Proc. Natl. Acad. Scd. USA
Vol. 73, No. 12, pp. 4392-4396, December 1976 Biochemistry
Subnuclear systems for synthesis of simian virus 40 DNA in vitro (DNA replication/chromatin)
HOWARD J. EDENBERG,* M. ANWAR WAQAR,t AND JOEL A. HUBERMAN*t * Department of Biology, Massachusetts, Institute of Technology, Cambridge, Mass. 02139; and t Department of Medical Viral Oncology, Roswell Park Memorial Institute, Buffalo, New York 14263
Communicated by John M. Buchanan, September 15,1976 We have developed two subnuclear systems ABSTRACT for synthesis of DNA of simian virus 40 in vitro. We prepare chromatin from infected cells by the method of Hancock [(1974) J. Mol. BioL 86, 649-6631; these "chromatin bodies" can be disrupted and large debris can be pelleted, leaving a supernatant ("soluble system"). Both chromatin bodies and the soluble system incorporate deoxyribonucleoside triphosphates into nucleoprotein complexes that contain simian virus 40 DNA. The DNA labeled in short pulses sediments in neutral sucrose gradients slightly faster than mature simian virus 40 DNA, as expected for replicating intermediate. When rebanded in alkaline sucrose gradients, about half of the radioactivity is found in short strands (200-300 nucleotides) and half in longer strands (up to full viral size) When these systems are supplemented with a cytoplasmic preparation from HeLa cells, synthesis is stimulated about 5-fold, and the short strands are converted into strands of up to full viral length as well as into covalently closed circles. These subnuclear DNA-replicating systems should be useful for biochemical fractionation and characterization of some of the proteins required for DNA replication.
Eukaryotic DNA replication has been the object of much study (for review see ref. 1), but many questions about the structure of replication forks and the proteins and cofactors that carry out replication remain unanswered. This is in large part due to the complexity of eukaryotic cells, and although the use of isolated nuclei has been helpful in studies of replication, such nuclear systems are still both complex and difficult to dissociate into simpler components. For this reason, we have developed two subnuclear DNAsynthesizing systems from simian virus 40 (SV40)-infected monkey cells. There are several advantages to studying the replication of papovaviruses like SV40. The chromosome is small and can be manipulated and analyzed far more easily than the enormous chromosomes of the host cell. It is therefore possible to more fully characterize the products of synthesis in vitro and compare them to the products of replication in vivo. And, with the exception of a viral protein involved in the initiation of replication (2-5), the replication of SV40 DNA requires host proteins and resembles the host chromosomal replication in nearly every way yet noted (1). Thus our subnuclear DNA-synthesizing systems should be excellent probes of cellular DNA replication. In this paper we describe the two subnuclear DNA-synthesizing systems we have developed: one using "chromatin bodies" that retain some nuclear structure although they lack nuclear membranes, and a "soluble system" made from the chromatin bodies by physical disruption and pelleting of the large debris. We present our initial characterization of these systems, and data from them, which in the simplest interpretation supports a model of replication discontinuous on only one side of the replication fork. The soluble systems we describe Abbreviations: SV40, simian virus 40; buffer A, 0.1 M sucrose, 0.2 mM KH2PO4 (pH 7.5); buffer B, 10 mM each of Tris, EDTA, NaCl (pH 8).
should make it possible to identify and isolate proteins and other factors involved in eukaryotic DNA replication. MATERIALS AND METHODS Cells and Infection. BSC-1 cells were allowed to become confluent in 100-mm plastic petri dishes (Falcon Plastics) in Dulbecco's modified Eagle's medium supplemented with 5% fetal calf serum (both from Grand Island Biological Co.). The medium was removed and 1 ml of fresh medium containing SV40 (grown at low multiplicity of infection from plaquepurified virus) at a multiplicity of 5-10 plaque-forming units per cell was added. After 1 hr at 370 in 5% CO2, 9 ml of fresh medium was added and incubation was continued at 370 in 5%
CO2.
Preparation of In Vitro Systems. At 38-44 hr after infection, some cells were labeled with [3H]thymidine (20-100 /.Ci/ml, 40-60 Ci/mmol, New England Nuclear) for the desired time. The dishes were rinsed two to three times with ice-cold phosphate-buffered saline (0.137 M NaCl, 2.7 mM KCI, 8 mM Na2HPO4, 1.5 mM KH2PO4). The cells were resuspended with a rubber policeman, mixed with unlabeled cells, and pelleted for 4 min at 500 X g, at 4°. Chromatin was prepared by either of two slightly different modifications of the Hancock (6) procedure; the one used in most experiments is given here. All procedures were carried out at 0-4°. The cell pellet was resuspended in 15 ml of buffer A (0.1 M sucrose, 0.2 mM KH2PO4, pH 7.5) and centrifuged at 800 X g for 8 min. This was repeated and the pellet was resuspended in 2-3 ml of buffer A, adjusted to 0.2% Nonidet P-40 (Shell Chemicals) by addition of 0.5% Nonidet P-40 in buffer A, and kept on ice for 10 min. Buffer A was added to 15 ml total, and the material was centrifuged at 800 X g for 8 min. The pellet was resuspended in 15 ml of buffer A and pelleted twice more; dithiothreitol was added to 2 mM. The material in the final pellet had the appearance of greatly swollen nuclei; however, electron micrographs of thin sections through the pelleted material (taken by Elaine V. Lenk of the MIT Electron Microscope Facility) revealed the absence of nuclear membranes, as reported by Hancock (6). Consequently, this material will be referred to as "chromatin bodies." Chromatin bodies were further disrupted in a Dounce homogenizer (Kontes Glass) with 25-100 strokes of the tight (B) pestle. Any unbroken chromatin bodies and large debris were pelleted twice (3 min each at 800 X g). (In some cases the solution was very viscous and more vigorous centrifugation was required.) The supernatant contained no nuclei or debris visible in the phase contrast microscope; this supernatant is referred to as the "soluble" system. In Vitro Synthesis. The chromatin bodies or soluble system were diluted as desired with buffer A and reaction components were added as concentrated solutions. The standard reaction mixture contained sucrose (a 0.08 M) and KH2PO4 (a 0.16 mM) from buffer A as well as 70 mM KCl; 7 mM MgCl2; 10
4392
Proc. Natl. Acad. Sci. USA 73 (1976)
Biochemistry: Edenberg et al. mM N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid (Hepes); 100 ,uM each of three unlabeled deoxyribonucleoside triphosphates; the fourth dNTP, A-32P-labeled at 6-15 M
(10-93 ,Ci/ml, New England Nuclear); 5 mM rATP; and 0.35 mM each rGTP, rCTP, and rUTP: all at pH 7.8. Incubation was at 37'. Chases were performed by adding a 60 to 100-fold excess of the appropriate dNTP. For analysis of chromatin, the reaction was stopped by adding EDTA to 40 mM and layering the entire mixture on a neutral sucrose gradient (details in legend of Fig. 1). Analysis of DNA Synthesized In Vitro. Reactions were stopped by adjusting to 0.6% sodium dodecyl sulfate, 10 mM EDTA; SV40 DNA was selectively extracted by the method of Hirt (7). Aliquots of the Hirt supernatant were spotted onto filter papers and washed at least three times with cold 1 M HC1, once each with ethanol and acetone, and then dried. Radioactivity was measured in a liquid scintillation counter. For size analysis of DNA strands, the Hirt supernatant was
first sedimented through a neutral sucrose gradient (details in figure legends). The peak fractions of viral-sized DNA were pooled, dialyzed against buffer B (10 mM each of Tris, Na2EDTA, NaCI, pH 8), then concentrated by dialysis against 25% Carbowax 6000 (Union Carbide) in buffer B. The material was adjusted to 0.2 M NaOH and sedimented through an alkaline sucrose gradient (10-30% sucrose in 0.2 M NaOH, 0.8 M NaCI) for 6-7 hr at 45,000 rpm, 22° in the SW50.1 rotor. For analysis of superhelix density, viral-sized DNA pooled and dialyzed as above was adjusted to a final volume of 7.5 ml in buffer B and a final refractive index of 1.388 (at 220) with CsCl; 270 ,g/ml of ethidum bromide was present. This material was centrifuged for at least 40 hr at 37,000 rpm in the Spinco 50 Ti rotor. For density analysis after incorporation of BrdUTP, viralsized DNA pooled and dialyzed as above was mixed with 0.8 ml of 0.5 M KH2PO4, 0.45 M KOH, and then adjusted with 'Ao strength buffer B and CsCl to 8 ml at a final refractive index of 1.409 (220). Centrifugation was carried out for at least 40 hr at 37,000 rpm in the Spinco 50 Ti rotor. Fractions from gradients, or aliquots thereof, were spotted onto filter papers and processed as described above. Agarose gels (1.5%, 14 cm long) were run in 40 mM Tris, 1 mM EDTA, 20 mM Na acetate, 0.5 mg/liter of ethidium bromide, at 60-80 V; the DNA was visualized by fluorescence of the ethidium bromide. Radioactivity in gel slices was measured in Scintiverse (Fisher Scientific). Digestion with the restriction
endonuclease HindIII (Miles Laboratories) was carried out overnight at 370 in 50 mM NaCI, 6 mM Tris-HCI, 6 mM MgCl2, 1 mg/ml of bovine serum albumin (pH 7.5), with an excess of enzyme.
Cytoplasmic Extract. A cytoplasmic extract from HeLa cells, prepared by homogenizing a swollen cell pellet and centrifuging for 1 hr at 30,000 X g, then 1 hr at 100,000 X g, was the generous gift of J. K. Fraser. The protein concentration of the final supernatant is about 20 mg/ml (J. K. Fraser, Ph.D dissertation, MIT, 1976). This extract was dialyzed against buffer A.
RESULTS Preparation of Subnuclear Systems. Isolation of chromatin from SV40-infected monkey cells in solutions of very low ionic strength allowed both the chromatin bodies and the soluble system prepared from them to continue synthesizing SV40DNA in vitro. The chromatin bodies synthesize 2- to 7-fold more DNA (normalized to the amount of in vivo prelabel) than the soluble system; the nature of the DNA synthesized is essentially
4393
Table 1. Triphosphate dependence of incorporation into chromatin bodies in vitro Standard reaction mixture* t= 0 (on ice) t= 15 min
-dTTP
-dTTP, dGTP, dCTP -rATP
-rATP, rGTP, rCTP, rUTP -rGTP, rCTP, rUTP
32P/3H
% Controlt
0.14 1.20 0.13 0.08 0.62 0.56 1.01
0 100 0 0 45 40 82
* The standard reaction mixture contained 70 mM KCl; 86 mM sucrose; 7 mM MgCl2; 5 mM rATP; 0.35 mM each rGTP, rCTP, and rUTP; 100 AM each dGTP, dCTP, and dTTP; and 10 AM [32P]dATP (2 Ci/mmol). All changes from standard conditions are noted. Incubation was at 370 for 15 min. t The corrected percent incorporation was determined after subtracting the zero time value from each reaction.
the same, as described below. The Dounce homogenization used to prepare the soluble system from chromatin bodies is capable of breaking a variable fraction of the SV40 chromosomes. Since replicating chromosomes would be expected to be more sensitive to shear than mature SV40 chromosomes, excessive shearing might be one factor limiting the extent of synthesis in the soluble system. Reaction Conditions. Optimum conditions for DNA synthesis in vitro were similar for both systems. The optimum for KCI is about 60-80 mM; concentrations of 70 mM are routinely used. The magnesium optimum is dependent upon the concentration of ribo- and deoxyribonucleoside triphosphates in the reaction mixture; in the presence of 5 mM rATP and 0.35 mM each rGTP, rCTP, and rUTP the optimum concentration of MgCl2 was relatively sharp, centering on 7 mM. The presence of the four rNTPs enhanced synthesis at 7 mM MgCl2 (Table 1); even when MgCl2 was lowered to compensate for the absence of rNTPs, the chromatin bodies synthesized less DNA than under standard conditions (56%). The soluble system,
however, synthesized more DNA at 1.5 mM MgCl2 in the absence of the four rNTPs than in the standard mixture (140%). This is one of the few differences noted between these systems, and its significance is not clear. The role of rNTPs in DNA synthesis in vitro is not yet understood. Both subnuclear systems are completely dependent on the addition of dNTPs (Table 1). The standard reaction mixture was used for both the soluble system and the chromatin bodies, even when cytoplasmic extracts were present. These extracts contain dNTPs, as indicated by the loss of total dependence upon added dNTPs (and probably contain rNTPs as well). This complicates estimating the degree of stimulation by cytoplasmic factors, since we have not yet determined the extent of dilution of the [32P]dNTPs used for labeling. These systems lose activity during incubation at 370 in the absence of dNTPs and cytoplasm; the extent of loss is variable (70-100% in 15 min) and its cause was not explored. Kinetics of DNA Synthesis. In the absence of cytoplasmic extract the incorporation is linear with time for the first few minutes, and then decreases in rate. Incorporation by chromatin bodies generally ceases (or slows dramatically) by about 10-15 min; the soluble system generally levels sooner, but in some experiments it too has continued for over 15 min. The presence of cytoplasmic extract increases the initial rate of DNA synthesis, and in addition seems to allow the reaction to proceed for a longer period, as much as 30 min or more. (The incorpo-
Biochemistry: Edenberg et al.
4394
30
Proc. Natl. Acad. Sci. USA 73 (1976) 5
2)
-
a
jq
I25-
ba:;20-
lo
o -
02
x
5.0
// A
0
0.5
CL
}T
a-
0 '5
1
X
A
'Q x
4r
5
10
a-
(j
5
_0
5
'0N -
iL
I1
10
20 25 -30 15 Fraction number FIG. 1. Sedimentation of soluble system components through a neutral sucrose gradient. A soluble system (from cells labeled with [-l]thymidine for 15 min in vivo) was incubated at 370 in the standard reaction mixture containing [a-32P]dTTP (22.5 gCi/ml, 17.9 Ci/mmol) in the absence of cytoplasm. After 45 min, EDTA was added to 38 mM and the reaction mixture was cooled on ice, then layered over a 9.5-ml gradient (10-40% sucrose in 0.2 mM NaH2PO4, pH 7.5, over a 1-ml shelf of 80% sucrose in the same buffer) in a polyallomer tube. Centrifugation was in the SW41 rotor at 40,000 rpm for 120 min at 4°. Aliquots from each fraction were processed as described. (-) :fH (in vivo); (0) 32P (in vitro).
ration at each time point is given by the 32P/3H ratios in Figs. 3 and 4.)
Fractionation of the Soluble System on a Sucrose Gradient. The sedimentation profile of the protein-DNA complexes responsible for synthesis in vitro in the soluble system is shown in Fig. 1. A variable portion of the prelabeled material, containing both cellular and unit-length viral DNA, sedimented very rapidly. In Fig. 1 the major part of the 3H prelabel forms a sharp peak with a leading shoulder. DNA extracted from the peak sediments (>80%) as form I in alkaline sucrose gradients, while DNA extracted from the shoulder sediments as SV40 replication intermediates in neutral and alkaline gradients. Thus, under our conditions, as under other conditions (8-11), replicating 5V40 chromosomes sediment faster than mature SV40 chromosomes. The 32p label incorporated in vitro sediments predominantly with the leading shoulder of prelabel which contains replication intermediates, as expected for a continuation of DNA synthesis. Characteristics of DNA Made In Vitro. Most of the DNA made in vitro sediments more rapidly than form I DNA in neutral sucrose gradients (Fig. 2), as would be expected for incorporation into replication intermediates. After longer incubations (>15 min), especially in the presence of cytoplasmic extract, the peak of DNA made in vitro sediments noticeably more slowly (Fig. 2), much of the DNA at about the same rate as form I DNA. In the absence of cytoplasm some DNA may sediment more slowly than form I, and some remains faster sedimenting even at 30 min. This sedimentation behavior may be related to the closer approach to completion of synthesis achieved in the presence of cytoplasm and lack of joining of synthesized fragments in its absence (see below). In order to determine which portions of the SV40 DNA molecule were being replicated in vitro, viral-sized DNA synthesized by chromatin bodies in the presence of cytoplasmic extract was digested with the restriction enzyme HindIII and electrophoresed on an agarose gel. The fluorescence pattern of the chromatin body preparation was identical to that of purified SV40 form I, digested in parallel. The uniform prelabel in vivo
Fraction number FIG. 2. Sedimentation behavior of DNA synthesized in vitro. A chromatin body preparation (from cells labeled 14.5 min in vivo with [3H]thymidine) was incubated at 370 in a standard reaction mixture ([32P]dATP at 23 MACi/ml, 1.8 Ci/mmol) in which 48% of the volume was the cytoplasmic extract of HeLa cells described. At the times indicated aliquots were removed and viral DNA extracted (7), layered on neutral sucrose gradients (10-45% sucrose in 0.1 M Tris, 0.05 M EDTA, 0.85 M NaCl, pH8), and centrifuged at 45,000 rpm for 225 min at 220 in the SW 50.1 rotor. Aliquots of each fraction were acid-precipitated as described. (a) 0.6 min incorporation in vitro; (b) 15.1 min incorporation in vitro. (0) 3H in vivo label; (0) 32p in vitro label.
corresponded exactly with the fluorescent bands, as did most of the in vitro radioactivity. (Some of the in vitro label trailed the bands slightly, particularly near fragment A; this could conceivably be due to the presence of replication forks.) The ratio of in vitro radioactivity to the uniform prelabel was highest (9.7) in the A fragment (around the terminus of replication) and lowest (3.6) in the B + C doublet (around the origin). Such a pattern is consistent with a relative deficiency (or lack) of initiation of new rounds of replication. Size of Strands Made In Vitto. Viral-sized DNA pooled from neutral sucrose gradients was sedimented through alkaline sucrose gradients to determine the size of the DNA strands made in vitro. At all times from 0.5 to 15 min of labeling in vitro, the DNA made in the absence of cytoplasmic extract by both chromatin bodies and the soluble system (Fig. 3) contains approximately half long strands (up to full viral length) and half short strands (about 200-30 nucleotides, data not shown). This pattern is identical in both systems and is maintained while the 32P/3H ratio is increasing (i.e., while DNA synthesis is continuing). In neither case is there significant chasing of short strands into longer strands (Fig. 3). The addition of a cytoplasmic extract to either system dramatically changes the results. Both the chromatin bodies and the soluble system (Fig. 4), when supplemented with the extracts, show a peak of short strands after brief incubations in vitro (up to 2-4 min), but with longer incubations this peak becomes relatively smaller while the peak of longer strands becomes larger. Again, the pattern is identical in both systems, except that the rate at which the peak of short strands is ob-
Proc. Natl. Acad. Sci. USA 73 (1976)
Biochemistry: Edenberg et al.
4395
partially substituted DNA. This is consistent with the even distribution of radioactivity between Okazaki fragments and longer strands, and is similar to data reported for another in vitro system (12). As expected for DNA made semiconservatively, sonication of the DNA to a size about l' that of SV40 shifts most of the in vitro label to nearly fully heavy density in alkaline CsCl (data not shown). DISCUSSION 8n
1
m in
U4
x 6
P/:
P/H
A
15
5
25
35
45
5
15
25
35
45
Fraction number
FIG.
3.
Size distribution of DNA synthesized in vitro by the
soluble system (without cytoplasm). A soluble system (from cells la-
beled with
[2H]thymidine for 9.8 mn
in viva) was incubated at 370
[12P]dATP (72 mCi/ml
in the standard reaction mixture containing
at 9.2 Ci/mmol). At the times noted, aliquots were removed and viral
DNA was extracted and sedimented through neutral sucrose gradients as described in the legend of Fig. 2. The peak of viral-sized DNA was
pooled, dialyzed, and concentrated as described. The samples were adjusted to 0.2 M NaOH, and after 30
mmn layered on 10-30%6 alkaline
sucrose gradients and centrifuged at 45,000 rpm for 6 hr at 220
in the
SW 50.1 rotor. The duration of each pulse and the 32P/3H ratio (rep-
resenting the amount of DNA synthesized) are shown.
(-) :H
(in
viva); (0) 32p (in vitro).
scured is greater in the chromatin bodies, consistent with their
more rapid DNA synthesis. In the presence of cytoplasmic extract the short strands chase completely into longer DNA (Fig. 4) without further
incorporation of label.
Completion of Molecules. In the absence of cytoplasmic
extract, virtually no radioactive dNTPs are incorporated in vitro into covalently closed molecules (form I). When cytoplasmic factors are present, dNTPs are incorporated into form I. In general, chromatin bodies incorporate a higher proportion of the in vitro dNTPs into form I than does the soluble system; this may be due to damage sustained by the SV4O chromosomes during preparation of the soluble system, as previously noted. In one 30-mma incorporation with chromatin bodies and cytoplasmic extract at least 25% of the in vitro label was converted to covalently closed molecules, whereas a parallel incubation without cytoplasmic extract incorporated less than 1% into form I.
In an
experiment with the soluble system, in which care was
taken to minimize shear, 3-4% of the in vitro label was converted to
form
I.
Semiconservative DNA Synthesis. When incubation in vitro is carried out in the presence of BrdUTP instead of dTTP, the DNA made by chromatin bodies (without cytoplasmic extract) is denser than unsubstituted DNA. When centrifuged to equilibrium in alkaline CsCl, about half of the radioactivity is associated with nearly fully substituted DNA and about half with
In this paper we report methods for preparation of two related subnuclear systems for SV40 DNA replication, as well as preliminary characterization of the two systems. The soluble system we have developed offers many advantages over nuclear in vitro systems, since it should prove far more amenable to fractionation. DNA synthesis in vitro in nuclear sonicates from bovine lymphocytes has been reported (13), but in that system the characteristics of thereaction were unusual. Recently another subnuclear DNA-synthesizing system similar to ours has been independently developed (14). The DNA synthesis in our subnuclear systems appears to be a continuation of the replication process started in vivo. Incorporation during short incubations is into nucleoprotein complexes cosedimenting with those that contain replication intermediate DNA of SV40. The DNA extracted from short incubations cosediments with replication intermediate DNA of SV40. After longer incubations in the presence of cytoplasmic factors a fraction of the label is incorporated into mature (form I) SV40 DNA. DNA synthesis in these systems is semiconser-
vative. One of the interesting characteristics of the DNA synthesized in both our systems, in the absence of cytoplasm, is the fact that regardless of length of incubation or length of chase the synthesized DNA contains half long strands (up to the size of linear viral strands) and half short strands (200-300 nucleotides). In the presence of cytoplasm, the short strands chase into long
strands. These short strands thus resemble the Okazaki fragments that have previously been detected during papovavirus DNA replication in vivo (15, 16) and in nuclear systems (17-25). Other investigators working with nuclear systems have also noticed that cytoplasmic factors stimulate DNA synthesis and joining (18, 20, 23). The equal distribution of radioactivity between long and short strands in the absence of cytoplasm suggests that synthesis of only one of the daughter strands is through Okazaki fragments (semidiscontinuous synthesis). The fact that this is seen in both systems despite differences in the rate and extent of synthesis argues that such a distribution is not due to a coincidental balancing of the rates of synthesis and joining of Okazaki fragments. Francke and coworkers (17-19) have concluded that synthesis in their nuclear in vitro polyoma DNA-synthesizing system is semidiscontinuous; Hershey and Taylor have come to the same conclusion in a different in vitro synthesizing system (12). We note that the equal distribution of label into long and short strands could be consistent with totally discontinuous synthesis provided that the joining of half of the short strands is both very rapid (even in vitro) and not dependent upon cytoplasmic factors. The fact that our soluble system can be sedimented through neutral sucrose gradients to partially purify and separate replicating SV40 chromosomes, and that the replicating SV40 chromosomes can continue synthesis in vitro (details to be published elsewhere), suggests that this system provides an excellent starting point from which it should be possible to purify and characterize the enzymes that participate in DNA
4396
Biochemistry: Edenberg et al.
Proc. Natl. Acad. Sci. USA 73 (1976)
7 x
-3
a-0
0)
2 ~
A, I0~~~ ,
C 2
54- 4 -3
-2
5
15
25
35
0-% 4 0
0
M
Fraction number FIG. 4. Size distribution of DNA synthesized in vitro by the soluble system in the presence of cytoplasmic factors. A soluble system (from cells labeled with [3H]thymidine for 14.5 min in vivo) was incubated at 370 in a standard reaction mixture ([32PjdATP at 23 pCi/ml, 1.8 Ci/mmol) in which 48% of the volume was the cytoplasmic preparation of HeLa cells described. Aliquots were processed as described in the legend of Fig. 3. (-) :H (in vivo); (0) 32p (in vitro).
synthesis, as well as to characterize the proteins that bind to the replicating chromosomes and the structure of the replicating chromosomes. We thank Michael Reszka and Daniel Perlman for helping, and Elaine V. Lenk for the electron microscopy of chromatin bodies. This research was supported by grants from the National Science Foundation and National Institutes of Health. 1.
2. 3. 4. 5.
Edenberg, H. J. & Huberman, J. A. (1975) Annu. Rev. Genet. 9,245-284. Tegtmeyer, P. (1972), J. Virol. 10, 591-598. Francke, B. & Eckhart, W. (1973) Virology 55, 127-135. Francke, B. & Hunter, T. (1974) J. Virol. 13,241-243. Chou, J. Y., Avila, J. & Martin, R. G. (1974) J. Virol. 14, 116-
124. 6. Hancock, R. (1974) J. Mol. Biol. 86, 649-663. 7. Hirt, B. (1967) J. Mol. Biol. 26,365-369. 8. White, M. & Eason, R. (1971) J. Virol. 8,363-371. 9. Hall, M. R., Meinke, W. & Goldstein, D. A. (1973) J. Virol. 12, 901-908. 10. Sen, A. & Levine, A. J. (1974) Nature 249,343-344. 11. Cremisi, C., Pignatti, P. F., Croissant, 0. & Yaniv, M. (1976) J.
Virol. 17,204-211. 12. Hershey, H. V. & Taylor, J. H. (1974) Exp. Cell Res. 85, 7988. 13. Thompson, L. R. & Mueller, G. C. (1975) Biochim. Biophys. Acta 14.
378,344-353. Su, R. T. & DePamphilis, M. L. (1976) Fed. Proc. 35, 1376
abstr. 15. Fareed, G. C. &. Salzman, N. P. (1972) Nature New Biol. 238, 274-277. 16. Laipis, P. J. and Levine, A. J. (1973) Virology 56,580-594. 17. Francke, B. & Hunter, T. (1974) J. Mol. Biol. 83,99-121. 18. Francke, B. & Hunter, T. (1975) J. Virol. 15,97-107. 19. Francke, B. & Vogt, M. (1975) Cell 5,205-211. 20. Otto, B. & Reichard, P. (1975) J. Virol. 15, 259-267. 21. Magnusson, G., Pigiet, V., Winnacker, E. L., Abrams, R. & Reichard, P. (1973) Proc. Natl. Acad. Sci. USA 70,412-415. 22. Qasba, P. K. (1974) Biochem. Biophys. Res. Commun. 60, 1338-1344. 23. DePamphilis, M. L. & Berg, P. (1975) J. Biol. Chem. 250, 4348-4354. 24. Magnusson, G. (1973) J. Virol. 12, 609-615. 25. Qasba, P. K. (1974) Proc. Natl. Acad. Sci. USA 71, 10451049.
Vol. 73, No. 12, pp. 4392-4396, December 1976 Biochemistry
Subnuclear systems for synthesis of simian virus 40 DNA in vitro (DNA replication/chromatin)
HOWARD J. EDENBERG,* M. ANWAR WAQAR,t AND JOEL A. HUBERMAN*t * Department of Biology, Massachusetts, Institute of Technology, Cambridge, Mass. 02139; and t Department of Medical Viral Oncology, Roswell Park Memorial Institute, Buffalo, New York 14263
Communicated by John M. Buchanan, September 15,1976 We have developed two subnuclear systems ABSTRACT for synthesis of DNA of simian virus 40 in vitro. We prepare chromatin from infected cells by the method of Hancock [(1974) J. Mol. BioL 86, 649-6631; these "chromatin bodies" can be disrupted and large debris can be pelleted, leaving a supernatant ("soluble system"). Both chromatin bodies and the soluble system incorporate deoxyribonucleoside triphosphates into nucleoprotein complexes that contain simian virus 40 DNA. The DNA labeled in short pulses sediments in neutral sucrose gradients slightly faster than mature simian virus 40 DNA, as expected for replicating intermediate. When rebanded in alkaline sucrose gradients, about half of the radioactivity is found in short strands (200-300 nucleotides) and half in longer strands (up to full viral size) When these systems are supplemented with a cytoplasmic preparation from HeLa cells, synthesis is stimulated about 5-fold, and the short strands are converted into strands of up to full viral length as well as into covalently closed circles. These subnuclear DNA-replicating systems should be useful for biochemical fractionation and characterization of some of the proteins required for DNA replication.
Eukaryotic DNA replication has been the object of much study (for review see ref. 1), but many questions about the structure of replication forks and the proteins and cofactors that carry out replication remain unanswered. This is in large part due to the complexity of eukaryotic cells, and although the use of isolated nuclei has been helpful in studies of replication, such nuclear systems are still both complex and difficult to dissociate into simpler components. For this reason, we have developed two subnuclear DNAsynthesizing systems from simian virus 40 (SV40)-infected monkey cells. There are several advantages to studying the replication of papovaviruses like SV40. The chromosome is small and can be manipulated and analyzed far more easily than the enormous chromosomes of the host cell. It is therefore possible to more fully characterize the products of synthesis in vitro and compare them to the products of replication in vivo. And, with the exception of a viral protein involved in the initiation of replication (2-5), the replication of SV40 DNA requires host proteins and resembles the host chromosomal replication in nearly every way yet noted (1). Thus our subnuclear DNA-synthesizing systems should be excellent probes of cellular DNA replication. In this paper we describe the two subnuclear DNA-synthesizing systems we have developed: one using "chromatin bodies" that retain some nuclear structure although they lack nuclear membranes, and a "soluble system" made from the chromatin bodies by physical disruption and pelleting of the large debris. We present our initial characterization of these systems, and data from them, which in the simplest interpretation supports a model of replication discontinuous on only one side of the replication fork. The soluble systems we describe Abbreviations: SV40, simian virus 40; buffer A, 0.1 M sucrose, 0.2 mM KH2PO4 (pH 7.5); buffer B, 10 mM each of Tris, EDTA, NaCl (pH 8).
should make it possible to identify and isolate proteins and other factors involved in eukaryotic DNA replication. MATERIALS AND METHODS Cells and Infection. BSC-1 cells were allowed to become confluent in 100-mm plastic petri dishes (Falcon Plastics) in Dulbecco's modified Eagle's medium supplemented with 5% fetal calf serum (both from Grand Island Biological Co.). The medium was removed and 1 ml of fresh medium containing SV40 (grown at low multiplicity of infection from plaquepurified virus) at a multiplicity of 5-10 plaque-forming units per cell was added. After 1 hr at 370 in 5% CO2, 9 ml of fresh medium was added and incubation was continued at 370 in 5%
CO2.
Preparation of In Vitro Systems. At 38-44 hr after infection, some cells were labeled with [3H]thymidine (20-100 /.Ci/ml, 40-60 Ci/mmol, New England Nuclear) for the desired time. The dishes were rinsed two to three times with ice-cold phosphate-buffered saline (0.137 M NaCl, 2.7 mM KCI, 8 mM Na2HPO4, 1.5 mM KH2PO4). The cells were resuspended with a rubber policeman, mixed with unlabeled cells, and pelleted for 4 min at 500 X g, at 4°. Chromatin was prepared by either of two slightly different modifications of the Hancock (6) procedure; the one used in most experiments is given here. All procedures were carried out at 0-4°. The cell pellet was resuspended in 15 ml of buffer A (0.1 M sucrose, 0.2 mM KH2PO4, pH 7.5) and centrifuged at 800 X g for 8 min. This was repeated and the pellet was resuspended in 2-3 ml of buffer A, adjusted to 0.2% Nonidet P-40 (Shell Chemicals) by addition of 0.5% Nonidet P-40 in buffer A, and kept on ice for 10 min. Buffer A was added to 15 ml total, and the material was centrifuged at 800 X g for 8 min. The pellet was resuspended in 15 ml of buffer A and pelleted twice more; dithiothreitol was added to 2 mM. The material in the final pellet had the appearance of greatly swollen nuclei; however, electron micrographs of thin sections through the pelleted material (taken by Elaine V. Lenk of the MIT Electron Microscope Facility) revealed the absence of nuclear membranes, as reported by Hancock (6). Consequently, this material will be referred to as "chromatin bodies." Chromatin bodies were further disrupted in a Dounce homogenizer (Kontes Glass) with 25-100 strokes of the tight (B) pestle. Any unbroken chromatin bodies and large debris were pelleted twice (3 min each at 800 X g). (In some cases the solution was very viscous and more vigorous centrifugation was required.) The supernatant contained no nuclei or debris visible in the phase contrast microscope; this supernatant is referred to as the "soluble" system. In Vitro Synthesis. The chromatin bodies or soluble system were diluted as desired with buffer A and reaction components were added as concentrated solutions. The standard reaction mixture contained sucrose (a 0.08 M) and KH2PO4 (a 0.16 mM) from buffer A as well as 70 mM KCl; 7 mM MgCl2; 10
4392
Proc. Natl. Acad. Sci. USA 73 (1976)
Biochemistry: Edenberg et al. mM N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid (Hepes); 100 ,uM each of three unlabeled deoxyribonucleoside triphosphates; the fourth dNTP, A-32P-labeled at 6-15 M
(10-93 ,Ci/ml, New England Nuclear); 5 mM rATP; and 0.35 mM each rGTP, rCTP, and rUTP: all at pH 7.8. Incubation was at 37'. Chases were performed by adding a 60 to 100-fold excess of the appropriate dNTP. For analysis of chromatin, the reaction was stopped by adding EDTA to 40 mM and layering the entire mixture on a neutral sucrose gradient (details in legend of Fig. 1). Analysis of DNA Synthesized In Vitro. Reactions were stopped by adjusting to 0.6% sodium dodecyl sulfate, 10 mM EDTA; SV40 DNA was selectively extracted by the method of Hirt (7). Aliquots of the Hirt supernatant were spotted onto filter papers and washed at least three times with cold 1 M HC1, once each with ethanol and acetone, and then dried. Radioactivity was measured in a liquid scintillation counter. For size analysis of DNA strands, the Hirt supernatant was
first sedimented through a neutral sucrose gradient (details in figure legends). The peak fractions of viral-sized DNA were pooled, dialyzed against buffer B (10 mM each of Tris, Na2EDTA, NaCI, pH 8), then concentrated by dialysis against 25% Carbowax 6000 (Union Carbide) in buffer B. The material was adjusted to 0.2 M NaOH and sedimented through an alkaline sucrose gradient (10-30% sucrose in 0.2 M NaOH, 0.8 M NaCI) for 6-7 hr at 45,000 rpm, 22° in the SW50.1 rotor. For analysis of superhelix density, viral-sized DNA pooled and dialyzed as above was adjusted to a final volume of 7.5 ml in buffer B and a final refractive index of 1.388 (at 220) with CsCl; 270 ,g/ml of ethidum bromide was present. This material was centrifuged for at least 40 hr at 37,000 rpm in the Spinco 50 Ti rotor. For density analysis after incorporation of BrdUTP, viralsized DNA pooled and dialyzed as above was mixed with 0.8 ml of 0.5 M KH2PO4, 0.45 M KOH, and then adjusted with 'Ao strength buffer B and CsCl to 8 ml at a final refractive index of 1.409 (220). Centrifugation was carried out for at least 40 hr at 37,000 rpm in the Spinco 50 Ti rotor. Fractions from gradients, or aliquots thereof, were spotted onto filter papers and processed as described above. Agarose gels (1.5%, 14 cm long) were run in 40 mM Tris, 1 mM EDTA, 20 mM Na acetate, 0.5 mg/liter of ethidium bromide, at 60-80 V; the DNA was visualized by fluorescence of the ethidium bromide. Radioactivity in gel slices was measured in Scintiverse (Fisher Scientific). Digestion with the restriction
endonuclease HindIII (Miles Laboratories) was carried out overnight at 370 in 50 mM NaCI, 6 mM Tris-HCI, 6 mM MgCl2, 1 mg/ml of bovine serum albumin (pH 7.5), with an excess of enzyme.
Cytoplasmic Extract. A cytoplasmic extract from HeLa cells, prepared by homogenizing a swollen cell pellet and centrifuging for 1 hr at 30,000 X g, then 1 hr at 100,000 X g, was the generous gift of J. K. Fraser. The protein concentration of the final supernatant is about 20 mg/ml (J. K. Fraser, Ph.D dissertation, MIT, 1976). This extract was dialyzed against buffer A.
RESULTS Preparation of Subnuclear Systems. Isolation of chromatin from SV40-infected monkey cells in solutions of very low ionic strength allowed both the chromatin bodies and the soluble system prepared from them to continue synthesizing SV40DNA in vitro. The chromatin bodies synthesize 2- to 7-fold more DNA (normalized to the amount of in vivo prelabel) than the soluble system; the nature of the DNA synthesized is essentially
4393
Table 1. Triphosphate dependence of incorporation into chromatin bodies in vitro Standard reaction mixture* t= 0 (on ice) t= 15 min
-dTTP
-dTTP, dGTP, dCTP -rATP
-rATP, rGTP, rCTP, rUTP -rGTP, rCTP, rUTP
32P/3H
% Controlt
0.14 1.20 0.13 0.08 0.62 0.56 1.01
0 100 0 0 45 40 82
* The standard reaction mixture contained 70 mM KCl; 86 mM sucrose; 7 mM MgCl2; 5 mM rATP; 0.35 mM each rGTP, rCTP, and rUTP; 100 AM each dGTP, dCTP, and dTTP; and 10 AM [32P]dATP (2 Ci/mmol). All changes from standard conditions are noted. Incubation was at 370 for 15 min. t The corrected percent incorporation was determined after subtracting the zero time value from each reaction.
the same, as described below. The Dounce homogenization used to prepare the soluble system from chromatin bodies is capable of breaking a variable fraction of the SV40 chromosomes. Since replicating chromosomes would be expected to be more sensitive to shear than mature SV40 chromosomes, excessive shearing might be one factor limiting the extent of synthesis in the soluble system. Reaction Conditions. Optimum conditions for DNA synthesis in vitro were similar for both systems. The optimum for KCI is about 60-80 mM; concentrations of 70 mM are routinely used. The magnesium optimum is dependent upon the concentration of ribo- and deoxyribonucleoside triphosphates in the reaction mixture; in the presence of 5 mM rATP and 0.35 mM each rGTP, rCTP, and rUTP the optimum concentration of MgCl2 was relatively sharp, centering on 7 mM. The presence of the four rNTPs enhanced synthesis at 7 mM MgCl2 (Table 1); even when MgCl2 was lowered to compensate for the absence of rNTPs, the chromatin bodies synthesized less DNA than under standard conditions (56%). The soluble system,
however, synthesized more DNA at 1.5 mM MgCl2 in the absence of the four rNTPs than in the standard mixture (140%). This is one of the few differences noted between these systems, and its significance is not clear. The role of rNTPs in DNA synthesis in vitro is not yet understood. Both subnuclear systems are completely dependent on the addition of dNTPs (Table 1). The standard reaction mixture was used for both the soluble system and the chromatin bodies, even when cytoplasmic extracts were present. These extracts contain dNTPs, as indicated by the loss of total dependence upon added dNTPs (and probably contain rNTPs as well). This complicates estimating the degree of stimulation by cytoplasmic factors, since we have not yet determined the extent of dilution of the [32P]dNTPs used for labeling. These systems lose activity during incubation at 370 in the absence of dNTPs and cytoplasm; the extent of loss is variable (70-100% in 15 min) and its cause was not explored. Kinetics of DNA Synthesis. In the absence of cytoplasmic extract the incorporation is linear with time for the first few minutes, and then decreases in rate. Incorporation by chromatin bodies generally ceases (or slows dramatically) by about 10-15 min; the soluble system generally levels sooner, but in some experiments it too has continued for over 15 min. The presence of cytoplasmic extract increases the initial rate of DNA synthesis, and in addition seems to allow the reaction to proceed for a longer period, as much as 30 min or more. (The incorpo-
Biochemistry: Edenberg et al.
4394
30
Proc. Natl. Acad. Sci. USA 73 (1976) 5
2)
-
a
jq
I25-
ba:;20-
lo
o -
02
x
5.0
// A
0
0.5
CL
}T
a-
0 '5
1
X
A
'Q x
4r
5
10
a-
(j
5
_0
5
'0N -
iL
I1
10
20 25 -30 15 Fraction number FIG. 1. Sedimentation of soluble system components through a neutral sucrose gradient. A soluble system (from cells labeled with [-l]thymidine for 15 min in vivo) was incubated at 370 in the standard reaction mixture containing [a-32P]dTTP (22.5 gCi/ml, 17.9 Ci/mmol) in the absence of cytoplasm. After 45 min, EDTA was added to 38 mM and the reaction mixture was cooled on ice, then layered over a 9.5-ml gradient (10-40% sucrose in 0.2 mM NaH2PO4, pH 7.5, over a 1-ml shelf of 80% sucrose in the same buffer) in a polyallomer tube. Centrifugation was in the SW41 rotor at 40,000 rpm for 120 min at 4°. Aliquots from each fraction were processed as described. (-) :fH (in vivo); (0) 32P (in vitro).
ration at each time point is given by the 32P/3H ratios in Figs. 3 and 4.)
Fractionation of the Soluble System on a Sucrose Gradient. The sedimentation profile of the protein-DNA complexes responsible for synthesis in vitro in the soluble system is shown in Fig. 1. A variable portion of the prelabeled material, containing both cellular and unit-length viral DNA, sedimented very rapidly. In Fig. 1 the major part of the 3H prelabel forms a sharp peak with a leading shoulder. DNA extracted from the peak sediments (>80%) as form I in alkaline sucrose gradients, while DNA extracted from the shoulder sediments as SV40 replication intermediates in neutral and alkaline gradients. Thus, under our conditions, as under other conditions (8-11), replicating 5V40 chromosomes sediment faster than mature SV40 chromosomes. The 32p label incorporated in vitro sediments predominantly with the leading shoulder of prelabel which contains replication intermediates, as expected for a continuation of DNA synthesis. Characteristics of DNA Made In Vitro. Most of the DNA made in vitro sediments more rapidly than form I DNA in neutral sucrose gradients (Fig. 2), as would be expected for incorporation into replication intermediates. After longer incubations (>15 min), especially in the presence of cytoplasmic extract, the peak of DNA made in vitro sediments noticeably more slowly (Fig. 2), much of the DNA at about the same rate as form I DNA. In the absence of cytoplasm some DNA may sediment more slowly than form I, and some remains faster sedimenting even at 30 min. This sedimentation behavior may be related to the closer approach to completion of synthesis achieved in the presence of cytoplasm and lack of joining of synthesized fragments in its absence (see below). In order to determine which portions of the SV40 DNA molecule were being replicated in vitro, viral-sized DNA synthesized by chromatin bodies in the presence of cytoplasmic extract was digested with the restriction enzyme HindIII and electrophoresed on an agarose gel. The fluorescence pattern of the chromatin body preparation was identical to that of purified SV40 form I, digested in parallel. The uniform prelabel in vivo
Fraction number FIG. 2. Sedimentation behavior of DNA synthesized in vitro. A chromatin body preparation (from cells labeled 14.5 min in vivo with [3H]thymidine) was incubated at 370 in a standard reaction mixture ([32P]dATP at 23 MACi/ml, 1.8 Ci/mmol) in which 48% of the volume was the cytoplasmic extract of HeLa cells described. At the times indicated aliquots were removed and viral DNA extracted (7), layered on neutral sucrose gradients (10-45% sucrose in 0.1 M Tris, 0.05 M EDTA, 0.85 M NaCl, pH8), and centrifuged at 45,000 rpm for 225 min at 220 in the SW 50.1 rotor. Aliquots of each fraction were acid-precipitated as described. (a) 0.6 min incorporation in vitro; (b) 15.1 min incorporation in vitro. (0) 3H in vivo label; (0) 32p in vitro label.
corresponded exactly with the fluorescent bands, as did most of the in vitro radioactivity. (Some of the in vitro label trailed the bands slightly, particularly near fragment A; this could conceivably be due to the presence of replication forks.) The ratio of in vitro radioactivity to the uniform prelabel was highest (9.7) in the A fragment (around the terminus of replication) and lowest (3.6) in the B + C doublet (around the origin). Such a pattern is consistent with a relative deficiency (or lack) of initiation of new rounds of replication. Size of Strands Made In Vitto. Viral-sized DNA pooled from neutral sucrose gradients was sedimented through alkaline sucrose gradients to determine the size of the DNA strands made in vitro. At all times from 0.5 to 15 min of labeling in vitro, the DNA made in the absence of cytoplasmic extract by both chromatin bodies and the soluble system (Fig. 3) contains approximately half long strands (up to full viral length) and half short strands (about 200-30 nucleotides, data not shown). This pattern is identical in both systems and is maintained while the 32P/3H ratio is increasing (i.e., while DNA synthesis is continuing). In neither case is there significant chasing of short strands into longer strands (Fig. 3). The addition of a cytoplasmic extract to either system dramatically changes the results. Both the chromatin bodies and the soluble system (Fig. 4), when supplemented with the extracts, show a peak of short strands after brief incubations in vitro (up to 2-4 min), but with longer incubations this peak becomes relatively smaller while the peak of longer strands becomes larger. Again, the pattern is identical in both systems, except that the rate at which the peak of short strands is ob-
Proc. Natl. Acad. Sci. USA 73 (1976)
Biochemistry: Edenberg et al.
4395
partially substituted DNA. This is consistent with the even distribution of radioactivity between Okazaki fragments and longer strands, and is similar to data reported for another in vitro system (12). As expected for DNA made semiconservatively, sonication of the DNA to a size about l' that of SV40 shifts most of the in vitro label to nearly fully heavy density in alkaline CsCl (data not shown). DISCUSSION 8n
1
m in
U4
x 6
P/:
P/H
A
15
5
25
35
45
5
15
25
35
45
Fraction number
FIG.
3.
Size distribution of DNA synthesized in vitro by the
soluble system (without cytoplasm). A soluble system (from cells la-
beled with
[2H]thymidine for 9.8 mn
in viva) was incubated at 370
[12P]dATP (72 mCi/ml
in the standard reaction mixture containing
at 9.2 Ci/mmol). At the times noted, aliquots were removed and viral
DNA was extracted and sedimented through neutral sucrose gradients as described in the legend of Fig. 2. The peak of viral-sized DNA was
pooled, dialyzed, and concentrated as described. The samples were adjusted to 0.2 M NaOH, and after 30
mmn layered on 10-30%6 alkaline
sucrose gradients and centrifuged at 45,000 rpm for 6 hr at 220
in the
SW 50.1 rotor. The duration of each pulse and the 32P/3H ratio (rep-
resenting the amount of DNA synthesized) are shown.
(-) :H
(in
viva); (0) 32p (in vitro).
scured is greater in the chromatin bodies, consistent with their
more rapid DNA synthesis. In the presence of cytoplasmic extract the short strands chase completely into longer DNA (Fig. 4) without further
incorporation of label.
Completion of Molecules. In the absence of cytoplasmic
extract, virtually no radioactive dNTPs are incorporated in vitro into covalently closed molecules (form I). When cytoplasmic factors are present, dNTPs are incorporated into form I. In general, chromatin bodies incorporate a higher proportion of the in vitro dNTPs into form I than does the soluble system; this may be due to damage sustained by the SV4O chromosomes during preparation of the soluble system, as previously noted. In one 30-mma incorporation with chromatin bodies and cytoplasmic extract at least 25% of the in vitro label was converted to covalently closed molecules, whereas a parallel incubation without cytoplasmic extract incorporated less than 1% into form I.
In an
experiment with the soluble system, in which care was
taken to minimize shear, 3-4% of the in vitro label was converted to
form
I.
Semiconservative DNA Synthesis. When incubation in vitro is carried out in the presence of BrdUTP instead of dTTP, the DNA made by chromatin bodies (without cytoplasmic extract) is denser than unsubstituted DNA. When centrifuged to equilibrium in alkaline CsCl, about half of the radioactivity is associated with nearly fully substituted DNA and about half with
In this paper we report methods for preparation of two related subnuclear systems for SV40 DNA replication, as well as preliminary characterization of the two systems. The soluble system we have developed offers many advantages over nuclear in vitro systems, since it should prove far more amenable to fractionation. DNA synthesis in vitro in nuclear sonicates from bovine lymphocytes has been reported (13), but in that system the characteristics of thereaction were unusual. Recently another subnuclear DNA-synthesizing system similar to ours has been independently developed (14). The DNA synthesis in our subnuclear systems appears to be a continuation of the replication process started in vivo. Incorporation during short incubations is into nucleoprotein complexes cosedimenting with those that contain replication intermediate DNA of SV40. The DNA extracted from short incubations cosediments with replication intermediate DNA of SV40. After longer incubations in the presence of cytoplasmic factors a fraction of the label is incorporated into mature (form I) SV40 DNA. DNA synthesis in these systems is semiconser-
vative. One of the interesting characteristics of the DNA synthesized in both our systems, in the absence of cytoplasm, is the fact that regardless of length of incubation or length of chase the synthesized DNA contains half long strands (up to the size of linear viral strands) and half short strands (200-300 nucleotides). In the presence of cytoplasm, the short strands chase into long
strands. These short strands thus resemble the Okazaki fragments that have previously been detected during papovavirus DNA replication in vivo (15, 16) and in nuclear systems (17-25). Other investigators working with nuclear systems have also noticed that cytoplasmic factors stimulate DNA synthesis and joining (18, 20, 23). The equal distribution of radioactivity between long and short strands in the absence of cytoplasm suggests that synthesis of only one of the daughter strands is through Okazaki fragments (semidiscontinuous synthesis). The fact that this is seen in both systems despite differences in the rate and extent of synthesis argues that such a distribution is not due to a coincidental balancing of the rates of synthesis and joining of Okazaki fragments. Francke and coworkers (17-19) have concluded that synthesis in their nuclear in vitro polyoma DNA-synthesizing system is semidiscontinuous; Hershey and Taylor have come to the same conclusion in a different in vitro synthesizing system (12). We note that the equal distribution of label into long and short strands could be consistent with totally discontinuous synthesis provided that the joining of half of the short strands is both very rapid (even in vitro) and not dependent upon cytoplasmic factors. The fact that our soluble system can be sedimented through neutral sucrose gradients to partially purify and separate replicating SV40 chromosomes, and that the replicating SV40 chromosomes can continue synthesis in vitro (details to be published elsewhere), suggests that this system provides an excellent starting point from which it should be possible to purify and characterize the enzymes that participate in DNA
4396
Biochemistry: Edenberg et al.
Proc. Natl. Acad. Sci. USA 73 (1976)
7 x
-3
a-0
0)
2 ~
A, I0~~~ ,
C 2
54- 4 -3
-2
5
15
25
35
0-% 4 0
0
M
Fraction number FIG. 4. Size distribution of DNA synthesized in vitro by the soluble system in the presence of cytoplasmic factors. A soluble system (from cells labeled with [3H]thymidine for 14.5 min in vivo) was incubated at 370 in a standard reaction mixture ([32PjdATP at 23 pCi/ml, 1.8 Ci/mmol) in which 48% of the volume was the cytoplasmic preparation of HeLa cells described. Aliquots were processed as described in the legend of Fig. 3. (-) :H (in vivo); (0) 32p (in vitro).
synthesis, as well as to characterize the proteins that bind to the replicating chromosomes and the structure of the replicating chromosomes. We thank Michael Reszka and Daniel Perlman for helping, and Elaine V. Lenk for the electron microscopy of chromatin bodies. This research was supported by grants from the National Science Foundation and National Institutes of Health. 1.
2. 3. 4. 5.
Edenberg, H. J. & Huberman, J. A. (1975) Annu. Rev. Genet. 9,245-284. Tegtmeyer, P. (1972), J. Virol. 10, 591-598. Francke, B. & Eckhart, W. (1973) Virology 55, 127-135. Francke, B. & Hunter, T. (1974) J. Virol. 13,241-243. Chou, J. Y., Avila, J. & Martin, R. G. (1974) J. Virol. 14, 116-
124. 6. Hancock, R. (1974) J. Mol. Biol. 86, 649-663. 7. Hirt, B. (1967) J. Mol. Biol. 26,365-369. 8. White, M. & Eason, R. (1971) J. Virol. 8,363-371. 9. Hall, M. R., Meinke, W. & Goldstein, D. A. (1973) J. Virol. 12, 901-908. 10. Sen, A. & Levine, A. J. (1974) Nature 249,343-344. 11. Cremisi, C., Pignatti, P. F., Croissant, 0. & Yaniv, M. (1976) J.
Virol. 17,204-211. 12. Hershey, H. V. & Taylor, J. H. (1974) Exp. Cell Res. 85, 7988. 13. Thompson, L. R. & Mueller, G. C. (1975) Biochim. Biophys. Acta 14.
378,344-353. Su, R. T. & DePamphilis, M. L. (1976) Fed. Proc. 35, 1376
abstr. 15. Fareed, G. C. &. Salzman, N. P. (1972) Nature New Biol. 238, 274-277. 16. Laipis, P. J. and Levine, A. J. (1973) Virology 56,580-594. 17. Francke, B. & Hunter, T. (1974) J. Mol. Biol. 83,99-121. 18. Francke, B. & Hunter, T. (1975) J. Virol. 15,97-107. 19. Francke, B. & Vogt, M. (1975) Cell 5,205-211. 20. Otto, B. & Reichard, P. (1975) J. Virol. 15, 259-267. 21. Magnusson, G., Pigiet, V., Winnacker, E. L., Abrams, R. & Reichard, P. (1973) Proc. Natl. Acad. Sci. USA 70,412-415. 22. Qasba, P. K. (1974) Biochem. Biophys. Res. Commun. 60, 1338-1344. 23. DePamphilis, M. L. & Berg, P. (1975) J. Biol. Chem. 250, 4348-4354. 24. Magnusson, G. (1973) J. Virol. 12, 609-615. 25. Qasba, P. K. (1974) Proc. Natl. Acad. Sci. USA 71, 10451049.
