Structural Proteomics of Macromolecular Assemblies

1
Structural Proteomics of Macromolecular
Assemblies using Mass Spectrometric
Approach Sayan Gupta, PhD. Center for
Synchrotron Biosciences, Case Western Reserve
University National Synchrotron Light Source,
Brookhaven National Laboratory
2
Introduction

• Understanding composition, structure and
  dynamics and assembly of biological macromolecule
  is highly important.
• Advances in NMR and MX provides the foundation
  of this area
• Structural Genomics initiatives like Target
  Selection, High-Throughput Cloning, Expression
  and Purification and High-Throughput Structure
  Determination and Bioinformatics Approach
  correlate genomes sequence and structure of most
  protein domains
• Solving macromolecular structure and connecting
  structure to function for macromolecular
  assemblies are the most important challenges
  today
• Combination of biophysical techniques (NMR, MX
  of domain structure, Cryo-EM, Electron
  tomography, Cross linking, FRET, Spin Label EPR
  etc) are necessary to solve this problem
• Footprinting or Covalent Labeling is one of the
  well established approach for probing
  macromolecular structure in solution. H/D MS
  Exchange and Cross-linking are provide
  information about back-bone dynamics and
  Interaction interfaces

3
Multidisciplinary Approach

• Footprinting MS
• Local conformation change
• Binding interfaces
• Large Complexes
• Time Resolved

• Cross-linking MS
• Binding interfaces
• Large Complexes

• Cryo-EM
• Low Resolution
• Large Complexes gt200kD

• H/D exchange MS
• Backbone conformation
• Binding interface

• High-Throughout
• Structure (NMR, MX)
• High resolution
• Local and global
• lt50kD

• Small Angle
• X-ray Scattering
• Global Conformation
• Time resolved
• Large Complexes

• Computational
• Approach
• Homology Modeling
• Ab initio Modeling
• Docking

4
Footprinting Approach
Binding Isotherm
Site Specific Kd
Strand cleavage
Control
Protection
X
Kinetic progress curve
X
Protection
Site Specific kobs
dsDNA Protein
dsDNA (Control)
Gel Electrophoresis
Enzyme DNAase I Modifying agent DMSO, UV
radiation Cleavage reagent Hydroxyl radical
5
Chemical Generation of OH radical
Fenton Chain Reaction
Initiation
Fe2 H2O2
Fe3 HO? OH Fe3 H2O2
Fe2 HO2? H H2O2
HO? HO2? H2O HO2?
O2? H Fe3
HO2? Fe2 O2
H Fe3 O2?
Fe2 O2 Fe2 HO? H
Fe3 H2O
Propagation
Termination
1mM Fe(II)-EDTA, 0.3 H2O2, 10mM Ascorbate
Incubation at RT for 0-30min
Sample 20-80mM in Tris buffer 0.3 1.0 H2O2
Surface Modification Cleavage
Quenching by Thio-urea
6
Laser photolysis of H2O2
hu
H2O2 2HO? HO?
H2O2 H2O HO2? HO2?
H2O2 H2O O2
HO? 2HO?
H2O2 2HO?
Glutamine (radical scavenger)
NdYAG, 266nm,3-5ns pulse 1-100 pulse at
2mJ/pulse 20-80mM in Phosphate buffer
Protein Sample 20-80mM in Phophate buffer 0.3
1.0 H2O2
Surface Modification
7
Use of ionization radiation to generate OH radical
X-ray, g-ray
H2O H2O? edry, or
H2O H2O? HO?
H edry eaq H2O
HO? H? HO? eaq
OH 2HO? H2O2 eaq
O2 O2? H? O2
HO2?
10-14 sec
H2O
H2O
HO? production 10-12 sec
Secondary reaction products
8
X-ray mediated HO? radical Footprinting
X-ray
Synchrotron Source
Bending magnet
Beamline End-station
Insertion device
Linear Accelerator
100-500m
Booster

• Property of Synchrotron X-ray
• High brightness and flux
• Broad energy range (5-50keV)

Storage Ring
9
Beamline Configuration for Synchrotron
Footprinting
X-ray ring wall
V valves IG ion gauges/ion pumps BS
bremsstrahlung shields Be beryllium
BS1
Mirror
X-28C hutch
V3
IG2
Be window 2 3
10-7 torr
He (1 torr) -filled extendable beam-pipe
BS4
IG2
Be window 4
Heavy jack
BS6
Motorized table
Mirror upgrade has result in gt 120 fold increase
in effective X-ray dose
10
End-Station
Mirror
Beam at X28C
End station Equipments
Time Resolved Study
Sample Exposure
11
Advantage of X-ray radiolysis

• High Flux Beam
• Shorter exposure time
• Reduced sample damage
• Fast time resolved studies possible
• No use of additional chemical
• Stable sample conditions
• Experiment under optimum activity
• Dilute solution of macromolecule
• No direct interaction of X-ray with sample
• Reduced sample damage
• Use of various buffer solution
• Experiment under physiological conditions
• Experiment under optimum activity

• Dose can be varied precisely by exposure time
  variations.
• Dose can be regulated by mirror and attenuator
  in the X-ray pathway
• Added flux density overcomes extrinsic quencher
  in biological samples.
• Added flux density overcomes intrinsic quencher
  in large complexes.
• High flux density provides shorter exposure for
  in vivo experiments.

12
Monitoring X-ray Radiolysis
Fluorophore degradation
Rate constant of Flurophor degradation reflects
effective X-ray dose
13
Reaction with Proteins and Peptides
HO? is slightly electrophilic radical
Hydrogen Abstraction
Side chain H - abstraction
Side chain H - abstraction can lead to
backbone cleavage
Main chain H- abstraction can lead to
backbone cleavage. 100 fold slower
14
Reaction with Proteins and Peptides
Oxidation of Met
15
Reaction with Proteins and Peptides
Hydroxylation of Phe
Hydroxylation of Tyr
16
Multiple modification products
Modification of His
16Da
-22Da
-22Da
Modification of Trp
HO-trp Mixture of isomers 16Da
2-HO-trp 16Da
N-Formylkynurenine 32Da
Kynurenine 4Da
17
Reactivity order of the amino acids that are
modifiable by Synchrotron X-ray C gt M gt W gt Y gt F
gt C - C gt H gt L gt I gt R gt K gt V gt S gt T gt P gt Q gt
E gt D gt N gt A gt G
18
X-ray Radiolysis and Mass Spectrometry
HQGVMVGMGQK (m/z of 2 ion 586.3)
Exposed protein (0 100ms)
Protease Digestion
Peptide fragments
LC/MS
TIC (total ion chromatogram)
Calculate the modification from SIC (selected ion
chromatogram)
UM
Fraction Unmodified
? M UM
Identify sites of modifications by MS/MS
19
Site of Modification
y5
y3
y2
y1
y6
y4
MS/MS sequencing of the peptide
y5
y5 16
b2
y3 16
y3
b2
y4 16
b3
b3
b4
y4
b616
b4
b6
y6
y6 16
y1
y1
20
Change in RT of the modification products
EETLMEYLENPK
Native peak
Modified peaks
M
L
P
MIFAGIK
Native peak
Modified peaks
M
K
21
Rate of Modification - Dose Response
?UM
vs. Exposure Time
Fraction Unmodified
M ?UM
Multiple experimental data are fitted globally to
a single exponential decay y y0 A . exp (-
k.x ) k rate constant (sec-1)
Rate constant (sec-1) reflects both Reactivity
Solvent Accessibility
22
Solvent Accessibility and Side Chain Modification
Cyt c
Ubiquitin
23
Protein Footprinting Scheme
X-ray radiolysis (ms exposures)
MS analysis (LSMS, MSMS)
Modification
O
S
S
Native
OH
RA
Oxidized
Protein
?OH
30ms
?OH
20ms
X
O
10ms
S
S
0ms
x
x 16
m/z
Ligand
Model Building
24
Challenges in protein footprinting

• Data analysis takes longer time
• Complexity in the digest of large protein complex
• Suppression of modification reactions
• Low abundance of modification products

Recent advancements

• Automation in Data analysis and Software
  development
• Use of high sensitivity and high resolution MS
  instruments
• New methods to detect low abundance modification
  products
• Use of intense X-ray and controlled X-ray
  exposure

25
Activation of AVP by its cofactors DNA-Protein
Interaction
26
Study of Actin Cofilin Interaction
Actin-cofilin interaction in cell regulates
cellular processes such as cell motility, cell
division and cell morphology
Kamal JK, et. al. Proc. Natl. Acad. Sci. USA.
(2007), 104, 7910-7915.
27
Actin Cofilin Interaction
28
Superposition of footprinting results in the
structure
Significant protection Moderate protection No
change Increased reactivity
Cofilin binds at the cleft between subdomains 1
and 2. It induces closure of the nucleotide cleft
29
Modeling Protein Binary Complex using MS data

• Footprinting provide surface mapping of solvent
  exposed residues in undergoing conformational
  changes and binding interactions
• Molecular Docking program in conjunction with
  the footprinting result to generate high
  resolution structures of protein complexes
• Molecular docking program that compares surface
  complimentarity between interacting protein can
  generate high resolution model when target and
  template had 50 sequence homology. Accuracy
  dropped significantly when the identity is lt30.
• The performance of the docking programs is
  highly dependent on the nature of binding site.
  Identification of interface that could guide the
  docking steps and recognition of driving forces
  defining the energy minimization steps are
  crucial to the modeling.

30
Footprinting Experiment
Step 1
Region of conformation reorganization identified
Binding interface identified
Selecting residue from the footprinting probe
within the interface to act as positive
constraint in the docking strategy
Bound conformation of individual proteins derive
by homology modeling
Selecting residues outside the interface to act
as negative constraints in the docking energy
Examine the electrostatic potential surface to
judge the dominant mode of binding
Electrostatic or desolvation
Step 2
Protein-protein docking using ClusPro
Docking by DOT (grid based shape-complementarity
)
Positive constraints (attract)
Negative constraints (block)
No constraints
Top 20,000 conformations (shape complementarity)
Energy filtering (electrostatic and desolvation)
Top 2000 conformations (energy score)
http//nrc.bu.edu/cluster/clusdoc.html
Clustering
Top 10 conformation (cluster size)
1. Consistency of interface residues with
footprinting residues (FICS score) 2. Interface
parameter (SASA planarity, Gap Volume Index in
the acceptable range)
CHARM energy minimization
Testing Models
1. Identification of interface H-bond/salt
bridge 2. Structural correlation with function
FINAL MODEL
31
Actin / Profilin No Footprinting Test 1
X-ray structure
Modeled structure
S2
S4
Electrostatic hits2000
S3
S1
Rank 1
Dominant electrostatic interaction
Top five models Rank 1 Rank 2 Rank 3 Rank 4 Rank
5
32
Actin-DNaseI No Footprinting Test 2
Modeled Structure
X-ray Structure
Electrostatic hits 0
Rank 2
Dominant Desolvation Free energy
Electrostatic hits 2000
Electrostatic hits 0
Top five models Rank 1 Rank 2 Rank 3 Rank 4 Rank
5
33
Actin-Cofilin Interaction Footprinting
X-ray Test 3
Dominant Desolvation Free energy FICS
0.29/1.0 CS 65 RMSD ABP 6.7 RMSD Interface
2.9
Dominant Desolvation Free energy FICS
0.21/1.0 CS 33 Rank 6 had low RMSD
Dominant Desolvation Free energy FICS
0.33/1.0 CS 69 RMSD ABP 6.1 RMSD Interface 3.7
34
Actin - Cofilin No Experimental Constraint
Rank 2
Rank 3
Rank 1
FICS 0.1/0.2 CS 30
35
Actin - Cofilin ATTRACT Constraint FP data (79,
87, 88, 91, 362, 367, 371)
Actin/Cofilin BLOCK Constraint Regions from SD3
and SD4
Rank 1
Rank 1
FICS 0.14/0.2 CS - 48
FICS 0.2/0.2 CS 57
36
Actin-Cofilin BLOCK ATTRACT Constraint
MODEL OF ACTIN-COFILIN COMPLEX
We identified three key functionally relevant
interactions at the actin-cofilin interface. The
hydrogen bonding His87-Ser89, may ontribute to
the observed pH dependency of cofilin binding.
Rank 1
CS - 68
Conventional model didnt show up in the top 10
structures
37
FootPrinting/Rosetta Project Experimentally-Guid
ed ab initio Structure Determination
Rosetta
Footprinting
Structure prediction via ab initio modeling
algorithm
Provides protein surface accessibility
information
Rosetta Footprinting
Guide dog
Increased speed/accuracy via elimination of
invalid ab initio models
38
Flow of Experimental plan

• Select Initial Protein Structural Initiative
  targets
• Type I Targets with structures determined for
  calibration of FP/Rosetta strategy
• Type II Targets with diffraction data available
  but no phasing information for evaluation of
  FP/Rosetta strategy via molecular
  replacement/refinement
• All targets are prepared and subjected to
  biochemical analysis at
• All targets sent for exposure and Mass
  Spectrometry analysis
• Rosetta models will be generated for Type II
  targets and filtered using footprinting data
• These models will be used for Molecular
  Replacement phasing of diffraction data and for
  comparison with final PDB depositions

39
Rosetta/Footprinting MR StrategyProof-of-Concept
Study

• Small protein test case (76 aa)
• Rosetta models (71 aa) generated with FP (blue)
  or without FP (green) data
• Very accurate backbone prediction from Rosetta
  (1 Å rmsd) with FP
• Five models from with FP or without FP selected,
  each with minimal energy scores
• Successful MR with all five with FP models
• All five without FP models failed MR

X-axis Ca rmsd compared to Ubiquitin xtal
structure Y-axis energy score Red Minimized
crystal structure Blue Rosetta models with
incorporation of FP data Green Control models
with default Rosetta setting
40
In vivo Footprinting
Protection due to subunit association in E. coli
16S rRNA in a H2O2 / Fe(II)-EDTA footprinting
experiment is done .
Structure of Ribosome
What happens inside the cell?
Adilakshmi, T. et al. Nucleic Acid Research
(2006)
41
Experimental Strategy for In vivo Footprinting
Grow E.coli (MRE600) cells
Harvest and re-suspend in buffer ( 6ml/mg wet
cells)
Snap-freeze as 5-10ml aliquots in PCR tubes
Primer extension analysis with 32P labeled primers
Frozen cell pellets are exposed at - 34 to -38º C
Isolate RNA from the frozen exposed cells

cDNA products were separated in denaturing PAGE.
Observe protection pattern was compared with ASA
and in vitro studies
Fragmentation of cellular RNA by X-ray. 200-300
ms exposure time for optimum cleavage
42
In vivo Footprinting - Results

• The in vivo footprint of 16S rRNA in frozen
  cells were similar to those obtained in vitro.
• The in vivo footprints were consistence with
  the predicted accessibility of the RNA backbone
  to hydroxyl radical.

In vivo
In vitro
43
In vivo Footprinting - Results
nt 1218-1242
nt 483-508
Figure Comparison of in vivo and in vitro
footprinting of 16S rRNA. (A B) The intensity
of bands in sequencing gels 16S domains. Red
lines, irradiation in vivo (250ms exposure) and
the black lines irradiation in vitro (30 ms
exposure). (C D) Difference in cleavage
protection mapped in the secondary structure. Red
circles, greater protection in vivo than in
vitro black circle, enhanced cleavage in vivo.
44
Rapidly dividing cells require thousand of new
ribosomes. Assembly of individual subunit
requires careful regulation of rRNA
transcription, ribosomal protein synthesis and
the precise interaction of rRNA with 20 or 30
unique ribosomal proteins.
How the complex assembles in cell?
Adapted from http//rna.ucsc.edu/rnacenter/noller
Structure of intact 70S ribosome.
Use of X-ray OH radical footprinting
45
Rapidly dividing cells require thousand of new
ribosomes. Assembly of individual subunit
requires careful regulation of rRNA
transcription, ribosomal protein synthesis and
the precise interaction of rRNA with 20 or 30
unique ribosomal proteins.
How the complex assembles in cell?
Adapted from http//rna.ucsc.edu/rnacenter/noller
Structure of intact 70S ribosome.
Use of X-ray OH radical footprinting