Oxygen toxicity and the health and survival of eukaryote cells: A new piece is added to the puzzle

My first thought was “Superoxide dismutase: of what use is that to a cell?” It was 1976 and I was a graduate student in search of a topic for a “literature” seminar. Fortunately I chose to present the story of this exotic-sounding protein, “SOD,” to my colleagues and was rewarded with the beginnings of a long-term fascination with the concept that our aerobic (O₂-based) lifestyle is fraught with oxidative (O₂-based) hazards. This concept is now firmly linked to topics of enduring interest and importance such as aging, cancer, ischemia (anoxic tissue damage), aerotolerance, and immune system function. In this issue of PNAS, Luk et al. (1), using brewer's yeast cells, show how a key eukaryote antioxidant enzyme, mitochondrial Mn-cofactored SOD (SOD2), acquires its Mn cofactor and possibly its active conformation via a previously uncharacterized nuclear-encoded protein, thus providing both a tool for experimental oxy-radical stress manipulations and a more detailed understanding of how eukaryote cell oxygen defenses are assembled.

By the 1960s it was clear that reactive oxygen species (ROS), namely, partially reduced O₂ derivatives such as hydrogen peroxide (H₂O₂), the hydroxyl radical (^*OH), and the superoxide anion radical (O₂^-) formed by ionizing radiation could oxidize and damage cell proteins, lipids, and nucleic acids. However, there was little convincing evidence that either O₂-mediated damage or ROS were produced by normal respiratory (dioxygen-reducing) cellular processes. This changed in 1969 when McCord and Fridovich (2) demonstrated that the catalytic activity of a long-known red blood cell protein of unknown function, erythrocuprein, was to specifically dismute O₂^- to H₂O₂ and O₂. If O₂^- were not a real and present danger, why would red cells contain an enzyme specifically degrading it? Because this seminal report also presented a simple and sensitive assay for SOD, numerous surveys looked for SOD in everything from whales to tube worms to Archaea. The conclusion: Almost all aerobic cells and many anaerobes contain one or more highly conserved SOD protein. It is now widely accepted that the necessary flood of water produced in a normal respiring prokaryote cell or mitochondrion by the protein-bound tetravalent reduction of O₂ is often accompanied by a noxious trickle of partially reduced O₂ (free ROS). SOD enzymes, along with peroxide-destroying catalases and peroxidases, are the cell's principal ways of minimizing that trickle.

Almost all SOD proteins fall into four classes: Cu-Zn SODs (SOD1s) are typical of eukaryote cytosols, Fe-SODs of many prokaryotes, and MnSODs (SOD2s) of many prokaryotes, chloroplasts, and the mitochondrial matrix. In mammals, extracellular (EC) SODs (SOD3s) (3) are also present. The SOD2s found in the chloroplasts and mitochondria of most eukaryote cells have high sequence homologies to the MnSODs found in prokaryotic bacteria but almost no similarity to the Cu-Zn-cofactored SOD1s found in their own cytosols. All catalyze the dismutation of two superoxide anions to O₂ + H₂O₂ at superoxide diffusion-limited rates, making them the fastest enzymes known.

Superoxide dismutase acquires its Mn cofactor by a previously uncharacterized protein.

The inducible homotetramer SOD2s found in the mitochondria of organisms from yeasts to humans are particularly important antioxidant defenses, because they rest close against the inner membrane of the mitochondrion, and the eukaryote respiratory chain, the chief source of intracellular ROS, is embedded in and operates across that membrane. Mitochondrial SOD2s can increase in response to ROS (4) and seem to be essential in preventing ROS-linked respiratory and other mitochondrial protein damage (5-9), cell ionizing-radiation resistance (10-11), minimizing the occurrence of certain cancers and tumors (12-14), retarding cell senescence and aging (5, 6, 15-17), preventing myocardial and neurodegenerative effects (6, 7, 9, 18), decreasing the lethal effects of sepsis and septic shock (19), and deferring or preventing apoptosis (programmed cell death) (5-7). Mouse model results seem to confirm that ROS derived from mitochondria over time produces widespread oxidative damage if there is insufficient SOD2 present (20). Cytosolic SOD1 may also assist in mitochondrial ROS defense, because it has been found in the space between the inner and outer mitochondrial membranes in yeast (21). The preceding reports are only a sampling of the literature.

It is intriguing that Mn ions complexed both in SOD2 and in nonprotein SOD analogues (22) in bacteria, mitochondria, and chloroplasts defend against the ill effects of the Earth's atmospheric O₂, whereas Mn ions complexed in the Hill reagent present in the same chloroplasts produce all of Earth's atmospheric O₂.

Mitochondrial SOD2 is synthesized in the cytoplasm from a nuclear gene, the apoprotein passes into the mitochondrial matrix, and the essential Mn cofactor is somehow bound to three histidines and one asparagine in each peptide subunit (23). Mn ions are transported into yeast cells by the natural resistance-associated macrophage protein (Nramp) family high-affinity Mn transporter Smf1p associated with the cell membrane (24). Two years ago Luk and Culotta (25) reported that the Nramp Mn ion transporter in yeast, Smf2p, was essential to the delivery of Mn to the mitochondrial matrix as well as to other intracellular compartments. Thus, unlike Smf1p, Smf2p seems to be a general-purpose, vesicle-associated intracellular Mn transporter. Luk et al. (1) now present good evidence that a yeast nuclear gene that they designate mtm1 (manganese trafficking factor for mitochondrial SOD2) is essential for the loading of the SOD2 apoprotein with the Mn necessary to produce fully active SOD2 in the mitochondrion (1). The protein product of mtm1, Mtm1p, localizes to the mitochondria and seems to be a previously uncharacterized member of the mitochondrial family of metal carrier proteins. The authors show that the mitochondria of yeast mutants with a defective mtm1 gene may contain normal levels of Mn and SOD2 apoprotein, but active SOD2 does not form. Thus the Mtm1p seems to be a Mn-inserting protein specific for mitochondrial SOD2 activation. Fig. 1 illustrates our current understanding of SOD2 activation and function in brewer's yeast (Saccharomyces cerevisiae). A number of questions immediately arise. Does Mtm1p fold apoSOD2, bind and insert Mn into it, or both? Given the distance between the four Mn-liganding residues (His-52, His-107, Asp-194, and His-198) in the SOD2 subunit polypeptide, Mn incorporation without or before polypeptide folding is unlikely. How similar is Mtm1p to the CCS chaperone protein-loading SOD1 with Cu in the cytosol (26) in which the folding of CCS resembles the folding of the SOD1? A defective mtm1 gene results in elevation of both total mitochondrial Fe and Mn (1). Is this a direct effect of Mtm1p loss or a compensatory effect in response to the resultant low levels of active SOD2 or high levels of ROS? Is Mtm1p specific for both Mn and SOD2? Both the +2 and +3 valences of Mn have crystal and hydrated ionic radii that are almost identical to the analogous +2 and +3 Fe ions; thus, sites and chelators specifically binding Mn⁺³ often bind Fe⁺³ almost equally (they have similar stability constants). Does mature Mtm1p specifically bind one or more of these four ions? Can it strip Fe or Mn from nonspecific protein complexes or from the common ferri- or mangani-chelates formed with citrate, oxalate, or pyrophosphate present in the mitochondrial matrix? Heme is reported to down-regulate mitochondrial SOD2 (27): Can Mtm1p ferrate the heme precursor protoporphyrin IX?

Fig. 1.

Diagram of the postulated mechanisms of localization and activation of Mn SOD (SOD2) in a cell of brewer's yeast (S. cerevisiae) based on the findings of Luk et al. (1). c.m., cell or plasma membrane; m.o.m., mitochondrial outer membrane; m.i.m., mitochondrial inner membrane; TCA, tricarboxylic acid cycle generating reducing equivalents for the electron transport (respiratory) chain (sinuous red line); ROS, primarily O₂^- and H₂O₂; GPX, glutathione peroxidase; GSH, reduced glutathione monomer; GSSG, oxidized glutathione dimer; Smf1p, high-affinity Mn uptake protein; Smf2p, hydrophobic protein moving Mn into mitochondria and other intracellular vesicles; Mtm1p, SOD2-activating protein first reported by Luk et al. (1).

The reasons for particular cell or tissue abnormalities associated with very low (or high) expressed SOD2 levels can now be sought in (i) SOD2 gene apoprotein synthesis and regulation (27, 28), (ii) general intracellular Mn transport (Smf2p), (iii) high-affinity Mn uptake from the environment by the yeast cell protein Smf1p, and (iv) specific SOD2 Mn loading and activation via the mtm1 gene and Mtm1p protein. Likewise, these known genes and proteins can be used to experimentally produce cells with a variety of different lesions leading to very low or elevated mitochondrial SOD2. Mutants lacking Mtm1p will be particularly useful if the action of Mtm1p proves to be truly SOD2-specific.

The identification by Luk et al. (1) of the human mtm1 gene homolog, CGI-69, strongly supports the fundamental nature of Mtm1p's activity and its likely relevance to human ROS and SOD2 pathologies as well as those of a wide range of plants, animals, and lower eukaryotes. Clearly there is a lot left to be learned about SOD2 translocation, activation, and regulation, but the way is now open for developing Smf2p, Mtm1p, and CGI-69 protein assays and gene probes and screens to better understand the delicate balance of mitochondrial ROS threat and defense.