Abstract
Prohibitin (PHB) is a critical protein involved in many cellular activities. In brain, PHB resides in mitochondria, where it forms a large protein complex with PHB2 in the inner TFmembrane, which serves as a scaffolding platform for proteins involved in mitochondrial structural and functional integrity. PHB overexpression at moderate levels provides neuroprotection in experimental brain injury models. In addition, PHB expression is involved in ischemic preconditioning, as its expression is enhanced in preconditioning paradigms. However, the mechanisms of PHB functional regulation are still unknown. Observations that nitric oxide (NO) plays a key role in ischemia preconditioning compelled us to postulate that the neuroprotective effect of PHB could be regulated by NO. Here, we test this hypothesis in a neuronal model of ischemia–reperfusion injury and show that NO and PHB are mutually required for neuronal resilience against oxygen and glucose deprivation stress. Further, we demonstrate that NO post-translationally modifies PHB through protein S-nitrosylation and regulates PHB neuroprotective function, in a nitric oxide synthase-dependent manner. These results uncover the mechanisms of a previously unrecognized form of molecular regulation of PHB that underlies its neuroprotective function.
SIGNIFICANCE STATEMENT Prohibitin (PHB) is a critical mitochondrial protein that exerts a potent neuroprotective effect when mildly upregulated in mice. However, how the neuroprotective function of PHB is regulated is still unknown. Here, we demonstrate a novel regulatory mechanism for PHB that involves nitric oxide (NO) and shows that PHB and NO interact directly, resulting in protein S-nitrosylation on residue Cys69 of PHB. We further show that nitrosylation of PHB may be essential for its ability to preserve neuronal viability under hypoxic stress. Thus, our study reveals a previously unknown mechanism of functional regulation of PHB that has potential therapeutic implications for neurologic disorders.
Introduction
Prohibitin (PHB) also known as prohibitin1 (PHB1, herein PHB) is a nuclear encoded protein that localizes in brain mitochondria (Zhou et al., 2012) and is essential for life, as germline PHB gene deletion leads to embryonic death in Caenorhabditis elegans (Artal-Sanz et al., 2003) and in mice (Park et al., 2005). PHB resides in the inner mitochondrial membrane where it forms a high-molecular weight complex with its close family member, PHB2, to serve as a scaffold essential for several mitochondrial functions (Merkwirth and Langer, 2009; Thuaud et al., 2013). PHB is involved in the maintenance of mitochondrial morphology (Steglich et al., 1999; Nijtmans et al., 2000; Merkwirth et al., 2008; Anderson et al., 2018), dynamics regulation (Merkwirth et al., 2008), and respiratory chain complex assembly and stability (Anderson et al., 2018, 2019). Studies from us and others have demonstrated that PHB expression is beneficial in counteracting the oxidative stress caused by glutamate in neurons (Zhou et al., 2012) and the inflammatory response-associated oxidative stress and injury in mouse intestinal epithelium (Theiss et al., 2009). When expressed even at modest levels in mice, PHB exhibits a robust neuroprotective effect against ischemic brain injury (Kahl et al., 2018). However, despite these functional characterizations, little is known about how this important protein is functionally regulated.
Nitric oxide (NO) is an important signaling molecule. Recent studies showed that NO is involved in the development of ischemic tolerance, also termed ischemic preconditioning (IPC), a phenomenon in which a short-term, mild ischemic insult to any organ induces the development of substantial resistance to a subsequent major ischemic attack (Murry et al., 1986). Involvement of NO produced by all three isoforms of NO synthase (NOS) in IPC signaling has been substantiated in numerous studies. Inducible NOS-derived NO plays an obligatory role in ischemic preconditioning in mouse brain (Cho et al., 2005). Moreover, genetic deletion of either neuronal NOS (nNOS) or endothelial NOS renders mice insensitive to IPC stimulation and abolishes IPC-mediated neuroprotection, demonstrating an essential role of NO in the protective effect of IPC (Atochin et al., 2003). Together, these studies underscore the importance of NO in IPC. However, the mechanisms and the targets involved in NO-mediated neuroprotection remain elusive.
Based on the requirement of NO for IPC induction, IPC upregulation of PHB expression (Zhou et al., 2012), and the neuroprotection afforded by PHB expression, we hypothesized the existence of a mechanistic link between NO and PHB in neuroprotection. Here, we test this hypothesis using oxygen and glucose deprivation (OGD), an in vitro neuronal model of ischemic stress. We show that NO and PHB are mutually required to maintain cell viability in hypoxic stress, and that mechanistically NO directly interacts with PHB, resulting in PHB protein S-nitrosylation in its sole Cys residue. Furthermore, we show that endogenous NO formation during neural activity is able to modulate the protective effect of PHB via nitrosylation.
Materials and Methods
Materials
Tissue culture reagents (medium, serum, salt supplements, and solutions) were purchased from Thermo Fisher Scientific. Protein S-nitrosylation reagents (N-[6-(biotinamido)hexyl]−3′-(2'-pyridyldithio) propionamide and streptavidin-agarose beads) were from Thermo Fisher Scientific, and NO donors were from Cayman Chemicals. Other chemicals were from Sigma-Aldrich. Anti-PHB antibodies were from Thermo Fisher Scientific. Precast SDS polyacrylamide gels were from Bio-Rad.
Primary cortical neuronal culture
The experimental procedures for primary neuronal culture using timed pregnant mice were approved by the Institutional Animal Care and Use Committee of Weill Cornell Medicine. Primary neuronal cultures were prepared following an established procedure, as described previously (Zhou et al., 2012). Briefly, cortices from day 16 embryos (C57BL/6) were dissected out into digestion buffer containing 0.25% trypsin and 0.15 µg/ml DNase I. The tissue was incubated at 37°C for 4 min followed by triturating using a fire-polished glass Pasteur pipette. The dissociated cells were diluted into 10 ml of HBSS plus 10% FBS. The suspension was collected and centrifuged. The pellet was resuspended into Neurobasal medium supplemented with 2% B27. The cells were counted and seeded onto poly-d-lysine-coated plates at a density of 8 × 105/ml for incubation at 37°C. The medium was changed twice a week, and the cultures were maintained for 5 or 12 d, depending on the purpose of the experiments.
siRNA transfection into primary neurons
siRNA specific for mouse PHB (si-PHB, GGGACUCAUUUCCUCAUCCtt; Thermo Fisher Scientific) was introduced into primary neurons by using Lipofectamine 3000 (Thermo Fisher Scientific) following the protocol from manufacturer. siRNA with scrambled sequences that do not have a homolog to known mammalian sequence was used to serve as a negative control. Primary neurons were plated into six-well plates or 15 cm dishes, transfected with siRNA on day 7, and cultured for 5 more days before being used for experiments.
Oxygen and glucose deprivation and cell death assessment
Injury in neuronal culture was produced by OGD, as described previously (Zhou et al., 2012). To start OGD, medium in neuronal cultures (12 days in culture) was replaced with buffered salt solution without glucose and cells were subjected to OGD for 4 h in a hypoxia chamber (Billups-Rothenberg) by flushing the chamber with 95% N2 and 5% CO2. Sham cultures were treated with buffer containing glucose under normoxia and served as controls. After OGD treatment, cells were washed, and normal culture medium was added and returned to the incubator for an additional 24 h before cell viability assessment by morphologic criteria as described previously (Zhou et al., 2005) with modifications. For nuclear morphology assessment, cells were first fixed in 4% paraformaldehyde for 10 min followed by incubation in PBST (PBS buffer with 0.1% Triton X-100) for 10 min. The cells were then incubated with DAPI (0.5 µg/ml in PBS) for 10 min at room temperature. The cells were then imaged under a fluorescence microscope equipped with a DAPI filter (Nikon). Both DAPI and corresponding bright-field images (BFIs) from 10–15 randomly selected fields were taken at 20× (50–100 cells/field) for each plate of cells. Merged DAPI images and BFIs of the same field were used to identify neurons and exclude them from counting cell debris and astrocytes. Live or dead cells were determined based on nuclei morphology in DAPI images and cell body-associated neurite intactness in BFI. Cells with fragmented or condensed nuclei and fragmented neurites in continuity with the cell body were counted as dead, while cells with uncondensed nuclei and intact cell bodies and associated neurites were considered alive. In case there were cells with intact nuclei but fragmented neurites, they were excluded from counting either as alive or dead. In some experiments as indicated, cell viability was also assayed by MTS cellular metabolism measurement with reagents from Promega.
Quantitative real-time PCR analysis of mouse PHB expression in neuronal cultures
PHB mRNA levels after DPTA (diethylenetriaminepentaacetic acid) stimulation were analyzed by quantitative PCR (qPCR), as described previously (Kahl et al., 2018). Briefly, total RNAs from primary neurons were extracted by Invitrogen TRIzol (Thermo Fisher Scientific) according to the manufacturer instructions. After cDNA synthesis from total RNA (5 µg), qPCR was performed by SYBR Premix EX Taq in a Bio-Rad IQ5 PCR System. PCR was performed under the following conditions: denaturation at 95°C for 2 min, 40 cycles of denaturation at 95°C for 15 s, annealing at 60°C for 20 s, and extension at 72°C for 20 s, followed by a final extension at 72°C for 5 min. Finally, the melting curve analysis was performed to confirm that a single product was amplified without primer dimers interfering with the reaction. The comparative Ct method (2-ΔΔCt) was used for the relative mRNA quantification.
Protein nitrosylation assay
A biotin switch assay (Jaffrey et al., 2001) was used to detect the protein Cys-S nitrosylation status of PHB. The assay was performed according to the standard procedure (Forrester et al., 2009). Briefly, lysates of neuronal culture or brain tissue were prepared in HEN buffer (100 mm HEPES, 1 mm EDTA, 0.1 mm neocuproine, pH 8.0). After protein concentration determination, 0.6 mg of protein (1.8 ml) were aliquoted into each tube, blocking reagents added [final concentration: 2.5% SDS, 0.1% methyl methane thiosulfonate (MMTS)], and incubated for 20 min at 50°C to block free-thiol groups. Excess MMTS was removed by cold acetone precipitation (three volumes) followed by 70% acetone washes. The pellets were resuspended in HENS buffer (HEN with 1% SDS) andthe free nitrosothiols were reduced to thiols with ascorbate (20 mm) in the presence of N-[6-(biotinamido)hexyl]−3′-(2′-pyridyldithio)propionamide (biotin-HPDP; 2.5 mg/ml). An aliquot of the samples was saved before the next step to serve as a loading control (input) in the detection step. Biotinylated proteins were pulled down with streptavidin-agarose beads, and the pellet was washed four times with wash buffer (25 mm HEPES, pH 7.5; 100 mm NaCl; 1 mm EDTA; 0.5% Triton X-100; 600 mm NaCl). The proteins on beads were then eluted by incubating with elution buffer (10 mm HEPES, pH 8.0; 0.1 mm EDTA; 1% β-mercaptoethanol), followed by a quick spin for 30 s. The supernatants were collected, mixed with SDS-PAGE sample loading buffer, and loaded onto 12% polyacrylamide gels (Bio-Rad). The nitrosylated proteins were detected by Western blot analysis.
Western blot
After protein concentration determination with DC reagents (Bio-Rad), equal amounts of proteins were loaded on premade 12% polyacrylamide gel (Bio-Rad), separated, and transferred to PVDF membranes. Membranes were incubated with specific antibodies in appropriate dilutions after blocking with 5% dry milk in PBS for 1 h at room temperature. The membranes were washed three times with PBST (PBS + 0.1% Tween-20), incubated with appropriate secondary antibodies for 1 h, and washed, and images were acquired with a CLx Imaging Station (LI-COR). Band intensity was analyzed using Image Studio software (version 3.1; LI-COR).
PHB protein sequence alignment analysis
Amino acid sequences of PHB1 and PHB2 for all species from human to fly and fungus were obtained from the GenBank database. Sequences were aligned using the Clustal Omega program from UniProt (www.uniprot.org/align).
cGMP measurement in neuronal culture
The cGMP content in neuronal cells was measured with a Parameter cGMP assay kit (R&D Systems) following manufacturer protocols. Briefly, cells from different treatments were washed, lysis buffer was added, and cells were frozen–thawed two times. After a brief spin to remove cell debris, the cell lysates were collected for assay. The lysate samples were added to the assay plate and incubated with HRP-conjugated cGMP in fixed amount and an antibody against cGMP for 3 h. After four washes, color developing reagents were added to the plate and the absorbance was read at 450 nm. The cGMP content was derived by calculating from a standard curve constructed with a serial dilution of known concentration of cGMP.
NO measurement with fluorescent indicator DAF-FM
NO generated from nNOS on NMDA receptor activation was detected using the fluorescent NO indicator Invitrogen DAF-FM diacetate (5 mm DMSO stock; catalog # D-23 844, Thermo Fisher Scientific). Neurons [day in vitro 14 (DIV14)] were pretreated with l-NAME (100 μm) for 2 h followed by bicuculline (Bic; 50 μm) for 1 h. Cells were washed with warm medium twice and then incubated with DAF-FM for 30 min. After washing three times with PBS to remove any extracellular dye, the cells were imaged with an ImageXpress Pico Automated Cell Imaging System (Molecular Devices). Fluorescence intensity of the cells was quantitated with ImageJ software after the subtraction of background fluorescence in cells without the dye.
Experimental design and statistical analysis
All the experiments were designed with proper controls. The NO donor experiments were vehicle controlled. Neuronal viability/OGD experiments were sham controlled. In a biotin switch assay, ascorbic acid omission, or NOS inhibitor, was used as a negative control and a NO donor was used as a positive control, respectively. In siRNA-mediated PHB knockdown, a scrambled sequence that had no effect on cell viability and PHB expression was used as control. We performed a neuronal cell viability assay based on cell morphology changes after OGD. To avoid bias, the cell plates being assayed were coded and the identities of the cells were kept blind for designated personnel in image analyses. At least 10 randomly selected fields for each cell type were imaged, and at least 200 cells per cell type were analyzed. Three to five independent experiments were performed. For statistical analyses, all data are presented as the mean ± SEM. Statistical differences between two groups were evaluated by unpaired, two-tailed t test. ANOVA with Tukey's post hoc test was used to compare differences across multiple groups. Differences were considered significant at p < 0.05.
Results
Increased PHB protein level is a prerequisite in NO-mediated maintenance of cell viability in neurons exposed to OGD
To establish an in vitro model of NO-mediated neuroprotection, we first screened several NO donors with different half-lives for their ability to preserve cortical mouse neurons viability under conditions of OGD. Among the NO donors tested, DPTA-NONOate (DPTA) displayed the strongest neuroprotection (Fig. 1A). Therefore, in this study we used DPTA as an NO donor. Taking advantage of this NO-stimulated neuroprotection model, we tested whether NO released from DPTA was able to enhance cell viability in association with increased PHB protein levels in neurons exposed to OGD. We incubated neuronal cultures with DPTA at different concentrations and assayed cell viability and PHB protein levels. We found that DPTA resulted in a significant decrease of OGD-induced cell death, in a dose-dependent manner (Fig. 1B). In addition, DPTA increased neuronal PHB protein content (Fig. 1C,D), mimicking the effect of IPC in vivo (Zhou et al., 2012). Interestingly, quantitative real-time PCR with mRNA samples from a set of DPTA-treated samples showed that PHB transcription was unchanged within the time frame of NO-induced PHB protein increase (Fig. 1E), suggesting that NO-induced PHB upregulation does not occur at the transcriptional level. Together, these data demonstrate that donor-derived NO is able to raise PHB protein levels and provide neuroprotection against OGD.
To understand how NO induces PHB protein increase without enhancing transcription, we preincubated neurons with the protein synthesis inhibitor cycloheximide (CHX; 100 μm) for 2 h preceding DPTA addition, followed by OGD treatment and cell viability assessment. As expected, DPTA-induced PHB protein increase was completely blocked by CHX (Fig. 1F,G), suggesting that NO-mediated PHB upregulation occurs at the protein synthesis level. Furthermore, when cells were incubated with CHX first, followed by DPTA, NO-mediated neuroprotection was abolished (Fig. 1H), suggesting that protein synthesis is needed for NO-mediated neuroprotection.
To specifically demonstrate the role of PHB upregulation in NO-mediated neuroprotection, we downregulated PHB expression with siRNA before exposing neurons to NO and OGD. Transfection of PHB-specific siRNA for 5 d reduced endogenous PHB to ∼50% of a scrambled si-control (si-ctrl; Fig. 1I,J), but did not affect neuronal viability in normal conditions (Fig. 1K), consistent with our previous study (Zhou et al., 2012). Conversely, neurons with reduced PHB were more sensitive to OGD compared with neurons treated with scrambled control siRNA (Fig. 1K). Strikingly, the neuroprotective effect of DPTA was completely abolished by PHB downregulation. These results demonstrate that normal PHB protein levels are required for NO-mediated neuroprotection and that PHB participates in the neuroprotective mechanisms downstream of NO signaling.
NO is essential for PHB expression-mediated neuroprotection
Having established that PHB is required for NO-mediated neuroprotection, we next investigated the role of NO in PHB-mediated neuroprotection. We used l-NAME, a nonselective NOS inhibitor, to deplete NO in neurons followed by hypoxic stress. In neuronal cultures without OGD, NO depletion did not affect cell viability (Fig. 2A), indicating that acute NO depletion in normal conditions is not harmful to cells and that no toxic effects of l-NAME were apparent at the dose used. However, l-NAME treatment resulted in greater neuronal death than vehicle in OGD. The increased cell death was not attributable to altered PHB protein levels, as PHB levels were unchanged in all treatment groups (Fig. 2B,C). These results demonstrate a critical role of NO in PHB neuroprotection, independent of PHB levels.
To further demonstrate that the effects of NO depletion on neuronal survival in OGD are dependent on PHB, we overexpressed PHB by lentivirus (LV) transduction. As expected, without OGD, NO depletion did not cause a decline in cell viability (Fig. 2D, Sham panel), and, when cells were exposed to OGD without l-NAME, overexpression of PHB with LV–PHB transduction provided neuroprotection (Fig. 2D). Strikingly, l-NAME completely abolished the neuroprotection afforded by LV–PHB transduction, without affecting protein overexpression (Fig. 2E,F). These results clearly demonstrate the role of NO for PHB protective function, as increasing PHB protein levels in the absence of NO is insufficient for neuroprotection. Thus, NO is necessary for the protective effect of PHB and, conversely, endogenous PHB content is likely required for NO preconditioning.
NO effects on PHB neuroprotection are not mediated by the soluble guanylyl cyclase/cGMP pathway
Next, we sought to understand how NO modulates PHB function in neuroprotection. NO was originally identified as a vasodilating factor from endothelium through activation of guanylyl cyclase to produce the second messenger cGMP (Murad, 1986; Palmer et al., 1987; Russwurm and Koesling, 2004). More recently, an alternative NO signaling pathway has been discovered in which NO directly reacts with the free thiol group (−SH) of amino acid cysteine to form S-nitrosocysteine, a post-translational protein modification process termed protein S-nitrosylation (Hess et al., 2005; Nakamura and Lipton, 2016; Stomberski et al., 2019). To delineate which pathway is involved in NO modulation of PHB, we first assayed cGMP production in response to DPTA. As anticipated, we observed a time-dependent increase of cellular cGMP with DPTA (Fig. 3A), demonstrating that the cGC/cGMP system is functional in our experimental model. We then used ODQ, a potent cGC inhibitor (Garthwaite et al., 1995), to investigate the role of cGMP on the protective effect of DPTA. ODQ (10 μm) blocked the increase in cGMP induced by DPTA (Fig. 3B) without altering PHB expression (Fig. 3C,D), but did not prevent the protective effects of DPTA in OGD (Fig. 3E). These data indicate that NO-mediated PHB functional modulation does not depend on soluble guanylyl cyclase (sGC) signaling and suggest that protein S-nitrosylation may be involved instead.
PHB is nitrosylated on its highly conserved unique cysteine (Cys69) residue
Protein S-nitrosylation occurs at the thiol group of Cys residues in target proteins. We analyzed PHB amino acid sequences of various species to determine whether they contain Cys residues that could be nitrosylated. Sequence alignment shows that there is one highly conserved Cys at position 69 (Cys69) in vertebrate species, but not in invertebrates, yeasts, or plants (Fig. 4). On the other hand, vertebrate PHB2 does not contain a Cys residues, indicating that PHB2 is not susceptible to S-nitrosylation (Fig. 4). Interestingly, in invertebrate animal species the situation is reversed, and there is one conserved Cys residue in PHB2, but not in PHB (Fig. 4, bottom). The high degree of conservation of the single Cys residues in the animal PHB/PHB2 complex suggests that it may play a role in the function of the complex.
We next assessed whether Cys69 of PHB is nitrosylated using a well established biotin switch method (Forrester et al., 2009) and found that PHB is nitrosylated in neurons in basal conditions (Fig. 5A,B). Treatment with l-NAME to inhibit endogenous NO production resulted in reduced PHB nitrosylation by ∼45% (Fig. 5C,D). Conversely, treatment with the NO donor DPTA increased the level of PHB nitrosylation by approximately onefold from baseline (Fig. 5A,B). Omission of ascorbic acid from the biotin-HPDP reaction buffer resulted in weak or undetectable nitrosylated PHB (Fig. 5A,B), demonstrating that PHB contains a bona fide S-nitrosylated cysteine residue (Derakhshan et al., 2007). To further verify the above result, we measured PHB content in the leftover fraction of avidin bead pulldown, which contains all the proteins that are not nitrosylated. The addition of DPTA resulted in significantly less PHB in the unbound (non-nitrosylated) fraction (Fig. 5E,F), confirming that more PHB is nitrosylated in the presence of the NO donor. Together, these data demonstrate that NO leads to PHB S-nitrosylation in a NOS-dependent manner.
Decrease in endogenous PHB protein level leads to enhanced PHB nitrosylation
It is known that total PHB deletion is lethal to cells but that neurons with partial PHB knockdown remain healthy in normal conditions, without extra stress (Fig. 1I–K), a phenomenon with an unclear biological basis. Based on the fact that PHB protective function requires the presence of NO (Fig. 2) and that PHB is nitrosylated at baseline (Fig. 5), we postulated that nitrosylated PHB levels may be altered in neurons with decreased PHB. To test this hypothesis, we assayed PHB nitrosylation in neurons transfected with si-PHB and control siRNA. si-PHB treatment for 5 d resulted in a significant decrease of PHB protein levels (42 ± 5% of si-ctrl; Fig. 6A,B). Intriguingly, the fraction of nitrosylated PHB in neurons with si-PHB was substantially elevated (68 ± 4% of si-ctrl; Fig. 6A,C). On the other hand, the nitrosylation of β-actin was unchanged (Fig. 6D–F), suggesting that the elevation of nitrosylated PHB when total PHB is reduced is a specific event. Together with the results shown in Figure 1, the data suggest that increasing PHB nitrosylation could play a compensatory role that support normal cellular activities, when total PHB is decreased.
Neuronal activity-induced NO modulates PHBS-nitrosylation and neuroprotection
In order to increase endogenous neuronal NO production, we treated neurons with bicuculline Bic (50 μm for 1 h), a GABA receptor antagonist that promotes neuronal synaptic activity by dampening inhibitory signals (Johnston, 2013). We first assayed NO production in neurons after Bic stimulation with a NO-specific fluorescence dye DAF-FM diacetate (Kojima et al., 1998). NO level was increased in neurons after 1 h of incubation with Bic, and NO increase was blocked by l-NAME, as expected (Fig. 7A,B). We next assayed PHB nitrosylation by the biotin switch assay after Bic incubation, and found that Bic induced higher PHB nitrosylation relative to vehicle-treated neurons (Fig. 7C,D). This increase was completely abolished by simultaneous treatment of neurons with Bic and l-NAME (Fig. 7C,D), demonstrating the requirement of NOS for the activity-dependent increase of PHB S-nitrosylation. Further, to test whether Bic-induced PHB nitrosylation is associated with neuroprotection, we subjected neurons to Bic treatment followed by OGD. In baseline conditions, Bic did not affect cell viability (Fig. 7E). However, Bic-treated neurons survived OGD better than vehicle controls, while l-NAME abolished the protective effect of Bic (Fig. 7E). Together, these data suggest that neuronal activity regulates NOS-dependent PHB nitrosylation and contributes to neuroprotection.
Discussion
PHB plays significant roles in diverse cellular processes (Artal-Sanz et al., 2003; Merkwirth and Langer, 2009; Merkwirth et al., 2012) or in pathologic conditions such as cancer, inflammation, obesity, and diabetes (Wang et al., 2002; Theiss et al., 2007; Ande et al., 2016a,b; Fan et al., 2017; Mishra and Nyomba, 2017). Work from our laboratory has demonstrated that PHB has a potent neuroprotective role against brain ischemic injury (Zhou et al., 2012; Kurinami et al., 2014; Anderson et al., 2018, 2019; Kahl et al., 2018). However, little is known about how PHB is regulated in neurons.
In this study, we explored the relationship of NO, PHB, and neuroprotection using an in vitro model of ischemia-reperfusion injury. We show that in order for PHB to be protective, adequate protein levels are required but not sufficient, as NO is also required. Thus, NO and PHB are mutually interacting for neuronal survival under stress conditions. We further demonstrate that NO modulates PHB function through protein S-nitrosylation and that this post-translational modification is NOS dependent; and that in neurons PHB protein decline elevates PHB nitrosylation, potentially compensating for the partial loss of PHB. Finally, we show that PHB nitrosylation can be stimulated by increased neuronal synaptic activity in a NOS-dependent manner. These findings reveal a previously unrecognized form of functional regulation of PHB by NO.
In the brain, NO has been established as a neurotransmitter (Garthwaite et al., 1988) and is involved in the control of synaptic functions through modulating neurotransmitter release and plasticity (Brenman and Bredt, 1997; Taqatqeh et al., 2009; Neitz et al., 2011, 2014). In many cases, NO exerts its functions by binding to and activating sGC, which generates the multifunctional second messenger cGMP. For example, NO-mediated cGMP is involved in activity-dependent excitatory synapse development (Nikonenko et al., 2013) and is important for neurite growth and synapse remodeling after axotomy (Cooke et al., 2013). In addition to this classical signaling pathway, NO-mediated post-translational modifications, in particular protein S-nitrosylation, have become increasingly recognized as a form of functional regulation of target proteins (Jaffrey et al., 2001; Hess et al., 2005; Foster et al., 2009). For example, S-nitrosylation of Parkin inhibits its ubiquitin EIII ligase activity and compromises its protective function (Chung et al., 2004). On the other hand, protein S-nitrosylation is essential for normal homeostatic assembly and plasticity of GABAergic synapses (Dejanovic and Schwarz, 2014). In the case of PHB, our results indicate a positive role of S-nitrosylation, since PHB nitrosylation status dictates PHB neuroprotective capacity.
In our model, NO executes two separate but related functions leading to neuroprotection against OGD in neurons. One is the enhanced PHB translation to elevate the total PHB protein levels in neuronal culture that recapitulates what occurs in vivo (Zhou et al., 2012). The other is the direct interaction of NO with PHB, resulting in the PHB nitrosylation (Fig. 5). The finding that elevating PHB protein levels without NO is not neuroprotective (Fig. 2) adds a layer of regulation in PHB function and underscores a critical role of NO in cell viability. However, it remains to be elucidated how NO renders PHB functional. One possible mechanisms suggested by our results is through protein nitrosylation. The nitrosylation assay in neurons with reduced PHB supports this hypothesis. Neurons maintain normal viability even when endogenous PHB is reduced by half (Fig. 1I–K). We found that in neurons with low PHB, the proportion of nitrosylated PHB is much higher than that in control cells (Fig. 6), possibly compensating for the decrease in total PHB and suggesting that nitrosylation can be dynamically regulated to support cellular activities. In addition, it is possible that NO regulates PHB function indirectly through nitrosylation of other proteins, which in turn regulates PHB function. Our results also show that PHB protein levels are critical in this regulation. When endogenous PHB is decreased, the effect of NO on neuroprotection is lost (Fig. 1). Therefore, it appears that both PHB and NO are needed to maintain proper neuronal activities. Further work with more specific approaches is needed to fully understand the molecular mechanisms of this PHB regulation.
Cys69 is within the β2 strand and outside of the PHB/PHB2 binding interface predicted by an in silico structure model (Winter et al., 2007). Therefore, it is unlikely that PHB nitrosylation status affects its interaction with PHB2. S-Nitrosylation may affect PHB structure, but it is difficult to hypothesize the precise effect of nitrosylation on PHB structure and tertiary conformation at this stage, because the structures of PHB and PHB/PHB2 complex have not yet been resolved. Another possible way that S-nitrosylation could affect PHB function is through modulating its binding partners. PHB has been reported to bind numerous other proteins. For example, PHB interacts with subunits of complex I (Bourges et al., 2004), it forms a complex with histone regulator A, a specific H3.3 chaperone required for embryonic development, to serve as a key factor in human embryonic stem cell self-renewal (Zhu et al., 2017), and binds to C3 to enhance complement activation (Mishra et al., 2007). In addition, we have previously shown that PHB binds to the essential mitochondrial phospholipid cardiolipin and stabilizes its content, promoting respiratory chain supercomplex assembly and bioenergetic function (Anderson et al., 2018). Thus, it is possible that nitrosylation of PHB changes its conformation or its ability to interact with other proteins and lipids, allowing it to modulate mitochondrial function.
Proteomic approaches have identified stimuli that induce increases in PHB protein in cells or tissues (Smalla et al., 2008; Bernstein et al., 2012; Poitelon et al., 2015). However, the underlying mechanisms responsible for regulating PHB content remain unknown. In this study, we demonstrate that the NO-mediated increase in PHB protein does not occur at the transcriptional level as PHB mRNA is unchanged by NO stimulation. Our data support an alternative mechanism in which PHB is increased by translational activation because inhibition of protein synthesis by CHX abolishes the NO-stimulated increase in PHB. These data suggest a translational regulatory mechanism governing PHB function in various stress conditions, but further biochemical studies are needed to dissect the exact mechanism by which NO stimulates PHB synthesis from the existing pool of mRNA.
Another interesting observation is that, depending on the species, a conserved unique cysteine residue is present either in PHB or PHB2, but not in both (Fig. 4). Thus, covalent linkage between PHB and PHB2 through a disulfide bridge formation cannot occur. Since both PHB and PHB2 are evolutionarily highly conserved proteins, “unfit” mutations leading to a second Cys residue and a potential disulfide bridge formation in PHB/PHB2 dimer may be harmful and eliminated. The one cysteine per PHB/PHB2 dimer configuration could carry evolutionary significance and mediate important interactions of the PHB/PHB2 complex.
In summary, we show here a previously unrecognized direct interaction between NO and PHB in NO-mediated neuroprotection, which results in PHB S-nitrosylation modulated by endogenous NOS activity. These findings reveal a novel regulatory mechanism of PHB, a mitochondrial protein critical for maintaining oxidative phosphorylation in conditions of cellular stress (Anderson et al., 2018). However, more work is needed to understand the role of PHB nitrosylation in neuronal brain physiology and in disease conditions.
Footnotes
The authors declare no competing financial interests.
This work was supported in part by National Institutes of Health Grants R01-NS-067078 (to P.Z.) and R01-NS-034179 (to C.I.).
- Correspondence should be addressed to Ping Zhou at piz2001{at}med.cornell.edu