Reprint of “Improved spectral resolution in COSY 1H NMR spectra of proteins via double quantum filtering”

Vol. 117, No. 2, 1983 December 16, 1983

IMPROVED

BIOCHEMICAL

RESEARCH COMMUNICATIONS Pages

SPECTRAL

M.

AND BIOPHYSICAL

+*

Rance

IN COSY 1H NMR QUANTIJM FILTERING

RESOLUTION DOUBLE

SPECTRA

OF

G. Bodenhausen+, s!.?wensen+, + R.R. Ernst and K. Wiithrich*

G.

, O.M.

PROTEINS

479-485

VIA

Waqner*,

* Institut fiir Molekularbioloqie und Biophysik, Eidgenassische Technische Hochschule, CH-8093 Ziirich, Switzerland t

Received

Laboratorium Eidqenossische CH-8092

fiir

Physikalische Chemie, Technische Hochschule, Ziirich, Switzerland

4, 1983

October

SUMMARY: A double quantum filter is inserted into a two-dimensional correlated (COSY) 1~ NMR experiment to obtain phase-sensitive spectra in which both cross peak and diagonal peak multiplets have anti-phase fine structure, and in which the cross peaks and the major contribution to the diagonal peaks have absorption lineshapes in both dimensions. The elimination of the dispersive character of the diaqonal peaks in phase-sensitive, double quantum-filtered COSY spectra allows identification of cross peaks lyinq immediately adjacent to the diaqonal, which represents a siqnificant improvement over the conventional COSY experiment.

Two-dimensional

correlated

employed

as

an

networks

of

the

early

showed

resolution resolved

further

use

of

are

COSY peaks

the

obtained

and

couplinq

of

usually

experiment and

constants

the

diagonal

peaks

of a protein peaks

to

over

to have

obtain the

example,

479

recent

with

the

resonance

be

A limitation

in

experiments

arises

made. COSY

cross

analysis. quantum

lineshapes.

more diqital

and

peaks

The

which

in

the

near

paper (DQF)

spectra entire

spectra

the

present

filter

The

quantitative

(1).

regions

phase-sensitive same

the

reqions,

spectral

a double

coupling

data, high

lineshapes the

been

While

unambiquous

different

spectral

applyinq

For

conventional

important

amenable

usefulness

have

diagonal

a consequence, not

of

peaks

can

the

with

peaks

spectral

of

spin-spin

of

(7).

time

(4-6).

recorded

crowded

some

the

presentations

cross

for

proteins

advantaqes

in

diagonal

of in

spectra

presentations

As

the

value

even

has

delineation

individual

spin-spin

enhancement

demonstrates

cross

peaks

macromolecules.

diagonal

the

the

of

(l-3)

residues

COSY

phase-sensitive

cross

cause

of

acid

important

structure be

the

absolute

provide

can

(COSY)

for

amino used

fine

because

of

individual

phase-sensitive

measurements

can

method

that

assiqnments

the

efficient

investigations

work

spectroscopy

in

(8,9)

in

which spectrum

0006-291X/83 $1.50 Copyright 0 1983 by Academic Press, Inc. All rights of reproduction in any form reserved.

Vol. 117, No. 2, 1983

thus

becomes

and

FUNDAMENTAL

CONSIDERATIONS:

the

The

components

respect

to

are

no

absorption

the

of

the

to

anti-phase

presence

the in

alternating

vicinity

signs,

the

incompletely diagonal

peak

spectra

typically

experiment

to

into the

convert

it

is

magnetization, else

on

a cross character absorption however, filter

order

other

peak.

back spin

Both

and,

some in

on

furthermore, As

dispersive systems

is the

involved

diaqonal

lineshapes.

spin

which

in

is then

and

cross

can

be

phased

to

multiple contributions

than

have

two

spins.

480

the

cycling

the

second

reconverted rise

cancel the

to

then

to

the

Fortunately

when

protein

COSY

a third

90°

COSY second of

pulse the

pulses

pathway quantum pulse

filter, into

anti-phase

a diagonal

peak,

or

rise

to

giving

have

simultaneously quantum

peaks

in-phase

to

coherence,

multiplets

The

have

in

a double

quantum

in-phase

cross

a coherence

givinq

peak

of

they

by

phase

double

lineshapes

periods.

to

the

(4-6).

created

by

2D then

a conventional

converted

spin,

the

pure

involves

detect

conventional

more

(8,9)

Using

intensity

identify

peaks peaks

immediately

same in

anti-phase with

cross

(11,12).

of a directly

2D

for

diaqonal

selectively

the

the for

tend

of

all

signs

and

since

appropriate

coherence one

to

pulse

the

variation

detection

coherence By

to

of of

either the

quantum

possible

coherence,

90°

transfer

different

and

technique

second

coherence.

The

occurs

the

from

dimensions),

difficult

the

to

filter

the

magnetization quantum

relative

multiple

a particular

anti-phase

that

quantum

receiver

through

so

the

phased

components

components,

quantum

are

Furthermore,

cancellation

following

single

it

no

have

the

mixing

that

SO

of

sinusoidal

evolution

multiplet

of

and

peak,

frequency

the

the

alternating

cross

mode.

diaqonal.

by

as

magnetization

both

makes

phase

t2

arise

pulse,

exhibit

contrast,

enhanced

mixinq

peaks

from the

peak

In

multiple

immediately

double

cross

the

generated

cross

result

the

component

the

in

tails

of

are

generates

2D dispersion

peaks

multiplet

The

back

pure

that

are

following

to

from

same

spectra.

period,

experiment

anti-phase

the

evolution

this

the

resonance

DQF-COSY the

referred

of

therefore

during

resolved.

to

When

dispersion

the

have

lineshapes

cross

long

by

the uses

the

usually use

peaks

spin

which

in

the

multiplets

magnetization

lie

pulse

peak

be

and

of

cross

one

absorption

diagonal

which

peak

The

is

on

from

experiment is

result

a diaqonal

binomial.

will

in peaks

coupling

longer

peaks

pulse

unaffected

of cross

(i.e.

diagonal

the

of

the

90°

is

ratios.

spin;

ratios

tl

diagonal

magnetization

coupled

The

interval

new

extracted

COSY

which

intensity

anti-phase

and

The

be

A conventional

limitations

(10).

and

can

second

forementioned

multiplet

with

information

the

magnetization

binomial

analysis

detailed conformation

and

following

in-phase

for

90°-tl-90°-t2.

period,

pulse.

RESEARCH COMMUNICATIONS

molecular

sequence:

detection

AND BIOPHYSICAL

accessible

assignments

pulse

BIOCHEMICAL

anti-phase pure

spectroscopy diagonal these

2D (13,14)

peaks

pass

contributions

Vol. 117, No. 2, 1983

are of

weak

and,

their

judged

an

can work

practical

use

experiments

small

of

DISCUSSION:

for

COSY

of

the

cross

lineshapes

of

the

impossible

diagonal to

observe

spectrum,

spectrum of

to

that

is

It in

the peaks

is

of

the

and

Fig.

The same

the

obvious

that

conventional

filterinq

followinq

the

with

in

hiqh

the

field

portion

1B presents

lowest

a

contour

fraction the

COSY lie

the

illustrated

be

which

and

diagonal.

quantum

sample. to

not

(BPTI).

2H20,

same

do peaks,

the

In

plot

in

chosen

cross

(8,9).

Because

lines cross

double

inhibitor

BPTI

intensities. peaks

close

proteins

the

was

broad

very

a contour

of

tolerable.

the

trypsin

1A shows

are

to

lines of

RESEARCH COMMUNICATIONS

relative

showed

sinqlet

studies

spectra

peak

DOF-COSY

also

spectrum

each

corresponding

almost

Fig.

below,

located

pancreatic

quantum-filtered

plotted

when

of for

presented

peaks

molecules

basic

a phase;sensitive

double

even

DQF-COSY the

the

diaqonal

suppression

on AND

the

AND BIOPHYSICAL

results in

analyzed

with

effective

RESULTS

of

be

provides

the

character

enhancement

latter

Earlier

of

from

anti-phase

cause the

BIOCHEMICAL

pure

level

of 2D dispersion

spectrum

make

vicinity

of

it the

1

1

Wl (pm)

Wl (wm)

3

3

5 i w2

Fig.

1

n

5

i

5

(PPm)

3

1

WZ (wm)

Comparison of the spectral region from 0.5 to 5.0 ppm of (A) a conventional, phase-sensitive COSY spectrum of basic pancreatic trypsin inhibitor (BPTI) with (B) the corresponding region in a phasesensitive, double quantum-filtered (DQF) COSY spectrum. The data were recorded at 360 MHz in a 2OmM BPTI solution, in 2H20, p2H 4.6, T=36%. The same filter functions were applied to both data sets, i.e. a phase-shifted sine bell window with phase shifts of n/8 along t2 and n/4 along tI. The data are presented as contour plots, where both positive and negative levels have been plotted. The COSY spectrum was recorded with 738 tI values, from 4 ns to 117 ms, and the DQF-COSY spectrum with 672 tl values, from 4 us to 106 ms. In both spectra 64 transients were averaged for each tl value. Note that the singlet resonances of Met 52 at 2.13 ppm and the solvent at 4.65 ppm were suppressed by the double quantum filter, althouqh some of the tl noise of the solvent resonance remains.

481

Vol. 117, No. 2, 1983

diagonal.

This

where

the

problem

tails

of

In containing The

to

the

of

using

Fourier

tails

lonq

of

the

introduced.

Althouqh

dispersion

by

of

aromatic

and

2B.

were

The

made

shown

basis

of

7.06

conventional

assignments

of

time

new and

as

and

qlycine many

peak

multiplets

protons the

Phe

the

distinction

in

both

as

tyrosine

cross

type can for

22 4-5

ratio frequency

6--~

couplings,

and

S-n

that serine

a-,5,

illustrated

by

positive

4 3,5-2,6 3,5

characteristic dimensions

a few

for

close

to

the

rather

than

the

and

aromatic

E-C

of

the

expected. both

peak 22

482

fine from

the

intensities

Phe

the

unambiguously

couplings

diagonal

methylene

indicate

for

a cross the

on

for &-E

examples

as

neqative for

spin spectrum

COSY

peaks

within

coupling

triplets

and

BPTI

aromatic

be

ppm

examples

from

protons,

indicate

2A

are groups,

chains.

be

the

the

4 can

Other

couplinqs

available

Phe

of

changed

7.31

cross

are

side

Figs.

quantum-filtered of

couplings.

long

of

phenylalanine

information

for

double

normally

a

(17).

distinction

of

couplinq of

the

in

DQF-COSY

be

at

experiments

the

one-dimensional

Phe

is

2C is

spectrum

all

of

been

represents

The

must

peaks

prior

dispersion

Fiq.

shown NMR

almost

resonance

2C,

the

within

the

a triplet

C4H

Fig.

peaks a-6,

cross

earlier

in

E-G

couplinqs

peaks

and

the

in useful

threonine

The cross

that the

the

assignment

2C the

illustrated be

earlier

Fig.

from

to

other

COSY

In

tryptophan

rise c(-u,

an

follows

such

give

intensity

data.

should

couplings,

and

instance

the

the

clearly

using

of

shows

have

in

(16,17).

assiqnment

28

completely

spectra

effort, simulations

the

the

peaks

Presented

in

by

troublesome

result

protons

with'considerable spectral

while

necessitates

COSY

BPTI,

Methods)

eliminate

this

mode.

scale.

and

cross

to

this

aromatic

immediate

concluded

chains,

which

one

the the

and

it

As

side

ago,

2C allows

the

experiment

value

over

in

ppm,

absolute

the

(15),

expanded

Fig.

Materials

principle

of

difficult

suppressed, for

filter

an

data;

1 ppm

spectrum

peaks.

COSY

largely

in

made

around

peaks.

on

diagonal (see

improvement

and

identified

the

spectrum;

Fig.

protons,

is

DQF-COSY

techniques in

in

the

spectrum

lineshapes

possible

data

of

some

decoupling

is

plotted

of

been

a pseudo-echo

the

portion

significant

it

using

presentation

have

region cross COSY

is

conventional

2D absorption

spectral with

functions

the

peaks

the

the

window

RESEARCH COMMUNICATIONS

phase-sensitive

tails

of

diagonal of

the

of

the

interfere

resonances,

dispersion

transformation

distortions

of

part

optimized

in

peaks

proton this

AND BIOPHYSICAL

pronounced

methyl

aromatic of

of

most

2A a region

interpretation

results

is strong

Fig.

the

presence

BIOCHEMICAL

4-5

coupling.

Fiq.

of

2C.

The

a doublet

for

The

peaks

4 and is

triplet

structure

made, is

cross

cross the

5 protons. the clearly

the

2,6 for When

+1:0:-l observed

Vol. 117, No. 2, 1983

BIOCHEMICAL

AND BIOPHYSICAL

RESEARCH COMMUNICATIONS

6.6

0, (PPml

w, (Pm

24

z-4

Z8

28 7.8

WzlPPm)

7.4

7.0 W,(wml

Phe4

*, (wml 7.4

1

, 7.8

1 7.4

7.0

6.6

W2(wmi Fiq.

2

Comparison of the aromatic reqion in (A,B) conventional, phasesensitive COSY spectra of BPTI and (C) a DQF-COSY spectrum. Same data sets as in Fiq. 1. Positive and negative levels are shown in the contour plots. The COSY spectrum A and the DQF-COSY spectrum were processed with identical filter functions, i.e. phase-shifted sine bell windows with phase shifts of n/8 along t2 and n/4 alonq tl. The COSY spectrum B employed stronger resolution enhancement, with a sine bell-squared window function phase-shifted by n/16 applied along t2 and a sine bell phase-shifted by n/32 along tl. Note that while the dispersive diaqonal has been siqnificantly suppressed, artefacts have been introduced in the cross peak multiplets. In the DQF-COSY spectrum the chemical shifts are indicated for the complete spin system of Phe 4 and for two protons of the non-rotating Phe 22 (17). Note that doublet and triplet fine structures in the cross peaks can be distinquished from the different spacinq of the fine structure components: while the spacing for doublets is equal to J. it amounts to 2J for triplets. For example, the cross peak between protons 4 and 5 of Phe 22 has a 1:0:-l triplet fine structure in both dimensions. 483

6.6

Vol. 117, No. 2, 1983

BIOCHEMICAL

A few comments experiment

versus

principle,

due to the

quantum

coherence,

factor

the

usually

restriction

the

rather

to the intensity determine

the

factors

comparable

by the

of tl

into to the

of the coherence

transfer

account,

peak dispersion

level the

conventional

size

which

through is

double

reduced

of the cross tails,

by a

In practice, is

in the normal

sense

peaks

as these

can be usefully

sensitivity

In

COSY experiment

ratio

and by the

DQF-COSY

be made.

COSY experiment.

of a conventional

noise

contour

of the

should

signal-to-noise

of the diagonal lowest

sensitivity

of a DQF-COSY experiment

sensitivity

not

by the amount

sensitivity

RESEARCH COMMUNICATIONS

COSY experiment

to the conventional

effective

determined

the relative

conventional

of two relative

however,

these

the

about

AND BIOPHYSICAL

plotted.

but

relative

criteria Taking

of the DQF-COSY experiment

is

COSY experiment.

MATERIALS AND METHODS: The sample used in the experiments was a 20 mM solution of basic pancreatic trypsin inhibitor, BPTI (Trasylol@, obtained as a gift from Farbenfabriken Bayer, Leverkusen), in 2H20, p2H 4.6. The experiments were performed at 36OC on a Bruker AM-360 spectrometer. The conventional COSY spectrum was obtained as described elsewhere (7). The DQF-COSY spectrum was recorded using the following pulse sequence (8,9): (to - 90:

- tl

- 90:

- r - 90; - acquisition($))4n

7 was set to be 4 us to allow The relaxation delay to was 1 s and the interval time for switching rf phases between the second and third pulses. The basic four step phase cycling to select for double quantum coherence consisted of cycling the third pulse as $=x,y,-x,-y and cycling the data routing in the ASPECT-2000 computer as I!J=x,-y,-x,y (18). As judqed by the degree of suppression of singlet lines, better performance was obtained with further phase cycling by repeating the entire four step cycle four ti;es with the phases of all pulses and the data routing incremented together by 90 each time, giving a sixteen step cycle (18,19). The DQF-COSY spectrum shown in this paper was obtained with the transmitter placed at the low field end of the spectrum, and thus there was no need for quadrature detection in the wl dimension. Cuadrature detection in w1 can be achieved for a phase-sensitive presentation by incrementing the phase of the first pulse by 90° for successive tl values, and processing the data as described elsewhere (7). Alternatively, a second data set can be recorded using the same tl values but with the phase of the first pulse advanced by 90°, and the data should then be processed as described by Bachmann et al. (20) and States et al. (21). The data was multiplied by phase-shifted sine bell window functions prior to Fourier transformation in each of the two dimensions. ACKNOWLEDGEMENTS: We thank Dr. D. Neuhaus for several helpful discussions. O.W.S. acknowledges support from the Danish Natural Science Research Council (J.Nr. 11-3933). This research was supported by the Swiss National Science Foundation (project 3.284-0.82) and by the Kommission zur Forderung der wissenschaftlichen Forschunq (project 1120). REFERENCES: 1. Aue, W.P., Bartholdi, E. and Ernst, 2. Nagayama, K., Kumar, A., Wiithrich, J.Magn.Reson. 40, 321-334.

R.R. (1976). K. and Ernst,

484

J.Chem.Phys. R.R. (1980).

64,

2229-2246.

Vol. 117, No. 2, 1983

3. 4. 5. 6. 7.

8. 9. 10. 11. 12. 13.

BIOCHEMICAL

AND BIOPHYSICAL

RESEARCH COMMUNICATIONS

Bax, A. and Freeman, R. (1981). J.Maqn.Reson. 44, 542-561. Nagayama, K. and Wiithrich, K. (1981). Eur.J.Biochem. 114, 365-374. Wagner, G., Kumar, A. and Wiithrich, K. (1981). Eur.J.Biochem. 114, 375-384. Wider, G., Lee, K.H. and Wiithrich, K. (1982). J.Mol.Biol. 155, 367-388. Marion, D. and Wiithrich, K. (1983). Biochem.Biophys.Res.Commun. 113, 967-974. Piantini, U., S@rensen, O.W. and Ernst, R.R. (1982). J.Am.Chem.Soc.104, 6800-6801. Shaka, A.J. and Freeman, R. (1983). J.Magn.Reson. -51, 169-173. Selrensen, O.W., Eich, G.W., Levitt, M.H., Bodenhausen, G. and Ernst, R.R. Prog.NMR Spectros. (in press). (1983). Wokaun, A. and Ernst, R.R. (1977). Chem.Phys.Lett. =,407-412. Bodenhausen, G., Koqler, H. and Ernst, R.R. (1983). J.Magn.Reson. (submitted). Braunschweiler, L., Bodenhausen, G. and Ernst, R.R. (1983). Mol.Phys. 48, 535-560.

14. 15. 16. 17.

18. 19. 20. 21.

Smrensen, O.W., Levitt, M.H. and Ernst, R.R. (1983). J.Maqn.Reson. 54, (in press). Bax, A., Freeman, R. and Morris, G.A. (1981). J.Magn.Reson. 43, 333-338. Snyder, G.H., Rowan, R.111, Karplus, S. and Sykes, B. (19751.Biochemistry 14, 3765-3777. Biophys.Struct.Mechanism Wagner, G., DeMarco, A. and Wiithrich, K. (1976). 2, 139-158. Bax, A., Freeman, R. and Kempsell, S.P. (1980). J.Am.Chem.Soc. 102, 4849-4851. D-1. and Richards, R.E. (1975). Proc.Roy.Soc.Ser.A 344, 311-320. Hoult, Bachmann, P., Aue, W.P., Miiller, L. and Ernst, R.R. (1977). J.Magn.Reson. 2, 29-39. States, D.J., Haberkorn, R.A. and Ruben, D.J. (1982). J.Magn.Reson. 48, 286-292.

485