Requirement of CD9 on the Egg Plasma Membrane for Fertilization

See allHide authors and affiliations

Science  14 Jan 2000:
Vol. 287, Issue 5451, pp. 321-324
DOI: 10.1126/science.287.5451.321


CD9 is an integral membrane protein associated with integrins and other membrane proteins. Mice lacking CD9 were produced by homologous recombination. Both male and female CD9−/− mice were born healthy and grew normally. However, the litter size from CD9−/− females was less than 2% of that of the wild type. In vitro fertilization experiments indicated that the cause of this infertility was due to the failure of sperm-egg fusion. When sperm were injected into oocytes with assisted microfertilization techniques, however, the fertilized eggs developed to term. These results indicate that CD9 has a crucial role in sperm-egg fusion.

CD9, expressed in a wide variety of cells (1, 2), is an integral membrane protein belonging to a family of tetraspan-membrane proteins (TM4) (3) and is reported to play a role in cell adhesion, cell motility, and tumor cell metastasis (4). CD9, which can associate with integrins such as α3β1 and α6β1, is suggested to be a possible co-factor of the integrins (5, 6). CD9 also associates with the membrane-anchored form of heparin-binding EGF-like growth factor (proHB-EGF) and up-regulates its biological activities (7).

To study the physiological role of CD9 in vivo, we produced CD9−/− mice by targeted disruption of the CD9 gene using embryonic stem (ES) cell technology (8). A part of the third exon and all of the fourth exon, which encode second to third membrane-spanning domains, were deleted in the targeting vector (Fig. 1A). Chimeric males were bred with C57BL/6J females to produce homozygous mice for the deletion. Homozygous mutant mice were identified by Southern blot analyses of genomic DNA (Fig. 1B). Two independent ES cell lines carrying a mutated allele generated chimeric mice transmitting the mutated alleles to the progeny. Both lines of mice had essentially identical results. Breeding yielded the predicted number of null mice at a mendelian frequency. Flow cytometrical analyses of bone marrow cells proved an absence of CD9 expression in the mice carrying the deletion (Fig. 1C). Both male and female homozygous mutant mice were born and grew normally, demonstrating that there were no detrimental effects in cell adhesion or cell growth in CD9−/− mice.

Figure 1

Generation of CD9-deficient mice. (A) Gene-targeting strategy. Exons and introns are represented by vertical bars and horizontal lines, respectively. A neomycin-resistance gene driven by a PGK promoter (NEOr) and a diphtheria toxin A fragment driven by a MC1 promoter (DTA) are indicated as white boxes. Restriction sites: A, Apa I; B, Bgl II; Bm, Bam HI; E, Eco RI; N, Not I; X, Xba I; Xh, Xho I. (B) Hybridization of Eco RI–digested genomic DNA from wild-type (+/+), heterozygous CD9 (+/–), or homozygous CD9 (–/–) mice with the 5′- and 3′-external probes. (C) Flow cytometry of bone marrow cells in CD9+/+, CD9+/–, and CD9−/− littermates stained with fluorescein isothiocyanate–conjugated anti-mouse CD9 antibody (KMC8.8). Data represents monocyte-rich fractions gated by forward scatter and side scatter. (D) Fecundity of CD9+/+, CD9+/–, and CD9−/− females and males was examined by caging one male to one or two females with various genetic backgrounds (9). The average litter size ± SEM is shown. The numbers in parentheses indicate the numbers of mating pairs. Asterisk indicates no litters. (E) Micrograph of oocytes collected from CD9−/− females. First polar bodies (arrowheads) are seen in all of the oocytes.

However, CD9−/− females were infertile, whereas CD9−/− males showed normal fertility when mated with wild-type mice. Mating experiments (9) showed that there was no difference in the frequency of copulation plugs between CD9−/− and CD9+/− females. However, the number of pups born from CD9−/− females was <2% of the rates of wild-type and heterozygote mice (Fig. 1D). Histochemical analysis of the ovaries of CD9−/− females showed no morphological abnormalities, and the follicles of all developmental stages were present, including corpora lutea indicative of past ovulation (10). The females responded to hormone treatment, and they ovulated as many eggs as wild-type females (average of 19 ± 1.9 in CD9−/− and 22 ± 3.0 in CD9+/+mice). First polar bodies were observed in 97% (37 of 38) of oocytes collected from CD9−/− females 12.5 to 13 hours after human chorionic gonadotropin (hCG) injection (Fig. 1E), and the oocyte chromosomes were confirmed to be in metaphase II by the observation of whole-mount specimens (11). Therefore, lack of ovulation or oocyte maturation did not cause the infertility.

We then conducted in vitro fertilization (IVF) experiments to analyze the infertility of CD9−/− female mice (12). Oocytes collected from CD9−/− and CD9+/+ females were inseminated with wild-type sperm. When examined 6 hours after insemination, sperm penetrated the zona pellucida of eggs from both lines. However, almost all of the inseminated CD9−/− eggs (153 of 154) failed to fuse with sperm and to form pronuclei, whereas 57% of the CD9+/+eggs fused with sperm and reached the pronuclear stage (Fig. 2, A and B). In normal conditions, zona-penetrated sperm fuse with the egg plasma membrane and initiate egg activation that includes the release of cortical granule contents leading to modification of the zona pellucida to prevent subsequent sperm penetration (13). However, 65% of the CD9−/− eggs accumulated more than five sperm in their perivitelline space (Fig. 2A).

Figure 2

Phenotype of CD9-deficient eggs in in vitro fertilization experiments. [(A) and (B)] Zona-intact eggs were used. Oocytes, collected from CD9−/− and CD9+/+ females 13 to 15 hours after the injection of hCG, were inseminated with wild-type sperm. (A) Representative micrographs and (B) a summary of the rates of fertilization of CD9+/+ and CD9−/− eggs were obtained 6 hours after insemination. The number of sperm that penetrated the zona pellucida of CD9+/+ or CD9−/− eggs was determined by staining with lacmoid (24). CD9−/− eggs failed to fuse with sperm and remained in metaphase II, whereas CD9+/+ eggs fused with sperm, and pronuclei (arrows) were observed. [(C) and (D)] IVF was performed using zona-free eggs. (C) Similar numbers of sperm bound to the surface of CD9+/+ and CD9−/− eggs following washing 1 hour after insemination. In (A) and (C), magnification (×400) shows live eggs under phase contrast view. (D) Fusing ability of CD9+/+ and CD9−/− eggs with wild-type sperm. The number of sperm fused with CD9+/+ or CD9−/− eggs was determined 1 and 6 hours after insemination, respectively.

To define whether CD9−/− eggs have the ability to bind sperm at the plasma membrane, IVF was carried out using eggs that were freed from the zonae pellucidae (12). In this condition, many sperm can directly interact with the egg surface. No difference was seen in the number of sperm that bound to the surface of zona-free CD9−/− and CD9+/+ eggs (Fig. 2C). In contrast, sperm-egg fusion (14) was greatly impaired in the CD9−/− eggs: the fusion rate dropped from 98 to 4% and from 100 to 21%, when judged 1 and 6 hours after insemination, respectively (Fig. 2D). Thus, CD9 functions in sperm-egg fusion (15).

The rare success of fertilization in vitro (Fig. 2) and the appearance of healthy offspring in vivo (Fig. 1D) suggest that the deficiency of CD9 affects only sperm-egg interaction but not the developmental processes that follow. To test this, sperm were injected directly into the cytoplasm of CD9−/− oocytes [intracytoplasmic sperm injection (ICSI)] (16). CD9−/−oocytes injected with wild-type sperm showed normal implantation efficiency in the uteri of pseudopregnant females, and the resulting embryos developed to term with rates similar to those of wild-type mice (Table 1).

Table 1

Development of CD9−/− oocytes after direct injection of sperm into egg cytoplasm (ICSI) and embryo transfer. Results are from two replicate experiments.

View this table:

Protein immunoblot analysis with monoclonal antibody (mAb) to the CD9 antigen indicated that CD9 is expressed in the oocyte membrane, and the CD9 has a molecular size similar to that of CD9 in F9 mouse embryonal carcinoma cells (Fig. 3A). Immunofluorescent staining showed the localization of CD9 on the oocyte plasma membrane (Fig. 3B). When anti-CD9 mAb was added to the in vitro fertilization system, the fertilization of wild-type eggs was inhibited (Fig. 3C). Unlike antibodies to integrin α6 subunit (17), antibodies to CD9 did not block sperm binding to the egg plasma membrane. Although the inhibition of sperm-egg fusion was not as complete as the results seen in CD9−/− eggs, the antibodies to CD9 showed the inhibition of sperm-egg fusion, and as a result, an accumulation of many sperm inside the perivitelline space was observed. These results reinforce the involvement of CD9 in sperm-egg fusion.

Figure 3

Immunological studies of egg CD9 and the role of CD9 in fertilization. (A) Protein immunoblotting with an anti-CD9 mAb (KMC8.8). We lysed 37 eggs and 2 × 103 embryonic carcinoma F9 cells with 10 mM CHAPS solution (5) and loaded them on an SDS gel. (B) Immunofluorescence micrographs of CD9 on the egg plasma membranes. Zona-intact mouse eggs collected from CD9+/+ and CD9−/− females were stained with anti-mouse CD9 mAb and the fluoresceinated second antibody. Original magnification, ×200. (C) Effect of anti-CD9 antibody on IVF of zona-intact eggs. Anti-CD9 mAb (KMC8.8) or control mAb (anti–Thy-1; PharMingen) was added at the concentrations indicated 30 min before insemination. The number of eggs at the two-cell stage were counted for assessment of fertilization 24 hours after insemination. (D) Co-immunoprecipitation of CD9 with integrin α6β1 from eggs and F9 cell lysates. We lysed 106 eggs or 3 × 105 F9 cells with 10 mM CHAPS solution, and then immunoprecipitated the lysates with anti-integrin α6β1 mAb (GoH3), anti-CD9 mAb, or control mAb (25). Immunoprecipitated materials were electrophoresed in SDS-gels under nonreducing conditions and immunoblotted by anti-CD9 antibodies.

It was postulated that the sperm surface protein fertilin, a member of the ADAM family, functions in sperm-egg fusion (13, 18). However, fertilin-β knockout mice showed that fertilin functions in sperm-egg binding (19). Integrin α6β1 has been noted as a sperm receptor on egg plasma membrane (17, 20) and binds to fertilin through fertilin's distintegrin domain. We examined whether CD9 forms a complex with integrin α6β1 on egg plasma membrane by co-immunoprecipitation assay. CD9 co-precipitated with anti-integrin α6 antibody in lysates from both mouse eggs and F9 cells (Fig. 3D), indicating that CD9 physically associates with integrin α6β1 on egg plasma membrane, as shown in other cell lines (7).

The integrin family provides a physical link between the extracellular matrix and the cell cytoskeleton and transduces signals, eliciting changes in the intracellular pH, cytoplasmic calcium level, phospholipid metabolism, protein tyrosine and serine/threonine phosphorylation, and expression of certain genes (21). Recent studies suggest that integrin-associated transmembrane proteins, including CD9 and TM4, may also participate in integrin-mediated signaling (22). We have shown here that CD9 associates with integrin α6β1 in eggs. Therefore, integrin α6β1 may transduce signals to CD9 and initiate, or otherwise promote, fusion. However, CD9 may directly function in membrane fusion. In support of this possibility, it should be noted that some anti-CD9 or anti-TM4 antibodies block virus-mediated syncytium formations where the involvement of integrin is not clear (23).

Our results show that CD9 is a crucial factor for mouse oocytes in fertilization. CD9−/− mice may serve to elucidate the precise mechanism of sperm-egg fusion and the role of CD9-integrin complex.

  • * These authors contributed equally to this work.

  • To whom correspondence should be addressed. E-mail: emekada{at}


View Abstract

Stay Connected to Science

Editor's Blog

Navigate This Article