REGULATION OF CELL PROLIFERATION AT THE ONSET OF DNA … · 2013-05-23 · Regulation of onset of DNA synthesis 173 regulate it. For example, removal of extracellular phosphate or

J. Cell Sci. Suppl. 4, 171-180 (1986)Printed in Great Britain © The Company of Biologists Limited 1986

171

REGULATION OF CELL PROLIFERATION AT THE ONSET OF DNA SYNTHESIS

A R T H U R B. P A R D E E , D ON A LD L . C O PPO C K a n d H E N R Y C. YANG

Department o f Pharmacology, H arvard M edical School and Division o f Cell Growth and Regulation, D ana-Farber Cancer Institute, 44 Binney St, Boston, MA 02115, USA

INTRODUCTION

General outlines are now emerging as to how cellular proliferation is regulated. Growth factors provide external signals that govern proliferation. These factors interact with their receptors to set in motion intracellular signals, sometimes called second messages, which in turn activate specific genes including oncogenes and initiate key biochemical events.

Duplication of all of a cell’s initial components furnishes the substance for eventual cell duplication. These components come together to permit the readily observed biochemical events, such as the onset of DNA synthesis at the beginning of S phase and morphological terminal events of nuclear and cellular divisions. Relatively few growth-factor-regulated biochemical processes are critical for cell proliferation, amongst this myriad of events during the cell cycle. All present evidence suggests that these controlling events occur in only one part of the cycle, G\, that culminates in the onset of DNA synthesis. All that follows - in S, Gz, M and cytokinesis — proceeds and goes to completion independently of extracellular regulatory factors. Homeostatic control mechanisms must balance the events in the different phases, as well as those in G\ itself. However, these controls are secondary in contrast to the few primary regulations of the most critical processes, of which onset of DNA synthesis is a major one. In this essay, the control of onset of DNA synthesis will be discussed. We will discuss primarily results obtained with some fibroblastic rodent lines such as 3T3.

PROLIFERATION CONTROL IN RELATION TO THE CELL CYCLE

Arrest of proliferating cells selectively in any part of the cell cycle other than G\ is not possible by manoeuvres that are dependent on physiological agents (as opposed to inhibitory drugs). Furthermore, cessation of growth under physiological con­ditions stops cells in intact animals or in culture with unduplicated DNA contents. That is, arrested cells exist in the quiescent (G0) state before they can make DNA. Furthermore, the derangements of proliferation control found in neoplastic cells are all Go/Gi-related. These pieces of information focus attention on the Gq/G x con­dition when proliferation and its derangements are considered.

Cell cycle progression

M G\ S (?2 A/---------R - > --------------------- ►— ► >

C -----------------yV

tGo

__ 1________________ I______________ L-6 0 6

Time (h)

Fig. 1. Progression through the cell cycle. Stages of the cell cycle and major regulatory points are indicated at times representing the fastest proliferating mouse 3T3 cells. Note that duration of Gi period before R is highly variable in a population and depends on growth conditions. The diagram also indicates convergence of pathways from mitosis by cycling cells and recovery of initially quiescent cells.

This topic is described conventionally (Baserga, 1985) in the framework of a temporal sequence leading from quiescent cells into and through the cell cycle. The interval from Go to the onset of DNA synthesis (S ) can now be divided into four subsections, as shown in Fig. 1. Cells are stimulated to emerge from quiescence into a competent state (C), which has been studied in most detail with fibroblastic 3T3 cells (Stiles, 1983). Then during a period of about 6h they come to a point (V), at which they are similar to cycling cells that have just completed mitosis (M ). Following V (or M) cells usually require several more hours in G\ to prepare for DNA synthesis. They then reach a control point (R ), beyond which neither growth factors nor transcription and rapid translation are required (Yang & Pardee, 1986). Between R and the onset of DNA synthesis (start of 5 phase) is a period for organization of the machinery for DNA synthesis. Once DNA synthesis has started, the rest of the cycle — S, Gz,M and cytokinesis - follows independently of external physiological controls. Cells that have started DNA synthesis will reach the next G\ phase, and then will either proceed through another cycle if conditions are adequate, or fall back into quiescence (Go) from which they will emerge only when restimulated.

Hydroxyurea and aphidicolin, which block ribonucleotide reductase and DNA polymerase, respectively, prevent the synthesis of DNA, and the events that follow, such as mitosis and cytokinesis. Similarly, blocking cells in mitosis with agents such as colcemid also prevents cytokinesis. These results suggest that each event is dependent upon prior events in the cycle. But there is no method for stopping cells in S, Gz or M specifically that depends upon normal physiological conditions such as growth factors or high cell density.

Although these statements are based mostly on studies with fibroblasts, similar events have been described for other kinds of cells such as lymphocytes (for a recent review, see Baserga, 1985). Not every factor required for cell growth is a physio­logical regulator. Many components are needed for growth, but all do not normally

172 A. B. Pardee, D. L. Coppock and H. C. Yang

_l________L14 16

Regulation of onset of DNA synthesis 173regulate it. For example, removal of extracellular phosphate or potassium ions causes cells to stop growing, but there is little likelihood that these ions are crucial to controlling the decision as to whether or not a cell will grow. It is much more likely that the levels of factors such as platelet-derived growth factor and somatomedin C are critical for determining whether or not normal cells will grow. The same factors seem to be less essential for cancer cells, thereby permitting neoplastic growth.

Various sorts of events are under study during each of these successive steps, as listed in Table 1. We cannot in this brief essay begin to summarize all of this information, and how they relate to each of the half-dozen parts of the cell cycle. We will discuss some that are involved in the onset of DNA synthesis, in particular during the interval between the termination of cell division (M ) (or the V point) and the beginning of S phase. We will work progressively backwards from each effect to its prior event.

THE ONSET OF DNA SYNTHESIS

DNA synthesis begins suddenly, many hours after a quiescent cell is stimulated or a cycling cell has completed mitosis. The remainder of the cell cycle then proceeds according to a quite exact time schedule, compared to the highly variable Go to S interval. The onset of DNA synthesis is usually measured by determining incor­poration of radioactive thymidine into trichloracetic acid insoluble material (DNA). This method is subject to numerous possible errors, particularly if quantified as counts incorporated rather than by autoradiography, which determines whether or not a cell is incorporating thymidine. The amount of thymidine incorporated depends upon factors such as intracellular pools, activity of the thymidine transport mechanism, and activities of enzymes including thymidine kinase; but it is still by far the most convenient and most frequently used method. If one exposes an exponen­tially growing population of cells to thymidine, about half the cells are observed to incorporate radioactivity immediately, as shown by autoradiography (Fig. 2). These are the cells initially in S phase and synthesizing DNA. Thereafter, more cells become labelled, as they enter 5 phase. So as time progresses the curve measures transit into 5 of cells from earlier parts of the cycle, initially from G\ and later from Gz. G\ cells constitute the majority of this remaining population, as can be shown in studies in which colcemid is added to prevent eventual entry of the Gz and M cells into 5 phase.

Table 1. Cell proliferation components1. Growth factors2. Membrane receptors3. Second messengers4. Activated genes5. Messenger RNAs6. Proteins7. Enzymes

174 A. B. Pardee, D. L. Coppock and H. C. YangWhat event is it that suddenly permits a cell to start making DNA? Many of the

required enzymes are present in Gj cells; and the availability of DNA precursors does not seem to be limiting, since when these are supplied to cells made permeable to deoxynucleoside triphosphates new DNA only appears in the same time frame as in intact cells (Castellot, Miller & Pardee, 1978), which must make their own deoxynucleoside triphosphates through a series of enzyme-catalysed events starting with the ribonucleoside diphosphates.

T h e final incorporation of deoxynucleotides into DNA depends upon at least a dozen enzymes and other proteins, as demonstrated with lower organisms primarily by Kornberg (1980) and co-workers. In higher organisms a multiprotein ‘replitase’ complex contains many of the known enzymes that are required for DNA synthesis (Reddy & Pardee, 1980). Such a complex with a molecular weight of about 5 x l 0 6 has been purified about 10-fold and shown to contain nascent DNA as well as a considerable number of enzymes involved in DNA synthesis (Noguchi, Reddy & Pardee, 1983). Until these enzymes and associated proteins are assembled the complex should not function. A reasonable hypothesis is that the final step leading to onset of DNA synthesis is production of one of these enzymes or proteins. Indeed, a number of enzymes rise dramatically at the same time as DNA synthesis starts (Coppock & Pardee, 1985; Szyf et al. 1985). Their appearance is not dependent on DN A synthesis, unlike histones, which are made coordinately with DNA (Sariban, Wu, Erickson & Bonner, 1985). However, none of these has been demonstrated

Fig. 2. Entry of an exponential population of cells into S phase. Exponential populations of 3T3 cells received [3H]thymidine ([3H ]dThd) at time zero. Half also received 0-1 jUgml- cycloheximide (CH M ), which was removed after 4h (O ). The other culture did not receive cycloheximide (• ) . Samples were taken for autoradiography at intervals as shown. The data are plotted as percentage of cells with labelled nuclei. (Modified data from Campisi & Pardee (1984).)

Regulation of onset of DNA synthesis 175to provide the essential final element for incorporation of deoxynucleotides into DNA.

Interestingly, this complex was not found in G\ phase cells, but only in S phase cells; its appearance correlated with the time at which DNA synthesis started (Reddy & Pardee, 1980). Formation of this complex from its components could thus be the signal that is required to trigger DNA synthesis. Remarkably, enzymes of the complex were not even found in the nuclei of pre-S cells, but rather in their cytoplasm; they were located in the nuclei only after DNA synthesis had begun (Reddy & Pardee, 1980). Possibly the nucleus permits facilitated diffusion of these enzymes shortly before DNA synthesis starts. An alternative and simpler hypothesis, however, is that the enzymes can penetrate freely into the nucleus but they are trapped there only when they are able to assemble into the multienzyme complex (See Feldherr, 1985). The assembly of enzymes into the nucleus just at the time when DNA synthesis starts requires explanation.

Studies in vivo also suggest interaction of enzymes required for DNA synthesis. Remarkably, an inhibitor of one of the enzymes blocks the activity of another within the intact cell, although not inhibiting this second enzyme in extracts. Thus, aphidicolin, which is an inhibitor of solubilized DNA polymerase but not of thymidylate synthase, strongly inhibited both of these activities in intact cells. Hydroxyurea, a specific inhibitor of ribonucleotide reductase and novobiocin, a preferential inhibitor of topoisomerase, similarly also blocked thymidylate synthase within the intact cell (Reddy & Pardee, 1983). The degree of thymidylate synthase inhibition has been found by others (Chiba, Bacon & Chiba, 1984; Nicander & Reichard, 1985) to be slightly less than that of the primarily inhibited enzyme. This partial inhibition was under different experimental conditions, and in both cases it was very considerable.

Possible explanations for this cross-inhibition in vivo are: (1) the enzymes are all associated in a multienzyme complex and inhibition of one allosterically affects another, as modulated by protein conformational changes (Reddy & Pardee, 1983). (2) The inhibitor of one enzyme modulates various deoxynucleotide pools, with inhibitory effects upon thymidylate synthase. There is, however, no evidence for large modifications of pool components, or of inhibition by any such components on thymidylate synthase.

Another sort of data supporting the idea of a multienzyme complex for DNA synthesis is a phenomenon called channelling (Reddy & Pardee, 1982). Labelled ribonucleoside diphosphate was preferentially incorporated into trichloroacetic acid- insoluble material relative to the incorporation of deoxynucleoside triphosphate, the proximal DNA precursor. One explanation for this is that enzymes of the pathway to DNA, from ribonucleoside diphosphate reductase may be so closely coupled that only a molecule of ribonucleoside diphosphate entering this sequence is able to be incorporated.

Although such a preferential incorporation into TCA-insoluble material of per- meabilized cells has been demonstrated (Reddy & Pardee, 1982), it has been argued that the product formed is RNA rather than DNA, in which case it is not surprising

176 A. B. Pardee, D. L. Coppock and H. C. Yangthat deoxy-triphosphates do not interfere (Spyrou & Reichard, 1983). More recent work (Reddy, Klinge & Pardee, 1986) shows that the product formed by these cells, which must be permeabilized in order to permit entry of labelled and charged compounds, is not RNA but actually a co-polymer of ribo- and deoxynucleotides. Although this is an artifactual product and not authentic DNA, neither is it RNA. Furthermore, incorporation of ribonucleoside diphosphate into only the deoxy­nucleotides of this product clearly showed selective chanelling. The concept that the enzymes leading from the ribo-precursors into DNA are closely coupled is thus supported.

Tight coupling (exclusion of deoxytriphosphates) can be demonstrated with permeabilized cells, but this can be loosened so as to allow incorporation of both precursors (Reddy & Pardee, 1980). Various experiments with intact cells clearly demonstrate both formation of deoxynucleotide pools (important in regulation of ribonucleotide reductase) and their incorporation into DNA, and therefore leaky channelling (Nicander & Reichard, 1983).

Further work is required in order to demonstrate the degree of channelling in intact cells, and that it is via a multienzyme complex involving all of the enzymes from ribonucleoside diphosphate reductase to DNA polymerase. Many data are, however, very hard to explain without this hypothesis. The sudden onset of DNA synthesis is plausibly accounted for by assembly of some or all of these enzymes into a replitase complex at the time DNA synthesis starts.

THE RESTRICTION POINT

The G\ (M to S) interval and the latter part of the considerable longer interval from Go to 5 is bipartite. Before the restriction (R) point, which is located about 2h before the onset of DNA synthesis, cells require transcription and rapid protein synthesis as well as somatomedin C in order to move on and start DNA synthesis. In contrast, cells beyond the R point are independent of growth factors, of transcription and of rapid protein synthesis. An illustration of this phenomenon is seen (Fig. 2) after addition of the translational inhibitor cycloheximide to a population of exponen­tially growing cells. Cells continued to enter S for approximately 2 h (Campisi et al. 1982). Cells originally located earlier in the cycle were blocked. A variety of experiments show timing corresponding to a critical R point event being reached at about 2 h before onset of DNA synthesis, and also a good correlation between loss of requirement for the growth factor somatomedin C and loss of transcriptional inhi­bition (80% ) by 5,6-dichlororibofuranosylbenzimidazole (DRB) (Yang & Pardee, 1986). Cells became independent of translation, as determined using inhibitors of protein synthesis an hour later (but they are not independent of an isoleucine requirement until S phase starts) (Wynford-Thomas, La Montagne, Marin & Prescott, 1985). These results are generally consistent with transcriptional events .that are completed 2h before S, leading to translations completed an hour later.

D e novo synthesis of several enzymes such as thymidylate synthase and thymidine kinase occurs after R (Coppock & Pardee, 1985), and is apparently under the same control as onset of DNA synthesis.

Regulation of onset of DNA synthesis 177The important point here is that cells initiated DNA synthesis after a period of an

hour during which neither rapid transcription nor (rapid) translation were im­portant. This result suggests that morphological changes, possibly associated with rearrangements of the DNA synthetic apparatus, were occurring during this con­siderable time interval. As described above, enzymes for DNA synthesis both move from cytoplasm to nucleus and assemble into a readily sedimented high molecular weight complex at about this time.

It should be noted that the interval following R is not subject to physiological regulation of cell proliferation. The extracellular factors have lost their ability to stimulate or block progression of cells into the cycle by the time the cells have reached the/? point, but protein synthesis inhibitors such as cycloheximide are useful for further fractionation of this last part of the pre-S phase.

Das (1981) has reported that cells held at the G\/S interface with hydroxyurea lose ability to make DNA when released, with a half-life of about 5h. The loss is kinetically consistent with a prereplicative state that is lost as a unit. In general, cells that are arrested in their cycle are in a dynamic state. They do not long remain so, unlike an electric clock that has been detached from its power source. Molecules are in dynamic states of varying stabilities. So, after a stimulus necessary for progression in the cycle is removed some molecules are degraded, and time is later required for their resynthesis. Other synthetic processes go on at the same time, resulting in a different balance of cell components. This instability of certain cell components makes interpretations of position in the cycle on the basis of interrupted timing of events quite complex and difficult to interpret.

REGULATION BEFORE THE RESTRICTION POINT

A major process for control of cell proliferation is located in G\ before the restriction point. During this interval of several hours somatomedin C is the only factor required by 3T3 cells (Leof, Van Wyck, O’Keefe & Pledger, 1984; Campisi & Pardee, 1984). Somatomedin C like insulin autophosphorylates a tyrosine in its cell membrane receptor. Little is known regarding the biochemical events triggered by somatomedin C beyond the relevance of rates of protein synthesis and degradation.

The ras oncogene is activated during this same interval, as shown by increases in its mRNA (Campisi et al. 1984). The relationship between growth factors and this oncogene remains to be unravelled. The timing of ras activation is very different from the timing of activation of fos and ras oncogenes, which occurs shortly after cells in quiescence are stimulated, and thus half a dozen hours before activation of ras. Transcriptional events are required during this earlier part of G\, up to the R point (Yang & Pardee, 1986). A few mRNAs are made specifically during this same interval (Linzer & Nathans, 1983). Since mRNAs for enzymes such as thymidine kinase appear at the end of G\ phase (Stuart, Ito, Stewart & Conrad, 1985), some of these messages possibly correspond to proteins required for DNA synthesis.

Not only is transcription required for transit through the early part of G\ phase, but rapid translation is also important. Concentrations of cycloheximide that in­completely block total protein synthesis nearly completely block transit specifically

178 A. B. Pardee, D. L. Coppock and H. C. Yangthrough this part of the cell cycle. We have proposed that this is because a particular protein is required, i.e. one that is quite labile and hence must be synthesized rapidly in order to increase in amount in the face of its continued degradation (Rossow, Riddle & Pardee, 1979; Campisi, Medrano, Morreo & Pardee, 1982). Remarkably, this requirement for rapid protein synthesis was absent in a variety of cells transformed either with carcinogens or with RNA viruses (Campisi et al. 1982). These results suggest that stabilization of such a protein has an important role in the loss of growth control in cancer cells.

Having postulated an important regulatory protein that must be made in G\, that is unstable in normal cells, and that is stabilized in tumour cells, we initiated a search for it using two-dimensional electrophoresis gels. One such protein was identified, of Mr 68 000 and isoelectric point 6-3 (Croy & Pardee, 1983). This protein is made in larger amounts by tumour cells, and appears to be more stable. Its functional role is unknown, but it is being studied further with regard to its genetics, location and activity.

Not only does the onset of DNA synthesis depend differently upon rapid protein synthesis in normal versus transformed cells, presumably of an unstable protein such as p68, but also the activities of enzymes such as thymidine kinase, which appear at this time, are similarly controlled (Stuart et al. 1985; Coppock & Pardee, 1985). A regulatory mechanism that is dependent upon this labile protein appears to trigger a set of events including the appearance of new enzymes as well as the onset of DNA synthesis.

EVENTS LEADING QUIESCENT CELLS INTO THE CYCLE

Up to this point, we have discussed events culminating in the onset of DNA synthesis and initiated after M phase, by cycling cells. Quiescent cells are in a state (Go) that is different from the G\ state of cycling cells, even though the latter have the same unduplicated DNA content as do Go cells. Extra biochemical events are required for cells to emerge from quiescence, and these require extra time. G0 cells essentially need, as a first step, to reactivate their machinery for making proteins, including stimulation of messenger RNA enzymology, and this requires a period of several hours. Cycling Gi cells already have active protein and mRNA synthetic machinery after they emerge from mitosis, and so they prepare more quickly for events that lead to synthesis of DNA. To activate these quiescent cells and bring them back into the cycle occupies half a dozen hours, during which the synthesis of RNA and protein can increase considerably, and higher activities of various systems for transporting metabolites into cells can appear (see Baserga, 1984).

Activation of quiescent cells is another highly controlled event, dependent upon different factors such as platelet-derived and epidermal growth factors, and possibly insulin. Furthermore, a number of genes are activated, as shown by the appearance of new mRNAs, including those coded by oncogenes such as myc and fos. It is important here to note that the whole cast of characters for activation of quiescent cells is quite different from the corresponding factors, oncogenes and proteins such

Regulation of onset of DNA synthesis 179as p53 (see Baserga, 1985), involved in the preparation of cells to begin DNA synthesis. The events required for this activation have recently been reviewed (Stiles, 1983) and will not be discussed here. There now appear to be two critical regulatory processes required to bring quiescent cells into full proliferation: (1) those involved in initial activation; (2) those involved in the later onset of DNA synthesis. The relative importance of these two processes for cancer in its various stages remains to be elucidated.

In conclusion, clearly a great deal has been learned during the past dozen years about the relation of the cell cycle and its control to cell proliferation. Discoveries of growth factors, their receptors, oncogenes, new messages and new proteins, to say nothing of enzymes such as protein kinases have led to a much deeper understanding of the proliferation of normal cells and the derangement of cells that have become neoplastic.

This work was aided by a USPHS grant no. GM24571. The authors are indebted to Marjorie Rider for preparing the manuscript.

REFERENCESB a s e r g a , R. (1984). Growth in size and cell DNA replication. Expl Cell Res. 151, 1-5.B a s e r g a , R. (1985). The Biology o f Cell Reproduction. Cambridge, MA: Harvard University Press.Cam pisi, J ., G r a y , H. E., P a r d e e , A. B., D e a n , M. & S o n e n s h e in , G . E. (1984). Cell-cycle

control of c-myc but not c-ras expression is lost following chemical transformation. Cell 36, 241-247.

Cam pisi, J ., M e d r a n o , E. E., M o r r e o , G. & P a r d e e , A. B . (1982). Restriction point control of cell growth by a labile protein: Evidence for increased stability in transformed cells. Proc. natn. Acad. Sci. U.SA. 79, 436-440.

Cam pisi, J . & P a r d e e A. B. (1984). P ost-transcrip tional control of the onset of DNA synthesis by an insulin-like grow th facto r. Molec. cell. Biol. 4, 1807-1814.

C a s t e l l o t , J. J . J r , M i l l e r , M . R. & P a r d e e , A. B. (1978). Animal cells reversibly permeable to small molecules. Proc. natn. Acad. Sci. U.SA. 75, 351-355.

C h ib a , P., B a c o n , P. E. & C o ry , J. G. (1984). Stu dies directed tow ard testin g the “C hanneling” hypothesis - ribonu cleotides -> DNA in leukem ia L1210 cells. Biochem. biophys. Res. Commun. 123, 656-662.

C oppock, D. L. & P a r d e e , A . B. (1985). Regulation of thymidine kinase activity in the cell cycle by a labile protein, jf. cell. Physiol. 124, 269-274.

C r o y , R. G. & P a r d e e , A. B. (1983). Enhan ced synthesis and stabilization of Mt 68,000 protein in transform ed BALB/c-3T3 ce lls : C andidate for restriction point control of cell g row th. Proc. natn. Acad. Sci. U.SA. 80, 4699-4703.

D a s , M. (1981). Initiation of nuclear DNA replication: evidence for formation of committed prereplicative cellular state. Proc. natn. Acad. Sci. U.SA. 78, 5677-5681.

F e l d h e r r , C. M. (1 9 8 5 ). The uptake and accumulation of proteins by the cell nucleus. BioEssays 3 , 5 2 - 5 5 .

K o r n b e r g , A. (1980). DNA Replication. San Francisco: W. H. Freeman.L e o f , E. G., V a n W yk, J . J ., O ’K e e fe , E. J. & P le d g e r , W. J. (1984). Epidermal growth factor

(EGF) is required only during the traverse of early G\ in PDGF stimulated density-arrested Balb/c-3T3 cells. Expl Cell Res. 14, 202-208.

L in z e r , D. I. H. & N a th a n s , D. (1983). Growth-related changes in specific mRNAs of cultured cells. Proc. natn. Acad. Sci. U.SA. 80, 4271-4275.

N ic a n d e r , B . & R e ic h a r d , P. (1983). Dynamics of pyrimidine deoxynucleoside triphosphate pools in relationship to D N A synthesis in 3T6 mouse fibroblasts. Proc. natn. Acad. Sci. U.SA. 80, 1347-1351.

180 A. B. Pardee, D. L. Coppock and H. C. YangN ic a n d e r , B. & R eic h a r d , P . (1985). R elations betw een synthesis of d eoxyribonucleotides and

D N A rep lication in 3T6 fibroblasts. J . biol. Chem. 2 6 0 , 5376-5381.N o g u c h i, H., R e d d y , G. P . V. & P a r d e e , A. B . (1983). Rapid incorporation of label from

ribonucleoside diphosphates into DNA by a cell-free high molecular weight fraction from animal cell nuclei. Cell 3 2 , 443-451.

Reddy, G. P. V ., K lin ge , E . M. & Pardee, A. B. (1986). Ribonucleotides are channeled into a mixed D N A -RN A polymer by permeabilized hamster cells. Biochem. biophys. Res. Commun. (in press).

R e d d y , G. P . V. & P a r d e e , A. B. (1980). Multienzyme complex for metabolic channeling in mammalian DNA replication. Proc. natn. Acad. Sci. U.SA. 77, 3312-3316.

R e d d y , G. P . V. & P a r d e e , A. B. (1982). Coupled ribonucleoside diphosphate reduction, channeling. & incorporation into DNA of mammalian cells. J . biol. Chem. 2 5 7 , 12526-12531.

R e d d y , G. P . V. & PARDEE, A. B. (1983). Inhibitor evidence for allosteric interaction in the replitase multienzyme complex. Nature, Lond. 3 0 4 , 86-88.

Rossow, P . W., R id d l e , V. G. H. & P a r d e e , A. B. (1979). Synthesis of labile, serum-dependent protein in early G\ controls animal cell growth. Proc. natn. Acad. Sci. U.SA. 7 6 , 4446-4450.

S a r ib a n , E., Wu, R . S . , E r ic k so n , L. C. & B o n n e r , W. C. (1985). Inter-relationships of protein and DNA syntheses during replication of mammalian cells. Molec. cell. Biol. 5 , 1279-1286.

S py r o u , G. & R e ic h a r d , P. (1983). R ibon ucleotides are not channeled in to DNA in p er­m eabilized m am m alian cells. Biochem. biophys. Res. Commun. 115 , 1022-1026.

S t il e s , C. D. (1983). The molecular biology of platelet-derived growth factor. Cell 33, 653-655.S t u a r t , P., I t o , M., S t e w a r t , C . & C o n r a d , S . E. (1985). Induction of cellular thymidine

kinase occurs at the mRNA level. Molec. cell. Biol. 5, 1490-1497.S z y f , M ., K a p l a n , F., M a n n , V., G il o h , H., K e d a r , E. & R a z in , A. (1985). Cell cycle-

dependent regulation of eukaryotic DNA methylase level. J . biol. Chem. 2 6 0 , 8653-8656.W y n fo r d - T h o m a s , D., L a M o n ta g n e , A ., M a r in , G. & P r e sc o t t , D. M . (1985). Location of

the isoleucine arrest point in CHO and 3T3 cells. Expl Cell Res. 158 , 525—532.Y a n g , H. C. & Pa r d e e , A. B. (1986). Insulin-like growth factor I regulation of transcription and

DNA replicating enzyme induction necessary for S phase entry. J . cell. Physiol, (in press).