epr @eth Molecular and Structural Biology V: Studying ...€¦ · We need an unpaired electron 2 e Three basic types of systems HN O CH2 O Radical enzymes Metalloproteins Spin labels

Molecular and Structural Biology V: Studying Macromolecules by NMR and EPR

Gunnar Jeschke, HCI F227, [email protected]

epr@eth

Two Lectures · general introduction to EPR techniques & intrinsic paramagnetic centers in biological systems (19.4.)

· spin labeling & structure modeling (26.4.)

Script F in-depth discussion, reference for future research work (epr.ethz.ch/education.html)

One tutorial · simulating EPR spectra with EasySpin (two short examples)

· analyzing DEER data in terms of a distance distribution (two examples)

· rotamer library simulation of spin labels and comparison to DEER data

· localization of a spin label site in a protein

Examination content will be specified at the end of semester

1MSBV-EPRSpectroscopy

The focus is on information from EPR and its use in structural biology, not on inner working and theory of EPR

(see „The EPR part of the ETH Magnetic Resonance lecture script“ at epr.ethz.ch/education.html)

(31.5.)

We need an unpaired electron

2

e

Three basic types of systems

HN

O

CH2

O

Radical enzymes Metalloproteins Spin labels

· electron transfer reactions in

cell energetics and metabolism

· electron transfer reactions and

catalysis of reactions

· studies of structure and dynamics

on diamagnetic macromolecules

Fe

S

S

XNiOC

NC

NC

S

S

H2C

CH2

H2C

H2C N O

SSH3C

O

O

Electronic structure, identity of nuclei, proton coordinates Probing the environment

Nanometer-range distance distributions

tyrosylradical

MTSSL

[FeNi] hydrogenase

MSBV-EPRSpectroscopy

What is a SOMO?

Singly occupied molecular orbital

3

e

n

e0

e-1

e-2

e1

HOMO

LUMO SOMO

Visualization of the SOMO of a tyrosyl radical

The SOMO can be probed by hyperfine couplings to nuclei

1· H hyperfine couplings related to spin density on the adjacent heavy atom

· g tensor related to global properties of the SOMO via spin- coupling orbit


The SOMO and free radical EPR spectra

4

e

n

3330 3340 3350 3360

Magnetic field [mT]Magnetic field [mT]

Magnetic field [mT]

336 338 340 342 344 346 348

336 338 340 342 344 346 348

Magnetic field [mT]

3330 3340 3350 3360

X band: 9.5 GHz W band: 94 GHz

HN

O

CH2

O

hyperfine

hyperfine gx

gz

gy

still no gresolution

hyperfine-dominated better g resolution, worse hyperfine resolution

NH

O

H2C

N


5

An overview of microwave bands and interactions

m.w. bands S X Q W

1mK 4.2 K 48 KThermal energy

Electron Zeeman interaction

Zero field interaction

Hyperfine interaction

10 10 10 10 104 6 8 10 12

Dipole-dipole interactionbetween weakly coupled electron spins

distance 10 5 3 1 nm

Homogeneous EPR linewidths

Homogeneous ENDOR andESEEM linewidths

10 10 10 10 100 2 4 6 8

kHz MHz GHz

10 1010

THz

Hz

Hz

12

S band 2-4 GHz

X band 9-10 GHz

Q band 33-35 GHz

W band 94-95 GHz

· for electron group spin > 1/2 (more than one unpaired electron)

· resolution limit depends on sample preparation

· most valuable source of EPR restraints on structure

· hyperfine resolution better when measuring on the nuclear spin


More than one unpaired electron

6

e

e

HOMO

Triplet state (S = 1)

· the two magnetic moments couple through space

(dipole-dipole coupling)

· they are both spatially distributed in their orbitals

· for heavier atoms, especially transition metals,

zero-field splitting has a spin-orbit contribution

· this causes zero-field splitting

(typically 300 MHz¼ 2 GHz)

magnetic field (mT)

250 300 350 400 450

D = 1000 MHz

E = 0 MHzD

2D

x,y

z

Axial symmetry

Orthorhombic symmetry

D = 1000 MHz

E = 100 MHz

2D

magnetic field (mT)250 300 350 400 450

y

xz

D - 3E

D + 3E


More than one electron in metal centers

7

5 III IIBare 3d metal ion (Fe , Mn ) Weak ligand field Strong ligand field

S = 5/2 (Hund’s rule) S = 5/2 (high spin)S = 1/2 (low spin)

e

e

e

ee

00 100100 200200 300300 400400

Magnetic field [mT]Magnetic field [mT]

g = 2.00

g = 5.99

g = 6.75

g = 5.25

g = 1.98

IIIHigh-spin Fe : large ZFS and effective g values IIILow-spin Fe : smaller g dispersion

axial symmetry Type II (P450 )camorthorhombic symmetry Type I (Myoglobin-CN)

200 200400 400600 600800 800

Magnetic field [mT] Magnetic field [mT]

g = 3.45

g = 1.89

g = 0.93

g = 2.41

g = 2.25

g = 1.91


Fingerprinting metal ions

8

100 200 300 400

IICu

IIIFe high-spin

IIMn

magnetic field (mT)

X-band EPR of common metal ions IIDifferent types of Cu centers

S = 1/2, I = 3/2

S = 5/2ZFS >> Electron Zeeman

S = 5/2, I = 5/2ZFS << Electron Zeeman

260 280 300 320 340 360

Magnetic field [mT]

260 280 300 320 340 360

Magnetic field [mT]

Type II

Undistorted Type I Distorted Type I

particulate methanemonooxygenase

poplarplastocyanin

stellacyanin

260 280 300 320 340 360

Magnetic field [mT]

II· the half-filled shell (3d5) of Mn makes the g tensor and

hyperfine coupling almost isotropic


9

Some are invisible

Kramers and non-Kramers ions

e

e

e

e

Large ZFS may split all levels by more than the microwave frequency

Non-Kramers ions may be “EPR silent”

2· to first order ZFS contribution is proportional to mS

II 6 III 6 II 8· typical cases: Fe (3d , S = 2), Co (3d , S = 2), Ni (3d , S = 1) in their high-spin states

· low-spin states of ions with an even number of unpaired electrons are diamagnetic (S = 0)

Þ usually, metal ions are only seen when they have an odd number of unpaired electrons

· for half-integer spin (S = 1/2, 3/2, 5/2, 7/2), there are ±m pairs of levels that are degenerate in zero field: S Kramers ions

· whatever spectrometer you have, there is a field/frequency combination where you can excite transitions of Kramers ions

· for integer spin (S = 1, 3, 5), all levels are split to first order by ZFS at zero field (unless symmetry is axial): non-Kramers ions

· if ZFS is larger than microwave frequency plus electron Zeeman interaction at maximum field,

no transition can be excited for non-Kramers ions


10

Weakly coupled electron spins

Exchange coupling

Dipole-dipole coupling

· arises from overlap of the SOMO's of two electrons

- binding overlap « antiferromagnetic coupling « DE =DE < DE =DEab ba aa bb

- anti-binding overlap « ferromagnetic coupling « DE =DE > DE =DEab ba aa bb

· strong exchange coupling (J > gm B /h)B 0

- antiferromagnetic: diamagnetic singlet ground state

- ferromagnetic: paramagnetic triplet ground state

local field invertedlocal field

r r N

S

Exte

rna

l ma

gn

eti

c fi

eld

for weak g anisotropy

w /2p » 52.04 MHz at r = 1 nmdd 12

e

e

Unless orbitals strongly overlap,exchange coupling is negligible at r > 15 Å

Exchange coupling decreases exponentially with distance r


11

Interactions and the information that they provide

electron Zeeman fingerprinting of radical type or metal coordination

nuclear Zeeman identification of nuclei that give rise to hfi

nuclear quadrupole binding situation of the nucleus for I > 1/2 (chemical shift is not available)

hyperfine distribution of the SOMO (reactivity)

distance of protons from the center of spin density

zero-field fingerprinting of triplet type or metal coordinationspin state for metal ions (low or high spin)

exchange orbital overlap (important for electron transfer)

dipole-dipole distances in the nanometer range (15 - 100 Å)

Name Information

e

e

n

n

n

n

n

n

B0


Measuring hyperfine couplings

12

14N

14N

14N63,65Cu (spin center)

resolved in EPR

resolved in EPR

possibly resolved in EPR, elseENDOR

ESEEM orHYSCORE ESEEM or

HYSCORE

ENDORor HYSCORE

ENDORunresolved

17oxygen is normally invisible, but can be made visible with O if the problem justifies the expense


What is CW EPR?

13

DB0

DBpp

DV

DV

microwave source

reference arm

bias

attenuator

resonator

magnet

modulationcoils

1

23

f

m.w.diode

PSD

modulationgenerator

Sig

nalcirculator

phase

Points to remember

Field modulation

· signal increases linearly with modulation amplitude, until it starts to broaden (use 2 G amplitude at the beginning)

· signal increases proportionally to the square root of microwave power (factor 2 per 6 dB less attenuation)

until it starts to broaden, level off, and eventually to decrease again (use 20 dB attenuation at the beginning)


14

When can and should CW EPR be applied?

CW EPR is the first experiment to be applied to any unknown sample

Hardware requirements: basic CW EPR spectrometer (widely available, cheap)

Sensitivity : radicals >1 mM to 10 mM

metal ions >10 mM to 100 mM

Aggregation state : liquid & solid

Special requirements : liquid polar solvents (aqueous buffer) require special sample geometry for best sensitivity (flat cells or bundles of capillaries) if utmost sensitivity is not an issue, a capillary will do nicely

Information : type of paramagnetic center (may require high field) large hyperfine couplings rough idea on relaxation by playing with microwave attenuation spin quantification (comparison of double integral with the one of a reference sample)


15

What is ENDOR?

Electron nuclear double resonance

e

n

p

p/2 p

p

t tT

microwave

radiofrequency"frequency-swept NMR"

Davies ENDOR

The is observed as a function of echo amplitude radio frequency

Davies ENDOR in water

in deuterated water

Difference spectrum

simulation

N

NN

NCo

N

N

N

N

OO

O

2.3

A

110°

H

CO2Et

O

O

CO2Et

O

EtO2C

O

EtO2C

b

hemoglobin model compound


16

When can and should ENDOR be applied?

ENDOR is applied if hyperfine couplings are unresolved in CW EPR and too large for ESEEM/HYSCORE

Hardware requirements: pulse EPR, radiofrequency channel



Aggregation state : solid (liquid state requires rarely available CW ENDOR)

Special requirements : longitudinal relaxation time of at least 10 ms signals of different isotopes overlap at X band, high field may be required in some cases

Information : large and moderately sized hyperfine couplings nuclear Zeeman frequency nuclear quadrupole coupling


17

What is HYSCORE?

Hyperfine sublevel correlation

e

n

1n( H)1

-n( H)0

n1

1

2n(H)

A /23

p/2 p/2 p/2

p

t tt1 t2

wa wa

wb wb

aa aa

ab ab

ba ba

bb bb

A1

A2

n2

· correlates frequencies of the same nucleus for the

a and b state of the electron spin

· the 1D version without the p pulse is called 3-pulse ESEEM

e

-3µ r

large couplings small couplings

curved correlation ridgescontain informationon hyperfine anisotropy

1Þ H-electron spin distance


18

When can and should HYSCORE/ESEEM be applied?

HYSCORE is applied if hyperfine couplings are unresolved in CW EPR

Hardware requirements: pulse EPR



Aggregation state : solid

Special requirements : transverse relaxation time of at least 100 ns anisotropic hyperfine couplings hyperfine coupling of the same order of magnitude as twice the nuclear Zeeman frequency

Information : small and moderately sized hyperfine couplings nuclear Zeeman frequency nuclear quadrupole coupling

1 separation of isotropic and anisotropic hyperfine contributions ( H distances)


19

-30 -20 -10 0 10 20 300

10

20

30

-30 -20 -10 0 10 20 300

10

20

30

n2 /MHz

n2 /MHz

n1 / MHz

n1 / MHz

gz

Ni

OGlna‘147

Hg

Hg

Hg Hb1

Hg Hb2

SO

SO

3

3

-

-

N

N

N

NNi

NN

NN

Nin+

HN

H2NOC

H3C

–OOC

O

CH3

COO–

COO–

COO–

–OOC

O

H

H

O

H2NGln

a'147F430

gb

aBr

13C

HINDERBERGER D. et al., Angew. Chem. Int. Ed. 2006, 45, 3602-3607

3-bromopropane sulfonate

Active center of the enzyme 13C signals in HYSCORE: hyperfine coupling Hyperfine coupling reveals

the binding mode

Inhibitor

· binds to the enzyme (step 1)

relative orientation of

hyperfine and g tensor

can be determined

· cannot be reduced to methane

(step 2 blocked)

Experiment

Simulation

13C

13C

13C

Þ

e

n

Binding mode of an inhibitor to methyl-coenzyme M reductase


20

What is DEER?

p p p

p

t1

t

t = 0

t1 t2 t2

wA

wB

2

pg

intermolecularbackground(exponential decay)

Numberof spinsin the object

1/Dn3

(~ r )

V(t

)/V

(0)

F(t

)

0 5 100

0.5

1

t (µs)0 2 4 6 8

0.6

0.8

1

t (µs)

Hex

Hex

NO•

N

O

O

N O•

N

O

O(CH2)6OMe

(CH2)6OMe

D

Model compound

Primary data Form factor

PANNIER M, VEIT S, GODT A, JESCHKE G, SPIESS HW, J. Magn. Reson. 2000, 142, 331

MARTIN RE et al., Angew. Chem. Int. Ed. 1998, 37, 2834

Tikhonov

regularizationDistance distribution

JESCHKE G et al. Appl. Magn. Reson. 2006, 30, 473

JESCHKE G et al. J. Magn. Reson. 2002, 155, 72

2 3 4 5 6

0

0.01

0.02

0.03

r (nm)

P(r

)

e

e

The is observed echo amplitudeas a function of time t


21

When can and should DEER be applied?

DEER is applied to measure distances in the range from 15 Å up to 60 (membrane proteins) or even 100 Å(world record at 160 Å in fully deuterated GroEL)

Hardware requirements: pulse EPR, second microwave frequency (ELDOR) or arbitrary waveform generator (AWG)


metal ions >50 mM to 0.5 mM

Aggregation state : solid

Special requirements : transverse relaxation time of at least 500 ns (unless distance is very short) absence of exchange coupling for straight distance determination orientation of spin-spin vector to magnetic field uncorrelated to spectral selection (for straight distance analysis)

Information : distance distributions or, at the long limit, mean distances between electron spins number of spins in the same macromolecule or complex orientation of the spin-spin vector in the molecular frame (high field, larger effort)


22

DEER example: Localization of the N-terminal domain in LHCII

Primary data Distance distributions

0 0.5 1 1.5 2t (µs)

3

10

9

7

4

3

11

12

14

22

29

34

4

7

9

10

11

12

14

22

29

34

2 4 6 8

r (nm)r (nm)

2 4 6 8

22 44 66 88

2 4 6 8

experimentaldistance distribution

rotamer predictionfrom crystal structure

DOCKTER, C; MULLER, AH; DIETZ, C; VOLKOV A; POLYHACH Y; JESCHKE, G; PAULSEN H J. Biol. Chem 2012, 287, 2915

Unrestrained

DEER restrained


The art of sample preparation

23

Concentration

Oxygen

Cryoprotectant

Sample freezing

· too high concentration in liquid state : exchange broadening (stay < 1 mM¼ 200 mM for radicals, < 2 mM for metal ions)

· too high concentration in solid state : dipolar broadening, shorter phase memory time (stay below 200 mM/1 mM)

· at very high (local) concentration, hyperfine structure may collapse (exchange narrowing)

3· O is a paramagnetic line-broadening agent, especially in unpolar solvents, detergent micelles, and lipid bilayers2

· biomacromolecules don’t like ice crystals, structure distortion and precipitation may occur

· immersion of the tube in liquid nitrogen: freeze-quench to 80 K in a few seconds, limited by gas bubbles (poor heat conduction)

· weaker effects in the solid state, but relaxation times may shorten

· 10% glycerol may suffice for liposomes, 25% for soluble proteins, 50% makes freezing simple

· immersion of the tube into iso-pentane or ethanol cooled to 120 K: freeze quench to below glass transition in shorter time

· DMSO can be used for DNA/RNA

· spraying of the sample onto a silver wheel that rotates in liquid nitrogen, collection of the “snow”: about 40 ms freeze time


Optimizing relaxation time for pulsed EPR

24

Long T (T ), but short T2 m 1

Prolonging the low-temperature limit of the phase memory time Tm

· transverse relaxation limits resolution and pulsed EPR sensitivity

· nuclear spin diffusion generates fluctuating hyperfine fields Þ dominating phase memory loss mechanism for electron spin

in the low-temperature limit

· concentration of nuclei with high gyromagnetic ratio must be reduced: deuteration helps

use D O in the buffer2

use d -glycerol as cryoprotectant8

deuterate recombinant protein by using D O in the growth medium2

deuterate recombinant protein even better by feeding deuterated glucose in minimal medium

reconstitute membrane protein into deuterated lipids (or solubilize in deutrated detergent)

· check, whether concentration limits T by instantaneous diffusion m

(for DEER to measure very long distances, 100 mM may be too much)

· if all is done and it still does not suffice, work in the absence of oxygen (if you can)

· too long T requires long waiting times between experiments, optimum 100 ms to 1 ms1

· T attains a low-temperature limit (~50 K for radicals, ~10 K for S = 1/2 metal ions) 2

increasing expenseand effort


Spin labeling


Measurement

Labeling and

sample preparation

· where to label?

· which label?

· what matrix (solvent, detergent, liposome, nanodisc, deuteration)?

· what concentration?

· what frequency?

· how much sample?

· which experiment and what parameters?

· optimization of spectrometer and probehead

Data analysis· what can be trusted?

· model-free or model-based?

· restraints and their error bars

· information beyond distances

Structure

modeling

· label conformation problem

· sparse constraint problem (which approach?)

· extent of coarse graining

· uncertainty of models

Site-directed spin labeling of proteins and RNA


8 Å

bC -SH

bC -S

N OS

S

O

O

H3C

S

O

N

protein

backbone

engineered

cysteine

spin

label

Proteins RNA

O

HN N

S

O

NH

N O·

O. DUSS, M. YULIKOV, G. JESCHKE, F. H.-T. ALLAIN Nature Comm. 5, 3669 (2014)W.L. HUBBELL, C. ALTENBACH, ET AL.

Alternative types of labeling


Cofactor labeling Metal ion substitution

II IIMn substitution for Mgin hnDnaB helicase

TEMPO-labeled cobalamin (vitamin B12)bound to BtuB

ADP

Mg/Mn

T. WIEGAND ET AL. Angew. Chem. Int. Ed. 2017, 56, 3369 –3373B. JOSEPH ET AL. Angew. Chem. Int. Ed. 2015, 54, 6196 –6199

Nitroxide labels


N O

R

Dehydro-Proxyl

N O

R

Proxyl

N OR

TEMPO

N OHOOC

H2N

TOAC

H

N

R

O.

z

y

x

14A( N)

14A( N)

orientation dependenceof g value

The nitroxide spectrum depends on orientation... ...and on polarity of the environment

· Proxyl preferred because of stability

and relative rigidity

· methyl group replacement by ethyl

or spirohexyl groups is advantageous

for relaxation and stability - but tedious

N O.

z

y

x

H

-d+

d

· +· -N-O « N -O

influenceshyperfinecoupling

influencesg tensor

hydrogen bonding

Nitroxide spectra and dynamics


Fast regime Slow tumbling

2A' » 3.0 mTzz

2A' » 6.8 mTzz

tr tr

10 ps 4 ns

32 ps 10 ns

100 ps 32 ns

316 ps 200 ps

1 ns 316 ns

3 ns 1 µs

d

d

I0

I0

· nitroxide spectra are sensitive on the

time scale of sidegroup dynamics

· the actual dynamics is more complex

than isotropic rotational diffusion

· in many cases, semi-quantitative analysis

in terms of spectral parameters A’ or dzz

suffices

X-band CW EPR spectra for isotropic Brownian rotational diffusion

Nitroxide motion - What really happens


ps ns ms ms

-1210

-1410

-1010

-810

-610

-410

-210 s

backbonelibration

bondvibration

librationin exposed

labels

transitions betweenrotameric states

collectivesegmental

backbone motionglobalprotein

tumbling large-scale backbone motion

NH

CC

CO

SS

N

O

c1

c2

c3

c4

c5

x

y

C

N

O

CA

CB

SG

S1

C4

C3

C2

C1 C8

C9

N1C5

C6

C7

O1

131

72

Rotamer ensembles in T4 Lysozyme

jump motion between rotamers

time-resolved CW EPR

relaxation

Nitroxide rotamer libraries


Spin label conformations are (semi-)discrete

Principle of rotamer library prediction of spin label conformations

MD simulation of unrestricted MTSSL spin label side chain

c1 c2 c3 c4 c5MD trace

rotamer library

NH

CC

CO

SS

N

O

c1

c2

c3

c4

c5

x

y

C

N

O

CA

CB

SG

S1

C4

C3

C2

C1 C8

C9

N1C5

C6

C7

O1

MMM 2017.1www.epr.ethz.ch/software.html

3

Matlab-based open-source freeware

· rotamer populations for the unrestricted label

· + non-bonded label-macromolecule interaction from Lennard-Jones potential Þ relative free energies of bound rotamers

· via Boltzmann distribution: ensemble of rotamers with populations and partition function

Boltzmann inversionrelative free energies of unbound rotamers

Paramagnetic quenchers relax nitroxides


N NNO O

OO=O O=OO=O· ··· ··

Diffusing paramagnetic species

before and after collisionduring

electrons in overlappingorbitals are indistinguishable

observer spin up observer spin down ¯

· most easily detected via change in T by progressive power saturation1

· at high concentration, shortening of T leads to line broadening (T £ T )2 2 1

· the environment (macromolecule, lipids) may shield the nitroxide from

such collisions Þ accessibility measurements

A et al., Proc. Natl. Acad. Sci. USA 91, 1667-1671 (1994)LTENBACH

Ni

OHN

HN

O

O

O

OH2

OH2NiEDDA

III IIGd and Cu labels


H RN

ON N

NN

Gd(III)

OO

O

O

OO

[Gd(DOTA)]

N

NN

O OO

OOO O O

Gd(III)

O

HN R

-[Gd(DTPA)]2-[Cu(EDTA)]

û broader EPR spectra, faster relaxation ü III chemically more stable (especially Gd )

û larger size ü spectroscopically orthogonal

û not suitable for assessing dynamics, polarity, and accessibility

Trityl labels


S

SS

S

O OH

SS

SSO

HO

SS

SS O

OR

û hard to synthesize, not commercially available ü chemically more stable than nitroxides

û larger size than nitroxides ü spectroscopically orthogonal

ü very narrow spectral line up to high fieldsû not suitable for assessing dynamics, polarity, and accessibility

Linker chemistry for spin labels


SSH3C

O

O L I

LHN

OLN

O

O

Thiol-specific linkers

Linkers to unnatural amino acids

most selective, short,but labile attachment

selective at pH 6.5¼ 7.5somewhat bulky

may label primary aminesif thiol groups are inaccessible or missing

MTS Maleiimide Iodoacetamide

O

+N

OH2N

O

-H2O

NO

N

O

12-48 h at pH 4, not all proteins like that catalyst may reduce nitroxide label

Ketoxime chemistry Click chemistry

I[Cu ]

NH

O

N3

N O+NH

O

N

N N

NO

Choice of labeling sites and site scan


· well accessible sites with many rotamers and large partition

function are preferable

· helix surface sites are often suitable

Progressive power saturation


0 5 10 15 20 25 300

0.5

1

1.5

2

Microwave power (mW)

Sig

nal i

nte

nsi

ty (

a.u

.) air (20% O )2

nitrogen

PDB# 2BHW

V228

Microwave power P is increased and the amplitude I of the central line measuredmw 0

Example: High oxygen accessibility of a lipid-exposed residue in plant light-harvesting complex LHCII

· amplitude A and homogeneity parameter e are of no concern

· the accessibility parameter P removes line broadening effects (d ) and normalizes0

to power conversion of the given spectrometer/probehead (reference measurement)

· the half-saturation power P quantifies the relaxation enhancement1/2

· protein complex is detergent-solubilized

· sample is contained in a gas-permeable plastic (TPX) capillary

· sample equilibrates with the composition of an external gas stream in less than a minute


|b b ñS I|b a ñS I

|a b ñS I|a a ñS I

saturating

microwave

wI

wI

w2

w0

Overhauser Dynamic Nuclear Polarization (DNP)

áI ñz (0)

áI ñz

e = = 1 - x·f·s·gS

gI

Polarization

enhancement

(0)áI ñ = áI ñ 1 + z z ( )s r

r wt

(0)áS ñ-áS ñz z

(0)áS ñz

(0)áS ñz

(0)áI ñz

{ { {{x f s g /gS I

s = w - w2 0

r = w + 2w + w2 I 0

(0)w = 1/T + rt 1n

Transferring electron polarization to nuclear transitions

DE = hn

pb

pa

· s is maximum if relative

diffusion rate matches

nuclear resonance frequency

Spin Label

Water molecule

Boltzmann distribution

· works at physiological temperature

· no deuteration required

Overhauser DNP water accessibility measurements


bila

yer

bila

yer

(A,B)141

BtuCD BtuCD-F

PDB# 1L7V PDB# 2QI9

(A,B)141

(A,B)141 (extruded liposomes)

mw power [dBm]

P /

P0

-10 -5 0 5 10 15 200.7

0.8

0.9

1

5 mM

BtuCD

BtuCD-F

Opening of an inner gate of an ABC transporter on binding of the substrate-binding protein

Dependence of water accessibility on immersion depth in a lipid bilayer

¾ Overhauser DNP

pp simulation

W2R1

W22R1

~A11R1

· a few mL of sample at

a concentration of

10-100 mM suffice


What is DEER? A reminder

p p p

p

t1

t

t = 0

t1 t2 t2

wA

wB

2

pg

intermolecularbackground(exponential decay)

Numberof spinsin the object

1/Dn3

(~ r )

V(t

)/V

(0)

F(t

)

0 5 100

0.5

1

t (µs)0 2 4 6 8

0.6

0.8

1

t (µs)

Hex

Hex

NO•

N

O

O

N O•

N

O

O(CH2)6OMe

(CH2)6OMe

D

Model compound

Primary data Form factor

PANNIER M, VEIT S, GODT A, JESCHKE G, SPIESS HW, J. Magn. Reson. 2000, 142, 331

MARTIN RE et al., Angew. Chem. Int. Ed. 1998, 37, 2834

Tikhonov

regularizationDistance distribution

JESCHKE G et al. Appl. Magn. Reson. 2006, 30, 473

JESCHKE G et al. J. Magn. Reson. 2002, 155, 72

2 3 4 5 6

0

0.01

0.02

0.03

r (nm)

P(r

)

e

e

The is observed echo amplitudeas a function of time t


Long-range distance distribution restraints by DEER

0 0.5 1 1.5 2

0.7

0.8

0.9

1

t (µs)

0 0.5 1 1.5 2

0.8

0.9

1

t (µs)r (nm)

( ) ( )

( )2

2

2

2

)(

rPdr

d

tDrKP

PG

-

= r + haa

Mean square deviation

Roughness

r =

h =

Minimize

(Tikhonov regularization)

background correction

V(t

)/V

(0)

F(t

)/F

(0)

log r

log h

a = 71.3opt

Primary DEER data

Form factor

L curve

Distance distribution

~20-40 mM Bax 87R1/126R1 in mitochondria-like lipid vesicles (34 GHz, 150 W, 20 mL oversized sample)

Ÿ enforce P(r) ³ 0

DeerAnalysis 2016www.epr.ethz.ch/software/index

-1.4 -1.2 -1 -0.8 -0.6 -0.4-15

-10

-5

0

5

10

15

20

2 3 4 5 6

0

2

4

6

8

10

10-3´

Secondary structure information from spin-labeling site scans


A , d , P and P vary periodically in a site scan through an a-helix or b-sheet zz 0 O NiEDDA2

295 300 305 310 315 320 3253.4

3.45

3.5

3.55

3.6

3.65

Residue number

A [m

T]

zz

+External loop eL4 of Na /proline symporter PutP of E. coli

+· loop model based on homology (Na /glucose symporter vSGLT), secondary structure information, and a few DEER

distance restraints

without

secondary structure restraints

with

Hybrid structure determination


NATHANIEL VARGAS, Glendale

If a single method does not provide sufficient restraints

This may blur the boundary between experimental structure and ab initio model

· combine experimental data from different techniques

! for each restraint subset, be aware of its uncertainties

! be careful about your assumptions: the solution may not be a single conformation

· stabilize the solution by computational chemistry information

Stem loops /SL1 SL3

exp.

conformer L

conformer R

8642distance [nm]

O. DUSS, E. MICHEL, M. YULIKOV, M. SCHUBERT, G. JESCHKE, F. H.-T. ALLAIN Nature 509, 588-592 (2014)

Conformer L Conformer R

Hybrid structure - Types of experimental informationmodeling


· atomic resolution structures of domains or complex components (x-ray, NMR, cryoEM)

in the or in a same state different state

· EPR distance distribution restraints

information on width of an ensemble of conformations or the presence of distinct conformations

· small-angle scattering curves (SAXS, SANS)

low resolution, restrain global shape

· other EPR restraints

(secondary structure, accessibility/bilayer immersion depth)

· other NMR restraints

(secondary structure propensities, pseudo-contact shift information on label distribution)

· cross-linking restraints

only subsets may apply if distinct conformations exist


Hybrid structure - Types of ab initio informationmodeling

· bond length, bond angles

· clash avoidance (repulsive part of non-bonded interaction in a molecular force field)

· Ramachandran-allowed backbone torsion angles

· fragment-library information (Rosetta)

· homology information

· secondary structure prediction

· molecular force fields beyond repulsive part of non-bonded interactions

· ab initio folded structure

Assumptions that one can make... ... and their reliability

The larger the system and the more distributed its conformation, the more critical are assumptions with low reliability

The larger the system and the more distributed its conformation, the more assumptions must be made

Uncertainty and inaccuracy of spin-label based restraints


probability density isosurface

for spin label localization

from five distance constraints

MMM prediction

of spin label

rotamers

reference labeling sites68

89

10986

72

Exercise: GPS-like localization of 131R1 in T4 lysozyme Accuracy test of label-to-label distance predictions (Å)

Rotamer library 30 pairs T4L 62 pairs mixed

MD/Charmm

MC/UFF

MC/UFF CaSd

2.3

2.4

1.7

3.0

3.0

2.6

· approaches by others (mtsslWizard, PRONOX) perform

on a similar level

· MD simulation usually performs slightly worse, after

special parametrization slightly better

JESCHKE G Progr. Nucl. Magn. Reson. Spectr. 2013, 72, 42-60.

ISLAM SM, ROUX B J. Phys. Chem. C 2015, 119, 3901-3911.

The error can be reduced by overdetermination, but EPR distance restraints are usually sparse


Sparse distance restraints & structure modeling

Concept: (Semi)rigid bodies joined by flexible linkers

· 6 degrees of freedom (3 translation/3 rotation)

for each rigid body beyond the first one

· 2(N-1) free torsion angles (f, y) for an N-residue peptide

Beware of cis peptides!

· side groups are predicted by SCWRL4

Dx,Dy,Dz

ÅaÅg

,,

Åb

KRIVOV GG, SHAPOVALOV MV, DUNBRACK RL Proteins 2009, 77, 778-795.

Example: Combining x-ray crystallography, SAXS, and DEER


Second pair of FnIII domains of integrin α6β4

· six rigid-body parameters from 13 DEER restraints

· the two individual domains crystallize,

but the domain-linker-domain construct does not

· two restraints to center of 21-residue linker

· the SAXS shape does nor reveal orientation of the

globular domains

· Monte-Carlo linker modeling based on residue-specific

Ramachandran plots

· CRYSOL for testing models against SAXS curves

· SAXS curves used for detecting structural changes by

spin labeling

The problem

The approach

1.2 1.25 1.3 1.35 1.4 1.452.4

2.5

2.6

2.7

2.8

2.9

cSAXS

sD

EE

R(Å

)

interdomain model 84



linker ensemblewidth dominatedby lack of restraints

N. ALONSO-GARCÍA et al. Acta Cryst. D 2015, 71, 969-985

Restraint-augmented homology modeling


0 10 20 30 40 50 600

1

2

3

4

5

6

Restraint number

Dis

tance

(nm

)

0 10 20 30 40 50 600

1

2

3

4

5

6

Restraint number

Dis

tance

(nm

)

+Modeling of Na /proline symporter PutP based on homology and DEER restraints

+· crystal structure of the Na /glucose symporter vSGLT was known

· only about 20% sequence homology, different number of transmembrane helices

· 68 DEER distance restraints for “helix end” pairs

with Daniel Hilger & Heinrich Jung

Restraint matching for aligned residuepairs in vSGLT

Restraint matching of the DEER-augmentedhomology model

Large-scale conformational change by elastic network models


Residue-level elastic network model (ENM) Hinge motion of chaperonin GroELwith simulated DEER data

Ca

experimental distance longer

experimental distance shorter

· force constants of the springs depend on Ca-Ca distance

· type of motion recognized, but model does not have

atomic resolution

· may not work as well as for other types of motion

· network is deformed along its normal modes

by forces that are proportional to

the mismatch of distance restraints

ZHENG W, BROOKS BR Biophys. J. 2006, 90, 4327-36.

· label-label distances can be used as well

JESCHKE G J. Chem. Theor. Comp. 2012, 8, 3854-63.

x-ray (1AON/1OEL) ENM with 20 restraints

Rigid-body docking


f

q

180°

Dx

Dy

C2

X'

Y'

Z'X

Y

Z

A

B

+ +Dimer structure of Na /H antiporter NhaA What you assume to be rigid, may move

· 9 distance restraints determine 4 free parameters

· full grid search in parameter space

· protomer structure assumed to be rigid (PDB# 1ZCD)

D. HILGER, YE. POLYHACH, E. PADAN, H. JUNG, G. JESCHKE, Biophys. J. 2007, 93, 3675-3683.

side view

crystal packing effect in structure 1ZCD

our model

new cryo-EMstructure

M. APPEL, D. HIZLAN, K. R. VINOTHKUMAR, C. ZIEGLER, W. KÜHLBRANDT, J. Mol. Biol. 2009, 386, 351-365.

Modeling of intrinsically disordered domains


Ÿ bond lengths and bond angles

Ÿ 2(N-1) torsion angles fi,yi for a peptide

with N residues

Ÿ without constraints sampling of solution

space is unfeasible for N > 15… 20

Ÿ even with constraints loop closure

between two anchor residues requires

steering the loop to the second anchor

Ÿ preferences for backbone dihedral angles

Ÿ side chain rotamer preferences as encoded

in SCWRL4

Ÿ distance distribution restraints

Ÿ secondary structure propensities

(NMR chemical shifts or periodicity

of EPR parameters or )

Ÿ lipid bilayer immersion depths (membrane proteins)

Reliable information

Free parameters

180° unless cis peptide

probability distributionknown from residue-specific Ramachandran plots}

Tryptophan

S. HOVMÖLLER, T. ZHOU, T. OHLSON

Acta Cryst. D 2002, 58, 768-776

f

y

180

-180-180 180

20 40 60 80 100%

cumulative probability

Intrinsically disordered domains: Uncertainty versus flexibility


p27KID: from EnsembleNMR-restrained MD and its central structure

p27KID: recoveredEnsemblefrom 56 simulated DEER restraintsand 21 secondary structure restraints

p27Kip1: in complexCrystal structurewith Cdk2 and ensemble obtained from56 DEER and 21 secondary structure restraints

Ÿ uncertainty about spin label

conformation and lack of restraints

translate into larger ensemble width

Ÿ global shape is reproduced, but at

low resolution

Þ the width of EPR-derived ensembles

is an upper bound on the conformation

space that is actually sampled

G. JESCHKE, Proteins 2016, 84, 544-560

Documents

epr @eth Molecular and Structural Biology V: Studying ...€¦ · We need an unpaired electron 2 e Three basic types of systems HN O CH2 O Radical enzymes Metalloproteins Spin labels