Prohibitin reduces mitochondrial free radical production and protects brain cells from different injury modalities

Abstract

Prohibitin is an essential mitochondrial protein that has been implicated in a wide variety of functions in many cell types, but its role in neurons remains unclear. In a proteomic screen of rat brains in which ischemic tolerance was induced by electrical stimulation of the cerebellar fastigial nucleus, we found that prohibitin is upregulated in mitochondria. This observation prompted us to investigate the role of prohibitin in neuronal death and survival. We found that prohibitin is upregulated also in the ischemic tolerance induced by transient ischemia in vivo, or oxygen-glucose deprivation in neuronal cultures. Cell fractionation and electron-microscopic immunolabeling studies demonstrated that prohibitin is localized to neuronal mitochondria. Upregulation of prohibitin in neuronal cultures or hippocampal slices was markedly neuroprotective, whereas prohibitin gene silencing increased neuronal vulnerability, an effect associated with loss of mitochondrial membrane potential and increased mitochondrial production of reactive oxygen species. Prohibitin upregulation was associated with reduced production of reactive oxygen species in mitochondria exposed to the complex I inhibitor rotenone. In addition, prohibitin protected complex I activity from the inhibitory effects of rotenone. These observations, collectively, establish prohibitin as an endogenous neuroprotective protein involved in ischemic tolerance. Prohibitin exerts beneficial effects on neurons by reducing mitochondrial free radical production. The data with complex I activity suggest that prohibitin may stabilize the function of complex I. The protective effect of prohibitin has potential translational relevance in diseases of the nervous system associated with mitochondrial dysfunction and oxidative stress.

Figure 1.

Identification of PHB by 2-D gel electrophoresis and mass spectrometry. A, Mitochondrial proteins from FN- and DN-stimulated rat brain were separated on 2-D gels, and protein spots were visualized by Ruby Red staining. The open arrow indicates the spot that corresponds to PHB with a probability of 100%. The closed arrow indicates a spot that corresponds to annexin V binding protein with 60% probability. B, Western blot of mitochondrial proteins from FN- and DN-stimulated rat brain confirming that PHB is upregulated in mitochondria of brains preconditioned with FN stimulation. The mitochondrial marker cytochrome c oxidase (COX IV) was used to assure equal gel loading. Each lane contains 10 μg of mitochondria. C, Quantification of band intensity in B. *p < 0.05; n = 7/group. Error bars indicate SEM.

Figure 2.

PHB is upregulated in models of ischemic preconditioning in vivo and in vitro. A, PHB is upregulated in mitochondria of brains preconditioned by BCCAO. B, Quantification of band intensity in A (*p < 0.05; n = 7). C, PHB is upregulated in cell lysates from neuronal cultures preconditioned with OGD. D, Quantification of band intensity in C (*p < 0.05; n = 7/group). E, Neurons preconditioned with a nonlethal OGD episode are more resistant to damage produced by lethal OGD. Cell viability was assessed by morphological criteria 24 h after lethal OGD (*p < 0.05; n = 4/group with >1000 cells/group). Error bars indicate SEM.

Figure 3.

Neuronal localization of PHB. A, Subcellular fractionation of mouse neocortex. PHB is present mainly in the mitochondrial fraction and is not observed in the nuclear or cytosolic fraction. COX IV, histone H3, GAPDH, and calreticulin were used as mitochondrial, nuclear, cytosolic, and membrane markers, respectively. Results shown are representative from three separate experiments. B, C, Electron micrographs showing the immunoperoxidase localization of PHB in neuronal and glial profiles in the rat somatosensory cortex. B, PHB immunoreactivity (ir) is associated with inner and outer membranes of a mitochondrion (Pm) in a dendrite (P-den), located near an intensely labeled plasma membrane (arrow) contacted by an unlabeled terminal (U-te). Other neuronal processes within the neuropil contain unlabeled mitochondria (Um). C, PHB labeling is located on the plasmalemma (arrow) of a glial process (P-g) that sheaths small neuronal process containing mitochondria, which are either PHB-labeled (P-m) or unlabeled (U-m). Scale bar, 500 nm. D, In neuronal cultures, PHB immunoreactivity is colocalized with the mitochondrial marker MitoTracker Red.

Figure 4.

Expression of human PHB is neuroprotective. A, PHB expression improves viability of neurons exposed to STS (0.25 μm for 1 h). Noncondensed nuclei identify viable neurons (*p < 0.05; n = 4/group, 250 cells/group). B, PHB expression improves viability in neurons exposed to X/XO (X, 150 μm; XO, 15 mU for 1 h) (*p < 0.05; n = 4/group, 250 cells/ group). C, AAV mediated PHB-myc-GFP expression in hippocampal slices. D, Myc-tagged PHB is expressed in high level in the slices infected with AAV-PHB-myc but not control vector (AAV-ctrl). Anti-PHB antibody was used to detect the PHB-myc tag fusion protein. E, OGD-induced cell death assessed by PI staining is lower in slices that expressed PHB than in controls. Slices infected with control (AAV-ctrl) and PHB (AAV-PHB-myc) viruses were treated with OGD and imaged. The same slices were then treated with high dose of NMDA to induce maximum cell death (MAX) (see Materials and Methods for details). F, Measurement of PI intensity in slices after OGD (*p < 0.05; n = 3/group, 15–20 slices/group). Error bars indicate SEM.

Figure 5.

Downregulation of endogenous PHB by RNAi increases neuronal vulnerability. A, si-PHB-2 downregulates endogenous PHB expression, while si-control and si-PHB-1 have no effect. B, si-PHB-2 decreases PHB protein band intensity by >75% (*p < 0.05; n = 4). C, “Knockdown” of PHB by si-PHB-2 enhances the increase in caspase-3 activity induced by STS compared with si-Ctrl and si-PHB-1 (*p < 0.05 from respective vehicle; ^#p < 0.05 from si-PHB-1 and si-Ctrl; n = 6/group). D, si-PHB-2 reduces the viability of neuronal cultures exposed to X/XO, assessed by trypan blue exclusion (*p < 0.05 from si-PHB-1; n = 4/group, 2500 cells/group). E, si-PHB2 enhances the decrease in relative membrane potential (JC1 fluorescence ratio) induced by X/XO (*p < 0.05; n = 5). Error bars indicate SEM.

Figure 6.

siRNA-induced PHB downregulation enhances glutamate neurotoxicity. A, Endogenous PHB downregulation in neuronal cultures is present 13 d after si-PHB-2 treatment. B, si-PHB-2 reduces PHB band intensity by 55% (*p < 0.01; n = 3 separate experiments). C, D, si-PHB-2 increases ROS production induced by glutamate (25 μm for 5 h) and assessed by MitoSOX (*p < 0.05; n = 3 separate experiments). Magnification: 200×. E, si-PHB-2 enhances glutamate induced neuronal death and assessed by the MTS reduction assay (*p < 0.01; n = 3 separate experiments). Error bars indicate SEM.

Figure 7.

Expression of PHB in neurons suppresses ROS production induced by complex I, but not complex III, inhibition and preserves complex I activity. A, The increase in MitoSOX fluorescence induced by rotenone (5 μm) is attenuated in neurons expressing PHB. n = 5 /group, 120 cells/group. B, Similar results were obtained when rotenone-induced ROS production was assessed by DHE (5 μm). n = 4 /group, 85 cells/group. C, PHB expression in neurons has no effect on the ROS production induced by the complex III inhibitor antimycin A (1 μm). n = 5 /group, 95 cells/group. D, Effect of rotenone on complex I activity in mitochondria isolated from PC12 cells transfected with PHB or control vector. Complex I activity is resistant to rotenone inhibition in PHB-expressing cells (*p < 0.05; n = 4/group). The inset illustrates PHB upregulation in transfected PC12 cells. Error bars indicate SEM.