Introduction

Colwellia psychrerythraea belongs to a group of strictly psychrophilic gamma proteobacteria, which thrive at temperatures below 5 °C (D’Aoust and Kushner 1972; Huston et al. 2000; Deming et al. 1988). It has been isolated from stably cold marine environments, including deep sea and Arctic and Antarctic sea ice (Junge et al. 2002). Brine channels within the sea ice provide a unique ice-associated life habitat for C. psychrerythraea and some other bacteria as well as some viruses, algae, protists, flatworms, and small crustaceans (Mock and Thomas 2005). Organisms living in or on sea ice face several extreme environments: concentrated salinity, low temperatures, high UV light, and dissolved gases (Thomas and Dieckmann 2002). C. psychrerythraea can be exposed to high levels of reactive oxygen species (ROS) through their own metabolism as well as the extracellular ROS within the brine channel by exposure to cold temperatures, increased solubility of oxygen, and high UV light (Neftel et al. 1984; King et al. 2005; Cooper and Zika 1983; Mock and Thomas 2005). To enhance membrane fluidity at the low temperatures, the bacterium produces extra polyunsaturated fatty acids and these can undergo lipid peroxidation by ROS (Halliwell and Gutteridge 1984; Methe et al. 2005; Barriere et al. 2001). Several antioxidant enzymes are encoded in the genome of C. psychrerythraea including superoxide dismutases (SODs) and catalases, indicating that the bacterium does have methods of coping with the oxygen toxicity (Methe et al. 2005). Since little is understood about oxidative stress in psychrophilic organisms in general, one key antioxidant enzyme produced in C. psychrerythraea, glutathione reductase (GR), was the focus of this study.

Reduced glutathione (GSH) is a tripeptide composed of glutamate, cysteine, and glycine that has numerous important functions within cells. It contains an unusual peptide linkage between the amine group of cysteine and the carboxyl group of the glutamate side chain (Hopkins 1929). Glutathione is ubiquitous in animals, plants, and microorganisms, and, being water soluble, is found mainly in the cell cytosol and other aqueous phases of living systems (Kosower and Kosower 1978; Lomaestro and Malone 1995; Ritz and Beckwith 2001). It is known to protect cells from toxins such as free radicals by serving as a reductant (Grant et al. 1996). The sulfhydryl group of GSH can be used to reduce peroxides. The resulting oxidized form of GSSG consists of two molecules of disulfide bonded together (GSH). Glutathione often attains millimolar levels inside cells, which makes it one of the most highly concentrated intracellular antioxidants (Kehrer and Lund 1994). GSH takes part in many different intracellular processes, including maintenance of reduced thiol groups, protection from oxygen-induced cell damage, and generation of deoxyribonucleotide precursors for DNA synthesis (Holmgren 1976; Kehrer and Lund 1994). GSH is regenerated in an NAD(P)H-dependent reaction catalyzed by glutathione reductase (Grant and Dawes 1996) as indicated:

$${\text{GSSG }} + {\text{ NAD}}\left( {\text{P}} \right){\text{H }} + {\text{ H}}^{ + } = \, 2{\text{GSH }} + {\text{ NAD}}\left( {\text{P}} \right)^{ + }.$$

Its role in replenishing pools of GSH is critical for maintaining a reducing environment within the cell (Huseby et al. 2009). The active site of GR is a redox-active disulfide bond which receives electrons from the bound FAD cofactor. In Escherichia coli, GR is a member of the dimeric FAD-containing thiol reductase family. It shares significant homology with the thioredoxin reductases from higher eukaryotes (Ritz and Beckwith 2001). One interesting aspect of glutathione metabolism in E. coli is that the ratio of reduced to oxidized glutathione does not appear to change significantly in mutants that lack glutathione reductase (Tuggle and Fuchs 1985).

There have been a small number of studies characterizing prokaryotic glutathione reductases such as from E. coli (Scrutton et al. 1987), Xanthomonas campestris (Loprasert et al. 2005), Streptococcus mutans (Yamamoto et al. 1999), Pseudomonas aeruginosa (Perry et al. 1991), Rhodospirillum rubrum (Libreros-Minotta et al. 1992), and the cyanobacterium Anabaena (Jiang et al. 1995). In this study, the glutathione reductase from the psychrophillic bacterium, C. psychrerythraea, was recombinantly expressed in E. coli and was characterized by investigating the biochemical, molecular, catalytic, and kinetic properties of the enzyme. Furthermore, in vivo complementation studies in a glutathione reductase-deficient E. coli strain were conducted to validate its functionality in a heterologous system over a broad range of temperatures.

Materials and methods

E. coli glutathione reductase mutant strain construction

To avoid background glutathione reductase (GR) activity, E. coli strains, BL21 (λDE3) and JM105, were subjected to disruption of the gene encoding glutathione reductase (gor). The gor mutant of BL21(λDE3) strain was used for overexpression and purification of C. psychrerythraea GR (Killens-Cade et al. 2014), and the gor mutant strain of JM105 was generated for the growth studies described in this work. GR was inactivated in the E. coli strains using the PCR-based method of one-step inactivation of chromosomal genes described by (Datsenko and Wanner 2000). Plasmid, pKD3, which contains a chloramphenicol resistance gene cassette, was used as the template for generation of PCR fragments encoding the chloramphenicol resistance marker and homologous DNA regions for the target gene (gor). The primers consisted of 60 bases, which include 40 that are homologous to the flanking region of the target genes and 20 bases that are complementary to the chloramphenicol resistance cassette (see Online Resource 1). The PCR products were gel purified (Qiagen) for electroporation. pKD46 is a helper plasmid that carries the phage λ Red recombinase that is l-arabinose inducible and can be cured from cells when the incubation temperature is raised to 37 °C. JM105 and BL21 (λDE3) were transformed with pKD46 and plated in LB supplemented with ampicillin (100 μg/ml) and 0.3 % glucose agar and were incubated overnight at 30 °C. Transformants carrying pKD46 were grown in 10 ml SOB medium containing 0.2 % of l-arabinose and ampicillin (100 μg/ml) at 30 °C until OD600 reached 0.6. The cultures were spun down at 9500 rpm in a Beckman C1015 rotor for 10 min and washed three times with ice-cold 10 % glycerol to make the cells electrocompetent. The cells carrying pKD46 were suspended in 100 μl of 10 % glycerol and 50 μl aliquots were placed into two separate microfuge tubes. The gel-purified gor-specific PCR products (concentration ~1 μg) were added to 50 μl of resuspended cells and no DNA was added to the other 50 μl of cells as a negative control. Electroporation was conducted by using a Gene Pulser (Bio-Rad, Hercules, CA) set at 2.5 V, 200 Ω, and 25 μF according to the manufacturer’s instructions. Shocked cells were added to 1 ml of pre-warmed SOC medium and incubated for 1 h at 37 °C. 100 μl of the cell suspension was then plated onto LB agar containing chloramphenicol (35 μg/ml) and incubated at 37 °C overnight. The remainder of the cells was incubated overnight at room temperature before plating the next day. The colonies were screened for mutation by isolating the DNA using the GNOME DNA Kit (Bio 101, Vista, CA), digesting with BstEII, and PCR amplifying with primers specific to the inner and outer region of the gene (see Online Resource 1 for primer sequences). The verified strains were designated MLJ700 for the gor-deficient JM105 strain and MLJ600 for the gor-inactivated BL21 (λDE3) strain (Table 1).

Table 1 Bacterial strains used in this study
Full size table

Cloning and purification of glutathione reductase from C. psychrerythraea strain 34H

The C. psychrerythraea GR gene was cloned into the pET-21b plasmid and overexpressed in the gor-deficient E. coli BL21 (λDE3) strain MLJ600 at room temperature, and the GR protein was purified as previously described (Killens-Cade et al. 2014). The specific activities of the recombinant GR at each purification step were determined and are presented in Online Resource 2. Purified recombinant GR was used for characterization.

To investigate the functionality of C. psychrerythraea GR in a heterologous system, it was cloned into the pTrc99A vector. The gene encoding C. psychrerythraea GR was amplified from genomic DNA using the following primers targeted to C. psychrerythraea 34H: forward primer containing an NcoI site 5′-CTAATGAGTGAACCCATGGCACAA-3′ and the reverse primer containing an SacI site 5′-GCAATCAACATTGAGCTCGCTTAA-3′. The following PCR amplification program was used: 35 cycles consisting of denaturation at 95 °C for 30 s, annealing at 56 °C for 30 s, and extension at 72 °C for 1 min. The C. psychrerythraea GR gene was amplified using iProof High-Fidelity DNA polymerase (BioRad). The resulting PCR product was gel extracted and ligated to the pTrc99A plasmid previously digested with NcoI and SacI. The ligated construct was transformed into E. coli XL-Blue competent cells (Stratagene, Santa Clara, CA). Plasmid DNA was isolated from the transformants using Qiagen Minipreps, and the plasmids were screened for inserts by visualization of DNA agarose gels. The cloned GR gene sequence was verified by sequencing (Eurofins MWG Biotech, High Point, NC).

Molecular weight determination

The molecular weight of the C. psychrerythraea GR was calculated using Mac vector software, and the estimation of the molecular weight under denaturing conditions was determined by SDS-PAGE analysis as previously described (Killens-Cade et al. 2014). In this study, a quantitative determination of the molecular weight was conducted using MALDI-TOF mass spectrometry analysis (NC State Mass Spectrometry Facility, Raleigh, NC). The native molecular weight of the recombinant C. psychrerythraea GR was determined by gel filtration chromatography using a HiPrep 16/60 Sephacryl S-200 HR column (GE healthcare, Pittsburgh, PA) equilibrated with solution A (0.05 M sodium phosphate buffer with 0.15 M NaCl at pH 7.0). The elution of recombinant C. psychrerythraea GR was compared to the elution of gel filtration standards (thyroglobin) (bovine):670 kDa; γ-globulin (bovine):158 kDa; ovalbumin (chicken): 44 kDa; myoglobin (horse): 7 kDa; vitamin B12: 1.35 kDa (Bio-Rad, Hercules, CA).

Enzyme assays, kinetic studies, and protein estimation

GR activity was measured spectrophotometrically with a Shimadzu Spectrophotometer UV2401-PC (Nakagyo-Ku, Kyoto, Japan) at various temperatures. The GR reaction mix (3 ml) included 75 mM potassium phosphate, 2.6 mM ethylenediaminetetraacetic acid, 1 mM oxidized glutathione, 0.09 mM β-nicotinamide adenine dinucleotide phosphate (reduced form), and 0.13 % (w/v) bovine serum albumin (Mavis and Stellwagen 1968). One unit of enzyme activity was defined as the amount of glutathione reductase that oxidizes 1 µM NADPH per min. K M and V max values for NADPH were calculated using the UVProbe software (Shimadzu, Japan) for kinetic studies. Activities were determined at 5 different NADPH concentrations (6, 13, 25, 50, 100 µM) in reactions containing 1 mM oxidized glutathione. Baker’s yeast glutathione reductase used for the mesophilic enzyme comparison was purchased from Sigma-Aldrich (St. Louis, MO). Protein concentrations were routinely estimated using the Bradford method with bovine serum albumin as the standard (Bradford 1976).

Sequence analysis and UV–visible spectra determination

For gene sequence analysis, orthologs of GR from other organisms were identified by BLAST sequence analysis by using the GeneBank database at the National Center for Biotechnology Information. Predicted amino acids sequences were aligned using the CLUSTAL W program. UV–visible spectra were recorded on a Shimdazu Spectrophotometer UV2401-PC at 25 °C. The reactions were prepared in 100 mM potassium phosphate buffer and 3.4 mM EDTA at pH 7.6. Purified C. psychrerythraea GR was reduced with 0.067 mM DTT under anaerobic conditions and GR was oxidized with 1 mM GSSG aerobically.

E. coli in vivo complementation study

The bacterial strains and plasmids used in this study are listed in Table 1. The ability of the C. psychrerythraea GR gene to complement an E. coli gor-deficient strain was examined by growth in LB media treated with the oxidizing agent diamide, where restoration of growth indicates complementation of E. coli GR. The plasmid, pTrc99A, was transformed into the E. coli strains JM105 and MLJ700 (JM105:Δgor), and pMLJ40 (C. psychrerythraea GR in pTrc99A) was transformed into MLJ700. Cultures grown overnight at 37 °C were used to inoculate 30 ml of LB broth, and they were subsequently grown at either 18, 25 or 37 °C shaking at 200 rpm. When OD600 measurements were between 0.15 and 0.25 (at early log phase), IPTG (0.1 mM) was added to every culture for induction. 0.2 mM or 0.3 mM of diamide was also added to some cultures to induce oxidative stress. The cultures were harvested (5 ml of each culture) once the non-diamide-treated cultures reached OD600 measurements between 0.8 and 1.0 (at mid to late log phase). The harvested cells were broken with 1.2 ml of B-per (Thermo Scientific, Waltham, MA) with lysozyme (10 mg/ml) and cell debris was pelleted using centrifugation. The cell extracts were assayed for GR activity at the same temperature as the growth studies that had been conducted.

To determine if C. psychrerythraea GR would provide protection against hydrogen peroxide or menadione which induces oxidative stress, E. coli complemented with C. psychrerythraea GR was challenged either with hydrogen peroxide (2 and 4 mM) or menadione (0.05 and 0.08 mM). The culture samples were prepared at room temperature as described above and monitored for growth.

Statistical analysis

Statistical analyses were conducted for the OD 600 nm values of the final time points for the growth study at each temperature using SAS 9.4 version. All estimates of sample variability are expressed as means of standard error (n = 3). Student’s two-tailed t tests (p value < 0.05) were performed to determine whether the growth responses were significantly affected by the levels of diamide or strain type at the three different temperatures used in the experiment. A two-way ANOVA with replication (n = 3) was performed for each temperature to validate whether the responses of the growth with respect to strain type were significantly different from the responses to the amount of diamide cells were exposed to and the possible interaction between these two factors. The analysis was considered significantly different if the strains, the diamide treatments, and the interaction had p values below 0.01. Graphs of marginal means were generated to visualize the characteristics of each factor (strain and diamide effect) and the interaction that may occur between them.

Results

C. psychrerythraea glutathione reductase sequence analysis

Glutathione reductase is a member of the pyridine nucleotide-disulfide oxidoreductase family of flavoenzymes, containing two active site electron acceptors, FAD, and a redox-active disulfide (Meister and Anderson 1983). C. psychrerythraea GR was compared with other GR sequences from various sources including bacteria, plant, and human (blood cells) using the Clustal W program (Fig. 1) to identify sequence similarities. C. psychrerythraea GR showed a high degree of sequence identity with GR sequences from E. coli (66 %) and human blood cells (49 %), but showed less identity to Arabidopsis GR (34 %). C. psychrerythraea GR has the conserved FAD binding motif GxGxxG(x17)E and NADPH binding motif (GXGYIAX18RX5R) that are typically present in the other GR homologs and are characteristic of flavoenzymes in general (Rescigno and Perham 1994; Jiang et al. 1995; Dym and Eisenberg 2001; Scrutton et al. 1990; Loprasert et al. 2005).

Fig. 1
figure 1

Alignment of the amino acid sequence of C. psychrerythraea glutathione reductase with other GRs from various organisms. Colwellia psy: Colwellia psychrerythraea, E. coli: Escherichia coli, Human: human blood cell, Cyanobacterium: Cyanobacterium anabaena PCC7120, Xanthomonas cam: Xanthomonas campestris, Pseudomonas aer: Pseudomonas aeruginosa, Arabidopsis tha: Arabidopsis thaliana. The x marks indicate the conserved motif for FAD binding, and the open circles mark the NADPH binding motif

Full size image

Recombinant C. psychrerythraea glutathione reductase molecular weight determination and spectral properties

The purified C. psychrerythraea GR enzyme migrated as a single band in SDS-PAGE with an apparent molecular weight of around 50 kDa (Killens-Cade et al. 2014). A molecular weight of 48.7 kDa was determined using MALDI-TOF mass spectrometry, which was similar to Mac Vector software’s estimation of 48.9 kDa. The molecular weight of the non-denatured enzyme as determined by gel filtration through a Sephacryl S-200 HR column (GE Health Care, Piscataway, NJ) was 95.24 kDa, suggesting that the recombinant C. psychrerythraea GR is a homodimer (Online Resource 3). The color of the purified C. psychrerythraea GR enzyme was yellow, and it had an absorption spectrum characteristic of the other flavoproteins (Lopez-Barea and Lee 1979; Garcia-Alfonso et al. 1993; Macheroux 1999), displaying peaks at 376 and 462 nm and a deep trough at 404 nm when it is in the oxidized form (Fig. 2).

Fig. 2
figure 2

Absorption spectra of C. psychrerythraea glutathione reductase under reduced (black line) and oxidized (gray line) conditions. 10 uM of purified GR was reduced with 0.067 mM DTT and oxidized with 1 mM of GSSG. The enzyme was dissolved in 100 mM potassium phosphate buffer and 3.4 mM EDTA at pH 7.6

Full size image

C. psychrerythraea glutathione reductase kinetic studies and heat stability

The kinetic parameters of the recombinant C. psychrerythraea GR were determined by comparing it with the mesophilic baker’s yeast GR using NADPH as a substrate when the enzymes were assayed at 4 and 25 °C (Table 2). The kinetic values of baker’s yeast GR assayed at 25 °C in this study are very similar to values reported in previous studies (Mavis and Stellwagen 1968; Massey and Williams 1965). As expected, both psychrophilic and mesophilic GR enzymes tend to exhibit lower K m values at their preferred environmental temperature (Lonhienne et al. 2001). The temperature dependence of the K m values allows the fine kinetic optimization of the enzymes to adapt to their environmental condition by modifying the thermodynamic properties of weak interactions for substrate binding (Aghajari et al. 1998; Feller and Gerday 2003). At 4 °C, C. psychrerythraea GR had a k cat three times higher and a K m four times lower compared to the mesophilic baker’s yeast GR. These properties of the enzyme described above exhibit a typical psychrophilic behavior (Bentahir et al. 2000). The heat stability of C. psychrerythraea GR was also assessed. The thermal half-life of the C. psychrerythraea GR was estimated to be 40, 8, and less than 2 min at 50, 55, and 60 °C, respectively. Also, 75 % of the activity remained after incubation at 45 °C for 3 h, and more than 90 % of the activity remained after incubation at 37 °C for 3 weeks.

Table 2 Comparison of kinetic parameters for recombinant C. psychrerythraea glutathione reductase and baker’s yeast glutathione reductase at different temperatures
Full size table

Construction and verification of the E. coli glutathione reductase mutant strains

To investigate the ability of the C. psychrerythraea GR to function in a heterologous system, the GR gene of E. coli strain JM105 was disrupted. The isolated strains (along with the parental type strain) were subjected to PCR analysis for verification of the targeted gene disruption. For the PCR analysis, sheared genomic DNA from each isolate and the control strain were used as the DNA templates and primers that are specific to either the outer region of the genes or located within the disrupted genes were used (see Online Resource 1). The expected sizes of the GR-disrupted DNA were 1.1, 1.35, 1.45, and 1.7 kb when the following primer combinations were used: 5′ gr:Cm specific and 3′ gr:Cm specific primers, 5′ outer region and 3′ gr:Cm specific primers, 5′ gr:Cm specific and 3′ outer region primers, and 5′ outer region and 3′ outer region primers, respectively. The expected DNA size when using the parental strain as the template DNA for the 5′ outer region and 3′ outer region primers was about 1.95 kb. PCR analysis of the parental strain (JM105) showed the expected sizes of the DNA fragments, and one of the isolates did produce PCR fragments that matched the expected sizes (Fig. 3). The verified strain was named MLJ700. GR enzyme assays indicated that negligible GR activity was observed in MLJ700, while the wild-type parental strain, JM105, showed approximately 45-fold greater activity (Table 3).

Fig. 3
figure 3

PCR verification of the E. coli gor mutant (MLJ700) using different combinations of gr:Cm specific or outer region specific primers

Full size image
Table 3 Glutathione reductase activities of the E. coli strains used in this study
Full size table

C. psychrerythraea glutathione reductase can functionally complement a glutathione reductase-deficient E. coli strain

To characterize the effect of heterologous expression of C. psychrerythraea GR at various temperatures, MLJ70 (JM105 with pTrc99A), and MLJ701 (gor-deficient JM105 transformed with pTrc99A), and MLJ702 (gor-deficient JM105 transformed with C. psychrerythraea GR/pTrc99A) were grown aerobically at three different temperatures (18, 25, and 37 °C) in LB media. Once the OD600 reached between 0.15 and 0.25 (early log phase), a final concentration of 0.1 mM IPTG was added for the induction of GR followed by the addition of 0.2 or 0.3 mM diamide as a source of oxidative stress. All cultures for each temperature condition were harvested when MLJ70 cultures (parental strain) that had not been treated with diamide reached mid to late log phase (OD600 between 0.8 to 1.0) and enzyme assays were performed on the cell-free extract to determine their GR activity. The results for the growth study presented in Figs. 4 and 5 demonstrated that diamide does affect the growth of all three strains regardless of the temperature, and these factors interact significantly at 25 and 18 °C. (see p values of diamide and strain*diamide in Online Resource 4).

Fig. 4
figure 4

Demonstration of functional complementation of E. coli glutathione reductase mutants with C. psychrerythraea glutathione reductase using oxidative stress growth studies. The experimental cultures were inoculated with enough inoculum to give a starting OD600 of 0.06. IPTG and diamide were added in the culture when the OD600 was between 0.15 and 0.25. Cells were harvested at the end points of the growth to investigate the GR activity (see Table 4). Growth studies were conducted at three different temperatures: 37 °C (a), 25 °C (b), and 18 °C (c). The error bars indicate the standard error of triplicate assay repeats. DA diamide

Full size image
Fig. 5
figure 5

The graphs of estimated marginal means were generated by averaging the three replications for the OD 600 values of the last time point for the growth study. Diamonds indicate the MLJ70 (parental strain), squares MLJ701(GR mutant), and triangles MLJ702 (overexpressing C. psychrerythraea GR in GR mutant) OD values

Full size image

Data were further analyzed in detail to determine if there was a significant difference between strains at a specific diamide concentration or between diamide concentrations for a specific strain (Online Resource 5). The gor-deficient strain (MLJ701) did not show a significant inhibition in its growth compared to the parental strain (MLJ70) when they were grown at moderate temperatures (25 and 37 °C) under non-oxidative stress conditions (p value > 0.05). However, when grown at low temperature (18 °C), MLJ701 growth was ~30 % lower than MLJ70 (p value < 0.05). In general, for all three temperature conditions, the GR mutant E. coli strain expressing C. psychrerythraea GR (MLJ702) showed some degree of restored growth in the oxidative stress cultures (diamide-treated cultures), especially at 25 and 18 °C as indicated by the significant p values <0.05. Interestingly, at 18 °C, the MLJ702 strain exhibited a high degree of protection from oxidative stress (no significant difference in growth) when treated with 0.2 mM diamide (p value = 0.7131) and growth was only reduced by ~15 % when treated with 0.3 mM diamide (p value < 0.05) compared to the non-diamide-treated condition of the same strain. At the same temperature, the GR mutant strain’s (MLJ701) growth was inhibited (~70 %) when treated with 0.2 or 0.3 mM diamide compared to the non-diamide-treated condition (p value < 0.05). Furthermore, when C. psychrerythraea GR is expressed in the E. coli gor mutant (MLJ702) incubated at 18 °C with diamide treatment, higher growth is observed compared to wild type (MLJ70) (p value < 0.05).

To further analyze the contribution of C. psychrerythraea GR activity to growth restoration, GR enzyme assays were conducted on cell-free extracts for each condition previously mentioned (Table 4). Regardless of the temperature or diamide concentration, the overall GR activities of GR-deficient strain MLJ701 were very low, while the GR wild-type strain MLJ70 showed moderate activity. The C. psychrerythraea GR expressing strain MLJ702 showed distinguishably higher activity than MLJ70 for 18 and 25 °C incubation conditions, while activity was reduced by ~20-fold for the 37 °C incubation condition.

Table 4 Glutathione reductase activities of E. coli strains for complementation growth studies
Full size table

Similar to the effects of diamide, certain types of ROS are also known to oxidize thiol groups; therefore, E. coli heterologously expressing C. psychrerythraea GR at room temperature was also treated with hydrogen peroxide and menadione to investigate if the enzyme would provide some protection. However, unlike the diamide treatment, it was determined that no growth restoration was observed for the gor mutant heterologously expressing C. psychrerythraea GR (data not shown) when the cells were treated with either hydrogen peroxide or menadione.

Discussion

There are several studies that report a high amount of dissolved oxygen within sea ice and a potential for production of ROS caused by UV radiation (Thomas and Dieckmann 2002; Cooper and Zika 1983; McMinn et al. 2005). Damage to cellular components such as proteins, lipids, DNA, and RNA by ROS to the point of cell death is also well documented (Yu 1994; Halliwell and Gutteridge 1984). However, few studies have investigated the antioxidant enzymes utilized by bacteria living in extreme conditions found in brine channels of sea ice. Examination of the properties and activity of glutathione reductase from C. psychrerythraea conducted in this study have provided new understanding of oxidative stress defense mechanisms used by psychrophilic bacteria.

The recombinant C. psychrerythraea GR exhibits common characteristics of other GRs, such as the use of NADPH (but not NADH) as a substrate, a yellowish color, and peak absorbance at 376 and 462 nm, which are consistent with the characteristics of flavoproteins (Colman and Black 1965; Garcia-Alfonso et al. 1993; Rendón et al. 1986; Lopez-Barea and Lee 1979).

As anticipated, it also has psychrophilic properties. It has more than ten times the catalytic efficiency (k cat /K m) of baker’s yeast GR at 4 °C. Despite catalytic inhibition due to low temperatures, high catalytic efficiency and metabolic flux are achieved by a high k cat and low K m (Feller 2013). In addition to having high activity at 4 °C, it also has significant stability at 37 °C, retaining activity for 3 weeks. This finding suggests that recombinant C. psychrerythraea GR has relatively high thermostability compared to the other reported psychrophilic enzymes (Cavicchioli et al. 2002; Yamanaka et al. 2002; Oikawa et al. 2005). A more specific comparison that demonstrates its high thermostability is that GR from an Antarctic ice microalga, Chlamydomonas sp., retains only about 5 % activity after incubation at 4 °C for the same period of time (Ding et al. 2007).

To verify the activity of recombinant C. psychrerythraea GR in vivo, it was heterologously expressed in a gor-deficient E. coli strain. The results from this study showed that C. psychrerythraea GR was able to restore the growth of an E. coli gor mutant to some extent when it was cultured with diamide. Diamide is a mild oxidizing compound known to quickly react with and oxidize thiols (Kosower and Kosower 1995). GR protects thiol-containing compounds from oxidation by maintaining a high cellular GSH/GSSG (reduced/oxidized glutathione) ratio (Grant et al. 1996; Carmel-Harel and Storz 2000; Yamamoto et al. 1999). Although the activity of glutathione reductase appears to be essential for maintaining a high glutathione content (Kunert et al. 1990), it was previously shown that E. coli GR mutants did not have a significantly different GSH/GSSG ratio suggesting that oxidized glutathione can be reduced independently of GR in E. coli (Tuggle and Fuchs 1985; Kunert et al. 1990). Consistent with these previous findings, our results show that when the gor mutant and parental strain are incubated at 25 or 37 °C without diamide, there is no significant difference in growth (p value > 0.05). However, there is a detectable difference in growth (~30 %) between these strains when cultured at 18 °C (p value < 0.05). This suggests that the other native thiol protective systems are not as effective at 18 °C. Furthermore, when the gor mutant (MLJ 701) and parental strain (MJL70) were tested with 0.2 and 0.3 mM diamide, growth of the gor mutant was substantially less than the parental strain at 18, 25, and 37 °C (p values < 0.05), which suggests that GR is responsible for protecting thiols that would be oxidized by diamide. However, when C. psychrerythraea glutathione reductase was heterologously expressed in the E. coli gor mutant (MLJ702) with diamide treatments at 18 °C, only a minor reduction in growth (~0–15 %) was observed compared to non-diamide treated suggesting that oxidized glutathione was reduced by cold active C. psychrerythraea GR, while the other native thiol protective systems in E. coli were not as effective at 18 °C. Overall, it was demonstrated that there was some degree of growth restoration of the diamide-treated cultures of MLJ702 compared to MLJ701 at each temperature, even though the growth difference at 37 °C is not considered significant (p value > 0.05). The effect of C. psychrerythraea GR expression was most pronounced at 18 °C with diamide treatments (p value < 0.05) (Fig. 5). Complementation of MLJ702 restored and even improved the growth above MLJ70 (parental strain, wild type) levels at 18 °C during the diamide-treated incubation. These observations were also supported by two-way ANOVA analysis (Online Resource 4); the p values of strain*diamide at low temperature incubations (25 and 18 °C) indicated that there was a significant interaction between strains and diamide treatments. This is further evidence that C. psychrerythraea GR is an active thiol-protective enzyme when the cells are under cold stress.

Besides demonstrating the ability of C. psychrerythraea GR to restore the growth of the E. coli gor mutant, the effectiveness of the enzyme in vivo was also supported by enzyme activity assays of cell-free extract. The data presented in Table 4 shows that at 25 °C GR activity of the complemented mutant was more than 15-fold higher than at 37 °C (whether the culture was treated with diamide or not), and at 18 °C GR activity was more than 20-fold higher than at 37 °C. These values when compared to Figs. 4 and 5 show a positive trend between GR activity and growth restoration.

Despite the stability of the purified C. psychrerythraea GR, the activity of the cell-free lysate of E. coli grown at 37 °C was about 15- to 20-fold lower than for cell lysate from cultures grown at 25 and 18 °C for the gor mutant stain with heterologously expressed C. psychrerythraea GR (MLJ702). Likely, the C. psychrerythraea GR does not fold correctly at 37 °C during expression, but once it is properly folded at lower temperatures it can stably function at 37 °C for weeks. Furthermore, it has been shown that at 37 °C compared to 18 °C, there is less C. psychrerythraea GR activity detected (Table 4), suggesting that more inactive protein is produced and turned over than active protein (Feller et al. 1998). Despite lower activity at 37 °C compared to 25 and 18 °C, the growth of the mutant strain was still restored by gor complementation.

Consistent with the overall results of the C. psychrerythraea GR complementation study, it was previously shown that in Streptococcus mutants, GR provided protection from diamide, but not hydrogen peroxide or menadione (Yamamoto et al. 1999; Becker-Hapak and Eisenstark 1995). As with S. mutants, it is presumed that H2O2 and menadione can increase intracellular peroxide (O2 2−) or superoxide (O −2 ) levels, respectively (Pocsi et al. 2005), but they might further propagate other ROS instead of disrupting the GSH/GSSG level, or ROS may be tightly regulated by other antioxidant enzymes before compromising the thiol redox status (Lushchak 2001).

Since C. psychrerythraea GR is a cold-adapted protein and also demonstrated activity in a heterologous E. coli system at mesophilic temperatures, there is potential for use of this antioxidant enzyme in biotechnological applications. For instance, when expressed in plants, C. psychrerythraea GR may have utility in improving the protection of protein structure and prevention of cellular damage from certain environmental stresses such as exposure to cold temperatures, radiation, and dehydration.