Parasites

The gametocyte-producer clone cl15cy1 (HP) of P. berghei ANKA was used as the reference parasite line [32]. In addition, the following mutant lines of the ANKA strain were used: 2.33, a non-gametocyte producer (NP) line [33] and 137cl8 (RMgm-15, www.pberghei.eu), a mutant lacking expression of P48/45 [4].

Generation of mutants deficient in expressing 6-cys family members

To disrupt genes encoding different members of the 6-cys family, we constructed a number replacement constructs using plasmid pL0001 (www.mr4.com) which contains the pyrimethamine resistant Toxoplasma gondii (tg) dhfr/ts as a selectable-marker cassette (SC). Target sequences for homologous recombination were PCR amplified from P. berghei genomic DNA (ANKA, cl15cy1) using primers specific for the 5′ or 3′ end of the different 6-cys genes (see Table S1 for the sequence of the different primers). The PCR–amplified target sequences were cloned in plasmid pL0001 either upstream or downstream of the SC to allow for integration of the construct into the genomic target sequence by homologous recombination. DNA constructs used for transfection were obtained after digestion of the replacement constructs with the appropriate restriction enzymes (Table S1). Replacement constructs pL1138 (p47) and pL0123 (p36), were constructed using replacement plasmid pD_B.D_T∧H.D_B [34] and plasmid pL0121 (p47&48/45) was constructed in the previously described replacement plasmid for disruption of pb48/45 (plasmid p54 is renamed here to pL1137; [4]). This plasmid was made by exchanging the 5′ pb48/45 targeting sequence with the 5′ targeting sequence of pb47. The p230pII replacement construct pL0120 is a derivative of plasmid pL0016 [35] containing the tgdhfr-ts SC, gfp (under control of the pbeef1aa promoter and 3′UTR of pbdhfr/ts) and p230p 5′ and 3′ targeting sequences [36]. Transfection, selection and cloning of mutant parasite lines were performed as described [32], [37] using P. berghei ANKA cl15cy1 as the parent reference line. For all mutants with an observable phenotype, mutants were generated and selected in two independent transfection experiments (Table S1). Of each transfection experiment we selected one cloned line for further genotype and phenotype analysis. Correct integration of the construct into the genome of mutant parasites was analysed by standard PCR analysis and Southern blot analysis of digested genomic DNA or of FIGE separated chromosomes [32]. PCR analysis on genomic DNA was performed using specific primers to amplify either part of the wild type locus (primers WT1 and 2) or the disrupted locus (primers INT1 and 2). See Table S2 for the sequence of these primers.

Analysis of expression by Northern and Western analysis

Total RNA was isolated from the different blood stage parasites of the gametocyte-producer clone cl15cy1 of P. berghei ANKA (HP), the non-gametocyte producer line 2.33 (NP) and the different mutant lines according to standard methods. To determine stage-specific transcription of the 6-cys family members, Northern blots containing RNA from different blood stages were hybridised with different gene specific probes, which were PCR-amplified using the primers shown in Table S2 (primer pairs WT1+ 2). To detect expression of the P48/45 protein we used polyclonal antiserum raised against recombinant P. berghei P48/45 as described [4]. For detection of P47 we generated the following polyclonal antiserum; a fragment of the Pb47 ORF (encoding amino acids 80–411) was PCR-amplified using primers L964 and L965 (Table S2) and cloned into the NdeI/BamHI sites of the expression vector pET-15b (Novagen) providing an N-terminal 6-Histidine tag. Polyclonal antiserum was raised in New Zealand rabbits by injection of 200 µg of gel-purified recombinant protein. Boosting was carried out subcutaneously with 3-weeks intervals using 200 µg protein in incomplete Freund's adjuvant. Serum (P47) obtained 2 weeks after the third boost was immuno-purified on immobilised purified recombinant P47. To detect P48/45 and P47 in the different mutant lines, total protein samples of purified gametocytes were fractionated on non-reducing 10% SDS polyacrylamide gels.

Phenotype analysis of parasite lines lacking expression of 6-cys gene family members

The fertility of wild type and mutant gamete populations was analysed by standard in vitro fertilisation and ookinete maturation assays [4], [17] from highly pure gametocyte populations [38]. The fertilisation rate of gametes is defined as the percentage of female gametes that develop into mature ookinetes determined by counting female gametes and mature ookinetes in Giemsa stained blood smears 16–18 hours after in vitro induction of gamete formation. Fertility of individual sexes (macro- and micro-gametes) was determined by in vitro cross-fertilisation studies in which gametes are cross-fertilised with gametes of lines that produce only fertile male (Δp47; 270cl1) or only fertile female gametes (Δp48/45; 137cl1 [4], [17], [39]. All fertilisation and ookinete maturation assays were done in triplicate on multiple occasions in independent experiments. In vivo ookinete, oocyst and salivary gland sporozoite production of the mutant parasites were determined by performing standard mosquito infections by feeding of Anopheles stephensi mosquitoes on infected mice [40]. Oocyst numbers and salivary gland sporozoites were counted at 7–10 days and 21–22 days respectively after mosquito infection. For counting sporozoites, salivary glands from 10 mosquitoes were dissected and homogenized in a homemade glass grinder in 1000µl of PBS pH 7.2 and sporozoites were counted in a Bürker-Türk counting chamber using phase-contrast microscopy [41]. Infectivity of sporozoites was determined by infecting mice through bites of 25–30 infected mosquitoes at day 21–25 after mosquito infection.

The formation of exflagellation centres (i.e. male gamete interactions with red blood cells) was determined by adding 10µl of infected tail blood to 100–300 µl of standard ookinete culture medium pH 8.2 to induce gamete formation. Ten minutes after induction of gamete formation a droplet of 5–10 µl was placed on a cover slip and analysed under a standard light microscope (40× magnification) as a hanging-drop using a well slide. When red blood cells were settled in a monolayer, the number of exflagellating male gametocytes was counted that form or did not form exflagellation centres. An exflagellation centre is defined as an exflaggelating male gametocyte with more than four tightly associated red blood cells [27]. The formation of exflagellation centres was performed using tail blood collected at day 6 or 7 from mice that were infected with 10⁵ parasites without treatment with phenylhydrazine. For quantification of male-female interactions tail blood was collected from phenylhydrazine-treated mice with high numbers of gametocytes [42]. Tail blood (10µl) was collected at gametocytemias ranging between 4–8% and added to 100µl of standard ookinete culture medium pH 8.2 to induce gamete formation. Ten minutes after induction of gamete formation, the cell suspension was placed in a Bürker-Türk counting chamber and during a period of twenty minutes the male-female interactions were scored using a phase-contrast light microscope at a 40× magnification. Attachments of males to females were scored if the male had active (attachment-) interactions with the female for more than 3 seconds. Penetration of a female by the male gamete was scored as a fertilisation event.

Polymorphisms and sequence divergence of the Plasmodium 6-cys genes

Pairwise alignments were generated between the orthologous sequences of p48/45, p47 and p230 genes in P. berghei, P. yoelii and P. chabaudi; sequences were obtained from PlasmoDB (http://www.plasmodb.org version 6.1; see Table S3 for the accession numbers of the 6-cys gene family members). Complete gene sequences for a number of these genes were obtained from the Sanger Institute (A. Pain, personal communication). Maximum-likelihood estimates of rates of non-synonymous substitution (dN) and synonymous substitution (dS) between pairwise alignments were generated using the PAML algorithm (version 3.14; [43], [44]) using a codon-based model of sequence evolution [45], [46], with dN and dS as free parameters and average nucleotide frequencies estimated from the data at each codon position (F3×4 MG model [47]). For this analysis we assumed a transition/transversion bias (i.e. kappa value) that had been estimated previously and found to be similar in case of P. falciparum and P. yoelii, i.e. 1.53 [48]. A sliding window analysis of dN/dS ratios was performed of p230, p47 and p48/45 from the three rodent parasites. We analysed the dN/dS values of these genes across their length by analysing sequentially 300bp of the gene in 150bp steps. This analysis is essentially the same as the calculation of π (i.e. the number segregating or polymorphic sites) described for p48/45 in distinct P. falciparum isolates described by Escalante et al. [29]. We obtained the single nucleotide polymorphisms (SNPs) data identified from field and laboratory isolates of P. falciparum (excluding all P. reichenowi SNPs) from PlasmoDB (www.PlasmoDB.org). The alignment of these SNPs along the different genes (to scale) was extracted from the Genome Browser page of PlasmoDB. The locations of the SNPs were aligned onto the schematic representation of the 6-cys genes of the rodent parasites. It should be noted that the alignment of the p230 gene of the different Plasmodium species was only possible around 1008bp after the putative start site. In order to determine which residues of p230, p47 and p48/45 genes were under positive selection in the rodent malaria parasites, a Bayes Empirical Bayes (BEB) analysis was performed using sequences from the 3 rodent genomes and was calculated as described in Yang et al. [49]. To test which genes were undergoing positive selection the likelihood ratio test (LRT) was performed using a comparison of site specific models of evolution [50], [51]. This test compares a ‘nearly neutral’ model (without any residues under positive selection) and a ‘positive selection’ model (with residues under positive selection and therefore under adaptive evolution). Both models assume that there are different categories of codons, which evolve with different speeds. The ‘nearly neutral’ model assumes two categories of sites at which amino acid replacements are either neutral (dN/dS = 1) or deleterious (dN/dS<1). The ‘positive-selection’ model assumes an additional category of positively selected sites at which non-synonymous substitutions occur at a higher rate than synonymous ones (dN/dS>1). Likelihood values indicate how well a model fits to the analyzed alignment and answers the question if the ‘positive selection’ model fits better to the analyzed alignment than the ‘nearly neutral’ model.

Animal ethics statement

All animal experiments were performed after a positive recommendation of the Animal Experiments Committee of the LUMC (ADEC) was issued to the licensee. The Animal Experiment Committees are governed by section 18 of the Experiments on Animals Act and are registered by the Dutch Inspectorate for Health, Protection and Veterinary Public Health, which is part of the Ministry of Health, Welfare and Sport. The Dutch Experiments on Animal Act is established under European guidelines (EU directive no. 86/609/EEC regarding the Protection of Animals used for Experimental and Other Scientific Purposes).

Four out of ten members of the 6-cys family of P. berghei are specifically transcribed in gametocytes

Ten members of the 6-cys family have been identified in Plasmodium and are found in all Plasmodium species (Table S3). We analysed the transcription profile of the 10 members during blood stage development of P. berghei by Northern blot analysis and combined this analysis with a search of publicly available literature, transcriptome and proteome datasets. This method established that multiple members are transcribed in gametocytes of which four members, p48/45, p47, p230, p230p, are transcribed exclusively in the gametocyte stage (Fig. 1A). The gametocyte specific expression of p48 and p230p has been shown before [4], [8]. Transcription of p38 occurs both in gametocytes and in asexual blood stages as has also been reported [8], whereas p12 is transcribed in all blood stages. The relative weak band observed in gametocytes might be due to low contamination of the gametocyte preparation with asexual blood stages (gametocyte samples always contain a small degree of contamination with schizonts when density gradients are used for gametocyte purification). Transcription of p41 and p12p show a complex pattern of multiple transcripts in all blood stages. The close paralogue pair p36 and p36p have quite different transcriptional profiles: p36p is not transcribed in blood stages but transcription is exclusive to sporozoites [14], [15] whereas p36 is transcribed both in gametocytes (Fig. 1B; [8], [52]) and in sporozoites [14], [15].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 1. Expression of the 10 members of the 6-cys family of Plasmodium.

A. Northern blot analysis of transcription of the 10 P. berghei genes during blood stage development of a gametocyte non-producer (NP) and a high producer (HP) line. The left panel shows the four genes that are exclusively expressed in gametocytes. P36 and p36p are shown in the right panel since they are also expressed in the sporozoite stage (see B). As (loading) controls Northern blots were hybridized to probes recognising LSU rRNA (87R primer) and the gametocyte specific gene p28. Lanes: 1) NP asynchronous blood stages (ABS); 2) NP schizonts (Schz); 3) HP asynchronous blood stages; 4) HP purified gametocytes (Gam). B. Transcription and protein expression of the 10 genes determined by RNA and proteomic analyses (G = gametocyte; F = female gametocyte; M = Male gametocyte; Bl = blood stage; Mz = merozoite; Sp = sporozoite). References: 1 [4]; 2 [52]; 3 [17]; 4 [18]; 5 [71]; 6 [72]; 7 [73]; 8 [74]; 9 [12]; 10 [75]; 11 [10]; 12 [13]; 13 [8]; 14 [16]; 15 [14]; 16 [41]; 17 [76]; 18 [15].

https://doi.org/10.1371/journal.ppat.1000853.g001

Since no polyclonal or monoclonal antibodies exist for most of the 6-cys family members of P. berghei, except for P48/45 [4], P47 (this study) , P36 and P36p [14], data on expression of these proteins in different life cycle stages mainly comes from large-scale proteome analyses. For most members of the 6-cys family which have been detected by proteome analysis, the presence of the protein coincides with transcription of its gene (Fig. 1B). The exclusive presence of P48/45, P47, P230 and p230p in the proteomes of gametocytes corresponds to the transcription pattern of their respective genes. The presence of P48 and P47 in P. berghei gametocytes has been confirmed using polyclonal antibodies against these proteins (Fig. S1; [4]). P12, P38 and P41 have been detected in the proteome of merozoites which agrees with their transcription in the asexual blood stages and with their identification in the raft-like membrane proteome of the P. falciparum merozoite surface [13]. Also the presence of P36 in proteomes of both gametocytes and sporozoites [41], [52] and P36p in sporozoites [14], [41] fits with the transcription profile of these genes. Up to now only P12p has not been detected in any proteome of Plasmodium. Comparison of the transcription and expression patterns of the 10 conserved members of the 6-cys family of P. berghei with those of P. falciparum from large scale transcriptome and proteome analyses demonstrates that the expression patterns are conserved between the rodent and human parasite (Fig. 1B) and also confirms that four out of the 10 members are specific to the gametocyte stage.

Three out of 4 members of the 6-cys family of P. berghei that are specifically transcribed in gametocytes play a role in fertilisation

We previously reported the functional analysis of mutant P. berghei parasites that were deficient in expressing P48/45, generated by targeted disruption of p48/45 through a double crossover homologous recombination event [4]. Here we have used the same approach, schematically shown in Fig. 2A, to disrupt 5 other members of the 6-cys family that are transcribed in gametocytes. We excluded p12, p12p, p41 and p36p from this analysis since the results obtained from transcriptome and proteome analyses indicate a role for the first three of these genes during the asexual blood stage development (Fig. 1B). We have previously demonstrated in both, P. berghei and P. falciparum, that P36p is involved in liver-cell infection and disruption of its gene had no effect on development of gametes and fertilisation [15], [53]. Mutant parasite lines have been generated deficient in P47 (Δp47), P230 (Δp230), P230p (Δp230p), P38 (Δp38) or P36 (Δp36) and for each gene, mutants were selected from two independent transfection experiments (Table S1). Two different Δp230p mutant lines were generated, Δp230p-I and Δp230p-II, differing in which regions of 230p have been disrupted. In mutant Δp230-I a fragment is deleted from the second 6-cys domain (i.e. first 894aa still present) onwards whereas in mutant Δp230-II the deleted fragment includes part of the first 6-cys domain (i.e. first 492 amino acids still present). In addition we generated a mutant line deficient in the expression of both P48/45 and P47 (Δp48/45&Δp47). Correct disruption of the target-genes was verified by diagnostic PCR analysis (Fig. 2B) and Southern blot analysis of separated chromosomes and/or digested genomic DNA (data not shown). To demonstrate that the mutant parasite lines were deficient in expression of the targeted gene we analysed transcription of the corresponding genes by Northern blot analysis using mRNA collected from purified gametocytes (Fig. 2B). No transcripts of p47 and p38 could be detected in ΔP47 and Δp38 mutants, and no p48/45 and p47 transcripts are present in the DKO mutant Δp48/45&Δp47. Only small, truncated transcripts were detected for p230 and p230p in gametocytes of the Δp230 and Δp230p lines and also in Δp36 a truncated p36 transcript was found. Full length transcripts of wt p230 and p230p are 8.5 and 9.5 kb respectively, whereas truncated transcripts are approximately 2.5 kb in size. Since several of the disrupted genes are organised as pairs within the genome (i.e. p230&p230p and p48/45&p47), we analysed whether disruption of one member of a pair affected transcription of the other gene. For Δp48/45 parasites it has been shown before that disruption of p48/45 had no effect on expression of its paralog P47 [4]. In this study we similarly show for p47, p230 and p230p that disruption had no effect on transcription of its paralogous member (Fig. S1 A&B). In addition to the transcription analysis of the disrupted genes, we analysed the presence or absence of the proteins P47 and P48/45 in the mutant parasites by Western analysis using polyclonal antiserum (Fig. S1C). P47 is present in wt gametocytes and gametocytes of the Δp48/45 but is absent in Δp47 and Δp48/45&Δp47 gametocytes. P48/45 is present in wild type and absent in the Δp48/45&Δp47 gametocytes.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 2. Generation and analysis of mutants lacking expression of different members of the 6-cys family of genes.

A. Schematic representation of the replacement construct used for disruption of the target genes by double cross-over homologous recombination. Correct integration of the construct results in disruption of target gene as shown (replacement locus) and is analysed by PCR (see B) using the primers INT1, 313, INT2 and 692 as shown in the figure and Table S1 and S2. Black boxes: the target regions of the 6-cys genes; grey box: the tgdhfr/ts selectable marker cassette. B. PCR analysis of correct disruption of the 6-cys genes and analysis of transcription of the genes in wild type and mutant (ko) parasite lines. PCRs were performed with primers that specifically amplify either the 5′ (INT1 and 313) or 3′ (INT2 and 692) regions of the disrupted locus (int). In addition PCR's to amplify the intact open reading frame (orf) were performed using genomic DNA of wild type parasites as a control (wt). The double knockout mutant Δp48/45&Δp47 was checked for both p47 and p48/45. Control PCR amplifying the gametocyte specific p28 gene (C). Northern blot analysis of transcription was performed using RNA extracted from gametocytes of wild type (wt) or mutant parasites. Blots were hybridised with 6-cys specific gene probes that were obtained by PCR amplification (see Table S2). As a control Northern blots were hybridized to a probe recognising the gametocyte specific gene p28. The sizes of transcripts (kb) are shown next to the Northern blots.

https://doi.org/10.1371/journal.ppat.1000853.g002

We next analysed the phenotype of the different mutant lines during gametocyte and gamete development as well as during fertilisation, ookinete and oocyst formation using standard assays for phenotype analysis of the sexual- and mosquito stages of P. berghei. Surprisingly, three of the six mutants lacking expression of genes that are transcribed in gametocytes did not exhibit a phenotype that was different from wild type parasites during these stages of development. These mutants, Δp230p, Δp38 or Δp36, showed a normal growth of the asexual blood stage (data not shown), sexual development and development of the mosquito stages up to the mature oocysts (Table 1). All these mutant lines produced wild type numbers of gametocytes and gametes and showed normal fertilisation rates as measured by in vitro zygote/ookinete production (Table 1; Fig. 3). In contrast to the absence of a discernable fertilisation phenotype with the Δp230p, Δp38 and Δp36 mutants, we found that the capacity of fertilisation is severely affected in the other three mutants, (Fig. 3A). Specifically, Δp47, Δp230 and Δp48/45&Δp47 lines showed a fertilisation rate that was reduced by more than 99.9% compared to wt, as shown by the inhibition of zygote/ookinete production in vitro (Table 1; Fig. 3A). These mutants produced normal numbers of mature gametocytes during blood stage development. The analysis of in vitro gamete formation (exflagellation of males; emergence of female gametes from the erythrocyte) by light-microscopy also revealed that the process of gametocyte and gamete formation was not affected, resulting in the production of motile male gametes and female gametes, emerged from the host erythrocyte by more than 80% of the mature gametocytes (Table 1). At 16–18h after activation of gamete formation, the in vitro cultures of Δp47, Δp230 and Δp48/45&Δp47 lines contained many (clusters of) unfertilized, singly nucleated, female gametes. This phenotype of a strong reduction of fertilisation despite the formation of male and female gametes closely resembles the phenotype of Plasmodium parasites lacking P48/45 [4]. As had also been previously observed with the P48/45 deficient mutant, the fertilisation rate of gametes of the three mutant lines seems to be more efficient in the mosquito compared to in vitro fertilisation [4]. Compared to wild type parasites, the in vivo fertilisation of the mutants is reduced by 93–98% as calculated by ookinete and oocyst production in mosquitoes (Table 1), whereas the reduction of in vitro fertilisation rate is greater than 99.9%. Infections of naïve mice through bite of 20–30 mosquitoes infected with parasites of Δp47, Δp48/45&Δp47DKO and Δp230 parasites, resulted in blood stage infections containing only gene disruption mutants (i.e. mutant genotype and no ‘wild type’ parasites), as determined by PCR and Southern analysis of genomic DNA (results not shown). These results show that gametes of all three mutant lines still have a low capacity to fertilise, resulting in the production of viable and infective ookinetes, oocysts and sporozoites. Moreover, the results obtained with the double knock-out mutant Δp48/45&Δp47 indicate that the few fertilisation events in single knock-out mutants deficient in expression of either P47 or P48/P45 (this study and [4]) cannot be explained by a compensation effect due to its paralogous protein because the Δp48/45&Δp47 mutant still shows a comparable, albeit greatly reduced, ability to fertilise and to pass through the mosquito.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 3. Fertilisation rates and male and female fertility of mutants lacking expression of different members of the 6-cys family of proteins.

The fertilisation rate is defined as the percentage of female gametes that develop into mature ookinetes (ookinete conversion rates); 1 and 2 indicate mutants obtained from independent transfection experiments. A. Self-fertilisation rates of the different mutants, showing wild type fertilisation rates of mutants Δp230p, Δp36 and Δp38. B. Cross-fertilisation rates in assays in which gametes of the Δp47, Δp230 and Δp48/45& Δp47 mutants (that were affected in their fertilisation rate) were crossed with fertile females of Δp48/45. Δp47 males are fertile and fertilise Δp48/45 females at wild type rates whereas Δp230 males are infertile. Δp230 females are fertile and are fertilised by Δp47 males at wild type levels. Gametes of both sexes of the Δp48/45& Δp47 mutant are infertile.

https://doi.org/10.1371/journal.ppat.1000853.g003

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Gametocyte/gamete production, fertilisation rate and development in mosquitoes of different mutants that lack expression of members of the 6-cys family of proteins.

https://doi.org/10.1371/journal.ppat.1000853.t001

P230 plays a role in male gamete fertility and P47 in female gamete fertility

Fertility of the male and female gametes produced by the mutant lines can be determined by in vitro cross-fertilisation studies, where gametes are cross-fertilised with gametes of parasite lines that produce either only fertile male gametes or female gametes. Such an approach was used to establish that Δp48/45 parasites produced infertile male gametes, whereas the female gametes are completely fertile [4]. We performed different in vitro cross fertilisation experiments to determine whether the reduced fertilisation capacity of the Δp47 and Δp230 mutants was due to affected male gametes, female gametes or to both sexes. Gametes of both mutants were cross-fertilised with female gametes of Δp48/45 (males are infertile) to determine male fertility of Δp47 and Δp230. Male gametes of Δp47 were able to fertilise Δp48/45 females (at wild-type levels) whereas the males of Δp230 were unable to fertilise the Δp48/45 females (fertilisation rates <0.01%; Fig. 3B). These results demonstrate that male gametes of Δp47 are viable with wild type fertilisation capacity and therefore the fertilisation defect of Δp47 must be due to infertile females. The normal fertility of male gametes of Δp47 has also been shown in previous studies in which the males of this mutant have already been used in other cross-fertilisation studies [17], [39], [54], [55]. The lack of fertilisation in the crossing experiments of gametes of Δp230 with Δp48/45 shows that P230 plays a role in male fertility. In order to test the fertility of Δp230 females we crossed the gametes of this line with the fertile male gametes of Δp47 (as mentioned above the females are infertile). We find that Δp47 male gametes are able to fertilise Δp230 female gametes in a manner identical to their ability to fertilise Δp48/45 females (Fig. 3B). This demonstrates that female gametes of Δp230 have a fertility that is comparable to wild type female gametes and that the fertilisation defect is the result of infertile males. Crossing experiments performed with gametes of the double knockout mutant, Δp48/45&Δp47 with gametes of either Δp230, Δp47 or Δp48/45 did not result in increased fertilisation rates (<0.01%), demonstrating that gametes of both sexes are infertile in the double knock-out mutant (Fig. 3B).

Infertile Δp230 males form exflagellation centres but do not attach to females and fertile males do not attach to infertile Δ47 females

In P. falciparum it has been shown that male gametes lacking P230 expression have a reduced capacity to adhere to red blood cells, as measured by the formation of ‘exflagellation centres’ [27]. We therefore examined the ability of P. berghei male Δp230 gametes to attach to erythrocytes, by microscopic examination of exflagellation centre formation under standardized in vitro conditions. In these experiments 76–92% of exflagellating wt males and 72–90% exflagellating Δp230 male gametocytes, formed such centres (Table 2), indicating that in contrast to P. falciparum Δp230 in P. berghei both wt and Δp230 male gametes have a similar ability to interact with red blood cells. Gametocytes that did not form exflagellation centres were often floating on/above the red blood cell layer during exflagellation. Further analysis of single, free male gametes of Δp230 revealed that they were highly motile and often attach to red blood cells, producing characteristic red blood cell shape deformations due to the active interactions between the male gamete and the erythrocyte. Male gametes lacking expression of P48/45 do not attach to female gametes as has been previously shown by analysing male-female interactions by light microscopy [4], [5]. We therefore analysed the interactions between male and female gametes of Δp230 or Δp47, between 10 and 30 minutes after induction of gamete formation using phase-contrast microscopy. In wt parasites attachment of males to females was readily detected with a mean of over 25 attachments during a 20 minutes period of observation, with a mean of more than 6 confirmed fertilisations (i.e. male gamete penetrations; Table 2). In preparations of gametes of both Δp230 and Δp47 not a single fertilisation event was detected and the number of male and female gamete attachments was drastically reduced (Table 2). We observed that while male gametes of both mutants undergo active interactions with red blood cells and platelets, attachment of males to female gametes are hardly ever observed. These results show that P230 like P48/45 is a male fertility factor involved in recognition or attachment to females and that P47 is a female fertility factor involved in recognition or adherence by the male gamete. Whether P48/45 and P230 once on the surface of the male gamete directly interact with P47 on the surface of the female gamete is unknown. Unfortunately, repeated immuno-precipitation experiments with anti-P. berghei P48/45 antibodies and wt gamete preparations, in order to identify interacting partners, were unsuccessful (data not shown).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. The interactions of Δp230 and Δp47 male gametes with red blood cells (exflagellation centres) and female gametes (attachment and fertilisation).

https://doi.org/10.1371/journal.ppat.1000853.t002

Sequence polymorphism of Plasmodium proteins involved in fertilisation

Analyses of sequence polymorphisms of p48/45, p47 and p230 of laboratory and field isolates of P. falciparum has provided evidence that these proteins are under positive selection [28], [29], [30], [31]. We analysed synonymous (dN) and non-synonymous (dS) polymorphisms of p48/45, p47 and p230 by comparing these genes in three closely related rodent parasites P. berghei, P. yoelii and P. chabaudi by making use of the newly available gene sequences (www.PlasmoDB.org version 6.1). The updated dN/dS values for these genes obtained here, which is commonly used as an indicator of positive selection, were in all comparisons higher than the mean dN/dS value of all genes within the respective genomes (Table S4). However, only the dN/dS ratio of p47 in the P. berghei/P. yoelii comparison showed a significant difference with the mean dN/dS value (0.82 compared to the mean dN/dS of 0.26). Overall, P47 is in the top 4–6% of fastest evolving proteins in the rodent parasite genomes as compared to top 10–16% for P230 and 15–50% for P48/45 (Table S4). In addition, we have used the likelihood ratio test (LRT) to analyse if these genes were undergoing neutral or positive selection (see Materials and Methods). This test shows that p47 is indeed under positive selection (P = 0.006) when comparing the site/residue specific models of evolution.

We next examined sequence mutations in the same genes in more detail by performing a comparative dN/dS ratio analysis across these genes using small and corresponding regions of these genes using a ‘sliding window analysis’ (i.e. 300bp in 150bp intervals; Fig. 4; Table S4). This analysis showed that p47 has an exceptionally elevated dN/dS value (i.e. 1–2) in one area corresponding to the truncated B-type domain II as defined by [7]. Interestingly, although P230 had a relatively low overall dN/dS value (0.33–0.44), the sliding window analysis revealed that P230 contains several areas where the dN/dS ratio is higher than 1.0 with an increased ratio in all 3 species in particular around the B-type domain IV as defined by Gerloff et al. (2005). In order to analyse similarities in the location of sequence polymorphism between P. falciparum and the three rodent parasites, we aligned all known single nucleotide polymorphisms (SNPs) described for P230, P47 and P48/45 in P. falciparum (i.e. www.PlasmoDB.org; [56], [57], [58]) with the dN/dS ratios determined by the ‘sliding window analysis’ (for details see Materials and Methods; Fig. 4). Interestingly, the elevated dN/dS ratios of p47 domain II and domain IV of P230, both correspond with the location of high SNP densities in the orthologous P. falciparum genes. These findings would suggest that similar regions in the p47 and p230 genes of rodent parasites and P. falciparum are subject to positive selection. To predict which residues of the three P. berghei genes are under positive selection we performed a Bayes Empirical Bayes analysis (BEB; [49]). This analysis calculates dN/dS values (ω values) on each residue of a particular protein when the genes encoding these proteins are compared in least 3 similar species and an ω>1 indicates positive selection on a residue. For P47 ten residues were identified undergoing positive selection with ω values ranging between 4 and 7 (Table S5). Nine of these 10 residues are confined to the first two domains of P47 including the region B-type domain II. In P48/45 four residues were identified (ω values ranging between 1 and 2) and for P230 only one amino acid (ω = 1.3). Interestingly, this one residue in P230 (i.e. residue 845V) maps to the corresponding region of the P. falciparum P230, domain IV, where 6 of the 27 non-synonymous polymorphisms described by Gerloff et al. map (Table S4).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 4. Polymorphisms and sequence divergence across p230, p48/45 and p47.

Schematic representation of p230, p47 and p48/45 (shown to scale). A- and B-type recurring domains (green and grey respectively; [7]) are shown and the numbering of domains (I–XIV) are shown as according to [7]. The putative Signal Peptide (SP) is indicated in red. Above each gene the locations of all single nucleotide polymorphisms (SNPs) are shown as identified in different P. falciparum strains in PlasmoDB (www.PlasmoDB.org; August 2009). Dark blue diamonds: non-synonymous polymorphisms; Light blue diamonds: synonymous polymorphisms. Below each gene the dN/dS ratios are shown across the length of the three rodent Plasmodium orthologs. This dN/dS analysis is performed using a ‘sliding-window’ analysis, where 300bp of corresponding DNA sequence was compared at 150bp intervals. The gene from each species has been compared to the same gene of the other species; Red: P. berghei against P. chabaudi; Blue: P. berghei against P. yoelii; Green: P. yoelii against P. chabaudi. The complete gene sequence is only available for P. berghei and P. chabaudi; The 5′ end of all three rodent parasite p230 genes is shorter than those of the P. falciparum p230 and therefore alignment of the P. falciparum to the rodent p230s is only possible ∼1kb after the start site.

https://doi.org/10.1371/journal.ppat.1000853.g004