Amyloid Fibril-Induced Structural and Spectral Modifications in the

Dec 11, 2012 - Laetitia Lemoine , Per-Göran Gillberg , Marie Svedberg , Vladimir Stepanov ... Erik Donovan Hedegård , Nanna H. List , Hans Jørgen A...
0 downloads 0 Views 4MB Size
Download PDF
Letter pubs.acs.org/JPCL

Amyloid Fibril-Induced Structural and Spectral Modifications in the Thioflavin‑T Optical Probe N. Arul Murugan,*,† Jógvan Magnus Haugaard Olsen,‡ Jacob Kongsted,‡ Zilvinas Rinkevicius,† Kestutis Aidas,§ and Hans Ågren† †

Division of Theoretical Chemistry and Biology, School of Biotechnology, Royal Institute of Technology, SE-10691 Stockholm, Sweden ‡ Department of Physics, Chemistry and Pharmacy, University of Southern Denmark, Campusvej 55, DK-5230 Odense M, Denmark § Department of General Physics and Spectroscopy, Faculty of Physics, Vilnius University, Sauletekio al. 9, LT-10222 Vilnius, Lithuania S Supporting Information *

ABSTRACT: Motivated by future possibilities to design target molecules for fibrils with diagnostic or therapeutic capability related to amyloidosis diseases, we investigate in this work the dielectric nature of amyloid fibril microenvironments in different binding sites using an optical probe, thioflavin-T (THT), which has been used extensively to stain such fibrils. We study the fibrilenvironment-induced structural and spectral changes of THT at each binding site and compare the results to the fibril-free situation in aqueous solution. All binding sites are found to show a similar effect with respect to the conformational changes of THT; in the presence of the fibril, its molecular geometry tends to become planarized. However, depending on the dielectric nature of the specific binding site, a red shift, blue shift, or no shift in the absorption spectra of THT is predicted. Interestingly, the experimentally measured red shift in the spectra is seen only when THT binds to one of the core or surface-binding sites. It is found that the dielectric nature of the microenvironment in the fibril is strongly nonhomogeneous. SECTION: Biophysical Chemistry and Biomolecules fluorescence properties would be important for fibril imaging in brains, and much medical research has been devoted to this. Such a development of optical probes from bench-to-bedside can be done experimentally but is generally highly timeconsuming. In that respect, it might prove advantageous to rely on multiscale modeling techniques that can provide the insight necessary to cut down the cost and time for the development of such probes. Understanding the dielectric nature of the microenvironments in fibrils and the nature of molecular probes interacting with fibrils could provide us knowledge to design useful guess precursor molecules that can eventually evolve as “fibrilindicator” molecules.3,4,8 To design a better probe, knowledge about different possible binding sites as well as the dielectric nature of the microenvironments of these binding sites is needed because the optical properties of the indicator molecules are dictated by the dielectric nature of the microenvironments. The present study aims to contribute to the subject of understanding the dielectric nature of fibrils and

he self-aggregation of amyloid fibrils to form waterinsoluble plaques in human tissue and organs has been attributed to a number of so-called “conformational diseases” such as Alzheimer’s and Parkinson’s diseases.1 Early stage detection of these fibrils and prevention of the fibril growth would contribute significantly to the treatment of these diseases. Many organic molecules have been proposed and tested for their binding affinity to fibrils, some of them having a tendency to inhibit the fibril growth.2 In addition, a number of these molecules show specific changes in their absorption spectra or enhancement in their fluorescence intensity that leads to the possibility to use them as potential optical probes to identify the presence of fibrils in human tissue or in test samples.3−5 Thioflavin-T (THT) and congo red are such molecules well-known as optical and fluorescent probes for fibrils,3−8 and there are by now numerous reports available that discuss staining amyloid fibrils using these molecular probes. However, these molecules are charged and have small lipophilicity, which prevents them from crossing the blood brain barrier, something that obviously limits their use as probe molecules for fibrils in the brain. The structural tuning of these molecules to increase the lipophilicity but at the same time retaning their fibril-binding ability and fibril-specific optical and

T

© 2012 American Chemical Society

Received: November 14, 2012 Accepted: December 11, 2012 Published: December 11, 2012 70

dx.doi.org/10.1021/jz3018557 | J. Phys. Chem. Lett. 2013, 4, 70−77

The Journal of Physical Chemistry Letters

Letter

solvatochromic shift, is not as substantial as it has been reported in the case of other optical probes. For example, in the case of Reichardt’s dye, a solvatochromic shift as much as 9730 cm−1 has been reported from a nonpolar to a water solvent.18 Similarly, in the case of Michler’s ketone, a positive solvatochromic shift that amounts to 4240 cm−1 has been reported when the solvent polarity is switched. However, in the present case, a solvatochromic shift of only 690 cm −1 (associated with a red shift by 12 nm) is reported for a change in solvent from water to chloroform.16,17 Concerning the emission spectra, the position of the band remains the same independent of any change in the solvent polarity. However, the fluorescence intensity has a strong dependence on the solvent nature. The lowest fluorescence intensity has been registered in water solvent while the highest fluorescence intensity has been registered in glycerol solvent.16,17 The fluorescence intensity and the quantum yield have been reported to be strongly dependent on the solvent polarity as well as its viscosity and are associated with the conformational dynamics of the molecule in its excited state. The chargeseparated excited state is stabilized in its twisted conformation (with ϕ = 90°) while the ground state is more stable in a twisted form with ϕ = 35°.6,7 Here the angle, ϕ defines the relative orientation of the two aromatic moieties. When THT is excited it can either emit from the twisted excited state or undergo torsional relaxation to a near-planar conformation, which is a nonradiative relaxation pathway. However, such torsional relaxation is limited in highly viscous solvents like glycerol, and the nonradiative relaxation pathway is hindered, which eventually leads to an increase in fluorescence intensity. There exist some semiempirical calculations that report the ground- and excited-state potential energy surface as a function of the torsional angle between the dimethylaniline ring to the N-methylated benzothiazole fragment, which suggests that the absorption spectra of THT will have a strong dependence on this torsional angle.16,17 Similar to what has been reported in the case of nonpolar solvents, a red shift in the absorption spectrum has been observed when THT binds to amyloid fibrils.16,17 In addition, there is a manifold increase in its fluorescence intensity. However, only little is known about the binding mode of THT in fibrils as well as the responsible factors for the shift in the absorption spectra and increase in the fluorescence intensity when compared with THT in aqueous solution. Here we will contribute with a detailed investigation on the microscopic level discussing the different THT binding sites of the fibrils as well as the fibril-environment-induced structural and spectral modifications resulting from the binding of THT at these sites. For this, we will rely on an integrated approach, thereby providing insight into the physical origin of the observed shifts in the THT absorption spectra. With the integrated approach, we model first the binding sites of THT in fibrils, followed by studies of the fibril-induced structural and spectral modifications in THT. The approach employs the tools in the given sequence: First, different binding sites of THT in the fibrils have been identified using molecular docking. For each of the THTs docked at different binding sites of the fibril, subsequent molecular dynamics (MD) and Car− Parrinello hybrid quantum mechanics/molecular mechanics (QM/MM) calculations have been carried out. The molecular dynamics simulations account for the long time scale THTinduced structural and conformational changes in the fibrils, whereas the Car−Parrinello QM/MM molecular dynamics approach19 accounts for the short-time-scale intramolecular

the nature of the interaction of a molecular probe, THT, with fibrils. We analyze different aspects of the THT interaction with fibrils and the fibril environment-dependent structural and optical properties in THT. Below, we describe some key literature facts related to the structure of the fibrils and the use of THT as a molecular probe for fibrils. To design target binding molecules for diagnostic approaches and fibrillation inhibitors for therapeutic approaches, it is important to have the 3D structure of the fibril available. Along this line, many experimental studies based on X-ray diffraction and NMR measurements have been reported that provide some information about the secondary structure of the fibrils.9 Recently, a high-resolution cryoelectron microscopy-based structure has been proposed for this fibril,10 and a 42 residue amyloid-β peptide has been found to be the major component. The residues 18−42 contribute to the β-strand-turn-β-strand motif, whereas the residues 1−17 form a flexible segment of the fibril.10 There is also a very recent report of crystal structure for the oligomer (particularly tetramer) of amyloid β peptide.11 THT is the most popular molecule that has been used to stain the deposit of amyloid fibrils in human tissue and body organs. For the molecular structure of THT, refer to Figure 1.

Figure 1. Molecular structure of THT.

The interaction of THT and its derivative molecules with fibrils has been studied in some detail experimentally and by molecular dynamics simulations.12−14 Multiple binding sites in the fibrils have been proposed by using fluorescent and radio-ligand binding assay titrations with the fibrils12 and subsequent positron emission tomography. These results have also been supported by independent molecular dynamics simulations that identified three binding sites of THT with the fibril. Two of the binding sites were identified on the surface of the fibril on either side of the grooves and another one at the ends of the β-sheet.14 The former two sites have been identified as high-capacity micromolar-affinity binding sites, whereas the latter one is associated with a low-capacity nanomolar-affinity binding site. A recent report based on docking and molecular dynamics also confirms these two binding sites, and in addition it recognizes two core binding sites.15 It is not yet clear which binding site is responsible for the experimentally reported optical property of THT specific for the fibril environment.16,17 Moreover, it is not clear what happens to the molecular structure and conformation of THT when it is bound to these different binding sites. There exist detailed experimental reports on the absorption and emission spectra of THT in aqueous or glycerol solutions as well as in other solvents and when it is bound to fibrils.16,17 The absorption maximum (λmax) of THT in water solvent is reported to be between 400 and 412 nm, which red shifts with increasing concentration of glycerol.16,17 In solvents of varying polarity, a general observation is that the maximum in the absorption spectrum is shifted toward shorter wavelengths when increasing the solvent polarity. However, the shift in the absorption spectra, which is generally referred to as a 71

dx.doi.org/10.1021/jz3018557 | J. Phys. Chem. Lett. 2013, 4, 70−77

The Journal of Physical Chemistry Letters

Letter

Figure 2. Initial configurations for THT in fibril simulations, as obtained from docking studies.

vibrations and fibril-environment-induced geometrical changes in the THT molecule. In the Car−Parrinello QM/MM MD, THT has been treated at the level of density functional theory while the fibrils, ions, and water molecules (the solvent) were included in the MM region. The effective Hamiltonian in the CPMD calculation includes the van der Waals and electrostatic interaction between the QM and MM subsystems and so accounts for the polarization of THT by the fibril and solvent environment. Finally, for various snapshots of THT (80 in number) in the fibrils obtained from the Car−Parrinello QM/MM MD trajectories, we have carried out QM/MM response calculations20 to compute the excitation energy of THT. We have employed a developmental version of Dalton software for this.21 Here again the THT was treated at the level of density functional theory with fibril, ions, and water solvent included in the MM region. In particular, we employed a Coulombattenuated B3LYP functional22 that is capable of describing excitations of charge-transfer-type. We have employed the Turbomole-TZVP basis sets.23 The reasons for employing such combination of density functional theory and basis sets to describe the excitations for the optical probes are described in detail in our previous works.24,25 In the QM/MM response calculations, the interaction between THT and the bioenvironment is described using an embedding technique where an electrostatic interaction between these two subsystems is accounted for.20 This set of property calculations will be referred to as MM-1. In addition to this, we have also carried out a set of calculations for THT alone without including the media which basically separates out the indirect contributions to the optical property from the environment. This set of calculations will be referred to as MM-0. In particular, MM-0 and MM-1 level calculations are carried out for 80 snapshots

extracted from the hybrid QM/MM MD. Overall, the computational approach employed for modeling the structure and optical property is similar to that we have adopted in the ref 24, which describes the modeling an optical probe, Nile red, in the cavity of the β-lacto globulin protein. A similar integrated approach has been also carried out for THT in water to model the structure and optical property of THT in aqueous solution. A detailed description of various techniques used in this study is given in the Supporting Information for this letter. Below we discuss different aspects of fibril-induced structural and spectral tuning in THT. Binding Sites of THT and Its Solvation Shell Structure in Fibrils. Until recently, the issue of multiple binding sites of the fibrils was not clearly understood. Previously, intrinsic fluorescence assay experiments suggested a single binding site in the fibrils for THT.12 However, recent microassay titration experiments propose multiple binding sites; in particular, three possible binding sites have been identified for THT in fibrils among which two have been reported to be micromolar affinity highcapacity binding sites, whereas the third has been reported to be a nanomolar affinity site.13 In addition, molecular dynamics simulations of fibrils with a number of THT and solvent molecules have identified THT localization in three distinct binding sites of the fibrils.14 Two of the sites were located on either side of the grooves of the fibrils, whereas in the third binding mode, the THT is stacked to the fibril in the fibrilgrowth direction. It is important to remember that all of these identified binding sites are surface binding sites. However, molecular dynamics simulations possess problems in identifying core binding sites because the time scale associated with the binding to these is significantly larger than that corresponding to surface binding sites. In addition, the binding into the corebinding sites requires some structural adjustment by the 72

dx.doi.org/10.1021/jz3018557 | J. Phys. Chem. Lett. 2013, 4, 70−77

The Journal of Physical Chemistry Letters

Letter

Figure 3. Solvation shell structure of THT in water and in different binding sites of fibrils. Shown are only for sites 2, 5, and 6. The solvent molecules up to the second coordination shell are shown. The definition of solvation shells is based on the probe-all-atom and solvent center of mass radial distribution function.

biostructure (here a fibril) and is usually associated with significant conformational changes around the binding site. Without such structural adjustment in the protein, the ligand molecule would experience steric repulsion, something that limits these sites to be recognized by MD simulations. However, these sites can be identified from molecular docking studies. As it is reported recently, a docking and moleculardynamics-based approach could recognize some of the core binding sites.15 Thus, the integrated approach we have employed starts with molecular docking as the first step and has the ability to recognize both core and surface binding sites. The blind docking employed in our study recognizes altogether six different unique binding sites on the fibrils, out of which three are core binding sites and the other three are surface binding sites. All binding sites and the mode of binding of THT in these sites are shown in Figure 2. As can be seen, the modes 1, 3, and 4 correspond to surface binding sites, whereas the modes 2, 5, and 6 correspond to core binding sites. Moreover, the modes 1 and 6 might show therapeutic value due to the tendency to inhibit fibril growth because the addition of a new peptide to this pentamer to yield a hexamer will be perturbed to some extent by the THT molecule. So, increasing the binding energy of THT in these two binding sites has therapeutic values for amyloidosis diseases. In the remaining four sites, the THT has probably diagnostic values. Another important point to consider for the binding modes is that in modes 1 and 2 the THT is bound to the fibril in the perpendicular direction of the fibril growth, whereas in all other cases it binds to the fibril in the fibril growth direction. It is also worthwhile to mention that the population of binding sites 2, 3, 4, and 5 increases with the fibril size along the fibril growth direction. However, this does not happen for the binding sites 1 and 6. We have also analyzed the solvation shell structure of THT in different binding sites of the fibril. We have defined the solvation shell structure for the probe based on the probe-allatoms and solvent center of the mass radial distribution function. Snapshots of solvation shell structures (up to the

second coordination shell) of THT in water solvent and in a few binding sites (sites 2, 5, and 6) are shown in Figure 3. As it can be seen in this Figure, the solvation shell structure of THT is substantially modified by the fibril environment. The exposure of THT to the water environment is controlled by the specific nature of the binding site. We have also computed the average number of solvent molecules in the solvation shell for THT in water and at the different binding sites; the values are presented in Table 1. As can be seen, the number of solvent Table 1. Average Number of Water Molecules in the First Solvation Shell of THT system THT THT THT THT THT THT THT

in in in in in in in

water site 1 site 2 site 3 site 4 site 5 site 6

NH2O 114 54 38 51 67 60 28

molecules reduces to 50% in the case of sites 1, 3, 4, and 5, whereas in the case of sites 2 and 6 it is reduced to 25% when compared with the value in water solvent. Molecular Structure of THT in Aqueous Solution and in Fibrils. The THT molecular geometry includes two aromatic moieties, namely, benzothiazole and dimethylaminobenzene connected through a central C−C bond. The torsional rotation around the C−C bond makes it possible for this molecule to exist in numerous conformations that differ with respect to the relative orientation of the aromatic groups. As in polyphenyl-like systems, there are two competing interactions that determine the molecular conformation in the gas phase: The conjugation stabilizes the planar conformation while the steric repulsion between the N-methyl group with ortho-hydrogens of the phenyl group stabilizes a twisted conformation. In addition, in 73

dx.doi.org/10.1021/jz3018557 | J. Phys. Chem. Lett. 2013, 4, 70−77

The Journal of Physical Chemistry Letters

Letter

also presented along with the average values in Table 1. The observed difference in the average angles ϕ1 and ϕ2 has to be attributed to the presence of the methyl group in one side of the aromatic group, which contributes with a larger repulsive contribution when the molecule tends to become planar. This is the reason for the average ϕ1 always being larger than ϕ2. Large values for the angles, ϕ1 and ϕ2, for THT are seen in the case of water solvent when compared with the fibril-bound condition, which suggests that the molecular geometry is more twisted in the water solvent. Interestingly, at the different binding sites of the fibril, the angle ϕ1 can vary between 19.5 and 31.0°, whereas the angle ϕ2 can vary between 17.6 and 31.4°. This clearly demonstrates the heterogeneity in the microenvironment of the fibrils and reveals the complexity in characterizing the dielectric nature of the fibrils. A careful look into the average values of ϕ1 and ϕ2 suggests that THT is more planar when it is bound to sites 1, 2, and 3, whereas it has more twisted geometry at the binding sites 4, 5, and 6. The angle ϕ3 for THT also has nonzero values between 10.9 and 13.3° in either water solvent or under fibril-bound conditions, which suggests that the NMe2 group also has a near-planar geometry with respect to the phenyl group, and this conformational degree of freedom is not sensitive to the dielectric nature of the environment. The conformational equilibrium between the twisted and planar conformers (characterized by angles ϕ1 and ϕ2) is to a large degree controlled by the dipole moment of the conformers, and a more twisted conformer is stabilized in the polar water solvent. The conformational equilibrium shifting toward a near planar conformer under the fibril-bound condition suggests a hydrophobic-like microenvironment in the fibril. Overall, even though we do not observe any significant changes in the average bond lengths along the conjugation pathway, we report substantial geometrical changes in THT with respect to the conformational structure when it is bound to the fibril. Optical Properties of THT in Aqueous Solution and Fibrils. In the last subsection, we discuss in some detail the fibrilenvironment-induced structural changes in THT. Such mediainduced geometrical changes in the molecular probes significantly contribute to a number of properties, for example, the polarizability, hyperpolarizability, two photon absorption cross sections, and solvatochromic shift.27 We have previously reported such media-induced contributions to optical properties in a number of molecular probes such as phenol blue, nile red, quinolinium betaine, and stilbazolium merocyanine.24−26 In particular, in the case of merocyanines, even the sign (i.e., either a red or blue shift) of the solvatochromic shift is dictated by such structural modifications of these dye molecules.28 Overall, the contributions to the molecular property in question due to media effects are usually discussed in terms of indirect and direct effects. The former one refers to the environmentinduced structural changes of the molecular probes, whereas the latter corresponds to direct solute−media interactions. The integrated approach we have employed allows us to dissect these two contributions. Results based on model MM-0 include the contributions only from indirect media effects, whereas the results based on the model MM-1 include the total contributions from both direct and indirect media effect. However, the results from MM-0 includes at the same time all environment-induced structural changes, and it is not possible to separate out contributions from the individual conformational degree of freedoms. In particular, it would be interesting to explore how the changes in the dihedral angles ϕ1 and ϕ3

solvents or in other media, the dipole moment of the molecular conformation also plays a role. A conformer with large dipole moment, that is, a twisted conformer, is stabilized in polar solvents due to dipole−dipole interaction. A similar observation has been reported by us in the case of o-betaine, where the gasphase molecular geometry has a near-planar geometry, whereas in water solvent the conformer was twisted.25 Overall, the conformational geometry of THT-like molecules in a particular media is dictated by three dominant interactions, namely, (i) conjugation, (ii) steric repulsion, and (iii) solute−solvent interactions that are dominated by electrostatic interactions. From previous studies of molecular probes, it is known that the media effects on the molecular geometry can be substantial.24−26 Obviously, such media-induced geometrical changes in the molecular probes might also contribute significantly to various molecular properties, and this will be discussed in relevance to THT in the following section. Because the THT is a molecule of relatively large size, there exist many internal degrees of freedom for this molecule. However, based on the topic of our interest, only a few of these degrees of freedom are relevant for discussion. Here we will consider the bond length changes along the conjugation pathway as well as the angle between the two aromatic moieties and the angle between the dimethyl amino and phenyl groups. Usually, the electronic structure along the conjugation pathway is very sensitive to the dielectric nature of the environment, which drives changes in the molecular geometry, and this in turn controls various properties including the optical ones. We have computed the average bond lengths along the conjugation pathway for THT in water solvent and at the different binding sites of the fibril. Interestingly, we have not observed any significant difference in the average bond lengths of THT in aqueous solution or under the fibril-bound conditions, whereas we observe significant differences with respect to the conformational geometry of THT in these two environments. We have characterized the conformational geometry of THT using three different dihedral angles, ϕ1, ϕ2, and ϕ3. The first two angles define the relative orientation of the two aromatic moieties, whereas the third one defines the relative orientation of the NMe2 group with respect to the phenyl aromatic group. The dihedral angle, ϕ1, is defined by the atoms 1, 2, 3, and 4 (for labeling of atoms, see Figure 1), and the angles ϕ2 and ϕ3 are defined between the atoms 5, 2, 3, 6 and 11, 8, 9, 12 (or 7, 8, 9, 10). The average dihedral angles for ϕ1, ϕ2, and ϕ3 for THT in water and at the different binding sites of the fibrils are given in Table 2. The averaging has been done over all configurations of THT from hybrid QM/MM MD trajectory. The standard errors associated with these dihedral angles are Table 2. Average Dihedral Angles, ϕ1, ϕ2, and ϕ3, of THT in Water Solvent and in Various Binding Sites of Fibrils along with Their Standard Errors angle (deg) system THT THT THT THT THT THT THT

in in in in in in in

water site 1 site 2 site 3 site 4 site 5 site 6

ϕ1

ϕ2

ϕ3

33.4(0.4) 19.5(0.3) 19.7(0.3) 24.0(0.4) 29.9(0.4) 31.0(0.4) 29.5(0.4)

31.7(0.4) 17.8(0.3) 17.6(0.3) 23.2(0.4) 26.9(0.4) 28.9(0.4) 31.4(0.4)

13.1(0.3) 12.5(0.3) 11.7(0.3) 10.9(0.3) 11.1(0.3) 13.3(0.4) 12.0(0.4) 74

dx.doi.org/10.1021/jz3018557 | J. Phys. Chem. Lett. 2013, 4, 70−77

The Journal of Physical Chemistry Letters

Letter

contributions from the environment are not very significant. However, the MM-1-based results for THT at the different binding sites yield a red, blue or no significant shift when compared with the corresponding in-water calculations. Interestingly, experimentally only a red shift in the absorption spectra of THT is seen under the fibril-bound condition when compared with the in-water observations. Our current study sheds light onto the complexity in interpreting the experimental absorption spectra of some molecules when bound to proteins that have multiple binding pockets. In the present case, a red shift in the absorption spectra is seen for THT at the binding sites 2 and 3, whereas in site 5 a blue shift is seen. It is worthwhile to recollect that the conformational geometry of THT was near-planar in former two sites than in the latter site. In the sites, 1, 4, and 6, no significant shift in λmax is observed. We speculate that the experimentally observed red shift can be explained by assuming that the residence time of THT is larger at the binding sites 2 and 3 than in the remaining other binding sites. Moreover, the predicted red shifts are underestimated at both red-shifting sites, and, particularly at site 2, the fibrilinduced shift in the absorption maximum is predicted to be 18 nm compared with 38 nm, as reported experimentally.16,17 One potential reason for this discrepancy between experimental and calculated absorption maxima could be attributed to the force field used for the fibril in the QM/MM response calculation (i.e., the standard PARM99 force field). To investigate if this could be the possible explanation for the differences observed between the calculated and experimental results for the red shifts, we have here constructed a new polarizable and more accurate force-field for the fibril based on quantum mechanical calculations. The procedure is discussed in detail in the computational section given in the Supporting Information. Because of the computational expense associated with calculations employing a polarizable force field, we have carried out the excitation energy calculations for only 25 configurations taken from the Car−Parrinello QM/MM MD simulation for the binding of THT at the site 2. As expected, this indeed gives rise to some improvement in the predicted absorption maximum, leading to an increase in the predicted shift of 7 nm; that is, the total red shift is predicted to be 25 nm. However, even if this increases the agreement between the predicted and experimental shift, the change is still small to fully explain the deviations between the predicted and observed red shift. Another reason for the observed discrepancy between the predicted and measured shifts in the excitation energy could be that we in our predictions account only for the electrostatic interactions caused by the fibrils on the THT probe. In a hydrophobic environment, probe−environment dispersion interactions may play a significant role, and such interactions have not been accounted for in our calculations of the excitation energies. In fact, neglect of such interactions may in nonpolar environments lead to underestimated predictions for the excitation energies, as previously observed by us.24,26 Thus, we may attribute the underestimation in the predicted excitation energies as being mainly due to missing dispersion interactions in the model calculations. Finally, we have also characterized the nature of the relevant excitation by plotting the HOMO and LUMO molecular orbitals involved in the lowest energy excitation. From this, the electron density appears to be localized over both of the aromatic moieties, which suggest that the excitation is mostly a π ⇒ π* transition, with a small charge transfer from the dimethyl amino group to

individually contribute to the excitation energy. For this, we have carried out “static” calculations where we have optimized the molecular geometry of THT for varying ϕ1 and ϕ3 between 0 and 90° in increments of 10°. For these geometry optimizations, the Gaussian09 software is used.29 On the basis of these optimized geometries, we have further calculated the excitation energy using time-dependent density functional theory at the level of B3LYP/6-31+G(d,p). The results are shown in Figure 4. Interestingly, as it can be seen, the

Figure 4. Dependence of absorption maximum on variation of the dihedral angles ϕ1 and ϕ3 between the range 0−90°.

planarization of the conformation with respect to ϕ1 and ϕ3 results in opposite effects. The former one yields a red shift in the absorption spectra, whereas the latter one yields a blue shift. A similar effect is also seen when the molecule becomes more twisted with respect to the ground-state minimum energy structure. However, the changes in absorption maximum due to the angle ϕ3 are not very steep when compared with the case of ϕ1. As discussed above, there are changes with respect to both of these dihedral angles in THT in the case of water solvent and at the different binding sites of the fibrils. Because these two angles give rise to opposing effects in the calculated absorption maximum, it would be very difficult to predict the outcome. This is exactly what is seen in the case of the average excitation energy of THT in the different binding pockets of the fibrils when compared with the in-water results − we refer to the values of λmax of THT given in Table 3. The standard error associated with the values presented in Table is between 1 and 2 nm. Moreover, the results from the MM-0 model for THT in different binding sites of the fibril differ from the in-water MM0 result by only 5 nm, and this suggests that indirect Table 3. Average Excitation Energies (in nm) in Water Solvent and Fibril-Induced Spectral Shifts for THTa method system THT THT THT THT THT THT THT

in in in in in in in

water site 1 site 2 site 3 site 4 site 5 site 6

QM/MM-0

QM/MM-1

shift

384 387 382 388 379 381 382

382 380 400 392 374 360 374

NS red red NS blue NS

a

Shifts are abbreviated as red shift (red), blue shift (blue), and no shift (NS). 75

dx.doi.org/10.1021/jz3018557 | J. Phys. Chem. Lett. 2013, 4, 70−77

The Journal of Physical Chemistry Letters

Letter

(8) Frid, P.; Anisimov, S. V.; Popovic, N. Congo Red and Protein Aggregation in Neurodegenerative Diseases. Brain Res. Rev. 2007, 53, 135−160. (9) Tycko, R. Progress Towards a Molecular-Level Structural Understanding of Amyloid Fibrils. Curr. Opin. Struct. Biol. 2004, 14, 96−103. (10) Luhrs, T.; Ritter, C.; Adrian, M.; Riek-Loher, D.; Bohrmann, B.; Dobeli, H.; Schubert, D.; Riek, R. 3D Structure of Alzheimer’s Amyloid-β(1−42) Fibrils. Proc. Natl. Acad. Sci. U.S.A. 2005, 102, 17342−17347. (11) Streltsov, V. A.; Varghese, J. N.; Masters, C. L.; Nuttall, S. D. Crystal Structure of the Amyloid-β p3 Fragment Provides a Model for Oligomer Formation in Alzheimer’s Disease. J. Neurosci. 2011, 31, 1419−1426. (12) Krebs, M. R. H.; Bromley, E. H. C.; Donald, A. M. The Binding of Thioflavin-T to Amyloid Fibrils: Localisation and Implications. J. Struct. Biol. 2005, 149, 30−37. (13) Lockhart, A.; Ye, L.; Judd, D. B.; Merritt, A. T.; Lowe, P. N.; Morgenstern, J. L.; Hong, G.; Gee, A. D.; Brown, J. Evidence for the Presence of Three Distinct Binding Sites for the Thioflavin T Class of Alzheimer’s Disease PET Imaging Agents on Beta-Amyloid Peptide Fibrils. J. Biol. Chem. 2005, 280, 7677−7684. (14) Wu, C.; Wang, Z.; Lei, H.; Duan, Y.; Bowers, M. T.; Shea, J. E. The Binding of Thioflavin-T and Its Neutral Analog BTA-1 to Protofibrils of the Alzheimer A16−22 Peptide Probed by Molecular Dynamics Simulations. J. Mol. Biol. 2008, 384, 718−729. (15) Rodríguez-Rodríguez, C.; Rimola, R.; Rodríguez-Santiago, L.; Ugliengo, P.; Á lvarez-Larena, Á .; Gutiérrez-de-Terán, H.; Sodupe, M.; González-Duarte, P. Crystal Structure of Thioflavin-T and Its Binding to Amyloid Fibrils: Insights at the Molecular Level. Chem. Commun. 2010, 46, 1156−1158. (16) Sulatskaya, A. I.; Maskevich, A. A.; Kuznetsova, I. M.; Uversky, V. N.; Turoverov, K. K. Fluorescence Quantum Yield of Thioflavin T in Rigid Isotropic Solution and Incorporated Into the Amyloid Fibrils. PLoS One 2010, 5, e15385. (17) Maskevich, A. A.; Stsiapura, V. I.; Kuzmitsky, V. A.; Kuznetsova, I. M.; Povarova, O. I.; Uversky, V. N.; Turoverov, K. K. Spectral Properties of Thioflavin T in Solvents with Different Dielectric Properties and in a Fibril-Incorporated Form. J. Proteome Res. 2007, 6, 1392−1401. (18) Reichardt, C. Solvatochromic Dyes as Solvent Polarity Indicators. Chem. Rev. 1994, 94, 2319−2358. (19) Laio, A.; VandeVondele, J.; Rothlisberger, U. A Hamiltonian Electrostatic Coupling Scheme for Hybrid Car-Parrinello Molecular Dynamics Simulations. J. Chem. Phys. 2002, 116, 6941. (20) Olsen, J. M. H.; Aidas, K.; Kongsted, J. Excited States in Solution Through Polarizable Embedding. J. Chem. Theory Comput. 2010, 6, 3721−3734. (21) DALTON, A Molecular Electronic Structure Program, 2011. http://daltonprogram.org/ (accessed December 14, 2012). (22) Yanai, T.; Tew, D. P.; Handy, N. C. A New Hybrid ExchangeCorrelation Functional Using the Coulomb-Attenuating Method (CAM-B3LYP). Chem. Phys. Lett. 2004, 393, 51−57. (23) Schäfer, A.; Horn, H.; Ahlrichs, R. Fully Optimized Contracted Gaussian Basis Sets for Atoms Li to Kr. J. Chem. Phys. 1992, 97, 2571. (24) Murugan, N. A.; Kongsted, J.; Rinkevicius, Z.; Ågren, H. Color Modeling of Protein Optical Probes. Phys. Chem. Chem. Phys. 2012, 14, 1107−1112. (25) Murugan, N. A.; Ågren, H. Role of Dynamic Flexibility in Computing Solvatochromic Properties of Dye-Solvent Systems: Ortho-Betaine in Water. J. Phys. Chem. A 2009, 113, 2572−2577. (26) Murugan, N. A.; Kongsted, J.; Rinkevicius, Z.; Ågren, H. Modeling the Structure and Absorption Spectra of Stilbazolium Merocyanine in Polar and Non-polar Solvents Using Hybrid QM/MM Techniques. J. Phys. Chem. B 2010, 114, 13349−13357. (27) Gorman, C. B.; Marder, S. R. An Investigation of the Interrelationships Between Linear and Nonlinear Polarizabilities and Bond-Length Alternation in Conjugated Organic Molecules. Proc. Natl. Acad. Sci. U.S.A. 1993, 90, 11297−11301.

the benzothiazole group. The HOMO and LUMO molecular orbitals involved in this excitation are shown in Figure 1S of the Supporting Information. Using an integrated approach, we have characterized the heterogeneous dielectric nature of the microenvironments created by fibrils. We first identified the important binding sites for the THT probe molecule in the fibrils and characterized them as core binding or surface-binding sites. Under the fibril bound conditions, the THT molecule tends to become more planar when compared with the case of a water solution. We found that the spectral properties of THT at different binding sites lead to either a red or a blue shift when compared with THT in aqueous solution. This suggests that the optical property is dictated by the direct media effect characteristic for a specific binding pocket rather than the indirect media effect arising due to the fibril-induced geometrical changes in THT. We believe that our findings can bring a new level of insight necessary to design new probe molecules for amyloid fibrils with ramifications for the development of diagnostic and therapeutic approaches to conformational diseases.

■

ASSOCIATED CONTENT

S Supporting Information *

Computational details for the integrated approach employed in this study. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was supported by a grant from the Swedish Infrastructure Committee (SNIC) for the project “Multiphysics Modeling of Molecular Materials”, SNIC 023/07-18. J.K. thanks the Danish Center for Scientific Computing (DCSC), The Danish Councils for Independent Research (STENO and Sapere Aude programmes), the Lundbeck Foundation, and the Villum foundation for financial support.

■

REFERENCES

(1) F.E. Cohen, F. E.; Kelly, J. W. Therapeutic Approaches to Protein-Misfolding Diseases. Nature 2003, 426, 905−909. (2) Härd, T.; Lendel, C. Inhibition of Amyloid Formation. J. Mol. Biol. 2012, 421, 441. (3) Nilsson, K. P. R. Small Organic Probes as Amyloid Specific Ligands: Past and Recent Molecular Scaffolds. FEBS Lett. 2009, 583, 2593−2599. (4) Buell, A. K.; Esbjörner, E. K.; Riss, P. J.; White, D. A.; Aigbirhio, F. I.; Toth, G.; Welland, M. E.; Dobson, C. M.; Knowles, T. P. J. Probing Small Molecule Binding to Amyloid Fibrils. Phys. Chem. Chem. Phys. 2011, 13, 20044−20052. (5) Hawe, A.; Sutter, M.; Jiskoot, W. Extrinsic Fluorescent Dyes as Tools for Protein Characterization. Pharm. Res. 2007, 25, 1487−1499. (6) Stsiapura, V. I.; Maskevich, A. A.; Kuzmitsky, V. A.; Uversky, V. N.; Kuznetsova, I. M.; Turoverov, K. K. Thioflavin T as a Molecular Rotor: Fluorescent Properties of Thioflavin T in Solvents with Different Viscosity. J. Phys. Chem. B 2008, 112, 15893−15902. (7) Munishkina, L. A.; Fink, A. L. Fluorescence as a Method to Reveal Structures and Membrane-Interactions of Amyloidogenic Proteins. Biochim. Biophys. Acta, Biomembr 2007, 1768, 1862−1885. 76

dx.doi.org/10.1021/jz3018557 | J. Phys. Chem. Lett. 2013, 4, 70−77

The Journal of Physical Chemistry Letters

Letter

(28) Dominguez, M.; Rezende, M. C. Towards a Unified View of the Solvatochromism of Phenolate Betaine Dyes. J. Phys. Org. Chem. 2010, 23, 156−170. (29) 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. et al. Gaussian 09, revision A.02; Gaussian, Inc.: Wallingford, CT, 2009.

77

dx.doi.org/10.1021/jz3018557 | J. Phys. Chem. Lett. 2013, 4, 70−77