Structural studies of protein fibrillation...pellets at pH 5 was established, resulting in aggregation of highly b‐sheeted prion proteins • The process was monitored by SAXS, DLS

Structural studies of protein fibrillation

D.Svergun, EMBL-Hamburg

Am. Soc.HematologyAm. Soc.HematologyAm. Soc.Hematology

E. coli Biofilm AJC1/FlickrE. coli Biofilm AJC1/FlickrE. coli Biofilm AJC1/Flickr

pdb‐code 1d0r, GLP1

• Amyloid diseases (Alzheimers, Parkinsons…)

• Functional Fibrils (Antimicrobial, Biofilm, Spider silk)

• Biopharmaceutical stability (insulin, glucagon, …)

• Self‐assembly bio‐systems • Drug delivery (Degarelix) • Nano‐material: the strength of steel

Amyloid(‐like) fibrils

GNNQQNY fibrilsA.E. Langkilde

aSNA. van Maarschalkerweerd

Fibril Structure

Jimenez et al, 2002, PNASHuman Insulin

GNNQQNY fibrilsA.E. Langkilde

Nelson et al, 2005, NaturePeptide/yeast prion protein

Fitzpatrick et al, 2013, PNASTTR 105-115

aSNA. van Maarschalkerweerd

20 µm

Cross-β-sheets formationThe amyloid fibrils share common structural properties: twisted and unbranched fibrils typically have a diameter of about 100 Å and highly variable lengths up to several microns. Furthermore, mature fibrils are suggested to be composed of intertwined protofibrils built from two to three intertwining protofilaments,each with a typical diameter of 15–50 Å.

The most dominant repeating features of presumably all amyloid fibril structures are suggested cross‐β‐sheets with the β‐strands running perpendicular to the fibril axis and an inter‐strand spacing of about 4.8 Å. Atomic resolution structures of amyloid‐like hexameric peptide segments, which form such cross‐β‐sheets, have been determined.

Dynamic protein structural equilibria

A hypothetical monomeric protein (green) has an intrinsically disordered N‐terminus (lighter green, a number of potential conformations are shown). The N‐terminus may become ordered upon dimer formation (red:purple). Under particular experimental conditions the protein may fibrillate (the fibrils are sketched in orange colors, not to scale with the monomeric protein). From B. Vestergaard, Arch. Biochem. Biophys. 602 (2016) 69

Tht fluorescence to detect amyloids

Thioflavin T fluorescence is often used as a diagnostic of amyloids. When ThT binds to beta sheet‐rich structures, such as those in amyloid aggregates, the dye displays enhanced fluorescence and a characteristic red shift of its emission spectrum.

fibril

Fibrillation prone conformation Nucleus?

On‐pathway oligomer

Fibrillation related off‐pathway oligomer

Amorphous aggregate

Native unfolded monomer

Native interactions

Partially (un)folded (membrane vicinity)

?

Amorphous aggregate

FIBRIL strain 1

FIBRIL strain 2

Alternative fibrillation pathways

Which species is cytotoxic?What is the molecular mechanism?

How do they interfere?Structural elucidation of intermediates

is of paramount importance but extremely challenging

Mechanism includes multiple states in equilibrium

fibril




Amorphous aggregate


Native interactions


?

Amorphous aggregate

FIBRIL strain 1

FIBRIL strain 2


0

1000

2000

3000

4000

5000

6000

0 5 10 15 20

ThT-

emis

sion

[rfu

]

T im e [h]

y = (m 1)+((m 3 )/(1+exp(-(m 0-m ...

E rrorValue207.13-119.03m 1 262.244315 .1m 3

0 .341 878.1 073m 5 0 .303 511.5 765m 6

NA9.4279 e+6ChisqNA0 .96673R

SAXS

Mixtures SAXS data – additive

fibril




Amorphous aggregate


Native interactions


?

Amorphous aggregate

FIBRIL strain 1

FIBRIL strain 2


0

1000

2000

3000

4000

5000

6000

0 5 10 15 20

ThT-

emis

sion

[rfu

]

T im e [h]

y = (m 1)+((m 3 )/(1+exp(-(m 0-m ...

E rrorValue207.13-119.03m 1 262.244315 .1m 3

0 .341 878.1 073m 5 0 .303 511.5 765m 6

NA9.4279 e+6ChisqNA0 .96673R

SAXS

fibril




Amorphous aggregate


Native interactions


?

Amorphous aggregate

FIBRIL strain 1

FIBRIL strain 2


fibril



FIBRIL strain 1

0

1000

2000

3000

4000

5000

6000

0 5 10 15 20

ThT-

emis

sion

[rfu

]

T im e [h]

y = (m 1)+((m 3 )/(1+exp(-(m 0-m ...

E rrorValue207.13-119.03m 1 262.244315 .1m 3

0 .341 878.1 073m 5 0 .303 511.5 765m 6

NA9.4279 e+6ChisqNA0 .96673R

Itot=xIstart+yIintermediate+zIfibrilSAXS advantages:- Solution + few restrictions on

experimental conditions- Time resolved analysis of the

structure and quantitative characterization of evolving mixtures

• Prion protein: purified on‐path oligomers (SAXS, DLS and negative stain EM)

• Insulin (the first shape analysis of unseparatable intermediate)

• Transthyretin (joint use of SAXS with hydrogen exchange mass spectrometry)

• α‐synuclein (shape analysis of intermediates and correlation with their cytotoxicity)

SAXS applications to study intermediates of amyloid fibrillation

• The cellular prion protein (PrPC) may form a misfolded isoform (PrPSc) that assembles into amyloid fibrils.

• As increased levels of oxidative stress have been linked to prion diseases, a metal‐induced oxidation of human PrP (90–231) was studied

• An in vitro conversion assay based on aerobic incubation of PrP in the presence of elemental copper pellets at pH 5 was established, resulting in aggregation of highly b‐sheeted prion proteins

• The process was monitored by SAXS, DLS and negative stain EM

L. Redecke et al. J. Struct. Biol. 157 (2007) 308–320.

Oligomers formed during oxidative prion protein aggregation

• Location and covalent modification of histidine and methionine residues in rhPrP (90–231) after metal‐induced oxidation

• NMR structure of recombinant human PrP corresponding to residues 125–228 is presented (residues 90–124 display a flexible random coil structure). Locations of histidine (red) and methionine residues (purple) supposed to act as a primary target for oxidative damage are highlighted



Negative stained electron micrographs of oligomerized rhPrP (90–231); the bar, 50 nm(A) Monomers can hardly be visualized (B) Oligomers O1, which are formed after incubation of monomeric rhPrP (90–231) for 3 h at 37 C in the presence of Cu0, exhibit a regular appearance with a diameter of 8.2 ± 2.4 nm. (C) The extension of the incubation period up to 12 h leads to oligomers O2, which are characterized by a wider distribution in size and shape (diameter of about 30–45 nm).


SAXS on purified monomer O1. Blue, the average shapeGrey, most typical shapeRed, space‐filling model of the monomer


Schematic mechanism of PrP aggregation induced by Cu0. A mechanism of oxidative damage of rhPrP (90–231) includes a direct oxidation of the accessible methionines by H2O2 as well as oxidation of the histidines catalyzed by Cu2+ ions bound to the PrP molecule. The oxidized PrP molecules are converted into a β‐sheet enriched structure, forming stable oligomeric intermediates (about 25 mers and 100 mers) as well as precipitating aggregates.

EMBL SAXS X33 beamline, 1979-2012

This is how it worked in 2005

X33 SAXS beamline of the EMBL (DORIS-3, bending magnet)

Hardware-independent analysis block

Automated data take and analysis pipeline

Simultaneous SAXS-WAXS data collection with two PILATUS detectors

High throughput robotic sample handling

Full remote control and user access

19

First automated SAXS sample changer for solutions (2007)

First multi-module Pilatus detector outside SLS (2007)

First automated SAXS data analysis pipeline (2008)

First remote SAXS experiment (2009)

Worked hard until the very last day of DORIS-3 (21.10.2012)

October 21, 2012, 19:35

Last bunch of X33 users (EMBO SAXS Course) on the last day of DORIS-3 operation

Data collection at X33, EMBL‐Hamburg

Human insulin fibrillation

Vestergaard, B., Groenning, M., Roessle, M., Kastrup, J.S., de Weert, M.V., Flink, J.M., Frokjaer, S., Gajhede, M. & Svergun, D.I. (2007) PLoS Biol. 5, e134

The initial step has been suggested to proceed through a non-native, partly unfolded, monomeric intermediate, which forms an oligomeric nucleus prior to elongation of protofilaments.

Insulin fibrillation is a huge problem. Amyloid deposits were observed in patients after subcutaneous insulin infusion as well as in vitro, where insulin easily fibrillates, causing problems during production, storage, and delivery of insulin-based drugs.

Fibrillation of insulin0 hours: monomers

9 hours: mature fibrils

Scattering and shape of the intermediate

Growth rate of fibrils is proportional to volume fraction of intermediates

Monomers

Fibrils

Intermediate

SAXS detects three components

Component1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

log (Eigenvalue)

5

6

7

8

5 g/l 20% acetic acid 0.5M NaCl 45˚C

Fibrillation of insulinOligomers are fibrillation nuclei and potential targets against amyloidosis

Assembly of protofilaments Formation of mature fibrils from the helical precursors (5-6 units) from intertwinning protofilaments

Vestergaard, B., Groenning, M., Roessle, M., Kastrup, J.S., de Weert, M.V., Flink, J.M., Frokjaer, S., Gajhede, M. & Svergun, D.I. (2007) PLoS Biol. 5, e134

Transthyretin (TTR) aggregation

The process of amyloidosis is linked to tissue degeneration, yet amyloid fibrils themselves may not mediate the cytotoxicity. In fact several studies support a cytotoxic role for lower molecular mass soluble species formed during TTR amyloidogenesis. Only limited treatments are available for the TTR amyloid diseases and it is of utmost importance to gain insight into the molecular processes underlying TTR amyloidosis in order to stimulate further therapeutic development.Here, SAXS combined with and hydrogen exchange mass spectrometry (HXMS) was used to describe an unexpectedly dynamic TTR protofibrilstructure which exchanges protomers with highly unfolded monomers in solution.

TTR is a 55 kDa homotetrameric β‐sheet‐rich protein mainly produced in the liver, being responsible for the transport of the hormone thyroxine and vitamin A.

Senile Systemic Amyloidosis Familial Amyloidotic PolyneuropathyFamilial Amyloidotic CardiomyopathyFamilial Leptomeningeal Amyloidosis

TTR amyloid diseases are associated with four protein aggregation “gain of toxic function” diseases.

Groenning et al. & Vestergaard (2015) Scientific Reports, 5:11443

TTR dH2O___ TTR 3.5 min….. TTR 10 minTTR unfolded (11% ac.ac.)

SAXS on TTR fibrillation• Pre‐ini MW ≈ 3.3 monomers• Pre‐ini ≠ xtal• Pre‐ini ≠ xtal+dimer/monomer• Pre‐ini = 85% xtal + 15% unfolded



~2900 kDa~215 protomers



In‐fibril

In‐tetramer

HXMS:Solvent protectionin protofibtils ismuch lower thanin tetramer

TTR fibrillation proceeds via addition of considerably unfolded monomers, and the continuous presence of amyloidogenic structures near the protofibril surface makes the fibrillar state interchange with the solution state.

Fibrillation of α‐synuclein (αSN)

Giehm, L., Svergun, D.I., Otzen, D.E. & Vestergaard, B. (2011) PNAS USA, 108, 3246

Aggregation of αSN leads to Parkinson decease. The fibril formation process characterized by SAXS reveals that there exists an intermediate oligomer formed by several partially unfolded αSN molecules, and the mature fibril is formed by association of these oligomers.

0

1000

2000

3000

4000

5000

6000

0 5 10 15 20

ThT-

emis

sion

[rfu

]

T im e [h]

y = (m 1)+((m3)/(1+exp(-(m0-m ...

ErrorValue207.13-119.03m1 262.244315.1m3

0 .341878.1073m5 0 .303511.5765m6

NA9.4279e+6ChisqNA0.96673R

Interactions of the oligomer with vesicules

0

20

40

60

80

100

0 5 10 15 20 25 30

Calcein release Oligomer [%]Calcein release Monomer [%]

Cal

cein

rele

ase

[%]

Time [min]

DOPG vesicles

Oligomer isolated via SEC, low conc.

Monomer, low conc.

Oligomeric and monomeric αSN were isolated via gel filtration of a fibrillating sample after 6 hours incubation, where significant amounts of the oligomer are present. To examine whether the oligomer could permeabilise DOPG vesicles, the release of the fluorescence probe calcein entrapped in vesicles at self-quenching concentrations was monitored. The oligomer (and not the monomer) is shown to penetrate into the vesicles, i.e. the oligomer is potentially cytotoxic.

Analysis of SAXS data from evolving multi‐component systems is not trivial

K

kjkkmjm sFvsI

1)()(

We have M curves Im(sj), each measured in J points, and they correspond to K components in the evolving mixture (1 ≤ m≤M, 1 ≤ j≤J, 1 ≤ k≤K). Each of the measured curves should be represented as a linear combination

where vkm is the volume fraction of the k-th component in m-th curve and Fk(sj) is the scattering from the component. In general, neither vkm nor Fk(sj) are known a priori, but we know that all values should be non-negative, K may be extracted from SVD. Quite often, (some of) the Fk(sj) can be measured, pre-computed or guessed.

Decomposition into components

22

21)()(min FvsFvsI

K

kkkmm

The cost function to be minimized can be written as

Here, the first term in this cost function is the distance measure between the data and its approximation (overall discrepancy). The second, a “regularization term” may be added to govern a generalized solution (e.g. prevent overfitting to some local noisy effects).

If either all vkm or all Fk(s) are known, the other unknown set is computed simply by “normal” least squares. ALS starts e.g. with tentative Fk(s), computes vkm‘s, makes them feasible (e.g. positive), recomputes Fk(s), makes them feasible, etc.

Alternative least squares (ALS)

Multivariate Curve Resolution using Alternating Least Squares (MCR‐ALS)Herranz‐Trillo F, Groenning M, van Maarschalkerweerd A, Tauler R, Vestergaard B, Bernadó P. (2017) Structure, 25, 5‐15.

0

1 0 0 0

2 0 0 0

3 0 0 0

4 0 0 0

5 0 0 0

6 0 0 0

0 5 1 0 1 5 2 0

ThT-

emis

sion

[rfu

]

T im e [h ]

y = (m 1 ) + ( ( m 3 )/(1 + e x p (- ( m 0 -m . . .

E rr o rV a lu e2 0 7 .1 3-1 1 9 .0 3m 1 2 6 2 .2 44 3 1 5 .1m 3

0 .3 4 1 8 78 .1 0 7 3m 5 0 .3 0 3 5 11 .5 7 6 5m 6

N A9 .4 2 7 9 e + 6C h is qN A0 .9 6 6 7 3R

Reduce Ambiguities: Use of constrains: non‐negativity and closure)

COSMiCSComplex Objective Structural analysis of Multi‐Component Systems

Reduce Ambiguities: Use of multiple data matrices (standard: different data sources)

Absolute scale(I vs s)

Kratky plot(I·s2 vs s)

Porod plot(I·s4 vs s)

Holtzer plot(I·s vs s)

(semi‐) Automatic combinationreducing ambiguityOptimizingdiscriminative power by requiring overall fitin different intensityrepresentations (differentweighting functions).

Proof of concept: Human Insulin

B. Vestergaard et al. (2007) PLoS Biology

Inclusion of complementary data for the decomposition

Possible additional data to include: ThT fluorescence, CD etc

0

1 0 0 0

2 0 0 0

3 0 0 0

4 0 0 0

5 0 0 0

6 0 0 0

0 5 1 0 1 5 2 0

ThT-

emis

sion

[rfu

]

T im e [h ]

y = (m 1 ) + ( ( m 3 )/(1 + e x p (- ( m 0 -m . . .

E rr o rV a lu e2 0 7 .1 3-1 1 9 .0 3m 1 2 6 2 .2 44 3 1 5 .1m 3

0 .3 4 1 8 78 .1 0 7 3m 5 0 .3 0 3 5 11 .5 7 6 5m 6

N A9 .4 2 7 9 e + 6C h is qN A0 .9 6 6 7 3R

Reduce Ambiguities: Use of multiple data matrices (standard: different data sources)

Direct shape determination of an intermediate

Let us consider a case when the initial and final states are known/measured but there is a major (unknown) intermediate component (often happens for amyloid formation, viral assembly etc):

Ik(s)= vinit,kIinit(s) + vfin,kIfin(s) + vikIi(s)

We know neither the volume fractions nor the intensity of the intermediate but we can assume that this intensity corresponds to (an unknown) shape and try to reconstruct this shape and fractions by fitting all available data with the above linear combination.

Components/curves Volume fractions Simulated curves

Shape determination of the intermediateA shape determination program DAMMIX is developed to reconstruct the shape of the unknown intermediate. It works like DAMMIF but looks for a shape that yields the intensity Ii(s) providing the overall best fitting to all available data sets with the values of the volume fractions varying between zero and unity:

MIN ∑k [Ik(s) - {vinit,kIinit(s) + vfin,kIfin(s) + vikIi(s)}]2

The use of DAMMIX on the published data reporting a helical structural nucleus for insulin amyloid fibrils (Vestergaard et.al. (2007), PloS Biol. 5, 1089)

P.V.Konarev & D.I.Svergun (2017), submitted

Shape determination of various intermediates

Multiple assembly states of lumazine synthase (data of Zhang et al, 2006, J Mol Biol 362, 753)

Dimer-tetramer NGF equilibrium (data ofCovaceuszach et al 2015, Biophys J. 108, 687)

Summary

SAXS is clearly the method of choice to study amyloid formation processes

SAXS studies are readily combined with the results of other methods

There are methods available to globally analyse and fit the SAXS data, detect the intermediates and reconstruct their shape

Fibrillation is a slow processand the kinetics and influence of various factors can easily be studied in real time

Acknowledgments

EMBL BioSAXS Group and especially: C. Blanchet, C. Jeffries, D. Franke, A.Kikhney, H.Mertens, A.Panjkovich, M. Graewert, N.Hajizadeh

Former BioSAXS Group members: M. Petoukhov, P.Konarev, M.Roessle

EMBL-HH: S.Fiedler University of Copenhagen: B.Vestergaard,

A.Langkilde, M.Groenning, L.Giehm University of Hamburg: L.Redecke, C.Betzel University of Montpellier: P. Bernado

Structural studies of protein fibrillation...pellets at pH 5 was established, resulting in aggregation of highly b‐sheeted prion proteins • The process was monitored by SAXS, DLS

Documents