Replication of non-repetitive DNA during early, middle and late S phase

Experimental Cell Research 158 (1985) 333-341

Replication of Non-repetitive DNA during Early, Middle and Late S Phase The General

Class and the Subset

LEO J. GRADY, MYRNA Wadsworth

Center for Laboratories

Transcribed

L. SANDERS and WAYNE P. CAMPBELL and Research, New York State Department Albany, NY 12201, USA

of Health,

Synchronized cultures of Chinese hamster ovary cells were pulse-labelled with 5bromodeoxyuridine (BrdU) during early (O-2.0 h), middle (254.0 h) and late (4.5-6.0 h) S phase in two successive cell cycles. In each case, the DNA containing BrdU in both strands was duplicated at the same time in both cycles and was isolated for further characterization by centrifugation in CsCl gradients. These DNAs were then radiolabelled by nick-translation and used in either DNA-DNA or RNA-DNA hybridization experiments. In the DNA-DNA experiments, advantage was taken of the substantial rate increases attainable in high concentrations of dextran sulfate to obtain complete reassociation curves with relatively small amounts of material. Assuming that no unresolved low repetition frequency components exist, renaturation kinetics suggest that early replicating DNA contains a greater proportion of non-repetitive sequences than DNA synthesized at later times, the order being early>middle>late. However, in terms of complexity the nonrepeated DNA duplicated early had only 74% of the diverse sequences present in logphase cells, whereas that replicated in middle and late S phase had 82 and 79.5%, respectively. It therefore appears that while DNA synthesized at different times in S phase may contain varying proportions of non-repetitive sequences, when their diversity is taken into account very few of these sequences (25% or less) exhibit temporal control of replication. Finally, measurements with total cell RNA indicated that the transcribed fraction of non-repeated DNA showed a slight preference for replication in early S phase. 0 1985 Academic Press, Inc.

There is a substantial body of data consistent with the idea that DNA replication in mammalian cells is under some type of temporal control. The evidence takes several forms: (1) density labelling experiments showing that DNA replicated at a given time in one S phase tends to be replicated at the same time in a subsequent S phase [l-4]; (2) when DNA duplicated during a particular interval of S phase (usually the first half) is damaged, changes have been observed in cells with respect to toxicity [5], mutation frequency [6, 71, the frequency of cell transformation [8], and in the synthesis of later replicating DNA [9]; (3) in CHO cells the genes coding for ribosomal RNA have been shown to replicate in early S phase [lo]; (4) also in CHO cells, the light bands of Giemsa-stained chromosomes replicate in the first half of S phase and the dark bands in the last half [ 111;and (5) the genes coding for several proteins have been found to replicate at distinct times during S phase [12-161. 22-858336

Copyright @I 1985 by Academic Ress, Inc. All rights of reproduction in any form resewed 0014-w7/85 $03.00

334 Grady, Sanders and Campbell In contrast to the above, more conflicting results have been obtained from studies designed to determine if there is a particular time of replication for general classes of DNA, such as repetitive and non-repetitive sequences and the sequences transcribed into RNA. In this regard, whereas Comings & Mattoccia [17], and Holmquist et al. [ 1I] found no differences in the distribution between repetitive and non-repetitive sequences of early and late replicating DNA, May & Belle [3] obtained evidence that some repeated sequences are preferentially replicated during the second half of the S phase. With respect to that portion of the genome transcribed into RNA, neither Meltz & Okada [I81 nor Holmquist et al. [ 1l] were able to detect any specific timing of the replication of these DNA sequences. It must be noted, however, that the experiments were concerned primarily with repetitive DNA which consists of families of closely related, but not identical sequences. It is thus possible that the RNA hybridized not with the region of the genome from which it was transcribed, but with one or more closely related components. If the replication of these similar sequences was spread across S phase, the result would be to mask the time of replication of the actual sequence transcribed. More recently, in a study focusing on the non-repetitive DNA where such problems of interpretation should be lessened, Belle reported that the transcribed fraction was predominantly replicated in the first half of the S phase [ 191. In the studies with repetitive and non-repetitive DNA referred to above, the region of the genome duplicated in a particular portion of S phase was marked by incorporating either a radiolabel or density label during the appropriate time period. However, a potential drawback to this procedure is that only about 50% of the DNA so labelled in one S phase replicates at the same time in a second S phase [ 1, 3,4]. Whether this should be attributed to a lack of temporal control or to asynchrony in the culture is still unclear. In either case, it is difficult to evaluate any effects this phenomenon may have on experimental results. To circumvent this problem, we previously introduced a method which utilizes density labelling during the same interval in two successive S phases as a means for isolating a more pure population of sequences replicated at specific times [4]. In the present investigation this method was employed to further pursue the question of the timing of replication of the general classes of repetitive and nonrepetitive DNA, as well as of the transcribed fraction of the non-repetitive sequences. MATERIALS

AND METHODS

Density Labelling and DNA Isolation from Synchronized Cells Chinese hamster ovary (CHO) cells were generously provided by Dr R. A. Tobey of the Los Alamos Scientific Laboratories. The procedures for cell growth and synchronization, as well as for isolating DNA from various intervals of S phase, have been described in detail [4]. Basically, the cells were synchronized by the isoleucine deprivation-hydroxyurea method of Tobey [20] which arrests the cells in early S phase with less than 4% of the DNA replicated [21]. After release from the hydroxyurea block, the cells were pulse-labelled with S-bromodeoxyuridine (BrdU, 6 pg/ml) during Exp Cell Res 158 (1985)

Replication of non-repetitive DNA

335

either early (O-2 h), middle (2.5-4.0 h), or late (4.5-6.0 h) S phase in two successive cell cycles. DNA containing BrdU in both strands was replicated at the same time in both cycles and was isolated by banding twice in CsCl gradients. Prior to banding, the DNA was sheared by passage through a 27 g hypodermic needle at maximum finger pressure. The starting density for the first CsCl gradient was 1.715 g/cc, and for the second was 1.76 &cc. Control DNA was obtained from log phase cells (3-4~ Id/ml) grown for 30 h in the presence of BrdU.

Radiolabelling of DNA Purified DNA from each period of S phase, as well as control BrdU-containing DNA was radiolabelled in vitro by nick-translation 1221.Using [‘HldCTP (New England Nuclear, 15-30 Ci/mM) as the labelled precursor, specific activities of 34~ lo6 cpm/pg were obtained. The average piece size was 300-400 nucleotides as determined by gel electrophoresis.

Isolation of Non-repeated DNA To obtain non-repetitive sequences, approx. 0.5 ug of total DNA in 50 pl of 0.03 M phosphate buffer (PB) was denatured by heating for I.5 min in boiling water, followed by rapid cooling in an ice bath. The sample was then brought to a final volume of 0.2 ml containing 45 % dextran sulfate and 1.O M NaCl and incubated to a Cat equivalent to 2000 (these conditions result in a 1000-fold rate increase compared to that observed in 0.12 M PB at 60°C [23]). At this point, the sample was diluted IO-fold in 0.12 M PB at 60°C and passed over a 0.5 cc column of hydroxylapatite (HAP) at 60°C. The column was washed ten times with 1.0 ml 0.12 M PB and the fractions containing the radioactive, singlestranded (ss) DNA were pooled. After addition of 10 pg of sheared E. coli DNA as carrier, the solution was brought to 1.O M CsCl to precipitate the dextran sulfate [23]. The precipitate was packed by low-speed centrifugation, the supematant removed and the DNA then pelleted by centrifugation for 20 h at 62 K and 20°C in a Beckman SW65 rotor. The pellet was resuspended in 0.1 ml 0.03 M PB and residual CsCl removed by microdialysis against 0.03 M PB using a Millipore V series filter. Recoveries were better than 80 % of the ssDNA originally eluted from the HAP column.

Isolation of RNA Total cell RNA was isolated from log-phase cells by the guanidinium thiocyanate method of Chirgwin et al. using the modification in which RNA is separated from the guanidinium thiocyanate by centrifugation through a cushion of 5.7 M CsCl [24]. Following resuspension in the presence of 0.5% sarkosyl, the RNA was precipitated with cold ethanol. The precipitate was held overnight at -20°C collected by centrifugation and dissolved in 0.05 M Tris-HCI, pH 7.0, containing 0.1 M NaCl and 0.02 M MgC&. Any remaining DNA was removed by treating for 30 min at 37°C with 20 CLg/mlof DNase processed to remove traces of RNase [25]. Sodium dodecyl sulfate (SDS) was then added to 0.5 % and proteinase K to 250 &ml and the incubation continued for 45 min. The RNA was carried through several ethanol precipitations and finally resuspended in 0.01 M Tris-HCI, pH 7.0 containing 0.01 M EDTA.

DNA-DNA

Hybridization

The reassociation kinetics of DNA from early, middle and late S phase, as well as log-phase control DNA were measured in reactions containing 45 % dextran sulfate and 1.0 M NaCl prepared as described above. Reactions were run in 1.5 ml plastic tubes and the solutions were layered with liquid paraffin to prevent evaporation. The large increase in the reaction rate under these conditions (see above) made it possible to obtain complete reassociation curves using relatively small amounts of DNA. The concentrations employed in the homologous reactions were: Cot 0.625-120,O.l &ml; Cot 125-lo”, 40 @ml. Ahquots were withdrawn from the samples at different times and diluted 300-fold into 0.12 M PB at 60°C. The fraction of the DNA hybridized was then determined by passage over a column of HAP [26]. Experiments in which unlabelled DNA from the different intervals of S phase were used to drive reactions containing trace amounts of labelled, non-repetitive log phase DNA were conducted in a manner similar to the above. In this case the unlabelled DNAs were sheared by sonication to the same average piece size as the nick-translated DNAs. The driver and tracer concentrations were 40 and 0.04 @ml, respectively. Exp Cell Res 158 (1985)

336 Grady, Sanders and Campbell

Fig. I. Reassociation kinetics of DNA from (A) log-phase cells and sequences replicated in (B) early;

(C) middle and (0) late S phase. All Car values have been adjusted so as to be equivalent to incubation in 0.12 M phosphate buffer at 60°C [35]. In each panel the solid line depicts the best tit to the data as determined by non-linear least squares analysis.

RNA-DNA Hybridization The extent to which the transcribed fraction of the non-repetitive DNA was duplicated early and late in S phase was determined in RNA-excess hybridization experiments using total cell RNA from log-phase cells and trace amounts of either early or late replicating non-repetitive DNA prepared as above. All other details were as previously described 1271.

Data Analysis Non-linear least-squares analysis of the hybridization data was carried out on a microcomputer using the NNLSQ program (CET Research Group, Ltd) [28]. It was assumed that DNA-DNA reactions followed second-order kinetics [29, 301 and that RNA-DNA reactions followed first-order kinetics [31, 321.

RESULTS The initial measurements were made with BrdU-containing DNA from logphase cells for the purpose of obtaining a reference with which to compare the reassociation of DNA from different intervals of S phase, and also to verify that nick-translated, BrdU-substituted DNA did not show anomalous reassociation behavior. The results are presented in fig. 1A. Within the range of Cot examined, the kinetics were consistent with three main components: (1) a fairly repetitious class comprising 14.2 % of the genome; (2) a moderately repeated fraction equivalent to 21.2 % of the DNA; and (3) a non-repeated portion equal to 64.5 % of the ExpCell

Res 158 (1985)

Replication

Table 1. Summary of reassociation

measurements phase cells and from various portions of S phase

of non-repetitive

DNA

337

with total DNA from

log-

Log

Early

Middle

Late

1st Repetitive component ha

14.2% 0.212 kO.076

20.7 % 0.161 fO.055

7.6% 0.518 f0.455

27.7% 0.128 kO.025

2nd Repetitive component k2

21.2% 3.18~10-~ +5.70x 10-4

18.9% 1.6x lo-* &4.4X 10-3

25.5% 1.86x10-3 f3.21 x lO-4

Non-repetitive component k3

64.5% 1.64x lo-4 +1.22x1oP

73.3% 3.09x lo-4 +2.92x lo-’

45.8% 1.08x lO-4 ?1.06~10-~

78.9% 7.67x lO.-4 +5.78x 10-s

For reference, if E. coli (MW 2.7~ 104 is used for comparison and 3X lOI* is taken as the MW of the haploid CHO genome, then the calculated rate constants for the non-repeated components in the above cases are as follows:

Log, 64.5 % non-repetitive Early, 78.9 % non-repetitive Middle, 73.3 % non-repetitive Late, 46.8 % non-repetitive

Calculated

Observed

1.45x 10-4 1.77x10-4 1.65~10-~ l.o5xlo-4

1.64x 10-4 7.67x 1O-4 3.09x 10-4 1.08x lO-4

a Data corrected to standard conditions of 0.12 M phosphate buffer at 60°C [35] and for the fraction of the tracer not reactible (5-10%).

genome. Analogous results (data not shown) were obtained with DNA from logphase cells that lacked BrdU and that were labelled in culture with [3H]TdR. The data are also in fairly good agreement with those of Benz & Burki [331, who reported values of 14, 28, and 56%, respectively, for similar components in their detailed study of Chinese hamster M3-1 cell DNA. The reassociation kinetics of DNA isolated from early, middle and late S phase are shown in fig. 1B-D. In each case, the data illustrated included measurements with at least two DNA preparations isolated from completely independent synchronized cultures. The values for the different components extracted by nonlinear least squares analysis are summarized in table 1. The major features that emerge from these experiments are that DNA from early S phase had no readily discernible middle repetitive component and that non-repetitive sequences accounted for a greater proportion of early than late replicating DNA. These conclusions must be tempered, however, by the observation that early nonrepetitive DNA reassociated approx. four times faster than expected and thus it is possible that an unresolved component of low repetition frequency is present. If so, both of the above conclusions would be affected to some extent. In all of the other instances, the observed and calculated rate constants for the non-repetitive DNA varied by a factor of less than two. Exp CellRes 158(1985)

338 Grady, Sanders and Campbell

“- 3A

Fig. 2. Complexity measurements of the non-repetitive sequences duplicated in early, middle and late S phase. Radiolabelled non-repeated sequences purified from the DNA of log-phase cells were employed as tracer in reactions driven by unfractionated, unlabelled DNA from each time period. (A) Log-phase driver; (B-D) early, middle and late drivers, respectively. C,r values were adjusted as in fig. 1. The non-linear least squares best tit for a single component is shown in each instance by the solid line. Controls containing only the tracer DNA showed no reaction beyond the small non-specific binding (
The first set of experiments dealt with the proportion, on a weight basis, of early, middle and late replicating DNA that consisted of non-repetitive sequences. The next question to be addressed was the extent to which different non-repeated sequences are being duplicated during these intervals of S phase. It was reasoned that if a radiolabelled tracer DNA comprised of the non-repeated fraction from log phase cells, in which all of these sequences should be represented, was reacted with an excess of unlabelled DNA from a particular part of S phase, then only those tracer sequences that were also present in the unlabelled driver DNA would form duplexes. The percentage of the tracer hybridized when the reaction was run to completion would then provide an estimate of the sequence diversity of the non-repetitive DNA being replicated during the corresponding period of S phase. The outcome of such experiments is illustrated in fig. 2A for a homologous reaction using DNA from log-phase cells as driver, and in fig. 2 B-D for reactions in which the driver consisted of DNA replicated in early, Exp Cell Res 158 (1985)

.

Replication

of non-repetitive

DNA

339

middle and late S phases, respectively. The termination values for the reactions were approx. 74 % in the case of early replicating DNA, 82 % for DNA from midS phase, and 80% for late replicating DNA. It thus appears that on the basis of sequence diversity, only 25 % or less of the non-repetitive sequences are under temporal control. Nevertheless, if viewed from the perspective of coding capacity, this is equivalent to over a hundred thousand average size genes. To explore the possibility that the replication of the subset of non-repetitive sequences transcribed into RNA might exhibit more pronounced temporal control than that observed for non-repeated DNA in general, a series of RNA-DNA hybridization experiments were undertaken. The basic approach was to use large amounts of RNA from log phase cells, which should include representatives of all of the transcribed sequences, to drive reactions containing trace amounts of radiolabelled non-re:petitive DNA synthesized either early or late in S phase. Under such circumstances, the fraction of the DNA hybridized at the point where the reaction terminates should reflect the extent to which transcribed sequences are present in the DNA from each time period. The results of these experiments are presented in fig. 3. In all of the situations examined, the saturation values determined by nonlinear least squares methods were essentially the same as those obtained by visual inspection of the data. The calculated values were 10.0+0.7% when the non-repetitive tracer used came from log phase cells (fig. 3 A), 13.3f0.3 % when early replicating DNA was employed (fig. 3B), and 10.7kO.7 % in the case of DNA from late S phase (fig. 3C). However, it is necessary to correct the measurements with early and late replicating DNA to account for the fact that, in terms of sequence diversity, neither possesses a full complement of non-repeated sequences. When this adjustment is made (0.743 for early, 0.795 for late), the final values become 9.9 and 8.5%, respectively. Allowing for uncertainty in the calculations, a reasonable conclusion would be that about 10% of the transcribed non-repetitive sequences (equivalent to approx. 1% of the total non-repeated DNA) are preferentially replicated in early S phase. While this number may seem rather small, it represents enough information to code for thousands of average size messenger RNAs. DISCUSSION The possibility that the ratio of repeated and non-repeated DNA sequences might vary during S phase was previously investigated by several laboratories [3, 17, 111 with essentially negative results. In contrast, the data obtained in the course of the present experiments suggests that early replicating DNA contains a greater proportion of non-repeated sequences than does DNA duplicated at later times. To a large extent, the apparent differences can probably be ascribed to the fact that the current study used a different, and believed to be more stringent, criterion to define DNA replicated within a given interval of S phase (i.e., the incorporation of BrdU during the same period of two separate cell cycles) and Exp Cell Res 158 (1985)

340 Grady, Sanders and Campbell employed hybridization conditions under which reactions could be run to completion, a circumstance which permitted the reactibility of the DNA to also be taken into consideration. Turning our attention to the question of the complexity of the non-repetitive DNA duplicated in different intervals of S phase, to the best of our knowledge the only measurements currently available are those from the present work. Although they cannot be used to draw conclusions about events which occur at higher levels of chromosome organization, the data obtained indicate that in terms of sequence diversity no more than 25 % of the non-repeated DNA replicated under strict temporal control. Another general category of DNA on which interest has focused regarding its time of replication is the fraction transcribed into RNA, and both positive [19] and negative [ll, 181 results have been published. The negative data were obtained in experiments involving repetitive DNA and the problems inherent in their interpretation were briefly reviewed earlier and have also been discussed by Bello [19]. With respect to non-repeated sequences, it was observed in a study with human KB cells that the transcribed fraction as a class was replicated to a greater extent in early than in late S phase by a factor of about 3 : 1 [ 191. The difference detected in the present investigation, albeit in the same direction, was much less pronounced even before correction for the complexity of the early and late replicating non-repetitive DNA. At a higher level of resolution, but with a much more restricted population of sequences, data have recently been obtained using cloned DNA copies of approximately two dozen genes indicating that while some are replicated early and some late, transcription correlates with early replication [16]. It is also interesting to note that in a somewhat similar study it was found that the time of replication of a gene can change depending on whether or not it is transcriptionally active or inactive [34]. While the present measurements were aimed at the general class of transcribed sequences and thus did not approach the same level of sensitivity in terms of individual genes as the experiments just cited, there are no obvious inconsistencies in the two sets of data. The small difference in the transcribed sequences replicated in early and late S phase detected in the current work is more than enough to accommodate the observations made to date with cloned segments of the genome. Nevertheless, while not ruling out the possibility that some subset of transcribed sequences may be preferentially replicated at different stages of S phase, the present data strongly argue that transcription of a region of the genome is not synonymous with early replication. This material is based on work supported in part by the National Science Foundation under Grant No. PCM-8112211.

REFERENCES 1. Mueller, G C & Kajiwara, K, Biochim biophys acta 114 (1966) 108. 2. Taylor, J H, Myers, T L & Cunningham, H L, In vitro 6 (1971) 309. Exp Cell Res 158 (1985)

Replication

of non-repetitive

DNA

341

3. May, M S & Bello, L J, Exp cell res 83 (1974) 79. 4. Grady, L J & Sanders, M L, Exp cell res 138 (1982) 480. 5. Grisham, J W, Greenberg, D S, Kaufman, D G & Smith, G J, Proc natl acad sci US 77 (1980) 4813. 6. Watanabe, M & Horikawa, M, Mutation res 71 (1980) 219. 7. Chang, H L & Baserga, R, J cell physio192 (1977) 333. 8. Tsutsui, T, Barrett, J C & Ts’o, P 0 P, Cancer res 39 (1979) 2356. 9. Hamlin, J L, Exp cell res 112 (1978) 225. 10. Stambrook, P J, J mol biol 82 (1974) 303. 11. Holmquist, G, Gray, M, Porter, T & Jordan, J, Cell 31 (1982) 121. 12. Lo, S C, Ross, J & Mueller, G C, Biochim biophys acta 608 (1980) 103. 13. Furst, A, Brown, E H, Braunstein, J D & Schildkraut, C L, Proc natl acad sci US 78 (1981) 1023. 14. Epner, E, Rifkind, R A & Marks, P A, Proc natl acad sci US 78 (1981) 3058. 15. D’Andrea, A D, Tantravahi, U, Lalande, M, Perle, M A & Latt, S A, Nucleic acids res 11 (1983) 4753. 16. Goldman, M A, Holmquist, G P, Gray, M C, Caston, L A & Nag, A, Science 224 (1984) 686. 17. Comings, D E & Mattoccia, E, Proc natl acad sci US 67 (1970) 448. 18. Meltz, M L & Okada, S, Exp cell res 67 (1971) 90. 19. Bello, L J, Exp cell res 146 (1983) 79. 20. Tobey, R A, Methods in cell biology (ed D M Prescott) vol. 6, p. 67. Academic Press, New York (1973). 21. Walters, R A, Tobey, R A & Hildebrand, C E, Biochem biophys res commun 64 (1976) 212. 22. Rigby, P W J, Dieckmann, M, Rhodes, C & Berg, P, J mol biol 113 (1977) 237. 23. Weider, R & Wetmur, J G, Biopolymers 20 (1981) 1537. 24. Chirgwin, J M, Przybyla, A E, MacDonald, R J & Rutter, W J, Biochemistry 18 (1979) 5294. 25. Grady, L J, Campbell, W P & North, A B, Anal biochem 101 (1980) 118. 26. Grady, L J & Campbell, W P, Cancer res 35 (1975) 1559. 27. Grady, L J, Campbell, W P & North, A B, Nucleic acids res 7 (1979) 259. 28. Christian, S D & Tucker, E E, American lab 14 (1982) 31. 29. B&ten, R J & Kohne, D E, Carnegie inst Wash yearbook 65 (1967) 78. 30. Wetmur, J G & Davidson, N, J mol biol 31 (1968) 349. 31. Britten, R J, Carnegie inst Wash yearbook 67 (1969) 332. 32. Galau, G A, B&ten, R J & Davidson, E H, Cell 2 (1974) 9. 33. Benz, R D 8t Burki, H J, Exp cell res 112 (1978) 155. 34. Calza, R E, Eckhardt, L A, DelGiudice, T & Schildkraut, C L, Cell 36 (1984) 689. 35. B&ten, R J & Smith, J, Carnegie inst Wash yearbook 68 (1970) 378. Received August 30, 1984 Revised version received December 4, 1984

Exp CellRes 158(1985)