Lecture 11

how find structure?

1. PDB - 80,000 structures
2. x-ray crystallography - not simple, can't do within a week, cost time money effort, and no guarante to get structure
3. make one by using homology modeling, if lucky enough

• but novel protein? amyloid beta doesn't crystallize
• use a template - don't need to invent wheel all the time!
• the template is the homologous protein for which we already know structure

how claim 2 structures are similar?

• RMSD - root mean squared division
• structurally align as much as possible
• then measure distance
• homework - calculate RMSD - "c-alpha" "all atom rmsd"

structure building

• how to use template to model homologous sequence
• sequence alignment - like 85% done
• backbone chain optimization - change template to account for gap - done by energy minimization technique
• then optimize side chain
• then evaluate how correct your projection is through theoretical ways and experimental ways - USE BOTH
• making model is easy - validating is not easy
  • theoretical: hydrophobic stays on one side, phillic on the other

troubleshooting

• protein quality check software, molecular dynamics, monte carlo, energy minimization
• if homology low - fundamental problem

journal club

• conformational sampling - building reliable models - THE BOTTLENECK
• computational cost increases dramatically with chain length
• longer aa, becomes longer beads on a string instead of lobular, can break up into modules and
• rosetta algorithm - crowd sourcing to figure out inter domain interaction, David Baker 2011 "Crowdsourcing to write algorithms"

CASP - protein structure prediction competition

• homology modelling has thus far beaten energy minimization every time
• protein docking capri
• most accurate = lowest energy - free energy scatter plots
• refined natives = blue points
• energy landscape

protein folding

• achieve most thermodynamically stable state
• leviathan paradox
• folded state goes through local intermediates but eventually reach global minimum
• lattice model
  • beads in a lattice
  • off-lattice model
• all atom model

S6 folding

• energy landscape
• transition state
• enthalpy - heat released
• entropy
• gibbs free energy
• phi-value - minimize this thermodynamic state

disulfide bonds tendamistat

• covalent interact btw residuues containing sulfide
• order of oxydation and reduction - order matters in way it folds
• limits folding space to make fast? kinetic roadblocks of transitional intermediate bonds to make slow?
• "go-model" simulation
• graph showing which residues interact with which

=beta hairpins

bbl folding

• type 0 folding no high free energy barrier - has one minimum, no high activation, no energy barrier
• type 1 - high free enrgy barrier, two state
• CL2 - well known protein that has two state

chaperonin assisted folding

• chaperones = GroEL protein
• makes crowded state around protein
• in vivo folding - geometrically confines - accellerates folding folding, lowering entropy
  • but crowdined environment retards folding
• so try folding in a cylinder

protein self interaction

• discrete molecular dynamics
• discrete models of potentials, reduces equations to set of algebraic eqs
• reduces computational intensity
• protein tested EAK peptide
• electrostatic interaction
• discreteMD - reduced computation for molecular dynamics
• 1 femtosecond = delta t to calculate energy = model interaction over 1 MILIsecond - takes months - discreteMD reducees by 2 ords magnitute

pathway

• alpha helix
• beta helix - rate limiting
• zinc-binding - plays significant role - stablizes interactions
• "molecular machines"