Bifunctional Chloroplastic DJ-1B from Arabidopsis thaliana is an Oxidation-Robust Holdase and a Glyoxalase Sensitive to H2O2

Members of the DJ-1 protein family are multifunctional enzymes whose loss increases the susceptibility of the cell to oxidative stress. However, little is known about the function of the plant DJ-1 homologs. Therefore, we analyzed the effect of oxidation on the structure and function of chloroplastic AtDJ-1B and studied the phenotype of T-DNA lines lacking the protein. In vitro oxidation of AtDJ-1B with H2O2 lowers its glyoxalase activity, but has no effect on its holdase chaperone function. Remarkably, upon oxidation, the thermostability of AtDJ-1B increases with no significant alteration of the overall secondary structure. Moreover, we found that AtDJ-1B transcript levels are invariable, and loss of AtDJ-1B does not affect plant viability, growth and stress response. All in all, two discrete functions of AtDJ-1B respond differently to H2O2, and AtDJ-1B is not essential for plant development under stress.


Introduction
α-dicarbonyls, such as glyoxal (GO) and methylglyoxal (MG), are toxic compounds produced during glycolysis, metal-catalyzed glucose auto-oxidation, and lipid peroxidation. When they react with proteins, they form advanced glycation end-products (AGEs), which have been implicated in the progression of diseases such as diabetes, atherosclerosis, and neurological disorders [1][2][3]. Plants also accumulate sugar-derived MG as a byproduct of the Calvin cycle [4]. GO and MG are being removed by a glutathione (GSH)-dependent two-enzyme system, consisting of glyoxalase I (GLYI) and glyoxalase II (GLYII) [5]. A third glyoxalase, initially termed GLYIII [6], is a GSH-independent enzyme. This GLYIII protein is a member of the DJ-1/ThiJ/PfpI superfamily, and it is conserved across all kingdoms. In humans and plants, this enzyme is termed DJ-1, and it is known to exhibit a variety of cellular functions. Human DJ-1 protects rat neurons against oxidative stress [7], has chaperone [8],

Thermal Unfolding
The thermal stability of AtDJ-1B was assessed by monitoring the change in the absorbance of circularly polarized light at 222 nm as a function of temperature. Immediately following spectra acquisition by CD as described above, the same samples were subjected to a temperature gradient from 15-85 • C at a ramp of 1 • C min −1 , and the absorption at 222 nm measured at 0.2 • C intervals. The linear slopes of the initial and final baselines of the sigmoidal unfolding curves were fitted with equation 1. For equation 1, A is ellipticity in mdeg (θ), m and k are the pre-unfolding intercept and slope respectively, T is temperature, n and l are the post-unfolding intercept and slope respectively.

Glyoxalase Assay
The glyoxalase activity of AtDJ-1B was estimated by monitoring the degradation of its substrate, glyoxal, in function of time. The amount of glyoxal in the sample was measured by derivatizing it with 1,2-diaminobenzene, as described for methylglyoxal [25].
A reaction mix containing 7.62 mM glyoxal and 571 nM AtDJ-1B in Assay Buffer A was incubated for 2 h at 30 • C on a thermoblock with mixing. Every 20 min an aliquot of the reaction mix was taken, 2-fold diluted in Assay Buffer A, and 1,2-diaminobenzene (Sigma-Aldrich, Overijse, Belgium) and HClO 4 were added to give a final concentration of 186 nM AtDJ-1B, 575 µM 1,2-diaminobenzene (Sigma-Aldrich, Overijse, Belgium) and 0.5 M HClO 4 . After 5 min mixing, the absorbance of the derivatized sample at 340 nm was measured. For each time point measurements from control standard solutions of 0, 2, 4, 6, 8, 10 mM of glyoxal derivatized in the same manner as the samples containing AtDJ-1B were taken for generation of a standard curve. The assays were performed for three experimental replicates with averaging of two technical replicates.

Chaperone Assay
AtDJ-1B protein sample was buffer exchanged to Assay Buffer B and either reduced with 5 mM TCEP or oxidized by treatment with either 2-fold or 10-fold molar excess H 2 O 2 to AtDJ-1B of 2:1 and 10:1 for 1 h at 25 • C. The samples were again buffer-exchanged into Assay Buffer B.
The thermal unfolding of citrate synthase (CS; from porcine heart; Sigma-Aldrich, Overijse, Belgium) was induced at 44 • C in Assay Buffer B with 0.24 µM citrate synthase (20 µg mL −1 ) and either 3-fold molar excess of Hsp90 (Sigma-Aldrich, Overijse, Belgium) as positive control, 5-fold molar excess of lysozyme (from chicken egg white, (Sigma-Aldrich, Overijse, Belgium)) as negative control, or 5-fold or 20-fold molar excess of AtDJ-1B protein. Aliquots (17 µL) were taken every 20 min and added to a cuvette containing the CS Assay Buffer to a final volume of 500 µL. Citrate synthase (CS) activity was followed by monitoring the rate of decrease of acetyl-CoA at 233 nm.

Oxidation of AtDJ-1B for Mass Spectrometric Analysis
AtDJ-1B was reduced with 50 mM TCEP for 30 min at room temperature and excess of TCEP was removed by Bio-Spin ® (Bio-Rad Laboratories N.V., Temse, Belgium) to a buffer containing 50 mM HEPES pH 7.3, 0.5 M NaCl. 73 µg of protein (25 µM) was mixed with 125-fold molar excess of dimedone prior to the addition of 10-fold molar excess of H 2 O 2 , and incubated at 37 • C for 1 h. H 2 O 2 and dimedone were removed by Bio-Spin™ and the protein concentration determined. Iodoacetamide (IAM) was added at 400-fold molar excess (i.e., final concentration 8 mM) and the sample incubated for 1 h at room temperature in the dark.

Determination of Cysteine Oxidation States by Mass Spectrometry
For the identification of modified residues in AtDJ-1B, 10 µg of desalted proteins were denatured by methanol/chloroform precipitation and digested O/N with trypsin or chymotrypsin at 30 • C in 50 mM NH 4 HCO 3 (pH 8.0).
After dissolving in eluent C (0.1% (v/v) trifluoroacetic acid in 2% (v/v) acetonitrile (ACN)), peptides were directly loaded onto reversed-phase pre-column (Acclaim PepMap 100, Thermo Fisher Scientific, Waltham, MA, USA) and eluted in backflush mode. Peptides were separated using a reversed-phase analytical column (Acclaim PepMap RSLC, 0.075 × 250 mm, Thermo Scientific), equilibrated in eluent A (0.1% (v/v) hydrofluoric acid (FA) with 4% eluent B (0.1% (v/v) FA in 80% (v/v) ACN, with a linear gradient of 4%-27.5% eluent B (0.1% hydrofluoric acid in 98% acetonitrile) for 100 min, 27.5%-40% eluent B for 10 min, 40%-95% eluent B for 1 min and holding at 95% for the last 10 min at a constant flow rate of 300 nl/min on an EASY-nLC 1000 ultra performance liquid chromatography (UPLC) system (Thermo Fisher Scientific, Waltham, MA, USA). Orbitrap Fusion Lumos tribrid mass spectrometer (Thermo Fisher Scientific, Waltham, MA, USA) was used for the analysis of the resulting peptides, which were then subjected to NanoSpray Ionization (NSI) source followed by tandem mass spectrometry (MS/MS) in Fusion Lumos coupled online to the UPLC. Orbitrap at a resolution of 120,000 was used for intact peptide detection. Peptides were selected for MS/MS using HCD (higher energy collisional dissociation) setting as 30; Orbitrap at a resolution of 30,000 was used for ion fragment detection. The the top 20 precursor ions above a threshold ion count of 5.0 3 in the MS survey scan with 20.0 s dynamic exclusion were subjected to a data-dependent procedure that alternated between one MS scan followed by 20 MS/MS scans. The electrospray voltage applied was 2.1 kV. MS1 and MS2 spectra were obtained with an AGC target of 4E5 ions and a maximum injection time of 50ms and an AGC target of 5E4 ions and a maximum injection time of 100 ms, respectively. The m/z scan range for MS scans was 350 to 1500. MS/MS data processing was performed using Sequest HT search engine within Proteome Discoverer 2.2 against a homemade protein database containing the recombinant AtDJ-1A and AtDJ-1B sequences. Trypsin or chymotrypsin was specified as cleavage enzyme allowing up to 2 missed cleavages, 5 modifications per peptide, and up to 7 charges. Mass error was set to 10 ppm for precursor ions and 0.2 Da for fragment ions. Oxidation on Met, sulfenic-dimedone, sulfinic or sulfonic on Cys were considered as variable modifications. False discovery rate (FDR) was assessed using a fixed value PSM validator and thresholds for protein, peptide and modification site were specified at 1%. Site of covalent modification were manually validated. The mixed disulfides were identified by the use of the DBond software (Hanyang University, South Korea) [26].

Gene Expression Levels
The transcript level changes upon various treatments were obtained from several RNA Sequencing (RNA-Seq) experimental datasets, including: 3-h high light stress on cat2-2 plants [27], treatment of Col-0 plants with 50 µM Antimycin A [28], Restricted Gas and Continuous Light (RGCL) treatment [29]  Data visualization was performed using Heatmapper [31], genes were hierarchically clustered using Euclidean distance with average linkage.

Plant Material
All mutants used in this study are SALK T-DNA insertion lines in Col-0 background: dj1a (SALK_049637), dj1b-4 (SALK_046449), dj1b-9 (SALK_093414). All lines were genotyped prior to further analysis and homozygous plants were selected for phenotyping and confirmation of gene expression levels by RT-qPCR.

RT-qPCR
To confirm that T-DNA lines are true knockout lines, their RNA was extracted from three-week-old plants rosettes using TRIzol solubilization and extraction [32] followed by RNeasy Plant Mini Kit (Qiagen, Venlo, The Netherlands). First strand cDNA synthesis was performed using iScript cDNA Synthesis Kit (Bio-Rad Laboratories N.V., Temse, Belgium) using 1 µg of total RNA used as input material. The 5-fold diluted cDNA and gene-specific primers (Supplementary Table) were used for RT-qPCR performed by iCycler iQ (Bio-Rad Laboratories N.V., Temse, Belgium), with SYBR Green I Master Kit according to manufacturer's instructions. Data was analyzed by qBASE+ (Biogazelle, Zwijnaarde, Belgium), using ELONGATION FACTOR 1α (EF-1α) and POLYUBIQUITIN 5 (UBIQ5) as reference genes. For each data point three biological and three technical replicates were used.

Growth Conditions and Plant Stress Assays
To assess the effect of high light, plants were grown in soil for 21 days in a controlled chamber (100 µmol·m −2 ·s −1 light intensity,16 h/8 h light/dark, 21 • C, 50% relative humidity). Three-week-old plants were transferred o high light (600 µmol·m −2 ·s −1 ) for 72 h. The bright-light pictures of the plants, as well as measurements of photosystem II (PSII) maximum efficiency (Fv'/Fm') using an Imaging-PAM-Series chlorophyll fluorescence system (HeinzWalz, Effeltrich, Germany) were taken every 24 h.
The Restricted Gas exchange and Continuous Light (RGCL) treatment was used to trigger photorespiratory stress in plants growing on 1 /2 MS agar plates [34]. 21 days after vernalization, plates (either 1 /2 MS or 1 /2 MS + 3-AT) were sealed with multiple layers of Parafilm ® M (Bemis Company Inc., Oshkosh, WI, USA) in order to restrict gas exchange and transferred to continuous light (100 µmol·m −2 ·s −1 light intensity, 21 • C, 70% relative humidity) for 10 days. The bright-light pictures of the plants, as well as measurements of photosystem II (PSII) maximum efficiency (Fv'/Fm') using an Imaging-PAM-Series chlorophyll fluorescence system (HeinzWalz, Effeltrich, Germany) were taken every 2-3 days.

AtDJ-1B Contains Multiple Oxidant-Sensitive Cysteines
The conserved catalytic cysteine of human DJ-1 is recognized as being highly prone to oxidation, with the most commonly observed oxidation state of this cysteine being sulfinic acid [35]. Arabidopsis AtDJ-1B is sulfenylated in planta [20], but nothing was known about its sulfenylation sites. AtDJ-1B is a pseudodimer containing two conserved DJ-1/PfpI domains in which there are eight cysteines, six within the N-terminal subunit, and two in the C-terminal subunit ( Figure 1A). Of these eight cysteines, two correspond to the conserved catalytic cysteines of homodimeric human DJ-1 (Cys109 and Cys314), and seven are predicted to be solvent-exposed based on a homology model of AtDJ-1B (Cys129 being the sole buried residue). Hence, we sought to characterize the extent of cysteine oxidation in recombinantly expressed AtDJ-1B after exposure to H 2 O 2 . Mass spectrometry (MS) analysis of AtDJ-1B treated for 1 h with a 10-molar excess of H 2 O 2 showed, as expected, sulfinylation of the conserved active-site cysteines (Cys109 and Cys314) ( Figure 1B). Sulfinylation also occurred at Cys110, and to a lesser extent at Cys129 and Cys339 ( Figure 1B).

AtDJ-1B Contains Multiple Oxidant-Sensitive Cysteines
The conserved catalytic cysteine of human DJ-1 is recognized as being highly prone to oxidation, with the most commonly observed oxidation state of this cysteine being sulfinic acid [35]. Arabidopsis AtDJ-1B is sulfenylated in planta [20], but nothing was known about its sulfenylation sites. AtDJ-1B is a pseudodimer containing two conserved DJ-1/PfpI domains in which there are eight cysteines, six within the N-terminal subunit, and two in the C-terminal subunit ( Figure 1A). Of these eight cysteines, two correspond to the conserved catalytic cysteines of homodimeric human DJ-1 (Cys109 and Cys314), and seven are predicted to be solvent-exposed based on a homology model of AtDJ-1B (Cys129 being the sole buried residue). Hence, we sought to characterize the extent of cysteine oxidation in recombinantly expressed AtDJ-1B after exposure to H2O2. Mass spectrometry (MS) analysis of AtDJ-1B treated for 1 h with a 10-molar excess of H2O2 showed, as expected, sulfinylation of the conserved active-site cysteines (Cys109 and Cys314) ( Figure 1B). Sulfinylation also occurred at Cys110, and to a lesser extent at Cys129 and Cys339 ( Figure 1B).

Oxidized AtDJ-1B Becomes More Thermostable
To assess the effect of the redox state of AtDJ-1B on its structural stability, we used circular dichroism (CD). We compared the relative secondary structure content of reduced and oxidized AtDJ-1B using increasing molar ratios of H 2 O 2 ( Figure 2A). The secondary structure of AtDJ-1B was not significantly altered upon oxidation, and both reduced and oxidized AtDJ-1B are characterized by 36-37% α-helical and 15-17% β-strand content, in agreement with previous findings, and comparable to what was observed for human DJ-1 [36]. generated using the I-TASSER structural prediction server, wherein the crystal structure of AtDJ-1D (PDB ID, 3UK7) was applied as a threading template. (B) Number of peptides (#PSM) detected by mass spectrometry that contained sulfenylated (-SOH) and sulfinylated (-SO2H) cysteines, either in a sample reduced by 5 mM TCEP or oxidized by 10-fold molar excess of H2O2 for 1 h.

Oxidized AtDJ-1B Becomes More Thermostable
To assess the effect of the redox state of AtDJ-1B on its structural stability, we used circular dichroism (CD). We compared the relative secondary structure content of reduced and oxidized AtDJ-1B using increasing molar ratios of H2O2 (Figure 2A). The secondary structure of AtDJ-1B was not significantly altered upon oxidation, and both reduced and oxidized AtDJ-1B are characterized by 36-37% α-helical and 15-17% β-strand content, in agreement with previous findings, and comparable to what was observed for human DJ-1 [36].
To evaluate the effect of oxidation on the thermostability of AtDJ-1B, we determined the melting temperature (TM) of AtDJ-1B by following the change in ellipticity at 222 nm in function of temperature ( Figure 2B). After fitting the data with a Boltzmann Sigmoidal equation, we obtained a TM of 59.7 °C for reduced AtDJ-1B. For oxidized AtDJ-1B, the TM increased to 67.6 °C and 69.3 °C following treatment with 2-fold and 10-fold molar excess of H2O2, respectively. This demonstrates that the oxidized form of AtDJ-1B is significantly more thermostable than the reduced form. To evaluate the effect of oxidation on the thermostability of AtDJ-1B, we determined the melting temperature (T M ) of AtDJ-1B by following the change in ellipticity at 222 nm in function of temperature ( Figure 2B). After fitting the data with a Boltzmann Sigmoidal equation, we obtained a T M of 59.7 • C for reduced AtDJ-1B. For oxidized AtDJ-1B, the T M increased to 67.6 • C and 69.3 • C following treatment with 2-fold and 10-fold molar excess of H 2 O 2 , respectively. This demonstrates that the oxidized form of AtDJ-1B is significantly more thermostable than the reduced form.

Oxidation Inactivates the Glyoxalase Activity of AtDJ-1B
One of the demonstrated functions of DJ-1 proteins is as a GSH-independent glyoxalase enzyme [9,37,38]. To measure the glyoxalase activity of AtDJ-1B, we monitored the consumption of glyoxal in function of time through derivatization with 1,2-diaminobenzene ( Figure 3). Then, under near-steady-state conditions (8 mM glyoxal) the effect of increasing molar ratios of H 2 O 2 on the glyoxalase activity of AtDJ-1B was evaluated. Pre-treatment of AtDJ-1B with a 2-fold molar excess H 2 O 2 resulted in a decrease of the relative activity to~64% of the activity of the reduced enzyme, and a further decrease to~16% after treatment with 10-fold molar excess H 2 O 2 . This dose-dependent oxidative inactivation of AtDJ-1B was further confirmed by complete inactivation of glyoxalase activity following treatment with a 100-fold molar excess of H 2 O 2 . Oxidation of AtDJ-1B by treatment with 5 mM diamide (disulfide bond formation) also elicited a decrease to~17% of full activity.
thermostability. (A) Comparative CD spectra of AtDJ-1B reduced with TCEP (blue trace), and DJ-1B oxidized by treatment with either a 2-fold (orange) or 10-fold (red) molar excess of H2O2. A slight spectral shift can be observed for the H2O2-treated protein relative to the reduced protein, though this spectral difference does not ultimately relate to any significant change in secondary structure. (B) Thermal unfolding of AtDJ-1B fitted to the Gibbs-Helmholtz equation. The change in ellipticity at 222 nm was followed by CD as a function of temperature for reduced AtDJ-1B (5 mM TCEP), and oxidized AtDJ-1B (either 2-fold, or 10-fold molar excess of H2O2). Data were converted to fraction of unfolded protein ([U]/[F]) and baseline-subtracted as described in Materials & Methods.

Oxidation Inactivates the Glyoxalase Activity of AtDJ-1B
One of the demonstrated functions of DJ-1 proteins is as a GSH-independent glyoxalase enzyme [9,37,38]. To measure the glyoxalase activity of AtDJ-1B, we monitored the consumption of glyoxal in function of time through derivatization with 1,2-diaminobenzene ( Figure 3). Then, under nearsteady-state conditions (8 mM glyoxal) the effect of increasing molar ratios of H2O2 on the glyoxalase activity of AtDJ-1B was evaluated. Pre-treatment of AtDJ-1B with a 2-fold molar excess H2O2 resulted in a decrease of the relative activity to ~64% of the activity of the reduced enzyme, and a further decrease to ~16% after treatment with 10-fold molar excess H2O2. This dose-dependent oxidative inactivation of AtDJ-1B was further confirmed by complete inactivation of glyoxalase activity following treatment with a 100-fold molar excess of H2O2. Oxidation of AtDJ-1B by treatment with 5 mM diamide (disulfide bond formation) also elicited a decrease to ~17% of full activity. Figure 3. Glyoxalase activity of AtDJ-1B is lost after oxidation. Reduced and oxidized AtDJ-1B were prepared by pretreatment with either TCEP, varying molar ratios of H2O2: protein, or diamide. The reductant/oxidant was then removed prior to the activity assay. Displayed are observed rate constant (kobs) values averaged from experimental triplicates with ± standard deviation (SD) indicated.

Oxidation does not Affect the Chaperone Activity of AtDJ-1B
Hsp31, a close homolog of AtDJ-1B in Saccharomyces cerevisiae, has previously been demonstrated to function as a holdase and confer stress resistance [39][40][41]. To explore whether AtDJ-1B is capable of similar functionality, we assessed the ability of AtDJ-1B to act as a holdase towards citrate synthase Figure 3. Glyoxalase activity of AtDJ-1B is lost after oxidation. Reduced and oxidized AtDJ-1B were prepared by pretreatment with either TCEP, varying molar ratios of H 2 O 2 : protein, or diamide. The reductant/oxidant was then removed prior to the activity assay. Displayed are observed rate constant (k obs ) values averaged from experimental triplicates with ± standard deviation (SD) indicated.

Oxidation does not Affect the Chaperone Activity of AtDJ-1B
Hsp31, a close homolog of AtDJ-1B in Saccharomyces cerevisiae, has previously been demonstrated to function as a holdase and confer stress resistance [39][40][41]. To explore whether AtDJ-1B is capable of similar functionality, we assessed the ability of AtDJ-1B to act as a holdase towards citrate synthase (CS). Here we found that AtDJ-1B protects CS against thermal inactivation at 44 • C, with 5-fold molar excess of AtDJ-1B preserving 35% of CS activity after 40 min heat treatment compared to only 10% preservation when using a 5-fold molar excess lysozyme as a negative control ( Figure 4A). Although AtDJ-1B offered less protection than the positive control of Hsp90 (preserving 68% of activity at a 3-fold molar excess), the extent to which AtDJ-1B protected CS against inactivation was proportional to the relative concentration of AtDJ-1B, with 20-fold molar excess resulting in greater protection than 5-fold excess ( Figure 4A). Considering a putative role of AtDJ-1B in the oxidative stress response, we evaluated the impact of oxidation on the ability of AtDJ-1B to act as a holdase. Remarkably, no significant difference was observed between the chaperone effectiveness of reduced (TCEP-treated) and oxidized (H 2 O 2 -treated) AtDJ-1B, with all samples showing a similar level of protection against thermal inactivation of CS ( Figure 4B).
Although AtDJ-1B offered less protection than the positive control of Hsp90 (preserving 68% of activity at a 3-fold molar excess), the extent to which AtDJ-1B protected CS against inactivation was proportional to the relative concentration of AtDJ-1B, with 20-fold molar excess resulting in greater protection than 5-fold excess ( Figure 4A). Considering a putative role of AtDJ-1B in the oxidative stress response, we evaluated the impact of oxidation on the ability of AtDJ-1B to act as a holdase. Remarkably, no significant difference was observed between the chaperone effectiveness of reduced (TCEP-treated) and oxidized (H2O2-treated) AtDJ-1B, with all samples showing a similar level of protection against thermal inactivation of CS ( Figure 4B).

DJ-1B-Deficient Plants are Phenotypically Identical to Wildtype
Having established the possible activities of DJ-1B and their redox regulation, we sought to determine the phenotype of Arabidopsis T-DNA dj-1b insertion lines. Two SALK T-DNA dj1b lines (further referred to as dj1b-4 and dj1b-9, see Materials and Methods) and one SALK T-DNA dj-1a line were genotyped and confirmed to be homozygous. The dj1b-4 and dj1b-9 mutants contain the T-DNA insert in the 4th exon and the 4th intron, respectively, while in the dj1a line the insert is located in the promoter region (Supplementary Figure S1). The lack of gene transcripts was validated by reverse transcription quantitative PCR (RT-qPCR) (see Supplementary data).
When grown in soil in a growth chamber (16 h of light, 8 h of dark, 100 µmol·m −2 ·s −1 light intensity), the dj1b and dj1a mutants were phenotypically identical to wildtype Col-0 plants ( Figure 5). Since dj1a mutants were previously reported to show increased susceptibility to high light stress [16], we sought to confirm these findings and checked whether plants lacking AtDJ-1B show the same phenotype. After 3 days of high light stress treatment, both wildtype and mutant plants had no leaf lesions and their photosystem II efficiency, inferred from Fv'/Fm' ratios, remained unchanged ( Figure 6).
When grown in soil in a growth chamber (16 h of light, 8 h of dark, 100 μmol·m −2 ·s −1 light intensity), the dj1b and dj1a mutants were phenotypically identical to wildtype Col-0 plants ( Figure  5). Since dj1a mutants were previously reported to show increased susceptibility to high light stress [16], we sought to confirm these findings and checked whether plants lacking AtDJ-1B show the same phenotype. After 3 days of high light stress treatment, both wildtype and mutant plants had no leaf lesions and their photosystem II efficiency, inferred from Fv'/Fm' ratios, remained unchanged ( Figure  6)   Figure 6. Photosystem II efficiency of mutants subjected to high light treatment was the same as that of wild type (WT). Three-week-old plants were exposed to high light intensities (600 μmol·m −2 ·s −1 ) for 72 h and the Fv'/Fm' ratios were measured. Each data point represents average from biological triplicates ± SD.
To further characterize the phenotypes of dj1a and dj1b mutants, we grew them on ½ Murashige-Skoog (MS) agar medium for 3 weeks and assessed their rosette sizes when plants were subjected to different stress agents in the medium: NaCl, mannitol, methyl viologen, 3-amino-1,2,4-triazole. In all cases, mutants grown on control ½ MS medium or subjected to stress were characterized by the same rosette size as wildtype (Figure 7). Moreover, when plants were subjected to photorespiratory stress by a Restricted Gas exchange and Continuous Light (RGCL) treatment [34] for up to 10 days, the Fv'/Fm' ratio of the mutants was the same as for wildtype (Figure 8). Taken together, these results indicate that the loss of DJ-1A or DJ-1B does not result in altered growth rates and photosynthetic capacities when exposed to the described abiotic stress conditions. Figure 6. Photosystem II efficiency of mutants subjected to high light treatment was the same as that of wild type (WT). Three-week-old plants were exposed to high light intensities (600 µmol·m −2 ·s −1 ) for 72 h and the Fv'/Fm' ratios were measured. Each data point represents average from biological triplicates ± SD.
To further characterize the phenotypes of dj1a and dj1b mutants, we grew them on 1 /2 Murashige-Skoog (MS) agar medium for 3 weeks and assessed their rosette sizes when plants were subjected to different stress agents in the medium: NaCl, mannitol, methyl viologen, 3-amino-1,2, 4-triazole. In all cases, mutants grown on control 1 /2 MS medium or subjected to stress were characterized by the same rosette size as wildtype (Figure 7). Moreover, when plants were subjected to photorespiratory stress by a Restricted Gas exchange and Continuous Light (RGCL) treatment [34] for up to 10 days, the Fv'/Fm' ratio of the mutants was the same as for wildtype (Figure 8). Taken together, these results indicate that the loss of DJ-1A or DJ-1B does not result in altered growth rates and photosynthetic capacities when exposed to the described abiotic stress conditions.

Transcript Levels of DJ-1B are Stress-Independent
Since various oxidative stress treatments trigger transcriptome-wide changes in Arabidopsis mRNA levels [42], we analyzed several RNA-Seq datasets, aiming to determine whether mRNA levels of all Arabidopsis DJ-1 homologs are also susceptible to such treatments (Figure 9). We concentrated on treatments triggering oxidative stress, such as RGCL [34], high light [27], methyl viologen treatment, and Pseudomonas syringae infection. In addition to wildtype Col-0, we also analyzed the transcriptome of cat2-2 mutants, which are commonly used as stress-inducible systems to study oxidative stress in vivo [43].
The RNA-Seq dataset analysis confirmed the earlier finding that AtDJ-1A is upregulated under stress [16]. The same trend was observed for AtDJ-1E, whose mRNA levels were also upregulated in the majority of the stress conditions. Remarkably, for AtDJ-1F, which shares 76% protein sequence identity with AtDJ-1E, the mRNA levels are significantly decreased upon stress. The mRNA levels of the other three AtDJ-1 homologs are less susceptible to stress treatments; in particular AtDJ-1B transcription levels are not affected. This result is in line with previous studies of plant DJ-1 homologs showing different expression patterns and transcriptional responses to stress [44].

Transcript Levels of DJ-1B are Stress-Independent
Since various oxidative stress treatments trigger transcriptome-wide changes in Arabidopsis mRNA levels [42], we analyzed several RNA-Seq datasets, aiming to determine whether mRNA levels of all Arabidopsis DJ-1 homologs are also susceptible to such treatments ( Figure 9). We concentrated on treatments triggering oxidative stress, such as RGCL [34], high light [27], methyl viologen treatment, and Pseudomonas syringae infection. In addition to wildtype Col-0, we also analyzed the transcriptome of cat2-2 mutants, which are commonly used as stress-inducible systems to study oxidative stress in vivo [43].
The RNA-Seq dataset analysis confirmed the earlier finding that AtDJ-1A is upregulated under stress [16]. The same trend was observed for AtDJ-1E, whose mRNA levels were also upregulated in the majority of the stress conditions. Remarkably, for AtDJ-1F, which shares 76% protein sequence identity with AtDJ-1E, the mRNA levels are significantly decreased upon stress. The mRNA levels of the other three AtDJ-1 homologs are less susceptible to stress treatments; in particular AtDJ-1B transcription levels are not affected. This result is in line with previous studies of plant DJ-1 homologs showing different expression patterns and transcriptional responses to stress [44].  [29], or two lines (Col-0 and cat2-2) were compared to each other. The RNASeq experiments from which the data was obtained, as well as the detailed descriptions of the treatments are referred to in Materials and Methods. Colors represent log2 fold change.

Discussion
We report reduced AtDJ-1B as a glyoxalase with specific activity of 600 nmol·min −1 ·mg protein −1 (Supplementary Table). For non-reduced AtDJ-1B, a specific activity of 310 nmol·min −1 ·mg protein −1 was reported [12]. The glyoxalase activity of AtDJ-1B is ~35 times lower than for AtDJ-1D when glyoxal is used as substrate (Supplementary Table) [12]. This lower glyoxalase activity could be due to the absence of a conserved histidine, which has been suggested to facilitate proton transfer and

Discussion
We report reduced AtDJ-1B as a glyoxalase with specific activity of 600 nmol·min −1 ·mg protein −1 (Supplementary Table). For non-reduced AtDJ-1B, a specific activity of 310 nmol·min −1 ·mg protein −1 was reported [12]. The glyoxalase activity of AtDJ-1B is~35 times lower than for AtDJ-1D when glyoxal is used as substrate (Supplementary Table) [12]. This lower glyoxalase activity could be due to the absence of a conserved histidine, which has been suggested to facilitate proton transfer and glyoxal stereospecificity [45]. In plants under biotic and abiotic stress, and after H 2 O 2 treatment, the intracellular concentration of glyoxals has been shown to increase from 100 to 2000 µmol/g tissue [46,47]. When methylglyoxal is photoreduced by photosystem I during photosynthesis, it donates electrons to O 2 , producing superoxide (O 2 •− ) [48]. Thus, by removing superoxide-generating glyoxals, enhanced glyoxalase activity contributes to cellular protection during oxidative stress. However, we clearly showed inactivation of the glyoxalase activity of AtDJ-1B by H 2 O 2 , the predominant oxidant under oxidative stress conditions (Figure 4). To our knowledge, this is the first report of a DJ-1 protein losing its glyoxalase activity upon oxidation. Moreover, as the GLYI/II enzyme system catalyzes glyoxal/methylglyoxal detoxification with far greater catalytic efficiency than even AtDJ-1D [49], we propose that in plants the majority of glyoxals is detoxified by GLYI/II enzymes, rather than by H 2 O 2 -sensitive DJ-1 proteins.
AtDJ-1C has been reported with no glyoxalase activity [12], and its knockout results in non-viable plants [18]. Therefore, we sought to explore possible alternative functions for AtDJ-1B, which shares its chloroplastic localization with AtDJ-1C. We found that AtDJ-1B acts as an oxidation-robust holdase, though a molar excess of AtDJ-1B is needed to protect CS from thermal inactivation. Since porcine CS used for the holdase activity assay is not its natural client protein, it still remains unclear whether the holdase function of AtDJ-1B would be its real physiological function. Furthermore, it is known that dedicated chaperones suppress aggregation of client proteins at stoichiometric concentrations [50], whereas AtDJ-1B showed holdase activity at 5 to 20-fold molar excess. Similarly to human DJ-1 [8], AtDJ-1B could have specific client proteins in Arabidopsis that still have to be identified.
Furthermore, human DJ-1 has been demonstrated to be a redox-dependent holdase, with the oxidized form of DJ-1 able to suppress α-synuclein aggregation significantly better than the reduced form, offering 20 to 60% more protection [8]. On the other hand, the yeast homolog Hsp31 is a redox-independent holdase and this chaperone activity is required for protecting yeast from cytotoxic stress [39]. We also found the holdase activity of AtDJ-1B to be redox-independent. Remarkably, we observed that the thermostability of AtDJ-1B increases upon oxidation, which indicates that the oxidized form of AtDJ-1B is more likely than reduced AtDJ-1B to retain its chaperone activity under sustained heat stress. Other DJ-1 family members have also been shown to be highly thermostable, with a T M values ranging from 64 to 77 • C [36,51,52]. Here, we determined a comparatively lower T M for AtDJ-1B of 59.7 • C, and observe an increase in T M of approximately 8-10 • C for oxidized AtDJ-1B. Similar increase in thermostability has also been observed for human DJ-1 and drosophila DJ-1β, with respective T M increases of 13.3 and 11.5 • C upon oxidation in the presence of a 7-fold molar excess of H 2 O 2 [51].
Finally, it is worthy of mention that both human and prokaryotic DJ-1 homologs also have deglycase activity [53][54][55], and it was proposed to be one of the core physiological functional roles of DJ-1 proteins [56]. As glyoxalase function is required for deglycase activity, when glyoxalase activity is lost upon oxidation, the deglycase function will also be abolished. In both its role as a deglycase and as a holdase, it is reasonable to assume that AtDJ-1B has preferential target proteins. Further understanding of the in vivo role of AtDJ-1B will require the identification of its interaction partners in Arabidopsis.
In contrast to AtDJ-1C, whose loss results in infertile plants with numerous developmental defects [18], AtDJ-1B loss does not influence plant viability. We also found that dj1b plants show wild-type resistance to stress, at least under the conditions tested. Moreover, while it was previously reported that AtDJ-1A deletion leads to increased susceptibility to high light [16], the phenotype of the dj1a SALK T-DNA line described here is identical to wildtype, which directly contradicts earlier findings [16].
The Arabidopsis DJ-1 family originates from one ancestor gene and diverged into six genes through both whole-genome and tandem duplications [57]. The same mechanism is responsible for higher duplication rates in plants in general, when compared to other eukaryotes [58,59]. After duplication, proteins evolve new functions; one of the copies might acquire a novel function, or if the ancestral protein is multifunctional, the duplicates divide the original function, for instance by differential expression patterns [59]. The analysis of RNASeq data presented in this paper shows that Arabidopsis DJ-1 proteins differ greatly in their expression patterns. These differences will be better understood after the analysis of their cis-regulatory elements in promoter sequences, like described for other plant species [44,60,61]. Moreover, since the mRNA levels are not representative for the protein levels, especially during dynamic phases such as cellular differentiation and stress response [62], it would be very informative to quantify all DJ-1 protein levels, both in stress and control conditions. Generating multiple mutant lines and biochemically characterizing all six Arabidopsis DJ-1 proteins will explain to what degree their functions overlap and whether the findings about the role of AtDJ-1B presented here are applicable to the other homologs as well.

Conclusions
In summary, we showed for the first time that Arabidopsis DJ-1B is a bifunctional protein, having both glyoxalase and holdase activity. Importantly, these two functions are differently regulated by H 2 O 2 ; while the glyoxalase activity is lost upon oxidation, the holdase activity is not affected by H 2 O 2 . We also, for the first time, report the dj1b plant phenotype and prove that AtDJ-1B is not necessary for viability, development, or stress resistance of Arabidopsis plants, which might be due to redundant functions of all DJ-1 homologs. To explain why Arabidopsis has retained six DJ-1 homologs, the physiological relevance of the bifunctionality of AtDJ-1B still need to be confirmed in vivo and compared to those of the other AtDJ-1 proteins.

Conflicts of Interest:
The authors declare no conflict of interest.