Crucial Roles of Glu60 in Human Neuroglobin as a Guanine Nucleotide Dissociation Inhibitor and Neuroprotective Agent

Mammalian neuroglobin (Ngb) protects neuronal cells under conditions of oxidative stress. We previously showed that human Ngb acts as a guanine nucleotide dissociation inhibitor (GDI) for the α-subunits of heterotrimeric Gi/o proteins and inhibits reductions in cAMP concentration, leading to protection against cell death. In the present study, we created human E60Q Ngb mutant and clarified that Glu60 of human Ngb is a crucial residue for its GDI and neuroprotective activities. Moreover, we investigated structural and functional properties of several human Ngb mutants and demonstrated that the neuroprotective effect of human Ngb is due to its GDI activity and not due to its scavenging activity against reactive oxygen species.


Introduction
Globins are iron porphyrin complex (heme)-containing globular proteins that bind reversibly to oxygen (O 2 ) and, as such, play an important role in respiratory function. Mammalian neuroglobin (Ngb) is widely expressed in the brain [1][2]. The iron atom in the heme prosthetic group of Ngb normally exists in either the ferrous (Fe 2+ ) or the ferric (Fe 3+ ) redox state. Both the ferric and ferrous forms of Ngb are hexa-coordinated to endogenous protein ligands, namely proximal and distal His residues, and O 2 displaces the distal His residue of ferrous Ngb to produce ferrous O 2 -bound Ngb [3]. The ferrous O 2 -bound form of Ngb that exists under normoxia is converted to the ferric conformation during oxidative stress, inducing large tertiary structural changes [4]. Mammalian Ngb proteins can protect neurons from hypoxic-ischemic insults and protect the brain from experimentally induced stroke in vivo [5][6][7][8].
Hypotheses of the neuroprotective mechanism of human Ngb have been reported previously [9][10][11][12][13][14]. Initially, Ngb was suggested to be an O 2 storage protein [1]. However, the low concentration (in the micromolar range) of Ngb in brain tissues except for the retina perhaps argues against a role for Ngb in storing and carrying significant amounts of O 2 . Alternatively, Ngb may act as an intracellular scavenger of reactive oxygen species (ROS) and/or nitric oxide [15][16][17][18][19]. The reaction of ferric Ngb with hydrogen peroxide does not generate the highly reactive cytotoxic ferryl (Fe 4+ ) species [18]. This property may be beneficial under conditions of oxidative stress.
To investigate other functions of human Ngb under conditions of oxidative stress, we previously performed yeast two-hybrid screening using human Ngb as a bait and identified flotillin-1, a lipid raft microdomain-associated protein, as a binding partner of human Ngb [20]. We demonstrated that human Ngb is recruited to lipid rafts by interacting with flotillin-1 only during oxidative stress and that lipid rafts are crucial for neuroprotection by Ngb [21]. We found that human ferric Ngb, which is generated under oxidative stress conditions, binds exclusively to the GDP-bound form of the a-subunits of heterotrimeric G i/o proteins (Ga i/o ), which is present in lipid rafts and inhibits adenylate cyclase activity [22], thereby acting as guanine nucleotide dissociation inhibitor (GDI) for Ga i/o and inhibiting the reduction of intracellular cAMP concentration to protect against cell death [10,21,23]. By contrast, we previously showed that human ferrous ligand-bound Ngb under normoxia does not interact with Ga i/o and does not have GDI activity [10,21,23]. We recently demonstrated that human Ngb acts as a non-receptor-mediated oxidative stress-responsive sensor for signal transduction in the brain [10,21,24].
Although Ngb was originally identified in mammalian species, it is also present in non-mammalian vertebrates [25,26]. We found that zebrafish Ngb does not exhibit GDI activity [27]. In order to clarify residues of human Ngb that are crucial for its GDI activity, we prepared human Ngb mutants with a focus on residues differing between human and zebrafish Ngb and on exposed residues with positive or negative charges on the protein surface [27]. We showed that human E53Q, R97Q, E118Q, and E151N Ngb mutants, which did not function as GDI proteins, did not rescue cell death under oxidative stress conditions [8,27], indicating that Glu53, Arg97, Glu118 and Glu151 of human Ngb are crucial residues for its GDI activity and that the GDI activity of human wild-type (WT) Ngb is tightly correlated with its neuroprotective activity. Furthermore, Matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) mass spectrometry (MS) analysis of tryptic peptides derived from a cross-linked complex between human WT Ngb and Ga i1 , which is a member of the Ga i/o family [22], revealed cross-linking between Glu60 (Ngb) and Ser206 (Ga i1 ), and between Glu53 (Ngb) and Ser44 (Ga i1 ) [23]. Glu53 as well as Arg97, Glu118 and Glu151 of Ngb are conserved only among boreotheria mammals [28], whereas Glu60 is highly conserved throughout vertebrates. As shown in Figs. 1A and 1B, Glu53 and Glu60 of human Ngb are located in and near the CD-D region, where large tertiary structural changes are induced by the conversion of ferrous O 2 -bound Ngb to ferric Ngb during oxidative stress [4].
In the present study, to clarify the role of Glu60 of human Ngb upon neuroprotective activity, we created human E60Q Ngb mutant. We clarified that the Glu60 of human Ngb is a crucial residue for its GDI and neuroprotective activities. Moreover, we investigated the structure and ROS scavenging activities of human E53Q, R97Q, E118Q, and E151N Ngb mutants as well as the E60Q mutant and demonstrated that the neuroprotective effect of human WT Ngb is due to its GDI activity and not due to its scavenging activities against ROS. [8,5'-3 H]GDP (20-50 Ci/mmol) was purchased from PerkinElmer Life Sciences (Boston, MA). Horse heart myoglobin (Mb) and N-acetylcysteine (NAC) were obtained from Sigma-Aldrich (St. Louis, MO) and Wako Pure Chemical Industries (Osaka, Japan), respectively. Preparation of recombinant human Ngb proteins Plasmids for human Ngb were prepared as described previously [10,27]. A QuikChange TM site-directed mutagenesis system (Stratagene, La Jolla, CA) was used for site-directed mutagenesis. The constructs were confirmed by DNA sequencing (FASMAC Co., Ltd., DNA sequencing services, Atsugi, Japan). Overexpression of each Ngb was induced in E. coli strain BL 21 (DE 3) by treatment with isopropyl-b-D-thiogalactopyranoside (IPTG) for 4 h, and each Ngb protein was purified as described previously [8,10,21,27]. In brief, soluble cell extracts were loaded onto DEAE sepharose anion-exchange columns equilibrated with buffer A (20 mM Tris-HCl, pH 8.0). Ngb proteins were eluted from columns with buffer A containing 150 mM NaCl, and further purified by passage through Sephacryl S-200 HR gel filtration columns. Ngb proteins were then applied to a HiTrap Q HP column (GE Healthcare Biosciences, Piscataway, NJ), eluted with a 0-500 mM linear NaCl gradient in buffer A. Purified Ngb was dialyzed overnight against phosphate-buffered saline (PBS). Endotoxin was removed from the protein solutions by phase separation using Triton X-114 (Sigma-Aldrich) [29,30]. Trace amounts of Triton X-114 were removed by passage through Sephadex G25 gel (GE Healthcare Biosciences) equilibrated with PBS. The protein concentration of human ferric Ngb was determined spectrophotometrically using an extinction coefficient of 122 mM 21 cm 21 at the Soret peak.

Circular dichroism (CD) spectra
CD spectra in the far UV region were measured with a spectropolarimeter (J-805; JASCO Co., Tokyo, Japan) at 20uC. The samples were measured at a concentration of approximately 5 mM in 50 mM sodium phosphate buffer (pH 7.4). The path length of the cells used for the measurements was 1 mm. The molar ellipticity (deg cm 2 dmol 21 ) was determined on the mean residue basis. The a-helix content (f H ) was calculated according to Chen et al. [31] by the following equation:

Denaturation assays
Guanidine hydrochloride (GdnHCl)-induced denaturation experiments were carried out in 50 mM sodium phosphate buffer (pH 7.4), containing various concentrations of GdnHCl. The solutions contained 5 mM protein and were incubated for at least 4 h. CD spectra from 200 to 250 nm were measured. The fractional denatured population (f D ) under each condition was estimated by the following equation: The free energy of denaturation, DG, was calculated by the following equation: When DG varies linearly with the GdnHCl concentration, [GdnHCl], DG H2O , extrapolated to DG at [GdnHCl] = 0, can be estimated by the following equation: where m GdnHCl is the slope of the linear relation between DG and [GdnHCl].

Preparation of recombinant human Ga i1 protein
The DNA fragment containing the human Ga i1 subunit (residues 1-354) was amplified by PCR and cloned into the pET151/D-TOPOH vector (Invitrogen, Carlsbad, CA) to be expressed as human WT Ga i1 protein fused to a TEV protease recognition site directly after an N-terminal tag of six histidine residues (His 6 -tag) [21]. The resulting Ga i1 was expressed in E. Coli strain BL21 (DE3) by induction with IPTG and purified by using a nickel affinity column (His?BindH resin; Novagen, Madison, WI), as described in [21]. Then, the sample was incubated with His 6tagged TEV protease (MoBiTec GmbH, Göttingen, Germany) and loaded onto a His?BindH column to separate the cleaved Ga i1 from the cleaved His 6 -tag, any uncleaved protein, and His 6 -tagged TEV protease, as described in [21].

[ 3 H]GDP dissociation assays
GDP dissociation assays were performed, as described in [21]. Cell culture SH-SY5Y cells (CRL-2266) were obtained from the American Type Culture Collection (ATCC; Manassas, VA) and maintained in a 1:1 mixture of Dulbecco's modified Eagle's medium (DMEM) and Ham's F-12 nutrient mixture containing 2.5 mM glutamine, supplemented with 10% (v/v) fetal bovine serum (FBS), 100 U/ml penicillin, and 100 mg/ml streptomycin (all from Invitrogen) in a humidified atmosphere containing 5% CO 2 at 37uC. The medium was changed every 4 days, and the cultures were split at a 1:20 ratio once a week. Cultured cells were induced to differentiate into a neuronal phenotype by treatment with 10 mM retinoic acid (Sigma-Aldrich) over a period of 6 days (media were exchanged every 3 days during sub-culture). Differentiation was verified by monitoring macroscopic changes to the cells.

Protein transduction by Chariot
Protein transduction was performed by using Chariot TM (Active Motif, Carlsbad, CA) as described previously [8,21]. Each purified Ngb protein (3 mg per well) was incubated in the presence of diluted Chariot for 30 min at room temperature. Next, the mixture was added to differentiated SH-SY5Y cells that had been washed in DMEM without serum. DMEM without serum was added and the cells were incubated at 37uC for 1 h; FBS was then added to a final concentration of 2%. The cells were incubated at 37uC for another 2 h to allow Ngb internalization.

Oxygen-glucose deprivation (OGD)
The day before experiments, differentiated SH-SY5Y cells were plated on a poly-D-lysine-coated 96-well tissue culture plate at a density of 5610 5 cells / mL. Protein transduction was performed by Chariot TM . An in vitro model of ischemia was created by maintaining SH-SY5Y cells under OGD for 16 h, followed by 24 h of recovery. In brief, the standard culture medium was replaced with a glucose-free OGD buffer (154 mM NaCl, 5.6 mM KCl, 5.0 mM HEPES, 3.6 mM NaHCO 3 , and 2.3 mM CaCl 2 at pH 7.4). Hypoxia was induced in a multi-gas incubator (Astec, Fukuoka, Japan; set to 1% O 2 , with 5% CO 2 and 94% N 2 ) at 37uC for 16 h. After hypoxia, the culture medium was replaced with standard culture medium ( a 1 : 1 mixture of Dulbecco's modified Eagle's medium and Ham's F-12 containing 2.5 mM glutamine, supplemented with 10% (v/v) fetal bovine serum), and the cells were incubated at 37uC for 24 h under normoxia (95% air/5 % CO 2 ).

Transfection of human Ngb expression vector into SH-SY5Y cells and treatment of cells with hydrogen peroxide
The eukaryotic expression vector pcDNA3.1 (Invitrogen) for human Ngb was prepared as described previously [21]. A QuikChange TM site-directed mutagenesis system (Stratagene, La Jolla, CA) was used to introduce the E60Q substitution and the construct was confirmed by DNA sequencing (FASMAC Co., Ltd., DNA sequencing services). Differentiated SH-SY5Y cells were plated on poly-D-lysine coated 96-well plates at a density of 5.0610 5 cells/mL for 24 h. The pcDNA3.1-human WT or E60Q Ngb expression vector or control vector (pcDNA3.1 empty vector) was transfected by using Lipofectamine TM 2000 (Invitrogen) according to the manufacturer's instructions. After 24 h of transfection, hydrogen peroxide was added at 100 mM and cells were incubated for 24 h.

Western blot analyses
Protein transduction or transfection of expression vector was confirmed by Western blot analyses using rabbit anti-Ngb (FL-151) polyclonal antibody (Santa Cruz Biotechnology, Santa Cruz, CA), mouse anti-b-actin monoclonal antibody (Sigma-Aldrich). After washing, membranes were incubated with an HRP-linked F(ab') 2 fragment of donkey anti-rabbit IgG or an HRP-linked whole antibody of sheep anti-mouse IgG (GE Healthcare Biosciences). Proteins were visualized using ECL TM western blotting detection reagents (GE Healthcare Biosciences). Chemiluminescent signals were detected using a LAS-4000 mini luminescent image analyzer (GE Healthcare Biosciences).

Hydroxyl radical scavenging assay
The hydroxyl radical scavenging activities were measured and analyzed, as described previously [19,32]. In this system, hydroxyl radicals were generated by the Fenton reaction. Briefly, the reaction mixture included 100 mL of 0.75 mM 1,10-phenanthroline, 200 mL of 5 mM PBS (pH 7.2), 100 mL of 0.75 mM FeSO 4 , 50 mL of 0.01% hydrogen peroxide, and 50 mL of Milli-Q-purified water. The reaction was initiated by addition of hydrogen peroxide. After incubation at 37uC for 60 min, the absorbance of the mixture at 536 nm was measured, as A f . The hydroxyl radical scavenging activity was calculated by the following equation: where A 0 is the absorbance using Milli-Q-purified water instead of hydrogen peroxide, A s is the absorbance using sample (100 mg/ mL) instead of hydrogen peroxide, and A x is the absorbance using sample (100 mg/mL) instead of Milli-Q-purified water.

Structural analyses of human E60Q Ngb mutant
Initially, we evaluated the effects of the E60Q mutation on the electronic state of the heme group by measuring the absorption spectra of the E60Q Ngb mutant. Fig. 2A shows the UV-visible spectra of the ferric, ferrous deoxy, and ferrous carbon monoxide (CO)-bound forms of human E60Q Ngb. Because ferrous O 2bound Ngb is unstable and is converted into ferric Ngb very rapidly due to autoxidation [3], stable ferrous CO-bound Ngb was used as a model for ferrous O 2 -bound Ngb. The wavelengths of the Soret peaks of human ferric, ferrous deoxy, and ferrous CObound E60Q Ngb were 413, 425, and 418 nm, which were the same as those of human WT Ngb (413, 425, and 418 nm), respectively [33], demonstrating that the E60Q mutation of human Ngb did not perturb the electric state of the heme group. Moreover, the absorption ratio of the Soret peak and at 280 nm suggested that the Ngb mutant bind heme just as tightly as the WT protein (Table 1).
Next, to examine the effect of the E60Q substitution upon secondary structure, we measured the far UV CD spectra of the ferric forms of human WT or E60Q Ngb. As shown in Fig. 2B, human WT and E60Q Ngb proteins exhibited two negative broad peaks around 222 and 208 nm, which are characteristic of an ahelical structure. The a-helical content of the E60Q Ngb protein was estimated to be 67.9%, which is almost identical to that of human WT Ngb (68.9%) ( Table 1). These results showed that the secondary protein structure is not affected by the amino acid substitution.
Alterations in equilibrium stability caused by the E60Q substitution were quantified in a GdnHCl-induced denaturation experiment. The GdnHCl denaturation process was followed by monitoring the ellipticity value at 222 nm, which reflects structural changes in the whole protein. As shown in Fig. 2C, the ferric forms of WT and E60Q Ngb showed the cooperative transition curves. The transition curve for GdnHCl denaturation of E60Q Ngb was similar to the curves for human WT Ngb, indicating that the globular structure of E60Q Ngb was as stable as those of the WT Ngb.

Glu60 of human Ngb is a crucial residue for its GDI and neuroprotective activities
To examine the effect of the E60Q mutation of human Ngb upon the release of GDP from Ga i1 , we measured the rates of GDP dissociation in the absence or presence of human Ngb. As shown in Fig. 3, [ 3 H]GDP release from [ 3 H]GDP-bound Ga i1 was inhibited by human ferric WT Ngb in the presence of an excess amount of unlabeled GTP. By contrast, human ferric E60Q Ngb did not function as the GDI for Ga i1 (Fig. 3).   Next, we used SH-SY5Y cells differentiated into a neuron-like type to evaluate whether human Ngb protects cells against ischemia in an in vitro OGD model of in vivo ischemia-reperfusion insult. Protein transduction was achieved by using the protein delivery reagent Chariot, which can efficiently deliver a variety of proteins into several cell lines in a fully biologically active form [8,21,[34][35][36], and was confirmed by Western blot analyses (Fig.  4A). MTS assays showed that cell survival was significantly Human WT or E60Q Ngb was transfected into differentiated cells using Chariot. Following OGD/recovery, cell viability was measured by MTS (B) and TBE assays (C). All data are expressed as means 6 SEM from three independent experiments, each performed in triplicate. Data were analyzed by one-way ANOVA followed by Tukey-Kramer post hoc tests. *P,0.05, **P,0.01. (D) Western blot analyses of SH-SY5Y cell lysates after transfection. Control vector, human WT or E60Q Ngb expression vector was transfected into differentiated SH-SY5Y cells with Lipofectamine. The cells were then incubated for 24 h. Cell lysates were analyzed on 15.0% or 12.5% SDS/PAGE and by Western blot analyses using rabbit anti-Ngb polyclonal antibody or mouse anti-b-actin monoclonal antibody, respectively. (E) Effect of the E60Q mutation in human Ngb on SH-SY5Y cell death caused by hydrogen peroxide. Differentiated SH-SY5Y cells transfected with control vector, human WT or E60Q HNgb expression vector with Lipofectamine were treated with hydrogen peroxide, and cell viability was measured by MTS assay. All data are expressed as means 6 SEM from six independent experiments, each carried out in triplicate. Data were analyzed by one-way ANOVA followed by Tukey-Kramer post hoc tests. **P,0.01. doi:10.1371/journal.pone.0083698.g004 enhanced by the transduction of human WT Ngb into SH-SY5Y cells (Fig. 4B). TBE assays also showed that protein transduction of human WT Ngb via Chariot resulted in a significant increase in cell viability (Fig. 4C). These results suggest that human WT Ngb is effective in rescuing SH-SY5Y cell death induced by the OGD model. By contrast, the E60Q Ngb mutant, which lacked GDI activity, did not significantly rescue cell death under oxidative stress conditions (Figs. 4B and 4C). Moreover, the protective effect of human WT Ngb, but not human E60Q Ngb, was also confirmed by the following experiment: a pcDNA3.1-human WT or E60Q Ngb expression vector, or a control vector (pcDNA3.1 empty vector) was transfected into SH-SY5Y cells by Lipofectamine and the protective effects of Ngb proteins against hydrogen peroxide-induced cell death was tested. Expression of human Ngb proteins was confirmed by Western blot analyses (Fig. 4D). As shown in Fig. 4E, MTS assays showed that human WT Ngb enhanced cell survival. By contrast, human E60Q Ngb did not protect SH-SY5Y cells against cell death (Fig. 4E). Taken together, we conclude that the Glu60 of human Ngb is crucial for its GDI and neuroprotective activities.
The neuroprotective effect of human WT Ngb is not due to its scavenging activities against ROS We previously showed that human E53Q, R97Q, E118Q, and E151N Ngb mutants, which do not function as GDI proteins, do not rescue cell death under oxidative stress conditions [8,27]. In the present study, we evaluated the effects of these mutations on the electronic state of the heme group and on the secondary structures of the ferric form of the Ngb mutants by measuring the UV-visible spectra and the far UV CD spectra, respectively. The absorption spectra of human Ngb mutants were nearly identical to, and the a-helical contents were almost the same as, that of human WT Ngb (Table 1). Alterations in equilibrium stability caused by the substitutions were quantified in a GdnHCl-induced denaturation experiment. The denaturation transition curves were fitted by a single two-state model to obtain the thermodynamic parameters of Ngb. As shown in Table 1, m GdnHCl values and DG H2O values of human Ngb mutants were almost the same as that of human WT Ngb, indicating that the globular structures of Ngb mutants were as stable as that of the wild-type Ngb. These data verified that these mutations do not induce significant structural changes, indicating that Glu53, Arg97, Glu118, and Glu151 are crucial residues for the GDI and neuroprotective activities of human Ngb.
In order to gain further insight into the neuroprotective mechanism of Ngb under conditions of oxidative stress, we measured the scavenging activities against the hydroxyl radical, which is the most reactive oxidant among ROS. Our data clearly showed that hydroxyl radical scavenging activities of human Ngb mutant proteins are almost identical to that of the WT Ngb and are much higher than that of Mb (Fig. 5). These results suggest that the neuroprotective effects of human WT Ngb are not due to their scavenging activities against ROS.

Discussion
In the present study, we clarified the significance of Glu60 of human Ngb in terms of GDI and neuroprotective activities. It should be noted that the amino acid sequence surrounding Glu60 in human Ngb has a motif homologous to those of the R6A-1 peptide and KB-752 peptide, which interact with GDP-bound Ga i1 (Fig. 6). R6A peptide was isolated by in vitro selection using mRNA display to identify a novel peptide sequence that binds with high affinity to Ga i1 and was minimized to a 9-residue sequence (R6A-1) that retains high affinity and specificity for the GDP-  Homologous residues based on the sequence alignment among human Ngb and some peptides containing core motif for Ga binding are highlighted in yellow. Numbers above the sequences correspond to those of the residues of human Ngb. Gaps in the sequences are indicated by dashes. Binding sites in human Ngb and Ga i1 identified by MS analyses combined with chemical cross-linking of the proteins are listed: Glu53 and Glu60 of human Ngb are cross-linked to Ser44 and Ser206 of Ga i1 , respectively [23]. The specific Ga i1 amino acid residues that interact with amino acid residues in KB-752 peptide (Trp5, Phe8, and Glu11) are listed based on the structure of the KB-752 peptide bound to Ga i1 [41]. Residues in the switch II (a.a. 199-219) of Ga i1 were highlighted in red boxes. doi:10.1371/journal.pone.0083698.g006 bound state of Ga i1 [37]. It has previously been reported that the R6A-1 peptide interacts with Ga subunits representing all four G protein classes (Ga i/o , Ga s , Ga q/11 , and Ga 12/13 ) and binds to switch II (a.a. 199-219) of Ga i1 [38,39]. R6A-1-like peptide (8.1.08; Fig. 6) was then selected from an mRNA display library of peptides based on R6A-1 [40]. On the other hand, other GDPbound Ga i -binding peptides including KB-752 were identified by screening from a phage display peptide library and showed strong sequence similarity around the motif TWX(E/D)FL [41]. X-ray crystal structural determination of the Ga i1 /KB-752 peptide complex revealed that the conserved motif in KB-752 peptide interacts with the switch II of Ga i1 [41]. Fig. 6 shows the partial sequence alignment among human Ngb, R6A-1 peptide, R6A-1like peptide (8.1.08) and KB-752 peptide. As shown in Fig. 6, a hydrophobic residue (L/V) is conserved at the 4 th position, along with residues (E/D)(F/Y)L at the 8 th -10 th positions. Because our previous MS analyses combined with chemical cross-linking of the proteins clarified that the Glu60 of the Glu-Phe-Leu motif of human Ngb is located in the switch II of Ga i1 like R6A-1 and KB-752 peptides [23], the corresponding region including Glu60 in human Ngb may function as the core motif for Ga binding.
In the present study, we showed that human Ngb mutants, which do not function as GDI, have almost the same hydroxyl radical scavenging activity as human WT Ngb. Moreover, we previously demonstrated that zebrafish Ngb, which does not act a GDI [27], and chimeric ZHHH Ngb, which acts as a GDI [27], have almost the same hydroxyl radical scavenging activity as human WT Ngb [32]. Because the human Ngb mutants and zebrafish Ngb cannot protect cells against oxidative stress, these results suggest that the scavenging activity of human WT Ngb against ROS is not essential for its neuroprotective activity. Moreover, although it has been reported that both human H64V Ngb and Mb generate very reactive cytotoxic ferryl (Fe 4+ ) species upon treatment of the ferric form with peroxide [18], we recently showed that neither Mb nor human H64V Ngb enhanced cell death [21]. Taken together, we conclude that the neuroprotective effect of human Ngb is due to its GDI activity and not to its scavenging activity against ROS.