The notion that human α-synuclein is an intrinsically disordered monomeric protein was recently challenged by a postulated α-helical tetramer as the physiologically relevant protein structure. The fact that this alleged conformation had evaded detection for so many years was primarily attributed to a widely used denaturation protocol to purify recombinant α-synuclein. In the present paper, we provide in-cell NMR evidence obtained directly in intact Escherichia coli cells that challenges a tetrameric conformation under native in vivo conditions. Although our data cannot rule out the existence of other intracellular protein states, especially in cells of higher organisms, they indicate clearly that inside E. coli α-synuclein is mostly monomeric and disordered.
- in-cell nuclear magnetic resonance (in-cell NMR)
- intrinsically disordered protein (IDP)
Structural features of human α-synuclein
α-Synuclein is expressed abundantly in brain dopaminergic neurons and its aggregation into amyloid fibrils has been correlated strongly with the onset of PD (Parkinson's disease) and neurodegeneration [1,2]. α-Synuclein contains 140 amino acids and can be divided into three main regions according to its primary sequence. The N-terminal region (residues 1–60) is known to interact with lipid vesicles [3,4], the hydrophobic NAC (non-amyloid-β component) region (residues 61–95) is responsible for protein aggregation  and the negatively charged C-terminus (residues 96–140) is reported to counteract α-synuclein aggregation [6–8]. The structural properties of α-synuclein in its soluble form have been investigated extensively. From an early purification study under native conditions by Lansbury and co-workers in 1996, for which SEC (size-exclusion chromatography), native PAGE, sucrose gradient centrifugation and MALDI–TOF (matrix-assisted laser-desorption ionization–time-of-flight)-MS were used, α-synuclein was described as a disordered monomer with an apparent molecular mass of 14.5 kDa . Spectroscopic characterization by CD and FTIR (Fourier-transform IR) spectroscopy did not detect regions of stable secondary structure or a hydrophobic protein core. In the following years, these findings were corroborated by several other groups [10–19]. NMR was later employed to characterize the monomeric solution state of α-synuclein at atomic resolution. 13C chemical shift data demonstrated that the first ~100 residues had propensities to populate transient α-helical structures [20,21]. The C-terminal region of α-synuclein was determined to be less well defined, although short stretches of β-turn-like conformations were identified [17,20]. As expected for a monomeric disordered protein, 15N relaxation data revealed increased backbone amide dynamics compared with those of folded globular proteins . NMR spectroscopy, SAXS (small-angle X-ray scattering), DLS (dynamic light scattering) and fluorescence spectroscopy delineated a solution conformation that was not fully compatible with an extended random-coil state [6,13,22–24]. In turn, transient intramolecular long-range contacts between the N- and C-terminus, as well as the NAC region and C-terminus, of α-synuclein were identified and shown to prevent the central NAC region from spontaneous aggregation into oligomeric fibrils [6,22]. Together, these data established that α-synuclein existed in defined structural ensembles of interconverting monomers, similar to what has also been observed for many other IDPs (intrinsically disordered proteins) [25,26].
Monomeric compared with tetrameric α-synuclein
The disordered monomeric conformation of α-synuclein was challenged recently by two papers that proposed a different structure under physiological conditions [27,28]. These proposed a tetrameric state of the protein with well-defined α-helical segments. Moreover, it was suggested that this α-synuclein conformation had remained undetected to date mainly because of a denaturation protocol that is employed by many groups to purify recombinant α-synuclein. It was stated that the use of denaturing agents or boiling of the bacterial cell lysates during the initial purification steps of recombinant α-synuclein could destroy the tetramer and result in the accumulation of the disordered monomer [27,28]. Endogenous α-synuclein isolated from RBCs (red blood cells), different cultured neuronal cell lines and human tissue under non-denaturing conditions was reported to occur predominantly as a native tetramer of approximately 58 kDa, on the basis of analytical ultracentrifugation, native PAGE, transmission scanning electron microscopy and in vitro cross-linking of the protein . A similar hypothetical α-synuclein tetramer was also purified from Escherichia coli, although in this case the α-synuclein construct contained ten additional amino acids at its N-terminus . CD spectra of α-synuclein purified under non-denaturing conditions suggested a high α-helical content in both studies. Monomeric α-synuclein was shown to bind to lipid vesicles, leading to the formation of stable α-helical segments within its first 100 residues [29,30]. The alleged α-synuclein tetramer was reported to bind membranes with higher affinity than the disordered monomer . Finally, the proposed native tetrameric form of α-synuclein displayed a reduced tendency to form amyloid fibrils and was non-toxic when added to cultured neuronal cells [27,28]. According to 13Cα and 1Hα NMR chemical shift data, the N-terminally extended tetrameric form of α-synuclein exhibited ‘significant’ helical propensities at different N-terminal positions and within the central NAC region . 15N-edited NOESY–HSQC (heteronuclear single-quantum coherence) experiments and CD spectroscopy measurements were reported to confirm these findings. It was noted that two-dimensional 1H-15N correlation NMR spectra of tetrameric α-synuclein failed to detect the characteristic features of α-helical secondary structure, in contrast with what has been observed for monomeric α-synuclein bound to micelles [29,30]. This behaviour was attributed to the dynamic transient nature of the helical regions in the tetramer .
Conspicuously, the overall appearance of the two-dimensional NMR spectra of the α-synuclein tetramer closely resembles previous in vitro NMR spectra of the monomeric protein [6,7,31,32]. We have used the NMR chemical shift data of the α-synuclein tetramer that have been deposited in the Biological Magnetic Resonance Data Bank (BRMB #17665)  to emulate a two-dimensional 1H-15N correlation spectrum of this protein state. We then compared this virtual NMR spectrum with a two-dimensional 1H-15N correlation spectrum that we recorded in vitro on a sample of recombinant monomeric α-synuclein, purified under denaturing conditions, but resuspended in the same buffer used for collecting the deposited tetramer chemical shifts. A superposition of these NMR spectra (Figure 1A) shows that they are very similar. Differences were only seen for the deposited chemical shifts of Tyr39 and Leu113. Line broadening of Ser42, Asn103 and of residues in the glycine region of the NMR spectrum was due to known unfavourable chemical exchange effects at this temperature and pH (i.e. 25°C and pH 7.4). These observations are in line with previously published data . For two-dimensional 1H-15N NMR spectra recorded on the same sample at 10°C, previously line-broadened resonances were clearly visible (results not shown). These data suggest that the reported differences in chemical shift values between the alleged α-synuclein tetramer and the monomeric form of the protein primarily arise from different solution conditions.
Possible oligomeric states of α-synuclein have been discussed previously. Even the seminal paper by Lansbury and co-workers, in which the first biophysical characterization of α-synuclein was reported, raised the question of whether α-synuclein might exist as an oligomer . The reasons for this were based on concerns about the unusual migration behaviour of α-synuclein on SEC columns and native PAGE (apparent molecular mass ~58 kDa, measured relative to globular protein standards). These observations were interpreted as the monomeric form of α-synuclein having a more elongated shape than the globular protein standards. Further support for this notion was provided by sucrose gradient ultracentrifugation analyses and native PAGE experiments at different acrylamide concentrations, with which molecular masses of ~20 kDa were determined . Higher apparent molecular masses (~57 kDa) have also been obtained for β-synuclein, another synuclein isoform with high sequence homology within the N-terminal protein region, and for other IDPs [33–35]. In 2012, six research groups teamed up to carefully reinvestigate the monomer/tetramer controversy . Employing different native and denaturing purification methods in parallel, α-synuclein from mammalian central nervous system and red blood cells, as well as from E. coli, was prepared. The consortium reported that α-synuclein from both types of preparations had similar chromatographic elution profiles and electrophoretic mobility properties, according to molecular mass standards of folded and unfolded proteins. Moreover, the two-dimensional 1H-15N NMR features of α-synuclein purified under non-denaturing conditions closely matched those of protein preparations that were obtained from denaturing purification procedures , which collectively argued against the natively folded tetramer conformation. These observations are in agreement with a recent study that stipulated the requirements for N-terminal α-synuclein acetylation and a non-denaturing purification protocol in the presence of glycerol and the non-ionic detergent BOG (octyl β-D-glucopyranoside) in order to recover an oligomeric more α-helical form of α-synuclein from E. coli . Because α-synuclein is not naturally acetylated in bacteria, a fission yeast N-terminal acetylation B complex had to be co-expressed for efficient modification. These results also contradict the hypothesis of a folded α-synuclein tetramer in non-acetylation-competent E. coli cells.
Insights from in-cell NMR
Because possible differences in α-synuclein conformations caused by denaturing or non-denaturing purification protocols are at the centre of the monomer/tetramer controversy, direct in-cell NMR measurements of α-synuclein in intact E. coli cells are well poised to provide insights into the ‘native’ protein state(s) inside live bacteria. Prokaryotic in-cell NMR experiments do not require any form of sample purification or cell lysis and are therefore ideally suited to investigate the high-resolution structural propensities of α-synuclein in an undisturbed cellular environment [38–42]. In-cell NMR measurements involve NMR isotope labelling (15N and/or 13C) during the induction period of recombinant protein expression and direct NMR readouts on the intact cell slurry. Two-dimensional in-cell NMR spectra then provide atomic-resolution details about the structural features of the overexpressed intracellular protein. A number of in-cell NMR studies of α-synuclein inside E. coli cells have been reported [21,43–48]. In all of these studies, in-cell NMR spectra of α-synuclein (1H-15N and 13CO-15N) and in vitro reference correlations of the protein purified under denaturing conditions looked remarkably similar. Our own in-cell NMR data corroborate this notion (Figure 1B). We, and others, interpret these findings as an indication for the preservation of the monomeric disordered protein state inside the crowded cytoplasm of live bacteria. Proponents of the tetramer hypothesis argue that these and other E. coli in-cell NMR spectra resemble cross-peak patterns of the folded tetramer, which they interpret as pointing towards a native oligomeric α-synuclein state inside bacterial cells .
To correlate the intracellular protein concentration of overexpressed α-synuclein in our in-cell NMR samples to the experimentally obtained NMR signal intensities, we employed quantitative Western blotting and two-dimensional NMR signal integration (Figures 1B and 1C). These measurements established that intracellular α-synuclein was present at a concentration of ~300 μM, which corresponded well to the ~250 μM of α-synuclein in-cell NMR signals that we recorded (see the Supplementary Online Data at http://www.biochemsoctrans.org/bst/040/bst0400950add.htm for details). This indicated that the majority of α-synuclein inside E. coli contributed to the NMR signals detected. If these in-cell NMR measurements were to report the presence of a folded tetramer, the overall NMR signal quality would be greatly impaired by the much lower tumbling rate in this high-viscosity environment . Pielak and co-workers recently presented an elegant demonstration of this concept by recording in-cell NMR spectra of α-synuclein (14 kDa) covalently fused to ubiquitin (5 kDa) . Only NMR signals of the disordered α-synuclein portion of this construct could be detected and these superimposed well on to reference cross-peaks of the disordered monomer. A similar scenario would be expected for the postulated tetramer for which the first 100 residues were speculated to form the α-helical core region, while the protein C-terminus remained more flexible . No such NMR characteristics were observed in in-cell NMR experiments.
However, in-cell NMR spectra of α-synuclein did reveal regions of site-selective line broadening. These mapped to different portions of α-synuclein, interspersed throughout the primary amino acid sequence in a non-continuous fashion (Supplementary Figure S1B at http://www.biochemsoctrans.org/bst/040/bst0400950add.htm). In macromolecular crowded in vitro solutions (305 mg/ml BSA), α-synuclein displayed similar patterns of line broadening (Supplementary Figure S1A), which indicated that unspecific interactions with BSA (in vitro), or with cellular components in E. coli (in-cell) and/or conformational/chemical exchange effects could be causing this behaviour. To rule out chemical exchange contributions, we resorted to two-dimensional heteronuclear (13CO-15N) in-cell experiments, which do not involve exchangeable protons . Direct carbon-detected 13CO-15N experiments have previously been employed to characterize bacterial in-cell α-synuclein samples . Because 13CO-15N spectra display greater overall chemical shift dispersion, their use for residue-resolved analyses is advantageous over 1H-15N correlations. In addition, 13C chemical shifts of backbone carbonyls are highly sensitive to changes in protein backbone conformations and are widely used as unbiased indicators for protein secondary structure [50,51]. Similar to the previous in-cell NMR study employing 13CO-15N experiments, our in vivo NMR data revealed α-synuclein 13CO-15N cross-peaks that closely matched the in vitro reference pattern of the disordered monomer (Figure 1B, right-hand panel). Minor chemical shift differences (δΔCO <0.07 p.p.m., with the exception of His50 which reported a slightly more basic intracellular pH) were detected throughout the protein (Figure 1D). Overall, no alterations in secondary structure were detected. Matching regions of site-selective line broadening were preserved under BSA-crowded in vitro conditions, although to a lesser extent (Supplementary Figure S1B). Moreover, when we lysed the incell NMR samples and recorded NMR spectra of the resulting extracts, the majority of the broadened NMR signals were recovered at nearly identical peak positions compared with in vitro reference spectra (Supplementary Figures S2C–S2D at http://www.biochemsoctrans.org/bst/040/bst0400950add.htm). Cell lysis was performed using a mild protocol  and without boiling or denaturing the extract. According to quantitative NMR peak volume analyses, the amount of α-synuclein that was present in this extract (~260 μM) was comparable with the amount detected by in-cell NMR measurements. Together, this indicated that α-synuclein exhibited a disordered monomeric conformation inside live E. coli cells.
We have presented in vitro and in-cell NMR evidence that call into question the postulated tetramer conformation of α-synuclein. Although, strictly speaking, we cannot rule out other intracellular α-synuclein conformations, especially in higher eukaryotic cells, in-cell NMR data in E. coli strongly support the idea that at least in this cellular environment α-synuclein displays structural and dynamic properties of an intrinsically disordered monomer.
P.S. is a recipient of an Emmy Noether Programme grant [grant number SE-1794/1-1] by the Deutsche Forschungsgemeinschaft (DFG).
We thank Dr Linda Ball for carefully reading the paper and helpful comments, Dr Stamatios Liokatis and Dr Honor M. Rose for stimulating discussions, and Dr Peter Schmieder and Monika Beerbaum for NMR infrastructure support.
Intrinsically Disordered Proteins: A Biochemical Society Focused Meeting held at University of York, U.K., 26–27 March 2012. Organized and Edited by Jennifer Potts (York, U.K.) and Mike Williamson (Sheffield, U.K.).
Abbreviations: BOG, octyl β-D-glucopyranoside; HSQC, heteronuclear single-quantum coherence; NAC, non-amyloid-β component; SEC, size-exclusion chromatography
- © The Authors Journal compilation © 2012 Biochemical Society