|Home | About | Journals | Submit | Contact Us | Français|
Mutations in superoxide dismutase (SOD1) cause amyotrophic lateral sclerosis (ALS), a neurodegenerative disease characterized by loss of motor neurons. With conformation specific antibodies, we now demonstrate that misfolded mutant SOD1 binds directly to the voltage-dependent anion channel (VDAC1), an integral membrane protein imbedded in the outer mitochondrial membrane. This interaction is found on isolated spinal cord mitochondria and can be reconstituted with purified components in vitro. ADP passage through the outer membrane is diminished in spinal mitochondria from mutant SOD1-expressing ALS rats. Direct binding of mutant SOD1 to VDAC1 inhibits conductance of individual channels when reconstituted in a lipid bilayer. Reduction of VDAC1 activity with targeted gene disruption is shown to diminish survival by accelerating onset of fatal paralysis in mice expressing the ALS-causing mutation SOD1G37R. Taken together, our results establish a direct link between misfolded mutant SOD1 and mitochondrial dysfunction in this form of inherited ALS.
Amyotrophic lateral sclerosis (ALS) is a progressive adult-onset neurodegenerative disorder characterized by the selective loss of upper and lower motor neurons in the brain and spinal cord (Cleveland and Rothstein, 2001). The typical age of onset is between 50 to 60 years, followed by paralysis and ultimately death within 2-5 years after onset (Mulder et al., 1986). Most instances of ALS are sporadic lacking any apparent genetic linkage, but 10% are inherited in a dominant manner. Twenty percent of these familial cases have been attributed to mutations in the gene encoding cytoplasmic Cu/Zn superoxide dismutase (SOD1) (Rosen et al., 1993). Although multiple hypotheses have been proposed to explain mutant SOD1-mediated toxicity (Ilieva et al., 2009), the exact mechanism(s) responsible for motor neuron degeneration remains unsettled.
Mitochondrial dysfunction has been proposed to contribute to disease pathogenesis. Histopathological observations of disturbed mitochondrial structure have been reported in muscle of both sporadic and familial ALS patients (Hirano et al., 1984a; Hirano et al., 1984b; Sasaki and Iwata, 1996, 2007) and in mutant SOD1 mouse models expressing dismutase active (Dal Canto and Gurney, 1994; Higgins et al., 2003; Kong and Xu, 1998; Wong et al., 1995), but not inactive mutants (Bruijn et al., 1997). Moreover, functionality of mitochondria has been reported to be affected in spinal cord and skeletal muscles of human sporadic ALS or familial ALS patients (Dupuis et al., 2003; Echaniz-Laguna et al., 2002; Vielhaber et al., 1999; Wiedemann et al., 2002), as well as in some ALS mouse models (Damiano et al., 2006; Mattiazzi et al., 2002; Nguyen et al., 2009).
A proportion of the predominantly cytosolic SOD1 has been reported to localize to mitochondria in certain contexts. In both rodent models and patient samples, mutant SOD1 is present in fractions enriched for mitochondria derived from affected, but not unaffected, tissues (Bergemalm et al., 2006; Deng et al., 2006; Liu et al., 2004; Mattiazzi et al., 2002; Vande Velde et al., 2008; Vijayvergiya et al., 2005) and a clear temporal correlation between mitochondrial association and disease progression was shown for multiple mutant SOD1s (Liu et al., 2004). Purification of mitochondria, including floatation steps that eliminate protein only aggregates, coupled with protease accessibility has demonstrated mutant SOD1 deposition on the cytoplasmic-facing surface of spinal cord mitochondria (Liu et al., 2004; Vande Velde et al., 2008). Sensitivity to proteolysis and immunoprecipitation with an antibody specific for misfolded SOD1 further indicated that misfolded forms of dismutase active and inactive SOD1 are deposited onto the cytoplasmic face of the outer membrane of spinal cord mitochondria (Vande Velde et al., 2008). This is accompanied by altered accumulated levels of a few mitochondrial proteins, reduced import of multiple mitochondrial proteins, and reduced complex I activity (Miller, Vande Velde and Cleveland, unpublished), albeit none of these appear to be direct effects of mutant SOD1.
Oxidative phosphorylation requires the transport of metabolites, including ADP, ATP and inorganic phosphate across both mitochondrial membranes. Located in the outer mitochondrial membrane, the voltage-dependent anion channel (VDAC), known as mitochondrial porin, assumes a crucial position in the cell, controlling metabolic crosstalk between the mitochondrion and the rest of the cell, thus regulating the metabolic and energetic functions of mitochondria (Shoshan-Barmatz et al., 2006; Shoshan-Barmatz et al., 2008). Of the three VDAC isoforms (VDAC1-3), VDAC1 is the most abundant in most cells. VDAC1 is a primary contributor to ATP/ADP flux across the outer mitochondrial membrane (Colombini, 2004; Lemasters and Holmuhamedov, 2006). Initially named somewhat misleadingly as a channel for anions, it is also responsible for import/export of Ca2+ (Gincel et al., 2001) and other cations (Benz, 1994; Colombini, 2004), adenine nucleotides (Rostovtseva and Colombini, 1997; Rostovtseva and Bezrukov, 1998) and other metabolites (Hodge and Colombini, 1997). Indeed, it has been demonstrated that silencing VDAC1 expression in a cultured cell line using shRNA resulted in reduced ATP production and a decrease in cell growth (Abu-Hamad et al., 2006).
VDAC1 is also a key player in mitochondria-mediated apoptosis. VDAC1 has been implicated in apoptotic-relevant events, due to serving as the target for members of the pro- and anti-apoptotic Bcl2-family of proteins (Arbel and Shoshan-Barmatz, 2009; Shimizu et al., 1999) and due to its function in the release of apoptotic proteins from the intermitochondrial membrane space (Abu-Hamad et al., 2009; Shoshan-Barmatz et al., 2006; Shoshan-Barmatz et al., 2008; Tajeddine et al., 2008). VDAC1 has also been implicated in Parkinson's disease as a direct target for Parkin-mediated polyubiquitylation and mitophagy (Geisler et al., 2010).
Starting from recognition that a proportion of misfolded, mutant SOD1 is bound to the cytoplasmic face of the outer membrane of mitochondria in affected tissues (Liu et al., 2004; Rakhit et al., 2007; Vande Velde et al., 2008), we now identify damage to spinal cord mitochondria to arise through direct binding of misfolded SOD1 onto the cytoplasmic-facing domain of VDAC1, thereby inhibiting its conductance.
To investigate potential interactions between mutant SOD1 and VDAC1, mitochondria from rats expressing wild type human SOD1 (hSOD1wt) or either of two different ALS-linked SOD1 mutants, a dismutase active hSOD1G93A and a dismutase inactive hSOD1H46R, were highly purified by repeated centrifugation steps (summarized in Fig. 1A) including a final density gradient flotation step to eliminate any contaminating protein only aggregates (proteins sediment downward in these conditions because of their higher density), as previously described (Vande Velde et al., 2008). Immunoblotting of immunoprecipitates generated after addition of an SOD1 antibody to solubilized mitochondrial lysates revealed that a proportion of VDAC1 was co-precipitated with dismutase active and inactive mutant SOD1, but not wild type SOD1 (Fig. 1B). Parallel immunoprecipitations with a VDAC1 antibody confirmed co-precipitation of both hSOD1G93A and hSOD1H46R with VDAC1 (Fig. 1D). Binding to VDAC1 was a property only of spinal cord mitochondria, as no association of mutant SOD1 was seen with purified brain mitochondria from the same animals using immunoprecipitation with SOD1 (Fig. 1C) or VDAC1 (Fig. 1E) antibodies. This latter finding is consistent with prior efforts that had demonstrated that mutant SOD1 associates with the cytoplasmic face of the outer membrane of mitochondria in spinal cord, but not other tissue types (Liu et al., 2004; Vande Velde et al., 2008). Moreover, mutant SOD1 binding to VDAC1 is inversely correlated with the level of hexokinase-I, a known partner that binds to VDAC1 exposed on the cytoplasmic mitochondrial surface (Abu-Hamad et al., 2008; Azoulay-Zohar et al., 2004; Zaid et al., 2005), with hexokinase accumulating to much higher level in brain than spinal cord mitochondria (Fig. 1F).
To test the nature of the interaction between mutant SOD1 and VDAC1, immunoprecipitation was performed with a SOD1 antibody that recognizes a “disease-specific epitope” (DSE) that is unavailable on correctly folded SOD1 (Cashman and Caughey, 2004; Paramithiotis et al., 2003; Urushitani et al., 2007), but is present on misfolded mutant SOD1s in inherited ALS (Rakhit et al., 2007). Using one such antibody (DSE2), age-dependent deposition of mutant SOD1 onto the cytoplasmic face of spinal cord mitochondria has been shown to reflect association of misfolded SOD1 (Vande Velde et al., 2008). We exploited this antibody to examine if the SOD1 associated with VDAC1 is bound through misfolded SOD1. Liver, brain and spinal cord cytosolic and mitochondrial fractions purified from symptomatic rats expressing mutant hSOD1G93A were immunoprecipitated (see schematic in Fig. 2A) with the DSE2 antibody, which recognizes an epitope in the electrostatic loop of hSOD1 (between residues 125-142) that is buried in normally folded SOD1. Misfolded mutant SOD1G93A was not detectable in the soluble fraction of any tissue, but was immunoprecipitated from the spinal cord, but not liver or brain, mitochondrial fractions (Fig. 2B).
Solubilized spinal cord mitochondria purified from presymptomatic and symptomatic rats expressing either of two different SOD1 mutants, dismutase active hSOD1G93A and dismutase inactive hSOD1H46R, as well as hSOD1wt were immunoprecipitated with the DSE2 antibody and co-immunoprecipitated components identified by immunoblotting. An age-dependent increase in misfolded SOD1 was seen for both mutants, with a significantly higher proportion of the dismutase inactive SOD1H46R in a misfolded conformation. In samples from symptomatic animals, VDAC1 coprecipitated together with the misfolded mutant SOD1, as revealed by immunoblotting of immunoprecipitates (Fig. 2C). This association was selective for VDAC1, as misfolded mutant SOD1 did not co-immunoprecipitate with any of three other mitochondrial proteins examined (Fig. 2C), including two additional outer mitochondrial membrane proteins with domains facing the cytoplasm: TOM40, the 40 kDa component of Transport across the Outer Membrane (TOM) complex mediating all protein import from the cytoplasm to the mitochondria, and VDAC2, a second voltage-dependent anion channel isoform that has been estimated to represent 7 % (kidney) to 25 % (brain) of accumulated VDAC (Yamamoto et al., 2006). It also did not co-precipitate cyclophilin-D, an important component of the permeability transition pore.
Furthermore, in order to determine which cells accumulate the misfolded form of SOD1, we performed immunostaining using the DSE2 antibody. Spinal cords from loxSOD1G37R mice at different stages of the disease were subjected to immunostaining with DSE2 antibody (Fig. 2D). The accumulation of misfolded SOD1 dramatically increased with disease progression. Although little accumulation of misfolded SOD1 is found by disease onset, it was preferentially found within motor neurons. During disease progression, a dramatic increase of misfolded SOD1 was apparently accumulated in other cells as well and probably also extracellularly. Throughout disease a proportion of the misfolded SOD1 was co-localized with mitochondria of motor neurons and other cells, starting at onset and increasing with disease progression (Fig. 2D).
To test if binding of mutant SOD1 affects VDAC1 function, VDAC1 was purified from spinal cords of non-transgenic rats (Fig. 3A) and reconstituted into a planar lipid bilayer (Fig. 3A) using conditions previously demonstrated to yield polarized VDAC1 membrane insertion such that the VDAC1 surface exposed on the cis side is the surface exposed to the cytosol when inserted into the mitochondrial outer membrane (Gincel et al., 2001; Israelson et al., 2005). Activity of individual channels was measured as a function of time by the ions passing across the bilayer in response to an applied voltage gradient. This revealed that in the absence of SOD1, VDAC1 was stably in a fully open state (4 nS at 1 M KCl (Shoshan-Barmatz et al., 2006)) and remained so for extended periods.
Mutant SOD1 proteins hSOD1G93A, hSOD1G85R, as well as hSOD1wt, were expressed using baculovirus and purified (Fig. 3C) (Hayward et al., 2002). Wild type SOD1, even at the highest added concentration (8 μg/ml), had no effect on VDAC1 conductance when added on either cis or trans sides of the membrane (Fig. 3E,I). However, addition of purified recombinant hSOD1G93A or hSOD1G85R (Fig. 3C) substantially reduced VDAC1 channel conductance (Fig. 3F,G). Both mutant SOD1s modified VDAC1 conductance only when added to the cis side (Fig. 3F,G), but not the trans side (Fig. 3J,K) of the bilayer, indicating that mutant SOD1 interacts with what would correspond to the cytosolic face of VDAC1 inserted into the outer mitochondrial membrane. Use of multichannel recordings revealed that not only did mutant SOD1 significantly lower the maximum voltage gated conductance of individual channels, it also provoked a stable, reduced level of VDAC1 conductance at all applied voltages (Fig. 3L-N). In order to determine if this interaction is specific for mutant SOD1, the effect of another aggregating protein (α-synuclein) was tested on bilayers containing reconstituted VDAC1. Even when added to levels 25 times greater than an amount of mutant SOD1 that markedly affected VDAC1 conductance (Fig. 3F,G), neither wild type nor mutant α-synuclein affected VDAC1 channel activity at any voltage (Fig. 1S).
Since both dismutase active and inactive SOD1 mutant proteins reduced VDAC1 channel conductance for K+ and Cl- (Fig. 3), we next tested whether mitochondrial conductance across the outer mitochondrial membrane was affected in animals chronically expressing mutant SOD1. To do this, we examined the uptake into mitochondria of adenine nucleotides (Fig. 4A) which are known to be transported by VDAC1 (Lemasters and Holmuhamedov, 2006; Rostovtseva and Colombini, 1997). Freshly isolated spinal cord and liver mitochondria from SOD1G93A rats were incubated (for 1 min.) with radiolabeled [3H]ADP and the amount of imported ADP was measured by scintillation counting after rapid filtration to remove the unincorporated ADP. Co-incubation with 1 mM of the VDAC1 inhibitor DIDS (4, 4′-diisothiocyanostilbene-2, 2′-disulfonic acid) demonstrated that ~2/3rd of the ADP uptake was through VDAC1 (Fig. 4B,C). Compared to mitochondria from non-transgenic animals, uptake of ADP by spinal cord mitochondria from SOD1 mutant expressing animals was selectively and progressively inhibited, yielding ~40% inhibition of VDAC1-dependent uptake (~25% overall inhibition of ADP uptake) by a symptomatic stage (Fig. 4B). Inhibition of ADP uptake was selective to spinal mitochondria as liver mitochondria from the same hSOD1G93A animals retained normal ADP import at all ages examined (Fig. 4C).
To test if inhibition of ADP import seen in spinal cord mitochondria from mutant SOD1 animals could be generated solely from mutant SOD1 binding to the cytoplasmic face of those mitochondria, purified recombinant SOD1 proteins (hSOD1wt, hSOD1G93A and hSOD1G85R) (Fig. 3C) were added to mitochondria purified from spinal cords or livers of non transgenic rats (Fig. 5A). Although a proportion of each of the recombinant SOD1s associated with both spinal cord and liver mitochondria (Fig. 5D), accumulation of radiolabeled Ca2+ (presumably through the action of the mitochondrial calcium uniporter) into spinal cord or liver mitochondria was not affected by the addition of wild type or mutant SOD1 (Fig. 5C). On the other hand, VDAC1-mediated ADP accumulation into the same spinal cord or liver mitochondria was inhibited by both hSOD1G93A and hSOD1G85R mutants, but not hSOD1wt (Fig. 5B). This inhibition corresponded to a proportion of misfolded SOD1 associated with those mitochondria after incubation with either mutant, but not wild type SOD1, as demonstrated by immunoprecipitation of intact mitochondria with the DSE2 antibody to misfolded SOD1 (Fig. 5E). In contrast, wild type SOD1 associated with the same mitochondria was not recognized by this misfolded SOD1 antibody (Fig. 5E), consistent with its retention of normal folding and/or import into those mitochondria (Fig. 5E).
Since we have established that 1) mutant SOD1 interacts directly with VDAC1 thereby inhibiting VDAC1 conductance (Fig. 3), 2) spinal cord mitochondria from SOD1 mutant animals have progressive loss of ADP uptake, and 3) misfolded mutant SOD1 binds to normal mitochondria in vitro accompanied by selective loss of ADP conductance (Fig. 5), we examined how reduced level and activity of VDAC1 affect disease course in SOD1G37R mutant mice. To do this, we exploited mice heterozygous for disruption of the VDAC1 gene (producing what is effectively a null allele – (Weeber et al., 2002)). These mice accumulate about half the normal level of VDAC1 protein (Fig. 2S), while overall ADP conductance of spinal mitochondrial isolated from VDAC1+/- mice is reduced by ~25% (Fig. 3S) relative to wild type mice. After mating with SOD1G37R mice, sex matched cohorts of mice and their littermates carrying the SOD1G37R transgene and either one or two active VDAC1 alleles were obtained and followed for disease onset, progression and survival. Measurement of ADP conductance of spinal mitochondria from SOD1G37R/VDAC1+/- mice revealed a reduction to a level comparable to that corresponding to complete deletion of VDAC1 (Fig. 3S).
A simple and objective measure of disease onset and early disease progression was applied by initiation of weight loss, reflecting denervation-induced muscle atrophy. While timing of progression from onset through either early (Fig. 6E) or late (Fig. 6F) disease phases was only modestly affected by reduction of VDAC1 levels, disease onset (Fig. 6A, D) and progression to an early disease point (Fig. 6B) were accelerated by 41 and 45 days, respectively, in SOD1G37R/VDAC1+/- mice (183 ± 22 and 230 ± 28 d) compared with their SOD1G37R littermates (224 ± 19 and 275 ± 25 d). Moreover, age at which end stage disease was reached was also reduced by an average of 59 days (Fig. 6C; SOD1G37R/VDAC1+/- mice (310 ± 42 d) compared with their SOD1G37R littermates (369 ± 32 d)). A similar reduction in age of onset and life span was also observed for SOD1G37R/VDAC1-/- mice (Fig. 4S), demonstrating that reduction in VDAC1 activity does affect SOD1 mutant-dependent pathogenesis, primarily by accelerating an early step in disease onset or spread.
We have demonstrated here in floated spinal cord mitochondria from mutant SOD1 expressing animals that both misfolded dismutase active or inactive SOD1 mutants bind directly and selectively to the cytoplasmically exposed face of VDAC1. Both dismutase active and dismutase inactive, but not wild type, SOD1 binding to VDAC1 reduces channel conductance, as demonstrated for K+ and Cl- ions by electrophysiological recording and for ADP by inhibition of normal ADP accumulation into mitochondria. Channel conductance was not affected in liver mitochondria (where misfolded SOD1 does not accumulate). Mutant association and conductance inhibition is replicated in spinal cord mitochondria purified from mutant expressing animals beginning pre-symptomatically and increasing in severity during disease progression contemporaneous with increased accumulation of misfolded mutant SOD1. The clear implication from this is that only the misfolded portion of SOD1 is able to affect the channel, thereby partially blocking metabolite flux across the outer mitochondrial membrane. Reduced conductance by VDAC1 will decrease ATP synthesis, increase the ADP/ATP ratio in the cytosol, and reduce membrane potential (as outlined in Fig. 7). Chronic mitochondrial dysfunction can in turn drive generation of damaging reactive oxygen species that could drive further SOD1 misfolding through chemical damage to it, as has been previously documented selectively in spinal cords from mutant SOD1 animals (Liu et al., 2004; Vande Velde et al., 2008). Thus, our evidence demonstrates that reduced VDAC1 conductance, and correspondingly reduced respiration rate (Lemasters and Holmuhamedov, 2006), are direct components of intracellular damage from mutant SOD1.
Moreover, not only does mutant SOD1 lower VDAC1-dependent ADP conductance by half as much as does complete VDAC1 deletion (Fig. 3S), further reduction in conductance (by VDAC1 gene inactivation) significantly accelerates disease onset (but not progression), reducing survival by more than two months for both VDAC1 heterozygous and homozygous mice. Intracellular targets for SOD1 damage beyond VDAC1 have been proposed (Ilieva et al., 2009), including aberrant glutamate handling from delayed synaptic glutamate recovery by astrocytes (Rothstein et al., 1995), mutant damage in the extracellular space following aberrant co-secretion with chromogranin (Urushitani et al., 2006), endoplasmic reticulum stress from inhibition of the ERAD pathway by mutant SOD1 binding to the integral membrane protein derlin (Nishitoh et al., 2008), and excessive production by microglia of extracellular superoxide following mutant SOD1 binding to the small G protein Rac1 and its subsequent stimulation of NAPDH oxidase (Harraz et al., 2008). Moreover, it was recently proposed that misfolded SOD1 damage to mitochondria can induce morphological changes and cytochrome c release in the presence of Bcl-2 (Pedrini et al., 2010). To those hypotheses, we proposed that the partial blockage of the VDAC1 channel by direct association with misfolded SOD1 would make motor neurons more vulnerable to any of these additional stresses derived either from mutant SOD1 acting within motor neurons, astrocytes, microglia and possibly additional neighboring non-neuronal cells. Indeed, in the presence of reduced VDAC1 conductance such pathways can play roles in pathogenesis, as we have shown that mutant SOD1-mediated disease still ensues in VDAC1 null mice.
Surprisingly, in the absence of VDAC1, we have found a 60% residual ADP conductance which seems most likely to be contributed by compensatory VDACs or VDAC-like activity(ies). Although no other VDAC isoform is known to be overexpressed in VDAC1 null mice, VDAC2 has been shown to exist in two forms that differ in conductance and selectivity (Xu et al., 1999). It is plausible that in the absence of VDAC1, VDAC2 exists predominantly in a high conductance state, as a compensatory mechanism. This mechanism should now be tested by purifying VDAC2 from VDAC1 knockout mouse, and testing its channel properties in lipid bilayers.
The compromise in mutant SOD1-mediated VDAC1 conductance that we have found offers a mechanistic explanation for alteration in mitochondrial electron transfer chain complexes and the capacity to consume oxygen and synthesize ATP previously reported in one mutant SOD1 expressing mouse line (Jung et al., 2002; Kirkinezos et al., 2005; Mattiazzi et al., 2002). The recent report that association of hSOD1G93A and hSOD1G85R with motor neuron mitochondria reduces capacity of the electron transfer chain to limit Ca2+-induced Ψm depolarization (Nguyen et al., 2009) is also fully compatible with altered adenine nucleotide transport across the outer mitochondrial membrane as the initiating deficit. So too is the report of reduced ability of mitochondria from SOD1G93A and SOD1G85R mice to survive repetitive Ca2+ addition (Damiano et al., 2006).
VDAC1 has been proposed to be the mediator for ROS release from the intermitochondrial spaces to the cytosol (Han et al., 2003; Madesh and Hajnoczky, 2001). Moreover, hexokinase (known to interact with VDAC1) has been shown in cell culture to decrease ROS release when over-expressed, thereby reducing intracellular levels of ROS (Ahmad et al., 2002; da-Silva et al., 2004). The relatively low level of hexokinase in spinal cord as compared to that in brain (Fig. 1F) may therefore be a component of selective vulnerability. This is also consistent with the selective association of misfolded mutant SOD1 with VDAC1 on the cytoplasmic face of mitochondria from spinal cord, but not liver or brain. Although both tissues accumulate high levels of mutant SOD1 (Liu et al., 2004; Vande Velde et al., 2008), prior findings show that misfolded mutant SOD1 is bound to the cytoplasmic face of spinal cord mitochondria, while apparently imported into the intermembrane space of mitochondria from cortex of the same animals and not associated with liver mitochondria at all (Vande Velde et al., 2008). Another factor likely underlying the differences in mutant SOD1 association with mitochondria, and therefore potentially factors underlying selective vulnerability, is that mitochondria from different tissues (and which retain different functional properties) have different protein compositions (Bailey et al., 2007; Mootha et al., 2003), including hexokinase levels. This is accompanied by intrinsic differences in O2−· production, lipid peroxidation, DNA oxidation and Ca2+ accumulation capacity (Sullivan et al., 2004).
Our finding that VDAC1 is one of the targets for misfolded SOD1 within the nervous system raises substantial implications for the mechanism underlying premature degeneration and death of motor neurons. A variety of apoptotic stimuli are known to trigger cell death by modulation of VDAC1 (Abu-Hamad et al., 2008; Shoshan-Barmatz et al., 2006; Tsujimoto and Shimizu, 2002; Yagoda et al., 2007; Zaid et al., 2005; Zamzami and Kroemer, 2003; Zheng et al., 2004), implicating VDAC1 as a component of the apoptotic machinery. Although VDAC1 proteins have been reported to be dispensable for Ca2+ and oxidative stress-induced permeability transition pore (PTP) opening (Baines et al., 2007), siRNA-mediated reduction in VDAC1 has supported VDAC1 as an indispensable protein for endostatin-, cisplatin- and selenite-induced oxidative stress induced PTP opening and apoptosis (Tajeddine et al., 2008; Tomasello et al., 2009; Yuan et al., 2008). Moreover, VDAC1 was recently shown to be involved in staurosporine- and ceramide-induced cell death downstream of BAD and BCL-XL (Roy et al., 2009) and curcumin induced apoptosis by cooperating with Bax in the release of AIF from mitochondria (Scharstuhl et al., 2009). Since VDAC1 is one of several targets for a cholesterol-like small molecule (TRO19622) that can protect motor neurons from SOD1 mutant-mediated death in culture and modestly delay disease onset in SOD1 mutant mice (Bordet et al., 2007), it now seems likely that its efficacy may be through direct effect on VDAC1.
Finally, it is well established that although motor neurons are the final targets in ALS, mutant damage within astrocytes and microglia contributes to driving rapid disease progression (Beers et al., 2006; Boillee et al., 2006a; Boillee et al., 2006b; Clement et al., 2003; Yamanaka et al., 2008a; Yamanaka et al., 2008b). In this context, we show here that little accumulation of misfolded SOD1 is found by disease onset, preferentially within motor neurons. However, during disease progression a dramatic increase of misfolded SOD1 is observed accumulated in other cells as well and probably extracellularly. Interestingly, mitochondrial dysfunction(s) within mutant astrocytes has been reported to cause acute motor neuron death in astrocyte-motor neuron co-cultures (Cassina et al., 2008) and astrocytes expressing mutant SOD1 have been reported to induce mitochondrial dysfunction within motor neurons (Bilsland et al., 2008). Coupling these findings with the appearance of aberrant mitochondria within motor neurons in multiple animal models of SOD1 mutant mediated ALS (Bendotti et al., 2001; Jaarsma et al., 2001; Kong and Xu, 1998; Wong et al., 1995) and the association of mutant SOD1 with mitochondria within affected tissues, we propose that misfolded SOD1 association directly with VDAC1 represents a primary event of damage within motor neurons.
Transgenic rats expressing hSOD1wt (Chan et al., 1998), hSOD1G93A (Howland et al., 2002) and hSOD1H46R (Nagai et al., 2001) were as originally described. All animal procedures were consistent with the requirements of the Animal Care and Use Committee of the University of California.
Mice heterozygous for the mutant human SOD1G37R transgene (LoxSOD1G37R) (Boillee et al., 2006b) were crossed with mice heterozygous for a VDAC1 gene disruption (Weeber et al., 2002). Mice were genotyped by PCR for the presence of the mutant SOD1 transgene (Williamson and Cleveland, 1999) and using a four-primer multiplex PCR for the presence of VDAC1 (Weeber et al., 2002), as previously described.
For survival experiments, SOD1G37R, VDAC1+/- mice were always compared with their contemporaneously produced SOD1G37R, VDAC1+/+ littermates. Time of disease onset was retrospectively determined as the time when mice reached peak body weight, early disease was defined at the time when denervation-induced muscle atrophy had produced a 10% loss of maximal weight, and end-stage was determined by paralysis so severe that the animal could not right itself within 20 seconds when placed on its side, an endpoint frequently used for SOD1 mutant mice and one that was consistent with the requirements of the Animal Care and Use Committee of the University of California.
Mitochondria were purified as previously described (Vande Velde et al., 2008). Tissues were homogenized on ice in 5 volumes of ice-cold homogenization buffer (HB) composed of 210 mM mannitol, 70 mM sucrose, 1 mM EDTA-(Tris) and 10 mM Tris-HCl (pH 7.2). Homogenates were centrifuged at 1000 × g for 10 min. Supernatants were recovered, and pellets were washed with ½ volume HB and centrifuged at 1000 × g. Supernatants were pooled and centrifuged at 12,000 × g for 15 min to yield a crude mitochondrial pellet. The supernatant was used to make cytosolic fractions by further centrifugation at 100,000 × g for 1 hour. The mitochondria were gently resuspended in HB and then adjusted to 1.204 g/ml Optiprep (iodixanol) and loaded on the bottom of a polycarbonate tube. Mitochondria were overlaid with an equal volume of 1.175 g/ml and 1.079 g/ml Optiprep and centrifuged at 50,000 × g for 4 h (SW-55; Beckman). Mitochondria were collected at the 1.079/1.175 g/ml interface and washed once to remove the Optiprep. Optiprep stock solution was diluted in 250 mM sucrose, 120 mM Tris-HCl (pH 7.4), 6 mM EDTA plus protease inhibitors.
For activity assays, spinal cords were homogenized in 5 volumes of ice-cold homogenization buffer (HB) on ice. Homogenates were centrifuged at 1000 × g for 5 min. Supernatants were recovered, and centrifuged again at 1000 × g for 5 min. Supernatants were centrifuged at 12,000 × g for 10 min to yield crude mitochondrial pellets. These mitochondria were gently resuspended in HB and then adjusted to 12 % Optiprep (iodixanol) and centrifuged at 17,000 × g for 10 min (SW-55; Beckman). The majority of the myelin (at the top of the sample) was removed and the mitochondria were washed once with HB (without EDTA) to remove the Optiprep.
Liver was homogenized in 5 volumes of ice-cold homogenization buffer (HB) on ice. Homogenates were centrifuged at 1000 × g for 5 min. Supernatants were recovered, and centrifuged again at 1000 × g for 5 min. Supernatant was centrifuged at 12,000 × g for 10 min to yield a crude mitochondrial pellet. These mitochondria were resuspended in HB (without EDTA) and centrifuged again at 12,000 × g for 10 min. The pellet was resuspended in a small volume of HB without EDTA.
Reconstitution of VDAC into a planar lipid bilayer (PLB), single channel current recording, and data analysis were carried out as previously described (Gincel et al., 2001). Briefly, PLB were prepared from soybean asolectin dissolved in n-decane (50 mg/ml). Only PLB with a resistance greater than 100 GΩ, were used. Purified protein (about 1 ng) was added to the cis chamber. After one or a few channels were inserted into the PLB, the excess protein was removed by perfusion of the cis chamber with 20 volumes of a solution to prevent further incorporation. Currents were recorded under voltage-clamp using a Bilayer Clamp BC-525B amplifier (Warner Instrument Corp.). The currents were measured with respect to the trans side of the membrane (ground). The currents were low-pass, filtered at 1 kHz and digitized on-line using a Digidata 1200 interface board and pCLAMP 6 software (Axon Instruments, Inc.). Sigma Plot 6.0 scientific software (Jandel Scientific) was used for curve fitting. All experiments were performed at room temperature.
We would like to thank Neil Cashman (University of British Columbia) for generously providing us with DSE2 antibodies, William Craigen (Baylor College of Medicine) for VDAC1 knockout mice and Larry Hayward (UMass Medical School) for wild type and mutant SOD1 baculovirus stock. This work has been supported by grants from the NIH (R37 NS27036). A.I. has been supported by EMBO Long-Term Fellowship and by a postdoctoral fellowship from IsrALS. D.W.C. receives salary support from the Ludwig Institute for Cancer Research.
Please see Supplemental Information for the following experimental procedures: Protein Purification, Immunoprecipitation, DSE2 antibodies, Immunostaining, Ca2+ and ADP accumulation by mitochondria and Immunoblotting.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.