SEARCH

SEARCH BY CITATION

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Malaria parasites have two actin isoforms, ubiquitous actin1 and specialized actin2. Actin2 is essential for late male gametogenesis, prior to egress from the host erythrocyte. Here, we examined whether the two actins fulfil overlapping functions in Plasmodium berghei. Replacement of actin2 with actin1 resulted in partial complementation of the defects in male gametogenesis and, thus, viable ookinetes were formed, able to invade the midgut epithelium and develop into oocysts. However, these remained small and their DNA was undetectable at day 8 after infection. As a consequence sporogony did not occur, resulting in a complete block of parasite transmission. Furthermore, we show that expression of actin2 is tightly controlled in female stages. The actin2 transcript is translationally repressed in female gametocytes, but translated in female gametes. The protein persists until mature ookinetes; this expression is strictly dependent on the maternally derived expression. Genetic crosses revealed that actin2 functions at an early stage of ookinete formation and that parasites lacking actin2 are unable to undergo sporogony in the mosquito midgut. Our results provide insights into the specialized role of actin2 in Plasmodium development in the mosquito and suggest that the two actin isoforms have distinct biological functions.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Transmission of malarial parasites to the mosquito and the ensuing differentiation and population expansion phases in the vector represent a very complex developmental programme in the Plasmodium life cycle, which requires the initiation of sexual stage development in the blood of the vertebrate host. Although the duration of gametocytogenesis within the red blood cell differs between Plasmodium species, the cell cycle of mature gametocytes is suspended until taken up by a mosquito host with a blood meal (Baker, 2010). Immediately after uptake, the male and female gametocytes are activated and develop into fertile gametes, which fuse to form a zygote about 1 h later. An unusual aspect is that meiosis immediately follows zygote formation, resulting in the development of a cell containing four haploid genomes. The round zygote is then remodelled: a microtubule organizing centre (MTOC) is laid down close to the cell membrane and from this, microtubules are nucleated and an elongation of the cell is initiated through their extension. This form of the cell is called a retort. Gradual elongation results in the formation of the polar ookinete within the next ∼ 15 h (Sinden, 1999). The ookinete is a motile cell, using gliding motility to traverse midgut epithelial cells (Vlachou et al., 2004). Once it reaches the basal side of the epithelial cell, extensive reorganization of the parasite follows, resulting in the formation of the round oocyst, which is surrounded by a capsule. In this cell, a series of synchronous mitotic divisions (∼ 13) in a syncytium lead to the formation of several thousand nuclei in each oocyst (Sinden, 1999). The apicoplast and mitochondrion are also replicated (Stanway et al., 2009) and massive protein synthesis takes place followed by the formation of secretory organelles. During the final phases, about 10–12 days after the blood meal, budding produces the individual sporozoites. Breakdown of the oocyst capsule finally results in the release of the sporozoites into the haemocoel (Aly and Matuschewski, 2005).

As all Plasmodium parasites, P. berghei expresses two conventional actin isoforms, the ubiquitous actin1 (PBANKA_145930) and a specialized actin2 (PBANKA_103010) (Wesseling et al., 1988; Gardner et al., 2002; Schuler and Matuschewski, 2006). [The gene PBANKA_103010, was previously referred to as actin II (Deligianni et al., 2011) and actin 2 (Kooij et al., 2012); it will be named actin2 in this article. Ubiquitous actin, PBANKA_145930, called actin I, will be named actin1]. Both actin isoforms have the ability to form actin filaments in vitro (Schmitz et al., 2005; 2010; Schuler et al., 2005; Skillman et al., 2011). While the actin1 gene is essential, actin2 can be deleted in P. berghei (Deligianni et al., 2011). Our previous detailed genetic and cell biological analysis revealed that actin2 has a critical role in male gametogenesis. Mutant male gametocytes were severely blocked in egress, and, while axonemes were formed, these were not activated. As a result exflagellation was reduced to almost undetectable levels, and only a very small number of ookinetes were occasionally formed. However, not a single oocyst was ever detected after feeding the mutant to mosquitoes (Deligianni et al., 2011). While this study employing classical gene deletion revealed critical roles for actin2 in male gametogenesis, additional important functions later in the Plasmodium life cycle, for example during the stages in the mosquito vector, remain unknown.

Here, we investigated whether open reading frame (ORF) swapping with that of ubiquitous actin1 could restore the function of actin2. Using this isoform complementation approach we could show that resulting parasites were able to form ookinetes, albeit at reduced numbers. Surprisingly, the resulting oocysts were completely blocked in sporogony. To study this second function of actin2 we employed genetic crosses. We demonstrate that in addition to its role in male gametogenesis actin2 has an essential function during early ookinete formation, and lack of actin2 ultimately manifests as a complete block in sporogony.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Replacement of actin2 with actin1 leads to a reduction of ookinete formation

A longstanding question is whether the two Plasmodium actin isoforms have related molecular functions or whether they play distinct roles in the cell. Because of the critical function of actin1 for the growth of asexual blood stages, the phase of the life cycle where transfection is performed, we focused our study on modifying the actin2 gene. Our strategy aimed at replacing the actin2 ORF with that of actin1. For this we used a recipient line where the complete ORF of actin2 had been deleted and the mutant strain had been recycled to remove the resistance cassette (Kooij et al., 2012) (Fig. S1). Similar to our original actII(−) mutant (Deligianni et al., 2011) this strain was incapable of exflagellation.

The recipient line was transfected with a construct containing the actin1 ORF under the control of the actin2 5′- and 3′-flanking regions (FR) resulting in the act2rep mutant. Because the recipient actin2::mCherry mutant already contained several kbp of foreign DNA, we also verified that we could complement the actin2 deletion by an analogous construct, containing the actin2 ORF under its own 5′ and 3′FR, thus restoring the correct expression of the gene in the act2com mutant (Fig. 1A, Fig. S1). For both constructs recombinant parasites harbouring the predicted genetic modifications were obtained and cloned using limiting dilution (Fig. S1).

figure

Figure 1. Expression of actin1 cannot replace the function of actin2 in oocyst development.

A. Schematic representation of the complementation (act2com) and replacement constructs (act2rep). The recipient strain where the ORF of the actin2 gene had been deleted contained the actin2 5′FR directly upstream of the coding region of mCherry, a GFP cassette followed by actin2 3′FR. This strain was transfected with two identical constructs except for the ORF, which was actin2 (blue arrow) and actin1 (orange arrow) in act2com and act2rep respectively. The ORF was inserted between 2.7 kb of the actin2 5′FR and 728 bp of the actin2 3′FR and the vector in addition contained the human DHFR encoding gene. A single cross-over integration in the 5′FR of the deleted actin2 gene resulted in integration of the transgenes in the actin2 locus, followed by the hDHFR cassette and the duplicated actin2 5′FR. For full description of the strategy and the verification of the mutants see Fig. S1B.

B. Expression of actin1 under actin2 regulatory elements results in significantly less efficient ookinete formation compared to WT and the act2com strain. ***P <0.0001 Student's t-test.

C. Expression of actin1 under actin2 regulatory elements leads to a complete block in oocyst formation. Experiments 1 and 2 were performed by feeding A. gambiae mosquitoes on infected mice (darker blue and green). The oocysts were counted at day 12. In experiment 3 ookinetes (3 × 105 of each strain) from in vitro cultures were fed using a membrane feeder (lighter blue and green) and oocysts scored at day 11. No oocysts were detected in the act2rep fed mosquitoes. Act2com blue circles, act2 rep green squares. ***P <0.0001 Kruskal–Wallis non-parametric test.

D. RT-PCR of cDNA originating from act2com and act2rep mixed blood stages. The A2UF primer binding to the actin2 5′FR was used in both samples while the reverse primers were specific for each ORF (green arrows in A). The act2rep-specific fragment was amplified slightly more efficiently (1.6-fold) than the act2com fragment, determined by measuring the intensity of the bands and normalizing the two cDNA samples with primers specific for Pbmap2 (PBANKA_093370). The primers were normalized using WT and act2rep gDNA respectively.

Download figure to PowerPoint

We then investigated the fertility of the two lines. Exflagellation of the act2rep parasites was roughly 10-fold reduced compared to the act2 com strain (J. Vahakoski et al., unpublished). Ookinete conversion of the act2com strain was restored to WT numbers (Fig. 1B), confirming that the complementation was successful. The conversion of the act2rep strain, on the other hand, was reduced more than sixfold, but ookinetes were always detected, and we were able to obtain moderate numbers of ookinetes from act2rep. This is in contrast to the actin2 deletion mutants where only occasionally very few ookinetes were detected (Deligianni et al., 2011). These findings show that actin1 can partially restore the essential functions of actin2 during male exflagellation when its expression is controlled by the actin2 regulatory elements. The fact that the act2rep mutant formed ookinetes allowed us to investigate the role of actin2 in the mosquito midgut stages.

Replacement of actin2 with actin1 blocks transmission through the mosquito

To determine whether the ookinetes were able to develop into oocysts we fed mosquitoes on act2com- and act2rep-infected mice. In two independent experiments oocysts were detected in mosquitoes fed with act2com, while we did not detect a single oocyst in the act2rep-infected mosquitoes (Fig. 1C). We confirmed this result by feeding mosquitoes on equal numbers of in vitro cultured ookinetes, to exclude the possibility that the results were due to substantially fewer ookinetes having formed in the act2rep strain. Again, oocysts were only found in the act2com-infected midguts and no oocysts were detected in the mosquitoes fed with act2rep ookinetes (Fig. 1C). Taken together, these data suggested that the act2rep ookinetes were not able to complete sporogony.

One interpretation of these results is that actin1 is not produced in amounts similar to actin2. It is not possible to directly determine the amount of actin1 protein produced from the actin2 locus, due to the fact that it is also produced from the WT actin1 gene and is present in the same cellular compartment (Deligianni et al., 2011). We therefore performed RT-PCR (reverse transcription PCR) experiments to determine whether both transgenes are equally transcribed into mRNA (Fig. 1D). cDNA was prepared from mixed blood stages of the two strains, and specific primers were used to amplify fragments corresponding to each of the two transgenes. The results indicate that transcript levels of the actin2 and actin1 transgenes are similar in the act2com and act2rep strains respectively. Thus gene expression regulated by the actin2 cis-elements is, at least in this case, independent of the identity of the ORF.

Act2rep oocysts are formed but do not undergo sporogony

The failure to detect act2rep oocysts in these experiments could be due either to a failure of the ookinetes to traverse the midgut epithelium or to arrested oocyst development. It is known that motility of the ookinete is necessary for traversal of the epithelium (see for example Dessens et al., 1999; Vlachou et al., 2004; Siden-Kiamos et al., 2006). To exclude that the act2rep ookinetes were affected in their motility, we used a motility assay employing matrigel (Moon et al., 2009). Our results show that there is no detectable difference in the motility of the act2rep ookinetes compared to WT (Fig. 2A, Movies S1 and S2). The speed was measured and found to be 5.7 μm min−1 (mean of nine ookinetes) which is similar to that of WT ookinetes (6 μm min−1). In order to recognize even small oocysts, we labelled the oocysts with an antibody against the oocyst capsule protein Cap380 (Srinivasan et al., 2008) and stained the nuclei for DNA. Mosquitoes were fed on mice infected with the two strains. Oocysts were detected as early as day 3 after feeding in the act2com but not in the act2rep-fed mosquitoes, a result possibly due to the very small number of oocysts developing in this strain (Fig. 2B,a). At day 8, very few oocysts (7 in 11 midguts) were detected, which were all significantly smaller than the developing act2com oocysts and which contained no detectable DNA (Fig. 2B,b,d). Eleven days after blood feeding, a few act2rep oocysts (5 in 20 midguts) were detected; the difference in size of the oocysts of the two strains was even more pronounced (Fig. 2B,c,e), which was verified by measurements of the area of the oocyst (Fig. 2C). We confirmed the apparent loss of the genomic DNA in the act2rep oocysts by performing PCR of infected midguts using primers for the gapdh gene, which specifically amplifies parasite DNA. This showed that the act2rep DNA was lost while, as expected, act2com DNA was present at all three time points (Fig. 2D).

figure

Figure 2. The act2rep strain forms abnormal oocysts.

A. Gliding of WT and act2rep ookinetes in Matrigel™ is illustrated with representative frames from time-lapse videos (Movies S1 and S2). The black asterisk marks the apical end of the ookinete. Scale bar, 20 μm.

B. act2com (a,b,c) and act2rep (d,e) oocysts labelled with the Cap380 (green) antibody and stained for DNA (blue). At 3 days (a) only act2com oocysts were detected. At 8 (b,d) and 11 days (c,e) the act2 rep oocysts are small and their DNA hardly visible comparable to the developed act2com oocysts. Act2com oocysts have developed into big oocysts at day 11 and the DNA has replicated. Scale bar, 10 μm.

C. The area of the oocysts at day 3, 8 and 11. Act2 com: red circles, act2rep: grey squares. The act2rep oocysts do not grow in size beyond 3 days. **P <0.005, ***P <0.001, Student's t-test.

D. The act2rep is lost during oocyst development. Genotyping of mosquitoes fed with the two strains separately. PCR was performed on genomic DNA isolated from dissected mosquito midguts using primers specific for the P. bergheigapdh gene and the ribosomal S7 gene of A. gambiae as control. Samples correspond to mosquitoes dissected 3, 8 and 11 days after blood feeding.

Download figure to PowerPoint

We also fed naïve mice on mosquitoes infected with these two strains. As expected, no infection was established from the act2rep-infected mosquitoes, while the act2com-infected mosquitoes repeatedly produced infections in all mice with an average pre-patent period of 5 days (data not shown).

Taken together, these results suggest that the functions of actin1 and actin2 in the mosquito vector are distinct. Furthermore, it also suggests that actin2 has two consecutive essential functions. In addition to the previously described role in male gametogenesis (Deligianni et al., 2011), this actin isoform is also critical for sporogony and, importantly, this second function is independent of the role in gametogenesis.

Actin2 is expressed during ookinete development

We previously reported that the actin2 gene was specifically transcribed in gametocytes of both sexes and not in asexual forms during the blood stages, and that the protein is present only in male gametocytes (Deligianni et al., 2011). Our new experiments indicated that actin2 was expressed during additional stages of the parasite life cycle. To investigate this we performed RT-PCR analysis on mature ookinetes, and mosquito midguts containing oocysts (Fig. 3A). The actin2 transcript was detected in the ookinete sample, but not in oocysts. Actin1, on the other hand, was constitutively transcribed during oocyst maturation.

figure

Figure 3. Expression of actin2 in mosquito midgut stages.

A. Reverse transcription PCR analysis of expression of actin2 in ookinetes (ook) and oocyst containing midguts at 3, 8 and 12 days after blood feeding. Primers specific for ubiquitous actin1 was used for normalization.

B. Actin2 is localized in the cytoplasm of mature ookinetes. (a–c) IFA of a WT ookinete using the actin2 antiserum (a, b, green). (d–f) Labelling of an ookinete expressing the GFP–actin2 fusion protein with anti-GFP antibody (d,e, green) and the 13.1 antibody recognizing Pbs21 (d, red). DNA was stained with TO-PRO (a,c,d,f, blue). Images are projections from confocal stacks. Scale bar, 5 μm.

C. GFP–actin2 expression during zygote to ookinete development. Left panel: live imaging of GFP and corresponding bright-field images (BF). Brightness of signal has been adjusted for clarity and is not comparable between images. Right panel: Developing ookinetes were fixed and labelled with antibodies recognizing GFP and Pbs21. The time of each sample is indicated on the left of each panel (h after seeding of the culture). Pictures were taken using an epifluorescence microscope.

Download figure to PowerPoint

We next performed immunofluorescence assays (IFA) of ookinetes to determine the localization of actin2. We made use of a strain expressing a GFP–actin2 fusion protein encoded by an episomal construct as previously described (Deligianni et al., 2011). The episome also carries a gene conferring resistance to the antifolate pyrimethamine, as a positive selection for its presence in the parasites. Ookinetes of this strain developed normally from parasites cultivated in the presence of a low concentration of the antifolate and expressed the GFP–actin2 fusion protein, visualized by labelling with an antibody against GFP. In addition, an antiserum directed against actin2 was also employed. In both cases the protein was detected in the cytoplasm, in a diffuse pattern with no obvious discrete localization, and it was excluded from the nucleus (Fig. 3B).

We next followed expression of the GFP–actin2 fusion protein during development from zygotes to ookinetes. Samples from an ookinete culture were removed at different time points after seeding. At the early time points (3, 6, 9 h) live cells were imaged using an epifluorescence microscope. At later time points (12, 15, 18, 21 h) the samples were fixed and labelled with antibodies directed against GFP and the surface protein Pbs21 (Fig. 3C). The protein was detected already at 3 h after the seeding of the culture. In the retort forms (12 and 15 h after seeding the culture) the fusion protein was mainly present in the round residual cell, while comparatively less was present in the extension of the cell. At later stages the label was detected in the cytoplasm of the cell, with no discrete localization.

Attempts to detect the protein in oocysts using either the GFP fusion strain or the antiserum were unsuccessful (data not shown). In conclusion, actin2 is expressed during the first critical phase of mosquito colonization, i.e. maturation of Plasmodium ookinetes.

Parasites lacking actin2 are lost during oocyst development

We previously characterized a knockout mutant of actin2, named actII(−), which was blocked in male gametogenesis and thus formed negligible numbers of ookinetes and no oocysts upon transmission to mosquitoes (Deligianni et al., 2011). This mutant was used to investigate the function of actin2 in midgut stages by performing genetic crosses to WT parasites. In these experiments the male and female gametocytes will form a mixture of WT ookinetes and heterozygous ookinetes, resulting from a cross between WT male gametocytes and actII(−) females (Fig. 4A). Mosquitoes were fed on mice, which had been double-infected with WT and actII(−) parasites. In parallel, an in vitro culture of ookinetes was seeded with blood from the same animal. Diagnostic PCR was performed to follow the presence of the mutant during oocyst development (Fig. 4B). As expected, both WT and mutant were detected in the ookinete sample, while oocysts bearing the mutant actII(−) gene were only present at the initial phase of oocyst development; the mutant genotype was no longer detected in oocysts at day 8 or in salivary gland sporozoites. Consequently, when infected mosquitoes were fed to naïve mice, only the WT genotype was detected (data not shown).

figure

Figure 4. Genetic cross between WT and actII(−) mutant gametocytes and genotyping of resulting oocysts reveals the loss of the actinII genomic DNA during oocyst development.

A. Schematic of the cross between WT (blue) and actII(−) gametocytes (red), of which only the females are fertile. WT male gametocytes can fertilize either WT or actII(−) females. Two populations of ookinetes are formed, those being homozygous WT for the actin2 locus and those being heterozygous actin2/+. Mosquitoes were fed on mice double infected with both strains. Females are depicted as filled circles and the males as a wavy line.

B. Genotyping with primers specific for mutant actII(−) locus (top) and for WT actin2 locus as control (bottom). Samples were: blood of a double infected mouse (BS); an in vitro ookinete culture seeded from the same mouse (Ook); midguts dissected at 3 and 8 days after feeding mosquitoes on infected mice (Ooc); salivary gland sporozoites at day 21 after feeding (Sp).

Download figure to PowerPoint

Oocysts derived from heterozygous ookinetes are blocked in sporogony

Since in these experiments normal proliferation of WT oocysts obscured any phenotype of the mutant oocysts we next performed an in vitro genetic cross of the actII(−) strain with the strain Δp47, which only produces fertile males (van Dijk et al., 2010). An ookinete culture was seeded with blood from mice infected with each strain separately, and the culture was then fed to mosquitoes, as it has been observed that the Δp47 strain can inefficiently develop into oocysts when gametocytes are fed to mosquitoes (van Dijk et al., 2010). All ookinetes resulting from the cross actII(−)p47 are heterozygous for the actin2 gene (actin2/+, in the following the female genotype is always listed first in crosses and progeny), as the actII(−) mutants form very few fertile male gametocytes (Deligianni et al., 2011). Parallel cultures of the individual strains were used to verify that these two strains were unable to form ookinetes on their own (data not shown).

The ookinetes from the cross actII(−)p47 were fed to mosquitoes in parallel with WT ookinetes. At day 3 post feeding the oocysts were labelled with the Cap380 antibody and stained for DNA. The numbers of oocysts were comparable (WT: mean 17 oocysts/midgut, prevalence 90%, n = 19; actII(−)xΔp47 mean 12 oocysts/midgut, prevalence 78%, n = 18). The heterozygous oocysts were of equivalent size to the WT (Fig. 5A), but substantial variations in the DNA content of these oocysts were noted, ranging from values similar to the WT to others with very weak DNA. At day 10 the heterozygous oocysts were considerably smaller than WT (Fig. 5A and B); DNA was not detectable or the amount was significantly less than in WT oocysts of the same age (Fig. 5A,e, inset and Fig. 5C) suggesting that DNA replication was affected in the heterozygous oocysts. This is consistent with the result from the cross of the mutant to the WT where the genomic DNA of the mutant was lost during oocyst development (Fig. 4B).

figure

Figure 5. Genetic cross between Δp47 and actII(−) mutant gametocytes results in aborted oocyst development.

A. Comparison of WT and actII(−)p47 oocysts at after feeding of an in vitro ookinete culture. Oocysts were labelled with the Cap380 antibody and DNA was stained with TO-PRO. The sizes of the oocysts are similar at day 3 after feeding (a–c). DNA content was variable in the oocysts derived from the genetic cross, two extreme cases are shown with almost WT level (left) and hardly any visible DNA (right). Scale bar, 5 μm. (d,e) Comparison at day 10 after feeding. The actII(−)p47 oocysts have remained small and the DNA is not visible (inset in e). Scale bar, 10 μm.

B. Size of oocysts measured as the area of the largest section of each oocyst imaged in the confocal microscope. ***P <0.0001 Student's t-test.

C. DNA content measured as the area of DNA of the largest section of each oocyst imaged in the confocal microscope. ***P <0.0001 Student's t-test. The number of oocyst measured in B, C: WT n = 82 and actII(−)p47 n = 80, pooled data from four independent experiments.

Download figure to PowerPoint

Mosquitoes harbouring these abnormal oocysts were allowed to feed on naïve mice. As expected, in none of the four independent experiments was an infection established, whereas the transmission of WT parasites was normal with a pre-patent period of 5 days (data not shown).

actII(−)p47 heterozygous ookinetes contain the tetraploid value of DNA

One explanation for the above results could be that meiosis in the zygote had been aberrant, similar to what has been described for the nek-2 (Reininger et al., 2009), nek-4 (Reininger et al., 2005) and misfit (Bushell et al., 2009) deletion mutants. Therefore, we measured the DNA content of ookinetes from the actII(−)p47 cross (Fig. 6A and B). The values were found to be equal to WT. Hence, the abnormal DNA values in the oocysts were not a direct result of a failed meiosis in the zygote.

figure

Figure 6. actII(−)p47 heterozygous ookinetes have normal DNA content but lack actin2.

A. Measurement of DNA ploidy. Fixed WT and actII(−)p47 ookinetes were stained with Hoechst 33342. Fluorescence intensities were measured using Image J and normalized to haploid parasites in the same field. n = 20. n.s., non-significant, Student's t-test.

B. Representative examples of WT and Δp47xactII(−) ookinetes. DNA was stained with Hoechst 33342. Scale bar, 5 μm.

C. Actin2 is absent from ookinetes lacking the female derived actin2 gene. Western blot analysis of extracts of WT and actII(−)p47 ookinetes. The blot was probed with the antiserum against actin2 and, as a loading control, with the antiserum against the SET protein.

Download figure to PowerPoint

actII(−)p47 heterozygous ookinetes lack actin2

Next, we investigated whether the heterozygous ookinetes that resulted from the actII(−)p47 cross (actin2/+) expressed actin2 at levels comparable to the WT. For this we probed a Western blot containing extracts of the ookinete populations with the antiserum directed against actin2. Actin2 was not detected in the progeny ookinetes from the cross, while it was detected in the WT (Fig. 6C). This result was unexpected as the WT actin2 allele carried by the Δp47 males is present in these ookinetes. These results suggest that actin2 in ookinetes is encoded specifically by the maternal gene.

Overexpression of GFP–actin2 blocks female development

Preliminary experiments in our laboratory had suggested that overexpression of the GFP–actin2 fusion protein had a negative effect on zygote development. The fusion protein is encoded by an episome containing the standard tgdhfr/ts gene providing resistance against pyrimethamine in the WT background; this line is called gfp::actin2. By increasing the concentration of antifolate in the mice, in which the parasites are maintained, there are more copies of the episome and thus the expression of the fusion protein is increased (Fig. S2). When ookinete cultures were seeded with blood from mice treated with a high concentration of the drug, here named HE (high expression, for details see Experimental procedures), ookinetes did not develop (Table S1). However, when the antifolate concentration was reduced ookinetes developed. Notably, under the HE conditions exflagellation was normal, indicating that GFP–actin2 overexpression does not affect male gametogenesis. Ookinete formation could be rescued by crossing the HE gfp::actin2 gametocytes to the Δp48/45 strain which produces fertile females, but infertile males (van Dijk et al., 2010). In marked contrast, no ookinetes were formed after crossing the HE gfp::actin2 line to the Δp47 mutant producing fertile males only (Table S1). Taken together these results indicated that the overexpression of the fusion protein negatively affects the female gametocytes or gametes.

Overexpression of GFP–actin2 inhibits zygote development

The negative effect of overexpression of GFP–actin2 on ookinete development in the female cell could be due to (i) failure of egress, (ii) an inability of the female gamete to be fertilized by the male or (iii) to defects in zygote maturation to ookinetes. To distinguish between these possibilities, we used a combination of the antibody labelling of Pbs21, a surface antigen on female gametes and zygotes, and the nuclear dye TO-PRO, which permits distinction between the small haploid nucleus of the activated female and the considerably larger diploid nucleus of the zygote. Samples from a culture seeded with the gfp::actin2 HE line were labelled with these reagents and an antibody recognizing GFP (Fig. 7A). Activated females, labelled with both antibodies and having a small nucleus were detected along with zygotes characterized by their bigger nuclei and strong GFP signal. These observations suggest that overexpression of GFP–actin2 permits development of female gametes and that zygotes are also formed. However, the latter do not develop into the retort form (Fig. 7A) and thus ookinete development is arrested at this stage.

figure

Figure 7. A. Overexpression of GFP–actin2 causes development arrest at the zygote stage. Samples of an in vitro ookinete culture were fixed 3 h after seeding and labelled with the 13.1 antibody, recognizing the Pbs21 surface protein of female gametes and zygotes, and the anti-GFP antibody. The DNA was stained with TO-PRO and the size of the nucleus is an indicator of the stage of the cell. Top row: one female gamete (asterisk) and one zygote (arrowhead). Middle row: Zygote (arrowhead) with four distinct nuclei, suggesting that meiosis has taken place. Bottom row: Zygote (arrowhead). No retorts or ookinetes were detected in this culture after overnight incubation.

B. Immunolabelling of ookinetes resulting from the crosses of Δp45/48xgfp::actin2 (a–e) and actII(−)xgfp::actin2 (f–i) with the antiserum against actin2 and with anti-GFP. The gfp::actin2 strain was under HE conditions, and therefore homozygous zygotes of this strain do not develop to ookinetes. (a–e) One ookinete is in a cluster with a female gamete Δp45/48, which does not express GFP, and one gfp::actin2 female gamete and one zygote expressing GFP. Gametes are indicated with an asterisk in panel e. All cells express actin2. (f–i) One ookinete resulting from the actII(−)xgfp::actin2 cross. No cytoplasmic actin2 labelling is detectable, although background label of the secondary antibody at the cell periphery is present. Inset: enhanced intensity of peripheral label of (f). (a) and (f) were taken with the same exposure time in an epifluorescence microscope. Scale bar, 10 μm.

Download figure to PowerPoint

Actin2 is encoded by the maternally derived gene

We next performed a genetic cross of the gfp::actin2 strain HE (producing fertile males) to the actII(−) mutant (producing fertile females) (Table S1). The resulting ookinetes have a paternal actin2+ allele and a male-derived GFP–actin2 episome. In parallel we crossed the same males to the Δp45/48 females, yielding ookinetes homozygous for the WT allele of actin2. We carried out IFA using the anti-GFP antibody and the actin2 antiserum. The ookinetes from the actII(−)xgfp::actin2 cross (with the genotype actin2/+,gfp::actin2) lacked both actin2 and GFP, suggesting that the paternal actin2 allele was not expressed. In contrast, actin2 was detectable in the Δp45/48xgfp::actin2 (actin2+/+,gfp::actin2) ookinetes. Again, expression of the paternally inherited gfp::actin2 episome could not be revealed in these ookinetes (Fig. 7B). These results are consistent with our results from the actII(−)p47 cross, and strengthen the notion that actin2 is encoded by the maternal copy of the gene. These findings also indicate either that episomes are not segregated into the male microgametes or that no expression of the male-derived gene takes place after fertilization, as the fusion protein was undetectable in these heterozygous cells. In conclusion, our genetic data show that maternally inherited actin2 is essential for efficient zygote to ookinete conversion.

We previously showed that the actin2 promoter is active in female gametocytes, although the protein was not detectable in these cells (Deligianni et al., 2011). Together with the data in the present study the possibility arose that actin2 transcripts in the female gametocyte are translationally repressed, as has been shown for a number of target genes that are transcribed in the gametocyte but are translated only in the zygote or ookinete (Mair et al., 2006; 2010). We investigated this possibility by RT-PCR of total bound RNA bound to immunoprecipitated DOZI and CITH, two central factors involved in translational repression in gametocytes (Fig. 8). In two parasite lines DOZI::GFP and CITH::GFP (Mair et al., 2006; 2010) we found that a proportion of actin2 mRNA is bound to DOZI and CITH, similar to p25, a known target of translational repression. DOZI, which is expressed in gametocytes to exert its function, served as a negative control. Together, these results are consistent with translational repression of the actin2 transcript in female gametocytes, a hallmark of maternally encoded genes that exert critical functions in the zygote and/or ookinete.

figure

Figure 8. RNA immunoprecipitation RT-PCR of actin2 suggests translational repression in gametocytes. Immunoprecipitation of DOZI::GFP (top panel) and CITH::GFP (bottom panel) reveals association of actin2 mRNA with both translational repressors in gametocytes. p25 mRNA served as an established translationally repressed transcript, while dozi mRNA was used as a negative association control; it is translated in gametocytes. IP = immunoprecipitation. The lanes are: RT-PCR of mRNA isolated from gametocyte lysates (input), of immunoprecipitated material using antibodies against GFP and, as negative control, myc, or of material binding to beads only.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

The most important finding from our study is the identification of a second critical function of the actin2 isoform in ookinete formation and sporogony. We had previously determined that a knockout mutant of P. berghei actin2 was arrested in male gametogenesis, and therefore this strain produced very few ookinetes (Deligianni et al., 2011). Using two complementary approaches, we could extend our analysis of the function of actin2 and overcome the first developmental defect, i.e. the block in male gametogenesis of the knockout mutant (Deligianni et al., 2011). By replacing the actin2 ORF with that of actin1 we provide genetic evidence that the two actin isoforms have distinct cellular functions. Unexpectedly, actin1 under the control of the regulatory elements of actin2 could, at least partially, complement the observed defects in male gametogenesis, resulting in formation of motile ookinetes. These ookinetes were able to traverse the midgut epithelium and form oocysts, but these did not develop into sporogonic oocysts and there was no transmission to naïve mice, revealing that actin2 is also essential for sporogonic development of the oocysts. To consolidate these findings we performed genetic crosses to produce heterozygous ookinetes, lacking one copy of the actin2 gene. We could show that in this cross ookinetes do develop and they are able to cross the epithelium when fed to mosquitoes. Although oocysts are formed, these do not undergo sporogony, most likely because of a defect in DNA replication. This resulted in a complete block in transmission of the parasite by the mosquito.

These experiments reveal two aspects of actin2 function. First, the function in male gametogenesis can, to some extent, be executed by actin1, if this isoform is expressed under control of the actin2 cis-regulatory elements. In WT parasites actin1 and actin2 are expressed in the cytoplasm of male gametocytes (Deligianni et al., 2011). However, in actII(−) parasites the endogenous actin1 isoform is nevertheless unable to rescue the block in gametogenesis. This suggests, first, that actin2 has a very specific function, at a precise time point, in a specific cellular localization, or at a very distinct concentration, and that correct regulation of gene expression is critical for the protein to carry out this unique function. A similar requirement for faithful regulation of transcription of Plasmodium genes for the correct trafficking of the protein product has been described for several important Plasmodium genes, such as apical membrane antigen-1 (AMA-1) (Kocken et al., 1998), ring-infected erythrocyte surface antigen (RESA) (Rug et al., 2004), and subtilisin 2 (SUB2) (Child et al., 2013). Second, the function of actin2 in the mosquito stages leading to arrested oocyst formation cannot be substituted by expressing actin1 in its place. Taken together, these data indicate that the function of actin2 is distinct in the two different cell environments where it is required; in the male gametocytes the requirement is less stringent than at the zygote stage.

As the actII(−) knockout parasites formed only negligible numbers of ookinetes, we used genetic crosses as an independent approach to consolidate these findings. First, we performed gametocyte crosses between the mutant actII(−) and WT. This results in a mixed population of heterozygous ookinetes (actin2/+) and WT ookinetes. After feeding to mosquitoes, oocysts developed, but those oocysts carrying the mutant allele were shown to be completely lost during development. These findings were corroborated by independent crosses using a well-studied strain which produces only fertile males, the Δp47 strain (van Dijk et al., 2010). Such a cross results in a population of heterozygous ookinetes (actin2/+), all with an identical genotype. Interestingly, in these ookinetes the actin2 protein was not detected, suggesting that the protein is expressed from the female derived copy of the gene; therefore these ookinetes are phenotypically null mutants. When these ookinetes were fed to mosquitoes, oocysts were initially formed. This observation provides evidence that actin2 has no important role in ookinete gliding motility and cell invasion. Similarly to the phenotype of the act2rep mutant, these oocysts did not undergo sporulation and nor did transmission to naïve mice occur. These results, together with those of the act2rep mutant, suggest that the absence of functional actin2 results in aborted oocyst development at an early time point and is possibly due to failed DNA replication.

In addition to our previous study, in which actin2 was shown to be abundantly expressed in the male gametocyte (Deligianni et al., 2011) we determined that the protein is also present in the first mosquito midgut stages, i.e. female gametes, zygotes and ookinetes. We have not been able to detect the protein in oocysts while the actin2 transcript was detected in ookinetes but not in oocysts. Interestingly, a proteomic analysis of Plasmodium yoelii sporozoites detected the protein in salivary gland sporozoites (Lindner et al., 2013), leaving the question open for whether this protein has yet another function, perhaps again during host switch. Specific targeting of sporozoites by a site-specific recombination system, such as FLP/FRT mutagenesis (Lacroix et al., 2011), could reveal potential additional functions of actin2 during pre-erythrocytic development of the malarial parasite.

The critical time at which actin2 is needed was also investigated. We made use of an overexpression approach, in which we expressed a GFP–actin2 fusion protein regulated by the actin2 promoter region. The fusion protein was encoded by an episome and we could regulate its copy number, and thus the amount of protein produced, by varying the concentration of the antifolate pyrimethamine used for treatment of the infected mice. This method for varying the expression of a protein encoded by an episome has also been used in experiments of Plasmodium falciparum (Epp et al., 2008). At the standard drug concentration (here called HE) ookinetes were not formed. This was due to the loss of female fertility, as ookinete formation could be rescued by crossing to the Δp48/45 strain, which produces fertile females and infertile males. The HE parasites produced female gametes, but the resulting zygotes did not develop to the retort form. We interpret these data such that expressing high amounts of the GFP–actin2 protein inhibits the function of endogenous actin2, resulting in abortion of further zygote development. This toxic effect is due to the GFP tag, as experiments using an episome expressing actin2 without any tag resulted in completely normal ookinete development even in HE conditions (data not shown). A similar negative effect of expression of a tagged copy of a gene in addition to the WT copy has also been observed in the case of alpha-tubulin 2 (Kooij et al., 2005). In the cross actII(−)xgfp::actin2, on the other hand, where actin2 is not produced in the female gamete, zygotes do form. These develop into motile ookinetes able to form oocysts, whose development, however, is aborted early.

Our experiments also supported the notion that the actin2 found in ookinetes is derived from the female copy of the parental genome, as described above. This was apparent in two independent genetic crosses. In the actII(−)p47 no actin2 was detectable in the ookinetes, despite the fact that the Δp47 males carry the WT allele. When the HE gametocytes, carrying a WT allele of the gene in the fertile males, were crossed to actII(−) parasites, producing female gametes only, no actin2 was detected. On the other hand, when crossing the HE parasites to the Δp48/45 strain, producing females carrying the WT allele, actin2 was detected in the ookinetes. There is precedence for dependence on gene expression derived from only one of the parental genomes. LAP (or PCCp) proteins are encoded by the female-derived allele and are critical for subsequent oocyst development (Raine et al., 2007; Lavazec et al., 2009), while MISFIT (Bushell et al., 2009) was found to be expressed by the paternal allele. Interestingly, in all previous cases and similar to what we found here, the function of the protein is required in the gamete or zygote stages but the phenotype of the mutation is manifested during oocyst development. In the case of the LAP genes the sex-specific expression of the gene has been shown to be dependent upon translational repression (Mair et al., 2006; 2010; Saeed et al., 2013) in the female gametocyte. Here, we provide data suggesting that the actin2 transcript is also translationally repressed in the female gametocyte, thus explaining the strict requirement on the maternal gene for actin2 expression in female gametes, zygotes and ookinetes.

A number of P. berghei mutant parasites have been described with a block in oocyst development. Of particular interest is the P. berghei mutant carrying a deletion of the gene encoding the C-CAP protein (Hliscs et al., 2010). This mutant has a phenotype resembling the one we describe here. C-CAP proteins have homology to the C-terminal actin-binding domain of Cyclase Associated Proteins. These proteins bind actin monomers and prevent actin polymerization. The P. berghei mutant was viable in the blood stages but after transmission to the mosquito, small, degenerate oocysts were formed which did not sporulate (Hliscs et al., 2010). It is tempting to speculate that actin2 in the zygote has a role dependent upon the regulatory function of C-CAP; however, experimental data to support this hypothesis are currently lacking.

Our work here has shed new light on the function of the stage-specific actin isoform actin2. The protein has at least two distinct functions during the mosquito stages of the life cycle of the parasite, in male gametogenesis, as we have previously reported (Deligianni et al., 2011), and in oocyst development. An intriguing open question is whether actin2 forms filaments in these cells. Actin2 can form filaments in vitro (Skillman et al., 2011) but we have no evidence that actin filaments are required at the stages studied here. This issue may be investigated by introducing specific mutations in the actin2 gene which will interfere with filament formation. Future studies should also aim at identifying proteins binding actin2 in the stages of interest, since they may be shared between the two actin isoforms or fulfil rather specific regulatory functions.

Our study also provides an example, of how additional essential gene functions, which manifest further down the Plasmodium life cycle, can be revealed by two complementary classical genetic approaches. While alternative strategies, including sporozoite-specific targeted deletion (Lacroix et al., 2011) and promoter swap (Laurentino et al., 2011; Siden-Kiamos et al., 2011), have been established recently, controlled excision or downregulation of genes presently remain restricted to targets that are vital for asexual blood-stage growth and have shared functions in sporozoites or ookinetes respectively. In previous studies, genetic crosses to rescue defects of mutant parasites were employed to support distinct gene functions at specific life cycle checkpoints and to exclude additional defects later in development (e.g. Raine et al., 2007; Ganter et al., 2009). The identification of consecutive critical roles of actin2 encourages systematic assessment of the growing collection of mutant parasite lines displaying defects in mammal-to-mosquito transmission (Janse et al., 2011) through multiple comparative genetic crosses, as described here. Similarly, gene complementation remains under-utilized (Goldberg et al., 2011) and, as shown here for the partial rescue of the male gamete function of actin2 by the actin1 isoforms, may offer opportunities to gain a better understanding of parasite gene functions beyond the first developmental defect.

Experimental procedures

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Parasite strains and methods

The WT strain used was ANKA 2.34. The two mutants where actin2 had been deleted have been described; actII(−) strain (Deligianni et al., 2011) and act2::mCherry (Kooij et al., 2012). Genetic crosses were carried out using knockout mutants which only produced healthy males or females, Δp47 and Δp45/48 respectively (van Dijk et al., 2001; Khan et al., 2005). The parasites harbouring the episomal gfp::actin2 construct was described earlier (Deligianni et al., 2011).

Ookinetes were cultured in vitro as described before (Sinden, 1997). Ookinete conversion was assessed as previously described (Deligianni et al., 2011). Briefly, ookinetes, female gametes and zygotes were labelled with an anti-Pbs21 antibody and at least 100 cells were counted in each experiment.

Sporozoites were isolated from dissected salivary glands at day 21 after feeding. The salivary glands were crushed with a pestle and the sporozoites collected by centrifugation.

Mosquitoes used were Anopheles gambiae strain G3 and membrane feedings were done according to Sinden (1997). Alternatively, mosquitoes were fed directly on anaesthetized infected mice. Infected mosquitoes were kept at 19°C until dissection or feeding to naïve C57BL/6 mice.

The act2::mCherry (Kooij et al., 2012) was used as recipient strain for transfection of plasmids constructs, and transfection was carried out as described previously (Janse et al., 2006).

All animal work was carried out in full conformity with Greek regulations and the protocols approved by the Ethics committee of FORTH, Crete.

Complementation and replacement constructs of actin2

Act2com and act2rep constructs were made in the vector pSD141, a derivative of the pL0006 vector (de Koning-Ward et al., 2000; Billker et al., 2004). A simplified scheme of the strategy is shown in Fig. S1A. 2.7 kb of the 5′FR and 728 bp of the 3′FR of the P. berghei actin2 gene were amplified from gDNA using the primers indicated in Table S2. The fragments were cloned into the KpnI and NheI sites and the NotI and EcoRI sites respectively. For the act2rep construct, the P. berghei actin1 ORF was amplified from gDNA and cloned into NheI and NotI sites between the actin2 5′ and 3′FR of actin2. For the act2com construct a fragment comprising the actin2 ORF and the 728 bp actin2 3′FR was inserted downstream of the actin2 5′FR. Both plasmids were linearized with ClaI before transfection of the act2::mCherry parasite strain, which had been recycled to remove the resistance cassette (Kooij et al., 2012). Parasites were cloned by limiting dilution (Janse et al., 2006). Correct integration was verified by PCR genotyping (Fig. S1B, Table S2) and Southern blot analysis (Fig. S1C). For Southern blot genomic DNA was digested with BamHI and EcoRI. After electrophoresis, the DNA was transferred to a nylon filter and probed with a 2.7 kb actin2 5′FR PCR fragment. Probe labelling and detection was done using the DIG High Prime DNA Labelling and Detection Starter Kit II (Roche) according to the manufacturer's instructions.

Antibodies and antiserum against actin2

The anti-GFP antibody was purchased from Invitrogen. The 13.1 monoclonal antibody (Winger et al., 1988) and the antibody recognizing Cap380 (Srinivasan et al., 2008) have been described. An antiserum recognizing the SET protein was used as a loading control in Western blot (Pace et al., 2006). The blocking antibody AffiniPure Fab Fragment Donkey Anti-mouse IgG was from Jackson Immuno Research.

The antiserum recognizing actin2 was produced in mice immunized with protein purified from Escherichia coli. The complete ORF of actin2 was cloned in plasmid pET23d and expression was done in E. coli BL21(DE3)pLysS containing the plasmid pMICO (Christopherson et al., 2004). The protein was isolated from insoluble inclusion bodies. The antiserum was pre-adsorbed for 15 min with a crude extract of actII(−) gametocytes before use.

Secondary antibodies were anti-mouse (Alexa-488, Alexa-555) and anti-rabbit (Alexa-488, Alexa-546) conjugated to Alexa Fluor (Invitrogen) or anti-rabbit conjugated to Cy-3 (Jackson Research).

Immunofluorescence assays of in vitro cultured ookinetes using the actin2 antiserum

All steps were carried out at RT unless otherwise stated, and all washes were in phosphate-buffered saline (PBS). A total of 1–5 × 104 ookinetes were diluted in 400 μl of PBS and the samples were collected on poly-l-lysine-coated coverslips by centrifugation at 500 g for 10 min. The supernatant was removed and the samples incubated in 500 μl of fixative [4% formaldehyde, 0.1% Triton in microtubule stabilizing buffer (MTSB, 10 mM MES, 150 mM NaCl, 5 mM EGTA, 5 mM glucose, 5 mM MgCl2 pH 6.9)] for 10 min. After removal of the supernatant 200 μl of ice-cold methanol was added to each well and incubated at −20°C for 2 min. This was followed by two washes before blocking with Fab Fragment Donkey Anti-mouse IgG diluted 1:50 in PBS with 5% NGS (Normal Goat Serum) for 1 h at 37°C. After two washes the sample was incubated with the actin2 antiserum diluted 1:200 overnight at 4°C. For controls samples were incubated in parallel only in PBS with NGS. The samples were washed twice in PBS, before the secondary antibody anti-mouse Alexa-488 was added and incubated for 1 h. DNA was stained using either TO-PRO or Hoechst 33342. The samples were mounted in Vectashield. When stained with Hoechst the samples were viewed with a Zeiss Axioskop 2 plus microscope fitted with an Axiovert CCD camera (Zeiss). TOP-RO samples were analysed using a Zeiss LSM 510 confocal laser scanning microscope with Biorad lasers. Images were analysed with ImageJ software (http://rsbweb.nih.gov/ij/).

Regulation of GFP–actin2 expression and imaging of gfp::actin2 parasites

The mice harbouring parasites transfected with the episomal gfp::actin2 construct were treated continuously with pyrimethamine to avoid the loss of the episome during schizogony. Two different concentrations of pyrimethamine were used, called HE (14 μg ml−1 diluted in drinking water, pH 3.5–5) and LE (2.8 μg ml−1 diluted in drinking water, pH 3.5–5). Forty hours prior to the IFA the water containing pyrimethamine was replaced by sulfadiazine (2.5 μg ml−1) to enrich for gametocytes (Beetsma et al., 1998). To maintain the episome 100 μl pyrimethamine (2.5 μg ml−1) was injected i.p. once.

For LE experiments, samples of an ookinete culture were removed at 3, 6 and 9 h after seeding and imaged without prior fixation. Samples at later time points as well as samples of HE experiments were fixed for 15 min in 4% paraformaldehyde and labelled with an anti-GFP (Invitrogen) and the 13.1 antibodies to distinguish the developing ookinetes from the other parasite stages. DNA was visualized with Hoechst 33342. The samples were viewed as described above.

Determination of DNA content in ookinetes

Photographs of fixed ookinete samples stained with Hoechst 33342 were analysed with ImageJ software (http://rsbweb.nih.gov/ij/). Fluorescence intensities were normalized to haploid parasites in the same fields.

IFA of dissected midguts containing oocysts

Mosquitoes were dissected in PBS and their midguts immediately placed in fixative [4% formaldehyde (Polysciences), 0.2% saponin (Sigma) in PBS] and incubated on ice for 45–60 min. All following steps were carried out at room temperature, unless specifically mentioned. Washes and antibody dilutions were done in PBS with 0.2% saponin and 5% NGS. The midguts were washed twice for 15 min on a rocking platform and then incubated for 30 min. After addition of the primary antibody the midguts were kept at 4°C overnight. The guts were washed three times for 15 min and then incubated with the secondary antibodies for 1 h. DNA was stained with TO-PRO (Invitrogen) or Hoechst33342. The guts were mounted and viewed as described above.

RT-PCR analysis

cDNA was prepared as previously described (Deligianni et al., 2011). The primers used for RT-PCR analysis of WT ookinetes and oocysts are described in Table S2. The primers used to detect expression of actin2 and actin1 in act2com and act2rep parasites were a common forward primer binding to the 5′FR and specific primers for actin2 and actin1 ORF respectively. For details see Fig. 1A and Fig. S1A (green arrows) and Table S2.

Western blot

Enriched ookinetes were sonicated on ice, and the extracts loaded on a 12% SDS-PAGE gel. After gel electrophoresis the proteins were transferred to nitrocellulose membrane filters by electroblotting. The membrane filters were incubated with the primary antiserum directed against actin2, followed by anti-mouse antibodies conjugated with horseradish peroxide. The signal was detected using the SuperSignal West Pico solution (Pierce Biotechnology).

RNA immunoprecipitation RT-PCR

Plasmodium berghei DOZI::GFP and CITH::GFP lines have been described before, they behave as WT lines and express exclusively the C-terminally GFP-tagged protein (Mair et al., 2010). RNA immunoprecipitations were performed as described in Mair et al. (2010); briefly, gametocytes were enriched using a 49% Nycodenz cushion. Following parasite lysis, a monoclonal anti-GFP antibody was used to immunoprecipitate DOZI::GFP or CITH::GFP. Controls included a monoclonal α-myc antibody and a ‘no antibody’ control. Antibody–protein–RNA complexes were finally recovered with protein G-sepharose beads. RNA isolation was done with TRIzol, followed by cDNA synthesis using oligo dT oligonucleotides. PCR primers used for dozi (PBANKA_121770) were g0907 (TGTGGCTTTGGCTGGAAAA) and g0908 (AGCGCCAATTCCCTTGTG); primers for p25 (PBANKA_051500) were g0830 (ATGTACTATGCATGAAGTCGG) and g0831 (TGAACATGGGGTATCTCC); primers for actin2 (PBANKA_103010) were g1217 (TTTTTTAATGAATTAAGGGTAT) and g1218 (TCAAGAACGATACCTGTG).

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

We thank Dr I. Kursula and Dr F. Frischknecht for stimulating discussions. G. Vrentzos provided technical support and Anthi Georgopoulou assisted in producing the actin2 antiserum.

R.M. was funded by Intermal, a Marie Curie Initial Training Consortium (Grant No. PITN-GA-2008-215281 to I.S.-K.). K.L. and K.M. are members of the EVIMalaR network [Grant #201588 of the FP7 (HEALTH) programme of the European Commission] which funded in part this work. T.W.A.K. was supported by a Max Planck fellowship. G.R.M. was supported by FCT grant projects PTDC/BIA-BCM/105610/2008 and PTDC/SAU-MIC/122082/2010.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Supporting Information

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information
FilenameFormatSizeDescription
cmi12274-sup-0001-figs1.pdf513K

Fig. S1. Integration of actin2 and actin1 ORF in the actin2 locus.

A. Schematic representation of strategy used to obtain the act2com and act2rep strains expressing actin2 and actin1, respectively, under the control of actin2 5′ and 3′FR. In step 1 (reported previously, Kooij et al., 2012) the complete ORF of the WT actin2 gene was exchanged via a cross-over/ends-out homologous recombination with a fragment encoding a red fluorescence protein (mCherry), a GFP cassette and the hybrid resistance gene allowing positive–negative selection consisting of hDHFR and yeast cytosine deaminase-uracil phosphoribosyl transferase (yFcu) yielding the strain act2::mCherry. Next, parasites were selected to recycle the resistance cassette (Step 2). These parasites were used as a recipient strain for transfection of the plasmids pAct2com and pAct2rep. Integration was achieved via a single cross-over/ends-in homologous recombination in the actin2 5′FR. This resulted in the integration of the actin2 or actin1 ORF in the actin2 locus (Step 3). Red arrows: primers for genotyping of transfectants. Green arrows: Primers used for RT-PCR analysis of transfectants.

B. Genotyping using PCR. The primers are depicted as red arrows in A and the expected sizes of the fragments are indicated (red colour below map). Left and right integration was tested as well as possible contamination of act2::mCherry parasites. Primers specific for gapdh were used for quality control of the gDNA.

C. Southern blot analysis. Genomic DNA of the recipient strain act2::mCherry, and the act2com and act2rep was digested with BamHI and EcoRI and processed for Southern blot. A PCR fragment corresponding to actin2 5′FR was used as a probe. The restriction sites and the theoretical sizes of the fragments are indicated in A (grey colour above map). The positive bands are indicated with asterisks.

cmi12274-sup-0002-figs2.pdf143K

Fig. S2. Expression of GFP–actin2 encoded by the episomal construct. Blood-stage parasites were selected using two different concentrations of anti-folate (14 and 2.8 μg ml−1).

A. Representative pictures of gametocytes originating from mice treated with the high anti-folate concentration (a–c), and with the low concentration (d–f).

B. Measurement of the intensity of the GFP signal in gametocytes from mice treated with the two concentrations of anti-folate. The difference in intensity is significant (P <0.0001 Student's t-test).

cmi12274-sup-0003-movies1.avi3309K

Movie S1. Movie of gliding WT ookinete. The movie represents a time-lapse video of 20 min.

cmi12274-sup-0004-movies2.avi2411K

Movie S2. Movie of gliding act2rep ookinete. The movie represents a time-lapse video of 18 min.

cmi12274-sup-0005-tables1.pdf251K

Table S1. Genetic crosses of gfp::actin2 strain under LE and HE conditions.

cmi12274-sup-0006-tables2.pdf337K

Table S2. Primers used for DNA constructs, genotyping and RT-PCR.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.