Metallothionein: An Aggressive Scavenger-The Metabolism of Rhodium(II) Tetraacetate (Rh2(CH3CO2)4).

Anthropogenic sources of xenobiotic metals with no physiological benefit are increasingly prevalent in the environment. The platinum group metals (Pd, Pt, Rh, Ru, Os, and Ir) are found in marine and plant species near urban sources, and are known to bioaccumulate, introducing these metals into the human food chain. Many of these metals are also being used in innovative cancer therapy, which leads to a direct source of exposure for humans. This paper aims to further our understanding of nontraditional metal metabolism via metallothionein, a protein involved in physiologically important metal homeostasis. The aggressive reaction of metallothionein and dirhodium(II) tetraacetate, a common synthetic catalyst known for its cytotoxicity, was studied in detail in vitro. Optical spectroscopic and equilibrium and time-dependent mass spectral data were used to define binding constants for this robust reaction, and molecular dynamics calculations were conducted to explain the observed results.

Introduction

With the turn of the millennium, sales of consumer products based on xenobiotic metallic elements have increased globally. Examples of these metals are Ti, Ni, and Sn in cosmetics,[1−6] Sn, Ni, and rare earth elements in electronic cigarettes,[7−9] Cd in batteries, Hg in lamps, As in older microchips, and platinum group metals (PGMs; Pd, Pt, Rh, Ru, Os, and Ir) in jewelry and widely in the automotive industry.[10−13] In addition, occupational exposure and environmental pollution can occur as a direct result of metal mining and refining.[14,15] These metals have no known physiological role, and with multiple routes of exposure, pose an unknown and largely unavoidable risk to human health. In the case of metallotherapeutics incorporating a range of rare metals (e.g. Au, Ag, Pt, Pd, Rh, and others), humans are directly exposed to acute concentrations. This increase in mobilization of these rare metals results in widespread human exposure, spreading from concentrated points and diffusing across the globe.

The physiological response to these rare metals relies on the body’s own protective response mechanisms to metals. Because of their nucleophilicity and metal scavenging ability, biological thiols such as metallothioneins (MTs) and glutathione (GSH) are involved in toxic metal resistance, whether directly or due to a downstream response triggered by metal exposure. Reports of these PGMs in urban environments and prevalence and accumulation in marine life near industrial runoff indicate the impending omnipresence and perturbance in the food chain and in daily life.[16,17] The release of these metals into the atmosphere is traceable in arctic ice cores, indicating the widespread dispersion.[18−20] This is also reflected in the increased expression of MTs and the documented accumulation of PGMs initiated in a wide variety of marine, shellfish, seaweed, and plant MTs, to changes in the mineral environment. This metal accumulation in organisms can be used as a natural marker for environmental toxicity.[21−25] Regardless, while the introduction of such foreign metals triggers an apparent defensive response, renal toxicity can still occur (e.g. Cd in MT, itai–itai disease). The long-term storage of Cd by MT in the kidneys raises the question about whether these xenobiotic metals will follow this pathway and bioaccumulate. For instance, platinum levels persist many years after anticancer treatment in the blood of urban women.[26]

While the forms of these metal pollutants are generally metallic and considered stable, it is important to remember that these metals are not normally found on the surface of the earth and that their introduction into the ecosystem through mining (runoff), processing, and subsequent consumer product disposal will ultimately cause physiological changes in affected life forms. Without previous knowledge or history with this type of pollutant, it is important to understand the possible response of biological metalloproteins, such as MTs, with metals that humans and the natural world have not traditionally but are now becoming increasingly exposed to.

Metallothioneins

MTs are ubiquitous metal chelators found throughout nature. The primary sequence of this family of small proteins (∼6 kDa) consists of 30% cysteine residues, making these proteins extremely air-sensitive, considering natively, no disulfide S–S bonds are present. The mammalian βαMT 1A isoform is a two domain protein (N-terminal β domain, 9 cys; C-terminal α domain, 11 cys), whose 20 cysteine thiolates participate in metal binding, transport, and recycling of physiological metals (Zn(II), Cu(I)), controlling cellular metal sinks and donating cofactor metals to metalloproteins as required.[27,28] MTs also sequester toxic soft metals, such as Cd in the first MT isolated (equine).[29] A build-up of these toxic metals can result in renal failure and death.[29,30] The regulation of the MT gene is controlled by cellular concentrations of free metals.[27] This characteristic and its down-stream effects have been implicated as a mode of resistance by MT in cancerous cells in response to anticancer metallotherapeutic agents.

Except for a few examples,[31−34] mechanistic details of reactions of xenobiotic metal complexes with MTs are largely unknown. Mammalian MTs are intrinsically disordered when metal-free, but the titration of M(II) to saturation produces two metal-thiolate clusters, one in each domain, with the 20 cysteines involved in a network of bridged and terminally bound sulfurs. These metals can be displaced by softer metals with a higher binding affinity. The order of increasing binding constants of metals with free thiols is as follows: Zn(II) < Cd(II) < Pt(II) < Bi(III) < As(III) < Cu(I) ≈ Hg(II).[30] Hg(II), and Cd(II), for example, can displace the Zn(II) in native Zn-MT, releasing free Zn(II) into the cytoplasm.[35] The Cd(II), Zn(II), and Cu(I) forms of MT are the most understood as they are most commonly found in nature. Metalation with other toxic or heavy metals results due to exposure, whether through pollution or drug administration—rare occurrences of metal exposure that are not understood until they are observed in effect. It is becoming clear through studies of metalation, where M ≠ Cd(II) or Zn(II), alternate mechanisms are in play, notably with the nonclustering As(III) binding,[36,37] or complex deconstruction of cisplatin.[33] While metalation of MTs with metal ions is well understood in vitro, the interaction of these proteins with nontraditional metals and metal complexes-carrying ligands are not fully understood.

Exposure to Xenobiotic Metal Complexes from Therapeutic Agents

Since the discovery of platinum based anti-cancer drugs, developed chemoresistance in certain cancer types has been associated with increased expression of MT.[38−41] MTs can interfere with a chemotherapeutic agent before it can reach its desired target or act in converse to its desired cytotoxic effect. Cellular thiols such as GSH and MT are correlated with induction and expression in resistant cancers, although their direct causation to resistance is unclear. MT expression levels are affected by a variety of stress inducers[21,39,41−51] including metal therapeutics, for example Au and Ag nanoparticles[52−54] and Pd, Pt, and Rh particles airborne through mining and industrial exposure.[15] Cisplatin resistance and its relationship with MT’s aggressive metal binding behavior is the focus of much research.[31,33,55−60] MTs are rich in highly reactive cysteine thiolates, which are suspected of being involved in the interference of platinum drugs. The many stress inducers of MT may easily cause upregulation of the protein following a dosage of platinum drugs. This trend has been identified by looking at MT mRNA expression and cellular MT levels in a variety of cisplatin resistant cell lines compared to their parent lines.[38,52,59,60] In medicine, new metallocomplexes show promising results in clinical trials, many of them consisting of PGMs and related metals.[61] This present study focuses on this new realm of “nontraditional metalation” to draw attention to the mechanism behind the metalation of xenobiotic metals to mammalian MTs.

In this study, we turn to the specific exposure of humans to a xenobiotic metal complex, where the clearest examples come from well-defined cytotoxic agents. The antitumor agent and common chemical catalyst dirhodium(II) tetraacetate (Rh2(OAc)4) and its binding reactions to MT is the focus of the study described here.

Cytotoxic Dirhodium(II) Tetraacetate as a Model Metal Complex

The American Food and Drug Administration’s approval of cisplatin in 1978 opened exploration of the cytotoxicity of metal-based complexes for cancer treatment.[62−64] Dirhodium(II) carboxylate complexes [Rh2(O2CR)4 (R = Me, Et, Pr)] are 18 electron systems when coordinated by solvent, involving octahedral coordination of each of the rhodium metals. Dirhodium(II) carboxylates are mildly water soluble and display high in vivo antitumor activity against multiple tumor lines.[65−68] Similar to cisplatin, these dimeric compounds also cause disruption of DNA replication and transcription, as well as creating DNA adducts and cross-linking.[69−73] Most notably, these antitumor dirhodium complexes irreversibly bind free cysteinyl thiols, inhibiting thiol-dependent enzymes.[66,68,74] Unlike other amino acids, the reaction of cysteines with these rhodium complexes results in a breakdown of the carboxylate cage structure by thiolate replacement of the O donor ligands.[67,73] Their sensitivity to sulfur coordination from cysteines in biomolecular targets means rhodium(II) carboxylates can also act as radiosensitizers by depleting cellular thiol sources, lowering the cell’s ability to respond to oxidative stress.[74,75] Current research has extended the applications of rhodium compounds to medical imaging, radioactive isotope therapy, photodynamic therapy, and protein labeling, but the significant antitumor characteristics of these rhodium complexes remain of great interest.[65,74,75]

Dirhodium(II) tetraacetate (Rh2(OAc)4), Figure , is the simplest of the dirhodium(II) carboxylates. Rh2(OAc)4 is used widely as a synthetic catalyst in many organic reactions, a unique form of potential occupational exposure. It is used to catalyze C–H, X–H insertion and aromatic cycloaddition and has a relatively high solubility in aqueous solutions compared to Cu(II) acetate. Rh2(OAc)4 and related dirhodium complexes have demonstrated significant antitumor activity; Rh2(OAc)4 administered to mice dramatically decreased the size of tumors.[65−68] Tumor growth was inhibited by the interruption of DNA replication and protein transcription. Exhaled 14CO2 from 14C-labeled Rh2(OAc)4 indicated that the complex was decomposed within the organism. Urine analysis showed that ∼1% of the Rh was excreted, indicating that the remaining rhodium irreversibly bound to its target and accumulated. However, the exact target of Rh2(OAc)4 activity is unknown, and toxic renal side effects were high.[65−68]

Figure 1

Ball-and-stick representation of (Rh2(OAc)4). Teal represents Rh, red represents O, gray represents C, and white represents H.

Because these dirhodium(II) carboxylates are uniquely sensitive to cysteine binding, MT is a likely binding target. Another major cellular thiol, GSH, was found to rapidly form adducts with Rh2(OAc)4 both aerobically and anaearobically.[76,77] If cytotoxicity by Rh2(OAc)4 involves DNA adduct formation or inhibition of specific enzymes, then MT would be acting as a pretarget interference source. Alternatively, if the cytotoxicity involves the depletion of cellular thiols, then MT would be the desired target. This can provide beneficial information for the design of future chemotherapeutics.

Scope and Application of Results

To prepare for the inevitable increased exposure of humans to xenobiotic metal complexes from many sources, it is imperative to understand their physiological chemistry. For the success of metal-based chemotherapeutics, it is imperative to understand drug metabolism and resistance mechanisms. The strong correlation between cellular thiol induction and chemoresistance requires the study of the metal binding pathways to the key thiol molecules, especially ubiquitous MTs. In this paper, the systematic and rapid deconstruction of the tetraacetate ligands in Rh2(OAc)4, and the robust Rh2 binding by the 20 cysteines in apo-βαMT is reported. The metalation speciation was modeled using semiquantitative electrospray ionization mass spectrometry (ESI-MS) data to obtain relative Kf values. Time-dependent ESI-mass spectral data show the sequential loss of the Rh2 core ligands. The results and their impact on our understanding the cytotoxic and environmental effects of the bioconjugation of MT and dirhodium complexes are discussed. In the text, “MT” refers to the 20 cysteine mammalian MTs, unless otherwise specified. “βαMT” refers to the human isoform MT1a as described in the Methods.

Results and Discussion

Metalation Reactions of βαMT with Rh2(OAc)4: Optical Spectroscopic Properties

Figure , top, shows the UV–visible absorption spectral data for the titration of apo-βαMT with increasing mole equivalents of Rh2(OAc)4 added. Significantly, the UV–visible absorption data show an increased absorption at ∼300 nm, corresponding to the well-known cysteine S → Rh ligand-to-metal-charge-transfer (LMCT) band. Similar absorption is also observed with GSH, cysteine, methionine, and other biological thiols[76−82] and confirms the cysteine thiol as the binding moiety involved in the Rh2 complexation. This LMCT band is very strong and the transition from the blue of the metal solution to the yellow of the protein-bound product can be followed readily with the naked eye.

Figure 2

UV–visible absorption (top) and CD (bottom) spectra for a titration of 25 μM βαMT with the black arrow showing the direction of increasing molar equivalents of Rh2(OAc)4. The S → Rh LMCT band increases in absorbance at ca. 300 nm.

When compared to previous results involving Rh2(OAc)4 and the 9 cysteine apo-βMT,[78] the weak absorption caused by the Rh–Rh bond ∼580 nm is not apparent with the full, 20 cysteine apo-βαMT. This is likely due to the broad 20 cys S–Rh absorbance overlapping the weak Rh–Rh absorption. While previous knowledge leads us to presume that the Rh–Rh bond remains intact, we must rely on other data to confirm this.

In the case of the circular dichroism (CD) spectral data (Figure , bottom), it is surprising that no significant spectral intensities are observed during the titration of the full apo-βαMT. However, the lack of a distinct CD spectral intensity is likely due to the overlap from the presence of several structurally different species because the mass spectral and structural cysteine modification data indicate that the Rh2(OAc)4 has bound to the apo-βαMT. This contrasts the results previously reported for Rh2(OAc)4 binding to the 9 cysteine β-domain fragment, in which a single product is formed in a 1:1 fashion, as confirmed by ESI-MS, producing a strong CD signal.[78]

Early Reaction Steps and Initial Deconstruction of Rh2(OAc)4 by Apo-βαMT: Rapid Displacement of the Tetraacetate Ligands

Figure shows a series of ESI-mass spectra as a function of increasing molar ratio of Rh2(OAc)4/MT (from 0 to 1 mol equiv). The data shown in Figure provide a snapshot of speciation at very low molar ratios during the early stages of apo-βαMT (6875 Da) metalation, described by the reaction in 1. As previously demonstrated with cisplatin,[33] the apo-βαMT rapidly (within the dead time of the instrument) engulfs the incoming Rh2(OAc)4 molecule (441.99 Da) and then systematically shreds the ligands from the metal complex until only the Rh–Rh core is sequestered within its thiolate binding site. The mass spectral data show the presence of the incoming complete complex and the final isolated Rh2 core bound to the MT. The protein species with masses of 7316, 7517, and 7758 Da are representative of Rh2(OAc)4-MT, Rh2(Rh2(OAc)4)1-MT, and (Rh2(OAc)4)2-MT, respectively. These dominant masses gradually diminish as the acetate ligands are displaced, until the masses of the final products of Rh2MT and Rh4MT, at 7074 and 7273 Da, respectively, dominate. The change in mass coincides with the mass differences between the lost acetate moieties from the Rh2(OAc)4 bound protein species, where each acetate group has a mass of 59.04 Da.

Figure 3

Initial deconstruction of Rh2(OAc)4 by apo-βαMT observed with ESI-MS. Stepwise addition of 0.0–1.0 mol equiv of Rh2(OAc)4 to 20 μM apo-βαMT. Major species are indicated by dashed lines: black for apo-βαMT, red for Rh2MT, blue for Rh4MT, and green for species bound to at least one complete Rh2(OAc)4 molecule. Carboxymethylated MT species are mass-shifted by ∼60 Da and are represented with a triangle.

Spectroscopically, the Rh–Rh bond can be determined by a very weak absorbance at ∼580 nm in solution. However, this band is overlaid by the strong S to Rh LMCT, when the Rh2(OAc)4 binds both initially and finally to the thiols in MT. Previous studies with thiol complexation (e.g., GSH, methionine, and so on) indicate that the thiols coordinate to the axial positions and do not disrupt the Rh–Rh bond.[77,81,82] Because of the properties of MT metalation, we can infer that this bond has not been broken. If the metals were separated into mononuclear metal centers, it would be expected that the metal exchange reactions between MT molecules observed with Cd, Hg, Cu, and so forth would also be observed with Rh. This would result in a distribution of metals, including stoichiometric ratios with odd numbers. However, only pairs of Rh (i.e., Rh2) are ever observed in mass differences between species, indicating that the Rh–Rh remains locked in a bimetallic state. It remains unclear how the Cys thiols coordinate the Rh, whether terminal cysteines are involved, or bridging S coordination as suggested by Dunbar et al.[83]

Experimental and Simulated Formation of Rh2MT, Rh4MT, and Rh6MT

The formation of Rh2-metalated βαMT products can be described by three stepwise bimolecular reactions, as shown in Scheme . Each metalation event begins with encapsulation of the whole molecule, followed by rapid deconstruction of the ligand structure to the dirhodium core. The metalation continues with the formation of the ligand-stripped Rh4MT (7273 Da) and Rh6MT (7473 Da). Figure shows that a series of stepwise metalation reactions as from 0.0 to 2.0 mol equiv of Rh2(OAc)4 were added to apo-βαMT under equilibrium-binding conditions. The major product of the reaction is Rh4MT, which forms rapidly, with a small amount of Rh6MT.

Scheme 1

Proposed Bimolecular, Stepwise Metalation Reactions for Rh2(OAc)4 Binding to Apo-βαMT

K1–3 represent the stepwise equilibrium binding constants for each reaction.

Figure 4

ESI-mass spectral data recorded during the titration of 20 μM apo-βαMT with aliquots 0.0–2.0 mol equiv Rh2(OAc)4. The major species are indicated by dashed lines: black for apo-βαMT; red for Rh2MT; blue for Rh4MT; and green for Rh6MT. The inset bar graphs on the right of each spectra show simulated mass spectral data based on the model described in figure, using Kf values in Table . Carboxymethylated βαMT species are mass-shifted by ∼60 Da, are represented with a triangle, and are not included in the simulation.

Table 1

Relative Stepwise Equilibrium Binding Constants for the Reaction of Rh2(OAc)4 with Apo-βαMT at pH 7.4 from Simulated Data

MT(Rh2)n	stepwise log Kf
1	4.22
2	3.00
3	1.67

The further addition of 3.0 and 4.0 mol equiv of Rh2(OAc)4 did not result in any further significant changes to the mass spectra (see Supporting Information for data). This is unexpected as MT is known to be fluxional enough to adopt a high M(I) and M(II) to MT stoichiometry. Because metalation does not continue, it is clear that the structure has reached a maximum capacity. The intensity of each metalated state is related to their equilibrium rate constants, when described as bimolecular reactions, as shown in Scheme . These Ks will allow us to understand the distribution we see, with only a small fraction of highly metalated MT species.

Using the Hyperquad simulation speciation[84] software, the mass spectral abundances associated with these reactions were modeled based on the concentrations and mole equivalences of the Rh2(OAc)4 added, using the equations in Scheme . The three relative stepwise K values (shown in Table ) were calculated so that the simulated model matched the experimental data, Figure . The calculated abundances were used to simulate the mass spectra as seen in the red bar graph Figure insets. The simulated mass spectra closely match the experimental MS data over the course of the titration.

Figure shows the experimental speciation data extracted from Figure . The sequence of speciation clearly follows a typical noncooperative metalation mechanism[85] in which Rh2MT forms, followed by Rh4MT and by Rh6MT. The fitted results, the smooth lines in Figure , match the experimental data closely and provide the relative binding constants as shown in Table .

Figure 5

Experimental mass spectral speciation abundance data (square markers) and simulated data (lines) calculated from the fitted relative Kf values (Table ). The root mean square error = 0.04699 (apo-MT); 0.04878 (Rh2-MT); 0.07370 (Rh4MT); 0.028875 (Rh6MT).

Figure shows the absorbance data of the S → Rh LMCT band at 300 nm plotted against the addition of metal. The rise in absorbance is correlated with the appearance of metalated species (RhMT) in the experimental mass spectrum, Figure , indicating that the mass spectrum reflects the solution phase species.

Figure 6

Absorbance data of a titration from 0 to 2.0 mol equiv Rh2(OAc)4 overlapped with mass spectral speciation data of a titration of apo-βαMT from 0 to 2.0 mol equiv Rh2(OAc)4. The % absorbance (blue triangles) shows the increase at 300 nm from the absorbance data in Figure aligning with the % metalated species formed (black squares). The % speciation for apo-βαMT is shown with red circles.

Nontraditional Metalation and Metal Complex Deconstruction Is Evident from the Mass Spectra

Mass spectrometry is especially helpful in identifying biological targets.[86] ESI-MS is a soft ionizing, high-resolution, solution-phase technique that is ideally suited for tracking the metalation status of MT. The advantage of using ESI-MS is that it allows for confirmation of the presence of intermediate states in real time. These versatile properties have been widely exploited in native ESI-MS,[87] allowing detailed kinetic and equilibrium studies of the metalation of MTs.[78,85,87−98] Binding coefficients can be accurately extracted from these data. The observed kinetic binding of Pt(II)1–7 from cisplatin was first determined by chromatographic methods,[99] which were then enhanced by our ESI-MS modeling methods,[33] a similar methodology to that applied to this current dirhodium study. The stoichiometric and structural information provided by the ESI method is particularly important for MT because the lack of typical protein features and its high sensitivity to oxidation makes MT’s protein structure difficult to observe with crystallography, NMR, or other structure elucidating methods.

The conditions for cooperative and noncooperative bimolecular metalation of Zn(II), Cd(II), and Cu(I) to MT have been studied with consistent stoichiometric ratios identified primarily using ESI-MS, including traditional metalation to form stable cluster states M(I)6S9, M(II)3S9, and M(II)4S11.[37,85,88,89,91−93,100]

The presence of metal-thiolate (Zn(II), Cd(II), Cu(I)) clusters that form in higher ratios of these metals make MTs resistant to trypsin digestion.[94] However, it should be noted that As(III) binding is not considered to involve cluster formation. The As(III) bound to the isolated MT fragments M(III)3S9 in the β domain fragment and also M(III)3S9 in the α domain fragment and to the full protein forms M(III)6S18,[37] which means that the MT peptide flexibly accommodates a completely different style of binding than reported for the group 12 M(II) ions. Our results show that Rh2 binding also does not follow traditional cluster formation of Zn(II), Cd(II), and Cu(I), as it binds to MT in a rigid manner. MT is unable to perform its usual metal relocation, as described in Chart . Therefore, a novel MT binding mode is now reported for Rh2 in this study, which enforces the difference between the metalation of complexes and metal ions by MTs.

Chart 1

Proposed Pathway for Metalation of βαMT 1a by Rh2(OAc)4 at Physiological pHa

The two pathways shown apply to different configurations. Pathway A: Rh2MT binds an additional Rh2(OAc)4 which is deconstructed, forming Rh4MT of configuration I. No further metalation occurs at this point. Pathway B: Rh2MT binds an additional Rh2(OAc)4 which is deconstructed, forming Rh4MT involving a proposed alternate configuration II. This form can proceed to bind a further Rh2(OAc)4, producing Rh6MT.

The metalation of Rh2(OAc)4 to apo-βαMT can be categorized as nontraditional metalation, because the binding of MT involves chelating Rh2(OAc)4 followed by disassembly of the initial metal complex by displacement of the original coordinated ligands. ESI-MS has been used to observe this destructive effect MT has with cisplatin and Ru-based arene complexes.[31] This step-wise, sequential deconstruction of Rh2(OAc)4 has been reported previously with the 9 cysteine single apo-β-domain fragment of MT1a. Under those conditions, one discernible product (Rh2MT) is formed in a linear fashion.[78] In this present ESI-MS study, the full, two-domain protein has a higher sulfur availability (20 cysteines), which greatly increases the rate of reaction and the number of possible conformers (Rh2MT, Rh4MT, and Rh6MT) that can form. The ligand loss of the coordinating acetates and accumulation of the sequestered metal species show that Rh2(OAc)4 does not survive intact when bound to MT. This emphasizes the destructive effect MTs can have on both chemotherapeutically related complexes and xenobiotic metals in inorganic and bioinorganic forms.

pH-Dependent, Metal-Dependent Metalation Mechanism of MTs

Figure shows the time-dependent metalation of the 11 cysteine single domain apo-αMT fragment at pH < 2 with Rh2(OAc)4. This pH is extremely low; Cd(II) cannot bind to MT due to proton competition under these conditions. However, the results how, despite the extreme condition, the S–Rh bond formation forms favorably. Over the time course of the reaction (44 min), species ranging from [Rh2(OAc)4]1-MT to [Rh2(OAc)4]6-MT are observed. This is completely different when compared with the results from the titration carried out at pH 7.4 with the two-domain protein (Figure ), which resulted in rapid metalation and systematic ligand loss.

Figure 7

(Left) Time-dependent mass spectral data of apo-αMT at low pH (<2), recorded over 44 min following addition excess Rh2(OAc)4. Data acquisition times are (averaging times shown) at 0 min; 17−19; 23−25; 33−34; 34−35; and 42−44 min. Red arrows indicate mass shift resulting from the loss of (OAc)4. (Right) Labels for identifiable species in the mass spectral data and proposed formation pathways.

In Figure , the key species identified are apo-αMT, [Rh2(OAc)4]1-αMT, [Rh2(OAc)4]2-αMT, [Rh2(OAc)4]3-αMT, [Rh2(OAc)4]4-αMT, [Rh2(OAc)4]5-αMT, [Rh2(OAc)4]6-αMT, Rh2[Rh2(OAc)4]1-αMT, Rh2[Rh2(OAc)4]2-αMT, Rh2[Rh2(OAc)4]3-αMT, Rh2[Rh2(OAc)4]4-αMT, Rh2[Rh2(OAc)4]5-αMT, Rh2-αMT, and Rh4-αMT.

These data show that the unfolded, loose structure caused by the high [H+] greatly increases cysteine accessibility to the incoming Rh2(OAc)4 in this single domain fragment. In the presence of excess Rh2(OAc)4, this manifests as two distinct reaction pathways: (i) ligand replacement by thiolates following Rh2(OAc)4 binding, as described for the whole protein above, and (ii) proposed axial coordination of Rh2(OAc)4 by cysteine. These two reaction pathways are outlined in Figure , right. The acidic conditions minimize the nucleophilicity of the cysteinyl thiols and allow for the lower Cys/metal ratios to bind with metal ratios of up to 6× Rh2(OAc)4 per MT molecule. Despite the observation of this “supermetalation,” the cysteine replacement of the tetraacetates still occurs albeit significantly slower than at physiological pH.

The low pH experiment shows that while the unfolded, 11 cysteine apo-αMT can bind to the Rh2(OAc)4, it is unable to effectively displace the coordinating ligands of the complex. Instead, the unfolded peptide quickly binds to 1–6 Rh2(OAc)4 species which occupy and lock the cysteine thiolates from further reactivity. Significantly, the MS data recorded at pH < 2 demonstrate that Rh2(OAc)4 has a strong binding affinity, surpassing that of Cd, which is completely demetalated from MT at this pH. These acid-denatured conditions impede MT from attaining efficient metal complex metabolism. This contrasts the results at physiological pH, at which the more folded MT structure can orient its cysteines to displace the coordinated ligands more effectively than the unfolded protein. The key message from the data reported here is that the initial intrinsically disordered MT structure at physiological pH can flexibly entrap metal complexes and is necessary to further react, deconstruct, and isolate the metal for storage. We describe these properties as nontraditional binding of MTs, and we expand on this subject below.

Cysteine Modification by p-Benzoquinone and N-Ethyl Maleimide Titration: Structural Information Regarding Solvent Accessibility

The 2° and 3° structures of human MT are only defined by metalation,[101] as it exhibits no 2° structural elements or aromatic residues. A broad distribution of metalated and metal-free MT species exists in the cell.[88,89,92,93,102−104] Despite the sensitivity of the free cysteines to oxidation, fluorescence labeling has demonstrated the presence of metal free apo-MT in vivo in roughly equal quantities to holo-ZnMT, both in the oxidized and reduced form.[105] This confirms that metalation of MT is a posttranslational modification that can only occur once the protein is fully synthesized because of the cross-linking nature of metal–cysteine coordination that forms the metal-thiolate clusters. To better understand the metalation reaction, we use the fully demetalated apo-βαMT as the model for the partially metalated MTs expected in vivo.

Typically, protein structural information can be obtained by analyzing the charge state distribution patterns in the mass spectral data. Because of the small size of MTs, this method is not as effective, especially to probe the apo-protein or partially metalated states. More recently, however, structural information about MT has been obtained through the reaction profiles of cysteine modifiers that take advantage of MT’s cysteine rich structure. p-Benzoquinone (pBQ) and N-ethyl maleimide (NEM) are organic agents capable of binding covalently with free thiols, as shown in Scheme . Similarly to Ellman’s reagent (5,5′-dithiobis(2-nitrobenzoic acid)), this irreversible binding permits quantification of free thiols.[37,92,106−108] This is observed as a predictable mass change, but in addition with respect to MT and its 20 cysteines, the modified speciation distribution can also provide information about the solvent accessibility of the cysteines.[108] This technique has been used to understand metalation of apo-βαMT through partially metalated and fully metalated Cu6MT, Cu13MT, Cu20MT, and As6MT.[89,92]

Scheme 2

Structure and Reaction Mechanism Involved in the Formation of Covalent Bonds between Cysteine and Modifiers pBQ (A, Top) and NEM (B, Bottom)

In this present study, we used both pBQ and NEM in conjunction with mass spectrometry to quantify the number of free cysteines and to probe the surface structure and the solvent accessibility of the partially metalated states of Rh2MT, Rh4MT, and Rh6MT. Structural characteristics can be inferred from the reaction profile of modification observed in the mass spectral data.[106] As expected, the higher metalated Rh4 and Rh6-bound MT shows less solvent accessibility because of the lower spread of the p(BQ)-modified species. The most solvent accessible species, Rh2MT, shows two distinct distributions, indicating that two metalated states are created. From the change in intensity for the higher metalated states, we can infer that only one of the observed Rh2MT species is capable of further metalation, as described pictorially in Chart . This rigidity is reflected again with the lowered abundance of Rh4MT and Rh6MT and in the decrease of the relative log Kf values showing that progressive metalation is less favorable.

The data in Figure show modification of MT containing Rh2MT, Rh4MT, and Rh6MT by both pBQ (Figure A,B, left half) and NEM (Figure , right half). The Rh2MT, Rh4MT, and Rh6MT spectra measured prior to modification are shown in Figure , top (A,C). In B,D, these same species are present (identified by triangles) even though excess modifier was added. Both sets of modified species (Figure , lower) display a variety of normally distributed species, as shown in Figure in bar graph format (see Supporting Information for experimental data). As described by Irvine and Stillman, cysteine accessibility can be understood by the distribution of cysteine modified species.[106] A normal distribution represents statistically equal accessibility of the modifier to all available free cysteines. This means that the unbound cysteines are exposed. Therefore, the breadth or spread of the distribution tells us the number of available cysteines, and the shape of the distribution tells us if there is conformation specificity or equal accessibility of the free thiols. Figure B shows the distribution of the modified Rh2(BQ)MT species (red bars). The data indicate the presence of two species, one with x = 0–8 free cysteines and another with x = 10–15 free cysteines. This means that there are two distinct Rh2MT products: one coordinating the Rh2 with 12 cysteines and the other coordinating the Rh2 with 5 cysteines. This is harder to distinguish at increased concentrations as the abundancies of the two distributions overlap each other, Figure D. The steep rise of the modified Rh4(BQ)MT (blue) also displays a normal distribution with y = 0–7 free cysteines. The formation of a modified Rh6(BQ)MT species with z = 0–4 free cysteines also can be observed (green).

Figure 8

ESI-mass spectral data of modified and unmodified mixed metalated species. (A) Nonmodified Rh2-, Rh4-, and Rh6-MT, represented by a triangle. (B) Solution in A following addition of pBQ. (C) Nonmodified Rh2-, Rh4-, and Rh6-MT, represented by a triangle. (D) Solution in C following addition of NEM. The experimental data are shown in bar graph representation (see Supporting Information for experimental data). Persistent titrations to excess and long incubation times did not affect the resulting spectra indicating completed reaction. Rh2MT species shown in red, Rh4MT species in blue, and Rh6MT species in green.

There is a small discrepancy between the number of modified cysteines observed using pBQ and NEM. This is likely due to the reaction occurring at physiological pH, at which pBQ is not as effective as NEM.[108] Therefore, our conclusions about the number of free cysteines for the metalated MT is based on the greatest number of NEM-modified cysteines for each Rh-bound complex.

Three-Dimensional Models: Cysteine Accessibility throughout Metalation

Three-dimensional structures constructed using molecular mechanics (MM)/molecular dynamics (MD) modeling techniques from previously reported AsMT[109,110] and CdMT[111,112] metalated forms were adapted to examine the possible metal thiolate constructs MT may adopt when bound to dirhodium cores. The absence of aromatic and hydrophobic residues and the absence of disulfide bonds mean that the driving force for protein folding is the cross-linking that results from the metalation of the cysteines, stabilized by hydrogen bond formation. As the cysteine modification results and the MD structures indicate, metal-free cysteine thiolates are found to be solvent accessible,[113] while holo-cysteines are buried within the protein structure. MD calculations on the holo-Rh6MT show no change in the conformational geometry, demonstrating the holo-MT’s rigid status.

Representative structures of apo-βαMT, Rh2MT, Rh4MT, and Rh6MT are shown in Figure . Energy-minimized geometries using MD for apo-βαMT, Rh2MT, Rh4MT, and Rh6MT were constructed in Scigress 6.0.0 and geometry optimized using molecular mechanics (MM3) prior to MD calculations at 300 K for 500 ps. These structures are shown in ball and stick format in Figure and in full page format along with H-bond mapping in the Supporting Information.

Figure 9

Energy minimized MD models of apo-βαMT (A), Rh2MT (B–D), Rh4MT (E–G), and Rh6MT (H). Rh shown enlarged in teal, cysteines shown enlarged in yellow, C in gray, O in red, N in blue, and H in white. Structures are based on Cd7MT structures from previously reported,[111,112,116] using Scigress 6.0.0 details of the minimization are given in the text.

The apo-βαMT model, Figure A, shows the structure of the nonmetalated cysteines and how MT orients itself with respect to metalation. The energy minimization of apo-βαMT results in a compact globular structure, in accordance with mass spectral studies at physiological pH.[113] The cysteine residues (−SH shown with yellow spheres in Figure ) are oriented to the exterior when metal-free, but coalesce toward the interior of the protein when metal-bound. In the apo-βαMT structure, the backbone and cysteines arrange themselves in a similar fashion to a helix, with cysteines facing outward. This agrees with results from Irvine et al. and Rigby et al. that the cysteine solvent accessibility in the apo form facilitates metal binding.[101,114,115]

The driving force for organized structural formation in mammalian MTs is the metal-induced folding with each sequential M–SCYS chelation event.[101] Once bound to a metal, these cysteines become buried within the protein interior, shielding them from the cellular environment. The fully bound Rh6MT (Figure H, “Rh2β4αMT”) shows this clearly. The configuration of the peptide backbone shows that MT wraps around the metal core upon binding. This effect is observed in the partially metalated Rh2MT and Rh4MT structures as well, in Figure . When the Rh2 is bound to cysteines that nominally form a single domain (Figure ; Rh2βMT, Rh2αMTI, Rh2αMTII, and Rh4αMT), the metal-bound cysteines are buried, while the remaining metal-free cysteines maintain their external orientation. This reflects the globular protein structure seen in the apo-MT model, Figure A. However, these one-domain-bound MTs are fluxional and capable of further subsequent metalation. In contrast, for two-domain metal binding, like in the Rh2β2αMTI, Rh2β2αMTII, and Rh6MT, a dumb-bell-like structure typical of M(II) ion metalation is obtained.[88] When both domains are bound, the cysteine accessibility is greatly diminished. We consider that this accounts for the lower abundance of the Rh6MT species in the mass spectral data.

Nontraditional Metalation and Comparison of Binding Constants across Metals

Metal binding to MT is generally fluxional such that the metal ions can move from site to site in search of the lowest thermodynamic product (e.g., the well-known change in Cu(I)-MT emission signal when Cu(I) is titrated into Zn-MT 1A or Zn-MT 2A).[117] In the case of the dimeric rhodium binding, the bound cysteines form multiple metalated species (Chart ). The orientation of the initial metal binding in relation to the MT cysteines defines how the metalation proceeds. Because of the decreased abundance of Rh4MT and Rh6MT, the domain location and formation of Rh2MT determine whether the metalation continues to Rh6MT. The two distributions for the Rh2(BQ)MT-modified speciation data reflect these products. The decrease in relative log Kf’s demonstrates its rigidity as metalation progresses (Table ). From previous work with the single β domain fragment, we know that the β-domain accommodates 1 Rh2 moiety.[78] The 11 cysteines of the α reside can accommodate the remaining 2Rh2 if metalated efficiently (Chart , right pathway B). If not, the cysteines become too tangled to accommodate further metals (Chart , left pathway A).

The relative Kf’s reported in Table must be in absolute values, be greater than the Kf of Cd, as Rh2(OAc)4 binds MT at a lower pH than Cd. However, at this time, we do not know the absolute values of this complicated reaction. Directly comparing the metalation reaction between chemically different metals is difficult, because of both different binding geometries and the presence of competitive coordinating ligands. A comparison of metalation pathways can only be made meaningfully between isomorphous metals. Just as Cu(I) metalation of MT differs from Cd(II) metalation, metal ion accumulation like this in MT cannot be compared to the non-traditional, deconstructive process that MTs exert on ligated metal complexes, as described here for Rh2(OAc)4.

Conclusions

We report the reaction of a xenobiotic complex with MT as a test of the possible chelation properties of MTs with transition metal complexes. Rh2(OAc)4, a robust synthetic catalyst, binds intact to the metal-free apo-βαMT. The tetraacetate ligands are rapidly removed upon binding by the folded MT at physiological pH, in a manner reported previously for cisplatin. The initial reaction of the apo-βαMT involves the displacement of tetraacetate groups by the cysteinyl thiolates. The location and space within MT that the first Rh2 binds in the protein, which domain, governs the remaining space in which metalation can continue. This intermediate state can then proceed toward two structures, one of which allows the metalation of a third dirhodium moiety, forming Rh6MT, and the other pathway stops at Rh4MT. These branching pathways were determined using complementary techniques of cysteine modification and mass spectrometry, together providing speciation data for equilibrium reaction modeling.

The increase in global exposure to anthropogenic xenobiotic metals demands updated understanding of the metabolic effects caused. Rh and PGMs are known to exert cytotoxic and nephrotoxic activity, through metabolic process that can increase the bioavailability. Our report demonstrates the ease with which biological structures can greatly change the chemical activity of rhodium. A metal’s persistence through biological metabolism in ever changing chemical forms emphasizes the need for long-term study and caution for the continued utilization of these elements for human health.

Finally, of concern, we note that long term accumulation and storage of PGMs can provoke chronic toxic effects. Cd-bound MT has a reported half-life in the kidneys exceeding 20 years.[118] Because of the firm and rigid binding of Rh2(OAc)4 to MT, Rh would be expected to accumulate in a similarly dangerous manner, hindering normal MT function. The heavy and widespread use of PGMs and their related metals and contaminants pose a risk that cannot be ignored, as history would remind us with the use of lead in gasoline and paints causing widespread toxic effects. This long-term storage is also applicable for these xenobiotic metals. The bioaccumulation of metals can pose a significant health risk in the future.

Methods

Preparation of Apo-βαMT

Recombinant human MT 1a (rh-βαMT 1a, referred when used in this study as “MT” unless specified otherwise) was overexpressed with an S-tag in Escherichia coli. The S-tag was removed and the MT purified according to previously described methods.[89,111] The cleaved construct has the sequence GSMGKAAAACSCATGGSCTCTGSCKCKECKCNSCKKSCCSCCPMSCAKCAQGCVCKGASEKCSCCAKKAAAA. This construct contains two mutations when compared with the human MT 1A sequence archived on UniProt protein database, T27N and I51V. These mutations are commonly found in other mammalian MTs including other human isoforms (see Supporting Information Figure S2 for sequence comparison) and are not involved in metalation reactions of the cysteinyl thiolates.[119] The purified Cd-bound MT1a was demetalated by acidification (pH < 2), and the resulting free Cd(II) in solution was removed by centrifugal filtration (Amicon Ultra-4 3000 Da MWCO). Protein concentration was calculated by remetalating a measured aliquot with CdSO4 and determining the Amax at the 250 nm shoulder characteristic of the LMCT band of Cd–S bonds (ε = 89 000 L mol–1 cm–1).[120]

Because of the highly sensitive nature of MTs to oxidation, all solutions are thoroughly evacuated and saturated with Ar.

UV–Visible Absorption and CD Spectroscopy

UV–visible absorption and CD spectral data were acquired on Cary UV Bio50 and JASCO J-810 spectropolarimeters, respectively. Solutions were evacuated and backfilled with Ar gas prior to data collection and measured in a sealed, 1 cm × 1 cm quartz cuvette.

ESI-MS Studies

Solutions of Rh2(OAc)4 (Sigma-Aldrich) were prepared in deoxygenated, argon-aerated, deionized water. Aliquots containing known molar equivalents were added to the apo-MT solution immediately prior to mass spectral data acquisition using a micrOTOF II (Bruker Daltonics, Toronto). All solutions were at room temperature and thoroughly deaerated using vacuum evacuation followed by Ar saturation. NaI was used as the calibrant. Spectra were collected in positive ion mode, as a function of time following mixing. The settings used are described in the Supporting Information. The averaged spectra and data analysis were carried out using the maximum entropy application in Bruker DataAnalysis 4.2 program. The resulting spectral data were normalized, and the dominant species were identified by mass.

Cysteine Modification by pBQ and NEM

Stock solutions of pBQ and NEM (10 mM) were dissolved in 10% v/v methanol in water, and the solution vials were wrapped in aluminum foil to protect from photochemical degradation. Aliquots of the modifiers were added to protein samples containing Rh2, Rh4, and Rh6MT, and mass spectral data of the resulting products were obtained to identify the species present. The modifier was titrated to excess, until the mass spectra no longer changed between additions.

MD Models of Rh2MT, Rh4MT, and Rh6MT Structures

Molecular modeling calculations were carried out using Scigress Version 6.0.0 (Fujitsu Poland Ltd.). Structures, modeling parameters and sequence information were adapted from previously reported Cd7MT and As6MT models to build the Rh2MT models.[111,112,121] In brief, the Cd(II) ions were deleted and replaced with Rh(II) prior to molecular mechanics calculations (MM3). MD calculations were carried out at 300 K for 500 ps using the dielectric constant for water of 78.5. Metal thiolate structures in the α domain were based on bridging thiol arrangements in Cd7MT.

Methodology

Spectroscopic and mass spectral studies provide the identity of the binding moiety and detailed speciation data. The mass spectral data were simulated computationally to determine the stepwise equilibrium constants. Cysteine modification was used initially to determine the metalation stoichiometry by quantifying the number of cysteines not involved in binding the Rh2(OAc)4. However, these results directed the research focus toward understanding the protein structural changes in metalation, using MD modeling and mass spectral titrations under unfolded conditions. Altogether, these experiments fully document the metalation process by which xenobiotic dirhodium(II) tetraacetate is accumulated by MT.