BCHM 313 – Physical Biochemistrypldserver1.biochem.queensu.ca/~rlc/steve/313/Smith_313... ·...

BCHM 313 – Physical Biochemistry

Dr. Michael Nesheim (Coordinator)[email protected]

Rm. A210 Botterell

Dr. Steven Smith (Co-coordinator)[email protected]

Rm. 615 Botterell

Dr. Susan [email protected]. 623 Botterell

BCHM 313 – Physical Biochemistry

Topics1. Protein NMR – Smith

2. Macromolecular Crystallography – Yates3. Hydrodynamics – Nesheim

4. Equilibrium Binding – Nesheim5. Enzyme Kinetics – Nesheim6. Spectroscopy – Nesheim

EvaluationMidterm test – 35%Final exam – 65%

Protein NMR Spectroscopy

Determining three-dimensional structures and monitoring molecular interactions

(http://pldserver1.biochem.queensu.ca/~rlc/steve/313/)

Outline

• N-dimensional NMR• Resonance assignment in proteins• NMR-based structure determination• Molecular interactions

Reference textbooks:- Lehninger - others available in my office.

Nuclear Magnetic Resonance (NMR)

• MRI – Magnetic Resonance Imaging (water)• In-vivo spectroscopy (metabolites)• Soild-state NMR (large structures)• Solution NMR

• Chemical structure elucidation- Natural product chemistry- Synthetic organic chemistry – analytical tool of choice for chemists

• Biomolecular structural studies (3D structures)- Proteins- DNA and Protein/DNA complexes- Polysaccharides

• Molecular interactions- Ligand binding and screening (Biotech and BioPharma)

•Nobel prizes (3): Felix Bloch, 1952 (Physics); RichardErnest, 1991 (Chemistry); Kurt Wuthrich, 2002 (Chemistry)

Spectroscopy & Nuclear Spin•Absorption (or emission) spectroscopy (IR, UV, vis). Detects the absorption of radiofrequencies (electro-magnetic radiation) by certain nuclei in a molecule - NMR

• Unfortunately, some quantum mechanics are needed to understand it

• Only nuclei with spin number (I) ≠ 0 can absorb/emit electro-magnetic radiation

• Nuclei with even mass number & even number of protons (12C): I = 0• With even mass number & odd number of protons (14N): I = 1, 2, 3….• With odd mass number (1H, 15N, 13C, 31P): I = 1/2, 3/2 …….

• Spin states of nucleus (m) are quantified: mI = (2I + 1)

Thus, for biologically relevant nuclei, two spin states exist: 1/2, -1/2

Spin ½ Nuclei Align in Magnetic Fields

Energy

Efficiency factor-nucleus

Constants Strength ofmagnet

• In ground state all nuclear spins are disordered – no energy difference degenerate• Since nuclei are small magnets, they orient when a strong magnetic field is applied – small excess aligned with field (lower energy)

Bo

∆E = h γ Bo/2π

Intrinsic Sensitivity of Nuclei

Nucleus γ % Natural Relative Abundance Sensitivity

1H 2.7 x 108 99.98 1.0

13C 6.7 x 107 1.11 0.004

15N -2.7 x 107 0.36 0.0004

31P 1.1 x 108 100. 0.5

Resonance: Perturb Equilibrium

hν = ∆EH1

2. pump in energy

∆E = h γ Bo/2π


Constant Strength ofmagnet

β

α

1. equilibrium

∆E

β

α 3. non-equilibrium

Bo

Return to Equilibrium (Relax)

β

α5. equilibrium

hν = ∆E 4. release energy (detect)

β

α

∆E 3. Non-equilibrium

Magnetic Resonance Sensitivity

∆E = h γ Bo/2π


Constant Strength ofmagnet

Nβ

Nα= e-∆E/kT

Sensitivity (S) ~ ∆(population)

S ~ ∆N =

• ∆E is small - @ r.t. ~1:105

Thus, intrinsically low sensitivity *Need lots of sample

Increase sensitivity by increasing magnetic field strength

Energy/Frequency Relationship

∆E = hν∆E = hγB°/2π

ν = γB°/2π

For a particular nucleus:

Have to consider precession since: 1) nucleus has inherent spin and 2) application of a large external magnet generates a torque which results in precession

Bo

ωo µ

Precession occurs around B° at frequency,

termed Larmor frequency (ω).

ν = γB°/2π = ω

Reason why 14.1 Telsa magnet often called 600 MHz magnet.

Bulk Magnetization

α

β Mo

z

x

y

Bo

x

z

x

z

y

x

z

y

y

Two spins All spins Sum

Power of Fourier Transform

z

x

y

z

x

y

90° RF pulse

z

x

y

t

ω = γB°

t

A

Variation of signal at X axis vs. time

ωf FourierTransform

B°

NMR frequency

Pulse Fourier Transform NMR

z

x

y

z

x

y

90° RF pulse

z

x

y

t

ω1 = γB°

FourierTransform

ω2 = γB°

y

x

z

t

A

NMR time domain⇒ Variation in amplitude vs. time

NMR frequency domain⇒ Spectrum of frequencies

ω1f ω2

Pulse FT NMR Experiment

equilibration

90º pulse

detection of signals

Experiment

DataAnalysis

FourierTransform

Time domain (t)

equilibration acquisition (t)

FID

NMR terminology

• Magnetic field (B°) felt by each nucleus affected by its local electronic environment - big difference between B° (MHz) and Blocal (hundreds of Hz) ie. parts per million (ppm, δ)• Use relative scale and refer all signals in spectrum to a signal from a reference compound (DSS)

δ =ω - ωref

ωref

Summary

α

β Mo

z

x

y

Bo

x

z

y

All spins Sum

}

∆E = hγBo/2π

z

x

y

z

x

y

90° RF

z

x

y

t

ν = γB° = ω∆E = hν

0 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00t1 sec

FID – time domain

Summary (cont)

0 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00t1 sec

FID – time domain

234 233 232 231 230 229 228 227 226 225 224 223f1 ppm

frequency domain

010ppm

Chemical shift – relative scale

NMR terminology

Scalar and Dipolar Coupling

ThroughSpace

ThroughBonds

Coupling of nuclei gives information on structure

Resonance Assignment

CH3-CH2-OH

OH CH2 CH3

Which signal from which H atoms?

The key attribute: Use scalar and dipolar couplings to match the set of signals with the molecular structure

Proteins Have Too Many Signals!

~500 resonances

1H 1D NMR Spectrum of Ubiquitin

Resolve resonances by multi-dimensional experiments

1H (ppm)

Examples of Amino Acids

NMR experiment

equilibration

90º pulse

detection of signals

1D equilibration acquisition (t)

preparation mixingevolution (t1) acquisition (t2)

90º pulse

Same as 1Dexperiment

2D detect signals twice(before/after couple)

Transfers between coupled spins

2D

2D NMR: Coupling is the Key


90º pulse


2D detect signals twice(before/after couple)

Transfers between coupled spins

t1 t2t1t2

2D NMR Spectrum

t2t1


excitation

Pulse sequence

Coupled spins

Spectrum

Before mixing

After mixing

HA

HB

t1 t2

Either:

or:

The Power of 2D NMR

Resolving Overlapping Signals

1D

2D

2 signalsoverlapped

2 cross peaksresolved

Multi-Dimensional NMR


90º pulse


3D detects signals 3 times

t2t1

Built on the 2D Principle

evolution (t2) mixing

excitation

t3

Protein NMR: Practical Issues

• Magnet: homogeneous, high field - $$$$

• Electronics: stable, tunable

• Environment: temperature, pressure, humidity, stray fields

Hardware:

• Recombinant protein expression (E. coli, Pichia pastoris etc)

•Volume: 300 µL – 600 µL

• Concentration: 1D ~ 50 µM, nD ~ 1mM ie. @ 20 kDa, 1mM = 10 mg

• Purity: > 95%, buffers

• Sensitivity (γ): isotope enrichment (15N, 13C)

Sample Preparation:

Protein NMR: Practical Issues (cont.)

• Variables: buffer, ionic strength, pH, temperature

• Binding studies: co-factors, ligands

• No crystals!

Solution Conditions:

• up to 30 – 40 kDa for 3D structure determination

• > 100 kDa: uniform deuteration, residue and site-specific, atom-specific labeling

• Symmetry reduces complexity: 2 x10 kDa ≠ 20 kDa

Molecular Weight:

NMR Spectrum to 3D structure?

||012 1H (ppm)

Critical Features of Protein NMR Spectra

• The nuclei are not mutually coupled

• Regions of the spectrum correspond to different parts of the amino acid

• Tertiary structure leads to increased dispersion of resonances

• chemical shifts associated with each nucleus influenced by local chemical environment – nearby nuclei

Each amino acid gives rise to an independent NMR sub-spectrum, which is much simpler than the complete protein spectrum

Regions of a protein 1H NMR Spectrum

What would an unfolded protein look like?

Solutions to the Challenges

1. Increase dimensionality of spectra to better resolve signals: 1 ⇒ 2 ⇒ 3 ⇒ 4

2. Detect signals from heteronuclei (13C, 15N)

Better resolution of signals/chemical shifts not correlated nuclei

More information to identify signals Lower sensitivity to MW of protein

1D Protein 1H NMR Spectrum

Resolve Peaks by Multi-D NMR

A BONUS→regions in 2D spectra provide protein fingerprints

If 2D cross peaks overlap→ go to 3D

Basic Strategy to Assign Resonances in Protein

1.Assign resonances for each amino acid

2. Put amino acids in order- Sequential assignment (R-G-S, T-L-G-S)- Sequence-specific assignment

LT G S S R G

1 2 3 4 5 6 7

R - G - S - T - L - G - S

Acronyms for Basic Experiments

Differ Only in the Nature of Mixing

Scalar Coupling

Dipolar Coupling

Homonuclear

COSYCOrrelation SpectroscopY

Heteronuclear

TOCSYTOtal Correlation SpectroscopY

NOESYNuclear Overhauser Effect

(Enhancement) SpectroscopY

HSQCHeteronuclear

Hetero-TOCSY

NOESY-HSQC(thru-space)

(thru-bond)

Homonuclear 1H Assignment Strategy

• For proteins up to ~ 10 kDa

•Scalar couplings to identify resonances/spin systems/amino acids, dipolar couplings to place in sequence

• Based on backbone HN (unique region in 1H spectrum, greatest dispersion of resonances, least overlap)

• Concept: Build out from the backbone to identify the side-chain resonances (unique spin systems)

• 2nd dimension resolves overlap, 3D rare


Step 1: Identify Spin System

HN

αHN – C – C

–

C

=–

H O

COSY (3-bond)

Alanine

δCH3

–

––

HH H

TOCSY

–

H



HN

αH

β ’H

βH

γH

δCH3

δ’CH3COSY (3-bond)TOCSY

N – C – C

–

H – C – H

–

C – H

– –

H3C CH3

=–

H O

Leucine

–

H



HN

αH

β’H

βH

γH

δCH3

δ’CH3COSY (3-bond)TOCSY

N – C – C

–

H – C – H

–C – H

– –

H3C CH3

=–

H O

Leucine- closed circles

–

H

N – C – C

–

C

=–

H O

Alanine- open circles

–

––

HH H

–

HN

αH

δCH3

–

H


Step 2: Fit residues in sequenceMinor Flaw: All NOEs mixed together!

A B C D Z• • • • Intraresidue

Sequential

Medium-range(helices: Hα-HN (i, i + 3,4)))

Long Range

Use only these to make sequential assignments

•Sequential NOEs⇒ HN-HN (i, i + 1)⇒ Hα-HN (i, i + 1)


Step 2: Fit residues in sequence

HN

αH

β’H

βH

γH

δCH3

δ’CH3

COSY/TOCSY+

N – C – C

–

H – C – H

–C – H

– –

H3C CH3

=–

H O

–

H

N – C – C

–

C

=–

H O

–

––

HH H

–

HN

αH

βCH3

–

H

NOESY =

Extended Homonuclear 1H Strategy

• For proteins up to ~ 15 kDa

•Same basic idea as 1H strategy: based on backbone HN

• Concept: When backbone 1H overlaps ⇒ disperse with backbone 15N

• Use heteronuclear 3D experiments to increase signal resolution

1H 1H 15N

Solutions to the Challenges

1. Increase dimensionally of spectra to better resolve signals: 1 ⇒ 2 ⇒ 3 ⇒ 4

2. Detect signals from heteronuclei (13C, 15N)

Labeling with NMR-observable 13C, 15N isotopes Better resolution of signals/chemical shifts not

correlated nuclei More information to identify signals Lower sensitivity to MW of protein

Isotopic Labeling

• Require uniform 15N/13C labeling ie. Every carbon and nitrogen isotopically labeled

How?• Grow bacteria on minimal media (salts) supplemented with 15N-NH4Cl and 13C-glucose as soles sources of nitrogen and carbon

• lower yields than protein expression than on enriched media, therefore need very good recombinant expression system

Double Resonance Experiments

Increases Resolution/Information Content

Increases Resolution/Information Content

15N – Cα – C – 15N – Cα

H

OH

R

R

Heteronuclear NMR: 15N-Edited Experiments

3D Heteronuclear NMR: 15N-Edited Experiments

+

Extended Homonuclear 1H Strategy15N dispersed 1H-1H TOCSY

3 overlapped NH resonances(diagonal)

Same NH, different 15N

F1

F2F3

1H 1H 15Nt1 t2 t3

TOCSY HSQC

HN (ppm)

1 H (ppm)

15N – Cα – C – 15N – Cα

H

OH

R

R

Summary of Homonuclear Assignment Strategy

• for proteins up to ~10 kDa (2D homonuclear) and proteins up to ~ 15 kDa (15N-labeling and 3D)

• using scalar coupling-type experiments (COSY, TOCSY) assign spin systems/side-chain resonances

• Connect amino acids (identified based on spin systems) sequentially using NOE-type experiments and characteristic sequential NOEs (HN-HN (i, i+1); Hα-HN (i, i+1))

Heteronuclear (1H, 13C, 15N) Strategy

• for larger proteins (backbone assignment: ~70 kDa; full structure determination: ~40 kDa)

•Assign resonances (chemical shifts) for all atoms (except O)

•Handles overlap in backbone H regiondisperse with backbone C’, Cα, Hα ,Cβ, Hβ

• Heteronuclear 3D/4D increases resolution

• Works on bigger proteins because scalar couplings are larger

15N

1H 13C 1H 15N

Heteronuclear (1H, 13C, 15N) Strategy

Step 1: Sequence-specific backbone assignment

Assign backbone 1H, 15N, Cα, Cβ resonances/chemical shifts and sequentially link amino acids using partner scalar coupling experiments

Step 2: Side-chain assignment

Assign side-chain 13C & 1H resonances/chemical shifts using TOCSY-type 3D scalar coupling experiments

** Have complete list of chemical shifts for all 13C, 15N, 1H atoms in protein **

Heteronuclear (1H, 13C, 15N) Assignments

Backbone Experiments

Names of scalar experiments based on atoms detected

Consecutive residues!!NOESY not needed


Backbone ExperimentsCBCA(CO)NH HNCACB- inter-residue connectivity - intra-residue connectivity(HN to previous Cα, Cβ) (HN to own Cα, Cβ)

N – C – C=

H H O

H

N – C – C – N

H

=H H O

H – C – H

H

13Cβ chemical shift

13Cα chemical shift

common 15N and HN chemical shift in both experiments

(found on same 15N plane)

Search 15N planes for 13Cα and 13Cβ chemical shifts

N – C – C – N

R

=

H H O

H – C – H

H


Backbone Experiments

CBCA(CO)NH- inter-residue connectivity(HN to previous Cα , Cβ)

HNCACB- intra-residue connectivity

and possibly inter-residue(HN to own Cα , Cβ)

Start with unique residue1. Gly – only Cα2. Ala – upfield-shifted Cβ

(~18 ppm)3. Thr/Ser – downfield-shifted

Cα & Cβ which are close to each other


Side-chain Experiments

Multiple redundancies increase reliability


Key Points

• Enables the study/assignment of much larger proteins (up to ~100 kDa)

•Scalar coupling-type 3-dimensional experiments only

•Bonus: Amino acid identification and sequence-specific assignment all at once

• Most efficient but experiments are more complex

•Requires 13C, 15N enrichment (also 2H)⇒High expression levels on minimal media⇒ Increased cost ($150/g 13C-gluocose; $30/g 15NH4Cl)

Structure Determination Overview

List of chemical shifts for all nuclei in protein (1H, 13C, 15N)

NMR Experimental ObservablesProvide Structural Information

1. Backbone conformation from chemical shifts (Chemical Shift Index – CSI; Hα, Cα, Cβ, C’)

2. Hydrogen bond constraints

3. Backbone and side chain dihedral angle constraints from scalar couplings

4. Distant constraints from NOE connectivities

1. Chemical Shift Index

• Comparison of Hα, Cα, Cβ, C’ determined chemical shifts from protein to standard random coil chemical shift values

• Upfield-shifted Hα and Cβ and downfield-shifted Cα and C’ values indicate amino acid residues in an α-helical conformation (requires three consecutive residues displaying this pattern)

• Downfield-shifted Hα and Cβ and upfield-shifted Cα and C’ values indicate residues in an extended (β-strand) conformation

2. Hydrogen Bonds

C=O H-N

• Slow rate of exchange of labile HN with solvent•Protein dissolved in 2H2O; HN signals disappear with time•HN groups that are H-bonded (i.e. part of secondary structure) will exchange a lot slower than those in loops

3. Dihedral Angles from Scalar Couplings

• • • •6 Hz

Must accommodate multiple solutions→ multiple J values

4. 1H-1H Distances from NOEs

A B C D Z• • • • Intraresidue

Sequential

Medium-range(helices)

Long-range(tertiary structure)

Challenge is to assign all peaks in NOESY spectra- semi-automated processes for NOE assignment using

NOESY data and table of chemical shifts yet still significant amount of human analysis

Protein Fold without Full Structure Calculations

1. Determine secondary structure•CSI directly from assignments•Medium-range NOEs

2. Add key long-range NOEs to fold

Approaches to Identifying NOEs

• 15N- or 13C dispersed 1H-1H NOESY

3D1H

13C

1H1H

15N

1H

1H

15N

1H

13C

1H

13C

1H

13C

1H

15N

1H

15N

4D

1H 1H2D• 1H-1H NOESY

NMR Structure Calculations

Objective: Determine all conformations consistent with experimental data

• Programs that only do conformational search may lead to bad geometry use simulations guided by experimental data

• need a reasonable starting structure

•Distance restraints arrived at from NOE signal intensities signal is an average of all conformations

NMR Structure Calculations (cont)

1. NOE signals are time & population-averaged (ie. measured on entire sample over period of time)

2. Intensity of NOE signal ∝ 1H-1H distance (1/r6)

∴ NOE distance restraints are given a range of valuesstrong NOE: 0 - 2.8 Åmedium NOE: 2.8 – 3.5 Åweak NOE: 3.5 – 5.0 Å

NMR data not perfect: Noise, incomplete data multiple solutions (conformational ensemble unlike X-ray crystallography with one solution)

Variable Resolution of Structures

• Secondary structures well defined, loops variable• Interiors well defined, surfaces more variable• Trends the same for backbone and side chains

More dynamics at loops/surface Constraints in all directions in the interior

Assessing the Quality of NMR Structures

• Number of experimental constraints

• RMSD of structural ensemble (subjective!)

• Violation of constraints- number, magnitude

• Molecular energies

• Comparison to known structures: PROCHECK

• Back-calculation of experimental parameters

Summary of Protein NMRStructure Determination

Sample preparation with possible isotope labeling

Data collection (scalar coupling and dipolar coupling expts.

Resonance and sequence-specific assignments

Identification and quantification of NOE peaks and intensities and conversion to approx. 1H-1H distances

Generation of models consistent with NOE distance constraints, dihedral angle ranges, H-bond distances

Model improvement by inclusion of newly identified NOES using above mentioned models

NMR Structures – Now what?

Monitoring Molecular Interactions

15N-1H HSQC

NMR Provides

Site-specific

Multiple probes

In-depth info

Spatial distribution of responses can be mapped on structure

A14

L15

S16

W17

D18

V19 M20

Q21

G22

T23

I24

R25

K26

G27

F28

V29

N30

H31

A32

K33

I34 S35

Y36

C37


Titration followed by 15N-1H HSQC


Map of chemical shift perturbations on the Map of chemical shift perturbations on the

structure of protein?structure of protein?

Transcription factor (CBP) -oncoprotein (E2A) interaction- collaboration with Dr. David LeBrun (Pathology)


- Identification of ligand (E2A)-binding site on the structure of the KIX domain of CBP


Chemical Perturbation Mapping Structure

Ligand Binding

NMR timescale – 1 sec to 1 x 10-6 sec

1/koff = t >> 1 sec ⇒ slow exchange, superposition of spectra

1/koff = t << 1 x 10-6 sec ⇒ fast exchange, weighted average

A B Kdiss = [A]/[B] = koff/kon

kon

koff

Ligand Binding

- Another protein- Metal ion- Drug or chemical

P + L = PL Kdiss =[P] [L][PL]

Ligand Binding - exchange

E641, S642, and S670- Fast exchange(weighted average of free and bound populations)

T614- Intermediate-fast exchange

Ligand Binding

Ptot = P + PL

Ltot = L + PL

So……. Kdiss = [Ptot - PL] [Ltot - PL]

[PL]

Plot [Ltot]/[Ptot] vs “change” in NMR spectra

For fast exchange (weak binding):

Change = = shifting of resonances in spectra

For slow exchange (tight binding):

Change = = intensity changes in peaks

δobs - δ init

δsat - δ init

[ PL][Ptot]

Integral of peakobs

Integral of peakmax

[ PL][Ptot] of free and bound forms


Binding Constants by NMR

Fit change in chemical shift to binding equation

Molar ratio of d-CTTCA

Stronger Weaker

Protein DynamicsInteresting because……..

• Function requires motion/kinetic energy

• Entropic contributions to binding events

• Protein folding/unfolding

• Uncertainty in NMR and crystal structures

• Effects on NMR experiments: spin relaxation is dependent on motions know dynamics to predict outcomes and design new experiments

Characterizing Protein Dynamics

Parameters & Timescale

Dynamics from NMR Parameters

• Number of signals per atom: multiple signals for slow exchange between conformational states

A B

Populations ~ relative stability

Rex < ω (A) - ω (B)

Rate



• Linewidths: narrow = fast motions, wide = slow motions; dependent on protein molecular weight (MW)

Linewidths Dependent on Protein MW

A B A B

1H

1H

15N15N

1H

15N

• Same chemical shifts,same structure

• Linewidth determinedBy size of molecule

• Fragments have narrow linewidths




• Exchange of HN solvent: slow timescales (milliseconds to years!)

• requires local or global unfolding events• HN involved in H-bonds exchanges slowly• surface or flexible region: HN exchange rapidly




• Exchange of HN solvent: slow timescales (milliseconds to years!)

• NMR relaxation measurements (ps – ns; µs – ms)• R1 (1/T1) spin lattice relaxation rate (z-axis)

• R2 (1/T2) spin spin relaxation rate (xy-plane)• Heteronuclear NOE (15N-1H)

Dynamics to Probe the Originof Structural Uncertainty

Strong correlation

Weak correlation↑

↑↑↑

↑

- Measurements show if high RMSD is due to high flexibility (low S2)

NMR and CrystallographyNMR

• Can mimic biological conditions - pH, temp, salt

• information on dynamics

• monitor conformational change on ligand binding

• 2° structure derived from limited experimental data

• need concentrated sample - lots of protein; aggregation issues

• size limited – ~40kDa for full structure determination

• more subjective interpretation of data

• lack of quality factors -resolution and R-factor

X-ray

• Highly automated with more objective interpretation of data

• Quality indicators (resolution, R)

• Surface residues and water molecules well defined

• Huge molecules and assemblies can be determined

• non-physiological conditions – crystallization difficult

• need heavy-atom derivatives – production not always trivial

• snap-shot of protein in time –less indication of mobility

• flexible proteins difficult to crystallize

Identifying Unique NOEs

• Filtered/Edited NOE: based on selection of NOEs from two molecules with unique labeling patterns

1H 1H

13CUnlabeled

peptide

Labeledprotein

Only NOEs at the interface

• Transferred NOE: Used for weak interactions (ligand in excess) and based on NOEs from bound state passed to free state

Only NOEs from bound stateH

H

H

Hkon

koff

BCHM 313 – Physical Biochemistrypldserver1.biochem.queensu.ca/~rlc/steve/313/Smith_313... ·...

Documents

Transcript of BCHM 313 – Physical Biochemistrypldserver1.biochem.queensu.ca/~rlc/steve/313/Smith_313... ·...