Identification of residues crucial for the interaction between human neuroglobin and the α-subunit of heterotrimeric Gi protein

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 the decrease in cAMP concentration, leading to protection against cell death. In the present study, we used an eukaryotic expression vector driving high-level expression of human wild-type Ngb or Ngb mutants that either exhibit or lack GDI activities in human cells. We demonstrate that the GDI activity of human Ngb is tightly correlated with its neuroprotective activity. We further demonstrate that Glu53, Glu60, and Glu118 of human Ngb are crucial for both the neuroprotective activity and interaction with Gαi1. Moreover, we show that Lys46, Lys70, Arg208, Lys209, and Lys210 residues of Gαi1 are important for binding to human Ngb. We propose a molecular docking model of the complex between human Ngb and Gαi1.

In the present study, in order to reconfirm the results using the protein transfection reagent "Chariot", we used an eukaryotic expression vector which can provide the ability for high-level expression of human WT Ngb or Ngb mutants with or without GDI activities in human cells. We demonstrate that the GDI activity of human Ngb is tightly correlated with its neuroprotective activity. We also prepared site-directed mutants of Ngb or Gα i1 and investigated neuroprotective activities and protein-protein interactions by performing glutathione S-transferase (GST) pull-down assays. Moreover, we propose a model of the complex between human Ngb and Gα i1 based on our experimental results.

Results
Neuroprotective and Gα i1 -binding assays of E53Q, E60Q, R97Q, E118Q, and E151Q Ngb mutants. We used SH-SY5Y cells differentiated into a neuron-like type to investigate the neuroprotective mechanism of human Ngb under oxidative stress conditions. A pcDNA3.1-human WT or Ngb mutant expression vector, or a control vector (pcDNA3.1 empty vector) was transfected into SH-SY5Y cells by Lipofectamine Control vector, human WT or Ngb mutant 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% polyacrylamide-SDS gels and by Western blot analyses using rabbit anti-Ngb polyclonal antibody or mouse anti-β -actin monoclonal antibody, respectively. The arrow indicates the position expected for Ngb or β -actin. Molecular size markers (in kilodaltons) are shown at the left. (B) Effect of the mutation in human Ngb on SH-SY5Y cell death caused by hydrogen peroxide. Differentiated SH-SY5Y cells transfected with control vector, human WT or Ngb mutant expression vector with Lipofectamine were treated with hydrogen peroxide, and cell viability was measured by MTS assays. All data are expressed as means ± standard error of means (SEM) from four independent experiments, each carried out in triplicate. Data were analyzed by one-way ANOVA followed by Tukey-Kramer post hoc tests. **P < 0.01. (C) GST pull-down assays of human ferric WT Ngb or Ngb mutant with the GDP-bound truncated Gα i1 . GST, GST-human WT Ngb, or GST-human Ngb mutant was incubated with human GDP-bound Gα i1 in a buffer (pH 7.4). Western blot analyses were performed with anti-Gα i1 mouse monoclonal antibody. The arrow indicates the position expected for Gα i1 . Molecular size markers (in kilodaltons) are shown at the left.  (E) Effect of the mutation in human Ngb on SH-SY5Y cell death caused by hydrogen peroxide. Differentiated SH-SY5Y cells transfected with control vector, human WT or Ngb mutant expression vector with Lipofectamine were treated with hydrogen peroxide, and cell viability was measured by MTS assay. All data are expressed as means ± SEM from four independent experiments, each carried out in triplicate. Data were analyzed by oneway ANOVA followed by Tukey-Kramer post hoc tests. **P < 0.01. (F) GST pull-down assays of the human 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. 1A). Cell viability was measured by using 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium, inner salt (MTS). The absorbance at 490 nm was directly proportional to the number of living cells ( Supplementary Fig. S1). As shown in Fig. 1B, MTS assays showed that human WT Ngb enhanced cell survival. By contrast, human E53Q, E60Q, R97Q, E118Q, and E151N Ngb single mutants, which lack GDI activity, did not protect SH-SY5Y cells against cell death (Fig. 1B). These results are consistent with previous results using the protein transfection reagent "Chariot" 10,16 . Taken together, we reconfirmed that the GDI activity of human Ngb is tightly correlated with its neuroprotective activity.
Next, to characterize the protein-protein interaction between human Ngb and Gα i1 in vitro, we performed GST pull-down assays employing human Ngb fused to GST (GST-human Ngb). We previously demonstrated that GST-human ferric Ngb bound to the GDP-bound form of Gα i1 11,24 . These data are consistent with those obtained by surface plasmon resonance of non-tagged Ngb 14 , suggesting that the GST tag has no effect on protein-protein interactions between Ngb and Gα i1 . In the present study, because the truncated Gα i1 protein, which lacks the N-terminal disordered region (25 amino acids), was more stable than the full-length enzyme 25,26 , GST-human ferric Ngb, or GST was purified ( Supplementary Fig. S2), incubated with human truncated Gα i1 for GST pull-down assays and Western blot analyses were performed using antibody against Gα i1 . As shown in Fig. 1C, GST-human WT Ngb bound to the GDP-bound form of the truncated Gα i1 . Moreover, we found that human E53Q, E60Q, R97Q, E118Q, and E151N Ngb mutants did not, suggesting that Glu53, Glu60, Arg97, Glu118 and Glu151 residues of human Ngb are all indispensable for its interaction with Gα i1 .

Analyses based on a possible model of the complex between human Ngb and Gα i1 .
To gain further insight into the structure of the complex between human Ngb and Gα i1 , we created a molecular docking model of the complex based on the following two experimental results: i) Glu53, Glu60, Arg97, Glu118, and Glu151 of human Ngb are all crucial residues for its neuroprotective effect and interaction with Gα i1 ; ii) Glu53 and Glu60 of human Ngb were cross-linked to Ser44 and Ser206 of Gα i1 , respectively 23 . The molecular docking model was created by manually adjusting the strucutres of Ngb and Gα i1 to avoid steric hindrance between Ngb and Gα i1 and is shown in Fig. 2A.
Next, to evaluate this model, we prepared site-directed Gα i1 mutants at residues with positive or negative charges at the protein binding interface between human Gα i1 and Ngb: R86A, K180L, and E236Q Gα i1 single mutants, an E238N, E239Q Gα i1 double mutant, and a R208A, K209A, K210A Gα i1 triple mutant. For example, Lys180 of human Gα i1 was mutated to Leu because the corresponding residue of Lys180 of Gα i1 in human Gα s is Leu. An E238N, E239Q Gα i1 double mutant was generated because the corresponding residues of Glu238 and Glu239 of Gα i1 in human Gα s are Asn and Gln, respectively. As shown in Fig. 2B,C, R208A, K209A, K210A Gα i1 triple mutant did not interact with human ferric WT Ngb, whereas R86A, K180L, and E236Q Gα i1 single mutants, and the E238N, E239Q Gα i1 double mutant bound to Ngb to an extent similar to that of WT Gα i1 . These data suggest that Arg208, Lys209, Lys210, but not Arg86, Lys180, Glu236, Glu238, Glu239, of Gα i1 are crucial for the interaction with Ngb and that the binding model in Fig. 2A is not appropriate.
Functional analyses of Ngb mutants prepared on the basis of amino acid sequence alignment among various species of Ngb. We previously reported that Glu53, Arg97, Glu118, and Glu151 of human Ngb are conserved only among boreoeutheria of mammalia, and not among afrotheria, maetatheria, prototheria, aves, reptillia, amphibia, or osteichtyes 27 , suggesting that boreoeutheria Ngb proteins may protect neurons against oxidative stress-induced cell death. In the present study, we performed more detailed sequence comparisons by using sequence data of Ngb proteins from a greater number of species. As shown in Fig. 3A and Supplementary  Fig. S3, the residue corresponding to Arg97 of human Ngb is a Cys in megabat Ngb. And the residue corresponding to Glu151 of human Ngb is a Gln in bush baby Ngb ( Fig. 3A and Supplementary Fig. S3). Therefore, we created human R97C and E151Q Ngb mutants and investigated their structure, neuroprotective activity and interaction with Gα i1 .
Initially, we evaluated the effects of the R97C or E151Q mutation on the electronic state of the heme group by measuring the absorption spectra. As shown in Fig. 3B and Table 1, the wavelength of the Soret peak of human ferric R97C or E151Q Ngb was the same as those of human ferric WT Ngb, demonstrating that the R97C or E151Q mutation of human Ngb did not perturb the electric state of the heme group. Next, to examine the effect of the R97C or E151Q substitution upon secondary structure, we measured the far-UV circular dichroism (CD) spectra. As shown in Fig. 3C, human WT, R97C and E151Q Ngb proteins exhibited two negative broad peaks around 222 and 208 nm, which are characteristic of an α -helical structure. The α -helical contents of the R97C and E151Q Ngb proteins were estimated to be 68.9% and 73.4%, which is almost identical to that of human WT Ngb (71.4%) ( Table 1). These results showed that the secondary protein structure is not affected by the amino acid substitution.
Ngb mutant with GDP-bound Gα i1 . GST, GST-human WT Ngb, or GST-human Ngb mutant was incubated with human GDP-bound Gα i1 in a buffer (pH 7.4). Western blot analyses were performed with anti-Gα i1 mouse monoclonal antibody. The arrow indicates the position expected for Gα i1 . Molecular size markers (in kilodaltons) are shown at the left. (G) The effect of the R97C or E151Q mutation on dissociation of GDP from human GDP-bound Gα i1 . The amount of [ 3 H]GDP bound to Gα i1 in the absence of human ferric Ngb at 0 min was defined as 100%. All data are expressed as means ± SEM from three independent experiments. Data were analyzed by one-way ANOVA followed by Tukey-Kramer post hoc tests. *P < 0.05, **P < 0.01.
Next, we tested whether R97C and E151Q Ngb can protect SH-SY5Y cells against oxidative stress. We first confirmed that human WT, R97C, and E151Q Ngb proteins were expressed at a similar level in SH-SY5Y cells by Western blot analyses (Fig. 3D). MTS assays showed that the expression of human R97C or E151Q Ngb mutant in the cells rescued cell death under oxidative stress condition, as did that of human WT Ngb (Fig. 3E). Moreover, GST-human ferric Ngb, or GST was purified ( Supplementary Fig. S2), incubated with human truncated Gα i1 and Western blot analyses were performed using antibody against Gα i1 . As shown in Fig. 3F, GST pull-down assays showed that R97C and E151Q Ngb bound to the truncated Gα i1 as did the WT Ngb. Furthermore, to examine the effect of the R97C or E151Q mutation of human Ngb upon the release of GDP from Gα i1 , we measured the rates of GDP dissociation in the absence or presence of human Ngb. In the presence of an excess amount of unlabeled GTP, [ 3 H]GDP release from [ 3 H]GDP-bound Gα i1 was inhibited by human ferric WT Ngb (Fig. 3G). As shown in Fig. 3G, human ferric R97C and E151Q Ngb functioned as the GDI for Gα i1 as did human WT Ngb. These results suggest that the positive charge of Arg97 and the negative charge of Glu151 of human Ngb are not essential for the GDI and neuroprotective activities. Taken together, we conclude that Glu53, Glu60, and Glu118 of human Ngb are the only crucial residues for its activities.
Search for residues of Gα i1 crucial for the interaction with human Ngb. Next, because we found that Arg97 and Glu151 residues of Ngb are not involved in the interface with Gα i1 , we created a new molecular docking model of the complex between human Ngb and Gα i1 , which is depicted in Fig. 4A. To clarify binding sites in the complex between human Ngb and Gα i1 , we prepared K46A, K67A, and K70A Gα i1 single mutants. As shown in Fig. 4B,C, K46A and K70A Gα i1 single mutants did not interact with Ngb. Moreover, the R208A, K209A, K210A Gα i1 triple mutant also did not interact with Ngb (Fig. 2B,C). These results suggest that the model of Fig. 4A is correct. Moreover, K67A Gα i1 bound to Ngb as did WT Gα i1 (Fig. 4B,C), indicating that Glu118 of human Ngb interacts with Lys70, but not Lys67, of Gα i1 in a site-specific manner.

Discussion
In the present study, by using eukaryotic expression vectors for human Ngb proteins we demonstrated that the GDI activity of human Ngb is tightly correlated with its neuroprotective activity. GST pull-down assays using GST-fused human Ngb demonstrated that Glu53, Glu60, Arg97, Glu118 and Glu151 residues of human Ngb are all indispensable for its interaction with Gα i1 . However, R97C and E151Q Ngb single mutants, generated based on sequence alignment among various species of Ngb, showed that the positive charge of Arg97 and the negative charge of Glu151 of human Ngb are dispensable for both the GDI and neuroprotective activities. We therefore speculated that the lack of GDI and neuroprotective activities of the R97Q and E151N Ngb mutants may arise from structural alterations induced by the mutations and concluded that only the negative charges of Glu53, Glu60, and Glu118 of human Ngb are crucial for the GDI and neuroprotective activities. We measured the UV-visible and CD spectra of the WT and mutants but could not detect any significant structural differences among them. Further studies are in progress to clarify structural alterations between the WT and R97Q or E151N Ngb mutant.
Moreover, we identified residues of Gα i1 crucial for the interaction with human Ngb; Lys46, Lys70, Arg208, Lys209, and Lys210 residues of Gα i1 are important for its binding to human Ngb. The present results imply that electrostatic interactions between the negative charges of Glu53, Glu60, and Glu118 of human Ngb and the positive charges of Lys46, Lys70, Arg208, Lys209, and/or Lys210 of Gα i1 are crucial for the formation of a complex between Ngb and Gα i1 . This is consistent with the previous observation that electrostatic complementarity is an important factor for the interaction of Gα with its regulator 28 . Further more detailed studies are in progress to investigate the significance of Arg208, Lys209, and/or Lys210 of Gα i1 . Moreover, because the residue corresponding to Arg97 of human Ngb is a Gln in yak Ngb ( Supplementary Fig. S3), it is interesting to investigate whether yak Ngb has the GDI and neuroprotective activities.
The present results support that a molecular docking model of the complex between human Ngb and Gα i1 in Fig. 4A. In this model, Glu60 (Ngb), which was cross-linked to Ser206 (Gα i1 ) by chemical cross-linking 23 , interacts with Arg208, Lys209, and/or Lys210 of Gα i1 . The amino acid sequence surrounding Glu60 in human Ngb has a motif homologous to those of the R6A-1 peptide and the KB-752 peptide, which interact with GDP-bound Gα i1 16 . It has previously been reported that the R6A-1 and KB-752 peptides interact with the switch II (a.a. 199-219) of Gα i1 26,29,30 . The molecular docking model of the complex between human Ngb and Gα i1 shows that Glu60 of human Ngb is located near the switch II of Gα i1 , as are R6A-1 and KB-752 peptides, suggesting that the motif including Glu60 in human Ngb functions as the core motif for the binding with Gα i1 . As shown in Fig. 4A, Glu53 (Ngb), which was cross-linked to Ser44 (Gα i1 ) by chemical cross-linking 23 , binds to Lys46 of Gα i1 . X-ray structural data of Gα i1 show that Ser44 and Lys46 of Gα i1 are located in the vicinity of GDP 25 . This implies that Ngb could be positioned near the GDP-binding site in the Ngb-Gα i complex, which would inhibit the dissociation of GDP from the binding site. Moreover, it should be also noted that Glu53 of Ngb is conserved among boreotheria except for the harp or hooded seal (Supplementary Fig. S3). In seals Ngb is mainly localized in astrocytes, whereas other boreotheria Ngb exists in neuron 31,32 . The different localization of Ngb suggests that the function of seal Ngb may be different from that of other boreotheria Ngb. Figure 4A shows that Glu118 of Ngb interacts with Lys70 of Gα i1 . We previously carried out chemical cross-linking of the Ngb and Gα i1 complex by treatment with zero-length cross-linkers to identify the sites of interaction between Ngb and Gα i1 23 . In the peptide map of tryptic peptides derived from the cross-linked Ngb-Gα i1 complex, the MS peak of a Ngb peptide (amino acids 103-119 of Ngb), which was observed in the tryptic peptide map of Ngb, was missing or significantly decreased in intensity 23 , suggesting that the 103-119 amino acid residues of Ngb are involved in the Ngb-binding site for Gα i1 , but information about the detailed residues cross-linked in the complex of Ngb and Gα i1 remained unclear. The present results suggest that Glu118, which is located in the 103-119 region, of human Ngb interacts with Lys70 residue, which is located in the α -helical domain of Gα i1 , as shown in Fig. 4A. Strikingly, crystal structure of GoLoco motifs of LGN protein, which act as GDI for Gα i1 , elucidated that they interact with the switch II region and Tyr69 and Val72, which are close to Lys70, in the α -helical domain of Gα i1 33 . It is also worth noting that RGS domains, which bind to Gα i selectively, interact with some residues in the α -helical domain of Gα as well as the switch regions 28 .
Lys46, Arg208, Lys209, and Lys210 residues, which are crucial for the interaction with Ngb, and their sequential neighborhoods of Gα i1 are conserved among Gα i/o and Gα s . In contrast, Lys70 residue of Gα i1 is conserved among Gα i/o , but corresponds to Val in Gα s based on sequence alignment among members of the Gα family 13 . Moreover, sequential neighborhoods around Lys70 of Gα i/o are very different from those of Gα s . Therefore, the positive charge of the 70 th residue of Gα , which interacts with the negative charge of carboxylic group of Glu118 through electrostatic interaction, may be crucial for the Gα i/o -specific binding property of Ngb.

Methods
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 and Ham's F-12 nutrient mixture containing 2.5 mM glutamine, supplemented with 10% (v/v) fetal bovine serum, 100 U/ml penicillin, and 100 μ g/ml streptomycin (all from Invitrogen, Carlsbad, CA) in a humidified atmosphere containing 5% CO 2 at 37 °C. 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 μ M retinoic acid (Sigma-Aldrich, St. Louis, MO) 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.

Transfection of 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 11 . A QuikChange TM site-directed mutagenesis system (Stratagene, La Jolla, CA) was used for site-directed mutagenesis and the constructs were confirmed by DNA sequencing (FASMAC Co., Ltd., DNA sequencing services, Atsugi, Japan). Differentiated SH-SY5Y cells were plated on poly-D-lysine coated 96-well plates at a density of 5.0 × 10 5 cells/mL for 24 h. The pcDNA3.1-human WT or Ngb mutant 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 (100 μ M) was added and cells were incubated for 24 h.

Western blot analyses.
After cell lysates were resolved by electrophoresis on polyacrylamide-SDS gels, proteins were electroblotted onto Hybond-P PVDF membranes (GE Healthcare Biosciences, Piscataway, NJ), which were then blocked with phosphate-buffered saline (PBS) and 5% skim milk (Wako Pure Chemical Industries, Osaka, Japan) and incubated with rabbit anti-Ngb (FL-151) polyclonal antibody (Santa Cruz Biotechnology, Santa Cruz, CA), or mouse anti-β -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).

Preparation and purification of GST and a fusion protein of GST and Ngb. Human Ngb cDNA
was cloned into the pGEX-4T-1 vector (GE Healthcare Biosciences) to produce the fusion protein GST-Ngb 11,24 . A QuikChange TM site-directed mutagenesis system (Stratagene) was used for site-directed mutagenesis. The constructs were confirmed by DNA sequencing (FASMAC Co., Ltd., DNA sequencing services). Overexpression of GST-Ngb and GST alone (as a control) was induced in the Escherichia coli strain BL21 (DE3) (Novagen, Madison, WI) by treatment with isopropyl-β -D-thiogalactopyranoside (IPTG) for 4 h. Both GST-ferric Ngb and GST were purified by using glutathione-Sepharose 4B beads (GE Healthcare Biosciences) according to the manufacturer's instructions.
Preparation of recombinant human truncated Gα i1 protein. The DNA fragment containing the human truncated Gα i1 subunit (residues 26-354) was amplified by PCR and cloned into the pET151/D-TOPO ® vector (Invitrogen) to be expressed as human WT truncated Gα i1 protein (residues 26-354) fused to a TEV protease recognition site directly after an N-terminal tag of six histidine residues (His 6 -tag). A QuikChange TM site-directed mutagenesis system (Stratagene) was used for site-directed mutagenesis. The constructs were confirmed by DNA sequencing (FASMAC Co., Ltd., DNA sequencing services). The resulting Gα i1 was expressed in E. coli strain BL21 (DE3) by induction with IPTG and purified by using a nickel affinity column (His·Bind ® resin; Novagen), as described previously 11 . Then, the sample was incubated with His 6 -tagged TEV protease (MoBiTec GmbH, Göttingen, Germany) and loaded onto a His·Bind ® column to separate the cleaved Gα i1 from the cleaved His 6 -tag, any uncleaved protein, and His 6 -tagged TEV protease, as described previously 11 .