Functional linkages between replication proteins of genes 1 , 3 and 5 of Bacillus subtilis phage φ 29

Gene 1 product (gp1) of Bacillus subtilis phage φ29 has been shown to be involved in viral DNA replication in vivo, but the essential role is still unknown. As part of an ongoing effort to understand the role of gp1 in viral DNA replication, we investigated genetic interaction between gene 1 and other viral genes. Because φ29 mutants which do not produce functional gp1 show temperaturesensitive growth, we isolated temperature-resistant phages from the φ29 gene 1 mutants, and eventually, obtained nine extragenic suppressors. These suppressor mutations were located in two essential genes for φ29 DNA replication in vivo: gene 3 encoding terminal/primer protein (gp3) or gene 5 encoding viral singlestranded DNA binding protein (gp5). Most of these mutations resulted in single amino acid substitutions in the products. By trans-complementation assay, we confirmed that the absence of gp1 at non-permissive temperature can be compensated by the suppressors which have the single amino acid substitution in either gp5 or gp3. These results indicate that gp1 has functional relationship to gp5 and gp3. From the positions of amino acid substitutions in gp3, we propose its new regulatory subdomain at which other molecules including gp1 would interact with and regulate functions of gp3.


INTRODUCTION
Bacillus subtilis phage ϕ29 has a linear doublestranded DNA genome with terminal proteins covalently linked at each 5'-end (Hirokawa, 1972) (Fig. 1).Such protein-linked DNAs have also been found in other viruses (e.g., PRD1 and adenovirus), plasmids (e.g., pSKL1) and bacteria (e.g., Streptomyces).In the case of these viruses, DNA replication initiates via a proteinpriming mechanism, and ϕ29 has been studied as the most simple and typical case of the protein-priming replication (Meijer et al., 2001).
The genes of ϕ29 DNA replication proteins including DNA polymerase, gp3 (acts as terminal protein and primer protein), gp5 {single-stranded DNA binding (SSB) protein} and DBP (double-stranded DNA binding protein) are clustered in the left end early region of the genome (Fig. 1), and their roles have been investigated by in vitro experiments (Meijer et al., 2001).Among these proteins, DNA polymerase and gp3 encoded by viral gene 2 and gene 3, respectively, were found to be minimal essential components for in vitro duplication of ϕ29 genome.At the initiation of protein-primed DNA replication, a heterodimer composed of DNA polymerase and gp3 binds to the two replication origins located at both ends of ϕ29 linear genome, and the DNA polymerase covalently combines the first nucleotide (dAMP) with the hydroxyl group of Ser 232 provided by the gp3.This initiation step is greatly facilitated by DBP that is believed to help opening the termini of genomic DNA by forming a multimeric nucleoprotein complex at the replication origins.Subsequently, the same DNA polymerase catalyzes highly processive polymerization by strand displacement mechanism.Gp5 was reported to bind to the displaced single-stranded DNA (ssDNA) in vitro for protection from degradation by nucleases and for prevention from nonproductive binding of DNA polymerase to the ssDNA (Gutiérrez et al., 1991;Gascón et al., 2000).These roles of gp5 probably explain the stimulation of dNTP incorporation by ϕ29 DNA polymerase in vitro.
According to the genetic studies using conditional lethal mutants, products of ϕ29 gene 2, gene 3, gene 5 and gene 6 are essential for in vivo ϕ29 DNA replication (Talavera et al., 1972;Carrascosa et al., 1976;Mellado et al., 1980).In addition to these proteins, viral gene 1 product (gp1) is required for efficient ϕ29 DNA replication in vivo.ϕ29 sus1(629), an ochre mutant in gene 1, has a replacement of a CCA codon (Gln at position 6) by a nonsense TAA codon (Prieto et al., 1989).ϕ29 which bears this sus1 mutation does grow at 30°C but fails to replicate its genome at 42°C, indicating that gp1 is essential at 42°C but dispensable at 30°C for the growth (Carrascosa et al., 1976;Bravo and Salas, 1998).Gp1 has been reported to exhibit multiple characteristics including selfassociation (Bravo and Salas, 1998;Bravo et al., 2001;Serrano-Heras et al., 2003;Hashiyama et al., 2005), cellular membrane localization (Bravo and Salas, 1997;Serrano-Heras et al., 2003), RNA binding (Takeuchi et al., 1998) and interaction with gp3 (Bravo et al., 2000).Despite the extensive studies, the role of gp1 in DNA replication is still unknown.Suspecting from the various characteristics, gp1 may play the roles in partnership with other viral replication proteins.To clarify the roles of gp1 in ϕ29 DNA replication, we attempt to find relationships between gene 1 and other genes by suppressor analysis.We focused on the temperature-sensitive phenotype of ϕ29 gene 1 mutants and started to isolate temperature-resistant (Tr) revertants from it.Up to now, nine Tr mutants bearing the original gene 1 mutations were identified.These Tr mutants have second-site mutations in gene 3 or gene 5 which involved in ϕ29 DNA replication.The functional relationship between gp1, gp5 and gp3 will be discussed.

MATERIALS AND METHODS
Bacterial strains, plasmid, bacteriophages Bacterial and viral strains used in this study are listed in Table 1 including a series of newly constructed ϕ29 nonsense mutants (ns1, ns3, ns5 and their double mutants: "ns" stands for nonsense mutant and number shows the genes that have the nonsense mutation).Integration vector pSG1729 (Lewis and Marston, 1999) was obtained from Bacillus Genetic Stock Center.Escherichia coli DH5α was used for cloning.

Construction of B. subtilis strains expressing ϕ29
genes The xylose-inducible expression vector, pSG1729, allows the integration of a gene at amyE locus of B. subtilis genome.At first, ϕ29 gene 3 or gene 5 were PCR amplified from phage genomes using primers carrying AvrII and BamHI restriction sites (Table 2).After digestion with AvrII and BamHI, these PCR fragments and vector pSG1729 were ligated and cloned in E. coli DH5α to give pSG1729-gene 3 and pSG1729-gene 5.Although vector pSG1729 was developed to express N-terminal GFP fusion protein (Lewis and Marston, 1999), the gfpmut1 gene was unnecessary in this study.Thus, the gfpmut1 gene was excised and replaced by each of ϕ29 genes.To construct B. subtilis strains expressing both genes 5 and 1 or genes 3 and 1, gene 1 was PCR amplified from phage genome using primers carrying BamHI and ClaI restriction sites (Table 2).The gene 1 fragment was ligated with pSG1729-gene 5 or pSG1729-gene 3 previously digested with BamHI and ClaI.The ligation mixtures were then subjected to PCR using the Pori and Pbla primers (Table 2) to give pSG1729-gene 3/gene 1 and pSG1729-gene 5/gene 1 fragments.Using these vectors (pSG1729-gene 3 and pSG1729-gene 5) and PCR fragments (pSG1729-gene 3/gene 1 and pSG1729-gene 5/gene 1 fragments), B. subtilis 168 was transformed as described (Young and Spizizen, 1961), and transformants were selected on LB plates containing spectinomycin (50 μg/mL).To confirm the insertion sites, amyE region was PCR amplified using amyE_5 and amyE_3 primers (Table 2) and sequenced.Since Spo0A protein encoded by host genome inhibits ϕ29 DNA replication (Castilla-Llorente et al., 2006), the spo0A gene was disrupted by inserting neomycin resistant gene.

Constructions of ϕ29 mutants
We introduced mutations into ϕ29 genome by overlap extension PCR method using mutagenic oligonucleotides listed in Table 2.For example, ϕ29 ns5 was constructed as follows.A set of primers 29end and mu-R (ns5) in PCR 1, and 5895R and mu-L (ns5) in PCR 2 were used with ϕ29 genome DNA as template.The products of PCR 1 and PCR 2 were mixed and used as templates in PCR 3 which was primed by two primers 29end and 5895R.The products of PCR 3 bearing the nonsense mutation in gene 5 contain two FbaI sites at near the both ends.After FbaI digestion, the DNA fragments were purified from agarose gel.ϕ29 genomic DNA containing terminal proteins (TP-DNA) was prepared by incubating ϕ29 particles in the presence of 3 mM EDTA and 0.5% SDS for 30 minutes at 25°C.(Meijer et al., 2000)  tured at 37°C for 2.5 hours in LB medium containing 1% glucose were concentrated 5-fold in PPB buffer (0.5 M sucrose, 20 mM MgCl 2 and 20 mM Maleic acid pH 6.5) containing 0.1 mg/mL lysozyme.After incubation at 40°C for 15 minutes, the protoplast cells were collected in 1 mL of PPB buffer, and mixed with TP-DNA in the presence of 40% polyethylene glycol 4000 for 2 minutes at room temperature.The cells were resuspended in RTSB medium (3% Trypticase soy broth, 0.1 mg/mL L-glutamine, 0.3 M sodium succinate, 20 mM MgCl 2 and 1 mg/mL BSA), and incubated for 2 hours at 25°C to develop phage particles in the cells.The RTSB cultures were plated with indicator strain TT299 on LB plate containing 0.5% xylose and incubated at 30°C to obtain plaques of ϕ29 mutants.After single plaque isolation, left end early regions were PCR amplified using primers 29end and 5895R (Table 2) and sequenced.Other ϕ29 mutants used in this study were constructed by the same method using specific primers (Table 2).For the construction of ϕ29 ns1/ns5 double mutant, ns5 phage was used as template for PCR 1 and 2. For the construction of ϕ29 ns1/ ns3 double mutant, ns1 phage was used as template DNA for PCR 1 and 2, respectively.B. subtilis strain TT486 expressing ϕ29 gene 3 was used as host for the construction of ϕ29 ns3.B. subtilis strain TT541 expressing ϕ29 gene 3 and gene 1 was used as host for the construction of ns1/ns3 mutant.B. subtilis strain TT535 expressing ϕ29 gene 5 and gene 1 was used as host for the construction of ϕ29 ns1/ns5 mutant.

RESULTS
ϕ29 gene 1 mutants as parental strain for suppressor hunting We used three ϕ29 gene 1 mutants as parental strain for suppressor hunting: sus1(629), ns1 and mis1(L57S) (Table 1).ϕ29 sus1(629) and ns1 have the same nonsense mutation at the sixth codon in gene 1 and do not produce any detectable gene 1 product (gp1) in the infected cells as examined by western blot analysis using anti-gp1 serum (data not shown).Although ϕ29 sus1(629) strain has been used in several studies on gp1, we found that our stock had numerous additional mutations in the left end early region where most of the ϕ29 DNA replication genes locate (Fig. 1).Therefore, to simplify the analysis, we selected a clone, ns1 which does not have such additional mutations at least in the left end region.In contrast, ϕ29 mis1(L57S) have the missense mutation in gene 1 which results in an alteration of Leu 57 to Ser.The missense mutation was assumed to affect gp1's function since it substitutes one of the key amino acid in the putative coiled-coil region which is important for the self-association property of gp1 in vitro (Hashiyama et al., 2005).Western blot analysis showed that the amount of gp1-L57S was gradually decreased after infection while wild type gp1 accumulated with time, suggesting that stability of the mutant protein is reduced (data not shown).
Table 3 shows plating efficiency of these ϕ29 gene 1 mutants at 30°C and 42°C.Although all three strains could grow in our standard host strain SR22 at 30°C, the growth was greatly inhibited at 42°C.This growth inhibition at 42°C was almost completely restored when gp1producing B. subtilis TT275 were infected with these phages.These results indicated that these ϕ29 gene 1 mutants showed the temperature sensitive phenotype due to the absence of functional gp1.

Isolation of suppressors of gene 1 mutations
To isolate suppressors from these gene 1 mutants, approximately 10 5 -10 6 ϕ29 phages were plated on B. subtilis SR22 at 42°C in order to select for temperature-resistant (Tr) revertants.After single plaque isolation of the appeared plaques, the left end early region of ϕ29 genome (Fig. 1) was PCR amplified using a set of primers 29end and 5895R (Table 2), and their DNA sequences were determined.
Tr revertants were appeared at a frequency shown in Table 3.These revertants would have second-site, compensatory mutations (suppressor) or reverse mutations to wild type gene 1.To avoid picking up the revertants to wild type gene 1, we selected for slightly small plaques among the Tr revertants.Such precaution was taken on the assumption that suppressors might have intermediate phenotypes between the wild type and the parental phage strains.In the screen using sus1(629) and ns1, one or two Tr phages forming smaller plaque were isolated, respectively.In the screen using mis1(L57S), 13 Tr phages were isolated, and four were shown to be revertants to wild type gene 1.
As shown in Table 4, nine kinds of Tr phages retaining the original gene 1 mutations were obtained, and extragenic-and missense mutations were found in the three genes reported to be essential for ϕ29 DNA replication: gene 2, gene 3 and gene 5.Most of the second site mutations were isolated only once (Table 4), suggesting that suppressor hunting was not saturating and further suppressors will be obtained.
Absence of gp1 can be compensated by the suppressors which have amino acid substitutions in gp5 (viral SSB protein) Four Tr phages have secondsite mutations in gene 5 (Table 4).To verify whether these gene 5 alleles are suppressors of the gene 1 mutations, we conducted trans-complementation assay as follows.First, each of the gene 5 alleles was cloned into amyE locus of B. subtilis chromosome and expressed from the xylose inducible promoter (Table 1).Western blot analysis revealed that the levels of the synthesized gp5 were comparable with that between variants and wild type (data not shown).Then we examined whether the gp5 variants can support growth of the parental ϕ29 strain ns1 at 42°C.To avoid production of wild type gp5 from infected ns1 phage, we introduced nonsense mutation into the gene 5 in ns1 phage genome, resulting in ϕ29 ns1/ns5 double mutant (Table 1).Thus in this assay, ns1/ns5 phage cannot produce both gp1 and gp5, but gp5 is supplied in trans from the host genome infected.As shown in Fig. 2, wild type gp5 failed to support the growth of the ϕ29 ns1/ns5 at 42°C.In striking contrast, four gp5 variants were able to support the growth of the ϕ29 ns1/ns5.These results confirm that the identified gene 5 alleles are suppressors of nonsense mutation in gene 1, and indicate that the absence of gp1 at non permissive temperature of 42°C can be compensated by the amino acid substitutions in gp5.Although the gene 5-S77Y mutation was isolated from mis1(L57S) mutant (Table 4), it suppressed the nonsense mutation in gene 1 (Fig. 2).Therefore, the suppression by gene 5-S77Y mutation does not require gp1-L57S protein.The gene 2-D516H mutation was found in ns1-Tr1 in addition to gene 5-L16I mutation (Table 4).Although the effect of the gene 2-D516H mutation was not examined in this study, gp5-L16I alone was sufficient to suppress the nonsense mutation in gene 1 (Fig. 2).
Absence of gp1 can be compensated by the suppressors which have amino acid substitutions in gp3 (Terminal/Primer protein) Three Tr phages isolated from ϕ29 mis1(L57S) mutant have second-site mutations in gene 3 (Table 4).By similar transcomplementation assay as described above, we verified whether the identified gene 3 alleles were also suppressors of the gene 1 mutation.When each of the gene 3 alleles was cloned into B. subtilis chromosome under xylose promoter (Table 1), the induced levels of the gp3 in the cells were comparable with that between variants and  629) strain: gp5-A71T is one of such mutations which derived from parental sus1(629) strain.c Representative strains were shown.From ϕ29 mis1(L57S) mutant, we have recently obtained other three Tr strains in which second-site mutations were found in gene 5: two Tr strains have the mutation leading to gp5-S32T amino acid change, the other have the one leading to gp5-S74L amino acid change.At present, these two gene 5 alleles were not analyzed further.Fig. 2. The growth defects of ϕ29 phage in the absence of gp1 can be compensated by the amino acid substitutions in gp5.B. subtilis strains TT299 producing wild type gp5 (open circle), TT303 producing gp5-A71T/A86V (filled triangle), TT305 producing gp5-S77Y (filled circle), TT307 producing gp5-L16I (open triangle) and TT309 producing gp5-T82N (open square) were grown in LB medium containing 5 mM MgSO 4 at 42°C.Before 60 minutes of infection, xylose was added at a final concentration of 1% to express gene 5 alleles, and these cells were infected with ϕ29 ns1/ns5 mutant at a multiplicity of infection of 0.1.After 7 minutes of incubation with gentle shaking, the culture was diluted with the pre-warmed same medium to 10 4 -fold and incubated further at the same temperature.At indicated times plaque forming unit was measured by plating on indicator strain TT535.0 2 0 4 0 6 0 8 0 1 0 0 1 2 0 Time after infection (min.) Relative p.f.u.
1 10 100 wild type (data not shown).The ability of these gp3 variants to support growth of the ϕ29 gene 1 mutants at 42°C was examined with another ϕ29 strain, ns1/ns3 (Table 1).This strain was constructed from ns1 by inactivation of gene 3 with nonsense mutation so that the wild type gp3 would not be produced from the phage genome.Thus in this assay, the phage does not produce both gp1 and gp3, but gp3 was supplied in trans from the host genome infected.As shown in Fig. 3A, gp3-N69K and wild type gp3 were not able to support growth of ϕ29 ns1/ns3, while two gp3 variants, gp3-E78K and gp3-E100K, were able to support the growth of the mutant at 42°C.Although the two gene 3 alleles were isolated from ϕ29 mis1-L57S mutant (Table 4), gp3-E78K and gp3-E100K supported growth of the ϕ29 ns1/ns3.These results indicate that these two alleles of gene 3, gene 3-E78K and gene 3-E100K, can suppress the nonsense mutation in gene 1, and that the absence of gp1 can be compensated by the amino acid substitutions in gp3.Only gp3-N69K was not able to support the growth of ϕ29 ns1/ns3 (Fig. 3A).We confirmed that ϕ29 mis1(L57S)-Tr2, a Tr strain bearing the gene 3-N69K mutation (Table 4), formed smaller plaques at 42°C compared with other Tr strains isolated here (data not shown).Therefore, we suppose that the suppression by gp3-N69K is less effective than other suppressors.
As shown in Fig. 3A, wild type gp3 does not support the growth of ϕ29 ns1/ns3 mutant at 42°C.However, when pre-induced before infection, wild type gp3 did support the growth of the ns1/ns3 mutant (Fig. 3B).

The gp5 and gp3 variants retain their original activities
As described above, amino acid substitutions in gp5 or gp3 compensated the absence of gp1 at 42°C.To examine whether these mutant proteins retain their original functions, we tested if the wild type gp5 or gp3 could be replaced by their variants.For this purpose, we constructed gp5 or gp3-deficient ϕ29 mutants, ns5 and ns3, which has a nonsense mutation in gene 5 and gene 3, respectively (Table 1).These phages could not direct synthesis of gp5 or gp3 as determined by western blot analysis using anti serum against gp5 or gp3 (data not shown).
First, each of gp3-producing cells described above was infected with ϕ29 ns3, and growth of the phage was monitored.As shown in Fig. 4A, all of gp3 variants supported growth of the ϕ29 ns3 as well as the wild type gp3 did, indicating that gp3 variants, gp3-N69K, gp3-E78K and gp3-E100K, retained their original activities.
Upon conducting the same experiment for gp5 variants, we unexpectedly found that ϕ29 ns5 did not grow at 45°C but grow at 42°C and lower temperatures, showing temperature-sensitive growth (Tone et al., 2012).Therefore, each of gene 5-producing cells was infected with ϕ29 ns5 at 45°C (Fig. 4B).Three gp5 variants, gp5-L16I, gp5-S77Y and gp5-A71T/A86V, supported the growth of ns5 mutant almost as well as the wild type gp5, whereas gp5-T82N did not support the growth.These results indicate that gp5's activity is retained in gp5-L16I, gp5-S77Y and gp5-A71T/A86V, but affected in gp5-T82N at least under this condition.

DISCUSSION
Gene 1 product (gp1) is required for ϕ29 DNA replication at 42°C but dispensable at 30°C.Although several reports about gp1 have been published, its role in DNA replication is unclear.In this study, we identified gene 3 and gene 5 alleles as suppressor of gene 1 mutations, and showed that the absence of functional gp1 at 42°C was compensated by the amino acid substitutions in products of gene 5 (gp5) or gene 3 (gp3).These results demonstrate that gp1 has functional relationship with replication proteins gp5 and gp3.Because all gp5 or gp3 variants retained their original activities (Fig. 4), these suppressors are thought to be gain-of-function mutations.Only gp5-T82N did not support growth of gp5-deficient phage ns5 (Fig. 4B).This assay was performed at 45°C due to the phenotype of the ns5 mutant phage that can grow at 42°C in the absence of functional gp5 (Tone et al., 2012).We assumed that the gp5-T82N retained its original activity at temperature of 42°C where the suppression occurs.
Second-site mutations which suppress gene 1 defect were also found in other genes.Recently, we have isolated Tr phages from another ϕ29 gene 1 mutant, mis1(M53T).This mutant strain has missense mutation in gene 1 which results in an alteration of Met 53 to Thr and also showed temperature-sensitive growth: plating efficiency at 42°C was 51-fold lower than that at 30°C.Three Tr strains retaining the original gene 1-M53T mutation were obtained.DNA sequencing revealed that two strains had second-site and missense mutation in gene 6, resulting in an alteration of Thr 10 to Ser (T10S) or Glu 64 to Gly (E64G) in the product.The gene 6 product (double-stranded DNA binding protein; DBP) is also reported to be essential for ϕ29 DNA replication in vivo (Carrascosa et al., 1976), and initiation of ϕ29 DNA replication in vitro is greatly stimulated by DBP (Blanco et al., 1986).The other Tr strain has missense mutation in gene 1 leading to Ala 84 to Glu (A84E) substitution in addition to the original M53T mutation.It would be required to confirm that these gene 6 and gene 1 mutations are intergenic and intragenic suppressors of the gene 1 mutation, respectively.To perform the similar trans-complementation assay described in this study, ϕ29 double mutant which has mutations in gene 1 and in gene 6 is required.We have tried to construct the double mutant, but the trials are unsuccessful so far.
ϕ29 gp3 is a key protein for ϕ29 DNA replication: it acts as protein primer for initiation of DNA replication and is also the component of DNA replication origin as the "terminal protein" covalently linked to ϕ29 genome (Meijer et al., 2001).The crystal structure of gp3/ϕ29 DNA polymerase heterodimer complex was determined (Kamtekar et al., 2006).As shown in Fig. 5A, gp3 is composed of three domains; the N-terminal domain (1-73 amino acid) that is important for the DNA binding (Zaballos and Salas, 1989;Muñoz-Espín et al., 2010), the intermediate domain (74-172 amino acid) that extensively interacts

B
with TPR-1 subdomain of the polymerase, and the priming domain (173-266 amino acid) which protrudes into the DNA template tunnel of the polymerase.In this study, we found that absence of gp1 at 42°C was compensated by the gp3 variants with single amino acid substitutions (N69K, E78K, and E100K).The positions of the amino acid substitutions in gp3 are also shown in Fig. 5A.All the substitutions are located in or near the intermediate domain that is reported to confer specificity to the interaction with the DNA polymerase (Pérez-Arnaiz et al., 2007).However, these substitutions are located in the opposite side of the interacting surface with the polymerase (Fig. 5A).In fact, biochemical analyses using purified gp3 indicated that gp3-DNA polymerase complex formation and DNA binding ability were not affected by these amino acid substitutions (data not shown).Here, we propose that the intermediate domain of gp3 would be divided into two subdomains, the DNA polymerase interaction subdomain and a suggested regulatory subdomain including amino acid residues N69, E78, and E100 where external molecules including gp1 would be accessible (Fig. 5B).
Gp3 is known to interact with versatile molecules other than the DNA polymerase or DNA: in addition to gp1, gp16.7 (viral gene 16.7 product), cellular membrane, and gp3 itself were reported to interact with gp3 (Bravo and Salas, 1997;Bravo et al., 2000;Serna-Rico et al., 2000, 2003).In the gp3 variants, interactions of gp3 with these components could be affected.Considering the functions of these components, three possibilities of the suppression mechanism could be argued.First, interaction between the protein primer (gp3 associated with the DNA polymerase) and the terminal protein (gp3 covalently bound to the ϕ29 DNA) could be affected.This interaction is considered to be important for the initiation of ϕ29 DNA replication (Serna-Rico et al., 2000).Alternatively, subcellular localization of gp3 variants could be affected.In vivo localization of gp3 was observed near the bacterial membrane or the nucleoid (Bravo and Salas, 1997;Muñoz-Espín et al., 2010), and the integral membrane protein gp16.7 and gp1 were proposed to help gp3 localize near the membrane (Bravo and Salas, 1997;Bravo et al., 2000;Meijer et al., 2000;Serna-Rico et al., 2003;Muñoz-Espín et al., 2009).Finally, although interaction with gp3 is unknown, DBP (gene 6 product) which has been reported to promote the initiation of viral DNA replication could be involved in the suppression mechanism.This possibility could be supported by the fact that gene 6 mutations were identified as candidates for suppressor of a gene 1 mutation (M53T), suggesting a functional linkage among gp1, gp3 and DBP.To clarify the suppression mechanism by the gp3 variants in the absence of gp1, above mentioned protein-protein interactions and its influences on viral DNA replication or  (Bravo and Salas, 1997;Bravo et al., 2000;Serna-Rico et al., 2000, 2003).The intermediate domain of gp3 was divided into two subdomains, a DNA polymerase interaction subdomain and a proposed regulatory subdomain.In this model, gp1 regulates gp3's interactions by acting as a bridging molecule or as a competitive inhibitor, or by inducing conformational change of gp3.

B
A Genetic interactions among ϕ29 replication genes subcellular localization should be investigated.Also to fully understand the functional relationship between gp1 and gp3, involvement of gp1 in these interactions should be examined.Because physical interaction between gp1 and gp3 was reported (Bravo et al., 2000), we surmise that binding of gp1 influences gp3's interactions with other molecules by acting as a bridging molecule or as a competitive inhibitor, or by inducing conformational change of gp3 (Fig. 5B).
ϕ29 gp5 is a viral SSB protein which has been shown to be essential for its DNA replication in vivo (Talavera et al., 1972;Carrascosa et al., 1976;Mellado et al., 1980) and greatly stimulated ϕ29 DNA synthesis in vitro (Gutiérrez et al., 1991).In contrast to gp3 and DBP, gp5 has been reported to be involved in the elongation, but not in the initiation of viral DNA replication (Martín et al., 1989).Thus, the suppressor mechanism by gp5 variants might be different from that of the suppressors in gp3 and DBP.We found that ϕ29 ns1/ns5 double mutant constructed in this study cannot grow at 30°C, although each single mutant can grow at this temperature, indicating that mutations in gene 5 and gene 1 cause synthetic lethal phenotype (to be published elsewhere).This suggests that gp1 and gp5 have redundant or mutually compensating functions.Considering the affinity of gp1 to RNA (Takeuchi et al., 1998), it would be interesting to test its ability to bind DNA as well, and investigate its role in stabilizing single-stranded DNA during the elongation of viral DNA replication.Furthermore, in recent years, SSB proteins have emerged as key components in recruiting and organizing the replication machinery via multiple protein-protein interactions (Lohman and Ferrari, 1994;Shereda et al., 2008).Such interaction partner has not been identified for gp5.Given the genetic relationship presented here, it would also be interesting to assess physical interaction between gp1 and gp5, and synergistic action of the interaction.
By isolating suppressors of gene 1 mutations, the presence of functional relationship between gp1 and gp3, gp5, or DBP is uncovered.Because gp1 is required for viral DNA replication at 42°C and higher temperatures but dispensable at 30°C, we assume that gp1 supports the function of other replication proteins, such as gp3, gp5 and DBP at 42°C.Therefore, it is important to investigate the physical interaction between these proteins, and the functions of the variants identified here.We are now examining biochemical properties of these gp3 and gp5 variants.Through investigation of the suppressor mechanism, we expect that the role of gp1 in ϕ29 DNA replication will be clarified.
We would like to thank Dr. K. Asai, Dr. F. Kawamura and Dr. K. Matsumoto for useful discussions and advices.

Fig. 1 .
Fig.1.Schematic representation of ϕ29 genome.Filled circles represent covalently linked terminal proteins.The direction of transcription is indicated by arrows.Left end early region is magnified.The genes are represented as white boxes and their products are indicated.This diagram is modified from previous report(Meijer et al., 2000).SSB; single-stranded DNA binding protein.DBP; double-stranded DNA binding protein.

Fig. 3 .
Fig. 3.The growth defects of ϕ29 phage in the absence of gp1 can be compensated by the amino acid substitutions in gp3.(A) Growth of ϕ29 ns1/ns3 double mutant at 42°C in various gp3-producing cells.Experimental conditions were essentially the same as described in Fig. 2. B. subtilis strains TT486 producing wild type gp3 (open circle), TT475 producing gp3-N69K (filled circle), TT459 producing gp3-E78K (open triangle) and TT559 producing gp3-E100K (filled triangle) were infected with ϕ29 ns1/ns3 mutant.At the same time with infection, xylose was added at a final concentration of 0.1% to express gene 3 alleles.Plaque forming unit was measured by plating on indicator strain TT541.(B) Growth of ϕ29 ns1/ns3 double mutant at 42°C in wild type gp3-producing cells.Experimental conditions were the same as Fig. 3A.In B. subtilis strains TT486, wild type gp3 was induced simultaneously with infection (open circles) or before 60 minutes of infection (closed circles).

Fig. 5 .
Fig. 5. Structure of gp3-DNA polymerase complex and model for the role of gp1.(A) Ribbon representation of gp3/DNA polymerase heterodimer.This figure was made with the Swiss-PdbViewer software (http://www.expasy.org/spdbv/)using PDB code 2EX3.DNA polymerase is colored in blue.Gp3 N-terminal, intermediate, and priming domains are colored in pink, yellow, and green, respectively.The positions of the amino acid substitutions in gp3 are indicated by arrows and are colored in red.Note that PDB code 2EX3 does not contain data of several amino acids including Asn at position 69.Thus, the position of Asn 69 was predicted from the position of Ala 71 .(B) Schematic representation of gp3-DNA polymerase complex and a model for the role of gp1.Gp1 and other components which have been reported or suggested to interact with gp3 are shown(Bravo and Salas, 1997;Bravo et al., 2000;Serna-Rico et al., 2000, 2003).The intermediate domain of gp3 was divided into two subdomains, a DNA polymerase interaction subdomain and a proposed regulatory subdomain.In this model, gp1 regulates gp3's interactions by acting as a bridging molecule or as a competitive inhibitor, or by inducing conformational change of gp3.

Table 3 .
Effect of temperature on plating efficiency of ϕ29 gene gp1) were mixed with each of ϕ29 strains and top agar, and plated on LB plate containing 0.1% xylose.After incubation of the plates at 30°C or 42°C for about 13 hours, the plaques appeared were counted.The plating efficiency of each ϕ29 mutant on B. subtilis TT275 at 30°C is expressed as 1, and the relative plating efficiencies were shown.These values were the average of three separate experiments.

Table 4 .
Temperature-resistant (Tr)strains isolated from ϕ29 gene 1 mutants a All Tr strains have the original mutations in gene 1.b Although sus1(629)-Tr1 strain has another four mutations in the left end region, these are found in parental sus1(