Cell Cycle-dependent Regulation of Pyrimidine Biosynthesis*

De novo pyrimidine biosynthesis is activated in proliferating cells in response to an increased demand for nucleotides needed for DNA synthesis. The pyrimidine biosynthetic pathway in baby hamster kidney cells, synchronized by serum deprivation, was found to be up-regulated 1.9-fold during S phase and subsequently down-regulated as the cells progressed through the cycle. The nucleotide pools were depleted by serum starvation and were not replenished during the first round of cell division, suggesting that the rate of utilization of the newly synthesized nucleotides closely matched their rate of formation. The activation and subsequent down-regulation of the pathway can be attributed to altered allosteric regulation of the carbamoyl-phosphate synthetase activity of CAD (carbamoyl-phosphate synthetase-aspartate carbamoyltransferase-dihydroorotase), a multifunctional protein that initiates mammalian pyrimidine biosynthesis. As the culture approached S-phase there was an increased sensitivity to the allosteric activator, 5-phosphoribosyl-1-pyrophosphate, and a loss of UTP inhibition, changes that were reversed when cells emerged from S phase. The allosteric regulation of CAD is known to be modulated by MAP kinase (MAPK) and protein kinase A (PKA)-mediated phosphorylations as well as by autophosphorylation. CAD was found to be fully autophosphorylated in the synchronized cells, but the level remained invariant throughout the cycle. Although the MAPK activity increased early in G1, the phosphorylation of the CAD MAPK site was delayed until just before the onset of S phase, probably due to antagonistic phosphorylation by PKA that persisted until late G1. Once activated, pyrimidine biosynthesis remained elevated until rephosphorylation of CAD by PKA and dephosphorylation of the CAD MAPK site late in S phase. Thus, the cell cycle-dependent regulation of pyrimidine biosynthesis results from the sequential phosphorylation and dephosphorylation of CAD under the control of two important signaling cascades.

The activity of the de novo pyrimidine biosynthetic pathway closely parallels the growth rate of the cell and is highest during periods of rapid proliferation (1)(2)(3)(4)(5)(6)(7)(8)(9)(10)(11)(12)(13). Cells growing in culture synthesize pyrimidines at a rapid rate during the ex-ponential growth phase, but the pathway is much less active when the cells enter the stationary phase or have their growth arrested by serum deprivation (14 -18). In an earlier study, Mitchell and Hoogenraad (19) show that pyrimidine biosynthesis is maximally activated during the S phase of the cell cycle. Thus, the de novo biosynthetic pathway plays a dominant role in providing precursors for the synthesis of DNA that accompanies cell division. Furthermore, there is strong evidence (7,11,12) that the activation of the pathway is a prerequisite for the proliferation of tumor and neoplastic cells.
The flux through the pathway is governed (20,21) by the activity of carbamoyl-phosphate synthetase II (CPSase), 1 a component of CAD ( Fig. 1), the multifunctional protein (22)(23)(24) that initiates pyrimidine biosynthesis in mammalian cells. CP-Sase catalyzes the rate-limiting step, the formation of carbamoyl phosphate from bicarbonate, glutamine, and two ATP molecules. The enzyme is subject to a complex network of control mechanisms that include both allosteric regulation and phosphorylation. UTP, an end product of the pathway, is a feedback inhibitor, whereas 5-phosphoribosyl-1-pyrophosphate (PRPP), a substrate for both the parallel de novo purine biosynthetic pathway and a subsequent step in pyrimidine biosynthesis, is an activator (13,25,26). Carrey et al. (28)(29)(30)(31) made the important observation that purified CAD is phosphorylated at two sites ( Fig. 1) by cAMP-dependent protein kinase A (PKA). PKA phosphorylation has no effect on the catalytic activity but decreases UTP inhibition, a modification that would be expected to stimulate pyrimidine biosynthesis. However, subsequent studies (32) showed that PKA phosphorylation resulted in a decrease in the affinity for the activator, PRPP, an effect that would be expected to down-regulate the pathway. Thus, PKA phosphorylation alone cannot account for the activation of the pathway in rapidly growing cells.
MAP kinases (33) are ubiquitous components of the mitogenactivated cascade that result in cellular proliferation in response to growth factors. Stimulation of quiescent cells with epidermal growth factor results in the activation of the MAP kinase cascade and the phosphorylation of Thr-456 in the A1 subdomain ( Fig. 1) of the CAD CPSase (34). MAP kinase phos-phorylation both in vivo and in vitro has no effect on any of the catalytic activities of CAD but relieves UTP inhibition and stimulates PRPP activation of the CPSase. Both of these changes in allosteric regulation would be expected to increase the rate of pyrimidine biosynthesis.
We have recently shown (18) that the activation of pyrimidine biosynthesis in exponentially growing BHK 165-23 cells is a consequence of the MAP kinase-mediated phosphorylation of CAD and that its subsequent down-regulation as the culture became confluent is associated with dephosphorylation of the MAP kinase site and phosphorylation by PKA. To more precisely define the mechanism of growth-related activation and down-regulation of pyrimidine biosynthesis, the activity of the pathway and the role of the diverse control mechanisms was examined as BHK 165-23 cells traversed the cell cycle.

EXPERIMENTAL PROCEDURES
Cell Culture-BHK 165-23 (35), a baby hamster kidney cell line derived from BHK-21, was grown in Dulbecco's modified Eagle's medium (DMEM, Invitrogen) supplemented with 10% dialyzed fetal bovine serum and 2 g/ml gentamicin (Invitrogen). Two million cells were plated in T75 flasks containing 25 ml of media. The medium was changed every 2 days. Cells were counted using a hemocytometer, and viability was assessed by trypan blue staining.
Cell Synchronization-BHK cells were synchronized by serum deprivation as described previously (36). Briefly, 10 T-75 flasks of exponentially growing cells (60 -80% confluent) were serum-starved in DMEM, 0.1% fetal bovine serum for 24 h. The cells were then stimulated by the addition of 25 ml of DMEM supplemented with 10% fetal bovine serum and incubated at 37°C. The cultures were harvested at 2-h intervals over a period of 20 -24 h for pyrimidine biosynthesis assays and the preparation of cell extracts as previously described (18).
Pyrimidine Biosynthesis-The rate of pyrimidine biosynthesis in live cells was assayed using a modification of the methods devised by Huisman et al. (1) and Simmonds and co-workers (37). Cells were harvested by trypsin treatment and suspended at a concentration of 2.5 ϫ 10 6 cells/ml in 2.8 ml of DMEM, F-12 medium without bicarbonate. The assay was initiated by the addition of 0.33 ml of 49 mM 14 C-labeled sodium bicarbonate (3.75 Ci/mol). The final concentration of bicarbonate in the assay was 5.2 mM. After allowing the reaction to proceed with gentle shaking for 45 min at 37°C the medium was removed, and the reaction was quenched by the addition of 0.375 ml of 6% cold trichloroacetic acid. The incorporation of 14 C-labeled bicarbonate into UTP and CTP was immediately quantitated by HPLC. We found that pyrimidine biosynthesis proceeded linearly for at least 1 h.
HPLC Analysis of Nucleotides-The nucleotides were analyzed following the protocol of Dr. Melinda E. Wales, 2 using a Waters Model 2695 HPLC system equipped with an IN/US systems ␤-ram, Model 3 radioactivity flow detector. All reagents were ultrapure or HPLC grade. The trichloroacetic acid supernatant was extracted twice with an equal volume of freshly prepared 0.7 M tri-N-octylamine/Freon solution (376 l of 93% w/v tri-N-octylamine (Sigma) diluted to a total volume of 5 ml with freon (1,1,2-trichlorotrifluoroethane (Sigma)). The mixture was vortexed for 1 min, placed on ice for 30 min, and then centrifuged for 1 min at 10,000 rpm. The aqueous phase (100 l) was then immediately injected onto a Whatman Partisil 10 SAX column (250 ϫ 4.6 mm (Phenomenex)) equilibrated with 5 mM ammonium phosphate buffer at pH 2.7 (Buffer A). The nucleotides were eluted with a discontinuous gradient of 5 mM (NH 4 ) 3 PO 4 , pH 2.7 (Buffer A), and 750 mM (NH 4 ) 3 PO 4 , pH 3.8 (Buffer B), at a flow rate of 2 ml/min. Buffer B increased linearly to 50% in 20 min and then from 50 to 100% over the next 10 min. The absorbance at 254 and 280 nm and the radioactivity of the effluent were monitored continuously. The peaks were integrated using the IN/US systems software. The identity and total amount of each nucleotide were calculated from standard curves obtained using authentic standards. The amount of the radiolabeled nucleotides were calculated from the specific radioactivity of the 14 C-labeled sodium bicarbonate.
Enzyme Assays-The glutamine-dependent CPSase assay has been described (39). The 1-ml assay mixture contained 100 g of protein, 100 mM Tris-HCl, pH 8.0, 100 mM KCl, 7.5% Me 2 SO, 2.5% glycerol, 1 mM dithiothreitol, 3.5 mM glutamine, 20.2 mM aspartate, 1.5 mM ATP, 3.5 mM MgCl 2 , and 5 mM sodium 14 C-labeled bicarbonate (1.6 ϫ 10 6 Ci/ mol). The concentration of MgCl 2 was adjusted so as to maintain a 2 mM excess over the sum of the concentration of ATP, UTP, and PRPP. Aspartate transcarbamoylase activity was assayed by the colorimetric method (40). The aspartate transcarbamoylase assay mixture contained 100 g of protein, 5 mM carbamoyl phosphate, and 12 mM aspartate in a buffer consisting of 100 mM Tris-HCl, pH 8.0, 100 mM KCl, 7.5% Me 2 SO, 2.5% glycerol, and 1 mM dithiothreitol in a total volume of 1 ml. The MAP kinase (p44/p42 or Erk1/Erk2) activities were assayed using the New England Biolabs kit following the manufacturer's protocol. The phosphorylation of the MAP kinase substrate, glutathione S-transferase-Elk1 (2 g), was followed by Western blotting using specific anti-phospho-Elk1 antibodies (18). Alternatively, Erk1/Erk2 were assayed using phospho-Erk1/2 antibodies that recognize only the activated kinases (Promega). The PKA activity in the cell extract (20 -150 g of protein) was assayed using the PepTag PKA assay kit (Promega), which relies on a change in the electrophoretic mobility of a fluorescent Leu-Arg-Arg-Ala-Ser-Leu-Gly (Kemptide) derivative upon phosphorylation. Phosphorylation of the Kemptide could also be measured by electrophoresis on a 20% SDS-PAGE followed by immunoblotting using anti-phospho-serine antibodies.
CAD Phosphorylation-Previous studies show (18,41) that the extent of the phosphorylation of the MAP kinase site, Thr-456, could be quantitated by immunoblotting with phosphothreonine antibodies corrected for autophosphorylation. Similarly, the extent of phosphorylation of the CAD PKA sites (Ser-1406 and Ser-1859) could be quantitated using the phosphoserine antibodies. Autophosphorylation was measured by a previously described modification (41) of the "on membrane kinase assay" (42). The increase in phosphorylation of CAD, determined by quantitative immunoblotting with phosphothreonine antibodies, that occurred upon incubation of the protein with 0.5 mM ATP for 2 h was taken as a measure of the unmodified autophosphorylation site of CAD in cell extracts. Purified CAD incubated with and without ATP was used as a standard to determine the fractional autophosphorylation. Protein quantitation was performed by the Lowry method (43) and by scanning stained SDS-polyacrylamide gels calibrated with known amounts of bovine serum albumin as a standard using the software UNSCAN-IT (Silk Scientific Corp.). 2 M. E. Wales, personal communication.

FIG. 1. Structural organization and regulation of CAD.
The first three steps in the de novo pyrimidine biosynthetic pathway are catalyzed by the CPSase, aspartate transcarbamoylase (ATC), and dihydroorotase (DHO) domains of the multifunctional protein CAD, a complex consisting of six copies of a 243-kDa polypeptide organized into multiple domains and subdomains. The synthesis of carbamoyl phosphate involves four partial reactions that are catalyzed by the concerted action of the glutaminase domain (GLN) and two synthetase subdomains (CPS.A and CPS.B). The CPSase activity of CAD controls the flux through the pyrimidine pathway. The allosteric ligands, UTP, the inhibitor, and PRPP, the activator, bind to the B3 regulatory and by autophosphorylation (Autophos) subdomain of CPS.B. The CPSase activity is also controlled by phosphorylation of the MAPK and PKA1 sites. There is a second PKA target site (PKA2) unrelated to regulation. The formation of carbamoyl aspartate and dihydroorotase is catalyzed by the ATC and DHO domains, respectively.
Cloning of the Fluorescent Fusion Proteins-A commercial PKA clone (GeneStorm human PKA catalytic subunit from ResGen) was amplified by PCR using the primers 5Ј-taaaaggtaccatggggaacgcggcgaccgccaagaaaggc-3Ј, incorporating a KpnI restriction site upstream of the first codon, and 5Ј-tttttgatcatcaaaattcaccaaattcttttgcac-3Ј, incorporating a BclI restriction site downstream of the coding sequence. The purified PCR product was cleaved with KpnI and BclI (Invitrogen). The vector pCruz-GFP (Santa Cruz Biotechnology) was digested using KpnI and BglII and treated with thermosensitive alkaline phosphatase (Invitrogen). The purified fragments were ligated and transformed into Escherichia coli DH5␣ (Invitrogen). The plasmid encoded the PKA catalytic subunit with the green fluorescent protein appended to the amino end. The plasmid Npt7.5 Erk2 containing the murine Erk2 cDNA fragment, a generous gift of Dr. Melanie Cobb (Dept. of Pharmacology, University of Texas Southwestern Medical Center, Dallas, TX), was digested with NcoI and filled in with Klenow (Invitrogen). The purified fragment was digested with XbaI (New England Biolabs) and repurified. The vector pECFP-C1 (Clontech) was cleaved with BspE1, filled in with Klenow, digested with XbaI, and dephosphorylated with thermosensitive alkaline phosphatase. The vector and insert were ligated and transformed into DH5␣ cells. The construct encodes MAPK (Erk2) with the enhanced cyan fluorescent protein fused to the amino end. To construct the glutaminase (GLN) CPS fusion protein, pCK-CAD10 (39) encoding the full-length CAD cDNA in E. coli was cleaved with BamHI (a unique site introduced at the 3Ј end of the coding sequence) and BbsI for 2 h at 37°C. The resulting fragment encoding the entire GLN-CPSase domain ( Fig. 1) was purified, filled in with Klenow, and inserted into the pECFP-C1 vector (Clontech) that had been linearized by cleavage with BspE1 and filled-in. The ligation products were transformed into DH5␣-subcloning efficiency bacteria, and recombinants were verified by restriction digests and sequencing. The construct, pECFP-GLN-CPS, encoded CAD GLN-CPS with the enhanced cyan fluorescent protein fused to the amino end. For the construction of the full-length CAD fusion protein, pCK-CAD10 was cleaved with XbaI. The 3750-bp fragment containing the end of the CPS.B domain and the remainder of the CAD-coding sequences was inserted into pECFP-GLN-CPS that had been cleaved with XbaI. XbaI cleaves pECFP-GLN-CPS in the CPS.B domain and in the vector downstream of the insert. The ligation products were transformed into DH5␣. The construct, pECFP-CAD, encoded the entire CAD protein with the enhanced cyan fluorescent protein appended to the amino end. The fidelity of all constructs was verified by DNA sequencing (Wayne State University DNA Sequencing Facility).
Transfection and Fluorescence Microscopy-The E. coli transformants were cultured overnight in Luria broth, 100 g/ml kanamycin, and plasmids were purified using the Plasmid Maxi kit (Qiagen). BHK cells were transfected with LipofectAMINE Plus reagents (Invitrogen) using 2 g of the DNA following the manufacturer's protocol. After 5 h, the transfection medium was replaced with DMEM complete medium. The expression of the recombinant proteins was monitored by fluorescence microscopy and Western blotting using antibodies directed against, CAD, PKA, MAPK (p44/p42), and green or cyan fluorescent protein. The recombinant proteins were fluorescent and had the expected size on calibrated SDS gels. Fluorescence microscopy was carried out using a Zeiss Axiophot Upright Epifluorescent Digital Microscope (Karmanos Confocal Imaging Facility). With the available filters, it was not possible to detect the cyan chromophore at its maximum emission wavelength, but there is a significant secondary maximum that allowed its visualization with the green filters set. Photographs were obtained using a SPOT RT camera (Diagnostic Instruments, Inc.) coupled to a Macintosh Personal Computer with the SPOT software Version 3.0). The expression of cyclin A and cyclin E in the transfectants was visualized by fluorescence microscopy after immunostaining with cyclinspecific antibodies (44).

RESULTS
Cell Synchronization-Serum deprivation following well established protocols (36) was used to synchronize cultures of BHK165-23 cells. Replacing the complete media in exponentially growing cultures with the same media supplemented with only 0.1% serum resulted in the arrest of growth. Growth resumed 24 h later upon stimulation with 10% fetal bovine serum, at which point the cells began to traverse the cell cycle in synchrony. The periodic expression of cyclins is commonly used to identify the different cell cycle phases (45)(46)(47). In this study, cell cycle-dependent markers were used to monitor the progression of the culture through the initial phases of the cycle. Cyclin E expression, known to be confined to the G 1 phase, began to increase soon after serum stimulation, peaked at 4 h, and then declined. The inactivation of the retinoblastoma tumor suppressor protein by hyperphosphorylation is also known to occur during G 1 (46) and to remain phosphorylated through mitosis. In the synchronized BHK cultures, retinoblastoma phosphorylation preceded the expression of cyclin E, peaked at about 3 h, remained phosphorylated for 16 h, and then decreased. Cyclin A expression, reported to begin at the G 1 /S boundary and to reach the highest levels in late S phase (46,47), peaked at 12 h in BHK cells, whereas cyclin B, another S phase marker, peaked at about 13 h (Fig. 2). We concluded that the midpoint of the S phase of the cycle occurred at 10 -11 h after serum stimulation.
Pyrimidine Biosynthesis-The activity of the pyrimidine biosynthetic pathway was assayed by measuring the incorporation of 14 C-labeled bicarbonate into UTP and CTP. The basal level of pyrimidine biosynthesis, measured during serum deprivation (Fig. 3, 0 h) was 1.0 pmol/min/10 6 cells. The level remained constant until just before entry into the S phase of the cell cycle, at which point the rate increased to 1.85 pmol/ min/10 6 cells. The rate of pyrimidine biosynthesis decreased as the cells emerged from S phase. Thus, there was a strong correlation between the activity and the S phase of the cell cycle. Radiolabeled bicarbonate is also a precursor of purines, and there was a parallel increase in the incorporation of 14 Clabeled bicarbonate into ATP from 1.6 to 3.0 pmol/min/10 6 cells in mid-S phase.
The Intracellular Concentration of the Nucleotides-The intracellular concentration of the nucleotides (Table I) was measured by HPLC as the cells traversed the cell cycle. The absorbance at 254 nm was monitored (Fig. 4), and the amount of each nucleotide was determined from the area of the corresponding peaks using standard curves. The concentration was calculated assuming 1 l of intracellular fluid/10 6 packed cells (48). Although most of the nucleotides were well separated (Fig. 4), in the best cases CTP appeared as a small leading shoulder of the larger UTP peak. Because it proved impractical to reliably distinguish these species, the pyrimidine trinucleotide concentrations were expressed as a combined UTP and CTP.
In serum-starved cells (t ϭ 0 h), the concentration of UTP/CTP was 0.17 mM and remained reasonably constant for the duration of the cell cycle. This value was approximately half that observed in exponentially growing cultures of the same cells. A similar trend was observed for most of the other nucleotides. Notably, the ATP concentration in the synchronized cells was 5-fold lower than the value measured in the exponentially growing cells and did not significantly increase for 18 h post-serum stimulation. The observation that the nucleotide pools did not expand during the first round of cell division despite ongoing pyrimidine biosynthesis suggests that the newly synthesized nucleotides are utilized as rapidly as they are produced.
CAD Activity and Regulation-The activity of CAD CPSase, measured in the absence of allosteric ligands (Fig. 5A), increased ϳ2-fold from 1.2 nmol/min immediately after serum stimulation to 2.9 nmol/min at 22 h. We have shown previously (41) that when purified CAD is incubated with ATP, the CPS domain undergoes autophosphorylation, a phenomenon that is accompanied by a 2-fold increase in the CPSase-specific enzymatic activity. The CAD autophosphorylation, measured in synchronized cultures of BHK cells (Fig. 5A), was found to be 90 -95% complete and remained unchanged throughout the cell cycle. Thus, the increase in CAD CPSase activity cannot be attributed to an increase in autophosphorylation. Moreover, neither PKA nor MAP kinase phosphorylation affects the CPS-   4. Determination of intracellular nucleotide concentrations by HPLC. A 100-l aliquot corresponding to 5.6 ϫ 10 6 cells of an extract from an exponentially growing culture was fractionated on a Whatman Partisil 10 SAX column eluted with a discontinuous gradient of ammonium phosphate as described under "Experimental Procedures." The absorbance of the eluant was monitored at 254 nm, and the nucleotides were identified by chromatography of authentic standards. With the exception of UTP and CTP, which coeluted, all of the nucleotide tri-and diphosphates were well resolved. Several peaks in the profile remained unidentified. The nucleotide concentration was determined from the peak areas using standard curves. ase activity in the absence of allosteric ligands. Thus, it is likely that the approximate 2-fold increase in CPSase activity is simply a consequence of the increase in the CAD concentration in the cell that must occur during each turn of the cell cycle. This interpretation is supported by assaying aspartate transcarbamoylase, an unregulated CAD activity, which was also found to increase 2-fold by the end of the cell cycle.
These changes in CPSase activity cannot explain the activation of the pathway that occurred during the S phase of the cell cycle. For example, between 6 and 10 h, where pyrimidine biosynthesis was up-regulated by 85%, the CPSase activity increased only 5%. However, there were appreciable changes (Fig. 5B, Table II) in the allosteric properties of CPSase. As the cells approached S phase, the activation of the enzyme by 200 M PRPP increased about 2-fold, from 1240 to 2524%. In serum-starved cells, UTP (0.5 mM) inhibition of CPSase activity was 40%, but as the cells approached S phase, the effect of UTP was completely abolished. Both of these changes in allosteric regulation, which would be expected to increase the rate of pyrimidine biosynthesis, were fully or partially reversed by 14 h post-serum stimulation. Thus, the changes in the sensitivity of CPSase to PRPP and UTP closely paralleled the S phase and the activation of the pyrimidine biosynthetic pathway.
Kinase Activities and CAD Phosphorylation State-The MAP kinase activity (Fig. 6A) was lowest in early G 1 and increased 15-fold during the first 4 h after serum stimulation. The activation of MAP kinase persisted for several hours and then began to decrease just before entry into the S phase of the cell cycle. The levels remained at ϳ50% of the maximum value up to 20 h. Despite high levels of MAP kinase activity during G 1 , the CAD phosphothreonine remained unchanged until just before S phase and then increased 2-fold. The phosphothreonine antibodies also detect phosphorylation of the single CAD autophosphorylation site (Thr-1037), and this site was found to be nearly fully phosphorylated throughout the cell cycle. Because there was a background phosphorylation of 1 mol of phosphothreonine, a 2-fold increase can be attributed to the incorporation of 1 mol of phosphate into the single CAD MAP kinase site (Thr-456). This interpretation was reinforced by the observation that exposure of cells 10 h post-serum stimulation (Fig.  6C), at which point the CAD MAP kinase site was 40% modified, to 20 M PD98059 or UO126, two MEK (MAP kinase kinase) inhibitors, for 30 min resulted in a 35 and 40% reduction in phosphothreonine, respectively. These inhibitors had no effect on CAD phosphoserine. Similar results were obtained for cells harvested at other time points.
The PKA activity (Fig. 5B) was highest immediately after serum stimulation and began to decrease to 50% of the maximum value over the next 6 h. The inactivation of PKA was accompanied by a simultaneous decline in CAD Ser phosphorylation. Exposure of cells 2.5 h after serum stimulation to the PKA inhibitor, H89 (Fig. 6C), reduced the CAD serine phosphorylation by another 25%, with no effect on CAD phosphothreonine. The levels remained low until 16 h, at which point both the PKA activity and the PKA phosphorylation of CAD began to increase once again.

Intracellular Localization of the Kinases and CAD-For
phosphorylation to occur, the kinase and its substrate CAD must be co-localized within the same intracellular compartment. Plasmids were constructed encoding fusion proteins that had the green or cyan fluorescent protein appended to the amino  Previous studies (49) show that the transfection procedure results in synchronization of the subpopulation of cells that have taken up the plasmid. MAP kinase was found to be primarily localized in the nucleus, although an appreciable amount was cytoplasmic throughout the time course. By contrast, PKA was initially localized in the nucleus but was subsequently translocated into the cytoplasmic compartment. Interestingly, a significant fraction of CAD, initially cytoplasmic, had migrated into the nucleus by 24 h. A similar distribution was observed for the GLN CPS domain, indicating that the putative nuclear targeting signals are located within this region of CAD. Immunofluorescence microscopy (not shown) showed that the high levels of cyclin E were expressed at 0 and 12 h, whereas the levels were much lower in the 24-h cultures. These results suggested that the 0and 12-h samples represent G 0 /G 1 and that by 24 h, the transfected subpopulation had entered the S phase of the cell cycle. DISCUSSION The results reported here are in agreement with previous studies of synchronously growing cells showing that the de novo pyrimidine biosynthetic pathway is activated just before S phase and subsequently down-regulated at the transition between S and G 2 (Fig. 8). We found that the rate increased 1.9-fold from 1.0 pmol/min/10 6 cells in G 1 to 1.85 pmol/min/10 6 cells in mid-S phase. These rates are in good agreement with the rate of pyrimidine biosynthesis of 0.5 pmol/min/10 6 cells measured in asynchronous cultures of the same cells during the first 24 h after passage but are significantly lower than the rate of 4.2 pmol/min/ 10 6 cells measured in fully established, exponentially growing cells (18). The lower rate of pyrimidine biosynthesis after serum starvation can probably be attributed to the low intracellular concentration of the CPSase substrate, ATP.
The average values for the intracellular nucleotide triphosphates in mammalian cells, compiled by Traut in an exhaustive literature search (50), were found to be 567 Ϯ 460, 278 Ϯ 242, 3152 Ϯ 1698, and 468 Ϯ 224 M for UTP, CTP, ATP, and GTP, respectively. The nucleotide triphosphate concentrations, measured by HPLC analysis of extracts of exponentially grow-ing BHK cells (Table I), were for the most part within the expected range. It is noteworthy, that the nucleotide concentrations after stimulation of serum-deprived cells were appreciably lower that the values in exponentially growing cultures. Moreover, the nucleotide pools did not appreciably expand during the course of the cell cycle. Only modest increases were observed even after 20 h of post-serum stimulation (Table I). These results suggest that 1) the increased rate of pyrimidine biosynthesis nearly matches the rate of utilization of nucleotides for the synthesis of macromolecules during the S phase of the cell cycle, and 2) several cycles of cell division are apparently required before the nucleotide pools can be fully replenished after serum starvation. ATP, for example, was found to be 0.6 mM in serum-starved cells compared with a value of 2.9 mM obtained in fully established cultures. The low ATP concentration would be expected to result in an ϳ5-fold decrease 3 in CPSase activity in serum-starved cells, in reasonable agree- 3 The CPSase activity in the cell was calculated at the ATP concentrations of 0.4 and 2.9 mM using the Hill equation assuming a K m of 3.4 mM and a Hill coefficient of 1.3, values measured (41) for purified CAD in the absence of allosteric effectors. Because the kinetic parameters are dependent on the concentration of effectors, the calculation is only approximate. ment with the 7-fold decrease in the flux through the pyrimidine pathway compared with the rate in exponential cultures.
As described above, the increase in pyrimidine biosynthesis during the S-phase of the cell cycle cannot be attributed to an increase in the CPSase activity measured in the absence of allosteric effectors or to autophosphorylation of CAD. The CAD autophosphorylation site was nearly completely phosphorylated in vivo, representing a constant phosphothreonine background throughout the cell cycle. However, previous studies show (18) that the CPSase activity of CAD, and consequently the rate of pyrimidine biosynthesis, is controlled by PKA-and MAP kinasemediated phosphorylation of the CAD CPSase domain.
The activity of the pyrimidine biosynthetic pathway was up-regulated as the cells entered the S phase, reached a maximum at 11 h, and then decreased as the cells exited the S phase. The increase in the pyrimidine biosynthetic activity during S phase closely paralleled the response of CPSase to allosteric effectors. During S phase, UTP inhibition was abolished, and PRPP activation increased 2-fold. This altered response to allosteric effectors could readily account for the 1.9fold activation of the pathway.
Similar changes in allosteric regulation have been observed in asynchronous cultures in response to MAP kinase phosphorylation (18). In synchronized BHK cells, the MAP kinase activity increased rapidly and peaked at 5 h but remained elevated for the duration of the cycle. However, the phosphorylation of the MAP kinase site of CAD did not correlate with the highest MAP kinase activities, and no significant phosphorylation of this site could be detected for 7 h. The resistance of CAD to MAP kinase phosphorylation can probably be attributed to PKA phosphorylation since we have found 4 that prior phosphorylation of CAD by PKA blocks phosphorylation by MAP kinase. CAD serine phosphorylation closely paralleled the activity of PKA, highest initially after serum stimulation and decreasing during the next 7 h. A decrease in PKA activity has been previously shown to be a prerequisite for entry into S phase (44). These results suggest that the activating MAP kinase phosphorylation of CAD is delayed until the PKA site is dephosphorylated. Another factor that could potentially influence the phosphorylation of CAD may be its intracellular location relative to the kinases. MAP kinase was found by fluorescence microscopy to be distributed throughout the cell but was concentrated in the nucleus during G 1 and S. In contrast, PKA was initially localized in the nucleus but is translocated into the cytosol late in G 1 . CAD was cytosolic during G 1 and migrated into the nucleus, where there were high levels of MAP kinase and low levels of PKA during S phase. Experiments are currently under way to follow up on these interesting observations.
The down-regulation of the pathway occurred as cells emerged from the S phase at about 14 -15 h post-serum stimulation. At this point, the MAP kinase activity remained elevated, and the CAD MAP kinase site began to dephosphorylate. Moreover, at 15 h the phosphorylation of CAD by PKA began to increase once again. PKA phosphorylation is known to reduce the sensitivity of CAD to PRPP activation (32). Accordingly, a 1.7-fold decrease in the response to PRPP was observed (Fig. 5) as the cells exited the S-phase that coincided with the downregulation of the pyrimidine pathway.
In summary, we conclude that the cell cycle-dependent upand down-regulation of the rate of pyrimidine biosynthesis can be attributed to the sequential, concerted action of the PKA and MAP kinase signaling cascades. The phosphorylation and dephosphorylation of the regulatory sites on the CPSase domain of CAD alter the response to UTP and PRPP, allosteric effectors that control the flux through the pyrimidine pathway.