R.T. DEAN, * S. NAISH-BYFIELD, J.V. HUNT
* Present address: Heart Research Institute, Missenden Road, Camperdown, Sydney, N.S.W. 2050, Australia.
Cell Biology Research Group, Brunei University
Uxbridge, Middlesex UB8 3PH, UK
and S.P. WOLFF
Laboratory of Toxicology, Rayne Institute, University College
University Street, London WCIE 6JJ, UK
References
Free radicals are produced inevitably during cellular metabolism, such as by electron leakage from electron transport chains and redox enzymes. In addition, leukocytes, and perhaps some lymphocytes, can produce a triggered extracellular flux of free radicals by means of a specialized electron transport chain assembled on the cell surface and involving a low-potential cytochrome, cytochrome b-245.1,2 This triggered radical flux is probably important in defence against foreign organisms, and may also have a positive role in the normal function of the cells. The primary radical produced in biological systems is the superoxide radical, but this may undergo conversion into many other radicals, including the extremely reactive hydroxyl radical. Reactions with lipids and possibly proteins may also deposit relatively stable hydroperoxides, analogous to hydrogen peroxide (the dismutation product of superoxide radicals). All these peroxides are important in that they represent a further source of radical generation after cleavage reactions involving transition metals. Table 1 gives a crude summary of the nature and actions of the biologically important radicals occuring in normoxic conditions.
Studies on the mechanisms of action of such free radicals initially concentrated on lipids and DNA as targets. More recently it has become apparent that proteins are at least as important as targets. I will outline our studies on protein damage by radicals, and indicate some of its consequences for functional activity, proteolysis and cytolysis. I will start with the exterior of cells, move to intracellular proteins and then discuss proteins at the interface, the plasma membrane and their involvement in cytolysis.
Extracellular proteins are catabolized mainly intracellularly after endocytosis. However, limited extracellular fragmentation of proteins may be a necessary preliminary to such endocytosis in the case of polypeptides of the fixed extracellular matrices such as cartilage. These matrices present the substrates in a compact non-diffusible form, and this limitation has to be overcome before substantial degradation can occur. In addition, extracellular degradation of diffusible proteins by limited proteolysis can occur, though it is usually restricted by the large extracellular concentrations of proteinase-inhibitors. We have recently studied the influences of free radicals on the generation of fragments from cartilage proteoglycan and their influence on the proteinase-proteinase inhibitor balance.
In addition we have indicated that significant protein fragmentation may result in diabetes from radicals generated by autoxidation of extracellular glucose.
Our earlier studies4 on free radical attack on intact discs of nasal cartilage showed that defined radical fluxes (induced by gamma-radiolysis) could release macromolecules (biosynthetically labelled by 35sulphate) from the discs. Molecular characterization of these products revealed that virtually intact chondroitin sulphate was the primary product and implied that the main cleavage reactions were taking place on the polypeptide core, releasing oligopeptides attached to chondroitin sulphate side chains, which were themselves almost undegraded unless gross radical fluxes were used. Significant release of intact glycosaminoglycan could be obtained with doses of radicals corresponding to those that can be produced by a few million macrophages in a few hours, in other words, corresponding to biologically plausible doses, especially in inflammation. The limited degradation products of free radical attack on proteoglycans were characterized in more detail recently using purified proteoglycan monomer and intact proteoglycans as substrates. The data (Naish-Byfield and Dean, unpublished) confirmed that selective attack on the polypeptide core is a major event. A clear contrast in the nature of products released by chemical elimination reactions (glycosaminoglycan alone) and by radical attack (glycosaminoglycan attached to oligopeptide) was demonstrated.
Such fragments of connective tissue may possibly be generated in inflammatory conditions by the action of radical-generating cells such as macrophages. In the extracellular fluid they might then be subject to enzymic proteolysis. It has been suggested frequently that free proteinases may occur widely in these inflammatory circumstances because extracellular inhibitors such as alpha-l-proteinase inhibitor (a 1PI) are selectively inactivated by free radicals, leaving free proteinase activity. However, we (Dean, Nick and Schnebli, unpublished) found that free radicals generated from transition metals with hydrogen peroxide have roughly the same capacity to inactivate three proteinase inhibitors (a 1PI, secretory leukocyte proteinase inhibitor and eglin) as they do a relevant target enzyme (neutrophil elastase) when all are used at concentrations appropriate to the in vivo condition (around 5 M). Since many extracellular proteins probably have detectable quantities of transition metals loosely bound (e.g., copper on histidine residues) and since hydrogen peroxide is the dismutation product of the primary radical generated by triggered leukocytes, this radical generating system is probably relevant (Table 1). On the other hand, some peroxy radicals may be more selective in inactivating a 1PI (by virtue of its active site methionine, which is very susceptible to oxidation) than the other inhibitors and enzymes. Whether appropriate peroxy radicals occur in the extracellular fluid at appropriate places and times is rather debatable. So the suggestions that emphysema and other conditions in which extracellular proteolytic activity comes to exceed extracellular proteinase inhibitors may result from free radical affront are therefore insecure, though it is clear that localized environments may contain active proteinases.
As another extracellular event that may mark protein for complete intracellular proteolysis, we have studied the action of autoxidizing sugars on some soluble proteins. Partly consequent on radical generation, proteins are glycosylated. We now report5 that protein fragmentation takes place, with bovine serum albumin as target protein, using glucose and glyceraldehyde autoxidizing for one to eight days. This may cause accelerated catabolism of some proteins in poorly controlled diabetics.
The general point has been raised indirectly above that radical damage to proteins may be important in causing inactivation of protein function. If such non-functional proteins are to be removed, then it may be expected that some feature of the radical modification will lead to recognition of the protein by cellular degradative machinery. In the case of the extracellular released fragments, this recognition is probably based usually on the adsorptive endocytosis of the materials: this can result probably from the unfolding of the protein moiety and the relatively increased exposure of hydrophobic areas, though other more subtle changes may also be important.
In the case of intracellular proteins, radical-mediated damage to proteins may also be important in establishing basal rates of catabolism of intracellular proteins. The consequent residue modification, fragmentation with some associated new N-termini, and unfolding may all help to 'signal' that a damaged protein is more available for catabolism by a variety of routes. For instance, by cytoplasmic mechanisms involving ubiquitin conjugation or by lysosomal routes involving exposure of hydrophobic surfaces on the damaged proteins, which facilitate their uptake into the site of ultimate catabolism. We have demonstrated that several conditions of elevated radical flux lead to an increased rate of proteolysis of bulk long half-life proteins in cultured cells6 and of mitochondrially synthesized proteins in isolated mitochondria.7 A greater acceleration of intracellular protein degradation after radical attack can be observed in erythrocytes.8 The roles of the several routes for intracellular catabolism of radical modified proteins are not established yet, though the main relevant possibilities are summarized in Table 2.
A striking feature of studies of effects on radical fluxes on cell catabolism6 is that while low doses of radical flux may cause accelerated catabolism, which can be construed as a detoxifying protective antioxidant function, higher doses often simply lyse the cells. We9 have investigated this to some extent in relation to macrophage-mediated lysis of T. brucei. We noted an interesting difference in sensitivity to lysis by radicals generated from hydrogen peroxide between bloodstream and procyclic forms, which to some extent could be abolished by removal of the glycoprotein coat of the bloodstream forms. Surprisingly, this coat seemed to be conferring increased sensitivity to radical Lysis; this again indicates the possible importance of protein damage (as opposed to damage to other target molecules). Table 3 summarizes some evidence that macrophages can lyse T. brucei by means of their triggered radical production: the process can be inhibited by antioxidant enzymes that remove hydrogen peroxide (i.e., by catalase, or catalase with superoxide dismutase, but not by superoxide dismutase alone). It is interesting that the addition of metal is not needed in this system: it is provided by the medium. The failure of trolox (an amphiphilic water soluble analogue of tocopherol, vitamin E, the main lipid soluble chain-breaking antioxidant) to prevent cytolysis indicates that lipid peroxidation may not be critical in its mechanism.
We are investigating more closely the mechanisms involved in cytolysis of nucleated human macrophage cell lines by radicals generated from hydrogen peroxide and transition metals.10 We use a simple medium consisting of Hanks' Balanced salts solution, so that the addition of metal is unnecessary. Our concern is to establish whether protein or lipid damage, or both, are crucial early events committing cells to Lysis during such radical attack. We demonstrated that in the case of our cell lines, an event predictive (during radical attack) of Lysis and occurring earlier than any of those presently in the literature, is membrane depolarization. This occurs within a few minutes, while depletion of ATP and GSH is rather slow, and, indeed, not very extensive. Lysis follows the membrane depolarization events with a lag of the order of 1 hour. We argued that changes in membrane potential could reflect either lipid or protein damage. Using a very sensitive fluorimetric assay of lipid damage, we were able to detect lipid damage (as production of hydroperoxides) over a short period comparable with that of depolarization. However, we could completely abolish this lipid peroxidation (by the addition of antioxidants such as butylated hydroxytoluene) without affecting depolarization or subsequent Lysis significantly. In this respect the cells seemed to behave rather like the trypanosomes mentioned above. Therefore, amongst possible cell membrane targets, proteins seem to be the most likely site of primary damage leading to Lysis. In agreement with the comments above on functional inactivation of proteins, we have shown that the activity of certain ion pumping proteins in the cell membrane, notably the Na/K-AtPase, is depressed over an appropriate time scale. It seems that proteins may be critical targets in cytolysis by radicals.
Thus protein damage by radicals is critical in many biological processes. This damage may lead to functional inactivation, but usually the inactivated proteins are probably degraded, so that proteolysis forms a secondary defence. However, when the target proteins are critical for rapid homeostatic mechanisms (as in the case of transport proteins) or when the proteolytic defence and/or other antioxidant defences are overwhelmed, toxic events may ensue. These are probably important in many chronic pathologies such as artherosclerosis (where damage to LDL may be critical) and chronic inflammatory diseases (where connective tissue catabolism probably involves some radical-mediated damage). The toxic events may even be so disastrous as rapidly to lyse cells.
1. HALLIWELL, B. and J.M.C. GUTTERIDGE. 1986. Free Radicals in Biology and Medicine. Clarendon, Oxford.
2. WOLFF, S.P., A. GARNER and R.T. DEAN. 1986. Trends Biochem. Sci. 11: 27-31.
3. THOMAS, S.M., J.M. GEBICKI and R.T. DEAN. In press. Analyt. Biochem.
4. DEAN, R.T., C.R. ROBERTS and L. FORNI. 1984. Biosci. Repts. 4: 1017-1026.
5. HUNT, J.V., S.P. WOLFF and R.T. DEAN. In press. Biochem.
6. VINCE, G.S. and R.T. DEAN. 1987. FEBS Lett. 216: 253-256.
7. DEAN, R.T. and J.K. POLLAK. 1985. Biochem. Biophys: Res. Commun. 126: 1082-1089.
8. DAVIS, K.J.A. 1986. Free Radicals Biol. Med. 2: 155-173.
9. ROSSI, B.C. and R.T. DEAN. 1988. Exp. Parasitol. 65: 131-140.
10. RICHARDS, D.C.H., W. JESSUP and R. DEAN. In press. Biochem. Biophys. Acta.
Table 1. Some important Free radicals and related molecules in oxygenated biological systems
|
Radical |
Source/site |
Comments |
|
Superoxide radicals (O2) |
Electron transport chains (intracellular and in plasma membranes of some leukocytes, such as macrophages) |
Not a very reactive entity; gives rise to H2O2 by spontaneous and catalyzed dismutation |
|
Peroxyl radicals (RO2) |
Lipids, sugars and proteins that may be relatively stable |
Selective in reactivity; generate hydroperoxides |
|
Alkoxyl radicals (RO) |
Lipids, other macromolecules, reactions of and from peroxy radicals |
More reactive than peroxyl radicals in many contexts |
|
Hydroxyl radicals (OH) |
From some reactions of quinones and from reaction of transition metals with H202 (the Fenton reaction) |
Highly reactive; rather unselective in action |
|
H2O2 |
From dismutation of superoxide radicals and from some oxidase reactions |
Together with transition metals, generates hydroxyl radicals; diffusible through membranes |
|
Hydroperoxides (ROOH) |
On lipids and proteins |
Can be fragmented by transition metals to give alkoxyl and peroxyl radicals |
Table 2. Possible mechanism for enhanced protein degradation after radical damage
|
After limited radical attack |
|
|
|
polypeptide-histidine-polypeptide |
|
¯ |
|
polypeptide + aspartate-polypeptide |
|
| |
|
(enzymatic addition of N-terminal basic amino acid) |
|
¯ |
|
(lysine or arginine)-aspartate-polypetide |
|
| |
|
recognition by the ubiquitin-dependent proteolytic system (present in most cellular compartments) |
|
¯ |
|
enzymatic-proteolysis |
|
|
|
After extensive radical attack |
|
1) unfolding of proteins makes them: |
|
a) more susceptible to proteinases |
|
b) more hydrophobic, and therefore more likely to enter lysosomes from the cytoplasm, or endosomes (and hence later lysosomes) from the extracellular fluid |
|
2) Radical attack fragments proteins, and hence in the case of fixed proteins of extracellular matrices makes them diffusible and more accessible for endocytosis and degradation by cells such as macrophages |
Source: The basis for the ideas above are summarized in Dean, 1987, FEBS Lett. 220: 278-282; and Wolff et, al., 1986, TIBS 11: 27-31.
Table 3. Lysis of bloodstream forms of Trypanosoma brucei by the macrophage oxidative burst
|
Conditions
|
% specific lysis by macrophages |
||
|
Untriggered |
PMA-triggered |
||
|
Control |
0.3 (0.3) |
60.3 (1.4) |
|
|
|
(a) + catalase (2000 U/ml) |
0.0 (0.5) |
00.1 (0.9) |
|
|
(b) + superoxide dismutase (300 U/ml) |
1.0 (0.9) |
58.0 (2.2) |
|
|
(c) + catalase and superoxide dismutase (as above) |
0.2 (0.6) |
00.8 (1.1) |
|
trolox (1mM) |
2.1(0.6) |
56.1(5.1) |
|
Bloodstream trypanosomes (biosynthetically labelled in their protein with leucine) were co-cultured with macrophages (from 7 day infected mice) at a 10:1 macrophage:trypanosome cell ratio for 2 hours. The medium was M 199 plus 2.5% rat serum. Phorbol myristate acetate (PMA) was at 50 ng/ml. Results are means (with std. in brackets, for n = 3). From ref. 9.
