Pleiotropic effect due to targeted depletion of secretory rhoptry protein ROP2 in Toxoplasma gondii

The Apicomplexa comprise some of the most significant human pathogens including Plasmodium sp., Toxoplasma gondii and Cryptosporidium parvum; and the notorious veterinary parasites such as Eimeria, Theileria and Sarcosystis. These parasites cause heavy casualties among humans and livestock worldwide. A hallmark of this diverse group of obligate intracellular parasites is the presence of rhoptries. These are club-shaped, membrane-bound and electron-dense secretory organelles that are positioned at the apical end of the invasive stage of the parasite (Perkins, 1992; Sam-Yellowe, 1996). Although rhoptries are arguably the most distinctive secretory organelles in these protozoan parasites, and their secretory products are vaccine candidates, the biogenesis, phylogenetic origin and function of these organelles remain obscure.

T. gondii enters the host cell by an active, orientation-dependent penetration mechanism that is powered by the actin cytoskeleton of the parasite (Dobrowolski and Sibley,1996). Rhoptries discharge at the time of host cell invasion(Aikawa et al., 1977; Carruthers and Sibley, 1997)and the released proteins include those encoded by the rhoptry gene family(ROP1, 2, 3, 4, 6, 7 and 8) that range in size from 42-68 kDa (Ossorio et al., 1992; Joiner and Dubremetz, 1993; Beckers et al., 1994; Beckers et al., 1996). The function of these rhoptry proteins remains unknown, although they are thought to have roles in host cell attachment and invasion, establishment and maintenance of the parasitophorous vacuole membrane (PVM) and replication in the PV. It has been suggested that rhoptries secrete a lytic product or penetration-enhancer factor (PEF) to facilitate T. gondii invasion into the host cell (Lycke et al.,1975; Schwartzman,1986). ROP1 is associated with PEF, but disruption of the ROP1 gene showed no effect on invasion(Kim et al., 1993; Soldati et al., 1995). Similarly, a deletion of RAP1 rhoptry gene in Plasmodium falciparum did not affect parasite invasion or growth(Baldi et al., 2000).

We have previously shown that the T. gondii ROP2 protein is localized to the PVM with its N-terminal domain exposed to the host cell cytosol (Beckers et al., 1994). In vitro import assays of ROP2 deleted in the putative mitochondrial transit sequence in the N-terminus (residues 98-12) support the hypothesis that ROP2 interacts with the mitochondrial import machinery and mediates the tight association of host mitochondria to the PVM (Sinai et al., 2001). However,this model has not been tested in vivo.

Because rhoptries are assumed to be essential for parasite invasion,studies of rhoptry/gene function are especially complicated by the fact that T. gondii cannot survive outside the host cell, thus precluding a gene knockout strategy that ablates parasite invasion. This may potentially explain our inability in previous attempts to generate a ROP2 gene knockout (C. J. M. Beckers and K.A.J., unpublished). In addition, rhoptry genes are often part of multigene families, and/or are tandemly repeated(Beckers et al., 1996), further complicating a gene knockout approach. To circumvent these problems, we applied a recently developed ribozyme-modified antisense RNA strategy(Nakaar et al., 1999; Nakaar et al., 2000) to lower the expression of ROP2 gene. In this review we show that targeted depletion of ROP2 results in: (1) disruption of rhoptry biogenesis and an impairment of cytokinesis; (2) reduction in the association of host cell mitochondria with PVM of the parasite, and a reduced sterol uptake from the host cell; (3) reduced capacity of parasites to invade and replicate in human fibroblasts, and attenuation of virulence in mice. These data suggest that rhoptry discharge, and in particular the release of ROP2, is critical for parasite invasion, replication and parasite-host cell interaction.

A 2.2 kb DNA fragment encompassing the entire coding region of ROP2 was generated from a genomic clone pCBROP2. 1(Beckers et al., 1994) using a set of primers ROP2F (5′-gcctagg GCCTAGGGTACCTGCGCACGATTTAGTGGTC-3′) and ROP2R(5′-gcctaggGAATTCAAGCTCATCGCTGGCCCC-3′) in a PCR. The resultant fragment, flanked by AvrII sites, was cloned in the antisense orientation into the AvrII site of ribozyme-histone cassette, pNTPRZ(Nakaar et al., 1999),generating the construct pASRP2.

A fragment including the chloramphenicol acetyl transferase (CAT) gene was amplified by PCR with T3 (5′-GACTAGTAATTAACCCTCACTAAAGGG-3′) and T7 (5′-GAGCTCCAATTCGCCCGATC-3′) primers using the 220NTP3 cassette as DNA template (Nakaar et al.,1998). This sequence was inserted in a cohesive end ligation to the SpeI site of pASRP2 generating pASRP2-CAT. These constructs were verified by sequencing at the W. M. Keck Sequencing Center, Yale University School of Medicine. Other constructs, pminCAT and pAS-HXCAT, have previously been described (Nakaar et al.,2000).

RH strain of T. gondii was cultivated in Vero or human foreskin fibroblasts (HFF) cells. These cells were routinely grown in MEM andα-MEM, respectively, supplemented with antibiotics, 1 mM L-glutamine and 10% FBS (Gemini Bio-Products, Calabasas, CA). Typically, 10-20 μg of pASRP-CAT purified by a commercially available kit (Qiagen) was used to electroporate 10⁷ tachyzoites that were purified from host cells by syringe passage (21G needle) as previously described(Roos et al., 1994). After 24 hours, 20 μM chloramphenicol was added to the medium(Soldati and Boothroyd, 1993)and the medium containing the selection drug was changed every 3-4 days. When the monolayer of cells neared lysis, the parasites were sub-cultured onto fresh cells. After two passages, the parasites were purified and plated by limiting dilution into 96-well plates. Stable transformants were cloned and amplified into mass cultures.

Parasite replication was assayed by measuring the incorporation of[³H]-uracil (Amersham) into parasites as previously described(Nakaar et al., 1999). Invasion assays were performed using confluent cultures of HFF in 6-well plates that were inoculated with 10⁶ freshly lysed-out parasites. After 30 minutes, most of the parasites that had not invaded were washed off and fresh medium was added. The cultures were incubated overnight at 37°C and parasitophorous vacuoles were counted in 50-100 random light microscopy fields.

Between 10⁵ and 10⁶ parasites were separated on SDS-PAGE and the material was transferred onto nitrocellulose filters. The filters were probed with T34A7 (Sadak et al., 1988), anti-ROP2, 3, 4 antibody and with goat anti-mouse horseradish peroxidase polyclonal antibody as secondary antibody or with anti-NTPase antibody as previously described(Nakaar et al., 1999). Both primary and secondary antibodies were generally used in 1:1000 dilution. Detection was by ECL kit (Amersham Pharmaceuticals Biotech).

Immunofluorescent microscopy was done as previously described(Hoppe et al., 2000). For electron microscopy, monolayers of Vero cells were infected with RH or recently isolated stable clones expressing ROP2 antisense (ROP2AS)for 2-3 days until parasite vacuoles were detectable by light microscopy. Infected cultures are fixed in 2.5% glutaraldehyde in 0.1 M sodium cacodylate(pH 7.4) for 30 minutes and post-fixed in 1% osmium tetroxide in the same buffer. Fixed samples were processed for epon plastic embedding and thin sectioned for transmission electron microscopy using standard protocols. Morphometric analysis and mitotracker labeling of parasites were completed using a procedure similar to that previously described(Sinai et al., 1997).

A monolayer of HFF cells were infected with ROP2AS-1 and ROP2AS-10 parasites for 24-48 hours and fixed in 8% PFA in 0.25 M HEPES, pH 7.4 for 2 days at 4°C. Infected monolayers were scraped and pelleted in 10% fish skin gelatin, infiltrated overnight with 2.3 M glucose at 4°C, and frozen in liquid nitrogen. Ultrathin cryosections were obtained and incubated with T43A7 mAb (1:100), washed, and then incubated with 5 nm protein A-gold Conjugate (1:70). The washed sections were postfixed in 1%glutaraldehyde and contrasted with 1.8% methyl cellulose and 0.5% uranyl acetate and examined by transmission electron microscope.

The protocols of isolation of human LDL (density 1.019 to 1.063 g/ml) from fresh plasma by zonal density gradient ultracentrifugation, preparation of LPDS by ultracentrifugation of fetal bovine serum (FBS) and incorporation of the fluorescent lipid NBD-C into LDL have been described(Coppens et al., 2000).

To visualize fluorescent cholesterol acquired by T. gondii,synchronized infected cells were incubated in culture medium containing 10%LPDS. After 24 hours, cells were labeled with [NBC-C]-LDL and treated as previously described (Coppens et al.,2000). For cytochemical staining of β-hydroxysterols with filipin, infected cells were incubated in culture medium containing 10% FBS,fixed in paraformaldehyde before incubation with filipin and viewed by fluorescence microscope as previously described(Coppens et al., 2000).

All animals were treated and handled according to institutional established guidelines. A set of 10 Swiss-Webster mice (8-10 weeks old) in each group was injected intraperitoneally with 10⁵ parasites of ROP2AS-1,ROP2AS-7, ROP2AS-20 and RH parasites. Survival of mice was monitored twice daily.

To study the functions of ROP2 gene in T. gondii, we employed a recently developed ribozyme-modified antisense RNA strategy(Nakaar et al., 1999; Nakaar et al., 2000). Parasites were stably transfected with an antisense ROP2 gene construct coexpressed with the bacterial CAT gene as a selectable marker (Soldati and Boothroyd,1993). Initially, 50 chloramphenicol-resistant clones were isolated and expanded in mass cultures. Interestingly, most recombinants(>80%) failed to grow after several weeks under selection, suggesting a growth defect, or an inability to infect host cells. Of the surviving recombinants, five were randomly selected and screened by immunoblot for the presence or absence of ROP2 using a mAb T34A7(Sadak et al., 1988) that recognizes mature ROP2, 3, 4 (M_r=55, 59, 60 kDa,respectively). Four of five clones (ROP2AS-1, ROP2AS-7, ROP2AS-10,ROP2AS-20) displayed a dramatic reduction in endogenous ROP2 levels, with ROP2AS-7 and ROP2AS-20 exhibiting the most profound effect,whereas ROP2AS-8 clone was unaffected(Fig. 1A). The magnitude of ROP2 inhibition was comparable to other targeted antisense depletion studies that we have recently described (Nakaar et al., 1999). Although ROP2, ROP4 and ROP8 are related proteins,there are significant sequence differences between ROP2 and ROP4 (36.3%identity, 46.7% similarity) in comparison to ROP8 (73.4% identity, 78.3%similarity). The lack of isoform-specific antibody precluded the evaluation of ROP8 protein expression. However ROP3 and ROP4 levels were not significantly altered. In order to further rule out the nonspecific effects of the antisense RNA on ROP2 expression, we employed two plasmid constructs, pminCAT and pAS-HXCAT (Nakaar et al.,2000). The former serves as an irrelevant vector control. The latter antisense construct has previously been shown to drastically reduce the non-essential gene HXGPRT expression(Nakaar et al., 2000). As expected, stable, chloramphenicol-resistant parasites displayed no apparent reduction in ROP2 levels in either of the two vectors(Fig. 1B). Moreover, the expression of an abundant, soluble and secreted NTPase protein was not affected in these parasites. Taken together, these data confirm the specificity of the antisense approach in abrogating gene expression and is consistent with our previous observations(Nakaar et al., 1999; Nakaar et al., 2000; Ngô et al., 2003).

ROP2 is synthesized as a precursor of ∼64 kDa and processed to the mature 55 kDa form in the secretory pathway of the parasite en route to or upon arrival in the rhoptries (Sadak et al., 1988; Hoppe et al.,2000). Precursor ROP2 did not accumulate in antisense-expressing parasites, indicating as expected that ROP2 gene products were targeted by antisense RNA prior to processing(Fig. 1A). This is consistent with the notion that antisense RNA interacts with its target mRNA with consequent sequence-specific inhibition of gene expression, by a mechanism that may be similar to RNA interference (RNAi). RNAi has been described in diverse organisms including trypanosomes (Ngô, 1998) and much is now known about its mode of action (Cogoni et al., 1999; Dalmay et al., 2000; Djikeng et al., 2001; Bernstein et al., 2001). However, analysis of the genome and EST databases suggested that apicomplexan parasites may not have the machinery that supports RNAi.

Light microscopy indicated that the morphology of ROP2-deficient parasites is atypical in comparison to the characteristic banana shape of their wild-type (WT) counterparts. When host cells were infected, a significant number (>60%) of aborted parasite vacuoles were detected that deviated from the typical rosettes of WT parasites (data not shown). A portion (up to 30%)of the vacuoles appeared morphologically normal at the light microscopic level. Transmission electron microscopy (TEM) confirmed that aborted parasite vacuoles contained grossly enlarged parasites bearing multiple nuclei (2-7)that were arrested in the late stages of cell division(Fig. 2A,B)(Hager et al., 1999). The daughter cells were formed internally and released from the mother cell, but the assembly of the cytokinetic furrow to segregate the maturing cells appeared to be inhibited, as indicated by a large number of empty vesicles aligning at the base of the cytokinetic plane(Fig. 2A,B).

Immunofluorescence microscopy of extracellular(Fig. 2K-R) and intracellular(data not shown) parasites demonstrated both an alteration in the pattern(Fig. 2N-R) and a decrease in intensity (Fig. 2M,O-R) of ROP2 fluorescent labeling in the majority (>80%) of parasites. In some cases,however, the pattern and intensity of ROP2 staining in the antisense clone was not altered in comparison to WT RH (compare panels K and L). These data indicate that ROP2 expression can be specifically abrogated by an antisense RNA strategy, although there is phenotypic heterogeneity in the antisense clones.

Toxoplasma has differential mechanisms of sorting and secreting secretory proteins from the three regulated secretory organelles, i.e. micronemes, rhoptries and dense granules(Ngô et al., 2000). Whereas ROP2-deficient parasites exhibited no morphological defects in the biogenesis of dense granules and micronemes as assessed by qualitative immunofluorescence microscopy and electron microscopy (data not shown), the formation of mature rhoptries was specifically abrogated in greater than 80%of the parasites. In WT parasites, flask-shaped rhoptries were assembled and typically docked in lateral alignment near the apical pole of the cell (data not shown). Depleting the constitutive expression of the putative transmembrane protein ROP2 generated several notable rhoptry phenotypes(Fig. 2,Fig. 2). The extremity of phenotypes correlates with the level of ROP2 depletion, hence ROP2AS-7 and ROP2AS-20 exhibiting the more striking changes than ROP2AS-1 and ROP2AS-10.

In all ROP2AS clones examined, most rhoptries were no longer aligned in linear fashion and were not restricted to the apical half of the parasite (Fig. 2A,K-N). There was a qualitative reduction in rhoptry cellular density, albeit the quantification of organelle number per cell was hindered by the incomplete segregation of parasites. The categorical club shape of mature rhoptries was not formed, or maintained (Fig. 2C-F,H-Jand legends). In parasites with the higher level of ROP2 depletion(ROPAS-7 and ROPAS-20), rhoptries were infrequently observed and they exhibited an irregular shaped basal portion, and the distal neck was aberrant from the typical homogenous electron dense lumen. Parasites exhibiting a lower level of ROP2 reduction (ROPAS-1 and ROPAS-20) contained more rhoptry-like tubules but they were incompletely segregated and appeared as connected clusters(Fig. 2H-I). Immunolabelling by cryoelectron microscopy confirmed that these clusters contained ROP2/3/4 as detected by mAb T34A7 (Fig. 2H-J) and cathepsin B (data not shown). Cathepsin B (TgCP1) is a cysteine protease that is specifically localized to the lumenal core of the rhoptry basal portion (Que et al.,2002). All of these morphologic abnormalities in rhoptry shape were interspersed, even within parasites in the same vacuole, with rhoptries which were morphologically normal (Fig. 2G). Two major conclusions derive from these observations. First,ROP2 appears to be an important determinant in the normal formation of mature rhoptries. Second, parasites tolerate reduction of ROP2 expression levels only to a point, which is on the cusp between formation of morphologically normal and abnormal rhoptries.

To evaluate the hypothesis that secreted ROP2 mediates the tight association between host mitochondria and the PVM(Sinai et al., 1997; Sinai et al., 2001), we determined the association with host cell mitochondria qualitatively by mitotracker staining (Fig. 3A,B) and quantitatively by morphometric TEM(Fig. 3C-E). The association between host cell mitochondria and the PVM was significantly reduced in ROP2-deficient parasites in both analyses; most notable was a 90% reduction in the linear density of PVM membrane that was tethered to host mitochondria(Fig. 3E). The association of vacuolar membrane with endoplasmic reticulum appeared qualitatively unaffected. These results suggest that ROP2 is necessary and may be sufficient for mitochondrial association. Such association could potentially facilitate the salvage from the host cells of lipids for which the parasite is auxotrophic (Foussard et al.,1991).

Cholesterol is concentrated in parasite rhoptries, as determined by both biochemical and morphological data(Foussard et al., 1991; Coppens et al., 2000). We tested the hypothesis that the rhoptries may be involved in the trafficking or storage of lipids from the host cell. Cholesterol content, as detected by filipin, was lower in the ROP2-deficient parasites as compared with WT parasites (Fig. 4, compare panels A-A' with B-B').

We have recently demonstrated that the parasite can efficiently access cholesterol from host lysosomal compartments by an active mechanism, which is independent of vesicular fusion, and requires parasite viability(Coppens et al., 2000). Incubation of infected fibroblasts with fluorescent cholesterol(NBD-C,22-(N-(7-nitrobenz-2-oxa-1,3-diazol-4-yl)-amino)23,24-bisnor-5-cholen-3β)-ol)incorporated into low density lipoproteins (LDL) led to a lower uptake of NBD-C into ROP2AS parasites (Fig. 4, compare panels C-C' with D-D'). Hence, lowering ROP2 levels results in aberrant incorporation of cholesterol into rhoptries.

Antisense ROP2 clones were markedly impaired in cell invasion. To monitor invasion, the number of vacuoles established after infection was determined in three ROP2-deficient clones (ROP2AS-1, ROP2AS-7 and ROP2AS-20). Compared to control parasites, the ROP2-deficient clones produced 6-13-fold fewer vacuoles (Fig. 5A), indicating that antisense clones are not able to efficiently invade human cells. Once intracellular, all the ROP2 antisense clones, in comparison to control parasites, exhibited between 50-90% reduction in their ability to take up [³H] uracil after 24 hours postinfection (Fig. 5B). Although uracil incorporation taken over time is a more accurate measurement of parasite replication, the limited quantities of ROP2-depleted parasites(see below) coupled with the problem of host cell lysis by parasites precluded these experiments. To discriminate between reduced proliferation of intracellular parasites and diminished ability to invade host cell, we monitored the number and size of the vacuoles. In the ROP2AS-7 and ROP2AS-20 clones, the majority of the vacuoles had between 4-8 parasites compared to 16 parasites in the WT(Fig. 5C). These values correspond to 14-21 hours doubling times for the antisense clones, in comparison to approximately 7 hours for WT parasites. Hence, ROP2 deficiency not only compromised invasion of host cells by T. gondii but also its intracellular replication.

To test the effect of the ROP2-deficiency in vivo, naïve mice were challenged with ROP2AS parasites and survival was monitored for two weeks. These parasites are derived from the highly virulent parental RH strain(LD₅₀=1). All the mice in the control died by the sixth day after challenge, whereas in ROP2AS-1 and ROP2AS-20 groups there were no survivors at 8 days (Fig. 5D). In contrast, a third of mice survived a challenge with ROP2AS-7 by day 15 after infection. These data indicate that ROP2 deficiency attenuates the virulence of T. gondii in mice.

ROP2 expression gradually recovered in the ROP2AS clones,generally within 3-4 weeks (Fig. 6). Normalization of ROP2 levels correlated with restoration of the normal rhoptry morphology, organelle association, cholesterol uptake and virulence in mice. These data further confirm the association between low ROP2 levels and the phenotypes illustrated in Figs 2, 2, 3, 4, 5. This observation is also the probable explanation for variability in mouse virulence, and the capacity of even the ROP2AS-7 clone to cause a lethal in vivo infection(Fig. 5D).

Rhoptries are unique secretory organelles that are postulated to be essential to apicomplexan parasitism. The experiments reported here establish that the T. gondii cargo protein ROP2 is a major determinant of proper biogenesis and maintenance of rhoptry structure. The integrity of the rhoptries affects the ability of the parasites to invade and to establish a niche permissive for parasite replication within the host cell. These results are in contrast to the situation with ROP1, because disruption of the ROP1 gene altered neither rhoptry club shape nor parasite virulence(Soldati et al., 1995). Interestingly, a blockage of cell division (endodyogeny) coincides with the ablation of ROP2 expression. These results also provide further support of the involvement of secreted ROP2 in mediating the attachment of host cell mitochondria to the parasite vacuole(Sinai et al., 1997). Such a link has been postulated to serve as a conduit for the bulk transfer of lipids from the host for which the parasite is auxotrophic(Foussard et al., 1991; Coppens et al., 2000). The ROP2-deficient parasites are now less endowed with the ability to salvage sterols from the host cytosol. It is very probable that this reduced ability could result from impairment in parasite replication or from some alteration in the characteristics of the PVM, such as aberrant localization or function of a sterol-binding protein. Concomitant with these phenotypic changes,ROP2-depleted parasites are compromised in terms of virulence in a mouse model. Taken together, these pleiotropic effects are consistent with the essential function of ROP2 gene and the critical role of the rhoptries.

We have recently demonstrated that a tyrosine-based motif (YXXΦ) in the cytoplasmic tail of ROP2 mediates its faithful targeting to mature rhoptries most probably by interacting with the μ1 chain of the AP-1 adaptor complex(Hoppe et al., 2000). Dominant negative interference with the tyrosine-binding pocket of μ1 adaptin abolished its binding to transmembrane rhoptry proteins ROP2 and ROP4, and disrupted the steady state formation of mature rhoptries(Ngô, et al., 2003). Ablation of μ1 adaptin expression by antisense RNA altered rhoptry biogenesis and was detrimental to parasite survival. In contrast, deletion of the gene for the soluble protein ROP1 altered the honeycombed architecture of the rhoptry lumen in the basal portion, but not the defined flask shape and formation of rhoptries (Soldati et al.,1995). In Plasmodium, gene knockout of the rhoptry protein RAP1, which forms a hetero-complex with soluble RAP2 and RAP3,inhibits the delivery of the latter two cargo proteins to rhoptries(Baldi et al., 2000). Similarly, the transmembrane microneme protein MIC6 in Toxoplasma is proposed to complex with soluble MIC2 and MIC4 and is required for the delivery of these soluble proteins to micronemes(Reiss et al., 2001). Although we have no evidence to implicate ROP2 in hetero-complex formation, it is conceivable that other rhoptry proteins can associate with the transmembrane ROP2 to mediate the formation of mature rhoptries. Whether ROP2 functions independently or in conjunction with other proteins, it is intriguing that disruption of a transmembrane cargo interferes with the maturation of a regulated secretory organelle. These data indicate that ROP2 is an essential determinant of rhoptry shape and biogenesis.

Antisense RNA-mediated depletion of either ROP2 transmembrane cargo (this study) or μ1 adaptin (Ngô et al.,2003) disrupts normal rhoptry formation and consequently causes defects in cell division and growth. In contrast to parasites expressing ROP2 antisense RNA, the TGN and endosomal vacuoles are drastically distorted in AP-1-depleted parasites, but the rhoptries maintain their club shape. Although both manipulations generated aborted parasite vacuoles, the specific block in parasite division is distinct. AP-1-depleted parasites complete the formation of cortical membrane complex along the cytokinetic furrow, but the daughter cells are unable to segregate. Toxoplasma acquires cholesterols from the host cell by an endocytosis pathway mediated in part by the endosomal GTPase rab5 (Robibaro et al.,2002) and cholesterols are eventually trafficked to the rhoptries for secretion. Deficiency of cholesterol uptake in ROP2-depleted parasites may explain the lack of lipid materials needed for the synthesis of plasma membrane and cortical cisternae to complete cytokinesis. It remains to be determined whether a critical rhoptry component (e.g. protease) that is required for unzipping the daughter cell surfaces is missorted in AP-1-depleted parasites. Nevertheless, the antisense effects appear to be specific to functions of ROP2 and AP-1 adaptin.

It is important to note that in our experiments, ROP2AS clones in early passage were used. At this point, the diminution of ROP2 expression is very pronounced with the accrual of multiple phenotypes. Although the phenotype is variable, with only approximately 60% of the parasites displaying pronounced morphologic defects at the light microscopic level, all abnormalities observed are probably ascribable to an alteration in ROP2 expression. Supporting this assertion, when parasites were maintained in continuous passage for 3-4 weeks, ROP2 levels were restored leading to the recovery of defects in growth, morphology, association of host mitochondria with the PVM, and mouse virulence (Fig. 6). Hence, it is highly probable that recovery of ROP2 expression occurred after ROP2AS clones were inoculated into mice, explaining the variable and modest alteration in parasite virulence. Because the antisense strategy inherently does not usually generate complete inhibition or complete loss of function of a gene, it is not suitable for the creation of null mutants (Stuart and Wold,1985; van der Krol et al.,1988; Liu et al.,1994; Shi et al.,2000). In the case of ROP2AS, a clonal population expressing a basal level of ROP2 is probably favored for survival. Whether the eventual recovery of ROP2 expression involves an RNA suppressor mechanism or degradation of antisense RNA is not known, but loss of antisense RNA inhibitory effects is in concordance with the data in Leishmania and trypanosomes (Zhang and Matlashewski,1997; Ngô et al.,1998). These data further underscore the utility of an antisense RNA approach to study essential gene function, as deletion of a critical gene produces a lethal phenotype.

ROP2 is expressed at all invasive stages of the parasite (tachyzoite,bradyzoite and sporozoite) potentially reflecting its essential role in parasite biology. In addition, ROP2 and its family are potent humoral and T-cell antigens (Saavedra et al.,1996; Jacquet et al.,1999). It also carries B-cell epitopes because antibodies against the ROP2 antigen are present in more than 85% of T. gondii-seropositive individuals (Van Gelder et al., 1993). These findings, coupled with the essential nature of ROP2 in mediating pathogen-host cell interaction, infectivity and virulence validate this protein as a potential vaccine candidate. It is probable that rhoptry function is conserved in all related coccidian parasites raising the intriguing prospects that these organelles may serve as novel drug targets. The data presented herein may thus open up new avenues for chemotherapeutic and immunologic intervention against these devastating parasitic diseases.

We thank Marc Pypaert and Kim Murphy-Zachichi from the Center for Cellular and Molecular Imaging at Yale University for technical assistance. This work was supported by an NIH Minority Supplement to V.N., and NIH NRSA grants 5T32AI07404 and 5F32AI10044 to H.M.N.; and Public Health Service Grant ROI AI30060 from the NIH and a Scholar Award in Molecular Parasitology from the Burroughs Wellcome Fund to K.A.J.