Vital role for the Plasmodium actin capping protein (CP) beta-subunit in motility of malaria sporozoites

Successful malaria transmission from the mosquito vector to the mammalian host depends crucially on active sporozoite motility. Sporozoite locomotion and host cell invasion are driven by the parasite's own actin/myosin motor. A unique feature of this motor machinery is the presence of very short subpellicular actin filaments. Therefore, F-actin stabilizing proteins likely play a central role in parasite locomotion. Here, we investigated the role of the Plasmodium berghei actin capping protein (PbCP), an orthologue of the heterodimeric regulator of filament barbed end growth, by reverse genetics. Parasites containing a deletion of the CP beta-subunit developed normally during the pathogenic erythrocytic cycle. However, due to reduced ookinete motility, mutant parasites form fewer oocysts and sporozoites in the Anopheles vector. These sporozoites display a vital deficiency in forward gliding motility and fail to colonize the mosquito salivary glands, resulting in complete attenuation of life cycle progression. Together, our results show that the CP beta-subunit exerts an essential role in the insect vector before malaria transmission to the mammalian host. The vital role is restricted to fast locomotion, as displayed by Plasmodium sporozoites.


Introduction
The actin-based microfilament system drives motile processes, such as cell motility, cytokinesis and vesicle trans-port in eukaryotic cells. These processes require dynamic interconversion of pools of monomeric and filamentous actin (G-and F-actin respectively), regulated by a large number of accessory proteins (Carlier and Pantaloni, 2007). Capping protein (CP) is a heterodimeric protein that controls assembly at the barbed (fast growing) end of the actin filament in non-muscle cells (Wear and Cooper, 2004;Cooper and Sept, 2008). Its skeletal muscle variant, CapZ, links barbed ends of filaments to the Z-line. CP is a central component of actin polymerization-driven cell motility, as it restricts growth of a subset of filaments thereby allowing fast, directed polymerization from a pool of unpolymerized actin. It is enriched in the periphery of motile cells, where filament growth pushes the cell envelope forward (Amatruda and Cooper, 1992). CP is necessary for in vitro reconstituted actin-based motility (Loisel et al., 1999). In Saccharomyces cerevisiae, CP null mutants are viable but display disorganized actin patches (Amatruda and Cooper, 1992). These loss-of-function mutants are synthetic lethal with null mutants of the filament cross-linking protein fimbrin (Adams et al., 1993). CP is also a component of the dynactin complex, where it binds to the barbed end of the actin-related protein 1 (Arp1) minifilament (Schafer et al., 1994). CP gene inactivation in Dictyostelium, Drosophila and mouse, resulted in increased length and bundling of actin filaments, excessive ruffling, and loss of lamellipodia and explosive formation of filopodia respectively (Hug et al., 1995;Rogers et al., 2003;Mejillano et al., 2004). These reverse genetic approaches demonstrated a direct role of CP in actin-based motility. Genetic work in Drosophila further established an essential function of the barbed end-capping protein in early embryogenesis of a multicellular organism (Hopmann et al., 1996).
In this study, we characterized the cellular role of CPb in the unicellular eukaryote Plasmodium, the causative agent of malaria. Malaria remains the most important vector-borne infectious disease and particularly affects children in sub-Saharan Africa (Guinovart et al., 2006). During the complex life cycle in the mosquito vector and vertebrate host, the parasite follows a developmental program with alternating intracellular transformation and replication phases that lead to the formation of invasive stages. These stages, i.e. merozoites, ookinetes and sporozoites, are highly specialized for recognition of and entry into host cells, namely erythrocytes, the mosquito midgut epithelium and mammalian hepatocytes respectively. All malaria-associated symptoms and pathology originate exclusively from the pathogenic asexual red blood cell phase of the parasite life cycle. Plasmodium invasion differs fundamentally from receptor-mediated endocytosis, a hallmark of bacterial and viral host cell entry mechanisms (Sibley, 2004). Malaria and related parasites, such as Toxoplasma gondii, employ their own actin/myosin motor machinery to propel themselves into the host cell (Keeley and Soldati, 2004;Sibley, 2004). In addition, actin-based motility drives parasite locomotion and transmigration en route to the final target cell. Understanding the underlying molecular mechanisms has important implications for future malaria intervention strategies in order to target multiple stages and species simultaneously.
The actin motor machinery of the parasite features the short tailless motor MyoA (Meissner et al., 2002), tethered to the inner membrane complex by accessory proteins (Bergman et al., 2003), and very short polymers of actin that are linked to thrombospondin-related anonymous protein (TRAP)/MIC2-family invasins via aldolase (Buscaglia et al., 2003;Jewett and Sibley, 2003) and possibly other proteins. This arrangement mediates gliding on the substratum, apparently by moving F-actin-receptor complexes from the apical tip backwards along the parasite's longitudinal axis (Sibley, 2004;Schüler and Matuschewski, 2006a). The regulation of this motor machinery remains elusive. Plasmodium and related Apicomplexa encode only a fraction of the conventional microfilament regulators, with many protein families missing entirely (Baum et al., 2006;Schüler and Matuschewski, 2006b). Given the intrinsic instability of parasite actin polymers (Sahoo et al., 2005;Schmitz et al., 2005;Schüler et al., 2005) F-actin end-capping proteins are expected to be required for both orchestrated F-actin assembly and sustained filament stability. In Plasmodium genomes, two formin homology domains and two capping protein subunits can be identified, while WASP homology domains, an Arp2/3 complex and gelsolin-related proteins are apparently absent (Gordon and Sibley, 2005;Schüler and Matuschewski, 2006b).
Plasmodium and other apicomplexan parasites contain single copies of both CP subunits in their genomes. We previously identified CPb in a screen for transcripts that are upregulated during maturation of sporozoite infectivity, indicating that CPb/UIS17 (upregulated in infectious sporozoites gene 17) plays a particularly important role during malaria transmission from the mosquito to the mammalian host (Matuschewski et al., 2002). Notably, infectious sporozoites are the only Plasmodium stages that display fast (1-3 mm s -1 ) gliding motility (Matuschewski and Schüler, 2008). In this study, we could identify an essential function for CPb in sporozoite gliding locomotion and, as a consequence, colonization of Anopheles salivary glands. Our data suggest that F-actin capping by CP is vital for fast locomotion, a property of Plasmodium sporozoites that ensures efficient transmission to the vertebrate host.

Plasmodium capping protein
Protein sequences of the putative beta capping protein subunits from different Plasmodium species (PFE0880c and PB000641.00.0 for Plasmodium falciparum and P. berghei CPb respectively) share around 25% sequence identity with those of yeast, chicken and human (Fig.  S1A). These genes are among the most divergent within the family (Cooper and Sept, 2008). However, the majority of the key residues stabilizing the heterodimer (Yamashita et al., 2003) and those implicated in actin binding (Barron-Casella et al., 1995;Yamashita et al., 2003;Narita et al., 2006) are present (Fig. S1). The Plasmodium genomes also encode the corresponding putative capping protein alpha-subunit (CPa).
In order to evaluate the likelihood of these gene products being bona fide heterodimeric actin capping proteins we modelled their three-dimensional structure with SWISS-MODEL (Guex and Peitsch, 1997) using the chicken CapZ crystal structure as a template. A high fidelity model with < 1% Ramachandran outliers was readily generated (Fig. 1). In the P. berghei CPa/b homology model ( Fig. 1), conserved side chains maintain the overall architecture as well as the interactions across the subunit interface, much as previously described for CapZ (Yamashita et al., 2003).
The C-terminal 'tentacle' extension of CPb is implicated in actin binding by folding into an amphipathic a-helix that binds to a hydrophobic patch on actin (Narita et al., 2006). An additional function of the amphipathic a-helical C-termini of both CP subunits may be negative regulation of CP by binding of the tentacles through other proteins (Wear and Cooper, 2004). The modelled C-terminal tentacle of Plasmodium CPb has a similar amphipathic character (Fig. 1B). Interestingly, it is about twice as long as that of chicken CapZb, which is likely an adaptation to the structure of the parasite actin filament. In summary, our structural model is consistent with a heterodimeric Plasmodium capping protein that functions in actin polymer regulation.
To confirm that Plasmodium CP has actin capping activity we employed fluorescence microscopy of actin filaments (Fig. 2). Coexpression of the two Plasmodium capping proteins in Escherichia coli and co-purification of the a/b subunits resulted in highly purified P. berghei CP (Fig. 2B). Since Plasmodium actin forms only very short filaments (Schmitz et al., 2005;Schüler et al., 2005), we employed heterologous non-muscle b-actin to test purified proteins for capping activity in vitro. Upon addition of PbCP the length distribution of microfilaments shifted towards shorter filaments due to the barbed end capping activity of CP (Xu et al., 1999). PbCP was able to reduce the average length of polymers by half (Fig. 2). We conclude that PbCP displays bona fide F-actin capping activity in vitro.
In order to study the cellular role of Plasmodium CP for parasite propagation, we first tested transcript expression of both subunits in the three extracellular and invasive parasite stages, namely blood-stage merozoites that enter host erythrocytes, ookinetes that traverse the mosquito midgut epithelium and sporozoites that invade host hepatocytes (Fig. 3). As expected, the two genes are expressed in all motile stages similar to the motor protein MyoA. This expression pattern suggested important functions of Plasmodium CP in multiple extracellular life cycle stages.

Generation of cpb(-) parasites
We focused on the in vivo roles of CPb/UIS17 for parasite development in the two hosts. Reverse genetics in the rodent malaria model parasite P. berghei is particularly suited to study gene functions during Plasmodium life cycle progression. Initially, we targeted the endogenous PbCPb locus with an integration vector that disrupts the corresponding open reading frame. Successful stable insertion of the integration plasmid suggested that CPb is dispensable for propagation of the pathogenic blood stages (data not shown). A. Superposition of a structural model of the Plasmodium berghei CP heterodimer onto the crystal structure of the chicken CapZ heterodimer. The alpha-subunits of P. berghei and chicken CP are shown in light and dark green, and the beta-subunits of P. berghei and chicken CP are shown in yellow and gold respectively. The positions of the N-and C-termini are indicated. B. Amphipathic characters of the C-terminal tentacle extensions of parasite (Pb) and chicken (Gg) CPb. Key side-chains contributing to the amphipathic character are indicated.
To confirm this unexpected finding and to generate a genetically stable CPb loss-of-function parasite line, we constructed a targeting vector that contained the 5′ and 3′ CPb flanking regions separated by the mutant T. gondii dhfr/ts gene cassette as a positive selection marker (Fig. 4A). Upon transfection the target gene locus is replaced with the selection marker via a double homologous recombination event. We obtained a parental population that contained mixed parasites with the disrupted cpb(-) and the wild-type (WT) CPb locus (not shown). Next, we cloned parasite lines in vivo by serial dilution and intravenous injection into 15 NMRI mice as recipient animals. We obtained six clones: three WT, one mixed and two cpb(-) lines. Genotyping of the cpb(-) clones by Southern blot analysis (Fig. 4B) and replacement-specific PCR analysis (Fig. 4C) showed absence of the WT-specific signal, and successful disruption of the CPb locus. RT-PCR analysis with poly (A) + RNA from WT and mutant mixed blood stages confirmed the absence of CPb transcripts in the cpb(-) parasite lines (Fig. 4D). Successful generation of CPb loss-of-function parasite lines thus confirmed that this gene is not vital for propagation of the pathogenic asexual blood stages.
To test whether mutant parasites display a growth defect during blood-stage multiplication we performed an in vivo growth assay by intravenous injection of 1000 parasite blood stages into mice, followed by daily parasitaemia counts (Fig. 4E). Proliferation of cpb(-) parasites was indistinguishable from WT parasites suggesting that CPb is not essential in the pathogenic blood stages.

cpb(-) parasites are impaired in ookinete invasion of the mosquito midgut
We then tested the mutant parasites for transmission to the Anopheles vector. A fraction of asexual blood stages eventually enter sexual development giving rise to female and male gametes that fuse in the mosquito midgut to form a motile zygote, the ookinete (Kooij and  Presence of CPb, CPa and the unconventional class XIV myosin MyoA was tested for transcript expression in three invasive Plasmodium stages: merozoites (mer) that invade erthrocytes, ookinetes (ook) that penetrate the mosquito midgut and sporozoites (spz) that invade the mammalian liver. CPa contains eight introns that distinguish amplification from gDNA and cDNA. No amplification is detectable in control reactions lacking reverse transcriptase (-RT).

Fig. 4. Targeted gene disruption of P. berghei CPb.
A. Replacement strategy to generate the cpb(-) parasite. The wild-type (WT) CPb genomic locus is targeted with a SacII/KpnI-linearized replacement plasmid containing the 5′ and 3′ untranslated regions of the CPb open reading frame (ORF) and the Toxoplasma gondii dhfr/ts-positive selectable marker. Upon a double-cross-over event the CPb ORF is replaced by the selection marker. Replacement-specific Southern fragments and test primer combinations are indicated by arrows, and expected fragments are shown as lines. B, BglII; Sp, SpeI; S, SacII; K, KpnI. B. Diagnostic Southern blot analysis of two clones and WT parasites. The fragments recognized by the 5′ probe in the BglII digest are 9.96 and 12.0 kb for the WT and cpb(-) parasites respectively. Similarly, in the SpeI digest, the probe hybridizes to a 5.0 kb band in WT and a 3.98 kb band in cpb(-) parasites. C. Replacement-specific PCR analysis. Confirmation of the predicted gene targeting is achieved by specific primer combinations that can only amplify a signal from the recombinant locus. Black and grey arrows in (A) indicate primers that hybridize to regions in the plasmid backbone and within and outside the CPb ORF respectively. A WT-specific PCR reaction confirms the absence of residual wild-type parasites in the two clonal cpb(-) populations. D. Depletion of CPb transcripts in cpb(-) parasites. cDNA from wild-type and cpb(-) clone 1 merozoites was amplified at 35 PCR cycles. Note the absence of a CPb-specific signal compared with a transcript control (MyoA). E. cpb(-) parasites develop normally in the mammalian host. Asexual blood-stage development was determined by i.v. injection of 1000 infected erythrocytes. Parasitaemia of recipient animals (n = 5) was determined by daily quantification of Giemsa-stained blood smears. White bars, WT; grey bars, cpb (-). Shown are the mean values (Ϯ standard deviation). Matuschewski, 2007). Quantification of sexual stages in Giemsa-stained blood smears of infected mice showed that cpb(-) parasites are not affected in commitment to form mature gametes (Table S1). Similarly, we could not observe a loss-of-function phenotype in in vitro formation of ookinetes, the stage that penetrates the mosquito midgut epithelium (Table S1). Together, these findings suggest that transmission from the warm-blooded host to the invertebrate vector does not rely on functional CPb.
When we counted oocyst numbers of WT and mutant parasites at day 10 after mosquito infection we could detect a significant phenotype that was, however, not complete ( Fig. 5 and Fig. S2). The cpb(-) parasites produce substantially fewer oocysts as compared with the isogenic WT. This finding was confirmed by membrane feeding of equal numbers of in vitro cultured ookinetes to Anopheles mosquitoes (Fig. 5), suggesting that the observed reduced ookinete infectivity does not depend on the presence of the peritrophic membrane that is formed after blood meal ingestion. Most importantly, mutant ookinetes that successfully penetrated the midgut epithelium are capable of producing mature midgutassociated oocysts (Fig. 5). This observation prompted us to test the ability of ookinetes to glide in vitro ( Fig. 6 and Movies S1 and S2). In agreement with the impaired oocyst production rate we observed a reduction in ookinete motility. Quantitative analysis of ookinete velocities revealed an average speed of 3.7 (Ϯ 1.7) mm min -1 and 6.5 (Ϯ 2.0) mm min -1 for cpb(-) and WT ookinetes respectively (Fig. 6B). Nevertheless, mutant ookinetes transiently displayed fast locomotion (Fig. 6C). Together, cpb(-) ookinetes displayed productive motility and were  apparently capable of reaching the midgut epithelium (Fig. 5). From these findings we conclude that CPb plays an auxiliary role in ookinete penetration of the mosquito midgut epithelium.

cpb(-) sporozoites are non-motile and fail to invade salivary glands
To confirm the capacity of cpb(-) parasites to form sporozoites we counted midgut-associated sporozoites at various time points after infection ( Table 1). As expected, mutant parasites were able to produce substantial numbers of midgut sporozoites. In fact, the average number of midgut sporozoites was only moderately diminished as compared with the observed reduction in oocyst numbers (Fig. 5), consistent with a recent observation that Plasmodium development in the mosquito vector exhibits a strong density dependence, i.e. few oocysts produce disproportionally high numbers of sporozoites .
When we followed sporozoite maturation we observed a dramatic loss-of-function phenotype (Table 1); cpb(-) sporozoites lost their ability to invade salivary glands, the final target organ in the mosquito vector. Consistent with this finding, none of highly susceptible C57BL/6 mice or young Sprague/Dawley rats developed a malaria infection when cpb(-)-infected Anopheles mosquitoes were used to infect naïve animals by natural mosquito bites ( Table 2). This finding indicated an essential role of CPb during sporozoite maturation, a prerequisite to complete the life cycle and infect the mammalian host (Matuschewski, 2006).
To test whether ablation of salivary gland invasion was the exclusive cause for the observed interruption of transmission, we isolated large numbers of midgut-associated sporozoites and injected them intravenously into susceptible mice (Table 2). Again, all infected mice stayed free of malaria. Next, we tested whether attenuation of Plasmodium life cycle progression in the cpb(-) parasites can be reversed by inheritance of a WT copy during sporozoite formation. Previous work on the P. berghei LCCL/lectin adhesive-like proteins (LAPs) has established that heterozygous oocysts obtained by crossing mutant and WT parasites in vivo can rescue loss-of-function mutants (Raine et al., 2007). For the present study, we crossed cpb(-) and WT parasites and genotyped the mixed parasites in comparison with clonal parasites before and after mosquito transmission. As predicted, in mixed inoculations the cpb(-) genotype was recovered after life cycle completion (Fig. S3). These findings led us to conclude that the essential function of Plasmodium CPb is restricted to the insect vector stages, and this deficiency  can be rescued by transient complementation of one WT copy during oocyst development, where a heterozygous cell undergoes multiple rounds of replication prior to sporozoite budding. We finally studied sporozoite gliding locomotion. For this analysis, we isolated haemocoel sporozoites, the stage that is the most advanced in the mutant parasite lines. In marked contrast to WT sporozoites that perform continuous fast gliding ( Fig. 7 and Movie S3), cpb(-) sporozoites never displayed productive locomotion ( Fig. 7 and Movie S4). Notably, bending and flexing, a microtubule-dependent form of non-productive motility (Vanderberg, 1975), was frequently observed, corroborating the viability of mutant sporozoites and supporting a distinct defect in forward locomotion. In agreement with defective sporozoite locomotion, we detected substantial numbers of midgut-associated sporozoites throughout the mosquito lifespan (Table 1), suggesting that active gliding motility is important for efficient sporozoite egress out of oocysts.
In conclusion, the Plasmodium CPb subunit exerts a vital function during life cycle progression of the malaria parasite in the insect vector. Loss of CPb function results in non-motile sporozoites that fail to colonize the mosquito salivary glands and, hence, cannot be transmitted to the vertebrate host.

Discussion
The most important finding of our reverse genetics analysis of Plasmodium CPb is a vital role for life cycle progression of the unicellular obligate intracellular malaria parasite. Our systematic phenotypic characterization of P. berghei cpb(-) parasites revealed important cellular functions exclusively in locomotion of extracellular parasite stages. We establish that CPb is essential for sporozoite gliding locomotion, a function that is gradually acquired during sporozoite maturation and required for subsequent transmission to the mammalian host (Matuschewski, 2006). This stage-specific vital role of CPb correlates with fast motility, a feature observed only in mature, salivary gland-associated sporozoites. In analogy, we predict that the T. gondii CPb protein performs essential functions for motility of tachyzoites, Fig. 7. CPb is essential for sporozoite gliding locomotion. Quantification of motility patterns in midgut-and haemocoel-associated sporozoites from WT and the cpb(-) mutant (clone 1) (right). Shown is the percentage of non-motile, i.e. detached (white) and attached (light grey), sporozoites and sporozoites that display non-productive (dark grey) and productive (black) motility. In contrast to WT sporozoites that mature upon egress to the mosquito haemocoel cpb(-) parasites remain non-gliding. Track projections (projections) and phase micrographs at time point 0 s (phase) of representative motility patterns of WT haemocoel sporozoites are shown on the left. On coated glass slides a proportion of WT sporozoites glide in circles at an average speed of 1-3 mm s -1 (represented by bottom panels). In marked contrast, cpb(-) sporozoites display only non-productive motility patterns, such as bending, flexing and pendulum movement (represented by upper panels). Scale bars, 10 mm. the fast-replicating stage that infects all nucleated cells.
Unexpectedly, CPb is dispensable for asexual erythrocytic parasite growth, the life cycle phase that causes all malaria-related morbidity and mortality. Proliferation of cpb(-) parasites was indistinguishable from WT parasites. Two actin-dependent processes appear to be central to the intra-erythrocytic cycle, i.e. endocytic haemoglobin uptake (Smythe et al., 2008) and erythrocyte invasion (Miller et al., 1979). Since a defect in haemoglobin uptake, the major amino acid source for the parasite, likely translates into delayed and/or decreased merozoite formation, we postulate that an actin capping activity, if any, is either redundant or mediated by proteins other than CPb. Inhibitor studies established a requirement for parasite actin in merozoite invasion (Miller et al., 1979). In support of a non-vital role of CPb in actin capping during this process PbCPb transcripts are readily detectable in free merozoites. In a previous biochemical study three major F-actin binding proteins, i.e. HSP70 and two unknown proteins of 32 and 34 kDa, were isolated from Plasmodium knowlesi merozoites (Tardieux et al., 1998). The F-actin binding properties and the respective molecular sizes of the yet unidentified proteins are consistent with the heterodimeric CP protein. In addition, the forminlike protein 1 has been localized in the vicinity of the parasite restriction that forms upon cell entry, termed moving junction (Baum et al., 2008). Formin-dependent incorporation of G-actin monomers into the growing filament resembles a transient barbed end capping function that may partially compensate for the lack of CPb. However, it is important to note that merozoite invasion into erythrocytes is typically completed within 30 s. Therefore, a reduced speed does not necessarily translate into a detectable alteration of parasite propagation, since the intra-erythrocytic life cycle phase of P. berghei takes 24 h. Hence, either cpb(-) parasites may have a partial invasion defect that remains undetected in vivo or CPb is dispensable for the motility of merozoites, which appear to lack the capacity to glide on substrates entirely.
Direct proof for a role of CPb in cell motility comes from the intermediate phenotype observed in cpb(-) ookinetes that display an overall reduced motility, while all other functions, including zygote formation and viability as well as subsequent sporogony, remain unaffected. Membrane feedings of in vitro cultured ookinetes excluded a defect in penetration of the peritrophic membrane, which is formed exclusively after a blood meal and requires secretory parasite proteins, such as chitinase, to be permissive for ookinetes.
The observed intrinsic instability of Plasmodium actin (Schmitz et al., 2005;Schüler et al., 2005) is expected to result in rapid microfilament depolymerization. Therefore, the cellular role of CP as an F-actin stabilizer may be particularly important in this model eukaryote. Barbed end capping by CP is essential for fast actin/myosindependent gliding locomotion, as displayed by mature infectious Plasmodium sporozoites. Incidentally, this forward motility displaying an average speed of 1-3 mm s -1 is among the fastest gliding locomotions on substrates for any eukaryotic cell (Vanderberg, 1975). We propose that barbed end capping by capping protein is vital for rapid microfilament turnover to achieve high speed actin-dependent locomotion in vivo.

Experimental animals
Animals were from Charles River Laboratories. All animal work was conducted in accordance with European regulations and approved by the state authorities (Regierungspräsidium Karlsruhe).

Parasite transfection and genotypic analysis
For replacement of PbCPb we employed primers PbCPb_forI (5′-ATCCCCGCGGAGTACATGCAATATATACATATATATTCA ATGC-3′; SacII site is underlined) and PbCPb_revII (5′-ATAAGAATGCGGCCGCAATTTAGTTTTTTTATATGGTTATT TATTTATTACAG-3′; NotI site is underlined) for amplification of the 5′ flanking region, and PbCPb_revIII (5′-CC CCAAGCTTGAATGCGATTTTAGGGGCCAATACAATTAGC-3′; HindIII site is underlined) and PbCPb_revIV (5′-CGGGGTACCCGATTTTTTTTATTAATTCATCAAATTTTCCC-3′; KpnI site is underlined) for the 3′ flanking region, respectively, using P. berghei genomic DNA as template. Cloning into the P. berghei transfection vector (Thathy and Ménard, 2002) resulted in the plasmid pPbCPbREP. The targeting plasmid was linearized with SacII and KpnI, and parasite transfection, positive selection and parasite cloning were performed as described previously (Janse et al., 2006). Standard Southern blot analysis was performed with a commercial kit (DIG High Prime Labelling and Detection Starter kit II, Roche). For probe amplification we utilized primers PbCPb_forI and PbCPb_revII. We obtained two independent cpb(-) clonal parasite populations that were phenotypically identical. Detailed analysis was performed with one representative clone.

Transcript detection
For RT-PCR analysis, we isolated poly (A + ) RNA using oligo dT-columns (Invitrogen). For cDNA synthesis and amplification, we performed a two-step PCR using oligo dT primers (Ambion) and subsequent standard PCR reactions, using gene-specific primers.

Analysis of parasite development
Anopheles stephensi mosquito rearing and maintenance was carried out under a 14 h light/10 h dark cycle, 75% humidity and at 28°C or 20°C respectively. Blood-stage development was analysed in vivo in asynchronous infections using Naval Medical Research Institute (NMRI) mice. Gametocyte differentiation and exflagellation of microgametes were detected in mice before ookinete culture or mosquito feedings respectively.
Ookinete culture was conducted in RPMI 1640 medium with L-glutamine and 25 mM HEPES (Gibco) supplemented with 100 mM sodium bicarbonate; 100 mM hypoxanthine; 10% FCS; 50 mM xanthurenic acid; and 125 U ml -1 penicillin/ streptomycin. For ookinete culture, we treated Theiler's Original (TO) mice with phenylhydrazine (1.2 mg per mouse) 24 h before infection. Infected blood was collected by cardiac puncture 4 days after infection and added to 10 volumes of culture medium. After incubation for 24 h at 20°C, erythrocytes were lysed for 20 min in ice-cold ammonium chloride solution (170 mM), and ookinetes were washed and re-suspended in HBSS. For artificial membrane feeding of ookinetes, cultured ookinetes were mixed with 1 ml of blood from a naïve mouse and applied to a temperature-regulated glass feeder.
Ookinete motility was analysed by time-lapse microscopy. Cultured ookinetes were purified, mixed with mos20 cells and imaged on a Zeiss Axiovert M200 microscope (Carl Zeiss, Göttingen, Germany). Images were captured using a Cool-Snap HQ camera (Photometrics, Tucson, AZ, USA) and MetaMorph imaging software (Molecular Devices, Downingtown, PA, USA). Processing of all images was conducted utilizing the program ImageJ.
Sporozoites were dissected and analysed as described previously (Vanderberg, 1975). For determination of sporozoite infectivity, infected mosquitoes were dissected at days 14-30 after feeding. Sporozoites were liberated and injected intravenously at the numbers indicated into young Sprague/ Dawley rats or C57BL/6 mice respectively. Patency was determined by daily examination of Giemsa-stained blood smears.

Homology modelling
The homology model of the P. berghei CPa/b heterodimer was constructed based on sequence alignments with the respective chicken CapZ subunit sequences using SWISS-MODEL (Guex and Peitsch, 1997). The a-subunits are 23.8% identical and 42.3% similar, and the beta-subunits are 30.7% identical and 37.2% similar on the amino acid level. The model was superimposed onto the crystal structure of the chicken CapZ heterodimer (PDB entry 1izn). The betasubunit tailpieces shown in Fig. 1B represent the sequences 250-PDNQKYKQLQRELSQVLTQRQI-271 (GgCPb) from crystal structure 1izn and SKGNIQNELKSKLKKK (PbCPb) from the homology model based on that structure.

Recombinant protein expression and purification
Recombinant active P. berghei capping protein heterodimer was produced by coexpression of both alpha-and betasubunits from the same expression plasmid as described previously for chicken CP (Soeno et al., 1998), employing the pET-Duet-1 vector system (Novagen). Expression of the recombinant capping protein subunits, a N-terminally hexahistidine-tagged PbCPa and C-terminally S-tagged PbCPb, were induced using 0.5 mM IPTG. Recombinant capping proteins were purified from lysed bacterial pellets using HisTrap HP columns (GE Healthcare, Piscataway, NJ, USA) equilibrated with 20 mM Tris pH 8.0, 20 mM imidazole and 300 mM NaCl, 10% glycerol. After washing, proteins were eluted in 1 ml fractions with 20 mM Tris pH 8, 400 mM imidazole, 300 mM NaCl, 10% glycerol. Protein identity was confirmed by TOF-MS analysis. Capping protein concentrations were determined using the BCA Protein Assay Kit (Pierce) and stored as aliquots at -80°C. Non-muscle b-actin was purchased from Cytoskeleton (Frankfurt, Germany) and gelsolin from Sigma-Aldrich (Taufkirchen, Germany).

F-actin assays
As described previously (Xu et al., 1999), samples of actin (5 mM) were induced to polymerize by addition of 1 mM MgCl2 and 0.15 M KCl, and incubated at room temperature for 2-3 h. Actin polymers were supplemented with 100 nM rhodamine-phalloidin (Invitrogen) and incubated for 15 min at room temperature on coverslips in the presence of purified recombinant P. berghei capping protein. Samples were mounted in Vectashield (Vector Laboratories, Burlingame, CA, USA) and imaged using a 100¥ Fluoroplan oil immersion lens on a Zeiss Axiovert M200 microscope (Carl Zeiss, Göttingen, Germany). Images were captured using a CoolSnap HQ camera (Photometrics, Tucson, AZ, USA) and Meta-Morph imaging software (Molecular Devices, Downingtown, PA, USA).
Actin polymer length measurements were carried out using ImageJ software. Polymers were sorted into 0.5 mm bins and their length distributions plotted. Polymers shorter than 0.5 mm and longer than 20 mm were omitted from the analysis.