N-Terminus Binding Preference for Either Tanshinone or Analogue in

Apr 13, 2017 - ... of pH on Aβ 42 Monomer and Fibril-like Oligomers-Decoding in Silico of the Roles of p K Values of Charged Residues. Wei Zhao , Hon...
0 downloads 0 Views 2MB Size
Subscriber access provided by Binghamton University | Libraries

Article

N-terminus Binding Preference for either Tanshinone or Analogue in both Inhibition of Amyloid Aggregation and Disaggregation of Preformed Amyloid Fibrils---Toward Introducing A kind of Novel Anti-Alzheimer Compounds Mingyan Dong, Wei Zhao, Dingkun Hu, Hongqi Ai, and Baotao Kang ACS Chem. Neurosci., Just Accepted Manuscript • DOI: 10.1021/acschemneuro.7b00080 • Publication Date (Web): 13 Apr 2017 Downloaded from http://pubs.acs.org on April 14, 2017

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.

ACS Chemical Neuroscience 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 33

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

ACS Chemical Neuroscience

N-terminus Binding Preference for either Tanshinone or Analogue in both Inhibition of Amyloid Aggregation and Disaggregation of Preformed Amyloid Fibrils---Toward Introducing A kind of Novel Anti-Alzheimer Compounds Mingyan Dong†, Wei Zhao†, Dingkun Hu†, Hongqi Ai*,Baotao Kang* Shandong Provincial Key Laboratory of Fluorine Chemistry and Chemical Materials, School of Chemistry and Chemical Engineering, University of Jinan, Jinan, 250022, China Corresponding author: Hongqi Ai E-mail: [email protected] Corresponding author: Baotao Kang E-mail: [email protected]

These authors equally contributed to this work

1 Environment ACS Paragon Plus

ACS Chemical Neuroscience

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 33

ABSTRACT

Amyloid-β (Aβ40/Aβ42) peptide with length of 40 or 42 residues is naturally secreted as cleavage product of the amyloid precursor protein, and formation of Aβ aggregates in patient’s brain is a hallmark of Alzheimer’s disease (AD). Therefore disaggregation and disruption provide potential therapeutic approaches to reduce, inhibit, and even reverse Aβ aggregation. The disaggregation/inhibition effect of the inhibitors applies generally to both Aβ40 and Aβ42 aggregations. Here we capture the atomic-level details of the interaction between Aβ40/Aβ42 and either natural tanshinone compound TS1 or its derivative TS0, and observe novel results by using molecular dynamics simulations. We observe that the natural TS1 indeed inhibits the monomolecular Aβ42 (mAβ42) aggregation and disaggregates Aβ42 amyloid fibrils, being in good agreement with the experimental results. TS1 is favorable to stabilize mAβ40 and even Aβ40 fibril, playing an opposite role to that in the Aβ42 counterpart, however. TS0 can inhibit the misfolding of either mAβ40 or mAβ42 and disaggregate Aβ42 fibril but stabilize the Aβ40 fibril. Using a combination of secondary structural analysis, MM-PBSA binding energy calculations, and radial distribution functions computations, we find that both TS0 and TS1, especially the former, prefer to bind at the charged residues within disordered N-terminus with a scarce positive binding energy and disappear the characteristic C-terminal bend region of Aβ42 fibril, as well as twist the Aβ42 fibril seriously. It turns out to destabilize the Aβ42 fibril and enable the conversion of U-shaped Aβ42 fibril from the one-fold to the two-fold morphologies. The N-terminal binding preference helps us to identify N-terminal region as the specific epitope for specific inhibitors/drugs (such as TS0 and analogues), heralding unusual inhibition/disaggregation or stabilization mechanisms, and offering an alternative direction in engineering new inhibitors to treat AD.

2 Environment ACS Paragon Plus

Page 3 of 33

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

ACS Chemical Neuroscience

KEYWORDS Aβ, Tanshinone, Alzheimer’s disease, disaggregation/inhibition mechanism, preferential binding at Nterminus region

3 Environment ACS Paragon Plus

ACS Chemical Neuroscience

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 33

INTRODUCTION As the likely cause of Alzheimer's disease, the aggregation of amyloid β protein (Aβ) is widely studied via experimental and computational efforts1,2. Aβ40 and Aβ42 peptides are the major species in Aβ production with Aβ42 not only the predominant content in neuritic plaques of AD patients, but also a higher propensity in vitro to aggregate and form amyloid fibrils3, 4. Although the fibril structure of Aβ aggregated end product depends strongly on the environment and consequently has diverse polymorphic features, the U-shaped (β1 strand-turn-β2 strand motif/β hairpin) β-sheets’ protofilaments2, 5, 6 and Sshaped ones 7-9 remain two kinds of dominant themes however. For example, the theme of Aβ40 fibril is U-shaped although it has two kinds of morphologies, whereas that of Aβ42 fibril characterizes either Ushaped or S-shaped ones. Zhang et al's study10 revealed that the center region β1 of Aβ42 is first to aggregate, followed by the C (second) and N termini (last), consistent with that the hydrophobic C-terminus serves as an interface for Aβ42 aggregation and plays a smaller but important role in aggregation 11 and that the hydrophilic Nterminus is unstructured and flexible12. Different from S-shaped Aβ42 fibril7-9, NMR-determined Ushaped fibril is one-fold one (Architecture1, abbreviated as Ar1) characterizing a specific side chain packing arrangement-Phe19 packing against Gly38 within the β1-turn-β2, and a longer disordered N terminal region (DR, residues 1-16)12. For the two-fold Aβ42 symmetry fibril, it is composed of two βturn-β units and a characteristic intermolecular contact of Met35 with Gly3713(Ar2), where the specific side chain packing arrangement becomes Phe19 against Leu34 with a short N terminus (residues 110/13)14. In another two-fold and even possible three-fold fibrils14 with higher stability15, the C-terminal β-strand bends at Gly37-Gly38 to allow Ala42 to contact the side chain of Met35. Aβ40 also prefers to generate two or three-fold fibril morphology5, with Phe19 contacting with Leu34 and unstructured tenresidue length N-terminus. Therefore Aβ42 fibril in Ar2 is similar to two-fold Aβ40 fibril (Ar3)16. Aβ42 is

4 Environment ACS Paragon Plus

Page 5 of 33

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

ACS Chemical Neuroscience

characterized by such a bend structure centered at G37-G3817(sometimes including V3918) and a rigid C terminus,19 which may contribute to Aβ42 more aggregation and toxicity than Aβ40. Baram et al's study20 reproduced the existence of the bend structure (residues 37GG38) in newly produced Aβ42 fibril on the basis of Tycko’s Aβ40 Ar321, indicating the bend is the sine qua non of Aβ4217. In addition, the two additional residues in Aβ42, absent in Aβ40, had been verified to stabilize the neurotoxic, low-order oligomers in a non-β-sheet secondary structure13, 14, 16 and play an important role in amyloidogenesis and toxicity by combining to the rather flexible N-terminal residues22. The disordered residues of the Nterminus in Ar2 end at most at residue 1214, akin to that of Aβ40 Ar3 fibril21,23 but of Ar1. The most update study24 revealed that the non-contacting charged N-terminal residues help the transfer of entropy to the surrounding solvation shell and stabilize β-hairpin. Therefore intervention at different regions of Aβ peptide would produce different effects on the morphology, stability and neurotoxicity of Aβ aggregates. Besides immunotherapy, a substantial number of intervention efforts focus on developing various inhibitors to prevent the Aβ self-aggregation or disaggregate the preformed Aβ fibril to date25 since neurotoxicity is mainly associated with the formation of Aβ aggregates with β-sheet-rich structure. These common inhibitors can be divided into three classes, including (1) peptide(i.e. LVFF)

26

or

peptide derivatives27, (2)synthetic compounds (i.e., NQTrp)28, and (3) chemical compounds extracted from natural products (i.e., EGCG, curcumin, and resveratrol) 29-31. They inhibit the Aβ self-aggregation by efficient coassembling with hydrophobic β region of Aβ26,27,

32

and C-terminus29-31 as well as

hydrophilic interaction with charged residues at N-terminus28,25. For the natural inhibitor products, two most abundant components Tanshinone I (TS1) and Tanshinone IIA (TS2) of Tanshinone are confirmed experimentally33 as two powerful disaggregation and β-sheet-disrupting compounds (Figure 1).What is more, tanshinones can cross blood brain barrier 34 and are so far a very few small molecules to have the

5 Environment ACS Paragon Plus

ACS Chemical Neuroscience

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 33

ability of disaggregating preformed Aβ42 amyloid fibrils33. The disaggregation and inhibition mechanisms of such Tanshinone drugs remain obscured however. Thus predictions on the disaggregation/inhibition effects of TS1 and its derivative TS0 on the Aβ aggregates can guide ones to design new drugs. Toward this target, we observe that TS0 holds a more potent disaggregation ability than TS1 to the preformed Aβ42 amyloid fibril, by an amazing DR-preferential binding pattern. More importantly, DR is associated closely to the Aβ self-assembly and toxicity,

35, 36

whereas studies on

inhibitors by intervening DR still remain scarce.

5.341e-3

-5.341e-3

i+1 1

i

bend

Simulate

TS0 -5.519e-3

(1)

5.519e-3

β2 1β1 (2) 1 42 pAβ Fibril axis

i-2 i

TS1 -5.637e-3

5.637e-3 Simulate

β2 1 TS2

(1)

β1 1 pAβ40

(2)

Figure 1 Optimized structures of three inhibitors TS0, TS1, TS2 and structures of pAβ42 and pAβ40 from NMR PDB files (1) of refs[21 and 12] and simulations (2) with omitted disordered DR (Residues 1-8 for pAβ40 and 1-16 for pAβ42).Molecular electrostatic potential surfaces of TS0/1/2 from SCF density matrix (shown in medium grid) obtained at the B3LYP/6-31+G* level in combination with PCM model. Color scale changing from red to blue over each TS molecule stands for the electron-rich region shifting to electron-poor one, and value of the region is marked on the top of the color scale bar. For pentameric oligomers of Aβ40/42 (pAβ40/42), the crystal PDB structures alter a lot after they are simulated. For instance, intermolecular β1-β2 packing in pAβ40 between original i β1 chain and (i-2) β2 chain (iβ1-(i21 with serious twist6 and extended β1content37. Whereas there is a distinct 2)β2) becomes between iβ1-(i-1)β2 chains 18, 20 bend region in the β2 region of simulated pAβ42 observed and the bend makes the pAβ42 become iβ1-turn-(i+0.5) β2 bend motif (2) from original iβ1-turn-i+1β2 motif (1)12.

6 Environment ACS Paragon Plus

Page 7 of 33

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

ACS Chemical Neuroscience

RESULTS AND DISCUSSION

In this paper we probe the interactions between full-length Aβ40/Aβ42 and TS1/TS0 (Figure 1) with molecular dynamics (MD) simulations in explicit solvent. We observe a series of novel phenomena and propose a disaggregating mechanism for preformed Aβ42 amyloid fibrils, which differ from anyone supposed previously26-31,33,

38

. The full-length Aβ40/Aβ42 models include Aβ40/Aβ42 monomers and

preformed pentameric fibrils, and both are derived from experimental crystal PDB files21,12. The pentamer is chosen as the fibril-like model because it is verified sufficient to describe the binding characteristics of the various amyloid binding molecules50. The inhibitor TS1 is chosen due to its stronger ability than TS2 in inhibiting unseeded amyloid fibril formation and comparable ability to TS2 in disaggregating Aβ42 amyloids33. TS0 is a TS1 derivative with lack of a methyl and synthesized once by Wu group39 (Figure 1). Comparing the structures of TS1/TS2 and combining their inhibition ability, we expect that the TS0 performs more inhibitory efficacy than TS1 and thus is chosen as well. Results reveal that TS0 indeed does not let us down. TS0 is not only a powerful antiaggregation and β-sheetdisrupting compound, but also suggests a novel disaggregation mechanism. Inhibition of TS0/1 on the Monomeric Aβ42 (mAβ42) Aggregation. To verify the inhibition effect of TS0, we first employ a reliable scheme26 to calculate β-sheet content (Table S1 in the supporting information, SI) and contribution of each residue (Figure 2) to the β-sheet from averaging three parallel runs of miAβ42-TS0 system (i=1, 2, or 3) and one mAβ42 system (without TS0). Meanwhile we also compare these results to their TS1-involved counterparts.

7 Environment ACS Paragon Plus

ACS Chemical Neuroscience

mAβ42

i=1 i=2 i=3

0.8 0.4

Probablities of β−sheet

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 33

0.0 0

10

20

30

40

10

20

30

40

20

30

40

miAβ42-TS1

0.8 0.4 0.0 0

miAβ 42-TS0 0.8 0.4 0.0 0

10

Residue number

Figure 2 Contribution of each residue to the β-sheet of mAβ42 in the absence and presence of TS1/TS0. i=1, 2, 3 denote three parallel runs.

The probability results of β-sheet contents for mAβ42 simulation almost reproduce the structural characters of mAβ42 observed theoretically and experimentally. Five β-sheet regions, including F4-H6, Y10-H13(DR), L17-F19(β1), L34-V36 and V39-I41(β2), are in good agreement with Roychaudhuri et al’s experiment17 and Olubiyi and Strodel(O&S)'s

40

simulation with same force field. In addition,

residues 37-38(GG) excluded from the β2-content is also observed in Figure 2, consistent with the identification of unique characteristic of Aβ42 bend by other groups 17, 33, 40-42. The bend is almost absent in a folded mAβ4041. Our simulated result in Figure S1 of SI for mAβ40 further confirms this point that no bend is observed at the V36–G37 sequence but two β-contents locating in the ranges of 17LVFFA20 and 34MVGG37. In the presence of TS0/1, β-sheet (Table S1) in mAβ42 reduces greatly, indicating stronger inhibition effect of TS0/1 on mAβ42 and agreeing well with experimental observations33. As a criterion employed commonly to assess the stability and aggregation of Aβ26-31, β-sheet change is also employed to assess mAβ40+TS0/1 systems. Surprisingly, TS1 increases, instead of decreasing, 10% probability of β-sheet (from 17% to 27%) in mAβ40, indicating the enhancement of stability of mAβ40 in presence of 8 Environment ACS Paragon Plus

Page 9 of 33

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

ACS Chemical Neuroscience

TS1. TS0 decreases β-sheet content of mAβ40, as TS1 does to mAβ42. Probability calculation shows that the incremental β-sheet content mainly derives from the contributions of residues within DR region (see miAβ40-TS1 in Figure S1). To analyze the different effect of TS0/1 on mAβ42/40, we calculate the binding free energy (∆Eb in Table S2) and corresponding contribution from each residue. The ∆Eb of mAβ40-TS1 is -76.8 kJ/mol, larger than that (-54.8 kJ/mol)27 of LPFFD-mAβ40, and comparable to that (80.3 kJ/mol)43 of Curcumin-mAβ40. Interestingly, the latter two compounds indeed do as inhibitors and match the rule44 that the stronger the binding energy is, the better the inhibitory capacity will be, whereas TS1 does not. The abnormal effect of TS1 on mAβ40 implies that there should be other factors to determine the inhibitory or stable effect. Figure 3 reveals that the residues with stronger binding to the TS1 in mAβ40 locate in four regions (DR, β1, turn and β2) impartially, whereas those to TS1 in mAβ42 and to TS0 in both mAβ40 and mAβ42 mainly focus on locating the DR region. The DR-preferred location will decrease the intrapeptide van der Waals (vdW) energy from the DR, whereas the vdW energy from DR accounts for ca 30 per cent of total vdW contribution35. Therefore DR-preferred location, rather than the impartial location over the whole peptide, will disfavor to the stability of mAβ40, which accounts in part for that the unique binding feature of mAβ40-TS1 converts the disordered content of DR region in isolated mAβ40 to ordered β-sheet and results in more stability. The additional β-content from the DR is likely to increase the free energy barrier to convert the mAβ40 to its aggregation-prone state

36

and thus changes the oligomer size distribution, leading to new oligomer conformations and

hence changing the overall toxicity. In addition, hydrophobic (i.e., F4, Y10, V40, etc) and charged residues (i.e., K16, K28, D7, etc) in both mAβ40 and mAβ42 are two supreme ones for TS0/1 to bind. Meanwhile, most of binding energies from these charged residues are positive and amazing, opposite to the observed generally28,25.

9 Environment ACS Paragon Plus

ACS Chemical Neuroscience

Binding energy(kJ/mol)

K28

miAβ40−ΤS1

30

miAβ40−ΤS0

30

25

i=1 i=2 i=3

25

K16

20

15

10 D1

D1 R5

10

E22

R5

E3

20

15

5

K28

5

0

D23

0

-5

-5

F20

F4

-10

I31

Y10 0

10

20

30

V40

-10

40

miAβ42−ΤS1

30

Binding energy(kJ/mol)

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 33

F4

30

25

D7

0

E11 E22 10

20

I31 30

M35 V40 40

miAβ42−ΤS0

25

D7

20

E11

20

K16

15

15

10

E22 D23

D7

10

K28

E11

5 D1

5 0

0

-5

-5

F20

-10

F4 0

Y10 H13 10

20

Residues

V40 30

40

A2

-10

V39/V40

F20 0

10

20

30

40

Residues

Figure 3 Contribution of each residue to the binding free energy in miAβ40/42-TS1/0, in which the DR regions (residue sequences D1-S8 for mAβ40 and D1-K16 for mAβ42) are highlighted in green box.

Among all these miAβ42/40-TS1/0 complexes, m2Aβ42-TS1 is undoubtedly the most interesting sample due to its rich and arresting charged residues with either positive (D7, E11, K16) or negative (H13) binding energies (Figure 3, i=2). The positive binding energy of D7 derives from deformation of its carboxyl groups (-CO2) during its interaction with two TS1 drugs, inferred from the rotating -CO2 group (Figure 4 B,F). One oxygen(Od1) of -CO2 points toward and is sandwiched by the Oar atoms (Figure 4 I, F) of two TS1 drugs, and therefore rotation of -CO2 increases not only the steric barrier but also the electrostatic repulsion between -CO2 and two Oar atoms. Likewise, both -NH3 group of K16 (Figure 4 I, G) and -CO2 of E11 (Figure 4 I, H) are also apt to rotate and the rotation for -NH3 group keeps at constant 43.7~159.9 o. The 116.2 o difference just corresponds to 120 o of any of the three HzNz-Hz angles, indicating that the two adjacent Hz atoms exchange locations during the rotation of -NH3 group with one Hz H-bonding (2.106 Å) to Nδ atom of indole ring (H13) and another Hz to one carbonyl oxygen of TS1-1. The calculated distance between Nz and the mass center of TS1-1 is 6.462 Å (Figure 4 I, C), whereas the distance between Nz and the nearest carbonyl oxygen of TS1-1 is 4.673 Å. Therefore the week electrostatic interaction between the K16 and TS1-1 is at expense of breaking Hz-Nδ H-bond. The stronger H-bond than the electrostatic interaction leads to the positive binding energy of K16

10 Environment ACS Paragon Plus

Page 11 of 33

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

ACS Chemical Neuroscience

residue. The interaction between E11 and TS1-1 depends on the π-π interaction between one lone pair of Oe2 of E11 (Figure 4 I) and the π electrons of TS1-1 ring and, the H-bond (1.907 Å) between H(-Nζ) of H13 and another lone pair of the Oe2. Therefore, the rotation of carboxyl group (-CO2) in E11(Figure 4 H) not only results in the deformation of the residue itself but also breaks the H-bond and the π-π interaction, producing positive binding energy. On the contrary, the interaction between H13 and TS1 relies on the indole ring of H13 stacking against the ring TS1 in ca 3.693 Å distance, and therefore generates -9.1 kJ/mol binding energy. Thus these adverse binding contributions from charged residues can be attributed generally to serious deformation of the corresponding residue45. Studied individually, TS1 prefers to bind the F4, Y10, K16, F20 and V40 of both mAβ40 and mAβ42, in which most are the hydrophobic residues. In contrast,TS0 prefers to bind D1-D7, E11, E22, D23, K28 and V40 of both mAβ40 and mAβ42, in which most are hydrophilic and charged residues. In addition, D7 of mAβ42 is the unique preserved residue for TS1 to bind strongly in all the three parallel runs, whereas both D7 and E11 are two preserved ones for TS0 to bind, two times charged residues as much as that for TS1. Such preserved residues in mAβ40 are not obvious however. The inherent reason maybe derives from the less electrostatic potential in TS0 than in TS1 (Figure 1), rendering TS0 to binding more hydrophilic and charged residues than TS1. In fact, it makes no great differences for TS1 and TS0 in both structure (with or without a methyl group) and electrostatic potential (-5.519e-3 vs -5.314e-3). There appear great differences however in binding locations, residue types, and as result of effect on the mAβ40/42, which would be amazing and needs further investigation. The DR region in mAβ42 contains 16 unstructured residues12 and is the major region for hydrophilic and charged residues to locate, whereas the residues 916 in mAβ40 are ordered shape of β-sheet21, which hinders and reduces the free interaction between TS0/1 and the residues in the DR region.

11 Environment ACS Paragon Plus

ACS Chemical Neuroscience

It is noted that the result of these binding region is sensitive to the initial sampling for MD. Since two TS0 or two TS1 molecules are inserted the simulation box random, two TS1 molecules location in initial three miAβ42-TS1 systems are uncertain. Two TS1 molecules in m2Aβ42-TS1 system has been by chance closed to the DR in its unrestrained MD-starting conformations (Scheme 1), resulting in that both TS1 molecules mainly locate at the DR and produce positive binding property. The locations of two TS1 molecules in unrestrained MD-starting m1Aβ42-TS1 and m2Aβ42-TS1 are also different (not shown) and thus the binding regions (residues) are distinct (Figure 3). Anyway, the decrease of β-content in three miAβ42-TS1 systems is doubtless. (D) and (E) in Scheme 1 showcase the distinctions in location of two TS0 molecules, indicating the potential effect of the initial location of TS0 on the subsequent three parallel MD results. 16 12 8 4 0

E 3.482 1. 734

16 12 8 4 0

D 4. 861 2. 279

20 16 12 8 4

C 6. 462

12 8 4 0

B 3. 205

1.0 0.8 0.6 0.4 0.2 0.0

A 0

50

100

150

200

Time evolvement in ns

250

275

300

Dihedral angles in degree

D ista n c e in 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 12 of 33

200 160 120 80 40 0

H 88.4±37.4

200 160 120 80 40 0

159.9

G 43.7

200 160 120 80 40 0

F 93.9

0

50

100

150

200

250 275 300

Time evolvement in ns

Figure 4. Distance and dihedral angle evolutions over time in m2Aβ42-TS1 with data marked in blue from the 275-300 ns simulation. RMSD (A), Distances between mass centers of -CO2 group of D7 and of Oar atoms of two drugs (B), between Nz atom of K16 and mass center of Drug1 (C), between Hz(Nz) of K16 and Nδ of H13 (D), between Oe2 of E11 and H(Nζ) of H13 (E). Dihedral angles of ∠CαCβCGOd1 in D7 (F), ∠CDCENZHZ in K16 (G), ∠CβCDCGOe2 in E11(H). Residues D7, E11, H13 and K16 in licorice, and drugs TS1-1/2 in lines with Oar in beads are highlighted and shown in (I).

12 Environment ACS Paragon Plus

Page 13 of 33

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

ACS Chemical Neuroscience

Table 1 Number of TS0/1 locating at each regions of pAβ40/42 and, number of H-bonds and TS-TS stacking pairs DR β1 turn β2 β1-β2 H-bonds TS-TS Complexes 5.3(0) 8 5 1 1.7 0.6 1.3 pAβ40-TS1 2.4 (0.7) 3 3 1.7 3 1 1 pAβ40-TS0 3.1 8 6 2 1.3 0.6 pAβ42-TS1 2.7 6 8 1 1 pAβ42-TS0 The number in the parenthesis denotes H-bond number generated by the TS1/0 and the residues D23/K28. All the data obtained are the average of three parallel runs. For example, 0.7 H-bond indicates H-bond observed only in two parallel runs.

Disaggregation of TS0/1 on pentameric Aβ42 (pAβ42) fibrils The β-sheet contents in pAβ42-TS0/1 fibrils decrease by 8% and 11% (Table S3) compared to that in the bare pAβ42, indicating the potent disaggregation effect on pAβ4233. The TS0/1 in the pAβ40 systems reverses from it does in the pAβ42 counterparts. More specifically, TS0 increases the β-sheet content of pAβ40 by 4%, TS1 also preserves the β-sheet content (47%) as the same as that of the bare pAβ40, indicating that TS1 does not decrease the stability of pAβ40. To elaborate distinct effects of TS0/1 on pAβ40/42, we calculate the twist angles (Figure S2) of pAβ40/42 and the contributions of each residue to the binding energy (Figure 5) in absence and present of TS0/1. We find that the twist angle increases by 7.6 decreases by 3.2

o

o

after TS0 drugs are attached at pAβ40, but

when TS0 becomes TS1. The most TS0/1 drugs locate around the DR regions and

partly at the salt-bridge residues (Figure 5). In contrast, the number of TS1 (5, or 50%) in DR region is more than that of the TS0 (3, or 30%); the number in turn region reverses however (1.7(17%) vs 3 (30%); see Table 1). There are four-fifths of D23-K28 salt-bridge pairs broken and reaching out to the solutions, with only

(i=4)D23-(i=2)K28

salt-bridge preserved in the stable pAβ40 (i definition in Figure

1).Therefore these D23/K28 residues extending outward provide chances for TS0/1 to contact. Further, more contact between TS0 and the turn region heralds more H-bonding between TS0 and D23/K28. Comparisons show that there is 0.7 H-bond (averaged by three runs) generated in the turn region of

13 Environment ACS Paragon Plus

ACS Chemical Neuroscience

30

p1Aβ40−TS1

Binding energy(kJ/mol)

R5

R5 K28

E11

E22

-5

F4

D7

Y10 V12 10

50

E11

-10

-10 20

30

E3

Y10

0

40

V39

F19

10

20

30

F20 Y10 0

10

p2Aβ40−ΤS0

30 D1

E3

K28 D7

20

10

K28

E22 E11

E3

D23

R5

40

K16

20

D1

30

p3Aβ40−ΤS0

K28 30

20

20

40

p1Aβ40−ΤS0

V40

F4

-15

40

40 30

E22

D7

0 -10

K28

0

K16

10

0

0

Binding energy(kJ/mol)

E11

5

K16

one two three four five

p3Aβ40−ΤS1

10

D23

20

E3 10 D1

-20

10

E22

10

0

0

0 -10

-10

D7

25

10

20

30

40

10

20

30

40

H13

V36/G37 K16

-20 0

10

20

p2Aβ42−ΤS1

K16

E22

15

20

D7

15

30

K16

10

D23

40

p3Aβ42−ΤS1

20

30

R5

D1

D23

Y10

0

-10

p1Aβ42−ΤS1

D1

20

V40

F4

E11

0

Binding energy(kJ/mol)

15

p2Aβ40−ΤS1

30

20

K28

5

10 10

5 0

0 -5

0

-10

-5 -10

E22/D23

-15

F4

I41/A42

Y10/E11

0

-10 D1 F4 -20

I32 M35

E11

-20

D7

10

20

80

30

40

0

p1Aβ42−ΤS0

15

K16

10

60

10

R5 D7

20

30

40

p2Aβ42−ΤS0

E11 20 D1

E3/F4

5

40

0

30

10

20

H6

Residue

30

40

K16

30

40

p3Aβ42−ΤS0

D23

-10

D23

-20 0

K28

0

Y10

-15 E22

20

10

E3 Y10

10

20 D1 R5

-10 D1 E3

0

50

-5

0

0

Y10/E11

60

K28

K16

40

-20

M35

V18

-15 D1

-25

-20

Binding energy(kJ/mol)

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 33

10

20

30

40

E3

-20

Residue

0

D7

E11 10

20

30

40

Residue

Figure 5 Residue contributions to the binding free energies in pAβ40/42-TS1/TS0, in which the DR regions (residue sequences D1-S8 for pAβ40 and D1-K16 for pAβ42) are highlighted in green box. Five separate Aβ42/Aβ40 monomers in each pentamer are marked with one, two, …, five and shown in different colors.

pAβ40-TS0 (Table 1), whereas the number is zero for pAβ40-TS1, indicating that hydrophobic interaction between pAβ40 and TS1 is more than that between pAβ40 and TS0 in the turn region. The observation is further verified by Figure 5, in which the probability (~56.6%) of TS0-D23/K28 contact is far more than

14 Environment ACS Paragon Plus

Page 15 of 33

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

ACS Chemical Neuroscience

that (~33.3%) of TS1-D23/K28. In addition, the total H-bonds (2.4) between TS0 and pAβ40 is less than that (5.3) between TS1 and pAβ40, indicating that both pAβ40 and TS0 will hold more charged residues or group (i.e., carboxyl oxygens of TS0) to reach out to the solutions and the charged residues in DR region of pAβ40-TS0 is less than that of pAβ40-TS1. More charged residues in solutions for pAβ40-TS0 than for pAβ40-TS1 indicate incremental twist of pAβ40, being consistent with the observation that charged termini of peptide will enhance ca 30% twist angle relative to that with the neutral one30,46. It is obvious that the incremental twist angle in pAβ40-TS0 corresponds to the enhanced stability of pAβ40 fibril, inferring from the more β-sheet contents in pAβ40-TS0 (51%) than in pAβ40 (47%). Interestingly, the presence of either TS1 or TS0 increases the twist angle of pAβ42 remarkably (5.7 o/7.3 o of pAβ42TS1/0), but decreases the β-sheet contents in pAβ42-TS1/0, indicating the reduced stability of pAβ42TS1/0 and resultant antiaggregation effect on the pAβ42 fibril33. In order to probe the disaggregated mechanisms of TS1/0 to pAβ42 fibril but enhanced one to pAβ40, we display the secondary structures of pAβ40/42 (Figure 1) and pAβ40/42-TS1/0 in Figure 6, in which a clear bend region composed of 37GGV39 can be observed in pAβ42, rather than pAβ40,18-20 and distinct changes induced by TS1/0 are also presented. The attached TS1 mainly locates at the DR region of pAβ42 and with small amounts at D23-K28 and C-terminal regions (Figure 5). Note that there are 80% TS1 (Table 1) closing together in π-π stacking or hydrophobic interaction when some of them are interacting with the surrounding residues of the pAβ42, which account for why the disaggregation effect of TS1 on Aβ42 mature fibrils at a molar ratio of Aβ:TS1 = 1:2 is stronger than that at either 1:1 or 1:533. It is clear that lower ratio (i.e. 1:1) would imply that pAβ42 has still rooms for additional TS1 molecules to attach directly, whereas higher ratio (i.e., 1:5) only indicates the increase of the stacked TS1-TS1 pairs but makes no contributions to the direct TS1-pAβ42 interaction. This indicates that parts of TS1 molecules do not contact directly with the pAβ42 and the distribution of all the ten TS1 molecules

15 Environment ACS Paragon Plus

ACS Chemical Neuroscience

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 33

instead of those attached directly would be more informative for insight into the changes of pAβ42 secondary structure. Table S4 in SI indicates the distributions by calculating distances r1 and r2 of two pairs of mass-center (MC), where one MC comprises Cαs of five Leu17 in pAβ42(Gly9 in pAβ40), as the five residues are just the boundary between DR and β1 regions and taken as a pivot (O). Another two MCs are composed of ten TS0/1 drugs (TS) and

Cαs of five Asp1 (Dα) in the N-terminal tip,

respectively. Therefore r1 and r2 denote the distances between O and TS and between O and Dα, respectively. The comparison between r1 and r2 can unveil the overall distribution of TS0/1 relative to either Leu17 or Asp1. Results show that the r2 in pAβ42 is 1.38 nm, far shorter than the its native length and even that (1.67 nm) of pAβ40. By the way, the r2 is 1.3 nm in Xu et al's study24 calculated by using the MC of β-hairpin of mAβ40 as the pivot, which is comparable to our 1.67 nm although different pivots and models (Xu et al's mAβ40 vs our pAβ42) are employed. Noted that r2 in pAβ42 and pAβ40 are composed of 16 and 8 residues respectively. Therefore, shorter r2 in pAβ42 than in pAβ40 implies that the DR region in pAβ42 would look more like a canopy, covering on the trunk twined by β1 and β2 strands (Figure 6 (b)). The entanglement of β1 and β2 strands is small as the twist angles of an experimental mature Aβ42 fibril12 and of our present pentameric Aβ42 model are only 0.45

o

and 1.7 o, respectively.

The presence of either TS1 or TS0 can increase markedly the twist angle of the pAβ42, comparable to that of pAβ40, indicating that β1 and β2 strands entangle together currently. TS0/1 effect on the secondary structure of pAβ42 is mainly presented in the disappearance of bend region 37GGV38 and decrease of the pAβ42 stability. In all these simulations for three pAβ42-TS1/0 samplings, only the 37GGV38 region in p3Aβ42-TS0 preserves bend shape. Figure 6 (a3/b3) clearly displays the changes of the region along with the time evolution by comparing the dihedral angels composed of four Cα atoms of 37GGVV40. The dihedral angles in pAβ42 (original from PDB file of Lührs et al12) and p3Aβ42-TS0 become ca 0 o from original 180 o at about 12.5 ns (Figure 6(a3)) and 15.0

16 Environment ACS Paragon Plus

Page 17 of 33

ns (Figure S3), respectively. The angles in other five TS1/0 involved samplings preserve ca 0 o, which is favorable to generate the packing of F19 against G3812. In fact, the packing degenerates and becomes characteristic contact of iF19-iG37 (see the definition for i in Figure 1 and the packing in Figure 6(a1)) in the simulated pAβ 42 model, whereas G38 is at the bottom of the bend groove and thus V40 L17

V39

F19

G38

G37

(b1)

(a1) •

Fibril

80 40

pAβ 42

1.2

200

1.0

160

0.8

120

0.6

80

0.4

40

0.2

0

1.2

RMSD in Å

120

RMSD in Å

160

(b)

(b2)

(a2)

Angle in degree

(a) 200

Angle in degree

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

ACS Chemical Neuroscience

p1Aβ42-TS0/1

1.0 0.8 0.6 0.4

0 0

10

20

30

40

50

ns

0

20

40

60

80

100

ns

0.2

(b3)

(a3)

Figure 6.Structures of both pAβ42 (a) and a representative (b) of pAβ42-TS0/1. The characteristic bend regions 37GGVV40 are amplified in (a1/b1) and side-looked in (a2/b2), respectively. The fibril axis is marked in a red dot in (a) or circled dot in (b) and the TS0/1 drugs are shown in violet solvent. The coroniform DR regions in (a) and (b) are highlighted in tube to distinguish from the gray β-hairpin motif of pAβ42. The time evolutions of RMSD (in black) and of dihedral angles (in blue) composed of four Cα atoms of the bend region are displayed in (a3) for pAβ42 and in (b3) for the representative (b) of pAβ42-TS0/1. The four Cα atoms in (a) and (b) are shown in bead and with different colors, red for G37, green for G38, orange for V39 and blue for V40, respectively.

far away from the F19. G37 protrudes toward intramolecular F19 and meanwhile the side chains of V40 residues point to the intramolecular aromatic rings of F19 residues in hydrophobic interaction (Figure 6(a1)), that not only enhances stability of the Ar1 Aβ42 fibril but also makes it impossible for the intermolecular contact of G37-M35 to generate a Ar2 Aβ42 fibril13 by docking two protofilament subunits in opposite orientation14-16. In the presence of TS0/1, the iV40-iF19 interaction also disappears (excluding that in p3Aβ42-TS0, not shown).The packing iF19-iG37 becomes iF19-i+1G38 (Figure 6(b2)), indicating the forward shift of

17 Environment ACS Paragon Plus

ACS Chemical Neuroscience

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 33

the β1 region along the fibril axis and as a result, the fibril twist generates. As a unique complex with preserved bend region, p3Aβ42-TS0 has the packing with iF19 against

i+1G37,

also indicating a forward

shift of the β1 region and a fibril twist. Therefore, except for the case of p3Aβ42-TS0, the Aβ42 in the rest five pAβ42-TS0/1 complexes would be apt to generate further two-fold Aβ42 fibril Ar2 along with the disappearance of the bend region and forward shift of the β1 region. The C-terminal bend is regarded17 as the sine qua non of Aβ42 and as a result, facilitating its formation in Aβ40 creates a more Aβ42-like peptide; On the contrary, destabilizing the bend in Aβ42 makes this peptide “Aβ40-like”. Previous predictions revealed that the Ar1 and Ar2 Aβ42 fibrils would be symbiotic relationship before they grow up into a mature fibril37 although the Ar1 one is more stable and apt to be generated20, 47. The latest study observed that hexameric Aβ42 fibril still preserves bend region 37GGV3818 in the C-terminus in a small Aβ42 fibril whether it interacts with Amylin37 oligomers or not20, further highlighting the potent interfering effect of TS0/1 on the stability of Aβ42 fibril. The interference outcome offers a plausible way for conversion from Ar112 to the Ar2 one10-11 observed experimentally. TS0/1 interference effect on the pAβ42 morphology is driven by the DR-TS0/1 interaction and therefore indicating that β1 region is the core of twist. The β1 region was once argued as a rigid48 and immobile 13β-sheet core as well as the origin of aggregation10, thus both the β1 mobility and the twist of whole β-hairpin driven by the TS0/1 interaction at the DR region indicate strong interference ability of TS0/1 and special disaggregation mechanism. A novel disaggregation mechanism. It is clear that the interaction between TS0/1 and the DR region of pAβ42 drives the shift of the β1 region and consequent fibril twist around its fibril axis. The direct outcome is that β-sheet contents reduce and stability of pAβ42 decreases. The indirect outcome is that following lateral association of Aβ aggregates becomes difficult and thus inhibits fibril growth33 because the lateral association was argued as the major source of fibril elongation

18 Environment ACS Paragon Plus

49,50

. The twisted

Page 19 of 33

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

ACS Chemical Neuroscience

pAβ42 enhances a steric energy barrier for lateral association of additional Aβ peptides in the "docklock" mechanism51. In comparison with TS1, TS0 reduces more β-sheet contents and induces more serious twist of pAβ42, indicating more potent disaggregation ability. Moreover, the three samplings simulated individually show that almost all the ten TS0 molecules locate at the DR region (Table S4 in SI, Figure 5, and Figure 6 (b)), indicating a novel disaggregation mechanism and differing from anyone supposed or any inhibitor presented previously26-31, 33. The differences include three aspects. One is the binding region identification. The common characters observed26-31, 33 previously are the widely location of these inhibitors around four regions of Aβ42, DR, β1, turn and β2, rather than focusing on a specific region. This is, the binding features commonly as “multiple binding sites" on the Aβ surface26-31, 33 or in the core of the Aβ fibril observed recently32. Chebaro et al once suggested a preferential binding for several drug molecules but at β1 region (residues 17-21)26. Two, both the wide positive binding energies between charged residues (major in DR region) and the TS0, and huge ∆GGB value (454 kJ/mol) in mAβ42-TS0 further implies DR the preferential region for TS0 to locate, as the DR comprises the overwhelming majority of the charged residues. The interaction energies reported28-26 availably between charged residues and other inhibitors are not only commonly negative values however, but also involve far less charged residues than the hydrophobic ones. A interesting comparison can be made between present result and Lee' et al's52, in which a N-terminus binding preference was also observed. Unlike our mAβ42-TS0 observation, the Nterminus binding in Lee' et al's study52 is occurred by stacking erythrosine B (ER), a xanthene food dye, with aromatic side chains at the N-terminus region of mAβ40, and resulting in strong binding energy (negative value). Investigation for the structural difference between ER and TS0 would have its great significance for deep insight into the different N-terminus binding mechanisms, and targeting novel derivative inhibitors. A2V-mutated Aβ1-6 peptide was just confirmed experimentally to be effective in

19 Environment ACS Paragon Plus

ACS Chemical Neuroscience

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 33

interfereing with the formation of toxic Aβ42 oligomeric species by binding the DR of Aβ4253. The similarity for the bindings between LPFFD sequenc and the β1 region27 and between A2V-mutated Aβ16

peptide and DR53 encourage us to estimate the property of the latter binding energy as a negative one,

as well as the former was27. Three, the population distribution for ten TS0 molecules in the pAβ42-TS0 (Table S4 in SI) further confirms DR the special and unique region for TS0 to locate, although there is trifling TS0 locating closely to D23/K28, the charged residues as well. The disordered and charge-rich DR was clearly demonstrated by Xu et al24 that it can transiently interact with the central regions of mAβ40 through hydrophobic interactions and hydrogen bonding and, therefore enhances the β-hairpin stability of the amyloid β peptide in the central amyloidogenic region by transferring entropy to the surrounding solvation shell with its fluctuating charged N-terminal residues. Therefore TS0 locating at this region disturbs the positive role to stability of Aβ42, as a metal ion does. It is known that the Nterminus of Aβ is the primary metal binding site, and the metal binding could modulate Aβ selfassembly54, 55 just by decreasing entropy of the Aβ24. TS1 also prefers to locate the DR region but with part at the turn and β2 regions, different from the observations from Wang et al33, in which multiple binding sites of pAβ42 for TS1 binding were suggested to prevent the lateral association between different aggregates or disturb the stability of pAβ42 structure. They highlighted the significance of the C-terminal β-sheet and the hydrophobic residues, that interferes with lateral association of Aβ oligomers into higher-order aggregates by binding TS1. The difference derives from the pAβ42 model built. The model in Wang et al is based on the U-shaped Aβ9-40 conformation from Tycko’s experimental data21, which is quite different from the Aβ42 cystal configuration12. For example, the Asp23-Lys28 salt bridge in the Aβ40 oligomer structure is intramolecular21, whereas in Aβ4212 it is intermolecular47. The DR region in Aβ40 is composed of 8 residues and more ordered21, whereas that in Aβ42 comprises 16 residues and all of them are

20 Environment ACS Paragon Plus

Page 21 of 33

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

ACS Chemical Neuroscience

unstructured12. The packing between two β-strands in Tycko’s Aβ40 crystal PDB whereas that in Aβ42 is iβ1 against

12 i+2β2 .

21

is iβ1 against

i-3β2,

The fact of that r1 is unanimous shorter than r2 in the three

parallel piAβ40-TS1 runs (Table S4 in SI), indicates that if we use pAβ40 to build pAβ42, as Wang et al did33, then the TS1 molecules are more close to the β1/β2 region than the Asp1, declaring that TS1 molecules would prefer to close to the β strands, as argued by Wang et al. r1 is unanimous longer than r2 in the three parallel piAβ42-TS0 runs, indicating that, on the other hand, the TS0 molecules are more far away to the β1/β2 region than the Asp1 residues (N-terminal tip), also encouraging us to obtain above corollaries. Therefore the model difference employed heralds different disaggregation mechanisms obtained.

CONCLUSIONS The present prediction on the potent inhibitory/disaggregation ability of TS1 to Aβ42 reproduces the experimental observations of Wang et al33. The inhibitory/disaggregation mechanisms are different from each other however due to different models employed. Should the model of Aβ42 be built based on the architecture of Aβ40, similar mechanism to that of Wang et al is probably obtained. Meanwhile, present work reveals that TS1 can stabilize mAβ40 and even Aβ40 fibril, differing from its behavior to Aβ42 however since Aβ40 holds different architecture, especially short and relative ordered DR. As a derivative of TS1, TS0 can inhibit the misfolding of either mAβ40 or mAβ42 and disaggregate Aβ42 fibril but stabilize the Aβ40 fibril, further verifying the significance assessment for Aβ42/Aβ40 DR difference. TS0/1 prefers to interact with the charged residues of Aβ in positive binding energies and the DR of U-shaped Aβ40/42 becomes the major accommodation of TS0/1. The preferential DR binding is not only different from all those common inhibitors’ preference26-31, 33,32 in inhibition/disaggregation of Aβ40/42,

21 Environment ACS Paragon Plus

ACS Chemical Neuroscience

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 33

but also first observed in positive binding energy as far as we know. It modulates not only the stability of U-shaped Aβ40/42 conformation but also the possible morphology conversion of Aβ42 fibril. Although why TS0 is more prone to bind the DR than TS1 is not revealed completely, the prior trend of TS0>TS1>TS2 in binding DR can be affirmed absolutely. We hope that the trend can draw experimental chemists/pharmacists' attention to make new attempts based on TS0 and its further derivatives after validating the trend experimentally. Based on the identical DR binding preference, but distinct structural characteristics or different binding residues (TS0 prefers to bind the charged residues of DR in positive binding energy, whereas ER prefers to stack the aromatic residues in common negative binding energy52), the present study for TS0/1 offers an alternative for experimental chemists/pharmacists to further optimize and modify such DR preferential inhibitors, or balance both of TS0 and ER. Whether the prior trend is still valid to the S-shaped Aβ42 fibril needs to be further examined in the future due to the significant differences in not only the conformations (U-shape2, 5, 6 vs S-shape7-9) but also in the DR with different lengths of residue numbers (1612 vs 117-9). This would be another interesting topic.

METHODS Model Design Pentameric Aβ9-40 (pAβ9-40) and Aβ17-42 (pAβ17-42) are from PDB files of 2LMN21 and 2BEG12, respectively. After the missing residues 1-8 for Aβ40 and 1-16 for Aβ42 are supplemented by Chimera program56, full-length pAβ42 and pAβ42 are obtained, from which monomeric Aβ40 (mAβ40) and Aβ42 (mAβ42) are extracted. The structure of TS1 is built based on reference33 and TS0 is the derivative of TS1, in which the methyl of TS1 is substituted by a hydrogen. Then B3LYP/6-31G* method implemented in gaussian0957 is employed to optimize the two drug molecules. Both TS0 and TS1 (generic term, TS0/1) are shown in following Scheme 1, in which TS2 is also included for

22 Environment ACS Paragon Plus

Page 23 of 33

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

ACS Chemical Neuroscience

comparison. Since the disaggregation effect of TS1 on Aβ mature fibrils at a molar ratio of Aβ:TS1 = 1:2 is the strongest determined experimentally33, 1Aβ+2TS0/1 systems are built as simulated solutes.

A

B

D

E

C

F

Scheme 1. Sketches of TS1(A), TS0 (B), and TS2(C), and unrestrained MD-starting conformations of m2Aβ42TS1(D), m1Aβ40-TS0(E), and m3Aβ40-TS0(F) from simulated annealing. Yellow ball in (D-F) denotes the Asp1 of Nterminus. Purple ribbon stands for the helix content and bonds-shown are two TS1 or two TS0 molecules attached.

MD Simulation Details The force field parameters of TS0/1 are produced by PRODRG server58 for the next simulation, and shown in Tables S5 and S6 of SI. Then three parallel runs are performed for every 1Aβ+2TS0/1 system and one run for Aβ system without TS0/1. This is a successful scheme once used in another computational study31. 1Aβ+2TS0/1 system denotes any of the followings, piAβ40-TS1, piAβ40-TS0, piAβ42-TS1, piAβ42-TS0, miAβ40-TS1, miAβ40-TS0, miAβ42-TS1, and miAβ42-TS0 (i=1, 2, and 3), and the Aβ system without TS0/1 stands for any of the followings, mAβ40, mAβ42, pAβ40 and pAβ42. Without i subscript followed indicates the mean value of the TS0/1 involved systems. For example, pAβ40-TS1 presents the mean result of p1Aβ40-TS1, p2Aβ40-TS1 and p3Aβ40-TS1.

23 Environment ACS Paragon Plus

ACS Chemical Neuroscience

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 33

Simulated annealing (SA) simulations and molecular dynamics (MD) simulations are performed for all these systems using GROMACS-4.6 software package59 with gromos53a6 force field60. Each system is immersed into a cubic box of TIP3P water with at least 8 Å distance around the solute and an appropriate number of sodium counterions is added to maintain the electroneutrality. Twice as many TS1/TS0 as mAβ40/42/pAβ40/42 is inserted random in the box (see three samplings in Scheme1). The random method had been verified successfully in treating the interaction of drug and alzheimer amyloidβ peptide by several groups61-63. The van der Waals interactions are calculated using a cutoff of 1.0 nm. The non-bonded interaction pair list, with a cutoff of 1 nm, is updated every 5 fs. The particle mesh Ewald method64 is employed to treat the electrostatic interactions with a cutoff of 1.0 nm. The LINCS algorithm65 is used to constrain the lengths of all covalent bonds to reduce the calculation time. The water molecules are restrained using the SETTLE algorithm.66 A force constant of 1000 kJ/mol/nm is used for position restraints during the SA. For the subsequent unrestrained MD simulations, the same parameters are used as did for the restrained MD, except that the temperature is maintained at 310 K and no positional restraints are applied. The V-rescale temperature coupling67 is used to control the temperature at 310K. The berendsen pressure coupling method68 is applied to describe the barostat with constant pressure of 1 atm. All MD simulations are conducted using periodic boundary conditions. The simulation time is set from 100 ns to 300 ns, dependent on the system equilibrium. Root mean square deviation (RMSD) analysis is conducted on the base of the backbone of the m/pAβ42/40. Selected MD structures are showed by VMD69 software. SA details Due to the arbitrarily orientation of the first 8/16 residues in DR regions and the experimental stability of β-hairpin motif in fibrils, α carbon atoms (Cα) of residues 9-40(Aβ40)/1742(Aβ42) are position restrained and the rest are not restrained in the SA simulations for these U-shaped pAβ40/42 and their TS0/1 involved systems to obtain potential energy minima. SA with repeating

24 Environment ACS Paragon Plus

Page 25 of 33

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

ACS Chemical Neuroscience

heating/cooling cycles and covering a broad temperature range (from 310 up to 500 K) is employed for these fibril-like piAβ40/42 and piAβ40/42-TS0/1. In detail, the whole system (peptide, water, and ions with/without drug) is coupled to 310 K at 0 ps firstly and then the temperature is linearly risen to reach 500 K at 20 ps. After keeping constant temperature at 500 K for 40 ps, the temperature is linearly decreased from 500 K to 310 K at 70 ps. Subsequently the temperature is kept constant at 310 K for 30 ps until a 100 ps SA cycle is completed. After finishing 100 such SA cycles (10 ns), a very flexible DR region is obtained. Only the structure at 10 ns in the SA is employed as the original sample for bare pAβ40/42, whereas three structures collected at each 3 ns during the 10 ns are used as the original samples for the piAβ40/42-TS0/1 (i=1, 2, 3) in the following unrestrained MD simulations at constant 310 K because such a SA scheme was verified70 successfully for tri- and pentrameric Aβ42 systems. For the mAβ40/42 and their TS0/1 involved systems, the same SA scheme is performed except for the Cα position restraint. That is, there is no position restraint employed for these monomeric systems. After SA, the unrestrained MD-starting conformations are obtained.

For the TS0/TS1 involved

mAβ40/mAβ42 systems, the initial mAβ40/mAβ42 samplings for SA are from MD-equilibrated mAβ40/mAβ42 structures. Thus mAβ40 and mAβ42 still remain the characters of the corresponding crystal structures in the unrestrained MD-starting conformations, indicating these samplings are suitable for the subsequent unrestrained MD. For example, mAβ40 in m1Aβ40-TS0 and m3Aβ40-TS0 in Scheme1 are still reasonably consistent with that of two Aβ40 crystal structures, 1AML71 (pH = 2.8) and 1BA4

72

(pH =

5.1), in which the helix component is also dominant. That is, structure 1AML contains two helix regions (residues Gln15−Lys23 and Ile31−Met35), whereas 1BA4 has one long helix, being composed of residues from Gln15 to Val36. Analysis Means Secondary structure analysis is performed using the dictionary secondary structure of protein (DSSP) method73. The RMSD-based clustering method within a cutoff of 2.0 Å for

25 Environment ACS Paragon Plus

ACS Chemical Neuroscience

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 33

protein is used to generate the average structures. A hydrogen bond is considered to be formed if the distance between N(O) and O atoms is ≤ 3.5 Å and the angle of N-H…O is ≥ 150 ° 74. The binding free energy (∆Gb) of a TS1/TS0 molecule with either Aβ40/42 monomer or pentamer is estimated using the molecular mechanics/Poisson-Boltzmann surface area (MM/PBSA) method75,76. The calculational details employed for the MM/PBSA method are almost the same as that in the our previous paper45. Briefly, ∆Gb between a ligand (l) and a receptor (r) in a complex (c) is calculated as, ∆Gb = Gc – (Gr + Gl)

(1)

where Gc, Gr and Gl are the free energies of the complex (c), receptor (r) and ligand (l), respectively. Complex herein denotes separate Aβ42-TS0, Aβ40-TS0 Aβ42-TS1 or Aβ40-TS1. Correspondingly, receptor stands for separate Aβ42 or Aβ40 monomer in either pAβ42/pAβ40 or mAβ42/mAβ40, and ligand refers to either TS0 or TS1, respectively. Detailed calculations for Gc, Gr and Gl refer to the Ref 45. The final binding energy for each system listed in Tables or Figures is the mean of three separate trajectories. Calculational Schemes for Twist Angle Cα atoms of A30/K28 and V36/V36 residues in the top and bottom layers of pAβ40/pAβ42, respectively, are selected as reference points because the two residues are almost in the center of each chain and therefore least affected by the entropic fluctuations presenting on the tails of each chain of the layers. Then the intersection angles between lines linking the two reference points correspond to the twist angles. Finally, the angle is further normalized by the layer number so that an angle for two adjacent layers can be obtained. Whereas the angles in Figure S2 of SI denote the averaged result of three parallel runs.

ASSOCIATED CONTENT Supporting Information The Supporting Information is available free of charge on the ACS Publications website at DOI: xxx/acschemneuro.xxx

26 Environment ACS Paragon Plus

Page 27 of 33

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

ACS Chemical Neuroscience

Probability of secondary structures of mAβ40/42 in the absence and presence of TS1/TS0 in Table S1, the binding free energies between TS1/0 and mAβ40/42 and their partitions in Table S2, probability of secondary structures of pAβ40/42 in the absence and presence of TS1/TS0 in Table S3, distances(r1/r2) obtained at the largest distribution peaks in Table S4 (distribution details in Figure S4), force field parameters of TS0 and TS1 in Tables S5 and S6, contributions of each residues to the β-sheet of mAβ40 in the absence and presence of TS1/TS0 in Figure S1, twist angles θ and atomic identification listed in Figure S2, the time evolutions of RMSD and of dihedral angles composed of Cα atoms in 37GGVV40 in Figure S3. AUTHOR INFORMATION Corresponding Authors *E-mail: [email protected] *E-mail: [email protected] ORCID Hongqi Ai: 0000-0002-9933-390X Author Contributions Y. D., W. Z., D. H. and H. A. designed the research; D. Y., W. Z., and D. H performed the simulations. D. Y., W. Z., D. H., H. A. and B. K. analyzed the data. Y. D., W. Z., D. H. and H. A. prepared the manuscript. Funding We are thankful for financial support from National Natural Science Foundation of China (Nos. 20973084, 21211140340). Notes The authors declare no competing financial interest.

27 Environment ACS Paragon Plus

ACS Chemical Neuroscience

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 28 of 33

REFERENCES (1) Hamley, I. W. (2012) The amyloid beta peptide: a chemist's perspective. Role in Alzheimer's and fibrillization, Chem. Rev. 112, 5147-5192. (2) Tycko, R. (2011) Solid-State NMR Studies of Amyloid Fibril Structure, Annu. Rev. Phys. Chem. 62, 279-299. (3) Jarrett, J. T., Berger, E. P., and Lansbury, P. T. (1993) The carboxy terminus of the .beta. amyloid protein is critical for the seeding of amyloid formation: Implications for the pathogenesis of Alzheimer's disease, Biochemistry 32, 4693-4697. (4) Roher, A. E., Lowenson, J. D., Clarke, S., Woods, A. S., Cotter, R. J., Gowing, E., and Ball, M. J. (1993) beta-Amyloid-(1-42) is a major component of cerebrovascular amyloid deposits: implications for the pathology of Alzheimer disease, Proc. Nat. Acad. Sci. U.S.A. 90, 10836-10840. (5) Paravastu, A. K., Leapman, R. D., Yau, W.-M., and Tycko, R. (2008) Molecular structural basis for polymorphism in Alzheimer's β-amyloid fibrils, Proc. Nat. Acad. Sci. U.S.A. 105, 18349-18354. (6) Paparcone, R., and Buehler, M. J. (2009) Microscale structural model of Alzheimer A beta(1-40) amyloid fibril, Appl. Phys. Lett. 94, 243904-243903. (7) Xiao, Y., Ma, B., McElheny, D., Parthasarathy, S., Long, F., Hoshi, M., Nussinov, R., and Ishii, Y. (2015) Aβ(1–42) Fibril Structure Illuminates Self-recognition and Replication of Amyloid in Alzheimer’s, Nat. Struct. Mol. Biol. 22, 499-505. (8) Colvin, M. T., Silvers, R., Ni, Q. Z., Can, T. V., Sergeyev, I., Rosay, M., Donovan, K. J., Michael, B., Wall, J., Linse, S., and Griffin, R. G. (2016) Atomic Resolution Structure of Monomorphic Aβ42 Amyloid Fibrils, J. Am. Chem. Soc. 138, 9663-9674. (9) Wälti, M. A., Ravotti, F., Arai, H., Glabe, C. G., Wall, J. S., Böckmann, A., Güntert, P., Meier, B. H., and Riek, R. (2016) Atomic-resolution structure of a disease-relevant Aβ(1–42) amyloid fibril, Proc. Nat. Acad. Sci. U.S.A. 113, E4976-E4984. (10) Zhang, Y., Rempel, D. L., Zhang, J., Sharma, A. K., Mirica, L. M., and Gross, M. L. (2013) Pulsed hydrogen-deuterium exchange mass spectrometry probes conformational changes in amyloid beta (Aβ) peptide aggregation, Proc. Nat. Acad. Sci. U.S.A. 110, 14604-14609. (11) Dahlgren, K. N., Manelli, A. M., Stine, W. B., Baker, L. K., Krafft, G. A., and Ladu, M. J. (2002) Oligomeric and Fibrillar Species of Amyloid-β Peptides Differentially Affect Neuronal Viability, J. Biol. Chem. 277, 32046-32053. (12) Lührs, T., Ritter, C., Adrian, M., Riek-Loher, D., Bohrmann, B., Döbeli, H., Schubert, D., and Riek, R. (2005) 3D structure of Alzheimer's amyloid-β(1–42) fibrils, Proc. Nat. Acad. Sci. U.S.A. 102, 17342-17347. (13) Sato, T., Kienlen-Campard, P., Ahmed, M., Liu, W., Li, H., Elliott, J. I., Aimoto, S., Constantinescu, S. N., Octave, J.-N., and Smith, S. O. (2006) Inhibitors of Amyloid Toxicity Based on β-sheet Packing of Aβ40 and Aβ42, Biochemistry 45, 5503-5516. (14) Ahmed, M., Davis, J., Aucoin, D., Sato, T., Ahuja, S., Aimoto, S., Elliott, J. I., Van Nostrand, W. E., and Smith, S. O. (2010) Structural conversion of neurotoxic amyloid-[beta]1-42 oligomers to fibrils, Nat. Struct. Mol. Biol. 17, 561-567. (15) Rajadas, J., Liu, C. W., Novick, P., Kelley, N. W., Inayathullah, M., Lemieux, M. C., and Pande, V. S. (2011) Rationally designed turn promoting mutation in the amyloid-β peptide sequence stabilizes oligomers in solution, PloS One 6, e21776. (16) Török, M., Milton, S., Kayed, R., Wu, P., Mcintire, T., Glabe, C. G., and Langen, R. (2002) Structural and dynamic features of Alzheimer's Abeta peptide in amyloid fibrils studied by sitedirected spin labeling, J. Biol. Chem. 277, 40810-40815.

28 Environment ACS Paragon Plus

Page 29 of 33

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

ACS Chemical Neuroscience

(17) Roychaudhuri, R., Yang, M., Deshpande, A., Cole, G. M., Frautschy, S., Lomakin, A., Benedek, G. B., and Teplow, D. B. (2013) C-Terminal Turn Stability Determines Assembly Differences between Aβ40 and Aβ42, J. Mol. Biol. 425, 292-308. (18) Murakami, K., Irie, K., Ohigashi, H., Hara, H., Nagao, M., Shimizu, T., and Shirasawa, T. (2005) Formation and Stabilization Model of the 42-mer Aβ Radical:  Implications for the Long-Lasting Oxidative Stress in Alzheimer's Disease, J. Am. Chem. Soc. 127, 15168-15174. (19) Yan, Y., and Wang, C. (2006) Aβ42 is More Rigid than Aβ40 at the C Terminus: Implications for Aβ Aggregation and Toxicity, J. Mol. Biol. 364, 853-862. (20) Baram, M., Atsmon-Raz, Y., Ma, B., Nussinov, R., and Miller, Y. (2016) Amylin-A[small beta] oligomers at atomic resolution using molecular dynamics simulations: a link between Type 2 diabetes and Alzheimer's disease, Phys. Chem. Chem. Phys. 18, 2330-2338. (21) Petkova, A. T., Yau, W.-M., and Tycko, R. (2006) Experimental Constraints on Quaternary Structure in Alzheimer's β-Amyloid Fibrils, Biochemistry 45, 498-512. (22) Cheon, M., Hall, C. K., and Chang, I. (2015) Structural Conversion of Aβ17–42 Peptides from Disordered Oligomers to U-Shape Protofilaments via Multiple Kinetic Pathways, PloS Comput. Biol. 11, 1-23. (23) Petkova, A. T., Ishii, Y., Balbach, J. J., Antzutkin, O. N., Leapman, R. D., Delaglio, F., and Tycko, R. (2002) A structural model for Alzheimer's beta-amyloid fibrils based on experimental constraints from solid state NMR, Proc Natl Acad Sci U S A 99, 16742-16747. (24) Xu, L., Nussinov, R., and Ma, B. (2016) Allosteric stabilization of the amyloid-β peptide hairpin by the fluctuating N-terminal, Chem. Commun. 52, 1733-1736. (25) Doig, A. J., and Derreumaux, P. (2015) Inhibition of protein aggregation and amyloid formation by small molecules, Curr. Opin. Struct. Biol. 30, 50-56. (26) Takahashi, T. M., H. (2008) Peptide and protein mimetics inhibiting amyloid beta-peptide aggregation., Acc.Chem. Res. 41, 1309-1318. (27) Viet, M. H., Ngo, S. T., Lam, N. S., and Li, M. S. (2011) Inhibition of Aggregation of Amyloid Peptides by Beta-Sheet Breaker Peptides and Their Binding Affinity, J. Phys. Chem. B 115, 74337446. (28) Zhang, T., Xu, W., Mu, Y., and Derreumaux, P. (2014) Atomic and Dynamic Insights into the Beneficial Effect of the 1,4-Naphthoquinon-2-yl-l-tryptophan Inhibitor on Alzheimer’s Aβ1–42 Dimer in Terms of Aggregation and Toxicity, ACS Chem. Neurosci. 5, 148-159. (29) Zhang, T., Zhang, J., Derreumaux, P., and Mu, Y. (2013) Molecular Mechanism of the Inhibition of EGCG on the Alzheimer Aβ1–42 Dimer, J. Phys. Chem. B 117, 3993-4002. (30) Zhao, L. N., Chiu, S.-W., Benoit, J., Chew, L. Y., and Mu, Y. (2012) The Effect of Curcumin on the Stability of Aβ Dimers, J. Phys. Chem. B 116, 7428-7435. (31) Wang, Q., Ning, L., Niu, Y., Liu, H., and Yao, X. (2014) Molecular Mechanism of the Inhibition and Remodeling of Human Islet Amyloid Polypeptide (hIAPP1–37) Oligomer by Resveratrol from Molecular Dynamics Simulation, J. Phys. Chem. B 119, 15-24. (32) Chebaro, Y., Jiang, P., Zang, T., Mu, Y., Nguyen, P. H., Mousseau, N., and Derreumaux, P. (2012) Structures of Aβ17–42 Trimers in Isolation and with Five Small-Molecule Drugs Using a Hierarchical Computational Procedure, J. Phys. Chem. B 116, 8412-8422. (33) Wang, Q., Yu, X., Patal, K., Hu, R., Chuang, S., Zhang, G., and Zheng, J. (2013) Tanshinones inhibit amyloid aggregation by amyloid-β peptide, disaggregate amyloid fibrils, and protect cultured cells, ACS Chem. Neurosci. 4, 1004-1015. (34) Pardridge, W. M. (2009) Alzheimer’s disease drug development and the problem of the blood−brain barrier., Alzheimer's Dementia 5, 427-432.

29 Environment ACS Paragon Plus

ACS Chemical Neuroscience

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 30 of 33

(35) Nguyen, P. H., Sterpone, F., Pouplana, R., Derreumaux, P., and Campanera, J. M. (2016) Dimerization Mechanism of Alzheimer Aβ40 Peptides: The High Content of Intrapeptide-Stabilized Conformations in A2V and A2T Heterozygous Dimers Retards Amyloid Fibril Formation, J. Phys. Chem. B 120, 12111-12126. (36) Nguyen, P. H., Sterpone, F., Campanera, J. M., Nasica-Labouze, J., and Derreumaux, P. (2016) Impact of the A2V Mutation on the Heterozygous and Homozygous Aβ1–40 Dimer Structures from Atomistic Simulations, ACS Chem. Neurosci. 7, 823-832. (37) Dong, M., Paul, T. J., Hoffmann, Z., Chan, K., Hu, D., Ai, H., and Prabhakar, R. (2016) Structural and Material Properties of Amyloid Aβ40/42 Fibrils, ChemPhysChem 17, 2558-2566. (38) Murugan, N. A., Halldin, C., Nordberg, A., Långström, B., and Ågren, H. (2016) The Culprit Is in the Cave: The Core Sites Explain the Binding Profiles of Amyloid-Specific Tracers, J. Phys. Chem. Lett. 7, 3313-3321. (39) Liu, W., Zhou, J., Geng, G., Shi, Q., Sauriol, F., and Wu, J. H. (2012) Antiandrogenic, Maspin Induction, and Antiprostate Cancer Activities of Tanshinone IIA and Its Novel Derivatives with Modification in Ring A, J. Med. Chem. 55, 971-975. (40) Olubiyi, O. O., and Strodel, B. (2012) Structures of the Amyloid β-Peptides Aβ1–40 and Aβ1–42 as Influenced by pH and a d-Peptide, J. Phys. Chem. B 116, 3280-3291. (41) Urbanc, B., Cruz, L., Yun, S., Buldyrev, S. V., Bitan, G., Teplow, D. B., and Stanley, H. E. (2004) In silico study of amyloid β-protein folding and oligomerization, Proc. Nat. Acad. Sci. U.S.A. 101, 17345-17350. (42) Yun, S., Urbanc, B., Cruz, L., Bitan, G., Teplow, D. B., and Stanley, H. E. (2007) Role of Electrostatic Interactions in Amyloid β-Protein (Aβ) Oligomer Formation: A Discrete Molecular Dynamics Study, Biophys. J . 92, 4064-4077. (43) Ngo, S. T., and Li, M. S. (2012) Curcumin Binds to Aβ1–40 Peptides and Fibrils Stronger Than Ibuprofen and Naproxen, J. Phys. Chem. B 116, 10165-10175. (44) Cairo, C. W., Strzelec, A., Murphy, R. M., and Kiessling, L. L. (2002) Affinity-Based Inhibition of β-Amyloid Toxicity, Biochemistry 41, 8620-8629. (45) Dong, M., Li, H., Hu, D., Zhao, W., Zhu, X., and Ai, H. (2016) Molecular Dynamics Study on the Inhibition Mechanisms of Drugs CQ1–3 for Alzheimer Amyloid-β40 Aggregation Induced by Cu2+, ACS Chem. Neurosci. 7, 599-614. (46) Periole, X., Rampioni, A., Vendruscolo, M., and Mark, A. E. (2009) Factors That Affect the Degree of Twist in β-Sheet Structures: A Molecular Dynamics Simulation Study of a Cross-β Filament of the GNNQQNY Peptide, J. Phys. Chem. B 113, 1728-1737. (47) Miller, Y., Ma, B., and Nussinov, R. (2009) Polymorphism of Alzheimer's Aβ17-42 (p3) Oligomers: The Importance of the Turn Location and Its Conformation, Biophys. J . 97, 1168-1177. (48) Hoshino, M., Katou, H., Hagihara, Y., Hasegawa, K., Naiki, H., and Goto, Y. (2002) Mapping the core of the beta(2)-microglobulin amyloid fibril by H/D exchange, Nat. Struct. Biology 9, 332-336. (49) Han, W., and Schulten, K. (2014) Fibril Elongation by Aβ17–42: Kinetic Network Analysis of Hybrid-Resolution Molecular Dynamics Simulations, J. Am. Chem. Soc. 136, 12450-12460. (50) Gurry, T., and Stultz, C. M. (2014) Mechanism of Amyloid-β Fibril Elongation, Biochemistry 53, 6981-6991. (51) Esler, W. P., Stimson, E. R., Jennings, J. M., Vinters, H. V., Ghilardi, J. R., Lee, J. P., Mantyh, P. W., and Maggio, J. E. (2000) Alzheimer's Disease Amyloid Propagation by a Template-Dependent Dock-Lock Mechanism, Biochemistry 39, 6288-6295. (52) Lee, J., Kwon, I., Jang, S. S., and Cho, A. E. (2016) Investigation of the effect of erythrosine B on amyloid beta peptide using molecular modeling, J. Mol. Model. 22, 1-9.

30 Environment ACS Paragon Plus

Page 31 of 33

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

ACS Chemical Neuroscience

(53) Diomede, L., Romeo, M., Cagnotto, A., Rossi, A., Beeg, M., Stravalaci, M., Tagliavini, F., Di Fede, G., Gobbi, M., and Salmona, M. (2016) The new β amyloid-derived peptide Aβ1–6A2V-TAT(D) prevents Aβ oligomer formation and protects transgenic C. elegans from Aβ toxicity, Neurobiol. Dis. 88, 75-84. (54) Miller, Y., Ma, B., and Nussinov, R. (2010) Zinc ions promote Alzheimer Aβ aggregation via population shift of polymorphic states, Proc. Nat. Acad. Sci. U.S.A. , 21, 9490-9495. (55) Savelieff, M. G., DeToma, A. S., Derrick, J. S., and Lim, M. H. (2014) The Ongoing Search for Small Molecules to Study Metal-Associated Amyloid-β Species in Alzheimer’s Disease, Acc. Chem. Res. 47, 2475-2482. (56) Pettersen, E. F., Goddard, T. D., and Huang, C. C. (2004) UCSF Chimera -- a visualization system for exploratory research and analysis, J. Comput. Chem. 25, 1605-1612. (57) Frisch, M. J., Trucks, G. W., Schlegel, H. B., Scuseria, G. E., Robb, M. A., Cheeseman, J. R., Scalmani, G., Barone, V., Mennucci, B., Petersson, G. A., Nakatsuji, H., Caricato, M., Li, X., Hratchian, H. P., Izmaylov, A. F., Bloino, J., Zheng, G., Sonnenberg, J. L., Hada, M., Ehara, M., Toyota, K., Fukuda, R., Hasegawa, J., Ishida, M., Nakajima, T., Honda, Y., Kitao, O., Nakai, H., Vreven, T., Montgomery Jr., J. A., Peralta, J. E., Ogliaro, F., Bearpark, M. J., Heyd, J., Brothers, E. N., Kudin, K. N., Staroverov, V. N., Kobayashi, R., Normand, J., Raghavachari, K., Rendell, A. P., Burant, J. C., Iyengar, S. S., Tomasi, J., Cossi, M., Rega, N., Millam, N. J., Klene, M., Knox, J. E., Cross, J. B., Bakken, V., Adamo, C., Jaramillo, J., Gomperts, R., Stratmann, R. E., Yazyev, O., Austin, A. J., Cammi, R., Pomelli, C., Ochterski, J. W., Martin, R. L., Morokuma, K., Zakrzewski, V. G., Voth, G. A., Salvador, P., Dannenberg, J. J., Dapprich, S., Daniels, A. D., Farkas, Ö., Foresman, J. B., Ortiz, J. V., Cioslowski, J., and Fox, D. J. (2009) Gaussian 09, Gaussian, Inc., Wallingford, CT, USA. (58) Aalten, A. W. S. a. D. M. F. v. (2004) PRODRG: a tool for high-throughput crystallography of protein-ligand complexes, Acta Crystallogr 60, 1355-1363. (59) Van Der Spoel, D., Lindahl, E., Hess, B., Groenhof, G., Mark, A. E., and Berendsen, H. J. (2005) GROMACS: fast, flexible, and free, J. Comput. Chem. 26, 1701-1718. (60) Oostenbrink, C. V., A.; Mark, A. E.; van Gunsteren, W. F. (2004) A biomolecular force field based on the free enthalpy of hydration and solvation: The GROMOS force-field parameter sets 53A5 and 53A6, J.Comput. Chem. 25, 1656-1676. (61) Adler, J., Scheidt, H. A., Kruger, M., Thomas, L., and Huster, D. (2014) Local interactions influence the fibrillation kinetics, structure and dynamics of A[small beta](1-40) but leave the general fibril structure unchanged, Phys. Chem. Chem. Phys. 16, 7461-7471. (62) Tarus, B., Nguyen, P. H., Berthoumieu, O., Faller, P., Doig, A. J., and Derreumaux, P. (2015) Molecular structure of the NQTrp inhibitor with the Alzheimer Aβ1-28 monomer, Eur. J. Med. Chem. 91, 43-50. (63) Wu, C., Bowers, M. T., and Shea, J.-E. (2011) On the origin of the stronger binding of PIB over thioflavin T to protofibrils of the Alzheimer amyloid-β peptide: a molecular dynamics study, Biophys. J . 100, 1316-1324. (64) Darden, T. A., York, D., and Pedersen, L. (1993) Particle mesh Ewald: an N•log(N) method for Ewald sums in large systems, J. Chem. Phys 18, 1463-1472. (65) Hess, B., Bekker, H., Berendsen, H. J. C., and Fraaije, J. G. E. M. (1997) LINCS: a linear constraint solver for molecular simulations, J.Comput. Chem. 18, 1463-1472. (66) Miyamoto, S., and Kollman, P. A. (1992) Settle: An analytical version of the SHAKE and RATTLE algorithm for rigid water models, J. Comput. Chem. 13, 952-962.

31 Environment ACS Paragon Plus

ACS Chemical Neuroscience

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 32 of 33

(67) Bussi, G., Donadio, D., and Parrinello, M. (2007) Canonical sampling through velocity rescaling, J. Chem. Phys 126, 2007. (68) Berendsen, H. J. C., Postma, J. P. M., van Gunsteren, W. F., DiNola, A., and Haak, J. R. (1984) Molecular dynamics with coupling to an external bath, J. Chem. Phys. 81, 3684-3690. (69) Humphrey, W., Dalke, A., and Schulten, K. (1996) VMD: visual molecular dynamics, J Mol Graph 14, 33-38, 27-38. (70) Masman, M. F., Eisel, U. L., Csizmadia, I. G., Penke, B., Enriz, R. D., Marrink, S. J., and Luiten, P. G. (2009) In silico study of full-length amyloid beta 1-42 tri- and penta-oligomers in solution, J. Phys. Chem. B 113, 11710-11719. (71) Sticht, H., Bayer, P., Willbold, D., Dames, S., Hilbich, C., Beyreuther, K., Frank, R. W., and Rösch, P. (1995) Structure of Amyloid A4-(1–40)-Peptide of Alzheimer's Disease, Eur. J.Biochem. 233, 293-298. (72) Coles, M., Bicknell, W., Watson, A. A., Fairlie, D. P., and Craik, D. J. (1998) Solution Structure of Amyloid β-Peptide(1−40) in a Water−Micelle Environment. Is the Membrane-Spanning Domain Where We Think It Is?, Biochemistry 37, 11064-11077. (73) Kabsch, W., and Sander, C. (1983) Dictionary of protein secondary structure: Pattern recognition of hydrogen-bonded and geometrical features., Biopolymers 22, 2577-2637. (74) Man, H. V., Nguyen, P. H., Derreumaux, P., and Mai, S. L. (2014) Effect of the English Familial Disease Mutation (H6R) on the Monomers and Dimers of A beta 40 and A beta 42, ACS Chem. Neurosci. 5, 646−657. (75) Baker, N. A., Sept, D., Joseph, S., Holst, M. J., and McCammon, J. A. (2001) Electrostatics of nanosystems: Application to microtubules and the ribosome, Proc. Natl. Acad. 98, 10037-10041. (76) Kumari, R., Kumar, R., and Lynn, A. (2014) g_mmpbsa—A GROMACS Tool for HighThroughput MM-PBSA Calculations, J. Chem. Inf. Model. 54, 1951-1962.

32 Environment ACS Paragon Plus

Page 33 of 33

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

ACS Chemical Neuroscience

For Table of Contents Use Only

N-terminus Binding Preference for either Tanshinone or Analogue in both Inhibition of Amyloid Aggregation and Disaggregation of Preformed Amyloid Fibrils---Toward Introducing A kind of Novel Anti-Alzheimer Compounds Mingyan Dong†, Wei Zhao†, Dingkun Hu†, Hongqi Ai*,Baotao Kang* Shandong Provincial Key Laboratory of Fluorine Chemistry and Chemical Materials, School of Chemistry and Chemical Engineering, University of Jinan, Jinan, 250022, China Corresponding author: Hongqi Ai E-mail: [email protected] Corresponding author: Baotao Kang E-mail: [email protected]

These authors equally contributed to this work

Tanshinone derivative TS0 can disaggregate the fibril-like Aβ42 pentamer by binding at the charged and polar residues of Aβ42 N-terminal region.

+ten TS0s

33 Environment ACS Paragon Plus