ABSTRACT
Bacterial pathogens Neisseria meningitidis and Brucella abortus pose threats to human and animal health worldwide, causing meningococcal disease and brucellosis, respectively. Mortality from acute N. meningitidis infections remains high despite antibiotics, and brucellosis presents alimentary and health consequences. Superoxide dismutases are master regulators of reactive oxygen and general pathogenicity factors and are therefore therapeutic targets. Cu,Zn superoxide dismutases (SODs) localized to the periplasm promote survival by detoxifying superoxide radicals generated by major host antimicrobial immune responses. We discovered that passive immunization with an antibody directed at N. meningitidis SOD (NmSOD) was protective in a mouse infection model. To define the relevant atomic details and solution assembly states of this important virulence factor, we report high-resolution and X-ray scattering analyses of NmSOD and of SOD from B. abortus (BaSOD). The NmSOD structures revealed an auxiliary tetrahedral Cu-binding site bridging the dimer interface; mutational analyses suggested that this metal site contributes to protein stability, with implications for bacterial defense mechanisms. Biochemical and structural analyses informed us about electrostatic substrate guidance, dimer assembly, and an exposed C-terminal epitope in the NmSOD dimer. In contrast, the monomeric BaSOD structure provided insights for extending immunogenic peptide epitopes derived from the protein. These collective results reveal unique contributions of SOD to pathogenic virulence, refine predictive motifs for distinguishing SOD classes, and suggest general targets for antibacterial immune responses. The identified functional contributions, motifs, and targets distinguishing bacterial and eukaryotic SOD assemblies presented here provide a foundation for efforts to develop SOD-specific inhibitors of or vaccines against these harmful pathogens.
IMPORTANCE By protecting microbes against reactive oxygen insults, SODs aid survival of many bacteria within their hosts. Despite the ubiquity and conservation of these key enzymes, notable species-specific differences relevant to pathogenesis remain undefined. To probe mechanisms that govern the functioning of Neisseria meningitidis and Brucella abortus SODs, we used X-ray structures, enzymology, modeling, and murine infection experiments. We identified virulence determinants common to the two homologs, assembly differences, and a unique metal reservoir within meningococcal SOD that stabilizes the enzyme and may provide a safeguard against copper toxicity. The insights reported here provide a rationale and a basis for SOD-specific drug design and an extension of immunogen design to target two important pathogens that continue to pose global health threats.
INTRODUCTION
Superoxide dismutases are master scavengers of reactive oxygen species (ROS), which are unavoidable byproducts of aerobic life. ROS, such as superoxide anion (O2·−), H2O2, and HO·−, play critical and varied roles in biological processes ranging from aging and oncogenesis to pathogenesis and antibiotic action (1–4). ROS act as signals at low levels but as cytotoxins at higher levels (3, 5–8), so virtually all aerobic (and many anaerobic) organisms have evolved defenses for ROS detoxification. The superoxide dismutase enzymes, which catalyze disproportionation of O2·− radicals, are one such important antioxidant defense (9, 10). Three structurally distinct families employ alternate oxidation and reduction of active-site metal ion cofactors (11) (Mn/Fe, Ni, or Cu coupled to Zn) to protect different subcellular compartments (10). Ubiquitous Cu,Zn superoxide dismutase (SOD) family members exhibit a Greek-key β-barrel fold (12, 13) and can exist as monomers, dimers, or tetramers of distinct assemblies, depending on the particular enzyme (12, 14–17). All employ similar mechanisms to achieve their catalytic activity and remarkable diffusion-limited reaction rates (18–22). However, our mechanistic analyses comparing the periplasmic SOD from Photobacterium leiognathi with the cytoplasmic eukaryotic SODs suggested that some functional properties of bacterial (historically termed prokaryotic or P-class) and eukaryotic (E-class) SODs (19) evolved independently. Notably, periplasmic SODs are among several defenses used by pathogenic bacteria to protect themselves from the respiratory burst of their host's innate immune response and thus to support bacterial survival and replication within phagocytes (23–25).
Neisseria meningitidis and Brucella abortus are two major pathogens that continue to threaten public health and welfare. The opportunistic bacterium N. meningitidis colonizes the nasopharyngeal mucosa; its access to the blood can cause life-threatening infections. N. meningitidis is a leading cause of meningitis and septicemia, with an estimated 1.2 million cases of human meningococcal infection reported annually (26) but with nonuniform global distribution rates (27), serogroup prevalences (28), and affected populations (29). Another Gram-negative species, B. abortus, is one of a few Brucella species producing the zoonotic disease brucellosis, characterized by abortion, predominantly in ruminant animals, including cattle, bison, sheep, and goats. Furthermore, although human brucellosis is well controlled in the United States (∼100 cases reported annually), more than half a million new cases are diagnosed globally each year (30). Infection is mostly due to laboratory contact, animal handling, or consumption of tainted meat or dairy products; human manifestations of brucellosis range from flu-like symptoms to arthritis, epididymal/testicular inflammation, endocarditis, hepatitis, and/or meningitis (31). Brucellosis therefore remains a worldwide threat both to human health and to the livestock industry. Preventative strategies against meningococcal disease and brucellosis have met with some recent success. Vaccines against the major N. meningitidis serogroups infecting humans are now available, although their use is not routine, and their long-term efficacy has not been fully established in all populations (32). Similarly, although several B. abortus vaccines are available for cattle and small ruminants, their efficacy is variable and not protective against all Brucella species, with some vaccines complicating serological testing and/or causing abortion in a percentage of animals (33, 34). Further, no vaccine against human brucellosis currently exists (35). Thus, research on understanding the host immune response to N. meningitidis and B. abortus is needed.
Cytoplasmic superoxide dismutases protect aerobic organisms from endogenous O2·−-induced damage to cellular enzymes (3, 4), and periplasmic SODs protect bacteria from extracytosolic damage (36), which, although less well characterized, can arise from O2·− produced endogenously (37) or exogenously by phagosomal NADPH oxidases (38). Unsurprisingly, many pathogenic SODs are potent virulence factors (39–43). Previously, we found that N. meningitidis SOD (NmSOD) deletion mutants were much more sensitive than the wild type (WT) to extracellularly generated O2·− in vitro (44). Further, NmSOD mutant strains had increased phagocytic uptake into monocytes/macrophages (45) and were also less virulent in mice (44). Similarly, B. abortus SOD (BaSOD) deletion mutants were found to be susceptible to O2·−-mediated killing in vitro and by macrophages in mice (46) and exhibited decreased survival during the initial phases of infection (47). Furthermore, our work found that immunization with a synthetic peptide derived from BaSOD was immunoprotective in mice (48). Supporting and extending these previous results, we now show that neutralization of NmSOD with a monoclonal antibody (49) also protected mice during infection.
To better understand the molecular bases of virulence imparted by pathogenic SODs, we solved 7 high-resolution structures from N. meningitidis and B. abortus. By combining crystallography with small-angle X-ray scattering (SAXS) to obtain accurate structures, conformations, and assemblies in solution (50), we found that NmSOD adopts a P-class dimeric assembly, whereas BaSOD functions as a monomer. An auxiliary Cu-binding site spanning the NmSOD dimer interface was found to contribute to enzyme stability and may promote survival of N. meningitidis within phagocytes. The Cu site geometries within the NmSOD and BaSOD structures indicate a mixture of oxidation states that promote SOD function within the phagosome. Our combined structural, mutational, and biochemical analyses show that NmSOD K91 and K94 contribute to electrostatic substrate guidance to the catalytic Cu site. These data further highlighted motifs distinguishing P-class SODs from classic E-class SOD dimers. Collectively, these results expand our understanding of bacterial SODs on the molecular level and suggest testable experimental and therapeutic strategies to target bacteria to prevent and treat pathogenesis.
MATERIALS AND METHODS
Protein expression and purification.NmSOD was overexpressed in Escherichia coli strain QC779 (51) by using the vector pJSK205 (44). Cultures of E. coli in mid-log phase grown in LB medium were induced with 0.4 mM IPTG (isopropyl-β-d-thiogalactopyranoside), and protein expression proceeded overnight. Overexpressed SOD directed to the periplasm was isolated by osmotic shock using 20% sucrose, by precipitation using 65% (wt/vol) ammonium sulfate, and by subsequent centrifugation. The pellet was resuspended in 20 ml H2O and dialyzed against H2O for 4 h at 4°C. The sample was then centrifuged for 30 min (8,000 × g), and CuSO4 was added to the supernatant to achieve a final concentration of 1 mM. The supernatant was then heated to 65°C for 30 min and then centrifuged for 45 min (40,000 × g). The supernatant was dialyzed against 20 mM MES (4-morpholineethanesulfonic acid; pH 5.9) and loaded onto a Poros 20 HS (Applied Biosystems) cation exchange column. Protein was eluted with a linear NaCl gradient, and SOD-containing fractions were pooled and dialyzed into 20 mM Tris-HCl (pH 8.0) and then applied to a Poros 20 HQ (Applied Biosystems) anion exchange column. Protein was eluted with a linear NaCl gradient, and fractions containing SOD were pooled, concentrated, and applied to a HiLoad 16/600 Superdex 75 PG (GE Healthcare) gel filtration column. Peak fractions were pooled, and metal reconstitution with Cu and Zn was carried out by dialysis as described previously (52). BaSOD was purified as described previously (53, 54). SOD activity of the recombinant proteins was measured using a gel-based SOD assay as described previously (11), and active, metal-reconstituted SOD aliquots were frozen in liquid nitrogen and stored at −80°C. NmSOD mutants (E73A, K91Q, K91E, K91Q/K94Q, and K91E/K94E) were constructed in pJSK205 using a QuikChange site-directed mutagenesis kit (Agilent) and purified as described above for WT NmSOD. Homo sapiens SOD (HsSOD) was prepared as described previously (55).
Protein analysis.Prior to SAXS analyses, analytical gel filtration was performed at 4°C on BaSOD (loaded at 9 mg/ml) and HsSOD (loaded at 10 mg/ml) using a Superdex 75 10/300 GL column (GE Healthcare) and 20 mM Tris (pH 7.5) containing 500 mM NaCl with 2% glycerol. Molecular weight standards (Bio-Rad) were run in the same buffer for comparison. A similar elution profile for monomeric BaSOD was also obtained on a Sepharose G-75 column (120 by 2.5 cm) using 10 mM sodium-potassium phosphate (pH 6.4)–150 mM NaCl.
For metal analysis, samples of NmSOD were digested with concentrated nitric acid for 2 h at 65°C and were then diluted to 10% nitric acid with distilled H2O for inductively coupled plasma optical emission spectrometry (ICP-OES) analysis on a Vista-MPX spectrophotometer (Varian). Metal content was determined from 3 independent wavelengths by using standard curves (0.1 to 100 ppm) for Cu and Zn. The protein concentrations of the samples prior to digestion were determined by UV absorbance and were used together with the ICP-OES results to determine the metal/protein ratio for NmSOD samples.
Stability measurements were performed as described previously (56) by titration of purified WT and E73A NmSOD samples and WT HsSOD (final concentration of 20 μM) with guanidine-HCl (Gu-HCl) (from 0 to 7 M in 0.5 M steps) in the presence of phosphate-buffered saline (PBS; pH 7.4), subsequent vortex mixing, and overnight incubation at room temperature, followed by circular dichroism (CD) analysis. The CD sample cell path length was 0.1 cm, and the ellipticities were measured at 218 nm in an Aviv model 2o2SF stopped-flow CD spectrometer in kinetics mode. Three 30-s measurements were averaged to obtain the final values, and the following equation was used to convert the raw CD data into fraction-unfolded (Fu) data for normalization:
where Yn and Yd are the intercepts of the lines through the lower and upper plateau regions (low and high denaturant concentrations), respectively, and Y is the CD value at a given concentration of guanidine-HCl. Data were plotted as the fraction-unfolded value versus the guanidine concentration by using Origin 5.0 (Microcal Software Inc.) and fitted to a sigmoidal curve. Detailed analyses of the free-energy changes were not included, as previous work suggested that, because of metal binding, WT SOD folding and unfolding are not entirely reversible (56).
For pulse radiolysis measurements, all solutions were prepared using ultrapurified (Millipore) distilled water, and reaction rates of SOD with O2·− were measured using the 2-MeV Van de Graaff accelerator at Brookhaven National Laboratory. Dosimetry was calculated using a potassium thiocyanate (KSCN) dosimeter with a molar extinction coefficient of 7,950 M−1 cm−1 for (SCN)2− and a radiolytic yield (G value) of 6.13. The path length was 2 cm, and a radiation dose of 150 to 2,000 rads was used, yielding 1 to 20 μM O2·− per pulse. All activity measurements were carried out using aqueous solutions containing 10 mM formate and 10 mM phosphate at 25°C. The first measurement was made in the absence of EDTA, and subsequent measurements were made after adding 10 μM EDTA. No differences in the disproportionation rates were observed, ensuring that no free Cu was present to confound the results.
The radiolysis of water predominantly yields eaq−, OH, and H. These radicals can be converted to an O2·− radical by the following reactions: eaq− + O2 → O2·−; OH/H + HCO2− → H2O/H2 + CO2−; and CO2− + O2 → O2·− + CO2. The disappearance of O2·− was followed spectroscopically at 260 nm [ε(O2·−) = 1,800 M−1 cm−1 at 260 nm] in the absence and presence of micromolar concentrations of enzyme. EDTA and sodium formate were purchased from Sigma, and monobasic potassium phosphate was produced from Ultrex reagent (JT Baker, Inc.). Solution pH was adjusted by adding H2SO4 (doubly distilled from Vycor; GFC Chemicals Co.) and NaOH (Puratronic; Alfa-Ventron Chemicals).
The activities of BaSOD and bovine erythrocyte SOD (Sigma) were assessed by using a xanthine-xanthine oxidase system as previously described (57).
X-ray crystallography.Crystals of NmSOD (WT and mutants) and BaSOD were obtained using 2 to 2.3 M (NH4)2SO4 (pH 6 to 8.0) and 100 to 200 mM KCl and the hanging drop vapor diffusion method, with the exception of the NmSOD E73A mutant, which was crystallized from 20% polyethylene glycol 3000–50 mM Tris-HCl (pH 8.0). Prior to cryogenic cooling, crystals were soaked in a solution of 20% (vol/vol) glycerol, 75% (wt/vol) saturated (NH4)2SO4, 50 mM Tris (pH 7.5), 50 mM NaCl, 0.1 mM EDTA, and 0.02% NaN3, except for NmSOD E73A, which was soaked in mother liquor plus 20% ethylene glycol.
Diffraction data for all crystals were collected at the Stanford Synchrotron Radiation Laboratory (SSRL). Table S1 in the supplemental material lists the specific beamlines used for each data set. Data reduction and processing were carried out using the HKL package (HKL Research, Charlottesville, VA). Initial phases for NmSOD and BaSOD were obtained by molecular replacement using AMoRE (58) with Actinobacillus pleuropneumoniae SOD (ApSOD; RCSB Protein Data Bank [PDB] code 2APS) and a monomer of Salmonella enterica serovar Typhimurium SodCI (PDB code 1EQW), respectively. The best-scoring repositioned models were refined in CNS (59). Each subunit was treated as a rigid body for the initial refinements, and then the structures were refined using torsion angle simulated annealing at 5,000 K. Following these initial stages, the refinement of the models proceeded through cycles of manual rebuilding in XFIT (60) into sigma-A-weighted 2Fo-Fc and Fo-Fc electron density maps and then through positional and temperature factor refinement. Composite-omit simulated annealing maps were calculated for fitting when needed. The maximum likelihood target function, bulk solvent corrections, and anisotropic temperature factor corrections were used for the refinement cycles in CNS. Where resolution permitted, the structures were then further refined using SHELXL (61) and the same Rfree reflection set along with individual anisotropic temperature factors. PHENIX (62) was used in the final round of refinement for BaSOD.
SAXS.Solution scattering data on WT NmSOD and BaSOD were collected at SIBYLS beamline 12.3.1 at the Advanced Light Source of Lawrence Berkeley National Laboratory (63). Both samples were subjected to size exclusion chromatography using a Superdex 75 10/300 GL column (GE Healthcare) prior to small-angle X-ray scattering (SAXS) analysis. SAXS data were collected on 15-μl samples at 16°C, using automated sample loading from a 96-well format and a Mar charge-coupled-device (CCD) 165 detector with a sample-to-detector distance of 1.5 m. The X-ray wavelength (λ) was 1.03 Å, and the scattering vector q (q = 4π sin θ/λ, where 2θ is the scattering angle) range was ∼0.01 to 0.32 Å−1. X-ray exposures ranged from 0.5 to 5 s (for each sample and matching gel filtration buffer blank), and the scattering value of the buffer was subtracted from that of the protein. NmSOD was analyzed at 1.18, 1.3, and 2.35 mg/ml in phosphate-buffered saline (pH 7.4) containing 2% glycerol, and profiles superimposed well. X-ray exposure data for the 1.3-mg/ml sample were used for merging and further analyses. BaSOD was run at 0.85 and 1.7 mg/ml in a mixture containing 20 mM Tris (pH 7.5) and 500 mM NaCl with 2% glycerol, and data at 1.7 mg/ml were used for further analyses. Data sets were merged with PRIMUS (64). The P(r) function, Rg (real space approximation), and maximum distance (Dmax) were calculated with GNOM (65). DAMMIF (66) was used to generate 10 ab initio models each for NmSOD and BaSOD, and the data were then averaged using DAMAVER (67). Further details are tabulated in Table S2 in the supplemental material. FoXS (68) was used for theoretical profile computation and data fitting.
Molecular-image rendering and modeling.PyMOL (Schrodinger, LLC) and Chimera (69) were used for generating structural images. The ClustalOmega server (http://www.ebi.ac.uk/Tools/msa/clustalo/) was used for sequence alignment. Normal mode analysis (NMA) models were generated using default settings on the elNémo server (70).
Antibody, bacterial strains, and animal studies.Mouse monoclonal antibody HD1 was previously described (49). N. meningitidis MC58 (B:15:P1.7,16-2, ST-74, cc32) wild type and sodC knockout strains were cultured as described previously (44), and Western blotting was used to detect binding of HD1 to N. meningitidis protein extracts separated by denaturing or nondenaturing gel electrophoresis. Mice were challenged with 2 × 106 CFU intraperitoneal N. meningitidis bacteria coadministered with HD1 antibody or administered without antibody, as described previously (44). Survival was monitored for 4 days postinfection. All animal-related studies were conducted under guidelines approved by the IACUC of the USDA National Animal Disease Center or by the Home Office, United Kingdom.
NmSOD deletion mutants and HD1 epitope analysis.NmSOD deletion constructs were constructed by PCR. The encoded protein sequence of the C-terminal deletion mutant (lacking 9 residues) was MKYLLPTAAAGLLLLAAQPAMAHGHGHGHGHGHEHNTIPKGASIEVKVQQLDPVNGNKDVGTVTITESNYGLVFTPDLQGLSEGLHGFHIHENPSCEPKEKEGKLTAGLGAGGHWDPKGAKQHGYPWQDDAHLGDLPALTVLHDGTATNPVLAPRLKHLDDVRGHSIMIHTGGDNHSDHPAPLGGGG, and that of the internal-deletion mutant (lacking 71 residues) was MKYLLPTAAAGLLLLAAQPAMAHGHGHGHGHGHEHNTIPKGASIEVKVQQLDPVNGNKDVGTVTITESNYGLVFTPDLQGLSEGLHGFHIHENPSCEPKEKEGKLTAGLGAGGHWDPRMACGVIK.
Deletion mutants were expressed in E. coli as described above. To confirm gene expression, RNA was prepared, and reverse transcription-PCR (RT-PCR) was performed using oligonucleotide primers for the respective sodC genes. As previously reported, immobilized metal affinity chromatography can be used to purify histidine-rich SOD samples (71). Whole-cell lysates were applied to cobalt-Talon resin (Clontech), and activity of the eluted NmSOD samples was assessed as described previously (44). Western blotting was used to detect binding of HD1 to deletion mutants.
Protein structure and SAXS data accession numbers.The coordinates described in this paper were deposited into the PDB under the following codes: for NmSOD, 2AQN; for NmSOD E73A, 2AQP; for NmSOD K91Q, 2AQR; for NmSOD K91E, 2AQQ; for NmSOD K91Q/K94Q, 2AQT; for NmSOD, K91E/K94E, 2AQS; and for BaSOD, 4L05. SAXS data were deposited into the BIOISIS repository (http://bioisis.net) using codes NMSODP and BASODP.
RESULTS
Anti-SOD antibody provides protection from N. meningitidis infection.To test the hypothesis that NmSOD is a virulence factor that protects meningococci from free-radical-mediated host defenses, we examined the ability of a monoclonal anti-SOD antibody to passively protect mice against N. meningitidis challenge. Passive immunity provides immediate rather than delayed protection against infection and is currently used for postexposure prophylaxis for chicken pox (caused by varicella-zoster virus) and rabies in humans (72, 73). In our experiments, mice were challenged with a dose of 2 × 106 CFU of N. meningitidis (MC58), with or without coadministration of the monoclonal NmSOD-reactive HD1 antibody (49), and subsequently monitored for survival. Our results demonstrated that this NmSOD-neutralizing antibody conferred a robust level of passive protection, with a survival rate of 100% in HD1-treated mice at 4 days postinfection (Fig. 1).
NmSOD-reactive monoclonal antibody HD1 protects mice from N. meningitidis serogroup B infection. Mice were challenged with bacteria with (HD1) or without (control) coadministration of the HD1 antibody. As indicated by the survival data, HD1 conferred passive protection against experimental infection. d, days.
NmSOD conserves P-class dimer assembly.To better understand potential determinants of virulence for NmSOD, we solved the structure of the enzyme to 1.4-Å resolution (see Table S1 in the supplemental material; Rwork = 12.8%, Rfree = 19%). Each subunit of the NmSOD homodimer conserves the SOD Greek-key β-barrel fold (Fig. 2 and 3); the root mean square deviations (RMSDs) of Cα atoms between the NmSOD and bacterial SOD structures from A. pleuropneumoniae (PDB code 2APS) (74), Haemophilus ducreyi (PDB code 1Z9N) (75), P. leiognathi (PDB code 1YAI) (19, 76, 77), Salmonella Typhimurium (SodCI; PDB code 1EQW) (78), and Yersinia pseudotuberculosis (PDB code 2WWO) were all <1 Å. As suggested by comparative structural analyses of bacterial SOD sequences (see Fig. S1) (74), subunit interface residues characteristic of dimeric P-class SODs, together with the ring-like assembly of buried water molecules (Fig. 2A and 3B), were conserved in NmSOD (Fig. 3C). The surface buried by the dimer interface was calculated with PISA to be ∼915 Å2. This P-class dimer is the best-represented crystallographic assembly for bacterial SODs in the Protein Data Bank thus far (see Table S3).
Three-dimensional X-ray structures of NmSOD and BaSOD. (A) NmSOD is a P-class homodimer of conserved Greek-key β-barrel SOD subunits (β-strands in gold and α-helices in green). Buried interface water residues are depicted as red spheres, disulfide positions by S-S, and active-site Cu and Zn ions by bronze and silver spheres, respectively. (B) BaSOD is a monomer with the same conserved SOD fold (β-strands in blue and α-helices in purple). (C) Closeup of NmSOD (top) and BaSOD (bottom) active sites with metal-binding histidine and aspartic acid ligand side chains shown and the Cu ligands labeled. The redox-active Cu site and bridging histidine are modeled in two positions (oxidized and reduced) in BaSOD. (D) Experimental SAXS data (gold and blue) fit SAXS profiles (red and black) simulated from the crystallographic assemblies (NmSOD dimer and BaSOD monomer, respectively). Residuals (experimental/model intensity ratio) are depicted below profiles. (E) Real-space electron pair distributions for NmSOD and BaSOD reveal significant differences in overall shapes and dimensions of these two enzymes in solution, consistent with NmSOD P-class dimer assembly and monomeric BaSOD. (F) X-ray structures of NmSOD (left) and BaSOD (right) docked into ab initio envelopes modeled from SAXS data show agreement between crystalline and solution X-ray data. A.U., arbitrary units.
Comparative structural assemblies of SODs and SOD homologs are linked to sequence motifs. (A) Distinct quaternary arrangements of bacterial and eukaryotic SODs and a SOD homolog. Top and side views (relative to the β-barrel) depict Cα traces for 4 distinct, representative dimers, superimposed using one reference subunit each: P-class NmSOD (PDB code 2AQN), loop class (L-class) M. tuberculosis SOD (PDB code 1PZS), and E-class human SOD (HsSOD; PDB code 1PU0) dimers and a crystallographic B. subtilis SOD-like dimer (PDB code 1S4I), although this homolog is monomeric in solution. For comparison, the buried dimer interface surface areas for SOD homologs were computed with the PISA webserver (http://www.ebi.ac.uk/msd-srv/prot_int/cgi-bin/piserver): NmSOD, 915 Å2, M. tuberculosis SOD, 1,636 Å2, HsSOD, ∼704 Å2, and the B. subtilis SOD-like homolog, ∼347 Å2. PISA predicts B. abortus SOD to be monomeric. However, PISA also predicts the crystallographic BsSOD dimer to be stable. (B) NmSOD dimer structure and surface, bisected at the dimer interface, highlight P-class dimer-specific residues (red), which are located around the perimeter of the water-filled cavity buried in the dimer interface. Red residues correspond to red sequence motifs depicted in panel C. The star denotes the position of the NmSOD-specific cavity adjacent to the auxiliary Cu site (gold sphere). (C) Sequence alignment representing 4 SOD assembly classes: BaSOD (P-class monomer), NmSOD (P-class dimer), M. tuberculosis SOD (L-class dimer), and HsSOD (E-class dimer). Bacterial sequences start at the site of signal peptidase cleavage. Conserved sequence motifs (magenta, except where indicated) include the leucine “cork,” (replaced by a single residue insertion in fungi) starting the conserved LXXGXHG pattern, Cu ligands (gold H or gold/silver bridging H), Zn ligands (gold/silver bridging H or silver D/H, missing in some L-class SODs), an aspartic acid that indirectly bridges Cu and Zn ligands via hydrogen bonds, and the RX(A/V)CGV(I/V) motif (near the C terminus), which includes both the conserved active-site arginine (italic R) and the invariant second cysteine of the conserved disulfide bond. Discriminating features of E-class SODs include conserved residues involved in electrostatic guidance (indicated by purple italics), the first cysteine of the conserved disulfide bond (light green C) located 6 residues before a bridging histidine, and an aromatic or hydrophobic interface residue (yellow F, followed by G) located 2 residues after the second Cu ligand (Cu2). Discriminating features of bacterial SODs include distinct residues implicated in electrostatic guidance (purple italics in boxed sequence), the first cysteine of the conserved disulfide bond (dark green C) located 5 residues after Cu2, an insertion ∼7 residues after Cu2 in P-class SODs, and an insertion ∼10 residues after the fourth Cu ligand (Cu4) in L-class SODs. Discriminating features in P-class dimers but not monomers (red lettering) include a semiconserved YG motif, located 11 residues N terminal to the conserved leucine cork; a hydrophobic residue (typically V or L; V in NmSOD) 8 residues before the leucine cork; a hydrophobic residue, preceded by glycine, replaced in monomers with a residue capable of hydrogen bonding, located ∼4 residues after the cork leucine (L in NmSOD); and an aromatic tryptophan or tyrosine in P-class SOD dimers 7 to 8 residues before the conserved aspartate (Zn ligand in many SODs) replaced by a charged residue in other classes (W in NmSOD). Protective (red bar) and nonprotective (orange bars) peptide immunogens (48) are mapped below the BaSOD sequence. Approximate positions of conserved secondary structural elements are denoted below sequences as arrows (β-strands) and coils (α-helices). The blue arrows denote P-class secondary structure in bacterium-specific insertions, and the turquoise coil denotes an L- and P-class-specific secondary structure feature.
To determine whether the biological state of the enzyme within the crystal is accurately represented, we used SAXS, which allows determination of conformations and assemblies for macromolecules in solution (79, 80). NmSOD SAXS data agreed with the scattering profiles predicted from the crystallographic dimer, with a χ2free value (80) of 1.83 (Fig. 2D). The SAXS-derived electron pair distribution [P(r) plot] (Fig. 2E) and molecular envelope (Fig. 2F, left) models further indicated an elongated shape consistent with the crystal structure. SAXS measurements also indicated a particle density marginally lower than average (81) for NmSOD (see Table S2 in the supplemental material), with small deviations in the profile fitted for scattering angles greater than q = 0.17 (Fig. 2D). These observations, which are also supported by results of normal mode analyses (see Fig. S2 in the supplemental material) and by elevated values for the atomic displacement parameters (ADP; B-factors), suggested considerable protein flexibility in NmSOD, particularly within bacterial SOD-specific loop regions (see Fig. S3).
Electrostatic guidance in NmSOD speeds enzymatic activity.The crystal structure of NmSOD revealed that seven conserved, active-site residues mediate Cu (histidines H79, H81, H104, and H160) and Zn (H104, H113, H122, and aspartic acid D125) coordination (Fig. 2C; see also Fig. S1 in the supplemental material). The catalytic Cu site has trigonal planar ligation geometry characteristic of reduced Cu(I), as was also found in structures of E-class yeast SOD (PDB code 2JCW) (82) and P-class A. pleuropneumoniae SOD (PDB code 2APS) (74), among others. The average Cu(I)-ligand bond lengths (± standard deviation) for the 3 NmSOD subunits in the asymmetric unit were 2.0 ± 0.0, 2.0 ± 0.1, and 2.0 ± 0.1 Å for H79, H81, and H160, respectively. The separation distance between Cu(I) and H104 of 3.1 ± 0.1 Å indicates a broken bond to the bridging histidine, which ligates both metal ions in the oxidized enzyme. The Zn site exhibits a characteristic, distorted tetrahedral geometry, with an average H104-Zn bridging distance of 2.0 ± 0.1 Å.
Analyses of SOD structures and sequences suggested that residues promoting electrostatic guidance in bacterial SODs lie across the active-site channel from the functionally equivalent electrostatic loop (linking β-strands 7 and 8) of eukaryotic SODs (19, 74). The functionally equivalent NmSOD residues predicted from our structural alignment (see Fig. S1 in the supplemental material) include K91, E92, and K94, located in a bacterial SOD sequence insertion (relative to eukaryotic SODs) (Fig. 3C), forming a protruding β-hairpin at the ends of the P-class dimer (Fig. 2A). We constructed mutants to either neutralize (K91Q and K91Q/K94Q) or reverse (K91E and K91E/K94E) one or both charges and then used pulse radiolysis to evaluate the rate constants over a wide pH range (Fig. 4A). We found that while charge neutralization at residue 91 alone did not substantially affect the rate constant, the K91Q/K94Q double mutant reacted 1.5-fold more slowly than the WT. Moreover, reversing the charge at one or both residues decreased the rate constant substantially: at pH 7, the rate constants for K91E and K91E/K94E dropped 2.5-fold and 5-fold, respectively, relative to that for the WT (Fig. 4A). Mirroring analogous studies on the HsSOD enzyme (18), these results confirmed the involvement of these two lysine residues in the electrostatic attraction of the O2·− substrate to the active-site Cu ion in NmSOD and suggest a greater contribution for K94 than was seen for the equivalent residue in the monomeric E. coli enzyme (83). We also determined structures of these NmSOD electrostatic mutants at resolutions ranging from 1.65 to 1.8 Å (see Fig. S3 and Table S1). Apart from mutation site alterations (most notably in K91E/K94E), slight subunit rotations, and modestly increased ADPs (see Fig. S3B), all structures were similar to WT structures (see Fig. S3A). These findings suggest that fine-tuning of electrostatic guidance is not accomplished through major structural rearrangements in NmSOD and also confirm that the active-site geometry is maintained in the mutants described above.
Role of NmSOD residues in Cu binding and activity. (A) Identification of electrostatic guidance residues in NmSOD. The reaction rate constant for NmSOD electrostatic mutants as a function of pH was evaluated using pulse radiolysis. Whereas the E73A mutation has no effect on the reaction rate constant, reversing or neutralizing the charge at both proposed electrostatic residues, K91 and K94, caused a substantial decrease in the reaction rate constant. kobs, observed reaction rate constant. (B) NmSOD E73 contributes to protein stability, as assayed by monitoring guanidine-HCl unfolding of WT and E73A NmSOD, and of WT HsSOD, with circular dichroism (CD) spectroscopy. The unfolding midpoint for the E73A mutant is approximately 0.6 to 0.7 M lower than that for WT NmSOD. Asterisks denote data points omitted from the fitting. (C) Closeup view looking down the 2-fold dimer axis of WT NmSOD (top). The E73 and H133 side chains from each subunit of the dimer are positioned with tetrahedral coordination geometry around the Cu ion. Auxiliary Cu binding was abrogated in the E73A mutant (bottom), wherein only solvent molecules are observed at the equivalent site. (D) Superposition of NmSOD and heme-bound HdSOD (PDB code 1Z9P) reveals parallel structure and dimer assembly. (E) Closeup view of the area outlined by the dashed box in panel D revealing similarities in Cu binding by NmSOD E73 and H133 compared to heme binding by HdSOD H64 and H124.
A stabilizing, auxiliary copper site bridges the NmSOD dimer interface.Metal sites are key contributors of activity and stability in SODs. Surprisingly, ICP-OES analyses of NmSOD indicated a stoichiometry of 1.5 Cu ions per subunit, matching one in each active site and one additional Cu ion per dimer. Correspondingly, in the X-ray structures (Fig. 2 and 4), we observed a unique cross-dimer Cu-binding site, evidenced by a strong (>10σ) electron density peak in omit maps (Fig. 2A; see also Fig. S4B, C [top], and D in the supplemental material). This auxiliary Cu is tetrahedrally ligated by H133 and E73 from each subunit (Fig. 4C [top] and 5A), implying an oxidized Cu(II) state. Thus, the NmSOD structure appears to contain both oxidized and reduced Cu sites, suggesting that the geometry of these sites tunes their redox activity; the distorted active-site Cu ion geometry evolved to promote redox chemistry for catalysis, whereas the less-constrained geometry of the auxiliary site allows the Cu ion to remain oxidized, despite equivalent exposure to reduction by free electrons during the X-ray diffraction data collection. Interestingly, structural superposition of NmSOD with H. ducreyi SOD (HdSOD), which contains a unique heme-binding site (75, 84), revealed that the auxiliary Cu and heme Fe align, forming an unusual cross-dimer bridge (Fig. 4D and E).
Potential roles of NmSOD auxiliary Cu site. (A) Overview (top) and closeup views (bottom) of the binding pocket below the NmSOD auxiliary Cu site. Surface rendering suggested that the pocket volume (124.6 Å3) could accommodate small molecules. (B) Suggested mechanisms for virulence utilized in NmSOD upon phagocytosis. NmSOD counters reactive oxygen species through its disproportionation activity and buffers toxic Cu influx through its cross-dimer Cu binding site.
To probe the functional contributions of the auxiliary Cu in NmSOD, we constructed the E73A point mutant, which removed 2 ligands from the cross-dimer Cu-binding site. We found no major oligomeric state (see Fig. S4A in the supplemental material) or catalytic differences between WT and E73A NmSOD, aside from a slight rate increase for E73A NmSOD above pH 8.5 (Fig. 4A). Next, to assess whether the auxiliary Cu-binding site affects protein stability, we used circular dichroism (CD) spectroscopy to monitor NmSOD unfolding under conditions of increasing concentrations of guanidine-HCl (Gu-HCl) (Fig. 4B). Generally, the unfolding curves for WT and E73A NmSOD were less steep than that for WT HsSOD, suggesting somewhat less concerted and cooperative unfolding for the bacterial enzyme. WT NmSOD and HsSOD displayed similar unfolding characteristics, with the NmSOD unfolding midpoint (4.3 M Gu-HCl) marginally greater than that of HsSOD (3.9 M). E73A unfolded at a slightly lower concentration of guanidine-HCl (3.6 M), indicating a modest loss of stability upon removal of Cu-binding E73. Unfolding for E73A also began at lower Gu-HCl concentrations than for either WT enzyme, and complete unfolding of WT NmSOD required Gu-HCl concentrations higher that those required for WT HsSOD or E73A NmSOD. Overall, these results suggest that, although the auxiliary, dimer-spanning Cu-binding site is not necessary for dimer formation, it stabilizes the NmSOD assembly.
To understand how E73 imparts stability at the atomic level, we determined the crystallographic structure of E73A NmSOD to 1.3-Å resolution (Fig. 4C, bottom; see also Table S1 in the supplemental material) (Rwork = 13.5, Rfree = 18.3). We found that the E73A NmSOD structure was generally similar to the WT structure but with no Cu bound in the vicinity of A73 (Fig. 4C, bottom). Instead, Cu ions appeared to mediate crystal contacts between H148 of one E73A NmSOD dimer and its symmetry mate (see Fig. S4C). In contrast, mutation of electrostatic guidance residues did not affect the position of the auxiliary Cu ion in NmSOD (see Fig. S3B), despite differing crystal symmetries and solvent contents (see Table S1). These results suggest that the decreased stability arises from the lack of Cu binding rather than from gross structural differences.
BaSOD has a conserved active site but monomeric assembly.Sequence alignments suggested that BaSOD, in contrast to NmSOD, lacks canonical P-class dimer interface residues (see Fig. S1 in the supplemental material) (74), and nuclear magnetic resonance (NMR) results suggested that BaSOD was a monomer (53). Informed by these observations, we proceeded to solve the high-resolution X-ray crystallographic structure of BaSOD to 1.1-Å resolution (see Table S1) (Rwork = 11.3, Rfree = 13.8). Notably, the active-site electron density of BaSOD revealed a mixture of oxidized and reduced structural conformations, reflective of the enzyme mechanism, so we modeled 2 positions each for the Cu ion and its bridging histidine (Fig. 2C, bottom). In the oxidized state, the bridging H73-Cu(II) bond is 2.0-Å long, and upon reduction, active-site movements shift the H73 side chain 3.1 Å away from Cu(I) such that H73 can form a 2.5-Å hydrogen bond with a bound water molecule within the active site. The bond distances between the 2 alternate H73 conformations and the Zn ion are 1.9 and 2.1 Å. For the remaining ligands, the Cu(I)-ligand bond lengths are 2.0, 2.0, and 2.1 Å for H48, H50, and H128, respectively, and the Cu(II)-ligand bond lengths are 2.1, 2.2, and 2.2 Å for H48, H50, and H128, respectively. Like NmSOD, BaSOD retains the overall Greek-key β-barrel fold conserved in other SODs (Fig. 2B) (85), albeit with only one BaSOD molecule contained within the asymmetric unit.
BaSOD is active as a monomeric enzyme.BaSOD exhibited activity similar to that of the well-studied bovine SOD dimer (see Fig. S5A in the supplemental material), even though the crystal structure suggested that BaSOD is monomeric. To confirm this stoichiometry, we used both gel filtration chromatography (see Fig. S5B) and SAXS (Fig. 2D and E) to ascertain the monomeric biological assembly of the enzyme in solution. The SAXS profile (Fig. 2D) revealed that the BaSOD monomer in the crystallographic asymmetric unit matches the shape and assembly in solution (χ2free value of 1.30) and is distinct from the dimeric NmSOD SAXS profile; the P(r) plot and ab initio envelopes for BaSOD in solution (Fig. 2E and F) indicated a smaller, more globular shape that fit the monomeric crystallographic structure. Although diverse SOD enzymes retain the same subunit fold and Cu-mediated chemistry, their quaternary assemblies differ considerably, notably in bacteria (Fig. 3; see also Table S3). Although SOD monomers are less well represented than P-class dimers in the PDB, the BaSOD structure is similar to those of monomeric homologs E. coli SOD (PDB code 1ESO [86], with an RMSD of < 0.6 Å), S. enterica SodCII (PDB code 2K4W [87], with an RMSD of ∼1.4 Å), and Bacillus subtilis SOD-like protein (PDB code 1S4I [88], with an RMSD of ∼1.08), with discrepancies predominantly in loop regions. Overall, monomeric SODs are stable and active, despite the prevalence of their dimeric counterparts.
Epitope mapping onto NmSOD and BaSOD structures.The crystal structures and resulting structure-based sequence alignments of NmSOD and BaSOD enabled us to investigate the determinants of virulence within these enzymes. The anti-SOD antibody HD1 (49) bound strongly to NmSOD in denaturing or nondenaturing Western blots of N. meningitidis protein extracts, and we established that no cross-reactive meningococcal products bound HD1 by using a sodC knockout strain. Given the cross-reactivity of HD1 to several known bacterial SODs (49) and the high sequence conservation in the C terminus, we hypothesized that this region might contain the HD1 epitope. To test this prediction, we created 2 NmSOD deletion mutants (see Materials and Methods): a 9-residue C-terminal deletion mutant (which disrupts the disulfide loop) and an internal-deletion mutant that retained this C-terminal sequence (see Fig. S6 in the supplemental material). The mutant proteins were expressed in E. coli and were then tested for SOD activity and HD1 reactivity. Both proteins lacked disproportionation activity, but only the internal-deletion mutant bound HD1 in Western blots. These binding results suggest that the C-terminal residues contribute to the NmSOD epitope recognized by the HD1 antibody, which protected mice from succumbing to N. meningitidis infection (Fig. 1; see also Fig. S6). The 9-residue C-terminal NmSOD sequence (PRMACGVIK) includes 6 residues conserved across domains of life, as well as variable residues (XRXACGVIX) that may dictate the reactivity of HD1. In NmSOD, the first and third variable residues of this sequence flank the invariant active-site arginine to form a “PRM” motif shared by HD1-reactive (49), dimeric SODs from Actinobacillus and Haemophilus. In contrast, the monomeric BaSOD and E. coli SODs do not react with HD1 (49) and lack the “PRM” motif within the proposed C-terminal epitope of HD1 (see Fig. S6).
To understand the structural basis for protective peptide immunogens in BaSOD, we mapped the peptide immunogens we had previously tested (48) onto our structure (see Fig. S7 in the supplemental material). The protective BaSOD peptide (residues 56 to 67) (Fig. 3C, underlined in red) occurs in the loop insertion that is specific to the bacterial SODs (Fig. 3C, boxed) and that contains electrostatic guidance residues K60, D61, and K63. Comparison of this loop in the bacterial and HsSOD structures (and, by analogy, murine SOD structures) revealed an extended protrusion in BaSOD (and NmSOD) (see Fig. 6). Therefore, antibody binding to this BaSOD epitope would be expected to disrupt electrostatic guidance and/or sterically hinder access to the active site. In contrast, the 2 nonprotective peptides span residues 117 to 131 and 130 to 143 (see Fig. S7, orange). The former encompasses all of strand β7 (numbered going around the β-barrel [74]) and part of the core β-barrel framework forming the back wall of the active site (containing Cu ligand H128). This epitope is solvent exposed only partly; therefore, an immune response to the intact protein might not be expected. The latter nonprotective peptide encompasses part of the loop connecting strand β7 to strand β8 (the electrostatic loop of eukaryotic SODs). This surface-exposed loop is adjacent to the active site but is not critical to the enzymatic functioning of bacterial SODs, so antibody binding may impact substrate entry only modestly. These BaSOD peptide vaccines were developed based on hydrophobicity, flexibility, and antigenicity predictions, in the absence of any structural information (48). The high-resolution structures presented here provide detailed, specific frameworks to improve targeting for the development of more-effective vaccines.
DISCUSSION
Structural insights differentiating bacterial and eukaryotic SODs.The oligomeric state of a protein can control its physicochemical properties, behavior, and biological interactions and thus is key for informed design of drugs and vaccines. For example, HsSOD polymers genetically engineered for therapeutic use as anti-inflammatories exhibited a desirable extended serum half-life (89); in contrast, loss of assembly specificity in mutant HsSODs was associated with toxic aggregation in the degenerative motor neuron disease amyotrophic lateral sclerosis (ALS) (55, 56). Many distinct assemblies have been found for the SOD subunit fold (Fig. 3A): in prokaryotes, the P-class monomer and dimer, an alternative extended-loop dimer found in Mycobacterium tuberculosis SodC (90) (here termed loop or L class), and the B. subtilis SOD-like dimer (88) (see Table S3); in eukaryotes, a fungal monomer (91), the metazoan E-class dimer (12), and the dimer-of-dimers tetramer of extracellular SOD (17), as well as larger “dogbone” (12) and nanofilament (92) assemblies of HsSOD dimers. Determining the natural biological assembly state from a crystallographic structure can be challenging, especially in the case of SOD proteins, which display these different biological assemblies and many complex crystal packing arrangements. Therefore, complementary characterization in solution is crucial for accurately defining the physiological assembly state of a protein (93, 94), as we report here for the architecturally distinct virulence factors NmSOD and BaSOD (Fig. 2).
Exploitable physiological differences between bacterial and eukaryotic SODs stem from evolutionary structural dissimilarities. Here, on the basis of the new structures that we determined, we identify sequence motifs that discriminate among SOD homologs, assign stoichiometry, and connect gene to function (Fig. 3C; see also Fig. S1 in the supplemental material). Two conserved motifs identify most SODs: a C-terminal RX(A/V/S)CGV(I/V) sequence (Fig. 3C; see also Fig. S1) and the active-site histidine pattern, notably the first 2 Cu-ligating histidines separated by a single hydrophobic residue. So, how are bacterial SODs distinguished from those of eukaryotes? First, only E-class enzymes conserve an aromatic (or hydrophobic) interface residue (F50 in HsSOD) followed by glycine, positioned 2 residues after the second Cu-ligating histidine, as well as a salt-bridging arginine (R79 in HsSOD), positioned immediately before the third Zn-binding histidine. Another feature is the C-terminal shift of the first disulfide bond cysteine, arising from a disulfide loop insertion in bacterial SODs. In P-class SODs, this cysteine is positioned 5 residues after the second Cu-ligating histidine, while in E-class SODs, it is 6 residues before the third Cu-ligating (bridging) histidine (Fig. 3C). In bovine and HsSODs, mutation of oxidation-prone free cysteines (near the N terminus and Greek key connections) increased enzyme stability (12, 95). Notably, the P-class SODs lack free cysteines, which are selected against in the oxidizing periplasm (96, 97); this feature may contribute to the stability of NmSOD being higher than that of HsSOD (Fig. 4B). These insights extend our previous work detailing the molecular basis of the Greek key β-barrel evolution (98).
Given the diversity among microbial SODs, we also reevaluated sequence and structural features characteristic of their different assembly states. L-class bacterial SODs are clearly differentiated from their P-class counterparts by a large insertion between the last 2 β-strands, but how are P-class dimers and monomers distinguished? P-class dimer interfaces generally conserve 4 buried aromatic or hydrophobic residues encircling a buried water ring (19, 74–76) (Fig. 3B and C, red): (i) a lateral aromatic tryptophan or tyrosine (19, 74, 76, 78); (ii) its cognate hydrophobic residue (19, 74, 76, 78), typically V or L, in the opposing monomer; (iii) a YG motif present in the majority of P-class dimers, which forms hydrophobic packing interactions and a hydrogen bond with an adjacent D at the bottom of the interface (which variably interacts with residues adjacent to the YG) (19, 74, 76, 78); and (iv) a hydrophobic amino acid, which contributes to a hydrophobic cluster toward the top of the interface through interactions with 1 (19, 76, 78) or more of its cognate residues in the opposing monomer. This cluster is located just below the auxiliary copper site in NmSOD (Fig. 3B). In P-class monomers, however, these 4 conserved residues are frequently replaced with hydrophilic or hydrogen-bonding residues expected to destabilize hydrophobic interface packing (86), which was echoed experimentally by characterization of monomer-dimer hybrid proteins (99). These discriminating features identified from SODs of known structure suggest that the assembly class is predictable from sequence data (e.g., for uncharacterized archaeal SOD-like sequences or SODs from bacterial and eukaryotic viruses that conserve P-class [41] and E-class [100, 101] features, respectively). Identification, testing, and refinement of these predictive motifs show promise for distinguishing the SODs of pathogens from those of their hosts, for therapeutic targeting applications.
Significance of the auxiliary Cu site in NmSOD.The need to control ROS has been a major driver of pathogenic and eukaryotic evolution, with the extreme example of glutathione peroxidase incorporation of selenocysteine, via an altered stop codon (102). Cu-containing SODs evolved to protect against rising oxygen levels (103), thus allowing bacterial SODs to serve also as virulence factors protecting pathogens from ROS-mediated host defenses. Host phagocytes use an arsenal of strategies to combat bacterial invaders during engulfment and the phagosomal maturation stages of infection (Fig. 5B). During the latter process, toxic levels of phagosomal Cu may contribute to bactericidal activity (104), in part through ROS production and the resulting oxidative damage (105). N-terminal histidine- and/or methionine-rich Cu(I/II) and/or Zn(II) binding motifs in H. ducreyi and A. pleuropneumoniae SODs (71, 106, 107) may facilitate Cu acquisition or active-site transfer under conditions of scarcity (106) and may also serve to protect these pathogens from phagosomal Cu toxicity. Likewise, auxiliary metal (nickel and bismuth)-binding sites in other bacterial proteins are implicated in metal buffering, sequestration, and toxicity (108). The novel metal site at the NmSOD interface was unexpected, but its presence is consistent with observations indicating that microbial metalloproteomes are still incompletely characterized (109). These findings further our understanding of metal-buffering properties in key bacterial proteins.
Interestingly, the auxiliary NmSOD Cu site corresponds to the heme-binding site in HdSOD (75, 84) (Fig. 4D and E), with HdSOD ligands H64 and H124 positionally equivalent to NmSOD E73 and H133, although the HdSOD heme is bound asymmetrically. Since H. ducreyi cannot synthesize heme and must obtain it from its host, HdSOD was proposed to be involved in heme metabolism as a sensor or buffer against heme toxicity (75). On the basis of these similarities, we tested heme binding to NmSOD in vitro by using the same procedure (84) but found no evidence of bound heme. The auxiliary Cu- and heme-binding residues are not generally conserved in other SODs (see red squares in Fig. S1 in the supplemental material), highlighting the uniqueness of this interface region as a potential antimicrobial target. The solvent-exposed cavity (Fig. 3B and 5A) formed between the auxiliary Cu site and the rest of the P-class dimer interface makes an attractive target for small-molecule binding. For instance, an inhibitor targeted to this cavity could disrupt binding of the auxiliary Cu, thus compromising NmSOD stability. CO or NO binding to the heme in HdSOD did in fact induce conformational rearrangements of that enzyme, including altered active-site accessibility (110). Similarly, active-site accessibility and catalysis were affected by interface substitution mutants in the P. leiognathi SOD P-class dimer (111). The slightly extended pH range over which NmSOD E73A was active (Fig. 4A) suggests that NmSOD interface modulation might also alter catalytic activity through allosteric effects. Thus, our NmSOD structures (Fig. 4C and 5A; see also Fig. S3 and S4 in the supplemental material) provide an experimental basis for testing the impacts of the auxiliary Cu site on not only enzyme activity and stability but also the virulence and survival of this deadly pathogen.
Protective immunity via targeting of bacterial SODs.The monoclonal anti-SOD antibody HD1 recognizes not only meningococcal SODs but also ApSOD and Haemophilus SOD (49). Here, we showed that HD1 conferred an impressive level of passive protection against infection of mice with different strains of serogroup B meningococci (Fig. 1) and that the NmSOD C terminus is critical for HD1 binding (see Fig. S6 in the supplemental material). Although the epitope recognized by HD1 has not yet been comprehensively characterized, this antibody serves as both a useful experimental tool and a proof of concept for protective immunization. In contrast, we were unable to confer active protection in mice by the use of a standard preparation of recombinant NmSOD. Thus, reengineered NmSOD or peptide fragments may provide useful components for broadening coverage of existing vaccine formulations (112). Since the most common serogroups of N. meningitidis all express the sodC gene (44) and since the sequences of NmSODs are highly conserved, a SOD immunogen might also improve vaccine efficacy against related microbes.
Current animal vaccines for brucellosis are unsuitable for human use, and accidental exposure to these brucellosis strains can result in adverse effects (113). The illegitimate use of Brucella as a potential bioterrorism agent also underscores the need for further intervention strategies (114). Periplasmic SOD from B. abortus, a major cause of brucellosis, was identified as a major cattle serum-reactive protein that might serve as a Brucella vaccine immunogen (16, 48, 54). Vaccination with attenuated SOD deletion strains of B. abortus prevented abortion in cattle (115), and vaccination with BaSOD-overexpressing bacteria (116, 117) or with liposome-encapsulated BaSOD or nucleic acid BaSOD vaccines (118, 119) also conferred a protective advantage. Previous efforts identified one of three tested peptide immunogens from the BaSOD sequence as protective when administered in mice (48). Comparative high-resolution structural analyses of pathogen versus host proteins enable identification of peptides that promote immunogenicity against pathogens while minimizing production of host autoantibodies (see Fig. S7 in the supplemental material), a major impetus for this work.
Antigenic targets from new structures.Our structures provide direct insights for potential peptide-based vaccination efforts against both N. meningitidis and B. abortus. These P-class SODs contain protruding surface-exposed loops that include insertions relative to E-class SODs (Fig. 6C). Immunogenic peptide BaSOD A56-A67 (48) represents part of such a loop (see Fig. S7 in the supplemental material). Extending this peptide to N52-A67 (and, by analogy, to NmSOD N83-G98) to incorporate the entire loop exposed in bacterial SODs, but buried in the mammalian E-class dimer interface, might elicit an improved response. BaSOD K6-K17 (equivalent to NmSOD K37-K48) represents another such loop. Lastly, the C-terminal NmSOD sequence implicated in HD1 binding might provide another immunogen (see Fig. S6), as it also is exposed in P-class SODs but buried in the E-class dimer interface. The flexibility inherent in these sequences (Fig. 6B; see also Fig. S2) is important for the interactions of antibody complementarity-determining regions and protein epitopes (120–122). However, to maintain the appropriate secondary structure, peptide immunogens may benefit from structural restraints, designed on the basis of hydrogen bonding, loop conformations, and attachment to the parent protein, as revealed by our NmSOD and BaSOD structures.
Immunogen extension from new SOD structures. (A) Comparison of NmSOD and BaSOD with HsSOD. Identical residues (magenta) were mapped onto superposed HsSOD (light gray) with NmSOD (gold, left) or BaSOD (blue, right) molecular surfaces. Despite the large evolutionary separation between bacterial and mammalian SODs, BaSOD and NmSOD contain many residues identical to HsSOD residues, predominantly near the active site. (B) Cα atomic displacement parameters (ADP; B-factor) for NmSOD and BaSOD are plotted by residue. Increased mobility was noted predominantly in two regions specific to bacterial SODs (an extended loop, green, and a bacterial insertion, red). These two regions were suggested to be similarly mobile using normal mode analysis. Superimposed normal mode analysis model data (see Fig. S2 in the supplemental material) are inset in the plot for NmSOD (top, gold) and BaSOD (bottom, blue). (C) Superposition of single subunits from bacterial SOD (NmSOD, gold, left; BaSOD, blue, center) and HsSOD (gray) crystallographic structures also reveals that the aforementioned insertions/loops characteristic of bacterial SODs are nonconserved regions in 3-dimensional space. These potential immunogens were mapped onto aligned SOD subunits (right, in red and green).
Conclusions.The still-evolving vaccination programs and suboptimal treatments against N. meningitidis and B. abortus (34, 123) underscore the need for novel, additional approaches to intervention in pathogenesis, especially given the emergence of antibiotic resistance. Vaccine and drug design strategies directed at periplasmic SOD virulence factors (124) benefit from detailed knowledge of the structures, assemblies, and mechanisms of SODs from bacteria and their hosts. Both microbial and eukaryotic SODs function as master regulators of ROS and can contribute to human disease. For example, human cancer cells can evade cell death arising from ROS-induced mitochondrial damage, by metabolic reprogramming (125) that upregulates SOD to compensate for decreased Mn superoxide dismutase in mitochondria, which evolved from microbial ancestors. In ALS-linked mutant HsSOD proteins, Cu deficiency and flexibility are associated with misassembly, aggregation, and clinical severity (55). Here, we show that the SODs of pathogenic bacteria differ from the SODs of their hosts by several targetable means: distinct assembly states, flexible insertions, and, sometimes, auxiliary Cu sites that enhance stability. This work provides a framework for predictive understanding of SODs from sequence analysis, highlights similarities and differences between bacterial and eukaryotic enzymes, and describes notable determinants of virulence which can serve as therapeutic targets.
ACKNOWLEDGMENTS
This work was funded by NIH grant R01GM039345 (to E.D.G. and J.A.T.). J.A.T. is supported by a Robert A. Welch Distinguished Chair in Chemistry. Work in the Kroll laboratory was supported by The George John and Sheilah Livanos Charitable Trust. A.J.P. was supported in part through predoctoral NSF and Skaggs Institute for Chemical Biology fellowships and an NIH T32AG000266 postdoctoral training grant. M.D. was supported by a Canadian Institutes of Health Research postdoctoral fellowship.
We thank the beamline staff at the Stanford Synchrotron Radiation Laboratory (SSRL) for their help during data collection and for the use of the Structurally Integrated BiologY for Life Sciences (SIBYLS) beamline at the Advanced Light Source at Lawrence Berkeley National Laboratory. The SIBYLS beamline (BL12.3.1) is funded through the Integrated Diffraction Analysis Technologies (IDAT) program, supported by the Department of Energy and by NIH grant R01GM105404. Pulse radiolysis studies were carried out at the Center for Radiation Chemical Research, Brookhaven National Laboratory, which was supported under contract DE-AC02-98CH10886 with the U.S. Department of Energy and by its Division of Chemical Sciences, Office of Basic Energy Sciences. The contributions and technical assistance provided by Chiharu Hitomi, Andrew S. Arvai, Carey Kassman, and Zhujin Cao are greatly appreciated.
FOOTNOTES
- Received 24 June 2015.
- Accepted 29 September 2015.
- Accepted manuscript posted online 12 October 2015.
- Address correspondence to Elizabeth D. Getzoff, edg{at}scripps.edu.
↵* Present address: Ashley J. Pratt, Health Interactions, San Francisco, California, USA; Michael DiDonato, Genomics Institute of the Novartis Research Foundation, San Diego, California, USA; Carol A. Belzer, Center for Veterinary Biologics-Policy, Evaluation and Licensing, Biotechnology, Immunology and Diagnostics, Ames, Iowa, USA; Louisa B. Tabatabai, Roy J. Carver Department of Biochemistry, Biophysics & Molecular Biology, Iowa State University, Ames, Iowa, USA.
A.J.P. and M.D. contributed equally to this article.
Citation Pratt AJ, DiDonato M, Shin DS, Cabelli DE, Bruns CK, Belzer CA, Gorringe AR, Langford PR, Tabatabai LB, Kroll JS, Tainer JA, Getzoff ED. 2015. Structural, functional, and immunogenic insights on Cu,Zn superoxide dismutase pathogenic virulence factors from Neisseria meningitidis and Brucella abortus. J Bacteriol 197:3834–3847. doi:10.1128/JB.00343-15.
Supplemental material for this article may be found at http://dx.doi.org/10.1128/JB.00343-15.
REFERENCES
- 1.↵
- 2.↵
- 3.↵
- 4.↵
- 5.↵
- 6.↵
- 7.↵
- 8.↵
- 9.↵
- 10.↵
- 11.↵
- 12.↵
- 13.↵
- 14.↵
- 15.↵
- 16.↵
- 17.↵
- 18.↵
- 19.↵
- 20.↵
- 21.↵
- 22.↵
- 23.↵
- 24.↵
- 25.↵
- 26.↵
- 27.↵
- 28.↵
- 29.↵
- 30.↵
- 31.↵
- 32.↵
- 33.↵
- 34.↵
- 35.↵
- 36.↵
- 37.↵
- 38.↵
- 39.↵
- 40.↵
- 41.↵
- 42.↵
- 43.↵
- 44.↵
- 45.↵
- 46.↵
- 47.↵
- 48.↵
- 49.↵
- 50.↵
- 51.↵
- 52.↵
- 53.↵
- 54.↵
- 55.↵
- 56.↵
- 57.↵
- 58.↵
- 59.↵
- 60.↵
- 61.↵
- 62.↵
- 63.↵
- 64.↵
- 65.↵
- 66.↵
- 67.↵
- 68.↵
- 69.↵
- 70.↵
- 71.↵
- 72.↵
- 73.↵
- 74.↵
- 75.↵
- 76.↵
- 77.↵
- 78.↵
- 79.↵
- 80.↵
- 81.↵
- 82.↵
- 83.↵
- 84.↵
- 85.↵
- 86.↵
- 87.↵
- 88.↵
- 89.↵
- 90.↵
- 91.↵
- 92.↵
- 93.↵
- 94.↵
- 95.↵
- 96.↵
- 97.↵
- 98.↵
- 99.↵
- 100.↵
- 101.↵
- 102.↵
- 103.↵
- 104.↵
- 105.↵
- 106.↵
- 107.↵
- 108.↵
- 109.↵
- 110.↵
- 111.↵
- 112.↵
- 113.↵
- 114.↵
- 115.↵
- 116.↵
- 117.↵
- 118.↵
- 119.↵
- 120.↵
- 121.↵
- 122.↵
- 123.↵
- 124.↵
- 125.↵
- Copyright © 2015, American Society for Microbiology. All Rights Reserved.
