JOURNAL
OF
Vol. 32, No. 2
VIROLOGY, Nov. 1979, p. 386-393
0022-538X/79/11-0386/08$02.00/0
Replication of Polyoma DNA in Isolated Nuclei: Analysis of Replication Fork Movement GORAN MAGNUSSON* AND MAJ-GRETH NILSSON Medical Nobel Institute, Department of Biochemistry, Karolinska Institute, Stockholm, Sweden
Received for publication 20 March 1979
The movement of replication forks during polyoma DNA synthesis in isolated nuclei was analyzed by digesting newly synthesized DNA with the restriction endonuclease HpaII which cleaves polyoma DNA into eight unique fragments. The terminus of in vitro DNA synthesis was identified by cleaving newly completed molecules with HpaLL. The distribution of label in the restriction fragments showed that the in vitro DNA synthesis was bidirectional and had the normal terminus of replication. Analysis of replicative intermediates pulse-labeled in vitro further suggested that DNA synthesis in isolated nuclei is an ordered process similar to replication in intact cells. Replication forks moved with a constant rate from the origin towards the terminus of replication. The nonlinear course of the DNA synthesis reaction in the isolated nuclei seems to result from the random inactivation of replication forks rather than a decrease in the rate of fork movement. During the in vitro synthesis a replication fork could maximally synthesize a DNA chain about 1,000 nucleotides long. The results suggest that some replication forks might be initiated in vitro at the origin of replication. The replication of polyoma DNA has been studied in vitro with nuclei isolated from infected cells. In this system a detailed picture of the process for elongation of DNA chains has been obtained (6, 15). It was initially established (13) that the nuclei supported the semiconservative replication of substantial segments of the viral genome. However, no DNA strands were found that had been synthesized from start to finish in the isolated nuclei. Furthermore, the analytical methods used in that and subsequent studies did not allow a determination of the regions of the viral genome in which DNA was synthesized. Here we analyze newly synthesized DNA by the use of restriction endonuclease cleavage. Cleavage of DNA with restriction endonucleases allows the study of events in specific parts of a genome. The restriction endonuclease HpaII from Haemophilus parainfluenzae cleaves polyoma DNA into eight unique fragments that can be separated by gel electrophoresis. The eight fragments have been ordered into a physical map by Griffin et al. (7). DNA synthesis starts at a unique site in the genome, and from there the replication forks proceed in both directions until the two forks meet at the terminus of replication, which is 1800 from the origin on the circular map (3). To study the rate and extent of DNA synthesis in specific regions of the polyoma genome, viral DNA was labeled during the incubation of the isolated nuclei and then cleaved with HpalI.
The amount of radioactivity in the separated restriction fragments was then used to measure the number of replication forks that had moved through respective fragment during the labeling period.
386
MATERIALS AND METHODS For many of the details concerning cells, virus, and general methodology, previous publications should be consulted (11, 15, 19). Cells and virus. Growing cultures of 3T6 cells were infected at a multiplicity of 20 to 50 PFU/cell. Nuclei were isolated at about 26 h postinfection. The polyoma virus was of the A2 type (7) of the Pasadena largeplaque strain. Virus stocks were prepared by infecting primary mouse kidney cells at a multiplicity of about 0.01 PFU/cell with repeatedly plaque-purified virus. Chemicals. ['H]thymidine (20 Ci/mmol) and [a2P]dGTP (100 Ci/mmol) were obtained from New England Nuclear Corp. Prelabeling of DNA with [3H]thymidine.
[3H]thymidine was added directly to the medium to a
final concentration of 1 MM (20 pCi/ml). After about 4 h, the labeling was terminated by removing the medium and rinsing the cell monolayer twice with icecold Tris-buffered saline. The cells were immediately used for preparation of nuclei. In vitro synthesis and purification of DNA. Nuclei were prepared exactly as described before (15). After se(dimentation the nuclei were suspended in 1 volume of isotonic buffer. For in vitro synthesis of DNA, four parts of the nuclei suspension were mixed with one part of buffer containing the standard components of the reaction. Incubations were done at
VOL. 32, 1979
POLYOMA DNA REPLICATION
387
25°C and were stopped by the addition of 5 volumes recovered as form I. The corresponding value of 50 mM Tris-chloride (pH 8.0)-10 mM EDTA. The for the material isolated after 30 min of incubanuclei were lysed by the addition of sodium dodecyl tion was 3.7%. sulfate to lVi; and NaCl to 1 M. Viral DNA was Form I DNA from the two samples was then selectively extracted and purified as described, including centrifugation through neutral sucrose gradients cleaved with HpaII, and the digests were ana(9, 12). For purification of closed circular (form I) lyzed by gel electrophoresis. The 32P radioactivDNA, chromatography on benzoylated-naphthoylated ity of each DNA fragment was normalized to the size of the fragment by using the 'H radioactivity DEAE-cellulose was used (19). Restriction endonuclease digestion of DNA. as an internal standard. In Fig. 1 the specific :2P The restriction endonuclease from H. parainfluenzae radioactivity for each fragment is shown. In the (Hpall) was purified as described by Sharp et al. (16). figure the physical map of the circular polyoma DNA was concentrated by ethanol precipitation be- DNA molecule has been linearized by opening fore digestion. The reactions were carried out at 37°C it at the terminus of in vivo replication. The for 2 h in 10 mM Tris-chloride (pH 7.5)-10 mM MgCl2 molecules which had a completed replication and 1 mM dithiothreitol. The reactions were stopped by the addition of sodium dodecyl sulfate to 0.5%. The round during the first 10 min of incubation were samples were stored frozen and were heated to 50°C labeled exclusively in fragments 2 and 6. After 30 min of incubation some radioactivity was for 15 min before gel electrophoresis. Gel electrophoresis. Electrophoresis was carried present in fragments 1 and 7 in addition to out in cylindrical gels (6 by 150 mm) consisting of 2.9% fragments 2 and 6. It is clear that the termination acrylamide, 0.15% bisacrylamide, 0.5% agarose in 0.09 site in vitro was in the same region of the genome M Tris-borate buffer (pH 8.3), 0.0025 M EDTA, 10% as in vivo (3). Furthermore, the distribution of glycerol, and 0.5% sodium dodecyl sulfate (14). The radioactivity in the fragments was symmetrigels were run at 110 V for 7 h. The gels were then fractionated with a Gilson gel slicer and analyzed for cally arranged around the terminus, suggesting that the DNA synthesis was bidirectional. Only radioactivity. insignificant amounts of radioactivity were found in fragments 3 and 5 at the origin site for RESULTS DNA replication, located at the junction beOrigin and direction of polyoma DNA tween the two fragments. Consequently, the synthesis in vitro. For the analysis of the
progression of replication forks through the polyoma genome during DNA synthesis in isolated nuclei, it was necessary to establish that DNA replication in vitro had the same origin and direction of synthesis as the in vivo process. For this purpose we used the technique described by Danna and Nathans (4). Replicating DNA is radioactively labeled, and only mature (form I) DNA molecules are examined. After a short labeling period, the DNA will only be labeled at the terminus of replication. By increasing the pulse length, replicating molecules, which at the start of the labeling period were further from completion, will have time to finish replication. In this way radioactivity incorporated into completed molecules will form a gradient from the terminus towards the origin of replication. This gradient can be determined by cleavage of the labeled DNA with a restriction enzyme and measurement of the radioactivity in each fragment. Nuclei from polyoma-infected cells, prelabeled for several hours with [3H]thymidine, were incubated with [a-32P]dGTP under standard conditions for 10 or 30 min. Total viral DNA was applied to benzoylated-naphthoylated DEAE-cellulose, and form I was eluted with 1 M NaCl and further purified by centrifugation in CsCl-propidium diiodide density gradients. From the nuclei incubated for 10 min, 1.5% of the total 2p radioactivity in viral DNA was
18
. SH2I 0.
2 2
78 4 5 3
1
6
Fragment Order FIG. 1. Distribution of radioactivity in newly completed form I DNA molecules. Nuclei isolated from cells prelabeled with [3H]thymidine were incubated under standard conditions with [a-32P]dGTP for 10 or 30 min. Viral DNA was extracted, and form I DNA was purified and digested with HpaII restriction endonuclease. The resulting restriction fragments were separated by gel electrophoresis. The 3H radioactivity from the in vivo prelabeling and the 32p radioactivity incorporated during the in vitro incubation were measured, and the 32P/3H ratio for each fragment was calculated. This ratio is plotted versus respective fragment on a linearized physical map as determined by Griffin et al. (7). Symbols: O---O, 10-min incubation; -, 30-min incubation.
388
MAGNUSSON AND NILSSON
J. VIROL.
data show that DNA synthesis in isolated nuclei restriction fragments consisted of branched is an ordered process in which replication forks DNA molecules containing replication forks beprogress in the same direction as in replication tween two restriction sites and therefore had a in intact cells. larger mass than the normal linear fragments. It Flow of replication forks during in vitro is also shown below (Fig. 5) that the DNA repolyoma DNA synthesis. In the previous sec- covered from the peaks had full fragment length. tion, form I DNA molecules that had been comTo quantitate the synthesis of the different pleted in vitro were analyzed. A similar analysis restriction fragments, 32P/3H ratios were calcuwas done with the total viral DNA synthesized lated for uncleaved DNA and for each of the -in the nuclei. As mentioned above, more than fragments after correction for the 32P back95% of the in vitro-labeled DNA was replicative ground-level present between the peaks of radiointermediates; therefore, the contribution of la- activity (Table 1). The incorporation of :2P-label bel from form I DNA was ignored in this exper- into total DNA followed a nonlinear time course iment. Cells were prelabeled with [ 3H]thymidine before isolation of nuclei. The nuclei were then incubated for 10, 20, or 30 min. The non-radioactive nucleotides dATP, dCTP, and dTTP were present at 50 MM concentration, whereas [a32P]dGTP was present at 5 ,uM. Reactions were stopped by the addition of buffer containing EDTA and sodium dodecyl sulfate. The viral DNA was then selectively extracted and purified.
To analyze into what regions of the genome [a-:2P]dGTP was incorporated during the course of the reaction, the purified viral DNA was cleaved with HpaII restriction endonuclease. After the digestion, the DNA fragments were separated by electrophoresis (Fig. 2A, B, and C). The ;3H label introduced in vivo, serving as an internal standard, was recovered in seven peaks corresponding to HpaII fragments 1 to 7 (fragment 8 ran off the end of the gels). The amount of radioactivity in each of these peaks was proportional to the size of the restriction fragment (7). The ;12P radioactivity showed a different profile. It was recovered at positions corresponding to each of the HpaII fragments. In addition, 30 to 40% of the 32P-labeled DNA remained at the top of the gels, having a mobility less than that of HpaII-1. The relative amount of this material decreased during the course of the reaction. In other experiments the slowly migrating material could be eliminated by following a pulse-chase protocol (data not shown). These results suggest that the radioactivity at the top of the gels and between the positions of the
10
30
50
FradN
FIG. 2. HpaII cleavage pattern of in vitro-labeled DNA. Viral DNA prelabeled with [3H]thymidine was further labeled in vitro with [a-32P]dGTP for 10 (A), 20 (B), or 30 min (C). The purified DNA was digested with HpaII, and the resulting fragments were separated by electrophoresis in polyacrylamide gels. Symbols: 0--0, 3 H; * 32 p.
TABLE 1. Distribution of radioactivity in HpaII fragments of in vitro synthesized DNA"
12P/
Labeling period (min)
0-10 0-20 0-30
"The 32P/'H
Total
HpaII-1
HpaII-2
H radioactivity (x 10) ratio
HpaII-3
2.78 0.61 0.51 0.85 4.40 1.28 1.13 2.24 5.00 2.00 1.83 2.75 ratios are calculated from the experiment shown in
HpaIl-4
HpaII-5
HpaII-6
HpaII-7
0.95 2.19 3.25 Fig: 2.
1.80 2.65 2.83
0.79 2.01 2.79
1.16 3.24 4.22
VOL. 32, 1979
as previously observed (19). The 32P/3H ratios of the restriction fragments were in all cases lower than the corresponding ratio for uncleaved DNA. This result merely reflects the loss of material containing forks from the positions of the restriction fragments in the gels. That HpaII-l, for example, showed a lower recovery than the much smaller HpaII-7 is also consistent with this notion because a larger fragment has a higher probability of containing a replication fork than a small one. In general the 32P/3H ratios increased almost linearly with time during the incubation period. This increase probably represented the composite effect of incorporation of labeled nucleotides into DNA and an increase of the radioactivity recovered in linear restriction fragments. One exception to this pattern was HpaII-5 that had by far the highest :32P/^H ratio after the first 10 min of incubation, but then increased only about 1.5-fold during the continued reaction. The result suggests that early during the in vitro reaction, there was a relative abundance of replication forks close to the origin of replication, located at the junction between HpaII-3 and -5, and that this region of the genome later was depleted of replication forks. One tempting interpretation of the result is that initiation of DNA synthesis occurs in vitro, but that the process is inactivated faster than the elongation of DNA chains. To get a more sensitive deternination of the movement of replication forks, an experiment similar to the previous one was done. However, in this experiment the radioactive label was only introduced during the last 10 min of the reaction, instead of continuously from the start. The reactions were started in the presence of unlabeled nucleotides, dATP, dCTP, and dTTP present at 50 ,uM concentration, and dGTP at 5 ,uM concentration. During a 10-min pulse (0 to 10, 10 to 20, and 20 to 30 min, respectively), [a-YP]dGTP was added to a final dGTP concentration of 10 ,uM. Analysis of the viral DNA by sedimentation through alkaline sucrose gradients (Fig. 3A and B) showed that the 3H radioactivity introduced in vivo sedimented in two peaks: a major peak at 53S (form I DNA) and a minor peak at 16 to 18S (form II DNA). The 3P label from the 0- to 10-min pulse (Fig. 3A) appeared as an asymmetric peak extending from a position corresponding to full-length viral strands to the top of the gradient. The 3P label from the 20- to 30-min pulse (Fig. 3B) sedimented as a relatively sharp peak at about 16S and a second, smaller peak at the top of the gradient. The material in this second peak may have consisted of "Okazaki fragments," newly initiated DNA chains, or sim-
POLYOMA DNA REPLICATION
389
P
.X
B 0
It I I
10 Fra
20 30 onumber
FIG. 3. Sedimentation in alkaline sucrose gradients of viral DNA pulse-labeled in vitro. Nuclei isolated from cells prelabeled with [3H]thymidine were incubated for either 10 (A) or 30 min (B), with [oa-32P]dGTP present from 0 to 10 (A) or 20 to 30 min (B) after the start of the incubation. Viral DNA was selectively extracted and purified, and portions were sedimented through alkaline 5 to 20% sucrose gra, 32 p. H dients. Symbols: O--_O, 3H;
ply degradation products. The 32P-labeled viral DNA from the 10- to 20-min pulse had a sedimentation profile intermediate between the profiles shown in Fig. 3. The sedimentation profiles show that the average chain length increased during the reaction. The label in the chains was introduced during the last part of the incubations and thus reflected the synthesis of DNA behind replication forks that remained active during the reaction. Since mainly long DNA chains were synthesized during the later part of the reaction, it is again clear that the isolated nuclei largely support chain elongation. As in the previous experiment the viral DNA was cleaved with HpaII restriction endonuclease, and the resulting digests were analyzed by gel electrophoresis. We only show the gel profiles from the 0- to 10-min and 20- to 30-min pulses (Fig. 4A and B). They were similar to those shown in Fig. 2. Ratios of 32P/3H radioactivity were calculated for uncleaved DNA and the
J. VIROL.
390
MAGNUSSON AND NILSSONJ.Vo.
fragments (Table 2). The decreasing rate of the overall reaction. The recovery of HpaII fragments 1 to 7, labeled during the 0- to 10-min pulse, was similar to what was found previously (Table 1). Looking at the rates of synthesis of individual fragments, they generally decreased in parallel with the overall rate of synthesis. One obvious exception was the synthesis of HpaII-5. It started at a high level, and then (during 20 to 30 individual restriction
results show the
min)
decreased to about 10% of the initial rate.
However, relative 5 had
min
a
fragments, HpaIIsynthesis during 20 to 30
to the other
normal rate of
after the start of the incubation.
Does initiation of
new
replication
forks
occur
50 -A
nuclei? The high rate of fragment synthesis early during the in vitro reactions suggests that initiation might occur, but to answer the question we need to know whether HpaII-5 was completely synthesized during the in vitro reaction. In the previous section we in isolated
HpaII-5
that
DNA
contained
between two restriction sites
in the positions of normal fragments. If this notion is correct, labeled restriction fragments isolated from gels should be of full length. The extent of synthesis of the fragments in vitro could be determined by using a density label, in addition to a radioactive label, during the in vitro incubation. An experiment similar to the previous one was done. Viral DNA was density labeled with bromodeoxyuridine triphosphate instead of dTTP throughout 10 or 30 min of incubation and radioactively labeled with [a_-12p]dGTP either during the 0 to 10 or 20 to 30 min after the start of were
not recovered
restriction
the
incubation.
Hpall
30
labeled
argued that the replication forks
The
DNA
was
cleaved
with
restriction endonuclease, and the result-
ing fragments were separated electrophoretically, located by Cerenkov radiation, eluted electrophoretically from the gels, and concentrated by ethanol precipitation. Only the results of the analysis of fragments Hpall-3 and -5 are shown, but similar results were obtained with HpaII-6 and -7. The fragments were first analyzed by sedimentation through alkaline sucrose gradients. In Fig. 5 profiles of HpaII-3 (A and B) and HpaII5 (C and D) labeled during 0 to 10 min (A and
6r
C)
or
20 to 30
min
nuclei incubation
ing
0 to 10
(B and D) after the
are
min sedimented sharp :2p-labeled material had
peaks.
as
bulk of the
start of
shown. DNA labeled dur-
a
The
somewhat
velocity because of its density resulting from bromouracil
increased sedimentation increased 40
20
Fraction
substitution (8). It is clear that most of the :12p_
60
labeled DNA chains had the full
number
restriction
FicG.
HpaII cleavage pattern of in vitro pulsesynthesized in t'itro for 10 (A) or- 30 min (B) as described in the legend to Fig. 3 uwas cleaved with HpaII restriction endonuclease. The cleaved DNA uwas electrophoresed through poly. acrylamide gels.
fragments. However,
4.
labeled between 20 and 30
labeled DNA. Viral DNA
T'ABLE, 2. Distribution of radioactiuity
about 20% of the
[32
P]DNA
min was
length
of the
in the material
(Fig.
5B
and D)
shorter than full
length. The degree HpaII-3 and -5
of
bromouracil
was
substitution
then measured
by
HpaII fragments of DNA pulse-labeled in vitro" .2p/3H radioactivity (x 10) ratio
in
Labeling period (min)
0-10 10-20 20-30
The
12 P/H
TDNal
Hpall-I
HpaII-2
HpaHI-3
Hpall-4
Hpall-5
Hpall-6
Hpall-7
1.01 0.50 0.32
0.21 0.10 0.07
0.15 0.08 0.05
0.33 0.16 0.08
0.34 0.15 0.09
0.61 0.18 0.08
0.20 0.12 0.09
0.57 0.30 0.22
ratios are calculated from the experiment shown in Fig. 4.
in
centrifu-
VIOL. 32, 1979
Fracon nmber FIG. 5. Sedimentation in alkaline sucrose gradients of restriction fragments HpaII-3 and HpaII-5 fronm DNA pulse-labeled in vitro. Isolated nuclei were incubated as described in the legend to Fig. 4 for 10 (A and C) or 30 min (B, D). Bromodeoxyuridine triphosphate was present throughout the reactions, and [a-32P]dGTP was added during the last 10 min of the incubation as a pulse-label. After digestion of viral DNA with HpaII and separation of the DNA fragments, HpaII-3 (A and B) and HpaII-5 (C and D) uere isolated and sedimented through alkaline 5 to 20%/r sucrose gradients (6 h at 55,0f0 rpm at 4°C in a Beckman SW56 rotor). Symbols: O---O, 3H; v v 32p
gation in alkaline cesium sulfate density gradients (Fig. 6A, B, C, and D). The DNA labeled with 3H in vivo formed symmetrical peaks serving as internal references of buoyant density. HpaII-3 DNA (Fig. 6A and B) formed narrower peaks than HpaII-5 DNA (Fig. 6C and D), as expected from the difference in molecular weight. HpaII-3 (Fig. 6A) and HpaII-5 (Fig. 6C) labeled with 12P during the first 10 min of the reaction had a higher buoyant density and formed somewhat broader peaks than the 'Hlabeled DNA. The peak values of buoyant density were increased by 20 mg/cm: for HpaII-3 and 37 mg/cm3 for HpaII-5. After 30 min of incubation the "2P-labeled DNA formed substantially broader peaks (Fig. 6B and D) with peak values of buoyant density increased by 58 mg/cm3 for HpaII-3 and 54 mg/cm' for HpaII5. Knowing the adenine-thymine content of the two restriction fragments (7), the buoyant density in alkaline cesium sulfate gradients of DNA with complete substitution of thymine by bromouracil can be calculated (1, 10, 13) to be 83 and 79 mg/cm', respectively, for HpaII-3 and HpaII-5. The bromouracil substitution in HpaII-3 and -5 synthesized during 0 to 10 min was consequently 24 and 47%, respectively. The corresponding values for DNA labeled during 20
POLYOMA DNA REPLICATION
391
to 30 min were 70 and 68%. We believe that 70% represent full substitution under our experimental conditions, since no fragment we tested, including the small HpaII-7, had a higher degree of bromouracil substitution. From the length of HpaII-3 and -5 (890 and 410 nucleotides, respectively), the rate of chain elongation can be estimated to about 300 nucleotides during the first 10 min. During the next 20 min, the number of newly synthesized nucleotides in HpaII-3 had increased to 890, whereas the corresponding value for HpaII-5 was 400, representing the full length of the fragments. These values are corrected for a maximal bromouracil substitution of 70%. From these considerations it follows that the rate of DNA synthesis in the isolated nuclei is linear, when only the replication forks that remain active are measured. These replication forks seem to be able to synthesize a DNA chain at least 900 nucleotides long which is in agreement with the experiment presented in Fig. 1. In addition, the data suggest the possibility that some replication forks are initiated in vitro, since they are actively synthesizing DNA within 400 nucleotides from the origin of replication as late as 20 to 30 min after the start of the reaction.
I
I I
I
1020 3040 10 2030 40 FacUm mdm*
FIG. 6. Equilibrium centrifugation in alkaline cesium sulfate gradients of restriction fragments from DNA pulse-labeled in vitro. The restriction endonuclease fragments HpaII-3 (A and B) and HpaII-5 (C and D), generated from DNA labeled with bromodeoxyuridine triphosphate during 10 (A and C) or 30 min (B and D) incubation of isolated nuclei and radioactively labeled as described in the legend to Fig. 4, were centrifuged to equilibrium in alkaline cesium sulfate (1.45 g/cm3). Centrifugation was for 120 h at 25,000 rpm at 20°C in a Beckman SW50.1 rotor. The Ap values in the figure denote the increase in buoyant density calculated from refractive index 32P. measurements. Symbols: 0---0, 3H; *
392
MAGNUSSON AND NILSSON
DISCUSSION DNA synthesis in eucaryotic systems has been studied in detail with cells infected with papovaviruses. The viral genome which consists of a double-stranded circular DNA molecule is duplicated via a replicative intermediate in which DNA synthesis is initiated at a specific site and then continues bidirectionally until the two replication forks meet at a site opposite from the origin of replication (3, 4). The use of subcellular systems, isolated nuclei, and nucleoprotein preparations has led to further understanding of how DNA is synthesized at each replication fork (5, 6, 15, 17). In the present study we have used isolated nuclei and analyzed whether the replication forks move in the same way as during polyoma DNA synthesis in intact cells. We asked the following questions. (i) Does DNA synthesis in vitro terminate at the same site as the in vivo process? (ii) How are the replication forks distributed around the genome? (iii) What is the maximal length of a DNA chain synthesized by an individual replication fork, and does the synthesis occur at a constant rate? (iv) Is there any indication of initiation of new rounds of DNA synthesis? (i) It is clear that replicative intermediates that were successfully completed and transformed into closed circular molecules were terminated at the normal termination site (Fig. 1). The labeling pattern also shows that replication forks approach the termination site from both directions. After 10 min of incubation, in vitro radioactive label was detected only very close to the termination point. After 30 min of reaction, however, some molecules which at the start of the reaction had replication forks as far as 1,000 nucleotides from the termination point were successfully completed. Since only form I DNA was measured in this experiment, it is possible that even longer DNA chains were synthesized in vitro, but that those molecules had not yet been transformed into form I DNA. (ii) The distribution and movement of replication forks were analyzed by labeling the DNA radioactively during the incubation of the nuclei and then isolating the total viral DNA. Of the radioactivity more than 90% was present in replicative intermediates. The viral DNA was then digested with the restriction endonuclease HpaII, and the radioactivity was measured in the separated restriction fragments. By knowing the size and the location of the fragments in the genome, a measure of the distribution of replication forks should be obtained. It appears, however, that when replicative intermediates are digested with restriction enzyme, part of the DNA does not behave as normal linear frag-
J. VIROL.
ments. By using electrophoresis to separate the DNA fragments, those that contain a replication fork are lost. The quantitation in this type of experiment is therefore uncertain. Nevertheless, it is obvious from the results presented in Tables 1 and 2 that replication forks are rather evenly distributed around the genome. One exception was the incorporation of label into HpaIl-5 sequences that was quite high early during the reaction, but later dropped to normal values. These results are in accord with earlier observations (2; G. Bjursell, unpublished data) that replicative intermediates are evenly distributed between different stages of replication. However, different conclusions were reached from studies of simian virus 40 DNA replication, where a preponderance of "late" replicative intermediates was found (18). (iii) The movement and inactivation of replication forks were analyzed in an experiment in which DNA was density labeled with bromodeoxyuridine triphosphate throughout the reaction and, in addition, radioactively labeled during the last part of the reaction. This protocol was designed to analyze DNA synthesized at replication points that remained active throughout the reaction. After digestion with HpaII, restriction fragments which were of full length (Fig. 5) were isolated and analyzed by centrifugation in alkaline cesium sulfate gradients. The result, depicted in Fig. 6, shows that the density of the radioactively labeled DNA chains increased linearly with time, indicating that replication forks that remain active move with a constant speed throughout the reaction. The presence of lighter-than-average material in the gradients (Fig. 6B and D), consisting of full-length DNA chains with one light and one heavy part, might indicate that some replication forks either can pause and then resume synthesis or move quite slowly. The finding that HpaII-3 (Fig. 6B) can be completely synthesized in vitro again shows that replication forks can synthesize chains of at least 1,000 nucleotides in the nuclei (compare Fig. 1). (iv) The last question concerning initiation of new rounds of replication remains essentially unanswered. Initiation might occur, but the data are inconclusive. The high level of DNA synthesis that occurs close to the origin of replication early during the reaction (Table 2) could represent synthesis by replication forks initiated during the early part of the reaction. If this is true the rate of initiation then seems to drop rather quickly during the in vitro reaction. The information presented in Fig. 6 and Table 2 shows that even as late as between 20 and 30 min after the start of the incubation, sequences in HpaII5 are synthesized. From the same experiment it
VOL. 32, 1979
is clear that the synthesis of the DNA in HpaII5 takes less than 20 min, so the synthesis seen in HpaII-5 later than 20 min after the start of the reaction might in fact have been initiated during the first 10 to 15 min. A trivial explanation to this finding that we cannot rule out is that the replication forks were initiated in vivo and had synthesized DNA chains a few nucleotides long at the time the nuclei were isolated. In this case, however, we also have to assume that the replication forks pause for a substantial time period before they complete the synthesis of DNA in
HpaII-5. ACKNOWLEDGMENTS This work was supported by grants from the Swedish Cancer Society and Magnus Bergvall's Foundation. LITERATURE CITED 1. Baldwin, R. L., and E. M. Shooter. 1963. The alkaline transition of BU-containing DNA and its bearing on the replication of DNA. J. Mol. Biol. 7:511-526. 2. Bjursell, G., and G. Magnusson. 1976. Replication of polyoma DNA. Accumulation of early replicative intermediates during hydroxyurea inhibition. Virology 74: 249-251. 3. Crawford, L. V. C., C. Syrett, and A. Wilde. 1973. The replication of polyoma DNA. J. Gen. Virol. 21:515-521. 4. Danna, K. J., and D. Nathans. 1972. Bidirectional replication of Simian virus 40 DNA. Proc. Natl. Acad. Sci. U.S.A. 69:3097-3100. 5. Edenberg, H. J., M. A. Waqar, and J. A. Huberman. 1976;. Subnuclear systems for synthesis of Simian virus 40 DNA in titro. Proc. Natl. Acad. Sci. U.S.A. 73:43924396. 6. Eliasson, R., and P. Reichard. 1978. Primase initiates Okazaki pieces during polyoma DNA synthesis. Nature (London) 272:184-185. 7. Griffin, B. E., M. Fried, and A. Cowie. 1974. Polyoma DNA. A physical map. Proc. Natl. Acad. Sci. U.S.A. 71: 2077-2081.
POLYOMA DNA REPLICATION
393
8. Hirt, B. 1966. Evidence for semiconservative replication of circular polyoma DNA. Proc. Natl. Acad. Sci. U.S.A. 55:997-1004. 9. Hirt, B. 1967. Selective extraction of polyoma DNA from infected mouse cell cultures. J. Mol. Biol. 26:365-369. 10. Inman, R. B., and R. L. Baldwin. 1964. Helix-random coil transitions in DNA homopolymer pairs. J. Mol. Biol. 8:452-469. 11. Magnusson, G. 1973. Hydroxyurea-induced accumulation of short fragments during polyoma DNA replication. I. Characterization of fragments. J. Virol. 12:600608. 12. Magnusson, G., R. Craig, M. Narkhammar, P. Reichard, M. Staub, and H. Warner. 1974. Replication of polyoma DNA: effects of hydroxyurea and arabinosyl cytosine. Cold Spring Harbor Symp. Quant. Biol. 39: 227-233. 13. Magnusson, G., E.-L. Winnacker, R. Eliasson, and P. Reichard. 1972. Replication of polyoma DNA in isolated nuclei. II. Evidence for semi-conservative replication. J. Mol. Biol. 72:539-552. 14. Peacock, A. C., and C. W. Dingman. 1967. Resolution of multiple RNA species by polyacrylamide gel electrophoresis. Biochemistry 6:1818-1827. 15. Pigiet, V., R. Eliasson, and P. Reichard. 1974. Replication of polyoma DNA in isolated nuclei. III. The nucleotide sequence at the RNA-DNA junction of nascent strands. J. Mol. Biol. 84:197-216. 16. Sharp, P. A., B. Sugden, and J. Sambrook. 1973. Detection of two restriction endonuclease activities in Haemophilus parainfluenza using analytical agaroseethidium bromide electrophoresis. Biochemistry 12: 3055-3063. 17. Su, R. T., and M. L DePamphilis. 1978. Simian virus 40 DNA replication in isolated replicating viral chromosomes. J. Virol. 28:53-65. 18. Tapper, D. P., and M. L. De Pamphilis. 1978. Discontinous DNA replication: accumulation of simian virus 40 DNA at specific stages of replication. J. Mol. Biol. 120:401-422. 19. Winnacker, E.-L., G. Magnusson, and P. Reichard. 1972. Replication of polyoma DNA in isolated nuclei. I. Characterization of the system from mouse fibroblast 3'T'(i cells. ,J. Mol. Biol. 72:523-527.
OF
Vol. 32, No. 2
VIROLOGY, Nov. 1979, p. 386-393
0022-538X/79/11-0386/08$02.00/0
Replication of Polyoma DNA in Isolated Nuclei: Analysis of Replication Fork Movement GORAN MAGNUSSON* AND MAJ-GRETH NILSSON Medical Nobel Institute, Department of Biochemistry, Karolinska Institute, Stockholm, Sweden
Received for publication 20 March 1979
The movement of replication forks during polyoma DNA synthesis in isolated nuclei was analyzed by digesting newly synthesized DNA with the restriction endonuclease HpaII which cleaves polyoma DNA into eight unique fragments. The terminus of in vitro DNA synthesis was identified by cleaving newly completed molecules with HpaLL. The distribution of label in the restriction fragments showed that the in vitro DNA synthesis was bidirectional and had the normal terminus of replication. Analysis of replicative intermediates pulse-labeled in vitro further suggested that DNA synthesis in isolated nuclei is an ordered process similar to replication in intact cells. Replication forks moved with a constant rate from the origin towards the terminus of replication. The nonlinear course of the DNA synthesis reaction in the isolated nuclei seems to result from the random inactivation of replication forks rather than a decrease in the rate of fork movement. During the in vitro synthesis a replication fork could maximally synthesize a DNA chain about 1,000 nucleotides long. The results suggest that some replication forks might be initiated in vitro at the origin of replication. The replication of polyoma DNA has been studied in vitro with nuclei isolated from infected cells. In this system a detailed picture of the process for elongation of DNA chains has been obtained (6, 15). It was initially established (13) that the nuclei supported the semiconservative replication of substantial segments of the viral genome. However, no DNA strands were found that had been synthesized from start to finish in the isolated nuclei. Furthermore, the analytical methods used in that and subsequent studies did not allow a determination of the regions of the viral genome in which DNA was synthesized. Here we analyze newly synthesized DNA by the use of restriction endonuclease cleavage. Cleavage of DNA with restriction endonucleases allows the study of events in specific parts of a genome. The restriction endonuclease HpaII from Haemophilus parainfluenzae cleaves polyoma DNA into eight unique fragments that can be separated by gel electrophoresis. The eight fragments have been ordered into a physical map by Griffin et al. (7). DNA synthesis starts at a unique site in the genome, and from there the replication forks proceed in both directions until the two forks meet at the terminus of replication, which is 1800 from the origin on the circular map (3). To study the rate and extent of DNA synthesis in specific regions of the polyoma genome, viral DNA was labeled during the incubation of the isolated nuclei and then cleaved with HpalI.
The amount of radioactivity in the separated restriction fragments was then used to measure the number of replication forks that had moved through respective fragment during the labeling period.
386
MATERIALS AND METHODS For many of the details concerning cells, virus, and general methodology, previous publications should be consulted (11, 15, 19). Cells and virus. Growing cultures of 3T6 cells were infected at a multiplicity of 20 to 50 PFU/cell. Nuclei were isolated at about 26 h postinfection. The polyoma virus was of the A2 type (7) of the Pasadena largeplaque strain. Virus stocks were prepared by infecting primary mouse kidney cells at a multiplicity of about 0.01 PFU/cell with repeatedly plaque-purified virus. Chemicals. ['H]thymidine (20 Ci/mmol) and [a2P]dGTP (100 Ci/mmol) were obtained from New England Nuclear Corp. Prelabeling of DNA with [3H]thymidine.
[3H]thymidine was added directly to the medium to a
final concentration of 1 MM (20 pCi/ml). After about 4 h, the labeling was terminated by removing the medium and rinsing the cell monolayer twice with icecold Tris-buffered saline. The cells were immediately used for preparation of nuclei. In vitro synthesis and purification of DNA. Nuclei were prepared exactly as described before (15). After se(dimentation the nuclei were suspended in 1 volume of isotonic buffer. For in vitro synthesis of DNA, four parts of the nuclei suspension were mixed with one part of buffer containing the standard components of the reaction. Incubations were done at
VOL. 32, 1979
POLYOMA DNA REPLICATION
387
25°C and were stopped by the addition of 5 volumes recovered as form I. The corresponding value of 50 mM Tris-chloride (pH 8.0)-10 mM EDTA. The for the material isolated after 30 min of incubanuclei were lysed by the addition of sodium dodecyl tion was 3.7%. sulfate to lVi; and NaCl to 1 M. Viral DNA was Form I DNA from the two samples was then selectively extracted and purified as described, including centrifugation through neutral sucrose gradients cleaved with HpaII, and the digests were ana(9, 12). For purification of closed circular (form I) lyzed by gel electrophoresis. The 32P radioactivDNA, chromatography on benzoylated-naphthoylated ity of each DNA fragment was normalized to the size of the fragment by using the 'H radioactivity DEAE-cellulose was used (19). Restriction endonuclease digestion of DNA. as an internal standard. In Fig. 1 the specific :2P The restriction endonuclease from H. parainfluenzae radioactivity for each fragment is shown. In the (Hpall) was purified as described by Sharp et al. (16). figure the physical map of the circular polyoma DNA was concentrated by ethanol precipitation be- DNA molecule has been linearized by opening fore digestion. The reactions were carried out at 37°C it at the terminus of in vivo replication. The for 2 h in 10 mM Tris-chloride (pH 7.5)-10 mM MgCl2 molecules which had a completed replication and 1 mM dithiothreitol. The reactions were stopped by the addition of sodium dodecyl sulfate to 0.5%. The round during the first 10 min of incubation were samples were stored frozen and were heated to 50°C labeled exclusively in fragments 2 and 6. After 30 min of incubation some radioactivity was for 15 min before gel electrophoresis. Gel electrophoresis. Electrophoresis was carried present in fragments 1 and 7 in addition to out in cylindrical gels (6 by 150 mm) consisting of 2.9% fragments 2 and 6. It is clear that the termination acrylamide, 0.15% bisacrylamide, 0.5% agarose in 0.09 site in vitro was in the same region of the genome M Tris-borate buffer (pH 8.3), 0.0025 M EDTA, 10% as in vivo (3). Furthermore, the distribution of glycerol, and 0.5% sodium dodecyl sulfate (14). The radioactivity in the fragments was symmetrigels were run at 110 V for 7 h. The gels were then fractionated with a Gilson gel slicer and analyzed for cally arranged around the terminus, suggesting that the DNA synthesis was bidirectional. Only radioactivity. insignificant amounts of radioactivity were found in fragments 3 and 5 at the origin site for RESULTS DNA replication, located at the junction beOrigin and direction of polyoma DNA tween the two fragments. Consequently, the synthesis in vitro. For the analysis of the
progression of replication forks through the polyoma genome during DNA synthesis in isolated nuclei, it was necessary to establish that DNA replication in vitro had the same origin and direction of synthesis as the in vivo process. For this purpose we used the technique described by Danna and Nathans (4). Replicating DNA is radioactively labeled, and only mature (form I) DNA molecules are examined. After a short labeling period, the DNA will only be labeled at the terminus of replication. By increasing the pulse length, replicating molecules, which at the start of the labeling period were further from completion, will have time to finish replication. In this way radioactivity incorporated into completed molecules will form a gradient from the terminus towards the origin of replication. This gradient can be determined by cleavage of the labeled DNA with a restriction enzyme and measurement of the radioactivity in each fragment. Nuclei from polyoma-infected cells, prelabeled for several hours with [3H]thymidine, were incubated with [a-32P]dGTP under standard conditions for 10 or 30 min. Total viral DNA was applied to benzoylated-naphthoylated DEAE-cellulose, and form I was eluted with 1 M NaCl and further purified by centrifugation in CsCl-propidium diiodide density gradients. From the nuclei incubated for 10 min, 1.5% of the total 2p radioactivity in viral DNA was
18
. SH2I 0.
2 2
78 4 5 3
1
6
Fragment Order FIG. 1. Distribution of radioactivity in newly completed form I DNA molecules. Nuclei isolated from cells prelabeled with [3H]thymidine were incubated under standard conditions with [a-32P]dGTP for 10 or 30 min. Viral DNA was extracted, and form I DNA was purified and digested with HpaII restriction endonuclease. The resulting restriction fragments were separated by gel electrophoresis. The 3H radioactivity from the in vivo prelabeling and the 32p radioactivity incorporated during the in vitro incubation were measured, and the 32P/3H ratio for each fragment was calculated. This ratio is plotted versus respective fragment on a linearized physical map as determined by Griffin et al. (7). Symbols: O---O, 10-min incubation; -, 30-min incubation.
388
MAGNUSSON AND NILSSON
J. VIROL.
data show that DNA synthesis in isolated nuclei restriction fragments consisted of branched is an ordered process in which replication forks DNA molecules containing replication forks beprogress in the same direction as in replication tween two restriction sites and therefore had a in intact cells. larger mass than the normal linear fragments. It Flow of replication forks during in vitro is also shown below (Fig. 5) that the DNA repolyoma DNA synthesis. In the previous sec- covered from the peaks had full fragment length. tion, form I DNA molecules that had been comTo quantitate the synthesis of the different pleted in vitro were analyzed. A similar analysis restriction fragments, 32P/3H ratios were calcuwas done with the total viral DNA synthesized lated for uncleaved DNA and for each of the -in the nuclei. As mentioned above, more than fragments after correction for the 32P back95% of the in vitro-labeled DNA was replicative ground-level present between the peaks of radiointermediates; therefore, the contribution of la- activity (Table 1). The incorporation of :2P-label bel from form I DNA was ignored in this exper- into total DNA followed a nonlinear time course iment. Cells were prelabeled with [ 3H]thymidine before isolation of nuclei. The nuclei were then incubated for 10, 20, or 30 min. The non-radioactive nucleotides dATP, dCTP, and dTTP were present at 50 MM concentration, whereas [a32P]dGTP was present at 5 ,uM. Reactions were stopped by the addition of buffer containing EDTA and sodium dodecyl sulfate. The viral DNA was then selectively extracted and purified.
To analyze into what regions of the genome [a-:2P]dGTP was incorporated during the course of the reaction, the purified viral DNA was cleaved with HpaII restriction endonuclease. After the digestion, the DNA fragments were separated by electrophoresis (Fig. 2A, B, and C). The ;3H label introduced in vivo, serving as an internal standard, was recovered in seven peaks corresponding to HpaII fragments 1 to 7 (fragment 8 ran off the end of the gels). The amount of radioactivity in each of these peaks was proportional to the size of the restriction fragment (7). The ;12P radioactivity showed a different profile. It was recovered at positions corresponding to each of the HpaII fragments. In addition, 30 to 40% of the 32P-labeled DNA remained at the top of the gels, having a mobility less than that of HpaII-1. The relative amount of this material decreased during the course of the reaction. In other experiments the slowly migrating material could be eliminated by following a pulse-chase protocol (data not shown). These results suggest that the radioactivity at the top of the gels and between the positions of the
10
30
50
FradN
FIG. 2. HpaII cleavage pattern of in vitro-labeled DNA. Viral DNA prelabeled with [3H]thymidine was further labeled in vitro with [a-32P]dGTP for 10 (A), 20 (B), or 30 min (C). The purified DNA was digested with HpaII, and the resulting fragments were separated by electrophoresis in polyacrylamide gels. Symbols: 0--0, 3 H; * 32 p.
TABLE 1. Distribution of radioactivity in HpaII fragments of in vitro synthesized DNA"
12P/
Labeling period (min)
0-10 0-20 0-30
"The 32P/'H
Total
HpaII-1
HpaII-2
H radioactivity (x 10) ratio
HpaII-3
2.78 0.61 0.51 0.85 4.40 1.28 1.13 2.24 5.00 2.00 1.83 2.75 ratios are calculated from the experiment shown in
HpaIl-4
HpaII-5
HpaII-6
HpaII-7
0.95 2.19 3.25 Fig: 2.
1.80 2.65 2.83
0.79 2.01 2.79
1.16 3.24 4.22
VOL. 32, 1979
as previously observed (19). The 32P/3H ratios of the restriction fragments were in all cases lower than the corresponding ratio for uncleaved DNA. This result merely reflects the loss of material containing forks from the positions of the restriction fragments in the gels. That HpaII-l, for example, showed a lower recovery than the much smaller HpaII-7 is also consistent with this notion because a larger fragment has a higher probability of containing a replication fork than a small one. In general the 32P/3H ratios increased almost linearly with time during the incubation period. This increase probably represented the composite effect of incorporation of labeled nucleotides into DNA and an increase of the radioactivity recovered in linear restriction fragments. One exception to this pattern was HpaII-5 that had by far the highest :32P/^H ratio after the first 10 min of incubation, but then increased only about 1.5-fold during the continued reaction. The result suggests that early during the in vitro reaction, there was a relative abundance of replication forks close to the origin of replication, located at the junction between HpaII-3 and -5, and that this region of the genome later was depleted of replication forks. One tempting interpretation of the result is that initiation of DNA synthesis occurs in vitro, but that the process is inactivated faster than the elongation of DNA chains. To get a more sensitive deternination of the movement of replication forks, an experiment similar to the previous one was done. However, in this experiment the radioactive label was only introduced during the last 10 min of the reaction, instead of continuously from the start. The reactions were started in the presence of unlabeled nucleotides, dATP, dCTP, and dTTP present at 50 ,uM concentration, and dGTP at 5 ,uM concentration. During a 10-min pulse (0 to 10, 10 to 20, and 20 to 30 min, respectively), [a-YP]dGTP was added to a final dGTP concentration of 10 ,uM. Analysis of the viral DNA by sedimentation through alkaline sucrose gradients (Fig. 3A and B) showed that the 3H radioactivity introduced in vivo sedimented in two peaks: a major peak at 53S (form I DNA) and a minor peak at 16 to 18S (form II DNA). The 3P label from the 0- to 10-min pulse (Fig. 3A) appeared as an asymmetric peak extending from a position corresponding to full-length viral strands to the top of the gradient. The 3P label from the 20- to 30-min pulse (Fig. 3B) sedimented as a relatively sharp peak at about 16S and a second, smaller peak at the top of the gradient. The material in this second peak may have consisted of "Okazaki fragments," newly initiated DNA chains, or sim-
POLYOMA DNA REPLICATION
389
P
.X
B 0
It I I
10 Fra
20 30 onumber
FIG. 3. Sedimentation in alkaline sucrose gradients of viral DNA pulse-labeled in vitro. Nuclei isolated from cells prelabeled with [3H]thymidine were incubated for either 10 (A) or 30 min (B), with [oa-32P]dGTP present from 0 to 10 (A) or 20 to 30 min (B) after the start of the incubation. Viral DNA was selectively extracted and purified, and portions were sedimented through alkaline 5 to 20% sucrose gra, 32 p. H dients. Symbols: O--_O, 3H;
ply degradation products. The 32P-labeled viral DNA from the 10- to 20-min pulse had a sedimentation profile intermediate between the profiles shown in Fig. 3. The sedimentation profiles show that the average chain length increased during the reaction. The label in the chains was introduced during the last part of the incubations and thus reflected the synthesis of DNA behind replication forks that remained active during the reaction. Since mainly long DNA chains were synthesized during the later part of the reaction, it is again clear that the isolated nuclei largely support chain elongation. As in the previous experiment the viral DNA was cleaved with HpaII restriction endonuclease, and the resulting digests were analyzed by gel electrophoresis. We only show the gel profiles from the 0- to 10-min and 20- to 30-min pulses (Fig. 4A and B). They were similar to those shown in Fig. 2. Ratios of 32P/3H radioactivity were calculated for uncleaved DNA and the
J. VIROL.
390
MAGNUSSON AND NILSSONJ.Vo.
fragments (Table 2). The decreasing rate of the overall reaction. The recovery of HpaII fragments 1 to 7, labeled during the 0- to 10-min pulse, was similar to what was found previously (Table 1). Looking at the rates of synthesis of individual fragments, they generally decreased in parallel with the overall rate of synthesis. One obvious exception was the synthesis of HpaII-5. It started at a high level, and then (during 20 to 30 individual restriction
results show the
min)
decreased to about 10% of the initial rate.
However, relative 5 had
min
a
fragments, HpaIIsynthesis during 20 to 30
to the other
normal rate of
after the start of the incubation.
Does initiation of
new
replication
forks
occur
50 -A
nuclei? The high rate of fragment synthesis early during the in vitro reactions suggests that initiation might occur, but to answer the question we need to know whether HpaII-5 was completely synthesized during the in vitro reaction. In the previous section we in isolated
HpaII-5
that
DNA
contained
between two restriction sites
in the positions of normal fragments. If this notion is correct, labeled restriction fragments isolated from gels should be of full length. The extent of synthesis of the fragments in vitro could be determined by using a density label, in addition to a radioactive label, during the in vitro incubation. An experiment similar to the previous one was done. Viral DNA was density labeled with bromodeoxyuridine triphosphate instead of dTTP throughout 10 or 30 min of incubation and radioactively labeled with [a_-12p]dGTP either during the 0 to 10 or 20 to 30 min after the start of were
not recovered
restriction
the
incubation.
Hpall
30
labeled
argued that the replication forks
The
DNA
was
cleaved
with
restriction endonuclease, and the result-
ing fragments were separated electrophoretically, located by Cerenkov radiation, eluted electrophoretically from the gels, and concentrated by ethanol precipitation. Only the results of the analysis of fragments Hpall-3 and -5 are shown, but similar results were obtained with HpaII-6 and -7. The fragments were first analyzed by sedimentation through alkaline sucrose gradients. In Fig. 5 profiles of HpaII-3 (A and B) and HpaII5 (C and D) labeled during 0 to 10 min (A and
6r
C)
or
20 to 30
min
nuclei incubation
ing
0 to 10
(B and D) after the
are
min sedimented sharp :2p-labeled material had
peaks.
as
bulk of the
start of
shown. DNA labeled dur-
a
The
somewhat
velocity because of its density resulting from bromouracil
increased sedimentation increased 40
20
Fraction
substitution (8). It is clear that most of the :12p_
60
labeled DNA chains had the full
number
restriction
FicG.
HpaII cleavage pattern of in vitro pulsesynthesized in t'itro for 10 (A) or- 30 min (B) as described in the legend to Fig. 3 uwas cleaved with HpaII restriction endonuclease. The cleaved DNA uwas electrophoresed through poly. acrylamide gels.
fragments. However,
4.
labeled between 20 and 30
labeled DNA. Viral DNA
T'ABLE, 2. Distribution of radioactiuity
about 20% of the
[32
P]DNA
min was
length
of the
in the material
(Fig.
5B
and D)
shorter than full
length. The degree HpaII-3 and -5
of
bromouracil
was
substitution
then measured
by
HpaII fragments of DNA pulse-labeled in vitro" .2p/3H radioactivity (x 10) ratio
in
Labeling period (min)
0-10 10-20 20-30
The
12 P/H
TDNal
Hpall-I
HpaII-2
HpaHI-3
Hpall-4
Hpall-5
Hpall-6
Hpall-7
1.01 0.50 0.32
0.21 0.10 0.07
0.15 0.08 0.05
0.33 0.16 0.08
0.34 0.15 0.09
0.61 0.18 0.08
0.20 0.12 0.09
0.57 0.30 0.22
ratios are calculated from the experiment shown in Fig. 4.
in
centrifu-
VIOL. 32, 1979
Fracon nmber FIG. 5. Sedimentation in alkaline sucrose gradients of restriction fragments HpaII-3 and HpaII-5 fronm DNA pulse-labeled in vitro. Isolated nuclei were incubated as described in the legend to Fig. 4 for 10 (A and C) or 30 min (B, D). Bromodeoxyuridine triphosphate was present throughout the reactions, and [a-32P]dGTP was added during the last 10 min of the incubation as a pulse-label. After digestion of viral DNA with HpaII and separation of the DNA fragments, HpaII-3 (A and B) and HpaII-5 (C and D) uere isolated and sedimented through alkaline 5 to 20%/r sucrose gradients (6 h at 55,0f0 rpm at 4°C in a Beckman SW56 rotor). Symbols: O---O, 3H; v v 32p
gation in alkaline cesium sulfate density gradients (Fig. 6A, B, C, and D). The DNA labeled with 3H in vivo formed symmetrical peaks serving as internal references of buoyant density. HpaII-3 DNA (Fig. 6A and B) formed narrower peaks than HpaII-5 DNA (Fig. 6C and D), as expected from the difference in molecular weight. HpaII-3 (Fig. 6A) and HpaII-5 (Fig. 6C) labeled with 12P during the first 10 min of the reaction had a higher buoyant density and formed somewhat broader peaks than the 'Hlabeled DNA. The peak values of buoyant density were increased by 20 mg/cm: for HpaII-3 and 37 mg/cm3 for HpaII-5. After 30 min of incubation the "2P-labeled DNA formed substantially broader peaks (Fig. 6B and D) with peak values of buoyant density increased by 58 mg/cm3 for HpaII-3 and 54 mg/cm' for HpaII5. Knowing the adenine-thymine content of the two restriction fragments (7), the buoyant density in alkaline cesium sulfate gradients of DNA with complete substitution of thymine by bromouracil can be calculated (1, 10, 13) to be 83 and 79 mg/cm', respectively, for HpaII-3 and HpaII-5. The bromouracil substitution in HpaII-3 and -5 synthesized during 0 to 10 min was consequently 24 and 47%, respectively. The corresponding values for DNA labeled during 20
POLYOMA DNA REPLICATION
391
to 30 min were 70 and 68%. We believe that 70% represent full substitution under our experimental conditions, since no fragment we tested, including the small HpaII-7, had a higher degree of bromouracil substitution. From the length of HpaII-3 and -5 (890 and 410 nucleotides, respectively), the rate of chain elongation can be estimated to about 300 nucleotides during the first 10 min. During the next 20 min, the number of newly synthesized nucleotides in HpaII-3 had increased to 890, whereas the corresponding value for HpaII-5 was 400, representing the full length of the fragments. These values are corrected for a maximal bromouracil substitution of 70%. From these considerations it follows that the rate of DNA synthesis in the isolated nuclei is linear, when only the replication forks that remain active are measured. These replication forks seem to be able to synthesize a DNA chain at least 900 nucleotides long which is in agreement with the experiment presented in Fig. 1. In addition, the data suggest the possibility that some replication forks are initiated in vitro, since they are actively synthesizing DNA within 400 nucleotides from the origin of replication as late as 20 to 30 min after the start of the reaction.
I
I I
I
1020 3040 10 2030 40 FacUm mdm*
FIG. 6. Equilibrium centrifugation in alkaline cesium sulfate gradients of restriction fragments from DNA pulse-labeled in vitro. The restriction endonuclease fragments HpaII-3 (A and B) and HpaII-5 (C and D), generated from DNA labeled with bromodeoxyuridine triphosphate during 10 (A and C) or 30 min (B and D) incubation of isolated nuclei and radioactively labeled as described in the legend to Fig. 4, were centrifuged to equilibrium in alkaline cesium sulfate (1.45 g/cm3). Centrifugation was for 120 h at 25,000 rpm at 20°C in a Beckman SW50.1 rotor. The Ap values in the figure denote the increase in buoyant density calculated from refractive index 32P. measurements. Symbols: 0---0, 3H; *
392
MAGNUSSON AND NILSSON
DISCUSSION DNA synthesis in eucaryotic systems has been studied in detail with cells infected with papovaviruses. The viral genome which consists of a double-stranded circular DNA molecule is duplicated via a replicative intermediate in which DNA synthesis is initiated at a specific site and then continues bidirectionally until the two replication forks meet at a site opposite from the origin of replication (3, 4). The use of subcellular systems, isolated nuclei, and nucleoprotein preparations has led to further understanding of how DNA is synthesized at each replication fork (5, 6, 15, 17). In the present study we have used isolated nuclei and analyzed whether the replication forks move in the same way as during polyoma DNA synthesis in intact cells. We asked the following questions. (i) Does DNA synthesis in vitro terminate at the same site as the in vivo process? (ii) How are the replication forks distributed around the genome? (iii) What is the maximal length of a DNA chain synthesized by an individual replication fork, and does the synthesis occur at a constant rate? (iv) Is there any indication of initiation of new rounds of DNA synthesis? (i) It is clear that replicative intermediates that were successfully completed and transformed into closed circular molecules were terminated at the normal termination site (Fig. 1). The labeling pattern also shows that replication forks approach the termination site from both directions. After 10 min of incubation, in vitro radioactive label was detected only very close to the termination point. After 30 min of reaction, however, some molecules which at the start of the reaction had replication forks as far as 1,000 nucleotides from the termination point were successfully completed. Since only form I DNA was measured in this experiment, it is possible that even longer DNA chains were synthesized in vitro, but that those molecules had not yet been transformed into form I DNA. (ii) The distribution and movement of replication forks were analyzed by labeling the DNA radioactively during the incubation of the nuclei and then isolating the total viral DNA. Of the radioactivity more than 90% was present in replicative intermediates. The viral DNA was then digested with the restriction endonuclease HpaII, and the radioactivity was measured in the separated restriction fragments. By knowing the size and the location of the fragments in the genome, a measure of the distribution of replication forks should be obtained. It appears, however, that when replicative intermediates are digested with restriction enzyme, part of the DNA does not behave as normal linear frag-
J. VIROL.
ments. By using electrophoresis to separate the DNA fragments, those that contain a replication fork are lost. The quantitation in this type of experiment is therefore uncertain. Nevertheless, it is obvious from the results presented in Tables 1 and 2 that replication forks are rather evenly distributed around the genome. One exception was the incorporation of label into HpaIl-5 sequences that was quite high early during the reaction, but later dropped to normal values. These results are in accord with earlier observations (2; G. Bjursell, unpublished data) that replicative intermediates are evenly distributed between different stages of replication. However, different conclusions were reached from studies of simian virus 40 DNA replication, where a preponderance of "late" replicative intermediates was found (18). (iii) The movement and inactivation of replication forks were analyzed in an experiment in which DNA was density labeled with bromodeoxyuridine triphosphate throughout the reaction and, in addition, radioactively labeled during the last part of the reaction. This protocol was designed to analyze DNA synthesized at replication points that remained active throughout the reaction. After digestion with HpaII, restriction fragments which were of full length (Fig. 5) were isolated and analyzed by centrifugation in alkaline cesium sulfate gradients. The result, depicted in Fig. 6, shows that the density of the radioactively labeled DNA chains increased linearly with time, indicating that replication forks that remain active move with a constant speed throughout the reaction. The presence of lighter-than-average material in the gradients (Fig. 6B and D), consisting of full-length DNA chains with one light and one heavy part, might indicate that some replication forks either can pause and then resume synthesis or move quite slowly. The finding that HpaII-3 (Fig. 6B) can be completely synthesized in vitro again shows that replication forks can synthesize chains of at least 1,000 nucleotides in the nuclei (compare Fig. 1). (iv) The last question concerning initiation of new rounds of replication remains essentially unanswered. Initiation might occur, but the data are inconclusive. The high level of DNA synthesis that occurs close to the origin of replication early during the reaction (Table 2) could represent synthesis by replication forks initiated during the early part of the reaction. If this is true the rate of initiation then seems to drop rather quickly during the in vitro reaction. The information presented in Fig. 6 and Table 2 shows that even as late as between 20 and 30 min after the start of the incubation, sequences in HpaII5 are synthesized. From the same experiment it
VOL. 32, 1979
is clear that the synthesis of the DNA in HpaII5 takes less than 20 min, so the synthesis seen in HpaII-5 later than 20 min after the start of the reaction might in fact have been initiated during the first 10 to 15 min. A trivial explanation to this finding that we cannot rule out is that the replication forks were initiated in vivo and had synthesized DNA chains a few nucleotides long at the time the nuclei were isolated. In this case, however, we also have to assume that the replication forks pause for a substantial time period before they complete the synthesis of DNA in
HpaII-5. ACKNOWLEDGMENTS This work was supported by grants from the Swedish Cancer Society and Magnus Bergvall's Foundation. LITERATURE CITED 1. Baldwin, R. L., and E. M. Shooter. 1963. The alkaline transition of BU-containing DNA and its bearing on the replication of DNA. J. Mol. Biol. 7:511-526. 2. Bjursell, G., and G. Magnusson. 1976. Replication of polyoma DNA. Accumulation of early replicative intermediates during hydroxyurea inhibition. Virology 74: 249-251. 3. Crawford, L. V. C., C. Syrett, and A. Wilde. 1973. The replication of polyoma DNA. J. Gen. Virol. 21:515-521. 4. Danna, K. J., and D. Nathans. 1972. Bidirectional replication of Simian virus 40 DNA. Proc. Natl. Acad. Sci. U.S.A. 69:3097-3100. 5. Edenberg, H. J., M. A. Waqar, and J. A. Huberman. 1976;. Subnuclear systems for synthesis of Simian virus 40 DNA in titro. Proc. Natl. Acad. Sci. U.S.A. 73:43924396. 6. Eliasson, R., and P. Reichard. 1978. Primase initiates Okazaki pieces during polyoma DNA synthesis. Nature (London) 272:184-185. 7. Griffin, B. E., M. Fried, and A. Cowie. 1974. Polyoma DNA. A physical map. Proc. Natl. Acad. Sci. U.S.A. 71: 2077-2081.
POLYOMA DNA REPLICATION
393
8. Hirt, B. 1966. Evidence for semiconservative replication of circular polyoma DNA. Proc. Natl. Acad. Sci. U.S.A. 55:997-1004. 9. Hirt, B. 1967. Selective extraction of polyoma DNA from infected mouse cell cultures. J. Mol. Biol. 26:365-369. 10. Inman, R. B., and R. L. Baldwin. 1964. Helix-random coil transitions in DNA homopolymer pairs. J. Mol. Biol. 8:452-469. 11. Magnusson, G. 1973. Hydroxyurea-induced accumulation of short fragments during polyoma DNA replication. I. Characterization of fragments. J. Virol. 12:600608. 12. Magnusson, G., R. Craig, M. Narkhammar, P. Reichard, M. Staub, and H. Warner. 1974. Replication of polyoma DNA: effects of hydroxyurea and arabinosyl cytosine. Cold Spring Harbor Symp. Quant. Biol. 39: 227-233. 13. Magnusson, G., E.-L. Winnacker, R. Eliasson, and P. Reichard. 1972. Replication of polyoma DNA in isolated nuclei. II. Evidence for semi-conservative replication. J. Mol. Biol. 72:539-552. 14. Peacock, A. C., and C. W. Dingman. 1967. Resolution of multiple RNA species by polyacrylamide gel electrophoresis. Biochemistry 6:1818-1827. 15. Pigiet, V., R. Eliasson, and P. Reichard. 1974. Replication of polyoma DNA in isolated nuclei. III. The nucleotide sequence at the RNA-DNA junction of nascent strands. J. Mol. Biol. 84:197-216. 16. Sharp, P. A., B. Sugden, and J. Sambrook. 1973. Detection of two restriction endonuclease activities in Haemophilus parainfluenza using analytical agaroseethidium bromide electrophoresis. Biochemistry 12: 3055-3063. 17. Su, R. T., and M. L DePamphilis. 1978. Simian virus 40 DNA replication in isolated replicating viral chromosomes. J. Virol. 28:53-65. 18. Tapper, D. P., and M. L. De Pamphilis. 1978. Discontinous DNA replication: accumulation of simian virus 40 DNA at specific stages of replication. J. Mol. Biol. 120:401-422. 19. Winnacker, E.-L., G. Magnusson, and P. Reichard. 1972. Replication of polyoma DNA in isolated nuclei. I. Characterization of the system from mouse fibroblast 3'T'(i cells. ,J. Mol. Biol. 72:523-527.
