Journal of General Virology (1992), 73, 1703-1707. Printedin Great Britain
1703
A leucine zipper structure present in the measles virus fusion protein is not
required for its tetramerization but is essential for fusion
Robin Buckland, Etienne Maivoisin, Philippe Beauverger and Fabian Wild*
Immunologie et Stratbgie Vaccinale, Institut Pasteur de Lyon, Avenue Tony-Garnier, 69365 Lyon, Cedex 07, France
The biological role o f a leucine zipper m o t i f present in
the measles virus fusion (F) protein has been investigated. This m o t i f is present in all paramyxovirus F
proteins, all coronavirus spike proteins and many if not
all retrovirus envelope proteins. By analogy to its role
in certain transcription factors, it has been suggested
that the m o t i f may be responsible for the oligomeriza-
tion o f these viral m e m b r a n e proteins. In this study,
one, two or four heptadic leucines in the m o t i f were
substituted using site-directed mutagenesis. We found
that fusion is prevented when all four heptadic leucines
present in the m o t i f are mutated whereas cellular
transport and the oligomeric state of the F protein are
unaffected.
Introduction
in their env glycoproteins (Delwart & Mosialos, 1990),
again close to the transmembrane region. It has been
suggested that the leucine zipper motifs could contribute
to the oligomeric structure of these viral proteins
(Buckland & Wild, 1989; Delwart & Mosialos, 1990;
Tucker et al., 1991). Oligomerization of fusogenic viral
envelope glycoproteins is involved not only in their
cellular transport (Gething et al., 1986 b; Kreis & Lodish,
1986) but also in fusion pore formation (Spruce et al.,
1991). Disruption of oligomerization at both levels could
affect fusion. In order to investigate the biological role of
such motifs in paramyxovirus F proteins we have
undertaken site-directed mutagenesis to substitute the
heptadic leucines in the motif present in the measles
virus (MV) F protein.
Enveloped viruses possess fusogenic envelope (env)
glycoproteins responsible both for the initial fusion of
viral and target membranes necessary for virus penetration and for subsequent cell-to-cell infection. The best
studied viral env glycoprotein is the haemagglutinin
(HA) of influenza virus (Wiley & Skehel, 1987; Stegmann et al., 1989; White, 1990). In this virus, as in
paramyxoviruses, coronaviruses and retroviruses, a
fusion (F) peptide liberated via a conformational change
in protein structure makes the primary contact with the
target membrane but is unlikely to be solely responsible
for the fusion process (Gething et al., 1986a). Fusion
begins with the formation of a 'fusion pore' which is
thought to be no more than between 1 and 2 nm wide at
first but later dilates, 'slowly and haltingly', to more than
100 nm (Spruce et al., 1991). However, the molecular
basis for these events is poorly understood.
The leucine zipper (Landschultz et al., 1988), formed
by two parallel amphipathic ~-helices, which structurally
resembles a coiled-coil (O'Shea et al., 1989), has been
shown to mediate the dimerization of certain transactivators (Landschultz et al., 1989; Hope & Struhl,
1987). Paramyxovirus F and coronavirus spike fusogenic
proteins all possess a leucine zipper motif in a region
predicted to be an amphipathic ~-helix 10 amino acids
N-terminal from the transmembrane region (Buckland &
Wild, 1989; Britton, 1991) and many retroviruses
including human T cell lymphotropic virus type 1
(HTLV-1), HTLV-2, bovine leukaemia virus, Moloney
murine leukaemia virus and human immunodeficiency
virus type 1 (12 are listed) also contain zipper-like motifs
0001-0876 © 1992 SGM
Methods
lmmunofluorescence. Immunofluorescencestudies were performedon
HeLa cells which bad been infected with wild-type(wt) vaccinia virus
(VV) then transfected with the appropriate plasmid. Cells were
detached using PBS/lmM-EDTA,dried on microscopeslides and then
fixedwith acetone at - 20 °C for 10 min. The presenceof the F antigen
was examined using an anti-F protein monoclonal antibody (MAb)
prepared against the denatured F~ antigen (de Vries et al., 1988).
Fusion assay. Two hours after infection (m.o.i of 0-2 p.f.u./cell) of
HeLa cells with the MV HA-expressingVV recombinant (Drillien et
al., 1988) the cells were transfected by the lipofectin method (BRL)
with 10 p~gof either the wt F cDNA or a mutated form cloned into the
pgptA6plasmid. This allowedthe expressionof the cDNA from the VV
11K late promoter (Stunnenberget al., 1988). Twenty hours later the
cells were examinedmicroscopicallyfor the presenceof syncytia(Wild
et al., 1991) and then detached, fixed in acetone and examined by
immunofluorescenceusing anti-MV F protein MAbs.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sat, 17 Jun 2017 16:42:01
1704
R. Buckland and others
Site-directed mutagenesis. This was performed using the Bio-Rad
Mutagene Kit and all mutants were checked by DNA sequencing.
Preparation of VV recombinants. To generate VV recombinants
expressing the modified glycoprotein, the wt and mutated versions of
an MV F protein cDNA (Buckland et al., 1987) were first placed in the
pgptA6 VV pplsmid under the control of the 11K VV promoter
(Stunnenberg et al., 1988). VV recombinants were then isolated as
described by Falkner & Moss (1988).
Fluorescence-activated cell sorter (FACS) analysis. HeLa cells were
infected (m.0.i. of 0.2 p.f.u./cell) with VV recombinant viruses
expressing either the wt F protein sequence or the 4L mutation. After
24 h the cells were detached using PBS/l m ~ - E D T Aand washed in
MEM containing 5% foetal calf serum (FCS). Cells were then
resuspended in PBS, 1% BSA, 0.1 % sodium azide at 2 x 106cells/ml.
Aliquots (200~1)were incubated with anti-F protein MAb 186
(Malvoisin & Wild, 1990) for 30 min at 4 "C, washed, incubated with
anti-mouse fluorescein isothiocyanate for a further 30 min at 4 T ,
washed again and the cells were examined.
Sucrosegradientsedimentationanalysisof wt Fprotein and 4L Fprotein
oligomerization and cleavage. [35S]Methioninewas used to label both wt
and 4L VV recombinant F proteins in HeLa cells. The cells were lysed
in the presence of 100 m~-Tris-HC1pH 8.0, 100 m~-NaCl,250 mMoctyl glucoside (Sigma) and 10 m~-iodoacetamide.Aliquots (0-4ml) of
the cleared lysate were loaded onto 10 m15 to 20% (wlv) linear sucrose
gradients containing 100 mht-NaC1, 50 m~-Tris-HC1pH 8.0,25 mMoctyl glucoside, which were centrifuged in an SW41 rotor at 38000
r.p.m. for 20 h at 4 OC and 1 ml fractions were taken (pumped from the
bottom).
SDS-PAGE. Fractions from the sucrose gradients were diluted in
buffer containing 10 m~-Tris-HC1pH 8.0, 100 m~-NaC1and 25 mMoctyl glucoside and immunoprecipitatedusing anti-F protein MAb 186
(Malvoisin & Wild, 1990). Staphylococcus aureus Protein A bound to
Sepharose beads (Pharmacia) was used to collect antibody-antigen
complexes. Antigen was released from these complexes by boiling for
2 min in SDS-PAGE sample buffer containing 2% 2-mercaptoethanol
and loaded onto a 10% SDS-polyacrylamide gel which, after drying,
was exposed to photographic film (Amersham Hyperfilm-Bmax).
Viruses and cells. HeLa cells were cultivated in MEM containing 5%
FCS plus antibiotics (penicillin and streptomycin). The Copenhagen
strain of VV and recombinants derived from this strain were used.
Results
To address the question of whether the leucine zipperlike sequence in the MV F protein is required for fusion
we used site-directed mutagenesis to substitute 1, 2 or 4
heptadic leucines in the motif (but maintaining the orhelicity) in an MV F protein cDNA (Buckland et al.,
1987) (Fig. 1). It should be noted that methionine and
alanine are hydrophobic whereas serine is hydrophilic
but uncharged. To study fusion, we transfected the
various plasmids into HeLa cells previously infected
with a recombinant VV expressing the MV HA. The
cells were observed after a further 20 h for syncytium
formation. Mutation of the first and second heptadic
leucines did not affect fusion but mutation of all four
heptadic leucines (the 4L mutant) prevented fusion
Fig. I . (a) F protein, the wt zipper sequence. (b) 1L mutant sequence
which differs from wt F protein in having the fourth heptadic leucine of
the zipper mutated to an alanine. (c) 2L mutant sequence which is as 1L
but also has the third heptadic leucine mutated to a methionine. (d)4L
mutant sequence which is as 2L but also has the first and second
heptadic leucines both mutated to serine. (e) Sequence of the
(LSSLLSL)3 peptide (Lear et al., 1988). Parts (f) to (h) show helical
wheel representations of sequences (a), (d) and (e).
Table 1. Fusion activity of MV wt F and mutant F proteins
VV-F plasmid
Fusion*
Immunofluorescence
wt F
1L
2L
4L
+
+
+-
+
+
+
+
* HeLa cells previously infected with the VV-HA recombinant were
transfected with one of the VV-F plasmids and syncytium formation
was monitored.
completely (Table 1). Immunofluorescence studies on the
cells confirmed that in each case the F antigen was
expressed. To study whether the abolition of fusion was
due to a block at the level of transport we examined the
surface expression of the mutated antigen by FACS
analysis. HeLa cells were infected with VV recombinants expressing either the wt F protein or the 4L mutant.
The expression of the F antigen at the cell surface was
examined by FACS analysis using anti-F protein MAbs
(Fig. 2). These results show that the 4L mutant is
transported to the cell surface and that it is in a native
conformation.
As the cellular transport of viral glycoproteins has
been shown to be dependent on oligomerization (Gething et at., 19866; Kreis & Lodish, 1986), these results
suggest that the oligomeric form of the MV F protein
required for transport is not affected by the 4L mutation.
However, the absence of fusion could be explained by an
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sat, 17 Jun 2017 16:42:01
M V fusion protein leucine zipper
(a)
t
U~
I
I
I
(b)
{
6o
t'
I
1705
!
i i~
40
10 °
10 ~
10 2
10 3
10 4
10 °
10 ~
10 2
10 3
[0 4
Logt0fluorescence
Fig. 2. FACS analysis of surface expression on intact cells of (a) the wt F protein and (b) 4L mutant VV recombinants. HeLa cells
infected at 0.2 p.f.u./cell were examined 24 h after infection for the surface expressionof F antigen using an anti-F protein MAb
recognizing a conformational epitope (Malvoisin & Wild, 1990). Non-infected cells were used as controls.
~)
~)
M 1 3 5 7 9 1 1 1 3 5 7 9
200K
11
97K
69K
• 46K
30K
Fig. 3. Lanes 1 to 11 refer to alternate fractions from sucrosegradients
used (1 is the bottom fraction) to separate oligomericforms of (a) wt F
protein and (b) the 4L mutant, immunoprecipitatedand loaded onto
10~ SDS-polyacrylamidegels in the presence of reducing agent. The
arrow indicates cleaved Ft. M, standards are shown in lane M.
The sedimentation of VV-expressed wt F protein in 5 to
20~o linear sucrose gradients gives a broad peak over the
lower part of the gradient (Fig. 3) representing tetramers
to higher-order oligomers. The 4L F protein mutant gives
a profile similar to that of the wt F protein suggesting
that both tetramerization of the MV F protein and its
higher oligomerization does not involve the leucine
zipper motif. In the activation of the MV F protein, the
Fo, Mr 62K, is cleaved to F1 (47K) and F2 (15K), the
cleavage products remaining associated via a disulphide
bridge. The presence of the FI band under reducing
conditions shows that the 4L mutant is cleaved in the
same manner as the wt F protein.
Discussion
inability to form higher oligomeric states or alternatively
by a lack of cleavage. To study this, 3sS-labelled F
protein was solubilized from wt or 4L VV recombinantinfected H e L a cells and, after centrifugation on sucrose
gradients, were immunoprecipitated with an anti-F
protein MAb, and analysed by S D S - P A G E (Fig. 3). The
paramyxovirus F protein is thought to be tetrameric
(Sechoy et al., 1987; Collins & Mottet, 1991) and we have
recently confirmed this for the MV F protein (E.
Malvoisin, R. Buckland & F. Wild, unpublished results).
Fusion does not take place in the case of the 4L F protein
mutant even though the polypeptide is cleaved and is
present at the cell surface in the same oligomeric state as
is the wt F protein. One explanation for this could be that
the mutated F protein can no longer make the association
with H A shown to be required for fusion to take place
(Wild et al., 1991). We are currently investigating this
possibility but another possible explanation is that
formation of the fusion pore itself is affected by the 4L
mutation. Ojcius & Young (1991) suggest that a bundle of
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sat, 17 Jun 2017 16:42:01
1706
R. Buckland and others
amphipathic s-helices (or fl-sheets) represents the only
three-dimensional structure needed to create membrane
pores. An amphipathic or-helix could be a common
structural motif not only for cytolytic pore formation but
also for enveloped virus fusion. Many proteins forming
oligomeric pores contain amphipathic ~t-helices (Ojcius
& Young, 1991) within which leucine zipper-like motifs
are often present (R. Buckland & F. Wild, unpublished
observations). Leucine could be predominant in such
helices not only because of its helix-forming propensity
and high hydrophobicity but also because its long
branched hydrophobic side-chain is optimal for filling
space between helices in a knobs-into-holes fashion
(Crick, 1953). Among the pore-forming proteins are
toxins which lyse cells through a 'barrel-stave' mechanism in which monomers are thought to bind to and
insert into the target membrane and then aggregate like
barrel staves surrounding a central water-filled pore that
increases in diameter through the progressive recruitment of additional monomers, the pore being formed by
their lateral oligomerization (Ojcius & Young, 1991). A
21 residue synthetic peptide composed exclusively of
leucine and serine residues and designed to be a
membrane-spanning amphipathic s-helix, (LSSLLSL) 3
(Fig. 1), forms ion channels in membranes (Lear et al.,
1988). This peptide is sufficiently long to traverse the
membrane and possesses a leucine zipper motif consisting of three heptad repeats. Whereas the pore formed by
the (LSSLLSL)3 peptide is thought to be a hexamer (Lear
et al., 1988), the fusion pore formed by the MV F protein
is probably a much higher oligomer. It has been
estimated (Henderson & Unwin, 1975) that eight ~helices spaced 1 nm centre-to-centre would only form a
1.6 nm diameter pore whereas the fusion pore of
influenza virus has been estimated to have a diameter of
more than 100 nm (Spruce et al., 1991).
The number of heptadic leucines in the (LSSLLSL)3
peptide (Fig. 1) could suggest that the fourth leucine in
the MV F protein zipper is redundant for pore formation
which would explain why fusion is unaffected in the 1L
mutant. It has also been shown that a conservative
substitution of just one heptadic leucine in thejun zipper
does not abolish dimerization (Kouzarides & Ziff, 1988),
but the two mutations to serine (the first and second
heptadic leucines of the zipper) might have their effect
because this involves the part of the motif with a 4,3
hydrophobic repeat that most resembles a coiled-coil.
The leucine zippers found amino-terminally adjacent
to the transmembrane region in paramyxoviruses,
coronaviruses and retroviruses are topologically well
placed for participating in fusion pore formation. We
propose that in MV fusion the fusion pore is formed
through the progressive recruitment of the tetrameric F
protein and that although lateral oligomerization is not
mediated exclusively by the leucine residues of the zipper
motif, in their absence, at the level of the target
membrane, a coiled-coil structure with the s-helices
packed together to form a water-tight fusion pore can no
longer be formed.
If amphipathic s-helices play a general role in viral
fusion they could be potential antiviral targets in
enveloped viruses. Fusion is an ideal point in viral
replication to be targeted, because if fusion can be
inhibited the cell-to-cell spread of infection is also halted.
We thank C. Bella for the FACS analysis, G. Joly for photography,
H. Stunnenberg for the pgptA6 plasmid, A. Osterhaus for the anti-F
protein MAbs, and B. Pope, J. Schmitt and E. Manet for helpful
discussions. R.B. is supported by the CNRS.
References
BRII-roN, P. (1991). Coronavirus motif. Nature, London 353, 394.
BUCKLAND, R. & WILD, F. (1989). Leucine zipper extends. Nature,
London 338, 547.
BUCKLAND, R., GERALD, C., BARKER, R. & WILD, Z. F. (1987). Fusion
glycoprotein of measles virus: nucleotide sequence of the gene and
comparison with other paramyxoviruses. Journal of General Virology
68, 1695-1703.
COLLINS, P. ~g MOTTET, G. (1991). Post-translational processing and
oligomerization of the fusion glycoprotein of human respiratory
syncytial virus. Journal of General Virology 72, 3095-3101.
CRICK, F. H. C. (1953). The packing of ct helices: simple coiled coils.
Acta crystallographiea 6, 689-697.
DE VRIES, P., VANBINNENDIJK, R. S., VANDER MAREL, P., VANWEZEL,
A. L., VOORMA, H. O., SUNDQUIST, B., UYTDEHAAG, F. G. C. M. &
OSTERI-IAUS, A. D. M. E. 0988). Measles virus fusion protein
presented in an immune-stimulating complex (iscom) induces
haemolysis-inhibiting and fusion-inhibiting antibodies, virusspecific T cells and protection in mice. Journalof General Virology69,
549-559.
DELWART, E. & MOSIALOS, G. (1990). Retroviral envelope glycoproteins contain a leucine zipper-like repeat. AIDS Research and Human
Retroviruses 6, 703-706.
DRILLIEN, R., SPEHNER, D., KIRN, A., GIRAUDON, P., BUCKLAND, R.,
WILD, F. & LECOCO, J.-P. (1988). Protection of mice from fatal
measles encephalitis by vaccination with vaccinia virus recombinants encoding either the hemagglutinin or fusion proteins.
Proceedings of the National Academy of Sciences, U.S.A. 85, 12521256.
FALKNER, F. & Moss, B. (1988). E. coli gpt gene provides dominant
selection for vaccinia virus open-reading frame expression vectors.
Journal of Virology 62, 1849-1854.
GETItlNO, M.-J., DOt,'-_s,R., YORK, D. & WHITE, J. (1986a). Studies on
the mechanism of membrane fusion : site-specific mutagenesis of the
hemagglutinin of influenza virus. Journalof Cell Biology 102, 11-23.
GETI-IING, M.-J., McCAMMON, K. & SAMBROOK,J. (1986b). Expression
of wild-type and mutant forms of influenza hemagglutinin: the role
of folding in intracellular transport. Cell 46, 939-950.
HENDERSON, R. & UmVIN, P. N. T. (1975). Three-dimensional model
of purple membrane obtained by electron microscopy. Nature,
London 257, 28-32.
HOPE, I. & STRUnL, K. (1987). GCN4, a eukaryotic transcriptional
activator protein, binds as a dimer to target DNA. EMBO Journal6,
2781-2784.
KOUZARIDES,T. & ZIFF,E. (1988). The role of the leucine zipper in the
fos-jun interaction. Nature, London 336, 646-651.
KREIS, T. & LODISH, H. (1986). Oligomerization is essential for
transport of vesicular stomatitis viral glycoprotein to the cell surface.
Cell 46, 929-937.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sat, 17 Jun 2017 16:42:01
M V fusion protein leucine zipper
LANDSCHULTZ, W., JOHNSON, P. & McKslGrrr, S. (1988). The leucine
zipper: an hypothetical structure common to a new class of DNA
binding proteins. Science 240, 1759-1764.
LANDSCHULTZ, W., JOHNSON, P. & MCKNIGHT, S. (1989). The DNA
binding domain of the rat liver nuclear protein C/EBP is bipartite.
Science 243, 1681-1688.
LEAR, J. D., WASSERMAN,Z. R. ~; DEGRADO, W. F. (1988). Synthetic
amphiphilic peptide models for protein ion channels. Science 240,
1177-1181.
MALVOISIN, E. & WILD, F. (1990). Contribution of measles virus fusion
protein in protective immunity: anti-F monoclonal antibodies
neutralize virus infectivity and protect mice against challenge.
Journal of Virology 64, 5160-5162.
OJclos, D. & YOUNG, J. (1991). Cytolytic pore-forming proteins and
peptides: is there a common structural motif? Trends in Biological
Sciences 16, 225-229.
O'SHEA, E., RUTKOWSKI,R. & KIM, P. (1989). Evidence that the leucine
zipper is a coiled-coil. Science 243, 538-543.
SECHOY,O., PHILPOTT, J. & BIENVENUE,A. (1987). F-protein-F-protein
interactions within the Sendal virus identified by native bonding on
chemical cross-linking. Journal of Biological Chemistry 262, 1151911523.
SPRUCE, A., IWATA,A. & ALMERS, W. (1991). The first milliseconds of
the pore formed by a fusogenic viral envelope protein during
1707
membrane fusion. Proceedings of the National Academy of Sciences,
U.S.A. 88, 3623-3627.
STEGMANN, T., DOMS, R. & HELENIUS, A. (1989). Protein-mediated
membrane fusion. Annual Review of Biophysics and Biophysical
Chemistry 18, 187-121.
STUNNENBERG, H., LANGE, H., PHILIPSON, L., VAN MILTENBURG, R. &
VAN DER VLIET, P. (1988). High expression of functional adenovirus
DNA polymerase and precursor terminal protein using recombinant
vaccinia virus. Nucleic Acids Research 16, 2431-2444.
TUCKER, S., SRINIVAS, R. & COMPANS, R. (1991). Molecular domains
involved in oligomerization of the Friend murine leukemia virus
envelope protein. Virology 185, 710-720.
WHITE, J. (1990). Viral and cellular fusion proteins. Annual Review of
Physiology 52, 675-697.
WILD, T. F., MALVOISIN, E. t~. BUCKLAND, R. (1991). Measles virus:
both the haemagglutinin and fusion glycoproteins are required for
fusion. Journal of General Virology 72, 439-442.
WILEY, D. C. t~; SKEHEL, J. J. (1987). The structure and function of the
hemagglutinin membrane glycoprotein of influenza virus. Annual
Review of Biochemistry 56, 365-394.
(Received 24 January 1992; Accepted 4 March 1992)
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sat, 17 Jun 2017 16:42:01