A Genetic Analysis of the Functional Interactions within Mycobacterium tuberculosis Single-Stranded DNA Binding Protein

Single-stranded DNA binding proteins (SSBs) are vital in all organisms. SSBs of Escherichia coli (EcoSSB) and Mycobacterium tuberculosis (MtuSSB) are homotetrameric. The N-terminal domains (NTD) of these SSBs (responsible for their tetramerization and DNA binding) are structurally well defined. However, their C-terminal domains (CTD) possess undefined structures. EcoSSB NTD consists of β1-β1′-β2-β3-α-β4-β451-β452-β5 secondary structure elements. MtuSSB NTD includes an additional β-strand (β6) forming a novel hook-like structure. Recently, we observed that MtuSSB complemented an E. coli Δssb strain. However, a chimeric SSB (mβ4-β5), wherein only the terminal part of NTD (β4-β5 region possessing L45 loop) of EcoSSB was substituted with that from MtuSSB, failed to function in E. coli in spite of its normal DNA binding and oligomerization properties. Here, we designed new chimeras by transplanting selected regions of MtuSSB into EcoSSB to understand the functional significance of the various secondary structure elements within SSB. All chimeric SSBs formed homotetramers and showed normal DNA binding. The mβ4-β6 construct obtained by substitution of the region downstream of β5 in mβ4-β5 SSB with the corresponding region (β6) of MtuSSB complemented the E. coli strain indicating a functional interaction between the L45 loop and the β6 strand of MtuSSB.


Introduction
Single-stranded DNA binding protein (SSB) binds singlestranded DNA in a sequence independent manner during major DNA transactions such as DNA replication, repair and recombination [1][2][3][4][5]. Besides their crucial function in DNA transactions, they protect transiently generated single-stranded DNA (ssDNA) from nucleases or chemical attacks [6]. The eubacterial SSBs contain subunits with a similar basic fold, but may exhibit variations in their quaternary association [7]. SSBs possess an oligonucleotide-binding fold (OB-fold) in the N-terminal domain responsible for their oligomerization and DNA binding. The conserved C-terminal acidic tail of SSBs is important in proteinprotein interactions [8][9][10][11]. One of the features of EcoSSB, important for its in vivo function, is the dynamic transition in its modes of DNA binding [6,12]. SSB binds to ,35 nucleotides by two of its subunits known as SSB 35 mode and is required for unlimited cooperatively. While all the four subunits bind to ,56 or ,65 nucleotides in a limited cooperative manner known as SSB 56 or SSB 65 modes, respectively [13][14][15][16].
The crystal structures of SSB in free and DNA bound forms have provided valuable information to understand their function [17,18]. EcoSSB monomer consists of an N-terminal domain (,115 amino acids) of defined structure, and the C-terminal domain whose three dimension structure is not available. The tertiary structure of the N-terminal domain of EcoSSB is defined by the presence of b1-b19-b2-b3-a-b4-b45 1 -b45 2 -b5 secondary structure elements (Fig. 1). In the X-ray crystal structure, one of the b hairpin loops (L 45 ) with well-defined electron density connects b4 and b5. Structural studies of EcoSSB suggested that its quaternary association is mediated by the L 45 loops as well as by the six-stranded b-sheets formed by the dimers [17]. Furthermore, the L 45 loop undergoes a significant change upon binding to DNA [18]. Functional importance of this movement, however, remains unclear.
MtuSSB shares ,30% identity and ,39% similarity with EcoSSB in its primary sequence. The secondary structure involved in OB-fold is very similar in the two SSBs except for the presence of a novel b6 strand (numbered according to EcoSSB, 17) downstream of the b5 in MtuSSB (Fig. 1). While both the SSBs share overlapping tertiary structures, there are notable variation in their quaternary associations due to the presence of the b6 strand in MtuSSB [19]. Although a role for b6 strand in providing stability through the formation of a clamp like structure has been suggested in the mycobacterial SSBs [19][20][21] its biological importance is unknown.
Recently, using an in vivo assay wherein replication of the resident ssb support plasmid in an E. coli strain deleted for its chromosomal copy of ssb gene could be selectively blocked, we showed that overexpression of MtuSSB complemented E. coli [22]. However, a chimeric SSB (mb4-b5), wherein the b4-b5 region (which possess the L 45 loop) of EcoSSB was replaced with the corresponding secondary structure elements of MtuSSB, did not complement the strain [22]. This suggested that the L 45 loop might be involved in specific interactions within MtuSSB. In this study, we have designed additional chimeric constructs to uncover the importance of such interactions between the MtuSSB L 45 loop and the novel b6 strand for its function in E. coli.

Complementation analysis
The complementation assays were performed using a recently described revised plasmid bumping method [22]. Briefly, the pBAD based expression constructs were introduced into E. coli RDP317-1 harboring pHYDEcoSSB as support plasmid (ColE1 ori, Cam R , whose replication is dependent on the presence of isopropyl-b-D-thiogalactopyranoside, IPTG) and the transformants were selected on LB agar containing Kan, Amp and 0.02% arabinose (or Kan, Amp and 0.5 mM IPTG, as control). The isolated colonies were streaked on LB agar containing Kan and Amp with various concentration of arabinose.

Growth curve analysis
Freshly isolated transformants were inoculated in LB containing Kan, Amp and 0.02% arabinose to obtain late stationary phase cultures; and inoculated at 0.1% level in LB containing Kan, Amp and arabinose (as indicated) in the honeycomb plates. The growth was recorded at 600 nm using Bioscreen C growth reader (OY growth, Finland) at 37uC on an hourly basis. Average values (6SEM) were plotted.

Microscopic studies
Freshly isolated transformants of E. coli Dssb strain harboring pBAD based SSB constructs were grown to log phase (7-9 h in 2 ml LB containing arabinose). Bacterial cells were collected by centrifugation, fixed with 4% paraformaldehyde, kept on poly-Llysine treated multi-well slide, washed with PBS and visualized in fluorescence microscope (ZEISS, Axio Imager) with a 1006 objective lens [22].

Experimental rationale and generation of SSB chimeras
The N-terminal domain of EcoSSB is defined by b1-b19-b2-b3a-b4-b45 1 -b45 2 -b5 as its secondary structure elements (Fig. 1A). The N-terminal domain of MtuSSB, in addition possesses a b6 strand ( Fig. 1), which causes a notable variation in its quaternary structure by the formation of a clamp like structure at the dimeric interface of the interacting subunits [19]. The C-terminal domains of both the SSBs possess acidic tails important in protein-protein interactions during various DNA transactions [8][9][10][11].
Recently, we observed that MtuSSB sustained E. coli for its essential function of SSB [22]. However, the mb4-b5 SSB, EcoSSB-NheI-Rp ccctgacgaccagctagcatctgcatg [22] EcoSSB-Fp ggaattcaccatggccagcagagg [22] EcoSSB-XmaI-Fp agcgaatatctggcccggggttctcaggtt This work MtuSSB-NheI-FP ttgggccttcgctagcgtacgccaccgc [22] MtuSSB-NheI-Rp gcggtggcgtacgctagcgaaggcccaa [22] pTrc-Bcl-Rp ggctgttttggcggatgagaga [22] pTrc-Fp taacaagcttacacaggaaacag [22]  wherein amino acids 74 to 111 (comprising b4, b45 1 , b45 2 and b5 strands) were replaced with the corresponding region of MtuSSB, failed to sustain E. coli despite its normal oligomerization and DNA binding properties. Another chimera, mb1-b5 wherein the b1-b5 elements of EcoSSB were replaced with the corresponding elements of MtuSSB, conferred filamentation phenotype to E. coli. However, the mb1-b6 SSB with the entire N-terminal domain of MtuSSB (i. e. including the b6 strand) fused to the C-terminal domain of EcoSSB, functioned well in E. coli [22]. These observations suggested specific interaction of b4-b5 region of MtuSSB with the b6 region of MtuSSB. To study the functional importance of such an interaction and to further our understanding of the structure-function relationship of eubacterial SSBs, we generated additional chimeric SSBs (Fig. 2). The mb4-b5 SSB was modified to generate mb4-b5 (acidic), and mb4-b6 SSBs. One of the distinctive features of the region between the b4 and the b5 strands of MtuSSB is that, unlike EcoSSB, it possesses a number of acidic residues (Fig. 1A). Hence, these residues were changed to EcoSSB specific sequences in a chimera designated mb4-b5 (acidic) by mutating E 90 , T 91 , E 95 , K 96 , E 103 , D 105 , and E 106 within MtuSSB region of b4-b5 to T 90 , D 91 , Q 95 , D 96 , V 103 , N 105 and V 106 , respectively. To generate mb4-b6, MtuSSB sequence corresponding to amino acids 74-111 in mb4-b5 was extended to 131 to include b6 of MtuSSB. Among other constructs, mb1-a contained the first 73 amino acids (consisting of b1-a structural elements) from MtuSSB and the amino acid 74 to the end from EcoSSB. In mb6 SSB, the b6 strand and the downstream spacer sequences of MtuSSB (amino acid 114 to 133) substituted the corresponding region of EcoSSB. The remainder of the sequences (the N-terminal region consisting of the first to 113 amino acids and the C-terminal region (amino acids 134 to the end) were from EcoSSB. The mb6-CTD contains the b6 strand and the C-terminal region (amino acid number 114 to the end) from MtuSSB whereas, the N-terminal region (the first 113 amino acids) from EcoSSB. Lastly, the mCTD construct contains only the C-terminal region from MtuSSB (amino acid number 129 to the end) and the N-terminal and the spacer sequences (first 128 amino acids) of EcoSSB. More details of generation of these constructs are provided in Methods S1 and Table S1.

Oligomerization of the chimeric SSBs
All SSBs were purified and analyzed by gel filtration chromatography to determine their oligomerization status (Fig. 3). Elution profile of the chimeric SSBs was very similar to those of the wildtype EcoSSB and MtuSSB suggesting that they folded properly and formed homotetramers.

DNA binding properties
To demonstrate the DNA binding abilities of various SSB constructs, we performed electrophoretic mobility shift assays (EMSA) using 32 P labeled 79mer DNA. Using this assay (Fig. 4), EcoSSB and MtuSSB form a faster migrating complex under limiting SSB concentration (Complex I). As the concentration of SSB increases, a second slower migrating band (Complex II) appears. Based on their mobility, these complexes potentially correspond to the SSB 56/65 and SSB 35 modes of DNA binding, respectively. More importantly, within the detection limits of this assay, all the chimeric SSBs reveal DNA binding similar to the parent SSBs (compare panels 4A and 4B with 4C to 4H), suggesting that the quaternary structures of the chimeric SSBs are largely unaffected by the mutational manipulations performed to generate them.

Functionality of SSB chimeras in E. coli
Recently, we described a sensitive assay to assess the functionality of a test SSB using a modification of the original 'plasmid bumping method' [22,25]. In the revised assay, the test ssb construct (on a ColE1 ori plasmid, Amp R ) is introduced in a Dssb (ssb::kan) strain of E. coli (RDP317-1, Kan R ) harboring a plasmid borne support of wild-type ssb on another ColE1 ori plasmid, pHYDEcoSSB (Cam R ). The replication of pHYDEcoSSB is dependent on the presence of IPTG. Hence, withdrawal of IPTG from the growth medium results in the loss of the support plasmid (pHYDEcoSSB) and failure of the strain growth unless sustained by the test SSB. Growth of the original transformants of the test ssb plasmid on plate lacking IPTG, together with the loss of Cam R phenotype, suggests that the test ssb complemented the Dssb strain of E. coli for its function of SSB. An advantage of this assay is that the in vivo activity of even a weakly functioning SSB can be assessed (fitness disadvantage of the test ssb, if any, is avoided by selectively blocking replication of the original ssb support plasmid).
Using this method, we checked the in vivo activity of various SSB constructs subcloned into a ColE1 ori (Amp R ) plasmid wherein their expression was inducible by arabinose (the pBAD series of constructs, Table 1). As shown in Fig. 5A, all constructs showed expression of the corresponding SSBs in E. coli TG1. Subsequently, to check for their in vivo function, the ssb constructs were introduced into RDP317-1 strain (Kan R ) harboring pHYDEcoSSB (Cam R ), and the transformants were selected on Kan, Amp and 0.02% arabinose plates either containing or lacking IPTG. An analysis of the plating efficiencies (obtained from the ratios of transformants on the -IPTG to +IPTG plates) is shown in Table 2. The mb4-b5(acidic) SSB did not complement the Dssb strain of E. coli suggesting that conversion of mb4-b5 SSB to mb4-b5(acidic) SSB does not make it functional in E. coli. However, transplantation of the b6 region of the MtuSSB into the mb4-b5 construct in mb4-b6, resulted in an efficient rescue of the Dssb strain of E. coli suggesting a functional interaction between the b4-b5 and the b6 regions of MtuSSB. Interestingly, substitution of the unstructured region of EcoSSB downstream of its b5 region with the b6 region of MtuSSB in mb6 SSB, maintained its activity suggesting that the b4-b5 region of EcoSSB is tolerant of its downstream sequences.
In vivo complementation by various SSB constructs was further validated by streaking of the freshly obtained transformants (Fig. 5B) on plates containing either IPTG (as control) or varying concentrations of the inducer (0.002-0.2% arabinose). As expected from the replication of the pHYDEcoSSB support plasmid in the presence of IPTG, all transformants showed growth on the +IPTG plate. Like the vector control (sector 1), neither the mb4-b5 nor the mb4-b5 (acidic) complemented the Dssb strain at any of the arabinose concentrations (sectors 4 and 5). Further, the results of the growth curve analyses (Fig. 6) of the strains harboring SSBs that sustained E. coli are also consistent with the plating efficiency data. Weakly functioning SSBs, in general, resulted in longer lag phases when expression of SSBs was induced with 0.002% arabinose (panel ii). These differences were, however, lost in cultures induced with 0.02% or 0.2% arabinose (panels iii and iv) which result in higher level of expression of these SSBs (Fig. S1). As a control, when the growth curve analyses were carried out in the absence of inducer, arabinose (Fig. 6, panel i) none of the cultures grew confirming that the phenotypes observed in Table 2, and Figs. 5B and 6 (panels ii and iii) are due to the plasmid borne SSBs. The longer lag phases in Fig. 6 (panel ii) could be a stress related phenomenon. Interestingly, we observed that the weakly functioning SSBs also conferred temperature and cold sensitive phenotypes to E. coli for growth at 42uC and 30uC, respectively (Fig. 7). These phenotypes could also be suppressed upon induction of SSB expression with higher concentrations of arabinose. It may also be noted that even under these conditions (temperatures of 42uC or 30uC), the mb4-b5 and mb4-b5 (acidic) failed to complement the E. coliDssb strain (Figs. 7A and 7B, sectors 4 and 5, respectively).

Microscopic analyses
In our earlier study microscopic analyses of the fixed E. coli cells revealed that the mb1-b5 SSB, a poorly functioning SSB, resulted in a notable filamentation phenotype [22]. On the other hand, SSBs that functioned, but not as well as EcoSSB, resulted in a slightly elongated cell phenotype. As before, MtuSSB showed a phenotype of slightly elongated cells (Fig. 8, compare panels d and  a). However, the mb4-b6 SSB showed a more pronounced phenotype of the elongated cells (compare panel j with a). The mb1-a SSB showed a weak phenotype of the elongated cells (compare panels m with a). Interestingly, as in Figs. 6 and 7, overexpression of the SSBs suppressed these phenotypes (compare panels d with e and f; j with k and l; m with n and o).

Discussion
Determination of the three-dimensional structure of MtuSSB by X-ray crystallography revealed that while its structure at the tertiary level is very similar to that of EcoSSB, it shows significant variations at the level of quaternary interactions [19]. A notable difference seen at the level of tetramerization of MtuSSB is the presence of a clamp like structure formed by the b6 strand of the mycobacterial SSB [19]. However, it has so far remained unclear as to what the biological significance of this unique structural element of MtuSSB is.
The L 45 loop in EcoSSB has been shown to undergo a conformational change upon DNA binding and suggested to be important for its cooperative binding [17,18]. In addition, the computational analyses suggested that the movements of L 45 loop in EcoSSB, MtuSSB, and Streptomyces coelicolor SSB are different [21]. Our observation shows that the mb4-b5 construct wherein the L 45 loop (of MtuSSB origin) is intact does not function in E. coli but the mb4-b6 SSB wherein a small region (b6) downstream of b5 was also included, does. Together with the biophysical and computational analyses [17,18,19,21], these observations highlight the importance of the functional interactions of the L 45 loop with the b6 region. And, some of these interactions may well contribute to the stability of the MtuSSB tetramer predicted from the crystal structure analysis [19]. However, it should also be said that our present study does not allow us to comment on the mechanistic details of such interactions for the SSB function in vivo.
How crucial is the species specificity of these interactions (in the context of SSB tetramer) for SSB function? When we changed this region of EcoSSB with the corresponding region of MtuSSB in the context of E. coli L 45 loop, we did not detect a significant defect in the chimeric SSB (mb6), suggesting that the interactions of the L 45 loop with its downstream sequence are more tolerant in EcoSSB. In the context of M. tuberculosis L 45 , when the entire upstream region of MtuSSB was provided, such as in the mb1-b5 SSB i. e., wherein the N-terminal domain (b1-b5) of EcoSSB was replaced with the corresponding sequence from MtuSSB, it did sustain E. coli viability but the growth was poor and it resulted in a filamentation phenotype [22]. These observations suggest that the context of both the upstream and the downstream regions (with respect to the L 45 loop of MtuSSB) is biologically significant. Lack of either of the regions compromises SSB function in a context dependent manner. However, the chimeras mb1-a and mb6-CTD, wherein the entire region upstream of, or downstream of the loop L 45 (of EcoSSB), respectively are from MtuSSB, functioned well in E. coli (as did the mb6). These observations suggest that in EcoSSB, any interactions mediated by the L 45 are more tolerant of the neighboring sequences. This is further indicated by the observation (Fig. 6, panel ii) that the construct mCTD (EcoSSB harboring only the CTD from MtuSSB) functioned nearly as well as the mb6 (harboring only the b6 of MtuSSB) or the mb6-CTD (harboring the entire region downstream of L 45 , from MtuSSB). An availability of the three-dimensional structures of the chimeric SSBs may further our understanding of the interactions L 45 establishes within SSB.  (2) or supplemented (+) with 0.02% arabinose, and grown further for 3 h. Cells were harvested and processed as described [22]. Cell-free extracts (,10 mg total protein) were resolved on SDS-PAGE (15%). (B) Transformants of E. coli RDP 317 harboring chimeric SSBs obtained in the presence of IPTG were suspended in LB and streaked on LB-agar (Kan, Amp) containing IPTG or arabinose (0.002-0.2%) and incubated at 37uC for ,12 h. Sectors: 1, pBAD vector; 2, pBADEcoSSB; 3, pBADMtuSSB; 4, pBADmb4-b5; 5, pBADmb4-b5(acidic); 6, pBADmb4-b6; 7, pBADmb1-a; 8, pBADmb6; 9, pBADmb6-CTD;10, pBADmCTD. doi:10.1371/journal.pone.0094669.g005   Finally, the modification [22] of the 'plasmid bumping' assay [25] we recently developed has been useful in determining the efficacy of SSB mutants in sustaining E. coli even when they are compromised in their function, and provided with a convenient approach to study the structure-function relationship of the various structural elements of the eubacterial SSBs.