Zinc inhibition of cellular energy production: implications for mitochondria and neurodegeneration

Authors


Address correspondence and reprint requests to Ian J. Reynolds, Department of Pharmacology, University of Pittsburgh, W1351 Biomedical Science Tower, Pittsburgh PA 15261, USA. E-mail: iannmda@pitt.edu

Abstract

An increasing body of evidence suggests that high intracellular free zinc promotes neuronal death by inhibiting cellular energy production. A number of targets have been postulated, including complexes of the mitochondrial electron transport chain, components of the tricarboxylic acid cycle, and enzymes of glycolysis. Consequences of cellular zinc overload may include increased cellular reactive oxygen species (ROS) production, loss of mitochondrial membrane potential, and reduced cellular ATP levels. Additionally, zinc toxicity might involve zinc uptake by mitochondria and zinc induction of mitochondrial permeability transition. The present review discusses these processes with special emphasis on their potential involvement in brain injury.

Abbreviations used
AIF

apoptosis-inducing factor

AM

acetoxymethyl

ETC

electron transport chain

GAPDH

glyceraldehyde-3-phosphate dehydrogenase

GPDH

glycerol 3-phosphate dehydrogenase

KGDHC

α-ketoglutarate dehydrogenase complex

Km

equilibrium affinity constant

LADH

lipoamide dehydrogenase

Δψm

mitochondrial membrane potential

MT

metallothionein

MPT

mitochondrial permeability transition

TCA

tricarboxylic acid cycle

[Zn2+]i

intracellular free Zn2+ concentration

Interest in Zn2+-mediated brain injury is motivated by evidence implicating Zn2+ as a neurotoxin in models of stroke, epilepsy, mechanical trauma, and Alzheimer's disease (Choi and Koh 1998). The precise mechanism of cytotoxicity is unknown, but emerging evidence suggests that Zn2+ kills neurons through the inhibition of ATP synthesis (Weiss et al. 2000). The notion that Zn2+ affects energy production is admittedly not a new one; indeed, investigators revealed that Zn2+ impedes mitochondrial function very soon after mitochondria themselves could be properly studied (Hunter and Ford 1955). There has, however, been a resurgence of interest in this general area due to key advances over the last decade. First, a number of seminal reports have allowed a far better understanding of how zinc is regulated under physiological conditions and how disturbances in zinc homeostasis can lead to neuronal injury. Second, neuroscientists have greatly clarified how energy-producing systems such as mitochondria participate in the events leading to neuronal death. Consequently, investigators are now well positioned to explore the impact of zinc on cellular energy production, and how these events may be related to neurodegeneration. This topic is the principal concern of the present review. Initially, we assess data suggesting that zinc interferes with glycolysis, the tricarboxylic acid cycle (TCA), and the mitochondrial electron transport chain. Several related issues also bear consideration, such as mitochondrial generation of reactive oxygen species (ROS) and possible induction of mitochondrial permeability transition (MPT). We then discuss mitochondrial transport of Zn2+ and possible regulation of metabolism by the metallothionein family of zinc binding proteins. Finally, we examine approaches used in the determination of intracellular free zinc concentrations, which is a critical component of any hypothesis concerning the mechanism of zinc toxicity. More expansive treatments of the neurobiology of Zn2+ are cited herein.

Zinc in the brain

Zinc is abundant in the brain with levels at ∼200 ng/mg protein. As suggested by Frederickson (1989), it is useful to consider three distinct pools of cellular zinc in the CNS. The largest fraction, 80% or more, exists in tight coordination with intracellular proteins and is generally considered immobile. A second pool of zinc is sequestered in the vesicles of certain neurons and is readily detected with metal-sensitive histochemical stains. Vesicular zinc can account for up to 10% of cellular totals, is always colocalized with glutamate, and, like glutamate, can be released into the synapse with neuromodulatory effects. The third pool of zinc is that which is free in the cytoplasm as unbound, ionic form. Although the concentration of intracellular free zinc ([Zn2+]i) in the cytoplasm of resting cells is thought to be very low or even nonexistent (Outten and O'Halloran 2001), [Zn2+]i may be elevated during injurious stimuli (Choi and Koh 1998). It is clear that zinc toxicity requires elevated [Zn2+]i, as strategies that prevent Zn2+ entry (Weiss et al. 1993; Sensi et al. 1997; Kim et al. 2000), augment Zn2+ efflux (Palmiter and Findley 1995; Tsuda et al. 1997; Kim et al. 2000), or increase intracellular buffering capacity (Palmiter 1995; Kelly et al. 1996) all ameliorate Zn2+-mediated cell death. For reasons discussed below, obtaining precise measurements of [Zn2+]i are problematic. Nonetheless, one study suggests a peak intracellular LC50 of 300 nm in cultured cortical neurons (Canzoniero et al. 1999). When reviewing the existing body of literature, it is important to consider what [Zn2+]i is actually experienced by neurons during even the most severe in vivo insults, as a number of studies used amounts of zinc that would now be considered supraphysiologic in the context of brain injury.

What is the source of neurotoxic Zn2+?

While it is accepted that Zn2+ accumulates in and contributes to the death of neurons in various injury models, the source of this toxic Zn2+ is not entirely clear. A number of metal transporters probably regulate Zn2+ flux across the plasma membrane under physiological conditions (McMahon and Cousins 1998), but [Zn2+]i accumulation during neuronal injury occurs chiefly through voltage-sensitive calcium channels, glutamate receptors, and sodium–calcium exchange (Sensi et al. 1997; Cheng and Reynolds 1998). Initially, it was thought that in vivo[Zn2+]i-mediated injury resulted from excessive release of vesicular Zn2+ into the synapse, which then translocated into postsynaptic neurons with toxic consequences (Koh et al. 1996). More recent evidence suggests that while Zn2+ apparently translocates into neurons from the extracellular space, it does not come from the histochemically reactive, synaptically released vesicular pool (Lee et al. 2000). Lethal [Zn2+]i may also be reached through the oxidation of intracellular zinc binding proteins (Aizenman et al. 2000), and several reports indicate that [Zn2+]i is a critical mediator of oxidative damage in whole animal models (Cuajungco and Lees 1998). While the precise source remains ambiguous, it is clear that [Zn2+]i-mediated neuronal injury requires the mobilization and redistribution of Zn2+ already present in the brain- hence its classification as an endogenous neurotoxin.

Zn2+ inhibition of glycolysis

Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and phosphofructokinase are two glycolytic enzymes that might be impaired when cells experience elevated [Zn2+]i. Zn2+ shows IC50s of 400 nm and 1.5 µm for purified GAPDH and phosphofructokinase, respectively (Ikeda et al. 1980; Krotkiewska and Bana 1992). Zn2+ inhibition of GAPDH could be especially deleterious to the cell, given its critical role in the regulation of glycolysis. In cultured cortical neurons, Sheline et al. (2000) showed that elevated [Zn2+]i resulted in (i) buildup of glycolytic metabolites upstream of GAPDH (i.e. dihydroxyacetone phosphate and fructose-1,6-bisphosphate); (ii) depletion of downstream intermediates (1,3-biphosphoglycerate and pyruvate); (iii) reduction of cellular ATP levels; and (iv) neuronal death that was ameliorated by the addition of downstream substrates such as pyruvate. All of these observations supported the authors' initial hypothesis that cellular Zn2+ overload results in the inhibition of GAPDH. Several inconsistencies were observed however, chief among these was an unexpected decrease in NAD+ levels. GAPDH catalyzes the conversion of glyceraldehyde-3-phosphate to 1,3-bisphosphoglycerate, reducing NAD+ to NADH in the process, so one would not predict a fall in NAD+ if Zn2+ were inhibiting only GAPDH. This lead the authors to speculate that, in addition to inhibiting GAPDH, high [Zn2+]i activated an NAD+ catabolizing enzyme that could not be identified. While the precise mechanism(s) remains unclear, these provocative results leave little doubt that Zn2+ impedes glycolysis. In any case, the importance of these findings relies on the degree to which cultured neurons recapitulate in vivo energy production. It may be that only cultured neurons rely heavily on glycolysis while brain energy is derived primarily from mitochondrial oxidative phosphorylation.

Zn2+ inhibition of tricarboxylic acid cycle

Brown et al. (2000) noted that rat liver mitochondria respiring on complex-I substrates (glutamate and malate) appeared more sensitive to Zn2+ inhibition than those respiring on the complex-II substrate succinate. Consequently, the authors suspected that a more sensitive inhibition must occur at or upstream of complex-I. In agreement with their hypothesis, they found that Zn2+ inhibited the α-ketoglutarate dehydrogenase complex (KGDHC) of the TCA cycle. An apparent Ki of ∼1 µm was calculated, but the authors reasoned that this estimation represented an upper limit because (i) Zn2+ inhibition of KGDHC increased with time, but the apparent Ki was calculated from just the initial reaction period, and (ii) the presence of contaminating metals may have resulted in a spurious rightward shift in their concentration response curve. In support of the latter argument, addition of metal chelator gave as much as a twofold increase in enzyme activity. A true Ki was therefore estimated at ∼100 nm. Using KGDHC isolated from porcine heart, the site of inhibition was subsequently identified as a catalytic disulfide of the lipoamide dehydrogenase (LADH) subunit (Gazaryan et al. 2002). Moreover, LADH inhibition by Zn2+ was associated with augmented ROS production (see below).

Zn2+ inhibition of the electron transport chain

Using submitochondrial particles isolated from rabbit heart, Skulachev (Skulachev et al. 1967) first described Zn2+ inhibition of the electron transport chain, possibly at multiple sites, the most sensitive of which was thought to be somewhere between cytochromes b and c1. These initial findings were generally supported by subsequent studies using bovine heart mitochondria (Nicholls and Malviya 1968; Kleiner and von Jagow 1972; Kleiner 1974). A more precise localization of Zn2+ interaction remained a matter of speculation for a number of years until two reports independently concluded that Zn2+ reversibly inhibited the bc1 complex purified from bovine heart (Lorusso et al. 1991; Link and von Jagow 1995). Moreover, both groups established a Km for Zn2+-bc1 near the IC50 for Zn2+ inhibition of electron transfer. Specifically, Link and von Jagow (1995) concluded that (i) zinc inhibited the Q cycle, possibly near the Qp center, (ii) Zn2+ binding was inhibited by low pH, and (iii) Zn2+ inhibition of bc1 activity and 65Zn binding to bc1 both occurred in the range of 100–200 nm (i.e. well within estimations of pathophysiolgical [Zn2+]i). They further speculated that the putative Zn2+ binding site is a component of the proton channel at the hydroquinone center of bc1. Most recently, crystallographic techniques demonstrated that Zn2+ binds avine bc1 at four sites proximal to the stigmatellin binding locale (Berry et al. 2000). This was predicted to inhibit bc1-quinone binding as well as the flow of protons from the Q0 site to the intermembrane space.

Recent work in our laboratory suggests that Zn2+ similarly inhibits isolated rat brain mitochondria (Dineley et al. 2001). We found that Zn2+ (∼200 nm) inhibited O2 consumption and reduced Δψm in brain mitochondria using substrates for complexes I, II and glycerol-3-phosphate dehydrogenase. Such an extensive shutdown of substrate usage is most parsimoniously explained by inhibition at a single site common to all three substrate pathways, i.e. at or downstream of ubiquinone. Because we found no evidence for inhibition of complex IV, we concluded that Zn2+ inhibition of the neural ETC occurs somewhere in the vicinity of ubiquinone-cytochrome c. Finally, it should be noted that although the majority of literature supports Zn2+-bc1 interaction, Zn2+ has been proposed to inhibit electron transfer at other sites in the ETC (Skulachev et al. 1967; Nicholls and Malviya 1968).

Inhibition of the ETC should result in dissipation of Δψm in intact cells. Several studies have reported mitochondrial depolarization in neurons subjected to toxic elevations of [Zn2+]i. Sensi et al. (1999) showed that toxic [Zn2+]i accumulation through AMPA/KA receptors resulted in mitochondrial depolarization in murine cortical neurons. Using the zinc ionophore pyrithione to increase [Zn2+]i in rat cortical neurons, we demonstrated that neurotoxic [Zn2+]i can be achieved with no change in Δψm (Dineley et al. 2000). Why these studies generated contrasting results is unclear. It could be that Zn2+ entering through AMPA/KA receptors preferentially targets mitochondria, while ionophore-induced elevations of [Zn2+]i are more effectively handled by intracellular buffers. In any case, more studies are necessary to clarify how [Zn2+]i affects mitochondria in intact cells.

Apart from the ETC per se, Zn2+ inhibits cytosolic glycerol-3-phosphate dehydrogenase (GPDH) with an IC of ∼ 1 µm at pH 7.4 (Maret et al. 2001). Cells also contain a flavin-linked GPDH isoform that is bound to the outer face of the inner mitochondrial membrane (Nelson and Cox 2000). Mitochondrial GPDH (mGPDH) catalyzes the NADH-linked reduction of ubiquinone, and serves as the major route for shuttling reducing equivalents from the cytosol into neural and skeletal mitochondria. We are unaware of any studies supporting or refuting inhibition of mGPDH by Zn2+, and our own data using glycerol-3-phosphate as a mitochondrial substrate does not allow discrimination between direct inhibition of mGPDH by Zn2+, or if Zn2+ is acting at bc1 as already described. Were Zn2+ inhibition of mGPDH to occur however, then cytosolic NADH produced by glycolysis could not deliver electrons to the electron transport chain.

Zn2+ and mitochondrial permeability transition

Perturbed mitochondria may undergo permeability transition (MPT), which is the opening of a relatively large, non-selective pore. The pore allows the passage of species of up to ∼1.5 kDa, and its opening may be a critical step in initiating both apoptotic and necrotic cell death. Though the pore remains incompletely characterized, there is some understanding of the factors regulating its activity. Mg2+, adenine nucleotides, low pH, and the immunosuppressant cyclosporin A all inhibit the pore, while loss of Δψm, high matrix calcium, and oxidative stress all contribute to MPT. Consequences of MPT include mitochondrial swelling, efflux of mitochondrial calcium stores, as well as the release of numerous molecules from mitochondria, including glutathione, cytochrome c and apoptosis-inducing factor (AIF) (Zoratti and Szabo 1995; Zamzami and Kroemer 2001). A number of reports investigated Zn2+ involvement in MPT with varying results.

Using rat liver mitochondria respiring in sucrose buffer, Hunter and Ford first observed Zn2+-induced (30 µm) swelling that was inhibited by Mg2+ (Hunter and Ford 1955). An extension of these studies showed this swelling did not occur in anaerobic conditions (Hunter et al. 1956). Initial reports of insulin-induced swelling in liver mitochondria were later attributed to Zn2+ contamination (Cash et al. 1968). In a series of studies using bovine heart mitochondria, Brierley and colleagues showed that Zn2+ caused swelling and increased permeability to Mg2+, K+, Na+, and Cl (Brierley 1967; Brierley and Knight 1967; Brierley et al. 1968). Interestingly, Zn2+ did not increase sucrose permeability in these experiments, suggesting that while Zn2+ altered membrane properties, MPT did not occur. In another study, Zn2+ (2–10 µm) was shown to induce swelling and loss of GSH in liver mitochondria co-treated with agents that inhibit ATP synthesis (e.g. valinomycin, FCCP, Ca2+ and oligomycin) (Wudarczyk et al. 1999). Consistent with MPT, both swelling and loss of GSH were inhibited by Mg2+ or CsA. However, the effect of [Zn2+]i in the absence of these other agents was not determined. Moreover, others reported no evidence of Zn2+-induced swelling or evidence of MPT in liver mitochondria treated with Zn2+ (Brown et al. 2000).

Using isolated brain mitochondria, Jiang et al. (2001) recently claimed evidence for Zn2+-induced MPT. At concentrations as low as 10 nm, Zn2+ caused swelling as well as the release of cytochrome c and apoptosis inducing factor. Considering the low concentrations used, these results certainly attest to the potency of Zn2+. A metal buffering system was not used in this study however, so it is not clear how such low [Zn2+] was controlled. Given the reputation of Zn2+ as a ubiquitous and problematic contaminate of everything from distilled water to allegedly pure reagents, one would expect an unbuffered solution to contain nanomolar Zn2+ even before any deliberate additions (Frederickson 1989). Our recent work (Dineley et al. 2001) used a KCl-based recording solution buffered with EGTA, and free Zn2+ was calculated considering only chelation by EGTA. This would almost certainly lead us to overestimate the actual [Zn2+], as many other species in the recording solution are also effective metal chelators. Inhibition of O2 consumption and loss of Δψm was observed at [Zn2+]≈ 200 nm, but both effects were rapidly and almost completely restored by addition of excess EGTA. If Zn2+ induced MPT, then one would not expect restoration of these parameters with simple metal chelation, particularly if critical elements such as cytochrome c were lost from the mitochondria. Moreover, addition of cyclosporin A gave no protection. Taken together, our results are inconsistent with the substantial mitochondrial changes expected to accompany permeability transition in brain mitochondria.

It is presently not clear why some reports generated evidence for Zn2+-induced MPT while others did not. Assay conditions are almost certainly a critical factor: our informal observations suggest that mitochondria in sucrose-based buffers behave substantially different from those incubated in KCl-based solutions (T. Votyakova and I. Reynolds, unpublished observations). There is also an emerging appreciation of tissue-specific differences between mitochondria isolated from different tissues. Berman et al. (2000) reported that rat liver mitochondria swelled much more robustly and released glutathione in response to agents that induce MPT, while similarly treated rat brain mitochondria exhibited little swelling and did not release glutathione. Moreover, mitochondria isolated from different brain regions may have considerable differences in their sensitivities to Ca2+-induced MPT (Friberg et al. 1999). Clearly, more studies are required to resolve these contradictions in the context of brain injury.

Zn2+-induced generation of ROS

There is evidence for ROS accumulation and lipid peroxidation in neurons subjected to toxic [Zn2+]i, and treatment with antioxidants ameliorates [Zn2+]i-induced neuronal injury in culture (Kim et al. 1999; Noh et al. 1999; but see Sheline et al. 2000). Given that mitochondria are believed to be the primary source of oxidative stress in neurons, and considering the large body of evidence implicating Zn2+ in mitochondrial dysfunction, it is reasonable to speculate that Zn2+ augments mitochondrial ROS production. In support of such a mechanism, elevated [Zn2+]i via AMPA/kainate receptor activation resulted in mitochondrial free radical production in murine cortical cultures (Sensi et al. 1999). Zn2+-inhibition of the TCA cycle may also instigate ROS production: the LADH subunit of KGDHC catalyzes the oxidation of NADH by O2, directly generating superoxide and H2O2. Using enzymes purified from isolated porcine heart mitochondria, Gazaryan et al. (2002) recently showed that Zn2+ accelerates this reaction fivefold. However, other sources besides or in addition to mitochondria might account for ROS accumulation in [Zn2+]i toxicity. For example, the superoxide generating enzyme NADPH oxidase is activated by high [Zn2+]i in cultured neurons and astrocytes, resulting in elevated cellular ROS and cytotoxicity (Noh and Koh 2000).

To directly determine if isolated brain mitochondria accumulate ROS upon exposure to Zn2+, we used the H2O2-sensitive Amplex red assay in spectrofluorometric recordings. We found that 200 nm Zn2+ caused approximately a twofold increase in ROS accumulation in mitochondria respiring on the complex-I substrates glutamate and malate (Dineley et al. 2001). We have not identified the exact mechanism responsible for Zn2+-induced ROS production. However, in addition to its ability to bind Zn2+ with high affinity, bc1 is considered the major source of mitochondrial superoxide. Therefore, it is plausible that Zn2+-inhibition of brain mitochondria at bc1 results in increased accumulation of ROS. Interestingly, while inhibition of O2 consumption and dissipation of Δψm were reversed by chelation of Zn2+ with EGTA, the same maneuver did not reduce ROS accumulation. Based on these data alone, there may be differences in the way Zn2+ affects ROS compared with how it dissipates Δψm and diminishes O2 consumption. The nature of these differences is presently unknown.

Of course, any direct impact of Zn2+ on TCA enzymes would require elevation of matrix [Zn2+], either by mobilization of Zn2+ already present in the mitochondria, or via entry of Zn2+ from the cytosol. As to the former, we can only speculate that oxidative stress could liberate Zn2+ bound to matrix proteins. With respect to the latter, the uptake of appreciable amounts of Zn2+ by mitochondria, at least on an acute time scale (< 1 h), seems unlikely for reasons discussed below.

Do mitochondria transport Zn2+?

Delineating site(s) of inhibition inevitably raises the question of possible Zn2+ uptake by mitochondria. Given that mitochondria are semiautonomous organelles capable of transcription and translation and that there is a critical need for Zn2+ in such processes, it is plausible that the mitochondrial membrane bilayer possesses transporters responsible for tightly regulating matrix zinc. A number of Zn2+specific transporters have been localized to the plasma membrane, Golgi body and intracellular vesicles, but there is no evidence that any of these are explicitly associated with mitochondria (McMahon and Cousins 1998; Kambe et al. 2002). As previously mentioned, numerous plasma membrane pathways permeable to Ca2+ can also permit Zn2+ flux. Mitochondrial calcium pathways might behave similarly, and indeed two studies are often cited as evidence of uniporter mediated Zn2+ uptake (Brierley and Knight 1967; Saris and Niva 1994). Brierley and Knight (1967) used atomic absorption to measure Zn2+ uptake by bovine heart mitochondria. They found that mitochondria accumulated Zn2+ from a sucrose buffer supplemented with high Zn2+ (20 µm or more) and devoid of other cations, and Zn2+ uptake was inhibited by the addition of Mg2+. Insofar as Zn2+ may be forced into the matrix under extreme conditions, it is otherwise difficult to draw physiologically useful conclusions from such experiments. Others observed ruthenium red inhibition of Zn2+-induced swelling in rat liver mitochondria energized by succinate (Saris and Niva 1994), suggesting that the effects of Zn2+ on mitochondria require uniporter-driven Zn2+ uptake. However, it is important to note that even high concentrations of added Zn2+ (e.g. 40 µm) caused only a modest degree of swelling. Thus, direct evidence of Zn2+ uptake into mitochondria is very limited.

We recently showed that Zn2+-induced loss of Δψm and inhibition of O2 utilization were rapidly reversed by a membrane impermeant chelator, suggesting that the effects of Zn2+ did not require influx (Dineley et al. 2001). We found that ruthenium red partially protects isolated brain mitochondria from depolarization induced by ∼200 nm Zn2+. Protection was lost at higher concentrations of zinc (> 1 µm), indicating either some non-specific protection unrelated to uniporter activity, or that inhibition of electron transport occurs at several sites, some of which do not require Zn2+ import. In any case, loss of ruthenium red protection is unlikely to be a consequence of simple displacement by Zn2+, because ruthenium red is not thought to act in a competitive fashion. In other experiments we used Zn2+-sensitive fluorophores in an unsuccessful attempt to generate direct evidence for Zn2+ uptake by isolated brain mitochondria. As all of our experiments were performed on a time scale of minutes, it is possible that we did not observe Zn2+ transport that may occur over longer periods. Nevertheless, we could find no evidence supporting mitochondrial import of Zn2+ through the uniporter, or indeed by any other mechanism. Furthermore, published data in support of such a mechanism(s) are scant and not wholly convincing.

Metallothioneins and energy

Because the vast majority of intracellular zinc is protein bound, it is conceivable that zinc–protein complexes might serve as important regulators of cellular metabolism. In this regard, metallothionein (MT) proteins may present ideal candidates: they reversibly bind up to seven Zn2+ ions, their small size (6–7 kDa) suggests a relatively high degree of intracellular mobility, and they are expressed in all mammalian tissues. The eukaryotic family includes four isoforms designated MT I–IV. MTs I and II are present in all cell types, MT III is expressed largely in neurons, and MT IV is found in epithelial tissue (reviewed by Palmiter 1998). A relatively long history of investigation has yielded few clues regarding the precise function(s) of this family of metalloproteins. However, MT participation in energy metabolism was initially postulated by Beattie et al. (1998), who claimed that knockout mice lacking MT I and II develop obesity. Subsequently, Jiang et al. (1998) reported that MT binds ATP in 1 : 1 stoichiometry, and suggested a direct interaction between the two molecules that facilitated the transfer of Zn2+ from MT. Other findings demonstrated that MT may be imported into isolated liver mitochondria, and that MT-derived Zn2+ inhibits mitochondrial O2 utilization (Ye et al. 2001). These reports are remarkable in that they support a direct link between zinc homeostasis and cellular energy status, as well as suggesting a profound physiological function for these otherwise poorly characterized proteins. However, as pointed out by Palmiter (1998), it must be noted that Beattie and colleagues did not use control mice of the appropriate genetic background, and that MT I/II knockouts developed in other labs do not exhibit obesity. Moreover, Zangger et al. (2000) have challenged direct ATP–MT interaction: using five different methods (NMR spectroscopy, titration calorimetry, gel-filtration chromatography, affinity chromatography, and ultrafiltration) they found no evidence to support the findings of Jiang et al. (1998). Finally, it is difficult to imagine a critical role for MTs in energy production given that MT I/II null mice grow normally (Palmiter 1998). In summary, ascribing to MTs any specific bioenergetic role beyond their ability to control [Zn2+]i seems premature.

Measuring [Zn2+]i

When experimenting with isolated enzymes or organelles, it is important to consider what concentration of Zn2+ accurately reflects [Zn2+]i achieved during in vivo brain injury. Thus, the design of pathophysiologically relevant experiments depends heavily on the ability to measure [Zn2+]i with confidence. Our work with isolated brain mitochondria used an EGTA buffer to control free zinc, and we observed substantial alterations of mitochondrial function at [Zn2+]≈ 100–200 nm. A recent study using Zn2+-sensitive fluorophores in cultured neurons demonstrated that [Zn2+]i may rise as high as several hundred nanomolar during injurious stimuli (Canzoniero et al. 1999). According to this estimation, our conditions ostensibly reflect a pathophysiological setting. As previously mentioned however, estimation of [Zn2+]i in live cells is difficult. A variety of techniques may be employed, though each has some critical limitation. Fractionation approaches for example disrupt cellular compartments and inevitably mix different zinc pools. Radioisotope techniques with 65Zn2+ do not allow the distinction between free zinc and bound zinc, and electrophysiology is usually not feasible due to the small currents carried by Zn2+ when compared with Ca2+, for example. The use of ion-sensitive fluorophores is probably the best and simplest method for measuring [Zn2+]i. Although it is easy enough to use Zn2+-sensitive dyes as an indication of [Zn2+]i changes, converting fluorescence values into precise [Zn2+]i is far more complicated (Dineley et al. 2002). Most fluorophore-based ion measurements load cells with dye via passive diffusion of acetoxymethyl (AM) ester derivatives. Once inside the cell the esters are cleaved, resulting in a trapped membrane impermeant species. This approach however, often yields intracellular dye concentrations in excess of 10−3 m (Tsien 1999). Because neurobiologically relevant [Zn2+]i in all probability does not exceed 10−6 m, intracellular dye vastly exceeds available Zn2+. Unfortunately, the classic equation used to derive [ion]i from fluorescence is inapplicable because it assumes that the concentration of free ion is not reduced through binding with intracellular dye (Grynkiewicz et al. 1985). In theory, one could use binding isotherms that do consider ion depletion (Kenakin 1993); however, such an accommodation requires that intracellular dye concentration be known with relative precision, which in itself is a difficult parameter to establish. The upshot of this is that although Zn2+ sensitive fluorophores are useful tools for examining relative changes in [Zn2+]i, extrapolations to exact values of [Zn2+]i must be viewed carefully (Sensi et al. 1997; Cheng and Reynolds 1998; Canzoniero et al. 1999; Sensi et al. 1999).

Conclusions

The zinc biologist faces numerous challenges at this point. It is established that [Zn2+]i interacts with targets essential to glycolysis, the TCA cycle, and mitochondrial electron transport. These interactions and their putative ramifications are summarized in Fig. 1. However, which if any of these mechanisms are critical for the execution of neurons remains unresolved. An attempt to clarify these issues might feature the following mission statements: (i) to estimate with greater accuracy the amount of Zn2+ that accumulates in neurons during injury models so that pathophysiologically relevant experiments are designed; (ii) to use neural models so that tissue specific-differences in phenomena related to energy production (e.g. mitochondrial swelling and MPT) are better understood in the context of brain injury; (iii) to determine if and how mitochondrial Zn2+ is increased during [Zn2+]i toxicity so that the comparative vulnerability of matrix targets vs. cytoplasmic targets can be assessed; (iv) to increase understanding of the mechanisms that serve to protect cells from elevated [Zn2+]i and how they might be manipulated to increase neuronal resistance to toxic [Zn2+]i; (v) perhaps most importantly, to progress toward advanced models with the purpose of establishing how [Zn2+]i-induced neurotoxicity is relevant in the clinical setting.

Figure 1.

Potential bioenergetic targets for Zn2+ inhibition. Zn2+ may compromise cellular energy production through the inhibition of glycolysis, the tricarboxylic acid cycle (TCA) and/or the mitochondrial electron transport chain. Possible consequences of elevated [Zn2+]i include reduced O2 consumption, reduced Δψm, reduced ATP levels, increased generation of ROS, and neuronal death. The impact of Zn2+ on mitochondrial permeability transition, as well as the mechanisms responsible for mitochondrial Zn2+ uptake remain unclear.

Acknowledgements

Studies in the authors' laboratory were supported by NIH grant NS34138 (IJR) and a predoctoral fellowship from the American Heart Association (KED).

Ancillary