Purification and Characterization of Yeast Orotidine 5’-Monophosphate Decarboxylase Overexpressed from Plasmid PGU2*

Orotidine 5’-monophosphate decarboxylase (ODC- ase) has been overexpressed in yeast 15C cells transformed with a plasmid carrying the URA3 gene that encodes ODCase. Twenty g of cells having ODCase activity equal to 30 mg of pure enzyme per liter of cell culture were obtained after 9 h of galactose induction. To remove yeast proteases, a 60-90s ammonium sul- fate fractionation step plus the addition of EDTA as an inhibitor of metallopeptidases was necessary. The pu- rification protocol yielded ODCase that was protease-free and stable to storage at 4 “C for 16 months. The pure enzyme had a specific activity of 40 units/mg in 50 mM phosphate buffer, pH 6, and could be stored at -20 “C in 20% glycerol with retention of full activity for more than 2 years. The enzyme had a K,,, for orotidine 5’-monophosphate of 0.7 p~ at pH 6 and 25 “C. The molecular weight of the plasmid-derived ODCase monomer electrophoresis on polyacrylamide gels 29,500. ODCase through sucrose density gradients as a monomer of about 30 kDa at low protein concentration the of ligands at the catalytic site. An in the sedimentation the ODCase ODCase in a single band typical of a protein of 46

Orotidine 5'-monophosphate decarboxylase (ODCase) has been overexpressed in yeast 15C cells transformed with a plasmid carrying the URA3 gene that encodes ODCase. Twenty g of cells having ODCase activity equal to 30 mg of pure enzyme per liter of cell culture were obtained after 9 h of galactose induction. To remove yeast proteases, a 60-90s ammonium sulfate fractionation step plus the addition of EDTA as an inhibitor of metallopeptidases was necessary. The purification protocol yielded ODCase that was proteasefree and stable to storage at 4 "C for 16 months. The pure enzyme had a specific activity of 40 units/mg in 50 mM phosphate buffer, pH 6, and could be stored at -20 "C in 20% glycerol with retention of full activity for more than 2 years. The enzyme had a K,,, for orotidine 5'-monophosphate of 0.7 p~ at pH 6 and 25 "C. The molecular weight of the plasmid-derived ODCase monomer determined by electrophoresis on denaturing polyacrylamide gels was 29,500. ODCase sedimented through sucrose density gradients as a monomer of about 30 kDa at low protein concentration and in the absence of ligands that bind at the catalytic site. An increase in the sedimentation rate could be induced by increasing the ODCase concentration or by adding ligands that are competitive inhibitors. ODCase sedimented in a single band typical of a protein of 46 kDa at the highest protein concentration studied or in the presence of 50 mM phosphate or 933 pM substrate (orotidine 5'-monophosphate) or product (UMP). A dimer sedimenting as a protein of about 64 kDa occurred in the presence of 50 p~ 6-azauridine 5'-monophosphate or 2 p~ 1-(5'-phospho-~-~-ribofuranosyl) barbituric acid, competitive inhibitors of ODCase. These results resemble the ligand-induced subunit association of the ODCase domain of bifunctional UMP synthase and support the use of yeast ODCase as a model for ODCases from other species. ' The abbreviations used are: OMP, orotidine 5'-monophosphate; ODCase, orotidine 5'-monophosphate decarboxylase; BMP, 1-(5'phospho-P-o-ribofuranosy1)barbituric acid aza-UMP, 6-azauridine 5'-monophosphate; PMSF, phenylmethylsulfonyl fluoride; SDS, sodium dodecyl sulfate, MES, 2(N-morpho1ino)ethanesulfonic acid. dine 5'-monophosphate (UMP), the last step in the de nouo pyrimidine biosynthetic pathway, is catalyzed by orotidine 5'monophosphate decarboxylase (ODCase, EC 4.1.1.23). In mammals ODCase is part of the bifunctional protein, UMP synthase, which also contains the preceeding enzymatic activity in the pathway, orotate phosphoribosyltransferase (1). The deduced amino acid sequence of yeast ODCase ( 2 ) shares 53-54% homology with the sequence of the ODCase domain of mouse (3) and human (4) UMP synthases, which are themselves 90% identical. In addition, the deduced amino acid sequences of ODCases from several other species share several well conserved regions with yeast ODCase (5).
In the 1970's ODCase was purified from commercial bakers' yeast by several groups (6)(7)(8). The amount of ODCase in crude yeast extracts is small, and therefore, large, bulky preparations were required to obtain pure protein. The purification described by Brody and Westheimer (6) yielded 15 mg of pure ODCase from 12 pounds of pressed bakers' yeast using a large (400 ml) affinity column and two additional column chromatography steps. Such laborious purification schemes to prepare ODCase from commercial yeast made obtaining the hundreds of milligrams of ODCase required for structural and mechanistic studies involving crystallography and NMR spectroscopy forbidding.
In 1987 Lue et al. (9) constructed a vector for overexpression of URA3, the gene which codes for yeast ODCase. This plasmid, pGU2, carries the URA3 gene under control of the promoter for GALl, along with a region of the 2-micron plasmid needed for maintenance of the plasmid at high copy number in yeast, and the GAL4 gene which codes for the GAL4 protein (9). The GAL4 protein promotes transcription from the promoter for GALl in response to galactose (lo), thus allowing the expression of plasmid-derived proteins to be induced by adding galactose to the growth medium. Workers in Roger Kornberg's group (9) reported yields of 30 mg of ODCase per liter of yeast culture with a single purification step using yeast strain Sf657-2D transformed with plasmid pGU2.
We have utilized plasmid pGU2 to optimize expression of ODCase in a similar yeast strain, 15C, which was supplied to us by the Kornberg laboratory. Modifications to the published protocols for purification of yeast ODCase (6,9) were essential in order to eliminate protease activities so that large amounts of highly pure, stable ODCase could be obtained.
Both the conserved amino acid sequence and the availability of large amounts of pure enzyme make yeast ODCase an ideal model for studying the structure and mechanism of ODCases. In reports from this laboratory, yeast ODCase purified by the protocol described herein has been crystallized for structural studies (11) and used in catalytic mechanism studies of the binding of a I3C-labeled inhibitor using NMR spectroscopy (la), as well as in kinetic studies of "'C isotope 12662 This is an Open Access article under the CC BY license.

Yeast Orotidine 5'-Monophosphate Decarboxylase
effects (13). In this report we present details of the overexpression, purification, and characterization of plasmid-derived ODCase. Its structural and kinetic properties are compared with those reported for ODCase purified from commercial bakers' yeast. In addition, we examine the sedimentation of yeast ODCase in sucrose gradients and compare the effects of ligands in promoting different aggregation states of the enzyme with those observed for UMP synthase (14-16).

RESULTS AND DISCUSSION*
Overexpression of Yeast ODCase-We obtained plasmid pGU2 and yeast strain 15C from Andrew Buchman in Roger Kornberg's laboratory, who informed us that they no longer used strain Sf657-2D for overexpression of ODCase. Therefore, modifications to their published procedure for overexpression of yeast ODCase from strain Sf657-2D cells carrying the pGU2 plasmid (9) were necessary to obtain a high level of ODCase expression in yeast strain 15C. Using the published protocol (9), yeast 15C cells grew poorly in unsupplemented YP medium reaching an OD660nm of only 1.5 instead of 4 as was reported for Sf657-2D cells, and only a slight increase in the total ODCase activity occurred when cells grown in unsupplemented YP medium were harvested after 2 h of galactose induction (Table I).
Our improved protocol included addition of 2% sucrose to the YP medium for the initial growth of the cells and induction with galactose for 9 instead of 2 h. With these modifications yeast 15C cells grew to OD660. m of 2.53 in YP medium plus 2% sucrose prior to galactose induction, with a 2.6-fold increase in the cell wet weight when compared to growth in YP medium without sucrose (Table I). Harvesting cells after 9 h of galactose induction yielded an increase in the total ODCase activity of 4.4-fold in cells grown in YP medium without sucrose and 6.8-fold in cells grown in YP medium plus 2% sucrose, compared with the activity in cells grown in YP medium without sucrose which were harvested after only 2 h of galactose induction (Table I). The observed increase was due to an increase in both the total protein and the amount of ODCase. Our modified protocol typically yielded 18-22 g of cells (wet weight) and ODCase activity equal to about 30 mg of pure ODCase per liter of YP medium, as was reported for Sf657-2D cells transformed with plasmid pGU2 (9).
Purification and Stabilization of ODCase-We found that the ODCase protein overexpressed from plasmid pGU2, when purified by previous protocols (6,9), was highly susceptible to proteolysis after storage for as little as 1 month at 4 "C (Fig. 1). These ODCase preparations contained high levels of protease activities, which completely destroyed the ODCase activity after 2 months of storage at 4 "C. Since we were interested in crystallizing the enzyme and conducting mechanism studies (both of which could require that the enzyme be kept for long periods a t 4 "C or higher temperatures), it was necessary to stabilize the enzyme against proteolysis so as to maintain an active, homogeneous protein preparation.
In order to design a strategy for removing proteases, we first determined which of the three classes of yeast protease activities, distinguished by their pH optima (17,18 as was done by Lue et al. (9), the cell lysate still contained significant amounts of basic and neutral protease activities. These two classes of protease activities include proteinase B and carboxypeptidase Y, both neutral proteases, as well as a number of metallopeptidases which are either neutral or slightly basic proteases (17). Yeast strain 15C contains apep4 mutation, which makes it deficient in both proteinase B (19) and carboxypeptidase Y (20). Therefore, we reasoned that metallopeptidases might comprise a significant portion of the residual protease activity in the enzyme preparations. Adding 2 mM EDTA in addition to the other protease inhibitors reduced the neutral protease activity measured in cell lysates by 70-75%. Including an ammonium sulfate precipitation step and collecting the 60-90% precipitated protein completely eliminated the remaining neutral and basic protease activities, as well as the small amount of residual acidic protease activity. The purified ODCase contained no measurable protease activity and was stable against proteolysis when stored for up to 16 months at 4 "C (cf. Fig. 2, lane 12 with Fig. 1, lane 4 ) . We have added 2 mM EDTA to the protease inhibitor mixture containing 1 mM PMSF, 1 PM pepstatin A, and 0.6 pM leupeptin and have routinely added this mixture to each buffer solution used during the ODCase purification.
Yeast ODCase overexpressed from plasmid pGU2 was purified to homogeneity in three purification steps (Table I1 and Fig. 2). The CM-52 cellulose column removed a small amount of contaminating proteins remaining after elution from the Affi-Gel Blue column (cf. Fig. 1, lunes 1 and 2). In a typical purification, about 60 mg of ODCase with a specific activity of 39-43 units/mg (at 25 "C in phosphate buffer, pH 6) or 70-80 units/mg (at 37 "C in Tris-HC1 buffer, pH 7.4) was obtained from 4.8 liters of cell culture with a recovery of 30-40% (Table 11). Recovery is dependent on the binding capacity of the Affi-Gel Blue column, which diminishes with repeated use. The specific activity at pH 6 is consistent with the 35- 40 units/mg reported for other yeast purifications (6,9). The pure ODCase monomer had a molecular mass of 29,500 daltons as estimated from its migration on a 12% SDS-polyacrylamide gel (Fig. 2, lanes 9-12). This value is close to the 29,000 daltons reported by Lue et al. (9) and is consistent with the value of 29,200 calculated for the amino acid sequence deduced from the nucleotide sequence of the URA3 gene (2). This value is higher than the 27,500 reported by Brody and Westheimer (6) for ODCase purified from commercial bakers' yeast. While this difference may reflect a variation in the molecular weight standards used for estimating the weight of the OD-Case monomer on SDS-polyacrylamide gels, we suspect that the difference could also be due to proteolysis of the enzyme prepared from autolyzed commercial bakers' yeast without the addition of protease inhibitors (6).
Subunit Association Studies-Previous studies of ODCase purified from commercial bakers' yeast have indicated that the enzyme exists in its native state as a simple dimer of 51 kDa (8). In contrast, a monomer and dimer form of the ODCase domain of mouse UMP synthase expressed in yeast cells have been reported (14). Studies of bifunctional UMP synthase isolated from mouse Ehrlich ascites cells have established that this UMP synthase can exist in three distinct conformational forms: a 3.6 S monomer, a 5.1 S simple dimer, and a 5.6 S dense dimer (15, 16). The aggregation states were identified by sedimentation of UMP synthase through sucrose gradients containing various ligands that bind to the enzyme. Data from these studies allowed Traut et al. (16) to suggest that a simple dimer, the 5.1 S species, is produced by ligand binding to the ODCase catalytic site and that formation of a dense dimer, the 5.6 S species, is promoted by effector binding to a noncatalytic, regulatory site.
In order to investigate the subunit association and aggregation state of yeast ODCase, the protein was subjected to centrifugation through 5-20% sucrose density gradients under a variety of experimental conditions. The effects of phosphate and enzyme concentration on the sedimentation of yeast ODCase were first examined in order to determine optimal conditions for studying the effect of nucleotide ligands on the subunit association of ODCase. We observed that ODCase stored in 50 mM sodium phosphate buffer migrated through sucrose gradients at a molecular mass intermediate to that expected for a monomer (29.5 kDa) or a dimer (59 kDa). A similar effect of orthophosphate on the sedimentation of UMP synthase has been attributed to the rapid equilibration of the monomer and dimer forms of the enzyme (15). When protein from an ammonium sulfate precipitation step (ODCase specific activity = 15 units/mg) was loaded onto a sucrose gradient in the amount of 10, 50, or 200 pg (equivalent to about 2, 10, and 40 pg of pure ODCase), a single ODCase activity peak was observed in each gradient. The apparent molecular masses of ODCase in gradients loaded with the three protein concentrations were 29, 36, and 46 kDa, respectively. We attribute this change in apparent mass to an equilibrium between the monomer and dimer forms of ODCase induced by a mass action effect as the ODCase concentration is increased. In order to eliminate the effects that protein concentration and phosphate have on the aggregation and subunit association of yeast ODCase, sucrose gradients to determine the effect of nucleotide ligands were run using a small amount of pure ODCase that had been dialyzed against Tris buffer.
When pure ODCase (2.2 pg) was sedimented through a 5-20% sucrose gradient in the absence of any nucleotide, OD-Case migrated as a monomer of about 30 kDa (Fig. 3). Under identical conditions, ODCase sedimented as a dimer of about 64 kDa in a gradient containing either 50 p~ aza-UMP or 2 p~ BMP (Fig. 3). Gradients containing either OMP or UMP at a concentration of 933 p~ both yielded an ODCase activity peak at a position in the gradient corresponding to a molecular mass of 46 kDa, which is intermediate between the monomer and dimer (Fig. 3). Since the ODCase activity loaded onto the gradient was sufficient to consume all of the OMP present during the course of the 40-h centrifugation, this intermediate species probably resulted from binding of UMP to ODCase in both cases.
In summary, changes in the aggregation and/or subunit association state of yeast ODCase can be influenced by protein concentration, phosphate, and nucleotide ligands. Two distinct species of the yeast ODCase were identified a 30-kDa monomer observed at low enzyme concentration in the absence of phosphate and other ligands, and a 64-kDa dimer observed in the presence of aza-UMP or BMP.
Based on gel filtration and sucrose density gradient studies, the native state of ODCase purified from commercial bakers' yeast was considered to be a dimer of 51 kDa (8). The fact that a monomer of the yeast ODCase in solution had not been observed previously can be explained in light of our results, which suggest that the 51-kDa species of ODCase probably corresponds to the 46-kDa intermediate observed in the presence of UMP, phosphate, or at high protein concentration. The previous protocol (8) used phosphate buffer during the purification and sucrose density gradient centrifugation, which can cause ODCase to migrate a t a higher apparent molecular mass. Also, the protein concentrations used during sucrose gradient centrifugation and gel exclusion chromatography could have been sufficient to promote subunit association. In each of the subsequent purifications of yeast ODCase (6,9,21) in which aza-UMP was used to elute the enzyme from affinity columns, no attempt was made to remove residual aza-UMP from the enzyme. As we have shown, a substantial amount of the inhibitor remains bound to the enzyme and must be removed by sequential dialysis (see Fig. 4 and "Results" in Miniprint). Therefore, in these cases the residual bound aza-UMP may have been sufficient to promote subunit association of the enzyme and preclude observation of the monomer form of ODCase.
Changes in the subunit association of yeast ODCase in the presence of ligands that bind to the catalytic site of the enzyme are very similar to those reported for the ODCase domain (14) and the intact UMP synthase (15, 16) from mouse. Such changes in the aggregation state of multisubunit enzymes have been suggested as a means of regulating enzymatic activity (22). Indeed, Traut et al. (16) have hypothesized that the 5.6 S dense dimer is promoted by effector binding to a regulatory site and is the only form which has ODCase activity. While we have observed both a monomer and a dimer form of yeast ODCase, we have no evidence for a species equivalent to the dense dimer species observed in UMP synthase. In this respect, yeast ODCase more closely resembles the isolated ODCase domain of the bifunctional mouse UMP synthase. Results in this report have shown that properties of the plasmid-derived ODCase are very similar to those of the enzyme prepared from commercial bakers' yeast. In addition, our results have shown that yeast ODCase closely resembles the ODCase domain of UMP synthase in its subunit association properties, thus adding to kinetic and sequence data which support its use as a model for ODCases. Therefore, cell extracts containing the plasmid-derived yeast ODCase are excellent sources for obtaining large quantities of pure enzyme for studies on the structure and mechanism of ODCases.