Half of Saccharomyces cerevisiae Carbamoyl Phosphate Synthetase Produces and Channels Carbamoyl Phosphate to the Fused Aspartate Transcarbamoylase Domain*

The first two steps of the de novopyrimidine biosynthetic pathway in Saccharomyces cerevisiaeare catalyzed by a 240-kDa bifunctional protein encoded by theura2 locus. Although the constituent enzymes, carbamoyl phosphate synthetase (CPSase) and aspartate transcarbamoylase (ATCase) function independently, there are interdomain interactions uniquely associated with the multifunctional protein. Both CPSase and ATCase are feedback inhibited by UTP. Moreover, the intermediate carbamoyl phosphate is channeled from the CPSase domain where it is synthesized to the ATCase domain where it is used in the synthesis of carbamoyl aspartate. To better understand these processes, a recombinant plasmid was constructed that encoded a protein lacking the amidotransferase domain and the amino half of the CPSase domain, a 100-kDa chain segment. The truncated complex consisted of the carboxyl half of the CPSase domain fused to the ATCase domain via the pDHO domain, an inactive dihydroorotase homologue that bridges the two functional domains in the native molecule. Not only was the “half CPSase” catalytically active, but it was regulated by UTP to the same extent as the parent molecule. In contrast, the ATCase domain was no longer sensitive to the nucleotide, suggesting that the two catalytic activities are controlled by distinct mechanisms. Most remarkably, isotope dilution and transient time measurements showed that the truncated complex channels carbamoyl phosphate. The overall CPSase-ATCase reaction is much less sensitive than the parent molecule to the ATCase bisubstrate analogue,N-phosphonacetyl-l-aspartate (PALA), providing evidence that the endogenously produced carbamoyl phosphate is sequestered and channeled to the ATCase active site.

alyzes the synthesis of carbamoyl phosphate for both pyrimidine and arginine pathways. In most prokaryotes, a single enzyme produces this metabolite for the two pathways (1)(2)(3). In contrast, eukaryotes possess two specific enzymes providing carbamoyl phosphate; one for aspartate transcarbamoylase (ATCase, EC 2.1.3.2) in the pyrimidine pathway, and the other for ornithine transcarbamoylase (OTCase, EC 2.1.3.3) in the arginine pathway or urea cycle. Although in prokaryotes these enzymes are independent monofunctional proteins (4), some of them are consolidated into multifunctional proteins in eukaryotes (5).
The best known example of such an organization is the mammalian CAD protein that catalyzes the first three reactions of the pyrimidine pathway, thus carrying the CPSase, ATCase, and the dihydroorotase (DHOase, EC 3.5.2.3) activities (6 -8). These enzymatic activities correspond to different protein domains that can be dissociated through mild proteolysis without loss of catalytic activity (9 -11).
CPSase can either use ammonia provided exogenously or endogenous ammonia produced by glutamine hydrolysis (13). The hydrolysis of glutamine requires the presence of a glutamine amidotransferase domain, which is either part of the same polypeptide (eukaryotes) (6 -8, 15-18), or a separate subunit noncovalently associated with the synthetase subunit (prokaryotes) (4,19). The N-and C-terminal regions of the synthetase domain from all organisms examined so far, show a significant degree of sequence homology (15-18, 20 -22), an observation which was interpreted to mean that the gene coding for these enzymes evolved through a process of gene duplication, differentiation, and fusion (20,21). The two domains corresponding to the two halves of the synthetase domain are called CPS.A and CPS.B. Unexpectedly, it was recently discovered (23) that each of these two domains of CAD CPSase are able to independently catalyze the formation of carbamoyl phosphate.
In Saccharomyces cerevisiae, the multifunctional protein that catalyzes the first steps of the pyrimidine pathway, possesses only the CPSase and the ATCase domains (15,24,25). However, it also contains an inactive domain (pDHO) homologous (15) (50) to functional DHOases. The three domains of this protein are linked in the same order as in CAD, that is CPSase, pDHOase, and ATCase (Fig. 1). In terms of allosteric regulation, the yeast complex also shows properties that are intermediary between those of bacteria and mammals (Fig. 1). In the yeast complex, both CPSase and ATCase are sensitive to feedback inhibition by UTP (26,27), whereas in CAD only the CPSase activity responds to this allosteric effector (28). In addition, the CPSase activity of CAD is activated by 5Ј-phosphoribosyl pyrophosphate (29), a metabolite that has no influence on the yeast enzyme. Both multifunctional proteins exhibit channeling of carbamoyl phosphate from the CPSase catalytic site where it is synthesized to that of ATCase where it is used as a substrate (30 -33) for the formation of carbamoyl aspartate.
In this study, a recombinant plasmid encoding a truncated yeast multifunctional protein (C B ApD) lacking the glutamine amidotransferase and CPS.A domains was constructed and expressed in Escherichia coli. Unlike the mammalian constructs thus far studied, the yeast CPS.B domain in this protein remains fused through the pDHO to the ATCase domain. The analysis of its functional properties showed that, indeed the CPS.B domain is enzymatically active and that many of the interdomain interactions responsible for regulation and channeling persist in the construct.

EXPERIMENTAL PROCEDURES
Plasmids and Strains-The yeast pDHO and ATCase coding sequences were inserted into a vector derived from pEK81 (35). The insert is located immediately downstream of the pyrBI promoter, which controls the expression of E. coli ATCase. The plasmid pC4 (4.0-kb) is a derivative of pRS315 (36) containing the ura2 gene (37). The E. coli mutant L673 strain (38), defective in carA and carB, as well as the Lon-protease, was a gift from Dr. Carol Lusty (Public Health Research Institute of the City of New York). The genes carA and carB encode the small and large subunits of E. coli carbamoyl phosphate synthetase, respectively. The yeast strain MD171-1C (a, cpa2, ura3, fur4), isolated by M. Dennis-Duphil (University of Toulouse, France) was derived from the wild-type strains FL-100 and from FL-233-3C (39) and lacks the arginine-specific CPSase.
Recombinant DNA Methods-Transformation and preparation of competent E. coli cells was carried out by the Hanahan procedure (40). DNA fragments were gel-purified following electrophoresis in 0.8% agarose by extraction onto glass beads (Gene Clean kit). Restriction digests, ligations, and other DNA methods were carried out using standard protocols (41).
Cell Growth and Preparation of Cell-free Extracts-Cells harboring the recombinant plasmid were routinely grown from a single colony in 2 ϫ YT medium supplemented with 100 mg/liter ampicillin. For induction of recombinant proteins under control of the pyrBI promoter, the transformed L673 cells were grown in a minimal medium consisting of 6 g/liter Na 2 HPO 4 , 3 g/liter KH 2 PO 4 , 1 g/liter NH 4 Cl, 5 g/liter casamino acids (DIFCO 0230), 4 g/liter glucose, 0.5 mg/liter ZnSO 4 ⅐7H 2 O, 0.5 mg/ml FeSO 4 ⅐7H 2 O, 0.1 mM CaCl 2 , 1 mM MgSO 4 ⅐7H 2 O, 10 mg/liter tryptophan, supplemented with 12 mg/liter uracil and 100 mg/liter ampicillin. Under these conditions, there is sufficient uracil to sustain growth for about 19 -21 h, after which time, growth is slowed and the recombinant protein is expressed. Growth was monitored spectrophotometrically at 600 nm. The cells were harvested in late exponential phase or early stationary phase by centrifugation at 3,000 ϫ g for 30 min in a Centrikon T-124 centrifuge. The cells were resuspended in 50 mM Tris-HCl, pH 7.5, 1 mM EDTA, 1 mM dithiothreitol, 5% glycerol, and disrupted by sonication three times for 1 min on ice, using a Biosonik III sonifier set at 20,000 kHz. The sonicate was cleared by centrifugation at 12,000 ϫ g for 30 min at 4°C. These extracts were dialyzed to eliminate all the metabolites, including nucleotides, which might interfere with enzyme determinations. The yeast strain MD171-1C was grown (32) as described previously.
Construction of the Recombinant Plasmid pSV-C B ApD-The 5.6-kb plasmid pHL-Y12, which encodes the pDHO and ATCase domains of yeast, is expressed under control of the pyrBI promoter (34). The 14.0-kb plasmid pC4-URA2 contains the yeast ura2 gene encoding the bifunctional CPSase-ATCase complex. The cDNA encoding the CPS.B and part of the pDHO domains was amplified by PCR using pC4-URA2 as a template. The 5Ј primer incorporated a NdeI site, whereas the 3Ј primer contained a HpaI site. The primers were designed so that when cleaved with the restriction enzymes, the PCR product could be ligated FIG. 1. Organization and allosteric regulatory properties of the enzymes catalyzing the first reactions of the pyrimidine pathway in prokaryotes and eukaryotes. This scheme shows the functional domains that comprise the proteins that catalyze the initial steps in the de novo pyrimidine biosynthetic pathway, the amidotransferase or glutaminase domain (GLNase), the CPSase synthetase domain consisting of two subdomains (CPS.A and CPS.B), the dihydroorotase domain (DHOase), and ATCase. The activities are associated with separate polypeptide chains in E. coli, whereas in mammals (CAD) all are consolidated on a single multifunctional polypeptide. Mammalian ATCase lacks the regulatory subunit (Reg) found in the E. coli protein. In yeast, the CPSase and ATCase domains are carried by a single polypeptide, but in this case the active DHOase that is encoded by a separate gene is replaced with an inactive DHOase homologue (pDHO). The truncated yeast protein, C B ApD, constructed for this study begins at residue 962 with the sequence MVLGSG. The scheme also shows the allosteric effectors that regulate the activity of these proteins and the locus of regulation represented by arrows.
in frame to the large fragment of pHL-Y12 plasmid cleaved with the same restriction enzymes, resulting in the 7.1-kb recombinant pSV-C B ApD. The PCR insert and junctions were sequenced at the Wayne State University Core Facility by the Sanger dideoxy method using 4 complimentary synthetic oligonucleotides and a double-stranded DNA template.
Protein Methods-The C B ApD protein was partially purified by chromatography of 5 ml of dialyzed cell free extracts on a Q-Sepharose fast-flow column equilibrated with 50 mM Tris-HCl, pH 7.5, 1 mM EDTA, 1 mM dithiothreitol (resuspension buffer). The recombinant C B ApD was eluted with a linear 0 -0.5 M sodium chloride gradient. The ATCase and CPSase activities coeluted and the active fractions were pooled and precipitated with 3.6 M ammonium sulfate. The precipitate was dissolved in 500 l of the resuspension buffer, and directly applied to a Sephacryl S-300HR column equilibrated with the same buffer. The active fractions were pooled and stored at Ϫ20°C. The protein was estimated by scanning stained SDS gels to be 30% pure. This preparation was then used to perform all enzymatic assays.
Protein concentrations were assayed by the Lowry method (42). SDS-gel electrophoresis was carried out on 10% polyacrylamide gels using the Laemmli procedure (43). For analytical gel filtration, the partially purified protein was chromatographed on a 45 ϫ 2-cm calibrated Sephacryl S-300 column equilibrated with 0.05 M Tris-HCl, pH 7.5, 1 mM dithiothreitol, and 5% glycerol. The elution volume was determined by assaying CPSase activity.
Enzyme Assays-The ATCase activity was assayed as described by Denis-Duphil et al. (44). The standard conditions used were 30 mM [ 14 C]aspartate (0.03 Ci/mol), 10 mM carbamoyl phosphate, and 50 mM Tris-HCl, pH 7.5. The assays were conducted at 30°C for 10 min. The CPSase activity of the C B ApD complex was assayed in the presence of 5 g of E. coli ATCase catalytic subunits to efficiently trap all the unstable carbamoyl phosphate formed. The standard conditions used were 50 mM Tris-Ac, pH 7.5, 100 mM KCl, 100 mM NH 4 Cl, 150 mM [ 14 C]sodium bicarbonate (0.168 Ci/mol), 20 mM magnesium acetate, 10 mM ATP, and 50 mM aspartate. The assays were conducted at 30°C for 30 min. It was verified that the E. coli ATCase activity present in the extracts of L673 cells was negligible compared with the activity resulting from the expression of the recombinant. For some experiments, the carbamoyl phosphate formation was measured directly by converting the product to urea (78). The bicarbonate-dependent ATPase activity and the utilization of ATP in the overall reactions were assayed by measuring the time-dependent accumulation of ADP using a coupled enzymatic assay described previously (45,69).
The overall CPSase-ATCase activity of C B ApD complex was assayed using the standard conditions described above but without the addition of E. coli ATCase catalytic subunits. One unit of enzyme activity is defined as the amount of protein that catalyzes the formation of 1 mol of carbamoyl aspartate in 1 h. The sensitivity of CPSase and ATCase to the feedback inhibitor UTP was assayed under the standard conditions described above in the presence of varying concentrations of this effector.
Carbamoyl Phosphate Channeling-The channeling of carbamoyl phosphate from the catalytic site of CPSase to that of ATCase was examined in two ways (31,32). The kinetics of the overall CPSase-ATCase reaction of the truncated yeast protein was measured in the presence of saturating concentrations of CPSase substrates and aspartate as described above. These experiments were designed to determine the transient time (), a measure of the time necessary to reach the steady state phase of the coupled reaction (32,46). Isotopic dilution by exogenously added unlabeled carbamoyl phosphate of the [ 14 C]carbamoyl phosphate synthesized by CPSase was measured as described previously (32). The coupled assays were conducted under the same conditions that were used for assaying CPSase activity. For comparison, channeling was also measured in permeabilized yeast cells, strain MD171-1C, in the presence of saturating ornithine, so that carbamoyl phosphate must diffuse through the aqueous phase to OTCase. Partially purified proteins or E. coli cell extracts would have been ideal controls but the yeast complex has not been expressed in E. coli.

Construction and Expression of the Recombinant Plasmid pSV-C B ApD
The region of the ura2 cDNA sequence extending from nucleotides 3776 to 5763, which encodes the CPS.B domain and part of the pDHO, was amplified by PCR using the plasmid pC4 as a template. The 5Ј primer incorporated a NdeI site, whereas the 3Ј primer had a HpaI site. The PCR product was purified and then cleaved with these two restriction enzymes. The 5.6-kb plasmid pHL-Y12 (34), which encodes the yeast pDHO and ATCase domains, was cleaved with NdeI and HpaI, and the linear fragment was gel-purified and ligated to the PCR product. The fidelity of the construct was confirmed by DNA sequencing of the insert and all of the junctions. The resulting 7.1-kb pSV-C B ApD plasmid was transformed into the E. coli strain L673, a uridine auxotroph that lacks the carA and carB genes encoding the small and large subunits of E. coli CPSase, respectively.
The pSV-C B ApD recombinant plasmid encodes a protein, C B ApD, that possesses the C-terminal half of CPSase (C B ), beginning at methionine 962 (ura2 numbering), linked to the pDHO (pD) and the ATCase (A) domains. E. coli L673 cells were transformed with pSV-C B ApD and grown under various conditions including the presence and absence of uracil and arginine. Remarkably, transformants were able to grow in the absence of these two metabolites, a result which indicated that the cells were able to synthesize carbamoyl phosphate. The strict requirement for these metabolites for the growth of the untransformed L673 strain was verified. These results indicate that C B , the C-terminal half of the CPSase synthetase domain is able to produce carbamoyl phosphate. Indeed, the cell-free extract of the transformed strain exhibits appreciable carbamoyl phosphate synthesizing activity (0.085 mol⅐mg Ϫ1 h Ϫ1 ).
The C B ApD protein, purified as described under "Experimental Procedures," had a specific enzymatic activity of CPSase and ATCase of 0.34 mol⅐mg Ϫ1 h Ϫ1 and 3 mol⅐mg Ϫ1 h Ϫ1 , respectively. The molecular mass of the protein determined under denaturing conditions by electrophoresis on calibrated SDS gels was 136 kDa. The size of the oligomer was estimated by gel filtration to be 814 kDa indicating that the protein is probably a hexamer.
Characterization of the CPSase Activity of C B ApD Initial analysis of cell extracts of the L673 transformants, demonstrated that C B ApD requires (Table I) all of the carbamoyl phosphate synthetase substrates. Because the protein lacks the glutamine amidotransferase domain, the CPSase activity is dependent on the presence of NH 4 Cl as a nitrogen donating substrate. The residual activity, which is observed in the presence of glutamine, probably results from hydrolysis by endogenous glutaminases, which are present in the extract as previously observed (31,47). Because the CPSase activity is measured using the CPSase-ATCase coupled assay in this experiment, aspartate is also required.
The CPSase saturation curves of the partially purified C B ApD recombinant enzyme for all three substrates are shown

Feedback Inhibition by UTP
In the wild-type yeast bifunctional protein, both CPSase and ATCase activities are feedback inhibited by the end product UTP. The influence of this allosteric effector on the activities of C B ApD was investigated, and the results are shown in Fig. 3. The effect of UTP on the CPSase activity of the truncated yeast protein is very similar to that observed for the entire CPSase complex (31,32) when normalized for differences in K m by making measurements at substrate concentrations, which gave similar values of [S]/K m . In contrast, the sensitivity of C B ApD ATCase activity to UTP is totally abolished.

Does C B ApD Function as a Carbamoyl Phosphate
Synthetase or as a Carbamate Kinase?
Some organisms such as Pseudomonas aeruginosa (51,52), Mycoplasma hominis (53), Streptococcus faecium (54), and Streptococcus faecalis (55) possess carbamate kinase, an enzyme belonging to the arginine dihydrolase pathway that can produce ATP and carbamate from carbamoyl phosphate and ADP. The equilibrium constant of this reaction is such that carbamoyl phosphate is generated from ATP and carbamate spontaneously formed by the mixture of bicarbonate and ammonium ions (55). It was thus possible that the truncated CPSase from yeast forms carbamoyl phosphate via a carbamate kinase-like mechanism. In this case, C B ApD would be able to catalyze only the third of the three CPSase partial reactions cited above. This possibility was tested in several ways: Preincubation of Bicarbonate and Ammonium Chloride-Ammonium chloride and bicarbonate react spontaneously, but slowly, to produce carbamate. The chemical equilibrium constant of this reaction (56) at 37°C is shown in Equation 1.
Consequently, the initial rate of the carbamate kinase reaction should be enhanced by preincubation of ammonium chloride and bicarbonate through accumulation of the substrate carbamate. To test this possibility, these two substrates were preincubated for 30 min at 37°C before starting the enzymatic reaction by the addition of enzyme and ATP-Mg 2ϩ . The result of this experiment, clearly showed that the preincubation had no influence on the initial rate of the enzyme-catalyzed reaction. The progress curve obtained with preincubated substrates is virtually the same as that obtained when the substrates were mixed at zero time.
CPSase Inhibitors-The compound, ␣,␤-methylene-ATP is thought to specifically inhibit the first partial reaction catalyzed by CPSase (57), the bicarbonate-dependent ATP hydrolysis. Consequently, the activity of CPSase is inhibited by this compound, whereas that of carbamate kinase should be insensitive to it. The carbamoyl phosphate-synthesizing activity of C B ApD was found to be inhibited greater than 95% at 2 mM concentrations of this ATP analogue (data not shown).
Ap5A and Ap3A are good competitive inhibitors of CPSases (57-59), but the former is known to be the more potent inhibitor perhaps because the distance between the two adenosine rings is longer. We found in this study that Ap5A is indeed a much more efficient inhibitor of C B ApD CPSase activity than Ap3A. However, a recent report (60) questioned the validity of this criterion in distinguishing CPSases and carbamate kinases, and the issue of whether either of these inhibitors are effective toward carbamate kinases needs further study.
Stoichiometry of the Reactions-The hallmark of a true carbamoyl phosphate synthetase is that two ATP molecules are consumed for each mole of carbamoyl phosphate synthesized. In contrast, carbamate kinases use a single ATP to enzymatically phosphorylate carbamate formed in a spontaneous chemical reaction from bicarbonate and ammonia. In the absence of ammonia, C B ApD catalyzes a bicarbonate-dependent ATPase reaction, the first partial reaction in the overall synthetic scheme (see the Introduction), and no carbamoyl phosphate is formed. If the same reaction is carried out in the presence of saturating ammonia, the rate of ATP consumption increases two-fold from 0.33 mol/hr/mg to 0.71 mol/hr/mg. The rate of ATP consumption in the presence of ammonia is two-fold higher than the overall rate of carbamoyl phosphate synthesis, 0.34 mol/hr/mg, strongly suggesting that 2 mol of ATP are utilized for each mole of carbamoyl phosphate synthesized. By comparison (Table II) Reversibility of Carbamoyl Phosphate Synthesis-Another defining characteristic that distinguishes CPSases from carbamate kinases is that coupling of the biosynthetic reaction to the hydrolysis of a second mole of ATP drives the reaction in the thermodynamically unfavorable direction of carbamoyl phosphate formation. In contrast, carbamate kinases are freely reversible. To test the reversibility of the truncated yeast molecule, the formation of carbamoyl phosphate was measured versus the concentration of the protein. The assays were not coupled to ATCase or OTCase, so that carbamoyl phosphate formed accumulated in the reaction mixture. E. coli CPSase and S. faecalis carbamate kinases served as controls. Although the reaction catalyzed by the E. coli protein (Fig. 4) was linear over the entire range of concentrations tested, the carbamate kinase catalyzed formation of carbamoyl phosphate rapidly reached equilibrium and leveled off as the carbamoyl phosphate accumulated, and the rate of the reverse reaction approached that of the forward direction. The progress curve for the synthesis of carbamoyl phosphate by the yeast protein (Fig.  4) was also linear and closely resembles the E. coli CPSase progress curve.
Taken together, these results indicate that the truncated yeast CPSase is acting as a true CPSase and not as a carbamate kinase, which catalyzes only the second ATP-dependent partial reaction.

Channeling of Carbamoyl Phosphate
It was previously shown that the wild-type yeast bifunctional complex exhibits channeling, a process by which the carbamoyl phosphate synthesized at the catalytic site of the CPSase domain is transferred to the catalytic site of the ATCase domain where it is used as a substrate (31,32). This phenomenon would be expected to require a precise structural arrangement of the domains within the complex. Consequently, it was interesting to examine whether the two catalytic domains in C B ApD can still support the channeling of carbamoyl phosphate. Channeling was assessed by the initial reaction rate measurements and isotopic dilution assays as described under "Experimental Procedures." The time course of the CPSase-ATCase coupled reaction (Fig. 5A) in the presence of the CPSase substrates and aspartate, gave a transient time () of 70 s for the C B ApD complex. Although higher than the value of 25 s for the wildtype protein (31), the transient time for the truncated complex is appreciably lower than that expected for independent enzymes catalyzing sequential reactions. For example, in the coupled reaction catalyzed by CPSase and OTCase, enzymes which are not physically associated, carbamoyl phosphate must diffuse through the solvent from one active site to the other. In this instance, the transient time measured (Fig. 5A) under the same experimental conditions (31, 32) is 180 s. Thus, on the basis of this criteria, some channeling of carbamoyl phosphate is still operative in the truncated bifunctional protein. The isotope dilution experiment (Fig. 5B) gave similar results. The observed dilution profile of C B ApD is intermediary between the curves obtained in the case of the wild-type complex and the CPSase-OTCase coupled reaction. This result confirms that the carbamoyl phosphate is partially sequestered within the C B ApD complex. a In this analysis, ATP hydrolysis and carbamoyl phosphate formation were measured in the same reaction mixture and differs from the related experiment that measured bicarbonate-dependent ATPase activity in the presence and absence of NH 3 described in the text.

Effect of PALA on the ATCase Reaction
PALA, a bisubstrate analogue (61), strongly inhibits the binding of carbamoyl phosphate to the ATCase domain. When the ATCase activity of the C B ApD complex was assayed (Fig.  6), using carbamoyl phosphate and aspartate in the presence of PALA, the activity was nearly completely abolished at the lowest concentration, 0.02 mM, of the inhibitor tested. Remarkably, the coupled reaction initiated by the addition of aspartate and the CPSase substrates, was virtually unaffected by concentrations of PALA up to 1 mM. Thus, endogenously synthesized carbamoyl phosphate competes much more effectively with the bisubstrate analogue than does exogenously added carbamoyl phosphate. DISCUSSION The results reported here show that the C-terminal half of the yeast CPSase, C B , a part of the yeast bifunctional complex, is able to catalyze the synthesis of carbamoyl phosphate. Several lines of evidence showed that the recombinant protein functions as a true CPSase not a carbamate kinase. Although the inhibitor studies and the failure to detect a lag associated with formation of carbamate from NH 3 and HCO 3 Ϫ are suggestive, the best evidence is that the biosynthetic reaction is irreversible and requires 2 mol of ATP for each mole of carbamoyl phosphate formed. The isolated CPS.A and CPS.B domains of the homologous mammalian CAD protein (23) are also catalytically active, suggesting that this may be a general feature of CPSases. This molecule described here is unique in that in contrast to the mammalian "half-molecules," the CPS.B domain remains fused to the pDHO and ATCase domains.
The observation that a single 60-kDa domain of both mammalian and yeast carbamoyl phosphate synthetases are fully functional enzymes is interesting in view of recent reports that showed that the archaebacteria, Pyrococcus abyssi (59) and Pyrococcus furiosus (77), possess a CPSase whose size is less than half that of the known CPSases. This surprising discovery lends further credence to the hypothesis (18,21,22,(62)(63)(64) that the two halves of CPSase originated as the consequence of an ancestral gene duplication. The observation that small molecules, like the truncated yeast and mammalian CPSases are functional, suggests that the capacity to synthesize carbamoyl phosphate probably predated the gene duplication and fusion that occurred before the divergence of the eubacterial and eukaryotic lineages.
Although the mammalian CPS.B domain monomer can catalyze both ATP-dependent reactions (23), the homodimer (45) is required for the overall synthesis of carbamoyl phosphate. In showing the rate of citrulline formation in the sequential reactions catalyzed by the pyrimidine-specific CPSase and the separate endogenous OTCase, which has a comparable K m value for carbamoyl phosphate (32). The latter reaction was measured in permeabilized yeast cells (strain MD 171-1C), which lack the arginine-specific CPSase. In this case, the transient time () is 180 s and the steady state concentration of carbamoyl phosphate normalized for differences in the steady state rate of product formation was 10 M. The formation of carbamoyl aspartate by the C B ApD complex (q) and the wild-type yeast CPSase-ATCase complex (E) was assayed (panel B) after initiating the reaction with saturating [ 14 C]bicarbonate, MgATP, ammonium chloride, and aspartate in the presence of increasing concentrations of unlabeled carbamoyl phosphate. As a negative control, the same experiment was carried out with the C B ApD complex, except in this case the assay mixture contained OTCase and saturating ornithine was substituted for aspartate. The formation of citrulline (Ⅺ) by the unlinked OTCase provides a measure of isotopic dilution in a coupled reaction where channeling cannot occur. this model (Fig. 7), the function and juxtaposition of the two monomers in the homodimer is the same as that of CPS.A and CPS.B in the parent molecule. Thus, in the truncated yeast protein, the CPS.B domains are likely to dimerize. Sequence homology (65) and molecular modeling (66) of yeast ATCase suggests that three domains associate to form the active enzyme. Consequently, the yeast C B ApD complex would be expected to be a 3 2 hexamer (Fig. 7) held together by dimeric interactions between the CPS.B domains and trimeric interactions between the ATCase domains. The molecular mass, 870 kDa, as determined by size exclusion chromatography is consistent with a hexameric structure.
This half CPSase still shows apparent positive cooperativity for the utilization of the substrate ATP. The measured Hill coefficient (2.7 Ϯ 0.08) is comparable to that obtained for the parent enzyme (2.5 Ϯ 0.04). The cooperative substrate binding suggests that intersubunit interactions occur within the hexameric complex. The homotropic transitions in the wild-type and truncated complexes may reflect the fact that the overall synthesis of carbamoyl phosphate involves two ATP-dependent partial reactions, which occur on different domains as shown (67)(68)(69)(70)(71)(72)(73)(74) for E. coli CPSase. It is also noteworthy that the S 0.5 value for ATP is about 20 times lower in the truncated CPSase than in the wild-type complex. A similar observation was made in the case of the isolated mammalian CPSase domains (23) suggesting that the catalytic sites may be somewhat constrained in the native CPSase. The x-ray structure of E. coli CPSase (75) suggests a hypothesis in which carbamate formed on CPS.A diffuses through a long narrow tunnel to the active site of CPS.B where it is phosphorylated to form carbamoyl phosphate. Thus, if the tunnel is necessary for carbamoyl phosphate synthesis, a similar passage must be present in the homodimer formed by the C B ApD complex.
UTP inhibited the CPSase activity of the truncated molecule to the same extent as the wild-type species, whereas UTP inhibition of the ATCase was completely abolished. It is interesting to consider the differential UTP sensitivity of the functional domains in the truncated complex in view of the regulatory properties (37) of a series of 16 yeast CPSase-ATCase regulatory mutants that showed the same behavior. The CP-Sase activity remained sensitive to UTP, whereas the ATCase no longer responded to the effector. Similarly, a series of hamster-yeast CPSase-pDHO-ATCase chimeras exhibited the same regulatory properties (34). The differential loss of the CPSase and ATCase regulation suggests that the mechanisms of feedback inhibition of the two activities might be of a different nature. The feedback inhibition of the ATCase activity may be a genuine allosteric process analogous to that operative in E. coli ATCase in which the regulatory signal is transmitted from a distal regulatory site to the catalytic site. This type of long range signal transmission typically involves changes in interdomain interactions that may be impaired in the truncated complex. In contrast, the influence of UTP on the CPSase activity may result from a more direct interaction with the CPS.B catalytic domain, B2. Kinetic studies (76) suggest that the major locus of regulation of E. coli CPSase is the catalytic site on CPS.B. In this regard it is interesting to note that the binding site for the feedback inhibitor, UMP, tentatively identified in the crystallographic structure of E. coli CPSase (75) is relatively close to the catalytic site. A phosphate ion thought to bind to the UMP allosteric site on the B3 regulatory subdomain is 23 Å from the ATP binding site on the B2 catalytic subdomain and only about 14 Å from the B2 B3 interface.
Remarkably, the truncated yeast bifunctional complex still supports channeling of carbamoyl phosphate from the CPS.B active site to that of the ATCase, although the process is not as efficient as that observed in the native yeast complex. Both the transient time measurements and the isotope dilution experiments showed that the extent of channeling by the C B ApD complex is intermediate between the tightly coupled yeast complex and the sequential reactions catalyzed by CPSase and OTCase in which the intermediate is not sequestered. Because the channeling of carbamoyl phosphate is probably a consequence of the close juxtaposition of the CPS.B and ATCase catalytic sites (Fig. 7), the persistence of channeling suggests FIG. 6. The Effect of PALA on the activity of the ATCase domain of the C B ApD complex. The effect of the bisubstrate analogue, N-phosphoacetyl-L-aspartate (PALA) on the ATCase reaction catalyzed by C B ApD was determined by assaying with saturating carbamoyl phosphate (10 mM) and aspartate (40 mM) (E) as described under "Experimental Procedures." PALA inhibition of carbamoyl aspartate formation in the coupled reaction, initiated with saturating bicarbonate, MgATP, ammonium chloride, and aspartate (q) was also measured. that the three-dimensional structure of the truncated complex is such that the interactions of the carboxyl half of the CPSase and the ATCase domain still occur. This result suggests that in the wild-type complex, whose three-dimensional structure is still unknown, the C-terminal domain of CPSase interacts with ATCase.
The effect of PALA on the ATCase activity of the C B ApD complex also strongly supports the contention that carbamoyl phosphate is sequestered within the truncated yeast complex. PALA proved to be a potent ATCase inhibitor when the activity was measured directly in the presence of PALA. In contrast, the ATCase activity of the complex was largely unaffected when the substrate carbamoyl phosphate was generated endogenously by the CPSase domain of the truncated yeast protein. This observation is especially striking, because the calculated steady state concentration of carbamoyl phosphate in the system was determined by transient time measurements to be only 0.2 M in the coupled assay, compared with the concentration of 10 mM carbamoyl phosphate used when the ATCase activity was directly assayed. Irvine and Carrey (33) also observed decreased PALA inhibition of the coupled reaction catalyzed by the mammalian CAD complex. The simplest explanation of this result is that the effective concentration of carbamoyl phosphate in the vicinity of the active site is much higher in the coupled reaction as a consequence of channeling, so that the intermediate competes much more effectively with PALA.