L-PGDS [lipocalin-type PGD (prostaglandin D) synthase] is a dual-functional protein, acting as a PGD2-producing enzyme and a lipid transporter. L-PGDS is a member of the lipocalin superfamily and can bind a wide variety of lipophilic molecules. In the present study we demonstrate the protective effect of L-PGDS on H2O2-induced apoptosis in neuroblastoma cell line SH-SY5Y. L-PGDS expression was increased in H2O2-treated neuronal cells, and the L-PGDS level was highly associated with H2O2-induced apoptosis, indicating that L-PGDS protected the neuronal cells against H2O2-mediated cell death. A cell viability assay revealed that L-PGDS protected against H2O2-induced cell death in a concentration-dependent manner. Furthermore, the titration of free thiols in H2O2-treated L-PGDS revealed that H2O2 reacted with the thiol of Cys65 of L-PGDS. The MALDI–TOF (matrix-assisted laser-desorption ionization–time-of-flight)-MS spectrum of H2O2-treated L-PGDS showed a 32 Da increase in the mass relative to that of the untreated protein, showing that the thiol was oxidized to sulfinic acid. The binding affinities of oxidized L-PGDS for lipophilic molecules were comparable with those of untreated L-PGDS. Taken together, these results demonstrate that L-PGDS protected against neuronal cell death by scavenging reactive oxygen species without losing its ligand-binding function. The novel function of L-PGDS could be useful for the suppression of oxidative stress-mediated neurodegenerative diseases.
- prostaglandin D synthase
- reactive oxygen species (ROS)
Oxidative stress induced by the excessive production of ROS (reactive oxygen species) has been associated with ischaemia/reperfusion injury [1,2]. ROS are produced during and after cerebral ischaemia and cerebral infarction, have been shown to have a neurotoxic impact on the brain, and have also been implicated in neurological disorders including dysautonomia, Alzheimer's disease and Parkinson's disease [3,4]. A number of studies have suggested that the excessive generation of ROS underlies neuronal degeneration under diverse neurological conditions [5–8]. Among ROS, H2O2 is well known to be involved in neuronal cell death . Oxidative stress also modifies, reversibly or irreversibly, cellular components such as DNA, lipids and proteins. Thus, to understand the protective mechanism against oxidative stress in the central nervous system, it is necessary to elucidate the relationship between oxidative stress and the expression level of key proteins.
In the brain, L-PGDS [lipocalin-type PGD (prostaglandin D) synthase] is synthesized mainly at the rough endoplasmic reticulum membrane of arachnoid cells, chorioid plexus cells and oligodendrocytes, and then is secreted into the CSF (cerebrospinal fluid) as a second major protein in human CSF . L-PGDS is a dual-functional protein, acting as a PGD2-synthesizing enzyme and as an extracellular transporter for lipophilic ligands . The tertiary structure of L-PGDS exhibits a single eight-stranded antiparallel β-sheet closed back on itself to form a continuously hydrogen-bonded β-barrel and an α-helix [12,13]. Human L-PGDS possesses two free thiol groups, those of Cys65 and Cys167, and Cys65 is the active centre for the catalytic function of L-PGDS . Protein thiol has been indicated to react with various kinds of active oxygen, indicating that protein thiol modification by ROS is important in biological defence mechanisms [15,16]. These reports prompted us to investigate possible oxidative stress-mediated thiol modification of free cysteine residues in human L-PGDS.
The L-PGDS concentration increases in the CSF of patients several weeks after a stroke as well as in some tumour patients . The CSF level of L-PGDS is significantly lower in patients with schizophrenia , a brain tumour , bacterial meningitis  and normal pressure hydrocephalus . L-PGDS is up-regulated in acute and massive brain injury resulting from neonatal hypoxic ischaemic encephalopathy . The L-PGDS levels in the CSF of patients with aneurysmal subarachnoid haemorrhage transiently increase to be 2-fold over normal within 5 days . All of these reports suggest that the L-PGDS concentration is strongly related to various pathophysiological conditions. In the present study we demonstrate the protective effect of human L-PGDS against ROS-induced neuronal cell death by scavenging the active oxygen via the thiol of Cys65.
DTNB [5,5′-dithiobis-(2-nitrobenzoic acid)], trichloroacetic acid, and guanidinium chloride were purchased from Wako Pure Chemical. Thrombin and DAPI (4′,6-diamidino-2-phenylindole) were purchased from Sigma. Sinapic acid was obtained from Aldrich. All other chemicals were of analytical grade.
Expression of recombinant L-PGDS in Escherichia coli and purification
The open reading frame for human L-PGDS, which is composed of 190 amino acid residues (GenBank® accession number M61900) , or that for mouse L-PGDS, composed of 189 amino acid residues (GenBank® accession number X89222); , was ligated into the BamHI/EcoRI sites of the expression vector pGEX-2T (GE Healthcare Bio-Sciences) . The N-terminal 22 amino acid residues corresponding with the putative secretion signal peptide of both L-PGDSs were truncated. WT (wild-type) human L-PGDS possesses four cysteine residues (Cys65, Cys89, Cys167 and Cys186), and two of them, Cys89 and Cys186, form a disulfide bridge that is highly conserved among most lipocalins. C65A/C167A (ϵ280=25900 M−1·cm−1), C89A/C167A/C186A (ϵ280=25700 M−1·cm−1) and C89A/C186A (ϵ280=25500 M−1·cm−1) human L-PGDS, and C89A/C186A (ϵ280=22461 M−1·cm−1) and Y63S/T67S/C89A/C186A (ϵ280=20892 M−1·cm−1) mouse L-PGDS were expressed in E. coli BL21(DE3) (Toyobo). Site-directed mutagenesis was performed using a QuikChange™ site-directed mutagenesis kit (Stratagene). The DNA sequences were confirmed with a CEQ™ 8000 Genetic Analysis System (Beckman Coulter) after cycle sequencing with a SequiTherm Cycle Sequencing kit (Epicentre Technologies). Each mutated L-PGDS was expressed as a glutathione transferase fusion protein. The fusion protein was bound to glutathione–Sepharose 4B (GE Healthcare Bio-Sciences) and incubated with thrombin (1 unit/1 μl) to release the L-PGDS. These recombinant proteins were further purified to apparent homogeneity by HiLoad 16/60 Superdex 75 (GE Healthcare Bio-Sciences) chromatography and dialysed against PBS.
SH-SY5Y cells, which are human neuroblastoma cells, were cultured in Dulbecco's modified Eagle's medium/Ham's F12 nutrient mixture (Gibco) containing 10% FBS (fetal bovine serum; Gibco), 100 units/ml penicillin, 100 μg/ml streptomycin and 0.25 μg/ml amphotericin B (Gibco) in a humidified atmosphere of 5% CO2 and 95% air at 37°C.
SH-SY5Y cells were cultured in 96-well tissue culture plates at a density of 3×104 cells/well. After 48 h of cultivation, the appropriate amounts of H2O2 and/or L-PGDS were added to the culture medium. The cells were further cultured for 24 h, and then WST-8 solution (Dojindo Laboratories) was added to the culture medium; thereafter, the cells were incubated for 4 h at 37°C. The absorbance of the formazan produced from WST-8 was measured at 450 nm by using a microplate reader, Model 680 (Bio-Rad Laboratories).
LDH (lactate dehydrogenase) activity was measured by use of a Cytotoxicity Detection KitPLUS (Roche Diagnostics) according to the manufacturer's instructions. Absorbance was measured at 492 nm by using a Microplate reader Lucy2 (Anthos).
Detection of DNA fragmentation by agarose gel electrophoresis
SH-SY5Y cells were cultured for 24 h in the presence or absence of H2O2 (100 μM). Genomic DNA was isolated from the cells by using Isogen (Nippon Gene) following the manufacturer's instructions, and resuspended in TE buffer [10 mM Tris/HCl and 1 mM EDTA (pH 8.0)]. The purified DNA was loaded on to a 2% agarose gel in TAE [40 mM Tris/acetate, 1 mM EDTA (pH 8.0)] buffer and, following electrophoresis, it was stained with 200 μg/ml ethidium bromide to visualize the bands with a 254-nm UV transilluminator.
Cells were seeded on plates coated with 100 μg/ml poly-D-lysine, and then treated with H2O2 (100 μM) for 24 h. The cells were washed with PBS and fixed for 10 min with 4% (w/v) paraformaldehyde in phosphate buffer (pH 7.4). Then, cells were washed three times with PBS and stained with 1 μg/ml DAPI in PBS including 0.05% Tween 20. Fluorescence was detected using a fluorescence microscope (Nikon).
Preparation of RNA and quantification of mRNA levels
Total RNA was extracted by use of TriPure Isolation Reagent (Roche Diagnostics) according to the manufacturer's directions. First-strand cDNAs were synthesized by using ReverTra Ace Reverse Transcriptase (Toyobo) primed by random-hexamer as described previously . Quantification of mRNA levels was carried out by using a real-time PCR system (LightCycler, Roche Diagnostics) and Fast SYBR Green Master Mix (Roche Diagnostics) with gene-specific primer sets: forward, 5′-CCAACTTCCAGCAGGACAAG-3′ and reverse, 5′-CCACAGACTTGCACATGGAC-3′ for L-PGDS, and forward, 5′-CCAACCGCGAGAAGATGA-3′ and reverse, 5′-CCAGAGGCGTACAGGGATAG-3′ for β-actin. All data were obtained from three independent experiments and transcription levels of all genes were normalized to the level of the β-actin gene used as the internal control.
siRNA (small interfering RNA) study
The following Stealth™ siRNAs for L-PGDS and Stealth™ NC (negative control) siRNA were purchased from Invitrogen: L-PGDS siRNA#1, 5′-CAGGACUUCCGCAUGGCCACCCUCU-3′ and L-PGDS siRNA#2, 5′-GACUUCCGCAUGGCCACCCUCUACA-3′. At 1 day prior to the transfection, SH-SY5Y cells were plated at a density of 2×104 cells/well on a 96-well culture plate. The cells were transfected with either siRNA or the NC siRNA (20 nM) by using X-tremeGENE siRNA Transfection Reagent (Roche Diagnostics) according to the manufacturer's instructions. After 1 day of transfection, H2O2 (100 μM) was added and the cells were then cultured for a further 1 day. RNA was extracted and mRNA levels were then measured by quantitative PCR as described above.
Expression of L-PGDS in SH-SY5Y cells
The expression vector of human L-PGDS was constructed previously [25a]. Site-directed mutagenesis of Cys65 to alanine (C65A) was carried out as described above. SH-SY5Y cells were transfected with FuGENE6 Transfection Reagent (Roche Diagnostics) according to the protocols prescribed by the manufacturer. At 1 day after transfection, the transfectants were incubated in medium containing 100 μM H2O2 for 24 h. Cell viability and LDH activity were measured as described above.
Western blot analysis
Cells were lysed in RIPA buffer containing 50 mM Tris/HCl (pH 8.0), 150 mM NaCl, 0.1% SDS, 0.5% sodium deoxycholate, 1% (v/v) Nonidet P40 and 1% (v/v) Triton X-100 along with protease inhibitor cocktail (Nacalai Tesque), and then centrifuged at 12000 g for 20 min at 4°C. The proteins were separated on an SDS/PAGE (12.5% gel) and then transferred on to PVDF membranes (Immobilon P, Millipore) for Western blot analysis by the use of the SNAP i.d. Protein Detection System (Millipore). Blots were incubated with anti-L-PGDS (1:2000 dilution; 1B7 ), anti-FLAGM2 (1:5000; Sigma) and anti-actin monoclonal antibody (1:3000; AC-15; Sigma), and then incubated with HRP (horseradish peroxidase)-conjugated secondary antibody (1:1000; Santa Cruz Biotechnology). Immunoreactive signals were detected by the use of a Luminata Forte Western HRP substrate (Millipore). Protein concentrations were measured with a Pierce BCA (bicinchoninic acid) protein assay kit (Thermo Scientific).
Determination of free thiol groups
L-PGDS (5 μM) was incubated with 50 μM H2O2 for 0–24 h or H2O2 (0–100 μM) in PBS at 37°C for 24 h. The reaction was stopped by adding trichloroacetic acid to a final concentration of 10% (v/v). The sample tubes were placed on ice for 30 min and then centrifuged at 29000 g for 30 min at 4°C. The supernatant was removed, and the pellet was dissolved in 500 μl of 50 mM Tris/HCl buffer (pH 8.0) containing 6 M guanidinium chloride and incubated for 15 min at room temperature (25°C). Free thiol groups were determined by the reaction with DTNB, which was added at 5-fold excess. The number of free thiol groups was calculated from the absorbance of the released TNB (5-thiobenzoic acid) anion (ϵ412=13600 M−1·cm−1 ).
The samples (200 μl) were desalted by using a C4 Zip Tip (Millipore), and eluted in 70% acetonitrile. MALDI–TOF (matrix-assisted laser-desorption ionization–time-of-flight)-MS analysis was performed with an Ultraflex TOF/TOF (Bruker Daltonics) operating under FlexControl 2.0 (Bruker Daltonics) in the positive-ion linear mode. Acceleration voltages of 25 kV were applied. Ions ranging from 10 to 22 kDa were registered. Approximately 300 single spectra were summarized. Data analysis was performed with FlexAnalysis 2.0. The MALDI–TOF mass spectrometer was equipped with a standard nitrogen laser (337 nm), and data were acquired in the positive-ion mode with external calibration. Before MALDI–TOF-MS analysis, both C89A/C167A/C186A and C65A/C167A L-PGDS at 5 μM were either treated or not with 50 μM H2O2 at 37°C for 24 h. Then the proteins were mixed with saturated sinapinic acid in acetonitrile/0.1% trifluoroacetic acid [1:3 (v/v)] and externally calibrated. Prepared proteins were loaded on to a single spot of an AnchorChip (Bruker Daltonics). The function of AnchorChip was to condense each protein in a very small area so that a higher signal intensity could be obtained even with a low concentration of analyte on the target plate.
BV (biliverdin) and all-trans-RA (retinoic acid) were each dissolved in DMSO to give a 2 mM stock solution. The concentrations were determined spectroscopically based on their respective molar absorption coefficients of ϵ377 in methanol for BV (51500 M−1·cm−1)  and ϵ336 in ethanol for RA (45000 M−1·cm−1) . C89A/C167A/C186A L-PGDS at 5 μM was reacted with 50 μM H2O2 for 24 h at 37°C in PBS. The reaction sample was dialysed against PBS, and the oxidized L-PGDS was thus obtained. Various concentrations of BV or RA were added to untreated or oxidized L-PGDS in PBS for pH 7.4, and the final concentration of each protein was adjusted to 1.5 μM. After incubation at 25°C for 30 min, the intrinsic tryptophan fluorescence was measured as described previously . The quenching of tryptophan fluorescence caused by non-specific interactions with each ligand was corrected by use of 1.5 μM N-acetyl-L-tryptophanamide. The dissociation constants (Kd) for binding between the ligands and each L-PGDS were calculated using the method described previously .
Structural model of human L-PGDS
Homology modelling was performed by using the SWISS-MODEL workspace [32,33]. The homology model of WT human L-PGDS was built by using the solution structure of the C65A mouse L-PGDS (PDB code 2RQ0 ) as a template to illustrate the environment around Cys65 in human L-PGDS. Human L-PGDS (168 residues) shows approximately 75% amino acid sequence identity with mouse L-PGDS. The model obtained was evaluated by PROCHECK . Graphical representations were prepared using PyMOL (http://www.pymol.org).
Data are expressed as means±S.D. The statistical significance between the control and the experimental group was assessed using Student's t test. P<0.05 was considered to be significant.
Cytotoxicity in H2O2-treated human neuronal SH-SY5Y cells
The cytotoxicity of H2O2 towards human neuronal SH-SY5Y cells was investigated by performing the WST-8 assay, which is a general measure of cellular metabolic activity. The results show that H2O2 at 0 to 10 μM had a negligible effect on cell viability (Supplementary Figure S1A at http://www.BiochemJ.org/bj/443/bj4430075add.htm). However, at higher concentrations of 30–100 μM, H2O2 reduced the cell viability in a concentration-dependent manner. These results were almost consistent with the results reported previously [36–39]. The IC50 value was determined to be 49.5±1.4 μM by using the curve-fitting program GraFit Version 3.0 (Erithacus Software).
Since we observed the H2O2-induced cell death of SH-SY5Y cells, we next investigated whether these cells had undergone apoptosis. We examined the nuclear morphologies of dying SH-SY5Y cells by using a fluorescent DNA-binding agent, DAPI. When the cells were incubated with 100 μM H2O2 for 24 h, they displayed the typical morphological features of apoptosis, that is, chromatin condensation in their nuclei (Supplementary Figure S1B). We also examined the ability of H2O2 to cause DNA fragmentation in SH-SY5Y cells and showed that a DNA ladder appeared upon electrophoresis when genomic DNA prepared from H2O2-treated SH-SY5Y cells, but not from untreated cells, was used (Supplementary Figure S1C). These results suggest that H2O2 potentially induced apoptosis in the SH-SY5Y cells.
Protective effect of human L-PGDS against H2O2-induced neuronal cell death
We first investigated the gene expression of human L-PGDS in SH-SY5Y cells. After the cells had been treated with H2O2, RNA was isolated and gene expression was measured by performing quantitative PCR. L-PGDS was expressed in the SH-SY5Y cells, and its expression level was enhanced in an H2O2 concentration-dependent manner (Figure 1A). These results were consistent with the results obtained from Western blot analysis by use of anti-L-PGDS and anti-actin antibodies (Figure 1B).
Next, in order to investigate the effect of human L-PGDS on H2O2-induced neuronal cell death, we prepared three kinds of mutants, i.e. C89A/C167A/C186A, C89A/C186A and C65A/C167A L-PGDS (Figure 1C). As shown in Figure 1(D), a 50 μM concentration of H2O2 reduced the viability of the neuronal cells by 50±1.7% as compared with that of the untreated ones. However, C89A/C167A/C186A L-PGDS, in which Cys65 was retained, protected against H2O2-induced cell death in a concentration-dependent manner. That is, the cell viability was recovered to 77±1.5% of wild-type by C89A/C167A/C186A L-PGDS (5 μM). In addition, C89A/C186A L-PGDS (5 μM) also protected against the H2O2-induced cell death to almost the same degree as C89A/C167A/C186A L-PGDS (78±2.2%). On the other hand, C65A/C167A L-PGDS (5 μM) did not show any protective effect against H2O2-induced cell death. These results, taken together, indicate that L-PGDS expression was enhanced in the neuronal cells by the oxidative stress and that L-PGDS had the ability to protect against the neuronal cell death. Furthermore, the essential role for Cys65 was demonstrated.
Association of human L-PGDS level with cell viability in H2O2-treated neuronal cells
Next, we examined the protective effect of endogenous L-PGDS in H2O2-treated SH-SY5Y cells by employing the siRNA-knockdown method to reduce the level of endogenous L-PGDS. First, to investigate the knockdown efficiency of L-PGDS expression by siRNA, we transfected SH-SY5Y cells with either of two siRNAs, i.e. siRNA#1 or siRNA#2, for human L-PGDS. The mRNA level of L-PGDS was significantly decreased by more than 50% by either siRNA as compared with that when treated with the NC siRNA (Figure 2A). Moreover, either L-PGDS siRNA suppressed the level of L-PGDS protein as compared with that obtained with the NC siRNA, although the actin level, an internal control, was not altered in any sample (Figure 2B). We used the more suppressive siRNA#1 (L#1) for further study. The cells were transfected with either L#1 or NC siRNA, and then cultured with 100 μM H2O2 for 24 h. The cell viability of the NC siRNA-treated cells was decreased approximately 43% as compared with that of untreated cells (Figure 2C). Moreover, when the siRNA-transfected cells were treated with H2O2, their viability was decreased further, approximately 62%, as compared with that of the untreated ones (Figure 2C). Furthermore, the LDH activity of the NC siRNA-treated cells was increased approximately 158% as compared with that in the untreated cells (Figure 2C). When the cells were transfected with L#1 siRNA and incubated with H2O2, their LDH activity was slightly enhanced, being approximately 184%, as compared with that in the untreated cells (Figure 2C). When we utilized L-PGDS siRNA#2 instead of L#1 (Supplementary Figure S2 at http://www.BiochemJ.org/bj/443/bj4430075add.htm), almost the same results were obtained.
Then, to confirm that L-PGDS was involved in the prevention of H2O2-induced cell death of SH-SY5Y cells, we expressed FLAG-tagged WT or C65A L-PGDS (C65A) protein in SH-SY5Y cells, and then incubated the cells in medium containing H2O2. First, we used Western blot analysis to detect the heterologous expression of FLAG-tagged L-PGDS WT or C65A mutant protein in the cells. When the cells were transfected with L-PGDS WT or C65A mutant vector, both L-PGDS WT and C65A mutant proteins were detected by using anti-FLAG monoclonal antibody (Figure 3A). In contrast, no corresponding signal was observed in the empty vector-transfected cells, although the actin level was not altered in any sample (Figure 3A). Moreover, L-PGDS was expressed in the empty vector-transfected cells, but a stronger signal was detected when the cells were transfected with either L-PGDS WT or C65A mutant vector (Figure 3A).
Next, cells were transfected with either the L-PGDS WT or C65A mutant vector and then cultured with 100 μM H2O2 for 24 h. The viability of the empty vector-transfected cells in the presence of H2O2 was decreased approximately 48% compared with that in its absence (Figure 3B). Moreover, when the L-PGDS WT vector-transfected cells were treated with H2O2, their viability was approximately 66% of that of the empty vector-transfected ones (Figure 3B). In contrast, no apparent preventive effect was observed when the cells were transfected with the L-PGDS C65A mutant vector (Figure 3B). Moreover, LDH activity was elevated approximately 167% when the empty vector-transfected cells were incubated with H2O2 (Figure 3B). It is noteworthy that, when the cells were transfected with the L-PGDS WT vector, LDH activity was decreased to approximately 136% of the value for the empty vector-transfected cells (Figure 3B). In contrast, no repressive effect on LDH activity was observed when the cells were transfected with the L-PGDS C65A mutant vector and treated with H2O2 (Figure 3B). These results indicate that human L-PGDS plays a critical role in the prevention of H2O2-induced cell death of SH-SY5Y cells.
Titration of free thiol in L-PGDS with DTNB
As we considered Cys65 to be a key amino acid to reduce the neuronal cell death induced by H2O2, we investigated the thiol modification of cysteine residues of L-PGDS by DTNB. Figure 4(A) shows the free thiol titration of three human L-PGDS mutants, C89A/C186A (open circle), C89A/C167A/C186A (closed circle) and C65A/C167A L-PGDS (open triangle), by various concentrations of H2O2. In the absence of H2O2, the free thiol concentration of C89A/C186A L-PGDS was 8.13±0.32 μM. This concentration was decreased in a concentration-dependent manner by treatment with H2O2. Moreover, when the cells were treated with 50 μM H2O2, the free thiol concentration reached a plateau corresponding to approximately 40% of that by the non-treatment with H2O2. In case of C89A/C167A/C186A L-PGDS, the free thiol concentration was estimated to be 3.96±0.14 μM, and was decreased in an H2O2 concentration-dependent manner. Free thiol almost disappeared by the treatment with 50 μM H2O2. In the case of C65A/C167A L-PGDS, however, no free thiol was detected, even in the absence of H2O2. Figure 4(B) shows the time course of changes in the free thiol concentration of L PGDS mutants caused by treatment with 50 μM H2O2. The free thiol concentrations of both C89A/C186A (open circle) and C89A/C167A/C186A (closed circle) L-PGDS mutants were decreased in a time-dependent manner after the initiation of treatment with H2O2, and reached a plateau at 12 h. These results show that H2O2 reacted with the thiol of only Cys65 of human L-PGDS, but not with that of Cys167.
Thiol modification of Cys65 of human L-PGDS by S-oxidation
To assess the oxidative status of Cys65 of L-PGDS, we subjected C89A/C167A/C186A and C65A/C167A L-PGDS to linear positive-ion MALDI–TOF-MS analysis. The MALDI–TOF-MS spectrum of C89A/C167A/C186A L-PGDS showed a peak at m/z 18746, whereas by the treatment with 50 μM H2O2, C89A/C167A/C186A L-PGDS showed a peak at m/z 18778 which corresponded to an increase in the mass of 32 Da (Figure 5A). On the other hand, C65A/C167A L-PGDS, lacking the active-centre Cys65, showed a peak at m/z 18778 and did not show any change in molecular mass by treatment with 50 μM H2O2 (Figure 5B). These results demonstrate that Cys65 oxidized by H2O2 to its sulfinic acid (-SO2H) form, i.e. Cys65 of human L-PGDS is involved in the scavenging of ROS.
Tryptophan fluorescence quenching of oxidized L-PGDS by small lipophilic molecules
We reported previously that L-PGDS is an extracellular transporter protein for small lipophilic molecules such as retinoids, BV, bilirubin and thyroid hormones [11,40,41]. So, to investigate whether the oxidized L-PGDS possessed the ability to bind small lipophilic molecules, we measured the fluorescence quenching of intrinsic tryptophan residues of L-PGDS with various concentrations of BV or all-trans-RA. Human L-PGDS contains three tryptophan residues, located at positions 43, 54 and 112, and these tryptophan residues are considered to contribute to the fluorescence quenching of L-PGDS by binding with small lipophilic molecules. Both untreated and oxidized C89A/C167A/C186A L-PGDS showed BV concentration-dependent fluorescence quenching (Figure 6A). In the presence of 10 μM BV, both fluorescence intensities decreased to less than 10% of that in the absence of BV. From the quenching curves, the Kd values of BV of untreated and oxidized C89A/C167A/C186A L-PGDS were calculated to be 91±10 μM and 110±20 μM respectively. In the case of RA, again both untreated and oxidized C89A/C167A/C186A L-PGDS showed fluorescence quenching, which occurred in an RA concentration-dependent manner, and the Kd values calculated were 1520±20 μM and 770±320 μM respectively (Figure 6B). These results show that oxidized L-PGDS retained its ability to bind small lipophilic molecules.
Critical roles for amino acid residues around Cys65 of human L-PGDS in scavenging ROS
To investigate the difference in the reactivity toward H2O2 between human and mouse L-PGDS, we constructed a structural model of WT human L-PGDS (Figure 7A). The model obtained for human L-PGDS had 97.9% of its amino acid residues within the favoured or allowed regions of the Ramachandran ϕ–ψ plots, indicating a good quality of the predicted model. Figures 7(B) and 7(C) show magnifications of the area around Cys65, which is S-oxidized by ROS, in human and mouse L-PGDS. In human L-PGDS, Ser63 and Ser67 were respectively arranged above and below Cys65, whereas in the mouse L-PGDS, Tyr63 and Thr67 were respectively situated above and below Cys65. The other amino acid residues around Cys65 in mouse L-PGDS, such as Trp43, Ser45, Leu48, Ser81, Phe83 and Arg85, were located in the same position as in human L-PGDS. So we prepared Y63S/T67S/C89A/C186A mouse L-PGDS, i.e. a human-type mouse L-PGDS, and compared the efficiency as a ROS scavenger of human, mouse and human-type mouse L-PGDSs. Figure 7(D) shows the changes in free thiol concentration of C89A/C167A/C186A human (open circle), C89A/C186A mouse (closed circle) and Y63S/T67S/C89A/C186A mouse (open triangle) L-PGDS incubated with various concentrations of H2O2. In all cases, the free thiol concentrations decreased in a concentration-dependent manner by treatment with H2O2. From these curves, the IC50 values of H2O2 against human, mouse and human-type mouse L-PGDS were calculated to be 12.1±0.7, 38.4±5.7 and 19.3±1.9 μM respectively (GraFit Version 3.0). These results indicate that human L-PGDS showed higher reactivity towards H2O2 than mouse L-PGDS and that the amino acid modification around Cys65 in the mouse L-PGDS corresponding with human L-PGDS led to increased reactivity towards H2O2. Thus we concluded that the environment around Cys65 in L-PGDS is essential for the reaction with H2O2.
Apoptosis is the process of cell death characterized by cell shrinkage, nuclear condensation, DNA fragmentation and membrane blabbing. Neurons are very delicate, being susceptible to oxidative stress, and such stress is tightly associated with cellular damage in a variety of human diseases including neurodegenerative disorders [42,43]. ROS, including H2O2, superoxide anion and hydroxyl radicals, damage biological molecules such as lipids, proteins and DNA, which leads to apoptotic or necrotic cell death . H2O2 has been extensively utilized as an inducer of oxidative stress in in vitro cell culture systems . The treatment of cells with H2O2 results in an imbalance in energy metabolism and harmful effects of hydroxyl and peroxyl radicals on membrane lipids and proteins. Much experimental evidence indicates the occurrence of oxidative stress in the pathogenesis of various neurodegenerative diseases . Thus the removal of excessive ROS or suppression of their generation is critical for the prevention of oxidative cell death in various diseases.
L-PGDS was first identified in the rat brain as a PGD2-synthesizing enzyme. L-PGDS was reported to be one of the most abundant proteins in human CSF and is expressed in the mammalian central nervous system in locations such as the choroid plexus and leptomeninges, and in oligodendrocytes [10,21]. Interestingly, L-PGDS, as a protein of the lipocalin family, binds small lipophilic molecules, such as retinal and RA [11,40], BV, bilirubin [13,41,47] and gangliosides . Therefore L-PGDS has been considered as a unique bifunctional, i.e. multifunctional, protein acting as both a PGD2-synthesizing enzyme and a lipophilic molecule-binding protein, i.e. a lipocalin. In the present study, we found a novel property of L-PGDS as a ROS scavenger in vitro. L-PGDS expression was increased in the H2O2-treated SH-SY5Y cells (Figures 1A and 1B), and a reduction in the L-PGDS level was seemed to correlate with an increase in H2O2-induced cell death (Figure 2C). Prostaglandins are known to be involved in the regulation of neuronal cell death. 15d-PGJ2 (15-deoxy-Δ12,14-prostaglandin J2), one of the PGD2 metabolites, has a potential to suppress H2O2-induced cell death of mouse neuronal N18D3 cells  and rat PC12 cells . In contrast, Kondo et al.  reported that 15d-PGJ2 induced the apoptotic cell death of SH-SY5Y cells. Therefore the roles for 15d-PGJ2 in the regulation of neuronal cells death are controversial, indicating the requirement of further study for elucidation of the roles for PGs on the regulation of neuronal cell death. In the present study, exogenously added prostaglandins, including PGD2 which is non-enzymatically metabolized to 15d-PGJ2, had no effect on the viability of H2O2-treated SH-SY5Y cells (Supplementary Figure S3 at http://www.BiochemJ.org/bj/443/bj4430075add.htm). Moreover, the L-PGDS C65A mutant, lacking PGD2-synthesizing activity, could not decrease the H2O2-induced cell death of SH-SY5Y cells (Figures 1D and 3B). Thus the PGD2-synthesizing activity of L-PGDS was not related to the protection against H2O2-induced cell death of SH-SY5Y cells, indicating that the ability of L-PGDS to afford this protection was due to some other feature of this protein. As shown in Figure 1(D), L-PGDS suppressed the H2O2-induced death of SH-SY5Y cells in a dose-dependent manner by scavenging ROS. So far, several studies have been reported related to the protective effects of lipocalin proteins against oxidative stress. Tear lipocalin has been shown to bind lipid-peroxidation products and also to be up-regulated in human NT2 neuronal precursor cells during oxidative stress . α1-Microglobulin binds toxic haem  and has antioxidant catalytic activities . Apolipoprotein D is an acute-response protein with a protective effect on the brain, and its expression is enhanced by oxidative stress in the brain . However, the molecular mechanism of the scavenging effects of lipocalins remains unclear.
In the present study, the ROS-scavenging mechanism effected by L-PGDS was clearly shown to be the oxidation of the thiol group of its Cys65, which resulted in the conversion of this cysteine residue into cysteine sulfinic acid (-SO2H). This is the first report to clarify the ROS-scavenging mechanism by lipocalin family proteins. In addition, the S-oxidized L-PGDS still possessed the binding property for small lipophilic molecules such as BV and RA (Figure 6), and, conversely, the complex of L-PGDS and RA also possessed ROS-scavenging capability (Supplementary Figure S4 at http://www.BiochemJ.org/bj/443/bj4430075add.htm). Furthermore, human L-PGDS exhibited higher reactivity towards H2O2 than did mouse L-PGDS, and the amino acid modification around Cys65 in mouse L-PGDS, making it correspond to human L-PGDS, led to increased reactivity towards ROS (Figure 7). These results suggested to us that the specific amino acid composition near the active site of human L-PGDS was responsible for the more prominent protective action against ROS by the human L-PGDS than by the mouse one. L-PGDS is not only a multifunctional protein, but also a multitasking protein able to simultaneously perform the crucial functions of both ROS scavenging and the binding of small lipophilic molecules. Thus L-PGDS is a unique multitasking protein in the lipocalin family proteins.
In summary, we found that L-PGDS plays protectively against the H2O2-mediated apoptosis in human neuronal SH-SY5Y cells by scavenging ROS. Free radicals produced from H2O2 bind to the active centre of L-PGDS, indicating that L-PGDS might be useful for the suppression of oxidative stress-mediated neurodegenerative diseases.
Mao Yamada, Ko Fujimori and Takashi Inui designed the research; Mao Yamada, Ayano Fukuhara, Ko Fujimori, Hidemitsu Nakajima and Takashi Inui performed the biochemical research; Yuya Miyamoto performed the structural modelling; Toshihide Kusumoto performed the MS analysis; and Ko Fujimori and Takashi Inui wrote the paper.
This work was supported in part by Grants-in-Aid for Scientific Research (B) and (C) [grant numbers 17300165 and 21500428 (to T.I.)], by Grants-in-Aid for Scientific Research on Innovative Areas [grant number 21200076 (to T.I.)] and by a Special Research Grant (to T.I.) from Osaka Prefecture University.
We thank Nobuko Uodome (Osaka Bioscience Institute, Osaka, Japan) for technical assistance.
Abbreviations: BV, biliverdin; CSF, cerebrospinal fluid; DAPI, 4′,6-diamidino-2-phenylindole; 15d-PGJ2,15-deoxy-Δ12,14-prostaglandin, J2; DTNB, 5,5′-dithiobis-(2-nitrobenzoic acid); HRP, horseradish peroxidase; LDH, lactate dehydrogenase; L-PGDS, lipocalin-type prostaglandin D synthase; MALDI–TOF, matrix-assisted laser-desorption ionization–time-of-flight; NC, negative control; PGD, prostaglandin D; RA, retinoic acid; ROS, reactive oxygen species; siRNA, small interfering RNA; TNB, 5-thiobenzoic acid; WT, wild-type
- © The Authors Journal compilation © 2012 Biochemical Society