A Model for Compartmentation of De Novo and Salvage

Proc. Nat. Acad. Sci. USA Vol. 72, No. 9, pp. 3333-3336, September 1975 Biochemistry A model for compartmentation of de novo and salvage thymidine nucleotide pools in mammalian cells* (HeLa cells/DNA replication/thymidine labeling) DANIEL KUEBBING AND RUDOLF WERNERt Department of Biochemistry, University of Miami School of Medicine, Miami, Florida 33152 Communicated by Sidney P. Colowick, June 6,1975

ABSTRACT When [3H]thymidine is added to HeLa cells in addition, 16 gM thymidine. HAT medium (2, 3) con- grown in thymidine-free medium it is incorporated into DNA tained 100 1AM hypoxanthine, 1 jsM amethopterin (Lederle), almost immediately at full specific activity, blocking any fur- 16 ,uM thymidine, and 10 ,uM glycine. HAT medium with- ther incorporation of de novo synthesized thymidine nucleo- HA in MEMT or tides. Apparently, de novo nucleotide pools [Baumunk, C. N. out thymidine is termed medium. Growth & Friedman, D. L. (1971) Cancer Res. 31, 1930-35] and sal- HAT medium for several months had no effect on the results vage nucleotide pools are compartmentalized in HeLa cells. described, and HeLa cells retained their ability to grow in After one generation in the presence of thymidine, the cells MEM. The doubling time was identical in all three media begin utilizing nucleotides from both pools to about equal ex- (22-24 hr). Cell numbers were determined in replicates of tent. the experimental flasks by trypsinization and counting with Most studies of DNA synthesis in mammalian cells use thy- a hemocytometer at the time of the experiment. Variation in midine as radioactive label. Prior to labeling, the cells are cell number was less than 10%. usually grown in thymidine-free medium, and depend on de Incorporation Experiments. Falcon flasks were seeded novo synthesis of thymidine nucleotides. To be able to incor- with cells in the appropriate medium and incubated over- porate [3H]thymidine the cells have to use salvage pathway night. Labeling was initiated, after pouring off the growth enzymes. This switch in pathways cannot be considered medium, by- quickly adding the same type of medium con- steady-state. In bacteria, the presence of steady-state or non- taining labeled nucleoside (2.4 gM, 1-5 Ci/mmol, New En- steady-state conditions has been shown to affect the type of gland Nuclear). The residual unlabeled thymidine, adhering DNA that is labeled (1). In order to study the possible effects to cells grown in MEMT or HAT medium, amounted to less of such conditions in mammalian cells, we investigated the than 10% of the added labeled nucleoside. The labeling me- incorporation of thymidine into HeLa cells in three differ- dium had been pre-equilibrated to 370 in 95% air/5% CO2. ent media, Eagle's minimal essential medium (MEM) (con- After incubation at 370 for the appropriate time, incorpora- taining no thymidine), HAT (containing hypoxanthine, am- tion was terminated by pouring off the radioactive medium, ethopterin, and thymidine), and MEMT (containing only rinsing the monolayer once with ice-cold Hanks' saline to re- thymidine). In MEM, the cells depend solely on de novo move excess protein, and then adding ice-cold 5% trichloro- synthesized nucleotides; labeling with thymidine, therefore, acetic acid/1% sodium dodecyl sulfate. This procedure took represents a non-steady-state condition. In HAT medium, on 5-10 sec. The samples were filtered through glass fiber fil- the other hand, the de novo pathway is blocked by am- ters, rinsed first with 5% trichloroacetic acid and then meth- ethopterin; the cells exclusively use exogenous thymidine via anol, dried, and counted by liquid scintillation spectroscopy the salvage pathway, and incorporation of [3H]thymidine using Omnifluor (New England Nuclear). can be observed under steady-state conditions. In the third Cs2SO4 Density Analysis. After stopping nucleoside in- medium, MEMT, the cells have a choice between thymidine corporation with ice-cold phenol/ethanol (2/70%) the mo- nucleotides derived from the salvage and the de novo path- nolayers were rinsed once with 95% ethanol to remove traces ways. The kinetics of thymidine incorporation observed in of phenol, and the residual ethanol was removed by evapo- three as as the ration. Cells (5 X 106) were lysed by the addition of 2 ml of these media, well analysis of the intracellular 0.05 M Tris-HCI/0.05 M EDTA (pH 7.4), 0.2 ml of 20% Sar- thymidine nucleotide pools, suggest a compartmentation of kosyl NL-97 (Geigy), and 0.2 ml of nuclease-free Pronase intracellular salvage and de novo thymidine nucleotide (Calbiochem, 2 mg/ml). This mixture was incubated at 370 pools. for 1 hr. KOH (3 M, 0.2 ml) was then added, and the lysate MATERIALS AND METHODS was heated at 50-60° for 30 min to hydrolyze RNA. After chilling, the lysate was neutralized by the addition of 0.5 ml HeLa S-3 cells (ATCC CCL 2.2) were maintained in Eagle's of 1 M Tris-HCl (pH 7.4) and 0.55 ml of 1 M HCl, and di- Minimal Essential Medium (MEM Gibco F-11) supplement- alyzed extensively against 0.05 M Tris-HCI/0.05 M EDTA ed with 10% calf serum (Gibco 617H1). MEMT contained, (pH 7.4) in the cold. The DNA was then sheared using a Branson sonifier (4 X 30 sec, 60 W). After the addition of 2000-4000 cpm of heavy and light density markers, the mix- Abbreviations: MEM, Eagle's minimal essential medium + 10% calf 5 serum; MEMT, MEM + 16 MAM thymidine; HAT medium, MEM + ture was denatured by incubation for min at 1000 and 100 MM hypoxanthine/1 MM amethopterin/16 MM thymidine/10 quick chilling in ice. (Uniformly [14C]thymidine- and MM glycine; HA medium, HAT without thymidine. [14C]bromodeoxyuridine-labeled HeLa cell DNA was pre- * This paper is dedicated to Dr. Karl H. Slotta on the occasion of pared by growing cells in HA medium containing one of the his 80th birthday. two labeled nucleosides.) A 2.75-ml aliquot of this solution t To whom reprint requests should be sent. was then mixed with 2.30 ml of saturated Cs2SO4 solution, 3333 Downloaded by guest on September 29, 2021 3334 Biochemistry: Kuebbing and Werner Proc. Nat. Acad. Sci. USA 72 (1975)

U~~~~~~~~~~ 2F -3000 | U 8 103 ~~0 U~~~~~~~ IL I M I E ~~~~~MEMT U 0 rerset59X105 celNna2-acnfak achfakTa n

0 2 ~0 40 tI < ME MT M IN UTE 9 FIG. 1. Kinetics of [3Hjthymidine incorporation. Each point represents 5 X i05 cells in a 25-m Facnfak ahfakws incubated with 3 ml of medium containing radioactive thymidine II I~ (2.4 gM; 5 Ci/mmol). The inset depicts the same data with an ex- panded time scale. and the mixture was centrifuged in a Spinco type 40 rotor FRACTION NUMBER for 65-70 hr at 34,000 rpm and 5°. To prevent sticking of FIG. 2. Density gradient analysis in Cs2SO4. Each gradient contains the DNA from 5 X 106 cells in a 75-cm2 flask labeled for the DNA, the polyallomer tubes had been boiled for 1 hr in one hour with [3H]bromodeoxyuridine (1 Ci/mmol) in the indicat- 0.01 M EDTA (pH 7.0) and then stored at 40 in distilled ed media. The broken lines marked H and L represent the posi- water. Six-drop fractions were collected from the bottom of tions of marker DNAs fully substituted with bromodeoxyuridine the tube directly onto filter papers in scintillation vials, and thymidine, respectively. dried, and counted by liquid scintillation spectroscopy. Nucleotide Pool Analysis. Falcon dishes (30 mm diame- If the cells are grown in MEMT medium (i.e., in the pres- ter) were seeded with approximately 5 X 105 cells in 5 ml of ence of thymidine) for one generation prior to the addition the appropriate medium and incubated overnight. Intracel- of [3H]thymidine, the label is incorporated at one-half the lular thymidine nucleotide pools were labeled by incubating rate observed in cells grown in MEM or HAT medium. This the cells for 30 min in medium containing [3H]thymidine result suggests that the cells have adapted to the presence of (2.4 ,uM, 12 Ci/mmol). Nucleotides were extracted from the thymidine and are now using nucleotides derived from both cells either directly or after further incubation in HA medi- pathways to about an equal extent. Again the initial rate of um by quickly aspirating all the medium and adding 0.5 ml incorporation is lower, indicating the presence of a preexist- of ice-cold 5% trichloroacetic acid. The acid extracts were ing pool of nucleotides. centrifuged to remove cell debris and extracted three times When cells are labeled for 1 hr with [3H]bromodeoxyuri- with 3 ml of water-saturated ether. Thirty microliters of dine after growth in each of the three media and the DNA is each sample were chromatographed on PEI-cellulose thin analyzed in Cs2SO4 gradients (Fig. 2), the labeled DNA iso- layer sheets (Brinkman, Westbury, N.Y.) together with 5 Al lated from cells grown in MEM or HAT medium, after de- of a solution containing 5 mg/ml each of dTTP, dTDP, and naturation, has a density indistinguishable from that of DNA dTMP (4). in which thymidine is fully substituted by bromodeoxyuridine. Salvage nucleotides, therefore, must have completely RESULTS bypassed the de novo synthesized nucleotides, confirming Figure 1 compares the kinetics of [3H]thymidine incorpora- the conclusion drawn from the previous experiment. The la- tion into cells growing in MEM, HAT, or MEMT medium. beled DNA, isolated from cells grown in MEMT medium, In MEM, the rate of incorporation is essentially linear from bands at an intermediate density, indicating that both de time zero, suggesting that thymidine does not pass through novo synthesized thymidine nucleotides and salvage bromo- preexisting de novo synthesized thymidine nucleotide pools, deoxyuridine nucleotides had been used for its synthesis. which Baumunk and Friedman have estimated to be suffi- If cells grown in HAT medium are incubated in the ab- cient for about 5 min of DNA synthesis (5). In HAT medi- sence of thymidine (HA medium) prior to labeling with um, on the other hand, the initial rate of thymidine incorpo- [3H]thymidine, a much shorter lag in incorporation rate is ration is lower, reflecting the mixing of radioactive thymi- seen than in HAT medium alone, suggesting that most of the dine with preexisting intracellular salvage thymidine nu- salvage thymidine nucleotide pool has been depleted (Fig. cleotides. The final rates of incorporation, however, are 3). (The higher rate of incorporation observed in HA medi- identical in both media, suggesting that the addition of thy- um, as compared to MEM or HAT medium, may represent midine to cells not previously exposed to thymidine com- the accumulation of forks in replicons initiated in the ab- pletely prevents the utilization of de novo synthesized nu- sence of thymidine.) To examine the depletion of the sal- cleotides. vage pool directly, HeLa cells were first labeled with Downloaded by guest on September 29, 2021 Biochemistry: Kuebbing and Werner Proc. Nat. Acad. Sci. USA 72 (1975) 3335

PYRIMIDINE RIBONUCLEOT IOES SA LVAOE dTTP 5000 POOL -J U HAT -HA

THYMIOINE

I U * 000eO RE PLICATI ON M IN E MEM | P RECURSOR o* I I~~~~ OE NOVO POOLP dTTP POO L 10 20 DNA FIG. 5. Model for the organization of the intracellular thymi- M INUTE S dine nucleotide pools in HeLa cells. FIG. 3. [3H]Thymidine incorporation after treatment with anm- DISCUSSION ethopterin (HA medium). Conditions were as described in Fig. observed that the de (.) Control cells grown in MEM medium. (*) Cells first grown i1. Baumunk and Friedman (5) novo HAT medium and then incubated in HA medium for 4 hr prior 1in dTTP pool in HeLa cells cannot be depleted in the presence thymidine incorporation. of amethopterin. To explain these results, they proposed that there exist two separate dTTP pools, only one of which [3H]thymidine for 30 min (sufficient to label the dTTP potsol (amounting to less than 5% of the dTTP) is readily available to full specific activity), and then the nucleotide pools werre for replication. Our results are compatible with this view. In extracted and analyzed chromatographically, either immbe- addition, we have shown that cells grown in MEM prefer ex- diately or after incubation in HA medium. As seen in Fig. i4, ogenous thymidine to de novo synthesized nucleotides, the tritium-labeled salvage dTTP pool of cells grown iin thereby completely excluding de novo nucleotides from in- MEM, HAT, or MEMT medium prior to exposure tto corporation. Furthermore, in contrast to the de novo pool, [3H]thymidine can be depleted. In different experiments thie the salvage thymidine nucleotide pool can be depleted in the amount of depletion within the first hour after the removaal absence of thymidine (Fig. 4). These results may obe ex- of thymidine varied from 40 to 75%. Within a single experi - plained by the model shown in Fig. 5. According to this ment, the depletion observed in MEM- or HAT-grown cellis model, there exist three separate intracellular nucleotide was identical. (The dTMP and dTDP pools together accounit pools, a de now pool, a salvage pool, and a small replication for only about 10% of the total thymidine nucleotide pocil precursor pool. In MEM, where cells have no salvage pool, and are depleted with similar kinetics but to a greater extenIt only a small portion of the de novo synthesized dTTP is than the dTTP pool.) The salvage pools of cells labeled ifn present in the small replication precursor pool. We do not MEM or HAT medium are essentially the same size, but th,te know whether this pool contains thymidine nucleotides as pool in MEMT-grown cells is only about one-half this size dTTP or some other form of activated dTMP (6, 7). When The salvage pools in all three media are depleted to th(e thymidine is added to cells growing in MEM, a new dTTP same final size. pool (salvage pool) is created. In contrast to the nucleotides in the de novo dTTP pool, the nucleotides of the salvage pool freely mix with the replication precursor pool. As soon 33x 103 as a dTTP is it * MEM salvage pool established, immediately prevents entry of de novo synthesized nucleotides into the repli- * HAT cation precursor pool, most probably by blocking their syn- -J * thesis through feedback inhibition (8). The very small num- w MEMT ber of de novo-derived nucleotides that are present in the replication precursor pool at -the time thymidine is first 0. added for about 15 sec of would not no- a.) (enough replication) .a. ticeably affect the initial incorporation rate of exogenous .-103 1 thymidine. After cells are exposed to thymidine for one gen- I2- *13 0 eration, they adapt to its presence and return to using de novo synthesized nucleotides in addition to the exogenously e supplied thymidine. This adaptation may reflect the achievement of a new 2 4 steady-state in the control of de novo I synthesis of nucleotides. We consider it unlikely that the subclone of HeLa cells HOURS IN HA MEDIUM (S3) used in our experiments has a much smaller de novo FIG. 4. Depletion of thymidine nucleotide pools. HeLa cells thymidine nucleotide pool than the subclones used by Bau- were labeled with [3H]thymidine for 30 min in the indicated culture media. The medium was then changed to HA (O hr), and thy- munk and Freedman (5). Furthermore, de novo thymidine midine nucleotides were extracted and analyzed at the indicated nucleotide pools in other cell lines have been reported to be times as described. Each point represents the average of two sepa- similar in size (10, 11). However, if HeLa S3 cells did have a rate cultures. much smaller de novo dTTP pool, the same results would Downloaded by guest on September 29, 2021 3336 Biochemistry: Kuebbing and Werner Proc. Nat. Acad. Sci. USA 72 (1975) have been obtained. This would not affect the conclusions This work was supported by the Comprehensive Cancer Center concerning the control of the de novo pathway except that of Greater Miami, Grant no. NIH SBO4 5POlCA1439502. D.K. was all de novo synthesized nucleotides may mix with the repli- the recipient of an NSF predoctoral fellowship; R.W. an Estab- cation precursor pool. lished Investigator of the American Heart Association. The residual salvage nucleotide pool observed in cells incubated in HA medium may be located in cells that are not 1. Werner, R. (1971) Nature 230,570-572. engaged in DNA replication. However, Adams (9) reported 2. Szybalski, W., Szybalska, E. H. & Ragnie, G. (1962) Nat. Can- that cer Inst. Monogr., no. 7, p. 75. only cells in S and early G2 phase can take up thymi- 3. Littlefield, J. W. (1964) Science 145,709. dine. Therefore, most of the labeled thymidine should be 4. Randerath, K. & Randerath, E. (1967) in Methods in Enzy- contained in cells actively involved in DNA synthesis. An al- mology, eds. Grossman, L., & Moldave, K. (Academic Press, ternative explanation for the incomplete depletion may be New York), Vol. XII, part A, pp. 342-343. that the overall rate of DNA synthesis and, consequently, the 5. Baumunk, C. N. & Friedman, D. L. (1971) Cancer Res. 31, rate of depletion of the thymidine nucleotide pool in S phase 1930-1935. cells is reduced if the concentration of nucleotides falls 6. Werner, R. (1971) Nature New Biol. 233,99-103. below a certain level. This interpretation would explain the 7. Pollock, J. M., Jr. & Werner, R. (1975) Biochem. Blophys. Res. slow further depletion of the dTTP pool that is seen on long- Commun. 63,699-705. er 8. Kelly, W. N. (1972) in Growth, Nutrition, and Metabolism of incubations with amethopterin (Fig. 3). Cells in Culture, eds. Rothblatt, G. H. & Cristofalo, V. J. (Ac- Other workers have reported that the de novo dTTP pool ademic Press, New York), Vol. 1, pp. 211-241. in different cell lines can be depleted in the presence of am- 9. Adams, R. L. P. (1969) Exp. Cell Res. 56,49-54. ethopterin (9-11). The difference between these cell lines 10. Adams, R. L. P., Berryman, S. & Thomson, A. (1971) Biochim. and HeLa cells may lie in the tightness of control preventing Biophys. Acta 240, 455-462. mixing between nucleotide pools. 11. Fridland, A. (1974) Cancer Res. 34, 1883-1888. Downloaded by guest on September 29, 2021