Amyloid Beta Aggregation in the Presence of Temperature-Sensitive Polymers

The formation of amyloid fibrils is considered to be one of the main causes for many neurodegenerative diseases, such as Alzheimer’s, Parkinson’s or Huntington’s disease. Current knowledge suggests that amyloid-aggregation represents a nucleation-dependent aggregation process in vitro, where a sigmoidal growth phase follows an induction period. Here, we studied the fibrillation of amyloid β 1-40 (Aβ40) in the presence of thermoresponsive polymers, expected to alter the Aβ40 fibrillation kinetics due to their lower critical solution behavior. To probe the influence of molecular weight and the end groups of the polymer on its lower critical solution temperature (LCST), also considering its concentration dependence in the presence of buffer-salts needed for the aggregation studies of the amyloids, poly(oxazolines) (POx) with LCSTs ranging from 14.2–49.8 °C and poly(methoxy di(ethylene glycol)acrylates) with LCSTs ranging from 34.4–52.7 °C were synthesized. The two different polymers allowed the comparison of the influence of different molecular structures onto the fibrillation process. Mixtures of Aβ40 with these polymers in varying concentrations were studied via time-dependent measurements of the thioflavin T (ThT) fluorescence. The studies revealed that amyloid fibrillation was accelerated in, accompanied by an extension of the lag phase of Aβ40 fibrillation from 18.3 h in the absence to 19.3 h in the presence of the poly(methoxy di(ethylene glycol)acrylate) (3600 g/mol).


Introduction
A wide range of neurodegenerative diseases, such as Alzheimer's, Parkinson's or Huntington's disease, is related to misfolding and aggregation of particular amyloid proteins, leading to the formation of insoluble fibrils [1]. Despite the fact that native soluble proteins associated with these illnesses have very different amino acid sequences and, therefore, different structures, the resulting insoluble aggregates are mainly straight, unbranched and share a common β-sheet secondary structure [2][3][4][5][6][7]. The fibrillation kinetics of the amyloid protein has been broadly studied and can be described by a sigmoidal curve featuring three characteristic regions known as the lag phase, the growth phase and the saturation phase or "plateau region" [3,4,8].
At the very early stage of the lag phase, when no (larger, detectable) aggregates are present, native soluble proteins combine to form primary nuclei, which are referred to as oligomeric species and protofibrils characterized by a significant β-sheet conformation (Figure 1a) [9,10]. These primary nuclei then elongate (Figure 1b), and a secondary nucleation catalyzed by the surface of a growing (c) monomer-dependent secondary nucleation; (d) fragmentation creates new elongation sites in the system; and (e) transition of a thermoresponsive polymer from the hydrated coil to the collapsed state. Prior to heating, the polymer chain (brown) is more hydrated. When the cloud point temperature (TCP) has been reached, the chain collapses to a compact globule by releasing most of the hydration water (blue), and further aggregation of the collapsed chains is possible [3,[12][13][14][15][16]. LCST: lower critical solution temperature.
Beside small molecules [17,18] and proteins [19][20][21][22], especially polymers are promising candidates for the inhibition of the fibrillation as has been shown for the poly-(L)-lysine (PLL) [23], dendrimers [11,24,25] and polymeric nanoparticles [26][27][28]. Thus, the presence of bilayer membranes [3,29], micellar aggregates [3,[30][31][32][33] and nanoparticles [27,34,35] has shown to significantly alter the aggregation process, where especially the interplay between Aβ with artificial membranes is often seen as an important factor of the Aβ fibrillation in vivo [3]. In vitro aggregation of the amyloid protein can be delayed by means of interaction with, e.g., phospholipid membranes, leading to a reduction of the amyloid-formation from initially 2-9 h in the presence of liposomes made from dioleoyl-phosphatidylcholine (DOPC) in the gel phase [29]. Furthermore, surfactant micelles interfere with the formation of the Aβ fibrils in a concentration-dependent manner [3,[30][31][32][33], as in the presence of sodium dodecyl sulfate (SDS) the lag time for a solution containing 150 mM Aβ40 is decreasing from 7.5 ± 2.0-1.9 ± 0.2 h at 0.9 mM SDS and 1.2 ± 0.2 h at 2.75 mM SDS (below its critical micelle concentration (cmc)), whereas a higher concentration of SDS (12.5 mM, above the cmc) suppressed fibrillation [33]. The same effect was observed for Aβ-nanoparticle systems [27,34,35], where the presence of 0.05 or 1.1 mg/mL of amine-modified polystyrene nanoparticles in a 16 mM Aβ system changed the lag time from 138 ± 20 min to around 50 and 500 min, respectively, also reporting the formation of stickier and shorter fibrils [27]. However, contradictory results have been achieved with regard to the interference of polymers and polyelectrolytes with Aβ fibrillation kinetics [36].
To the best of our knowledge, the influence of thermoresponsive polymers on the fibrillation pathways of amyloids has not yet been investigated. The current study provides a first report on the (c) monomer-dependent secondary nucleation; (d) fragmentation creates new elongation sites in the system; and (e) transition of a thermoresponsive polymer from the hydrated coil to the collapsed state. Prior to heating, the polymer chain (brown) is more hydrated. When the cloud point temperature (T CP ) has been reached, the chain collapses to a compact globule by releasing most of the hydration water (blue), and further aggregation of the collapsed chains is possible [3,[12][13][14][15][16]. LCST: lower critical solution temperature.
Beside small molecules [17,18] and proteins [19][20][21][22], especially polymers are promising candidates for the inhibition of the fibrillation as has been shown for the poly-(L)-lysine (PLL) [23], dendrimers [11,24,25] and polymeric nanoparticles [26][27][28]. Thus, the presence of bilayer membranes [3,29], micellar aggregates [3,[30][31][32][33] and nanoparticles [27,34,35] has shown to significantly alter the aggregation process, where especially the interplay between Aβ with artificial membranes is often seen as an important factor of the Aβ fibrillation in vivo [3]. In vitro aggregation of the amyloid protein can be delayed by means of interaction with, e.g., phospholipid membranes, leading to a reduction of the amyloid-formation from initially 2-9 h in the presence of liposomes made from dioleoyl-phosphatidylcholine (DOPC) in the gel phase [29]. Furthermore, surfactant micelles interfere with the formation of the Aβ fibrils in a concentration-dependent manner [3,[30][31][32][33], as in the presence of sodium dodecyl sulfate (SDS) the lag time for a solution containing 150 mM Aβ 40 is decreasing from 7.5˘2.0-1.9˘0.2 h at 0.9 mM SDS and 1.2˘0.2 h at 2.75 mM SDS (below its critical micelle concentration (cmc)), whereas a higher concentration of SDS (12.5 mM, above the cmc) suppressed fibrillation [33]. The same effect was observed for Aβ-nanoparticle systems [27,34,35], where the presence of 0.05 or 1.1 mg/mL of amine-modified polystyrene nanoparticles in a 16 mM Aβ system changed the lag time from 138˘20 min to around 50 and 500 min, respectively, also reporting the formation of stickier and shorter fibrils [27]. However, contradictory results have been achieved with regard to the interference of polymers and polyelectrolytes with Aβ fibrillation kinetics [36].
To the best of our knowledge, the influence of thermoresponsive polymers on the fibrillation pathways of amyloids has not yet been investigated. The current study provides a first report on the influence thermoresponsive polymers exert on the fibrillation of an amyloid protein, mixed noncovalently into a fibrillating amyloid system. An important requirement for such an experiment is the correct choice of the lower critical solution temperature (LCST) behavior of the polymer, which should be chosen close to the temperature, where amyloid aggregation is observed (experimentally in vitro). As the LCST is also referred to as the cloud point temperature (T CP ), where the conformation changes from a hydrated coil to a collapsed globule by the release of most of the hydration water (coil-globule transition; see Figure 1e) [16], it is tempting to speculate about the effects of such a transition on the amyloid aggregation, most of all with respect to the macroscopic kinetic aggregation behavior.
Here, we describe the design and syntheses of two different types of polymers (poly(oxazolines) and poly(methoxy di(ethylene glycol)acrylates)), both showing an LCST at temperatures close to the in vitro-aggregation temperature of amyloids. Subsequently, we study their effect on amyloid aggregation probed by conventional aggregation assays. A special focus was placed on the comparison of the two different hydrophilic polymers.
Aβ 40 peptide was synthesized using the standard F-moc solid phase synthesis strategy at the core unit "Peptid-Technologien" of the Medical Faculty of the University of Leipzig [52].
The oxazoline monomers were stored over calcium hydride and were freshly distilled before usage. All solvents were freshly distilled and degassed by bubbling with nitrogen for at least 20 min prior to usage. Acetonitrile (ACN) was dried with calcium hydride by boiling for several hours followed by distillation. AIBN was recrystallized from ethanol twice before usage.
MALDI-TOF-MS analysis was carried out using a Autoflex III Smartbeam (Bruker Daltonics, Bremen, Germany) equipped with a nitrogen laser (337 nm) working in linear and reflection modes. The obtained data were evaluated using flexAnalysis software (Version 3.0) (Bruker Daltonics, Bremen, Germany). The matrix solution was prepared by dissolving 1,8,9-anthracenetriol (dithranol) in THF at a concentration of 20 mg/mL. The polymer was dissolved in THF (10 mg/mL) and mixed with sodium trifluoroacetate (10 mg/mL in THF). The ratio between the matrix, the analyte and the salt was 100:10:1.
Gel permeation chromatography (GPC) for the poly(methoxy di(ethylene glycol)acrylates) was performed on a Viscotek GPCmax VE 2002 (Malvern, Crowthorne, UK) using a H HR -H Guard-17360 precolumn and a GMH HR -N-18055 column with THF as the solvent and a VE 3580 IR detector for refractive index determination. A polystyrene standard (M P = 1000-115,000 g/mol) was used for external calibration. Column and detector temperatures were held at 22 and 35˝C, respectively, and the flow rate was set to 1 mL/min. The concentration of all samples was 3 mg/mL. GPC measurements of the POx were performed on a Viscotek GPCmax VE 2001 (Malvern, Crowthorne, UK using a H HR -H Guard-17369 and a GMH HR -N-18055 column with DMF as the eluent at 60˝C and via detection of the refractive index with a VE 3580 RI detector from Viscotek at 35˝C. The external calibration was done using a polystyrene standard (M P = 1000-115,000 g/mol). The concentration of all samples was 5 mg/mL, and the flow rate was 1 mL/min. Turbidimetry measurements were done using the UV-Vis spectrometer HP 8543 (Hewlett-Packard, Waldbronn, Germany). By coupling with a Peltier element HP 89090A (Hewlett-Packard, Waldbronn, Germany), controlled heating with a heating rate of 0.5˝C/min was possible. The observed wavelength was λ = 500 nm. For all measurements in water, a 1 wt % solution of the polymer was used. Measurements in buffers were done as follows: for the poly(methoxy di(ethylene glycol)acrylates), sodium phosphate buffer (25 mmol/L), containing sodium chloride (150 mmol/L) at pH = 9.2, was used. T CP was detected at 50% of transmission. The POx were measured in sodium borate buffer (50 mmol/L, pH = 9.0) as 1 wt % solutions. Only the P(iPrOx-nBuOx) copolymers were measured as 0.25 wt % solutions due to their poor solubility. T CP was chosen as the onset temperature, when the first drop of the transmission was observed.
Fluorescence measurements for the fibrillation kinetics of the Aβ 40 peptide in the absence and in the presence of the polymers were measured on a Tecan infinite M200 microplate reader (Tecan Group AG, Männedorf, Switzerland). Data were analyzed using standard protocols from the literature [53]. Thioflavin T (ThT) was used as a fibril indicator, because it shows increasing fluorescence intensity at λ = 482 nm by binding to β-sheet-rich structures, such as the formed Aβ 40 fibrils [54]. Measurements were performed in a 96-well plate. Mixtures of Aβ 40 and the polymers were dissolved in sodium phosphate buffer (25 mmol/L) at pH 9.2, containing sodium chloride (150 mmol/L), ThT (20 µmol/L) and 0.01% NaN 3 to prevent bacterial growth. The used concentrations were: Aβ 40 : 230 µM; polymer 3a: 196 µM; polymer 3b: 113 µM; polymer 3c: 260 µM; and polymer 6c: 230 µM. One hundred fifty microliters of each solution were pipetted into the wells. The temperature was 37˝C; the excitation wavelength was set to 450 nm; and the emission was measured at 485 nm every 30 min for at least three days. Two different shaking protocols were used: 30 min cycles of 5 min shaking (at a 1-mm shaking amplitude), 5 min waiting, 5 min shaking followed by the measurement (Protocol I) and 25 min and subsequent 10 s shaking followed by the measurement (Protocol II). The fluorescence intensity was fitted using the following equation [53]: where t char is the characteristic time, where the fluorescence intensity reaches half its maximum and the lag time is given by t char -2τ with τ as the inverse of the rate constant [53]. Thin-layer chromatography (TLC) was performed using "Merck silica gel 60" plates (Merck, Darmstadt, Germany). Spots on the TLC plate were visualized using oxidizing agent "blue" stain or UV light (254 or 366 nm). "Blue" stain was prepared as follows: (NH 4 ) 6 Mo 7 O 24¨4 H 2 O (1 g) and Ce(SO 4 ) 2¨4 H 2 O (1 g) were dissolved in a mixture of distilled water (90 mL) and concentrated sulfuric acid (6 mL). Subsequent column chromatography was carried out using high purity-grade Merck 60 (230-400 mesh particle size) silica gel (Darmstadt, Germany).

General Procedure for the Syntheses of the Poly(methoxy di(ethylene glycol)acrylates) 3
As an example, the synthesis for a projected molecular weight of M n = 3800 g/mol is described.
The reaction was carried out using a round-bottomed flask equipped with a magnetic stirrer, a rubber septum and a balloon filled with argon. Prior to the reaction, a mixture of monomer 1 (242.60 mg, 234.80 µL, 1.39 mmol), chain transfer agent (CTA) 2 (15.10 mg, 0.063 mmol), AIBN (1.04 mg, 0.0063 mmol) and DMF (0.34 mL) was bubbled with argon for 25 min and subsequently placed into a preheated oil bath at 70˝C. The reaction was stirred for six hours before it was opened to air and cooled by means of an ice bath. The resulting bright yellow polymer was precipitated into n-hexane (3ˆ70 mL) and dried in a high vacuum within two days. The polymeric product 3 was characterized via size exclusion chromatography (SEC), matrix-assisted laser desorption/ionization mass spectrometry (MALDI-TOF MS) (for isotopic pattern simulation see Supplementary Materials, Figure S1), 1 H-NMR (see Supplementary Materials, Figure S2) and turbidimetry (Table 1). Thus, the expected final polymer was obtained as truly proven via these experimental methods. 1

Syntheses of the Poly(oxazolines) 6-9
Similar to known procedures [47,55], 2-isopropyl-2-oxazoline 4 (iPrOx) and 2-n-butyl-2-oxazoline 5 (nBuOx), served as the monomers [56] for the cationic ring-opening polymerization for the POx. Polymers 6, 7 and 9 were all initiated with propargyl tosylate, but polymers 6 and 9 were quenched with water, whereas polymer 7 was quenched with N,N-diethylamine. The synthesis of polymer 8 was initiated with methyl trifluoromethanesulfonate and quenched with N-methylpropargylamine. Results for the characterization and the measured LCSTs for all polymers are given in Table 1 (turbidimetry, molecular weights and polydispersities). Detailed procedures for all reactions and the corresponding characterizations are given in the Supplementary Materials.

General
Procedure for the Syntheses of the Poly(oxazolines) 6-8 2-Isopropyl-2-oxazoline (2.00 g, 2.10 mL, 17.63 mmol), dry ACN (8.81 mL) and the appropriate amount of initiator (propargyl tosylate for polymers 6 and 7, methyl trifluoromethanesulfonate for polymer 8) were added to a Schlenk tube, which was subsequently sealed with a rubber septum. The mixture was stirred at room temperature for one hour and consecutively for 48 h at 80˝C. Living chain ends were quenched by the addition of the respective quencher (6: water; 7: N,N-diethylamine; 8: N-methylpropargylamine) and further stirring for 24 h at 60˝C. After evaporation of the solvent, the residue was dissolved in dichloromethane DCM (5.0 mL) and was extracted with water (5ˆ30.0 mL). The combined aqueous phases were back extracted using DCM (10ˆ30.0 mL). Subsequently, the organic phases were combined and dried over sodium sulfate. After filtration, most of the solvent was removed, and the remaining viscous solution was precipitated three times in a cold mixture of diethyl ether/n-hexane (1:1) to obtain the pure polymer. A representative 1 H-NMR spectrum is given in the Supplementary Materials for polymers 6, 7 and 8 ( Figures S3-S5). 1  2.6. General Procedure for the Syntheses of Poly(2-isopropyl-2-oxazoline-grad-2-n-butyl-2-oxazoline) 9 The procedure was done as described for the homopolymerization of 2-isopropyl-oxazoline. A mixture of 2-isopropyl-2-oxazoline (0.54 mL, 0.51 g, 4.50 mmol), 2-n-butyl-oxazoline (63.60 mg, 0.50 mmol), propargyl tosylate (50.86 µL, 61.80 mg, 0.29 mmol) and ACN (2.50 mL) was stirred for one hour at room temperature in a Schlenk tube. After stirring for 48 h at 80˝C, the reaction was quenched by the addition of water (20.90 µL, 20.90 mg, 1.16 mmol). The reaction was stirred for a further 24 h at 60˝C, and the work-up was done as described for the poly(2-isopropyl-2-oxazoline). A representative 1 H-NMR spectrum is given in the Supplementary Materials ( Figure S6), as well as a table with full characterization of the synthesized copolymers (initial monomer ratio, molecular weight as determined from 1 H-NMR and GPC, polydispersity index (PDI) and copolymer composition as determined from 1 H-NMR) (
As a representative example, the MALDI-TOF of polymer 3a is shown in Figure 2, displaying three different series (Series S1-S3), corresponding to the repeating unit (difference~174 Da) and different types of attached ions.  Figure S1).
As a representative example, the MALDI-TOF of polymer 3a is shown in Figure 2, displaying three different series (Series S1-S3), corresponding to the repeating unit (difference ~174 Da) and different types of attached ions. The first series at 3047.179 Da can be assigned to poly(methoxy di(ethylene glycol)acrylate) with a formula of [HOOCC2H4(C8H14O4)16S3C5H9]Na + . The main signal of the multiplet chosen for Series S2, which appears at 3069.360 Da, can be assigned to [NaOOCC2H4(C8H14O4)16S3C5H9]Na + , whereas the minor Series S3 appears at 3084.645 Da and is assigned to the polymer with the formula of [NaOOCC2H4(C8H14O4)16S3C5H9]K + . For simulation of the isotopic pattern, check the Supplementary Materials ( Figure S1). Further characterization via NMR spectroscopy is shown in the Supplementary Materials ( Figure S2), finally proving both the end groups and the true structure of polymer 3. Similarly, the poly(oxazolines) 6-9 were characterized via NMR spectroscopy and GPC (see Supplementary  Materials Figures S3-S6)

Turbidimetry of the Polymers 3, 6-9
In preparation of the subsequent fibrillation experiments (amyloid aggregation in the presence of polymers), first-hand experimental information on the LCST influence of the molecular weight and the addition of salts (buffer solutions are required for respective fibrillation conditions), the concentration of the polymer via appropriate extensive turbidimetric measurements was accomplished. First experiments were done with PiPrOx (6, 7 and 8), whose LCST behavior is well documented in the literature [41,42,47,50], followed by poly(methoxy diethylene glycol)acrylate, Figure 2. MALDI-TOF spectrum of poly(methoxy di(ethylene glycol)acrylate) (3a). S1, Series 1.

Turbidimetry of the Polymers 3, 6-9
In preparation of the subsequent fibrillation experiments (amyloid aggregation in the presence of polymers), first-hand experimental information on the LCST influence of the molecular weight and the addition of salts (buffer solutions are required for respective fibrillation conditions), the concentration of the polymer via appropriate extensive turbidimetric measurements was accomplished. First experiments were done with PiPrOx (6, 7 and 8), whose LCST behavior is well documented in the literature [41,42,47,50], followed by poly(methoxy diethylene glycol)acrylate, which is less intensely studied in the literature [45,61,62]. The results of the turbidimetric measurements are given in Table 1. Table 1.
Measured T CP for the synthesized poly(methoxy diethylene glycol)acrylates (3), poly(2-isopropyl-2-oxazoline) (PiPrOx) (6-8) and P(iPrOx-nBuOx) (9) in sodium borate buffer and in water. For the homopolymers, 1 wt % solutions were used, whereas for the copolymers, 0.25 wt % solutions were used due to poor solubility in water. The samples were heated at a rate of 0.5 K/min. PDI = Polydispersity Index  1 Ratios of the monomers (4 and 5) within the copolymers as calculated from 1 H-NMR; 2 the 1 H-NMR signal from the methylene group next to the alkyne group at δ = 4.05 ppm was used as a reference for polymers 6, 7 and 9, whereas for polymer 8, the signal from the methyl-initiator group at δ = 3.05 ppm was used; 3 measured in DMF with polystyrene (M P = 1000-115,000 g/mol) as the standard; 4 sodium phosphate buffer (pH = 9.2, 150 mmol NaCl) was used instead with a polymer concentration of 230 µmol/L; 5 no LCST detectable under the used conditions; 6 due to the higher PDI, the drop in transmission is much less pronounced in comparison to the other samples; polymer 6d exhibits two LCSTs when measured in buffer solution; this might be due to its relatively high polydispersity; 7 bimodal distribution according to GPC.
For the PiPrOx polymers (6, 7 and 8), the same influence of the molecular weight on the LCST was found as described in the literature [46,55,63]. Thus, an increase in the molecular weight led to a decrease of the LCST. By varying the molecular weight, it was possible to tune the LCST in a range of˘5 K. PiPrOx 6a with a molecular weight of 2600 g/mol (GPC) exhibits an LCST of 47.5˝C (in H 2 O), whereas PiPrOx 6c with a molecular weight of 11,000 g/mol (GPC) displays an LCST of 36.4˝C (in H 2 O). The measured curves of some of the PiPrOx are shown in Figure 3a and illustrate the influence of the molecular weight on the LCST. A significant influence of the used buffer on the LCST was observed when changing to sodium borate buffer as the solvent (required for the amyloid fibrillation assays; 50 mmol/L, pH = 9.0), as the presence of the sodium borate provoked a decrease of the LCST in the range of 1-4˝C, except for polymer 8. As the ions in the buffer are able to weaken the hydrogen bonds between the polymer chains and water molecules, the release of water is facilitated, consequently shifting the LCST to lower temperatures (salting out effect). Figure 3b exemplary shows the measured curves for the polymer 6a in pure water and in sodium borate buffer as a comparison.
Furthermore, the concentration dependency of the LCST was measured for polymer 7 in the range from 0.25-1.5 wt % in water. The results are presented in the Supplementary Materials (Table S2), and the measured curves are shown in Figure 3c. Furthermore, the concentration dependency of the LCST was measured for polymer 7 in the range from 0.25-1.5 wt % in water. The results are presented in the Supplementary Materials (Table  S2), and the measured curves are shown in Figure 3c. For the lowest concentration (0.25 wt %), an LCST of 48.2 °C and for the highest concentration (1.5 wt %) an LCST of 41.3 °C were measured. Apparently, a change of the concentration by 0.25 wt % has a significant effect on the LCST, especially at lower concentrations.  As expected, the copolymerization with the more hydrophobic nBuOx resulted in a decrease of the LCST as described in the literature [47]. Because of the high amount of nBuOx from 21%-35%, the LCST was reduced to 27.2 and 15.3˝C respectively, illustrating the strong influence of the hydrophilic/hydrophobic ratio of a polymer on its LCST. By measuring the samples in sodium borate buffer, the LCST could be further decreased to 14.2˝C (see Supplementary Materials, Figure S7).
In a similar manner, LCST data were obtained via turbidimetric measurements for the polyacrylates 3 as explained in the experimental part, showing the same unexpected influence of the molecular weight on the LCST as described in the literature [59,61,62] (see Table 1). Thus, LCSTs ranging from 34.4˝C (polymer 3a) to 45.1˝C (polymer 3c) could be determined, demonstrating a strong increase ofthe T CP with increasing molecular weight [62].
Upon determining the LCST in sodium phosphate buffer at pH = 9.2, keeping the polymer concentration at 230 µM, polymer 3a does not display an LCST under the applied conditions (measurement range = 25-70˝C). Cloud point temperatures of samples 3b and 3c were detected at 52.7˝C (40.4˝C in water) and 49.5˝C (45.1˝C in water), indicating that the addition of ions stabilize the macromolecules in solution, however with a less pronounced effect for higher molecular weights (3b: ∆T cp = 12.3˝C; 3c: ∆T cp = 4.4˝C) (see Supplementary Materials for details on the measured curves, Figures S8 and S9).

Fibrillation of Aβ 40 in the Presence of Polymers 6 and 3
The choice of the polymers for the amyloid fibrillation measurements was based on the reverse thermoresponsive behavior of the POx and polyacrylates. These polymers are water and buffer soluble, and the T cp could be adjusted over a wide range of temperatures. Polyacrylates 3a, 3b and 3c (displaying molecular weights below and above the molecular weight of Aβ 40 ) serve as a good selection for the Aβ aggregation studies due to the different LCSTs. PiPrOx 6c was chosen due to the molecular weight of the polymers 3c and 6c being close to each other (14,600 g/mol and 11,000 g/mol, respectively), thus probing whether there is an influence on the fibrillation of the Aβ 40 due to the structural differences of the polymers.
Upon addition of an LCST-type polymer to an aggregating amyloid, a number of possible scenarios can be considered. Thus, by mixing of the polymer with the Aβ protein, the polymer can either inhibit or enhance fibrillation, as observed before [36]. At T < T LCST and a concentration C amyloid < C polymer , the amyloid could evenly distribute within the polymer phase, in turn preventing the amyloid/amyloid contact formation and, thus, reducing or even eliminating fibrillation. At a concentration C amyloid < C polymer , the polymer could also form a protective layer around the native proteins, and at C amyloid > C polymer , the polymer might boost the fibrillation by acting as a catalytic surface. If present above its LCST, several scenarios are possible, e.g., that the polymer can hinder the fibrillation of the amyloid by evenly binding the native proteins to its surface and therefore preventing the formation of bigger oligomers or alternatively accelerating the fibrillation by acting as a seed.

Effect of Poly(oxazolines) on Aβ 40 Aggregation
For the first fibrillation measurements, equimolar amounts of Aβ 40 (wild-type, WT) and the corresponding polymer were dissolved in sodium phosphate buffer (25 mM, pH = 9.2, containing 150 mM NaCl) at a concentration of 230 µmol/L for both Aβ 40 and the polymer. Tracking of the fibril growth was done by using the fluorescence dye thioflavin T (ThT) (20 µmol/L), as ThT binds to β-sheet-rich structures showing an increasing fluorescence at λ = 482 nm, which in turn allows the measurement of the kinetics of the fibril growth [54]. Measurements were done at 37˝C using a well plate, which allowed the triple measurement of each sample simultaneously (Protocol I, see Methods). The time-dependent development of the ThT at λ = 482 nm is shown in Figure 4.  Thus, the faster fibril growth upon the addition of polymer 6c can be attributed to a shortening of the lag phase of fibrillation. As under the used conditions the LCST of 6c was 36.2 °C, the polymer was above its LCST and consequently precipitated during the measurement, in turn acting as a seed for the Aβ40 monomers and favoring oligomerization.  Thus, the faster fibril growth upon the addition of polymer 6c can be attributed to a shortening of the lag phase of fibrillation. As under the used conditions the LCST of 6c was 36.2˝C, the polymer was above its LCST and consequently precipitated during the measurement, in turn acting as a seed for the Aβ 40 monomers and favoring oligomerization.

Effect of Poly(methoxy di(ethylene glycol)acrylates) on Aβ 40 Fibrillation
Measuring amyloid aggregation via the ThT fluorescence assay (Protocol II, see Methods) in the presence of the acrylates 3a, 3b and 3c was accomplished in 230 µM Aβ 40 buffered solutions, mixing with buffered solutions of the polymers with different chain lengths and concentrations ( Figure 5). In all cases, the lag time was reduced, and the saturation phase of fibrillation was achieved more quickly. Furthermore, the slope of all three aggregation curves became steeper, in comparison to WT Aβ 40 , implying that the presence of the polymer is boosting the protein fibrillation and enhancing secondary nucleation and the elongation rates. Remarkably, the duration of the lag phase seems to be decreased in the case of the polymers 3b and 3c, but slightly increased from t = 18.3 h to t = 19.3 h in the case of the polymer 3a, indicating that polymer 3a, having the closest molecular weight compared to the Aβ40 (~4300 g/mol), is able to interact with the native protein interfering with the primary nucleation process. This phenomenon is extremely interesting due to the prospective control over the Aβ40 fibrillation upon the change of the molecular weight or the concentration of the polymer and will be studied in detail in the upcoming publications. Nevertheless, the WT Aβ40 peptide shows the longest characteristic times compared to the peptide/polymer mixtures.
The summarized values for the PiPrOx, as well as for the poly(methoxy di(ethylene glycol)acrylates) are given in Table 2, as well as the cloud points for those polymers under the conditions that were used for the fibrillation experiments.  Remarkably, the duration of the lag phase seems to be decreased in the case of the polymers 3b and 3c, but slightly increased from t = 18.3 h to t = 19.3 h in the case of the polymer 3a, indicating that polymer 3a, having the closest molecular weight compared to the Aβ 40 (~4300 g/mol), is able to interact with the native protein interfering with the primary nucleation process. This phenomenon is extremely interesting due to the prospective control over the Aβ 40 fibrillation upon the change of the molecular weight or the concentration of the polymer and will be studied in detail in the upcoming publications. Nevertheless, the WT Aβ 40 peptide shows the longest characteristic times compared to the peptide/polymer mixtures.
The summarized values for the PiPrOx, as well as for the poly(methoxy di(ethylene glycol)acrylates) are given in Table 2, as well as the cloud points for those polymers under the conditions that were used for the fibrillation experiments.

Discussion of Aβ 40 Aggregation Experiments
The influence of polymers, such as neutral polymers [36], polyelectrolytes [36], dendrimers [11,24,25] or polymeric nanoparticles [26][27][28], on amyloid aggregation has been studied. Often, retarding effects on the fibrillation of Aβ have been observed. Amyloid formation has been found to be promoted by glycosaminoglycans [64] and positively-charged polymers, such as poly(diallyldimethylammonium chloride) (PDDA), poly(ethylenimine) (PEI) and poly(lysine) hydrobromide [36]. Most relevant to our situation is the study by Assarsson et al., in which the effect of charged polymers and polyamino acids on the Aβ fibril formation was studied [36]. While an accelerating effect of all positively-charged polyelectrolytes on Aβ fibril formation was found, neutral (poly(threonine) and negatively-charged polymers poly(acrylic acid sodium salt) (PAA) and poly-glutamic acid sodium salt did not interfere with the kinetics of Aβ fibrillation. In contrast, all neutral polymers investigated in our study also accelerated the fibrillation kinetics of Aβ 40 . In agreement with the aforementioned results, the lag time of Aβ fibril formation decreased, and the growth phase showed a steeper slope. This is akin to the situation where the Aβ concentration is systematically increased [65]. Table 2. Measured data for the fibrillation kinetics of the Aβ 40 fibrillation in the presence of the polymers 3a, 3b, 3c (Fluorescence Protocol II) and 6c (Fluorescence Protocol I, see Methods) at 37˝C. The characteristic time and lag time vary between the different protocols, as shaking times varied between the two protocols. Shaking increases the probability of molecular contacts, thus enhancing aggregation.

Name
T CP (˝C) WT concentration 1 It should finally be mentioned that the concentration of the polymers added to the Aβ preparations is too low to influence the Aβ fibrillation by crowding through excluded volume effects. However, somehow, it is likely that the presence of the polymer leads to an increase in the local Aβ concentration, which favors nucleation of the peptide and leads to the observed increased fibrillation kinetics.

Conclusions
Amyloid aggregation was studied in the presence of hydrophilic polymers, able to display an LCST in the range between 35 and 45˝C. For this effect, we have synthetically prepared a large number of hydrophilic polymers, adjusting the LCST to 35-45˝C by either homo-or co-polymerization. Poly(oxazolines) 6-9, prepared via cationic ring opening polymerization, served as one class of hydrophilic polymer, whereas poly(methoxy di(ethylene glycol)acrylates) 3a-3c prepared via RAFT polymerization were the second class of polymers. For all polymers, a significant influence of the end groups, as well as the molecular weight on the final LCSTs was observed, restricting the investigations to a total of four polymers (the acrylates 3a, 3b and 3c, as well as the poly(oxazoline) 6c) for the amyloid-fibrillation experiments. Furthermore, the used buffers did show a significant influence on the LCSTs of the polymer in solution, with a significant increase in the case of the poly(acrylates) 3b and 3c and a decrease in the case of the poly(oxazoline) 6c. The polymers (6c, 3b, 3c) do show a reduction of the lag time of amyloid-aggregation, indicative of an enhanced fibrillation. Although one polymer 3a did not show an LCST in the used buffer solution, an increase of the lag time of the fibril formation from 18.34-19.33 h was observed.
In summary, several factors can be considered as being influential for the polymer/amyloid aggregation in the current system: (1) enhanced nucleation due to Aβ-cluster formation at the polymer chains at temperatures below the LCST [36,66]; (2) steric shielding of the amyloids as a consequence of polymer adsorption (polymers 3a, 3b and 3c) [15,[30][31][32]; (3) enhanced nucleation of amyloid aggregation via the formed polymer particles at temperatures above the LCST (polymer 6c) [3,27]. Among the discussed possibilities, the enhanced nucleation by all of the used polymers seems the most probable scenario, as in all cases, fibrillation of the amyloids is enhanced, irrespective of their chemical or physical nature. Research is proceeding in this direction, as currently, other mechanisms cannot be ruled out.