Proteomic Identification of the Involvement of the Mitochondrial Rieske Protein in Epilepsy


Address correspondence and reprint requests to Dr. A. Popa-Wagner at Department of Neurology, Ernst-Moritz-Arndt-Universität Greifswald, Ellernholzstr. 1-2, 17487 Greifswald, Germany. E-mail:


Summary: Purpose: Kindled seizures are widely used to model epileptogenesis, but the molecular mechanisms underlying the attainment of kindling status are largely unknown. Recently we showed that achievement of kindling status in the Sprague–Dawley rat is associated with a critical developmental interval of 25 ± 1 days; the identification of this long, well-defined developmental interval for inducing kindling status makes possible a dissection of the cellular and genetic events underlying this phenomenon and its relation to normal and pathologic brain function.

Methods: By using proteomics on cerebral tissue from our new rat kindling model, we undertook a global analysis of protein expression in kindled animals. Some of the identified proteins were further investigated by using immunohistochemistry.

Results: We report the identification of a modified variant of the Rieske iron-sulfur protein, a component of the mitochondrial cytochrome bc1 complex, whose isoelectric point is shifted toward more alkaline values in the hippocampus of kindled rats. By immunohistochemistry, the Rieske protein is well expressed in the hippocampus, except in the CA1 subfield, an area of selective vulnerability to seizures in humans and animal models. We also noted an asymmetric, selective expression of the Rieske protein in the subgranular neurons of the dorsal dentate gyrus, a region implicated in neurogenesis.

Conclusions: These results indicate that the Rieske protein may play a role in the response of neurons to seizure activity and could give important new insights into the molecular pathogenesis of epilepsy.

Epilepsy is a chronic disorder characterized by recurrent seizures occurring over long periods. The most studied model of chronic limbic epilepsy is kindled seizures, a condition achieved by the repeated administration of an initially subconvulsant stimulus that ultimately results in the generation of seizure activity (1). Although the attainment of the kindling criterion is known to require time to develop, the precise developmental period has not been identified.

We recently reported that optimal achievement of the kindling criterion in the Sprague–Dawley rat is associated with a critical interstimulus interval of 24–26 days. Specifically, a high proportion (>90%) of animals reached full kindling status after only two subconvulsant doses of pentylenetetrazole (PTZ; 30 mg/kg) given 25 days apart; we dubbed this model the critical time window kindling model (2). The identification of this long, well-defined developmental interval for inducing kindling status makes possible a dissection of the cellular and genetic events underlying this phenomenon and its relation to normal and pathologic brain function.

With this model, we have shown that the high-susceptibility period is associated with increased expression of a protease [tissue plasminogen activator (tPA)], a structural protein [microtubule-associated protein 1B (MAP1B)] and an axonal growth-associated protein (GAP43) in the hippocampus. Although this work suggests that the persistent expression of several classes of brain plasticity–associated proteins may be required for the maintenance of kindling status, it seemed likely that a number of other proteins might be involved, the identification of which could further our understanding of the molecular mechanisms that underlie the pathogenesis of epilepsy. Based on a proteomic analysis of the brains of PTZ-treated rats, we here report, for the first time, that a modification of the Rieske iron-sulfur protein in hippocampal neurons is strongly dependent on the development of kindling status.




For the kindling experiments, we used 50 male Sprague–Dawley rats, aged 3–4 months, that were maintained on a 12-h light/dark cycle and allowed free access to food and water. The body weights ranged from 320 to 400 g.

Administration of pentylenetetrazole

To induce full kindling status, 30 rats were treated i.p. with subconvulsant doses of PTZ (30 mg/kg) given 25 days apart (2). Seizure severity was scored according to Racine (3), whereby the maximum score of 5 refers to rearing and falling accompanied by generalized seizures. A group of 20 rats of similar age received i.p. physiologic saline and served as the control group. The mortality rate in this model was very low (2). Ethical approval for these experiments was granted by the university animal experimentation ethics board according to the requirements of the German National Act on the Use of Experimental Animals.


After survival times of 25 days after the last kindling session, the brains were removed and bisected midsagittally. One half of each brain (left and right hemispheres were collected alternately) was fixed in 4% paraformaldehyde, 50 mM phosphate buffer (pH 7.2) for 24 h, cryoprotected in 20% sucrose, 10 mM phosphate-buffered saline (pH 7.2), and stored at –70°C. This hemisphere was used for immunohistochemistry experiments; the hippocampus and the neocortex from the other hemisphere were used for proteomic and immunoblotting analyses. For protein analysis, individual hippocampi were homogenized in a lysis buffer containing 8 M urea, 2 M thiourea, 4% CHAPS, 65 mM DTT, and 40 mM Tris.

2D-PAGE and immunoblotting

First 100–500 μg of protein was separated by isoelectric focusing by using IEF strips with a pH range of 3–10 (IPGphor system; Pharmacia, Uppsala, Sweden). Proteins were then separated on 12% sodium dodecylsulfate–polyacrylamide gel electrophoresis (SDS-PAGE) gels and silver- and/or Coomassie Blue- (Roti Blue; Roth, Karlsruhe, Germany) stained.

For immunoblotting, SDS-PAGE–separated proteins were transferred to nitrocellulose membranes by semidry blotting. The membrane was then incubated with a monoclonal anti-OxPhos Complex III subunit FeS antibody (1:1,000, mouse clone 5A5; Molecular Probes, Leiden, The Netherlands), and the antigen–antibody complex was detected by incubating the membrane with goat anti-mouse alkaline phosphatase immunoglobulin G (IgG) followed by chemoluminescent detection by exposing to x-ray film. Although the primary antibody was directed toward the human antigen sequence, by Western blotting, it also crossreacts with the rat antigen.

After identification of the Rieske protein with mass spectrometry (see later), Coomassie blue–stained spots from the region of interest were cut out of the gel, transferred to nitrocellulose membranes, and analyzed for the presence of the Rieske protein.

Mass spectrometry

In-gel digestion was performed as described by Shevchenko et al. (4), with minor modifications. The protein spots were excised from the gel, washed, and dried. Proteins were digested by using trypsin. Peptides were extracted from the gel matrix with 5% formic acid and with 50% acetonitrile/5% formic acid, pooled, and concentrated. After purification with ZipTips (C18-ZipTip; Millipore, Bedford, MA, U.S.A.), aliquots were analyzed on an α-cyano-4-hydroxycinnamic acid/nitrocellulose matrix by using a Reflex III MALDI-TOF mass spectrometer (Bruker Daltonic, Bremen, Germany). Sequence verification of tryptic peptides was performed by nanoelectrospray tandem mass spectrometry (QSTAR Pulsar I; Applied Biosystems/MDS Sciex, Foster City, CA, U.S.A., and nanoelectrospray needles from BioMedical Instruments, Zoellnitz, Germany).


Free-floating sections (25 μm) were cut on a freezing microtome, collected in Petri dishes, and processed for immunohistochemistry as previously described (2) by using the following antibodies: (a) monoclonal antibody anti-OxPhos Complex III subunit FeS (1: 1,000, mouse clone 5A5; Molecular Probes); (b) mouse monoclonal antibody to the mitochondrial oxidative phosphorylation component, adenosine triphosphate (ATP) synthase, β subunit, also known as F1 complex β subunit (MAB3494, 1:2,000; Chemicon, Hofheim, Germany); (c) mouse monoclonal antibody to the mitochondrial oxidative phosphorylation component, cytochrome oxidase, subunit I (clone 1D6, 1:1,000; Molecular Probes). Color development ensued by using 3,3′-diaminobenzidine/hydrogen peroxide (ABC system, Vectastain Elite Kit; Vector). To visualize nuclei, some sections were counterstained with methyl green. The staining pattern of the monoclonal antibody anti-OxPhos Complex III subunit FeS also was verified on human sections from control tissue.

Light microscopy

For light microscopy, a Nikon microscope was used. Images (768 × 1,024 pixels) were captured electronically by using a Spot CCD camera (Optronics). The digital images were arranged and labeled by using Adobe Photoshop version 8.0 and printed by using a Kodak XLS 8000 digital printer.


Mass-spectrometric identification of the Rieske protein in brain homogenates of rats treated with PTZ

After two-dimensional gel electrophoretic analysis of hippocampal homogenates from kindled animals (Fig. 1B) and controls (Fig. 1A), a complex spot pattern could be imaged by silver staining. In the alkaline region of the 2-D gels, one spot appeared only in samples from kindled rats (Fig. 1B, thick arrow). This spot was analyzed by MALDI-TOF mass spectrometry and identified as being the Rieske protein by searching a database for proteins. This identification was verified by sequencing of two peptides (VPDFSDYR and EIDQEAAVEVSQLR) by using nanoelectrospray tandem mass spectrometry. Database searches (NCBI nr, non-redundant protein database), using the MASCOT software from Matrix Science (5) with carboxymethylation of cysteine and methionine oxidations as variable modifications, led to the identification of the Rieske iron-sulfur protein (p < 0.05). Based on the molecular mass of ∼27 kDa, we detected the apparently unprocessed protein (theoretical pI, 8.9; In seeking the processed protein (theoretical pI, 7.67; MM, 21.5 kDa) in this region, we identified two additional spots as Rieske protein, both of which probably represent unprocessed forms of the protein (marked by arrows in Fig. 1A and B). Analyzing this region by Western blotting revealed two dominant spots from the control rats (Fig. 1C) and one additional, more-alkaline locus in the sample from kindled rats (Fig. 1D, thick arrow). However, no obvious changes were noted in the total amount of the Rieske protein in control and kindled rats as detected by one-dimensional Western blotting (data not shown), nor were there differences after the animals had achieved full kindling status. On PTZ withdrawal, however, after 2 months, the pI of the Rieske protein reverted to the more acidic value, suggesting either that the change is reversible or that the altered protein had been degraded and replaced with the normal one.

Figure 1.

Electrophoretic analysis of hippocampal homogenates of control and kindled rats. Left: A portion of a two-dimensional sodium dodecylsulfate–polyacrylamide gel electrophoresis (SDS-PAGE) gel showing the silver staining of the hippocampal homogenates in controls (A) and kindled animals (B). Arrows, The spots identified by immunoblotting (C, D); the thick arrow indicates the spot identified by mass spectrometry as a modified form of the Rieske protein in pentylenetetrazol (PTZ)-treated animals (B). Note the appearance of a new variant of the Rieske protein in kindled animals having a slightly lower molecular weight and a shift in pI toward a more alkaline value after PTZ treatment (D, thick arrows). E: Long-term changes in the body weight of rats treated with PTZ (solid circles) and controls (open circles). Note that with increasing numbers of seizure episodes, the difference between groups became statistically significant (N1 = 30 for kindled animals and N2 = 20 for controls), *p = 0.05 by ANOVA.

Long-term changes in body weight

Because of the involvement of the Rieske protein in energy metabolism, we measured the body weights after priming the animals with two doses of PTZ, followed by scheduled doses of PTZ every 25th day, over a period of 216 days. During this period, the kindled animals gained more weight than did the control group (Fig. 1E). Although the difference between the groups was small, it became statistically significant, with an increasing number of seizure episodes in the kindled group (p = 0.05).

Immunohistochemical analysis of Rieske protein in rat tissue

By immunohistochemistry, the Rieske protein was strongly expressed in the CA2 region of the hippocampus of control and kindled animals (Fig. 2C) in CA3 neurons (Fig. 2B) and in some interneurons in the polymorphic layer (Fig. 2B, inset, arrow). The Rieske antigen was not detectable in the granule cell layer of the dentate gyrus (Fig. 2B, inset, arrow). However, we noted a selective Rieske-like staining of neuronal somata in the subgranular zone of the dentate gyrus (Fig. 2B, inset, arrowhead). Moreover, the immunostaining was confined exclusively to the dorsal hippocampus (i.e., the distribution of the Rieske protein in the dentate gyrus was highly asymmetrical). Most important, no detectable expression of Rieske protein–like immunoreactivity was found in the hippocampal CA1 region (Fig. 2A and C). The same CA1 region, however, abundantly expressed other markers for mitochondrial energy metabolism located either in complex IV, such as cytochrome oxidase (COX) (Fig. 2D and E), or in the complex V, such as ATP synthase (Fig. 2F). COX and ATP synthase were likewise well expressed in all other hippocampal regions, including the granule cell layer of the dentate gyrus and the polymorphic layer (shown for COX only in Fig. 2D). In the cell, the Rieske protein had a punctate distribution in the cytoplasm (Fig. 2C, inset), much like that of another mitochondrial protein, ATP synthase (Fig. 2F, inset).

Figure 2.

Immunohistochemical analysis of Rieske protein in rat hippocampus. A, D: Low-power magnification of the distribution of the Rieske protein (A) and cytochrome oxidase (COX) (D) in the rat hippocampus. The Rieske protein was well expressed in the CA3 region (B). In the dentate gyrus, we noted sporadic staining of neuronal somata in the subgranular layer (B, inset, arrowhead) and frequent staining of neurons in the polymorphic layer (B, inset, arrow). In contrast, the CA1 region remained devoid of signal (A, C). However, other mitochondrial proteins such as COX (D, E) and ATP synthase (F) were well expressed in the CA1 region as well. At higher magnification, also note the dot-like localization of Rieske protein (C, inset) and of ATP synthase (F, inset) in the neuronal cytoplasm. To visualize nuclei, some sections were counterstained with methyl green (A–C) CA3, CA2, CA1, and DG, hippocampal regions; gcl, granule cell layer; pml, polymorphic layer. Bar, 50 μm (B, C, E, F); 200 μm (A, D).


In this work, we have shown that seizure activity in rats leads to pronounced alterations in the Rieske iron-sulfur protein in specific brain regions. Specifically, a shift toward a more alkaline pI range in the isoelectric point of the Rieske protein occurs in kindled rats. Work is in progress to determine whether the change in the overall charge of the protein is due to a posttranslational modification of the protein.

The Rieske protein is a high-potential iron-sulfur cluster [(2Fe-2S)](ISP) that is part of the mitochondrial cytochrome bc1 complex (ubiquinol:cytochrome c reductase, or complex III), a multifunctional membrane protein complex that catalyzes electron transfer from ubiquinol to cytochrome c but also is involved in proton translocation, peptide processing, and superoxide generation (6). Mutations affecting mitochondrial oxidative phosphorylation complexes have been linked to a large number of clinical phenotypes. For instance, impaired complex III activity due to a gene deletion in cytochrome b is associated with oxidative phosphorylation deficits in parkinsonism (7), and a stop-codon mutation in the cytochrome b gene causes a form of mitochondrial encephalomyopathy (8).

Posttraumatic epilepsy has been modeled by the intracerebral deposition of iron compounds such as hemosiderin, which occurs as a consequence of the extravasation of erythrocytes after injury (9,10). Abnormally high iron levels have been found in patients with epilepsy (11). Likewise, it has been proposed that elevated circulating iron is a risk factor for epilepsy (12). Interestingly, partial seizures have been reported in Bull Terriers with augmented serum iron levels (13). Furthermore, intracortical or amygdalar injection of Fe (III) salts or heme compounds into the rodent brain induces recurrent seizures and epileptic discharges in the EEGs (14–16).

The most conspicuous histopathologic changes occur in the hippocampus, a region heavily implicated in the development of seizures in both humans and animal models. The CA1 subfields of the hippocampal formation in particular have been recognized as being extremely susceptible to damage after seizure activity in humans (17–19) and in rats (20), especially when seizures are accompanied by hypoxia (i.e., a low-energy state) (21,22). The selective reduction of Rieske protein levels in the CA1 region in our rat model lends support to the hypothesis that low-energy states may facilitate seizure development. Conversely, the Rieske protein may have a protective role for regions that are rich in this protein and known to be less prone to damage, such as CA3 and CA2. Equally important is the asymmetric expression of the Rieske protein in the subgranular zone of the dentate gyrus, a region known to be involved in neurogenesis (23). However, it is not clear at this time how the Rieske protein can be related to neurogenesis. The dorsoventral asymmetry in the hippocampus has been documented after seizure activity in mice (24) and seems to be important for the acquisition of spatial memory (25,26). Furthermore, the modified form of the Rieske protein appears to have caused a reduction in mitochondrial activity, the consequence of which would be an accumulation of bodily adipose tissue in kindled rats. This hypothesis is supported by the gain of weight in rats subjected repeatedly to episodes of seizure activity.

Our results are in line with a recent report demonstrating that spontaneous seizure activity induced by pilocarpine treatment of rats causes a decline in the activities of complex I and complex IV in CA1 and CA3, but not in the dentate gyrus (27). Furthermore, positron emission tomography demonstrates hypometabolism in the epileptogenic zone of patients with temporal lobe epilepsy that correlates with impaired oxidative metabolism in the CA3 region (28).


This study suggests a potentially important role of the Rieske protein, in particular, and of mitochondrial energetics, in general, in epileptogenesis.


Acknowledgment:  This research was supported by grants from the Deutsche Forschungsgemeinschaft to A.P.W. (Po359/5-1) and A.N. (No 120/11-1).