Thermodynamic analysis of DNA binding by a Bacillus single stranded DNA binding protein

Background Single-stranded DNA binding proteins (SSB) are essential for DNA replication, repair, and recombination in all organisms. SSB works in concert with a variety of DNA metabolizing enzymes such as DNA polymerase. Results We have cloned and purified SSB from Bacillus anthracis (SSBBA). In the absence of DNA, at concentrations ≤100 μg/ml, SSBBA did not form a stable tetramer and appeared to resemble bacteriophage T4 gene 32 protein. Fluorescence anisotropy studies demonstrated that SSBBA bound ssDNA with high affinity comparable to other prokaryotic SSBs. Thermodynamic analysis indicated both hydrophobic and ionic contributions to ssDNA binding. FRET analysis of oligo(dT)70 binding suggested that SSBBA forms a tetrameric assembly upon ssDNA binding. This report provides evidence of a bacterial SSB that utilizes a novel mechanism for DNA binding through the formation of a transient tetrameric structure. Conclusions Unlike other prokaryotic SSB proteins, SSBBA from Bacillus anthracis appeared to be monomeric at concentrations ≤100 μg/ml as determined by SE-HPLC. SSBBA retained its ability to bind ssDNA with very high affinity, comparable to SSB proteins which are tetrameric. In the presence of a long ssDNA template, SSBBA appears to form a transient tetrameric structure. Its unique structure appears to be due to the cumulative effect of multiple key amino acid changes in its sequence during evolution, leading to perturbation of stable dimer and tetramer formation. The structural features of SSBBA could promote facile assembly and disassembly of the protein-DNA complex required in processes such as DNA replication.


Background
Nearly all cellular nucleic acid transactions, including DNA replication, repair and recombination require the activity of a single stranded DNA binding protein (SSB) [1][2][3][4][5][6][7]. SSB proteins and are found throughout nature and their functional importance is underscored by their presence in prokaryotes, archaea, and eukaryotes including mammals [1]. Among its multifaceted roles, upon binding to ssDNA, SSB prevents the reformation of duplex DNA making it possible for other enzymes such as DNA polymerase to use ssDNA as substrate. In addition, the binding of SSB-type proteins protects the ssDNA molecules from attack by intracellular nucleases. Although not possessing intrinsic enzymatic activity in and of themselves, SSB proteins are known to influence the activities of many enzymes as well as to organize the multi-protein complexes required for processes such as DNA replication, recombination and DNA repair [8][9][10][11].
The function of SSB during DNA replication has been extensively studied in E. coli, which serves as the prototypical model system for prokaryotes and eukaryotes alike. In E. coli, the large nucleoprotein replication initiation complex is stabilized by single stranded DNA binding protein, following which DNA is unwound by the DnaB helicase protein. Efficient DNA unwinding activity of DnaB protein in progression of the replication fork in E. coli is strongly dependent on the continued action of a cognate SSB [12,13]. SSB works in concert with DnaB helicase, DNA primase, and DNA polymerase III holoenzyme during E. coli DNA replication [5,9,12,14,15]. Phage λ DNA replication requires the participation of host E. coli SSB as well [16,17]. In archaea and eukaryotes, its functional homolog, Replication Protein A (RPA), carries out the role of organizing and stabilizing the replisome during DNA replication [1,3,10,[18][19][20][21].
Vital to its function in DNA metabolism is the structure of SSB. In the Gram-negative bacteria, SSB is homotetrameric, with each monomer contributing a single ssDNA-binding domain to the functional form. The eukaryotic RPA is composed of three subunits (RPA70, RPA32, and RPA14) and functions as a heterotrimer through the use of four ssDNA-binding domains [2,3,18].
Unlike E. coli SSB, single stranded DNA binding protein from bacteriophage T4, the gene 32 protein, is a monomer. T4 gene 32 protein can form multimers at high concentration induced by high salt and high pH [22]. Kim and Richardson demonstrated that the bacteriophage T7 SSB, the gene 2.5 protein, is a dimer [23]. The T7 gene 2.5 SSB appears to bind DNA as a dimer. The ssDNA binding affinities of both T4 and T7 SSBs are lower than that observed with E. coli SSB. Despite these differences, ssDNA binding of SSB proteins using OB fold-domains (oligosaccharide/oligonucleotide binding domains) appears to be universal throughout all systems described to date [1].
The E. coli SSB is highly cooperative in ssDNA binding that is influenced by salt concentration [24,25]. Recent studies indicate that SSB has at least two distinct modes of ssDNA binding [26]. The binding is modulated by monovalent salts. At very low salt concentration (<10 mM NaCl), SSB binds ssDNA using two of its four subunits in a highly cooperative manner and occludes only 35 nucleotides [(SSB) 35 mode]. On the other hand, at higher salt concentrations (>200 mM NaCl), it binds to ssDNA using all four subunits and protects~65 nucleotides [(SSB) 65 mode]. It is not clear how the ssDNA binding is altered between 10 and 200 mM NaCl. Higher-order forms of SSB in ssDNA bound states, based on high resolution electron microscopic studies of SSB-ssDNA complex, have also been reported [27]. Chrysogelos and Griffith discovered that repeated freezing-thawing of E. coli SSB leads to the formation of unique strings of tetramers [28].
Gram-positive bacterial protein sequences do not form a monophyletic group, but are intermixed with plasmid and phage sequences [29,30]. Gene organization in these organisms can differ from that observed in Gramnegative E. coli and these organisms may contain multiple paralogues [31,32]. Sequence analysis indicated that Gram-positive SSBs have a highly conserved nearlyidentical (>90% identity) N-terminal ssDNA binding as well as monomer-monomer interaction domains but they differ to some extent from the Gram-negative SSBs.
We have investigated the structure and ssDNA binding of a Gram-positive bacillus SSB (SSB BA ) in order to understand its mechanism of action of SSBs in these organisms. We present here a report of a Gram-positive SSB that utilizes a novel structural mechanism for protein-DNA interaction using a transient tetramer formation.

Results
The single-stranded DNA binding protein ORF of B. anthracis (BAS5326) was identified by BLAST search of the annotated sequenced genome of B. anthracis Stern strain [33,34]. The ORF encodes a polypeptide of 172 amino acid residues with a predicted molecular weight of 19.2 kDa.

Sequence analysis of SSB BA
The amino acid sequence of the N-terminal ssDNA binding and protein-protein interaction domains responsible for dimer and tetramer formation of SSB BA were compared with the sequences of a Gram-positive (Bacillus anthracis) and Gram-negative (E. coli and Salmonella typhimurium) SSB proteins. Multiple-alignment of these sequences, using ClustalW2, is shown in Figure 1. In general, a high degree of sequence homology among SSBs was observed only in the N-terminal two-thirds of the proteins. In the N-terminus two-third, sequence homology (identity + similarity) of SSB BA was observed with Gram-positive SSBs and was estimated to be~54%, whereas, it was ≥90% among Gram-negative SSBs. Amino acid residues 1-46 constitute the major portion of the SSB core domain and contain the OB DNA binding fold. The second region of homology observed was between residues 51-104 ( Figure 1). A third area of homology was also observed in the last six residues with the sequence, DDDLPF, which corresponds to the acidic carboxy-terminal domain characteristic of eubacteria; this sequence is required for interaction with recombination, replication and repair machineries [35,36]. The Cterminal region lacked significant homology even among Gram-negative SSBs in the region between the residues 100-166 among these SSBs [37].

Purification of SSB BA
Recombinant SSB BA was highly soluble when expressed in E. coli. It was purified using a combination of ammonium sulfate fractionation as well as conventional ion exchange chromatography. These steps resulted in homogenous SSB BA (Figure 2A). In SE-HPLC analysis of purified SSB BA (100 μg/ml) the protein eluted as a single peak with an elution volume consistent with that of a monomer ( Figure 3).
To test the biological activity of SSB BA we measured its ability to stimulate its cognate DnaB, DnaB BA , using a FRET based DNA helicase assay. It was based on the ability of SSB BA to stimulate DNA unwinding activity of its cognate DNA helicase, DnaB BA ( Figure 2). In the absence of SSB BA , DNA unwinding by DnaB BA was limited ( Figure 2B) and was greatly stimulated in the presence of SSB BA ( Figure 2C). The stimulation of the DNA helicase activity of DnaB BA in the presence of the purified SSB BA was >10 fold which was as expected for the cognate SSB [13].

Mechanism of ssDNA binding by SSB BA
Fluorescence anisotropy-based titration of a DNAfluorophore by the DNA binding protein is a direct methodology for determining affinity of protein-DNA complexes. Following this approach, 5′-Fluorescein-labeled oligo(dT) 20 (Fl(dT) 20 ) was used as a fluorescence anisotropy probe for analyzing SSB BA and ssDNA interactions. In order to determine the binding constant for SSB BA , the interaction of SSB BA with 5′-flourescein labeled oligo (dT) 20 was examined at 25°C and in buffer B containing 0.1 nM Fl-(dT) 20 , 25 mM KCl and 5 mM Mg +2 . Fl-(dT) 20 was titrated with SSB BA until saturation in anisotropy was observed. The anisotropy values at various SSB BA concentrations were used to create a binding isotherm as a semilog plot as shown in Figure 4. At very low SSB BA concentrations, very small anisotropy changes and a flat plateau (~43 mA) were observed that were attributed to Fl-(dT) 20 . Upon further addition of SSB BA , the anisotropy value increased with an increase in SSB BA •Fl-(dT) 20 complex formation. A sigmoid binding isotherm with saturation binding at high SSB BA concentration was observed with maximum anisotropy of 184 ± 5 mA. At higher SSB concentration (≥ 1 μM), anisotropy did not change significantly (data not shown). Nonlinear regression analysis, using a sigmoidal dose-response equation, of the data allowed for determination of the SSB BA concentration at which 50% of the ligand was in bound form (EC 50 ); that value corresponds to the apparent dissociation constant. The K D for SSB BA •Fl-(dT) 20 complex was 1.0 ± 0.1 x 10 −9 M. The Hill coefficient was 1.6 ± 0.6 indicating a possible binding of one to two molecules of SSB BA to Fl-(dT) 20 , which is , Pseudomonas aeruginosa (SSB-PA) were aligned using ClustalW2 program. The color coding is as follows: red, basic; blue hydrophobic; green, hydrophilic; orange, neutral; pink, acidic; and light green, proline. Notations of the secondary structures in SSB BA are as defined by Murzin [38] with the β strands of the OB-fold labeled 1-5 (Light green box) and the α-helix (dark green box). Locations of the important hydrophobic tryptophans (in E. coli) are indicated by red up arrows. The residues in the monomer-monomer interface of the dimer in E. coli SSB sequence are indicated in yellow boxes. Changes of Gln77 !Leu and Gln111!Phe should destabilize dimer-dimer interaction important for the tetramer formation [39].
not surprising considering the fact that SSB BA does not form a stable tetramer. A simple hyperbolic fit of the anisotropy data was also carried out (Figure 4 inset). The estimated K D value was 1.2 ± 0.8 x 10 −9 M. In order to determine the correct method of analysis of binding data, we have examined the "Goodness of Fit" using both models. The R 2 value for the sigmoidal dose-response model was found to be 0.98, whereas, for the simple hyperbolic model was 0.94. The Ftest using the values of degrees-of-freedom and absolute sum of squares, gave an F-value of 22.4 which correlated well with the R 2 values and clearly demonstrated that the sigmoidal dose-response model as the correct model/equation for fitting anisotropy binding data for such analysis.

Thermodynamics of ssDNA binding
In order to understand the thermodynamics of SSB BA •ssDNA binding interactions, we have analyzed SSB BA binding to ssDNA at different temperatures over a range of 20-37°C. The temperature-dependent binding isotherms for SSB BA and Fl-(dT) 20 are presented in Figure 5A. Higher anisotropy values observed at 20 or 25°C were due to glycerol used in this assay. As the temperature increased, an overall decrease in anisotropy value was observed for both the free and bound oligonucleotides. This overall decrease in anisotropy value is attributed to the decrease in viscosity of the solution with increasing temperature which was somewhat attenuated by adding glycerol. With a decrease in (b) FRET helicase analysis of DnaB BA in the presence and absence of SSB BA : Emission spectra of the substrate (4.2 nM), after 15 min incubation with SSB BA (3 μg/ml) DnaB BA (0.5 μg/ml), and both SSB BA and DnaB BA at 37°C; (c) Kinetic analysis of helicase activity: The helicase substrate was rapidly mixed with indicated protein(s) and fluorescence emission at 662 ± 8 nm was recorded as a function of time for 500 s using a Slow Kinetic mode in PC1 spectrofluorometer using Vinci software (ISS Inc. Champaign, IL). Using FRET spectra of native and heat denatured substrate, we predetermined that 1% decrease in FRET is equivalent to~3.1 pmol of nucleotide (or bp) unwinding of duplex DNA.

Figure 3
Size exclusion HPLC analysis of SSB BA .Size exclusion HPLC analysis of SSB BA . The native molecular mass of SSB BA Fraction V was investigated using a TSK GS3000SW gel filtration column using A-100 as running buffer. Twenty micrograms of SSB BA was injected and the column was eluted at 0.4 ml/min. and 0.4 ml fractions were collected. The gel filtration standards were BSA (68 kDa) and ovalbumin (44 kDa) and lysozyme (15 kDa). temperature, the viscosity as well as the anisotropy values appeared to increase [40]. Analysis of the binding curve data for K D showed the dissociation constants increased steadily from 20°c to 37°C. We did not observe any significant change in the Hill coefficient with temperature. This increase in K D could likely be due to the dissociation of the protein•DNA complex at higher temperatures. The DnaA BA •DNA complex was most stable at 20-25°C.
The dissociation constants obtained at varying temperatures were used to evaluate the thermodynamic properties of DNA binding. We have plotted the dissociation constants using the Van't Hoff equation, lnK D = −ΔH°/RT, where ΔH°is the enthalpy change and T and R are the temperature and gas constant respectively, with the dissociation constants derived from 20, 25, 30, and 37°C ( Figure 5B). The plot is linear for temperatures 20°C to 37°C and it diverges from linearity below 20°C. The slope of the Van't Hoff plot was used to derive the change in enthalpy (ΔH°) at 25°C (33.9 kJ mol −1 ). The change in entropy (ΔS°) was calculated to be 56.9 J mol −1 K −1 . Thus, it appears that the formation of the SSB BA •ssDNA complex has a strong entropic or hydrophobic component to the overall protein-DNA interaction.
In addition to hydrophobic interaction between SSB BA and ssDNA, we analyzed the contribution of ionic interactions, if any, in the complex formation. Fluorescence anisotropy was used determine the K D of formation of SSB BA •ssDNA complex as a function of NaCl concentration. The binding isotherm at different salt concentrations was generated using SSB BA titration of ssDNA at 25°C ( Figure 6A). Each titration curve fits according to a single site binding isotherm. The data shows that highest affinity binding occurs at 0-25 mM NaCl. The dissociation constant increased steadily from 0-250 mM NaCl. The most striking change was in the values of the Hill coefficient. At 0 mM NaCl, it was 1.9 ± 0.6 and steadily decreased to 0.9 ± 0.1 at 150 mM NaCl and beyond. It could indicate that at 0 mM NaCl, SSB BA was forming a predominantly dimeric structure which transformed into monomeric at higher NaCl concentration. However, at very low ionic strength, non-specific protein-protein interactions could not be ruled out. A thermodynamic linkage plot for ssDNA binding as a function of NaCl ( Figure 6B) was generated from the K D values obtained from Figure 6A. The presence of a negative slope was indicative of a net ion release [41]. The data were analyzed using the following equation to determine the number of ions released upon binding: The analysis suggests that upon SSB BA •ssDNA complex formation, only one ion was released from the protein-DNA interface. These results appeared to indicate a small but significant contribution of ionic interaction in the ssDNA binding.

Structural analysis of SSB BA by homology modeling
The SSB BA sequence was further analyzed for secondary structure using Rosetta software (http://robetta.org/fragmentsubmit.jsp). Rosetta analysis indicated that there are at least five significant β strand structures and a single α-helix in the N-terminal half of the molecule (data not shown). The structure between the residues 101-170 appeared to be a random coil. These secondary structures are consistent with known features of SSB EC monomer, as determined earlier by X-ray crystallography [37,39,42].
Alignment of SSB BA sequence with sequences of other prokaryotic SSBs did not provide any clue to the basis of  its ssDNA binding properties ( Figure 1). Therefore, the three-dimensional structure of SSB EC derived from Xray crystallography [37] was explored. The significant sequence homology of SSB BA with E. coli SSB allowed us to develop a putative three dimensional model of SSB BA using homology based modeling. The initial modeling was done using the SWISS-MODEL server [43,44]. Further refinement of the model by energy minimization was carried out using SYBYL 8.1 (Tripos Inc., St. Louis, MO) molecular modeling software using the SSB crystal structure, PDB ID: 1QVC, as the structure template [37]. The structures for E. coli SSB EC , 1QVC and SSB BA were visualized using PyMOL (The PyMOL Molecular Graphics System, Version 1.3, Schrödinger LLC). Structures of both SSBs are presented in PyMol in Figure 7.
SSB proteins are known to bind ssDNA through their oligonucleotide/oligosaccharide binding fold (OB fold) as described by Murzin [38]. The OB fold is characterized by a β-barrel consisting of five β-strands capped by an α-helix. Despite sequence differences between SSB BA and SSB EC , the OB fold observed in SSB EC remained intact in SSB BA including the β-turn regions, particularly L45 between β sheets 4 and 5 ( Figure 7). It has been shown that the β sheet 1 of SSB EC with the sequence VNKVILV is in the monomer-interface of the SSB EC dimer [39]. In SSB BA , this β sheet remains partially intact (NKVILV) with the loss of the Val5 residue.
However, the His56 of one monomer in SSB EC forms a hydrogen bond with Asn6 and the carbonyl oxygen of Leu83 of another monomer, which is essential for a stable dimer/tetramer formation. Although Asn6 (Asn2 in SSB BA ) and Leu83 (Leu76 in SSB BA ) remained conserved, one of the most important residues, His56 in SSB EC was altered to Ile (Ile47) in SSB BA (Figure 1). It should be noted that in the temperature-sensitive E. coli mutant, ssb-1, His56 was mutated to Tyr56 leading to the ts-phenotype. E. coli ssb-1 ts-mutant does not form a stable tetramer at non-permissive temperature [5,45]. Thus, the lack of this His residue in SSB BA will likely hinder a stable dimer formation.
The SSB EC tetramer is formed by the interaction of two dimers [37,39]. The dimer-dimer interface involves two six-stranded surfaces, each comprised of β1, β4, and β5 from two monomers. The structure of SSB BA , as shown in Figure 7, could form such a tetramer interface, had it not been for the difficulty associated with the dimer formation. It has been shown with a number of SSB crystal structures that a network of hydrogen bonds among the side chains in this six-stranded interface is necessary for a stable tetramer formation. The residues that were shown to be important in SSB EC for this network of hydrogen bond formation are Lys8, Tyr79, Gln77, Glu81, and Gln111. Sequence comparison ( Figure 1) between SSB EC and SSB BA indicated that all of . Both of these structures were generated using PyMol. In SSB EC , Trp55, Trp89, and Phe61 are important in the ssDNA binding, which are replaced by Phe47, Tyr82, and Trp53 respectively in SSB BA . Most notably, His56 residue of SSB EC , required for monomer-monomer interaction is replaced by Ile48 in SSB BA [39]. these residues in SSB BA underwent changes and are as follows: Lys8!Arg2, Tyr79!Gly71, Gln77!Leu69, Gln83!Arg75, and Gln111!Phe104. Although all of the changes may not be significant, three of these five changes are significant in terms of hydrogen bond formation. Therefore, these amino acid changes in SSB BA are likely to impede tetramer formation further. Taken together, inhibition of both monomer-monomer interaction leading to dimer formation as well as dimerdimer interaction leading to tetramer formation, the amino acid sequence of SSB BA does not support formation of stable dimer or tetramer.
Analysis of the structure of ssDNA binding pocket in SSB BA Single-stranded DNA binding by prokaryotic SSBs has been shown by several groups to be carried out exclusively by tetrameric forms of SSBs containing four OB folds or dimeric forms with each monomer containing two OB folds [37,39,46]. Thus, the presence of four OB folds in SSBs appears essential for high affinity ssDNA binding. Our studies indicated that SSB BA bound ssDNA with very high affinity (1.0 ± 0.1 x 10 −9 M) even though it did not appear to form a stable tetramer in the absence of DNA at the concentration range examined (Figures 3 & 7).
The amino acid residues in SSB BA that are homologous to the residues in other SSBs, particularly SSB EC , that are known from crystallographic studies to bind to ssDNA were analyzed. Several hydrophobic residues, Trp89, Trp55, and Phe61, in SSB EC have been identified as involved in ssDNA binding through base stacking interactions [39]. In Helicobacter pylori SSB (SSB HP ), Phe37, Phe50, Phe56, and Trp84 are involved in base stacking interactions with ssDNA [47]. These hydrophobic residues (Phe36, Phe43, Trp53, and Tyr81) with changes remained conserved in SSB BA (Figure 1). As shown in Figure 8, these aromatic side chains are exposed in the DNA binding groove of the OB fold so that ssDNA bases could have stacking interactions. In addition, ssDNA binding to SSB requires a large number of positively charged residues for the formation of ionic bridges with the phosphodiester backbone of ssDNA. A large number of Arg and Lys residues were observed in and around the groove as shown in Figure 8. Among these, Arg9, Lys12, Arg17, Arg42, Arg54, Lys55, Lys65, Lys66, Lys87, Arg88, Arg103 appeared to be in close proximity of the DNA binding groove and form ionic bridges with the phosphodiester backbone. Consequently, the high affinity ssDNA binding observed with SSB BA could be due to basic as well as aromatic residues in its DNA binding groove.

Subunit structure of SSB BA in the SSB BA •ssDNA complex
Our studies demonstrated that SSB BA was capable of highaffinity binding of ssDNA. Was it possible that the protein could form higher order structures, possibly tetramers, upon binding long ssDNA templates? Tetrameric E. coli SSB binds ssDNA in two distinguishable forms; SSB 35 and SSB 65 [26,48]. The SSB 35 Figure 7A using the program PyMOL. Hydrophobic, positive, and neutral potential surface are displayed and colored green, red, and white, respectively. template. In the SSB 35 form, only two subunits of the tetramer make contacts with the DNA, whereas, in SSB 65 form all four subunits of the tetramer make contacts with the DNA. These two forms can be distinguished by fluorescence resonance energy transfer (FRET) using a long ssDNA labeled in each end with Cy3 and Cy5 fluorophores [26]. The SSB 65 form produces high FRET and the SSB 35 form produces attenuated FRET. This approach was utilized using a seventy nucleotide, (dT) 70 , oligonucleotide labeled with a 3′ Cy3 fluorophore and a 5′ Cy5 fluorophore (Cy5-(dT) 70 -Cy3) to test our hypothesis that SSB BA forms higher order structures upon DNA binding. With this ssDNA substrate, we anticipated that SSB BA in SSB 35mode would produce attenuated FRET and SSB 65 -mode would produce high FRET. Binding of SSB EC to Cy5-(dT) 70 -Cy3 oligonucleotide was first analyzed to test the validity of our assay. The results are presented in Figure 9. FRET was measured using 515 nm excitation wavelength and 665 nm emission wavelength. The emission intensity was corrected for Cy5 contribution to 665 nm emission. FRET was defined by (F-F 0 )/F 0 where F 0 and F are the corrected 665 nm emission intensities of the 200 nM 5′-Cy5 (dT) 70 Cy3-3′ oligonucleotide in the absence and presence of SSB respectively. In the titration with SSB EC , sharp increase in FRET with initial titration with SSB EC reaching a peak at 0.44 μM concentration was seen. Upon further titration with SSB EC , the FRET decreased substantially and reached a plateau at concentrations higher than 1.1 μM. These results with SSB EC are comparable to that observed by Roy et al. [26]. However, our experiments required a somewhat higher concentration of SSB EC than reported by Roy et al. [26], which could be due to different buffer systems and ssDNA concentration. Thus, at 0.44 μM SSB EC, the high FRET (SSB EC ) 4 -(dT) 70 complex was observed and at or above 1.1 μM SSB EC , intermediate FRET (SSB EC ) 8 -(dT) 70 complex was seen.
Next, SSB BA binding to Cy5(dT) 70 Cy3 oligonucleotide was analyzed. The Cy5(dT) 70 Cy3 oligonucleotide was titrated with SSB BA as described above for SSB EC . Similar to SSB EC , FRET increased linearly with increasing concentration of SSB BA , and reached a plateau at 0.8 μM Surprisingly, only a high FRET form of (SSB BA )-(dT) 70 complex was observed. However, initial slope of the plot is very different from that observed with SSB EC which is a stable tetramer. Therefore, it appeared that the high FRET form of the complex was formed but required higher concentration of SSB BA . The high FRET form of (SSB EC )-(dT) 70 complex is a tetrameric complex in which all four monomers bind ssDNA. The high FRET form of (SSB BA )-(dT) 70 complex should have the same oligomeric structure as the high FRET form of (SSB EC )-(dT) 70 complex. As SSB BA lacked stable tetramer formation, it required higher (~two fold) SSB concentration to form the high FRET complex. The reason is that this complex is not formed by a single binding event as is the case with SSB EC complex. It involves four binding and one structural rearrangement steps. As individual monomers are binding, there are four separate binding constants (K1, K2, K3, and K4) involved in the (SSB BA ) 4 -(dT) 70 complex formation. It will be erroneous to assume that they are all 1 nM. Certainly K1 is 1 nM as observed in Figure 4. Other binding constants (K2, K3, and K4) are likely to be higher than 1 nM due to steric hindrance which is particularly important for K4 involving the binding of the fourth monomer. The (dT) 70 is small and can accommodate only four SSB BA resulting in progressive lack of sufficient open DNA for second, third and fourth SSB BA monomers to bind. Therefore, it is very likely that K2 and K3 are higher and K4 is substantially higher than 1 nM which explains a higher SSB BA concentration to achieve a high FRET complex. Based on this reasoning of the (SSB BA ) 4 -(dT) 70 complex formation, as described above, our data actually supports the model in Figure 10. Lack of observation of an intermediate FRET form of the (SSB BA )-(dT) 70 complex in this study indicated that a (SSB BA ) 8 -(dT) 70 complex probably did not form in appreciable amount even with high proportional levels of SSB BA . Based on these reasonings of the (SSB BA ) 4 -(dT) 70 complex formation, as described above, our data led to the proposal of the model presented in Figure 10 for the (SSB BA ) 8 -(dT) 70 complex.

Discussion
SSB protein is required for a variety of processes such as DNA replication, recombination and DNA repair despite its lack of any enzymatic activity [5,9,49]. Among its multifaceted cellular activities, a common feature of all of these processes is to bind ssDNA with high affinity and protect it from reannealing and/or degradation. Of emerging importance is SSB's role in protein-protein interaction during various DNA transactions. Most studies involving SSB proteins demonstrated that ssDNA wraps around a tetrameric form of SSB. SSB BA does not form a stable tetramer E. coli SSB EC is a stable tetramer with high solubility and tremendous thermal stability [1]. The majority of prokaryotic cellular SSBs are homotetramers, where each monomer harbors an OB fold. However there are exceptions. SSB DR from Deinococcus radiodurans is a homodimer, where each monomer is quite large and contains two OB folds [37][38][39]46]. Each OB fold is capable of binding ssDNA independently. In both cases, a stable SSB protein complex with four OB folds is required for ssDNA binding. SSB BA was found to be not tetrameric at or above ambient temperature by size exclusion HPLC (Figure 3). This physicochemical property of SSB BA is closely comparable to the T4 bacteriophage SSB, which is monomeric.

Molecular basis of SSB BA structure
Sequence alignment and three dimensional structure of SSB BA , generated by homology-based modeling were utilized to probe the molecular basis of ssDNA binding (Figures 1). Sequence alignment and secondary structure prediction (data not shown) clearly indicated the presence of an OB fold in SSB BA , which is a characteristic of SSBs and required for high affinity ssDNA binding (Figure 4).
Both sequence alignment and homology modeling (Figures 1 & 8) indicated lack of several residues that are important for monomer-monomer and dimer-dimer interactions leading to the formation of a stable tetramer in SSB EC . His56 (as well as Glu54) of β sheet 3 (Figure 1) in SSB EC plays an important role by forming a hydrogen bond with Asn7 in β sheet 1, the carbonyl group in Leu84 and Thr100 at the base of loop L45. Notably, in E. coli temperature-sensitive mutant, ssb-1, His56 is mutated to Tyr56 [5]. This mutant does not form a stable tetramer with respect to monomers at nonpermissive temperatures [5,45,50]. Thus, a lack of the corresponding His residue (His!Ile change) in SSB BA may be one of the important contributors to its structure. Moreover, in B. anthracis the sequence Glu54. Trp55.His56 in the E. coli β sheet 3 is altered to Asp46. Phe47.Ile48 (Figure 1). This change did not alter the β sheet structure but may have altered the contributions of these residues in the monomer-monomer interaction in the stable dimer and tetramer formation. It appears that although β sheet 1 remained conserved in SSB BA, this region lacks the valine residue which may have attenuated the interaction of the β sheet 1 with β sheet 1′ in the monomer-monomer interface of the dimer. The shortening of the N-terminus in SSB BA may also have deleterious effect in the interactions involving H-bonds in this region and can contribute to the lack of tetramer formation. Taken together, the amino acid residue substitutions in SSB BA , as described above, are likely contributed to the disruption of monomer-monomer interaction leading to dimer formation.
The dimer-dimer interface in the SSB EC tetramer is primarily a six-stranded β sheet-mediated. Residues that are important in SSB EC for the network of hydrogen bond formation at the dimer-dimer interface are Lys8, Tyr71, Gln77, Glu81, and Gln111 [39]. Sequence comparison ( Figure 1) between SSB EC and SSB BA indicated that all of these residues in SSB BA underwent alteration and are as follows: Lys8!Arg3, Tyr79!Gly71, Gln77!Leu69, Gln83!Arg75, and Gln111!Phe104. Some of these changes are chemically significant leading to possible disruption of the network of hydrogen bond formation that is required for a stable tetramer formation. In addition, Gln77 and Gln111 are located in the dimer-dimer interface and have been implicated in the tetramer formation. As described earlier, Gln111 is altered to Phe104 in SSB BA . An equally significant change is observed with Gln77 which is changed to Leu69 in SSB BA . Taken together, these changes in amino acid sequence may disrupt both monomer-monomer and dimer-dimer interactions leading to a monomeric SSB BA at a physiological temperature.

Energetics of SSB BA •ssDNA binding
Protein-DNA recognition and binding involve complex interactions. Earlier, we have used fluorescence anisotropy analysis of DNA binding by E. coli DNA primase and determined the thermodynamic parameters of protein-DNA interaction (38). We have used a similar analysis to probe the ssDNA binding by SSB BA . In addition, we have analyzed contribution of electrostatic and ionic interactions in the binding by analyzing the dependence of binding on the ionic strength of the environment. Together, these two analyses provided a detailed picture of the forces in SSB BA •ssDNA binding.
The K D values were determined at different temperatures (20-37°C) ( Figure 5). Our data showed that although SSB BA was able to bind DNA at a wide range of temperatures, it bound with the highest affinity at 20-25°C. The free energy change for SSB•ssDNA association was −23 kJ mol −1 at 25°C. Using the two equations: ΔG°= −RT lnK D and ΔG°= ΔH°−TΔS°and the slope of the plot, we determined that ΔS°was~188 J mol −1 K −1 in this temperature range.
We determined the K D value of SSB BA binding to ssDNA at different salt concentrations (0-200 mM NaCl) ( Figure 6A). The binding is progressively weakened with an increase in ionic strength. The K D values were then analyzed using a linkage plot to determined ionic interactions in the binding. The negative slope of the plot in Figure 5B indicated a release of ions from. Our analysis determined the release of one Na + and one Clion during the binding process. In addition, our results also pointed out that SSB BA likely formed a tetrameric species at 0 mM NaCl and became monomeric at higher NaCl concentration. It is perhaps possible that the tetramer formation could be dependent on the ionic strength.
Thus our results suggest that ionic interaction or salt bridge formation between the protein and the DNA made specific contribution to the overall free energy change. In order to determine the contribution we first extrapolated K D value of the complex at infinite salt concentration (K D ∞ ) by nonlinear regression of K D versus log[NaCl] plot (data not shown). The value of ΔG°i onic was −8 kJ mol −1 .

Mechanisms of ssDNA binding by SSB BA
Despite differences between its Gram-negative counterpart, SSB BA bound to ssDNA with high affinity (Figure 4). The ssDNA binding affinity (K D ) for a SSB BA monomer binding to a small oligonucleotide was 1.0 ± 0.1 x 10 −9 M at 25°C. Even though many changes in amino acid sequence of SSB BA directly relate to ssDNA binding, such as Trp55!Phe47, Trp90!Tyr81, Phe61!Trp53, the changes are not drastic enough to alter ssDNA binding ( Figure 8). Three dimensional structure as well as electrostatic surface potential in Figure 8 indicates that ssDNA binding remained unperturbed. A temperature-sensitive mutant of E. coli SSB, ssb-1, is unable to form a stable tetramer at a non-permissive temperature [5,45,50]. This mutant is also defective in supporting DNA replication at non-permissive temperature. Thus, it appears a SSB tetramer formation is a prerequisite for DNA replication. Consequently, we sought to explore whether ssDNA template could influence the ability of SSB BA to form tetramers upon DNA binding. A likely possibility is that SSB BA is capable of forming a normal tetrameric structure containing four OB-folds, as seen in other SSBs, upon sufficiently long ssDNA. This possibility was examined using a recently developed FRET assay for SSB•ssDNA interaction [26].
Previous studies with SSB EC have shown that its SSB 35 and SSB 65 binding modes can be distinguished by a FRET assay [26]. Both of these ssDNA binding modes require a tetrameric (or di-tetrameric) structure of bound SSB. A similar FRET assay was used to probe the structure of SSB BA in ssDNA bound state. As the ssDNA binding constant of SSB BA is very high, the possibility that it may form a tetramer only in the ssDNA bound state was investigated. As shown in Figure 9A, SSB EC formed both SSB 35 and SSB 65 structures with the 5′-Cy5 (dT) 70 Cy3-3′ oligonucleotide as evidenced by FRET analysis. As described earlier, SSB 35 Figure 9B). In addition, the slope of the plot with SSB BA is different from that of SSB EC . As SSB EC is a stable tetramer, the high FRET complex formed rapidly with increasing SSB concentration and it formed much slowly with SSB BA because of the lack of a stable tetramer formation. Our results appeared to indicate (i) a tetrameric structure of SSB BA in the SSB BA -ssDNA complex, and (ii) the SSB BA -ssDNA complex was formed only in the SSB 65 mode. Due to high affinity of ssDNA binding, perhaps four monomers can bind the oligo(dT) 70 prior to the tetramer formation. Once this multi-SSB BA complex is formed, the bound SSB BA monomers undergo conformational transition and form tetrameric structure in SSB 65 mode. A hypothetical model is proposed in Figure 10. In this proposed model, all four monomers in the tetramer would first bind to the ssDNA, which would likely lead to the formation of the SSB 65 complex and prevent the formation of a SSB 35 complex. In the case of SSB BA , a two-fold higher concentration of protein was needed to observe the high-FRET complex. We believe this is due to the following reasons. First, the SSB BA is in essence a mutant form of SSB EC and as a result its ssDNA binding mechanism is likely somewhat different. Second, higher concentration of SSB BA might have favored the binding of all four monomers to the ssDNA. Initial slope of the plot in Figure 9B appeared to support this pathway.
SSB BA appears to undergo structural transformations which may support its high affinity binding to ssDNA. Its structure is due to a cumulative effect of multiple changes in key amino acid residues in its sequence which resulted in the loss of stable tetramer formation. Nonetheless, the SSB BA bound oligo(dT) 20 with high affinity as shown in Figures 4, 5, 6. Therefore, multiple monomers will bind to oligo(dT) 70 due to its long size. Thus, it is reasonable to assume that four monomers are binding to a long ssDNA (≥70 nucleotides). FRET data presented in Figure 9 established that SSB 65 -like structure is being formed upon oligo(dT) 70 binding. Therefore, the ssDNA binding is leading to the formation of an SSB 65 complex in which ssDNA is bound to a SSB EClike tetrameric structure. We have proposed a hypothetical model, presented in Figure 10, which may explain the mechanism of formation of a SSB BA tetramer assembly upon ssDNA binding which require further studies of such complex formation. The proposed model represents a stepwise process by which SSB BA can achieve high affinity DNA binding through a tetramer formation. This mechanism of SSB-ssDNA complex formation and its reversal may aid in the rapid removal of SSB, a necessary step, by enzymes such as a DNA polymerase during DNA replication as well as in other processes. In essence, SSB BA could actually be more effective than its tetrameric orthologs in executing its multifaceted functions in cellular DNA transactions.

Conclusions
Our studies suggest that the structural properties of SSB BA differ from that of its Gram-negative counterpart, SSB EC , and that furthermore its structure is modulated in the presence an ssDNA template. It is noteworthy that despite complexities in structure and oligomerization, SSB BA retains high-affinity ssDNA binding, which is its primary function. Its unique structure may be due to the cumulative effect of multiple key amino acid changes in its sequence during evolution, leading to alteration of stable dimer and tetramer formation. In the presence of a long ssDNA (≥70 nucleotides) appears to form with SSB BA a SSB 65 complex in which ssDNA is bound to all four SSB monomers in a tetrameric structure. A proposed model may explain the mechanism of such SSB BA -ssDNA complex formation through a transient tetramer formation. This model indicates that SSB BA may be more efficient in assembly and disassembly of the protein-DNA complex particularly during DNA replication. The physiological consequence(s) of the unusual structural dynamics of SSB BA , could be significant. Further studies are required to fully elucidate the role of protein•DNA and protein•protein interactions on SSB BA protein structure.

Nucleic acids and other reagents
Ultra pure nucleotides were obtained from GE Biosciences (Piscataway, NJ) and were used without further purification. All other chemicals used to prepare buffers and solutions were reagent grade and were purchased from the Fisher Chemical Company (Pittsburgh, PA).
HPLC ion exchange columns, ion exchange chromatography matrix, and the Bio-Cad 20 HPLC instrument were from Applied Biosystems Inc., Woburn, MA. The gel filtration column, TSK gel 3000SW, was from Tosoh Bioscience, King of Prussia, PA. Custom oligonucleotides for PCR and fluorescently labeled oligonucleotides were from Sigma-Aldrich (St. Louis, MO).

Buffers
Lysis buffer contained 25 mM Tris-HCl, (pH 7.9), 10% sucrose, 250 mM NaCl, and 0.001% NP40. Buffer A contained 25 mM Tris-HCl (pH 7.5), 5 mM MgCl 2 , 10% glycerol, 5 mM DTT, and NaCl in mM as indicated in the subscript. Buffer B, used for all fluorescence studies, contained 20 mM Hepes-NaOH (pH 7.5), 5 mM MgCl 2 and 1 mM DTT and 25 mM (unless otherwise indicated) ultrapure NaCl. In temperature and salt titration experiments, buffer B containing 5% ultrapure glycerol was used. Buffers for fluorescence measurements were prepared with HPLC-grade water (with minimal background fluorescence), fluorescence grade reagents, and filtered through a 0.2 μm nylon filter, examined for background fluorescence and Raman spectrum before use in anisotropy measurements. Background fluorescence was subtracted where necessary.

Cloning and expression of SSB BA
The SSB BA gene was amplified by PCR using B. anthracis genomic DNA, obtained as a gift from Dr. Theresa M. Koehler of the University of Texas Houston Health Science Center, Houston (33,34). This ORF codes for a 172 amino acid polypeptide with a predicted molecular weight of 19.2 kDa. The amplified gene was cloned into a pET29b vector (Novagen, Inc., Madison, WI) under the control of a T7 promoter (pET29b-SSB BA recombinant plasmid). The presence of the correct insert was confirmed by DNA sequencing. The SSB BA protein was over-expressed in E. coli strain BL21(DE3)RIL (Agilent Technologies Inc., Santa Clara, CA) harboring pET29b-SSB BA plasmid. Cells harboring the recombinant plasmid were grown in 2X-YT media containing 50 μg/ml of kanamycin, 20 μg/ml of tetracycline and 12 μg/ml of chloramphenicol with shaking at 37°C to an optical density at 600 nm of 0.4. IPTG (isopropyl-β-D-thiogalactopyranoside) was added to a final concentration of 0.25 mM. The cells were shaken for an additional two hours at 25°C, then harvested by centrifugation for 10 min at 5,000 x g. The cells were resuspended in 2.5% of the original culture volume of lysis buffer at 4°C and stored at −80°C until further use.

Purification of SSB BA
Cells were thawed, adjusted to pH 8.0 with 1 M Tris base, and lysed using 0.25 mg/ml lysozyme, 5 mM MgCl 2 , 5 mM spermidineÁHCl, and 2.5 mM DTT via incubation at ambient temperature for 60 min. The mixture was Dounce homogenized followed by centrifugation. The lysate was centrifuged at 43,000 x g for 30 min at 23°C . The supernatant was precipitated overnight using 0.25 g/ml ammonium sulfate at 4°C. This precipitate was collected by centrifugation at 43,000 x g for 30 min at 4°C, and dissolved in buffer A 0 (Fraction II). Fraction II was clarified by centrifugation at 43000 x g for 30 min. All steps were carried out at ambient temperature unless otherwise indicated.
The salt concentration of Fraction II was adjusted to the conductivity of buffer A 50 by diluting with buffer A 0 . The protein fraction was then passed through a 5 ml POROS-Q column equilibrated with buffer A 50 . SSB BA protein was eluted with a 150 ml gradient from A 100 to A 500 . The SSB BA fractions, identified by SDSPAGE, were pooled (fraction III). The salt concentration of Fraction III was adjusted to the conductivity of buffer A 50 by diluting with buffer A 0 . Diluted Fraction III was bound to a 5 ml S-Sepharose column equilibrated with A 50 . SSB BA was eluted using a 150 ml gradient from A 100 to A 500 . Fractions containing SSB BA were identified by SDS-PAGE and combined (Fraction IV). The Fraction IV, adjusted to 0.25 g/ml ammonium sulfate, was incubated on ice for two hours that resulted in the selective precipitation of SSB BA . The SSB BA precipitate was collected by centrifugation at 43000 x g for 60 min at 0-1°C. The pellet (Fraction V) was resuspended in 10 ml of buffer A 100 . Homogeneity was assessed by SDS-PAGE.

Assay of SSB biological activity
The standard assay, based on the stimulation of DNA helicase ativity of DnaB protein, was carried out in 1 ml of buffer B containing 1 mM DTT, 25 mM KCl, 3.5 mM ATP and 4.2 nmol of the 55 bp partial duplex substrate containing the following oligonucleotides [51]: Italicized nucleotides denote non-complimentary bases that create the fork structure of the duplex. Fluorescence emission spectra of the samples, before and after reaction, were recorded between 550-750 nm with 519 nm excitation with 8 nm slit-width. Reaction was initiated by adding 0.5 μg/ml DnaB BA helicase to the reaction mixture and incubated for 15 min at 37°C and FRET was measured. SSB BA (3 μg/ml) was added to the reaction mixture where indicated. DnaB BA helicase unwinding of the duplex led to inhibition of the FRET between Cy3 and Cy5. SSB BA was required for efficient helicase action of DnaB BA which was the basis of the assay. By using native and heat denatured substrates, it was determined that 1% decrease in FRET is equivalent to 3 pmol DNA unwinding in terms of base pairs (bp).

Steady-state fluorescence measurements
Fluorescence anisotropy was measured to investigate DNA binding by SSB BA in solution [40,52]. Fluorescence measurements were carried out using a steadystate-photon counting spectrofluorometer, PC1 with Vinci software, from ISS Instruments (Champaign, IL) and Fluoromax4-TCSPC with time-resolved fluorescence from Horiba Instruments Inc. (Edison, NJ). Excitation and emission slits were adjusted to 8 nm to maximize intensity counts [53]. Temperature during measurements was maintained using a programmable Peltier-controlled cuvette holder from Quantum Northwest Inc. (Seattle, WA).

Fluorescence anisotropy analysis of equilibrium ssDNA binding
The oligonucleotide was diluted to a concentration of 1 nM and titrated with SSB BA within a concentration range of 0.1 nM to 1μM. The sample was incubated for 2 min after each addition and thermostated at 25°C. Anisotropy measurements were carried out in general as described above except the excitation wavelength was 480 nm and emission anisotropy was measured at 540 nm. The standard deviation for the anisotropy values was <5 mA. Anisotropy at each titration point was measured three times for 10 s. and averaged. The total fluorescence intensity did not change significantly (≤10%) with increase in SSB BA concentration. Therefore, fluorescence lifetime changes, or the scattered excitation light, did not affect the anisotropy measurements.
Anisotropy; A; is defined as : where, G is the instrumental correction factor for the fluorometer and it is defined by G ¼ I hv =I hh I vv , I vh , I hv and I hh represent the fluorescence signal for excitation and emission with the polarizers set at (0 0 , 0 0 ), (0 0 , 90 0 ), (90 0 , 0 0 ) and (90 0 , 90 0 ) respectively.
The interaction of SSB BA with the labeled oligonucleotide can be represented as follows: Where, R is the ligand i.e., labeled oligonucleotides and P is SSB BA .
At equilibrium, K A , the equilibrium association constant can be given as Fraction of the binding sites occupied, f, can be represented as f ¼ occupied binding sites ½ = total binding sites Substituting for [RP] and rearranging the equation we get Similarly, equilibrium dissociation constant K D (K D = 1/K A ) can be expressed as Thus, K D can be further defined as the protein concentration at which half of the sites are occupied when ligand concentration is constant, as in the present case. Non-linear regression analysis of the anisotropy plot (anisotropy vs. log[SSB BA ]) was carried out using Prism 6.1 software (GraphPad Software Inc., San Diego, CA) and the concentration of SSB BA required to bind 50% of oligonucleotides was computed using the following equation:- where, Y MIN and Y MAX are the anisotropy values at the bottom and top plateaus respectively. X represents log [SSB BA ] (where [SSB BA ] represents total concentration of SSB BA ) and X 0 is the X value when the anisotropy is halfway between the top and the bottom of the plot and n app is the Hill coefficient.
FRET analysis of ssDNA binding by SSB EC and SSB BA FRET analysis was used to monitor ssDNA binding by SSB as described (22). Reaction mixtures were assembled on ice and incubated at 25°C for 5 min before FRET analysis. Reaction mixtures contained 40 nM labeled Cy5-(dT) 70 -Cy3 oligonucleotide and the indicated amount of SSB EC or SSB BA in a total volume of 1 ml. SSB BA or SSB EC titrations were performed with PC1 spectrofluorometer with the monochromator set at 515 nm for excitation for the Cy3