Apicomplexan parasites divide and replicate through a complex process of internal budding. Daughter cells are preformed within the mother on a cytoskeletal scaffold, endowed with a set of organelles whereby in the final stages the mother disintegrates and is recycled in the emerging daughters. How the cytoskeleton and the various endomembrane systems interact in this dynamic process remains poorly understood at the molecular level. Through a random YFP fusion screen we have identified two Toxoplasma gondii proteins carrying multiple membrane occupation and recognition nexus (MORN) motifs. MORN1 is highly conserved among apicomplexans. MORN1 specifically localizes to ring structures at the apical and posterior end of the inner membrane complex and to the centrocone, a specialized nuclear structure that organizes the mitotic spindle. Time-lapse imaging of tagged MORN1 revealed that these structures are highly dynamic and appear to play a role in nuclear division and daughter cell budding. Overexpression of MORN1 resulted in severe but specific defects in nuclear segregation and daughter cell formation. We hypothesize that MORN1 functions as a linker protein between certain membrane regions and the parasite's cytoskeleton. Our initial biochemical analysis is consistent with this model. Whereas recombinant MORN1 produced in bacteria is soluble, in the parasite MORN1 was associated with the cytoskeleton after detergent extraction.
The protozoan phylum Apicomplexa is extremely species rich and members of this group are found as parasites in virtually every vertebrate and many invertebrate animals. Several species cause life-threatening diseases in humans including malaria (Plasmodium) and AIDS-associated encephalitis or gastroenteritis (Toxoplasma and Cryptosporidium). Apicomplexans are intracellular parasites and each reproductive cycle is initiated through a motile stage, the zoite. This stage invades the host cell in a complex process depending on the parasite's gliding and secretory machinery (Sibley, 2004). Although the zoite form is a structurally conserved start and end point of each cycle, the course of intracellular development has considerable morphological and functional diversity across the phylum. The parasites achieve this diversity through a remarkable flexibility of their cell and division cycle. Toxoplasma divides in a way most similar to animal cells: DNA replication is followed by nuclear division and cytokinesis resulting in two new zoites (Sheffield and Melton, 1968). Plasmodium and Eimeria proceed through several rounds of DNA synthesis and nuclear division prior to cytokinesis (Dubremetz, 1975). Sarcocystis omits nuclear division and cytokinesis for multiple rounds yielding a cell with a polyploid nucleus. This nucleus divides and segregates concomitantly with the formation of multiple daughter cells (Speer and Dubey, 1999; Vaishnava et al., 2005).
The ability of the parasites to form varying numbers of progeny (from two to thousands) is linked to their unique budding mechanism. Daughter cells pre-form within the mother as buds, are endowed with a set of organelles and emerge from the mother cell. Throughout the diverse division modes, budding is invariably linked to the final round of mitosis. Both number and localization of buds depends on the number and position of centrosomes (Dubremetz, 1975; Morrissette and Sibley, 2002a; Vaishnava et al., 2005). A critical step towards building daughter cells is the initiation of a new inner membrane complex (IMC). These flattened double-membranous cisternae delineate the forming daughters and are homologous to the alveolae found in ciliates and dinoflagellates (Cavalier-Smith, 1993). The IMC is structurally organized by a meshwork formed by a group of proteins weakly related to intermediate filaments (Gubbels et al., 2004; Hu et al., 2002a; Mann and Beckers, 2001). Initially the IMC is flexible allowing for growth but proteolytic processing of IMC components is then thought to stabilize the structure in the mature cell (Hu et al., 2002a; Mann et al., 2002). In addition to providing a pellicle, the IMC also plays a critical role as an anchor for the gliding apparatus of the parasite (Gaskins et al., 2004).
The IMC also interacts with a set of 22 subpellicular microtubules. Both structures are in intimate physical contact through an organized array of intramembranous particles (Morrissette et al., 1997), and (at least initially) grow in a coordinated fashion during budding (Hu et al., 2002a). Whether microtubule growth drives IMC extension or vice versa is less clear. Pharmacological ablation of microtubules had strong effects on the formation of daughter cells (Shaw et al., 2000). Budding is not completely abolished but is uncoupled from nuclear and organellar segregation. The extend to which budding is affected also varies depending on the parasite species and drug concentration (Morrissette and Sibley, 2002b; Vaishnava et al., 2005).
Whereas the structural elements of daughter cell formation have been characterized in some detail, how they interact and how this interaction is controlled and timed is poorly understood. In particular, how the various membranous systems and organelles tie in with the complex cytoskeleton remains elusive. In this study we describe a Toxoplasma gondii protein that participates in the assembly of the IMC. This protein, which is conserved among apicomplexans, caps the anterior and posterior end of the IMC in a ring-like structure. During cell division these rings are dynamic and contract during the final phase of budding. Thus the membrane occupation and recognition nexus (MORN)-associated ring may divide and segregate nuclei and organelles. Detergent fractionation experiments showed that this otherwise soluble protein is associated with the cytoskeleton.
Identification of two membrane occupation recognition nexus (MORN) proteins in Toxoplasma gondii
To identify parasite genes based on the subcellular localization of their products we have used a genetic strategy based on expression cloning (Gubbels et al., 2004). This screen used a library of T. gondii genomic DNA fragments fused to the yellow fluorescent protein (YFP) coding region. Clone 52C19 was initially isolated as a tag associated with the plasma membrane (data not shown). Rescue and sequence analysis revealed an open reading frame encoding a protein with similarity to membrane occupation recognition nexus (MORN) repeat proteins. Searching the T. gondii genome database (www.toxodb.org) with the cloned sequence identified two related genes (one of them tagged in the screen). Transcription of both loci was confirmed in tachyzoites by RT-PCR and 5′ and 3′ RACE-PCR (see Materials and Methods for details). Both proteins consist of 23 amino acid MORN repeats (Fig. 1). MORN1 (accession number DQ181547) and MORN2 (DQ181548, tagged in 52C19) encode predicted proteins with 39.8% sequence similarity and 24.5% identity. MORN1 shows a high level of sequence conservation across the phylum of Apicomplexa and homologs were identified in all species for which substantial sequence information is available (Fig. 1). MORN2 does not appear to have clear homologs. Based on its conservation MORN1 was studied in further detail.
Antibodies raised against recombinant MORN1 label the apical and posterior end of the inner membrane complex
MORN repeat proteins play a role in membrane-membrane and membrane-cytoskeleton interactions (Shimada et al., 2004; Takeshima et al., 2000). To explore the function of MORN1 in apicomplexans antisera were raised against purified recombinant protein. Sera were tested by western blotting using purified recombinant protein and T. gondii protein lysate. In both a band with an apparent molecular mass of 41 kDa was detected; this band was absent using pre-immune sera and no reactivity towards recombinant MORN2 (HM2) was detected (Fig. 2A). The observed mass matches the mass predicted for the MORN1 protein (40.9 kDa). To locate the protein within the parasite, fibroblast cultures infected with T. gondii tachyzoites were fixed and reacted with the MORN1 antiserum. A complex staining pattern was observed with MORN1 present at the posterior and anterior ends of the cell as well as in a single dot in the mid-cell region (Fig. 2B). Labeling at the posterior end was considerably stronger. In parasites undergoing cell division, the complexity of the pattern increased, as additional lines and dots were observed (Fig. 2E). To provide reference, cells were double labeled with an antibody specific to IMC1, a structural element of the IMC (Mann and Beckers, 2001). The MORN1 structures precisely colocalize with the apical and posterior end of the IMC in mature parasites and in the forming daughter cells (Fig. 2D,G, arrowheads indicate the apical end of cells). Interestingly, weak but reproducible labeling of an additional structure within the IMC was observed (arrow in Fig. 2G). This structure could possibly represent the micropore, an opening in the pellicle and the site of endocytosis (Nichols et al., 1994). No robust molecular markers are currently available to confirm this assignment.
MORN1 is a component of the centrocone, a specialized structure of the intranuclear mitotic spindle
To identify the localization of MORN1 at mid-cell, DNA was stained with DAPI to label the nucleus. MORN1 was present at the anterior side within the nucleus (Fig. 3A). In some cells two nuclear MORN1 dots were observed (Fig. 3B). The DNA content of individual nuclei in DAPI-stained preparations was measured by image analysis (see Materials and Methods for details) and scored for the number of nuclear MORN1 dots per nucleus. As shown in Fig. 3D, nuclei with two MORN1 dots harbor, on average, double the amount of DNA of single MORN1 dot nuclei (n=105, P<0.0001). Furthermore, in U-shaped nuclei undergoing division the two dots were associated with the leading edge of the daughter nuclei (Fig. 3C). These observations are consistent with MORN1 localization to the centrocone. The centrocone is a specialized structure of the apicomplexan nucleus associated with the intranuclear mitotic spindle (Dubremetz, 1975; Sheffield and Melton, 1968). The spindle directly adjacent to the centrosome can be detected using tagged tubulin transgenes (Striepen et al., 2000). In double labeling assays using an antibody to α-tubulin this segment of the spindle colocalizes with the nuclear MORN1 (arrowheads in Fig. 3E-G). Spindles are readily labeled in the closely related coccidian parasite Sarcocystis neurona (Vaishnava et al., 2005). As shown in Fig. S1 in supplementary material, double labeling experiments in developing S. neurona schizonts localize MORN1 (green) to the nuclear envelope precisely coinciding with the positions of the multiple spindles (α-tubulin, red) found in this organism. Sarcocystis merozoites emerging at the end of the developmental process show labeling indistinguishable from T. gondii tachyzoites.
MORN1-YFP transgenic lines were established to study the function of MORN1 during parasite development. Initial attempts using transgenes driven by the strong tubulin promoter failed to yield stable transformants (see discussion of MORN1 over-expression below). Toxicity was not due to the bulky YFP tag as transfection with constructs employing epitope tags were equally detrimental (data not shown). We reasoned that strong and/or constitutive expression might be toxic and decided to introduce the native promoter upstream of the coding region to provide appropriate expression. Transfection with this construct yielded stable fluorescent transformants at a low frequency, which were isolated by cell sorting (Southern and western blot analyses showed non-homologous insertion; data not shown).
MORN1-YFP transgenics had a normal growth rate and presented YFP localization identical to the pattern observed with the antibody (Fig. 3H). Optical sectioning, deconvolution and three-dimensional (3D) projection revealed that MORN1 structures at the posterior end of cells and buds were rings rather than lines (see Movies 1 and 2 in supplementary material). Rings could also be observed at the electron microscopic level. Fig. 3N shows a cross section through the posterior end of a tachyzoite with a ring of gold particles as a result of labeling with the MORN1 antibody. In higher magnifications of the posterior end of tachyzoites gold labeling is found on the inner (cytoplasmic) face of the IMC (Fig. 3L,M).
To provide further subcellular references additional transgenes were introduced by transfection into the MORN-YFP clone resulting in labeling with red fluorescent protein of the nucleus (H2B-mRFP), apicoplast (FNR-RFP), centrosome (centrin-RFP), and Golgi (GRASP-RFP). The Golgi, and especially the apicoplast, show association with the centrosome of the mitotic spindle during cell division (Pelletier et al., 2002; Striepen et al., 2000; Vaishnava et al., 2005). Consistent with centrocone localization of nuclear MORN1, close proximity was observed between MORN1-YFP and GRASP-RFP, FNR-RFP and centrin-RFP (arrows in Fig. 3I-K). These observations were confirmed by electron microscopy. Fig. 3O shows a longitudinal section through the apical half of a tachyzoite. A central nucleus (N) with a centrocone (black arrow) as well as the outline of an adjacent daughter bud (arrowheads) are visible. Fig. 3P and Q show two higher magnifications of centrocones. Labeling is consistently apparent in the triangular section of the centrocone (arrow) as well as in the dense material underlying the centrocone within the nucleoplasma (double arrowheads). The dense material on both sides of the membrane is of comparable density and structure, suggesting a common composition supported by the MORN1 location. Also note that the cisternae of early forming IMC (arrowheads) show MORN1 labeling at their ends on the side of the membrane facing the centrocone.
MORN1-associated structures are highly dynamic during mitosis and budding
To gain further insight into the behavior of MORN1-associated structures parasites double labeled with MORN1-YFP and histone H2B-mRFP (nucleus) were time-lapse imaged. Infected cultures were observed for 5 hours at 37°C and stacks of optical sections were taken in both channels every 10 minutes. Fig. 4 shows a panel of selected frames; the entire movie is available in the supplementary material (Movie 3). Two double-labeled tachyzoites are visible in Fig. 4A, a single dot of centrocone label is present in each nucleus (arrow in Fig. 4D) and two faint rings, the first signs of newly forming daughter buds, are seen in close proximity (double arrows). The centrocone label then increases in size and intensity and stretches into a bar perpendicular to the longitudinal axis of the parasite (Fig. 4E,F). The daughter rings move apart in association with both ends of the bar. Next the bar splits into two discrete dots (Fig. 4G). Concurrently the daughter rings further increase in size and a central (conoidal) ring becomes visible (arrow in Fig. 4G). The conoidal ring is pushed away from the nucleus towards the apex of the mother cell thus establishing the hollow cylinder shape of the IMC (arrows Fig. 4H). The posterior daughter rings then push back onto the nucleus engulfing and dividing it into two equal halves in the process (Fig. 4I,J, also see Movie 4 in the supplementary material). Finally, both rings markedly constrict, coinciding with fission of the nucleus into two daughter nuclei (Fig. 4K). Deconvolution microscopy also revealed the presence of a second MORN1 structure in the nucleus (highlighted by arrow in Fig. 4I). Labeling was considerably weaker when compared with the centrocone. This structure moves within the nucleus during mitosis and finally comes to rest at a position in the middle between the two emerging daughter nuclei.
Movements of MORN1 structures were quantified by image analysis of each deconvolved stack. Distances between the daughter centrocone and the apical (red) and posterior (green) MORN1 ring of the daughter and the posterior ring of the mother (black) as well as the diameter of the posterior daughter ring (blue) were measured (see Fig. 4B,C) and plotted over time (distances were recorded for both daughters and the average is shown). The conoidal and the posterior MORN1 ring start moving at the same time. After the nucleus slides back initially as previously observed (Radke et al., 2001) it is held at a constant mid-cell position. Note that constriction of the ring coincides with the end of the longitudinal movements.
MORN1 interacts with the parasite's cytoskeleton
The MORN1 ring is highly dynamic during daughter cell formation, however, MORN1 does not contain any ATPase domains suggesting that the force for this movement is most probably generated through interaction with other proteins. We therefore hypothesized that MORN1 might interact with elements of the parasite's cytoskeleton. The T. gondii cytoskeleton is resistant to extraction with Triton X-100 and deoxycholate (Mann et al., 2002; Morrissette et al., 1997) and co-fractionation in extraction assays has been a hallmark of proteins associated with the cytoskeleton (Gaskins et al., 2004). MORN1 showed marked resistance to Triton X-100 extraction and 82% of the protein remained in the pellet fraction (Fig. 5B, this was highly reproducible, n=3, s.d.=8.4%). By contrast, deoxycholate solubilized 80% of the MORN1. Note that recombinant MORN1 protein is soluble in 1% Triton X-100 up to a concentration of 100 μg/ml (data not shown).
The force driving the ring constriction could potentially be generated by the actin/myosin system. Several differentially localized myosins have been characterized in T. gondii. Myosin A is found at the periphery of the cell and is essential for gliding motility and invasion (Meissner et al., 2002). Myosin C (MyoC) has been previously suggested to have a role in cytokinesis (Delbac et al., 2001). To test if myosin C could be a candidate interactor and motor for the MORN1 ring, parasites stably expressing a Myc-tagged version of MyoC (a kind gift from D. Soldati, University of Geneva, Switzerland) were double labeled with antibodies to MORN1 (red) and Myc (green). In dividing tachyzoites MyoC forms distinct rings at the posterior end of the daughter buds that precisely colocalize with the MORN1 rings (arrows and insets in Fig. 5C,D,F). As shown here and reported previously (Delbac et al., 2001), MyoC staining is not limited to these rings.
Overexpression of MORN1 severely perturbs parasite nuclear division and cytokinesis
Based on the morphological and functional characterization of MORN1 we hypothesized that the overexpression toxicity observed for this gene might be due to disruption of parasite cell division. To test this hypothesis transient transfection experiments were performed using MORN1 transgenes driven by the strong tubulin promoter and parasites were subjected to immunofluorescence or live cell microscopy 24-48 hours after transfection (MORN1-RFP is shown but MORN1-YFP transgenics behaved identically). Parasites overexpressing MORN1-RFP grow to considerable size but fail to divide (compare Fig. 6B and D, which are shown at the same magnification). These cells contain a large unsegregated nucleus (compare Fig. 6C with D upper panel). The DNA content of parasite nuclei was measured 48 hours after transient transfection as described above and plotted against MORN1-RFP expression (Fig. 6E). The mean DNA content of MORN-RFP expressors (+) was on average 11-fold higher (n=120, P<0.0001) than that of untransfected cells (-). This suggests that MORN1 overexpression does not interfere with DNA replication yet prevents nuclear segregation and budding. Cells were double labeled with IMC3-YFP or YFP-TUB to highlight IMC and microtubules. Some new IMC seemed to form in MORN1 overexpressors, however, no discernable daughter buds could be identified (compare Fig. 6G and I). By contrast, overexpressors produced multiple well-formed microtubular skeletons per cell [arrow in Fig. 6K highlights one of the intensely labeled conoids (Hu et al., 2002b)]. However, the subpellicular microtubules frayed broadly and were not cupped as observed in wild-type parasites (see dotted lines in Fig. 6L,M). MORN1 in overexpressors typically localized to three intensely fluorescent rings (Fig. 6A,F,J), which showed no association with the newly formed microtubular skeletons but were labeled with an IMC marker (arrows in Fig. 6G,H).
Pharmacological analysis of MORN1 dynamics
To further study the interactions of the MORN1 rings with the cytoskeleton, pharmacological experiments were performed using actin- or tubulin-disrupting drugs, cytochalasin D and oryzalin, respectively. Cytochalasin-resistant host cells [KB100 Cyt1 a kind gift from David Sibley, Washington University (Dobrowolski and Sibley, 1996; Toyama, 1984)] were infected for 2 hours with parasites and then treated with 0, 2.5, 5 and 10 μM cytochalasin D, fixed after 24 and 48 hours and stained for MORN1. No difference in the number or morphology of MORN1 rings was observed regardless of the drug concentration (data not shown). Note that the actin/myosin-dependent motility and invasion of T. gondii is completely abolished by 1 μM cytochalasin (our controls) (Dobrowolski and Sibley, 1996). T. gondii-infected cultures were also treated with 0.05, 0.25 or 2.5 μM oryzalin. 2.5 μM oryzalin completely abolished any intracellular parasite development. At lower concentrations parasites grew to considerable size as previously reported but no clear buds were visible. However, individual sheets of IMC were formed, each associated with a region of MORN1 staining (arrowheads in Fig. 6N-P). Interestingly, the duplication of the centrocone was highly susceptible to oryzalin treatment; only a single centrocone (arrow in Fig. 6P) can be detected in the polyploid nucleus. Taken together these observations suggest that microtubules, but not actin, play a role in the biogenesis of the MORN1 ring and IMC.
MORN repeat domains are found in proteins of organisms across the tree of life. MORN repeats are often combined with additional enzymatic domains as in the phosphatidylinositol-4-phosphate 5-kinases found in many plants (Ma et al., 2004) or the human Rab5 guanine exchange factor ALS2 (Hadano et al., 2001). Most importantly in the context of this work, MORN motifs play critical roles in several proteins with roles in the organization of membranous and cytoskeletal structures. Examples are the junctophilins, which are critical for the tight appositions of endoplasmic reticulum and plasma membrane in excitable cells (Takeshima et al., 2000). Loss or mutation of these proteins prevents junctional complex assembly and leads to neuronal dysfunction (Holmes et al., 2001; Takeshima et al., 2000). MORN proteins have also been implicated in the biogenesis of the sperm flagellum (Ju and Huang, 2004; Satouh et al., 2005). Since no enzymatic activity has been associated with the MORN structure it has been suggested that MORN repeats act as protein-protein or protein-phospholipid binding domains. This would be consistent with the observation that MORN proteins are frequently part of larger protein complexes (Ju and Huang, 2004; Kunita et al., 2004; Satouh et al., 2005).
In this study we show that T. gondii MORN1 is associated with a set of subcellular structures that play key roles in parasite mitosis, daughter cell formation and cytokinesis. These structures fall into two classes: the nuclear centrocone and several IMC-associated rings (see Fig. 2 and schematic representation in Fig. 7). This suggests a role for MORN1 in capping and or anchoring the ends of the IMC cisternae. Such a role is supported by the tight colocalization of MORN1 and IMC (Fig. 2), and the coordinated movements of both structures during intracellular development (Fig. 4). Furthermore, disruption of the MORN1 rings through MORN1 overexpression prevents budding (Fig. 6). MORN1 could bind directly to the IMC membrane, however, MORN1 lacks a membrane-spanning domain or consensus motifs for lipid modification. Previous studies on junctophilins have shown that MORN repeats by themselves can provide attachment to the cytoplasmic face of the ER through interaction with phospholipids (Takeshima et al., 2000). Alternatively, MORN1 could interact with a component of the membrane skeleton that associates with the inner face of the IMC. Our electron microscopic analysis localizes the MORN1 ring to the inside of the IMC where the membrane skeleton is found (Gaskins et al., 2004; Mann and Beckers, 2001). Lastly, an integral IMC membrane protein could serve as an adapter analogous to TgGAP50, which anchors the actin/myosinA gliding machinery to the IMC (Gaskins et al., 2004). The current data do not favor one model over the other. Identifying molecular interactors of MORN1 will help to solve this question and is the focus of our future work.
The posterior MORN1/IMC ring is highly dynamic during mitosis and cell division. It moves along the longitudinal axis and constricts perpendicular to this axis at the end resulting in nuclear division and cytokinesis (Fig. 4, red and green arrows in Fig. 7D,F). How does the MORN1 ring move in the absence of motor domains in the MORN1 sequence? The tubulin and actin/myosin cytoskeletons of the parasite provide two obvious candidates. The subpellicular microtubules are organized by the apical polar ring and grow by addition of monomers at the distal end (Russell and Burns, 1984). Microtubular outgrowth seems critical to budding, as daughter cell formation is impeded under oryzalin (Morrissette and Sibley, 2002b; Shaw et al., 2000; Stokkermans et al., 1996). Daughter cell IMC still forms under oryzalin but is present as flat sheets rather than the usual hollow cylinders (Stokkermans et al., 1996) (Fig. 6). Interestingly, the ends of these sheets are still associated with MORN1 (see Fig. 6). Based on these observations we hypothesize that subpellicular microtubules are not essential to initiate new IMC and MORN1 complexes, however, they are critical to shape and extend the structure. This idea is supported by the MORN1 overexpression data where multiple intact microtubular skeletons form in the absence of functional MORN1 rings. However, in the absence of a functional MORN1 rings the distal ends of the microtubules fray and are no longer tied into the typical barrel shape.
Although microtubular growth provides a convincing model for longitudinal movement of the MORN rings during mitosis (see Fig. 7 for a schematic outline) it is unclear how microtubules could be responsible for its constriction (Fig. 4). Constrictive actin-myosin rings are well characterized elements of the cytokinesis machinery (Glotzer, 2005). Recently myosin C and actin were shown to localize to the anterior and posterior end of the parasite (Delbac et al., 2001; Patron et al., 2005). Furthermore over-expression of myosin myoB/C-tail caused a cell division phenotype (Delbac et al., 2001). In this study we have shown that MORN1 and MyoC colocalize in the posterior division ring. When MORN1 rings are disturbed by MORN1 over-expression we no longer detect myosin rings (data not shown), and budding is inhibited. These data are consistent with a model (Fig. 7F) for budding in which MORN1 acts as a linker between the posterior end of the IMC and an internal constrictive ring formed by MyoC. However, MORN1 ring formation and budding are resistant to cytochalasin D treatment [our data and a previous electron microscopic study (Shaw et al., 2000)]. Although this result does not support involvement of actin it also does not exclude it. Several clearly actin-dependent processes appear to be resistant to cytochalasin D; this includes the constriction of the cleavage furrow in fission yeast (Pelham and Chang, 2002). Further work is needed to fully delineate the components and mechanism of this constrictive ring.
MORN1 represents the first molecular marker for the centrocone. The centrocone is a unique apicomplexan structure associated with the intranuclear spindle that has been described at the electron microscopic level (Dubremetz, 1973; Dubremetz, 1975; Sheffield and Melton, 1968). It is cone shaped and delineated by the nuclear envelope with an opening at the apical end where the centrosomes are found on the cytoplasmic side (see Fig. 7E for a schematic representation). Depending on the stage in division the cone may be separated from the nucleoplasm by a membrane, which is continuous with the nuclear envelope. This membrane is perforated by spindle microtubules, which emanate from the cone and are organized by the centrosome. Electron-dense material fills the cone in between the microtubules and underlies the centrocone on the nucleoplasmic site where the kinetochores are attached (Dubremetz, 1973). MORN1 is found in both of these locations (Fig. 3O-N, Fig. 7E). MORN1 overexpression prevents the formation of new centrocones as the cells progress through the cell cycle. Formation of new centrocones is equally dependent on microtubules. Whereas multiple IMC-MORN1 structures are formed under oryzalin treatment only a single centrocone is found within the large polyploid nucleus. This is consistent with the view that the centrocone cannot be formed de novo and requires duplication, which in turn requires a functional mitotic spindle.
Spindles are generally restricted to mitosis. We were therefore surprised to find a MORN1-labeled structure consistent with the location of the centrocone in every tachyzoite regardless of its stage in the cell and division cycle. We have recently described persistence of short spindles and centrocones through interphase in the closely related parasite S. neurona (Vaishnava et al., 2005). As expected, MORN1 co-localizes with spindles in S. neurona (Fig. S1 in supplementary material). The new observation in T. gondii suggests that centrocone persistence (and with it potentially persistence of kinetochore attachment) could be a general theme of apicomplexan nuclear organization.
Materials and Methods
Parasites and cells
RH-strain T. gondii tachyzoites were passaged in confluent human foreskin fibroblasts as described previously (Roos et al., 1994). Sarcocystis neurona merozoites were passaged in primary bovine turbinate cells as described previously (Vaishnava et al., 2005). MORN1-YFP parasites were isolated after transfection with plasmid 5′morn1MORN1-YFP/DHFR/3′-morn1 under pyrimethamine selection and enrichment using a MoFlo high-speed cell sorter (DAKO/Cytomation, Fort Collins, CO, USA) as described previously (Gubbels et al., 2004). The following lines have been described previously: IMC3-YFP (Gubbels et al., 2004), YFP-α-tubulin (Striepen et al., 2000) and Myc-myosin C (Delbac et al., 2001).
Recovered sequences were searched against www.ToxoDB.org, www.PlasmoDB.org, www.CryptoDB.org, www.GeneDB.org and the NCBI non-redundant database. Sequences were aligned using ClustalW (www.ebi.ac.uk/clustalw), and prepared for publication using BoxShade (http://bioweb.pasteur.fr/seqanal/interfaces/boxshade.html). Identity and similarity numbers were calculated using MatGat software Version 2.03 (Campanella et al., 2003).
5′- and 3′-RACE of MORN1 and MORN2
Total RNA was extracted from RH tachyzoites using Trizol (Invitrogen). RT-PCR, 5′-RACE and 3′-RACE were performed using the SMART RACE cDNA amplification kit (BD BioSciences) following the manufacturer's instructions using primer 1 and 2 (MORN1), and 3 and 4 (MORN2; see supplementary material Table 1 for the sequences of primers used in this study). Amplified fragments were cloned into pCR2.1 using the TOPO-TA cloning protocol (Invitrogen).
Plasmid tubMORN1-YFP/sagCAT was cloned by PCR amplification of the MORN1 ORF from genomic DNA using primers 5 and 6 and replacement of the first YFP in tubYFPYFP/sagCAT (Gubbels et al., 2003) by BglII/AvrII digestion. 5′morn1-MORN1-YFP/sagCAT was generated by amplifying the 1920 bp preceding the start-codon using primers 7 and 8 and replacing the tub-promoter in tubMORN1-YFP/sagCAT by PmeI/BglII digestion. Plasmid 5′morn1-MORN1-YFP-3′dhfr/sagDHFR/3′morn1 was constructed by amplification of 5′morn1-MORN1-YFP-3′dhfr using primers 9 and 10 and cloned at the ApaI/ClaI sites into pKO-DHFR (see below) resulting in 5′morn1MORN1-YFP/DHFR. The 3′morn1 genomic region was amplified (primers 11 and 12) and inserted by XbaI/NotI digestion into plasmid 5′morn1MORN1-YFP/DHFR resulting in 5′morn1-MORN1-YFP-3′dhfr/sagDHFR/3′morn1. pKO-DHFR was constructed from the pKO plasmid (Martin Gastens, unpublished; this pKS-based plasmid carries a T. gondii tub-CAT-sag1 cassette flanked on both sides by a multiple cloning site); first the tub promoter was replaced with the sag promoter from the sagCATsag cassette by HindIII/NsiI digestion. The CAT gene was replaced with DHFR-TS m2m3 ORF using NsiI/PacI digestion [primers 14 and 15; the internal BglII site within DHFR-TS was deleted using primer 13 in the megaprimer approach (Lai et al., 2003)].
Plasmid morn1CentrinYFP/sagCAT was constructed by amplifying the centrin ORF from cDNA (primers 16 and 17) and cloning via BglII/AvrII digestion into morn1-MORN1-YFP/CAT. Plasmid tubH2B-mRFP/sagCAT was constructed by replacing YFP (BglII/AvrII) in tubYFP-mRFP/CAT with H2B from tubH2B-YFP/CAT [kindly provided by K. Hu and D. S. Roos, University of Pennsylvania (Hu et al., 2004)]. The monomeric mRFP originated from the pDsRed-monomer-N1 plasmid (BD BioSciences) and was PCR amplified (primers 18 and 19) deleting the internal BglII site using the megaprimer approach (primer 20). The amplified mRFP was cloned AvrII/AscI into a tubYFP-YFP(AscI)/sagCAT plasmid (Michael Kurth, M.J.G. and B.S., unpublished). Plasmid tubGRASP55-RFP/sagCAT was cloned by excision of GRASP55 from tubGRASP55-YFP/sagCAT (Pelletier et al., 2002) (kindly provided by G. Warren, Yale University, USA) and replacing ferredoxin NADH reductase (FNR) in tubFNR-RFP/sagCAT (Striepen et al., 2000).
Recombinant protein expression, antibody production and affinity purification
The MORN1 ORF was amplified from genomic DNA using primers 21 and 22 and cloned by ligation independent cloning into plasmid pAVA421 (Alexandrov et al., 2004) as described previously (Aslanidis and de Jong, 1990) to generate a six His-tagged N-terminal fusion. Full-length MORN2 was similarly cloned (primers 23 and 24). Recombinant fusion proteins were purified in the presence of 6 M urea on Ni2+-NTA resin (Qiagen, Hilden, Germany). Polyclonal antibody was generated by rabbit immunizations (Cocalico Biologicals, Reamstown, PA, USA). MORN1 antibodies were affinity purified against purified His-tagged protein cross-linked to activated CNBr Sepharose 4B (Sigma, St Louis, MO, USA) as described previously (Harlow and Lane, 1988).
Immunofluorescence and microscopy
Images were taken using a DM IRB inverted microscope (Leica) equipped with a PL-APO 100×/1.4 NA oil emersion lens and a Hamamatsu C4742-95 camera and adjusted for contrast using Openlab software (Improvision). For immunofluorescence we used the following antibodies: affinity purified anti-MORN1 (1:100), monoclonal 45.15 anti-IMC1 (1:1000; a kind gift from Gary Ward, University of Vermont, USA (Mann and Beckers, 2001), rabbit anti-GFP (1:5000; Torrey Pines Biolabs, CA, USA), monoclonal antibody 12G10 anti-α-tubulin (1:10; a kind gift from Jacek Gaertig, University of Georgia, USA) (Jerka-Dziadosz and Frankel, 1995), monoclonal anti-Myc epitope tag 9E10 (1:1000). Goat anti-rabbit and goat anti-mouse antibodies conjugated with Alexa Fluor 488 or Alexa Fluor 546 (Molecular Probes, Eugene, OR, USA) were used as secondary antibodies at 1:200 dilution. DNA was visualized by staining with 4′,6-diamidino-2-phenylindole (DAPI; 2 μg/ml in PBS) for 5 minutes. To measure the DNA content of individual nuclei, the intensity of DAPI staining was measured in situ by image analysis. Images were recorded with constant exposure time and within the linear range of the CCD (the contrast was not adjusted). Nuclei were defined as objects by generating binary image masks using Openlab software. The mean pixel intensity was multiplied by the area for each mask to obtain a cumulative intensity measurement in arbitrary units. More than 50 nuclei were measured for each sample and statistical significance was evaluated using the Student's t-test. For deconvolution, specific time-lapse experiments (as indicated), were performed on a DeltaVisionRT microscope (Applied Precision, Issaquah, WA, USA). SoftWoRx software was used for image analysis and presentation.
Vero cells infected for 24 hours with T. gondii were resuspended using 5 mM EDTA in PBS, pelleted, and fixed for 15 minutes at room temperature with 2% paraformaldehyde and 0.1% glutaraldehyde in 0.2 M sodium phosphate buffer pH 7.4. The pellet was dehydrated in ethanol at -20°C and embedded in LRWhite (London Resin Co, Berkshire, UK). Thin sections were collected on carbon-coated grids and saturated for 30 minutes with 2.5% non fat dry milk and 0.1% Tween 20 in PBS (PBS-MT). The grids were floated successively for 1 hour each on affinity purified rabbit anti-MORN1 serum diluted 1:40, followed by protein A 10 nm gold diluted in PBS-MT, with PBS washes between each step. The grids were then stained with uranyl acetate and lead citrate and observed with a Jeol 1200EX electron microscope operated at 80 kV.
Detergent extractions were performed essentially as described previously (Gaskins et al., 2004). Extracts were spun for 20 minutes at 4°C, supernatants were recovered and pellets solubilized in SDS extraction buffer. Samples were run on 10% Bis-Tris gels and transferred to nitrocellulose. Blots were probed with the following antibodies: rabbit anti-MORN1 (1:1000), monoclonal 45:15 anti-IMC-1 (1:2000); monoclonal 12G10 anti-α-tubulin (1:500), rabbit anti-GFP (1:5000) followed by either goat anti-rabbit or goat anti-mouse conjugated to horseradish peroxidase (1:3000; Bio-Rad). Enzyme activity was visualized using enhanced chemiluminescence (Pharmacia) and signals were recorded on X-ray film or directly quantified on a GeneGnome (Syngene, Cambridge, UK).
This work was funded in part by grants from NIH-NIAID to B.S. and a postdoctoral fellowship to M.J.G. from the American Heart Association. We thank Julie Nelson for help with cell sorting, Véronique Richard (SCME-UM2) for immunoelectron microscopy specimen preparation. We are also grateful to D. Soldati, G. Warren, D. S. Roos, L. D. Sibley, G. Ward, and J. Gaertig for antibodies, plasmids and parasite strains, D. Soldati and L. D. Sibley for discussion, and J. Gaertig for critical reading.