|Home | About | Journals | Submit | Contact Us | Français|
Entamoeba histolytica is an intestinal protozoan parasite that causes amoebic dysentery and liver abscess. Phagocytosis by the parasite is a critical virulence process, since it is a prerequisite for tissue invasion and establishment of chronic infection. While the roles of many of the proteins that regulate phagocytosis-related signaling events in E. histolytica have been characterized, the functions of lipids in this cellular process remain largely unknown in this parasite. In other systems, phosphatidylinositol (3,4,5)-trisphosphate (PIP3), a major product of phosphoinositide 3 kinase (PI3-kinase) activity, is essential for phagocytosis. Pleckstrin homology (PH) domains are protein domains that specifically bind to PIP3. In this study, we utilized glutathione S-transferase (GST)- and green fluorescent protein (GFP)-labeled PH domains as lipid biosensors to characterize the spatiotemporal aspects of PIP3 distribution during various endocytic processes in E. histolytica. PIP3-specific biosensors accumulated at extending pseudopodia and in phagosomal cups in trophozoites exposed to erythrocytes but did not localize to pinocytic compartments during the uptake of a fluid-phase marker, dextran. Our results suggest that PIP3 is involved in the early stages of phagosome formation in E. histolytica. In addition, we demonstrated that PIP3 exists at high steady-state levels in the plasma membrane of E. histolytica and that these levels, unlike those in mammalian cells, are not abolished by serum withdrawal. Finally, expression of a PH domain in trophozoites inhibited erythrophagocytosis and enhanced motility, providing genetic evidence supporting the role of PI3-kinase signaling in these processes in E. histolytica.
Entamoeba histolytica is an intestinal protozoan parasite that causes amoebic dysentery and liver abscess. A high incidence of E. histolytica infection is found in developing countries and is associated with low basic hygiene standards and a lack of water sanitation (reviewed in reference 20). Since E. histolytica is mainly a human parasite (40), improvements in sanitation may help prevent the fecal-oral spread of this pathogen; however, overpopulation, scarcity of clean water, and socioeconomic shortcomings impede the progress of such improvements in developing countries (36). This lack of progress supports an elevated need for the development of improved prevention, diagnosis, and treatment for dysentery caused by E. histolytica. This requires a better understanding of the basic biology of this parasite.
E. histolytica infections are contracted by ingestion of its multinucleate infective cysts from fecally contaminated food or water (reviewed in reference 39). Upon excystation in the small intestine, motile trophozoites move to the bowel lumen, where bacteria, erythrocytes, and host cell debris serve as food sources which are taken up by phagocytosis. Infection is established when trophozoites adhere to the intestinal wall, destroy colonic epithelium, and occasionally disseminate via the hematogenous route to extraintestinal sites.
Phagocytosis is recognized as an important virulence function in this parasite. For example, several studies suggest a connection between exposure to intestinal bacteria and increased virulence in E. histolytica (3, 46). Furthermore, transcriptional profiling of E. histolytica exposed to Escherichia coli revealed increased gene expression of a protein kinase, an ABC transporter, a Rho family GTPase, and Hsp90, which may collectively modulate virulence in this parasite (6). Finally, phagocytosis-deficient mutants of E. histolytica exhibit reduced virulence (27), and an avirulent Entamoeba species, E. dispar, carries out limited phagocytosis compared to the virulent species (28). Therefore, understanding the molecular mechanisms of phagocytosis in E. histolytica may provide insight into factors that contribute to virulence.
Two recent proteomic screens of purified E. histolytica phagosomes have revealed proteins that may be involved in the processes of phagosome biogenesis in this parasite (22, 26). A putative phosphoinositide 3 kinase (PI3-kinase) was identified as one of the signaling proteins that physically associate with E. histolytica phagosomes. PI3-kinases belong to a family of proteins that generate signaling phosphoinositides (PIs) phosphorylated on hydroxyl groups. These include phosphatidylinositol 3-phosphate (PI3P), phosphatidylinositol (3,4)-bisphosphate (PIP2), and phosphatidylinositol (3,4,5)-trisphosphate (PIP3) (10). Phosphorylated PIs are capable of regulating phagocytosis by recruiting proteins via specific lipid-recognition domains. Examples of such protein motifs include FYVE finger domains, which specifically bind PI3P and pleckstrin homology (PH) domains, some of which specifically interact with PIP3 (reviewed in reference 17). The possibility that PI3-kinase and its products may be involved in the regulation of phagocytosis in E. histolytica is also supported by the finding that small-molecule inhibitors of PI3-kinase block uptake of phagocytic targets (2, 14, 30).
In a previous study, a recombinant glutathione S-transferase (GST)-labeled FYVE finger protein domain was used to localize PI3P in E. histolytica trophozoites (30). It was demonstrated that PI3P accumulated in forming erythrophagosomal cups of E. histolytica. This localization was confirmed by Nakada-Tsukui et al. (24) using live-cell imaging of trophozoites expressing green fluorescent protein (GFP)-labeled FYVE-finger domains. GFP-labeled PH domains have demonstrated that PIP3 regulates phagocytosis in neutrophils (7), macrophages (1), and a nonpathogenic soil amoeba, Dictyostelium discoideum (9, 13). However, little is known about the evolution of PIP3 on cellular membranes in E. histolytica, particularly during phagocytosis. Therefore, in this study, we investigated the spatiotemporal characteristics of PIP3 distribution during endocytosis using both GFP- and GST-tagged biosensors containing a PH domain derived from Bruton's tyrosine kinase (Btk). This PH domain specifically binds to PIP3 (32, 35).
Entamoeba histolytica trophozoites (strain HM-1:IMSS) were cultured axenically in TYI-S-33 (8) in glass screw-cap tubes at 37°C.
The biosensor construct was obtained from a pEGFP-N1 parent plasmid modified by the manufacturer (SignaGen Laboratories, Gaithersburg, MD) to contain the cDNA encoding the Bruton's tyrosine kinase (Btk) pleckstrin homology (PH) domain cloned between the EcoRI and BamHI sites upstream of the sequences encoding enhanced green fluorescent protein (EGFP; herein referred to as GFP). The DNA encoding PHBtk was PCR amplified from the plasmid template using the following pair of primers: 5′-CCGGATCCTCCAGAAAGAAG-3′ and 5′-CCGAATTCGGTTTTAAGCTTCC-3′. The PCR product was subcloned into the pCR2.1-TOPO plasmid vector (Invitrogen, Carlsbad, CA). After digestion with BamHI and EcoRI, the PH domain-encoding DNA fragment was ligated into the polylinker region downstream of and in frame with the sequence encoding GST in the pGEX-5x-1 expression vector (Amersham Biosciences, Piscataway, NJ). Successful construct generation was confirmed by restriction enzyme analysis and sequencing.
GST and GST-PHBtk fusion proteins were expressed in Escherichia coli BL21 (Amersham/GE Biosciences, Piscataway, NJ) and purified using glutathione-Sepharose affinity chromatography as previously described (37). The purity of GST and GST fusion proteins was assessed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and silver staining as described previously (29). Western blot analysis was performed as described previously (44) to verify the authenticity of GST-tagged proteins using a 1:10,000 dilution of anti-GST antibodies (Chemicon, Temecula, CA). The relative concentration of GST or the GST-tagged fusion protein was estimated using spectrophotometry (Biophotometer; Eppendorf, Westbury, NY).
The GFP-PHBtk construct was amplified from the PH-domain-containing pEGFP-N1 template by PCR using the following primers containing BglII and SalI restriction sites: 5′-CCAGATCTAAATGGCCGCAG-3′ and 5′-CCGTCGACTTACTTGTACAGC-3′. The sequence of GFP alone, used for the generation of the control cell line, was obtained from the same plasmid and was amplified using the following primers: 5′-CCGGATCCATGGTGAGCAAGG-3′ and 5′-CCGTCGACTTACTTGTACAGCTCGTC-3′. The amplified products were subcloned into the pCR2.1-TOPO plasmid vector (Invitrogen). Subsequently, DNA encoding GFP-PHBtk or GFP (control) was ligated into the E. histolytica expression vector pGIR209 (31) (gift of W. A. Petri, Jr., University of Virginia, Charlottesville, VA), which had been digested with BglII and SalI. This vector confers G418 resistance on trophozoites and allows the inducible expression of exogenous proteins via the addition of tetracycline to the medium. It is normally cotransfected with a second vector, pGIR308, that confers hygromycin resistance and encodes the tetracycline repressor (31), allowing tetracycline-inducible expression. Successful construct generation was confirmed by PCR, restriction enzyme analysis, and sequencing.
A log-phase culture of E. histolytica was stably transfected with the expression vector pair by electroporation as described previously (15). Transfectants were selected and maintained by the addition of 6 μg ml−1 G418 and 15 μg ml−1 hygromycin to the medium. Stable transfectants were maintained using the same selection agents. GFP or GFP-PHBtk expression was induced by the addition of 5 μg ml−1 tetracycline to the culture medium 24 h prior to experiments. The authenticity of the GFP-tagged protein was verified by Western blot analysis of E. histolytica cell lysates as described previously (30) using a 1:1,000 dilution of mouse anti-GFP antibodies (Invitrogen).
For time-lapse video confocal microscopy of live trophozoites expressing GFP-PHBtk or GFP, cells were incubated on ice for 10 min to release them from the glass surface of the culture tube, pelleted by centrifugation at 500 × g for 5 min at 25°C, and placed in warmed phosphate-buffered saline (PBS) in glass chamber slides for observation using a LSM 510 laser confocal microscope (Carl Zeiss Inc., Thornwood, NY). To observe phagocytosis, GFP-PHBtk-expressing cells and control trophozoites were exposed to human red blood cells (hRBCs) (U.S. Biologicals, Swampscott, MA) during the imaging session. Time-lapse images were collected at 5-s intervals. The instant of adhesion to an erythrocyte was designated time zero (t0). Image analysis was carried out via xyz plots generated with Image J software (National Institute of Health, Bethesda, MD). A supplemental movie was compiled using QuickTime 7 Pro (Apple Computer Inc., Cupertino, CA).
For immunofluorescence (IF) microscopy using GST-PHBtk, E. histolytica trophozoites were incubated with hRBCs (ratio of E. histolytica to hRBCs = 1:100) for 1 min or with lysine-fixable tetramethyl rhodamine isothiocyanate (TRITC)-dextran (0.3 mg ml−1) (10 kDa; Invitrogen) for 2, 10, 30, or 60 min. Cells were subsequently fixed, permeabilized, and stained with recombinant GST or GST-PHBtk as described previously (30). Trophozoites stained with GST alone served as controls. Alexa Fluor 488-conjugated rabbit anti-GST antibody (1:2,000) (Invitrogen) or Texas Red-conjugated goat anti-GST antibody (1:1,000) (Rockland Immunochemicals, Gilbertsville, PA) was used for GST-PHBtk immunodetection.
For comparison, wild-type cells were also stained with anti-PIP3 antibody (Echelon Biosciences, Salt Lake City, UT) per the manufacturer's protocol, with modifications. Briefly, cells were fixed in glass screw-cap tubes for 24 h in 4% (vol/vol) paraformaldehyde in TYI-S-33 at 37°C. Cells were scraped from the glass and harvested by centrifugation (500 × g, 5 min, 25°C). Cells were washed three times in Tris-buffered saline (TBS) and then permeabilized in 0.5% (vol/vol) saponin in phosphate-buffered saline (PBS) for 15 min at room temperature. Cells were washed three times in TBS and blocked for 30 min at 37°C in 10% (vol/vol) goat serum/TBS. Cells were incubated with or without mouse anti-PIP3 antibody (5 μg ml−1) in TBS for 1 h at 37°C with rotation and washed three times in 1% (vol/vol) goat serum/TBS. Cells were then incubated with Alexa Fluor 488 goat anti-mouse/TBS (1:1,000) for 30 min at 37°C.
To confirm the authenticity of GST-PHBtk staining, cells were treated with a range of concentrations (60, 100, 250, and 500 nM) of the PI3-kinase inhibitor wortmannin (Sigma Chemicals, St. Louis, MO) for 15 or 30 min prior to fixation and decoration with the PIP3 biosensor. Cell viability after wortmannin treatment was estimated using trypan blue exclusion (0.5 mg ml−1).
Cells stained with GST-PHBtk or anti-PIP3 antibody were mounted in glycerol/PBS (1:1) solution on microscope slides and observed using a Zeiss LSM 510 confocal microscope.
Measurements of phagocytosis of hRBCs in E. histolytica cell lines were carried out according to the methods of Voigt et al. (43). Fluid-phase pinocytosis was measured using the fluorescent fluid-phase marker FITC (fluorescein isothiocyanate)-dextran (10 kDa; 2 mg ml−1) (Sigma Chemicals) as previously described (45). A correction for background GFP fluorescence was achieved by subtracting the initial (time zero) fluorescence values from all other fluorescence values.
Adhesion of E. histolytica amoebae to erythrocytes was assayed using a previously described rosette formation assay (33). Briefly, trophozoites (1 × 105) and erythrocytes (1 × 106) were mixed in 500 μl of TYI-33, centrifuged (200 × g, 4°C, 5 min), and incubated for 30 min on ice. Following incubation, 450 μl of the supernatant was removed and 50 μl trypan blue was added. A portion of the resulting mixture (10 μl) was used for counting on a hemacytometer, and at least 200 trophozoites were scored per trial. Adherent amoebae were defined as those with three or more bound erythrocytes (33).
Motility assays were carried out according to the methods of Zaki et al. (47), with modifications. To prepare the motility chambers, 8 ml of complete TYI-S-33 medium, supplemented with 0.75% (wt/vol) agarose, was poured into 60-mm petri dishes and allowed to solidify at room temperature. Transgenic trophozoites were induced to express GFP or GFP-PHBtk by tetracycline treatment for 24 h. Wild-type and transgenic trophozoites were then incubated in incomplete medium lacking tryptone and yeast extract for 1 h. Following this incubation, 5 × 105 cells (suspended in incomplete medium) were placed in a trough (2 by 30 mm) that had been cut into the solidified complete medium. A coverslip (22 by 40 mm) was placed over the trough, and the plate was incubated at 37°C for 1 h in 5% CO2. Migrations were visualized using a Zeiss LSM 510 confocal microscope. Measurements of migration distance were determined using Zeiss LSM 510 image analysis software.
All values are given as means ± 1 standard deviation (SD). Statistical analyses were performed using GraphPad Instat V.3 with one-way analysis of variance (ANOVA) and a Tukey-Kramer multiple comparison test. P values less than 0.001 were considered extremely significant. P values less than 0.01 were considered highly significant, and P values between 0.01 and 0.05 were considered statistically significant.
In order to observe the spatial and temporal localization of PIP3 in E. histolytica, we generated cell lines that stably expressed GFP or GFP-tagged PHBtk. Expression of the exogenous proteins was confirmed by Western blotting of cell lysates prepared from the transfected trophozoites with anti-GFP antibody (Fig. (Fig.1).1). A number of protein bands nonspecifically reacting with the anti-GFP antibody were also visible on the Western blots (Fig. (Fig.1).1). However, it is unlikely that these proteins were degradation products of the exogenous protein, since they were also evident before induction with tetracycline.
The dynamic distribution of PIP3 was then observed by microscopy in the transformed cell lines. First, microscopy revealed that expression of GFP or GFP-PHBtk was variable among transformed trophozoites within the population (Fig. (Fig.2;2; also, see Movie S1 in the supplemental material, which shows GFP-PHBtk-expressing cells). Others have observed variable expression of GFP-labeled proteins in populations of transformed trophozoites. For example, trophozoites expressing GFP-tagged FYVE finger domains, derived from an E. histolytica protein, EhFP4, exhibited variable expression, as evidenced by differences in fluorescence intensity from cell to cell (24). Second, observation of randomly moving trophozoites revealed a diffuse intracellular staining in quiescent cells and a prominent localization of the PIP3-specific probe in extending pseudopodia (Fig. 2C and F; also, see Movie S1 in the supplemental material). In contrast, control cells expressing GFP alone exhibited cytosolic fluorescence, which was excluded from forming pseudopodia (Fig. 2B and E, arrows). Phagocytosis and cell motility share many features, including the requirements for membrane extension formation and localized actin polymerization (23). Therefore, localization of PIP3 to randomly extending pseudopodia in E. histolytica was not surprising. Untransfected wild-type amoebae exhibited minimal background fluorescence (Fig. (Fig.2A2A).
To gain insight into the timing of PIP3 evolution during phagocytosis of hRBCs, E. histolytica trophozoites expressing GFP-PHBtk were exposed to erythrocytes and visualized by time-lapse fluorescence imaging. The ingestion of a single erythrocyte occurred over a 60- to 80-s time interval, as evidenced by the observation of three independent phagocytic events. Each phagocytic event began with trophozoite binding to the erythrocyte, followed by the extension of pseudopodia around the particle and phagosomal closure with subsequent internalization of the newly sealed phagosome (Fig. (Fig.33).
Accumulation of GFP-PHBtk was not evident at the site of erythrocyte attachment, suggesting that PIP3 is not involved in the earliest stages of particle binding and adhesion (Fig. (Fig.3,3, t0). However, a peak in PIP3 fluorescence intensity was detected during the formation of deep and nearly sealed erythrocyte-containing membrane invaginations (Fig. (Fig.3,3, t40). The rise in PIP3-specific signal at the cup was most obvious in three-dimensional plots generated from corresponding single-plane confocal images, in which fluorescence intensity is represented as the height of the xyz plot (Fig. (Fig.4,4, t40). This accumulation was transient, lasting less than 10 s, and was followed by a rapid dissipation of the signal during the internalization of the phagosome (Fig. (Fig.33 and and4,4, t50 to t70). These data are consistent with known PIP3 dynamics that occur during the phagocytosis of particles in neutrophils (7) and D. discoideum (9).
Microscopic comparison of erythrophagocytosis in GFP-PHBtk-expressing trophozoites with that in wild-type or GFP control cells indicated that actual phagocytic events were rare in the transgenic cells. This suggested a functional defect in phagocytosis and prompted us to quantify phagocytosis in E. histolytica cells expressing GFP-PHBtk.
Trophozoites were exposed to hRBCs for 10 min, after which extracellular erythrocytes were lysed hypotonically with distilled water and the level of ingested heme was measured by spectrophotometry. Erythrophagocytosis in GFP-PHBtk-expressing trophozoites was inhibited by 78% and 69% compared to the wild-type and GFP control cells, respectively (Fig. (Fig.5A).5A). This suggests that expression of the GFP-tagged PH domain exerts a dominant negative effect on this cellular process.
To gain further insight into the phagocytosis defect, we measured the ability of wild-type and transgenic cell lines to adhere to erythrocytes using a standard rosette assay (33). Adhesion is considered an initial step in phagocytosis. Interestingly, cells expressing GFP-PHBtk bound erythrocytes as efficiently as wild-type or GFP-expressing control cells (Fig. (Fig.5B).5B). Therefore, it is likely that only later steps in phagocytosis, such as cup formation and closure, are inhibited by GFP-PHBtk expression. These data are consistent with microscopy data indicating that the GFP signal was not enhanced at the site of adhesion to erythrocytes (Fig. (Fig.33 and and4,4, t0). These data are also consistent with a previous observation that chemical inhibitors of PI3-kinase can block internalization but not binding of erythrocytes (30).
To determine if the cell line expressing GFP-PHBtk possessed a general endocytic defect, we measured uptake of a 10-kDa fluid-phase marker, FITC-dextran. Interestingly, uptake of this marker was not affected by the expression of the GFP-PHBtk (Fig. (Fig.5C).5C). Importantly, the differential effect of GFP-PHBtk expression on phagocytosis versus pinocytosis suggested that inhibition of phagocytosis was specific.
Since the GFP-tagged biosensor accumulated in the membrane extensions of randomly moving trophozoites (Fig. (Fig.2;2; also, see Movie S1 in the supplemental material) and since PIs are involved in regulating cell motility in other systems (reviewed in reference 18), we also examined motility in E. histolytica trophozoites expressing GFP-PHBtk. Interestingly, under-agar motility assays indicated that cells expressing the PIP3 biosensor displayed enhanced motility compared to GFP-expressing or wild-type cells (Fig. (Fig.6),6), supporting the notion that PIP3-based signaling is important in motility in E. histolytica and that such signaling pathways may be altered by the expression of GFP-PHBtk.
Since expression of GFP-PHBtk altered cellular functions in E. histolytica trophozoites, it was necessary to confirm the localization of PIP3 using untransformed cells. It was recently shown that recombinant glutathione-S-transferase (GST)-labeled FYVE finger protein domains and fluorescent anti-GST antibody could be used to localize PI3P in E. histolytica trophozoites (30). This provided the impetus to develop an analogous method to localize intracellular PIP3, using a GST-tagged PHBtk domain (GST-PHBtk). Others have shown that recombinant GST-PHBtk can bind specifically to PIP3 in vitro (32, 35). Importantly, Salim et al. (35) also demonstrated that fluorescent anti-GST antibody does not interfere with the interaction of GST-PHBtk and liposomes containing PIP3. Together, these data suggest that the GST-tagged pleckstrin homology domain of Btk is conformationally and functionally intact and supports its use as a lipid biosensor.
Bacterial expression of GST and GST-PHBtk and subsequent affinity purification yielded 28-kDa and 49-kDa proteins, respectively (Fig. (Fig.7A).7A). The larger protein (49 kDa) had a molecular weight equivalent to the combined molecular weights of the GST tag and the PHBtk domain. The authenticity of the GST and GST-PHBtk recombinant proteins was verified by Western blot analysis using anti-GST antibodies (Fig. (Fig.7B7B).
To test the utility of the PIP3 biosensor, E. histolytica trophozoites were fixed, permeabilized, and stained with GST-PHBtk or GST as previously described (30). Imaging revealed that the biosensor uniformly decorated plasma membranes, producing a high-intensity signal along the cell periphery and giving it a ring-like appearance (Fig. (Fig.8A,8A, panel ii). Some intracellular staining was also evident (Fig. (Fig.8A,8A, panel ii). This ring-like fluorescence pattern was different from that observed in trophozoites expressing the GFP-tagged probe. However, images of GST-PHBtk-stained cells were obtained by confocal optical sectioning and represent a single confocal plane, while images of transgenic trophozoites were obtained in live-cell mode and represent total fluorescence. Therefore, the dissimilarity is likely the result of various image capture methods. A similar difference was noted (24) when trophozoites stained with GST-tagged FYVE finger domains (30) were compared to those expressing GFP-tagged FYVE finger domains (24). Cells stained with the control GST protein lacking a PHBtk domain exhibited minimal fluorescence (Fig. (Fig.8A,8A, panel i).
We also compared the pattern of staining obtained with GST-PHBtk to that obtained by standard immunofluorescence microscopy using an anti-PIP3 antibody. Decoration of trophozoites with an anti-PIP3 antibody resulted in uniform peripheral fluorescence with some intracellular staining (Fig. (Fig.8B,8B, panel ii). This pattern was similar to that of GST-PHBTK-stained cells, supporting the authenticity of the staining pattern with the recombinant probe.
As an additional test for the specificity of biosensor-PIP3 interaction, we utilized a known inhibitor of PI3-kinase activity, wortmannin. Interaction of a PI3P-specific probe, GST-2XFYVE, with trophozoite membranes was maximally inhibited by 60 nM wortmannin (30). Therefore, this concentration was used as a starting point to test the response of the PIP3-specific probe, GST-PHBtk, to PI3-kinase inhibition. E. histolytica trophozoites were treated with increasing concentrations of wortmannin (60 nM, 100 nM, 250 nM, and 500 nM) for 15 or 30 min prior to fixation and decoration with the biosensor, and the intensity of fluorescence was measured. Treatment with 60 nM wortmannin did not significantly inhibit staining with GST-PHBtk (data not shown). Staining with the GST-tagged probe was not inhibited by treatment with 100 nM wortmannin for 15 min and was slightly inhibited by treatment with 100 nM wortmannin for 30 min (Fig. (Fig.9A,9A, panels i and iv, and B). Treatment with higher concentrations of wortmannin inhibited GST-PHBtk staining throughout trophozoites in a dose- and time-dependent manner (Fig. (Fig.9A,9A, panels ii, iii, v, and vi, and B), suggesting that GST-PHBtk authentically binds to products of PI3-kinase. Importantly, the viability of trophozoites exposed to wortmannin was not affected, since trypan blue exclusion demonstrated that 96.7% cells remained viable after treatment with 500 nM wortmannin for 30 min (data not shown). In other systems, similarly high concentrations of wortmannin (200 nM  and 500 nM [Btk-PH redistribution assay; Thermo Scientific]) were used to abolish the interaction between PIP3 and PH domains.
PIP3 is virtually undetectable in quiescent mammalian cells and transiently increases upon stimulation with growth factors and chemokines (38). Since trophozoites grown in culture medium, which is supplemented with adult bovine serum, may be exposed to such growth factors, we tested whether serum withdrawal would alter the localization or intensity of biosensor staining. E. histolytica cells were cultured in serum-free medium for 14 h, fixed, stained with GST-PHBtk, and visualized using fluorescence microscopy. Serum starvation inhibited GST-PHBtk staining in the plasma membrane only slightly (Fig. 10A). Quantitative measurement of membrane fluorescence indicated no statistical differences between populations of cells that were grown with and without serum (Fig. 10B). Similarly, probe-specific staining was not altered by a 1 h of incubation of trophozoites in PBS (data not shown). Importantly, trophozoites remained viable during 14 h of serum starvation or 1 h of PBS incubation (data not shown). While it is possible that these conditions were not sufficient to induce quiescence in trophozoites, the data suggest that E. histolytica exhibits high steady-state levels of PIP3.
To confirm the subcellular localization of PIP3 during phagocytosis obtained using expression of GFP-PHBtk, E. histolytica trophozoites were exposed to hRBCs for 1 min. Subsequently, the cells were fixed, permeabilized, and stained with GST-PHBtk or with the GST control protein as previously described (30). Different stages of erythrocyte uptake were captured using laser-scanning confocal microscopy. GST-PHBtk labeled membranes were associated with invaginations surrounding hRBCs (Fig. 11A) and nearly sealed erythrophagocytic cups (Fig. 11B). However, no probe-specific staining was observed in internalized erythrophagosomes located just beneath the plasma membrane (Fig. 11C and H, arrowheads), which suggests that PIP3 dissipates from the phagosomal membrane after closure is complete. These data are consistent with those obtained by live-cell microscopy of cells expressing GFP-PHBtk (Fig. (Fig.33 and and44).
Erythrocyte membranes were not stained with the GST-PHBtk, as evidenced by the extracellular hRBCs captured in the same confocal plane as the labeled E. histolytica cell (Fig. 11G, asterisk), demonstrating that GST-PHBtk staining was specific for E. histolytica. Control cells that were stained with the GST protein alone did not exhibit any fluorescence at the sites of interaction with erythrocytes (Fig. 11D and I, arrowheads). As expected, erythrophagosomal cups were rare in wortmannin-treated trophozoites; however, partially engulfed hRBCs were occasionally observed (Fig. 11E and J, arrowhead). In these instances, GST-PHBtk staining was not visible at the cup (Fig. 11E, arrowhead). Together, these results suggest that PIP3 localizes to phagosomes during formation of the cup up to the point of closure. The data also suggest that PIP3 may not participate in later stages (maturation) of phagocytosis in E. histolytica.
To confirm a lack of involvement of PIP3 in fluid-phase uptake, E. histolytica trophozoites were exposed to lysine-fixable tetramethyl rhodamine isothiocyanate (TRITC)-dextran (10 kDa) for 2, 5, 10, and 60 min prior to fixation and staining with GST-PHBtk. Similar to a previous report describing a lack of association of PI3P with pinosomes (30), the PIP3-specific probe (Fig. 12B) did not colocalize with fluid-phase compartments (Fig. 12A) at any time (only the 60-min time point is shown). This observation is in accordance with data showing that expression of GFP-PHBtk did not inhibit uptake of FITC-dextran (Fig. (Fig.5C).5C). Together, these data indicate that PIP3 may not be involved in the uptake of fluid phase from the surrounding medium.
In this study we utilized GFP- and GST-labeled PH domains to examine the subcellular localization of PIP3 in E. histolytica. We showed that these PIP3-specific biosensors accumulated in membrane extensions in E. histolytica during random movement and phagocytosis. Microscopy also revealed that trophozoites may possess high steady-state levels of PIP3. We also demonstrated that expression of the GFP-tagged PHBtk domain inhibited phagocytosis and enhanced motility in this pathogen, supporting a role for PIP3 in these cellular functions. Since the GST-tagged probe did not colocalize with fluid-filled vesicles and since expression of the GFP-tagged probe did not inhibit trophozoite-erythrocyte interactions or uptake of fluid-phase, PIP3 may not be involved in adhesion of trophozoites to hRBCs or in fluid-phase pinocytosis in E. histolytica.
It was previously demonstrated that PI3P was present on early erythrophagosomes in E. histolytica (30). In the present study we showed that PIP3 similarly accumulates during phagosome biogenesis. These observations raise the question of why both of these lipids appear early in phagocytosis in E. histolytica. One possibility is that PIP3 and PI3P may recruit different downstream protein targets to facilitate vesicle formation and activate actin cytoskeletal rearrangement. Both lipids may also be necessary to facilitate the interaction of the proteins with which they bind. In support of this, in mammalian cells, Akt is recruited to the plasma membrane by virtue of the interaction of its PH domain with PIP3 (41). At the membrane, Akt interacts with one of its effector proteins, ProF, which is recruited to the plasma membrane by virtue of the interaction of its FYVE finger domain with PI3P (12). Thus, the simultaneous presence of PIP3 and PI3P at the membrane may allow effective assembly of interacting partners in a signaling pathway.
Although variable expression of GFP-PHBtk was observed among transgenic trophozoites within a single population, measurements of the whole population of mutants revealed a phagocytosis defect. These data suggest that it might be possible to perturb PI-based signaling by overexpression of protein domains that bind to phosphorylated products of PI. The observation also indicates that phagocytosis in E. histolytica is very sensitive to such perturbation. Importantly, uptake of fluid phase pinocytosis was not inhibited in the same population of transformants, which supports the authenticity of the observed phagocytic defect. A similar reduction in the uptake of phagocytic targets, but not of a fluid-phase marker, was observed after expression of GFP-tagged FYVE finger domains in E. histolytica (24). The latter study further supports the notion that overexpression of lipid binding domains can be used to study PI-based signaling in E. histolytica.
In other systems, overexpression of PH-containing proteins does indeed interfere with PI-based signaling (5, 42). For example, Varnai et al. demonstrated that expression of GFP-PHBtk in fibroblasts inhibited activation of Akt, a well-characterized downstream target of PIP3 (42). It is possible that interaction of the chimeric probe with PIP3 causes titration of endogenous lipids, thus preventing the recruitment of downstream protein targets such as Akt. In addition to lipids, PH domains have been shown to interact with a variety of proteins in vitro; whether they interact with such proteins in vivo has only been confirmed for a few candidates (21, 34). Therefore, in addition to titration of lipid, expressed PHBtk may also sequester proteins that normally interact with PH domains. This, too, would interrupt relevant signaling cascades.
The activity of PI3-kinase is balanced by the activity of phosphatases, such as PTEN or SHIP, which can remove the phosphate moiety from PIP3, converting it to PIP2 (reviewed in reference 19). Therefore, titration of PIP3 by GFP-PHBtk may not only inhibit recruitment of Akt but also block the action of phosphatases, resulting in aberrantly high cellular levels of PIP3 and low cellular levels of PIP2. Turnover of PIP3 is critical to phagocytosis, since PIP3 is proposed to regulate pseudopod extension, whereas PIP2 is proposed to regulate actin remodeling (reviewed in reference 25). Thus, it is conceivable that altering PIP3/PIP2 ratios could inhibit phagocytosis, as seen in the GFP-PHBtk-expressing cell line.
How is it possible, however, that expression of a PH domain inhibits phagocytosis while enhancing another PI-dependent cellular function, namely, motility? Interestingly, loss of PTEN function is common during tumorigenesis (4), and this deficit promotes motility and invasiveness. In such tumor cells, PIP3 is not converted to PIP2, which leads to persistent PIP3-specific signaling and enhanced motility. Similarly, if the chimera protected PIP3 from the action of cellular phosphatases in trophozoites, cell movement might have been promoted.
In mammalian cells, PIP3 is transient and its levels can be reduced by serum starvation (38). Surprisingly, PIP3 staining was only slightly inhibited after serum starvation in E. histolytica trophozoites. Although we cannot rule out the possibility that this culture condition did not induce a quiescence-like state in E. histolytica, the data suggest that PIP3 exhibits high steady-state levels in this organism. Since turnover of PIP3 is important for cellular processes such as motility (4), apparent high steady-state levels of PIP3 in E. histolytica might be achieved by a continuous rapid phosphorylation and dephosphorylation cycle.
Although different from levels in mammalian cells, high steady-state levels of PIP3 in E. histolytica are similar to those seen in a related, nonpathogenic amoeba, D. discoideum (48). It is not known if PIP3 in D. discoideum is concentrated at the plasma membrane, since all localization studies were conducted with GFP-PH domain-expressing amoebae and live-cell imaging (9, 48). The reason that PIP3 exists as a highly persistent pool of lipids in these organisms is not known. However, several important physiological processes in these lower eukaryotes may require such high steady-state levels of PIP3. Given its role in phagocytosis, PIP3 may be required for the continuous uptake of nutrients, a common characteristic of these protozoa. PI3-kinase signaling also plays a role in cell proliferation and survival (reviewed in reference 11). Therefore, the stability of PIP3 in E. histolytica and D. discoideum may also be related to their single-cell nature, for which rapid and continuous divisions are an important goal.
In summary, this study characterized for the first time the spatiotemporal distribution of PIP3 in E. histolytica and demonstrated that this lipid is important for early steps of erythrophagocytosis. Since this process represents an important virulence function in E. histolytica, this study contributes to our understanding of the molecular events that underlie pathogenic mechanisms in this parasite. The downstream interacting partners of phosphoinositides in E. histolytica are not known. However, identification and functional analyses of PI3P- and PIP3-binding proteins will provide additional insight into pathogenesis. Finally, given that PI3-kinase based signaling may be perturbed in the cell line expressing GFP-PHBtk, further study of this mutant will provide the opportunity to use a genetic approach to gain insight into cellular signaling pathways and virulence.
We thank W. A. Petri, Jr. (University of Virginia, Charlottesville, VA), for the pGIR209 and pGIR308 expression plasmids. We thank Terri F. Bruce (Clemson University, Clemson, SC) for guidance in measuring motility.
The project described was supported by grant no. R01AI046414 from the National Institute of Allergy and Infectious Diseases to L.A.T. and a Sigma Xi Grant-in-Aid to Y.A.B. This material is based upon work supported by CSREES/USDA, under project number SC-1700312 (Technical Contribution No. 5678 of the Clemson University Experiment Station).
The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institute of Allergy and Infectious Diseases, the National Institutes of Health, or the USDA.
We thank the members of the Temesvari laboratory for critical reading of the manuscript and helpful discussions.
Editor: F. C. Fang
Published ahead of print on 9 November 2009.
†Supplemental material for this article may be found at http://iai.asm.org/.