Amyloids are broadly investigated protein misfolding products with characteristic β-sheet assemblies that have an important role in neurodegenerative diseases (e.g., Alzheimer's disease). While they are usually visualized by staining with Thioflavin-T, Congo Red, or other fluorescent markers, it still arouses a controversy over possible staining molecule influence on the amyloid structure or aggregation process. In this work we present, for the first time, the polarization analysis of two-photon excited autofluorescence of amyloids and confirm that polarization dependence of the observed emission can be correlated with the orientation of fibrils. We show the potential of two-photon excited autofluorescence for resolution of molecular organization of fibrils within amyloid superstructures. This label-free method is compatible with two-photon imaging already applied in investigation of neurodegeneration model in mice.
Amyloids are broadly investigated protein misfolding products with characteristic β-sheet assemblies that have an important role in neurodegenerative diseases (e.g., Alzheimer's disease). While they are usually visualized by staining with Thioflavin-T, Congo Red, or other fluorescent markers, it still arouses a controversy over possible staining molecule influence on the amyloid structure or aggregation process. In this work we present, for the first time, the polarization analysis of two-photon excited autofluorescence of amyloids and confirm that polarization dependence of the observed emission can be correlated with the orientation of fibrils. We show the potential of two-photon excited autofluorescence for resolution of molecular organization of fibrils within amyloid superstructures. This label-free method is compatible with two-photon imaging already applied in investigation of neurodegeneration model in mice.
Amyloids are insoluble aggregates
of misfolded proteins, rich in β-sheet structures forming filaments
and fibers. They are the hallmark of Alzheimer’s disease (AD),
Parkinson’s disease (PD), and a range of other neurodegenerative
disorders as well as amyloidoses affecting various organs and systemic
conditions with amyloid deposits in multiple organs.[1,2] Diagnosing of the disease and investigation of the origins of amyloidogenesis
require imaging in vivo and needs to be performed
for complex, densely packed biosystems. Optical fluorescence microscopy
is a method of choice in such cases. Dyes such as Thioflavin-T (ThT)
and Congo Red have been widely used as selective fluorescent labels
for amyloid formation studies as they exhibit distinct intermolecular
interaction with β-sheet-rich, mature fibrils and only weak
binding to the oligomeric intermediates.[3−5] Specific ordering of
ThT on amyloid fibrils allowed application of polarization sensitive
techniques to study amyloids organization, as was shown by Brasselet
et al.[6] However, the presence of a label
may affect the aggregation processes or structural properties of amyloids.
As was shown for ThT, the presence of the dye interferes with the
aggregation process and this arouses controversy over fiber-binding
modes.[7−9] Moreover, a multitude of factors affects ThT molecule
binding to amyloids.[5] Thus, analysis of
the fiber structure based on a label-free technique would be beneficial.
Interestingly, autofluorescence characterized by emission maxima located
between 340 and 450 nm, dependent on the excitation wavelength, with
lifetimes of approximately 1–3 ns was reported as an intrinsic
optical property of amyloid fibers for a variety of amyloidogenic
proteins (e.g., Aβ1–42, α-synuclein,
insulin).[10−12] It should be emphasized that the phenomenon does
not require the presence of multiple covalent bonds conjugation or
the presence of aromatic residues necessary for traditional fluorescent
molecular systems.[10,11,13,14] The origin of this luminescence of amyloids
is still poorly understood and several explanations have been discussed,
which point at the delocalization of electrons through hydrogen bonds
or proton transfer as possible mechanisms.[15] Recently, Grisanti et al. showed by numerical simulations that the
main contribution to excitation bands of autofluorescence of amyloids
lacking aromatic amino acids comes from stabilization of nπ* states with excitation localized on amide groups.[14] Simultaneously, rigidity of the fibrils introduced by cross
β-sheet structure enhances the radiative decay channel, which
results in blue-green fluorescence.[10] However,
the determination of what contributes to autofluorescence in real
amyloids, with various side chains, is still a challenge. Nevertheless,
amyloids autofluorescence utility in in vivo imaging
of amyloid deposits was already confirmed.[16]Here, we demonstrate, for the first time, that amyloids autofluorescence
presents incident light polarization-dependent excitation and emission,
which can be correlated with the orientation of amyloid fibrils. We
study amyloid structures under polarization-sensitive two-photon excited
fluorescence microscope (ps-TPFM). Two-photon excitation provides
deeper optical penetration, lower scattering, and lower phototoxicity
of imaging, in comparison to the one-photon based excitation microscopy,[17,18] and its advantages were already successfully implemented in in vivo imaging of amyloids deposits in ADmouse models,
stained with two-photon absorbing fluorophores.[19] Moreover, the two-photon excitation presents reduced angular
photoselection of fluorophores, which results in higher precision
in determination of molecular organization in a sample, as was shown
with ps-TPFM in artificial lipid vesicles and cell membranes imaging.[20] Further, we explore the application of two-photon
excited autofluorescence to resolve the three-dimensional organization
of fibrils in higher order amyloid structures in aqueous environment.
The ps-TPFM was already used by our group in the studies of DNA strands
organization and proved to provide detailed information on local ordering
of DNA helices in various self-organized liquid crystal phases.[21−23] Here, we image amyloid spherulites: larger-scale and β-sheet
rich aggregates, which are seen in the post-mortem brains of Parkinson’s and Alzheimer’s patients.[24] Such spherical structures are also found in,
e.g., Creutzfeldt–Jackob disease, but their exact origin and
biological importance is yet to be discovered.[25,26] Johansson et al. reported that spherulitic amyloid deposits can
be detected on the basis of their intrinsic optical properties, including
nonlinear optical processes such as second harmonic generation (SHG).[27] We show that ps-TPFM can provide additional
information about local orientation of fibrils in a spherulite. Finally,
we compare the information provided by two-photon excited autofluorescence
(2PAF) with two-photon excited fluorescence imaging of amyloids labeled
with ThT and see clear differences of labeling various components
of amyloid systems.We obtained amyloid spherulites from bovineinsulin powder in denaturing
conditions, as described in the Supporting Information section. Spherulites were imaged under a polarized light microscope
with crossed polarizers, and characteristic anisotropic patterns with
Maltese crosses were observed (Figure a). Such patterns arise due to the regular and radially
growing arrangement of amyloid fibrils.[25] Moreover, spherulites present a disordered/isotropic core, with
the size dependent on the overall size of a spherulite. When imaged
under scanning electron microscope (SEM), the spherulites reveal cracking,
due to the dehydration, with organized radial lines from the periphery
to the disordered center of a spherulite (Figure b).
Figure 1
(a) Polarized light optical microscopy image
of different sizes
of spherulites with crossed polarizers. Scale bar 100 μm. (b)
SEM image of dehydrated and cracked spherulites. Scale bar 10 μm..
(a) Polarized light optical microscopy image
of different sizes
of spherulites with crossed polarizers. Scale bar 100 μm. (b)
SEM image of dehydrated and cracked spherulites. Scale bar 10 μm..The one-photon excited emission spectra of spherulite
solution
were measured (Figure S1a) and identified
as the amyloid autofluorescence with the maxima at 425 nm observed
for 375 nm excitation, in agreement with previous reports.[11] Samples were then excited by a femtosecond mode-locked
Ti-sapphire laser with the excitation wavelength set to 810 nm. We
observed an approximately quadratic dependence of two-photon excited
fluorescence (2PEF) on the incident laser power (Figure S1b), which confirms the two-photon character of the
observed processes. The presence of two-photon excited autofluorescence
of label-free spherulites enabled their imaging in which we illuminated
them with varying directions of the incident light polarization and
analyzed the direction of 2PEF polarization (Figure ). In detail, we performed raster scans with
the polarization angle of incident light, α, aligned with the Y or X axis in the sample plane, as denoted
with EX white arrows in the respective plots. The second white arrow,
denoted as EM, corresponds to the direction of emission components
(horizontally and vertically polarized emission components). A comprehensive
description of our microscope setup and the imaging process is available
in the Supporting Information (Figure S2).
Figure 2
Two-photon autofluorescence intensity raster scans of a label-free
insulin spherulite. Polarization of the excitation beam (Ex) and emission
(Em) is denoted with white arrows, and the excitation wavelength was
set to 810 nm. The same spherulite imaged under polarized light microscope
with crossed polarizers is available in the Supporting Information (Figure S3).
Two-photon autofluorescence intensity raster scans of a label-free
insulinspherulite. Polarization of the excitation beam (Ex) and emission
(Em) is denoted with white arrows, and the excitation wavelength was
set to 810 nm. The same spherulite imaged under polarized light microscope
with crossed polarizers is available in the Supporting Information (Figure S3).Regions of the strongest two-photon excited luminescence were observed
to be arranged in patterns corresponding to the excitation beam polarization
angle α. The maximum 2PEF intensity was observed when EX and
EM are parallel, which confirms that the absorption and emission transition
dipole moments are parallel. Localization of the highest 2PEF intensities
for respective scans in Figure confirm radial organization of emitting dipoles. Comparison
with Figure and the
described radial growth of fibrils from the spherulite core[27,28] leads to a conclusion that both absorption and emission transition
dipole moments are parallel to the long axis of amyloid fibrils. Characteristic
features of amorphous core surrounded by highly ordered radial structures,
observed under standard polarized light microscope (see Figure S3) were also visible under TPFM. However,
weak autofluorescence was collected also from the full area of the
nonbirefringent core, where high ordering is not expected.Several
spots on selected spherulites were chosen to perform a
full polarization analysis, where horizontally and vertically polarized
emission dependence on the excitation beam polarization angle α
were collected: I(α)
and I(α), respectively
(Figure ). Due to
the photoselection and with an assumption that absorption and emission
angles of the molecular transition dipoles are collinear, the highest
emission intensity is observed when the transition dipole moment of
the molecule is aligned parallel to the polarization of the incident
light. Simultaneous detection of X and Y components of 2PF enables
us to reveal information about molecular ordering as in isotropic
samples X and Y components should be perpendicular to each other,
whereas for a defined orientation of emitting molecules they should
be parallel to the direction of emitting dipoles. To reveal information
about molecular order of amyloid spherulites and interpret the angular
dependence of I(α)
and I(α), we adapted
the mathematical model introduced previously to describe ps-2PEF of
dyes labeling biological membranes and DNA structures.[21,29,30] Two-photon excited fluorescence
intensity of a single molecule is proportional to the probability
of two-photon excitation and emission collection. The excitation probability
depends on |μ⃗abs × E⃗|4, where μ⃗abs is the absorption transition dipole moment and E⃗ is the polarization of incident light field. The
probability of the emission is defined by |μ⃗em × μ⃗I|2, where μ⃗em and μ⃗I correspond to the
emission dipole and the analysis axis (detection polarization), respectively.
Analysis of the X and Y components of two-photon excited fluorescence
intensity is then performed with the equation:where the intensity is integrated over the
orientation distribution of μ⃗abs(Ω) and μ⃗em(Ω)
of emitters, where Ω is the solid angle. The open cone distribution
describes orientational distribution of a fluorophore on a fiber.
It is characterized by the following parameters: half aperture of
a cone, Ψ, with a variable thickness ΔΨ, and its
orientation angle Φ in a XY sample frame, which is collinear
with a long axis of an amyloid fibril, as shown in Figure a. In the fitting procedure
we took into account the polarization distortions introduced by a
dichroic mirror and high numerical aperture objective. All the details
of the measurement and fitting of angular dependence of 2PF are given
in the Supporting Information.
Figure 3
2PEF XY raster
scan of label-free spherulite excited with vertically
polarized light and polar graphs from spots denoted on the scan. Experimental
data are presented as dots (I, red; I, blue), and solid lines present the fitting
of the angular dependence of two-photon excited emission. Polarization
of the excitation beam (Ex) and emission (Em) are denoted with white
arrows. The excitation wavelength was set to 810 nm.
Figure 4
(a) Open cone model of the conical distribution of the emission
dipole of the dye (half angle, Ψ) in respect to the long fibril
axis (dashed line). Rotation of the fibril in the XY microscope sample
plane is described by the Φ angle. Aberrations of Ψ due
to the molecular rotations in filaments are described by ΔΨ.
(b) Schematic amyloid fibril model where red and blue arrows represent
the β-sheets oriented perpendicularly to the long axis of the
protofilament. Emitting dipole binding sites are denoted with pink
double-headed arrow. Θ stands for the angular distribution of
the protofilaments that build a fibril (maximal deviation from the
long axis of the fibril).
2PEF XY raster
scan of label-free spherulite excited with vertically
polarized light and polar graphs from spots denoted on the scan. Experimental
data are presented as dots (I, red; I, blue), and solid lines present the fitting
of the angular dependence of two-photon excited emission. Polarization
of the excitation beam (Ex) and emission (Em) are denoted with white
arrows. The excitation wavelength was set to 810 nm.(a) Open cone model of the conical distribution of the emission
dipole of the dye (half angle, Ψ) in respect to the long fibril
axis (dashed line). Rotation of the fibril in the XY microscope sample
plane is described by the Φ angle. Aberrations of Ψ due
to the molecular rotations in filaments are described by ΔΨ.
(b) Schematic amyloid fibril model where red and blue arrows represent
the β-sheets oriented perpendicularly to the long axis of the
protofilament. Emitting dipole binding sites are denoted with pink
double-headed arrow. Θ stands for the angular distribution of
the protofilaments that build a fibril (maximal deviation from the
long axis of the fibril).The high organization of amyloid fibrils near the spherulites periphery
allows us to determine autofluorescence transition dipole moments
with respect to the long axis of a fibril. We performed fitting of I(α) and I(α) collected from highly organized
regions of several spherulites (an example is given in Figure , spots 1, 2, 3, 4) and obtained
Ψ and ΔΨ values. The average relative angle between
the long fibril axis and the transition dipole moment of autofluorescence
Ψ was equal to 29°, with ΔΨ = 6°. It should
be emphasized that mature amyloid fibrils are composed of several
protofilaments that envelope one another in rope-like structure giving
rise to helical twist and diverse amyloid morphology.[31] The number of the coiled protofilaments determines the
overall fibril cross-sectional width and rigidity.[32] The angular distribution of the protofilaments that build
a fibril can be described with angle Θ, which specifies maximal
deviation from the long axis of the mature fibril, as presented schematically
in Figure b. Therefore,
the Θ angle contributes to the measured value of Ψ, together
with Ψ0, which expresses the deviation of the dye
from the long axis of the protofilament. In order to estimate values
of θ we prepared individual insulin fibrils grown in similar
conditions as fibrils in spherulites (details available in Supporting Information) and performed their imaging
under atomic force microscope (AFM). The AFM images (Figures S5a,c and S6a,b) clearly demonstrate the presence
of a twisted ribbon conformation with the pitch distance within the
range 85–134 nm, width of 22.4 ± 5.3 nm, and height of
3.8 ± 1.5 nm (Figures S5a,c and S6a,b, Table S1). The angle Θ determined for such fibrils was in the
∼7–18° range, which corresponds with previous reports.[6] The 2PAF distribution angle Ψ = 29°
determined in our studies for fibrils packed within a spherulite is
higher than θ of individual fibrils; therefore, within the spherulite
the Ψ0 value can be estimated to fall into the ∼11–22°
range. Of note is that the final morphology of amyloids is strongly
dependent on the growth conditions. In order to test the influence
of salt concentration, we incubated insulin at varying salt concentrations,
i.e., 0, 50, and 100 mM NaCl (more details in Supporting Information). AFM analysis revealed that ionic
strength impacts on fibrils morphology, especially on the twist pitch
distance (Figure S5, Table S1). With the
increase of salt concentration, the pitch distance tends to increase
and, finally, at 100 mM NaCl, fibrils form untwisted morphology. These
results further imply that the salt concentration may alter the final
distribution angle Ψ of grown spherulites.In order to
explain the 2PEF angular distribution, we decided to
compare it with superstructures of similar size and incubation conditions,
but labeled with broadly studied Thioflavin-T (ThT) dye. Regions of
the strongest two-photon excited luminescence observed in both ThT-labeled
spherulites and label-free samples presented surprisingly similar
patterns for corresponding α and 2PF polarizations (see Figure S7 for ThT-labeled spherulites, compare
with Figure ). As
was seen previously,[6,33] due to the highly anisotropic
structure of amyloids and specific binding of ThT perpendicular to
β-strands of a fibril, ThT can reproduce the orientation of
the fibrils. As ThT was previously applied in polarization-sensitive
imaging of amyloids,[6] we adapted the molecular
orientation of the dye being perpendicular to the long axis of the
protofilaments. Similarly as in the case of 2PAF, we obtained Ψ
and ΔΨ values by fitting I(α) and I(α) collected from highly organized regions of several
ThT-labeled spherulites (see Figure S8,
spots 1, 2, 3, 4). The average relative angle between the long fibril
axis and the transition dipole moment of ThT molecules Ψ was
equal to 30° and was almost the same as Ψ determined for
2PAF. Thus, we confirmed that highly polarized 2PEF corresponds to
the arrangement of fibrils within spherulites. The average value of
Ψ contains information about internal structure of a fibril,
with specific coiling of protofilaments, as denoted by the θ
angle. On the basis of a protofibril mean cross-section equal 30 ×
40 Å, reported by Saibil et al., fibrils composed of up to 6–10
protofilaments enveloping each other were estimated to have θ
values in range ∼7–18°, which is also expected
to be observed for fibrils in spherulites.[31]The polarization dependence of amyloid autofluorescence resolved
in our work leads to better understanding of the origin of that process.
Alignment of autofluorescence transition dipole moments parallel to
the long fibril axis (with discussed deviation due to the helical
structure of fibrils) supports the photophysical mechanism of amyloid
autofluorescence proposed by Grisanti et al.[14] Their theoretical analysis pointed at the deplanarization of amide
groups and extended carbonyl bonds as a main contribution to red-shifted
excitation bands in amyloid autofluorescence, as well as to the H-bond
structure of the cross β arrangement in amyloids as responsible
for restriction of nonradiative decay channels of autofluorescence.
Both of these effects may be valid for the explanation of the highly
polarized autofluorescence within amyloid superstructures.We
show that two-photon excited autofluorescence of amyloids is
highly polarized and distributed within ∼30° around the
long axis of the fibrils. Comparison with polarization analysis of
two-photon excited fluorescence of ThT bound to similar fibrils shows
the same conical distribution of fluorophores. Thus, the observed
distribution may be dependent on protofibril conformation in amyloid
fibrils. Two-photon excitation assures the restriction of excitation
to the focal point of the objective; thus, two-photon excited autofluorescence
will enable label-free, yet precise, localization of amyloids. Moreover,
the determination of relative angle between the long axis of amyloid
fibrils and autofluorescence transition dipole moments presented in
our work allows for application of autofluorescence for determination
of local organization of amyloid fibrils in various structures. The
presented results provide therefore a contribution toward new methods
to image amyloids organization and broaden our knowledge about amyloids
autofluorescence.
Authors: José L Jiménez; Ewan J Nettleton; Mario Bouchard; Carol V Robinson; Christopher M Dobson; Helen R Saibil Journal: Proc Natl Acad Sci U S A Date: 2002-07-01 Impact factor: 11.205
Authors: Joanna Olesiak-Banska; Marta Gordel; Katarzyna Matczyszyn; Vasyl Shynkar; Joseph Zyss; Marek Samoc Journal: Nanoscale Date: 2013-09-24 Impact factor: 7.790
Authors: Pu Chun Ke; Ruhong Zhou; Louise C Serpell; Roland Riek; Tuomas P J Knowles; Hilal A Lashuel; Ehud Gazit; Ian W Hamley; Thomas P Davis; Marcus Fändrich; Daniel Erik Otzen; Matthew R Chapman; Christopher M Dobson; David S Eisenberg; Raffaele Mezzenga Journal: Chem Soc Rev Date: 2020-07-07 Impact factor: 54.564
Authors: Fiona T S Chan; Gabriele S Kaminski Schierle; Janet R Kumita; Carlos W Bertoncini; Christopher M Dobson; Clemens F Kaminski Journal: Analyst Date: 2013-02-18 Impact factor: 4.616