https://orcid.org/0000-0001-9901-1997
Figures
Abstract
Colonisation of mosquitos by the malarial parasite is critically reliant on the invasive ookinete stage. Ookinete invasion of mosquito is coordinated by the apical complex, a specialised parasite structure containing components for secretion, attachment and penetration. While studies have investigated cytoskeletal and secretory elements, it is currently unknown if signalling modules are present or functional at the apical complex. Here we elucidate the role of a cryptic cyclic nucleotide-binding protein which we name CBP-O. PbCBP-O showed a marked localisation to the ookinete apex and disruption of the protein severely compromised ookinete invasion of mosquitos. Domain dissection analysis revealed that the N- and C-termini have distinct functions. Intriguingly, PbCBP-O exhibits dual binding specificity to both cGMP and cAMP. Our findings suggest the apical tip of the ookinete is a platform to transduce cyclic nucleotide signals essential for malaria parasite transmission.
Author summary
Malaria parasites complete their life-cycle within mammalian and mosquito hosts requiring specialised forms to invade and establish infection in host tissue. In order to transmit to the mosquito, the parasite forms a motile ookinete that uses the apical end to attach and penetrate the mosquito midgut membrane. Previous studies have identified several proteins localised to the apex of the ookinete. However, the functions of many of these proteins are unknown. Here, we dissect the role of a putative cyclic nucleotide-binding protein, CBP-O, in establishing ookinete infection within the mosquito. We discovered that in the absence of the protein, efficiency of transmission is significantly compromised. Interestingly the ability of ookinetes to move was not affected, but the parasites were unable to pass through the midgut membrane. We also reveal that both the N- and C- terminal domains of the protein perform distinct functions and are both required for parasite transmission. Sequence and spatial conservation of the protein across the phylum Apicomplexa suggests an equivalent and crucial function of CBP-O in these parasites.
Citation: Kwecka D, Wang Z, Eigminas E, Liu L, Regan JC, Kim C, et al. (2025) A divergent cyclic nucleotide binding protein promotes Plasmodium ookinete infection of the mosquito. PLoS Pathog 21(9): e1013467. https://doi.org/10.1371/journal.ppat.1013467
Editor: Mathieu Brochet, Université de Genève Centre Médical Universitaire: Universite de Geneve Faculte de Medecine, SWITZERLAND
Received: January 30, 2025; Accepted: August 18, 2025; Published: September 2, 2025
Copyright: © 2025 Kwecka et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All data are in the manuscript and/or supporting information files (ST2).
Funding: This work was supported by University of Edinburgh startup funds and Wellcome Trust Seed Award (212398/Z/18/Z) to N.P and the UKRI Biotechnology and Biological Sciences Research Council (BBSRC: BB/M010996/1) to D.K. Wellcome Trust:https://wellcome.org/ UKRI-BBSRC: https://www.ukri.org/councils/bbsrc/ Funders did not play any role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Malaria continues to be a major cause of childhood disease resulting in 263 million cases and almost 600,000 deaths in 2023 [1]. The Apicomplexan parasite, Plasmodium, responsible for causing malaria, completes its complex lifecycle in a vertebrate host where it causes disease and, a mosquito vector that is necessary for parasite transmission. Transmission is initiated when the mosquito takes a blood meal containing Plasmodium male and female gametocytes. Upon encountering the mosquito environment gametocytes are activated and develop into gametes that fertilise to form a zygote. Within 16–24 hours zygotes transform into polarised, crescent-shaped ookinetes. Ookinetes are specialised for seeking, interacting with and penetrating the insect midgut epithelium for successful colonization of the mosquito.
Cyclic nucleotide signalling mediated by guanosine 3’,-5’-cyclic monophosphate (cGMP) and adenosine 3’,-5’-cyclic monophosphate (cAMP) is essential for maintaining the invasive life style of Plasmodium [2,3]. In ookinetes, coordination of cGMP synthesis by guanylate cyclase β and hydrolysis by phosphodiesterase PDEδ is critical for timely regulation of cGMP effector kinase Protein Kinase G (PKG) and governs ookinete morphology and motility [4,5]. GCβ mutant ookinetes show a strong motility defect which can be recapitulated by pharmacological inhibition of PKG demonstrating a critical function for cGMP mediated signalling cascade in ookinete gliding [6]. While gliding motility is essential to power ookinete displacement, signalling processes regulating ookinete passage into mosquito tissue are poorly understood.
Only three cellular effectors of cyclic nucleotide signalling are predicted in the Plasmodium genome which include PKG, cAMP-dependent protein kinase (PKA) and a cyclic nucleotide domain containing protein, designated as exchange protein activated by cAMP or EPAC [2]. While EPAC is not essential for the P. falciparum asexual blood stages and accumulates mutations in culture adapted parasites, the gene locus is preserved in field isolates and rodent parasite species [7–9]. This suggests EPAC is either required for in vivo growth of asexual stages or is important for parasite transmission. A large-scale screen of the molecular architecture of the apicomplexan apical complex identified P. berghei homologue of PfEPAC, PBANKA_1025300 localised to a specific apex compartment in all invasive or zoite stages of the parasite: merozoites, ookinetes and sporozoites [10]. Zoites of the Apicomplexa phylum are defined by a specialised structure at their polar tip called the apical complex [11]. The apical complex is not only important for maintaining the cytoskeletal structure of the zoite, it equips the parasite with a channel to secrete molecules that facilitate motility, as well as attachment and invasion of the host cell [12–15].
The distinct concentration of PBANKA_1025300 to the apex of the ookinete implies a critical function at the apical complex; however, it is unknown whether PBANKA_1025300 binds cAMP or coordinates ookinete motility and invasion. In this study, we systematically dissected the role of putative EPAC in ookinete function and discovered the protein regulates efficiency of ookinete invasion of the mosquito midgut epithelium. We reveal distinct functions of the N-terminal and C-terminal domains and provide the first direct evidence that the protein exhibits dual-specificity to both cAMP and cGMP. Finally, we found no evidence the protein is a guanine exchange factor and therefore renamed EPAC to cyclic nucleotide binding protein in ookinetes, or CBP-O.
Results
CBP-O is localised at the tip of ookinetes and enables mosquito infection
We first interrogated if PBANKA_1025300 is the predicted rap guanine nucleotide exchange factor [16]. Mammalian isoforms of EPAC are divided into a regulatory and a catalytic domain (Fig 1A). The regulatory domain contains the cyclic nucleotide binding domain or CNB (two CNBs in the case of EPAC2) and a Dishevelled, Egl-10/pleckstrin (DEP) domain. The catalytic portion is formed by the domains involved in the guanine nucleotide exchange, namely Ras exchange motif (REM) domain, RAS association (RA) domain, and CDC25-homology domain (CDC25HD) [17,18]. In contrast, the domain architecture of the putative CBP-O in Pb or Pf does not support function as a guanine exchange factor (Fig 1A). Consistent with findings noted for PF3D7_1417400 by Patel, Perrin and colleagues, we found the N-terminus of PBANKA_1025300 contains no recognizable domains and, structures required for the recruitment of Rap1 GTPase are absent. However, the C-terminus contains four predicted cyclic nucleotide binding domains (CNBs), out of which CNB1–3 are consecutive and the final CNB4 caps the protein. Additionally, a fifth highly degenerate domain, CNBX is inserted between CNB3 and 4. This suggests PBANKA_1025300 might be a novel cyclic nucleotide responsive protein which has acquired functions specific to parasitic life.
(A) Schematic representation of the domains of mammalian EPAC isoforms composed of a regulatory and a catalytic portion. PbCBP-O shows an uncharacterised N-terminal domain, followed by five putative cyclic nucleotide binding domains (1-4 and X). DEP = Dishevelled, Egl-10/pleckstrin, cyclic nucleotide binding domain = CNB, Ras exchange motif = REM, RAS association = RA and CDC25-homology domain = CDC25HD. (B) Live imaging of GFP-tagged PbCBP-O shows expression at the ookinete’s apical tip. Scale bar = 5μm.
Tagging PbCBP-O by adding GFP to the C-terminus revealed the protein is predominantly expressed as a distinct punctum at the apical tip of ookinetes, with some diffuse staining in the cell body as well (Figs 1B and S1A). This is consistent with previous studies showing the localization of PbCBP-O and its Toxoplasma orthologue (TGME49_219070) at the apex of the conoid [10]. The appearance of a small variation in positioning of CBP-O in our study compared to ookinetes from previous findings might be attributed to use of the ookinete plasma membrane marker P28 in the Koreny study [10]. Alternatively, spinning disc confocal imaging on live ookinete cells without labelling in our study, potentially captured apical protrusion of ookinetes. Interestingly, localization of CBP-O in our study is akin to the T. gondi orthologue in the extruded conoid [10]. Nevertheless, both studies confirm the expression of CBP-O at the Apicomplexan conoid. The conoid is located within the apical complex, a specialised structure containing the machinery required for invasion such as secretory organelles and specialised cytoskeletal elements [19].The highly specific apical localization of PbCBP-O in zoites implies a role during parasite invasion.
To determine the function of PbCBP-O in zoites, we deleted the complete ORF using CRISPR-Cas9 and replaced the locus by a GFP cassette constitutively expressing across the whole life cycle (Fig 2A). Diagnostic PCR confirmed successful integration of the repair fragment at the cbp-o locus (S1B Fig). Successful deletion of pbcbp-o indicated that it is not essential for asexual stage development and when compared to wild-type, PbΔCBP-O parasites grew at a similar rate over three replication cycles (Fig 2B). PbΔCBP-O parasites exhibit no change in parasitaemia and form gametocytes in mice, and therefore it is unlikely that PbCBP-O has a crucial function during the asexual stage. These observations concur with previous studies that showed P. falciparum CBP-O is dispensable for asexual growth in vitro [7,9]. To examine the role during transmission, PbΔCBP-O gametocytes were obtained from sulfadiazine treated mice and investigated for their ability to activate in response to xanthurenic acid in ookinete media. No defects in exflagellation were observed (Fig 2C) and 24 hours later when gametocyte conversion into ookinetes was assessed, no defect was observed in the ability to form ookinetes (Fig 2D). However, when Anopheles stephensi mosquitoes were fed on mice infected with wild-type or PbΔCBP-O parasites and midguts dissected 7 days later to check for oocyst presence, we observed a striking 98% decrease in oocyst load (Fig 2E).
(A) Schematic of strategy used for deletion of pbcbp-o. (B) In asexual blood stages, both wild-type and PbΔCBP-O parasites grow at a similar rate. Parasitaemia over three replication cycles. Error bars, ± SD and n = 3 independent experiments. (C) Exflagellation is not affected in PbΔCBP-O parasites (n = 2 independent experiments, paired t-test). (D) Ookinetes are formed in the absence of PbΔCBP-O parasites (n = 1). (E) PbΔCBP-O parasites fail to infect the mosquito midgut. Images show a mosquito midgut and infected green oocysts (dissections performed at day 7 post infection, wild-type n = 40, PbΔCBP-O n = 43, unpaired t-test). In all relevant panels, statistical significance is indicated as: ns = not significant, and **** p < 0.0001. For panels B, C, D and E, medians are indicated.
CBP-O deficient ookinetes are motile but cannot cross the midgut epithelium
Gliding motility is critical for midgut traversal of ookinetes [20,21]. We assessed the speed of PbΔCBP-O ookinetes using an in vitro Matrigel-based assay. The ookinetes displayed a 3- D meandering motion and, speed of both wild type and PbΔCBP-O ookinetes were comparable (Fig 3A; S1 and S2 Movies). Moreover, over 80% of mutant and wildtype ookinetes showed the characteristic left-handed chiral movement (S3 Fig; S3 and S4 Movies). To determine if the defect in oocyst production was due to deficient ookinete development in vivo, mosquitoes were fed on mice infected with PbΔCBP-O parasites and 19 hours after transmission, mosquito midguts were dissected to assess ookinete presence. No difference in ookinete numbers was observed between wild-type and PbΔCBP-O parasites (Fig 3B) confirming that PbΔCBP-O can produce mature ookinetes in vivo. Secretion of adhesins and proteases from micronemal organelles through the apical complex facilitates ookinete motility, escape from the blood bolus and breaching of the midgut epithelium [22–25]. However, total levels of micronemal proteins such as CTRP, WARP and Chitinase in PbΔCBP-O remained comparable to wild type ookinetes (Fig 3C). Nor did we observe any differences in chitinase secretion.
(A) PbΔCBP-O ookinetes show regular motility. Average speed is 3.07 and 3.2 μm/min for wild-type and ΔCBP-O, respectively (wild-type n = 38; PbΔCBP-O n = 48, representative data from 3 independent experiments, unpaired t-test). (B) Comparable ookinete maturation in the mosquito midgut is observed in PbΔCBP-O and wildtype parasites. Nineteen or twenty-one hours post-infection, six midguts were scored per replicate (n = 2). Significance was analysed using an unpaired t-test. (C) Expression of micronemal proteins including CTRP, WARP and chitinase are unaffected in the absence of CBP-O. Actin is used as loading control (left panel). Chitinase secretion is maintained in PbΔCBP-O ookinetes. Western blot of culture supernatants from wild type and PbΔCBP-O ookinetes show similar levels of chitinase secretion (right panel), representative data from 2 independent experiments, (D) Immunofluorescence showing decreased ookinete presence in the basal side of mosquito midguts. Midguts were fixed and stained with P25 (plasma membrane marker) and DAPI 24-25 hours post-infection. P25+ parasites were quantified, and single dots show number of ookinetes counted per midgut (n = 20 midguts per genotype, unpaired t-test.). Scale bar = 200μm. (E) Quantification of salivary gland sporozoites. Each data point is an average of five infected mosquitoes (representative data from two independent experiments). (F) Expression of An. stephensi SRPN6 mRNA relative to An. stephensi rps7 in mosquitoes that were fed uninfected blood (control), fed with wt or PbΔCBP-O infected blood (Data are mean ± SD, n = 25 midguts per condition, pooled from 4 independent experiments, unpaired t-test with Welch’s correction). (G) Quantification of sporozoites isolated from salivary glands nineteen days post ookinete injections. Each data point is an average of five infected mosquitoes (representative data from n = 2 independent experiments, unpaired t-test). In all relevant panels, statistical significance is indicated as: ns = not significant, ** = p < 0.01, *** = p < 0.001 and **** = p < 0.0001.
Ookinetes penetrate the midgut epithelium and, passage into the lumen between the midgut epithelium and basal lamina layer is obligatory to establish a mosquito infection. To examine if PbΔCBP-O crossed the midgut epithelial layer, infected midguts were dissected 24–25 hours after transmission followed by an immunofluorescence assay. The blood bolus was removed, midguts washed in PBS and, P25 antibody staining was used to count the parasites that had crossed the epithelium. Strikingly, removal of CBP-O resulted in a significant decrease in ookinete numbers in the midgut epithelium and, consequently there were fewer sporozoites in the salivary glands (Fig 3D & 3E). To understand if the decreased infection in midguts is due to failure of ookinetes to invade the midgut epithelium or parasite elimination by the mosquito immune system, we analysed expression of An. stephensi SRPN6 immune gene. AsSRPN6 is a protease inhibitor that is transcriptionally activated in response to ookinete invasion [26]. Compared to uninfected blood-fed mosquitoes, midguts containing wildtype parasites had significantly elevated AsSRNP6 mRNA expression (Fig 3F). Notably, PbΔCBP-O failed to induce levels of SRNP6 compared to wildtype parasites indicating that the mosquito immune system doesn’t perceive the mutant ookinetes.
To test whether bypassing the midgut barrier would rescue the mutant phenotype and allow PbΔCBP-O to mature into salivary gland sporozoites, we generated mutant and wildtype ookinetes in vitro and injected 700 ookinetes into the mosquito haemocoel. When salivary gland sporozoites were assessed 19 dpi, comparable numbers of sporozoites were obtained from wildtype and PbΔCBP-O ookinete injections (Fig 3G).
Taken together, our experiments reveal that ookinetes require CBP-O to invade the mosquito midgut and the protein may regulate efficient traversal through midgut epithelium.
The CNB and N-terminal domains are required for efficient midgut invasion
PbCBP-O contains an N-terminal region without any recognisable domains, but five predicted cyclic nucleotide binding domains (CNB) span the second half of the protein (Fig 4A). To investigate the contributions of the N- and C- regions, three domain deletion mutants tagged with GFP were generated using CRISPR-Cas9 in a parasite line (1804cl1) that constitutively expressed red fluorescence protein [27]. The mutants include: (i) removal of the distal and most conserved CNB, ΔCNB4: 2940–3136 a. a. deleted, (ii) all five CNBs removed ΔCNB1–4/X:1716–3136 a. a. deleted and (iii) the N-terminus excision ΔN-term: 1–1694 a. a. deleted. Similar to PbΔCBP-O, the domain mutants displayed wild-type rates of ookinete transformation (Fig 4B). However, both N-terminal deletion or complete removal of CNBs decreased oocyst load intensity by 96.63% and 96.5%, respectively, signifying the crucial role of the domains in establishing a midgut infection (Fig 4C). The individual CNB4 mutant exhibited no defects in oocyst load (Fig 4C), suggesting that the remaining CNBs are sufficient to maintain CBP-O function. Interestingly, while the infection prevalence of all mutant lines is comparable to the wildtype, the intensity of infection is significantly reduced in ΔCBP-O, ΔCNB1–4/X and ΔN-term mutants (Fig 4C) suggesting that CBP-O ensures efficient invasion of the midgut. When we tested for the ability of parasites to penetrate the midgut epithelium, significant proportion of ΔCNB1–4/X and ΔN-term ookinetes failed to cross the epithelial barrier (Fig 4D and 4E). Accordingly, reduced midgut invasion and oocyst intensity impacted sporozoite load in salivary glands (Fig 4F). These findings demonstrate that both the N-terminal region and CNB domains are important for CBP-O to regulate ookinete passage into mosquito tissue.
(A) Schematic representation of full length and GFP tagged mutant PbCBP-O lacking N-terminus (ΔN-term) or cyclic nucleotide binding domains (ΔCNB). All mutants were generated in a background line constitutively expressing RFP. (B) Ookinete maturation of full length ΔCBP-O and the individual domain mutants (ΔN-term, ΔCNB1-4/X and ΔCNB4) are comparable to wildtype. n = 3 independent experiments. (C) Oocyst numbers in midguts infected with ΔCBP-O (null), ΔN-term, ΔCNB1-4/X, ΔCNB4 and wild-type 12 days post-infection. A dot represents the number of oocysts in a single midgut (n = number of midguts dissected, unpaired t-test). Pie chart shows infection prevalence (Fischer’s exact test). (D) Immunofluorescence imaging of ookinetes and early oocysts in mosquito midguts infected with ΔCBP-O (null), ΔN-term, ΔCNB1-4/X, ΔCNB4 and wild-type 24 hours post-infection. Scale bar = 200μM. (E) Quantification of ookinetes/early oocysts in the midgut. n = 10 midguts per genotype. (F) Quantification of salivary gland sporozoites. Each data point is an average of five infected mosquitoes. In all relevant panels, statistical significance is indicated as: ns = not significant, * p = 0.05, ** = p < 0.01, *** = p < 0.001 and **** = p < 0.0001. Experiments in C, D, E and F were independently repeated twice, and data is shown from one representative experiment.
Given the intriguing localisation of CBP-O to the extreme apex of the ookinete, we asked whether the domain mutants could reach the ookinete tip. Live microscopy of ΔCNB1–4/X and immunofluorescence imaging of ΔCNB4 revealed that in the absence of the CNBs, localization of the protein to the apical tip of the ookinete is not disrupted (Figs 5 and S4A). However, upon removal of the N-terminal domain (ΔN-term), the truncated protein remained within the cell body and was no longer targeted to the tip (Fig 5). Immunofluorescence imaging of merozoites in mature schizonts revealed similar localisation features (S4B Fig). While we are not aware of a consensus sequence or structure determining protein trafficking to the apical tip, our results reveal that the N-terminal domain secures CBP-O localisation to the apex in both merozoites and ookinetes.
Live fluorescence microscopy with GFP-tagged CBP-O reveals requirement of the N-terminus for targeting CBP-O to the ookinete apex. The full-length and CNB domain mutant protein (ΔCNB1-4/X) maintain apical localisation. Scale bar = 5μM.
CNB domains of CBP-O reveal structural and sequence variation
Recognition of the secondary messengers, cGMP and cAMP, by cyclic nucleotide binding domains can transduce both internal and external signals to regulate diverse cellular responses [28]. Functional CNBs are commonly composed of a variable number of α-helices, the three major ones being αA, αB and αC, and an eight strand β-barrel structure that functions as a basket. The cyclic nucleotide binds at the base of the β-barrel protected from phosphodiesterases while the αC helix caps the bound cyclic nucleotide to increase binding affinity. The Phosphate Binding Cassette (PBC), the key structural motif responsible for binding to the cyclic nucleotide, comprises a short helical turn or P-helix and a loop between β6 and β7 [28].
To investigate whether the CNBs of PbCBP-O adopt the predicted architecture, their primary sequences and AlphaFold models (Fig 6A and 6B) were compared to the crystal structures of the most carboxy-terminal CNB domain (PfCNB-D) of P. falciparum cGMP-dependent protein kinase (PfPKG) and the CNB domains of P. falciparum cAMP dependent kinase regulatory subunit (PfPKA-R) [29,30]. Initial analysis suggests that the CNBs of PbCBP-O have an eight β-barrel structure with flanking helices as in PfCNB-D of PfPKG and PfCNB2 of the regulatory domain in PfPKA, with the PBC located inside the β-barrel and the capping αC helix (Fig 6B). However, the sequence alignment revealed a few key differences (Fig 6A). CNB3 shows a large insertion between β4 and β5, and CNBs 1–3 showed extra helices close to αC. Moreover, sequence analysis of CNB2 indicated that it lacks key PBC residues required for cGMP binding. Previous co-crystal structures of CNB domains from PfPKG bound with cGMP revealed a conserved glutamate and arginine within the PBC that mediate binding to the cyclic nucleotide [31]. These residues are conserved in CNB1, 3 and 4, but in CNB2 they have been replaced with a histidine (His2046) and a serine (Ser2051). Mammalian and Plasmodium PKGs also present a highly conserved threonine (PfCNB-D Thr494) following the arginine (PfCNB-D Arg493) in the PBC, whereas the PfPKA- R-CNB domains contain an alanine at the corresponding position. The Thr494 in PfPKG mediates high affinity interaction to the cyclic nucleotide, and when mutated there is a dramatic decrease in PKG activity [32,33]. CNB1, 3 and 4 of PbCBP-O present a threonine (Thr1813, Thr2237 and Thr3024, respectively) following the highly conserved arginine, suggesting similarity to PfCNB-D and that they may bind cGMP. On the other hand, CNB2 shows an isoleucine (Ile2056) at the corresponding position. Taken together, these residue differences suggest that CNB2 cannot bind cyclic nucleotides. Another determining feature of CNBs is a capping interaction to enclose the cyclic nucleotide in a hydrophobic pocket and shield it from solvent. In PfCNB-D, capping is mediated by Arg484 in the PBC and Gln532/Asp533 in the α-C helix [29]. In PfPKA-R the capping is mediated by a tyrosine residue (Tyr424) also located in the αC helix [30]. In PbCBP-O, the capping mechanism in CNB1,3 and 4 might be similar to PfPKA-R, due to an equivalently positioned tyrosine (Tyr1853, Tyr2275 and Tyr3062, respectively) in the αC helix. PbCBP-O includes a degenerate CNB (CNBX, located between CNB3 and CNB4) with a lower degree of conservation than the other four CNBs. While the Alphafold model shows an eight β-barrel structure with flanking helices as in other CNBs (Fig 6B), this domain lacks all key residues for binding cyclic nucleotide. From our structure-based analysis, we infer that CNB1, 3 and 4 of CBP-O contain a blend of features from CNB domains of PfPKA and PfPKG and could support cyclic nucleotide binding.
(A) Sequence alignment of the five CNBs of PbCBP-O with CNB-D of PfPKG and CNB2 of PfPKA-R illustrates that the domain organization is conserved. α-helices are shown in peach, β-barrels in green and the phosphate binding cleft (PBC) in magenta. Black arrows refer to the conserved glutamate (E), arginine (R) and threonine (T) in the PBC of PfCNB-D (E484, R493 and T494) and the PbCBP-O CNB1 (E1803, R1812, T1813), CNB3 (E2227, R2235, T2237) and CNB4 (E3014, R3023, T3024). The orange arrow marks the Tyrosine (Y) capping residue for PbCBP-O CNB1 (Y1853), CNB3 (Y2275) and CNB4 (Y3062) which is equivalent to the Tyrosine capping residue for PfPKA- R (Y424). Underlined residues highlight the conserved glycine triad. (B) Comparison of Alphafold model predictions of PbCBP-O-CNB1–4 and X to the crystal structure of PfCNB-D bound with cGMP (PDB ID: 4OFG) and CNB2 of PfPKA-R bound with cAMP (PDB ID: 5KBF). The PBC and αC helix of PfCNB-D/PfPKA-R -CNB2 are coloured in yellow and red respectively. The bound cGMP and cAMP are shown in sticks. A disordered insertion is predicted between β5 and β6 barrels of CNB3. Images were created in PyMOL. (C) cAMP and cGMP pull down assay with recombinantly expressed 6XHis tagged CNB-D from PbPKG and CNB1, 3 and 4 of PbCBP-O revealing that CNB3 binds both cAMP and cGMP. (D) Lysate from mature schizonts expressing GFP tagged ΔCNB1–4/X and ΔN-term isoforms of CBP-O were incubated with cAMP or cGMP agarose. Western blot analysis (α-GFP) revealed that upon removal of the N-terminal, the protein retains the capacity to bind cyclic nucleotides, while in the absence of CNB1–4/X, the mutant protein is unable to associate with cAMP or cGMP. n = 4 and 3 independent replicates for (C) and (D) respectively, and data is shown from a single representative experiment.
CNBs exhibit binding to both cAMP and cGMP
The sequence alignment predicts that PbCBP-O contains three functional CNB domains. However, as the domains displayed features characteristic of both PKA and PKG, cyclic nucleotide specificity could not be confidently predicted by structural modelling. To examine their binding capabilities, 6X his-tagged CNB1, CNB3 and CNB4 were recombinantly expressed in E. coli and CNB-D from PbPKG was used as control. PbCNB2 and PbCNBX was excluded from our analysis since they presented features of a degenerate CNB domain incapable of cyclic nucleotide binding. Recombinant PbPKG-D, PbCNB1 and PbCNB3 were easily recovered from the soluble fraction, but PbCNB4 aggregated into inclusion bodies (S5 Fig). To recover PbCNB4, inclusion bodies were lysed in 8M urea followed by buffer exchange through dialysis to reconstitute the protein in PBS.
To analyse cyclic nucleotide binding properties, recombinant protein forms of PbPKG-D, PbCBP-O’s CNB 1,3 and 4 were incubated with agarose beads linked to either cAMP or cGMP. Cyclic nucleotide bound proteins were recovered by heating with 1X SDS loading buffer and analysed by western blotting. As expected PbPKG-D showed strong binding to cGMP, but also displayed affinity to cAMP beads. This is not surprising since PfPKG-D is known to demonstrate low affinity binding to cAMP with a Kd = 7.4μM [29] and cAMP ligand density of the affinity matrix is 6 mM. While CNB1 and CNB4 from CBP-O lacked the ability to bind either cAMP or cGMP, CNB3 showed affinity for both cAMP and cGMP (Fig 6C). The results obtained for PbCNB4 were surprising as the domain is predicted to have higher conformational conservation compared to CNB1, 2 and 3. However, purification of PbCNB4 from inclusion bodies might have failed to reconstitute the correct folding, and therefore lacked binding. Cyclic nucleotide binding capacity of PbCNB3 is intriguing and demonstrates that the large insertion between β4 and β5 does not prevent binding to either cAMP or cGMP. Finally, to investigate whether the CNB domains are functional in vivo, we took advantage of the ΔCNB1–4/X and ΔN-term mutant lines and utilised schizonts containing mature merozoites that permitted larger preparative cultures and greater number of cells expressing CBP-O compared to ookinetes (~14n vs 1n). Protein lysates from schizont parasites expressing ΔCNB1–4/X and ΔN-term isoforms were incubated with cAMP or cGMP agarose and examined for binding capacity. While CNB1–4/X removal prevented protein binding to cAMP and cGMP, the N-term mutant retained cyclic nucleotide binding activity (Fig 6D).
Taken together these results demonstrate that the C-terminus of PbCBP-O is indeed a cyclic nucleotide binding domain and at least one of the predicted CNBs, CNB3 can directly interact with both cAMP and cGMP.
Discussion
PKA and PKG are the only well-characterised cyclic nucleotide effectors in Plasmodium and show high specificity to cAMP and cGMP respectively [2,3,34]. However, a third putative cyclic nucleotide binding protein is encoded in the Plasmodium genome, whose function remained elusive. In this study, we show that CBP-O is critical for P. berghei ookinetes to invade the mosquito midgut and transmit into the vector.
Domain analysis of PbCBP-O revealed the presence of four putative cyclic nucleotide binding (CNB) domains at the C-terminus of the protein; it also highlighted the absence of recognisable guanine nucleotide exchange factors. PbCBP-O lacks the domain necessary to catalyse the exchange of inactive GDP for GTP, nor does the protein contain the RA domains necessary for binding to RAP GTPases. Therefore, we propose PbCBP-O is not a homologue of the human guanine exchange factor, EPAC, which concurs with previous findings by Patel, Perrin and others for the P. falciparum homologue [9]. To gain insight into function of CBP-O, we utilised P. berghei parasites and deleted the complete ORF of the gene. Similar to observations in P. falciparum [9], we detected no impact on the asexual life cycle. It is interesting to note that the cbp-o gene has been reported to contain loss-of-function mutations in long-term culture adapted strains, such as P. falciparum 3D7 or HB3 [7]. Curiously, culturing of single or multiple genotype P. falciparum clinical isolates over several months also led to the emergence of loss-of-function mutations [8,35]. In light of our findings that PbCBP-O promotes parasite midgut invasion, it would be useful to assess the mutational status of PfCBP-O in culture adapted P. falciparum strains routinely used for mosquito transmission studies.
Notably, we show PbCBP-O is important for transmission into the mosquito and when the protein is absent, ookinetes reach maturity but fail to traverse the mosquito midgut epithelium effectively. Recent studies revealed that the conoid is conserved in Plasmodium [10,36]. Even when molecular composition of the conoid is maintained in different Plasmodium zoites, essentiality of components could be dependent on the different host environments and tissues encountered by the zoites. Accordingly, although CBP-O is expressed at the tip of merozoites and ookinetes, we observed invasion defects only in ookinetes. Ookinetes encounter a hostile gut environment and to ensure survival, they might require higher invasion competence to penetrate and cross the mosquito gut epithelium.
Our in vitro and in vivo studies demonstrated that the putative cyclic nucleotide binding domains of PbCBP-O are indeed functional and bind to both cAMP and cGMP. Interplay of cAMP and cGMP signalling in the related Apicomplexan, T. gondi, regulates parasite egress where PKA activity prevents premature egress of tachyzoites [37,38]. Moreover, TgPKA is required for shutting down PKG dependent tachyzoite motility to allow replication, when the parasite encounters the intracellular environment of the host cell. Similar to Toxoplasma, P. falciparum PKG regulates parasite egress [39]. In contrast, PfPKA activity does not influence merozoite egress and is essential specifically for erythrocyte invasion [9]. While there are lineage and developmental stage-specific differences, it is clear that cyclic nucleotide signalling is crucial for zoite motility and invasion. Intriguingly, PbCBP-O is expressed at the extreme apex of the ookinete tip and, in close proximity lies the ookinete extrados site (OES) which is the localisation site of GCβ or the cGMP synthesizer [40]. Considering that PbCBP-O localises close to the site of cGMP production, it is tempting to speculate that CBP-O regulates transduction of cGMP signals by either sequestering and/or sensing the cyclic nucleotide. Until now the only cGMP effector studied in Plasmodium was PKG and, in ookinetes the kinase promotes gliding motility [6]. High cGMP level is sustained in ookinetes and activity of the cGMP hydrolyser, PDEδ is presumably required for rounding up after the ookinete reaches the basal lamina of the midgut [4,5]. PbCBP-O mutant ookinetes show normal movement implying that requirement for CBP-O is subsequent to initiation of gliding motility (Fig 3A; S1 - S4 Movies). This suggests that spatial and temporal cGMP activity powers both ookinete motility and penetration of the mosquito via distinct transducers. In Plasmodium ookinetes, microneme secretion is controlled by PKG to support gliding motility [41]. Microneme exocytosis in ookinetes also drives adhesion, and mosquito midgut traversal [22,42,43]. We did not observe defects in expression of several micronemal proteins, nor in chitinase secretion, a micronemal protein required for ookinetes to cross the peritrophic matrix (PM), the structure which encases the infected blood meal in the mosquito midgut [24,44]. Moreover CBP-O is not required for the asexual lifecycle, therefore a role in microneme expression or secretion is unlikely. However, micronemes show broad cargo diversity in the related Apicomplexan, T. gondi [45] and, further examination will determine if cohorts of ookinete-specific micronemes are CBP-O controlled.
Alternatively, cGMP action is restricted to regulating motility and cAMP signalling takes control during crossing of the epithelial cell barrier. The role of cAMP in ookinete biology remains unexplored. However, Adenyl cyclase β, a cAMP producer, and both the cAMP binding and catalytic subunit of protein kinase A (PKA-R and PKA-C) are expressed in mature ookinetes implying a functional cAMP signalling module in ookinetes [6,46]. The ability of CBP-O to bind cAMP reveals another putative player that could coordinate epithelial layer crossing by the ookinete in a cAMP dependent fashion. Further biochemical studies with isothermal titration calorimetry could determine the selectivity and specificity of CBP-O’s CNB domains to cGMP or cAMP. Nevertheless, our data reveals a temporal nature of cyclic nucleotide signalling with discrete roles in ookinete motility and penetration of mosquito tissue.
While investigating the individual contribution of the N- and C-terminus, we discovered the N-terminus is required for positioning of CBP-O to the ookinete apex and, the C-terminus ensures binding to cyclic nucleotides. Previously published bioinformatic analysis of CNB domain containing proteins in eukaryotic pathogens identified amino acid variations in the nucleotide binding pocket and significant number of insertions within the CNB domain [47]. The authors hypothesized that the divergences could facilitate pathogen specific protein-protein interactions or alter nucleotide binding properties. Here we provide the first experimental evidence that, despite the amino acid variations and an insert of >20 amino acids, CNB3 domain of CBP-O retains the capacity to bind cyclic nucleotides (Fig 6C). Our findings will encourage further structural and biochemical studies to determine nucleotide specificity of the divergent CNB domains of CBP-O. Experiments should explore whether the divergence allows CBP-O to interact with cGMP and cAMP simultaneously or sequentially and if CBP-O regulates transition of cyclic nucleotide signalling from cGMP to cAMP. In ookinetes this relay could connect initiation of gliding motility to zoite invasion of the mosquito midgut epithelium. Moreover, it will be interesting to investigate if the T. gondi orthologue performs a similar role by linking cGMP and cAMP signalling to regulate tachyzoite motility and host cell invasion.
In conclusion, our findings expand the repertoire of functional cyclic nucleotide effectors in Plasmodium and demonstrate that PbCBP-O presents affinity for both cAMP and cGMP. Moreover, CBP-O conservation throughout Apicomplexa will provide opportunities to mechanistically understand how the zoite apex in these parasites utilise cyclic nucleotide signalling for host cell interaction and invasion.
Materials and methods
Ethics statement
All animal work was performed in accordance with the UK Animals (Scientific Procedures) Act 1986 (amended in 2012) and with European Directive 2010/63/EU on the Protection of Animals Used for Scientific Purposes. All procedures were approved by the University of Edinburgh’s animal welfare and ethics board, and UK’s Home Office (project license: PPL P04ABDCAA6). P. berghei parasites were maintained in CD-1 mice (Charles River) weighing between 22–32 g.
Generation of targeting constructs, parasite transfection and diagnostic PCRs
Transgenic parasite lines were generated using the CRISPR-Cas9 system. For all PCR reactions Q5 polymerase (NEB) was used and plasmid assemblies were performed using the NEBuilder HiFi DNA Assembly (NEB) or T4 ligase (NEB).
Vector (named ENP119 here) expressing the gRNA from the PyU6 promoter, a cassette that contains the selectable markers hdhfr and yfcu and a humanised SpCas9 driven from the Pbeef1-α promoter was linearised with BsmBI to allow incorporation of the gRNA [48].
All homology arms ranged between 800 and 1200 base pairs. To generate Pb∆CBP-O parasites the 5’ homology arm was amplified using EU503 and EU504 and digested with HindIII and SacII. For the 3’ homology arm primers EU492 and EU493 were used and digested with KpnI and NheI. These fragments were inserted sequentially into ENP119, to which the gRNA (EU564) was previously inserted. A gfp cassette flanked by Pbeef1-α promoter and Pbcalmodulin 3’utr was ligated into the vector using KpnI and SacII restriction enzyme sites. The plasmid (ENP142) was transfected into the HPtbb and 1804cl1 parental lines.
All domain mutants and GFP tagged CBP-O were created in 1804cl1 parental line. To generate CBP-O(∆CNB4) parasites, primers EU704 and EU705 were employed to amplify the 5’ homologous arm and, gfp was amplified with EU751 and EU752 from plasmid pG548. The previously constructed ENP142 vector was digested with HindIII and KpnI to remove the 5’ homology arm and the GFP cassette used for generating ∆CBP-O lines, leaving only the 3’ homology arm. The new 5’ homology arm and gfp were then ligated into this vector. Subsequently, HindIII and EcoRI were employed to release the fragment containing the homology arms and the GFP tag which was ligated into ENP119 containing the gRNA primer EU701. To generate CBP-O(∆CNB1–4/X), the 5’ homologous arm was amplified with primers EU749 and EU750 from genomic DNA and the GFP tag was amplified with EU751 and EU752 from plasmid pG548 [49]. The gene targeting construct was assembled similarly to CBP-O(∆CNB4) except for the gRNA (EU700). To generate CBP-O(∆N-ter), primers EU884 and EU885 were used to amplify the 5’ homology arm and the 3’ homology region was amplified with primers EU888 and EU889, and the GFP tag was amplified from pG548 using EU886 and EU887. The pieces were ligated into vector ENP119 containing the gRNA primer EU883.
C-terminal GFP tagged CBP-O parasite lines were generated using the PL0031 vector using single crossover homologous recombination containing 1 kb of the cbp-o C-terminus amplified by using primers EU706 and EU707. The vector was linearised with BamHI prior to transfection.
Schizonts for transfection were obtained from culture, and selection of mutant parasites was performed as previously published [50]. Correct integration of the targeting plasmid and gene disruption/modification was verified by diagnostic PCRs with primers listed in S1 Table.
P. berghei infections and phenotyping
Mice were administered phenylhydrazine (12.5 mg/L) by intraperitoneal injection two days prior to infection. Infections were established by injecting 150–300 μl of cryopreserved stocks. Mice used for transmission experiments were not administered phenylhydrazine. Parasitaemia was monitored by Giemsa staining of thin smears from a drop of tail blood. Parasites were observed under a standard light microscope with a 100X oil objective.
For generating mature schizonts, infected red blood cells were added to schizont media (RPMI with 25 mM HEPES, 20% FBS, 10 mM sodium bicarbonate, 0.38 mM hypoxanthine and 100 U/ml penicillin, 100 μg/ml streptomycin) and cultured over night at 37°C shaking at 40 RPM. For enrichment of gametocytes, mice were provided sulfadiazine (30 mg/L) in drinking water when parasitaemia reached between 1–6%. 48 hours later, gametocytes were retrieved through cardiac puncture. For ookinete cultures, gametocyte enriched blood was placed in ookinete media (RPMI with 25 mM HEPES, 20% FBS 4 mM sodium bicarbonate, 100 µM Xanthurenic acid, pH 7.8). Ookinetes were cultured for 24 hours at 21°C and mature ookinetes were assessed by Giemsa smears. Ookinete conversion rates were calculated from Giemsa staining or immunofluorescence using FITC-P25 antibody (see immunofluorescence assay methods). Conversion rates were determined by comparing mature ookinete (MO) numbers to total number of round (RO), retort (RE) and mature ookinetes (
To analyse exflagellation rates, blood obtained from cardiac puncture was immediately placed in exflagellation media (RPMI with 25mM HEPES, 4mM Sodium bicarbonate, 100uM XA, pH 7.8) at 21°C and 10 μl were placed in a haemocytometer. After 15 minutes, exflagellation centres, defined as cell clumps, were counted using a 10X objective with a light microscope.
Transmissions and mosquito dissections
For transmissions, 100–300 female Anopheles stephensi mosquitoes were starved for 24 h and subsequently fed on anaesthetised, infected mice with a parasitaemia between 1–6%. Mosquitoes were dissected 6–14 days after transmissions to count oocyst burden in the mosquito midgut, or 19–21 days to remove the salivary glands. Fluorescent oocysts and salivary glands were observed using a Leica MZ 10F dissecting microscope, and images were taken with the Leica DFC3000 G camera and the Leica X software. Oocyst numbers were counted using ImageJ.
For sporozoite counts, salivary glands were dissected in 1X PBS 19–21 days after infection. Complete sets of salivary glands (2x3 lobes) from 5 mosquitoes were transferred to a microcentrifuge tube containing 100 μl of 1X PBS and were homogenized with a pestle for one minute. The microcentrifuge tube was spun for 3 minutes at 3500 g and the supernatant was removed leaving 15μl to resuspend the sporozoites. Sporozoite numbers were counted using a haemocytometer.
For transmission bite-backs, mosquitoes were allowed to feed on anaesthetised naive mice for 10 minutes between days 19–21 after their infection. Parasitaemia was monitored through Giemsa smears from day 3 until a parasitaemia of 3%, or if no parasites were observed twelve days after the bite, mice were culled.
For in vivo ookinete conversion assays, mosquitoes were infected with wild-type or ∆CBP-O parasites and 19 hours later midguts were isolated in 1X PBS. Midguts were cut lengthwise to remove the blood bolus which was transferred to a microcentrifuge tube. The solution was centrifuged at 1000 g for 2–3 minutes and the cell pellet was smeared in a glass slide. Ookinetes in the blood bolus were counted via Giemsa staining.
Ookinete motility
Ookinetes were purified on a MACS LD magnetic column (Miltenyi). The column was pre-wet with enriched PBS (1X PBS supplemented with 20mM glucose, 4mM NaHCO3 and 1% BSA). The ookinete culture was passed twice through a 27G needle to dissociate any ookinete clusters, and subsequently added to the column. The column was washed with 4 ml ePBS and ookinetes eluted by passing another 4ml of ePBS after removing the column from the magnet. An extra 2ml of ePBS were added and a plunger used to ensure all parasites were retrieved. Enriched ookinetes were resuspended 1:1 ratio with Matrigel. 10 μl of this solution were added to a glass slide and a coverslip was added and sealed with nail varnish. The samples were placed in the 21°C incubator for 30 minutes. Ookinetes were imaged in the Zeiss Imager.Z2 microscope fitted with the Prime BSI camera (Photometrics). The 40X objective was used to take sequential images every 20 seconds for 15 minutes and only ookinetes whose tracks were detectable for at least 5 minutes were considered for analysis. Images were processed with ImageJ using the MTrackJ plugin to retrieve speed values.
Ookinete invasion assay
Mosquitos were fed on infected mice and 19–24 h later midguts were retrieved as described above (see transmissions and mosquito dissections). Midguts were either directly imaged under the microscope or by immunofluorescence assays. For immunofluorescence, after removal of blood bolus, midguts were fixed in 4% formaldehyde (Fisher Scientific) with 0.075% glutheraldehyde (Sigma) for 1 hour at room temperature. Then midguts were washed two times in 1X PBS and blocked in 3% BSA/PBS for 1 hour. Parasites embedded in the midgut were stained using FITC-P25 antibody incubated overnight at 4°C. The next morning, midguts were washed twice with 1X PBS for 15 minutes and incubated in 1X PBS supplemented with 10 μM Hoechst (Fisher Scientific) for 15 minutes. Excess Hoechst was removed with two washes of 1X PBS for 15 minutes. Midguts were then dried and flattened against the glass slides and mounted in clear Fluoromount G mount (Southern Biotech) overnight using a 22mm coverslip. Images were acquired with the Leica DM 2000 LED microscope, fitted with the Leica EL6000 external light source and a 20X objective. Images were taken with the Leica DFC3000 G camera using the Leica X software. Z-stacks were taken with a 20X objective on the Zeiss Axio Observer microscope fitted and Prime BSI camera (Photometrics) with a depth of 1μm and between 20–30 slices to cover the whole midgut. Images were analysed using ImageJ.
For direct imaging of infected midguts, mosquitoes were infected with RFP fluorescent parasites. After mosquito infection and midgut dissection, midguts were incubated in 1X PBS with 10 μM Hoechst for 15 minutes. Midguts were then washed twice with 1X PBS, and subsequently dried and flattened by creating a lengthwise cut using a needle and placing a 22mm coverslip. Midguts were imaged as described above.
Immunofluorescence (IFAs) assays of schizonts and ookinetes and, live microscopy of ookinetes
For IFAs, 1ml of cultured ookinetes or mature schizonts were spun down at 2000 g to pellet cells, the supernatant was removed, and the cell pellet was smeared on a glass slide. Smears were fixed in in 4% formaldehyde (Fisher Scientific) with 0.075% glutheraldehyde (Sigma) in PBS for 10 minutes, washed three times in 1X PBS and permeabilised with 0.1% Triton X-100/PBS for 10 minutes at room temperature. Slides were blood with 3% BSA/PBS for 1 hour at room temperature. The primary antibody was diluted to a desired concentration in 3% BSA/PBS and slides were incubated with it overnight at 4°C. Slides were then washed three times in 1X PBS for 10 minutes to remove excess primary antibody. Secondary antibody was diluted (1:2000) in 3% BSA/PBS and applied to slides and incubated for 1 hour and 30 minutes at room temperature. Slides were washed two times in 1X PBS for 15 minutes. The slides were mounted using 30μl of Fluoromount-G with DAPI (Fisher Scientific) and a 40mm coverslip and left overnight to set. Images were acquired using a 100X oil objective on the Leica DM 2000 LED microscope, fitted with the Leica EL6000 external light source and images were taken with the Leica DFC3000 G camera using the Leica X software. Images were analysed on ImageJ.
For live imaging of ookinetes, 10 μl of ookinete culture was pipetted onto a glass slide and subsequently covered with a 44 mm coverslip. Ookinetes were imaged using a 100X oil objective on a Nikon Spinning Disk Confocal microscope fitted with a Prime BSI camera (Photometrics). Z-stacks of 15–20 slices were acquired with a depth of 0.3 μm. Images were analysed on ImageJ.
Microneme proteins and secretion assay
For determination of micronemal protein expression, ookinete cultures of wild-type and ∆CBP-O parasites were spun down at 400 g for 5 minutes and supernatant removed. Pellets of ookinetes generated from 106 gametocytes were resuspended in erythrocyte lysis buffer (1.5M NH4Cl, 0.1M KHCO3, 0.01M EDTA) and incubated on ice for 20 minutes. The solutions were then centrifuged at 2800 rpm at 4°C and washed with 20 ml of cold PBS and centrifuged again to pellet the parasites. Pellets were lysed in RIPA buffer (10 mM Tris-HCl pH 8.0, 140 mM NaCl, 1 mM EDTA, 1% Triton X- 100, 0.1% sodium deoxycholate, 0.1% SDS, 1mM DTT) complemented with protease inhibitor cocktail (Complete Mini EDTA free tablets, Roche). Clarified lysates were resuspended in Laemmli sample buffer and separated by SDS-PAGE on a 4–20% Mini-PROTEAN TGX stain-free gel (Bio-Rad). Samples were then transferred onto nitrocellulose membranes (Amersham Bioscience) and proteins detected though immunoblotting using 1:6000 rabbit anti-CTRP, 1:6000 rabbit anti-chitinase, 1:4000 anti-WARP and 1:1000 rabbit anti-actin diluted in 5% milk/PBST [51]; and secondary antibodies (1:4000 goat anti-rabbit in 5% milk/PBST). Results were visualised with ECL kit (Amersham). For microneme secretion assay, ookinetes (from a 2% gametocytemia culture with 70% ookinete conversion rate) were purified and on a LD50 magnetic column and incubated in PBS for 2 hours [51]. The cells were spun down at 700 × g and the supernatant collected, filtered through a 2 µm filter and concentrated on an Amicon-ultra centrifugal filter, 30 K cut-off (Millipore), mixed with Laemmli sample buffer and Chitinase secretion was analysed by western blotting as described above.
Ookinete injections into mosquito
Ookinetes were purified from in vitro cultures as described previously and resuspended in ookinete medium to achieve a concentration of 700 ookinetes/100 nl. The thorax of thirty A. stephensi mosquitoes (5–7 day old) were injected with 700 ookinetes each using glass capillary needles and the Nanojet III microinjector (Drummond Scientific). Salivary gland sporozoites were counted at 19dpi as described above.
Quantitative real-time PCR
Four biological replicates of A. stephensi mosquito infections were performed with P. berghei 1804cl1 and ∆CBP-O (in 1804cl1 background) parasites and uninfected control. Blood-engorged midguts were dissected 22–24 hours post-blood feed. Twenty-five mosquito midguts were pooled for each group and stored in RNAlater (Thermofisher) at -80°C. Total RNA was isolated from homogenised midguts using the miRVana isolation kit (Thermofisher), treated with TURBO DNAse (TURBO DNA-free, Thermofisher) and cDNA synthesised using SuperScript IV Reverse Transcriptase (Thermofisher) according to manufacturer’s protocols. Transcript expression was quantified using Power SYBR green (Applied Biosystems) and gene specific primers (S1 Table) on a QuantStudio 5 Real-Time PCR System (Applied Biosystems). SRPN6 gene expression was normalized against Asrps7 using the ∆∆Ct method.
Recombinant protein expression
Coding sequences of PKG-D, CNB1,3 and 4 of PbCBP-O were amplified from P. berghei ANKA genomic DNA using primer pairs EU831/EU832 for PKG-D, EU839/EU840 for CNB1, EU835/EU836 for CNB3 and EU837/EU838 for CNB4 (S1 Table). PKG-D, CNB3 and 4 encoding DNA segments were cloned into Xho/NdeI site and CNB1 into XhoI/NheI of pET28a (Novagen) allowing expression of the CNB domains with a N-terminal 6X-histadine tag.
The pET28a:CNB1 was transfected into BL21-CodonPlus-RIL (Novagen) competent cells E. coli cells and pET28a:PKG-D, pET28a:CNB3 and pET28a:CNB4 were transfected into Rosetta DE3 (Novagen) competent cells. Cells were grown overnight in terrific broth (TB) at 37°C and then diluted 1:100 and grown in TB until an OD600 of around 0.7. CNB1, CNB3 and PKG-D expression was induced by adding IPTG at 0.5mM for 5 hours at 30°C. CNB4 expression was induced by adding 0.1mM IPTG overnight at 21°C. Cells were spun at 4°C at 1800 rpm for 10 minutes to pellet the cells. The supernatant was removed and pellet stored at -80°C until further processing.
Solubility tests were performed by lysing bacterial pellets in 1X BugBuster
(Novagen) complemented with protease inhibitors (complete Mini EDTA free tablets, Roche) and 0.5 μl benzonase (Sigma). Lysis was performed at room temperature for 30 minutes and then spun at 14,000 rpm for 15 minutes. Supernatant was mixed with 6x Laemmli sample buffer and run on an SDS-polyacrylamide gel. CNB1, CNB3 and PbPKG-D were detected in the soluble fraction while CNB4 was found in the insoluble fraction.
To recover CNB4 protein, the bacterial pellet was lysed with 2ml of 1X BugBuster and protease inhibitors and 0.2 μl of Benzonase by rotating at 30 rpm for 30 minutes at room temperature. The mixture was centrifuged for 10 minutes at maximum speed (30,000 g) at 4°C and supernatant removed. The pellet was washed twice in 1ml of 1X BugBuster and then spun at maximum (30,000 g) speed for 5 minutes at 4°C. To the final pellet 2 ml of buffer A were added (1xPBS, 8M Urea, 1mM DTT, protease inhibitor cocktail and 0.2 μl of benzonase), and mixed by vortexing for 1–2mins. The suspension was mixed by axial rotation (30 rpm) for 120 mins at 4°C. The suspension was centrifuged at 30,000 g for 15 mins and the supernatant was isolated. The sample was then dialysed to remove urea and allow refolding of the protein. The lysate was added to the dialysis cassette which was immersed in 500 ml of cold buffer B (1xPBS, 1M Urea, 1mM DTT) and dialysed for 2 hours at 4°C when the buffer was exchanged for fresh buffer and dialysed for another 2 hours. The dialysis cassette was then immersed into buffer C (1X PBS, 1mM DTT) and dialysed overnight at 4°C. The dialysed sample was mixed with glycerol to a final concentration of 20% and stored in the freezer.
Cyclic nucleotide binding assay
Recombinant proteins: Soluble proteins were obtained by lysing bacterial pellets in 1.5 ml of buffer D (50 mM Hepes pH 7.4, 1 mM EDTA, 1% Nonidet P-40, 10 mM NaF, 0.1 mM sodium orthovanadate, 1 mM dithiotreitol) supplemented with protease inhibitor cocktail (Complete Mini EDTA free tablets, Roche). Samples were sonicated 3 times, 1.5 μl of Benzonase were added and sample were mixed by axial rotation at 4°C for 30 minutes. Samples were centrifuged for 10 minutes at 14,000 rpm at 4°C. CNB4 was obtained as described above and diluted to a 10% glycerol concentration. The cAMP and cGMP agarose gel matrix (8-AEA-cAMP-Agarose and 8-AET-cGMP-Agarose, BioLog) were diluted 1:1 in sterile PBS. For PKG-D, CNB1 and CNB3, matrix was washed in buffer D and for CNB4 gel was equilibrated in 1xPBS containing 10% of glycerol. For each assay 200 μl of the lysates or purified CNB4 were incubated with 100 μl of cAMP and cGMP agarose matrix for 2 hours and mixed by axial rotation (30 rpm) at 4°C. Matrix was spun down at 100 g for 30 seconds and the supernatant removed and kept as the flow through. Three washes were done with 1ml of buffer D or 1xPBS (for CNB4) rotating for 5 minutes. For elution, the washed matrix was resuspended in 1x Laemmli sample buffer and heated for 15 minutes at 75°C. Elutes were separated on a 15% SDS-polyacrylamide gel and transferred to nitrocellulose membrane. Membrane was blocked in 5% skim milk in PBST buffer for 1h at room temperature and incubated with rat α-His antibody (Clonetech) at 1:6000 dilution for 1 h at room temperature. Membrane was washed 3X15 min with PBST and incubated with HRP-conjugated α-rat IgG (CST) at 1:5000 dilution for 1h at room temperature. Membranes were washed three times in PBST followed by a final PBS wash and proteins visualized by ECL advanced kit (GE healthcare).
Endogenous CBP-Os: Pellets containing 1x 107 cells of GFP-tagged CBP-O(∆N-ter) or CBP-O(ΔCNB1–4/X) expressing parasite lines from mature schizont stages with visible segmented merozoites were lysed in buffer D for 30 minutes on ice, and subsequently spun at 4°C for 15 minutes at 14,000 rpm. A small fraction of the lysate was resuspended in 6X Laemmli sample buffer to keep as the input sample, and the rest of the lysate was used for the binding assay. For this, 50 μl of previously PBS diluted cAMP and cGMP agarose gel matrix were washed in 200 μl buffer D. Parasite lysate was then incubated with the washed matrix for 2 hours at 4°C mixing by axial rotation (30 rpm). The beads were then spun down at 100 g for 1–2 minutes and supernatant was removed and kept as the flow-through sample. The matrix was washed with 500 μl buffer D on a rotator for 5 minutes for a total of three times. The washed matrix was resuspended in 1X Laemmli sample buffer and heated at 75 °C for 15 minutes. Bound elutes were separated by SDS-PAGE on a 4–20% Mini-PROTEAN TGX stain-free gel (Bio-Rad). Western blot analysis was performed as previously described using rabbit anti-GFP antibody (Chromotek) at 1:1000 dilution and HRP- conjugated anti-rabbit IgG (CST).
AlphaFold structure predictions
The crystal structures of Plasmodium falciparum PKG-CNB-D or 4OFG [27] and PfPKA-R-CNB2 or 5KBF [30] were obtained from the RCSB Protein Data Bank (https://www.rcsb.org). The predicted structures of P. berghei CBP-O-CNB1–4 and CNBX were generated in the AlphaFold Protein Structure Database https://alphafold.ebi.ac.uk [52]. Structure predictions for P. berghei CBP-O-CNB1–4 and CNBX were generated using the ColabFold default MSA pipeline [53], monomer model and three cycles in the Google Colaboratory. The generated structures were annotated in PyMOL.
Supporting information
S1 Fig. Schematic representation of strategy to generate PbCBP-PO transgenic lines.
(A) CBP-O was C- terminally tagged with GFP employing a single crossover strategy where the gene was linearised using a BamHI site. (B) Full-length knock out was generated using CRISPR/Cas9 with a guide RNA mapping to the middle of the gene. The locus was replaced by a constitutively active GFP. Positions of oligonucleotides used for genotyping are indicated by arrows and identified by specific EU numbers. Genotyping by the respective diagnostic PCRs and expected sizes are indicated by agarose gel electrophoresis and a table respectively.
https://doi.org/10.1371/journal.ppat.1013467.s001
(TIF)
S2 Fig. Schematic representation of strategy to generate PbCBP-PO mutant lines.
All mutant lines were created in constitutively expressing RFP line. (A) Full-length knock out (Δcbp-o) was generated using the strategy as represented in SF1A. (B) cbp-o(Δcnb1–4/x), (C) cbp-o(ΔN-term) and (D) cbp-o(Δcnb4) transgenic lines were generated by replacing the indicated loci with gfp resulting in a GFP-tagged domain mutant. Positions of oligonucleotides used for genotyping are indicated by arrows and denoted by specific EU numbers. Genotyping by diagnostic PCRs and expected sizes are indicated.
https://doi.org/10.1371/journal.ppat.1013467.s002
(TIF)
S3 Fig. Proportion of ookinetes displaying left-handed helical motion.
Ookinetes showing motility for at least 5 minutes were examined. n = 46 for each genotype and data is pooled from two independent replicates.
https://doi.org/10.1371/journal.ppat.1013467.s003
(TIF)
S4 Fig. Contribution of N- and C-terminal domains on protein localisation.
(A) Deletion of CNB4 does not impact protein localisation to the apical tip of ookinetes. Immunofluorescence microscopy of mature ookinete expressing GFP-tagged CBP-O protein lacking the terminal CNB4 domain. (B) N-terminus is required for targeting CBP-O to the merozoite tip. The full-length and CNB domain mutant protein (ΔCNB1–4/X) maintain apical localisation. Immunofluorescence microscopy of mature schizonts expressing the various domain mutant proteins.
https://doi.org/10.1371/journal.ppat.1013467.s004
(TIF)
S5 Fig. Recombinant protein expression of PbCBP-O’s cyclic nucleotide binding domains.
Western blot analysis of His-tagged recombinant CNB1, CNB3 and CNB4 proteins. CNB1 and 3 were isolated from solution fractions after lysis under non-denaturing conditions. CNB4 was isolated from inclusion bodies in 8M Urea buffer, followed by refolding dialysis.
https://doi.org/10.1371/journal.ppat.1013467.s005
(TIF)
S1 Movie. Motion of wildtype ookinetes (over 15 mins) in Matrigel tracked using MTrackJ plugin.
https://doi.org/10.1371/journal.ppat.1013467.s006
(AVI)
S2 Movie. Motion of PbΔCBPO ookinetes (over 15 mins) in Matrigel tracked using MTrackJ plugin.
https://doi.org/10.1371/journal.ppat.1013467.s007
(AVI)
S3 Movie. Magnification displaying left-handed helical motion of wildtype ookinetes.
https://doi.org/10.1371/journal.ppat.1013467.s008
(AVI)
S4 Movie. Magnification displaying left-handed helical motion of PbΔCBPO ookinetes.
https://doi.org/10.1371/journal.ppat.1013467.s009
(AVI)
S1 Table. Sequences of oligonucleotides used in the study.
https://doi.org/10.1371/journal.ppat.1013467.s010
(DOCX)
S2 Table. Values used to build graphs and statistical analysis.
https://doi.org/10.1371/journal.ppat.1013467.s011
(XLSX)
Acknowledgments
We thank the Centre Optical Instrumentation Laboratory (COIL) at the University of Edinburgh for training and use of their microscopes, Ronnie Mooney for providing mosquitoes, Dina Vlachou and Julia Cai (Imperial College) for helpful discussions and protocols for mosquito injections. We acknowledge Vector and Eukaryotic Pathogen Database (VEuPathDB) for their invaluable support for this study.
References
- 1. Abossie A, Yohanes T, Nedu A, Tafesse W, Damitie M. Prevalence of Malaria and Associated Risk Factors Among Febrile Children Under Five Years: A Cross-Sectional Study in Arba Minch Zuria District, South Ethiopia. Infect Drug Resist. 2020;13:363–72. pmid:32104008
- 2. Baker DA, Drought LG, Flueck C, Nofal SD, Patel A, Penzo M, et al. Cyclic nucleotide signalling in malaria parasites. Open Biol. 2017;7(12):170213. pmid:29263246
- 3. Perrin AJ, Patel A, Flueck C, Blackman MJ, Baker DA. cAMP signalling and its role in host cell invasion by malaria parasites. Curr Opin Microbiol. 2020;58:69–74. pmid:33032143
- 4. Hirai M, Arai M, Kawai S, Matsuoka H. PbGCbeta is essential for Plasmodium ookinete motility to invade midgut cell and for successful completion of parasite life cycle in mosquitoes. J Biochem. 2006;140(5):747–57.
- 5. Moon RW, Taylor CJ, Bex C, Schepers R, Goulding D, Janse CJ, et al. A cyclic GMP signalling module that regulates gliding motility in a malaria parasite. PLoS Pathog. 2009;5(9):e1000599. pmid:19779564
- 6. Brochet M, Collins MO, Smith TK, Thompson E, Sebastian S, Volkmann K, et al. Phosphoinositide metabolism links cGMP-dependent protein kinase G to essential Ca²⁺ signals at key decision points in the life cycle of malaria parasites. PLoS Biol. 2014;12(3):e1001806. pmid:24594931
- 7. Claessens A, Affara M, Assefa SA, Kwiatkowski DP, Conway DJ. Culture adaptation of malaria parasites selects for convergent loss-of-function mutants. Sci Rep. 2017;7:41303. pmid:28117431
- 8. Claessens A, Stewart LB, Drury E, Ahouidi AD, Amambua-Ngwa A, Diakite M, et al. Genomic variation during culture adaptation of genetically complex Plasmodium falciparum clinical isolates. Microb Genom. 2023;9(5):mgen001009. pmid:37204422
- 9. Patel A, Perrin AJ, Flynn HR, Bisson C, Withers-Martinez C, Treeck M, et al. Cyclic AMP signalling controls key components of malaria parasite host cell invasion machinery. PLoS Biol. 2019;17(5):e3000264. pmid:31075098
- 10. Koreny L, Zeeshan M, Barylyuk K, Tromer EC, van Hooff JJE, Brady D. Molecular characterization of the conoid complex in Toxoplasma reveals its conservation in all apicomplexans, including Plasmodium species. PLoS Biol. 2021;19(3):e3001081.
- 11. Levine ND. Progress in taxonomy of the Apicomplexan protozoa. J Protozool. 1988;35(4):518–20. pmid:3143826
- 12. Blackman MJ, Bannister LH. Apical organelles of Apicomplexa: biology and isolation by subcellular fractionation. Mol Biochem Parasitol. 2001;117(1):11–25. pmid:11551628
- 13. Cova MM, Lamarque MH, Lebrun M. How Apicomplexa Parasites Secrete and Build Their Invasion Machinery. Annu Rev Microbiol. 2022;76:619–40. pmid:35671531
- 14. Hu K, Johnson J, Florens L, Fraunholz M, Suravajjala S, DiLullo C. Cytoskeletal components of an invasion machine--the apical complex of Toxoplasma gondii. PLoS Pathog. 2006;2(2):e13.
- 15. Qian P, Wang X, Guan C, Fang X, Cai M, Zhong C-Q, et al. Apical anchorage and stabilization of subpellicular microtubules by apical polar ring ensures Plasmodium ookinete infection in mosquito. Nat Commun. 2022;13(1):7465. pmid:36463257
- 16. Dawn A, Singh S, More KR, Siddiqui FA, Pachikara N, Ramdani G, et al. The central role of cAMP in regulating Plasmodium falciparum merozoite invasion of human erythrocytes. PLoS Pathog. 2014;10(12):e1004520.
- 17. de Rooij J, Zwartkruis FJT, Verheijen MHG, Cool RH, Nijman SMB, Wittinghofer A. Epac is a Rap1 guanine-nucleotide-exchange factor directly activated by cyclic AMP. Nature. 1998;396(6710):474–7.
- 18. Lee K. Epac: new emerging cAMP-binding protein. BMB Rep. 2021;54(3):149–56. pmid:33298248
- 19. Dos Santos Pacheco N, Tosetti N, Koreny L, Waller RF, Soldati-Favre D. Evolution, Composition, Assembly, and Function of the Conoid in Apicomplexa. Trends Parasitol. 2020;36(8):688–704. pmid:32487504
- 20. Siden-Kiamos I, Ecker A, Nybäck S, Louis C, Sinden RE, Billker O. Plasmodium berghei calcium-dependent protein kinase 3 is required for ookinete gliding motility and mosquito midgut invasion. Mol Microbiol. 2006;60(6):1355–63. pmid:16796674
- 21. Vlachou D, Zimmermann T, Cantera R, Janse CJ, Waters AP, Kafatos FC. Real-time, in vivo analysis of malaria ookinete locomotion and mosquito midgut invasion. Cell Microbiol. 2004;6(7):671–85. pmid:15186403
- 22. Dessens JT, Beetsma AL, Dimopoulos G, Wengelnik K, Crisanti A, Kafatos FC, et al. CTRP is essential for mosquito infection by malaria ookinetes. EMBO J. 1999;18(22):6221–7. pmid:10562534
- 23. Kadota K, Ishino T, Matsuyama T, Chinzei Y, Yuda M. Essential role of membrane-attack protein in malarial transmission to mosquito host. Proc Natl Acad Sci U S A. 2004;101(46):16310–5.
- 24. Tsai YL, Hayward RE, Langer RC, Fidock DA, Vinetz JM. Disruption of Plasmodium falciparum chitinase markedly impairs parasite invasion of mosquito midgut. Infection and Immunity. 2001;69(6):4048–54.
- 25. Yuda M, Yano K, Tsuboi T, Torii M, Chinzei Y. von Willebrand factor A domain-related protein, a novel microneme protein of the malaria ookinete highly conserved throughout Plasmodium parasites. Molecular and Biochemical Parasitology. 2001;116(1):65–72.
- 26. Abraham EG, Pinto SB, Ghosh A, Vanlandingham DL, Budd A, Higgs S. An immune-responsive serpin, SRPN6, mediates mosquito defense against malaria parasites. Proc Natl Acad Sci U S A. 2005;102(45):16327–32.
- 27. Lin J, Annoura T, Sajid M, Chevalley-Maurel S, Ramesar J, Klop O, et al. A novel “gene insertion/marker out” (GIMO) method for transgene expression and gene complementation in rodent malaria parasites. PLoS One. 2011;6(12):e29289. pmid:22216235
- 28. Berman HM, Ten Eyck LF, Goodsell DS, Haste NM, Kornev A, Taylor SS. The cAMP binding domain: An ancient signaling module. Proceedings of the National Academy of Sciences. 2005;102(1):45–50.
- 29. Kim JJ, Flueck C, Franz E, Sanabria-Figueroa E, Thompson E, Lorenz R. Crystal structures of the carboxyl cGMP binding domain of the Plasmodium falciparum cGMP-dependent protein kinase reveal a novel capping triad crucial for merozoite egress. PLoS Pathog. 2015;11(2):e1004639.
- 30. Littler DR, Bullen HE, Harvey KL, Beddoe T, Crabb BS, Rossjohn J, et al. Disrupting the Allosteric Interaction between the Plasmodium falciparum cAMP-dependent Kinase and Its Regulatory Subunit. J Biol Chem. 2016;291(49):25375–86. pmid:27738107
- 31. El Bakkouri M, Kouidmi I, Wernimont AK, Amani M, Hutchinson A, Loppnau P, et al. Structures of the cGMP-dependent protein kinase in malaria parasites reveal a unique structural relay mechanism for activation. Proc Natl Acad Sci U S A. 2019;116(28):14164–73. pmid:31239348
- 32. Altenhofen W, Ludwig J, Eismann E, Kraus W, Bönigk W, Kaupp UB. Control of ligand specificity in cyclic nucleotide-gated channels from rod photoreceptors and olfactory epithelium. Proc Natl Acad Sci U S A. 1991;88(21):9868–72.
- 33. Weber IT, Shabb JB, Corbin JD. Predicted structures of the cGMP binding domains of the cGMP-dependent protein kinase: a key alanine/threonine difference in evolutionary divergence of cAMP and cGMP binding sites. Biochemistry. 1989;28(14):6122–7. pmid:2550070
- 34. Lasonder E, More K, Singh S, Haidar M, Bertinetti D, Kennedy EJ, et al. cAMP-Dependent Signaling Pathways as Potential Targets for Inhibition of Plasmodium falciparum Blood Stages. Front Microbiol. 2021;12:684005. pmid:34108954
- 35. Stewart LB, Diaz-Ingelmo O, Claessens A, Abugri J, Pearson RD, Goncalves S, et al. Intrinsic multiplication rate variation and plasticity of human blood stage malaria parasites. Commun Biol. 2020;3(1):624. pmid:33116247
- 36. Bertiaux E, Balestra AC, Bournonville L, Louvel V, Maco B, Soldati-Favre D, et al. Expansion microscopy provides new insights into the cytoskeleton of malaria parasites including the conservation of a conoid. PLoS Biol. 2021;19(3):e3001020. pmid:33705377
- 37. Jia Y, Marq J-B, Bisio H, Jacot D, Mueller C, Yu L, et al. Crosstalk between PKA and PKG controls pH-dependent host cell egress of Toxoplasma gondii. EMBO J. 2017;36(21):3250–67. pmid:29030485
- 38. Uboldi AD, Wilde M-L, McRae EA, Stewart RJ, Dagley LF, Yang L, et al. Protein kinase A negatively regulates Ca2+ signalling in Toxoplasma gondii. PLoS Biol. 2018;16(9):e2005642. pmid:30208022
- 39. Collins CR, Hackett F, Strath M, Penzo M, Withers-Martinez C, Baker DA, et al. Malaria parasite cGMP-dependent protein kinase regulates blood stage merozoite secretory organelle discharge and egress. PLoS Pathog. 2013;9(5):e1003344. pmid:23675297
- 40. Gao H, Yang Z, Wang X, Qian P, Hong R, Chen X, et al. ISP1-Anchored Polarization of GCβ/CDC50A Complex Initiates Malaria Ookinete Gliding Motility. Curr Biol. 2018;28(17):2763-2776.e6. pmid:30146157
- 41. Fang H, Gomes AR, Klages N, Pino P, Maco B, Walker EM, et al. Epistasis studies reveal redundancy among calcium-dependent protein kinases in motility and invasion of malaria parasites. Nat Commun. 2018;9(1):4248. pmid:30315162
- 42. Dessens JT, Sidén-Kiamos I, Mendoza J, Mahairaki V, Khater E, Vlachou D, et al. SOAP, a novel malaria ookinete protein involved in mosquito midgut invasion and oocyst development. Mol Microbiol. 2003;49(2):319–29. pmid:12828632
- 43. Kariu T, Ishino T, Yano K, Chinzei Y, Yuda M. CelTOS, a novel malarial protein that mediates transmission to mosquito and vertebrate hosts. Mol Microbiol. 2006;59(5):1369–79. pmid:16468982
- 44. Dessens JT, Mendoza J, Claudianos C, Vinetz JM, Khater E, Hassard S. Knockout of the rodent malaria parasite chitinase pbCHT1 reduces infectivity to mosquitoes. Infect Immun. 2001;69(6):4041–7.
- 45. Kremer K, Kamin D, Rittweger E, Wilkes J, Flammer H, Mahler S, et al. An overexpression screen of Toxoplasma gondii Rab-GTPases reveals distinct transport routes to the micronemes. PLoS Pathog. 2013;9(3):e1003213. pmid:23505371
- 46. Lal K, Prieto JH, Bromley E, Sanderson SJ, Yates JR 3rd, Wastling JM, et al. Characterisation of Plasmodium invasive organelles; an ookinete microneme proteome. Proteomics. 2009;9(5):1142–51. pmid:19206106
- 47. Mohanty S, Kennedy EJ, Herberg FW, Hui R, Taylor SS, Langsley G, et al. Structural and evolutionary divergence of cyclic nucleotide binding domains in eukaryotic pathogens: Implications for drug design. Biochim Biophys Acta. 2015;1854(10 Pt B):1575–85. pmid:25847873
- 48. Zhang C, Gao H, Yang Z, Jiang Y, Li Z, Wang X, et al. CRISPR/Cas9 mediated sequential editing of genes critical for ookinete motility in Plasmodium yoelii. Mol Biochem Parasitol. 2017;212:1–8. pmid:28034675
- 49. Philip N, Waters AP. Conditional Degradation of Plasmodium Calcineurin Reveals Functions in Parasite Colonization of both Host and Vector. Cell Host Microbe. 2015;18(1):122–31. pmid:26118994
- 50. Janse CJ, Ramesar J, Waters AP. High-efficiency transfection and drug selection of genetically transformed blood stages of the rodent malaria parasite Plasmodium berghei. Nat Protoc. 2006;1(1):346–56. pmid:17406255
- 51. Philip N, Vaikkinen HJ, Tetley L, Waters AP. A unique Kelch domain phosphatase in Plasmodium regulates ookinete morphology, motility and invasion. PLoS One. 2012;7(9):e44617. pmid:22957089
- 52. Jumper J, Evans R, Pritzel A, Green T, Figurnov M, Ronneberger O, et al. Highly accurate protein structure prediction with AlphaFold. Nature. 2021;596(7873):583–9. pmid:34265844
- 53. Mirdita M, Schütze K, Moriwaki Y, Heo L, Ovchinnikov S, Steinegger M. ColabFold: making protein folding accessible to all. Nat Methods. 2022;19(6):679–82. pmid:35637307