Functional Roles of the N- and C-Terminal Regions of the Human Mitochondrial Single-Stranded DNA-Binding Protein

Biochemical studies of the mitochondrial DNA (mtDNA) replisome demonstrate that the mtDNA polymerase and the mtDNA helicase are stimulated by the mitochondrial single-stranded DNA-binding protein (mtSSB). Unlike Escherichia coli SSB, bacteriophage T7 gp2.5 and bacteriophage T4 gp32, mtSSBs lack a long, negatively charged C-terminal tail. Furthermore, additional residues at the N-terminus (notwithstanding the mitochondrial presequence) are present in the sequence of species across the animal kingdom. We sought to analyze the functional importance of the N- and C-terminal regions of the human mtSSB in the context of mtDNA replication. We produced the mature wild-type human mtSSB and three terminal deletion variants, and examined their physical and biochemical properties. We demonstrate that the recombinant proteins adopt a tetrameric form, and bind single-stranded DNA with similar affinities. They also stimulate similarly the DNA unwinding activity of the human mtDNA helicase (up to 8-fold). Notably, we find that unlike the high level of stimulation that we observed previously in the Drosophila system, stimulation of DNA synthesis catalyzed by human mtDNA polymerase is only moderate, and occurs over a narrow range of salt concentrations. Interestingly, each of the deletion variants of human mtSSB stimulates DNA synthesis at a higher level than the wild-type protein, indicating that the termini modulate negatively functional interactions with the mitochondrial replicase. We discuss our findings in the context of species-specific components of the mtDNA replisome, and in comparison with various prokaryotic DNA replication machineries.


Introduction
Single-stranded DNA-binding proteins (SSBs) are essential components in DNA metabolic processes, including replication, repair and recombination. In addition to their protective singlestranded DNA-coating properties, it has been demonstrated that SSBs from distantly-related species have far more complex roles that include the organization and/or mobilization of all aspects of DNA metabolism (reviewed in [1]). In eukaryotic cells, there are two compartmentalized SSBs: replication protein A is found in the nucleus, whereas mtSSB is found in the mitochondrion. Despite sharing a similar single-stranded DNA-binding domain (oligonucleotide/oligosaccharide binding domain or OB-fold) together with bacterial and viral SSBs [2,3,4,5,6], and performing analogous functions in their respective cellular compartments, mtSSB and RPA are not related evolutionarily. mtSSBs are homologues of eubacterial SSBs, whose prototype is the well studied Escherichia coli SSB (EcSSB) (Fig. 1), a fact that is in agreement with the endosymbiont theory of mitochondrial origin [7].
At the mitochondrial DNA (mtDNA) replication fork, mtSSB interacts functionally with DNA polymerase c (pol c) and mtDNA helicase (also known as Twinkle) to promote mtDNA replication. In vitro, the Drosophila melanogaster mtSSB (DmmtSSB) stimulates 15-to 20-fold the DNA polymerase and 39-59 exonuclease activities of Drosophila pol c (Dmpol c) on a singly-primed single-stranded DNA (ssDNA) template [8]. The human mtSSB (HsmtSSB) stimulates the DNA unwinding activity of the human mtDNA helicase (HsmtDNA helicase) [9], and is required for strand-displacement DNA synthesis in the presence of human pol c (Hspol c) and HsmtDNA helicase [10]. In vivo, deleterious mutations in the mtSSB gene cause loss of mtDNA that results in developmental arrest in Drosophila [11], and growth limitation due to mitochondrial dysfunction in Saccharomyces cerevisiae [12]. Our group has shown that the knockdown of the endogenous DmmtSSB in Drosophila Schneider cells results in mtDNA copy number reduction and growth retardation [13]. The overexpression of the wild-type protein rescues the phenotype, whereas a ssDNA-binding mutant of DmmtSSB is unable to do so. Very recently, the knockdown of HsmtSSB in human HeLa cells was shown to promote a gradual decline in mtDNA copy number and a severe reduction in 7S DNA synthesis [14].
In this report, we examine the biochemical and physical properties of four HsmtSSB proteins: the mature full-length protein (notwithstanding the mitochondrial presequence, HsmtSSBwt), a deletion variant lacking the first 9 residues in the N-terminus (HsmtSSBDN), a deletion variant lacking the last 7 residues in the C-terminus (HsmtSSBDC), and a variant that lacks both termini (HsmtSSBDNDC). The target regions are of particular interest for several reasons: 1) they represent two of the few regions of significant amino acid sequence variability between mtSSBs and bacterial SSBs -the N-terminal extension is absent in bacterial SSBs, whereas the C-terminus is short and uncharged in mtSSBs (Fig. 1); 2) they appear disordered in the crystal structure of HsmtSSB [3], suggesting flexibility and/or dynamism of these regions without apparent interactions with the ssDNA-binding domain; and 3) in EcSSB, bacteriophage T7 gp2.5 and T4 gp32 SSBs, the C-terminal region interacts with other components of DNA metabolic processes, and regulates ssDNA binding negatively. We evaluate our findings in the context of functions at the mitochondrial replication fork, and discuss them in comparison with other DNA replication systems.

Results
Purification of amino-and carboxyl-terminal deletion variants of HsmtSSB, and determination of oligomeric state In order to study the possible roles of the N-and C-terminal regions of HsmtSSB at the mtDNA replication fork, we produced three deletion variants together with HsmtSSBwt: HsmtSSBDN, HsmtSSBDC, and HsmtSSBDNDC. The overexpression of untagged recombinant proteins in E. coli using the pET-11a system resulted in high levels of soluble proteins, except for HsmtSSBDC; the extraction and solubility of HsmtSSBDC was dependent on inclusion of a dodecyl-maltoside detergent. From this point on, the purification of the proteins was identical. We modified earlier protocols for purification of DmmtSSB [8] and HsmtSSB [15] in order to obtain highly pure proteins (Fig. 2). In particular, substitution of velocity sedimentation, used previously as a final step, by phosphocellulose chromatography followed by assay of ATP-dependent dsDNA unwinding activity in each fraction, was a critical improvement towards the elimination of a highly active bacterial contaminant (see ''Materials and Methods'' for details).
With near-homogeneous HsmtSSBwt and the terminal deletion variants in hand, we evaluated the possible consequences of the lack of the termini on its physical and biochemical properties. SDS-PAGE of HsmtSSBwt, DN, DC, and DNDC reveals polypeptides of ,15, 14, 14, and 13 kDa, respectively (Fig.2). To investigate their oligomeric state in solution, we employed hydrodynamic methods to estimate native molecular mass. We observed single peaks for each recombinant protein both in velocity sedimentation and in Superdex 75 gel filtration (Fig. 3). The sedimentation coefficient was 4.2 S for all of the proteins, and the Stokes' radii were 3.4, 3.2, 3.2, and 3.0 nm respectively for Figure 1. Sequence alignment of animal mtSSBs with E. coli SSB, and mutagenesis of HsmtSSB. Thirteen mtSSB sequences (notwithstanding the mitochondrial presequence) from various species across the animal kingdom were aligned with the E. coli SSB sequence using Clustal X [49]. Only the representative mtSSB sequences from humans (GenBank accession: NP_003134), fruitfly (Drosophila melanogaster; GenBank accession: AAF16936), silkworm (Bombyx mori; GenBank accession: ABF51293), and frog (Xenopus laevis; GenBank accession: NP_001095241) are shown. The gray boxes denote the N-and C-terminal regions of animal mtSSB that are not conserved in E. coli SSB. The amino acid residues indicated in bold in the HsmtSSB sequence were targeted for deletion mutagenesis. Amino acid residues 13-123 in the HsmtSSB sequence comprise the OB-fold domain (see the text for details). doi:10.1371/journal.pone.0015379.g001

Terminal deletion variants of HsmtSSB bind to ssDNA with similar affinities
We proceeded to evaluate the HsmtSSB deletion variants by examining their biochemical activities as compared to the wildtype protein. Using a gel mobility shift assay (GMSA), we examined the apparent DNA-binding affinities of HsmtSSBwt and the terminal deletion variants using a 48-mer ssDNA oligonucleotide (Fig. 4), whose size is close to the binding-site size determined previously for HsmtSSB [16]. A titration of the proteins at 50 mM NaCl revealed no significant differences in ssDNA-binding affinities between HsmtSSBwt and the deletion variants, with apparent K d s of ,5 nM. Interestingly, the lack of the C-terminal region does not appear to interfere with the ssDNAbinding efficacy of HsmtSSB, in contrast to EcSSB and T7 gp2.5. A recent report by Kozlov et al. [17] shows that an increased ssDNA binding of EcSSB lacking the C-terminus is observed only at 100 and 200 mM NaCl, and not at low salt concentration (20 mM NaCl). We also performed the GMSA assays with the HsmtSSB proteins at 20 and 100 mM NaCl, but failed to observe any differences in binding affinities (data not shown). At the 20-100 mM range, salt does not seem to affect the ssDNA-binding affinity of either the wild-type or variant forms of HsmtSSB.

Salt-dependent stimulation of Hspol c by HsmtSSB terminal deletion variants
To evaluate the capacity of the HsmtSSB variants to stimulate the DNA polymerase activity of Hspol c in vitro, we first examined the effects of KCl concentration on DNA synthesis by reconstituted Hspol c holoenzyme on singly-primed M13 DNA in the presence and absence of HsmtSSBwt (Fig. 5). In the absence of HsmtSSBwt, Hspol c activity is stimulated ,2 fold as the concentration of KCl in the reaction increases from 20 to 100 mM, in agreement with previously published data [18]. Unlike Dmpol c stimulation by DmmtSSB that occurs over a broad range of KCl concentrations (0-130 mM) [8], the stimulation of Hspol c by HsmtSSBwt is only observed at 20 mM KCl, and reaches its maximum at a concentration of HsmtSSBwt sufficient to cover the entire singly-primed M13 DNA substrate (according to our GMSA data). At 50 mM KCl, the presence of HsmtSSBwt at low levels promotes a slight stimulation of DNA synthesis by Hspol c, but it becomes somewhat inhibitory at higher levels. At 100 mM KCl, where the activity of Hspol c alone is highest, the presence of HsmtSSBwt is completely inhibitory. The maximal stimulation of DNA synthesis by Hspol c in the presence of HsmtSSBwt at 20 mM KCl is only moderate, albeit ,3-fold higher than the activity of Hspol c alone at 100 mM KCl. Judging by the fact that the ssDNA-binding affinity of HsmtSSBwt does not change over the range of 20-100 mM KCl, the data suggest that increasing ionic strength inhibits the ability of Hspol c to displace HsmtSSB from ssDNA template during the course of in vitro DNA synthesis.
Next, we evaluated the ability of the terminal deletion variants of HsmtSSB to stimulate DNA synthesis by Hspol c (Fig. 6). Overall, the effect of KCl concentration is similar among the variants: at 20 mM, DNA synthesis is stimulated; at 50 mM, we observe no stimulation and/or slight inhibition; at 100 mM, Hspol c is inhibited completely. Notably, under low salt conditions, HsmtSSBDN, DC, and DNDC show 1.4-to 2-fold higher stimulation of Hspol c as compared to HsmtSSBwt. This increased stimulation is not apparent at low concentrations of the HsmtSSBs, but it is clearly evident and reproducible at saturating levels. The data argue that the N-and C-terminal regions of HsmtSSB have functionally inhibitory roles on its ability to stimulate the DNA polymerase activity of Hspol c, suggesting a modulatory role in vivo.

Stimulation of DNA unwinding activity of HsmtDNA helicase by terminal deletion variants of HsmtSSB
As shown previously, HsmtSSBwt stimulates the DNA unwinding activity of HsmtDNA helicase in vitro [9]. We examined the effects of the terminal deletion variants in stimulating the DNA unwinding activity of HsmtDNA helicase using as substrate pBSKS+ ssDNA (2,958 nt), to which was annealed a 60-mer ssDNA oligonucleotide that creates a 40-nt 59-single-stranded tail for helicase loading followed by 20 nt of paired sequence. First, we asked if varying KCl concentrations produced the same pattern of stimulation of HsmtDNA helicase by HsmtSSBwt as compared to its stimulation of Hspol c (Fig. 7). We found that HsmtDNA helicase shows a slightly better DNA unwinding activity at 100 mM KCl as compared to that at 20 and 50 mM KCl. We then evaluated various potassium salts, including phosphate, We extended our analysis by titrating HsmtSSBwt, DN, DC, and DNDC in DNA unwinding assays conducted at 50 mM KCl (Fig. 8). None of the concentrations of the HsmtSSBs used were sufficient to cause any dsDNA destabilization in the absence of helicase (Fig. 8A). HsmtDNA helicase shows maximal DNA unwinding activity in the presence of 100 nM HsmtSSB, a concentration corresponding to coating of ,80% of the ssDNA substrate (according to our GMSA data). No significant differences in stimulation were observed between HsmtSSBwt and deletion variants; stimulation of the HsmtDNA helicase was ,8 fold at the highest HsmtSSB concentrations.

Discussion
Biochemical studies of the mtDNA replisome demonstrate that pol c and the mtDNA helicase are stimulated by mtSSB. We have shown previously that DmmtSSB can stimulate ,20-fold in vitro DNA synthesis by Dmpol c over a broad range of KCl concentrations, primarily by enhancing primer recognition and binding [8,19]. Similarly, DmmtSSB stimulates ,15-fold the 39-59 exonuclease activity of Dmpol c over the same KCl range [8]. Furthermore, HsmtSSB has been shown to stimulate the DNA unwinding activity of the HsmtDNA helicase [9] and together with Hspol c and HsmtDNA helicase, reconstitutes a minimal  mtDNA replisome in vitro [10]. These data indicate that mtSSBs serve an important role in initiation and elongation of DNA synthesis in mtDNA replication, consistent with the observation that the disruption of the Drosophila (lopo) and yeast (rim1) mtSSB genes results in loss of mtDNA and respiratory capacity, and consequently, developmental lethality or impaired growth, respectively [11,12].
mtSSBs are homologues of EcSSB, with which they exhibit a high degree of amino acid sequence conservation in the OB-fold domain (Fig. 1), and share similar physical, biochemical, and structural properties [3,16,19,20]. Animal mtSSBs, however, evolved at least two different sequences from their eubacterial counterpart: their Nand C-termini. After import into mitochondria, the mitochondrial presequence is cleaved, producing the mature mtSSB protein. The mature polypeptide in humans contains 9 amino acids at the Nterminus that are absent in EcSSB, and only 7 residues at the Cterminus, in contrast with the long acidic C-terminal tail of its eubacterial homologue; these potentially protrude from the ssDNAbinding core without any detectable secondary structure [3]. Here, we sought to analyze the functional importance of the N-and Cterminal regions of HsmtSSB in the context of mtDNA replication. We purified HsmtSSBwt, HsmtSSBDN, HsmtSSBDC, and HsmtSSBDNDC to near-homogeneity and showed that all of the proteins form tetramers in solution, indicating that the termini are not critical for the folding or stability of HsmtSSB, as predicted by the crystallographic data [3].
Next, we analyzed the ssDNA-binding properties of the terminal deletion variants using a GMSA approach. The lack of either or both termini did not alter the ssDNA-binding affinity of HsmtSSB, which is relatively high as observed for SSBs from various sources [21]. This is particularly relevant because we have shown previously that ssDNA-binding variants of DmmtSSB fail to stimulate Dmpol c efficiently in vitro, and promote mtDNA depletion and cell growth defects in Drosophila Schneider cells [13]. Our findings distinguish biochemically the role of the Cterminal region of mtSSBs as compared to prokaryotic forms. The crystal structures of viral, eubacterial, eukaryotic nuclear and mitochondrial SSBs show that these proteins share a common structural domain for binding to ssDNA, the OB-fold [2,3,4,5,6], although they share no sequence homology. In addition, eubacterial and bacteriophage SSBs do share another structural feature: a long acidic C-terminal tail, which is essential for DNA replication and viability of the organisms [22,23,24]. The removal of the C-termini of EcSSB, T7 gp2.5 and T4 gp32 increases their ssDNA-binding affinities significantly [22,24,25] but at the same time, abolishes interactions with other components of their cognate replication machinery [1,26,27,28]. Recently, Shereda et al. [29] showed that proteins that interact with the C-terminus of E. coli SSB share a similar structural surface where the interaction occurs, demonstrating the presence of a signature contact structure. Interaction with the C-terminus of E. coli SSB via this signature structure appears to be highly regulated, because progressive truncations from its C-terminal end cause a progressive loss of both physical and functional interactions. Marintcheva et al. [30] showed that the acidic C-terminus of T7 gp2.5 and ssDNA actually compete for binding to the ssDNA-binding cleft of the protein, which is located in the N-terminal OB-fold domain. A functional model proposes that in absence of DNA, the C-terminal region binds to the ssDNA-binding cleft, and is then displaced upon ssDNA binding, rendering it available for protein-protein interactions. Such an interaction between the C-terminus and the ssDNA-binding cleft is suggested to create an electrostatic shield that protects the binding cleft from random charged surfaces inside the cell. Our data shows clearly that the C-terminal region of HsmtSSB does not influence the ssDNA-binding affinity of the protein, suggesting that this region of the eubacterial-like mtSSBs serves a role that differs from those of EcSSB, T7 gp2.5 and T4 gp32. Indeed, in this sense, HsmtSSB resembles the nuclear replication protein A (reviewed in [31]); binding of random charged molecules to the ssDNA-binding cleft of eukaryotic SSBs is thus most likely prevented by a distinct mechanism.
To investigate further the functional properties of the N-and Cterminal regions of HsmtSSB, we performed stimulation assays of Hspol c and HsmtDNA helicase under varying conditions. We found that the stimulation of the DNA polymerase activity of  Hspol c by HsmtSSBwt is moderate (,3 fold) and observed only at low KCl concentrations (20 mM). Increasing KCl concentrations resulted in concentration-dependent inhibition, suggesting that electrostatic forces govern the functional interactions between Hspol c and HsmtSSB. Interestingly, in vitro stimulation (up to 20 fold) of the DNA polymerase activity of Dmpol c by DmmtSSB is observed over a broad range of KCl concentrations (0 to 130 mM) [8]. We speculate that this difference in the human and Drosophila systems likely reflects the different subunit composition of the mammalian and insect pol cs. Hspol c is a heterotrimeric enzyme comprising one catalytic and two accessory subunits (ab 2 ) [32,33]; the a and dimeric b subunits can be expressed in heterologous systems and purified independently [15,18], and the holoenzyme is subsequently reconstituted in vitro. In contrast, Dmpol c has a heterodimeric composition, with one catalytic and one accessory subunit (ab) [34]; both folding and stability are interdependent, as evidenced by the fact that in a heterologous system, the two subunits must be co-expressed to reconstitute the Dmpol c holoenzyme [35]. Furthermore, subunit interactions in the insect enzyme occur at multiple sites along the polypeptides [36]. That the differences in pol c stimulation by the fly and human mtSSBs results from differences in pol c structure is supported by the fact that both DmmtSSB and EcSSB stimulate similarly Dmpol c [37]. Additional support for this hypothesis is also provided by a recent report, which shows that each protomer of the dimeric human accessory subunit serves distinct roles in DNA synthesis by Hspol c [38]. Thus, we speculate that the function of the Hspol c-b dimer in the human mtDNA replisome is at least in part performed by DmmtSSB in the Drosophila system. Further investigation of the mechanism of pol c stimulation by mtSSB is clearly warranted to promote understanding of the species-specific roles of these proteins at the mtDNA replication fork.
Our analysis of stimulation of Hspol c by the deletion variants of HsmtSSB revealed an interesting and surprising feature: the lack of the N-and/or C-terminus of HsmtSSB increases its capacity to stimulate Hspol c under low ionic strength conditions. This suggests that both termini may actually modulate the DNA polymerase activity of Hspol c by inhibiting its stimulation by HsmtSSBwt. Whether this modulation is mediated through physical or functional interactions only remains to be determined, but it is clear that the relevant interactions do not involve a positive regulation, as is the case for EcSSB, T7 gp2.5 and T4 gp32, and their respective DNA polymerase partners at the replication fork. In considering the electrostatic forces that may govern functional interactions, we examined the predicted isoelectric points (pIs) for various structural elements in Hspol ca and wild-type HsmtSSB. We were especially interested in a fragment of the spacer region domain of Hspol c-a assigned as the intrinsic processivity (IP) sub-domain in the recent crystal structure [33], because earlier studies from our lab suggested that residues in this sub-domain are important for the functional interaction between Drosophila pol c and its cognate mtSSB [39]. Consistent with the hypothesis that the functional interactions between Hspol c and HsmtSSB are electrostatic, we found a pI of 8.2 for wild-type HsmtSSB, and 5.7 for the IP sub-domain of Hspol c-a, suggesting that increasing salt in the pol assays disrupts the electrostatic forces that allow Hspol c to displace HsmtSSB from ssDNA. The only other domain of Hspol c-a that carries an overall negative charge is the accessory-interacting domain (pI 4.7), which interacts tightly with the proximal protomer of the pol c-b dimer. In examining the terminal deletion variants of HsmtSSB we found pIs of 9.0 (DC), 9.2 (DN) and 9.5 (DNDC), suggesting that lack of either or both termini causes a significant increase in the overall positive charges of the protein that may strengthen its interactions with Hspol c, stimulating DNA synthesis and allowing Hspol c to displace it from the ssDNA. At the same time, the ssDNA-binding affinities of the variant HsmtSSBs are unaffected over the range of 20-100 mM KCl ( Fig. 4 and data not shown). As the HsmtSSB becomes more positively charged without its termini, the electrostatic forces between Hspol c and HsmtSSB likely increase at 20 mM KCl, increasing the stimulation and the ability of Hspol c to displace the HsmtSSB variants more easily. However, whereas ssDNA binding by HsmtSSB is stable at 50 mM KCl, this salt concentration is apparently sufficient to disturb interactions between Hspol c and both the wild-type and variant forms of HsmtSSB, giving rise to its inhibitory effects. Although the effects we observe are modest, it seems possible that given the fluctuating ionic conditions that occur in the mitochondrion [40,41], they may play a role in regulating the interactions between Hspol c and HsmtSSB in vivo.
In contrast with Hspol c, the stimulation of HsmtDNA helicase by HsmtSSB proteins is not salt-dependent, and the deletion variants of HsmtSSB do not show a higher stimulatory effect than HsmtSSBwt on the DNA unwinding activity of HsmtDNA helicase. Taking into consideration that EcSSB may not stimulate HsmtDNA helicase as well as its cognate SSB [9], this result argues that the N-and C-terminal regions of HsmtSSB are not involved directly in the 8-fold stimulation of HsmtDNA helicase. However, it has been shown that EcSSB can replace T7 gp2.5 in vitro to stimulate the T7 DNA polymerase holoenzyme activity on a singly-primed ssDNA template and in strand-displacement assays (the latter also involves the function of T7 gp4 primase-helicase), but it fails to promote either coupled leading and lagging strand synthesis in vitro, or the growth of bacteriophage T7 mutants lacking the gp 2.5 gene, both of which require the coordinated function of the T7 replisome [42,43]. Therefore, our assays may be limited in assessing biochemically the possible defects of HsmtSSB variants. In any case, one might argue that the mtDNA replication fork most likely comprises other unidentified components in addition to pol c, mtDNA helicase and mtSSB, especially given the complexity of the myriad processes that occur in the mitochondrion [44], and the various modes of mtDNA replication that operate in vivo [45,46,47], which ensure appropriate mtDNA copy number and mitochondrial gene expression. Thus, physiological analysis of these and other mtSSB mutants, in addition to development of new in vitro assays that reconstitute fully the mtDNA replisome, will be informative in understanding the mechanism of mtDNA replication.

Mutagenesis and purification of HsmtSSB proteins
The HsmtSSB deletion variants were constructed by cloning of PCR fragments containing the coding region for HsmtSSBDN, DC and DNDC into the NdeI site of the pET11a vector. PCRs were performed using the coding region of the mature HsmtSSBwt cloned in pET11a vector as template, Pfu DNA polymerase (Stratagene) and standard laboratory methods. The oligonucleotides used for PCR mutagenesis were: 59CCCGGGCATatgCTT-GAAAGATCCCTGAATCG39 and 59CCCGGGCATATGCT-ACTCCTTCTCTTTCGTCTGG39 for HsmtSSBDN; 59CCCG-GGCATATGGAGTCCGAAACAACTACCAG39 and 59CCC-GGGCATATGctaACTCAGAAATATAATATTATCAG39 for HsmtSSBDC; and 59CCCGGGCATatgCTTGAAAGATCCC-TGAATCG39 and 59CCCGGGCATATGctaACTCAGAAAT-ATAATATTATCAG39 for HsmtSSBDNDC. The underlined sequences correspond to NdeI restriction sites, and the lower case letters indicate the sites where mutations were introduced into the HsmtSSB cDNA to create new start and stop codons. BL21(DE3) cells containing pET-11a plasmid expressing HsmtSSBwt and deletion variants were grown at 37uC with aeration in L-broth containing 0.1 mg/mL of ampicillin. When the bacterial cell culture reached an optical density of 0.6 at 595 nm, isopropyl b-D-1-thiogalactopyranoside was added to 0.2 mM, and the culture was incubated further for 3 hr. Cells were harvested by centrifugation, washed in 50 mM Tris-HCl, pH 7.5, 10% sucrose (Tris-sucrose), frozen in liquid nitrogen, and stored at -80uC. All further steps were performed at 0-4uC, and all buffers contained 5 mM DTT, 1 mM phenylmethylsulfonyl fluoride, 10 mM sodium metabisulfite, and 2 ug/mL leupeptin. The cell pellet was thawed on ice, and cells were resuspended in 1/25 volume of original cell culture in Tris-sucrose and lysed by addition of 5 X lysis buffer (1 M NaCl, 10 mM EDTA, and 10% sodium cholatefor HsmtSSBDC purification, 7.5% n-dodecyl-b-D-maltoside instead of 10% sodium cholate) to a final 1 X concentration, followed by incubation for 30 min on ice and freezing in liquid nitrogen. After thawing on ice, the suspension was centrifuged at 175006g for 30 min. The supernatant (soluble Fr I) was loaded onto a Blue Sepharose column equilibrated with 10 column volumes (CV) of 35 mM Tris-HCl, pH 7.5, 10% glycerol, 2 mM EDTA, 0.2 M NaCl at a packing ratio of 5-7 mg of total protein per mL of resin. The column was washed with 1 CV of equilibration buffer and 3 CV of 35 mM Tris-HCl, pH 7.5, 10% glycerol, 2 mM EDTA, 0.25 M NaSCN. The bound protein was eluted with 8 CV of a 0.4-1.2 M NaSCN linear gradient, followed by a final elution step of 1.5 M NaSCN buffer (2 CV). Fractions containing HsmtSSB were pooled (Fr II) and dialysed against buffer containing 60 mM KPO 4 , pH 7.6, 10% glycerol, 2 mM EDTA (Fr IIb) (for HsmtSSBDC, dialysis buffer contained 40 mM KPO 4 , pH 7.6). Fr IIb was then loaded onto a phosphocellulose column equilibrated with dialysis buffer at a packing ratio of 0.5 mg of total protein per mL of resin. The column was washed with 2.5 CV of the same buffer and the proteins were eluted with 5 CV of a 60-150 mM KPO 4 linear gradient, followed by a final step of 350 mM KPO 4 buffer (2 CV). The HsmtSSB proteins typically elute at ,80 mM KPO 4 . Pooled fractions (Fr III) were concentrated to ,1 mg/mL of protein in a Centricon-30 spin concentrator (Amicon) treated with 5% Tween 20 (Fr IIIb). Fr IIIb was frozen in liquid nitrogen and stored at 280uC.

Purification of Hspol c and HsmtDNA helicase
Recombinant human pol c-a exoand pol c-b were prepared from Sf9 and bacterial cells, respectively, as described by Oliveira and Kaguni [15]. Recombinant human mtDNA helicase was prepared from Sf9 cells, as described by Ziebarth et al. [48].

Glycerol gradient sedimentation and gel filtration
HsmtSSBwt and variants (100 mg) were layered onto preformed 12-30% glycerol gradients (10 mL) containing 35 mM Tris-HCl, pH 7.5, 100 mM NaCl, 2 mM EDTA, 5 mM DTT, 1 mM phenylmethylsulfonyl fluoride, 10 mM sodium metabisulfite, and 2 ug/mL leupeptin. Centrifugation was at 264,0006g for 63 hrs at 4uC in a Beckman SW41 rotor. Fractions were analyzed by SDS-PAGE and Coomassie blue staining. For Superdex 75 gel filtration, 200 mg HsmtSSBwt and variants were chromatographed on a column equilibrated with the buffer described above containing 8% glycerol at a flow rate of 0.25 mL/min at 4uC. Fractions were analyzed by SDS-PAGE and Coomassie blue staining to confirm UV trace recordings. Standard protein markers used in both procedures were: bovine serum albumine (BSA, 4.85 S, R S = 3.61 nm), carbonic anhydrase (CA, 3.23 S, R S = 2.43 nm) and lysozyme (LYS, 1.91 S, R S = 1.5 nm). The data were plotted as S value versus fraction number to obtain a sedimentation coefficient, and as R S value versus the peak of protein elution in mL to obtain the Stokes radii for the HsmtSSB proteins. The native molecular mass of the proteins were calculated using the formula: MW = 3.909 x S value x R S .

ssDNA binding and gel mobility shift assay
Reaction mixtures (20 mL) contained 20 mM Tris-HCl, pH 7.5, 1 mM DTT, 4 mM MgCl 2 , 50 mM NaCl, 36 fmol 59-end-labeled 48-mer, and the indicated amounts of the HsmtSSB proteins. Incubation was at 20uC for 10 min. Samples were processed and electrophoresed in 6% native polyacrylamide gels. The amounts of shifted and free oligonucleotide were quantitated as follows: % ssDNA bound = (V S /(V S +V F )) 6100, where V S represents the volume of the shifted and V F the volume of unshifted oligonucleotide in the sample lane of interest.

DNA unwinding assays
Reaction mixtures (50 mL) contained 20 mM Tris-HCl, pH 7.5, 10% glycerol, 500 mg/mL bovine serum albumin, 10 mM DTT, 4 mM MgCl 2 , 3 mM ATP, 50 mM KCl (unless stated otherwise), 0.4 nM of DNA unwinding substrate, 3.5 nM of mtDNA helicase (as hexamer), and the indicated concentrations of HsmtSSB proteins. The reactions were pre-incubated at 37uC for 10 min prior to the addition of the helicase. Once the helicase was added, the reactions were incubated further at 37uC for 30 min and then stopped by the addition of 5 mL of 10 X stop solution (6% SDS, 100 mM EDTA, pH 8.0), followed by 5 mL of 10 X loading buffer (50% glycerol, 0.25% bromophenol blue). DNA products were fractionated from substrate by electrophoresis in a 22% polyacrylamide gel (59:1 acrylamide/bisacrylamide) using 1 X TBE (90 mM Tris-HCl-borate, 2 mM EDTA) at 600 V for ,30 min. After electrophoresis, the gel was dried under vacuum with heat, and exposed to a Phosphor Screen (Amersham Biosciences). The data were analyzed by scanning the Phosphor Screen using a Storm 820 Scanner (Amersham Biosciences), and the volume of each band were determined, and background subtracted, by computer integration analysis using ImageQuant version 5.2 software (Amersham Biosciences). For all reactions, DNA unwinding is defined as the fraction of radiolabeled DNA species that is single-stranded (product), as follows: % unwinding = (V P / (V S + V P )) 6100, where V P represents the volume of the product and V S the volume of unreacted substrate in the sample lane of interest.