Effect of Mg2+ on the DNA Binding Modes of the Streptococcus pneumoniae SsbA and SsbB Proteins*

The effect of Mg2+ on the binding of the Streptococcus pneumoniae single-stranded DNA binding (SSB) proteins, SsbA and SsbB, to various dTn oligomers was examined by polyacrylamide gel electrophoresis. The results were then compared with those that were obtained with the well characterized SSB protein from Escherichia coli, SsbEc. In the absence of Mg2+, the results indicated that the SsbEc protein was able to bind to the dTn oligomers in the SSB35 mode, with only two of the four subunits of the tetramer interacting with the dTn oligomers. In the presence of Mg2+, however, the results indicated that the SsbEc protein was bound to the dTn oligomers in the SSB65 mode, with all four subunits of the tetramer interacting with the dTn oligomers. The SsbA protein behaved similarly to the SsbEc protein under all conditions, indicating that it undergoes Mg2+-dependent changes in its DNA binding modes that are analogous to those of the SsbEc protein. The SsbB protein, in contrast, appeared to bind to the dTn oligomers in an SSB65-like mode in either the presence or the absence of Mg2+, suggesting that it may not exhibit the pronounced negative intrasubunit cooperativity in the absence of Mg2+ that is required for the formation of the SSB35 mode. Additional experiments with a chimeric SsbA/B protein indicated that the structural determinants that govern the transitions between the different DNA binding modes may be contained within the N-terminal domains of the SSB proteins.

The naturally transformable Gram-positive bacterium Streptococcus pneumoniae has two single-stranded DNA-binding (SSB) 2 proteins, designated SsbA and SsbB (Fig. 1). The SsbA protein is similar in size to the extensively studied SSB protein from Escherichia coli, SsbEc, a homotetrameric, non-sequence-specific, single-stranded DNA (ssDNA)-binding protein that is involved in many aspects of DNA metabolism in E. coli (1,2). The SsbB protein, in contrast, is a smaller protein that is specifically induced during natural transformation in S. pneumoniae (3,4). These results suggest that the SsbA protein may serve as a general SSB protein involved in routine DNA functions (analogous to the SsbEc protein) and that the SsbB protein may be a specialized SSB protein used primarily during natural transformation. Consistent with this idea, a recent survey revealed that those naturally transformable Gram-positive bacteria that are related to S. pneumoniae generally have two ssb-like genes, whereas the non-naturally transformable Gram-negative bacteria that are related to E. coli have only a single ssb gene (5).
We have recently amplified the ssbA and ssbB genes from S. pneumoniae genomic DNA, developed efficient expression systems, and purified the SsbA and SsbB proteins to apparent homogeneity (1,6). In our initial investigations, we found that the SsbA and SsbB proteins, like the SsbEc protein, formed stable homotetramers in solution (7). However, although the ssDNA binding properties of the SsbA protein appeared to be similar to those of the SsbEc protein, the ssDNA binding characteristics of the SsbB protein were quite different. For example, although the SsbB protein was able to bind to the shorter oligomer dT 50 with an affinity similar to that of the SsbEc and SsbA proteins, our results indicated that two SsbEc or SsbA tetramers were able to bind to the longer oligomer dT 75 , whereas only a single SsbB tetramer was able to bind to this ssDNA. The apparent differences in the stoichiometries of binding to dT 75 and other longer oligomers indicated that the SsbB protein interacts with ssDNA in a manner different from that of the SsbEc and SsbA proteins (7).
The ssDNA binding properties of the SsbEc protein have been shown to exhibit a complex dependence on solution conditions (2). In general, at lower DNA binding densities and higher monovalent or divalent salt concentrations, the SsbEc protein binds in either the SSB 56 or the SSB 65 mode in which all four subunits of the tetramer interact with ssDNA, occluding ϳ56 or 65 nucleotides of ssDNA/tetramer, respectively (since the physical distinction between the SSB 56 and SSB 65 modes is not clear, they will be referred to collectively here as the SSB 65 mode). At higher DNA binding densities and lower salt concentrations, however, the SsbEc protein can bind in the SSB 35 mode in which only two subunits of the tetramer interact with ssDNA, occluding ϳ35 nucleotides of ssDNA/tetramer (2).
Our initial analysis of the ssDNA binding properties of the SsbA and SsbB proteins was based on polyacrylamide gel shift assays that were carried out using the standard gel electrophoresis running buffer, TBE (Tris borate (pH 8.5)/EDTA) (7). However, SSB proteins have been found to act as accessory factors for a variety of enzymes involved in DNA metabolism, many of which are dependent on Mg 2ϩ for activity. We have therefore examined the effect of Mg 2ϩ on the ssDNA binding properties of the SsbA and SsbB proteins. The ssDNA binding reactions were carried out in a typical enzyme reaction buffer (25 mM Tris acetate (pH 7.5)) with various concentrations of magnesium acetate, and the resulting complexes were analyzed by electrophoresis in polyacrylamide gels that contained the same buffer and magnesium acetate concentrations as in the reaction solutions. The results of this analysis have led to an explanation for the previously observed differences in the ssDNA binding properties of the SsbA and SsbB proteins. In addition, a new chimeric SSB protein was prepared as a first step toward identifying the structural basis for the differences in the ssDNA binding properties of the SsbA and SsbB proteins.

EXPERIMENTAL PROCEDURES
Materials-S. pneumoniae SsbA protein (1) and SsbB protein (6) were prepared as described. E. coli SSB protein was from Promega. dT n oligomers were from Invitrogen; 32 P-end-labeled dT n oligomers were prepared using [␥-32 P]ATP (Amersham Biosciences) and T4 polynucleotide kinase (New England Biolabs). G-250 BioSafe Coomassie Brilliant Blue protein gel stain was from Bio-Rad. dT n concentrations were determined by absorbance at 260 nm using the extinction coefficient 8.4 mM Ϫ1 cm Ϫ1 (2).
Preparation of the Chimeric SsbA/B Protein-A NruI restriction site was introduced into each of the previously described pETssbA (1) and pETssbB constructs (6) using oligonucleotide-directed mutagenesis (8); the NruI site was designed such that although nucleotide changes were made in the coding sequence of both ssbA and ssbB genes, the corresponding amino acids were not altered. The constructs were then digested with NdeI and NruI to generate fragments encoding the N-terminal region of the SsbA protein (amino acids 1-105) and the C-terminal region of the SsbB protein (amino acids 105-131 plus the pET21a vector), respectively. These fragments were ligated together to give the final construct, pETssbA/B. The insert was sequenced and found to be identical to the anticipated ssbA-(1-105)/ssbB-(105-131) chimeric sequence. The protein encoded by this insert will be referred to as the SsbA/B protein in the results described below.
The SsbA/B protein was expressed in E. coli strain Rosetta(DE3)-pLysS (Novagen). Competent Rosetta(DE3)pLysS cells were transformed with pETssbA/B and selected for growth on LB/carbenicillin/chloramphenicol plates. A single Rosetta(DE3)pLysS/pETssbA/B colony was used to inoculate 2ϫ YT broth (10 ml) containing ampicillin (50 g/ml) and chloramphenicol (34 g/ml), and the resulting culture was incubated overnight at 37°C. This overnight culture was used to inoculate 1 liter of 2ϫ YT broth/ampicillin/chloramphenicol medium, and the cells were grown at 37°C to an A 600 of 0.8. Isopropyl-1-␤-D-thiogalactopyranoside (1 mM) was added to induce expression of the SsbA/B protein. After 2.5 h at 37°C, the cells were collected by centrifugation, suspended in a 2:1 ratio of 50 mM Tris-HCl (pH 7.5), 10% glycerol, 1 mM EDTA, 3 mM spermidine resuspension buffer to wet cell weight, and frozen in liquid nitrogen. The cell suspension was stored at Ϫ80°C.
The SsbA/B protein was purified by a procedure similar to that described previously for the SsbA (1) and SsbB proteins (6), with the following modification. Instead of utilizing a Polymin P precipitation, the SsbA/B protein was precipitated from the crude lysate fraction with ammonium sulfate at a final concentration of 0.25 g/ml. The ammonium sulfate precipitation step was then followed by MonoQ and heparin chromatography steps (1,6). The purified SsbA/B protein (greater than 95% homogeneity) was stored in 20 mM Tris-HCl (pH 7.5), 10% glycerol, 1 mM dithiothreitol, 0.1 M NaCl.
Polyacrylamide Gel Electrophoresis Mobility Shift Assays-The ssDNA binding reaction solutions contained 25 mM Tris acetate (pH 7.5), 5% glycerol, 1 mM dithiothreitol, and the concentrations of magnesium acetate, dT n , and SSB protein given in the legends for Figs. 2-8 and 10. The reaction solutions were incubated at 37°C for 15 min. Aliquots (20 l) were removed from each reaction solution and added to 2 l of gel loading solution (0.25% bromphenol blue, 40% sucrose). The aliquots were analyzed by electrophoresis on 5% native polyacrylamide gels using a buffer system consisting of 25 mM Tris acetate (pH 7.5) and the same concentration of magnesium acetate as in the reaction solutions. Bands corresponding to unbound and SSB-bound dT n oligomers were visualized by autoradiography (these reactions contained 32 P-endlabeled dT n oligomers), and bands corresponding to unbound and dT nbound SSB protein were visualized by G-250 BioSafe Coomassie Brilliant Blue staining (these reactions contained only unlabeled dT n oligomers).

Binding Modes of the SsbEc Protein
In the initial set of experiments, the complexes that were formed between the SsbEc protein and the oligomers dT 75 and dT 35 were examined to establish whether polyacrylamide gel electrophoresis could be used to analyze the binding modes of SSB proteins. The binding reactions were carried out in a reaction buffer consisting of 25 mM Tris acetate (pH 7.5) and either 0 or 10 mM magnesium acetate, and the resulting complexes were analyzed by electrophoresis in polyacrylamide gels that contained the same buffer and magnesium acetate concentrations as in the reaction solutions.
Forward Titrations with dT 75 -A set of forward titration experiments was carried out in which a fixed concentration of 32 P-end-labeled dT 75 was incubated with increasing concentrations of SsbEc protein.
The complexes that were resolved on the polyacrylamide gels were visualized by autoradiography to monitor the changes in the mobility of the dT 75 that occur when it binds to SsbEc protein.
When increasing concentrations of SsbEc protein were added to dT 75 in the absence of Mg 2ϩ , an initial complex with a gel mobility lower than that of unbound dT 75 was formed at the lower protein concentrations (Fig. 2). A further increase in the concentration of SsbEc protein resulted in the disappearance of this initial complex and the appearance of a second complex of even lower gel mobility (Fig. 2). These results are similar to those that were obtained previously using the TBE buffer system and indicated that two tetramers of SsbEc protein were able to bind to dT 75 at the higher protein concentrations in the absence of Mg 2ϩ (7).
A different result was obtained when increasing concentrations of SsbEc protein were added to dT 75 in the presence of Mg 2ϩ (10 mM) (Fig.  2). In this case, an initial complex was again formed at the lower protein concentrations. However, in contrast to the results that were obtained in the absence of Mg 2ϩ , there was no indication of the formation of the second complex in the presence of Mg 2ϩ , even at the highest concentration of SsbEc protein examined (Fig. 2). These results indicated that only a single tetramer of SsbEc protein was able to bind to dT 75 in the presence of Mg 2ϩ .
These results suggested that in the absence of Mg 2ϩ , the SsbEc protein was able to bind to dT 75 in the previously described SSB 35 mode, with two of the four subunits of each of the two bound tetramers interacting with the dT 75 (occluding 35 nucleotides/tetramer), whereas in the presence of Mg 2ϩ , the SsbEc protein was binding to dT 75 in the SSB 65 mode, with all four subunits of a single tetramer interacting with the dT 75 (occluding 65 nucleotides) (2). Reverse Titrations with dT 35 -A complementary set of reverse titration experiments was carried out using the oligomer, dT 35 , which is only long enough to bind to two of the four subunits of the SsbEc tetramer (2). In these reactions, a fixed concentration of SsbEc protein was incubated with increasing concentrations of dT 35 . The complexes that were resolved on the polyacrylamide gels were visualized by protein staining to monitor the changes in the mobility of the SsbEc protein that occur when it binds to dT 35 .
When the SsbEc protein was incubated with increasing concentrations of dT 35 in the absence of Mg 2ϩ , a new band with a gel mobility greater than that of the free SsbEc protein was formed (Fig. 3). A parallel reverse titration experiment that was carried out using 32 P-end-labeled dT 35 confirmed that the new band corresponded to an SsbEc-(dT 35 ) n complex (gel not shown). The concentration of dT 35 that was required to convert the free SsbEc protein to the complex corresponded to approximately one dT 35 molecule/tetramer of SsbEc protein. These results indicated that in the absence of Mg 2ϩ , the SsbEc tetramer was able to bind a single dT 35 molecule (through two of the four subunits of the tetramer) to form an SsbEc-dT 35 complex (the additional negative charges that are contributed by the phosphoryl groups of the bound dT 35 presumably increase the mobility of the SsbEc-dT 35 complex relative to that of the free SsbEc protein).
A different result was obtained when the SsbEc protein was incubated with increasing concentrations of dT 35 in the presence of Mg 2ϩ (10 mM) (Fig. 3). In this case, an initial complex was formed at the same concentration of dT 35 that was observed for the formation of the SsbEc-dT 35 complex in the absence of Mg 2ϩ . However, in contrast to the results that were obtained in the absence of Mg 2ϩ , a further increase in the concentration of dT 35 resulted in the disappearance of the first complex and the appearance of a new band with even greater gel mobility (Fig. 3). A parallel reverse titration experiment that was carried out using endlabeled dT 35 confirmed that both the first and the second bands corresponded to SsbEc-(dT 35 ) n complexes and that the second complex contained approximately twice as much dT 35 as the first (gel not shown). These results indicated that in the presence of Mg 2ϩ , the SsbEc tetramer was able to bind two dT 35 molecules (through all four subunits of the tetramer) to form an SsbEc-(dT 35 ) 2 complex (the additional negative charges contributed by the phosphoryl groups of the second dT 35 presumably increase the mobility of the doubly liganded complex relative to that of the singly liganded complex).
The results of the reverse titrations were consistent with those from the forward titrations and indicated that the SsbEc protein was able to bind to the dT n oligomers in the SSB 35 mode (with two of four subunits interacting with dT n ) in the absence of Mg 2ϩ and in the SSB 65 mode (with all four subunits interacting with dT n ) in the presence of Mg 2ϩ (10 mM). These results are consistent with previous fluorescence studies of the SsbEc protein (see "Discussion") and indicated that the Mg 2ϩ -polyacrylamide gel electrophoresis method described here could be used to analyze the binding modes of the SsbA and SsbB proteins.

Binding Modes of the SsbA Protein
Forward and reverse titration experiments analogous to those described above for the SsbEc protein were used to analyze the DNA binding modes of the SsbA protein.
Forward Titrations with dT 75 -When increasing concentrations of SsbA protein were added to 32 P-end-labeled dT 75 in the absence of Mg 2ϩ , an initial complex with a gel mobility lower than that of unbound dT 75 was formed at the lower protein concentrations (Fig. 4). A further increase in the concentration of SsbA protein resulted in a decrease in this initial complex and the appearance of a second complex of even lower gel mobility (Fig. 4). These results were similar to those that were obtained for the SsbEc protein (Fig. 2) and to those that were obtained previously for the SsbA protein using the TBE buffer system (7) and indicated that two tetramers of SsbA protein were able to bind to dT 75 in the absence of Mg 2ϩ .
When increasing concentrations of SsbA protein were added to dT 75 in the presence of Mg 2ϩ , an initial complex was again formed at the lower protein concentrations (Fig. 4). However, in contrast to the results that were obtained in the absence of Mg 2ϩ , there was no indication of the formation of the second complex in the presence of Mg 2ϩ , even at the highest concentration of SsbA protein examined (Fig. 4). These results were again similar to those that were obtained for the SsbEc protein (Fig. 2) and indicated that only one tetramer of SsbA protein was able to bind to dT 75 in the presence of Mg 2ϩ .
Reverse Titrations with dT 35 -When the SsbA protein was incubated with increasing concentrations of dT 35 in the absence of Mg 2ϩ , a single complex with a gel mobility greater than that of the free SsbA protein  was formed (Fig. 5). These results were similar to those that were obtained with the SsbEc protein (Fig. 3) and indicated that in the absence of Mg 2ϩ , the SsbA tetramer was able to readily bind a single dT 35 molecule to form an SsbA-dT 35 complex (the slight increase in the apparent mobility of this complex at higher dT 35 concentrations suggests that a second dT 35 may bind weakly to the SsbA protein under these conditions).
When the SsbA protein was incubated with increasing concentrations of dT 35 in the presence of Mg 2ϩ , an initial complex was formed at the same concentration of dT 35 that was observed for the formation of the SsbA-dT 35 complex in the absence of Mg 2ϩ (Fig. 5). However, in contrast to the results that were obtained in the absence of Mg 2ϩ , a further increase in the concentration of dT 35 resulted in the disappearance of the first complex and the appearance of a new complex with even greater gel mobility (Fig. 5). These results were again similar to those that were obtained with the SsbEc protein (Fig. 3) and indicated that in the presence of Mg 2ϩ , the SsbA tetramer was able to bind two dT 35 molecules to form an SsbA-(dT 35 ) 2 complex.
The similarity of the results that were obtained in the forward and reverse titration experiments, both in the absence and in the presence of Mg 2ϩ , indicated that the SsbA and SsbEc proteins undergo similar Mg 2ϩdependent changes in the modes in which they bind to ssDNA.

Binding Modes of the SsbB Protein
Forward and reverse titration experiments were also carried out to analyze the DNA binding modes of the SsbB protein.
Forward Titrations with dT 75 -When increasing concentrations of SsbB protein were added to 32 P-end-labeled dT 75 in the absence of Mg 2ϩ , a complex with a gel mobility lower than that of unbound dT 75 was formed at the lower protein concentrations (Fig. 6). However, in contrast to the results that were obtained with the SsbA and SsbEc proteins, there was no indication of the formation of a second complex with dT 75 by the SsbB protein in the absence of Mg 2ϩ , even at the highest concentration of SsbB protein examined (Fig. 6). These results are similar to those that were obtained previously using the TBE buffer system and indicated that only one tetramer of SsbB protein was able to bind to dT 75 in the absence of Mg 2ϩ (7).
A similar result was obtained when increasing concentrations of SsbB protein were added to dT 75 in the presence of Mg 2ϩ (Fig. 6). In this case, an initial complex was formed at the lower protein concentrations, but there was no indication of the formation of a second complex with dT 75 by the SsbB protein, even at the highest concentrations of SsbB protein examined. These results indicated that only one tetramer of SsbB protein was able to bind to dT 75 in the presence of Mg 2ϩ .
Reverse Titrations with dT 35 -When the SsbB protein was incubated with increasing concentrations of dT 35 in the absence of Mg 2ϩ , a new band with a gel mobility greater than that of the free SsbB protein was formed (Fig. 7). However, in contrast to the results that were obtained with the SsbEc and SsbA proteins in the absence of Mg 2ϩ , a further increase in the concentration of dT 35 resulted in the disappearance of the first band and the appearance of a new band with even greater gel mobility (Fig. 7). A parallel reverse titration experiment that was carried out using 32 P-end-labeled dT 35 demonstrated that both the first and the second bands corresponded to SsbB-(dT 35 ) n complexes (gel not shown). These results suggested that in the absence of Mg 2ϩ , the SsbB tetramer was able to bind two molecules of dT 35 (through all four subunits of the tetramer) to form an SsbB-(dT 35 ) 2 complex.
A similar result was obtained when the SsbB protein was incubated with increasing concentrations of dT 35 in the presence of Mg 2ϩ (Fig. 7). However, in this case, the first and second complexes appeared concurrently (rather than sequentially) at the lower dT 35 concentrations. A further increase in the dT 35 concentration resulted in the disappearance of the first complex and an increase in the intensity of the second complex. A parallel reverse titration experiment that was carried out using end-labeled dT 35 confirmed that both bands corresponded to SsbB-(dT 35 ) n complexes (gel not shown). These results suggested that in the presence of Mg 2ϩ , the SsbB tetramer was able to bind two dT 35 molecules (through all four subunits of the tetramer), perhaps with positive intrasubunit cooperativity, to form an SsbB-(dT 35 ) 2 complex.
The results from both the forward and the reverse titrations indicated that the SsbB protein differed from the SsbEc and SsbA proteins in that it appeared to be able to bind to the dT n oligomers with all four subunits of the tetramer in either the presence or the absence of Mg 2ϩ .

Dependence of DNA Binding Modes on Mg 2؉ Concentration
To more precisely define the Mg 2ϩ dependence of the DNA binding modes of the SsbEc, SsbA, and SsbB proteins, the binding of these pro-  teins to dT 75 was examined over a range of Mg 2ϩ concentrations (0 -10 mM). The reactions were carried out with an excess of SSB protein relative to dT 75 so that complexes containing either one SSB tetramer (SSB 65 -like mode) or two SSB tetramers (SSB 35 -like mode) could be identified.
The dependence of the binding of the SsbEc protein to dT 75 on Mg 2ϩ concentration is shown in Fig. 8. In the absence of Mg 2ϩ , the dT 75 was almost completely converted into the second complex with only a small amount of the first complex apparent. This was consistent with the results in Fig. 2 and indicated that two SsbEc tetramers were able to bind to dT 75 under these conditions. At 1 mM Mg 2ϩ , the band corresponding to the second complex was fainter, and there was a smear leading down to the position where the first complex was found. This indicated that the second complex was able to form under these conditions but was less stable than in the absence of Mg 2ϩ . At 2.5 mM Mg 2ϩ , the dT 75 appeared mainly as a band at the position of the first complex with an upward smear, suggesting that some dissociation of the second complex had occurred during electrophoresis. At 5 mM Mg 2ϩ , the dT 75 appeared as a discrete band at the position corresponding to the first complex, with no indication of the formation of the second complex. This result was essentially identical to that observed at 10 mM Mg 2ϩ and indicated that the first complex was the dominant complex for the SsbEc protein at the higher Mg 2ϩ concentrations.
The dependence of the binding of the SsbA protein to dT 75 on Mg 2ϩ concentration is shown in Fig. 8. In the absence of Mg 2ϩ , the dT 75 was largely converted into the second complex, although a band corresponding to the first complex was also visible. This was consistent with the results in Fig. 4 and indicated that two SsbA tetramers were able to bind to dT 75 under these conditions. At 1 mM Mg 2ϩ , the dT 75 appeared as a discrete band at the position of the first complex, with no indication of the formation of the second complex. This result was essentially identical to those that were obtained at the higher Mg 2ϩ concentrations and indicated that the first complex was the dominant complex for the SsbA protein in the presence of Mg 2ϩ .
The dependence of the binding of the SsbB protein to dT 75 on Mg 2ϩ concentration is shown in Fig. 8. As expected from the results shown in

Binding Modes of a Chimeric SsbA/B Protein
Proteolysis studies have shown that the SsbEc protein consists of an N-terminal domain that contains the subunit tetramerization and ssDNA binding sites and a C-terminal domain that is terminated by an acidic tail (2). A comparison of the primary sequences reveals that the N-terminal regions of the SsbA and SsbB proteins (amino acids 1-106) are highly similar to the corresponding region of the SsbEc protein (amino acids 1-116) (Fig. 1). However, the sequence comparison indicates that the C-terminal region of the SsbB protein is significantly shorter than that of either the SsbEc or the SsbA protein (Fig. 1). This suggested that the shorter C-terminal domain may be responsible for the differences between the ssDNA binding properties of the SsbB protein and those of the SsbA and SsbEc proteins. To explore this idea, a chimeric SSB protein was constructed in which the C-terminal domain of the SsbB protein (amino acids 105-131) was joined to the N-terminal domain of the SsbA protein (amino acids 1-105). The resulting chimeric SsbA/B protein (15,039 Da/monomer) was similar in size to the native SsbB protein (14,926 Da/monomer) and significantly smaller than the native SsbA protein (17,350 Da/monomer) (Fig. 9). The binding of the SsbA/B protein to dT 75 is shown in Fig. 10.
When increasing concentrations of SsbA/B protein were added to 32 P-end-labeled dT 75 in the presence of Mg 2ϩ , a complex with a gel mobility lower than that of unbound dT 75 was formed (Fig. 10). The mobility of this complex was greater than that of the complex formed with the SsbA protein but essentially identical to that of the complex formed with the SsbB protein (Fig. 10). This result was consistent with the similar sizes of the SsbA/B and SsbB proteins and indicated that the complex corresponded to the binding of a single SsbA/B tetramer to the dT 75 . There was no indication of the formation of a second complex under these conditions, even at the highest concentration of SsbA/B protein examined (Fig. 10). These results were similar to those that were obtained with the SsbA and SsbB proteins and indicated that a single tetramer of SsbA/B protein was able to bind to dT 75 in the presence of Mg 2ϩ (Fig. 10).
When increasing concentrations of SsbA/B protein were added to  dT 75 in the absence of Mg 2ϩ , an initial complex with a gel mobility lower than that of unbound dT 75 was formed at the lower protein concentrations (Fig. 10). The mobility of this initial complex was again greater than that of the complex formed under these conditions with the SsbA protein but essentially identical to that of the complex formed with the SsbB protein (Fig. 10). This indicated that the initial complex corresponded to the binding of a single SsbA/B tetramer to the dT 75 . A further increase in the concentration of SsbA/B protein resulted in a diminishment in this initial complex and the appearance of a second complex of even lower gel mobility (Fig. 10). This result indicated that two tetramers of SsbA/B protein were able to bind to dT 75 in the absence of Mg 2ϩ . Under the same conditions, a second complex was also observed with the SsbA protein, whereas there was no indication of the formation of a second complex with the SsbB protein (the increased gel mobility of the second SsbA/B complex relative to the second SsbA complex was presumably due to the smaller size of the SsbA/B protein) (Fig. 10). The band corresponding to the second SsbA/B complex did appear to be fainter than that for the second SsbA complex, suggesting that the second SsbA/B complex may be less stable than the second SsbA complex under these conditions. Nevertheless, these results indicated that the replacement of the C-terminal domain of the native SsbA protein with the shorter C-terminal domain from the SsbB protein does not fundamentally alter the binding of the protein to dT 75 .

DISCUSSION
The results presented here demonstrate that Mg 2ϩ -dependent changes in the DNA binding modes of the SsbEc protein can be monitored by Mg 2ϩ -polyacrylamide gel electrophoresis. The experimental strategy consisted of forward titration reactions in which increasing concentrations of SsbEc protein were added to a fixed concentration of dT 75 and reverse titration reactions in which increasing concentrations of dT 35 were added to a fixed concentration of SsbEc protein. In the absence of Mg 2ϩ , the forward titration results indicated that two SsbEc tetramers were able to bind to dT 75 , implying that the SsbEc tetramers were binding in the SSB 35 mode under these conditions with only two of the four subunits of each tetramer interacting with the dT 75 . This conclusion was supported by the reverse titration results, which indicated that only one dT 35 was able to bind to the SsbEc tetramer in the absence of Mg 2ϩ , suggesting that only two of the four subunits of the SsbEc tetramer were able to bind to ssDNA under these conditions. In the presence of Mg 2ϩ , in contrast, the forward titration results indicated that only one SsbEc tetramer was able to bind to dT 75 , suggesting that the SsbEc tetramer was binding in the SSB 65 mode under these conditions with all four subunits of the tetramer interacting with the dT 75 . This conclusion was also supported by the reverse titration results, which indicated that two molecules of dT 35 were able to bind to the SsbEc tetramer in the presence of Mg 2ϩ , suggesting that all four subunits of the tetramer were able to bind to ssDNA under these conditions. These results are consistent with previous fluorescence studies, which have shown that the SSB 35 mode of the SsbEc protein is generally favored in the absence of or at low concentrations of Mg 2ϩ , whereas the SSB 65 mode is favored at higher Mg 2ϩ concentrations (2).
The SsbA protein behaved similarly to the SsbEc protein in the forward and reverse titration experiments, both in the presence and in the absence of Mg 2ϩ . This indicates that the SsbA protein undergoes Mg 2ϩ -dependent transitions between an SSB 35 -like binding mode and an SSB 65 -like binding mode that are analogous to those of the SsbEc protein. In contrast, the results with the SsbB protein indicated that one tetramer of SsbB protein was able to bind to dT 75 in the forward titrations and that two dT 35 molecules were able to bind to the SsbB tetramer in the reverse titrations, both in the presence and in the absence of Mg 2ϩ . These results suggested that the SsbB protein differed from the SsbEc and SsbA proteins in that it appeared to bind to the dT n oligomers in an SSB 65 -like mode, even in the absence of Mg 2ϩ .
It has been proposed that the ability of the SsbEc protein to bind to ssDNA in the SSB 35 mode at low salt concentrations is due to a high degree of negative intrasubunit cooperativity among the DNA binding sites of the SsbEc tetramer (2,7). For example, it has been shown by fluorescence analysis that when the SsbEc protein is titrated with increasing concentrations of dT 35 at low salt concentrations, one molecule of dT 35 binds to two subunits of the tetramer with high affinity, but the binding of a second molecule of dT 35 to the remaining two subunits is greatly reduced (9). At higher salt concentrations, however, the SsbEc protein is able to readily bind two molecules of dT 35 , indicating that the negative intrasubunit cooperativity is reduced or eliminated under these conditions (9). Although the molecular basis for the negative intrasubunit cooperativity is not known, it has been suggested that it is due at least in part to electrostatic repulsions between the phosphoryl groups of the segments of ssDNA that are bound in the first and second subunits and the segments of ssDNA bound in the third and fourth subunits of the SsbEc tetramer and that the relief of negative intrasubunit cooperativity that is observed at higher salt concentrations may be due to a reduc- tion in this electrostatic repulsion which results from the binding of cations to the ssDNA (2,9).
The reverse titration gel shift results reported here, which indicated that the SsbEc protein was able to bind one molecule of dT 35 in the absence of Mg 2ϩ and two molecules of dT 35 in the presence of Mg 2ϩ , are consistent with the fluorescence results described above. Thus, the results that suggested that the SsbB protein was able to bind two molecules of dT 35 in either the presence or the absence of Mg 2ϩ indicate that the SsbB protein does not exhibit the pronounced negative intrasubunit cooperativity at low salt concentrations that has been found for the SsbEc protein. An absence of negative intrasubunit cooperativity and the resulting stabilization of an SSB 65 -like binding mode would also account for the results that suggested only a single SsbB tetramer was able to bind to dT 75 , in either the absence or the presence of Mg 2ϩ . In either case, the binding of an SsbB tetramer to dT 75 in an SSB 65 -like mode would prevent a second SsbB tetramer from binding to that dT 75 molecule.
Our new results provide an explanation for our previously observed differences in the apparent stoichiometries for the binding of the SsbB protein and the SsbEc and SsbA proteins to various dT n oligomers (7). Since our original analysis was carried out using the standard gel electrophoresis buffer, TBE, which contains no Mg 2ϩ , it is likely that the SsbA and SsbEc proteins were able to bind to the dT n oligomers in an SSB 35 mode, whereas the SsbB protein was restricted to binding to the oligomers in an SSB 65 -like mode. The new results indicate that the binding properties of the SsbB protein and those of the SsbEc and SsbA proteins are more similar in the presence of Mg 2ϩ in that all three proteins appear to bind to dT n oligomers in an SSB 65 -like mode under these conditions. However, the reverse titration experiments indicate that even in the presence of Mg 2ϩ , there are differences between the SsbB protein and the SsbA and SsbEc proteins in terms of the intrasubunit cooperativity for ssDNA binding. Although the presence of Mg 2ϩ appears to reduce the negative intrasubunit cooperativity for ssDNA binding by the SsbEc and SsbA proteins, the SsbB protein may actually bind ssDNA with positive intrasubunit cooperativity in the presence of Mg 2ϩ .
The most obvious structural difference between the SsbB protein and the SsbA and SsbEc proteins is in the length and composition of the C-terminal domains of these proteins. It is conceivable that the differences in the C-terminal domains could affect the degree of intrasubunit cooperativity and the relative stabilities of the different binding modes of the various SSB proteins. However, our results indicated that the chimeric SsbA/B protein (in which the entire C-terminal domain of the SsbA protein was replaced with the shorter C-terminal domain of the SsbB protein) undergoes Mg 2ϩ -dependent transitions between an SSB 35 -like binding mode and an SSB 65 -like binding mode that are similar to those of the native SsbA protein. This suggests that the apparent inability of the SsbB protein to readily bind in an SSB 35 -like mode in the absence of Mg 2ϩ is not due simply to the shorter length of the C-terminal domain of the SsbB protein. It remains possible that the C-terminal SsbB domain may interact differently with the N-terminal domain of the SsbB protein than it does with the N-terminal domain of the chimeric SsbA/B protein and that this interaction modifies the ssDNA binding properties of the SsbB protein. However, the similarity of the DNA binding properties of the SsbA and SsbA/B proteins suggests that the primary structural determinants that govern the stability of the DNA binding modes may be contained within the N-terminal domains of the various SSB proteins.
It has been suggested that the different DNA binding modes of the SsbEc protein may be used selectively for different functions in the cell (2). The finding that the SsbA protein exhibits Mg 2ϩ -dependent changes in its binding modes that are similar to those of the SsbEc protein provides further support for the proposal that the SsbA protein may be the S. pneumoniae analog of the SsbEc protein, a general purpose SSB provides involved in routine DNA functions. Although the SsbB protein is strongly induced during natural transformation in S. pneumoniae, the specific role of the SsbB protein in this process is not known. However, it is conceivable that the apparent preferential formation of the SSB 65 -like binding mode by the SsbB protein may reflect an adaptation that enhances the ability of the SsbB protein to function during natural transformation in S. pneumoniae.  , and either no magnesium acetate (Ϫ Mg 2ϩ ) or 10 mM magnesium acetate (ϩ Mg 2ϩ ). The reactions were analyzed by polyacrylamide gel electrophoresis using a gel running buffer consisting of Tris acetate (pH 7.5) and the same concentration of magnesium acetate as in the reaction solutions. The bands corresponding to unbound dT 75 (ssDNA) and the various (SSB) n -dT 75 complexes were visualized by autoradiography.