constants were obtained from two similar experiments. In the first, binding isotherms were fit to the titration of apoMT with Zn²⁺ in the presence of FluoZin-3 and RhodZin-3. The second was based on competition between either FluoZin-3 or RhodZin-3 and MT for Zn²⁺.

Figure 1.

Structure of FluoZin-3

Besides their implications for metallothionein as a cellular participant in Zn²⁺ trafficking, the results presented by Maret and Krezel suggest that FluoZin-3 might be used to image the presence and intracellular location of Zn₇-MT. Considering the apparent centrality of MT-bound Zn²⁺ in cellular Zn²⁺ trafficking hypothesized by these authors and others, it was important to reexamine their determination that Zn₇-MT contains a weak Zn²⁺ binding site [3–6]. The results in the present paper address the validity of this conclusion and show that Zn₇-MT isolated at pH 7 from rabbit liver does not react significantly with FluoZin-3 and that Zn²⁺ ions but that the protein taken to low pH and then restored to pH 7 reacts as described by Maret and Krezel [12].

Experimental Methods

Materials

Sephadex G-75, G-25, and G-15 were from GE Healthcare Bio-Sciences AB, DEAE-cellulose and 5,5′-dithio-bis(2-nitrobenzoic acid) were provided by Sigma. Nitrilotriacetate and 2-mercaptoethanol were obtained from Aldrich Chemical Co. FluoZin-3^™ was obtained from Invitogen. The source for Chelex-100 resin was BioRad. TCEP was from Hampton Research. The thiosemicarbazone ligands, 3-ethoxy-2-oxobutyraldehyde-bis(thiosemicarbazone) and 3-ethoxy-2-oxobutyraldehyde-bis(⁴N-dimethyl-thiosemicarbazone) were gifts from Harold G. Petering.

Isolation of Zinc Metallothionein

Female New Zealand white rabbits (5–6 lbs.) were injected with 2 mL of 0.15 M ZnSO4 every 24 hours for 8 consecutive days. On the eighth day, the rabbits were anaesthetized with ketamine followed by a lethal dose of sodium pentobarbital (120 mg/kg) into the heart. The livers were removed immediately from the rabbits and washed with cold saline to remove excess blood, divided into about 100 g samples and stored at −80 °C until further use. About 100 g of frozen rabbit liver was thawed and minced into small pieces. Tissue was mixed with homogenizing buffer (0.25 M sucrose, 0.01 M 2-mercaptoethanol, 5 mM Tris-Cl, and PMSF) and homogenized. The homogenate was centrifuged in a Sorvall high speed centrifuge (SS34 rotor at 20,000 rpm) for 20 min followed by ultracentrifugation in a Beckman L-70 ultracentrifuge (70TI rotor at 31,000 rpm) and the final supernatant was loaded onto a 4×85 cm Sephadex G-75 column equilibrated with 20 mM Tris-Cl, pH 7.4, and 5 mM 2-mercaptoethanol and eluted at 4 °C. The MT fractions were located by measuring the Zn²⁺ profile of the eluate. Altogether about 1–1.5 liters of MT solutions were pooled. The two isoforms, MT1 and MT2, were separated by anion exchange HPLC using a semi-preparative DEAE column. The pooled MT solutions were loaded onto the column and MT was eluted with a gradient of 5–300 mM Tris-Cl at pH 7.4. MT-containing fractions were determined by measuring the concentrations of Zn²⁺ and thiol as described below. The overall yield of MT2 was 700–800 μg protein/g wet weight liver. Experiments described in the text used MT2 primarily and MT1 for some of the stability constant experiments.

Metallothionein Characterization

Isolated MT was characterized by measuring the metal and sulfhydryl contents in the sample. The metal concentration was obtained by atomic absorption spectrophotometry and the sulfhydryl concentration was determined using 5,5-dithiobis-2-nitrobenzoic acid (DTNB) assay [14]. The concentration of MT was also determined spectrophotometrically at 220 nm (ε₂₂₀=1.59 × 10⁵ M⁻¹cm⁻¹).

Preparation of Modified Metallothionein (MT*)

Concentrated HCl (12 N) was added to 1–2 mL of Zn7-MT to a final concentration of 1.2 M HCl, reducing the pH to −0.4. The solution was left at low pH for 20 min and then neutralized to pH 7.5 by titration with 10 N KOH. In other experiments a range of hydrogen ion concentrations was used to test the pH dependence of the reaction.

Properties of Ligands, H₂KTS, H₂KTSM₂ NTA, and FluoZin-3

Four competing ligands were used to study the Zn²⁺ binding properties of metallothionein: H₂KTS, H₂KTSM₂ NTA, and FluoZin-3. All of these ligands form one-to-one complexes with Zn²⁺. Zn-KTS has a pH dependent log stability constant of 5.9 at pH 7.4 and a maximum absorbance at 417 nm with extinction coefficient of 11,300 M⁻¹ cm⁻¹, as determined by titration of the ligand with a standard solution of ZnCl₂ [15]. The related complex, Zn-KTSM_2, has been characterized with a log stability constant at pH 7.4 of 9.73 and an extinction coefficient of 15,500 M⁻¹ cm⁻¹ at 434nm [15]. Nitrilotriacetic acid (NTA) has an intermediate Zn(II) binding affinity. The log apparent stability constant of Zn-NTA at 25° C, pH 7.4, and in 0.1 M KCl has been reported as 8.11 [16]. FluoZin-3 is a Zn²⁺ responsive fluorophore with a related log apparent stability constant of 7.8 [13].

Reaction of MT Species with Competing Ligands

Native and non-native metallothionein (1.6 μM protein), and ZnCl₂ (1.54 μM) were reacted with FluoZin-3 (5.8 μM) anaerobically in 50 mM Hepes buffer, 50 mM KNO3, pH 7.4 at 25° C. Fluorescence intensity was measured at 517 nm with excitation at 492 nm using an Hitachi F-4500 spectrofluorimeter. The reactions of Zn₇-MT and Zn₇-MT* with other ligands were conducted under similar conditions as described in the text. The reaction involving H₂KTS was monitored with a Beckman Model DU 640 spectrophotometer. In later confirmatory reactions involving FZ and NTA, TCEP, a thiol reducing agent, substituted for anaerobiosis [12]. Kinetic traces were averages of 3 reactions. Error (standard deviation) bars for points along the reactions of Zn₇-MT* are now provided. Final extent of reaction was recorded as average ± standard deviation.

Acid-Base Titration of Zn₇-MT

Zn₇-MT (5.8 μM) in 5 μM Tris buffer, 0.1 M KCl, pH 7.4 was titrated at room temperature with 12 N HCl to pH 0.6. UV absorbance and pH (Fischer Scientific Accumet, AR 10) were recorded after each addition of HCl. At the conclusion of the titration, the sample was incubated for 20 min and then back-titrated with 10 N NaOH to about pH 7.6.

Determination of Stability Constants of Zn₇-MT with Competing Ligands

The ligands H₂KTSM₂ and NTA establish measurable equilibria with Zn₇-MT as measured by the formation of ZnKTSM₂ (445 nm) or the loss of Zn²⁺ from the protein (215 nm), respectively [17]. The latter is possible because NTA does not absorb strongly in the UV. Two alternative hypotheses guided the analysis of the equilibrium data. Either all MT sites bind Zn²⁺ with equivalent affinity, within error, or the sites in each cluster associate with Zn²⁺ with the same affinity. In the first approach, Zn₇-MT is treated as Zn-MT in the reaction,

$$\text{Zn}-\text{MT}+\text{L}\stackrel{\text{K}}{\rightleftharpoons }\text{Zn}-\text{L}+\text{MT}$$ (7)

and K represents the equilibrium constant for the reaction per Zn²⁺, assuming each metal ion is bound with equal affinity. The overall equilibrium constant for ligand substitution reaction 1 and the knowledge of the apparent stability constants of Zn-KTSM₂ and Zn-NTA permit one to calculate the apparent stability constant for Zn²⁺ bound to MT under the conditions of the experiment.

In the second, it has been determined that the kinetics of reaction of Zn₇-MT with H₂KTSM₂ and NTA are biphasic and independent of competing ligand concentration [17,18]. This behavior is consistent with the unitary reaction of each Zn-thiolate cluster with these ligands. That is, all binding sites in each protein domain display the same reaction kinetics. From this result, it is hypothesized that with competing ligands each cluster metal binding site has effectively the same apparent stability constant for Zn²⁺. Analyzing the kinetic data as first order reactions, one can immediately assign a fraction of overall reaction to each reaction phase. For reactions that went to completion, the fraction of reaction data for each phase grouped about 0.5 (theoretically 0.43 and 0.57 for the two clusters). In this situation, one arbitrarily assigns the faster step to either cluster. Then, given the initial concentrations of each Zn-sulfhydryl cluster and the extent of reaction of each kinetic phase at equilibrium, the apparent stability constant of each cluster or kinetic class can be calculated as in reaction 7.

Results

Preparation of Zn₇-MT from Rabbit Liver

Zn₇-MT was prepared from rabbit liver as described in the methods section. Two peaks were obtained from DEAE chromatography, representing MT1 and MT2. The MT2 fraction was characterized for protein content (220 nm absorbance), Zn²⁺ concentration (atomic absorption spectrophotometry) and sulfhydryl concentration (DTNB analysis). Typical protein used in this study had a SH/Zn²⁺ of 2.9 ± 0.1, agreeing closely with the theoretical value of 2.86. In addition, PAGE electrophoresis showed that the product was greater than 95% pure based on runs with 0.2–10 μg protein (data not shown).

Preparation of Zn₇-MT* from Zn₇-MT

Maret and Krezel prepared Zn₇-MT* by expressing human MT2 in Escherichia coli as an intein fusion protein without addition of metal ions to the growth medium [12]. The construct was purified by affinity chromatography and the intein was cleaved using 0.1 M DTT. The purified protein was acidified to pH 1 using 1.2 N HCl and purified by G-25 chromatography equilibrated with 0.01 N HCl (pH 2). Final neutralization was achieved by adding small amounts of apoMT to neutral reaction buffer in the presence of Zn²⁺.

We subjected rabbit liver Zn₇-MT 2 to acid conditions below pH 0, which released all of the bound Zn²⁺, and then restored the sample to pH 7.4. After this acid-base cycle the product, designated as Zn₇-MT*, retained its full complement of sulfhydryl groups as measured with DTNB. Both migrated at the same rate and displayed only one band, indicative that the low pH incubation did not cause frank peptide cleavage (data not shown). Zn²⁺ binding properties of the starting protein were then compared with those of the protein that had gone through the acid-base cycle.

Characterization of Zn₇-MT and Zn₇-MT* Using Various Competing Ligands

FluoZin-3

Anaerobic reaction of excess FluoZin-3 with Zn₇-MT removed 0.22 μM Zn²⁺ or 0.14 mol Zn²⁺ per mol MT after a 120 min incubation period (Figure 2). The lack of reactivity alerted us to the fact that Zn₇-MT isolated from tissue did not behave as reported by Maret and Krezel [12]. In comparison, Zn₇-MT* reacted with FluoZin-3 as described by these authors such that 1.0 ± 0.07 mol of Zn²⁺ per mol MT* was extracted over a 120 min incubation period. Based on the stability constants for Zn-FZ and the weak Zn²⁺ binding site observed by Maret and the concentrations of protein and FZ used in this experiment, we expected that 0.81 mol Zn²⁺/mol MT* would be removed. Thus, 20% more Zn²⁺ reacted than anticipated.

Figure 2.

Reaction of FluoZin-3 with ZnCl₂ (▲), Zn₇-MT (◆) or Zn₇-MT* (■). Conditions: 1.6 μM protein (29 μM Zn²⁺), 5.0 μM FZ in anaerobic 50 mM Hepes buffer, containing 50 mM KNO3, pH 7.4 at 25° C. ZnCl₂ concentration, 1.54 μM.

The kinetics of the Zn₇-MT* reaction were multi-phasic and could be fit in part with two first order kinetic steps. A third very rapid reaction phase occurred in the time of mixing and comprised about 60 % of the overall reaction. With some preparations, the fast step reached 100% of the entire reaction. FluoZin-3 was not in sufficient excess so that it would compete with the second and third Zn²⁺ sites (reactions 5 and 6). Thus, the greater than expected removal of Zn²⁺, 1.0 not 0.8 Zn²⁺/mol, and the multiphasic kinetics suggest that Zn₇-MT* may be a mixture of species.

In order to broadly establish the differential behavior of Zn₇-MT and Zn₇-MT*, both proteins were reacted with other appropriate ligands that also have modest affinity for Zn²⁺

H₂KTS

A second ligand, H₂KTS, was employed to probe for a modestly bound Zn²⁺ ion in Zn₇-MT. H₂KTS forms a complex with Zn²⁺ that is readily detected spectrophotometrically. No Zn²⁺ became bound to this competing ligand after 120 min incubation of H₂KTS with Zn₇-MT under anaerobic conditions (Figure 3). With Zn₇-MT*, it was expected that at equilibrium 0.61 mol Zn²⁺/mol MT* would be sequestered by H₂KTS. After 2 h, 0.78 ± 0.06 mol Zn²⁺/mol MT* was complexed by this ligand, most of it within 10 min, with the reaction still slowly proceeding at the end of this period.

Figure 3.

Reaction of H₂KTS with Zn₇-MT (dashed line) or Zn₇-MT* (solid line). Conditions: 23 μM protein (161 μM Zn²⁺), 660 μM H₂KTS in 50 mM Hepes and 100 μM TCEP, containing 50 mM KNO3, pH 7.4 at 25° C.

NTA

Both Zn₇-MT and Zn₇-MT* were reacted with NTA and the reaction analyzed using gel filtration chromatography as done by Maret and Krezel [12]. Samples were incubated anaerobically with NTA for 120 min and then chromatographed over Sephadex G-25 (Figure 4). The protein fractions were analyzed for Zn²⁺ and sulfhydryl concentration and their ratio used to assess the state of the protein. Zn₇-MT retained its native ratio of 2.9, whereas, the ratio of MT* protein increased to 3.4± 0.4, consistent with the removal of 1.1 Zn²⁺ per mol MT* (20 SH:6 Zn²⁺).

Figure 4.

Reaction of NTA with Zn₇-MT(A) or Zn₇-MT* (B). Conditions: 0.56 μM protein, 3.9 μM Zn²⁺), 200 μM NTA in 50 mM Hepes and 480 μM TCEP, containing 50 mM KNO3, pH 7.4 at 25° C. SH (▲); Zn (■)

Ligand Concentration Dependent Reactivity and Stability of Zn₇-MT

The experiments above show that Zn₇-MT is unreactive with competing ligands under conditions that select for a binding site with log stability constant of 7.8. We examined the ligand concentration-dependent reaction of Zn₇-MT 1 and 2 with NTA and H₂KTSM₂ to determine whether there might be intermediate binding affinity sites in the protein. With NTA, we followed the change in Zn₇-MT absorbance at 215 nm as the indicator of the extent of reaction. With H₂KTSM₂, the visible absorbance of Zn-KTSM₂ served to monitor the extent of reaction. As seen in Figure 5, biphasic reaction kinetics were observed. Li and Otvos had previously reported this result with NTA [18]. Changing the concentration of either ligand did not affect the measured kinetics. In both reactions, it was plausible to hypothesize that each kinetic phase represented the reaction of Zn²⁺ ions bound in one of the protein’s Zn-thiolate clusters.

Figure 5.

Reaction of Zn₇-MT with H₂KTSM₂: pseudo-first order plot of the reaction kinetics observed at 434 nm. Conditions: 1.0 μM Zn₇-MT and 2.0 mM H₂KTSM₂ in 5 mM Tris-Cl and 0.1 M KCl at pH 7.4 and 25° C.

Next, the apparent stability constants of Zn²⁺ ions bound to MT were determined. Figure 6 shows the results when all seven sites are considered to have identical Zn²⁺ equilibrium binding properties. Over a wide range of ligand concentrations–15–500 μM H₂KTSM₂, which extracts 18–70% of protein bound Zn²⁺, and 0.54–10 mM NTA, which binds 5–65% of the Zn²⁺–the calculated stability constants at pH 7.4 and 25° C with Zn₇-MT 1 and 2 were within error independent of the initial concentration of the reactants and of the number of metal ions removed from MT 1 and 2 (Table 1). The average and standard deviation of all of the data for the reaction of Zn₇-MT I and 2 with these ligands was (1.73 ± 0.63) × 1011. The similarity of the calculated constants as a function of ligand concentration supported the hypothesis that all the binding sites had, within error, the same affinity for Zn²⁺.

Figure 6.

Apparent stability constants (K_app) per Zn²⁺ of Zn₇-MT1 and 2 as a function of competing ligand concentration (H₂KTSM₂ and NTA) assuming 7 indistinguishable sites. Conditions as in Figure 5. H₂KTSM₂ with Zn₇-MT1 (◆) and Zn₇-MT2 (■) and NTA with Zn₇-MT1 (▲).

Table 1.

Apparent Stability Constants for Zn²⁺ Bound to Zn₇-MT

Reactants	Apparent Stability Constant × 1011 pH 7.4 and 25°C
Zn7-MT 1, H2KTSM2	7 sites: 2.17 0.58
Zn7-MT 1, NTA	7 sites: 1.25 0.31
Zn7-MT 2, H2KTSM2	7 sites: 1.70 0.66
Zn7-MT 1 and 2, H2KTSM2 and NTA	7 sites: 1.73 0.63
Zn7-MT 1 and 2, H2KTSM2 and NTA	faster phase, β-domain (3 sites): 1.77 0.79 slower phase, α-domain (4 sites): 1.66 0.84
Zn7-MT 1 and 2, H2KTSM2 and NTA	faster phase, α-domain (4 sites): 1.55 1.29 slower phase, β-domain (3 sites): 1.56 0.75

The rates of reaction of NTA and H₂KTSM₂ with Zn₇-MT were biphasic (Figure 5) and the extent of reaction attributable to each phase was also approximately equal. It was assumed but not proven that the faster step was due to the reaction of the β-domain (Zn₃S₉) with the ligands. Using the additional information of the fraction of the overall reaction in each step in relation to the total Zn²⁺ assigned to each step, both phases were characterized by identical stability constants, (1.77 ± 0.78)×10¹¹ for the β-domain and (1.66 ± 0.84)×10¹¹ for the α-domain (Zn₄S₁₁) (Figure 7). The other assignment of domains to kinetic phases led to indistinguishable results with values of (1.56 ± 0.75)×10¹¹ and (1.55 ± 1.29)×10¹¹ for the β- and α-domain clusters, respectively.

Figure 7.

Apparent stability constants (K_app) per Zn²⁺ of Zn₇-MT1 and 2 as a function of competing ligand concentration (H₂KTSM₂ and NTA) assuming the two kinetic phases of reaction represent the two Zn-thiolate clusters, Zn₃-β domain, faster reaction (A) and Zn₄-α domain, slower reaction (B), and that Zn²⁺ ions bind indistinguishably within each cluster. Conditions as in Figure 5. H₂KTSM₂ with Zn₇-MT1 (◆) and Zn₇-MT2 (■) and NTA with Zn₇-MT1 (▲).

Acid-base Titration of Zn₇-MT

Zn₇-MT* was prepared by acidifying Zn₇-MT to about pH −0.4 and then restoring the pH to 7.4. Therefore, it was of interest to examine the acid-base titration characteristics of this process. Typical acid-base titrations of Zn₇-MT span pH 7–8 to 2–3 and display a single, continuous process of Zn²⁺ dissociation [19,20]. Few studies have reported the Zn²⁺ reassociation process. According to Figure 8, when the acid titration was taken to pH 0.6, two steps were resolved upon neutralization. The first step followed the path of the acidification and accounted for 85% of the overall change in 215 nm absorbance (Zn-sulfhydryl bonding) or 6/7 of the bound Zn²⁺ assuming a linear relationship between Zn²⁺ and 215 nm absorbance. The second occurs between pH 5–7 and represented a weakly bound fraction of about 1 Zn²⁺ per mol MT*.

Figure 8.

Acid-base titration of Zn₇-MT. Acidification with HCl (◆) and neutralization with NaOH (■) in 5 mM Tris-Cl buffer, 0.1 M KCl at room temperature .

Kinetics and Extent of Reaction of Zn₇-MT with Hydrogen Ion

Zn₇-MT was incubated at pH 3, 2.7, 1.8, and 1.0 for a series of times, rapidly neutralized, and then reacted with FluoZin-3 as an indicator of the degree of conversion of Zn₇-MT to Zn₇-MT*. Like pH −0.4, the native samples reacted at pH 1.0 and 1.8 were converted to Zn₇-MT* within 20 min. Indeed, at pH 1.0, the reaction was complete within 3 min, the quickest time for acidification and neutralization. At pH 2.7 and 3.0 the reactions were time dependent. For example, the fractional conversion to Zn₇-MT* after incubation at pH 2.7 for 20, 60, and 120 min was 0.68, 0.96, and 1.12 mol Zn²⁺/mol MT, respectively.

Discussion

A recent paper by Maret and Krezel provided evidence for multi-step binding of Zn²⁺ to apo-metallothionein [12]. Four stability constants were calculated, consistent with concerted binding of 4 Zn²⁺ ions to the α-domain of the protein (reactions 3–6) and step-wise addition of Zn²⁺ to the β-domain. The last step, in which the Zn²⁺ with the lowest affinity for the protein binds, was detected with a fluorescent ligand, FluoZin-3 (conditional stability constant log K = 7.8). The authors argued that the detection of a weak Zn²⁺ binding site in MT implies that MT can serve as donor of Zn²⁺ to proteins with weak as well as strong affinity for Zn²⁺. This property would confer on MT the capacity to serve as a broad spectrum, general source of Zn²⁺ for the proteome during Zn²⁺ trafficking processes.

FZ is best known for its use in cellular Zn²⁺ sensing [13]. A major issue in the application of fluorescent probes for Zn²⁺ is that increases in cellular fluorescence after incubation with sensors such as FZ can not be unambiguously ascribed to reaction with particular Zn²⁺ pools [21]. If FZ reacts with a low affinity Zn²⁺ binding site in MT, this result encourages the hypothesis that FZ might selectively image the presence and location of Zn₇-MT in vivo.

We revisited the reaction of FluoZin-3 with Zn₇-MT because earlier studies of the reaction of Zn²⁺ with the protein suggested that metal ions in the 2 clusters were bound with large affinity and reacted in concerted fashion with competing ligands [9,10,17,18]. Using Zn₇-MT freshly prepared at pH 7, we were unable to show that the protein reacted significantly with FZ and other ligands with weak to modest affinity for Zn²⁺, including H₂KTS, and NTA and Zincon (data not shown) (Figures 3 and 4).

Previous studies of the pH dependent dissociation of Zn²⁺ from Zn₇-MT characterized the process in terms of a single equilibrium constant, consistent with the assumption that all Zn²⁺ ions react equivalently with H⁺ [19,20]:

$$\text{Zn}-\text{MT}+{\text{mH}}^{+}\rightleftharpoons {\text{H}}_{\text{m}}-\text{MT}+{\text{Zn}}^{2+}$$ (8)

Similarly, equilibrium competition studies involving the reaction of Zn₇-MT with metal binding ligands (L) of known stability with Zn²⁺ indicated that each Zn²⁺ effectively displays similar affinity for the protein [9,10]. Thus, according to Figures 6 and 7, the reactions of a wide range of concentrations of NTA and H₂KTSM₂ with Zn₇-MT 1 and 2 could either be successfully analyzed by assuming a single stability constant for each Zn²⁺ in the entire protein or equivalent constants for each Zn²⁺ in the α- or β-domains. Using the first approach, apparent stability constants (K) per Zn²⁺ of (1.7±0.6) × 10¹¹ at pH 7.4 and 25° C were determined over a range of 5–70% of the total, bound Zn²⁺. No weak binding site (ca 10⁸) was detected (Figure 6). We also analyzed the equilibria in terms of their 2 component kinetic phases, reasonably attributing these reactions to the two Zn-thiolate clusters (Figure 7). The partial extent of reaction assigned to each phase at equilibrium was consistent with each cluster binding Zn²⁺ with similar affinity per Zn²⁺ as seen in the hydrogen ion titration results (Figure 8). Assigning the 2 kinetic steps of the reaction to the α- and β-domains similarly resulted in each domain cluster having respective K values per Zn²⁺ of (1.7±0.8) × 10¹¹ and (1.8±0.8) × 10¹¹ . If the assignment of kinetic phases was switched, the resultant log K data were indistinguishable from those obtained above.

An interesting feature of these results is that they support much earlier data that indicated that the two metal ion-thiolate clusters display distinct, often unitary kinetic behavior. These early studies suggested that the titration of apo-MT with Zn²⁺ and other metal ions resulted in cooperative formation of cluster structures in the absence of other intermediate metal ion-domain stoichiometries [22]. Those results agree with our findings described above. However, the presence of a protease in the reaction mixture to hydrolyze any unreacted peptide compromised the simplicity of the experiment and may have driven the system toward protease resistant cluster formation. Later studies revealed a different picture. Addition of Co²⁺ to MT is clearly non-cooperative as intermediate Co₁-, Co₂- and Co₃- species have been identified [23,24]. Similarly, an electrospray ionization mass spectrometric analysis of the titration of MT with Cd²⁺ also revealed a range of intermediate Zn_n-MT (n = 1–7) species [25].

These stability constants lie within an order of magnitude of 3 of the 4 constants (reactions 3–5) defined in the work of Maret and Krezel as well as one determined with the use of the amino-carboxylate chelating agent, Bapta (log K = 11.2, pH 8.0 or 10.4 extrapolated to pH 7.4) [11,12]. Earlier, a stability constant of 11.7 was reported, based on pH titration data and extrapolation of the results to pH 7.4 [20].

It seemed clear from the stability constant results that the protein used in our laboratory was different from that employed by Maret and Krezel. The two groups prepare Zn₇-MT in different ways. Our protein was isolated from rabbit liver, induced to make MT by Zn²⁺ injection, using a procedure that maintained the pH at about 7. In contrast, Maret and Krezel obtained their protein from a bacterial expression system in a process that included step acidification to pH 1 followed by gel filtration at pH 2 [12].

We discovered that acidification of our protein, Zn₇-MT, to low pH and subsequent neutralization converted it to Zn₇-MT* with properties like those of the Maret protein. Zn₇-MT* reacted with FZ, H₂KTS, NTA, and Zincon, removing slightly more than 1 Zn²⁺ per mol of protein (Figures 2–4). Zn₇-MT was almost unreactive.

The observed pH-dependent conversion of Zn₇-MT into Zn₇-MT* has the following properties. The pH dependent Zn²⁺ dissociation reaction was not reversible as seen in Figure 8. The dissociation of Zn²⁺ followed a smooth absorbance vs. pH curve, centered at pH 4. In contrast, the reassociation reaction displayed two steps, most of it overlaid with Zn²⁺ dissociation and about 15% occurring between pH 5 and 7, indicative of the presence of a weaker Zn²⁺ binding site. The transition between protein forms occurred faster at higher H+ concentration. Thus, at pH 2 or lower, the reaction that produced Zn₇-MT* was complete within 20 min and probably occurred even faster. In contrast at pH 3, 20 min incubation caused partial conversion that increased as incubation time at pH 3 was prolonged. Notably, at pH 3 and below, Zn²⁺ is nearly or fully dissociated from the protein. Therefore, the key reaction that changes the properties of Zn₇-MT involves the apo-protein structure. But the consequences of this reaction affect subsequent Zn²⁺-apo-MT interactions that occur as the protein is restored to pH 7.

We have not found conditions that restore Zn₇-MT* to Zn₇-MT in a preliminary survey. Neither mild heating nor urea treatment of Zn₇-MT* alters its reactivity with competing ligands. This provisional finding suggests that some property of the protein backbone changes at low pH, that this property is retained upon neutralization to pH 7, and that the modified protein backbone determines the protein fold about the two clusters. We do not know what the alteration in structure may be. Protonation reactions are anticipated to be fast, non rate limiting, and reversible; so some other process must be involved.

Previous studies indicate that strongly acidic conditions can catalyze cis-trans isomerization of peptide bonds [26,27]. In particular, protonation of the amide nitrogen of X-proline peptide bonds favors the single bond resonance structure. This can facilitate rapid cis-trans interconversion. We hypothesize that in moderate to strong acid, cis-trans isomerization of key amide linkages along the MT backbone convert MT to MT*. The conserved Proline 38 within the α-domain is a possible site of acid catalyzed conformational change [12,28]. The loss of metal binding stabilization of the MT peptide conformation at low pH would facilitate such a reaction. Neutralization would effectively lock the peptide bond in place as the amide-bond assumes its planar partial double bond character. Further inquiry will be needed to decide whether cis-trans backbone amide isomerization accounts for the empirical pH effect on Zn²⁺ binding properties of MT* that is observed.

Bacterially expressed MT protein is widely used as a source of MT protein in physico-chemical studies and preparations commonly involve a low pH step. Because a variety of properties of the two metal-thiolate clusters may differ in MT vs. MT*, it will be important for investigators to determine whether their protein behaves like pH 7-isolated or pH-modified metallothionein. .

Besides the intrinsic interest in Zn₇-MT* species and the nature of its modification in relation to Zn₇-MT, it is important to know whether it can be formed in vivo under other conditions. For example, change in configuration of His-Pro amide linkages can be catalyzed by protonation of the imidazole group near pH 7 [29]. If such reactions can occur, perhaps involving thiolate protonation, then Maret and Krezel’s contention that Zn₇-MT* may participate in Zn²⁺ trafficking with partners that have a wide range of equilibrium affinities for Zn²⁺ may be physiologically relevant [12]. This becomes a particularly interesting possibility considering that the metallothionein pool is unsaturated in many cells and would be the preferential MT species for such reactions [30–32].

If Zn₇-MT* can be formed under mild, physiologically significant conditions, the interpretation that a weak Zn²⁺ binding site in Zn₇-MT* provides the protein with a capability to donate Zn²⁺ to a wide range of cellular acceptors needs some qualifications. First, it is likely that GSH may readily compete for Zn²⁺ weakly bound to MT, as its stability constant with Zn²⁺ is about 10⁸ [33].

$${\text{Zn}}_7-\text{MT}+\text{GSH}\rightleftharpoons \text{Zn}-\text{SG}+{\text{Zn}}_6-\text{MT}+{\text{H}}^{+}\text{K}={10}^8/{10}^{7.7}$$ (9)

Thus, under physiological conditions of mM GSH, cellular MT* might at best exist as Zn₆-MT*, without its weakly bound Zn²⁺.

Furthermore, it is recognized that the cellular MT pool is partially unsaturated with Zn²⁺ [30–32]. Even under conditions of exposure to elevated concentrations of Zn²⁺ that induce the synthesis of MT, cells contain metal-unsaturated MT [32]. Thus, under equilibrium conditions, Zn²⁺ will populate the higher affinity sites, leaving vacant weaker Zn²⁺ binding sites. Thus, it is difficult to see how Zn²⁺ could remain selectively bound in cells to a low affinity binding site in the protein.

Abbreviations

Bapta

1,2-bis(o-aminophenoxy)ethane -N,N,N′,N′-tetra acetic acid

DTNB

5,5′-dithio-bis(2-nitrobenzoate)

FZ or FluoZin

2,2′-((2-(2-(2((carboxylatomethyl) amino)-5-(2,7-difluoro-6-oxido-3-oxo-3H-xanthen-9-yl)phenoxy)ethoxy)-4-methoxyphenyl) azanediyl)diacetate tetra potassium salt

H₂KTS

3-ethoxy-2-oxobutyraldehyde-bis(thiosemicarbazone)

ME

2-mercaptoethanol

MT

metallothionein

MT*

H⁺ ion modified MT

NTA

nitrilotriacetate

PMSF

phenyl-methyl-sulphonyl fluoride

S

sulfhydryl

TCEP

tris(2-carboxyethyl)phosphine