To maintain the cellular functionality, damaged or modified proteins have to be recognized and degraded. On the other hand, proteins after they have served their purpose have to be removed from the cell; therefore, proteins are submitted to a permanent turnover. During evolution several cellular mechanisms performing those tasks have been developed, including the proteasomal system.
The proteasomal system is a very complex and several-fold regulated “degrading machinery” of cells, consisting of the 20S “core” proteasome and several different regulator proteins. It works closely together with the ubiquitinating system, actively labelling proteins by polyubiquitination for proteasomal degradation.
Although the recognition and removing of modified proteins is a major task of the proteasomal system, it is involved in many other processes including immune response (antigen presentation) and cancerogenesis.
Many of the intracellular functions might be still unknown, particularly specific regulatory functions of the cellular protein homeostasis via the ubiquitin-proteasome-system (UPS) as well as the regulation of protein ubiquitination itself. In this review, we summarize the known facts and reasonable hypotheses concerning the proteasomal system focussing on the degradation of oxidized proteins.
PROTEIN OXIDATION IN MAMMALIAN CELLS
The main intracellular sources of the primary radical superoxide (O2•−) are the mitochondria. Superoxide is a byproduct of the oxidative phosphorylation: it results from electrons that are “leaking” mostly out of the complexes I and III and are transferred to molecular oxygen1, 2. The superoxide is moderately reactive, but due to the interaction with other oxidants a number of potentially reactive metabolites are formed. The nitric oxide radical (•NO)3, 4 is one of these components, produced by one of the intracellular nitric oxide synthases (NOS). In mammalian cells, an inducible (iNOS)5, a neuronal (nNOS)6, an endothelial (eNOS)7, and a mitochondrial NOS8, 9 is found, releasing •NO by converting arginine to citrulline while consuming O2 and NADPH10. The relatively long half-life and the ability to permeate membranes make •NO to an important intra- and intercellular messenger11. The interaction of both radicals is resulting in the generation of peroxynitrite (ONOO−)12, 13, a very reactive anion. Peroxynitrite is not a radical, but in its protonated form, the peroxynitric acid, decomposes extremely fast while releasing both nitryl (•NO2) and hydroxyl radicals (•OH)14, 15, two of the most aggressive radicals found in cellular systems, able to induce protein modification. So, two protein modifications often used to quantify the amount of oxidative stress in cells are the result of this reaction: 3-nitrotyrosine16, 17 (by •NO2) for nitrosative stress, and protein carbonyls18–20 (by •OH), the most often used protein related “stress marker”21.
To avoid oxidative damage or modifications of cellular molecules, structures or whole organelles, threatening the viability of the cell, several antioxidative “lines of defence” developed during evolution. These three lines of the antioxidative capacity of the cell are depicted in Fig. 1. Normally, the antioxidative defence is sufficient to preserve protein homeostasis. If, on the other hand, the cellular antioxidative capacity is overwhelmed by reactive species, the result is defined as “oxidative stress”26, 27. A protein that is oxidatively modified runs through three different stages of damage28, depending of the amount of damage it sustained. The first stage is represented by a slight oxidation and a marginal reduced activity. Sometimes such proteins show an increased solubility that can be attributed to the induction of negative charges by oxidation and the resulting increased polarity29, 30. The second stage is an increased oxidation accompanied by disfolding of the protein. These proteins show no more activity and a decreased solubility caused by the exposure of their hydrophobic inner structures to the surface. These proteins are the preferred substrates for enzymatic degradation. If such proteins are not degraded in time and/or the cell suffers stronger oxidative stress, further oxidation might occur. This is possibly accompanied by cross-linking to other proteins, perhaps induced by products of lipid peroxidation like the bifunctional aldehydes HNE31, 32 or MDA33, 34. Severe oxidized and cross-linked proteins have reached the third stage of oxidative modification and are not longer accessible for proteolytic degradation. Such proteins tend to build hydrophobic and insoluble aggregates and are preliminary stages of the so-called lipofuscin35–37, an indegradable and nonexocytosable material; in the literature sometimes also referred to as “age pigment”38 or “ceroid”39. If the proteolytic systems of the cell40–42 fail to degrade oxidized proteins in time, the likelihood of cross-linking events increases and lipofuscin is formed. Since lipofuscin is able to inhibit the proteasomal system43–45, thus reducing the degradation of other oxidized proteins and again increasing the amount of heavily oxidized and cross-linked indegradable proteins, this vicious cycle results finally in a self-catalyzing formation of lipofuscin37. Lipofuscin is continuously accumulated in lysosomes, thus decreasing besides proteasomal also lysosomal protein turnover36, 46.
As shown, oxidative stress results in an intracellular formation of oxidatively damaged proteins that might jeopardize the cellular functionality. Therefore, oxidized/misfolded proteins have to be rapidly recognized and degraded. Several proteolytic systems could be responsible for that, such as the cytoskeleton-localized calpains47, 48, the lysosomal cathepsins49, 50, and the proteasomal system. It was shown, that the 20S proteasome, the catalytic center of the proteasomal system, is responsible for the degradation of about 90% of all intracellular oxidatively damaged proteins, showing the highest substrate specificity for proteins that are in the second stage of damage (see earlier).
The 20S “core” proteasome, discovered in 1968 by Harris and called “cylindrin”51, is a barrel-shaped structure (about 160 × 100 Å252), build of two homologous rings, an α- and β-ring, arranged in the order αββα. Each ring contains seven different subunits (α1-α7, respectively β1-β7) summarizing to an overall proteasomal mass of about 700 kDa (Fig. 2). The α-rings provide the substrate specificity by recognizing and binding of exposed hydrophobic structures. In addition, the N-terminal parts of the α1-3, α6, and α7 strictly regulate the entrance to the inner proteolytic chamber, build by the β-subunits. How exactly this entrance is regulated remains still unknown, but low doses of SDS seem to induce a slight unfolding of these “gating”-structures, resulting in an increased substrate turnover caused by better substrate accessibility54, 55. The inner proteolytic chamber has a maximal diameter of about 53 Å56, whereas three of the β-units show proteolytic activity: β1 (a peptidyl-glutamyl-peptide-hydrolising or caspase-like), β2 (a trypsin-like), and β5 (a chymotrypsin-like activity), each of them having the active center in form of a threonine residue at the N-terminus of its protein chain. Unfolded proteins are normally cleaved into peptides with a length of 2–30 amino acids, showing three distribution peaks at 2–3, 8–10, and 20–30 amino acids57. This protein degradation by the 20S “core” proteasome is ATP-independent. The proteasome shows a relatively high intracellular concentration of about 1–20 μg per 1 mg of soluble protein and has an intracellular half-life of about 5 days (measured in HeLa cells)58. Among other, the whole proteasome assembly is dependent from the proteasome maturation protein (POMP)59, 60, composing the multicatalytic complex from inactive precursors, activated by intramolecular autolysis61. Phylogenetic trees date the evolutionary “invention” of the proteasome back in the time before archaea and eukaryotes split up62. The 20S “core” proteasome is found both in the cytosol and the nucleoli of the mammalian cell, making an exception of the nucleoli and the ER lumen63. In the cytosol it is often associated with the cytoskeleton and the ER. It can be transported from the cytosol into the nucleus via an unidirectional karyoperin α,β import guided by nuclear localization signals64, 65. Homogenous distributions of both nuclear and cytosolic 20S fractions are only visible after dissolution of the nuclear envelope in mitosis66. Investigations of the intracellular distributions of so-called proteasomal regulators, particles binding facultative to the proteasome influencing specificity and/or activity, showed same results67.
γ-Interferon secreted in inflammatory reactions by CD4+ and CD8+ T-cell induces the expression of the so-called inducible forms of β1i, β2i, and β5i replacing the constitutive subunits during de novo synthesis68, whereas the constitutive forms of β1, β2, and β5 are downregulated69, 70. Since this is a very quick process, many intermediate forms can be observed71, showing large variation in fragment length produced by the cell during protein degradation. The average length of fragments produced by a proteasome containing all three βi-subunits, the immunoproteasome, is about 8–10 amino acids, and thus ideal for being presented as antigen by the MHC I molecules on the surface of infected cells72, 73. The immunoproteasome shows a shorter half-life than the constitutive form of about 27 h58.
Today, some mathematical models57, 74, 75 of proteasomal protein degradation have been performed, and meanwhile even programs are available that can predict the cleavage-pattern of a degraded protein76 if the flanking ±5 amino acids of the potential cleavage-site of the protein are known. How exactly the proteasomal degradation of a substrate is performed has to be investigated in detail, especially how the protein is feed stepwise into the inner chamber. One hypothesis concerning the substrate feeding was proposed by Kisselev et al.77, called “bite and chew”, suggesting mutual activation and inactivation while substrate binding and cleaving, but experiments using proteasomal inhibitors showed newer that there are no allosteric interactions of the β-rings active centers78. The cleavage pattern seems to be determined by substrate binding to inner structures of the proteasome52.
REGULATORS OF THE 20S “CORE” PROTEASOME
Three main proteasomal regulators are known that are able to bind to one or both of the α-rings of the 20S “core” proteasome: 11S (also termed as PA28 or REG)79, 80, the recently discovered PA200 (Blm10)81, 82 and the 19S (PA700 or RP)83, 84 regulator.
The 11S regulator is γ-interferon inducible like the proteasomal βi-subunits and is also involved in the cellular immune response. It is a hexa- or heptameric ring structure build of one or two of three different subunits α, β, and γ. The cytosolic 11S regulator contains α and β subunits in the form 3α3β (hexameric) or 3α4β, respectively, 4α3β (heptameric), and is especially expressed in tissues involved in immune response85. The nuclear form is a heptamer containing also γ-subunits, which are only found in the nucleus86, showing highest concentrations in the brain87. In contrast to the cytosolic 11S regulator, which activates all three proteolytic sites of the proteasome, the nuclear form only stimulates the trypsine-like activity localized on the β2-subunit. The cytosolic 11S form is involved in MHC I antigen presentation, the function of the nuclear form is still not clear. All forms attach to the α-rings of the 20S proteasome, whereas the proteasomal gate and the regulators pore are superimposed opening a channel into the proteolytic chamber resulting in increased activity.
The PA200 activator is a 200-kDa single-chain cap-shaped protein, able to bind to one α-ring of the “core” particle, while contacting all α-subunits except α7 and having a base diameter of about 100 Å and an axially length of 60 Å, performing proteasomal activation via opening the gated 20S channel to the inner proteolytic chamber81. Redistribution of PA200 in the nucleus of γ-irradiated HeLa cells to punctuate patterns, reminding to the reactions of many DNA repairing proteins, suggest a participation of the protein in DNA repair, confirmed by experiments showing a hypersensitivity of yeast having PA200 mutations to bleomycin82; in between 20S-PA200 complexes attached to chromatin after γ-irradiation have been shown as well82. Until now, three different forms of PA200 are identified via their mRNA: PA200i, ii, and iii, but only PA200i seems to bind the 20S “core” proteasome88.
The 19S regulator, a particle, build of 17 different subunits with single masses ranging from 15 to 112 kDa89 summarizing to a molecular weight of about 700 kDa. These subunits are divided into two general classes of subunits: the Rpt-family with and the Rpn-family without ATPase activity. The 19S regulator has a base ring, binding to one of the 20S α-rings, build of the subunits Rpt1-Rpt6, as well as Rpn1, Rpn2, and Rpn10. The substrate-entry regulating lid, attached to the base ring of 19S, contains the subunits Rpn3-Rpn9, Rpn11, and Rpn12.
Each of these regulators can bind to one or both ends of the 20S “core” proteasome. Furthermore, hybrid proteasomes86, 90 were found, binding two different regulatory proteins at the same time (Fig. 3). These hybrid forms show an enhanced hydrolysis of small peptides without significant elevation of proteolysis92 and an altered cleavage pattern than the normal 26S proteasome, suggesting an involvement in antigen presentation.
THE UBIQUITIN-PROTEASOME SYSTEM
A 20S “core” proteasome, binding a 19S regulator to each of its ends, is called 26S proteasome, a huge proteolytic complex with an overall mass of about 2 MDa. The 26S proteasome shows a changed substrate specificity and an up to 20-fold increased proteolytic activity compared to 20S93. This is achieved by the opening of the α-ring gate-structures for better substrate accessibility. Both binding of the 19S and substrate degradation of the 26S proteasome are strictly ATP-stimulated, since in an ATP-depleted environment the 19S regulator detaches from the 20S “core” particle. The main substrates of the 26S proteasome are polyubiquitinated proteins, even though the 26S is able to degrade other substrates as well. A central role in this process are played by the 19S subunits Rpt2 (opening the gate of the 20S proteasome, controlling substrate entry and release)94, Rpt5 and Rpn10 (both binding polyubiquitinated substrates), and Rpn11, a Zn2+-dependent metalloprotease, which releases ubiquitin monomers from 26S substrates95.
Ubiquitination, the enzyme controlled attachment of a 76-amino acid polypeptide called ubiquitin (Ub) to a target protein, is a reversible modification, changing the activities or the fate of the according target.
The best-known function is the role of Ub as degradation signal. In this case, at least three further Ub molecules have to be linearly attached to the first Ub on the target molecule via the Lys48 residue of Ub. Even though a chain of four Ub molecules attached are sufficient to degrade a protein, usually longer Ub-chains (about 12 molecules) on a target protein are found. Experiments comparing the length of the attached Ub-chain and the strength of the “degradation signal” have been performed by Thrower et al.96 resulting in a strong nonlinear increase for elongation from 2 to 4 Ub molecules (about 100-fold), but only an increase lower than 10-fold for an Ub-chain elongation from 4 to 12 molecules. From 4 to 8 Ub molecules in the chain attached to the doomed target a 6.6-fold increase in the affinity to the 26S proteasome was found.
Corresponding investigations showed that Lys63-linked polyubiquitin chains, monoubiquitin are not involved in protein degradation but in processes like DNA-repair, signal transduction, and regulation of transcription97.
Substrate ubiquitination is performed by a chain of three different protein classes: E1, E2, and E3 enzymes. The first step is the ATP-consuming binding of Ub to E1. E1 transfers Ub to E2 and E3 catalyses by binding both E2-Ub and the substrate the ubiquitination of the substrate, in which Ub is bound to a lysine-residue. After this step, E2 and E3 are released and the cycle repeats until the substrate is polyubiquitinated. The substrate specificity here is delivered by E3 class of enzymes present as several hundreds of different forms, whereas E1 is a single form and E2 several dozens98. The UPS is involved in many cellular functions including cell cycle regulation, immune response, and ER-associated degradation of misfolded proteins (“quality control”)56.
Now it has been shown that the E1-E2-E3-machinery is not sufficient for the polyubiquitiniation of a target protein but an additional factor is needed99, since the same E3-enzyme would both be able to attach Ub to the target protein and to a growing chain of Ub molecules. This additional factor has been termed “E4” and was so far found in yeast, mouse, and human cells100–102.
One actual model is discussed by Hoppe103, in which monoubiquitination performed by E1-E3 “activates” an enzyme involved in specific cellular functions, whereas E4 “inactivates” the enzyme by elongation of the attached Ub-chain, thus marking it for 26S-proteasomal or lysosomal degradation.
THE DEGRADATION OF OXIDIZED PROTEINS—A FUNCTION OF THE 20S PROTEASOME
As already stated, the large majority of oxidatively damaged proteins both in the cytosol and the nucleus of mammalian cells are removed by the 20S proteasome. Experimental investigations to determine the degradation rate of oxidized proteins in mammalian cells have been performed by our group17, 41, 104–108 and by others109–113, by applying external stress using different oxidative agents, including H2O2, ONOO−, SIN-1, and paraquat. In all these cases it was demonstrated that inhibition or removal of proteasomes resulted in a loss of the ability of cells to degrade oxidized proteins114, 115. Interestingly, the loss of a functional ubiquitin is not affecting the removal of oxidized proteins116, 117. Since it is assumed that various oxidants are involved in the formation of various forms of oxidized proteins, it seems to be clear that there should be a common recognition mechanism. It was proposed by the groups of Stadtman and Davies118–120 and later confirmed by our group121–124 that this common recognition motive is the exposure of hydrophobic amino acids to the surface of the protein, which means at least a partial unfolding of the substrate. The α-rings of the 20S “core” proteasome are responsible for substrate recognition and binding. The exact mechanism is still unclear, also how the gated channel of the 20S proteasome is opened or whether additional factors are required. Investigations of substrate binding, release and degradation of side-on and bottom immobilized proteasomes revealed that one entry to the proteolytic chamber is sufficient for substrate access and release of the products125. A positive cooperativity of substrate binding has been shown: The proteasome can degrade two substrates at the same time, as well as it can cleave proteins without accessible “ends” while showing endoproteolytic behavior126. It is supposed that cleavage products leave the proteasome mainly via the (opposite) substrate entry but it could be possible to release short peptides through gaps between the α- and β-rings that are still too small to allow substrate access127.
Oxidative stress in its connection to aging and many pathologies is a complex network of oxidant formation pathways and the antioxidative capacity of the cell.
Since the 20S proteasome is a complex multicatalytic protease, highly regulated, and all these components are also exposed to stress-related changes, the exact knowledge of the complex mechanisms of the cellular regulation of the UPS is important. A further understanding of the removal of oxidized proteins in cells will deepen our understanding of the pathogenesis of such diseases as neurodegenerative diseases and cancer.
One of the results of this basic research is the successful clinical usage of proteasome inhibitors: The devastating effects of proteasome inhibition in fast growing and angiogenesis-dependent tumor cells are much worse than in normal body tissue; thus proteasome inhibition became an interesting new approach in cancer therapy. One example is the proteasomal inhibitor bortezomib (PS-341, a boronic acid derivative128, available as the medicament Velcade129), which is used especially in multiple myeloma treatment130 showing encouraging results.
The functions of the ubiquitiniating systems, especially the substrate recognition and binding of the highly specific E3-enzymes might become an important role in the therapeutic regulation of cellular protein equilibrium, especially in new approaches in cancer therapy. One of these therapeutic attempts is the specific inhibition of single E3 ligases: the tumor suppressor protein p53 is normally bound to its specific E3 ligase MDM2 in an inactive form unable to induce apoptosis and normally the intracellular level of free p53 is very low. However if nutlin-3 is added, a low-molecular MDM2 antagonist, the intracellular concentration of free p53 increases resulting in apoptosis131. But this “strategy” is not new; the human papilloma virus expresses the E3 ligase E6-AP that provides degradation of p53 resulting in cervical tumors132. Furthermore, the hepatitis B virus encodes its own proteasome inhibitor, the protein X133 interfering with p53, as well and thus promoting hepatic carcinogenesis. The clinical application of artificial E3 ligases, E3 inhibitors might give access to new therapeutic approaches to pathologies resulting from shifted protein homeostasis.
Further experimental results might reveal today unknown strategies in preventing or solving intracellular protein aggregates and thus a slowing down or perhaps a rate-limiting factor in postmitotic aging.