The in Vitro Replication of DNA Containing the SV40 Origin

The SV40 genome contains a single origin of DNA replication within a small (5243 base pairs) covalently closed double-stranded DNA genome and is organized into minichromosomes containing histones derived from the host cell. With the exception of the virus-encoded large T antigen, all the proteins required for SV40 replication are supplied by the host cell. Thus, many of the essential host proteins for SV40 DNA replication probably play a similar role in cellular DNA replication. The establishment of an in vitro SV40 DNA replication system by Li and Kelly (1) has been pivotal in allowing the characterization of the host proteins which play a role in SV40 replication. This review will focus on the pathway of SV40 DNA replication and the proteins involved in the DNA synthesis reactions. Other reviews of eukaryotic DNA replication have also appeared recently (2-5).

The SV40 genome contains a single origin of DNA replication within a small (5243 base pairs) covalently closed double-stranded DNA genome and is organized into minichromosomes containing histones derived from the host cell. With the exception of the virus-encoded large T antigen, all the proteins required for SV40 replication are supplied by the host cell. Thus, many of the essential host proteins for SV40 DNA replication probably play a similar role in cellular DNA replication. The establishment of an in vitro SV40 DNA replication system by Li and Kelly (1) has been pivotal in allowing the characterization of the host proteins which play a role in SV40 replication. This review will focus on the pathway of SV40 DNA replication and the proteins involved in the DNA synthesis reactions. Other reviews of eukaryotic DNA replication have also appeared recently (2-5).

A Model of the DNA Replication
Fork during SV40 DNA Synthesis In order to introduce the proteins involved in carrying out DNA synthesis, a model of the replication fork is presented in Fig. 1, and a summary of their properties is presented in Table I. A number of ATP-dependent pre-DNA synthetic reactions at the core origin of replication, carried out by T antigen, summarized by Borowiec et al. (6), lead to DNA unwinding in the presence of a single-stranded DNA binding protein and topoisomerase I. At some point prior to the extensive unwinding of the duplex from the core origin, T antigen may interact with the DNA pol a-DNA primase complex which initiates RNA-primed DNA synthesis (7,8).
As shown in Fig. I, DNA pol cu-primase, on the lagging strand, synthesizes short RNA-primed DNA fragments in the presence of the three-subunit human SSB (HSSB).' Joining of these "Okazaki fragments" to form a completed daughter strand requires the action of a 5'-+3'-exonuclease and RNase H to remove the RNA primers, filling in of the gaps by a polymerase, and sealing the nicks by DNA ligase (9). Finally, topoisomerase II can decatenate the daughter molecules (9, 10).
DNA pol 6 is shown carrying out elongation of the leading daughter strand in the presence of HSSB and in conjunction with the protein factors PCNA and activator 1 (Al) (ll),' also identified as RF-C (12).
Since DNA polymerases cannot initiate DNA chains de nouo (13), DNA primase is required to synthesize small oligoribonucleotide primers. In a coupled reaction, these primers ' The abbreviations used are: SSB, single-stranded binding protein; HSSB, human single-stranded binding protem; PCNA, pfoliferating cell nuclear antlgef Al, activator 1; pol, polymerase; RF, repbcatwe facto!. S.-H. Lee, A. Kwong, Z.-Q. Pan, and J. Hurwitz, J. Bcol. Chem., submitted for publication. are immediately extended by the DNA p01 a complexed to primase. In contrast, pol d lacks DNA primase activity and extends RNA primer ends poorly. Thus, the pol a-DNA primase complex is responsible for the synthesis of the first DNA segment, which could be considered the first lagging strand Okazaki fragment but is subsequently utilized as the primer for the initiation of leading strand synthesis by pol 6. In the presence of PCNA and Al, pol d elongates the leading strand, while pol cr-primase continues synthesis of the lagging strand as T antigen unwinds the duplex.

Role of the Single-stranded DNA Binding Protein
No other SSB examined can efficiently replace the threesubunit SSB (Table I) isolated from human cells (HSSB) in the SV40 replication reaction (14,15,56). A similar threesubunit SSB from yeast partially replaces the HSSB in replication (16). This is in contrast to the T antigen-mediated unwinding reaction for which SSBs isolated from Escherichia coli, adenovirus, and herpes simplex virus all substituted for HSSB (17). The SSBs that support the elongation of primed DNA templates by both DNA pol N and pol d are the HSSB and yeast SSB.3 Pol LY was inhibited by all SSBs tested with the exception of the three-subunit SSBs (human and yeast), while pol 6 was active in the presence of all SSBs tested. Furthermore, monoclonal antibodies directed against either the human 70-kDa subunit, which binds DNA, or the human 34-kDa subunit block the replication reaction, suggesting that both subunits are essential (18).
The sequence of the human 34-kDa subunit has been determined (19). In yeast and human cells, the state of phosphorylation of this subunit appears to be cell cycle-regulated (16,20). The role of phosphorylation of the 34-kDa subunit in replication is unknown.

Involvement
of PCNA PCNA is present in the cell nucleus at locations that correspond to the sites of DNA synthesis, which suggested that it plays a role in DNA replication (21,22). This was further supported when SV40 DNA replication in vitro was shown to require PCNA (23, 24). PCNA was required for synthesis of the leading DNA strand, and in its absence only DNA products from the lagging strand template, 150 nucleotides long, accumulated (23,25,26). In addition, an accessory factor for DNA pol 6 (27) was found to be identical to PCNA (25,28,29). Based on these results, it was suggested that SV40 DNA synthesis was carried out on the lagging strand by DNA pol a-primase complex, and on the leading strand by PCNA and DNA ~016 (30,31).

Monopolymerase and Dipolymerase Systems
After initiation of the two nascent strands by DNA primase and limited elongation by DNA pol (Y, the switch to leading strand synthesis by pol 6 occurs. We have called this the dipolymerase phase of the replication reaction. In the SV40 replication system, Tsurimoto and Stillman (12,32) have shown that PCNA and a multi-subunit protein, RF-C, are both essential for leading strand synthesis. In the absence of either one, or both, of these proteins SV40 DNA replication a S.-H. Lee, unpublished results. SV40 DNA Replication was reduced 3-4-fold. The synthesized DNA was derived from the lagging strand, and the length of these products resembled Okazaki-sized fragments. These results indicate that both PCNA and RF-C promote leading strand synthesis.
Prior to these findings, Wobbe et al. (14) demonstrated that extensive bidirectional replication of SV40 origin-containing (ori+) DNA could be achieved by the combined action of T antigen, HSSB, topoisomerase I and the pol a-primase complex (the monopolymerase system). The monopolymerase system supported the synthesis of Okazaki-sized fragments, arising from lagging strand templates, and longer products, derived from the leading strand (9).
The laboratories of Stillman and Kelly (24,26) reported that replication reactions carried out with SV40 ori+ DNA, T antigen, topoisomerase I, HSSB, and pol cy-primase complex predominantly synthesized small Okazaki fragments, arising from the lagging strand template. This may be explained by the finding that, in the monopolymerase system, the efficient synthesis of long DNA products is dependent on high levels of pol a-primase (0.2-0.4 unit), while smaller DNA fragments are predominantly synthesized in the presence of lower levels (0.02-0.04 unit) of pol a-primase (11).
The monopolymerase system was unaffected by PCNA or the addition of antibodies that neutralized PCNA. However, DNA synthesis using crude extracts of HeLa cells was inhibited more than 90% by neutralizing antibodies to PCNA. This inhibition was substantially reversed by PCNA (33). These results prompted our laboratory to further fractionate crude extracts, selecting for DNA synthesis dependent on both PCNA and T antigen. This resulted in the isolation of three protein fractions which were called elongation inhibitor and activators I and II. These three protein factors were subsequently identified as poly(ADP-ribose) polymerase, Al (a multisubunit protein which appears to be identical to RF-C), and PCNA-dependent ~016, respectively (ll).' The monopolymerase system was blocked by the binding of poly(ADPribose) polymerase at the ends of DNA chains, resulting in the accumulation of small Okazaki fragments which arose from the lagging strand template. However, in the presence of Al, ATP, and PCNA, poly(ADP-ribose) polymeraseblocked ends were rapidly elongated by ~016.
In the absence of poly(ADP-ribose) polymerase, DNA synthesis using the dipolymerase system is dependent on the amount of pol a-DNA primase added. In the presence of low levels of pol cY-primase (0.02 unit and below), DNA synthesis with the dipolymerase system was totally dependent on Al, PCNA, and pol 8.' Under these conditions, long DNA products were formed, and virtually no Okazaki fragments were detected. This is probably due to the efficiency with which the pol &PCNA-Al system can bind to and elongate low levels of primer ends. Thus, the majority of the labeled products that accumulated were due to leading strand synthesis. In the presence of high levels of pol a-primase (0.2-0.4 unit), which can synthesize both short and long products even in the absence of the pol 6 system, the effects of pol 6, PCNA, and Al on DNA synthesis were less apparent. The addition of poly(ADP-ribose) polymerase increased the rate of synthesis of long DNA products, and this effect was especially evident after incubation periods of 5 or 10 min. However, based on these findings, poly(ADP-ribose) polymerase is not essential for leading strand synthesis. The Role of Activator 1 in DNA Synthesis The multi-subunit Al selectively bound to primer ends and increased the affinity of pol 6 for primer ends about lo-fold and decreased the amount of PCNA required for pol 6 as much as loo-fold.
Al contains an intrinsic DNA-dependent ATPase activity. However, its binding to primer ends was stimulated only 2-fold by ATP. The DNA-dependent ATPase activity of Al (RF-C) could be stimulated 3-4-fold by PCNA (34), and the further addition of HSSB stimulated the DNAdependent ATPase activity of Al about 2-fold.* The action of Al and PCNA resembles the role played by the accessory proteins that participate in the elongation of primed templates by the T4 DNA pol (35, 36) and the E. coli DNA pol III systems (13). In the T4 (37-40), E. coli (13,(41)(42)(43), and human systems, the action of the accessory proteins requires ATP hydrolysis. The effects of the accessory proteins, Al and PCNA, in the pol 6 system also require ATP. As in the prokaryotic systems, there is a precise order in the formation of the elongation complex with primed templates. The stepwise addition of the accessory proteins in these reactions is summarized in Fig. 2. In the E. coli system, the y6 subunits can bind to primed DNA, and the product can be filtered through a sizing column to separate free protein from the DNA-protein complex. In the next step, the addition of dnaN to the complex required the presence of ATP. After gel filtration, the complex supported elongation after the addition of pol III.
In the T4 phage system, the T4 DNA pol must be added to the 44/62-DNA complex prior to the addition of the gene 45 product (40). The ATP-dependent reaction is involved in the binding of the gene 45 product to the 44/62-43 (T4 DNA pol) DNA complex.
The order of addition of the accessory proteins that support the elongation of primed templates in the pol 6 system more closely resembles the E. coli system than the T4 system. The ATP-dependent step, involving the binding of PCNA to the Al-primed DNA complex, precedes the binding of pol b. Functionally, the pol b accessory proteins resemble auxiliary proteins of both prokaryotic systems. The 44162, yb, and Al (RF-C) proteins are all multi-subunit DNA-dependent ATPases. The gene 45, &UN, and PCNA proteins all require ATP for their association with other corresponding proteins.

Proceseivity in the Elongation Reaction
The T4 and E. coli prokaryotic polymerase accessory proteins convert their respective DNA polymerase from an enzyme that acts with moderate processivity to an enzyme that acts processively over long stretches of DNA. It has been postulated that accessory proteins act as clamps, increasing the association of the polymerase to primer-template (44). In the pol &catalyzed elongation reaction, PCNA is essential for DNA synthesis, and the addition of increasing levels of Al stimulated this reaction. However, the size of products was unaffected by the increased concentrations of Al. Rather, the level of nucleotide incorporation detected in the presence of PCNA and ~016 alone was low, and the amount of nucleotide incorporated was markedly increased in the presence of Al.* This suggests that while PCNA may be acting as a protein clamp to increase processivity, the role of Al is to bring about the binding of pol 6 to the primer terminus, presumably through its interaction with PCNA bound to Al. It is possible that additional factor(s) will be found that will contribute to the processivity of pal 6.

Initiation of SV40 DNA Replication
The presynthetic reactions, mediated by T antigen and leading to the unwinding of the DNA duplex, are essential for the initiation of DNA synthesis. Experiments carried out with crude HeLa cell extracts, as well as with purified proteins (T antigen, topoisomerase I, HSSB, and ATP), reveal a 10-15min lag phase before unwound DNA accumulates (3,45,56). With purified proteins, up to 70% of origin-containing duplex DNA can be unwound (18). Bullock et al. (46) examined the initiation reaction using crude extracts of HeLa cells. Preincubation of crude extracts with T antigen, ATP, and SV40 origin-containing DNA resulted in the accumulation of unwound DNA. Irrespective of the length of preincubation, pulse labeling, followed by a chase with unlabeled dNTPs, indicated that DNA regions neighboring and including the origin region served as the initial templates. This indicates that the preinitiation complex was fixed near the origin. Since there appears to be an interaction between T antigen and DNA primase-pol a complex, as well as species specificity in this interaction (47), it is possible that T antigen unwinding and the positioning of pol cY-primase complex on the DNA are linked. Other factors that might be responsible for this positioning of the preinitiation complex are unknown.

Lagging Strand Model in SV40 Replication
Models to explain the variation in size of lagging strand products have been proposed in the T4 replication system (40,44,48,49) which we find attractive for a number of reasons. In these models, as the T4 pol and accessory protein complex elongates primers on the lagging strand it remains bound to the primase-helicase complex, which acts to unwind the duplex at the replication fork. They propose that the release of the polymerase and accessory proteins from the 3'-OH end of the completed Okazaki fragment signals the initiation of a new primer by the primase-helicase in the complex. Studies with forked templates showed that the size of the Okazaki fragments increased in the presence of low concentrations of primase and polymerase (49). Similar effects have been observed in the SV40 monopolymerase and dipolymerase replication systems.
The model of Richardson et al. (40), applied to the SV40 system (Fig. 3), suggests the following features. The movement of T antigen governs the rate of chain growth on both leading and lagging strands. As the T antigen unwinds the duplex, it generates a loop on the lagging strand which is sequestered by the HSSB. In this model, the pol a-primase complex which is bound to T antigen synthesizes a short oligoribonucleotide immediately behind the T antigen on the single-stranded DNA which is devoid of HSSB. The RNA primer synthesized just behind the T antigen is then elongated by pol LY, which traverses through the HSSB. As the elongation reaction proceeds, the unwinding activity of T antigen generates a single-stranded loop of DNA. The pol a-catalyzed elongation reaction continues until the new 3'-end becomes juxtaposed behind the preceding Okazaki fragment, which signals the release of pol CY. This activates the primase to initiate the primer for the next Okazaki fragment formation.
In this model, synthesis on the leading strand is catalyzed by the pol 6 system and is limited by the migration rate of T antigen and its generation of the leading strand template. This model predicts that the rate of lagging strand synthesis and T antigen unwinding govern the size of Okazaki fragments. Since low concentrations of pol a-primase yield longer Okazaki fragments in the mono-and dipolymerase SV40 DNA replication systems, the length of the single-stranded loop on the lagging strand (Fig. 3) must be larger under these conditions, resulting in the replication of longer lengths of HSSBcoated DNA. For this to occur, it is possible that pol o(primase can dissociate from T antigen. In the presence of high levels of pol cu-primase complex, its association with T antigen is favored, in the presence of low levels, more T antigen free of pol cY-primase complex may exist, leading to the creation of longer single-stranded loops. After being reassociated with T antigen, pol cY-primase traverses these loops, resulting in longer Okazaki chains.

Perspectives on the SV40 Replication System
As recently reviewed by Prives (50), phosphorylation and dephosphorylation of T antigen plays a critical role in controlling the activity of T antigen. The origin binding proteins X0 (.52), dnaA (51), and T antigen are all required in high concentrations for activation of their well defined origin sequence. However, a specific eukaryotic origin and a T antigen equivalent have remained elusive. The discovery that structure rather than specific sequences in DNA govern primosome assembly (54, 55) and can increase the facility with which origins are activated (53) may be important in defining eukaryotic origins. Thus, the identification of the cellular equivalent of T antigen may require the identification of an origin structure-specific binding protein rather than an origin sequence-specific binding protein. :. :. i: i: 9. 10.