|Home | About | Journals | Submit | Contact Us | Français|
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.
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.
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.
Nearly all cellular nucleic acid transactions, including DNA replication, repair and recombination require the activity of a single stranded DNA binding protein (SSB) [1-7]. SSB proteins and are found throughout nature and their functional importance is underscored by their presence in prokaryotes, archaea, and eukaryotes including mammals . 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 [8-11].
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 [1,3,10,18-21].
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 . Kim and Richardson demonstrated that the bacteriophage T7 SSB, the gene 2.5 protein, is a dimer . 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 .
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 . The binding is modulated by monovalent salts. At very low salt concentration (<10mM 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 (>200mM 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 200mM 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 . Chrysogelos and Griffith discovered that repeated freezing-thawing of E. coli SSB leads to the formation of unique strings of tetramers .
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 Gram-negative E. coli and these organisms may contain multiple paralogues [31,32]. Sequence analysis indicated that Gram-positive SSBs have a highly conserved nearly-identical (>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 (SSBBA) 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.
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 19.2kDa.
The amino acid sequence of the N-terminal ssDNA binding and protein-protein interaction domains responsible for dimer and tetramer formation of SSBBA 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 Figure1.1. 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 SSBBA 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 (Figure1).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 C-terminal region lacked significant homology even among Gram-negative SSBs in the region between the residues 100–166 among these SSBs .
Recombinant SSBBA 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 SSBBA (Figure (Figure2A).2A). In SE-HPLC analysis of purified SSBBA (100 μg/ml) the protein eluted as a single peak with an elution volume consistent with that of a monomer (Figure (Figure33).
To test the biological activity of SSBBA we measured its ability to stimulate its cognate DnaB, DnaBBA, using a FRET based DNA helicase assay. It was based on the ability of SSBBA to stimulate DNA unwinding activity of its cognate DNA helicase, DnaBBA (Figure (Figure2).2). In the absence of SSBBA, DNA unwinding by DnaBBA was limited (Figure (Figure2B)2B) and was greatly stimulated in the presence of SSBBA (Figure (Figure2C).2C). The stimulation of the DNA helicase activity of DnaBBA in the presence of the purified SSBBA was >10 fold which was as expected for the cognate SSB .
Fluorescence anisotropy-based titration of a DNA-fluorophore 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 SSBBA and ssDNA interactions. In order to determine the binding constant for SSBBA, the interaction of SSBBA with 5′-flourescein labeled oligo(dT)20 was examined at 25°C and in buffer B containing 0.1 nM Fl-(dT)20, 25mM KCl and 5mMMg+2. Fl-(dT)20 was titrated with SSBBA until saturation in anisotropy was observed. The anisotropy values at various SSBBA concentrations were used to create a binding isotherm as a semi-log plot as shown in Figure Figure4.4. At very low SSBBA concentrations, very small anisotropy changes and a flat plateau (~43mA) were observed that were attributed to Fl-(dT)20. Upon further addition of SSBBA, the anisotropy value increased with an increase in SSBBA·Fl-(dT)20 complex formation. A sigmoid binding isotherm with saturation binding at high SSBBA concentration was observed with maximum anisotropy of 184 ± 5mA. At higher SSB concentration (≥ 1μM), 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 SSBBA concentration at which 50% of the ligand was in bound form (EC50); that value corresponds to the apparent dissociation constant. The KD for SSBBA·Fl-(dT)20 complex was 1.0 ± 0.1 x 10−9M. The Hill coefficient was 1.6 ± 0.6 indicating a possible binding of one to two molecules of SSBBA to Fl-(dT)20, which is not surprising considering the fact that SSBBA does not form a stable tetramer. A simple hyperbolic fit of the anisotropy data was also carried out (Figure (Figure44 inset). The estimated KD value was 1.2 ± 0.8 x 10−9M. In order to determine the correct method of analysis of binding data, we have examined the “Goodness of Fit” using both models. The R2 value for the sigmoidal dose–response model was found to be 0.98, whereas, for the simple hyperbolic model was 0.94. The F-test using the values of degrees-of-freedom and absolute sum of squares, gave an F-value of 22.4 which correlated well with the R2 values and clearly demonstrated that the sigmoidal dose–response model as the correct model/equation for fitting anisotropy binding data for such analysis.
In order to understand the thermodynamics of SSBBA·ssDNA binding interactions, we have analyzed SSBBA binding to ssDNA at different temperatures over a range of 20–37°C. The temperature-dependent binding isotherms for SSBBA and Fl-(dT)20 are presented in Figure Figure5A.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 temperature, the viscosity as well as the anisotropy values appeared to increase . Analysis of the binding curve data for KD 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 KD could likely be due to the dissociation of the protein·DNA complex at higher temperatures. The DnaABA·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, lnKD= −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 (Figure5B).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 (33.9kJmol−1). The change in entropy (ΔS°) was calculated to be 56.9Jmol−1K−1. Thus, it appears that the formation of the SSBBA·ssDNA complex has a strong entropic or hydrophobic component to the overall protein-DNA interaction.
In addition to hydrophobic interaction between SSBBA and ssDNA, we analyzed the contribution of ionic interactions, if any, in the complex formation. Fluorescence anisotropy was used determine the KD of formation of SSBBA·ssDNA complex as a function of NaCl concentration. The binding isotherm at different salt concentrations was generated using SSBBA titration of ssDNA at 25°C (Figure (Figure6A).6A). Each titration curve fits according to a single site binding isotherm. The data shows that highest affinity binding occurs at 0–25mM NaCl. The dissociation constant increased steadily from 0–250mM NaCl. The most striking change was in the values of the Hill coefficient. At 0mM NaCl, it was 1.9 ± 0.6 and steadily decreased to 0.9 ± 0.1 at 150mM NaCl and beyond. It could indicate that at 0mM NaCl, SSBBA 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 (Figure6B)6B) was generated from the KD values obtained from Figure Figure6A.6A. The presence of a negative slope was indicative of a net ion release . The data were analyzed using the following equation to determine the number of ions released upon binding:
The analysis suggests that upon SSBBA·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.
The SSBBA 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 SSBEC monomer, as determined earlier by X-ray crystallography [37,39,42].
Alignment of SSBBA sequence with sequences of other prokaryotic SSBs did not provide any clue to the basis of its ssDNA binding properties (Figure (Figure1).1). Therefore, the three-dimensional structure of SSBEC derived from X-ray crystallography  was explored. The significant sequence homology of SSBBA with E. coli SSB allowed us to develop a putative three dimensional model of SSBBA 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 . The structures for E. coli SSBEC, 1QVC and SSBBA were visualized using PyMOL (The PyMOL Molecular Graphics System, Version 1.3, Schrödinger LLC). Structures of both SSBs are presented in PyMol in Figure Figure77.
SSB proteins are known to bind ssDNA through their oligonucleotide/oligosaccharide binding fold (OB fold) as described by Murzin . The OB fold is characterized by a β-barrel consisting of five β-strands capped by an α-helix. Despite sequence differences between SSBBA and SSBEC, the OB fold observed in SSBEC remained intact in SSBBA including the β-turn regions, particularly L45 between β sheets 4 and 5 (Figure (Figure7).7). It has been shown that the β sheet 1 of SSBEC with the sequence VNKVILV is in the monomer-interface of the SSBEC dimer . In SSBBA, this β sheet remains partially intact (NKVILV) with the loss of the Val5 residue. However, the His56 of one monomer in SSBEC 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 SSBBA) and Leu83 (Leu76 in SSBBA) remained conserved, one of the most important residues, His56 in SSBEC was altered to Ile (Ile47) in SSBBA (Figure (Figure1).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 SSBBA will likely hinder a stable dimer formation.
The SSBEC 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 SSBBA, as shown in Figure Figure7,7, 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 SSBEC for this network of hydrogen bond formation are Lys8, Tyr79, Gln77, Glu81, and Gln111. Sequence comparison (Figure (Figure1)1) between SSBEC and SSBBA indicated that all of these residues in SSBBA 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 SSBBA are likely to impede tetramer formation further. Taken together, inhibition of both monomer-monomer interaction leading to dimer formation as well as dimer-dimer interaction leading to tetramer formation, the amino acid sequence of SSBBA does not support formation of stable dimer or tetramer.
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 SSBBA bound ssDNA with very high affinity (1.0 ± 0.1 x 10−9M) even though it did not appear to form a stable tetramer in the absence of DNA at the concentration range examined (Figures (Figures33 & 7).
The amino acid residues in SSBBA that are homologous to the residues in other SSBs, particularly SSBEC, that are known from crystallographic studies to bind to ssDNA were analyzed. Several hydrophobic residues, Trp89, Trp55, and Phe61, in SSBEC have been identified as involved in ssDNA binding through base stacking interactions . In Helicobacter pylori SSB (SSBHP), Phe37, Phe50, Phe56, and Trp84 are involved in base stacking interactions with ssDNA . These hydrophobic residues (Phe36, Phe43, Trp53, and Tyr81) with changes remained conserved in SSBBA (Figure (Figure1).1). As shown in Figure Figure8,8, 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 Figure8.8. 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 SSBBA could be due to basic as well as aromatic residues in its DNA binding groove.
Our studies demonstrated that SSBBA was capable of high-affinity 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; SSB35 and SSB65[26,48]. The SSB35 form binds approximately 35 nt and the SSB65 form binds approximately 65 nt on a (dT)70 template. In the SSB35 form, only two subunits of the tetramer make contacts with the DNA, whereas, in SSB65 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 . The SSB65 form produces high FRET and the SSB35 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 SSBBA forms higher order structures upon DNA binding. With this ssDNA substrate, we anticipated that SSBBA in SSB35-mode would produce attenuated FRET and SSB65-mode would produce high FRET. Binding of SSBEC to Cy5-(dT)70-Cy3 oligonucleotide was first analyzed to test the validity of our assay. The results are presented in Figure Figure9.9. FRET was measured using 515nm excitation wavelength and 665nm emission wavelength. The emission intensity was corrected for Cy5 contribution to 665nm emission. FRET was defined by (F-F0)/F0 where F0 and F are the corrected 665nm emission intensities of the 200 nM 5′-Cy5(dT)70Cy3-3′ oligonucleotide in the absence and presence of SSB respectively. In the titration with SSBEC, sharp increase in FRET with initial titration with SSBEC reaching a peak at 0.44 μM concentration was seen. Upon further titration with SSBEC, the FRET decreased substantially and reached a plateau at concentrations higher than 1.1 μM. These results with SSBEC are comparable to that observed by Roy et al. . However, our experiments required a somewhat higher concentration of SSBEC than reported by Roy et al. , which could be due to different buffer systems and ssDNA concentration. Thus, at 0.44 μM SSBEC, the high FRET (SSBEC)4-(dT)70 complex was observed and at or above 1.1 μM SSBEC, intermediate FRET (SSBEC)8-(dT)70 complex was seen.
Next, SSBBA binding to Cy5(dT)70Cy3 oligonucleotide was analyzed. The Cy5(dT)70Cy3 oligonucleotide was titrated with SSBBA as described above for SSBEC. Similar to SSBEC, FRET increased linearly with increasing concentration of SSBBA, and reached a plateau at 0.8 μM Surprisingly, only a high FRET form of (SSBBA)-(dT)70 complex was observed. However, initial slope of the plot is very different from that observed with SSBEC which is a stable tetramer. Therefore, it appeared that the high FRET form of the complex was formed but required higher concentration of SSBBA. The high FRET form of (SSBEC)-(dT)70 complex is a tetrameric complex in which all four monomers bind ssDNA. The high FRET form of (SSBBA)-(dT)70 complex should have the same oligomeric structure as the high FRET form of (SSBEC)-(dT)70 complex. As SSBBA 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 SSBEC 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 (SSBBA)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 Figure4.4. 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 SSBBA resulting in progressive lack of sufficient open DNA for second, third and fourth SSBBA 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 SSBBA concentration to achieve a high FRET complex. Based on this reasoning of the (SSBBA)4-(dT)70 complex formation, as described above, our data actually supports the model in Figure Figure10.10. Lack of observation of an intermediate FRET form of the (SSBBA)-(dT)70 complex in this study indicated that a (SSBBA)8-(dT)70 complex probably did not form in appreciable amount even with high proportional levels of SSBBA. Based on these reasonings of the (SSBBA)4-(dT)70 complex formation, as described above, our data led to the proposal of the model presented in Figure Figure1010 for the (SSBBA)8-(dT)70 complex.
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.
E. coli SSBEC is a stable tetramer with high solubility and tremendous thermal stability . The majority of prokaryotic cellular SSBs are homotetramers, where each monomer harbors an OB fold. However there are exceptions. SSBDR from Deinococcus radiodurans is a homodimer, where each monomer is quite large and contains two OB folds [37-39,46]. 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. SSBBA was found to be not tetrameric at or above ambient temperature by size exclusion HPLC (Figure (Figure3).3). This physicochemical property of SSBBA is closely comparable to the T4 bacteriophage SSB, which is monomeric.
Sequence alignment and three dimensional structure of SSBBA, generated by homology-based modeling were utilized to probe the molecular basis of ssDNA binding (Figures (Figures1).1). Sequence alignment and secondary structure prediction (data not shown) clearly indicated the presence of an OB fold in SSBBA, which is a characteristic of SSBs and required for high affinity ssDNA binding (Figure (Figure44).
Both sequence alignment and homology modeling (Figures (Figures11 & 10) indicated lack of several residues that are important for monomer-monomer and dimer-dimer interactions leading to the formation of a stable tetramer in SSBEC. His56 (as well as Glu54) of β sheet 3 (Figure (Figure1)1) in SSBEC 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 . This mutant does not form a stable tetramer with respect to monomers at non-permissive temperatures [5,45,50]. Thus, a lack of the corresponding His residue (His→Ile change) in SSBBA 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 (Figure1).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 SSBBA, 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 SSBBA 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 SSBBA, as described above, are likely contributed to the disruption of monomer-monomer interaction leading to dimer formation.
The dimer-dimer interface in the SSBEC tetramer is primarily a six-stranded β sheet-mediated. Residues that are important in SSBEC for the network of hydrogen bond formation at the dimer-dimer interface are Lys8, Tyr71, Gln77, Glu81, and Gln111 . Sequence comparison (Figure (Figure1)1) between SSBEC and SSBBA indicated that all of these residues in SSBBA 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 SSBBA. An equally significant change is observed with Gln77 which is changed to Leu69 in SSBBA. Taken together, these changes in amino acid sequence may disrupt both monomer-monomer and dimer-dimer interactions leading to a monomeric SSBBA at a physiological temperature.
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 SSBBA. 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 SSBBA·ssDNA binding.
The KD values were determined at different temperatures (20–37°C) (Figure (Figure5).5). Our data showed that although SSBBA 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 −23kJmol−1 at 25°C. Using the two equations: ΔG°= −RT lnKD and ΔG°=ΔH°−TΔS° and the slope of the plot, we determined that ΔS° was ~188Jmol−1K−1 in this temperature range.
We determined the KD value of SSBBA binding to ssDNA at different salt concentrations (0–200mM NaCl) (Figure (Figure6A).6A). The binding is progressively weakened with an increase in ionic strength. The KD values were then analyzed using a linkage plot to determined ionic interactions in the binding. The negative slope of the plot in Figure Figure5B5B indicated a release of ions from. Our analysis determined the release of one Na+ and one Cl- ion during the binding process. In addition, our results also pointed out that SSBBA likely formed a tetrameric species at 0mM 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 KD value of the complex at infinite salt concentration (KD∞) by nonlinear regression of KD versus log[NaCl] plot (data not shown). The value of ΔG°ionic was −8kJmol−1.
Despite differences between its Gram-negative counterpart, SSBBA bound to ssDNA with high affinity (Figure (Figure4).4). The ssDNA binding affinity (KD) for a SSBBA monomer binding to a small oligonucleotide was 1.0 ± 0.1 x 10−9M at 25°C. Even though many changes in amino acid sequence of SSBBA directly relate to ssDNA binding, such as Trp55→Phe47, Trp90→Tyr81, Phe61→Trp53, the changes are not drastic enough to alter ssDNA binding (Figure (Figure8).8). Three dimensional structure as well as electrostatic surface potential in Figure Figure88 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 SSBBA to form tetramers upon DNA binding. A likely possibility is that SSBBA 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 .
Previous studies with SSBEC have shown that its SSB35 and SSB65 binding modes can be distinguished by a FRET assay . 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 SSBBA in ssDNA bound state. As the ssDNA binding constant of SSBBA is very high, the possibility that it may form a tetramer only in the ssDNA bound state was investigated. As shown in Figure Figure9A,9A, SSBEC formed both SSB35 and SSB65 structures with the 5′-Cy5(dT)70Cy3-3′ oligonucleotide as evidenced by FRET analysis. As described earlier, SSB35 represented the intermediate-FRET complex and SSB65 represented the high-FRET SSB-ssDNA complex. At a low SSB to dT70 ratio, it formed the high FRET complex and at a high SSB to dT70 ratio, it formed the intermediate FRET complex. In the FRET analysis, SSBBA formed only a high-FRET complex but not the intermediate-FRET complex (Figure (Figure9B).9B). In addition, the slope of the plot with SSBBA is different from that of SSBEC. As SSBEC is a stable tetramer, the high FRET complex formed rapidly with increasing SSB concentration and it formed much slowly with SSBBA because of the lack of a stable tetramer formation. Our results appeared to indicate (i) a tetrameric structure of SSBBA in the SSBBA-ssDNA complex, and (ii) the SSBBA-ssDNA complex was formed only in the SSB65 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-SSBBA complex is formed, the bound SSBBA monomers undergo conformational transition and form tetrameric structure in SSB65 mode. A hypothetical model is proposed in Figure Figure10.10. 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 SSB65 complex and prevent the formation of a SSB35 complex. In the case of SSBBA, 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 SSBBA is in essence a mutant form of SSBEC and as a result its ssDNA binding mechanism is likely somewhat different. Second, higher concentration of SSBBA might have favored the binding of all four monomers to the ssDNA. Initial slope of the plot in Figure Figure9B9B appeared to support this pathway.
SSBBA 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 SSBBA bound oligo(dT)20 with high affinity as shown in Figures Figures4,4, ,5,5, ,6.6. 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 Figure99 established that SSB65-like structure is being formed upon oligo(dT)70 binding. Therefore, the ssDNA binding is leading to the formation of an SSB65 complex in which ssDNA is bound to a SSBEC-like tetrameric structure. We have proposed a hypothetical model, presented in Figure Figure10,10, which may explain the mechanism of formation of a SSBBA tetramer assembly upon ssDNA binding which require further studies of such complex formation. The proposed model represents a stepwise process by which SSBBA 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, SSBBA could actually be more effective than its tetrameric orthologs in executing its multifaceted functions in cellular DNA transactions.
Our studies suggest that the structural properties of SSBBA differ from that of its Gram-negative counterpart, SSBEC, and that furthermore its structure is modulated in the presence an ssDNA template. It is noteworthy that despite complexities in structure and oligomerization, SSBBA 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 SSBBA a SSB65 complex in which ssDNA is bound to all four SSB monomers in a tetrameric structure. A proposed model may explain the mechanism of such SSBBA-ssDNA complex formation through a transient tetramer formation. This model indicates that SSBBA 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 SSBBA, could be significant. Further studies are required to fully elucidate the role of protein·DNA and protein·protein interactions on SSBBA protein structure.
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).
Lysis buffer contained 25mM Tris–HCl, (pH 7.9), 10% sucrose, 250mM NaCl, and 0.001% NP40. Buffer A contained 25mM Tris–HCl (pH 7.5), 5mM MgCl2, 10% glycerol, 5mM DTT, and NaCl in mM as indicated in the subscript. Buffer B, used for all fluorescence studies, contained 20mM Hepes-NaOH (pH 7.5), 5mM MgCl2 and 1mM DTT and 25mM (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.
The SSBBA 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 19.2kDa. The amplified gene was cloned into a pET29b vector (Novagen, Inc., Madison, WI) under the control of a T7 promoter (pET29b-SSBBA recombinant plasmid). The presence of the correct insert was confirmed by DNA sequencing. The SSBBA protein was over-expressed in E. coli strain BL21(DE3)RIL (Agilent Technologies Inc., Santa Clara, CA) harboring pET29b-SSBBA plasmid. Cells harboring the recombinant plasmid were grown in 2X-YT media containing 50μg/ml of kanamycin, 20μg/ml of tetracycline and 12μg/ml of chloramphenicol with shaking at 37°C to an optical density at 600nm of 0.4. IPTG (isopropyl-β-D-thiogalactopyranoside) was added to a final concentration of 0.25mM. The cells were shaken for an additional two hours at 25°C, then harvested by centrifugation for 10min 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.
Cells were thawed, adjusted to pH 8.0 with 1M Tris base, and lysed using 0.25mg/ml lysozyme, 5mM MgCl2, 5mM spermidine·HCl, and 2.5mM DTT via incubation at ambient temperature for 60min. The mixture was Dounce homogenized followed by centrifugation. The lysate was centrifuged at 43,000 x g for 30min at 23°C. The supernatant was precipitated overnight using 0.25g/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 A0 (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 A50 by diluting with buffer A0. The protein fraction was then passed through a 5ml POROS-Q column equilibrated with buffer A50. SSBBA protein was eluted with a 150ml gradient from A100 to A500. The SSBBA fractions, identified by SDSPAGE, were pooled (fraction III). The salt concentration of Fraction III was adjusted to the conductivity of buffer A50 by diluting with buffer A0. Diluted Fraction III was bound to a 5mlS-Sepharose column equilibrated with A50. SSBBA was eluted using a 150ml gradient from A100 to A500. Fractions containing SSBBA were identified by SDS-PAGE and combined (Fraction IV). The Fraction IV, adjusted to 0.25g/ml ammonium sulfate, was incubated on ice for two hours that resulted in the selective precipitation of SSBBA. The SSBBA precipitate was collected by centrifugation at 43000 x g for 60 min at 0–1°C. The pellet (Fraction V) was resuspended in 10ml of buffer A100. Homogeneity was assessed by SDS-PAGE.
The standard assay, based on the stimulation of DNA helicase ativity of DnaB protein, was carried out in 1ml of buffer B containing 1mM DTT, 25mM KCl, 3.5mM ATP and 4.2nmol of the 55bp partial duplex substrate containing the following oligonucleotides :
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 550–750nm with 519nm excitation with 8nm slit-width. Reaction was initiated by adding 0.5 μg/ml DnaBBA helicase to the reaction mixture and incubated for 15min at 37°C and FRET was measured. SSBBA (3 μg/ml) was added to the reaction mixture where indicated. DnaBBA helicase unwinding of the duplex led to inhibition of the FRET between Cy3 and Cy5. SSBBA was required for efficient helicase action of DnaBBA 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).
Fluorescence anisotropy was measured to investigate DNA binding by SSBBA in solution [40,52]. Fluorescence measurements were carried out using a steady-state-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 8nm to maximize intensity counts . Temperature during measurements was maintained using a programmable Peltier-controlled cuvette holder from Quantum Northwest Inc. (Seattle, WA).
The oligonucleotide was diluted to a concentration of 1 nM and titrated with SSBBA within a concentration range of 0.1 nM to 1μM. The sample was incubated for 2min after each addition and thermostated at 25°C. Anisotropy measurements were carried out in general as described above except the excitation wavelength was 480nm and emission anisotropy was measured at 540nm. The standard deviation for the anisotropy values was <5mA. Anisotropy at each titration point was measured three times for 10s. and averaged. The total fluorescence intensity did not change significantly (≤10%) with increase in SSBBA concentration. Therefore, fluorescence lifetime changes, or the scattered excitation light, did not affect the anisotropy measurements.
where, G is the instrumental correction factor for the fluorometer and it is defined by
Ivv, Ivh, Ihv and Ihh represent the fluorescence signal for excitation and emission with the polarizers set at (00, 00), (00, 900), (900, 00) and (900, 900) respectively.
The interaction of SSBBA with the labeled oligonucleotide can be represented as follows:
Where, R is the ligand i.e., labeled oligonucleotides and P is SSBBA.
At equilibrium, KA, the equilibrium association constant can be given as
Fraction of the binding sites occupied, f, can be represented as
Substituting for [RP] and rearranging the equation we get
Similarly, equilibrium dissociation constant KD (KD=1/KA) can be expressed as
Thus, KD 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[SSBBA]) was carried out using Prism 6.1 software (GraphPad Software Inc., San Diego, CA) and the concentration of SSBBA required to bind 50% of oligonucleotides was computed using the following equation:-
where, YMIN and YMAX are the anisotropy values at the bottom and top plateaus respectively. X represents log[SSBBA] (where [SSBBA] represents total concentration of SSBBA) and X0 is the X value when the anisotropy is halfway between the top and the bottom of the plot and napp is the Hill coefficient.
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 5min before FRET analysis. Reaction mixtures contained 40 nM labeled Cy5-(dT)70-Cy3 oligonucleotide and the indicated amount of SSBEC or SSBBA in a total volume of 1ml. SSBBA or SSBEC titrations were performed with PC1 spectrofluorometer with the monochromator set at 515nm for excitation for the Cy3 donor and with the monochromator set at 665nm emissions for the Cy5 acceptor. Slit width was 8nm.
B. anthracis: Bacillus anthracis; ssDNA: Single stranded DNA; DTT: Dithiothreitol; SDS: Sodium dodecyl sulfate; PAGE: Polyacrylamide gel electrophoresis; rNTP: Ribonucleotide triphosphate; BSA: Bovine serum albumin; SSB: Single stranded DNA binding protein; SSBBA: SSB of Bacillus anthracis; DnaBBA: DnaB helicase of Bacillus anthracis; FRET: Fluorescence resonance energy transfer; mA: Millianisotropy; kDa: Kilo Dalton.
The author(s) declare that they have no competing interests.
EB-F was involved in the design of the study, experiments, and manuscript preparation. JK was involved in the design of the study and carried out purification and fluorescence experiments. SB was involved in the design of the study and overall direction of the project. All authors read and approved the final manuscript.
Authors gratefully acknowledge support of this work by grants from the National Institute of Allergy & Infectious Diseases, NIH and UMDNJ Foundation. Authors wish to thank Ms. S. Rotoli and K-Y Luu for DNA helicase assay of SSBBA, Ms. J. Debski for critical review of the manuscript, Ms. Julia Crawford for F-test, Mr. Robert McBride of Educational Media, UMDNJ-SOM for illustration. We also thank the anonymous reviewers for helpful suggestions and comments.