Introduction

Superoxide dismutases (SODs, EC 1.15.1.1) are metalloenzymes that are widely expressed in prokaryotic and eukaryotic cells. SODs catalyse the conversion of superoxide radicals (O2·) to H2O2 and O2 and play a crucial role in the primary cellular defence system against oxidative stress1. Four types of SODs have been reported according to their metal cofactors: manganese SOD (Mn-SOD), iron SOD (Fe-SOD), copper/zinc SOD (Cu/Zn-SOD) and nickel SOD (Ni-SOD)2,3,4. Of these, Mn-SOD and Fe-SOD are closely related as their protein structures differ from the other two groups and were recently assigned into one family5. Mn-SOD and Fe-SOD are prevalent in bacteria, mitochondria and chloroplasts and likely evolved from a common ancestor3,5. Mn-SOD and Fe-SOD are typically homodimers or homotetramers and generally require either Mn2+ or Fe2+ to perform their biological activities; however, some SODs, referred to as cambialistic SODs, can accommodate both ions and are often found in anaerobic organisms5,6.

Over the past 40 years, SODs have attracted tremendous attention and are widely used in the pharmaceutical, cosmetic, food, agriculture and environmental protection industries due to their excellent antioxidant properties7,8,9,10,11,12. At one time, SODs were isolated from animal or plant sources, but the focus has recently shifted to microbial sources since these organisms can be easily induced and cultivated on a large scale3. For industrial applications, thermostability is a major requirement for commercial SODs, because thermal denaturation is a common cause of enzyme inactivation. Fortunately, the isolation of thermostable SODs from thermotolerant or thermophilic microorganisms offers a rapid and effective means to address this issue. To date, many thermostable SODs have been identified from thermophiles and hyperthermophiles, such as Aquifex pyrophilus13, Sulfolobus solfataricus14, Aeropyrum pernix15, Pyrobaculum aerophilum16, Bacillus stearothermophilus17, Chloroflexus aurantiacus18, Thermomyces lanuginosus19, Bacillus licheniformis20, Chaetomium thermophilum21 and Thermus thermophilus22. Recent efforts to improve the thermal resistance of SODs through chemical modification, gene recombination and simulated SOD compounds have achieved considerable success23,24. Bioinformatic methods have also been used to guide the bioengineering process25. Nevertheless, the generality of these methods is limited because the determinants of protein thermostability are numerous. Previous studies indicated that SOD thermostability was affected by many factors, such as ion pairs, hydrogen bonds, hydrophobicity, polar interactions, molecular weight and secondary structures25,26,27,28,29. Optimising the thermostability of a SOD using more than one strategy could offer significant progress but also faces great challenges29.

Geobacillus thermodenitrificans NG80-2 is a crude oil-degrading thermophilic facultative anaerobe isolated from a deep-subsurface oil reservoir in Northern China and grows at 45 to 73°C (optimum 65°C)30. Analysis of the NG80-2 genome sequence revealed the presence of three putative sod genes, Fe/Mn-sod (GTNG_2215, referred as sodNG2215 in this study), Mn-sod (GTNG_2400) and Cu/Zn-sod (GTNG_2884), which are also present in other species of the Geobacillus genus31. Of these three genes, sodNG2215 encodes an unusual SOD comprised of 449 amino acids and is considerably larger than other identified SODs. Apart from the conserved functional SODA domain that belongs to the Fe/Mn-SOD family, SODNG2215 contains an extra 244-amino-acid N-terminal domain (NTD), distinguishing it from the homologous enzymes from other mesophilic species of the Bacillus genus. Interestingly, this type of NTD is present in all SODNG2215-like SODs from in Geobacillus and contains one or two repeating segments. In this study, the role of the NTD in the activity, structure, thermal resistance and stress resistance of the host protein was investigated in a detailed manner. The work presented here reveals a special and novel mechanism for protein thermostability and provides a feasible new method to engineer heat-resistant SODs.

Methods

Materials and reagents

The primers were synthesised by the AuGCT Biotechnology Corporation (Beijing, China). Taq DNA Polymerase, dNTP and restriction enzymes were purchased from TaKaRa (Dalian China). T4 DNA ligase was obtained from Promega (Mannheim, Germany) and the DNase I was procured from Roche (Basel, Switzerland). The Chelating Sepharose Fast Flow column and High Molecular Weight Standards were purchased from Amersham Biosciences (Hammersmith, UK). Phenylmethane-sulfonyl fluoride (PMSF) was purchased from Sigma (St. Louis, Mo, USA). All other chemicals and reagents were from Shanghai Sangon, China.

Bacterial strains and culture conditions

G. thermodenitrificans NG80-2 (CGMCC 1228) was grown in Luria-Bertani (LB) medium at 60°C with shaking. Bacillus subtilis BSn5 and Escherichia coli BL21 (DE3) (Novagen) were grown in LB medium at 37°C with shaking. When necessary, 50 mg kanamycin/L was added to the medium.

Bioinformatics methods

The alignment of SODs from G. thermodenitrificans NG80-2 (YP_001126309), B. subtilis BSn5 (ADV92836) and their closely related homologs was performed using the ClustalW program32. Phylogenetic trees were constructed using the neighbour-joining (NJ) method. Bootstrap values (expressed as percentages of 500 replications) greater than 50% are indicated above the branches. Sequence logos were generated with Weblogo33.

Cloning and plasmid construction

Genomic DNA from NG80-2 and BSn5 was extracted as described previously30. The primers used in this study are listed in Supplementary Table S1. The PCR was initiated by denaturation at 94°C for 3 min followed by 30 cycles of 94°C for 30 s, 55°C for 45 s and 72°C for 2 min and a final extension at 72°C for 5 min. Sodr was amplified using PCR with sod_N-GTNG_2215 and sod_C-BSn5 as the templates. The PCR product was digested with NdeI/HindIII (for sodNG2215 and sodANG2215), NdeI/BamHI (for sodABSn5 and sodr), or EcoRI/XhoI (for sodBSn5) and the resulting fragment was ligated into pET-28a (Novagen) to generate the recombinant plasmid. The presence of the insert in the recombinant plasmid was confirmed by sequencing using an ABI 3730 automated DNA sequencer (ABI, Foster City, USA).

Protein expression and purification

E. coli BL21 carrying the recombinant plasmid was grown in LB medium containing 50 mg/L kanamycin to an A600 nm of 0.6. Protein expression was induced for 3 h using 0.1 mM IPTG at 37°C (for SODNG2215 and SODANG2215), 0.05 mM IPTG at 25°C (for SODBSn5 and SODABSn5) and 0.1 mM IPTG at 30°C (for SODr).

All steps in the following procedures were performed at 4°C. E. coli cells were harvested with centrifugation at 10,000 × g for 10 min, washed with lysis buffer (50 mM Tris-HCl, pH 8.0; 300 mM NaCl; 10 mM imidazole), resuspended in the same buffer containing PMSF (1 mM) and lysozyme (1 g/L) and sonicated (Hielscher UP200s ultrasonic processor, 20 kHz, 0.5 cycle, 90% amplitude). After the addition of RNase A (10 μg/mL) and DNase I (5 μg/mL), the lysate was incubated on ice for 15 min and the crude extract was obtained using centrifugation at 18,000 × g for 30 min. The crude extract was applied to a Chelating Sepharose Fast Flow column according to the manufacturer's instructions. Unbound proteins were removed with wash buffer (50 mM Tris-HCl, pH 8.0; 300 mM NaCl; 20 mM imidazole). His-tagged recombinant proteins were eluted with elution buffer (50 mM Tris-HCl, pH 8.0; 300 mM NaCl; 250 mM imidazole) and dialysed against 50 mM Tris-HCl (pH 8.0) containing 20% glycerol.

The protein concentration was determined using the Bradford method34. Sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS-PAGE) was performed according to the method described by Laemmli35.

SOD activity assay

SOD activity was measured using the method of Beauchamp and Fridovich36. Briefly, the 3-ml reaction mixture contained 13 mM L-methionine, 63 μM nitroblue tetrazolium (NBT), 1.3 μM riboflavin, 10 μM EDTA-Na2 and 10 μL purified enzyme in 50 mM potassium phosphate buffer (pH 7.8). The test tubes were exposed to a source of light. The reduction of NBT was monitored after 10 min at 560 nm. One unit of SOD activity was defined as the amount of enzyme that caused 50% of maximum inhibition of the NBT reduction. All assays were performed in triplicate and average values were reported. Activity was estimated as a percentage of the maximum.

Metal reconstitution and analysis

The apoprotein was prepared from purified SOD using dialysis against a solution of 20 mM 8-hydroxyquinoline, 2.5 mM guanidium chloride, 50 mM Tris and 0.1 mM EDTA at pH 3.8 for 24 h at 4°C. The SOD apoprotein was then reconstituted through dialysis with 50 mM Tris buffer (pH 7.8) containing 1 mM ferrous ammonium sulphate or 1 mM manganese chloride for 18 h at room temperature. Then, the solution was subject to further dialysis against the same buffer to remove excess amounts of metal salts37.

Metal analysis and quantification were performed using a 180-80 Polarized Zeeman atomic absorption spectrophotometer. A manganese standard curve was obtained using 0.1% nitric acid and Mn2+ concentrations of 1.0, 2.0 and 4.0 μg/L. Similarly, an iron standard curve was obtained using 0.1% nitric acid and Fe2+ concentrations of 1.0, 2.0 and 4.0 μg/L. The absorbencies of manganese and iron were measured at 297.5 and 248.3 nm, respectively. The SOD samples were diluted in nanopure water with 0.1% nitric acid until the absorbance was within the range of the standard curves. The samples were thoroughly dialysed in 20 mM Tris (pH 8.0) prior to the assay.

Effects of temperature and pH on SOD activity

To determine the optimum temperature, SOD activity was measured in the standard reaction mixture at pH 7.8 at temperatures ranging from 20 to 80°C. To determine the optimum pH, SOD activity was measured at the optimum temperature in the pH range of 3.0 to 10.0 using sodium citrate (pH 3.0–8.0), Tris-HCl (pH 8.0 and 9.0), or glycine-NaOH (pH 9.0 and 10.0) buffers. Activity was measured as a percentage of the maximum.

Thermostability test

The thermal stability of SOD was tested after incubating samples at 20, 30, 40, 50, 60, 70 and 80°C for 10, 20, 30, 40, 50 and 60 min without substrate. The remaining activities of enzymes were detected at their optimum temperature and pH and calculated as the percentage of the maximum SOD activity.

Analytical ultracentrifugation

Sedimentation velocity experiments were performed in a ProteomeLab XL-1 Protein Characterization System (Beckman Coulter). All interference data were collected at a speed of 42,000 rpm in an An-60 Ti rotor at 4°C. A set of 200 scans was collected at 6-min intervals. The proteins were prepared in 50 mM potassium phosphate buffer plus 150 mM NaCl at pH 6.8 for SODNG2215, SODANG2215 and SODr and at pH 7.8 for SODBSn5 and SODABSn5. The data were analysed using the program SEDFIT (version 11.8) in terms of a continuous c(s) distribution.

Effects of inhibitors, detergents and denaturants on SOD activity

The effects of inhibitors (ethylenediaminetetraacetic acid (EDTA) and β-mercaptoethanol (β-ME)), detergents (sodium dodecyl sulphate (SDS)) and denaturants (urea and guanidine guanidine hydrochloride) on SOD activity were investigated according to the method described by Zhu38.

Circular dichroism (CD)

Circular dichroism was measured on a Jasco J-715 circular dichroism spectrometer (Japan). First, 0.14 mg/mL SODBSn5 and 0.4 mg/mL other proteins including SODNG2215, SODANG2215 SODABSn5 and SODr in 20 mM potassium phosphate buffer (pH 8.0) was placed in a 1 mm quartz cell. All spectra were recorded at a scanning rate of 100 nm/min with 0.1 nm wavelength steps from 190 to 250 nm and three accumulations.

Structural modelling

The structure of SODANG2215, the functional domain of SODNG2215, was modelled via the SWISS-MODEL server (http://swissmodel.expasy.org/). The structure of a closely related iron SOD from the strain M. thermautotrophicus Delta H, which has a 52% sequence identity and 67% sequence similarity with SODANG2215, was used as a template. The stereochemical and overall quality of the final model was assessed using PROCHECK and ProSA. In the resulting model, 89% of the residues were in the most favoured regions and no residues were located in disallowed regions. The PROCHECK overall g factor evaluating all torsion angles and bond lengths was 0.08, indicating a high-quality model. The ProSA Z score of the SODANG2215 model was calculated as −6.03. Z scores for experimentally determined X-ray structures of proteins with a similar size (around 207 amino acids) lie in the range of −1 to −10.8. The Z score of this model was thus within the range of scores typical for similarly sized native proteins.

Results

Bioinformatic analysis of SODNG2215

SODNG2215 possesses a unique repeat-containing N-terminal domain (NTD) that differs from its closely related homologues from Bacillus. The functional domains (SODA) of SODNG2215-like SODs from Geobacillus are conserved (preliminary sequence identity: 84–93%) and are also highly similar to their counterparts from Bacillus (preliminary sequence identity: 64–71%). Alignment of the SODNG2215-like SODs from 13 thermophilic Geobacillus strains showed that these SODs have an extra N-terminal segment despite their various lengths (Supplementary Fig. S1). A blast analysis revealed that these segments share significant amino acid similarity with each other but not with other sequences in GenBank. Further alignment of these N-terminal segments revealed that the eight SODs (subgroup 1) harbour two repeat sequences within their extra N-terminal segments, whereas the other four (subgroup 2) only have a single repeat sequence (Fig. 1b, c). Due to the presence of these extra segments in SODNG2215-like SODs from all thermophilic Geobacillus but not in SODs from any Bacillus species, these repeat-containing domains were thus proposed to be involved in the thermostability maintenance of SODs. However, whether the repeat sequence forms a special secondary structure remains unknown. The phylogenetic trees of both the N- and C-terminal domains of Geobacillus SODs exhibited highly similar topological structures (Fig. 1a,b), indicating that the sod genes potentially gained their extra NTDs before the divergence at nodes N1 and N2 occurred.

Figure 1
figure 1

Phylogenetic analysis of C-terminal (functional) fragments (a) and N-terminal fragments (b) of SODs from G. thermodenitrificans NG80-2 (YP_001126309), B. subtilis BSn5 (ADV92836) and their closely related homologs. The accession number of each SOD in GenBank is labelled prior to the species name. The proteins used for the activity test in this study are highlighted in red. The sequence logos of the repeat sequences were constructed from alignments of SOD N-terminal fragments from the 13 Geobacillus strains (c). The letter size is proportional to the degree of amino acid conservation. The deletion (gap) is represented by the letter “B”.

Gene manipulation, expression and purification of SOD variants

To explore the role of the SODNG2215 NTD, we constructed five recombinant clones (Fig. 2a) to comparatively study the thermostability, temperature-induced conformational changes and stress resistance of the corresponding recombinant proteins. First, the full-length sodNG2215 and the NTD-deleted form (sodANG2215) were separately cloned to study the effects of the NTD in its natural host. To further investigate its effects on other Fe/Mn SODs, the NTD was fused to a Fe/Mn SOD from a mesophilic strain of B. subtilis BSn5 (SODBSn5), which shares 66% amino acid sequence identity with SODNG2215. The resulting construct (sodr) contained the DNA segments encoding the SODNG2215 NTD and the SODA domain of SODBSn5. The SODs from Bacillus are the closest phylogenetically to the SODs from Geobacillus (Fig. 1a) and typically exhibit poor heat-resistance39. The full-length sodBSn5 and sodABSn5 genes encoding the C-terminal SODA domain were also cloned as negative controls.

Figure 2
figure 2

A schematic illustration of the various constructs (a) and the SDS-PAGE analysis of the purified SOD proteins (b).

Proteins were stained with Coomassie brilliant blue. Ladder, standard protein size marker. The expected sizes of SODNG2215, SODANG2215, SODBSn5, SODABSn5 and SODr were 54.0 kD, 26.6 kD, 37.3 kD, 26.3 kD and 53.6 kD, respectively.

SDS-PAGE analysis showed that all five SOD proteins were expressed accordingly to their expected sizes (Fig. 2b).

Removing the NTD does not alter the metal ion specificity of the host SOD

To investigate the metal ion preference of SODNG2215 and its NTD-deleted form, SODANG2215 Fe2+- and Mn2+-reconstituted enzymes were prepared. As shown in Supplementary Table S2, an atomic absorption analysis demonstrated that each monomer of the native SODNG2215 was associated with 0.55 ± 0.03 atoms of iron and 0.10 ± 0.01 atoms of manganese, whereas 0.73 ± 0.02 atoms of iron and 0.81 ± 0.03 atoms of manganese were associated with the Fe2+- and Mn2+-reconstituted SODNG2215 enzymes per subunit, respectively. The Fe2+- and Mn2+- reconstituted SODNG2215 exhibited specific activities of 769.1 ± 17.2 and 1158.4 ± 29.3 U/mg, respectively; the apoprotein was inactive. The Mn2+-reconstituted SODNG2215 was approximately 40 to 50% more active than the native and Fe2+-reconstituted SODNG2215, suggesting that SODNG2215 is a cambialistic SOD that can accommodate both Fe2+ and Mn2+ as cofactors with manganese preferred over iron in regard to enzymatic activity. The metal ion specificity study of SODANG2215 revealed similar characteristics only with less enzymatic activity than SODNG2215 (Supplementary Table S2). This result indicates that the NTD has no effect on the metal ion specificity of SOD.

Oligomerisation form and composition of SOD are not altered by the NTD

Analytical ultracentrifugation analysis (Fig. 3) indicates that either the full length SODNG2215 or the NTD-deleted form of SODANG2215 exist mainly in a tetrameric form. SODBsn5 and its counterparts, SODABsn5 and NTD-fused SODr, were found in dimeric and monomeric forms with no tetramer detected. The extremely increased peak area ratio of the tetramer compared with the monomer indicates most protein molecules of SODNG2215 and SODANG2215 tend to form multimers, whereas all three SODBsn5-derived proteins prefer to exist in monomeric forms. It thus seems that the functional SODA domain of the SOD variants determines the formation of oligomerisation, whereas the NTD has no effect on the oligomerisation form or the monomer-multimer ratio.

Figure 3
figure 3

Analysis of the oligomerised forms of five SOD proteins.

The size of SODNG2215, SODANG2215, SODBSn5, SODABSn5 and SODr (all at 0.05 mM) were determined using analytical ultracentrifugation. The estimated molecular masses (kDa) are provided above the peaks.

NTD contributes to host thermophilicity with a slight alteration in pH profiles

The optimum active temperature (OAT) was determined by testing the SOD activity at temperatures ranging from 20 to 80°C (Fig. 4a). The OAT for SODNG2215 was 70°C, which is close to that of other thermophile-derived SODs (50–70°C)17,40 but lower than that of the hyperthermophilic SODs (87–95°C)14,16. SODANG2215, SODBSn5 and SODABSn5 exhibited mesophilic properties with OATs of 30°C, 35°C and 30°C, respectively. Intriguingly, SODr exhibited optimal activity at 55°C and retained 46% of its maximum activity even at 80°C. Clearly, the NTD-fused proteins of SODNG2215 and SODr are considerably more thermophilic than their counterparts without the NTD (SODANG2215, SODBSn5 and SODABSn5).

Figure 4
figure 4

Effects of temperature and pH on SOD activity.

The temperature profiles were determined by assaying the activity of purified SOD at temperatures ranging from 20 to 80°C (a). The activity at the optimum temperature was defined as 100%. Each point represents the mean (n = 3) ± standard deviation. The activity of purified SOD was evaluated in buffers ranging from pH 3 to 10 (b). The activity at the optimum pH was defined as 100%. Each point represents the mean (n = 3) ± standard deviation.

Although the relative activities were used for the comparison of the thermophilicities of the variant SODs, the real activities of these recombinants are quite different. The initial enzymatic activities of SODNG2215, SODANG2215, SODBSn5, SODABSn5 and SODr investigated at 20°C are 800.9 ± 18.3, 837.1 ± 19.2, 682.3 ± 14.6, 598.8 ± 16.2 and 678.3 ± 15.1 U/mg, respectively, whereas the maximum at their individual OATs rise to 1132.2 ± 20.6 (70°C), 915.9 ± 18.9 (30°C), 933.0 ± 19.1 (35°C), 752.5 ± 17.5 (30°C) and 1055.1 ± 20.3 (55°C) U/mg, respectively.

The pH profiles of the SOD activity are presented in Figure 4b. SODNG2215 and SODANG2215 exhibited maximum activities at pH 4 and 5–6, respectively, indicating that they are both acidophilic. Outside their optimum pH ranges, the activities of both enzymes quickly decreased. To our knowledge, SODNG2215 is the most acidophilic, thermophilic, cambialistic SOD that has been reported to date. In contrast, SODBSn5, SODABSn5 and the NTD fused SODr were all neutrophilic with an optimum pH of around 8. Therefore, the NTD showed negligible effects on the pH preference of SOD activities.

NTD plays an important role in maintaining the thermostability of SODs

An optimum active temperature assay demonstrated that the presence of the NTD alters the thermophilicity of SODs. Therefore, we further examined the role of the NTD in SOD thermostability.

As shown in Figure 5, all enzymes were very stable at a low temperature (20°C) and almost fully active after 60 min. However, the enzymes exhibited variable stability when the temperature was greater than 40°C. SODANG2215, SODBSn5 and SODABSn5, which are mesophilic, also demonstrated poor thermostability when the temperature was greater than 40°C. All three enzymes lost approximately 90% of their activity after 60 min above 60°C and were almost inactive after 10 min at 80°C.

Figure 5
figure 5

Thermostability of purified SODNG2215 (a), SODANG2215 (b), SODBSn5 (c), SODABSn5 (d) and SODr (e).

The enzyme was pre-incubated at various temperatures (20–80°C, steps of 10°C) and aliquots were withdrawn every 10 minutes to test the residual activities at the optimum temperature and pH by using the standard assay described in the Methods section. The activity of non-heated SOD was defined as 100%. Each point represents the mean (n = 3) ± standard deviation.

In contrast, the thermophilic SODNG2215 exhibited excellent thermostability over a range of temperatures below 70°C and still remained 70% active after 60 min at 70°C. Even at 80°C, SODNG2215 remained 52% active after 10 min and almost maintained at this level until 60 min. SODr, the NTD fused SODABSn5, also exhibited good thermostability (retained 80% and 61% active at 50°C and 60°C, respectively, after 60 min), indicating that this enzyme acted more like a moderate thermophilic Mn-SOD from B. licheniformis20.

Meanwhile, the half-lives and D-values deduced from thermostability assays clearly showed that the NTD fused SODs were much more stable than their counterparts without the NTD especially at high temperatures (Supplementary Table S3).

CD spectroscopy analysis reveals the enhanced conformational stability of NTD-fused SODs

Far-UV CD spectra of five SODs exhibited a positive maximum at 192 nm and double-lobed negative peaks at 208 nm and 222 nm, which are characteristic of the high α-helix content of the protein41,42 (Fig. 6). However, temperature-dependent CD measurements clearly indicated that the conformational changes of SODNG2215 and SODr are quite small compared with the other three counterparts, which is consistent with the results from the temperature-dependent activity assays. The CD signal intensity almost did not decrease further when the temperature rose from 70 to 90°C. Additionally, strikingly reduced ellipticity signals at 192 nm and 208 nm for SODANG2215, SODBSn5 and SODABSn5 resulted in a sharp isodichroic point at 202 nm, indicating a two-state helix-coil transition43. In contrast, the isodichroic points for SODNG2215 and SODr are not obvious. When the Spectra was measured at 5°C intervals (Supplementary Fig. S4), an apparent isodichroic point was observed at 201 nm for both SODNG2215 and SODr. In addition, the magnitudes at 222 and 208 nm decreased very slightly when the temperature crossed from 70°C to 85°C.

Figure 6
figure 6

Temperature-dependent CD spectra of SODNG2215 (a), SODANG2215 (b), SODBSn5 (c), SODABSn5 (d) and SODr (e).

The NTD enhances the host proteins' abilities to resist inhibitors, detergents and denaturants

The effects of various inhibitors, detergents and denaturants on SOD activity were examined using EDTA, β-ME, SDS, urea and guanidine hydrochloride (Fig. 7 and Supplementary Table S4). SODs fused with the NTD are considerably more resistant to these stresses than their counterparts lacking the NTD. Of these stresses, the NTD confers an extraordinary ability to resist denaturants. When tested with urea or guanidine hydrochloride at a final concentration of 2.5 M, 72–93% of the relative activities of SODNG2215 and SODr were maintained. In contrast, only 43–64% of their activities remained in their counterparts without the NTD. Additionally, SODNG2215 and SODr maintained greater than 45% of their relative activities in SDS at a concentration of 1%, whereas the other SODs were completely repressed.

Figure 7
figure 7

Effects of inhibitors, detergents and denaturants on SOD activity.

The enzyme was incubated with each inhibitor, detergent, or denaturant at various final concentrations in 50 mM sodium phosphate buffer (pH 7.8) at 25°C for 30 min. Residual activities were measured by the standard assay as described in the Methods section. The reaction mixture without inhibitor, detergent, or denaturant was used as a control and defined as 100%. The values are the mean ± standard deviation from three separate replicates. No enzymatic activities of SODANG2215, SODBSn5 and SODABSn5 were detected with 1% SDS.

Structural modelling and analysis reveal a conserved backbone structure shared by SODANG2215 and its homologs

The structures of several SODs, including those from M. thermautotrophicus (1MA1, tetrameric), A. pernix (3AK3, tetrameric), P. aerophilum (3EVK, tetrameric), S. acidocaldarius (1B06, tetrameric) and S. solfataricus (1WB8, dimeric), which fall under the same branch as SODANG2215, have been reported.

The structure of SODANG2215, the functional domain of SODNG2215, modelled based on the structure of a closely related iron SOD from M. thermautotrophicus Delta H, which has 52% sequence identity and 67% sequence similarity with SODANG2215. The model of SODANG2215 demonstrated good superimposition with 1MA1, 3AK1, 3EVK, 1B06 and 1WB8 with rmsd values of 0.061, 0.513, 0.501, 0.697 and 0.763, respectively. In addition to the similar backbone structure, the metal binding residues (Fig. 8) are also conserved among the thermophilic Fe-SODs.

Figure 8
figure 8

The model of SODANG2215 generated based on an iron SOD from M. thermautotrophicus Delta H (a). Superposition of the SODANG2215 model with 1MA1, 3AK1, 3EVK, 1B06 and 1WB8 (b) as well as structures (1ISA and 1MSD) of SODs from E. coli and humans (c).

Discussion

Three types of SODs (Fe/Mn-SOD, Mn-SOD and Cu/Zn-SOD) are found in thermophilic Geobacillus, of which Mn-SOD was identified as being considerably thermostable38. In this study, we investigated the thermostability of SODNG2215, a cambialistic SOD from the facultative anaerobic G. thermodenitrificans NG80-2, which can function with both Fe2+ and Mn2+ as cofactors. Cambialistic SODs can be divided into two groups. One group exhibits approximately equal activity in the Mn2+ and Fe2+ forms, whereas the other group prefers manganese over iron in regard to enzymatic activity44. The former group is often found from anaerobes and the latter is from aerobes or facultative anaerobes15. Clearly, SODNG2215 belongs to the latter group.

Geobacillus is a phenotypically and phylogenetically coherent genus of thermophilic bacilli and was recently separated from the genus Bacillus45. Sequence analysis has revealed that the SODNG2215-like proteins from Geobacillus all contain unique NTDs that are not present in other mesophilic Bacillus species. The NTDs are thought to be involved in maintaining the thermostability of these SODs (Fig. 1 and Supplementary Fig. S1). Of considerable interest is a 40-amino-acid sequence that is repeated either once or twice within the NTDs and is exclusive to SODNG2215 homologs from Geobacillus. The phylogenetic analysis of currently known thermostable Fe-, Mn- or Fe/Mn- SODs from various microorganisms showed that these SODs were roughly split into three major branches except for a Mn-SOD from Thermoascus aurantiacus and a Fe-SOD from Thermosynechococcus elongates, with that the phylogeny and metal specificity of these thermophilic SODs are not congruent (Supplementary Fig. S2). The Fe/Mn-SOD (SODNG2215) and Mn-SOD (GTNG_2400) from G. thermodenitrificans NG80-2 are located in two distinct subgroups. Given their differing protein sizes and ion requirements, it appears that these two SODs may have divergently evolved. SODNG2215 and one of its closely related homologs from Geobacillus sp. EPT3, which was previously identified as a Mn-SOD38, cluster into a same branch with five other SODs derived from different species. However, SODNG2215 and Mn-SOD from EPT3 may adopt completely different thermophilic adaptation strategies compared to five other SODs because the NTD deleted SODANG2215 is more like a mesophilic enzyme. In this study, we provided strong evidence that the NTD containing two repeat sequences contributed to SOD thermostability. Further work to determine the crystal structure and characterise the NTDs containing only a single repeat sequence will help to elucidate the specific role that these sequences play in thermostability.

The finding that the optimum active temperature of the NTD-deleted SODANG2215 decreased dramatically from 70 to 30°C suggests that the NTD plays an essential role in maintaining the high thermotolerance of SODNG2215. In contrast, when the NTD was fused with SODABSn5, a mesophilic SOD, the resulting enzyme (SODr) exhibited remarkably enhanced thermophilicity. The detailed assessment of thermostability also demonstrated that the thermostability of SODr is significantly enhanced compared with SODBSn5 and SODABSn5. All these findings indicate that the NTD of SODNG2215 is capable of enhancing the thermophilicity as well as the thermostability of other mesophilic SOD homologs. Futhermore, the comparison of five SOD proteins' Teq values calculated according to the Equilibrium Model (Supplementary Fig. S5 and Table S5) indicated that the NTD also increased the estimated optimum working temperature range46,47. It is worth mentioning that although SODBSn5 has a small N-terminal peptide containing 78 residues that is absent in SODABSn5, it only slightly affects the thermophilicity and enzymatic activity.

Some previous studies on cambialistic SODs found that they were more thermostable when reconstituted with manganese rather than iron15,48. The thermostability of Mn2+- or Fe2+-reconstituted SODNG2215 and SODANG2215 were also investigated in the present work (Supplementary Fig. S3). The trends chart of the thermostability of both enzymes at 70°C revealed that the Mn2+-reconstituted proteins are considerably more stable than the Fe2+-reconstituted SODs and as stable as the native SODs. Importantly, when the NTD is deleted, the Mn2+-reconstituted SODANG2215 still exhibited poor stability at high temperatures and its thermostability was even worse than the Fe2+-reconstituted SODNG2215, which contains the NTD. Thus, the effect of the NTD is much more important than the ion constitution in regards to the thermostability of SODNG2215.

Enzymes isolated from thermophilic (50–80°C) or hyperthermophilic (>80°C) microorganisms are generally more thermostable and more resistant to enzyme inhibitors, protein detergents, pH and other denaturing agents compared with SODs from mesophilic (25–50°C) or psychrophilic (<25°C) microorganisms49,50. These enzymes therefore have the potential to be widely used in industrial capacities. Factors contributing towards thermostability of proteins are numerous and complex, such as certain types of amino acids51, residue contributions to VdW interaction energies52, increase in hydrophobicity53,54, differences in compactness55 and packing density56, significant changes in hydrogen bonding networks57, enhancement of secondary structure propensity58, extensive ion pair interactions59, differences in the number of amino acids comprising the loops60 and even a single amino acid mutation far from the active site24,61. Therefore, it is extremely difficult to bioengineer a specific enzyme with improved thermostability by using a “universal” method. The discovery that fusion with the NTD from SODNG2215 can significantly improve the thermostability of host SODs provides an easy and feasible method to generate heat-resistant SOD proteins. A similar example has been reported that the N-terminal domain of a multi-functional cellulase EGXA increased the enzyme's thermostability (only tested on 45°C) to a slight extent, while the mechanism might be different since the N-terminal domain was identified as a cellulose-binding domain and only selectively enhanced enzymatic activity and thermostability towards some of substrates62. In addition to thermostability enhancement, the NTD also significantly improved the stress resistance of the host SOD proteins. For all of the tested stresses of inhibitors, detergents and denaturants in this study, SODNG2215 and the NTD-fused SODr both demonstrated enhanced resistance compared with their counterparts without the NTD (Fig. 7 and Supplementary Table S4).

All known thermophilic Fe-SODs except the SODNG2215-like proteins from Geobacillus are similar in protein size (20 kD), backbone structure and metal binding residues (Fig. 8 and Supplementary Fig. S2). Despite such similar features, their thermal stability is remarkably different. It was believed that α-helical content, α-helix length and even the salt bridges formed by the charged residues, such as Arg and Glu, are important factors for Fe-SOD thermostability. In contrast, the β-strand is negatively correlated via a principal component analysis of five thermophilic SODs and six mesophilic counterparts25. Superimposing this model with SOD structures (1MA1, 3AK1, 3EVK, 1B06 and 1WB8) revealed that these tertiary structures are well conserved, especially the regions around the ion binding site. Evidence from metal reconstitution analysis and density gradient centrifugation indicated that the NTD of SODNG2215 did not affect the metal ion specificity and polymerisation of the protein. Additionally, both temperature-dependent CD analysis and temperature-dependent activity assays demonstrated that this domain greatly contributes to the thermo-tolerance of the host protein. We thus assumed that the NTD might provide an outer envelope that covers the temperature-sensitive hydrophobic residues or cavities on the surface of the active SODA “core” and improves the formation of hydrogen bonds or polar interactions between the monomers without affecting the interactions in the inner SODA “core”, which contributes the metal binding site and is important for tetramer formation. Additionally, artificially fusing the NTD to the mesophilic SODABSn5, a close homolog of SODANG2215, led to a moderately thermophilic enzyme (OAT changed from 30 to 55°C). In contrast, SODNG2215 can tolerate higher temperatures, indicating that the NTD has evolved to be a well-matched heat-resistant partner to SODANG2215 over a long period.

In conclusion, to our knowledge, various factors determine protein thermostability. Some proteins even evolved more than one strategy to maintain their thermal tolerance. Herein, a new mechanism of protein thermostability defined by a unique peptide was discovered. The ability of the SODNG2215 NTD to improve the heat and stress resistance of host proteins could be highly useful in generating thermostable SODs for industrial applications. Work to determine the crystal structure of these enzymes is underway and could help to clearly reveal the mechanism of the improved thermostability conferred by the NTD to SODs.