Purification and Characterization of Ornithine Transcarbamoylase from Saccharomyces cereuisiae*

Ornithine transcarbamoylase (OTCase) has been pu- rified in 100-rng quantities from a plasmid-containing, enzyme-overproducing strain of Saccharomyces cere- visiae. The specific activity of the homogeneous enzyme is 2.5-fold above that previously reported. The molecular weight and partial specific volume of OTCase were determined by sedimentation equilibrium in solutions containing HzO and DzO. Data from two rotor speeds were simultaneously fit using nonlinear least squares analysis with multiple independent variables giving a molecular weight of 110,000 f 2,200 and a partial specific volume of 0.732 2 0.006 ml g-’. The ultraviolet absorption spectrum of OTCase gives a spe- cific absorbance at 280 nm of 0.36. This low value is consistent with a small number of aromatic residues. Amino acid analysis, fluorescence, and multicomponent analysis yield 1 tryptophan, 4 tyrosine, and 24 phenylalaninelpolypeptide chain. From an analysis of the circular dichroic spectrum, it was determined that OTCase contained 22% a-helix, 43% &sheet, 8% p- turn, and 27% random structure. The fluorescence of the single tryptophan/polypeptide chain has an emis- sion maximum at 320 nm, indicating a hydrophobic environment.

The first committed step in the biosynthesis of L-arginine in saccharomyces cerevisiae is catalyzed by ornithine transcarbamoylase. OTCase' (EC 2.1.3.3) catalyzes the condensation of t-ornithine and carbamoyl phosphate to form Lcitrulline. The first committed step in the degradation of Larginine is catalyzed by arginase. Arginase (EC 3.5.3.1) catalyzes the hydrolysis of L-arginine to L-ornithine and urea. In yeast, both of these enzymes exist in the cytoplasm. Hence, when L-ornithine and L-arginine levels are high, the potential exists for the coupling of the two pathways into a futile urea * This work was supported in part by United States Public Health Service Research Grant GM 28731 from the National Institutes of Health and a grant from the Research Corporation. A preliminary account of a portion of this work was presented at the 74th Annual Meeting of the American Society of Biological Chemists, June 5-9, 1983, San Francisco, CA (Eisenstein and Hensley, 1983a) and the 15th Annual Meeting of the Federation of European Biochemical Societies, July 24-29, 1983, Brussels, Belgium (Eisenstein andHensley, 1983b). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduerbement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The study of the regulation of enzyme activity in this system is important because, for the first time, the potential exists for a detailed analysis of the linkage between changes in subunit interaction energy and the modulation of enzyme activity in a well defined, discrete multienzyme complex. An understanding of the thermodynamic linkage in this system will give insight into linkage phenomena in other multienzyme complexes, which due to their enormous size and complexity are less likely to be resolved on a detailed level (Reed, 1974(Reed, , 1981Hammes, 1981;Reed and Oliver, 1982).
Although a preliminary thermodynamic characterization of the formation of the regulatory complex has been made indirectly using enzyme kinetic data (Penninckx and Wiame, 1976), the opportunity to explore these linkage relationships directly requires that the thermodynamic and structural attributes of OTCase and arginase be precisely determined. Consequently, the availability of large quantities of highly purified enzymes is mandatory. In this paper, we describe a rapid, large scale purification for yeast OTCase and present preliminary studies defining some of its thermodynamic and structural properties.

RESULTS
Purification of OTCase-OTCase has previously been purified from wild type s. cereuisiae, strain 21278b, in 1-2-mg quantities (Penninckx et al., 1974). In order to undertake the planned chemical and physical studies of the regulation of OTCase activity by arginase in the multienzyme complex, gram quantities of enzyme are essential. Attempts to scale up the previous procedure using a newly developed enzyme-  Kalckar (194;). 1 unit = 1 pmol of citrulline formed per mg of OTCase/min. 'Specific activity was determined at 10 mM ].-ornithine and 10 mM rarhomoyl phosphate at 30 "C using a modification ofthe method of Elodi (1954) (E. Eisenstein and P. Hensley, unpuhlished data).
"Protein concentration was determined from the extinction coefficient of homogeneous OTCase, ( = 0 3 6 ml mg" cm" at 280 nm. "Since I'enninckx e t a[. (1974) measured protein concentration hy the Kalckar method, the specific activity they quoted was too high. (!sing the extinction coefficient of homogeneous OTCase, the specific activity of the enzyme produced by their protocol would be 224 units mg". overproducing strain, lCllllC, proved unsuccessful. Therefore, the procedure of Penninckx et al. (1974) was modified, and a summary of the new steps and yields in this protocol is presented in Table I. These purification steps repeatedly yield about 100 mg of highly purified enzyme with a specific activity of 600 f 100 units mg" of OTCase from 150 g of yeast cells, wet weight. This value is actually 2.5 times higher than that previously reported.:' Homogeneity of OTCase resulting from a typical purification procedure was demonstrated by submitting samples to polyacrylamide gel electrophoresis under both denaturing and nondenaturing conditions. Fig. 4 shows Coomassie blue R-250-stained gels run as described under "Experimental Procedures." M o k c u l a r Weight-To determine the molecular weight of purified OTCase, equilibrium sedimentation studies were undertaken as described under "Experimental Procedures." In order to calculate a molecular weight from this data, the partial specific volume o f t h e molecule must he known. For a homogeneous, noninteracting system, these parameters may he determined at sedimentation equilibrium in solutions made up in H,O and D,O (Edelstein and Schachman, 1967). Additionally, these experiments were performed at two rotor speeds. Fig. 5 shows the data from equilibrium sedimentation ' "The specific activity reported here, 600 & 100 units mg-l, is determined at pH 8.5 and 30 "C and in the presence of 10 mM I.ornithine and 10 mM carbamoyl phosphate. The amount of OTCase present is hased on the extinction coefficient for purified OTCase of ~1.1) = 3.6. Specific activity is expressed as micromoles of' citrulline produced per mg/min under the ahove conditions. Penninckx et al. (1974) have reported a specific activity of 11,900 units mg" in the presence of 7 mM 1.-ornithine and 10 mM carhamoyl phosphate at pH 8.5 and 15 "C. They determine the amount of enzyme present in their purified pool hy Kalckar's method (1947). This procedure overestimates the amount of OTCase present hv 3-fold. Furthermore, they denote specific activity as micromoles of citrulline produced per mg/ h. Working under the conditions of Penninckx et a/. (1974). i.e. 1) using 7 mM 1.-ornithine and 10 mM carhamoyl phosphate, 2) at 15 "C, 3 ) estimating the enzyme concentration hy Kalckar's method, and 4) defining specific activity as micromoles of citrulline produced per mg/h, the specific activity of enzyme isolated as discussed ahove is 29.500 units mg", or an increase of ahout 2.5-fold over 11,900. in the form of protein concentration uersus radius for the four different experimental conditions. In contrast to the method of data analysis suggested by Edelstein and Schachman (196T), it was possible to calculate both M, and E from the primary data using nonlinear least squares analysis with the multiple independent variables: radius, rotor speed, and solvent density. The value calculated for the molecular weight of OTCase was 110,000 ? 2,200 and for the partial specific volume was 0.732 f 0.006 ml g".
In the analysis of nonlinear data by least squares methods, there are no formal statistical tests for goodness of fit; therefore, the precision and accuracy of parameter estimation are determined qualitatively. The random distribution of resid-

Purification and Characterization
of Yeast OTCase 5141 uals, narrowness of confidence limits, and whether the determined values are physically reasonable are used to indicate t,he validity of parameter estimation (Osborne et al., 1977;Turner et al., 1981;Osborne et al., 1982). In addition to the narrow confidence limits determined for the molecular weight and the partial specific volume of OTCase, the distribution of residuals is random, and the residuals themselves are small in magnitude as shown in the bottom of Fig. 5. The subunit molecular weight for OTCase was 37,000 & 3,700 based on SDS-polyacrylamide gel electrophoresis (Fig.  4A), which is consistent with the trimeric structure proposed by Penninckx et al. (1974).
Spectral Properties-The UV absorption spectrum of OT-Case shown in Fig. 6 is given as specific absorbance versus wavelength based on the value of the extinction coefficient of &?
= 3.6. This number appeared low for a protein the size of OTCase. This low extinction, as well as the high degree of fine structure in the 280-260 nm region, prompted a determination of the content of aromatic amino acids in OTCase. Fitting the second derivative of the UV absorption.spectrum of OTCase in GdmCl to the second derivatives of UV absorption spectra for model compounds of tryptophan, tyrosine, and phenylalanine residues in proteins in GdmCl (Levine and Fedesici, 1982), values of 1.0 tryptophan, 3.77 tyrosine, and 16.67 phenylalanine/polypeptide chain were found. Only 1 tryptophan/chain accounts for the low extinction coefficient for OTCase, and the apparent presence of 17 phenylalanine and only 4 tyrosine is consistent with the fine structure in the spectrum. Since only 1 tryptophan/chain in OTCase was found, the fluorescence properties of the enzyme were examined. The fluorescence excitation and emission spectra of OTCase are shown in Fig. 7. These spectra were recorded in the ratio mode using 3-nm slit widths. For the excitation spectrum, emission was monitored at 320 nm, and for the emission spectrum, the excitation wavelength was 295 nm. The emission spectrum is characterized by a maximum a t 320 nm, indicating a hydrophobic tryptophan environment in OTCase.
To estimate the fraction of a-helix, @-sheet, 0-turn, and random secondary structure in OTCase, the circular dichroic spectrum was recorded in the far UV region (240-200 nm) as described under "Experimental Procedures." The experimentally obtained spectrum for OTCase is shown in Fig. 8. The two curves represent the fit of the dat,a by the two methods described under "Experimental Procedures." Using the reference spectrum technique (Chang et al., 1978), OTCase was estimated to contain 23% a-helix, 37% P-sheet, 10% 0-turn, and 30% random structure. Alternatively, using the method in which the experimental spectrum was compared to a library of CD spectra of proteins with known secondary structure (Provencher and Glockner, 1981), OTCase was estimated to contain 22% a-helix, 43% @-sheet, 8% P-turn, and 27% random structure.
Amino Acid Composition-The amino acid composition of OTCase is given in Table 11. Integral values for amino acids present are based on extrapolation to zero time from three timed hydrolysis samples or from averages from three runs.  Values represent averages from 48-and 72-h hydrolyses. e Value for tyrosine represents the average of 24-, 48-, and 72-h hydrolyses since there was no indication of degradation at 48 and 72 h. Additionally, this value is in excellent agreement with 3.77 Tyr/ chain determined using the second derivative of the ultraviolet absorption spectrum of OTCase in GdmCl (Levine and Federici, 1982).
Values for the partial specific volume of each amino acid were taken from Cohn and Edsall (1943).
Tryptophan was determined by fluorescence measurements using known concentrations of OTCase and L-tryptophan (Pajot, 1976). It was found that there was 0.99 tryptophan/ polypeptide chain in the enzyme. This value is in excellent agreement with the multicomponent analysis of the second derivative of the absorption spectrum in GdmC1.
Cysteine was determined by mixing 5,5'-dithiobis-(2-nitrobenzoic acid) with OTCase previously treated with urea and NaBH4 to denature the enzyme and reduce any disulfide groups (Cavallini et al., 1966). The absorbance a t 412 nm was recorded and compared with the molar extinction coefficient of 12,000 for the nitrobenzenethiol anion (Flavin, 1962). This method gave 0.98 cysteine/polypeptide chain in OTCase.

DISCUSSION
Under appropriate physiological conditions, OTCase and arginase form a regulatory multienzyme complex in yeast. In this complex, OTCase activity is completely inhibited while the activity of arginase is unaffected. Prior to a detailed analysis of the modulation of enzyme activity which results from the interaction of these two enzymes, both enzymes must be available in large quantities and their chemical and physical attributes known.
A summary of the newly developed purification procedure for OTCase is given in Table I. The procedure involves three column chromatography steps and may be accomplished rapidly. Using this protocol, 100 mg of highly purified OTCase can be isolated from 150 g of plasmid-containing, OTCaseoverproducing yeast cells. In this strain, OTCase represents 6-8% of the soluble protein. The homogeneity of the protein isolated using this procedure was verified by polyacrylamide gel electrophoresis under denaturing and nondenaturing conditions (Fig. 4). Coomassie blue R-250-stained gels containing up to 25 wg of protein/lane show single bands.
The specific activity of yeast OTCase isolated as described above is 2.5 times that previously reported (Penninckx et al., 1974). This increase in specific activity prompted the question of whether the kinetic properties of OTCase prepared as described here are the same as those previously reported. The K,,, for L-ornithine was determined in the presence of 10 mM carbamoyl phosphate, and the previously noted phenomenon of reduced enzyme activity a t L-ornithine concentrations above 5 mM was confirmed (Messenguy et al., 1971). Data analysis in terms of a model wherein L-ornithine acts as an uncompetitive inhibitor of OTCase (Cleland, 1979b) revealed a K , for L-ornithine of 0.90 k 0.28 mM and a K i of 23 -t 8.7 mM. The theoretical value for V , in the absence of apparent substrate inhibition was 825 f 100 units mg". Varying carbamoyl phosphate in the presence of 5 mM L-ornithine to minimize apparent substrate inhibition yielded a X , of 0.2 2 0.03 mM and a V , of 690 k 20 units mg". These values for the K , of L-ornithine and carbamoyl phosphate are different from those previously reported. Simon and Stalon (1977) determined a K , for L-ornithine of 0.5 mM and a K,,, for carbamoyl phosphate of 0.02 mM. The differences observed in the K , values for L-ornithine and carbamoyl phosphate may be due to the fact that these workers used partially purified enzyme. Observations in this laboratory indicate that the specific activity of yeast OTCase is increased, and the K , for L-ornithine is decreased by the addition of bovine serum albumin to the assay. This phenomenon has also been observed for the catalytic subunit of aspartate transcarbamoylase (Porter et al., 1969).
A rigorous test for homogeneity is equilibrium sedimentation in the analytical ultracentrifuge. From Equation 1, the determination of the molecular weight of a monodisperse protein from its concentration distribution at sedimentation equilibrium requires that its partial specific volume, is, be known. Several methods exist for the determination of the partial specific volumes of proteins. Two methods were utilized here: solvent density perturbation sedimentation (Edelstein and Schachman, 1967), and a method using the partial specific volumes for the individual amino acids (Cohn and Edsall, 1943) which have been determined from the amino acid composition. The density perturbation technique makes use of a linear transform (In c uersus r 2 ) of the primary data from a sedimentation experiment ( e uersus r ) . The slopes of the resulting two straight lines give two equations and two unknowns (Mr and 5) which can be analyzed directly. However, it is always advisable, when possible, to analyze the primary data directly in terms of the desired parameter values since, without proper weighting, the error in the determination of parameters when data is transformed does not directly reflect the error in the primary data (Wilkinson, 1961;Dowd and Riggs, 1965;Cleland, 1979a;Hensley et al., 1981;Klotz, 1982). Consequently, the data from this experiment were analyzed directly in terms of the two exponential equations given in Miniprint (Equations 1 and 2). This requires nonlinear least squares analysis in which the dependent variable is a function of several independent variables. In this case, the dependent variable is protein concentration, which is a function of the independent variables: radius, solvent density (for H20 or D20), and rotor speed (10,006 or 15,006 rpm). This forces the protein concentration distributions to fit a single molecular weight and partial specific volume under four different conditions. If the protein was not a single, monodisperse species, the errors in the parameter values would be large, and the distribution of residuals would be nonrandom. However, in this case, the distribution of residuals is random (Fig. 5). Moreover, the value for the molecular weight is determined to within 2% and the partial specific volume to within 0.8%. Hence, this approach to the determination of molecular weight and partial specific volume, in addition to determining the two parameter values with excellent precision and accuracy, suggests an easily performed test for protein homogeneity and nonideality.
The spectral properties of OTCase were also examined. Fig.  6 shows the ultraviolet absorption spectrum of OTCase. An interesting feature of this spectrum is the unusual fine structure in the area of 280-260 nm. This is due to a relatively high abundance of phenylalanine. The low specific absorbance at 280 nm is due to the low content of tryptophan (l/polypeptide chain) and tyrosine (4/polypeptide chain). The number of tryptophan/polypeptide chain was established using two methods. The second derivative of the ultraviolet absorption spectrum in GdmCl was used to calculate the ratio of aromatic amino acids in OTCase (Levine and Federici, 1982). Results indicated the tryptophan to tyrosine to phenylalanine ratio was 1 to 4 to 17. These results were confirmed by amino acid analysis except in the case of phenylalanine, where amino acid analysis gave 24/chain. Since the molar extinction of phenylalanine is considerably less than tryptophan and tyrosine, it is likely that this optical method would be less precise for phenylalanine than for tryptophan or tyrosine. The number of tryptophan/polypeptide chain was also determined from fluorescence measurements (Pajot, 1976). A value of 0.99 tryptophan/polypeptide chain was calculated in this manner for yeast OTCase. One tryptophan/polypeptide chain is consistent with the value obtained from the ultraviolet second derivative method.
The fluorescence properties of this tryptophan are shown in Fig. 7. The emission maximum of 320 nm indicates tryptophan should be in a hydrophobic environment in the protein. A single tryptophan in yeast OTCase ought to be experimentally very useful. For instance, both the absorption and fluorescence properties of this unique chromophore are altered when the high affinity bisubstrate analogue &N-(phosphonacetyl)-L-ornithine binds at the active site of OTCase (Eisenstein and Hensley, 1983b). Therefore, this tryptophan will be very important in monitoring ligand binding events and potentially for monitoring subsequent conformational changes due to active site ligand binding.
The circular dichroic spectrum of OTCase is shown in Fig.  8. This spectrum has been analyzed in terms of secondary structural classes by two methods. The dotted line corresponds to the fit of the experimental spectrum to reference spectra for the various secondary structures derived from proteins whose crystal structures are known (Chang et al., 1978). The solid line is the fit of the experimental spectrum directly to a linear combination of the CD spectra of 16 reference proteins whose crystal structures are known (Provencher and Glockner, 1981). The latter method appears more successful in minimizing the variance of the fit. Utilizing this method, the secondary structure was determined to be 22% a-helix, 43% @-sheet, 8% @-turn, and 27% random structure. The method of Chang et al. (1978) gave 23% a-helix, 37% P-sheet, 10% Pturn, and 30% random structure.
The amino acid composition of yeast OTCase also has some unusual features. First, as discussed above, there is a single tryptophan/polypeptide chain. Second, there is a unique cysteine. In all OTCases examined to date, there exists 2-3 cysteine/polypeptide chain. Marshall and Cohen (l972,1980a, 1980b, 1980c have shown in bovine liver, in Streptococcus faecalis, and in Streptococcus faecium, one of these sulfhydryl groups is at or near the active site. Studies in this laboratory have demonstrated that reaction of the sulfhydryl group in yeast OTCase with p-chloromercuribenzoate and N-ethylmaleimide completely inhibits enzyme activity, similarly suggesting that this sulfhydryl may be at or near the active site: In order to elucidate directly the molecular basis of the regulation of OTCase activity in the multienzyme complex, both OTCase and arginase are needed in homogenous form and their physical properties known. The results discussed here represent the first step in attaining these goals. The large scale purification and characterization of arginase are currently underway. -