The heat shock protein Hsc70-3 mediates an anti-apoptotic response critical for Plasmodium evasion of Anopheles gambiae immunity

ABSTRACT The mosquito immune system limits Plasmodium infection and malaria transmission. Plasmodium falciparum evades the mosquito defense response by expressing the surface protein P. falciparum P47 (Pfs47) that inhibits midgut epithelial nitration by interacting with the mosquito midgut P47 receptor (P47Rec). However, the mechanism by which P47Rec suppresses caspase-mediated nitration is unknown. Here, we show that epithelial invasion by Pfs47 knockout parasites is followed by an extrusion of cells undergoing caspase-mediated apoptosis, which triggers the release of hemocyte-derived microvesicles (HdMv’s), known to promote complement activation and ookinete lysis. In contrast, invasion by P. falciparum parasites that express Pfs47 accelerates the extrusion of damaged cells and disrupts caspase activation and HdMv release. The Anopheles gambiae heat shock protein 70 cognate 3 (Hsc70-3), the ortholog of Drosophila binding immunoglobulin protein, was identified as a molecular partner of P47Rec. Silencing of Hsc70-3 promotes lysis of Pfs47 wild-type parasites through a caspase S2-dependent mechanism. The interaction of P47Rec activates a Hsc70-3-mediated anti-apoptotic response that prevents caspase activation. The interaction of P47Rec with Hsc70-3 is necessary for P. falciparum to evade the mosquito early immune responses that target the ookinete stage. IMPORTANCE Malaria transmission by Anopheles gambiae mosquitoes is very effective, in part because the parasite expresses a surface protein called Pfs47 that allows it to evade the mosquito immune system. Here we investigate how this protein changes the response of mosquito midgut epithelial cells to invasion by the parasite. Pfs47 is known to interact with P47Rec, a mosquito midgut receptor. We found that Pf47Rec inhibits caspase-mediated apoptosis by interacting with the Hsc70-3. This disrupts nitration of midgut epithelial cells invaded by the parasite and the release of hemocyte-derived microvesicles, which are critical for effective activation of the mosquito complement system that eliminates the parasite.

Parasite development in the mosquito is complex and relies on a series of molec ular interactions between the parasite and its vector.Plasmodium gametes undergo fertilization in the midgut lumen and differentiate into motile ookinetes that actively traverse the midgut epithelium.Ookinetes differentiate into oocysts as they reach the basal lamina.Oocysts grow continuously and generate thousands of sporozoites that are released into the hemolymph 2-3 weeks later.Sporozoites that invade the salivary glands can be transmitted when the mosquito bites another person.
Studies using Plasmodium berghei, a murine malaria model, revealed that Anopheles gambiae mosquitoes mount an orchestrated and potent anti-plasmodial response that effectively kills most ookinetes by activating the mosquito complement system (3,4).Ookinete invasion damages mosquito midgut cells, triggering a caspase-mediated apoptotic cell death response (5).The induction of caspase activity is mediated by the c-Jun N-terminal kinase (JNK) signaling pathway and triggers the expression of NADPH oxidase 5 (NOX5) and heme peroxidase 2 (HPX2), two enzymes that potentiate the nitric oxide synthase-mediated epithelial nitration response to ookinete invasion (4).Ookinetes also disrupt the peritrophic matrix, a protective layer of chitin polymers and proteins that surrounds the blood bolus, allowing direct contact between the gut microbiota and the midgut epithelium (6,7).Bacterial immune elicitors trigger the synthesis and systemic release of prostaglandins by gut epithelial cells, which attracts hemocytes to the basal surface of the midgut (8).When hemocytes encounter the nitrated basal lamina, they release hemocyte-derived microvesicles (HdMv's) that promote complement activation mediated by the thioester-containing protein 1 (TEP1).TEP1 binds to the ookinete surface, forming a complex that ultimately lyses the parasite (3,9).Interestingly, although TEP1 eliminates P. berghei efficiently in A. gambiae, P. falciparum ookinetes have a powerful mechanism to evade TEP-1-mediated killing in this mosquito vector (5,10,11).
A series of studies using a combination of genetic mapping, linkage group selection, and functional genomics identified Pfs47 as the gene that allows P. falciparum to evade elimination by the mosquito complement-like system (10).Pfs47 is a cell surface protein that disables the mosquito complement system by suppressing the midgut nitration response following ookinete invasion (10).Mosquito midgut cells invaded by wild-type (WT) P. falciparum (NF-54) ookinetes expressing a Pfs47 haplotype compatible with the vector fail to activate JNK signaling, elicit weak caspase activity, and do not mount an effective nitration response (5).In contrast, invasion by Pfs47 knockout (KO) parasites elicits strong caspase activity, followed by induction of HPX2 and NOX5, and robust epithelial nitration resulting in a strong TEP1 activation (5,10), similar to the response to P. berghei infection (3,4).
P. falciparum isolates from different geographic areas vary in their capacity to infect different anopheline species (12,13).In general, parasite isolates are more compatible with mosquitoes from the same geographic regions (13,14), suggesting adaptation by a genetic selection of parasites by local vectors (15).We have shown that amino acid polymorphisms in the central domain of Pfs47 are major determinants of P. falciparum compatibility with different anopheline species (16).The lock and key model was proposed, in which Pfs47 is thought to be the "key" that allows the parasite to "turn off" the mosquito detection system by interacting with a specific mosquito receptor ("the lock") (14).Only those parasites with a Pfs47 haplotype compatible with the midgut receptor can evade the mosquito immune system and be transmitted.The mosquito Pfs47 receptor (P47Rec) was identified based on recombinant Pfs47 (rPfs47) interaction with midgut extracts in far-Western blots (11).The P47Rec localizes immediately below the microvilli (submicrovillar region) of midgut cells and exhibits a reticular distribution pattern, suggesting that it may interact with the cytoskeleton (11).Recombinant P47Rec binds with high affinity to Pfs47, and silencing P47Rec reduced P. falciparum infection, indicating that the interaction of Pfs47 with P47Rec is critical for parasite survival (11).Notably, the P47Rec from a given anopheline species binds with higher affinity to Pfs47 from sympatric parasites, supporting the hypothesis that Pfs47 is a key mediator of parasite adaptation to compatible anophelines (11,14).The P47Rec is a novel protein whose function and molecular interactions are not known.Here, we show how the lack of Pfs47 expression affects the response of midgut epithelial cells to P. falciparum ookinete invasion.We also provide direct evidence that P47Rec interacts with the heat shock protein 70 cognate 3 (Hsc70-3) and that this interaction affects caspase-mediated apoptosis in invaded cells and the survival of P. falciparum ookinetes.

Pfs47-knockout P. falciparum ookinetes are lysed in the mosquito midgut
To gain new insights on how Pfs47 affects midgut invasion, we compared the dynam ics of Anopheles gambiae midgut invasion by WT and Pfs47KO P. falciparum ookinetes using immunofluorescence (IFA) microscopy.Mosquito midguts were dissected every 2 h between 22 and 32 h post-feeding (PF) to establish whether there were differences in the timing of invasion.The timing of midgut invasion was similar for both parasites (Fig. S1), with few invasion events between 22 and 24 h PF, followed by a substantial increase between 26 and 28 h PF (Fig. S1).Based on these observations, all subsequent cell biology experiments were analyzed in midguts dissected at 28 h PF.At 28 h PF, the proportion of fragmented ookinetes was significantly higher in Pfs47KO (69%) than in WT parasites (13%) (P < 0.0001, Mann-Whitney) (Fig. 1a and b), confirming that Pfs47 expression is critical for ookinetes to evade lysis by the mosquito immune system.

Caspase-mediated apoptosis in response to invasion by WT and Pfs47KO ookinetes
We confirmed that epithelial caspase activity was significantly higher (fourfold) in midguts invaded by Pfs47KO than by WT ookinetes (P < 0.01, t-test) (Fig. 1c), as previously reported (5), and in agreement with higher levels of midgut nitration (10).To better understand how Pfs47 affects caspase activation in invaded cells, we performed IFAs using an antibody to detect enzymatically active caspase.The proportion of ookineteinvaded cells protruding toward the midgut lumen that were caspase positive was significantly higher in midguts infected with Pfs47KO ookinetes (80%) than with WT (35%) (P < 0.0001, Mann-Whitney U test) (Fig. 1d; Table S1), and caspase-mediated cell death was restricted to cells that have been invaded by the parasite (Fig. 1e).Further more, we also evaluated DNA fragmentation in parasite-invaded cells by terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) assay (Fig. 1f; Table S2).The proportion of ookinete-invaded midgut cells with DNA fragmentation was also signifi cantly higher (75%) for Pfs47KO than for WT (40%) parasites (P < 0.0001, Mann-Whitney) (Fig. 1f and g).

Pfs47 accelerates the extrusion of ookinete-invaded epithelial cells and decreases microvesicle release
Ookinete midgut traversal causes irreversible damage to the invaded epithelial cells that results in cell death.Damaged cells detach gradually from the epithelium and ultimately bud off into the midgut lumen (Fig. 2a, left panel).Neighboring healthy cells extend pseudopodia-like projections and stretch their cytoplasm to gradually close the space left by the extruding cell (17).When the damaged cell detaches, these extensions leave a flowerlike (rosette) structure in the epithelium with the ookinete that traversed the epithelium at the center (Fig. 2a, right panel).Although the time of onset of midgut invasion was similar between WT and Pfs47KO parasites (Fig. S1), the proportion of ookinete-associated rosettes at 28 h PF was significantly higher for the WT (77%) than for Pfs47KO (61%) parasites (P < 0.0001, Mann-Whitney) (Fig. 2b; Table S3), indicating that damaged cells are extruded faster when they are invaded by parasites that express Pfs47.Presumably, this faster extrusion restricts the time that damaged cells have to activate caspases and nitrate the basal lamina.This, in turn, would reduce the release of HdMv's, an event necessary for effective activation of the mosquito complement-like system (18).
We evaluated the release of HdMv after infection of WT and Pfs47KO ookinetes by IFA, as previously described by Castillo et al. (18).The proportion of extruding cells positive for HdMv was significantly higher for Pfs47KO (42%) than for WT (11%) parasites (Fig. 2c and  d; P < 0.0001, Fisher's exact test).

Identification of molecular partners of P47Rec
The interaction of recombinant P47Rec with homogenate of midgut proteins was explored using far-Western blot analysis to uncover potential partners.Previous studies showed that P47Rec is present in the insoluble pellet fraction of mosquito midgut homogenates, which includes cell membrane and cytoskeleton-associated proteins (11).The pellet fraction was incubated with a buffer containing 0.1% Triton X-100 to separate soluble and insoluble fractions and further reduce the complexity of the sample.Western blots revealed that most of the P47Rec protein remained in the Triton X-100 insoluble (I) fraction (Fig. 3a).This fraction was used to identify potential proteins interacting with recombinant P47Rec.
Proteins in the Triton X-100 insoluble fraction were subjected to SDS electrophoresis under denaturing/non-reducing conditions (samples were heated to 70°C after the addition of SDS loading buffer) and transferred to a PVDF membrane.The membrane was then probed with recombinant His-tagged P47Rec protein that was detected with  anti-His antibodies.P47Rec interacted strongly with a major protein of approximately 70 kDa and weakly with a protein of higher molecular weight, 140 kDa (Fig. 3b).To establish the identity of the protein(s) interacting with P47Rec, the Triton X-100 insoluble fraction was subjected to two-dimensional (2D) gel electrophoresis, transferred to a PVDF membrane, and probed with recombinant P47Rec.
We were surprised to find that the P47Rec interacted with seven major protein spots, all migrating at approximately 70 kDa but with different isoelectric points (pIs), ranging from 4.0 to 9.0.Another group of weaker spots with a molecular weight of approximately 75 kDa and with different pIs was also detected (Fig. 3c, left panel).The observed differences in pIs could be explained by post-translational modifications of the same polypeptide, such as protein phosphorylation or ADP ribosylation.Six spots were excised from the polyacrylamide gel and subjected to mass-spectrometry (Fig. 3c, right panel and file S1).Based on bioinformatic analysis of the detected peptides, we restricted our analysis to six proteins with molecular weights from 65 to 75 kDa.These were detected in the Spot 1 (1 protein), Spot 2 (5 proteins), and Spot 5 (2 proteins) (Fig. 3c), whereas in the spots 3, 4 and 6 only proteins with discrepant molecular weights were detected.Notably, the proteins with the highest number of peptides were three members of the HSP70 family, AGAP004192 (nine unique peptides; 72,743 kDa), AGAP002076 (seven unique peptides; 71,376 kDa) and AGAP004944 (four unique peptides; 70,592 kDa).The mature protein AGAP004192 has a predicted isoelectric point of 5.09, corresponding to the spot that was excised.This protein is the ortholog of Drosophila Hsc70-3.

A. gambiae Hsc70-3 interacts with P47Rec
We prioritized Hsc70-3 because we have previously shown that reducing the expres sion of this gene by double-stranded RNA (dsRNA)-mediated silencing significantly reduced P. falciparum infection (19).Hsc70-3 is a molecular chaperone of the heat-shock protein family with potent anti-apoptotic activity (20,21).To establish whether Hsc70-3 mediated the interaction of recombinant P47Rec with the immobilized midgut extract, we knocked down Hsc70-3 expression by systemic injection of Hsc70-3 dsRNA.When Hsc70-3 mRNA expression was reduced by 88% (Fig. 3d), the signals for the major 70-kDa band and the weaker band around 140 kDa were no longer detectable (Fig. 3e), suggesting that Hsc70-3 is the main protein in the midgut extract that interacts with recombinant P47Rec.Furthermore, we expressed recombinant Hsc70-3 protein in Escherichia coli and confirmed that recombinant Hsc70-3 binds to recombinant P47Rec immobilized in a membrane after gel electrophoresis (Fig. 3f).The interaction of these proteins was further confirmed by reciprocal enzyme-linked immunosorbent assay (ELISA) (Fig. 3g and h) in which serial dilutions of Hsc70-3 or P47R were used to address reciprocal binding to the immobilized counterpart.The binding was saturable, and the estimated IC 50 for Hsc70-3 and P47R were 1.23 and 0.77 µM, respectively (Fig. 3g and h).These data indicate that the interaction between Hsc70-3 and P47R is specific.

A. gambiae Hsc70-3 mediates inhibition of apoptosis by P. falciparum ookinete invasion
Silencing of Hsc70-3 significantly increased the number of fragmented ookinetes in the midgut (Fig. 4a) and, in agreement with our previous report (19), it dramatically reduced the number of P. falciparum oocysts (Fig. 4b).The orthologs of Hsc70-3 in Drosophila, also known as binding immunoglobulin protein (BiP) in mammals, are strong inhibitors of apoptosis (20).The effect of silencing of Hsc70-3 on caspase-mediated apoptosis in P. falciparum-infected midguts was investigated.Reducing Hsc70-3 expression resulted in a 2.8-fold increase in caspase activity (Fig. 4c) in response to ookinete midgut invasion, indicating that Hsc70-3 inhibits apoptosis in A. gambiae midguts infected with P. falciparum.Caspase S2 (CASP-S2) regulates expression of HPX2 and NOX5, two key enzymes that potentiate midgut nitration in response to ookinete invasion, and silencing CASP-S2 enhances survival of Pfs47KO P. falciparum parasites (5).We investiga ted whether the decrease in parasite survival following Hsc70-3 silencing is mediated by CASP-S2.Silencing Casp-S2 alone had no effect on P. falciparum WT oocyst (Fig. 4d), as previously reported (5) (Fig. 1c and d).However, co-silencing Hsc70-3 and CASP-S2 reverts the decrease in oocyst numbers observed when Hsc70-3 alone is silenced (Fig. 4d).This indicates that Hsc70-3 is upstream of CASP-S2 and that, in the absence of Hsc70-3, CASP-S2 is an important effector of anti-plasmodial immunity.We propose that P. falciparum Pfs47 interacts with the mosquito Pf47Rec, which inhibits apoptosis by interacting with Hsc70-3, a suppressor of caspase activation.

DISCUSSION
Previous studies have shown that Pfs47 inhibits midgut epithelial nitration by interacting with the P47Rec, a response critical for efficient activation of the mosquito complementlike system (11).Here, we uncovered how P. falciparum Pfs47 disrupts the mosquito antiplasmodial response by accelerating the extrusion of cells damaged by ookinete invasion.Enterocytes are extruded from the epithelium through the formation of a basal contractile actin ring that maintains epithelial integrity by bringing together neighboring cells, giving rise to a rosette-like structure.Epithelial invasion by Pfs47KO parasites was associated with damaged cells undergoing caspase-mediated apoptosis that often also exhibited DNA fragmentation (Fig. 1d and g).Pfs47KO ookinetes triggered HdMv release (Fig. 2c and d) and were lysed (Fig. 1a and b) as they encountered the mosquito hemo lymph, resembling the response to infection with P. berghei parasites (10).In contrast, cells invaded by WT ookinetes are seldom captured in the process of being extruded from the epithelium (Fig. 2a and b), and they are often negative for caspase expression and for DNA fragmentation (Fig. 1c and d).WT parasites were mostly associated with "rosette" structures left behind by cells that have already been extruded.These structures preserve the integrity of the midgut barrier by bringing together healthy neighboring cells to fill the space left as the damaged cell buds off into the midgut lumen (Fig. 1b).
Collectively, these results suggest that Pfs47 limits caspase activation, epithelial nitration and HdMv release, at least in part, by accelerating the extrusion of damaged cells.In Drosophila, both caspase-dependent and caspase-independent epithelial cell extrusions have been documented, under the regulation of different signaling pathways (22).In zebrafish, UV light damage triggers caspase-dependent cell extrusion regulated by JNK signaling (22), similar to the response of midgut epithelial cells invaded by P. berghei or Pfs47KO P. falciparum ookinetes, in which epithelial nitration involves activation of the JNK pathway (5,23).
A specific interaction of P47Rec with the chaperone protein Hsc70-3 was identi fied using a combination of far-Western blots, ELISA, and dsRNA silencing (Fig. 3).In mammals, BiP can undergo several different post-translational modifications, includ ing sulfenylation, glutathionylation, ADP-ribosylation, phosphorylation, AMPylation, and citrullination that regulate its biological activity, turnover, and protein availability (24)(25)(26).Thus, post-translational modifications of Hsc70-3 may be responsible for the multiple 70-kDa spots with different pIs observed in 2D gels of immobilized mosquito midgut extracts.BiP can localize at the cell surface, where it interacts with proteins that control different cell signaling pathways, including the α2-macroglobulin-induced signal transduction (24,27), and the cytosolic isoform of BiP has been shown to improve cell survival during ER stress (28).
The Hsc70-3 orthologs in Drosophila and mammals are both strong inhibitors of apoptosis (24).We found that silencing of Hsc70-3 increased midgut caspase activity in invaded cells and decreased parasite survival.Furthermore, co-silencing of Hsc70-3 and the effector caspase CASP-S2, whose activity mediates the killing of Pfs47KO ookinetes (5), rescued the survival of WT oocysts, indicating that Hsc70-3 favors parasite survival by inhibiting caspase activation.One cannot discard the possibility that P47Rec may also engage other signal transduction pathways that trigger a swift cell death response so that the invaded cell does not have enough time to activate an effective nitration response.We propose a working model in which Pfs47, on the surface of ookinetes, interacts with the mosquito P47Rec during midgut invasion and accelerates the extrusion of the invaded cell.The interaction of P47Rec with Hsc70-3 inhibits midgut epithelial nitration by suppressing CASP-S2 activation, and thus prevents the release of HdMv, which is necessary for activation of the mosquito complement-like system.We conclude that P. falciparum parasites expressing Pfs47 activate a Hsc70-3-mediated anti-apoptotic response and accelerate the dynamics of cell death and extrusion of invaded enterocytes, favoring Plasmodium evasion of early immune responses that target the ookinete stage.

Midgut infections and ookinete counting
Female mosquitoes were infected by membrane feeding with reconstituted human blood, consisting of human RBC mixed with human serum (60:40).Mature gametocyte cultures 14-16 days old were adjusted for 0.25% gametocytemia and used for feeding.Midguts were dissected 10 days after feeding, and oocysts were stained with 0.2% mercurochrome in phosphatebuffered saline (PBS) and counted by light microscopy.Median oocyst numbers were compared using the Mann-Whitney U test.All assays were confirmed with two or three independent experiments.

Immunofluorescence microscopy
After infection with P. falciparum, midguts were dissected at different time points ranging from 22 to 32 h, as indicated in the figures.The blood bolus was carefully removed by longitudinally opening the midguts on ice-cold PBS.After fixation with 4% paraformalde hyde in PBS for 1 h at RT, the midguts were washed two times with 0.1% Triton X-100 (Sigma-Aldrich) in PBS (PBST), followed by blocking for at least 1 h with 5% bovine serum albumin (BSA) (Sigma-Aldrich), 0.1% gelatin (Sigma-Aldrich) in PBST.

Caspase activation
After the first blocking, the midguts were incubated with rabbit anti-DCP-1 (1:250) and anti-Pfs25, followed by counterstaining with goat anti-mouse Alexafluor 488 conjugated antibody (1:1,000; Thermo Fisher Scientific) and goat anti-rabbit Alexafluor 594 (1:1,000; Thermo Fisher Scientific).For the TUNEL assay, after washing off the secondary antibody, the samples were incubated with the TUNEL reagent (TMR red, Roche/Sigma) for 30 min at 37°C.As described above, the samples were rinsed with PBST and counterstained with Hoechst/phalloidin.

Microvesicle staining
To stain hemocytes, female mosquitoes were injected with 69 nL of 100-µM Vybrant CM-DiI cell labeling solution (Thermo Fisher Scientific) 2 days prior to P. falciparum infection.The midguts were processed as described for ookinete staining.Images were obtained using a Leica TCS SP8 confocal microscope (Leica Microsystems) ×63 oil immersion objective.Images were taken using sequential acquisition and variable z-steps.Images were processed using Imaris v.9.0.0 (Bitplane AG) and Adobe Photoshop CC (Adobe Systems).

Recombinant proteins
The Hsc70-3 (AGAP004192) was PCR amplified from a P. falciparum-infected midgut complementary DNA (cDNA) library, and the amplicon was cloned into a pET17 vector (EMD Millipore) between NdeI and XhoI sites by In-Fusion (Clontech), with primers containing an N-terminus his.tag excluding the signal peptide (F′-AAGGAGAT ATA CAT ATG CAT CAC CAT CAC CAT CAC gag gaa aag aag gaa aag gac a and R′-G CCGGATCTGCTCGA TTATTA CAGCTCATCCTTCAGCTCG). Pf47R recombinant protein was produced as described before (11).The proteins were expressed in Rosetta (DE3) E. coli (EMD Millipore).Cultures were induced with 1-mM isopropyl β-D-thiogalactopyranoside (Sigma-Aldrich), at 37°C overnight, 250 rpm, and the protein in the soluble fraction was purified by nickel affinity chromatography.After purification and dialysis, the proteins were stored in PBS and 20% glycerol.

Western and far-Western blots
SDS-PAGE was performed under mild denaturing/non-reducing conditions.Cytoplasm and membrane cytoskeletons (five midgut equivalent) were mixed with NuPAGE LDS sample buffer (Thermo Fisher Scientific) and heated to 70°C after adding SDS loading buffer.Proteins were separated on a 4%-12% NuPAGE Bis-Tris gel (Invitrogen) and transferred to PVDF membranes using the iBlot blotting system (Invitrogen).The gels were stained with GelCode Blue Stain Reagent (Thermo Fisher Scientific).After blotting, the PVDF membrane was blocked with 5% non-fat dry milk in PBS, 0.1% Tween 20 (PBST), for 1 h, at room temperature (RT).Membranes were washed with PBST and incubated with a rabbit IgG purified anti-6398 (recP47R) at 2 µg/mL, for 1 h at RT, then washed with PBST, and incubated with an anti-rabbit HRP (Thermo Fisher Scientific).Subsequently, the bands were developed with Pico or Dura substrate (Thermo Fisher Scientific) and exposed in Kodak Film.For the one-dimensional far-Westerns, after blocking, the blots were incubated with recPf47R (150 nM) for 1 h at RT or with no recombinant protein as a control.The blots were washed and incubated with anti-his tag (1:1,000; Sigma-Aldrich) and mouse HRP (1:25,000, Thermo Fisher Scientific).For the 2D far-Westerns, the insoluble membrane fraction was prepared as described above.Protein pellet equivalent to 300 midguts was separated in 3-10 pH using a Novex IEF gel system (Thermo Fisher Scientific).After blotting, the membrane was probed with recP47R, as described above.The spots were aligned with a Coomassie-stained 2D gel, and the bands were excised for mass spectrometry analysis.For far-Western blot analysis after gene silencing, dsRNA was injected as described below: briefly, 48 h after injection of dsRNA for Hsc70-3 (dsHsc-3) or dsRNA for lacZ (dsLacZ), mosquitoes were fed in human serum, and the midguts dissected 28 h after feeding as described above.

dsRNA-mediated gene knockdown
As previously described, the knockdown of Hsc70-3 and CaspS2 was performed by injection of specific double-stranded RNA (5,19).In short, 3-to 4-day-old female A. gambiae were cold-anesthetized and injected with 69 nL of a 3 µg/µL dsRNA solution specific for each gene.dsRNA was generated from a cDNA template from A. gambiae and the MEGAscript RNAi kit (Thermo Fisher Scientific).Primers were designed specifically for targeted genes and are listed as follows: Hsc70-3 was amplified again using internal primers containing T7 promoters (F′-taatacgactcactatagggGCCTTTCATCCACTCCATCT and R′-taatacgactcactatagggCTTGGGTGGTGTATGTGTGTG).For CaspS2 (F′-taatacgact cactatagggCGTGAGCAAAGAGGATCACA and R′-taatacgactcactatagggACACGTGAGTCAG CAAGGTG)

RNA extraction, cDNA synthesis, and qPCR analysis
RNA samples from midguts were obtained using TRIzol (Thermo Fisher Scientific) according to manufacturer instructions.One microgram of total RNA was used for cDNA synthesis using Quantitect Reverse Transcription Kit (Qiagen) according to the manufac turer's instructions.Gene expression was assessed by quantitative PCR (qPCR) using the resulting cDNA as a template.We used the DyNamo SYBR green qPCR kit (Thermo Fisher Scientific) with the following targetspecific primers: CaspS2 (F′-CGACAAAACAAAGCAC GAGA, R′-TGCCATCGCCACAATTACTA) and Hsc70-3 (F'-TGAGCTGTAAGTGCCATCGG, and R'-AGTTGGGCTGTGAAGTACCG); the assay was run on a CFX96 Real-Time PCR Detection System (Bio-Rad, Hercules, CA, USA).Relative quantitation results were normalized against A. gambiae ribosomal protein S7 (RpS7) as an internal standard and were analyzed using the ΔΔ Ct method (Livak and Schmittgen, 2001, Pfaffl, 2001).Statistical differences in fold change values between the conditions were determined using the Mann-Whitney test (GraphPad, San Diego, CA, USA).Each independent experiment was done with three biological replicates (three pools of 15 tissues) for each condition tested, and each one was analyzed by qPCR in duplicates (technical replicates).At least three independent experiments were performed.

ELISA
To test the interaction of Hsc70-3 and recP47R, we performed ELISA, where one protein was immobilized and the other was used as bait.The recombinant proteins (50 µL at 2 µg/mL) were immobilized in a 96-well ELISA plate (Immunolon microtiter 96-Well plates, Thermo Fisher Scientific VWR) in duplicate in carbonate buffer (15-mM Na 2 CO 3 , 35-mM NaHCO 3 , pH 9.6), overnight at 4°C.The plates were washed 10 times with TBST and blocked with 5% milk in 0.1% Triton X-100 in TBS (TBST) for 1 h at RT. Plates were incubated with 100 µL of increasing concentrations of the recombinant proteins diluted in blocking buffer for 1 h at RT.After 10 washes with TBST, the plates were incubated with 1 µg/mL of anti-DELK antibody (Thermo Fisher Scientific) or anti-FLAG antibody (Sigma-Aldrich) in blocking buffer for 1 h at RT.The plates were then washed 10 times with TBST and incubated with 100 µL of goat anti-mouse IgG conjugated to HRP (1:15,000; Thermo Fisher Scientific) secondary antibodies for 1 h at 37°C.The plates were washed 10 times, and the detection was performed using TMB Ultra (Thermo Fisher Scientific).After 30 min, an equal volume of sulfuric acid (0.16 M) was added, and the absorbance was read at 450 nm using the Cytation 5 plate reader (Biotek).

Caspase assay
Caspase activity was measured using the caspase-3 kit assay (Biovision), as described in Ramphul et al. (5).Midguts infected with the WT or Pf47KO were dissected 28 h PF, and a pool of five midguts was macerated in 50 µL of the kit provided lysis buffer.After centrifugation at 10,000 g for 10 min at 4°C, 50 µL of the supernatant was added to 50 µL of fresh lysis buffer.Twentyfive microliters of the diluted supernatant was added to the reaction buffer containing the fluorometric caspase-3 substrate DEVD_AFC, followed by incubation in the dark for 2 h at 37°C.Fluorescence readings were taken at 400-nm excitation, and 505-nm emission, and the control readings were subtracted for background.Alternatively, midguts from mosquitoes injected with dsLacZ or dsHsc-3 28 h PF were used for the reaction.

FIG 1
FIG 1 The role of Pfs47 in P. falciparum ookinete integrity and midgut cell death activation.(a) P. falciparum ookinetes stained for Pfs25 (green) and DNA (Hoechst), blue.Scale bar: 2 µm.(b) Comparison of the percentage of fragmented WT and Pfs47KO ookinetes 28 h PF in individual midguts (Mann-Whitney U test).****P < 0.0001.(c) Caspase activity of midgut pools 28 h PF with WT or Pfs47KO.The activity of the midgut invaded with the WT was set as reference.The result is an average of four independent experiments (t-test).**P < 0.01.(d) Effect of WT or Pfs47KO invasion on caspase activation.Cells were stained with anti-active caspase antibody, and the percentage of invaded cells positive for caspase on individual midguts was calculated, including midguts with three or more invasion events (see TableS1).Horizontal blue line indicates medians that were compared using the Mann-Whitney U test.****P < 0.0001.(e) Detection of enzymatically active caspase in midguts infected with WT (left) or Pfs47KO (right).Ookinete, green; actin, cyan; caspase, red; nucleus, blue.(f ) terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) assay of an infected midgut 28 h PF. red, TUNEL; green, ookinete; blue, nucleus.(g)Quantification of the proportion of TUNEL-positive cells associated with an ookinete on individual midguts (see TableS2).The line indicates medians that were compared using the Mann-Whitney U test.**P < 0.01, ****P < 0.0001.Horizontal blue lines represent the median.The number of individual midguts shown in panels b, d, and g are indicated as n.

FIG 3
FIG 3 Characterization of the interaction between P47R and Hsc70-3 in midgut membrane protein preparations.(a) Coomassie stained gel (left) and Western blot of P47R in midgut membrane extracts (right).(Left) Coomassie-stained gels.(b) Binding of P47Rec to cytosolic (c) and Triton X-100 insoluble fraction (I).Midgut samples were collected 28 h post-feeding.The protein equivalent of five midgut extracts was resolved in the gel.(Left) Coomassie-stained gel; (center) far-Western blot; (right) control blot, where no bait was added.(c) Binding of P47Rec to midgut membrane insoluble proteins in a two-dimensional Western blot.Excised spots analyzed by mass spectrometry are indicated by the red rectangle in the figure.The blue rectangle corresponds to the region excised as negative control.The major proteins found are summarized in the in the right panel.Complete information on mass spectrometry results is presented in File S1.(d) quantitative PCR quantification of Hsc70-3 mRNA in midguts after injection of Hsc70-3 dsRNA (dsHsc70-3) or dsRNA for lacZ (dsLacZ).(e) Effect of dsRNA-mediated silencing of the Hsc70-3 in the binding of P47Rec in a far-Western blot of midguts 28 h PF.Control mosquitoes were injected with dsLacZ (LcZ); (right) loading control: after stripping, the membrane was re-probed with the anti-actin horseradish peroxidase (HRP) antibody.(f ) Recombinant P47Rec (500 ng) was subjected to SDS-PAGE and blotted to a polyvinylidene fluoride (PVDF) membrane for far-Western blot analysis.Recombinant Hsc70-3 was used as bait at 150 nM in 5% milk diluted in 0.1% Triton X-100 in PBS.Hsc70-3 binding was detected with an anti-KDEL antibody.(g) Enzyme-linked immunosorbent assay (ELISA) binding of recombinant P47Rec at increasing concentrations to immobilized recombinant Hsc-70-3.(h) ELISA binding of recombinant Hsc-70-3 at increasing concentrations to immobilized recombinant P47Rec.All Western and far-Western blots were detected by chemiluminescence.C, cytosolic fraction; I, Triton X-100 insoluble fraction; MW, molecular weight ladder; S, Triton X-100 soluble fraction.

7 FIG 4
FIG 4 Suppression of apoptosis mediated by Hsc-3 and the impact on midgut infection.(a) Comparison of the percentage of fragmented ookinetes 28 h PF in midguts dissected from mosquitoes injected with dsLacZ or dsHsc70-3.Medians were compared using the Mann-Whitney U test.***P < 0.001.(b) Comparison of oocyst numbers 10 days PF in mosquitoes injected with dsLacZ or dsHsc70-3.Medians were compared using the Mann-Whitney U test.****P < 0.0001.(c) Caspase activity of midgut pools 28 h PF with dsLacZ or dsHsc70-3.The activity of the infected midgut in the LacZ condition was set as reference.The result is an average of three independent experiments (t-test).*P < 0.01.(d) Comparison of oocyst numbers 10 days PF in mosquitoes injected with dsLacZ/dsLacZ, dsHsc70-3/dsLacZ, dsCasS2/dsLacZ, or dsCasS2/dsHsc70-3.The line indicates medians that were compared using the Mann-Whitney U test.**P < 0.01, ****P < 0.0001.