Transcript Time Domain EPR: Membrane-binding Proteins Using R1 from EPR as a probe of the Structure-function and the Dynamicsfunction relation in biology Graduate Students: Post Docs: Faculty: Tamara.

Time Domain EPR:
Membrane-binding Proteins
Using R1 from EPR as a probe of the
Structure-function and the Dynamicsfunction relation in biology
Graduate Students:
Post Docs:
Faculty:
Tamara Okonogi
Andy Ball
Michael Gelb
Robert Nielsen
Ying Lin
Kate Pratt
Stephane Canaan
Kepeng Che
Supported by NSF and NIH
Time Domain EPR: Outline
• Time Domain: Saturation Recovery and
Pulsed Electron Double Resonance Methods
– Comparison with CW methods
– Spectrometer, experiment, data
• Theory of relaxation Rates
• Application to the Spin Relaxant Method
• Using site directed mutagenesis
– Orienting a Membrane-binding Protein
– Determine an Oxygen gradient in a membrane
CW Power Saturation
The method to obtain spin-lattice relaxation rates using CW
methods. Plot the Peak to Peak height as a function of
microwave power (or really amplitude).
Details of CW Power Saturation
Y  C
Peak-to-Peak height
1
3
 
2
2
h1

2 

h1
 1 

P
2 

P2 = R1 R 2
A product of spin-spin and spin-lattice relaxation rates.
CW
3
O
C NH2
O
C
C
O
O
15
N
O
O
C O
C
O
O
Cr
O
O
R 2  R 2+Crox  R -Crox
  CROX 
2
Field (Gauss)
O
C
O
C
O
TD ESR Spectrometer
Pulsed Bridge
Phase
control
Amplifier/Digitizer
Balance
control
Sig.1
Sig.2
0
90
A
B
C
D
0
0
90
90
F
G
observe osc.
E
pump osc.
IF
0
90
0
90
H
RF
LO
RF
LO
IF
Free Induction Decay
Y'
X'
Z'
Y'
X'
Z'
Measures spin-spin
dephasing
Pulsed Saturation Recovery
Y'
X'
Z'
Y'
X'
Z'
Measures
relaxation to
equilibrium
Pulsed Electron-Electron Double Resonance
pSR; the effect of a relaxant
R1  R1+Crox  R1-Crox   CROX 
OH
Collision with Oxygen
14
N
O
Redfield Theory (or BWRT)
Relaxation rate theory began with Bloch and Wangness,
and was amplified by Alfred Redfield to be a complete
theory for the effects of dynamics and stochastic processes
on spins in condensed matter. Relaxation theory is often
called BWRT (Bloch Wangness and Redfield Theory).
BWRT Predicts Spin-Lattice (R1) and SpinSpin (R2) relaxation rates.
Relaxation Rates are related to Relaxation times:
R1  1 and R2  1
T1
T2
What Redfield Theory (BWRT) Uses
• There is a system Hamiltonian H s
• BWRT requires a bilinear operator, which couples
the spin system (S) to the lattice (or bath), F
– The Hamiltonian is: H  S  F
• F is the fluctuating variable that causes the spins
to have a fluctuating environment
• The fluctuation of the lattice, coupled to the spin
system, then causes the spins to relax or dissipate
the absorbed (microwave or r.f.) energy, nonradiatively.
A problem: R2 diverges
Nielsen, R. D. and Robinson, B. H. "A Novel Relaxation Equation of Motion". J. Physical
Chemistry 2004; 108: 1589-1600.
The coupling Hamiltonian (e.n.d.) is:
H  S z   I z  F  where F  a 
1
2
1  3cos  
2
The orientation variable,  , is a stochastic function of time.
The correlation function (at high temperature) is:

1
4
1  3cos   t   1  3cos   0 
2
2
t
1 c
 e
5
This shows the statistical origin of the rotational correlation time.
Exponential decay of the correlation function with time is typical
of such functions.
CW Spectra
CTPO
O
C NH2
14
N
O
D.A.Haas, C. Mailer, and B.H. Robinson, Biophysical J. (1993) 64, 594
R2 from BWRT,
diverging
Relaxation Rates (MHz)
R2 rates from
Kubo Theory
R2 rates from
modified BWRT
log  c 
R1 does not diverge
• R1 for the electron and R1 for the nitrogen nucleus
in a nitroxide spin label as a function of rotational
correlation times can be computed from BWRT.
• If R2 diverges for correlation times longer than a
few nanoseconds how can we rely on the theory to
give us R1 values out to milliseconds and beyond?
• The Problem: Why does the theory fail for R2
rates but not for R1 rates?
• It is important to understand why R1 works and to
understand why R2 fails.
Nitroxide
Nitrogen
Spin Lattice
Relaxation
Rates
With O2
and Without O2
Relaxation Rates (sec-1)
Electron spin-lattice
relaxation rates:
Correlation Time (sec)
B.H. Robinson, D. A. Haas and C. Mailer (1994). Science 263(5146): 490-3.
Mechanisms of R1
• Sum the rates from statistically independent
processes
1

– Spin-Rotation, rate goes as c
– Electron-Nuclear Dipolar Coupling
• Electron rate peaks at the spectrometer frequency
• Nuclear rate Peaks at coupling a 2
– Oxygen relaxation (used later)
– Empirical “Spin-Diffusion” process

1
8
• Just a catch-all effect, goes at 
c
• Partially due to spins local to the nitroxide
The ideal form of the solution is:
The actual form of the solution
is a bit more complicated:
Sx  Sx
Sx  Sx
 R2 t

e
t 0
t 0

 ae
R2 t
 ae
R2 t

2


The two rates are: R2  1   1   fo2  1 1  1   2 c fo 2 

2 c
2 c 
 2 c 
The slower rate dominates.
0.1
f0  5 104 sec1
Solution (Signal)
2 c f 0  1
Black: two rate solution

Blue: eal  R2  rate
Green: BWRT rate
10
Time (sec)
Dominant Rate in all limits
In the fast motion limit:
R2   c 
2
fo
In the slow motion limit:
1
R2 
c
 2 
BWRT gives only the fast motion limit, which predicts that the rate
goes to infinity as the correlation time goes to infinity. The new
theory avoids this and correctly predicts coherent oscillations of Sx
(at frequency fo) as the interaction becomes coherent, in the no
motion limit.
The rates in the two limits may be “combined” into a rate that
does cover both motional regimes (for both R1 and R2):
R2
c
f
2
1  2  c f o 
2
o
R1
 c  fo2
2
2
1   c   s    c  fo 
New Term in both R2 and R1 rates
Spin Labeled-Fatty Acids in DOPC
Spin-Lattice Relaxation rates for varied
Doxyl-Steric Acids in DOPC.
• Different spectrometer
frequencies (from 2 to
35 GHz) with the best
possible single
effective correlation
time.
• A poor fit. The
frequency dependence
of simple isotropic
rotational motion is
incomplete.
Data from Jin and Hyde
SL at 5 position
SL at 12 position
SL at 16 position
Same Data Different Model
• Improving the model to
include anisotropic
dynamics.
• For simplicity the
anisotropy ratio was
kept constant.
• Improved agreement
indicates the need to
improve the model, and
the frequency
dependence of the
relaxation rates can ruleout some incorrect
models.
Relaxation rates from 60 different experiments
Correlation among all
the data and the
model. Model has 1
adjustable parameter
(the mean rotational
correlation time) for
each sample at all 5
different frequencies
and two different
isotopic forms.
Membrane Binding Proteins
Bee venom phospholipase
Oriented on a membrane
surface by
Site Directed Mutagenesis
EPR spin relaxant method
Lin, Y., Nielsen, R., Murray, D.,
Hubbell, W. L., Mailer, C., Robinson, B. H.
and Gelb, M. H. Science 1998; 279 (5358):
1925-9
Labeling a protein (PLA2) with a Spin Probe
Use site directed mutagenesis techniques to
prepare proteins with a single properly placed
cytsteine.
General Reaction for adding relaxants
PLA2
C
O
PLA2
+
O
H3C S
SH
S
C
S
S
CH2
CH2
N
N
.O
.O
The protein should contain only one cysteine for labeling.
Protein labeled at only one site at a time per experiment.
Relaxant Method:
Nitroxide Spectra depend on concentration
of relaxants
R1  R     Rlxnt 
Spin-Lattice: T1-1 or R1
R2  R     Rlxnt 
Spin-Spin:
o
1
o
2
T2-1 or R2
Rates are increased by the same amount due to
additional relaxing agents (relaxants).
P2  R1 R2  R     Rlxnt R     Rlxnt 
o
1
R  R
o
2
o
1
o
2
R     Rlxnt 
o
2
P2  P2  P  R  R
0
2
o
1
o
2
    Rlxnt 
Human (HGIIA)
Secretory Phospholipase sPLA2
A highly charged (+20 residues) lipase, 14kDa
protein
And a highly charged (-70 mV) membrane
All exposure data was determined by SR and pELDOR directly
measuring spin-lattice relaxation rates.
O
C NH2
14
N
O
CW Spectra of hGIIA on Micelles
CTPO
hGIIA
CW Spectrum of site N70C with CROX
Probing the hGIIA protein surface potential using CW and
TD EPR
 [CROX ]  o [CROX ]









 zcroxo e
kBT
e

ro 8 [ NaCl ] 
 kBT






Spin Lattice Relaxation Rates for sl-sPLA2
rates from pSR and
pELDOR for CTPO
Power Saturation Curves site S120C
O2 Relaxant: hGIIA on LUV
TD
CW
Compare O2 Relaxant Effects
from TD-SR and CW
Summary of Vesicle data
• Large protein surface charge determined by
CW and TD data
• Complete protection from Crox for all EPR
data
• Oxygen effect reduced relative to solution
• Light scattering occurs
Aggregation model
~50 enzymes (36 Angstrom diameter)
LUV of DOPM (100 nm diameter)
TD data, Vesicles vs. Mixed Micelles
Vesicle
(DTPM)
Mixed
Micelles
hGIIA-sPLA2 on mixed micelles
Crox
3
O
C
C
O
O
O
C O
C
O
O
O
Cr
O
C
O
C
O
O
NiEDDA
O
C
O
Ni
O
2+
C
O
NH
CH2
CH2
NH
sPLA2 on Membrane
View from
membrane
Yellow:
Hydrophobic
Residues
Blue: Charged
(pos) residues
Orientation perpendicular to that
predicted by M. Jain.
Anchored by hydrophobic
residues. Charges not essential
sPLA2 Conclusions
• sPLA2 causes the vesicles to aggregate.
Explains much other data and misconceptions about
the kinetics and processive nature of sPLA2 action.
• sPLA2 was oriented on micelles (instead) using spinlattice relaxation rates alone.
Orientation different from that of another model.
• Hydrophobic residues are the main points of contact.
• Charges provide a general, non-specific attraction.
•Substrate binding site identified by orientation on the
mixed micelles
The WALP Proteins
WALP proteins are single alpha helical
membrane-spanning proteins.
The sequence is 23 residues long:
HCO-NH-G-WW-L-(AL)8-WW-A-CO-NH2
Leucine and Alanine are both hydrophobic.
In a membrane this forms a single turn alpha
helix.
The membrane, di-oleic (DO) PC, is about
28-30 Ang thick. The two outer Tryptophans
(W) are about 30 Angs apart. The membrane
will stretch (or shrink) to accommodate the
protein.
Demmers et al: J. Biol. Chem., 276, 34501-34508, 2001
WALP23
The sequence is 23 residues long: HCO-NH-G-WW-L-(AL)8-WW-A-CO-NH2
Subcyznski et al. Biochemistry,
2003, 42, 3939
WALP23-sl CW spectra
CW EPR spectra of spin labeled WALP23 at various
positions.
Oxygen Transport Parameter
The Oxygen transport parameter is the change
in the spin-lattice relaxation rate due to oxygen
collision-relaxation
R O


1  2 
 O




2  ,

Depends on transport
properties (e.g. Diffusion)
of Oxygen in the local
environment of the spin
label
where
R1  O2   R1 O2   R1 O2   1  O2   1  O2 


T1 
T1 






Typical WALP/DOPC Saturation
Recovery EPR
With Oxygen
CW
Without Oxygen
Walp23 in DOPM: Oxygen Transport Parameter
R2e
R1e 
1
T1e
From SR
Estimated
from the
CW line
width
Walp23 in DOPC: Oxygen Transport Parameter
R1e 
1
T1e
From SR
R2e Estimated from the CW line width
Ratio Parameter*

 R1
  ln 
 R1

*Altenbach, C. et al. PNAS (1994)

O2  

Ni
  
91 (5), 1667-71.
Conclusions
• The gradient of the oxygen transport parameter, measured on
WALP 23, is ideal as a ruler for determination of spin label
position in membranes.
• The spin-lattice and spin-spin relaxation rates show dependence
on local mobility of the spin label in the bi-layer.
• The oxygen transport parameter cannot be separated into its two
components: the oxygen concentration and transport-dependent
coefficient.
• The ratio parameter, designed to cancel out transport effects,
provides a profile of relative relaxant concentration.
• Ratio parameter can be used to position nitroxide in the
membrane.