Bacterial strains, media, and chemicals

Escherichia coli K-12 TOP10 cells (Invitrogen, Carlsbad, CA), grown in Luria-Bertani (LB) medium, were used for preparation of plasmid DNA. Escherichia coli BL21(DE3) strain grown in LB medium was used for recombinant protein expression. S. pneumoniae 774A, isolated from CSF of a child with meningitis [37] was grown routinely in Brain Heart Infusion (BHI) broth or on Horse Blood Agar (HBA) plates, at 37°C in an atmosphere of 5% CO₂. The chemically defined medium with (CDM⁺), or without (CDM^-), (S)-lysine (200 µg ml^-1) was that described by van de Rijn and Kessler [38], supplemented with choline chloride (5 µg ml^-1), asparagine (50 µg ml^-1), and sodium pyruvate 250 µg ml^-1) as described by Moscoso et al. [1]. Ampicillin was used at a final concentration of 50 µg ml^-1, chloramphenicol at 10 µg ml^-1, erythromycin at 0.1 µg ml^-1, and 5-bromo-4-chloro-3-indolyl-β-D-galactopyranoside (X-Gal) at 25 µg ml^-1. For growth studies, cells from freshly grown single colonies on HBA plates were inoculated into 3 ml of CDM⁺ and 3 ml CDM^-. Cultures were incubated statically for 16 hours in an atmosphere of 5% CO₂ at 37°C. Growth was assessed by monitoring the optical density (Absorbance) of duplicate cultures on an hourly basis using a Klett-Summerson photoelectric colorimeter (filter no. 54, spectral range 520-580 nm).

Recombinant DNA techniques

Routine DNA manipulations were performed using standard techniques [39]. Plasmid DNA was purified using Wizard plus SV DNA purification system (Promega, Madison, WI). PCR amplifications were performed with Phusion High-Fidelity DNA Polymerase (Finnzymes Oy, Finland) or high-fidelity Platinum Taq DNA polymerase (Invitrogen, Carlsbad, CA). DNA derived from PCR reactions was purified using the UltraClean PCR Clean-up Kit (Mo Bio Laboratories, Inc.). Oligonucleotides were purchased from GeneWorks Pty. Ltd. (Hindmarsh, South Australia, Australia).

Construction of the S. pneumoniae 447A ΔdapA mutant strain

Overlapping extension PCR [40] was used to generate a DNA fragment carrying the erythromycin resistance (Em^R) gene flanked by regions up and downstream of S. pneumoniae 447A dapA. Briefly, primer pairs dap.F/dapERM.R and dapERM.F/dap.R (Table 1) were used to amplify DNA flanking the region to be deleted from the chromosome of S. pneumoniae 447A, and primers pVA838.F/pVA838.R (Table 1) were used to amplify the Em^R gene from plasmid pVA838 [41]. The products of these three PCR reactions (100 ng each) served as template in overlapping extension PCR using primers dap.F/dap.R (Table 1) to generate a linear construct, which was cloned into pGEM-T Easy (Promega, Madison,WI), introduced into E. coli K-12 TOP10 cells and confirmed by sequencing. The pGEM-T Easy construct was used as a template in a PCR with primers dap.F/dap.R, to amplify the linear allelic replacement DNA fragment, which was introduced into S. pneumoniae 447A by transformation. The ΔdapA mutation was confirmed by PCR using primer pairs in which one primer flanked the targeted region and the other primed within the Em^R gene (OCD52/dapERM.R and OCD53/dapERM.F). The PCR products were sequenced using primers OCD52 and OCD53 (Table 1).

Transformation of S. pneumoniae 447A

Bacteria were grown in c-CAT medium (1% w/v Casamino acids, 0.5% w/v Tryptone, 0.5% w/v NaCl, 1% w/v Yeast Extract, 16 mM K₂HPO₄, 0.2 % w/v glucose, 15 μg ml^-1 glutamine) at 37°C to OD₆₀₀ of 0.25-0.30. Cells were diluted 1/10 in 10 ml CTM medium (c-CAT containing 0.2% BSA and 1 mM CaCl₂), grown at 37°C to OD₆₀₀ of 0.10, collected by centrifugation and resuspended in 1 ml of 15% v/v glycerol prepared in CTM adjusted to pH 7.8. 100 μl aliquots of cell suspension were stored at -80°C until required. For transformation, 100 µl of cells were thawed on ice, 1 ml of CTM-pH 7.8 and 100 ng of synthetic competence-stimulating peptide 1 (CSP-1) [42] were added and cells incubated at 37°C for 13 min. DNA was added and cells were incubated at 32°C for 35 min then incubation continued for 3 hours at 37°C. Transformation mixture was plated out on HBA containing 0.1 µg ml^-1 erythromycin and incubated overnight at 37°C in an atmosphere of 5% CO₂.

Expression and purification of Sp-DHDPS

Recombinant Sp-DHDPS was expressed, purified and assessed by SDS-PAGE and mass spectrometry to be >95% homogeneous as described previously [43].

Coupled kinetics assay

Kinetic studies were conducted using the coupled-assay method as previously described [16,36,44,45]. Data were collected on a Cary UV-Vis spectrophotometer (Varian) connected to a Peltier cell to maintain a constant temperature of 30°C. Assays were performed in 1.5 ml semi-micro acrylic cuvettes with a path length of 10 mm. (S)-ASA was synthesized according to the methods of Roberts et al. [46]. A standard assay contained 20 nM of DHDPS, 250 mM HEPES pH 8.0, 0.2 mM NADPH, varied concentrations of pyruvate and (S)-ASA, 75 µg ml^-1 of DHDPR (purified from E. coli) in a final volume of 0.8 ml. Cuvettes were pre-incubated at 30°C for 8 min before the reaction was initiated via the addition of (S)-ASA. Rates were determined from the initial linear portion of the data collected. The background degradation of NADPH was factored into rate calculations and each data point was measured in triplicate within a <10% error margin.

Circular dichroism (CD) spectroscopy

CD spectroscopy experiments were performed on an Aviv Model 420SF CD spectrometer using a 1.0 nm bandwidth. CD spectra of Sp-DHDPS were recorded at a protein concentration of 4.5 µM solubilized in 20 mM Tris-HCl pH 8.0 and 150 mM NaCl. Wavelength scans spanning 195-240 nm were measured using a step size of 0.5 nm with a 2 sec averaging time in a 1 mm stoppered quartz cuvette as reported previously [16,36,47]. The resulting spectra were analyzed using the CONTINLL algorithm and SP43 database employing the CDPRO software package [48,49]. Thermal denaturation experiments were monitored at 222 nm over a temperature range of 4–90°C collecting data at 1°C intervals with a 5 s averaging time. Given that DHDPS requires chaperones to fold [50], the denaturation of DHDPS enzymes is irreversible in vitro, and therefore the apparent melting temperature (T_M^app), or midpoint of the transition between folded to unfolded state, was determined empirically from the ordinate maximum of the first derivative of the thermal denaturation profile [36].

Analytical ultracentrifugation

Analytical ultracentrifugation (AUC) studies were conducted in a XL-I analytical ultracentrifuge (Beckman Coulter) using 12 mm double sector cells with quartz windows loaded into either an An-60 Ti 4-hole rotor or An-50 Ti 8-hole rotor at a temperature of 20°C. Sp-DHDPS samples for all centrifugation runs were solubilized in 20 mM Tris-HCl pH 8.0 and 150 mM NaCl. For sedimentation velocity experiments, 380 µL of sample and 400 µL of reference (20 mM Tris-HCl pH 8.0 and 150 mM NaCl) were employed and absorbance versus radial profiles were generated at 40,000 rpm in continuous mode using a step size of 0.003 cm and a radial range of 5.8-7.3 cm without averaging. Data were collected at wavelengths of 210 nm (296-740 nM Sp-DHDPS) or 230 nm (4.5 µM Sp-DHDPS). The absorbance versus radii profiles at different time points were fitted to single discrete species or a continuous size-distribution model using the program SEDFIT [51,52] (available from www.analyticalultracentrifugation.com). The program SEDNTRP [53,54] was used to determine the partial specific volume of Sp-DHDPS (0.7475 ml g^-1), buffer density (1.005 g ml^-1) and buffer viscosity (1.021 cp) at 20°C. For sedimentation equilibrium experiments, 120 µL of reference solution and 100 µL of sample at three initial protein concentrations (i.e. 296 nM, 355 nM and 740 nM) were centrifuged at rotor speeds of 10,000 rpm and 18,000 rpm. Initial absorbance versus radial profiles were measured at 210 nm between 6.8 and 7.2 cm in step mode using a 0.001 cm step size and 3 averages until sedimentation equilibrium was attained (t ~ 24 hours). At sedimentation equilibrium, detailed absorbance versus radius scans were taken using a step size of 0.001 cm over a radial range of 6.8 to 7.2 cm with 15 replicates. The resulting absorbance versus radial position profiles at multiple sample concentrations and rotor speeds were globally fitted to a single discrete species model or various self-associating models (including monomer-dimer, monomer-trimer, dimer-tetramer, trimer-hexamer and monomer-dimer-tetramer models) using the program SEDPHAT [55] (available from www.analyticalultracentrifugation.com).

Dynamic light scattering

Dynamic light scattering measurements were made on an ALV 5022F DLS (ALV, Germany). Protein samples were analyzed at a final concentration of 59 μM. Measurements were taken over a 30 s time period with 10 replicates at 20°C. The samples were illuminated with a HeNe laser (633 nm) and experiments were conducted at a scattering angle of 90 degrees. The in-built ALV analysis software was used to determine average radii.

Crystallization

Crystallization of Sp-DHDPS was performed as previously described [40]. Briefly, crystals were obtained by employing the hanging-drop vapor-diffusion method. 1 µl protein solution (10 mg ml^-1) and 1 µl precipitant solution were equilibrated against 1 ml reservoir solution [0.2 M ammonium chloride, 20% (w/v) PEG 6000, 0.1 M MES pH 6.0] in 24-well Linbro plates at 20°C. Crystals were soaked briefly in cryoprotectant composed of 0.2 M ammonium chloride, 20% (w/v) PEG 6000, 0.1 M MES pH 6.0 and 20%(w/v) glycerol and were then flash-frozen using liquid nitrogen.

Structure determination

X-ray diffraction experiments were carried out at the Australian Synchrotron, Victoria, Australia on the MX1 beamline using Blu-Ice [56]. A 1.9 Å resolution data set was integrated and merged with HKL2000 [57] and scaled with SCALA [58]. The crystals were initially assigned to the tetragonal crystal system, space group P4₂2₁2 with unit cell dimensions a = b = 105.5 Å and c = 62.4 Å, and an overall R_merge of 6%. Molecular replacement was carried out using the program PHASER [59] with the structure of dihydrodipicolinate synthase from B. anthracis (PDB ID: 1XL9) [22] as the search model. One molecule was located in the asymmetric unit. The structure was refined using the program PHENIX [60] with rigid body refinement, simulated annealing and atomic displacement parameters (ADP) refinement. Iterative model building was carried out using the program COOT [61] followed by ADP refinement. Even though the 2F_o-F_c electron density map was clear, there were spurious streaks in the F_o-F_c difference maps and refinement stalled with a R_work of 24 % and a R_free of 32 %. Examination of the cumulative intensity plot from SCALA suggested that the data were affected by crystal twinning. The data were inspected further by running the program phenix.xtriage. Since there are no twin laws possible in the above crystal symmetry, over-merging of pseudo-symmetric or twinned data may have been present. The diffraction data were re-integrated in space group P1 using XDS [62] and further analyzed with phenix.xtriage. The presence of seven pseudo-merohedral twin operators were found, two 4-fold and five 2-fold. The twin law giving the lowest R_obs was -h,l,k superimposing on h,k,l. The data were scaled in space group P2 (cell dimensions a = 64.4 Å, b = 105.3, Å, c = 105.5 Å, β = 90.0° degrees) and refined as above with the inclusion of the above two-fold twin law and four crystallographically-independent subunits (residues 2-299 and 2-297) in the asymmetric unit leading to values for R_work of 19.7 % and R_free of 22.9 % (this model has been deposited with PDB ID: 3H5D). However, inspection of the diffraction data and the structural model provided clear evidence for 2₁ screw axes perpendicular to the monoclinic unique b axis. The twinning axis is parallel to the crystallographic two-fold axis (which is a pseudo-4₂ or 2₁ screw axis with significant intensity violations), leading to reflections −k,h,l being overlapped with h,k,l and to the observed pseudo-tetragonal 4/mmm Laue symmetry. The final model, now in space group P22₁2_1, was solved by molecular replacement using MOLREP [63] with the model PDB ID: 3HIJ [36] together with amplitude-based twin refinement with NCS restraints using REFMAC5 [64]. The final model (R_work = 16.5% and R_free = 21.2%) comprises of two monomers (residues A2 – A299 and B2 – B297, and A302 – A311 and B302 – B311), 235 water molecules, 4 potassium ions, 4 glycerol molecules and 2 MES ions. The pseudo-tetragonal symmetry leads to two possible choices of asymmetric unit, one comprising a pair of subunits forming the ‘weak’ dimer interface, the other comprising a pair of subunits forming the ‘tight’ interface – the usual asymmetric unit seen in other tetrameric DHDPS structures. Both were tested, the latter being confirmed as correct. The crystallographic two-fold axis generates the tetramer. Residues 300-301 of chain A and residues 298-301 of chain B were not observed in electron density maps, and as a result of crystal packing, it is not unequivocally clear whether the C-terminal residues 302-311, which were very clearly defined in electron density maps once the twinning and space group problems were solved, indulge in domain swapping to a neighboring tetramer or remain with their own tetramer, However, electrospray ionization mass spectrometry analysis of dissolved crystals of recombinant Sp-DHDPS indicates that the entire protein (residues 1-311) is intact. A summary of the crystallographic data collection and refinement statistics is provided in Table 2.

Coordinates

Coordinates and structure factors for the final model are accessible via the Protein Data Bank (PDB ID: 3VFL).

S. pneumoniae 447A ΔdapA mutant requires (S)-lysine for growth

To validate DHDPS as a promising drug target in Streptococcus pneumoniae, the gene encoding this enzyme (i.e. dapA) was deleted from strain 447A. The ΔdapA mutant was generated by homologous recombination as described in the Materials and Methods. Cultivation of the ΔdapA and wild-type strains on nutrient-rich media, such as HBA, showed that their size and morphology were indistinguishable (data not shown). To determine whether the ΔdapA mutant required (S)-lysine for growth, wild-type and mutant strains were grown in Chemically Defined Medium CDM⁺ and CDM^- (where + and - indicates the presence and absence of 200 µg ml^-1 (S)-lysine, respectively) at 37°C in an atmosphere of 5% CO₂. Analysis of growth rates of wild-type and ΔdapA mutant strains showed that the rate of growth of the mutant in CDM⁺ was not significantly different to that of the wild-type strain (Figure 2). However, the ΔdapA mutant was unable to grow in CDM^- whereas the wild-type strain reached comparable cell density in CDM⁺ and CDM^- (Figure 2). This demonstrates that the dapA gene, encoding DHDPS, is essential for the growth of S. pneumoniae in the absence of lysine.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 2. Growth phenotype of the dapA knockout strain of S. pneumoniae.

Growth of wt S. pneumoniae 447A (control) and the ΔdapA mutant strain in the presence and absence of 20 mM (S)-lysine. Growth was assessed by monitoring the optical density (Relative Absorbance) of duplicate cultures on an hourly basis as described in the Materials and Methods.

https://doi.org/10.1371/journal.pone.0083419.g002

Secondary structure and stability of Sp-DHDPS

Sp-DHDPS was expressed and purified to >95% homogeneity as described previously [43]. The enzyme was subjected to circular dichroism (CD) spectroscopy in aqueous solution to assess the secondary structure of the recombinant product. The CD spectrum (Figure 3A) shows a broad minimum spanning 208 nm to 222 nm, suggesting Sp-DHDPS adopts a mixed α/β secondary structure in solution [16,36,47]. This assertion was confirmed by fitting the CD spectrum of Sp-DHDPS to the CONTINLL algorithm and SP43 database using the CDPRO software suite [48,49]. The nonlinear best-fit demonstrates a significant proportion of α-helix and β-strand (Table 3). The calculated secondary structure composition of Sp-DHDPS is comparable to that of DHDPS enzymes from other bacteria, including E. coli (Table 3). Next we assessed the stability of recombinant Sp-DHDPS in solution by conducting thermal denaturation experiments monitored by CD at 222 nm (Figure 3B). The resulting thermal denaturation profile reveals that the enzyme unfolds via a single transition with an apparent melting temperature (T_M^app) of 72°C (Figure 3B and Table 4). Recombinant DHDPS from E. coli (Ec-DHDPS) also unfolds via a single transition, but with a significantly lower T_M^app of 59°C (Figure 3B and Table 4). These data therefore indicate that Sp-DHDPS is markedly more stable in solution than Ec-DHDPS. We were thus interested in unraveling the molecular mechanism for this enhanced thermostability by assessing the quaternary structure of Sp-DHDPS in solution.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 3. Circular dichroism spectroscopy of Sp-DHDPS.

(A) CD spectrum of Sp-DHDPS () plotted as the molar circular dichroism (Δε) as a function of wavelength. The solid line represents the nonlinear least squares fit using the CONTINLL algorithm and SP43 database within the CDPRO software suite [48.49]. The best-fit resulted in the secondary structure composition reported in Table 3. (B) Thermal denaturation profiles of Sp-DHDPS () and Ec-DHDPS () plotted as mean residue ellipticity (MRE) versus temperature. The apparent melting temperature (T_M^app), or midpoint of the transition between folded and unfolded states, was determined from the ordinate maximum of a plot of the first derivative of the MRE as a function of temperature to yield 59°C for Ec-DHDPS and 72°C for Sp-DHDPS.

https://doi.org/10.1371/journal.pone.0083419.g003

Quaternary structure of Sp-DHDPS in aqueous solution

To gain further insights into the solution stability of Sp-DHDPS, sedimentation velocity studies were conducted in the analytical ultracentrifuge. Absorbance versus radial data profiles at different time points were fitted to a continuous size-distribution, c(s), model [51.52,65,66]. The resulting c(s) distribution for Sp-DHDPS at an initial concentration of 4.5 µM is shown in Figure 4A. The c(s) distribution reveals a single peak with a s_20,w value of 7.2 S that is consistent with a tetrameric species (Table 4) [34,36]. The corresponding c(M) distribution (Figure 4B) confirmed this assertion and shows a single peak with a molar mass of 133 kDa, which closely matches the theoretical mass of the Sp-DHDPS tetramer (135 kDa). Sedimentation studies at a 6-fold lower protein concentration (i.e. 740 nM) were conducted and the resulting c(s) nonlinear least-squares fit also reveals the presence of a single peak (Figure 4B) with a s_20,w value of 7.0 S, consistent with the enzyme existing primarily as a tetramer [26,27,34,36]. These sedimentation velocity analyses suggest that Sp-DHDPS exists as a very stable tetramer in solution. Sedimentation equilibrium experiments were subsequently performed in the analytical ultracentrifuge to examine the strength of subunit interactions and the resulting data at multiple protein concentrations fitted to various equilibrium schemes. Not surprisingly, the optimal global nonlinear least-squares fit was obtained for a dimer-tetramer equilibrium model (Figure 5, solid lines) with a tetramer-dimer dissociation constant (K_D⁴²) of 1.7 nM. Interestingly, the calculated K_D⁴² for Sp-DHDPS is considerably tighter than that obtained for the previously characterized Ec-DHDPS tetramer (Table 4) [34]. We next set out to determine the average hydrodynamic radius of the Sp-DHDPS tetramer in solution using dynamic light scattering.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 4. Sedimentation velocity and dynamic light scattering analyses of Sp-DHDPS.

(A) Continuous sedimentation coefficient, c(s), distribution of Sp-DHDPS at 4.5 µM plotted as a function of standardized sedimentation coefficient (s_20,w) with RMSD and runs test Z values for the best-fit of 0.008 and 6.6, respectively. (B) Continuous mass, c(M), distribution of Sp-DHDPS at a concentration of 4.5 µM plotted as a function of molar mass with RMSD and runs test Z values of 0.008 and 6.6 respectively. (C) c(s) distribution of Sp-DHDPS at 740 nM, with RMSD and runs test Z values of 0.02 and 1.64, respectively. Data were analyzed employing the program SEDFIT [51,52,65,66] using a resolution (N) of 200 and a sedimentation coefficient range of 0.1-12 S or molar mass range 1.0-250 kDa with a P-value of 0.95. Above Panels A-C: Residuals plotted as a function of radial position resulting from continuous size-distribution best-fits shown in panels A-C. (D) A distribution plot of the average unweighted hydrodynamic radii of Sp-DHDPS at a concentration of 59 µM determined using dynamic light scattering (DLS). The results of the distribution plot reveal an average hydrodynamic radius of 4.5 ± 0.2 nm.

https://doi.org/10.1371/journal.pone.0083419.g004

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 5. Sedimentation equilibrium analyses of Sp-DHDPS.

Absorbance at sedimentation equilibrium is plotted as a function of radial position for Sp-DHDPS at an initial concentration of (A) 296 nM; (B) 355 nM and (C) 740 nM. Data was measured at 210 nm and rotor speeds of 10,000 rpm () and 18,000 rpm (). The resulting global nonlinear least squares fit to a dimer-tetramer equilibrium model is shown as solid lines and yielded a K_D^4à2 of 1.7 nM with a global reduced χ² of 0.5. Above Panels A-C: Residuals plotted as a function of radial position resulting from the global nonlinear best-fit analysis to a dimer-tetramer model for data at 10,000 rpm () and 18,000 rpm ().

https://doi.org/10.1371/journal.pone.0083419.g005

The hydrodynamic radius of the Sp-DHDPS tetramer

The quaternary structure and hydrodynamic radius of Sp-DHDPS were studied by dynamic light scattering (DLS) experiments at an enzyme concentration of 59 μM [3.5 × 10⁴-fold above the K_D^4→2 of Sp-DHDPS (Table 4)]. The resulting data were analyzed by the method of cumulants (second-order) and were found to be reproducible between runs and independent of the choice of fit range. Figure 4D shows a regularized distribution fit employing the in-built software (ALV). Analysis of data resulted in a range of hydrodynamic radii that centered on an average value of 4.5 ± 0.2 nm. The average hydrodynamic radius was consistent between runs and was not affected by the choice of fit parameters (correlation function limits or radius limits). By comparison, the Stokes radius of the tetramer calculated from sedimentation velocity studies is 4.3 nm, which is in excellent agreement with the DLS experiment.

Enzyme kinetic properties of Sp-DHDPS

To characterize the catalytic properties of the Sp-DHDPS tetramer, we employed the DHDPS-DHDPR coupled assay [45] at an enzyme concentration of 20 nM, which is ~12-fold greater than the K_D^4→2 (Table 4). Initial rates were measured with fixed pyruvate concentrations of 0.5 mM, 1.0 mM, 2.0 mM, 4.0 mM, 8.0 mM and 16.0 mM and varying ASA concentrations (0-0.48 mM). The resulting data were expressed initially as Lineweaver-Burk plots (Figure 6A), which displayed a characteristic series of parallel lines indicating that Sp-DHDPS follows a Ping-Pong kinetic mechanism [67]. This mechanism has been demonstrated for DHDPS orthologs from other bacterial species, including Ec-DHDPS [45,68,69]. The data were subsequently plotted to produce Michaelis-Menten profiles (Figure 6B) and fitted to bi-substrate kinetic models (with and without substrate inhibition), namely the ternary complex and Ping-Pong mechanism employing ENZFITTER software (Biosoft). The global nonlinear regression analysis yielded a best-fit to a Ping-Pong model with no substrate inhibition (Figure 6B, solid lines) that resulted in a R² = 0.98 and the kinetic parameters reported in Table 5. The K_M of Sp-DHDPS for (S)-ASA is similar that for the E. coli enzyme (Table 5). However, the K_M for pyruvate is 10-fold higher than that for Ec-DHDPS, and the catalytic turnover (k_cat) is 6-fold lower for Sp-DHDPS compared to the E. coli ortholog (Table 5).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 6. Enzyme kinetic profiles of recombinant Sp-DHDPS.

Initial velocity was measured as a function of (S)-ASA concentration. Experiments were conducted at fixed pyruvate concentrations of () 0. 5 mM, () 1.0 mM, () 2.0 mM, () 4.0 mM, () 8.0 mM and () 16.0 mM. (A) Lineweaver-Burk plots showing multiple parallel lines; diagnostic of a ping pong kinetic mechanism [25,67,69]. (B) Michaelis-Menten plots of data shown in A, where solid lines represent the global nonlinear best-fit to a Ping-Pong mechanism (without substrate inhibition) using ENZFITTER software, resulting in a R² = 0.98 and the enzyme kinetic parameters summarized in Table 5.

https://doi.org/10.1371/journal.pone.0083419.g006

Crystal structure of Sp-DHDPS

In order to gain further insight into the stability and kinetic behavior at the atomic level, we next sought to determine the crystal structure of Sp-DHDPS. The collection of high-resolution synchrotron X-ray data has been reported recently [43]. Crystals were pseudo-merohedrally twinned to give the illusion of tetragonal symmetry, with a twin fraction of 0.46 (Table 2). Moreover, the apparent 4₂ screw axis was characterized by significant systematic absence violations, which led after abortive attempts in P4₂2₁2 to solution and refinement in the monoclinic space group P2₁ where there is ambiguity with respect to the twin law. However, with two well-defined 2₁ screw axes perpendicular to the pseudo-tetragonal axis, final refinements proceeded successfully in the orthorhombic space group P22₁2₁ with the twinning axis being parallel to the pseudo-tetrad. Only in this final assignment of crystallographic symmetry did the C-terminal residues 302-311 (see below) become well defined. The structure has been refined at 1.9 Å resolution (Table 2) to an R_work of 16.5 % (R_free of 21.2 %). Two residues, V147 and Y114 in both chains A and B lie in disallowed Ramachandran conformers, but are clearly defined in the electron density. Both residues are located at the ‘tight’ dimer interface (Figure 7) and are held in strained conformations by hydrogen bonding and hydrophobic packing at this interface. The strained conformation of the highly conserved Y114 (Figure 1B) has been observed in all DHDPS structures determined to date, including those from plants [20-36]. However, the strained conformation of V147, which is poorly conserved (Figure 1B), appears to be unique to Sp-DHDPS.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 7. Crystal structure of Sp-DHDPS determined at a resolution of 1.9 Å.

The enzyme crystallizes as a tetramer comprised of four identical subunits labeled A, B, C & D with subunits AB or CD connected via the ‘tight’ dimer interface, and subunits AC or BD connected via the ‘weak’ dimer interface.

https://doi.org/10.1371/journal.pone.0083419.g007

Consistent with the solution studies, Sp-DHDPS forms a homotetrameric structure (PDB ID: 3VFL) that is depicted in Figure 7. Each monomer unit is folded to form an N-terminal (β/α)₈-barrel (residues 2-229) with a C-terminal extension comprised of three helices (residues 232-292). The C-terminal residues 302-311 lie in a groove formed by the ‘tight’ dimer interface (i.e. the interface between chains A & B or C & D in Figure 7). As with other DHDPS structures, including E. coli DHDPS [25,31], the active site of Sp-DHDPS is located within the β-barrel of each subunit. Many of the residues adopt a similar spatial arrangement to that of the Ec-DHDPS structure [PDB ID: 1YXC] [25] (Figure 8A). The key lysine residue, K168, which forms a Schiff base with the pyruvate substrate, is present alongside residues Y140, R145, and G192, which play a pivotal role in the cyclization after reaction with the second substrate (S)-ASA [25,70] (Figure 8A). Two of the residues forming the catalytic triad, which serves to shuttle protons to and from the active site (i.e. Y140 and T50), are also present in a similar orientation to the equivalent residues from the E. coli ortholog (Figure 8A). However, the third residue of the catalytic triad, Y114, undergoes a ~70° rotation in the active site of Sp-DHDPS relative to the equivalent residue of the E. coli enzyme (i.e. Y107), although the functionally important –OH group is similarly positioned. This residue interdigitates across the ‘tight’ dimer interface forming a hydrophobic stack with Y113 from the adjacent monomer (Figure 8B). This 70° rotation serves to change the interaction to a π stacking between the two tyrosine residues.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 8. Comparison of the active sites of Sp-DHDPS (yellow) and Ec-DHDPS (pink).

(A) An overlay of the active site with residues labeled according to Sp-numbering. (B) Image demonstrating the approximate 70° rotation of Y114 in the active site of Sp-DHDPS compared to the equivalent residue in Ec-DHDPS (Y107).

https://doi.org/10.1371/journal.pone.0083419.g008

Relative to the Ec-DHDPS structure (PDB ID: 1YXC) [26], insight into the enhanced stability observed for Sp-DHDPS in solution is revealed by close inspection of the interactions stabilizing the ‘tight’ dimer interface, and in particular, the ‘weak’ dimer interface (Figure 7, interface AC or BD). Analysis of the solvent inaccessible surface areas (SISA), calculated using the Protein Interfaces, Surfaces and Assemblies (PISA) program [71], reveals that the ‘tight’ dimer interface of Sp-DHDPS (Figure 7, interface AB or CD) buries 1350 Ų (Table 6, Figure 9A), slightly more than that calculated (1290 Ų) for the equivalent interface of the Ec-DHDPS structure (Table 6, Figure 9C). Although both enzymes contain the same number of residues (i.e. 38 in total) at this interface, Sp-DHDPS contains a larger proportion of residues forming hydrogen bonds. In contrast, the ‘weak’ dimer interface has a total SISA of 800 Ų (Table 6, Figure 9B), which is significantly greater than the equivalent interface of the E. coli structure (500 Ų) (Table 6 & Figure 9D). The substantially larger buried surface area in Sp-DHDPS is due to a greater number of residues forming contacts at this interface. In addition, it is interesting to note that this enzyme also contains three residues that form salt bridge interactions, a feature that is absent in Ec-DHDPS (Table 6 & Figure 9D).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 9. Comparison of the oligomeric interfaces of DHDPS from S. pneumoniae (panels A & B) and E. coli (panels C & D).

Noncovalent interactions and solvent-inaccessible surface area (SISA) were calculated using the program PISA [71]. The ‘tight’ dimer interface of Sp-DHDPS (panel A) buries 1350 Ų of SISA compared to Ec-DHDPS (panel C), which buries 1290 Ų of SISA. The ‘weak’ dimer interface of Sp-DHDPS (panel B) buries 800 Ų of SISA compared to 500 Ų for Ec-DHDPS (panel D). White, grey and black shaded regions correspond to interfacing, hydrogen bonding and ion bonding residues, respectively.

https://doi.org/10.1371/journal.pone.0083419.g009