A New Method for Quantitative Immunoblotting of Endogenous a-Synuclein
Andrew J. Newman, Dennis Selkoe, Ulf Dettmer*
From the Center for Neurologic Diseases, Brigham and Women’s Hospital and Harvard Medical School, Boston, Massachusetts, United States of America

Abstract
b-Sheet-rich aggregates of a-synuclein (aSyn) are the hallmark neuropathology of Parkinson’s disease and related synucleinopathies, whereas the principal native structure of aSyn in healthy cells - unfolded monomer or a-helically folded oligomer - is under debate. Our recent crosslinking analysis of aSyn in intact cells showed that a large portion of endogenous aSyn can be trapped as oligomers, most notably as apparent tetramers. One challenge in such studies is accurately quantifying aSyn Western blot signals among samples, as crosslinked aSyn trends toward increased immunoreactivity. Here, we analyzed this phenomenon in detail and found that treatment with the reducible aminereactive crosslinker DSP strongly increased aSyn immunoreactivity even after cleavage with the reducing agent bmercaptoethanol. The effect was observed with all aSyn antibodies tested and in all sample types from human brain homogenates to untransfected neuroblastoma cells, permitting easy detection of endogenous aSyn in the latter, which had long been considered impossible. Coomassie staining of blots before and after several hours of washing revealed complete retention of aSyn after DSP/b-mercaptoethanol treatment, in contrast to a marked loss of aSyn without this treatment. The treatment also enhanced immunodetection of the homologs b- and c-synuclein and of histones, another group of small, lysine-rich proteins. We conclude that by neutralizing positive charges and increasing protein hydrophobicity, amine crosslinker treatment promotes adhesion of aSyn to blotting membranes. These data help explain the recent report of fixing aSyn blots with paraformaldehyde after transfer, which we find produces similar but weaker effects. DSP/bmercaptoethanol treatment of Western blots should be particularly useful to quantify low-abundance aSyn forms such as extracellular and post-translationally modified aSyn and splice variants.
Citation: Newman AJ, Selkoe D, Dettmer U (2013) A New Method for Quantitative Immunoblotting of Endogenous a-Synuclein. PLoS ONE 8(11): e81314. doi:10.1371/journal.pone.0081314 Editor: Philipp J. Kahle, Hertie Institute for Clinical Brain Research and German Center for Neurodegenerative Diseases, Germany Received September 10, 2013; Accepted October 16, 2013; Published November 20, 2013 Copyright: ß 2013 Newman et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Funding: This work was supported by a research grant from the Fidelity Biosciences Research Initiative (www.fidelitybiosciences.com)to DS and a research grant from the Michael J Fox Foundation (www.michaeljfox.org) to UD and DS. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing Interests: The authors have declared that no competing interests exist. * E-mail: ulf_dettmer@harvard.edu

Introduction
The synuclein family consists of a-, b- and c-Synuclein (aSyn, bSyn, cSyn), small soluble proteins with a molecular weight below 15 kDa. aSyn is by far the best known and most studied among the three, due to its strong link with Parkinson’s disease (PD), the second most common human neurodegenerative disease. Point mutations and copy number variants in the aSyn gene cause familial forms of PD [1–4], and aSyn is the major component of Lewy bodies, intraneuronal protein aggregates that are the cytopathological hallmark of both familial and sporadic PD [5]. Its central role in PD pathogenesis has attracted considerable research into the biochemical, biophysical and functional properties of aSyn. However, despite more than two decades of aSyn research, essential features, such as molecular function, interaction partners, structure, and even sub-cellular localization, are still contentious. Recent developments indicate that results of studies of aSyn are dependent on the model system used: standard recombinant expression of aSyn in bacteria, for example, provides large quantities of protein suitable for biophysical and structural analyses, but this protein may insufficiently replicate the posttranslational modifications, folding and oligomerization of the endogenous protein in neurons [6–10]. Analysis of aSyn in

eukaryotic cells, on the other hand, is limited by relatively low abundance and various technical challenges. For example, Western blot analysis of endogenous aSyn in immortalized cell lines, even those of neuronal origin such as SH-SY5Y, had long been considered almost impossible due to extremely low expression levels [11]. A recent publication, however, has challenged this view by demonstrating that clear and strong detection of total endogenous aSyn from cultured cells is facilitated by fixation of proteins on blotted membranes with low concentrations of paraformaldehyde (PFA) [12], a method previously applied to the study of hemoglobin [13]. The authors suggest that this treatment prevents washing off of synuclein from PVDF or nitrocellulose membranes. We applied their method in a previous study of aSyn crosslinking, and it enabled meaningful Western blot comparisons of monomeric and oligomeric aSyn species without vs. with crosslinking [12]. Here, we report that immunodection of aSyn monomer can be improved even further – by more than 100% – with the application of reducible amine-reactive crosslinkers such as dithiobis[succinimidylpropionate] (DSP) followed by reductive cleavage (5% bME), prior to SDS-PAGE and electroblotting. In a variety of samples from neuroblastoma cell lines to human brain, we demonstrate that our novel method of improved aSyn monomer immunodetection is superior to the PFA
1 November 2013 | Volume 8 | Issue 11 | e81314

PLOS ONE | www.plosone.org

Quantitative Immunoblotting of a-Synuclein

treatment of blots. Purified recombinant aSyn transferred to PVDF membranes and stained with Coomassie dye before and after hours of washing was virtually 100% retained on the membrane with DSP/bME treatment, while untreated aSyn was nearly completely lost. Moreover, we observed the same enhancement of monomer immunodetection with the synuclein homologs b- and cSyn, as well as with another group of small, lysine-rich proteins, the histone family, suggesting that this new method may also benefit work outside the field of aSyn research.

Electrophoresis & Immunoblotting
All materials were purchased from Invitrogen unless stated otherwise. Total protein concentrations were determined by BCA assay (Thermo Scientific) according to manufacturer’s directions. Samples were prepared for electrophoresis by diluting with 4X NuPAGE sample buffer with or without 20% b-mercaptoethanol followed by boiling for 10 min. Routinely, 15–30 mg of sample were electrophoresed on NuPAGE 4–12% Bis-Tris gels with NuPAGE MES-SDS running buffer and the SeeBlue Plus2 molecular weight markers. After electrophoresis, gels were blotted onto Immobilon-PSQ 0.2-mm PVDF (Millipore) for 90 min at 400 mA constant current at 4uC in 25 mM Tris, 192 mM glycine, 20% methanol transfer buffer. Unless stated otherwise, posttransfer membranes were treated with 0.4% paraformaldehyde (PFA) in PBS for 30 min at RT and then rinsed with water [12]. Blocking was in 0.2% iBlock (Applied Biosystems) in PBS with 0.1% (v/v) Tween-20 (PBST) for $30 min. Membranes were then incubated in primary antibody in 0.2% iBlock in PBST with 0.02% sodium azide for either 1 h at RT or overnight at 4uC. Membranes were then washed 3610 min in PBST and incubated in secondary antibody in 0.2% iBlock in PBST followed by washing 3610 min in PBST and developing with ECL Prime (GE Healthcare-Amersham Biosciences) according to the manufacturer’s instructions. To visualize proteins on PVDF, membranes were briefly incubated in Ponceau solution or Coomassie (PageBlue Protein Staining Solution, Thermo Scientific; application on dried membranes).

Materials and Methods Cell culture
All materials were purchased from Invitrogen unless stated otherwise. All cells were cultured at 37uC in a 5% CO2 atmosphere. Human erythroid leukemia cells (HEL; ATCC number TIB-180) were cultured in RPMI 1640 (ATCC modification) supplemented with 10% fetal bovine serum (Sigma), 10 units/mL penicillin and 10 mg/mL streptomycin (Pen-Strep). Human neuroblastoma cells BE(2)-M17 (called M17D; ATCC number CRL-2267) and 3D5, an M17D-derived line overexpressing wild type human aSyn regulated by the tetracycline-off system [14], were cultured in high-glucose DMEM supplemented with 10% FBS, 2 mM L-glutamine and Pen-Strep. Doxycycline was added to cell medium to a final concentration of 100 ng/mL to repress aSyn expression. The human neuroblastoma cell line SHSY5Y (ATCC number CRL-2266) was cultured in a 1:1 mixture of high-glucose DMEM and F-12 supplemented with 10% FBS, 2 mM L-glutamine and Pen-Strep.

Antibodies
2F12, a monoclonal antibody (mAb) to aSyn, was produced inhouse [7]. Other aSyn mAbs were 15G7 [15] and Syn1 (Clone 42, Becton-Dickinson), and polyclonal antibody (pAb) C20 (Santa Cruz) was also used. Other antibodies were mAb EP1537Y to bsynuclein (Novus Biologicals), pAb 6169 to c-synuclein (Abcam), mAb 8226 to b-actin (Abcam), mAb AA2 to b-tubulin (Sigma), mAb to calmodulin (05-173, Millipore), mAb 71.1 to GAPDH (Sigma), pAb to DJ-1 ([16]), pAb to RAN (4462, Cell Signaling), pAb to 14-3-3 (9063, Abcam), pAb to UCH-L1 (ab1761, Millipore), pAb to Histone H3 (9715, Cell Signaling), and pAb to trimethyl Histone H3 Lys27 (ABE44, EMD Millipore).

Sample preparation from cultured cells and brain tissues
Our method of generation of cytosolic lysates from cultured human cell lines and primary rat neurons has been described [7]. In this study, membranes were routinely pelleted by centrifugation at 100,000 g for 60 min at 4uC. Brain homogenates from frozen mouse brain and human brain were generated by thawing, dounce homogenization and 15 s sonication (Sonic Dismembrator model 300, Fisher Scientific, Waltham, MA; microtip setting = 40). Brains were dissected from C57BL/6 male mice aged 11–13 weeks (Jackson Laboratory, Bar Harbor, ME) and stored at –80uC. A sample of human cerebral cortex from a neurologically normal subject was stored at –80uC.

Semiquantitave densitometry of Western blot signals
Scanned Western blots were analyzed with ImageJ software, version 1.47 [17]. Pictures were inverted and background signal from an empty lane subtracted to obtain the actual signals for each lane.

Ethics Statement
The use of human brain samples was approved under protocol number 1999P001180 (’Aging in the Brain: Role of the Fibrous Proteins’) by the Partners Human Research Committee (PHRC), the Institutional Review Board (IRB) of Partners Research Management. Informed written consent for the use of postmortem brain tissue was obtained from the families donating the patients’ brains to research when they enrolled in our CND Tissue Donation Program. The neuropathologist obtained consent at the time of autopsy. Rodent samples were acquired under protocol number 05022 (’Mouse Models for Parkinson’s Disease’), approved by the appropriate IACUC, the Harvard Medical Area Standing Committee on Animals.

Results Discrepancy between immunoblotting of aSyn from crosslinker-treated and untreated lysates
We recently published a detailed crosslinking analysis of aSyn in intact cells [7] that led us to conclude that a large portion of endogenous cytosolic aSyn exists in physiological oligomers, principally apparent tetramers, consistent with previous studies from our lab [6] and others [8–10]. In contrast to these findings, some labs have recently published evidence that leads them to support the earlier model that physiological aSyn exists in cells and tissues principally as unfolded monomers. However, these studies either did not include the intact-cell crosslinking approach [18], or the trapping of aSyn at oligomeric (e.g., dimeric) positions by crosslinking was considered unspecific [19]. One challenge in studying aSyn by crosslinking is achieving equality of Western blot signals between crosslinker- and control-treated samples. In most
2 November 2013 | Volume 8 | Issue 11 | e81314

Crosslinkers
Dithiobis[propionimidate] (DTBP), dithiobis[succinimidylpropionate] (DSP) and disuccinimidylglutarate (DSG) were from Pierce. Crosslinkers were dissolved in DMSO to a 50X stock concentration prior to addition to sample. Detailed crosslinking protocols (in vivo and in vitro with cells and on recombinant protein) are in [7].
PLOS ONE | www.plosone.org

Quantitative Immunoblotting of a-Synuclein

of our experiments, the crosslinker-modified aSyn showed somewhat greater immunoreactivity than did non-crosslinked aSyn [7]. We speculated that unmodified aSyn may be washed off of Western blot membranes, whereas crosslinker modification might prevent this loss. This hypothesis emerged from the finding of Lee and Kamitani that 0.4% paraformaldehyde (PFA) fixing of blot membranes prevented the apparent loss of aSyn monomers during immunoblotting, thereby enhancing their immunodetection [12]. By including their PFA step in our standard protocol for aSyn crosslinking analysis, we achieved quantitatively meaningful results (‘conservation of aSyn immunochemical matter’) to a much greater extent: the combined immunoreactivity of all aSyn species after crosslinking roughly approximated that of the monomer alone before crosslinking. However, even with the routine use of PFA treatment, we still found some inconsistencies in the relative aSyn immunoreactivities of untreated and crosslinker-treated samples. For example, we observed certain trends in aSyn immunodetection, as exemplified by primary rat neurons (Fig. 1a, two left panels are short and long exposures). First, treatment of intact neurons with increasing amounts of the reducible amine crosslinker DSP resulted in a stepwise increase in aSyn monomer immunoreactivity (lanes 2 and 3) compared to the vehicle (DMSO) treated control (lane 1), until at a higher DSP concentration, aS-60 (aSyn tetramer) was trapped (lane 4). Second, low amounts of the non-reducible amine crosslinker DSG trapped small amounts of aSyn oligomers but still increased monomer detection (lane 6) compared to DMSO alone (lane 1). Third, a higher DSG concentration trapped more oligomeric aSyn (lane 7), but this increase was not accompanied by a corresponding decrease in detected monomer compared to DMSO alone (lane 1). Fourth, an even higher DSG concentration led to appearance of a high MW aSyn-reactive smear, primarily at the expense of the monomer signal (lane 8). Densitometry of these blots (Fig. 1b, left and middle panel) supported these qualitative impressions and suggested that the relative immunoreactivity of the crosslinker-treated and untreated samples that one observes depends on the exposure time of the blot. Most importantly, longer blot exposures reveal the presence of HMW smears at higher DSG concentrations, thus exaggerating the total aSyn immunoreactivity in those lanes (see lane 8). A parallel analysis for the well-known dimeric protein DJ-1 revealed better conservation of total immunoreactivity between the treatments for this control protein and no appearance of high MW smears (Western blot: Fig. 1a, right panel; densitometry: Fig. 1b, right panel).

Lysine modification increases aSyn immunoreactivity by strengthening attachment to blot membranes
As Western blotting is only a semi-quantitative method, given the potential for signal saturation and overestimation of diffuse immunoreactivity, the increase in total aSyn immunoreactivity in the crosslinked samples was not unexpected, particularly in those with larger amounts of trapped oligomers and upon long exposure times. Nevertheless, the increase in aSyn monomer signal at crosslinker amounts that trapped only trace amounts of aSyn oligomers (Fig. 1a, lanes 2, 3 and 6) was unanticipated. This phenomenon was reminiscent of our difficulty in achieving equality of Western blot signals between crosslinker- and control-treated samples when initially blotting for aSyn without PFA treatment. We therefore reasoned that Western blotting preferentially detects crosslinker-modified aSyn, whether monomeric or oligomeric, even when PFA fixation is used. We hypothesized that PFA treatment of blots alone does not fully prevent aSyn from washing off during development, whereas robust crosslinker-modification of the lysine residues in the protein
PLOS ONE | www.plosone.org 3

may be sufficient. To test this, we treated human erythroid leukemia (HEL) cells (which express high endogenous levels of aSyn [7]) with 1 mM of DSP in vivo or DMSO only, then generated high-speed cytosols, normalized for protein concentration, and cleaved the reducible crosslinker by adding 5% bME before samples were boiled for SDS-PAGE. By Western blot analysis on PFA treated blots, we observed a markedly enhanced signal for aSyn in the DSP/bME-treated sample (Fig. 2a, right panel). Importantly the control protein DJ-1 did not show a similar effect (Fig. 2a, middle panel), and Ponceau staining of the PVDF membrane did not indicate higher protein loading in the DSP/ bME lane (left panel). We then tested our paradigm on the primary neurons and found that in addition to in vivo (not shown) even in vitro DSP treatment (addition of DSP to cytosolic lysates) followed by bME cleavage enhanced the aSyn immunosignal (Fig. 2b, second panel from the left, compare left and right lanes). Signal for the control monomeric protein b-tubulin (Fig. 2b, right panel, upper part) was not affected by DSP treatment, but bSyn and cSyn (Fig. 2b, third panel from the left and right panel, bottom part) showed similar marked enhancement of monomer by DSP/ bME as did aSyn. Ponceau staining (Fig. 2b, left panel) and blotting for b-tubulin (Fig. 2b, right panel, top part) ruled out higher protein loading in the DSP/bME lanes, as expected. Our next step was to test the DSP/bME treatment on the aSyn protein source most relevant to the study of Parkinson’s disease, human brain (Fig. 2c, left two lanes labled ‘H.s.’). Again, we observed a strong signal enhancement for aSyn monomers with DSP/bME treatment of brain homogenates (Fig. 2c, lowest panel). A similar effect was achieved with mouse brain homogenates (Fig. 2c, right two lanes labled ‘M.m.’). The control proteins b-tubulin and b-actin behaved differently, as their signals were unaffected (btubulin) or decreased (b-actin) by DSP/bME treatment. Based on the PFA result of Lee and Kamitani [12], we had initially hypothesized that increased immunodetection of aSyn by improved blotting procedures was a result of the aSyn protein not washing off the blot. However, alternate explanations such as enhanced antibody accessibility to aSyn were possible although unlikely, as all aSyn antibodies tested so far led to similar results. To rule out this possibility by avoiding the need for immunodetection, we turned to pure aSyn protein. Purified recombinant aSyn [20] and purified hen-egg lysozyme as a control [7] were incubated with DSP/bME or DMSO alone (Fig. 2d). All four conditions (two proteins, each with two treatments) were subjected to SDS-PAGE followed by transfer to a PVDF membrane. One part of the membrane (containing all four samples) was dried immediately after transfer, reducing the risk of subsequent protein wash-off (Fig. 2d, left panel). The other half of the membrane (also containing all 4 samples) was incubated overnight in our standard wash buffer (PBS with 0.1% Tween-20; PBST), then briefly washed in water and dried as well (Fig. 2d, right panel). We visualized the proteins on the dried membranes by brief incubation in Coomassie (see Materials and Methods), which we identified as the best method because it stained untreated and DSP/bME-treated aSyn similarly well. (Ponceau staining underestimated protein in crosslinker-treated samples, as it binds to positive charges, which are neutralized by the crosslinker; see Fig. 2a.). We found that unmodified aSyn, while fully present immediately after transfer (Fig. 2d, left half), was almost entirely gone after overnight washing of the blot (right half). In striking contrast, DSP/bME-treated aSyn was detectable to a similar extent with and without overnight washing. The control protein, lysozyme, was largely unaffected by any of the conditions. We conclude that the enhancing effect of DSP/bME on aSyn immunodetection is indeed a result of less washing-off of the
November 2013 | Volume 8 | Issue 11 | e81314

Quantitative Immunoblotting of a-Synuclein

Figure 1. Discrepancy between immunoblotting of aSyn from crosslinker-treated and untreated lysates. A. Intact primary cortical rat neurons (13 days in vitro) were incubated with DMSO only (-) or a gradient of 0.1, 0.5 and 1.0 mM DSP or DSG, sonicated in PBS supplemented with protease inhibitors and spun at 100,000 g for 60 min. The resultant cytosols (30 mg) were blotted with aSyn antibody Syn1 (left panel: short exposure; middle panel: long exposure) and with anti-DJ-1 as a control (right panel). Arrows indicate the crosslinker-trapped aSyn species that have been described before [7] as well as the DJ-1 dimer (2-mer) and monomer (1-mer). B. Densitometric analysis of the blots shown in Fig. 1a using ImageJ (see Materials and Methods). Bars indicate the total immunoreactivity of each lane as a percentage of the signal from DMSO-only treated sample (lane 1), after subtraction of the background from the empty lane 5. doi:10.1371/journal.pone.0081314.g001

crosslinker-modified aSyn compared to untreated protein. Moreover, the Coomassie staining in Fig. 2d shows that the DSP/bME method quantitatively retains aSyn on the membrane.

In vitro incubation of lysates with 2 mM DSP followed by bME reduction allows optimal aSyn detection
To optimize the DSP/bME method, we focused on human erythroid leukemia (HEL) cells, which are rich in endogenous aSyn, more consistent, and more readily available than primary
PLOS ONE | www.plosone.org 4

neurons or brain samples [7]. We compared the cleavable ˚ crosslinkers DSP (spacer length 12.0 A) and dithiobis[propionimi˚ date] (DTBP, spacer length 11.9 A) (Fig. 3a). We tested concentration gradients of 0.5, 2.0 and 8.0 mM of each crosslinker vs. DMSO alone (-) and both in vivo and in vitro. DTBP/bME treatment had a similar but weaker effect as DSP/bME, and DSP concentrations higher than 2 mM had no additional benefit, suggesting saturation of the mechanism. Overall, we observed no major differences between the in vivo and in vitro treatments

November 2013 | Volume 8 | Issue 11 | e81314

Quantitative Immunoblotting of a-Synuclein

Figure 2. Lysine modification increases aSyn immunoreactivity by strengthening attachment to blot membranes. A. HEL cells were crosslinked in vivo with the cleavable crosslinker DSP (1 mM) in comparison to DMSO-only treatment (-). Cytosols were prepared (post-20,000 g) and boiled in sample buffer with 5% (v/v) b-mercaptoethanol (bME), which cleaves the disulfide bond of DSP (DSPßME: reduced DSP). Samples (30 mg) were blotted with aSyn antibody 15G7. Equal protein loading was visualized by blotting for DJ-1 and by Ponceau (Ponc.) staining. B. Primary cortical rat neurons (13 days in vitro) were treated with 1 mM DSP in vitro (cytosolic lysates) in comparison to DMSO-only (-) treatment. Cytosols (post 100,000 g) were boiled in sample buffer with 5% bME and blotted with aSyn antibody 2F12, as well as antibodies to bSyn, cSyn and b-tubulin (b-tub.). Ponceau (Ponc.) stain of the blot membrane is shown on the left. C. Human (H.s.; 15 mg) and mouse (M.m., 30 mg) brain homogenates (PBS soluble fraction; post 100,000 g) were treated in vitro with DSP or DMSO-only (-), followed by boiling in sample buffer containing 5% bME. Blots were developed with aSyn antibody Syn1 as well as antibodies to b-tubulin, b-actin, GAPDH and DJ-1. D. Purified recombinant aSyn and purified hen-egg lysozyme (lyz) were crosslinked at a concentration of 500 ng/mL with 2 mM DSP (DSPßME) in comparison to DMSO only (-), followed by boiling in sample buffer with 5% bME. After separation by SDS-PAGE, samples were transferred to PVDF membranes which were dried immediately (left panel: ‘no wash’) or after washing overnight in PBST (‘wash’). Dried membranes were briefly stained with Coomassie Blue. 14, 17 and 28 kDa molecular weight markers are visible in the outer left lane. doi:10.1371/journal.pone.0081314.g002

(Fig. 3A). As treatment of lysates is generally easier than that of intact cells, we focused on the in vitro approach for further method optimization and on the more efficient compound, DSP. We checked whether the protein concentration during in vitro crosslinking affected the degree of aSyn detection on Western blots (Fig. 3b) and saw no clear trend with lysate protein concentrations of 2.4, 3.3 and 4.5 mg/mL for the two aSyn antibodies we tested, 2F12 and 15G7, using a DSP gradient paradigm analogous to

Fig. 3a. Independent of the lysate protein concentration, we found a strong enhancement of aSyn immunodetection by DSP/bME treatment that plateaued at 2 mM DSP, while immunodetection of GAPDH was unaffected. However, we subsequently chose a cytosol protein concentration range of between 2 and 5 mg/mL for in vitro treatments. We next confirmed the strong aSyn enhancement by DSP/bME using two more aSyn antibodies, C20 and Syn1, while the signals of several control proteins were more

PLOS ONE | www.plosone.org

5

November 2013 | Volume 8 | Issue 11 | e81314

Quantitative Immunoblotting of a-Synuclein

weakly enhanced (DJ-1, Ran) or unaffected (Calmodulin, 14-3-3, b-actin) (Fig. 3c). All samples shown in Fig. 3 were treated with bME at 5% final concentration, ruling out the possibility that the reducing agent alone is responsible for the increased immunodetection. This was confirmed when we boiled untreated lysates in the presence or absence of bME and subsequently blotted for aSyn (not shown).

DSP/bME treatment allows facile detection of aSyn from cultured cells
Finally, we tested how well our method detects aSyn in readily available sources whose aSyn expression levels have previously been underestimated, such as cultured cell lines of neuronal origin. We compared the levels of aSyn immunoblot detection with vs. without DSP/bME treatment in the tet-off aSyn overexpressing line 3D5 [14], in the absence (Fig. 4a, top two panels, lanes 1 and 2) and presence (lanes 3 and 4) of doxycycline, to their parental cell line M17D expressing only endogenous aSyn (lanes 6 and 7). Western blotting confirmed very strong expression in induced 3D5 cells compared to doxycycline-repressed 3D5 cells, which still expressed more aSyn than the parental M17D line due to incomplete repression. We observed the aSyn signal-enhancing effect of DSP/bME treatment independently of expression level (compare ‘-‘ lanes to ‘+’ lanes). Detection of the control protein DJ1 also benefited from DSP/bME treatment to some degree in this experiment, while b-actin detection was again reduced somewhat. Consistent with the report of Lee and Kamitani, PFA treatment alone of the PVDF membranes after transfer already improved detection of endogenous aSyn from these neuroblastoma cells. Nevertheless, additional incubation of the lysates with DSP/bME

further enhanced the aSyn signal markedly. We obtained similar results for another commonly studied neuroblastoma cell line, SHSY5Y (Fig. 4b, lanes 3 and 4). The relative levels of aSyn detected in M17D and SH-SY5Y cells and in human brain homogenates was not changed by DSP/bME treatments, i.e., the aSyn signals were proportionally enhanced regardless of aSyn abundance in the respective source (Fig. 4b, compare long blot exposure for neuroblastoma cells to short exposure for human brains). To compare systematically the effects of DSP/bME treatment of protein lysates and PFA treatment of blot membranes, we analyzed DSP/bME- and vehicle-treated lysates of SH-SY5Y cells (Fig. 4c) and M17D cells (Fig. 4d) on PVDF membranes with and without PFA treatment. All four aSyn antibodies used routinely in our lab (2F12, 15G7, Syn1, C20) showed the aSyn-enhancing effect of PFA in both cell lines (Figs. 4c and 4d, compare lane 1 to lane 3 in all aSyn panels). DSP/bME treatment, however, led to substantial additional aSyn signal on PFA-treated membranes (Figs. 4c and 4d, compare lane 3 to lane 4 in all aSyn panels). In fact, the strong aSyn signals caused by DSP/bME treatment of the lysates were not significantly enhanced further by additional PFA treatment of the blots (Figs. 4c and 4d, compare lane 2 to lane 4 in all aSyn panels), suggesting that the two treatments do not have additive effects. Furthermore, we tested these same samples on multiple types of transfer membrane – 0.2 mm pore PVDF (our standard), 0.45 mm pore PVDF and 0.2 mm pore nitrocellulose – and found no differences among these membrane materials (data not shown). While signals for most control proteins were either unchanged (UCH-L1, b-tubulin) or reduced (GAPDH) by DSP/ bME treatment, we did observe an immunodetection enhancing effect on the histone protein H3 using an antibody to total H3 as

Figure 3. In vitro incubation of lysates with 2 mM DSP followed by bME reduction allows optimal aSyn detection. A. HEL intact cells (in vivo) or lysates (in vitro) were incubated with DMSO only (-) as well as gradients of 0.5, 2.0 and 8.0 mM DSP and DTBP, respectively. Cytosols (post100,000 g) were boiled in sample buffer/5% bME and blotted with aSyn antibody 2F12. Identical exposures of the same blot are shown; film was cut at dotted line. B. HEL cell cytosols (post-100,000 g) at three different protein concentrations (2.4, 3.3., 4.5 mg/mL) were incubated with DMSO only (-) as well as a gradient of 0.5, 2.0 and 8.0 mM DSP and DSG. Samples were normalized to 2.4 mg after quenching, then boiled in sample buffer plus 5% bME and blotted with aSyn antibodies Syn1 and 15G7 as well as an antibody to GAPDH. C. HEL cell cytosols (post-100,000 g) were treated with DMSO-only (-) or DSP, quenched, boiled in sample buffer plus 5% bME and analyzed by blotting for aSyn (mAb Syn1 and pAb C20), Calmodulin, DJ-1, Ran, 14-3-3 and b-actin. doi:10.1371/journal.pone.0081314.g003

PLOS ONE | www.plosone.org

6

November 2013 | Volume 8 | Issue 11 | e81314

Quantitative Immunoblotting of a-Synuclein

Figure 4. DSP/bME treatment facilitates easy detection of aSyn from cultured cells. A. Cytosols (post-100,000 g) of 3D5 aSyn tet-off cells in induced (3D5I) or repressed state (3D5R) as well as lysates of the parental neuroblastoma cell line M17D were crosslinked with 2 mM DSP, followed by boiling in sample buffer with 5% bME. Western blot analysis was performed for aSyn (C20), DJ-1 and b-actin. (Note that the DJ-1 blot shows residual C20 signal from a previous exposure.) B. Cytosols (post-100,000 g) of neuroblastoma cell lines M17D and SH-SY5Y as well as the PBS-soluble fraction of human brain homogenates were incubated in 2 mM DSP, followed by boiling in sample buffer with 5% bME and blotting for aSyn (2F12) as well as b-tubulin. A long and a short exposure of the 2F12 blot are shown. C. Cytosols (post-100,000 g) of SH-SY5Y cells were crosslinked with 2 mM DSP, followed by boiling in sample buffer/5% bME and blotting for the indicated antibodies. Membranes were cut at dotted lines after protein transfer and the left half was incubated in 0.4% PFA for 30 min, while the right half was blocked immediately. Membranes were reassembled before film development; PFA-treated and -untreated halves shown were exposed identically. D. M17D cells analyzed analogously to SH-SY5Y cells in Fig. 4c. In addition to aSyn and UCH-L1, histone protein H3 was detected using a total H3 as well as an H3 lysine 27-methylation-specific antibody (H3 K27met). doi:10.1371/journal.pone.0081314.g004

well as one specific for trimethyl H3 (lysine 27) (Fig. 4d). This finding is of interest because histone proteins share certain features with aSyn – small size (aSyn: 140 aminoacids; H3: 136 aminoacids) and high lysine content (aSyn: 15; H3: 13) – suggesting that enhancement of immunodetection by lysine modification may be a common theme among small, lysine-rich proteins.

Discussion
Here, we present a simple method to markedly increase Western blot signals of aSyn by improving its retention on blot membranes. Our method involves protein modification by the cleavable amine-reactive crosslinker DSP followed by reductive cleavage of the crosslinks. In our optimized protocol, intact cells or cell lysates (2–5 mg/mL protein concentration) are treated with 2 mM DSP at 37uC for 30 min, followed by addition of 50 mM Tris (pH 7.6) to quench remaining crosslinker, and then boiling in dodecyl sulfate sample buffer containing 5% bME. Higher concentrations of DSP did not further enhance aSyn immunore-

activity. Beneficial effects of bME treatment alone were ruled out. A recently published method of improving aSyn immunodetection, i.e., incubation of blot membranes in 0.4% PFA [12], is less effective than our DSP-based method, and treatment with PFA in addition to DSP did not improve immunodetection beyond the effect of DSP/bME alone. The study describing the PFA method begins with the observation that endogenous aSyn in neuroblastoma cells is easily detected by fluorescent immunocytochemistry but not by standard Western blotting, despite the generally higher sensitivity of the latter technique [12]. Noting a previously reported method using the fixative PFA to improve immunodetection of hemoglobin, the authors applied this to Western blots of aSyn and found that the discrepancy between immunocytochemistry and WB was resolved [13]. Regarding the mechanism of the effect, they showed that the PFA method reduced aSyn wash-off from blot membranes. In our recent aSyn study [7], we observed discrepancies between aSyn immunodetection in crosslinked and non-crosslinked samples: aSyn immunodetection was greatly reduced in non-crosslinked samples [7]. With the PFA method, we could partially overcome
7 November 2013 | Volume 8 | Issue 11 | e81314

PLOS ONE | www.plosone.org

Quantitative Immunoblotting of a-Synuclein

this problem, suggesting that PFA could partially substitute for crosslinker modification and that the crosslinker, similarly to PFA, improved retention of aSyn on Western blots. However, even with PFA treatment of blot membranes, we found that untreated or lightly crosslinked samples routinely yielded lower aSyn monomer immunodetection than more highly crosslinked samples (see Fig. 1). This indicated that crosslinker modification improved retention of aSyn to a greater extent than the PFA method, even when the crosslinker is cleaved to yield just monomers. Indeed, DSP/bME treatment generally led to stronger aSyn signals relative to PFA treatment alone, while the PFA method was always superior to completely untreated samples (e.g., Fig. 4c and 4d). Nearly complete retention on membranes of recombinant aSyn treated with DSP/bME after hours of washing, in contrast to the substantial loss of untreated protein, confirms that the DSP/ bME method, like PFA treatment, effects stronger immunodetection by enhancing retention on the blot membrane rather than through epitope accessibility. Moreover, the close to complete retention observed suggests that further enhancement of aSyn immunodetection may not be necessary. The mechanism underlying the reported PFA effect is not selfevident [13]. The chemical inertness of PVDF and nitrocellulose makes formation of covalent bonds with aSyn, by reaction with PFA, improbable. As proteins bind to blot membranes through non-covalent, primarily hydrophobic interactions [21,22], small hydrophilic proteins like aSyn are not expected to bind as strongly as larger, more hydrophobic proteins. It is thus possible that PFA exerts its stabilizing effect on aSyn through formation of covalent bonds with other proteins that adhere more firmly to the blot membrane. However, Lee and Kamitani showed that the PFA method is effective even with purified aSyn. We similarly show that purified aSyn is retained on membranes to a greater extent with DSP/bME treatment than without, indicating that modification of lysine residues is sufficient for retention (Fig. 2d). Another amine-reactive, reducible crosslinker, DTBP, also enhanced aSyn immunoblotting though not as greatly as DSP; this difference may be due to the positive charge of the amidine product of the DTBPlysine reaction instead of the uncharged amide product of the DSP-lysine reaction. Furthermore, the strong effect of DSP/bME treatment, despite cleavage of the crosslinker, makes intermolecular crosslinking of aSyn an unlikely explanation for the PFA effect. We hypothesize that both PFA and DSP/bME effects occur through a combination of: a) masking the positive charges of lysines (of which aSyn has 15), thereby increasing aSyn hydrophobicity and blot membrane binding; and b) adding to aSyn a net hydrophobic moiety large enough to confer stronger interaction with the blot membrane. As non-immunological stains - Coomassie Blue dye - also show that crosslinker-modified aSyn can be fully retained on membranes, unlike unmodified aSyn (Fig. 2d), the enhanced immunodetection is probably not mediated by improved epitope accessibility. The closely similar results with all aSyn (as well as b- and cSyn) antibodies we tested further corroborate this conclusion. It was surprising that, with some exceptions such as GAPDH, modification of lysines did not negatively interfere with immunodetection of most proteins tested. The DSP/bME effect was even observed for the Syn1 antibody, whose recognized epitope contains lysines. Based on our experience with aSyn, we postulated that other small, lysine-rich proteins like the aSyn homologs bSyn and cSyn and histonefamily proteins might similarly benefit from DSP/bME treatment. Indeed, we observed enhanced Western blot detection upon DSP/

bME treatment for all three synucleins and the histone H3 (other histones not tested) (Figs. 2b and 4d). Our results support a conclusion from the Lee and Kamitani paper, that the field must reconsider the notion that immortalized cell lines are necessarily very low in aSyn expression. Our data show that they indeed have lower expression levels than brain homogenates, but aSyn is believed to constitute up to 1% of cytosolic proteins in neural tissue [23], and thus, the relatively lower levels detected by our method for the neuroblastoma cell lines M17D and SH-SY5Y may still be relatively high compared to other cellular proteins. Therefore, we propose using our new method in future studies of aSyn using Western blotting, at least in comparison to traditional Western blot protocols or the PFA-alone method, especially when analyzing sources with low levels or rare forms of aSyn such as secreted aSyn, low abundance splice variants, or possibly rare modifications of aSyn. Similarly, our method ensures detection of nearly all aSyn present at 14 kDa after transfer, allowing more quantitatively correct study of, for example, the relative abundance of high molecular weight aggregates and unaggregated oligomers by Western blotting (as shown and discussed for the PFA method in [12]). Finally, we propose DSP/bME treated samples to be the correct control to estimate total aSyn levels in comparison to the different oligomeric species trapped by uncleaved crosslinkers when performing crosslinking analyses of physiological aSyn oligomerization. If one used a vehicle-only control instead of DSP/bME, the consequent loss of aSyn monomer would give the impression that the crosslinked sample contained additional aSyn that was generated by the crosslinking. It should be remembered, however, that treatment with DSP/bME or PFA may mask antibody epitopes, although we did not observe such an effect with any of our aSyn, bSyn, cSyn or histone antibodies, nor for most of our control antibodies. However, each antibody needs to be assessed independently, regardless of its target. An additional caveat of our method is that certain standard biochemical analyses of protein lysates treated with DSP/bME may be altered, such as assays to determine protein concentrations or Ponceau membrane staining, in which masking of lysines interferes with binding of the dye to proteins. To overcome these limitations, we suggest normalizing samples before DSP treatment and visualizing proteins on blot membranes by methods less dependent on the presence of positive charges. In conclusion, the present method should be useful for all synuclein researchers who want to study, by Western blotting, the total synuclein in a given cellular, tissue or in vitro sample rather than just the small fraction detected using standard techniques.

Acknowledgments
We thank Mei-Chen Liao for the generation of primary rat neurons, Nicole Exner and Christian Haass for generously providing mAb 15G7, Maxwell Wang for confirming key observations, Eric Luth for help with human and mouse brain samples and critical revision of the manuscript, Tim Bartels for constant advice and critical revision of the manuscript, and all members of the Selkoe, LaVoie and Young-Pearse labs in the Center for Neurologic Diseases for helpful discussions.

Author Contributions
Conceived and designed the experiments: AJN DS UD. Performed the experiments: AJN UD. Analyzed the data: AJN DS UD. Contributed reagents/materials/analysis tools: AJN DS UD. Wrote the paper: AJN DS UD.

PLOS ONE | www.plosone.org

8

November 2013 | Volume 8 | Issue 11 | e81314

Quantitative Immunoblotting of a-Synuclein

References
1. Kruger R, Kuhn W, Muller T, Woitalla D, Graeber M, et al. (1998) Ala30Pro ¨ ¨ mutation in the gene encoding alpha-synuclein in Parkinson’s disease. Nat Genet 18: 106–108. doi:10.1038/ng0298-106. 2. Polymeropoulos MH, Lavedan C, Leroy E, Ide SE, Dehejia A, et al. (1997) Mutation in the alpha-synuclein gene identified in families with Parkinson’s disease. Science 276: 2045–2047. 3. Zarranz JJ, Alegre J, Gomez-Esteban JC, Lezcano E, Ros R, et al. (2004) The ´ new mutation, E46K, of alpha-synuclein causes Parkinson and Lewy body dementia. Ann Neurol 55: 164–173. doi:10.1002/ana.10795. 4. Singleton AB, Farrer M, Johnson J, Singleton A, Hague S, et al. (2003) alphaSynuclein locus triplication causes Parkinson’s disease. Science 302: 841. doi:10.1126/science.1090278. 5. Spillantini MG, Schmidt ML, Lee VM, Trojanowski JQ, Jakes R, et al. (1997) Alpha-synuclein in Lewy bodies. Nature 388: 839–840. doi:10.1038/42166. 6. Bartels T, Choi JG, Selkoe DJ (2011) a-Synuclein occurs physiologically as a helically folded tetramer that resists aggregation. Nature 477: 107–110. doi:10.1038/nature10324. 7. Dettmer U, Newman AJ, Luth ES, Bartels T, Selkoe D (2013) In vivo crosslinking reveals principally oligomeric forms of a-synuclein and b-synuclein in neurons and non-neural cells. J Biol Chem 288: 6371–6385. doi:10.1074/ jbc.M112.403311. 8. Wang W, Perovic I, Chittuluru J, Kaganovich A, Nguyen LTT, et al. (2011) A soluble a-synuclein construct forms a dynamic tetramer. Proc Natl Acad Sci U S A 108: 17797–17802. doi:10.1073/pnas.1113260108. 9. Trexler AJ, Rhoades E (2012) N-Terminal acetylation is critical for forming ahelical oligomer of a-synuclein. Protein Sci Publ Protein Soc 21: 601–605. doi:10.1002/pro.2056. 10. Westphal CH, Chandra SS (2013) Monomeric synucleins generate membrane curvature. J Biol Chem 288: 1829–1840. doi:10.1074/jbc.M112.418871. 11. Matsuo Y, Kamitani T (2010) Parkinson’s disease-related protein, alphasynuclein, in malignant melanoma. PloS One 5: e10481. doi:10.1371/journal.pone.0010481. 12. Lee BR, Kamitani T (2011) Improved Immunodetection of Endogenous a-Synuclein. PLoS ONE 6: e23939. doi:10.1371/journal.pone.0023939. 13. Suzuki Y, Takeda Y, Ikuta T (2008) Immunoblotting conditions for human hemoglobin chains. Anal Biochem 378: 218–220. doi:10.1016/j.ab.2008.04.008. 14. Ko LW, Ko HHC, Lin WL, Kulathingal JG, Yen SHC (2008) Aggregates assembled from overexpression of wild-type alpha-synuclein are not toxic to human neuronal cells. J Neuropathol Exp Neurol 67: 1084–1096. doi:10.1097/ NEN.0b013e31818c3618. 15. Kahle PJ, Neumann M, Ozmen L, Muller V, Jacobsen H, et al. (2000) Subcellular localization of wild-type and Parkinson’s disease-associated mutant alpha -synuclein in human and transgenic mouse brain. J Neurosci Off J Soc Neurosci 20: 6365–6373. 16. Baulac S, LaVoie MJ, Strahle J, Schlossmacher MG, Xia W (2004) Dimerization of Parkinson’s disease-causing DJ-1 and formation of high molecular weight complexes in human brain. Mol Cell Neurosci 27: 236–246. doi:10.1016/ j.mcn.2004.06.014. 17. Schneider CA, Rasband WS, Eliceiri KW (2012) NIH Image to ImageJ: 25 years of image analysis. Nat Methods 9: 671–675. 18. Burre J, Vivona S, Diao J, Sharma M, Brunger AT, et al. (2013) Properties of ´ native brain a-synuclein. Nature 498: E4–6; discussion E6–7. doi:10.1038/ nature12125. 19. Fauvet B, Mbefo MK, Fares MB, Desobry C, Michael S, et al. (2012) aSynuclein in central nervous system and from erythrocytes, mammalian cells, and Escherichia coli exists predominantly as disordered monomer. J Biol Chem 287: 15345–15364. doi:10.1074/jbc.M111.318949. 20. Kamp F, Exner N, Lutz AK, Wender N, Hegermann J, et al. (2010) Inhibition of mitochondrial fusion by a-synuclein is rescued by PINK1, Parkin and DJ-1. EMBO J 29: 3571–3589. doi:10.1038/emboj.2010.223. 21. Starita-Geribaldi M, Sudaka P (1990) Fast electrotransfer of human serum proteins in their native state from polyacrylamide thin gradient gels reinforced by textiles to polyvinylidene difluoride membranes. Rapid electrotransfer from thin gradient gels reinforced by textiles. Bioseparation 1: 111–117. 22. Wahlgren M, Arnebrant T (1991) Protein adsorption to solid surfaces. Trends Biotechnol 9: 201–208. 23. Iwai A, Masliah E, Yoshimoto M, Ge N, Flanagan L, et al. (1995) The precursor protein of non-A beta component of Alzheimer’s disease amyloid is a presynaptic protein of the central nervous system. Neuron 14: 467–475.

PLOS ONE | www.plosone.org

9

November 2013 | Volume 8 | Issue 11 | e81314