Characterization of a nontypeable Haemophilus influenzae thermonuclease

Nontypeable Haemophilus influenzae (NTHi) has been shown to form biofilms, comprised of extracellular DNA (eDNA), in the middle ear and bronchus during clinical infections. Studies in our laboratory have shown that NTHi possesses a homolog of Staphylococcus aureus thermonuclease (staphylococcal thermonuclease), NTHi nuclease (NTHi Nuc, HI_1296). This enzyme had similar size, heat stability, and divalent cation requirements to those of the staphylococcal homolog as determined by light scattering and circular dichroism spectroscopy. Small angle X-ray scattering (SAXS) analysis suggested an overall shape and substrate-binding site comparable to those of staphylococcal nuclease. However, NTHi Nuc was approximately 25-fold more active in fluorescence resonance energy transfer (FRET) activity assay than staphylococcal thermonuclease. Homology modeling implicates shorter NTHi Nuc loops near the active site for this enhanced activity.


Introduction
Nontypeable Haemophilus influenzae (NTHi) are frequently found as a member of the normal upper respiratory tract bacterial flora. This species can be a frequent cause of airway infections, including otitis media in children, and sinusitis and acute exacerbations of chronic bronchitis in adults [1]. NTHi has been shown to be capable of forming biofilms in vitro and in the upper and lower human respiratory tract during human disease [2][3][4]. Bacterial biofilm matrix is an elaborate network of molecules, which can include pili, polysaccharides, double stranded extracellular DNA (eDNA), and bacterial and host derived substances that help shape and secure the biofilm to an inanimate or host surface [5]. Recent  and dying neutrophil, and appear to be active similar to stationary biofilms [6,7]. The matrix protects the underlying bacteria from assault by the host immune response and antibiotic treatment thus contributing to the recalcitrant nature of biofilm to treatment [8]. An open reading frame with high homology to staphylococcal thermonuclease, called NTHi nuclease (NTHi Nuc, HI_1296), is present in the sequenced genome of H. influenzae strains 2019, KW20 Rd and 86-028NP. We have previously found that this nuclease is an important factor in remodeling NTHi biofilm [9].
In this study, we present biophysical and functional characterization of NTHi Nuc. FRET based enzyme kinetic studies suggested that NTHi Nuc was~25 times more active than staphylococcal thermonuclease. SAXS analysis indicated that NTHi Nuc had a structure and DNA binding cleft similar to those of staphylococcal thermonuclease. Like the staphylococcal thermonuclease, this enzyme retained activity after heating to 65˚C and required a divalent cation, magnesium or calcium, for optimal activity. This stability and increased activity of NTHi Nuc could play a role in efficacy of NTHi biofilm remodeling.

Bacteria and culture conditions
Escherichia coli (E. coli) BL21 Ã (DE3) was grown in Luria-Bertani (LB) medium with or without agar and supplemented with IPTG as needed.

Cloning, expression and purification of NTHi Nuc
NTHi Nuc without signal sequence was expressed with a cleavable 6x His-tag in pET151/ D-TOPO (Life Technologies Corporations) in BL21 Ã (DE3) E. coli cells and induced with 1.5 mM IPTG (Invitrogen) at 18˚C. Cell cultures were pelleted and frozen at -20˚C in preparation for lysis. The frozen cells were resuspended in 100 mM Tris, pH 9.1, 5 mM CaCl 2 , 200 mM NaCl, 50 mM imidazole (lysis buffer) plus a mini cOmplete protease inhibitor cocktail tablet (Roche) then homogenized with an Emulsiflex 2000. His-tagged NTHi Nuc was bound to Ni-NTA affinity resin (Qiagen) and washed with the lysis buffer then eluted with elution buffer (lysis buffer with 250 mM imidazole). The eluted protein was concentrated and mixed with lab-purified Tobacco Etch Virus protease at 1:200 in 4˚C for 16 hours to cleave the His-tag. The cleaved products were passed over a Ni Sepharose column (HisTrap FF, GE Healthcare) and the flow-through was loaded onto a gel filtration column (Superdex 75 25/60, GE Healthcare) to separate NTHi Nuc from contaminants and to exchange buffer into the final purification and reaction buffer: 100 mM Tris, pH 9.1, 5 mM CaCl 2 , 200 mM NaCl, 5 mM DTT.

FRET assay
Nuclease enzyme activity was measured by a fluorescence resonance energy transfer activity study (FRET) assay using a Tecan Infinite M200 Pro. The FRET substrate was a singlestranded 30-mer oligonucleotide (Cy3-CCGCGAGAAACCAAGCACAGAGCACCGAAGA-BHQ_2) with the 5' end modified with Cy3 fluorophore and the 3' end modified with Black Hole Quencher 2 [10]. The substrate was incubated with either NTHi Nuc, S. aureus thermonuclease (Sigma-Aldrich) or RNase-free bovine pancreatic DNase I (New England Bioscience, positive control), both enzymes in DNase I reaction buffer (New England Bioscience), at 26˚C, and then the absolute fluorescence was measured every 10 s for 5 min. Relative activity was calculated by, is the reaction velocity of the enzyme of interest and nM is the concentration (nM) of the enzyme of interest. v d is the reaction velocity of staphylococcal thermonuclease and nM d is the staphylococcal thermonuclease concentration (nM). v/nM of pancreatic DNaseI was calculated based on published data [11]. Each study shown was performed a minimum of six times on multiple NTHi Nuc samples.

Melting point analysis
The DynaPro NanoStar in DLS mode was used to conduct melting temperature experiments in both 50 μl and 4 μl plastic cuvettes on filtered NTHi Nuc samples. Rubber caps were used to seal the sample chambers during temperature ramps to minimize sample evaporation. Scattering intensity and hydrodynamic radius data were recorded as temperature was ramped at 1.0˚C/min from 25˚C until the protein unfolded. A sudden rapid increase of these parameters from the baseline marks the onset of protein unfolding. These T onset values with standard deviations were calculated using the linear fit analysis module in Dynamics v7. Melting point data was validated with Jasco J-815 Circular Dichroism Spectropolarimeter (21-Q-1, Starna Cells Inc.) [12]. Temperature was increased by 1˚C/min and spectra collected from 260-190 nm with data pitch of 1 nm, bandwidth of 1 nm, data integration time of 2 sec and scanning speed of 100 nm/min. Circular Dichorism (CD) spectrum measurements were also taken at fixed temperatures using the same parameters but with 10 measurements/spectra/temperature for higher accuracy. Purified NTHi Nuc was diluted to 0.5 mg/ml in phosphate buffer (PBS + 0.003% Tween-20) to reduce buffer background in CD experiments.

Homology modeling
A homology model of NTHi Nuc was generated using the hm_model macro in the program YASARA (YASARA Biosciences). High-resolution (1.5-1.9 Å) crystal structures of staphylococcal nuclease were used as templates and the best of five models (in terms of dihedral angles and packing scores) after simulated annealing minimization was selected as the homology model.

Small-angle X-ray scattering (SAXS)
Purified NTHi Nuc was dialized against SAXS buffer (100 mM Tris, pH 7 or 9, 150 mM NaCl, 5 mM CaCl, 5 mM DTT). SAXS data were collected at the SIBYLS beamline (beamline 12.3.1) at the Advanced Light Source, Lawrence Berkeley National Laboratory. The data were collected at 10˚C at three concentrations in the range of 1.2-3.7 mg/ml for the sample at pH 7 and 0.4-1.2 mg/ml for the sample at pH 9. The samples were progressively exposed to radiation for 0.5, 1, 2 and 4 seconds. Radiation damage was observed for the highest exposure for all samples and those data were not used for further analysis. Scattering data from the dialysate buffer were collected and subtracted from the protein solution scattering data. No concentration-dependent scattering was noted based on data processed from three different concentrations. Data processing was performed using the PRIMUS program package [13]. Analysis of the Guinier region (low-q region) of the scattering data for all samples at all concentrations and exposures (except the highest) was linear, suggesting that the proteins did not undergo aggregation or interparticle interference. The radius of gyration, R g , was evaluated using the Guinier approximation at low q where qR g < 1.3 [14]. The program GNOM [15] was used to compute the pair distribution function, P(r), and the maximum dimension of the macromolecule, D max . Experimental SAXS data were compared with simulated scattering curves in the program CRYSOL [16] for the homology model of NTHi Nuc and the crystal structure of Staphylococcal thermonuclease (pdb id 1STN). Ab initio molecular envelopes were generated from the experimental scattering data using both DAMMIF [17] and GASBOR [18].
Only the highest concentration and highest exposure scattering data without radiation damage were used for ab initio reconstructions. Eight molecular envelopes were obtained using identical parameters but different randomly chosen values to seed the calculation. The coordinates for the models from eight runs were superimposed, averaged and filtered in the programs SUPCOMB [19] and DAMAVER [20] to create an average model, which represented the general structural features of each reconstruction, and a filtered model, which represented the core conserved features of all reconstructions. The derived molecular envelopes were visualized in PyMOL (v.1.6; Schrödinger LLC).

Statistical analysis
Statistical analysis was performed with Prism 6 software (GraphPad Software, Inc., La Jolla, CA). Values that met a p value cutoff of 0.05 were considered statistically significant.

NTHi Nuc (HI_1296) is homologous to staphylococcal thermonuclease
Complete amino acid sequence of NTHi 2019 Nuc is shown in Fig 1A with the signal sequence highlighted. Fig 1B shows that NTHi Nuc is homologous to staphylococcal thermonuclease with 35% conserved amino acids and E-value of 2e-10. Enzymatic and physicochemical characterization of NTHi nuclease. Enzyme activity studies were performed with PCR grade staphylococcal thermonuclease (Sigma-Aldrich) and highly purified recombinant NTHi Nuc without signal sequence. Signal sequence was determined by LipoP 1.0 analysis. NTHi Nuc possessed both single-stranded (Fig 2) and doublestranded DNA [9] nuclease activity. NTHi Nuc was a highly efficient nuclease with greater than 1500 fold increased activity compared to bovine pancreatic DNase I and greater than 25 fold increased activity compared to staphylococcal thermonuclease (Table 1). NTHi Nuc activity was greatest near pH 9, similar to staphylococcal thermonuclease, however NTHi Nuc retained some activity over a range of pH 5-9.5 (data not shown). NTHi Nuc activity was also temperature dependent with maximum at 37˚C. Of note, NTHi Nuc remained active at 4˚C and 25˚C for several months in sterile solution and retained endo-exonuclease activity above 45˚C in pH = 8.5; heating to 60˚C partially denatured the protein. Heat-denatured NTHi Nuc, when cooled to 25˚C, re-established its activity and was able to degrade DNA (Fig 3). NTHi Nuc and staphylococcal thermonucleases are most active near their pIs [21], similarly both staphylococcal thermonuclease and NTHi Nuc are competitively inhibited by thymidine  3,5-bisphosphate (pdTp), a DNA mimic [22]. Half and equimolar concentrations of pdTp impacted initial rates of endonuclease activity (Table 1). However, at low substrate concentration at 300 seconds, half-molar nuclease-pdTp reactions had approximately twice the activity of equimolar nuclease-pdTp reactions (Fig 4).

Divalent cation dependence
EDTA chelation quenched NTHi Nuc activity (Fig 2), highlighting divalent cation requirement for its activity. Addition of Ca 2+ or Mg 2+ restored NTHi Nuc activity (data not shown). Either Ca 2+ or Mg 2+ was required, but not both, as NTHi Nuc activity was unchanged when both Ca 2+ or Mg 2+ were added (data not shown). Cobalt, an inhibitor of staphylococcal thermonuclease [23], not only aggregated NTHi Nuc structure (Fig 5. Panel A), but also quenched its activity, similar to addition of EDTA (Fig 5. Panel B).   Fig 6A shows scaled SAXS scattering curves of NTHi Nuc in pH = 7 (blue) and 9 (red). There were no gross structural changes at those two pH. Fig 6B and Fig 6C show Guinier plot for NTHi Nuc at pH = 7 and 9, respectively. Rg was 17.4 Å at both pH. Fig 6D shows parabolic Kratky plots in pH 7 (blue) and 9 (red), showing some flexible region. Fig 6E shows pairwise distribution function plots at pH 7 (blue) and 9 (red), with maximum dimensions of 58.3 Å at pH 7 and 59.6 Å at pH 9. https://doi.org/10.1371/journal.pone.0197010.g006

Static and dynamic light scattering
Static and Dynamic Light Scattering (SLS/DLS) were used to determine the biophysical characteristics of NTHi nuclease. Molecular weight determination from SLS shows no oligomerization at pH 7 (Table 2). At pH 9, hydrodynamic radius increased, indicating a larger open conformation ( Table 2). Protein stability at various pH was reported as melting temperatures, T onset, in Table 3. NTHi Nuc was most stable around pH 9 with a reported T onset of 53˚C. Comparatively, the T onset of staphylococcal thermonuclease is 45.8˚C at pH 7 (calculated from published data; [24]). At pH 7, excess pdTp increased NTHi Nuc thermal stability, reported as a T onset , from 40˚C to 48˚C (Table 3). In addition to pH dependence, NaCl and divalent cations were assayed for effects on stability of NTHi Nuc. DLS showed that NTHi Nuc was stable between 50 mM and 1M of NaCl, below and above physiological concentrations (data not shown). When Ca 2+ was replaced with Co 2+ , NTHi Nuc aggregated (Fig 5. Panel A) and activity was lost (Fig 5. Panel B). Passing NTHi Nuc + Co 2+ through a 0.22 μm membrane filter removed these aggregates and re-aggregation can be quantified in real time using DLS (Fig 5. Panel A).

Small-angle X-ray scattering
The Guinier analysis resulted in R g values of 17.4 Å at both pH 7 and pH 9. The radius of gyration of the homology model as calculated by CRYSOL was 17.0Å indicating that NTHi Nuc was a monomer in solution. Pairwise distance distribution function, P(r), showed maximum dimension, D max , of 58.3 and 59.6 Å (Fig 6) at pH 7 and 9, respectively. These data indicate that despite the significant difference in activity of NTHi Nuc at pH 7 and 9, there were no global conformation changes. The ab initio model had features, such as the active site cavity, that were remarkably similar to that of staphylococcal nuclease (Fig 7) indicating similarity between NTHi nuclease and staphylococcal thermonuclease. However, the three loops, including the O loop near the calcium ion, forming the metal binding and active site in NTHi Nuc homology model were predicted to be shorter than the corresponding loops in staphylococcal nuclease (Fig 8).

Discussion
A number of bacterial species responsible for causing human disease including Staphylococcus aureus, Neisseria gonorrhoeae, Streptococcus sanguinis, Serratia marscens, Campylobacter jejuni, Enterococcus faecalis and Bacillus anthracis encode a thermonuclease. The functional properties and structure of the S. aureus thermonuclease has been studied for almost 50 years [26]. Studies have shown that S. aureus and N. gonorrhoeae encode nucleases, which play a role in remodeling the extracellular DNA (eDNA) matrix of the biofilm formed by the respective organisms [27,28]. In addition, it has been shown that the nucleases of N. gonorrhoeae, S. aureus and S. sanguinis protect the microbes from neutrophil extacellular trap-mediated bactericidal activity [29][30][31]. Our previous study evaluated the role of NTHi Nuc in biofilm by NTHi [9]. This study further characterizes the NTHi enzyme in detail. NTHi Nuc has 35% identity and 55% similarity to staphylococcal thermonuclease. Similar to the staphylococcal thermonuclease, NTHi Nuc could be inactivated by heating to 65˚C, but could refold to its active form at 25˚C. Unlike staphylococcal nuclease, which has no cysteines, NTHi Nuc has three cysteines, two of which were predicted to be on neighboring beta strands in the homology model, capable of forming a disulfide bond that could potentially contribute to its high thermostability. NTHi Nuc was approximately 25 and 1500 times more potent than staphylococcal thermonuclease and DNase I, respectively. Low resolution SAXS analysis indicated that NTHi Nuc had a similar binding pocket and general shape as staphylococcal thermonuclease, but predicted to have shorter loops forming the metal binding and active sites compared to those of staphylococcal thermonuclease, including the O loop. Previous studies have shown that the deletion of staphylococcal thermonuclease O loop had increased stability compared to the staphylococcal thermonuclease with an active site mutation [32,33]. Therefore, the absence of the O loop in NTHi Nuc could play a role in the enhanced activity of the NTHi Nuc. The other two shorter loops in NTHi Nuc active site could also allow space for larger substrates, which may also contribute to the enhanced activity of the NTHi Nuc. The melting temperatures and optimal pH for enzymatic activity are similar for both enzymes. NTHi Nuc required divalent cations, Mg 2+ or Ca 2+ , for its activity. NTHi Nuc can digest both double-and singlestranded eDNA.
Our previous work has shown that NTHi nuc expression is crucial for biofilm remodeling and dispersal. We were surprised that while the nuclease was regulated, there was only 1.52-fold increased gene expression in planktonic phase compared to those of biofilm phase. The high potency of NTHi Nuc suggests that this tight regulation of expression was necessary to disperse organisms from biofilm and was required to maintain biofilm homeostasis.
The use of exogenous nucleases has been suggested to treat chronic biofilm infections in the respiratory tract in illnesses such as cystic fibrosis [34,35]. Given the potency of the NTHi nuclease, it may find a role in management of such disease processes. In addition, the NTHi nuclease could find usefulness in industrial processes such as production of flavor enhancers [36] and in treatment of food processing equipment contaminated with DNA containing biofilms [37].