The Bacillus subtilis Phage (cid:1) 29 Protein p16.7, Involved in (cid:1) 29 DNA Replication, Is a Membrane-localized Single-stranded DNA-binding Protein*

The functional role of the (cid:1) 29-encoded integral membrane protein p16.7 in phage DNA replication was studied using a soluble variant, p16.7A, lacking the N-ter-minal membrane-spanning domain. Because of the pro-tein-primed mechanism of DNA replication, the bacteriophage (cid:1) 29 replication intermediates contain long stretches of single-stranded DNA (ssDNA). Protein p16.7A was found to be an ssDNA-binding protein. In addition, by direct and functional analysis we show that protein p16.7A binds to the stretches of ssDNA of the (cid:1) 29 DNA replication intermediates. Properties of protein p16.7A were compared with those of the (cid:1) 29-encoded single-stranded DNA-binding protein p5. The results obtained show that both proteins have different, non-over-lapping functions. The likely role of p16.7 in attaching (cid:1) 29 DNA replication intermediates to the membrane of the infected cell is discussed. Homologues of gene 16.7 are present in (cid:1) 29-related phages, suggesting that the proposed role of p16.7 is conserved in this family of phages.

The functional role of the 29-encoded integral membrane protein p16.7 in phage DNA replication was studied using a soluble variant, p16.7A, lacking the N-terminal membrane-spanning domain. Because of the protein-primed mechanism of DNA replication, the bacteriophage 29 replication intermediates contain long stretches of single-stranded DNA (ssDNA). Protein p16.7A was found to be an ssDNA-binding protein. In addition, by direct and functional analysis we show that protein p16.7A binds to the stretches of ssDNA of the 29 DNA replication intermediates. Properties of protein p16.7A were compared with those of the 29-encoded single-stranded DNA-binding protein p5. The results obtained show that both proteins have different, non-overlapping functions. The likely role of p16.7 in attaching 29 DNA replication intermediates to the membrane of the infected cell is discussed. Homologues of gene 16.7 are present in 29-related phages, suggesting that the proposed role of p16.7 is conserved in this family of phages.
Studies on DNA replication and related processes have provided detailed insights in the function of many proteins involved in these processes (for review, see Ref. 1). Despite this, little is known about the in vivo organization of DNA replication. To gain a better insight in this fundamental process, we studied the in vivo DNA replication of the Bacillus subtilis bacteriophage 29 (2). The detailed knowledge of its in vitro mechanism of DNA replication (for reviews, see Refs. 3 and 4) made 29 an attractive system for this study.
The genome of 29 is a linear double-stranded DNA (dsDNA) 1 of 19,285 base pairs that contains a terminal protein (TP) covalently linked at each 5Ј end. Fig. 1A shows a schematic representation of the genetic and transcriptional organization of the 29 genome. Regulation of 29 DNA transcription, which can be divided into an early and a late stage, has been studied extensively in vivo as well as in vitro (for reviews, see Refs. 3 and 5). The late expressed genes, all transcribed from a single operon present in the central part of the genome, encode the structural proteins of the phage, proteins involved in morphogenesis, and those required for lysis of the infected cell. The early expressed genes are present in two operons that flank the late operon. The early operon located at the left side of the 29 genome encodes the transcriptional regulator protein p4 and various proteins that are directly involved in phage DNA replication, such as the DNA polymerase, TP, singlestranded DNA-binding protein (SSB), double-stranded DNAbinding protein, and protein p1. The operon located at the right side of the 29 genome encodes, in addition to proteins p17 and p16.7, four putative proteins of unknown function.
A schematic overview of the in vitro 29 DNA replication mechanism is shown in Fig. 1B. Initiation of 29 DNA replication occurs via a so-called protein-primed mechanism (reviewed in Refs. 3, 4, and 6). The TP-containing DNA ends constitute the origins of replication. Initiation of DNA replication starts by recognition of the origin by a heterodimer formed by the 29 DNA polymerase and the primer TP. The DNA polymerase then catalyzes the addition of the first dAMP to the primer TP. Next, after a transition step, these two proteins dissociate, and the DNA polymerase continues processive elongation until replication of the nascent DNA strand is completed. Replication, which starts at both DNA ends, is coupled to strand displacement. This results in the generation of socalled type I replication intermediates consisting of full-length double-stranded 29 DNA molecules with one or more singlestranded DNA (ssDNA) branches of varying lengths. When the two converging DNA polymerases merge, a type I replication intermediate becomes physically separated into two type II replication intermediates. Each of these consists of a fulllength 29 DNA molecule in which a portion of the DNA, starting from one end, is double-stranded, and the portion spanning to the other end is single-stranded.
Over the last decades convincing evidence has been presented that replication of bacterial genomes, including that of resident plasmids and infecting phages, occurs at the bacterial cell membrane (for review, see Ref. 7). Also, replication of 29 DNA takes place at the membrane of the infected cell (2,8,9). Gene 16.7, present in the early expressed operon located at the right side of the 29 genome (see Fig. 1A), encodes an integral membrane protein of 130 amino acids. The efficiency of in vivo 29 DNA replication is affected in the absence of protein p16.7 (10). In this work we analyzed the functional role of p16.7 in 29 DNA replication using purified p16.7A, a soluble variant of p16.7 lacking the N-terminal transmembrane-spanning domain. We found that protein p16.7A can functionally substitute the 29 SSB p5 in in vitro 29 DNA amplification assays, suggesting that it is a ssDNA-binding protein. This inference was further supported by direct assays such as electron microscopy and gel retardation studies. Thus, in addition to a classical SSB p5, 29 encodes a membrane-localized ssDNA-binding protein. Contrary to the SSB p5, p16.7A has no helix-destabilizing activity, and p16.7 is not synthesized in stoichiometric amounts in infected cells. These and other results show that p16.7 and SSB p5 have non-overlapping functions. Based on the properties determined in this work together with the features described before, it is most likely that p16.7 attaches 29 DNA to the membrane of the infected cells by binding to the stretches of ssDNA present in the replication intermediates.
PCR Techniques-PCR reactions were carried out with proofreadingproficient Vent DNA polymerase (New England Biolabs, Beverly, MA) using conditions as described before (10). Oligonucleotides were purchased from Isogen Bioscience BV (Maarsen, The Netherlands).
Overexpression and Purification of p16.7A-Protein p16.7A was overexpressed and purified using a Ni 2ϩ -nitrilotriacetic acid resin column as described before (10). 29 Protein-primed Initiation, TP-DNA Replication, and Amplification Assays-These assays were performed as described before (14). The reaction mixtures of the 29 TP-DNA replication assays contained the indicated amount of protein p16.7A or SSB p5. The amplification assays were stopped after an incubation period of 20 min.
Psoralen Cross-linking and Spreading of DNA Molecules for Electron Microscopy-Replication reactions were carried out in the absence or in the presence of p16.7A or SSB p5 as described below. After 30 min at 30°C the samples were stopped by adding 0.05 volumes of 4,5Ј,8trimethylpsoralen (200 g/ml in 100% ethanol) on ice. The samples were then irradiated with 366-nm UV light on ice for 1 h with 2 psoralen additions, as described by Sogo and Thoma (15). These cross-linking conditions were sufficient to produce essentially complete cross-linking of the DNA molecules in the absence of protein. After psoralen crosslinking, the samples were digested with proteinase K (500 g/ml) for 2 h at 56°C and extracted with phenol, and the DNA was precipitated with ethanol. Denaturation and spreading of the psoralen cross-linked DNA for electron microscopy were carried out according the BAC technique, as described by Sogo et al. (16). Electron micrographs were taken with a Philips 420 electron microscope at 80 kV, routinely at a magnification of 20,000-fold.
Gel Mobility Shift Assays-Unless stated otherwise, the incubation mixtures contained, in a final volume of 20 l, 25 mM Hepes, pH 7.5, 4% Ficoll 400, 1 mM EDTA, 0.1 mg/ml bovine serum albumin, 10 mM dithiothreitol, the indicated labeled DNA fragment, and the indicated amount of protein p16.7A or 29 SSB p5. After incubation for 10 min at 4°C, the samples were subjected to electrophoresis in 4% non-denaturing polyacrylamide (80:1) gels containing 12 mM Tris acetate, pH 7.5, and 1 mM EDTA and run at 4°C using a running buffer containing 12 mM Tris acetate, pH 7.5, and 1 mM EDTA at 70 V for 6 h. Next, the gels were dried and autoradiographed.
Glycerol Gradients-Twenty g of protein p16.7A was subjected to a 15-30% linear glycerol gradient containing 50 mM Tris-HCl, pH 7.5, 1 mM EDTA, and 7 mM ␤-mercaptoethanol and run as described before (14). After fractionation of the gradient, aliquots of each fraction were analyzed by SDS-PAGE and gel retardation assays.
Oligonucleotide Degradation and Helix Destabilization Assays-Oligonucleotide degradation assays and the helix destabilization assays, the latter using 125 ng of primed M13mp18 DNA, were carried out as described by Gascón et al. (17). Oligonucleotide Ϫ40 Universal and the same with a 5Ј (5 thymidines) or a 3Ј (5 adenines) extension were end-labeled with T4-polynucleotide kinase using [␥-32 P]ATP for 1 h at 37°C.
Preparation of Crude Cell Extracts, Western Blot, and Quantification of p16.7A and SSB p5-To determine the intracellular levels of the viral proteins p16.7 and SSB throughout the course of the 29 infection cycle, B. subtilis cells were infected with phage 29 mutant sus14(1242) as described above. At different times after infection, 1.5-ml samples were withdrawn and processed as described before (10). Known amounts (ng) of the corresponding purified proteins (p16.7A or SSB p5) were run in the same gel to determine the standard curve. Polyclonal antisera from rabbits against 29 p16.7 or against 29 SSB p5 were diluted 2,500 and 1,000 times, respectively.

Protein p16.7A Has No Effect On in Vitro 29 DNA Initiation of Replication-Previous results
show that the absence of protein p16.7 affects the efficiency of in vivo 29 DNA replication, especially at early infection times (10). Protein p16.7 might stimulate 29 DNA replication by enhancing the rate-limiting step of initiation of DNA replication. To study this possibility in vitro, 29 DNA replication initiation assays were performed in the absence or presence of increasing amounts of purified p16.7A. Although efficient initiation requires the presence of 29 template TP-DNA, the DNA polymerase is also able to carry out the TP-deoxynucleotidylation reaction in the absence of TP-DNA (18). Therefore, in vitro 29 DNA initiation experiments were performed in reaction mixtures either lacking or containing template TP-DNA and in the absence or presence of increasing amounts of protein p16.7A. No significant effect of p16.7A on the initiation reactions was obtained (not shown), indicating that p16.7A plays no role in the in vitro initiation of 29 DNA replication.
Protein p16.7A Can Functionally Substitute the SSB p5 in 29 DNA Amplification Assays-The possibility that p16.7A has a role in 29 DNA replication after the initiation step was analyzed by studying the possible effects of protein p16.7A in an in vitro 29 DNA amplification system. This system, which allows the amplification of very low amounts of 29 template DNA, requires the following four 29-encoded proteins: DNA polymerase, TP, double-stranded DNA-binding protein p6, and the SSB p5 (19). It has been described that omission of SSB p5 from the reaction mixtures results in the generation (19) and amplification (20) of short 29 DNA products. Esteban et al. (20) demonstrated that the short 29 DNA products are of palindromic nature and that they are caused by a DNA polymerase template-switching event during replication. Binding of SSB p5 to the displaced stretches of ssDNA present in type I DNA replication intermediates avoids the DNA polymerase to switch template, thus preventing the generation of short DNA products. Interestingly, we found that the addition of increasing amounts of protein p16.7A to reaction mixtures lacking SSB p5 resulted in the synthesis of increasing amounts of full-sized 29 DNA and a concomitant decrease in the amounts of short DNA products (Fig. 2). In fact, the generation of the short DNA products was fully prevented when the reaction mixtures contained 14 or 28 M protein p16.7A. These results show that, under the conditions tested, the presence of protein p16.7A prevents the DNA polymerase from switching template and, therefore, that it can functionally substitute the SSB p5. The most likely explanation for these results is that the p16.7A protein binds to the stretches of displaced ssDNA present in the type I replication intermediates. The amount of full-sized 29 DNA synthesized in the presence of 14 or 28 M protein p16.7A is slightly lower than that compared in the presence of 7 M. Probably, this decrease is due to binding of protein p16.7A to double-stranded 29 template DNA at these elevated p16.7A concentrations (see below), causing a small effect on the efficiency of the 29 DNA amplification.
Protein p16.7A Protects the Displaced ssDNA from Psoralen Cross-linking-Psoralen can cross-link portions of a dsDNA molecule that are free of protein (15) as well as folded-back regions in ssDNA (21). The presence of the 29 SSB p5 was shown to prevent the displaced ssDNA of the 29 DNA replication intermediates from psoralen cross-linking, demonstrating that it binds to ssDNA (22)(23)(24). To study whether protein p16.7A indeed binds to the ssDNA portions of 29 DNA replication intermediates, as indicated by the results of the 29 DNA amplification assays described above, we used the psoralen cross-linking technique. Thus, 29 DNA replication reactions, carried out in the absence or presence of either protein p16.7A or SSB p5, were treated with psoralen as described under "Materials and Methods." Next, after the reaction products were treated with proteinase K, extracted with phenol, and purified, they were spread under denaturing conditions and analyzed by electron microscopy. As expected, in all three samples full-length dsDNA molecules with one or more ssDNA tails (type I) and full-length DNA molecules formed by a dsDNA portion of variable length from one DNA end plus an ssDNA portion spanning to the other DNA end (type II) were observed. A representative type I replication intermediate of each sample is shown in Fig. 3. As described before (22)(23)(24), the displaced ssDNA regions of replication intermediates produced in the absence (Fig. 3A) or the presence of SSB p5 (Fig. 3C) appeared as collapsed and well unfolded structures, respectively. Fig. 3B shows that the ssDNA portions in replication intermediates produced in the presence of protein p16.7A also had a well unfolded structure. These results, therefore, show that protein p16.7A, like the SSB p5, prevents the ssDNA from psoralen cross-linking, demonstrating that it binds to the ssDNA portions of the replication intermediates generated during 29 DNA replication.
Protein p16.7A Has Higher Affinity for ssDNA Than the SSB p5-Binding of p16.7A to ssDNA was further analyzed by gel mobility shift assays. For this purpose, the 175-base pair right- end fragment of the 29 genome was end-labeled with 32 P (see "Materials and Methods"), heat-denatured, and used in gel retardation assays in the absence or presence of increasing amounts of purified protein p16.7A. Although retardation of the ssDNA fragment was observed in the presence of 45 nM p16.7A, full retardation of all the ssDNA molecules required a p16.7A concentration of 360 nM (Fig. 4A). Similar results were obtained with various other DNA fragments (results not shown). To gain an insight in the global ssDNA binding of p16.7A compared with that of the well studied 29 protein p5, the experiment shown in Fig. 4A was carried out in parallel using purified SSB p5. Whereas in agreement with previously published results (25)(26)(27), some retardation of part of the ssDNA molecules occurred in the presence of 3.3 M SSB p5, full retardation of all the ssDNA molecules required a concentration of 13.3 M (Fig. 4B). Together, these results show that the ssDNA binding activity of p16.7A is about 50 times higher than that of the 29 SSB. To study possible binding of protein p16.7A to dsDNA, the same fragment of the 29 genome used in Fig. 4, A and B, but in its double-stranded form, was used in gel retardation assays. The results, presented in Fig. 4C, show that p16.7A binds to dsDNA, although the amount of p16.7A required to obtain full retardation of all the dsDNA molecules was about 20-fold higher than that required to bind the same DNA molecules in their single-stranded form. The observation that a smear of retarded DNA species is observed at low p16.7A concentrations (45-180 and 180 -720 nM for ss-and dsDNA, respectively) indicates that the nucleoprotein complex formed at these concentrations is rather unstable. In addition, the increasing mobility shift caused by the various concentrations of p16.7A protein analyzed suggests the binding of more than one protein molecule per DNA molecule. This inference is further supported by the findings that a considerable amount of protein p16.7A is required (i) to prevent the generation of short ssDNA products in amplification assays lacking the SSB p5 (see above) and (ii) to cover the complete circular M13 ssDNA in order to prevent binding of the 29 DNA polymerase (see below).
To confirm that the observed retardation was due to protein p16.7A and not to a possible minor contaminant, the purified p16.7A protein was subjected to a glycerol gradient, after which aliquots of the gradient were analyzed for ssDNA binding by gel retardation. The results presented in Fig. 4D show that the ssDNA binding activity was restricted to those fractions that contained protein p16.7A.
Protein p16.7A Binds to Circular ssDNA-The following approach was used to study possible binding of p16.7A to circular ssDNA. The 29 DNA polymerase has strong affinity for naked ssDNA (28) but does not bind ssDNA when it is covered with the SSB p5 (17,23). In addition, free 29 DNA polymerase, but not when bound to M13 ssDNA, can degrade a single-stranded oligonucleotide due to its 3Ј-5Ј exonucleolytic activity. M13 DNA that is complexed with an ssDNA-binding protein, therefore, is unable to trap the 29 DNA polymerase, which can be measured by the 3Ј-5Ј exonucleolytic activity. Thus, 29 DNA polymerase was added to either naked M13 ssDNA or the M13 ssDNA that was preincubated with increasing amounts of protein p16.7A. Then, 1 min after a 5Ј labeled oligonucleotide was added to the mixtures, samples were analyzed for degradation of the oligonucleotide. The assays were carried out in parallel using SSB p5. Fig. 5 shows, as expected, that the oligonucleotide was degraded by the 29 DNA polymerase when the M13 ssDNA trap was omitted (lane 2), but it was not degraded when the reaction mixtures contained naked M13 ssDNA (lane 3). In agreement with previously published results (17), preincubation of the M13 ssDNA with increasing amounts of SSB p5 resulted in increasing levels of degradation of the oligonucleotide (lanes 4 -7). Similar results were obtained when the M13 ssDNA had been preincubated with protein p16.7A (lanes 9 -12). The oligonucleotide was not degraded when it was only incubated with the highest concentration of SSB p5 or p16.7A (lanes 8 and 13, respectively). This excludes the possibility that the observed degradation of the oligonucleotide would be the consequence of a contaminant exonuclease in the purified protein preparations. These results, therefore, indicate that protein p16.7A binds to circular M13 ssDNA and that this prevents it from 29 DNA polymerase binding.

Protein p16.7A Has No Helix-destabilizing Activity and Has No Stimulatory Effect On in Vitro 29 TP-DNA Replication-
Possible helix destabilization activity of protein p16.7A was studied by its ability to displace a 5Ј-labeled 17-mer oligonucleotide hybridized to its complementary sequence in circular M13 ssDNA molecules. This substrate was incubated without or with increasing amounts of protein p16.7A, after which the samples were analyzed by polyacrylamide gel electrophoresis. The experiments were carried out in parallel using 29 SSB p5. As shown in Fig. 6, incubation in the absence of protein did not result in release of the labeled oligonucleotide, indicating that the hybrid substrate was stable throughout the experiment. In agreement with previously published results (17,24), the 29 SSB p5 was able to displace the labeled oligonucleotide from the M13 DNA. However, no displacement of the oligonucleotide was detected when the hybrid substrate was incubated with protein p16.7A up to a concentration of 112 M. To facilitate the oligonucleotide displacement, these assays were also per-formed using an oligonucleotide that contained, in addition to the 17 complementary nucleotides, either a non-complementary 5Ј-(5 thymidines) or a 3Ј-(5 adenines) tail. Contrary to the SSB p5, protein p16.7A was also unable to displace these oligonucleotides (results not shown). Because protein p16.7A is able to bind M13 ssDNA (see above), these results strongly suggest that protein p16.7A has no helix-destabilizing activity.
The 29-encoded TP and the DNA polymerase are the only two proteins required in a minimal in vitro 29 DNA replication system (29). Contrary to the 29 DNA amplification system, the minimal replication system requires high concentrations of template TP-DNA and is limited to one or two rounds of 29 DNA replication. To study a possible effect of protein FIG. 4. Protein p16.7A has higher affinity for ssDNA than the 29 SSB p5. Gel mobility shift assays were used to study the binding of purified protein p16.7A (A) and SSB p5 (B) to 5Ј end-labeled ssDNA molecules and to study binding of protein p16.7A to 5Ј end-labeled dsDNA molecules (C) corresponding to the 175-base pair right-end fragment of the 29 genome. The assays were performed as described under "Materials and Methods" in the presence of the indicated amounts of protein. The electrophoretic mobility of uncomplexed and DNA-protein complexes are indicated. D, Twenty g of purified p16.7A was subjected to a 30 -15% glycerol gradient and separated into 25 fractions. Part of each aliquot was analyzed by SDS-PAGE (upper part). The relative amount of p16.7A in each aliquot was determined and plotted in the graph shown in the lower part (squares). In addition, part of each aliquot was analyzed in gel retardation assay. The ability of each fraction to retard an ssDNA fragment is also presented graphically in the lower part of the figure (diamonds).
p16.7A in this system, in vitro 29 DNA replication assays were performed in the absence or presence of protein p16.7A using the SSB p5 as a control. The amount of incorporated [␣-32 P]dAMP was determined (Fig. 7A), after which the samples were subjected to alkaline-agarose gel electrophoresis to determine the size of the synthesized DNA (Fig. 7B). As reported previously (23,25), the SSB p5 stimulated the in vitro 29 DNA replication, especially at long incubation times. However, no stimulatory effect on replication was observed in the presence of protein p16.7A in this assay.
Relative Amounts of Protein p16.7 and SSB p5 Synthesized during the Infection Cycle-The intracellular amount of protein p16.7 accumulated in B. subtilis cells throughout the infection cycle has been determined before by quantitative immunoblotting (10). Although it has been described that the SSB p5 is one of the most abundantly expressed 29-encoded proteins in infected cells (25), a kinetic analysis of the accumulation of the SSB p5 during the infection cycle had not been performed. We were especially interested in the ratio between p16.7 and SSB p5 molecules per cell during the course of the infection cycle. Thus, samples of a B. subtilis culture were withdrawn at different times after infection and used for quantitative immunoblotting using polyclonal antibodies against protein p16.7 or against the 29 SSB p5. Fig. 8A shows the p16.7 and SSB p5 signals obtained in the Western blots. Both protein p16.7 and the SSB p5 were detected as soon as 9 min after infection. Overexposure of the blots shown in Fig. 8A and Western blots in which larger amounts of cell extracts were used allowed the detection of both proteins at 6 min after infection (results not shown). These latter Western blots were used to determine the amounts of the accumulated proteins at this early infection time (see below). In agreement with earlier observations (10), the p16.7 level increased only moderately during the middle stage of infection (10 -30 min) and remained almost constant at late infection times. On the contrary, a great increase in the amount of accumulated SSB p5 was observed during the middle stage of infection (10 -30 min). Densitometric analyses of the Western blot signals were used to calculate the number of protein p16.7 and SSB p5 molecules per infected cell at the various times analyzed (Fig. 8B). Whereas an infected cell contained a maximum of about 180,000 molecules of p16.7, the SSB p5 reached levels up to about 2 ϫ 10 6 . Fig. 8C shows that at 6 min after infection, the ratio between the SSB p5 and protein p16.7 is about 2, but this value increased rapidly during the next 10 min of infection to a value of about 12. DISCUSSION p16.7 Is a Membrane-localized, Dimeric, ssDNA-binding Protein-Here we demonstrated by functional and direct analysis that protein p16.7A has high ssDNA binding activity and that it binds ssDNA without an apparent sequence preference. Thus, protein p16.7A is an ssDNA-binding protein, and hence, the genome of 29 encodes for two ssDNA-binding proteins, protein p16.7 and SSB p5. Previously, we demonstrated that protein p16.7A forms dimers in solution, and in vivo crosslinking experiments suggested that native p16.7 also exists as a dimer in infected cells (10). In addition, we demonstrated that the native protein p16.7 is an integral membrane protein and FIG. 5. Protein p16.7A prevents binding of 29 DNA polymerase to circular M13 ssDNA. The assays were carried out using the 5Ј-labeled 17-mer oligonucleotide (Ϫ40 Universal). The indicated amounts of p16.7A (lanes 9 -13) or SSB p5 (lanes 4 -8) were incubated for 10 min with 400 ng of M13 ssDNA. Next, 29 DNA polymerase was added, and the mixtures were incubated for 75 s at 30°C, after which the reactions were stopped and processed as described under "Materials and Methods." M13 ssDNA and 29 DNA polymerase were omitted in lanes 8 and 13 (56 M SSB p5 and protein p16.7A, respectively) to study the possible presence of nucleases in the purified protein batches. The positions of the undegraded 17-mer oligonucleotide (17-mer) and its degradation products are indicated.
FIG. 6. Protein p16.7A has no helix destabilizing activity. Circular M13 ssDNA molecules to which the labeled 17mer oligonucleotide (Ϫ40 Universal) was hybridized was used as substrate. The mixtures were incubated for 60 min at 37°C with the indicated amount of SSB p5 or protein p16.7A, after which the samples were analyzed by SDS-PAGE. HD, heat-denatured control. The positions of the hybrid substrate (M13 ssDNA/17-mer) and the displaced 17-mer are indicated. that the N-terminal region of the protein, constituting a transmembrane-spanning domain, is responsible for its membrane localization (2,10). Together with the results obtained in this work, we conclude that the native protein p16.7 is a membranelocalized, dimeric, ssDNA-binding protein. To the best of our knowledge this is the first prokaryotic protein described with such characteristics.
29 Encodes for Two ssDNA-binding Proteins with Different Characteristics-In most systems SSBs stimulate DNA replication (for reviews, see Refs. 1, 30, and 31). This stimulation is the result of direct and/or indirect effects that SSBs have on DNA replication. The direct effect, reported for several SSBs, involves specific interactions between the SSB and its cognate DNA polymerase. The indirect effect can be the consequence of one or more of the following mechanisms. First, in various prokaryotic and eukaryotic replication systems in which the DNA polymerase lacks strand displacement activity, the SSB may help to load the helicase (for reviews, see Refs. 32 and 33). Second, upon binding, SSBs protect ssDNA from nuclease degradation. Third, it prevents non-productive binding of the DNA polymerase to ssDNA. Fourth, it prevents the formation of DNA secondary structures, e.g. loops, hairpins, or triple helices, which generally impede progression of the DNA polymerase. And finally, cooperative binding of the SSB to the ssDNA strand reduces the energy demand required for unwinding the dsDNA at the replication fork. Thus, the helix destabilization activity inherited by many SSBs helps to increase the DNA elongation rate.
Convincing evidence has been provided that the 29 gene 5, present in the early expressed operon located in the left part of the 29 genome (Fig. 1A), encodes an SSB (for reviews, see Refs. 3, 6, 34, and 35). The 29 SSB p5 stimulates in vitro 29 DNA replication, especially at long incubation times (23,25). Because no evidence has been obtained that the 29 SSB has direct interactions with the 29 DNA polymerase, the stimulation is most likely due to indirect effects. The following features of 29 SSB have been described that, at least in part, explain this stimulatory effect. First, binding of the 29 SSB protects ssDNA against nuclease degradation (25,26). Second, it prevents non-productive binding of the 29 DNA polymerase to ssDNA (17,23). Third, it prevents and removes the formation of DNA secondary structures (23,24). Fourth, it assists in unwinding the dsDNA at the replication fork due to its helix destabilization activity (Refs. 17 and 24, this work, and Fig. 6). And finally, it is required in in vitro 29 DNA amplification assays when low amounts of 29 template TP-DNA are used (19).
Here we demonstrated that protein p16.7A has high ssDNA binding activity. Contrary to SSB p5, however, protein p16.7A has no helix destabilization activity, and its presence in the minimal in vitro 29 DNA replication assay did not stimulate DNA replication. These results indicate that, although both proteins bind ssDNA, they have non-overlapping functions. This view is further supported by the following results. First, the SSB p5 is essential for in vivo 29 DNA replication (36), indicating that protein p16.7 cannot replace all of the functions of the SSB p5. Second, although both proteins are expressed early after infection, much lower amounts of protein p16.7 are synthesized than the SSB p5 during the middle and late stages of phage infection. The rapid amplification of the 29 genome during the middle stage of the infection cycle generates large amounts of ssDNA. The huge increase of SSB p5 synthesis during this infection stage is in agreement with the view that   FIG. 7. Protein p16.7A has no stimulatory effect on in vitro 29 DNA replication. Replication reactions using TP-DNA as template were carried out as described under "Materials and Methods" in the absence (squares) or the presence of either 6 g of protein p16.7A (circles) or SSB p5 (diamonds). After incubation for the indicated time at 30°C, the samples were first used to measure the amount of incorporated [␣-32 P]dATP, from which the relative replication activity was calculated (A). Next, the samples were run in an alkaline-agarose gel to determine the lengths of the synthesized DNAs (B). The position of unit length 29 DNA is indicated.
the SSB p5, as most other SSBs, is required in stoichiometric amounts.
Model of the Function of Protein p16.7 in in Vivo 29 DNA Replication-As a consequence of its protein-primed mechanism of replication and the absence of lagging strand synthesis, the 29 DNA replication intermediates contain very long stretches (up to Ͼ10,000 nucleotides in length) of displaced ssDNA (see Fig. 1B). Binding of protein p16.7A to these stretches of ssDNA was demonstrated directly by electron microscopy analysis. Also, the observation that protein p16.7A prevented the accumulation of short DNA products in the in vitro 29 DNA amplification assays lacking SSB p5 strongly indicates that it binds to these stretches of ssDNA. Importantly, these latter results also show that the binding of p16.7A to the stretches of ssDNA hardly interfered with DNA replication in the dynamic 29 DNA amplification system. Rather, under these conditions, protein p16.7A had a positive effect on the efficiency of full-length 29 DNA synthesis. Based on these results, the characteristics of p16.7A described above, and taking into account that native protein p16.7 is membrane-localized (10), we propose, as schematically presented in Fig. 9, that the principal role of p16.7 in vivo involves the attachment of replicating 29 DNA molecules to the membrane of infected cells. Although not shown in the model, it is possible that both protein p16.7 and SSB p5 will be bound simultaneously to the ssDNA portions of the 29 DNA replication intermediates under natural conditions. The percentage of the 29 ssDNA portions that is bound by either protein probably varies during the infection cycle because of the different expression patterns of these proteins. Nevertheless, based on the relative ssDNA binding activities, it is most likely that p16.7 binds more rapidly to a displaced ssDNA strand of a 29 replication intermediate than the SSB p5, especially at early infection times when the amount of p16.7 in the infected cell, compared with the SSB p5, is relatively high. Interestingly, in vivo 29 DNA replication is affected, especially at early infection times in the absence of protein p16.7 (10), indicating that it has an important role during early infection times. Attachment of 29 DNA replication intermediates to the membrane by protein p16.7 will result in compartmentalization of the phage DNA within the infected cell. Besides a possible role of p16.7 in the organization of in vivo 29 DNA replication, compartmentalization of replicating 29 DNA probably stimulates in vivo 29 DNA replication directly.
The proposed role of p16.7 in in vivo 29 DNA replication may explain several observations made by Ivarie and Pène (8), who demonstrated for the first time that 29 DNA replication occurs at the membrane of infected cells. These authors showed (i) that it took at least 10 min before the parental infected 29 DNA molecules became attached to the membrane, (ii) that Cell extracts, prepared by sonication of samples withdrawn at the indicated times, were separated by SDS-PAGE (10 -20% linear gradient) and subsequently analyzed by Western blotting as described under "Materials and Methods." The amount of cell extracts loaded to detect the SSB p5 was 4-fold lower than those used for the detection of protein p16.7. Films developed after various exposure times, such that the signals were in the linear range, were used for quantitative analysis of the accumulated proteins. B, graphic representation of the absolute amounts of p16.7 and SSB p5 molecules per cell during the infection cycle. C, graphic representation of the ratio of SSB p5 and protein p16.7 during the infection cycle.
early expressed viral protein(s) was required for membrane association, (iii) that replicating 29 DNA molecules were attached to the membrane and that the fully replicated doublestranded 29 DNA molecules were first released from the membrane before they were packaged into the phage particles, and (iv) that phage DNA of a 29 mutant containing a temperature-sensitive mutation in gene 2 (later shown to encode the DNA polymerase) did not become associated to the membrane when infected at the non-permissive temperature. Replication of 29 DNA does not start until about 10 min after infection (2,8), implying that until this time no ssDNA-containing replication intermediates are present in the infected cell. This explains why the infecting 29 DNA molecules, present in its double-stranded form, are not associated to the membrane at the very early infection times even though p16.7 is already detected as soon as 6 min after infection (this work and Ref. 10). Also the observation that the replicating DNA molecules, but not the fully replicated phage DNA molecules, are attached to the membrane is in agreement with the proposed role of protein p16.7. By immunofluorescence microscopy we have indeed confirmed that the majority of the replicated doublestranded 29 DNA molecules present at late infection times were no longer localized at the membrane. Rather, these molecules were found within the bulk of the host nucleoid (2). Finally, although the possibility, proposed by Ivarie and Pène (8), that the 29 gene 2 product, the DNA polymerase, could be directly involved in attachment of replicating 29 DNA to the membrane still holds, another perhaps more logic explanation can be considered. Replication will not occur in the absence of a functional DNA polymerase. As a consequence, the infecting DNA, despite the synthesis of protein p16.7 and other viral proteins, will remain double-stranded, which explains why it did not become attached to the membrane. By immunofluorescence microscopy it was shown that the first rounds of 29 DNA replication localized nearly always to the distal end of the bacterial nucleoid both in the presence or absence of protein p16.7. However, whereas a few minutes later, phage DNA replication was found to occur at multiple sites at the membrane in the wild-type situation, phage DNA replication remained restricted to the initial replication site for a prolonged time in the absence of p16.7. These results indicated that protein p16.7 is involved in the efficient distribution of replicating phage DNA from its initial to additional replication sites in the infected cell. During the middle stage of infection protein p16.7 localizes throughout the membrane, and this pattern is consistent with the localization of the viral DNA at these times (2). This may suggest that the in vivo distribution of phage DNA is a passive process due to the dispersed p16.7 localization. However, active distribution of phage DNA within the infected cell cannot be ruled out.
Gene 16.7 Is Conserved in 29-related Phages-The generation of replication intermediates containing large stretches of displaced ssDNA is a characteristic feature of the proteinprimed mechanism of DNA replication. Thus, it is conceivable that gene 16.7 is conserved in the genome of other phages that replicate by the protein-primed mechanism. Phage 29 is the paradigm of a family of Bacillus phages that all use the proteinprimed mechanism of DNA replication (for review, see Ref. 6). These 29-like phages can be divided into three groups. The first group includes, in addition to 29, phages PZA, 15, and BS32. The second group comprises B103, Nf, and M2Y (M2), and the third, most distantly related group, contains GA-1 as its sole member. Except for phages Nf and M2Y, the DNA sequences of the right part of these phage genomes are available. Analysis of these DNA sequences revealed that they all contain an open reading frame whose deduced protein sequence (ranging from 130 to 133 amino acids) shares significant similarity to that of the 29 protein p16.7. In addition, Western blot analysis using polyclonal antibodies against p16.7 of 29 provided evidence that phage Nf (group II) encodes a p16.7 homologue (10). Together, these data show that gene 16.7 is conserved in most and probably all 29-related phages, sup- porting the view that the proposed mechanism of p16.7-mediated attachment of replicating phage DNA to the membrane of the infected cell is conserved in all these phages.
In summary, we found that the integral membrane protein p16.7 of 29 is an ssDNA-binding protein. Based on the results reported here and those reported previously (2,10), it is most probable that the principal role of p16.7 involves the attachment of replicating 29 DNA to the membrane of the infected cell, resulting in the efficient distribution of replicating 29 DNA from its initial to additional replication sites. Probably, this mechanism is conserved in all members of the 29 family of phages. As such, these results are an important contribution to the better understanding of protein-primed 29 DNA replication.