nmr random coil index & protein dynamics
TRANSCRIPT
NMR Random Coil Index
& Protein Dynamics
Mark Berjanskii Edmonton, Alberta
November 19th, 2013
Outline
• RCI background
• Backbone RCI development
• Backbone RCI application
• Side-chain RCI
• Side-chain RCI application
RCI background
• What is protein dynamics?
• Why study protein dynamics?
• Methods of studying fast
equilibrium motions
• RCI history and basics
Proteins in textbooks are
shown as static structures
Proteins are dynamic
Prion protein
Dynamic pictures Static picture
Stevens DJ, Walter ED, Rodríguez A, Draper D, et al. (2009)
PLoS Pathog 5(4): http://www.rcsb.org/pdb/explore.do?structureId=2k1d
Equilibrium and non-equilibrium
dynamics
• Equilibium dynamics - thermal motions, happens in thermal
equilibrium state
• Non-equilibium dynamics - happens when protein is displaced
from its equilibrium state by an event (e.g. ligand binding, etc)
http://www.ch.cam.ac.uk/person/pjb91
Papoian G A PNAS 2008;105:14237-14238
Equilibrium motions are involved in protein
function
Borrowed from http://www.itqb.unl.pt/labs/protein-modelling/activities/upseknas
Ligand binding by serine protease subtilisin
Non-equilibrium motions are important too
Borrowed from http://www.itqb.unl.pt/labs/protein-modelling/activities/abc-transporters
The ATP-binding cassette (ABC) transporters change
conformations upon ATP binding and hydrolysis
Time-scales of protein motions
Functional aspects of protein flexibility. Teilum K, Olsen JG, Kragelund BB.
Cell Mol Life Sci. 2009 Jul;66(14):2231-47
Today we discuss only
fast equilibrium motions
Role of fast motions in thermodynamics
of protein-ligand interactions
Association and dissociation
constants:
Standard Gibbs free energy:
Molecular modelling,
rational drug design
Largest contributions Random Coil Index
What do we want to know about
a motion of a protein group?
• How fast? (time-scale, frequency)
• How far? (amplitude)
• Where does it move? (direction)
Methods of studying fast
equilibrium dynamics
X-Ray B-factor
Sources of B-factor uncertainty
- multiple conformations in different unit cells (internal static disorder),
- protein model errors,
- the contribution of more than one atom to a particular electron density,
- intermolecular crystal packing contacts,
- low temperature of X-ray experiments,
- non-equilibrium conformations
- the degree to which the electron density is spread out,
- uncertainty for each atom position
High uncertainty
Low uncertainty
mean squared displacement
B-factor info content:
Ampliutude: YES
Time-scale: No
Direction: No
Disorder in NMR ensembles (per residue RMSD)
• Number of restraints (harder to get for mobile regions)
• Model convergence (depends on length of TAD and CD steps)
• Number of structures in NMR ensemble
• Bias in structure-based assignment of ambiguous NOEs
• Unrealistic force-field (no water, no electrostatic term,
simplified Van der Waals interactions)
• Inclusion of unrecognized spin-diffusion NOEs
• Requires full structure determination
NMR RMSD info
content:
Ampliutude: YES
Time-scale: No
Direction: No
Protein motions can be visualized by Molecular Dynamics
simulations
Correlated changes in hydrogen bonds:
Comparison of MD conformations
with active and inactive
conformations of J domain
MD info content:
Ampliutude: YES
Time-scale: YES
Direction: YES
MD disadvantages:
1) Theoretical method
2) Requires a 3D structure
3) Depends on the quality of a force-field and starting structure
4) Time-consuming (weeks – months)
5) Does not fully explore conformational space
6) Can characterize only fast motions (<100-1000 ns)
MD solves Newton's equation of motion:
Fi=miai
Fast protein dynamics by NMR relaxation 15N relaxation experiments Amplitude and time-scale of motions
Model-free analysis
Disadvantages:
1) Requires a 3D structure
2) Poor separation of rates of overall tumbling and internal
dynamics for ns-motions
3) Sensitive to NMR signal intensity
4) Opposing effects arising from motions on different time-
scales for both transverse and longitudinal rates.
5) Insensitive to motions that do not change orientation of N-H
vector with respect to external magnetic field.
NMR relaxation info content:
Ampliutude: YES
Time-scale: YES
Direction: No
Random Coil Index
NMR signals from groups in protein random coils
are located in particular rigions of NMR spectra
http://www.biochem.ucl.ac.uk/~rharris/BCSB/NMR/hsqc.html
Folded protein Unfolded "random coil" protein
Random Coil
Chemical Shifts
What are random coil chemical
shifts? -> Statistical coil chemical shifts
Originate from an energy-weighted ensemble of conformations, in which a single residue
(1) is not involved in non-local interactions with other residues and
(2) can occupy any of the regions of the Ramachandran plot with a certain probability specified by the Boltzmann distribution
Unblocked statistical-coil tetrapeptides and pentapeptides in aqueous
solution: theoretical study. Jorge A. Vila, Daniel R.Ripoll, Hector A. Baldoni & Harold A. Scheraga. Journal of Biomolecular NMR, 24: 245–262, 2002.
BIG old idea
• Random coils are highly flexible
• If a residue has chemical shifts close to
random coil chemical shifts (=small
secondary shifts), it should be very
flexible too.
What are secondary chemical
shifts?
Difference between experimental chemical
shifts and random coil chemical shifts
Your protein
chemical shifts
Reference random coil
chemical shifts
- = Secondary
Chemical
Shifts
Random Coil Index
Random coil chemical
shifts
Secondary chemical shifts
Measured
chemical shifts
Random Coil Index
Random Coil Index definition
Random Coil Index (RCI) – evaluates the proximity of residue structural
and dynamic properties to the properties of flexible random coil rigions
from NMR chemical shifts
Outline
• RCI background
• Backbone RCI development
• Backbone RCI application
• Side-chain RCI
• Side-chain RCI application
Not as easy as it
sounds
The first program "Dynamr" has
failed
Sequence corrected random coil
Ca chemical shifts Experimental Ca chemical shifts
|DdCa|
Double smoothing via three-
point moving average
K-DdCa
CSI-based scaling of (K-DdCa ) for helices and
b-strands
Dynamr worked for helical proteins
and failed to distinguish a b-strand from coil
Why did Dynamr fail?
• Ca secondary chemical shifts are small in b-
strands.
S. Spera and A. Bax: An empirical correlation between protein backbone conformation
and Ca and Cb chemical shifts.
J. Am. Chem. Soc. 113, 5490-5492 (1991).
Problem with false positives
PyJ
Possible solution? • use more than one chemical shift
• find a better mathematical expression
• optimize the way of combining
chemical shifts into a single parameter
RCI development
Testing various
empirical equations
Generating reference
flexibility profiles
by MD
Optimization of
empirical equations Method testing
and validation
Eight empirical equations have been
tested ADdCa+BDdCb+CDdHa+DDdCO+EDdHN+FDdN
A(DdCa)-1+B(DdCb)-1+C(DdHa )
-1+D(DdCO )-1+E(DdHN )
-1 +F(DdN )
-1
(DdCa)A+(DdCb)B+(DdHa )
C+(DdCO )D+(DdHN )
E +(DdN )
F
(DdCa)A (DdCb)B (DdHa )
C (DdCO )D (DdHN )
E (DdN )F
e(DdCa)A e(DdCb)B e (DdHa )C e (DdCO )D e (DdHN )E e (DdN )F
e(DdCa)A + e(DdCb)B + e (DdHa )C + e (DdCO )D + e (DdHN )E + e (DdN )F
Ae(DdCa) + Be(DdCb) +Ce (DdHa ) +De (DdCO ) +Ee (DdHN ) +Fe (DdN )
|A(DdCa)+B(DdCb)+C(DdHa )+D(DdCO )+E(DdHN )+F(DdN )|-1
Generation of reference
flexibility profiles by MD
Too many proteins, too little time
• os.system('xmgrace %s -hdevice PostScript -hardcopy -printfile %s.ps' % (data,data))
• os.system('xmgrace %s -autoscale none -hdevice PostScript -hardcopy -printfile %s.ps' % (data,data))
• os.system('xmgrace %s -autoscale y -world 0 0 %s 1 -hdevice PostScript -hardcopy -printfile %s.ps' % (data,L_last_residue,data))
• os.system('g_analyze -f %s -b %s -e %s > %s/%s' % (output,L_analysis_start,L_analysis_end,RMSF_folder,"logfile"))
• os.system('rm %s/logfile' % RMSF_folder)
• os.system('rm %s' % (N_index))
• os.system('rm %s' % (output))
• os.system('g_analyze -f %s -b %s -e %s > %s/%s/%s' % (output,L_analysis_start,L_analysis_end,RMSF_folder,G_MSD_folder,"logfile"))
• os.system('mv %s/%s/logfile logfile.bak' % (RMSF_folder,G_MSD_folder))
• os.system('rm %s' % (N_index))
• os.system('rm %s' % (output))
• os.system('rm %s' % NH_index)
• os.system('mv %s %s' % (energy_log,new_name_for_log))
• os.system('mv %s %s' % (output,new_name_for_plot))
• os.system('g_chi -s %s -f %s -corr %s/%s/%s_order.xvg -psi -phi -o %s/%s_order.xvg -p %s/%s_order.pdb -g %s/%s -shift -b %s -e %s' % (L_topology_file,L_traj_name,L_output_directory,L_output_directory_order,L_output_file,L_output_directory,L_output_file,L_output_directory,L_output_file,L_output_directory,L_output_file,L_start,L_end))
• os.system('rm *"#"*')
• os.system('rm -r %s' % RMSD_dir)
• os.system('mkdir %s' % RMSD_dir)
• os.system('rm -r %s' % Energy_dir)
• os.system('mkdir %s' % Energy_dir)
• os.system('rm -r %s' % RGYR_dir)
• os.system('mkdir %s' % RGYR_dir)
• os.system('rm -r %s' % NH_order_dir)
• os.system('mkdir %s' % NH_order_dir)
• os.system('rm -r %s' % DIH_dir)
• os.system('mkdir %s' % DIH_dir)
• os.system('g_rama -f %s -s %s -b %s -e %s -o %s/%s' % (traj_name,topology_file,analysis_start,DIH_analysis_end,DIH_dir,traj_name))
• os.system('g_filter -f %s -s %s -nf %s -oh %s_high_pass_%s.xtc -ol %s_low_pass_%s.xtc' % (traj_fit_name,topology_file,filter_value,Traj_name_no_ext,filter_value,Traj_name_no_ext,filter_value))
• os.system('rm -r %s/%s' % (RMSF_dir,G_MSD_dir))
• os.system('rm -r %s' % RMSF_dir)
• os.system('mkdir %s' % RMSF_dir)
• os.system('mkdir %s/%s' % (RMSF_dir,G_MSD_dir))
• os.system('rm -r %s' % NH_order_dir_filter_off)
• os.system('mkdir %s' % NH_order_dir_filter_off)
• os.system('rm -r %s' % NH_order_dir_filter_on)
• os.system('mkdir %s' % NH_order_dir_filter_on)
• os.system('rm -r %s/%s' % (G_CHI_dir_filter_off,DIH_order_dir))
• os.system('rm -r %s' % G_CHI_dir_filter_off)
• os.system('rm -r %s' % G_CHI_dir)
• os.system('mkdir %s' % G_CHI_dir)
• os.system('mkdir %s' % G_CHI_dir_filter_off)
• os.system('mkdir %s/%s' % (G_CHI_dir_filter_off,DIH_order_dir))
• os.system('rm -r %s/Filter_%s' % (Shift_dir,filter_value))
• os.system('rm -r %s' % (Shift_dir))
• os.system('mkdir %s' % Shift_dir)
• os.system('mkdir %s/Filter_%s' % (Shift_dir,filter_value))
• os.system('mv corrpsi*.xvg %s/%s/.' % (G_CHI_dir_filter_off,DIH_order_dir))
• os.system('mv corrphi*.xvg %s/%s/.' % (G_CHI_dir_filter_off,DIH_order_dir))
• os.system('mv histo-psi*.xvg %s/%s/.' % (G_CHI_dir_filter_off,DIH_order_dir))
• os.system('mv histo-phi*.xvg %s/%s/.' % (G_CHI_dir_filter_off,DIH_order_dir))
• os.system('rm -r %s/%s' % (G_CHI_dir_filter_on,DIH_order_dir))
• os.system('rm -r %s' % (G_CHI_dir_filter_on))
• os.system('mkdir %s' % (G_CHI_dir_filter_on))
• os.system('mkdir %s/%s' % (G_CHI_dir_filter_on,DIH_order_dir))
• os.system('rm -r %s/Filter_%s' % (Shift_dir,filter_value))
• os.system('mkdir %s/Filter_%s' % (Shift_dir,filter_value))
• os.system('mv corrpsi*.xvg %s/%s/.' % (G_CHI_dir_filter_on,DIH_order_dir))
• os.system('mv corrphi*.xvg %s/%s/.' % (G_CHI_dir_filter_on,DIH_order_dir))
• os.system('mv histo-psi*.xvg %s/%s/.' % (G_CHI_dir_filter_on,DIH_order_dir))
• os.system('mv histo-phi*.xvg %s/%s/.' % (G_CHI_dir_filter_on,DIH_order_dir))
• os.system('rm *"#"*')
os.system('xmgrace %s -hdevice PostScript -hardcopy -printfile %s.ps' % (data,data))
os.system('xmgrace %s -autoscale none -hdevice PostScript -hardcopy -printfile %s.ps' %
(data,data))
os.system('xmgrace %s -autoscale y -world 0 0 %s 1 -hdevice PostScript -hardcopy -printfile
%s.ps' % (data,L_last_residue,data))
os.system('g_analyze -f %s -b %s -e %s > %s/%s' %
(output,L_analysis_start,L_analysis_end,RMSF_folder,"logfile"))
os.system('rm %s/logfile' % RMSF_folder)
os.system('rm %s' % (N_index))
os.system('rm %s' % (output))
os.system('g_analyze -f %s -b %s -e %s > %s/%s/%s' %
(output,L_analysis_start,L_analysis_end,RMSF_folder,G_MSD_folder,"logfile"))
os.system('mv %s/%s/logfile logfile.bak' %
(RMSF_folder,G_MSD_folder))
os.system('rm %s' % (N_index))
os.system('rm %s' % (output))
os.system('rm %s' % NH_index)
os.system('mv %s %s' % (energy_log,new_name_for_log))
os.system('mv %s %s' % (output,new_name_for_plot))
os.system('g_chi -s %s -f %s -corr %s/%s/%s_order.xvg -psi -phi -o %s/%s_order.xvg -p
%s/%s_order.pdb -g %s/%s -shift -b %s -e %s' %
(L_topology_file,L_traj_name,L_output_directory,L_output_directory_order,L_output_file,L_output_directory,L_output_file,L_output_dir
ectory,L_output_file,L_output_directory,L_output_file,L_start,L_end))
os.system('rm *"#"*')
os.system('rm -r %s' % RMSD_dir)
os.system('mkdir %s' % RMSD_dir)
os.system('rm -r %s' % Energy_dir)
os.system('mkdir %s' % Energy_dir)
os.system('rm -r %s' % RGYR_dir)
os.system('mkdir %s' % RGYR_dir)
os.system('rm -r %s' % NH_order_dir)
os.system('mkdir %s' % NH_order_dir)
os.system('rm -r %s' % DIH_dir)
os.system('mkdir %s' % DIH_dir)
os.system('g_rama -f %s -s %s -b %s -e %s -o %s/%s' %
(traj_name,topology_file,analysis_start,DIH_analysis_end,DIH_dir,traj_name))
os.system('g_filter -f %s -s %s -nf %s -oh
%s_high_pass_%s.xtc -ol %s_low_pass_%s.xtc' %
(traj_fit_name,topology_file,filter_value,Traj_name_no_ext,filter_value,Traj_name_no_ext,filter_value))
os.system('rm -r %s/%s'
% (RMSF_dir,G_MSD_dir))
os.system('rm -r %s' % RMSF_dir)
os.system('mkdir %s' % RMSF_dir)
os.system('mkdir %s/%s' % (RMSF_dir,G_MSD_dir))
os.system('rm -r %s' % NH_order_dir_filter_off)
os.system('mkdir %s' % NH_order_dir_filter_off)
os.system('rm -r %s' % NH_order_dir_filter_on)
os.system('mkdir %s' % NH_order_dir_filter_on)
os.system('rm -r %s/%s' % (G_CHI_dir_filter_off,DIH_order_dir))
os.system('rm -r %s' % G_CHI_dir_filter_off)
os.system('rm -r %s' % G_CHI_dir)
os.system('mkdir %s' % G_CHI_dir)
os.system('mkdir %s' % G_CHI_dir_filter_off)
os.system('mkdir %s/%s' % (G_CHI_dir_filter_off,DIH_order_dir))
os.system('rm -r
%s/Filter_%s' % (Shift_dir,filter_value))
os.system('rm -r %s' %
(Shift_dir))
os.system('mkdir %s' % Shift_dir)
os.system('mkdir %s/Filter_%s' % (Shift_dir,filter_value))
os.system('mv corrpsi*.xvg %s/%s/.' % (G_CHI_dir_filter_off,DIH_order_dir))
os.system('mv corrphi*.xvg %s/%s/.' % (G_CHI_dir_filter_off,DIH_order_dir))
os.system('mv histo-psi*.xvg %s/%s/.' % (G_CHI_dir_filter_off,DIH_order_dir))
os.system('mv histo-phi*.xvg %s/%s/.' % (G_CHI_dir_filter_off,DIH_order_dir))
os.system('rm -r %s/%s' %
(G_CHI_dir_filter_on,DIH_order_dir))
os.system('rm -r %s' % (G_CHI_dir_filter_on))
os.system('mkdir %s' % (G_CHI_dir_filter_on))
os.system('mkdir %s/%s' %
(G_CHI_dir_filter_on,DIH_order_dir))
os.system('rm -r %s/Filter_%s' % (Shift_dir,filter_value))
os.system('mkdir
%s/Filter_%s' % (Shift_dir,filter_value))
os.system('mv corrpsi*.xvg %s/%s/.' %
(G_CHI_dir_filter_on,DIH_order_dir))
os.system('mv corrphi*.xvg %s/%s/.' %
(G_CHI_dir_filter_on,DIH_order_dir))
os.system('mv histo-psi*.xvg %s/%s/.' %
(G_CHI_dir_filter_on,DIH_order_dir))
os.system('mv histo-phi*.xvg %s/%s/.' %
(G_CHI_dir_filter_on,DIH_order_dir))
os.system('rm *"#"*')
MD set-up MD analysis
GromPy A: a Python wrapper for Gromacs.
Generation of trajectories.
Protein structure (PDB format) Secondary structure info (GromPy format)
Simulation set-up. (1) Temperature (2) Force-field (3) Water type (4) Water box
(5) Flags and length of MD simulation steps (6) Time-step, etc.
Simulation:
- Conversion of pdb file to Gromacs format and setting up restraints
- Identification of total protein charge
- Protein solvation
- Minimization of protein and water
- Setting the type and number of ions and adding them to the system
- Restrained minimization with decreasing strength of protein restraints
- MD equilibration of water and ions
- MD equilibration of water, ions and protein side-chains
- MD equilibration of water, ions, protein side-chains and flexible backbone
- MD production run
Trajectories, energy file, RMSD for every step, secondary structure file
GromPy B: a Python wrapper for Gromacs. Analysis and
conversion of trajectories.
GromPy A output files
Analysis and conversion set-up:
(1) Part of trajectory to analyze, (2) Names of output directories, (3) Flags for
trajectory conversion and analysis steps, etc
Trajectory conversion:
Compact trajectory, PDB files, water removal
Analysis:
Stability of simulation: Time series for (1) Energies, (2) RMSD, (3)
Radius of gyration.
Dynamics parameters: (1) NH order parameters, (2) y and f order
parameters, (3) RMSF per residue, (4) Fluctuations of y and f
Pictures: Postscript and Grace formats
Coefficient optimization Experimental chemical
shifts
Random coil chemical
shifts MD results
|DdCa|,|DdCb|,|DdHa|,|DdCO|,|DdHN|,|DdN|
Three-point moving average
smoothing
Correlation
analysis Calculating RCI with all combinations of A,B,C,D,E,F:
|A(DdCa)+B(DdCb)+C(DdHa )+D(DdCO )+E(DdHN )+F(DdN )|-1
Sorting combinations of A,B,C,D,E,F
based on correlation coefficients
Average each type of coefficients for the
combinations with the best correlations
Extracting weighting coefficients
Combination number
Co
rrela
tio
n c
oeff
icie
nt
Co
rre
lati
on
co
eff
icie
nt
0.007
Current RCI expression RCI = Q * |A(DdCa)+B(DdCb)+C(DdHa )+D(DdCO )+E(DdHN )+F(DdN )|
-1
Q – is a scaling coefficient that makes the sum of
weighting coefficients identical for all cases of
incomplete assignments.
RCI protocol Experimental shifts: Ca, Cb, CO, N, HN, Ha Random coil values of chemical shifts
Reference correction by RefCor Neighboring residue correction, i+1, i-1, i+2, i-2
Calculation of secondary chemical shifts for Ca, Cb, CO, N, HN, Ha
Combining secondary chemical shifts using the RCI equation.
Random Coil Index
Filling gaps by averaging Dd of residues i+1, i-1 or i+2, i-2
Smoothing by three-point moving averaging
Hertz, Extreme & End-Effect corrections
Smoothing by three-point moving averaging
Before and after the RCI treatment
PyJ
2 3
1
Occurrence of several random
coil shifts in loops is higher
PyJ
2 3
1
1 2 3
Nuclei used: CA, CB, CO, N, NH, HA
RCI performance
RCI info content:
Ampliutude: YES
Time-scale: No
Direction: No
Who cares?
Outline
• RCI background
• Backbone RCI development
• Backbone RCI application
• Side-chain RCI
• Side-chain RCI application
RCI applications
RCI validation of NMR
RMSD
RCI correlation with NMR RMSF of
1D3Z – 0.60
1XQQ – 0.75
1ILF
2JHB
CORE BINDING FACTOR BETA
2JHB 1ILF
1D3Z 1XQQ
Ubiquitin
RCI can be used to detect over- and under-
restrained regions in NMR ensembles
RCI validation of MD
simulations
MD validation by RCI Chicken prion
RCI vs. Model-Free order paramter
How can RCI supplement Model-Free S2 ?
- Unfolded proteins: structure is needed for model-free
- Large proteins with overlapped signals: the error in peak position is smaller than in peak intensity
- Proteins at high pH: NHs are weak, needed for MF, not needed for RCI
- Proteins with intermediate exchange: peaks are too weak to measure relaxation, but not shifts
- Multi-domain proteins: overall motion is too complex for MF, not for RCI
- Poor separation between overall and internal motions on ns time-scale
Unlike MF, RCI is not sensitive to the overall motions
RCI can assess mobility where NMR relaxation S2 can't
(Nuclease)
RCI agrees with MD RMSF for the intermediate
exchangeregin in nuclease
RCI of large multi-domain proteins
32kD HIV-1 Gag (283 aa)
HIV-1 Gag
Correlation coefficient - 0.74
RCI of partially unfolded proteins
Octa-peptide
repeats
RCI webserver and stand-
alone program
RCI web. server
http://wishart.biology.ualberta.ca/rci
http://www.randomcoilindex.com
RCI web. server
RCI web. server
RCI stand-alone version http://wishart.biology.ualberta.ca/download/rci/
RCI impact on protein NMR
community
Backbone RCI publications
Impact factor: 11
Citations: 125
Impact factor: 3
Citations: 43
Impact factor: 8
Citations: 33
Impact factor: 8
Citations: 39
Total number of citations is by 2013 is 240
RCI implementation in other programs
Ad Bax's group
NIH
Bethesda, MD
US
RCI of apomyoglobin molten
globule
(B) Secondary structure and RCI-S2 parameter for the transient MG state and
the N state. The secondary structure of the MG state was predicted by TALOS+
(30) from An external file that holds a picture, illustration, etc.
The rectangles depict the location of helical structure in each state; the
thickness of each rectangle is proportional to the population of helix. The
hatched lines indicate the small population of transient helical structure in the
C- and E-helix regions of MG.
(C) Changes in secondary structure accompanying the N[left arrow over right
arrow]MG transition, mapped to the structure of holoMb. Residues predicted
to be helical by TALOS+ are red. The population of helix in the MG ensemble is
indicated by the tube radius, with a larger radius indicating higher population.
Flexible regions with RCI S2 < 0.7 are blue, and coil regions with S2 > 0.7 are
green.
Measurement of protein unfolding/refolding kinetics and structural characterization of hidden intermediates by NMR
relaxation dispersion. Proc Natl Acad Sci U S A. 2011 May 31;108(22):9078-83. 2011 May 11.
Peter Wright
The Scripps Research Institute
La Jolla, CA
US
RCI of transiently formed state of
a T4 lysozyme mutant
Bouvignies G, Vallurupalli P, Hansen DF, Correia BE, Lange O, Bah A, Vernon RM, Dahlquist FW, Baker D, Kay LE
Nature. 2011 Aug 21;477(7362):111-4. doi: 10.1038/nature10349.
Solution structure of a minor and transiently formed state of a T4 lysozyme mutant.
University of Toronto
Toronto, Canada
DynaMine optimization to predict protein
flexibility from sequence
From protein sequence to dynamics and disorder with DynaMine.
Cilia E, Pancsa R, Tompa P, Lenaerts T, Vranken WF.
Nat Commun. 2013 Nov 14;4:2741
Outline
• RCI background
• Backbone RCI development
• Backbone RCI application
• Side-chain RCI
• Side-chain RCI application
Side-chain RCI
Side-chain RCI is more challanging due
to variability in side-chain structure
Phe78
1 2
f y
3) Collective motion 1) Rotations around
side-chain torsion
angles
2) Rotation around local backbone torsion angles
What side-chain motions do
we want to predict?
MD RMSF of total side-chain
motions
Existing protocol for backbone RCI could
not be applied to side-chains
I will skip slides about side-
chain RCI development
Side-chain RCI expressions
Side-chain RCI performance
Localization of PyJ side-chains with low (<0.11, violet color) and
high (>0.15, green color) RCISC values. Rigid side-chains (low
RCISC) are located primarily in the PyJ core (A), whereas flexible
side-chains (high RCISC) are mostly solvent exposed (B).
Side-chain RCI correlates with accessible surface area
Side-chain RCI webserver and paper
http://www.randomcoilindex.ca
Outline
• RCI background
• Backbone RCI development
• Backbone RCI application
• Side-chain RCI
• Side-chain RCI application
Application of side-chain RCI
RCI can help to identify mechanism
of total side-chain motions
Side-chain RCI detects dynamic changes
due to transient interactions Decrease in side-chain mobility due to
interactions with N-terminus
90°
Increase in
side-chain
mobility in PrP
"rigid" loop
Increase in side-
chain mobility on
the exposed side
of PrP helix 3
STAT1
TAZ2 domain
Side-chain RCI detects enthropy-entropy compensation
upon protein ligand interactions
Visualizing Side Chains of Invisible
Protein Conformers by Solution NMR
University of Toronto
Toronto, Canada
Side-chain mobility from 13C chemical shifts. RCI plots and for the native (a) and intermediate (b)
states of the L24A FF domain as a function of residue number. The first nine residues are omitted
as they are completely flexible. RCIBB and RCISC values are shown in blue and red, respectively,
along with the secondary structural elements in each of the states.
Bouvignies G, Vallurupalli P, Kay LE., J Mol Biol. 2013 Nov 8. pii: S0022-2836(13)00700-6
Summary • We have developed a new method, Random Coil
Index or RCI, to predict mobility of protein backbone
and side-chain groups from NMR chemical shifts
• The RCI method is not limited by the protein overall
tumbling
• The method does not require a 3D structure
• The RCI method does not require special
experiments beyond standard NMR assignment
experiments
• Side-chain RCI can be applied to all 19 side-chain
bearing residues
• Comparison of backbone and side-chain RCI can
identify the dominant mechanism of side-chain
motions