Since our publication (Bartels et al., 2011), we have conducted extensive experiments using both Clear Native (CN) PAGE and in-vivo crosslinking of endogenous cellular α-synuclein. Regarding the former method, the migration of endogenous human α-synuclein in lysates of untransfected M17D neuroblastoma cells occurs at ~66 kDa (we cannot be certain from just CN-PAGE that it represents a tetramer), whereas bacterial recombinant (unfolded) α-synuclein monomer always migrates at ~50 kDa. We have observed this migration difference of the two species on CN gels in many experiments since publication (an example had already been shown in Supp. Fig 5A of Bartels et al., 2011). This consistently lower migration of bacterial than cellular α-synuclein in CN gels was also visible (but not commented upon) in Fig. 4C of Fauvet et al. (compare lanes 1-3 to lanes 5-14), although the difference appears less pronounced in their CN gel system. We have further observed that a low level of overexpression of α-synuclein in the M17D cells specifically augments the endogenous ~66 kDa band (confirming that...  Read more

Since our publication (Bartels et al., 2011), we have conducted extensive experiments using both Clear Native (CN) PAGE and in-vivo crosslinking of endogenous cellular α-synuclein. Regarding the former method, the migration of endogenous human α-synuclein in lysates of untransfected M17D neuroblastoma cells occurs at ~66 kDa (we cannot be certain from just CN-PAGE that it represents a tetramer), whereas bacterial recombinant (unfolded) α-synuclein monomer always migrates at ~50 kDa. We have observed this migration difference of the two species on CN gels in many experiments since publication (an example had already been shown in Supp. Fig 5A of Bartels et al., 2011). This consistently lower migration of bacterial than cellular α-synuclein in CN gels was also visible (but not commented upon) in Fig. 4C of Fauvet et al. (compare lanes 1-3 to lanes 5-14), although the difference appears less pronounced in their CN gel system. We have further observed that a low level of overexpression of α-synuclein in the M17D cells specifically augments the endogenous ~66 kDa band (confirming that it is α-synuclein), whereas higher overexpression leads to the additional appearance of the ~50 kDa band, which co-migrates with bacterial α-synuclein. The latter result is consistent with the co-migration of overexpressed α-synuclein in HEK cell transfectants with bacterial α-synuclein seen by Fauvet et al. (their Fig. 5C). Collectively, these results raise the possibility (as yet unproven) that in some cell types which express low levels of α-synuclein (e.g., M17D), high overexpression leads to α-synuclein that is not folded and thus co-migrates with the unfolded bacterial protein. In addition, we have recently observed some cell types expressing high levels of endogenous α-synuclein in which this protein runs at ~50 kDa, co-migrating with bacterial α-synuclein. The reason for this variability in CN migration among endogenous α-synuclein species is unclear and being investigated. We thus agree with the conclusion of Fauvet et al. that CN-PAGE gels “do not provide an accurate estimate of the size and quaternary structure distribution of α-synuclein.”

Accordingly, we believe that CN-PAGE, despite its convenience and widespread availability, cannot be used to establish the native assembly state of α-synuclein. This is a statement that had already been made in our paper (Bartels et al., 2011). Even though we initially used native PAGE as an indicator of a higher α-synuclein assembly state, the central conclusion of our publication was based on three key approaches: STEM and analytical ultracentrifugation (AUC) analyses, and in-vivo crosslinking in cells expressing only endogenous α-synuclein.

Regarding the crosslinking approach, although we used a DSS method in our paper (and Fauvet et al. followed this), extensive subsequent efforts to optimize the efficiency of α-synuclein crosslinking in vivo led us to try several different crosslinking reagents and protocols. From this work, we have found that the cell-permeant crosslinker, DSG (7.7 Angstrom linker), applied at 0.5 or 1.0 mM for 30 minutes at room temperature, results in the trapping of a major endogenous α-synuclein species migrating at ~60 kDa on SDS-PAGE (consistent with the size of a tetramer), accompanied by the partial depletion of the ~14 kDa monomer band. Importantly, simultaneous application of this new protocol to crosslink the known dimeric protein DJ-1 similarly yielded some SDS-stable dimers and a reduction in monomers, although, again, this was not 100 percent efficient (as expected with in-vivo crosslinking protocols). Such control crosslinking of a known endogenous oligomeric protein was not performed by Fauvet et al., and this precludes an estimate of the efficiency of their crosslinking experiments. In any event, we now recommend the use of DSG at 1 mM for more efficient trapping of α-synuclein oligomers, and in our hands, this method has now consistently yielded a ~60 kDa α-synuclein band as the major in-vivo species.

Nevertheless, Fauvet et al. report that in DSS-treated HEK cell transfectants, they detected dimers as the main oligomeric species, but also a ~58 kDa (tetramer or trimer) band, accompanied by some decrease in the monomer (see their Fig. 7A, right half). They then discuss published reports that α-synuclein dimers can occur under various stress-linked conditions (e.g., oxidative stress), but such conditions are not relevant to their crosslinking experiments in healthy cells, or to ours. They interpret their in-vivo crosslinking of apparent dimers, tetramers, and “high molecular weight” oligomers (Fig. 7A) as potentially “reflecting a crosslinking artifact, resulting in the observation of diffusion-limited crosslinking reactions of originally separate monomers.” However, they then acknowledge that “alternatively, our data may also reflect the fact that the dimer could be the prevalent α-synuclein species in cells, suggesting it could serve as the basic building block for a very complex further assembly mechanism.” Thus, the interpretation of their successful in-vivo crosslinking of oligomers in at least their overexpressing HEK cells (Fig. 7A) remains unclear.

Fauvet et al. did not detect oligomers when applying DSS to endogenous α-synuclein in untransfected HEK cells (although faint bands at dimer and trimer/tetramer positions are apparent in Fig. 7B, lane 5), but they also did not show the monomer level before DSS. Thus, the efficiency of crosslinking under their experimental conditions was not established. Our new in-vivo crosslinking data (using DSG) since publication lead us to conclude that we can consistently trap endogenous, tetramer-sized (~60 kDa) oligomers in native (non-transfected) neural and non-neural cells. We believe that until the above-discussed variability in the CN gel migration of α-synuclein from different cell types and in different lysis conditions can be resolved, in-vivo crosslinking is a more reliable method to assess the oligomeric state of α-synuclein, and our improved crosslinking method produces evidence of the existence of tetramers and some dimers and trimers in intact cells, just as we had originally reported (see Bartels et al., Fig. 1C, right panel).

As regards the data of Fauvet et al. on purified red blood cell (RBC) α-synuclein, our results disagree with theirs. With our purification protocol, we have repeatedly obtained evidence of an α-helix-rich oligomeric form as the principal α-synuclein species that can be purified from RBC and other mammalian cell types, and have not observed any significant degree of random coil structure. On the other hand, we always do see random coil structure in our bacterially expressed α-synuclein, just as Fauvet et al. do. Upon request, we offered help to Fauvet et al. for their initial attempts to reproduce our purification protocol after our publication, but the authors neither reported any problems they had with our protocol nor informed us about it in their paper this week. It is possible that their additional column purification step that exposes the sample to 10 mM imidazole could partially denature native α-synuclein and contribute to the differences in α-synuclein secondary structure that the two labs report; we will now address this question.

Indeed, we have found in past experiments that the addition of small amounts of certain organic solvents can denature our helically folded α-synuclein preparations. Until further investigation of their new RBC purification protocol, we cannot say what leads to the striking difference in results. It is interesting to note that their CD spectra of purified RBC α-synuclein appear to display a somewhat more secondary structure than that of their bacterial samples (compare Fig. 8F to Fig. 1C). Whether this is due to some helical folding (i.e., the remains of a natively folded α-synuclein) or β-sheet folding (i.e., aggregated protein) should be investigated by deconvolution of the spectra. We acknowledge that our purification method can be challenging, as we had specifically addressed in the troubleshooting section of our detailed published protocol. Experiments to improve yield and consistency have been an intense focus of our lab in the past months. On the other hand, we always do see random coil structure in our bacterially expressed α-synuclein, just as Fauvet et al. do.

Importantly, Fauvet et al. also state that “in a blood sample from a different donor that was purified using the protocol developed by Bartels et. al., we observed a heat and SDS-stable α-synuclein-positive band migrating at ~46 kDa…. Using Western blotting, oligomer specific ELISA, immunoprecipitation, and mass spectrometry, we confirmed that this additional band corresponds to an oligomeric form of α-synuclein.” While they are correct that we had not reported in our paper the detection of SDS-resistant α-synuclein oligomers in our denaturing gels, our subsequent experiments on RBC α-synuclein have shown some level of SDS-stable dimers, trimers, and tetramers that are recognized by several different α-synuclein antibodies. We speculate that a fraction of endogenous oligomeric α-synuclein in RBC (and perhaps in other cell types) has relative resistance to conventional SDS denaturation, while most oligomers may be depolymerized to monomers by heating in SDS.

Finally, Fauvet et al. state they were unable to purify sufficient α-synuclein to perform sedimentation equilibrium-analytical ultracentrifugation (AUC) analysis. All four reviewers of our paper had strongly recommended this technique as providing (if it could be accomplished) the most unambiguous approach for establishing the mass (and thus assembly state) of endogenous α-synuclein purified from cells under non-denaturing conditions. The mass of ~57,800 that we obtained is very close to that of four N-acetylated monomers (as we showed the cellular α-synuclein monomer to be). Our AUC was complemented by another non-denaturing method to estimate the mass of a protein complex, STEM analysis. We performed STEM on α-synuclein purified under non-denaturing conditions from both RBC and human neuroblastoma cells (in each case, analyzed for α-helical structure by CD spectroscopy and confirmed to lack significant contamination by Coomassie staining and mass spectrometry) and obtained an approximate mean mass of the purified protein of ~55 kDa. These two key methods (or other unbiased instrumental analysis) have not yet been pursued by Fauvet et al., making the claimed monomeric state of their native α-synuclein samples not certain. The negative ELISA data the authors present (Fig. 6) could be a matter of differences in the antibody 211 epitope exposure of natively folded α-synuclein and the in-vitro aggregated bacterial α-synuclein on which this assay is based. Also, we are not clear why the authors chose 211 as the capture/detection antibody in this ELISA, given its poor reported sensitivity (Mollenhauer et al., 2008).

In conclusion, perhaps Fauvet et al. and we can agree that the assembly state of α-synuclein in cells may be complex and dynamic, and it is possible—even likely—that α-synuclein can interconvert into more than one assembly state. The preponderance of our data then and now suggests to us that α-synuclein occurs in cells in oligomeric states, principally tetramers, that are different than the unfolded recombinant monomers routinely obtained from bacteria. The widely confirmed observation (including by Fauvet et al. and us) that unfolded α-synuclein achieves an α-helical structure when it binds small lipid vesicles is, in our view, unlikely to be an irrelevant in-vitro finding, but rather likely to occur as such in the crowded environment of the cytoplasm. The dynamic nature of α-synuclein and the potential heterogeneity of its assembly forms among cells means that substantial further analyses will be needed before its native states are clearly defined. We continue to believe that these native states are not solely identical to that of the bacterially expressed monomer, as Fauvet et al. claim.

References:
Bartels T, Choi JG, Selkoe DJ. α-Synuclein occurs physiologically as a helically folded tetramer that resists aggregation. Nature. 2011 Sep 1;477(7362):107-10. Abstract

View all comments by Tim Bartels
View all comments by Ulf Dettmer
View all comments by Eric Luth
View all comments by Dennis Selkoe

Our experience working with α-synuclein is similar to Lashuel and colleagues’ in that we always recover α-synuclein as monomers. We have never observed tetramers using similar biochemical and biophysical techniques in the analyses of recombinant α-synuclein purified from E. coli, from cultured primary neurons from rat and mouse, from non-neuronal cells overexpressing the protein, from control or from transgenic mice overexpressing wild-type or mutant α-synuclein, from human Parkinson’s disease tissue, or from brain tissue affected by other synucleinopathies.

Bartels et al. claimed that α-synuclein tetramers underwent little or no amyloid-like aggregation in vitro as opposed to the recombinantly expressed monomers that aggregate readily into β-pleated-rich amyloid fibrils. Based on this observation, they speculated that destabilization of the helically folded tetramer precedes α-synuclein misfolding and aggregation in synucleinopathies. However, using primary neurons generated from non-transgenic wild-type mice, a more physiological system, we recently showed that...  Read more

Our experience working with α-synuclein is similar to Lashuel and colleagues’ in that we always recover α-synuclein as monomers. We have never observed tetramers using similar biochemical and biophysical techniques in the analyses of recombinant α-synuclein purified from E. coli, from cultured primary neurons from rat and mouse, from non-neuronal cells overexpressing the protein, from control or from transgenic mice overexpressing wild-type or mutant α-synuclein, from human Parkinson’s disease tissue, or from brain tissue affected by other synucleinopathies.

Bartels et al. claimed that α-synuclein tetramers underwent little or no amyloid-like aggregation in vitro as opposed to the recombinantly expressed monomers that aggregate readily into β-pleated-rich amyloid fibrils. Based on this observation, they speculated that destabilization of the helically folded tetramer precedes α-synuclein misfolding and aggregation in synucleinopathies. However, using primary neurons generated from non-transgenic wild-type mice, a more physiological system, we recently showed that endogenous α-synuclein is readily recruited by small amounts of synthetic α-synuclein preformed fibrils into Lewy bodies and Lewy neurites (Volpicelli-Daley et al., 2011). Pathological insoluble α-synuclein species were detected as monomers, dimers, and other higher oligomeric species. Thus, irrespective of whether α-synuclein exists as monomers or tetramers, the endogenous protein can be readily recruited to form α-synuclein pathology in primary neurons.

View all comments by Virginia Lee

The report from the collaborative Lashuel group is a careful but partial rebuttal of two recent reports (Bartels et al., 2011 and Wang et al., 2011) that concluded that α-synuclein in cells exists as a stable tetramer. In the current report, Fauvet et al. carefully reconfirm what was known for a while by α-synuclein aficionados. That, regardless of the source, α-synuclein does not behave like a globular protein and resolves as something that appears much larger than a monomer. Thus, Fauvet’s study re-establishes that native gel electrophoresis or chromatography are insufficient indicators of tertiary structures of α-synuclein. This is important because some studies mistakenly conclude that α-synuclein behaves like a globular protein.

Looking at these three reports, it is tempting to conclude that only one of the views is correct. However, resolution of the issues will likely require additional studies. For example, while the studies with the A140C mutant suggest that α-synuclein...  Read more

The report from the collaborative Lashuel group is a careful but partial rebuttal of two recent reports (Bartels et al., 2011 and Wang et al., 2011) that concluded that α-synuclein in cells exists as a stable tetramer. In the current report, Fauvet et al. carefully reconfirm what was known for a while by α-synuclein aficionados. That, regardless of the source, α-synuclein does not behave like a globular protein and resolves as something that appears much larger than a monomer. Thus, Fauvet’s study re-establishes that native gel electrophoresis or chromatography are insufficient indicators of tertiary structures of α-synuclein. This is important because some studies mistakenly conclude that α-synuclein behaves like a globular protein.

Looking at these three reports, it is tempting to conclude that only one of the views is correct. However, resolution of the issues will likely require additional studies. For example, while the studies with the A140C mutant suggest that α-synuclein dimers retain the aberrant mobility characteristics (see Fig. 1 in Fauvet et al., 2012), it is possible that tetramers of α-synuclein could behave as a globular protein and migrate with the monomer on native gels and size exclusion chromatography. In this regard, Fauvet does not fully negate previous studies from Selkoe’s group (Bartels et al., 2011) and Petsko’s group (Wang et al., 2011). This is because both Bartels et al. and Wang et al. used an additional sophisticated approach, such as scanning transmission electron microscopy (STEM) or electron microscopy (EM) to support the existence of “tetramers” in selected α-synuclein preps. However, one is struck by the fact that the heterogeneity seen with STEM/EM analysis is not sufficiently reflected in the native gel or chromatographic studies. Specifically, both STEM (Bartels et al.) and EM (Wang et al.) studies show, in addition to tetramers, significant presence of dimers, trimers, and higher-order oligomers; while only one peak is seen with native gel and chromatographic studies. Fauvet’s report does not account for these observations, either, and we are left to wonder whether some portion of α-synuclein does exist as not only tetramers, but also as dimers and trimers.

All in all, the details of what α-synuclein looks like in cells may be just an interesting “biophysical” puzzle, except that Bartels et al. and Wang et al. also show that the “native” α-synuclein preps (ones retaining tetramers) do not form fibrils. This point is also not addressed by Fauvet et al., but further studies of this issue may have the most significant impact on understanding the pathogenesis of α-synucleinopathies. A clue to resolving this issue could be related to the previous studies showing that some small molecules, such as baicalein, can stabilize oligomers of α-synuclein that are unable to form fibrils. Possibly, such small molecules could be made by the cell.

Finally, all of these studies examine α-synuclein that is extracted from cells. Thus, it is difficult to know what is the native state of α-synuclein in intact cells. In fact, Wang et al. hypothesize that the recovery of the “tetramers” could be α-synuclein concentration dependent. This would imply that α-synuclein may not be in “tetramers” in some parts of the neuron.

View all comments by Michael K. Lee

We shouldn’t lose sight of the most important aspect of this controversy, which is that α-synuclein can, and does, take on different forms in response to the local environment, and likely occupies multiple states in vivo. In other words, it is a moving target, and to try and categorize any one form as “the structure” is a mistake. We know that synuclein associates with cell membranes, and there is recent published work that suggests it may be involved in shepherding SNARE complex formation. (By the way, the crystallized SNARE complex is a heterotetrameric parallel four-helix bundle, similar to what we have proposed for tetrameric α-synuclein.) I found it interesting that Lashuel and coworkers found higher-order (dimer, trimer-tetramer) α-synuclein crosslinking in the crowded and very heterogeneous environment of the cell. Even from their results, it is clear that α-synuclein self-associates. I suspect that the tetrameric form may represent a stable non-toxic “storage” of soluble α-synuclein at high concentrations, as one might find in neurons or erythrocytes. What might be even...  Read more

We shouldn’t lose sight of the most important aspect of this controversy, which is that α-synuclein can, and does, take on different forms in response to the local environment, and likely occupies multiple states in vivo. In other words, it is a moving target, and to try and categorize any one form as “the structure” is a mistake. We know that synuclein associates with cell membranes, and there is recent published work that suggests it may be involved in shepherding SNARE complex formation. (By the way, the crystallized SNARE complex is a heterotetrameric parallel four-helix bundle, similar to what we have proposed for tetrameric α-synuclein.) I found it interesting that Lashuel and coworkers found higher-order (dimer, trimer-tetramer) α-synuclein crosslinking in the crowded and very heterogeneous environment of the cell. Even from their results, it is clear that α-synuclein self-associates. I suspect that the tetrameric form may represent a stable non-toxic “storage” of soluble α-synuclein at high concentrations, as one might find in neurons or erythrocytes. What might be even more interesting would be a thorough mass spectral analysis of any hetero-crosslinked species that might occur in vivo. What proteins (or other molecules) are closely associated with α-synuclein? This could be an important step in understanding the function of this protein, which is still the major issue to be resolved.

View all comments by Thomas Pochapsky

α-Synuclein Tetramers: The New Kid in Town or a Ghost?
α-Synuclein is critically involved in Parkinson's disease, where it accumulates in an aggregated state in Lewy body inclusions of degenerating neurons. The direct involvement of α-synuclein in cellular demise has been highlighted by the autosomal dominant phenotype caused by rare missense mutations and gene multiplications in the α-synuclein gene, SNCA, in some families. α-Synuclein aggregates over time in solution under physiological conditions, and the process exhibits concentration-dependent characteristics. In cytoplasmic inclusions, α-synuclein is heavily phosphorylated on Ser129, and this modification has been suggested to induce the cytotoxic phenotype. Accordingly, therapeutic strategies have been focused on 1) the rational reduction of SNCA expression; 2) inhibition of α-synuclein aggregation, and 3) inhibition of the Ser129 directed kinase.

Lansbury’s group demonstrated in 1996 that recombinant human α-synuclein was natively unfolded, but able to acquire α-helical structure upon incubation with...  Read more

α-Synuclein Tetramers: The New Kid in Town or a Ghost?
α-Synuclein is critically involved in Parkinson's disease, where it accumulates in an aggregated state in Lewy body inclusions of degenerating neurons. The direct involvement of α-synuclein in cellular demise has been highlighted by the autosomal dominant phenotype caused by rare missense mutations and gene multiplications in the α-synuclein gene, SNCA, in some families. α-Synuclein aggregates over time in solution under physiological conditions, and the process exhibits concentration-dependent characteristics. In cytoplasmic inclusions, α-synuclein is heavily phosphorylated on Ser129, and this modification has been suggested to induce the cytotoxic phenotype. Accordingly, therapeutic strategies have been focused on 1) the rational reduction of SNCA expression; 2) inhibition of α-synuclein aggregation, and 3) inhibition of the Ser129 directed kinase.

Lansbury’s group demonstrated in 1996 that recombinant human α-synuclein was natively unfolded, but able to acquire α-helical structure upon incubation with SDS-micelles or apolar solvents. This was demonstrated by gel filtration and CD spectroscopy. Ever since, it has been generally accepted—and reproduced in countless laboratories—that α-synuclein is an unfolded monomeric protein under physiological conditions, though the monomer is also able to acquire significant levels of α-helical structure in its approximately 100 N-terminal residues upon interaction with synthetic acidic phospholipid vesicles or anionic surfactant micelles.

Two papers challenged this view in 2011, and suggested that the native form of α-synuclein is a folded tetramer, which is unable to aggregate. Such an observation would have significant implications: Rather than focusing on aggregation, expression, or phosphorylation, a straightforward and novel therapeutic strategy could be to stabilize the native tetramer. Such an approach has already been achieved for transthyretin using the small molecule Tafamidis.

First, Dr. Selkoe’s group in Boston (Bartels et al., 2011) isolated α-synuclein from erythrocytes, and suggested that it was present as a tetramer. Their data relied to a large extent on native gel electrophoresis. However, this technique is not optimal because unfolded recombinant α-synuclein, despite being approximately 14.5 kDa, migrates abnormally due to its lack of persistent structure. This makes it difficult to distinguish a compact tetramer from an extended monomer on hydrodynamic radius alone. More significantly, Bartels et al. presented circular dichroism data that showed an α-helical signature indistinguishable from recombinant α-synuclein supplemented with liposomes. The authors took pains (using phosphate analysis and addition of surfactant to strip off lipids) to ensure that all lipids were removed from the α-synuclein preparation prior to the CD spectroscopic analysis. Importantly, the “native” α-synuclein was unable to aggregate. Although the number of methodological analyses were limited because of the relatively low yield of α-synuclein from erythrocytes, the genie was out of the bottle.

Secondly, a group headed by Dr. Hoang from Indianapolis purified α-helix-rich α-synuclein from E. coli with properties similar to the species purified from erythrocytes. The purification protocol for the expressed human GST-α-synuclein fusion protein was gentle because the binding of the GST moiety to the affinity column could be broken by reduction. Subsequently, the GST part was removed by gentle and specific proteolysis followed by a final gel filtration step. The resulting α-synuclein peptide had a 10-amino-acid N-terminal extension remaining from the GST fusion protein. Although this is a relatively minor change, it is not clear whether this extension can affect the association properties of α-synuclein; removal of the C-terminal 30-40 residues makes (bacterially expressed) α-synuclein aggregate more readily, and is more toxic than the full 140-residue protein in vivo. The high yield of α-synuclein from E. coli allowed for a series of analyses comprising gel filtration, chemical cross-linking, negative stain electron microscopy, circular dichroism spectroscopy, NMR spectroscopy, and aggregation assays. The Hoang group’s analyses indicated an α-helical-rich structure compatible with a tetramer, and corroborated the lack of aggregation behavior of this species. Surprisingly, α-synuclein with the disease-causing mutations A30P, E46K, or A53T all formed less compactly folded structures and demonstrated more highly aggregating behavior. The interpretation was that the lowered stability of the mutant tetrameric species led to higher concentrations of monomers, which could subsequently aggregate.

The paper made great efforts to explain why former analyses had missed this folded species. However, their statement, “The unfolded form can be induced by heating, chemical treatment, or dilution, and our preliminary data also suggest that too high a concentration of α-synuclein appears to favor species with less helical content that, over time, aggregate into amyloid fibrils,” suggests it may take a good deal of luck to purify, store, and analyze this species.

In response to these observations, this thorough study has just been published that critically tries to reproduce the data published in the above papers. The study, headed by Dr. Lashuel, represents the combined work of several laboratories with a long record in studying α-synuclein, including Drs. Masliah and El-Agnaf. They both study recombinant α-synuclein purified from GST-α-synuclein as presented in the Wang paper, and erythrocyte-derived human α-synuclein by methods similar to the Bartels paper. The recombinant protein was analyzed by native gel electrophoresis, gel filtration combined with dynamic light scattering, NMR spectroscopy, and CD spectroscopy. The authors make a systematic effort to introduce positive controls for their analyses. These include a recombinant α-synuclein dimer (based on an engineered disulfide bridge) and an N-acetylated species (this species is abundant in eukaryotic samples). Based on these techniques, they are unable to identify any species compatible with an α-helical tetramer, but only an unfolded monomer, despite using constructs similar to those of the Hoang group. Moreover, α-synuclein extracted from different types of human tissues (human control, Alzheimer's disease, dementia with Lewy body brain tissue, and tissue from wild-type, α-synuclein transgenic and knockout mice) all migrate like unfolded α-synuclein on native gels. On cells expressing human α-synuclein, they use an ELISA technique, with the same antibody used for capture and detection, to test for the tetramer. This test did not give any indication of polymers, including tetramers, in the extract. Chemical crosslinking of the cells with the short length crosslinker DSS followed by SDS-PAGE reveals a polymeric pattern that is not incompatible with a higher-ordered polymer because the majority of the α-synuclein immunogen migrates as a dimer, although higher-order species also are detected. Chemical crosslinking reactions are not always 100 percent efficient, and a tetramer may just appear on SDS-PAGE as a stabilized dimer. The authors did not crosslink their monomeric standard, which could have provided evidence of the significance of these bands. Similarly, data were to a large extent obtained with the α-synuclein purified from erythrocytes.

A few differences stand out when looking into the two papers on recombinant α-synuclein. The Hoang group uses a hypertonic buffer (150 mM NaCl + 100 mM Hepes) pH 7.4 but added 10 percent glycerol and 0.1 percent of the detergent n-octyl-β-glucopyranoside. They test that the detergent does not affect the behavior of the tetramer, but a putative effect of the 10 percent glycerol is not tested. By contrast Fauvet et al. perform bacteria lysis in a hypotonic buffer at slightly basic pH 8.3, but their lysis of eukaryotic cells was done by hypotonic lysis at physiological pH, and the Fauvet paper specifically tested the protocol by Bartels. However, such differences are small and the experimental approaches appear appropriate.

What can we conclude based on the studies? All labs involved have an impressive track record in conducting thorough peer-approved work, but their conclusions are essentially incompatible. One may say that time will tell, but the question may be too important to wait, especially because this may govern how large sums are going to be spent on neuroprotective strategies in the coming years. One approach could be for a funding agency to fund a collaborative project among the groups involved that could compare the α-synuclein purified in the different labs, e.g., by exchange of postdocs who can perform the purification in the collaborator’s lab. This will be a way to rapidly shed some light on this tetramer species. Like the Aβ field, the α-synuclein field is awash with species at different levels of aggregation and association (from monomer through dimer to various levels of protofibrils/oligomers and fibrils proper). Clarifying this issue is an urgent task—and a fascinating glimpse into the many different states that these freaky proteins can apparently assume.

References:
Bartels T, Choi JG, Selkoe DJ. α-Synuclein occurs physiologically as a helically folded tetramer that resists aggregation. Nature. 2011 Sep 1;477(7362):107-10. Abstract

View all comments by Poul Henning Jensen

The study of Bartels et al. on α-synuclein is very interesting and provocative, and has the potential to further the understanding of the functional role of α-synuclein. However, given that the study was performed primarily with α-synuclein derived from red blood cells (RBCs), it will be important to investigate if the behavior of α-synuclein is similar in the central nervous system, where the lipid content in the plasma membranes is rather different. It is worth noting that utilizing native gels, the monomer of α-synuclein can be readily identified in human brains. The behavior of α-synuclein in RBCs might be unique to the peripheral compartments. It will be also important to determine if other investigators are able to reproduce such results in RBCs versus the CNS utilizing the scanning transmission electron microscopy approach.

View all comments by Eliezer Masliah

These are interesting observations presented by Bartels et al. The data suggest that equilibrium may exist between a tetrameric aggregation-incompetent α-synuclein species, and an aggregation-competent monomer. Shifting the equilibrium to the monomeric form would be expected to promote the formation of amyloid and Lewy bodies. Consistent with their notion that certain oligomeric forms of α-synuclein resist aggregation, we have also previously identified aggregation-incompetent, non-toxic assemblies in certain "pathology-free" regions of the A53T transgenic mouse brain (Tsika et al., 2010). The size of the α-synuclein assemblies documented in that study ranged from ~225 to 50 kDa. This size estimate was obtained by size exclusion chromatography, and is an interpretation based on the elution profile of globular protein standards. It is also important to note that we were likely detecting the most stable species under the specific conditions used for brain extracts (100,000 x g centrifugation of protein soluble in 1 percent Triton...  Read more

These are interesting observations presented by Bartels et al. The data suggest that equilibrium may exist between a tetrameric aggregation-incompetent α-synuclein species, and an aggregation-competent monomer. Shifting the equilibrium to the monomeric form would be expected to promote the formation of amyloid and Lewy bodies. Consistent with their notion that certain oligomeric forms of α-synuclein resist aggregation, we have also previously identified aggregation-incompetent, non-toxic assemblies in certain "pathology-free" regions of the A53T transgenic mouse brain (Tsika et al., 2010). The size of the α-synuclein assemblies documented in that study ranged from ~225 to 50 kDa. This size estimate was obtained by size exclusion chromatography, and is an interpretation based on the elution profile of globular protein standards. It is also important to note that we were likely detecting the most stable species under the specific conditions used for brain extracts (100,000 x g centrifugation of protein soluble in 1 percent Triton X-100). It is unclear whether the tetrameric species documented by Bartels et al. are stable under these extraction conditions. It will be of future interest to determine if such tetrameric species identified by Bartels et al. exist in human neurons, and if this particular form of α-synuclein is the predominant species in healthy brain.

View all comments by Harry Ischiropoulos
View all comments by Joseph Mazzulli

What I find most compelling about this work is the purification of helical α-synuclein from a cellular source. It is a long-standing hypothesis, based on conserved protein domains, that the active form of synuclein should be helical, but it has been difficult to directly demonstrate this species in vivo. Evidence that the native form is tetrameric, and that the tetramer has higher lipid affinity than monomer, is also intriguing. Identification of this conformer should help facilitate efforts to understand the normal function of α-synuclein.

It remains to be seen whether the helical tetramer is the predominant cellular species in brain neurons, an important question with relevance to Parkinson’s disease, which is not addressed directly by this study. I personally favor the idea that α-synuclein in the living neuron exists in conformational equilibrium, and that this equilibrium is influenced by phospholipids, fatty acids, and other environmental factors. Maintenance of this equilibrium might regulate synuclein function, while its perturbation could lead to protein misfolding...  Read more

What I find most compelling about this work is the purification of helical α-synuclein from a cellular source. It is a long-standing hypothesis, based on conserved protein domains, that the active form of synuclein should be helical, but it has been difficult to directly demonstrate this species in vivo. Evidence that the native form is tetrameric, and that the tetramer has higher lipid affinity than monomer, is also intriguing. Identification of this conformer should help facilitate efforts to understand the normal function of α-synuclein.

It remains to be seen whether the helical tetramer is the predominant cellular species in brain neurons, an important question with relevance to Parkinson’s disease, which is not addressed directly by this study. I personally favor the idea that α-synuclein in the living neuron exists in conformational equilibrium, and that this equilibrium is influenced by phospholipids, fatty acids, and other environmental factors. Maintenance of this equilibrium might regulate synuclein function, while its perturbation could lead to protein misfolding and aggregation. Future experiments will hopefully illuminate these issues.

View all comments by Julia George

Dr. Selkoe’s group has, in my opinion, made a breakthrough on the protein chemistry of α-synuclein. The finding that this protein indeed exists natively as a structured oligomer is very exciting and opens the field up for new experiments to elucidate its molecular biology and potentially also for new strategies for therapy development. I would like to add a few thoughts on biophysics to the ongoing discussion on this forum.

First, it does not come as a surprise that α-synuclein indeed natively adopts α-helical secondary structure, because secondary structure prediction algorithms predict substantial helix propensity in the N-terminal and central regions of the 140-residue isomer. This is in contrast to typical intrinsically unstructured proteins (IUPs) for which little ordered secondary structure (only “coil”) is predicted. The fact that α-synuclein previously has been deemed as unstructured has in this regard been somewhat of an enigma (to me). The present results clarify the issue very nicely.

Second, I still think that the stoichiometry of the present oligomer (tetramer)...  Read more

Dr. Selkoe’s group has, in my opinion, made a breakthrough on the protein chemistry of α-synuclein. The finding that this protein indeed exists natively as a structured oligomer is very exciting and opens the field up for new experiments to elucidate its molecular biology and potentially also for new strategies for therapy development. I would like to add a few thoughts on biophysics to the ongoing discussion on this forum.

First, it does not come as a surprise that α-synuclein indeed natively adopts α-helical secondary structure, because secondary structure prediction algorithms predict substantial helix propensity in the N-terminal and central regions of the 140-residue isomer. This is in contrast to typical intrinsically unstructured proteins (IUPs) for which little ordered secondary structure (only “coil”) is predicted. The fact that α-synuclein previously has been deemed as unstructured has in this regard been somewhat of an enigma (to me). The present results clarify the issue very nicely.

Second, I still think that the stoichiometry of the present oligomer (tetramer) might not be the last word, although in particular, the sedimentation equilibrium experiments provide strong support for a tetramer. A full structure determination by crystallography or nuclear magnetic resonance might eventually confirm the tetramer. However, the stoichiometry is, at present, not that important. What is important is that a native oligomeric folded state has been identified.

Third, it is not surprising that the STEM indicates some heterogeneity, as oligomers actually can be expected to dissociate in diluted samples under STEM. (I would actually have expected to see more monomeric species, but maybe these cannot be detected.) The apparent presence of a 30-kDa species is presumably due to such dissociation and possibly also an indication that the native tetramer is a dimer of dimers (containing two different monomer interaction surfaces).

Fourth, the question has been raised whether α-synuclein really adopts the folded oligomeric state also in the central nervous system. I think that this, in fact, is very likely, and that the described oligomer indeed is a native conformation in all cell types. I think so not only because the authors test a number of cell lines including neuroblastoma cells and mouse cortex tissue, but also because it would appear strange from a biological point of view if the physiologically relevant species in the human brain should resemble test-tube recombinant peptides rather than the natively folded state of the protein found in cell lines and tissue samples.

View all comments by Torleif Hard

We thank Julia George and Torleif Hard for their helpful comments and agree with both comments.

The four points outlined by Dr. Hard are compelling and make sense, and they reflect our own views about α-synuclein based on the new information.

View all comments by Tim Bartels
View all comments by Dennis Selkoe

Reply to Selkoe Comment
In Dr. Selkoe's comment above, he writes: "First, Dr. Lashuel suggested that one should run the denatured recombinant protein side by side with cell extracts. That is just what we did, and showed in Supp. Fig. S5 A and B. These two figures demonstrate the distinct migration of the cell-derived α-synuclein versus both recombinant α-synuclein and denatured cell-derived α-synuclein."

It would be great if the authors could provide more details about the exact conditions they used to generate the data shown in this figure, more specifically how they prepared the denatured α-synuclein in Figs. S5 A and B. Which denaturing agents or methods were used? How stable is the denatured protein and reproducibility of the data? This information is not in the manuscript, supplementary materials, or figure legend, and is essential for reproducing this particular experiment.

Also, it would be great if the authors could explain the discrepancy between the data in Fig. 1B and supplementary Fig. S5. In figure 1B, they indicated, as quoted below from Nature...  Read more

Reply to Selkoe Comment
In Dr. Selkoe's comment above, he writes: "First, Dr. Lashuel suggested that one should run the denatured recombinant protein side by side with cell extracts. That is just what we did, and showed in Supp. Fig. S5 A and B. These two figures demonstrate the distinct migration of the cell-derived α-synuclein versus both recombinant α-synuclein and denatured cell-derived α-synuclein."

It would be great if the authors could provide more details about the exact conditions they used to generate the data shown in this figure, more specifically how they prepared the denatured α-synuclein in Figs. S5 A and B. Which denaturing agents or methods were used? How stable is the denatured protein and reproducibility of the data? This information is not in the manuscript, supplementary materials, or figure legend, and is essential for reproducing this particular experiment.

Also, it would be great if the authors could explain the discrepancy between the data in Fig. 1B and supplementary Fig. S5. In figure 1B, they indicated, as quoted below from Nature paper, that monomeric α-synuclein runs below the 14 kDa molecular weight marker in a clear native gel, whereas in Fig. S5, they showed that it migrates at a much higher molecular weight in the same gel system.

The better resolution CN-PAGE without Coomassie dye also revealed small amounts of apparent monomers running below the 14 kDa molecular weight markers (Fig. 1B, lanes 1-4,6).

View all comments by Hilal Lashuel