Cyclic AMP signalling controls key components of malaria parasite host cell invasion machinery

Abstract

Cyclic AMP (cAMP) is an important signalling molecule across evolution, but its role in malaria parasites is poorly understood. We have investigated the role of cAMP in asexual blood stage development of Plasmodium falciparum through conditional disruption of adenylyl cyclase beta (ACβ) and its downstream effector, cAMP-dependent protein kinase (PKA). We show that both production of cAMP and activity of PKA are critical for erythrocyte invasion, whilst key developmental steps that precede invasion still take place in the absence of cAMP-dependent signalling. We also show that another parasite protein with putative cyclic nucleotide binding sites, Plasmodium falciparum EPAC (PfEpac), does not play an essential role in blood stages. We identify and quantify numerous sites, phosphorylation of which is dependent on cAMP signalling, and we provide mechanistic insight as to how cAMP-dependent phosphorylation of the cytoplasmic domain of the essential invasion adhesin apical membrane antigen 1 (AMA1) regulates erythrocyte invasion.

Fig 1. Conditional disruption of ACβ expression.

(A) Schematic representation of the Cas9-mediated creation of the ACβ-HA:loxP line in DiCre-expressing P. falciparum parasites and subsequent RAP-induced deletion of the modified gene. Double-headed arrows represent the regions amplified by PCR in (B). Red arrowheads represent loxP sites, lollipops represent translational stop codons, black boxes indicate the position of an open reading frame downstream of ACB and light blue boxes indicate regions of re-codonised sequence. (B) Diagnostic PCR verifying successful integration of both repair constructs and successful genetic excision upon treatment with RAP in an ACβ-HA:loxP clone. Early rings (0–4 h post invasion) were treated with RAP or vehicle only (DMSO), and genomic DNA from schizonts (approximately 40 h posttreatment) was used in these PCRs. (C) Western blots confirming HA tagging and RAP-induced ablation of ACβ expression in the ACβ-HA:loxP line. In addition to a signal of the predicted intact ACβ molecular mass, bands of lower molecular mass were also detected, likely due to proteolysis. Antibodies to the ER protein BiP (PF3D7_0917900) were used as a loading control. (D) IFA showing localisation of ACβ-HA₃ in close proximity to the rhoptry protein ARO (PF3D7_0414900) in schizonts, and ablation of expression by RAP treatment. Over 99% of all RAP-treated ACβ-HA:loxP schizonts examined by IFA were HA-negative in three independent experiments. Scale bar, 10 μm. ACβ, adenylyl cyclase beta; ARO, armadillo repeats only protein; BiP, binding immunoglobulin protein; Cas9, CRISPR-associated protein 9; DIC, differential interference contrast; DiCre, dimerisable Cre-recombinase; ER, endoplasmic reticulum; HA₃, triple hemagglutinin; hDHFR, human dihydrofolate reductase selectable marker; IFA, immunofluorescence assay; loxPint, loxP containing intron; RAP, rapamycin; RR, recodonised-region; SERA2, serine repeat antigen 2; sgRNA, single guide RNA.

Fig 2. Conditional disruption of PKAc expression.

(A) Schematic representation of the SLI strategy used to produce the PKAc-HA:loxP line and RAP-induced disruption of the gene. Double-headed arrows represent the regions amplified by PCR in (B). Red arrowheads represent loxP sites, lollipops represent translational stop codons, and light blue boxes indicate regions of re-codonised sequence. glmS was not exploited in these experiments. (B) Diagnostic PCR analysis verifying successful SLI to produce the PKAc-HA:loxP line and successful excision of floxed sequences upon treatment with RAP. Rings (about 20 h post invasion) were RAP or DMSO treated for 2 h, and genomic DNA from schizonts (about 20 h post-treatment) was used in these PCRs. (C) Western blots showing expression (DMSO) and ablation (RAP) of PKAc-HA₃ in PKAc-HA:loxP parasites. Expression of GAPDH (PF3D7_1462800) is shown as a loading control. (D) IFA showing the diffuse localisation of PKAc-HA₃ (DMSO) and the loss of expression upon RAP treatment. Over 99% of all RAP-treated PKAc-HA:loxP schizonts examined by IFA were HA-negative in three independent experiments. (E) Electron micrograph of a segmented RAP-treated PKAc-HA:loxP schizont from high-pressure frozen, freeze-substituted plastic sections. Inset: image of an entire PKAc-HA:loxP schizont showing the typical morphology of a mature schizont prior to PVM rupture. Main image: a more detailed view of two of the merozoites within the schizont. Scale bar, 500 nm. AmpR, ampicillin resistance cassette used for plasmid selection in bacteria; DIC, differential interference contrast; DiCre, dimerisable Cre-recombinase; FV, food vacuole; GAPDH, glyceraldehyde 3-phosphate dehydrogenase; GAP45, glideosome-associated protein 45; glmS, glucosamine-6-phosphate riboswitch ribozyme; HA₃, triple hemagglutinin; hDHFR, human dihydrofolate reductase selectable marker; IFA, immunofluorescence assay; IMC, inner membrane complex; loxPint, loxP containing intron; Myc, c-myc tag; N, nucleus; NeoR, neomycin resistance selectable marker; PKAc, catalytic subunit of cAMP-dependent protein kinase; PPM, parasite plasma membrane; PVM, parasitophorous vacuole membrane; R, rhoptries; RAP, rapamycin; RBC, red blood cell membrane; RR, recodonised-region; SERA2, serine repeat antigen 2 gene; SLI, selection-linked integration; T2A, thosea asigna virus 2A peptide.

Fig 3. ACβ and PKA are both essential for parasite proliferation.

(A) Giemsa-stained blood films showing ring-stage parasites following egress of DMSO-treated ACβ-HA:loxP and PKAc-HA:loxP parasites (left) and the absence of rings following egress of RAP-treated parasites. Scale bar, 5 μm. (B) Growth curves showing changes in parasitaemia of ACβ-HA:loxP and PKAc-HA:loxP parasites treated with DMSO (vehicle only control) or RAP. Means from three replicates are plotted. Error bars, SD. (C) Schematic representation of the approach used to genetically complement the PKAc-HA:loxP line by Cas9-mediated introduction of a RAP-inducible, trimethoprim (TMP)-stabilised HA-tagged PKAc transgene at the p230p locus to create the PKAc-HA_DDDcomp:loxP line. Double-headed arrows represent the regions amplified by PCR in S4C Fig. Red arrowheads represent loxP sites, lollipops represent translational stop codons, and light blue boxes indicate regions of re-codonised sequence. (D) Western blots showing the RAP-inducible switch from expression of PKAc-HA₃ from the PKAc locus to expression of TMP-stabilised PKAc-HA₃_DDD from the P230p locus. Note the decreased mobility of the PKAc-HA₃_DDD resulting from its fusion to the DDD. GAPDH is shown as a loading control. Some lower molecular weight products are detected in the RAP TMP lane, likely due to incomplete stabilisation of all protein species. (E) Growth curve showing rescue of growth of PKAc-HA₃–deficient parasites by TMP-mediated stabilisation of PKAc-HA₃_DDD. Means from three replicates are plotted. Error bars, SD. Data associated with this figure can be found in the supplemental data file (S1 Data). ACβ, adenylyl cyclase beta; BsdR, blasticidin resistance selectable marker; Cas9, CRISPR associated protein 9; DDD, DHFR degradation domain; DHFR, dihydrofolate reductase; EGFP, enhanced green fluorescent protein; GAPDH, glyceraldehyde 3-phosphate dehydrogenase; HA₃, triple hemagglutinin; PKA, cAMP-dependent protein kinase; PKAc, catalytic subunit of cAMP-dependent protein kinase; RAP, rapamycin; TMP, trimethoprim.

Fig 4. PKA, cAMP, and Epac are not required for egress.

(A) Western blots data monitoring egress kinetics of DMSO- and RAP-treated PKAc-HA:loxP and ACβ-HA:loxP schizonts. The slower onset of detection of SERA5 p50 in RAP-treated ACβ-HA:loxP parasites indicates delayed or impaired egress in the absence of cAMP. Blots are representative of two biological repeats, which are both quantified in S4A Fig. (B) Quantification of the proportion of schizonts rupturing in 30-min videos of DMSO- and RAP-treated parasites. For each video, one parasite population (DMSO or RAP) was stained with Hoechst (indicated in blue on the plots). The p-values derive from paired t tests. (C) Quantification of the mean time taken for the DMSO- and RAP-treated parasites from (B) to progress to egress, as measured by visual analysis of the same video microscopy data shown in panel (B). The p-values derive from paired t tests. For all data in (B) and (C), each point is the mean for one population (DMSO or RAP) from a single video (50–100+ schizonts). Ten videos were quantified from at least three independent experiments. (D) Schematic representation of the selection-linked targeted homologous recombination-based approach used to disrupt the PfEpac gene. Lollipops represent translational stop codons. Validation of this line by PCR is shown in S1E Fig. (E) Western blot data indicating normal rupture of PfEpac-deficient schizonts. Data associated with this figure can be found in the supplemental data file (S1 Data). AmpR, ampicillin resistance cassette used for plasmid selection in bacteria; cAMP, cyclic AMP; Epac, exchange protein directly activated by cAMP; hDHFR, human dihydrofolate resistance selectable marker; KO, knockout; NeoR, neomycin resistance selectable marker; PKA, cAMP-dependent protein kinase; p50, processed 50 kDa form; RAP, rapamycin; SERA5, serine repeat antigen 5; T2A, thosea asigna virus 2A peptide.

Fig 5. Invasion is critically dependent on cAMP and PKAc, but calcium mobilisation and microneme secretion are not.

(A) Quantification of invasion, merozoite-induced erythrocyte surface deformation, and echinocytosis observed by video microscopy following rupture of DMSO- and RAP-treated PKAc-HA:loxP and ACβ-HA:loxP schizonts. At least 20 videos per condition were quantified. Statistical significance was assessed by t test; ns indicates not significant (p > 0.05), whereas **** indicates p < 0.0001. (B) Electron micrograph of a RAP-treated PKAc-HA:loxP merozoite attached to the surface of an erythrocyte from high-pressure frozen, freeze-substituted plastic sections. Left: apical attachment of a merozoite to the surface of the erythrocyte. Scale bar, 500 nm. Right: a more detailed view of the electron-dense attachment region (arrowed) showing the close association of the apical end of the parasite and the erythrocyte membrane. Scale bar, 100 nm. (C) Electron tomography of the attachment region between a RAP-treated PKAc-HA:loxP merozoite and the RBC surface. Image is a sum of 30 central sections from the tomogram. Arrows indicate a region of apparent thickening of the RBC membrane. Full-tilt series in S7 Movie. Scale bar, 250 nm. (D) Super-resolution immunofluorescence imaging of PKAc-HA:loxP merozoites attached to the RBC surface. For the DMSO-treated group, parasites’ medium contained 1 μM cyotchalasin D to arrest invasion at tight junction formation. Additional images are shown in S3 Fig. Scale bar, 2 μm. (E) Induction of calcium mobilisation using 100 μM zaprinast in synchronous Fluo-4–loaded late-stage schizonts assayed by fluorimetry. The signal was normalised to DMSO carrier (0% signal) and 20 μM A23187 ionophore (100% signal). Means from six technical replicates (three samples from two biological replicates) are plotted. Error bars, SD. (F) IFA showing re-localisation of AMA1 from micronemes to the merozoite periphery in DMSO- and RAP-treated PKAc-HA:loxP schizonts. Quantification of 100 imaged schizonts from three individual biological replicates indicated no significant difference between peripheral and punctate staining of AMA1 between the two treatments (DMSO 55.88% ± 2.25% punctate, 41.21% ± 2.23% peripheral, and RAP 55.86% ± 2.93% punctate, 41.14% ± 1.93% peripheral). IFA analysis was performed on highly synchronous cultures that were treated with 20 μM E64 approximately 44 h post invasion for approximately 4 h. Scale bars, 5 μm. Data associated with this figure can be found in the supplemental data file (S1 Data). AMA1, apical membrane antigen 1; cAMP, cyclic AMP; DIC, differential interference contrast; E64, cysteine protease inhibitor; IFA, immunofluorescence assay; M, merozoite; ns, not significant; PKAc, catalytic subunit of cAMP-dependent protein kinase; RAP, rapamycin; RBC, red blood cell.

Fig 6. Efficient surface shedding of AMA1 requires cAMP and PKA.

(A) Western blot of culture supernatants from a time course of egressing DMSO- and RAP-treated PKAc-HA:loxP parasite cultures. Progression of egress is indicated by detection of the SERA5 p50 fragment. Shedding of surface adhesins AMA1, EBA175, Rh2b, and MSP1 are monitored by their detection in supernatants using the indicated antibodies. For each western blot, a representative image from one of three independent experiments is shown. The full-length blots used to produce this figure are shown in S4B Fig. Densitometry analyses of three biological replicates are shown in S4C Fig and S4D Fig. No significant differences in EBA175 shedding were observed after one hour, but 4.2 ± 1.6-fold less AMA1 was shed in RAP- compared with DMSO-treated PKAc-HA:loxP parasites. Blots are representative of three biological repeats. (B) Western blots indicating the presence of SERA5 p50 and shed surface adhesins in supernatants from egressing DMSO- and RAP-treated ACβ-HA:loxP parasites. A single time point was used because the RAP-treated population are slower to egress. The full-length blots used to produce this figure are shown in S4B Fig. Blots shown are representative of two biological repeats. (C) Left: western blot of culture supernatants from a time course of egressing 3D7 parasites showing unaltered AMA1-shedding kinetics in the presence or absence of the invasion inhibitor cytochalasin D (1 μM). Right: Giemsa-stained blood films confirming the cytochalasin D–mediated block in invasion by an absence of ring-stage parasites in the treated cultures. Scale bar, 5 μm. AMA1, apical membrane antigen 1; cAMP, cyclic AMP; EBA175, erythrocyte binding antigen 175; PKA, cAMP-dependent protein kinase; p50, processed 50 kDa form; RAP, rapamycin; Rh2b, reticulocyte binding protein homologue 2b; SERA5, serine repeat antigen 5.

Fig 7. cAMP- and PKA-dependent phosphorylation of invasion-related proteins and induction of a structural change in the cytoplasmic tail of AMA1 by Ser₆₁₀ phosphorylation.

(A) S-curves representing the phosphosites detected by mass spectrometry in DMSO- versus RAP-treated ACβ-HA:loxP and PKAc-HA:loxP schizont/merozoite preparations. The most enriched site in each of the indicated invasion-related proteins is labelled. Data shown are from three technical triplicates and are representative of three biological repeats. (B) Volcano plots showing the changes in detection of phosphosites between DMSO- and RAP-treated ACβ-HA:loxP and PKAc-HA:loxP. The negative log₁₀ transform of the p-value–derived Welch-corrected t test comparing three DMSO- and three RAP-treated replicates is plotted against the log₂-transformed fold change in reporter ion intensity (DMSO/RAP). Significantly altered sites (p < 0.05) in the ACβ and PKAc proteins are indicated in magenta and blue, respectively. Data shown are from the three technical triplicates shown in panel (A) and are representative of three biological repeats. (C) Circular dichroism (CD) spectra of recombinant AMA1_cyt with and without treatment with recombinant mammalian PKA. (D) Overlay of a selected region of a 2D ¹H-¹⁵N HSQC NMR spectra of unphosphorylated AMA1_cyt (red) and AMA1_cyt _pSer₆₁₀ (blue). The residues are indicated below their positions on the AMA1_cyt_pSer₆₁₀ spectra. Phosphorylation induced changes in the amide chemical shifts of the targeted Ser₆₁₀ as well as surrounding residues. (E) Secondary structure prediction of the AMA1 cytoplasmic tail sequence (Ser₆₁₀ indicated in green) as calculated by TALOS+ and Chemical Shift Index compared with random coil based on the ¹H, ¹⁵N, ¹³C chemical shifts of AMA1_cyt _pSer₆₁₀. Extended conformation is presented as arrows and alpha-helical conformation as a cylinder. No secondary structure elements are predicted for unphosphorylated AMA1_cyt. Data associated with this figure can be found in the supplemental data file (S1 Data) and supporting table 1 (S1 Table). ACβ, adenylyl cyclase beta; AMA1, apical membrane antigen 1; AMA1_cyt, AMA1 cytosolic domain; ARO, armadillo repeats only protein; cAMP, cyclic AMP; CD, circular dichroism; HSQC, heteronuclear single quantum coherence; IMC, inner membrane complex; MyoA, myosin A; MyoE, myosin E; NMR, nuclear magnetic resonance; PKA, cAMP-dependent protein kinase; PKAc, catalytic subunit of cAMP-dependent protein kinase; PKAr, regulatory subunit of cAMP-dependent protein kinase; RAMA, rhoptry-associated membrane antigen; RAP, rapamycin; ROM4, rhomboid protease 4; RON2, rhoptry neck protein 2; TALOS+, prediction of protein backbone and sidechain torsion angles from NMR chemical shifts program.