T Residues of the Human Chaperone DNAJB6 Are

Jul 19, 2018 - The development of a protocol to achieve highly reproducible kinetic ... of a few more residues from APP,(54) which can be overcome usi...
1 downloads 0 Views 2MB Size
Subscriber access provided by READING UNIV

Article

Conserved S/T-residues of the human chaperone DNAJB6 are required for effective inhibition of A#42 amyloid fibril formation Cecilia Månsson, Remco T. P. van Cruchten, Ulrich Weininger, Xiaoting Yang, Risto Cukalevski, Paolo Arosio, Christopher M. Dobson, Tuomas P.J. Knowles, Mikael Akke, Sara Linse, and Cecilia Emanuelsson Biochemistry, Just Accepted Manuscript • DOI: 10.1021/acs.biochem.8b00353 • Publication Date (Web): 19 Jul 2018 Downloaded from http://pubs.acs.org on July 20, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 25 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

Conserved S/T-residues of the human chaperone DNAJB6 are required for effective inhibition of Aβ β 42 amyloid fibril formation Cecilia Månsson1, Remco T P van Cruchten1, Ulrich Weininger2, Xiaoting Yang1, Risto Cukalevski1, Paolo Arosio3, Christopher M Dobson3, Tuomas Knowles3,5, Mikael Akke2, Sara Linse1 and Cecilia Emanuelsson1* 1 Department of Biochemistry and Structural Biology, Center for Molecular Protein Science, Lund University, Lund, Sweden, 2Department of Biophysical Chemistry, Center for Molecular Protein Science, Lund University, Lund, Sweden, 3Department of Chemistry and Center for Misfolding Diseases, University of Cambridge, Cambridge, UK 4Current address: Department of Chemistry and Applied Biosciences, ETH Zurich, Switzerland. 5Cavendish Laboratory, Department of Physics, JJ Thomson Avenue, CB3 0HE, Cambridge UK * Corresponding author: E-mail: [email protected] ABSTRACT: The human molecular chaperone DNAJB6, an oligomeric protein with a conserved S/T-rich region, is an efficient suppressor of amyloid fibril formation by highly aggregation-prone peptides such as the Aβ and polyQ peptides associated with Alzheimer’s and Huntington’s disease, respectively. We previously showed that DNAJB6 can inhibit the processes through which amyloid fibrils are formed due to strong interactions with aggregated forms of Aβ42 that become sequestered. Here we report that the concentration-dependent capability of DNAJB6 to suppress fibril formation in ThT fluorescence assays decreases progressively with an increasing number of S/T substitutions, with an almost complete loss of suppression when 18 S/T-residues are substituted. The kinetics of primary nucleation in particular are affected. No detectable changes in the structure are caused by the substitutions. Also the binding of DNAJB6 to Aβ42 decreases with the S/T substitutions, as determined by surface plasmon resonance and microscale thermophoresis. The aggregation process monitored using NMR spectroscopy showed that DNAJB6, in contrast to a mutational variant with 18 S/T-residues substituted, can maintain monomeric Aβ42 soluble for an extended time. The inhibition of the primary nucleation is likely to depend on hydroxyl groups in side chains of the S/T-residues and hydrogen bonding with Aβ42 is one plausible molecular mechanism although other possibilities cannot be excluded. The loss of ability to suppress fibril formation upon S/T to A substitution is previously observed also for polyQ peptides, suggesting that the S/T-residues in the DNAJB6-like chaperones have a general ability to inhibit amyloid fibril formation by different aggregation-prone peptides.

INTRODUCTION Neurodegenerative disorders encompass a group of more than 50 different syndromes 1 linked to protein aggregation, including Alzheimer’s disease (AD), the most prevalent form of human dementia. The self-assembly of the amyloid-β (Aβ) peptide into amyloid fibrils and smaller oligomeric aggregates is strongly implicated in this condition and to memory loss and cognitive impairment 2–6. The Aβ peptide is a cleavage product of the Amyloid β Precursor Protein (APP) and two of the more prominent forms consist of 40 (Aβ40) and 42 (Aβ42) residues. The Aβ40 peptide is about ten times more abundant than Aβ42 but the latter is more aggregation prone and the primary component of amyloid plaques found in the brains of AD patients 7–9. Toxic oligomers are believed to be causative agents for AD 10,11, whereas fibrils can serve as catalytic surfaces for oligomer generation 12 and are thus likely to be involved in the spreading of the disease through tissue 13–15 although the exact size, shape and composition of the disease-causing species remain unclear 16,17. The transition from a supersaturated solution of a monomeric peptide or protein to a solution composed mainly of

ACS Paragon Plus Environment

Biochemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

fibrils is a complex process involving several mechanistic steps and strongly dependent on the environment. Amyloid fibril formation is a nucleated self-assembly process 18,19, where initial aggregates are formed by primary nucleation and subsequently elongate by addition of monomers. In addition, new aggregates can also be formed by secondary processes, involving fibril fragmentation or surface-catalyzed nucleation, the latter being particularly important in the case of Aβ 12,20 as well as several other amyloidogenic proteins such as IAPP 21 and αsynuclein 22. Extensive effort has been put into the identification and characterization of inhibitors of amyloid fibril formation, with reported inhibitors ranging from small molecules 23 such as polyphenols or lipid-based compounds to proteins antibodies 24 that have been shown to inhibit Aβ fibril formation in vitro and certain molecular chaperones 25–27. Under cellular conditions a network of molecular chaperones safeguards the proteome from misfolding, aggregation and proteotoxicity. A subset of 20-30% of the approximately 300 human chaperones have been found to have altered expression levels in the human brain during aging and in Huntington’s, Alzheimer’s and Parkinson’s diseases 15,28. The human chaperone DNAJB6 was identified as an efficient suppressor of polyQ aggregation in a cell-based screen covering numerous different human chaperones 29. We have later shown that sub-stoichiometric concentrations of DNAJB6 can efficiently suppress amyloid fibril formation by both polyQ and Aβ peptides, and that DNAJB6 inhibits very effectively in particular the primary nucleation step 30–32. The physiological relevance of this highly efficient aggregation suppression is emphasized by the very recent findings that DNAJB6 can rescue pathological changes in a mouse model of Huntington’s disease, resulting in the delayed appearance of aggregates in the brain, attenuated symptoms and prolonged lifespan 33. This inhibition of polyQ aggregation is dependent on a unique region which is conserved among DNAJB6 homologues and was designated “the SSF-SST region” when first described in the close DNAJB6-homologue DNAJB8 29. Due to the side chains of the serine (S) and threonine (T) residues, this S/T-rich region exposes an array of up to 18 hydroxyl groups, which is of particular interest in the context of the compound polyphenol (-)-epigallocatechin-3-gallate (EGCG), a potent inhibitor of fibril formation by both Htt-Ex1 34 and Aβ 35, with six hydroxyl groups being crucial for its function 34. The interaction between inhibitors and aggregation-prone peptides is often reliant on multiple intermolecular interactions and may, in addition to hydrogen bonding (as in the case of for example in polyphenols and self-complementary peptide fragments 36), also contain contributions from electrostatic interactions (as in the case of Congo Red), the hydrophobic effect and van der Waals' interactions (as in the case of lipids and selfcomplementary peptide fragments). Each monomer subunit of the DNAJB6 oligomers consists of three domains: (i) the Nterminal domain (residues 1–71), which is α-helical and also called the J-domain, (ii) the middle domain (residues 72–189), which is rich in glycine and phenylalanine residues and probably structurally disordered and (iii) the C-terminal domain (residues 190–241), which is predicted to have predominantly β-sheet structure 31. The S/T-rich region (residues 155–195) is located largely within the disordered middle domain, with a partial overlap with the Cterminal domain. The level of sequence conservation among DNAJ proteins is low, except for the J-domain that defines the DNAJ protein family of over 40 human homologues distributed into the DNAJA, DNAJB and DNAJC subfamilies 37. DNAJB6 belongs to the DNAJB subfamily, whose approximately 15 members show low sequence identity in the C-terminal domain, e.g. only 12% identity between human DNAJB6 and its paralogue DNAJB1 (Hdj1) for which the structure has been determined to atomic resolution 38. The S/T-rich region and 2 ACS Paragon Plus Environment

Page 2 of 25

Page 3 of 25 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

the oligomerization are features absent in other members of the DNAJ protein family and observed only in close homologues to DNAJB6 29. In this study, we used ThT fluorescence together with kinetic analysis to study the conversion of Aβ42 from its initially supersaturated monomeric form into its predominantly fibrillar state in the absence and presence of DNAJB6. The ability to suppress aggregation was found to be reduced with the number of S/T-to-A substitutions and was almost completely abolished when 18 S/T-residues were substituted. Thus we here report a striking similarity in the effect of S/T-substitution on the DNAJB6 chaperone activity, as previously observed for polyQ 33, towards another highly aggregation-prone peptide which is also one of the most well studied, due to the aggressiveness of Aβ42 and its huge impact with increasing incidence rates of AD. This emphasizes that the S/T residues are vital for the DNAJB6 chaperone activity, a result that is important particularly in the context of the common mechanisms. Thermophoresis experiments showed no detectable affinity of DNAJB6 for Aβ42 when 18 S/T-residues were substituted. The ability of DNAJB6 to maintain Aβ42 in solution, monitored also by observing the monomeric Aβ42 population by NMR spectroscopy, was also reduced by S/T-substitution. Our results show that the S/T-rich region is the underlying reason that DNAJB6 is able to inhibit primary nucleation, maintain Aβ42 in solution and suppress the formation of amyloid fibrils. EXPERIMENTAL PROCEDURES Recombinant expression and purification of proteins and peptides. The amino acid numbering used here refers to the amino acid sequence of DNAJB6 isoform B (UniProt KB accession number O75190). Mutants were generated with 6, 13 or 18 S/T to A substitutions in the conserved S/T-rich region (aa 155–195), designated ST6A, ST13A, and ST18A. One deletion mutant (∆132–183) was also generated, with 21 S/T residues and in total 35 residues deleted and designated ∆ST, and two more substitution mutants designated ST26A and ST5A_190-5, with 26 and 5 S/T to A substitutions, respectively. Expression of DNAJB6 and its mutant variants protein was performed at the Lund Protein Production Platform, Lund University, Sweden (http://www.lu.se/lp3) as previously described 31,32. Human Aβ(M1-42), here called Aβ42, (with an N-terminal methionine residue to initiate translation, thus corresponding to UniProtKB accession number P0506, aa 671–713) was stored in aliquots as lyophilized monomers purified from inclusion bodies following bacterial expression as described previously 39. Immediately before use, each aliquot of monomeric Aβ42 was dissolved in 6 M guanidine hydrochloride pH 8.5 and subjected to gel filtration on a Superdex 75 column HR 10/30 column (Amersham Biosciences, Amersham, UK) with a flow rate of 0.7 ml/min of 20 mM sodium phosphate, 200 µM EDTA and 0.02% NaN3, pH 8.0. UV absorbance at 280 nm was used to monitor the elution of the peptides and to determine the concentration using the molar extinction coefficient ε =1400 mol–1 cm–1. Aggregation kinetics. For measurements of Aβ42 fibril formation, a 3 µM solution of Aβ42 in its monomeric form (freshly prepared to ensure the absence of aggregates) was mixed with DNAJB6 or its mutational variants at different molar ratios (1:0.005; 1:0.01; 1:0.05 and 1:0.1) of Aβ42 to DNAJB6 in the same buffer as used for the gel filtration step in the Aβ42 preparation process. All samples were then supplemented with 10 µM Thioflavin T (ThT) for detection of fibrils. Samples were prepared on ice and 70 µl aliquots were transferred to microplate wells (Corning® 96 well plate, half-area black/clear bottom polystyrene, nonbinding surface, Corning 3881 (Corning Incorporated Life Sciences, Acton, USA) in triplicate. The nonbinding (NBS™) surface is a Corning proprietary treatment 3 ACS Paragon Plus Environment

Biochemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

technology used on polystyrene microplates to create a nonionic hydrophilic surface (polyethylene oxide-based) that minimizes molecular interactions. These PEGylated plates were chosen based on the finding that no Aβ42 binding could be detected to PEGylated surfaces by QCMD 40. The ThT fluorescence was measured every 120, 150 or 300 s at quiescent conditions at 37°C using a Fluostar Optima plate reader (BMG Labtech, Offenburg, Germany) with excitation at 440 nm and emission at 480 nm. The fluorescence data were normalized, with the lowest value in the dataset set to 0 and the highest to 1 for each reaction separately. The halftime (t½) was determined from each curve after fitting non-linear sigmoidal dose-response curves (variable slope) to the normalized data at the point in time where the first derivative is at a maximum. Within each experiment these values were divided by the average values of three replicates of Aβ42 reactions without DNAJB6. The ratios obtained were averaged between experiments and the standard error of mean (SEM) was calculated. The concentration of monomeric Aβ42 remaining in solution at different time points along the fibril formation in PEGylated plates at 3 µM total Aβ42 concentration, without or with DNAJB6 or its mutational variants, was determined by collecting samples at the end of the lag phase and after reaching the plateau region, and the monomer concentration estimated by SDS PAGE after removing the fibrils by centrifugation. Kinetic analysis of self-assembly reactions. To evaluate the effect of each of DNAJB6 and its mutational variants the specific microscopic events involved in the aggregation processes, the reaction profiles were fitted to integrated rate laws for filamentous growth 12,18,19,41,42 considering different microscopic rate constants in the absence and presence of different concentrations of molecular chaperone according to the following equation:

where M(t) is the total fibril mass at the time t and the kinetic parameters B±, C±, κ, κ∞ and are functions of the microscopic rate constants k+k2 and knk2, where kn, k+ and k2 are the primary nucleation, fibril elongation and secondary nucleation rate constants, respectively. Microscale thermophoresis (MST). To evaluate the binding between Aβ42 and DNAJB6 with microscale thermophoresis (MST), an aliquot of purified Alexa488-labeled Aβ42 (4 µM) was used to prepare samples in a buffer composed of 20 mM sodium phosphate pH 8.0, 150 mM NaCl, 0.005% Tween-20 with Aβ42 at 0.5 µM and DNAJB6 or its mutational variant ST18A, at 16 concentrations varying from 0 to 50 µM. Samples were placed in low-binding capillaries (MST Premium Coated from Nanotemper Technologies, Munich, Germany) and mounted in a Monolith NT.115 Instrument (Nanotemper Technologies, Munich, Germany) operated at 37 °C. Thermophoresis measurements were repeated in cycles over 8 h using 15% LED power and 20 and 40% MST-power. The data were fitted to calculate the KD values as previously described 43. Nuclear Magnetic Resonance (NMR) spectroscopy. Aβ42 (labeled with 15N and13C) was expressed in minimal medium as described previously 44. All NMR samples contained 20 µM 15 N-13C labeled Aβ42 in 20 mM sodium phosphate buffer, pH 8.0, and a subset of samples also contained 2 µM of either DNAJB6 or its mutational variant S/T18A. Series of 13C-edited 1 H 1D spectra, which filter out contributions from the unlabelled DNAJB6, were acquired over time at a static magnetic field strength of 11.7 T and a temperature of 37 °C. Spectra were processed with NMRPipe 45 and the intensities of methyl groups integrated using Matlab. Using the same basic setup, a series of DEST experiments 46 were run at different 4 ACS Paragon Plus Environment

Page 4 of 25

Page 5 of 25 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

time points, using a saturation time of 400 ms, saturation strengths of 200, 400, and 800 Hz, and offsets between 1–30 kHz. Pulsed-field gradient diffusion experiments were acquired with a diffusion time of 410 ms, gradient lengths of 1 ms and gradient strengths between 1.8– 56.5 G/cm. The amount of Aβ42, and DNAJB6, in solution was determined by acid hydrolysis after filtration to remove fibrils. Electrophoresis. Non-denaturing polyacrylamide gel electrophoresis (PAGE) was performed using precast 3–12% Native-PAGE Bis-Tris Gels (Life Technologies Europe BV, Stockholm, Sweden) and SDS-PAGE was performed with 12% gels, that were stained with colloidal Coomassie Brilliant Blue 47 and scanned with the Image Scanner III using the software Labscan/IQTL (GE Healthcare Lifesciences, Uppsala, Sweden). Circular Dichroism (CD) spectroscopy. CD spectra were recorded in the far-UV range for DNAJB6 and its mutational variants at 0.3 mg/ml or 0.2 mg/ml in CD buffer (5 mMNaPB pH 8 and 40 mMNaF) in quartz cuvettes with an 1 mm path length at 37°C and 75°C using a Jasco J-815 (Jasco Analytical Instruments, Easton, MD) CD spectrometer. Three spectra were recorded between 250 and 185 nm and averaged. The background spectrum of the buffer was subtracted and the molar ellipticity calculated in each case. Cryo-EM imaging. The samples were prepared as described above for recording the aggregation kinetics and collected at the plateau region for 3 µM Aβ42 with or without 0.3 µM DNAJB6 or its mutational variants. Specimens, prepared in a controlled environment vitrification system to ensure a stabilized environment and avoid loss of solvent, were deposited with a thin layer of liquid on top of a lacey carbon filmed copper grid less than 300 nm thick and plunged into liquid ethane (-180°C), then transferred to and stored in liquid nitrogen until imaging. An Oxford CT3500 cryoholder and its workstation was used to transfer the specimens into the electron microscope (Philips CST6A20 BioTWIN Cryo) equipped with a post-column energy filter (Gatan GIF 100). The acceleration voltage was 120 kV. Images were recorded digitally with a CCD camera under low electron dose conditions and fibril measurements done with Digital Graph software (Gatan, Inc., Pleasanton, CA, US). Surface Plasmon Resonance (SPR). Measurements were performed using a Biacore 3000 (GE Healthcare, Uppsala, Sweden) instrument in 20 mM sodium phosphate, 200 µM EDTA, 0.005% Tween20, pH 8.0 at a flow rate of 10 µl/min and Aβ42 immobilized on a CM5 carboxymethylated dextran sensor chip (GE Healthcare, Uppsala, Sweden). One flow channel was blocked directly using ethanolamine to serve as negative control. Immobilization of Aβ42 in the other three flow channels was performed at approximately pH 3 by dilution of freshly prepared monomers into 10 mM sodium acetate pH 2.8. The association was monitored under constant flow of 1 µM of DNAJB6 or its mutational variants for 10 min, then buffer was flowed over the chip for 110 min to study the dissociation from the immobilized Aβ42. To remove background signals and instrument drift, the response of the control flow channel without immobilized Aβ42 was subtracted from the values obtained when the protein was flowed over channels with immobilized Aβ42. From this background-corrected signal the baseline was subtracted. By injecting 10 mM HCl or 2% SDS for 3 min followed by buffer flow for 87 min tightly bound proteins were removed and surface regenerated. RESULTS Effects of S/T to A substitutions on the DNAJB6 inhibition of Aβ42 fibril formation. To investigate the role of the S/T-rich region in DNAJB6, a series of mutational variants was 5 ACS Paragon Plus Environment

Biochemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Fig. 1 The S/T-rich region in the DNAJB6 chaperone and the effects of DNAJB6 and its mutational variants on the process of Aβ42 fibril formation. (A) Outline of the amino acid sequence of DNAJB6, highlighting the region (residues 132-195) rich in serine (S) and threonine (T) residues which were subjected to alanine substitution. (B) ThT fluorescence assays show aggregation reaction profiles fitted by integrated rate laws of 3 µM Aβ42 in the absence and presence of 1:0.005, 1:0.01, 1:0.05 and 1:0.1 equivalents of wildtype DNAJB6 (WT). (C-H) Aggregation reaction profiles as in (B) for the DNAJB6 mutational variants (designated S/T6A, S/T13A, S/T18A, S/T∆, S/T26A and S/T5A_190-5). Note that the scale of the time axis varies between panels and that of B and C is logarithmic. (I) Image showing a model obtained as output after crosslinkassisted docking of one copy of Aβ42 (PDB code 2NAO in ribbon, green) to the structural model of a DNAJB6 dimer (Söderberg et al 2018, in space-fill, grey). The S/T-residues in region 132-195 are high-lighted in pale pink, and the S/T-residues 190 and 192-5 that surround the possible peptide-binding cleft in dark pink.

created and their effects on the aggregation kinetics of Aβ42 fibril formation were evaluated (Fig. 1). As outlined in Fig. 1A in addition to the sequence of wildtype DNAJB6 (referred to as WT), three mutational variants with an increasing number of alanine substitutions in the S/T-rich region (referred to as S/T6A, S/T13A, S/T18A) were generated. Three other specific mutational variants were also created, (i) S/T∆, with a deletion of 21 S/T-residues, a sequence with 35 residues was deleted), (ii) S/T26A, with all 26 S/T residues in the region 132–195 substituted by alanine and (iii) S/T5A_190-5, with just the 5 S/T-residues corresponding to residues 190 and 192-195 substituted by alanine. Whereas in the deletion mutant (S/T∆) both the oligomerization and secondary structure was perturbed, all the S/T-substituted variants of DNAJB6 appear structurally unchanged (Fig. S1). Thus, the structure of the S/T-substituted variants of DNAJB6 resemble the wild type and are therefore relevant for evaluation of the effect of the presence of the S/T-residues without disturbance from structural effects. The capacity of DNAJB6 and its mutational variants to inhibit fibril formation by Aβ42 was evaluated by monitoring ThT fluorescence as a function of time, a method that relies on the enhanced fluorescence intensity of ThT when bound to aggregates containing regions of 6 ACS Paragon Plus Environment

Page 6 of 25

Page 7 of 25 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

β-sheet structure. The concentration of Aβ42 was 3 µM and the various chaperone concentrations shown in Fig. 1B-H correspond to molar ratios of Aβ42 to DNAJB6 ranging from 1:0.005 to 1:0.1. In the absence of DNAJB6, the high aggregation propensity of Aβ42 leads to a reaction half-time (t1/2) for fibril formation of less than 0.5 h, whereas in the presence of the highest (but still sub-stoichiometric) amounts of DNAJB6 (WT, Fig. 1B) the half-time for fibril formation is increased to nearly t1/2 ≈ 100 h. The finding that the aggregation profiles are parallel with similar slopes means that DNAJB6 delays the reaction by increasing the lag-time of Aβ42 aggregation, without a detectable effect on the growth phase of the aggregation reaction. This behavior is characteristic for inhibition of the primary nucleation step in the aggregation process 41,48,49. For the mutational variants of DNAJB6 (Fig. 1C-H, note the differences in scale on the xaxes), the aggregation reaction profiles show that the potency to suppress the fibril formation of Aβ42 declines in proportion to the number of substitutions (6, 13, 18) in the S/T-rich region (Fig. 1C-E). The mutational variant of DNAJB6 with 6 substitutions (S/T6A, Fig. 1C) is quite similar in efficacy to DNAJB6 (WT, Fig. 1B), whereas those with 13 (S/T13A, Fig. 1D) and 18 substitutions (S/T18A, Fig. 1E) show a considerably lower degree of inhibition. The variants with the deletion of 21 S/T residues (∆ST, Fig. 1F) and with the 26 substitutions (S/T26A, Fig. 1G) both show a similar loss of inhibition activity as the variant with 18 substitutions (S/T18A, Fig. 1E). In these three cases also the slopes of the aggregation profiles are changed, indicating that the S/T-substitutions have reduced substantially the capacity to inhibit primary nucleation but have retained some capacity to inhibit secondary nucleation (see below). Substitution of only the 5 S/T-residues in the end of the S/T-rich region (S/T5A_190-5, Fig. 1H) has resulted in a large activity loss compared to substitution of 6 S/Tresidues in beginning of the S/T-rich region (S/T6A, Fig. 1C). The S/T-residues 190, 192, 193, 194 and 195 are located in the segment of the S/T-rich region which is part of a predicted β-sandwich fold and which, according to a structural model of DNAJB6 50, is located in close proximity to a peptide-binding cleft at the monomer-monomer interface, as shown in the image of the DNAJB6 dimer in Fig. 1I, with S/T-residues 190 and 192-5 high-lighted in dark pink. Approximate half-time values for fibril formation are summarized for comparison in Table 1, further emphasizing the difference in inhibition of fibril formation between DNAJB6

Table 1 Half-time values for Aβ β42 fibril formation in absence and presence of DNAJB6 or its mutational variants. Approximate half-time (t1/2)-values for the process of fibril formation by Aβ42 are taken from the aggregation reaction profiles in the ThT fluorescence assay in Fig. 1 to compare the inhibitory effect of DNAJB6 and its mutational variants. Approximate t1/2-values (h) Molar ratio Aβ42 to DNAJB6

1:0

1:0.1

1:0.05

1:0.01

1:0.005

WT

0.5

100

50

10

1.5

S/T6A

0.5

60

40

1.2

0.8

S/T13A

0.5

5

3

1

1

S/T18A

0.5

2.5

1.5

1

1

S/T∆

0.5

2

1

1

1

S/T26A

0.5

1.5

1

1

1

S/T5A_190-5

0.5

15

7

1

1

7 ACS Paragon Plus Environment

Biochemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

WT and its mutational variants. In samples with DNAJB6 WT and its mutational variants fibril formation was completed at different time-points, at which the samples yet appeared similar in terms of concentrations of Aβ42 (Fig. S2), low enough to indicate that equilibrium had been reached and with no detectable differences in cryo-TEM in the fibrils formed (Fig. S3). S/T to A substitution reduces the ability to inhibit the primary nucleation pathways in Aβ42 fibril formation. The development of a protocol to achieve highly reproducible kinetic data and the derivation of an analytical solution to the coupled differential equations that govern amyloid fibril growth has allowed the rates of the microscopic processes to be extracted from macroscopic measurements of the aggregation kinetics 51. Using this approach, we previously showed that DNAJB6 inhibits both primary and secondary nucleation events in the aggregation process of Aβ42 peptide, with a much larger effect on the primary nucleation rate constant (kn) relative to the secondary nucleation rate constant (k2) 32,42. Here, we took the Aβ42 aggregation profiles shown in Fig. 1 and quantified the relative changes in lag-time as well as in the rate constants for primary and secondary nucleation pathways, both in the absence and presence of different concentrations of DNAJB6 and its mutational variants (Fig. 2). The results show that the lag-time is increased by a factor 100 in the presence of DNAJB6 (WT, Fig. 2A) compared to its absence, and that this effect on the lag-time is lowered by the S/T substitutions in the order S/T6A, S/T5A_190-5, S/T13A, S/T18A and S/T26A. Strikingly, it is especially the capacity to inhibit the primary nucleation pathways that is affected by S/T substitutions within DNAJB6, with relative changes in the inhibition of primary and secondary pathways, shown in Fig. 2B and 2C being, respectively, 10-17 and 10-2. The ability of the S/T-substitutional variants to inhibit the primary and secondary nucleation pathways follow essentially the same order as the lag-time, with the inhibition efficiency reduced progressively with the number of S/T to A substitutions in S/T6A, S/T13A, S/T18A. The rate constant associated with the primary nucleation pathway is lower by a factor of 1015 for S/T18A, meaning that its capacity to inhibit primary nucleation pathways is nearly completely abolished. Even more strikingly, it is particularly the substitution of the S/T-residues 190 and 192-5 that reduces the capacity to inhibit primary nucleation, since the rate constant for

Fig. 2 Differences between DNAJB6 and its mutational variants with respect to lag-time and rate constants in the inhibition of Aβ42 fibril formation. Data taken from Fig. 1 show the differences between the wildtype DNAJB6 (WT) and the mutational variants (designated S/T6A, S/T13A, S/T18A, S/T∆, S/T26A and S/T5A_1905), with respect to lag-time (A) and the primary (B) and the secondary (C) nucleation rate constants. Data are presented as ratios with respect to the behaviour in the absence of DNAJB6. The concentration of Aβ42 was 3 µM and the concentrations of DNAJB6 and its mutational variants correspond to the molar ratios 1:0.005, 1:0.01, 1:0.05 and 1:0.1 of Aβ42 to the DNAJB6 chaperone.

8 ACS Paragon Plus Environment

Page 8 of 25

Page 9 of 25 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

primary nucleation was reduced more substantially for the S/T5A_190-5 than for S/T13A (Fig. 2B), whereas the opposite for found for the secondary nucleation (Fig. 2C). We concluded that the role of the S/T residues for DNAJB6 function is an interesting a very definitive result and the mutational variant S/T18A, with drastically reduced capacity to inhibit the fibril formation by Aβ42, was selected for further characterization of the interactions between Aβ42 and DNAJB6. Interactions between Aβ42 and DNAJB6 or the mutational variant S/T18A. Using microscale thermophoresis (MST), an approach to determine the interactions and binding affinities of molecular species in solution that is based on the fact that different types of particles can exhibit different responses to the presence of a temperature gradient 52,53, the interactions between Aβ42 and DNAJB6 were evaluated. A fixed concentration of Alexafluor488-labeled Aβ42 (0.5 µM) was used and increasing concentrations, up to 50 µM, of DNAJB6 or its mutational variant S/T18A. Labelling of Aβ42 at the N-terminus retards aggregation in a similar manner as the addition of a few more residues from APP 54, which can be overcome using a low molar ratio of labeled peptide to unlabeled peptide 55. In the present study, the labelled peptide was used to estimate the binding equilibrium between Aβ42 and DNAJB6, and the obtained parameters are valid for the Alexafluor488-labelled Aβ42. The results showed that there was a distinct difference between on the one hand the wildtype DNAJB6 (WT, Fig. 3A), to which Aβ42 was observed to bind and on the other hand

Fig. 3 Thermophoresis study of the interaction of Aβ42 with DNAJB6 or its mutational variant S/T18A. Microscale thermophoresis (MST) traces for 0.5 µM Aβ42 at various concentrations of wildtype DNAJB6 (WT, A) and the mutational variant (S/T18A, B), to which no binding is detected. Insets show the fluorescence change during the heating period (from 0-20 s) as a function of DNAJB6 concentration. To calculate the affinity between Aβ42 and DNAJB6 WT, the data for 0.5 µM Aβ42-Alexa488 at various concentrations of wildtype DNAJB6 ranging from 0 to 50 µM were plotted, with linear (C) and logarithmic (D) scale on the xaxis. The lines are fits to the data of a 1:1 Langmuir isotherm with a KD of 3 µM.

9 ACS Paragon Plus Environment

Biochemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 25

the S/T-substituted variant, to which no binding could be detected (S/T18A, Fig. 3B). To calculate the affinity between Aβ42 and DNAJB6 the MST-data were plotted (Fig. 1C, measurements were repeated over a time window of 8 h, with no significant difference being evident and the data shown are the average over all time points). The solid lines represent fits to a 1:1 Langmuir isotherm with KD = 3 µM. Fitting to each individual data set yielded similar KD values with an average and standard deviation of 2.9±0.7 µM. We also evaluated the interactions by surface plasmon resonance (SPR) and found that immobilized Aβ42 exposed to DNAJB6 or its mutational variants showed typical association curves. Although the rate constants and equilibrium binding constants could not be determined since the signals returned to baseline level, within the time frame of the dissociation phase, the amplitudes of the association curves were observed to decrease with the degree of S/T substitution in the order WT > S/T6A > S/T13A > S/T18A (Fig. S4). Thus, both the ability to confer substoichiometric inhibition of primary nucleation (Fig. 1) and the ability to directly interact with Aβ42 (Fig. 3, Fig. S3) were strongly affected by the S/T substitutions. The interactions between Aβ42 and DNAJB6 were next evaluated by monitoring the Aβ42 monomer concentration in solution by NMR spectroscopy, using doubly isotope-labeled (15N and 13C) Aβ42 in absence or presence of DNAJB6 WT or S/T18A (Fig. 4) and observing the methyl group intensities in 13C-edited 1H NMR spectra. In the absence of DNAJB6, the signal from monomeric Aβ42 displayed a characteristic sigmoidal transition from 100% to close to 0%, with a reaction mid-point of t1/2 ≈ 0.4 days. In the presence of DNAJB6 WT (molar ratio Aβ42 to DNAJB6 1:0.1), a transition was observed with a mid-point of ≈ 17 days such that over half the concentration of Aβ42 (12 µM, ≥600-fold supersaturation) remained monomeric solution for an extended period of time, by contrast, for S/T18A only ca. 10% of the Aβ42 signal was observed after a transition with a reaction mid-point of ≈ 4 days (Fig. 4A). The NMR spectra of Aβ42 in solution displayed the characteristics of a disordered peptide, with no significant changes of either chemical shifts or line widths over time (Fig. 4B), indicating Fig. 4 DNAJB6 maintains the Aβ42 monomer soluble for extended time as monitored by NMR. (A) 13C edited integrated 1H signal of 20 µM 15N13C Aβ42 was followed over time, for samples containing Aβ42 only (black) or Aβ42 + 2 µM of DNAJB6 (WT, blue) or the S/Tsubstituted mutational variant (S/T18A, red). Inset shows a magnified view the first 24 h of Aβ42 only. A sigmoidal function was used for regression resulting in t/2 = 17.2 ± 0.2 days, cooperativity factor k = 0.23 ± 0.01 /day and a remaining Aβ42 monomer fraction = 62.1 ± 0.5 % (for WT), t/2 = 4.4 ± 0.1 days, k = 0.33 ± 0.01 /day, Aβ42 monomer = 10 ± 5% (for S/T18A), t/2 = 0.43 ± 0.01 days, k = 14 ± 1 /day (Aβ42 only), this was fitted to decay to zero intensity. (B) 13C edited 1H methyl groups of 20 µM 15N13C Aβ42 in the presence of 2 µM DNAJB6 WT. The first 10 spectra of day one are shown in blue (left y-axis), the last 10 spectra of day 76 are shown in cyan (right y-axis). All samples are in 20 mM sodium phosphate, pH 8.0, 10% (v/v) D2O at 37°C.

10 ACS Paragon Plus Environment

Page 11 of 25 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

that the signals observed arise predominantly from unfolded and unbound monomeric Aβ42. Moreover, we measured translational diffusion of Aβ42 in solution during the aggregation process using pulsed-field gradient experiments, and the diffusion coefficient did not change significantly with time. The diffusion coefficient of 2.69 ± 0.16 is in complete agreement with that of monomeric Aβ42. Taken together, the results show that the species observed in solution during the entire aggregation experiment is disordered monomeric Aβ42, with no detectable (