31P NMR in lipid membranes.CSA recoupling.

Ludovic BERTHELOT, Dror E. WARSCHAWSKI& Philippe F. DEVAUX 1

1 Laboratoire de physico-chimie moléculaire des membranes biologiques UPR 9052

IntroductionIntroduction

31P NMR experiments have been carried out withliposomes containing lipid mixtures or red blood cellmembranes. We used MAS with a rotating speed of5kHz, and recoupling of the CSA by rotationsynchronized π-pulses. We have been able toseparate the lipids on a 2D-spectrum according totheir polar headgroup. The goal is to attribute thephase of each lipid by comparing the cross-sections ofthe spectrum with a static spectrum.

Alpine conference on solid state NMR,Chamonix, 12-16 September 1999

Lipid Phases and NMRLipid Phases and NMR

• Structure of lipids investigated

• Lipid polymorphism in water : [1]

lamellarphase

invertedhexagonalphase (HII)

fluid Lα

gel Lβ

31P : natural abundance 100%I=1/2

O

O

O

O

O P

O

O

ON+_

O

O

O

O

O P

O

O

ONH3

_ +

O P

O

O

ONNH

OH

O

+_

PCPC

SMSM

PEPE

(phosphatidylcholin)

(phosphatidylethanolamine)

(sphingomyelin)

Static 31P NMR of phospholipids in waterStatic 31P NMR of phospholipids in water

ωω ωωββ

CS iso CSA== ++−−

∆ω∆ω3 1

2

2cos

for each phospholipid, theprecession frequency ωCS

depends on the orientation of thepolar headgroup with the field [2].

∆ω∆ωCSA chemical shift anisotropy

31P is in an anisotropic environment and undergoesrapid anisotropic motions.

The effective chemical shift tensor has acylindrical symmetry

$σσ

$σσσσ

σσσσ

==

⊥⊥

⊥⊥

0 0

0 0

0 0 //

ωω iso isotropic chemical shift

e.g. : static spectra ofDMPC at 10°C and 30°C

(gel-fluid transition : 23°C).

-40-30-20-1090 80 70 60 50 40 30 20 10 0 ppm

10°C

30°C

The result of the integration withrespect to β is characteristic of the lipid

phase [2] :

gel

fluid

inverted hexagonal

Micelles ororganic solvent

0 -100100

frequency (ppm)

Magic Angle SpinningMagic Angle Spinning

B0

ωωr/2ππ = 5kHz

ααm ααm = 54°44’

CSA is averaged outto zero

ωω ωω ωω ωωωω ωω

CS iso r r

r r

t C t C t

S t S t

( ) cos( ) cos( )

sin( ) sin( )

== ++ ++++ ++

1 2

1 2

2

2

for a rotation period of 200µs

ωω ωωCS iso==

Thus one obtains a high resolution spectrum, withone line for each headgroup.

The averaging is macroscopic instead ofmicroscopic.

2D recoupling2D recoupling

30°C

60°C

46°C

37°C

PCPE

recoupled indirect

dimension

decoupleddirect

dimension(MAS)

Correlation of a decoupleddimension (high

resolution, MAS) with arecoupled dimension [3] .

ωω ωω ωω1 1 2∝∝ == ++ ++CSstat

iso C C

We need :

so that the indirectdimension corresponds to a

static spectrum.

e.g. : static spectra of aPC/PE mixture at different

temperatures

A method derived from Tycko et al., 1989 [4]A method derived from Tycko et al., 1989 [4]

• Rotation synchronized ππ-pulses

(4 π-pulses)

We want :1

0 1 2tp t t dt C C

r

t

CSCSstat

iso

r

( ) ( ) ( )ωω χχ ωω χχ ωω∫∫ == == ++ ++

•p(t) is even

• p t t dt p t t dtt

r

t

r

trr r

( ) cos( ) ( ) cos( )ωω ωωχχ

0 0

12

12 2

2∫∫ ∫∫== ==

In fact : ωω ξξ ωω χχ ωωCS iso CSstatt( ) == ++

ξ ξ isotropic scaling factor

χχ anisotropic scaling factor

• The pulse program

+ 1H decoupling during the acquisition

π/2 π/2 π πππ

for τ1=39µs and τ2=89µs at 5000Hz,ξ = 0ξ = 0

χ = 0.393χ = 0.393

• no decoupling during the evolution time.

• no need for cross-polarization (31P is naturally

abundant).

• the direct dimension is proportional to a static

spectrum : the elements of the principal axis

tensor may be extracted directly.

• we just need a simple MAS probe to run our

experiments (no switch angle or speed).

Result #1DOPC/DOPE/cholesterol [5]

Result #1DOPC/DOPE/cholesterol [5]

vertical cross-

sections :

ωωr/2ππ=5000Hz ; χχ=0.393

37°C

PE PC

30°C

60°C60°C

46°C37°C

46°C

37°C

30°C

horizontal cross-section

of 2D

Result #2Ghosts of red blood cells

Result #2Ghosts of red blood cells

Hz

-2.5-2.0-1.5-1.0-0.52.0 1.5 1.0 0.5 0.0 ppm

-2000

-1500

-1000

-500

2000

1500

1000

500

0

PCPESM+

-12-10-8-6-4-214 12 10 8 6 4 2 0 ppm

PE SM+(0,60 ppm)

PC(0,00 ppm)

ωωr/2ππ=5000Hz ;χχ=0.393 ; 30°C

PCPS?

PE SM+

-2-12 1 0 ppm

MAS Hz5000

horizontal cross-sectionof 2D

vertical cross-sections : cholesterol 25

total phospholipids 56 in which PC 23

PE 20PS 11PI 2

SM 18others 1

average composition of lipids in humanerythrocytes (%)

ConclusionsConclusions

• a good signal/noise ratio even for biologicalsamples (experiment time : 3h30 ; ghosts :overnight).

• a good resolution in the direct dimension : lipidsare separated according to their polar headgroup.

• a quantitative narrowing of the recoupledspectrum with the temperature, corresponding tothe lamellar-hexagonal transition.

• recoupled spectra cannot besuperimposed on canonical staticspectra need to decouple duringmixing time, or to compare with staticnon decoupled spectra

-40-20100 80 60 40 20 0 ppm

static DMPC

spectrum

1H decoupled

non decoupled

• recoupled spectra for our ghosts do not exhibit aproper lineshape technical deficiencies (MASstability, amplifier power) ; probably also finitepulse length[6] and “ring down” effects.

[1] Cevc G. “ Phospholipids handbook “, 1993, Marcel Dekker,inc.

[2] Seelig J. “ 31P nuclear magnetic resonance and the headgroup structure of phospholipids in membrane ”, Biochim.Biophys. Acta, 1978, 515 : 105-140.

[3] Schmidt Rohr K. & Spiess H.W. “ Multidimensional solid-state NMR and polymers ”, 1994, Academic Press.

[4] Tycko R., Dabbagh G. & Mirau P.A. “ Determination ofchemical-shift-anisotropy lineshapes in a two dimensionalmagic-angle-spinning NMR experiment ”, J. Magn. Reson.,1989, 85 : 265-274.

[5] Moran L. & Janes N. “Tracking phospholipid populations inpolymorphism by sideband analyses of 31P magic anglespinning NMR”, Biophys. J., 1998, 75 : 867-879.

[6] Ishii Y. & Terao T. “Manipulation of nuclear spin hamiltoniansby rf-field modulations and its applications to observation ofpowder patterns under magic angle spinning”, J. Chem. Phys.,1998, 109 : 1366-1374.

ReferencesReferences

AcknowledgmentsAcknowledgments

This work was supported by CNRS. The authors wish to thank Pr.Geoffrey Bodenhausen and his whole team for fruitful discussions.