Williamson Folding And Design Flashcards
Anfinsen experiment
What did this prove
Why don’t all refolding?
Bovine ribonuclease A
1960s
Denaturant and reductant
Removal saw refolding
Some can’t because of pro peptides etc.
Assistance chaperones- PDI, PPI (proline isomerisation)
Folding vs aggregation
Inclusion body formation
Eqm between folded and unfolded
Unfolded protein response with hydrophobic patches
Ab initio prediction
Why is it tricky
Solution
The ‘protein folding problem’
Can we predict from sequence what it is going to do
Bioinformatics- Chou Fasman
Force fields for protein non covalent interactions too crude
Solution to thread sequences through known structures e.g. Phyre
Thermodynamics of protein folding
G = H - TS
Enthalpy- making bonds is exothermic, large negative when folding
Entropy- will be less positive as the protein folds
For folding to be favourable, G needs to be negative to be spontaneous
How many bonds broken to unfold?
About 1-5 bonds broken
This increases the positive entropy
Decreases the negative of the enthalpy
Means that delta G is less negative
3 main entropic contributions
Conformational entropy (-) Hydrophobic interactions (+, stop water ordering) Disulphide bonds (+ stop disorder)
1 main enthalpic contributions
Electrostatic interactions (large -)
Folding/ unfolding curves
Mid point of 4.5M urea
Transverse urea gradient electrophoresis
Unfolded state gives less migration during native page
Can also be done with CD, fluorescence, NMR shift change
Conformational stability from denaturation curves
Keq is calculated for each point in the TS (U/F)
This then means that G can be calculated for each point
G = -RTln Keq
G is then plotted against denaturation
The Y intercept gives Gh2o which is the protein stability
Levinthals paradox
Do proteins go through all conformations to reach lowest energy?
Only backbone, only phi and psi
2 conformations, only a and b in ramachandran
10^45 conformations
Would take 10^25 years for a protein to fold
The universe is 1.4 x10^10 yrs old
Folding energy landscape
Experiments observed in distinct and different intermediates
Unfolded protein
Falls into molten globules as energy decreases
Multiple different intermediate pathways
Eventually reaches either a folded or unfolded low energy state
Methods for studying proteins folding
Native page, sedimentation, size exclusion chromatography
Secondary by Far CD and D2O exchange
Tertiary by Fl, CD (near), NMR, ANS (dye to hydrophobic patch) and reverse phase HPLC
Name the 5 protein theories
Anfinsen spontaneous refolding Hydrophobic collapse- promote water disorder, entropic Framework model Nucleation growth Jigsaw model
Hydrophobic collapse
The protein folds to hide its hydrophobic patches
Decreases the order of water around hydrophobic regions
Makes this entropic ally favourable
Would make the reaction more spontaneous
Framework model
Secondary structures form independently
Then dock together
Anfinsen spontaneous folding
Protein spontaneous
Eqm between folded and unfolded states
Jigsaw model
Same puzzle can be formed from different starting pieces
So a number of different structures could form first
Nucleation growth
Something initially folds
Rest of folding happens around the nucleation
Bit like crystal folding
Trapping intermediates by disulphide bonds
BPTI has 3 S-S bonds
Can encourage folding by adding GSSG
Add iodoacetic acid at different points
This will iodoacetylate the free Cys residues
As time goes on, should be less iodoacetylation
Can analyse samples by ion exchange chromatography
Mass spec sequencing to see what order they fold in?
BPTI folding pathway
Which pathways does this support?
The peaks in absorbance of ion chromatography indicate the bonds forming
BPTI showed multiple folding pathways
14-38 bonds needs to be broken
Turned into 5-14 or 5-38 before -> 5-55
The 14-38 is then reformed last
Rate limiting step by isomerisation of bonds by PDI
Support for jigsaw?
Protein folding by mutagenesis studies
E.g. Mutate residue in helix to see if in transition state
Measure free energy change
If not folding in TS, then no difference apart from final energy higher
If is in the TS, then both TS and end G will be higher
Calculated by change in TS/change in NS
If 1, then is in TS
Study of CI2 protein inhibitor by energy analysis
Which model does this support?
Showed folding directly from denatured state
Secondary and tertiary formed at same time
All elements partially formed at TS
Nucleation CONDENSATION
Folding of CI2 in terms of energy diagram
Shown with A16, L49 and I57 in CI2 come together during folding
At the TS the entropic penalty rises fast than the enthalpic benefit
Means that at TS, G gets more positive
Nucleation condensation model
What does CI2 give a model for?
Weak local nucleus stabilised by a critical number of long range interactions
Large extended nucleus
Consolidation of the extended nucleus and structure occurs at the same time
CI2 a model for a FOLDON
D2O exchange to show protein folding
Saturated with d2O at ph 6, all D Folding by diluting in water Increase to pH 10 - this speeds exchange Only unfolded regions exchange Returned to pH 6 to continue folding HSQC spectra tuned for N, H Less H folded quicker, %H plotted for each SS element
Native state HX
Protein gradually denatured
At time points, proteins removed from denaturation and D exchange measured
Shows unfolding in stages
Each event is 2-state and cooperative
Series of smaller unfolding give a large unfolding
Each is a foldon
HX instead measured by mass spec
What is the advantage of this?
Protein Deuteration ph6
Left to fold for different time amounts
Then ph10 pulse
Unfolded will exchange
Gives QUENCHING ADVANTAGE pH 2.5 at 0 degrees
This can’t be done with NMR as a folded state is needed
Peptide digestion and ESI MS
H:D ration per peptide. Which parts of Seq folded first
Mass spec HX pulse of ribonuclease H1 results
Which model did this support?
As time increases, time for folding before pulse increases
Peptides will be heavier
Those that show the quickest m/z increase have folded first
Allows low concs
Folding occurs in 2-state stepwise manner and cooperative
Showed sequential foldons, supported nucleation condensation
Protein folding conclusion
First foldon random search
Foldons small to avoid levinthals paradox
First foldon is most stable, forms by NC
Large would face paradox, small foldons wouldn’t be able to overcome entropy barrier
What can delay protein folding?
How did this influence folding pathway models?
Aggregation Proline isomers Disulphide bonds Non native hydrophobic clustering Partial heme misligation
Likely that these events were mistaken for multiple folding pathways
So which models actually fit protein folding?
Framework extreme example of foldons then docking
Some evidence of hydrophobic collapse to form first foldon
Nucleation condensation generally accepted
Yamasaki 2013
Temperature jump NMR Heating and cooling system 46 -> 12 degrees Scans over 12s PCA Rate constant 0.2-0.7 s Use to measure refolding
Proline isomerisation
Prolyl peptide bond must be right
30% in cis form
Protein with one cis-Pro will show a fast and slow rate
Slow needs to be converted from trans first
Double jump experiment
Cis protein denatured and refolded with and without PPI
Enzyme increases folding rate
Disulphide bond formation
Active site for PDI has 2 Cys residues
Can bind to free Cys and catalyses bonds
This oxidises (joins) the protein bonds and reduces the PDI
Can also catalyse the reduction of a non native bond to replace it (e.g. BPTI pathway)
Scrambled ribonuclease assay
Where is PDI found?
Denature ribonuclease with urea + reductant (DTT)
Oxidation and urea removal -> INACTIVE
Add more reductant to undo the bonds
Then add PDI
This time bonds reform in right places -> ACTIVE
converts RNA -> ribonucleotides
Increase in A260 nm
PDI found in ER as bonds are in secreted proteins
Structure of PDI
Two a and two b domains
U shape
Cleft between alpha is active site
(a, a’, b, b’)
How is in vivo folding different to in vitro?
1 mg/ml vs. 340
In the cells protein can form whilst on ribosome, different to denaturation experiments
Molecular chaperones
Macro molecular crowding
4 different crowding agents on refolding on lysozyme
Drops refolding yield due to interference
Favours aggregation
Increases rate of hydrophobic collapse
Ovalbumin, BSA, Ficoll 70 and Dextran 70
Co translational folding
Synthesis may be slower than folding
Ras and DHFR domains linked by linker region to be flexible
Measured Ras and DHFR activity
In vitro poor when fused
In e.coli less than 2% folding of fused protein
Rabbit reticulocyte lysate- 90% effective folding
Co translational folding prevents domain interference
Molecular chaperones
Small hsps- no ATP use, may need help dissociating
Chaperones- hsp40, hsp70. Need ATP.
Chaperonins- hsp60, GroEL
The 3 pathways of chaperone folding
Hsp70/60 independent- trigger factor protects
Hsp70 - shield hydrophobic patches
Hsp60- sequestered by groEL / groES
Trigger factor
Domains
Unbound is a monomer or dimer
Interacts with folded proteins or vacant ribosomes
Association accelerated by hydrophobic patch
Might remain associated after its dissociation from ribosomal binding site
PPIase, C terminal and Ribosome binding
Open groove along whole protein to increase surface area
Liu et al- PPIase no activity, but assists folding of protein to crevice
Hsp70
Function
Structure
Binds and stabilises nascent peptides
ATP for release
Protein may fold on release or be transferred to other protein
Important during heat shock response
Opening and closing by allosteric reactions upon ATP binding
NBD and peptide binding domain with alpha helical lid
Hsp70 cycle
ATP increases affinity for nascent peptides
Hydrolysis closes lid
Undo incorrect folding to encourage native structure
Release of ADP is assisted by GrpE, Hsp110 etc.
Protein is either folded or unfolded
Experiment to determine hsp70 specificity
13 AAs peptide that overlap Cross link to membrane Add chaperone to bind Wash Western blot with antibodies Mao the binding sites onto native protein- specific for buried sites
GroEL/ES
2 heptamer rings make up GroEL
Binds protein by hydrophobic interactions
Binds of ATP causes elongation and opening of ring cavity
GroES lid binds
Replaced by hydrophilic surface, forces protein to fold by entropy
ATP hydrolysis
Affinity for groES lost, affinity of trans domain for protein lost
There is cooperative binding in the two chambers
90 degrees out of phase
3 domains of groEL
L234, L337, V259 L263 V264
Make hydrophobic patch
Polar residues moved to surface during conformational change
Apical domains moves up
Intermediate and Equatorial
Why design new proteins?
New and novel catalytic activities
Improve existing
New structures
3 general methods to design new proteins
Rational design
Directed evolution
Catalytic antibodies
Rational design
Prior knowledge
Modifies existing or
De novo- new site on existing scaffold or new scaffold
Modifying existing site
Lactate dehydrogenase -> malate dehydrogenase
PPI converted into protease by inserting catalytic triad
Gave a good rate enhancement 10^8
But had a low Km value
Meant that Km/kcat wasn’t large, inefficient
De novo design of new site on existing scaffold
Making thioredoxin cleave PNPA His catalytic nucleophile, less stable intermediate to rapid hydrolysis Computational search of surface Two mutants made: F12H, Y70A the other contained L17H too Very poor catalysts
Kemp elimination
Reaction
2 designs
Ring opening of benzisoxazole by a base
Designed in silico
Glu base to abstract proton from carbon
1: Glu, Phe for pi stacking, Lys H bond donor
2: Asp:His dyad, Trp stacking and Ser
3 aims of creating kemp enzyme
Base to abstract proton from carbon
Hydrogen bond donor to stabilise -ve charge on oxygen
Stabilisation of planar state by pi stacking
Results of using TIM barrel for kemp reaction
2 designs
2 mutants designed to prove mechanism
Big kcat/km is efficient Mutant 1: 2.5 x10^5 2: 1.4 x10^5 Good enzyme is around 10^6 Further improved by direct evolution E->Q mutant removes activity Asp makes His better nucleophile, so removal lowers activity
Coiled coils for de novo design
Self assembled cage particle
Coiled coils used as specificity between coils
Heterodimers and homotrimers
Folding driven by need to maximise coiled coil interactions
Sheet curves
Can attach enzymes etc. To inside or outside
Directed evolution
No prior knowledge
Fitness landscape- islands of function within sequence space
Mutations to reach ‘peak’ fitness
Random mutagenesis induced by error prone PCR
Success dependent on screen or selection process
Need both for accuracy and high throughput
Better to screen for catalysis than substrate binding
In vivo screen
dsRED
Fluorophore will slow half time
Plated and most red colony after 24h selected and rein put into PCR
Managed to get short half life, but the relative brightness decreased
E x Quantum yield
Shows that need to be conscious of other parameters changing
In vitro screen
Protein of interest fused to g3p on protein surface
Gene for protein of interest in tandem with g3p to display
Easy to screen for activity to immobilised substrate
Isolate and amplify the phages that stick
Remove metal cofactors to allow binding but not catalysis e.g. Zn
Eluted by adding cofactor again
Smart libraries
High throughout screens can’t always be done
Smaller libraries with high hit frequency
Targeting sites of substrate binding by mutations
But this needs prior knowledge
Saturation mutagenesis of these sites can be combined with cycles of random mutagenesis
Combined rational design and direction evolution
Kemp reaction again
7 rounds of directed evolution
Catalyst with 200x increase in kcat/Km
Most effective was R7 mutant
Analysis of R7 mutant for the Kemp reaction
Combined rational design and directed evolution
Catalytic E101 and K222 H donor not changed
Changes in adjacent residues for fine tuning
Some hydrophobic -> polar changes to hold K222 in correct position?
Catalytic antibodies
How they work
How to make them
Abzymes
Only prior knowledge of mechanism needed
Lowers TS energy by stabilising, or destabilising substrate
Prepare antibody to haptenic resembling TS
Introduces strain
Covalently link haptene to carrier protein to induce Ab formation in B cells
B cells + myeloma -> hybridoma cell line
Ester hydrolysis by Abzymes
Incoming OH around 0.6 A longer than TS
Needs to be stable enough
Not completely accurate
Success depends on how good the approximation is
Chelatase mechanism by antibodies
Antibody raised against n-methylmesoporphyrin
The alkylation of the pyrrole nitrogen distorts the ring
Antibody has a Zn insertion rate similar to enzyme
3 applications of novel antibodies
Novel catalysis of useful reactions e.g. Aldolase antibody for synthetic chemistry
Drug clearance- antibody for cocaine overdose, stable for 3 weeks, recycled unlike conventional antibodies
Pro drug activation- variation on ADEPT. Used instead of enzyme, avoids risk of immunogenicity which would limit repeat administrations.
One arm -> cancer antigen, other arm activates pro drug -> toxic drug
How good are antibodies?
Best antibodies approach least efficient enzymes
Limitations of Abzymes
Improvement strategies 3
Imperfect TSA
Lack of catalytic machinery
Rigid compared to enzymes
Mimics TSA for uncatalysed reaction
SDM to add catalytic residues, needs high res structure of complex
Direction evolution to improve catalytic efficiency
Hyper variable regions to increase flexibility