Ribonucleotide Reductase Activity and Deoxyribonucleoside Triphosphate Metabolism during the Cell Cycle of S49 Wild-type and Mutant Mouse T-lymphoma Cells*

We investigated deoxyribonucleoside triphosphate metabolism in 549 mouse T-lymphoma cells synchro- nized in different phases of the cell cycle. 549 wild-type cultures enriched for G1 phase cells by exposure to dibutyryl cyclic AMP (Bt2cAMP) for 24 h had lower dCTP and dTTP pools but equivalent or increased pools of dATP and dGTP when compared with exponentially growing wild-type cells. Release from BtzcAMP arrest resulted in a maximum enrichment of S phase occurring 24 h after removal of the Bt2cAMP, and was accompanied by an increase in dCTP and dTTP levels that persisted in colcemid-treated (GdM phase enriched) cultures. Ribonucleotide reductase activity in permeabilized cells was low in GI arrested cells, increased in S phase enriched cultures and fur- ther increased in G2/M enriched cultures. In cell lines heterozygous for mutations in the allo- steric binding sites on the M1 subunit of ribonucleotide reductase, the deoxyribonucleotide pools in S phase enriched cultures were larger than in wild-type 549 cells, suggesting that feedback inhibition of ribonu- cleotide reductase is an important mechanism limiting the size of deoxyribonucleoside triphosphate pools. The M1 and M2 subunits of ribonucleotide reductase from wild-type 549 cells were identified on two-di- mensional


Ribonucleotide Reductase Activity and Deoxyribonucleoside
Triphosphate Metabolism during the Cell Cycle of S49 Wild-type and Mutant Mouse T-lymphoma Cells* (Received for publication, May 14, 1984) Daniel A. Albert$ and Lorraine J. Gudast We investigated deoxyribonucleoside triphosphate metabolism in 549 mouse T-lymphoma cells synchronized in different phases of the cell cycle. 549 wildtype cultures enriched for G1 phase cells by exposure to dibutyryl cyclic AMP (Bt2cAMP) for 24 h had lower dCTP and dTTP pools but equivalent or increased pools of dATP and dGTP when compared with exponentially growing wild-type cells. Release from BtzcAMP arrest resulted in a maximum enrichment of S phase occurring 24 h after removal of the Bt2cAMP, and was accompanied by an increase in dCTP and dTTP levels that persisted in colcemid-treated (GdM phase enriched) cultures. Ribonucleotide reductase activity in permeabilized cells was low in GI arrested cells, increased in S phase enriched cultures and further increased in G2/M enriched cultures.
In cell lines heterozygous for mutations in the allosteric binding sites on the M 1 subunit of ribonucleotide reductase, the deoxyribonucleotide pools in S phase enriched cultures were larger than in wild-type 549 cells, suggesting that feedback inhibition of ribonucleotide reductase is an important mechanism limiting the size of deoxyribonucleoside triphosphate pools.
The M 1 and M2 subunits of ribonucleotide reductase from wild-type 549 cells were identified on two-dimensional polyacrylamide gels, but showed no significant change in intensity during the cell cycle. These data are consistent with allosteric inhibition of ribonucleotide reductase during the G1 phase of the cycle and release of this inhibition during S phase. They suggest that the increase in ribonucleotide reductase activity observed in permeabilized S phase-enriched cultures may not be the result of increased synthesis of either the M 1 or M 2 subunit of the enzyme.
The enzyme ribonucleoside diphosphate reductase (EC 1.17.4.1) catalyzes reduction of ribonucleoside diphosphates to deoxyribonucleoside diphosphates, which are then converted to the precursors for DNA synthesis. Ribonucleotide * This work was supported by Grant R01 CA27953 from the National Cancer Institute (L. J. G.) and by Grant CA30387 from the National Cancer Institute (D. A. A.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
$ reductase functions in the only de m u 0 pathway for in vivo synthesis of all four deoxyribonucleoside triphosphates. The mammalian enzyme is composed of two subunits; the M1 protein (89 kilodaltons) has distinct binding sites that are involved in the allosteric regulation of enzyme activity, whereas the M2 subunit (55 kilodaltons) contains a hydroxyurea-sensitive site (1, 2). Functionally, both subunits are necessary for catalytic activity. Ribonucleotide reductase is under complex allosteric feedback control, and it has been suggested that this could provide a self-regulated flow of dNTPs for DNA synthesis ( Fig. 1) (3). This model proposes activation by ATP for pyrimidine diphosphate reduction, by dTTP for guanosine diphosphate reduction, and by dGTP for adenosine diphosphate reduction. Feedback inhibition is also proposed in this model-dTTP inhibits pyrimidine biosynthesis, dGTP inhibits pyrimidine and its own biosynthesis, and dATP inhibits reduction of all four nucleoside diphosphates (4-8).
Cells synchronized in different phases of the cell cycle might reflect this allosteric regulation of their dNTP pools. Thus, during periods of active DNA synthesis (S phase), pools of dCTP and dTTP might be relatively high, whereas dATP pools might be lower; during the GI and GB phases of the cell cycle, the opposite might occur. In addition, if allosteric regulation were the sole mechanism responsible for cell-cyclespecific alterations in dCTP production, no significant changes in ribonucleotide reduction would be expected in synchronized permeabilized cells from specific phases of the cell cycle.
We have tested each of these predictions in wild-type and two partially characterized mutant S49 cell lines with abnormal allosteric binding sites on the M1 subunit of ribonucleotide reductase (11, 12) using dibutyryl cyclic adenosine monophosphate and colcemid to generate populations enriched for three portions of the cell cycle: the GI, S, and G2/M phases. CO'S modification of Eagle's medium containing heat-inactivated 10% horse serum at 37 "C in a 10% CO, (in air) atmosphere.
The hydroxyurea-resistant cell line was selected from wild-type S49 by culturing cells in the presence of progressively higher concentrations of hydroxyurea, an inhibitor of ribonucleotide reductase (13). This cell line will be described in greater detail in another report.' Briefly, the line was selected by exposure of wild-type S49 cell to incrementally increasing concentrations of hydroxyurea from 50 p M to 1 mM over a 6-month period. The resulting line was 20-fold resistant to the cytotoxic effects of hydroxyurea. These cells were resistant to hydroxyurea inhibition and had 4-fold elevated CDP reductase activity and similarly elevated dNTP pools.
Cell-cycle Analysis-Experimental cell suspensions (5 ml) containing 5-10 X lo5 cells/ml were centrifuged and resuspended in 1-3 ml of hypotonic solution containing 0.05 mg/ml propidium iodide, 0.1% sodium citrate, and 0.1% Triton X-100 (14). After staining, cells were analyzed on an argon-laser (488 nM) cytofluorimeter designed by Shapiro (15). The distribution of cells in the GI, S, and G2/M phases were determined by counting 1000 cells and integrating under the left slope of the GI peak and the right slope of the G2/M peak. The proportion in GI equals 2 X left slope/total; the proportion in G,/M equals 2 X right slope/total; and the proportion in S phase is the total -(GI + G2/M)/total.
Determination of Deoxyribonucleotide Pook"2.5-5 X lo7 cells were washed in phosphate-buffered saline, centrifuged in a microfuge, resuspended in 200 pl of ice-cold 0.1 M potassium phosphate buffer, pH 7.2, and an aliquot was counted on a Coulter model ZBI. 1.5 N perchloric acid (40 pl) was added, followed 60 s later by the addition of 80 p1 of 1.2 N KOH (final pH = 7.0), and samples remained on ice for 5-10 min, followed by a centrifugation a t 4 "C in a microfuge. These extracts were analyzed for ribonucleotides. For deoxyribonucleotide-pool measurements, periodation of the extract with 0.02 M sodium periodate for 90 min a t 37 "C was performed by the method of Garrett and Santi (16). The reaction was terminated by the addition of 0.03 M rhamnose. Extracts were analyzed on a Beckman Altex high-performance liquid chromatograph using an SAX Partisil column with a 0.38 M ammonium phosphate, 1.5% acetonitrile, pH 3.45, elution buffer. Peaks were measured a t both 280 and 254 nM and concentration was proportional to peak height above base-line.
Purification of MI-The M1 subunit of ribonucleotide reductase was purified according to the method of Gudas et al. (17). Briefly, this involved harvesting and washing 2-3 X lo9 cells in ice-cold phosphate-buffered saline, followed by sonication in 50 mM Tris, 2 mM dithiothreitol, pH 7.4. RNA and DNA were removed by streptomycin sulfate precipitation (0.65%) and the supernatant was precipitated with 40% ammonium sulfate. This precipitate was dialyzed or desalted in Tris buffer and loaded onto a 3-ml column of dextran blue Sepharose. The column was washed with Tris buffer and M1 was eluted at a salt concentration of between 50 and 500 mM NaCl. This eluate was placed over a 1-ml column of dATP Sepharose washed with 0.1 M KC1 in Tris buffer and then eluted with a linear gradient D. A. Albert, manuscript in preparation. Two-dimensional Gels-Two-dimensional gels were performed by Blue, or, if the material analyzed was radioactive, autoradiographed.
[35S]Methionine labeling was performed by incubating 5 X 10' cells in 100 pCi/ml of [35S]methionine for 3 h in methionine-free media containing dialyzed horse serum.
Ribonucleotide Reductase Assay-Ribonucleotide reductase activity was measured by the conversion of CDP to deoxyCDP by permeabilized cells or partially purified cell extracts. 1.2 X IO7 cells were permeabilized by exposure to 1% Tween 80 by the method of Lewis et al. (19). The assay mixture included final concentrations of 39 mM Hepes: 6.6 mM ATP, 8 mM MgC12, 22 mM dithiothreitol, 50 p M CDP, and 0.42 pCi of ["CICDP in a final volume of 300 $1 after addition of 60 p1 of a 5-fold concentrated mixture. To assay the partially purified enzyme preparation, 5 mM NaF was added to the other chemicals to diminish phosphatase activity. The reaction was terminated by boiling the samples for 4 min. 6.0 mg of Crotalus atrox venom/assay was then added, and samples were incubated for 4 h a t 37 "C. Samples were diluted with 500 pl of HZO, loaded onto 1 ml of borate Dowex 1 columns, and [14C]deoxycytidine product was eluted from the column with 2-3 ml of water. Samples were then counted by liquid scintillation. Fig. 2

. Exponentially growing cells are
The abbreviation used is: Hepes, 4-(2-hydroxyethyl)-l-piperazineethanesulfonic acid.  (Fig. 2 A ) . S49 cells reversibly arrest in the GI phase of the mammalian cell cycle when exposed to cyclic AMP or Bt2cAMP (20). Cells exposed to 500-1000 ~L M Bt2cAMP accumulate in GI (about 90%) within 24 h of Bt,cAMP addition, with the remaining 10% in G2/M (Fig. 2B), with no loss in cell number or viability, as assessed by either microscopic examination or [35S]methionine incorporation into cell protein. S49 cells arrested in the G1 phase may be released from this arrest by the removal of Bt,cAMP. This release results in a maximal enrichment of S phase cells (about 60%) 24 hr after Bt,cAMP removal (Fig.  2C). Colcemid (0.5 pg/ml) synchronizes S49 cells in the G2/ M phases of the cycle. This arrest is maximal at 8 h (about 90%) (Fig. 2 0 ) and subsequently results in an increasing proportion of tetraploid cells.
Nucleotide Pools-Deoxyribonucleoside triphosphate pools in synchronized populations of cells are shown in Table I. We observed a 50% reduction in the dCTP and dTTP pools, but normal dATP and dGTP pools in GI arrested wild-type S49 cells, as compared to exponentially growing cells. In populations enriched for S phase cells, the dCTP pool was elevated, and the dTTP pool was slightly depressed, as compared to exponentially growing wild-type S49 cells. A decline in the dATP pool was observed in S phase-enriched cells (Table I, Fig. 3). In G2/M enriched wild-type S49 cells, the dCTP and dTTP pools were approximately the same as those in exponentially growing cells, whereas the dATP pool was much larger ( Table I).
The mutant dGuo-L cells exhibited the same pattern of dNTPs during GI arrest that wild-type cells did. By contrast, the dGTP pool of this dGTP-resistant mutant increased throughout S and G2/M. The dTTP pool during S phase was considerably higher than that of wild-type cells as well (Table  I). This suggests that feedback inhibition by dGTP is an important control limiting the size of the dTTP and dGTP pools during S phase. The mutant dGuo-200-1 also had low dCTP and dTTP pools in GI similar to wild-type cells. However, there was a striking increase in all four dNTP pools during S phase which persisted through Gz/M ( Table I) suggest that dATP feedback inhibition is important in limiting the dNTP pools during the cell cycle.
In addition, the ATP pool varied during the cell cycle (Table   I). In wild-type S49 cells arrested in the GI phase of the cycle, the ATP pool was only about 60% of that in S phase-enriched populations, whereas the GTP pool variation was less than 10% (data not shown). Dose-titration experiments revealed no augmentation of the ATP concentration by Bt2cAMP, indicating that it does not contribute to the adenine nucleotide pool. Ribonucleotide Reductase Activity-The activity of the enzyme ribonucleotide reductase was measured by using the substrate CDP. CDP reduction in permeabilized S49 wildtype cells paralleled the dCTP and dTTP pool sizes. As shown

Cell-cycle Reductave Activity and Triphosphate Metabolism
in Tahle 11, CDP reductase activity declined 50% or more in G I arrested cells, whereas S phase enriched populations equalled or exceeded the activity in exponentially growing control cultures. Colcemid-arrested G2/M cells had activity that was further increased over control cultures. Enzvme activitv (per mg of protein) measured in sonicates of cellcvcle-enriched cultures was similar to the data from permeahilized cells except, for the results of cells in the G2/M phase ( Table 11).
Identification  Fig.   4. Contaminating proteins were eluted at lower ATP concentrations with some M1. In the 20 mM ATP-elution fraction, M1 can he seen in relationship to other proteins that remain hy comparison of a two-dimensional gel of the elut ion fraction with one showing a [""Slmethionine-labeled whole-cell extract (Fig. 5 ) . At 60 mM ATP, only a horizontal triplet (the third density is less intense and is poorly seen) was eluted. Lf'hen aliquots of eluted protein are recornhined with the dextran blue Sepharose column flow-through (which contains the "L suhunit hut no M1 suhunit), CDP reductase activitv can he reconstituted. This act.ivitv corresponds to the intensity of M1 on the gels. When fractions of purified ribonucleotide reductase M1 suhunit were mixed with ['"'SJmethionine-labeled cell extracts and run on a two-dimensional gel. the Coomassie blue-stained purified M1 could be aligned with corresponding radiolabeled spots.
Idcntificntion of the M:! Suhunit of Rihonuclrotidc Rcductnsc on Two-dimensional Gds-The M2 suhunit of rihonucleot ide reductase has been tentatively identified on two-dimensional gels (Fig. 6 ) GI arrested (top p a d ) , S phase enriched size estimates of M2 and corresponded in PI and molecular weight to a spot identified as M2 in fibroblasts by Lewis and Srinivasan (21).
Cell-cycle-enriched Two-dimensional Gels-Using the previously described cell-cycle-phase enrichment procedures, we attempted to discern a difference in either the M1 subunit or the putatively identified M2 subunit spot in gels from [35S] methionine-labeled cultures enriched for G1, S, or GJM phase cells (Fig. 6). Enriched cultures were labeled identically and the same amount of radioactive protein was loaded on each gel; thus, the relative spot intensity should reflect the amount of that protein in those cells. In G1-arrested cells, both the M1 and M2 spots were present, but we were unable to detect a difference in spot density when compared with S phase or G2/M phase enriched cells.

DISCUSSION
The regulation of deoxyribonucleoside triphosphate synthesis is closely linked to cellular-DNA synthesis. Inhibitors of ribonucleotide reductase such as hydroxyurea (22) or the feedback inhibitors dATP and dGTP decrease deoxyribonucleoside triphosphate pools and inhibit DNA synthesis (23). In addition, alterations in the normal pool sizes have been implicated in mutagenesis by potentiating mispairing (24-26).
Thus, it appears that a closely regulated continuous flow of dNTPs is necessary for accurate DNA synthesis and that allosteric regulation might be involved in this regulation.
The deoxyribonucleotide pool data in GI arrested wild-type S49 cells is consistent with allosteric control of ribonucleotide reductase in G1. dCTP, dTTP, and ATP are consistently low and dATP and dGTP are normal or elevated in G, arrested wild-type S49 cells when compared with either exponentially growing controls, S phase enriched populations, or G2/M enriched populations. S phase enriched wild-type cultures exhibited an increase in dCTP, dTTP, and ATP consistent with activation of ribonucleotide reductase, but a decrease in dATP was less regularly observed during S phase. However, we did observe a decrease in dATP level which occurred 4 h after release from BtzcAMP-induced GI arrest (Fig. 3). Previous studies have also shown dNTP-pool-size changes during the cell cycle. Skoog et al. (27,28) documented increased pools of all four deoxynucleoside triphosphates during S phase, with dCTP exhibiting the largest increase. Studies in lymphocytes have shown similar results (29, 30).
We also employed two mutant cell lines: dGuo-L, a cell line heterozygous for a mutant ribonucleotide reductase that is dGTP feedback resistant, and dGuo-200-1, a cell line heterozygous for a mutant ribonucleotide reductase that is dATP feedback resistant. If dATP and dGTP were inhibitory effector molecules for deoxyribonucleoside triphosphate synthesis, then these mutant cell lines might fail to show the "inhibitory" pattern of deoxyribonucleotide pool levels that we observed in wild-type S49 cells arrested in G1. This was not the caseboth mutant cell lines showed depressed pools of dCTP and dTTP and normal dATP pools in GI arrested populations. However, a suggestion of abnormal cell cycle regulation of dNTP pools came from the S phase and Gz/M phase enriched mutant cell lines. Both the dTTP and dGTP pools were elevated in dGuo-L cells suggesting a lack of dGTP feedback (middle panel), and G2/M enriched (bottom panel) cell cultures, respectively. Cultures from cycle enriched populations were labeled with [35S]methionine as stated under "Experimental Procedures." Labeled cells were resuspended in a small volume of lysis buffer and 1 X IOfi cpm were loaded on each gel. Gels were exposed for 7 days each. Insets show expanded views of ribonucleotide reductase M1 and M2 subunits.
inhibition. dGuo-200-1 cells had very elevated pools of dCTP and dTTP in S phase, suggesting a decreased ability to slow dNTP production by dATP feedback inhibition. It should be noted that the mutants are heterozygous for the mutant ribonucleotide reductase M1 subunit; thus, they might not exhibit drastically disordered cell-cycle regulation of dNTP pools.
Ribonucleotide reductase activity in the absence of allosteric effector molecules was measured by the permeabilized cell CDP reductase assay. G1 arrested wild-type S49 cells had significantly less CDP reductase activity than either exponentially growing controls or cells synchronized in S or G2/M phase (Table 11). Conversely, S and G2/M phase enriched cultures showed CDP reductase activity that was greater than that measured in exponentially growing cells. These data are consistent with the results of Kucera et al. (31) and Murphree et al. (32). In contrast, Lewis et al. (19) found that, in permeabilized hamster cells, ribonucleotide reductase activity increased during S phase but declined rapidly at the end of S phase. If the activity of ribonucleotide reductase were regulated solely by allosteric effector molecules during the cell cycle, then the enzyme activity in permeabilized cells should be the same throughout the cell cycle; our data thus suggest a nonallosterically mediated increase in ribonucleotide reductase activity in S phase.
There are many possible mechanisms for a nonallosteric increase in CDP reductase activity. Five possibilities (not necessarily exclusive) that have been suggested are: 1) the compartmentalization of ribonucleotide reductase into a multienzyme complex including the enzymes necessary to process precursor nucleotides into deoxynucleotide substrates for DNA polymerase (33, 34), 2) increased quantity of the M2 subunit of ribonucleotide reductase (35), 3) nuclear translocation of ribonucleotide reductase (36), 4) changes in the endogenous hydrogen donor system (31, 37), and 5) association with the nuclear matrix (38).
We examined one of these possibilities, that an increased quantity of either the M1 or M2 subunit of ribonucleotide reductase is responsible for the increased activity observed in S phase cells. We identified M1 and tentatively identified M2 on two-dimensional gels, and then compared the intensity of these spots on gels of [35S]methionine-labeled synchronized populations in different phases of the cell cycle. The spots that correspond to the M1 and M2 subunits of ribonucleotide reductase did not change by greater than %fold in intensity on two-dimensional gels from [35S]methionine labeled cultures enriched for GI phase, S phase, or G2/M phase wildtype S49 cells. Thus, we conclude that the increased ribonucleotide reductase activity seen in S phase cells compared with GI arrested cells is not due to a large induction of the synthesis of either the M1 or M2 subunit.
Our data are consistent with previous observations on the cell-cycle control of ribonucleotide reductase activity by Eriksson and Martin (35). They noted a 6-fold increase in ribonucleotide reductase activity in S phase that appeared to be due to increased M2 rather than M1 subunit activity. Our data suggests that this may not be due to an increased quantity of M2 protein.
In summary, we have provided evidence that allosteric control of ribonucleotide reductase activity accounts for some of the observed changes in deoxyribonucleoside triphosphate pools during the cell cycle, especially during the G1 phase. Nonallosteric activation of the enzyme must occur in S phase, but we were unable to demonstrate induced synthesis of either the M1 or M2 subunit of ribonucleotide reductase as a source of increased activity in S uersus GI phase cells. Further study will be necessary to ascertain the mechanism by which ribonucleotide reductase activity is increased in S phase cells.  , .