Fusogenic activity of SIV (Simian Immunodeficiency Virus) peptides located in the GP32 NH2 terminal domain

Vol. 175, No. 3, 1991

BIOCHEMICAL

March 29, 1991

FUSOGENIC

AND BIOPHYSICAL RESEARCH COMMUNICATIONS Pages 8724379

ACTIVITY OF SIV (SIMIAN IMMUNODEFICIENCY VIRUS) LOCATED IN THE GP32 NH2 TERMINAL DOMAIN

I. Martin, F. Defrise-Quertain, T. Saermark,* A. Burny,il R. Brasseur,

Laboratoire

“Laboratoire

Received

V. Mandieau, N. M. Nielsen,* J-M. Ruysschaert, and M. Vandenbranden

de Chimie-Physique des Macromol&cules aux Interfaces Universite Libre de Bruxelles, 1050 Bruxelles, Belgium

*EEC concerted

action/

Medical physiology 2200 Kobenhavn

de Chimie

January

Biologique,

14,

PEPTIDES

B, Panum Institute, N, DENMARK

CP206/2,

Blegdamsvej

Universite Libre de Bruxelles, Genbse, Belgium

3C, DK-

1640 Rhode-St-

1991

Peptides of 12, 16 and 24 amino acids length corresponding to the NH2 terminal sequence of SIV gp32 were synthesized. Fluorescence energy transfer studies have shown that those peptides can induce lipid mixing of SUV (Small Unilamellar Vesicles) of various compositions at pH 7.4 and 37°C. LUV (Large Unilamellar Vesicles) were shown to undergo fusion, provided they contained PE in their lipid composition. This work is an attempt to determine how the fusogenic activity depends on the structure of the peptide inserted into a lipidic environment. The peptides secondary structure and orientation in the lipid bilayer were determined using Fourier Transform infrared spectroscopy (FTIR). They adopt mainly a &sheet conformation in the absence of lipids. After interaction with DOPC SUV, the l&sheet is partly converted into a-helix oriented obliquely with respect to the membrane interface. We bring here evidence that this oblique orientation is a prerequisite to the fusion process. 0 1991 Academic Press. Inc.

Membrane molecular with

fusion

mechanism

model

of membrane

membranes

destabilization fusion,

is one of the primary has

of the lipid

were

described

domain” peptides

However,

there is no obvious

Our strategy

understood.

demonstrated

bilayer.

poorly

that

Non bilayer

of the hemagglutinin corresponding

membrane

structures,

to this

domain between

events which

to a better understanding

is based on the use of synthetic

(1). However,

fusion

peptides

In influenza into

induce

fusion

occur prior

work

requires

the

for membrane

or proteins

could

virus, the so-called

the lipid

the conformational

the

Extensive

responsible

by which

is able to insert

relationship

and the destabilizing

attempt to contribute

remains

and induce fusion is unknown.

synthetic

fusion domain

infection

(2) but the mechanism

generate such structures “fusion

fusion

event of virus

of

bilayer

(3) and

liposomes properties

(4). of the

to fusion. This work

is an

of such a relationship. peptides

corresponding

to the fusogenic

sequence of the SIV gp32. We report here that the capacity of such synthetic depends on the lipid

composition

0006-291X/91 $1.50 Copyright 0 1991 by Academic Press, Inc. All rights of reproduction in any form reserved.

peptides

of vesicles, the structure

872

to induce

fusion of liposomes

and the mode of insertion

of

Vol.

175,

No.

BIOCHEMICAL

3, 1991

the peptides

into lipid

the mean orientation Peptides lipid

with

fusogenic

membrane.

MATERIAL

model membranes. of o-helix activity

AND

BIOPHYSICAL

Polarized

and B-sheet

infrared

in the lipid

were demonstrated

The role of this unusual orientation

RESEARCH

COMMUNICATIONS

spectroscopy bilayer

to be obliquely

spectra allows

to be determined. oriented

into the

in the fusion process is discussed.

AND METHODS

Material: Bovine brain phosphatidylethanolamine (PE), egg phosphatidylcholine (PC), soya bean phosphatidylinositol (PI) and cholesterol (Chol) were purchased from SIGMA Chemical Company (Saint Louis, USA). 1,3-diazol) Dioleoylphosphatidylcholine N-(Nitrobenzo-2-oxa (DOW, phosphatidylethanolamine (NBD-PE) and N-(lissamine rhodamine B sulfonyl) phosphatidylethanolamine (Rh-PE) were from Avanti Polar Lipid Inc. (Birmingham, Ala, USA). Peptide synthesis: Peptides (F1G.I) were synthesized using a commercially available peptide synthesizer (model Biolynx, Pharmacia Biochrome, Cambridge UK) and the preweighted Fmoc amino acid OPFP esters supplied for this machine (Pharmacia Biochrome, Cambridge UK). The acylation rate was monitored by the Bioplus software using the release of anionic dye (Acid violet 17.3 mg pr lOOm1 dimethylformamide and 0.14 ml diisopropylethylamine) at 600nm. The principle is known as counter ion distribution monitoring, CDM, and is described in (17). The linkers used resulted in release of peptide amide (Ultrosyn C, Pharmacia Biochrome, Cambridge UK). The peptides were cleaved from the resin using trifluoroacetic acid with the addition of 2% anisol and 2% ethanedithiol for 2h followed by ether precipitation. The peptide was purified to more than 95% purity by HPLC on a TSK l20T reverse phase column (7.5 x 300mm) (Pharmacia, Sweden).The peptides typically eluded at 65% acetonitrile (between 65 and 75%) using a linear gradient over 90 minutes from 0 to 80% acetonitrile in 0.1% trifluoroacetic acid. The sequence was verified by protein sequencing on an Applied iosystem sequencer. The peptides were dissolved in DMSO at concentration of 8 loVesicles preparation: -Multilamellar vesicles (MLV) were obtained by vortexing a lipid film in a buffer (10 mM Hepes, 150 mM NaCl, pH 7.4). -Small unilamellar vesicles (SUV) were prepared by sonication of the MLV with a Branson Sonifier B12 for 15 min. The sonicated suspension was centrifuged at 8000xg for 10 min to remove titanium and residual multilamellar vesicles. -Large unilamellar vesicles (LUV) were prepared according to the extrusion procedure (5) using an Extruder (Lipex Biomembranes Inc., Vancouver, Canada). Briefly, freezed and thawed MLV were extruded 10 times through two stacked polycarbonate membranes with a pore size of 0.1 m (Nuclepore Corp., Pleasanton, CA, IJSA.). Liposome fusion assay: Lipid mixing was determined by changes in fluorescence intensity resulting from fluorescence energy transfer between the probes NBD-PE and Rh-PE, as described (6). Fluorescence was monitored using a Jobin Yvon JY3D spectrofluorimeter with excitation and emission slits of 4nm. Both probes were introduced into the lipid film and SUV or LUV were prepared as described above. Labeled liposomes containing both probes at 0.6% (molar ratio) each, were mi ed in a l/9 mole ratio with probe free liposomes at a final lipid concentration of 3 lo- %M. The initial fluorescence of the l/9 (labeled/unlabeled) suspension was taken as 0% fluorescence and the 100% fluorescence was determined using an equivalent concentration of vesicles prepared with 0.06% of each fluorescent phospholipid. The suspensions were irradiated at 470nm and the NBD fluorescence was recorded at 530nm. Infrared attenuated total reflection spectroscopy (ATR): Spectra were recorded with a Perkin Elmer infrared spectrophotometer FTIR 1720X equipped with a Perkin Elmer microspecular reflectance accessory (ref P.E 221-0357) and a polarizer mount assembly 873

Vol.

175,

No.

3, 7991

BIOCHEMICAL

AND

BIOPHYSICAL

RESEARCH

COMMUNICATIONS

with a silver bromide element (ref P.E.L 106-0249). The internal reflection element was a Germanium plate (50X20X2mm) with an aperture angle of 45”, yielding 25 internal reflections. A complete description of the procedures allowing the determination of the peptide secondary structure and of its orientation is given in (7) and (8). Mu_[4iluyers formation: SIV peptides, dissolved in DMSO, were added to DOPC SUV (13 10 M)in buffer at a molar lipid/peptide ratio of 65/l. After an incubation of 1 hour at 37”c, aggregated and/ or fused liposomes were centrifuged (Sigma 1OlM centrifuge) at 10 000 X g for 15 min at room temperature. A flotation sucrose gradient (30%-2%) was used to separe the lipid-peptide complex from the free peptides or the free liposomes. After centrifugation (Ultracentrifuge Beckman L765) at 35 OOOXg at 4°C for 16 hours in a SW60 Beckman rotor, the gradient was fractionated and the fractions were tested for the presence of PC using the calorimetric test of Boehringer Mannhein. Liposomes were collected and dialyzed 36 hours at 4°C against tridistilled water (microdialyzer, Pierce) to eliminate sucrose and salts which interfere with the IR analysis. Oriented multilayers were obtained by slow evaporation of the liposomes under N2 on one side of the germanium plate. To differentiate between the a-helix and the random structures, the multilayers were exposed 3 hours to D20 - saturated N2 (7). RESULTS Lipid

mixing

assays.

Fluorescence (Fig.1) rapid

energy

studies show that addition

to SUV of DOPC or PC/PE/SM/Chol lipid

mixing

compositions The

transfer

presence

of

(DOPC/Chol) maximum

lipid

which

Surprisingly,

labeled

the fusion mixing

and unlabeled

in

DOPC

whose curvature

with

neutral

SUV

the peptide

and stability

fusion

better

activity

mimick

when

DOPC liposomes

PE is replaced

with

PC

the lipid

and

l/.75

(data not shown).

The

PC/PE/SM/Chol,

mammalian

a

plasma cell.

length increases.

membrane

vesicles (LUV),

structure.

Fusion

liposomes (fig 3). No lipid mixing

in the lipid

formed

during

:52.6, 26.3, 21% w/w) bilayer

forms hexagonal

the fusion mechanism

*SIVITlZaa:GLY-VAL-PHE-VAL-LEU-GLY-PHE-LEU-GLY-PHE-LEU-ALA 1 5 *sIVsFIlCaa:GLY-VAL-PHE-VAL-LEU-GLY-PHE-LEU-GLY-PHE-LEU-ALA 1 5 -THR-ALA-GLY-SER 15 16 *sI~ZI;aa:GLY-VAL-PHE-VAL-LEU-GLY-PHE-LEU-GLY-PHE-LEU-ALA 1 5 -THR-ALA-GLY-SER-ALA-MET-GLY-ALA-ALA-SER-LEU-THR 15 20

FIGURE (BK28).

l/.5

at was

or on

(data not shown).

process at 37°C and pH 7.4. PE

structures

ratio

on large unilamellar

(PC /SM/Chol

The presence of phosphatidylethanolamine the fusion

a

For all the liposome

of

made

lipids of a typical

37°C was only observed with the PC/PE/SM/Chol

lipidic

a molar

the extent of fusion decreases when the peptide

We also determined

observed

in

of the SIV peptides

is observed

contains the major

vesicles (Fig.2).

induces

at 37°C than at 20°C (data not shown).

vesicle

activity

type SIV peptides

(26.3, 26.3, 26.3, 21 % w/w)

is more efficient

cholesterol

reduces of

composition

between

tested, the fusion

of the wild

seems to be necessary for phases which

10

12

10

10

24

1. Amino acid sequences of peptides from the gp32 N-terminus

874

are transient

(9)

of SIV,,,

Vol.

175,

No.

3, 1991

BIOCHEMICAL

AND

% Fusion

RESEARCH

COMMUNICATIONS

% Fusion

70

a suv

BIOPHYSICAL

70SUV

DOPC

PC/PE/SM/Chol

60,

50 i 40

i

30.

SlVWTlDaa

I

WTlGas

20 WT24aa 10

0

2

4

6

8

(mi$

10

0

2

4

6

6

10

Time

12

Time

FIGURE 2. Fusion percentage of small SIVWT12aa, SIVWTl63 and SIVWT24aa

unilamellar vesicles (SUV) after addition of peptides at pH = 7.4 and 7°C. The lipid 3 peptide concentration is 1.3 IO- M. The molar

concentration is 3 IO M and the lipid/peptide ratio is 25. (a)Fusion of DGPC liposomes in the presence of SIVWT peptides. (b)Fusion of PC/PE/SM/Chol (26.3, 26.3, 26.3, 21 % w/w) liposomes

in the presence

of

SIVWT peptides.

Structure

and

In order

to investigate

spectroscopy

orientation

methods unordered

cm-‘)

bonding with

components

the structure

(10). Vibrational

band (1600-1700 hydrogen

of the fusogenic

a band

structures

are sensitive fitting

(7),(8).

bilayer.

in a lipid

or peptide,

procedure

The combination allows

secondary

Moreover,

we used infrared

enhancement

Total

O-sheet

Reflection

I

SlWTlPaa

SIVWTlGaa

0

5

10

15

20

25

FIGURE 3. Fusion percentage of large unilamellar vesicles (LUV) of PC/PE/SM/Chol (26.3,26.3,26.3,21 % w/w) $fter addition of SIVWT peptides at 37-C a-y pH 7.4. The lipid concentration is 3 lo- M and the peptide concentration is 1.3 10 M. The molar lipid/peptide ratio is 25.

875

and

infrared

% Fusion ,Ol

1

assessment of various

such as a-helix,

the use of Attenuated

the Amide

as this involves specific

of resolution

the quantitative structure

bilayer

and particularly

to the secondary structure

groups.

or peptide

in the lipid

of the peptide

bands of protein

of the C=O

of protein

peptides

(mid

Vol.

175, No. 3, 1991

BIOCHEMICAL

AND BIOPHYSICAL

RESEARCH COMMUNICATIONS

TABLE 1. Proportion of the different secondary structures of SIVWTl2aa, SIVWT16aa and SIVWT24aa in the absence and in the presence of lipid (SUV of DOPC). The molar livid/vevtide ratio is 65 a-helix

SAMPLE

0 48

67 34

33 19

SIVWT.16aa + DOPC

9 39

76 58

15 3

SIVWT24aa + DOPC

9 45

53 20

38 35

allows

spectra to be recorded

gained about the orientation orientation

of different

of the a-helix

the peptide Polarization

on ordered

structures

and D-sheet structures

bond corresponding

and perpendicular

between

%

SIVWT12aa + DOPC

spectroscopy

(0”)

random %

P-sheet

%

of protein

incident

as the dichroic

the C=O band and a normal

and information

or peptide

was determined

to the C=O group. Spectra

(90”) polarized

was expressed

bilayers

light

ratio

from the orientation with

plate surface

The

of

parallel

respect to the ATR

Rat,. = A9p/AOo.

to the ATR

(1 I). The mean

were recorded

with

to be

plate.

mean angle

is calculated

from Rat,

(7). IR spectra of SIVWT solution

in

DMSO

conformation The

on the ATR

of DOPC

1650 cm-l

significant

SUV

to the lipid/peptide

ratio Ratr=

A9O/AO

is neither

intermediate D-sheet

revealed

evaporation

of a concentrated

that the peptides

the appearance

structure (table

orientation

of a new

(40 to 50%) content 1). The

adopt

a g-sheet

percentage

large

peak centered

structure

(Fig 4). This

is accompanied

of a-helix

by a

structure

is

ratio (table 2).

with

the incident

=1.04 corresponding

parallel

structure

orientations

by direct

of the presence of an a-helix

structure

From the spectra recorded helix

induces

of the a-helix

of the B-sheet

proportional

plate

characteristic

increase

decrease

alone obtained

(table 1).

presence

around

peptides

nor

light polarized

to the helical

perpendicular

in the lipid

corresponds

to the lipid

bilayer.

The dichroic

to an orientation

did not change whatever

acyl

indicates chain

ratio Ratr=2.3

parallel

the lipid/peptide

at 90” and O”, the dichroic

structure

to the lipid

but

that the Qadopts

an

associated to the bilayer.

These

ratio.

DISCUSSION Our results show that synthetic N-terminal promote

fusion

properties similar peptides

domain

of lipid

attributed

observation (12), giving

peptides

of 12, 16 and 24 residues corresponding

of the transmembrane vesicles,

glycoprotein

demonstrating

further

influenza

experimental

of SIV,

to the

are able to

that at least a part of the fusogenic

tv gp32 depends on a limited was made with

(gp32)

stretch of its N-terminal

peptides

support

sequence in viral fusion.

876

(4) and more recently

for the role of N-terminal

domain. with

A HIV

fusogenic

Vol.

BIOCHEMICAL

175, No. 3, 1991

p-turn

P-sheet

AND BIOPHYSICAL

random

a-helix

RESEARCH COMMUNICATIONS

p-sheet

a

1700

1680

1640

1660

1620

1600

b _--. ,,,’ -.\:,\

.‘L_ ,.,.--.._ ,>’ :\ // ,:’J’ ‘/ ,I

,.../~ ,I ..:. 2,’ (,'. ,' ,,..._

9

',.,.. -. ...'

./ <...z-.-.,~::r..-~~~~~-.z+--

--... *;&..L;~ ~~~h~&+s"-‘.

/ _--....

._-

-..-- ._...,_ ~~;IIzz--";w

---

i

.___._

,.,“.., \\ i' /

. . .. . . .;--=--=a:

-----_

\, !

1, '\

‘>' '~ ,/ ,,,, ,..,, j,' j>,.' .i .: . . .. ,/' :. ...

, '<

___-

',

:

k,

,.

,.... .\ ;c-... .---..ycL -~:~-~~~,.~~

y--w----~--~-=2-~=m+a

rm-~rrrr7-rr-rp-rrrTi~l-~r~~r~r~l 1660

>\,, \

,:. Ii

-I ..:.I--__ -...- :::Li--:y..~m~u

---.-c----z---

,.-.

1660

1640

, , , , , -1 Icm

1620

1600

FIGURE 4. Curve fitting of the amide I’ region of the SIVWT12aa at pH 7.4 in the absence of lipid (a) and in the presence of SUV DOPC (b). The result of the fitting appears under the curve. The vertical dotted lines define the region of the spectrum assigned to the different secondary structures. The sum of the components is represented by the dotted spectrum.

In contrast shorter

to studies on influenza,

one,

fusogenic,

we show

here

the 12 amino

that

where

longer

12 residues

and

model

residues peptide

has been proposed of SUV towards

of their surrounding

TABLE

peptides

than

are highly

This result suggests that in the

“fusion

the higher

domain” fusogenic

could be shorter. capacity

A

of the 12

(13).

The susceptibility curvature

to explain

were more fusogenic

16 residues

acids being the most efficient.

case of SIV (and perhaps other viruses) the so-called theoretical

peptides

lipid

fusion

or destabilization

bilayer.

is attributed

Here, we show that when

to the strong the appropriate

2. Proportion of the different secondary structures of SIVWT12aa presence of lipid (SUV of DOPC) for different lipid/peptide ratios

SIWT12aa lipid/peptide 18 36 65

+ DOPC ratio

a-helix % 17 35 48

877

P-sheet % 83 47 34

random % 0 18 19

in the

Vol.

175, No. 3, 1991

lipid composition of LUV

is chosen, LUV

is probably

unsaturated

in

destabilization

bilayer

several

component

infrared

membrane

fusion

conformational

transition

similar

to a-helical

data

processes

property

(9).

to undergo

which

The

a

is thought

need

for

to

intrinsic

fusion. that

from R-sheet to a-helix

seems not to be a specific

is known

PE with

in vesicle fusion suggests a possible role

indicate

transition

Indeed,

HII phase, a structure

factors in viral membrane

spectroscopy

S-sheet

fusion too. The decreased stability

of cell membranes,

to hexagonal

factors (PE or surface curvature)

for those additional Our

RESEARCH COMMUNICATIONS

due to the presence of phosphatidylethanolamine.

from

a role

AND BIOPHYSICAL

are able to undergo

acyl chains, a natural

phase transition play

BIOCHEMICAL

fusogenic

peptides

when interaction

was demonstrated of the fusion

undergo

a

with lipids occurs. A

with influenza

peptides (14) but

domain since it has also been observed

in signal sequence peptides (15) and some model peptides (16). As observed

in the spectra, even at high

separated spectral components the conversion

to a-helix

should reflect lipid

bilayer

lipid-peptide

in the a-helix is uncomplete.

sheet. Indeed,

it

conformations,

could

structure,

is hard

The

occurence

coexist

reason why the peptide

within

should

that

two

well

enlarge

well that

components

one penetrating

into the

in the aqueous phase as a B-

defined

a-helix

the same short sequence

gradually

showing

of two spectral

populations,

the other, remaining

to believe

there are two major

and B-sheet spectral domains,

the presence of two separate peptide in an a-helical

ratio,

its a-helix

and

D-sheet

and we see no obvious

domain

while

lipid/peptide

ratio is increasing. When fused vesicles were not separated infrared

spectrosopy,

idea that peptides

spectra reveal a major

having not interacted

In the SIV peptide, the C-terminus. domain

from

g-sheet

spectral component,

to our previous an unusual

modelling

orientation

has been demonstrated

peptide

to the lipid bilayer

was calculated

However,

to conclude

orientation

of a peptide

spectroscopy orientation.

populations

The prediction

small differences that the fusogenic orientation.

in the lipid

is unable to discriminate

(or more)

that fusogenic it should

a fixed uniaxial

capacity orientation giving

878

orientations

its orientation

several peptides

as the gp32 NH2-terminal

having

domain

induce

A tilt angle

depends

on the

that infrared and a mixture

an average

gives only a static view of the phenomenon changing

not

in this work.

be kept in mind

orientations

for various

to

for SIV. This is in accordance

bilayer,

energies

synthesized

interface.

spectroscopy

between different

segment is actually

We recently

same hydrophobicity

with

approach

in calculated

the

According

of residues might

at the lipid-water

found by infrared

it is generally

helices of proteins.

distribution

with the mean angle of orientation even if it is tempting

supporting

to be associated to the fusion

segments and surface-seeking studies, this asymmetric

of the fusogenic

of 52” from the normal

prior

increases along the helix from the N-terminus

of a series of viruses (13) and to signal peptides (11) whereas

observed in transmembrane

of two

on a sucrose gradient

with vesicles form a separate B-sheet population.

the hydrophobicity

This distribution

free peptides

oblique and the

could

reflect

the fact

quickly

around

a mean

the same length

and the

but for which

the calculated

Vol.

BIOCHEMICAL

175, No. 3, 1991

orientation between fusion peptides

into the lipid the oblique

orientation

(to be published).

Another

in the SIV glycoprotein

expressed in mammalian syncytia

bilayer

formation

has been modified. of the peptides way to verify precursor

RESEARCH COMMUNICATIONS

A good correlation

and their capacity this correlation

using directed

This

work

is currently

was observed

to induce

liposomes

is to integrate

mutagenesis.

cells and the capacity of such modified

is measured.

results seem encouraging

AND BIOPHYSICAL

The precursor

glycoproteins

in progress

these is

to induce

and preliminary

( Burny A., personal communication.)

ACKNOWLEDGMENTS This work was performed with the financial support of BNB (Banque National de Belgique), FNRS (Fonds National de la Recherche Scientifique ), IRSIA (Institut pour 1’Encouragement de la Recherche Scientifique dans 1’Industrie et I’Agriculture) and Smith Kline Beecham. One of us (R.B.) is a Research Associate of the National Fund for Scientific Research. We gratefully thank NIH (NIAID grant Al-27136-01Al) and the Commission of the European Communities (SC 1000195) for a continued financial support.

REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. IO. 11. 12. 13. 14. 15. 16. 17.

White J., Kielian M., and Helenius A. (1983) Quart. Rev. Biophys. 16, 151-195. Verkleij A.J. (1984) Biochim. Biophys Acta 779, 43-63. Harter C., James P., Biichi T., and Brunner J., (1989) J. Biol. Chem. 264, 6459-6464. Wharton S.A., Martin S.R., Ruyrok R.W.H., Skehel J.J.,and Wiley D.C. (1988) J. Gen. Virol. 69, 1847-1857. Hope M.J., Bally M.B., Webb G., and Cullis P.R. (1985) Biochim. Biophys. Acta 812, 55-65. Struck D. K., Hoekstra D., and Pagan0 R.. (1981) Biochemistry 20, 4093-4098. Cabiaux V., Brasseur R., Wattiez R., Falmagne P., Ruysschaert J.-M., and Goormaghtigh E. (1989) J. Biol. Chem 264, 4928-4938. Goormaghtigh E., Cabiaux V., and Ruysschaert J.-M. (1990) Eur. J. Biochem. 193, 409-420. Ellens H., Siegel D., Alford D., Yeagle P. Boni L., Lis L., Quinn P., Bentz J. (1989) Biochemistry 28, 3692-3703. Fringeli U. R., Giinthard M. H. in “Membrane Spectroscopy.“, ed. E. Grell Springer-Vellay pp 270-332, (1981). Goormaghtigh E., Martin I., Vandenbranden M., Brasseur R., Ruysschaert J.-M. (1989) Biochem. Biophys. Res. Comm. 158, 610-616. Rafalski M., Lear D., and DeGrado W.F. (1990) Biochemistry 29, 7917-7922. Brasseur R., Vandenbranden M., Cornet B., Burny A., and Ruysschaert J. -M. (1990) Biochim. Biophys. Acta, 1029267-273. Lear J.D., and DeGrado W.F. (1987) J. Biol. Chem. 262, 6500-6505. Roise D., Horvath S.J., Tomich J.M., Richards J.M., Schatz G. (1986) EMBO J. 6,1327-1332. Subbarao N., Parente R., Szoka F., Nadoudi L., Pongracz K. (1987) Biochemistry 26, 2964-2972. Salisbury S.A., Treemeer E.J., Davies J.W., Owen D. (1990) J. Chem. Sot. Chem. Commun. 538-540.

879