Plasmodium parasites are fertilized in the mosquito midgut and develop into motile zygotes, called ookinetes, which invade the midgut epithelium. Here we show that a calcium-dependent protein kinase, CDPK3, of the rodent malarial parasite (Plasmodium berghei) is produced in the ookinete stage and has a critical role in parasite transmission to the mosquito vector. Targeted disruption of the CDPK3 gene decreased ookinete ability to infect the mosquito midgut by nearly two orders of magnitude. Electron microscopic analyses demonstrated that the disruptant ookinetes could not access midgut epithelial cells by traversing the layer covering the cell surface. An in vitro migration assay showed that these ookinetes lack the ability to migrate through an artificial gel, suggesting that this defect caused their failure to access the epithelium. In vitro migration assays also suggested that this motility is induced in the wild type by mobilization of intracellular stored calcium. These results indicate that a signalling pathway involving calcium and CDPK3 regulates ookinete penetration of the layer covering the midgut epithelium. Because humans do not possess CDPK family proteins, CDPK3 is a good target for blocking malarial transmission to the mosquito vector.
When mosquitoes feed on an infected vertebrate host, sexual stages of malarial parasites are transferred to the mosquito midgut. In the blood meal, they are fertilized and develop into motile zygotes, called ookinetes, which migrate towards the midgut epithelium. On the way, the ookinetes meet with a biological barrier, the peritrophic membrane or matrix (PM). The PM is formed over the midgut epithelium after a blood meal and is believed to have a role in protecting the epithelium from invasion by pathogens (Berner et al., 1983). After traversing this barrier, the ookinetes invade the epithelial cells and migrate towards the basal lamina to form oocysts, within which numerous sporozoites develop (Han et al., 2000; Vlachou et al., 2004; Baton and Ranford-Cartwright, 2005). The efficiency of infection of the mosquito midgut is critical for malarial parasites to succeed in transmission to the mosquito vector. Understanding of parasite–host interactions in midgut invasion process is therefore important for developing strategies to block parasite transmission.
Calcium-dependent protein kinase (CDPK) has been identified in plants, green algae, ciliates and apicomplexan parasites including malarial parasites, but not in mammals. CDPK has a calmodulin-like domain with four EF-hand motifs (Harper and Harmon, 2005) and is activated by calcium ions in the absence of calmodulin or phospholipids. In this feature it is different from calmodulin-dependent protein kinase and protein kinase C, which play central roles in calcium signalling in mammals. Recent genomic analysis has revealed that 5 CDPK isoforms (CDPK1–5) and some structurally related kinases are encoded in the genome of the human malarial parasite, Plasmodium falciparum (Ward et al., 2004). Among them, CDPK4 is expressed in the gamete/gametocyte stage and has a critical role in male gametogenesis that is induced by mosquito xanthurenic acid and following intracellular calcium mobilization (Billker et al., 2004). The function of other CDPK isoforms remains unclear, but they are thought to have important roles in calcium signalling in the parasite life cycle.
In apicomplexan parasites, calcium signalling has an important role in regulation of their invasive motility. In Toxoplasma gondii tachyzoites and Cryptosporidium parvum sporozoites, chelating of intracellular calcium ions inhibits secretion of apical organelles including micronemes (apical organelles that are involved in parasite migration and host cell invasion), and treatment with a calcium ionophore conversely promotes the secretion (Lovett et al., 2002; Lovett and Sibley, 2003; Chen et al., 2004; Wetzel et al., 2004). Also in malarial sporozoites, liver invasive forms, it has been reported that a rise in intracellular calcium regulates discharge of micronemes (Mota et al., 2002). In T. gondii a CDPK isoform (TgCDPK1) is produced in the tachyzoite stage. CDPK activity is involved in the parasite's cell attachment and migration (Kieschnick et al., 2001), and these events are concomitant with the rise of intracellular calcium (Silva-Neto et al., 2002; Reininger et al., 2005), suggesting that calcium signalling involving CDPKs has an important role in regulation of invasive motility.
Here we show that an isoform of plasmodium CDPK, CDPK3 (Li et al., 2000), is produced in the ookinete stage of a rodent malarial parasite, P. berghei, and regulates ookinete motility for invasion of the mosquito midgut. We show that CDPK3 is necessary for ookinetes to access epithelial cells by crossing the layer between a blood meal and the midgut epithelium. We suggest that a signalling pathway involving intracellular calcium and CDPK3 activates ookinete motility for traversing this layer, permitting ookinete invasion of the midgut.
CDPK3 is produced in the ookinete
pfCDPK3 (PFC420w) was identified in P. falciparum as the third CDPK-family protein in plasmodium species, and its transcription was reported to occur in the sexual stage (Li et al., 2000). However, recent EST and DNA microarray studies have showed that P. berghei CDPK3 is transcribed in the ookinete stage (Abraham et al., 2004; Vontas et al., 2005). We also analysed our P. berghei EST database and confirmed that six ESTs for PbCDPK3 were included in ookinete ESTs.
We prepared polyclonal antibodies and investigated the expression profile of this gene. In Western blot analysis, CDPK3 was not detected in the blood stage, but was found in the late stage of ookinete formation (at 16 h after beginning of ookinete culture) and increased with maturation of the ookinetes (Fig. 1A). This pattern of expression was similar to that of ookinete micronemal proteins including CTRP (Yuda et al., 1999a) and WARP (Yuda et al., 2001), suggesting that this CDPK isoform is involved in ookinete invasion of the mosquito midgut. Indirect immunofluorescent microscopy confirmed that PbCDPK3 is produced in the ookinete stage (Fig. 1B; for a negative control using the CDPK3 antibodies, see also Fig. 2E).
Disruption of the CDPK3 gene
To investigate the function of CDPK3, targeted disruption of cdpk3 was performed in P. berghei. The CDPK3 locus was replaced with a construct containing the pyrimethamine-resistant DHFR-TS (dihydrofolate reductase-thymidylate synthase) gene, and disruptant parasites were selected with pyrimethamine (Fig. 2A). To exclude that the phenotype might be judged by a single clone, two clones [cdpk3(-)1 and cdpk3(-)2] were obtained from independent disruption, and in each clone the expected replacement of the cdpk3 locus was confirmed by polymerase chain reaction (PCR) (data not shown) and genomic Southern blot analysis (Fig. 2B). Disruption of CDPK3 was further confirmed by Western blot analysis (Fig. 2D) and indirect immunofluorescence analysis (Fig. 2E) in cultured ookinetes.
cdpk3(-) parasites showed normal exflagellation rates (data not shown). They developed into ookinetes in vitro as efficiently as wild-type parasites (Table 1). These ookinetes were morphologically normal as observed with Giemsa staining (Fig. 2C), and production and distribution of CTRP (circumsporozoite- and TRAP-related protein), a microneme protein essential for ookinete motility (Dessens et al., 1999; Yuda et al., 1999b; Templeton et al., 2000), was normal (Fig. 2E, bottom panels). In addition, disruptants developed into mature ookinetes in the mosquito midgut (data not shown).
Table 1. CDPK3 disruption reduced the number of formed oocysts in the mosquito midgut.
. Ookinetes were produced by in vitro culture. The highest ratios of produced ookinetes to red blood cells are shown.
. Midguts were dissected 14 days after blood feeding and examined under a transmission microscope to count the number of oocystes (n = 15).
. Sporozoites were collected separately from midguts (n = 15) and salivary glands (n = 30) 20–24 days after feeding and counted. Values are means (± SEM) of three independent experiments.
Mosquitoes were fed on mice infected with wild-type or cdpk3(-) parasite populations.
NT, not tested.
210 ± 36
141 444 ± 5 556
18 113 ± 1 249
1.51 ± 0.51
1 074 ± 643
907 ± 275
1.36 ± 0.40
1 289 ± 320
1 057 ± 23.4
1.28 ± 0.51
1 228 ± 256
975 ± 260
CDPK3 is necessary for transmission to mosquitoes
To assess the transmission efficiency of the disruptants, mosquitoes were fed on infected mice and oocyst number in the midgut was counted. The number of oocysts was greatly reduced in disruptants, on average to less than 1% of that of the wild type (Table 1). Sporozoites in oocysts were also decreased to approximately 1% of wild-type numbers, indicating that the cdpk3 disruption did not affect sporozoite formation in oocysts. The relative percentage of disruptants was increased in the salivary gland (approximately 5% of wild type). This is probably because the number of sporozoites that can infect salivary glands reached an upper limit in the wild type. To assess liver infectivity, the salivary gland sporozoites were inoculated intravenously into rats and the increase in parasitaemias was monitored (Fig. 3). The increase in the parasitaemias was similar between disruptants and wild type, indicating that the liver infectivity (and also the infectivity to the erythrocyte) of disruptants is normal. These results demonstrated that CDPK3 has a critical role in the parasite's infection of the mosquito midgut and is required solely for this step.
CDPK3 is necessary for ookinete migration to the epithelial cells
To investigate the cause of the reduced oocyst number, we examined whether the disruptant ookinetes can invade midgut epithelial cells. Mosquitoes were fed on infected mice, and midguts were dissected after 21 h. Sections of these midguts were stained with toluidine blue, which has been reported to stain epithelial cells injured by ookinete invasion deeply, and the number of injured epithelial cells was counted (Fig. 4A and B). When mosquitoes were fed on mice infected with wild type, approximately 3% of epithelial cells (302/10248) were injured. In contrast, less than 0.1% of epithelial cells (8/10485) were injured with the cdpk3(-) parasite. This result indicates that most ookinetes with the gene disruption became unable to invade midgut epithelial cells.
We further investigated which step of the invasion was affected in the disruptants. Mosquitoes were fed on infected mice and the location of disruptant ookinetes in the midgut was examined at 21 h after blood meal by transmission electron microscopy (TEM). In the wild type, 35% of ookinetes invaded epithelial cells, and half of them had already reached the basal lamina (Table 2; Fig. 4C, upper panel). In contrast, with cdpk3(-) parasites, all but one ookinete (118/119) were detected in the midgut lumen (Table 2). Approximately half of them were located on the luminal side of the layer between the microvillous surface of epithelial cells and the ingested blood (Fig. 4C, lower panel). Taken together, these results indicate that CDPK3 is necessary for ookinetes to cross this layer and gain access to the epithelial cell surface.
Table 2. Localization of ookinetes in mosquito midguts.
Location of ookinetes
Wild type (%) (n = 43)
cdpk3(-) (%) (n = 119)
Mosquito midguts were fixed 21 h after a blood meal including wild-type or cdpk3-disrupted parasites. Ookinetes were classified according to their location detected by electron microscopy.
Between erythrocytes and epithelial cells
Attached to epithelial cells
Attached to basal membrane
CDPK3 is required for ookinete migration through Matrigel
For analysis of ookinete motility, we generated green fluorescent protein (GFP)-expressing wild-type and cdpk3(-) parasites. First, GFP-expressing wild-type parasites were prepared (see Experimental procedures and Figure S1A). The transgenic parasites (wt/gfp) produce GFP throughout the life cycle including the ookinete stage (Fig. 5A, upper panel). These wt/gfp parasites showed essentially the same growth rates and infectivities as the wild type throughout the life cycle (data not shown). Next, GFP-expressing cdpk3(-) parasites were generated by mating of wt/gfp and cdpk3(-) parasites (see Experimental procedures and Fig. S1B). The obtained cdpk3(-)/gfp (Fig. 5A, lower panel) showed almost the same phenotypes as the original cdpk3(-) parasites, including the decrease in oocyst number (Table 1).
Using these mutant parasites, we assessed the motility of ookinetes. Cultured GFP-expressing wild-type or cdpk3(-) ookinetes were inoculated to the top of a layer of gelled basement membrane matrix (Matrigel) in a 24 well plate and were allowed to migrate. After 8 h the number of ookinetes that entered the Matrigel and the depth they reached were measured by optical and confocal microscopy (Fig. 5B and C). Soon after inoculation, almost all wild-type ookinetes entered the gel and started moving towards the bottom of the layer (Supplementary material Videos 1 and 3). In contrast most cdpk3(-) ookinetes remained on the gel surface (indicated by an arrowhead). The small proportion of disruptant ookinetes (6%) that succeeded in entering the gel remained near the surface (see representative confocal microscopy images in Fig. 5C). These results indicated that cdpk3(-) ookinetes are severely impaired in the ability to migrate through a gel structure, resulting in failure to cross the layer covering the surface of the midgut epithelium.
Real time images of cdpk3(-) ookinetes showed that they are able to move on the gel surface (Supplementary material Videos 2 and 4). This indicates that without CDPK3 ookinetes have a surface motility that is distinguishable from their ability to traverse a gel.
After another 3 days in culture, all wild-type ookinetes had changed to early stage oocysts regardless of the depth they had reached (Fig. 5D, left panel). The cdpk3-disrupted ookinetes formed aggregates and also changed to early oocysts on the surface of the gel (Fig. 5D, right panel), showing that they retain the ability to transform into oocysts.
Ookinete migration through the matrix is inhibited by an intracellular calcium chelator and kinase inhibitors
The requirements for Ca2+ and kinase activity for ookinete migration were examined (Fig. 6). Intracellular and extracellular calcium chelators were added to the medium, and the number of ookinetes that entered the Matrigel was counted. The extracellular calcium chelator, ethyleneglycol bis (2-aminoethylether) tetraacetic acid (EGTA), had no effect on ookinete migration. However, the intracellular calcium chelator, BAPTA [1,2-bis (o-aminophenoxy) ethane-N,N,N′,N′-tetraacetic acid tetra (acetoxymethyl) ester] inhibited ookinete invasion of Matrigel in a dose-dependent manner. At the concentration required for 70% inhibition of intragel mobility (100 µM), ookinetes still actively moved on the gel surface (data not shown). These results indicated that ookinete migration through the Matrigel requires mobilization of stored calcium ions.
Next, different types of kinase inhibitors were added to the medium. Ookinete movement was inhibited by staurosporine (a broad-spectrum protein kinase inhibitor), W-7 (a calmodulin antagonist), and KN-93 (an inhibitor for calmodulin kinase II). ML-7, a specific inhibitor for myosin light chain kinase, had no effect on ookinete migration. Taken together, these results showed that ookinete migration through the matrix required both kinase activity and mobilization of stored Ca2+.
This study has demonstrated that a Plasmodium CDPK, CDPK3, has a crucial role in malarial transmission to the mosquito vector, being required for ookinete migration to midgut epithelial cells. TEM observation revealed that cdpk3 disruptant ookinetes stopped in the layer covering the surface of the epithelium and were unable to access the epithelial cells. Matrigel in vitro migration assays indicated that disruptants lack the ability to migrate through gel structure, suggesting that this motility is necessary to cross the midgut layer.
In mosquitoes the PM is formed over the midgut epithelium after feeding and separates the epithelial cells from a blood meal. The function of this layer remains unclear, but it is suggested that it has a role in protecting the epithelium from invasion of pathogens and facilitating blood digestion. It has been reported that in Anopheles stephensi, the PM is detected in the midgut lumen at 12 h after a blood meal, and its formation is completed by 48 h (Berner et al., 1983). Because mature ookinetes have not been formed at 12 h after a blood meal, this implies that the PM, although incomplete, is already present, when ookinetes start migration to the epithelium. On the contrary, it has been reported that PM is sometimes not found in A. stephensi (Schneider et al., 1986). In our TEM observation, laminar structure of the PM was not observed clearly at 21 h after a blood meal. The surface of the epithelium was covered with a layer containing electron dense coarse material, presumably derived from digested erythrocytes, and most disruptant ookinetes stopped at the luminal side of this layer. At present it is unclear whether this layer is an immature PM or has an independent origin, but our results strongly suggest that a gel structure is formed over the midgut epithelium at this time and acts as a barrier against ookinete invasion. Activation of a signalling pathway involving CDPK3 may be necessary to activate the ability to migrate through this layer.
Although disruptant ookinetes almost lack ability to move through this layer, they seemed still to be able to migrate through a blood meal. In fact, our TEM study showed that the proportion of ookinetes detected in the layer between a blood meal and the epithelium was higher in disruptants (52%) than in the wild type (25%) (Table 2). This suggests that the disruptants moved to this layer as wild types did but could not proceed further, being blocked by it. In vitro images also support this observation. While most disruptant ookinetes could not enter the Matrigel, they still actively moved on its surface (Supplementary material Video 2). Therefore, it appears that the motility through a blood meal and the motility to cross the layer over the epithelial cell surface are distinct. CDPK3 might be necessary for ookinetes to switch modes of motility from that for migrating through a blood meal to that for migrating through the gel structure. Tests with calcium chelators strongly suggest that this switching is triggered by mobilization of an internal store of Ca2+, which may lead to activation of CDPK3. At present it is unclear what triggers this pathway, but considering that ookinete motility is observed even in an artificial gel, the signalling pathway might be activated by mechanical contact of ookinetes with an gel structure that blocks their movement. To investigate what factors induce calcium mobilization, imaging of intracellular calcium dynamics is in progress.
The present study revealed that the regulation of ookinete motility by CDPK3 has a critical role in malarial transmission to the mosquito vector. So far, little is known about ookinete motility to access the midgut epithelium, including motility to penetrate the PM. Our study showed that ability to move through a gel structure is necessary in this step and that this motility is regulated by a signalling pathway involving calcium and CDPK3. It has been reported that in T. gondii activation of CDPK is necessary for the parasite's invasion of the host cell. Our results demonstrated that CDPK is necessary for regulation of invasive motility in a malarial parasite, indicating that CDPK is a potential drug target for preventing parasite motility and blocking malarial infection.
Female BALB/c mice (6–10 weeks old, Japan SLC, Hamamatsu, Japan) infected with P. berghei ANKA strain were prepared by peritoneal injection of infected blood that was stored at −70°C. Ookinetes were prepared as follows. Stored infected blood was injected intraperitoneously into mice that were made anaemic by phenyl-hydrazine treatment. About 6 days later, the infected blood was collected from the mice and diluted 10-fold with RPMI1600 medium (Gibco, Gaithersburg, MD) containing 20% FCS and penicillin/streptomycin, and incubated at 20°C for a further 20–24 h. Ookinete formation was checked by Giemsa staining. Midgut and salivary gland sporozoites were collected from infected mosquitoes and counted as described previously (Kariu et al., 2002).
Antibody preparation and Western blot analysis
The C-terminal portion of CDPK3 (amino acid residues 488–554) was produced as a glutathione S–transferase fusion protein using the pGEX 6p-1 system (Amersham Bioscience, Piscataway, NJ). The primers used were 5′-CGGGATCCGC CTTTAGGGTATTTGATGTAGAC-3′ and 5′-CCGCTCGAGT TAATACTTTAGTTTCATCATTTCGC-3′. The recombinant protein was purified with a glutathione Sepharose column (Amersham Bioscience) and used for immunization of rabbits. Specific antibodies were purified from the rabbit antiserum as described previously (Ishino et al., 2004). Western blot analysis was performed as described previously (Kariu et al., 2002). Briefly, samples were separated by SDS-electrophoresis and transferred onto a polyvinylidene difluoride (PVDF) membrane. After blocking with 1% BSA-PBS solution, the membrane was incubated with 10 µg ml−1 anti-CDPK3 antibody, followed by incubation with anti-rabbit IgG antibody conjugated to alkaline phosphatase (Bio-Rad, Hercules, CA). Signals were detected with BCIP (5 bromo-4-chloro-3-indolyl phosphate) and NBT (nitroblue tetrazolium) (Bio-Rad).
Immunofluorescence microscopy was performed as described previously (Kariu et al., 2002). Briefly, purified parasites on a glass slide were fixed in acetone for 2 min. The slide was incubated with anti-CDPK3 antibodies or anti-CTRP antibody and then with fluorescein isothiocyanate (FITC)-conjugated secondary antibody (Zymed, South San Francisco, CA). The samples were examined with a fluorescent microscope (Olympus, Tokyo, Japan).
Targeted disruption of cdpk3
For construction of a targeting vector for cdpk3 disruption, two DNA fragments of the gene were amplified by PCR using P. berghei genomic DNA as template. The primer sets used are as follows: 5′-AAAGAGCTCCTTCTTAAATTTGTCCACT TGCC-3′ and 5′-AAAGGATCCGCAGCTAAACCAAAGTC GATG-3′; 5′CCGCTCGAGGAATAATATCTCAGAAGAGGCT AAG-3′ and 5′-GCCGGTACCGCCTCATTAAAACAAGAGTC TACAAC-3′. The amplified fragments were ligated to either side of the selectable marker gene (pyrimethamine-resistant P. berghei DHFR-TS gene) in pBluescript (Stratagene, La Jolla, CA). Gene targeting experiments were performed as described previously (Yuda et al., 1999b).
Genomic Southern hybridization
Genomic DNA of P. berghei (2 µg) was digested with HindIII, separated on 1.5% agarose gel, and then transferred onto a nylon membrane. DNA fragments were amplified by PCR using genomic DNA as template with the following primer pair: 5′-GCTGGAACTCCTTATTATGTTGCC-3′ and 5′-GCT TCAATCCACTATTTTCCAAACC-3′. The amplified DNA was labelled with alkaline phosphatase and used for a probe. Signals were detected using a CDP-Star chemiluminescent detection reagent (Amersham Bioscience).
Counting injured midgut epithelial cells and determination of ookinete location in the midgut
Mosquito midguts were dissected at 21 h after a blood meal. They were fixed in 0.1 M phosphate buffer (pH 7.4) containing 1% paraformaldehyde-2.5% glutaraldehyde (TAAB, Berkshire, UK) for 90 min on ice, dehydrated in ethanol, and embedded in Quetol 651 (Nishin EM, Tokyo, Japan). To count injured epithelial cells, serial semi-thin sections were obtained from the posterior portion of the midgut, stained with toluidine blue solution, and examined under an optical microscope. The location of ookinetes was determined by TEM. Ultrathin sections were prepared from the same embedded midguts and stained with 2% uranyl acetate and Reynald's lead citrate. Ookinetes were examined in each section with a HITACHI H-800 transmission electron microscope at 100 kV.
Generation of GFP-expressing cdpk3(-) parasites
Green fluorescent protein-expressing parasites were generated as follows (Fig. S1A). To make a GFP expression cassette driven by the hsp70 promoter, the 1.2 kb upstream region and 1.0 kb downstream region of hsp70 were ligated to either side of the GFP coding sequence. To make a targeted insertion vector, this expression cassette was ligated into the downstream region of the pyrimethamine-resistant DHFR-TS gene. A transgenic GFP-expressing parasite clone (named WT : GFP) was obtained by inserting this construct downstream of the dhfr/ts locus by homologous recombination. GFP-expressing cdpk3(-) parasites were prepared by genetic cross of WT : GFP and cdpk3(-) parasites (Fig. S1B). Briefly, cdpk3(-) parasites and GFP-expressing parasites (ratio 9:1) were injected into a mouse intravenously. Mosquitoes were fed on this mouse, and sporozoites were collected from mosquito salivary glands and injected into a rat intravenously. When parasitaemia attained ∼1%, erythrocytes infected by GFP-expressing parasites were sorted by an Epics Altra (Beckman Coulter, Fullerton, CA), and injected into a fresh rat. Finally, cdpk3(-)-GFP-expressing parasites were separated by limiting dilution.
Evaluation of sporozoite liver infectivity
Liver infectivity of sporozoites was evaluated as described previously (Ishino et al., 2004). Briefly, salivary gland sporozoites (30 000 each) were suspended in medium 199, and then injected intravenously into 3-week-old female Wistar rats (n = 5; Japan SLC). Parasitaemia was checked by Giemsa-stained blood smears twice daily from 3 days after inoculation.
Evaluation of ookinete migration activity in vitro
Green fluorescent protein-expressing wild-type and cdpk3(-) ookinetes were prepared by in vitro culture as described above. These ookinetes were diluted with oocyst culture medium containing 20% FCS, 14.7 mM hypoxanthine, 44 µM PABA, penicillin/streptomycin (modified from Al-Olayan et al., 2002) to 1.0 × 106 ookinetes in 500 µl medium, added to a 24 well plate (Nunc, Napierville, IL) coated with Matrigel (Clontech, Bedford, MA), and incubated at 20°C. Eight hours later, the number of ookinetes located inside the Matrigel was counted under a fluorescent inverted microscope (Nikon, Tokyo, Japan). Images of ventral sections were obtained using a LSM 5 LIVE confocal laser scanning microscope (Carl Zeiss, Jena, Germany). For reagent treatment assays, ookinetes were pretreated with BAPTA-AM (Calbiochem, La Jolla, CA), EGTA (Dojindo Laboratories, Kumamoto, Japan), Staurosporine (Calbiochem, La Jolla, CA), W-7 (Calbiochem), KN-93 (Calbiochem) or ML-7 (Calbiochem) at indicated concentrations for 10 min at room temperature. Then each ookinete solution was added to Matrigel-coated wells.
We would like to thank Sumina Kido for TEM; Tomomi Kato and Izumi Kaneko for technical assistance. We also thank Gerard R. Wyatt for reading the manuscript. This study was supported by a grant-in-aid for Scientific Research on Priority Areas to YC (16017243,14207011,16659110), and to MY (16390124, 16659111), of the Ministry of Education, Science, Culture, and Sports of Japan. It was also supported by grants from the Japan Science and Technology Agency to YC.
The following supplementary material is available for this article online:
Fig. S1. Construction of GFP-expressing cdpk3(-) parasites.
A. Targeted insertion of GFP-expressing cassette into the dhfr/ts locus. The GFP coding region (green box) was ligated to the 1.2 kb upstream region and the 1.0 kb downstream region of the HSP70 gene (blue line). This expression cassette was inserted between the pyrimethamine-resistant DHFR-TS gene and its downstream region. This construct was inserted downstream of the dhfr/ts locus by homologous recombination.
B. Generation of GFP-expressing cdpk3(-) parasites. Mosquitoes were fed on a mouse infected with GFP-expressing wild-type parasites (WT : GFP) and cdpk3(-) (KO) parasites in the ratio of 1:9. In the mosquito midgut, they mate and generate ookinetes with three genotypes. As cdpk3(-)/cdpk3(-) parasites hardly infect mosquitoes, the proportion of cdpk3(-) parasites in all GFP-expressing sporozoites was expected to be near 50%. About 20 days later, sporozoites were collected from salivary glands and intravenously injected into a rat. At blood stages, wt/gfp and cdpk3(-)/gfp parasites were sorted using GFP as an indicator, and then injected into a new rat. Finally, cdpk3(-)/gfp parasites were cloned by limiting dilution.
Video 1. Wild-type ookinetes migrated through the Matrigel. Wild-type ookinetes were added to the surface of the Matrigel. They entered the Matrigel soon after inoculation and actively moved through the gel towards the bottom (see also Video 3). A series of confocal images was taken at 30 s interval using a LSM 5 LIVE confocal laser scanning microscope (Carl Zeiss). All serial confocal microscopic images within the gel were combined. Duration: 45 min.
Video 2.cdpk3(-) ookinetes migrated on the Matrigel but did not enter the gel. A series of confocal images including the gel surface were taken at 30 s interval and were combined (see also Video 4). Duration: 45 min.
Videos 3 and 4. Migration of wild-type (3) and cdpk3(-) (4) ookinetes was visualized in vertical (X–Z) sections. The original data are the same as in Videos 1 and 2.