Cosolute Effects on Amyloid Aggregation in a ... - ACS Publications

Article

Subscriber access provided by SUNY DOWNSTATE

Cosolute Effects on Amyloid Aggregation in a Non-Diffusion Limited Regime: Intrinsic Osmolyte Properties and the Volume Exclusion Principle Brian Stephen Murray, Joseph Rosenthal, Zhongli Zheng, David Isaacson, Yingxi Zhu, and Georges Belfort Langmuir, Just Accepted Manuscript • DOI: 10.1021/acs.langmuir.5b00254 • Publication Date (Web): 24 Mar 2015 Downloaded from http://pubs.acs.org on April 1, 2015

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 free 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 accessible to all readers and 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.

Langmuir 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 26

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

Langmuir

Cosolute Effects on Amyloid Aggregation in a NonDiffusion Limited Regime: Intrinsic Osmolyte Properties and the Volume Exclusion Principle Brian Murray1, Joseph Rosenthal2, Zhongli Zheng3, David Isaacson2, Yingxi Zhu3 and Georges Belfort*1 1

Department of Chemical and Biological Engineering and Center for Biotechnology and

Interdisciplinary Studies, and 2Department of Mathematical Sciences, Rensselaer Polytechnic Institute, Troy, NY 12180 3

Department of Chemical and Biomolecular Engineering, University of Notre Dame, Notre

Dame, Indiana 46556

ABSTRACT: The effects of cosolutes on amyloid aggregation kinetics in vivo are critical and not fully understood. To explore the effects of cosolute additives, the in vitro behavior of destabilizing and stabilizing osmolytes with polymer cosolutes on the aggregation of a model amyloid, human insulin, is probed using experiments coupled with an amyloid aggregation reaction model. The destabilizing osmolyte, guanidine hydrochloride (GuHCl), induces biphasic behavior on the amyloid aggregation rate exhibited by an enhancement of the aggregation

ACS Paragon Plus Environment

1

Langmuir

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 2 of 26

kinetics at low concentrations of GuHCl (< 0.6 M) and a reduction in kinetics at higher GuHCl concentrations.

Stabilizing osmolytes, glycerol, sorbitol trimethylamine N-oxide, slow the rate

of aggregation by reducing the rate of monomer unfolding.

Polymer cosolutes,

polyvinylpyrrolidone 3.5 kDa and 40 kDa, delay amyloid aggregation mainly through a decrease in the nucleation reaction. These results are in good agreement with the volume exclusion principle for polymer crowding and supports the need to include conformational rearrangement of monomers prior to nucleation. Using fluorescence correlation spectroscopy, we demonstrate that amyloid aggregation is non-diffusion limited, except during fibril accumulation in the presence of high concentrations of long chain polymers. Lastly, the neutral surface area of osmolytes correlates well with the time to initiate fibril formation, tlag, which implicates an intrinsic osmolyte property underlying preferential interactions. INTRODUCTION Aggregation of amyloidogenic proteins is implicated in multiple life threatening, debilitating diseases, such as Alzheimer’s and Parkinson’s disease.1 However most of the in vitro experiments that have studied protein aggregation have focused on a system that is dissimilar to in vivo conditions. The concentration of macromolecules (proteins, lipid membranes, etc.) in vivo can reach up to 400 mg/ml and is expected in such a concentrated environment to directly affect protein aggregation kinetics differently as compared with the same reaction in vitro.2, 3 In addition, osmolytes (sugars, polyols, urea, etc.) exist up to molar concentrations, well above the concentration required for influencing protein folding.4, 5, 6, 7 The interactions by which macromolecules and osmolytes affect protein folding are generally well understood. Macromolecules reduce the free volume available for protein unfolding


2

Page 3 of 26

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

Langmuir

through excluded volume interactions,8, 9, 10 whereas osmolytes can stabilize a more compact protein structure through preferential exclusion from the protein surface.5, 11, 12, 13, 14, 15 Further, the effects of cosolutes on amyloid aggregation have been studied across a range of amyloids with various solution conditions.6, 16, 17, 18 Typically empirical models are used to quantify the solution condition’s effect on aggregation.19 Here, we have developed a kinetic reaction model based on the earlier work of Lee and colleagues to quantify the effects of well-studied cosolutes on the microscopic rate constants of amyloid aggregation.20 To investigate the behavior of cosolutes on the rate of amyloid aggregation, we selected the model amyloid human insulin, as many others have done,6, 19, 21, 22, 23, 24, 25, 26, 27, 28 along with cosolute conditions for which the native protein folding/unfolding equilibrium is welldocumented. Polyols such as glycerol and sorbitol stabilize the folded state of proteins at submolar concentrations.29, 30 At equal cosolute concentrations, sorbitol increases the change in free energy of unfolding (∆∆G) about twice as much as glycerol via a thermodynamic mechanism of enthalpic stabilization.31 Trimethylamine N-oxide (TMAO) enthalpically stabilizes ∆∆G to a greater extent relative to the polyols due to a combination of multiple stabilization mechanisms: (1) preferential exclusion from the protein’s surface as the polyols do, and (2) a reduction in water’s hydrogen bonding ability.32, 33 Guanidine hydrochloride (GuHCl) destabilizes the protein’s folded state, resulting in a negative ∆∆G, through a mechanism of preferential inclusion at the protein’s surface.34, 35, 36 These interactions of preferential inclusion and exclusion can be quantified by the thermodynamic parameter “preferential interaction coefficient”, Γ, which is defined as the partial derivative of solute molality with respect to protein molality at constant temperature, pressure and chemical potential, µ.37


3

Langmuir

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 4 of 26

Similar to stabilizing osmolytes, hydrophilic polymers such as polyvinylpyrrolidone (PVP) encourage native folded protein states.31 Through a mechanism of excluding the volume available to proteins (steric exclusion), polymers with a large degree of polymerization (~ 100) such as PVP with a molecular weight of 40,000 Da (PVP40) can have a significant effect on protein structure even at low volume fractions (φ).38 Polymers with a small degree of polymerization (~ 10) (e.g. PVP3.5) still support a folded structure, however to a lesser degree.38 Here the cosolute effects on the aggregation kinetics of the model amyloid, human insulin, are measured using absorbance at 600 nm (A600), a well-established method for tracking insulin aggregation.6, 20, 21, 23 We use a modified version of a reaction model previously developed to explore the reaction rate constants for insulin aggregation.20 This revised model includes two additional concepts − conformational rearrangement of monomers prior to nucleation and fragmentation of fibrils. The reaction rates obtained from best fits are analyzed to determine the effects that different classes of cosolutes (stabilizing osmolytes, destabilizing osmolytes, and polymers) have on different stages of amyloid aggregation. Ultimately, the modeled aggregation parameters are related to the aforementioned thermodynamic parameters, along with one intrinsic parameter for osmolytes. MATERIALS Polyvinylpyrrolidone, 40 kDa, (PVP40), (CAS No. 9003-39-8, Sigma Aldrich, St. Lois, MO), polyvinylpyrrolidone, 3.5 kDa, (PVP3.5) (276142500, ACROS Organics, Pittsburgh, PA), guanidine hydrochloride (GuHCl) (CAS No. 50-01-1, Sigma Aldrich), trimethylamine N-oxide (TMAO) (CAS No. 1184-78-7, Sigma Aldrich), D-sorbitol, (CAS No. 50-70-4, Sigma Aldrich), and glycerol, (CAS No. 56-81-5, Sigma Aldrich) were purchased from the given companies.


4

Page 5 of 26

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

Langmuir

The insulin was obtained from Novo Nordisk (Gentofte, Denmark). All other chemicals were purchased from Sigma-Aldrich. 0.2 µm SFCA filter were purchased from Fischer Scientific (Cat. No. 09-740-37F, Fisher Scientific, Pittsburgh, PA). METHODS Protocol for insulin solution preparation: All insulin solutions were prepared fresh for each experiment in 0.01 M phosphate buffer, 0.0027 M KCl, 0.137 M NaCl and adjusted to pH 1.6 to a final concentration of 2 mg/ml (344 µM) insulin monomer using a protocol developed to increase the reproducibility of insulin aggregation.21 A cosolute was added at this point to the desired concentration, and the solution was again brought to pH 1.6. Fluorescence probe, 5Carboxyrhodamine 6G succinimidyl ester (5-CR6G) (C-6127, Invitrogen, Grand Island, NY) was used to label insulin for the FCS experiments. 5-CR6G was attached to the C terminal end of insulin by following a commonly adopted protein labeling protocol.39 5-CR6G labeled insulin at an extremely low concentration of 3.5×10-3 µM was added to plain insulin aqueous solution of total 2.0 mg/ml insulin concentration for the FCS experiments. Insulin aggregation protocol: Insulin aggregation, with or without 5-CR6G labeled insulin monomers, was performed by adding 1 ml samples of insulin solution as prepared above to glass vials and heated to 65°C ± 1°C. Samples were removed periodically during aggregation and cooled to RT. Three technical replicates were prepared, and the data in Figs. 1 & 2 are all of the collected data points. A600 was measured using a Hitachi U 2000 spectrophotometer (U 2000 Double-Beam UV/VIS spectrophotometer, Hitachi Instruments Inc., Danbury, CT). Fluorescence correlation spectroscopy (FCS): FCS was set up on an inverted microscope (Zeiss Axio A1, Carl Zeiss, Germany) equipped with an oil-immersion objective lens (Plan


5

Langmuir

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 6 of 26

Aprochromat 100x, NA = 1.4) as detailed elsewhere.40 The tiny fluctuations, I(t), in fluorescence intensity, due to the motion of fluorescence labeled insulin in and out of the laser excitation volume, with an Ar laser (λ = 488 nm, Melles Griot, Rochester, NY) was measured by two photon counting modules (H7422-40, Hamamatsu, Japan) independently in a confocal detection geometry. The auto-correlation function, G(τ) of measured I(t) as = 〈 〉 41, 42, 43 , 〈 〉

where = − 〈〉, was thus obtained by using a multichannel FCS

data acquisition board and its software (ISS, Champaign, IL)41, 42, 43 via cross-correlation analysis, which removes the artifacts from detectors, to extract the diffusion coefficient, D, and concentration of fluorescent insulin, [c], in buffer solutions added with different co-solutes. The excitation focal volume was calibrated at the room temperature (∼ 25oC) by Rhodamine 6G of known D (= 280 µm2/s) in a dilute aqueous solution to be ϖ ~ 260 nm in the lateral dimension and z ~ 2 µm in the vertical dimension. The measured G(τ) was fitted to: = . 1 +

#$ %

&

1+

#$ (. '

&

(1)

to obtain D.40, 43 THEORY A four phase Aggregation Reaction Model (ARM) for insulin conversion from monomer to fibril is presented here. The model consists of monomer unfolding to an aggregation prone conformation, nucleation via monomer addition, fibrillation with both monomer and oligomer reactive species, and fragmentation of fibrils. The reaction scheme is based on a previously published insulin aggregation model combined with details from more recent publications, including conformational rearrangement and fragmentation.20, 22, 27, 44


6

Page 7 of 26

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

Langmuir

The model is summarized by the following species reactions: (1) monomer interconversion between native state and unfolded, aggregation-prone state with the rate constants kmon and kmon-, (2) oligomer production by reversible addition of unfolded monomer only with the rate constants knu and knu-, (3) fibril accumulation by reversible addition of species A1 through A5 to a growing end with the rate constants kfb and kfb, and (4) fibril fragmentation of one fibril into two fibrils until fibril concentration reaches 0.4 mg/ml with the rate constant kfrag. The key assumptions for the model are as follows: (1) all fibrils are accounted for as F due to publications from Serio and colleagues as well as Heldt and coworkers showing fibrils of all length undergo reactions only at the two fibril ends,45, 46 (2) reverse rate constants and the fragmentation rate constant are fit to the control run data, and are then fixed for the remaining aggregation runs (due to their insensitivity from a sensitivity analysis, Fig. S1), and (3) kfrag(t) is defined as λfrag if t ≤ t20% and 0 otherwise (t20% is the time until fibril concentration reaches 0.4 mg/ml). Morris and colleagues have shown fibril fragmentation ceases when human insulin fibril concentration reaches 0.4 mg/ml under very similar solution conditions.44 The species reactions and fluxes are given in Scheme 1. Using MatLab (The MathWorks, Natick, MA), nonlinear least-squares regression was utilized to minimize the sum of squared errors between the experimental data and those predicted by the model through estimation of kmon, kmon-, knu, knu-, kfb, kfb-, kfrag, and tlag (time to 5% of the final asymptote value) for the control run (no cosolute).


7

Langmuir

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 8 of 26

Scheme 1. Summary of species reactions and fluxes Reaction Diagram

(2)

i = 1, 2, …, 5

(3)

i = 1, 2, …, 5

(4) (5)

Species Reactions )*( = −+,-. )

(6)

)* = +,-. − / +.0,1 − +.0, − +34, )

(7)

12

)*1 = −+.0,1 + +.0,1 − +34,1 )

i = 2, 3, …, 5

)5 = +.0, + 63789 5 )

(8) (9)

Species Fluxes +,-. = 6,-. *( − 6,-. *

(10)

+.0,1 = 6.0 * *1 − 6.0 *1

i = 1, 2, …, 5

(11)

+34,1 = 634 *: 5 − 634 5

i = 1, 2, …, 5

(12)

RESULTS AND DISCUSSION Control Aggregation: The aggregation kinetics of human insulin with destabilizing and stabilizing osmolytes (Fig. 1) and polymer cosolute additives (Fig. 2) are shown. The results for


8

Page 9 of 26

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

Langmuir

kmon, knu, kfb, and tlag are given in Figs. 3 & 4, while the values for kmon-, knu-, kfb- and kfrag are kept constant at 3.6×10-3 h-1, 6.7×10-3 h-1, 4.5×102 h-1 and 1.3×10-3 h-1, respectively. During solving, the rate constants were constrained to be within three orders of magnitude in both directions of the control values to ensure error minima were not found in unreasonable parameter space. All of the fitting parameters fell well within this constraint. Aggregation with Osmolytes: Destabilizing osmolyte: GuHCl is a destabilizing osmolyte that denatures insulin monomer purportedly via a preferential inclusion mechanism.34, 35, 36 Adding GuHCl to the reaction mixture induces a non-linear cosolute-concentration dependent behavior on the rate of amyloid conversion from monomer to fibril. From Fig 1A, we observe that upon addition of GuHCl, aggregation kinetics are sped-up (smaller tlag) and slowed-down (greater tlag) for CGuHCl < 0.6 M and > 0.6 M, respectively. Also, for CGuHCl < 0.6 M, the values for kmon, knu and kfb are mostly all significantly above the values with no-cosolute (Fig 3). Notably these values for GuHCl are at the only cosolute concentrations where a significant increase in all three rate constants is observed. Above CGuHCl = 0.6 M, kmon and kfb return to baseline values, while knu is slightly decreased compared with the control (Fig. 3). Stabilizing osmolytes: Osmolytes that stabilize the insulin monomer have a different effect on aggregation kinetics compared with destabilizing osmolytes. All stabilizing osmolytes − glycerol, sorbitol and TMAO − either cause no change or slow down the aggregation process and increase tlag at all concentrations. Glycerol has minimal influence on insulin aggregation at CGlycerol < 1 M (Fig. 1B). At CGlycerol = 1.1 M, aggregation is delayed from tlag = 3.2 ± 0.1 to 4.1 ± 0.7 h. For sorbitol on the other hand, the aggregation reaction is delayed relative to the control


9

Langmuir

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 26

for all cosolute concentrations (Fig. 1C) and this is mainly due to a lowering of knu and kfb (Fig. 3B & 3C). Analogous aggregation effects from sorbitol occur at approximately half the molar concentration of glycerol.

Figure 1. Insulin Aggregation with Osmolytes. Aggregation kinetics of 2 mg/ml (344 µM) human insulin measured with A600 at T = 65°C and pH = 1.6 with (A) GuHCl ( ), (B) glycerol ( ), (C) sorbitol ( ) and (D) TMAO ( ) all compared with the control (× ×). Symbol colors correspond with cosolute concentrations given in the legend. The solid lines represent the ARM fit to the data.

All the kinetic aggregation runs with TMAO as a cosolute additive, except at the lowest concentration of 0.16 M, are delayed within error relative to the control (Fig. 1D). For CTMAO = 0.49 M, tlag is significantly delayed to 6.9 ± 0.6 h. Aggregation is increasingly delayed to tlag = 9.6 ± 0.5 h as the concentration of TMAO is further increased to CTMAO = 0.67 M. The significantly slowed aggregation is modeled by a greater than 3.5 times decrease in kmon and a 3fold decrease in knu for CTMAO at 0.49 M and a further decrease in both kmon and knu for CTMAO at


10

Page 11 of 26

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

Langmuir

0.67 M with accompanying decreases in kfb (Fig. 3). The greatest stabilization by osmolytes used in this study occurs at CTMAO at 0.67 M (Fig. 1).

Figure 2. Insulin Aggregation with Polymers. Aggregation kinetics of 2 mg/ml human insulin measured with A600 at T = 65°C and pH = 1.6 with (A) PVP40 ( ) and (B) PVP3.5 ( ). Symbol colors correspond with cosolute concentrations given in the legend. The solid lines represent the ARM fit to the data.

Aggregation with Polymers: To determine the effect of chain length of hydrophilic polymers on amyloid aggregation rates, PVP40 and PVP3.5 were selected. Interestingly from Fig. 2A, addition of PVP40 at CPVP40 = 0.63 mM to the insulin aggregation reaction increases the aggregation reaction rate somewhat. While at higher concentrations up to 2.5 mM, tlag clearly decreases with increasing concentration of PVP40. From a fit of the ARM for CPVP40 > 0.63 mM, kmon does not exhibit a predictable trend, knu decreases, and kfb decreases slightly with concentration of PVP40 (Figs. 4Ai, Bi & Ci, respectively). Similarly, PVP3.5 exhibits the same effects on insulin aggregation as CPVP40 > 1 mM, albeit at ∼10-fold higher concentrations (Fig. 2B). The values for knu correspondingly decrease with CPVP3.5, in addition to an initial increase with a subsequent decline for kfb (Fig. 4).


11

Langmuir

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 12 of 26

Figure 3. Osmolyte Reaction Rates. (A) kmon the monomer unfolding rate constant, (B) knu the nucleation rate constant, (C) kfb, the fibrillation rate constant, and (D) tlag are plotted against Ccosolute for GuHCl, glycerol, sorbitol, and TMAO (symbols in legend). Error bars represent the fit of ± 1 standard deviation to the aggregation data. Horizontal black solid line and dashed lines represent the control parameter (no cosolute) and ± 1 standard deviation of the control parameters.

Modeling Insights: Besides estimates of the rate constants, another important feature of the model is its ability to predict the temporal concentration of the small oligomeric species present during the lag phase of aggregation. Here, the native insulin monomer unfolds under first order kinetics with the rate constant kmon into an active conformation that exists at a concentration comparable to the native species (Fig. 5A). The unfolded monomer converts to dimers and subsequent oligomers with first order kinetics governed by knu, which results in dimer concentrations an order of magnitude lower than the native monomer concentration. Higher imer oligomers have concentration profiles orders of magnitude lower than the monomer concentration and are thus not shown. Additionally, the integral from 0 h to tfinal (time to end of aggregation run) of the native monomer curve (AMon,Fold) can provide insight into the degree to which cosolutes stabilize or destabilize the native monomer. TMAO, previously shown to be the


12

Page 13 of 26

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

Langmuir

most stabilizing osmolyte,6 increases AMon,Fold more than 6-times (Fig. 5B). The destabilizing osmolyte decreases AMon,fold at least 33% for both concentrations below 0.6 M, indicating a significant destabilization of the native monomer. Furthermore, a correlation between the nucleation reaction rate and tlag was previously reported.18 Our ARM observes a similar correlation between the nucleation rate, knu, and tlag (R2 = 0.9790, Fig. 5C).

Figure 4. Polymer Reaction Rates. (A) kmon the monomer unfolding rate constant, (B) knu the nucleation rate constant, (C) kfb, the fibrillation rate constant, and (D) tlag are plotted against Ccosolute for PVP40 and PVP3.5 (same symbols in Fig. 2). The inserts (Ai – Di) are an expansion of the dashed region. Error bars represent the estimate of the fit associated with ± 1 standard deviation of the empirical parameters. Horizontal black solid line and dashed lines represent the control parameter (no cosolute) and ± 1 standard deviation of the control parameters.

Relating Thermodynamic Parameters to Aggregation Behavior: Correlating and ultimately predicting the complex effects of cosolutes on amyloid aggregation from a single thermodynamic or structural measurement are desirable. To this end, we demonstrate in Fig. 6 that the relative preferential interaction coefficient, ∆∆Γµ, and the neutral surface area, SAo, of osmolytes both correlate with the lag time, tlag (or the nucleation rate constant knu) (Fig. 6). ∆∆Γµ, is defined as follows: the preferential interaction coefficient with a generalized macromolecule (represents insulin), Γµcosolute, minus that with the solvent, Γµsolvent, is divided by (Γµcosolute - Γµsolvent) at


13

Langmuir

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 14 of 26

Ccosolute = 0 M. The data for ∆∆Γµ was estimated from a previous publication listing Γµ values for various cosolutes with a model protein using vapor pressure osmometry.37

Figure 5. Modeling Insights. (A) Temporal concentration profiles for the control aggregation run of the native folded monomer, the unfolded monomer and the dimer belong to the primary y-axis. The ARM fit to the control aggregation and the control aggregation data correspond with the secondary y-axis. Symbols given in the legend below. (B) Integral of the native, folded monomer from 0 h to tfinal for TMAO and GuHCl. (C) ln knu correlated with tlag. Symbols given in the legend to the right. Error bars represent the fit from ± 1 standard deviation of the model. Horizontal black solid line and dashed lines represent the control parameter (no cosolute) and ± 1 standard deviation of the control parameters for (B).

A good correlation between tlag and ∆∆Γµ for osmolytes for which data is available excluding the two high concentration data points (CGuHCl > 0.6 M) (R2 = 0.8515, Fig. 6A) is obtained.37 However, the good correlation is lost once these two data points for GuHCl are included in the analysis (R2 = 0.1943). The possibility remains that there is a crossover effect for ∆∆Γµ > 100%, i.e. tlag reaches a minimum at ∆∆Γµ ~ 105%.


14

Page 15 of 26

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

Langmuir

The second correlation is from an intrinsic structural property of the osmolytes. The neutral surface area, SAo, of an osmolyte has previously been shown to correlate strongly with tlag.6 Reported by Street and colleagues, SAo is determined computationally using PyMOL with a probe radius of 1.4 Å and atomic coordinates for the osmolytes from the HICUP database.47 Due to the intrinsic nature of SAo, selecting a single osmolyte concentration for the correlation would be arbitrary. Instead the slope of a linear regression between tlag and Ccosolute, ∆tlag, is plotted against SAo and yields a strong correlation (R2 = 0.9530, Fig. 6B). Because of the biphasic behavior of GuHCl and the fact that CGuHCL > 0.6 M does not correlate with ∆∆Γµ, a feature not observed by any other osmolyte, the linear regression between tlag and CGuHCl was taken only for CGuHCl < 0.6 M. If all concentrations of GuHCl are used to determine ∆tlag, the correlation between ∆tlag and SAo is lost.

Figure 6. Relating Thermodynamic Properties to Kinetic Effects. (A) Correlation between tlag and ∆∆Γµ estimated from Anderson et al. for stabilizing osmolytes and CGuHCl < 0.6 M (R2 = 0.8515). The R2 value for the correlation with all points is 0.1943. (B) Correlation between the slope of tlag v Ccosolute (∆tlag) with the neutral accessible surface area of the osmolyte (SAo). Symbols given in the legend below. Error bars represent the fit from ± 1 standard deviation of the model.

Diffusion Effects on Aggregation: For diffusion limitations to influence amyloid aggregation, the Damköhler number (Da = kL2D-1), is expected to be > ~1. k and L are defined as a reaction rate constant and the length scale for diffusion, respectively. To investigate these limitations, insulin monomer diffusion coefficients (D) are experimentally measured using FCS in the presence of cosolutes, and k is taken from the ARM. Here we examine two cases: (1) diffusion


15

Langmuir

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 16 of 26

limitations during the nucleation reaction in the absence of cosolute, and (2) the diffusion limitations of PVP40 at 100 mg/ml on the fibrillation reaction. In case (1), knu = 1.4×10-1 mM h-1 (Fig. 3B), D = 90 µm2/s (Fig. 7), and L is estimated to be 40 nm, which is a conservative estimate based on a particle diameter of 2.75 nm for insulin monomer22 and a volume fraction of 0.002.48 From these parameters, Da = 6.9×10-10 indicating no diffusion limitations on the nucleation reaction. For the nucleation reaction to be diffusion limited, D would have to decrease 9 orders of magnitude or the reaction rate increase by the same amount.

Figure 7. Diffusional Limitations on Aggregation. Insulin monomer diffusion coefficient, D, as a function of cosolute concentration, Ccosolute, measured using FCS. Symbols given in the legend to the right. Error bars represent 1 standard deviation from at least 3 measurements.

The fibrillation reaction with CPVP40 = 100 mg/ml represents conditions much closer to a diffusion limitation. For case (2), kfb = 1.6×106 mM h-1 (Fig. 4Ci), D = 19 µm2/s (Fig. 7), and L is estimated to be 100 nm, which is based on the fibril size distribution from Pease and colleagues22 and a comparable volume fraction to case (1).48 The increase in k and decrease in D from case (1) to case (2) results in Da = 2.4×10-1, a situation where diffusion effects are more prominent. In general, diffusion will only affect the fibrillation reaction for cosolutes that decrease D close to an order of magnitude. The nucleation reaction remains uninfluenced by diffusion in all cases.


16

Page 17 of 26

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

Langmuir

Further Discussion: Osmolytes and polymers can have wildly different effects on amyloid aggregation, as seen by the differing influence on the aggregation and reaction rate constants (Figs. 1-4). Osmolytes, specifically GuHCl, can have varied effects on protein aggregation depending on the osmolyte concentration (Fig. 1A). At CGuHCl < 0.6 M the rate of amyloid formation increases via a sharp increase in kmon, knu, and kfb (Figs. 1A & 3). This behavior correlates well with other osmolytes that are known to behave through a preferential inclusion/exclusion mechanism (R2 = 0.8515, Fig. 6A).32, 34, 35, 36, 49 However at CGuHCl > 0.6 M, the aggregation effects of GuHCl are different from CGuHCl < 0.6 M (Figs. 1A & 3). In addition, the correlation between ∆∆Γµ and tlag is lost (R2 = 0.1943) indicating that GuHCl at high concentrations no longer affects amyloid aggregation solely through the preferential inclusion mechanism. The stabilizing osmolytes examined here generally affect the amyloid aggregation via decreases in both kmon and knu, a result in good agreement with the known mechanisms for the stabilizing effect of polyols and TMAO.32, 49 The osmolyte with the largest stabilizing effect, TMAO, is known to affect protein stability through multiple mechanisms.32 Here we showed that TMAO was the most stabilizing in terms of lowering kmon and knu (Fig. 3A & B). Further, TMAO increased AMon,Fold greater than 6 times compared with the control, far more than any other osmolyte. Polymers delay amyloid aggregation via a different mechanism than that of osmolytes. PVP40 and PVP3.5 both slow nucleation and extend tlag except for PVP40 at 0.63 mM (Fig. 2). Thus, according to the ARM, knu, and kfb both increase slightly at low concentrations then decrease with increasing concentration (Fig. 4). This implies the polymer’s mechanism for effecting amyloid growth involves changes to solely the later stages of aggregation, an effect not generally


17

Langmuir

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 18 of 26

observed for osmolytes (Figs. 3 & 4). Notably PVP3.5 begins effecting aggregation at a concentration 10 fold greater than PVP40 (Fig. 2). Further the volume fraction, φ, for PVP40 at 100 mg/ml compared with PVP3.5 at 100 mg/ml is about 10 times higher (φPVP40,100mg/ml = 0.016, φPVP3.5,100mg/ml = 0.0014). Therefore the effects of low concentrations of polymers on amyloid aggregation scale with φ and not concentration, a result that agrees with known mechanisms of polymer effects on protein folding.38 Despite nearly an order of magnitude decrease in D from CPVP40 = 100 mg/ml (Fig. 7), the nucleation reaction remains unaffected by diffusion and hence is non-diffusion limited. Only the fibrillation reaction at CPVP40 = 100 mg/ml limited. Only the fibrillation reaction at CPVP40 = 100 mg/ml approaches diffusion limitations, indicating the rate limiting reaction for amyloid aggregation, i.e. the nucleation reaction, is not influenced by polymeric induced diffusion limitations. Besides the rate constants, fitting the ARM to the experimental kinetic conversion data also provides temporal concentration profiles of each i-mer (Fig. 5A). This information can predict when critical species are most prevalent during amyloid aggregation. In Alzheimer’s disease, the smallest synaptotoxic species was demonstrated to be the dimer.50 The ability to predict when this difficult-to-quantify oligomer is at its highest concentration from macroscopic aggregation measurements is a feature of this ARM that others lack.17, 44 Additionally, the AMon,Fold for GuHCl and TMAO can be estimated (Fig. 5B). Consistent with TMAO’s ability as a stabilizing osmolyte, AMon,Fold drastically increases compared with the control. Addition of GuHCl slightly decreases AMon,Fold at 0.26 M GuHCl, which is in agreement with GuHCl being a destabilizing osmolyte (Fig. 5B).


18

Page 19 of 26

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

Langmuir

It is desirable to be able to relate changes in aggregation behavior directly with biophysical rates predicted from the model. As has been shown previously, the reaction rate at the nucleation step correlates well with tlag.18 Here we show that knu correlates very well with tlag (R2 = 0.9790, Fig. 5C) and no other rate constant correlates as strongly as with tlag (data not shown). Clearly this implicates the nucleation reaction as the rate-limiting step of insulin amyloid formation. Ultimately, it would be powerful to be able to predict the aggregation effects of a cosolute before quantifying it experimentally. Here, it is shown that ∆∆Γµcosolute is in good correlation (R2 = 0.8515) with tlag for stabilizing osmolytes for which data are available and low concentrations of a destabilizing osmolyte (CGuHCl < 0.6 M) (Fig. 6A). Preferential inclusion/exclusion, which can be an interpretation of Γµ, is a well-studied mechanism for osmolyte effects on aggregation.6, 14, 18, 36, 37, 49, 51, 52

The correlation shown in Fig. 6A supports this mechanism. Interestingly, at

CGuHCl > 0.6 M the correlation is lost (R2 = 0.1943) indicating that at higher concentrations of GuHCl, preferential inclusion is no longer the only effect GuHCl has on amyloid aggregation. Additional mechanism(s) likely contribute to GuHCl’s complex effect, and further work is necessary to elucidate its full behavior. As has previously been shown,6 and is reinforced here (Fig. 6B), SAo strongly correlated with tlag for insulin aggregation (R2 = 0.9530). SAo is ideal for correlating with osmolyte effects on aggregation since it is an intrinsic property that can be calculated by knowing only the structure of the osmolyte. Additionally, SAo was previously used to explain protein unfolding free energies by relating the free energy of water transfer from a protein backbone to an osmolyte. This leads to the conclusion that aggregation behavior, which is presumed to be controlled by preferential inclusion/exclusion of water and osmolytes to the protein backbone,6, 14, 18, 36, 37, 49, 51,


19

Langmuir

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

52

Page 20 of 26

is in turn governed by the neutral surface area (or the non-neutral surface area) of the

osmolyte. Thus osmolyte effects on amyloid aggregation can be predicted once the structure of the osmolyte is known. CONCLUSIONS Amyloid aggregation is undoubtedly affected in vivo in ways not currently captured by in vitro experiments. We show that destabilizing osmolytes, stabilizing osmolytes and polymers have drastic effects on the aggregation of a model amyloid, insulin, through A600 measurements and analysis with an Amyloid Reaction Model (ARM). GuHCl increases the rate of aggregation at CGuHCl < 0.6 M, mainly through increases in kmon and knu. Contrary to behavior at CGuHCl < 0.6 M, aggregation is delayed at CGuHCl > 0.6 M, modeled by a decrease in knu. Polyols delay aggregation through a joint decrease in kmon and knu. Sorbitol consistently delays aggregation at molar concentrations half that of glycerol. TMAO reduces aggregation the most of any osmolyte, by means of a reduction in kmon, knu and kfb. Interestingly, all osmolytes effect aggregation so that tlag correlates with ∆∆Γµ except for CGuHCl > 0.6 M, which implies that the preferential inclusion/exclusion mechanism, thought to be the dominant mechanism for osmolyte interaction with proteins, no longer holds at these conditions. Polymers inhibit aggregation differently from osmolytes in that they consistently increase tlag. These effects decrease knu, and initially increase kfb but continually decrease kfb thereafter. Further, the mechanism appears in good agreement with the volume exclusion principle since the concentration required for affecting aggregation for PVP3.5 is about 10 fold higher than the concentration for PVP40, which corresponding with a 10 fold reduction in φ for PVP3.5 at equal concentrations. Nearly an order of magnitude reduction in D result in no meaningful difference


20

Page 21 of 26

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

Langmuir

in the diffusion limitations of the nucleation reaction, which is the rate limiting step of amyloid oligomer aggregation. Only > 9 order of magnitude decrease in D would result in diffusion limitations on the nucleation reaction. Fibril growth, on the other hand is much closer to diffusion limitations. While ∆∆Γµ correlates well with tlag, SAo has an even stronger correlation with tlag (R2 = 0.9530). It is likely that this relationship underlies the intrinsic property of osmolytes which governs the preferential interactions that effect protein folding and aggregation, which in turn can be used to predict osmolyte effects on aggregation.


21

Langmuir

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 22 of 26

AUTHOR INFORMATION Corresponding Author * To whom correspondence should be addressed: Georges Belfort; Tel.: +1 5182766948, Fax: +1 5182764030, Email: [email protected]. ACKNOWLEDGMENT We thank Arne Staby (Novo Nordisk A/S, Denmark) for supplying the human insulin. Thanks are also due to Mirco Sorci for valuable laboratory assistance, along with Shekhar Garde and Vasudevan Venkateshwaran for useful discussion (Rensselaer Polytechnic Institute).

This

research was supported by NSF Grant CTS-64-00610. ZZ and YZ also acknowledge the financial support by NSF (CMMI-1129821). SUPPORTING INFORMATION A sensitivity analysis of the Aggregation Reaction Model parameters and the reaction rate constants for the data shown in Figs. 1 and 2 are given in the SI. This material is available free of charge via the Internet at http://pubs.acs.org. REFERENCES 1. Chiti, F.; Dobson, C. M. Protein misfolding, functional amyloid, and human disease. Annual review of biochemistry 2006, 75, 333-66. 2. Zimmerman, S. B.; Minton, A. P. Macromolecular crowding: biochemical, biophysical, and physiological consequences. Annual review of biophysics and biomolecular structure 1993, 22 (1), 27-65. 3. Zimmerman, S. B.; Trach, S. O. Estimation of macromolecule concentrations and excluded volume effects for the cytoplasm of Escherichia coli. Journal of molecular biology 1991, 222 (3), 599-620. 4. Senske, M.; Törk, L.; Born, B.; Havenith, M.; Herrmann, C.; Ebbinghaus, S. Protein Stabilization by Macromolecular Crowding through Enthalpy rather than Entropy. Journal of the American Chemical Society 2014. 5. Politi, R.; Harries, D. Enthalpically driven peptide stabilization by protective osmolytes. Chemical Communications 2010, 46 (35), 6449-6451.


22

Page 23 of 26

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

Langmuir

6. Nayak, A.; Lee, C. C.; McRae, G. J.; Belfort, G. Osmolyte controlled fibrillation kinetics of insulin: New insight into fibrillation using the preferential exclusion principle. Biotechnology progress 2009, 25 (5), 1508-14. 7. Zou, Q.; Bennion, B. J.; Daggett, V.; Murphy, K. P. The molecular mechanism of stabilization of proteins by TMAO and its ability to counteract the effects of urea. Journal of the American Chemical Society 2002, 124 (7), 1192-1202. 8. Ellis, R. J.; Minton, A. P. Protein aggregation in crowded environments. Biological chemistry 2006, 387 (5), 485-497. 9. White, D. A.; Buell, A. K.; Knowles, T. P.; Welland, M. E.; Dobson, C. M. Protein aggregation in crowded environments. Journal of the American Chemical Society 2010, 132 (14), 5170-5175. 10. Minton, A. P. Influence of macromolecular crowding upon the stability and state of association of proteins: predictions and observations. Journal of pharmaceutical sciences 2005, 94 (8), 1668-1675. 11. Rösgen, J.; Pettitt, B. M.; Bolen, D. W. Protein folding, stability, and solvation structure in osmolyte solutions. Biophysical Journal 2005, 89 (5), 2988-2997. 12. Harries, D.; Rösgen, J. A practical guide on how osmolytes modulate macromolecular properties. Methods in cell biology 2008, 84, 679-735. 13. Cheung, M. S.; Klimov, D.; Thirumalai, D. Molecular crowding enhances native state stability and refolding rates of globular proteins. Proceedings of the National Academy of Sciences of the United States of America 2005, 102 (13), 4753-4758. 14. Timasheff, S. N. Protein-solvent preferential interactions, protein hydration, and the modulation of biochemical reactions by solvent components. Proceedings of the National Academy of Sciences 2002, 99 (15), 9721-9726. 15. Pai, S. S.; Przybycien, T. M.; Tilton, R. D. Protein PEGylation Attenuates Adsorption and Aggregation on a Negatively Charged and Moderately Hydrophobic Polymer Surface. Langmuir 2010, 26 (23), 18231-18238. 16. Kim, Y. S.; Randolph, T. W.; Stevens, F. J.; Carpenter, J. F. Kinetics and energetics of assembly, nucleation, and growth of aggregates and fibrils for an amyloidogenic protein Insights into transition states from pressure, temperature, and co-solute studies. Journal of Biological Chemistry 2002, 277 (30), 27240-27246. 17. Sukenik, S.; Harries, D. Insights into the disparate action of osmolytes and macromolecular crowders on amyloid formation. Prion 2012, 6 (1), 26-31. 18. Sukenik, S.; Politi, R.; Ziserman, L.; Danino, D.; Friedler, A.; Harries, D. Crowding Alone Cannot Account for Cosolute Effect on Amyloid Aggregation. PloS one 2011, 6 (1). 19. Nielsen, L.; Khurana, R.; Coats, A.; Frokjaer, S.; Brange, J.; Vyas, S.; Uversky, V. N.; Fink, A. L. Effect of Environmental Factors on the Kinetics of Insulin Fibril Formation: Elucidation of the Molecular Mehcanism. Biochemistry 2001, 40, 6036-6046. 20. Lee, C. C.; Nayak, A.; Sethuraman, A.; Belfort, G.; McRae, G. J. A three-stage kinetic model of amyloid fibrillation. Biophys J 2007, 92 (10), 3448-58. 21. Heldt, C. L.; Sorci, M.; Posada, D.; Hirsa, A.; Belfort, G. Detection and reduction of microaggregates in insulin preparations. Biotechnology and bioengineering 2011, 108 (1), 23741. 22. Pease, L. F., 3rd; Sorci, M.; Guha, S.; Tsai, D. H.; Zachariah, M. R.; Tarlov, M. J.; Belfort, G. Probing the nucleus model for oligomer formation during insulin amyloid fibrillogenesis. Biophys J 2010, 99 (12), 3979-85.


23

Langmuir

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 24 of 26

23. Sorci, M.; Grassucci, R. A.; Hahn, I.; Frank, J.; Belfort, G. Time-dependent insulin oligomer reaction pathway prior to fibril formation: cooling and seeding. Proteins 2009, 77 (1), 62-73. 24. Knowles, T. P.; Waudby, C. A.; Devlin, G. L.; Cohen, S. I.; Aguzzi, A.; Vendruscolo, M.; Terentjev, E. M.; Welland, M. E.; Dobson, C. M. An analytical solution to the kinetics of breakable filament assembly. Science 2009, 326 (5959), 1533-7. 25. Nayak, A.; Sorci, M.; Krueger, S.; Belfort, G. A universal pathway for amyloid nucleus and precursor formation for insulin. Proteins 2009, 74 (3), 556-65. 26. Munishkina, L. A.; Ahmad, A.; Fink, A. L.; Uversky, V. N. Guiding protein aggregation with macromolecular crowding. Biochemistry 2008, 47 (34), 8993-9006. 27. Vestergaard, B.; Groenning, M.; Roessle, M.; Kastrup, J. S.; van de Weert, M.; Flink, J. M.; Frokjaer, S.; Gajhede, M.; Svergun, D. I. A helical structural nucleus is the primary elongating unit of insulin amyloid fibrils. PLoS biology 2007, 5 (5), 1089-1097. 28. Gibson, T. J.; Murphy, R. M. Inhibition of insulin fibrillogenesis with targeted peptides. Protein science : a publication of the Protein Society 2006, 15 (5), 1133-41. 29. Gekko, K. Calorimetric study on thermal denaturation of lysozyme in polyol-water mixtures. Journal of biochemistry 1982, 91 (4), 1197-1204. 30. Kumar, A.; Attri, P.; Venkatesu, P. Effect of polyols on the native structure of αchymotrypsin: A comparable study. Thermochimica Acta 2012, 536, 55-62. 31. Sukenik, S.; Sapir, L.; Harries, D. Balance of enthalpy and entropy in depletion forces. Current Opinion in Colloid & Interface Science 2013, 18 (6), 495-501. 32. Ma, J.; Pazos, I. M.; Gai, F. Microscopic insights into the protein-stabilizing effect of trimethylamine N-oxide (TMAO). Proceedings of the National Academy of Sciences of the United States of America 2014, 111 (23), 8476-8481. 33. Miyawaki, O.; Dozen, M.; Nomura, K. Thermodynamic analysis of osmolyte effect on thermal stability of ribonuclease A in terms of water activity. Biophysical Chemistry 2014, 185, 19-24. 34. Greene, R. F.; Pace, C. N. Urea and guanidine hydrochloride denaturation of ribonuclease, lysozyme, α-chymotrypsin, and β-lactoglobulin. Journal of Biological Chemistry 1974, 249 (17), 5388-5393. 35. Brems, D. N.; Brown, P. L.; Heckenlaible, L. A.; Frank, B. H. Equilibrium denaturation of insulin and proinsulin. Biochemistry 1990, 29 (39), 9289-93. 36. Schneider, C. P.; Trout, B. L. Investigation of Cosolute-Protein Preferential Interaction Coefficients: New Insight into the Mechanism by Which Arginine Inhibits Aggregation. Journal of Physical Chemistry B 2009, 113 (7), 2050-2058. 37. Anderson, C.; Courtenay, E.; Record, M. Thermodynamic expressions relating different types of preferential interaction coefficients in solutions containing two solute components. The Journal of Physical Chemistry B 2002, 106 (2), 418-433. 38. Soranno, A.; Koenig, I.; Borgia, M. B.; Hofmann, H.; Zosel, F.; Nettels, D.; Schuler, B. Single-molecule spectroscopy reveals polymer effects of disordered proteins in crowded environments. Proceedings of the National Academy of Sciences 2014, 111 (13), 4874-4879. 39. Hentz, N.; Richardson, J.; Sportsman, J.; Daijo, J.; Sittampalam, G. Synthesis and characterization of insulin-fluorescein derivatives for bioanalytical applications. Analytical Chemistry 1997, 69 (24), 4994-5000.


24

Page 25 of 26

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

Langmuir

40. Wang, S.; Chang, H.; Zhu, Y. Hysteretic Conformational Transition of Single Flexible Polyelectrolyte under Resonant AC Electric Polarization. Macromolecules 2010, 43 (18), 74027405. 41. Sukhishvili, S. A.; Chen, Y.; Muller, J. D.; Gratton, E.; Schweizer, K. S.; Granick, S. Surface diffusion of poly(ethylene glycol). Macromolecules 2002, 35 (5), 1776-1784. 42. Magde, D.; Elson, E.; Webb, W. W. Thermodynamic Fluctuations in a Reacting System− Measurement by Fluorescence Correlation Spectroscopy. Physical Review Letters 1972, 29 (11), 705-708. 43. Rigler, R.; Mets, U.; Widengren, J.; Kask, P. Fluorescence Correlation Spectroscopy with High Count Rate and Low-Background - Analysis of Translational Diffusion. Eur. Biophys. J. Biophys. Lett. 1993, 22 (3), 169-175. 44. Morris, R. J.; Eden, K.; Yarwood, R.; Jourdain, L.; Allen, R. J.; MacPhee, C. E. Mechanistic and environmental control of the prevalence and lifetime of amyloid oligomers. Nature communications 2013, 4, 1891. 45. Serio, T. R.; Cashikar, A. G.; Kowal, A. S.; Sawicki, G. J.; Moslehi, J. J.; Serpell, L.; Arnsdorf, M. F.; Lindquist, S. L. Nucleated conformational conversion and the replication of conformational information by a prion determinant. Science 2000, 289 (5483), 1317-1321. 46. Heldt, C. L.; Zhang, S. Q.; Belfort, G. Asymmetric amyloid fibril elongation: A new perspective on a symmetric world. Proteins-Structure Function and Bioinformatics 2011, 79 (1), 92-98. 47. Street, T. O.; Bolen, D. W.; Rose, G. D. A molecular mechanism for osmolyte-induced protein stability. Proceedings of the National Academy of Sciences 2006, 103 (38), 13997-14002. 48. Hao, T.; Riman, R. E. Calculation of interparticle spacing in colloidal systems. Journal of Colloid and Interface Science 2006, 297 (1), 374-377. 49. Gilman-Politi, R.; Harries, D. Unraveling the Molecular Mechanism of Enthalpy Driven Peptide Folding by Polyol Osmolytes. Journal of Chemical Theory and Computation 2011, 7 (11), 3816-3828. 50. Shankar, G. M.; Li, S.; Mehta, T. H.; Garcia-Munoz, A.; Shepardson, N. E.; Smith, I.; Brett, F. M.; Farrell, M. A.; Rowan, M. J.; Lemere, C. A.; Regan, C. M.; Walsh, D. M.; Sabatini, B. L.; Selkoe, D. J. Amyloid-beta protein dimers isolated directly from Alzheimer's brains impair synaptic plasticity and memory. Nature medicine 2008, 14 (8), 837-42. 51. Auton, M.; Bolen, D. W.; Rosgen, J. Structural thermodynamics of protein preferential solvation: osmolyte solvation of proteins, aminoacids, and peptides. Proteins 2008, 73 (4), 80213. 52. Kane, R. S.; Deschatelets, P.; Whitesides, G. M. Kosmotropes form the basis of proteinresistant surfaces. Langmuir 2003, 19 (6), 2388-2391.


25

Langmuir

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 26 of 26

TABLE OF CONTENTS GRAPHIC


26

Cosolute Effects on Amyloid Aggregation in a ... - ACS Publications

Recommend Documents