New insights into the interaction between the gp120 and the HIV receptor in human sperm (human.sperm/gp120/galactoglycerolipid/antigalactosylceramide/seminolipid/spermatogonia)

Journal of Reproductive Immunology41 (1998) 213–231

New insights into the interaction between thegp120 and the HIV receptor in human sperm

(human.sperm/gp120/galactoglycerolipid/antigalactosylceramide/seminolipid/spermatogonia)

Alessandra Brogi a, Rivo Presentini b, Elena Moretti a,Michelina Strazza a, Paola Piomboni a,

Elvira Costantino-Ceccarini c,*a Istituto Biologia Generale, Uni6ersita Siena, Siena 53100, Italy

b Scla6o Diagnostics S.r.l., Siena 53100, Italyc Centro Studio Cellule Germinali, CNR, Via Pendola 62, Siena 53100, Italy

Abstract

The human immunodeficiency virus (HIV) can infect some cell types which lack CD4.Galactosylceramide, a glycolipid present in the nervous system and colonic epithelial cells,has been implicated in the virus entry in these cells. Our data demonstrate that the HIVsurface glycoprotein gp120 binds to the galactosyl-alkyl-acylglycerol (GalAAG), a glycolipidstructurally related to galactosylceramide present on the surface membrane of the spermato-zoa. In this paper, we review our previous data and further confirm the specificity of theinteraction between this galactoglycerolipid and the gp120. Consistent with the structuralsimilarity to galactosylceramide, the sperm GalAAG is capable of specifically binding thegp120. The specificity of the binding of antibodies anti-galactosylceramide and the gp120 tothe sperm extract and to the purified GalAAG fraction prepared from the same extract hasbeen demonstrated utilizing an ELISA assay which favors sensitivity and specificity. Immu-nofluorescence and immunoelectron microscopy data show a different localization for theGalAAG and its sulfated form the seminolipid (SGalAAG). The GalAAG is preferentiallylocalized in the equatorial segment and the middle piece of the sperm tail, while theseminolipid is widely distributed on the membrane of the spermatozoa. These data indicatethat human sperm express on their surface membrane a glycolipid similar in structure togalactosylceramide, the receptor for HIV identified in the CD4− cells, that could function as

* Corresponding author.

0165-0378/98/$ - see front matter © 1998 Elsevier Science Ireland Ltd. All rights reserved.PII S0165-0378(98)00060-6

A. Brogi et al. / Journal of Reproducti6e Immunology 41 (1998) 213–231214

a HIV receptor and possibly be implicated in its transmission. © 1998 Elsevier ScienceIreland Ltd. All rights reserved.

Keywords: gp120; HIV receptor; Human sperm

1. Introduction

Human semen is the main vehicle for sexual transmission of the humanimmunodeficiency virus (HIV); however, the role of the sperm cells ascarriers of HIV has been long challenged due to several discordant datareported in the literature. Using the same methodologies (electron mi-croscopy or polymerase chain reaction (PCR) for detection of viral DNA orRNA, different laboratories have obtained contrasting results (Borzy et al.,1988; Bagasra et al., 1988, 1990; Bagasra and Freund, 1990; Pudney, 1990;Anderson et al., 1990; Baccetti et al., 1991; Mermin et al., 1991; VanVoorhis et al., 1991; Anderson et al., 1992; Scofield et al., 1992; Baccetti etal., 1994, 1996). Dussaix et al. (1993) have shown that spermatozoa infectedin vitro and then cocultured with CD4+ T lymphocytes are capable oftransmitting the infection. Using immunoelectron microscopy and PCRanalysis, Baccetti et al. (1994, 1996) were able to demonstrate the presenceof HIV in germ cells but also that virus particles can be transferred in vitrointo oocytes. An in situ hybridization study preceded by polymerase chainreaction on testicular tissue obtained from AIDS patients showed thatHIV-1 selectively infects spermatogonia, spermatocytes and the rare sper-matids found in the testes of infected people (Nuovo et al., 1994). Using thesame technique, Bagasra et al. (1994) have detected proviral DNA in thesperm of infected men at various stages of the disease.

The modality of entry of HIV into the sperm cells remains a debatedquestion since the presence of CD4 on the membrane of spermatozoa hasalso been a contradictory issue (Wolff and Anderson, 1989; Anderson et al.,1990; Bagasra et al., 1990; Gobert et al., 1990; Anderson, 1992; Scofield,1992). Alternatively it has been proposed that sperm can bind HIV throughthe mechanism related to class II MHC-CD4 molecules (Ashida andScofield, 1987; Miller and Scofield, 1990; Scofield et al., 1994). The data ofDussaix et al. (1993) seem to have ruled out the presence of the CD4 ongerm cells, since the infection in vitro of spermatozoa was not blocked bypreincubation with anti-CD4 antibodies. The absence of CD4 epitopes intesticular tissue of infected men has been reported by Nuovo et al. (1994).

We have previously reported (Brogi et al., 1995, 1996) that a glycolipidmolecule, present on the surface membrane of human spermatozoa, inter-acts specifically with the gp120 and suggest that this molecule could

A. Brogi et al. / Journal of Reproducti6e Immunology 41 (1998) 213–231 215

function as receptor for HIV in human spermatozoa. On the basis of thebiochemical analysis (Brogi et al., 1996), this glycolipid molecule has beenidentified as galactosylalkylacylglycerol, the non-sulfated form of the semi-nolipid, a sulfoglyceroglycolipid known to be present in the testes andspermatozoa of mammals (Ishizuka et al., 1973). These results have beenconfirmed and extended by Gadella et al. (1998) who have shown thepenetration of recombinant gp120 into monolayers containing hydroxy fattyacid (HFA)-galactosylceramide or galactosyl-alkyl-acylglycerol (GalAAG)by surface pressure experiments.

2. Materials and methods

2.1. Immunofluorescence

Fresh motile sperm, obtained from healthy donors, were selected byswim-up migration. Semen samples were washed once in phosphate bufferedsaline (PBS) and centrifuged at 500×g for 10 min. The pellet fraction,containing the spermatozoa, was removed and normal goat serum (NGS)added to a final concentration of 3%. The sperm cells, kept in suspension bygentle agitation, were incubated on ice with the monoclonal antibody MabO1 (Sommer and Schachner, 1981). Incubation with the primary antibodieswas carried out on ice to prevent internalization. After 20 min the sperma-tozoa were washed three times in PBS containing 3% NGS, smeared onglass slides and fixed for 15 min in 4% paraformaldehyde (PFA) on ice.Glass slides were then washed in PBS and incubated at room temperature(RT) for 20 min with a FITC conjugated goat anti-mouse IgM m chainspecific (Sigma) diluted 1:64 in PBS containing 3% NGS. After washing inPBS, the glass slides were mounted and specimens observed in a LeitzAristoplan light microscope.

2.2. Lipid extraction

Lipids from human spermatozoa were extracted as reported by Ishizukaet al. (1973). Semen, obtained from healthy donors, was washed three timesin PBS, the pellets resuspended in PBS, pooled and 20 volumes of chloro-form/methanol (2:1 v/v) were added. The lipid extract was filtered and theresidue extracted again with 10 volumes of chloroform/methanol (1:1 v/v).To the extract, separated by filtration, the chloroform/methanol 2:1 ratiowas re-adjusted by adding the appropriate amount of chloroform. The firstand second extracts were washed separately with 1/5 and 1/4 of 0.25 MNaCl, respectively. The lower organic phases were pooled and concentrated


to dryness under a stream of N2. The lipid extract was separated on TLCSilica gel 60 (Merck) using chloroform/methanol/water as solvent (65:25:4by volume). The fractions of interest were purified by preparative thin layerchromatography (TLC) using the solvent system described above. Lipidswere located by exposing the sides of the plate to iodine vapors while theremaining part of the plate was covered with aluminum foil. The unexposedband corresponding to the lipid fraction of interest was scraped and elutedfrom the silica gel by extraction with chloroform/methanol (ratios from 2:1and 1:1 v/v); each extraction was carried out for 10 min and centrifuged at2500×g for 10 min. The purified pooled fractions were dried under N2 andresuspended in chloroform/methanol 2:1 (v/v). Fractions were aliquoted,dried under N2 and stored at −80°C. A new aliquot was used for eachexperiment.

2.3. ELISA studies

Stock solutions of cholesterol and phosphatidylcholine (PC) were pre-pared at 100 mg/ml, kept at −40°C, and diluted 1:10 before use. The lipidmixture, to be used in the ELISA assay, was prepared by mixing a solutioncontaining cholesterol and PC (1:3 v/v) and 60 m l were added to a tubecontaining 1 mg of the GalAAG, for convenience called the F6 fraction. Thecombined lipid mixture was diluted to 1 ml with ethanol and used asantigen. In the final solution, the lipids were present in the following ratio:cholesterol/PC/F6 (0.15:0.45:1 w/w/w).

Microwells (Titertek Microplate cod. 78.590.03 Flow) were coated with50 m l of the lipid mixture and evaporated under the hood with gentleagitation. The wells were then saturated overnight at RT with 250 m l ofTris–HCl 10 mM pH 7.4, 150 mM NaCl, 1% BSA, 0.1% merthiolate(TBSM-BSA) and then incubated for 2 h at 37°C with 100 m l of themonoclonal antibodies Mab O1, Mab O4 (an antisulfatide which alsorecognize seminolipids) (Bansal et al., 1989; Gadella et al., 1994), or dilutedgp120, as indicated. Two other antigalactosylceramide antibodies were alsoused: the monoclonal IgG from Boehringer Mannheim (Indianapolis, IN)and the polyclonal from the Advanced Immunochemical Service (LongBeach, CA). Wells were washed five times with TBSM-BSA containing0.02% Tween 20 with an automatic ELISA Washer 2 (Sclavo). The bindingof Mab O1 and Mab O4 was measured using horseradish peroxidase (HRP)conjugated affinity purified goat anti-mouse m chain specific diluted 1:1000(Sigma). The polyclonal sheep anti-gp120 (MRC AIDS Directed Pro-gramme Reagent Repository) was purified and HRP conjugated in ourlaboratories, and used at a dilution of 5 mg/ml, with incubation for 2 h at37°C. The wells were washed five times with (TBSM-BSA) containing 0.02%


Tween 20 then developed for 10 min with 100 m l of tetramethylbenzidine(TBM) (Liquid Substrate System Sigma). The reaction was stopped with anequal volume of 2N H2SO4. Optical densities were measured at 450 nm inELISA Microtiter Reader SR 400 AT (Sclavo). All determinations were runin triplicate. Blanks and control wells coated with the mixture of choles-terol/PC in the absence of the F6 fraction were always included in eachassay plate.

2.4. Immunoelectron microscopy

Fresh motile sperm were washed in PBS, and incubated at 4°C for 20 minin PBS containing 3% NGS. The sperm samples were then centrifuged,resuspended in PBS-BSA 0.1%, NGS 1% containing the Mab O1 or MabO4 diluted 1:10 and incubated for 30 min at 4°C to prevent internalizationof the antibodies. After careful washing in PBS, the samples were fixed inPFA (2% in PBS) and then treated with glycine (0.2 M in PBS) for 30 minat RT and, after centrifugation, incubated with the secondary antibody IgM(m chain specific) conjugated with 10 nm colloidal gold (Sigma) diluted 1:10in PBS-BSA 0.1%, NGS 1% for 2 h at RT. After washing the samples werepostfixed with 2.5% gluteraldehyde–PFA 2% in PBS for 1 h at 4°C andpost-fixed in 1% osmium tetroxide in cacodylate buffer 0.1 M pH 7.2. Thesamples were dehydrated and embedded in Epon-Araldite. Ultrathin sec-tions were cut with an Ultranova Reichert Jung microtome and observed atthe electron microscope (Philips CM10).

2.5. Immonohistochemistry of the testicular tissue

Testicular tissue from a 22-year-old man who had undergone surgery fortumor of the testis, was fixed for 2 h in 1% glutaraldehyde, 0.15 Mphosphate buffer pH 7.4 and washed in the same buffer containing increas-ing concentrations of sucrose (5–20%) (PS). The tissue was embedded inPS/AmesTM O.C.T. compounds diluted 2:1 in buffer and frozen at −30°C.Cryosections 4–8 mm thin, were air dried, treated with 0.2 M glycine in PBSand then with PBS 1% BSA, 5% NGS. The sections were incubatedovernight at 4°C with the Mab O1 diluted 1:40 and then processed asdescribed in Section 2.1.

2.6. Primary cultures of rat Sertoli and germ cells

Mixed cultures of Sertoli cells and germ cells were prepared as describedby Ziparo et al. (1980) and plated on glass coverslips. After 5 days thecoverslips were washed in PBS, fixed in 3% PFA and incubated with theMab O1 as described above.


3. Results

Indirect immunofluorescence microscopy using the anti-galactosylce-ramide antibody Mab O1 which recognized the GalAAG on live spermcells, obtained by the swim-up procedure, showed the positive staining to belocalized on the middle piece of the tail (Fig. 1A) and the equatorialsegment (Fig. 1B). It was estimated that 30–35% of the spermatozoa werestained. A different localization of the seminolipid was evident when thesperm cells were stained with the anti-sulfatide (Mab O4). The seminolipidappears to be distributed equally well on the head region and on the tail ofthe spermatozoa (Fig. 1C). The different localization of the two glycolipidson the membrane was confirmed by immunoelectron microscopy. As ob-

Fig. 1. Immunofluorescence of human sperm stained with Mab O1 and Mab O4. Live cellswere incubated with Mab O1 or O4 diluted 1:20, and binding was detected with fluorescein-conjugated goat antimouse IgM. (A) Binding of the O1 antibodies is localized on the anteriorportion of the middle piece of the tail, and (B) on the equatorial segment. (C) Binding of theO4 antibodies is distributed on the entire surface of the sperm cells. Magnification, ×1500.


Fig. 2. Localization of the O1 antibodies in immunoelectronmicroscopy. (A) Cross-sectionand (B) longitudinal section of the sperm tail. The gold particles are exclusively localized onthe surface membrane of the spermatozoa. No gold particles can be detected at the level ofthe principal piece. AX, axomeme; M, mithochondria; pp, principal piece. Magnification: (A)×58500, (B) ×31500.

served in immunofluorescence, the GalAAG is mainly localized on themiddle piece of the tail (Fig. 2A, B), and the equatorial segment (notshown) while the seminolipid can be detected on the entire surface mem-brane (Fig. 3A, B, C). The density of the gold particles (Fig. 3A, B, C, D)bound to the membrane indicates that the seminolipid (SGalAAG) isexpressed at a higher concentration compared to the GalAAG. Moreover,the immunoelectron microscopy clearly shows that both lipids are presenton the surface membrane of the spermatozoa such that the functionalepitopes are available for interaction with the gp120.

To identify the glycolipid fraction recognized by the antibody in immu-nofluorescence, we performed an immuno-TLC of the total lipid extract andfound that the Mab O1 reacted with a glycolipid fraction (as determined bythe orcinol staining) that, according to the mobility on TLC and based onadditional biochemical analysis, was classified as the GalAAG (data notshown) (Brogi et al., 1996).

The specificity of the binding of the Mab O1 to the lipid fraction wasanalyzed utilizing the ELISA method developed in our laboratory (Brogi etal., 1996). In this assay, the lipid antigen is immobilized on the solid phasein mixture with the non-antigenic lipids (cholesterol and PC in fixed ratio)


(Frey et al., 1993), improving the orientation and accessibility of theantigenic lipids in aqueous phase, and thus favoring both specificity andreproducibility of the assay.

Saturable curves of binding were obtained when increasing concentra-tions of Mab O1 were added to microwells containing constant amounts ofthe antigenic lipid mixture. The binding was detectable at 5–10 pmol/mland saturable at 80–100 pmol/ml (Fig. 4A). The same concentrations of anon-correlated purified IgM antibody did not bind to the glycolipid fraction

Fig. 3. Localization of the O4 antibodies in immunoelectronmicroscopy. Gold particles aredistributed: (A) on the apical region (×30000), (B) on the plasma membrane surroundingthe mitochondrial helix (×23500), (C) on the plasma membrane of the principal piece(×43000), and (D) magnification of a piece of the sperm tail shown in (C) ×78000.Cross-sections of the tail show a considerable concentration of the gold particles indicatingthat the seminolipid is present in high concentration. N, nucleus; A, acrosome; M,mithochondria; Ax, axomeme.


Fig. 4. Binding of the Mab O1 and gp120 to F6. Microwells were coated with 50 m l of asolution containing cholesterol:PC:F6 (0.15:0.45:1 mg/ml), binding was assayed in ELISAand detected as described in Section 2. The lipid mixture was incubated with increasingconcentrations of: (A) Mab O1 (�), or an IgM non-correlated antibody (anti-IL-1) (D)ranging from 2 to 100 pmol/ml. (B) gp120 ranging from 2 to 80 pmol/ml (�). (A, B) Controlwells containing non-antigenic lipid mixture (cholesterol/PC 0.15:0.45 mg/ml) (�). The datarepresent the mean9S.E. of two independent experiments done in triplicate.

(Fig. 4A), indicating that the binding of the Mab O1 to the lipid fraction isspecific. The non-antigenic lipid mixture run as control did not react withthe antibody at any of the concentrations used (Fig. 4A).


Binding of the gp120 to the GalAAG could be expected, based on thestructural similarity with the galactosylceramide. This hypothesis was ini-tially confirmed in immuno-dot blot and then in the ELISA assay (Brogi etal., 1996). The results obtained show that increasing concentrations ofgp120 bind in a dose dependent fashion to the F6 lipid fraction. Thebinding was detectable at a concentration less than 5 pmol/ml (Fig. 4B) andmaximum binding capacity was reached at 100 pmol/ml, indicating that thebinding of the gp120 to the F6 fraction was due to saturation of specificbinding sites.

In order to exclude that minute amounts of the seminolipids couldcontaminate the F6 fraction and be responsible for the binding of the gp120(Vos et al., 1994), an ELISA assay was carried out utilizing the Mab O4which is specific for sulfatides and seminolipids and does not recognize thegalactosylceramide or the GalAAG (Bansal et al., 1989; Gadella et al.,1994). The results obtained (Fig. 5) show that increasing concentrations ofthe Mab O4 bind to the seminolipid standard, but not to the F6 immobi-lized on the same plate, indicating that binding of the gp120 is not due tocontaminating seminolipids.

To determine if the binding of gp120 depends on a specific gp120–F6interaction, increasing concentrations of unlabeled gp120 were incubated

Fig. 5. Binding of the Mab O4 to seminolipid standard and F6. Microwells were coated with50 m l of the seminolipid standard (1 mg/ml) or F6 (1 mg/ml) in mixture with the non-anti-genic lipids cholesterol/PC (0.15:0.45 mg/ml), binding was assayed by ELISA and detected asdescribed in Section 2. Binding of increasing concentrations of Mab O4 ranging from 0.5 to500 pmol/ml to seminolipids (�) and F6 (�). Each value is the mean of two independentassays run in triplicate9S.D.


with HRP labeled gp120. Binding of gp120 was inhibited in a concentra-tion-dependent manner; in fact, an equimolar amount of unlabeled gp120(10 pmol/ml) was sufficient to inhibit by 50% the binding of the HRP gp120to the F6 fraction, while an excess amount (100 pmol/ml) completelyinhibited the binding of HRP labeled gp120 (Fig. 6A). When the Mab O1was tested for its ability to inhibit the binding of the gp120 to the F6fraction, incubation of gp120 with increasing concentrations of the antibodycompletely blocked the binding of the gp120 to the F6 fraction (Fig. 6B). Infact, 100% of inhibition was obtained with 150 pmol/ml of antibody (Fig.5B). Binding of the gp120 to the F6 fraction was not blocked by anon-specific IgM (Fig. 6B). The ability of the gp120 to inhibit the bindingof the Mab O1 was also tested. However, due to the limited amounts ofgp120 available, an excess of gp120 could not be used, but an inhibition of30% was observed when the gp120 was used at the concentration of 100pmol/ml (data not shown). These experiments demonstrate that the interac-tion of the gp120 and the Mab O1 to the F6 fraction is specific and that thebinding site for the two ligands on the F6 fraction is identical. The identityof the binding site of the GalAAG with that of galactosylceramide is furthersupported by the competition studies between the IgG monoclonal antibod-ies known to block the HIV infection in epithelial cell lines (Fantini et al.,1993) and the Mab used in these studies.

To determine if the glycolipid molecule indicated as a potential receptorfor HIV-1 is also present on immature germ cells, human testes wereanalyzed by immunohistochemistry for the binding of the Mab O1.Cryosections of testicular tissue treated with the Mab O1 show clearpositive staining of the seminiferous tubules preferentially localized on themembrane of large round cells at the periphery of the tubule, where thespermatogonia are located (Fig. 7A). Lamina propria and Sertoli cells werenot stained. In the absence of the primary antibody, no positive stainingwas observed (Fig. 7C). Light micrographs of the tubules show a relativelypreserved tissue where immature germ cells can be identified (Fig. 7B, D).The specificity of the binding of the antibodies to the germ cells was alsodemonstrated in mixed cultures of rat spermatogonia and Sertoli cells. Thespermatogonia, growing on top of the Sertoli cells are intensively stainedwith antibodies, while no staining of Sertoli cells is observed (Fig. 8B).

4. Discussion

The most common method of transmission of AIDS involves transfer ofHIV via semen. However, the cell source and the molecular mechanism


Fig. 6. Competition assay. Microwells were coated with the purified F6 fraction in mixturewith non-antigenic lipids and incubated: (A) with 10 pmol/ml of HRP labeled gp120 in thepresence of increasing concentrations of unlabeled gp120; (B) with 200 pmol/ml of unlabeledgp120 in presence of increasing concentrations of Mab O1 (�) and non-correlated Mab IgM(�). Binding was assayed by ELISA and detected as described in the text. Each value is themean of two independent assays run in triplicate9S.D.

involved are not yet unequivocally identified, probably due to the longlasting controversy on whether sperm should be considered potential carri-ers of HIV-1.


Recently, several laboratories have identified the virus or the pro-viralDNA in spermatozoa isolated from AIDS patients and/or from spermsinfected in vitro (Bagasra et al., 1994). Moreover, the infection does notseem to be limited to the sperm cells but viral nucleic acids have beendemonstrated in the spermatogonia and their progeny (Nuovo et al., 1994),suggesting that the sperms could acquire the HIV-1 during spermatogenesis.It seems unlikely that infection of the spermatogonia is the result ofadvanced stages of AIDS since the virus was not detected in the Sertolicells, Leydig cells or endothelial cells.

The modality of interaction and entry of HIV in the spermatozoa remainsto be clarified since the data of Dussaix et al. (1993) seem to have ruled outthe involvement of CD4, by demonstrating that preincubation of spermato-zoa with an anti-CD4 antibody is not sufficient to block the infection.

Fig. 7. Immunocytochemistry of human seminiferous tubules with Mab O1. Cryosections4–8 mm thin were incubated with Mab O1 diluted 1:40 and binding was detected withflurescein-conjugated goat antimouse IgM. (A) The fluorescence is localized on the mem-brane of round cells at the periphery of the tubules where the spermatogonia are located(arrow). (C) In the absence of primary antibody, no staining was observed. ((B), (D)) Lightmicrographs of the cryosections show a relatively preserved tissue where some of thespermatogonia can be recognized (arrow). Magnification, ×650.


Fig. 8. Localization of the O1 antibodies in mixed cultures of rat germ and Sertoli cells. (A)The spermatogonia are the only cells recognized by the antibodies. Sertoli cells seen in (B)are not stained with the antibodies. Magnification, ×1500.

We have previously reported that sperm cells express a glycolipid struc-turally related to galactosylceramide on their surface membrane (Baccetti etal., 1994; Brogi et al., 1995, 1996); we have now identified, in the total lipidextract prepared from human sperm, a glycolipid component capable ofbinding the gp120. Biochemical characterization suggests that this glycolipidcontains a galactose residue linked to the glycerol backbone. Two classes ofglyceroglycolipids have, so far, been identified in the mammalian tissue: the


mono and digalactosyl diglycerides, mainly present in the nervous tissue andthe seminolipids present in testes and spermatozoa (Ishizuka et al., 1973).Based on the chromatographic mobility on TLC, this glycolipid has beenclassified as GalAAG, the non-sulfated form of the seminolipid (SGalAAG)(Brogi et al., 1995, 1996).

The immunocytochemical analysis performed on fresh motile spermato-zoa showing preferential localization of the staining of the Mab O1 in themiddle-piece of the tail completely agrees with the data recently reported byBagasra et al. (1994) who have shown by IS-PCR that sperm cells stainpositively for HIV-1 in the middle piece. Furthermore, the localization ofthe GalAAG in the equatorial region (Fig. 1B) supports a functional rolefor this lipid since this region has been identified both as the fusogenic siteof the sperm plasma membrane as well as the binding site of the gp120 (Artset al., 1994; Gadella et al., 1998).

The functional role of GalAAG on the sperm membrane is also sup-ported by the immunocytochemical data obtained with the cryosection ofhuman seminiferous tubules which show that staining of the O1 antibodiesis restricted to the germ cells and does not involve Sertoli cells (Fig. 7A).These findings correlate very well with the localization of viral nucleic acidsin spermatogonia and their progeny, but not in the accessory cells of testesof AIDS patients (Nuovo et al., 1994).

The specificity of the binding of this glycolipid to the Mab O1 and thegp120 has been determined using a modified ELISA system, which favorsthe sensitivity and reproducibility of the assay. Our data unequivocallyshow that the binding of the Mab O1 and the gp120 to the F6 fraction isspecific; in fact, the molar ratio between the antigen, the ligands and theligands in competition are indicative of highly specific binding. The resultsobtained with immunoelectron microscopy clearly show that bothSGalAAG and GalAAG are located, although distributed differently, onthe surface membrane of the spermatozoa. These data have been confirmedand extended by Gadella et al. (1998) who have demonstrated by surfacepressure experiments that lipid monolayers containing HFA-ceramide andGalAAG were as effective in the insertion of gp120, while monolayerscontaining SGalAAG were less effective.

The binding of gp120 to liposomes containing GalCer, and other struc-turally related glycosphingolipids GalAAG, SGalCer, and SGalAAG hasindicated that conformation is important for the gp120 recognition (Gadellaet al., 1998).

In summary, we have demonstrated the presence, on the surface mem-brane of the human sperm cells, of a glycolipid molecule that couldfunction, based on the structural similarity with the galactosylceramide, as


a HIV-1 receptor, assigning to the GalAAG and SGalAAG the role ofalternative receptors for HIV-1, a role now well established for GalCer,SGalCer (Bath et al., 1991; Harouse et al., 1991; Yahi et al., 1992; Fantiniet al., 1993). These results have important implications in understanding themodality of the HIV-1 entry in spermatozoa.

Efficient entry of the HIV virus into the host cells is dependent on thebinding of the viral external envelope glycoprotein to CD4, the primaryvirus receptor. However, CD4 alone is not sufficient and one or moreadditional molecules present on the membrane and termed coreceptors areneeded to complete membrane fusion and the virus entry process (Berson etal., 1996; Doranz et al., 1996; Feng et al., 1996). Whether GalAAG alsorequires an accessory molecule or whether it can serve as an accessorymolecule itself remains to be established.

Acknowledgements

We thank Prof. Baccio Baccetti for stimulating discussion throughout thiswork. We also would like to thank the MRC AIDS Directed ProgrammeReagent Repository for providing the gp120 and the polyclonal anti gp120.Giancarlo Gatti and Giuseppe Braga are thanked for the art work.Thiswork was supported by Istituto Superiore di Sanita AIDS Research Project9403-08.

References

Anderson, D.J., Wolff, H., Pudney, J., Whenhao, Z., Martinez, A., Mayer, K., 1990.Presence of HIV in semen. In: Alexander, N.J., Gebelnick, H.L., Spieler, J.M. (Eds.),Heterosexual transmission of AIDS. Wiley-Liss, New York, pp. 167–180.

Anderson, D.J., 1992. Mechanisms of HIV-1 transmission via semen. J. NIH Res. 104,106–108.

Anderson, D.J., O’Brien, T., Politch, J.A., Martinez, A., Seage, G.R. III, Padian, N.,Horsburgh, C.R. Jr., Mayer, K.H., 1992. Effects of the disease stage and zidovudinetherapy on the detection of HIV-1 in semen. JAMA 267, 2769–2774.

Ashida, R.A., Scofield, V.L., 1987. Lymphocyte major histocompatibility complex-codedclass II structures may act as sperm receptors. Proc. Natl. Acad. Sci. USA 84, 3395–3399.

Arts, E.G.J.M., Jager, S., Hoekstra, D., 1994. Evidence for the existence of lipid-diffusionbarriers in the equatorial segment of human spermatozoa. Biochem. J. 304, 211–218.

Baccetti, B., Benedetto, A., Burrini, A.G., Collodel, G., Elia, G., Piomboni, P., Renieri, T.,Sensini, C., Zaccarelli, M., 1991. HIV particles detected in spermatozoa of patients withAIDS. J. Submicrosc. Cytol. Pathol. 23, 339–345.


Baccetti, B., Benedetto, A., Burrini, A.G., Collodel, G., Costantino-Ceccarini, E., Crisa,N., Di Caro, A., Estenoz, M., Garbuglia, A.R., Massacesi, A., Piomboni, P., Renieri,T., Solazzo, D., 1994. HIV-1 particles in spermatozoa of patients with AIDS and theirtransfer into the oocyte. J. Cell Biol. 127, 903–914.

Baccetti, B., Collodel, G., Piomboni, P., 1996. The debate on the presence of HIV-1 virusin human spermatozoa. Drug Discov. Des. 5, 129–142.

Bagasra, O., Freund, M., Weidmann, J., Harley, G., 1988. Interaction of human im-munodeficiency virus with human sperm in vitro. J. Acquired Immune Defic. Syndr. 1,431–435.

Bagasra, O., Freund, M., 1990. In vivo and in vitro studies of HIV-1 and human sperm.In: Alexander, N.J., Gebelnick, H.L., Spieler, J.M. (Eds.), Heterosexual Transmission ofAIDS. Wiley-Liss, New York, pp. 155–161.

Bagasra, O., Freund, M., Condoluci, D., Heins, B., Whittle, P., Weidmann, P., Comito, J.,1990. Presence of HIV-1 in sperm of patients with HIV/AIDS. Molec. Androl. 2,109–125.

Bagasra, O., Farzadegan, H., Seshamma, T., Oakes, J.W., Saah, A., Pomerantz, R.J., 1994.Detection of HIV-1 proviral DNA in sperm from HIV-1-infected men. AIDS 8, 1669–1674.

Bansal, R., Warrington, A.E., Gard, A.L., Ranscht, B., Pfeiffer, S.E., 1989. Multiple andnovel specificities of monoclonal antibodies O1, O4 and R-mAb used in the analysis ofoligodendrocyte development. J. Neurosci. Res. 24, 548–557.

Bath, S., Spitalnick, S.L., Gonzales-Scarano, F., Silberberg, D.H., 1991. Galactosylce-ramide or a derivative is an essential component of the neural receptor for humanimmunodeficiency virus type 1 envelope glycoprotein gp120. Proc. Natl. Acad. Sci. USA88, 7131–7134.

Berson, J.F., Long, D., Donanz, B.J., Rucker, J., Jirik, F.R., Doms, R.W., 1996. A seventransmembrane domain receptor involved in fusion and entry of T-cell tropic humanimmunodeficiency virus type-1 strains. J. Virol. 70, 6288–6295.

Borzy, M.S., Connell, R.S., Kiessling, A., 1988. Detection of human immunodeficiencyvirus in cell-free seminal fluid. J. Acquired Immune Defic. Syndr. 1, 419–424.

Brogi, A., Presentini, R., Piomboni, P., Collodel, G., Strazza, M., Solazzo, D., Costantino-Ceccarini, E., 1995. Human sperm and spermatogonia express a galactoglycerolipidwhich interacts with gp120. J. Submicrosc. Cytol. Pathol. 27, 1–5.

Brogi, A., Presentini, R., Solazzo, D., Piomboni, P., Costantino-Ceccarini, E., 1996. Inter-action of human immunodeficiency virus type 1 envelope glycoprotein gp120 with agalactoglycerolipid associated with human sperm. AIDS Res. Hum. Retroviruses 12,483–489.

Doranz, B.J., Rucker, J., Yi, Y., Smyth, R.J., Samson, M., Peiper, S.C., Parmentier, M.,Collman, R.G., Doms, R.W., 1996. A dual-tropic primary HIV-1 isolate that uses fusinand b-chemokine receptors CKR-5, CKR-3, and CKR-2b as fusion cofactors. Cell 85,1149–1158.

Dussaix, E., Guetard, D., Dauguet, C., D’Ameida, M., Auer, J., Ellrodt, A., Montagnier,L., Auroux, M., 1993. Spermatozoa as potential carriers of HIV. Res. Virol. 144,487–495.

Fantini, J., Cook, D.G., Nathanson, N., Spitalnik, S.L., Gonzalez-Scarano, F., 1993.Infection of colonic epithelial cell lines by type 1 human immunodeficiency virus isassociated with cell surface expression of galactosylceramide, a potential alternativegp120 receptor. Proc. Natl. Acad. Sci. USA 90, 2700–2704.


Feng, Y., Broder, C.C., Kennedy, P.E., Berger, E.A., 1996. HIV-1 entry cofactor: func-tional cDNA cloning of a seven-transmembrane domain. G-protein coupled receptor.Science 272, 872–877.

Frey, W.H. II, Schmalz, J.W., Perfetti, P.A., Norris, T.L., Emory, C.R., Ala, T.A.,1993. Silica-ELISA method improves detection and quantitation of minor glycolipidcomponents in lipid mixtures and of other antigens. J. Immunol. Methods 164, 275–283.

Gadella, B.M., Gadella, T.M.J., Colabrander, B., vanGolds, L.M.G., Lopes-Cardozo, M.,1994. Visualization and quantification of glycolipid polarity dynamics in the plasmamembrane of the mammalian spermatoon. J. Cell Sci. 107, 2151–2163.

Gadella, B.M., Hammache, D., Pieroni, P., Colenbrander, B., van Gold, L.M.G., Fantini,J., 1998. Glycolipids as potential binding sites for HIV: topology in the sperm plasmamembrane in relation to the regulation of membrane fusion. J. Reprod. Immunol. (thisvolume).

Gobert, B., Amiel, C., Tang, J., Barbarino, P., Bene, M.C., Faure, G., 1990. CD4 likemolecules in Human sperm. FEBS Lett. 261, 339–342.

Harouse, J.M., Bath, S., Spitalnik, S.L., Laughlin, M., Stefano, K., Silberberg, D.K.,Gonzalez-Scarano, F., 1991. Inhibition of entry of HIV-1 in neural cell lines by anti-bodies against galactosylceramide. Science 253, 320–323.

Ishizuka, I., Suzuki, M., Yamakawa, T., 1973. Isolation and characterization of a novelsulfoglycolipid, ‘Seminolipid’, from boar testis and spermatozoa. J. Biochem. 73, 77–87.

Mermin, J.H., Holodniy, M., Katzenstein, D.A., Merigan, D.C., 1991. Detection of hu-man immunodeficiency virus DNA and RNA in semen by the polymerase chain reac-tion. J. Infect. Dis. 164, 769–772.

Miller, V.E., Scofield, V.L., 1990. Transfer of HIV by semen: role of sperm. In: Alexan-der, N.J., Gebelnick, H.L., Spieler, J.M. (Eds.), Heterosexual Transmission of AIDS.Wiley-Liss, New York, pp. 147–154.

Nuovo, G.J., Becker, J., Simsir, A., Margiotta, M., Khalife, G., Shevchuk, M., 1994.HIV-1 nucleic acids localize to the spermatogonia and their progeny. Am. J. Pathol.144, 1142–1148.

Pudney, J., 1990. Caveats associated with identifyng HIV using transmission electronmicroscopy. In: Alexander, N.J., Gebelnick, H.L., Spieler, J.M. (Eds.), HeterosexsualTransmission of AIDS. Wiley-Liss, New York, pp. 197–204.

Scofield, V.L., 1992. Sperm as vector and cofactors for HIV-1 transmission. J. NIH Res.105, 108–111.

Scofield, V.L., Clisham, R., Bandyopadhyay, L., Gladstone, P., Zamboni, L., Raghupa-thy, R., 1992. Binding of sperm to somatic cells via HLA-DR: modulation by sulfatedcarbohydrates. J. Immunol. 143, 1718–1724.

Scofield, V.L., Rao, B., Broder, S., Kennedy, C., Wallace, M., Graham, B., Poiesz, B.J.,1994. HIV interaction with sperm. AIDS 8, 1733–1736 [letter].

Sommer, I., Schachner, M., 1981. Monoclonal antibodies (O1 to O4) of oligodendrocytesurfaces. An immunocytological study in the central nervous system. Dev. Biol. 83,311–317.

Van Voorhis, B.D., Martinez, A., Mayer, K., Anderson, D.J., 1991. Detection of HIV-1in semen from seropositive men using culture and polymerase chain reaction deoxyri-bonucleic acid amplification techniques. Fert. Steril. 55, 588–594.


Vos, P.J., Lopes-Cardozo, M., Gadella, B.M., 1994. Metabolic and functional aspect ofsulfogalactolipids. Biochim. Biophys. Res. Commun. 1211, 125–149.

Wolff, H., Anderson, D.J., 1989. HIV in sperm. Hautarzt 40, 737–740.Yahi, N., Baghdiguian, S., Moreau, H., Fantini, J., 1992. Galactosylceramide (or a closely

related molecule) is the receptor for human immunodeficiency virus type 1 on humancolon epithelial HT29 cells. J. Virol. 66, 4848–4854.

Ziparo, E., Geremia, R., Russo, M.A., Stefanini, M., 1980. Surface interaction in vitrobetween Sertoli cells and germ cells at different stages of spermatogenesis. Am. J. Anat.159, 385–388.

..

Documents

New insights into the interaction between the gp120 and the HIV receptor in human sperm (human.sperm/gp120/galactoglycerolipid/antigalactosylceramide/seminolipid/spermatogonia)