The Single Superoxide Dismutase of Rhodobacter capsulatus Is a Cambialistic, Manganese-Containing Enzyme

Recombinant Rhodobacter SOD displays enzymatic activity with both iron and manganese.

We first examined the dependence of SOD_Rc activity on the nature of the bound metal by expressing the Rhodobacter sod gene in E. coli under the control of the T7 RNA polymerase promoter. The sod coding region was amplified by PCR by using the pA22 plasmid (8) as a template and oligonucleotides 5′-GATCGCTGCCATGGCTTTCGAACTTCC-3′ and 5′-TTTCATCCGTCCATCGAGCTCACATCC-3′ as 5′ and 3′ primers, respectively. The generation of NcoI and SacI sites (underlined) permitted cloning of the amplified DNA into compatible sites of the pET-28a expression vector to yield pESOD28. A SacI-SphI fragment of about 900 bp carrying the sod gene was excised from pESOD28 and ligated into compatible sites of the pET-32b vector to introduce ampicillin resistance. The resulting plasmid, pESOD3228, was used to transform the E. coli sodA sodB strain QC774/DE3 (32).

Transformants were grown aerobically and induced by following the protocols suggested by the pET System manual (Novagen). Cells were broken by sonic oscillation during 30 s in a Branson Sonifier 250, and the corresponding extracts were analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) by using the buffer system described by Laemmli (20). The results shown in Fig. 1A indicate that most of the recombinant protein sedimented with the inclusion bodies fraction. Pellets were thus collected, washed, and dissolved by following the recommended protocols (pET System; Novagen). All buffers were supplemented with 5 mM dithiothreitol (DTT).

• Open in new tab
• Download powerpoint

FIG. 1.

Reconstitution of recombinant R. capsulatus SOD expressed in E. coli. Cells of strain QC774/DE3 transformed with plasmid pESOD3228 were induced with isopropyl-β-d-thiogalactopyranoside, collected, and broken by sonic oscillation. (A) Soluble fractions (lanes 1 and 2) and pellets (lanes 3 and 4) corresponding to 50 μl of culture, collected before (lanes 1 and 3) and after (lanes 2 and 4) induction, were resolved by SDS-PAGE and stained with Coomassie brilliant blue. Molecular masses (in kilodaltons) and electrophoretic mobilities of protein standards are shown on the left. (B and C) SOD_Rc was isolated from inclusion bodies and diluted in the absence (lanes 3) or in the presence of Mn²⁺ (lanes 4) or Fe²⁺ (lanes 5). Protein contents were determined by a dye binding assay (33). Aliquots corresponding to 1 μg of protein were subjected to native PAGE and in situ activity staining at pH 8.0 (B) or 6.0 (C), essentially as described by Beauchamp and Fridovich (3). Lanes 1 and 2 were loaded with 1 μg of purified aero-SOD_Rc and 1 μg of purified photo-SOD_Rc, respectively.

Solubilized apo-SOD was refolded by dilution (1:10) against 0.5 mM MnSO₄ or 0.5 mM Fe(NH₄)₂(SO₄)₂ in 20 mM Tris-HCl (pH 8.0), 1 mM DTT, and 10% (vol/vol) glycerol. The solution was then loaded onto a Q-Sepharose HP column (Pharmacia Äkta System) and eluted with a NaCl gradient. Protein peaks were dialyzed against 20 mM Tris-HCl (pH 8.0), 5 mM DTT, 5 mM EDTA, 2 mM ascorbic acid, and 10% (vol/vol) glycerol to eliminate unspecifically bound metal. All solutions were previously sparged with N₂.

SOD activity, as determined by in situ staining of native polyacrylamide gels (3), could be recovered by reconstitution with either Mn²⁺ or Fe²⁺. Although the activity of the Fe form of the enzyme was hardly detectable when developed at pH 8.0, it increased significantly as the pH of the assay mixture was lowered (Fig. 1B and C). A similar behavior has been observed with SODs from other sources (43).

The reconstituted enzymes were characterized with respect to quaternary structure and metal content. A single subunit of 22 kDa, corresponding to the deduced molecular mass of the protein encoded by the R. capsulatus sod gene (8), was visualized by SDS-PAGE, while gel filtration analysis on a Superdex G 200 HR 10/30 column indicated that the two reconstituted dismutases were dimers of identical polypeptides (data not shown). Atomic absorption photometry revealed that MnSOD contained 0.96 ± 0.05 atoms of Mn per monomer and no detectable Fe (<0.06 atoms per monomer) while FeSOD carried 0.86 ± 0.10 atoms of Fe per monomer with no detectable Mn (<0.01 atom per monomer).

The results indicate that, despite sequence similarity with FeSODs, the single SOD of R. capsulatus displays higher enzymatic activity when reconstituted in vitro with Mn²⁺. Since the enzyme is functional to various degrees with either of the two metals, it is not clear at this stage which cofactor is preferentially bound to SOD in vivo under the different lifestyles of the microorganism.

The culture conditions modulate metal uptake by Rhodobacter SOD.

To determine the identity of the metal cofactor incorporated by SOD_Rc in vivo, the enzyme was purified from both aerobically and photosynthetically cultured R. capsulatus 37b4 cells to yield aero-SOD_Rc and photo-SOD_Rc, respectively. Aerobic or phototrophic conditions were established by growing the cells at 32°C in malate mineral medium (40) either in an air-sparged 15-liter Microferm fermentor (New Brunswick Scientific Co.) under high stirring (500 rpm) or in illuminated screw-cap bottles. Bacteria were grown to an optical density of about 0.8 at 660 nm, harvested by centrifugation, washed, and suspended in 20 mM Tris-HCl (pH 8.0), 5 mM EDTA, 10 mM DTT, 2 mM MgCl₂, 0.1 mM phenylmethylsulfonyl fluoride, 20 μg of RNase/ml, 10 μg of DNase/ml, and 10% (vol/vol) glycerol. Cells were finally disrupted by sonic oscillation, and lysates were cleared by centrifugation (60 min at 140,000 × g). SOD_Rc was isolated by ammonium sulfate fractionation between 70 and 90% of saturation, followed by chromatography on a Pharmacia Äkta System (Q-Sepharose High Performance, Phenyl Sepharose CL-B4). SOD_Rc was purified 209-fold from aerobic cultures and 380-fold from photosynthetically grown cells, with yields of 34 and 40%, respectively.

Filtration chromatography showed that native SOD_Rc isolated from both respiratory and phototrophic cells behaved as a dimer, as had been previously observed for the reconstituted enzymes. Aero-SOD_Rc contained 93% manganese and only 7% iron per monomer, but photo-SOD_Rc increased its iron content up to 40% (Fig. 2A), suggesting that metal incorporation is influenced by culture conditions, most likely by the oxygen tensions during growth (1, 15). At pH 7.8, the specific activities of purified photo-SOD_Rcs (2,500 to 3,966 U/mg) represent 45 to 71% of that displayed by the aerobic form of the enzyme (5,573 U/mg).

• Open in new tab
• Download powerpoint

FIG. 2.

Correlation between the catalytic activity and the metal present in SOD of R. capsulatus. Native and reconstituted SOD_Rc forms were prepared as indicated in the text. (A) Manganese (light gray) and iron (dark gray) contents of the various SOD forms were measured by atomic absorption spectrometry. (B and C) The activities of SOD_Rc proteins isolated from aerobic cultures (aero) and from two independent photosynthetic cultures (phot1 and phot2) and those obtained from recombinant SOD_Rc reconstituted with either Mn or Fe were determined by published procedures (25) at pH 7.8 (B) or 6.2 (C). Bars represent the specific activities of each dismutase form at the two pH values. Standard errors in metal determination and activity measurements were less than 5 and 10%, respectively.

Superoxide disproportionation mediated by FeSOD or MnSOD involves a solvation step that makes the reaction pH dependent (6, 34). We therefore measured the catalytic activities of native and recombinant forms of SOD_Rc at two different pH values. The reaction rate exhibited by a recombinant SOD_Rc reconstituted with Mn was 47-fold higher than that of FeSOD_Rc at pH 7.8 (Fig. 2B), while at pH 6.2 this ratio declined to ∼3.5 (Fig. 2C). In particular, the specific activity of the SOD reconstituted with Fe increased almost one order of magnitude as the pH of the medium was lowered (Fig. 2C). A comparison of the properties of reconstituted dismutases with those displayed by aero-SOD_Rc and photo-SOD_Rc suggests that the latter enzymes are made up of mixtures of homogenous dimers (MnSOD or FeSOD) and heterodimeric forms in various proportions. However, the fall in catalytic competence experienced by all native SOD_Rc species at acidic pH (Fig. 2B and C) could not be accounted for by any simple combination of enzyme variants and metal distribution. It is likely that heterodimers have specific activities that are different from those of single-metal dismutases and affected by the pH in an unpredicted way, but further research is required to substantiate this assumption. The results also suggest that knowledge about the effects of pH on dismutase activity might contribute to an understanding of the molecular basis of metal discrimination and function in the FeSOD and MnSOD families, although the physiological significance of this pH dependence is not obvious. Internal pH is probably regulated within narrow limits as R. capsulatus shifts among different growth conditions, but little is known about the homeostatic mechanisms operating in these phototrophic bacteria under changing environments.

The monomeric unit of MnSOD and FeSOD is characterized by two domains (11, 21, 35). The N-terminal portion of the molecule includes two α-helices in close contact, while the C-terminal domain forms a globular-shaped structure surrounding a core of three β-strands. The active-site metal is bound at the interface between apolar side chains contributed by the β-sheet and the N-terminal α-helices. The metal cofactor presents a bipyramidal trigonal geometry and is ligated by two histidines and an aspartate in the equatorial plane, with a third histidine and a coordinated solvent as axial ligands.

The high degree of sequence and structural conservation among dismutases of these two families precludes the use of standard sequence homology analyses, which are unable to identify subtle conformational differences, such as small variations in distance among important residues in the folded protein or minor changes in metal coordination chemistry. However, careful comparison of aligned sequences of MnSODs and FeSODs combined with assessment of structural data led several authors to identify a few specific amino acids, located close to the cofactor binding site, as major determinants in metal discrimination (17, 27). In particular, the glutamine residue present in the outer coordination sphere that forms a hydrogen bond with the active-site solvent molecule has been singled out as a “crucial factor” for metal selectivity (41). For typical FeSODs, this glutamine is contributed by the α2-helix (corresponding to position 68 in the E. coli MnSOD), whereas in the majority of MnSODs, the isofunctional residue (Gln146 in E. coli FeSOD) belongs to the C-terminal end of the β2-strand. However, the atypical SOD from Sinorhizobium meliloti displays an iron-type sequence but preferentially incorporates Mn in vivo (30) and the cambialistic SOD of P. gingivalis has the solvent-coordinating glutamine at the N-terminal position (1, 2). Moreover, the results obtained by exchanging this functional residue between the α2-helix and β2-sheet locations through site-directed mutagenesis indicate that it is not the only determinant of metal selection in FeSODs and MnSODs (16, 22, 32).

At the metal binding site of R. capsulatus SOD, the crucial glutamine interacting with the coordinated solvent is most likely provided by the N-terminal portion of the protein, as in canonical FeSODs (8). However, the present study shows that the native dismutases found in both aerated and photosynthetic cultures contain Mn as the predominant cofactor (Fig. 2) and that purified recombinant SOD_Rc is indeed more active when reconstituted with Mn (Fig. 1B and C and 2). The cofactor environment of Mn-reconstituted SOD_Rc appears to be very similar to that of E. coli MnSOD, the major difference being a higher degree of symmetry in the equatorial ligand-metal interactions of the Rhodobacter dismutase, as revealed by high-field electron paramagnetic resonance (39). The overall results indicate that SOD_Rc is a MnSOD with cambialistic properties and that the corresponding gene should be renamed sodA. They also show that structural and functional predictions inferred from sequence comparisons have to be handled with caution for this class of SODs.

Another criterion used to classify SODs within the Fe and Mn families is their differential sensitivities to certain inhibitors. Rapid inactivation of FeSOD, but not MnSOD, by H₂O₂ has been ascribed to the oxidation of an essential tryptophan residue in the vicinity of the iron binding site (4, 42, 46), although the actual identity of the reactive amino acid is still a matter of debate (7, 27, 37, 45). A recombinant SOD_Rc reconstituted with Mn was virtually insensitive to H₂O₂, whereas Fe incorporation led to a holoenzyme that was inhibited 60 to 70% by a similar treatment (data not shown). On the other hand, native SOD_Rc was shown to be affected by H₂O₂ exposure, although metal contents were not determined at the time (8). When purified dismutases of known metal compositions were exposed to H₂O₂, the observed inactivation levels always exceeded those predicted from the theoretical fractions of FeSOD in the mixtures (data not shown), once again suggesting the presence of heterodimers that were sensitive to H₂O₂ deleterious action.

Previous reports have shown that metal incorporation in cambialistic SODs depends on metal availability and contingent growth conditions (24, 26, 44). Even in E. coli, competition between iron and manganese for nascent MnSOD polypeptide chains occurs in vivo, leading to copurification of variously enriched dismutases (5). In other facultative bacteria, manganese is preferentially incorporated in the presence of oxygen, whereas anaerobiosis facilitates iron binding (1, 15). This is indeed the behavior exhibited by R. capsulatus cells, which accumulate Fe-enriched SOD under phototrophic conditions and mostly MnSOD in aerated media (Fig. 2). This trend might be facilitated by Fe release from bacterioferritin under photosynthetic culture regimes (29). Then, Rhodobacter fluctuations between aerobiosis and the microaerobic environment prevailing during photosynthetic growth, together with the concomitant changes in metal availability, might have encouraged the evolution of a cambialistic SOD that can bind (and function with) either metal, depending on the external conditions imposed upon the growing bacteria.