Osmiophilic bodies are membrane-bound vesicles, found predominantly in Plasmodium female gametocytes, that become progressively more abundant as the gametocyte reaches full maturity. These vesicles lie beneath the subpellicular membrane of the gametocyte, and the release of their contents into the parasitophorous vacuole has been postulated to aid in the escape of gametocytes from the erythrocyte after ingestion by the mosquito. Currently, the only protein known to be associated with osmiophilic bodies in Plasmodium falciparum is Pfg377, a gametocyte-specific protein expressed at the onset of osmiophilic body development. Here we show by targeted gene disruption that Pfg377 plays a fundamental role in the formation of these organelles, and that female gametocytes lacking the full complement of osmiophilic bodies are significantly less efficient both in vitro and in vivo in their emergence from the erythrocytes upon induction of gametogenesis, a process whose timing is critical for fertilization with the short-lived male gamete. This reduced efficiency of emergence explains the significant defect in oocyst formation in mosquitoes fed blood meals containing Pfg377-negative gametocytes, resulting in an almost complete blockade of infection.
Transmission of malaria parasites to the mosquito vector is initiated upon the ingestion of the sexual stage gametocytes during a blood meal. Within the mosquito, gametocytes rapidly differentiate to produce male and female gametes that are released from the red blood cells and fuse to form the fertilized zygote. Over the ensuing 24–36 h, the zygote transforms into a motile ookinete, and infection within the mosquito vector becomes firmly established when the ookinete invades the midgut epithelium to form an oocyst on the haemolymph side of the midgut. It is within these oocysts that the sporozoites (the stage infectious to vertebrates) develop (for review, see Janse and Waters, 2004; Baton and Ranford-Cartwright, 2005). The sexual stages of Plasmodium therefore provide the crucial link in parasite transmission between vertebrate hosts.
The initial process of sexual differentiation involves the development of gametocytes within the bloodstream of the infected host, from a morphological stage indistinguishable from trophozoites (Stage I) to a highly specialized, infective mature gametocyte (Stage V) (Field and Shute, 1956; Hawking et al., 1971; Carter and Miller, 1979; Alano and Billker, 2005). During sexual differentiation, specialization of the male microgametocyte and the female macrogametocyte is reflected in production of sex-specific proteins with many fewer gametocyte-specific proteins common to both sexes, as recently highlighted in a proteome analysis conducted on separated male and female P. berghei gametocytes (Khan et al., 2005).
At an ultrastructural level, differences between microgametocytes and macrogametocytes become apparent from Stage III of maturation, when the female begins to show a marked proliferation of mitochondria, ribosomes, Golgi vesicles and dense spherules (Langreth et al., 1978; Jensen, 1979; Sinden, 1982). Sexual dimorphism becomes more discernable by Stage IV, with differentiation between the sexes occurring in both nuclear and cytoplasmic compartments (Sinden, 1982). At this stage osmiophilic bodies, which are small ovoid to elongated electron-dense organelles derived from Golgi vesicles, become apparent by electron microscopy at the periphery of macrogametocytes, and occasionally in microgametocytes (Sinden, 1982; Ponnudurai and Lensen, 1986), and they progressively increase in number as the macrogametocytes continue on to full maturity.
In natural infections, mature (Stage V) macrogametocytes escape from the erythrocyte only within the mosquito midgut. Osmiophilic bodies have been postulated to be involved in this emergence process, not only because of their location and timing of appearance and abundance (Rudzinska and Trager, 1968; Sinden, 1982), but also for their resemblance to dense granules present in the asexual stages of Plasmodium, which release their contents into the parasitophorous vacuole shortly after merozoite invasion (Aikawa et al., 1990; Culvenor et al., 1991). Consistent with this hypothesis are observations in several Plasmodium species and in Haemoproteus (Aikawa et al., 1969; Bradbury and Roberts, 1970; Bannister et al., 2005) that osmiophilic bodies establish physical contacts with the gametocyte pellicle through ducts, and that in P. gallinaceum (Aikawa et al., 1969) and in other blood parasites of the Apicomplexan lineage, their number declines after gamete emergence from the erythrocyte (Bradbury and Roberts, 1970; Sterling and DeGiusti, 1972). The efficiency of macrogametocyte escape from the erythrocyte is greater than for microgametocytes, which frequently remain trapped within the host erythrocyte plasma membrane (Sinden, 1983; Aikawa et al., 1984).
Currently the only protein that has been associated with osmiophilic bodies is Pfg377 in Plasmodium falciparum. Pfg377 is a gametocyte-specific protein encoded by a single exon of 9360 bp located on chromosome 12 (Alano et al., 1995). Homologues of this highly hydrophilic protein have been found in human, rodent, avian and primate malaria species (http://www.plasmodb.org), and the protein encoded by the P. berghei orthologue was detected only in the proteome of purified female gametocytes (Khan et al., 2005). The presence of an N-terminal signal sequence possibly targets this molecule to the osmiophilic bodies and, given its timing of expression around the onset of osmiophilic body appearance, it has been speculated that Pfg377 may be involved with the formation or development of this organelle (Severini et al., 1999).
In order to investigate the role of Pfg377 in gametocyte and osmiophilic body development, we targeted the gene for disruption. The resulting mutants were compared with wild-type (wt) parasites in ultrastructural studies and in their ability to generate a productive infection in the mosquito. The results presented here are consistent with a pivotal role for Pfg377 in the formation of osmiophilic bodies, and for the efficient escape of macrogametocytes from the erythrocyte once ingested by the mosquito.
Disruption of the pfg377 locus does not affect gametocyte production or sex ratio
To disrupt the pfg377 locus on chromosome 12, the plasmid construct p377R (Fig. S1) was used to transfect a P. falciparum 3D7 line that is effective at producing gametocytes capable of infecting Anopheles mosquitoes. Resulting transfectants that were resistant to WR99210 drug selection after two cycles of drug relief/drug exposure were cloned. Genomic DNA (gDNA) extracted from three parasite clones, derived from two independent transfection experiments (clones Δ377/E4 and Δ377/B7 were derived from the same transfection), was digested with PvuII and BamHI and hybridized with a pfg377 probe (Fig. 1, Fig. S1). The 1.8 kb fragment hybridizing with the pfg377 probe in wt 3D7 was absent in all Δpfg377 parasite clones (Fig. 1), in which it was replaced by two fragments of 2.7 and 2.8 kb. In addition, an intense 3.8 kb fragment was also visible in the Δpfg377 parasites, which is consistent with the presence of multiple copies of the plasmid integrated into the locus. It is well documented that several copies of a plasmid can integrate into the malaria parasite genome (Miller et al., 2002; McCoubrie et al., 2007). It is, however, difficult to distinguish this event from the presence of residual episomal plasmid, as in plasmid rescue experiments multiple inserted plasmids can transform E. coli after undergoing recombination. In our analysis plasmids could indeed be recovered in such experiments using DNA from Δpfg377 clones, although around half of the recovered plasmids had different restriction enzyme digest patterns to the original transfection vector (data not shown). It is also known that episomal plasmids are rapidly lost from transfected populations in the absence of drug selection (O'Donnell et al., 2001; 2002). Furthermore, Southern blot analysis of chromosomes from the uncloned Δpfg377 line separated by pulsed-field gel electrophoresis and probed with the pfg377 region present in the transfection plasmid hybridized to a single band corresponding to chromosome 12 (on which pfg377 resides), while it failed to detect bands representative of episomes, which typically run as multiple bands beneath this chromosome (data not shown). Finally, polymerase chain reaction (PCR) amplification of gDNA using oligonucleotides specific for the 5′ and 3′ boundaries of the integrated plasmid (data not shown), confirmed the genomic structure of the integration event predicted by the Southern blot analysis, indicating that all Δpfg377 parasites after cloning harboured the integration event.
The ability of parasite cultures of the Δpfg377 clones to produce gametocytes was compared with the wt line. As parasites are known to lose their ability to produce gametocytes during continuous in vitro propagation (Bhasin and Trager, 1984; Graves et al., 1984), the wt line in all subsequent analyses was one that had been grown for the same length of time in culture as the transgenic parasites (denoted 3D7 cyc2). While conversion rates to gametocytes were variable in different experiments, as is routinely observed in P. falciparum cultures, no consistent, dramatic differences in gametocyte production and morphology were observed in Giemsa-stained smears of the Δpfg377 clones and the 3D7 cyc2 wt line.
Giemsa-stained smears from cultures containing predominantly stages IV and V gametocytes, where sexual dimorphism is more readily detectable, were examined to determine the proportion of male and female gametocytes in the different clones (Table 1). Analysis was based on sex-specific morphological characters, such as presence of a more compact nuclear shape, bluer cytoplasmic staining and less dispersed haemozoin granules in female gametocytes compared with males. Gametocytes of the wt and the Δpfg377 parasite clones showed comparable strongly female-biased sex ratios as is typically observed with P. falciparum (Ranford-Cartwright et al., 1993).
Table 1. Sex ratio of gametocytes used for ultrastructural analysis.
The number of male and female gametocytes were determined by analysis of Giemsa-stained smears from cultures containing predominantly stages IV and V gametocytes.
Pfg377 is not expressed in the Δpfg377 parasite clones
To determine whether disruption of the pfg377 locus resulted in lack of expression of the Pfg377 protein, Western blot and immunofluorescence analysis analyses were performed. Western blot analysis on extracts of stages IV–V gametocytes with antiserum raised against Pfg377 fragment B (from Pfg377 amino acid 666–1146) confirmed the presence of Pfg377 protein only in wt parasites (Fig. S2). The presence of multiple bands reacting to the anti-Pfg377 antiserum in this parasite line indicate that Pfg377 is fragmented. This confirmed preliminary immunoprecipitation experiments with Stage IV gametocytes (R. Carter, unpublished), and analysis of proteomic data on Pfg377, in which the N- and C-terminal portions were recovered from different protein bands (Lasonder et al., 2002). It is possible that the amino-terminal portion of Pfg377 could still be expressed in the Δpfg377 parasites, producing a truncated protein of 610 amino acids (65–70 kDa). However, analysis of the same extracts with antiserum A2, raised against the N-terminal portion of Pfg377 (from amino acid 184–385), detected no fragment of this size on Western blots, in either soluble or pellet fractions (data not shown). A peptide fragment of 35 kDa was detected by antiserum A2 in both wt and Δpfg377 parasites, but a fragment of the same size was also recognized by antiserum B (Fig. S2), which was raised using the central portion of Pfg377, with no obvious homology to that used to raise antiserum A2. It is therefore most likely that this represents a cross-reacting polypeptide, rather than a truncated form of Pfg377.
Immunofluorescence analysis was conducted on acetone-fixed smears of parasite cultures containing gametocytes mostly at stages IV and V of maturation (Fig. 2). Expression of the gametocyte-specific protein Pfg27 was detected in gametocytes of the wt 3D7 cyc2 gametocytes and in all Δpfg377 clones (Fig. 2). In contrast anti-Pfg377 antibodies stained gametocytes of the wt clone only and not those produced by the Δpfg377 clones. Taken together, these analyses indicate that the pfg377 locus had been successfully disrupted on two independent occasions, and that the resulting Δpfg377 parasite lines produce gametocytes that can develop and appear morphologically normal in the absence of wt Pfg377 protein.
Osmiophilic bodies are rare or absent in female gametocytes of Δpfg377 parasites
Pfg377 appears at the onset of the biosynthesis of osmiophilic bodies, and is the only known molecular marker of these organelles. An ultrastructural analysis was conducted on gametocytes of the Δpfg377 clones to investigate possible consequences of the absence of this protein on the presence and morphology of osmiophilic bodies. Gametocytes at stages IV and V of maturation were obtained from clones 3D7 cyc2, Δ377/E4 and Δ377/B7, and analysis of Giemsa-stained smears, as well as IFA using antisera raised against Pfg377 B fragment (for wt parasites), confirmed that the majority of parasites subjected to ultrastructural analysis were indeed female gametocytes in all parasite clones (Table 1). Ultrastructural analysis of gametocyte ultrathin sections indicated that osmiophilic bodies, clearly distinguishable by their oval shape, electron-dense content and distinctive subpellicular location, could be readily observed in gametocytes of the wt parasite clone 3D7. Organelles were predominantly concentrated at the apical tips of the gametocyte, as classically reported, but they were also readily detectable in cell transversal sections (Fig. 3A). This result was obtained from analyses of 3D7 specimens from two independent preparations.
In contrast, examination of several sections from gametocyte preparations of Δpfg377 parasite clones Δ377/E4 and Δ377/B7, specifically chosen as they showed the gametocyte apical ends, failed to detect osmiophilic bodies in the vast majority of Δpfg377 gametocytes (Table 2). These gametocytes had clear female-specific ultrastructural features such as a large, round-shaped nucleus containing a nucleolus-like denser region (Fig. 3B and C), an elaborate mitochondrion (Fig. S3), and aggregated haemozoin granules (Fig. 3C). Osmiophilic bodies were never observed in sections of clone Δ377/B7. In an experiment examining independent gametocyte sections, a small number of normal osmiophilic bodies were detected in four out of 97 sections of Δ377/E4 gametocytes (Table 2 and Fig. S3C). However, in all four cases ultrastructural details were not sufficient to confirm or refute the possibility that these sections were taken from the rarer male gametocytes, in which low numbers of osmiophilic bodies are reported to be present (Ponnudurai and Lensen, 1986). In summary, ultrastructural analysis indicated that osmiophilic bodies were absent or very significantly reduced in female gametocytes of two Δpfg377 clones (Table 2, P < 0.0000001 for both clones examined). Although this approach cannot be taken as a truly quantitative measurement, these data strongly support the hypothesis that the Pfg377 protein is required for the formation or maturation of osmiophilic bodies.
Table 2. Presence and distribution of osmiophilic bodies (OBs) in independent gametocyte ultrathin sections in wt and Δpfg377 clones.
No. of gametocyte sections examined
Sections with OBs (%)
Gametocytes with ≥ 5 OBs
Gametocytes with < 5 OBs
Significant difference in OB number compared with wt (3D7 cyc2) control: Fisher's exact test, P < 0.0000001.
ΔPfg377 parasites emerge less efficiently from red blood cells during gametogenesis
In order to investigate possible effects of the defective complement of osmiophilic bodies on gametogenesis, this process was examined in Pfg377-deficient gametocytes in vitro and in vivo. Male gametogenesis in vitro was not significantly affected, with similar numbers of exflagellation events observed in wt and Δpfg377 lines (data not shown). The ability of female gametocytes to undergo gametogenesis and to fully emerge from the erythrocyte was first examined in vitro. In these experiments, emergence was examined by inducing gametogenesis and then scoring over the ensuing 15 min the ability of fixed gametes to react with monoclonal antibody 11E3 (Fig. S4). This antibody recognizes an epitope of Pfs230 (Read et al., 1994) accessible on the surface of emerged gametes, whereas staining in gametocytes requires permeabilization for reactivity. This experiment showed that the ability to ‘round up’ (the first step of gametogenesis) was unaffected in the Pfg377-defective gametocytes. Importantly, however, it indicated that the mean per cent of female gametes that were extracellular 15 min after induction was significantly higher in wt 3D7 compared with any of the Δpfg377 clones (Fig. 4; one-tailed t-test assuming unequal variance, P = 0.016 for Δ377/G5; P = 0.006 for both Δ377/E4 and Δ377/B7). At this time, 45–80% (mean 58%) of wt 3D7 gametes were extracellular, whereas extracellular female gamete numbers in all three of the Δpfg377 clones were significantly lower, ranging from 17% to 35% (means of 35%, 29%, 26% for Δ377/G5, Δ377/E4 and Δ377/B7 respectively).
As triggering of gametogenesis involves signals from the mosquito that are absent in vitro, further investigations were performed in vivo. Gametocytes from 3D7 cyc2 and Δ377/E4 were fed to mosquitoes, and their development in the mosquito midgut followed for 6 h after feeding (Fig. 5). An antibody against the gamete and zygote surface protein Pfs25, and one specific for the erythrocyte membrane, were used to monitor the progress of gamete emergence in this time-course experiment. By 1 h post feeding, the majority (58%) of wt parasites were found to be extracellular, whereas a significantly smaller proportion (33%) had fully emerged from the red blood cell in the Δ377/E4 line (P = 0.01). By 6 h post feeding, a significantly higher proportion of Δ377/E4 gametes were still intracellular and significantly fewer parasites were extracellular compared with the 3D7 cyc2 line, in which a mean of 77% had now fully emerged [P = 0.0001 (extracellular) and P = 0.006 (intracellular)]. Substantial impairment of gamete emergence in the Δpfg377 line was clearly shown by the observation that the proportion of extracellular gametes remained significantly lower in the Δpfg377 line compared with wt parasites in the subsequent 5 h, such that the fraction of female gametes of the Δpfg377 line that were extracellular at 6 h was less than the proportion of wt parasites that had fully emerged at 1 h post feeding. It should be noted that while it is possible that absence of Pfg377 somehow alters the reactivity of the Δpfg377 gamete surface to the antibodies used in these experiments, thus leading to an overestimation of the number of non-emerged gametes in the Pfg377-defective lines, we consider this unlikely as the antibodies used to score emergence were different for the in vitro and in vivo experiments.
In conclusion, despite the fact that experimental conditions in vitro and in vivo are drastically different, both experiments gave similar results and indicate that the Δpfg377 female gametes are significantly impaired and less efficient at emerging from their red cells than the wt parasites, suggesting that they are consequently significantly less accessible for fertilization by the short-lived motile male gametes.
Δpfg377 parasites have significantly reduced infectivity to mosquitoes
To investigate whether the observed reduction in the fraction of fully emerged Δpfg377 gametes would still be able to support mosquito transmission, we analysed the ability of the three Δpfg377 lines to produce oocysts in mosquitoes. Accordingly, gametocytes of wt 3D7 cyc2 and gene-disrupted parasite clones Δ377/B7, Δ377/E4 and Δ377/G5 were fed to Anopheles stephensi mosquitoes via membrane feeding. In total, five experimental infectious feeds were performed, with each Δpfg377 parasite clone tested on at least three occasions (Fig. 6, Table S1).
Despite the relatively long time that the 3D7 wt transfection control (3D7 cyc2) had been in continuous culture, reasonable infection levels were obtained with this line in the five experimental feeds (infection prevalence ranged from 25% to 50%), with oocyst numbers on mosquito midguts ranging from 0 to 45. This compares with infection prevalence of over 90% with oocyst intensities of 50–100 for the 3D7 parent line (data not shown). In contrast, parasites of the three Δpfg377 clones were significantly impaired in establishing an oocyst infection in mosquitoes (Fig. 6, Table S1). The mean prevalence of infection across all experiments was significantly lower for each of the Δpfg377 clones compared with the cyc2 control (P = 0.005 (Δ377/G5), P = 0.0014 (Δ377/E4), P = 0.0013 (Δ377/B7)). Oocyst intensities were also significantly lower in all (clones Δ377/G5 and Δ377/B7) or almost all (clone Δ377/E4) experiments (Table S1). These results show that the Δpfg377 cloned lines Δ377/B7, Δ377/E4 and Δ377/G5 are significantly impaired in their ability to infect mosquitoes in terms of both prevalence and intensity of oocyst infection.
Reversion events explain the majority of mosquito infections in Δpfg377 clones
While the statistical evidence clearly demonstrates that Pfg377 and osmiophilic bodies play an important role in mosquito infectivity, the occasional mosquito infection observed with the Δpfg377 clones (particularly clones Δ377/G5 and Δ377/E4) prompted us to analyse if this was due to genetic reversion at the disrupted pfg377 locus. As integration of p377R into the pfg377 locus had occurred by a single crossover event, excision of the insertion vector could reconstitute an intact locus expressing the wt gene product, as described in previous Plasmodium gene disruption experiments (Sultan et al., 1997; van Spaendonk et al., 2001). PCR reactions with primers diagnostic for the disrupted or intact pfg377 locus were thus performed on whole midgut DNA material taken from mosquitoes that had become infected with the different parasites (Table 3). All mosquitoes infected with 3D7 cyc2 gave a PCR product only when using primers specific for intact pfg377, confirming as expected that only the wt pfg377 locus was present. From four of the five mosquitoes infected with clone Δ377/E4, and from six out of the nine mosquitoes infected with Δ377/G5 both the wt (revertant) and the disrupted locus could be amplified (Table 3), indicating that indeed reversion can and did occur in these parasites with a disrupted pfg377 gene. In addition, only the wt (revertant locus) could be amplified from one mosquito infected with clone Δ377/G5 (harbouring 1 oocyst). On the other hand, only the disrupted locus could be amplified from one mosquito infected with clone Δ377/E4 (harbouring a single oocyst), and from two mosquitoes infected with clone Δ377/G5 (harbouring one oocyst and four oocysts). For clone Δ377/B7, only one oocyst was observed in three membrane-feeding experiments but no PCR product could be obtained, precluding any conclusion.
Table 3. Results of PCR analysis of parasite DNA from infected mosquitoes.
No. of oocysts on midgut
PCR product from pfg377 locus
Where two or more oocysts and both alleles are present it is not possible to assign genotype unambiguously, as the same result would be obtained for two heterozygous oocysts (each with wt and disrupted alleles), two homozygous oocysts (one homozygous for the wild-type allele and the other homozygous for the disrupted allele), or a heterozygous and (either) homozygous oocyst.
In summary, this analysis revealed that reversion had occurred in at least two of the Δpfg377 lines, and a degree of reversion to a wt pfg377 locus could explain the presence of oocysts in 11 out of 14 mosquitoes infected with the Δpfg377 lines. While this confirms that disruption of pfg377 severely impairs the ability of parasites to infect mosquitoes, parasites in three infected mosquitoes contained only the disrupted locus, suggesting that this event alone is not sufficient to completely abrogate infection.
Although the gametocyte-specific protein Pfg377 has been associated with the osmiophilic bodies in Plasmodium, its relative importance in gametocyte/gamete development and mosquito infectivity had not been addressed previously. Furthermore, the function of osmiophilic bodies remained speculative. The results from this work provide strong evidence not only for a fundamental role of Pfg377 in osmiophilic body development but also that osmiophilic bodies and Pfg377 are required for efficient emergence of macrogametocytes from the red blood cell in the mosquito midgut, a process required for the establishment of a mosquito infection. As Pfg377 is present only in female gametocytes, it is not expected to play a role in emergence of male gametes or microgametogenesis.
Southern blot analysis and immunofluorescence and Western blot experiments using anti-Pfg377 antibodies demonstrated successful disruption of the pfg377 gene in two independent transfections, resulting in loss of wt Pfg377 protein in transgenic parasites. To ensure that the resulting phenotypes of these transgenic parasites could be related back to the disruption of pfg377 and not to some aberrant genetic event, the wt parasites in this study were carefully maintained in culture in parallel. Although complementation of the pfg377 KO clones with a full-length pfg377 gene could more definitely prove this, it is not feasible because of the very large size of the gene (9.3 kb).
Neither the number of gametocytes, nor their sex ratio, was affected by disruption of pfg377 in any of the three Δpfg377 parasite lines. However, electron microscopy analysis on female gametocytes showed that osmiophilic bodies were undetectable in one Δpfg377 clone, and in another Δpfg377 clone appeared in significantly fewer gametocyte sections, and in significantly reduced numbers per section compared with wt gametocytes. These results provide a strong indication that the biogenesis, maturation or accumulation of the osmiophilic content of such organelles is severely affected in gametocytes lacking the Pfg377 protein.
Depletion of osmiophilic bodies in the Δpfg377 gametocytes was found to have relevant physiological consequences on the process of gametogenesis. Although the initial step of gamete ‘rounding up’, which is common to both sexes, and exflagellation of male gametes proceeded without significant differences in wt and Δpfg377 lines, emergence of female gametes in the pfg377 disrupted lines was significantly affected both in vitro and in vivo. In wt parasites, around 60% of female gametes became extracellular by 15 min after induction of gametogenesis in vitro. In contrast, the number of extracellular gametes was significantly reduced in all three of the Δpfg377 parasite lines, to approximately half that seen in the wt parasites. The failure of Δpfg377 parasites to emerge efficiently was confirmed with in vivo experiments in A. stephensi. At the first time point examined, 1 h after the start of a blood meal containing infectious gametocytes, almost 60% of wt female gametes had emerged from the erythrocytes, but only approximately half this number had successfully emerged in the Δpfg377 parasite line. Although absence of Pfg377 and reduced/absent number of osmiophilic bodies did not completely abolish emergence of the Δpfg377 female gametes, impaired emergence of the defective parasites had dramatic consequences on further parasite development in the mosquito. Mosquito transmission experiments indeed showed that the Δpfg377 lines generated a marked reduction in oocyst prevalence and intensity of infection compared with wt parasites, with mean prevalence reduced by 77–95%, and median oocyst intensity reduced by 88–92%. Reduction is probably even more pronounced, considering that reversion events reconstituting a wt pfg377 locus were detected by PCR analysis in the majority of oocysts formed in mosquitoes fed the Δpfg377 lines, with only three oocysts formed from Δpfg377 gametes.
The observed drastic effects on mosquito transmission caused by inefficient but not absolute abolishment of gamete emergence can be explained by considering that timing and efficiency of emergence is a critical parameter in gamete recognition and fertilization. Furthermore, gamete to ookinete development is a substantial population bottleneck in parasite development in the mosquito (Vaughan et al., 1994; Alavi et al., 2003; Baton, 2005) where even relatively small defects in gamete formation and recognition may be significantly amplified. It has been estimated from experimental infectious feeds that only 7% of female gametes develop into ookinetes, and of these 90% fail to cross the midgut epithelium to form oocysts (Baton, 2005), i.e. less than 0.5% of ingested macrogametocytes in a typical membrane feeding experiment develop to oocyst stage. It is not therefore necessary to reduce macrogamete numbers to zero to reduce the probability of oocyst infection to zero. In this respect, a single experiment aiming to pinpoint the developmental blockade of the Pfg377-defective parasites showed an 86% reduction in ookinete numbers 24 h after the feed in midguts infected with Δpfg377/E4 compared with the wt parasites (data not shown). This observation supports the hypothesis that impaired emergence of the Δpfg377 gametes results in a dramatic decrease in fertilization.
The present study investigated the cellular and molecular mechanisms of emergence of female gametes, which are less characterized than those governing the spectacular events associated to emergence of the motile male gametes (Alano and Billker, 2005). The significantswelling and enlargement of the female gamete at the onset of gametogenesis is thought to be one mechanism of destabilization of the erythrocyte (Sinden et al., 1978; Aikawa et al., 1984). However, the peculiar presence of osmiophilic bodies in female gametes, no longer present or much reduced in number in the emerged female gamete, has led to speculation that the discharge of their content is the major contributor to gamete egress from the erythrocyte. The work presented here provides the first functional evidence that the drastic reduction or disappearance of osmiophilic bodies in female gametocytes owing to disruption of the pfg377 locus greatly affects the process of gamete emergence. Pfg377-negative female gametes remain trapped within or closely associated with the erythrocyte in greater numbers than wt female gametes, and reach the extracellular environment at a much-reduced rate. The observed reduced efficiency of emergence of these gametes is critical for fertilization, as male gametes, which exflagellate at comparable levels in the Δpfg377 and wt parasites, are not able to survive and be functional for the prolonged time required to fertilize the Δpfg377 macrogametes. The absence of mitochondria in male gametes limits internal energy stores and lifetime in these cells. Indeed fertilization between gametes is usually complete 15–20 min after initiation of gametogenesis (Carter et al., 1979), which suggests that the significant reduction in efficiency in emergence of Δpfg377 female gametes is probably sufficient to block the majority of fertilization events.
The function of Pfg377 in osmiophilic body physiology is still unclear. While orthologues of Pfg377 have been found in rodent, avian and primate malarias, none are apparent in other Apicomplexan parasites, including Cryptosporidium parvum, C. hominis, Toxoplasma gondii and Theileria annulata. Genome-wide searches on these Apicomplexans using profile hidden Markov models of the conserved N- and C-terminal regions of the Plasmodium g377 family members failed to identify any significant homology, although this approach did produce significant hits to other annotated proteins in P. berghei and P. chabaudi (data not shown). It is noticeable that the female gamete-specific presence of osmiophilic bodies, and their disappearance at gamete emergence, is a highly conserved feature of sexual differentiation in Plasmodia, as are the cellular events of gametogenesis in both sexes, despite the major differences in gametocyte maturation among Plasmodium species. It is therefore likely that sequence conservation of pfg377 orthologues reflects a conserved role of this protein, and might help in identifying functional domains, which may represent an attractive target for malaria parasite transmission blocking strategies. We are nevertheless aware that the observed ability of a very minor fraction of gametes lacking the Pfg377 protein to escape the erythrocyte and be fertilized highlights the need to identify multiple transmission blocking target molecules.
Generation of Δpfg377 parasite lines
Nucleotides corresponding to bp 206–1842 of the coding sequence of pfg377 were amplified from P. falciparum 3D7 gDNA using oligonucleotides 377F and 377R and the resulting 1.6 kb product was cloned into the BglII and XhoI sites of pHH1 (Reed et al., 2000). Primer sequences are listed in Table S2. This plasmid, termed p377R, contains the human dihydrofolate reductase gene as a selectable marker (Fig. S1).
The infectious gametocyte-producing P. falciparum clone 3D7 (Walliker et al., 1987) was used to derive transgenic parasite lines lacking Pfg377. Parasites were cultured and synchronized using standard protocols (Trager and Jensen, 1976; Lambros and Vanderberg, 1979). For transfection, ring stage parasites (∼5% parasitaemia) were electroporated with 80 μg of purified plasmid DNA as previously described (Fidock and Wellems, 1997) and drug selection with 2.5 nM WR99210 was initiated 48 h later. Parasites containing integrated forms of the plasmid were obtained by repeated drug cycling, whereby WR99210 was removed for 3 weeks before reapplying. Selected parasites were cloned by limiting dilution using standard protocols (Rosario, 1981). The control line 3D7 cyc2 was treated in a similar manner with mock transfection (no plasmid in the mixture), similar subculturing and cloning procedures and was maintained in parallel for the same amount of time as the disrupted lines.
PCR and Southern blot analysis
Genomic DNA was extracted from asexual blood-stage parasites (Coppel et al., 1987) or from individual mosquito midguts containing oocysts (Ranford-Cartwright et al., 1991). PCR and Southern blot analysis of restriction fragments were performed using standard procedures (Sambrook et al., 1989). To determine the integrity of the pfg377 locus in transgenic and wt parasites, the primers 377iF1 and 377iR1 (Fig. S1) were used to amplify gDNA, generating a 1.8 kb product from an intact pfg377 locus only. Primers 377iF1 and PbDT3′R or primers CAM5′int and 377iR1 amplify 1.8 kb fragments from parasite gDNA containing a disrupted pfg377 locus only. For amplification of oocyst DNA an additional nested amplification reaction was performed (Ranford-Cartwright et al., 1993), using nested primers 377iR-N1 and 377iF-N1 to detect an intact pfg377 locus, or nested primer 377iR-N1 with PbDT3′R in a hemi-nested reaction to detect a disrupted pfg377 locus. All primer sequences are listed in Table S2.
Indirect immunofluorescence assay of gametocytes
Acetone-fixed smears of mature gametocytes, cultured according to standard methods (Carter et al., 1993), were incubated with a mixture of rat anti-Pfg377 antibodies raised against region B (residue 666–1146) of the protein (diluted 1:100) (Alano et al., 1995) (Fig. S1). A rabbit antiserum specific for the gametocyte protein Pfg27, expressed from the onset of gametocytogenesis throughout the later gametocyte stages, was used as a positive control (1:500) for gametocytes irrespective of age and sex. After incubation with a mixture of fluorescein isothiocyanate (FITC)- and tetramethylrhodamine isothiocyanate (TRITC)-conjugated secondary antibodies (1:200 and 1:500 respectively), slides were washed and mounted in Vectashield antifade reagent (Vector Laboratories) with 500 ng ml−1 4,6-diamindino-2-phenylindole (DAPI). Parasites were visualized by fluorescence microscopy with a Leitz DMR fluorescent microscope equipped with filters BP 340–380 (DAPI), BP 470–490 (fluorescein), and BP 515–560 (rhodamine), with the same fields photographed using the different filters.
Transmission electron microscopy
Gametocytes at stages IV and V of maturation were purified on Percoll gradients and processed according to Perry and Gilbert (1979). Cells were fixed overnight at 4°C with 2.5% glutaraldehyde, 2% paraformaldehyde and 2 mM CaCl2 in 0.1 M sodium cacodylate buffer (pH 7.4). Parasites were washed in cacodylate buffer and post-fixed with 1% OsO4 in 0.1 M sodium cacodylate buffer for 1 h at room temperature, treated with 1% tannic acid in 0.05 M cacodylate buffer for 30 min and rinsed in 1% sodium sulphate in 0.05 M cacodylate buffer for 10 min. Fixed specimens were washed, dehydrated through a graded series of ethanol solutions (30–100% ethanol) and embedded in Agar 100 (Agar Scientific, UK). Ultrathin sections prepared using a MT-2B Ultramicrotome (LKB – Pharmacia) were stained with uranyl acetate and lead citrate and examined using an EM 208 Philips electron microscope.
Emergence of gametes in vitro
A monoclonal antibody (mAb) 11E3 (a gift from Richard Carter, University of Edinburgh), which recognizes an epitope of Pfs230 on the gametocyte/female gamete surface was used (Read et al., 1994). This mAb is unable to react with fixed gametocytes (stages II–IV) and intracellular gametes, unless permeabilized with 0.1% Triton X-100 for 10 min, whereas it reacts with the surface of extracellular gametes (Fig. S4). Gametogenesis was induced in gametocyte cultures of wt 3D7, Δ377/G5, Δ377/B7 and Δ377/E4 by lowering the temperature to 30°C and increasing pH to 8.2 (Micks et al., 1948). Samples were removed 15 min after the induction and immediately fixed in suspension with 4% paraformaldehyde/0.075% glutaraldehyde. Following incubation with mAb 11E3 (1:150 dilution) and then TRITC-labelled anti-mouse IgG (1:200) in the presence of Hoechst to stain nuclei, counts were performed under fluorescence microscopy. Female gametes were identified by positive Hoechst fluorescence, and round shape (under bright field illumination), and then further classified as extracellular by a positive red (rhodamine) fluorescence, or as intracellular by the absence of red fluorescence (Fig. S4). Immature, uninduced gametocytes still present in the samples, clearly recognizable by the elongated shape, consistently failed to show any fluorescent reaction.
Emergence of gametes and development of ookinetes in vivo
To monitor the emergence and development of parasites from gametocytes through to ookinetes, 6–10 mosquitoes were dissected at 1 and 6 h after the infectious feed, corresponding to peak gamete and developing ookinete numbers (Baton, 2005), and their blood meal contents were prepared individually for IFA using a modification of a previously published protocol (Chege and Beier, 1994; Baton, 2005), details of which can be found in Online supplementary material. Antibodies recognizing the protein Pfs25 were used to label specifically the different stages of malaria parasite development within the mosquito (Chege and Beier, 1994). Pfs25 is expressed at very low levels within the cytoplasm of immature and mature gametocytes but undergoes significant upregulation and redistribution upon transformation into macrogametes (Vermeulen et al., 1985). From herein through the subsequent differentiation of the zygote into the ookinete, Pfs25 is found predominantly on the parasite surface. Thus by double-labelling parasites with antibodies specific for Pfs25 and for human glycophorin A (CD235a), one of the major proteins of the erythrocyte membrane, it was possible to classify the parasites as female gametes or zygotes (collectively defined as round forms), or as ookinetes, and to assess whether they were retained intracellularly (red cell membrane visible around parasite identified by glycophorin A-specific rhodamine fluorescence), were in the process of emerging (rhodamine fluorescence indicating red cell around one end of the parasite, but not completely encircling it) or had fully emerged from the red blood cell (no rhodamine fluorescence).
P. falciparum infectious feeds
P. falciparum gametocytes for mosquito infectivity studies were induced in vitro using standard procedures (Carter et al., 1993), and were fed via membrane feeders to non-bloodfed A. stephensi female mosquitoes according to standard protocols (Carter et al., 1993). Mosquitoes were dissected 10 days after the infectious blood meal and the midguts examined for the presence of oocysts. Oocyst numbers were counted with a 40× objective (400× magnification) on a Leitz compound microscope.
The infection prevalences (presence of any oocysts on the mosquito midgut) obtained with wt and Δpfg377 clones were compared using the χ2 or Fisher's exact tests as appropriate. Means were compared using a heteroscedastic Student's t-test (unequal variance), with arcsin transformation of the percentages. The distributions of oocyst numbers on mosquito midguts were all found to be non-normal, making mean-based parametric statistical analysis inappropriate (Medley et al., 1993). Oocyst intensities were compared between parasite clones by χ2 comparisons of oocyst distributions fitted to negative binomials using the Genmod procedure in the sas version 8e for Windows statistical package (Bell and Ranford-Cartwright, 2004). Corrections for multiplecomparisons were made where necessary using the Dunn-Šidák correction. Fisher's exact tests were performed using R × C (M.P. Miller, 1997, http://www.marksgeneticsoftware.net/), and all other statistical tests were performed using sas version 8.2.
The mean percentages of intracellular, extracellular and emerging Pfs25-positive parasites were compared at each time point using a heteroscedastic Student's t-test (unequal variance), with arcsin transformation of the percentages.
We thank Fiona McMonagle, Elizabeth Peat and Keith Scott for maintenance of the mosquito colonies, Luke Baton for help with parasite cultures and mosquito dissection, Dr L. Piro and Professor G. Girelli, Centro trasfusionale Dipartimento di Biopatologia umana, University of Rome ‘La Sapienza’, the Australian Red Cross Blood Service, and Dr M. Peterkin, Dr A. Docherty and the staff at the Glasgow and West of Scotland Blood Transfusion Service, for the provision of human blood and serum. Work at ISS, Rome, was supported by EU FP6 grant SIGMAL (Contract LSHP-CT-2004 012174), by the European Network of Excellence BioMalPar (Contract LSHP CT-2004-503578), and by the Italian Ministry of Health. This work was also supported by the National Health and Medical Research Council of Australia. B.S. Crabb is an International Research Scholar of the Howard Hughes Medical Institute.