Superproduction and rapid purification of Escherichia coli aspartate transcarbamylase and its catalytic subunit under extreme derepression of the pyrimidine pathway.

A strain of Escherichia coli has been constructed which greatly overproduces the enzyme aspartate transcarbamylase. This strain has a deletion in the pyrB region of the chromosome and also carries a leaky mutation in pyrF. Although this strain is a pyrimidine auxotroph, it will grow very slowly without pyrimidines if a plasmid containing the pyrB gene is introduced into it. Derepression occurs when this strain exhausts its uracil supply during exponential growth. Under extreme derepression, aspartate transcarbamylase can account for as much as 60% of the total cellular protein. This host strain/plasmid system can be utilized for the rapid purification of wild-type aspartate transcarbamylase or plasmid-born mutant versions of the enzyme. This system is particularly well-suited for analysis of the latter since the control of overproduction resides exclusively on the bacterial chromosome. Therefore, any plasmid bearing the pyrBI operon can be made to overproduce aspartate transcarbamylase in this host strain. Based on this system, a rapid purification procedure has been developed for E. coli aspartate transcarbamylase. The purification scheme involves an ammonium sulfate fractionation followed by a single precipitation of the enzyme at its isoelectric point. In a similar fashion, this strain can also be employed to produce exclusively the catalytic subunit of the enzyme if the plasmid only carries the pyrB gene. This system may be adapted to overproduce other proteins as well by using this host strain and the strong pyrB promoter linked to another gene.


Superproduction and Rapid Purification of Escherichia coli Aspartate Transcarbamylase and Its Catalytic Subunit under Extreme
Derepression of the Pyrimidine Pathway* (Received for publication, April 15,1985) Sean F. Nowlan and Evan R. KantrowitzS

From the Department of Chemistry, Boston College, Chestnut Hill, Massachusetts 02167
A strain of Escherichia coli has been constructed which greatly overproduces the enzyme aspartate transcarbamylase. This strain has a deletion in the pyrB region of the chromosome and also carries a leaky mutation in pyrF. Although this strain is a pyrimidine auxotroph, it will grow very slowly without pyrimidines if a plasmid containing. the pyrB gene is introduced into it. Derepression occurs when this strain exhausts its uracil supply during exponential growth. Under extreme derepression, aspartate transcarbamylase can account for as much as 60% of the total cellular protein. This host strain/plasmid system can be utilized for the rapid purification of wild-type aspartate transcarbamylase or plasmid-born mutant versions of the enzyme. This system is particularly well-suited for analysis of the latter since the control of overproduction resides exclusively on the bacterial chromosome. Therefore, any plasmid bearing the pyrBI operon can be made to overproduce aspartate transcarbamylase in this host strain. Based on this system, a rapid purification procedure has been developed for E. coli aspartate transcarbamylase. The purification scheme involves an ammonium sulfate fractionation followed by a single precipitation of the enzyme at its isoelectric point. In a similar fashion, this strain can also be employed to produce exclusively the catalytic subunit of the enzyme if the plasmid only carries the pyrB gene. This system may be adapted to overproduce other proteins as well by using this host strain and the strongpyrB promoter linked to another gene.
Because of recent advances in recombinant DNA technology, such as site-directed mutagenesis (Zoller and Smith, 1982), it is now relatively easy to isolate mutant versions of enzymes with single amino acid substitutions. These techniques are allowing enzymologists to use these mutant enzymes as powerful tools to investigate the mechanism of catalysis and allosterism (Dalbadie-McFarland et al., 1982;Winter et al., 1982). In order to characterize these mutant enzymes, it would be of great value to be able to rapidly purify them. As part of our analysis of mutant versions of Escherichia coli aspartate transcarbamylase, we have developed a 3 Recipient of fellowships from the Sloan Foundation and the Camille and Henry Dreyfus Foundation. To whom all correspondence should be addressed. new E. coli strain/plasmid system which will superproduce aspartate transcarbamylase. As a direct consequence of the high levels of enzyme in this system, a new purification procedure has been devised which is simple and extremely rapid.
E. coli aspartate transcarbamylase catalyzes the committed step in pyrimidine biosynthesis: the reaction of aspartate and carbamyl phosphate to form carbamyl aspartate and phosphate. This enzyme contributes to the control of the rate of the pyrimidine pathway by a combination of genetic, metabolic, and allosteric means. Beckwith et al. (1962) found that aspartate transcarbamylase can be derepressed 100-to 400fold. Although a p-independent attenuation mechanism has been proposed (Roof et al., 1982;Navre and Schachman, 1983;Turnbough et al., 1983) to account for this genetic control, by comparison to other related systems it is unlikely that attenuation alone can account for this level of regulation (Bertrand and Yanofsky, 1976;Jackson and Yanofsky, 1973). Metabolically, aspartate transcarbamylase is inhibited by CTP, the end-product of the pyrimidine pathway (Gerhart and Schachman, 1965), and is activated by ATP, the product of the parallel purine pathway (Gerhart and Schachman, 1965;Yates and Pardee, 1956). Finally, aspartate transcarbamylase exhibits homotropic cooperativity with both of its substrates, carbamyl phosphate and aspartate (Gerhart and Pardee, 1962;Bethel1 et al., 1968). E. coli aspartate transcarbamylase is composed of twelve polypeptides of two types. The larger 33,000 molecular weight chain is the product of thepyrB gene, while the smaller 17,000 molecular weight chain is the product of the pyrI gene. Three of the larger or catalytic chains spontaneously combine to form a catalytic trimer and two smaller or regulatory chains combine to form a regulatory dimer (Burns and Schachman, 1982a;Burns and Schachman, 1982b). These subunits then spontaneously assemble to form the holoenzyme which is composed of two catalytic subunits and three regulatory subunits (Rosenbusch and Weber, 1971;Foltermann et d., 1984).
The pyrB and pyrl genes are contiguous on the E. coli chromosome at 96 min on the genetic map (Bachmann, 1983;Pauza et al., 1982). The regulatory subunits which are necessary for the allosteric interactions of the enzyme are not required for catalytic activity since all the enzymatic activity resides on the catalytic subunit. Strains which have the pyrl gene deleted are not pyrimidine auxotrophs since they produce functional catalytic trimers (Foltermann et al., 1984). Gerhart and Holoubek (1967) were able to purify large amounts of aspartate transcarbamylase from an E. coli strain which contained a leaky mutation in the pyrF gene. When this strain was grown in medium containing limited amounts of pyrimidines, derepression occurred when the pyrimidine supply was exhausted. Using this strain, approximately 30 mg of crude enzyme was obtained per liter of culture. These authors were able to obtain approximately 15 mg of pure enzyme/liter of culture by a seven-step purification procedure employing one column chromatography step.
The purification of mutant versions of aspartate transcarbamylase has proven to be difficult (West et al., 1985), especially if activity is reduced or abolished. Here, we report the development of a new strain of E. coli which overproduces aspartate transcarbamylase. Furthermore, this strain will overproduce either wild-type or a mutant aspartate transcarbamylase as long as the pyrBI operon is carried on a plasmid, since control for the overproduction of the enzyme is located on the bacterial chromosome. To insure that the overproduced enzyme is the exclusive product of the pyrB operon carried on the plasmid, pyrB has been deleted from the chromosome. The yield of enzyme from this strain is approximately 10-20fold higher than the original Gerhart and Holoubek strain (Gerhart and Holoubek, 1967). This high level of expression has made possible a new simple purification of both the wildtype and mutant versions of the enzyme without any column chromatography.

EXPERIMENTAL PROCEDURES
Materials-Carbamyl phosphate, L-aspartate, agar, agarose, ampicillin, potassium dihydrogen phosphate, and carbamyl L-aspartate were obtained from Sigma. Enzyme-grade ammonium sulfate was purchased from Schwarz/Mann, and casamino acids were purchased from Difco. The casamino acids were tested to verify the absence of sufficient tryptophan and uracil to support the growth of tryptophan and pyrimidine auxotrophs, respectively. Restriction endonucleases were obtained from either Bethesda Research Laboratories or New England Biolabs and used according to the supplier's recommendations. Carbamyl phosphate was purified by precipitation from 50% (v/v) ethanol and stored dessicated at -20 "C (Gerhart and Pardee, 1962).
Enzyme Actiuity-The transcarbamylase activity was measured at 25 "C and pH 8.3 by either a colorimetric (Pastra-Landis et al., 1981) or pH-stat method (Wu and Hammes, 1973). pH-stat assays were carried out with a Radiometer TTT80 titrator and an ABU80 autoburette. Enzymatic assays were performed at 4.8 mM carbamyl phosphate and 30 mM aspartate.
Bacterial Strains-The various bacterial strains used in this work are listed in Table I. All bacteria growth was carried out at 37 "C. The E. coli strain U39a, obtained from J. Wild, Texas A & M University, was used to construct the overproducing strain. Preparation of Bacterial Extracts-Bacteria were grown overnight to an absorbance of approximately 1.50 at 560 nm in M9 medium (Miller, 1972) supplemented with 0.5% casamino acids and the appropriate amounts of uracil. The bacterial culture was centrifuged at 10,000 X g for 10 min, resuspended in 0.01 M Tris-acetate, pH 8.3, to one-fifth the original volume, and sonicated for 3 min using a Branson sonicator (Model 200) equipped with a microtip. The disrupted ceIls were centrifuged for 25 min at 20,000 X g and the supernatant was used directly for the pH-stat assay. PI Transduction-Transductions utilizing P1, cml, clrlOO were performed as previously described (Nowlan and Kantrowitz, 1983).
Polyacrylamide Gel Electrophoresis-The Ornstein and Davis procedure was used for nondenaturating polyacrylamide gel electrophoresis (Ornstein, 1964;Davis, 1964), while the Laemmli procedure was used for denaturating polyacrylamide gel electrophoresis in the presence of sodium dodecyl sulfate (Laemmli, 1970).
Determination of Protein Concentration-Concentration of pure wild-type aspartate transcarbamylase holoenzyme was determined by absorbance measurements at 280 nm using an extinction coefficient of 0.59 cm2/mg (Gerhart and Holoubek, 1967). All other protein concentrations were determined by the Bio-Rad version of Bradford's dye-binding assay (Bradford, 1976).

RESULTS
Construction of pEK2"Plasmid pEK2, which carries the pyrBI operon in pUC8, was created by restricting pUC8 and pPBhl04 (Roof et al., 1982) with PstI and Sal1 followed by ligation. The ligated mixture was then transformed into competent U39a cells (Maniatis et al., 1982) followed by selection on M9 medium plates supplemented with 0.5% casamino acids and 40 pg/ml ampicillin. Rapid plasmid purification (Maniatis et al., 1982) was performed on a number of candidates. A plasmid was selected which exhibited the proper size, as judged by agarose gel electrophoresis as well as the appropriate restriction pattern.
Construction of pEKl7-A substantial portion of the gene for the regulatory chain of aspartate transcarbamylase (pyrZ) was deleted from pEK2 to generate pEK17 which produces only the catalytic subunit of the enzyme. This was accomplished by first restricting pEK2 into three fragments with BglII and BamHI. The largest fragment, 4.5 kilobases, is comprised of the pUC8 vector, the entire pyrB gene, and its control region, along with a fragment corresponding to approximately one-third of the pyrl gene. This fragment, with complementary ends, was ligated upon itself followed by transformation into U39a. A number of colonies were selected that were Amp' and Ura+. The candidate plasmids from these strains were isolated and one was chosen which had the appropriate size and restriction pattern. This plasmid, pEKl7, was then transformed into EK1104 to create a strain which would overproduce only the catalytic subunit of aspartate transcarbamylase.
Construction of pEK20-Plasmid pEK2O was identical to pEK2 except that the DNA sequence in the region corresponding to tryptophan 209 of the catalytic chain was converted to a tyrosine codon by site-directed mutagenesis. Details of the construction of this plasmid will be described elsewhere.
Construction of EK1102 (ApyrB,trp)-Strain U39a was subjected to methanesulfonic acid ethyl ester mutagenesis (Miller, 1972), followed by two selective ampicillin/cycloserine selections (Kantrowitz et al., 1981) in M9 medium supplemented with 0.5% casamino acids and 30 pg/ml uracil. After the selections, Trp auxotrophs were detected as minute colonies on M9 medium plates supplemented with 0.5% casamino acids, 30 pg/ml uracil, and 0.8 pg/ml tryptophan. Introduction of episome F'123, which carries the entire trpABCDE operon, confirmed that the Trp mutation was in the region of the chromosome which could be cotransduced with pyrF.
Construction of EK1104 (ApyrB,pyrF*)-In order to overproduce aspartate transcarbamylase from a plasmid carrying the pyrBI gene, a strain was constructed which had both a deletion in the pyrB gene and also carried the leaky pyrF (pyrF*) defect originally employed by Gerhart and Holoubek (1967). The construction of this strain was accomplished by introducing this pyrF' allele into EK1102 by P1 transduction and selection for Trp+. A number of Trp+, Ura-candidates Superproduction of E. coli Aspartate Tramcarbamylase were then tested for the presence of pyrF* by transformation of competent cells with pEK2, a pUC8-based plasmid carrying the entire pyrBI operon. Colonies that exhibited Amp' were selected. Transformants with a pyrF' allele would grow normally on M9 medium plates supplemented with casamino acids (Trp-free) and ampicillin, while transformants with a pyrF* gene yielded extremely slow growing colonies. The ratio of pyrF' to pyrF+ colonies was in agreement with the 44% cotransduction frequency found between pyrF and the nearby Trp operon (Bachmann, 1983;Wu, 1966).
Expression of Aspartate Tramcarbamylase-The level of expression of aspartate transcarbamylase can be elevated greatly by derepression of the pathway (Gerhart and Holoubek, 1967). The original strain constructed for the overproduction of the enzyme utilized a leaky mutation in the pyrF gene along with a second copy of the pyrBI gene carried on an episome (Gerhart and Holoubek, 1967). For comparative purposes, we used strain EK005 which carries the same pyrF' allele but is not diploid for pyrBI. As shown in Table 11, when strain EK005 is grown in the presence of 12 pg/ml uracil, there is a 28-fold increase in the level of aspartate transcarbamylase over a comparable strain which has no defects in the pyrimidine pathway (EK017). Under these conditions, the uracil is exhausted during late exponential growth phase, at which time the pathway undergoes derepression. If a plasmid which carries the pyrB gene is inserted into a strain with no other pyrimidine pathway defects (U39a/pEK2), the level of the enzyme increases by a factor of 8. This increase is presumably due to the existence of multiple copies of the gene within the cell. The pyrB gene has been deleted from the chromosome in strain U39a; therefore, the aspartate transcarbamylase produced will be the exclusive gene-product of the pyrB gene carried on the plasmid.
By introducing the pyrF* allele into strain U39a, overproduction of aspartate transcarbamylase can occur under conditions of limited uracil. EK1104, a pyrimidine auxotroph, has the pyrF' allele in a U39a background. The introduction of pEK2 into this strain eliminates the pyrimidine requirement but this strain grows poorly without added pyrimidines due to the pyrF* defect. When this strain (EK1104/pEK2) is grown with limiting 12 pg/ml uracil, drastic overproduction of aspartate transcarbamylase occurs. The derepression causes a 35-fold increase in the level of aspartate transcarbamylase compared with U39a/pEK2. Under these condi- tions, approximately 60% of the cell protein is aspartate transcarbamylase (see Fig. 1).
Expression of the Aspartate Tramcarbamylase Catalytic Subunit-Strain EK1104/pEK17 will produce only the catalytic subunit of aspartate transcarbamylase from the intact pyrB gene. As shown in Table 11, the specific activity observed for EK1104/pEK17 is twice that observed for EK1104/pEK2. Gel electrophoresis, under nondenaturing conditions (Ornstein, 1964;Davis, 1964), of cell extracts from this strain shows no detectable holoenzyme.
Effect of Uracil Concentration on Aspartate Transcarbamylase Levels-The levels of aspartate transcarbamylase in these strains are variable depending on the uracil concentration. Fig. 2 compares the production of aspartate transcarbamylase from strains U39a and EK1104, both containing pEK2. When no uracil is present in the medium, EK1104/ pEK2 produces almost no detectable aspartate transcarbamylase because this strain barely grows without uracil. Under these same conditions, U39a/pEK2 produces approximately 5 times as much enzyme as a "wild-type" strain. Derepression of the pathway occurs in these strains during exponential growth, if the initial uracil concentration is below approximately 20 pg/ml. The fraction of aspartate transcarbamylase to total cell protein increases dramatically at levels of uracil below 20 pg/ml. As seen in Fig. 2, if the concentration of uracil is increased above 20 pg/ml, the fraction of aspartate A B C FIG. 1. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis of samples during the three stages of the purification of wild-type aspartate transcarbamylase. The two major bands in each gel correspond to the 33,000 molecular weight catalytic and 17,000 molecular weight regulatory chains of aspartate transcarbamylase. Lane A is a sample of the supernatant after the sonication step; lane B is a sample after the 65% ammonium sulfate fractionation; and lane C is a sample of the final material after the pH 5.9 isoelectric precipitation. In each case, approximately 800 units of activity (pmol of carbamyl aspartate formed/h) were loaded onto the gel.
[Uracil], pg/ml FIG. 2. Effect of uracil concentration on levels of aspartate transcarbamylase produced. The levels of aspartate transcarbamylase were measured in strains U39a and EK1104 both of which carried the pyrBI gene on pEK2. All cultures were innoculated at exactly the same density and grown for 20-22 h at 37 "C. The cell extract was prepared as described (see "Experimental Procedures") followed by pH-stat assays to determine aspartate transcarbamylase activity. Specific activity is reported in units of mmol of carbamyl aspartate/h/mg.

TABLE I11
Purification of aspartate transcarbamylase from EKllO4/pEK2 Cells were harvested from 200 ml of EK1104/pEK2 grown in M9 medium supplemented with 0.5% casamino acids, 12 pg/ml uracil and 25 pg/ml ampiciliin. These cells were suspended in 10 ml of 0.01 M Tris-acetate buffer, pH 8.3. The cell mixture was disrupted by sonication followed by ammonium sulfate and isoelectric precipitation as described (see "Results").
Step transcarbamylase diminishes. At high levels of uracil, little if any overproduction of the enzyme occurs.
Purification of the Enzyme-Purification of the aspartate transcarbamylase from the overproducing strain was carried out as follows. Typically, 1 g of cells was suspended in 5 ml of cold 0.1 M Tris-C1 buffer, pH 9.2, followed by sonication at 4 "C to disrupt the cells. The resulting mixture was centrifuged at 25,000 x g for 30 min. Ammonium sulfate was added to the sonication supernatant to 65% of saturation in order to concentrate the enzyme. The purification was completed by a single isoelectric precipitation at pH 5.9 (Gerhart and Holoubek, 1967). The final preparation was shown to be homogeneous by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (Laemmli, 1970). Fig. 1 and Table I11 summarize a typical purification of wild-type aspartate transcarbamylase from the EK1104/pEK2.
Purification of a Mutant Aspartate Transcarbamylase-Plasmid pEK20 was transformed into the overproducing strain to test the feasibility of using this strain to purify mutant versions of aspartate transcarbamylase. Plasmid pEK20 is identical to pEK2 except that it has a single base substitution in the pyrB gene resulting in a modified catalytic chain with tyrosine substituted for tryptophan at position 209. When strain EK1104/pEK20 was grown under derepression conditions, substantial over expression of the mutant aspartate transcarbamylase was observed. The mutant enzyme accounted for 53% of the total cellular protein in crude

IV
Purification of mutant aspartate transcarbamylase from EK1104/pEK20 Cells were harvested from 200 ml of EK1104/pEK20 grown in M9 medium supplemented with 0.5% casamino acids, 12 pg/ml uracil and 25 pg/ml ampicillin. These cells were suspended in 10 ml of 0.01 M Tris-acetate buffer, pH 8.3. The cell mixture was disrupted by sonication followed by ammonium sulfate and isoelectric precipitation as described (see "Results"). cell extracts. Purification of the mutant enzyme was accomplished using the identical procedure as outlined above for the wild-type enzyme. A summary of the purification is shown in Table IV.

DISCUSSION
With the development of in vitro mutagenesis techniques, multiple forms of a particular protein or enzyme with single amino acid substitutions can now be constructed readily (Zoller and Smith, 1982). Although the actual genetic manipulations often can be accomplished in a short period of time, the purification and biochemical characterization of these mutant proteins and enzymes is often very time consuming. To analyze the catalytic and regulatory properties of aspartate transcarbamylase, we have begun to generate a series of single amino acid substitution mutants. In order to purify some of these mutants a rather complex purification scheme is often required (West et al., 1985). By constructing a special strain of E. coli which overproduces this enzyme in large quantities, the purification is vastly simplified.
The strain EK1104, which carries a deletion in the pyrB region of the chromosome as well as a leaky pyrF allele, is ideal for the production of large quantities of aspartate transcarbamylase. Furthermore, the only way that this strain can produce aspartate transcarbamylase activity is if it carries a plasmid encoding the pyrB gene. Of particular significance is that this strain will overproduce aspartate transcarbamylase from any plasmid which carries the pyrBI gene, mutant or wild-type. Since the control of pyrBI expression is solely on the bacterial chromosome, no genetic manipulations of strain EK1104 have to be performed to cause overproduction of the aspartate transcarbamylase holoenzyme; this strain can be transformed with a plasmid which carries the pyrBI operon and then grown under limited uracil conditions. This system is therefore ideal for purifying mutant enzymes produced by in vitro mutagenesis techniques in which the mutation is introduced directly into the pyrB or pyrI genes carried on a plasmid.
The level of expression of aspartate transcarbamylase in this overproducing strain is exceptional. Under conditions of maximal derepression, aspartate transcarbamylase accounts for more than 60% of the total cell protein. Although this value seems exceedingly high, it is consistent with the enzyme levels expected in a strain which contains a multicopy plasmid under pyrimidine derepression. Strain EK005, which has a single copy of the pyrB gene on the chromosome, can also be derepressed in a similar fashion, although this strain exhibits 15-fold lower enzyme levels than EK1104/pEK2. The introduction of the plasmid alone accounts for a 15-fold increase in enzyme levels.
E. coli strains that have the pyrl gene deleted produce only active catalytic trimers (Foltermann et al., 1984); therefore, if a plasmid containing only the pyrB gene is introduced into EK1104, this strain should overproduce the catalytic subunit of the enzyme. To test this hypothesis, plasmid pEK17 was constructed. Approximately two-thirds of the pyrl gene has been deleted in this plasmid. As shown in Table 11, the transcarbamylase activity in EK1104/pEK17 was even higher than in the strain which overproduces the holoenzyme. Since the specific activity of the isolated catalytic subunit is approximately double that of the holoenzyme, these results indicate approximately the same level of expression of either catalytic subunit or holoenzyme.
Purification of wild-type aspartate transcarbamylase from this overproducing strain is relatively simple. The cells are disrupted and a cell extract was obtained from which the proteins are concentrated by ammonium sulfate precipitation. The aspartate transcarbamylase is then purified by a single precipitation at its isoelectric point. Using this procedure, we have been able to obtain gram quantities of pure enzyme as judged by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Specific activity and kinetic properties of the enzyme produced in this fashion are identical to that found for enzyme produced by the original procedure (Gerhart and Holoubek, 1967).
The overproducing strain that was constructed in this work was designed for use with both wild-type aspartate transcarbamylase as well as mutant versions of the enzyme in which a mutant pyrBI operon is carried on a plasmid. Using plasmid pEK2O which carries such a mutation, purification of a mutant version of aspartate transcarbamylase was accomplished (see Table IV). Although the purification of this particular mutant was carried out under exactly the same protocol as for the wild-type, mutant enzymes which have amino acid substitutions that affect the enzyme's isoelectric point will require variations to this procedure. However, because of the very high expression of the mutant enzyme in the overproducing host strain, it is unlikely that significant problems would be encountered in the purification.
High expression induced by derepression of EK1104 may be utilized for production of other proteins if their structural genes were linked downstream from the pyrB control region. Therefore, this system is not restricted to the production of large quantities of aspartate transcarbamylase.