Ornithine decarboxylase (ODC) is the first and rate-limiting enzyme in the biosynthesis of polyamines, low-molecular-mass aliphatic polycations that are ubiquitously present in all living cells and are essential for fundamental cellular processes. Most cellular polyamines are bound, whereas the free pools, which regulate cellular functions, are subjected to tight regulation. The regulation of the free polyamine pools is manifested by modulation of their synthesis, catabolism, uptake and excretion. A central element that enables this regulation is the rapid degradation of key enzymes and regulators of these processes, particularly that of ODC. ODC degradation is part of an autoregulatory circuit that responds to the intracellular level of the free polyamines. The driving force of this regulatory circuit is a protein termed antizyme (Az). Az stimulates the degradation of ODC and inhibits polyamine uptake. Az acts as a sensor of the free intracellular polyamine pools as it is expressed via a polyamine-stimulated ribosomal frameshifting. Az binds to monomeric ODC subunits to prevent their reassociation into active homodimers and facilitates their ubiquitin-independent degradation by the 26S proteasome. In addition, through a yet unidentified mechanism, Az inhibits polyamine uptake. Interestingly, a protein, termed antizyme inhibitor (AzI) that is highly homologous with ODC, but retains no ornithine decarboxylating activity, seems to regulate cellular polyamines through its ability to negate Az. Overall, the degradation of ODC is a net result of interactions with regulatory proteins and possession of signals that mediate its ubiquitin-independent recognition by the proteasome.
The polyamines spermidine, spermine and their precursor putrescine are small aliphatic polycations found in all living cells. Polyamines are fundamental for a variety of important cellular processes, the main ones being the processes associated with cell growth and proliferation. Depletion of cellular polyamines results in inhibition of cellular proliferation that, in most cases, can be completely reversed by readdition of polyamines to the growth medium. Although their mechanism of action in supporting important cellular functions is mostly unknown, it is clear that maintaining intracellular polyamine levels is essential for cellular wellbeing. It is therefore not surprising that cells utilize sophisticated regulatory mechanisms for maintaining optimal polyamine levels. At the heart of this regulation is an autoregulatory circuit that mediates ornithine decarboxylase (ODC) levels through polyamine-stimulated expression of a regulator, antizyme (Az), that stimulates ODC degradation and itself is regulated by antizyme inhibitor (AzI). In this paper, I focus on this regulatory circuit with an emphasis on the way Az and AzI affect ODC degradation and on the mechanism of their regulation.
ODC is degraded by the 26S proteasome in a ubiquitin-independent manner
In its functional form, ODC is a homodimer harboring two active sites that are shared at the interface between the two subunits . The active dimer is loosely bound, being in equilibrium with inactive monomers [2,3]. ODC is one of the most rapidly degrading proteins in mammalian cells. Although the vast majority of short-lived proteins are targeted for proteasomal degradation by their conjugation to ubiquitin [4,5], ODC is degraded by the proteasome in a ubiquitin-independent manner. This conclusion is based on the observations that (i) ODC is efficiently degraded in vitro, in reticulocyte lysate depleted of the ubiquitin-activating enzyme E1, (ii) ODC is degraded in fraction-2 of reticulocyte lysate, which lacks ubiquitin and the entire enzymatic machinery required for conjugating ubiquitin to target proteins, and (iii) ODC activity decays rapidly in ts85 cells harboring a thermosensitive ubiquitin-activating enzyme when grown at the non-permissive temperature [6,7]. Studies performed in vitro [8,9] and in vivo in yeast cells harboring mutant proteasomes  demonstrated that ODC is degraded by the 26S proteasome.
Az stimulates ODC degradation
Although ODC is subjected to basal degradation, this degradation is greatly stimulated by polyamines. Early studies from the Canellakis laboratory have demonstrated that polyamines induce an ODC inhibitory activity that has characteristics of a short-lived 26.5 kDa protein . This remarkable early description in which the name antizyme was given to this activity, was followed by enormous efforts in the Hayashi laboratory resulting in the cloning of Az cDNA and gene [12,13], paving the way to our current understanding of the fascinating autoregulatory circuit that controls ODC and cellular polyamines. Az binds to transient ODC monomers with higher affinity than they bind each other. Recent studies suggest the structural and energetic basis for this higher affinity [14,15]. The interaction of Az with ODC subunits results in two outcomes: (i) inactivation of ODC, and (ii) stimulation of ODC degradation. The inactivation is rationalized by the demonstration that the Az-contacting surface overlaps with the ODC homodimerization surface, thus inhibiting the formation of the active ODC homodimer [14,15]. Although inactivation of ODC is a dramatic outcome, it is an intermediate step in the process of stimulating ODC degradation. In yeast, it was recently demonstrated that polyamines directly affect the antizyme-mediated degradation of ODC by the proteasome .
Sequences that mediate ODC degradation
On the basis of comparing the sequence of the short-lived mouse ODC protein and the stable Trypanosoma brucei ODC [17–19] and on deletion analysis , it was demonstrated that two segments of ODC are required for its rapid degradation. The first, harboring amino acids 117–140, is required for Az binding. The second, encompassing the most C-terminal 37 amino acids of the mammalian enzyme (amino acids 424–461), was suggested to act as a degron. When it is appended to the C-terminus of the stable trypanosome ODC or even to ODC-unrelated stable proteins, it converts them into rapidly degrading proteins [17,18,21–23]. Sequence analysis of a stable ODC variant from HMOA cells demonstrated that it contains a C441W replacement . Mutating this cysteine residue in the wild-type protein inhibited ODC degradation, revealing its importance for the function of the C-terminal degron . Two structural elements are required for efficient degradation of a protein. The first mediates recognition and association with the proteasome (in most cases, this function is provided by covalently attached polyubiquitin chains). The second is an unstructured segment in the target protein that enables its initial infiltration into the proteolytic cavity of the proteasome [26,27]. In most cases, these two functions are separate. In ODC, it was suggested that both functions reside within the C-terminal degron that was predicted to be unstructured [28,29]. Although the C-terminal degron of ODC was demonstrated to be sufficient in conferring lability when appended to otherwise stable proteins, its efficiency in the context of native ODC is limited, unless interaction with Az is manifested. Antibodies prepared against the C-terminal degron recognized ODC much better upon interaction with Az. It was therefore suggested that the C-terminal degron is buried within the active ODC homodimer, and that it becomes exposed due to conformational changes imposed by the interaction of ODC with Az [30,31]. However, recent crystallographic and NMR spectroscopy analysis demonstrated that the C-terminal degron of ODC remains structurally unchanged upon interaction with Az. Moreover, it was suggested that an adjacent segment, encompassing residues 395–421, which becomes disordered and non-visible in the ODC–Az structure, may be involved in targeting the ODC–Az complex to the proteasome [14,15]. This possibility was supported further by the demonstration that the C-terminal deletion mutant, lacking amino acids 424–461, interacts with the proteasome as efficiently as the wild-type enzyme . These studies suggest that, although they are adjacent, there is functional separation between the segment that mediates proteasomal recognition (amino acids 395–421) and the most C-terminal unstructured segment (amino acids 424–461) that is responsible for the initiation of the infiltration of ODC into the proteasome. Interestingly, Az also contains two distinct domains required for stimulating ODC degradation: a C-terminal domain that is required for ODC binding and an N-terminal domain that stimulates ODC degradation [32–34]. Although the mode of action of the C-terminal domain is obvious, that of the N-terminal segment is obscure. It could, for instance, be required for the exposure of the proteasome recognition segment of ODC (a testable possibility), or it can compose, together with the recognition segment of ODC, a complete degradation signal. This possibility is supported by the structure-based observation demonstrating close proximity between the proteasome recognition surfaces of ODC and Az within the ODC–Az heterodimer . The proteins that become rapidly degraded when the mammalian C-terminal degron is appended to them, do not bind Az. It must therefore be assumed that, in these chimeric proteins, the mammalian C-terminal degron is sufficiently exposed without the need for interaction with Az. However, the interpretation of this result that assumes the proteasome-recognition role to the C-degron is incompatible with the recent claim that the most C-terminal 37 amino acids of the mammalian enzyme do not function as the proteasome recognition signal . Interestingly, the mammalian enzyme is stable when expressed in trypanosome cells even in the presence of rat Az, suggesting that it is not recognized by the trypanosome proteasome or that trypanosome cells contain an inhibitory factor(s) . These possibilities can be validated by testing the degradation of trypanosome and mammalian ODCs using purified proteasomes. It was similarly demonstrated that yeast ODC is rapidly degraded in yeast cells, but is stable when expressed in mammalian cells even together with yeast Az, most probably because it is not recognized by the mammalian proteasome . Yeast ODC lacks a segment homologous with the C-terminal degron of the mammalian enzyme. The rapid degradation of yeast ODC in yeast cells is mediated by an N-terminal segment that is unique to the yeast enzyme and whose deletion stabilizes the protein . This N-terminal segment that seems to be exposed upon interaction with Az, is unstructured but is not sequence-specific, and can be transferred and impose rapid ubiquitin-independent degradation on some stable proteins . Therefore, although it is clear that this N-terminal degron fulfills the function of initial threading into the proteolytic cavity of the proteasome, it is not clear whether and how it mediates proteasomal recognition. Interestingly, although the C-terminal degron is essential for the degradation of mouse ODC in yeast cells, the integrity of the Az-binding site is dispensable. This suggests that, in yeast cells, the C-terminal degron of mouse ODC is exposed without requiring interaction with Az [37,39]. Yeast and mammalian ODC also differ in their degradation in cells harboring different mutations that affect the proteolytic activities of the yeast proteasome [10,39].
Degradation of monomeric ODC by the 20S proteasome
The Az-stimulated degradation of ODC, which responds to intracellular polyamine concentrations, is manifested by the action of the 26S proteasome. It was demonstrated that monomeric ODC can be subjected to degradation by the 20S proteasome via a pathway regulated by the flavoenzyme NAD(P)H quinone oxidoreductase (NQO1) . This pathway is not affected by the intracellular concentration of polyamines, does not require the C-terminal degron and does not rely on interaction with Az or ubiquitination. In this pathway, transient ODC monomers are spared from default 20S proteasomal degradation by interaction with NQO1, interaction that is disrupted by dicoumarol resulting in stimulated degradation. Since this degradation of ODC is manifested (at least in vitro) by the 20S proteasome, which lacks unfolding activity, it must be assumed that the degraded ODC monomers are mostly disordered. This is also compatible with demonstrations that intrinsically disordered proteins can be degraded by the 20S proteasome [41–43].
Az is the central element in the autoregulatory circuit that controls cellular polyamine levels via modulating ODC degradation and polyamine uptake activity [44,45]. Synthesis of Az occurs via a polyamine-stimulated programmed +1 ribosomal frameshifting that subverts the translating ribosome from one to a second open reading frame, skipping a stop codon to generate a mature functional Az protein [46,47]. In yeast, it was demonstrated that polyamine sensing by the growing nascent Az stimulates decoding of its mRNA . The tight regulation of Az's expression by the level of polyamines may suggest that Az exerts its regulatory functions exclusively by regulating cellular polyamines. Several studies have implied that Az may regulate cell growth and proliferation also by regulating the expression of genes that do not belong to the polyamine metabolism [49–55]. However, it was demonstrated that as long as the supply of polyamines is secured, functional Az can be expressed to high level without affecting cellular proliferation. This finding negates these studies and supports the notion that Az regulates cellular proliferation only by regulating cellular polyamines . However, this does not exclude the possibility that Az regulates other proteins that are not involved in regulating cell proliferation. Translation of Az is initiated at one of two initiation codons, with the second being the main initiation site due to its localization within a favorable Kozak sequence context [57,58]. Although both forms inhibit ODC activity and stimulate ODC degradation, only the longer form, initiating translation at the first initiation codon, is localized in the mitochondria. It was suggested that Az provokes apoptosis through a mitochondrion-mediated mechanism . However, a specific role for mitochondrial Az has not been described to date. A 1:1 ratio is required for inhibition of ODC activity by Az. In contrast, stimulation of ODC degradation by increasing its affinity to the proteasome  is catalytic, suggesting that Az is recycled . Indeed, in an in vitro degradation reaction, Az remains stable while presenting ODC to the proteasome. In mammalian cells, however, Az is a rapidly degraded protein, but its degradation is independent of that of ODC and is ubiquitin-dependent . Similarly, it was demonstrated that, also in yeast cells, Az is subjected to ubiquitin-dependent degradation . Interestingly, this study has demonstrated that, in addition to being stimulated by polyamines at its synthesis, polyamines also regulate Az levels by inhibiting its degradation. It is presently unknown whether polyamines affect Az's degradation also in mammalian cells. It is also unknown at what stage of the degradation process polyamines exert their effect and which ubiquitin ligase(s) are involved. The above-described Az is the first discovered member of a family of proteins that contains at least two additional isoforms and is therefore termed Az1. Az1, which is abundantly expressed, exhibits similar tissue distribution to Az2, the second member that displays higher evolutionary conservation but is expressed at significantly lower levels . A third member, Az3, is expressed predominantly in testicular haploid germinal cells [63,64]. Like Az1, Az2 also inhibits ODC activity, targets ODC to degradation (in cells, but not in an in vitro degradation reaction) and inhibits polyamine uptake activity [65,66]. In contrast, Az3 inhibits polyamine uptake and ODC activity, but fails to stimulate ODC degradation .
Like Az, AzI was also first described as an activity capable of negating Az functions . Its cloning revealed that AzI is highly homologous, although not identical with ODC and that it lacks ornithine-decarboxylating activity [68,69]. The ability of AzI to inhibit Az activity relies on its higher affinity towards Az compared with that of ODC . A structural study revealing the crystal structure of AzI, combined with a biochemical analysis, demonstrated that, owing to fewer interactions at the dimer interface, smaller buried surface area and lack of symmetry of the interactions between residues located at the monomers, AzI exists as a monomer under physiological conditions . This higher availability of AzI monomers partially explains the higher accessibility of AzI subunits to Az. Although most residues involved in Az binding are either identical or conserved between ODC and AzI, there are two substitutions (Asn327 and Tyr331 of ODC by Ala325 and Ser329 of AzI) that may explain further the increased affinity of AzI to Az. Moreover, it also seems that the formation of the AzI–Az complex may be energetically favorable . The active site of ODC was demonstrated to reside at the interface between the two subunits . It is most likely that the monomeric existence of AzI and its inability to bind pyridoxal 5′-phosphate (PLP), the cofactor of ODC , also explain the lack of ornithine-decarboxylating activity as it contains amino acids required for the formation of the active site of ODC. Although it lacks homology with the C-terminal degron of ODC, AzI is a rapidly degraded protein. However, in contrast with ODC and like Az, the degradation of AzI is ubiquitin-dependent . Although Az binds to AzI with high affinity, it does not stimulate its degradation. On the contrary, interaction with Az actually stabilizes AzI by preventing its polyubiquitination . This may suggest that interaction with Az imposes conformational changes that may interfere with its accessibility to the ubiquitination machinery. The structure of the AzI–Az complex suggests that lysine residues that might be ubiquitinated are not shielded by the interacting Az, suggesting therefore that Az might affect the interaction between AzI and the relevant E3 ubiquitin ligase, and not its ubiquitination . An AzI variant, encoded by edited AzI mRNA that displays increased affinity to Az, was suggested to be a potential driver of hepatocellular carcinoma and perhaps of other human cancers . In contrast, a polymorphic splice variant of AzI was suggested to delay fibrosis via a polyamine-independent pathway .
Cellular levels of polyamines are tightly regulated, since they must be maintained at an optimal level. Deviations from this optimal level are deleterious, resulting in extreme cases in growth arrest or cell death. In the center of this regulation is an autoregulatory circuit, which controls ODC degradation and polyamine-uptake activity (Figure 1). At the heart of this circuit is the degradation of ODC that is unique in being manifested by the 26S proteasome in a ubiquitin-independent manner. There are few other proteins that were demonstrated to be degraded without requiring ubiquitination; however, they are also subjected to ubiquitin-dependent degradation, whereas ODC is degraded only via the ubiquitin-independent pathway. Moreover, ODC is the only protein for which an alternative mechanism for proteasomal recognition was described. This raises the question of why there is such an exception. Why is E3 ubiquitin ligase activity not regulated by polyamines? It was proposed previously that ubiquitin-independent recognition by the proteasome may be a historic residue of degradation by the proteasome, a proteolytic machine that evolutionarily preceded ubiquitination, or that such an alternative mechanism provides some independence of the cellular ubiquitin metabolism . However, the observation that Az and AzI, two key regulators of ODC degradation are degraded in a ubiquitin-dependent manner and, at least in yeast, Az's degradation is regulated by polyamines , weakens these possibilities. It is therefore also reasonable to assume that this ubiquitin-independent mechanism was actually selected and evolutionarily maintained. What advantage is attributed by this unique mechanism, requires original thinking and demonstration. Ubiquitinated substrates or free polyubiquitin chains were demonstrated to inhibit ODC degradation, suggesting competition for the recognition elements of the 26S proteasome . The 26S proteasome binds polyubiquitinated substrates either via receptors that are integral part of its structure or by receptor proteins that associate with it transiently . None of these receptors is essential and it is mostly unknown which receptor mediates the recognition of individual ubiquitinated proteins. The recent demonstration that a mutant ODC lacking the C-terminal 37 amino acids that initiates penetration into the proteasome is still capable of binding to the proteasome , may serve as a useful tool for identifying the proteasomal subunit(s) that recognize(s) ODC. In contrast with the ubiquitin-independent degradation of ODC, the degradation of its regulators, Az and AzI, is ubiquitin-dependent. In addition to responding to polyamines at its synthesis, yeast Az is also stabilized by polyamines, contributing further to an increase in its intracellular concentration . It will be of interest to determine whether polyamines also inhibit Az's degradation in other species, and whether AzI is similarly controlled. It will be of interest to determine how polyamines regulate the degradation of these two proteins; do they affect accessibility of relevant lysine residues for ubiquitination, or do they regulate the activity of the relevant ubiquitin ligases? Although ODC is a clear target of Az, it is possible that cells contain additional targets. As Az also regulates transport of polyamines across the plasma membrane, it is possible that a protein or proteins involved in this process are regulated by Az. Identification of such target(s) may also shed light on the process of polyamine transport, a process that, at least in mammalian cells, is mostly unresolved.
Regulation of ODC degradation.
The Author declares that there are no competing interests associated with this manuscript.