Structure, dynamics and molecular interactions of
biological macromolecules by NMR
Winter School on Structural Cell Biology, CEITEC, Brno, Feb 9-13, 2015
Michael Sattler
http://www.nmr.ch.tum.de
http://www.helmholtz-muenchen.de/stb
http://www.bnmrz.org
Solution NMR methods to study protein complexes
 Ligand binding: CSP
 Optimized isotope labeling and NMR experiments
 Spin labeling: PRE (NMR), solvent PREs (sPRE)
 Large proteins, complexes, domain arrangements
Integrated structural biology of protein-RNA interactions
 Intron RNA recognition by multi-domain splicing factors (splicing regulation)
 [ Cooperative mRNA recognition by Sxl/UNR (translational regulation) ]
Outline
Structure/imaging from molecules to animals
Dynamics, timescales
Size,spatialresolution
Proteins,
domains
Protein complexes
Molecular machines
Cell
Cryo EM
EM tomography
Light
microscopy
Light
microscopy
NMR
10-9 1
X-ray NMR
103 [s]
MRI
Animal
MRI
SAXS, SANS
Staticpicture,
snapshots
Dynamics:
regulation
Chemical
Biology
Small molecules
Why solution state NMR?
Nature 2007
Nature 2007
Nature 2011
Nature 2009
In vitroHeLa
Biomolecular NMR
• Structure determination of biomacromolecules
 no crystal needed, native-like conditions: solution,
macromolecular crowding, “in cell” NMR (Xenopus oocyctes)
 nucleic acids: difficult to crystallize, affected by crystal packing
• Ligand binding and molecular interactions in solution
 “Band shift” in NMR fingerprint - with residue/amino acid resolution !!!
• Characterization of dynamics and mobility (ps  days)
 conformational dynamics  enzyme turnover, kinetics, folding
• Molecular weight: X-ray: >200 kDa,
NMR: de novo structure <50 kDa, but: binding/dynamics: 900 kDa
•  NMR and X-ray crystallography are complementary
0
50
100
150
200
250
0
50
100
150
200
250
140 160 180 200 220 240 260 280 300 320 340
T2[ms]
residue
www.pdbe.org
www.rcsb.org
Effect of exchange/dynamics on NMR spectra
A B NMR time scale: chemical shift,
i.e. resonance frequencies
k1
A  B
k1
kex = k1+ k1
• Exchange process can be binding, conformational exchange, chemical reaction…
• Line widths and resonance frequencies depend on the exchange rates and
frequency differences  of the interconverting states
• Exchange can allow transfer of magnetization in 2D NOESY-type experiments
• Rate constants can be determined, for conformational or binding equilibrium,
chemical reaction, ….
A, B: resonance frequency
A B
average
slow exchange
kex<<
fast exchange
kex >>
coalescence
kex~
Göbl et al Sattler Prog NMR Spectrosc (2014)
Effect of dynamics on NMR spectra
slow exchange
kex<<
fast exchange
kex >>
coalescence
kex~
NMRtoolstostudydynamics
Time scaleNMR expt.
Mittermaier & Kay, TiBS 2009
Population: 3:1
ωA–ωB = 100 Hz
k1
A  B
k1
kex = k1+ k1
EXSY
CPMG
RDC
PRE
spin relax.
10ms
1s
1ms
100 μs
10 μs
ns-ps
100ms
Two-site exchange: protein/ligand interactions by NMR
Limit Rates Populations Line broadening
Fast kA,B >> ( pA =  pApB kk
This can be used to determine, e.g. residue pKa values or dissociation constants Kd.
kon
P + L  PL
koff
KD= [P][L] / [PL] = kB/kA
kA = kon [L]; kB = koff
B = protein-ligand complex PL
A = free protein P
fast exchange
slow exchange
Slow pA/pB = areaA/areaBkA,B << (  k
kex = kon [L] + koff
Ligand binding in NMR titrations (fast exchange)
Fraction bound
[PL] ~  obs- free = f([Ltot])
bound
free
Binding in fast exchange
on the NMR chemical shift time scale
[P] < KD
Binding of a RG-rich peptide to SMN Tudor domain
equilibrium dissociation constant KD
from binding isotherm
koff
P + L  PL
kon
KD= [P][L] / [PL] = kB/kA
kA = kon [L]; kB = koff
B = protein-ligand complex PL
A = free protein P
KD > [P] (μM-mM)  KD can be fitted
Viral B2 protein dimer: inhibitor of RNAi
Chao et al, Williamson (2005) NSMB
19 bp dsRNA
R54
K62
K47
K47'
R54'
R36'
B2 dimer
C
2
C'
2'
1'
R36
K62'
10A
N
N'
Lingel et al (2005) EMBO reports
Ligand binding in NMR titrations (slow exchange)
Binding stoichiometry
(from inflection point of titration curve)
5’ GCAGCACGACUUCUUCAAGTT 3’
3’ TTCGUCGUGCUGAAGAAGUUC 5’
Binding in slow exchange
on the NMR chemical shift time scale
[P] > KD
Kd < [P] (nM)  binding stoichiometry can be determined
1:1 binding of B2 protein
dimer to a 21 nt dsRNA
1:1 binding of B2 protein
dimer (250 μM) to a 21 nt dsRNA
Ligand binding - stoichiometry
• Stoichiometry can only be correct if protein concentration is accurately determined!
KD = 100 nM  KD << [P]
NMR titrations – large complexes
Binding of a small ligand to a large protein:
Bound state may be broadened beyond detection.
Identification by NMR Spectroscopy of Residues at Contact Surfaces in Large, Slowly Exchanging Macromolecular Complexes.
Matsuo, et al & Wagner (1999) JACS 121, 9903-4.
Line shape simulation of a system in slow exchange with a bound
state having a large molecular weight.
The fraction of free protein is 0.5. Rf=23 s-1;Rb=250 s-1; koff = 25 s-1
Chemical shift differences ∆δ = 0 (black) and 500 Hz (red).
The resonance position of the free state, ω1=250 Hz.
Kinetics and thermodynamics from NMR line shape analysis
intercept = log(kB/h) = ΔSǂ/R
• kex is obtained from measuring transverse relaxation / linewidth fitting
• Temperature dependence allows to determine activation enthalpy and entropy
based on Arrhenius/Eyring transition state theory
Eyring equation
Bain A.D. Prog NMR Spectroscopy (2003) 43, 63-103.; Kessler H. Angew Chem (1970) 9, 219-235.
∆Hǂ
reaction coordinate
energy
Exchange spectroscopy (EXSY)
diagonal peak
cross peak
dMz(t) = (R+K) Mz(t)
aII(t) = aSS(t) = ½ exp{-t} [1 + exp(2kext)]
aIS(t) = aSI(t) = ½ exp{-t} [1  exp(2kext)]
ρ = R1 = 1/T1
kex = kAB+ kBA
For symmetric 2-site exchange:
kAB = kBA; kex = kAB + kBA
equal population: pA = pB
kAB
A ← B
kBA
kex = kAB+ kBA
→
1
DP
CP
2
 
 
   
 










 
0
0
,
,
,
,
zB
zAt
zB
zA
M
M
e
tM
tM KR













1
1
1
0
0
1
T
T
R 








BAAB
BAAB
kk
kk
K
A B
Cyclophilin A - proline cis/trans isomerase, HIV capsid protein
Exchange spectroscopy
Kern et al, PNAS 2002
cis
trans
diagonal peak
cross peak
NMR of large protein complexes: ClpP
Sprangers R et al. Kay LE PNAS 2005;102:16678-16683
Conformational exchange in ClpP
1H, 13C methyl TROSY NMR
of U-[15N,2H], Ile δ1-[13C,1H] ClpP
Exchange spectroscopy to quantitate the F,S exchange process
Sprangers R et al. Kay LE PNAS 2005;102:16678-16683
Two NMR signals disappear upon
mutation of a single Ile
• Ile δ1 sees 2 different
conformational states
• Confirmed by temperature
dependence
Ligand detected NMR screening:
Saturation Transfer Difference (STD)
• Saturation Transfer Difference (STD) NMR
• WATER-LOGSY, T2, diffusion filters, …
• Little amount of target protein needed
• No size limitation for target protein
• Provides binding epitope mapping  SAR
• Detect micromolar binders (KD 10-3-10-8)
or competition for nanomolar ligands
B Meyer et al , Angew Chem 1999; JACS 2001
reference 1D
STD
1D with T2 filter
reference
STD
STD with T2 filter
ProteinonlyProtein+ligand
NOE and ROE
NOESY ROESY
Extreme narrowing limit
Positive NOE
W2 >W0
Slow tumbling limit
Negative NOE
W2 << W0
2D NOESY spectra of a small ligand free and
bound to a protein
Free ligand (aDMA)
Negative NOE cross peaks,
Positive diagonal peaks
Protein-ligand complex*
(aDMA/SMN Tudor)
Positive NOE cross peaks
Positive diagonal peaks
Tripsianes et al, Nature Struct Mol Biol (2011)
*with large excess of protein to observe mainly bound ligand
1mM ADMA, 4mM SMN Tudor
1mM ADMA
Isotope edited/filtered experiments
13C,15N
12C,14N
13C,15N
12C,14N
editing filtering
13C,15N
12C,14N
edited/filtered
NOESY experiments
Select 13C/15N Reject 13C/15N
Principle combinations of editing/filtering
1H-[12C]
or
1H-[13C]
Intermolecular
NOEs
1H-[13C]
editing
1H-[12C]
filtering
Intermolecular
NOEs
ω1-edited, ω2-filtered NOESYω2 filtered NOESY
t2t1
Editing
or
Filtering
Editing
or
Filtering
1H
15N
13C
1H-[12C]
filtering
ω1
ω2
ω1
ω2
Editing/filtering can be applied before t1 and/or t2  1 and/or 2-edited/filtered correlations
Isotope filtered 2D NOESY
1H-[12C]or1H-[13C]
1H-[12C]
Intermolecular
NOEs
Triple 13C filter (2x 13Caliphatic, 13Caromatic) , single 15N filter
∆′+∆′′ ~ 8ms
1D (protein + RNA)
1D filter experiment
(RNA only)
Sattler, Schleucher, Griesinger, Prog NMR Spectrosc (1999)
3D edited/filtered NOESY of protein-RNA complex
1H 13C (t1) 1H(-13C) (t2)  NOE  filter  1H(-12C/14N) (t3)
1H(-13C)proteinω2
1H(-12C/14N) RNA ω3
3’ splice site recognition in constitutive splicing
• Essential early step in pre-mRNA splicing
• Regulation of alternative splicing during spliceosome assembly
• Cooperative recognition of 3’ splice site by U2AF and SF1
65 35 Exon2Exon1 Intron
5’ ss 3’ ss
splicing
Complex E
Structural modules at the 3’ splice site
UACUAAC
QUA2 KH
5’ 3’
U2AF65
RRM2 RRM1
RRM3
RRM
U2AF35
P
UUUUUUU AG
SF1
Liu, Luyten et al.
Science (2001)
Kielkopf et al.
Cell (2001)
Selenko et al
Mol. Cell. (2003)
Ito et al. EMBO J. (1999); Sickmier et al Mol.Cell (2006); Mackereth et al Sattler Nature (2011) i
Dynamics in multi-domain protein interactions
RNA
closed/autoinhibited
preformed domain
contacts
independent domains
RNA looping
domain contacts,
additional domain binding
(intra- or intermolecular)
RBD1 RBD2
independent binding

free bound
Multi-domain dynamics Multiple register binding
Mackereth & Sattler, Curr Op Struc Biol 2012
sliding,
multiple registers
specific,rigid
multiple
binding sites
NMR approaches for studying large complexes
• 3D structure of subunits available (X-ray, NMR, ROSSETTA)
• Subunit-selective/segmental labeling
Sortase A ligation, optimized 2H-labeling
 sensitivity, spectral simplification
• Conformational dynamics: NMR relaxation
• Domain interfaces/intramolecular contacts
Chemical shift perturbations (CSP)
PREs (spin labeling): ~ 1/r6, <20Å
Solvent PREs
• Domain arrangements
Residual dipolar couplings (RDCs)
Small angle scattering (SAS)
• Structure calculations
Joint refinement NMR data / scoring with SAXS/SANS
Simon, et al Angew. Chem (2010) ; Madl et al JACS (2010) ; Madl et al Angew. Chem (2011);
Madl et al J Struct Biol (2011); Hennig & Sattler Protein Sci (2014); Göbl et al Prog NMR Spec (2014)
RDCs
SAXS/SANS
PRE
(spin label)Solvent
PRE
Py tract RNA recognition by U2AF65 RRM1-RRM2
U2AF65/Py tract RNAU2AF65
RRM2 RRM1
UUUUUUU
0
50
100
150
200
250
0
50
100
150
200
250
140 160 180 200 220 240 260 280 300 320 340
Transverserelaxation
T2[ms]
“Flexibility”
U2AF65 RRM1-RRM2
+ U9 RNA
U2AF65
UACUAAC UUUUUUU AG
SF1
U2AF35
• U2AF is an essential splicing factor, required for intron Py tract RNA recognition
• U2AF65 RRM1-RRM2 necessary and sufficient for Py tract RNA binding
• Two structural domains, connected by a flexible linker
flexible
linker
Residue Number
Py tract (U2 introns) 3’ ss
Subunit-selective labeling
Utility of 2H-labeling: SF1/U2AF/RNA (74 kDa)
TROSY 600 MHz, 0.2 mM, 16 hours
[50%-2H,15N] U2AF65 + SF1 + RNA (74 kDa)
TROSY 600 MHz, 0.2 mM, 2 hours
[U-2H,15N] U2AF65 + SF1 + RNA (74 kDa)
U2AF65SF1
UACUAAC UUUUUUU
Gardner & Kay (1997) JACS 119 7599
Goto et al. (1999) J. Biomol. NMR 13 369-374
13CH3 Val, Leu
E.coli
13CH3 1 Ile
Random fractional 2H-labeling
• Grow bacteria in 70-90% D2O
 random fractional (60-80%) 2H-
labeling
• Cost-effective
• But: presence of 13CHx isotopomers
 combine with CH multiplicity filters
Sibille et al (2002) JACS 124 14616-25
Gardner & Kay (1998) Ann Rev Biophys Biomol Struct 27 357-406
[3-2H]-13C -ketoisovalerate
[3,3-2H]-13C -ketobutyrate
H3
13C―13CD2― 13C―13CO2
-
O
║
13CD― 13C―13CO2
-
O
║
H3
13C
H3
13C
Metabolite Amino Acid
Random fractional deuteration and methyl-selective 1H,13C labeling
ILV labeling:
methyl-13C,1H for Ile, Leu, Val
Ollerenshaw, et al Kay JBNMR 2005
Residual dipolar couplings (RDCs)
Residual dipolar coupling
Dij ~ 1/rij
3 <(3cos2θ-1)>
!= 0 in anisotropic solution
B0
N
H
In anisotropic solution:
• D!=0  orientation
• Weak (10-4) alignment in dilute (3-5%) liquid crystalline medium
Domain orientation from RDC data
Template structure
with optimized
domain geometry
Randomize linker
and/or initial
domain orientation
Simulated annealing
no positional constraints
for unknown regions
Cluster structures
and analyze relative
domain orientation
z
x
y
180°x 180°y 180°z
25% 25% 25% 25%
Clustering of domain orientation:
• Structural noise
• # restraints per domain
• Forces NCS vs. RDC
Simon, et al (2010) Angew. Chem.
Domain orientation with two alignment tensors
X-ray NMR
Homology
model
RRM1
RRM2
Tensors
Phage PEG Phage PEG Phage PEG
Important factors for
domain orientation:
• Structural noise (input)
• # RDCs per domain
• Coupling to reference
Simon, et al (2010) Angew. Chem.
How to make your protein paramagnetic:
Metal-binding proteins
 Paramagnetic metals binding sites
 PRE, PCS, RDC
Paramagnetic tags (spin labels)
 nitroxide radicals
 lanthanide-binding peptide tags
 protein fusions with LBTs
 covalently linked to cysteines,
 4-thio-uracyl, 2’ amino (RNA)
 PRE, PCS, RDC
Soluble paramagnetic agents
 nitroxide radicals, ions, chelates
 Solvent PRE
PRE, PCS, RDC
Metal-binding
protein
Tag
Peptide
Fusion
Solvent PREs
Madl. et al Angew Chemie (2009, 2011); Otting JBNMR 2008
Göbl et al Prog NMR Spec (2014)
NMR restraints from paramagnetic effects
Spin labeling of proteins and nucleic acids

IPSL 3-[2-iodoacetamido]-proxyl
N
CH3
CH3
CH3
CH3
O
CH2
O
CNH S
N
N
R
O
4-thiouracil proxyl
Varani JACS (1998)
IPSL
3-[2-iodoacetamido]-proxyl
RNA spin labeling:
Chemically synthesize thiouracil RNA oligo
Protein spin labeling:
Recombinant protein with single
Cys mutant proteins
 site-directed mutagenesis
MTSL often used (EPR, NMR)
IPSL chemically more stable,
but also less reactive


MTSL
Interdomain distance restraints from PREs
(paramagnetic relaxation enhancement)
• PRE ~ r6 (electron-spin distance)
•  long-range distance restraints (<20 Å)
•  multiple single-Cys mutants of protein ( molecular biology)
• Measure transverse PRE R2
PRE from sample with oxidized (Ipara) and reduced (Idia) spin label
Paramagnetic
(Ipara)
Reference expt
(Idia)







 2
c
2
H
c
c6
2
B
22
H
PRE
2
τω1
3τ
4τ
r
1
μg1)γS(S
15
1
R
 
PREdia
PRE
dia
dia
para
RR
R
R
I
I
22
2
2
exp




Measuring distances between domains
Spin label
Domain 1 Domain 2
Ligand
IPSL 3-[2-iodoacetamido]-proxyl

Spin labeling
Battiste & Wagner Biochemistry (2000); Simon, et al Angew. Chem. (2010); Madl et al J Struct Biol (2011)
Measuring 1HN PRE as 2 directly
Donaldson et al Kay, J.Am.Chem.Soc. (2001)123, 9843–9847. Iwahara et al.Clore J Mag Res (2007) 184,185–195
• 2-point measurement of exponential decays
• accurate, systematic errors (3JHN,Ha) cancel
• Set Ta=0,
• Tb=1.15/(R2,dia+2)
to minimize error in 2
PRE in the presence of exchange/dynamics
Assume:
Otherwise, if:
need to now  and kex
PRE may become independent of r
Fast exchange allows detection of minor species
Clore & Iwahara, Chem Rev (2009)
Battiste & Wagner Biochemistry (2000), Simon et al Angew Chem (2010)







 226
222
2
1
3
4
1
g1)S(S
15
1
cH
c
cBH
PRE
r
R



6
22
2 1
3
4
1
*Å370 







cH
c
cPRE
R
r



• Distance calibration: linear approximation for 0.2 < Iox/Ired < 0.8
• Estimate c from (R2/R1)ox and (R2/R1)red
Note: c refers to the electron-nuclear spin vector!
• Grid search for correlation time c for each SL
c=12ns, R2
dia 50Hz; ± 4Å
Distance r [Å]
Ipara/Idia
Paramagnetic Relaxation Enhancement (PRE)
 
PREdia
PRE
dia
dia
para
RR
R
R
I
I
22
2
2
exp




Experimental vs. calculated distance
Back-calculated distance [Å]
PREderivedtargetdistance[Å]
Estimation of the electron-spin correlation time c
• Need to determine/estimate c from (R2/R1)ox and (R2/R1)red
• Grid search for correlation time c for each SL
Spin label flexibility and c of the electron - HN vector
SL318
Flexibility of the spin label
• Consider internal flexibility and conformational space
sampled by the spin label by a ensemble
representation (i.e. 4 copies per spin label site)
• ensemble averaged distance restraints during
structure calculations
Simon, et al Angew. Chem. (2010); Hennig et al, Sattler Methods Enzym (2015)
Iwahara, Schwieters, Clore JACS (2004) 126,5879-5896
Structure calculation from RDC + PRE data
273
287
318
281
209
155
164
171
188
187
Interdomain
distance restraints
Domain arrangements from PRE data
• Individual domain structures available
• Spin labeling  paramagnetic relaxation enhancements (PRE)
•  distance restraints to define interdomain arrangement
Simon, et al Angew. Chem. (2010); Madl et al JACS (2010) ; Mackereth et al Nature (2011)
Iwahara, Schwieters, Clore, JACS 126, 8579 (2004); Clore & Iwahara, Chem Rev (2009);
IPSL
PRE data define the domain arrangements
RRM1 RRM2residueRRM1 RRM2residue
No RNA
RRM2
RRM1
Open and closed conformations of U2AF65
Bound to U9 RNA

RRM2 RRM1
Mackereth et al Nature (2011)
open closed
Solution conformation differs from crystal structure
RRM1 RRM2
NMR
RRM1
RRM2’
RRM1’
RRM2
RNA
RRM1+2 ∆linker
RRM2
RRM1
Importance of using solution methods
for studying multidomain proteins
X-ray
Sickmier et al Mol. Cell (2006)
RDCexp[Hz]RDCexp[Hz]
RDCcalc [Hz]
RDCcalc [Hz]
UUUU
UUUUAAAA
UUUUAAAAAAAAUUUU
UUUUAAAAUUUU
UUUUUUUUU
UUUUUUUUUUUUU
KD /μM
33
18
16
7.1
1.3
0.35
RRM2 RRM1
U4
U4A4
U4A8U4
U4A4U4
U9
U13 always
binds
variable
binding
Conformational shift measures Py tract “strength”
Human U2 introns
Py tract 3’ ss
Pytract“strength”
UUUU
UUUUAAAA
UUUUAAAAAAAAUUUU
UUUUAAAAUUUU
UUUUUUUUU
Kd /μM
33
18
16
7.1
1.3
unbound
U4
U4A4
U4A8U4
U4A4U4
U9
Population shift of distinct domain arrangements
Pre-existing “bound” conformations in free RRM1-RRM2
RNA bound
free
PRE calculated PRE measured
Consistent with
free structure
Consistent with
population of bound
form
r [Å]
RRM1-RRM2
(compact & non-compact states)
0 20 40 60 80
Pairdistributionfunction(a.u.)
Free U2AF65 samples non-compact conformations
• Small Angle Scattering data indicate non-compact conformations in free RRM1,2
 free RRM1,2 is an ensemble of compact and non-compact states
• In contrast, RRM1,2/RNA is compact
detached/
non-compact
SAXS P(r)
RRM1-RRM2/RNA
(compact)
Ensemble of RRM1,2 based on NMR and SAS data
Ensemble with randomized
linker
Unbiased/unrestrained selection of
conformations and populations
Monte-Carlo based error analysis
Jie-Rong Huang, Martin Blackledge
experiment
calculated
RDC
PRE
SAXSRRM1
RRM1
full ensemble (pool) selected
open
closed
Center-of-mass of RRM2: cross validation
open
closed
Crystal
structure
x
Ensemble of free states selected from NMR & SAXS
open
closed
open
closed
Pool
Selected (NMR & SAXS)
open
(1%)
RRM1
RRM2
open
RRM2
closed
Enriched conformations
(difference F-E)
Center-of-mass Y coordinate [Å]
Center-of-massXcoordinate[Å]
Huang et al, JACS (2014)
RRM2
closed
RRM2
open
Ensemble of free states selected from NMR & SAXS
Hydrophobic
surface that
remains
accessible
electrostatic
interactions
~50% of conformations are encounter-like, i.e. compact domain
arrangement (consistent with 15N NMR relaxation data)
RRM1
RRM2 ensemble
Huang et al, JACS (2014)
Modulation of encounter-like domain interactions
• PRE for spin-labeled A318C RRM1,2 at different salt concentrations
• Encounter-like charged interactions are salt dependent
0 mM, 50 mM, 150 mM, 200 mM
Complex mechanisms of RNA recognition in solution
closed, inactive
open, active
RRM2 RRM1
RRM2 RRM1
Autoinhibition by linker  proof-reading
Dynamic ensemble of inactive states  conformational entropy
Conformationalselection
Inducedfit/fly-casting
RRM2RRM2
Key recognition elements in the ternary complex
Non-canonical
SXL-RNA contacts
Internal
base
stacking
Ternary “triple
zipper”
contacts
SXL
UNR-
CSD1
msl2-
mRNA
Large induced fit of the RNA ligand and Sxl/CSD domain arrangement
Hennig, et al, Nature (2014)
Structure validation in solution by NMR – UNR-CSD1
Ternary “triple zipper” contacts
0.2
183
193
203
213
223
233
243
252
0.4
0.6
0
UNR-CSD1
Relative domain orientations in solution from NMR RDCs
RDC data in ternary complex agree with domain orientation in crystal structure
-40
-20
0
20
40
-50 -30 -10 10 30
SXL phages HN-N
QRDC
= 0.41
ExperimentalRDC[Hz]
60 Couplings
Back calculated RDC [Hz]
D = -12.41
R = 0.17
SXL
-6
-2
2
6
-5 -3 -1 1 3
SXL phages N-C
QRDC
= 0.44
47 Couplings
Back calculated RDC [Hz]
ExperimentalRDC[Hz]
CSD1 phages HN-N
Back calculated RDC [Hz]
QRDC
= 0.39
-60
-20
20
60
-60 -40 -20 0 20 40
31 Couplings
D = -14.80
R = 0.28
CSD1
ExperimentalRDC[Hz]
Joint NMR/X-ray refinement possible
(collaboration with Claudio Luchinat)
Structure validation in solution by SAXS and SANS
• SAXS and SANS data fully corroborate the
crystal structure in solution
• MW I(0) = 34.2 kDa, expected: 33.5 kDa
DAMMIN-2.5
-1.5
-0.5
0.5
1.5
0 0.1 0.2 0.3 0.4
SAXS data
fit
Rg = 2.29 nm
2
= 0.77
18-mer
log(I)
Q (Å-1
)
SAXS
1
H-SXL
2
D-CSD1
1
H-RNA
42 % D2O
1
H-SXL
2
D-CSD1
1
H-RNA
70 % D2O
-4
-3
-2
-1
0.05 0.15 0.25 0.35
log(I)
SANS data
fit
2
= 1.09
2
D-SXL
1
H-CSD1
1
H-RNA
42 % D2O
2
= 2.17
2
D-SXL
1
H-CSD1
1
H-RNA
70 % D2O
-4
-3
-2
-1
0.05 0.15 0.25 0.35
log(I)
SANS data
fit
2
= 0.92
180°
-4
-3
-2
-1
0.05 0.15 0.25 0.35
log(I)
Q (Å
-1
)
SANS data
fit
2
= 1.35
Q (Å
-1
)
-4
-3
-2
-1
0.05 0.15 0.25 0.35
log(I)
SANS data
fit
f
Q (Å
-1
) Q (Å
-1
)
MONSA
shape
SANS
 


Summary
SAXS/SANS RDCs + PRE
EPR (PELDOR)
spin-label
• Structure and dynamics of protein complexes in solution:
• RDCs for relative domain arrangements
• PREs/ spin-labeling for long-range distance restraints
• PELDOR/DEER to measure specific distances and detect dynamics
• Sensitive, no limitations by molecular weight, spin-labeling required
• Solvent PRE to detect and refine domain interfaces
• Simple to measure, no protein modification required, dynamics affects analysis
• SAXS as complementary technique
• Detect conformational equilibria/dynamics
• Joint structural refinement
• Need to combine with additional experimental data to reduce/resolve ambiguities
• Structural dynamics of multi-domain RNA
binding proteins is important for their functional
activity
• Cooperative binding of multiple RNA binding
domains (RBDs) expands the protein-RNA
interaction network to regulate diverse biological
functions with a limited set of RBDs:  proteinRNA
recognition code
Conclusions
• Integrated structural biology –solution techniques, i.e. NMR, SAXS,
SANS to study dynamics of multi-domain proteins and complexes
(ensemble)
RNA binding
protein 1
RNA binding
protein 2
Funding