Identification of residues within two regions involved in self-association of viral histone-like protein p6 from phage theta29.

Protein p6 of Bacillus subtilis phage theta29 is involved in the initiation of viral DNA replication and transcription by forming a multimeric nucleoprotein complex with the phage DNA. Based on this, together with its abundance and its capacity to bind to the whole viral genome, it has been proposed to be a viral histone-like protein. Protein p6 is in a monomer-dimer-oligomer equilibrium association. We have identified protein p6 mutants deficient in self-association by testing random mutants obtained by degenerated polymerase chain reaction in an in vivo assay for dimer formation. The mutations were mainly clustered in two regions located at the N terminus, and the central part of the protein. Site-directed single mutants, corresponding to those found in vivo, have been constructed and purified. Mutant p6A44V, located at the central part of the protein, showed an impaired dimer formation ability, and a reduced capacity to bind DNA and to activate the initiation of O29 DNA replication. Mutant p6I8T has at least 10-fold reduced self-association capacity, does not bind DNA nor activate O29 DNA initiation of replication. C-terminal deletion mutants showed an enhanced dimer formation capacity. The highly acidic tail, removed in these mutants, is proposed to modulate the protein p6 self-association.

DNA transactions, such as DNA replication and transcription, require higher-order DNA-protein complexes, the assembly of which is sometimes facilitated by proteins with an architectural function. This is the case of the nucleoid-associated proteins, of which the ones most extensively studied are HU, H-NS, IHF, and FIS. They all bind and bend DNA, sharing some common characteristics such as being very abundant, having a small size, and playing a pleiotropic role. The function of these proteins requires formation of dimers or oligomers. Thus, HU is a homodimer in most bacteria, although it is a heterodimer in Escherichia coli (1,2), IHF is a heterodimeric protein encoded by two genes, himA (3) and hip (4), and FIS is a homodimer (5,6). On the other hand, H-NS monomers undergo self-association to form tetramers (7). The N-terminal domain of H-NS is involved in oligomerization (8,9), and the oligomeric structure of H-NS is necessary for recognition of intrinsically curved DNA and bending (10).
Protein p6 of Bacillus subtilis phage Ø29 is required in vivo for viral genome replication (11,12) and repression of transcription from the early promoter C2 (13). In vitro studies have shown that protein p6 is involved both in replication and transcription; it stimulates the initiation and the transition to elongation steps of Ø29 DNA replication (14,15), represses the early C2 promoter (16), and regulates the switch between early and late transcription (17). These functions are accomplished by the formation of a nucleoprotein complex in which the DNA adopts a right-handed toroidal conformation, and thus wrapping around a multimeric protein p6 core (18).
The number of copies of protein p6 in B. subtilis cells at late times of Ø29 infection has been calculated to be 6.6 ϫ 10 5 , enough to bind the entire viral progeny DNA (19). This, together with the ability to bind in vitro to the whole Ø29 genome, led us to propose a structural role in compacting and organizing the viral genome (20). The amounts of other histonelike proteins are cell cycle-dependent; thus, for E. coli cells in logarithmic growth, the most abundant ones are FIS and HU, with 1.2 ϫ 10 4 and 1.5 ϫ 10 5 copies per cell, respectively, whereas at stationary phase the most abundant is IHF with 1.2 ϫ 10 4 copies per cell (21). With these amounts, only a minor part of the bacterial genome would be bound by these proteins. Sedimentation equilibrium studies have shown that protein p6 is in a monomer-dimer equilibrium that shifts to higher association states at the millimolar concentrations found in vivo (19). These oligomeric structures have been observed by transmission electron microscopy (22), and their structure, as deduced by image processing, is compatible with that described for the path followed by the DNA in the protein p6⅐DNA complex (18). Thus, it has been proposed that protein p6 could act as a scaffold organizing the DNA into the appropriate configuration. Protein p6 binding to DNA is highly cooperative 1 and extends throughout the whole Ø29 DNA forming multiple complexes of very heterogeneous sizes; the minimal size ranges from ϳ130 base pairs (bp) 2 observed by electron microscopy after psoralen cross-linking, to ϳ80 -90 bp, shown by protection of micrococcal nuclease digestion (20). Thus, multiple protein-DNA and protein-protein interactions are required to stabilize the complex, suggesting that association equilibria among protein p6 oligomers would modulate their interaction with DNA.
In this work we have searched regions involved in protein p6 self-interaction. The characterization of deletion mutants has suggested that the N-terminal region is involved in protein p6 self-association, and the C-terminal acidic region interferes with it. We have obtained a collection of protein p6 random mutants unable to form dimers by using an in vivo self-association assay. The mutations were mainly clustered in two regions, located at the N-terminal and central parts of the protein. This allowed us to design site-directed mutants, which have been tested in vitro. The results obtained indicate that protein p6 dimer formation is drastically reduced by mutation I8T, whereas mutation A44V impairs dimerization and decreases the activation of Ø29 DNA initiation of replication.
Construction of Phage cI Repressor Fusions-The plasmid pBF21, containing the cI gene under the control of a tandem lacUV5 promoter-operator region (10), was digested with HindIII and EcoRV to remove the cI gene fragment encoding the oligomerization domain (OD) (Fig. 1). Plasmids p⌬cIp6 and p⌬cIp6m were made by a polymerase chain reaction (PCR) step cloning procedure. Wild-type (wt) gene 6, cloned in plasmid pPR55w6 (23), and mutant gene 6R6A, from plasmid pPR55R6A (23), were obtained by PCR by using primers designed to introduce HindIII and EcoRV restriction sites 5Ј-GAAAGTGGGAAAG-CTTTATGGCAA-3Ј and 5Ј-CCTTCTCTTGTGATATCATCATTCAGC-3Ј, respectively. PCRs were carried out with 1 M oligonucleotides, 0.2 g of pPR55w6 or pPR55R6A as templates, 100 M dNTP, and 2 units of Vent polymerase on its reaction buffer. PCR fragments were cloned into digested pBF21, to generate plasmids p⌬cIp6 and p⌬cIp6m, containing chimeric genes encoding for the DNA binding domain of the cI repressor, cI(DBD), fused to those encoding for p6wt protein or to mutant protein p6R6A, respectively (Fig. 1). The wt and mutant gene 6 were sequenced from the plasmids.
Random Mutagenesis in Gene 6 -Random mutagenesis of R6A single mutant gene 6 was performed on plasmid p⌬cIp6m. The oligonucleotides used generated HindIII and EcoRV restriction sites (5Ј-GGCTC-CAAGCCAAGCTTTATG-3Ј and 5Ј-CGGAATGGACGATATCATCA-3Ј). To obtain random mutations, PCR was carried out under two different conditions: one had limited dATP (20 M) and the other had additionally 0.5 mM MnCl 2 (24,25). Thus, the reactions contained 1 M oligonucleotides; 0.2 g of plasmid; 1.25 mM MgCl 2 ; 5 units of Taq polymerase; 100 M each of dCTP, dGTP, and dTTP; and 20 M dATP, in the absence or in the presence of 0.5 mM MnCl 2 in Taq polymerase buffer. The mutated fragments were digested with HindIII and EcoRV and cloned into plasmid pBF21 digested with the same enzymes. E. coli 71-18 lacI q cells (26) were transformed with these constructions and tested for phage development as described below. The mutated gene 6 was sequenced in those clones showing a lytic phenotype.
Phage Development-E. coli 71-18 lacI q cells transformed with plasmid pBF21 or the derivatives described above, were grown at 37°C up to an optical density of 0.6 at 620 nm in LB medium containing ampicillin (100 g/ml). The production of chimeric proteins was induced by addition of 250 l of 100 mM IPTG to 0.4 ml of culture. After induction, cells were infected with 146 hypervirulent phage (27), essentially as described (10). After 5-min incubation, 5 ml of LB medium containing ampicillin (100 g/ml) was added to the infected cultures, which were further incubated at 37°C. At the indicated times, 0.4-ml aliquots were centrifuged and 150 l of supernatant was mixed with 30 l of chloroform. The number of plaques forming units per milliliter in each supernatant was determined using E. coli 71-18 lacI q as the recipient strain.
Two oligonucleotides 5Ј-CGACCTGCAGGGATCCGGCC-3Ј and 5Ј-CTAATACTAGAAGCTTCTCTTGTG-3Ј were used to generate BamHI and HindIII restriction sites to clone the fragments containing mutated gene 6 into plasmid pPLc28 (28) under the control of P L promoter. The lysogen E. coli K-12⌬H1⌬trp encoding the thermosensitive cI857 repressor (28) was electroporated with the recombinant plasmids to obtain a library of random mutants. The mutations were determined by sequencing the corresponding gene 6 in selected clones from the library.
Purification of Protein p6 Deletion Mutants-Cultures of E. coli K-12⌬H1⌬trp cells transformed with the recombinant plasmids pMJn⌬5 and pMJn⌬13 (29) and pMJc⌬14 and pMJc⌬16 (30) were induced for 1 h at 42°C. After centrifugation, 20 g of cells was ground with alumina and extracted with buffer A (50 mM Tris-HCl, pH 7.5, 1 mM EDTA, 7 mM ␤-mercaptoethanol, 5% glycerol) with 0.5 M NaCl. Alumina and cell debris were removed by centrifugation. Polyethyleneimine (PEI) was added to the supernatant up to 0.25%, after adjusting absorbance at 260 nm to 120 units/ml, and centrifuged for 10 min at 10,000 ϫ g to remove DNA. The supernatant was made 0.35 M NaCl with buffer A; deletion mutant proteins were recovered in the pellet after centrifugation as above. To eliminate PEI, pellets were resuspended in buffer A containing 0.4 M NaCl, and the proteins were precipitated with ammonium sulfate up to 70% saturation. Pellets containing N-terminal deletion mutants, p6N⌬5 and p6N⌬13, were resuspended in buffer A and applied to a DEAE-cellulose column, whereas those containing C-terminal deletion mutants, p6C⌬14 and p6C⌬16, were applied to a phosphocellulose column. They were eluted with salt concentrations ranging from 75 to 200 mM NaCl. The mutant proteins were further subjected to heparin-agarose chromatography and eluted at salt concentrations from 50 to 350 mM NaCl. Minor contaminants were eliminated using Centricon-50 from Amicon; after this step the purity of the mutant proteins was estimated to be 95% by SDS-polyacrylamide gel electrophoresis (PAGE) followed by Coomassie Blue staining.
Purification of Protein p6 Site-directed Mutants-Cultures of E. coli K-12⌬H1⌬trp cells, transformed with plasmids encoding protein p6 site-directed mutants, were induced for 2 h at 42°C. After centrifugation, 6 -8 g of cells was ground with alumina, and the DNA was removed as described above. The mutant proteins were extracted from the PEI pellet by washing with 1 M NaCl buffer A and precipitated with ammonium sulfate to up to 70% saturation. Pellets containing mutant proteins were resuspended in buffer A and applied to a phosphocellulose column, except p6I8T, which was applied to a DEAE-cellulose column. The different protein p6 mutants were eluted with buffer A containing NaCl concentrations from 75 to 200 mM. Minor contaminants were eliminated by 15-30% glycerol gradient centrifugation. After this step all protein p6 single mutants were more than 90% homogeneous as estimated by SDS-PAGE followed by Coomassie Blue staining.
Initiation of DNA Replication Assay-The reaction mixture contained (in 25 l) 50 mM Tris-HCl, pH 7.5; 10 mM MgCl 2 ; 20 mM ammonium sulfate; 1 mM dithiothreitol; 10% glycerol; 0.25 M [␣-32 P]dATP; 0.3 g of phage Ø29 terminal protein-DNA complex, purified as described (31); 20 ng of terminal protein; 20 ng of Ø29 DNA polymerase; and the indicated amounts of either protein p6wt or protein p6 mutants. After incubation for 15 min at 15°C, the reaction was stopped by adding up to 10 mM EDTA and 0.1% SDS. The samples were filtered through Sephadex G-50 spin columns in the presence of 0.1% SDS. The initiation complex formed was analyzed by SDS-PAGE as described (31). Quantification was performed by using a Fuji BAS-1500 image analyzer.
Cross-linking with Glutaraldehyde-Glutaraldehyde cross-linking was carried out essentially as described (32). Either wt or mutant p6 proteins, 5 M each, were incubated at room temperature in 20 mM triethanolamine, pH 8.0, 50 mM NaCl, and 300 M glutaraldehyde. After 45-min incubation, the reaction was stopped by adding Tris-HCl, pH 7.5, up to 150 mM. Samples were analyzed by SDS-Tricine-PAGE (33).
Sedimentation Equilibrium Analysis-The experiments were performed in a Beckman Optima XL-A analytical ultracentrifuge equipped with absorbance optics, using an An60Ti rotor. Protein p6 was equilibrated in 50 mM Tris-HCl, pH 7.5, 10 mM MgCl 2 , and 50 mM NaCl. Protein p6 (100 l of 100 M) was centrifuged at 25,000 rpm, and absorbance scans at 280 nm were taken at sedimentation equilibrium. The equilibrium temperature was 4°C. High speed sedimentation (42,000 rpm) was conducted afterward for baseline correction.
Whole cell apparent weight average molecular weights (M w,a ) were determined by fitting a sedimentation equilibrium model for a single sedimenting solute to individual datasets with the program EQASSOC (supplied by Beckman, Ref. 34). The partial specific volume of protein p6 was 0.728 ml/g, calculated from the amino acid composition of the protein deduced from the gene 6 sequence (35).
DNase I Footprinting-DNase I footprinting was carried out essentially as described (36). The indicated amounts of either wt or mutant p6 proteins were incubated with 2 ng of Ø29 DNA left terminal fragment, 259 bp long. The DNA fragment was obtained by PCR in a reaction containing 0.2 g of Ø29 DNA as template, 1 M oligonucleotides 5Ј-GCGCTTTAAAGTAAGCCCCCACCCTC-3Ј and 5Ј-GCCCACA-TACTTTGTTGATTGG-3Ј, 100 M dNTP, and 2 units of Vent polymerase in its reaction buffer. The last oligonucleotide was previously 5Јlabeled with [␥-32 P]ATP using T4 polynucleotide kinase to obtain the Ø29 DNA terminal fragment internally labeled. The PCR product was digested with DraI to restore the Ø29 DNA terminal end. DNase I digestion was carried out for 5 min at 15°C in 25 l of a solution containing 10 mM MgCl 2 . The reaction was stopped by addition of EDTA, pH 8.0, up to 20 mM final concentration. DNA was precipitated with ethanol in the presence of RNA carrier (0.25 mg/ml) and then subjected to denaturing electrophoresis in 6% (w/v) acrylamide gel.
Circular Dichroism-Circular dichroism spectra were recorded using a Jasco J-720 spectropolarimeter using a 0.2-cm path length cuvette. Proteins were used at room temperature at a concentration of 0.13 g/l in 10 mM Tris-HCl, pH 7.5. The final spectrum in each case is the mean of 10 measurements with a step size of 0.2 nm recorded from 260 to 200 nm. Data were base line-corrected by substraction of that of the Tris buffer.

Self-association of Protein p6 Deletion
Mutants-It has been reported that N-terminal deletion mutants of protein p6 lacking 5 or 13 amino acids (p6N⌬5 and p6N⌬13, respectively) were not able to interact with a Ø29 DNA terminal fragment as detected by DNase I footprinting (29). Because DNA binding is highly cooperative 1 and requires the formation of large oligomers, the lack of DNA binding could be due to a deficiency in protein p6 self-association; therefore, we tested by glutaraldehyde cross-linking the capability of both deletion mutants to self-associate. Fig. 2 shows that dimer formation was highly impaired in the p6N⌬5 mutant, being about 10% that of the p6wt protein. In addition, dimers were not detected with the p6N⌬13 mutant (not shown). On the other hand, a 14-amino acid C-terminal deletion mutant of protein p6 (p6C⌬14) bound DNA 2-fold better than the p6wt protein (30). Two C-terminal deletion mutants of 14 and 16 amino acids (p6C⌬14 and p6C⌬16, respectively) were tested for dimer formation by glutaraldehyde cross-linking. As Fig. 2 shows, the ability of dimer formation in the p6C⌬16 mutant protein was about 2-fold that of the p6wt protein; a similar result was obtained with the mutant protein p6C⌬14 (not shown). Therefore, the observed DNA binding properties of p6N⌬5, p6N⌬13, p6C⌬14, and p6C⌬16 mutants could be explained by their self-association capacities. These results indicate that the N-terminal region of protein p6 is required for glutaraldehyde cross-linking, suggesting that this region is involved in dimer formation; however, glutaraldehyde cross-linking is favored when the C-terminal region is removed, suggesting that this region impairs dimer formation.
Self-interaction of Protein p6 in Vivo-To further investigate the protein p6 regions involved in self-association, we carried out an in vivo assay in which we used the region encoding the N-terminal domain of cI repressor as a reporter gene for dimerization (10). Phage cI repressor is a two-domain protein that binds DNA as a dimer; the N-terminal part is the DNA binding domain, cI(DBD), whereas the C-terminal one is the oligomerization domain, cI(OD), which is required for an efficient binding to the operator. Thus, E. coli cells expressing only cI(DBD) (⌬cI) were sensitive to hypervirulent phage 146 infection, whereas those with intact cI were immune. Replacement of the cI(OD) by another protein provides a self-interaction assay for that protein.
We have used the plasmid pBF21 (10), expressing the phage cI repressor under the control of the lac promoter, to construct p⌬cIp6 plasmid encoding a fusion protein containing the ⌬cI and the protein p6wt (Fig. 1). The dimerization capacity of protein p6 was assayed by determining the development of infecting 146 phage in E. coli cells harboring the p⌬cIp6 plasmid upon induction with IPTG. Cells expressing the ⌬cIp6 fusion protein were not immune to 146 infection (Fig. 3), behaving as the control cells expressing ⌬cI. This result was unexpected, because protein p6 self-interacts in vitro. The expression of ⌬cIp6 fusion protein was confirmed by Western blot analysis with protein p6 antiserum (not shown). An explanation for the lytic phenotype could be that the non sequencespecific DNA binding ability of protein p6 (37) prevented the binding of the ⌬cI to the operators. To avoid this, plasmid p⌬cIp6m (see Fig. 1) was constructed, expressing a fusion containing the ⌬cI and the protein p6 mutant defective in DNA binding p6R6A (38). Cells harboring the plasmid p⌬cIp6m showed, upon induction, a time course development of 146 phage similar to that of the control cells expressing intact cI repressor. In these cases, the number of pfu/ml was about 10 4fold lower than that obtained in cells expressing ⌬cI (Fig. 3). Therefore, we can conclude that the mutant protein p6R6A functionally replaces the cI(OD) allowing dimerization of the ⌬cI. This strongly suggests that protein p6 self-associates in vivo.
Random Mutagenesis of Gene 6 and in Vivo Selection of Protein p6 Mutants Defective in Self-interaction-The in vivo assay described above sets up the basis to select mutants in self-association. Because ⌬cIp6m fusion protein confers immunity to 146 infection, random mutagenesis was performed on mutant gene 6R6A contained in plasmid p⌬cIp6m. The PCR products were cloned as above, under the control of the lac promoter, in-frame to the cI(DBD). Therefore, chimeric proteins consisting of ⌬cI and random mutants of protein p6R6A were expressed. Individual colonies were tested for phage development as above, and clones showing a lytic phenotype were selected as candidates defective in protein p6 self-interaction. From 200 individual clones, 35 showed a lytic phenotype. Fig. 4 shows the development of 146 phage in selected lytic clones, which was about 10 4 -fold higher than that of cells expressing ⌬cIp6m. The expression of the fusion proteins was assessed by Western immunoblotting analysis using antiserum against protein p6 (not shown).
The entire nucleotide sequence of the mutated gene 6 was determined from the 35 lytic clones. Four of them had a premature stop codon. Most of the mutants had more than three amino acid changes, and only five carried double or triple mutations as shown in Fig. 5. Amino acid sequence comparison with the p6R6A protein (p6m) showed that the mutations were mainly clustered in two regions: N-terminal, around position 8 (region I, Fig. 5), like those in p6m8 and p6m22 mutant proteins, and a central region around position 44 (region II, Fig. 5), like p6m23 and p6m41 mutant proteins. Mutant protein p6m94 carried a single mutation in each region.
Residues Involved in Self-association of Protein p6 in Vitro-The non-random distribution of in vivo selected mutants strongly suggested the involvement of regions I and II in protein p6 self-interaction. To assess directly the involvement of individual residues in protein-protein interaction, we designed site-directed mutations on positions 7, 8, 43, and 44, where wt residues were replaced by those found in mutants selected in vivo (Fig. 5), namely E7V, I8T, Q43R, and A44V. The sitedirected mutagenesis was performed by PCR on plasmid pPR55w6, encoding wt gene 6 (23). The mutated genes were cloned into an expression vector under the control of the bacteriophage P L promoter. Mutants were overproduced and purified up to at least 90% homogeneity. Protein p6 mutants were tested for activation of initiation of phage Ø29 DNA replication in an in vitro assay with purified proteins (14). Fig.  6 shows the results obtained with the mutant proteins p6E7V, p6I8T, p6Q43R, and p6A44V. Although the mutants p6E7V and p6Q43R showed an activation similar to that of the p6wt protein, the p6A44V mutant protein exhibited about a 6-fold reduced capacity to activate the initiation reaction. No activation was detected by the p6I8T mutant. The deficiency observed in the activation of the initiation reaction in p6A44V and p6I8T mutant proteins could be due to an impaired selfassociation and/or DNA binding. Because we have selected these mutants in an in vivo self-association assay, we first studied the dimerization properties of the protein p6 mutants. The effect of the p6A44V and p6I8T mutations in self-interaction was tested, as a first approach, by glutaraldehyde crosslinking. The amount of dimers formed by the p6A44V and p6I8T mutant proteins was 50% and 7%, respectively, of that found in the p6wt protein (Fig. 7). These results prompted us to further study the self-interaction capability of the protein p6 mutants by analytical ultracentrifugation. Sedimentation equilibrium studies had shown that protein p6, in the 1-100 M range, is in a monomer-dimer equilibrium with a dimerization constant (K 2 ) of 2 ϫ 10 5 M Ϫ1 that shifts to higher association states at higher protein concentration (19). Fig. 8 shows that, at 100 M and 4°C, the M w,a ) of protein p6wt was 25,400, whereas those of p6A44V and p6I8T mutant proteins were 20,600 and 15,500, respectively, the latter being close to the theoretical value for the wt protein p6 monomer (11,900). Assuming that mutant proteins have the same self-association behavior as p6wt (19), the dimerization constant of p6A44V is slightly lower than that of the p6wt 3 ; however, the K 2 of p6I8T is, at least, 10-fold lower than that of the p6wt protein. 3 Therefore, we can conclude that A44V mutation slightly affects protein dimerization, which is almost abolished by the mutation I8T.
Protein p6 forms a nucleoprotein complex with phage Ø29 DNA terminal fragments, in which the DNA wraps around a multimeric protein core (39). The formation of the complex can be detected by a characteristic DNase I footprint pattern, in which strong hypersensitivities, with a periodicity of 24 bp, are located in between protected regions (36). As shown in Fig. 9, protein p6A44V required a 2.5-fold higher protein amount (0.5 g) than protein p6wt (0.2 g) to form the complex; this may reflect the reduced capacity to form dimers of p6A44V mutant. In addition, we can observe that the digestion pattern formed with p6A44V is slightly different than that observed with p6wt. As it could be expected from the previous results, p6I8T mutant protein failed to bind DNA even at a 20-fold higher protein concentration than that of the p6wt (Fig. 9). DISCUSSION Protein p6 of B. subtilis phage Ø29 binds DNA forming a multimeric nucleoprotein complex (39) that is required for activation of Ø29 DNA initiation of replication (37,38), repression of transcription from early C2 promoter (13,16), and regulation of viral switch between early and late transcription (17). Protein p6 in solution forms elongated oligomers from preformed dimers, at the in vivo estimated protein concentration (19). Oligomeric structures have been proposed to provide the scaffold on which the DNA folds into the appropriate configuration. Protein p6 binding to DNA requires multiple protein-DNA and protein-protein interactions. Thus, failure to bind DNA of N-terminal deletion mutants of 5 or 13 amino acids in protein p6 (29), can be explained by their impaired or lost capacity, respectively, in dimer formation.
We have looked for other regions involved in protein p6 self-association by using an in vivo system based on the immunity of cells expressing cI repressor to phage infection. If the cI(OD) is removed, immunity is lost and the infecting phage can develop lytic cycle. Thus, replacement of the cI(OD) by other protein provides a self-interaction test for this protein.
When ⌬cI was fused to the gene encoding p6R6A mutant protein, immunity was achieved, suggesting that the protein selfinteracts in vivo. Thus, random mutants of protein p6R6A, obtained by degenerated PCR, were tested for self-association. Although the proteins contained multiple mutations, they were not randomly arranged, but mainly clustered in two regions located at the N-terminal and the central region of protein p6 (regions I and II in Fig. 5). Therefore, single site-directed mu-3 G. Rivas, personal communication. tants of protein p6wt corresponding to some of those found in vivo were constructed and assayed in vitro. Taking into account that the in vivo selected mutant proteins p6m23 and p6m41 (Fig. 5) share the Q43R mutation, we would expect an involvement of Gln 43 in protein p6 self-interaction. However, this was not confirmed by in vitro experiments with the corresponding site-directed mutant. The mutant protein p6Q43R did not show any significant difference with the wt protein in the activation of Ø29 DNA initiation of replication or glutaraldehyde crosslinking. A possible explanation for this result could be that the in vivo assay is more restrictive. In addition, the single mutation may not be enough to show an impaired dimerization, and additional changes might be required, like those found in the in vivo selected mutants.
The involvement of region II in dimer formation is shown by p6A44V mutant protein, which has a slightly reduced capacity to form dimers. In agreement with this, DNase I footprinting (Fig. 9) shows that p6A44V has a DNA binding affinity lower than that of the p6wt protein and an impaired activation of Ø29 DNA initiation of replication.
The involvement in dimer formation of region I (see Fig. 5) was expected from the results obtained with protein p6 Nterminal deletion mutants. The p6I8T mutant protein failed to activate the Ø29 DNA initiation of replication, DNA binding was not detected, and dimer formation was reduced at least 10-fold. Mutant p6 proteins are folded, as shown in Fig. 10, where the circular dichroism spectra of the wt and mutant p6 proteins do not show significant differences in their secondary structure. Altogether, these results suggest that Ile 8 is involved in the self-interaction of protein p6.
It has been reported that a C-terminal, 14-amino acid truncated p6 protein has a DNA binding activity 2-fold higher than that of the p6wt protein (30). We have now shown that this truncated protein has increased its capacity to form dimers. Protein p6 has a very acidic C terminus, where 10 out of 19 residues are either Asp or Glu (see Fig. 5). Therefore, it seems likely that the C-terminal acidic region modulates protein p6 self-association, interacting with a basic region. There is a cluster of basic residues overlapping region II (see Fig. 5), although interaction with region I, which also contains basic residues, cannot be ruled out. This interaction could contribute to the dynamic nature of the protein p6 binding to DNA (20).
In summary, we have found two regions involved in vivo in self-association of phage Ø29 protein p6, the N-terminal and a central region. Site-directed mutation A44V, located at the central region, impairs the protein p6 self-association. The N-terminal mutation I8T has a more drastic effect and completely abolished dimer formation.