TrfA-dependent inner membrane-associated plasmid RK2 DNA synthesis and association of TrfA with membranes of different gram-negative hosts.

TrfA, the replication initiator protein of broad-host-range plasmid RK2, was tested for its ability to bind to the membrane of four different gram-negative hosts in addition to Escherichia coli: Pseudomonas aeruginosa, Pseudomonas putida, Salmonella enterica serovar Typhimurium, and Rhodobacter sphaeroides. Cells harboring TrfA-encoding plasmids were fractionated into soluble, inner membrane, and outer membrane fractions. The fractions were subjected to Western blotting, and the blots were probed with antibody to the TrfA proteins. TrfA was found to fractionate with the cell membranes of all species tested. When the two membrane fractions of these species were tested for their ability to synthesize plasmid DNA endogenously (i.e., without added template or enzymes), only the inner membrane fraction was capable of extensive synthesis that was inhibited by anti-TrfA antibody in a manner similar to that of the original host species, E. coli. In addition, although DNA synthesis did occur in the outer membrane fraction, it was much less extensive than that exhibited by the inner membrane fraction and only slightly affected by anti-TrfA antibody. Plasmid DNA synthesized by the inner membrane fraction of one representative species, P. aeruginosa, was characteristic of supercoil and intermediate forms of the plasmid. Extensive DNA synthesis was observed in the soluble fraction of another representative species, R. sphaeroides, but it was completely unaffected by anti-TrfA antibody, suggesting that such synthesis was due to repair and/or nonspecific chain extension of plasmid DNA fragments.

Identification of a potential membrane-targeting region of the replication initiator protein (TrfA) of broad-host-range plasmid RK2.

Plasmid RK2 codes for two species of the replication initiator protein TrfA (33 and 44 kDa). Both polypeptides are strongly associated with membrane fractions of Escherichia coli host cells (W. Firshein and P. Kim, Mol. Microbiol. 23, 1-10, 1997). We investigated the role of a 12-amino-acid hydrophobic region (HR) in the membrane association of TrfA. Epitope-tagged polypeptide fragments of TrfA that contained HR were expressed and found to be associated with membrane fractions. Site-directed mutagenesis of trfA revealed that changes of specific amino acids in HR can affect both TrfA association with the membrane and its ability to support replication of an RK2 oriV plasmid in vivo. These results are consistent with the hypothesis that membrane association of TrfA is functionally relevant and that the HR region of TrfA is involved in membrane association and DNA replication in vivo.

Probable identification of a membrane-associated repressor of Bacillus subtilis DNA replication as the E2 subunit of the pyruvate dehydrogenase complex.

Two Bacillus subtilis lysogenic libraries were probed by an antibody specific for a previously described membrane-associated inhibitor of B. subtilis DNA replication (J. Laffan and W. Firshein, Proc. Natl. Acad. Sci. USA 85:7452-7456, 1988). Three clones that reacted strongly with the antibody contained an entire open reading frame. Sequencing identified one of the clones (R1-2) as containing the E2 subunit of the pyruvate dehydrogenase complex, dihydrolipoamide acetyltransferase. An AT-rich sequence in the origin region was identified initially as the site to which extracts from the R1-2 clone were bound. This sequence was almost identical to one detected in Bacillus thuringiensis that also bound the E2 subunit but which was involved in activating the Cry1 protoxin gene of the organism, not in inhibiting DNA replication (T. Walter and A. Aronson, J. Biol. Chem., 274:7901-7906, 1999). However, the exact sequence was not as important in B. subtilis as the AT-rich core region. Binding would occur as long as most of the AT character of the core remained. Purified E2 protein obtained by use of PCR and an expression vector reacted strongly with antibody prepared against the repressor protein and the protein in the R1-2 clone, but its specificity for the AT-rich region was altered. The purified E2 protein was capable of inhibiting membrane-associated DNA replication in vitro, but anti-E2 antibody was variable in its ability to rescue repression when added to the assay.

Isolation of an inner membrane-derived subfraction that supports in vitro replication of a mini-RK2 plasmid in Escherichia coli.

Previous results have demonstrated that the inner, but not the outer, membrane fraction of Escherichia coli is the site of membrane-associated DNA replication of plasmid RK2, a broad-host-range plasmid capable of replication in a wide variety of gram-negative hosts (K. Michaels, J. Mei, and W. Firshein, Plasmid 32:19-31, 1994). To resolve the inner membrane replication site further, the procedure of Ishidate et al. (K. Ishidate, E. S. Creeger, J. Zrike, S. Deb, G. Glauner, T. J. MacAlister, and L. I. Rothfield, J. Biol. Chem. 261:428-443, 1986) was used to separate the inner membrane into a number of subfractions, of which only one, a small subfraction containing only 10% of the entire membrane, was found to synthesize DNA inhibited by antibody prepared against the plasmid-encoded initiation protein TrfA. This is the same subfraction that was also found to bind oriV and TrfA to the greatest extent in filter binding assays (J. Mei, S. Benashski, and W. Firshein, J. Bacteriol. 177:6766-6772, 1995).

Plasmid replication and partition in Escherichia coli: is the cell membrane the key?

The DNA-membrane complex has been the subject of intensive investigation for over 35 years as the possible site for DNA replication in the prokaryotic cell and the site through which newly synthesized chromosomes are segregated into daughter cells. However, the molecular mechanisms which control these phenomena are, for the most part, poorly understood despite genetic, biochemical, and morphologic evidence in favour of their existence. This is probably due to the transient nature and non-covalent interactions that occur between DNA and the membrane. In addition, there is a paucity of knowledge concerning the nature of the membrane receptors for DNA and whether the membrane plays simply a structural or metabolic role in the two processes. Plasmids can provide important insights into the role of the membrane in replication and partitioning because the plasmid life cycle is relatively simple, with replication occurring during the cell cycle and partitioning during cell division. The replicon model of Jacob et al. (1963, Cold Spring Harbor Symp Quant Biol 28: 329-348) still represents a good conceptual framework (with modifications) to explain how plasmid replication and partitioning are linked by the membrane. In its simplest form, the model focuses on specific membrane binding sites (possibly along the equator of the cell) for plasmid (or bacterial) replication, with the membrane acting as a motive force to separate the newly synthesized replicons and their attached sites into daughter cells. Indeed, proteins involved in both plasmid replication and partitioning have been found in membrane fractions and some plasmids require membrane binding for initiation and an active partitioning. We propose that several factors are critical for both plasmid DNA replication and partitioning. One factor is the extent of negative supercoiling (brought about by an interplay of various topoisomerases, but most importantly by DNA gyrase). Supercoiling is known to be critical for initiation of DNA replication but may also be important for the formation of a partition complex in contact with the cell membrane. Another factor is the presence of specific subdomains of the membrane which can interact specifically with origin DNA and possibly other regions involved in partitioning. Such domains may be induced transiently or be present at all times during the cell cycle.

Interactions of the origin of replication (oriV) and initiation proteins (TrfA) of plasmid RK2 with submembrane domains of Escherichia coli.

It has been possible to locate a submembrane domain representing less than 10% of the total membrane that appears to be responsible for sequestering some essential components required for plasmid RK2 DNA replication. This subfraction, whose cellular location in the membrane prior to extraction is still unknown, is derived from the inner membrane fraction, since it possesses enzyme marker activity (NADH oxidase) exclusively associated with the inner membrane. The subfraction was detected by a modification of the methods of Ishidate et al. (K. Ishidate, E. S. Kreeger, J. Zrike, S. Deb, B. Glauner, T. MacAlister, and L. I. Rothfield, J. Biol. Chem. 261:428-443, 1986) in which low pressure in a French pressure cell and lysozyme were used to preserve the supercoil plasmid DNA template during cell disruption. This was followed by successive cycles of sucrose gradient sedimentation and flotation density gradient centrifugation to reveal a number of subfractions, including the one of interest. The characteristics of plasmid interaction with the subfraction include the presence of supercoil DNA after extraction, the binding of the origin of plasmid replication (oriV) in vitro, and the association of the two plasmid-encoded initiation (TrfA) proteins (encoded by overlapping genes). However, another peak, the outer membrane fraction, also binds oriV in vitro, contains plasmid DNA in vivo, and associates with the TrfA initiation proteins. Nevertheless, it contains much less of the initiation proteins, and the specific activity of binding oriV is also much reduced compared with the other subfraction. There is a strong correlation between the association of the TrfA initiation proteins with a particular membrane fraction and the binding of oriV in vitro or plasmid DNA in vivo. Since the proteins are known to bind to repeated sequences in oriV (S. Perri, D. R. Helinski, and A. Toukdarian, J. Biol. Chem. 266:12536-1254, 1991; M. Pinkney, R. Diaz, E. Lanka, and C. M. Thomas, J. Mol. Biol. 203: 927-938, 1988), it appears that the initiation proteins themselves could be responsible, at least in part, for the association of plasmid DNA to the membrane.

TrfA-dependent, inner-membrane-associated plasmid RK2 DNA synthesis in Escherichia coli maxicells.

Previous results of experiments in which plasmid-encoded proteins were selectively labeled in ultraviolet sensitive "maxicell" mutants suggested that the essential initiation proteins of RK2 (33 and 43 kDa) were bound to the inner membrane of Escherichia coli (D. Kostyal et al., 1989, Plasmid 21, 226-237). However, in the present studies using a specific polyclonal antibody against the TrfA initiation proteins, significant levels of these proteins were also detected for the first time in the outer membrane fraction as well as the inner membrane fraction. Only in the cytosol fraction were the initiation proteins relatively absent. In order to determine whether initiation and replication were also associated with either or both of the membrane fractions, it was necessary to develop a replicating system more active than the one previously extracted from minicell membranes, which did not separate the membrane into its component parts (J. A. Kornacki and W. Firshein, 1986, J. Bacteriol. 167, 319-326). In addition, it was also necessary to devise an extraction procedure that did not degrade the supercoil DNA template during the separation of the inner from the outer membrane fraction. Both criteria were met, first by the use of maxicells containing a miniplasmid derivative of RK2 and second by disrupting cell envelopes in the French pressure cell using low pressure. Under these conditions, not only were the two major membrane fractions separated successfully from the cytosol fraction, but supercoil DNA template was also preserved in both fractions, detergents were avoided, and replication was significantly higher than that described in the earlier experiments. TrfA-dependent initiation of DNA replication was associated primarily with the inner membrane fraction.

Replication of a Bacillus subtilis oriC plasmid in vitro.

We constructed an in vitro replication system specific for a Bacillus subtilis oriC plasmid using a soluble fraction derived from cell extracts of B. subtilis. DNA polymerase III and two initiation proteins, DnaA and DnaB, were required for in vitro replication as observed in vivo. Both upstream and downstream DnaA box regions of the dnaA gene were required as cis-elements for in vitro synthesis of the B. subtilis oriC plasmid as well as for in vivo activity. The replication was semi-conservative and only one round of replication occurred within 15 min. These results indicate that in vitro replication faithfully reproduced in vivo replication. To elucidate the site of initiation and the direction of replication, we analysed replicative intermediates generated in vitro in the presence of various concentrations of ddGTP by two methods. First, analysis of restriction fragments around the dnaA gene showed a high level of incorporation of the radioactive substrate, indicating that replication began within the vicinity of the dnaA gene. Second, using 2-dimensional gel electrophoresis, bubble arcs were detected only on fragments containing the DnaA box region downstream of the dnaA gene, indicating that the initiation site resided within this region. The distribution of the bubble arcs suggested that both bidirectional and undirectional replication occurred in vitro.

Temporal expression of a membrane-associated protein putatively involved in repression of initiation of DNA replication in Bacillus subtilis.

A Bacillus subtilis membrane-associated protein that binds specifically to the origin region of DNA replication may act as an inhibitor of DNA replication (J. Laffan and W. Firshein, Proc. Natl. Acad. Sci. USA 85:7452-7456, 1988). This protein, originally estimated to be 64 kDa, had a slightly lower molecular size (57 kDa), as determined by sodium dodecyl sulfate-polyacrylamide gel electrophoresis during these studies. The size difference may be due to processing that results in modification of the protein. The protein can be extracted from both cytosol and membrane fractions, and the amounts in these fractions vary during the developmental cycle of B. subtilis. A complex pattern of expression in which significant levels were detected in spores was revealed; levels decreased dramatically during germination and increased after the first round of DNA replication. The decrease during germination was due to protease activity, as demonstrated by the addition of protease inhibitors and radioactive-labeling chase experiments. During vegetative growth, the protein levels increased until stationary phase, after which there was another decrease during sporulation. The decrease during sporulation may be partially due to sequestering of the protein into forespores, since as the putative repressor protein decreased in the mother cell, it increased in the forespores. However, protease activity was also involved in the decrease in the mother cell. The changes in expression of this protein are consistent with its role as a repressor of initiation of DNA replication. Additional studies, including sequence analysis and further antibody analysis, show that this protein is not a subunit of the pyruvate dehydrogenase complex. This relationship had been a possibility based upon the results of others (H. Hemila, A. Pavla, L. Paulin, S. Arvidson, and I. Palva, J. Bacteriol. 172:5052-5063, 1990).

Characterization of a multienzyme complex derived from a Bacillus subtilis DNA-membrane extract that synthesizes RNA and DNA precursors.

The activity of a variety of enzymes involved in the synthesis of RNA and DNA precursors was found to copurify with initiation of DNA replication activity. These enzymes included ribo- and deoxyribonucleoside kinases, kinases for their phosphorylated intermediates, and ribonucleoside diphosphate reductase. This precursor-synthesizing complex is part of a Bacillus subtilis DNA-membrane extract originally shown to contain all of the enzymes and template necessary for initiation of DNA replication (J. Laffan and W. Firshein, J. Bacteriol. 169:2819-2827, 1987). Although the complex incorporated deoxyribonucleoside triphosphates into DNA, deoxyribonucleosides were incorporated even faster, suggesting catalytic facilitation. Both ribonucleosides and deoxyribonucleosides were found by thin-layer chromatography separation to be converted by the complex into their mono-, di-, and triphosphate derivatives. Ribonucleotides were incorporated into DNA via the action of ribonucleoside diphosphate reductase. Some regulatory mechanisms of the kinase system may also be retained by the complex. Electron microscope studies revealed that the precursor-synthesizing-initiation subcomplex is contained within a particulate fraction consisting of different-size vesicles resembling liposomes and that these particles may be structurally important in maintaining the synthetic activity of the subcomplex.

Replication of an RK2 miniplasmid derivative in vitro by a DNA/membrane complex extracted from Escherichia coli: involvement of the dnaA but not dnaK host proteins and association of these and plasmid-encoded proteins with the inner membrane.

A DNA/membrane complex extracted from a miniplasmid derivative of the broad host range plasmid RK2 cultured in Escherichia coli capable of synthesizing new plasmid supercoiled DNA in vitro was treated with antibodies that were made against or reacted with the dnaA and dnaK host-encoded proteins, respectively. Anti-dnaA protein antibody inhibited total plasmid DNA synthesis significantly and the synthesis of supercoil plasmid DNA almost completely. In contrast, anti-dnaK protein antibody and nonimmune serum had little or no effect on total plasmid DNA synthesis. Both proteins were found to be present in the inner but not outer membrane fraction of E. coli. A variety of miniplasmid-encoded proteins which had previously been found in the DNA/membrane complex have also been localized to the inner but not outer membrane fraction. These include an essential initiation protein of 32 kDa (and an overlapping protein of 43 kDa coded for by the same gene), as well as a 30-kDa protein that may be linked to incompatibility functions. Various extraction methods were used to distinguish between the associated and the integral nature of the plasmid-encoded proteins. The results demonstrated that the essential replication proteins (32 and 43 kDa) as well as the 30-kDa protein was tightly bound to the inner membrane. Computer analysis of the amino acid sequence of the 32 (and 43)-kDa protein revealed a hydrophobic region that is only half that normally required to span the membrane. Other interactions are discussed with respect to attaching this protein to the membrane.

Origin-specific DNA-binding membrane-associated protein may be involved in repression of initiation of DNA replication in Bacillus subtilis.

Previous binding studies with labeled double-stranded Bacillus subtilis DNA fragments to a protein blot of renatured Bacillus membrane proteins showed selective binding of two adjacent origin fragments to a 64-kDa protein. The selective binding of the 64-kDa protein could be blocked by prior incubation of the blots with a specific polyclonal antibody. An in vitro replication system derived from a B. subtilis DNA-membrane complex showed initiation activity without addition of exogenous enzymes or template. When the complex was first incubated with the 64-kDa antibody or with its Fab fragments, initiation activity was enhanced. Antibodies to several other Bacillus membrane proteins as well as nonspecific antibodies did not show any significant stimulatory effect. A heavy-density-label experiment indicated that the complex initiated multiple rounds of replication in the presence of the 64-kDa antibody but not in its absence. The 64-kDa antibody plus an initiation inhibitor (streptovaricin) showed only repair and elongation activity. The 64-kDa protein may act in vivo as a repressor/regulator of initiation activity.

Membrane protein binding to the origin region of Bacillus subtilis.

Binding of membrane proteins extracted from Bacillus subtilis to an 11.6-kilobase region containing the origin of replication was examined by Western blotting (protein blotting) procedures. Two adjacent origin probes in the double-stranded form (spanning a length of 4 kilobases) were found to bind very strongly to a 63-kilodalton (kDa) protein in that they resisted dissociation after a high-concentration salt wash. This region encompasses both a site implicated in initiation in vivo and a gene coding for a DNA gyrase subunit (gyrA). In contrast, flanking origin and nonorigin double-stranded probes were dissociated after washing with a high salt concentration. Another protein of 67 kDa bound less intensely to the putative initiation site but not to the gyrA region. All of the origin and nonorigin probes in the double- or single-stranded form were found to bind nonspecifically to a subset of 10 to 12 proteins of 50 to 60 separated by gel electrophoresis after a low-concentration salt wash. They ranged in size from 14 to over 100 kDa (including 63 kDa). However, in contrast to the double-stranded forms, most of the single-stranded probes resisted dissociation from the protein subset after a high-concentration salt wash.

DNA replication by a DNA-membrane complex extracted from Bacillus subtilis: site of initiation in vitro and initiation potential of subcomplexes.

A DNA-membrane complex extracted from Bacillus subtilis was studied further as a model system for initiation of bacterial DNA replication in vitro. Of three subcomplexes purified from the crude complex by a combination of CsCl and sucrose gradient centrifugation, the synthetic capability of only one was inhibited significantly by streptovaricin, a known inhibitor of RNA primer formation. A selective enrichment in the level of this subcomplex was obtained by manipulating a thymine-requiring mutant. The synthetic capabilities of an enriched and nonenriched DNA-membrane complex were compared in the presence and absence of streptovaricin. Although the rate and extent of DNA synthesis per unit of protein were approximately the same in the absence of the antibiotic, there was a much greater inhibition of synthesis shown by the enriched complex in the presence of streptovaricin. Although the amount of DNA present in the putative initiation subcomplex was less than 0.3 to 0.4% of the total DNA present in the crude complex, such DNA, except for a few quantitative differences, was still representative of genomic DNA. Newly synthesized DNA hybridized to specific origin- and non-origin-derived restriction fragments of the B. subtilis genome. However, when an elongation inhibitor (ddCTP) was added, hybridization of such DNA to almost all of the nonorigin fragments disappeared or was reduced drastically, whereas origin region hybridization patterns remained strong. The highest level of hybridization in the origin region occurred with a BamHI (B7) restriction fragment of 5.6 kilobases that has been implicated by others as one site initiation in vivo (N. Ogasawara, M. Seiki, and H. Yoshikawa, Nature (London) 281:702-704, 1979; S. J. Seror-Laurent and G. Henckes, Proc. Natl. Acad. Sci. USA 82:3586-3590, 1985).

Replication of plasmid RK2 in vitro by a DNA-membrane complex: evidence for initiation of replication and its coupling to transcription and translation.

The following results with an in vitro replication system utilizing a plasmid RK2 DNA-membrane complex indicate that the essential trfA-encoded replication protein of RK2 is present and active in the complex. (i) A complex extracted from a conditional replication mutant of RK2, which contains a temperature-sensitive mutation in trfA, displayed extensive DNA synthesis at the permissive temperature but little activity at the restrictive temperature. A control wild-type RK2 complex showed no inhibition of DNA synthesis at the restrictive temperature. (ii) Analysis of plasmid-encoded proteins revealed that the trfA-specified replication protein and other proteins which may be involved in the replication and maintenance of RK2 are located physically in the complex. Semiconservative plasmid DNA replication by the DNA-membrane complex was indicated by density shift experiments; DNA synthesized in the presence of a heavy-density precursor banded primarily in a heavier-density area of a neutral CsCl density gradient and consisted mostly of heavy- and light-density single-stranded DNA as determined by alkaline CsCl density gradient centrifugation. Plasmid RK2 DNA replication by the DNA-membrane complex appears to be coupled to transcription and translation as indicated by the following results: the inhibitory effects of chloramphenicol on both DNA and protein synthesis by the complex; the stimulation of replication by components normally required for protein synthesis (tRNA and all the common amino acids); the synthesis of RNA and protein by the complex; and the synthesis of specific RK2-encoded proteins.

Gene control in broad host range plasmid RK2: expression, polypeptide product, and multiple regulatory functions of korB.

The korB gene of broad host-range plasmid RK2 prevents host-cell lethality by kilB and negatively controls RK2 replication. We precisely mapped the limits of korB to a region near korA, an autoregulated gene involved in control of several RK2 genes. The following results show that korA and korB are cotranscribed from the korA promoter: Mutants deleted for the korA promoter fail to express korB, even with korA function supplied in trans; the korA promoter is nonessential to korB if a heterologous promoter is present; and RNA produced in vivo has both korA- and korB-specific sequences. Analysis of polypeptides synthesized from wild-type and mutant korB plasmids in maxicells revealed that korB encodes a 52-kDa polypeptide, whose activity is extremely sensitive to changes in its carboxyl terminus but relatively unaffected by replacement of its amino terminus. The minimal korB-encoding region allowed us to identify two new regulatory functions, both of which duplicate previously known functions of korA. First, korB alone was found to control the kilB1 component of kilB, thus resolving the paradox of korA-independent control of kilB. Second, analysis of polypeptides from the korA-korB region in the presence and absence of korB, and studies with the korA promoter fused to the chloramphenicol acetyltransferase structural gene (cat) showed that korB, like korA, autoregulates expression of the korA-korB operon. We suggest that korA and korB gene products act as co-repressors in the control of certain RK2 genes.

Detection of displacement ("D") loops with the properties of a replicating intermediate synthesized by a DNA/membrane complex derived from the low-copy-number plasmid RK2.

A significant fraction of the plasmid DNA extracted from an RK2 miniplasmid DNA/membrane complex after incubation in vitro with appropriate substrates and cofactors contains "D" or displacement loops in one of at least three loci in the genome. These include the origin of replication and two transposon regions that code for kanamycin and tetracycline resistance. The relationship of these loops to early replicating intermediates for plasmid and transposon replication is discussed.

Proteins encoded by the trans-acting replication and maintenance regions of broad host range plasmid RK2.

The broad host range plasmid RK2 has previously been found to contain three separate regions of the genome involved in replication and maintenance in Escherichia coli (C. M. Thomas, R. Meyer, D. R. Helinski, 1980, J. Bacteriol. 141, 213-222). They include the origin of replication (oriRK2) and the trfA region which encodes a trans-acting function required for replication. The third region (trfB), although not essential for replication, supplies a function involved in the maintenance of plasmid RK2. Using the maxicell system of labeling plasmid-specific proteins, we have identified all of the proteins encoded by two miniplasmid derivatives of RK2 which contain only the regions oriRK2, trfA, and trfB. To determine which region specifies each protein, RK2/mini-ColE1 hybrid plasmids were used which contain various restriction fragments of the mini-RK2 replicon. The trfA region appears to encode three proteins designated A1 (39,000 MW), A2 (31,000 MW), and A3 (14,000 MW). Analysis of proteins synthesized by plasmids containing deleted forms of the trfA region indicates that the A2 protein is the essential trfA-encoded replication protein of plasmid RK2. The proteins A1 and A3 may be the products specified by the genes tra3 (involved in transmissibility) and kilB1 (involved in host-cell viability) which also map in the trfA region. The trfB region specifies two proteins designated B1 (36,000 MW) and B2 (30,000 MW). These may be the products of the two kil-override (kor) genes located in the trfB region which have been implicated in plasmid maintenance.

Initiation of DNA replication in vitro by a DNA-membrane complex extracted from Bacillus subtilis.

Initiation of DNA replication has been observed in vitro with a DNA-membrane complex extracted from Bacillus subtilis. Antibiotics known to interfere with various aspects of initiation inhibited DNA synthesis significantly in vitro, whereas a mutant resistant to one inhibitor failed to respond to its presence. The inhibitory effects occurred primarily when the immediate RNA precursors (ribonucleoside triphosphates) were present in the assay solution but not significantly when the precursors were omitted. Complexes extracted from a temperature-sensitive initiation mutant were almost incapable of synthesizing DNA at the restrictive temperature but displayed extensive synthesis at the permissive temperature. A strong indication of semiconservative DNA synthesis was obtained in vitro after density-shift experiments involving incubation of the complex with a heavy-density DNA precursor, followed by neutral and alkaline CsCl density gradient centrifugation. A significant amount of chain elongation or repair (or both) was also observed.