Eur. J. Biochem. 139, 627-635 (1984) 0 FEBS 1984

Inhibition of cell proliferation by interferons 2. Changes in processing and stability of newly synthesized DNA in human lymphoblastoid (Daudi) cells

Graham MOORE, Dirk R. GEWERT, and Michael J. CLEMENS Cancer Research Campaign Mammalian Protein Synthesis and Interferon Research Group, Department of Biochemistry, St George’s Hospital Medical School, London

(Received August 30/December 6, 1983) - EJB 83 0948

The inhibition of proliferation of Daudi cells in culture by human interferons is characterized by a change in the kinetics of labelling of different size classes of newly synthesized DNA. Initially, labelled precursors are incorporated exclusively into small DNA (Okazaki fragments) in both control and interferon-treated cells, as revealed by alkaline sucrose gradient centrifugation. In the interferon-treated cells, there is enhanced labelling of this small DNA after short periods of incorporation and slower conversion to larger DNA size classes, in comparison with the DNA of control cells. This effect is apparent after 12 h of interferon treatment, coincident with the onset of the inhibition of cell proliferation. It becomes progressively more marked up to 4 days, by which time cell growth has ceased completely.

Experiments using bromodeoxyuridine as a density label and analysis of radioactive DNA on caesium chloride/caesium sulphate gradients also reveal that some newly replicated DNA may be unstable and may turn over within a few hours of its synthesis. The label derived from DNA breakdown is efficiently reincorporated into newly synthesized molecules. It is suggested that interferon treatment inhibits DNA replication by activating DNA turnover rather than by directly inhibiting synthesis. This effect, together with the progressive retardation of conversion of Okazaki fragments to larger DNA, may lead to the eventual failure of cell proliferation.

Highly purified preparations of mammalian interferons have been shown to exert multiple biological effects on their target cells. In addition to the widely studied inhibition of viral replication, the anticellular actions of both naturally occurring [I - 31 and cloned interferon species [4] are of considerable interest. There is, however, very little information concerning the molecular mechanisms by which the replication of the cellular genome is regulated during growth inhibition by interferons. Some studies have suggested an inhibition of DNA synthesis, based on impaired incorporation of [3H]thymidine, but at least in some cases this reflects severely impaired thymidine uptake and thymidine kinase activity rather than a true effect on DNA replication [5 - 81. We are carrying out a detailed analysis of the changes in cellular functions which may be responsible for the antiproliferative effect of human in- terferons on Daudi cells in culture.

Daudi cells are human lymphoblastoid cells containing multiple copies of the Epstein-Barr virus genome. They are highly sensitive to the inhibitory effects of interferons on cell growth [7, 9- 111 and treatment with 100 units/ml of a-interferons causes a 40 - 50 % inhibition in the rate of cell proliferation after two days [8]. There is, however, no more than a 10 - 15 % inhibition of the overall rate of DNA synthesis per lo6 cells after this period of treatment, as measured by flooding the intracellular dTTP pool with excess labelled thymidine or by assessing incorporation of [3H]dCTP into DNA in permeabilized and lysed cell systems [I I]. There is also little inhibition by interferon treatment of the incorporation of alkali-stable radioactivity into DNA after pulse-labelling the cells with [3H]deoxyadenosine [8].

The clear inhibition of cell proliferation combined with the lack of change in the rate of DNA synthesis gives rise to a

paradox which can only be explained by one or more of three possibilities: namely, accumulation of DNA in cells, loss of newly replicated DNA from the cells (e.g. by cell death) or preferential degradation of a fraction of the newly replicated sequences. The first of these possibilities has been ruled out by analysis of the cell cycle distribution in Daudi cells by fluorescence-activated cell sorting. This shows very little change in the distribution of cells around the cycle and certainly no increase in DNA content after interferon treatment [lo, 1 I]. In this paper we consider the other two possibilities and present evidence which suggests that the stability of newly replicated DNA may be decreased during inhibition of cell proliferation in response to interferon treatment. We have also observed a change in the kinetics of synthesis of different size classes of DNA and the conversion of the initial products of replication (Okazaki fragments) to larger molecules. Our results indicate for the first time a possible mechanism by which cell prolife- ration may be regulated in this system. Preliminary com- munications of the data have been presented [12, 131.

MATERIALS AND METHODS

Materials

Preparations of purified human lymphoblastoid a-inter- ferons were as described previously [7, 8, 1 I] and were kindly provided by Drs K. H. Fantes and M. D. Johnston (Wellcome Research Laboratories, Beckenham, Kent, England). These preparations contained at least five different species of a-interferons at 3 - 7 x lo7 NIH international reference units/mg protein. Radiochemicals were from New England Nuclear or Amersham International. Dipyridamole, i.e. 2,6-

628

bis(diethano1amino) - 4,s -dipiperidino-pyrimido- [5,4- dlpyri- midine was supplied by Boehringer Ingelheim Ltd. Aphidicolin was a generous gift from ICI. Caesium chloride and caesium sulphate were purchased from Bethesda Research Laboratories and Sigma Chemical Co., respectively.

Cell culture and interferon treatment

Daudi cells were propagated at 37 "C in stationary suspen- sion culture as described previously [7, 8, 111. Cells were normally maintained at a density of 1-6 x 105/ml. For in- terferon treatment batches of exponentially growing cells at about 1.5 x 105/ml were split into two and the interferon preparation added to one of the flasks to give up to 100 NIH international reference units/ml, as indicated in the legends to figures and tables. Unless otherwise stated the cells were then incubated for 48 h in the presence or absence of the interferons. The efficacy of interferon treatment was monitored both by inhibition of cell proliferation and by the inhibition of [3H]thymidine incorporation into DNA observed when tracer doses of this precursor were used [7, 8, 111.

Incorporation of precursors into DNA and acid-soluble pools

DNA was labelled by incorporation of radioactivity from [3H]thymidine (20 - 30 Ci/mmol) or [3H]deoxyadenosine (1 5 - 40 Ci/mmol) into acid-insoluble material using tech- niques already published [7, 81. The details of precursor concentrations and incubation times are specified in the legends to individual figures and tables. When [3H]deoxyadenosine was the precursor an alkaline-digestion step was included in the procedure for sample preparation, to eliminate labelled RNA, as described previously [S].

Measurements of uptake and loss of radioactivity from intracellular acid-soluble pools were made by precipitating 300-pl aliquots of cell suspension (washed free of extracellular radioactivity) with an equal volume of 10 % trichloroacetic acid containing 0.5 % sodium pyrophosphate. Acid-insoluble material was removed by centrifugation in a microfuge and 500 pl of the supernatants were diluted to 2 ml with water and counted in Tritosol aqueous scintillation fluid [14]. The pre- cipitates from these experiments were dissolved in 0.6ml of 0.3 M NaOH and their radioactivity also determined by counting in Tritosol fluid.

Alkaline sucrose density gradient analysis of newly synthesised DNA

Samples (lml) of cell suspensions were incubated with labelled nucleosides for various times and then immediately lysed by addition of 0.4 M NaOH, 1 % N-dodecylsarcosine and 0.01 M EDTA (final concentrations) [IS]. 0.9-ml samples were layered directly on 38-ml 5 - 20 % alkaline sucrose gradients containing the above components plus 0.1 M NaCl. The gradients were then centrifuged in a Beckman SW27 rotor for 17 h at 4 "C at 27000 rpm to separate Okazaki fragments from large DNA. The gradients were collected from the bottom, 0.4mg of bovine serum albumin was added to each fraction, and the samples were then precipitated with trichloroacetic acid, washed and the radioactivity counted as described [7, 8, 1 I]. Approximate sedimentation coefficients of the various DNA size classes were obtained by calibration of the gradients with 3H-labelled DNA size markers, i.e. SV40 components I and I11 (53 S and 16 S) and phage I DNA (40 S), from Bethesda Research Laboratories Ltd (UK).

Extraction of DNA

Labelled DNA was extracted from cells as described by Schlaeger [16]. Cells were sedimented (600 x g, 20 min) and resuspended in 1 ml of distilled water. Lysis buffer was added to give 50mM Tris/HCl, pH 7.8, 10mM EDTA and 1 % (w/v) N-dodecyl-sarcosine, followed by pronase and proteinase K (1 mg/ml each, final concentration) in a final volume of 2 ml. After incubation at 37°C for 16h, another 3ml of Tris/ EDTAIN-dodecylsarcosine buffer was added and the mixture extracted by vigorous shaking with 10ml of chloroform/ isoamyl alcohol (24: 1, v/v). The aqueous layer recovered after centrifugation for 10 min at 10000 x g was retained, the organic phase was re-extracted with another 4ml of Tris/EDTA/ N-dodecylsarcosine buffer and the second aqueous layer combined with the first. The material was dialysed overnight against 0.1 x NaCl/Cit (NaCl/Cit = 0.15 M NaCl/O.O15 M tri- sodium citrate, pH 7.0) at 4°C and then precipitated with 2.5 vol. ethanol at - 20 "C for at least 24 h. The DNA was centrifuged (I0000 x g, 20 min), dried and resuspended in 5 ml of 0.1 x NaCl/Cit. For analysis on alkaline caesium gradients the DNA was denatured by adding NaOH to 0.1 M.

Buoyant density gradients

DNA was separated according to buoyant density in 5-ml alkaline caesium chloride/caesium sulphate gradients as de- scribed by Schlaeger [16]. Centrifugation was for 48 h at 233 000 x gin Beckman SW 50.1 or Sorvall AH-650 rotors. The gradients were fractionated by pumping from the bottom and the densities of selected fractions were measured to calibrate each gradient. DNA was precipitated from gradient fractions as described above.

RESULTS

Effects of interferon treatment on the size distribution of newly replicated DNA

The kinetics of labelling and the size characteristics of newly replicated DNA have been compared in control and interferon- treated Daudi cells. In order to avoid the complication of the effect of interferon treatment on thymidine uptake and dTTP specific radioactivity [7, 111, [3H]deoxyadenosine was used as the labelled precursor. Fig. 1 shows the distribution on alkaline sucrose gradients of radioactivity in DNA labelled with [3H]deoxyadenosine for up to 10 min. Centrifugation under these denaturing conditions separates rapidly labelled small DNA fragments ('Okazaki fragments') from larger size classes with sedimentation values greater than 40s [15, 17, 181. After very short labelling times, radioactivity is incorporated ex- clusively into the small fragments. Subsequently, label begins to appear in the larger DNA. Comparison of the DNA from control and 2-day interferon-treated cells shows that after 3 - 8 rnin a higher proportion of the total radioactivity remains in the form of small fragments in the DNA of interferon-treated cells. This appears to be a consequence of a delay in the conversion of small DNA to larger size classes which occurs by ligation [19-211. Summation of the radioactivity in the fractions at the top of each gradient shows that within 1 rnin after the addition of the radioactive precursor the small DNA fragments become maximally labelled in both interferon- treated and control cells (Fig.2A, B). However, there is a

629

40

30

50 1 3min F 10 min

50-53s 40s 16s

- + i + -

4oc A h I1

Fraction Number

501 6 min

30

Fraction Number

Fig. 1. Analysis of the distribution ofradioactivity in different size classes of newly synthesised DNA from control and interferon-treated cells. Control and 48-h interferon-treated (100 unitslml) cells, at 5.2 x lo5 and 3.4 x 10' cells/ml respectively, were labelled at 37°C with 20 pCi/ml of [3H]deoxyadenosine (40 Ci/mmol) for (A) 30 s, (B) 90 s, (C) 3min, (D) 6min, (E) 8min and (F) 10min. The cells were lysed immediately as described in Materials and Methods and the lysates were loaded onto 5 - 20 alkaline sucrose gradients. Centrifugation was carried out at 27000 rpm for 17h and the gradients were fractionated and radioactivity determined as described in Materials and Methods. Sedimentation was from right to left. (OA) Control cells; (A----A) interferon-treated cells. The gradients were calibrated with DNA size markers which sedimented to the positions indicated in F

25 - 30 % increase in the total radioactivity present as small fragments in interferon-treated cells relative to control cells (Fig. 2B). Using ['H]deoxyadenosine as precursor, at labelling times of 6min or more there is no difference between control and interferon-treated cells in the amount of radioactivity. incorporated into total DNA on alkaline gradients (data not shown), in confirmation of our previous results with this nucleoside [8]. Thus, although there is an initial delay in the appearance of radioactivity in large DNA in interferon-treated cells (Fig. 2C), the subsequent rate of labelling of this DNA is the same as in control cells.

The change in the early kinetics of labelling of different DNA size classes occurs at the onset of inhibition of cell growth in response to interferon treatment, as measured by the first

appearance or a difference in cell number. Cells which have been treated for as little as 12 - 18 h with interferons already exhibit a change in the distribution of radioactivity on alkaline sucrose gradients after a pulse of Smin with labelled de- oxyadenosine (Fig. 3A, B). The inhibition of the conversion of small fragments to larger DNA size classes then becomes progressively more marked with the length of interferon treatment so that by 4 days a large proportion of the radioactivity incorporated into DNA remains in the form of small fragments even at 10 min after the addition of labelled precursor (Fig. 3C). By 4 days of interferon treatment cell growth has ceased completely [l I] although clearly some DNA synthesis is still occurring.

We have attempted to analyse the kinetics of labelling of different sub-classes of the larger DNA by centrifuging the alkaline gradients for a shorter time (data not shown). Under these conditions, no reproducible differences were observed between control cells and interferon-treated cells in the kinetics of labelling of these sizes of DNA. It was clear, however, that the kinetics of labelling of various large DNA size classes are complex, as some Okazaki fragments appear to be converted rapidly to very large DNA molecules while others initially give rise to intermediate sizes of product. By 2 h of labelling in both control and interferon-treated cells most of the fragments were converted to DNA which sediments at greater than SOS on alkaline sucrose gradients.

It has been reported that misincorporation of dUMP into DNA in place of dTMP, caused by treatment of cells with deoxyuridine or addition of dUTP to isolated nuclei, results in subsequent excision of the uracil and cleavage of the DNA at the apyrimidinic sites [22 - 241. This produces an increase in the number of small fragments which may thus be mistaken for normal intermediates of DNA replication. However in Daudi cells treatment with 10pM deoxyuridine failed to cause any large change in the distribution of radioactivity between rapidly labelled DNA size classes (data not shown). Further- more, experiments with [3H]deoxyuridine indicate that this precursor is not incorporated significantly into DNA, other than via its prior conversion to dTMP by thymidylate syn- thetase, in either control or interferon-treated Daudi cells. These results suggest either that Daudi cells have a very efficient dUTPase which prevents any sizeable increase in the dUTP pool or that some other discriminating mechanism prevents incorporation of this precursor into DNA. It would seem unlikely therefore that the increase in the proportion of rapidly labelled small DNA fragments in interferon-treated cells is a result of an increase in the misincorporation of dUMP and its subsequent excision.

The relationship of rates of D N A synthesis to rates of cell proliferation

Interferon-treated Daudi cells exhibit a progressively slower rate of proliferation in culture over the first 3 days [l 11. Table 1 shows the increases in cell doubling time which develop as a function of (A) interferon concentration and (B) time of exposure to 100 U/ml of interferons. At 48 h, the time at which the majority of our measurements have been made, the doubling time of the cells treated with this concentration of interferons is increased from the normal value of about 25 h to 40 - 60 h. In an exponentially growing cell population, in which all cells are proliferating and there is minimal cell death, the rate of replicative DNA synthesis can be related to the cell doubling time as follows.

630

5 0 - B

- 5 40-

c 30-

._ U

0 z

0 2 4 6 8 100 2 4 6 6 10 0 2 4 6 6 1 0 Time (mini Time (mini

Fig. 2. Kinetics of synthesis and processing of Okazaki fragments into large DNA in control and interferon-treated cells. The radioactivity in fractions 1 - 6 (large DNA) (.A) and fractions 9- 12 (Okazaki fragments) (A-A) in the gradients shown in Fig. 1 was summed and plotted as a function of labelling time for (A) control cells and (B) interferon-treated cells. In (C) the percentage of newly synthesised DNA in the form of Okazaki fragments is shown for each time of labelling. (04) Control cells; (0-0) interferon-treated cells

18h A

1 1 1 1 I I I I I 1

5 min label

4 0 - c 9 6 h 10 min label

20

10- " I

4 8 1 2 Fraction number

Fig. 3. Analysis of the distribution of radioactivity in different size classes ofnewly synthesised DNA as a function of lengths of time of treatment with interferons. Exponentially growing control cells and cells trated with interferons (100 units/ml) for (A) 12 h, (B) 18 h and (C) 96 h were labelled with 10 pCi/ml of [3H]deoxyadenosine (40 Ci/mmol) for 5 min (A, B) or 10min (C). Cell lysis, gradient centrifugation and determi- nation of radioactivity were as described in Fig. 1. (0-) Control cells; (A----A) interferon-treated cells

Table 1. Cell doubling times and rates of DNA synthesis in control and interferon-treated Daudi cells Daudi cells were incubated in the presence or absence of the indicated concentrations of interferons for the times shown. Cell numbers were determined in a haemocytometer at daily intervals and doubling times were calculated from the slopes of plots of log (cell number) versus time. Note that the data in (A) and (B) are from separate experiments. In (C) the measured rates of DNA synthesis, derived from the rates of [3H]thymidine incorporation using a saturating dose of the precursor [12], are compared with the rates expected for cell doubling times of 25.5 h (control cells) and 40 - 60 h (interferon-treated cells), calculated as described in the text. A cellular DNA content of 1.2 pg/lOs cells and dTMP content of 28 % (by weight) have been used to convert fractional rates of DNA synthesis into pmol dTMP h-'/1OS cells

A. Interferon concn Cell doubling time at 48 h

units/ml 0 1

10 50

100

h 25 29 33 41 62

B. Time of treatment with Cell doubling time 100 units/ml interferons

h

24 48 12 96

120

21 41 56 no proliferation no proliferation

C. Interferon treatment Rate of DNA synthesis

measured expected

pmol dTMP h-'/1OS cells

None 31.7 27.8 100 units/ml, 48 h 44.8 12.2-11.4

631

Assuming a constant fractional rate of DNA synthesis, k,, and no degradation of cellular DNA, the rate of accumulation of DNA mass, m, is given by:

dm - = k , . m dt

Thus the ratio of DNA mass at time t , m,, to DNA mass at time 0, mo, is:

After one cell doubling time ( t d )

Thus,

or, In 2

k, = -. td

(3)

(4)

A knowledge of the cell doubling time will therefore give an expected value for k,, which may be compared with the measured rate of DNA synthesis. Any value for the latter in excess of the expected k, will imply a failure of one of the above assumptions, i.e. no cell death or no DNA degradation or loss.

When the rates of DNA synthesis in control and interferon- treated Daudi cells are measured using a saturating con- centration of t3H]thymidine as precursor [I 11, the rate for control cells is in reasonable agreement with that calculated from the doubling time using Eqn ( 5 ) (Table 1). Thus the majority of DNA synthesised remains stable in these cells, as expected. However when the measured and expected rates of DNA synthesis are compared for interferon-treated cells a large discrepancy (2.6 - 3.7-fold) is obtained (Table l), since the measured rate of DNA synthesis is close to that of the control cells whereas the doubling time is much longer. This discrep- ancy can only be explained if (a) DNA is accumulating faster than the cells are dividing; (b) a proportion of newly replicated DNA in interferon-treated cells is unstable; or (c) some DNA is lost from the cells within a few hours of its synthesis. We have shown, by fluorescence-activated cell sorting and by direct analysis, that DNA does not accumulate in interferon-treated Daudi cells [ll]. Furthermore, cell viability and integrity are maintained at 48 h of interferon treatment [ll]. We therefore sought evidence for the turnover or loss of newly synthesized DNA using a variety of techniques.

DNA turnover in interferon-treated cells

In agreement with our previous observations [8, 111, the initial rate of incorporation of thymidine into DNA in control and interferon-treated cells is identical when the precursor is supplied at a concentration of 100 pM (Fig. 4). In our previous work incorporation was only measured over a labelling time of up to 1 h. However, as shown in Fig. 4, with prolonged times of incubation a divergence in rates of incorporation of label becomes apparent between the two populations of cells, as would be expected if a proportion of the newly synthesised DNA were being turned over or lost from the interferon-treated cells. This delayed decrease in incorporation of label in the interferon-treated cells is not a consequence of a slowly developing inhibition of the cells by 100 pM thymidine since a similar time course is observed when cells are preincubated with the unlabelled precursor for 6 h before label is added (cf. Fig. 4A and B). We can find no evidence for the loss of any

Fig. 4. Incorporation of r3H]thymidine into DNA over prolonged time courses. (A) Control cells (initial density 4.2 x 10s/ml) and 36-h interferon-treated cells (100 units/ml) (initial density 2.4 x 1OS/ml) were incubated at 37°C with a saturating concentration of [3H]thymidine (100 pM, 4pCi/ml, see [Ill). At the times indicated the cell densities were determined in a haemocytometer and duplicate 0.3-ml aliquots of cell suspensions were acid-precipitated for determi- nation of radioactivity in DNA (Materials and Methods). (B) The same procedure was adopted except that the cells were preincubated for 6 h with 100 pM unlabelled thymidine before addition of the [jHIthymidine. (-) Control cells; (0-) interferon- treated cells

Time (h)

I 2 3 4 Time (h)

Fig. 5. Kinetics of turnover of the intracellular thymidine nucleotidepool and stability of label in newly synthesised DNA in control and interferon- treated cells. Control cells and cells treated for 48 h with 100 units/ml interferon were centrifuged (600 x g, 20 min) and resuspended in two aliquots of 2ml of fresh medium each at 2 x 106/ml (to conserve labelled thymidine). [3H]Thymidine was then added to the prewarmed cell suspensions to give 20 pCi/ml and incubations were continued for 1 min. The suspensions were immediately diluted with 40ml each of ice-cold medium containing 5 pg/ml dipyridamole (to block further thymidine uptake) with or without 1 p g / d aphidicolin (to inhibit further DNA synthesis) and centrifuged as above. The cells were washed once with a further 40 ml of the same medium by resuspension and centrifugation and were finally resuspended at 4 x 105 /d in 10 ml of cold dipyridamole-containing medium, with (----) or without (-) aphidicolin. The suspensions were warmed to 37°C and duplicate 0.3 ml aliquots precipitated with 0.3 ml of cold 10 % (w/v) trichloroacetic acid/0.5 % (w/v) sodium pyrophosphate at the times indicated. Radioactivity in the acid-soluble nucleotide pools (0) and in acid-insoluble DNA (0) was determined as described in Materials and Methods. (A) Control cells; (B) interferon-treated cells

632

1.6 t \ 30 ?

1 1 A Control II

:; 20-A ‘ I

10 -

f

0 4& , I J

30 B I F N

A- - -A pulse -chase

5 10 15 20

20

10

0 5 10 15 20

20 I I‘ / : J .

- 5 10 15 20

Fraction Number Fraction Number

Fig. 6. Buoyant density gradient analysis of labelled DNA after pulse-chases with bromodeoxyuridine. Two separate experiments are shown which are representative of several performed with different batches of cells. Cells were incubated without (A and C) or with interferons (lOOunits/ml) for 1 day (B) or 2 days (D). DNA was labelled by incubation of 9-ml aliquots for 30 min with [3H]thymidine (3 pCi/ml for control cells, 18 pCi/ml for interferon-treated cells). The cells were then centrifuged (600 x g, 20 min) and either lysed for extraction of pulse-labelled DNA (Materials and Methods), or resuspended in 9 ml of fresh medium (without radioactivity) containing 1 pM unlabelled bromodeoxyuridine and incubated at 37 “C for a further 3 h. The cells were then centrifuged and the DNA extracted as above. The purified DNA was denatured in 0.1 M NaOH and subjected to alkaline CsCl/Cs,SO, buoyant density gradient analysis as described in Materials and Methods. The gradients were calibrated by measuring the densities of selected fractions ( 0 4 ) . The specific radioactivities (cpm/A,,, unit) of the total DNA extracted from each batch of cells were: (A) pulse label, 12.4 x lo’; pulse-chase, 11.5 x lo5; (B) pulse label, 5.96 x lo5; pulse-chase, 6.09 x lo5; (C) pulse label, 15.6 x lo’; pulse- chase, 16.2 x lo’; (D) pulse label, 5.86 x 10’; pulse-chase, 5.06 x lo5. (0-0) 30-min pulse labelled DNA; (A----A) DNA after 3-h chase with bromodeoxyuridine

newly synthesised DNA from 48 h interferon-treated cells. When DNA was labelled for 30 min with [3H]thymidine and the cells then incubated for a further 6 h no loss of acid-insoluble radioactivity from the cells into the medium occurred (data not shown).

In an attempt to demonstrate DNA turnover during pulse- chase experiments with whole cells we employed the inhibitor of thymidine uptake dipyridamole [25], which rapidly blocks the membrane transport of the [3H]thymidine and only permits a few minutes of further incorporation of radioactivity into DNA. This incorporation results from the utilization of intracellular nucleotides labelled with thymidine already taken up by the cell. In the presence of 5pg/ml dipyridamole, radioactivity in the intracellular dTTP pool is sufficient for only 5-l0min of further DNA labelling in both control and interferon-treated cells (Fig. 5). This is observed when detailed time courses of both the loss of acid-soluble radioactivity and the increase of acid-insoluble radioactivity are examined and indicates a similar rate of turnover of the nucleotide pool in both cell populations. Loss of thymidine nucleotides from the

intracellular pool is largely prevented by the inhibitor of DNA replication aphidicolin (Fig. 5). In spite of the effectiveness of dipyridamole as an inhibitor of thymidine uptake, however, such an experiment provides no evidence for instability of newly labelled DNA in interferon-treated cells since no de- crease of acid-insoluble radioactivity in DNA labelled for only a few minutes is observed between 30 min and 4 h after the addition of dipyridamole (Fig. 5). A similar conclusion was reached in experiments where DNA was pulse labelled for longer periods (up to 1 h) and the cells then subjected to a further incubation in the presence of dipyridamole for up to 24 h (data not shown).

The failure to observe a loss of radioactivity from pre- labelled DNA in the presence of dipyridamole could be due to very efficient salvage pathway enzyme activity in Daudi cells with the consequent reincorporation of label derived from DNA breakdown back into newly synthesised DNA. In order to test this possibility, we employed bromodeoxyuridine den- sity labelling and caesium sulphate/caesium chloride buoyant density gradient analysis as a means of distinguishing between

633

- 3H deoxyadenosine A- -A 3H deoxyadenosine

30 and Budr

20

10

5 10 15 20

Fraction Number Fig. I . Effect of bromodeoxyuridine on buoyant density of newly synthesised DNA in control and interferon-treated cells. Control cells (A) (6.5 x 105/ml) and 2-day interferon-treated (100 units/ml) cells (B) (3.4 x lo5/&) were labelled for 90min with [3H]deoxyadenosine (IOpCi/ml) in the presence (A----A) or absence (0-) of unlabelled bromodeoxyuridine (1 pM). The incubation volume was 5ml. DNA was extracted and analysed on alkaline CsCl/Cs,SO, buoyant density gradients as described in Fig. 6

old and new DNA molecules. Cells were labelled with [3H]thymidine for 30min, then washed and a portion of them further incubated with 1 pM bromodeoxyuridine for 3 h. As illustrated in Fig. 6, when this approach was used DNA labelled with [jHIthymidine in control cells remained stable and of low density during the bromodeoxyuridine chase. In marked contrast, a substantial proportion of the DNA labelled during the 30-min pulse in 1-day or 2-day interferon-treated cells was lost from the low-density peak and the label appeared in high- density DNA synthesised during the bromodeoxyuridine chase. There was no significant change in the total amount of label in DNA during this chase. This evidence strongly suggests that some newly synthesised DNA in interferon-treated cells does indeed turn over and that there is a very efficient reincorporation of the label back into DNA. As a control, we have shown that there is no difference between control and interferon-treated cells in the extent of incorporation of bromodeoxyuridine into heavy DNA. Fig. 7 indicates that when label and bromodeoxyuridine were administered simul- taneously to the cells all of the incorporated radioactivity was located at a buoyant density of 1.75 g/ml in both cases. Normal DNA, not containing bromodeoxyuridine, has a density of 1.68 g/ml on these gradients. This experiment also confirms that both control and interferon-treated cells incorporate radioactivity into DNA predominantly by replicative rather than repair synthesis since the latter process would give rise to some low-density labelled DNA even when bromodeoxy- uridine is present. An additional observation, the significance of which is not yet certain, can also be made from these

experiments. As both Fig. 6 and 7 illustrate, a small fraction of pulse-labelled DNA in interferon-treated cells centrifuges at a higher equilibrium buoyant density than the bulk DNA even in the absence of any bromodeoxyuridine. The nature of this DNA, which presumably has a higher dGtdC content than bulk DNA, is currently being studied in our laboratory.

DISCUSSION

Studies of the mechanism of DNA replication in eukaryotic cells and their viruses have shown that synthesis on the retrograde arm of the DNA template involves initiation by polymerization of RNA primers, elongation of Okazaki frag- ments by DNA polymerase-a and excision of the RNA primers in a two-step process (reviewed in [26]). The Okazaki fragments are then completed by a modified DNA polymerase in a step requiring one or more cytosol protein factors [20, 26-28] before eventual ligation to the 5' ends of the adjacent DNA sequences. The mechanism of synthesis on the forward arm of the DNA template is less well understood since some studies suggest that nascent DNA strand synthesis is continuous [17, 29 - 321, whereas other reports indicate discontinuous syn- thesis on this arm also [15, 19-21]. Our results suggest that Daudi cells use a discontinuous mode of synthesis at replication forks on both strands of the template. Although we have not attempted to determine whether all the rapidly labelled small DNA molecules are true Okazaki fragments (i.e. sequences primed with short lengths of RNA) their size and kinetics of labelling suggest very strongly that they are true intermediates of DNA replication.

The analysis of newly synthesised DNA from interferon- treated cells shows a delay in the processing of the rapidly labelled small fragments to give larger DNA at early times after pulse labelling, but no extensive accumulation of such fragments at later times. When [3H]deoxyadenosine is the precursor, the subsequent rate of labelling of the larger DNA size classes on alkaline gradients is identical in control and interferon-treated cells. These results suggest that neither initiation nor elongation of nascent chains can be sufficiently slowed to limit the overall rate of DNA replication. Impairment of ligation of nascent DNA chains, in the absence of reduced rates of initiation or elongation, might be expected to result in continuous accumulation of DNA in small size classes, but only if all the newly replicated DNA is stable. The failure of DNA intermediates to accumulate continuously in interferon-treated cells may therefore be a consequence of their rapid turnover. Indeed this DNA could be vulnerable to nuclease attack because it is less rapidly or less efficiently ligated. However, the kinetics of overall DNA turnover are too slow and the extent of turnover too great to be accounted for by degradation of Okazaki fragments alone (Fig. 6) and there must also be loss of larger DNA replication intermediates. These might also be subject to defective ligation, although the resolving power of alkaline sucrose gradients is inadequate to characterize such an inhibition in interferon-treated cells and other techniques are required to answer this question.

Inhibition of ligation of Okazaki fragments occurs during DNA replication in cells treated with inhibitors of protein synthesis [29, 30, 331 or with agents such as hydroxyurea and aphidicolin [34, 351. Thus fragment joining is particularly sensitive to inhibition of translation, changes in availability of deoxynucleotides or impaired activity of DNA polymerase-a. Ligation of nascent DNA is also influenced by the availability of polyamines [36], by agents which induce DNA damage [37]

634

and by the activity of poly(ADP-ribose) polymerase [38,39]. In our system interferons could affect any one of these mech- anisms to cause transient accumulation of Okazaki fragments.

Our conclusion that some of the newly replicated DNA in interferotl-treated cells is unstable and is degraded within a few hours of its synthesis is supported by the following obser- vations. (a) The rate of DNA synthesis is too fast (by a factor of 2-4-fold) for the rate at which DNA is accumulating in cells treated with interferons for 48 h (Table 1). (b) DNA synthesis, assayed by three different methods, continues almost un- impaired in cells which show a substantial inhibition of proliferation [8, 111. (c) The rates of turnover of the in- tracellular dTTP pool are similar in control and interferon- treated cells (Fig. 5). (d) The initial rate of DNA labelling in interferon-treated cells declines markedly during prolonged incubation with [3H]thymidine (Fig. 4). (e) A proportion of the radioactivity which is incorporated into the DNA of interferon- treated cells during a 30-min pulse is subsequently found in density labelled DNA synthesized during a 3-h chase in the presence of bromodeoxyuridine; this is not observed in control cells (Fig. 6). The present results do not allow us to calculate an exact half-life of the DNA which is degraded. The latter will necessitate a more detailed study of the kinetics of the turnover using the approach described in Fig. 6. In the absence of such information it is, nevertheless, possible to calculate a mean turnover rate for total DNA in interferon-treated Daudi cells (applicable if all DNA is equally susceptible to degradation). The discrepancy between rates of synthesis and accumulation of DNA is of the order of 27-33pmol dT h-’/105 cells (Tablel). Taking figures of 28% (by weight) for the dTMP content [40] and 1.20pg/105 cells for the amount of DNA in Daudi cells [ll], this rate of degradation is equal to a fractional rate (kd) of 0.026-0.032yg h-l/yg DNA, or a half-life of 22 - 27 h [(ln 2)/k,]. If only a sub-fraction of the total DNA is turning over, as would be the case if newly replicated DNA but not parental strands were subject to degradation, then the half- life of this population of molecules would be much shorter. This possibility is a likely one since the data provided in Fig.6 indicate that up to half of the DNA labelled in a 30-min pulse can turn over in the subsequent 3 h.

The present data do not distinguish between a possible interferon-activated DNase activity in Daudi cells and a change in the susceptibility of the DNA to degradative enzyme(s). As discussed above, the latter change could occur if there were altered processing of newly replicated DNA or a modification of its assembly into chromatin. There is evidence that newly replicated DNA is more susceptible than older DN,A to nuclease degradation [41, 421 (and M. J. C., unpublished observations), perhaps because of the time required for its assembly into nucleosomes [43,44]. The sensitivity of this DNA to nuclease digestion depends on the availability of newly synthesized histones and other proteins for chromatin assembly [33,44]. It is possible that the effects of interferon treatment on Okazaki fragment accumulation and DNA stability may be related, possibly through a loss of essential proteins. Such effects could account for the failure of certain cell types to complete S phase within the normal time after exposure to interferons [45 - 471.

The results presented in this paper make it clear that interferon treatment can affect a number of functions as- sociated with the cell nucleus. There is relatively little infor- mation in the literature concerning this aspect of interferon action. It is known that interferon treatment can prevent or reverse cellular growth transformation by DNA tumour viruses such as SV40 [48], bovine papilloma virus (BPV) [49] or

Epstein-Barr virus [50] and can inhibit the biochemical trans- formation of cells by exogenous cloned DNA containing selectable genes (e.g. thymidine kinase) [51, 521. Both types of transformation probably require stable integration or re- arrangement of DNA sequences in the cellular genome, events which may share some steps in common with DNA replication. It has not been established whether increased degradation or inefficient integration of the foreign DNAs are responsible for the interferon effects although in the case of BPV-transformed cells a loss of viral sequences in response to interferon treatment has been reported [49]. A further intriguing finding is the presence of a novel species of interferon-induced (2’-S’)oligo(A) synthetase and of several (2’-S’)oligo(A) binding proteins in the nuclei of Ehrlich ascites tumour cells [53]. There is no evidence that the classical (2’-S’)oligo(A)-ribonuclease mechanism is responsible for the growth inhibition of Daudi cells. Nevertheless, the possibility of other functions for nuclear (2’-5’)oligo(A) or of a relationship between a nuclear DNase and the cytoplasmic RNase involved in the antiviral effects of interferons remains a fascinating one.

Relatively little attention has been paid in the literature to possible regulatory roles of endogenous DNases in eukaryotic cells. In the light of our conclusion that DNA degradation to acid-soluble products is activated in interferon-treated Daudi cells this potential mechanism for controlling cell proliferation deserves further investigation.

M. J. C. holds a Career Development Award from the Cancer Research Campaign and this work is supported by a grant from this source. G. M. and D. R. G. thank the Science and Engineering Research Council and the Medical Research Council respectively, for research studentships. We thank Vivienne Tilleray for skilled technical assistance, Barbara Bashford for the drawings and Vivienne Marvel1 for preparation of the manuscript.

REFERENCES

1. Knight, E. (1976) Nature (Lond.) 262, 302-303. 2. Gresser, I., De Maeyer-Guignard, J., Tovey, M. G. & De Maeyer,

3. Taylor-Papadimitriou, J. (1980) in Interferon 2 (Gresser, I., ed.)

4. Evinger, M., Maeda, S. & Pestka, S . (1981) J. Biol. Chem. 256,

5 . Brouty-Boye, D. & Tovey, M. G. (1978) Zntervirology 9,243 - 252. 6. Fuse, A. & Kuwata, T. (1978) J. Nut1 Cancer Znst. 60,1227- 1232. 7. Gewert, D. R., Shah, S. & Clemens, M. J. (1981) Eur. J . Biochem.

8. Gewert, D. R., Moore, G. & Clemens, M. J. (1983) Biochem. J .

9. Hilfenhaus, J., Damm, H. & Johanssen, R. (1977) Arch. Virol. 54,

10. Silverman, R. H., Watling, D., Balkwill, F. R., Trowsdale, J. &

11. Gewert, D. R., Moore, G., Tilleray, V. J. & Clemens, M. J. (1984)

12. Moore, G., Gewert, D. R. & Clemens, M. J. (1982) Biochem. SOC.

13. Moore, G. & Clemens, M. J. (1983) Biochem. SOC. Trans. 11,361 -

14. Fricke, U. (1975) Anal. Biochem. 63, 555-558. 15. Krokan, H., Cooke, L. & Prydz, H. (1975) Biochemistry 14,

16. Schlaeger, E.-J. (1978) Biochem. Biophys. Res. Commun. 81,s - 18. 17. Friedman, C. A,, Kohn, K. W. & Erickson, L. C. (1975)

18. Berry, D. E. & Collins, J. M. (1981) Biochim. Biophys. Acta 653,

E. (1979) Proc. Nail Acad. Sci. USA 76, 5308-5312.

pp. 13- 46, Academic Press, London.

21 13 - 21 14.

116, 487-492.

214, 983 - 990.

271 - 277.

Kerr, I. M. (1982) Eur. J. Biochem. 126, 333-341.

Eur. J . Biochem. 139, 619-625.

Trans. 10, 95.

362.

4233 - 4237.

Biochemistry 14, 4018- 4023.

160-168.

635

19. Gautschi, J. R. & Clarkson, J. M. (1975) Eur. J . Biochem. 50,

20. Tseng, B. Y. & Goulian, M. (1975) J. Mol. Biol. 99, 317-337. 21. Krokan, H., Wist, E. & Prydz, H. (1977) Biochim. Biophys. Acta

22. Graftstrom, R. H., Tseng, B. Y. & Goulian, M. (1978) Cell 15,

23. Brynolf, K., Eliasson, R. & Reichard, P. (1978) Cell 13,573 - 580. 24. Sedwick, W. D., Kutler, M. & Brown, 0. E. (1981) Proc. Natl

25. Plagemann, P. G. W. & Richey, D. P: (1974) Biochim. Biophys.

26. DePamphilis, M. L. & Wassarman, P. M. (1980) Annu. Rev.

27. Fraser, J. M. K. & Huberman, J. A. (1978) Biochim. Biophys. Acta

28. Nagata, K., Enomoto, T. & Yamada, M.-A. (1981) Biochim.

29. Werner, D. & Maier, G. (1975) Eur. J . Biochem. 54, 351 - 358. 30. Venkatesan, N. (1977) Biochim. Biophys. Acta 478, 437-453. 31. Miller, M. R., Castellot, J. J. & Pardee, A. B. (1978) Biochemistry

32. Funderud, S., Andreassen, R. & Haugli, F. (1978) Cell 15,1519-

33. Seale, R. L. & Simpson, R. T. (1975) J. Mol. Biol. 94,479-501. 34. Radford, I. R., Martin, R. F. & Finch, L. R. (1982) Biochim.

35. Yagura, T., Kom, T. & Seno, T. (1982) Eur. J. Biochem. 123,15-

36. Teraoka, H. & Tsukada, K. (1980) Biochem. Biophys. Res.

37. Wintersberger, E. & Mudrak, I. (1982) FEBS Lett. 147,215 -219.

403-412.

475, 553-561.

131 - 140.

Acad. Sci. USA 78, 917-921.

Acta 344, 263 - 305.

Biochem. 49, 627- 666.

520, 271 - 284.

Biophys. Acta 653, 316- 330.

17, 1073-1080.

1526.

Biophys. Acta 696, 145 - 153.

21.

Commun. 95, 638 - 643.

38. James, M. R. & Lehmann, A. R. (1982) Biochemistry 21,4007-

39. Creissen, D. & Shall S (1982) Nature (Lond.) 296, 271 -272. 40. Brawerman, G. & hapiro, H. S. (1962) in Comparative Bio-

chemistry (Florkin, M. & Mason, H. S., eds) p. 140, Academic Press, New York.

41. Hershey, H. W. &Taylor, J. H. (1975) Exp. Cell Res. 94,339 - 350. 42. Akiyoshi, H. (1981) J. Biochem. (Tokyo) 90, 1597-1601. 43. Herman, T. M., DePamphilis, M. L. & Wassarman, P. M. (1979)

44. Roufa, D. J. & Marchioni, M. A. (1982) Proc. Natl Acad. Sci. USA

45. Panniers, L. R. V. & Clemens, M. J. (1981) J. Cell Sci. 48, 259-

46. Lundblad, D. & Lundgren, E. (1981) Int. J. Cancer 27,749-754. 47. Lundblad, D. & Lundgren, E. (1982) Biochem. Biophys. Res.

48. Todaro, G. J. & Baron, S . (1965) Proc. Natl Acad. Sci. USA 54,

49. Turek, L. P., Byrne, J. C., Lowy, D. R., Dvoretzky, I., Friedman, R. M. & Howley, P. M. (1982) Proc. Natl Acad. Sci. USA 79,

50. Doetsch, P. M., Suhadolnik, R. J., Sawada, Y., Mosca, J. D., Flick, M. B., Reichenbach, N. L., Dang, A. Q., Wu, J. M., Charubala, R., Pfleiderer, W. & Henderson, E. E. (1981) Proc. Natl Acad. Sci. USA 78, 6699-6703.

51. Otsuka, H., Qavi, H. &Kit, S. (1982) Antiviral Research 2, 301 - 311.

52. Dubois, M.-F., Vignal, M., LeCunff, M. & Chany, C. (1983) Nature (Lond.) 303, 433-435.

53. St. Laurent, G., Yoshie, O., Floyd-Smith, G., Samanta, H., .Sehgal, P. B. & Lengyel, P. (1983) Cell 33, 95- 102.

4013.

Biochemistry 18, 4563 - 4571.

79, 1810- 1814.

279.

Commun. 105, 1569-1576.

752- 756.

7914-7918.

G. Moore, Departement de Biologie Moleculaire, Institut Pasteur, 25 - 28 Rue du Docteur-Roux, F-75 724 Paris-Cedex-15, France D. R. Gewert, Division of Infectious Diseases, Hospital for Sick Children, 555 University Avenue, Toronto, Ontario, Canada M5G 1x8 M. J. Clemens, Department of Biochemistry, St George’s Hospital Medical School, Cranmer Terrace, Tooting, London, England SW17 ORE