Superoxide dismutases (SODs, EC 184.108.40.206) belong to an important group of antioxidant metalloenzymes. Multiple SODs exist for scavenging of reactive oxygen species (ROS) in different cellular compartments to maintain an intricate ROS balance. The present study deals with molecular and biochemical characterization of CuZn SOD encoded by LOC_Os03g11960 (referred to as OsCSD3), which is the least studied among the four rice isozymes. The OsCSD3 showed higher similarity to peroxisomal SODs in plants. The OsCSD3 transcript was up-regulated in response to salinity, drought, and oxidative stress. Full-length cDNA encoding OsCSD3 was cloned and expressed in Escherichia coli and analyzed for spectral characteristics. UV (ultraviolet)–visible spectroscopic analysis showed evidences of d–d transitions, while circular dichroism analysis indicated high β-sheet content in the protein. The OsCSD3 existed as homodimer (∼36 kDa) with both Cu2+ and Zn2+ metal cofactors and was substantially active over a wide pH range (7.0–10.8), with optimum pH of 9.0. The enzyme was sensitive to diethyldithiocarbamate but insensitive to sodium azide, which are the characteristics features of CuZn SODs. The enzyme also exhibited bicarbonate-dependent peroxidase activity. Unlike several other known CuZn SODs, OsCSD3 showed higher tolerance to hydrogen peroxide and thermal inactivation. Heterologous overexpression of OsCSD3 enhanced tolerance of E. coli sod double-knockout (ΔsodA ΔsodB) mutant and wild-type strain against methyl viologen-induced oxidative stress, indicating the in vivo function of this enzyme. The results show that the locus LOC_Os03g11960 of rice encodes a functional CuZn SOD with biochemical characteristics similar to the peroxisomal isozymes.
Abiotic stress conditions are detrimental to all living organisms, and plants being sessile are more prone to stress-induced damage . An important aspect of most stress conditions is reactive oxygen species (ROS)-mediated cellular damage. ROS such as superoxide radical (), hydrogen peroxide (H2O2) and hydroxyl radical (·OH) are cellular metabolic byproducts and physiologically important; however, when their levels are elevated (e.g. in stress conditions), it leads to ‘oxidative stress’ . Cellular ROS homeostasis is important and maintained by several enzymatic and non-enzymatic antioxidant mechanisms . Unlike other ROS, the radical predominantly contributes towards oxidative stress in an indirect manner. It spontaneously dismutates to H2O2 that subsequently generates highly reactive ·OH radical in Fe2+-mediated Fenton reaction . Both H2O2 and ·OH radical are detrimental to several cellular components namely membrane lipids, nucleic acids and proteins . Therefore, scavenging is important and constitutes the very first line of defense, where superoxide dismutases (EC 220.127.116.11) play a crucial role in reducing and ·OH levels [5–7]. Superoxide dismutases efficiently catalyze the disproportionation of by increasing the rate constant of the spontaneous dismutation by >1000-fold . Based on the metal cofactors, SODs are categorized into four types: CuZn, Mn, Fe, and Ni SOD. Plants contain only three SOD types localized to cytosol (CuZn SOD), chloroplast (CuZn SOD and Fe SOD), mitochondria (Mn SOD), and peroxisomes (CuZn SOD and Mn SOD) [6,8,9]. CuZn SODs (also referred to as CSDs) are reported to be the most abundant isozymes in plants . CuZn SODs generally exist as homodimers containing non-covalently linked Cu2+ and Zn2+ ions in each subunit ; however, the homo-tetrameric  and monomeric forms  have also been reported. While Cu2+ is important for the catalytic activity of the enzyme, Zn2+ is involved in its dimerization and structural stabilization [12,13]. Hydrophobic interactions between the subunits do not contribute towards enzyme activity as the monomers are also catalytically active [12,14]. Although the various CuZn SOD isozymes show sequence heterogeneity, structural and functional important residues and secondary structure elements are highly conserved .
The CuZn SOD isozymes have been studied and characterized in a variety of plants [7,14–21]. Furthermore, the role of SODs in protection against oxidative stress including overexpression for enhanced stress tolerance in transgenic plants has also been studied [7,20,22–27]. However, recent reports on certain CuZn SODs from plants have provided new information about their unusual stability [15,28], specific activity , and biochemical/biophysical characteristics of mono- and dimeric forms . Such beneficial characteristics of enzymes are often desirable for agricultural or industrial applications [15,28]. Therefore, analyses of uncharacterized/putative isozymes would provide information of their biochemical/biophysical characteristics vis-à-vis other known SODs, and new insights into their role in stress responses.
Rice has four CuZn SODs encoded by different genomic loci: OsCSD1 (LOC_Os03g22810, cytosolic), OsCSD2 (LOC_Os08g44770, chloroplastic), OsCSD3 (LOC_Os03g11960), and OsCSD4 (LOC_Os07g46990, cytosolic). Most previous studies have been focused primarily on OsCSD1 and OsCSD2 isozymes [11,18,29]. Information available on OsCSD3 include limited transcript expression data under certain abiotic stresses [21,30–32]. Therefore, functional characterization of OsCSD3 protein, its characteristics vis-à-vis other CSD isozymes, and association with stress tolerance in rice would be worth exploring.
Here, we present the first report on the functional characterization of OsCSD3 encoded by LOC_Os03g11960 locus (Figure 1) in the rice genome. We demonstrated that this gene was up-regulated in response to multiple abiotic stresses. Recombinant protein purified from Escherichia coli was found to be a homodimeric CuZn SOD enzyme, active over a broad pH range and relatively more tolerant to H2O2 and thermal inactivation. Interestingly, the OsCSD3 also exhibited bicarbonate-dependent peroxidase activity. More importantly, this enzyme showed significant protection against oxidative stress in both wild-type and sod double-knockout (ΔsodA ΔsodB) mutant strains of E. coli. Thus, the rice OsCSD3 is a functional CuZn SOD with peroxisomal features.
Schematic diagram of the rice Os03g11960 CuZn superoxide dismutase gene.
Plant material and experimental conditions
Rice genotype NSICRc106 (salt tolerant) used in this study was obtained from International Rice Research Institute (IRRI, Philippines). Rice seedlings were grown hydroponically in Hoagland media (Himedia, India) in MLR-351H plant growth chamber (Sanyo, Japan) under a 14 h light and 10 h dark period. Following growth conditions used: light intensity: 150 µmol m−2 s−1 (photon flux density), temperature: 28 ± 1°C (light period) and 26 ± 1°C (dark period) and humidity: 65%. For stress treatment, the 6-day-old rice seedlings were subjected to salinity (150 mM sodium chloride, NaCl), drought (15% polyethylene glycol, PEG), and oxidative stress (10 µM methyl viologen, MV). Shoot tissue was collected at different time-points from control and stressed seedlings, frozen in liquid nitrogen, and stored at −70°C until further use. Chemicals and reagents used (if not specifically mentioned) were from Sigma–Aldrich (U.S.A.), and common molecular biology protocols were followed as per Sambrook and Russel .
RNA isolation, full-length cDNA isolation and cloning
Total RNA was isolated from rice shoot tissue by the Trizol (Invitrogen, U.S.A.) method and assessed for quality and quantity. RNA preparation was treated with DNase I (Roche Diagnostics, Germany) to remove DNA contamination and it was subsequently heat-inactivated. Total RNA (10 µg) was reverse transcribed using SuperScript II reverse transcriptase (Invitrogen, U.S.A.) and a mixture of anchored oligo(dT)35 and random nonamers (dN9) (New England Biolabs, U.S.A.) as per the protocol recommended by the manufacturer. The quantity of complementary DNA (cDNA) was estimated on a UV-1800 spectrophotometer (Shimadzu, Japan) and used for quantitative real-time PCR analysis.
Full-length OsCSD3 cDNA was PCR amplified using a forward (Os11960-ForP: 5′-GGGAATTCCATATGATGGCAGGGAAAGCCGGCGGCC-3′) and a reverse (Os11960-RevP: 5′-CGGAATTCTTAAACTGCAGATCGAAGTCCAATGATAC-3′) primer (Figure 1) designed using the sequence available at the Rice Genome Annotation Project website (http://rice.plantbiology.msu.edu) . Underlined bases indicate the sites for restriction enzymes in the forward (NdeI) and reverse (EcoRI) primers for cloning compatibility in pET28a(+) vector (Novagen, U.S.A.). PCR was carried out using Pwo DNA polymerase (Roche Diagnostics, Germany) on a Mastercycler Gradient PCR machine (Eppendorf, Germany) using following conditions: initial denaturation at 94°C (5 min), 35 cycles of 94°C for 1 min, 60°C for 1 min, 72°C for 1.5 min, and final extension at 72°C (5 min).
The PCR product was purified, double-digested with NdeI and EcoRI, and ligated to pET28a(+) linearized with the same enzymes. The ligation mixture was transformed into E. coli (DH5α) cells and transformants were selected on LB agar containing kanamycin (25 µg ml−1). Transformants were screened by colony PCR for the presence of insert (Os03g11960 cDNA), and also confirmed by restriction analysis of plasmid. The full-length cDNA product and insert from the recombinant plasmid (referred to as pET28a(+)-OsCSD3) of the two clones was sequenced and submitted to the GenBank database (accession numbers: KF953542, KF953543). One of these (KF953542) was used for heterologous expression of rice OsCSD3 in E. coli BL-21(DE3) strain. The rice cDNA was also cloned in pMAL-c5x plasmid (New England Biolabs, U.S.A.) using the same restriction enzymes (NdeI and EcoRI) for protein expression and stress tolerance studies in E. coli sodA sodB double mutant.
Quantitative RT-PCR analysis
Quantitative RT-PCR analysis for comparative transcript abundance of four rice CuZn SODs was carried out using the following sets of primers: (1) OsCSD3 (ForP: 5′-CAACGGCTGCAACTCTACCGGGCC-3′ and RevP: 5′- GGTCCTTTATGAAGATATCTGCAACAC-3′), (2) OsCSD1 (ForP: 5′-GCATGTCAACTGGGCCACACTACA-3′ and RevP: 5′-CATGGATATTAGCAACACCATCTTC-3′), OsCSD2 (ForP: 5′-TGGGTGCATATCAACAGGACCACA-3′ and RevP: 5′-GGTTGCCTCAGCTACACCTTCAGC-3′), and OsCSD4 (ForP: 5′-CATGTCAACTGGACCACACTTCAA-3′ and RevP: 5′-ATTGACATTAGCAACACCATCTGC-3′). Wherever possible, primers were designed at the exon–exon junctions with amplicon size <200 bp. Primer specificity was assessed by the Primer-BLAST program available at the National Centre for Biotechnology Information (NCBI) website (http://www.ncbi.nlm.nih.gov/tools/primer-blast).
Quantitative RT-PCR analysis was carried out on a Master cycler ep realplex PCR instrument (Eppendorf, Germany) using SYBR Green Jumpstart Taq Ready mix (Sigma–Aldrich, U.S.A.) following the recommended protocol. Following thermal cycling conditions were used: 94°C (2 min), 45 cycles of 94°C (10 s), 60°C (15 s) and 68°C (20 s). Specificity of the PCR amplification was assessed by melting curve analysis as well as on 2.0% agarose gel. Relative transcript abundance was estimated by Schmittgen and Livak  using Actin-2 (LOC_Os10g36650) as a reference gene (ForP: 5′-CTAGTGGACGTACTACTGGTATTG-3′ and RevP: 5′-GATCCCTACCAGCAAGATCAAGAC-3′) and validated by using a second reference gene, GAPDH (LOC_Os08g03290) (ForP: 5′-GTGACAGCAGGTCGAGCATCTTCG-3′; RevP: 5′-GTCGATGACACGGTTGCTGTAACC-3′). Analysis was carried out using two independent biological replicates, where 8–10 rice seedlings were pooled for RNA extraction in each sample. All the PCRs were run in triplicates and appropriate controls including the no template control were included for each set of analysis. Statistical analysis of the qPCR data in control and treated samples was carried out by Student's t-test and the differences were considered significant only when the P < 0.05.
Construction of E. coli sod double-knockout (ΔsodA ΔsodB) mutant strain
Single-gene sodA and sodB knockouts of E. coli (BW25113) were obtained from Keio collection . The mutants contained removable (FLP recombinase-mediated) kanamycin resistance gene (KanR) cassette in place of the sod gene(s). Plasmid pCP20 that features temperature-sensitive origin of replication, a heat-inducible Flp-recombinase, and an ampicillin resistance gene (AmpR) as selection marker was used for removal of KanR cassette. E. coli sod double-knockout (ΔsodAΔsodB) strain was generated using the FLP-FRT recombination system as per the protocol described by Narita and Peng  with minor modifications. Briefly, E. coli sodA mutant containing KanR cassette (in place of sodA gene) was transformed with pCP20 plasmid, followed by selection on LB agar ampicillin plates at 30°C. Few positive cells were grown at 43°C in LB media, serially diluted, and plated on LB agar plates for Flp recombinase-mediated removal of KanR cassette, and curing of pCP20 plasmid. Loss of KanR and AmpR was confirmed by patching clones on LB agar plates containing no or appropriate antibiotics. These sodA cells were designated as sodA-KanS (S: sensitive). E. coli sodB mutant was transduced with P1 phage, and the lysate was used to transduce the E. coli sodA-KanS cells as per the standard protocol. The transformed cells were selected on LB agar kanamycin plates to identify sodAsodB double-knockout mutants.
Oxidative stress tolerance in E. coli sod double-knockout (ΔsodAΔsodB) mutant and wild-type strains
Effect of MV-induced oxidative stress on growth of E. coli sod double-knockout mutant (ΔsodAΔsodB) and wild-type BL-21(DE3) strains was carried out as per Goulielmos et al. . Briefly, E. coli double-mutant cells containing either pMAL-c5x (control vector) or pMAL-c5x-OsCSD3 were grown at 37°C in LB medium containing kanamycin (25 µg ml−1) and ampicillin (50 µg ml−1). Cultures were diluted (1 : 100) in LB–Kan–Amp media supplemented with 0.2 mM CuCl2 and ZnCl2, grown further until absorbance (A600 nm) reached ∼0.2, and induced with 0.5 mM isopropyl β-d-thiogalactopyranoside (IPTG) for 4 h, for expression of recombinant OsCSD3 in E. coli. The cultures were again normalized to absorbance (A600 nm) ∼0.2, treated with different concentrations (0.0–0.025 mM) of MV in the presence of 0.5 mM IPTG, for 20 h at 25°C, and growth was measured at regular intervals. The cultures were also diluted serially in 0.85% saline and analyzed by spot test (10 μl of respective dilutions were spotted) on LB agar plate containing both kanamycin (25 µg ml−1) and ampicillin (50 µg ml−1). Oxidative stress tolerance was also analyzed in wild-type E. coli strains, BL-21(DE3) containing pET28a(+) (control vector), or pET28a(+)-OsCSD3. Transformed E. coli cells were induced with 0.1 mM IPTG for 16 h at 20°C and normalized to absorbance (A600 nm) ∼0.2. These cells were treated with different concentrations (0.0–0.2 mM) of MV and absorbance (A600 nm) was measured at regular interval. Additionally, the cultures were serially diluted in saline and analyzed by spot test (10 μl of respective dilutions were spotted) on LB agar plate containing kanamycin (50 µg ml−1). The experiment was repeated three times in triplicates. Statistical analysis was carried out by Student's t-test and the differences were considered significant only when P < 0.05.
Lipid peroxidation analysis
Lipid peroxidation was determined by Panat et al.  with minor modifications. The assay is based on quantification of malondialdehyde (MDA), an end product of lipid peroxidation that reacts with 2-thiobarbituric acid (TBA) to produce a red colored adduct. Briefly, E. coli cells were harvested, washed (twice) with phosphate buffer saline (PBS, pH 7.4), resuspended in 300 µl of PBS, and lysed by sonication in a Branson Digital Sonifier 450 (Branson Ultrasonics Corporation, U.S.A.) using the following settings: amplitude: 35%, ON time 2 s, OFF time 2 s, processing time 5 min. The samples were kept on ice (4°C) during the sonication. Sample volume of 300 µl was mixed with 900 µl of TBA reagent (0.375% TBA, 0.25 M HCl, 15% trichloroacetic acid and 6 mM Na2EDTA) and incubated at 95°C for 30 min. Samples were cooled to room temperature and centrifuged at 13 680×g for 20 min at 4°C. MDA equivalents were estimated in the supernatant by measuring the fluorescence (excitation/emission: 530/590 nm) on a multi-well-plate reader (Infinite M200, Tecan, U.K.). The experiment was repeated three times in triplicates. The lipid peroxidation was expressed as µmoles of MDA equivalents mg−1 protein using 1,1,3,3-tetramethoxypropane (TMP) as a standard.
The plant CuZn SODs (cytosolic, peroxisomal and chloroplastic) similar to the rice OsCSD3 protein were identified by the Basic Local Alignment Search Tool, BLASTP , and sequences from 9 monocots and 12 dicots (Supplementary Table S1) were retrieved from the GenBank database at the NCBI website (http://www.ncbi.nlm.nih.gov). Parameters like molecular weight (Mw) and theoretical isoelectric point (pI) of the sequences were estimated by the ‘Compute pI/Mw tool’ at the ExPASy web site (http://web.expasy.org/compute_pi/). Multiple sequence alignment was done by ClustalX software  using default gap opening and gap extension penalties and edited by BioEdit Software . Sequence divergence and genetic relationships were inferred using the Molecular Evolutionary Genetic Analysis software (MEGA version 4) . Phylogenetic analysis was done by the neighbor-joining  method and statistical analysis was performed by the bootstrap method .
The molecular model of the rice OsCSD3 was generated using the SWISS-MODEL  work space (http://swissmodel.expasy.org/workspace) using a dimeric target sequence of rice SOD, and Solanum lycopersicum CuZn SOD crystal structure (PDB ID: 3PU7) as the template. Template structure included both Cu2+ and Zn2+ cofactors. Structural superposition of the model onto the template was carried out using the molecular modeling software O . Figures of the molecular models were rendered using computer program PyMOL .
Purification of recombinant protein
E. coli BL-21(DE3) cells containing the recombinant plasmid pET28a(+)-OsCSD3 were grown overnight at 37°C in LB medium with kanamycin (25 µg ml−1) and diluted to 1 : 100 in the fresh LB-Kan medium supplemented with 0.2 mM CuCl2 and ZnCl2. Cells in mid-log phase were induced with 0.1 mM IPTG at 20°C, harvested after 16 h by centrifugation (4500×g, 10 min), resuspended in the resuspension buffer (50 mM potassium phosphate; 100 mM potassium chloride, KCl; and 1 mM phenylmethanesulfonyl fluoride, PMSF, pH 8.0), and lysed by sonication with following settings: amplitude: 35%, ON time 2 s, OFF time 2 s, processing time 15 min (repeated twice). The samples were kept on ice (4°C) during the sonication. Supernatant and pellet fractions were separated by centrifugation (10 000×g, 20 min, 4°C), resolved on 15% SDS–PAGE, and stained with Coomassie Blue R-250 (Sigma, U.S.A.). The recombinant rice protein was purified from the supernatant fraction obtained from a total of 4.0 gFW−1 (gram fresh weight) of IPTG-induced E. coli cells. Affinity purification was carried out using 0.5 ml Ni-NTA affinity resin (Qiagen, Germany) as per the recommended protocol and analyzed on 15% SDS–PAGE. The purified protein was dialyzed against resuspension buffer using 2 kDa MWCO dialysis tube (Sigma–Aldrich, U.S.A.) and concentrated using 3 kDa MWCO centrifugal columns (Vivaspin, GE Healthcare, U.S.A.). Protein samples were quantitated by the Bradford method  using bovine serum albumin (BSA) as a standard.
Determination of metal content in the protein
Copper and zinc content of the protein was estimated on a VG PQExCell Inductively Coupled Plasma-Mass Spectrometer (ICP-MS) instrument (VG Elemental, U.K.). Briefly, recombinant protein (1 mg) was dialyzed against deionized water using 2 kDa MWCO dialysis tubing (Sigma–Aldrich, U.S.A.). The protein sample (in a silica crucible) was charred at 500°C in a blast furnace and the inorganic ash was resuspended with 1% supra-pure nitric acid solution for ICP-MS analysis. The metal content was expressed as the number of Cu2+ and Zn2+ ions per protein subunit.
The native molecular weight of the recombinant OsCSD3 was determined by gel filtration chromatography on a Superdex™ 75 10/300 GL chromatography column (GE Healthcare) preequilibrated with potassium phosphate buffer (10 mM potassium phosphate; 100 mM KCl, pH 8.0). The column was precalibrated with following standard proteins, BSA (66.5 kDa), chicken egg albumin (45 kDa), carbonic anhydrase (29 kDa), and cytochrome C (12.4 kDa). The native molecular weight was estimated by plotting a calibration curve of the log of molecular weight of the standard proteins versus their respective Ve/Vo values (Ve: elution volume, Vo: void volume). The molecular weight of the monomeric subunit was determined by 15% SDS–PAGE using prestained protein molecular weight standard (New England Biolabs, U.S.A.). A standard curve of the log of molecular weight of the standard proteins versus their Rf values (Rf: migration distance of the protein/migration distance of the dye front) was used for estimating the subunit molecular weight.
Spectrophotometric characteristics of the recombinant OsCSD3 were analyzed at room temperature (25°C) on a double-beam UV-1800 spectrophotometer (Shimadzu, Japan). Absorbance spectra of protein (1 mg ml−1) in potassium phosphate buffer (10 mM, pH 8.0) containing KCl (10 mM) were recorded between wavelengths of 200 and 800 nm. Circular dichroism (CD) analysis was carried out on a Jasco J-815 CD spectropolarimeter (Jasco, Japan) at 20°C, using protein sample (concentration: 0.15 mg ml−1) in potassium phosphate buffer (10 mM, pH 8.0) containing KCl (10 mM). CD spectra was recorded between wavelengths 190–260 nm using a quartz cuvette (path length: 0.2 cm) with the following settings: scan speed: 20 nm min−1, data integration time: 2 s, data pitch: 0.1 nm, and band width: 1.0 nm. The measurements from five scans were averaged and corrected for the sample buffer. Protein concentration estimated by the Bradford method  was used to determine the molar ellipticity.
In solution, SOD activity of OsCSD3 was determined by the nitro blue tetrazolium (NBT) reduction assay in a multi-well-plate assay . In brief, purified protein (0–2 µg) was incubated in 200 µl assay volume containing sodium phosphate buffer (50 mM, pH 7.8), reduced nicotinamide adenine dinucleotide (NADH, 78 µM), NBT (50 µM), and EDTA (0.1 mM). Reaction was initiated by the addition of phenazine methosulfate (final concentration: 3.3 µM) and absorbance (A560 nm) was measured for 5 min on a multi-well-plate reader (Infinite M200, Tecan, U.K.). The biochemical assays were carried out using 1 U of the protein. One unit (1 U) of SOD activity is defined as the amount of protein required to inhibit the NBT reduction by 50%.
The effect of pH was measured in two ways (1: assaying enzyme at different pH and 2: incubation of enzyme at a pH followed by activity measurement at different incubation time) using the following buffers: sodium phosphate (pH 7.0–8.0), Tris–HCl (pH 7.0–9.0), and sodium bicarbonate (pH 9.0–10.8). Effect of temperature on the stability on the protein was evaluated as described earlier  with minor modification. For instance, the recombinant protein was preincubated in 50 mM sodium phosphate buffer (pH 7.8) at different temperatures (25, 37, 45, 50, 60, and 70°C) before the aliquots were taken for activity measurement. For measuring the effect of inhibitors like diethyldithiocarbamate (DDC), sodium azide (NaN3), and hydrogen peroxide (H2O2), the protein (5 U) was preincubated with increasing concentration of DDC (0.0–2.0 mM), NaN3 (0.0–10.0 mM), and H2O2, (0.0–10.0 mM) for 30 min. Enzyme activity was measured as described above.
In-gel SOD activity assay of OsCSD3 was carried out on non-denaturing as well SDS polyacrylamide gel as described by Chen et al. . Protein samples were analyzed by electrophoresis on 15% SDS–PAGE [10% for non-denaturing polyacrylamide gel electrophoresis (PAGE) gels]. The gel was rinsed in distilled water for 15 min (thrice) and incubated in ‘solution A’ (50 mM sodium phosphate buffer, pH 8.0 + 28 µM riboflavin + 28 mM tetramethylethylenediamine) for 30 min. Then, ‘solution B’ (50 mM sodium phosphate buffer, pH 8.0 + 1 mM NBT) was added and the gel was exposed to light (20 min) for color development and photographed on the gel-documentation system (Syngene, U.K.).
Bicarbonate-dependent peroxidase activity of OsCSD3 was determined by monitoring oxidation of dichlorodihydrofluorescein (DCFH) to 2′,7′-dichlorofluorescein (DCF) as per Zhang et al. . In brief, assay was carried out at 25°C in potassium phosphate buffer (100 mM, pH 7.4) containing 0.1 mM diethylenetriaminepentaacetic acid (DTPA), 50 µM DCFH, 25 mM sodium bicarbonate (NaHCO3), and increasing amount (0–10 µg) of protein. The reaction was initiated by the addition of H2O2 (1 mM) and fluorescence signal (emission: 524 nm; excitation: 480 nm) of the product DCF was monitored for 5 min. The analysis was carried out on a multi-well-plate reader instrument (Infinite M200, Tecan, U.K.). The formation of DCF was also monitored by recording the absorbance spectra on a UV-1800 spectrophotometer (Shimadzu, Japan).
Rice OsCSD3 CuZn SOD is up-regulated in response to abiotic stresses
Quantitative RT-PCR analysis estimated the basal transcript levels of four CuZn SOD isozymes in rice seedlings. Transcripts of OsCSD1 and OsCSD2 showed higher abundance and collectively represented >50% of the total CuZn SOD transcript (Figure 2A). The OsCSD3 transcript was ∼20% of the total transcript (Figure 2A). The rice CuZn SODs were up-regulated in response to abiotic stress conditions with minor variations in expression pattern. OsCSD3 showed >1.5-fold to 2.5-fold up-regulation under salinity, drought and oxidative stress (Figure 2B). OsCSD1 overall showed a similar pattern with enhanced transcript levels at 48 h under oxidative stress (Figure 2C). OsCSD2 transcript level were ∼1.5-fold up-regulated, except at 24 h (>2-fold) time-point under salt stress and ∼2-fold at 48 h in response to oxidative stress (Figure 2D). OsCSD4 showed higher up-regulation than other three isozymes under drought and salt stress (Figure 2E).
Quantitative RT-PCR analysis of rice CuZn SODs.
OsCSD3 showed high similarity to peroxisomal CuZn SODs
The OsCSD3 cDNA was successfully PCR amplified (Supplementary Figure S1) cloned, sequenced, and submitted to GenBank (accession numbers: KF953542 and KF953543). The NSICRc106 OsCSD3 polypeptide contained 164 amino acid residues with a predicted molecular weight of 16.5 kDa. The protein sequence showed 99.3% identity with Nipponbare OsCSD3 sequence with single variation at 151 position (A151T). The OsCSD3 showed both length and sequence heterogeneity with CuZn SOD isozymes from other plants (Supplementary Figure S2). Comparison of OsCSD3 with rice cytosolic CuZn SODs (OsCSD1 and OsCSD4) showed heterogeneity at ∼37% sites and length variation at N-terminal region (Supplementary Figure S2).
The multiple sequence alignment of CuZn SOD sequences from 9 monocots and 12 dicots (Supplementary Table S1) identified conserved and variable regions (Supplementary Figure S2). Phylogenetic analysis placed the SODs into three clusters: I (cytosolic), II (chloroplastic), and III (peroxisomal) supported by high bootstrap values, where monocot and dicot sequences showed distinct divergence (Supplementary Figure S3). The OsCSD3 CuZn SOD was placed in the monocot specific sub-cluster within cluster III. This cluster also included AtCSD3 from Arabidopsis thaliana (known to localize in peroxisomes)  and high isoelectric point SODs (PthipI-SODC1 and Pthip-SODC2) from Populus . CuZn SODs in peroxisomal cluster III exhibited higher divergence (p-distance: 0.239) than the cytosolic (p-distance: 0.149) and chloroplastic isozymes (p-distance: 0.224). The theoretical pI of CuZn SODs including OsCSD3 protein (pI: 6.82) in this cluster were generally higher than cluster I and II sequences, in both monocots (6.38–7.19) and dicots (6.22–7.87). Sequence comparison of monocot and dicot CuZn SODs from peroxisomal cluster III revealed high conservation of important residues involved in coordination with Cu2+ (His-55, His-57, His-72, and His-129) and Zn2+(His-72, His-80, His-89 and Aps-92), disulfide bond formation (Cys-66 and Cys-155) and surface interactions (Supplementary Figure S4). Sequence heterogeneity was more predominant towards N-terminal than C-terminal or middle region of the CuZn SODs (Supplementary Figure S4). Despite overall similarity to peroxisomal isozymes, the characteristic peroxisomal targeting signals could not be detected in OsCSD3.
Recombinant OsCSD3 is a homodimer and contains Cu2+ and Zn2+ cofactors
Rice OsCSD3, cloned in pET28a(+), was overexpressed in E. coli BL-21(DE3) as described in Experimental section. Recombinant protein (∼18 kDa) was expressed in cells containing pET28a(+)-OsCSD3 plasmid (lane I, Figure 3A) and was not seen in uninduced cells (lane U, Figure 3A). The yield of the recombinant OsCSD3 was ∼0.6 mg gFW−1 of IPTG-induced E. coli cells. In-gel SOD activity assay on non-denaturing PAGE showed that recombinant OsCSD3 is enzymatically active (lane 2, Figure 3B). The SOD activity of OsCSD3 was lost upon treatment with 10 mM β-mercaptoethanol (lane 4, Figure 3C), suggesting the importance of disulfide bond in the functionality of this protein. These results showed that LOC_Os03g11960 locus in the rice genome contains a coding sequence that expressed to SOD of expected molecular weight (18.7 kDa: 16.5 kDa, OsCSD3 and 2.2 kDa from pET28a(+) plasmid) in E. coli. Gel-filtration chromatography analysis showed that protein elutes as a single peak, corresponding to a molecular mass of ∼36 kDa (Supplementary Figure S5). Since the estimated size of the subunit was ∼18 kDa (Figure 3C), the OsCSD3 is suggested to be a homodimer in native form. ICP-MS analysis confirmed the presence of both Cu2+ and Zn2+ metal cofactors at concentration of 0.905 ± 0.18 and 0.945 ± 0.16 per protein subunit. Collectively, these results show that the rice protein is a CuZn SOD.
Overexpression, purification, and in-gel activity analysis of recombinant OsCSD3.
Recombinant OsCSD3 showed spectral signatures of a functional structure of CuZn SOD
The OsCSD3 isozyme contained a single tyrosine residue (Tyr-37) and six phenylalanines (at positions 29, 30, 54, 59, 73, and 107) and no tryptophan. The UV (ultraviolet) absorption spectra showed specific peaks associated with these aromatic amino acid residues in the protein (Figure 4A). The spectra in the visible region showed evidence of electronic transitions in two regions: (a) ∼620–680 nm, indicative of d–d electronic transition for Cu2+ coordination and (b) ∼380–410 nm, attributed to ligand-to-metal interaction between imidazole ring of the histidine (His-62) and Cu2+ (Figure 4B). CD spectroscopic analysis revealed higher β-sheet and low α-helix content in the rice protein (Figure 4C). These results indicated that the OsCSD3 has all the spectral signatures that qualify it to be a CuZn SOD.
UV–visible and CD spectroscopy analysis of OsCSD3.
OsCSD3 is a CuZn SOD with higher tolerance to thermal inactivation and H2O2
Purified OsCSD3 showed a specific activity of 4500 ± 500 U mg−1 protein (Supplementary Figure S6). The enzyme showed maximum activity at pH 9.0; however, it was active over a broad pH range (7.0–10.8) with 60–80% residual activity (Figure 5A). Stability of OsCSD3 was comparable from pH 7.0 to 10.0, and 40–60% loss in activity was observed beyond pH 10.0 within 18 h of incubation (Figure 5B). OsCSD3 showed tolerance to thermal inactivation with no substantial loss of activity at 50°C and a T½ value (temperature at which 50% of activity is lost) of 62°C (Figure 5C). The rice SOD retained 20% activity even on 40 min preincubation at 80°C.
Biochemical characteristics of OsCSD3.
OsCSD3 showed a differential response to SOD-specific inhibitors. SOD activity of OsCSD3 was inhibited by DDC (a CuZn SOD-specific inhibitor) in a concentration-dependent manner (IC50 0.5 mM), with complete inhibition at 2 mM (Figure 6). The OsCSD3 was relatively insensitive to H2O2 as it exhibited an IC50 (H2O2) value of ∼7.6 mM (Figure 6). OsCSD3 activity remained unaffected by sodium azide (NaN3) (Figure 6). OsCSD3 also exhibited bicarbonate-dependent peroxidase activity as assessed by enzyme-mediated oxidation of DCFH to DCF (Supplementary Figure S7A,B).
Effect of SOD inhibitors on OsSCD3 activity.
OsCSD3 enhanced oxidative stress tolerance of E. coli sod double mutant and wild type
The ability of rice OsCSD3 to protect E. coli cells from MV-mediated oxidative stress was investigated. Growth of wild-type BL-21(DE3) cells containing pET28a(+) or pET28a(+)-OsCSD3 was evaluated in the presence of increasing MV concentration. Overexpression of OsCSD3 enhanced the tolerance to MV-mediated oxidative stress (up to 0.2 mM MV) in the wild-type BL-21(DE3) cells (Figure 7A,C). However, the wild-type cells contain inherent SOD isozymes SODA and SODB which constitutes ∼97% SOD activity in the cell. To evaluate the protection conferred solely by OsCSD3, a sod double-knockout (ΔsodA ΔsodB) mutant was generated (detailed in Experimental section), and loss of both SODA and SODB activities was confirmed by the in-gel SOD assay (Supplementary Figure S8A). E. coli sodAsodB double mutant showed high sensitivity to MV treatment (Supplementary Figure S8B) than the sodA (Supplementary Figure S8C) and sodB single mutants (Supplementary Figure S8D).
Analysis of oxidative stress tolerance of E. coli BL-21(DE3) wild-type and sod double-knockout (ΔsodAΔsodB) mutant cells containing empty vector or recombinant vector expressing rice OsCSD3.
The pET28a(+)-OsCSD3 plasmid construct was compatible for heterologous expression in E. coli BL-21(DE3) host, but not suitable for sod double mutant (ΔsodA ΔsodB) as it lacked T7 RNA polymerase-dependent overexpression, and contained selection marker (kanamycin) similar to pET28a(+). Hence, plasmid vector pMAL-c5x [containing maltose-binding protein (MBP) tag] that was suitable for T7 RNA polymerase-independent expression and contained an ampicillin selection marker was used. Additionally, the presence of N-terminal MBP tag kept the recombinant OsCSD3 in the soluble form. For functional analysis in E. coli double mutant, plasmid pMAL-c5x-OsCSD3 was transformed into sod double mutant and activity of the OsCSD3 was confirmed (Supplementary Figure S8E). Growth of sod double-knockout mutant containing pMAL-c5x or pMAL-c5x-OsCSD3 was evaluated in the presence of increasing MV concentration. The double mutant showed enhanced tolerance at 10 µM MV (Figure 7B,D). The sod double mutant was also evaluated for oxidative stress-induced lipid peroxidation. MV treatment resulted in increased lipid peroxidation in the sodA sodB double mutant. However, the cells expressing OsCSD3 showed reduced lipid peroxidation compared with the control (Figure 8). This indicates that the OsCSD3 is able to protect the E. coli cells from the MV-mediated oxidative damage.
Analysis of lipid peroxidation in E. coli sod double-mutant (ΔsodA ΔsodB) cells containing empty vector (pMAL-c5x) or pMAL-c5x-OsCSD3 grown in control (no MV) and MV-treated (10 µM MV) conditions.
Homology model shows structural features conserved in OsCSD3
The homology model of OsCSD3 was generated using the crystal structure of S. lycopersicum CuZn SOD (PDB ID: 3PU7) as a template (Figure 9). The two sequences showed 64% identity (E-value of 2.26834 × 10−39; Supplementary Figure S9) in the template identification module of Swiss model workspace. The template structure, refined to 1.8 Å resolution, included both Cu and Zn cofactors. Homology model was a homodimer covering the residue range 19–160 with QMEA-N-score 4 and QMEAN4-Z-score of 0.782 and −0.21, respectively , and showed the Greek key β-barrel structural core typical of CuZn SODs. The alpha carbons of residues 19–31, 37–160 of both the chains of the model were superposed onto the corresponding residues of the template, thus omitting a beta turn where there is an insertion of a proline residue in case of rice model (Figure 9A). Structural superposition showed that the positions of active site residues (His-55, His-57, His-72, His-129, His-80, His-89 and Asp-92 involved in coordination with the metal cofactors (Cu and Zn), cysteines (Cys-66 and Cys-155) involved in intra-subunit disulfide bond), and catalytically important second-sphere residues (Asp-133 and Arg-152) were highly conserved. The two structures superposed to a root-mean-square deviation (RMSD) of 0.08 Å. After the superposition, active site of the model could very well accommodate both the metal cofactors (Cu and Zn) from the template (Figure 9B).
Homology model of 3D structure of OsCSD3 generated using SWISS-MODEL work space.
Oxidative stress mediated by ROS is minimized by coordinated action of multiple antioxidant enzymes in plants [2,6]. Since superoxide radical participates in generation of several ROS and RNS (reactive nitrogen species), its dismutation catalyzed by SOD reduces their levels [3,55]. Multiple CuZn SOD isozymes dismutate the membrane impermeable to minimize the oxidative stress in organelles actively involved in oxidative metabolism [6,8,56]. Despite importance of peroxisomes in oxidative metabolism , the peroxisome-specific SOD isozymes have been relatively less studied.
The present study involves functional characterization of a putative CuZn SOD encoded by LOC_Os03g11960 in the rice. We present evidence to suggest that rice OsCSD3 showed sequence, structural and biochemical characteristics similar to peroxisomal SODs in plants. Different SODs contribute towards antioxidant activity at different physiological stages  and stress conditions [25,27,30,32,58,59]. Isozymes specific to subcellular compartments are important for ROS homoeostasis within and for ROS signaling between the organelles . Changes in SOD activity at one location may affect gradient of H2O2 (also a signaling molecule) leading to modulation of redox-sensitive pathways . Therefore, for a better insight into dynamics of oxidative stress management at cellular level, it is necessary to understand role of each isozyme.
Among the four rice CuZn SODs, the chloroplastic isozyme was analyzed for tissue-specific expression pattern , while the cytosolic isozyme was expressed in E. coli to study certain enzyme characteristics . A rice Fe-SOD was also analyzed for stress responsiveness at transcript level . However, these SODs have not been characterized in detail for biochemical/biophysical properties. Similarly, OsCSD3 has not been studied thoroughly, except few reports on effect of anoxia , salinity , ozone  and gamma radiation  on OsCSD3 transcript levels. Stress responsiveness of this locus under drought and oxidative (observed in this study) indicates this to be a general abiotic stress responsive gene.
The rice OsCSD3 showed homodimeric subunit organization, metal cofactor composition (1 Cu2+and 1 Zn2+ per subunit) and predicted structural features similar to other plant CuZn SODs [7,12,14]. In silico analysis showed the presence of single intra-subunit disulfide bond between Cys66 and Cys165, which seems crucial for the functional conformation of OsCSD3 protein. Additional biochemical evidences confirmed that OsCSD3 is indeed a CuZn type of isozyme that also exhibits bicarbonate-dependent peroxidase activity reported in some other plant CuZn SOD isozymes [14,15].
Despite high conservation in structurally and functionally important residues, certain sequence variations affecting important characteristics (stability, oligomerization dynamics, specific activity) have been reported in some SOD isozymes [7,14,15,28]. Such characteristics have potential for biotechnological applications of the proteins [14,15,28]. The OsCSD3 showed remarkable properties (wide pH range, thermal stability, higher H2O2 and tolerance) that might qualify this protein to be a stress tolerant enzyme. These features may be attributed to variations specific to this isozyme.
Among various CuZn SOD isozymes, peroxisomal types are generally reported to be more thermostable [7,17,61]. The thermostability of OsCSD3 was comparable to the peroxisomal isozymes. It retained >80% SOD activity at alkaline pH, generally not observed for CuZn SODs [15,18,62], suggesting that the enzyme is more tolerant to pH-induced changes on subunit interactions, and Cu2+ leaching that may affect stability/activity [14,15,56]. At 4–5 mM concentration, H2O2 mediated by oxidation of metal (Cu2+) coordinating amino acid residues causes irreversible inactivation of most CuZn SODs [7,14,15,17,63]. Higher H2O2 tolerance (∼30% activity at 10 mM) of OsCSD3 further corroborates its peroxisomal nature. As peroxisomes contain higher H2O2 levels, the isozymes localized in this organelle are suggested to have adapted for higher H2O2 levels . The alkaline pH optima and higher H2O2 tolerance of OsCSD3 seem suitable for functioning in the alkaline environment of plant peroxisomes . Despite overall sequence similarity to peroxisomal SODs, the OsCSD3 lacked typical C-terminal peroxisomal signal peptide . Furthermore, its N-terminal region could not be modeled due to lack of good similarity with the template (Supplementary Figure S9). It may be speculated that this enzyme has additional N-terminal region and the possibility of its involvement in an alternative targeting mechanism of this protein cannot be ruled out.
In addition to generation of H2O2 and ·OH radical, the , in a membrane permeable protonated form (), can also initiate lipid peroxidation and damage cellular membranes . Oxidative stress-mediated damage resulting into cellular toxicity and cell death can be minimized by scavenging of . OsCSD3-mediated enhanced tolerance of wild-type and sod double-mutant E. coli cells subjected to MV (a redox-cycling agent for generation) demonstrated the in vivo scavenging function of this enzyme, resulting into reduced oxidative damage including lipid peroxidation. However, OsCSD3 alone was not sufficient to confer tolerance to higher levels of oxidative stress in E. coli sod double mutant lacking SODA and SODB. The in vivo scavenging capability of other SOD isozymes in E. coli cells has also been previously demonstrated [7,38].
This study comprises the first report on the expression and characterization of the CuZn SOD encoded by LOC_Os03g11960 in rice genome. The locus is responsive to multiple abiotic stresses and codes for a functional homodimeric, CuZn SOD. It also exhibits peroxidase activity and is capable of conferring protection against oxidative stress as demonstrated in E. coli. Overall sequence similarity, several structural, biochemical features suggest it to be a peroxisomal isozyme. The peroxisomal SODs are relatively less studied in plants; however, these are important component of antioxidant system in normal physiology and stress conditions [8,65] and contribute to ∼18% of total SOD activity . The rice OsCSD3, with higher thermal stability and tolerance to H2O2, may serve as a better candidate for abiotic stress tolerance enhancement in plants. Furthermore, experimental validation of its cellular localization may provide insights into involvement of non-canonical targeting sequences/mechanisms in rice, which contains a high diversity of potential PTS sequences .
ampicillin resistance gene
bovine serum albumin
CuZn superoxide dismutase
inductively coupled plasma-mass spectrometer
kanamycin resistance gene
maltose binding protein
polyacrylamide gel electrophoresis
phosphate buffer saline
reactive oxygen species
R.P.S. planned and executed experiments, analyzed results and wrote paper. A.S. (Amol) conducted experiments and analyzed results. V.P. conducted bioinformatic studies and results analysis. H.S.M. analyzed results and wrote the paper.
A.S. (Ajay) is a principal investigator, conceived idea, conducted experiments, analyzed results, wrote the paper and communicated.
The present study was supported by the institutional funding of Bhabha Atomic Research Centre, Mumbai, India. No separate funding was obtained from any other National/International funding body for the present study.
We thank Dr S. Chattopadhyay and Dr N. Jawali for the encouragement and support during this study, and Dr Sheetal Uppal, MBD, BARC for suggestions and comments. The authors are thankful to the International Rice Research Institute (IRRI), Philippines for providing the rice seed material used in this study.
The Authors declare that there are no competing interests associated with the manuscript.