Intermolecular transmission of superoxide dismutase 1 misfolding in living cells

Expression of Misfolded SOD1 Induces Misfolding of wtSOD1 in Human Cells.

We selected two disease-specific epitopes (DSEs) of wtSOD1 (DSE1a and DSE2) (Fig. 1A), specifically recognizing regions inaccessible to antibody binding in natively structured wtSOD1 but exposed on the molecular surface of the misfolded isoforms (21, 22). DSE1a comprises residues 145–151 [the SOD1 epitope of the dimer interface, SEDI, previously reported (23)] with cysteic acid replacing Cys146. Substitution of this sulfonic acid derivative in SEDI was based on the reasoning that sulfhydryl Cys146, exposed by Cys57–Cys146 intrachain disulfide bond reduction accompanying dimer dissociation (24), is a ready substrate for oxidative modifications. 10C12, the DSE1a mAb used in these studies, specifically binds to in vitro oxidized SOD1 and disease-associated misfolded SOD1 as determined by ELISA, IP, and IHC (22). DSE2 comprises residues 125–142 that form a segment of the SOD1 electrostatic loop, a structural element that is extruded and interacting with β-barrel elements in crystal structures of aggregated SOD1 (25). 3H1, the DSE2 mAb used in these studies, specifically binds to SOD1 that has been misfolded in vitro by denaturants and mild oxidation, and to disease-associated misfolded SOD1 by IP, IF, and IHC (21, 22).

G127X and G85R SOD1 proteins are partially misfolded, enzymatically inactive SOD1 molecular species, which do not stoichiometrically bind structure-promoting copper or zinc ions (26, 27). Moreover, the Cys146 necessary for formation of the monomer-stabilizing intrachain disulfide bond is deleted in G127X. Relative to wtSOD1, the instability of G85R has been demonstrated by its considerably reduced unfolding temperature in calorimetric studies (28). Equivalent data are unavailable for G127X, but equilibrium molecular dynamics simulation showed greater regional structural fluctuation compared with wild-type protein (Fig. S1A).

For these studies, we used an experimentally tractable cell culture system that did not overexpress human wtSOD1 (associated with spontaneous misfolding in vitro and in vivo (29, 30)]. Transient transfection-mediated expression of G127X or G85R in HEK-293FT (HEK) cells induced misfolding of endogenously expressed human wtSOD1, as observed by IF microscopy with the DSE mAb 3H1 (Fig. 1 C–F) and by IP with the DSE mAbs 3H1 and 10C12 (Fig. 2A). To ensure that these observations are not merely due to stressors associated with the transfection process, all experiments include an empty vector (EV) control in which cells are transfected with noncoding DNA. G127X expression in HEK cells permitted the unambiguous colocalization of mutant and misfolded wtSOD1, showing coincident immunoreactivity for GX-CT and DSE2-3H1 in 21% of transfected cells (2,000 GX-CT⁺ cells counted). Inspection of IF images show that within doubly positive cells immunoreactivity for G127X and misfolded wtSOD1 were not entirely congruent, and immunoreactivity for 3H1 had a punctate appearance, both of which may reflect differences in trafficking, consolidation, or degradation of the mutant and wild-type misfolded proteins. Because G85R SOD1 possesses the DSE epitopes, distinction of mutant and wtSOD1 misfolded species was not possible by IF. Molecular association of misfolded wtSOD1 with G127X was demonstrated by co-IP of mutant and misfolded wtSOD1 in nondenaturated HEK lysates from mutant-transfected HEK cells probed with GX-CT-, 10C12- or 3H1-coupled magnetic beads and detected on subsequent immunoblotting with pan-SOD1 pAbs. IP of wtSOD1 by DSE mAbs in G85R-transfected cells is consistent with similar misfolding induction or coincident IP of wtSOD1 with natively misfolded G85R protein (which possesses the DSE epitopes). Misfolded wtSOD1 was not detected in HEK cells transfected with the EV pFUW (Figs. 1F and 2A), so wtSOD1 misfolding was specifically induced by expression of the mutant SOD1 species and not due to cell stressors inherent in transfection.

The misfolded status of SOD1 isoforms in co-IP was also tested by digestion with protease K (PK; Fig. 2B). Natively structured wtSOD1 is highly resistant to PK, surviving exposure at concentrations up to 1.0 mg/mL (31), but G127X and G85R were sensitive to PK at low concentrations of <1.0 μg/mL (Fig. 2B). DSE mAb-immunoprecipitable wtSOD1 also displayed dramatic sensitivity to PK digestion at <1.0 μg/mL, consistent with enhanced PK access of the polypeptide backbone presumably due to structural loosening of the misfolded wtSOD1 isoform. The PK sensitivity of wtSOD1 in G85R transfection lysates confirmed that G85R induces misfolding of wtSOD1 and that co-IP is not merely due to physical association between natively folded wtSOD1 and G85R mutant protein. Thus, the presence of misfolded mutant SOD1 in the cytosolic compartment is associated with conformational conversion of human wtSOD1, as revealed by misfolding epitope exposure and marked protease sensitivity.

Mutant and wtSOD1 Form Nonnative Interchain Disulfide Linkages.

In natively structured wtSOD1, monomers are stabilized by the Cys57–Cys146 intrachain disulfide bond. As Cys146 is deleted in G127X, Cys57 becomes available for nonnative disulfide linkages (24). In addition, Cys6, which is at the dimer interface of wtSOD1, is solvent-exposed in this natively monomeric mutant SOD1. Cys111 is also resident at the molecular surface. Our generation of specific complementary pAb probes for GX-CT and WT-CT afforded an opportunity to unambiguously determine the participation of G127X and wtSOD1 in multimers containing nonnative disulfide bonds. GX-CT direct immunoblots of G127X-transfected HEK cell lysates under nonreducing conditions revealed massive higher-order multimers of G127X and a minimal homodimer species band (≈33 kDa; Fig. S2, IV), all converting to monomer upon β-mercaptoethanol (β-ME) reduction (Fig. S2, II). These findings, which indicate that G127X can readily form nonnative interchain disulfide bonds, are in good agreement with analyzed CNS tissue from a G127X human patient (26) and G127X and L126Z transgenic mice (32, 33). Nonreducing immunoblots also revealed a ≈35-kDa band immunoreactive with GX-CT and WT-CT, which also converted to monomers upon reduction. This is consistent with a G127X-wtSOD1 heterodimer stabilized by nonnative interchain disulfide bonds. Only trace wtSOD1 is incorporated in G127X higher-order multimers, a finding confirmed by solubility experiments that show that G127X is the major SOD1 species present in detergent-resistant aggregates (Fig. S3A). Under nonreducing conditions the monomeric SOD1 band apparently migrated as a doublet, similar to previous studies (33), resolving to a single band with reduction and suggesting that wtSOD1 can also undergo intramolecular oxidative modifications.

Higher-order multimers and a prominent dimeric species at ≈34 kDa were also observed in G85R-transfected HEK cells, resolving to monomeric G85R and wtSOD1 in a reducing system (Fig. S2); the presence of WT-CT sequence in G85R obviated resolution of wtSOD1 in heterodimeric or multimeric species. Transfection of HEK cells with G127X SOD1 in which all Cys residues were mutated to Ser (designated “C-less G127X”) revealed preserved conformational conversion of wtSOD1 (Fig. S3B), albeit at nonsignificantly lower levels than for Cys-containing G127X-transfected cells. Nonreducing immunoblots of lysates from these cells show C-less G127X almost exclusively in monomeric form, confirming that the 33-kDa, 35-kDa, and high molecular weight G127X-containing species are due to nonnative interchain disulfide linkages. We conclude that nonnative disulfide bonds are a consequence, and not a cause, of SOD1 misfolding, although they may stabilize misfolded species for co-IP.

W32 Is Involved in a Species-Specific Induction of wtSOD1 Misfolding by G127X SOD1.

Motor neuron disease in mutant human SOD1 transgenic murine models of ALS can be accelerated by cross-breeding on transgenic mice expressing human wtSOD1, in which coaggregation of mutant and human wtSOD1 is observed (34–36). Endogenous murine SOD1 in these transgenic models seems to be essentially inert to aggregation (26, 36), although minor protective effects of mouse SOD1 have been observed (36). The epitopes for DSE1a and DSE2 are identical between mouse and human and can be exposed by oxidation-induced misfolding (Figs. S4A and S5). However, DSE immunoreactivity was not observed in transfected murine N2a neuroblastoma cells expressing abundant G127X protein (Fig. 3 A and B). Likewise, DSE mAb-immunoprecipitation was also not observed in lysates of N2a cells (Fig. S4B) compared with that of lysates from human HEK cells. Human restriction of G127X-induced wtSOD1 misfolding was confirmed in three human cell lines (HEK, HeLa, and SH-SY5Y) and three mouse cell lines (N2a, Min6, and B16; Fig. S4B). Notably, HEK and HeLa cells are mesenchymal, whereas SH-SY5Y cells are derived from a human neuroblastoma, suggesting that any cell type expressing human SOD1 is susceptible to mutant-induced misfolding.

Further inspection of the sequence revealed a strikingly nonconservative substitution of tryptophan (W) at position 32 in human SOD1 by serine in mouse SOD1. W32 is the only tryptophan in the human protein and has previously been identified as a site of oxidative modification and a potentiator of aggregation (37). Moreover, W32 is highly solvent exposed, ranking in the 89th percentile for solvent exposure among nonredundant tryptophans in the Protein Data Bank. G127X and G85R constructs with a W32S substitution had markedly reduced ability to convert wtSOD1 compared with these mutant proteins (Fig. 3 C and D), suggesting that W32 directly participates in the wtSOD1 conformational conversion process.

Misfolded wtSOD1 Is Sustained After Induction by Misfolding-Competent SOD1 Molecular Species.

To understand whether induction of wtSOD1 misfolding was dependent on the continued presence of mutant SOD1, HEK cells were transfected with G127X and monitored over time for levels of wtSOD1 misfolding. As a negative control, EV transfection and the extended duration of cell culture was not associated with any measurable wtSOD1 misfolding (Fig. S6). Maximal abundance of G127X protein was observed at 48 h, after which progressive depletion results in an undetectable level by day 6 (Fig. 4A). Transfection with G127X/W32S displayed a protein abundance time course similar to that of G127X (Fig. 4B). However, G127X expression induced wtSOD1 immunoreactivity with mAbs 3H1 and 12C12, which was detectable at day 2 and persisted to day 10, whereas expression of the G127X/W32S was not associated with detectable wtSOD1 misfolding at any point in the time course (Fig. 4 A and B), consistent with the W32-restricted wtSOD1 conformational conversion being directly mediated by the mutant species, and not due to nonspecific effects such as protein overexpression or chaperone titration. Moreover, the sustained high level of misfolded wtSOD1 at day 10 at least six half-lives (38) after peak G127X expression, and despite increased protease sensitivity of the misfolded species as determined in Fig. 2, suggests the possibility that misfolded wtSOD1 may propagate independently of mutant G127X. We further investigated this notion by overexpressing human wtSOD1, which has been noted in prior studies to be associated with a proportion of misfolded molecules (29, 30). In contrast to G127X, misfolded wtSOD1 possesses the C-terminal misfolding-specific epitopes recognized by 3H1 and 10C12, regardless of whether the protein originates from the transfection construct or endogenous wtSOD1 that has been induced to misfold. To effectively disentangle putative “converting” from “converted” species of wtSOD1, we compared time course concentrations of misfolded SOD1 in HEK cells transfected with wtSOD1 and wtSOD1/W32S constructs. Immunoreactivity with 3H1 and 10C12 peaked at 2 d for both constructs (Fig. 4C), as with G127X. However, at day 10 in the time course, wtSOD1/W32S-transfected cell lysates show a statistically significant decline in 3H1 and 10C12 immunoreactivity compared with wtSOD1 transfection (P < 0.0001; Fig. 4 C and D), consistent with the incompetence of the W32S species to induce a sustained wtSOD1 misfolding propagation.

G127X Induces wtSOD1 Misfolding in a Recombinant Cell-Free System.

Induced misfolding of wtSOD1 prompted us to investigate the role of the intracellular milieu in the process, including non-SOD1 macromolecules. Purified recombinant wtSOD1 (Fig. S7) was incubated with and without preaggregated recombinant G127X for 24 h with agitation at 37 °C in Hepes-buffered saline in the presence or absence of DTT (50 mM) and EDTA (5 mM) to simulate the reducing and extensively metal cation-buffered intracellular environment. Misfolded wtSOD1 was measured in a Biacore surface plasmon resonance assay by capture with 3H1 mAb. G127X significantly potentiated wtSOD1 misfolding in the solution containing DTT and EDTA (Fig. 5A), although a limited background of spontaneous misfolding of wtSOD1 was observed under these conditions. Conversely, in the solution without DTT or EDTA, there was no induction of wtSOD1 misfolding either in the presence or absence of G127X (Fig. 5B). The misfolded SOD1 generated in the cell-free system was sensitive to degradation by PK (Fig. 5D), as also observed for misfolded wtSOD1 in cell culture (Fig. 2B). Coincubation with purified G127X/W32S instead of G127X induced considerably less wtSOD1 misfolding (Fig. 5E), confirming the protective effect of the W32S substitution. These results suggest that wtSOD1 as a metal replete dimer can still intermittently expose its intrachain disulfide bond for reduction and “trapping” in a partially misfolded state recognized by the 3H1 electrostatic loop mAb, and also indicates that natively misfolded G127X protein can facilitate this process. Reduction of the monomer disulfide bond has been previously shown to precipitate further structural disorganization in wtSOD1 and increase its propensity for aggregation (24). Moreover, no other macromolecule seems to be essential for the SOD1 conformational conversion reaction, consistent with direct physical interaction between isoforms. Although the solution conditions used in these experiments approximate the reducing and metal cation-buffered intracellular milieu, we cannot be sure that they recapitulate the cytosolic environment exactly.

Immunoprecipitation.

Refer to SI Materials and Methods for detailed methods. Briefly, transfected cells were lysed, and 100 μL of cell lysate was mixed with 10 μL of antibody-coupled M-280 Tosyl-activated magnetic Dynabeads, then incubated for 3 h at room temperature with constant rotation. Beads were washed three times and boiled in SDS sample buffer containing 1% β-mercaptoethanol for 5 min. One microliter of lysate was added directly into SDS sample buffer, boiled, and used as a pre-IP control. The generation of mouse monoclonal DSE antibodies used in IP experiments is described elsewhere (21, 22).

Protease Analyses of Misfolded SOD1.

HEK cells were transiently transfected with G127X- or G85R-SOD1 for 48 h. Cells were lysed without protease inhibitors, and 400-μL aliquots from each lysate were digested with a concentration series of proteinase K for 30 min at 37 °C. Digests were terminated by the addition of protease inhibitor mixture and phenylmethylsulfonyl fluoride to a final concentration of 5 mM.

Statistical Analyses.

At least five independent experiments were subjected to statistical analysis. The nonparametric Mann-Whitney test was used to determine differences between IP experiment quantitation in Figs. 2A and 5 and Figs. S2 and S3, comparing the pairs of independent samples. In experiments involving multiple groups (more than two cell lines, species, groupings, etc., in Figs. 4B and 6) the nonparametric Kruskal-Wallis test was used. The Bonferroni correction was applied to multiple comparisons; significance was set at P < 0.05 in all cases and indicated by asterisks (*). Statistical analyses were performed using SPSS 17.0 and XLSTAT 2008.