The present study was the first to report activation of mammalian ODC in normal tissues by GTP or by any other potential allosteric effector. Some prokaryotic ODCs such as biosynthetic ODC of E. coli (Applebaum et al. 1977, Anagnostopoulos & Kyriakidis 1996) and Lactobacillus 30a ODC (Oliveira et al. 1997), and ODCs from some lower eukaryotes such as S. cerevisiae (Tyagi et al. 1981) are activated allosterically by GTP. GTP-activatable ODC has been detected also in tumor cell lysates prepared from mouse epidermal papillomas (O"Brien et al. 1986, O"Brien et al. 1987), human squamous cell carcinoma (Hietala et al. 1988), colorectal adenocarcinoma (Hietala et al. 1990), gastric cancer (Okuzumi et al. 1991) or colorectal carcinoma (Matsubara et al. 1995). In these cases, as well as in the present work, it is not known whether GTP is a real allosteric effector of ODC or if it activates ODC indirectly via other regulatory protein(s).
The ability of GTP to activate ODC varied in lysates prepared from different parts of brain, but the activation was always higher when antizyme was removed from the lysate by gel-filtration chromatography. The activation was highest in the cerebellum lysates both before and after gel-filtration (2- and 4.2-fold). ODC from the cortex and hippocampus displayed lowest activation, but in the latter the total enzyme activity was highest. In ammonium sulphate precipitated tumour lysates GTP activation was at highest 8-fold, but generally less than 2-fold (Hietala et al. 1988, Hietala et al. 1990).
The heat sensitivity of brain ODC was comparable with that of mouse epidermal tumour GTP-activatable ODC, which is more stable than the normal epidermal ODC (O"Brien et al. 1986, O"Brien et al. 1987) or kidney ODC that was used as a control in this study. However, GTP has a substantial protective effect against heat inactivation of the GTP-activatable tumour ODC, but this effect was not observed with the brain enzyme which does not necessarily indicate differences between brain and tumour ODC, but may imply differences in the utilization or regulation of degradative machinery in different tissue lysates.
Tumour ODC was activated only by GTP and dGTP (O"Brien et al. 1986, O"Brien et al. 1987) whereas brain ODC could be activated by both purine nucleotide triphoshates, and their deoxyforms. GTP was highly favoured in activation over ATP, the K1/2 value for GTP was 2 µM and for ATP 40 µM. Tumour ODC has a lower K1/2 for GTP, 0.1 µM, although the order of magnitude was similar. These GTP concentrations are significantly lower than typical GTP concentrations in cells which range from 50 µM to 200 - 300 µM (Otero 1990, Jinnah et al. 1993). In that case ODC may always be maximally activated by GTP in cells if activation is not prevented e.g. by antizyme. We detected activation clearly in all brain lysates when antizyme was dissociated from ODC and removed by gel-filtration chromatography. Agreeing with previous data, it appeared that GTP-activatable ODC is inhibited more readily by antizyme. On the other hand, ammonium sulphate precipitation and gel-filtration chromatography decreased GTP concentration in the lysates and might have made the activation possible, i.e. most of the enzyme may have been “activated” in intact cells and raw lysates, and this activation could have been at first reversed by the removal of GTP, which would have made the enzyme again more clearly GTP-activatable. However, GTP-activation was apparently irreversible, since even after extensive dialysis activity in a sample preincubated with GTP was higher than in a sample preincubated without GTP. These samples were not activated by GTP after dialysis and the possibility remains that GTP only stabilized the enzyme in preincubation, although this may not be likely because there was no stabilizing effect by GTP on brain ODC when heat stability was tested. Furthermore, also tumour enzyme has been shown to be irreversible activated by GTP and in this case the enzyme sample preincubated without GTP was activated after dialysis (O"Brien et al. 1987). GTP activated brain ODC mainly by increasing the Vmax value indicating enhanced catalytic efficiency. In contrast, GTP effected mainly Km values of tumour ODC (O"Brien et al. 1986, O"Brien et al. 1987, Hietala et al. 1988).
There are several explanations how GTP-activatable ODC is brought about. At the very beginning of this work it was known that human and mouse genomes contain several ODC-like sequences and it was considered whether a normally silent ODC gene could be expressed in tumour cells. This was later shown to be virtually impossible. The human genome contains two highly ODC–like loci and another of them is a pseudogene, that can not be activated by simple mutation (Hickok et al. 1990, Radford et al. 1990). Recently cloned ODC-like protein is moderately identical to ODC (54 %) and might even posses some ODC activity (Pitkänen et al. 2001). However, it has a valine residue in a position corresponding to Cys-360, which means that its ODC-like catalytic efficiency would almost certainly be greatly diminished from that of ODC (Coleman et al. 1993), and that catalysis would probably not be specific for the L-ornithine or for decarboxylation (Jackson et al. 2000). Nevertheless, lack of Cys-360 may do ODC activity resistant to DFMO, because Cys-360 is the major binding residue of DFMO (Poulin et al. 1992). This could explain the low residual ODC activity we detected in the brain lysates after DFMO inhibition. In tumour cells ability to be activated by GTP could be a consequence of mutation in ODC gene, but GTP-activatable ODC in brain can not be explained by mutation and it is not plausible that two basicly different mechanisms could results in the formation of GTP-activatable ODC. It seems more apparent that GTP activation is a result of tissue specific posttranslational modification or tissue specific function of regularoty protein(s).
In principle, GTP could bind to ODC directly affecting its catalytic function, or GTP could activate a GTP-binding protein which mediates the activation of ODC, or it could be a substrate for a kinase that phosphorylates ODC. The first alternative is not likely. Structures and amino acid sequences of GTP-binding sites from several proteins are well known (Geyer & Wittinghofer 1997, Takai et al. 2001). ODC does not contain obvious putative binding sites for GTP, and although point mutations could create areas resembling the binding sites, they can not explain the presence of GTP-activatable enzyme in the brain as noted above. It is not plausible either that a posttranslational modification could create a binding site. The alternative suggesting that GTP acts as a substrate for a kinase appears to be more realistic. There are kinases that accept both ATP and GTP as a substrate (Richert et al. 1979, Payne & Dahmus 1993), and also kinases that accept deoxynucleotides as a substrate (Levy-Favatier et al. 1987, Payne & Dahmus 1993). The kinase using GTP as a substrate need not phosphorylate ODC directly. It could be involved in a signaling cascade leading finally to the phosphorylation of ODC. On the other hand, the non-hydrolyzable GTP analog GTP[γ -S] can also activate tumour ODC (O"Brien et al. 1987) although it certainly can not be used as a substrate for a kinase. The effect of GTP[γ -S] on brain ODC was not tested in this work. The third and also a very plausible, alternative is that GTP activates at first GTP-binding regulatory protein that regulates ODC directly or activates a signaling pathway leading to ODC phosphorylation.
Properties of GTP-activatable brain ODC agree well with those reported for the phosphorylated ODC from macrophage cell line (Reddy et al. 1996). The phosphorylated enzyme was more stable and displayed a higher Vmax value. Phosphorylation would also explain irreversibility, or reversibility in some cases, depending whether phosphatases and kinases are present. e.g. in the brain, fractions of ODC could be unphosphorylated and bound to antizyme and it could be phosphorylated when cells are lysed and antizyme removed. Alternatively, cell lysis could expose ODC to phosphatases that dephosphorylate the enzyme, which then again could be phosphorylated and activated. This scheme requires that kinase co-purify with ODC in the gel-filtration chromatography. On the other hand, if GTP activation is a result of direct interaction with regulatory protein, the regulatory protein should as well co-purify with ODC. It should be pointed out, that the different properties of tumour and brain ODC are understandable and to a certain extent expected. The enzyme may be phosphorylated at different residues or could be regulated by different regulatory proteins in the brain and tumors which have very different characteristics to tissues. The phosphorylation pattern in the tumours may even be aberrant.
Why is there GTP-activatable ODC in the brain and in some tumours? CNF is a particular tissue when it comes to polyamine metabolism. Polyamines have there specific functions in the regulation of ion channels and glutamine-activated receptor channels (Haghighi & Cooper 1998, Oliver et al. 2000). Polyamines have an essential role in the development and differentiation of CNS and their metabolism is altered in various pathological conditions (Bernstein & Müller 1999). Antizyme concentration in CNS is higher than elsewhere (Laitinen et al. 1985, Gritli-Linde et al. 2001). Brain is together with testis the tissues where transgenic mice overexpressing ODC have the most significant changes in putrescine and polyamine levels (Halmekytö et al. 1991a, Halmekytö et al. 1993). GTP-activatable ODC could be used to modulate some of the CNS-specific functions of polyamines or it could be important during restricted time in the development of neurons. In tumours ODC regulation may be working aberrantly and in the corresponding non-malignant cells GTP-activatable ODC could most likely exists, if it existed at all in this cell type, only during development and differentiation, since tumour cells are generally poorly differentiated.
Clarifying of the origin and function of GTP-activatable ODC would be aided by the purification of the enzyme. If the ability to be activated by GTP is lost during purification, it is likely that effector proteins or kinases are required. Potential roles of phosphorylation in ODC activation could be studied using [γ -32P]-GTP and/or [γ -32P]-ATP and immunoprecipitating ODC from labelling reactions. One way to address these questions could be to study whether recombinant ODC is activated in brain lysate, which even could be depleted from endogenous ODC activity. If exogenous ODC were activated, site-directed mutagenesis could be employed to elucidate the mechanism of GTP-activation. Dominant negative mutant forms of signaling proteins could be tested for the ability to prevent GTP-activation. An example of signaling proteins that might be involved in the phosphorylation and/or GTP-activation of ODC are members of MAP kinase pathway. MAP kinases themselves phosphorylate serine and threonine residues in many growth-related – but also in other type of – proteins (Pearson et al. 2001).