Requirement of a Cytoplasmic Fraction for Synthesis of SV40 Deoxyribonucleic Acid in Isolated Nuclei*

50% of the SV40 DNA in the process of replication lysates of infected BSC-1 cells by conversion to covalently closed, superhelical DNA). Fractionation of the lysate into nuclear and cytoplasmic components blocked 99% of the synthesis of SV40(1) DNA in the purified nuclei. The reconstituted system, made by adding back the cytoplasmic fraction before incubation at 30”, completely restored the in vitro level of SV40(1) DNA synthesis. Preliminary characterization of the activity found in the cytoplasmic fraction suggested it was a soluble, heat-labile protein (or proteins) with a minimum molecular weight of about 30,000 and an active sulf’hydryl group. The activity was present in both infected and uninfected monkey cells, and at a lower level in mouse, hamster, and human cell lines. Neither serum starvation nor cycloheximide treatment of cells diminished the activity in the cytoplasmic fraction. Purified cytoplasmic DNA polymerase from KB cells did not substitute for the cytoplasmic fraction which was required for elongation of newly synthesized DNA strands. In the absence of the cytoplasmic fraction, conversion of 4 S DNA into longer strands was inhibited, and SV40(RI) DNA appeared to be broken specifically at the replication forks.


BERG
From the Department of Biochemistry, Stanford University Medical Center, Stanford, California 94305 About 50% of the SV40 DNA in the process of replication (SV40(RI) DNA) completed rephcation in lysates of infected BSC-1 cells by conversion to covalently closed, superhelical SV40 DNA (SV40(1) DNA). Fractionation of the lysate into nuclear and cytoplasmic components blocked 99% of the synthesis of SV40(1) DNA in the purified nuclei. The reconstituted system, made by adding back the cytoplasmic fraction before incubation at 30", completely restored the in vitro level of SV40(1) DNA synthesis.
Preliminary characterization of the activity found in the cytoplasmic fraction suggested it was a soluble, heat-labile protein (or proteins) with a minimum molecular weight of about 30,000 and an active sulf'hydryl group. The activity was present in both infected and uninfected monkey cells, and at a lower level in mouse, hamster, and human cell lines. Neither serum starvation nor cycloheximide treatment of cells diminished the activity in the cytoplasmic fraction. Purified cytoplasmic DNA polymerase from KB cells did not substitute for the cytoplasmic fraction which was required for elongation of newly synthesized DNA strands. In the absence of the cytoplasmic fraction, conversion of 4 S DNA into longer strands was inhibited, and SV40(RI) DNA appeared to be broken specifically at the replication forks. 4348 In the accompanying paper (1) we demonstrated that a lysate from SV40-infected monkey cells can convert replicating SV40 DNA molecules (SV40(RI) DNA)' to covalently closed, superhelical DNA (SV40(1) DNA) in vitro. Further fractionation of this system revealed an absolute requirement for the cytoplasmic fraction. Purified nuclei in the presence of a buffered salts solution containing deoxynucleotides and an ATP regenerating system were unable to elongate further the daughter strands and make SV40(1) DNA. The reconstituted system, made by adding back the cytoplasmic fraction, was just as active as the unfractionated lysate. This system now provides a complementation assay for "cytoplasmic fractors" required in SV40 DNA replication in vitro.

Virus and Cells
Growth of SV40 virus and established monkey kidney cell lines were described in the accompanying paper (1). Other cell lines used in this work were grown the same way. Established mouse cell lines, 3T3 and 3T6, were from Howard Green (2), and SV3T3, an SV40-transformed 3T3 cell iine, was from George Todaro (3). Primary African green monkey kidney cells (AGMK) were purchased from Flow Laboratories. T22, an SV40-transformed monkey cell was obtained from Janet Butel (4). Hamster cell line BHK212 was obtained from Michael Stoker (5) and SV28, an SV40-transformed BHK cell line (6) was supplied by George Stark. David Sedwick provided the KB human cell line (5).

Analysis of DNA Synthesis in Vitro
Techniques for the preparation of lysates of SV40-infected cells, assay of SV40(1) DNA, neutral and alkaline sucrose gradient sedimentation, equilibrium centrifugation in CsCl-ethidium bromide density gradients, preparation of viral DNA standards, and the conditions for DNA synthesis in vitro are all described in the accompanying paper (1).

Preparation of Cytoplasmic Fractions
A lysate of uninfected, unlabeled cells (3 x 10' cells/ml) was freshly prepared for each experiment. A low speed cytoplasmic fraction was the supernatant fraction obtained by sedimenting the nuclei at 1200 x g for 5 min at 2'. The ability of this fraction to restore synthesis of SV40(1) DNA in purified nuclei was destroyed by freezing, but a high speed cytoplasmic fraction prepared by further sedimentation at 100,000 x g for 1 hour at 2" and adjustment of the supernatant fraction to 25 mM KC1 could be frozen in Dry Ice and ethanol and stored at -20" for at least 2 weeks with no loss of activity. This fraction contained 4 to 6 mg of protein per ml (Lowry assay (7)).

Conditions for Assaying Cytoplasmic Fractions
Two methods were used to prepare nuclei from SV40-infected cells that required the addition of a cytoplasmic fraction to convert SV40(RI) to SV40(1) DNA. Triton X-100 Method-All steps were carried out at 2-4" using an ice bath whenever possible. Cell lysates were prepared from infected cells that had their SV4O(RI) DNA prelabeled with ['Hlthymidine for 4.5 min as described previously (1).  (1). Pelleting the nuclei did not impair their ability to carry out this conversion if the supernatant fraction and nuclei were recombined (Table I). If the pelleted nuclei were resuspended instead in hypotonic buffer (about two-thirds of the total lysate volume), the activity was only slightly diminished.
However, if the lysate was first diluted 20-fold in a hypotonic buffer, the nuclei were about 18% as active as a cell lysate while those resuspended in the low speed cytoplasmic fraction contained 64% of the activity found in a cell lysate (Table I). Therefore, under certain conditions viral DNA synthesis in the nuclei was dependent upon addition of a soluble factor (or factors) in the cytoplasmic fraction. Two methods were developed for purifying nuclei that were dependent upon the cytoplasmic fraction for synthesis of SV40(1) DNA. Both of these methods (described in detail under "Experimental Procedures") produced nuclei with 1 to 6% of the activity observed in a cell lysate ( Table I). Addition of a low speed cytoplasmic fraction from an equivalent number of cells resulted in a 5-to 7-fold increase in the actual amount of SV40(1) DNA synthesized; this corresponded to 85 to 100% of the activity observed in an unfractionated cell lysate. Synthesis of SV40(1) DNA in isolated nuclei was prevented by washing nuclei in 20 mM Hepes (pH 7.8) and 0.5 mM dithiothreitol containing 5 mM KC1 or less and no MgCl,. However, complete restoration of DNA synthesis was observed only when nuclei had been washed in 50 to 100 mM KCl. Therefore nuclei were first washed in a low ionic strength buffer followed by a high ionic strength buffer containing 0.5 mM MgCl,; a divalent cation which decreased nuclear aggregation and increased nuclear stability as judged by their ability to synthesize viral DNA after 2 to 3 hours at 0". Addition of 0.5 mM CaCl, to the final wash resulted in about 20% less activity in the reconstituted system, but addition of ethylene glycol bis(a-aminoethyl ether)-N,N'-tetraacetate (EGTA), a preferential Ca+ chelator (ll), made no difference in the degree of inactivation or reactivation of isolated nuclei. The requirement for a low ionic strength wash to remove the cytoplasmic fraction from nuclei could be circumvented by including as little as 0.01% of either Triton X-100 or Brij-58 in the wash solution. Nuclei treated in this manner required a second wash to remove excess detergent which depressed synthesis of SV40(1) DNA by increasing the amount of SV40(RI) that was converted to SV40 (11). Including sucrose in the wash solutions resulted in a 10% increase in SV40(1) DNA made in the reconstituted system. The amount of SV40(1) DNA synthesized in the reconstituted system was always proportional to the concentration of high speed cytoplasmic fraction added (0.6 to 6 mg of protein per ml; i.e. bovine serum albumin equivalents using Lowry method (7)). However, the proportionality constant relating the conversion of SV4O(RI) DNA to SV40(1) DNA and the concentration of cytoplasm varied between 0.5 and 0.9 among different cytoplasmic preparations (Fig. 1). This might reflect a difference in the relative concentration of several active components.

4349
Whereas the high speed cytoplasmic fraction could never be diluted more than lo-to 15-fold, an unfractionated cell lysate diluted 40-fold with a proportional volume of assay mix and hypotonic buffer showed no decrease in SV40(1) synthesis (Fig. 1). Nuclei recovered from a lysate diluted lo-fold in this manner retained 60% of their capacity to make SV40(1) DNA. As indicated above, the soluble factor found in the cytoplasmic fraction is apparently associated with nuclei in the presence of KC1 and MgCl, concentrations used in our standard assay. Viral DNA Synthesis in Isolated Nuclei in Absence of Cytoplasmic Fraction-Although this paper only presents the data obtained with Triton X-loo-prepared nuclei, the same experiments performed on nuclei prepared using the low salt method gave essentially identical results. SV4O(RI) DNA, prelabeled in uiuo with [3H]thymidine, was not converted into either SV40(1) or covalently closed dimeric DNA by purified nuclei incubated at 30" in the absence of the cytoplasmic fraction ( Fig. 2A). Addition of the cytoplasmic fraction after 30 min of incubation failed to stimulate synthesis of SV40(1) DNA. Apparently prelabeled SV40(RI) DNA underwent some form of decomposition; it no longer sedimented in its normal position of 25 to 27 S (see Fig. 4A, Ref. 1) but was now found in three major peaks of 22 S, 20 S, and 17 S (Fig. 3A). The one fraction difference that distinguishes the 17 S peak from the 16 S SV40(11) DNA marker was found in 10 separate experiments.
Elongation of the daughter DNA strands was inhibited by the absence of the cytoplasmic fraction. About 50% of the label in nascent daughter strands of SV40(RI) DNA, labeled in oiuo for 30 s at 25", sedimented at 4 S in an alkaline sucrose gradient (Fig. 4A) as expected from a discontinuous model of DNA synthesis (12). Purified nuclei isolated from these cells and incubated in the absence of the cytoplasmic fraction converted most of the prelabeled 4 S DNA into longer strands (Fig. 4B). However, only a slight increase in the length of daughter strands occurred, In contrast to the obvious disappearance of 4 S DNA in Fig. 4, no significant growth of daughter strands was detected when longer labeling times were used to label SV4O(RI) and 4 S DNA was not preferentially labeled (e.g. 4 min at 37", see To evaluate further the effect of the cytoplasmic fraction on viral DNA synthesis, the endogenous SV40 (1) (1) DNA after dilution of the unfractionated cell lysate or the cytoplasmic fraction. The unfractionated cell lysate (1) was diluted with the appropriate volume of standard assay mix and hypotonic buffer required to maintain their concentrations at the standard levels (O---O).
The cytoplasmic fraction was diluted with hypotonic buffer before addition to purified nuclei (04).
In both cases the conversion of prelabeled SV4O(RI) to SV40(1) DNA was measured. The dilution factor is the ratio of the concentrations of the diluted lysate or cytoplasm to the undiluted fraction. purified nuclei were prepared and incubated at 30" with both [a-32P]dATP and [cr-3ZP]dCTP (1). In the absence of an added cytoplasmic fraction, the 32P label was rapidly incorporated into the DNA of an SDS supernatant for 10 min and then stopped. Whereas the initial rate was only slightly reduced relative to the reconstituted system, the extent of incorporation was reduced 3-fold (Fig. 5). Cellular DNA synthesis (the SDS pellet) responded in a similar manner. Less than 1% of the "P label in the SDS supernatant was incorporated into SV40(1) or covalently closed dimeric DNA (Fig. 6A). About 80% of the '*P label sedimented at 3 to 5 S in an alkaline sucrose gradient (Fig. 7A) and at 3 to 7 S in a neutral sucrose gradient (Fig. 8A). This material corresponds in size to the short nascent DNA fragments that are intermediates in SV40 DNA chain elongation (12). Its appearance in a neutral sucrose gradient may have resulted from strand displacement (13) or from degradation of the replication forks during incubation at 30". The remaining 20% of the incorporated 3T sedimented at about 22 S in a neutral sedimentation gradient (Fig. 8A) and presumably accounts for the label associated with strands longer than 3 to 5 S in an alkaline sucrose gradient but never longer than one viral genome length (Fig. 7A). A comparison of Fig. 8 A (14). About 10% of the prelabeled SV40(1) DNA was converted to SV40(11) DNA after 1 hour at 30" suggesting the presence of a low level of endonuclease activity.

Viral DNA Synthesis in Isolated Nuclei in Presence of Cytoplasmic Fraction
The reconstituted system, made by restoring the cytoplasmic fraction to the nuclei prior to incubation at 30", completed replication of SV40(RI) DNA as effectively as the unfractionated lysate system described previously (1). About 55% of the SV40(RI), prelabeled in uiuo with [3H]thymidine, was converted to SV40(1) DNA in vitro (Fig. 2B). No difference in sedimentation behavior was found between SV40(1) DNA produced in the reconstituted system and that made in a cell lysate (1) using alkaline (Fig. 2s) and neutral (Fig. 3B) sucrose gradients and equilibrium centrifugation in CsCl-ethidium bromide gradients (data not shown). Normal elongation of the nascent daughter strands was also observed in the reconstituted system. Purified nuclei isolated from infected cells, prelabeled with [3H]thymidine for 30 s at 25", and incubated with the cytoplasmic fraction converted 99% of the prelabeled 4 S DNA into longer strands; 75% of the label remaining in DNA containing single strand interruptions was one viral genome in length (Fig. 4C).
In vitro DNA synthesis was also measured by incorporation of [a-'T]dATP and [a-3ZP]dCTP. About 35% of the 32P label was found in SV40(1) DNA (Fig. 6B) and 80% of the newly made strands that still had free ends were one viral genome in length (Fig. 7B). As previously observed in unfractionated cell lysates (l), incorporation of labeled deoxyribonucleotides in vitro preferentially labeled SV40(RI) DNA that was early in replication at the time that nuclei were isolated. Therefore, following a l-hour incubation in vitro, about 55% of the 3zP label appeared as SV40(RI) in neutral sucrose gradients (Fig.  8B) in an advanced stage of replication (Fig. 7B). Since the fraction of SV40 (1) 3 (center). Sedimentation in neutral sucrose gradients of prelabeled SV40 DNA incubated in vitro. A, purified nuclei incubated in the absence of the cytoplasmic fraction as described in Fig. 2. B, nuclei incubated with the added low speed cytoplasmic fraction. Gradients remained constant throughout the in vitro incubation, no significant endonucleolytic activity was present in the reconstituted system.
Although DNA synthesis in a reconstituted system mimicked that observed in a cell lysate, the reconstituted system often incorporated more in the in vitro label (Fig. 5). This difference generally appeared as 3 to 7 S material on a neutral sucrose gradient (Fig. 8B) (1) [SH]DNA was always used as the criterion for optimal conditions in vitro.
The amount of cellular DNA in the SDS supernatant was estimated from the amount of [?P]DNA found when using nuclei from mock infected cells and, with nuclei from infected cells, the fraction of viral DNA was determined by DNA-DNA hybridization (15). In the reconstituted system, about 10 to 15% of the 'rP label was in cellular DNA whereas with purified nuclei incubated alone about 30% of the labeled DNA was cellular. The labeled cellular DNA was distributed throughout a neutral sucrose gradient with about 25% concentrated in the 3 to 7 S region.
were centrifuged in an SW 56 rotor for 3.25 hours, 4', at 55,000 rpm and collected from the bottom. SV40 (1)

Preliminary
Characterization of Active Cytoplasmic Fraction The cytoplasmic fraction was prepared from several cell lines and tested for the presence of the activity which restores the conversion of SV40(RI) to SV40(1) DNA in nuclei purified from SV40-infected BSC-1 cells (Table II). Since the activities from uninfected and infected BSC-1 cells were equivalent, a viral gene product was not responsible for the observed activity. The cytoplasmic fraction from other monkey cell lines was fully active whereas the same fractions from the human, mouse, and hamster cell lines tested were 75 to 25% as active as that from an equivalent number of BSC-1 cells. One surprising result was the very low activity found in two SV40-transformed cell lines, T22 and SV3T3. However, the significance of this difference must await further studies on the efficiency of the isolation procedure and stability of the various cytoplasmic fractions.
The requirement for the cytoplasmic fraction was not met by gelatin (0.2 to 2 mg/ml), bovine serum albumin (0.2 to 2 mg/ml), cytochrome c (0.2 to 2 mg/ml), or a mixture containing 120 PM each of ATP, UTP, CTP, and GTP. Furthermore, neither the high speed supernatant fraction from an Escherichia coli lysate (10) which supports conversion of a single strand circular Ml3 parental DNA to a double strand replicative form, nor E. coli DNA polymerase I (8), nor the cytoplasmic DNA polymerase from KB cells (9) exhibited any activity. by guest on March 24, 2020 http://www.jbc.org/ Downloaded from However, both a freshly prepared cytoplasmic fraction from KB cells and the cytoplasmic extract from which the KB cell DNA polymerase was purified (9) contained 75% as much cytoplasmic factor activity as did BSC-1 cell cytoplasm. Either the cytoplasmic polymerase is not the active agent or it requires additional factors.
The cytoplasmic fraction from cells that were not actively synthesizing DNA was also active. DNA synthesis in mammalian cells can be arrested by holding a confluent monolayer in low concentrations of serum (16,17). Confluent monolayers of BSC-1 cells were kept in Dulbecco's Modified Eagle's medium plus 0.1% calf serum for 4 days at which time autoradiographic analysis of a [3H]thymidine pulse showed that only 1% of the cells were active in DNA synthesis. The monolayers were then infected with SV40 for 1 hour at a multiplicity of 40 and further incubated in the same medium.
Cell lysates from these virus-infected, serum-starved cells incorporated about 2.5 times less [SH]thymidine into viral DNA during a 4-min pulse than did a lysate from cells grown in the usual manner.

TABLE II
Activity of cytoplasmic fractions from several cell lines A low speed cytoplasmic fraction was prepared from each cell line as described under "Experimental Procedures" and assayed for its ability to support SV40(1) DNA formation when incubated with nuclei (low salt method) from infected BSC-1 cells. Activity was measured as the conversion of prelabeled SV4O(RI) ISH]DNA to SV40 (1) [BH]DNA relative to that observed using BSC-1 cell cytoplasm.

Cell line
Cell type Activity   However, in vitro conversion of prelabeled SV40(RI) to SV40(1)  24",its viral DNA replication (18,19), but the level of activity in then with trypsin inhibitor for 5 min before incubatian with nuclei.
the cytoplasmic fraction did not diminish. Exposure of infected Heat treatment: HSC was incubated at either 24" for 35 min or 80" for BSC-1 cells to 20 wg/ml of cycloheximide (19) between 33 and 5 min. N-Ethylmaleimide (NEM) treatment: a cell lysate was incubated with 10 mM NEM with and without 12 mM dithiothreitol (DTT). 36 hours after infection reduced the rate of viral DNA synthesis HSC was incubated with NEM for 5 min at 24" then DTT was added in uivo to 15% of its normal value. (Cells were incubated for 1 for 5 min before assaying for cytoplasmic factor activity. p-Chloromerhour in 50 PCi of [SH]thymidine, 10 ml of complete medium, curibenzoate (PCMB) treatment was the same as NEM treatment 27", 6% CO,.) The low speed cytoplasmic fraction, isolated except 5 mM PCMB and 25 mM &mercaptoethanol (BME) were from infected cells that had been treated with cycloheximide, used. Since PCMB reacts reversibly with sulfhydryl groups, it was not was equivalent to the same fraction from untreated cells in completely inactivated by excess BME. BME alone is sufficient to both its ability to support the in vitro conversion of SV40(RI) convert some SV40 (1)  although their size did not require a soluble factor (or factors) found in the cytoplasmic increase to the length of SV40 DNA. Therefore, in the absence fraction in order to convert SV40(RI) to SV40(1) DNA. Alof the cytoplasmic fraction, nuclei permit the joining of though this activity was found in the cytoplasmic fraction, it relatively complete 4 S DNA to daughter strands, but further presumably is associated with the nucleus in the intact cell as elongation is inhibited and newly made 4 S DNA labeled in suggested by the fact that a cell lysate can be diluted under our vitro accumulates. The cytoplasmic fraction may be involved standard assay conditions without a decrease in the synthesis in the synthesis of DNA in the gaps created by discontinuous of SV4O(I) DNA (Fig. 1). However, this factor is not bound synthesis, or the excision of RNA primers and subsequent tightly to DNA since it was completely removed by washing joining of DNA pieces. The rapid initial rate of DNA synthesis nuclei in a low ionic strength buffer. The effectiveness of (Fig. 5) suggests that synthesis of 4 S pieces previously nonionic detergents in purifying nuclei further suggests that initiated in viuo is uneffected by removal of the cytoplasmic increasing the permeability of the outer nuclear membrane fraction. may be important in releasing all the residual factor (21). The cytoplasmic fraction may also be involved in preventing Although purified nuclei incubated in the absence of the the breakdown of replicating molecules. SV4O(RI) DNA, prelacytoplasmic fraction did not synthesize SV40(1) DNA, some beled in uiuo, changed its sedimentation behavior in neutral viral DNA synthesis did occur. This synthesis, which was sucrose gradients from a broad peak at 26 S to three peaks at 22 completed within 10 min, accounted for only one-third as much S, 20 S, and 17 S (Fig. 3A). Only the 22 S peak contained DNA labeled DNA in the SDS supernatant as found in the reconsti-synthesized in vitro and was apparently SV~()(RI) at an advanced stage in replication. One hypothesis is that SV40(RI) DNA in the absence of the cytoplasmic fraction was subjected to single strand endonucleolytic degradation at its replication forks. Assuming one single-stranded template region at each fork and on opposite sides (22), the postulated nuclease would produce relaxed circular molecules with a tail (the 20 S material), linear molecules greater than one genome in length (the 17 S material), and release short pieces of DNA only where the newly made 4 S pieces had not yet been joined to the growing daughter strand. Treatment of polyoma RI with Neurospora crassa single strand-specific endonuclease partially confirms these predictions (24). Note that the prelabeled daughter strands in SV40(RI) DNA were not degraded (Fig.  3A). The cytoplasmic fraction may contain an inhibitor of a single strand-specific endonuclease which prevents degradation of the single strand regions exposed during discontinuous DNA replication.
An alternate hypothesis is that the template strands of replicating molecules are nicked, causing them to sediment more slowly in neutral sucrose gradients and release newly made DNA at the replication forks by a strand displacement mechanism before it can be joined to the daughter strands. Such a phenomenon may result from the general inhibition of DNA synthesis is suggested by in uiuo studies with hydroxyurea (25, 26) and 5-fluorodeoxyuridine (27). Addition of the cytoplasmic fraction to purified nuclei completely restores the nuclei's ability to complete replication of SV40(RI). The reconstituted system behaves essentially the same as an unfractionated cell lysate; a lysate converts about 55% of the SV40(RI) to SV40(1) at about one-third the rate found in uiuo (1). The active factors in the cytoplasmic fraction have not yet been identified although they are found in uninfected cells and one of them is a heat-labile protein with an active sulfhydryl group. Cytoplasmic DNA polymerase could not replace this activity.
The function and properties of the cytoplasmic fraction described in this paper are similar to those of a cytoplasmic fraction that stimulates incorporation of labeled nucleotides into isolated HeLa cell nuclei (28,29), an activity also reported in other systems (6,(30)(31)(32). However, the HeLa cell activity is found during S phase but not Gl (28,29). Initiation of DNA synthesis in nuclei from Gl cells by cytoplasm from S phase cells has also been reported (33) but not confirmed (28). Since none of the replication factors involved have been purified or well characterized, it is not known whether they correspond to any of the DNA polymerases (9,34,35) or ligases (36, 37) already described.
One problem in unraveling the biochemistry of nuclear DNA replication lies in our inability to define the nature of both the substrates and products of synthesis. Viral DNA replication in isolated nuclei provides a way to overcome this problem since SV40 and polyoma DNA molecules have been well characterized and portions of their normal in uiuo replication process (14,22,38) have been duplicated in isolated nuclei (39,40) and cell lysates (1,41). Results with Triton X-100 isolated nuclei from SV40 infected African Green monkey kidney cells (40) are consistent with our results obtained in the absence of the cytoplasmic fraction. These in vitro systems should permit a functional analysis of the various factors involved.