Investigations of Macromolecular Structure and Function

1
Investigations of Macromolecular Structure and
Function

• II IV Protein folding, mis-folding and
  deposition

2
Learning Outcomes and Biophysical Techniques

• Understand the mathematical basis of the protein
  folding problem.
• Be able to describe the roles of macromolecular
  complexes in protein folding and secretion, and
  the methods that have enables us to understand
  their function.
• Be able to describe the structural biology of
  amyloid fibre formation and the methods used to
  study them.

3
Protein folding why is it important?
haemoglobin
collagen
globin
polyhedrin
Dobson, Nature 2003 426 884 -890
b-amyloid protein
4
Anfinsens Refolding Experiment
5
Anfinsens Refolding Experiment Interpretation
6
Anfinsen (1972, Nobel Prize acceptance speech)

• "The native conformation is determined by the
  totality of interatomic interactions and hence by
  the amino acid sequence, in a given environment."

7
The Protein Folding Problem Levinthals paradox

• How many conformations
• could a protein chain adopt?

8
Ramachandran plot

• Shows the allowed combinations of f and y
• Restrictions are due to steric clashes between
  adjacent sidechain groups
• Considerably less than full flexibility allowed.

9
The Protein Folding Problem Levinthals paradox

• there are about 3-5 combinations of f/y for each
  amino acid that would be acceptable to its
  neighbour.
• so its 3-5n conformations that a protein can
  adopt (n number of amino acids)
• 100 amino acids 4100 i.e about 1 x 1060
  conformations

10
Levinthal continued

• If a protein can sample 1014 conformations per
  second (average frequency of bond rotation is
  about 10-14s, 0.01 ps)
• Then it would take our 100 residue protein 1 x
  1045 seconds, or 5 x 1038 years to sample all its
  options.
• HOW DO PROTEINS EVER FOLD????

11
Balls roll downhill, and Levinthal is satisfied

• Levinthals problem is removed if the folding
  pathway is biased towards the correct structure.
• i.e. once you start moving in the right
  direction, youll get there

correctly folded state is the lowest energy state
12
Lowest energy state?

• Recall all the forces of macromolecular
  interaction
• Covalent bonds
• Electrostatics
• Hydrogen bonds
• van der Waals
• Hydrophobic effect
• If not optimal then each of these forces
• contributes an energy cost to the system.
• If optimal then the forces above are
• satisfied and the energy is minimal

13
Folding theory Energy landscapes

• formation of local structure stabilises
• An intermediate called the transition state
  ensemble is thought to exist
• Metastable
• Key contacts
• Same topology as fully folded structure
• enables collapse to native fold

energy
14
Intelligent typewriter analogy

• gd ytfg e uagkkt frgf j hjd pfhzgv kt
• af wwjs a dagroi tosj s jhr hrpie jr
• ht phis a dagthe shel a hui ehfelr mw
• is this a dagger that I see before me?

15
But do proteins fold on their own in the cell?
No! They have outside assistance
16
Co-translational and post-translational protein
folding.
co-translational
post-translational
17
Co-translational hide the hydrophobic patches
must keep the protein soluble to stop it
aggregating
18
Post-translationalshuffle the cards

• Protein disulphide isomerases
• enables multiple disulphide bridge containing
  proteins to exchange S-S bridge partners
• Peptidyl-prolyl isomerase
• enables cis-trans isomerization at X-Pro linkages

19
Protein disulphide isomerase

• ER localized
• Catalyses reduction and re-oxidation of
  disulphide bridges i.e. break and reform covalent
  bonds
• Isolated and purified byAnfinsen
• catalyses S-S bond exchange
• RNase A refolding in the presence of PDI occurs lt
  2 minutes c.f. 10 hours in absence

20
Peptidyl-prolyl isomerase

• catalyse the trans-cis isomerization of ca. 10
  of Pro-X bonds
• Dont break covalent bonds!
• 3 families, structurally and functionally
  distinct

a) FK506 binding proteins (bind FK506)
b) cyclophilins (bind CsA)
c) parvulin
21
Post-translationalComplete reshuffle
22
Electron microscopy helps understand Anfinsen
cage proteins.

• light microscopy has a limit of 1000x
  magnification i.e. a resolution of 0.2 mm
• no good for protein structure!
• electron microscope is essentially identical to a
  light microscope except that it uses electrons
  rather than light
• electron wavelength is 0.05Å so high resolution
  images are possible in theory
• electrons are damaging though

23
EM what it looks like and how it works
24
EM versus X-ray
25
Theoretical resolution
array of gold particles (bright spots), distance
between gold particles in the field is 0.2 nm (2
Angstroms)
but gold particles can withstand massive
electron doses, whereas proteins cannot
26
Negative stain EM Protects Proteins but
Resolution Drops

• Proteins are damaged by both the electron beam
  itself and the vacuum conditions
• Sample very rapidly degrades weak signal to
  noise ratio.
• massive increase in signalnoise from staining
• image the shadow, rather than the specimen
• much lower resolution, possibly artefacts from
  dehydration and flattening by the stain.

27
Cryo-EM

• Solves 2 major problems
• damage to the specimen by irradiation
• dehydration necessary for stability in the high
  vacuum
• Sample on grid in vitrified ice
• ice with no ice crystals freezing occurs within
  ms
• grid is maintained in the cryo-state by liq-N2

piston
forceps
EM grid
Liquid ethane, cooled by liquid nitrogen
28
Cryo-EM

• Solves 2 major problems
• damage to the specimen by irradiation
• dehydration necessary for stability in the high
  vacuum
• Sample on grid in vitrified ice
• ice with no ice crystals
• grid is maintained in the cryo-state by liq-N2
• The native, hydrated structure is observed
  directly
• Introduces major problem
• low contrast
• low signalnoise.

overcome by intense computational treatment (way
too complex)
29
What does something look like under cryo-EM
conditions

• In EM the goblet would appear like
• each line is a contour of equal density
• thin lines are low density structures
• thick lines are high density structures

3D specimen
30
EM So how do you get 3D data from a non
crystalline specimen? You need alternative views
of the object

• Each view contains unique information
• Each view contains information that is shared
  with at least one other view

31
EM So how do you get 3D data from a non 3D
specimen? Single particle imaging
Collect hundreds/thousands of images
32
EM So how do you get 3D data from a non 3D
specimen? Single particle imaging
Collect images that show the same view
33
EM So how do you get 3D data from a non 3D
specimen? Single particle imaging
Average images to improve the signal to noise
ratio
34
EM So how do you get 3D data from a non 3D
specimen? Single particle imaging
Use image information from one view that is
present in another view to improve the
resolution and produce a final image
35
Single particle imaging of proteins needs patience
Light coloured entities are protein
500 Å
36
Averaging and orientation of single particle
images
Identify thousands of single particles and
remove noise (computationally)
Representative of gt0000s
protein
detergent shell
37
Assemblies to fold proteinsCryo-EM can support
X-ray data

• This is the EM data for a large GroEL.GroES
  complex.
• Clearly consists of rings or layers of protein
  dense regions (white) with less dense regions
  (dark).

38
Assemblies to fold proteinsCryo-EM can support
X-ray data

• Determined by Helen Saibil (Birkbeck) to 30Å
• GroEL alone (gt 800 kDa)
• 4 rings, 7-fold symmetry
• GroES/GroEL complex (gt 1 MDa)
• bullet shaped, 5-rings

39
Conformational trapping of GroELGroES

• Why do we want to do this?
• Image proteins in different states
• How do we do this?
• Substrate analogues

ATP
40
GroESGroEL trapping

• GroES/GroEL complexed with ADP, ATP, AMP-PNP

41
Assemblies to fold proteins X-ray structures of
GroESGroEL

• GroES and GroEL x-ray structures
• Combined with EM images.

equatorial domains apical domains intermediate
domains GroES in blue
42
GroELGroES catalytic cycle structure
43
GroELGroES catalytic cycle function
44
Co-translational membrane insertion Cryo-EM can
extend X-ray data.

• Membrane (and secreted) proteins are inserted
  (secreted) co-translationally through a protein
  conducting channel (PCC)
• RibosomemRNAtRNASecYEG complex (minimally)
• Conserved-ish architecture in the ER of
  eukaryotes and the plasma membrane of bacteria
  and archaea.
• A massive complex gt 3 MDa

translating ribosome
mRNA
nascent chain
inserted membrane protein
membrane
(SecYEG)2
45
EM X-ray help understand the structure and
function.

• ribosome structure
• solved by X-ray crystallography
• SecYEG - monomer
• solved by X-ray
• but remember its a functional dimer
• but how do we fit the two together? its like a
  jigsaw without a cover image

46
Single particle EM reconstructions of a
ribosomemRNAtRNAPCC complex

• EM density rendered as a surface
• A, P, E the three transitional sites on the
  ribosome

E
P
A
30S ribosome
nascent chain
50S ribosome
SecYEG dimer
47
SecYEG X-ray structure fitted to EM density
SecYEG dimer fitted front-to-front
SecYEG dimer fitted back-to-back
Good fit
Poor fit
48
Ribosome X-ray structure fitted to EM density
P
A
49
EM shows how SecYEG and ribosome are connected
connecting loops
50
EM and X-ray provide a model for co-translational
insertion
51
Summary

• Protein folding pathways are complex both in
  vitro and in vivo.
• Structural resolution of the cellular machinery
  involved in protein folding and insertion is
  being resolved by electron microscopy and X-ray
  diffraction.
• EM is a powerful low resolution technique
• Complements and extends X-ray data particularly
  when applied to large complexes and membrane
  proteins

52
Investigations of Macromolecular Structure and
Function

• IV Protein mis-folding and deposition

53
Regulation of protein folding is critical
Dobson, Nature 2003 426 884 -890
54
What happens when folding goes wrong?

• Proteins can aggregate and form fibril structures
• Wild type (normal), truncated or mutant
• Insoluble
• Potentially auto-inducing
• Underlie many disease states
• All such diseases are called amyloidoses or
  b-fibrilloses

55
The amyloidoses

• What are they?
• What do they look like to the pathologist?
• What do they look like to the structural
  biologist
• What is happening at the level of protein folding
  and structure?
• Can we understand why?
• Can they be prevented?

56
Proteins and Amyloidoses
57
Creutzfeldt Jakob Disease

• Rare spontaneously occurring neuro-degenerative
  disorder
• Variant form caused (probably) by ingestion of
  bovine products from BSE infected cattle
• i.e. a transmissible form
• Infectious agent is not
• A bacterium
• A virus
• Infectious agent is a protein

58
Creutzfeldt Jakob Disease

• Prion protein (PrP).
• Soluble protein (GPI-linked) of no well
  understood function (k/o mouse has abnormal
  synaptic function)
• Protein is found deposited (i.e. insoluble) in
  CJD and vCJD.
• Injection of this form of the prion protein can
  cause recapitulation of the disease.

But what is happening at the level of protein
structure?
59
The amyloidoses

• What are they?
• What do they look like to the pathologist?
• What do they look like to the structural
  biologist
• What is happening at the level of protein folding
  and structure?
• Can we understand why?
• Can they be prevented?

60
Amyloid deposits in vivo look the same
Congo red stain. Positive red staining around the
large central artery
Diagnostic green birefringence under polarized
light.
61
The amyloidoses

• What are they?
• What do they look like to the pathologist?
• What do they look like to the structural
  biologist
• What is happening at the level of protein folding
  and structure?
• Can we understand why?
• Can they be prevented?

62
Amyloid deposits by EM look the same

• Long (mm)
• Unbranched
• Consistent diameter (10 nm)

(Makin, 2005)
63
Amyloid fibres are the same by EM

• EM is low resolution but this suggests that the
  overall architecture of fibres is a common one.
• 115 Å repeat in a direction parallel to that of
  the fibre axis
• What about the underlying atomic structure?
• If the fibre contains a repeating arrangement of
  proteins then it should diffract X-rays

115 Å repeat
64
Amyloid deposits look the same by X-ray
diffraction

• the fibres are aligned prior to being placed in
  the path of the X-ray beam.
• common diffraction pattern cross-b

4.8 Å meridional
fibre axis
ca. 10 Å equatorial
transthyretin
IAP
(Sunde, 1997)
65
The amyloidoses

• What are they?
• What do they look like to the pathologist?
• What do they look like to the structural
  biologist
• What is happening at the level of protein folding
  and structure?
• Can we understand why?
• Can they be prevented?

66
Tertiary structure does NOT determine
amyloidogenic ability.
Prion protein structure (PrP)
Transthyretin structure (Ttr)
67
What does the X-ray diffraction tell us?
4.8 Å meridional
fibre axis
ca. 10 Å equatorial
transthyretin
IAP

• There are three main structural repeat distances
  in the fibre
• 4.8 Å repeat in a direction parallel to that of
  the fibre axis
• 10 Å repeat in a direction perpendicular to the
  fibre axis
• 115 Å repeat in a direction parallel to that of
  the fibre axis

68
Fibre structure from diffraction

• 4.8 Å meridional . That is there is a repeating
  structural unit with a spacing of 4.8 Å, which is
  parallel with the fibre axis
• 4.8 Å is the spacing of b-strands in a b-sheet

4.7 Å
4.8 Å
69
Fibre structure from diffraction

• 10 Å equatorial perpendicular to the fibre axis
  
• spacing of b-sheets is apporximately 10 Å
• Exact repeat distance/reflection will reflect
  side-chain composition of the sheets

10 Å
10 Å
70
Model of protofilament

• repeat distance along fibre axis by EM 115 Å
• meridional reflection (b-strand separation) of
  4.8 Å
• suggests 24 b-strands per repeat of the fibre
• (115/4.8)
• b-sheets in all proteins are twisted
• twist of 15 per strand would produce a complete
  360 twist per repeat of the fibre (360/24)

71
Model of protofilament
protofilament axis
115 Å
protofilament end-on
4.8 Å
72
How many protofilaments in a fibre?

• Look at fibres end-on by single particle imaging
  electron microscopy.

average
average and apply symmetry
hundreds of images
4?
Serpell 1995
73
Model of amyloid fibrils
fibre axis
10 Å
74
Amyloidoses at the level of protein structure and
folding

• Non b-sheet proteins (as well as those inherently
  rich in b-sheets) convert into b-sheet rich
  fibrils
• What is the force that initiates this?
• What are the stabilising forces?

75
Current Model

• The molten globule is key
• The molten globule is a partially unfolded
  intermediate
• H-bond formation must be critical
• Hydrophobic effect must be involved
• But what drives the transition from a to b or
  from c to b?

Selkoe, Nature 2003 426 900-904
76
Even small peptides can form amyloid fibrils

• Seven amino acid
• GNNQQNY
• p-stacking of aromatics stabilises
• shows that the core element of a fibre could be
  very simplistic indeed

77
Support for model from CD

• Circular dichroism spectroscopy
• Absorbance spectroscopy
• Related to secondary structure
• Determine
• protein secondary structure percentages
• protein stability
• folding rates (e.g. Year III lectures)
• proteinligand interactions

78
CD theory (not difficult honest)

• chiral molecules will absorb left and
  right-handed plane polarised light (enantiomers)
• all relative to glyceraldehyde
• for small molecules (amino acids and sugars)
  predictable
• chiral molecules also absorb light that is
  circularly polarised
• not predictable
• left and right circularly polarised passed
  through a sample will be differentially absorbed

79
Circularly polarized light?
80
CD data analysis
Right circularly polarized

Left circularly polarized
if AbsR AbsL
if AR ? AL
Circularly polarized
Elliptically polarized Degree of ellipticity q
81
CD data analysis

• A ecl (Beer-Lambert Law)
• DA Decl
• q DA x 32982
• Ellipticity is a measure of protein secondary
  structure

82
Amyloidogenic mutations may access molten globule
state more easily

• CD of amyloidogenic lysozyme variants (causes
  systemic fatal amyloidosis)
• Wild-type is much more heat stable than variant
  Asp-67-His

WT
D67H
wild-type
amyloidogenic variant
Booth et al, Nature 1997 Nature 385 787-793
83
CD spectroscopy also shows increasing b-sheet

• Microtubule associated protein tau
• Forms paired helical filaments (PHFs) in
  Alzheimer's disease.
• Native protein relatively unstructured.
• Adopts ß-structure in the PHF mode

84
FT-IR spectroscopy can give comparable information

• Fourier Transform Infra-Red spectroscopy
• Bond stretching and bond angle vibration
• Motions on the 1013 Hz scale (pretty fast)
• l of 10-5 metres (infra-red) will be absorbed

85
Not all bonds stretch and bend at the same
frequency
Amide I C O stretch
Amide II N - H bend C - N stretch
86
FT-IR and secondary structure

• Amide I and Amide II exact frequencies are
  associated with local conformation i.e. 2ndary
  structure (e.g. dipole effects)
• Amide I _at_ 1650 - 1658 cm-1 a-helical
• Amide I _at_ 1620 1640 cm-1 b-sheet
• Fourier deconvolution of spectra can be used to
  assign 2ndary structure

87
FTIR spectroscopy of tau also shows increasing
b-sheet and fibre diffraction confirms classic
cross-b pattern

• Native protein relatively unstructured.
• Adopts ß-structure in the PHF mode

88
Current Model

• transition from a to b or from c to b could be
  driven by
• environmental factors (radicals, pH change)
• protein mutation
• prior presence of mis-folded protein

Selkoe, Nature 2003 426 900-904
89
References

• Sumner Makin, O and Serpell, L.C. (2005) Febs
  Journal 272 5950-5961 (amyloid fibre structures)
• Nature 426 (2003) p 884-900 (3 reviews on protein
  folding and unfolding)
• Saibil and Ranson (2002). TiBS 27 627-632 (EM
  and protein structure)