| Ornithine decarboxylase: Expression and regulation in rat brain and in transgenic mice | ||
|---|---|---|
| Prev | Chapter 2. Review of the literature | Next |
Ornithine decarboxylase (ODC, EC 4.1.1.17) catalyzes conversion of L-ornithine to putrescine. The enzyme is specific for the L-isoform of ornithine; Km value for the substrate is ≈ 0.09 mM (Pegg & McGill 1979, Seely et al. 1982b, Coleman et al. 1993, Osterman et al. 1994). Mammalian ODC can act also on lysine and very inefficiently on arginine (Pegg & McGill 1979, Osterman et al. 1995a), but these reactions are not of significance for cellular metabolism. All known ODCs require pyridoxal 5´-phosphate (PLP) as a cofactor. The mammalian enzyme binds the cofactor relatively weakly and it can be removed by a rather simple procedure (Poulin et al. 1992). Km value for the PLP is 0.2 - 0.3 µM (Obenrader & Prouty 1977, Seely et al. 1982b).
Active mammalian ODC is homodimer with 2-fold symmetry. Monomers are comprised of two domains. Residues 46-283 form an α/β barrel domain and the remaining carboxy-terminal residues form a β -sheet domain including two separate sheets. Two active sites in the dimer are formed at the interface between the α/β barrel domain of one monomer and the β -sheet domain of the other subunit (Tobias & Kahana 1993, Osterman et al. 1995a, Osterman et al. 1995b, Kern et al. 1999, Almrud et al. 2000). Subunits have molecular weight of about 51 kDa and the polypeptide chain consists of 461 amino acids, an exception being hamster ODC that has 455 amino acids (Grens et al. 1989, Yao et al. 1995). Several of those amino acids have been shown to be essential or of importance for catalytic activity. Lysine-69 binds PLP through a Schiff base (Poulin et al. 1992) and may play a role in the proper positioning of the substrate for efficient decarboxylation (Osterman et al. 1999). The irreversible inhibitor DFMO forms covalent adducts mainly with Cys-360 and to some extent with Lys-69 (Poulin et al. 1992) indicating that these amino acid residues are located at or very close to the active site. Cys-360 may function as a proton donor in catalysis (Kern et al. 1999, Jackson et al. 2000), other potential proton donors are Lys-69 and more unlikely His-197. Cys-360 is needed for the reaction specifity. When it is mutated catalytic rate is diminished and the major reaction is a decarboxylation-dependent transamination resulting in the formation of pyridoxamine 5’-phosphate (PMP) and γ -aminobutyraldehyde instead of PLP and putrescine (Jackson et al. 2000). This may be due to the protonation of C4’-carbon in PLP instead of Cα-carbon in ornithine. Gly-387 is essential for the formation of dimer (Tobias et al. 1993), but the reason for this is not clear (Kern et al. 1999). Enzyme activity is decreased when acidic Asp-88, Glu-94, Asp-233 or most significantly, Glu-274 are mutated (Osterman et al. 1995a). Glu-274 interacts with N1-nitrogen of PLP and forms an acidic cluster with Asp-88 and Asp-233 that, with three bound water molecules form a network of hydrogen bonds that probably influences the electron-withdrawing properties of the cofactor (Kern et al. 1999). Two other residues shown to have an effect on catalytic activity are Lys-115 and Lys-169 (Lu et al. 1991, Tsirka & Coffino 1992). The latter is involved in a salt-bridge stabilizing dimer (Kern et al. 1999), but no role for the former is known.
The catalytic mechanism of ODC is typical to all PLP-dependent decarboxylases (Kern et al. 1999, Osterman et al. 1999, Jackson et al. 2000). In the absence of substrate a holoenzyme contains an internal aldimine where the PLP is bound to the active-site lysine-69 by a Schiff-base linkage. The ornithine substrate reacts at first with the cofactor via transaldimination reaction to form an external aldimine. This is followed by the release of CO2 and the formation of a quinonoid intermediate that is protonated to form again external aldimine, now consisting of putrescine and PLP bound to each others via a Schiff-base linkage. Putrescine is then dissolved from the active site and PLP forms a Schiff base with Lys-69. PLP-dependent enzymes achieve reaction specifity by positioning specific residues or molecules (like Cys-360 in ODC) that interact with each groups surrounding the Cα carbon of the substrate in a geometry that favours a particular bond cleavage (Jansonius 1998). Otherwise, any of three bonds (the fourth one is in a Schiff base) around the Cα carbon might be cleaved, enabling a broad range of reactions including transamination, racemization, retro-aldo cleavage and deamination in addition to decarboxylation. In decarboxylases transaldimination leads to the orientation of the α carboxylate perpendicular to the pyridine ring of PLP leading to cleavage of the bond between the Cα carbon and carboxylate.
Complete structure and nucleotide sequence of ornithine decarboxylase gene from mammalians is known for mouse (Coffino & Chen 1988, Katz & Kahana 1988), rat (van Steeg et al. 1988, Wen et al. 1989, van Steeg et al. 1990), human (Fitzgerald & Flanagan 1989, van Steeg et al. 1989, Hickok et al. 1990, Moshier et al. 1990), and bovine (Yao et al. 1998). In addition, the complete amino acid sequence (Srinivasan et al. 1987, Grens et al. 1989) and nucleotide sequence of the promotor region (GenBank accession number X53906) is known for hamster ODC. All mammalian ODC genes have 12 exons and 11 introns. Location of exon-intron boundaries is identical in all ODC genes. The transcription unit is relatively short, 6 – 8 kb depending on species. The first intron is considerably longer than others, 2.0 – 3.0 kb, and contains potential or demonstrated regulatory elements. These elements as well as those upstream in the promoter region are discussed in detail in the next chapter. Exons range from 85 bp to about 900 bp in length.
Mammals have several ODC gene-like sequences in their genomes, but apparently only one functional gene, others being pseudogenes. Active mouse ODC gene has been localized to chromosome 12 (Cox et al. 1988), functional hamster ODC gene is in the chromome 7 (Tonin et al. 1987), human gene in the chromosome 2 (Winqvist et al. 1986, Hsieh et al. 1990, Radford et al. 1990) and rat gene in the chromosome 6 (Deng et al. 1994).
Northern analysis has revealed that rodents express two species of ODC mRNA (Berger et al. 1984, Kontula et al. 1984, Gilmour et al. 1985, van Kranen et al. 1987) whereas in human (Hickok et al. 1987, Radford et al. 1990) and bovine (Yao et al. 1995) only single ODC mRNA has been detected. The longer rodent mRNA is 2.6 – 2.7 kb in length, ant the shorter 2.1 – 2.2 kb. They are the result of the use of two separate polyadenylation signals present at the 3’ untranslated region. Human ODC mRNA has two polyadenylation signals as well, but apparently only that giving the shorter transcript is used.
The open reading frame encoding mammalian ODC is 1383 nucleotides long the only exception being hamster ODC mRNA with the coding region of 1365 nucleotides. The coding region is highly conserved. Mouse and rat amino acid sequences differ only at 14 sites from each others. Hamster and bovine ODCs are the most divergent mammalian ODCs. They have 56 differences in amino acid sequences. The 5’-untranslated region of ODC mRNA is exceptionally long. Hamster has the ODC 5’ leader sequence slightly shorter than 300 nucleotides, those of other mammals exceed 300 nucleotides in length. The leader sequence is GC-rich and has been suggested to form secondary structures with a high free energy and stability (Brabant et al. 1988, Katz & Kahana 1988, Wen et al. 1989). The region also contains a small open reading frame as does AdoMetDC mRNA. This ORF may have a role in the regulation of translation, although the nucleotides flanking the translational start site do not conform perfectly well to the consensus sequence for the translation initiation (Kozak 1989).
ODC is one of the most highly regulated enzymes in eukaryotic organisms. The enzyme activity is induced rapidly up to several hundred fold by a great variety of factors stimulating cell growth and proliferation. These factors can be such as hormones, tumor promoters and growth factors. In almost all cases the increase in activity is accompanied by roughly equivalent changes in the amount of enzyme protein, and thus ODC appears not to be generally regulated by post-translational modifications or by allosteric effectors. The accumulation of ODC protein is controlled in gene transcription, mRNA translation and enzyme degradation. In addition, ODC activity is specifically inhibited by antizyme protein before degradation and negatively feed-back regulated by polyamines. (reviewed in (Davis et al. 1992, Shantz & Pegg 1999)
Some increase in the amount of ODC mRNA is detected in virtually all cases when ODC activity is stimulated. This may mostly be due to increased transcription, but in several cases stabilization of ODC mRNA has also been detected (Laitinen et al. 1984, Berger & Porter 1986, Hölttä et al. 1988, Chen & Chen 1992). Induction of ODC activity by growth factors is maybe the most apparent case where induction is mainly a result of elevated ODC mRNA levels (Feinstein et al. 1985, Greenberg et al. 1985). Also application of phorbol ester and tumor promotor 12-O-tetradecanoylphorbol 13-acetate (TPA) leads to rapid increase in ODC mRNA followed slightly later (Gilmour et al. 1985, Verma et al. 1986a, Verma 1988) or even simultaneously (Verma et al. 1986b, Hsieh & Verma 1989) by increase in ODC activity.
Signaling pathways leading to ODC induction are only partly known. Protein kinase C and phospholipase C have been suggested to be involved in mediating TPA stimulated induction of ODC (Verma et al. 1986b). Protein kinase C activity is required also for ODC induction by oxidative stress (Otieno & Kensler 2000). Activation of protein kinase A has been shown to lead to a rapid elevation of ODC gene transcription (Abrahamsen et al. 1992). In human endothelial cell line ECV304, p44/42 MAPK pathway was required for induction of ODC by any stimulus employed, i.e. serum, histamine and ATP (Flamigni et al. 2001).
Whatever the nature of transcription inducing stimulus, its effect has to be mediated to ODC promotor and responsive elements. The ODC promotor region is typical to ubiquitously expressed genes. It contains TATA and CAAT boxes, although the latter is relatively poorly conserved and several GC-rich areas that are potential binding sites for the members of Sp/Krüppel-like transcription factor family regulating expression of various house-keeping genes, but also tissue-specific and growth- or differentiation related expression of target genes (Black et al. 2001). From this family, Sp1 (Li et al. 1994, al-Asadi et al. 1995, Kumar et al. 1995) and inhibitory Sp3 (Kumar & Butler 1997) as well as two Krüppel-like factors ZBP-89 (Law et al. 1998) and ZBP-99 (Law et al. 1999) have been shown to bind to the ODC promotor and to regulate expression of reporter or endogenous ODC gene. Other transcription factors demonstrated to bind and regulate ODC promotor are c-Myc/Max dimers (Bello-Fernandez et al. 1993, Pena et al. 1993, Tobias et al. 1995, Walhout et al. 1997), and WT1 tumor suppressor (Moshier et al. 1996, Li et al. 1999a). The ODC promotor has even been used in viral gene transfer vectors to achieve c-Myc- and N-Myc-regulated protein expression in tumor cells (Pawlik et al. 2000, Iyengar et al. 2001). Also cAMP-responsive element mediates regulation of the ODC promotor and binds proteins from nuclear extracts, but the identity of these proteins is not clear and there are contradictory reports on involvement of CREB-binding protein family (Palvimo et al. 1991, Abrahamsen et al. 1992, Palvimo et al. 1996). In addition, a negative regulative element has been localized to the ODC promotor (Zhao et al. 2000). This element is a putative Ets-binding site, but a protein of 55-60 kDa specifically bound to it was not antigenically related to c-Ets-1. Members of Ets family mediate responsiveness to extracellular signals, including mitogens and phorbol esters (Wasylyk et al. 1998). It is known that phorbol esters do not regulate the ODC promoter via AP-1 transcription factor as they do to many promotors (Kim et al. 1994, Mar et al. 1995), but regulation may be mediated at least partially by the general transcription initiation complex organising at the TATA box (Reddig et al. 1996). And finally, although the ODC promotor does not contain a consensus androgen response element, androgen receptor has been suggested in one study to bind on the promotor region (Bai et al. 1998).
The many cases cellular increase in ODC activity and protein content can not be explained merely by the increased transcription or stabilization of ODC mRNA, and not by stabilization of ODC protein, indicating that regulation may take place also at translational level. Translational regulation appears to have relatively the most significant role when ODC activity is induced by amino acids (Kanamoto et al. 1987, Chen & Chen 1991, Chabanon et al. 2000), insulin (Blackshear et al. 1987, Manzella et al. 1991) or, at least in some experimental systems, by hypotonic shock (Poulin & Pegg 1990, Lövkvist- Wallström et al. 1995, Lövkvist-Wallström et al. 2001). When it comes to negative regulation, inhibition of ODC mRNA translation by polyamines has been observed in several studies (e.g. Kameji & Pegg 1987b, Persson et al. 1988, Kanamoto et al. 1991, Lövkvist et al. 1993).
In the study of translational regulation of ODC, much of the work has focused on the long 5’-untranslated region (5’UTR) of its mRNA. The 5’UTR is composed of two distinct segments (Brabant et al. 1988, Katz & Kahana 1988, Wen et al. 1989). There is a 5’ proximal GC-rich segment, spanning the first 180 nucleotides, which is predicted to form a very stable hairpin structure and which contains a small upstream open reading frame close to it 3’ end. The second segment covers the rest of 5’UTR and is relatively unstructured. Although most genes encode mRNAs with short unstructured 5’UTRs, oncogenes and genes involved in cellular proliferation often encode mRNAs with longer 5’UTRs that may form secondary structures (Kozak 1989, Gray & Henze 1994). In rabbit reticulocyte lysates, translation of both mouse and rat ODC containing the full length 5’UTR was reduced by 95% compared to mRNA containing very short 5’ leader sequence (Ito et al. 1990, Manzella & Blackshear 1990, Kashiwagi et al. 1991, Van Steeg et al. 1991). The conserved GC-rich region in the 5’ end of 5’UTR repressed translation to the same extent as entire 5’UTR (Manzella & Blackshear 1990, Van Steeg et al. 1991). Similarly, in cultured cells, the expression of reporter genes was inhibited by up to 99% when the full length ODC 5’UTR was inserted immediately before the initiation codon (Grens & Scheffler 1990, Manzella & Blackshear 1990, Shantz et al. 1994). The GC-rich 5’ end of the leader sequence was again almost as effective as the full length 5’UTR in suppressing protein synthesis (Grens & Scheffler 1990, Manzella & Blackshear 1990, Shantz et al. 1996b).
The 3’UTR of ODC mRNA is also relatively long (300 nucleotides), but has less potential to form stable secondary structures than 5’UTR. However, it has been demonstrated in various expression systems that the 3’UTR may interact with the 5’UTR of ODC mRNA in such a way that the repressive effect of the 5’UTR on translation is relieved (Grens & Scheffler 1990, Lorenzini & Scheffler 1997). Interestingly, according to a recent study, the hypotonic induction of ODC is highly dependent on the presence of 3’UTR, but not on the presence of 5’UTR in ODC mRNA (Lövkvist-Wallström et al. 2001). In the same study it was pointed out that ODC 3’UTR contains region corresponding to AU-rich elements (AREs). These elements are located in the 3’UTR of a number of growth-related mRNAs, including mRNAs coding for a number of protooncogenes and cytokines (Chen & Shyu 1995).
All mammalian ODCs have a small internal open reading frame in their 5’UTR. The sequence around the AUG codon lacks the –3 purine but does contain a G residue in position +4 thought to be necessary for efficient translation (Kozak 1991, Kozak 1992). The predicted peptide sequence is not as conserved as that of the AdoMetDC internal ORF, but contains 10 amino acids in the most mammalian ODC mRNAs. The internal ORF is located approximately 150 nucleotides from the 5’ end of mRNA. Its translation has not been demonstrated, but the mutation of its initiation codon appears to increase the translation of major reading frame of reporter construct (Grens & Scheffler 1990, Manzella & Blackshear 1990, Shantz et al. 1996b). In contrast, no difference in translational efficiency between mRNAs containing wild-type or mutated AUG in 5’UTR was detected in in vitro translation reactions (Van Steeg et al. 1991). This may just indicate that some cellular factor responsible for the release of translation inhibition is limiting in reticulocyte lysates.
Eukaryotic initiation factor 4F (eIF4F) mediates cap-dependent translation. It is thought to be able to unwind secondary structures in the 5’UTRs, thereby facilitating ribosome binding to the 5’ end of mRNA (Gingras et al. 1999, Pestova et al. 2001). The recruitment of ribosome to mRNA is the rate-limiting step of translation under most circumstances and a primary target for translational control. One subunit of eIF4F is eIF4E. It functions directly in the recognition of the mRNA 5’ cap structure. A cell line overexpressing eIF4E is transformed (Lazaris-Karatzas et al. 1990) and ODC activity is increased by over 30-fold compared to wild-type cell line (Shantz & Pegg 1994, Shantz et al. 1996b). DFMO (Shantz & Pegg 1994) and expression of dominant negative ODC mutant (Shantz et al. 1996a) are able to partly revert the transformed phenotype, but the formation of phenotype is likely contributed also by the enchanced expression of various oncogenes. Overexpression of eIF4E has been suggested to promote translation of all mRNAs that contain a long and structured 5’UTR (Lazaris-Karatzas et al. 1990). For example c-sis, c-lck and c-myc proto-oncogenes have this kind of mRNAs. Very interestingly, it has been reported that ODC mRNA is translated also using capindependent mechanism (Pyronnet et al. 2000). In the a cap-independent translation ribosome binds to an internal ribosome entry sire (IRES) in 5’UTR. This mechanism appears to be functioning in ODC mRNA translation during the G2/M transition of cell cycle maybe providing sufficient levels of polyamines before mitosis.
Low levels of polyamines are necessary for general protein synthesis but excessive polyamine levels inhibit the translation of most mRNAs. Translation of ODC mRNA is both stimulated and inhibited at lower concentrations of polyamines than translation generally (Kameji & Pegg 1987b, Persson et al. 1988, Ito et al. 1990). The reducing effect of increased polyamine content on ODC mRNA translation has been observed both in reticulocyte lysates (Kameji & Pegg 1987b, Persson et al. 1988, Ito et al. 1990) and in cells in culture (Kanamoto et al. 1991, Kameji et al. 1993, Lövkvist et al. 1993). Removing or truncating the 5’UTR from ODC mRNA abolishes the polyamine effect according to several studies (Ito et al. 1990, Kashiwagi et al. 1991, Lövkvist et al. 1993). However, the exact location of any polyamine responsive element in the 5’UTR has yet to be defined. Furthermore, two studies have failed to detect any effects of polyamines on translation of ODC mRNA in vitro (Van Steeg et al. 1991) or reporter construct containing the ODC 5’UTR in vivo (Grens & Scheffler 1990). In addition, it has been reported that the polyamine mediated regulation of ODC expression is independent on 5’UTR (van Daalen Wetters et al. 1989b) or both 5’ and 3’UTR (Lövkvist-Wallström et al. 2001). It has been speculated that in these cases polyamines may induce rapid cotranslational degradation of ODC. Taken together, it is likely that the regulation of ODC has cell specific characteristics, and that intracellular levels of other factors combine with changes in polyamines to influence translation.
ODC induction is nearly always accompanied by a similar increase in the amount of ODC protein. According to the general concept, post-translational modification and allosteric effectors do not have a significant role in ODC regulation. However, these might be of importance in specific tissues or physiological conditions. For example, in several tumor lysates ODC activity can be stimulated by GTP suggesting either allosteric regulation or GTP-dependent post-translational modification or involvement of GTP-binding regulatory protein (O"Brien et al. 1986, O"Brien et al. 1987, Hietala et al. 1988, Hietala et al. 1990). So far the only demonstrated post-translational modification of mammalian ODC is phosphorylation (Rosenberg-Hasson et al. 1991, Worth et al. 1994).
The first step in recognition of ODC as a phosphoprotein was successful in vitro phosphorylation by casein kinase II (Meggio et al. 1984). The phosphorylation site was later identified to be Ser-303 (Rosenberg-Hasson et al. 1991). Phosphorylation of Ser-303 in COS-cells or in vitro did not have an effect on enzyme activity or stability (Meggio et al. 1984, Rosenberg-Hasson et al. 1991, Kanamoto et al. 1993). Ser-303 was the only phoshorylation site detected in COS cells, but in a transformed macrophage cell line RAW264 ODC was phosphorylated at, at least two serines and two threonines (Worth et al. 1994, Reddy et al. 1996). Another of serines was at the position 303. Purified phosphorylated ODC was more stable than unphosphorylated and its Vmax was 1.5-fold higher (Reddy et al. 1996).
In human keratinocytes phosphorylated ODC has been observed to be preferentially associated with insoluble cellular proteins (Pomidor et al. 1999). The reason for this localization is not known. A fraction of ODC protein has been reported to translocate to the cell surface membrane during cell activation and tranformation (Heiskala et al. 1999). This translocation was dependent on p47phox-related membrane targeting sequence (Nauseef et al. 1993) comprising amino acids 165-172 in ODC. When Ser-167 of ODC was mutated to alanine, the mutant ODC was unable to move to the cell surface. It is not clear why ODC translocates to the membrane and it is unclear whether the Ser-167 is actually phosphorylated. In the p47pbox protein phosphorylation of the corresponding serine supports translocation (Huang & Kleinberg 1999), but is not strictly essential (DeLeo et al. 1995). In any case, the function of postranslational modifications in ODC are likely to be like this, to be needed for some specific property in some specific situation.
ODC is a very labile protein. The half-life of enzyme activity is one of the shortest known for mammalian enzymes, it can be as short as 10 – 20 minutes (Seely et al. 1982a, Isomaa et al. 1983) and at its longest only from one hour to two hours (Hayashi et al. 1996). The short half-life of activity is due to rapid degradation of enzyme protein, because the halflife of immunoreactive ODC is only slightly longer than that of enzyme activity (Isomaa et al. 1983, Laitinen et al. 1984). Rapid turnover is essential for sensitive regulation and rapid induction of enzyme activity. When general protein synthesis in cells is increased, the amount of labile proteins increases much more rapidly than that of stable proteins (Tabor & Tabor 1984).
ODC is degraded via an exceptional pathway (reviewed in Hayashi 1995, Murakami et al. 2000, Coffino 2001b). Inhibitor protein antizyme, synthesis of which is greatly induced by polyamines (Fong et al. 1976, Heller et al. 1976), binds to the ODC monomer, inhibits its activity and targets the protein for rapid degradation by the 26S proteasome complex (Murakami et al. 1992a). Proteasome 26S is the main neutral protease of the cell and it was for a long time thought to degrade only proteins tagged with the poly-ubiquitin chain (Verma & Deshaies 2000). ODC was the first exception to this rule and to date only the cyclin-dependent kinase inhibitor p21cip1 (Sheaff et al. 2000) and the NS2 protein of parvovirus minute virus (Miller & Pintel 2001) have been found to be degraded by the 26S proteasome in ubiquitin-independent manner. Immunoremoval of the proteasome from cell extracts or use of a proteasome inhibitor clasto-lactacystin β -lactone in cell cultures almost completely inhibited the degradation of ODC showing that the proteasome 26S pathway is in deed the very major pathway for ODC degradation (Murakami et al. 1999).
The 26S proteasome is a multisubunit complex, consisting of a central proteinase called the 20S proteasome, and two terminal regulatory subcomplexes, termed PA700 (also called 19S complex) and PA28 (11S regulator) (Murakami et al. 2000). The 20S unit is a 700 kDa cylinder-shaped particle having multiple catalytic centers located within a hollow cavity of the cylinder. The 26S complex contains one regulatory subunit at its ends, these can both be PA700s or PA28s or also heterocomplexes exist containing both PA700 and PA28. However, homo-PA28 proteasome is incapable of degrading ODC (Tanahashi et al. 2000). Other variants of the 26S, but not 20S, proteasome degrade ODC in vitro in the presence of ATP and antizyme. Most likely, degradable proteins are at first recognised by the regulatory subunits of the proteasome complex and then ulfolded and fed to inner cavities of the 20S core particle (Hershko & Ciechanover 1998). The 26S proteasome irreversibly inactivates ODC prior to its degradation (Murakami et al. 1999). The inactivation, possibly due to unfolding, is coupled to sequestration of ODC within the 26S proteasome. This process requires antizyme and ATP, but not proteolytic activity of the proteasome. Antizyme is generally recycled and is only seldomly degraded (Tokunaga et al. 1994). The purified 26S proteasome can slowly degrade ODC even in the absence of antizyme, but the physiological significance of this degradation is unclear (Murakami et al. 2000).
Mutagenesis and structural studies have revealed the elements of ODC that are responsible for the interaction with antizyme and for ODC degradation. The carboxy- terminal region encompassing amino acids 423 – 462 has been shown to be important for both unstimulated and polyamine-stimulated degradation of ODC, although other areas also may be involved (Li & Coffino 1992, Li & Coffino 1993, Mamroud-Kidron et al. 1994). PEST regions, i.e. areas rich in proline, glutamate, serine and threonine, are typical to many rapidly degraded proteins (Rogers et al. 1986). The carboxy-terminal region of ODC contains one of the two PEST regions, but removal of the last five amino acids outside of the PEST region (Ghoda et al. 1992) or a single amino acid exchange in the PEST region (Cys441 to TRp) (Miyazaki et al. 1993) also stabilizes ODC, suggesting that this PEST sequence is not the only determinant of ODC instability. Another internal PEST region (amino acids 298 – 333) does not appear to be associated with ODC instability (Ghoda et al. 1992). In addition, an internal region of ODC (amino acids 117 – 140) is required for antizyme-binding and thus for antizyme-dependent degradation (Li & Coffino 1992). This region is highly conserved in mammals, avians, and amphibians whose ODC are rapidly degraded by an antizyme-dependent process (Hayashi et al. 1996).
The most likely interpretation of the observations presented above is that the putative “degradation signal” of ODC is located at the carboxy-terminal region which is normally masked, but exposed by attachment of antizyme and finally recognised in some way by the 26S proteasome (Li & Coffino 1993). On the other hand, according to the crystal structure of the human full-length ODC this model may be too simplified (Almrud et al. 2000). It appears that the carboxy-terminus from residue 401 is not buried by the structural core of homodimeric ODC. However, the binding of antizyme to the ODC monomer may yet induce other conformational changes to make ODC more susceptible to proteolysis or make the carboxy-terminus accessible still further to aminoterminus.
ODC degradation is regulated equally strictly as synthesis and provides another means to control polyamine levels in cells. There are several observations about increased stability of ODC during induction of enzyme activity. Androgen elevates ODC activity in kidneys dramatically and this is partly explained by the 4 – 10 –fold increase in half-life of ODC (Seely et al. 1982a, Isomaa et al. 1983) whereas hypotonic shock results in a 3 – 6 fold increase in half-life (Poulin & Pegg 1990, Tohyama et al. 1991). The most important factors enhancing ODC degradation are polyamines (Glass & Gerner 1986, Hölttä & Pohjanpelto 1986, Kanamoto et al. 1986) that, as mentioned earlier, induce synthesis of antizyme (Fong et al. 1976, Heller et al. 1976).
Mammalian ODC was thought to be lacking any allosteric effectors untill it was obseved that the enzyme in mammalian epidermal papillomas can be activated by GTP (O"Brien et al. 1986). Activation was detected in raw cell lysates and could be a result of a real allosteric interaction or post-translational modification requiring GTP, or involvement of GTP-binding regulatory protein. Later GTP activation has been detected in human squamous cell carcinoma (Hietala et al. 1988), colorectal adenocarcinoma (Hietala et al. 1990), gastric cancer (Okuzumi et al. 1991) and colorectal carcinoma (Matsubara et al. 1995). It was possible to separate GTP-activatable ODC from non-activatable enzyme by gel-filtration chromatography (O"Brien et al. 1987) suggesting co-purifying regulatory or modifying protein(s) or allosteric interaction that has considerable effect on the conformation of enzyme. GTP-activatable ODC displays a higher Km value than “normal” ODC and GTP activates the enzyme by decreasing Km close to or below normal level. In some cases GTP effects also Vmax or, in some rare cases, GTP activates ODC effecting only Vmax. GTP-activatable enzyme was more heat-stabile and more resistant to DFMO inhibition. It has been suggested that GTP-activatable ODC could be advantageous to tumor cell growth (Hietala et al. 1988), because it appears to be more stabile and its activity may be more readily deregulated. However, according to experimental data, it seems that the detection of GTP-activatable ODC in tumors may indicate more favourable patient prognosis (Matsubara et al. 1995).