Human Reproduction vol.13 no.3 pp.682–689, 1998
Human gamete fusion can bypass β1 integrin
requirement
Ya Zhong Ji1,3, Jean Philippe Wolf2,
Pierre Jouannet1 and Morgane Bomsel3,4
1Laboratoire
de Biologie de la Reproduction, Hôpital Cochin, 123,
Bd Port Royal, 75014 Paris, 2Laboratoire de Biologie de la
Reproduction, Faculté de Médecine de Bobigny, 74 rue Marcel
Cachin, Bobigny 93017 and 3Institut Cochin de Génétique
Moléculaire, U 332, 22, rue Méchain, 75014 Paris, France
4To
whom correspondence should be addressed
Since α6β1 integrin has been shown to function as a
sperm adhesion receptor in the mouse, we investigated the
potential role of β1 integrin in the gamete fusion process in
humans. The expression of β1 integrin was morphologically
analysed by indirect immunofluorescence and confocal
microscopy. A homogeneous and intense staining was
detected at the plasma membrane, and in some subcortical
vesicles of germinal vesicle stage oocytes (GV). β1 almost
disappeared from oolemma and cytoplasm of metaphase I
(MI) oocytes, but was re-expressed as asymmetrical patches
at the plasma membrane of metaphase II stage oocytes
(MII). A functional fusion assay based on Hoechst or
calcein-AM dye transfer from one gamete to the other
showed that maturing oocytes were able to fuse with an
increasing number of spermatozoa (11–22 from GV to MII
respectively), and that fused spermatozoa co-localized with
β1 integrin patches. Human gamete fusion was only partially inhibited either by RGD-containing peptide
(GRGDTP), or by blocking anti-human β1 integrin monoclonal antibody (DE9), with a maximum of 50% inhibition.
Despite the combined addition of GRGDTP and blocking
mouse anti-human β1 integrin DE9 in the assay, a complete
inhibition of fusion could not be achieved. A mouse polyclonal antibody raised against human oocyte membranes
was more potent in inhibiting the fusion. Since β1 integrin
expression at the plasma membrane was not correlated to
oocyte fusibility, and since it was only partially inhibited
by DE9 and/or RGD peptide, we suggest that human
gamete fusion can bypass the β1 requirement. β1 integrin
certainly participates in human gamete fusion by acting in
co-operation with multiple integrin/disintegrin couples or
another cofactor, not yet identified.
Key words: β1 integrin/human gamete fusion
Introduction
The fusion mechanism during mammalian fertilization remains
unclear at the molecular level. It contains two steps that can
be dissociated experimentally: the binding of the spermatozoon
682
to the oocyte plasma membrane and the actual gamete fusion,
i.e. the merging of the sperm and oocyte membranes
(Yanagimachi, 1994).
Integrins are a family of cell surface adhesion receptors
which mediate cell–cell and cell–extracellular matrix interactions. Integrins are heterodimeric glycoproteins composed
of α and β subunits, and their binding to their ligand (disintegrin) on the cell partner can affect various cellular mechanisms,
such as cell differentiation and gene expression (Hynes, 1992;
Miyamoto et al., 1995b; Schwartz et al., 1995). The presence
of integrin and disintegrin on mammalian gamete has already
been assessed.
In oocytes, integrin β1, α3, α5 and α6 subunits were
detected in unfertilized mouse eggs by their mRNA and by
subunit-specific antibodies (Tarone et al., 1993). αv, α5 and
β1 subunits were detected in Syrian hamster eggs using
detergent extracts by enzyme-linked immunoabsorbent assay
(ELISA) and α2, α5, α6 and αv were shown to be expressed
in human and hamster oocytes using immunobeads (Fusi et al.,
1993). In the mouse, β1, α2, α5 and αv integrin subunits are
present in the oocyte. Oocyte maturation was associated with
a redistribution of α5 and vitronectin receptor (αvβ3), as shown
by immunoprecipitation and confocal fluorescence microscopy
(Evans et al., 1995a). In a recent study, integrin subunits α2,
α5, α6 and α2, α4, α5 and α6 were revealed in the plasma
membrane of human and hamster oocytes respectively, by
indirect immunofluorescence labelling (De Nadai et al., 1995),
but the β1 subunit was not found in human or in hamster
oocytes, in this study.
In spermatozoa, a complex of sperm surface antigens, fertilin
αβ, was described in the guinea pig spermatozoa. The α subunit
shares homology with certain viral fusion glycoproteins, and
fertilin β subunit shares high sequence similarity with a family
of soluble disintegrins found in snake venom, especially for
an RGD-like motif (Primakoff et al., 1987; Blobel et al., 1992;
Wolfsberg et al., 1995).
The role of integrins in sperm–egg interaction remains
elusive. Penetration of human or hamster spermatozoa in zonafree golden hamster eggs was completely inhibited by RGDcontaining peptides (Bronson and Fusi, 1990). In the zonafree hamster egg penetration assay, human sperm binding to
the oolemma was inhibited by echistatin, a disintegrin known
to block the binding of fibronectin and vitronectin to their
respective integrin receptors (α5β1, αvβ3). However in the
same conditions, oocyte–sperm fusion was not inhibited
(Bronson et al., 1995). In the mouse egg, integrin α6β1
functions as a sperm receptor for binding (Almeida et al.,
1995). Taken together, these morphological and functional
results suggest that in the sperm–egg binding step, an integrin–
© European Society for Human Reproduction and Embryology
Human fertilization bypasses β1 integrin requirement
disintegrin couple is involved. However, additional cofactors
may be required for the gamete fusion step, as recently
described for other types of fusion mechanism (Feng et al.,
1996).
This study aimed to clarify the role of β1 integrin in
the human gamete fusion process. We first investigated the
expression of β1 integrin subunit at the human oocyte plasma
membrane and in the cytoplasm during oocyte nuclear maturation using immunolabelling and confocal fluorescence microscopy. We then studied the functional role of β1 integrin in
human gamete membrane fusion step using a real time fusion
assay. Finally, the localization of fused spermatozoa was
compared to that of β1 integrin at the human oocyte plasma
membrane.
Materials and methods
Human oocytes
Human oocytes were collected as previously described (Ji et al.,
1997). Briefly, unfertilized metaphase II human oocytes (MII) were
collected 48 h after insemination. Germinal vesicle stage (GV) and
metaphase I stage (MI) oocytes, unsuitable for intracytoplasmic sperm
injection, were donated by patients. According to the National Ethical
Committee recommendations, they were asked to sign an informed
consent form. Zona pellucida was mechanically removed to prevent
any membrane alteration by enzymatic or chemical treatment (Kellom
et al., 1992).
Functional fusion assay
Zona-free GV, MI and MII stage oocytes were routinely loaded with
Hoechst 33342 (Sigma, St Louis, MI, USA) at 5 µg/ml for 5 min,
and inseminated with 5000 motile spermatozoa in 30 µl of Ménézo
B2 medium (CCD) at 37°C under equilibrated oil (BDH, Nogent sur
Marne, France). After 18 h, oocytes were washed in Ménézo B2
medium, and fusion was evaluated by Hoechst transfer from the oocyte
cytoplasm to spermatozoa chromatin detected by epifluorescence
microscopy under UV illumination (Almeida et al., 1995; Conover
and Gwatkin, 1988). Alternatively, fusion was evaluated by calceinAM (Molecular Probes, Eugene, OR, USA). In this case, spermatozoa
were preloaded with calcein-AM which accumulated in the sperm
cytoplasm. Calcein-AM was used at 30 ng/ml, a concentration 100
times lower than the one usually used (Weston and Parish, 1992).
Fusion extent was estimated by redistribution of spermatozoa fluorescence to the oocyte cytoplasm. In some cases, blocking mouse antihuman β1 integrin monoclonal antibody (DE9, UBI, Lake Placide,
NY, USA) (Bergelson et al., 1992), non-blocking mouse anti-human
β1 integrin monoclonal antibody (LM534, Chemicon International,
Temecula, CA, USA) (Giancotti and Ruoslahti, 1990), or immunoglobulin G (IgG) control (Sigma) were added as indicated either in the
insemination medium, being present during the entire length of the
assay, or only with the oocyte in a pre-incubation step for 1 h
followed by three washes. In this latter case, free monoclonal
antibodies were absent from the gamete interaction step. To investigate
the possible role of RGD-containing peptide in human gamete fusion,
GV, MI and MII zona-free oocytes were inseminated with RGDcontaining peptide (GRGDTP, Sigma), or RGL-containing peptide
(GRGLSLSR, Sigma) as control. When indicated, DE9 and GRGDTP
were added together during the fusion assay.
Fluorescence confocal microscopy
Human oocytes at GV, MI and MII stages were fixed immediately
after mechanical zona removal by 2% paraformaldehyde in PHP
medium for 20 min. PHP medium (pH 7.4) was composed of
phosphate-buffered saline, Ca21 and Mg21, buffered with 25 mM
HEPES (Sigma) and supplemented with 1% polyvinylpyrrolidone
(Sigma). Free aldehydes were neutralized by incubation in PHP
medium supplemented with NH4Cl 75 mM and glycine 20 mM for
15 min. Surface labelling was performed by DE9 in PHP medium
supplemented with 2% BSA (Sigma) for 30 min, and revealed with
biotinylated goat anti-mouse IgG (Vector Laboratories, Burlingame,
CA, USA) and streptavidin–fluorescein isothiocyanate (FITC) (Pierce
Chemical Co., Rockford, IL, USA). Control oocytes were labelled
only with biotinylated anti-mouse IgG and streptavidin–FITC. Oocytes
were mounted on slides in moviol containing DABCO 10 mg/ml
(Bomsel et al., 1989) and analysed with a Bio Rad 1000 fluorescence
confocal microscope equipped with an argon/krypton and a UV laser
for Hoechst staining. Consecutive sections were separated by 0.7 µm.
For cytoplasmic β1 integrin detection, all incubation medium was
supplemented with 0.02% saponine (Sigma) as permeabilization agent
(Bomsel, 1997).
Co-localization of β1 integrin and fused spermatozoa
Unfertilized MII or GV zona-free human oocytes were loaded with
Hoechst 33342 and inseminated as described above. Seventeen hours
later, oocytes with fused spermatozoa were surface labelled after
fixation as above and analysed by fluorescence confocal microscopy
using UV and FITC excitation.
Preparation of oocyte enriched plasma membrane fraction and
production of antiserum
Zona-free human oocytes were lysed by repeated cycles of freeze–
thawing from –80 to 4°C. Plasma membrane enriched fractions were
prepared by ultracentrifugation (Ultracentrifuge TLA103, Beckman,
Palo Alto, CA, USA) at 50 000 g, 4°C for 15 min (Bomsel et al.,
1988). The pellet was layered on a nitrocellulose membrane (Bio
Rad, Ivry sur Seine, France) in three equal dot–blots of 232
mm2 each. The nitrocellulose membrane was stored at –20°C until
immunization. A control preparation with biotinylated membranes
detected by streptavidin–HRP and chemoluminescence showed that
the membranes were actually collected in the pellet. The generation
of antiserum was performed in BALB/C mouse as described by
Bazin, 1983. Briefly, a dot–blot nitrocellulose was solubilized, and
injected i.p. in each mouse. The animals were killed after three
injections, 2 weeks apart, and the sera were collected and analysed
in the morphological and functional tests described above. Preimmune serum was also collected and used as control in these assays.
Since the pre-immune serum contained irrelevant mouse IgGs, it was
also used as control for potential binding of mouse IgG to Fcγ
receptors on the oolemma.
Results
Expression of β1 integrin on human oocytes at various stage
of maturation
Indirect immunolabelling using DE9, a mAb anti-β1 integrin
known to inhibit its adhesion functions, and analysis by
confocal microscopy revealed the presence of β1 integrin at
the human oocyte plasma membrane for all GV, MI and MII
stage oocytes (n 5 6/stage). However, striking variations both
in β1 integrin quantity and surface pattern were clearly visible
during the maturation process (Figure 1A: a, b). Indeed, plasma
membrane β1 integrin staining was intense and uniform at the
GV stage oocyte (Figure 1A), whereas it almost disappeared
683
Y.Z.Ji et al.
Figure 1. (A) Confocal microscopy localization of human oocyte plasma membrane and cytoplasm β1 integrin during maturation. Zona-free
human oocytes at germinal vesicle (GV), metaphase I (MI) and metaphase II (MII) stage were surface labelled (a, b) for β1 integrin subunit
by indirect immunostaining. For β1 integrin intracellular detection (c, d), saponin was added during the labelling procedure. (a, c) Single
focal plan; (b, d) projection of consecutive optical sections of half oocyte, 0.7 µm apart. (B) Co-localization of fused spermatozoa with β1
integrin subunit at the human oocyte plasma membrane. After insemination, human MII oocytes with fused spermatozoa were surface
labelled for β1 integrin as in (A) (a, c). (a) plasma membrane β1 integrin subunit detected under fluorescein isothiocyanate excitation;
(b) fused spermatozoa detected under UV excitation.
at the MI stage (Figure 1A). At the MII stage, surface β1
integrin staining was again detected, but it was asymmetrically
distributed in patches (Figure 1A: MIIa, MIIb). To investigate
further the modifications of surface β1 integrin distribution at
the various oocyte stages, we next analysed the β1 integrin
intracellular pool, using a permeabilization agent during the
labelling procedure. When 0.02% saponin was added to all
incubation media, β1 integrin was detected not only at the
plasma membrane, but also in the cytoplasm of human oocytes
at all three stages (n 5 8/stage). As expected, plasma membrane
β1 integrin distribution remained identical as described above
in the absence of saponin. More importantly, in MI and
MII stages, β1 integrin did not accumulate intracellularly
(Figure 1A). Control oocytes, incubated only with biotinylated
anti-mouse IgG and streptavidin–FITC, showed very weak
signal corresponding to autofluorescence (data not shown).
Furthermore, this striking β1 integrin pattern was not due to
mouse IgG binding to the Fcγ receptor on the oolemma, since
labelling with irrelevant mouse IgG, performed under similar
conditions, gave no detectable signal above the autofluorescence background.
684
Human oocyte fusibility during maturation
To investigate in more detail the molecules involved in human
gamete fusion, we set up a functional fusion assay. This assay
was based on the transfer of Hoechst dye from oocyte to
spermatozoa DNA (Conover and Gwatkin, 1988). The extent
of fusion was also monitored with another marker calcein-AM
at 30 ng/ml that was preloaded into the spermatozoa. Fusion
evaluation as a result of calcein transfer from spermatozoa to
oocyte gave results similar to Hoechst (Weston and Parish,
1992). Sperm motility was not impaired at this concentration.
Zona-free human GV, MI and MII stage oocytes (n 5 12/stage)
were inseminated with human spermatozoa. Spermatozoa fused
with human oocytes at all maturational stages, but with
increasing efficiency from 11.4 6 3.4 and 14.7 6 2.6 to
21.8 6 4.5 per oocyte for GV, MI and MII stages, respectively
(see control, Figure 2).
Co-localization of β1 integrin and fused spermatozoa
To evaluate the possible involvement of β1 integrin in human
gamete interaction, we monitored the respective distribution
of fused spermatozoa and β1 integrin patches at the oocyte
Human fertilization bypasses β1 integrin requirement
Figure 2. Inhibition of human sperm–oocyte fusion by RGDcontaining peptide during oocyte maturation. Zona-free germinal
vesicle (GV), metaphase I (MI) and metaphase II (MII) stage
human oocytes were loaded with Hoechst 33342, and inseminated
with human spermatozoa (control), or in the presence of GRGDTP
(RGD), or GRGLSLSR (RGL) at 30 µM. Fusion was detected by
Hoechst transfer from the oocyte cytoplasm to spermatozoa
chromatin evaluated by fluorescence microscopy.
plasma membrane. After insemination, MII oocytes (n 5 10)
with fused spermatozoa were surface labelled with DE9 as
indicated above. By fluorescence confocal microscopy, plasma
membrane β1 integrin distributed into patches (Figure 1B: a).
In the same oocyte, fused spermatozoa were asymmetrically
distributed (Figure 1B: b). Strikingly, fused spermatozoa colocalized with plasma membrane β1 integrin patches, as
indicated by arrow heads, and definitely were not observed in
the patchless regions of the oocyte (Figure 1B). When fusion
was assessed on GV oocytes, similar analysis showed that the
fusion event did not modify the pattern of oocyte plasma
membrane β1 integrin observed prior to fusion (data not
shown).
Partial inhibition of human gamete fusion by DE9 and
GRGDTP
These morphological data indicated a potential role for β1
integrin in the actual gamete fusion process. To further investigate the requirement of β1 integrin in our fusion assay, we
used various approaches to inhibit integrin’s function, i.e. DE9
or RGD peptides. Forty-eight hour old unfertilized MII human
oocytes were inseminated as previously described. In the
presence of DE9 (n 5 6–9/concentration), the fusion was
inhibited in a concentration-dependent manner (Figure 3). This
inhibition increased linearly from 5 to 25 µg/ml. It reached a
plateau at 25 µg/ml with 13.2 6 2.9 fused spermatozoa. A
further increase in antibody concentration did not result in a
stronger inhibition of sperm–egg fusion. This inhibition was
specific, since LM534, a non-blocking mouse anti-β1 integrin
monoclonal antibody, as well as mouse IgG control antibody,
had no effect on sperm fusion. In addition, to determine
whether this inhibition resulted from the functional block of
β1-integrin at the oocyte or at the sperm cell surface, two
complementary protocols were run. In the first, gametes were
Figure 3. Partial inhibition of human sperm–oocyte fusion by
blocking anti-human β1 integrin monoclonal antibody (DE9). Zonafree human metaphase II (MII) oocytes were loaded with Hoechst
33342, and inseminated with human spermatozoa in the presence of
DE9 at increasing concentrations. Non-blocking anti-human β1
integrin monoclonal antibody (LM534) and mouse immunoglobulin
G (IgG) control were used as control antibody at the same
concentrations. Fusion was detected as in Figure 2.
Figure 4. Stronger inhibition of human sperm–oocyte fusion by
combination of GRGDTP (RGD) and blocking anti-human β1
integrin monoclonal antibody (DE9). Zona-free human metaphase II
(MII) oocytes were loaded with Hoechst 33342, and inseminated
with human spermatozoa in the presence of the control peptide,
RGL (30 µM): RGL, RGD (30 µM): RGD and DE9 (25 µg/ml):
DE9 or both: RGD 1 DE9. Alternatively, oocytes were preincubated with DE9 (25 µg/ml) and washed out from the excess of
unbound mAb before insemination with sperm: DE9 oocyte. Fusion
was detected as in Figure 2. For abbreviations, see Figures 2 and 3.
co-incubated with the anti-β1-integrin mAb and remained
present during the entire length of the assay, while in the
second, oocytes only were pre-incubated with the β1-integrin
antibody, washed free from unbound antibodies and then
inseminated with spermatozoa, leaving the incubation medium
free of soluble antibodies. As shown in Figure 4, both experiments did not give significantly different results. One can
therefore conclude that inhibition of sperm cell β1-integrin
was not involved in the fusion inhibition process seen here.
Human gamete fusion was not completely inhibited by DE9.
This may be due to a lack of total functional blocking efficiency
of DE9. From the literature, the additional involvement of
other integrins in the gamete fusion process could be postulated.
Many disintegrins contain RGD-like motifs involved in their
685
Y.Z.Ji et al.
binding to integrin, and RGD-containing peptides have been
shown to inhibit integrin/disintegrin interaction (Bronson and
Fusi, 1990). We therefore investigated the possible role of the
RGD motif in human gamete fusion by adding RGD containing
peptides to the fusion assay. Fusibility at all maturational GV,
MI and MII stage oocytes was tested. Whatever the stage,
sperm fusion to zona-free human oocytes was inhibited in a
concentration-dependent manner by GRGDTP from 2 to
30 µM. Complete inhibition was not obtained, even at 120 µM.
At 30 µM GRGDTP peptide, only 5.4 6 1.9, 8.8 6 2.5 and
11.1 6 3.3 spermatozoa remained fused per oocyte for GV,
MI and MII stage (n 5 6 to 12/stage), respectively (see RGD,
Figure 2). In the same conditions, an RGL-containing peptide,
i.e. GRGLSLSR, did not significantly modify the fusion at all
stages (see RGL, Figure 2).
Neither DE9 nor GRGDTP could completely inhibit human
gamete fusion. To increase the inhibition efficiency, DE9 and
GRGDTP were added together in fusion assay (n 5 6–12).
The inhibition reached a plateau at 25 µg/ml of DE9 and
30 µM of GRGDTP, but the inhibition was still not complete,
i.e. 6.1 6 2.4 spermatozoa remained fused per oocyte
(Figure 4).
Inhibition of gamete fusion with polyclonal antibody raised
against the plasma membrane proteins of human oocyte
Since interfering with integrin binding function could not
inhibit gamete fusion, we postulated that other plasma membrane proteins could be involved in this process. To get more
insights on this(ese) unknown cofactor(s), mouse polyclonal
antibodies were raised against partially purified human oocyte
plasma membranes. Indirect immunolabelling and confocal
microscope analysis showed that the mouse antiserum recognized specifically the human oocyte plasma membrane (data
not shown) and stained the oocyte in a similar asymmetrical
patchy distribution as already observed with the anti-β1 antibody. In contrast, control oocytes, labelled with non-immune
serum as primary antibody, or with only biotinylated antimouse IgG and streptavidin–FITC, revealed a very weak signal
corresponding to autofluorescence (data not shown). The lack
of specific binding with pre-immune serum demonstrated that
the detection of the β1 integrin patches did not rely on specific
Fcγ binding to the Fcγ receptor. When used in the functional
fusion test, the antiserum inhibited human gamete fusion in a
concentration-dependent manner (from 1/80 to 1/5), reaching
95% of inhibition, higher than that obtained with RGD and
mAb anti-β1. However, complete inhibition was never obtained
even with the dilution of 1:5 (Figure 5).
Discussion
In this work, we describe the presence of β1 integrin on the
human oocyte plasma membrane during the pre-ovulatory
maturation. The fusibility of the maturing human oocyte was
evaluated in a functional fusion test and compared to β1 integrin
expression. Furthermore, inhibition experiments induced by
various antibodies and peptides that block β1 integrin binding
activity allowed us to evaluate the β1 integrin involvement in
the human gamete interaction process.
686
Figure 5. Inhibition of human sperm–oocyte fusion by polyclonal
anti-human plasma membrane mouse antisera. Zona-free human
metaphase II (MII) oocytes were loaded with Hoechst 33342, and
inseminated with human spermatozoa in the presence of polyclonal
anti-human plasma membrane mouse antisera at increasing
concentrations. Fusion was detected as in Figure 2.
In a preliminary morphological experiment designed to
evaluate β1 integrin expression on the human oocyte plasma
membrane, β1 integrin could not be seen by a simple indirect
immunodetection technique using a primary anti-β1 antibody
and a secondary FITC-labelled anti-mouse IgG. In contrast,
when the detection was amplified by a biotinylated anti-mouse
IgG and avidin-FITC, β1 integrin was repeatedly detected
irrespective of the sequence of oocyte fixation and labelling.
The differences in labelling protocols could explain why De
Nadai et al. (1995), who did not include an amplification step,
were unable to detect β1 integrins either on hamster or on
human oocytes.
The second requirement of this study was to establish a
reliable functional fusion assay. We choose the technique of
Hoechst dye transfer from oocyte cytoplasm to sperm chromatin, detected by epifluorescence microscopy (Conover and
Gwatkin, 1988) and Almeida et al. (1995). This technique was
validated using another fusion reporter, i.e. by calcein-AM
transfer from spermatozoa to oocyte (Weston and Parish,
1992). Once loaded into the sperm cytoplasm, calcein was
trapped by esterification and could not diffuse in the oocyte
unless fusion had occurred. The same results were obtained
using either Hoechst or calcein-AM dye (data not shown). We
therefore used routinely the Hoechst transfer technique.
Unfertilized 2 day old and fresh oocytes were used for this
research. After depellucidation by a mechanical procedure,
both egg types retained a normal morphology, and kept their
fusiogenic property as previously shown (Ji et al., 1997).
Experiments performed with 2 day old unfertilized oocytes or
fresh ones provided the same results.
To investigate the involvement of β1 integrin in human
gamete fusion, we first characterized its expression in human
oocytes at various maturational stages. Intense β1 integrin
staining was seen at the surface of germinal vesicle stage
oocytes and in subcortical vesicles after membrane permeabilization. β1 integrin labelling disappeared almost completely
Human fertilization bypasses β1 integrin requirement
from MI stage (membrane and cytoplasm) but was again
detected in MII stage oocytes as asymmetrical patches at the
cell surface. The intracellular β1 integrin pool in MI stage
oocyte was negligible, indicating that disappearance of β1
integrin from the oocyte membrane of MI oocyte did not result
from an internalization process. It is more likely that β1
integrin was degraded from GV to MI stage oocytes, and that
its re-expression at the MII stage oocyte resulted from denovo synthesis. This type of modulation in the expression of
β1 integrin during development has also been reported in other
systems (Delcommenne and Streuli, 1995). The clustering of
β1 integrin detected at the MII stage was not an artifact
induced by the labelling procedure, since such patches were
not detected in identically treated GV and MI oocytes. It also
did not result from DE9 binding to the oolemma Fcγ receptor
since irrelevant mouse IgG used as primary antibody gave no
detectable signal. Furthermore, in these experiments, staining
was always performed after oocyte fixation, in conditions that
prevent protein redistribution (Mayor et al., 1993). Similar
asymmetrical distribution and change during oogenesis have
been observed for rat oocyte binding sites of sperm protein
DE. Although absent from GV stage oocytes with a diameter
smaller than 50 µm, it was expressed homogeneously at the
surface of larger GV oocytes and during maturation and growth
its localization on the oolemma changed progressively to
a patchy distribution (Cohen et al., 1996). However, the
consequences of the fluctuation in β1 integrin expression
during oocyte maturation remain to be elucidated. In addition
to gamete interaction, β1 integrins may be involved in functions
such as the corona cell–oocyte binding mechanism in GV
stage oocytes.
The patchy β1 integrin organization in human oocyte membrane parallels an increase in fusibility of the growing oocyte.
Indeed, the number of spermatozoa able to fuse per oocyte
increased during preovulatory meiosis. This is consistent with
the greater fusibility of human oocytes with a diameter
.110 µm as compared to smaller ones, as assessed after
subzonal sperm insemination (Wolf et al., 1995). This is also
consistent with the results of Zuccotti et al. (1995) for hamster
oocytes, which become fusion competent when their diameter
reaches ~20 µm, with a fusibility increasing with growth and
reaching a maximum at the metaphase of the second meiosis
division. The increased fusibility of growing oocytes may be
correlated either to the patchy organization of β1 or to integrin
conformational changes that could be regulated by some earlier
cellular signalling process such as calcium influx (Schwartz,
1993). The patchy organization of β1 integrin in MII oocytes
may reflect activation of the integrin, these clusters being
induced and stabilized by the cytoskeleton (Schwartz et al.,
1995). Indeed, surface signal transduction cascades are often
regulated by receptor oligomerization (Heldin, 1995), and
clustering at the cell membrane appears to be a general
activation mechanism (Kornberg et al., 1992; Miyamoto
et al., 1995a).
RGD peptides promote the recruitment of one integrin
(α5β1) to focal contacts induced by another integrin (α2β1)
on a substrate to which the first integrin did not bind (LaFlamme
et al., 1992). Both integrins, α2β1 and α5β1, acted in synergy
by cross talk between the two receptors of different families.
In this regard, fertilization in 23 undamaged MII eggs after
human sperm subzonal insemination into the perivitelline space
of human oocytes was prevented by cytochalasin D, suggesting
that the integrity of the actin network is mandatory for gamete
fusion (Ng et al., 1989).
Since β1 integrins are connected to the actin cytoskeleton,
one may suggest that the spermatozoa may bind to β1 integrin
and interfere with subcortical actin, a step required in the
fusion process. This hypothesis may explain why the fused
spermatozoa and β1 integrin patches overlap. The absence of
fused spermatozoa in the patchless oocyte regions strongly
suggests a preferential interaction between spermatozoa and
β1 integrin patches on the oocyte plasma membrane. But
strikingly, the human oocyte fusibility was not correlated to
its β1 integrin total content, suggesting that most detected
integrins were not required for the actual fusion process.
Furthermore, the β1 integrin patchy organization may not be
required for spermatozoa fusion with GV and MI stage oocytes.
However, our data do not rule out the existence of β1 integrin
micropatches below the detection limit, that could be sufficient
for gamete fusion.
Inhibition of integrin binding properties with mAb, RGD
peptide, or both, did not result in full inhibition of human
gamete fusion per se. Although the immune serum of mouse
was more potent in inhibiting the gamete fusion, some spermatozoa were still able to fuse. Similar results were reported in
the human–hamster system (Bronson et al., 1995), in the mouse
(Almeida et al., 1995), and for mouse spermatozoa fertilin, an
integrin ligand previously called PH30. PH30 β chain contains
the tripeptide QDE as an integrin ligand domain instead of
RGD. Peptides containing RGD or QDE sequences also
decreased the binding and fusion of spermatozoa with zonafree oocyte in the mouse but only by ~70% (Evans et al., 1995b;
Snell and White, 1996). Furthermore, and more importantly, in
the normal fertilization process, only one spermatozoon has
to penetrate the oocyte. Consequently, the presence of blocking
β1 antibody or RGD-containing peptides would probably not
prevent fertilization in vivo. It has been suggested in the mouse
that the inhibition of the fusion process by anti-integrin
antibodies was linked to the decrease of the number of bound
spermatozoa (Almeida et al., 1995). However, in contrast to
the mouse system, in the human very few spermatozoa (one
to five) bind to the oolemma during the interaction process. It
is therefore difficult to assess whether both inhibition of the
fusion and sperm binding to the oolemma are linked in the
human or not.
From our results, it appears that the fusion process can
bypass the requirement of the β1 integrin–disintegrin system.
Furthermore, if null mutations in β1 integrin genes produce
early embryonic lethality (Fässler and Meyer, 1995; Stephens
et al., 1995), they do not inhibit fertilization in mice. But it
cannot be excluded that β1 integrin mRNA may have been
provided as a maternal message to the null oocytes (Sutherland
et al., 1993). Most integrin ligands can interact with multiple
integrins and most integrins can bind more than one ligand,
illustrating a potential overlap in their function (Miyamoto
et al., 1996). Thus β1 integrins may be essential for gamete
687
Y.Z.Ji et al.
fusion, but should it be inhibited by a blocking antibody, other
redundant proteins may play their role, and take over its
function (Daneu et al., 1995; Ruoslahti, 1996).
A remaining unsolved issue is the specificity of sperm–egg
fusion. Spermatozoa do not bind or fuse to somatic cells
despite β1 integrin expression on a wide variety of cell types.
This specific sperm–egg interaction may rely on the regulation
of the binding domain of integrins, but could also involve an
interaction with additional protein or lipid factor. This is the
case in other fusion mechanisms such as HIV-infected cell
fusion with its CD41 target cell. Indeed, in addition to the
HIV envelope principal receptor CD4, a signalling molecule,
i.e. a 7TM chemokine receptor, acts as co-receptor to allow
fusion to occur (Feng et al., 1996). Accordingly, in our
experiment, an antiserum raised against partially purified
oocyte plasma membrane can block gamete fusion more
efficiently than the combination of GRGDTP and DE9
(Figure 4).
Taken together, the data suggest that β1 integrins are
involved in human gamete interaction but that a cofactor is
required for the gamete specificity and the fusion process. A
likely sperm receptor candidate at the oocyte plasma membrane
could be the 71 kDa surface protein we recently described in
the human oocyte. During its nuclear maturation, the 71 kDa
protein is over-expressed (Ji et al., 1997), accounting for 30%
of the total membrane protein content of the mature metaphase
II oocytes. Despite the additional data present here, the β1
integrin on the oocyte membrane during its nuclear maturation
remains still to be explained. Raising monoclonal antibodies
against the human oocyte membrane proteins may be an
efficient strategy in characterizing the oolemma co-receptor to
the spermatozoon.
Acknowledgements
We are grateful to Isabelle Bouchaert for her technical assistance in
the confocal microscopy. We thank all the colleagues of IVF Unit of
Cochin Hospital. Some of the data from this study were presented at
the 12th Annual Meeting of ESHRE (European Society of Human
Reproduction and Embryology). This research was partly supported
by grant 94229 from Direction de la Recherche Clinique, Assistance
Publique, by grant EA 1752 from DRED, and by the Fondation pour
la Recherche Médicale.
References
Almeida, E.A.C., Huovila, A.P.J., Sutherland, A.E. et al. (1995) Mouse egg
integrin α6β1 functions as a sperm receptor. Cell, 81, 1095–1104.
Bazin, H. (1983) Production of rat monoclonal antibodies with the LOU rat
non-secreting IR983F F myeloma cell line. In Peeters, H. (ed.), Protides of
the Biological Fluids. Vol. 29th Colloquium. Pergamon Press, Oxford and
New York, pp. 615–618.
Bergelson, J.M., Shepley, M.P., Chan, B.M.C. et al. (1992) Identification of
the integrin VLA-2 as a receptor for Echovirus 1. Science, 255, 1718–1720.
Blobel, C.P., Wolfsberg, T.G., Turck, C.W. et al. (1992) A potential fusion
peptide and an integrin ligand domain in a protein active in sperm–egg
fusion. Nature, 356, 248–252.
Bomsel, M. (1997) Transcytosis of infections human immunodeficiency virus
across a tight human epithelial cell line barrier. Nature Med., 3, 42–47.
Bomsel, M., De Paillerets, C., Weintraub, H. and Alfsen, A. (1988) Biochemical
and functional characterisation of three types of coated vesicles in bovine
adrenocortical cells. Implication in the intracellular traffic. Biochemistry,
27, 6806–6811.
688
Bomsel, M., Prydz, K., Patron, R.G. et al. (1989) Endocytosis in filter-grown
Madin–Darby canine kidney cells. J. Cell Biol., 109, 3243–3258.
Bronson, R.A. and Fusi, F. (1990) Evidence that an Arg-Gly-Asp adhesion
sequence plays a role in mammalian fertilization. Biol. Reprod., 43,
1019–1025.
Bronson, R.A., Gailit, G.J., Bronson, S. and Oula, L. (1995) Echistatin, a
disintegrin, inhibits sperm-oolemmal adhesion but not oocyte penetration.
Fertil. Steril., 64, 414–420.
Cohen, D.J., Munuce, M.J. and Cuasnicu, P.S. (1996) Mammalian sperm–egg
fusion: the development of rat oolemma fusibility during oogenesis involves
the appearance of binding sites for sperm protein ‘DE’. Biol. Reprod., 55,
200–206.
Conover, J.C. and Gwatkin, R.B.L. (1988) Pre-loading of mouse oocytes with
DNA-specific fluorochrome (Hoechst 33342) permits rapid detection of
sperm–oocyte fusion. J. Reprod. Fertil., 82, 681–690.
Daneu, E.H.,S. Aota, A.A. Van Kraats, K.M. et al. (1995) Requirement for
the synergy site for cell adhesion to fibronectin depends on the activation
state of integrin α5β1. J. Biol. Chem., 270, 21612–21618.
De Nadai, C., Fenichel, P., Donzeau, M. et al. (1995) Characterization and
role of integrins during gametic interaction and egg activation. Zygote, 4,
31–40.
Delcommenne, M. and Streuli, C.H. (1995) Control of integrin expression by
extracellular matrix. J. Biol. Chem., 270, 26794–26801.
Evans, J.P., Schultz, R.M. and Kopf G.S. (1995a) Identification and localization
of integrin subunits in oocytes and eggs of the mouse. Mol. Reprod. Dev.,
40, 211–220.
Evans, J.P., Schultz, R.M. and Kopf, R.S. (1995b) Mouse sperm–egg plasma
membrane interactions: analysis of roles of egg integrins and the mouse
sperm homologue of PH-30 (fertilin) beta. J. Cell Sci., 108, 3267–3278.
Fässler, R. and Meyer, M. (1995) Consequences of lack of β1 integrin gene
expression in mice. Genes Dev., 9, 1896–1908.
Feng, Y., Broder, C.C., Kennedy, P.E. and Berger, E.A. (1996) HIV-1 entry
cofactor: functional cDNA cloning of a seven-transmembrane, G proteincoupled receptor. Science, 272, 872–877.
Fusi, F.M., Vignali, M., Gailit, J. and Bronson, R.A. (1993) Mammalian
oocytes exhibit specific recognition of the RGD (Arg-Gly-Asp) tripeptide
and express oolemmal integrins. Mol. Reprod. Dev., 36, 212–219.
Giancotti, F.G. and Ruoslahti, E. (1990) Elevated levels of the α5β1 fibronectin
receptor suppress the transformed phenotype of Chinese hamster ovary
cells. Cell, 60, 849–549.
Heldin, C.H. (1995) Dimerization of cell surface receptors in signal
transduction. Cell, 80, 213–223.
Hynes, R.O. (1992) Integrins: versatility, modulation and signaling in cell
adhesion. Cell, 69, 359–362.
Ji, Y.Z., Bomsel, M., Jouannet, P. and Wolf, J.P. (1997) Modifications of the
human oolemma protein pattern during maturation. Mol. Reprod. Dev., 47,
120–126.
Kellom, T., Vick, A. and Boldt, J. (1992) Recovery of penetration ability in
protease-treated zona-free mouse. Mol. Reprod. Dev., 33, 46–52.
Kornberg, L., Earp, H.S., Parsons, J.T. et al. (1992) Cell adhesion or integrin
clustering increases phosphorylation of a focal adhesion-associated tyrosine
kinase. J. Biol. Chem., 267, 23439–23442.
LaFlamme, S.E., Akiyama, S.K. and Yamada, K.M. (1992) Regulation of
fibronectin receptor distribution. J. Cell Biol., 117, 437–447.
Mayor, S., Presley, J.F. and Maxfield, F.R. (1993) Sorting of membrane
components from endosomes and subsequent recycling to the cell surface
occurs by bulk flow process. J. Cell Biol., 121, 1257–1269.
Miyamoto, S., Akiyama, S.K. and Yamada, K.M. (1995a) Synergistic roles
for receptor occupancy and aggregation in integrin transmembrane function.
Science, 267, 883–885.
Miyamoto, S., Teramoto, H., Coso, O.A. et al. (1995b) Integrin function:
molecular hierarchies of cytoskeletal and signaling molecules. J. Cell Biol.,
131, 791–805.
Miyamoto, S., Teramoto, H., Gutkind, J. S. and Yamada, K. M. (1996)
Integrins can collaborate with growth factors for phosphorylation of receptor
tyrosine kinase and MAP kinase activation: roles of integrin aggregation
and occupancy of receptors. J. Cell Biol., 135, 1633–1642.
Ng, S.C., Bongso, A., Chang, S.I. et al. (1989) Transfer of human sperm into
the perivitelline space of human oocytes after zona-drilling or zona-puncture.
Fertil. Steril., 52, 73–78.
Primakoff, P., Hyatt, H. and Tredick-kiline, J. (1987) Identification and
purification of a sperm surface protein with a potential role in sperm–egg
membrane fusion. J. Cell Biol., 104, 141–149.
Ruoslahti, E. (1996) RGD and other recognition sequences for integrins.
Annu. Rev. Cell Dev. Biol., 12, 697–715.
Human fertilization bypasses β1 integrin requirement
Schwartz, M.A. (1993) Spreading of human endothelial cells on fibronectin
or vitronectin triggers elevation of intracellular free calcium. J. Cell Biol.,
120, 1003–1010.
Schwartz, M.A., Schaller, M.D. and Ginsberg, M.H. (1995) Integrins: emerging
paradigms of signal transduction. Annu. Rev. Cell Dev. Biol., 11, 549–599.
Snell, W. and White, J.M. (1996) The molecules of mammalian fertilization.
Cell, 85, 629–937.
Stephens, L.E., Sutherland, A.E., Klimanskaya, I.V. et al. (1995) Deletion of
β1 integrins in mice results in inner cell mass failure and peri-implantation
lethality. Genes Dev., 9, 1883–1895.
Sutherland, A.E., Calarco, P.G. and Damsky, C.H. (1993) Developmental
regulation of integrin expression at the time of implantation in the mouse
embryo. Development, 119, 1175–1186.
Tarone, G., Russo, M.A., Hirsch, E. et al. (1993) Expression of β1 integrin
complexes on the surface of unfertilized mouse oocyte. Development, 117,
1369–1375.
Weston, S.A. and Parish, C.R. (1992) Calcein: a novel marker for lymphocytes
which enter lymph node. Cytometry, 13, 739–749.
Wolf, J.P., Bulwa, S., Rodrigues, D. and Jouannet, P. (1995) Human oocyte
cytometry and fertilization rate after subzonal insemination. Zygote, 3,
101–109.
Wolfsberg, T.G., Straight, P.D., Gerena, R.L. et al. (1995) ADAM, a widely
distributed and developmentally regulated gene family encoding membrane
proteins with a disintegrin and metalloprotease domain. Dev. Biol., 169,
378–383.
Yanagimachi, R. (1994) Mammalian fertilization. In Knobil, E. and Neill,
J.D. (eds), The Physiology of Reproduction, 2nd edn. Raven Press, New
York, pp. 189–317.
Zuccotti, M., Yanagimachi, R. and Yanagimachi, H. (1991) The ability of
hamster oolemma to fuse with spermatozoa: its acquisition during oogenesis
and loss after fertilization. Development, 112, 143–152.
Received on July 25, 1997; accepted on November 17, 1997
689