Regulation of proliferating cell nuclear antigen during the cell cycle.

The proliferating cell nuclear antigen (PCNA), also known as cyclin and DNA polymerase delta auxiliary factor, is present in reduced amounts in nongrowing cells and is synthesized at a greater rate in the S phase of growing cells. The recently discovered involvement of PCNA in DNA replication suggested that this pattern of expression functions to regulate DNA synthesis. We have investigated this possibility further by examining the synthesis, stability, and accumulation of PCNA in HeLa cells fractionated by centrifugal elutriation into nearly synchronous populations of cells at various positions in the cell cycle. In these fractionated cells we found that there is an increase in the rate of PCNA synthesis with a peak in early S phase of the cell cycle, but the magnitude of the increase is only 2-3-fold. This change reflects similar changes in the amount of PCNA mRNA. The fluctuating synthesis of PCNA maintains this protein at a roughly constant proportion of the total cell protein, although the amount doubles/cell in the cell cycle. Consistent with this observation, the stability of PCNA does not differ significantly from that of total cellular protein in synchronized HeLa cells. We also observed that a maximum of one-third of the total PCNA is tightly associated with the nucleus, presumably in replication complexes, at the peak of S phase. We conclude that the cyclic synthesis of PCNA in cycling HeLa cells maintains PCNA in excess of the amount involved directly in DNA replication and the amount of the protein neither fluctuates significantly with the cell cycle nor is limiting for DNA synthesis.

The proliferating cell nuclear antigen (PCNA), also known as cyclin and DNA polymerase 6 auxiliary factor, is present in reduced amounts in nongrowing cells and is synthesized at a greater rate in the S phase of growing cells. The recently discovered involvement of PCNA in DNA replication suggested that this pattern of expression functions to regulate DNA synthesis. We have investigated this possibility further by examining the synthesis, stability, and accumulation of PCNA in HeLa cells fractionated by centrifugal elutriation into nearly synchronous populations of cells at various positions in the cell cycle. In these fractionated cells we found that there is an increase in the rate of PCNA synthesis with a peak in early S phase of the cell cycle, but the magnitude of the increase is only 2-3-fold. This change reflects similar changes in the amount of PCNA mRNA. The fluctuating synthesis of PCNA maintains this protein at a roughly constant proportion of the total cell protein, although the amount doubles/cell in the cell cycle. Consistent with this observation, the stability of PCNA does not differ significantly from that of total cellular protein in synchronized HeLa cells. We also observed that a maximum of one-third of the total PCNA is tightly associated with the nucleus, presumably in replication complexes, at the peak of S phase. We conclude that the cyclic synthesis of PCNA in cycling HeLa cells maintains PCNA in excess of the amount involved directly in DNA replication and the amount of the protein neither fluctuates significantly with the cell cycle nor is limiting for DNA synthesis. Several lines of research from different laboratories converged to produce our present understanding of the protein variously known as the proliferating cell nuclear antigen (PCNA),' cyclin, and the DNA polymerase 6 auxiliary factor (see Bravo, 1986;and Mathews, 1989 for reviews). PCNA was first described by Miyachi et al. (1978) as a nuclear antigen, restricted to proliferating cells, that reacts with sera from some patients with the autoimmune disorder systemic lupus erythematosus, hence the proliferating cell nuclear antigen, PCNA. Two years *This work was supported by National Cancer Institute Grant CA13106 (to M. B. M.) and by National Institutes of Health Postdoctoral Fellowship (CA08109) from the National Cancer Institute (to G. F. M.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
later Bravo and Celis (1980) observed that the amount of a protein on two-dimensional gels correlated with cell proliferation and that the protein was synthesized in greater amounts during the S phase of the cell cycle. This protein, designated cyclin , was subsequently shown to be identical to PCNA by this laboratory (Mathews et al., 1984a). More recent work demonstrated that PCNA functions in DNA replication (Prelich et al., 1987a) and established its identity with the DNA polymerase 6 auxiliary factor described by Downey and her colleagues Prelich et al., 1987b;Tan et al., 1986). We employ the designation PCNA for the protein because of the historical precedence of this term and to avoid confusion with the unrelated cyclins, which have been described in frogs (see Dunphy and Newport, 1988 for recent review), clams (Swenson et al., 1986), sea urchins (Evans et at., 1983), and yeast (Booher and Beach, 1988).
The growth-related properties of PCNA, as well as many of its chemical properties such as isoelectric point, molecular weight, and antigenicity, have been well conserved during evolution (Celis et al., 1984), suggesting that it plays a vital role in cellular metabolism. This inference has been upheld by investigations of its biochemistry. PCNA is required for replication of an SV40 DNA template in vitro in extracts from human 293 (Prelich et al., 1987a) or HeLa cells (Wold et al., 1988). Antibodies directed against PCNA reduce plasmid and chromosomal DNA replication in microinjected frog eggs (Zuber et al., 1989). PCNA stimulates the processivity of DNA polymerase 6 (Tan et al., 1986), and PCNA from calf thymus can specifically stimulate yeast DNA polymerase I11 (Burgers, 1988). PCNA is required for synthesis of the leading strand of SV40 DNA in human 293 cell extracts (Prelich and Stillman, 1988), possibly by overcoming the action of a cellular inhibitor to elongation of the replication fork (Lee et al., 1988).
This direct link to DNA metabolism, coupled with the cyclic pattern of PCNA synthesis in the cell cycle and the protein's relationship with cell proliferation, led to the suggestion that the abundance of PCNA may function as a means of regulation of DNA synthesis. Support for this idea was drawn from examinations of PCNA immunofluorescence showing localization at sites of DNA replication Bravo and MacDonald-Bravo, 1985;Madsen et al., 1987) and revealing a peak of staining intensity in the S phase of the cell cycle Madsen and Celis, 1985;Sadaie and Mathews, 1986;. However, it is apparent that results obtained by immunofluorescent staining for PCNA may have been misleading in that they seriously underestimate the amount of the protein in the cell.  recently demonstrated that previous studies of PCNA immunofluorescence employing methanol as a fixative missed a form of the protein that is easily detectable by other methods even in quiescent cells. This observation undermines the assumption that the Regulation of PCNA 13857 abundance of PCNA fluctuates with the cell cycle. Moreover, Wold et al. (1988) observed a constant level of PCNA in cycling HeLa cells, but this observation has not been reconciled with the previously reported changes in synthesis of the protein.
In this report we re-examine the synthesis, stability, and accumulation of PCNA during the cell cycle in HeLa cells synchronized by centrifugal elutriation. In contrast to the large activation of PCNA synthesis in cells traversing from Go to S phase (Almendral et al., 1987;Matsumoto et al., 1987;Jaskulski et al., 1988a), we find only a 2-3-fold variation in the synthesis of PCNA in cycling HeLa cells, with a peak in early S phase. This fluctuation of PCNA synthesis in the cell cycle reflects comparable changes in the level of PCNA mRNA. Although the synthesis of PCNA is cyclic, and the amount of the protein/cell increases, the ratio of PCNA to total cell protein remains roughly constant throughout the cell cycle. We also show that only a portion of the pool of PCNA protein is tightly associated with the nucleus, presumably in a replication complex, at the peak of S phase. We conclude that the cyclic synthesis of PCNA in HeLa cells serves to maintain a large pool of free PCNA throughout the cell cycle, and that PCNA does not exhibit a true "cyclin" pattern of abundance.

MATERIALS AND METHODS
Cell Culture-HeLa cells were grown in F13 medium (GIBCO) supplemented with 5% calf serum. When cultured after elutriation, 1.25 pg/ml fungizone was added to the medium.
Centrifugal Elutriation-Centrifugal elutriation of HeLa cells was an adaptation of a method described previously (Noga et al., 1986) with the modifications given below. Two liters of HeLa cells at a density at 5 X lo5 were harvested by centrifugation at 2000 rpm for 10 min in a Sorvall H-6000A rotor. All further operations were at 0-4 "C in PBS+ (phosphate-buffered saline containing 0.1% glucose, 0.3 mM EDTA, and 1% calf serum). The cell pellet was transferred to a 250-ml bottle and centrifuged again at 1200 rpm in the same rotor. The cell pellet was resuspended in 80 ml, and 5 ml of this unfractionated cell population were removed and placed on ice for analysis later. The remaining cells were loaded at 65 ml/min into a Beckman JE-1OX elutriator rotor running at 1550 rpm. Cells were removed from the rotor in 750-900-ml fractions by increasing the flow in increments of 11 ml/min. The first fraction, which was largely cellular debris, was discarded. Essentially all the cells were removed from the rotor at a flow rate of 190 ml/min. MgC12 was added to 2.5 mM, and the synchronous cell populations were concentrated by centrifugation a t 2500 rpm for 10 min. The cells were transferred to a 50-ml conical tube and brought to 50 ml with PBS+. The cell number in each fraction was determined with a hemacytometer, and aliquots were removed for further analyses. For elutriation progression experiments, the entire fraction was resuspended at 4 X lo5 cells/ml in F13 medium plus serum and stirred at 37 "C. At various times during growth, samples were removed for pulse labeling or other analyses as described below. Flow Cytometry-1O6 cells were removed for analysis by fluorescence-activated cell sorting (FACS). The cells were concentrated by centrifugation and resuspended in 0.3 ml of P B S . After the dropwise addition of 0.9 ml of 4 "C ethanol with mixing, the samples were left at 4 "C for at least 12 h. The fixed cells were washed twice with PBS+ and resuspended in 0.9 ml of the same solution. To this suspension was added 0.1 ml of a 50 pg/ml solution of propidium iodide (Sigma) in 38 mM sodium citrate at pH 7.0, and 10 pl of a 10 mg/ml solution of boiled RNase A. Then the sample was incubated at 37 "C for 30 min. Total fluorescence intensities were determined by quantitative flow cytometry using an Epics C System (Coulter Electronics) equipped with a 5-watt argon-ion laser. The analysis was based on the accumulation of 10,000 cells.
Methionine Labeling-For pulse labeling, IO6 cells were removed from each fraction and concentrated by centrifugation. The cell pellet was resuspended in 2 ml of methionine-free F13 medium containing 0.5-1 mCi of [35S]methionine (Translabel, ICN Radiochemicals). Growth was continued for the period indicated.

Pulse-Chase
Experiments-Three fractions representing the GI, S and G, stages of the cell cycle were resuspended in methionine-free medium (6 x IO6 cells/l5 ml) with 0.5 mCi/ml [35S]methionine. Cells were labeled for 1 h, concentrated by centrifugation, washed with phosphate-buffered saline, then resuspended in 15 ml of F13 medium plus 5% calf serum. Two additional cultures from each stage of the cell cycle were grown identically and used for l3H]thymidine labeling and for FACS analysis. Aliquots were taken from each culture during the chase period to prepare samples for immunoprecipitation or twodimensional gel analysis, FACS analysis, and [3H]thymidine labeling.
Preparation of Extractable PCNA-Cell lysates were prepared by lysis with a nonionic detergent (Nonidet P-40) as described previously (Mathews et al., 198413). The nuclear residue was discarded.
Assay of Total PCNA Synthesis-After two washes with phosphatebuffered saline, the cell pellet was resuspended in 150 pl of boiling dSDS (0.3% SDS, 1% /j"mercaptoethanol, 0.05 M Tris, pH 8.0) then 15 p1 of DNase/RNase (1 mg/ml DNase, 0.5 mg/ml RNase, 0.5 M Tris, pH 7.0, and 0.05 M MgCl,) was added. After incubation on ice for 1 min the samples were frozen at -70 "C. Proteins were fractionated on high resolution two-dimensional gels in the Cold Spring Harbor Laboratory Quest facility as described previously .
Isolation of Bound PCNA-Nuclei were prepared as described previously (Greenberg and Ziff, 1984). 3 X lo6 nuclei from each fraction were washed three times with an isotonic saline solution and tightly bound chromosomal proteins were prepared and analyzed by immunoblotting as described previously (Sadaie and Mathews, 1986) with a mouse monoclonal antibody to PCNA (19F4;Ogata et al., 1987) provided by E.M. Tan (Scripps Clinic and Research Foundation, San Diego, CA). Immunoprec@itation-Immunoprecipitations were performed with an equal amount of trichloroacetic acid precipitable radioactivity in each extract using a human autoantiserum to PCNA (provided by R. M. Bernstein, University of Manchester, Manchester, UK) as described previously (Mathews et al., 1984b). Control experiments indicated that both antibody and protein A immunosorbent were in excess. Immunoprecipitates were fractionated on a 15% SDS-polyacrylamide gel (Laemmli, 1970). Gels were quantitated by direct scanning of the dried gel with an Ambis Beta Scanner.
RNA Analysis-Cells from each fraction were lysed in 2 ml of Nonidet P-40 lysis buffer as described previously (Greenberg and Ziff, 1984). Nuclei and other insoluble material were removed by centrifugation for 5 min in a Sorvall GLC centrifuge. The supernatant was removed and added to another tube containing 2 ml of TSES (10 mM Tris, pH 7.6,0.15 M NaCl, 5 mM EDTA, and 1% SDS) and 4 ml of water-saturated phenol at pH 7.0. After extraction with phenol (twice), phenol/CHCls (once), and CHCl, (once), the samples were precipitated with ethanol. 20 pg of cytoplasmic RNA from each fraction were analyzed by Northern blot analysis with a 1% agarose gel as described by Maniatis et al. (1982). For RNA dot-blot analysis, RNA from each fraction was diluted with water to 50 pl, then 150 pl of a 1:l mixture of formaldehyde/20 X SSC was added. After heating at 68 "C for 15 min the samples were applied to a nitrocellulose filter in a Bethesda Research Laboratory HYBRI'DOT manifold with 20 X SSC washes. Hybridizations were conducted at 60 "C in 50% formamide, 5 X SSC, 0.1% Ficoll, 0.1% polyvinylpyrrolidone, 50 mM sodium phosphate, pH 6.5, 2 mM EDTA, 0.5% SDS, and 25 pg/ml polyadenylate. The filters were washed three times at 65 "C for 30 min each in 1 X SSC, 20 mM Tris, pH 7.4, 2 mM EDTA, and 0.5% SDS. In some cases the filters were treated with 10 pg/ml RNase A in 2 X SSC for 30 min at 37 "C.
The plasmid He-7, which codes for a cellular mRNA that does not fluctuate with the growth state of the cell (Kao and Nevins, 1983;Rittling et al., 1986), was provided by J. Nevins (Duke University, Durham, NC). A histone H4 genomic clone, pHu4A (Zhong et al., 1983), was provided by N. Heintz (Rockefeller University, New York, NY). High specific activity RNA probes were prepared by subcloning the desired insert into the vector pGEM-1 (Promega Biotec), which was transcribed as a linear plasmid with SP6 polymerase as described by Promega Biotec in the presence of [w3'P]UTP. DNA probes were prepared by random priming of gel purified DNA fragments with an oligolabeling kit (Pharmacia LKB Biotechnology Inc.) and [a-"P] dNTP.

RESULTS
Cell Fractionation-We fractionated cycling HeLa cells by centrifugal elutriation. This method separates cells on the basis of their sedimentation velocity and allows large numbers of nearly synchronous cells to be obtained from a mixed population. The smaller GI cells are removed first, followed by S phase cells, and the larger G2/M phase cells are removed last from the elutriator. The degree of synchrony obtained in each fraction from the elutriator can be assessed from the data presented in Fig. 1. A portion of each fraction of cells was pulse labeled with [3H]thymidine for 30 min to measure DNA synthesis. The peak of [3H]thymidine incorporation was in fraction 6 ( Fig. 1A). Another portion of cells from each fraction was fixed, and the distribution of cellular DNA contents within the fraction's population was determined by FACS (Fig. 1B). Fractions possessing the largest percentage of cells in each phase of the cell cycle were estimated from the FACS analysis. We estimate that 90% of the cells in fraction 2 possess a G, DNA content, about 70% of the cells in fraction 6 possess a DNA content typical of S phase cells, and 83% of the cells in fraction 10 possess a G1 DNA content.
We have fractionated up to lo9 cells in this manner and have routinely obtained results similar to those illustrated in Fig.  1. More importantly, this method has allowed us to obtain enriched populations of cells in different phases of the cell cycle while minimizing physiological stress of the sort that occurs when cells are synchronized by drugs or by serum deprivation. The results shown in Figs. 2 A , 3, A and D, 5, and 6 were obtained with the cell fractions that were examined in Fig. 1. PCNA Synthesis-PCNA synthesis was measured in each fraction of cells shown in Fig. 1. A portion of cells from each fraction was pulse labeled with [%3]methionine for 30 min, then the cells were lysed with a nonionic detergent, and the insoluble material was removed by centrifugation. The supernatant from this procedure, containing the extractable PCNA, was assayed for the incorporation of [%]methionine into PCNA by immunoprecipitation with an autoantibody directed against PCNA. The results of this experiment, displayed in an SDS-polyacrylamide gel, showed a slight increase in PCNA synthesis in fractions corresponding to early S phase cells ( Fig. 2 A ) . Although this particular experiment was not quantified, in similar experiments we measured a 2-3-fold increase in PCNA synthesis by direct scanning of the gel-fractionated immunoprecipitated material. In case elutriation itself might impose a transient stress that perturbs the results, PCNA synthesis was also examined throughout the cell cycle by post-elutriation progression. In this procedure, a synchronous population of cells from the elutriator was resuspended in fresh medium and allowed to proceed through the cell cycle. The position of the cells in the cell cycle was determined by [3H]thymidine incorporation and FACS analysis. At the indicated times, aliquots were removed and pulse labeled with [35S]methionine. PCNA synthesis was assayed by immunoprecipitation as described above. Shown in Fig. 2B are the results obtained in such an assay from one fraction, an S phase fraction. The [3H]thymidine counts incorporated a t each time point (shown beneath each lane) indicate that the cells entered the subsequent S phase of the cell cycle during the course of the experiment. The immunoprecipitates were also fractionated on a separate gel which was quantified by scanning directly: in good agreement with direct measurements on elutriated cells, a 2-fold change in synthesis with a peak in early S phase was observed. Similar results were obtained when a GI fraction was allowed to proceed through the cell cycle. T o deal with the possibility that these experiments might not have detected all the PCNA synthesized (because it is not fully extracted, for example) we also analyzed total cell proteins by two-dimensional gel electrophoresis. HeLa cells were nonionic detergent (0.5% Nonidet P-40) and insoluble material was removed by centrifugation. PCNA was immunoprecipitated from approximately 4 X lo5 cpm of trichloroacet.ic acid-insoluble radioactivity of each lysate with anti-PCNA autoantibody. The immunoprecipitates were resolved in a SDS-polyacrylamide gel. The gel was fixed, processed for fluorography, and exposed to x-ray film for 2 weeks. Only the portion of the gel containing the immunoprecipitated PCNA is shown. The lane designations correspond to the fractions shown in Fig. 1. B, PCNA synthesis by post-elutriation progression analysis. A synchronous population of S phase cells from the elutriator was returned to cell culture to continue growth. At 4-h intervals lo6 cells were removed and pulse labeled for 1 h with ["S]methionine.

A.
The cells were lysed with detergent, and PCNA was immunoprecipitated from each extract containing 6.5 X lo' trichloroacetic acid precipitable cpm. The immunoprecipitates were analyzed as in Fig.  2A  FACS analysis (not shown). The [3H]thymidine incorporation indicates that the peak of DNA synthesis was in fraction 5. Cells from each fraction were pulse labeled with [3sS]methionine followed by fractionation of total cell protein by twodimensional gel electrophoresis (Fig. 2C). The amount of PCNA synthesized in each fraction was quantified by densitometric scanning of the two-dimensional gels and the results (in parts/million) are presented in tabular form (Table I).
Again about a 2-fold change in the rate of PCNA synthesis in the cell cycle was observed. The results obtained by the three methods are in good agreement with each other as well as with the early results reported by Bravo and Celis (1980) with HeLa cells synchronized by mitotic selection. PCNA Levels-The results described in the preceding section indicate that there is a 2-3-fold cyclic increase in PCNA synthesis with a peak in early S phase, but the amount of the protein in the cell might vary by a larger factor due to differential stability. T o assay PCNA as a proportion of total cellular protein during the cell cycle, fractionated cells from the elutriator were extracted with a nonionic detergent, and equal amounts of protein from each fraction were analyzed by immunoblotting with a monoclonal antibody to PCNA . As shown in Fig. 3A, during the cell cycle the extractable PCNA remains nearly constant as a percentage of cellular protein. T o circumvent any possible artifacts associated with elutriation, a similar immunoblot analysis was also performed with samples from the post-elutriation progression fractions shown in Fig. 2B. This experiment also revealed little change in the amount of PCNA as a function of the cell cycle (Fig. 3 B ) . Furthermore, to ensure that no appreciable quantity of PCNA was missed by the extraction procedure, equal amounts of total cellular protein from each fraction of an elutriation were analyzed by immunoblotting with essentially identical results (Fig. 3C). The slight variais shown the trichloroacetic acid precipitable cpm of ['HJthymidine incorporated in each time point. C, PCNA synthesis during the cell cycle determined by two-dimensional gel electrophoresis. lo6 cells from each fraction of a n elutriation were pulse labeled for 1 h with 0.25 mCi/ml [%]methionine. An equal amount of trichloroacetic acid precipitable radioactivity from each fraction was separated by twodimensional gel electrophoresis in the Cold Spring Harbor Laboratory Quest facility. Fractionation in the first dimension was by isoelectric focusing from high (right) to low pH. Fractionation in the second dimension was by SDS-polyacrylamide gel electrophoresis from the top to bottom. Only the portion of interest of each gel is shown. The arrowhead denotes PCNA. Quantitation of PCNA synthesis by computer-aided densitometric scanning of the gels, as well as [3H]thymidine cpm incorporated in each fraction, are shown in Table I.

FIG. 3. Levels of PCNA protein in cycling HeLa cells. A,
level of extractable PCNA versus cell protein in elutriated cells. 18 pg of protein (Bradford, 1976) from the detergent lysate of each fraction described in Fig. 2A was separated in a SDS-polyacrylamide gel then transferred onto nitrocellulose. The blot was probed for PCNA with a mouse monoclonal antibody followed by '2511-labeled rabbit anti-mouse antibody. The immunoconjugate was detected by exposure of the filter to x-ray film for 4 days with an intensifying screen. The lane designations of T and 1-10 correspond to those for  Table 11) tions seen in Fig. 3, B and C were not reproducible in other immunoblots of these and other samples. Therefore, we conclude that the amount of PCNA changes little in relation to other cellular proteins during the cell cycle. Because the protein content of a cell rises through the cell cycle and the peak of PCNA synthesis is in early S phase, we anticipated that PCNA levels would increase at S phase when measured on a per cell basis. Fig. 3 0 shows an examination of the same samples as in A, loaded such that each lane contains extract from an equal number of cells. Consistent with expectations, the cellular content of PCNA remained steady during GI, increased during S phase, and remained at a fairly steady higher level during Gz.
PCNA Stability-Two replication-related proteins, thymidine kinase and topoisomerase 11, are destabilized as the cell exits from mitosis into the GI phase (Shirley and Kelly, 1988;Heck et al., 1988). It is possible that a brief period of instability might be a general property of proteins involved in DNA metabolism. Although the results presented in the preceding sections argue against the destruction of a large fraction of cellular PCNA at any stage of the cell cycle, they do not address the question directly. T o examine the stability of PCNA through the cell cycle, several fractions of cells from the elutriator were pulse labeled with ["S]methionine and grown for 20 h in the presence of a 100-fold excess of unlabeled methionine. The results achieved with an S phase fraction are shown in Fig. 4. The ["]thymidine radioactivity incorporated (Fig. 4A) and DNA content determined by FACS analysis (B) indicate that the cells entered the subsequent S phase of the cell cycle. As expected for cells progressing from one S phase to the next, the cell number roughly doubled during the course of the experiment (Fig. 4A). The quantity of radiolabeled PCNA remaining was determined at 4 hourly intervals during the chase period by immunoprecipitation from extracts containing equal amounts of total radioactivity (Fig. 4C). There appeared to be no change in the stability of PCNA in relation to total cellular protein as a function of the cell cycle. This observation was also verified by two-dimensional gel electrophoresis of whole cell extracts prepared at the same times during the chase (not shown).  determined the half-life for PCNA in 3T3 cells in the transition from growth to quiescence to be about 20 h. The half-life for PCNA in cycling HeLa cells observed here is similar to the half-life of total cellular protein, or about 8 h (Fig. 4A). The difference in the two estimates probably arises from differences in cell type and experimental protocols. Despite this difference, the important feature of both observations is that no abrupt change in the stability of PCNA was observed.
PCNA rnRNA-To study the level of PCNA mRNA during the cell cycle, we isolated cytoplasmic RNA from the synchronized cell populations shown in Fig. 1. A Northern blot containing equal amounts of total cytoplasmic RNA from each fraction (designated as in Fig. 1) was probed for PCNA mRNA, as well as for the mRNA coded for by pHe-7, which does not fluctuate with the cell cycle (Kao and Nevins, 1983;Rittling et al., 1986), and histone H4 mRNA (Zhong et al., 1983), expressed only during S phase. As expected, the concentration of the mRNA coded by pHE-7 did not vary, and the H4 mRNA concentration peaked in those fractions enriched for S phase cells (Fig. 5A). There was only a slight change in PCNA mRNA levels during the cell cycle. In order to estimate the magnitude of this change, RNA dot blots were probed for PCNA and He-7 sequences, and the fractions containing the minimum and maximum levels of PCNA mRNA were identified (Fig. 6B). Fractions 1 and 10, enriched GI and GZ populations, respectively, contained the lowest levels of PCNA mRNA; fraction 5, an early S phase population, contained the highest level of PCNA mRNA (see Fig.  1). RNA dots containing decreasing amounts of RNA from An aliquot of the chase sample was also assayed a t each time point for trichloroacetic acid precipitable ["S]methionine counts (M). The cell number of a third parallel culture intended for FACS analyses was determined by counting aliquots with a hemacytometer a t each time point these fractions were probed in Fig. 6C. This and three similar experiments reveal an approximately 2-fold change of PCNA mRNA levels as a function of changes in the cell cycle. The magnitude of this change reflects similar changes in the amount of PCNA protein synthesis in each fraction (see Figs.  1 and 2 A ) .
Bound PCNA-Recently, Bravo and Macdonald-Bravo (1987) inferred the existence of two different forms of PCNA. The free form stains diffusely in the nucleoplasm if formaldehyde is employed as a fixative but is not detected after methanol fixation. The bound form is tightly associated with chromatin and appears as bright foci after methanol fixation. DNase treatment has been used to release PCNA retained in the nuclear pellet after salt extraction (Sadaie and Mathews,

A.
T I 2 3 4 5 6 7 8 9 1 0 " " " " T I " " "  Fig. 1 and 20 pg of the RNA from each fraction was separated on a formaldehyde-agarose gel that was subsequently transferred to a nitrocellulose filter. The filter was probed by hybridization with radioactive RNA (PCNA and HE-7) and DNA ( H 4 ) probes. The lane designations refer to the fractions illustrated in Fig. 1. R, dot blot analysis of RNA from elutriated cells. Ten pg of formaldehyde-treated cytoplasmic RNA from each fraction was applied to a nitrocellulose membrane in a dot blotting manifold. One set each of the RNA dots was hybridized to radioactive RNA probes for PCNA and HE-7. After hybridization the RNA-RNA hybrids were treated with 10 pg/ml RNase A in 2 X SSC a t 37 "C for 30 min. The filter was exposed to an x-ray film with an intensifying screen for 48 h. C , quantitation of PCNA mRNA levels by dot blot analysis. RNA from selected fractions was spotted onto nitrocellulose in decreasing amounts from 10 pg to 0.5 pg and examined as in Fig. R. 1986). We employed this method to examine the bound PCNA from the synchronized cell population of Fig. 1. As shown in Fig. 6, about 10-fold more PCNA remains associated with the nuclear pellet from S phase cells than from GI or Gz phase cells. This tightly bound fraction is about 35% of the total PCNA in the S phase cell (see below), but presumably it includes the majority of PCNA that is actively involved in DNA replication, i.e. in replication complexes.

DISCUSSION
The term "cyclin" was first employed by Bravo and Celis to denote a protein that upon two-dimensional gel electrophoresis appeared to be synthesized in increased amounts in Regulation of PCNA T I 2 3 4 5 6 7 8 9 1 0 FIG. 6. Chromatin-bound PCNA in cycling HeLa cells. 3 X 10" nuclei prepared from the fractions shown in Fig. 1 were washed three times with an isotonic saline solution. Then the nuclear pellet was treated for 5 min with DNase followed by boiling in SDS gel loading buffer (Laemmli buffer). Proteins were fractionated in an SDS-polyacrylamide gel and electrophoretically transferred to a nitrocellulose membrane. PCNA was detected in each fraction by a mouse monoclonal antibody to PCNA followed by an '2sI-labeled rabbit anti-mouse antibody. The filter was exposed to x-ray film for 1.5 days with an intensifying screen. The lane designations correspond to those in Fig. 1. transformed or normal growing cells . Subsequently, Mathews et al. (1984a) showed that this protein is identical to an entity described previously by Miyachi et al. (1978) as the proliferating cell nuclear antigen, PCNA. Hunt and his colleagues (Evans et al., 1983) identified proteins, also called "cyclins," in developing sea urchin and clam embryos that exhibited a cyclic pattern of expression and implicated the protein in control of mitosis. Shortly after mitosis, sea urchin cyclin is rapidly degraded. Two proteins in mammalian cells, thymidine kinase and topoisomerase 11, turn-over rapidly in the transition from mitosis to GI in an analogous fashion . The turn-over of these proteins or cyclins may be related to a resetting of the cellular program after mitosis. Although the synthesis of PCNA fluctuates to a modest degree with the cell cycle, we show here that PCNA is not a true cyclin.
We examined PCNA synthesis in cycling HeLa cells synchronized by centrifugal elutriation. This method was chosen to avoid possible nonphysiological effects caused by other methods of cell synchronization. Our data agree with early work by Bravo and Celis (1980) showing a 2-3-fold increase of PCNA synthesis that peaks during the early S phase of cycling cells, although subsequent publications seem to show larger fluctuations (see for example, Celis and Celis, 1985;. We show here that this change reflects a comparable change in the PCNA mRNA level. Despite the change in synthetic rate for the protein, the amount of protein changes very little during the cell cycle. These results agree with the determination of Wold et al. (1988) that there is little change in the level of PCNA in HeLa cells as a function of the cell cycle. It appears that the cyclic pattern of synthesis of PCNA simply maintains the protein at a roughly constant level during the cell cycle.
The issue of PCNA regulation in cycling cells was confused by inferences drawn from the cyclic pattern of PCNA immunofluorescence. Because PCNA in immunofluorescence was only detected in S phase nuclei, it was assumed that the level of the protein fluctuates in a corresponding way during the cell cycle. However,  have recently shown that a large pool of presumably free PCNA remains undetected by immunofluorescence if methanol is used as a fixative. As shown here, although total PCNA changes little during the cell cycle, the fraction of PCNA that is tightly bound is maximal during S phase of the cell cycle, but this fraction is only 35% of the total PCNA. We believe that the bound PCNA represents the fraction actively engaged in replication which explains the cyclic immunofluorescent pattern.
The number of molecules of PCNA/cell can be estimated from our immunoblots by comparing the intensity of the signal from the various cell populations to that obtained with dilutions of purified PCNA on the same blot. Estimates of the number of molecules of PCNA/cell throughout the cell cycle are in Table 11. The ratio of PCNA to total cell protein (0.015%) appears to be constant throughout the cell cycle. This value is 4-fold lower than the value reported previously for HeLa cells (0.06%) by ; the discrepancy probably reflects differences in methods for estimating both PCNA and total cell protein (immunoblots and Bradford assays versus incorporation of [14C]amino acids). The estimate that about 35% of the total PCNA is tightly bound during the peak of DNA synthesis in S phase corresponds to about 2 X IO5 molecules/cell actively involved in DNA synthesis during this period. This number compares well with the estimated number of origins in a human cell, which is about 2 x lo4 to 2 x lo5 origins/diploid genome (Hand, 1978). Since HeLa cells are polyploid (Hay et al., 1985), and only a fraction of the origins are active a t any given time, the ratio must be considered approximate, but it seems likely that only one or a t most a few PCNA molecules are tightly bound at the growing fork during DNA synthesis.
It is widely accepted that the proteins involved in DNA metabolism are expressed in greater quantities in the S phase of the cell cycle (Johnson, 1984). This belief is somewhat exaggerated, since many of the proteins specified as S phase specific were identified by stimulation of quiescent cells from Go to S, and events in the Go/S transition are not always mirrored in the GI/S transition of the normal cell cycle.
Whereas the mRNA levels for these proteins increase in response to proliferative signals, there are few common features of their expression in the cell cycle. In Saccharomyces cerevkiae the mRNA levels for such proteins as thymidine kinase, DNA ligase, thymidylate synthetase, and DNA polymerase I peak at the GI/S boundary of the cell cycle (Johnston et al., 1987;White et al., 1987). In Schizosaccharomycespombe, however, the mRNA level for DNA ligase does not change during the cell cycle . In mammalian cells, experiments designed to reassess the cell cycle expression of some of these proteins through less severe means of achieving cell synchrony than serum starvation or drug treatment, have produced conclusions similar to those presented here. For example, recently it has been demonstrated that the large increase in thymidine kinase mRNA observed after stimulation of quiescent cells does not occur in cycling HeLa cells (Sherley and Kelly, 1988).
One might presume that the expression of proteins involved in DNA metabolism would be controlled by common mechanisms. We show here that one aspect of protein metabolism that is shared by topoisomerase I1 and thymidine kinase, their

TABLE I1
Levels of PCNAIcell throughout the cell cycle The quantity of total protein/cell was determined by Bradford assay of whole cell extracts. PCNA was estimated from immunoblots of total protein by comparison of signal intensities to that observed with decreasing quantities of purified PCNA (Fig. 3C). The molecular weight for PCNA was determined from the cDNA sequence to be 29,261 daltons (Almendral et al., 1987) rapid degradation after mitosis, does not occur for PCNA. Yet the magnitude of the change of PCNA mRNA and protein levels in the cell cycle is similar to that of DNA polymerase a in cycling cells (Thommes et al., 1986;Wahl et al., 1988).
Perhaps this similarity of expression reflects the close functional relationship of the two proteins. Is PCNA Involved in Regulation of DNA Synthesis?-PCNA synthesis is not tightly linked to DNA synthesis. Serum starved quiescent cells stimulated to grow by the addition of serum display elevated levels of PCNA synthesis even in the presence of the DNA synthesis inhibitors hydroxyurea or aphidicolin MacDonald-Bravo and Bravo, 1985;Almendral et al., 1987;Kurki et al., 1987). Cycloheximide does not prevent recruitment of PCNA into presumed DNA damage repair clusters in serum-starved quiescent cells . We have shown here that PCNA is present in fairly constant amounts during the cell cycle, but the amount that binds tightly to chromatin increases to a maximum during the peak of S phase. We interpret this result to be a consequence of recruitment of PCNA from the soluble pool into replication complexes. The manner in which PCNA is recruited into replication complexes remains unknown. A minor, more acidic, PCNA-related spot can be detected on two-dimensional gels, but the ratio of this minor component to the major PCNA spot neither varies as a function of cell growth (Mathews et al., 1984a) nor have any growth related post-translational modifications of PCNA been identified (Bravo and Celis, 1985;Sadaie and Mathews, 1986). Segregation of nascent PCNA from preexisting PCNA has also been investigated as a possible explanation for the migratory behavior of the protein but no evidence was found in support of that scenario .
Since PCNA appears to be in excess of the amount required even a t maximal levels of DNA synthesis, it seems unlikely that the amount of the protein exerts a strong influence on regulation of DNA synthesis in HeLa cells. Yet the amount of PCNA may have a bearing on DNA synthesis, since antisense oligodeoxynucleotides to PCNA inhibit growth of 3T3 cells (Jaskulski et al., 1988b). Adenovirus infection, which mobilizes the cell to replicate the viral genome, also activates synthesis of PCNA through the function of E1A (Zerler et al., 1986). Furthermore, DNA replication is generally accompanied by PCNA synthesis. These findings suggest that the cell requires PCNA for DNA synthesis and monitors a t least some aspect of PCNA metabolism.