Reaction Kinetics of Cyanide Binding to a Cobalt Schiff-Base

The kinetics of the reaction of cyanide with Co(II)N4[11.3.1] under anaerobic .... 6.13 mmol) was then slowly added, under argon, over the course of s...
0 downloads 0 Views 938KB Size
Download PDF
Article Cite This: Chem. Res. Toxicol. XXXX, XXX, XXX−XXX

pubs.acs.org/crt

Reaction Kinetics of Cyanide Binding to a Cobalt Schiff-Base Macrocycle Relevant to Its Mechanism of Antidotal Action Hirunwut Praekunatham, Linda L. Pearce,* and Jim Peterson*

Downloaded via NOTTINGHAM TRENT UNIV on August 13, 2019 at 20:43:52 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

Department of Environmental and Occupational Health, Graduate School of Public Health, The University of Pittsburgh, Pittsburgh, Pennsylvania 15261, United States ABSTRACT: The Co(II/III)-containing macrocycle, cobalt 2,12dimethyl-3,7,11,17-tetraazabicyclo-[11.3.1]-heptadeca-1(17)2,11,13,15-pentaenyl cation, or CoN4[11.3.1], is a potential cyanide-scavenging agent. The rate of reduction of Co(III)N4[11.3.1] by ascorbate is reasonably facile under pseudo-firstorder conditions; a second-order rate constant of 11.7(±0.4) M−1 s−1 was determined at 25 °C and pH 7.4, along with the activation parameters for the reaction (ΔH⧧ = 53.9(±0.8) kJ mol−1; ΔS −79(±3) J mol−1 K−1). It follows that any cyanide-decorporating capability of the cobalt complex should depend more on the cyanide-binding characteristics of Co(II)N4[11.3.1] than the oxidized form. The kinetics of the reaction of cyanide with Co(II)N4[11.3.1] under anaerobic pseudo-first-order conditions is rapid and resulted in a linear dependence on the cyanide concentration, kHCN = 8 × 104 M−1 s−1, with a nonlinear intercept of 420 s−1 at 10 °C, pH 7.6. The observed reaction rate increases significantly with increasing pH. A rate law is suggested, kobs = k′[X] + (kHCN + kCNKa/[H+])[HCN], where kCN is estimated to be ∼2 × 106 M−1 s−1. Activation parameters for the reaction with HCN (ΔH⧧ = 10.7(±0.4) kJ mol−1; ΔS⧧ = −153(±1) J mol−1 K−1) suggest an associative mechanism. In the presence of excess oxygen, i.e., at higher levels than free oxygen in vivo, the reaction rate was too fast to be measured, and the final product was the oxidized complex, Co(III)N4[11.3.1], where any cyanide ligands had been lost. This is much more rapid than the oxidation of the parent compound by oxygen, for which a second-order rate constant of 0.5(±0.02) M−1 s−1 at 25 °C was obtained. The study has gone some way toward enhancing our understanding of the reaction of Co(II)N4[11.3.1] with cyanide. The fast reaction rate implies a high efficacy of the cyanide-scavenging capability of the complex and further supports the suggestion stemming from our previous work that Co(II)N4[11.3.1] could prove to be a better and more cost-effective cyanide antidote than the FDAapproved hydroxocobalamin.

■

INTRODUCTION While a number of cobalt complexes have been shown to exhibit cyanide-scavenging properties, only hydroxocobalamin has, so far, been FDA labeled1 for use in ameliorating the harmful effects of cyanide toxicity in patients without causing life-threatening side effects.2,3 Despite its apparent safety and clinical success, cobalamin is expensive and requires the slow (∼15 min) intravenous administration of large volumes.4 Finding less expensive and more rapidly effective cyanide antidotes would clearly be beneficial. Recently, a cobalt complex of the Schiff-base macrocycle 2,12-dimethyl3,7,11,17-tetraazabicyclo[11.3.1]heptadeca-1(17)2,11,13,15pentaene (CoN4[11.3.1], Figure 1 inset) was found to be a viable option as a cyanide antidote.5 CoN4[11.3.1] is a lowmolecular-weight compound that binds two cyanide anions (cobalamin binds only one cyanide under physiological conditions) and is synthesized in a single step with affordable starting materials. In a sub-lethal mouse model, CoN4[11.3.1] was more effective at ameliorating cyanide toxicity when compared to cobalamin, its precursor cobinamide, or a cobaltcontaining porphyrin5−,7 and no long-term sequelae were observed in these mice more than a week later.6 Due to its relatively low molecular mass (∼4 times smaller than cobalamin),5,6 CoN4[11.3.1] can be expected to exhibit greater © XXXX American Chemical Society

Figure 1. Electronic absorption spectra of CoN4[11.3.1] derivatives, 1.00 cm path length, 20 °C. Main panel: Co(II)N4[11.3.1] (solid line), Co(III)N4[11.3.1](CN)2 (dashed line), and Co(III)N4[11.3.1] (dotted line). An aqueous 0.6 mM Co(III)N4[11.3.1] solution (0.1 M sodium phosphate buffer, pH 7.4) was reduced by the addition of sodium L-ascorbate to 3 mM under anaerobic conditions, and then HCN was added to 12 mM. Inset: The chemical structure of di- and trivalent CoN4[11.3.1] cations.

Received: April 18, 2019 Published: June 26, 2019 A

DOI: 10.1021/acs.chemrestox.9b00170 Chem. Res. Toxicol. XXXX, XXX, XXX−XXX

Article

Chemical Research in Toxicology

on the method of Long and Busch,19 adapted by Lacy et al.20 Briefly, CoBr2 (1.35 g, 6.17 mmol) and 2,6-diacetylpyridine (1.00 g, 6.13 mmol) were dissolved in 20 mL of degassed ethanol, and 0.5 mL of water was added under argon at room temperature. Deoxygenated 3,3′-diaminodipropylamine (0.857 mL, 6.13 mmol) was then slowly added, under argon, over the course of several minutes. The bluegreen color of CoBr2 and 2,6-diacetylpyridine gradually changed to dark red. Next, a small amount of glacial acetic (1 μL) was added before the solution was left to stir at 50 °C for 12 h. After cooling to room temperature, the resulting solution was filtered using a frittedglass funnel inside an Ar-filled glovebox (Vacuum Atmospheres, 9 the rate became so fast that it could not be measured by stopped-flow spectrophotometry. The pKa of HCN is 9.3 at 25 °C, and thus the concentration of cyanide anion (CN−) increases under more basic conditions in the range of our experiments. The percent cyanide anion as a function of pH is plotted alongside the observed rates of cyanide binding to Co(II)N4[11.3.1] (Figure 3C). There is a reasonable correspondence between the observed reaction rate and the amount of cyanide anion present, strongly suggesting CN− to be the predominant nucleophilic attacking species under more alkaline conditions. Under the neutral conditions at which the cyanide concentration dependence was measured, the reaction mixtures were pseudo-first-order in HCN (not CN−). It follows from the linear relationship observed (Figure 3B) that HCN was the predominant attacking species around neutrality. This pH dependence of the reaction appears to be in keeping with the observations of previous authors23 regarding the reaction of cobalt corrins with cyanide, in which both HCN and CN− may participate as attacking nucleophiles, but the reaction is faster in the case of the anion. Solutions of other cobalt compounds with four nitrogen donors in a square-planar arrangement and two other axial sites for ligand binding are known to have water (or other solvent) molecules in their axial positions. Deprotonation of water ligands to become hydroxyl ligands can lead to changes in the rates observed for ligand substitution reactions. The question therefore arises as to whether the changing nature of any axial ligands may contribute to the pH dependence of the reaction between Co(II)N4[11.3.1] and cyanide (Figure 3C). Anaerobic solutions of Co(II)N4[11.3.1] (0.6 mM) were prepared in 0.1 M sodium phosphate (pH 4.5−8.5) and 0.1 M sodium tetraborate (pH 9−11) buffers, with NaCl added to 0.3 mM to maintain ionic strength. These two stock solutions were then mixed together in various combinations to produce samples ranging in pH from 4.85 to 10.62 and their electronic absorption spectra recorded immediately. All of these spectra were essentially identical (not shown), appearing just like that

Co(III) and Co(II) complexes. Under pseudo-first-order conditions (10−50-fold excesses of sodium ascorbate) the rate of Co(III)N4[11.3.1] reduction was followed by observing the accompanying increase in absorbance at 460 nm in the stopped-flow spectrophotometer. A single phase was observed, and a linear fit of the measured pseudo-first-order rate constants versus the sodium L-ascorbate concentrations was obtained to determine a second-order rate constant of 11.7(±0.4) M−1 s−1 at 25 °C and pH 7.4 (Figure 2A). To determine the activation parameters of the reaction, the observed rates were measured at six different temperatures between 10 and 37 °C. The enthalpy (ΔH⧧) and entropy (ΔS⧧) of activation were then calculated based on a linear fit of k the data, plotted as ln Tobs vs 1 , an Eyring plot (Figure 2B). The T enthalpy of activation (ΔH⧧), obtained from the slope of the line, was found to be 53.9(±0.8) kJ mol−1; while, the entropy of activation (ΔS⧧), determined from the y-intercept, was found to be −79 (±3) J mol−1 K−1. As it has previously been observed that exposing a solution of the Co(II)N4[11.3.1] to air overnight resulted in the complete conversion of the compound to its oxidized form, it seemed important to determine the rate of oxidation of the reduced compound if only to ensure that it was not faster than the reduction by ascorbate. The oxidation of Co(II)N4[11.3.1] was performed under pseudo-first-order conditions using excesses of oxygen in solution made by volumetric dilutions of oxygen-saturated buffer with degassed buffer. The secondorder rate constant, determined from the linear fit of the data, was found to be 0.5(±0.02) M−1 s−1 at 25 °C (Figure 2C). The reaction may be somewhat complicated as the y-intercept is non-zero, indicating the rate law to have the form v = k′ + k″[Co(II)][O2]. Nevertheless, the second-order rate constant for this oxidation is ∼20-fold slower than that for the ascorbate reduction, and, as physiological ascorbate levels are ∼60 μM21 while free oxygen levels in blood are no more than ∼130 μM,22 it is clearly reasonable that in the absence of cyanide (and/or perhaps other ligands) one should expect CoN4[11.3.1] to be present in the circulation predominantly as a Co(II)containing derivative. It follows that the kinetics of the reaction between Co(II)N4[11.3.1] and cyanide are likely to prove of key importance with regard to the mechanism of antidotal action. Kinetics of Cyanide Binding to Co(II)N4[11.3.1] under Anaerobic Conditions. Following the addition of excess cyanide to Co(II)N4[11.3.1] under anaerobic conditions, a new absorption spectrum was obtained with a sharp maximum at 420 nm and a very broad feature extending from ∼530 nm to beyond 900 nm in the near-infrared (Figure 1, dashed line). Our group has previously shown the species responsible for the new electronic absorption bands to be a dicyanide adduct of an oxidized form of the complex, formally Co(III)N4[11.3.1](CN−)2, by EPR measurements.5 The kinetics of this reaction were examined by following spectral changes at both 420 nm (indicative of the final dicyano-Co(III) product) and 460 nm (indicative of the reduced Co(II) parent) by stopped-flow spectrophotometry. At the standard temperature of 25 °C, the reaction of Co(II)N4[11.3.1] with cyanide under anaerobic conditions was too fast to be observed. In order to slow the reaction, most experiments were performed with the temperature lowered to 10 °C, and the cyanide concentrations were lowered (only 10−20-fold excesses) as far as possible while still maintaining pseudo-first-order conditions. The rate of reaction, D

DOI: 10.1021/acs.chemrestox.9b00170 Chem. Res. Toxicol. XXXX, XXX, XXX−XXX

Article

Chemical Research in Toxicology

Figure 4. Time-dependent photodiode array spectra of the reaction of Co(II)N4[11.3.1] with cyanide at a 1:1 ratio under aerobic versus anaerobic conditions. The reaction occurred between Co(II)N4[11.3.1] (0.6 mM) in 0.05 M sodium phosphate buffer, pH 7.4, and HCN (0.6 mM) in the presence and absence of oxygen at 10 °C. After mixing 1:1 (v/v, resulting in 0.3 mM reagent concentrations), the reaction mixtures were measured to be pH 7.6. (A) Anaerobic conditions (0 μM O2): electronic absorption spectra at t = 0 ms (solid trace), 27 ms (dashed trace), and 513 ms (dotted trace); enlarged spectra between 400 and 500 nm (inset). (B) Aerobic conditions (350 μM O2): electronic absorption spectra at t = 0 ms (solid trace), t = 27 ms (dashed trace), and t = 513 ms (dotted trace); enlarged spectra between 400 and 500 nm (inset).

−153(±1) J mol−1 K−1. The small ΔH⧧ value found indicates a low enthalpic barrier to the reaction, typically found in the range of 20−150 kJ mol−1 for bimolecular reactions, and is certainly consistent with the large rate constant value determined (∼105 M−1 s−1).25 Kinetics of Cyanide Binding to Co(II)N4[11.3.1] under Non-Pseudo-First-Order Conditions. In order to investigate the portion of the reaction represented by the non-zero intercept (Figure 3B), the reaction of Co(II)N4[11.3.1] with cyanide was performed under anaerobic conditions (10 °C, pH 7.6) using a 1:1 ratio of [HCN]:[Co(II)N4[11.3.1]] (both 0.3 mM). The absorption spectrum in the 400−500 nm region exhibited remarkably small changes: a just-detectable increase during the first 27 ms time frame followed by a slightly greater decrease of the 460 nm band by 513 ms (Figure 4A). Any shift in the position of this maximum was marginal or non-existent, and the final product remained stable for several minutes at least (not shown). With solutions made up aerobically, but otherwise under the same reaction conditions, a similar initial small increase in absorbance at 460 nm was observed at 27 ms, but now followed by a further increase during the next 513 ms and emergence of a distinct shoulder in the near-ultraviolet, presumably due to oxidation of the complex to some Co(III) derivative(s) (Figure 4B). Importantly, in neither of these reactions employing a 1:1 ratio of reagents was there any evidence for a peak at 420 nm, which, when present, indicates the formation of a one-electron-oxidized dicyano-Co(III) complex.5 In summary, there is clearly a reaction under these conditions, which because of the prevailing stoichiometry must involve the formation of a monocyano Co(II) complex that appears to be stable under anaerobic conditions. Furthermore, even if this reaction were to be lost in the stopped-flow deadtime when cyanide is in excess, the accompanying spectral changes are not large enough to account for the non-zero intercept observed (Figure 3B). Reactions of Co(II)N4[11.3.1] with cyanide were also carried out with a 2:1 ratio of reactants (0.6 mM HCN:0.3 mM Co(II)N4[11.3.1], 10 °C, pH 7.6) (Figure 5). Under anaerobic conditions, an increase in absorbance at 420 nm (Figure 5, solid line) was observed and, concomitantly, the absorbance at 460 nm (Figure 5, dash-dotted line) decreased. This behavior closely resembled the results observed under pseudo-first-order conditions (Figure 3A), indicating the formation of the dicyano-Co(III) product. Again, there are no obvious clues in the data to suggest an explanation for the

previously obtained at pH 7.4 (Figure 1, solid trace). The absorption spectra of samples prepared at pH < 5.0 did exhibit loss of intensity on standing for a few minutes, probably due to some degradation of the Schiff-base macrocycle structure.24 In the range from pH ∼5 to 10.6, however, the samples and their spectra were stable while anaerobic conditions were maintained. Since changes in axial ligation, including conversion of bound water to hydroxyl ion, are expected to result in measured spectral differences, we found no evidence for any pH-dependent variation in axial ligands. Consequently, these results seem to support the idea that the observed pH dependence of the reaction kinetics (Figure 3C) is indeed solely due to CN− being the faster attacking species compared to HCN in the reaction of cyanide with Co(II)N4[11.3.1], similar to the mechanism observed for the substitution of aquocobalamin by cyanide.23 On a slightly different note, it has been shown that the Co(II)N4[11.3.1] complex isolated from acetonitrile solution has distorted square-pyramidal geometry, with a methyl isocyanide (solvent) ligand in the axial position.20 In aqueous media, the complex may exist in predominantly octahedral geometry with two axial (solvent) ligands. Clearly then, all of the data suggest the rate law takes the following form: kobs = k′[X] + (kHCN + kCNKa/[H+])[HCN], where k′ is a cyanide-independent term given by the yintercept (∼420 s−1) and X is some intermediate. At pH 7.6, 98% of the cyanide solution is present as HCN (with little contribution from the reaction with CN−), and thus we evaluated kHCN as a second-order rate constant from the slopes of the fits to find, respectively, 8.5(±0.5) × 104 and 8.0(±0.5) × 104 M−1 s−1 (pH 7.6, 10 °C) for the data obtained at 420 and 460 nm (Figure 3B). At pH 9, where roughly half the total cyanide exists as an anion, we used the previously determined rate constant, kHCN, and the cyanide concentration to estimate the rate constant for the reaction with the cyanide anion, kCN, to be ∼2 × 106 M−1 s−1. The reaction of cyanide binding to Co(II)N4[11.3.1] was found to be temperature dependent over the range of 10−25 °C. Using the Eyring equation, the reciprocal of the k temperature was plotted versus ln Tobs (Figure 3D). The enthalpy of activation (ΔH⧧) was calculated from the slope of a linear fit to the data, with a value of 10.7(±0.4) kJ mol−1 being determined. The entropy of activation (ΔS⧧) was calculated from the y-intercept of the line and found to be E

DOI: 10.1021/acs.chemrestox.9b00170 Chem. Res. Toxicol. XXXX, XXX, XXX−XXX

Article

Chemical Research in Toxicology

with Co(II). The resultant spectrum (not shown) closely resembled that of the nominally axial-ligand-free Co(III)N4[11.3.1] (dotted line, Figure 1). When this reaction was attempted under otherwise similar conditions ([HCN] = 3.0 mM, [Co(II)N4[11.3.1]] = 0.15 mM) but in the stopped-flow spectrophotometer, even at 10 °C the reaction was too fast to measure, occurring within the dead time of the instrument. Consequently, no rate constant for the aerobic reaction could be determinedbut the reaction with oxygen under these conditions appears to be faster than any of the other (cyanidedependent) reactions we consider here.

■

DISCUSSION In general, Co(III) complexes are well-known to be kinetically inert to ligand substitutions.9−11 Consequently, the question arises as to how certain Co(III) corrinoids (e.g., cobalamin) can bind cyanide efficiently enough to be antidotally effective. We have recently argued, however, that one does not necessarily need to invoke phenomena such as favorable kinetic trans effects to explain facile cyanide substitutions in these systems under physiological circumstances. The net reducing conditions in vivo favor reduction to Co(II) forms; therefore, the available cobalt-based cyanide scavengers almost certainly work by binding cyanide principally to their more substitution-labile Co(II) forms, followed by oxidation to kinetically stable Co(III) cyanide adducts.5,17,18 This last oxidation step would appear to prevent any systemic redistribution of the toxicant by trapping the cyanide in a nontoxic form suitable for excretion. In the current experiments, we have examined the physicochemical properties of CoN4[11.3.1] to ensure that it fits this model and seek to explore why it may have improved decorporation characteristics compared to other cobalt scavengers. In Vivo Oxidation−Reduction Considerations. Ascorbate is deemed to be one of the main reductants in body fluids and tissues in humans, with the estimated ascorbate steadystate concentration in blood plasma of ∼60 μM.21 The reduction of Co(III)N4[11.3.1] by ascorbate, has been found here to be reasonably facile, with a rate constant of 11.7(±0.4) M−1 s−1 at pH 7.4 and 25 °C. This rate constant is comparable to those observed for various other cobalt tetraamine complexes in their reactions with ascorbate (3.4−42 M−1 s−1).26 The rate constant for the oxidation of Co(II)N4[11.3.1] by oxygen was found to be 0.5(±0.02) M−1 s−1, almost 20-fold slower than the rate constant observed for the ascorbate reduction of the complex. Estimates of the effective free oxygen concentration in mammalian blood are ∼130 μM maximum,22 but the level of oxygen might be expected to be a little higher under circumstances where the consumption of oxygen is inhibited (e.g., during cyanide intoxication)limited to 6 times faster than the reoxidation. To summarize, whatever the oxidation state of the cobalt at the time of administration, in the absence of cyanide the steady-state derivative(s) of CoN4[11.3.1] in circulating blood will contain Co(II)as previously confirmed by EPR spectroscopic measurements.5

Figure 5. Reaction of Co(II)N4[11.3.1] with cyanide at a 1:2 ratio under aerobic and anaerobic conditions. The reactions between Co(II)N4[11.3.1] (0.3 mM after mixing) and HCN (0.6 after mixing) in 0.05 M sodium phosphate buffer, pH 7.6, in the presence and absence of oxygen (350 μM and 0 μM, respectively) at 10 °C were followed at 420 and 460 nm over a 500 ms time frame. Absorbance changes at 420 nm: anaerobic (solid line); aerobic (dashed line). Absorbance changes at 460 nm: aerobic (dotted line), anaerobic (dash-dotted line).

non-zero intercepts observed in the presence of excess cyanide (Figure 3B). If any confirmation were required, these differences in the results obtained at 1:1 and 2:1 ratios of reactants clarify the stoichiometry of the final product obtained when there is at least 2 equiv of cyanide per cobalt in the reaction mixture. Interestingly, under aerobic conditions, the absorbance at 420 nm also initially increased (Figure 5, dashed line), but it did not reach the same level as in the reaction performed anaerobically (Figure 5, solid line). The rate at which the 420 nm absorbance (aerobic) decreased was approximately equal to the rate at which the 460 nm absorbance (aerobic) also decreased (Figure 5, dotted line). The final absorbances obtained at 420 and 460 nm under aerobic conditions were unlike those observed under anaerobic conditions, providing clear evidence for the formation of different products. Without over-interpreting the data at this point, the simplest explanation for these findings is that there is competition between the second cyanide anion and molecular oxygen for the available sixth coordination position of Co(II). Oxygen wins this competition under the experimental conditions we employ here and directs the outcome toward a final product that is something other than Co(III)N4[11.3.1](CN−)2. Kinetics of Cyanide Binding to Co(II)N4[11.3.1] in the Presence of Oxygen. As we are interested in the potential antidotal activity of Co(II)N4[11.3.1] under physiological conditions, the effect of oxygen on the reaction of cyanide with the reduced cobalt complex was further investigated. While most of the oxygen in mammals is bound to hemoglobin, the circulating levels of “free” oxygen may increase during cyanide poisoning as the ability of cytochrome c oxidase to utilize oxygen will be compromised. Cyanide binding to Co(II)N4[11.3.1] in the presence of oxygen was initially assessed by electronic absorption spectrophotometry. HCN solutions (6 mM) were prepared aerobically, [O2] ∼250 μM, while Co(II)N4[11.3.1] (0.3 mM) was prepared anaerobically in 0.1 M phosphate buffer, pH 7.4. Equal volumes of the two solutions were then mixed in a cuvette, leading to the disappearance of the absorption peak at 460 nm associated F

DOI: 10.1021/acs.chemrestox.9b00170 Chem. Res. Toxicol. XXXX, XXX, XXX−XXX

Article

Chemical Research in Toxicology

absorption spectrum of the final product was essentially identical to the spectrum of the cyanide-free Co(III)N4[11.3.1] (Figure 1, dotted line). The overall reaction under the aerobic conditions employed was so fast that we could neither identify the exact pathways nor determine the rate constant(s). Since the Co(III)N4[11.3.1](CN−)2 formed under anaerobic conditions is stable and the direct reaction of Co(II)N4[11.3.1] with oxygen is slow, the aerobic reaction of the complex with cyanide must involve reaction of oxygen with the monocyano intermediate at a faster rate than addition of a second cyanide (Scheme 2). Furthermore, because the net

Kinetics and Plausible Antidotal Mechanism of Cyanide Binding to Co(II)N4[11.3.1]. It follows from the argument above that the binding of cyanide to Co(II)N4[11.3.1] is the functionally significant issue. In this study, the overall reaction, leading to formation of a dicyano final product, was found to be rapid with a second-order rate constant kHCN = 8 × 104 M−1 s−1 (pH 7.6) when measured at 10 °C. At 37 °C, the rate constant can be estimated from the activation parameters or slope of the temperature dependence (Figure 3C) to be ∼3 × 105 M−1 s−1. The results for the cyanide dependence of the reaction (Figure 3B) suggest a rate law: kobs = k′[X] + (kHCN + kCNKa/[H+]) [HCN] where the reaction of the anion (CN−) is faster than the molecular acid (HCN) which follows from the observed accessible pH dependence (Figure 3C) and X is some intermediate (see below). In the simplest plausible sequence of reactions seemingly consistent with both the present findings and previously reported observations, omitting any participation of solvent-derived ligands for convenience, formation of the dicyano product can be thought of as a two-step process (Scheme 1). Assuming rigorous square-planar geometry to be

Scheme 2. Plausible Mechanistic Pathway for the Aerobic Reaction of Co(II)N4[11.3.1] with an Excess of Cyanide at Neutral pHa

Scheme 1. Plausible (Minimal) Mechanism for the Reaction of Co(II)N4[11.3.1] with Excess Cyanide under Anaerobic Conditions at Neutral to Mildly Alkaline pHa

a

Again, for simplicity, any water-derived ligands have been omitted.

a

For simplicity, any water-derived ligands have been omitted.

aerobic and anaerobic reactions do not occur at the same rate, the aerobic being faster, it follows that the second step of the anaerobic reaction is rate determining (Scheme 1). Lack of cyanide binding to the cobalt complex in the presence of excess oxygen would certainly suggest that Co(II)N4[11.3.1] might not be an effective antidote against cyanide toxicity, but we have clearly shown that it does work effectively in vivo.5 Perhaps these seemingly conflicting observations can be rationalized by recognizing that the product of the reaction with oxygen (Scheme 2) can be rereduced to the initial cyanide-binding form by any available reductants. Moreover, once formed, the dicyano complex is a kinetically stable cyanide sink, oxygen does not displace the bound cyano ligands. We have previously demonstrated that in the presence of excess ascorbate, oxygen is turned over (possibly to H2O2) in a catalytic fashion by CoN4[11.3.1].5 Furthermore, when excess cyanide was added to the aforementioned reaction mixture, the oxygen turnover was effectively halted. Thus, the available evidence supports the argument that the physiologic level of reductants (ascorbate and other reducing compounds) must be high enough to ensure that available oxygen is unable to oxidize the complex to the (cyanide-free) Co(III) form and, possibly, the normal prevailing level of free oxygen can be efficiently lowered by the turnover reaction with CoN4[11.3.1]. Conclusion. This study has gone some way toward enhancing our understanding of the reaction of Co(II)N4[11.3.1] with cyanide. The fast rate of reaction of the complex with cyanide (kHCN ≈ 105 M−1 s−1) implies a high efficacy of its cyanide-scavenging capability. The present results further support the suggestion stemming from our previous work,5,6 that Co(II)N4[11.3.1] could prove to be a better and

unlikely, we do not know whether the initial reactant, Co(II)N4[11.3.1], is a square-pyramidal or octahedral complex (with the additional ligand(s) derived from the solvent water). That is, the first step could be either an association or a ligand substitution that we write as an equilibration because the presence of a facile reverse reaction could account for the nonzero y-intercept observed in the kinetic data obtained with excess cyanide (Figure 3B).27 The final product, Co(III)N 4 [11.3.1](CN − ) 2 , 5 is formed by the association, or substitution, of a second cyanide anion occurring with concomitant oxidation of the central cobalt ion to Co(III) (Figure 3A). Consideration of the effect of oxygen (see below) suggests that the second of these steps, with associated rate constant we henceforth denote kHCN, is rate determining. The large negative activation entropy obtained indicates a much more ordered transition state compared to the initial state. An associative mechanism with two particles coalescing into one is consistent with this negative entropy of activation. The nature of the intermediate (X) requires some further comment. Following the addition of a single cyanide anion to Co(II)N4[11.3.1] both the EPR5 and electronic absorption (Figure 4A) spectral changes are surprisingly subtle. This is so irrespective of whether the reaction involved is the association of the strong-field ligand, cyanide, or the substitution of a weaker water-derived ligand by the cyanide. Consequently, whether the intermediate should be considered a monocyano complex with an actual coordinate bond, or should more correctly be described as an intimate ion pair, remains a question for future studies. Effect of Oxygen on Cyanide Binding. When cyanide was added to Co(II)N4[11.3.1] in the presence of oxygen, the G

DOI: 10.1021/acs.chemrestox.9b00170 Chem. Res. Toxicol. XXXX, XXX, XXX−XXX

Article

Chemical Research in Toxicology

of the cobalt corrinoids: equilibrium constants for the reaction of some ligands with aquahydroxocobinamide. Inorg. Chim. Acta 209, 161−169. (14) Marques, H. M., Bradley, J. C., Brown, K. L., and Brooks, H. (1993) The Ligand-substitution Reactions of Aquahydroxocobinamide proceed through a Dissociative Interchange Mechanism. J. Chem. Soc., Dalton Trans., 3475−3478. (15) Chemaly, S. M., Florczak, M., Dirr, H., and Marques, H. M. (2011) Probing the nature of the Co(III) ion in corrins: a comparison of the thermodynamics and kinetics of the ligand substitution reactions of aquacyanocobyrinic acid heptamethyl ester and stable yellow aquacyanocobyrinic acid heptamethyl ester. Inorg. Chem. 50, 8719−8727. (16) Zelder, F. (2015) Recent trends in the development of vitamin B12 derivatives for medicinal applications. Chem. Commun. (Cambridge, U. K.) 51, 14004−14017. (17) Benz, O. S., Yuan, Q., Cronican, A. A., Peterson, J., and Pearce, L. L. (2016) Effect of Ascorbate on the Cyanide-Scavenging Capability of Cobalt(III) meso-Tetra(4-N-methylpyridyl)porphine Pentaiodide: Deactivation by Reduction? Chem. Res. Toxicol. 29, 270− 278. (18) Yuan, Q., Pearce, L. L., and Peterson, J. (2017) Relative Propensities of Cytochrome c Oxidase and Cobalt Corrins for Reaction with Cyanide and Oxygen: Implications for Amelioration of Cyanide Toxicity. Chem. Res. Toxicol. 30, 2197−2208. (19) Long, K. M., and Busch, D. H. (1970) Cobalt(II) complexes of the quadridentate macrocycle 2,12-dimethyl-3,7,11,17tetraazabicyclo[11.3.1]heptadeca-1(17),2,11,13,15-pentaene. Inorg. Chem. 9, 505−512. (20) Lacy, D. C., McCrory, C. C., and Peters, J. C. (2014) Studies of cobalt-mediated electrocatalytic CO2 reduction using a redox-active ligand. Inorg. Chem. 53, 4980−4988. (21) Koshiishi, I., and Imanari, T. (1997) Measurement of ascorbate and dehydroascorbate contents in biological fluids. Anal. Chem. 69, 216−220. (22) Pittman, R. N. (2011) Chapter 4, Oxygen Transport, in Regulation of Tissue Oxygenation, Morgan & Claypool Life Sciences, San Rafael, CA. (23) Reenstra, W. W., and Jencks, W. P. (1979) Reactions of Cyanide with Cobalamins. J. Am. Chem. Soc. 101, 5780−5791. (24) Constable, E. C., Corr, S., and Lewis, J. (1986) The hydrolytic behaviour of some pentadentate schiff-base type macrocycles. Inorg. Chim. Acta 116, 95−97. (25) Espenson, J. H., and Yiu, D. T. (2000) Kinetics and mechanisms of reactions of methyldioxorhenium(V) in aqueous solutions: dimer formation and oxygen-atom abstraction reactions. Inorg. Chem. 39, 4113−4118. (26) Tsukahara, K., Izumitani, T., and Yamamoto, Y. (1982) The Reduction of Cobalt(III) Complexes by Ascorbic Acid. II. The Kinetics and Mechanisms of the Reactions of Diaqua and Aquahydroxo Macrocyclic N4 Complexes. Bull. Chem. Soc. Jpn. 55, 130−135. (27) Espenson, J. H. (1981) Chemical kinetics and reaction mechanisms, McGraw-Hill, New York.

more cost-effective cyanide antidote than the FDA-approved hydroxocobalamin.

■

AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected]. Phone: 412-624-3328. *E-mail: [email protected]. Phone: 412-624-3442. ORCID

Linda L. Pearce: 0000-0002-0940-965X Funding

This work was supported by the National Institutes of Health CounterACT Program, National Institutes of Health Office of the Director (NIH OD), and the National Institute of Neurological Disorders and Stroke (NINDS), Award R21 NS089893 (to L.L.P. and J.P.). Notes

The authors declare no competing financial interest.

■

ABBREVIATIONS CoN 4 [11.3.1]; cobalt 2,12-dimethyl-3,7,11,17-tetraazabicyclo[11.3.1]heptadeca-1(17)2,11,13,15-pentaenyl cation(s); Cbi; cobinamide; Cbl; cobalamin; CoTMPyP; [[4,4′,4″,4‴-porphyrin-5,10,20-tetrayltetrakis(1-methylpyridiniumato](2−)]cobalt pentaiodide; PDA; photodiode array; PMT; photomultiplier tube

■

REFERENCES

(1) U.S. FDA (2006) Drugs@FDA: FDA Approved Drug Products, U.S. Food and Drug Administration, Silver Spring, MD. (2) Borron, S. W., Baud, F. J., Megarbane, B., and Bismuth, C. (2007) Hydroxocobalamin for severe acute cyanide poisoning by ingestion or inhalation. Am. J. Emerg. Med. 25, 551−558. (3) Megarbane, B., Delahaye, A., Goldgran-Toledano, D., and Baud, F. J. (2003) Antidotal treatment of cyanide poisoning. J. Chin. Med. Assoc. 66, 193−203. (4) Borron, S. W., Stonerook, M., and Reid, F. (2006) Efficacy of hydroxocobalamin for the treatment of acute cyanide poisoning in adult beagle dogs. Clin. Toxicol. 44, 5−15. (5) Lopez-Manzano, E., Cronican, A. A., Frawley, K. L., Peterson, J., and Pearce, L. L. (2016) Cyanide Scavenging by a Cobalt Schiff-Base Macrocycle: A Cost-Effective Alternative to Corrinoids. Chem. Res. Toxicol. 29, 1011−1019. (6) Cronican, A. A., Frawley, K. L., Straw, E. P., Lopez-Manzano, E., Praekunatham, H., Peterson, J., and Pearce, L. L. (2018) A Comparison of the Cyanide-Scavenging Capabilities of Some Cobalt-Containing Complexes in Mice. Chem. Res. Toxicol. 31, 259−268. (7) Benz, O. S., Yuan, Q., Amoscato, A. A., Pearce, L. L., and Peterson, J. (2012) Metalloporphyrin Co(III)TMPyP Ameliorates Acute, Sublethal Cyanide Toxicity in Mice. Chem. Res. Toxicol. 25, 2678−2686. (8) Pratt, J. M. (1972) Inorganic Chemistry of Vitamin B12, Academic Press, London. (9) Greenwood, N. N., and Earnshaw, A. (1997) Chemistry of the Elements, 2nd ed., Butterworth-Heinemann, Oxford, UK. (10) Housecroft, C. E., and Sharpe, A. G. (2012) Inorganic Chemistry, 4th ed., Pearson Education Ltd., Harlow, UK. (11) Huheey, J. E., Keiter, E. A., and Keiter, R. L. (1993) Inorganic Chemistry: Principles of Structure and Reactivity, 4th ed., Harper Collins College Publishers, New York, NY. (12) Hayward, G. C., Hill, H. A., Pratt, J. M., Vanston, N. J., and Williams, R. J. (1965) The chemistry of vitamin B 12. IV. The thermodynamic trans-effect. J. Chem. Soc. 1, 6485−6493. (13) Marques, H. M., Bradley, J. C., Brown, K. L., and Brooks, H. (1993) Placing hydroxide in the thermodynamic trans influence order H

DOI: 10.1021/acs.chemrestox.9b00170 Chem. Res. Toxicol. XXXX, XXX, XXX−XXX