α-synuclein C-terminal Truncation Modulates Its Cytotoxicity and Aggregation by Promoting Its N-terminal Interactions With Membrane and Chaperone

α-synuclein (α-syn) is the main protein component of Lewy bodies, the major pathological hallmarks of Parkinson’s disease (PD). C-terminal truncated forms of α-syn are found in the brain of PD patients, and reduce cell viability, as well as tend to form brils. Nevertheless, little is known about the molecular mechanisms underlying the role of C-terminal truncation on the cytotoxicity and aggregation of α-syn. The N-terminal domain of α-syn interacts with membranes and molecular chaperones, and these interactions are critical for the protein’s physiological function and its pathological effects in PD. Here, we use nuclear magnetic resonance spectroscopy, a co-otation assay, dynamic light scattering, and thioavin T uorescence to show that the truncation alters α-syn conformation, resulting in a new attractive interaction of the N-terminus with membranes and molecular chaperone, protein disulde isomerase (PDI). Remarkably, the truncated protein is more toxic to mitochondria than full-length protein and diminishes the effect of PDI on α-syn brillation. Our ndings reveal a modulatory role for the C-terminus in the cytotoxicity and aggregation of α-syn by interfering the interaction of the N-terminus with membrane and chaperone. pathological role of C-terminal truncation in

The compact conformation of monomeric α-syn is driven by long-range transient intramolecular interactions between the N-and C-termini (35,36). In cells, the N-terminal domain interacts with membranes and chaperones that modulates brillation. For instance, the domain regulates fusion of synaptic vesicles and subsequent neurotransmitter release, which are critical for normal brain function (37,38). Molecular chaperones, such as protein disul de isomerase (PDI) and Hsp70, bind the Nterminal domain, dissolve and degrade accumulated misfolded α-syn, highlighting the role of chaperones in preventing α-syn aggregation and regulating its physiological function (39)(40)(41). These observations raise the question of whether C-terminal truncations affect the interaction of the N-terminal domain with membranes and chaperones, which could affect its functionality.
Here, we investigate the effects of C-terminal truncation. Nuclear magnetic resonance spectroscopy (NMR) and co-otation data reveal a stronger interaction of N-terminus with membranes and with PDI, a major endoplasmic reticulum (ER) chaperone (42). The interaction may arise from the more extend conformation of CT-α-syn. The functional assay reveals severe mitochondrial membrane disruption by CT-α-syn compared to FL-α-syn in both neuronal cells and isolated mitochondria. Dramatically differential inhibitory effects of PDI on FL-α-syn and CT-α-syn brillation are also observed. Our results suggest that the C-terminus behaves like a 'guard' regulating α-syn function. The absence of this guard results in stronger membrane interaction, accelerated brillation and diminished chaperone inhibition, all of which could affect the etiology of PD.

Results
C-terminal truncation causes stronger interactions with membranes and PDI. α-syn forms an N-terminal αhelix upon binding negatively charged membranes, without involvement of the C-terminus(4), but it is not known if deleting the C-terminus affects binding. We prepared lipid mixtures comprising 1-palmitoyl-2oleoyl-sn-glycero-3-phosphocholine (POPC)/1-palmitoly-2-oleoyl-sn-glycero-3-phosphate (POPA) (1:1 in mole ratio) to obtain highly anionic liposomes. Binding of FL-and CT-α-syn was analyzed by co-otation, a common method for studying protein-membrane interaction. The principle is that bound and unbound proteins migrate differentially in iohexol gradients. Liposomes, and membrane associated proteins migrate to the top of the gradient and unbound proteins to the bottom ( Figure 1B). The fractions from top to bottom were collected and quanti ed by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) ( Figure 1C and S1A). Approximately 80% of the CT-α-syn binds the membrane whereas ~50% of FL-α-syn is bound state ( Figure 1D), indicating stronger association of CT-α-syn. α-syn associates with mitochondrial membranes, whose major component is the anionic phospholipid cardiolipin (CL) (43). We acquired 1 H-15 N heteronuclear single quantum coherence (HSQC) spectra of uniformly 15 N-enriched FL-and CT-α-syn in the presence of liposomes comprising POPC/CL (9:1 and 7:3 mole ratio, respectively). NMR provides binding information at the residue level because resonances from strongly bound residues show reduced intensities (44). Analysis of spectra acquired in the absence and presence of liposomes ( Figure 1E and S1B) indicate that, consistent with previous results, the N-terminus dominates binding (5). Data for CT-α-syn show that signals are more attenuated in the presence of liposomes containing CL, indicating a stronger binding of CT-α-syn to acidic lipids than FL-α-syn, consistent with co-otation data.
Molecular chaperones are another important binding partner of the N-terminal region (45). PDI interacts with the hydrophobic region of the N-terminal region, inhibiting aggregation (41,46). To investigate the impact of the C-terminal truncation on the interaction with PDI, we acquired 1 H-15 N HSQC spectra of uniformly 15 N-enriched FL-and CT-α-syn in the presence of PDI at a mole ratio of 2:1, which enabled us to assess intermolecular interactions at the residue level via chemical shift perturbations (CSPs) and intensity changes. PDI binds the rst 10 residues of FL-α-syn (segment 1) and 10 residues near Tyr39 (segment 2) (Figure 2A and B). Both sites are located at the N-terminus, consistent with other reports (41,45). Although the attenuation patterns of CT-and FL-α-syn in PDI are similar, there are changes in chemical shift and intensity (Fig 2C), suggesting the constructs have similar PDI-binding sites but different a nities. We quanti ed the dissociation constants (K D ) of FL-and CT-α-syn by tting CSPs values as a function of PDI concentration ( Figure 2D). Two segments of FL-α-syn show similar mM K D values, whereas segment 1 of CT-α-syn has a K D of ~39.3 μM, about an order magnitude increase in a nity of CT-α-syn for PDI.
C-terminal truncated α-syn exhibits a more extended conformation. Given the difference in membrane binding between FL-and CT-α-syn, we examined the constructs using PRE to determine whether deleting the C-terminus alters the ensemble of conformations. We prepared spin-labeled proteins by incorporating 1-oxyl-2,2,5,5-tetramethyl-delta3-pyrroline-3-methyl (MTSL), a paramagnetic spin label, into single cysteine variants (A90C) of 15 N-enriched FL-and CT-α-syn. Cross peaks arising from residues close to the label broaden due to enhanced transverse relaxation. As reported (35), intramolecular interactions between the N-and C-terminal regions lead to broadening proximal to the spin label and in residues within the N-terminal region, neighboring Tyr39 ( Figure 3A). Similar intramolecular long-range interactions are absent in CT-α-syn. Its N-terminal residues show weak PRE, suggesting a more extended conformation for CT-α-syn ( Figure 3B). We conclude that interactions between the N-and C-terminal regions cause the PRE effects observed in FL-α-syn.
Effects of C-terminus truncation on physiological and pathological functions. To better understand the membrane-related physiological differences between FL-and CT-α-syn, we evaluated the mitochondrial potential (ΔΨ m ) in living SK-N-SH cells using the dye, 5, 5', 6, 6'-tetrachloro-1, 1', 3, 3'tetraethylbenzimidazolocarbocyanine iodide (JC-1). In normal cells, JC-1 aggregates in the matrix due to the electrochemical potential gradient resulting in the red uorescence. Apoptotic events cause monomeric JC-1 to disperse throughout the cell resulting in green uorescence. Images were acquired after treatment with 50 μM FL-or CT-α-syn for 5 h. Mitochondria in untreated SK-N-SH cells stained bright red. Red uorescence decreased markedly in CT-α-syn-incubated cells concomitant with an increase in green uorescence, indicating collapse of the mitochondrial membrane potential ( Figure 4A). The green/red uorescence intensity ratio showed no signi cant difference between FL-α-syn-treated and control cells ( Figure 4B). Thus, the C-terminal truncation increases mitochondrial damage compared to the negative control and FL-α-syn.
Next, we isolated mitochondria from SK-N-SH cells. Given the dissipated membrane potential, we anticipated mitochondrial membrane permeabilization, allowing cytochrome c (cyto c) release (47,48). We assessed release after treatment with various concentrations of FL-and CT-α-syn after 1 h at 37 °C ( Figure 4C) via Western blotting (WB). Enhanced release relative to control mitochondria was detected with CT-α-syn. We did not observe signi cant release of cyto c compared to controls with FL-α-syn. This result lends support to the idea that CT-α-syn disrupts mitochondrial membrane, leading to cyto c release.
To understand the in uence of C-terminal truncation on membrane interactions, we used DLS to monitor formation of POPC/CL liposome clusters in the presence of FL-and CT-α-syn. FL-α-syn barely clusters liposomes while CT-α-syn dramatically increases particle size ( Figure 4D and S3). The results agree with our NMR data showing that CT-α-syn interacts more strongly with liposomes than FL-α-syn. Hence, the ability of CT-α-syn to cluster liposomes suggests a stronger membrane interaction, which could disrupt membranes.
To shed light on chaperone binding and CT-α-syn aggregation, we used Thio avin T (ThT) uorescence to monitor the rate of CT-α-syn aggregation in the absence or presence of an equal molar ratio of PDI and compared the rate to that of FL-α-syn. The ß-sheet structure of amyloid brils binds ThT dye enhancing uorescence (49). Figure 4E shows a typical nucleation-aggregation pro le, beginning with a lag phase, followed by an elongation phase, and ending in a plateau. In the absence of PDI, truncation of the Cterminus accelerates CT-α-syn aggregation with a decreased lag time compared to FL-α-syn ( Figure 4E, Table 1). This result points to an auto-inhibitory role for the C-terminus. As expected, the presence of PDI inhibits FL-α-syn aggregation(46, 50, 51), but inhibits CT-α-syn aggregation to a somewhat smaller extent, as re ected by an over 3-fold increase in lag time for FL-α-syn whereas less than 2-fold for CT-αsyn. Besides, half-time of aggregation (t 1/2 ) clearly show that PDI exerted a less inhibitory effect on CT-αsyn aggregation ( Figure 4F), though the seg-1 of CT-α-syn interacts more strongly with PDI.

Discussion
Misfolded, oligomeric and brillar α-syn contribute to neuronal toxicity and progression of PD (52)(53)(54). The proposed transient long-range interaction between the N-terminus and acidic tail of α-syn may inhibit bril formation (55). Releasing these contacts destabilizes the natively unfolded monomers potentiating aggregation (35,56). For instance, Ca 2+ , Cu 2+ and polyamine binding alters α-syn's long-range interaction exposing the amyloidogenic NAC region, resulting in enhanced brillation (57). We observe a similar conformational change for CT-α-syn. The results suggest that deleting the C-terminus disrupts the intramolecular long-range interaction between the N-and C-terminus such that CT-α-syn adopts a relatively extended conformation in which the N-terminus is more exposed (Figure 3B), resulting in faster aggregation of CT-α-syn compared to the full length protein. Moreover, the exposed N-terminus of monomeric CT-α-syn has a greater chance of recruiting brils(58). Consistent with previous reports, the Cterminus plays an auto-chaperoning role by shielding the amyloidogenic NAC region(20, 59-61).
Our nding that truncating the C-terminal region facilitates membrane binding contradicts a report purporting to show that the deletion has no effect(62). However, our NMR data ( Figure 1E) show attenuated signals of CT-α-syn in the presence of liposomes compared to FL-α-syn, supporting our cootation data ( Figure 1C-D) indicating more membrane bound CT-α-syn. The DLS data ( Figure 4D) also demonstrate a stronger interaction of CT-α-syn with membrane. α-syn binds negatively charged membranes via electrostatic interactions between the membrane surface and the positively charged Nterminus(63). Hence, exposing the N-terminus increases anionic membrane binding.
Membrane interaction is a common cause of membrane disruption by the intrinsically disordered proteins(64). After establishing the stronger interaction of CT-α-syn with arti cial mitochondrial membranes, we turned to the physiological effect on mitochondrial membrane potential in living SK-N-SH cells. Using JC-1 staining, the data ( Figure 4A-B) show CT-α-syn collapses the potential, which coincides with cyto c release ( Figure 4C). By contrast, FL-α-syn does not impact cyto c release or membrane potential. α-syn causes mitochondrial fragmentation, via direct interaction with mitochondrial membranes (65, 66). These fragmented mitochondria are present in neurons along with α-syn deposits(67). The stronger interaction of CT-α-syn with mitochondrial membranes may explain its disruptive ability.
Molecular chaperones help dissolve α-syn aggregates and prevent formation of pathological brils. For instance, Hsp70, a mammalian chaperone, breaks α-syn brils into monomers or oligomers (39,40). The common chaperone binding region of α-syn is in the N-terminus. Our data showing that PDI binds two segments in the N-terminus (Figure 2A-C) are consistent with the result of Yagi-Utsumi et al (41). In addition, we nd that segment 1 in CT-α-syn binds more tightly to PDI, while segment 2 has a similar binding a nity ( Figure 2D). The stronger interaction of segment 1 diminishes the inhibitory effect of PDI on CT-α-syn aggregation ( Figure 4E-F). Segment 2, comprising residues 36-42, is the master controller of α-syn aggregation at neutral pH(68). We hypothesize that the stronger interaction of PDI with segment 1 in CT-α-syn is the reason for the diminished inhibitory effect on aggregation. To test this idea, we changed Tyr39 to Glu (Y39E) in segment 2 to weaken its a nity for PDI. The mutation abolishes the interaction of segment 2 with PDI whereas segment 1 still bound tightly ( Figure S4A-C). While t 1/2 of the variant in the absence of PDI was too long to measure, the ThT uorescence assay indicates an increased aggregation rate in the presence of PDI, with t 1/2 of about 45 h (Table, Figure S4D-E). This observation suggests that segment 2 is a controller and segment 1 acts as accelerator of brillation. Our results reinforce the importance of residues near Tyr39 in manipulating aggregation, reveal the differential effect of segment 1 and 2 on the brillation upon binding PDI, and may provide a clue for chaperone-based design of brillation inhibitors.
Our results along with those others demonstrate that the C-terminus maintains α-syn's normal physiological function. It appears that the C-terminal region evolved to balance the interactions of the Nterminus with the membranes; strong interactions disrupt membranes causing toxicity, while weak interactions hobble neurotransmitter release. Chaperones help cells handle misfolded proteins, but when the C-terminus is truncated, α-syn brillation spins out of control. We propose a mechanism for how CT-αsyn might promote PD ( Figure 5). Long-range interaction between the N-and C-terminal regions keep αsyn in its normal physiology-relevant conformation. Uncontrolled brillation occurs upon C-terminal truncation. Moreover, the stronger interaction between the CT-α-syn and mitochondrial membrane induces collapse of the mitochondrial membrane potential causing membrane permeability, cyto c release and buildup of mitochondrial reactive oxygen species (mtROS), activating the apoptosome and NLRP3 in ammasome(69-74). The activation of the in ammasome causes more truncation of FL-α-syn in vivo resulting in a negative-feedback loop, amplifying truncation, which elicits neuronal cell death and PD (75)(76)(77). Our observations show the importance of maintaining intact α-syn in cells. We propose that reducing C-terminal truncation is a plausible way to prevent the onset or development of PD.

Conclusion
Truncating the C-terminus releases long-range interaction between domains. CT-α-syn exhibits a more extended conformation, resulting in stronger interactions with membranes and molecular chaperones. The truncation also causes mitochondrial damage and accelerates brillation, both are key aspects of pathogenesis. The differential inhibitory brillation effects of PDI on FL-and CT-α-syn reveal opposite roles for the two segments in brillation. The C-terminus regulates, either directly, via metal ion-and polyamine-binding, or indirectly by modulating membrane and chaperone interactions. Once truncated, its control over the N-terminus is released, triggering mitochondrial damage and pathological brillation. In summary, the C-terminal domain is a 'guard'. In other words, the C-terminal domain plays an autoinhibitory role in FL-α-syn by shielding the N-terminus, weakening interactions with membranes and chaperones. Our work also suggests that preventing truncation is a strategy for preventing the onset or development of PD and synucleinopathies(26, 27, 33, 78).

Materials And Methods
Protein expression and puri cation. α-syn mutants containing a deletion of the C-terminus (amino acids 1-99), single cysteine mutants and single glutamic acid mutant were generated by using site-directed mutagenesis. Natural abundance and 15  . FL-α-syn and FL-A90C were puri ed as described (79). For CT-α-syn and CT-A90C, cells were resuspended in 1 mM phenylmethanesulfonyl uoride (PMSF), 50 mM MES, pH 6.0 and lysed in a highpressure homogenizer (ATS Engineering). Cell lysates centrifuged at 48000g for 30 min at 4 °C. The supernatant was introduced into a SP cation exchange column (GE Healthcare) and the column eluted in a gradient of 0-1000 mM NaCl. Fractions containing the proteins were concentrated after SDS-PAGE analysis and then chromatographed on a Superdex 75 26/600 column (GE Healthcare) eluted with 50 mM MES and 200 mM NaCl, pH 6.0.
The human PDI gene was cloned into a pET-28a vector containing a His6-tag and a TEV protease cleavage site at the N terminus. The recombinant vector was transformed into E. coli BL21 (DE3). The cells were cultured in LB medium at 37 °C. When cell density reached an OD 600 of 0.8, PDI expression was induced with 1 mM IPTG for 7 h. Cells harvested by centrifugation were resuspended in 20 mM Tris, 500 mM NaCl, pH 8.0 with freshly prepared 10mg/L protease inhibitors (Sigma-Aldrich). After lysis and centrifugation, the soluble fraction applied to a HisTrap column (GE Healthcare), which was eluted with a linear gradient of 0-500 mM imidazole. Protein was further puri ed by anion exchange chromatography with a Resource Q column (GE Healthcare) using 20 mM Tris, pH 8.0 and eluted by a linear gradient of 0-1000 mM NaCl, followed by size-exclusion chromatography with a Superdex 200 16/600 column (GE Healthcare) equilibrated with 20 mM Tris, 300 mM NaCl, pH 8.0. Protein concentration was determined either by absorbance of aromatic amino acids at 280 nm with a NanoDrop spectrophotometer (extinction coe cients is 5.96 and 1.49 for FL-and CT-α-syn, respectively), or by Bradford assay using BSA as a standard. Puri ed proteins were desalted using a HiTrap column (GE Healthcare), lyophilized and stored at -80 °C.
Liposome co-otation assay. Liposomes were incubated with FL-or CT-α-syn for 30 min at room temperature. An equal volume of iohexol [80% (w/v)] was added. The sample in 40% (w/v) iohexol was transferred to a thick-wall centrifugation tube (Beckman) before 300 μl of liposome buffer, 375 μl of 30% (w/v) iohexol and 375 μl of 35% (w/v) iohexol were added in that order without mixing. A SW60Ti rotor (Beckman) was used for gradient centrifugation at 28000g for 150 min at 4 °C. Lipid bound proteins were recovered from the 0% and 40% iohexol. The mixture was then deconstructed from the top into nine equal fractions prior to SDS-PAGE analysis. Image J software (National Institutes of Health) was used to quantify band intensity. dissolved in 20 mM HEPES, 100 mM NaCl, 5 mM tris (2-carboxyethyl) phosphine (TCEP), pH 7.0, 10% (v/v) D 2 O in the absence, or presence of PDI at mole ratios from 10:1 to 1:1. Experiments were carried out at 15 °C on a Bruker Avance 800 or 850 MHz NMR spectrometer. Resonance assignments for α-syn are available from the BioMagResBank (entry number 16543). Data were analyzed with Sparky software. Intensity ratios were calculated by peak heights in the presence and absence of liposomes or PDI. CSPs of backbone amides were calculated according to Eq. 1(81), where Δδ H and Δδ N denote the chemical shift difference in the absence and presence of PDI in the 1 H and 15 N dimension, respectively. Where I dia , I para denote NMR signal intensity in the diamagnetic and paramagnetic state, respectively, the interval time T b −T a , was 17.2 ms for FL-α-syn and 20 ms for CT-α-syn.
Dynamic light scattering (DLS). FL-or CT-α-syn (20 μM) were incubated with 100 μM liposomes in a total volume of 1 ml of NMR buffer at room temperature for 30 min. Clustering activity was measured by DLS using a Protein Solutions DynaPro instrument (Malvern) in triplicate, with an average of 10 data points. Particle size distribution of liposomes for control experiments was monitored under same conditions. Cyto c release. At least 10 7 SK-N-SH cells were used per experiment. The mitochondria were prepared with a Mitochondria Isolation Kit (Solarbio) for cultured cells. The procedure is described previously(84). Brie y, freshly isolated mitochondria were resuspended in ice cold buffer (10 mM HEPES, 2 mM K 2 HPO 4 , 10 mM succinate, 250 mM sucrose, 1 mM ATP, 0.08 mM ADP, 1 mM DTT). Aliquots were incubated with FL-and CT-α-syn at 37 °C for 1h. After incubation, the mixture was centrifuged at 13000g at 4 °C for 10 min and the supernatant interrogated using SDS-PAGE followed by WB with an anti-cyto c monoclonal antibody (Abcam).
Mitochondrial membrane potential. SK-N-SH cells were incubated with 50 μM FL-and CT-α-syn 37 °C for 5 h. The JC-1 kit (Sigma-Aldrich) was used according to manual. Images were acquired with a confocal uorescence microscope (Leica TCS SP8). Fluorescence intensity was analyzed with Image J software.

Figure 5
Mechanism of CT-α-syn involvement in the pathogenesis of PD.
Interaction between FL-α-syn and membranes is critical for neurotransmitter release, and brillation of FLα-syn is inhibited by chaperones. When the C-terminus is truncated, the interaction with membranes and chaperones becomes stronger, leading to diminished inhibition of brillation and mitochondrial damage.
The in ammasome is then activated causing truncation of more α-syn, forms a vicious cycle that contributes to pathogenesis.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download. CtruncationSI20220112.docx