Molecular interactions between Anopheles stephensi midgut cells and Plasmodium berghei: the time bomb theory of ookinete invasion of mosquitoes
2000; Springer Nature; Volume: 19; Issue: 22 Linguagem: Inglês
10.1093/emboj/19.22.6030
ISSN1460-2075
Autores Tópico(s)Mosquito-borne diseases and control
ResumoArticle15 November 2000free access Molecular interactions between Anopheles stephensi midgut cells and Plasmodium berghei: the time bomb theory of ookinete invasion of mosquitoes Yeon Soo Han Yeon Soo Han Colorado State University, Pathology Department, Fort Collins, CO, 80523 USA Search for more papers by this author Joanne Thompson Joanne Thompson European Molecular Biology Laboratory, Meyerhofstrasse 1, D-69117 Heidelberg, Germany Present address: Leiden University Medical Centre, Department of Parasitology, L4-Q, Albinusdreef 2, 2333 ZA Leiden, The Netherlands Search for more papers by this author Fotis C. Kafatos Fotis C. Kafatos European Molecular Biology Laboratory, Meyerhofstrasse 1, D-69117 Heidelberg, Germany Search for more papers by this author Carolina Barillas-Mury Corresponding Author Carolina Barillas-Mury Colorado State University, Pathology Department, Fort Collins, CO, 80523 USA Search for more papers by this author Yeon Soo Han Yeon Soo Han Colorado State University, Pathology Department, Fort Collins, CO, 80523 USA Search for more papers by this author Joanne Thompson Joanne Thompson European Molecular Biology Laboratory, Meyerhofstrasse 1, D-69117 Heidelberg, Germany Present address: Leiden University Medical Centre, Department of Parasitology, L4-Q, Albinusdreef 2, 2333 ZA Leiden, The Netherlands Search for more papers by this author Fotis C. Kafatos Fotis C. Kafatos European Molecular Biology Laboratory, Meyerhofstrasse 1, D-69117 Heidelberg, Germany Search for more papers by this author Carolina Barillas-Mury Corresponding Author Carolina Barillas-Mury Colorado State University, Pathology Department, Fort Collins, CO, 80523 USA Search for more papers by this author Author Information Yeon Soo Han1, Joanne Thompson2,3, Fotis C. Kafatos2 and Carolina Barillas-Mury 1 1Colorado State University, Pathology Department, Fort Collins, CO, 80523 USA 2European Molecular Biology Laboratory, Meyerhofstrasse 1, D-69117 Heidelberg, Germany 3Present address: Leiden University Medical Centre, Department of Parasitology, L4-Q, Albinusdreef 2, 2333 ZA Leiden, The Netherlands *Corresponding author. E-mail: [email protected] The EMBO Journal (2000)19:6030-6040https://doi.org/10.1093/emboj/19.22.6030 Correction(s) for this article Molecular interactions between Anopheles stephensi midgut cells and Plasmodium berghei: the time bomb theory of ookinete invasion of mosquitoes15 March 2001 PDFDownload PDF of article text and main figures. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InMendeleyWechatReddit Figures & Info We present a detailed analysis of the interactions between Anopheles stephensi midgut epithelial cells and Plasmodium berghei ookinetes during invasion of the mosquito by the parasite. In this mosquito, P.berghei ookinetes invade polarized columnar epithelial cells with microvilli, which do not express high levels of vesicular ATPase. The invaded cells are damaged, protrude towards the midgut lumen and suffer other characteristic changes, including induction of nitric oxide synthase (NOS) expression, a substantial loss of microvilli and genomic DNA fragmentation. Our results indicate that the parasite inflicts extensive damage leading to subsequent death of the invaded cell. Ookinetes were found to be remarkably plastic, to secrete a subtilisin-like serine protease and the GPI-anchored surface protein Pbs21 into the cytoplasm of invaded cells, and to be capable of extensive lateral movement between cells. The epithelial damage inflicted is repaired efficiently by an actin purse-string-mediated restitution mechanism, which allows the epithelium to ‘bud off’ the damaged cells without losing its integrity. A new model, the time bomb theory of ookinete invasion, is proposed and its implications are discussed. Introduction Malaria infections impose a tremendous burden to human health in tropical regions worldwide, particularly in Africa. Although anopheline mosquitoes play a central role in transmission of the malaria parasite, Plasmodium, our understanding of the basic biology of vector–parasite interactions during the transit of the parasite through the mosquito is still limited. Upon ingestion, Plasmodium gametocytes differentiate into gametes within the midgut lumen, fertilization takes place and the zygote develops into an ookinete, which penetrates the midgut epithelium and forms an oocyst on the basal side. When the oocyst matures and ruptures, sporozoites are released into the haemolymph and invade the salivary glands, to be injected into a new host when the infected mosquito feeds again. Malaria transmission is possible because some parasites are able to develop and escape unharmed from the mosquito. There is increasing evidence that the mosquito midgut is an immune-competent organ, inducing the expression of several immune markers in response to Plasmodium infection. In Anopheles gambiae (Ag), mRNA levels of the antibacterial peptide defensin, a putative Gram-negative binding protein (AgGNBP), a chitinase-like domain-containing protein (ICHIT), a putative serine protease (ISPL5), a lectin-like protein (IGALE20) and nitric oxide synthase (AgNOS) increase in the midgut 24 h post-infection with Plasmodium berghei (Dimopoulos et al., 1997, 1998; Richman et al., 1997; reviewed by Barillas-Mury et al., 2000). Studies in Anopheles stephensi (As) have also shown transcriptional activation of AsNOS in the midgut and systemically in response to P.berghei infection. Furthermore, dietary provision of the NOS substrate L-arginine significantly reduces the percentage of Plasmodium-infected mosquitoes, while an NOS inhibitor significantly increases the number of oocysts that develop (Luckhart et al., 1998). These immune responses have been observed in mosquito strains of A.gambiae and A.stephensi susceptible to P.berghei infection. Previous studies of Aedes aegypti midgut invasion by the chicken malaria parasite Plasmodium gallinaceum reported that ookinetes invade a special type of epithelial cells expressing high levels of vesicular ATPase (vATPase), which were named ‘Ross cells’. The histochemical and electron microscopic analysis of the morphology of these cells indicated that they differ from other epithelial cells. They stain very poorly with toluidine blue (a basophilic dye), are less osmiophilic, contain minimal endoplasmic reticulum, lack secretory granules and have few microvilli (Shahabuddin and Pimenta, 1998). The same group has also reported that P.gallinaceum ookinetes invade a similar type of cell in A.gambiae midgut cells expressing high levels of vATPase (Cociancich et al., 1999). The conclusion that P.gallinaceum ookinetes do invade functionally related cells in both A.aegypti and A.gambiae was drawn from very few invasion events, due to the use of an unfavourable insect–parasite combination. The existence of ‘Ross cells’ has important implications. If ookinetes have developed a mechanism to invade a specific midgut cell type, this implies the existence of a recognition system. One could envisage that interactions between ookinetes and specific receptors, presumably present only on the surface of the ‘Ross cells’, are required for successful midgut invasion, and thus would represent potential target molecules to block disease transmission. In this study, the cell biology of the interactions between midgut cells and ookinetes during epithelial invasion are examined in the well-established A.stephensi–P.berghei malaria model, in which numerous ookinetes succeed in developing into oocysts. Conclusions are drawn after observing hundreds of invasion events in this compatible mosquito–parasite combination, where manipulations of the midgut tissue were minimized to prevent in vitro cell damage. Ookinetes were localized using monoclonal antibodies to Pbs21, an abundant surface protein, and their interactions with midgut epithelial cells were established by analysing the expression of several proteins, such as AsSTAT, vATPase, AsNOS, actin and a P.berghei subtilisin-like serine protease, PbSub2. The data show that, in A.stephensi mosquitoes, P.berghei ookinetes invade columnar midgut epithelial cells with microvilli, which are capable of mounting anti-parasitic responses, such as the induction of NOS expression, and that parasite invasion always results in cell death. Plasmodium berghei ookinetes do not invade the subset of cells expressing high levels of vATPase. Invaded cells are functionally and morphologically different from healthy midgut epithelial cells, but these differences are the consequence of the damage inflicted by the parasite and are not present before invasion. A new model of ookinete invasion, the time bomb theory, is proposed. Results Midgut epithelial cells invaded by ookinetes fail to translocate AsSTAT Two transcription factors, Gambif1 (gambiae immune factor 1), which belongs to the rel-family (Barillas-Mury et al., 1996), and AgSTAT, a member of the STAT family of receptors, have been characterized previously (Barillas-Mury et al., 1999). Both factors are activated in fat body cells following a bacterial challenge and are also expressed in the midgut epithelial cells of A.gambiae (Barillas-Mury et al., 1996, 1999). Pervanadate treatment (hydrogen peroxide and vanadate) is known to activate the STAT pathway in vertebrates and Drosophila through a ligand-independent mechanism (Zhong et al., 1994; Yan et al., 1996). AgSTAT is evenly distributed in the cytoplasm and nucleus of midgut cells, but pervanadate treatment results in a prominent translocation of AgSTAT from the cytoplasm to the nucleus, indicating that this pathway has been activated (Barillas-Mury et al., 1999). A polyclonal antiserum was generated previously using a region of AgSTAT with 54% amino acid identity (80% similarity) to the A.stephensi homologue, AsSTAT (C.Barillas-Mury, unpublished). The anti-AgSTAT antibody recognizes a single band of 78 kDa in adult A.stephensi female extracts (data not shown), with the same pattern as the immunoblot that has been reported for A.gambiae extracts (Barillas-Mury et al., 1999). This antibody also cross-reacts with A.stephensi midgut tissue in immunostainings and was used to determine the subcellular distribution of AsSTAT in P.berghei-infected midguts. The location of P.berghei ookinetes was determined using a monoclonal anti-Pbs21 antibody (Winger et al., 1988). Pbs21 is a surface protein expressed at high levels in P.berghei zygotes and ookinetes; it is present at lower levels in oocysts and absent in sporozoites and blood-stage parasites (Simonetti et al., 1993). Pbs21 has four epidermal growth factor-like domains, a putative C-terminal hydrophobic membrane anchor and sequence homology (45, 45 and 40%, respectively) to 25 kDa surface proteins found in P.falciparum (Pfs25), P.reichenvowi (Prs25) and P.gallinaceum (Pgs25) (Paton et al., 1993). Biochemical analysis indicates that Pbs21 is glycosylated and GPI anchored (Blanco et al., 1999). Double staining for AsSTAT and Pbs21 in midguts from females 8 h after feeding on ookinete cultures revealed that Pbs21 is present throughout the cytoplasm of some midgut epithelial cells (Figure 1A). In many cases, intact ookinetes could be observed emerging from cells with moderate Pbs21 cytoplasmic staining. Most probably this staining is due to secretion of this surface protein during the parasite's transit across the cell (Figure 1B). A small area rich in Pbs21 was often observed on the luminal surface of invaded cells, and could represent the site of parasite penetration. Arrowheads in Figures 1B and 2G indicate examples of such areas. The Pbs21 staining in some cells was very strong, although ookinetes could not be observed in the immediate vicinity; we surmise that in these cases the parasite may have been lysed intracellularly. AsSTAT was present in the cytoplasm and nucleus of most epithelial cells. However, all cells positive for Pbs21 were notable for having only a very low cytoplasmic level and lack of nuclear staining for AsSTAT (Figure 1B). To determine whether Pbs21-positive cells were capable of translocating AsSTAT, ookinete-infected midguts were treated in vitro with pervanadate before fixation. As expected, this treatment resulted in a dramatic redistribution of AsSTAT from the cytoplasm to the nucleus in most midgut cells. Again, however, AsSTAT staining was very weak in all Pbs21-positive cells, and was absent from the nucleus (Figure 1C). Figure 1.Immunofluorescence staining of midguts with anti-Pbs21 (green) and anti-STAT (red) antibodies, 8 h after feeding on an ookinete culture. (A) Some cells protrude to the luminal side and their cytoplasm stains with Pbs21. (B) An ookinete emerging from a Pbs21-positive cell. The invaded cell has very low AsSTAT expression and the nucleus does not stain. (C) Pbs21-positive cells fail to translocate AsSTAT, even after pervanadate treatment. Their location in the red channel is indicated by the arrows. The small arrowheads indicate some examples of nuclei with strong AsSTAT staining. Download figure Download PowerPoint Figure 2.Immunofluorescence staining of midguts with anti-NOS, anti-vATPase and anti-Pbs21. (A) DIC, NOS and Pbs21 staining of midguts 24 h after feeding on a healthy (control) or malaria-infected mouse. (B) Western blot of adult female homogenates from A.gambiae and A.stephensi stained with pre- or post-immune serum to Ag-vATPase. (C) Double staining of NOS (green) and vATPase (red). The cells expressing high levels of vATPase do not express NOS. The inset illustrates the NOS staining channel of a columnar cell and a small vATPase-overexpressing cell (indicated by the arrow) at higher magnification. The arrowheads indicate the nuclei: note the size difference between the two cell types and the lack of NOS staining in the small cell. (D–G) Triple stainings of infected midguts for Pbs21 (red), NOS (green) and vATPase (blue). (D) Homogeneous AsNOS cytoplasmic staining of epithelial cells at the posterior end of the midgut. (E) Small cells expressing high levels of vATPase were observed in the same sample as in (D); one is indicated by the arrow. (F) Ookinetes co-localize with protruding cells expressing high levels of NOS. (G) The protruding cells express the same level of vATPase as the uninvaded cells. (H) An ookinete that has reached the basal lamina is in close proximity to a vATPase-overexpressing cell. Download figure Download PowerPoint These results indicate that the ookinete-invaded cells are functionally different from other midgut epithelial cells. Two possible explanations exist. The invaded cells could belong to a special cell type that is targeted specifically by ookinetes, unable to activate the AsSTAT pathway and equivalent to the ‘Ross cells’ described for A.egypti. Alternatively, inability to activate the AsSTAT pathway may be the result of cell damage inflicted by the parasite during invasion. To test these hypotheses, a detailed analysis of the cellular morphology and the expression and distribution of several protein markers was performed. The invaded cells have a distinct morphology and overexpress NOS In all subsequent experiments, unless stated otherwise, midgut tissues were analysed 24 h after feeding either on naive or malaria-infected mice. Single or multiple immunofluorescence staining was detected by confocal microscopy in multiple optical sections, which were then merged to visualize the various markers at different planes of the midgut tissue. In an initial study, double immunofluorescence staining for AsNOS and Pbs21 was performed. AsNOS was detected using a commercial ‘Universal anti-NOS’ antibody (Oncogene, Cambridge, MA), raised against a conserved region of mouse iNOS that is known to cross-react with multiple species. In immunoblots, the antibody specifically detected a single band of the expected size (150 kDa) in A.gambiae and A.stephensi extracts from unfed females and pupae (data not shown). AsNOS is constitutively expressed in the cytoplasm of most midgut epithelial cells fed on normal blood. Morphological examination using differential interference contrast (DIC) revealed a smooth epithelium, in which epithelial cells of regular shape are present in a single plane, in a symmetrical distribution (Figure 2A). In contrast, following ookinete invasion, several cells were observed by DIC that protrude to the luminal side of the midgut, express very high levels of NOS and are closely associated with Pbs21-positive ookinetes (Figure 2A). The protruding cells are never observed in uninfected controls, suggesting that the morphological differences and very high NOS levels observed following ookinete invasion are the result of damage caused by the parasite. In ∼95% of the cases, healthy parasites that had reached the basal lamina co-localized with cells that protruded to the lumen and overexpressed NOS at 24 h post-feeding (Figure 2A and F). Damaged cells without associated parasites and healthy parasites (Figure 2H), not in contact with damaged cells, are also observed. Plasmodium berghei does not invade a subpopulation of cells expressing high levels of vATPase in A.stephensi To determine whether P.berghei ookinetes in A.stephensi invade the functional equivalent of the A.aegypti vATPase-expressing ‘Ross cells’ (Shahabuddin and Pimenta 1998), the expression of AsNOS and vATPase in ookinete-invaded cells was analysed in parallel. A rat polyclonal antiserum to the vATPase F subunit from A.gambiae was generated, as the A.aegypti anti-vATPase antiserum, kindly provided by Dr Shahabuddin, did not cross-react with A.stephensi in immunoblots or immunofluorescence staining (data not shown). In immunoblots of A.gambiae and A.stephensi whole female extracts, the anti-vATPase F rat antiserum detected a single band of the predicted size (∼13 kDa), which was not detected by the pre-immune serum of the same animal (Figure 2B). NOS and vATPase double stainings were performed in midguts of mosquitoes fed normal mouse blood. The anti-Ag-vATPase antibodies detected a subpopulation of cells that overexpress vATPase, but do not express AsNOS. These cells are small, flat, have a triangular shape, are only seen on the basal side of the epithelium and are more abundant in the posterior end of the midgut (Figure 2C). These morphological features make it hard to envisage how they could be successfully invaded by ookinetes. To detect protruding cells, ookinetes and the cells overexpressing vATPase simultaneously, triple staining for Pbs21, NOS and vATPase was performed. The posterior region, where the cells overexpressing vATPase are more abundant, was analysed initially to ensure that these cells were stained properly. Constitutive cytoplasmic expression of AsNOS in columnar epithelial cells (Figure 2D) and high levels of vATPase in the small triangular cells were observed (Figure 2E). Subsequently, images were obtained from a different area of the same midgut sample where ookinetes were present. Multiple cells, expressing high levels of AsNOS and associated with ookinetes, were observed protruding into the midgut lumen. The protruding cells were spherical, appeared to be about to ‘bud-off’ from the epithelium and remained attached only by a narrow area at the base (Figure 2F). The protruding cells stain only weakly for vATPase, with a similar intensity to that of uninvaded cells (Figure 2G). After scanning hundreds of ookinetes, it was possible to find only one that had reached the basal lamina and was in the same field of view and plane as a nearby cell overexpressing vATPase (Figure 2H). A cell protruding to the lumen was, however, also observed in close proximity by DIC (data not shown). The lack of co-localization of vATPase-overexpressing cells with ookinetes, therefore, again indicates that these are not the targets for ookinete invasion. Ookinete invasion results in extensive cell death To determine whether the protruding cells were irreversibly damaged, the terminal deoxynucleotidyl transferase-mediated deoxyuridine triphosphate nick end-labelling (TUNEL) DNA fragmentation assay was performed. In this assay, terminal transferase enzyme is used to incorporate fluorescence-labelled dNTPs at the free ends of genomic DNA. The efficiency of label incorporation depends on the number of free ends in the DNA strand. When the cell has been irreversibly damaged, DNA fragmentation generates numerous free ends, which result in strong fluorescence staining. Midgut cells that were damaged by a longitudinal incision made to remove the blood meal contents and create a flat midgut epithelial sheet provided a positive control. In the interior of the sheet, all ookinete-invaded cells protruded to the lumen and their nuclei were strongly TUNEL positive (Figure 3A). We have also observed in such preparations that all protruding cells have lower levels of AsSTAT expression and lack nuclear staining (data not shown). Figure 3.(A and B) DIC imaging, TUNEL assay (green) and DAPI nuclear staining (blue) of midgut epithelial cells 24 h post-feeding on malaria-infected mice. The specimen in (B) is the same as that in (A), at higher magnification, comparing the nucleus of a healthy cell with that of an invaded cell that protruded to the lumen. (C) Staining of protruding cells for NOS (red) and actin (Alexa-488 conjugated phalloidin in green). The invaded cells present actin aggregates and a loss of surface microvilli in the midgut lumen side (ml). Compare the microvilli density on the surface of an invaded cell (white box in the ml plane) with that of a healthy cell (white box in the bl plane). An actin ring is formed at the base of the cell, in the plane of the basal lamina (bl). The arrows indicate the direction of the actin ring constriction that would be required to bud off the cell into the midgut lumen. (D) Double staining of an invaded cell for actin (green) and Pbs21 (red). Notice the loss of microvilli (compare white boxes), the actin aggregates (indicated by the arrowheads), the constriction of the parasite at the point where it emerges from the cell and the actin ring at the base of the cell. (E) Lateral view of the parasite-invaded cell in (D). (F) Midgut staining 48 h after infection for NOS (green) and Pbs21 (red). Early oocysts can be observed in a healthy looking midgut that has been repaired. Download figure Download PowerPoint While the evidence for DNA fragmentation in invaded cells is clear, it is uncertain whether this is indicative of programmed cell death (apoptosis). DNA laddering tests of midgut DNA followed by a species-specific hybridization revealed nucleosome-like ladders indicative of apoptosis in the cells of the mouse blood meal, but not in the mosquito midgut cells, irrespective of the presence or absence of infective ookinetes (data not shown). Most probably, the amount of fragmented DNA from the damaged midgut cells is below the limit of detection of this assay. Thus, further analysis would be required to establish whether ookinetes are actively inducing apoptosis of the invaded cells. Gross changes in nuclear morphology were also often observed in protruding cells (Figure 3B). Whilst the nuclei of normal cells are circular, symmetrical and show homogeneous DNA distribution, those of invaded cells appear pycnotic, with irregular shapes. They present areas of DNA condensation and, in some cases, partial fragmentation of the nucleus. Dead cells ‘bud off’ into the midgut lumen as the epithelial damage is repaired through an actin purse-string-mediated restitution mechanism During experimental P.berghei infections in A.stephensi, it is not uncommon to observe >100 successful ookinete invasions within a single midgut. Our results indicate that all invaded cells die and that a single parasite often injures more than one cell. Furthermore, ookinete invasion causes this potentially massive damage at a time when the midgut is distended by the blood meal and high levels of proteases are present in the lumen. However, no difference in mortality between females fed on uninfected or infected mice is observed, with >96% of infected females surviving 48 h after infection. How does the midgut manage to repair such extensive damage, preserve the continuity of the epithelium and thus prevent direct contact between the haemolymph and the midgut contents? Disruptions of the mucosal lining of the gastrointestinal tract of vertebrates are resealed by the concerted movement of the surrounding cells, a process termed restitution, and not by cell division (Lotz et al., 2000). Epithelial sheets respond to injury by mobilizing their actin cytoskeleton through two different mechanisms: purse-string contraction and formation of lamellae. A detailed analysis of the actin cytoskeleton, using phalloidin conjugated to a fluorescent dye, was performed to determine the type of healing mechanism that takes place in the mosquito midgut epithelium. A dramatic redistribution of actin from the luminal to the basal side of the invaded midgut cells was observed (Figure 3C and D). Serial sections were taken across the epithelium from apical to basal, and were separated into two stacks before merging: the more apical stack including the midgut lumen (ml) and the more basal one the basal lamina (bl). In Figure 3C, panel ml visualized only protruding cells, which are NOS rich and depleted of microvilli (white box), except at their periphery. Panel bl includes the microvillar apical region of non-invaded cells (white box) as well as the more basal region of the invaded cells, which shows actin aggregates forming ring structures. In Figure 3D, panel ml, the two white boxes permit direct comparison of microvilli-rich non-invaded cells (peripheral) and a microvilli-poor invaded cell (central), in which the actin is aggregated around a Pbs21-stained parasite; panel bl clearly shows the actin ring. We propose that such rings are responsible for constricting the protruding cells at their base, allowing the epithelium eventually to pinch these cells off and expel them into the midgut lumen, whilst simultaneously closing the gap they would otherwise leave behind (Figure 5C). When several contiguous cells are damaged, they are expelled from the epithelium as a group (Figures 2F and 3C). A three-dimensional image of the invaded cell in Figure 3D was reconstructed using all sections and rotated to allow a lateral view of the same cell (Figure 3E). The cell protrudes towards the luminal side, the parasite itself is constricted by the actin at the basal surface of the cell and the adjacent cell has flattened at the site of contact with the injured cell. Parasite constrictions are seldom observed (see below); they could be either rare events or very transient ones. The epithelial repair mechanism is very effective; by 48 h after feeding, most of the damaged cells have been expelled to the lumen and early oocysts are observed developing in a healthy looking midgut (Figure 3F). P.berghei ookinetes are remarkably plastic and are capable of extensive lateral movement between cells A novel ookinete morphology has been described recently for P.gallinaceum, in which the central portion of the parasite forms an elongated narrow tube or stalk joining the anterior and posterior portions of the parasite. Stalk-form ookinetes were not observed in vitro, and thus it was postulated that they result from interactions of the parasite with the midgut epithelium (Vernick et al., 1999). Hundreds of Pbs21-stained parasites were examined to determine whether the stalk morphology also exists in P.berghei ookinetes as they traverse the A.stephensi midgut epithelium. More than 99% had the typical ‘banana shape’ (Figure 4A). Several examples of variations of the stalk shape were found, however, ranging all the way from moderate constrictions in the posterior end of the parasite (Figure 4B) to dramatic elongated stalks in the middle (Figure 4C). Figure 4.(A–D) Immunofluorescence staining of ookinetes with Pbs21 (red) and detection of haemozoin granules (yellow). (E and F) Double staining for Pbs21 (red) and actin (green). Note the close association of actin with the parasite constrictions (indicated by the arrows). (G) Immunofluorescence staining of cells following ookinete invasion with anti-NOS (red) and anti-Pbs21 antibodies (green), counterstained with DAPI (blue). The area where the Pbs21 trail is present was overexposed, as the signal was much weaker relative to that of the ookinete surface. (H) Double staining for actin (green) and Pbs21 (red). The same ookinete has undergone extensive lateral movement, damaging six consecutive cells. (I) Immunoblot of cultures containing schizonts and trophozoites with recombinant PbSub2 pre- and post-immune serum. (J and K) Triple staining for actin (green), Pbs21 (light blue) and subtilisin (PbSub2 in red). The arrows in (J) indicate the close contact between the PbSub2 aggregates and the actin cytoskeleton. All aggregates (including the one indicated by the arrow in K) are present in the cytoplasm of the invaded cell, and are presumably formed after PbSub2 is secreted by the parasite. Download figure Download PowerPoint Electron microscopic analysis of a stalk-shaped P.gallinaceum ookinete crossing laterally between two cells revealed that the point of entry of the stalk into the adjacent cell is surrounded by a halo of fine fibrillar material that resembles actin filaments (Vernick et al., 1999). To determine whether the stalk morphology correlated with interactions between ookinetes and the midgut actin cytoskeleton, double staining for Pbs21 and actin was performed. The stalk-shaped ookinetes were always found in the process of e
Referência(s)