Structural studies on the M. tuberculosis nucleoid-associated-protein, NapA, indicates DNA bridging mechanism

DNA binding analyses of NapA

Multiple sequence alignments of NapA homologs revealed that the N-terminal ~15 residues of these proteins are not conserved and, in fact, several NapA proteins lack this region altogether (Fig. 1A). Given that this sequence is not conserved, we generated an N-terminally truncated Mtb NapA construct, NapA(16-102), for structural and biochemical studies. Previous analyses indicated that NapA binds double stranded DNA (dsDNA) sites, 30 and 60 nucleotides in length and that it shows a preference for AT-rich DNA [37]. As the minimal DNA sequence is often the most readily crystallized when in complex with a DNA binding protein, we performed fluorescence polarization (FP) binding assays with 16mer, 20mer and 40mer dsDNA sites, to delineate the shortest DNA site bound with high affinity by NapA.

The FP binding assays revealed that Mtb NapA(16-102) bound 16mer, 20mer and 40mer DNA sites with apparent K_ds of 868 ± 99 nM, 170 ± 26 nM and 159 ± 19 nM, respectively (Fig. 1B). As these experiments were done with NapA(16-102), we used the FL NapA homolog from Mycobacterium sp. MBV951389.1, as a control to assess if the removed NapA N-terminal residues (Fig. 1A) are important for DNA binding. The Mycobacterium sp. MBV951389 NapA protein, herein termed NapA-H, bound the DNA sites with essentially the same affinities as NapA(16-102), with apparent K_ds of 780 ± 98 nM, 162 ± 14 nM and 211 ± 24 nM, for the 16mer, 20mer and 40mer, respectively (Supplementary Fig. S1). These data indicate that the N-terminal, nonconserved region in the Mtb protein is not essential for DNA binding and that NapA binds DNA sites with 16 base pairs (bps) with ~5-fold reduced affinity compared to a 20mer DNA site, while the 40mer DNA bound NapA with essentially the same affinity as the 20 bp site.

Structure of M. tuberculosis NapA

NapA shows no sequence homology to any characterized protein. To gain insight into NapA function we sought to obtain crystals. Apo crystals of Mtb NapA(16-102) or the NapA-H could not be produced. Hence, we generated C-terminally truncated constructs of Mtb NapA, as the C-terminal residues appeared to show low conservation and were predicted to be largely disordered (Fig. 1A). We obtained crystals of the Mtb NapA(16-92) protein and collected X-ray intensity data to 2.4 Å resolution. Available AI-based methods (AlphaFold2 or RoseTTAFold) or large language-based (ESMFold) predicted models of Mtb NapA failed to produce solutions in Molecular Replacement (MR) trials [45-46]. Thus, we purified selenomethionine substituted NapA(16-92) for phasing. These data were used to obtain a structure of apo NapA, which was refined to final R_work/R_free of 22.8%/24.4% to 2.4 Å (Fig. 2A-B; Table S1; Methods).

The NapA(16-92) structure revealed two NapA molecules in the crystallographic asymmetric unit (ASU), which combine to form a dimer. Density was present for residues 20-90 of one subunit and 20-91 of the other. Strikingly, crystallographic symmetry generates a NapA dimer-of-dimers from the dimer (Fig. 2B). Each NapA subunit harbors the topology (α1: residues 23-50, α2: 54-65, α3: 68-73, α4: 75-87) (Fig. 2B). A characteristic feature of the structure is the presence of the extended N-terminal helix, α1, that interacts with α1´ (where ´ indicates other subunit of the dimer) to form the dimer. In the dimer, the α2, α3 and α4 of the other subunit of the dimer make interactions with the N-terminal region of α1´ (Fig. 2B). The NapA dimer combines with another dimer to form the dimer-of-dimers with a central four helix bundle scaffold (Fig. 2B). To investigate if there are any structures with similarity to the NapA structure in the protein data bank, we performed structural similarity searches with SSM/Pdbfold [49]. These analyses did not uncover any structures with significantly similar folds to either the NapA monomer, dimer or dimer-of-dimers. The highest scoring structural similarities were limited to a few helices of NapA, which amounted to ~34 residues, with root mean square deviations (rmsds) of >3 Å for corresponding Cα atoms.

Structure of M. tuberculosis NapA-DNA complex

To gain insight into how NapA interacts with DNA we next carried out crystallization trials of NapA(16-102) and NapA-H with DNA sites ranging from 16 base pairs (bp) to 60 bp (Methods). Well diffracting crystals were not obtained with the NapA-H protein and DNA substrates, but crystals that diffracted to beyond 4.0 Å resolution were produced with NapA(16-102) in complex with the self-complementary 20mer DNA duplex, 5´-ATTATTAATTAATTAATAAT-3´ (Methods). X-ray intensity data to 3.87 Å resolution was collected and the structure was solved (Fig. 3A) by MR using our structure of the apo NapA dimer as a search model.

The NapA(16-102)-DNA structure contains one NapA dimer, which includes residues 21-84 of one subunit and 22-89 of the other subunit. Notably, crystallographic symmetry generates the same dimer-of-dimers that we observed in the apo NapA structure. The structure also contains density for two DNA molecules. Crystallographic symmetry generates duplex DNAs from these two DNA sites, which packed pseudo-continuously in the crystal (Fig. 3A). The density for one DNA site (labeled DNA1 in Fig. 3A) is well defined whereas the other (DNA2) is not as well resolved, but still evident. It should be noted that at this resolution the structure does not provide detailed information regarding protein residue-DNA contacts. However, it does reveals the position of the DNA relative to the protein, and hence shows which regions of NapA are poised to contact the DNA. In particular, the structure shows there are two regions that are adjacent to the DNA. Notably, these same regions are positioned next to the DNA backbone from each NapA dimer (Fig. 3B). These regions are comprised of residues from the loop between α1 and α2 and the N-terminus of α2, (herein called loop N element) and basic residues from α4´ (from the other subunit). Each NapA dimer contacts two DNA sites and consequently the NapA dimer-of-dimers contacts four DNA sites (Fig. 3B). Thus, the structure reveals a structural mechanism for DNA bridging.

Comparison of the NapA structure to the models predicted using RoseTTAfold, AlphaFold multimer and ESMFold [45-46] revealed similarities in the last three helices but not α1 (Supplementary Fig. S2A-D). Alphafold3 [47-48] was released after we completed the study and we then used it to predict a NapA dimer and dimer-of-dimers bound to two DNA sites. The predicted models were different from the crystal structure and showed individual NapA monomers binding along the DNA (Supplementary Fig. S3). As noted, the structure shows that NapA interacts solely with the DNA backbone. The finding that NapA only contacts the DNA backbone is consistent with what has been observed in other DNA complexes with NAPs that bind DNA nonsequence specifically [4-6]. Residues that are proximal to the DNA in the structure are from the two DNA binding elements. These include Arg56 and Thr55 from the loop N element and from α4, the side chains of Lys79, Arg82 and Arg83 (Fig. 3B). The DNA binding surface of NapA is also positively charged, hence likely playing a role DNA binding (Fig. 3C).

NapA does not contain an HTH

We noted that the α3-α4 component, with α4 residues proximal to the DNA, show some similarity to the classical helix-turn-helix (HTH) DNA-binding motif, commonly observed in both prokaryotic and eukaryotic DNA-binding proteins. Comparison of the NapA α3-α4 with the classical HTH structure from the 434 repressor [50-51], however, revealed they are different; the overlay resulted in an rmsd of 3.3 Å for only 21 aligned Cα atoms and a Z-score of 2.1 (Fig. 4A). These values indicate weak structural similarity, that are insufficient to classify the NapA α3-α4 unit as a HTH. Indeed, the NapA α4 helix is rotated ~45° to the perpendicular relative to helix 3, the recognition helix, of the 434 repressor (Fig. 4A). Inclusion of the DNA in the overlays also reveals that the 434 repressor HTH and the NapA α3-α4 motif assemble differently on DNA (Fig. 4B). The recognition helix of the 434 repressor HTH inserts deep within the major groove of the target DNA and by contrast, NapA α4 residues are positioned such they could only contact the DNA backbone (Fig. 4B). The larger rotation of the recognition helix/α2 of the HTH motif relative to α1 enables its insertion into the major groove, while the orientation of α4 in NapA allows it to track along the DNA backbone.

MP analyses of the NapA oligomeric state

Previous size exclusion studies by Datta et al. supported that NapA is a dimer [49]. Our apo and DNA-bound NapA crystal structures reveal the same dimer structure and also the same dimer-of-dimers. In the structure, the NapA dimer interface is formed by contacts between α1 and α1´ and is further stabilized by interactions between α2-α3-α4 of one subunit to α1´ of the other subunit (Fig. 5A). This interface buries ~790 Ų of protein surface from the solvent, which amounts to 26% of the total surface area of the dimer. The contacts within this dimer are largely hydrophobic with a few hydrogen bonds also contributing to formation of the interface. The hydrophobic contacts are composed of stacking interactions between Trp35 with the aliphatic atoms of Lys42´, Van der Waals contacts between Ile28 and Phe50´, hydrophobic interactions between residues Phe31 and Leu29 to Leu61´ and Leu64´ and between Leu29 to Ala71´. The hydrogen bonds in the interface are between α1 residue Glu32 and the side chains of Lys42´ and Tyr57´ (Fig. 5A). Two of these dimers combine to create a dimer-of-dimers via the formation of a central four-helix bundle. The additional interfaces within the four-helix bundle bury less than 260 Ų of protein surface from solvent. One interface involves contacts between Phe38 and Val41´ and hydrogen bonds between Lys37 and Glu44´ and the other is formed by interactions between two-fold related Phe38 residues and hydrogen bonds between Gln34 and Gln34´ (Fig. 5B-C).

To use another method from those previously employed [37] to assess NapA oligomer formation and to test our structural model, we carried out mass photometry (MP) analyses [52]. MP experiments report on single molecules and can determine accurate molecular weights (MW) of proteins in the range of 40 kDa to 5 MDa [52]. Because a NapA subunit and dimer are both below 40 kDa, we generated a maltose binding protein (MBP) tagged NapA fusion (Methods). We engineered a long linker between MBP and NapA of 30 residues, which when fully extended would span ~99 Å, to prevent possible interactions between MBP and NapA. MBP was utilized for this purpose because it is monomeric and adds 42.5 kDa in molecular weight (MW) [53]; Thus a MBP-NapA(16-102) fusion protein monomer, with a linker between MBP and NapA(16-102), has an expected MW of ~60 kDa. MP analyses of the purified MBP-NapA(16-102) protein at concentrations of 50 nM revealed that most of the sample was present as a dimer (109 ± 14 kDa) (63% compared to 25% monomer) while ~8% was present as a broad peak and thus likely a mixture of higher order oligomers (180 ± 60 kDa) (Fig. 5D). To test our structural model, we mutated residues the structure indicates as key in the formation of the main dimer interface, Leu29, Lys49 and Tyr57 (Fig. 5A). This mutant, MBP-NapA(16-102)(L29E-K42A-Y57E), was purified for MP experiments. Circular dichroism (CD) experiments showed that the mutant protein was folded (Supplementary Fig. S4A-B). MP analyses on the mutant protein revealed one major peak (99%) at 53 ± 23 kDa, consistent with a monomer (Fig. 5D). Consistent with this finding, glutaraldehyde crosslinking experiments on MBP-NapA(16-102) and MBP-NapA(16-102)(L29E-K42A-Y57E) revealed that the WT formed crosslinked oligomers, while the mutant showed no crosslinked oligomers (Supplemental Fig. S4C).

NapA is an abundant protein and hence is expected to be present at relatively high concentrations in cells (high nM to μM). Thus, we next tested the WT MBP-NapA(16-102) construct at a higher concentration (120 nM). These analyses revealed no monomers in the WT protein and showed 83% dimers and 13% dimer-of-dimers, with some aggregates (Supplementary Fig. S5A). By contrast, the mutant data revealed one peak with a molecular weight consistent with a monomer (Supplementary Fig. S5B). These data indicated the NapA mutant was monomeric even at elevated concentrations. Given that the WT protein displayed an increase in the amount of dimer-of-dimers at higher concentrations we next utilized a Macrofluidics HC add-on to the MP system to assess the oligomeric states at higher concentrations by rapid dilution. For these experiments, WT MBP-NapA(16-102) at 100 μM was rapidly diluted 10,000 fold and 5,500 fold prior to placement on the coverslip (Supplementary Fig. S5C). These samples revealed WT protein with 65% dimers and 34% dimer-of-dimers for the 10,000 fold dilution and 38% dimer and 61% dimer-of-dimers for the 5,500 fold dilution (Supplementary Fig. S5C). Thus, the data indicate that elevating the concentration of NapA leads to an increase in the amount of dimers-of-dimers species.

Test of the NapA-DNA structural model

The NapA-DNA structure revealed that residues poised to contact DNA are Thr55 and Arg56 from the loop N element and Lys79, Arg82 and Arg83 from α4´ (Fig. 3B). The functional importance of these residues is suggested by multiple sequence alignments (Fig. 1A), which reveal that these residues are among the most conserved residues in NapA homologs and moreover, lie within the two main clusters of conserved residues (Fig. 1A). However, a limitation of our NapA-DNA structure is that it was obtained at low resolution, which prevents detailed delineation of protein-DNA contacts. Hence, to test the structural model, we utilized three NapA mutant proteins, NapA(16-102)(T55A-R56A), NapA(16-102)(K79A-R82A-R83A) and NapA(16-92) in DNA binding assays. The latter mutant was tested to assess the possible role of the disordered C-terminal residues in DNA binding. FP binding experiments showed that NapA(16-92) bound the 20mer AT-rich DNA site with ~3.5 fold reduced affinity compared to NapA(16-102), which indicates that these residues, while not playing a central role in DNA binding, might contribute via electrostatics or that the C-terminal region could contribute to stability or cooperativity in some way (Fig. 6A). By contrast, NapA(16-102)(T55A-R56A) showed a significant decrease in DNA binding with an apparent K_d of 2.2 +/− 0.5 μM, (>10-fold reduction in binding compared to WT). NapA(16-102)(K79A-R82A-R83A) showed even weaker, non-saturable DNA binding (Fig. 6A). These data support the NapA-DNA structural model.

Our NapA-DNA structures revealed a molecular mechanism for DNA bridging, which previously had been alluded to in AFM studies [37] (Fig. 3B). To further interrogate the ability of NapA to bridge DNA we developed an approach similar to a previously reported bridge assay [54]. This approach utilizes a 32-bp biotinylated DNA site attached to streptavidin beads which captures a DNA binding protein and then follows with the addition of a fluoresceinated DNA molecule (F-DNA) to detect bridging. As our studies had indicated that NapA(16-102)(K79A-R82A-R83A) shows significantly reduced DNA binding, this mutant would be predicted to be impaired in DNA bridging. Thus, we performed DNA bridging assays on WT NapA(16-102) and the NapA(16-102)(K79A-R82A-R83A) mutant as well as a no protein control. For these experiments, the proteins (or the buffer control) were applied to the spin tubes containing the attached biotinylated DNA and washed. Subsequently, the F-DNA was added to each tube and the samples thoroughly washed. Visualization of the retained fluoresceinated DNA indicated bridging. While the control and NapA(16-102)(K79A-R82A-R83A) mutant showed little retained F-DNA, the WT NapA(16-102) sample revealed an F-DNA signal consistent with DNA bridging (Supplementary Fig. S6A-B). We note that the reduced bridging of the mutant is not due to an effect on oligomerization as glutaraldehyde crosslinking experiments of the NapA(16-102)(K79A-R82A-R83A) mutant revealed that it forms oligomers similar to WT (Supplementary Fig. S7). These data thus are consistent with previous studies showing bridging [37] and support that the DNA-binding residues identified in the structures are involved in this function.

Analysis of NapA DNA binding selectivity

Previous selective evolution of ligand by exponential enrichment (SELEX) type experiments indicated that NapA has a preference for AT-rich DNA sites. However, EMSA studies revealed a limited, 1.7-2-fold, enhancement of DNA binding of an AT-rich site over a GC-rich site [37]. Hence, to further investigate DNA binding selectivity by NapA, we carried out single molecule studies via C-trap experiments [55]. For these analyses we utilized a Halo-tagged NapA(16-102) protein. The Halo-tag allowed labeling with Janelia Fluor^® 503 and hence visualization of labeled NapA molecules bound to a tethered DNA site. We hypothesized that if NapA binds nonspecifically, we should observe binding along the length of a biotinylated λ-DNA probe that is 48502 bp in length and contains a mixture of A,T,C and G nucleotides. These analyses revealed that indeed, the labeled Halo-tagged-NapA(16-102) bound along the entire length of the λ probe (Fig. 6B-C). Increased forces on the DNA can impact DNA structure. We found that at forces of 1 pN and a slightly elevated force of 15 pN, NapA displayed the same overall binding to the DNA whereby NapA was also observed to bind along the entire DNA. These data collectively show that NapA is able to bind nonspecifically to target DNA that contains a near equal distribution of ATGC bases (the λ DNA site has an A/T/G/C content of 25%/26%/26%/23%) (Supplementary Fig. S8).

Our initial FP analyses were done with AT-rich DNA sites (Fig. 1A). To further probe nonspecific binding by NapA, we used FP to analyze NapA(16-102) binding to a 20mer DNA site containing GC bases (top strand: 5´- CTTGTTAGCTAGCTAACAAG-3´). The apparent K_d of NapA(16-102) for this DNA site was 196 ±40 nM (Supplementary Fig. S9). Thus, the observed K_d values for this site compared to the AT-rich DNA site reveals only a ~1.2-fold difference in affinity, supporting that NapA does not exhibit strong intrinsic sequence selectivity under these conditions. We note, however, that the presence of supercoiling has been shown to be important in NapA’s interaction with DNA and supercoiling could itself be impacted by DNA sequence [37]. Hence, the in vivo environment may provide a different landscape for NapA binding that is affected by sequence. Future studies will be needed to address this question.

Chemicals and reagents

Reagents used for protein purifications (buffers and salts) were all of analytical grade and obtained from Sigma-Aldrich (St. Louis, MO). Synthetic oligonucleotides (oligos), including fluoresceinated oligos, were obtained from Integrated DNA Technologies (IDT) except for the biotinylated λ DNA, which was purchased from LUMICKS (Waltham, MA). Artificial genes that were codon optimized for E. coli expression were obtained from Genscript Corporation and subcloned into pET15b using NdeI and BamHI sites. The native marker protein standard mixture used for calibration of Mass Photometry (MP) experiments was obtained from Thermo Fisher Scientific.

Expression and purification of NapA proteins

The genes encoding Mtb full length (FL) NapA, Mtb NapA(16-102), Mtb NapA(16-92), Mtb NapA(K79A-R82A-R83A), Mtb NapA(T55A-R56A), Mtb MBP-tagged NapA(16-102), Mtb MBP-tagged NapA(16-102) (L29E-K42A-Y57E), Mtb HaloTag^® NapA(16-102) and the FL NapA homolog from Mycobacteriaceae bacterium (MBV9513839) (NapA-H) were purchased from Genscript Corporation as codon optimized genes for E. coli expression and subcloned into pET15b. The resultant expressed proteins had thrombin cleavable N-terminal His-tags for purification (Piscataway, NJ, USA;http://www.genscript.com). E. coli C41(DE3) cells were transformed with the expression vectors. For protein expression, cells were grown at 37 °C in Luria-Bertani (LB) medium with 0.10 mg/mL ampicillin to an OD₆₀₀ of 0.3-0.4, then induced with 1 mM isopropyl β-d-thiogalactopyranoside (IPTG) at 15 °C overnight. FL Mtb NapA could be expressed in soluble form but with low yield. Hence, based on sequence alignments showing that the N-terminal ~15 residues are not conserved, NapA(16-102) was generated.

All NapA proteins were purified using the same general protocol. Specifically, cells (9 liter growth) were harvested by centrifugation and then resuspended in 100 mL of Buffer A (50 mM Tris-HCl pH 7.5, 800 mM NaCl, 5% (v/v) glycerol, 5 mM MgCl₂, 0.5 mM β-mercaptoethanol (βME)). Each 100 mL of resuspended cells were incubated for 30 min at 4°C (while stirring) with 200 μL protease inhibitor cocktail stock (100 μM aprotinin, 1 mM leupeptin, 1 mM pepstatin A) (Roche) and 40 U of a solution of 100 mg/mL DNase I (Roche). The resuspended cells were disrupted, first by sonication and then by microfluidizer. Cell debris was removed by centrifugation at 15,000 rpm at 4 °C for 60 min. Supernatants were loaded onto a cobalt NTA column (TALON Superflow histidine-tagged protein purification resin). The column was generated by pouring the resin into a Bio-Rad glass column and equilibrating before addition of the protein containing lysate and an overnight wash with Buffer A. The proteins were eluted in Buffer A in an increased gradient of imidazole. Fractions were analyzed by SDS-PAGE and those containing the protein were combined. The fractions were concentrated using Sigma-Millipore concentrators (Amicon) and subjected to a second purification step using size exclusion chromatography (SEC) on a Superdex 75 pg column and an ÄKTA prime plus. The buffer used for this step was 25 mM Tris-HCl pH 7.5, 150 mM NaCl, 5% (v/v) glycerol, 5 mM MgCl₂ and 0.5 mM βME.

Selenomethionine substituted NapA(16-92) was generated using the methionine inhibitory pathway. Specifically, a 500 mL overnight culture of LB with 0.10 mg/mL ampicillin was inoculated with a colony of the C41(DE3) cells containing the NapA(16-92)-pET15b expression vector. The next day the cells were spun and the supernatant discarded and resuspended in 50 mL of 1X M9 media with 0.10 mg/mL ampicillin. The cells were spun and then resuspended again (to ensure removal of LB). A final spin and resuspension of the cells in 1 X M9 media was done and these cells inoculated into 9 Liters of 1X M9 media with 0.10 mg/mL ampicillin (and 2 mM MgSO₄, 2.5% dextrose, 5 mM CaCl₂) and grown at 37 °C to an OD₆₀₀ of 0.25-0.30. Then inhibitory amino acids (100 mg/L of lysine, phenylalanine and threonine, 50 mg/mL of isoleucine, leucine and valine and 100 mg/mL selenomethionine) were added, the temperature reduced to 15 °C and the cells grown for 20 min. 1 mM IPTG was then added and the cells grown overnight. The next day the cells were spun and the selenomethionine protein was purified as per the WT except that 1.0 mM βME was included in the buffers. His-tags were removed by thrombin digestion overnight at room temperature (rt) using a thrombin cleavage capture kit (EMD Millipore; Catalog #69022). The cleaved His-tags were removed by loading the cleavage reaction onto a Ni-NTA column and collecting the flow through. Tag-free proteins were concentrated using centricons (Amicon) with a 10 kDa MW cutoff.

Crystallization and structure determination of apo Mtb NapA

Neither NapA-H nor NapA(16-102) produced apo crystals, hence we utilized NapA(16-92) in screens. Crystals were produced by mixing the protein (at 25 mg/mL) 1:1 with a crystallization condition of 23% (w/v) PEG 3350, 0.1 M sodium thiocyanate. These crystals were cryopreserved with the crystallization solution supplemented with 15% (v/v) glycerol and data were collected at ALS beamline 5.0.2 and processed with XDS [72]. Selenomethionine substituted NapA(16-92) was generated for phasing. Data were collected at ALS beamline 5.0.2 and processed with XDS and SAD phasing was done in Phenix Autosol [72-73]. The resultant density map revealed a dimer of the NapA(16-92) in the ASU. After the model was fit, refinement in Phenix was performed [73]. Multiple cycles of rebuilding in Coot [74] and refinement in Phenix [73] were done (version 1.20.1-4487-000) until the refinement converged. See Supplementary Table S1 for data collection and refinement statistics.

Crystallization and structure determination of Mtb NapA-DNA complexes

For crystallization trials of NapA-DNA complexes, Wizard screens I to IV, Natrix and Natrix 2, and cryo I and II were used for screening at rt by the hanging drop vapor diffusion method. His-tag free proteins, either Mtb NapA(16-102) or FL Nap-H were used for crystallization experiments with DNA duplex sites ranging from 16 bp to 60 bp. Several crystal forms of NapA-H bound to the 16mer and 20mer sites were obtained, but all diffracted to 20 Å or showed no diffraction. NapA(16-102)-DNA crystals were also obtained, but again, most did not diffract. Crystals were obtained of NapA(16-102) bound to a 20mer site that showed weak diffraction on in-house X-ray sources. These crystals were grown by mixing the protein (30 mg/mL) with the self complementary 20mer dsDNA site, 5´-ATTATTAATTAATTAATAAT-3´. Crystals were produced by using the crystallization reagent, 0.1 M Tris-HCl pH 6.5, 7.5% MPD at a ratio of 1:1 of the protein-DNA complex: crystallization reagent. A 1-2 s dip of the crystals in the crystallization supplemented with 15% (v/v) ethylene glycol prior to plunging in liquid nitrogen was used for cryo-preservation. X-ray intensity data were collected at ALS beamline 5.0.2. However, the best diffraction observed was to ~7 Å. Hence, dehydration was attempted by dipping the crystals in the cryo-solvent and then dragging them out of the drop for several seconds. This extended the diffraction to a resolution of >4.00 Å, though the data were anisotropic. Data were for these crystals were collected at ALS beamline 5.0.2 to 3.87 Å resolution and processed with XDS [72]. Molecular Replacement (MR) using the apo NapA(16-92) dimer was successful in providing an initial solution and refinement was then carried out. After an initial round of refinement, density was evident for two DNA molecules. The application of crystallographic symmetry on these DNA molecules creates dsDNA (with complementary bases) and two pseudo-continuous dsDNAs in the crystal. For the DNA molecule with the best density (DNA1) eight nucleotides were built and again, symmetry generated a continuous duplex DNA in the crystal from this DNA. In the other DNA molecule, the density was weaker and five nucleotides were conservatively built. See Supplementary Table S1 for data collection and refinement statistics.

Fluorescence polarization (FP) binding experiments

To measure DNA binding to M. tuberculosis NapA(16-102) and FL NapA-H, the following fluoresceinated DNA targets were used; 16mer (top strand: 5´-TATTAATTAATTAATA-3´), 20mer (top strand: 5´-ATTATTAATTAATTAATAAT-3´), 20mer/with GCs (top strand, 5´-CTTGTTAGCTAGCTAACAAG-3´), 40mer (top strand: 5´-TATATATTAATTAATTAATATATTAATTAATTAATATATA-3´). In each experiment, increasing concentrations of the NapA protein were titrated into the sample cell containing 1 nM of the DNA in a buffer of 25 mM Tris-HCl pH 7.5, 125 mM NaCl, 2.5% (v/v) glycerol. DNA binding analyses of the three mutant proteins, NapA(16-92), NapA(K79A-R82A-R83A), NapA(T55A-R56A), were done with the 20mer. Samples were excited at 490 nM and fluorescence emission was measured at 530 nM. The resultant data were plotted using KaleidaGraph version 4.5 for Mac; serial #8011073 (Synergy software) and the curves fit with single site binding mode. The binding affinity of protein for DNA was determined using the following equation: $P=Pf+(Pb-Pf)[M]∕(apparentKd+[M])$ where $\mathrm{P}$ is the measured polarization, $\text{Pf}$ is the polarization of free DNA, $\text{Pb}$ is the polarization of the completely bound DNA, apparent ${\mathrm{K}}_{\mathrm{d}}$ is the equilibrium dissociation constant, and $\mathrm{M}$ is the concentration of protein. $\text{Pb}$ and apparent ${\mathrm{K}}_{\mathrm{d}}$ were determined by nonlinear least-squares analysis. Three technical repeats were performed for each curve and the standard errors from the three affinities were determined.

DNA bridging assays

To test the ability of NapA to bind between DNA sites a bridge assay was employed using biotinylated DNA and a fluorescein labeled DNA (F-DNA). For these experiments, a biotinylated 32mer ds DNA site, (top strand: 5-BiosG/5´-ATAATTAATTAATTATATAATTAATTAATTAT-3´) (IDT), was reconstituted in 50 mM sodium cacodylate, pH 5.6, to create a duplex site. In the assay, streptavidin beads were washed in buffer B (20 mM sodium cacodylate pH 5.6, 100 mM NaCl) to create a suspension. 100 μL of the suspension was added to spin tubes (thrombin kit). The tubes were spun (14,000 rpm) to remove buffer leaving streptavidin beads. 20 μL of 50 μM biotinylated 32mer DNA was added and the tube spun to remove buffer. Next, 10-50 μL of 1 mg/mL NapA(16-102) protein (WT or NapA(16-102)(K79A-R82A-R83A) in buffer B were added to separate tubes. Buffer alone was added to another tube containing streptavidin beads as a control. The samples were spun to remove buffer and washed 2X with buffer B. Then 20 μL of fluorescein (56-FAM) labeled 64-mer dsDNA (5´-ATAATTAATTAATTATATAATTAATTAATTATATAATTAATTAATTATATAATTAATTAATTAT-3´) was added to each of the three tubes and each tube spun at 9,000 g to remove buffer. Each sample was washed 2X with buffer B. To assess if the F-labeled DNA had been retained indicating bridging, 20 μL of 8 M Urea was added and the samples spun and then 10 μL of gel sample running buffer was added to the flow through and the samples boiled. Each sample was run on a 1.5% agarose gel to assess the presence of the fluorescein labeled DNA. Gels were imaged using a Bio-Rad ChemiDocTM MP Imaging System, with the XcitaBlueTM Conversion Screen (Bio-Rad Cat# 1708182). The specific parameters selected were as follow: Flat Field = Applied (UV Trans orange); Application = Pro-Q Emerald 300 or Fluorescein; Excitation Source = UV Trans Illumination; Emission filter = Standard Filter; Software Version = 5.2.1. Experiments were done in duplicate.

Mass photometry (MP) experiments

MP experiments were done using a Refeyn OneMP instrument (Oxford, UK). The experiments were performed using microscope coverslips, cleaned by sequential sonication in isopropanol (HPLC grade, and Milli-Q H2O (15 min each)), followed by drying with a clean, pressurized air stream. Clean coverslips were assembled into the flow chamber, and silicone gaskets were positioned on the glass surface for sample loading to hold the sample drops with 4 X 4 wells prior to measurements. Contrast-to-mass calibration was achieved by measuring the contrast of proteins in the native marker protein standard mixture (NativeMark Protein Standard, Thermo Fisher Scientific) to generate a standard calibration curve; 1048 kDa, 480 kDa, 146 kDa, and 66 kDa Molecular masses were used to fit the calibration curve with an R2 value of 0.998 and a Max mass error of 11.3% in the Refeyn DiscoverMP v2023 R2 software. Calibration was applied to each sample measurement to calculate the molecular mass of each histogram distribution during analysis.

For these experiments, MBP-NapA(16-102) and the MBP-NapA(16-102) (L29E-K42A-Y57E) mutant were buffer exchanged into 25 mM Tris-HCl pH 7.5, 150 mM NaCl, and diluted in a buffer of 20 mM Tris-HCl pH 7.5, 150 mM NaCl to a final concentration of 100 nM prior to sample analysis with 2-fold dilution on a buffer droplet to a final concentration of 50 nM and 120 nM. To find focus, 10 μL of fresh buffer (20 mM Tris-HCl pH 7.5, 150 mM NaCl) adjusted to rt was pipetted into a well and the focal position was identified and locked using the autofocus function of the instrument. To obtain readings at higher concentrations, we used a MassFluidix HC add-on to the MP system. For these experiments, MBP-NapA(16-102) samples at 100 μM concentrations were rapidly diluted (10,000-fold and 5,500-fold) with the MassFluidix high concentration (HC) microfluidic system, flowed onto the measurement surface (coverslips) and the data acquired as for the regular MP readings (the mutant aggregated at these higher concentrations and so could not be analyzed). Movies of 60 s duration with 6000 frames (10 frame rate/Hz) were recorded using Refeyn AcquireMP version2023 (v2023) R1 software using normal measurement mode with regular image acquisition settings for data collection. Three technical replicates were performed for each experiments. All mass photometry movies of each measurement were processed and analyzed by Refeyn DiscoverMP version2023 R2 software, and Gaussian curves were fit to each histogram distribution, and the mass (kDa), sigma (kDa), and counts were determined.

Circular Dichroism (CD) experiments

For CD, WT MBP-NapA(16-102) and MBP-NapA(16-102) (L29E-K42A-Y57E), were each diluted 50-fold to a final concentration of 0.25 mg/mL with CD Buffer (50 mM NaH₂PO₄, 200 mM NaF, 5% (v/v) glycerol, 1 mM βME, pH 7.5). Readings were done for 5 replicates for each corresponding sample, including the control (Buffer CD), using a Circular Dichroism Spectrometer from AVIV Biomedical Inc. Data were collected using the AVIV 435 CD software.

Glutaraldehyde crosslinking experiments

For glutaraldehyde crosslinking of WT NapA(16-102), NapA(16-102)(K79A-R82A-R83A), WT MBP-NapA(16-102) and MBP-NapA(16-102) (L29E-K42A-Y57E) proteins at 0.3 mg/mL were used in a buffer consisting of 50 mM sodium cacodylate pH 6.5, 150 mM NaCl. Glutaraldehyde at a final concentration of 0.05% was added to the proteins and crosslinking assessed after varying time points. The reactions were quenched by addition of SDS running buffer and the samples run on 15% SDS-PAGE gels and visualized by staining with Simply Blue Safe Stain. The MW marker used was Bio-Rad, Precision Plus Protein Standard (catalog # 1610374). The crosslinking experiments were done in duplicates and revealed consistent results.

Single-molecule optical tweezer confocal imaging with LUMICKS C-trap

For C-trap experiments, HaloTag^® NapA(16-102) was labeled with JF503 ligand dye in a buffer composed of 25 mM Tris-HCl pH 8.0, 100 mM sodium chloride, 5 % (v/v) glycerol, and 0.5 mM βME in a protein: ligand molar ratio of 1:1 with 4 μM final concentration of each. The mixture was incubated at rt for 2 h and then the free dye was removed by passing the sample through Amicon^® Ultra 0.5 mL 10 kDa filter while concentrating the bound ligand and HaloTag^® NapA(16-102)-JF503. HaloTag-JF503 labeled NapA(16-102) protein was excited with a 488 nm laser and emission was collected with a 500–550 nm band pass filter. All data were collected with a 1.2 NA 60x water emersion objective and photons were measured with single-photon avalanche photodiode detectors. The laser power was set to 10% and 2D confocal scans were collected for each pixel of size 100 nm with a pixel dwell time of 1 ms. Data were acquired using LUMICKS Bluelake software version 1.6.16. Custom software from LUMICKS session viewer was employed for single-molecule data analysis and green pseudo-colors were applied for signal contrast visualization. 2D confocal images of the bound NapA(16-102)-halo-JF503 tagged on λ-DNA were created using Fiji (ImageJ) software for each 2D scan. Biotinylated double-stranded DNA (48,502 bp) λ-DNA, ~4.5 μm streptavidin-coated beads were used.

Single-molecule optical tweezer confocal microscopy experiments were performed at 25 °C combining two-color confocal fluorescence microscopy analyses on the LUMICKS C-trap with dual optical trap integrated confocal microscope using microfluidics with a 5-channel flow chamber (LUMICKS C-trap, Netherland). All 5 channels of the microfluidics chip were first passivated with 0.7% (w/v) BSA in Imaging buffer 30-60 min before starting the experiments. The optical trap movement was facilitated by a computer-controlled stage within a five-channel flow cell. Channels 1–3 were separated by laminar flow, and under a constant flow, a single streptavidin-coated polystyrene bead was caught in each trap using the dual optical tweezers in channel 1. The traps were then moved to channel 2 and biotinylated DNA was tethered between the streptavidin-coated bead traps as detected by an increase in the force reading. Flow was stopped, and the traps were moved to channel 3 containing only an imaging buffer where the presence of a single DNA tether was confirmed by the force-distance curve to check the integrity of the tethered λ-DNA and get the single tether. To investigate the binding of NapA(16-102)-halo-JF503 tagged on λ-DNA, optical traps tethering a λ DNA molecule (16.4 microns) under 1 to 15 pN of constant tension were moved slowly into channel-4 of the microfluidic flow cell loaded with 32 to 62.5 nM of Nap(16-102)A-halo-JF503 (i.e. excess NapA(16-102) to DNA) in an imaging buffer containing 25 mM Tris-HCl pH 7.6, 100 mM sodium chloride, 5% (v/v) glycerol, and 0.5 mM βME. Following 60-90 s incubation, the tether was moved to channel 3 containing the imaging buffer only for removal of nonspecific binding events and imaging.