Ornithine Decarboxylase from Neurospora crassa PURIFICATION, CHARACTERIZATION, AND REGULATION BY INACTIVATION*

Ornithine decarboxylase, a highly regulated enzyme of the polyamine pathway, was purified 670-fold from mycelia of Neurospora crassa that were highly augmented for enzyme activity. The enzyme is significantly different from those reported from three other lower eucaryotic organisms: Saccharomyces cerevisiae, Physarum polycephalum, and Tetrahymena pyriformis. Instead, the enzyme closely resembles the enzymes from mammals. The Mr = 110,000 enzyme is a dimer of 53,000 Da subunits, with a specific activity of 2,610 mumol per h per mg of protein. Antisera were raised to the purified enzyme and were rendered highly specific by cross-absorption with extracts of a mutant strain lacking ornithine decarboxylase protein. With the antisera, we show that the inactivation of the enzyme in response to polyamines is proportional to the loss of ornithine decarboxylase protein over almost 2 orders of magnitude. This is similar to the inactivation process in certain mammalian tissues, and different from the process in S. cerevisiae and P. polycephalum, in which enzyme modification, without proportional loss of antigen, accompanies enzyme inactivation. The N. crassa enzyme is therefore suitable as a microbial model for studies of the molecular regulation of the mammalian enzyme.

Ornithine decarboxylase, a highly regulated enzyme of the polyamine pathway, was purified 670-fold from mycelia of Neurospora crassa that were highly augmented for enzyme activity.
The enzyme is significantly different from those reported from three other lower eucaryotic organisms: Saccharomyces cerevisiae, Physarum polycephalum, and Tetrahymena pyriformis.
Instead, the enzyme closely resembles the enzyes from mammals.
The M, = 110,000 enzyme is a dimer of 53,000 Da subunits, with a specific activity of 2,610 pmol per h per mg of protein.
Antisera were raised to the purified enzyme and were rendered highly specific by cross-absorption with extracts of a mutant strain lacking ornithine decarboxylase protein.
With the antisera, we show that the inactivation of the enzyme in response to polyamines is proportional to the loss of ornithine decarboxylase protein over almost 2 orders of magnitude. This is similar to the inactivation process in certain mammalian tissues, and different from the process in S. cerevisiae and P. polycephalum, in which enzyme modification, without proportional loss of antigen, accompanies enzyme inactivation. The N. crassa enzyme is therefore suitable as a microbial model for studies of the molecular regulation of the mammalian enzyme.
Ornithine decarboxylase (EC 4.1.1.17) is a tightly regulated, rate-determining enzyme of polyamine biosynthesis. The enzyme has been purified to homogeneity from rat and mouse (l-4). These enzymes are dimers of M, = -54,000 subunits, and are low-abundance proteins in most cells. In contrast, the enzymes from three lower eucaryotes, Saccharomyces cereuisiae (5), Tetrahymena pyriformis (6) and Physarum polycephalum (7), vary greatly in molecular weight and specific activity, and none closely resembles the enzyme of mammals.
A prominent feature of the control of ornithine decarboxylase in all organisms is the inactivation of the enzyme (8,9). In most organisms studied, addition of polyamines causes inactivation (10-13). In the lower eucaryotes, P. polycephulum (14) and S. cereuisiae (5) inactivation, whereas in mammals, it is lost (11)(12)(13)15). In Neurospora crassa, we have studied the loss of enzyme activity and protein after physiological manipulation of polyamine pools (16). We inferred from these preliminary results that putrescine was the signal for enzyme inactivation and that enzyme protein was lost more slowly than activity.
In this paper, the purification and properties of N. crassa ornithine decarboxylase are described. Without the 75-fold augmentation of ornithine decarboxylase activity in the starting material, a 50,000-fold purification would have been required. The procedure routinely yielded l-2 mg of pure ornithine decarboxylase from 40 g of mycelium (dry weight equivalent), with an 11% yield.
The HPLC?-purified preparation contained polypeptides of M, = 53,000 and lesser amounts of others of M, = 44,000-47,000, visualized after SDS-polyacrylamide gel electrophoresis ( Fig. 2A). The lower molecular weight band(s) were labeled if the enzyme preparation was exposed to ['4C]difluoromethylornithine, which binds specifically and covalently to active ornithine decarboxylase molecules (Fig. 2B). In addition, Cleveland proteolytic digests (17) of the polypeptides in the HPLC-purified preparation showed very similar peptide patterns (data not shown). Thus the polypeptide species of the purified preparation were all ornithine decarboxylase or its derivatives.
The lower molecular weight polypeptides are derived by proteolysis from the M, = 53,000 polypeptide during the ammonium sulfate step of purification ( Fig. 2C). Difluoromethylornithine binding was used to determine the intrinsic specific activity of active ornithine decarboxylase  ' Percent purity was calculated using the average picomoles of difluoromethylornithine (DFMO) per unit of Western immunoblot of the crude extract ( l a n e 1 ) and the ammonium sulfate fraction ( l a n e 2) after sodium dodecyl sulfate gel electrophoresis. A cross-absorbed antiserum at a dilution of 1:1,000 was used.
molecules at each stage of purification (see "Experimental Procedures" in Miniprint). The intrinsic specific activity of ornithine decarboxylase is formally represented by the inverse of the value of picomoles of difluoromethylornithine bound/ unit of ornithine decarboxylase activity. The constancy of this value during purification (Table I) indicates that the specific activity of active ornithine decarboxylase molecules was not affected by the purification procedure. All of the protein of pure preparations bound [14C]difluoromethylornithine, indicating that no inactive ornithine decarboxylase molecules were present. Moreover, the specific activity of pure ornithine decarboxylase predicted by difluoromethylornithine binding in crude materials equaled its actual specific activity obtained after purification. Immunotitration was also used to test for inactive molecules. Almost identical units of activity were precipitated per microliter of antiserum (305 and 300, respectively) in the case of crude extracts and DE52-purified ornithine decarboxylase. Thus inactive ornithine decarboxylase molecules did not accumulate during the purification procedure. Under denaturing conditions, polyacrylamide gel electrophoresis revealed that the predominant species was an M , = 53,000 polypeptide ( Fig. 2A), indicating that native ornithine decarboxylase is a dimer. The pH optimum of the enzyme reaction was 7.1. The K,,, for ornithine was 350 p~, and the K,,, for pyridoxal phosphate was 0.16 PM. The Ki for the competitive inhibitor, a-methylornithine, was 280 pM. Arginine, spermidine, spermine, cadaverine, and lysine at a concentration of 2 mM failed to inhibit ornithine decarboxylase. Putrescine (2 mM) inhibited ornithine decarboxylase activity only 30%. Thus it is unlikely that ornithine decarboxylase activity is controlled directly by these metabolites in vivo. Dithiothreitol (2-5 mM) and the non-ionic detergent Brij 35 (0.01-0.1%) increased and stabilized purified ornithine decarboxylase activity, both during storage and during the enzyme reaction.
Pure ornithine decarboxylase displayed a series of isoelectric forms between pH 5.25 and 5.50 (Fig. 3). The same forms were observed in fresh crude extracts (Fig. 3), but their different proportions suggested some selectivity in the purification procedure. The quantitative results with [14C]difluoromethylornithine binding and immunotitration indicate that most or all ionic forms are active. The several forms of the enzyme can be seen in extracts of cells grown in minimal medium (data not shown) and thus do not reflect mistranslation during the polyamine starvation of cells used as a starting material. Multiple ionic forms of the enzyme have been seen in mouse kidney (4, 20). It is not certain whether more than one active copy of the gene is present in the mouse genome or whether allelic heterogeneity among animals or in heterozygotes prevails in these diploid organisms. Because there is only one active gene for ornithine decarboxylase in N. crmsa either increased or greatly diminished putrescine content. A strain carrying the aga mutation, grown in minimal medium and containing normal levels of putrescine and spermidine, had an ornithine decarboxylase activity of 0.15 units/mg of protein ( Table 2 in Miniprint). Cultures grown in medium supplemented with arginine cannot synthesize ornithine and thus become depleted of both putrescine and spermidine (16,18,19). These cultures had a maximally augmented ornithine decarboxylase activity of 3.8 units/mg, consistent with their lack of both putrescine and spermidine ( Table 2). Cultures grown in medium supplemented with the spermidine synthase inhibitor, cyclohexylamine (22), accumulated putrescine (16). They had an ornithine decarboxylase specific activity of 0.72 unit/mg (Table  2). (The steady-state enzyme activity is thought to be the net result of a higher rate of synthesis, owing to the depletion of spermidine, opposed by a higher rate of turnover of the enzyme induced by putrescine (16).) The units of ornithine decarboxylase activity precipitated per microliter of antiserum were very similar in the cyclohexylamine-and arginine-supplemented cultures and somewhat lower in the culture grown in minimal medium ( Table 2). The last observation has little significance at this point, owing to the low ornithine decarboxylase activity and antigen in these cultures. The data, therefore, reveal no inactive ornithine decarboxylase molecules in steady-state cultures, whether the putrescine content of the cells was very high or virtually nil.
Rapid inactivation of ornithine decarboxylase follows the restoration of ornithine to ornithine-starved cells (16), such as those used to purify the enzyme. In one such experiment, ornithine decarboxylase specific activity fell rapidly from 3.6 to 0.05 units/mg in 6 h (Fig. 4). Both ornithine decarboxylase activity and enzyme protein had 2-h half-lives after correction for the dilution caused by further growth after ornithine addition (Fig. 4). The same results were obtained with extracts made from sand-ground mycelia or from acetone powders, using four different antisera. We conclude that removal of protein is simultaneous with the disappearance of activity. The conclusion differs from our preliminary report, based on crude quantification of '251-immunoblots (16), that protein was lost somewhat more slowly than activity.
The requirement for protein synthesis during ornithine decarboxylase inactivation was examined by adding cycloheximide and ornithine simultaneously to a polyamine-starved strain. As previously seen in Neurospora (16), loss of ornithine decarboxylase activity and protein was greatly retarded under these conditions (Fig. 5 in Miniprint); in some experiments, the enzyme is entirely stable. Immunoblots of Ornithine Decarboxylase during Inactiuation-Immunoblots of sodium dodecyl sulfate-polyacrylamide and isoelectric focusing gels were used to reveal changes in the immunoreactive protein during inactivation. The expected augmentation of ornithine decarboxylase protein was seen in conditions of polyamine starvation, and ornithine decarboxylase protein was lost during inactivation. No antigenically active, lower molecular weight forms of ornithine decarboxylase appeared consistently during inactivation (Fig. 5 ) , even when the autoradiographs were overexposed. Immunoblots of the isoelectric focusing gel showed multiple ionic forms before and after the onset of inactivation. The most basic form (PI = 5.5) is lost more rapidly than the others (Fig. 3 ) . More study will show how selective the inactivation is, and whether one isoform is the actual substrate for the inactivation process. Fig. 6 (Miniprint) summarizes the correlation between the Ab50, a measure of ornithine decarboxylase protein (See "Experimental Procedures"), and specific activity during periods of enzyme inactivation, polyamine starvation, and steadystate growth in conditions of putrescine depletion and excess. The ratio of these parameters is constant among samples that vary in specific activity by lOO-fold, although deviations a t low activity and protein are obscured by the scales required to include all the points. The constant ratio between ornithine decarboxylase protein and activity was also seen in immunotitrations of crude extracts and partially purified ornithine decarboxylase preparations that varied in specific activity by 450-fold (see above).
Comparison of Eucaryotic Ornithine Decarboxylases-N. crassa ornithine decarboxylase differs markedly from the purified enzyme of other lower eucaryotes. The M , = 110,000 dimer is different from the M , = 86,000 monomer of yeast (5), the M , = 64,000 monomer of Tetrahymena ( 6 ) , or the M , = 80,000 dimer of Physarumpolycephalum isolated by Barnett and Kazarinoff (7). Moreover, the specific activities of the purified yeast ( 5 ) and Tetrahymena (6) enzymes (31 and 14 units/mg of protein, respectively) are 2 orders of magnitude lower than those of N. crmsa and P. polycephalum. In fact, the N. crassa enzyme, with its dimeric structure, subunit molecular weight (Mr = 53,000), and specific activity (2,610 units/mg of protein), is unique among lower eucaryotic ornithine decarboxylases in its close resemblance to that of mammals. Mammalian ornithine decarboxylases are all dimers of about Mr = 110,000 and have specific activities in the range of 1,400-3,200 units/mg of protein (1)(2)(3)(4).
The behavior of the N . crassa enzyme protein during inactivation differs from the case of yeast (5), in which no evidence of loss of the protein is found, and from P. polycephalum, in which enzyme modification, without proportional loss of protein, has been inferred (14,23). Again, the N . crassa enzyme resembles that of some mammalian systems such as Chinese hamster ovary cells (13) and mouse kidney (4) in showing near-proportional loss of protein and activity. Certain mammalian tissues, such as rat brain, heart, and liver, however, display an antienzyme, a stoichiometrically binding protein which inhibits the enzyme (24-26). The protein may be a controlling factor in these tissues, and indeed, loss of activity without comparable loss of enzyme protein is seen in them. No antienzyme has been detected in N. c r a s~a .~ It is possible that a rate-limiting modification of the protein precedes the disappearance of the N. crassa enzyme. This possibility is reinforced by our observation that cycloheximide interferes with polyamine-mediated enzyme inactivation (Ref. 16 and this paper). Whether this reflects a requirement for a noncovalent antienzyme-like binding agent (24-26) or a pro-G. R. Barnett  tein that covalently modifies ornithine decarboxylase is not known. We are currently exploring this matter by seeking mutations that affect the inactivation process.