The Involvement of His50 during Protein Disulfide Isomerase Binding

Rapid Report pubs.acs.org/biochemistry

The Involvement of His50 during Protein Disulfide Isomerase Binding Is Essential for Inhibiting α‑Syn Fibril Formation Priyatosh Ranjan and Ashutosh Kumar* Department of Bioscience and Bioengineering, Indian Institute of Technology Bombay, Mumbai 400076, India S Supporting Information *

magnetic resonance (NMR) spectroscopy and identified H50 as one of the important key residues involved in PDI binding. To understand the role of H50, NMR interaction studies were conducted with the H50Q familial mutant.14 We also investigated the aggregation of α-Syn WT and H50Q in the absence and presence of PDI using thioflavin T (ThT) fluorescence to understand aggregation kinetics and circular dichroism (CD) for structural conversion. Further, morphological characterizations of the aggregates were performed using atomic force microscopy (AFM). The 1H−15N heteronuclear single-quantum correlation (HSQC) two-dimensional (2D) NMR experiment represents the gold standard for monitoring residue-specific intermolecular interactions in proteins.15 Backbone amide protons are the most sensitive probes to variations in the chemical environment because of the intermolecular interaction that can be monitored by recording 2D 1H−15N HSQC spectra. Chemical shift changes between the free and complex state of a molecule are generally analyzed to probe the change in the chemical environment of the corresponding nucleus. For NMR studies, 15 N-labeled α-Syn WT along with the familial mutants H50Q and G51D, and unlabeled PDI were purified (Supporting Information). Low-molecular weight (LMW) forms of α-Syn16 were prepared at a concentration of 250 μM in 20 mM sodium phosphate buffer (pH 6.0). Earlier studies have shown that the LMW preparation of α-Syn mostly contains monomers along with some amount of low-order multimers.17 Even though the physiological pH is 7.4, under pathological conditions, a decrease in pH occurs in neurons affected by various neurodegenerative diseases, including PD, resulting in apoptosis.18 Therefore, the experiments in the work presented here were performed at pH 6.0. An overlay of 1H−15N HSQC spectra for α-Syn in the presence of increasing amounts of PDI exhibited a relatively narrow dispersion in the proton dimension, suggesting that PDI binding does not induce significant structural perturbations in α-Syn. However, binding of PDI caused observable chemical shift perturbations (CSPs) in α-Syn with many peaks shifting and/or completely disappearing during the course of the titration (Figure 1A and Figure S1). Significant line broadening leading to the disappearance of peaks was observed for residues V3−S9 and L38−V40. This indicated a strong binding (KD = 1−8 μM) (Figure S13) in the intermediate exchange regime19,20 (Figure 1C and Figure S2). Significant perturbations in the cross-peaks

ABSTRACT: An increased level of protein disulfide isomerase (PDI) is a protective response to various neurodegenerative disorders, including Parkinson’s disease. Interaction of PDI with α-synuclein (α-Syn) has been shown to inhibit its aggregation. Here, we report the residue-specific mapping of binding of PDI to α-Syn. We demonstrate that α-Syn N-terminal residues V3−S9 and L38−V40 bind more strongly to PDI than residues V49− V52 do, as do C-terminal residues E123−M127 and D135−E137. In addition, we show that residue H50 is key in preventing aggregation. These findings improve our understanding of PDI-protected aggregation of wild-type α-Syn and its H50Q familial mutant.

P

arkinson’s disease (PD) is the second most common neurodegenerative disorder, and the characteristic αsynuclein (α-Syn) fibrillar aggregates, also known as Lewy bodies, are found in the neurons of the substantia nigra.1 α-Syn is a small (140 residues), natively unfolded presynaptic protein of ∼14 kDa.2 It acquires α-helical secondary structure upon binding to lipid vesicles3 or the characteristic cross β-sheet conformation in the case of amyloid-like fibrils.4 The accumulation of unfolded or misfolded proteins leads to the increased levels of chaperones and foldases in the endoplasmic reticulum (ER),5 which is a salient pathological feature associated with various neurodegenerative disorders.6 Chronic ER stress, inducing cell death, has been reported during the overexpression of wild-type (WT)7,8 and mutant α-Syn.9 An elevated level of protein disulfide isomerase (PDI), a multifunctional stress protein abundant in the ER, has been reported in the brain of patients with PD and is found in Lewy bodies.10 PDI is a 55 kDa soluble protein that contains four thioredoxin-like βαβαβαββα domains (termed a, b, b′, and a′), a linker (x), and a C-terminal extension domain (c), which are arranged in the order abb′xa′c.11 Cheng et al.12 for the first time reported interaction between PDI and its domains with α-Syn and showed that a′ domain of PDI is essential and sufficient for inhibiting α-Syn fibril formation. Recently, Yagi-Utsumi et al.13 characterized substrate recognition by Humicola insolens (softrot fungus) PDI using human α-Syn as the model ligand; they reported interaction of α-Syn with PDI through its specific hydrophobic segment. However, detailed residue-specific mapping of binding of PDI to α-Syn remains unexplored. Probing the PDI binding sites on α-Syn will help us understand how PDI inhibits its aggregation. In this report, we have identified the putative PDI binding sites on α-Syn using nuclear © XXXX American Chemical Society

Received: March 28, 2016 Revised: May 2, 2016

A

DOI: 10.1021/acs.biochem.6b00280 Biochemistry XXXX, XXX, XXX−XXX

Biochemistry

Rapid Report

Figure 1. NMR chemical shift mapping of PDI binding sites on α-Syn WT and its familial mutant H50Q. Excerpt of the superimposed 1H−15N HSQC spectra of (A) α-Syn WT and (B) H50Q in the absence (colored black) and presence of different numbers of equivalents of PDI (0.1, green; 0.25, magenta; 0.5, red; 1, blue). I/I0 profiles of the 1H−15N HSQC NMR signals of (C) α-Syn WT and (D) H50Q in the presence of an equivalent concentration of PDI. Histograms of the chemical shift perturbation of (E) α-Syn WT and (F) H50Q by addition of PDI. PDI-induced chemical shift changes were calculated as [(5ΔδHN)2 + (Δδ15N)2]1/2, where ΔδHN and Δδ15N are the perturbations in the amide proton and nitrogen chemical shifts, respectively. Residues showing significant perturbations have been marked. Compared to α-Syn WT, no significant shift was observed for residues surrounding H50 in the case of the H50Q familial mutant.

Figure 2. Dynamics and aggregation studies of α-Syn in the absence and presence of PDI. 15N transverse relaxation rates (R2) for (A) α-Syn WT and (B) H50Q in the absence (black) and presence (red and blue) of PDI. (C) Aggregation kinetics of α-Syn WT and its H50Q familial mutant in the absence and presence of PDI monitored by ThT fluorescence. (D) CD spectra showing secondary structural changes during fibrillation of α-Syn WT and H50Q in the absence and presence of PDI. Over time, an increase in β-sheet content was observed for WT, H50Q, and H50Q in the presence of PDI. α-Syn WT in the presence of PDI showed no significant change in secondary structure and remained random coil during the course of aggregation.

observed with an increasing number of equivalents of PDI (Figure S3). PDI titration monitored by NMR showed that the residues encompassing H50 were the most perturbed, suggesting that the histidine residue could play a critical role in PDI binding. Previously, H50 of α-Syn was reported to play an important role in various metal binding studies.22,23 To understand the role of H50 in PDI binding, interaction studies were performed on the newly discovered H50Q familial mutant,14 where histidine at position 50 is replaced with glutamine. A series of 1H−15N HSQC spectra of 250 μM H50Q at pH 6.0 in the presence of increasing concentrations of

were observed for residues V49−V52 and residues at the Cterminus (E123−M127 and D135−E137), indicating weak binding in the fast exchange regime on the NMR time scale21 (Figure 1E and Figure S3). Amide cross-peak intensities and CSPs were also measured at substoichiometric concentrations of PDI to match the physiologically relevant concentration. Even at lower concentrations of PDI, amide cross-peaks for residues located at the N-terminus (residues 3−9 and 38−40) initially shifted and then broadened as indicated by lower I/I0 values (Figures S2 and S3). However, for residues 49−52 and residues located at the C-terminus, an increase in CSP was B

DOI: 10.1021/acs.biochem.6b00280 Biochemistry XXXX, XXX, XXX−XXX

Biochemistry

Rapid Report

phase (Figure 2C). CD spectroscopy was performed simultaneously along with ThT binding to probe the secondary structural changes during the course of aggregation. During amyloid fibril formation, α-Syn undergoes a conformational transition from its intrinsically disordered state to the β-sheet, as seen in CD experiments. Initially, both WT and H50Q, in the absence and presence of PDI, showed negligible ThT binding with random coil structure as indicated by the minima at ∼200 nm in CD spectra (Figure 2D). Over time, H50Q showed faster aggregation compared to that of WT as reported previously.16 H50Q took ∼18 h to form β-sheet, while WT converted to β-sheet at ∼40 h as seen in CD spectra. The aggregation kinetics of WT was significantly inhibited in the presence of PDI. Throughout the course of the aggregation process, the CD spectrum of α-Syn in the presence of PDI remained largely that of a random coil. However, in the case of H50Q, the aggregation profile was almost similar in the absence and presence of PDI. The aggregation lag times of H50Q in the absence and presence of PDI were calculated to be ∼13 and ∼17 h, respectively (Figure S8). The studies mentioned above suggest that H50 of α-Syn plays an important role in PDI binding, and this interaction appears to be essential for inhibiting fibril formation. The aggregation studies were also performed at pH 7.4, and the pattern was found to be similar to that at pH 6.0 (Figure S9). Under these experimental conditions, PDI alone did not contribute to fibril formation as indicated by ThT fluorescence and CD spectroscopy (Figure 2C and Figure S10A,B). The morphology of the aggregates generated in the absence and presence of PDI was monitored by AFM (Figure 3 and Figure S11). The aggregates of α-Syn in

PDI were recorded (Figure 1B and Figure S4). Similar to the case for WT, peaks corresponding to residues V3−S9 and L38−V40 initially shifted and then showed a significant reduction in intensity upon addition of an equivalent concentration of PDI (Figure 1D and Figures S2 and S3). An increase in the CSP was observed for the residues at the Cterminus (E123−M127 and D135−E137) with increasing concentrations of PDI (Figure 1F and Figure S3). However, compared to the case for WT, there was no change in chemical shifts for neighboring residues of Q50, suggesting that the histidine residue of WT α-Syn plays an important role in PDI binding. Moreover, the H50Q mutation does not affect the binding at the N- and C-termini. To determine the importance of other residues present in the region of residues 49−52, PDI titration experiments were performed in the case of another familial mutant, G51D.24,25 G51D showed a binding pattern similar to that of WT, suggesting that the G51D mutation does not abolish the binding in the region of residues 49−52 (Figure S5). The same sets of NMR experiments were also performed at pH 7.4 to match the physiological pH condition, and the binding pattern was found to be similar (Figures S6 and S7 and results of the Supporting Information). However, a difference in the strength of binding can be due to the fact that at neutral pH transient long-range interactions between the N- and Ctermini stabilize a closed conformation.26,27 At slightly acidic pH (i.e., pH 6.0), protonation of the acidic C-terminus can weaken these interactions and binding of PDI become comparatively stronger. The local mobility changes upon binding of PDI to α-Syn WT and its familial mutant H50Q were monitored by 15N transverse relaxation rates (R2). R2 rates are sensitive to millisecond to microsecond time scale motions and contain exchange contributions (Rex).19,28 Thus, R2 can measure significant effects from the rate of exchange between the free and bound forms. Upon addition of PDI to α-Syn WT (1:0.5 αSyn:PDI), an increase in the average R2 from 6.0 ± 0.5 s−1 to 10.1 ± 0.7 s−1 was observed for the residues at the N-terminus, namely, V3, F4, K6, L8, and S9 (Figure 2A). Similarly, an increase in R2 was also observed for Val 40, which showed prominent perturbation and line broadening during NMR titration experiments. Enhanced R2 values may be attributed to the chemical exchange between the free and bound forms and/ or a higher rigidity of the α-Syn backbone induced upon PDI binding. No such significant change in R2 was observed for residues 49−52 as well as residues at the C-terminus. This may be due to the fact that the chemical shift differences between the free and bound forms for these residues are not comparable to the exchange rate. In the case of the H50Q familial mutant, similar to the case for WT, an increase in the average R2 from 4.2 ± 0.2 s−1 to 8.7 ± 0.7 s−1 was observed for V3−S9 and V40 (Figure 2B). The assembly and propagation of amyloid fibrils formed in vitro were monitored by the binding of the amyloid-specific ThT dye that produces strong fluorescence upon binding to fibrillar species. To understand the effect of PDI binding on αSyn aggregation kinetics, the ThT binding assay was performed. For aggregation studies, the LMW forms of α-Syn WT and H50Q at 250 μM were prepared and incubated at 37 °C in 20 mM phosphate buffer containing 0.01% sodium azide (pH 6.0) in the absence and presence of an equivalent concentration of PDI. α-Syn aggregation follows a classical nucleation-dependent polymerization reaction in which an initial lag phase is followed by the rapid growth of amyloid fibrils, ending in a stationary

Figure 3. Aggregates of (A) α-Syn WT and (B) H50Q formed in the absence (left) and presence (right) of PDI monitored by AFM. Scale bars are 500 nm.

the absence of PDI revealed the formation of long, highly ordered dense fibrillar networks (Figure 3A, left panel, Figure S11A, and Figure S12A). However, in aggregation reaction mixtures containing both PDI and α-Syn, long fibrillar networks were rarely observed (Figure 3A, right panel, and Figure S11B). In the case of the H50Q familial mutant, fibrillar networks were seen both in the absence and in the presence of PDI, suggesting that PDI binding did not inhibit fibril formation (Figure 3B, Figure S11C,D, and Figure S12C,D). Amorphous aggregates were observed in the case of the samples containing PDI only, suggesting that PDI itself did not form the fibril during the course of aggregation (Figure S10C). In conclusion, we found that the N-terminus of monomeric α-Syn showed (V3−S9 and L38−40) a higher affinity for PDI (Kd = 1−8 μM) compared to that of residues V49−V52 (Kd = 26−32 μM) and that of the residues located at the C-terminus (Kd = 37−52 μM) (for details, see Figure S13). Previous studies have identified five β-strands within the fibril core of αSyn comprising residues 30−110.29,30 Residues 49−52 are suggested to adopt β-turn conformation during fibril formation. Even though the binding of PDI within this region is weak, the C

DOI: 10.1021/acs.biochem.6b00280 Biochemistry XXXX, XXX, XXX−XXX

Biochemistry

Rapid Report

binding is crucial for preventing the formation of β-strand within the core region and hence fibril formation. In the case of the H50Q familial mutant, where no PDI binding is taking place surrounding Q50, fibril formation is not perturbed even though the binding at the N- and C-termini is similar to WT. From the studies mentioned above, it can be postulated that binding of PDI within the residues surrounding H50 along with the residues at the N- and C-termini stabilizes the native conformation of α-Syn and does not allow it to form an unstable misfolded intermediate, thus inhibiting the fibril formation.

■

(7) Cooper, A. A., et al. (2006) Science 313, 324−328. (8) Gitler, A. D., et al. (2008) Proc. Natl. Acad. Sci. U. S. A. 105, 145− 150. (9) Smith, W. W., et al. (2005) Hum. Mol. Genet. 14, 3801−3811. (10) Conn, K. J., et al. (2004) Brain Res. 1022, 164−172. (11) Kemmink, J., et al. (1997) Curr. Biol. 7, 239−245. (12) Cheng, H., Wang, L., and Wang, C. C. (2010) Cell Stress Chaperones 15, 415−421. (13) Yagi-Utsumi, M., Satoh, T., and Kato, K. (2015) Sci. Rep. 5, 13909. (14) Appel-Cresswell, S., et al. (2013) Mov. Disord. 28, 811−813. (15) Williamson, M. P. (2013) Prog. Nucl. Magn. Reson. Spectrosc. 73, 1−16. (16) Winner, B., et al. (2011) Proc. Natl. Acad. Sci. U. S. A. 108, 4194−4199. (17) Ghosh, D., et al. (2013) Biochemistry 52, 6925−6927. (18) Harguindey, S., et al. (2007) Curr. Alzheimer Res. 4, 53−65. (19) Cavanagh, J., Fairbrother, W. J., Palmer, A. G., and Mark, R. (1996) Protein NMR Spectroscopy, Academic Press, San Diego. (20) Ziarek, J. J., et al. (2011) Methods Enzymol. 493, 241−275. (21) Vaynberg, J., and Qin, J. (2006) Trends Biotechnol. 24, 22−27. (22) Rasia, R. M., et al. (2005) Proc. Natl. Acad. Sci. U. S. A. 102, 4294−4299. (23) Lee, J. C., Gray, H. B., and Winkler, J. R. (2008) J. Am. Chem. Soc. 130, 6898−6899. (24) Kiely, A. P., et al. (2013) Acta Neuropathol. 125, 753−769. (25) Lesage, S., et al. (2013) Ann. Neurol. 73, 459−471. (26) Dedmon, M. M., et al. (2005) J. Am. Chem. Soc. 127, 476−477. (27) Bertoncini, C. W., et al. (2005) Proc. Natl. Acad. Sci. U. S. A. 102, 1430−1435. (28) Kay, L. E., Torchia, D. A., and Bax, A. (1989) Biochemistry 28, 8972−8979. (29) Heise, H., et al. (2005) Proc. Natl. Acad. Sci. U. S. A. 102, 15871−15876. (30) Vilar, M., et al. (2008) Proc. Natl. Acad. Sci. U. S. A. 105, 8637− 8642.

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.biochem.6b00280. Detailed description of materials and methods and supplementary figures (PDF)

■

AUTHOR INFORMATION

Corresponding Author

*Phone: 91-22-2576-7762. E-mail: [email protected]. Author Contributions

P.R. and A.K. designed the research. P.R. performed the experiments. P.R. and A.K. analyzed the data and wrote the manuscript. Funding

This work was supported by a Ramalingaswamy re-entry fellowship (BT/RLF/Re-entry/22/2010) from the Department of Biotechnology, Government of India, to A.K. P.R. is grateful to MHRD (Government of India) for his fellowship. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The authors acknowledge the HF NMR and Bio-AFM facility, funded by RIFC, IRCC, Indian Institute of Technology Bombay (IIT Bombay). We are also very thankful to the NMR Research Center, IISER, Pune. We thank Prof. Samir K. Maji (IIT Bombay) for the clones of α-Syn WT, H50Q, and G51D. We are also very thankful to Prof. Lloyd Ruddock (University of Oulu, Oulu, Finland) for sharing the PDI clones. We thank Prof. R. V. Hosur (UM-DAE CEBS, Mumbai University campus, Mumbai, India) for providing useful suggestions during the preparation of the manuscript. We thank Prof. Jeetender Chugh for help in conducting the NMR experiments at IISER, Pune. We thank Dr. Vaibhav Kumar Shukla and Dr. Sheeja V. Vasudevan for helping in data analysis. We thank Dr. Sarath Chandra Dantu and Nikita Malik for critical reading of the manuscript.

■

REFERENCES

(1) Spillantini, M. G., et al. (1997) Nature 388, 839−840. (2) Weinreb, P. H., et al. (1996) Biochemistry 35, 13709−13715. (3) Davidson, W. S., et al. (1998) J. Biol. Chem. 273, 9443−9449. (4) Serpell, L. C., et al. (2000) Proc. Natl. Acad. Sci. U. S. A. 97, 4897−4902. (5) Schröder, M., and Kaufman, R. J. (2005) Mutat. Res., Fundam. Mol. Mech. Mutagen. 569, 29−63. (6) Lindholm, D., Wootz, H., and Korhonen, L. (2006) Cell Death Differ. 13, 385−392. D

DOI: 10.1021/acs.biochem.6b00280 Biochemistry XXXX, XXX, XXX−XXX