MOLECULAR REPRODUCTION AND DEVELOPMENT 31:215-222 (1992)

Evidence for the Presence of an Integrin Cell Adhesion Receptor on the Oolemma of Unfertilized Human Oocytes F.M. FUS1,l M. VIGNALI,' M. BUSACCA,l AND R.A. BRONSON2 'III Department of Obstetrics and Gynecology, University of Milano, Ospedale Sun Raffaele, Milano, Italy; 'Division of Reproductive Endocrinology, Department of Obstetrics and Gynecology, State University of New York at Stony Brook, Stony Brook, New York ABSTRACT The Arg-Gly-Asp peptide (RGD), contained in several extracellular matrix proteins such as fibronectin, laminin, vitronectin, and collagen, is a tripeptide that plays a role as a recognition sequence in many cell-to-cell and cell-to-matrix adhesion mechanisms, through its interaction with several receptors of the integrin family. We previously described the ability of the oolemma of hamster oocytes to bind GRGDTP coupled to the surface of activated immunobeads and demonstrated that RGD-containing oligopeptides inhibit the adhesion of human and hamster spermatozoa to zona-free hamster oocytes and their subsequent penetration. In the present experiments,we show, utilizing immunobeads coated with an RGD-containing peptide (PepTiteTM 2000), that the oolemma of unfertilized human eggs is also able to recognize this adhesion sequence. The binding of PepTiteTM 2000coated immunobeads to the oolemma was inhibited by the oligopeptide GRGDTP as well as by fibronectin and laminin. When immunobeads were prepared with a PepTiteTM concentration of 10 pg/ml, GRGDTP 150 pg/ml, laminin 80 pg/ml, and fibronectin 60 pg/ml inhibited bead rosetting on the egg surface. These data suggest that a specific binding moiety for RGD is present on the human egg surface. The binding of fibronectin to the oolemma was also demonstrated by the rosetting of immunobeads coupled with antifibronectin antibody to human oocytes after their exposure to 1 mg/ml free fibronectin. Such binding of fibronectin to the oolemma could be inhibited by coincubation with a monoclonal antibody directed against the cell adhesion fragment of fibronectin. In addition, oolemmal rosetting of immunobeads coupled with a monoclonal antibody directed against the a 5 subunit, usually part of the fibronectin receptor VLA 5 (a5pl),provided additional evidence that a putative fibronectin receptor is present on the oolernma of human eggs. Key Words: RGD, a5, Immunobeads, lntegrins INTRODUCTION The integrin family consists of a series of related a p heterodimers (subunits of 95,000-200,000 Mr), involved in a variety of cell-to-matrix and cell-to-cell adhesion functions (Hemler, 1990). Integrins were originally organized into three distinct subfamilies based on

0 1992 WILEY-LISS, INC.

p-subunit type, p l , p2, and p3 (Hynes, 1987; Hemler et al., 1987). More recent evidence, however, that at least seven p subunits exist and that several a chains may associate with more than one p subunit makes a clear classification more difficult (Hemler, 1990).Members of the first subfamily, also called VLA proteins, each contain the p l subunit in association with one of at least six different a subunits (Hemler, 1990). The p l integrin subfamily includes receptors that bind to fibronectin, laminin, and collagen and that are expressed on many nonhematopoietic and leukocyte cell types (Springer, 1990). The VLA 5 protein (a5pl), demonstrated to be present on fibroblasts, endothelial cells, epithelial cells, platelets, thymocytes, and T lymphocytes (Wayner et al., 1988), was originally described as the mammalian fibronectin receptor (Yamada, 1989; Pytela et al., 1985a). This receptor, as well as other integrins, such as avpl, avp3, and avp5 (Pytela et al., 1985b; Springer, 1990); platelet glycoprotein IIb/IIIa (Hynes 1987); and perhaps aMp2 (Springer, 19901, recognize the aminoacid tripeptide Arg-Gly-Asp (RGD) as a ligand recognition sequence (Ruoslahti and Piershbacher, 1986). The RGD sequence has been shown to be present in several matrix proteins, such as fibronectin, vitronectin, type I collagen, and probably laminin (Ruoslahti and Piershbacher, 1986), and the addition of various RGD-containing oligopeptides competitively inhibits the attachment of a number of cell types to extracellular matrix proteins (Yamada and Kennedy, 1985; Plow et al., 1985). We previously demonstrated that the RGD sequence is involved in the adhesion of human spermatozoa to zona-free hamster eggs and in their subsequent penetration (Bronson and Fusi, 1990a,b), in that several fibronectin-derived RGD-containing peptides competitively inhibit sperm-egg interaction. The presence of a specific binding moiety for RGD on the oolemma of hamster eggs was also suggested by the adherence of immunobeads coated with an RGD-containing peptide

Received August 19,1991;accepted October 9, 1991. Address reprint requests to Richard A. Bronson, MD, Department of Obstetrics and Gynecology, Health Science Center, T-9, 060, Stony Brook, NY 11794-8091.

216

F.M. FUSI ET AL.

(GRGDTP) to the eggs, while bead matrix alone or beads coupled with a non-RGD containing fibronectinderived peptide did not. In humans, fibronectin has been shown to be present in seminal plasma (Vuento et al., 1980) and on fixed spermatozoa (Vuento et al., 1984), and its absence was suggested to possibly characterize morphologically abnormal spermatozoa of infertile men (Glander et al., 1987). We recently demonstrated that living human spermatozoa display F n on their surface after capacitation, and anti-Fn antibodies are able to inhibit partially the binding of human spermatozoa to zona-free hamster oocytes (Fusi and Bronson, 1991). In the present experiments, we utilized a synthetic RGD-containing peptide (PepTiteTM 20001, coupled to immunobead matrix, to demonstrate the presence of a n RGD binding moiety on the oolemma of unfertilized human oocytes. The specificity of this binding was confirmed by the ability of the oligopeptide GRGDTP and of the RGD-containing proteins fibronectin and laminin to inhibit PepTite-coated immunobead rosetting in a concentration-dependent manner. The binding of fibronectin to human oocytes was demonstrated by means of anti-Fn antibody-coated immunobeads, which rosetted to eggs after oocyte exposure to human serum fibronectin. Binding of anti-Fn immunobeads to the eggs could be partially inhibited by coincubation in medium containing a n antibody directed against the cell binding domain of Fn. In addition, we demonstrated oolemmal reactivity with a monoclonal antibody directed against a5,further suggesting that integrins may be present on human oocytes.

lano, Italy), for each PepTite-2000 concentration, was reconstituted in BWW and washed by centrifugation at 1,OOOg for 10 min, and medium was then removed by aspiration. Bead matrix was thereafter resuspended in 250 pl of each PepTite concentration, and incubated for 2 h r a t room temperature. After coating, PepTite solution was removed by centrifugation at l,OOOg, and beads were washed one time with 3 ml BWW. Bead suspension was centrifuged at 1,OOOg for 10 min and resuspended in 500 p1 BWW containing 5 mg/ml human serum albumin (HSA; Sigma Chem. Co, St. Louis, MO). As a control, the fibronectin-derived peptide C-8287, not containing the RGD sequence, was coupled to immunobeads using the same procedure utilized for antibody coupling to immunobeads, except that the pH of the coupling buffer was brought to 7, to the isoelectric point of oligopeptide, a s suggested by the manufacturer of immunobeads (Bio-Rad). This alternative approach was utilized in that, while PepTite was specifically designed to spontaneously adhere to inert supports, we preferred to utilize a covalent linkage to ensure C-8287 binding to immunobeads.

Antibody Coupling to Immunobeads Monoclonal antibody CDw49e, directed against the a,-integrin subunit was obtained from Immunotech (Marseille Cedex, France) and is a mouse IgG2c.u. A mouse myeloma protein of the same IgG subclass, UPClO (Sigma), was utilized as a control. In addition, as a further control, the mouse myeloma protein MOPC21, of the IgGl subclass, was used. Goat antihuman fibronectin polyclonal antibody F-1509 (Sigma) and normal goat serum (Dako, CopenMATERIALS AND METHODS hagen, Denmark) were utilized, respectively, to deterFibronectin-Related Peptides, Fibronectin, mine Fn binding to the oolemma of human oocytes and and Laminin as a control. All these antibodies were coupled to immuPepTite-2000TM, a synthetic peptide containing the nobead matrix to detect their binding to the egg plasma Arg-Gly-Asp sequence (RGD), was purchased from membrane by means of oolemmal rosetting of immunoTelios Pharmaceuticals (San Diego, CA). A stock solu- beads. The coupling of antibodies to activated Immunotion of 5 mg/ml PepTite was obtained by adding 1 ml beads was performed utilizing a modification of the 70% ethanol in phosphate-buffered saline (PBS) to the Bio-Rad protocol. The immunoglobulin preparations lyophylized product and stored at 4°C. (125 p,g antibody in 0.5 ml PBS) were dialyzed overThe fibronectin-derived peptides Gly-Arg-Gly-Asp- night at 4°C against a coupling buffer (0.003 M phosSer-Pro, containing the RGD sequence (GRGDTP; phate, pH 6.3). Two and one-half milligrams of actiSigma Chemical Co, St. Louis, MO) and Cys-Gln-Asp- vated immunobead matrix was reconstituted for each Ser-Glu-Thr-Arg-Thr-Phe-Tyr (C-8287; Sigma), not antibody in the same buffer, centrifuged, and resuscontaining the RCD sequence, were resuspended at a 1 pended in the antibody solution. After a 1h r incubation mg/ml concentration in PBS, aliquoted, and stored a t at P C , 0.5 mg of EDAC (l-ethyl-3-dimethylaminopro-20°C until used. Fibronectin (F-2006, Sigma Chemi- pyl carbodiimide HCl) (Bio-Rad) was added, and the cal Co.) and laminin (L-2020, Sigma Chemical Co.) suspension was incubated overnight at 4°C. Thereafter, were resuspended at a 4 mg/ml concentration in PBS, the beads were washed with PBS, followed by two aliquoted, and stored at -20°C. washings with 1.4 M NaCl/PBS and two additional washings with PBS. The beads were then resuspended PepTite2000TM Coupling to Immunobeads in 1ml PBS containing 5 mg/ml bovine serum albumin The stock solution of PepTite-2000 in ethanol was (BSA; Sigma). diluted at three different concentrations, 10 pg/ml, 25 The success of antibody binding to immunobeads was pg/ml, and 100 pg/ml, using Biggers-Whitten-Whit- checked by means of immunofluorescence. Fifty microtingham medium (BWW; Biggers et al., 1971). Two and liters of each bead preparation and 50 p1 of bead matrix one-half milligrams immunobead matrix (Bio-Rad, Mi- resuspended a t the same bead concentration were incu-

HUMAN OOLEMMAL INTEGRIN RECEPTORS bated for 1 h r at room temperature in the dark, after addition of goat antimouse Ig-FITC conjugated antibody (Sigma) or of rabbit antigoat immunoglobulins FITC-conjugated antibody (Sigma) at a 1:40 dilution in 0.5 ml microcentrifuge tubes. Thereafter, tubes were washed three times with PBS/BSA 5 mg/ml in a Biofuge A Microcentrifuge (Heraeus), and the pellet was resuspended in 50 p1 PBS/BSA. A drop of each bead suspension was placed on a slide, and immunobead fluorescence was checked by observation under f luorescent illumination.

217

were tranferred using a mouth pipette, and the dishes were incubated at 37°C 5% COz in air for 20 min. Thereafter, they were washed by serial passage in five droplets of BWW/HSA 5 mg/ml and scored utilizing a n inverted microscope (Olympus, Milano, Italy). Alternatively, they were transfered to slides and observed as whole mounts under phase-contrast illumination. The same individual pipetted all the eggs in each treatment group for all experiments performed in this study, and the same micropipette, of approximately 100-120 p m diameter, was used for all eggs washed in a n individual experiment. In that there was no rosetting of control immunobeads to the oolemma of human eggs, the oolemmal binding of any number of PepTite-coated immunobeads was considered positive. A quantitative grading system was established to score oolemmal bead rosetting as indicated in Table 1.

Human Oocytes Human oocytes were obtained from women undergoing ovum retrieval for gamete intrafallopian transfer (GIFT). Induction of superovulation in women undergoing GIFT at the San Raffaele Hospital of Milano was performed using human menopausal gonadotropins (HMG; Serono, Rome, Italy), after endogenous gonadoPepTiteTM-CoupledBead Binding tropin production was suppressed by Buserelin (SupreInhibition Studies fact; Hoechst Italia, L'Aquila, Italy), and were moniWe utilized the oligopeptide GRGDTP, fibronectin, tored by transvaginal ultrasound and serial E2 and laminin to study the competition between the solid determinations. When the lead follicle was 17-19 mm phase RGD coupled to beads via PepTite and RGDin diameter and serum estradiol was greater than 200- containing sequences in solution. For each preparation, 300 pg/ml per follicle, human chorionic gonadotropin a 30 p,1 drop of PepTite beads was added to a 30 pl drop (HCG; Profasi; Serono) was administered. Thirty-two to of BWW/HSA 30 mg/ml at double the concentration of thirty-six hours after HCG, oocytes were collected by GRGDTP, fibronectin, or laminin to obtain each of the means of transvaginal ultrasound-guided follicle punc- desired final concentrations of proteins. The proteinture. A maximum of six oocytes per patient was utilized free control was prepared by adding a 30 p1 drop of for the transfer, and the eggs in excess were used for our PepTite-coated beads to a 30 pl drop of BWW/HSA 30 study, because in vitro fertilization and cryopreserva- mg/ml. As a further control, we used the C-8287 peptide tion of embryos for further utilization by patients are at the same concentrations to test the effects of the not performed at our center. Oocytes that appeared to addition of a fibronectin-derived non-RGD-containing be mature, with cumulus and corona expanded, which protein. The eggs were then transferred to the drop, exhibited a first polar body and were free from blood incubated for 20 min at 37°C in 5% COP, washed as contamination, were selected to be replaced. Morpho- described above, and observed for bead rosetting. logically normal preovulatory eggs that had undergone In another series of experiments, to test the reversgerminal vesicle breakdown, with cumulus well ex- ibility of RGD-containing protein binding to the 00panded but with or without a first polar body and free lemma, oocytes were first preincubated for 1h r a t 37°C from or with minimal blood contamination, were uti- in 5%COz in a drop of BWW/HSA 30 mg/ml containing lized for our studies. Immature oocytes or eggs not of several concentrations of GRGDTP, fibronectin, or normal appearance were discarded. laminin. Thereafter, they were alternatively placed in Corona radiata cells and zona pellucida were re- a 60 pl drop containing PepTite beads and the same moved from each oocyte by means of short passages concentration of protein or were carefully washed (5-10 sec), performed using a mouth pipette, in a 50 p1 through five passages in BWW/HSA 5 mg/ml and exdrop of acidified BWW, and placed in a 35 mm polysty- posed to PepTite-coated immunobeads in the absence of rene dish (Falcon Becton Dickinson, Milano, Italy). RGD containing proteins. BWW was acidified by adding 0.1 N HC1 to BWWIHSA 5 mg/ml, to obtain a pH of 3.5. Thereafter, each oocyte Selection of the Method To Test Antibody was carefully washed by serial passage in at least five Binding to Human Oocytes 50 pl droplets of BWW/HSA 5 mg/ml and was then In that we had previously demonstrated Fcy receputilized for the experiments. tors both on hamster (Bronson et al., 1990) and human PepTiteTM-CoupledBead Binding to Zona-Free (Bronson et al., 1991) oocytes, we performed a prelimiHuman Oocytes nary experiment to determine whether Fc-mediated Thirty microliters of each PepTite-coated bead sus- binding of mouse immunoglobulins to human oocytes pension or of the control was added to a 30 pl drop of could be inhibited through Ig-coupling to a solid supBWW/HSA 30 mg/ml in a 35 mm dish (Falcon) and port. We utilized two mouse myeloma proteins, UPClO carefully mixed, so that beads appeared uniformly dis- and MOPC2l (Sigma) of the IgG2a and IgGl subpersed in the drop. The oocytes (one to four per drop) classes, respectively.

218

F.M. FUSI ET AL.

Zona-free human oocytes were placed in a 50 pl drop of BWW/BSA 30 mg/ml in which either murine myeloma protein was resuspended a t a concentration of 20 pg/ml and incubated for 45 min at 37°C in 5% CO, in air. Thereafter, eggs were washed by means of passage through five 50 pl drops of BWW/BSA 5 mg/ml. Onehalf milliliter of rabbit antimouse immunoglobulin immunobeads (Bio-Rad, Richmond, CA), heavy and light chain-specific, was washed by centrifugation in 5 ml PBS/BSA 5 mg/ml three times a t 1,500g for 4 rnin and resuspended in 1 ml of PBS/BSA 10 mg/ml. Eggs were transferred to a 50 pl drop of these rabbit antimouse immunobeads and incubated for 20 min at 37°C in 5% CO, in air. Thereafter, oocytes were washed a s described and observed for bead rosetting. In the same experiment, other oocytes were exposed for 20 min to beads coupled either with UPClO or MOPC21, washed as described above, and observed for bead rosetting.

Antifibronectin-CoupledBead Binding to Zona-Free Human Oocytes Human oocytes were placed in a 30 pl drop of either antifibronectin beads or nonimmune goat serum-coupled beads. They were first incubated for 20 rnin as previously described, and the absence of bead rosetting to the oolemma was confirmed. Thereafter, oocytes were placed in a drop of BWW/HSA 30 mg/ml containing 100 pg/ml, 200 pg/ml, or 1 mg/ml fibronectin and incubated for 1 h r at 37°C in 5% C 0 2 in air. After incubation, eggs were carefully washed through five 50 ml drops of BWW/HSA and placed again in a drop of either antifibronectin beads or normal goat serum-coupled beads. Incubation and scoring were performed as described above. To confirm the specificity of fibronectin binding to the oolemma of human oocytes, some eggs, preexposed to antifibronectin beads and negative for binding, were incubated for 1 h r alternatively in BWWIHSA 30 mg/ml containing 1mg/ml fibronectin or BWW/HSA 30 mg/ml containing 1 mg/ml fibronectin and 20 pg/ml of the antifibronectin monoclonal antibody 3E3 (Boehringer Manheim, Milano, Italy), a n antibody that specifically recognizes the cell binding sequence of fibronectin. Thereafter, eggs were washed and incubated with antifibronectin beads for 20 min, and bead rosetting to the oolemma was evaluated. Anti-a5 Bead Binding to Zona-Free Human Oocytes Zona-free human oocytes were incubated for 20 min at 37°C in 5%CO, in air in a 30 pl drop of immunobeads coupled with either UPC10, MOPC21, or bead matrix resuspended in PBS/BSA 5 mg/ml, so that each egg constituted its own control for a n t i d bead binding. After initial scoring, to exclude nonspecific bead oolemma1 binding, eggs were finally placed in a 30 p1 drop of anti-a5 beads and incubated for 20 min, and bead rosetting of the surface scored as described above.

Fig. 1. PepTite-coated (10kg/ml) immunobead rosetting on the surface of a zona-free human oocyte. x250.

RESULTS After removal of the zona pellucida, human oocytes were able to recognize RGD-containing peptides, in that immunobeads coated at several concentrations of PepTiteTM 2000 bound to the oolemma (Fig. 1). In contrast, bead matrix or immunobeads prepared utilizing a fibronectin-related, non-RGD-containing peptide, C-8287, did not rosette on the oolemmal surface (Table 1). To test the binding specificity of PepTite-coated beads, free fibronectin was added at several concentrations to the droplets containing oocytes and PepTite beads prepared with the 10 pg/ml and 100 pg/ml concentrations (Table 2). While 200 pg/ml fibronectin completely inhibited the binding of immunobeads coated with PepTite at 10 pg/ml to oocytes, incomplete inhibition was observed even a t a free fibronectin concentration of 2 mg/ml, when immunobeads were prepared with 100 pg/ml PepTite. We chose the bead preparation obtained utilizing PepTite 10 pg/ml for further studies. A concentration-dependent inhibition of rosetting of PepTite-coated immunobeads to the oolemma of human eggs was noted for fibronectin, laminin, and GRGDTP. On addition of fibronectin, we observed a limited inhibition at the 20 pg/ml concentration, a more significant inhibition at 40 pg/ml, and a complete inhibition from 60 pg/ml. For laminin, we obtained complete inhibition

HUMAN OOLEMMAL INTEGRIN RECEPTORS

219

TABLE 1. Binding of Immunobeads Coated at Different Concentrations With PepTite 2000, an RGD-Containing Peptide, to Zona-Free Human Oocytes Degree of oolemmal rosetting as judged by number of beads bindingb

Concentration of PepTite 2000 Number of eggs 10 pg/ml-coated beads 14 25 pg/ml-coated beads 6 100 pg/ml-coated beads 8 C8287-coated beadsa 7 Bead matrixa 6 aThe non-RGD-containing, fibronectin-derived peptide Sigma C8287 and uncoupled immunobead matrix were used as controls. b-, No bead binding; more than 50 beads.

+++ +++ +++

+++,

TABLE 2. Inhibition of PepTite Bead Rosetting to Zona-Free Human Oocytes by Addition of Free Fibronectin (Fn) i n the Medium at Different Concentrations Degree of oolemmal Concentration of fibronectin rosetting as judged PepTite 2000-coated in medium with by number of beads bead suspension binding a immunobeads Beads coated with PepTite 10 pg/ml

Beads coated with PepTite 100 pg/ml

Fn-free medium Fn 200 pg/ml Fn 500 pg/ml Fn 2 mg/ml Fn-free medium Fn 200 pg/ml Fn 500 pg/ml Fn 2 mg/ml

+++ -

+++ +++ ++

*

No bead binding; f,fewer than ten beads bound t o the oolemma; +, ten to 20 beads;

++, 20-50 beads; +++, more than 50 beads. using a n 80 Fg/ml concentration and, utilizing GRGDTP, at 200 pg/ml. No inhibition was observed when C-8287 was added to the medium (Fig. 2). In the following series of experiments, we tested whether the inhibition obtained utilizing the oligopeptide GRGDTP or the two proteins could be reversed by washing the oocytes after their exposure to several protein concentrations and by thereafter incubating them with PepTite beads in the absence of proteins. While the inhibition obtained with GRGDTP could be almost completely reversed, only a partial recovery of binding was obtained following washout of laminin, and no recovery was obtained with fibronectin (Table 3). These results suggest that fibronectin binding to human eggs is more intense and more stable than the binding of GRGDTP. Antibody coupling to immunobeads inhibits the interaction of immunoglobulins with Fc receptors present on the human egg oolemma. Rabbit antimouse immunoglobulin immunobeads rosetted to the oolemma of four eggs exposed to free MOPC2l and to five incubated in free UPC10, while bead rosetting of oocytes exposed to MOPC21- and UPC10-coupled immunobeads was not observed (four eggs for each). When the same oocytes

that had not reacted with mouse myeloma protein-coupled immunobeads were exposed to free immunoglobulins in solution, their interaction was demonstrated by means of antimouse immunobead rosetting to the surface. This finding demonstrated that these oocytes did not differ in their receptors from those exposed to free antibodies. Hence the coupling of antibodies to a solid support appears to impede the interaction between IgG Fc and oolemmal Fc receptors, perhaps by occluding the binding sites. To demonstrate the binding of fibronectin to human oocytes, they were preexposed to beads coupled with a goat polyclonal antifibronectin antibody (F15091, or normal goat serum, to exclude their nonspecific binding to the oolemma. Thereafter, eggs were exposed to fibronectin and successively to either of the bead preparations (Table 4). Eggs exposed to fibronectin bound antifibronectin beads, but there was no oolemmal rosetting by normal goat serum beads. This result indicated that fibronectin was not intrinsically present on the egg surface but was bound subsequently. Bead rosetting could be partially inhibited by adding to the solution a n antifibronectin antibody specifically directed against the fibronectin cell binding domain (Ta-

220

F.M. FUSI ET AL.

U

t-

+

’

Lamlnin GRGDTP Flbronectln Controla

0 40 LO 10 1 0 1 1 5 0 200 5 0 0

concentrations olnlml) Control consisted of c - 8 2 8 7 , a flbronectln-derived ollgopeptlde t h a t does not contain RGD

Fig. 2. Concentration-dependent inhibition of rosetting of PepTite-coated immunobeads by free fibronectin, laminin, or GRGDTP. The peptide C-8287, not containing the RGD sequence, was utilized as a control.

TABLE 3. Comparison of the Effect of Washing Oocytes Following Incubation With RGD-Containing Proteins on Their Inhibition of Binding of PepTite-Coated Immunobeads to the Oolemma vs. Their Effects During Coincubation* GRGDTP Laminin Fibronectin Concentration of Washing Coincubation Washing Washing Coincubation Coincubation RGD-containing Score No. Score Score No. Score No. No. Score No. proteins (pg/ml) No. Score 40 60 80 100 150 200

500

2 3 2 3 2 2 2

++ + +

*f

-

3 3 3 3 2 2 3

+++ +++ +++ ++ + + f

2 3 3 3 2 3 2

++ + -

3 3 3 3 2 3 2

++ ++ + + + -

-

3 4

3 2 2 2 2

?c

-

-

3 3 3 3 2 2 2

+f -

-

*Each egg was first incubated with GRGDTP,laminin, or fibronectin for 60 min, exposedto PepTite-coatedimmunobeads, scored, and subsequently washed in BWWIHSA 30 mg/ml and reexposed to PepTite beads. The rosetting of PepTitecoated (10 pg/ml) immunobeads to the oolemma was scored a s follows:-, no bead binding; k,fewer than ten beads bound 20-50 beads; more than 50 beads. to the oolemma; +, ten to 20 beads;

++,

+++,

ble 4),suggesting that fibronectin binding to the oolemma is RGD-dependent. The specific oolemmal rosetting of immunobeads coupled with anti-a 5 antibody, but not the murine myeloma UPC21, provides further evidence for a n integrin on the surface of zonafree human oocytes, which might act as a n RGD-dependent receptor (Table 5). DISCUSSION We previously presented evidence that the RGD recognition sequence is involved in the oolemmal adhesion and penetration of zona-free hamster oocytes by human and hamster spermatozoa (Bronson and Fusi, 1990a,b). In the present experiments, we showed t h a t the oolemma of human oocytes also is able to recognize RGD. The binding of this tripeptide to the oolemma appears to be specific, in that PepTite-coated immunobead rosetting to the egg surface could be inhibited by the addition of the RGD-containing proteins fibronectin and laminin as well as by GRGDTP. In addition, when beads were prepared using two different concentrations

of PepTite, their affinity for the oolemma differed. The binding of beads prepared with the 10 kg/ml PepTite concentration could be inhibited by relatively low fibronectin concentrations, while inhibition of binding of beads prepared with 100 kg/ml PepTite required a much higher fibronectin concentration. The fact that a different bead affinity for the oolemma was observed using two different PepTite preparations, as well as the concentration-mediated inhibition of binding by several RGD-containing proteins, suggests a receptor-mediated recognition of RGD. The binding of fibronectin to the egg surface appeared to be more avid than that of laminin or of the oligopeptide GRGDTP, since lower concentrations of F n were able to inhibit PepTite-bead rosetting of the oolemma, and this effect could not be eliminated by simple washing of the eggs. In mammals, several receptor-ligand interactions are RGD-dependent, and in humans the most important known RGD-recognizing receptors are five. Of these, a5pl (VLA 5) and avp1 recognize RGD in fibronectin; avb3 (VNR receptor) in vitronectin, fi-

HUMAN OOLEMMAL INTEGRIN RECEPTORS

221

TABLE 4. Binding of Antifibronectin-Coated Immunobeads to the Oolemma of Zona-Free Human Eggs Following Their Exposure to Fibronectin at Different Concentrations * Number of beads rosetting the egg surface AntiFn-coated Goat serum-coated immunobeads immunobeads Zona-free human eggs in 0 (four eggs) 0 (three eggs) the absence of fibronectin Zona-free human eggs 100 Fg/ml 0 (five eggs) 0 (two eggs) exposed to fibronectin 500 pg/ml 5-10 (five eggs) 0 (two eggs) 1 mg/ml 20-50 (five eggs) 0 (two eggs) Zona-free human eggs 0 (six eggs) 0 (one egg) exposed to fibronectin 1 mg/ml and antiFn antibody 20 fig * Immunobeads were coupled with F1509, a polyclonal antibody raised in goats against fibronectin, or with proteins from normal goat serum. TABLE 5. Binding of Immunobeads Coated With Anti-VLA5 Monoclonal Antibody to Zona-Free Human Oocytes* Oolemmal Oolemmal binding binding of control of anti-VLA5 Number of oocytes immunobeads" immunobeadsa ++/+++ MOPC21-coupledbeads 3 ++/+++ UPC10-coupled beads 5 Resuspended bead matrix 2 *Oocytes were first preexposed to immunobeads coupled with IgGl or IgG2a mouse myeloma proteins or uncoupled bead matrix as controls, to rule out nonspecific binding, and subsequently to anti-VLA5-coupledimmunobeads. fewer than ten beads bound aQuantitation of oolemmal rosetting: -, no bead binding; f, more than 50 beads. to the oolemma; +, ten to 20 beads; ++, 20-50 beads;

++/+++

+++,

bronectin, trombospondin, and von Willebrand factor; avp5 in fibronectin and vitronectin; and aIIbp3 in fibronectin, fibrinogen, and von Willebrand factor (Springer, 1990). This specificity suggests t h a t regions containing the RGD sequence in different proteins have a different conformation and that the secondary amino acid sequences modify, to a different extent, the activity of this cell binding domain (Cierniewski e t al., 1988). Several lines of evidence suggest that fibronectin is involved in mammalian and human reproduction. Fibronectin is involved in the resumption of meiosis of r a t oocytes (Hung et al., 1989), and the levels of fibronectin in human follicular fluid correlate well with follicular size and oocyte maturity, to such a n extent that it was supposed to be a useful marker of the potential of the oocyte to be fertilized (Tsuiki et al., 1988). In addition, fibronectin was shown to be involved in mouse embryo cleavage and probably implantation, in that antifibronectin antibodies, a s well as fibronectin- and RGDcontaining peptides, inhibit embryo development in a dose-dependent manner (Hung e t al., 1989). Cellular fibronectin was also demonstrated on human spermatozoa (Vuento et al., 1984; Glander et al., 1987), and the proportion of spermatozoa displaying fibronectin increases after sperm capacitation (Fusi and Bronson, 1991). In addition, we found t h a t antifibronectin anti-

bodies inhibit the oolemmal adhesion of zona-free hamster oocytes by human spermatozoa and their subsequent penetration, suggesting that the RGD-binding moiety on hamster and human oocytes could be a fibronectin receptor (Fusi and Bronson, 1991). In these experiments, we demonstrated, utilizing antifibronectin beads, that fibronectin binds to zona-free human oocytes, further supporting the hypothesis that a receptor-mediated, RGD-dependent recognition of fibronectin is a property of the human oolemma. Additional evidence that a n integrin containing the a5 subunit could be the fibronectin recognition receptor on human oocytes is provided by our observation that a n anti-& monoclonal antibody bound specifically to the oolemma. No integrins containing the a5 subunit are known other than VLA 5, and, although it is theoretically possible that a5 forms a complex with another p subunit, i t is likely that this receptor is VLA-5 (a5pl). This receptor is known to bind to fibronectin with only moderate affinity (kd 8 x lo7)(Yamada, 19891, and the relatively high fibronectin concentration (1 mg/ml) needed to obtain its oolemmal binding, as observed with antifibronectin antibodies bound to immunobeads, may be related to the limited affinity of free fibronectin for its receptor. Alternatively, the sensitivity of our immunobead preparation may be too limited

222

F.M. FUSI ET AL.

to allow the detection of lower fibronectin concentrations. The incomplete inhibition of fibronectin binding to the oolemma by an antibody directed against the cell attachment fragment could be explained on the basis of the observation that fibronectin-mediated cell adhesion may be mediated by more th an one molecular mechanism. The best characterized site of the fibronectin molecule for mediating adhesive functions is classically the “cell binding domain,” which interacts with fibronectin receptor via the RGD sequence, but a second cell binding non-RGD-dependent domain exists and acts sinergistically with the first (Yamada, 1989). This second cell binding domain is in part responsible for the much weaker activity of RGD-containing peptides compared with fibronectin on cell adhesion (Obara et al., 1987), but also different integrins with overlapping specificities may be involved. The combined evidence demonstrating the presence of fibronectin on the surface of capacitated spermatozoa (Fusi and Bronson, 1991)and of a putative fibronectin receptor on the oolemma of human eggs, in addition to the data demonstrating the involvement of RGD in human sperm-hamster oocyte interaction, would appear to indicate that this cell adhesion mechanism plays a physiologic role in gamete interactions.

REFERENCES Biggers JD, Whitten WK, Whittingham DG (1971): The culture of mouse embryos in vitro. In J C Daniel (ed). “Methods in Mammalian Embryology.” San Francisco: W.H. Freeman and Co, pp 8 6 1 1 6 . Bronson RA, Fleit HB, Fusi F (1990): Identification of a n oolemmal IgG Fc receptor: its role in promoting binding of antibody-labelled human sperm to zona-free hamster eggs. Am J Reprod Immunol 23:87-92. Bronson RA, Fusi F (1990a): Sperm-oolemmal interaction: role of the RGD adhesion peptide. Fertil Steril54:527-529. Bronson RA, Fusi F (1990b): Evidence that a n Arg-Gly-Asp (RGD) adhesion sequence plays a role in mammalian fertilization. Biol Reprod 43:1019-1025. Bronson RA, Fusi FM, Fleit HB (1991): Monoclonal antibodies identify Fca receptors on unfertilized human oocytes but not spermatozoa. J Reprod Immunol (in press). Cierniewski C, Swiatkowska M, Poniatowski J , Niewiarowska J

(1988): Anti-(Arg-Gly-Asp-Ser) antibody and its interaction with fibronectin, fibrinogen and platelets. Eur J Biochem 177:109-115. Fusi FM, Bronson RA (1991): Sperm surface fibronectin: Expression following capacitation. J Androl (in press). Glander HJ, Herrmann K, Haustein U F (1987): The equatorial Fibronectin band (EFB) on human spermatozoa: A diagnostic help for male fertility? Andrologia 19:456459. Hemler ME (1990): VLA proteins in the integrin family. Annu Rev Immunol 8:365-400. Hemler ME, Huang C, Schwarz L (1987): The VLA protein family. J Biol Chem 262:3300-3309. Hung IT,Tsuiki A, Yemini M (1989): Fibronectin in reproduction. Steroids 54:575-582. Hynes RO (1987): Integrins: A family of cell surface receptors. Cell 482549454, Obara M, Kang MS, Rocher-Dufour S, Kornblihtt A, Thiery JP, Yamada KM (1987): Expression of the cell binding domain of human fibronectin in E. coli. FEBS Lett 213:261-264. Plow EF, Piershbacher MD, Ruoslahti E, Marguerie GA, Ginsberg MH (1985): The effect of Arg-Gly-Asp-containingpeptides on fibrinogen and von Willebrand factor binding to platelets. Proc Natl Acad Sci USA 82S057-8061. Pytela R, Pierschbacher MD, Ruoslahti E (1985a): Identification and isolation of a 140 kd cell surface glycoprotein with properties expected of a fibronectin receptor. Cell 40:191-198. Pytela R, Pierschbacher MD, Ruoslahti E (1985b): A 125/115 kDa cell surface receptor specific for vitronectin interacts with the arginineglycine-aspartic acid adhesion sequence derived from fibronectin. Proc Natl Acad Sci USA 8257665770. Ruoslahti E, Piershbacher MD (1986): Arg-Gly-Asp: A versatile cell recognition signal. Cell 44:517-518. Springer T (1990):Adhesion Receptors of the immune system. Nature 346:425434. Tsuiki A, Preyer J, Hung T (1988): Fibronectin and glycosaminoglycans in human preovulatory follicular fluid and their correlation to follicular maturation. Hum Reprod 3:425429. Vuento M, Kuusela P, Virkki M, Koskimies A (1984): Characterization of fibronectin on human spermatozoa. Hoppe-Seyler’s Z Physiol Chem 365:757-762. Vuento M, Salonen E, Koskimies A, Stenman U (1980): High concentrations of fibronectin-like antigens in human seminal plasma. Hoppe-Seyler’s Z Physiol Chem 361:1453-1456. Wayner EA, Carter WG, Piotrowicz RS, Kunicki T J (1988):The function of multiple extracellular matrix receptors in mediating cell adhesion to extracellular matrix. J Cell Biol 107:1881-1891. Yamada KM (1989): Fibronectin domains and receptors. In D Mosher (ed): “Fibronectin.” New York: Academic Press Inc., pp 47-121. Yamada KM, Kennedy DW (1985): Amino acid sequence specificities of a n adhesive recognition signal. J Cell Biochem 28:99-104.