Alpha-synuclein ChimeraX tutorial

2D electron micrograph of alpha-synuclein fibrils
from Guerrero-Ferreira R, eLife 2018.

Tom Goddard
BCBB Training Seminars
Sept 11, 2019

We look at basic ChimeraX capabilities for display and analysis of atomic models using recent alpha-synuclein fibril structures from cryoEM. Using ChimeraX version 0.91 daily build. User's Guide and other tutorials.

Alpha-synuclein fibrils in Parkinson's disease

Alpha-synuclein is a small protein of 140 amino acids that aggregates in neuronal cells in Parkinson's disease. It is believed to form fibrils in intracellular compartments called Lewy bodies. At least 4 very different fibril atomic models have been determined by cryoEM in the last few years, all from recombinant preparations. Models from fibrils exctracted from brain tissue have not yet been determined. Six mutations (A30P, E46K, H50Q, G51D, A53E, and A53T) are associated with Parkinson's disease or alpha-synuclein aggregation, although these account for extremely few cases of disease. A truncated version of alpha-synuclein leaving residues 1-121 is much more prone to aggregation and is used in the cryoEM studies.

We will look at structures (PDB 6h6b, EMDB 0148 and 4276) described in

      Cryo-EM structure of alpha-synuclein fibrils.
      Guerrero-Ferreira R, Taylor NM, Mona D, Ringler P, Lauer ME, Riek R, Britschgi M, Stahlberg H.
      Elife. 2018 Jul 3;7. pii: e36402. doi: 10.7554/eLife.36402.
    

and (PDB 6rt0, EMDB 4994 and 4996)

      bioRxiv preprint, May 2019
      Two new polymorphic structures of alpha-synuclein solved by cryo-electron microscopy
      Ricardo Guerrero-Ferreira, Nicholas M.I. Taylor, Ana-Andrea Arteni, Pratibha Kumari, Daniel Mona,
      Philippe Ringler, Markus Britschgi, Matthias E. Lauer, Joeri Verasdock, Roland Riek, Ronald Melki,
      Beat H. Meier, Anja Böckmann, Luc Bousset, Henning Stahlberg
    

Fibril atomic model

open 6h6b hide atoms
show ribbons
Looking down fibril axis.
color /D,E,F,I,J tan
color /A,B,C,G,H pink
show /A ribbon only show /A atoms
style stick
color byhet

Details

  1. Fibril model has two proto-filaments.
  2. Each protofilament is stacked copies of alpha-synuclein monomer.
  3. Five layers in this model, but typical length is 100 to 1000 stacked monomers.

Sequence

Chain table from log. Click alpha-synuclein link to show amino acid sequence.
Ctrl-click ribbon to select residue.
Up-arrow key to select beta strand.
Ctrl-click background to clear selection.
Selection also shows on sequence view. label /a
label height 1.5
hide H
label delete

Details

  1. Sequence panel shows experiment sample had 121 amino acids.
  2. Only amino acids number 38 to 95 are in atomic model, others are disordered and not seen in cryoEM map.
  3. Black outline boxes in sequence indicate amino acids not in atomic model.
  4. Blue background in sequence panel is for beta strands.
  5. Selecting amino acids on structure highlights location in sequence.
  6. Selecting in sequence panel, highlights on 3d structure.

Binding

show ribbons
hide atoms
hbonds
~hbonds
show surfaces
or use Molecule Display toolbar
mlp
or use Molecule Display toolbar
hide /A,B,C,G,H surface
hide /A,B,C ribbon
Dimer interface is not very hydrophobic.
color bfactor
show /G,H atoms
Blue is highest resolution, red lowest.

Details

  1. Beta sheets are formed by strands from stacked alpha synuclein.
  2. Hydrogen bonds shown with hbonds command as dashed blue lines.
  3. Command mlp colors surface by hydrophobicity, yellow hydrophobic, blue hydrophilic.
  4. Interface between protofilaments has no hydrogen bonds and does not appear to be hydrophobic.
  5. In cryoEM bfactors are often not determined but some similar concept like local resolution is put into the PDB entry bfactor column. Not sure what was done for this structure.

Measurement

Rotation angle (degrees) 179.51228273
Shift along axis 2.44997643
RMSD between 806 atom pairs is 0.007 angstroms
Use distance mouse mode from toolbar and click two atoms
distance /H:38@HD12 /G:38@HD12
surface close ; hide ribbon
~distance
open 6h6b
align #2/G toAtoms #1/I reportMatrix true
tile
close #2

Details

  1. Layers in two protofilaments are offset by half a layer thickness.
  2. Layer half-thickness is 2.45 Angtroms from align command.

Disease mutations

We can show some the location of disease associated mutations A30P, E46K, H50Q, G51D, A53E, and A53T (from Guerrero-Ferreira, eLife 2018).

color :46,50,51,53 yellow
label :46,50,51,53 height 3
Alanine 53 at proto-filament interface. swapaa /I,G:53 GLU
label delete

Details

  1. 3 of 4 mutation sites are at protofilament interface.
  2. But they are not at the interface in other fibril models.
  3. Command swapaa (swap amino acid) mutates and sets sidechain position to avoid clashes or optimize hydrogen bonding.
  4. After mutating a residue molecular dynamics may be desirable to fix clashes requiring other software.

Polymorphism

Four very different alpha-synuclein fibril structures have been published (PDB 6h6b, 6cu8, 6rt0, 6rtb, compared in Guerrero-Ferreira, bioRxiv preprint May 2019). Compare two of them.

open 6rt0
show #2/J,E ribbon only
hide atoms
matchmaker #2/J to #1/I
tile matchmaker #2/J to #1/I
delete #1 & ~/I,G
delete #2 & ~/J,E
morph #1,2
Press red circle button at right of morph slider to record movie.

Details

  1. matchmaker command aligns handling sequence differences by first doing a sequence alignment to pair residues between two structures.
  2. morph command operates on entire model, so first delete all but 2 chains.
  3. Command "delete #1 & ~/I,G" means delete model #1 but not chains I and G.

cryoEM data

Look at cryoEM map EMDB 4994 used to determine atomic model PDB 6rt0.

open 4994 from emdb view initial #2
transparency #3 50
show #2 atoms,model
Use crop mouse mode to show thin slab of map.
volume #3 region 0,0,145,279,279,164
volume #3 level 0.01
volume #3 style mesh

Details

  1. 3D cryoEM map was made by Relion software from 2D images of fibrils.
  2. Map has 3.1 Angstrom resolution. Can be difficult to correctly position side chains.
  3. Disordered parts of molecule are not seen in map because thousands of different 2D views are averaged to get 3D map.