Logo of jvirolPermissionsJournals.ASM.orgJournalJV ArticleJournal InfoAuthorsReviewers
J Virol. Nov 2000; 74(22): 10658–10669.
PMCID: PMC110940

Life Cycle of an Endogenous Retrovirus, ZAM, in Drosophila melanogaster


ZAM is an env-containing member of the gypsy family of retrotransposons that represents a possible retrovirus of invertebrates. In this paper, we traced ZAM mobilization to get information about a potential path a retroelement may take to reach the germ line of its host. In situ hybridization on whole-mount tissues and immunocytochemistry analyses with antibodies raised against ZAM Gag and Env proteins have shown that all components necessary to assemble ZAM viral particles, i.e., ZAM full-length RNAs and Gag and Env polypeptides, are coexpressed in a small set of follicle cells surrounding the oocyte. By electron microscopy, we have shown that ZAM viral particles are indeed detected in this somatic lineage of cells, which they leave and enter the closely apposed oocyte. Our data provide evidence that the vesicular traffic and yolk granules in the process of vitellogenesis play an important role in ZAM transfer to the oocyte. Our data support the possibility that vitellogenin transfer to the oocyte may help a retroelement pass to the germ line with no need of its envelope product.

ZAM is a 8.4-kb retroelement that resides within the genome of Drosophila melanogaster (11). On the basis of sequence similarity and gene organization, ZAM is a member of a group of retrotransposons that bears a striking resemblance to the vertebrate retroviruses. These elements are flanked by long terminal repeats (LTRs) that direct the transcription of full-length RNAs representing potential templates for reverse transcription during mobilization. The LTRs flank three open reading frames (ORFs) analogous in position and coding potential to the retroviral gag, pol, and env genes (Fig. (Fig.1).1). Among the diverse classes of eukaryotic retrotransposons, the presence of a third env-like ORF (ORF3) is unique to ZAM and a small group of other members of this family, including gypsy, 297, 17.6, Idefix, and nomad in D. melanogaster (3, 8, 14, 19, 26), tom in Drosophila ananassae (25), Osvaldo in Drosophila buzzatii (15), TED in the lepidopteran Trichoplusia ni (5), and Yoyo in the medfly Ceratitis capitata (28). An envelope protein expressed in vivo has been identified for only three of these elements (gypsy, tom, and TED) (16, 21, 24, 25), and only one of them, gypsy, has been shown to date to have infectious properties (9, 22). Although retroviral Env proteins are known to be involved in viral infectivity through host cell receptor recognition and fusion of viral and cellular membranes, the role of the Env glycoproteins encoded by these elements is still unclear since no budding has ever been visualized for any of them.

FIG. 1
Genetic organization of the ZAM retroelement. The retrovirus-like gag, pol, and env ORFs are flanked by 5′ and 3′ LTRs. The pol and env riboprobes used for in situ hybridization experiments are shown, as are ZAM polypeptides used for preparing ...

ZAM was first identified as a spontaneous insertion at the white locus, giving rise to the wIR6RevI allele in a line of D. melanogaster subsequently called RevI (11). This mutation occurred in the course of a massive amplification of ZAM elements in this line due to their mobilization, which remains active in this stock of flies (3). The existence of RevI and its parental line, wIR6, which displays a low copy number of stable ZAM elements, offers a good genetic system where the control of ZAM mobilization and its relationship with its host genome may be studied. Indeed, we previously reported that ZAM transcription is active in RevI and inactive in wIR6. Two kinds of transcripts similar to mRNAs synthesized from a vertebrate retrovirus involved in a replication cycle were identified in RevI. One corresponds to a full-length genomic RNA, and the other corresponds to a subgenomic transcript of the ORF3 gene able to encode a protein which displays all the features of retroviral envelope proteins. Due to the presence of these transcripts in the course of ZAM mobilization, an important issue is to know whether its entire replication cycle is identical to that of infectious retroviruses and involves an extracellular step before ZAM integration in the genome.

We initially reported that ZAM is mobilized through a reverse transcription process occurring in the germ line of flies (11). In this paper, we searched for tissues where ZAM is transcribed, translated, and potentially assembled in viral particles.

We report that ZAM RNAs are detected in a very specific somatic lineage of cells located around the oocyte in the ovaries of the RevI line. Using polyclonal antisera raised against bacterial ORF1- and ORF3-encoded ZAM fusion proteins, we show that both proteins are coexpressed with the full-length ZAM RNAs in these follicle cells surrounding the oocyte. Furthermore, we provide evidence that particles of ZAM are formed in these follicle cells and pass to the oocyte via the vitelline granule traffic with no apparent need for its Env protein.


Fly stocks.

The wIR6 and RevI strains (low copy number and high copy number of ZAM, respectively) are from the collection of the Institut National de la Sante et de la Recherche Médicale U384.

In situ hybridization.

Embryos at different stages were collected on grape juice agar plates, and fly stocks were maintained on cornmeal-glucose-yeast media at 20°C.

Ovaries and testes were dissected in 1× phosphate-buffered saline (PBS). Dissected ovaries, testes, and embryos were fixed in heptane-saturated 4% paraformaldehyde–0.1 M HEPES (pH 6.9)–2 mM MgSO4–1 mM EGTA for 20 min. Ovaries were rinsed with PBT (PBS, 0.1% Tween 20) before proteinase K treatment was begun. Hybridization with ZAM env or pol digoxigenin-labeled RNA probes was performed at 55°C overnight and was followed by washes in hybridization solution (55.5% formamide, 0.25× SSC [1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate], 0.5 mg of heparin, 0.1 mg of salmon sperm DNA, and 0.1 mg of tRNA/ml, 0.1% Tween 20), in a 1/1 mixture of hybridization solution and PBT at 55°C for 30 min each, and in PBT at room temperature (two washes for 20 min each). The hybridized probe was detected using the Genius kit (Boehringer).

DNA constructs, protein purification, and generation of polyclonal antibodies.

The ZAM gag ORF was amplified with the Expand long-template PCR system (Boehringer) on RevI genomic DNA with oligonucleotide 1 (5′-GAGATCTCAAACAACTCGCTCCGTGTTA-3′; positions 1819 to 1839) and oligonucleotide 2 (5′-GGAATTCCTTCTATGTTGTGTAGCCC-3′; positions 2805 to 2823) (Fig. (Fig.1).1). Oligonucleotides 1 and 2 display at their 5′ ends BglII and EcoRI restriction sites, respectively. The gag PCR product was inserted into the pGEX4T2 vector (Pharmacia Biotech) for glutathione S-transferase (GST)-Gag fusion protein production in the bacterial BL26 strain. The GST-Gag fusion protein was purified by chromatography with glutathione immobilized on cross-linked 4% beaded agarose (Sigma). In order to test the anti-Gag polyclonal antibody, the ZAM gag PCR product was subcloned into the pRSETB vector for His-Gag fusion protein production (see Results).

A 0.7-kb BglII-DraI DNA fragment (Fig. (Fig.1)1) encoding the N-terminal part of ZAM Env protein was subcloned from the BH clone (11) into the pRSETC vector for histidine fusion protein production in the bacterial BL21 strain according to the manufacturer protocol (Invitrogen). The histidine-Env fusion protein was purified according to the manufacturer protocol by chromatography on a nickel affinity resin (Invitrogen).

The purified GST-Gag and histidine-EnvNt fusion proteins were used for generation of polyclonal antibodies in rabbits and rats, respectively (Eurogentec).

Whole-mount immunocytochemistry.

Ovaries were dissected in cold 1× PBS and fixed in 5% formaldehyde–1× PBS–50 mM EGTA–25% (vol/vol) heptane for 20 min. They were treated in methanol and copiously rinsed in 1× PBS. Immunodetections were performed with the ABC-Vectastain kit (Vector Biosys) according to the manufacturer protocol. Primary antibodies (pAbGag or pAbEnv) were added at 1/1,000 dilution. Preadsorbed secondary antibodies (goat anti-rabbit horseradish peroxidase (HRP) or goat anti-rat HRP) were added at a dilution of 1/400. After coloration ovaries were analyzed by optical microscopy (Zeiss Axiophot microscope).

For fluorescence stainings, we have used fluorescein isothiocyanate-conjugated antirat or Texas red-conjugated antirabbit antibodies at dilutions of 1/200 and 1/600, respectively (Molecular Probes). Ovaries were embedded in Mowiol 4.88 (Calbiochem) prepared as described by the manufacturer at pH 8.5. Whole-mount ovaries were scanned on the Leica confocal microscope. Optical sections were 2 μm thick.

Ultrastructural studies.

For ultrastructural studies 2- to 3-day-old flies were dissected in PBS, and the ovaries were quickly fixed for 2 h in ice-cold 5% glutaraldehyde–4% formaldehyde in 0.1 M cacodylate buffer at pH 7.2. Individual ovarian follicles were separated from the ovaries while in the fixative. Following a prolonged rinse in the same buffer, the ovarian follicles were postfixed for 2 h in 1% osmium tetroxide in 0.1 M cacodylate buffer at pH 7.2 and rinsed again in the same buffer. Ovarian follicles were then dehydrated in a graded series of alcohols, passed through propylene oxide, and eventually polymerized in epoxy resin for 3 days at 60°C.

For immunocytochemical detection of viral antigens, ovarian follicles were fixed for 2 h in 1% glutaraldehyde–4% formaldehyde in 0.1 M buffer at pH 7.2. After dehydration in alcohols, ovarian follicles were embedded in Unicryl resins and allowed to polymerize under a UV lamp at 4°C for 3 days. Sections were obtained with an LKB ultramicrotome and mounted over uncoated nickel grids. To detect the presence of viral antigens by gold immunocytochemistry, a number of ovarian follicles were dissected and fixed in formaldehyde and then incubated for 3 h in primary mouse (pAbGag) or rat (pAbEnv) antibodies diluted 1:500 in PBS. Ovarian follicles were then thoroughly rinsed in PBS and incubated for an additional hour at room temperature with either gold-tagged secondary goat anti-mouse immunoglobulin G (20 nM) or antirat (10 nM) antibodies (NCI) diluted 1:200 in PBS. Grids were conventionally stained with uranyl acetate and lead citrate and eventually observed in a Jeol EM transmission electron microscope.


ZAM is transcribed in the somatic follicle cells surrounding the oocyte.

The pattern of spatial expression of ZAM was determined by in situ hybridization with an antisense-specific riboprobe of the ZAM env gene labeled with digoxigenin. (Fig. (Fig.11 and and2;2; see Materials and Methods). This probe potentially recognizes the two ZAM transcripts identified through Northern blot analyses, i.e., the full-length 8.6-kb transcript and the 1.7-kb subgenomic transcript of the env gene (11). A ZAM-specific expression pattern was observed with this probe in the RevI strain, where ZAM mobilization is active. These transcripts were detected in the gonads and in the central nervous systems (CNS) of late embryos (>10 h). No hybridization was detected in early embryos (<8 h). The signal observed in the CNS was detected in almost all the embryos, while the prominent expression in gonads occurred in about 50% of the embryos, suggesting that this signal could be restricted to one sex (Fig. (Fig.2B).2B). When a sense strand-specific probe for the env gene of ZAM was used to probe RevI embryos, no signal was observed in the gonads of the embryos (data not shown).

FIG. 2
In situ hybridization to whole-mount embryos and ovaries to visualize the distribution of ZAM transcripts. (A) Schematic representation of an adult ovariole. The ovariole is composed here of the germarium (early stages of oogenesis) and later of two follicles ...

In situ hybridization of late embryos of the wIR6 strain, in which ZAM elements are stable, did not give any signal with the antisense-specific riboprobe of the ZAM env gene although a very faint hybridization may be detected in gonads after a very long exposure time (Fig. (Fig.2E).2E). This result corroborates those found by Northern blotting analysis (11), indicating that ZAM mobilization is accompanied by elevated RNA levels in RevI.

Since ZAM mobilization is known to occur in the germ line of flies, we then investigated ZAM transcription in the genital apparatus of adult flies. Testes and ovaries were dissected from wIR6 and RevI strains (see Materials and Methods) and subjected to in situ hybridization experiments with the riboprobes described in Fig. Fig.1.1. ZAM RNAs were visualized in RevI ovaries (Fig. (Fig.2C2C and D). Whatever the probe used, no transcript was detected in RevI or wIR6 testes or in wIR6 ovaries (Fig. (Fig.22F).

In insects, ovaries are composed of developing egg chambers arranged in tubular structures called ovarioles (Fig. (Fig.2A).2A). The Drosophila ovary consists of 15 to 18 ovarioles. Each ovariole contains a series of egg chambers at progressively more-advanced stages of oogenesis (10, 23). At the tip of each ovariole, the stem cells of the germ line and the follicle cell precursors reside in the germarium. During oogenesis, the germ line stem cells and follicle cells go through a defined set of division cycles and become organized into egg chambers, which progressively leave the germarium and continue developing as they move posteriorly within the ovariole. The mature egg chamber consists of the oocyte and 15 nurse cells, which are both surrounded by a monolayer of somatic follicle cells (13, 23).

During oogenesis, ZAM transcription occurs very early in the germarium of each ovariole and then is detected in the follicular cells of each egg chamber. However, ZAM RNAs are not present in all the follicle cells but are restricted to a patch of follicle cells located at the posterior side of the oocyte. ZAM expression persists until late stages of oogenesis (Fig. (Fig.2C2C and D). Similar experiments performed on RevI ovaries with the sense riboprobe did not reveal any hybridization signal (Fig. (Fig.22G).

Since the env riboprobe used to detect these transcripts did not allow discrimination between the presence of the full-length transcript and the presence of the ORF3 subgenomic transcript, a specific ZAM riboprobe of the pol gene (Fig. (Fig.1)1) was used for additional in situ hybridizations. The same pattern of expression as the one described above with the ORF3 probe was observed, indicating that full-length ZAM RNAs are present in the follicle cells (data not shown).

We then addressed the question whether proteins encoded by ZAM ORFs could be detected in the cells where ZAM transcripts have been visualized.

The ORF1-encoded ZAM polypeptide is present in cells where ZAM transcription is occurring.

Retroviral Gag proteins are synthesized from full-length RNAs as Gag and Gag-Pol fusion polyproteins and are assembled into virus-like particles in cells where these RNAs are detected. Gag structural polypeptides constitute the core of the viral particle. In order to know whether ZAM Gag proteins are synthesized in tissues where full-length transcripts have been detected, we performed immunocytochemistry experiments.

A purified bacterial GST-Gag fusion protein encompassing the whole length of ZAM Gag was prepared and injected into rabbits (see Materials and Methods). A polyclonal antibody denoted pAbGag, which potentially recognizes the pGag region (Fig. (Fig.1)1) of Gag, was obtained. We verified that the pAbGag antibody is raised against the Gag protein and not exclusively against the GST peptide present in the fusion protein. For that purpose, the gag ORF was subcloned into the Tag-histidine pRSETB vector and a histidine-Gag fusion protein was expressed. From Western blotting experiments, we found that pAbGag clearly reacts with the histidine-Gag fusion protein (data not shown).

We then examined the pattern of spatial and developmental accumulation of Gag products during different stages of Drosophila development where full-length ZAM RNAs had been previously identified. Using pAbGag for immunocytochemical experiments, we detected ZAM Gag proteins in all egg chambers of the RevI strain with the same distribution as ZAM RNAs (Fig. (Fig.3A3A and B). A strong immunostaining was revealed in a few follicle cells located at the posterior part of each egg chamber (Fig. (Fig.3A).3A). At stage 10, Gag immunostaining strongly underlined an area located at the frontier of the follicle cells and the oocyte. This signal tended to extend around the oocyte from stage 10 (Fig. (Fig.3B).3B).

FIG. 3
Expression of Gag and Env proteins during Drosophila oogenesis. Shown is the immunolocalization of Gag and Env proteins using polyclonal pAbGag and pAbEnv antibodies, respectively, in ovaries of RevI and wIR6 females. (A to C) Gag of ZAM revealed with ...

No immunostaining was detected in RevI embryos or larvae or when controls were performed using pAbGag antibody on wIR6 ovaries (Fig. (Fig.3C).3C). Preimmune serum on RevI ovaries did not produce any immunostaining (data not shown).

Translation of ZAM ORF3 is restricted to a defined developmental window of oogenesis.

ZAM encodes a subgenomic mRNA of 1.7 kb whose sequence predicts a protein with structural motifs typical of retroviral Env proteins, i.e., a signal peptide, a potential transmembrane domain, putative N-glycosylation sites, and cysteine residues. In order to determine whether this predicted Env protein is indeed synthesized in tissues where ZAM RNAs have been detected, immunocytochemistry experiments on late embryos, larvae, and dissected ovaries from RevI female strains were performed.

To this end, a bacterial histidine-ORF3 peptide fusion protein, in which the coding sequence for the ORF3 peptide extended from nucleotide 6385 to 7105 of ZAM ORF3, was synthesized (pEnv; Fig. Fig.1).1). After purification by chromatography on a nickel affinity resin, it was injected into rats (see Materials and Methods). A polyclonal antibody denoted pAbEnv, which recognizes the recombinant histidine-ORF3 peptide in Western blot analyses whereas the preimmune serum does not, was obtained (data not shown).

Using pAbEnv, we then determined the temporal and cell type-specific expression of ZAM Env protein in flies. No immunoreactivity was detected with pAbEnv in embryos or larvae (data not shown). Env proteins revealed with pAbEnv were only detected in the ovaries of the RevI strain (Fig. (Fig.3D3D and E). This translation of ZAM ORF3 was restricted to the very small patch of follicle cells surrounding the oocyte where ZAM transcripts were identified. However, the signal was detected in a more defined subset of these somatic cells since it was present at the posterior parts of stages 9 and 10 and not at earlier stages of follicle development. At a higher magnification a strong immunostaining highlighted the area along the follicle cells and the oocyte in addition to Env presence inside the follicle cells (E).

In order to verify the specificity of this immunostaining, controls were performed using the pAbEnv antibody and the preimmune serum on wIR6 and RevI ovaries, respectively. No Env protein was detected in ovaries of the wIR6 strain (Fig. (Fig.3F),3F), and no immunostaining was observed with the rat preimmune serum on RevI ovaries (data not shown).

Gag and Env proteins encoded by ZAM are coexpressed in the same somatic lineage.

To verify that Gag and Env proteins are indeed coexpressed in the follicle cells of RevI egg chambers, we used confocal microscopy on RevI ovaries stained with anti-Gag antibody and anti-Env antibody (Fig. (Fig.4).4).

FIG. 4
ZAM gag and env genes are coexpressed in the follicle cells of RevI follicles. Double staining of RevI follicles with Gag antibody (red) and Env antibody (green) in early stage 10 of oogenesis (upper panels) and in later stage 10 (lower panels). Bars: ...

As expected, double staining revealed that Gag and Env proteins are coexpressed at stages 9 and 10 A of oogenesis in the follicle cells located at the posterior part of the ovarian follicle (Fig. (Fig.4,4, upper panels). The detected fluorescence is specific to these follicle cells and almost absent within the other follicle cells surrounding the oocyte. At this stage of oogenesis, Gag and Env proteins are visualized within the cytoplasm of the cells.

At a later stage, around stage 10 B, of oogenesis (Fig. (Fig.4,4, lower panels). Gag and Env are detected as a thick line visualized at the boundary of the follicle cells and the oocyte. No Gag or Env signal was clearly detected further within the ooplasm of the oocyte or in the nucleus, which is located at the opposite side of this germ cell. At this stage of development, an Env signal persists within the follicle cells while the Gag product is no longer visualized. This picture of the Gag signal detected within the follicle cells and then concentrated at the follicle cell-oocyte border at a later stage of development is consistent with movement of Gag-containing particles between the two.

Viral particles are detected in the follicle cells of the RevI line.

Since all components necessary to assemble particles have been found in a small patch of cells clearly identified, we then searched for potential ZAM particles by an electron microscopy approach. A number of ovarian follicles from Drosophila strain RevI were examined by electron microscopy and compared with those from wIR6 flies. Ellipsoidal or ring-shaped particles about 45 nm in diameter with an electron-translucent center were regularly seen in the posterior follicle cells of ovarian follicles at stages 8 to 10. The region of the follicle cell cytoplasm most highly enriched in viral particles is the one close to the apical plasma membrane (Fig. (Fig.5A5A and B). Some particles have also been occasionally observed inside the nucleus.

FIG. 5
Ultrastructural identification of ZAM viral particles in ovarian follicles of the RevI strain from D. melanogaster. (A) A posterior follicle cell (fc) facing the oocyte (oo) from a stage 9 ovarian follicle is shown. Vm, vitelline membrane. Bar, 0.5 μm. ...

No viral particle was detected in any of the ovarian tissues or intercellular spaces examined in the wIR6 line (Fig. (Fig.5C).5C). This result is in good agreement with data presented above concerning the wIR6 line in which no ZAM mobilization or any ZAM RNA or Gag or Env polypeptides have been observed and strongly supports the idea that the presence of the particles detected in RevI is correlated with ZAM mobilization.

At stages 8 to 10 of oogenesis, Drosophila oocytes accumulate large amounts of yolk. Yolk precursor proteins are synthesized in the fat body and are transported via the hemolymph to the oocyte membrane, where they are subsequently taken up by endocytic vesicles. However, a significant amount of yolk protein is also synthesized in the follicle cells themselves (17).

Interestingly, most of the viral particles detected within the follicle cells have been seen in contact with the membrane enclosing the secretory granules containing the vitelline membrane precursors or even bound to the granule content itself (Fig. (Fig.6C).6C). In the wIR6 strain, no particles are associated with the secretory granules (Fig. (Fig.5C,5C, inset).

FIG. 6
Cytochemical detection of viral particles on vitelline membrane precursors and in yolk granules of the RevI strain. (A) A vitelline membrane precursor (pVm) along the apical end of a posterior follicle cell (fc) from a stage 9 ovarian follicle. Note the ...

Although the mobilization of ZAM occurs within the germ line of RevI, no virus budding along the apical plasma membranes of the follicle cells was observed in this study. In this context, it is interesting to note that the vitelline membrane precursors synthesized within the follicle cells are released from this somatic lineage and pass to the extracellular region bordering the oocyte. Thus, the viral particles may benefit from their association with the vitelline membrane precursors to sort out this somatic lineage. Viral particles detected in more developmentally advanced ovarian follicles are stockpiled along the apical follicle cell plasma membrane, as if extracellular release of residual viral particles would have indeed been impeded by completion of the vitelline membrane (Fig. (Fig.66B).

Viral particules were also detected within the oocyte at stages 8 to 10 of oogenesis. At these stages, the Drosophila oocyte is heavily involved in taking up vitellogenin from the hemolymph by receptor-mediated endocytosis (18). The endocytic apparatus at these developmental stages comprises a plethora of vesicles including coated vesicles, transitional yolk bodies, and mature yolk granules (6). Viral particles can be easily recognized by size and shape among these vesicles in the cortical ooplasm of RevI. Viral particles were located over the yolk granules, where they appear uniformly dispersed along the superficial layer (Fig. (Fig.7C)7C) or enclosed within a vesicular membrane (Fig. (Fig.6C6C and D).

FIG. 7
Immunocytochemical detection of Gag viral antigens. (A) The follicle cell-oocyte border from a stage 9 RevI ovarian follicle tested with anti-Gag antibody. fc, posterior follicle cell; N, follicle cell nucleus; oo, oocyte; Vm, vitelline membrane. Bar, ...

Gold immunocytochemical experiments localize Gag and Env proteins of ZAM at sites where particles accumulate.

To ascertain that the ring-shaped particles observed in the posterior-most follicle cells of the RevI ovaries are indeed due to ZAM expression, a number of ovarian follicles at stages 9 and 10 were treated for the immunocytochemical detection of ZAM proteins. When tested with anti-Gag antibodies, the most heavily labeled sites of the follicular epithelium appeared to be those cells that face the posterior pole of the oocyte (Fig. (Fig.7A).7A). Within the follicle cell cytoplasm gold label accumulated along the apical end, even though the basolateral borders were also labeled to some extent (Fig. (Fig.7A7A and B). Along the apical border, gold particles were preferentially associated with the vitelline membrane precursors or, extracellularly, with the deposited vitelline membrane.

In the cortical ooplasm, gold label appeared dispersed among endocytic vesicles (Fig. (Fig.7C).7C). Yolk granules were also labeled, but the gold particles over these organelles occurred more frequently inside the so-called superficial layer than within the enclosed main body (Fig. (Fig.7D).7D). In ovarian follicles at a more advanced developmental stage of oogenesis than stage 10, the label tended to gradually disappear both from the follicle cell cytoplasm and the oocyte (data not shown).

When tested with anti-Env antibodies, ovarian follicles appeared labeled over both the follicle cell cytoplasm and the oocyte (Fig. (Fig.8A),8A), with gold particles occurring along the apical follicle cell membrane (Fig. (Fig.8B)8B) and the cortical ooplasm among the endocytic vesicles (Fig. (Fig.8C).8C). Label appeared to persist along the oocyte plasma membrane even in ovarian follicles with a complete vitelline membrane and no endocytic uptake (Fig. (Fig.8D).8D).

FIG. 8
Immunocytochemical detection of Env viral antigens. (A) The follicle cell (fc)-oocyte border from a stage 10 RevI ovarian follicle exposed to anti-Env antibody. Gold grains are dispersed over the vitelline membrane (Vm). y, yolk granule. Bar, 0.5 μm. ...

As a general rule, the gold labeling due to anti-Env antibodies is low and does not spatially coincide with viral particles, indicating that the 45- to 50-nm particles may correspond to ZAM particles devoid of an envelope. Although the ZAM Env protein is associated with the plasma membrane, as expected for a functional Env protein, these results suggest that the Gag-Env interaction may not have an obligatory role for the cell-to-cell transmission of ZAM.


The present work reports data about the mobilization of retroviral particles produced in a somatic lineage and passing to the germ line. Analysis of the cell-to-cell transmission of the ZAM retroelement of Drosophila permits us to propose a possible mechanism for such a mobilization.

The mobilization of ZAM correlates with the production of all components necessary to assemble virus-like particles in the follicle cells.

A previous study had reported that ZAM displays all the structural features of a vertebrate retrovirus (11). However, this first study failed to determine whether the predicted products of ZAM were indeed synthesized in the course of its mobilization. This was achieved in this study.

Expression of retroviruses necessitates transcription of a full-length RNA and synthesis of retrovirally encoded proteins Gag, Pol, and Env. All these components encoded by ZAM have been detected in the ovaries of a strain where ZAM mobilization is known to be high, the RevI strain, and are absent in the wIR6 line, in which ZAM mobilization does not occur. Full-length ZAM transcripts have been detected in a group of cells of somatic origin that are the follicle cells surrounding the posterior part of each oocyte.

Polyclonal antibodies raised against the putative full-length Gag protein recognized a ZAM Gag product in adult ovaries in a distribution pattern similar to that of ZAM RNAs. Gag was detected in each follicle starting from early stages of oogenesis. At stages 9 and 10A, the antibody revealed Gag proteins at the border between the oocyte and the follicle cells, in addition to their location in the follicle cells. Later, in stage 10B, Gag proteins were present around the oocyte while they were then absent from the follicle cells. These data are consistent with movement of Gag-containing particles between the follicle cells and the oocyte. Antibodies raised against the ZAM Env led us to visualize the presence of such a protein in a specific group of follicle cells at the posterior part of the oocyte. Although ZAM RNA and Gag proteins have been visualized in these cells, the Env pattern of expression is detected in a more restrictive pattern of development since Env proteins are only present at stages 9 and 10 of oogenesis and are absent in earlier stages.

Virus-like particles of ZAM may benefit from exocytic and endocytic exchanges to pass from the follicle cells to the germ line.

Previous experiments had indicated that novel ZAM insertions frequently occur within the germ line of RevI (11). Owing to the fact that all the components necessary for ZAM's mobilization had been detected in a somatic lineage and that movement of Gag-containing particles had been suggested by our immunocytochemical approach, we then searched for a potential pathway leading ZAM to the oocyte. Through an ultrastructural study, ring-shaped or ellipsoidal viral particles of about 45 nm in mean diameter were detected in RevI. These particles are similar to defective human immunodeficiency virus particles that exhibit an electron-dense ring corresponding to a Gag protein not yet cleaved to yield the mature viral form (7). Several lines of evidence strongly argue that these particles correspond to ZAM particles: (i) they are detected within cells where immunostaining and confocal analysis with anti-Gag and anti-Env antibodies have revealed the presence of ZAM products, i.e., the follicle cells facing the posterior pole of the ovarian follicle; (ii) they are absent in wIR6, where no ZAM mobilization has been observed; (iii) immunogold cytochemistry with antibody pAbGag confirms this staining pattern by showing that the labeling of the follicular epithelium is primarily due to the ZAM product lying close to the vitelline membrane precursors.

This structural analysis brought three pieces of information that help to trace ZAM mobilization. First, the particles occur in close association with the vitelline membrane precursors along the apical cytoplasm of the posterior-most follicle cells. These data indicate that the particles will be able to sort out the follicle cells when these vesicles are secreted. Second, ZAM particles are also detected within the cortical ooplasm, indicating that ZAM particles have been able to pass from the follicle cells to the oocyte. In the oocyte, the particles display a very specific distribution. Indeed, almost all of them are embedded within the superficial layer of the yolk granules along the cortical ooplasm. Third, as for ZAM Gag, Env is observed within the follicle cells specifically along the apical follicle cell membrane. However, no budding within the extracellular compartment between the follicle cells and the oocyte has ever been detected.

These overall data support the following pathway by which ZAM particles enter the RevI germ line. The initial step is to form and accumulate ZAM particles in the follicle cells. At early stage 10 of oogenesis, these particles are secreted along the apical end of the follicle cells in close association with the vitelline membrane precursors. In more developmentally advanced ovarian follicles, extracellular release of residual particles is then impeded by completion of the vitelline membrane leading to viral particles stockpiled along the apical follicle cell plasma membrane. Once released into the follicle cell-oocyte interface, they are transferred to the oocyte and eventually conveyed to the yolk granules, where most of them are detected.

Surprisingly, the scenario deduced from our data supports the idea that ZAM may not need its envelope for an extracellular transmission. It has already been reported that retroviruses may not require their Env proteins for budding to take place. As an example, in polarized cells, the human immunodeficiency virus type 1 Gag protein has been found to direct budding from cell membranes with no necessity for the Env glycoproteins (2). The retroviral Gag proteins play a part in the incorporation of Env into the viral particle, but they also have the capacity for packaging foreign glycoproteins (4, 27). In that context, it is interesting to suggest that Gag proteins of ZAM could recognize the vitellogenin proteins as foreign glycosylated proteins and benefit from their release out of the posterior follicle cells to sort out this somatic lineage.

This potential way for a retroelement to pass from one cell to another may explain the results obtained with retroviruses expressed in a somatic lineage close to the germ line or other retroelements from insects such as gypsy. Indeed, in a genetic context permissive for gypsy mobilization, particles containing gypsy RNA have been described as preferentially clustered along the plasma membranes of the anterior follicle cells. When these cells were tested with anti-Env antibodies, gold labeling appeared almost exclusively associated with the plasma membrane but no viral particle budding or extracellular release from the follicle cells could be observed (12). In addition, recent data from a genetic approach have clearly demonstrated that invasion of the female germ line by gypsy retroviruses may occur in an Env-independent manner (1). Although no direct evidence was presented, the authors proposed that nonenveloped particles might enter the oocyte by endocytosis as a cytochemical tracer. One can predict that the vitellogenic traffic is potentially involved in gypsy mobilization, as suggested for ZAM from our observations.

What is then the function of the env genes of both these elements? Song et al. (21) reported that at least some gypsy elements can be enveloped and display infectious properties. Research into the role of ZAM Env and the formation of enveloped particles will certainly be the next step in understanding the ZAM life cycle. Indeed, if ZAM particles are “homed” by their Gag proteins to regions of the plasma membrane where the Env glycoproteins of ZAM reside, they may have been undetected in our experiments.

When ZAM is in the oocyte, the next step in the ZAM cycle is for it to reach the oocyte nucleus. Although this part of the ZAM cycle remains to be elucidated, the data reported in this paper bring the interesting observation that ZAM Gag may enter the nuclei of the follicle cells. Indeed, gold cytochemistry performed with pAbGag detected a nuclear staining in the follicle cells. These data could indicate that the ZAM Gag structural protein displays a specific motif responsible for directing the protein into the nuclei as already reported for foamy viruses (20). Is such a motif responsible for the entry of ZAM into the nucleus of the oocyte? Future experiments will have to clarify the pathway of ZAM to the oocyte nucleus.


We thank Tom Eickbush for reading a first version of the manuscript and providing helpful suggestions. We are grateful to members of the group, and especially Caroline Conte, for critically reading the manuscript and for valuable comments. We thank Francis Harper for his involvement in a preliminary electron microscopy analysis, Nathalie Gauthier for assistance in obtaining antibodies, M. Grammont for her schematic representation of an adult ovariole, and P. Giraud for assistance with confocal microscopy.

This work was supported by grants from the INSERM (U384), by a project grant from Programme Génome CNRS (Intégrité et plasticité des génomes), and from ARC 1999 to C.V. and partly by the Italian Ministry for Research and Technology in the University (MURST), Program on Cell Interactions, to F.G. and M.M.


1. Chalvet F, Teysset L, Terzian C, Prud'Homme N, Santamaria P, Bucheton A, Pélisson A. Proviral amplification of the Gypsy endogenous retrovirus of Drosophila melanogaster involves env-independent invasion of the female germline. EMBO J. 1999;18:2659–2669. [PMC free article] [PubMed]
2. Coffin J M, Hughes S H, Varmus H E. Retroviruses. Cold Spring Harbor, N.Y: Cold Spring Harbor Laboratory Press; 1997.
3. Desset S, Conte C, Dimitri P, Calco V, Dastugue B, Vaury C. Mobilization of two retroelements, ZAM and Idefix, in a novel unstable line of Drosophila melanogaster. Mol Biol Evol. 1999;16:54–66. [PubMed]
4. Dong J, Roth M G, Hunter E. A chimeric avian retrovirus containing the influenza virus hemagglutinin gene has an expanded host range. J Virol. 1992;66:7374–7382. [PMC free article] [PubMed]
5. Friesen P D, Nissen M S. Gene organization and transcription of TED, a lepidopteran retrotransposon integrated within the baculovirus genome. Mol Cell Biol. 1990;10:3067–3077. [PMC free article] [PubMed]
6. Giorgi F. Coated vesicles in the oocyte. In: Ockleford C D, Whyte H, editors. Coated vesicles. Cambridge, United Kingdom: Cambridge University Press; 1980. pp. 135–177.
7. Goto T, Nakai M, Ikuta K. The life-cycle of human immunodeficiency virus type I. Micron. 1998;29:123–138. [PubMed]
8. Inouye S, Yuki S, Saigo K. Complete nucleotide sequence and genome organization of a Drosophila transposable element, 297. Eur J Biochem. 1986;154:417–425. [PubMed]
9. Kim A, Terzian C, Santamaria P, Pélisson A, Prud'homme N, Bucheton A. Retroviruses in invertebrates: the gypsy retrotransposon is apparently an infectious retrovirus of Drosophila melanogaster. Proc Natl Acad Sci USA. 1994;91:1285–1289. [PMC free article] [PubMed]
10. King R C. Ovarian development in Drosophila melanogaster. New York, N.Y: Academic Press; 1970.
11. Leblanc P, Desset S, Dastugue B, Vaury C. Invertebrate retroviruses: ZAM, a new candidate in Drosophila melanogaster. EMBO J. 1997;16:7521–7531. [PMC free article] [PubMed]
12. Lécher P, Bucheton A, Pélisson A. Expression of the Drosophila retrovirus gypsy as ultrastructurally detectable particles in the ovaries of flies carrying a permissive flamenco allele. J Gen Virol. 1997;78:2379–2388. [PubMed]
13. Margolis J, Spradling A C. Identification and behavior of epithelial stem cells in the Drosophila ovary. Development. 1995;121:3797–3807. [PubMed]
14. Marlor R L, Parkhurst S M, Corces V G. The Drosophila melanogaster Gypsy transposable element encodes putative gene products homologous to retroviral proteins. Mol Cell Biol. 1986;6:1129–1134. [PMC free article] [PubMed]
15. Pantazidis A, Labrador M, Fontdevila A. The retrotransposon Osvaldo from Drosophila buzzatii displays all structural features of a functional retrovirus. Mol Biol Evol. 1999;16:909–921. [PubMed]
16. Pélisson A, Song S U, Prud'homme N, Smith P A, Bucheton A, Corces V G. Gypsy transposition correlates with the production of a retroviral envelope-like protein under the tissue-specific control of the Drosophila flamenco gene. EMBO J. 1994;13:4401–4411. [PMC free article] [PubMed]
17. Postlethwait J H, Giorgi F. Vitellogenesis in insects. In: Browder L W, editor. Developmental biology: a comprehensive synthesis. Vol. 1. New York, N.Y: Plenum Press; 1985. pp. 85–126. [PubMed]
18. Raikhel A S, Dhadialla T S. Accumulation of yolk proteins in insect oocytes. Annu Rev Entomol. 1992;37:217–251. [PubMed]
19. Saigo K, Kugimiya W, Matsuo Y, Inouye S, Yoshioka K, Yuki S. Identification of the coding sequence for a reverse transcriptase-like enzyme in a transposable genetic element in Drosophila melanogaster. Nature. 1984;312:659–661. [PubMed]
20. Schliephake A W, Rethwilm A. Nuclear localization of foamy virus Gag precursor protein. J Virol. 1994;68:4946–4954. [PMC free article] [PubMed]
21. Song S U, Gerasimova T, Kurkulos M, Boeke J D, Corces V G. An Env-like protein encoded by a Drosophila retroelement: evidence that gypsy is an infectious retrovirus. Genes Dev. 1994;8:2046–2057. [PubMed]
22. Song S U, Kurkulos M, Boeke J D, Corces V G. Infection of the germ line by retroviral particles produced in the follicle cells: a possible mechanism for the mobilization of the gypsy retroelement of Drosophila. Development. 1997;124:2789–2798. [PubMed]
23. Spradling A C. Developmental genetics of oogenesis. In: Bate M, Martinez-Arias A, editors. Development of Drosophila melanogaster. Cold Spring Harbor, N.Y: Cold Spring Harbor Press; 1993. pp. 1–70.
24. Szatkowski M, Friesen P D. The Env-like open reading frame of the baculovirus-integrated retrotransposon TED encodes a retrovirus-like Envelope protein. Virology. 1996;226:252–259. [PubMed]
25. Tanda S, Mullor J L, Corces V G. The Drosophila tom retrotransposon encodes an envelope protein. Mol Cell Biol. 1994;14:5392–5401. [PMC free article] [PubMed]
26. Whalen J H, Grigliatti T A. Molecular characterization of a retrotransposon in Drosophila melanogaster, nomad, and its relationship to other retrovirus-like mobile elements. Mol Gen Genet. 1998;260:401–409. [PubMed]
27. Young J A T, Bats P, Willert K, Varmus H E. Efficient incorporation of human CD4 protein into avian leukosis virus particles. Science. 1990;250:1421–1423. [PubMed]
28. Zhou Q, Haymer D S. Molecular structure of yoyo, a gypsy-like retrotransposon from the Mediterranean fruit fly, Ceratitis capitata. Genetica. 1997;101:167–178. [PubMed]

Articles from Journal of Virology are provided here courtesy of American Society for Microbiology (ASM)
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...