β-Hairpin Crowding Agents Affect α-Helix Stability in Crowded

Jan 13, 2016 - Bryanne Macdonald, Shannon McCarley, Sundus Noeen, and Alan E. van Giessen. Department of Chemistry, Mount Holyoke College, South ...
0 downloads 0 Views 892KB Size
Subscriber access provided by Flinders University Library

Article

#-Hairpin Crowding Agents Affect Alpha Helix Stability in Crowded Environments Bryanne Macdonald, Shannon Logan McCarley, Sundus Noeen, and Alan E. van Giessen J. Phys. Chem. B, Just Accepted Manuscript • DOI: 10.1021/acs.jpcb.5b10575 • Publication Date (Web): 13 Jan 2016 Downloaded from http://pubs.acs.org on January 18, 2016

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.

The Journal of Physical Chemistry B 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 35

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

The Journal of Physical Chemistry

β-hairpin Crowding Agents Affect Alpha Helix Stability in Crowded Environments

Bryanne Macdonald, Shannon McCarley†, Sundus Noeen†, and Alan E. van Giessen*

Department of Chemistry Mount Holyoke College South Hadley, MA 01075

* Corresponding author: Email: [email protected] Phone: (413) 538-2449 Fax: (413) 538-2327 †

These authors contributed equally.

ACS Paragon Plus Environment

1

The Journal of Physical Chemistry

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 35

Abstract The dense, heterogeneous cellular environment is known to affect protein stability. It is now recognized that attractive “quinary” interactions with other biomacromolecules in the cell, referred to as the crowding agents, play a significant role in determining the stability of the protein of interest or test protein. These attractive interactions can reduce or overcome the stabilizing effect of the excluded volume of the crowding agents. However, the roles of specific interactions, such as hydrogen bonding and side chain-side chain hydrophobic interactions, is still unclear. Here, we use molecular simulation to investigate the roles played by hydrophobic interactions and hydrogen bonding between a small helical test protein and equally-sized crowding agent proteins in a fixed β-hairpin configuration. The test protein and crowding agents are represented by a coarse-grained protein model and we use multicanonical molecular dynamics to study the folding thermodynamics of the test protein. Our results confirm that the stability of the test protein depends on the hydrophobicity of the crowding agents and that the stability of the test protein is reduced through favorable side chain-side chain interactions that preferentially stabilize the unfolded states. In addition, we show that when the intermolecular hydrophobic interactions are more favorable than the intramolecular hydrophobic interactions, the β-rich crowding agents can completely destabilize the test protein, causing it to adopt configurations with increased β-content and preventing it from forming its native helical state. Similarities between our results and those seen in the formation of amyloid fibrils are also discussed.

ACS Paragon Plus Environment

2

Page 3 of 35

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

The Journal of Physical Chemistry

Introduction The

cellular

environment

in

which

proteins

fold

is

highly

concentrated

with

biomacromolecules. This crowded environment is known to affect the stability of proteins relative to dilute solution.1 Quinary interactions play an important, though not yet fully understood, role in modulating protein stability.2 Past work on the effect that a crowded environment has on protein stability has used large, inert macromolecules to mimic the dense surroundings of the cell.3 Theoretical and computational studies often employ the use of inert spheres that are able to interact with a protein exclusively through excluded volume interactions,4 while experimental studies often use inert molecules, such as dextran or Ficoll,5 as crowding agents. It is now widely accepted that it is necessary to include attractive interactions between a protein of interest, referred to as the “test protein,” and the other macromolecules within a cell, termed the “crowding agents,” to more accurately depict the cellular milieu.6-12 Recently there have been several studies of protein stability in more realistic contexts that include not only excluded volume effects, but also attractive interactions, in silico,7,11 in vitro,6,20-25 and in vivo.6,20-25 These studies have begun to elucidate the effect that the chemical nature of the crowding agents has on the stability of the protein, and have shown that using proteins as crowding agents is a vital step towards more accurately modeling the cellular environment. However, these studies do not address the role of amino acid sequence in the crowding agents. While there has been work that varies the strength of the attractive force between a test protein and its local environment, the focus was either on protein association in the presence of spherical crowders8,26 or on protein stability in a chaperone cavity.27-29 To address this, we have begun a systematic study of the role of crowding agent hydrophobicity in affecting the stability of a helical test protein.

ACS Paragon Plus Environment

3

The Journal of Physical Chemistry

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 35

Our previous work30 used identical α-helical crowding agents to study the effect that crowding agent hydrophobicity has on the stability of a small α-helical test protein, (AAQAA)3. We found that the stability of the test protein depended on the hydrophobicity of the crowding agents through three mechanisms. For low crowder hydrophobicity, the test protein was stabilized relative to dilute solution via the dominant excluded volume effect. At intermediate crowding agent hydrophobicities, the test protein was destabilized by attractive side chain-side chain interactions that preferentially stabilized the unfolded states. At high crowder hydrophobic values, the protein was once again stabilized, this time by strong intermolecular attractions that caused the formation of a packed structure consisting of the test protein and four crowding agents. However, the helical nature of the crowding agents limited the role that hydrogen bonding played in these systems.

This work builds upon our previous work by examining the effect that hydrophobicity plays when the secondary structure of the crowding agents differs from that of the test protein. We maintain the identity of the test protein by using the α-helical protein (AAQAA)3, which has been studied both experimentally31,32 and computationally.33-35 For crowding agents, we use βhairpins of equal length and shape, varying the hydrophobicity for each simulation. In our model, each side chain is represented by a single interaction site of uniform size, regardless of the chemical identity of the residue. As a consequence, changes from one crowder sequence to another are a result solely of the hydrophobic interaction and not due to any size effects. As in our previous study, we recognize that while this is an advance over the use of physically

ACS Paragon Plus Environment

4

Page 5 of 35

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

The Journal of Physical Chemistry

unrealistic spherical crowders, our system still deviates from the true environment of the cell because of the uniformity of the crowding agents.

The presence of exposed β-sheet edges from the β-hairpin crowding agents resembles the ends of amlyloid fibrils that are an essential element of amyloidogenesis.36 Hence, the results of this work have implications in understanding protein-folding diseases such as Alzheizmer’s Disease by relating the amyloid propensity to the hydrophobicity of the pre-existing β-structure. In this, our work is similar in spirit to experimental research looking at the aggregation propensity of Amyloid-β in which replacement of valine 18 of the Aβ(1-40) peptide with more hydrophobic amino acids showed an increase in aggregation rates.37

Computational model and methodology The test protein and the crowding agents were both modeled using the coarse-grained peptide model of Bereau and Deserno,38 the same model used in our earlier work.30 Each residue is modeled by three backbone interaction sites and a single interactions site representing the side chain, shown schematically in Figure 1a. The implicit solvent scheme for the side chain-side chain interactions is based on an analysis of the Protein Data Bank.39 Charges are implicitly included in the statistics-based interaction potential. Each residue is assigned a hydrophobicity ε (0 ≤ ε ≤ 1), with 1 being the most hydrophobic. The parameter ε scales the strength of the hydrophobic attraction. One advantageous feature of this model is that each side chain is assigned the same size. While this differs from the true molecular nature of the side chains, this uniformity of size means that each of our crowding agents will be exactly the same size, regardless of sequence. Therefore, when the sequence of the crowders is changed from one

ACS Paragon Plus Environment

5

The Journal of Physical Chemistry

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 35

simulation to the next, the resulting changes in behavior are due solely to the strength of the hydrophobic interaction and not to size effects.

Simulation parameters are similar to those in our previous work,30 though the current systems are considerably larger: each simulation consisted of a single (AAQAA)3 test protein and twenty nine homopeptide crowding agent molecules of equal size. The seven crowding agents were chosen to include an excluded volume

b

a

only control, and six crowder sequences

C=O

that range from very hydrophilic (Lys)15,

NH CαH

(ε = 0.00) to very hydrophobic (Leu)15 (ε

R

C=O

= 1.00). The crowding agents and their NH

hydrophobicities are given in Table 1 in

Figure 1. a) The protein model. b) The folded test the results section. The simulation box protein (left), and two views of the β-hairpin had side lengths of 65 Å, approximately crowder (center, right), shown to scale. Though not explicitly included in the model, the location of the backbone H and O atoms are determined via the local geometry and are included for clarity. Protein images were generated using VMD.40

twice the length of the extended test protein (33.5 Å) and a total density of 300 mg/mL. Data for the end-to-end vector for each system are given in Figure S1.

In order to determine the influence that crowding agent secondary structure plays in the stability of the test protein, the crowding agents were fixed in a β-hairpin in the same manner as our previous work. The structures of the folded test protein and a crowding agent molecule are

ACS Paragon Plus Environment

6

Page 7 of 35

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

The Journal of Physical Chemistry

shown in Figure 1b. Our previous work used helical crowding agents. It was observed that hydrogen bonding between the test protein and the crowding agents was minimal, and this was ascribed to the limited number of available hydrogen bond donors on the crowding agents: only the residues at the ends of the crowder helices were able to form hydrogen bonds with the test protein. With the crowders fixed in the β-hairpin configuration, each residue is capable of forming hydrogen bonds.

Each system was simulated using the replica exchange statistical temperature molecular dynamics (RESTMD) algorithm.41,42 Simulation parameters are similar to those used in reference [(30)]. Briefly, RESTMD is a “flat histogram” algorithm after Wang and Landau43 that dynamically refines the statistical temperature T(U) throughout the course of the simulations. Once converged, the density of states, Ω(U), is calculated using the relation,

−1

−1

 ∂S(U)   ∂lnΩ(U)  T (U) =  =  ∂U   ∂U 

,

(1)

where kB = 1. Each replica α covers one of M temperature subranges (1 ≤ α ≤ M). Replica exchanges between subranges were attempted every 2000 timesteps. RESTMD determines the statistical temperature Tα (U) in each subrange on a progressively finer grid. The grid spacing is rescaled every 1.0 × 107 timesteps. After 1.0 × 108 timesteps, Tα (U) has converged to the statistical temperature and each replica samples its entire temperature subrange with a U

generalized weight Wα = exp{= Sα (U)} . Integrating eq. 1, Sα (U) = ∫ 1 / Tα (z)dz , produces the

ACS Paragon Plus Environment

7

The Journal of Physical Chemistry

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 35

microcanonical entropy Sα (U) . Each replica saved 25,000 configurations, separated by 2000 timesteps, for post-sumulation analysis using the statistical temperature weighted histogram analysis method (ST-WHAM).44

All simulations used M = 32 replicas. The temperature ranges were 150 to 450 K for the isolated test protein and 280 to 420 K for the test protein and crowding agents.

Results The temperature at which the fluctuations in the overlap parameter χ were at a maximum was taken to be the folding temperature of the test protein, Tf. The overlap parameter is defined by,

2 χ= 2 Nα − 5Nα + 6

Nα −3 Nα

∑ ∑ Θ (σ − r − r ) ij

N ij

.

(2)

i=1 j=i+3

In the helical native state, χ = 1. In eq 2, the sums are over all pairs of α-carbons that are at least 3 residues from each other. There are Nα total α-carbons, and the Heaviside function Θ is equal to 1 when the distance between α-carbons i and j, rij, is within σ = 0.5 Å of its distance in the native state, rijN , and is zero otherwise.

The average fluctuations in χ as a function of temperature are shown in Figure 2 for the test protein in dilute solution as well as in all seven crowding agent systems. In dilute solution, the test protein has a folding temperature of 322.4 K, indicated by the dashed vertical line. As in our

ACS Paragon Plus Environment

8

Page 9 of 35

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

The Journal of Physical Chemistry

previous work,30 the folding temperature of the test protein depends on the nature of the crowding agents. Folding temperatures are given in Table 1. Table 1. Details of the test protein and the seven crowding agent systems.a crowding

ε

Tf (K)

∆∆G

∆

-

322.4 (0.1)

-

89.1 (0.1)

(ExV)15

-

340.2 (0.9)

+0.45 (0.06)

17.9 (0.4)

(Lys)15

0.000

338.5 (0.8)

+0.40 (0.07)

16.7 (0.3)

(Ser)15

0.110

312.0 (0.9)

-0.06 (0.05)

12.5 (0.4)

(His)15

0.260

-

-0.17 (0.07)

-3.0 (0.4)

(Tyr)15

0.490

-

-

-

(Met)15

0.670

-

-

-

(Leu)15

1.000

-

-

-

agent dilute solution

a

The hydrophobicity of the crowding agents ε, the folding temperature of the test protein, the change in the

free energy of unfolding due to crowding (in kcal/mol) at T = 322.4 K are given along with entropy of unfolding (in cal/mol⋅K). The standard error is in parentheses.

In order to determine uncertainty values, the data were divided into four non-overlapping sequential subgroups. Averages were then determined for each subgroup. The standard deviation for all four averages was taken to be the uncertainty.

ACS Paragon Plus Environment

9

The Journal of Physical Chemistry

Dilute Solution Excluded Volume Crowders Lysine Crowders Serine Crowders Histidine Crowders Tyrosine Crowders Methionine Crowders Leucine Crowders T = 322.4 K

0.1

0.08

0.06

∆χ

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 35

0.04

0.02

0 280

300

320

340

360

380

400

420

Temperature (K) Figure 2. Temperature dependence of the fluctuations in the overlap parameter χ. The vertical dashed line is the folding temperature of the test protein in dilute solution. Data for the tyrosine, methionine, and leucine systems are coincident. There is a clear folding transition for the excluded volume only, lysine, and serine crowding agent systems. There is a large stabilization of the test protein for the excluded volume only and lysine crowders (which interact with the test protein only through excluded volume and hydrogen bonding interactions) relative to the dilute solution. The extent of this stabilization is less with β-hairpin crowders than with helical crowders: ∆T = +17.8 K versus +39.9 K for excluded volume crowders and ∆T = +16.1 K versus +37.2 K for lysine crowders.30 The similarity in folding temperatures in the excluded volume only and lysine systems indicates that the presence of hydrogen bonding by itself has little effect on the stability of the test protein.

ACS Paragon Plus Environment

10

Page 11 of 35

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

The Journal of Physical Chemistry

Serine crowders interact with the test protein through weak hydrophobic interactions and through hydrogen bonding and result in a lower folding temperature than for the macromolecular crowding control, regardless of crowders secondary structure. With β-hairpin serine crowders, the destabilization relative to the macromolecular crowding control results in a folding temperature lower than in dilute solution with ∆T = -10.4 K. This is in contrast to helical serine crowders, where the folding temperature for the test protein is higher than in dilute solution (but still lower than the excluded volume limit).30 In both cases, the destabilization of the test protein is due to favorable side chain-side chain contacts and not due to hydrogen bonding. At T = 322.4 K, the extended states have an average of 24 side chain-side chain contacts, versus 13 for the folded state. In contrast, the average energy of test protein hydrogen bonds shows a sharp change in the transition region while that of crowding agent hydrogen bonds remains largely unchanged. The average hydrogen bond energies are shown as a function of temperature in Figure S2 in the Supporting Information.

For crowding agents that are more hydrophobic than the test protein, we do not see the formation of a packed structure as with helical crowding agents. Instead, no stable helices were observed at any temperature. For the histidine crowder system, a small number of helical configurations were observed, though the helical native state is never thermodynamically stable. These configurations were used to determine the ∆∆G and ∆S values for the histidine crowder system given in Table 1. The few helical configurations do not contribute significantly to the overall energetics of the system, as shown by the hydrogen bond energies in Figure S2 in the Supporting Information.

ACS Paragon Plus Environment

11

The Journal of Physical Chemistry

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 35

For the three systems where the test protein adopts its native state at low temperatures, we find that the effects of the β-hairpin crowding agents on the folding transition are similar in magnitude to those of helical crowding agents. The foldability σ = (Tθ − Tf ) / Tθ , where Tθ is the collapse temperature, differs only minimally relative to dilute solution for each: σ = 0.0016 in dilute solution but σ = 0.0021 (0.0030) with excluded volume crowders, 0.0009 (0.0030) with lysine crowders, and 0.0011 (0.0024) with serine crowders, where the values with helical crowders are in parentheses. The cooperativity45 Ωdecreases for all three systems with folding transitions. In dilute solution Ω = 116.5, but is 47.2 (50.0) with excluded volume crowders, 50.1 (49.3) with lysine crowders, and 27.6 (123.8) with serine crowders.

The fractional helicity and the fraction of β-strand residues, θα and θ β respectively, are shown in Figure 3. θα is the fraction of helical hydrogen bonds while θ β is fraction of residues, exclusive of the end residues, in the β-sheet region: -150º < φ < -45º and 90º < ψ < 180º. The test protein shows mostly helical content at low temperatures in dilute solution as well as in the presence of excluded volume only, lysine, and serine crowders. A small amount of helical content is observed in the presence of histidine crowders, though it is not visible on the scale of Figure 3. No helical content is observed in the presence of tyrosine, methionine, and leucine crowders. However, the fraction of residues in the β-strand configuration in the presence of the more hydrophobic crowders is significantly larger than that in the presence of less hydrophobic crowders. At high temperatures, all systems have a value of θ β > 0.4 . For more hydrophobic crowders, the amount of β-strand residues increases gradually with decreasing temperature,

ACS Paragon Plus Environment

12

Page 13 of 35

never having a distinct transition from a random coil to aβ-strand-rich configuration. For less hydrophobic crowders, however, the fraction of β-strand residues decreases to near zero at low

1 0.8

θα

0.6 0.4 0.2 0 1 0.8 0.6

θβ

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

The Journal of Physical Chemistry

0.4 0.2 0 280

300

320

340

360

380

400

420

Temperature (K) Figure 3. The fraction of residues in the helical (top) and β-strand (bottom) configuration as a function of temperature. In the top figure, the data for the tyrosine, methionine, and leucine crowder systems are coincident. Colors are the same as in Figure 2. temperatures.

The radius of gyration as a function of the temperature is shown in Figure 4. Different behavior is observed for the test protein when in dilute solution and in the presence of less hydrophobic crowders than in the presence of more hydrophobic crowders. In dilute solution and in the presence of less hydrophobic crowding agents, the radius of gyration shows a transition

ACS Paragon Plus Environment

13

The Journal of Physical Chemistry

13

12

Radius of Gyration (Å)

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 35

11

10

9

8

7 280

300

320

340

360

380

400

420

Temperature (K) Figure 4. Temperature dependence of the radius of gyration of the test protein. The colors are the same as in Figure 2. from a compact state at low temperatures to an extended state at high temperatures. In the excluded volume only and lysine crowder systems, the test protein is on average smaller than in dilute solution at all temperatures, consistent with the predictions of macromolecular crowding.3 The presence of weak attractive hydrophobic interactions between the test protein and the serine crowders causes the unfolded, high-temperature states to have a larger radius of gyration than in dilute solution. In the presence of more hydrophobic crowders, the test protein adopts configurations with a large fraction of β-strand content that results in a radius of gyration that is larger than the unfolded states in dilute solution at all temperatures. The radius of gyration decreases with increasing temperature, paralleling the decrease in β-strand content.

ACS Paragon Plus Environment

14

Page 15 of 35

0.2

0.15

Distribution

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

The Journal of Physical Chemistry

0.1

0.05

0 6

8

10

12

14

Radius of Gyration (Å) Figure 5. Distribution of the radius of gyration at T = 322.4 K. Data are spaced every 0.1 Å and the sum of all data equals 1. For clarity, a smooth curve has been drawn through the data points. Colors are the same as in Figure 2.

The distribution of the radius of gyration at T = 322.4 K is shown in Figure 5. Figure 5 clearly shows the difference in the distributions for the unfolded states. Those for dilute solution show a maximum near 11 Å. As the strength of the test-protein-crowder hydrophobic attraction increases, the peak in the distribution of extended states increases from 11.4 Å for serine crowders to 12.4 for histidine crowders, to 12.7 for tyrosine, methionine, and leucine crowders.

ACS Paragon Plus Environment

15

The Journal of Physical Chemistry

These results are echoed in the free energy surfaces, plotted in Figure 6, for the eight different

180

180

180

180

90

90

90

90

2.0

1.6 0



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 35

-90 5

0

10

15

0

-90 5

Dilute Solution

10

15

0

-90 5

10

Excluded Volume

15

-90 5

Lysine

10

15

1.2

S erine

180

180

180

180

90

90

90

90

0

0

0

0

0.8

0.4

-90 5

10

15

-90 5

Histidine

10

15

Tyrosine

-90 5

10

Methionine

15

-90 5

10

15

0.0

Leucine

Radius of Gyration

Figure 6. The free energy surface (in kcal/mol) at T = 322.4 K of the test protein in dilute solution and that of the seven different crowder systems plotted as a function of the radius of gyration (in Å) and

(in degrees). Values are relative to the global minimum. The

minimum corresponding to a left-handed helix is indicated with an arrow on the excluded volume surface. systems as a function of the radius of gyration, Rg, and the parameter Qψ . In these plots, the minimum free energy at T = 322.4 K is set to zero for each surface. Consequently, when the folded state has the lowest free energy, the free energy represented is the change in free energy of unfolding, ∆∆G. The parameter Qψ has been adapted from Soto et al.46 and is an average of the ψ backbone dihedral angle,

ACS Paragon Plus Environment

16

Page 17 of 35

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

The Journal of Physical Chemistry

1 Qψ = N res − 4

N res −3

∑ψ

j

(3)

j=3

where Nres = 15 is the number of residues on the test protein. The two residues closest to the ends are excluded from the average. Qψ gives a measure of the secondary structure content of the test protein: for α-helices Qψ < −20 ; for random coils, −20 < Qψ < 60 ; and for β-strands, Qψ > 80 .

In dilute solution, there are three minima: a minimum corresponding to the (frayed) native state (Rg ≈ 7 and Qψ ≈ −40 ), a minimum corresponding to the unfolded extended states (Rg ≈ 12 and Qψ ≈ +110 ), and a minimum of ∆G = 0.03 kcal/mol at (Rg ≈ 7 and Qψ ≈ +70 ). As discussed below, this third minimum corresponds to a left-handed α-helix. In the excluded volume only and lysine crowding agent systems, the global minimum corresponds to the native state, which is more helical than in dilute solution. The ∆G of unfolding is 0.45 kcal/mol and 0.40 kcal/mol, respectively, with an uncertainty of 0.05 kcal/mol for both. The left-handed α-helix is clearly visible and is higher in free energy than the native state by 0.49 and 1.17 kcal/mol, respectively. Serine crowding agents destabilize the test protein relative to dilute solution. At T = 322.4 K, the test protein is above its folding temperature. The ∆G of unfolding is -0.06 kcal/mol and there is a barrier of 0.5 kcal/mol between the folded and unfolded states. The left-handed α-helices are 0.92 kcal/mol higher in free energy than the unfolded states. The minimum free energy path connecting the native state and the extended states for each surface is given in Figure S3 in the Supporting Information.

ACS Paragon Plus Environment

17

The Journal of Physical Chemistry

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 35

For the four most hydrophobic systems, the basin of unfolded states becomes narrower in the radius of gyration and larger in Qψ as the hydrophobicity increases. This basin corresponds to extended states of the test protein. This is true even for histidine crowders, whose average hydrophobicity is only slightly higher than the test protein (ε = 0.26 for histidine versus an average of ε = 0.234 for the test protein). In the three most hydrophobic systems, neither left- nor right-handed α-helices are observed. Though never thermodynamically stable, a small number of helical configurations were observed for the test protein in the presence of histidine crowders. Interestingly, the entropy of unfolding for these configurations is negative at T = 322.4 K. The unfolded state has a lower entropy than the folded state, though it must be emphasized that this includes the entropy of both the test protein and the crowding agents. This decrease in entropy is due to the hydrogen bonding of crowding agents to the test protein in the unfolded state and the loss of their independent translational and rotational motion. The consequent decrease in the entropy of the crowders is larger than the increase in the entropy of the test protein due to unfolding. In the folded state, no crowders are hydrogen bonded to the test protein.

Discussion Hydrogen bonds between the test protein and the crowding agents are only present when there are stabilizing side chain-side chain interactions. Figure 7 shows the total number of hydrogen bonds between the test protein and itself and between the test protein and the crowding agents. A hydrogen bond is counted if the C-H-O and the N-H-O bond angles are both less than 35° (where the locations of the H and O are determined by the local geometry)38 and if the C-N distance is less than 5 Å. For crowding agents with a lower hydrophobicity, and so a weaker attraction to the test protein, hydrogen bonding plays only a small role in affecting the stability of the test protein.

ACS Paragon Plus Environment

18

Page 19 of 35

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

The Journal of Physical Chemistry

In the presence of these crowders, almost all the hydrogen bonds are between the test protein and itself. For crowding agents with a higher hydrophobicity, the formation of a packed structure that increases the stability of the native helix is not observed. As noted elsewhere,30 this structure is unique to a helical test protein with helical crowding agents. Instead, more hydrophobic crowding agents destabilize the test protein to such an extent that the helical native state is never thermodynamically stable over the temperature range of interest. Almost all hydrogen bonds are

ACS Paragon Plus Environment

19

The Journal of Physical Chemistry

2.5

12

Number of β -segments

Number of H-Bonds

14 10 8 6 4 2 0

Number of 3- and 5-residue β-segments

4

3.5 3 280

2 1.5 1 0.5 0 1.4

4.5 Length of β -segment

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 20 of 35

300

320

340

360

380

400

420

1.2 1 0.8 0.6 0.4 0.2 0 280

300

320

340

360

380

400

420

Temperature (K) Figure 7. Top row: (left) The average number of test protein-test protein (solid lines) and test protein-crowder (dashed lines) hydrogen bonds and (right) the average number of β-segments (right). Bottom row: (left) The average length of a β-segments, in units of residues, and (right) the average number of 3-residue β-segments (solid lines) and 5-residue β-segments (dashed lines), all as a function of the temperature. The plot of the average length does not show data for the excluded volume and lysine systems. Colors are the same as in Figure 2.

formed between the test protein and crowding agents. Even at high temperatures, there are between 2 and 6 intermolecular hydrogen bonds.

It will be helpful to discuss the hydrogen bonding between the test protein and crowding agents in terms of β-segments. When residues i and i + 2 on the test protein are hydrogen bonded to the exposed hydrogen bond donors on the same crowding agent, it adopts a complementary β-

ACS Paragon Plus Environment

20

Page 21 of 35

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

The Journal of Physical Chemistry

strand configuration, which we refer to as a 3-residue β-segment. If residues i, i + 2, and i + 4 are all hydrogen bonded to the same crowder, a 5-residue β-segment is formed. Given the length of the crowding agents and their fixed β-hairpin configuration, only 3- and 5-residue β-segments can be formed. Representative configurations are shown in Figure S4 in the Supporting Information for the test protein in the presence of leucine crowders at T = 322.4 K.

Figure 7 shows the average number of β-segments as well as length of each β-segment. Note that the average length is only calculated for configurations with at least one β-segment and consequently has a minimum value of 3. In the presence of less hydrophobic crowders, the test protein forms almost no β-segments over the entire temperature range. In the presence of more hydrophobic crowders, the test protein forms few β-segments at high temperatures, less that 0.5 on average, but forms between 1.8 and 2.2 such segments at low temperatures. The average length of each segment also increases with decreasing temperature. This is due to the number of 5-residue segments increasing more rapidly than the number of 3- residue segments with decreasing temperature, as shown in the lower right of Figure 7. At low temperatures, for the histidine, tyrosine, and methionine systems there are nearly twice as many 5-residue β-segments as 3-residue β-segments. In the presence of leucine crowders, the test protein shows fewer and shorter β-segments at low temperatures compared with the other hydrophobic crowders. This is consistent with the behavior of the radius of gyration shown in Figure 4. Together, Figure 7 shows that when there are favorable test protein-crowding agent hydrophobic interactions, the test protein adopts extended configurations that allow it to increase its β-content through hydrogen bonding with the crowding agents. At the folding temperature in dilute solution, in the

ACS Paragon Plus Environment

21

The Journal of Physical Chemistry

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 35

presence of more hydrophobic β-hairpin crowders, the test protein has on average two βsegments that span four residues.

Our results show that increasing crowder hydrophobicity results in an increase in the β-strand content of the normally helical test protein. The formation of β-strands or β-sheets from otherwise non-β-rich proteins is an area of great interest to understanding protein-folding diseases, such as the formation of β-sheet-rich fibrils by the Amyloid-β protein (Aβ), implicit in Alzheimer’s Disease. The structural differences between our β-hairpin crowders and the ends of an amyloid fibril are clear: our hydrogen bonds are perpendicular to what would be the fibril axis while the true fibril end has hydrogen bonds parallel to it. In addition, for the amyloid fiber, the hairpin structure is due to side chain interactions, while for our crowders, the hairpin configuration is due to hydrogen bonding and is artificially fixed. Nevertheless, they share the common element of a fixed structure with exposed hydrogen bond donors that causes a non-βrich protein to hydrogen bond with the fixed fibril (or crowding agent) and increase its β content. It is generally agreed that Aβ fibril elongation occurs through a lock-dock mechanism.47-49 Experimental results show that an increase in hydrophobicity increases the amyloidogenic propensity through increases in aggregation rates47,50,51 or through a decrease in the free energy of fibrilization.51 This trend is not universal, however, with mutations that increase the hydrophobicity increasing the free energy of fibrilization.52 It has been suggested53 that strong side chain-side chain interactions may play a role in impeding the addition of a protein onto a fibril, though it is unclear if this is solely a kinetic effect or if they alter the free energy of fibrilization. Though the details of our system differ from fibril elongation, our results support the overall conclusion that an increase in the hydrophobicity of the preexisting β-sheet increases

ACS Paragon Plus Environment

22

Page 23 of 35

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

The Journal of Physical Chemistry

the propensity for the otherwise non-β-rich proteins to adopt a configuration high in β-content. This is true for both the dock, through an increased number of bonded crowders, and lock, through longer β-segments, parts of the elongation mechanism. Importantly, we see an increase in β-content in systems where the sequence of the seed fibril (our crowding agents) is unrelated to that of the protein of interest.

The hallmark of macromolecular crowding is the stabilization of the compact native state by destabilizing the unfolded, extended states. The presence of the crowding agents decreases the conformational entropy of the test protein, which affects the extended states more strongly than the compact states. Interestingly, as is seen in the free energy landscapes in Figure 6, the lefthanded α-helix is also stabilized by the presence of excluded volume only crowding agents. In this model, the left-handed helix is 21.5 kcal/mol higher in energy than the native state, or 3.1 kcal/mol per three residues. This falls in the lower end of the of 3-4 kcal/mol range predicted for the energy of a left-handed helix for an alanine tripeptide.54 However, the coarse-grained protein model does not accurately capture the steric interaction between the side chain and the backbone carbonyl group, so the stability of the left-handed helix seen here is overestimated. The potential energy surface for the test protein in dilute solution is given in Figure S5 in the Supporting Information. The left-handed helix is never thermodynamically stable, having a free energy difference of +0.03 kcal/mol at T = 322.4 K in dilute solution and of +0.49 kcal/mol in the presence of excluded volume crowders.

Changing the shape of the crowding agent from helical to β-hairpin reduces by half the magnitude of the thermal stabilization. The helical and β-hairpin crowding agents have similar

ACS Paragon Plus Environment

23

The Journal of Physical Chemistry

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 35

excluded volumes,55 4549 Å3 for the helices compared to 4967 Å3 for the β-hairpins, a difference of 9% (see Figure 1b). The difference in stabilization is due instead to the shape of crowding agents and the resulting void space distribution, which is known to affect the stability of a test protein.56,57 Christiansen et al.58 found that a dumbbell shaped crowders exerted a larger stabilizing influence than spheres of equal volume. Here, we find that the shape of the crowding agents also exerts an influence, but that the structures of the crowding agents, which are based on the secondary structures of proteins, make it difficult to predict a priori which will result in larger void spaces. We speculate that in our case, the more irregularly-shaped β-hairpin crowders result in larger void spaces than the cylindrically symmetric helices, but that the primary impact is on the entropy of the unfolded states. Larger void spaces allow for a larger number of configurations available to the unfolded protein and therefore a higher entropy. This is borne out by the entropies of unfolding: 16.0 cal/mol·K for helical excluded volume crowders, vs. 17.9 cal/mol·K for β-hairpin excluded volume crowders.

Conclusion The results of these simulations build upon those of our previous work.30 Together, they show that crowding agents affect the stability of a test protein through multiple mechanisms and that these mechanisms depend on both the secondary structure and the chemical nature of the crowding agent. The shape of the crowding agents plays a pivotal role in determining the amount of stabilization due to the entropy-driven excluded volume effect. Cylindrical α-helical excluded volume only crowders stabilize the test protein more than β-hairpin crowders. Relative to this stabilization, the introduction of weak attractive interactions destabilizes the test protein by a remarkably consistent amount irrespective of the shape of the crowder. Lysine crowders, which

ACS Paragon Plus Environment

24

Page 25 of 35

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

The Journal of Physical Chemistry

have no hydrophobic attractions but can form hydrogen bonds, destabilize the test protein by 1.7 K. Serine crowders, which have weak hydrophobic interactions and hydrogen bonding, destabilize the test protein through favorable side chain-side chain interactions that preferentially stabilize the unfolded states. The extent of destabilization relative to the excluded volume system is approximately 30 K for both crowder secondary structures. The difference in the excluded volume stabilization due to crowder shape results in the test protein being stabilized by helical serine crowders relative to dilute solution, but destabilized by β-hairpin serine crowders.

For more hydrophobic crowders, we do not see the formation of a stabilizing packed structure such as that formed with helical crowder. This is not unexpected, as it was noted that this structure was a result of the helical nature of both the native state of the test protein and the crowding agents and was therefore not expected to be a general phenomenon.

We find that hydrogen bonding between the test protein and crowding agents has an all-ornothing character about it. When the hydrophobicity of the crowding agent is less than that of the test protein, intermolecular hydrogen bonding has little effect. When the hydrophobicity of the crowding agent is more, however, the helical native state is destabilized to the point where it is virtually non-existent. Instead, the test protein hydrogen bonds with the crowding agents and adopts configurations that form β-sheets with the crowding agents. This suggests that mutations that increase the hydrophobicity of a pre-existing β-rich structure can induce a test protein to misfold and adopt a β–strand conformation. This is in agreement with experimental findings for Aβ and other amyloidogenic peptides.47,50,51 These findings also suggest that the same crowding agents may stabilize some proteins but destabilize others.

ACS Paragon Plus Environment

25

The Journal of Physical Chemistry

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 35

Finally, we anticipate that these results, together with those with helical crowding agents, shed light on the mechanisms through which biomacromolecules affect the stability of helical motifs in a crowded environment. Most globular proteins have a mix of secondary structures and their effect on helix stability will be through a combination of favorable side chain-side chain interactions and hydrogen bonding. Future work will look at crowding agents with mixed secondary structure and at heterogeneous systems with multiple types of crowding agents.

Acknowledgements A.E.v.G. gratefully acknowledges support from a Cottrell College Science Award from the Research Corporation.

Supporting Information The average end-to-end distance for each system is given in Figure S1. Figure S2 shows the hydrogen bonding energies for the histidine crowder system. The minimum free energy path across each free energy surface is given in Figure S3. Figure S4 gives representative configurations of the test protein / lecuine crowding agents system. Figure S5 is the potential energy surface of the test protein in dilute solution.

ACS Paragon Plus Environment

26

Page 27 of 35

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

The Journal of Physical Chemistry

REFERENCES (1)

Cheung, M. S. Where soft matter meets living matter – protein structure, stability, and folding in the cell. Curr. Opin. Struct. Biol. 2013, 23, 212-217.

(2)

Monteist, W. B.; Cohen, R. D.; Smith, A. E.; Guzman-Cisneros, E.; Pielak, G. J. Quinary structure modulates protein stability in cells. Proc. Natl. Acad. Sci. USA 2015, 112, 17391742.

(3)

Minton, A. P. Models for excluded volume interaction between an unfolded protein and rigid macromolecular cosolutes: Macromolecular crowding and protein stability revisited. Biophys. J. 2005, 88, 971-985.

(4)

Cheung M. S.; Klimov, D.; Thirumalai, D. Molecular crowding enhances native state stability and refolding rates of globular proteins. Proc. Natl. Acad. Sci. USA 2005, 102, 4753-4758.

(5)

Sasahara K.; McPhie, P.; Minton, A. P. Effect of dextran on protein stability and conformation attributed to macromolecular crowding. J. Mol. Biol. 2003, 326, 12271237.

(6)

Dhar A.; Girdhar, K.; Singh, D.; Gelman, D. H.; Ebbinghaus, E.; Gruebele, M. Protein stability and folding kinetics in the nucleus and endoplasmic reticulum of eukaryotic cells. Biophys. J. 2011, 101, 421-430.

(7)

Miklos A. C.; Sarkar, M.; Wang, Y.; Pielak, G. J. Protein crowding tunes protein stability. J. Am. Chem. Soc. 2011, 133, 7116-7120.

ACS Paragon Plus Environment

27

The Journal of Physical Chemistry

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

(8)

Page 28 of 35

Rosen, J.; Kim, Y. C.; Mittal, J. Modest protein-crowder attractive interactions can counteract enhancement of protein association by intermolecular excluded volume interactions. J. Phys. Chem. B 2011, 115, 2683-2689.

(9)

Benton L. A.; Smith, A. E.; Young, G B.; Pielak, G. J. Unexpected effects of macromolecular crowding on protein stability. Biochemistry 2012, 51, 9773-9775.

(10)

Feig M.; Sugita, Y. Variable interactions between protein crowders and biomolecular solutes are important in understanding cellular crowding. J. Phys. Chem. B 2012, 116, 599-605.

(11)

Wang Y.; Sarkar, M.; Smith, A. E.; Krois, A. S.; Pielak, G. J. Macromolecular crowding and protein stability. J. Am. Chem. Soc. 2012, 134, 16614-16618.

(12)

Harada R.; Tochio, N.; Kigawa, T.; Sugita, Y.; and Fieg, M. Reduced native state stability in crowded cellular environment due to protein-protein interactions. J. Am. Chem. Soc. 2013, 135, 3696-3701.

(13)

McGuffee S. R.; Elcock, A. H. Diffusion, crowding & protein stability in a dynamic molecular model of the bacterial cytoplasm. PLoS Comput. Biol. 2010, 6, e10000694.

(14)

Cossins, B. P.; Jacobson, M. P.; Guallar, V. A new view of the bacterial cystol environment. PLoS Comput. Biol. 2011, 7, e1002066.

(15)

Kurniawan N. A.; Enemark, S.; Rajagopalan, R. Crowding alters the folding kinetics of a

β-hairpin by modulating the stability of intermediates. J. Am. Chem. Soc. 2012, 134, 10200-10208.

ACS Paragon Plus Environment

28

Page 29 of 35

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

The Journal of Physical Chemistry

(16)

Predeus, A. V.; Gul, S.; Gopal, S. M.; Feig, M. Conformational sampling of peptides in the presence of protein crowders from AA/CG-multiscale simulations. J. Phys. Chem. B 2012, 116, 8610-8620.

(17)

Qin S.; Zhou, H.-X. FFT-based method for modeling protein folding and binding under crowding: Benchmarking on ellipsoidal and all-atom crowders. J. Chem. Theory Comput. 2013, 9, 4633-4643.

(18)

Yap, E.-H,; Head-Gordon, T. Calculating the bimolecular rate of protein-protein association with interacting crowders. J. Chem. Theory Comput. 2013, 9, 2481-2489.

(19)

Qin S.; Zhou, H.-X. Further development of the FFT-based method for atomistic modeling of protein folding and binding under crowding: Optimization of accuracy and speed. J. Chem. Theory Comput. 2014, 10, 2824-2835.

(20)

Ghaemmaghami S.; Oas, T. G. Quantitative protein stability measurement in vivo. Nat. Struct. Biol. 2000, 4, 879-822.

(21)

Ignatova, Z.; Krishnan, B.; Bombardier, J. P.; Marcelino, A. M. C.; Hong, J.; Gierasch, L. M. From the test tube to the cell: Exploring the folding and aggregation of a β-clam protein. Peptide Sci. 2007, 88, 157-163.

(22)

Inomata, K.; Ohno, A.; Tochio, H.; Isogai, S.; Tenno, T.; Nakase, I.; Takeuchi, T.; Futaki, S.; Ito, Y.; Hiroaki, H.; Shirakawa, M. High-resolution multi-dimensinal NMR spectroscopy of proteins in human cells. Nature 2009, 458, 106-109.

(23)

Ebbinghaus, S.; Dhar, A.; McDonald, J. D.; Gruebele, M. Protein folding stability and dynamics imaged in a living cell. Nat. Methods 2010, 7, 319-323.

ACS Paragon Plus Environment

29

The Journal of Physical Chemistry

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

(24)

Page 30 of 35

Crowley, P. B.; Chow, E.; Papkovskaia, T. Protein interactions in the Escherichia coli cystol: An impediment to in-cell NMR spectroscopy. ChemBioChem. 2011, 12, 10431048.

(25)

Schlesinger, A. P.; Wang, Y.; Tadeo, X.; Millet, O.; Pielak, G. J. Macromolecular crowding fails to fold a globular protein in cells. J Am. Chem. Soc. 2011, 133, 8082-8085.

(26)

Kim, Y. C.; Mittal, J. Crowding induced entropy-enthalpy compensation in protein association equilibria. Phys. Rev. Lett. 2013, 110, 208102.

(27)

Betancourt, M. R.; Thirumalai, D. Exploring the kinetic requirements for enhancement of protein folding rates in the GroEL cavity. J. Mol. Biol. 1999, 287, 627-644.

(28)

Jewett, A. I.; Baumketmer, A.; Shea, J.-E. Accelerated folding in the weak hydrophobic environment of a chaperonin cavity: Creation of an alternate fast folding pathway. Proc. Natl. Acad. Sci. USA 2004, 101, 13192-13197.

(29)

Sirur, A.; Knott, M.; Best, R.B. Effect of interactions with the chaperonin cavity on protein folding and misfolding. Phys. Chem. Chem. Phys. 2014, 16, 6358-6366.

(30)

Macdonald, B.; McCarley, S.; Noeen, S.; van Giessen, A. E. Protein-protein interactions affect alpha helix stability in crowded environments. J. Phys. Chem. B. 2015, 119, 29562967.

(31)

Scholtz, J. M.; York, E. J.; Stewart, J. M.; Baldwin, R. L. A neutral, water-soluble, αhelical peptide: The effect of ionic strength on the helix coil equilibrium. J. Am. Chem. Soc. 1991, 113, 5102-5104.

ACS Paragon Plus Environment

30

Page 31 of 35

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

The Journal of Physical Chemistry

(32)

Shalongo, W.; Dugad, L.; Stellwagen, E. Distribution of helicity within the model peptide Acetyl(AAQAA)3amide. J. Am. Chem. Soc. 1994, 116, 8288-8293.

(33)

Ferrara, P.; Apostolakis, J.; Calflisch, A. Thermodynamics and kinetics of folding of two model peptides investigated by molecular dynamics simulations. J. Phys. Chem. B 2000, 104, 5000-5010.

(34)

Chebaro, Y.; Dong, X.; Laghaei, R.; Derreumaux, P.; Mousseau, N. Replica exchange molecular dynamics simulations of coarse-grained proteins in implicit solvent. J. Phys. Chem. B. 2009, 113, 267-272.

(35)

Bereau, T.; Deserno, M.; Bachmann, M. Structural basis of folding cooperativity in model proteins: Insights from a microcanonical perspective. Biophys. J. 2011, 100, 27642772.

(36)

Straub, J. E.; Thirumalai, D. Toward a molecular theory of early and late events in monomer amyloid fibril formation. Annu. Rev. Phys. Chem. 2011, 62, 437-463.

(37)

Christopiet, T.; Hortschansky, P.; Schroeckh, V.; Ghürs, K.; Zandomeneghi, G.; Fändrich, M. Mutagenic analysis of the nucleation propensity of oxidized Alzheimer’s βamyloid peptide. Protein Sci. 2005, 14, 2125-2131.

(38)

Bereau T.; Deserno, M.. Generic coarse-grained model for protein folding and aggregation. J. Chem. Phys. 2009 130, 235106.

(39)

Berman, H. M.; Westbrook, J.; Feng, Z.; Gilliland, G.; Bhat, T. N.; Weissig, H.; Shindylov, N.; Bourne, P.E. The protein data bank. Nucleic Acid Res. 2000, 28, 235-242.

ACS Paragon Plus Environment

31

The Journal of Physical Chemistry

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

(40)

Page 32 of 35

Humphrey, W.; Dalke, A.; Schulten, K. VMD - Visual Molecular Dynamics. J. Molec. Graphics. 1996, 4, 33-38.

(41)

Kim, J.; Straub, J. E.; Keyes, T. Replica exchange statistical temperature molecular dynamics algorithm. J. Phys. Chem. B 2012, 116, 8646-8653.

(42)

Church, M. S.; Ferry, C. S.; van Giessen, A. E. Thermodynamics of peptide dimer formation. J. Chem. Phys. 2012, 136, 245102.

(43)

Wang F.; Landau, D. P. Determining the density of states for classical statistical models: A random walk algorithm to produce a flat histogram. Phys. Rev. E 2001, 64, 056101.

(44)

Kim J.; Keyes, T.; Straub, J. E. Iteration-free, weighted histogram analysis method in terms of intensive variables. J. Chem. Phys. 2011, 135, 061103.

(45)

Klimov, D.K.; Thirumalai, D. Cooperativity in protein folding: From lattice models with sidechains to real proteins. Fold. Des. 1998, 1, 127-139.

(46)

Soto, P.; Baumketner, A.; Shea, J.-E. Aggregation of polyalanine in a hydrophobic environment. J. Chem. Phys. 2006, 124, 134904.

(47)

DuBay, K. F.; Pawar, A. P.; Chiti, F.; Zurdo, J.; Dobson, C. M.; Vendruscolo, M. Prediction of the absolute aggregation rates of amyloidogenic polypeptide chains. J. Mol. Biol. 2004, 314, 1317-1326.

(48)

Nguyen, P. H.; Li, M. S.; Stock, G.; Straub, J. E.; Thirumalai, D. Monomer adds to preformed structured oligomers of Aβ-peptides by a two-stage dock-lock mechanism. Proc. Natl. Acad. Sci. USA 2007, 104, 111-116.

ACS Paragon Plus Environment

32

Page 33 of 35

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

The Journal of Physical Chemistry

(49)

Gurry, T.; Stultz, C. M.; Mechanism of amyloid-β fibril elongation. Biochemistry 2014, 53, 6981-6991.

(50)

Esler, W. P.; Stimson, E. R.; Ghilardi, J. R.; Lu, Y.-A.; Felix, A. M.; Vinters, H. V.; Mantyh, P. W.; Lee, J. P.; Maggio, J. E. Point substitution in the central hydrophobic cluster of human β-amyloid congener disrupts peptide folding and abolishes plaque competence. Biochemistry 1996, 35, 13914-13921.

(51)

Meinhardt, J.; Tartaglia, G. G.; Pawar, A.; Christopeit, T.; Hortschansky, P.; Schroeckh, V.; Dobson, C. M.; Vendruscolo, M.; Fändrich, M. Similarities in the thermodynamics and kinetics of aggregation of disease-related Aβ(1-40) peptides. Protein Sci. 2007, 16, 1214-1222.

(52)

Peim, A.; Hortschansky, P.; Christopeit, T.; Schroeckh, V.; Richter, W.; Fändrich, M. Mutagenic exploration of the cross-seeding and fibrillation propensity of Alzheimer’s βamyloid peptide variants. Protein Sci. 2006, 15, 1801-1805.

(53)

Takeda, T.; Klimov, D. K. Probing energetics of Aβ fibril elongation by Molecular Dynamics simulations. Biophysical J. 2009, 96, 4428-4437.

(54)

Tsai, M.; Xu, Y.; Dannenberg, J.J. Ramachandran revisited. DFT energy surfaces of diastereometric trialanine peptides in the gas phase and aqueous solution. J. Phys. Chem. B 2009, 113, 309-318.

(55)

Connolly, M. L. Computation of molecular volume. J. Am. Chem. Soc. 1985, 107, 11181124.

ACS Paragon Plus Environment

33

The Journal of Physical Chemistry

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

(56)

Page 34 of 35

Minton, A. P. Excluded volume as a determinant of macromolecular structure and reactivity. Biopolymers, 1981, 20, 2093-2120.

(57)

Homouz, D.; Stagg, L.; Wittung-Stafshede, P.; Cheung, M. S. Molecular crowding modulates folding mechanism of α/β protein apoflavodoxin. Biophys. J. 2009, 96, 671680.

(58)

Christiansen, A.; Wang, Q.; Samiotakis, A.; Cheung, M. S.; Wittung-Stafshede, P. Factors defining effects of macromolecular crowding on protein stability: An in vitro/in slico case study using Cytochrome c. Biochemistry 2010, 49, 6519-6530.

ACS Paragon Plus Environment

34

Page 35 of 35

TOC Image:

1 0.8

θα

0.6 0.4 0.2 0 1 0.8 0.6

θβ

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

The Journal of Physical Chemistry

0.4 0.2 0 280

300

320

340

360

380

400

420

Temperature (K)

ACS Paragon Plus Environment

35