NMR in Medicine and Biology

Transcription

1 NMR in Medicine and Biology MRI- Magnetic Resonance Imaging (water) In-vivo spectroscopy (metabolites) Solid-state t NMR (large structures) t Solution NMR Spectroscopy py Analytical biochemistry Comparative analysis Interactions between biomolecules Ligand screening (SAR by NMR) Structure determination Protein Dynamics

2 Protein NMR Detection of spin 1/2 nuclei 1 H, 13 C, 15 N, 31 P Only nuclei with spin number (I) 0 can absorb/emit electromagnetic radiation. Even atomic mass & number I = 0 ( 12 C, 16 O) Even atomic mass & odd number I = whole integer ( 14 N, 2 H, 10 B) Odd atomic mass I = half integer ( 1 H, 13 C, 15 N, 31 P) Nuclear spin angular momentum is a quantized property of the nucleus in each atom

3 Nuclei With Non-Zero Spin Align in Magnetic Fields Ho Alignment Energy Efficiency factornucleus parallel ΔE = h γ Ho anti-parallel Constants Strength of magnet

4 Resonance: Perturb Equilibrium i Ho p ap ΔE 1. equilibrium Efficiency factornucleus H1 hν = ΔE 2. pump in energy ΔE = h γ Ho p Constants Strength of magnet ap 3. non-equilibrium

5 Return to Equilibrium (Relax): Read Out Signals p ap ΔE 3. Non-equilibrium hν = ΔE 4. release energy (detect) p ap 5. equilibrium

6 Magnetic Resonance Sensitivity Sensitivity (S) ~ Δ(population) Efficiency factornucleus S ~ ΔN = Np N = e-δe/kt Nap e ΔE = h γ Ho ΔE is small At room temp., ΔN ~ 1:10 5 Intrinsically low sensitivity Need lots of sample Constants Strength of magnet Increase sensitivity by increasing magnetic field strength And high protein concentration

7 Intrinsic Sensitivity of Nuclei Nucleus γ % Natural Relative Abundance Sensitivity 1 H 2.7 x C 6.7 x N x P 1.1 x

8 The Classical Treatment: Nuclear Spin Angular Momentum Two spins All spins Sum Ho parallel anti-parallel Torque + int. motion = precession Precession around Z axis Larmor frequency (ω): ΔE = hγho ΔE = hν ν = γh 0 = ω excess facing down Bulk Magnetization

9 Pulse Fourier Transform NMR = 90ºx RF pulse t Ho Ho ω = γ H 0 f ω NMR frequency Fourier Transform A Variation of signal at X axis vs. time t

10 The Power of Fourier Transform + 90ºx RF pulse t ω1 = γ H 0 A ω2 = γ H 0 t f ω2 ω1 Fourier Transform NMR frequency domain Spectrum of frequencies NMR time domain Variation in amplitude vs time

11 The Pulse FT NMR Experiment Experiment 90º pulse (t) equilibration detection of signals Data Analysis Fourier Transform Time domain (t)

12 δ = ω - ω ref ω ref NMR Terminology ppm (parts per million), Universal Unit Universal Reference 1. Samples in water 2. Organic compounds The exact resonance frequency (chemical shift) is The exact resonance frequency (chemical shift) is determined by the electronic environment of the nucleus

13 NMR Scalar and Dipolar Coupling Through Space Through Bonds Coupling of nuclei gives information on structure

14 The 2D NMR Pulse Sequenceence + 1D + 1D = 2D [2nd preparation]

15 The 2D NMR Spectrum Pulse Sequence Spectrum t1 t2 Before mixing After mixing Coupled spins give rise to crosspeaks

16 The Power of 2D NMR: Resolving Overlapping Signals 1D 2 signals overlapped 2D 2 cross peaks resolved

17 Regions of the 1 H NMR Spectrum Water What would the unfolded protein look like?

18 Each side chain proton has characteristic ppm range of signal on the NMR spectrum. Aromatic wate er Imines Amides HCα HCβ, γ, δ,

19 Challenges NMR of Proteins Proteins have hundreds/thousands of signals Resonance assignment first (who do all these signals belong to?) Computer programs can be used to get assistance with resonance assignments Applications Folded protein? Measure binding constants Assess structural homology/effect of mutations Three-dimensional structure determination Measure flexibility/dynamics

20 H- Beyond Primary Structure: The Peptide Bond Peptide plane is flat ω angle ~180º Partial double-bond: Peptide bond -C - N- = O H -C = N- - O - - Resonance structures

21 Implications of Peptide Planes φ R C α H ψ C α Peptide planes H C α R ω angle varies little, φ and ψ angles vary alot Many φ/ψ combinations cause atoms to collide Each residue is sandwiched between two planes

22 Secondary Structure Local Conformation of Consecutive Residues Three low energy backbone φ/ψ combinations 1. Right-hand helix: α-, 3 10 helix (-40, -60 ) 2. Extended: antiparallel β-sheet (140, -140 ) 3. Left-hand helix (rare rare): ) α-helix (45, 45 ) Glycine is special- it has no side chain! Secondary structure is readily extracted from chemical shift assignments (Chemical Shift Index- CSI): ψ,φ calculation by Talos Software Hydrogen bonds between backbone atoms provides stability to secondary structures, NMR allows direct measurement of these bonds

23 Step 1: Identify Spin System

24 COSY: One coupling H H A N C H H B N C H H C N C

25 R-COSY: Add A 2nd Coupling H H H A N C C H H H B N C C H H C N C CH 3

26 DR-COSY: Add A 3rd Coupling H H H A N C C H H H H B N C C H H H C N C CH 3

27 TOCSY: All Coupled Spins H H H H A N C C C COOH H H H H H H H H B N C C C C C NH 3 H H H H H H C N C CH 3

28 Step 2: Fit Residues in Sequence A - B - C Peaks in NOESY spectra Same as scalar coupling peaks Peaks from residue i to i+1 A B (B C) Strong NOE Å Medium NOE Å Weak NOE Å

29 Distance Restraints for helix and sheet i+4 d αβ(i, i+3) d αn(i, i+3) d α(i)n(j) i+3 C N i+2 N C d NN(i, i+3) d αn (i, i+4) i i+1 C N N C i-1

30 NMR Structure Calculations Objective is to determine all conformations consistent with the experimental data Programs are available that perform automated assignment. NMR data are often not perfect: noise, incomplete data multiple solutions (conformational ensemble)

31 Assessing the Accuracy and Precision Variable Resolution Structures are reported Number of experimental restraints (A/P) Violation of constraints- number, magnitude (A) Compare model and exptl. parameters (A) Comparison to known structures: PROCHECK (A) Molecular energies (A) RMSD of structural ensemble (P)

32 Restraints and Uncertainty Large # of restraints = low values of RMSD The most important restraints are long-range

33 Accuracy/Precision Determined Differently for X- X-ray ray and NMR NMR RF Diffraction X-rays Pattern RF Resonance H0 Direct detection of atom positions Crystals Indirect detection of H-H distances In solution

34 Variability: Uncertainty and Flexibility in Experimental Uncertainty Structures X-ray NMR Avg. Coord. + B factor Ensemble Coord. Avg. Flexibility Mix static + dynamic Measure motions

35 Practical Issues: Sample Preparation 1D 100 μm Concentration: nd 500 μm Volume: 500 μl,, 200 μl 50 μm Cryoprobe 250 μm 20kDa 100 μm = 1 mg Purity > 95%, buffers Sensitivity (γ) isotope enrichment ( 15 N, 13 C)

36 Practical Issues: Molecular Weight *Symmetry reduces complexity* 8 x 10 kda 80 kda kda for 3D structure domains kda: residue-, site-, and atomspecific labeling, uniform deuteration TROSY (Transverse Relaxation Optimized [NMR] Spectroscopy)

37 Multi-Dimensional NMR: Built on the 2D Principle 90º pulse 3D- detect signals 3 times (t3) Same as 1D experiment t1 t2 t3 3D NMR Pulse Sequence Experiments are composites; Increase dimensionality Experiments are composites; Increase dimensionality of spectra to better resolve signals:

38 Acronyms For Basic Experiments Differ Only By The Nature Of Mixing Scalar Coupling Homonuclear COSY TOCSY Multiple Quantum Heteronuclear HSQC Hetero-TOCSY HMQC Dipolar Coupling NOESY NOESY-HSQC NOESY-HMQC

39 Double-Resonance Experiments Increases Resolution/Information Content t 15 N- 1 H HSQC R R - 15 N -C α -CO - 15 N -C α H H Detect signals from heteronuclei ( 13 C or 15 N) Better resolution of signals/chemical shifts not correlated between nuclei More information to identify signals Lower sensitivity to MW of protein

40 Large Scalar Couplings Less Sensitive to MW of the Protein Superior to 1 H homonuclear NMR: all J H-H < 20 Hz Mixing is faster so less time for signal to relax

41 Protein Folding and Fingerprints 15 N- 1 H HSQC 1 H COSY

42 15 N Dispersed 1 H- 1 H TOCSY 3 overlapped NH resonances with different side chains Add a 3rd dimension separating out Add a 3rd dimension separating out H N overlaps by their 15 N frequency

43 15 N Dispersed 1 H- 1 H TOCSY 3 overlapped NH resonances Same NH, different 15 N F2 F3 TOCSY HSQC F1 1 H 1 H 15 N t 1 t 2 t 3

44 Heteronuclear ( 1 H, 13 C, 15 N) Strategy One bond at a time - all atoms (except O) Even handles backbone 15 N 1 H overlaps disperse with backbone C C α H α C β H β Het. 3D/4D increases signal resolution 1 H 13 C 15 N 1 H Works on bigger proteins because one Works on bigger proteins because one bond scalar couplings are larger

45 Heteronuclear Assignments: Backbone Experiments Names of scalar experiments based on atoms detected Direct assignment of Consecutive residues (without NOESY )

46 Heteronuclear Assignments: Side Chain Experiments Multiple redundancies increase reliability

47 Approaches to Identifying NOEs 1 H- 1 H NOESY 2D 1 H 1 H 15 N- or 13 C-dispersed 1 H- 1 H NOESY 3D 3D 1 H 1 H 1 H 4D

48 Special NOESY Experiments Filtered, edited NOE: based on selection of NOEs from two molecules with unique labeling patterns. Labeled Unlabeled protein peptide Only NOEs at the interface Transferred NOE: based on 1) faster build-up of NOEs in large versus small molecules; 2) Fast exchange 3) NOEs of bound state detected at resonance frequencies of free state H H kon koff Only NOEs from bound state H H

49 NMR to Study Interactions Titration monitored by 15 N- 1 HHSQC HSQC Monitor the binding of molecules Determine binding constants (discrete off rates, on rates) Sequence and 3D structural mapping of binding interfaces

50 Binding Constants From Chemical Shift Changes Stronger Weaker Molar ratio of d-cttca Fit change in chemical shift to binding equation Arunkumar et al., JBC (2003)

51 Probe Binding Events by NMR 15 N-RPA32C + Unlabeled XPA N- 1 HHSQC HSQC 19 residues affected Discrete binding site RPA32C RPA32C + XPA 1-98 Signal broadening exchange between the bound and un-bound state Kd ~ 1 μm Mer et al., Cell (2000)

52 Map XPA Binding Site on RPA32C Using NMR N C Map chemical shift perturbations on the structure of RPA32C Can even map directly on to sequence with no structure Mer et al., Cell (2000)

53 Folding and Domain Structure Are domains packed together or independent? d RPA70 15 N A B 15 N 15 N A B Chemical shift is extremely sensitive If peaks are the same, structure is the same If peaks are different, the structure is 2 2 different but we 3 1 H 3 don t know how much 1 H 1 H 1 1 Arunkumar et al., JBC (2003)

54 Binding of histone H3 peptide to HP1 chromodomain requires H3K9 methylation Specific Non-specific Jacobs et al (2001) EMBO J.